Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
PETROLEUM
TECHNOLOGY
VOLUME 1-2
WILEY-INTERSCIENCE
WILEY-VCH
A JOHN WILEY & SONS, INC., PUBLICATION
Copyright # 2007 by John Wiley & Sons, Inc., and WILEY-VCH Verlag GmbH & Co. KGaA,
Weinheim. All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any
form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except
as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the
prior written permission of the Publisher, or authorization through payment of the appropriate
per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923,
(978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com. Requests to the Publisher
for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc.,
111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online
at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts
in preparing this book, they make no representations or warranties with respect to the accuracy or
completeness of the contents of this book and specifically disclaim any implied warranties of
merchantability or fitness for a particular purpose. No warranty may be created or extended by sales
representatives or written sales materials. The advice and strategies contained herein may not be
suitable for your situation. You should consult with a professional where appropriate. Neither the
publisher nor author shall be liable for any loss of profit or any other commercial damages, including
but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our
Customer Care Department within the United States at (800) 762-2974, outside the United States at
(317) 572-3993 or fax (317) 572-4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print
may not be available in electronic formats. For more information about Wiley products, visit our web
site at www.wiley.com.
10 9 8 7 6 5 4 3 2 1
CONTENTS VOLUME 1
v
CONTENTS VOLUME 2
vi
PREFACE
Wiley’s major reference works cover a multitude of subject disciplines and have
earned many accolades over the years for their well-structured, accessible arti-
cles, covering both fundamentals and cutting-edge topics. The Wiley Critical
Content Series was conceived in response to demand from customers who
have a strong interest in certain key topics covered within more than one of
Wiley’s reference works and who have asked to see the most relevant articles
brought together in one place. Wiley Critical Content: Petroleum Technology
is the first title to be published in this series. It features articles from two
of Wiley’s most prestigious encyclopedias—the Kirk-Othmer Encyclopedia of
Chemical Technology and Ullmann’s Encyclopedia of Industrial Chemistry.
Wiley Critical Content: Petroleum Technology provides comprehensive, up-
to-date information on all aspects of petroleum: properties and origin, explora-
tion, production and refining processes, economic issues, and environmental
and health concerns. The two volumes contain over fifty articles, organized
thematically:
Written by prominent scholars from around the world, this reliable reference
will be of particular interest to petroleum engineers, researchers in industry
and academia, and other professionals and consultants in petroleum-related
industries.
vii
CONTRIBUTORS
Colin Barker, University of Tulsa, Tulsa, M. Larry Campbell, Exxon Chemical Company,
Oklahoma, United States, Petroleum, Florham Park, New Jersey, United States,
Introduction Cyclohexane
Scott D. Barnicki, Eastman Chemical Martin Carrera, Amoco Corporation,
Company, Kingsport, Tennessee, United Naperville, Illinois, United States, Propylene
States, Separations Process Synthesis Clarence D. Chang, Mobil Research and
Calvin Bartholomew, Brigham Young Development Corporation, Princeton,
University, Provo, Utah, United States, New Jersey, United States, Fuels, Synthetic,
Catalyst Deactivation and Regeneration Liquid
Jürgen Bässler, Uhde GmbH, Dortmund, W. C. Cheng, W. R. Grace & Company,
Germany, Acetylene Columbia, Maryland, United States, Fluid
Arno Behr, Henkel KGaA, Düsseldorf, Catalytic Cracking, Catalysts and Additives
Germany, Hydrocarbons T. T. Peter Cheung, Phillips Petroleum
Hartmut Behringer, Hoechst Company, Bartlesville, Oklahoma, United
Aktiengesellschaft, Werk Knapsack, States, Cyclopentadiene and
Germany, Acetylene Dicyclopentadiene
L. E. Bell, UOP LLC, Des Plaines, Illinois, P. D. Clark, University of Calgary, Calgary,
United States, Fluid Catalytic Cracking, Units, Alberta, Canada, Sulfur and Hydrogen Sulfide
Regeneration Recovery
Rosemarie Berger, Ludwigshafen, Germany, Gerd Collin, DECHEMA e.V., Frankfurt am
Liquid–Liquid Extraction Main, Germany, Anthracene; Hydrocarbons;
Ronald L. Berglund, Terracon, Kingwood, Naphthalene and Hydronaphthalenes
Texas, United States, Emission Control, Axel Commichau, Mobil Europe Gas Inc., The
Industrial Hague, The Netherlands, Natural Gas
Dieter Biedenkapp, BASF Aktiengesellschaft, Werner Dabelstein, Deutsche Shell AG,
Ludwigshafen, Germany, Hydrocarbons Hamburg, Germany, Octane Enhancers
Geoff J. Bishop, Shell Aviation Ltd., London, Rula Deeb, Malcolm Pirnie, Inc., Emeryville,
United Kingdom, Aviation Turbine Fuels California, United States, Methyl Tert-Butyl
Eckhart Blass, Technische Universität Ether
München, München, Germany, Liquid–Liquid James R. Dickey, Lubricants Consultant,
Extraction Basking Ridge, New Jersey, United States,
Klaus Boehlke, BASF Aktiengesellschaft, Lubrication and Lubricants
Ludwigshafen, Germany, Waxes Wolfram Dietsche, BASF Aktiengesellschaft,
John T. Boepple, Nexant ChemSystems, White Ludwigshafen, Germany, Waxes
Plains, New York, United States, Michael F. Doherty, University of California,
Petrochemical Feedstocks Santa Barbara, Santa Barbara, California,
John K. Borchardt, Shell Development United States, Distillation, Azeotropic and
Company, Houston, Texas, United States, Extractive
Enhanced Oil Recovery T. J. Dougan, W. R. Grace & Company,
James A. Brient, Merichem Company, Columbia, Maryland, United States, Fluid
Houston, Texas, United States, Naphthenic Catalytic Cracking, Catalysts and Additives
Acids Mary Noon Doyle, Shepherd Chemical
P. F. Bryan, Chevron Research and Technology Company, Cincinnati, Ohio, United States,
Company, Richmond, California, United Naphthenic Acids
States, BTX Processing Lawrence J. Drew, U.S. Geological Survey,
Klaus Buckl, Linde AG, Höllriegelskreuth, Reston, Virginia, United States, Petroleum,
Germany, Acetylene Introduction
Narasimhan Calamur, Amoco Corporation, Günter Ebersberg, Degussa-Hüls AG, Marl,
Naperville, Illinois, United States, Propylene Germany, Acetylene
ix
x CONTRIBUTORS Vol. 1
Jörg Fabri, Deutsche Shell AG, Hamburg, Gerd Hohner, Hoechst Aktiengesellschaft, Werk
Germany, Octane Enhancers; Xylenes Gersthofen, Augsburg, Germany, Waxes
James R. Fair, The University of Texas at Hartmut Höke, Weinheim, Germany,
Austin, Austin, Texas, United States, Anthracene; Hydrocarbons; Naphthalene
Distillation and Hydronaphthalenes
Kevin J. Fallon, Washington Group Craig A. Hoyme, Eastman Chemical Company,
International, Inc., Cambridge, Kingsport, Tennessee, United States,
Massachusetts, United States, Ethylbenzene Separations Process Synthesis
William Fruscella, Unocal Corporation, El Chang Samuel Hsu, Exxon Research and
Segundo, California, United States, Benzene Engineering Company, Annandale, New
Dorothea Garbe, Haarmann & Reimer GmbH, Jersey, United States, Petroleum, Introduction
Holzminden, Germany, Hydrocarbons Walther W. Irion, Deutsche Shell AG,
Bruce C. Gates, University of California, Hamburg, Germany, Oil Refining,
Davis, Davis, California, United States, Environmental Considerations
Catalysis Michael Kavanaugh, Malcolm Pirnie, Inc.,
Eugen Gehrer, BASF Aktiengesellschaft, Emeryville, California, United States, Methyl
Ludwigshafen, Germany, Acylation and Tert-Butyl Ether
Alkylation Roland Kettner, Mobil Erdgas–Erdöl GmbH,
Heinz-Peter Gelbke, BASF Aktiengesellschaft, Celle, Germany, Natural Gas
Ludwigshafen, Germany, Ethylbenzene Jeffrey P. Knapp, E. I. du Pont de Nemours
Ulrich Graeser, Veba Öl AG, Gelsenkirchen, & Company, Wilmington, Delaware, United
Germany, Xylenes States, Distillation, Azeotropic and Extractive
Helmut Greim, Institut für Toxikologie und Christian Küchen, Institut für Wirtschaftliche
Umwelthygiene, Technische Universität Oelheizung, Hamburg, Germany, Heating Oil
München, Freising-Weihenstephan, Germany, G. R. Lappin, Albemarle Corporation, Baton
Naphthalene and Hydronaphthalenes Rouge, Louisiana, United States, Olefins,
Karl Griesbaum, Universität Karlsruhe (TH), Higher
Karlsruhe, Germany, Hydrocarbons Maryline Laugier, Malcolm Pirnie, Inc.,
Baruch Grinbaum, IMI (TAMI) Institute for Emeryville, California, United States, Methyl
Research and Development, DSBG, Haifa Bay, Tert-Butyl Ether
Israel, Drilling Fluids George M. Long, Institute of Gas Technology,
Georg Hammer, Mobil Erdgas–Erdöl GmbH, Chicago, Illinois, United States, Pipelines
Celle, Germany, Natural Gas Eckhard Loser, Bayer AG, Wuppertal,
Scott Han, Mobil Research and Development Germany, Isoprene
Corporation, Princeton, New Jersey, United Torsten Lübcke, Mobil Erdgas–Erdöl GmbH,
States, Fuels, Synthetic, Liquid Celle, Germany, Natural Gas
R. H. Harding, W. R. Grace & Company, Dieter Mayer, Hoechst Aktiengesellschaft,
Columbia, Maryland, United States, Fluid Frankfurt, Germany, Acetylene; Hydrocarbons
Catalytic Cracking, Catalysts and Additives Helmut Meinass, Linde AG, Höllriegelskreuth,
B. W. Hedrick, UOP LLC, Des Plaines, Illinois, Germany, Acetylene
United States, Fluid Catalytic Cracking, Units, Georg Michalczyk, DEA Mineralöl AG,
Regeneration Hamburg, Germany, Waxes
Werner Hefner, BASF Aktiengesellschaft, John J. Mooney, Environmental and Energy
Ludwigshafen, Germany, Acetylene Technology and Policy Institute, Wyckoff,
Franz-Leo Heinrichs, Hoechst New Jersey, United States, Emission Control,
Aktiengesellschaft, Werk Gersthofen, Automotive
Augsburg, Germany, Waxes Patrick Moran, U.S. Naval Academy,
Robert N. Heistand, Consultant, Englewood, Annapolis, Maryland, United States,
Colorado, United States, Oil Shale Corrosion and Corrosion Control
C. L. Hemler, UOP LLC, Des Plaines, Illinois, Sanford L. Moskowitz, American Economics
United States, Fluid Catalytic Cracking, Units, Group and Chemical Heritage Foundation,
Regeneration Abington, Pennsylvania, United States,
Albert M. Hochhauser, ExxonMobil Research Synthetic Organic Chemicals, Economic
and Engineering Company, Paulsboro, New Evaluation
Jersey, United States, Gasoline and Other Eckart Müller, Frankfurt am Main, Germany,
Motor Fuels Liquid–Liquid Extraction
Vol. 1 CONTRIBUTORS xi
Richard Müller, Degussa-Hüls AG, Marl, W. J. Reagan, UOP LLC, Des Plaines,
Germany, Acetylene Illinois, United States, Fluid Catalytic
Thomas Narbeshuber, BASF Cracking, Units, Regeneration
Aktiengesellschaft, Ludwigshafen, Germany, Herta Recknagel, Ruhrgas AG, Essen,
Acylation and Alkylation Germany, Natural Gas
Paul Natishan, Naval Research Laboratory, Arno Reglitzky, Deutsche Shell AG, Hamburg,
Washington, DC, United States, Germany, Octane Enhancers
Corrosion and Corrosion Control J. R. Riley, W. R. Grace & Company,
J. R. D. Nee, W. R. Grace & Company, Columbia, Maryland, United States,
Columbia, Maryland, United States, Fluid Fluid Catalytic Cracking, Catalysts and
Catalytic Cracking, Catalysts and Additives Additives
L. H. Nemec, Albemarle Corporation, Baton Heinrich Rischmüller, Technische Universität
Rouge, Louisiana, United States, Olefins, Clausthal, Clausthal-Zellerfeld, Germany,
Higher Exploration, Drilling, and Production
Hans-Joachim Neumann, German Petroleum Engineering
Institute, Clausthal-Zellerfeld, Germany, Winston K. Robbins, Exxon Research and
Natural Gas Engineering Company, Annandale, New
Otto S. Neuwirth, UK Wesseling, Wesseling, Jersey, United States, Petroleum,
Germany, Oil Refining, Environmental Introduction
Considerations Michael Röper, BASF Aktiengesellschaft,
Franz Nierlich, Oxeno Olefinchemie GmbH, Ludwigshafen, Germany, Acylation and
Marl, Germany, Methyl Tert-Butyl Ether Alkylation
E. Dickson Ozokwelu, Amoco Chemical F. S. Rosser, UOP LLC, Des Plaines, Illinois,
Company, Naperville, Illinois, United States, United States, Fluid Catalytic Cracking, Units,
Toluene Regeneration
Barbara Paczynska-Lahme, Consultant, Michael Sakuth, Hüls Infracor GmbH, Marl,
Osterode, Germany, Natural Gas Germany, Methyl Tert-Butyl Ether
Christian Paetz, Bayer AG, Leverkusen, J. D. Sauer, Albemarle Corporation, Baton
Germany, Hydrocarbons Rouge, Louisiana, United States, Olefins,
P. Palmas, UOP LLC, Des Plaines, Illinois, Higher
United States, Fluid Catalytic Cracking, Units, Alan W. Scaroni, Pennsylvania State
Regeneration University, University Park, Pennsylvania,
Peter Pässler, BASF Aktiengesellschaft, United States, Combustion Science and
Ludwigshafen, Germany, Acetylene Technology
Wolfgang Payer, Hoechst Aktiengesellschaft, Kuno Schädlich, VEBA OEL AG,
Werk Ruhrchemie, Oberhausen, Germany, Gelsenkirchen, Germany, Octane Enhancers
Waxes Hans Schmidt, Hoechst Aktiengesellschaft,
Udo Peters, Oxeno Olefinchemie GmbH, Marl, Werk Gersthofen, Augsburg, Germany,
Germany, Methyl Tert-Butyl Ether Waxes
Mark R. Pillarella, Air Products and Peter Schug, VEBA OEL AG, Gelsenkirchen,
Chemicals, Inc., Allentown, Pennsylvania, Germany, Octane Enhancers
United States, Natural Gas Ekkehard Schulte-Körne, Oxeno Olefinchemie
Edwin M. Piper, Piper Designs LLC, Littleton, GmbH, Marl, Germany, Methyl Tert-Butyl
Colorado, United States, Oil Shale Ether
Sarma V. Pisupati, Pennsylvania State Reza Sharifi, Pennsylvania State University,
University, University Park, Pennsylvania, University Park, Pennsylvania, United States,
United States, Combustion Science and Combustion Science and Technology
Technology Wolfgang Siegel, BASF Aktiengesellschaft,
Heinrich Predel, MiRO GmbH & Company Ludwigshafen, Germany, Acylation and
KG, Karlsruhe, Germany, Petroleum Coke Alkylation
Roger C. Prince, Exxon Research and Jeffrey J. Siirola, Eastman Chemical
Engineering Company, Annandale, Company, Kingsport, Tennessee, United
New Jersey, United States, Bioremediation States, Separations Process Synthesis
Günter Pusch, Technische Universität Thomas A. Simo, Lurgi Öl Gas Chemie GmbH,
Clausthal, Clausthal-Zellerfeld, Germany, Frankfurt, Germany, Xylenes
Exploration, Drilling, and Production Domien Sluyts, Bayer AG, Antwerpen,
Engineering Belgium, Liquid–Liquid Extraction
xii CONTRIBUTORS Vol. 1
PETROLEUM,
INTRODUCTION
1. Origin
1
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1518090702011811.a01.pub2
2 PETROLEUM, INTRODUCTION Vol. 1
(1,2) and there is a dualist theory incorporating both biological and inorganic
aspects (3). However, because inorganic processes generate racemic mixtures,
the presence of optically active compounds in oils, especially the multiringed
cycloalkanes (naphthenes), provides strong support for a biological hypothesis.
Oils also contain the so-called chemical fossils or biomarkers, compounds hav-
ing characteristic molecular structures that can be related to living systems.
The compounds include isoprenoids, porphyrins, steranes, hopanes, and
many others. The relative abundances of members of homologous series are
often similar to those in living systems. The strong odd preference in the
long-chain normal alkanes (C4) is particularly well documented (4). In addi-
tion, the lack of thermodynamic equilibrium among compounds (5), and the
close association of petroleum with sedimentary rocks formed in an aqueous
environment, suggests a low temperature origin. In this context, low tempera-
ture means less than a few hundred degrees Celsius as opposed to tempera-
tures in the 700–12008C range that characterize igneous processes involving
silicate melts. The elemental composition of petroleum (C, H, N, S, O), the iso-
topic composition of oils, and the presence of petroleum-like materials in more
recent sediments are consistent with a low temperature origin. The evidence
supporting a biological source for the material that generates petroleum is
extensive (6–8).
Organisms produce a wide range of organic compounds including signi-
ficant amounts of biopolymers like proteins, carbohydrates, and lignins, together
with a wide variety of lower molecular weight lipids (6–8). After the death of the
organism, all or part of this organic material may accumulate in aquatic envir-
onments where the various compounds have very different stabilities. Some are
metabolized in the water column by other organisms (including bacteria) and
only the biochemically resistant material is incorporated into sediments. Survi-
val of organic material depends on many factors but particularly the oxidizing or
reducing nature of the system. Preservation is strongly favored in anoxic sedi-
ments (9). However, the formation of a petroleum accumulation requires more
than just a concentration of the relatively low molecular weight hydrocarbons
that are present in more recent sediments. Although C2 –C10 hydrocarbons are
present in extremely low (parts per billion (ppb) level) concentrations in organ-
isms and sediments, these can account for up to 50% or more of the volume of
some crude oils.
Compounds that are not synthesized by organisms are also reported in
crude oils. Apparently these were formed from the available organic matter in
what is thought to be the main process of petroleum generation. As the organic
matter in sediments is buried in a reducing environment, and subjected to gra-
dually increasing temperature and pressure, petroleum is generated as an inter-
mediate in a transformation process that ultimately leads to methane and
graphite. Oxygen is first lost from the organic matter as carbon dioxide and
water. Continued low temperature (<8C) rearrangement takes place and the
immature organic matter is converted to high molecular weight, insoluble mate-
rial called kerogen. This becomes more aromatic and carbon-rich as it generates
petroleum and evolves toward graphite. In contrast, the solvent-extractable
lower molecular weight organic materials called bitumen or extractables
Vol. 1 PETROLEUM, INTRODUCTION 3
2. Composition
CH 3
CH 2 CH 3
CH 2 CH 2 CH 2 CH 3 CH CH 2 CH 2 CH 3 CH 2
H 3C CH 2 CH 2 CH 2 CH 3 CH 2 CH 2 CH 2
CH 3
CH 2 CH 2
CH 2 CH 3
CH 3
(4) (5) (6)
CH 2 SH CH 3 CH 2 CH 3
CH 3 CH 2 S CH 2 S
COOH
OH
COOH
N N N O
H H
oxygen compounds are more likely to be found in polar compounds such as pyr-
idines, pyrroles, phenols, carboxylates, amides, etc, than in nonpolar ones such
as ethers (61). The distribution and characteristics of these molecular species
account for the rich variety of crude oils.
2.1. Elemental Composition. On an atomic basis, H/C ratios range
from 1.5–2.0. The range of elemental composition of crude oil (16) may be
given as follows:
Nickel and vanadium are also generally present from a trace amount up to 1000
ppm. The value of using sulfur for petroleum classification can be clearly seen.
Whereas the range of hydrogen and carbon are quite narrow, sulfur, the princi-
pal heteroatom in crude oil, varies significantly. The ranges for the other two het-
eroatoms (O and N) are also fairly wide, however in most crude oils these
elements are nearly an order of magnitude lower than the sulfur level (Fig. 2).
For example, the average level of nitrogen in >9000 crude oils is 0.094%, the
average sulfur content is 0.65% (7). Although >90% of crude oils have nitrogen
levels <0.2%, some crude oils have > 1.5% nitrogen. As noted in Figure 2b, the
nitrogen concentrates into the highest boiling fractions. The trace metals Ni and
V are generally orders of magnitude higher than other metals in petroleum,
unless it is contaminated with coproduced brine salts (Naþ, Mg2þ, Ca2þ, Cl)
or iron corrosion products gathered in transportation.
2.2. Molecular Classes. The molecules in crude oil include several basic
structural types (Table 1, Fig. 1). Because they may contain from 1 to 100þ carbon
atoms and may occur in combination, the statistical potential for isomeric struc-
tures is staggering. For example, whereas there are just 75 possible paraffinic
10
Sulfur, % of fraction
C
0
(a)
Nitrogen, % of fraction
1.0
A
B
0
15 210 345 565
Boiling point, °C
C5 C10 C20 C40
(b)
Fig. 2. Distribution of compounds as a function of crude oil boiling point: (a) sulfur
where A, B, and C represent high (>2%), medium (ca 1.5%), and low (<0.1%) sulfur,
respectively, and (b) nitrogen, where A and B represent high (ca 0.5%) and low (<0.1%)
nitrogen, respectively.
10 PETROLEUM, INTRODUCTION Vol. 1
structures for C10, there are >105 isomers for C20. A few structures tend to dom-
inate the distributions of each isomer group, however.
The inclusion of naphthene and other aromatic rings introduces two addi-
tional dimensions, increasing the number of hydrocarbon isomers even further.
A three-dimensional array in which the molecules could be described in terms of
the number of aromatic rings, the number of naphthenic rings, and the number
of carbons in alkyl side chains has been proposed (61). Conceptually, this
amounts to describing a three-dimensional molecular mountain for hydrocar-
bons. There is also the potential of constructing similar mountains for heterocyc-
lics. The two-dimensional image in the naphthenic and aromatic dimensions has
been projected using sidebars to indicate the variation in alkyl substituents
among crude oils (7).
Molecular characterization of a whole oil is beyond the capability of most
analytical techniques. Distillation, however, can separate petroleum into mole-
cular weight fractions that simplify the task. Pioneering work with this
approach, sponsored by the American Petroleum Institute (API) starting in
1925, has led to the identification of hundreds of individual compounds in distil-
lation fractions of a single crude oil (62). Developments in chromatography
allowed oils to be fractionated by polarity as a second dimension. Under API
sponsorship, the U.S. Bureau of Mines extended separations and measurement
techniques to heavier fractions (63). At the same time, individual compounds
have been isolated and quantified from increasingly higher boiling fractions
(64–67). Techniques have been developed that use combinations of classical
open-column adsorption chromatography, gel permeation chromatography, and
ion-exchange separations to isolate fractions in which compounds could be iden-
tified by mass spectrometry.
Whereas neither distillation nor chromatography achieves perfect separa-
tions among groups, the fractions generated are amenable to molecular charac-
terization (66,67). An elegant argument has been made for the use of
distillation for primary separation which is combined with a solubility step
to achieve the atmospheric equivalent boiling point (AEBP) scale that covers
room temperature to 13708C (68). Because the distillations simplify analytical
complexity, the bulk of available molecular compositional data on petroleum
has been generated on sets of fractions defined by boiling points as gases
(C1 –C4), naphtha (initial bp 2108C), mid-distillate (210–3458C), vacuum
gas oil (345–5658C), and vacuum residuum (>5658C). The initial boiling
point (ibp) of the naphtha depends on the amount of C1 –C5 hydrocarbons dis-
solved in the oil. For characterization purposes, naphtha is usually depenta-
nized so that the ibp is about 328C. The material that boils above 3458C is
referred to as atmospheric resid and is further distilled under vacuum into
vacuum gas oil (VGO) and vacuum resid. Whereas the AEBP of the vacuum
resid is >5658C, the actual temperature of distillation does not exceed 3458C
to avoid thermal decomposition.
Crude distillations yield different quantities in each fraction. About the
same amounts are distilled into the middle distillate and vacuum gas oil from
conventional crude oils. More naphtha is distilled from light crude oils and more
vacuum residuum is obtained from heavy crude oils (Fig. 3). The typical distri-
bution of classes of petroleum compounds shows a significant shift with boiling
Vol. 1 PETROLEUM, INTRODUCTION 11
100
A
0
15 210 345 565
Boiling point, °C
C5 C10 C20 C40
Middle Vacuum
Naphtha Resid
distillate gas oil
Fig. 3. Cumulative volume distilled as a function of boiling point from A, light; B, inter-
mediate; and C, heavy crude oils (not including condensates).
point (Fig. 4). Whereas the lower boiling fractions are dominated by nonpolar
saturated hydrocarbons that exist in limited isomeric forms, the higher boiling
fractions increasingly contain a larger variety of classes, that have, in turn, an
increasing number of possible isomers. As the boiling point increases, aromatic
ring structures build in, first as naked rings, then more and more as rings
100
1 Ring
2 Ring
3 Ring
4+ Ring
Cumulative mass,%
50 C
0
15 210 345 565 705
Boiling point, °C
C5 C10 C20 C40
CH 3 CH 3
CH 2 CH 2 CH 2 CH CH 2 CH 2
CH 3 CH 2 CH 2 CH 3 CH 3 CH 2 CH 2 CH 3
CH 3
CH 2
CH 3
CH 3
S CH 2
CH 3 CH 3 CH CH 3
CH 3 SH
CH 2 CH 3
S
(25) (26) (27)
found in greater quantity than the parent unsubstituted structures and can be
present at levels above 2% (7). Fused-ring dicycloalkanes such as decahydro-
naphthalenes (decalins) and hexahydroindans are also common, but nonfused
bicylic naphthenes, eg, cyclohexyl cyclohexane, are not.
The numerous aromatics in petroleum naphtha begin with benzene, but
the C1 –C3 alkylated derivatives of benzene generally are present in larger
amounts. Toluene concentrations may reach nearly 2%; the combined xylene
isomers exceed 1%; benzene, however, rarely exceeds 1% of a whole crude oil
(7). Although present in lesser concentrations, each of the alkyl benzene homo-
logues through the 20 isomeric C4 alkyl benzenes have been isolated from
Ponca City crude, as have several of the C5-derivatives (76). Benzenes having
fused cycloparaffin rings (naphthenoaromatics) such as indan and tetralin
have been isolated along with a number of their methyl derivatives. Naphtha-
lene is included in the naphtha, whereas the 1- and 2-methyl naphthalenes and
higher homologues of fused two-ring aromatics appear in the mid-distillate
fraction.
Sulfur is the only heteroatom to be found in naphtha, and only at trace
levels (Fig. 2). A systematic study of organic sulfur compounds in Wasson
Texas crude, and to a lesser extent three other crude oils, culminated in the
identification of over 200 compounds, most of which were in the naphtha fraction
(24). However, the total of these compounds accounted for less than 1% of the
Vol. 1 PETROLEUM, INTRODUCTION 15
sulfur in the whole crude. In sour petroleum naphthas, 50–70% of the sulfur can
be found in the form of mercaptans, ie, thiols. Over 40 individual thiols have been
identified, including all the isomeric C1 to C6 compounds plus some C7 and C8
isomers and thiophenol (77). In sweet naphthas, the sulfur is distributed
between sulfides, ie, thio-ethers, and thiophenes. In these cases the sulfides
may constitute more than 50% of the sulfur compounds in the form of both linear
(alkyl sulfides) and five- or six-ring cyclic (thiacyclane) structures. Sulfur struc-
ture distribution mimics the hydrocarbons; ie, naphthenic oils having high
cycloalkanes have high thiacyclane content, etc. Typical alkyl thiophenes in
naphtha have multiple short side chains or exist as naphthenothiophenes (77).
Though rare, methyl and ethyl disulfides have been confirmed to be present in
some crude oils (78).
Mid-Distillates. As is indicated in Figure 4, saturates remain the primary
component in the mid-distillate fraction of petroleum, but aromatics, which
include simple compounds having up to three aromatic rings, and heterocyclics
are present and represent a larger portion of the total. Some raw middle distil-
lates are used directly as kerosenes, jet fuels, and diesel fuels; others are
cracked and hydroprocessed before use. Some compounds found in middle dis-
tillates are listed in Table 3. Within the saturates, the concentration of n-par-
affins generally decreases regularly from C11 to C20. The 2-methyl analogues
are sufficiently unique to be seen as distinct peaks in gc analyses of mid-distil-
lates. Few additional isoparaffins have been identified. However, the two iso-
prenoids (30) and (31) (see Fig. 6) are generally present in crude oils in
sufficient concentration to be seen as irregular peaks alongside the n-C17 and
n-C18 peaks in a gas chromatogram. These isoprenoids, believed to arise primar-
ily as fragments of ancient chlorophyll [1406-65-1] have relevance, as simple
CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 3
CH 3 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2
(28)
CH 3
CH CH 2 CH 2 CH 2 CH 2 CH 2 CH 2
CH 3 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 3
(29)
CH 3 CH 3 CH 3 CH 3
CH CH 2 CH CH 2 CH CH 2 CH
CH 3 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 3
(30)
CH 3 CH 3 CH 3 CH 3
CH CH 2 CH CH 2 CH CH 2 CH CH 3
CH 3 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2 CH 2
(31)
CH 3
CH
CH 3
H 3C H 3C CH 3
CH 3
H 3C CH 3
CH 3
S CH 3 S S
H
H 3C CH 2 CH 2 CH 3
C CH 2 CH
H 3C H H
H
CH 3
H 3C H 3C CH 3 CH CH 3
CH 3
H CH 3
H
H 3C CH 3
(a) (b)
distinct in the gc, these usually account for only a few percent of the saturates
measured by gc.
The bulk of VGO saturates consists of isoparaffins and especially naphth-
enes (Fig. 3). A few isoprenoid compounds, such as squalane, C30; lycopane,
C40; and carotanes, C40; have been detected. Analyses of petroleum waxes iso-
lated from the saturates show a parallel lower level of 2- and 3-methyl alkanes
as the most identifiable isoparaffins. Mass spectrometry techniques show that
the naphthenes contain from one to more than six fused rings. Having an aver-
age carbon number of C32, even the six-ring naphthenes have some alkyl sub-
stitution. For mono- and diaromatics, the alkyl substitution typically involves
one long side chain and several short methyl and ethyl substituents. Some spe-
cific tetracyclic naphthenes, including steranes, and pentacyclic naphthenes,
including hopanes, have been used as biomarkers (Fig. 7). These hopanes
and steranes have also been used as nondegradable conserved internal
markers for estimating biodegradation of crude oils during bioremediation
processes (83).
The aromatics in VGO may contain one to six fused aromatic rings that may
bear additional naphtheno rings and alkyl substituents in keeping with their boil-
ing range. Mono- and diaromatics account for about 50% of the aromatics in petro-
leum VGO samples. Studies by nmr indicate an average of 3.5 substituents per
monoaromatic, whereas ms techniques show the presence of up to four fused
naphthenic rings on some aromatic compounds (71). This is consistent with the
suggestion that these species originate from the aromatization of biogenic steroids
(7,84). Although they are present at lower concentration, alkyl benzenes and
naphthalenes commonly show one long side chain and multiple short side chains.
The fused 3þ ring aromatics in petroleum include both cata- and peri-
condensed structures (see Table 4, Fig. 8). The cata-condensed species are
those structures where only one face is shared between rings, the peri-condensed
molecules are those that share more than one face. The fused ring aromatics
form the class of compounds known as polynuclear aromatic hydrocarbons
(PAH) which includes a number of recognized carcinogens in the 4þ ring
family (85). Because of the potential health and environmental impact of PAH,
these compounds have been studied extensively in petroleum.
Vol. 1 PETROLEUM, INTRODUCTION 19
The total levels of three–six ring PAH in a petroleum VGO fraction range
from 2 to 25%. However, the concentrations of individual isomers that have been
quantified are generally expressed in parts per million (ppm). Although the most
abundant reported individual phenanthrene compounds appear to be the C1 –C3
derivatives, the average carbon number for the phenanthrenes in a VGO is
C32, indicating that the average side chain actually has 18 carbon atoms. The
analytical techniques that have been used to isolate individual compounds are
biased toward the least substituted aromatic structures, ie, the techniques that
facilitate identification of a few target compounds eliminate the majority of
multiring aromatic compounds.
Within petroleum certain aromatic structures appear to be favored. For
example, alkyl phenanthrenes outnumber alkyl anthracenes by as much as
components linked together with carbon and sulfur bridges rather than ever-
increasing fused-ring structures (88).
The Ni and V concentrated into the vacuum resid appear to occur in
two forms. From 10 to 14% of each of these two metals can be distilled in the
565– 7058C boiling range, where they exhibit the strong visible Soret bands
associated with the porphyrin structure. This tetrapyrrole structure (48, 49),
possibly derived from ancient chlorophyll, has been confirmed by a variety
of analytical techniques.
R′ R′ R′ R′
R R R R
N N N N
Ni V O
N N N N
R R′ R R′
R′ R′ R′ R′
(48) (49)
3. Nomenclature
Crude oils, complex mixtures of naturally occurring organic liquids, are difficult
to characterize in detail. Thus, many of the definitions used by the exploration,
production, and refining sectors of the petroleum industry to describe petroleum
and its products often lack precision. Even the term petroleum is poorly defined.
Although often used synonymously with crude oil, petroleum is also frequently
used to include natural gas (see NATURAL GAS) and even solid hydrocarbons. Defi-
nitions of materials are commonly given in terms of the processes used to obtain
them. Gasoline, for example, is the fraction of crude oil that distills between 15
and 2008C (60 and 3928F) (see GASOLINE AND OTHER MOTOR FUELS). Further compli-
cations arise because different parts of the petroleum industry use terms in dif-
fering ways. For example, wax may refer to material made up predominantly of
long-chain alkanes, or it may refer to esters of long-chain alcohols and acids (see
WAXES). Even the term hydrocarbons is used loosely indicating all the compounds
in crude oils, whether or not these include compounds of nitrogen, sulfur, and
oxygen.
Vol. 1 PETROLEUM, INTRODUCTION 23
Water corresponds to an API gravity of 10; crude oils fall between 10 and
608API. The most common crude oil values are in the 35–408 range.
Other terms relating to physical properties include viscosity; refractive
index; pour point, ie, the lowest temperature at which the oil flows; flash
point, ie, the temperature at which the oil ignites; and aniline point, ie, the
minimum temperature at which equal volumes of oil and aniline are comple-
tely miscible. These are determined under defined conditions established by
ASTM.
Natural gas production is generally given in cubic feet or cubic meters
(1000 ft3 ¼ 1 Mcf ¼ 28.3 m3). Reserves of a trillion cubic feet (Tcf) (28.3
1012 m3) or more form a giant gas field. Natural gas is called dry when methane
is the dominant hydrocarbon, and wet if it contains more than 4 L/100m3 of nat-
ural gas liquids (>0.3 gal/1000 ft3). When gas (or oil) has a bad odor owing to high
concentrations of hydrogen sulfide and volatile sulfur compounds it is called
sour. Sweet gas has no noticeable odor. For statistical purposes gas is commonly
reported as an equivalent amount of oil based on an equivalent heating capacity.
The conversion is normally made using 170 m3 (6000 ft3 ¼ 1 bbl) and leads to a
barrel of oil equivalent (boe).
Crude oils contain a wide range of hydrocarbons including straight and
branched chains, ring compounds, and aromatics, as well as more complex com-
pounds that incorporate nitrogen, sulfur, and oxygen (often called the NSOs),
and some nickel and vanadium. The straight-chain, normal alkanes, range
24 PETROLEUM, INTRODUCTION Vol. 1
from 1 to >100 carbon atoms. These are often called paraffins in the petroleum
industry because of the useful adjective paraffinic. Branched hydrocarbon chains
that are nominally built up from repeated isoprene units (2-methyl butane struc-
ture) are called isoprenoids or terpenoids, and the 19- and 20-carbon compounds,
named pristane and phytane, respectively, are frequently present in high con-
centrations. Isoprenoids also lead formally to saturated multiring structures.
Petroleum chemists use the obsolete word naphthenes for the compounds that
organic chemists call alicyclics. A better term, cycloparaffin, is used herein, lead-
ing to the adjective cycloparaffinic rather than naphthenic. Some of the charac-
teristic structures in this group can be directly related to molecules synthesized
by organisms. Whereas these have been called chemical fossils, it is more usual
to call them biological markers or biomarkers. Common examples include the
steranes and hopanes.
Aromatic hydrocarbons form a minor but important group of compounds
in crude oils and range from single-ring to multiring compounds. The latter
are called polycyclic aromatics (PAHs). Small aromatic molecules are environ-
mentally significant and BTEX is commonly used as an abbreviation for
benzene–toluene–ethyl benzene–xylenes. Multiringed compounds containing
both aromatic and saturated rings may be referred to in the older literature
as naphtheno-aromatics. The highest molecular weight fraction of crude oils
commonly contains ashphaltenes that are dark in color, NSO-rich, and very
aromatic.
Most crude oil is refined to provide useful products and the dominant
process is distillation (qv) (Table 5). Petroleum products produced by simple
distillation without the use of pressure, cracking, or catalysts are called straight
run. Residual material that has too high a molecular weight to distill forms
a residuum, often called by such names as asphalt. Naphtha (unrelated to
naphthenes) is a distillate of petroleum having a boiling range lower than
about 200 or 2608C (even occasionally up to 3508C). As a process intermediate,
naphtha includes the components used to formulate gasoline and the lighter
grades of fuel oils such as kerosene and diesel fuel oil. As a finished product,
naphtha usually denotes a more specific type of narrow boiling range material.
The terms naphtha and solvent may be used interchangeably. For example,
Varnish Makers’ & Painters’ (VM&P) naphtha has a range of 95–1508C. The
4. Petroleum Resources
Petroleum resources are distributed widely in the earth’s crust as gases, liquids,
and solids. The products derived from these naturally occurring resources
are used principally as energy sources, although substantial volumes serve as
feedstocks in the chemical, plastics, and other industries. Petroleum resour-
ces are found as natural gas, as a variety of liquids that are usually classified
as normal or heavy crude oils, and as semisolid and solid substances such as
asphalt, tar, pitch, gilsonite, and many others. The petroleum resources consid-
ered here are those liquid crude oils that can be produced through a conventional
wellbore by current primary, secondary, or tertiary (enhanced recovery) produc-
tion techniques and those unconventional crude oils that may be captured and
converted into conventional sources of crude petroleum by advancing production
technologies.
No method has been devised to estimate with complete accuracy the
amount of crude petroleum that ultimately will be produced from the world’s
conventional oil and gas fields. Degrees of uncertainty, therefore, should be
attached to all such estimates. These uncertainties can be expressed in several
ways, the most important of which is achieved by dividing a resource into var-
ious categories. Several petroleum resources classifications have been pro-
posed, and a comprehensive discussion of them (91), as well as the definition
used in the assessment of the undiscovered resources of the United States
(92), have been provided. Seven commonly used categories of resources are
given here.
Resources represent the total amount (including reserves) of petroleum that
exists in a form and amount such that economic extraction is currently or poten-
tially feasible.
Reserves constitute the petroleum that has been discovered and can be pro-
duced at the prices and with the technology that exist when the estimate is
made.
Proved reserves are estimates of petroleum reserves contained primarily in
the drilled portion of fields.
Indicated reserves constitute known petroleum that is currently producible
but cannot be estimated accurately enough to qualify as proved.
Inferred reserves are producible, but the assumption of their presence is
based on limited physical evidence and considerable geologic extrapolation.
Vol. 1 PETROLEUM, INTRODUCTION 27
This places them on the borderline of being considered undiscovered, and the
accuracy of the estimate is very poor.
Subeconomic resources constitute the petroleum in the ground that cannot
be produced at present prices and technology but may become producible at some
future date at higher prices or by improved technology.
Undiscovered resources are estimated totally by geological reasoning; no
evidence through drilling is available.
To various degrees, the conventional petroleum resources in many
parts of the world have been classified according to such a system. In certain
regions, only estimates of proved reserves are made routinely, whereas in
the United States, Canada, and certain other regions, estimates are made of
volumes of petroleum in each of these categories. In the United States and
Canada, estimates for several of these categories (in particular, proved reserves)
are made each year by governmental agencies. For other categories, eg, undis-
covered U.S. resources, many estimates have been prepared by various agencies,
committees, panels, and companies. In a similar manner, estimates for several
resource categories have been made for most other countries and for the world
as a whole.
4.1. World Reserves. Most of the large volume of crude petroleum con-
sumed in the world is extracted from only a small fraction of the total number of
oil fields discovered. The concentration of crude petroleum in a few large fields is
a consequence of the interaction of the geologic processes that create and trap
petroleum. Even though commercial quantities of petroleum have been discov-
ered in many localities around the world, there are enormous volume differences
in fields present in a single region and in the total volume of petroleum present in
different regions.
By far the largest known concentrations of conventional petroleum reserves
are in the Middle East, particularly in Saudi Arabia, the United Arab Emirates,
and Kuwait. The largest concentration of reserves is in the Burgan field (10.2
109 m3 (64.2 109 bbl)) in Kuwait (10), which contains about 68% of that coun-
try’s reserves. The second largest concentration of reserves is in the Ghawar
field (7.4 109 m3 (46.5 109 bbl)) in Saudi Arabia (93), which is about 18% of
that country’s reserves. In some regions, a large portion of the reserves may
not be contained in the largest field. However, the largest field usually contains
more than 10% of the total reserves of a region. More than 20,000 petroleum
fields have been discovered worldwide, and more than half of the world’s proved
reserves of ca 160.1 109 m3 (1006.8 109 bbl) of petroleum are contained in only
the 51 largest fields (93).
The Energy Information Administration estimates U. S. oil and gas
reserves, but does not systematically estimate worldwide reserves. They have
published a listing of international reserves as presented in two widely circu-
lated trade publications (Table 6). The world’s oil reserves are estimated to be
1.16 trillion barrels. The United States ranks 11th in the world for proved
reserves of crude oil. World reserves increased about 4% in 2003 owing to the
addition of former Soviet Union States.
Other large reserves other than Saudi Arabia’s are located in Iraq, whose
reserves are almost five times larger than that of the United States, Venezuela
and Canada have triple the amount.
28 PETROLEUM, INTRODUCTION Vol. 1
4.2. United States’ Reserves. The United States had the following
proved reserves as of December 31, 2003: crude oil, 21,891 106 barrels; dry
natural gas, 189,0444 109 ft3; and natural gas liquids, 7,459 106 barrels (94).
Table 7 listed the estimated annual oil reserve balances since 1993.
Crude oil proved reserves, reserve changes and production for individual
states for 2003 are listed in Table 8.
Proved reserves of crude oil decreased by 786 106 barrels in 2003. Total
discoveries are those reserves attributable to field extensions, new field discov-
eries, and new reservoir discoveries in old fields.
The majority of crude oil total discoveries were in new field discoveries in
the Gulf of Mexico Federal Offshore. New field discoveries accounted for 705
106 barrels of crude oil additions, 702 106 were in the Gulf of Mexico. Operators
Table 7. Total U.S. Proved Reserves of Crude Oil 106 barrels of 42 U.S. gal 1993–2003
New
Revisions Net of sales reservoir Change
Net and and New field discoveries Total Estimated Proved from
Adjustments revisions adjustments acquisitions Extensions discoveries in old fields discoveries Production Reserves Prior
Year (1) (2) (3) (4) (5) (6) (7) (8) (9) 12/31 (10) Year (11)
1993 271 495 766 NA 356 319 110 785 2,339 22,957 788
1994 189 1,007 1,196 NA 397 64 111 572 2,268 22,457 500
1995 122 1,028 1,150 NA 500 114 343 957 2,213 22,351 106
1996 175 737 912 NA 543 243 141 927 2,173 22,017 334
1997 520 914 1,434 NA 477 637 119 1,233 2,138 22,546 þ529
1998 638 518 120 NA 327 152 120 599 1,991 21,034 1,512
29
1999 139 1,819 1958 NA 259 321 145 725 1,952 21,765 þ731
2000 143 746 889 20 766 276 249 1,291 1,880 22,045 þ280
2001 4 158 162 87 866 1,407 292 2,565 1,915 22,446 þ401
2002 416 720 1,316 24 492 300 154 946 1,875 22,677 þ231
2003 163 94 257 398 426 705 101 1,232 1,877 21,891 786
a
Ref. 97.
b
Revisions and adjustments ¼ Col. 1þ Col. 2.
c
Net of sales and acquisitions ¼ acquisitions sales; NA ¼ not available.
d
Total discoveries ¼ Col. 5 þ Col. 6 þ Col. 7.
e
Proved reserves ¼ Col. 10 from prior year þ Col. 3þ Col. 4 þ Col. 8 Col. 9.
Notes: Old means discovered in a prior year. New means discovered during the report year. The production estimates in this table are based on data reported
on Form EIA-23, ‘‘Annual Survey of Domestic Oil and Gas Reserve’’ and Form EIA-64A, ‘‘Annual Report of the Origin of Natural Gas Liquids Production.’’
They may differ from the official EIA production data for crude oil, natural gas, and natural gas liquids for 2003 contained in the Petroleum Supply Annual
2003, DOE/EIA-0340(03) and the Natural Gas Annual 2003, DOE/EIA-0131(03).
Table 8. U. S. Crude Oil Proved Reserves, Reserves Changes, and Production by States, 106 barrels (42 U.S. gal) 2003a
Changes in reserves during 2003
30
Illinois 107 22 11 5 0 0 0 0 0 10 125
Indiana 15 5 1 0 0 0 0 0 0 2 19
Kansas 237 19 31 13 4 2 5 0 0 34 243
Kentucky 27b 2 2 4 0 0 0 0 0 2 25
Louisiana 501 3 57 75 46 38 31 0 6 63 452
Michigan 61 12 14 16 10 21 0 0 0 7 75
Mississippi 179 9 20 9 15 1 0 0 0 16 169
Montana 288 6 15 8 2 5 42 0 0 19 315
Nebraska 18 0 1 0 0 0 0 0 0 3 16
New Mexico 710 5 65 69 18 15 38 0 0 59 677
North Dakota 342 8 20 24 1 7 31 0 0 30 353
Ohio 67 9 4 8 1 0 0 0 0 5 66
Oklahoma 598 3 80 49 36 21 26 0 0 55 588
Pennsylvania 12b 3 6 1 0 0 1 0 0 2 13
Texas 5,015 63 362 178 639 242 77 3 2 364 4,583
Utah 241 1 7 18 24 25 1 0 0 12 221
West Virginia 13b 0 2 1 0 0 0 0 0 1 13
Wyoming 524 24 29 36 23 36 5 0 0 42 517
Federal Offshore 5,009 8 376 686 225 243 115 702 93 515 5,120
Pacific 565 1 10 2 0 0 22 0 0 30 566
(California)
Gulf of Mexico 4,088 5 289 616 201 230 93 698 91 426 4,251
(Louisiana)
Gulf of Mexico 356 2 77 68 24 13 0 4 2 59 303
(Texas)
Miscellaneousc 15 1 0 0 0 0 1 0 0 1 16
U.S. Total 22,677 163 1,486 1,392 1,107 709 426 705 101 1,877 21,891
a
Ref. 94.
b
Indicates the estimate is associated with a sampling error (95 percent confidence interval) that exceeds 20 percent of the estimated value.
c
Includes Arizona, Missouri, Nevada, New York, South Dakota, Tennessee, and Virginia.
31
Table 9. International Petroleum Supply and Consumption Summary, 106 barrels/daya
Projections
32
Western Europe 6.69 6.32 6.35 6.65 5.44 5.51 5.85 4.92 5.00 5.35
Japan 0.13 0.08 0.08 0.09 0.06 0.06 0.09 0.05 0.06 0.09
Australia and 0.66 0.95 0.96 1.01 0.86 0.89 0.95 0.83 0.86 0.93
New Zealand
Total industrialized 22.62 22.89 23.05 24.06 21.30 21.89 23.32 20.40 21.16 22.90
Eurasia
former Soviet Union
Russiae 8.34 9.89 9.98 10.64 10.63 10.90 11.87 10.76 11.11 12.20
Caspian Area 1.87 3.11 3.14 3.35 5.10 5.23 5.70 6.03 6.22 6.83
Eastern Europe 0.22 0.33 0.33 0.35 0.40 0.41 0.44 0.44 0.45 0.48
Total Eurasia 10.44 13.33 13.46 14.34 16.13 16.54 18.01 17.23 17.78 19.52
Developing
countries
OPECg
Asia 1.38 1.57 1.47 1.19 1.78 1.51 1.14 1.88 1.56 1.16
Middle East 20.95 26.08 24.45 19.84 38.02 32.37 24.44 46.42 38.47 28.62
North Africa 2.99 3.67 3.44 2.79 5.21 4.44 3.35 5.77 4.78 3.56
West Africa 1.98 2.51 2.36 1.91 3.67 3.13 2.36 4.52 3.74 2.78
South America 2.85 3.56 3.34 2.71 5.21 4.44 3.35 6.27 5.20 3.87
Non-OPEC
China 3.10 3.60 3.64 3.84 3.40 3.49 3.76 3.30 3.41 3.69
Other Asia 2.59 2.62 2.65 2.80 2.64 2.71 2.92 2.56 2.64 2.86
Middle Easth 1.81 2.22 2.24 2.37 2.51 2.57 2.77 2.69 2.78 3.01
Africa 2.94 3.72 3.75 4.04 5.31 5.44 6.00 6.36 6.56 7.31
South and Central 3.93 4.48 4.53 4.83 5.77 5.91 6.44 6.22 6.42 7.05
America
Total developing 44.52 54.05 51.87 46.31 73.52 66.02 56.52 86.00 75.57 63.93
countries
Total production 77.58 90.26 88.38 84.72 110.95 104.45 97.85 123.53 114.51 106.34
(conventional)
Productionh (non-
conventional)
U.S. (50 states) 0.00 0.00 0.00 0.01 0.00 0.00 0.19 0.00 0.00 0.31
other North 0.93 1.63 1.73 1.95 2.81 3.33 3.78 2.94 3.46 4.14
33
America
Western Europef 0.04 0.04 0.04 0.05 0.03 0.05 0.06 0.03 0.05 0.06
Asia 0.03 0.03 0.04 0.05 0.02 0.05 0.08 0.02 0.07 0.10
Middle Easth 0.03 0.03 0.12 0.12 0.03 0.21 0.28 0.04 0.25 0.35
Africa 0.21 0.16 0.23 0.26 0.18 0.28 0.41 0.19 0.32 0.47
South and Central 0.57 0.82 0.82 1.23 1.27 1.48 2.43 1.18 1.50 2.79
America
Total production 1.79 2.70 2.98 3.68 4.34 5.40 7.23 4.40 5.65 6.22
(nonconventional)
Total production 79.37 92.97 91.35 88.39 115.30 109.85 105.08 128.04 120.17 114.56
Consumptiond
Industrialized
countries
U.S. (50 states) 20.00 23.23 22.98 22.41 27.34 26.32 25.47 29.55 27.93 26.85
U.S. Territories 0.36 0.40 0.38 0.35 0.48 0.43 0.39 0.54 0.47 0.42
Table 9. ðContinuedÞ
Projections
Canada 2.17 2.39 2.30 2.15 2.68 2.62 2.39 3.16 2.80 2.55
Mexico 2.02 2.43 2.36 2.25 3.24 2.88 2.51 4.13 3.48 2.93
Western Europed 14.22 15.00 14.72 14.22 16.15 15.45 14.82 16.60 15.71 15.06
Japan 5.58 5.91 5.70 5.34 6.46 5.69 4.95 6.93 5.84 4.99
Australia and New 1.04 1.29 1.27 1.23 1.60 1.54 1.49 1.77 1.69 1.63
Zealand
Total industrialized 45.38 50.66 49.72 47.95 58.15 54.93 52.01 62.68 57.92 54.44
Eurasia
Former Soviet Union 4.18 4.46 4.39 4.26 5.94 5.74 5.55 6.73 6.45 6.24
34
Eastern Europef 1.42 1.58 1.56 1.53 1.93 1.89 1.85 2.15 2.09 2.05
Total Eurasia 5.59 6.04 5.95 5.79 7.88 7.63 7.41 8.88 8.54 8.28
Developing
countries
China 5.54 7.84 7.63 7.27 11.75 11.06 10.45 13.79 12.79 12.08
India 2.19 2.85 2.79 2.69 4.65 4.37 4.09 5.75 5.29 4.92
South Korea 2.17 2.57 2.51 2.39 2.96 2.75 2.56 3.21 2.93 2.72
Other Asia 5.74 7.37 7.28 7.11 9.76 9.47 9.20 11.07 10.66 10.35
Middle Eastf 5.58 6.90 6.83 6.73 8.49 8.34 8.21 9.28 9.08 8.93
Africa 2.72 3.16 3.13 3.07 4.26 4.13 3.99 4.85 4.66 4.49
South and Central 4.69 5.89 5.81 5.69 7.68 7.48 7.30 8.89 8.61 8.40
America
Total developing 28.64 36.57 35.98 34.94 49.55 47.59 45.80 56.83 54.01 51.90
countries
Total consumption 79.60 93.27 91.65 88.68 115.57 110.14 105.22 128.39 120.47 114.62
OPEC production 30.60 38.04 35.79 29.48 54.88 47.21 36.75 65.79 55.13 42.44
non-OPEC 48.77 54.93 55.56 58.91 60.42 62.64 68.33 62.25 65.04 72.13
production
net Eurasia exports 4.84 7.29 7.51 8.55 8.26 8.92 10.60 8.35 9.25 11.23
OPEC market share 0.39 0.41 0.39 0.33 0.48 0.43 0.35 0.51 0.46 0.37
a
Ref. 98.
b
Average refiner acquisition cost of imported crude oil.
c
Includes production of crude oil (including lease condensates, natural gas plant liquids, other hydrogen and hydrocarbons for refinery feedstocks, alcohol
and other sources.
d
Western Europe ¼ Austria, Belgium, Bosnia and Herzegovina, Croatia, Denmark, Finland, France, the unified Germany, Greece, Iceland, Ireland, Italy,
Luxembourg, Macedonia, Netherlands, Norway, Portugal, Slovenia, Spain, Sweden, Switzerland, United Kingdom, and Yugoslavia.
e
Caspian area includes Other Former Soviet Union.
f
Eastern Europe ¼ Albania, Bulgaria, Czech Republic, Hungary, Poland, Romania, and Slovakia.
g
OPEC ¼ Organization of Petroleum Exporting Countries - Algeria, Indonesia, Iran, Iraq, Kuwait, Libya, Nigeria, Qatar, Saudi Arabia, the United Arab
Emirates, and Venezuela.
h
Non-OPEC Middle East includes Turkey.
i
Includes liquids produced from energy crops, natural gas, coat, oil sands, and shale. Includes both OPEC and non-OPEC producers in the regional
breakdown.
j
Includes both conventional and non-conventional liquids production.
35
36 PETROLEUM, INTRODUCTION Vol. 1
discovered 426 106 barrels in extensions in 2003. This total was 13% less than
the previous year. New reservoir discoveries in old fields were 101 106 barrels,
34% less than the previous year.
Reserve additions are the sum of total discoveries, revisions, and adjust-
ments, and sales and acquisitions. In 2003 there were 1,091 106 barrels of
reserves additions, 48% less than the previous year.
The smaller than average volume of net revisions and adjustments and
negative net of sales acquisitions of crude oil proved reserves significantly
reduced reserves additions in 2003. Crude oil net revisions and adjustments
were 257 106 barrels, 77% less than 2002.
Production was estimated at 1,877 106 barrels in 2003.
4.3. World Petroleum Supply and Consumption. Historically, the
world’s petroleum production pattern can be related to geologic, economic,
and political factors. In the past, many countries have had large excesses in
production capacity, whereas since the 1990s, only countries in the Middle
East, such as Saudi Arabia, Kuwait, Iraq, and the United Arab Emirates,
have, in the short run, enough excess capacity to expand production of conven-
tional crude petroleum in any significant manner. In the Middle East, production
of petroleum is over five times the region’s consumption. On a much smaller
scale, Africa produces far more petroleum than it consumes. See Table 9 for
world supply and consumption projected out of 2025.
4.4. Outlook. Petroleum displaced coal as the principal source of energy
in the United States by 1948 and in the world by 1965 (99). The spectacular growth
in consumption of crude petroleum in the world during the middle and late
twentieth century is directly attributable to the ease with which petroleum can
be discovered, produced, transported, processed, and utilized (see ENHANCED OIL
RECOVERY). This growth has been so rapid that as much crude petroleum
(55.5 109 m3 (349.4 109 bbl)) was taken from the ground between 1976 and
1992 as was produced during the entire previous 119-yr period (1857–1975).
This rapid rate of expansion in production and consumption, coupled with the
finiteness of the conventional petroleum resource base, has from time to time led
some analysts to conclude that world petroleum production will peak in the near
future (100,101). Other analysts who examine such data forecast impending global
crisis as crude petroleum consumption declines and coal reclaims its former posi-
tion as the principal source of fossil energy (102).
The key factor influencing the varying interpretations is that although
there is an enormous volume of petroleum resources in the ground through-
out the world, it is found in deposits that differ in quality and quantity from
country to country. Only a small fraction of these resources are conventional
petroleum resources (160.1 109 m3 (1006.8 109 bbl)) and are in the category
of proved reserves. An additional 93 109 m3 (585 109 bbl) of conventional
petroleum is estimated to be undiscovered in the world. This estimate is of undis-
covered petroleum resources that are economic to produce by means of normal
production technology.
Although the world is not running out of petroleum, it is difficult to sum up
how much is available in the short run, as well as in the longer run, in light of
various possible future political and economic developments. The effect of a vari-
ety of sociopolitical forces now at work will be to reduce the world’s consumption
Vol. 1 PETROLEUM, INTRODUCTION 37
Athabasca tar sands in Alberta, Canada, are estimated to contain 360 109 m3
(2250 109 bbl) of in-place petroleum (103). This volume is equivalent to
the total inventory, ie, the combined cumulative production, reserves, and
undiscovered resources, of world conventional crude petroleum. In 1992,
however, only about 10% of total in-place petroleum resources was technically
recoverable (104).
Large unconventional resources of petroleum also occur as extra heavy
crude oils in the Orinoco belt, Venezuela, and in oil shale in the western United
States. Petroleum resources in the unconventional category, such as tar sands,
heavy crude oils, and oil shales, are located mostly in the Western Hemisphere,
as opposed to the conventional resources, which are located mostly in the
Middle East. Also, the in-place resources of these unconventional resources
are about twice as large as the in-place resources of conventional crude
petroleum. Although the recovery rates from these resources are low, improv-
ing technology may capture increasing volumes of these unconventional petro-
leum resources, thereby converting them into conventional petroleum
resources.
The irony underlying the current perceptions of the world’s petroleum
resource situation is that the world is not running out of combined petroleum
resources. However, the bulk of the world’s conventional petroleum resources
that are inexpensive to produce are concentrated in the Middle East, whereas
the much larger volume of higher cost unconventional petroleum resources,
such as tar sands and heavy oils, are located in Venezuela, western Canada,
and the western United States. The uneven distribution of the conventional
resources thus causes concern that in the short run political unrest in the Mid-
dle East could cause a catastrophic interruption of supply.
The world will never ‘‘run out’’ of petroleum, simply because there is so
much of it in the ground in so many different forms. However, the resources of
conventional crude petroleum are finite. These are the petroleum resources that
are very inexpensive to produce because they flow to the wellbore either directly
or by pumping after the application of standard well completion methods. There
is a more or less general agreement among analysts that the size of the inventory
of these resources is about 350 109 m3 (2200 109 bbl); the world is consuming
these resources at about 1%/yr. The primary question that faces the world is,
‘‘Has the pattern of exponential growth in consumption of petroleum that took
place between the end of World War II and 1973 become a relic of the past, or
could growth resume as world population continues to expand?’’ Analysis of
the pattern of world energy consumption shows that the world consumption
of crude petroleum may gradually increase even with increased efficiency in
the use of energy, simply as a result of population growth. However, these deve-
lopments could be dramatically altered by an increase in the price of energy
(105,106).
Another consideration of petroleum assessment analysts is whether, and
to what degree, the vast resources of unconventional petroleum in the world
can be captured by advances in petroleum production technologies, thereby
converting them into conventional sources of petroleum. It is a simple fact
that the in-place resources of petroleum in tar sands, heavy oils, and oil
shale can guarantee the future supply of petroleum for hundreds of years at
Vol. 1 PETROLEUM, INTRODUCTION 39
BIBLIOGRAPHY
‘‘Petroleum Origin’’ in ECT 1st ed., Vol. 10, pp. 97–109, by B. T. Brooks, Consultant;
‘‘Origin of Petroleum’’ under ‘‘Petroleum’’ in ECT 2nd ed., Vol. 14, pp. 838–845, by S. P.
Ellison, Jr., University of Texas; in ECT 3rd ed., Vol. 17, pp. 113–118, by C. Barker,
University of Tulsa; in ECT 4th ed., Vol. 18, pp. 346–352, by Colin Barker, University
of Tulsa; ‘‘Petroleum Origin’’ in ECT (online), posting date: December 4, 2000, by
Colin Barker, University of Tulsa; ‘‘Petroleum Composition,’’ in ECT 1st ed., Vol. 10,
pp. 92–97, by A. N. Sachanen, Socony-Vacuum Oil Co., Inc.; in ECT 2nd ed., Vol. 14,
pp. 845–855, by R. H. Hunt and M. J. O’Neal, Shell Oil Co.; ‘‘Petroleum–Composition,’’
in ECT 3rd ed., Vol. 17, pp. 119–131, by J. J. Elliott and M. T. Melchior, Exxon Research
and Engineering Co.; in ECT 4th ed., Vol. 18, pp. 352–370, by Winston K. Robbins and
Chang Samuel Hsu, Exxon Research and Engineering Company; ‘‘Petroleum Composi-
tion’’ in ECT (online), posting date: December 4, 2000, by Winston K. Robbins and Chang
Samuel Hsu, Exxon Research and Engineering Company; ‘‘Nomenclature in the Petro-
leum Industry’’ under ‘‘Petroleum’’ in ECT 3rd ed., Vol. 17, pp. 110–113, by C. Barker,
University of Tulsa; in ECT 4th ed., Vol. 18, pp. 342–346, by Colin Barker, University of
Tulsa; ‘‘Nomenclature in the Petroleum Industry’’ in ECT (online), posting date: December
4, 2000, by Colin Barker, University of Tulsa; ‘‘Petroleum Resources’’ under ‘‘Petroleum’’
in ECT 2nd ed., Vol. 14, 856–867, by E. R. Heydinger, Marathon Oil Co.; ‘‘Resources’’
under ‘‘Petroleum’’ in ECT 3rd ed., Vol. 17, pp. 132–142, by L. J. Drew, U.S. Geological
Survey; in ECT 4th ed., Vol. 18, pp. 469–479, by Lawrence J. Drew, U.S. Geological
Survey; ‘‘Petroleum Resources’’ in ECT (online), posting date: December 4, 2000, by
Lawrence J. Drew, U.S. Geological Survey.
CITED PUBLICATIONS
79. R. J. Hood, R. J. Clerc, and M. J. O’Neal, J. Inst. Pet. 45, 168 (1959).
80. J. S. Richardson and D. Miller, Anal. Chem. 54, 765 (1982).
81. St. Lanka and Sl. Hala, Erdoel Kohle Erdgas Petrochem. 11, 698 (1958).
82. S. W. Lee, S. Coulombe, and B. Glavincevski, Energy Fuels 4, 20 (1990).
83. R. C. Prince and co-workers, Environ. Sci. Technol. 28, 142 (1994).
84. J. F. McKay, J. H. Weber, and D. R. Latham, Anal. Chem. 48, 891 (1976).
85. A. Dipple, in C. E. Searle, ed., Chemical Carcinogens, ACS Monograph 173,
American Chemical Society, Washington, D.C., 1976, Chapt. 5, 245–314.
86. W. K. Seifert and R. M. Teeter, Anal. Chem. 42, 750 (1970).
87. G. N. George and M. L. Gorbaty, J. Am. Chem. Soc. 111, 3182 (1989).
88. I. A. Wiehe, Energy Fuels 8, 536 (1994).
89. R. H. Fish and J. Komlenic, Anal. Chem. 56, 510 (1984).
90. C. D. Pearson and J. B. Green, Energy Fuels 7, 338 (1993).
91. J. J. Schanz, Jr., 1978 Oil and Gas Resources—Welcome to Uncertainty, Resources
(Special Issue), Resources For The Future, Washington, D.C., 1978, 16 pp.
92. U.S. Geological Survey—U.S. Minerals Management Service, Estimates of
Undiscovered Conventional Oil and Gas Resources in the United States—A Part
of the Nation’s Energy Endowment, unnumbered report, U.S. Geological Survey
and the U.S. Minerals Management Service, Washington, D.C., 1989, 44 pp.
93. Technical data, Petroconsultants SA, Geneva, Switzerland, 1992.
94. U.S. Crude Oil, Natural Gas, and Natural Gas Liquid Reserves, 2003, Energy Infor-
mation Administration, U.S. Department of Energy, Washington, D.C.
95. Oil Gas J. 101, 46–47 (Dec. 22, 2003).
96. World Oil, 63 (Sept. 2004).
97. U.S. Crude Oil, Natural Gas, and Natural Gas Liquid Reserves, Annual Reports
1993–2002 DOE/EIA-0126, Energy Information Administration, U.S. Department
of Energy, Washington, D.C.
98. Annual Energy Outlook, Energy Information Administration, U.S. Department of
the Energy, Washington, D.C., 2005.
99. H. Enzer, W. Dupree, and S. Miller, Energy Perspectives, U.S. Dept. of the Interior,
Washington, D.C., 1975.
100. M. K. Hubbert, in Resources and Man, National Geological Sciences, National Re-
search Council, Washington, D.C., 1969, Chapt. 8.
101. D. H. Root and E. D. Attanasi, Natl. Res. Forum 4(2), 181 (1980).
102. W. S. Fyfe, M. A. Powell, B. R. Hart, and B. Ratanasthien, Nonrenewable Res. 2(3),
189 (1993).
103. World Petroleum Assessment, 2000, U.S. Geological Survey, Reston, Va.
104. N. Alazard and L. Montadert, Nonrenewable Res. 2(3), 197–206 (1993).
105. J. F. Brookout, Episodes 12(4), 257–262 (1989).
106. ‘‘The State of the Oil and Gas Industry,’’ Policy Analysis and Statistics, American
Petroleum Institute, www.api.com, accessed March 2005.
COLIN BARKER
University of Tulsa
WINSTON K. ROBBINS
CHANG SAMUEL HSU
Exxon Research and Engineering Company
LAWRENCE J. DREW
U.S. Geological Survey
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 43
EXPLORATION, DRILLING,
AND PRODUCTION
ENGINEERING
1. Origin, Deposition Mechanisms, Exploration, and Occurrence
of Crude Oil and Natural Gas
1.1. Crude Oil. Crude oil is the name given to all organic compounds
which are liquid under reservoir conditions. They can partly solidify at the surface
after expansion and cooling. Petroleum is a mixture of hydrocarbons in varying
proportions, which are accompanied by compounds of sulfur, oxygen, phosphorus,
and nitrogen. Minor constituents are porphyrins, ash-forming metal compounds
(usually sulfides of vanadium, nickel, copper, cobalt, molybdenum, lead,
chromium, and arsenic), as well as inorganic salts, hydrogen sulfide, and – as a
typical constituent of ‘‘crude oils’’ – water in varying amounts. The differing
mixing ratios preclude uniform chemical and physical properties. Elementary ana-
lyses show average carbon contents of ca. 79.5 – 88.5 wt %, hydrogen contents of
ca. 10 – 15.5 wt %, and impurity contents of up to a maximum of 5 wt %. Most
crude oils are optically active.
The essential ingredients of the petroleum mixture are alkanes, naphthenes,
and aromatics. The oxidation products naphthenic acids, terpenes, and phenols
are also present in some cases. According to the predominant constituents, a
distinction is made between paraffin-based crude oils and naphthene- or
asphalt-based crude oils. If both hydrocarbon types are contained in considerable
amounts in crude oil, it is termed mixed-based oil. The composition of crude oils on
a worldwide basis includes more than 30 % paraffins, at least 40 % naphthenes,
and ca. 25 % aromatics. Naphthene-based crude oils – that contain resinous and
asphaltic substances – are frequently found in the upper level of the Earth’s
crust, mixed-based oils in the middle, and paraffin-based oils in the deeper zones.
Near the surface, the crude oils lose part of their volatile constituents. Dur-
ing this, with chemical processes participating, more or less solid residues are
formed. Mainly ozokerite (earth wax) is formed from paraffinic crude oils and
natural asphalts from naphthenic crude oils.
For decades, petroleum has been mainly used as raw material for the pro-
duction of engine fuels, lubricants, and fuel oil. Because of the complex composi-
tions of crude oils, the chemical industry has developed many processes through
which petroleum has become the basis of numerous plastics. In 1990, more than
10 % of world petroleum production was used in petrochemistry.
Formation. Despite a century’s discussion with many contradictory opi-
nions and intensive research, the formation of petroleum can still not be satisfac-
torily explained. The discussion on the inorganic formation of crude oils continues
to this day. Russian experts, in particular Porfirev (56) do not exclude the fact that
petroleum is formed from inorganic matter (metal carbides and superheated
steam) even though it is generally agreed that the crude oils are predominantly
of organic origin. The investigations of potential petroleum source rocks in the
past decades have contributed considerably to this opinion. All deposits in the sedi-
mentary basins contain organic residues of terrestrial, limnic, fluvial, and marine
3) The ascent of marine currents from greater depths to the surface leads to the
substantial consumption of the scarce oxygen present in these currents as a
result of the biological productivity near the surface. From the organic resi-
dues that sink in the deep water lacking in oxygen, source rocks have been
formed. Examples of this occur off the Southwest African coast (Benguela
current) and off the west American coast (Humboldt current off Peru).
4) Source rocks are formed in the open ocean as a result of global climatic warm-
ing in association with the large transgressions. The many thick source rocks
in the Jurassic and Middle Cretaceous periods are attributable to this.
According to (60), the temperatures necessary for the genesis and maturation of
the petroleum are closely related to the geological time available for them.
Thus, according to Russian ideas, ca. 40 – 50106 a are required for the start
of petroleum formation at an earth temperature of at least 60 8C. The Californian
deposits mentioned, though, have formed at temperatures above 100 8C in only
1 – 2106 a (62).
Formation of Deposits. Deposit formation is the concentration of infi-
nite numbers of oil droplets which form in source rocks and are then squeezed
out of them to economically exploitable accumulations. Deposit formation is the
result of oil migration. The oil particles move during primary migration from
the argillaceous source rocks into the porous reservoir rocks, which usually
lie at a higher level. A considerable fractionation of the crude oils can occur
even during this movement as a result of selective filtration, adsorption, and
condensation (63,64).
Migration. In many cases, however, the true concentration of deposit
formation occurs only through migration in the bedding or horizontal secondary
migration, ie, through movement and accumulation within the permeable
reservoir horizons. With favorable facial conditions, lateral migration can be
very important for deposit formation. It can occur over distances of 10 km to a
maximum of 100 km. With changing facies, differential migration can occur, by
which the lighter hydrocarbons migrate to the higher reservoir beds. During the
course of this, the oils experience a certain buoyancy in relation to the formation
water present in the reservoir rocks because of their lower density. The oil is
therefore concentrated in the highest zones of the reservoir horizons, which
are sealed above by mudstones or salt beds. A gravity differentation of petro-
leum over edge water within the permeable reservoir rocks then occurs which
corresponds to the different densities of the individual components. Provided
gravitational separation is not prevented by capillary forces due to small pore
diameters, the edge water limits the oil column to the botom of the reservoir
(oil – water contact). If the oils are supersaturated with gas, a gas cap forms
over the petroleum.
Possible reservoir rocks are permeable and porous sandstones and lime-
stones. They have average porosities of ca. 10 % to a maximum of 40 %. The qual-
ity of reservoir rocks is determined, aside from their pore volume, by their
permeability, which is of decisive importance for the production capacity of wells.
Petroleum and gas deposits can occur in all geological formations, starting
from the Cambrian period, provided reservoir rocks are available that are accom-
panied or underlain by source rocks of appropriate maturity. Also the reservoir
rocks must possess trap structures from the overlaying of sealing horizons,
such as clay and salt beds. It is also important that the structural traps were
already present at the time of oil formation and survived during the subsequent
geological periods.
Reservoir Rocks. The quality of deposits is greatly affected by the type of
reservoir rock.
Wind-blown sands, for example, form the good gas reservoir beds
of the Rotliegendes. They spread from England to Poland over an area of
1000 – 3000 km2 with thicknesses of up to 200 m and maximum porosity
up to 20 %. Fluvial sands of meandering rivers can have net thicknesses of
48 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
6– 75 m. The Triassic sands of Prudhoe Bay with reserves of 2.8109 t oil belong
to this type. Both alluvial fans such as the Cretaceous sands of the Sirte Basin
in Libya and the beach and barrier sands parallel to the coastlines are good
reservoir beds. The latter include the Jurassic Piper field (United Kingdom).
The marine platform sands, which occur as transgressive or regressive
sequences, are equally good reservoir beds. Deep-water sands, occurring as
so- called deep-sea fans and turbidites, certainly spread over hundreds of kilo-
meters, but are of variable quality. This reservoir type includes the Pliocene
Sands in the Los Angeles Basin with a thickness up to 700 m and the Paleocene
Sands of the Viking Graben in the North Sea.
Carbonate reservoir beds can be of very differing quality. Thus reefs or reef
limestone debris can have excellent reservoir properties as in West Texas or
Alberta/Canada, with net thicknesses of 170 – 200 m, but they are mostly of lim-
ited extent and variable quality. The chalk forms only average-quality reservoir
beds. Mainly composed of coccoliths, the chalk is thin-bedded. The production
engineering of the Austin Chalk in West Texas, the Ekofisk field in the Norwe-
gian North Sea, or the reservoir of Reitbrook (Hamburg) is difficult. Owing to
their low permeability, these deposits can be extracted profitably only in conjunc-
tion with very fine fissuring.
Trap Structures. The type of trap structure encountered strongly
depends on the worldwide or regional tectonic processes in the Earth’s crust.
Movements at the plate edges lead to individual stress patterns with a variety
of structural phenomena as the result of forces of contraction and expansion.
The structures can, therefore, be of various sizes and shapes (65,66). The prin-
ciple trap structures, as listed below, can also occur in combination.
The types of structure possible as traps for petroleum and natural gas can
be of very different origin (see Fig. 2). There are anticlines (a), tiltings, faults (d),
salt intrusions, and discordant overlappings, and sometimes synsedimentary
facies differentiations which arose during the deposition of the beds. Permeabil-
ity barriers, arising as a result of secondary recrystallization and the formation
of new minerals in the pores, laterally seal the reservoir in such a way that a
special form of the facies reservoir is formed. This can often be observed in con-
junction with faults. The most important tectonic traps are the saddle-shaped
anticlines and flat, oval mound structures as well as monoclines originating
from antithetic faults. Reservoirs in association with overthrusts are usually
arched in the frontal zone and sometimes enrolled. They are known as rollover
structures. The first reservoirs discovered were found in association with salt
domes (d), where various kinds of structural traps are possible on the flanks
and the top of the dome. In the case of discordance reservoirs (c), porous horizons
are sealed off after tilting and partial removal of horizontally deposited younger
sediments. Other typical facies reservoirs are the reefs (b), which are usually
composed of corals, as well as traps which have arisen, eg, through the argilla-
tion of a sandstone with reservoir characteristics. In addition, there is a variety
of special forms and combined reservoir types. In all of these cases, it is common
that porous reservoir horizons in their top zone are covered and transposed with
sealing by clay rocks and salt layers.
Oil Exploration. The potential for petroleum and natural gas present in a
sedimentary basin must be proved by exploratory wells. The total costs of the
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 49
Fig. 2. Most important reservoir types and trap structures a) Under an updoming (anti-
cline); b) In a coral reef; c) Under overlapping strata (discordance); d) On a salt plug tank;
e) In a conformable downthrown fault; f) In a reverse downthrown fault.
magnetic minerals within them. Sediments, on the other hand, differ in their
susceptibility according to the proportion of such minerals in the strata. The
changes of the Earth’s magnetic field are recorded with the aid of proton magnet-
ometers and give, for example, indications of the shape and depth of the crystal-
line basement or of volcanic or magmatic intrusive bodies. Aeromagnetics was
developed in World War II to detect submerged submarines. The measuring
instruments are accommodated in the wingtips of the aircraft or towed behind
on cables. In their basic version, the measuring instruments are coils through
which current flows. The applied magnetic field is altered by the terrestrial mag-
netic field and the resulting anomaly is recorded.
Gravimetric Measurements. Gravimetric measurements are used to
measure the changes in the Earth’s gravitational field. Because of the different
densities of rocks, beds of relatively low density, such as salt rocks, with
negative anomalies can be demarcated from rocks of higher density, such as
intrusives or quartzites, with positive anomalies. As a result it is also possible
to detect blocks of identical rocks that are displaced tectonically by faults.
The typical measuring instrument is the spring balance, whose spiral spring
shows a different extension according to the density of the rocks. The
extension is recorded and interpreted (Fig. 3). Aerial gravimetry, combined
with aerial magnetics, is a rapid and cheap method of reconnaisance survey
for the petroleum exploration of extended areas.
Gravimetric methods are increasingly used again for the solution of special
problems. Computer model calculations can be carried out which lead to consider-
ably improved conclusions. Thus gravimetric measurements are an important
means for detecting petroleum reservoirs at the flanks of salt domes; for the defini-
tion of the drilling and casing program and the allowed deviation for deep gas wells;
and also for the interpretation of seismic surveys near salt domes, where the reflec-
tions are disturbed by the salt, in order to clearly determine the salt flanks.
Geoelectric Measurements. Geoelectric measurements of the natural
Earth currents (tellurics) in the form of magnetotellurics have not fulfilled
the hopes placed in them. The numerous interference factors in the industrial
countries, such as electrical energy in cross- country transmission lines and
on railway lines are as disadvantageous as the large spread of the results.
Magnetotelluries are therefore only rarely used.
Seismic Methods. Since the 1970s seismic methods have become the most
accurate and most frequently used exploration methods. This is mainly due to
digital recording and the many processing and interpretation programs avail-
able. If in the past, seismics was a subsidiary science of geology, today it must
be ranked of equal importance alongside it. The seismic methods permit very
accurate conclusions on the stratigraphic sequence and depth of geological
beds. These methods measure the transit time of artificially generated elastic
waves. The waves have usually been generated by the detonation of explosive
charges placed in boreholes (Fig. 4). Apart from this, other energy sources also
are used nowadays. The vibroseis method, has become established whereby a
steel plate mounted below a truck is pressed by the weight of the truck onto
the ground. This transmits controlled vibrations of a given frequency that are
produced by a vibrator installed in the truck, to the subsoil. Four to six vibrator
trucks operate simultaneously. With this method, measurements can be
52 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
performed on roads and tracks and therefore in populated areas. It is also more
rapid than explosion seismics, and cheaper.
The elastic waves spread in all directions and at bed boundaries, eg,
between claystone and sandstone or sandstone and carbonate. They are partly
refracted and partly reflected to the surface by the echo sounder principle.
The velocity of these seismic waves through the various geological strata is a
measure of the type of rock and depends on the rocks’ elasticity and density
53
Fig. 4. Reflection seismic surveying technique with two reflecting horizons, A and B.
54 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
(ca. 1000–2000 m/s in loose rocks and up to 6000 m/s in compact rocks). The more
accurately the actual velocity in each bed is known, the more accurately can the
depth of the reflection be calculated. Computer model calculations provide very
useful approximations. There are two different methods: the reflection seismic
method and the refraction seismic method, the latter was used almost exclusively
until the 1930s. The refraction seismic method is still used nowadays only for
wide-area surveying of deep-lying horizons, but is also called in for the determi-
nation of sonic velocities in the rocks and to solve special problems.
Seismic Refraction Method. In the seismic refraction method a part of
the energy waves is refracted at the refraction horizon, similar to light being
refracted at the interface between water and air. Some of the refracted waves
then continue to run along the bed boundaries and in the course of this generate
other waves that are refracted again to the surface. The measurement is of the
transit time from the detonation at the energy center until the refracted wave
impinges on the geophones set up on the surface at distances of up to ca.
25 km. The seismic refraction method requires large explosive charges with the
geophones at a relatively large distance and enables a relatively large depth of
penetration.
Seismic Reflection Method. The seismic reflection method uses, in con-
trast to the refracted wave, the waves directly thrown back by the reflecting hor-
izons. Depending on the depth of the reflecting horizons, the waves return to
the Earth’s surface after increasing time intervals and are received by groups
of geophones systematically laid out on the ground. They are then converted to
electromagnetic impulses and recorded on magnetic tapes in complicated digital
recording equipment. Recording the seismic signals on magnetic disks – having
already been pretreated in the recording truck by amplification, filtering, and
balancing of the vibrations – enables corrections to the raw data and various
investigations of the signals with simultaneous suppression of the interference
energy (noise) to be carried out at the geophysical playback centers. From the
transit time found for the emitted and reflected waves, the depth of the reflecting
bed can be determined, provided the velocity of the rock beds passed through,
which often shows rapid regional variations, is known. The depth is then half
the product of velocity and transit time. Indications of the velocity of the seismic
waves come from detailed calculations of the refraction and reflection seismic
methods, from seismic well-logging, and directly from a geophone being lowered
on a wine line in the bore hole. The geophone is positioned at the boundaries of
rock units and receives the waves produced in the usual manner off the well
site.
Aside from the compression waves which have up until now been used
almost exclusively for scanning the structural conditions in the subsoil, shear
waves also are gaining increasing importance today. They enable conclusions
to be made on the petrographic nature of the rocks and their contents. The com-
bination of compression and shear waves permits at least approximations to the
porosity of a reservoir rock and to the pore content.
Seismic Measurements in Offshore Regions. The reflection seismic mea-
surements can also be used in offshore regions. The seismic method based on
explosives previously used, has been abandoned and replaced because of
the great environmental damage, especially where a large fish stock exists; at
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 55
the same time the airgun method has established itself. In this method highly
compressed air is suddenly released. With the large number of available record-
ing vessels, seismic companies not only carry out commissions to order but, in all
seas, carry out self-financed measurements which they then offer for sale to the
oil companies for reconnaisance purposes.
Three-Dimensional (3 D) Method. In exploration seismics, the line
method has hitherto been used almost exclusively. According to the exploratory
problem definition, individual profiles spaced at a distance of 500 – 2000 m from
each other are intersected by crosslines having the same spacing. The 3 D
method, which is almost routine in the development of proved deposits, is also
being used more and more in exploration. The groups of geophones are placed
at profiles with distances of 25 m and the waves are generated by explosives
fired from various sides. The digitally recorded results are processed and
corrected, as in the line seismic method, and subsequently transmitted to work
stations. Because of the small distances between profiles and the possibility of
assembling profile lines according to one’s own ideas via shot points of neighbor-
ing lines (random line), the interpretation of the reflection horizons leads to
increased precision. Because of the high cost of equipment – several thousand
geophones are required for the geophone installation – and the large number of
staff used – up to 350 members – the method is very expensive. It pays, however,
especially during development of a productive area, since as a result of the close
surveying, false interpretations and consequently false structural descriptions
are restricted or even avoided. Characteristics of a 3 D measurement are given
below:
facies distribution, but also the settling and the related rise of temperature, as
well as the uplift phases with erosion of reservoirs already formed. A three-
dimensional kinetic model has attracted attention (69). It is based on a set of dif-
ferential equations for energy and mass balances that describe the course of var-
ious geological processes. The physical parameters, such as porosity and
permeability, and the thermal sequences are selected from available lithological
data and, as knowledge increases, can be adjusted iteratively.
Occurrences, Production, and Reserves. The petroleum occurrences
are associated with the sedimentary basins of the Earth. Their distribution is
in no way statistical. Even their tectonic classification is very diverse and
shows that deposit formation is not associated with only certain types of basin
(Fig. 5). The largest petroleum occurrences lie in basins that have subsided con-
siderably for instance as a result of the continental drift – approximately from
the Jurassic to the Middle Tertiary period – and, owing to a lack of contact
with oxygen-rich waters of polar origin and to a simultaneously warm climate
with high, oxygen-reducing water temperatures, have preserved large amounts
of organic material in the anoxic medium. The reserves in older geological forma-
tions are relatively small. The reason may be that there were fairly small
amounts of organic material in the shallow seas of the epicontinental region dur-
ing the flooding by the worldwide transgressions. But the smaller reserves in
older formations than those of the Mesozoic era may also be the result of thermal
evolution (conversion of oil into gas), the destruction of the reservoirs or their
sealing horizon by tectonic events and erosion. The sedimentary basins with pro-
spects are nowadays distributed both over the continents and over the regions of
the continental shelves covered by the seas (Fig. 6). The methods of exploration
and development are in both cases the same, for offshore wells the means of
transport and technical installations are adapted to the water medium. The dif-
ference lies in the greater costs for offshore wells and production installations.
Apart from the classical offshore fields in Lake Maracaibo (Venezuela) and
in Baku (Caspian Sea) that were already developed in the 1930s with modified
onshore installations, offshore exploration and production achieved increasing
importance only in the 1960s. Technical developments then became necessary
58
Fig. 6. The sedimentary basins of the Earth with the most important production areas.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 59
that enabled deposits to be developed even at a depth of more than 1000 m water.
In 1983, on the west coast of France, a well was drilled to almost 2000 m at a
water depth of 1714 m. Such depths of water cause difficulties for production,
which then is only possible by means of an underwater completion. This
means that the production equipment is totally installed on the ocean floor
and must be maintained by divers. The deepest producing field is the Jolliet
field in the Gulf of Mexico. It produces from a depth of 535 m of water and is
equipped with a tension leg platform, ie, a floating platform that is fixed to the
ground by steel wire ropes. In 1992 the Snorre field in the Norwegian North Sea
also went into production by means of a tension leg platform at a depth of up to
350 m of water. Further production installations according to the tension leg sys-
tem are being prepared up to ca. 1000 m (Auger and Marlin fields). Owing to high
investment costs, the systematic recovery of hydrocarbons at great depths of
water, in particular on the continental slope, depends on the development of
the market price of crude oil.
Production and Reserves. The prices of crude oil collapsed in the second
half of the 1980s. OPEC saw a possibility of stabilizing the prices only by fixing
production quotas. These efforts, however, had no particular success. The control
of production by OPEC as well as other economic measures in producing coun-
tries outside OPEC, influence production in such a way that the annual produc-
tion data do not give a clear picture of the productivity of a producing country.
Rather, the cumulative production gives only an impression of the geological and
technical capacities (Table 1). For better comparison, the annual production
for 1989 for the region concerned is indicated. World oil production for the
year 1989 was 3.434109 t. In 20 years, it rose only by ca. 20 %. In view of world-
wide industrial development this is only a moderate increase that is attributable
in part to energy saving efforts which started with the 1973/1974 energy crisis.
The ranking of the producing countries for 1989 shows that the Middle East
states Iran, Iraq, Kuwait, Saudi Arabia, and the United Arab Emirates lead
the list of producing countries with a production of 922106 t, followed by the
Soviet Union with 668106 t, and the United States and Canada together with
525106 t. Following behind are South and Central America (380106 t), Asia/
Oceania (365106 t), Africa (350106 t), and Europe (223106 t). The production
from offshore fields was about 25 % of world oil production (70). The proved world
oil reserves rose to ca. 164.30109 t through development, in particular, of off-
shore fields. Compared with reserves published in the 1970s (92.40109 t) this
is an increase of > 80 %. If the probable reserves determined according to the
Gaussian distribution curve are added, then there are currently 240109 t petro-
leum to be expected. Compared with the production of ca. 3.4109 t in 1990, this
gives a ratio of 72 : 1. Thus the ratio of 30 : 1 existing for decades has more than
doubled. Even taking only the proved reserves into account, the ratio is about
50 : 1. There may be various reasons for this. One could be the rapid sequence
of discovery and development of large deposits – above all offshore. These include
new discoveries in East Venezuela and in the Brazilian Campos Basin, as well as
in the Russian North Caspian Basin and in Saudi Arabia. Large reserves have
recently been developed in the Tarim Basin (West China) and in the Papua Basin
of New Guinea, in which up until now no economic oil discoveries have been
known. Another reason is certainly the nature of the recording or evaluation of
60 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Table 1. The most important crude oil producing countries; their total production up to
December 31, 1989 and their resources (assembled from various international data)
Resources
Bahrein 127 30
Iran 5898 11 750 3 020
Iraq 3 323 16 850 5 560
Kuwait 3 800 13 700 320
Neutral Zone 713 2 200 320
Oman 493 1 100 150
Qatar 684 650
Saudi Arabia 9 157 41 500 7 950
Syria 207 300
United Arab
Emirates 1 860 10 500 800
Others 207 750
Table 1. (Continued)
Resources
the proved reserves. The new assessment is based on the newly proven or re-
evaluated reserves in the oil fields, larger extent of the deposits, or improved
oil recovery. At the World Petroleum Congress in 1990, a rise of the reserves
by 45109 t to more than 164109 t, was reported, ie, a rise of ca. 33 %. Of
this, ca. 32109 t are apportionable to the Middle East and only 3109 t to
North and South America. Other authors quote a larger or smaller rise than
this. The reason for this may be that, in addition to the newly developed reserves,
new evaluations of the fields have been made in which, apart from the proved
reserves, probable and possible reserves have been accounted for. The reserves
according to MASTERS et al. (71) listed in Table 1 represent, despite all uncertain-
ties, a reliable order of magnitude. The size of the increase in reserves from new
discoveries also reflects the number of successful exploration wells. As stated
above, the saturation of the petroleum market since 1985 has led to a drastic
reduction of exploration activity. If in 1984 in the western countries, 1250
seismic crews and 6500 drilling rigs were in use, their numbers had fallen in
1989 to 447 and ca. 1924, respectively (1990: 2057, 1991: 1896). A speeding up
of exploration depends on the general improvement and stability of prices and
availability of reserves in the important consumer centers. The further pros-
pects for success in petroleum exploration are limited to where the assumed
profitability for foreseeable future ends or when the natural deposit potential
of liquid hydrocarbons in the subsoil of the earth is exhausted. The two aspects
are closely interlinked, since the estimation of potential reserves and even the
definition of a deposit are determined by profitability factors. The ca. 77109 t
of conventional crude oil still to be expected according to the most recent esti-
mates must be supplemented by the reserves from very heavy oil deposits and
oil (tar) sands (see Tar Sands), which can be assumed to be at least 100109 t.
An accurate demarcation from the ‘‘conventional’’ crude oils listed in Table 1
is not possible, since MASTERS et al. (71) also take into account the reserves
between 108 and 208 API, which other authors include in the low-quality heavy
oils. The heavy oil and bitumen deposits are found to the extent of 87 % in Canada
62 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
(Athabasca, Cold Lake, Peace River), the CIS (Volga – Urals, East Siberia), and
Venezuela (Orinoco). They represent a future reserve that can only be used
when, because of the depletion of the conventional fields, so- called tertiary recovery
methods are standard, eg, hot water injection or steam injection, to extract the oil
from the deposits. The prices by then must have exceeded the limit of $ 50/bbl (the
oil price in 1992 was ca. $ 20/bbl).
Regional Distribution of the Reserves. Middle East – Arabian (Persian)
Gulf. The largest reserves are in the countries of the Middle East and in the
Arabian Gulf. On 1st January, 1990, proved reserves of ca. 99109 t were estab-
lished and probable ones of 19109 t calculated (see Table 1). The first economic
oil find in recent times in this region was made in 1908 with the discovery of the
Iranian Masjid-i-Sulaiman field. Many subsequent discoveries have been added.
At present, Saudi Arabia, with more than 41.5109 t of proved reserves, followed
by Iran with 11.75109 t, Kuwait with ca. 13.7109 t, and Iraq with ca.
16.85109 t are leading. The largest deposits in the Middle East include the fields
of Burgan in Kuwait, Ghawar, and Safaniya in Saudi Arabia, Kirkuk and
Rumaila in Iraq, and the deposits of Gach Saran and Agha Jari in Iran.
Geologically, the oilfields of the Arabian Gulf are situated on the shelf area
in front of the pre-Cambrian massif outcropping in the west, and south of the
Arabian peninsula in the form of intensively block-folded crystalline. Continen-
tal and shallow-water sediments are found deposited between the Cambrian and
the Cretaceous systems on the slowly sinking shelf. The shelf is divided by
numerous North – South trends that are linked with basement highs. North
and east of the shelf, including the Taurus – Zagros – Oman mountains there fol-
lows a geosynclinal or plate-edge region. This region also received thick tertiary
sediments and during alpidic orogenesis was folded by tangential forces and
divided into structures with a northwest – southeast direction. In accordance
with the asymmetric structure of this trough, the large deposits on the Arabian
side mainly lie in Jurassic and Cretaceous beds, and the offshore deposits mainly
in Cretaceous reservoir rocks. The Persian – Iraqi fields produce in the more dee-
ply settled parts of the trough from Cretaceous and Tertiary horizons. On the
Arabian peninsula, anticline-like updomings dominate structurally, whilst in
the Persian Gulf salt structures and bending folds, and in the foothills of the Ira-
nian Zagros range anticlinal and overthrust structures can be found. The most
recent exploration activities in Saudi Arabia proved deposits with several hun-
dred million tonnes of oil in sandstones of the Paleozoic era. The Permian
Khuff, itself a reservoir bed of natural gas deposits, serves here as capping of
stratigraphic and structural traps. The oil is of Silurian origin. The Paleozoic
era could provide further potential in Saudi Arabia.
Europe. The petroleum deposits of Europe, apart from the North Sea,
are fully developed and, with production falling, are already in the phase of sec-
ondary and tertiary recovery methods. If the reserves of all onshore producing
countries are combined, proved reserves of ca. 0.7109 t can be established
together with exploratory reserves of ca. 0.5109 t. The United Kingdom and
Norway possess proved reserves of ca. 5.325109 t from offshore fields, and a
further ca. 2.34109 t exploratory reserves are expected. The comparison with
the Middle East, however, shows the small importance of the European states
as oil-producing countries. Even so, in 1974 the Federal Republic of Germany
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 63
produced more than 6106 t oil and was able to cover ca. 30 % of indigenous
requirements. Of the 823106 t oil reserves originally present, 61106 t proved
and probable oil reserves remained after a total production of 225106 t (Decem-
ber 31, 1991). The recovery factor of 35 % of the original reserves means that
537106 t remain unrecoverable. The reserves in Italy and the Netherlands,
with developed quantities of ca. 110106 t in each case and prospective reserves
of ca. 330106 t in Italy, are – on a worldwide view – of equally low importance as
those of Romania, with 190106 t and a prospective reserve of 235106 t. All
other countries on the European mainland are globally considered insignificant
as producing countries. The oil deposits of northern Germany are bound to
sediments of epicontinental floodings which occurred during the Jurassic and
Cretaceous periods. The narrow NNW – SSE striking troughs filled with Jurassic
deposits are caused by halokinetic movements. The Rhine Valley Graben as a
minor oil region dates from the ‘‘rifting’’ (continental drift) related-to young
tectonic movements. The European deposits are distributed over the foredeeps
of the Pyrenees, the Alps, and the Carpathians, although the deposits in the
Aquitaine Basin, the Molasse Basin (Bavaria), the Carpathian Basin, and
the Po Basin are not very large. The oil occurrences in Sicily, Spain, and in
the Aquitaine Basin of France are linked with Triassic, Jurassic, and Cretaceous
carbonates. The intramontane Pannonian Basin and the Paris Basin should be
mentioned because of their historic importance for the exploitation of oil in
Europe. The Lower-Saxony Basin also is of historic importance, since the
deposits of Ölheim (1881) and Wietze (1859) were discovered here almost
simultaneously with Colonel Drake’s success in the United States.
North Sea. The North Sea has been one of the most attractive exploration
regions for the international oil industry since the start of exploration in 1964.
All large oil companies, often together with recently founded local partners,
have introduced their worldwide know-how into development of this logistically
difficult region. Great depths of water, strong currents, and extremely harsh
weather conditions require the use of specially developed drilling and production
installations. Although the technology used can nowadays be regarded as
advanced, the damage to the platforms during construction and operation
shows the limits for the prospecting and production of hydrocarbons. The first
success in the North Sea was the discovery of the Ekofisk field on the Norwegian
Shelf in 1968. Two important successes in the British sector of the North Sea
were the discovery of Forties in 1970 and of Brent in 1971. About 17 oil fields
are in production in the Norwegian North Sea, and in the British North Sea
more than 100 oil and gas fields including numerous small ones. Norway’s lar-
gest fields are Ekofisk/Eldfisk, Gulfaks, and Statfjord, which have up to now pro-
vided ca. 80 % of the Norwegian oil production. In the United Kingdom the
Forties, Brent, Magnus, Ninian, Fulmar, and Statfjord (British part) fields pro-
vide more than 40 % of the British oil production. The five Danish and seven
Dutch offshore fields are negligible compared with those of Norway and the
United Kingdom. Geologically, the fields are linked with the large graben that
open up from the beginning of the Mesozoic era in conjunction with the rifting
of the Atlantic plate. The most important element is the Viking Graben which
collapsed between the East Shetland Platform and the Norwegian Vestland
Bogen. It runs for a length of ca. 455 km and has a width of ca. 50 km in the
64 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
south and 180 km in the north approximately on the midline between Norway
and the United Kingdom. In addition, there are special basins, some with
thick tertiary sediments, to which the oil is bound. The deposits are found in
sediments of the Triassic, the Jurassic, the Upper Cretaceous, and the Paleogene
periods. The traps are mostly the tilted fault blocks typical of rifting, that are
sealed by transgressively overlapping fine-grained Upper Cretaceous sediments.
The development of hydrocarbons is limited to the north by the 62nd degree of
latitude as a result of Norwegian legislation. Seismic measurements and specifi-
cally located wildcats have provided evidence that in the northern North Sea
(Tromsø, Haltenbanken) and in the Barents Sea, deposits both of oil and of nat-
ural gas can be expected. The time for specific development of these fields will
depend on the development of the reserves already developed and also on the
oil price. Even today, owing to expensive technology, a price in excess of $ 20/
bbl and tax relief by the producing countries is necessary for the continuation
of exploration in the North Sea. In 1991, the daily production of all North Sea
fields reached 192106 t.
Commenwealth of Independent States. Industrial oil recovery in the
territory of the CIS had already started at Baku in the middle of the last century
and reached its first peak around 1900, when North American production was
briefly surpassed. After a period of stagnation, a rapid upward development
started in the former Soviet Union after World War II with the development of
new oil reserves. The proved resources on January 1, 1990, were 13.2109 t, and
14.6109 t of exploratory reserves were expected. Accordingly, the total reserves
are more than 200 % higher than those of the United States. Until 1918, almost
the whole production came from the Caucasus mountains. In the 1950s the
Volga – Urals region and the Timan – Petschora Basin, adjacent to the north,
were developed. In addition, the discoveries in the North Caspian Basin
(Emba) and in the Mangyschlak trough of Turkmenia developed very rapidly,
but have passed their peak. In the 1960s and 1970s, the extended deposits
east of the Urals in the West Siberian Basin were discovered. Meanwhile 96 %
of the oil developed in West Siberia is produced in the Tjumen province. In the
southern provinces of Novosibirsk and Omsk, exploration has given disappoint-
ing results. In almost all oil provinces, as a result of technical deficiencies and
lack of oil field equipment, production is falling or remaining below expectations.
Proved reserves could be mobilized by the use of modern technology, such as the
horizontal drilling process in many reservoirs of low permeability. The opening of
the CIS to the West leads to the expectation of a strong involvement of interna-
tional oil companies which, with their know-how and their modern technology
can bring oil exploration and exploitation to a reliable high standard. There
are also justified expectations of the development of important new reserves in
Turkmenia and East Siberia, where only a small production of oil has been
achieved so far. Large oil discoveries in the pre-Caspian Depression promise
interesting exploration possibilities. In the recently discovered Thenguise,
Astrakhanskoe, and Karatshaganakskoe fields, oil is produced from carbonates
and clastics of the Paleozoic era at depths of more than 4000 m below a Permian
salt cover. With an area of 500 000 km2 and a sedimentation thickness of almost
20 000 m, good possibilities also arise here which may increase the number of
reservoirs in the CIS substantially. Geologically, the deposits in the Caucasus
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 65
mountains, the Caspian Sea, Galicia, the Fergana depression, and on Sachalin,
etc., are linked above all with Neogen horizons in prealpine depressions and
innermontane troughs. The deposits of the Volga – Ural area lie in Paleozoic
depressions of the Eastern – European Platform. The main reservoir rocks
here are Devonian and Permian reef limestones as well as sandstones of the Car-
boniferous period. In the large and oil-rich Timan – Pechora Basin, the reservoir
systems belong to Upper Devonian sandstones and chalks of the Lower Carboni-
ferous period. Mainly Permian and Mesozoic reservoir horizons are oil-bearing in
the deposits of the North Caspian Depression, in Kazakhstan, and Turkmenia.
In the deposits of the Western Siberian Platform, the hydrocarbons are assigned
to the central part of the Lower or Middle Cretaceous period. In the Molasse
Basin of the eastern Ural foreland, oil and gas deposits are linked with Upper
Jurassic sandstones and calcarenites.
Africa. In 1990, the proved and probable oil reserves of Africa were about
11.3109 t, and the exploratory ones ca. 5.0109 t. The resources are concen-
trated in Libya, Algeria, and Nigeria. There are further deposits of noteworthy
size in Egypt, Congo, Angola, Gabon, and Tunisia. The African continent was
long considered to have few prospects for hydrocarbons. The breakthrough
came in the 1950s with the almost simultaneous discoveries of the first large
deposits in Algeria and the Niger delta. The large discoveries in Libya followed
in the 1960s. Geologically, the oil deposits of Algeria and the smaller occurrences
in West Libya which are not yet producing, are linked to the Paleozoic basin
between the Precambrian Massifs of the mid-Sahara and the Atlas range. Fairly
large amounts of oil are found in the Neogene basins of the Syrte in Central
Libya and in the Niger delta. There are also older oil-producing strata, mostly
in the Cretaceous system. At the western edges of the African shield, between
several stable cratons, depression fields have been formed, on which marine
transgressions deposited thick sediments in the Mesozoic and Tertiary eras. In
some cases, between the Permo-Triassic system in Morocco and the Cretaceous
system in Gabon, salt deposits occurred that led, possibly in conjunction with the
continental drift, to diapirism. The basins not only cover parts of the current
mainland but reach far into the Atlantic. The coastal basins range as follows
from north to south: Spanish Sahara, Senegal – Mauritania, Ivory Coast,
Dahomey, Nigeria, Gabon, Congo, and Angola. In Eastern Africa, of the graben
systems which are largely linked to the rifting, only the Red Sea graben with
Miocene salt deposits and hydrocarbons below is of importance, as in the Gulf
of Suez, which is the only area of East Africa with fairly large reserves.
Libya, with 4.9109 t proved and probable reserves and ca. 1109 t pro-
spective, has the largest oil reserves of Africa. Cumulatively, ca. 2.6109 t oil has
been produced in Libya up to now. Geologically, Libya is divided into three large
basins: in the west the Murzuk basin, in the central part the Syrte basin, and in
the east the Kufra Basin. The oil originates for the greater part from reservoir
rocks of the Cretaceous (sandstones) and the Lower Tertiary (reef limestones)
systems, but also from the Cambrian and Ordovician systems of the Syrte
Basin. Libya’s only producing offshore field Bour, with a daily production of ca.
13 000 t (1991), has little importance. The development of hydrocarbons in North
Africa began with the discovery of the Zelten (now Nasser) field in 1959. Nasser
produces from Lower Eocene limestones of the Syrte Basin. With original
66 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
reserves of about 400106 t, it belongs, like Amal and Sarir, with ca. 1.2109 t, to
the largest fields of Libya. Libyan oil is low in sulfur and in great demand as a
light crude in European refineries.
Nigeria is the second largest producing country in Africa. The proved and
probable reserves are indicated as 2.8109 t and the prospective reserves as
1.1109 t oil. The fields are in the Niger Delta Basin that is filled with ca.
12 000 m of clastic rocks of the Tertiary period. In Nigeria, 15 giant oilfields
(according to U.S. definition, fields with more than 500106 bbl, ca. 80106 t)
have been proved, of which almost half lie offshore.
Algeria, with proved and probable reserves of ca. 1.4109 t, occupies the
third position among the producing countries of Africa. The prospective reserves
(ca. 240106 t) are quite insignificant. The oilfields lie in the Polignac Basin, near
the Libyan border, and in the Ghadames Basin. The fields in Polignac produce
from reservoirs in the Carboniferous, Devonian, and Cambro – Ordivician sys-
tems. In the Ghadames Basin likewise there are sandstones of the Cambro –
Ordivician, that are discordantly overlaid with evaporites.
North America. Geologically, the North American oil deposits are linked
with the large sedimentation region that appeared repeatedly since the Paleozoic
era. This basin forms the foredeep to alpine-type mountains in the west (Rocky
Mountains) and flattens to the east towards the basement area of the Canadian
shield. In the south, this sedimentary basin expands and becomes the thick Ter-
tiary Gulf basin. A lateral arm of the trough zone with Paleozoic rocks follows the
south-east edge of the Canadian shield to the mouth of the St. Lawrence river. In
accordance with this geological structure, the deposits in the more deeply sunken
western sections of the trough appear in Tertiary and Cretaceous horizons, and
in the higher-lying eastern marginal areas (mid-continent, eastern states with
Appalachians) in sandstones and limestones of the Paleozoic era. The deeply sun-
ken Gulf coast has its main deposits in Cretaceous and Tertiary strata. In the
Californian fields, the oil comes from Tertiary and Cretaceous horizons. The
new discoveries in Alaska occur mainly in Permo-Triassic sandstones. Other
oil discoveries, scarcely developed until now, in the arctic expanse of Canada
indicate several very promising basins with possibilities of reservoirs from the
Paleozoic era to the Tertiary period.
United States. On January 1, 1990, the United States had available
proved reserves of ca. 7.2109 t and prospective reserves of ca. 7.3109 t. Follow-
ing the producing countries of the Middle East and the CIS, the United States
are in the third position. The same ranking applies to the production in the
year 1990. The cumulative production of 24.7109 t crude oil was achieved
from 597 320 production wells. (The Middle East states produced ca. 26.4109 t
from only 7061 wells). The oil resources in the offshore region of Texas,
Louisiana, and California were about one quarter of the total reserves. Due to
results of exploration in the Gulf of Mexico an increase of the offshore reserves
can be anticipated from new discoveries in the deep water region, but this will
scarcely lead to compensation for the fall in reserves determined by production.
Before Texas, with the coastal strip on the Gulf of Mexico, advanced in 1930 to
become the leading petroleum province, the position of the leading producer was
occupied by Pennsylvania until the turn of the century, then by California, and
later by the mid- continent. On January 1, 1990, the order according to daily
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 67
production was Texas (ca. 306 179 t), Alaska (281 948 t), Louisiana (171 145 t),
California (152 840 t), and Oklahoma (48 900 t).
Canada. Only after World War II did Canada rise to become an impor-
tant producing country as a result of discoveries in Alberta. On January 1,
1990, the Canadian reserves amounted to ca. 2.1109 t, 7.9109 t of prospective
reserves were reported. The daily production of ca. 190 000 t is obtained from
39 159 production wells. The reserves are declining. However, numerous discov-
eries in arctic regions cannot be developed and therefore cannot be evaluated.
More than 80 % of the oil production occurs in Alberta. The hydrocarbons are pre-
sent in the foothills of the Rocky Mountains and in the Alberta Plain. The reser-
voirs for oil, as for natural gas, are present in highly stressed folds, mostly bound
to overthrusts, in limestones of the Upper Carboniferous period, more rarely in
beds of the Devonian or the Triassic periods. In the Alberta Plain, Paleozoic
reef limestones are the reservoir rocks for the oil. Occasionally Cretaceous sand-
stones are the reservoir rocks, overlaying the Paleozoic discordant.
Latin America. The most important deposits in Latin America are those
of Venezuela and Mexico; Trinidad, Ecuador, Argentina, Peru, Colombia, and
Brazil follow some way behind. The main deposits are in the foretroughs of the
cordilleras and their branches, and are bound to sediments from the Cretaceous
to the Neogene periods. In addition, fairly small coastal basins in Brazil and Argen-
tina are productive. The resource figures have increased due to new discoveries in
Venezuela and more recently in Columbia but most of all to the oil accumulations
developed in the jungle basins of Ecuador and Peru. Through the development of
the Marina – Austral Basin between Tierra del Fuego and the Falkland Islands,
the reserves of Argentina have increased considerably. The new reserves are not
yet in production because of their geographical position and high gas : oil ratio. The
high proportion of prospective reserves in Argentina of ca. 477106 t, as against
proved reserves of only ca. 350106 t still points to a good exploration potential.
As a result of the large oil deposits in the Bay of Maracaibo and on the Gulf
Coast of Mexico, the submarine reserves of Latin America reach 40 % of the
total resources. Further offshore deposits are off Trinidad and Brazil.
Far East. The most important producing regions of the Far East are
China, Indonesia, Brunei, and Australia. In addition, oil deposits are known,
for example, in Malaysia, Burma, Thailand, Japan, Taiwan, and New Zealand.
China. Only after World War II has China gained importance as a produ-
cing country. By the use of modern scientific techniques, the first small fields
were discovered in the 1950s. In three regions, groups of fields had been proved
by the end of the 1980s that originate from plate-tectonic movements. Up to
1980, 150 fields had been discovered. Of these, merely for reasons of size, only
ca. 20 fields are of importance. Even so, the proved reserves have grown to ca.
5.0109 t. The two largest groups of fields, Daqing in East China and Shengli
on the Yellow Sea, achieved annual productions of 55.6106 t and 33.5 106 t,
respectively in 1990; that is ca. 64 % of the Chinese production. In the Daqing
field, which was discovered in 1959, between 4000 and 5000 wells distributed
over an area of 1200 km2 are producing from Lower Cretaceous sandstones in
very shallow anticlinical deposits. Shengli produces in a group of more than 40
individual structures from Precambrian and Tertiary reservoir beds. Of the
numerous sedimentary basins of China, only a few are fully developed. The
68 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
and pipeline transport of the natural gas (see Natural Gas). The production
quantities reported are not a measure of the natural gas potential available
in the producing countries, which follows from the growth of the reserves. This
is evident on comparing the United States and the CIS In 1989, production in
the United States was 500109 m3, and that of the CIS 739109 m3. On
January 1, 1990, the proved and probable reserves of the United States were
ca. 8.41012 m3 and those of the CIS ca. 451012 m3. To these there must be
added exploratory reserves ca. 111012 m3 for the United States and ca.
351012 m3 for the CIS The disproportion between production and reserves in
the CIS originates in missing pipeline connections, line losses, and also marketing
problems. Other large natural gas producers are Canada, the Netherlands, Iran,
Venezuela, the United Kingdom, Romania, Mexico, China, and Germany. The
large natural gas potentials in the countries of the Arabian Gulf, in Nigeria, and
in Indonesia suffer from the distance of the markets, since a basic requirement
for transportation from remote countries to the trade centers is the liquefaction
of the natural gas. Because of the relatively small capacity of the liquid gas tankers
and the large distances to be covered, the economic use of liquid gas is often at or
below the break-even point. The natural gas resources of the world — proved and
probable — on January 1, 1990, were ca. 127.01012 m3, and the prospective
another ca. 1211012 m3. The CIS has developed its leading position through the
enormous new discoveries, above all in Western Siberia, and is followed by the Mid-
dle East, which also has a very powerful growth to record. The raising of the proved
gas reserves in these two large districts originates not only in new discoveries but
partly in the extension or higher evaluation of deposits already known and the
strengthened inclusion of petroleum gases, particularly in the Middle East and
Venezuela. While Africa and North America show a declining trend for the proved
reserves, Western Europe has improved its position in the international gas sector
by further new discoveries in the North Sea.
Regional Distribution of the Reserves. A breakdown of natural gas pro-
duction and resources according to region is given in Table 2.
Commonwealth of Independent States. The largest reserves in the
world, at 451012 m3, are those of the CIS The occurrences are widely spread
and in some cases combined with petroleum deposits.
Table 2. Natural gas production and resources up to December 31, 1989 (assembled
from various international data)
Resources
The first great gas boom was in the 1950s with the development of the fields
of Shebelinka in the Dnieper – Donetz Trough and Gazli in Central Asia, each
with primary resources of 450 – 500109 m3. The discoveries in the second
half of the 1960s, some of which were enormous, established the present position
of the CIS in the natural gas sector. Important deposits mentioned are those of
Krasny – Kholm (Orenburg) in the Volga – Urals district, where ca. 750109 m3
of gas have been discovered, and the Vuktyl and Layavozh deposits in the Petchora
Basin, linked with beds of the same age and each with ca. 500109 m3 of gas. In
Turkmenia, the Malai field with ca. 340109 m3, is one of the giants. Probably
the largest natural gas province in the world is in the north of the Western Siberian
platform. The gas occurring here in sandstones of the Cretaceous system (predomi-
nantly Cenomanian), in part several hundred meters thick, consists of almost pure
(97 – 99 %) methane. The structural traps are extended anticlines and mound-like
highs. About 220 fields have been developed in this area up to now.
This producing province is the most important for the future export of gas
to Europe. Its most extensive deposits include the Urengoy field, the largest
deposit proved so far with ca. 61012 m3 of initial gas reserves, and also the
large deposits of Jubileyny (ca. 21012 m3), Taz (ca. 1.151012 m3), Medvezhye
(ca. 1.05 1012 m3), and Yamburg (ca. 850109 m3) among others. At least a
further ten fields of this region have each reported primarily more than
100109 m3 of recoverable gas reserves. Two new discoveries in the Arctic
Ocean deserve attention. In 1987 in the Barents Sea, not far from the
Norwegian border, the Shtowmanowskaja field with an estimated reserve of
3 – 41012 m3 was discovered, and in 1988 in the Kara Sea, the Russanowskaja
field with 61012 m3. Both fields together are credited with 15 % of the future
production provided that the production and transportation to Nowaja Semlja
is possible under arctic conditions.
Middle East. Even if not so dominant as in the case of crude oil, the
areas on the Arabian Gulf, with proved natural gas reserves of ca. 681012 m3
play an important role in the possibility of future exports. However, a consider-
able fraction of these quantities is trapped in gas caps or is dissolved petroleum
gas linked with the large oilfields of this area and is therefore dependent on their
production. The gas cap of the Pazanan oilfield in Iran alone contains recoverable
quantities of ca. 1.41012 m3 of gas. With the recent discovery of enormous gas
accumulations in the Permian Khuff formation, very large dry gas deposits have
also been proved for the Middle East. The Kangan field in Iran is said to contain
41012 m3 of resources. The deposits available in the Permian formation are
sealed above by triadic salt beds. There are further deposits of this type, not
hitherto exploited, in Qatar. Iran also is involved in the northern part of the
Northwest Dome (North Field) – with 2.8 – 3.4 1012 m3 a ‘‘super giant’’ –
which belongs to the Permian Khuff. Notable natural gas reserves are also
known in Saudi Arabia and the Emirates.
North America. Despite higher quantities produced, the known North
American gas reserves have risen again. The reserves of 91012 m3 in the mid
1970s have risen to ca. 111012 m3 proved and probable. Of these, 8.41012 m3
in the mid 1980s were assigned to the United States alone, the traditional
land of natural gas. Its large-scale industrial use started here even before the
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 73
2. Reservoir Engineering
Fig. 7. Composition of a detritral rock (72) a) Grain (texture: sand granules 1/16 – 2 mm
diameter; silt 1/256 – 1/16 mm diameter); b) Pore fillings (texture: sand granules 1/16 mm
diameter; silt 1/16 – 1/256 mm diameter; clay < 1/256 mm); c) Cement; d) Fluid.
Vp
f¼
Vb
According to this definition, the porosity also includes the volume of isolated
pores (dead pores).
If all interconnected elements of the pore volume are taken into account, the
effective porosity is obtained. This is an indicator of the effective storage volume
of a rock. The sum of effective feff and dead fd porosities gives the total porosity:
feff þ fd ¼ f
76 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
For geometrical or hydrodynamical reasons, not all pores are available for
convective mass transport. A fixed portion of the cavity volume can be regarded
as dead-end pore volume.
The porosity depends essentially on four characteristics of the sedimentary
rock:
1) Particle shape
2) Degree of cementation
3) Particle size distribution
4) Particle arrangement
A B C
90° 60°
90° 90° 60° 60°
Fig. 9. Packing and porosity (73) A) Space-centered, cubic sphere packing (number
of contacts = 6, f ¼ 47.6 %); B) Hexagonal packing (number of contacts ¼ 8, f ¼ 39.5 %);
C) Densest packing of spheres (rhombohedral) (number of contacts = 12, f ¼ 25.9 %).
(Fig. 9 A), in which the particle midpoints form the corners of a cube, the inter-
particulate pore volume is exactly 47.6 % of the total volume. In this case each
particle has six points of contact with the neighboring particles. In the hexagonal
arrangement of particles (Fig. 9 B), the spherical particles of the upper lattice
plane are displaced by half the particle diameter relative to the lower lattice
plane. In this arrangement, the number of contacts rises from six to eight and
the porosity decreases to 39.5 %. If in the second lattice plane of the packing,
the row of spheres is displaced not only in the x-direction by d/2, but also
in the y-direction perpendicular to it, by d/2, the densest packing of spheres is
obtained (C). Here in the rhombohedral arrangement the particles of the upper
lattice plane are placed exactly in the holes of the lower plane and accordingly
each particle has twelve neighbors and the porosity is only 25.9 %. The porosity
of a natural packing of cemented sedimentary grains usually falls between the
limiting values of 39.5 and 25.9 %.
Pore Configuration Models. Although, at first glance, the shape of the
pores or fissures of a sedimentary rock has very little to do with regular geo-
metric forms, cylindrical capillaries or parallel plates are regarded as geometri-
cally similar to the pores and fissures and are chosen as models for calculating
the pore space properties. These very simple geometries are unable, however,
to reproduce the irregular variation of the pore or fissure cross section or of
the tortuosity of flow path, which are important characteristics of the pore
configuration.
In Figure 10, the commonest types of pore configuration models, up to the
most advanced representation (fractal network model), are shown. The capillary
bundle model (Fig. 10 A) is the most frequently applied pore configuration model,
mainly used for calculating the pore radius distribution, the saturation with
fluids, and the oil displacement process. A particular drawback of this model is
that a transverse flow between the capillaries is not possible. A pore configura-
tion model for fissured reservoir rocks derived from the plate model is the
78 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 10. Pore configuration models (74) A) Capillary bundle model; B) Brickstone model;
C) Intersphere network model; D) Cascade network model; E) Fractal network model
White color means pore space.
brickstone model (B), in which the individual blocks symbolize the porous matrix
and the gaps between them, the fissures. Here mass transfer between matrix and
fissure is possible. The intersphere network model (C) is used particularly for
determining the phase distribution in the pore space and the capillary forces.
With the cascade network model (D), dead-end pore effects, turbulence effects,
and mixing phenomena can be investigated.
According to the principle of self-similarity of geometric forms in nature,
relations are traceable as far as between the crystal forms in the microcosmos
and the geological outfits (morphology) of the Earth’s surface. In connection
with this, the fractal dimension is a scale-up factor between the characteristic
dimensions and the image scales. If the spherical shape (circular shape in
Fig. 10 E) is chosen as the characteristic elementary shape of the rock structure
and of the pore space, a fractal network for matrix and pore configuration can be
developed, whose specific internal surface and volume fraction are calculable and
correlatable with other physically measurable quantities such as electrical con-
ductivity and permeability.
An example of the use of pore configuration models for calculating pore
space properties is given here, based on the determination of the pore radius
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 79
2sHg cosyHg
pc ¼
r
where sHg is the surface tension of mercury and y the wetting angle of
contact.
In order to reconcile this measurement with actual pore space configura-
tions, it must be taken into account that each pore has at least two characteristic
dimensions – the pore entrance and the pore widening. The characteristic cross-
section determined by the mercury injection method corresponds to the pore
entrance radius.
Compressibility of the Pore Space. The compressibility c of a substance
is the change, with external pressure, of its volume relative to its initial volume.
For a solid the compressibility is negligibly small. A porous medium which is
built up of individual particles can, however, change its particle arrangement
under pressure and, therefore, is said to have considerable compressibility.
The deformability is discussed for the different forms of packing of spherical par-
ticles in Section 2.1.1.1.1. According to the studies of GEERTSMA (75), the pore
volume compressibility cf depends on the type of pore configuration (porosity of
the medium) and the effective pressure which is the difference of external pres-
sure pe and internal pore pressure pi.
1 qVp
cf ¼ ðÞ
Vp qð pe pi Þ
Figure 11 shows that the pore volume change reacts with increasing sen-
sitivity to an increase in the external pressure with decreasing porosity. The
porosity of sedimentary rock under in situ conditions (ie, under the influence
of the overburden pressure of the sediments) can, therefore, be up to 20 % smaller
than the porosity determined in the laboratory (77).
Physical Properties of the Pore Saturating Fluids. Gas, oil, and water
are the relevant pore saturating fluids, whose physical properties must be
known as a function of the pressure, the reservoir temperature, and the fluid’s
composition. The material properties of these multicomponent systems can
only be predicted for the gas state at low pressure with acceptable accuracy
and without an extensive experimental data base.
For liquids and phase mixtures, estimates of the material properties
are obtained in accordance with the theorem of corresponding states by
normalizing to conditions at the critical point. The state variables, pressure,
and temperature, are normalized by introducing the reduced pressure pr
80 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
14
12
10
0
0 0.04 0.08 0.12 0.16 0.20 0.24
Φ
p T
pr ¼ Tr ¼
pc Tc
Mr p
%g ¼
zRT
where Mr is the molecular mass, p the pressure, z the real gas factor, R the gas
constant, and T the temperature.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 81
1 %0 j
where the standard conditions for gases are: p0 = 0.1013 MPa and T0 273.15 K.
For liquids p0 = 0.1013 MPa and T0 = 293.15 K holds.
In the United States, 60 8F, corresponding to 288.65 K, is used as standard
temperature. The formation volume factor for gas Bg can be calculated directly
from the real gas equation:
zT p0
Bg ¼
pT0
To describe the volume behavior of a two-phase mixture of gas and oil, the two
formation volume factors can be added to obtain the two-phase formation volume
factor Bt, taking the gas – oil ratio (Rs) into account.
Bt ¼ Bo þ Bg ðRsi Rs Þ
1.0
0
from degassing below the bubble point pressure ( pb). Instead, it rises further as
a result of volume expansion on pressure release. This is due to the relatively
low dissolved gas volume in water (ca. 1 m3 gas/m3) compared with oil
(10 – 50 m3 gas/m3).
Compressibility. The isothermal compressibility of the pore fluids is
given by:
1 qv
c¼
v q pT
The pore fluids differ clearly in their compressibility behavior. While gases in
the pressure range 0.1 – 100 MPa are extraordinarily compressible (c = 105 –
108 Pa1), oils have a moderate to low compressibility (c = 108 – 1010 Pa1)
depending on their dissolved gas content, and water has a compressibility of ca.
1010 Pa1. The last figure is of the same order as the compressibility of the
rock pore volume. In calculating the compressibility of real substances, the nor-
malized thermodynamic functions of state, pressure, and temperature can be used.
For a gas mixture under real conditions, the compressibility equation
reduced to normalizedconditions is:
cpr 1 1 1 qz
cg ¼ ¼
ppc ppc ppr z q ppr
Here, instead of the critical data for a single gas, pseudocritical values ( ppc and
Tpc) are used as normalizing quantities.
p X
n
ppr ¼ ppc ¼ y j pc j
ppc 1
T X
n
Tpr ¼ Tpc ¼ y j Tc j
Tpc 1
84 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
To allow for the influence of pressure and temperature on the viscosity of a real
gas, the normalized representations of the reduced pressure and temperature
can again be employed.
The pressure and temperature dependence of the viscosity of reservoir oils
is shown in Figures 15 and 16, respectively. Above the bubble point pressure pb,
Fig. 16. Viscosity of oil as a function of temperature a) Heavy crude oil; b) Medium crude oil.
the viscosity linearly increases with the reservoir pressure. Below the bubble
point pressure, the oil viscosity falls steeply with increasing pressure or gas
content.
The temperature dependence of the viscosity of oil is characterized by an
exponential function, so that in a semilogarithmic plot versus the reciprocal of
the absolute temperature (see Fig. 16) a straight line is obtained. The slope of
this straight line is greater the higher the oil’s content of heavy components.
For the dependence of the viscosity of water on the temperature the same math-
ematical relationships apply. In addition, the viscosity of water increases with
increasing salt content (79).
Reserves. A detailed discussion of worldwide resources of oil and gas is
given in Sections and 1.2.3,, respectively.
Resources mean the total volume of hydrocarbon stored in reservoirs. With
oil, gas, or condensate reservoirs, the initial volume of the pore contents, the so-
called original oil in place (OOIP) or original gas in place (OGIP), is the first
quantity used for evaluating a resource. Since for physical, technological, and
economic reasons the resources are not completely recoverable, a second concept
is necessary for characterizing the profitability of a discovery: the reserves.
Reserves mean those quantities of hydrocarbons that at a given time are proved
on the basis of sufficient geological and reservoir engineering information and
can, in future, be produced with currently available methods of recovery and
applicable economic conditions. Reduced to a simple formula,
The recovery factors for oil reservoirs, apart from a few exceptions, are
between 20 and 45 % of the contents. For gas reservoirs, the recovery factor
can be up to 85 %. The degree of certainty in the determination of the reserves
increases with increasing development of a reservoir. According to a classifica-
tion of reserves made by the Committee of the World Petroleum Congress, ‘‘pro-
ven reserves’’ are spoken of when the probability of proof exceeds 90 %, and
‘‘unproven reserves’’ when that probability is between 10 and 90 % (80). The
86 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
unproven reserves are split by a median value of 50 % into probable and possible
reserves. The remainder are the speculative, so far undiscovered, possibly reco-
verable reserves.
Volumetric Determination of the Reservoir Contents. At the start
of development of an oil or gas reservoir, usually only the contours of the reser-
voir are known; apart from test analyses data, no adequate production and pres-
sure trends are available. The content of a reservoir can then be surveyed only by
volumetric methods. The contents related to standard conditions are calculated
from the following formulae:
Vb f ð1 Swi Þ
OOIP ¼
Boi
Vb f ð1 Swi Þ
OGIP ¼
Boi
where OOIP denotes original oil in place, OGIP denotes original gas in place, and
Swi is the initial water saturation.
The bulk volume of the reservoir Vb is determined from a geological model
of the reservoir. In this procedure the nonproductive regions of the strata must
also be taken into account such as clayey interlayers or lenses, water-bearing
belts, and the distinction between gas-, oil- and water-bearing zones (gas –
oil – water contacts) when delimiting the net rock volume from the gross rock
volume. In the evaluation of borehole measurements, so- called cut-off values
for porosity, saturation, clay content, and permeability are introduced to differ-
entiate between pay zone and nonproductive thickness and to determine the pay
zone thickness (Fig. 17). The black areas correspond to the porous, hydrocarbon-
bearing formation intervals that are included in the net layer thickness.
For displaying the geological models, various mapping methods are chosen.
For example, in Figure 18, the structural map of a reservoir is represented by
subsurface contour lines (that is in feed). For determining the net rock volume
of a reservoir, the reservoir body can be imagined to be divided into a large num-
ber of slices of equal thickness. Each of these slices can be represented by a geo-
metrically similar body such as a truncated cone, truncated pyramid, or other,
irregular shapes, allowing its volume to be calculated. A more elegant method
is the graphical evaluation of the subsurface contour planes. In this method
the areas of the contour line map determined by planimeter for the upper and
lower boundary of the reservoir are plotted against the relative depth in a system
of cartesian coordinates (77).
Furthermore, if the lower and upper boundaries of the reservoir contents –
the gas – oil and water – oil contacts, respectively — are taken into account, the
rock volume can be determined directly from the area – depth graph by means of
millimeter paper or planimetry. These computed or graphically-determined rock
volumes are inserted into the formulae for calculating the OOIP and OGIP. In
low-permeability reservoirs bearing formation water, the transition zone from
the completely water saturated zone to the oil bearing zone, containing irreduci-
ble water, is an uncertainty factor in the oil or gas resource calculation. The tran-
sition region is characterized by the capillary pressure curve.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 87
2.2. Flow Processes in Porous Media. The original state of a crude oil
or natural gas reservoir is characterized by the thermodynamic equilibrium of the
hydrocarbon phases and the capillary equilibrium of the wetting and nonwetting
phases, which in turn depends on the prevailing temperature, pore space pressure,
and rock properties. Not only the composition but also the properties of the oil, gas,
and water phase show a remarkable ‘‘memory’’ for the early history of the formation
of the deposit. For example, oils that lack in certain alkanes can be recognized as
intermediate products of biological degradation from an original sapropellitic
material. Light oils which have been exposed to oxidation as a result of tectonic
88 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
postsedimentary processes, might have been converted into medium-heavy oils, but
they still retain the chemical features (fingerprints) of their origins.
The initial distribution of the phases in the pore space is governed by the
equilibrium between interfacial forces and viscous forces during the migration pro-
cess. If a volatile oil enters a porous rock which is saturated with salt water and
flushes the pore space over extended periods then the resulting initial water
saturation in the deposit is normally greater than the adsorbed water saturation
(ie, the layer of water retained by physical adhesion to the interior surfaces of the
rock). If, in contrast, the migrating oil is composed of heavier hydrocarbon frac-
tions, it displaces the pore water almost down to the level of the absorbed water
saturation. After a reservoir has been filled, the materials are in dynamic rather
than static equilibrium. Diffusion and convection can influence the material equi-
librium and the distribution of the phases in the deposit.
Single-Phase Flow. If a fluid with little or no compressibility (water, oil)
is flowing in an infinitely large reservoir system towards a pressure sink (well)
then — after a transient phase — a steady-state condition is reached between the
quantity removed from the well and the quantity flowing in from the deposit.
Under these conditions, the pressure changes only as a function of space, but
not of time. However, during the transient state before the steady-state equili-
brium is reached the pressure also changes with time.
Steady-State Flow – Darcy’s Equation. In the laminar flow regime —
ie, in a homogeneously porous medium with a Reynolds number < 1 — Darcy’s
law holds; a single well in an infinitely large reservoir radial flow geometry
can be assumed
q k qp
¼v¼ ð2:1Þ
2prh m qr
where q is the flow rate, r the radius, h the thickness of the porous layer, v the
apparent velocity, k the permeability of the rock, m the dynamic viscosity, and p
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 89
the pressure. Integration of Equation (02.1) between the well radius rw and the
radius re of the area drained by the well gives the following for steady-state flow:
Boundary conditions: pe at re; pwf at rw are constant with time
2phk ð pe pwf Þ
qw ¼ ð2:2Þ
Bm ln re =rw
pe þ pwf
pm ¼
2
Ak ð pe pwf Þ
qm ¼
m l
2
Ak ð pe p2wf Þ
qw pwf ¼ qm pm ¼
m 2l
2
Ak ð pe p2wf Þ
qw ¼
Bm 2 pwf l
Fig. 19. Material balance in a cylindrical capillary of length Dx and cross-sectional area A.
90 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
divided by A Dx Dt
qð%vx Þ q%
þ ¼0
qx qt
The differential equation for the continuity law is obtained from the differ-
ence equation assuming small increments of time and space. The nabla operator
and the porosity in the storage volume DV are introduced so that this equation
can be applied to a porous medium and extended to three-dimensional flow
conditions. For horizontal flow of a medium of low compressibility, the extended
continuity equation becomes:
qðf%Þ
rð%vÞ þ ¼0 ð2:4Þ
qt
In one dimension:
k qp
vx ¼
m qx
q k qp qðf%Þ
% þ ¼0 ð2:5Þ
qx m qx qt
k q% q p q2 p q%
þ% 2 ¼
m qx qx qx qt
where k/m is the transmissibility of the pore fluid for a given porous medium.
As an initial condition it was assumed that the transmissibility k/f and the
porosity f are constant throughout the pore space system. The compressibility
equation
1 q%
c¼
% qp
for a fluid of low compressibility is introduced and used to replace the change in
density % by the change in pressure p.
"
k% q p 2 q2 p qp
c þ 2 ¼ %c
m qx qx qt
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 91
Both sides are multiplied by the product m/k%. The flow rate, at some dis-
tance from the well, is only in the range of a few centimeters per hour so the pres-
sure gradient p/x and the square of the pressure gradient are very small in
comparison with the other variables in Equation (2.5). This term can therefore
be neglected.
This then gives Equation (2.6) which is similar in form to the diffusivity
equation, and is referred to as the pressure propagation or hydraulic conductiv-
ity equation for the one-dimensional case.
q2 p fmc q p
¼ ð2:6Þ
qx2 k qt
Equation (2.5) can also be solved for other initial conditions. If the density of
the fluid phase is defined by a mass law, eg, the gas law for ideal gases, then the
following form of the hydraulic conductivity equation is obtained
q2 ð p 2 Þ fm q p
2
¼2 ð2:7Þ
qx k qt
q2 p 1 q p fmc q p
þ ¼ ð2:8Þ
qr2 r qr k qt
With a linear pressure sink as the well, and for infinite extension of the reservoir,
ie, re ¼ 1, the Boltzmann transformation can be used for solving Equation (2.8).
The following boundary conditions apply:
t¼0 r ¼ rw p ¼ pi
r ¼ re p ¼ pi
t ¼ t1 r ¼ rw p ¼ pwf
r = re p ¼ pi
where the symbol Ei denotes exponential integral and pwf the flowing pressure at
the bottom of the well.
In an infinitely large homogeneous reservoir system with a line source
as the well, the flowing pressure at any point can be calculated from
Equation (2.9) as a function of the radius r and the time t. The natural
92 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
logarithm can be used in place of the exponential integral provided the argu-
ment of the exponential integral is significantly smaller than the limiting value
of 0.01 (83).
qm gfmct rw 2
pwf ¼ pi þ ln ð2:10Þ
4phk 4kt
qm tp þ t
pws ¼ pi ln ð2:11Þ
4phk t
where tf ¼ tp denotes the flow duration and Dt the shut-in time (84). This equation
has been used to extrapolate the initial reservoir pressure pi from the graphical
presentation of the shut-in pressure plotted against the logarithm of the time
t þt
ratio ð p t ¼ 1Þ (Fig. 20).
For an infinitely large reservoir the initial pressure in the well will
t þt
be reconstituted after an infinite shut-in time pt ¼ 1: The gradient of the
tp þt
Fig. 20. Shut-in pressure pWS versus logarithmic time ratio t for extrapolation of
initial reservoir pressure pi (83).
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 93
qmB
m¼
4phk
qmB
kh ¼
4pm
In addition to this simple radial model with the selected initial and bound-
ary conditions many methods exist for finding solutions for specific reservoir con-
ditions, which can be applied to individual well and reservoir problems.
Individual solutions have therefore been found for the differential equations to
the following conditions; finite reservoir extension, near wellbore effects, frac-
tured reservoirs, stratified reservoirs, and partial penetration and particular
direction of the well in the producing formation.
Multiphase Flow. Capillary Pressure and Saturation. If there is
more than one phase in the pore space system of a reservoir rock then the distri-
bution of the phases is determined by the interfacial tensions between the fluids
and the solid phase. The equilibrium of forces between two liquid phases (oil and
water) and a solid phase is shown in Figure 21.
The wetting angle y is a measure of the wettability of the solid phase
by the oil and water phases. A value of 08 for y represents an ideally water
wet solid, whereas 1808 defines a completely oil wet surface. The distri-
bution of the phases for a pore cross section therefore can be idealized as
follows:
Fig. 21. Equilibrium of forces between two liquid phases (water and oil) and the solid
phase (77). sso = interfacial tension solid – oil; ssw = interfacial tension solid – water; swo
interfacial tension water – oil.
94 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 22. Idealized distribution of oil, water, and gas phase in a pore cross section.
a) Water; b) Gas; c) Oil; d) Rock.
– Gas, which is the least wetting, is positioned in the center of the pore duct
– The oil phase is distributed between them (Fig. 22)
Due to the adhesive forces the pressure in the wetting phase (water) is
lower than in the nonwetting phase (oil). The difference between these two pres-
sures is the capillary pressure pc where
pc ¼ po pw
Capillary pressure curves (see Fig. 23) are of fundamental importance for
the mobility of the phases in multiphase flow. They determine the irreducible,
and, therefore, fixed saturation of the wetting phase at infinite capillary pressure
and characterize the capillary ascent of the water table at the boundary of the
oil – water zones. The capillary pressure can be directly transferred into eleva-
tion by use of the capillary/gravity force equilibrium.
2sow cosy
¼ ð%w do Þgh
r
where sow denotes the interfacial tension between oil and water phase, r the
equivalent hydraulic radius of the porous system, %w the density of water, do
the density of oil, g the gravitational constant, and h the elevation above the
water table.
For the same capillary pressure, a higher water saturation is found in a
pore space system of lower permeability (a) than in a highly permeable
one (b). If it is assumed that water represents the wetting phase this means a
reinforced capillary imbibition of water from zones of high permeability into
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 95
Fig. 23. Capillary pressure versus wetting fluid saturation (85). a) Low permeability;
b) High permeability.
low permeability and therefore an improved displacement of oil from the pores of
lower permeability.
Relative Permeabilities. The mobilities of the individual phases in mul-
tiphase flow in a porous medium depend on the wettability of the rock, on the
interfacial tension between the fluids, and on their viscosities. A new flow coeffi-
cient, which is introduced into the phase flow equation as the relative permeabil-
ity, is needed to describe the different mobilities of the phases in the medium as a
function of their fractional saturation. Relative permeability kr is defined as the
ratio of the effective permeability of a phase (kw, ko, kg) in the presence of one or
more adjacent phases to a reference permeability. The absolute rock permeabil-
ity according to Darcy, or the permeability to water at 100 % water saturation,
or the permeability to oil for irreducible water saturation can be used as the
reference permeability. The relative permeabilities for a binary oil – water sys-
tem for a more strongly water-wet and a more strongly oil-wet rock are shown
in Figure 24. If the saturation of the wetting phase Sw is plotted on the abscissa
then the curves are displaced to the right with increasing water wettability. On
the other hand, the final value for the relative permeability of the water phase
increases if the water wettability decreases.
This type of representation can only be used for two-phase permeabilities;
for a three-phase system STONE (87) has proposed a paired representation of the
water – oil and gas – oil relative permeabilities.
96 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 24. Relative permeabilities for a binary oil – water system (86). A) Water-wet rock;
B) Oil-wet rock.
Apart from this, a ternary diagram is the only possible way of representing
the relative permeabilities of a three-phase system.
The flow equations for a two-phase system can therefore be represented
while taking account of the functions of the relative permeabilities and the capil-
lary pressure.
Oil:
kro k%o qSw o qw
r ro ¼ q
mo q pc qt qt
Water:
krw k%w qSw qo qw
r rw ¼ f
mw q pc qt qt
1
ð p þ %gzÞ
%
and kro and krw denote the relative permeabilities of oil and water, respectively.
Depending on the boundary conditions, this system of partial differential
equations, can, in general, only be solved numerically.
Drive Mechanisms and Recovery Efficiency. The energy reservoir for
the flow of oil, gas, and water in a porous medium consists of the compression
energy of the aquifer volume surrounding the deposit; the mechanical energy
of the gas liberation process; the influence of gravity; and the capillary pressure
effect.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 97
Fig. 25. Binary diagram for a gas – oil mixture (each phase consisting of one component)
(77). A = bubble point pressure (pressure depletion at room temperature); A0 = initial reser-
voir pressure; C = critical point.
change in the system also changes the properties of the phases in a way
which is disadvantageous to the oil recovery process; measures which
maintain the pressure are, therefore, favorable. Water drive is by far the
most effective drive mechanism; it can be introduced by artificial flooding
measures in locations where water cannot be fed from natural sources (ie,
inadequate aquifer size or inadequate energy replenishment due to low aqui-
fer permeability). The oil recovery process by natural aquifer expansion and
by the other drive mechanisms mentioned rate as primary extraction pro-
cesses. On the other hand, replenishment of the reservoir energy by water
or gas injection is one of the secondary extraction processes.
The recovery efficiency of a deposit is given by the percentage of the
cumulative production of the quantity of hydrocarbon originally present. The
performance of extraction and the recovery efficiency are determined largely
by the geological factors of the deposit and the physics of the oil recovery pro-
cess. The physical approach to oil recovery processes presupposes a micro-
scopic dimension, and the geological approach a macroscopic one; the
recovery factor must, therefore, include microscopic and macroscopic para-
meters. The total recovery factor or recovery efficiency ER is the product of
the microscopic recovery factor ED, the areal recovery factor EA, and the ver-
tical recovery factor (flood invasion efficiency) EI. Figure 26 shows that the
recovery efficiencies lie between 20 % and 40 % of the oil present, depending
on the characteristic drive mechanisms and oil viscosities. In gas reservoirs
the recovery efficiency lies between 80 and 90 % of the gas present. This is
Fig. 26. Oil recovery efficiency as a function of pressure for various drive mechanisms
(88). a) Depletion drive; b) External gas drive (small gas cap); c) External gas drive (large
gas cap); d) Water drive.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 99
attributable to the high mobility of the gas and the favorable form of the
volume expansion drive.
Microscopic Recovery Factor. BUCKLEY, LEVERETT, and WELGE (89–91)
have developed the theory of the displacement of a fluid by a second, immiscible
fluid in a homogeneous medium. This microscopic process determines which fluid
exhibits the stronger wetting properties towards the porous medium and which
has the greater mobility. The displacement process in the pore space is not as
perfect as the displacement of a fluid by a piston in a machine; this is because
a mobile residual phase saturation of the displaced fluid (oil) remains after the
displacement front has passed. Continued flooding can reduce this mobile resi-
dual phase saturation down to an irreducible, fixed percentage. Through a linear
flooding process it is possible to determine the saturation distribution S along the
flooding section, ie, in direction x (Fig. 27).
Fig. 27. A) Saturation distribution along flooding section (direction x); B) Fractional flow
for water along flooding section.
100 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
At the inlet cross section the water saturation is only slightly smaller than
the difference of 1 minus the physical residual oil saturation Sor. Up to the flood-
ing front, the water saturation decreases to a value Swf (front saturation),
depending on the mobilities of the phases. The difference between the ultimate
water saturation (1 Sor) and the actual water saturation at any section is
defined by the profile f, the saturation of mobile oil. The front is characterized
by a step change in the water saturation down to the irreducible water saturation
Swi. The proportions of the flowing phases change along this saturation profile
and can be expressed by the fractional flow f of one of the two flowing phases,
as it is shown for different times t1 and t2 in Figure 27 B.
qw 1
fw ¼ ¼
qo þ qw 1 þ M1
lw krw mo
M¼ ¼
lo kro mw
The fractional flow for water is a function of the mobility ratio M, ie, the
mobility of water kw/mw as the displacing phase to the mobility of the oil ko/mo
as the displaced medium. At the front the fractional flow of water falls to zero,
as only the mobile oil phase can flow ahead of the front. By using a continuity
approach in the porous medium the velocity has been determined at which the
front or the saturation profile moves through a pore continuum (90). If an infi-
nitely small volume unit is considered at the boundary between oil and water
then the difference between input and output water flow volume in the time
unit Dt must be equal to the change in saturation in the space element A Dx
(see Fig. 28).
Fig. 28. Mass balance for an infinitely small volume unit at the saturation boundary
between oil and water.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 101
saturation profile.
qx q q fw
¼
qt Af qSw
Integration between the limits x and 0 and the times t and 0 gives the equa-
tion for determining position x which has reached some arbitrary saturation
value at the time t after injection of a certain volume of water Wi.
W i q fw
x¼
Af qSw
Zt
Wi ¼ qdt
0
Fig. 29. Fractional flow curve for water, definition of the front saturation Swf and break
through saturation Swbt according to Welge (92).
Fig. 30. Concentric arrangement of four peripheral production with respect to one injec-
tion well, 5-point pattern.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 103
before breakthrough of the flooding front depends on the mobility ratio of the
front and the arrangement of the wells.
In a homogeneous producing formation the well spacing initially plays no
part. However, as the capacity of the producing formation to accept the injection
medium (water) depends on permeability, the well spacing cannot be set at any
arbitrary distance. There is little point in optimizing the well spacing on the
basis of a homogeneous deposit model, as in practice more complex deposit for-
mations are often found.
Vertical Recovery Factor or Flood Invasion Efficiency. If the oil pro-
ducing formation were to have no vertical subdivisions then the overall oil recov-
ery factor from this deposit would be the product of the microscopic and areal oil
recovery factors. However, the conditions under which the sediments were
deposited have normally led to a vertical, facially subdivided stratification,
with the result that the vertical recovery factor also has to be taken into account
when determining the overall oil recovery.
If water is invading into a stratified producing formation from an aquifer or
an injection well the strata fill in the order of their permeabilities. The most
permeable layer is flooded first.
DYKSTRA and PARSONS (93) have developed a simple theory for determining
the vertical recovery factor in a stratified producing formation. For this purpose
they have further simplified the microscopic displacement process as described
by BUCKLEY and LEVERETT and have assumed that no mobile residual oil remains
behind the front. However, unlike BUCKLEY and LEVERETT, they have taken into
account the fact that the velocity of the displacement front depends on the mobi-
lity ratio of the phases and on the position of the displacement front along the
flooding section. The higher the mobility of the displacing phase the more rapidly
the flooding front extends toward the end of the flooding section. The flooding
velocity at the inlet corresponds to the effective Darcy flow velocity of the oil
phase (see Fig. 31).
Fig. 31. Relative velocity of flooding front as function of fractional distance for different
mobilities of displacing phase (94). a) M ¼ 0.1; b) M ¼ 0.2; c) M ¼ 0.5; d) M ¼ 1; e) M ¼ 2; f)
M ¼ 5; g) M ¼ 10.
104 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 32. Cross section of a layered reservoir; position of flood front in the individual
layers at a certain time level t (94).
P
i P
n
xj
hj þ l hj
1 iþ1
EI ¼
P
n
hj
1
The position of the flooding front at time t in the incompletely swept strata
j ¼ i + 1, relative to the layer i in which the water has just broken through, is
determined by the equation
0:5
2 kj 2
M þ ki ð1 M Þ
xj M
¼
l 1M 1M
fl2 ðM þ 1Þmw
ti ¼
2kwi p
With the aid of these three equations the vertical oil recovery factor can be
determined at any given time. The product of the three factors gives the overall
recovery efficiency ER:
ED ¼ ED EA EI
plan measures to optimize production from the reservoir. The calculation meth-
ods used are the basis for predictions of future production performance and are a
strategic instrument of reservoir management. All methods of prediction are
based on the physical fundamentals of the recovery process and the flow pro-
perties of the fluids in a geologically complex reservoir body. The credibility
and reliability of these prediction methods depend on how closely the calculation
techniques model the reality of the recovery process and the reservoir. This
accuracy of the model must be checked by a model validation before use. Once
the model is chosen it must be adjusted to the production history of the reservoir
(model calibration). The better the model is calibrated, the more reliable the
prediction.
The following models for describing the reservoir behavior, with increas-
ing complexity of process description and/or reservoir characterization, can be
distinguished:
Material Balances Method. The tank model is the simplest form of recon-
struction of the reservoir contents. It is based on the conservation of mass and an
equation of state (88). All forms of material balance use the following approach:
Initial State:
N Boi
Vpi ¼
1 Swi
Vwi ¼ Vpi Swi
(for explanation of terms see the abbreviations list at the beginning of this
article).
1) The actual pressure exceeds the bubble point pressure pb but is lower
than pi.
2) The oil volume NpBo is produced. The volume lost is replaced by the expan-
sion of the rock matrix, the residual oil, and the initial water.
Vp ¼ Vpi ½1 cf ð pi pÞ
Vw ¼ Vwi ½1 þ cw ð pi pÞ
ðN Np ÞBo ¼ Residual oil
(for explanation of terms see the abbreviations list at the beginning of this
article).
State Below the Bubble Point Pressure pb:
1) The reservoir pressure p is less than the bubble point pressure pb.
2) Aside from the volume expansion of oil, initial water, and rock matrix, the
release of gas from solution occurs as a further drive process. To simplify
the calculation of the equilibrium behavior as far as possible, the compres-
sibility equations for the oil, water, and rock phases as well as the real gas
equation for the gas phase are used as material laws.
Vp ¼ Vpi ð1 cf ð pi pÞÞ
Vw ¼ Vwi ½1 þ cw ð pi pÞ
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 107
N Boi
N Rsi Bg N Rs Bg þ N Bo þ ðSwi ð1 þ cw pÞ 1 þ cf pÞ
1 Swi
¼ Gp Bg Np Rs Bg þ Wp Bw þ Np Bo
G p ¼ N p Rp
gives
Boi
N ðRsi Rs Þ Bg þ Bo þ ððcf þ Swi cw Þ p ð1 Swi ÞÞ
1 Swi
¼ Np ½ðRp Rs Þ þ Bo þ Wp Bw
Bo Boi
¼ co
Boi p
gives
Bg Bo Boi cf þ swi cw
N Boi ðRsi Rs Þ þ þ p
Boi Boi 1 Swi
Bg co Soi þ cw Swi þ cf
N Boi ðRsi Rs Þ þ p
Boi 1 Swi
¼ Np ðBo þ Bg ðRp Rs ÞÞ þ Wp Bw
108 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
with
co Soi þ cw Swi þ cf
ceff ¼
1 Swi
it follows that
Wp Bw
Np Bg ðRsi Rs Þ þ ceff pBoi N
¼ ð2:14Þ
N Bo þ Bg ðRp Rs Þ
Tz p0
Bg ¼
pT0
Gp ¼ Np Rp
qg
Rp ¼ when qg and q0 ¼ constant
q0
N Rsi Bg N Boi þ N Bo N Rs Bg ¼ Np Rp Bg þ Np Bo Np Rs Bg
with
Bo þ ðRsi Rs ÞBg ¼ Bt
Boi ¼ Bti
it follows that:
Np Bt Bti
¼ ð2:16Þ
N Bt þ Bg ðRp Rsi Þ
The index 1 refers to the gas – oil ratio derived from the material balance.
Change of saturation
OOIP ¼ NBoi
N Boi
Vp ¼
1 Swi
ðN Np ÞBo
So ¼ ð1 Swi Þ ð2:17Þ
N Boi
Np Bo
So ¼ ð1 Þ ð1 Swi Þ
N Boi
qg qo Rs
þ
qg Bg Bo
Rp2 ¼ ¼ qo
qo To ; po
Bo
The index 2 refers to the gas – oil ratio derived from flow equation.
qg Bo
Rp2 ¼ þ Rs
qo Bg
kg A
qg ¼ r pg
mg
ko A
qo ¼ r po
mo
kg mo Bo
Rp ¼ þRs ð2:18Þ
ko mg Bg
Gp2 ¼ Np Rp2
Fig. 33. Graphical method of solution in the Tarner method. a) Material balance; b) Flow
equation.
If the permeability is equal in all nodes (ie, in all blocks) and the areas of
flow are likewise identical (since the block width is constant), the general form of
flow equation for the Darcy law (see Section 2.2.1.1) can be written as follows:
Ak
qi ¼ pi ð2:20Þ
mx
Since the coefficients before Dpi are constant in all terms, Equation (2.21)
can be written in place of Equation (2.19)
For node 8 there then follows the pressure balance set up in Equation (2.21)
or rearranged to:
q5 þ q7 4 q8 þ q9 þ q12 ¼ 0
Such a pressure balance can be set up for each of the 13 nodes, in addition
radial flow conditions must be included in the pressure balance for nodes 1 and
15. The 15 equations with 15 unknowns are arranged in matrix form and solved
for the pressure values at the nodes with, eg, the Gauss – Jordan or Gauss – Sei-
del method. The calculation can also be performed with different permeabilities
in the individual blocks.
Model for Unsteady-State Flow. If the same model is used, but instead
of an incompressible fluid a compressible medium is injected into the well, then,
instead of a material balance for the individual volume flows in the nodes, a bal-
ance for the mass flows must be set up. As already derived earlier from the mass
balance in a cylindrical volume element, a variant of the diffusivity equation
must be used to describe unsteady-state flows. For linear flow through the
volume element of edge length Dx, the following differential equation can be writ-
ten for the change of flow along the distance Dx:
kA p kA p
t ð2:22Þ
m x xþx m x x
t ¼ Axfc p
After separating the variables Dx and Dt, the diffusivity equation for
unsteady-state flow is obtained (Eq. 2.23):
p k ð p=xÞ
¼ ð2:23Þ
t fmc x
p jþ1 2 p j þ p j1
¼
x2
114 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Equations (2.27) and (2.28) can give the same result only when the time
interval Dtn approaches zero. If the pressures at time n in all segments are
known, the pressure in element j at time n þ 1 can be calculated with Equa-
tion (2.27). This method is known as the explicit solution method. The advantage
of the explicit solution method is that it is very easy to program. However, it is
only accurate if very small time steps are chosen.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 115
In contrast to the explicit solution (Eq. 2.27) in which there is only one
unknown, ie, the pressure pj at time n þ 1, three unknown values are present
in Equation (2.28) at the end of the interval (ie, pj1, pj, and pj+1).
If, however, Equation (2.28) is formulated for all three nodes, j 1, j and
j þ 1, the three unknown pressures can be determined from the three equations.
The equations are arranged in matrix form and solved simultaneously either by
the Gauss – Seidel or by the Gauss – Jordan method. This procedure is known
as the implicit solution method and has the advantage of yielding sufficiently
accurate results even at quite large time intervals. For modeling problems in
multiphase flow, the saturation equation for each block of the reservoir model
must be solved in addition to the pressure equation. Hence, the chosen block
size of a reservoir model is influenced not only by the availability of geological
and petrophysical information but also by the accuracy of the numerical compu-
tation method. In regions of the reservoir where relatively rapid changes occur in
the transport characteristics — such as pressure gradient, transmissibilities, and
saturation changes — fine-mesh grids are necessary for numerical accuracy of
the solution method. For complex reservoir studies, simulation models with sev-
eral ten thousand blocks can be produced which require considerable computer
time. There are numerous simulation codes with a dynamicized fine grid size
that follows the movement of flood fronts.
The powerful computer programs currently available to the natural gas and
petroleum industry not only take into account simple processes such as gas dis-
solution, gravitational force, and capillary pressure effects but also can describe
the infuence of compaction in unconsolidated formations and the influence of
the diffusion and sorption of components. They can also treat particular rheologi-
cal properties of the phases, eg, the structural viscosity of polymer solutions in
polymer flooding the phase behavior during mixing processes, temperature-
controlled fractionation of materials (distillation and cracking), thermal conduction,
and thermal convection. A survey of common simulator types is given in Table 3.
Statistical Prediction Methods (Decline Curve Methods). Aside from
tank models or flow models to describe and predict reservoir behavior, statistical
prediction methods also can be used. These methods are based on observation of
the decline of production or pressure in a reservoir; analysis of the mathematical
trend; and extrapolation of the decline of production or pressure over the period
of prediction. The observed trend curves often show uniform patterns. This is due
to the fact that the change in the flow rate is proportional to the pressure change,
and the pressure change is linked logarithmically, as a result of the compressi-
bility relation, with the volume removed from a closed system. In many cases,
therefore, an exponential decline of the production rate with time can be derived
from the trend analysis.
To characterize the decline curve analysis method, some basic definitions
are given in advance (98). The fractional production decline is defined as the
change of flow rate divided by the initial flow rate
dq
¼ fractional production decline
q
The change of the fractional production decline with time is called the pro-
duction decline rate D.
dq
¼D
q dt
dq
D¼ ¼ a0 q0 ¼ a0
q dt
after integration
Zq Zt
dq
¼ a0 dt
q
q0 0
q
ln ¼ a0 t ¼ Dt
q0
q
¼ eDt
q0 ð2:29Þ
q ¼ q0 DNp ð2:30Þ
A plot of production rate q versus Np is shown in Figure 38. From this plot
the production decline rate D is also obtained from the gradient. As soon as
118 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 38. Exponential production decline, production rate versus cumulative production.
the production rate of a well falls below a marginal value determined by the cost,
production must be stopped. The corresponding cumulative amount produced
can be determined graphically by extrapolating the production decline curve to
the known limiting production rate qa. The difference between the maximum cu-
mulative production Npa and the present cumulative production Np gives the re-
serves of the well or reservoir.
Harmonic Production Decline. The defining equation for harmonic
production decline is:
dq
D¼0 ¼ a0 q
qdt
D0
D0 ¼ a0 q0 ; a0 ¼
q0
Zq Zt
dq 0
¼ a dt
q2
q0 0
1 1 D0
¼ t
q q0 q0
q0
q¼
1 þ D0 t
q
log ¼ logð1 þ D0 tÞ ð2:31Þ
q0
Next Page
Fig. 39. Harmonic production decline, production rate plotted against cumulative
production.
For harmonic production decline, the double logarithmic plot of the produc-
tion rate against the time function (1 + D0t ) gives a straight line of gradient 1.
Again, if a new quantity is introduced by combining q and t a semilogarithmic
relation between production rate and cumulative production is obtained after
integration (see Fig. 39).
tNp
Z
Np ¼ qdt
0
tN p
Zq Z
dq
¼ a0 qdt
q
q0 0
q D0
ln ¼ Np ð2:32Þ
q0 q0
dq D0
D¼ ¼ a0 qn a0 ¼
qdt qn0
Previous Page
Fig. 40. Hyperbolic production decline, production rate plotted against the time
function.
q0
q¼ ð2:33Þ
ð1 þ nD0 tÞ1=n
qn0 1 1
Np ¼ ð2:34Þ
ð1 nÞD0 qn1 0 qn1
Exploratory wells must be drilled in such a way that they provide accurate
information on the rock formation drilled through and on the deposit to be
explored. This is done by sampling the rocks, eg, in the form of drill cores, sam-
pling the materials present in the rocks (eg, water, oil, gas), gathering data from
a great variety of geophysical borehole measurements, carrying out productivity
tests, and then measuring the resulting behavior of the reservoir pressure.
Production wells must be drilled and completed in such a way that the cost
of subsequent production is minimal, as much as possible of the reservoir con-
tents can be produced economically, optimum use is made of its productive capa-
city, a minimum number of wells is required to develop the reservoir and trouble-
free production can be achieved, preferably over the reservoir’s lifetime (120).
Boreholes compete with mining methods as a means of prospecting for and
producing raw material and energy, and are in many cases, superior to them
from economic, operational, and safety viewpoints. Deep drilling engineering is
therefore becoming increasingly important with th development of deposits and
reservoirs that lie ever deeper and in inhospitable regions, under the oceans, or
in perpetual ice.
History (124–126). Deep drilling is far from being a new technology
(Fig. 41 ). As early as 600 B.C. CONFUCIUS reports that wells have been sunk in
China to a depth of about 500 m into salt brine with hammer drilling devices.
In Europe, both in ancient times and in the Middle Ages, wells were sunk to
obtain water and later for prospecting shallow ore lodes. Development of
drilling engineering and well depths that could be achieved stagnated for
many centuries. The search for raw materials (potash, iron ore, and coal) in the
course of the industrialization in the 19th century forced the development of
deep drilling engineering. At that time, drilling depths already reached
2000 m.
In Europe during the second half of the 19th century and the start of
the 20th century a extensive search for raw materials took place. It was pioneers
like VON OEYNHAUSEN, FAUCK, CRAELIUS and ANTON RAKY who encouraged the pro-
gression of deep drilling by, for example, developing high-speed rotary percus-
sion drilling with circulating drilling mud. Only then could the frozen
boreholes for shaft sinking of mines for coal, potash, and rock salt be produced.
The rotary drilling method, developed in the United States at the end of
the 19th century for the exploration and development of oil deposits, was
an important step forward. Since then, rotary drilling has been continuously
developed and today is still the dominant drilling method by which depths
of more than 9000 m have been reached in commercial boreholes. The
present depth record has been achieved with the Kola SG 3 well in the Com-
monwealth of Independent States, which reached a depth of 12 261 m, using
turbodrilling.
Rotary Drilling (127–131). The circulating viscoplastic drilling mud and
powerful roller cone drill bits are the most important characteristics of rotary
drilling. These, in combination with other typical features, have contributed to
the overriding importance of rotary drilling in the exploration and development
of deep and high-pressure reservoirs. Spectacular blowouts, which were numer-
ous in the past, are now rare.
The principle of mode of action in rotary drilling is as follows (Fig. 42):
By means of a rotary table (e), a torque is applied to the drill string that
consists of individual drill pipes (steel tubes). The drill string is suspended in
the derrick by a hook (b) attached to a traveling block. This hook, which allows
the drill string to rotate, can be moved vertically with a drawwork on the rig
floor. The drill string transfers the torque to the drill bit (n), whose impact on
the rock is controllable via the weight of the drill string and drill collars (m)
(massive thick-walled tubes arranged above the bit), which is not taken up by
the drilling hook.
The rotary table, which is a ‘‘surface drive’’, can be replaced by a hydraulic
drilling motor immediately above the bit. This downhole motor, which can be a
positive-displacement machine on the Moineau principle or a multistage turbine,
is installed in the drill pipe and driven by the circulating mud.
Drilling Tool. The drilling tool, which is adapted to the rock through
which the hole has to be sunk, performs the drilling work on the hole bottom.
This can be done by machine drilling (cutting bit), fragmentation drilling (roller
cone bit, Fig. 43 A – E), or abrasive drilling (diamond bit Fig. 43 G).
The key factor with regard to drilling costs is the rate of penetration (v). The
rate of penetration depends on the rotary speed of the bit n, the weight on bit
P, the drillability of the rock S, and the hole diameter D. In the empirical
relation
nP2
v¼ k
D2 S2
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 123
Fig. 42. Diagram of a rotary drilling system. a) Parked drill pipe; b) Drill hook;
c) Rotary swirrel; d) Flexible hose; e) Rotary table; f) Preventer; g) Vibrating screen
(shaker); h) Pit; i) Cement sheath; j) Casing; k) Hollow drill pipe; l) Circulating drilling
mud; m) Drill collar as part of the drill string; n) Roller cone bit; o) Rock strata;
p) Reservoir rock horizon.
124 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 43. Various types of bit A) Roller bit with coarse teeth (for soft formation); B) Roller
bit (for medium hard rock); C) Roller bit with fine teeth (for tough rock); D) Roller bit with
hard metal inserts (for very hard formations); E) Roller cone bit; F) Jet bit; G) Diamond
full-hole bit.
The drill string is loaded by tension (the drill pipes’ own weight), torsion,
bending (resulting from borehole deviations), constant or dynamic internal pres-
sure and buoyancy forces, and periodical vibration. In some cases, the super-
posed stresses in drilling can only be calculated with difficulty.
Drilling Mud. The cuttings are carried in the annulus between drill pipe
and borehole wall by the mud. A thixotropic liquid, containing bentonite and cel-
lulose for example, is used as the drilling mud. Its functions include cooling and
lubricating the drilling tool and cleaning the borehole bottom, discharging the
cuttings, and supporting the borehole wall (hydraulic casing). The mud is
pumped with piston pumps through the drill pipe to the bit, from which it
emerges at high speed to ensure that the cuttings are removed as completely
as possible. The cuttings and finer solids are removed from the mud at the sur-
face and the mud is then recycled by pumping it back again into the hole. Vibrat-
ing screens, centrifuges, and hydrocylones are used to seperate the solids
(Fig. 44).
The mud parameters (eg, the rheological properties, the mechanical, and
thermal state of the colloid system) are adjusted according to the drilling
speed, pump power, bit type, borehole diameter, pore pressure in the rock, tem-
perature, nature of the rock drilled (mud loss zones in the drilled formations),
borehole inflows, and solids content.
Fig. 44. Processing of drilling mud, schematic diagram. a) Flow measurement; b) Mud
pumps; c) Gas traps; d) Sieve; e) Centrifuge tank; f) Degasser tank (vacuum degasser); g)
Intermediate tank; h) Suction tank; i) Pill tank; j) Water tank; k) Reserve tank; l) Mixing
tank; m) Moineau pump; n) Centrifuge; o) Cuttings pit; p) Sludge pit; q) Container, cut-
tings sampling; r) Mixer; s) Separator sampler.
126 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Mud densities between 1.10 and 1.40 g/cm3 are usually sufficient for com-
pensating the reservoir pressure and the ‘‘trip margins’’, ie, the piston effect
and inertia forces during pulling the drill string. In special cases, however,
values below 1 g/cm3 or up to 2.5 g/cm3 are also required.
The carrying power of the mud for the cuttings is given by:
Fy ¼ Vð%G %mÞg
where F is the surface area of the cutting particles (cm2); V the volume
of the cuttings (cm3); %G, %m are the density of rock, and mud, respectively
(g/cm3), g is the acceleration due to gravity (dm/s2), and y the static shear stress
(dyne/cm2).
The filtration of water from the mud in porous rock strata, with the forma-
tion of a filter cake, proceeds according to Darcy’s law of filtration. The quantity
of filtrate is proportional to the permeability of the filter cake, the applied pres-
sure difference, and the active filter area, and inversely proportional to the fil-
trate viscosity and the filter cake thickness.
Pressure and temperature changes, particularly in deep holes, make it
difficult to assess and calculate the formation of the filter cake. The filter
cake can attain a high mechanical strength and cause drilling difficulties.
Prolonged filtration can also lead to considerable reservoir damage.
The electrical conductivity of the mud and its pH value and contents of
chloride, sulfate, calcium, and hydrogencarbonate ions are continuously moni-
tored. Additives used to adjust the mud system include:
There are water- and oil-based drilling muds. Aside from the technical dril-
ling requirements for the mud, the following requirements should be met: the
formations drilled through remain identifiable (mud-logging see below, indica-
tion of traces of gas and oil); borehole measurements, tests, core drilling, and
other borehole studies can be carried out; the mud causes as little damage as pos-
sible to the formations drilled through (obstruction, swelling, mud losses); and
corrosion of the drilling equipment is limited.
Drilling muds for use in geoscientific research such as in the German Con-
tinental Deep Drilling Program (kontinentales Tiefbohrprogramm, KTB) the
main well must satisfy not only the drilling engineering requirements but
also those of the geoscientists. For example, such drilling mud must have ade-
quate rheological properties, lubrication, and borehole stabilization without
influencing the geoscientific analysis or even disturbing it.
Data on the drilling parameters, mud properties and lithology are col-
lected and stored in a mud logging system by modern data acquisition and
a computer system for drilling rigs. A network of monitors, work stations,
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 127
Fig. 45. Continental Deep Drilling Program (KTB), drilling and casing diagram.
Fig. 46. Drilling rig UTB 1 for the 10 000 m main well of the Continental Deep Drilling
Program (Germany).
variant of directional drilling engineering. The size and performance data of the
drilling rig for a superdeep hole clearly exceed those for conventional extra-heavy
rigs.
Horizontal Drilling. The second main focus of development in deep dril-
ling engineering is directional drilling where the drilling of holes along a defined
course over long distances and at considerable depth in the underground rock is
now possible and economic (Fig. 48).
Offshore, virtually all wells are directional drilled holes, and the oil and gas
fields are developed with up to 60 wells per platform. Onshore, the classical
applications of horizontal deep drilling include recovery under residential
areas and traffic routes, borehole deviation in the event of damage, and relief
wells in the event of blowouts. A directional drilled hole is produced by using
mud-driven drilling motors. The deviation is created with a bent shaft (bent
sub) combined with the drilling motor, or the motor itself is tilted. In Western
Europe, 85 % of the onshore wells are directional drilled.
Inclination and azimuth (ie, the course of the borehole) are continually
measured and the values transmitted to the surface via pressure pulses through
the mud column. From the surface the azimuth can be corrected accordingly. The
continual checking, known as measuring while drilling (MWD), is crucial for
success. The azimuth is measured by magnetometer or, if in the neighborhood
of ferromagnetic masses, by gyro compass. The inclination is evaluated by accel-
erometers (pendulum). In addition, these instruments indicate the tool face,
ie, the angle between the desired direction and magnetic north (notch and
bead). The availability of powerful, slow-running drilling motors together
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 129
Fig. 47. A) Active steering vertical drilling system; B) Cross section of X; a) Flex sub;
b) Data pulser; c) Sensors; d) Top stabilizer; e) Electronic pack; f ) Generator; g) Control
valves; h) PDM – downhole motor; i) Nonrotating body; j) Hydraulic pistons with movable
ribs; k) Rotating shaft; l) Drill bit.
with the MWD technique are prerequisites for successful directional drilling
engineering.
The inclination build-up of a hole to be deviated into the horizontal can be
abrupt or gradual (see Fig. 49). At an inclination change of 108/10 m, the result-
ing deviated section has a radius of 6 m, whereas at 18/10 m, the radius is ca.
600 m. Holes have been deviated to reach destinations of up to 3000 m, and
holes exist which have been drilled horizontally for almost 1000 m.
3.2. Investigation of the Reservoir Rocks (144,145). With suitable
drilling tools, rock samples (mud samples, drill cores) are obtained from the
reservoir rock zone and can subsequently be examined comprehensively. Special
geophysical measurement (logging) and the first, brief productivity tests (drill-
stem tests) are also carried out.
Core Drilling. During core drilling, the drilling bit (full-hole bit) is
replaced by a core barrel with crown (ring bit). Single- or double-tube core barrels
130 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Fig. 48. Selected horizontal wells curvature. a) Texaco, 1981 (horizontal section
>300 m); b) Esso Resources Canada Ltd., Cold Lake 1978 – 1984 (horizontal section
1000 m); c) Esso Resources Canada Ltd., Norman Wells, 1981 (horizontal section 500 m);
d) Elf France, CR 163, 1986 (horizontal section 360 m); e) Elf France, LA 90/LA 91, 1980 –
1981 (horizontal sections 270 and 470 m); f) BP 1985 (horizontal section 324 m); g) Elf
Italy, RSM 6, 1982 (horizontal section 600 m); h) Elf Italy, RSM 211, 1986 (horizontal sec-
tion 600 m); i) Preussag, 1984 (horizontal section 200 m); j) Marsvoskoe, 1968 (horizontal
section 320 m); k) Elf France CLU 110, 1983 (horizontal section 330 m); l) Sohio, Y 20, 1986
(horizontal section 305 m); m) Sohio, B 30, 1985 (horizontal section 452 m); n) Sohio, JX 2,
1986 (horizontal sction 415 m); o) Sohio, BW 1, 1985 (horizontal section 579 m).
are used, some with diamond-type cutter heads. Single-tube core barrels have
the drawback that the core is directly exposed to the mud (core losses); With dou-
bletube core barrels this is prevented by the inner tube of steel or rubber
(Fig. 50). Wedgejaw catchers, flap catchers, or combinations of both prevent the
core from falling out. Orientation devices and core markings allow an accurate
determination of inclination (dip), course (strike), and pattern of the strata
drilled through.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 131
Fig. 50. Schematic diagram of double core barrel a) Coupling to drill pipe; b) Overshot
for pulling out with wireline; c) Outer tube, rotates core bit; d) Inner tube, nonrotating;
e) Bearing; f) Mud passage; g) Ball valve; h) Core catching spring; i) Core catching
cone; j) Mud channels.
Fig. 51. Comparison of oil and mining core drilling A) Oil core drilling (double tube core
bit); B) Mining core drilling (cable core bit).
During these tests, samples of the reservoir contents are also obtained, so
as to carry out chemical and physical studies. The results serve as a basis or aid
to planning decisions on: the completion of the well, well and reservoir treat-
ments, the production method, the surface installations, the further development
of the field, and exploitation strategy.
3.3. Casing and Cementation. Casing (146). The casing of a borehole
usually consists of several concentric strings of casing pipe, which are installed
according to the geological – engineering requirements at different depths
(Fig. 42). It protects the borehole wall against scouring and blowouts, seals
permeable horizons to prevent undesirable inflow or outflow, and prevents
sloughing from the formation into the hole. The casing is in part already inserted
during the drilling operation, in order to protect problem sections of borehole. It
represents a considerable cost component of the well.
With exploratory wells, the planning of the casing program is rather diffi-
cult, since knowledge of the formation is still imperfect. For production wells a
standard program is usually set up, in which tube dimensions, wall thicknesses,
and landing depths are optimized.
The casing can either be run to the surface and landed in the wellhead, or hung
in the borehole as a so-called liner in a string of casing already installed in the bore-
hole. To ensure a tight bond between casing string and formation, the casing is
cemented, either up to the surface or only over a certain length of the borehole.
The casing is subject to a multiplicity of loads, which generally arise in
combination:
1) Massive bond of casing and formation, in order to divert the forces result-
ing from the loading of the casing into the formation, ie, absorption of the
axial and radial stresses which act on the casing string
2) Isolation of permeable formations, in order to prevent flow between differ-
ent rock strata and, in particular, the contamination of the groundwater
3) Corrosion protection of the casing
The cement is mixed with water to form a cement slurry, and a wide
variety of additives are added to control the rheological properties, the
setting time, the final strength, etc. The slurry is then injected with high
pressure pumps through the borehole into the annulus between casing and
formation.
The cements used in the oil and gas industry are standardized according to
API. Table 4 gives some important characteristics of various API cements. Aside
from the cement quality, the technical execution of the cementing is of crucial
importance for the quality of the cement sheath. For control, the cement bond
log is used: this is an acoustic measurement method for determining the bonding
of cement and casing or of the cement with the formation.
A D, E, F G, H
The elevated temperature lowers the hardening time, so retarders and anti-
friction agents must be added.
3.4. Production Engineering (148,149). Objectives (150–153). The
purpose of the exploitation and production planning of hydrocarbon reservoirs
is to produce optimum amounts of saleable (transportable) products at mini-
mum cost and with close attention to all aspects of safety and ecology. In
initial stages of field development, with few wells and few drill-stem and pro-
duction tests available, the base for planning is not valid enough. However,
the data obtained from continuous production and field development lead to
continual refinement of production planning, in which reservoir and engineer-
ing are regarded as a synergism. Only in this way can an optimum overall
technical and economic result be achieved.
Production engineering has to solve the following problems in oil
production:
1) The time for water injection – when not using it to adjust the pressure
above the bubble point and/or for adjusting the dynamic level to below a
certain material stress, corresponding to the fatigue endurance limit of
the pump rods
2) Dependence of the productivity index (IPR) on viscosity of the oil and water
cut
3) Gas production and availability in the gas lift method
4) Advantages and disadvantages of the artificial lift methods
The safety and ecological aspects have always been important; nowadays it
is preeminantly so. In the second half of the 19th century, a blowout was proof of
discovery after finding the reservoir by drilling and a reason for celebration.
Thanks to the continual improvement of technical equipment and the regular
training of technical staff, blowouts have become rare, but they can never be
totally avoided due to human error.
Over wide areas, oil and gas production engineering have common princi-
ples, which are treated in the following section whereas specific questions of
production engineering for oil and gas are dealt with.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 137
Fig. 52. Production equipment of a typical gas well. a) Nozzle; b) Block-type Christmas
tree with five shut-off valves; c) Double P-seal; d) Hanger nipple with seat for plug;
e) Tubing seal; f) Flange slide valves; g) Tubing double flange with additional casing
seal; h) Bottom flange; i) Corrosion-inhibited protective liquid; j) API tubings; C 75 or
J 55; k) Intermediate casing; l) Circulation slide valve with seat for plug; m) End casing;
n) Tubing receptacle, overshot type, with seat for wireline plug; o) Tubing anchor; p) Pro-
duction packer; q) Perforation.
Fig. 53. Bottom-hole equipment in the reservoir zone. A) Main dolomite, uncased;
B) Carboniferous, casing with liner; a) Tubing, 2 7/8 inch (2.22 cm); b) Cement;
c) Production packer; d) Casing, 7 inch (17.8 cm); e) Liner hanger; f) Liner, 5 inch
(12.7 cm).
The method of tubing-coupled perforating (TCP) (Fig 55), which has been
available since ca. 1980, has captured a significant market, because of the fairly
large number of deep gas wells and the rapid development of horizontal drilling.
By this method the charge carriers are installed on the tubing string below the
packer and landed with the latter. They are then connected (wet) in a separate
run with an ignition mechanism on the cable, and then ignited. The advantages
of this method are
1) A greater shot density — 39 shots per meter with the TCP instead of
13 shots per meter conventionally
2) Considerably longer perforation sections, since the load limit of the cable
used in conventional perforation is no longer limiting
3) No hanging of the charge carriers in the inclined hole (> 708)
4) Greater safety, because the charges are sensitized underground
In practice the first two causes occur most frequently. The obstructions are
usually removed by chemical treatments. The most important chemical treat-
ment agents are: aqueous solutions of hydrochloric acid (5 – 15 wt %), hydrofluo-
ric acid (1 – 6 wt %), acetic acid, citric acid, and surfactants (all in concentrations
of 0.1 – 5 wt %).
Corrosion inhibitors are added to the acids. Hydrochloric acid is generally
used with carbonate- containing reservoir rock. Whose carbonate content should
be ca. 20 %. The hydrochloric acid partly dissolves the rock. Hydrofluoric acid,
usually in conjunction with hydrochloric acid, is used for dissolving silicates:
here, the hydrofluoric acid is added to the hydrochloric acid as NH4HF2.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 141
order to limit this effect, which is unfavorable for the success of the operation,
fluid loss agents are added that raise the viscosity and form a filter cake.
These agents cause the initially high fluid loss to decrease substantially after
the filter cake has built up. Later, these additives become ineffective due to
the influence of temperature and pressure, and the production stream reaches
the well unhindered.
In order to ensure the success of reservoir bed treatments, a sufficient depth
of penetration into the reservoir rock must be achieved. This penetration depth
must be greater with less permeable reservoir rock. Penetration depths of
40 – 70 m are usual with vertical or horizontal fractures.
The development of deep (ca. 5000 m), low-permeability gas reservoirs in
the Rotliegendes and Carboniferous has led to remarkable developments of
this technology. New developments now include a temperature-stable and suffi-
ciently viscous supporting fluid, the availability of pressure-resistant proppants,
and powerful equipment. For example, a treatment volume of 1100 m3 requires a
fracture length of 300 m, maximum pressure of 750 bar at a depth of 4200 m, and
15 pumping units with an output of 9000 hp.
Workover. An important chapter of constant concern for the production
engineer is workover technology. More than 106 production wells worldwide, of
which three quarters are oil wells, require continual upkeep and maintenance
and — repeatedly during the course of their use – repair. Thus the number of
workover hoists is three to four times the number of drilling rigs. In addition,
there are the modern systems, ie, the wireline technique (lowering activation
of tolds by wireline), the coiled tubing technique, and hydraulic lifting and low-
ering of the strings with the snubbing unit. Workover should always follow a
sequence of diagnosis, therapy (work program), operational execution, and mon-
itoring of success. Reliable measurement techniques and supervision of operation
and production are the prerequisites for correct diagnosis. The workover hoist,
equipped with rotary table or power swivel and pump unit, can carry out the
functions of lifting, lowering, rotation, and circulating fluid, likewise the snub-
bing unit. With the wireline technique, only running in and out is possible.
The coiled tubing (Fig. 57) enables lifting, lowering, and circulating. Rotation
Fig. 57. Schematic of a coiled-tube unit. a) Drive unit; b) Operator’s cabin; c) Coiled-
tube drum; d) Coiled tubing; e) Injector crane; f) Injector head; g) Preventer device; h)
Wellhead.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 143
is possible by using suitable downhole motors. The coiled tubing technology has a
high technical standard and a growing market, particularly also with regard to
horizontal drilling. Coiled strings with diameters of 1 inch (2.54 cm), 11/4 inch
(3.18 cm), 13/4 inch (4.44 cm) and 2 inch (5.08 cm) are available, and can be
used at a depth of up to 5500 m and in combination depending on the loading
situation.
Horizontal Wells (132,135). Horizontal drilling, which has been deve-
loped to a state of maturity since the beginning of the 1980s, makes great
demands on production engineering. Drilling engineering is able to pro-
duce, horizontal drilled sections of a few hundred to more than 1000 m in
any rock. In the deviation to the horizontal, the inclination build-up can be
abrupt or gradual (Fig. 46). Whether this value is 108/10 m (final radius
60 m) or 18/10 m (final radius 600 m) is very important for the subsequent
completion.
The essential reason for the rapid growth of horizontal drilling is that a
markedly larger zone of the reservoir is developed for production than with ver-
tical wells, so that a considerable (15 – 20-fold) production increase is achieved at
lower differential pressure; better sweep efficiency; and a higher recovery. In
heterogeneous, fissured rock the effect is even more significant. The potential,
especially if specific enhanced oil recovery (EOR) processes are incorporated in
this technology, is far from exhausted. In gas reservoirs, turbulence around
the borehole is avoided, and in underground storage, a capacity increase of
3- to 8-fold seems possible with horizontal drilling. There is still, however, a
considerable need to develop devices and methods specific to the process,
cf. Schlumberger coiled-tubing-logging (CTL). Completion in deviated and hori-
zontal wells can be carried out as follows:
flexure in the deviation zone, and the necessary adaptation of systems must be
carried out.
Oil Production Engineering (159–163). In oil production engineering, not
only the technical instruments and developments determine success. Efficiency
and overall profitability crucially depend on qualified supervision, damage eva-
luation, the logistics of material and distribution, and a sound operating organi-
zation. The use of tertiary methods (eg, steam injection and steam flooding)
signify additional tasks for production engineering: In steam injection wells,
and also in the hot production wells of the steam flooding projects, extreme
thermal stresses (140 bar, 340 8C) and corrosion problems in the underground
equipment must be kept under control.
Production engineering has powerful production processes and production
facilities available to it. Herewith, the particular conditions of each reservoir can
be taken into account. This in turn enables, firstly, the reservoir fluids to be pro-
duced as efficiently as possible with extensive use of the reservoir energy; and
secondly, production difficulties, caused for example by accumulated sand or
water or by corrosion, to be obviated in good time.
With the opening up and development of a reservoir, its most important
qualities are determined: reservoir pressure, permeability of the rocks, viscosity
of the production fluid, and the presence and characteristics of a natural water
drive. These are the foundations for planning production rates and methods, and
equipment.
In principle, the following production processes are available:
1) Flowing production
2) Gas lift
3) Centrifugal pumps
4) Piston pumps, sucker rod, or hydraulic
With these methods, ca. 3.25109 m3 crude oil per year are produced from
depths of up to 6000 m.
Flowing Production (156,164). As a result of production, a pressure
drop occurs which expands, funnel-shaped, around the hole. Under the influence
of this pressure gradient, the reservoir liquid flows to the borehole, enters the
casing through the perforation holes, rises up in the tubing string to the surface,
and flows at the surface by lines to a collection manifold at a central plant. Here
the output from the field is combined, dewatered, treated, and transported to the
refinery. The flow pressure at the depth of the reservoir bed is sufficient to over-
come the hydrostatic pressure of the production column and the frictional loss in
the tubing string and in the surface lines.
The flowing production rate is calculated as a function of bottom flow pres-
sure, tubing diameter, and pressure head according to the law of conservation of
energy, while taking into account the empirical parameters which determine the
two-phase flow.
The relationships become more complicated when the pressure in the
tubing string falls below the degassing pressure (bubble point) and free gas
evolves. Progressive pressure reduction gives rise to larger gas bubbles
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 145
Fig. 58. Flow patterns in two-phase flow. A) Bubble flow; B) Bubble flow (turbulent);
C) Slug flow; d) Froth Flow; E) Annular flow; F) Spray flow.
(see Fig. 58) and the liquid column breaks up (slug flow, Fig. 58 C), or the rising
gas bubbles entrain oil drops which, owing to friction at the tube walls, form an
upward-moving film (annular flow, Fig. 58 E) which finally changes into mist
flow, ie, the liquid film disappears and the liquid separates into droplets which
are mixed with the gas stream (Fig. 58 F).
The duration of this economically advantageous eruptive production phase
can be extended by a good matching of the production rate to the tube cross section.
Gas-Lift Production. In the gas-lift method, gas is injected into the
tubing string at one point or – to minimize the starting pressure – at several
successive points and in this way bubble or slug flow is initiated, as has been
described for eruptive production (see Fig. 59). This method is used continu-
ously or intermittently, depending on rate of flow, production rate, and gas
availability.
The eruptive production phase is often already supported with gas lift or is
switched to gas lift when the pressure gradient between borehole and reservoir
decreases. The applicability of this flexible production method is limited when
there is a shortage of gas or the gas compression costs are too high. Where suffi-
cient gas is available, the scope for application is wide because of the flexibility of
146 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
this method with regard to production rate and depth. In strongly deviated bore-
holes and if abrasive media are coproduced, gas lift is usually superior to other
production methods (no moving parts).
Gas lift valves controlled by the injection pressure are fitted in the tubing
string, into which they inject the lift gas. The pressure at the valve is controlled
via a pressure bellow which operates with or without spring support (according
to type). At the surface before installation the bellow is set to a certain opening
and closing pressure.
Since on start-up of a well in gas-lift operation, the back-pressure of the
whole liquid column requires a high injection pressure, several lift valves
are installed and distributed over the tubing string, whose opening pressures
are coordinated with each other.
Production with Subsurface Pumps (165,166). Subsurface pumps are
the most widespread form for the production of the oil, which is usually contami-
nated by considerable amounts of water. Types of pump used include electric
submersible centrifugal pumps and mechanical or hydraulic subsurface piston
pumps. The pumps are installed below the liquid level in the wells and, in diffi-
cult underground conditions, must overcome hydraulic heads of several hundred
to a few thousand meters (Fig. (60).
Electric Submersible Centrifugal Pumps (Fig. 57 B). These are multistage
centrifugal pumps that are submerged in the production liquid. They are
installed in the borehole, directly coupled with the motor, and suspended on
the tubing string. The pump unit comprises from bottom to top: electric motor
(bipolar asynchronous squirrel- cage, oil-immersed, j), centrifugal pump (with
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 147
Fig. 60. Different types of pump. A) Subsurface rod pump with walking beam drive;
B) Electric submersible centrifugal pump; C) Hydraulic subsurface piston pump with
pressure oil supply in the open system; a) Casing; b) Tubing string; c) Wellhead;
d) Pump rod; e) Subsurface pump; f) Drive pump, g) Production pump; h) Pressure oil
pump with i) Pressure raising pump on the suction side; j) Motor; k) Sealing adaptor;
l) Gas separator; m) Nonreturn valve; n) Tubing drain valve; o) Three-core cable;
p) Circulation piece; r) Pressure oil tank; s) High voltage transformer; t) Autotransformer;
u) Switch cabinet; v) Walking beam; w) Gearing; x) Tubing anchor; y) Production packer.
numerous stages e, f ), protector (sealing adaptor between motor and pump, k),
gas separator (for separating free gas from the produced liquid, l), electrical
switching device, and power cable (these limit the installation depth, since the
electric insulation resistance is temperature dependent).
The hydraulic head (hydrostatic pressure), frictional pressure losses,
and wellhead pressure determine the number of pump stages. Because of
possible changes in the hydraulic head, pumps with stable characteristics
are preferred.
At depths between 1000 and 4000 m, submersible centrifugal pumps cope
with minimum delivery rates between 260 and 50 m3/d and maximum delivery
rates between 340 and 150 m3/d.
Mechanical Rod Subsurface Pumps (Sucker Rod Pumping (167)) (Fig. 57 A).
Oil production was for a long time synonymous with the use of rod
pumps, which are installed as tubing pumps with the tubing or are landed at
the pumping depth as inserted pumps on the pump rods in the tubing string.
These pumps are driven from the surface by beam pumping units with motor,
148 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
gearing, and a pumping jack, that includes the typical walking beam with
‘‘horse’s head’’ on which the pump rod string is suspended. The weak point of
the system, the ‘‘pump rod string’’, requires particular attention. Liquid load,
rod string weight, vibrations, and mass acceleration, which occurs particularly
when the direction of motion reverses, determine the efficiency and the limits
of this kind of artificial lift.
The supply of types of subsurface pumps that differ according to material,
valve arrangement, piston area, and stroke length also take account of extreme
operating conditions. Valve function and pumping efficiency are periodically checked
by dynamometer measurements, ie, by recording the load uptake. Corrosion, sand,
or friction in deviated wells reduce the service life of pumps and rod string.
Hydraulic Subsurface Piston Pumps (Fig. 57 C). Like the electric submersi-
ble centrifugal pumps, these pumps are run underground, yet by high-pressure
fluid. The difficulties with the pump rod string are avoided by hydraulic
transmission of the automotive energy with an auxiliary liquid from the surface
to a pump installed in the borehole. A high-pressure pumping installation pro-
vides the ‘‘power oil’’ that is supplied to the individual producing wells and the
pump via the tubing string. The pump combines two identical sections in one
block, the upper of which operates as a motor which is driven by pressure oil
(f) and for its part drives the lower pump (g) which is hydraulically in complete
equilibrium.
The pressure oil is either taken from the produced oil and carefully cleaned
(open system) or is contained in a closed system. Usually the pressure oil is sup-
plied down the tubing and the flow back through the annulus or, where that is
not possible, it is fed through additional tubing strings. For repairs, the liquid
stream is reversed and the pump pumped to the surface (free pumps in contrast
to fixed pumps).
The range of application of hydraulic piston pumps exceeds that of rod
pumps by far. As with gas lift, a number of wells are connected to one power
oil high pressure pump station, respectively, so that an additional system of
lines is required. To account for the particular conditions of each well, a specially
devised flow and pressure control system is required.
With the electric submersible centrifugal pumps and rod pumps, neither
additional lines above ground nor an expensive control technique are used.
These basic differences are significant for economic efficiency (168).
Developments and Limitations. In addition there are jet pumps (mainly in
the United States); Moineau pumps with subsurface drive or driven by pump
rods from the surface; and turbine-driven submersible centrifugal pumps. How-
ever, these pumps are rarely used. The proven methods described above main-
tain their position, thanks to their high standard of technical maturity and
reliability. The correct choice of pump for the pressure and inflow conditions
(inflow performance) and the flow resistances in the production path (outflow
performance), both determined by the reservoir, and the product composition
(gas, oil) ensures a long and optimized service life. An example of this is Shell
Oil’s Reno field in Wyoming in the United States. There, since 1975, 190 wells
at depths ranging from 2000 to 4000 m have been converted from hydraulic
piston pumps to rod pumps for economic reasons alone. By using combined
pump rod strings of high-strength steel, a further developed heavy-duty beam
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 149
a b c
e f g h c
i j c
Well water
Gas
Oil
Water
Bacteria Injection well
Production well
Scale
Internal corrosion
Fig. 61. Flow diagram of an oilfield a) Compressor; b) Drying; c) Line; d) Gas separator;
e) Heater; f) Separator; g) Desalter; h) Tank; i) Secondary sedimentation; j) Filter.
(Mark II) drive, and subsurface pumps suitable for the great depths, 105 m3/d
wet oil with 30 % water are produced from 4420 m lifting depth. The field is mon-
itored by electronic dynamometers.
Collection and Treatment of Crude Oil (146). Production rates, crude
oil properties, and field conditions determine in what way and to what extent the
production of individual wells, groups of wells, parts of the overall field or several
neighboring fields are combined. Treatment at the field includes gas separation
and dewatering (Fig. 61).
Gas Separation. The degassing process starts already in the reservoir
and intensifies with the decrease of pressure and temperature through the tub-
ing string and the pipe system during production. Consequently, a two-phase
mixture enters the gas separators. To obtain a maximum liquid yield and, in par-
ticular, at high gas – oil ratios, a multistage separation should be aimed for.
However, apparatus cost and special requirements, such as, eg, gas delivery at
high pressure to the consumer, lead to the widespread use of two- and three-
stage separation.
Dewatering and Desalting. Since crude oil can still be produced at 95 %
water content, water – together with sand and contaminants – is removed as
early as possible to reduce transportation costs and corrosion. The refinery accep-
tance standard in Germany limits the total contamination (water and sludge) of
the crude oil generally to 1 % (salt content below 0.02 %). The water separates
only partly as ‘‘free water.’’ Water and oil mainly occur as an emulsion, which
must be broken by thermal, mechanical, chemical, or electrical methods, or by
a combination of these. At the same time, the emulsifiers present in the crude
oil (constituents that are concentrated at the interface) must be made ineffective.
The free water droplets must coalesce and, with the aid of gravitational or cen-
trifugal forces, separate.
150 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
Since the 1980s, a greatly simplified water separation method has evolved
from the variety of process combinations used. Particular features of separation
plants are:
1) Early feeding of the demulsifier in order to use the pipeline for internal
mixing
2) Moderate heating (‘‘cold treatment’’) to save thermal energy and reduce the
vaporization losses
3) Simple apparatus; the water separation can often be carried out in a single
tank, whereas previously a downstream fine purification in the gas-heated
treater or an electrical fine purification has been required
1) Borehole measurements for the early detection of danger and the subse-
quent establishment of a safe production rate which must not be
exceeded.
2) Retention of the sand by filters consisting of slotted tubes which can be
wound with one or more layers of wire or jacketed with a cylinder of plas-
tics-reinforced sand. When the product liquid flows into the filter, the en-
trained sand is filtered off, and then it deposits and/or fills the annulus
between filter and borehole wall. This effect is also achieved by artificially
jetted and injected sand packings (gravel pack, frac pack). The slot and
mesh widths of the filters and the particle size composition of the gravel
packings must be carefully matched to the particle size of the entrained
sand and the flow rates.
3) Artificial sand consolidation by two-component resins that are injected into
the formation. Shrinkage during the consolidation or flushing (flooding
after gelling) ensures the necessary permeability. Failures can force the
abandonment of a well.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 151
of tubing and packer. The necessity of complete and durable sealing makes extre-
mely high demands on the material and of machining precision. For gas wells,
special gas-tight tubing joints have been developed. In the selection of tubing
and its joints, the large stress resulting from high pressure and temperature
changes, causing large load reversals, must be taken into account.
The evolvement of reservoir pressure with time, its depth, the inflow area of
the well, the expected production period, and the possible liquid yield during pro-
duction determine the selection of tubing size and consequently the production
rate.
Packers seal off one or several producing zones from each other and from the
annulus. The sealing materials used must be temperature-stable, permanently
elastic, and gas- and H2S-resistant. For the further security of the well, the
annulus is usually filled with a protective liquid (inhibited water).
Essential elements of the tubing string are devices for well-engineering
operations, for the control and monitoring of production, and for remedial
measures, such as the injection of inhibitors or agents against hydrate formation.
These devices are installed, pulled out, and actuated on the wire line. Multizone
completions, borehole repairs, selective test and treatment operations, the killing
of wells, and gas-lift production require the selective, manipulable opening of the
tubing string. For this purpose sliding side door couplings are installed, which
are also actuated with tools on the wire line.
Production is controlled and monitored by control valves and measuring
sections. It is generally controlled as a function of rate or pressure. The rates
are determined by orifice meter measurements. (Fig. 63).
Fig. 63. Flow diagram of a natural gas drying plant with external cooling and pressure
increase. a) Prefractionator; b) Continuous heater; c) Glycol regeneration; d) Glycol pump;
e) Heat exchanger; f) Motor needle valve; g) Chiller (coolant NH3); h) Refrigeration plant;
i) Collector; j) Condenser; k) Piston compressor; l) Low-temperature separator (10 8C); m)
Glycol – oil separator; n) Condensate tank; o) Flare; p) Compressor plant; r) Measuring
section.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 153
Safety is particularly important. At the surface, for example, the valves are
normally kept open via a low pressure gas system. If the pressure drops in the
sensor system, these valves close the production line. With underground safety
valves, a safety valve kept open via a hydraulic control line closes the tubing
string as soon as the pressure in the control line falls. Other arrangements pre-
vent the freezing-up of control devices and lines at the surface by hydrate forma-
tion, and protect the surface installations from excessively high pressure.
Numerous exhausted natural gas fields are converted into natural gas
deep storage reservoirs, where demand and suitability allow (see Natural Gas).
Production engineering questions are also very important for storage reservoirs,
especially since withdrawal rates at up to ten times the production rate of a gas
deposit are aimed at.
Offshore Recovery (168,173). Liquid and gaseous hydrocarbons in
reservoirs under the sea bottom are recovered and produced by equipment
and devices which correspond in principle to those of onshore installations. A
distinction is made between supporting structures designed to incorporate dril-
ling rigs used only temporarily, and those to incorporate production and treat-
ment facilities which must be installed in the sea over the whole period of
exploitation of the deposit, frequently more than 20 years.
Because of the difficult environmental conditions and of the danger to the
staff from the processing of highly inflammable hydrocarbons in a very small
space (especially gases), the safety requirements for offshore units are particu-
larly high.
The positioning of the supporting structures over the defined locations in
the sea no longer presents difficulties in principle, since a very wide range of
navigation systems, such as satellite, Doppler, and radio navigation can be used.
Offshore wells require, for the distance between sea bottom and platform, a
special string of casing known as a riser, which must resist the weather condi-
tions, the motion of the sea, and the current. With exploration in ever deeper
water, platforms are being built and operated in ocean depths down to 600 m
(Fig. 64) and development and planning is already geared to ocean depths ca.
1200 m. The riser technology has also been pushed ahead; the safety engineering
aspects of continuous well monitoring under extreme weather conditions being a
priority. For the Shell Offshore East Coast Project a riser of 4000 feet (2121
meters) is planned in connection with a dynamically positioned drilling vessel.
Fig. 65. Mobile offshore drilling rigs. A) Submersible, water depth to 300 m; B) Jack-up
platform, water depth to 100 m; C) Drill ship, water depth to 500 m.
In deeper waters, in areas with shipping lanes, and for smaller fields, subsea
completion is sometimes used, in which wellhead and manifold are directly on
the sea bottom.
Supporting Structures for Offshore Drilling Rigs (Fig. 65). Barges
were the first constructional systems used for installing drilling equipment and
are still in use. These consist of flat hulls which are sunk (flooded) in several stor-
eys above each other in shallow water regions. Such installations must not be
exposed to currents and waves. The barge platforms are preferably used in
coastal swamp and mud flat regions, inland seas, and similar locations. These
installations can be used to water depths of up to ca. 10 m. They are dismantled
by refloating the barges.
Fixed Platforms. These are working platforms erected over the surface
of the sea on piles rammed into the bottom. Sometimes it can be appropriate
that only the derrick is erected on such a platform, while the remaining
equipment, materials, energy plants, crew accommodation, etc., are accommo-
dated on a ship anchored nearby. The field of use of these platform – ship
installations extends to ca. 30 m depth of water. However, this combined construc-
tion method can be used only in regions without difficult currents and weather
conditions.
Jack-Up Platforms (Fig. 62 B). Jack-up platforms are flat, multi-storey
barges that have three, four, six or more legs. As soon as these hulls have
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 155
been towed to their place of use, the legs are generally lowered vertically into the
sea bottom. The platform is raised on the legs so far above the sea floor (which
must be stable) that storm waves of maximum height 32 m (measured in centen-
nial storms) cannot beat under the hull of the barge.
These platforms are equipped for a payload of a few 1000 t of material,
depending on their operating location. The derrick and all necessary auxiliary
equipment, such as a generator (electric drive for the rotary table, hoist equip-
ment, and mud pumps), cementing, and mud installations, measurement ser-
vices, etc., are permanently installed. Crew rooms and a helicopter landing
deck are necessary support facilities.
Material is transferred from the supply ships by permanently installed
rotary cranes. The field of application of such rigs extends to about 100 m
depth of water. Because of good stability, the dependence of operations on the
weather is very small.
Semisubmersibles (Fig. 62 A) are floating offshore platforms that can be self-
propelled. The work platform rests on 3 – 8 slender columns, which themselves
rest on two long, parallel floats. The latter are usually submerged to
6 – 9 m during transportation to the location, to give more favorable sailing
conditions. At the location, sea water ballast is taken on and the draft increased
to about 20 – 25 m to lower the center of gravity. This results in small movements
with the heaving swell which together with the reduced water plane area pre-
vent large amplitudes of motion when in the drilling condition.
In addition, the semisubmersible is anchored to its location. Despite the use
of these measures there is an increased effect of waves and storms on these rigs
compared to other offshore units so that servicing is more restricted than of the
units previously described. The main fields of use are water depths between 50
and 600 m. The variable deck load reaches ca. 5000 t, depending on type. The
engineering installations required are the same as those of the jack-up platform.
Drillships (Fig. 62 C). Drillships were used very early in offshore engineer-
ing. Conventional ships were converted by erecting the derrick amidships and
also accommodating the other engineering installations on the ship. Under the
derrick a square recess (moonpool) was provided to make drilling possible. Gra-
dually special drillships with different shapes have been built for very diverse
applications. However, drillships are used only in small numbers since, owing
to their large motions with the heaving swell and despite dynamic positioning
over the borehole, they can be used only to a limited degree on the open seas
and oceans.
Demand also plays a part. Because of their greater hydrocarbon expecta-
tion, it is mainly the shelf edges which have been explored up to now, and the
deep sea hardly at all, so that drillships have largely been dispensed with. But
they offer considerable advantages in regions with fairly stable weather, because
of their high transit speed and fairly large carrying capacity (and therefore small
dependence on external supply).
The deep sea drillships ‘‘Joides Resolution,’’ and its predecessor the
‘‘Glomar Challenger’’ have undertaken studies of the deep sea bottom and its
underground as part of a broad program of geoscientific research.
Drilling rigs on fixed production platforms are used in fairly deep waters
where a field has already been explored by other drilling rigs and a larger
156 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
number of production wells must now be drilled. These platforms are already
prepared for later production. More than 40 holes can be drilled fanwise, direc-
ted to defined target points. The production installations for the extraction
of the hydrocarbons are installed later as required with the aid of rigging-up
ships.
Supporting Structures for Offshore Production Plants (Fig. 61).
The production platforms used for offshore oil and gas production are of the
same type as those for the drilling rigs of fixed platforms and semisubmersi-
bles. The first platforms were built as wooden structures, but since 1940
they have been constructed only of steel or concrete. All submersed steel
structures require corrosion protection in order to ensure a sufficient service
life. For this, direct protection paint and/or galvanization is used and in addi-
tion cathodic protection devices feeding external current or sacrificial anodes
are applied.
In the Norwegian and British sectors of the North Sea, several large con-
crete structures have been set down on the sea bottom. These consist of three hol-
low conical bodies of concrete, constructed on the sliding molding principle, that
contain, as an upper annex, multistorey production installations and personnel
accommodation; the derrick also is arranged there.
From these large structures, a fairly large number of deflected holes are
sunk. Two legs contain the production casing of the wells. The third leg receives
essentially the supply and line system. The base of the structure is provided with
a large number of concrete tanks in which the crude oil, salt water, and other
products are stored. The first stages of the structures were built in Norwegian
fjords, on a construction deck raised on sheet piles. Then, when the structure
had reached a certain stage, the installations were floated off and towed into dee-
per water, in order to allow immersion during the further weight increase result-
ing from concreting to the final height. The maximum total height of these
installations in the Brent and Statfjord fields is 250 m at a water depth down
to 175 m. A small wave pressure and short installation times at the location
(no ramming work) were reasons for constructing such structures, which, how-
ever, can be set down only on a suitably firm sea bottom. The largest offshore
platform in the world is 271 m high and has been installed in the Norwegian
Statfjord field.
Offshore Well Equipment. An important technical problem in drilling
at sea is the effect of the heaving swell and of the current on the sea bottom.
Movements and vibrations of the drilling rig in relation to the pipe strings
leading to the sea bed (riser and drill string) must be avoided. However,
these undesirable effects cannot completely be removed especially on floating
drilling rigs (semisubmersible and drillship). Devices are used to combat these
difficulties. They are installed in the hoist installations and in the casing
string (riser). For wells in shallow water areas there are no longer any special
problems. Here completion arrangements can be employed which are also in
use on land, with casing held in a ‘‘well protector’’ and led out above the
surface of the sea.
However, in fairly deep water and in the area of shipping lanes, underwater
completion must be used. This was first applied in 1943 in the 11.5 m deep water
of Lake Erie. Meanwhile, such systems, usually remote-controlled, have been
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 157
installed in a great variety of sea areas of the world. Installations that cannot be
reached by divers are not yet in use. Two types of completion exist: wet comple-
tion, in which all the components are in contact with the sea, and dry completion,
in which the components are in a pressure chamber under atmospheric condi-
tions, so that the personnel can carry out assembly and maintenance of the
installation by ‘‘submarine passenger ferry.’’ The criteria for the selection of sui-
table completion include water depth, weather conditions, areas with shipping
lanes, icebergs, mechanical properties of the sea-bottom and the possibility of
sediments wandering at sea bottom.
The use of divers, drilling rigs, or workover ships is required for necessary
repairs, possible replacement of the Christmas tree and the tubing, and workover
operations.
Logistics. As soon as the production of oil and gas in the offshore sector
is taken up, provision must be made for the transportation of these products.
Underwater pipelines and ships are primarily used. An essential prerequisite
for the smooth and secure operation of offshore drilling and production activ-
ities is the existence of a land base. The land base provides the maintenance
service, all supplies of materials, as well as the catering, intelligence, and
rescue services, and usually the exchange of personnel by helicopter, etc.
Acknowledgement
Chapter 1 is based on the article Erdöl und Erdgas, Bildung, Suche und
Vorkommen by Hans-Ulrich Hark { in the 4th edition of Ullmann’s.
Abbreviations
A cross-sectional area, m2
As swept area, m2
B formation volume factor
Bo,w,g formation volume factor for oil, water, gas
Bt two-phase formation volume factor
c compressibility coefficient, Pa1
cf pore volume compressibility, Pa1
co,w,g compressibility for oil, water, gas, Pa1
cpc pseudocritical compressibility
cpr pseudoreduced compressibility
cr reduced compressibility
C critical point
D production decline rate, s1
D hole diameter
EA areal oil recovery factor
ED microscopic oil recovery factor
EI vertical recovery factor
ER recovery efficiency
158 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING Vol. 1
f fractional flow
fo,w,g fractional flow for oil, water, gas
F surface area of cutting particles
g acceleration due to gravity, m s2
Gp gas produced, m3
h thickness, elevation, m
k permeability of rock, m2
k drillability coefficient
ko,w,g effective permeability for oil, water, gas, m2
kr relative permeability
kro,w,g relative permeability for oil, water, gas
l length, m
m gradient
Mr molecular mass
M mobility ratio
N initial oil volume, original oil in place, m3
Np oil produced, m3
Npa maximum cummulative production rate
OGIP original gas in place
OOIP original oil in place
p pressure, Pa
pb bubble point pressure, degassing pressure, Pa
pc capillary pressure, Pa
pe pressure at reservoir boundary, Pa
pm mean pressure
ppc pseudocritical pressure
ppr pseudoreduced pressure
pr reduced pressure
pwf flowing pressure in well, Pa
pws shut-in pressure, Pa
P weight on bit
q flow rate, m3 s1
qa limiting production rate
qo,w,g flow rate for oil, water, gas, m3 s1
r radius, m
re radius of area drained by the well
rw,c well radius, reservoir boundary radius, m
Rp produced oil – gas ratio
Rs gas – oil ratio
Rsi initial gas – oil ratio
S saturation
S drillability of the rock
So,w,g saturation oil, water, gas
Sor residual oil saturation
Swf front saturation
Swi initial water saturation, irreducible water saturation
t time, s
tp producing time, s
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 159
T temperature, K
Tc capillary temperature
Tpc pseudocritical temperature
Tr reduced temperature
v velocity, m s1
v apparent velocity
V cuttings volume, m3
Vb bulk volume, m3
VB total rock volume
Vp pore volume, m3
Wi volume of water, m3
Wp water produced, m3
x distance
z real gas factor
g Euler constant
y contact angle
Y static shear stress, dyne/cm2
m dynamic viscosity
mg viscosity of a real gas, Pa s
%o,w,g,f density of oil, water, gas, fluid, kg/m3
%m density of mud
s surface tension, interfacial tension
f porosity
fd dead porosity
feff effective porosity
F flow potential, Pa
lo,w,g mobility of oil, water, gas
Subscripts
i, 0 initial
r residual
b at bubble point
0 standard
j component j
f fluid
REFERENCES
GENERAL REFERENCES
SPECIFIC REFERENCES
56. V. B. Porfirev: ‘‘Inorganic Origin of Petroleum, Am. Assoc. Petrol. Geol. Bull. 58
(1974) 3 – 33.
57. B. Tissot, D. Welte: Petroleum and Occurrence, Springer Verlag, Heidelberg
1978.
58. D. H. Welte: ‘‘Organischer Kohlenstoff und die Entwicklung der Photosynthese auf
der Erde,’’ Naturwissenschaften 57 (1970) 17.
59. W. Schott: ‘‘Aufsuchen von Erdöl- und Erdgaslagerstätten’’ in A. Bentz, H. J. Martini
(eds.) Lehrbuch der Angewandten Geologie, vol. 2, part 1, Enke Verlag, Stuttgart
1968, pp. 563 – 671.
60. C. C. Cornelius: ‘‘Geothermal Aspects of Hydrocarbon Exploration in the North Sea
Area’’, NGU-Nor. Geol. Unters. 1975, no. 316, 29 – 67.
61. A. Hood, J. R. Castano: Organic Metamorphism: Its Relationship to Petroleum Gen-
eration and Application To Studies of Authigenic Minerals, United Nations ESCAP
CCOP, Techn. Bull. 8, 8 – 113.
62. G. Philippi: ‘‘The Influence of Marine and Terrestrial Source Material on The Com-
position of Petroleum,’’ Geochim. Cosmochim. Acta 38 (1974) 947 –966.
63. K. Krejci-Graf: ‘‘Moderne Anschauungen über die Entstehung des Erdöls,’’ Erdöl
Kohle Erdgas Petrochem. 13 (1960) 836 – 845.
64. U. Mann: ‘‘Revealing Hydrocarbon Migration Pathways,’’ Geol. Rundschau 78 (1990)
no. 1, 337 –348.
65. H. Klemme: Petroleum Basins – Classification and Characteristics,‘‘ J. Pet. Geol. 3
(1980) no. 187 – 207.
66. P. A. Ziegler: ‘‘Faulting and Graben Formation in Western and Central Europe,’’
in: P. Kent, M. H. P. Bott, D. P. McKenzie, C. A. Williams (eds.): ‘‘The Evolu-
tion of Sedimentary Basins,’’ Philos. Trans. R. Soc. Ser. London A 305 (1982)
113 –143.
67. W. Stahl, J. Koch: ‘‘13 C/12 C — Verhältnis norddeutscher Erdgase — Reifemerkmal
ihrer Muttersubstanzen,’’ Erdöl Kohle Erdgas Petrochem. 27 (1974) no. 10.
68. W. Stahl: ‘‘Carbon Isotope Ratios of German Natural Gases in Comparison with
Isotope Data of Gaseous Hydrocarbons from Other Parts of the World,’’ Adv. Org.
Geochem. Proc. Int. Meet. 6th (1973) 453 – 462.
69. D. H. Welte, M. A. Jükler: ‘‘Petroleum Origin and Accumulation in Basin
Evolution — A Quantitative Model,’’ Am. Ass. Petrol. Geol. Bull. 65 (1981)
1387 – 1390.
70. W. Rühl: Energiefaktor Erdöl, Edition Interform, Zürich 1989.
71. C. D. Masters, D. H. Root, E. D. Attanasi: ‘‘World Resources of Crude Oil and
Natural Gas,’’ Proceedings of World Petrol. Congress, Publ. J. Wiley and Sons,
Chichester 1991.
72. O. Serra: Fundamentals of Well Log Interpretation, vol. 2, Acquisition of logging
Data; Development in Petroleum Science, Elsevier, Amsterdam 1987.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 163
GENERAL REFERENCES
99. W. Arnold: Eroberung der Tiefe, VEB Dtsch. Vlg. f. Grundstoffind., Leipzig 1965.
100. D. Kett, Gas-Wasserfach 108 (1967) 221 – 229.
101. Oil Gas J. 73 (1975) Jan., 72.
102. Oil Gas J. 72 (1974) July, 33.
103. S. J. Pirson, Oil Reservoir Engineering, McGraw-Hill, New York 1958.
104. P. L. Randolph, World Oil 1974, Aug., 33 – 36; Sept., 65 – 71; Oct., 131 – 134; Nov. ,
81 – 84.
105. H. Rischmüller, Erdöl Erdgas Z. 19 (1966) 91 – 104.
106. H. Rischmüller, Erdöl Erdgas Z. 82 (1966) 90 – 99.
107. H. Rischmüller, H.-M. Tröndle: ‘‘Meerestechnik beim Ozeanbergbau,’’ in G. Dietrich
(ed.): Erforschung des Meeres, Umschau-Vlg., Frankfurt 1970, 267 – 285.
108. H. Rischmüller, Erdöl Erdgas Z. 88 (1972) 110 – 114.
109. K. Trenel, Erdöl Erdgas Z. 83 (1967) 188 – 191.
110. W. Tunn, Erdöl Erdgas Z. 87 (1971) July, 202 – 208.
111. E. Wichert, K. Aziz, Can. J. Chem. Eng. 49 (1971) 267 – 273.
112. R. A. Sumbatov, B. N. Khakhaev, V. N. Gramolin: ‘‘Drilling Rigs and Technical
Facilities for Superdeep Well Drilling,’’ Paper presented at the IGC, Washington
1989.
113. D. Noran: Oil Gas J. 73 (1975) Nov., 77 – 81.
114. Erdölinformationsdienst 28 (1975) no. 41.
115. Erdölinformationsdienst 29 (1975) no. 5.
116. W. Simmler, Gas-Wasserfach 109 (1968) 724 – 730.
117. Mannesmannröhren-Werke: Rohre für Öl- und Gasfelder, 6th ed., vols 1 – 3,
1962.
118. International Petroleum Encyclopedia, The Petroleum Publ. Co., Tulsa 1975.
SPECIFIC REFERENCES
119. R. D. Faniev: Abbau von Erdöl- und Erdgaslagerstätten, VEB Dtsch. Vlg. f. Grund-
stoffind., Leipzig 1963.
120. A. Mayer-Gürr: ‘‘Erdöl- und Erdgasgewinnung in der Bundesrepublik Deutschland,’’
Oel-Zeitschrift für die Mineralölwirtschaft 1 (1973).
121. K. K. Millheim: ‘‘The New Engineering Paradigm and the Emergence of Investiga-
tive Engineering,’’ J. Petrol. Technol. 1989, Aug.
122. Ullmann 4th ed., 11, pp. 19 – 40.
123. H. Rischmüller: ‘‘Tiefbohrtechnik als Schlüsselinstrument in der Energiewirtschaft,’’
Geowissenschaften 10 (1990) 317 – 323.
124. Ö. Aliquander: Das moderne Rotarybohren, VEB Dtsch. Vlg. f. Grundstoffind.,
Leipzig 1965.
125. W. Arnold: Eroberung der Tiefe, VEB Dtsch. Vlg. f. Grundstoffind., Leipzig 1983.
126. H. F. Spörker: ‘‘Entwicklungstendenzen der Tiefbohrtechnik,’’ Erdöl Erdgas Z. 88
(1972) 104 – 110.
127. K.-H. Grodde: Bohrspülungen und Zementschlämme in der Tiefbohrtechnik, Vlg. d.
Erdöl-Z. Otto Vieth, Hamburg 1963.
128. C. Marx, H. Rischmüller: ‘‘Drilling and Coring Techniques for Hard Rock,’’ in
H. J. Behr, F. G. Stehli, H. Vidal (eds.): Observation of the Continental Crust through
Drilling II, Springer, Berlin 1987, 149 – 160.
129. Maurer Engineering Inc.: Deep Drilling Cost Sensitivity Analysis, Study for Gas
Research Institute, Chicago 1989, GRI Contract No. 5088-260-1731.
Vol. 1 EXPLORATION, DRILLING, AND PRODUCTION ENGINEERING 165
KLAUS WEGGEN
Hemmingen, Federal Republic of Germany
GÜNTER PUSCH
Technische Universität Clausthal, Clausthal-Zellerfeld
Federal Republic of Germany
HEINRICH RISCHMÜLLER
Technische Universität Clausthal Clausthal-Zellerfeld
Federal Republic of Germany
Vol. 1 ENHANCED OIL RECOVERY 167
Excluding oil sands mining, enhanced oil recovery (EOR) accounted for 199,000 m3/
day (1.25 million b/day) of the world’s total 2000 oil production, 3 105m 3/day
(1.9 106 bbl/day) or 1.6% (1). In 2004, U.S. EOR production [105,740 m3/day
(663,451 bbl/day)] represented 12.2% of total production of crude oil plus gas
condensate (2). Most of the oil basins in the United States are mature.
These fields will require extensive enhanced oil programs to maintain current
production rates or, in some cases, merely to minimize rapidly increasing
depletion rates. For example, production in Prudhoe Bay, the largest U.S.
field, is declining 10% annually despite large investments in enhanced oil recov-
ery technology (3).
Primary and secondary oil recovery together recover only 25–50% of the oil
originally in place in a reservoir. After conventional oil recovery operations, the
U.S. will have 60 billion m3 (377 109) barrels of oil remaining after economic
production ceases. The U.S. Department of Energy has estimated as much as
43.3 billion barrels of oil could be produced using EOR technologies (4). The
most likely location of large new U.S. oil fields is Alaska and deep water offshore.
The costs of discovering and developing these reserves make unrecovered oil in
known fields an economically attractive target in part because much of the infra-
structure is also already in place.
Another factor promoting use of EOR technology in existing oilfields is
that countries with the greatest potential of containing undiscovered conven-
tional oil fields either limit Western investment in oil production (5).
The relatively low oil prices of 1981–2003 have resulted in significant
changes in the EOR technology being developed and field tested. Injection of
steam or oil-miscible gases remains of great interest. Micellar polymer flooding
can efficiently recover oil and was the focus of a large research and development
effort from 1970 to 1986, but this process remains relatively expensive and con-
sequently is seldom applied in field operations.
Using relatively low concentration surfactants as CO2 (6) and steam mobi-
lity control agents (foam) can improve recovery process cost effectiveness. Com-
binations of older EOR processes, eg, surfactant-enhanced alkaline
flooding and alkaline–surfactant–polymer flooding show promise of improved
cost effectiveness.
Enhanced oil recovery, the injection of fluids other than water to increase
oil recovery is but one means of recovering more oil from existing fields. Other
methods include better siting of injection and production wells due to improved
reservoir mapping technologies, infill drilling, horizontal wells, hydraulic frac-
turing, and acidizing.
With the increase of oil prices since 2003, major oil companies and national oil
companies have initiated quite large enhanced oil recovery projects usually using
carbon dioxide or steam as the injectants and are planning more such projects.
Using industrially generated CO2 for EOR while simultaneously sequestering
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0514080102151803.a01
168 ENHANCED OIL RECOVERY Vol. 1
the injected CO2 in the rock formation has the potential to produce more oil while
reducing emissions of this greenhouse gas (7).
Oil reservoirs are layers of porous sandstone or carbonate rock, usually sedimen-
tary. Impermeable rock layers, usually shales, and rock faults trap the oil in
the reservoir. The oil exists in microscopic pores in rock. Various gases and water
also occupy rock pores and are often in contact with the oil. These pores are inter-
connected with a complicated network of microscopic flow channels. The weight of
overlaying rock layers places these fluids under pressure. When a well penetrates
the rock formation, this pressure drives the fluids into the wellbore. The flow chan-
nel size, wettability of flow channel rock surfaces, oil viscosity, and other properties
of the crude oil determine the rate of this primary oil production.
As reservoir pressure is reduced by oil production, additional recovery
mechanisms may operate. One such mechanism is natural water drive. Water
from an adjacent more highly pressured formation is forced into the oil-bearing
formation by the pressure differential between the two formations. Another
mechanism is gas drive. Expansion of a gas cap above the oil as oil pressure
declines can drive additional oil to the wellbore. Produced gas may be reinjected
to maintain gas cap pressure as is done on the Alaskan North Slope. Additional
oil may also be produced by compaction of the reservoir rock as oil production
reduces reservoir pressure.
As the natural pressures in the reservoir decrease, oil production declines.
A pump is then used to maintain oil production at economic levels by drawing oil
to the surface and lowering the height of the fluid column in the wellbore. The
pressure of a column of fluid can decrease the rate of fluid entry into the wellbore.
Primary production typically recovers 10–25% of the oil originally in the
reservoir. Efficiency of primary production is related to oil properties, reservoir
properties, geometric placement of oil wells, and the drilling and completion
technology used to drill the wells and prepare them for production. Pumping
the well can maintain production at economic levels for years.
2.1. Waterflooding. Injection wells are used when the natural pres-
sures driving fluids to production wells are depleted and pumping alone is no
longer economical. Fluid injection repressurizes the reservoir, restoring a driving
force and promoting oil production. For economic reasons, water is the usual
injection fluid. Water injection or waterflooding is usually termed secondary oil
recovery. It accounts for 40% of total U.S. oil production. Additional oil recovery
by waterflooding is typically 15–25% of the oil originally in the reservoir.
Determining and using the optimum pattern and arrangement of production
and injection wells for a particular reservoir can have a significant effect on oil
recovery and production rates (8). Infill drilling and horizontal production wells
can drain oil reservoirs more efficiently (9). Horizontal injection wells have also
improved oil recovery by increasing volumetric sweep efficiency and increasing
fluid injection rates (10). One 600–1200-m long horizontal well can replace several
vertical wells decreasing both overall drilling and operational costs (11).
Vol. 1 ENHANCED OIL RECOVERY 169
There are two principal mechanisms of enhanced oil recovery: increasing volu-
metric sweep efficiency of the injected fluid and increasing oil displacement effi-
ciency by the injected fluid. In both, chemicals are used to modify the properties
of an injected fluid whether water, steam, a miscible gas, eg, CO2 or natural gas,
or an immiscible gas, usually nitrogen. Poor reservoir volumetric sweep effi-
ciency is the greatest obstacle to increasing oil recovery (12).
Wettability is defined as the tendency of one fluid to spread on or adhere to
a solid surface (rock) in the presence of other immiscible fluids (8). As many as
50% of all sandstone reservoirs and 80% of all carbonate reservoirs are oil-wet
(13). Strongly water-wet reservoirs are quite rare (14). Rock wettability can
affect its relative permeability to water and oil and thus influence fluid injection
rates, flow patterns of fluids within the reservoir, and oil displacement efficiency
(8,14,15). Alteration of rock wettability by adsorption of polar materials, eg, sur-
factants and corrosion inhibitors, or by the deposition of polar crude oil compo-
nents (16), can strongly alter the behavior of the rock (15).
When water is injected into a water-wet reservoir, oil is displaced ahead of
the injected fluid. Injection water preferentially invades the small- and medium-
sized flow channels or pores. As the water front passes, unrecovered oil is left in
the form of spherical, unconnected droplets in the center of pores or globules of
oil extending through interconnected rock pores. In both cases, the oil is comple-
tely surrounded by water and is immobile. There is little oil production after
injection water breakthrough at the production well (8).
In an oil-wet rock, water resides in the larger pores, oil exists in the smaller
pores or as a film on flow channel surfaces. Injected water preferentially flows
through the larger pores and only slowly invades the smaller flow channels
resulting in a higher produced water/oil ratio and a lower oil production rate
than in the water-wet case.
3.1. Injection Well Considerations. Fluid injection rate can have a
significant effect on oil recovery economics. Flow is radial from the wellbore
into the reservoir. Thus the region near the injection wellbore acts as a choke
for the entire reservoir.
Addition of surfactant to the injection water (17,18) can displace the oil
remaining near the well. The lower oil saturation results in an increase in the
water relative permeability (8). Therefore, a greater water injection rate may
be maintained at a given injection pressure. Whereas ultimate oil recovery
may not be increased, the higher water injection rate can increase oil production
rates improving oil recovery economics. The concentration of surfactant in the
injection water is relatively high (1–3%). However, the total amount of surfac-
tant used is not great because it is necessary only to displace the oil from a
2–3-m radius around the injection well.
Decreased injection rates resulting from formation damage, ie, reduction of
the rock fluid carrying capacity, near injection wells can reduce oil production
rates at offset (adjacent) production wells. Formation damage may result from
invasion of rock capillaries by solid particles in wellbore fluids during well dril-
ling and completion. Plugging of rock capillaries adjacent to fractures by fine
170 ENHANCED OIL RECOVERY Vol. 1
solid particles in fracturing fluids may also occur. Acidizing the rock immediately
adjacent to the wellbore can dissolve clays (qv), silica particles to improve the
rock permeability. However, treatments must be carefully designed to prevent
formation of precipitates that can plug rock flow channels.
Precipitate formation can occur upon contact of injection water ions and
counterions in formation fluids. Solids initially present in the injection fluid, bac-
terial corrosion products, and corrosion products from metal surfaces in the injec-
tion system can all reduce near-wellbore permeability. Injectivity may also be
reduced by bacterial slime that can grow on polymer deposits left in the wellbore
and adjacent rock. Strong oxidizing agents such as hydrogen peroxide, sodium
perborate, and occasionally sodium hypochlorite can be used to remove these bac-
terial deposits (19–21).
Formation damage can also be caused by chemical and physical interactions
of fluid and rock. Low salinity injection fluids are often preferred to obtain max-
imum viscosity from a given amount of water-soluble polymer. However, low sali-
nity fluids can cause swelling of water-expandable clays. This swelling reduces
the fluid-carrying capacity of rock flow channels. Because clays act as the
cementing medium in many sandstone formations, this swelling weakens cemen-
tation and can result in the release of mineral fine particles which can migrate to
constrictions and plug the flow channels.
Long-lasting stabilization of water-swelling clays may be achieved by using
materials, eg, hydroxyaluminum (22) and certain quaternary ammonium salt
polymers (23). Quaternary ammonium salt polymers have been used in drilling
fluids, completion fluids, acidizing treatments, and hydraulic fracturing as well
as in injection water and injected steam for enhanced oil recovery (23). Adsorp-
tion is long lasting and limited chemical treatment volumes are needed. Treat-
ment concentration is usually 0.1–1.0% by weight. For injection water and
injected steam, it is usually necessary to treat a 2–3-m radius from the injection
wellbore. The addition of potassium hydroxide to injection waters has also been
used to stabilize clays and maintain injection rates (24).
3.2. Injection Fluids. Whereas water is the most commonly used injec-
tion fluid, other fluids can provide higher oil recovery efficiency. Injecting gases
miscible with reservoir crude oil can result in low interfacial tension promoting
high oil displacement efficiency (25). The process of miscible gas flooding using
carbon dioxide (qv) is depicted in Fig. 1. The primary source of CO2 is high pres-
sure natural reservoirs of highly pure CO2. Any CO2 produced with the crude oil
is recovered, purified, and reused. However, CO2 produced in industrial opera-
tions, eg, separation from produced natural gas and coal gasification is increas-
ingly being used while use of flue gases from power plants, refinery units, and
other industrial operations is under consideration (26). For example, Norway’s
Sleipner Field project stores 1 million tonnes/year of CO2 associated with natural
gas production in a saline geologic aquifer. Norway’s carbon emission tax of 315
krone/tonne CO2 equivalent ($45/metric ton) encourages this storage (26). Other
suitable gases include natural gas and flue gas.
Overall, CO2 flooding can increase oil recovery by 7–15% of the oil originally
in place and can be sustained for 10–30 years (25). Improvement of CO2 flood per-
formance has resulted from advances in computer simulation of CO2 flood perfor-
mance and greater experience in handling carbon dioxide–water mixtures. Other
171
Fig. 1. Carbon dioxide flooding. The WAG process, in which a CO2 slug is followed by alternate water and CO2 injections, is usually
employed. The viscosity of the oil is reduced providing more efficient miscible displacement. (Courtesy of the U.S. Dept. of Energy).
172 ENHANCED OIL RECOVERY Vol. 1
recently reduced technology that can increase oil recovery includes horizontal
injection and production wells, four-dimensional (4D) seismic analyses to track
injected CO2 flow, automated field monitoring systems, and injecting larger
volumes of CO2 than were used in earlier EOR projects (27).
Many CO2 miscible EOR projects are located in the west Texas Permian
Basin where as much as two-thirds of the oil remains after waterflooding. An
incremental (10%) recovery is typical for Permian Basin CO2 floods, which
could correspond to as much as 0.5 109 m3 (3–4 109 bbl) (28). Pipelines (qv)
connect these large EOR projects to natural CO2 sources in Colorado and New
Mexico. Industrial point sources of CO2 have also been used for projects in
other areas.
Combining an in situ thermal enhanced oil recovery process (see below)
with CO2 injection and subterranean storage has been reported in the patent
literature (29).
The pressure–composition requirement for miscibility limits the oil reser-
voirs in which CO2 enhanced oil recovery has been applied. The low injected
fluid viscosity often results in poor volumetic sweep efficiency (see below).
Despite these concerns, a Department of Energy study indicates that carbon
dioxide injection has the potential of recovering 43 billion barrels of oil currently
‘‘stranded’’ in six mature U.S. oil-producing regions (30).
Supercritical CO2 (25,31) and various hydrocarbon injectants (25,32)
undergo physical interactions with crude oil that result in stripping out of the
low molecular weight components, which increases oil production (see SUPER
CRITICAL FLUIDS). The rapid or gradual development of miscibility with the remain-
ing crude oil constituents results in oil mobilization. Either partial or complete
miscibility with the oil may be developed depending on the nature of the injec-
tant, crude oil properties, and reservoir conditions, particularly temperature.
However, interaction of the injectant with the crude oil can alter rock wettability
and thus reduce injection rates and decrease oil recovery.
Another method of using CO2 is called cyclic CO2 stimulation or huff ‘n’puff
(33). A limited amount of CO2 is injected into a reservoir over hours or days. The
well is then shut in for a soak period of days to weeks to allow the CO2 to interact
with the crude oil, swelling the oil and reducing its viscosity. The well is then
opened, the CO2 provides a solution gas drive, and oil mobilized by the CO2
soak is produced.
3.3. Reducing Greenhouse Gas Emissions. Carbon dioxide EOR
could also play a role in reducing greenhouse gas emissions. Carbon dioxide cap-
ture and storage (CCS) underground could constitute as much as one-half of the
global emissions reduction by 2050, according to the International Energy
Agency (34). Carbon dioxide sequestration can provide two added economic
incentives for CO2 EOR in addition to the value of the additional oil recovered.
First, by reducing CO2 emissions, field operators can avoid being subject to pay-
ing CO2 emissions penalties. Second, they can receive credits for reducing their
CO2 emissions; credits that can be sold to other firms that have high levels of CO2
emissions. While these systems are being deployed in Norway and other
European countries, they are not yet in place in the U.S. Eight commercial CO2
sequestration projects are underway in the Norwegian sector of the North Sea,
Algeria, Alaska and Saskatchewan, Canada. Governments in Canada, Norway
Vol. 1 ENHANCED OIL RECOVERY 173
and other oil-producing areas are providing financial incentives for the under-
ground sequestration of CO2.
CCS involves three distinct processes: capturing CO2 from the gas streams
emitted during electric power production, industrial processes, or fuel proces-
sing; transporting the captured CO2; and storing it underground in rock forma-
tions where it cannot leak to the surface. Using this CO2 to increase oil
production would offset the costs of CCS and perhaps even render it profitable
(after factoring in governmental financial incentives).
3.4. Injection of Other Gases. Nonmiscible gases, eg, nitrogen, have
been used as EOR injection fluids. Increasingly, the nitrogen used is separated
at the oil field from produced natural gas rather than being produced in indus-
trial gases plants. Oil recovery mechanisms include volatilization of low molecu-
lar weight components of the crude oil and displacement of oil from the top of the
reservoir (27). The latter mechanism occurs as a result of gravity override of the
low density injectant.
Gas injection into a gas cap overlaying an oil reservoir is considered
an EOR method. The resulting repressurization of the reservoir promotes addi-
tional oil production. Reinjection of natural gas is responsible for a significant
fraction of Alaskan North Slope oil production. With the construction of a pipe-
line to transport North Slope natural gas to world markets, CO2 may be used
instead.
High temperature steam is also used for recovery of viscous crude oils
(28). Heat from the steam thins the oil, reducing viscosity and increasing mobi-
lity. The mobilized oil is produced at offset production wells. In heavy oil fields,
water flooding is often omitted and steam injection begun immediately after
primary production. Steam injection temperature is typically 175–2308C in
California oil fields. Injection temperature can reach 3008C in Canadian and
Venezuelan EOR projects.
The injection of large volumes of steam, steam flooding, also called steam-
assisted gravity drainage (SAGD), is used to mobilize oil, which is produced at
offset production wells. By reducing oil viscosity, the oil flows more readily
through the reservoir to production well bores. Smaller volumes of steam are
injected in the cyclic steam stimulation or huff ‘n’puff process (Fig. 2). Many
wells are placed on several cycles of steam stimulation and then used as injection
or production wells in steam flood projects. Large-scale steam projects can be
quite expensive. For example, a steamflood project in Oman’s Mukhaizna
heavy oil field is projected to cost > $2 billion and involve drilling 1800–2200
new wells (35). Production is expected to increase from the current 10,000 to
150,000 bbl/day.
An organic chemical may be added to the steam to provide additional oil
viscosity reduction in addition to the heat-thinning effect provided by injected
steam. For example, after the oil recovery rate in a steam flood reaches 25–75%
of the predicted maximum recovery rate, a hydrocarbon solvent may be injected
to recover additional heavy oil from an oil reservoir (36). Alternatively, steam or
hot water and solvent may be injected simultaneously in a ratio of > 5 : 1 (37). The
C1–5 alcohols have been used as additives to high temperature steam and injected
into heavy oil reservoirs in such a manner that the injectant does not form an
emulsion with the crude oil (38).
174
Fig. 2. Cyclic steam stimulation of an oil well: (a) steam, injected into a well over a period of days or weeks in a heavy oil reservoir, in-
troduces heat (huff) that, coupled with (b), alternate soak periods lasting a few days to allow (c) a production phase of weeks or months
(puff), thins the oil. This process may be repeated until production falls below a profitable level. (Courtesy of the U.S. Dept. of Energy).
Vol. 1 ENHANCED OIL RECOVERY 175
delay cross-linking (55) as can Cr(III) propionate (56). Encapsulation of the cross-
linking agent permits deeper penetration of gel-forming compositions into oil-
bearing formations (57). Injection of unhydrolyzed polyacrylamide followed by
in situ hydrolysis also delays cross-linking (58).
Cross-linked xanthan gums have also been used to reduce the permeability
of thief zones. Trivalent chromium is the preferred cross-linker (59). Cross-linker
effectiveness is less at high salinity. However, Cr(III) has been used in the field
at salinities as great as 166,000 ppm total dissolved solids (60).
Proper placement of the treatment fluid in the reservoir is critical to treat-
ment success. Careful sizing of the treatment and choice of injection rates are
required because overtreatment can cause plugging of the oil-containing rock
and excessive reduction of the injection rate. Even after a well-designed treat-
ment, the fluid injection rate is often significantly less than before well treat-
ment. Many successful applications of this technology in waterfloods and in
surfactant polymer floods have been reported. Wells in CO2 EOR projects have
also been treated using this technology.
Polymerization may also occur in situ. Reactive monomers, eg, acrylamide
in concentrations of 2–5 wt% and various additives including a free-radical poly-
merization initiator may be used (61,62). A difunctional monomer, eg, N,N0 -
methylenebis(acrylamide), can be added to the injected formulation to form a
cross-linked polymer in situ. Low viscosity aqueous monomer solutions can be
injected at higher rates and preferentially enter high permeability zones to a
greater extent than do non-Newtonian polymer solutions (63). If no difunctional
monomer is used, the viscous polymer mass may be slowly dissolved by injection
water increasing water viscosity and providing a second means of increasing oil
recovery.
Lignosulfonates may be cross-linked in situ using Cr(III) (64) or an acidic
gas, eg, CO2 (65). Cross-linked lignosulfonate can be an effective plugging
agent at high temperatures, such as in steam injection wells. Lignosulfonate con-
centration is usually 2–3 wt%. Blends of lignosulfonate and sodium silicate have
also been used (66). Other systems cross-linked in situ for water or stream
injection wells are phenol–formaldehyde (67), urea–formaldehyde (68), furfuryl
alcohol (69), formaldehyde resin plus sulfonated tannin extract (70), and formal-
dehyde resin and alkali kraft lignin (71).
Using foamed gels can reduce the overall amount of chemicals required to
form a given volume of gel within a subterranean formation. For example, a
foamed gel formed from a cross-linkable carboxylate-containing polymer, a reac-
tive transition metal cross-linking agent, polyvinyl alcohol, an aqueous solvent
and an added gas has been described (72).
Surfactant precipitation may be used for in-depth permeability reduction of
thief zones (73). Thief zones have a low oil saturation owing to the preferential
flow of injected fluids through high permeability rock. This process is based on
the sequential injection of a slowly propagating ionic surfactant followed by an
aqueous spacer containing no surfactant. Then a more rapidly propagating
ionic surfactant of the opposite charge type is injected. The oppositely charged
surfactants gradually mix in the high permeability portions of the reservoir
(thief zones) causing precipitation, which plugs flow channels reducing perme-
ability in the most flooded portions of the reservoir thereby diverting injectant
Vol. 1 ENHANCED OIL RECOVERY 177
to rock zones containing higher oil saturations. The economically limiting fac-
tors in the use of this process are cost and low propagation rate of the cationic
surfactant.
Rock surface chemistry can affect volumetric treatment effectiveness and
economics. Metal ion cross-linkers and cationic polymers can adsorb on mineral
surfaces, particularly clays, by ion-exchange (qv) processes. The subsequently
lower concentration of materials in solution decreases treatment effectiveness
by reducing the rate of polymer cross-linking and decreasing gel strength.
Both in situ cross-linking of partially hydrolyzed polyacrylamides (74) and
injection of quaternary ammonium salt polymers having long hydrophilic side
chains (75) have been used to reduce the permeability of water-producing
zones adjacent to production wells. This permeability reduction decreases the
produced water/oil ratio as does injection of polyacrylamide in high hardness
brine to reduce rock permeability (76). The polymers exist in saline solution as
tightly coiled chains and are readily adsorbed owing to relatively low solubility in
hard water. Subsequent injection of soft, low salinity water uncoils the adsorbed
polymer chains increasing water viscosity and reducing rock permeability. This
technology could also be used to reduce the permeability of thief zones adjacent
to injection wells. However, mechanical isolation of these zones may be necessary
for cost-effective treatments.
3.7. Polymer Flooding. Even in the absence of fractures and thief
zones, the volumetric sweep efficiency of injected fluids can be quite low. The
poor volumetric sweep efficiency exhibited in waterfloods is related to the mobi-
lity ratio, M, the mobility of the injected water in the highly flooded (low oil
saturation) rock, mw, divided by the mobility of the oil in oil-bearing portions
of the reservoir, mo (77,78). The mobility ratio is related to the rock permeability
to oil, kro, and injected water, krw, and to the viscosity of these fluids by the
following equation:
The terms hw and ho represent the viscosity of the aqueous and oil phases,
respectively.
The polymer flooding process is depicted in Fig. 3. The displacing or driving
fluid may be steam, supercritical carbon dioxide, hydrocarbon miscible gases,
nitrogen, or solutions of surfactants or polymers instead of water. The volumetric
sweep efficiency increases with lower mobility ratio values (79). A mobility ratio
of 1.0 or lower is considered optimum. The mobility of water is often high relative
to that of oil. Steam and oil-miscible gases, eg, supercritical carbon dioxide exhi-
bit even higher mobility ratios. Consequently, these more expensive injectants
can have low volumetric sweep efficiencies.
Mobility control agents reduce the mobility ratio. Virtually all polymer flood
field projects have used hydrolyzed polyacrylamides having molecular weights
of 1–5 106 or xanthan gum, a biopolymer (80). Variations in polymer molecular
weight and structure have been made to improve performance properties.
Relatively low (100 ppm for fresh water, 1000 ppm or more for saline systems)
polymer concentrations can significantly increase injected water viscosity.
178
Fig. 3. This polymer flooding method requires a preflush to condition the reservoir, the injection of a polymer solution for mobility control
to minimize channeling, and a driving fluid (water) to move the polymer solution and resulting oil bank to production wells. (Courtesy of the
U.S. Dept. of Energy).
Vol. 1 ENHANCED OIL RECOVERY 179
propoxylates (206) have been evaluated for this application. More recently,
anionic surfactants have been used (207–220). Alpha-olefin sulfonates (AOS)
have been found to possess good salt tolerance, chemical stability at elevated
temperatures, and good oil solubilization (210,221), whereas less salt tolerant
alkylaromatic sulfonates exhibit excellent chemical stability. The nature of
the alkyl group, the aryl group, and the aromatic ring isomer distribution can
be adjusted to improve surfactant performance under a given set of reservoir
conditions (222,223).
Surfactant blends can provide improve cost effectiveness compared to use of
a single surfactant. For example, the use of a blend of a synthetic polyisobutylene
with a sulfonate surfactant, alcohols or nonionic surfactants has been described
(224). Cosurfactant requirements can be minimized using a surfactant having a
short-branched hydrophobe or a branched-alkyl substituent on an aromatic
group (222,225) and a long ethoxy group chain (225). Blends of surfactants
optimized for seawater or reservoir brine salinity include linear alkylxylene
sulfonate–alcohol ether sulfate mixtures (226).
High (1–10%) concentrations of lignosulfonate have sufficient interfacial
activity to increase oil recovery from unconsolidated sands (227). Lignosulfonates
and petroleum sulfonates undergo a synergistic interaction resulting in ultralow
interfacial tension and substantially increased oil recovery (228).
The effect of temperature, pressure, and oil composition on oil recovery
efficiency have all been the subjects of intensive study (229). Surfactant propaga-
tion is a critical factor in determining the EOR process economics (230). Surfactant
retention owing to partitioning into residual crude oil can be significant compared
to adsorption and reduce surfactant propagation rate appreciably (231).
Various low cost sacrificial agents decrease surfactant adsorption on reser-
voir rock and increase the surfactant propagation rate. These agents include lig-
nosulfonates and chemically modified lignosulfonates (6,80,143). Alkaline
chemicals (198,199), particularly sodium silicate (232), which precipitate in
the presence of divalent metal ions, can increase the surfactant propagation
rate. Intermixing of polymer mobility control fluid with a previously injected sur-
factant slug can result in undesirable surfactant–polymer interactions affecting
interfacial behavior and reducing oil displacement efficiency (233).
3.10. Alkaline Flooding. Alkaline or caustic flooding involves injection
of high pH agents, eg, sodium hydroxide, sodium carbonate, or sodium silicate
solutions. At equivalent Na2O levels, the three alkaline agents gave equivalent
recovery of each of nine different crude oils in laboratory core floods (234). How-
ever, the use of buffered sodium carbonate rather than strong alkali can result in
reduced interaction with mineral surfaces. The lower reagent consumption can
reduce the amount of sodium carbonate required.
These chemicals generate surfactants in situ by reacting with organic acids
present in crude oil (232,235). Several oil recovery mechanisms may be operative.
Probably the most significant oil recovery mechanism for this process is lowering
of the capillary number (the ratio of viscous to capillary forces) through interfa-
cial tension reduction. Other possible mechanisms are altering rock wettability
(usually from oil-wet to water-wet), oil emulsification and entrapment resulting
in lower injectant mobility (in turn resulting in a greater injectant volumetric
sweep efficiency), oil emulsification and entrainment in the flowing aqueous
Vol. 1 ENHANCED OIL RECOVERY 185
phase, and possibly the solubilization of rigid films that may form at the oil–
water interface.
Caustic flooding chemicals are relatively inexpensive. However, project eco-
nomics are adversely affected by the large quantities that must be injected. The
high pH agents react with reservoir clays (236) and are precipitated by divalent
metal ions present in formation waters. Coinjection of a lignosulfonate (237) or a
polyacrylate (238) reduces precipitation. This precipitation has been used to
advantage to reduce adverse surfactant and polymer interactions with dissolved
divalent metal ions. Injecting a caustic preflush causes divalent metal ion
precipitation before beginning a micellar polymer flood (239). Ion-exchange pro-
cesses promoting solubilization of divalent metal ions limit the effectiveness of
preflushes injected before the caustic solution (240).
Including a surfactant in the caustic formulation (surfactant-enhanced
alkaline flooding) can increase optimal salinity of a saline alkaline formulation.
This can reduce interfacial tension and increase oil recovery (239,241,242). Both
nonionic and anionic surfactants have been evaluated in this application
(243,244). Encouraging field test results have been reported (245). Current
(2005) operating costs in a waterflood in Oklahoma’s Delaware-Childers Field
are $20–22 per barrel of recovered oil (245). An alkaline-surfactant-polymer
waterflood is being initiated that will increase the produced oil from 0.3 to 5%
of total produced fluids. At the estimated oil peak production of 4300 barrels/
day expected in 2007, the higher oil cut and reduced water pumping costs should
reduced operating costs to an estimated $10 per barrel of produced oil.
Surfactants evaluated in surfactant-enhanced alkaline flooding include
internal olefin sulfonates (244), linear alkylxylene sulfonates (246), petroleum
sulfonates (246), alcohol ethoxysulfates (242,247,248), and alcohol ethoxylates/
anionic surfactants (241). Water-thickening polymers, either xanthan or polya-
crylamide, can reduce injected fluid mobility in alkaline flooding (249) and
surfactant-enhanced alkaline flooding (244,248). Cross-linked polymers have
been used to increase volumetric sweep efficiency of surfactant–polymer–alkaline
agent formulations (250). The combined use of alkali, surfactant, and water-
thickening polymer has been termed the alkali–surfactant–polymer (ASP)
process.
4. Other Technologies
However, there is concern that the microbes can also enter lower permeability
zones containing higher oil saturations thereby reducing oil production. Control-
ling the rate and location of bacterial growth and chemical production can be dif-
ficult. Bacterial growth near wellbores has been a common problem causing
reduced injection rates and productivity. However, field test results have been pro-
mising (253). Variations of this technology continue to be patented (254).
The in situ combustion method of enhanced oil recovery through air injec-
tion (32,255,256) is a chemically complex process. There are three types of in situ
combustion: dry, reverse, and wet. In the first, air injection results in ignition of
crude oil and continued air injection moves the combustion front toward produc-
tion wells. Temperatures can reach 300–6508C. Ahead of the combustion front is
a 90–1808C steam zone, the temperature of which depends on pressure in the oil
reservoir. Zones of hot water, hydrocarbon gases, and finally oil propagate ahead
of the steam zone to the production well.
The oil zone is fairly cool, and in a viscous oil reservoir this can result in
little oil movement (liquid blocking). Reverse combustion, in which oil ignition
occurs near the production well, can avoid this problem. The combustion zone
moves countercurrent to the flow of air from the injection well. Oil flows through
heated rock and remains mobile. Reverse combustion requires more air and con-
sumes more oil than forward combustion.
In wet combustion, water is injected concurrently and alternately with air,
extending the steam zone and aiding heat transfer to the crude oil reducing oil
viscosity. This can decrease the injected air/produced oil ratio and improve pro-
ject economics.
Maintenance and propagation of the combustion front are problems. This
has led to a near-wellbore technology in which the same well is used for air injec-
tion and oil production. The combustion front needs to be propagated for a rela-
tively short distance (257).
5. Economic Aspects
World EOR production in 2000 averaged 199,000 m3/day (1.25 million b/day) the
world’s 2000 oil production (258). In 2004, U.S. EOR production was 105,740 m3/
day (663,451 bbl/day) (259). Due to the rapidly fluctuating price of oil it is difficult
to calculate the current value of this level of EOR production.
U.S. 2004 CO2 EOR production of crude oil was 205,877 b/day (260). The
price of crude oil must exceed US$25/bbl before CO2 injection becomes econom-
ical (261). A CO2 price sensitivity analysis indicates CO2 must be priced at less
that US$0.50 per thousand cubic feet before CO2 injection becomes more econom-
ical than conventional production. Incentives for reduction of CO2 emissions if
industrially generated CO2 is used would increase the CO2 price that could be
tolerated while achieving economic viability.
Injection of hydrocarbons accounted for 97,300 b/day of U.S. 2004 oil pro-
duction (260). The amount of crude oil produced by this tehnology has been
declining since 1992 due to the increasing price of natural gas. Notrogen injec-
tion accounted for 14,700 b/day of the 2004 U.S. oil production (260).
Vol. 1 ENHANCED OIL RECOVERY 187
All thermal recovery methods accounted for 345,514 b/day of U.S. 2004 oil
production with steam injection accounting for 98.6% of this amount (260).
Chemical injection: micellar polymer, polymer, alkaline chemicals, and surfac-
tants, accounted for only 60 b/day of U.S. 2004 oil production (260). The contri-
bution of other EOR technologies to U.S. oil production was negligible.
The value of the oil produced by EOR methods varies with crude oil prices.
Given recent oil price volatility, any financial estimates would have limited
value.
BIBLIOGRAPHY
‘‘Enhanced Oil Recovery’’ under ‘‘Petroleum’’ in ECT 3rd ed., Vol. 17, pp. 168–182, by H.
R. Froning, D. D. Fussell, and E. W. Heffern, Amoco Production Co.; in ECT 4th ed., Vol.
18, pp. 405–432, by J. K. Borchardt, Shell Development Co.; ‘‘Enhanced Oil Recovery’’ in
ECT (online), posting date: December 4, 2000, by J. K. Borchardt, Shell Development Co.
CITED PUBLICATIONS
20. F. Cusack, H. M. Lappin-Scott, and J. W. Costerton, Oil Gas J. 85, 87 (Nov. 9, 1987).
21. W. N. Hensel Jr., R. L. Sullivan, and R. H. Stallings, Pet. Eng. Int., 155 (May 1981).
22. M. G. Reed, J. Pet. Technol. 24, 860–864 (1972).
23. J. K. Borchardt, in J. K. Borchardt and T. F. Yen, eds., Oilfield Chemistry: Enhanced
Recovery and Production Stimulation, American Chemical Society, Washington,
D. C., 1989, p. 204.
24. R. D. Sydansk, J. Pet. Technol. 36, 1366 (1984).
25. F. Stalkup Jr., Miscible Displacement, Monograph Vol. 8, Society of Petroleum
Engineers, Dallas, Tex., 1983, p. 62.
26. B. M. Friedman, R. J. Wissbaum, S. G. Anderson, Oil Gas J., Aug. 23, 2004.
27. Anonymous, Oil Gas J., April 20, 2005.
28. D. F. Martin and J. J. Tabor, J. Pet. Technol. 44, 396 (1992).
29. U.S. Pat. 6,871,707 (March 29, 2005), J. M. Karanikas and co-workers, (Shell Oil
Company).
30. Anonymous, Oil Gas J., April 20, 2005.
31. L. Minssieux, in Ref. 23, p. 273.
32. R. T. Johansen, in Improved Oil Recovery, Interstate Oil Compact Commission,
Oklahoma City, Okla., 1983, pp. 91–112.
33. C. Bardon and co-workers, SPE Res. Eng. 9, 92 (1994).
34. D. Leblond, Oil Gas J., Jan. 10, 2005.
35. Anonymous, Oil Gas J., May 6, 2005.
36. U.S. Pat. 6,662,872 (Dec. 16, 2003), A. M. Gurtek and co-workers (ExxonMobil
Upstream Research Company).
37. U.S. Pat. 6,591,908 (July 15, 2003), T. N. Nasr (Alberta Science and Research
Authority).
38. U.S. Pat. 6,403,472 (Oct. 23, 2001), R. C. Richardson and co-workers (Texaco, Inc.).
39. Anonymous, Oil Gas J., Oct. 13, 2004.
40. U.S. Pat. 6,852,215 (Feb. 8, 2005), M. Y. Wen (ExxonMobil Upstream Research
Company).
41. N. C. Nahas, Oil Gas J., Oct. 4, 2004.
42. J. K. Borchardt, Coll. Surf. 63, 189 (1992) and references cited therein.
43. R. D. Sydansk and P. E. Moore, Oil Gas J., 405 (Jan. 20, 1992).
44. P. D. Moffitt, J. Pet. Technol. 45, 356 (1993).
45. J. K. Borchardt, in J. I. Kroschwitz, ed., Encyclopedia of Polymer Science and Engi-
neering, Vol. 10, John Wiley & Sons, Inc., New York, 1987, pp. 350–351.
46. A. Moradi-Araghi and P. H. Doe, SPE Reservoir Eng. 2, 189 (1987).
47. S. K. Nanda, R. Kumar, K. L. Sindhwani, and K. L. Goyl, ONGC Bull. 23, 175 (1986).
48. Ref. 47, p. 344.
49. M. Prats, Thermal Recovery, Monograph Vol. 7, Society of Petroleum Engineers,
Dallas, Tex., 1982.
50. R. K. Prud’homme, J. T. Uhl, J. P. Poinsatte, and F. Halverson, Soc. Pet. Eng. J. 23,
804 (1983).
51. U.S. Pat. 4,413,680 (Nov. 8, 1983), B. B. Sandiford, H. T. Dovan, and R. D. Hutchins
(to Union Oil Co., Calif.).
52. B. Sloat, Pet. Eng. Int., 20 (1977).
53. T. P. Lockhart and P. Albonico, SPE Prod. Engr. 9, 273 (1994).
54. D. J. Jordan, D. W. Green, R. E. Terry, and G. P. Willhite, Soc. Pet. Eng. J. 22, 463
(1982).
55. U.S. Pat. 3,687,200 (Aug. 29, 1972), W. G. Routson (to Dow Chemical Co.).
56. N. A. Mumallah, SPE Reservoir Eng. 3, 243 (1988).
57. U.S. Pat. 6,387,986 (May 14, 2002), A. Moradi-Araghi (no assignee).
58. U. S. Pat. 4,744,418 (May 17, 1988), R. D. Sydansk (to Marathon Oil Co.).
Vol. 1 ENHANCED OIL RECOVERY 189
96. Eur. Pat. 106,666 (Apr. 24, 1984), C. J. Philips (to Pfizer Corp.).
97. Can. Pat. 1,070,492 (Jan. 29, 1980), S. L. Wellington (to Shell Canada, Ltd.).
98. U.S. Pat. 4,454,620 (June 12, 1984), F. Dawans, D. Binet, N. Kohler, and D. V.
Quang (to Institut Francais du Petrole).
99. U.S. Pat. 4,481,316 (Nov. 6, 1984), S. Kanda and G. Kawamura (to Nitto Chemical
Industry Co.).
100. Jpn. Kokai Tokkyo Koho 61 275,337 (Dec. 5, 1986), S. Kanda and G. Kawamura
(to Nitto Chemical Industry Co.).
101. Jpn Kokai Tokkyo Koho 62 184,048 (Aug. 12, 1987), A. Niita, T. Arai, R. Funato, and
T. Sato (to Mitsui Cyanamid KK).
102. Jpn Kokai Tokkyo Koho 62 184,047 (Aug. 12, 1987), A. Niita, T. Arai, R. Funato, and
T. Sato (to Mitsui Cyanamid KK).
103. Jpn Kokai Tokkyo Koho 62 177,052 (Aug. 3, 1987), A. Niita, T. Arai, R. Funato, and
T. Sato (to Mitsui Cyanamid KK).
104. Jpn Kokai Tokkyo Koho 62 277,407 (Dec. 2, 1987), Y. Nakanishi (Mitsubishi Chemical
Industries KK).
105. Eur. Pat. Appl. 196,199 (Oct. 1, 1986), C. P. Judson (to Pfizer, Inc.).
106. Jpn Kokai Tokkyo Koho 61 136,545 (June 24, 1986), A. Niita and Y. Ito (to Mitsui
Toatsu Chemicals, Inc.).
107. Eur. Pat. Appl. 241,340 (Oct. 14, 1987), F. Contat and J. Boutin (to Rhône Poulenc
Chemicals).
108. Eur. Pat. 103,483 (Mar. 21, 1984), G. M. Holzwarth (to Exxon Research and
Engineering Co.).
109. J. C. Phillips, J. W. Miller, W. C. Wernau, B. E. Tate, and M. H. Auerbach, Soc. Pet.
Eng. J. 25, 594 (1985).
110. U.S. Pat. 6,316,614 (Nov. 13, 2001), D. H. Doherty (CP Kelco U.S., Inc.).
111. U.S. Pat. 4,457,372 (July 3, 1984), M. S. Doster, A. J. Nute, and C. A. Christopher
(to Texaco, Inc.).
112. Eur. Pat. Appl. 211,288 (Feb. 25, 1987), R. W. Vanderslice and P. Shanon (to Getty
Science and Development Co.).
113. A. Audibert, C. Noik, and J. Lecourtier, J. Can. Pet. Technol. 32(7), 53 (1993).
114. U.S. Pat. 4,439,332 (Mar. 27, 1984), S. Frank, A. T. Coscia, and J. M. Schmitt
(to American Cyanamid Co.).
115. U.S. Pat. 4,439,342 (May 8, 1984), J. K. Borchardt (to Halliburton Co.).
116. J. Klein and R. Heitzman, Makromol. Chem. 179, 1895 (1978).
117. F. H. Halverson and J. E. Lancaster, Macromolecules 18, 1139 (1985).
118. Fr. Pat. 2,495,217 (June 4, 1982), J. Boutin and F. Contat (to Rhône Poulenc
Industries).
119. G. Muller, J. P. Laine, and J. C. Fenyo, J. Polym. Sci. Polym. Chem. Ed. 17, 659
(1979).
120. J. S. Ward and F. D. Martin, Soc. Pet. Eng. J. 21, 623 (1981).
121. M. Szabo, J. Pet. Technol. 31, 553 (1979).
122. M. Szabo, J. Pet. Technol. 31, 561 (1979).
123. J. G. Southwick and C. W. Manke, SPE Reservoir Eng. 3, 1193 (1988).
124. C. L. McCormick, J. Macromol. Sci. Chem. Part A. 22, 955 (1985).
125. U.S. Pat. 4,644,020 (Feb. 17, 1987), G. A. Stahl (to Phillips Petroleum Co.).
126. P. H. Doe, A. Moradi-Araghi, J. E. Shaw, and G. A. Stahl, SPE Res. Eng. 2, 461
(1987).
127. C. L. McCormick and K. P. Blackmon, J. Polym. Sci. A, Polym. Chem. 24, 2635
(1986).
128. Ger. Pat. 3,627,456 (Feb. 26, 1987), T. Y. Ching (to Chevron Research Co.).
129. Eur. Pat. Appl. 233,533 (Aug. 26, 1987), R. G. Ryles (to American Cyanamid Co.).
Vol. 1 ENHANCED OIL RECOVERY 191
130. Ger. Pat. 3,220,503 (Dec. 1, 1983), P. Engelhardt, U. Greiner, H. Schmitz, W. Gulden,
and S. P. von Halasz (to Casella AG).
131. Jpn Kokai Tokkyo Koho 62 015,279 (Jan. 23, 1987), Y. Ito, A. Niita, Y. Sudo, and K.
Hayashi (to Mitsui Toatsu Chemicals, Inc.).
132. U.S. Pat. 4,540,498 (Sept. 10, 1985), M. M. Wu and L. E. Ball (Standard Oil Co., Ohio).
133. R. W. Dexter and R. G. Ryles, in Ref. 20, p. 102.
134. H. Pledger Jr., J. J. Meister, T. E. Hogen-Esch, and G. B. Butler, Polym. Prepr. Am.
Chem. Soc. Div. Polym. Chem. 22 (2), 72 (1981).
135. H. H. Neidlinger and C. L. McCormick, Polym. Prepr. Am. Chem. Soc. Div. Polym.
Chem. 20(1), 901 (1979).
136. J. J. Meister, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem. 25(1), 266 (1984).
137. U.S. Pat. 6,030,928 (Feb. 29, 2004), G. A. Stahl (Phillips Petroleum Company).
138. F. Friedmann, SPE Reservoir Eng. 1, 261 (1986).
139. Y. Cohen and F. R. Christ, SPE Reservoir Eng. 1, 113 (1986).
140. L. T. Lee, J. Lecourtier, and G. Chaveteau, in Ref. 20, p. 224.
141. S. A. Hong and J. H. Bae, SPE Res. Eng. 5, 467 (1990).
142. U.S. Pat. 4,627,494 (Dec. 9, 1986), G. Kalfoglou (to Texaco, Inc.).
143. U.S. Pat. 4,713,185 (Dec. 15, 1987), J. Howard and M. Stirling (to Canadian Patents
and Development Ltd.).
144. U.S. Pat. 4,284,517 (Aug. 18, 1981), C. S. H. Chen and E. W. Sheppard (to Mobil Oil
Corp.).
145. Brit. Pat. 2,148,356 (Mar. 4, 1987), J. H. Clint, P. K. G. Hodgson, and E. J. Tinley (to
British Petroleum PLC).
146. U.S. Pat. 4,574,885 (Mar. 11, 1986), R. L. Horton (to Phillips Petroleum Co.).
147. J. P. Heller, D. K. Dandge, R. J. Card, and L. G. Donaruma, Soc. Pet. Eng. J. 25, 679
(1985).
148. T. G. Arf and co-workers, SPE Reservoir Eng. 2, 166 (1987).
149. S. L. Wellington and H. J. Vinegar, J. Pet. Technol. 39, 885 (1987).
150. G. J. Parlani, K. A. McColloch, S. L. Warden, and D. R. Edens, J. Pet. Technol. 44,
984 (1992).
151. U.S. Pat. 6,105,672 (August 22, 2000), C. Deruyter (Institut Francais du Petrole).
152. F. Stalkup Jr., J. Pet. Technol. 36, 815–826 (1984).
153. J. P. Heller, in L. L. Schramm, ed., Foams: Fundamentals and Applications in the
Oil Industry, American Chemical Society, Washington, D.C., 1994, p. 201.
154. K. Mannhardt and J. J. Novosad, in Ref. 158, p. 259.
155. J. E. Hanssen and T. Holt, Rogoland Research Report, RF-80/93, Stavanger,
Norway, 1993, and references cited therein.
156. U.S. Pat. 3,185,634 (May 25, 1965), F. F. Craig Jr., and J. L. Lummus (to Pan
American Petroleum Corp.).
157. M. I. Kuhlman and co-workers, SPE Res. Eng. 7, 445 (1992).
158. J. K. Borchardt , in Ref. 20, pp. 181–204.
159. U.S. Pat. 4,706,752 (Nov. 17, 1987), L. W. Holm (to Union Oil Co. of California).
160. U.S. Pat. 4,706,750 (Nov. 17, 1987), J. J. Buckles (to Mobil Oil Corp.).
161. U.S. Pat. 4,502,538 (Mar. 5, 1985), S. L. Wellington, J. Reisberg, E. F. Lutz, and
D. B. Bright (to Shell Oil Co.).
162. U.S. Pat. 4,799,547 (Jan. 24, 1989), J. K. Borchardt (to Shell Oil Co.).
163. U.S. Pat. 4,380,266 (Apr. 19, 1983), S. L. Wellington (to Shell Oil Co.).
164. U.S. Pat. 4,739,831 (Apr. 26, 1988), L. A. Settlemeyer and M. J. McCoy (to Dow
Chemical Co.).
165. U.S. Pat. 4,676,316 (June 30, 1987), T. O. Mitchell (to Mobil Oil Corp.).
166. U.S. Pat. 6,439,308 (August 27, 2002), D. Wang (Da Qing Petroleum Administration
Bureau).
192 ENHANCED OIL RECOVERY Vol. 1
167. Can. Pat. 1,221,305 (May 5, 1987), N. F. Djabbarah (to Cities Service Oil and Gas
Corp.).
168. U.S. Pat. 4,733,727 (Mar. 29, 1988), A. H. Falls (to Shell Oil Co.).
169. L. W. Holm, J. Pet. Technol. 32, 1499 (1970).
170. L. W. Holm and W. H. Garrison, SPE Reservoir Eng. 3, 112 (1988).
171. L. D. Talley, SPE Reservoir Eng. 3, 235 (1988).
172. Brit. Pat. 2,239,278 (June 21, 1991), L. L. Schramm, C. Ayasse, K. Mannhardt, and
J. Novosad (to Alberta Oil Sands Technology and Research Authority).
173. G. P. Small, SPE Prod. Eng. 1, 388 (1986).
174. K. C. Hong and C. E. Stevens, SPE Reservoir Eng. 7, 407 (1992).
175. T. J. C. Fontaine, L. Hayes, and G. Reese, J. Can. Pet. Technol. 32, 44 (Nov. 1993).
176. P. J. Jesperson and T. J. C. Fontaine, J. Can. Pet. Technol. 32, 52 (May 1993).
177. D. K. Olsen and P. S. Sarathi, Proceedings of the Field Application of Foams for Oil
Production Symposium, U.S. Dept. of Energy, N.I.P.E.R., Bartlesville, Okla., 1993.
178. U.S. Pat. 5,333,687 (Aug. 2, 1994), W. T. Osterloh (to Texaco, Inc.).
179. E. E. Isaacs, F. C. McCarthy, and J. D. Maunder, SPE Reservoir Eng. 3, 565 (1988).
180. U.S. Pat. 4,693,311 (Sept. 15, 1987), H. M. Muijs and P. P. M. Keijzer (to Shell Oil
Co.).
181. Ger. Pat. 3,510,765 (Sept. 26, 1985), R. Janssen van Rosmalen, H. M. Muijs, and
P. P. M. Keijzer (to Shell Internationale Research Maatschappij BV).
182. U.S. Pat. 5,273,682 (Dec. 28, 1993), M. Danzik (to Chevron Research and Technology
Co.).
183. Ger. Pat. 3,734,075 (Apr. 4, 1988), H. M. Muijs and P. P. M. Keijzer (to Shell Inter-
nationale Research Maatschappij BV).
184. Ger. Pat. 3,503,532 (Aug. 8, 1985), W. S. Huang, Z. Z. Gassmann, J. T. Hawkins,
V. H. Schievelbein, and W. L. Hall (to Texaco Development Co.).
185. U.S. Pat. 4,699,214 (Oct. 13, 1987), H. P. Angstadt (to Sun Refining and Marketing
Co.).
186. PCT Int. Pat. Appl. 85 05,146 (1985), T. Lim (to Dow Chemical Co.).
187. U.S. Pat. 4,576,232 (Mar. 18, 1986), J. H. Duerksen, R. G. Wall, and J. D. Knight (to
Chevron Research Co.).
188. A. H. Falls, J. B. Lawson, and G. J. Hirasaki, J. Pet. Technol. 40, 95 (1988).
189. J. Novosad, B. B. Maini, and A. Huang, J. Can. Pet. Technol. 25, 42 (1986).
190. H. C. Lau and S. M. O’Brien, SPE Reservoir Eng. 2, 1177 (1988).
191. U.S. Pat. 4,597,442 (July 1, 1986), R. E. Dilgren, H. C. Lau, and G. J. Hirasaki
(to Shell Oil Co.).
192. U.S. Pat. 4,617,995 (Oct. 21, 1986), H. C. Lau (to Shell Oil Co.).
193. Fr. Pat. 2,557,198 (Juen 28, 1985), H. P. Angstadt, D. F. Rugen, and J. L. Cayias
(to Sun Marketing and Refining Co.).
194. E. E. Isaacs, J. D. Maunder, and L. Juan, in Ref. 20, p. 325.
195. B. B. Maini and V. Ma, J. Can. Pet. Technol. 25, 65–69 (1986).
196. H. P. Angstadt and H. Tsao, SPE Reservoir Eng. 2, 613 (1987).
197. U.S. Pat. 4,702,317 (Oct. 27, 1987), C. W. Shen (to Texaco, Inc.).
198. U.S. Pat. 4,727,938 (Mar. 1, 1988), H. C. Lau (to Shell Oil Co.).
199. H. C. Lau and J. K. Borchardt, SPE Reservoir Eng. 6, 470 (1991).
200. U.S. Pat. 5,105,884 (Apr. 21, 1992), R. D. Sydansk (to Marathon Oil Co.).
201. D. O. Shah and R. S. Schecter, eds. Improved Oil Recovery by Surfactant and
Polymer Flooding, Academic Press, Inc., New York, 1977.
202. H. K. van Poolen, Fundamentals of Enhanced Oil Recovery, PennWell Publishing
Co., Tulsa, Okla., 1980, p. 114.
203. J. C. Morgan, R. S. Schechter, and W. H. Wade, in K. L. Mittal, ed., Proc. Sect. 52nd
Colloid Surf. Sci. Symp., Vol. 2, Plenum Press, New York, 1979, pp. 749–775.
Vol. 1 ENHANCED OIL RECOVERY 193
240. U.S. Pat. 4,466,892 (Aug. 21, 1984), K. S. Chan and S. J. Majoros (to Standard Oil
Co., Ohio).
241. A. L. Bunge and C. J. Radke, Soc. Pet. Eng. J. 23, 657 (1983).
242. P. J. Shuler, D. L. Kuehne, and R. M. Lerner, J. Pet. Technol. 41, 80 (1989) and re-
ferences cited therein.
243. R. C. Nelson, J. B. Lawson, D. R. Thigpen, and G. L. Stegemeier, Proc. 1984 SPE
Enhanced Oil Recovery Symposium, Society of Petroleum Engineers, Dallas, Tex.,
1984, pp. 417–424.
244. R. C. Nelson and co-workers, SPE Res. Eng. 9, 217 (1994).
245. G. Moritis, Oil Gas J., March 28, 2005.
246. S. M. Saleem and M. J. Faber, Rev. Tec. INTEVEP 6, 133–142 (1986).
247. S. M. Saleem and A. Hernandez, J. Surf. Sci. Technol. 3, 1 (1987).
248. M. J. Pitts, Prepr. Am. Chem. Soc., Div. Pet. Chem. 33 (1), 169 (1988).
249. U.S. Pat. 4,502,541 (Mar. 5, 1985), J. B. Lawson and D. R. Thigpen (to Shell Oil Co.).
250. G. H. Neale, K. C. Khulbe, and V. Hornof, Can. J. Chem. Eng. 65, 700 (1987).
251. G. E. Jenneman, R. M. Knapp, M. J. McInerney, D. E. Menzie, and D. E. Revus, Soc.
Pet. Eng. J. 24, 31 (1984).
252. D. O. Hitzman, Microbial Enhancement of Oil Recovery—Recent Advances, Elsevier
Science Publishing Co., Inc., New York, 1991, pp. 11–20.
253. I. Lazar, Dev. Petr. Sci. 31, 485 (1991).
254. U.S. Pat. 5,758,270 (July 6, 2004), E. Sunde and co-workers (Statoil ASA).
255. P. B. Crawford and C. Cju, Improved Oil Recovery, Interstate Oil Compact Commis-
sion, Oklahoma City, Okla., 1983, pp. 251–309.
256. Ref. 214, pp. 41–55.
257. B. Nzekwu, R. J. Hallam, and G. J. J. Williams, SPE Reservoir Eng. 5, 163 (1990);
Ref. 224, pp. 41–55.
258. B. Williams, Oil Gas J., pp. 18–25 (Aug. 4, 2003).
259. U.S. Energy Information Agency, http://www.eia.doe.gov/emeu/international/
petroleu.html#IntlProduction.
260. Anonymous, Oil Gas J., pp. 45–52 (Apr. 12, 2004).
261. A. Y. Zekri and K. K. Jerbi, Oil & Gas Science and Technology–Rev. IFP 57,
pp. 259–267 (2002).
JOHN K. BORCHARDT
Shell Development Company
OIL SHALE
1. Introduction
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.15091216091605.a01
Vol. 1 OIL SHALE 195
Oil shale deposits were formed in ancient lakes and seas by the slow deposi-
tion of organic and inorganic remains. The geology and composition of the inor-
ganic minerals and organic kerogen components of oil shale vary with deposit
locations throughout the world (1) (see also PETROLEUM).
2. Reserves
Estimates of oil shale deposits by continent are given in Table 1 (2). Characteris-
tics of many of the world’s best known oil shales are summarized in Table 2 (3,4).
Oil shale deposits in the United States occur over a wide area (Table 3). The
most extensive deposits, covering ca 647,000 km2 (250,000 mi2), are the Devo-
nian-Mississippian shales of the eastern United States (5). The richest U.S. oil
shales are in the Green River formation of Colorado, Utah, and Wyoming. Typical
mineral and organic analyses for Green River oil shale are given in Table 4.
The Green River formation includes an area of ca 42,720 km2 (16,500 mi2),
and in-place reserves are ca ð0:5 1:1Þ 1012 m3 ðð3 7Þ 1012 bblÞ of which ca
80% are federally owned. The richest portion (85% of the reserve) of the Green
River formation is in the Piceance Basin of Colorado. The deposits in Utah and
Wyoming contain 10% and 5% of the reserves, respectively (6).
The Parachute Creek member contains the majority of the oil shale in
the Piceance Creek Basin and is ca 580 m thick at the depositional center
of the basin. The members of the Green River formation and the thickness of
the various zones are indicated in Figure 1. Organic and saline mineral con-
tents increase toward the depositional center of the basin. The rich Mahogany
zone extends across the Piceance Basin and into the Uinta Basin in eastern
Utah. In addition to its high contents of organic matter, the Parachute
Creek member contains large reserves of nahcolite[15752-47-3],NaHCO3,
and dawsonite [12011-76-6],NaAl(OH)2CO3, which are present in the deepest
parts of the basin.
Other properties
moisture, wt% 6.7–9.8 0.2–6 23.2 2.8 11.3 0.7 7.7 2.8 5.4–6.7 8.1 5.8
specific gravity 1.88–1.99 1.9–2.1 1.47 2.22 1.73 1.94 2.05 1.16 2.32–1.97 1.57 1.60
gross heating value, 5.230–6.904 5.439–6.987 11.950 5.791 8.577 9.113 4.728 30.669 3.766–6.908 4.209 16.07
J/gb
total carbon, wt% 14.78–19.46 12–17 25.67 12.82 18.74 23.45 10.50 70.54 10.75–16.58 15.5 36.8
total hydrogen, % 1.9–2.0 0.9–2.4 3.7 1.5 2.9 2.9 1.7 8.39 1.4–2.2 1.6 4.3
total sulfur, % 2.1–2.7 3.9–5.6 1.0 4.4 1.6 1.1 0.9 1.4 0.9–1.0 2.9 2.0
nitrogen, % 0.46–0.63 0.3–1.9 0.3 1.3 0.6 0.3 1.0 0.1
196
loss on ignition, at 31.4–38.9 20–24 41.9 21.3 32.1 38.0 18.6 91.7 21.1–28.5 56.2
9508C, %
Ash composition, wt%
SiO2 31.6–37.5 5.0–5.6 64.8 57.2 45.2 73.2 53.4 54.8–55.7 33.2
Fe2O3 3.5–5.8 7.6–9.8 10.7 12.2 5.5 8.1 9.9 6.8–5.7 6.6
Al2O3 8.6–13.0 9.8–12.6 12.5 19.5 2.3 12.1 24.3 17.9–15.0 8.9
CaO 15.7–26.7 1.3–3.9 1.9 1.1 18.9 2.0 3.4 8.9–13.8 33.7
MgO 5.6–7.4 2.0–3.7 0.6 0.8 17.4 1.0 4.1 6.1–3.7 9.5
Fischer assay oil
specific gravity, 208C 0.962 0.906 0.918 0.926 0.890 0.902 0.895 0.905 0.880 0.980 0.958
total carbon, wt% 78.73 84.60 83.40 84.95 84.81 84.21 84.72 84.32 85.6 80.8 83.4
hydrogen, % 9.69 12.50 11.37 11.85 11.65 11.29 12.54 11.89 12.3 10.4 10.7
sulfur, % 6.33 1.10 1.16 1.40 0.52 0.92 0.46 1.72 0.6 5.0 0.7
nitrogen, % 1.52 0.90 1.18 1.12 2.60 1.78 1.30 0.69 1.1 1.2 0.1
gross heating value, 40.074 42.547 43.070 41.773 42.447 42.723 42.677 42.539 43.932 39.748 39.790
J/gb
a
Ref. 3.
b
To convert J to cal, divide by 4.184.
Vol. 1 OIL SHALE 197
Sample preparation for the modified Fischer assay technique, a standard method
to determine the liquid yields from pyrolysis of oil shale, is necessary to achieve
reproducible results. A 100-g sample of >230 mm (65 mesh) of oil shale is
heated in a Fischer assay retort through a prescribed temperature range, eg,
ca 25.5–5008C, for 50 min and then soaked for 20 min. The organic liquid
which is collected is the Fischer assay yield (7). The Fischer assay is not an abso-
lute method, but a qualitative assessment of the oil that may be produced from a
given sample of oil shale (8). Retorting yields of greater than 100% of Fischer
assay are possible.
Upper aquifer
Uinta Thickness,
formation 0 – 500 m
Lean upper
oil shale
zone
Mahogany zone
aquifer
Lower
Leach zone
Saline zone
(nahcolite,
dawsonite,
etc)
member
Garden
Gulch
Thickness,
90 – 300 m
Douglas
member
Creek
Thickness,
0 – 140 m
Wasatch
formation Ancient
lake bed
A total material balance assay is a Fischer assay in which the retort gases
are collected. A complete material balance closure and yields in excess of those
expected from Fischer assay results are achieved. More complete descriptions of
both the Fischer assay and the Tosco material balance assay methods have been
reported (9).
4. General Properties
sequential A ! B ! C ð1Þ
Vol. 1 OIL SHALE 199
parallel A C
and
A B
alternate
C
4.2. Temperature and Product Yields. Most oil shale retorting pro-
cesses are carried out at ca 4808C to maximize liquid product yield. The effect
of increasing retort temperature on product type from 480 to 8708C has been stu-
died using an entrained bed retort (17). The oil yield decreased and the retort gas
increased with increased retorting temperature; the oil became more aromatic as
temperature increased, and maximum yields of olefinic gases occurred at about
7608C. Effects of retorting temperatures on a distillate fraction (to 3008C) are
given in Table 6.
4.3. Carbonate Decomposition. The carbonate content of Green River
oil shale is high (see Table 4). In addition, the northern portion of the Piceance
Creek basin contains significant quantities of the carbonate minerals nahcolite
and dawsonite. The decomposition of these minerals is endothermic and occurs
at ca 600–7508C for dolomite, 600–9008C for calcite, 350–4008C for dawsonite,
and 100–1208C for nahcolite. Kinetics of these reactions have been studied (19).
Carbon dioxide, a product of decomposition, dilutes the off-gases produced from
retorting processes at the above decomposition temperatures.
200 OIL SHALE Vol. 1
5. Retorting
Oil shales are solid minerals, impervious to the flow of fluids, and are generally
situated in deposits below the earth’s surface. Therefore, several process steps
must be undertaken to produce crude shale oil. In the case of the commonly
used above-ground retorting (AGR), these steps involve mining, crushing, and
heating. The grade (volume of oil per weight of rock) of most oil shales is low
(see Table 2), and large amounts of the oil shale rock must be processed to pro-
duce crude shale oil. Depending on the grade, 2 to 25 metric tons of oil shale must
be processed to produce one cubic meter of crude shale oil (0.4–4.6 short tons per
barrel of crude shale oil). In order to eliminate the costs of mining and material
handling, direct underground retorting (in situ retorting) has been considered as
an alternative to the conventional AGR.
Historically, direct combustion has been employed in which some of the
organic matter of the kerogen is combusted to provide the heat necessary for
retorting. Although these direct heat (DH) processes do not require a supplemen-
tal source of fuel, some of the kerogen is consumed and the gaseous products of
the kerogen decomposition are diluted with the products of combustion. In order
to obviate these shortcomings, indirect heat (IH) processes were developed in
which the heat required for retorting was supplied by hot gases or solids that
were heated externally. However, the IH processes do not utilize any of the
solid residual carbon or char resulting from kerogen decomposition and they
do require an external source of fuel.
There are numerous means of classifying the many processes that have
been employed to retort oil shale. In addition to the types of retorting, the
retorting process can be classified by the type of feed used and by the flows
within the retort. Types of oil shale feed can be classified as coarse, >5 mm
(>0:25 in:) or fine, <65 mm (<0:25 in:). The flows within the retort can be classi-
fied as concurrent, ie, all materials flowing in the same direction, or countercur-
rent, ie, the solid oil shale flowing in one direction, the air and gases in the
opposite direction. The coarse feed systems usually result in the disposal of
the raw shale fines as waste. The fine feed systems obviate the latter problem
but result in increased crushing costs and greater environmental impacts from
Vol. 1 OIL SHALE 201
Combustion air
Combustion air
preheat Spent shale zone:
spent shale residual carbon burned
cooling
Oxygen depleted
Pyrolysis zone:
kerogen conversion to
oil, gas, coke
retorted shale supplies much of the fuel for this process. The GCR is the forerun-
ner of most continuous AGR processes (Table 7).
PETROSIX. The PETROSIX technology is operated in the IH mode using
hot recycle gas as the heat-transport medium. The PETROSIX retort has only
one level of heat input, uses countercurrent flows, and uses a circular grate to
control the flow of solids (Fig. 3). The PETROSIX has been operated by Petrobras
(Brazil) since the 1950s and is one of the few retorting processes producing shale
oil in 1995.
Paraho. The Paraho retorting technology is similar to the PETROSIX
technology except that it can be operated in the direct heat (DH) mode. The
unique feature of the Paraho technology is the two levels of heat input (Fig. 4).
In the IH mode, the air blower shown in Figure 4 is replaced by a recycle gas
heater. The Paraho DH operation has been carried out near Rifle, Colorado
since the 1970s; operations to produce asphalt from shale oil are continuing.
Tosco II. The Tosco II retorting technology, developed by The Oil Shale
Corporation, represents IH technology employing concurrent flow using fine-
sized feedstock. It was tested in Colorado from the late 1950s until the early
1980s. The unique feature of the Tosco II process is the use of ceramic balls as
the heat-transport medium (Fig. 5). These 125-mm (0.5-in.) balls, larger than the
Vol. 1 OIL SHALE 203
Shale
distributor Electrostatic
precipitator Condenser
Cyclone
Separator Light
Heavy
shale
Pyrolysis shale Heavy Compressor oil
vessel Hot oil shale oil
gas
Discharge Waste
mechanism water
Cool gas
Retorted shale
facilities
Seal systems
finely crushed shale feed, are separated from the retorted shale, recycled,
reheated, and reused in the process.
UNISHALE B. The UNISHALE process, like the Paraho process, uses
lump feed and countercurrent flows, and can be operated in either the DH or
IH mode. The UNISHALE B process is an IH process that uses hot recycled
gas as the heat-transport medium (Fig. 6). The unique feature of the UNISHALE
processes is the rock pump. The solids move upward through the retort as
the vapors are moving downward. The rock pump was used in the UNISHALE
technology at Parachute, Colorado to produce more than 0:64 106 m3 (four
million barrels) of crude shale oil. Operations were shut down in 1991.
Lurgi. The Lurgi process, developed by Lurgi-Ruhrgas GmbH for mild
gasification of coal, has been tested as an oil shale retort. The Lurgi process
is similar to the Tosco II process in that it uses finely divided oil shale as a
feedstock and employs indirect heating and concurrent flows. The unique feature
of the Lurgi process is the use of hot combusted retorted shale as the heat-transport
medium (Fig. 7) and retorting is carried out by mixing the hot combusted shale
and raw oil shale in a screw conveyor. Because a portion of the retorted shale is
combusted to supply the process heat, the Lurgi process utilizes more of the
organic matter in the oil shale than other IH processes (see Table 7).
Superior. The Superior retort is different from all the other AGR pro-
cesses in that it consists of a slowly rotating circular grate instead of the vertical
204 OIL SHALE Vol. 1
Net product
Off-gas Gas
gas
Retort
Mist formation zone
Recycle gas
Recycle
blower
gas
Retorting zone Product oil
Retorted shale
Flue gas
to atmosphere
Ball Air
heater Fractionator Naphtha
Surge
Dilute phase fluid bed lift pipe
ea 648
°
sha ted Ho C
le ball t Accum- Resid
s
lator
Pyrolysis
as
drum Trommel
eg
482°C
t flu
Wa
Ho
rm
bal
ls
Hot
spent
shale
200 mesh
Spent (ca 74 µm)
shale spent shale
cooler to disposal
Raw Plant
shale in Hot recycled gas fuel gas
Recycled
Retort gas heater
gas
Oil
seal
Sulfer
R recovery
sh etor t plant
ale ed Venturi
ou scrubber Compressor
t Pump
Retorted oil Crude shale oil
Retorted shale
Retorted
shale
Upflow cooler
feeder
Recovery
Dust
cyclone
Oil Gaseous
Oil shale Waste and liquid
product
Condenser
Dust removal
Mixing screw cyclone
type retort Solids
surge
bin
Lift Solids
pipe to waste
Air + fuel
(if required)
Pre-heated
Hot recycled gas combustion air
Direct Air in
heating
low HV
burners
Liquid
seal
Oil
removal
Gas
compressor
low heating-
value surplus
retort gas out
Oil and water out
preheating. Because the gas composition varies within each of these zones, each
zone is separated by a baffle screen as the solids are transported around the
circle. The Superior retort was tested in Colorado in the late 1970s.
5.3. In Situ Retorting. True in situ retorting has been considered as a
means of avoiding the costs of mining, crushing, and surface disposal of spent
shale, and the associated environmental impacts of AGR. However, the imper-
vious nature of the oil shale formation and the overburden pressures have pre-
vented true in situ operations. Shale oil yields, the amount of oil produced
divided by the theoretical amount estimated to be in the oil shale rock, for
in situ retorting are usually half that experienced with AGR retorting. A true
in situ experiment, using drilling and resource fracturing procedures typical of
conventional petroleum development, was tried by the Energy Research Devel-
opment Administration (a forerunner of the U.S. Department of Energy) in 1975
in Rock Springs, Wyoming. No significant yields of shale oil were produced (22).
Other true in situ tests were conducted using the Equity BX superheated steam
process in Colorado, and Dow hot air process in Michigan; neither produced sig-
nificant yields of oil shale. It appears that true in situ retorting is not a practical
Vol. 1 OIL SHALE 207
Combustion
Gas
Air in Oil
Surface uplift
Preblast
surface
Overburden
Co
m
t
ga bu
fron
Oi
Oil Air flow l se stio
s n
shale Fire
bed
approach for the thick strata of oil shale normally situated deep below the
surface.
LOFRECO. The LOFRECO process, developed by Geokinetics, Inc.
(Utah) is a true in situ process. It is limited to relatively thin deposits of oil
shale situated beneath a relatively thin overburden (Fig. 9). The LOFRECO pro-
cess has been successful because the oil shale is rubblized in place by raising the
overburden. Retorting consists of direct combustion horizontally through the
rubblized formation, similar to that shown for the NTU retort (see Fig. 2).
Although the costs of mining and materials handling are obviated, the
LOFRECO process causes significant surface disturbance and results in oil
yields significantly lower than those obtained in the better controlled AGR
processes.
VMIS. The Vertical Modified in situ (VMIS) process consists of construct-
ing an underground retort of rubblized oil shale within the deep, thick deposits
situated in the Piceance Basin in western Colorado. In order to provide space for
the rubblization without upheaval of the overburden, a portion of the oil shale is
mined out and taken to the surface (Fig. 10). Retorting is carried out in the DH
mode exactly as shown in Figure 2. Steam and air are pumped into the top of the
VMIS retort, combustion proceeds down through the rubblized bed, and oil and
gas are pumped out from the bottom. Although yields are significantly lower
than those attained by AGR processes, the VMIS has demonstrated that modified
in situ can produce shale oil from thick deposits situated deep below the surface.
These operations have resulted in approximately one-fifth of the oil shale being
mined out to provide space for the VMIS rubblizing and retorting. A project
involving both VMIS and AGR processing, to utilize the mined-out shale, had
been planned, but was canceled in 1991 (24).
6.1. Properties. The composition of shale oil has depended on the shale
from which it was obtained as well as on the retorting method by which it was
208 OIL SHALE Vol. 1
produced. Properties of shale oils from various locations are given in Table 8. A
comparison of a Green River shale oil and a Michigan Antrim shale oil, retorted
under similar conditions, is given in Table 9.
Shale oil contains large quantities of olefinic hydrocarbons (see Table 8),
which cause gumming and constitute an increased hydrogen requirement for
upgrading. Properties for crude shale oil are compared with petroleum crude
in Table 10. High pour points prevent pipeline transportation of the crude
shale oil. Arsenic and iron can cause catalyst poisoning.
The primary difference in shale oils produced by different processing meth-
ods is in boiling point distribution. Rate of heating, as well as temperature level
and duration of product exposure to high temperature, affect product type and
yield (28). Gas combustion processes tend to yield slightly heavier liquid products
because of combustion of the lighter, ie, naphtha, fractions.
Carbon-to-hydrogen weight ratios for typical hydrocarbon fuels are natural
gas (methane), 3; gasoline, 6; crude oil, 6–7; shale oil, 7–8; Green River kerogen,
7; diesel and fuel oil, 8; residual oil, 10; and coal and coke, 12. A typical Green
River shale oil contains 40 wt% hydrocarbons and 60 wt% organic compounds
which contain nitrogen, sulfur, and oxygen. The nitrogen occurs in ring com-
pounds such as pyridines and pyrroles, as well as in nitriles, and these materials
comprise 60 wt% of the nonhydrocarbon organic components. Another 10 wt% of
these components is comprised of sulfur compounds which exist as thiophenes
and some sulfides and disulfides. The remaining 30 wt% is oxygen compounds
occurring as phenols and carboxylic acids (1).
6.2. Upgrading Shale Oil. Crude shale oil has a high (2 wt%) content
of organic nitrogen which acts as a catalyst poison, contains a large (20–50 wt%)
atmospheric residuum fraction, and has a high (>5
C) pour point (29,30).
Vol. 1 OIL SHALE 209
Table 10. Comparison of Green River Crude Shale Oil and Median U.S. Crudea
U.S. Bureau of Mines
Gas Simulated Union Oil Median U.S.
Property Tosco II combustion In situ in situ Co. A crude
distillation boiling
point, 8Cb
ibpc200 18 6 11–15 7 5 30
200–315 24 19 41–48 31 20 22
315–480 34 38 27–35 46 40 28
>480 24 37 9–14 17 35 20
pour point, 8C 1d, 15e 21–28 1 to 5 10–15 32 <15
specific gravity 0.927 0.934 0.892 0.910 0.940 0.850
8API 21 20 27 24 19 35
nitrogen, wt% 1.9 1.5–2.1 1.4–1.8 1.6 2.0 0.09
sulfur, wt% 0.7 0.8 0.7 0.6–0.9 0.9 0.6
oxygen, wt% 0.8 1.7 0.9
viscosity,
mm2 =s ð¼cStÞ
at 378C 22 59 8–15 21 46 6
1008C 04 07 06
saturates, vol% 30 59 41 60–100
olefins, vol% 50 16 37 <5
aromatics, vol% 20 25 22 0–40
carbon-to-hydro- 7–8 5–7
gen ratio
arsenic, ppm 40 <0:03
a
Refs. (1,27).
b
Values represent % of product.
c
ibp ¼ initial boiling point.
d
After a patented heat treatment which temporarily reduces pour point.
e
No heat treatment.
Prerefining crude shale oil to produce a synthetic crude that is compatible with
typical refineries generally is necessary (31–33). Prerefining to reduce organic
nitrogen content to low levels consists usually of either a delayed coking step
of the crude shale oil or residuum fraction, followed by one or more
hydrogenation steps, or a more severe direct hydrogenation of the crude shale
oil. Conditions for the hydrogenations are ca 4008C, 13.8 MPa (2000 psi) hydro-
gen partial pressure, and up to 356 standard cubic meters of hydrogen uptake
per cubic meter of shale oil (2000 ft3 at STP/bbl) (34). The nitrogen and sulfur
are converted to ammonia and elemental sulfur, and the hydrogen content of
the oil is increased. Upgraded shale oil is a desirable refinery feedstock. It is par-
affinic and characterized by low residuum, nitrogen, and sulfur (see FEEDSTOCKS).
Shale oil has been refined to produce gasoline, kerosene, jet fuel, and diesel
fuel (33). Different procedures have been tested to produce different product
states, eg, hydrotreating followed by hydrocracking for jet fuel production, hydro-
treating followed by fluid catalytic cracking for gasoline production, and coking
followed by hydrotreating for diesel fuel production. Production of military fuels
from the refining of 1590 m3 (10,000 bbl) of Paraho crude shale oil at the Gary
Western refinery in Colorado has been reported (35) and 15,900 m3 (100,000 bbl)
Vol. 1 OIL SHALE 211
Table 11. Properties of Shale Oil during the Unocal Upgrading Processa
Raw shale oil After particulates After arsenic After
from retort removal removal unicracking Syncrude
gravity, 8API 22 22 25 38
pour point, 8C 23.9 23.9 26.6 <3:9
particulates, 300 0 0 0
ppm
arsenic, ppm 25 25 0 0
distillation, 8C
initial 65.5 433
maximum 590 538
a
Ref. 36.
of Paraho shale oil has been processed under a U.S. Navy contract, at Sohio’s
Toledo refinery (33).
At the Parachute Creek Project, Unocal designed and operated an oil shale
upgrading unit to prerefine crude shale oil into syncrude, ie, upgraded shale oil
(36). The unit was designed to handle 1600 m3 (10,000 bbl) of crude shale oil per
stream day. More than 650,000 m3 (four million barrels) of syncrude were pro-
duced. Results of the Unocal shale oil upgrading process are given in Table 11.
The syncrude is compared with Arabian light crude oil in Table 12. Although
Arabian light is considered a premium crude oil among petroleum refiners, the
Unocal shale oil syncrude shows improvements in each of the characteristics
listed. Production of the conventional fuels and lubricants using Unocal shale
oil syncrude as the refinery feedstock is less difficult and less costly than using
Arabian light crude oil (see LUBRICATION AND LUBRICANTS).
7. Alternative Uses
Oil shale is an energy resource that produces a liquid fuel that can be used to
replace conventional crude oil or petroleum. However, the costs associated
with processing oil shale into conventional refined products are significantly
greater than that of processing conventional crude oil. In order to develop the
oil shale resource, other uses have been considered. These functions include
direct combustion to produce process heat for power generation, direct gasifica-
tion of the oil shale geological deposit, and special petrochemical production.
7.1. Direct Combustion. Direct combustion of oil shale has been used
to produce heat for power generation at specific sites and is being carried out
in Estonia and Israel.
In Estonia, most of the rich oil shale, 209 L/t kukersite, is burned as a
solid fuel to produce electric power (37). The kukersite, although technically
an oil shale, is actually similar to a high ash, low grade coal ideally suited
for this use.
In the remote Negev desert region of Israel, oil shale is being burned in a
fluidized-bed combustor to supply process heat and produce electric power.
Unlike the Estonian kukersite, this oil shale is lean, about 63 L/t, but is being
mined to access an underlying phosphate deposit (26).
7.2. Gasification. For significant conversion of shale oil or oil shale to
gaseous products, considerable hydrogen must be used. Hydrogasification is
the main process under consideration for gasification of oil shale. Hydroretorting
of oil shale has been studied extensively (38,39). Gasification of Colorado oil shale
in hydrogen and synthesis gas has been carried out at 524–7608C and at up to
38.3 MPa (5540 psig). Another study involves conventional oil shale retorting fol-
lowed by gasification of the resulting shale oil in a fluidized bed (40). The hydro-
gen is supplied by gasification of the coke on the spent shale and from the
gasifier. Other hydrogasification work includes the processing of Green River
and eastern Devonian oil shales (41,42). Because of their character, the eastern
Devonian shales produce less product oil, ca 35 wt% Fischer assay, than do
Green River shales. The hydrogasification process is claimed to recover about
90 wt% of the kerogen content. Additional oil shale gasification research at the
Laramie Energy Technology Center (U.S. DOE) has been reported (43–45).
7.3. Petrochemicals Production. Early Consideration. In the
1950s, the U.S. Bureau of Mines studied pyrolysis of both oil shale and shale
oil for the production of light olefins. High temperature retorts were used to
determine the effect of continuing the cracking, which is begun when the kerogen
is converted to shale oil (46). Low temperature shale oils are low in aromatic con-
tent, thus one goal has been the production of an aromatic-rich naphtha. High
temperature retorting of this type on Green River oil shales has the disadvantage
of the additional energy requirements of the endothermic carbonate decomposi-
tion, with over 50 wt% decomposition occurring at 8158C. This effect has little
importance for low carbonate oil shale, eg, the eastern U.S. Antrim oil shales
(41). Comparison of thermal cracking of conventional shale oil to high tempera-
ture retorting under the same conditions illustrates that naphtha production is
enhanced considerably by high temperature retorting (46).
Utilization of shale oil products for petrochemical production has been stu-
died (47–51). The effects of prerefining on product yields for steam pyrolysis of
shale oil feed and the suitability of Green River shale oil as a petrochemical feed-
stock were investigated. Pyrolysis was carried out on the whole oil, vacuum
distillate, and mildly, moderately, and severely hydrogenated vacuum distillates.
Specialty Chemicals. Specialty chemicals have been considered an eco-
nomically attractive means of using the oil shale resource. These specialty che-
micals consist of high value, niche market items (52) that utilize the high
Vol. 1 OIL SHALE 213
8. Environmental Issues
The plans to develop a commercial oil shale industry in the three-state region of
Colorado, Utah, and Wyoming in the 1970s raised the possibility of significant
adverse environmental, health, safety, and socioeconomic (EHSS) impacts. Pro-
cessing oil shale to produce oil on a large-scale commercial basis requires a large
amount of mining, crushing, material transport, and disposal operations.
Adverse EHHS impact could result from uncontrolled, or inadequately con-
trolled, large-scale oil shale operations. Without controls, significant amounts of
dust, ie, particulates, would be produced. Because the gas produced from kerogen
breakdown contains significant amounts of hydrogen sulfide and ammonia,
uncontrolled release, or direct combustion with no control technology, could
pose adverse health impacts and air pollution. The liquids produced from retort-
ing operations, ie, process water and crude shale oil, contain significant levels of
214 OIL SHALE Vol. 1
Table 14. Estimate of Colony Oil Shale Project Emissions Production, kg=ha,b
Solid Hydro- Carbon
Source SO2 NOx particulates carbons monoxide
crushing and conveying
primary crusher dust 0 0 27 0 0
collection system
final crusher dust 0 0 136 0 0
collection system
fine-ore storage dust 0 0 27 0 0
collection system
pyrolysis and oil recovery unit
preheat systems 529 2172 51 128 17.4
steam superheater ball 9 56 103 0.2 1.0
circulation systems
processed shale 0 0 116 0 0
moisturizing systems
hydrogen unit
reforming furnaces 146 244 4.9 0.8 4.6
gas–oil hydrogenation unit
reactor heaters 10 54 1.1 0.4 0.45
reboiler heater 2.7 17.7 0.4 0.04 0.3
naphtha hydrogenation unit
reactor heater 2.2 4 0.09 0.01 0.09
sulfur recovery unit
sulfur plants with 29
common tailgas plant
delayed coker unit
heater 21 35 0.7 0.14 0.64
utilities
boilers 51 246 5.4 1.8 1.4
Total 799.9 2828.7 472.59 131.39 25.88
a
Ref. 57.
b
Using no control technology.
1600 100
Density, lb/ft3
No added water
1440 90
1360 85
500 1500 2500
Compactive effort, kJ/m3
facilities. No violations for unauthorized air emissions were issued by the U.S.
Environmental Protection Agency during this time (62).
8.2. Water Quality. All commercial oil shale operations require substan-
tial quantities of water. All product water is treated for use and operations are
permitted as zero-discharge facilities. In the Unocal operation, no accidental
releases of surface water have occurred during the last four years of sustained
operations from 1986 to 1990. The Unocal Parachute Creek Project compliance
monitoring program of ground water, surface water, and process water
streams have indicated no adverse water quality impacts and no violations of
the Colorado Department of Health standards (62).
8.3. Solids. Proper handling and disposal techniques can obviate poten-
tial problems associated with the solid waste-retorted shale. Retorted shale dis-
posal and revegetation have posed no adverse environmental impacts at the
Unocal Parachute Project (62). Earlier studies carried out using Paraho and
Lurgi retorted shales indicated that these materials behave as low grade
cements (63,64) and can be engineered and compacted into high density materi-
als (Fig. 11) and water impervious structures (Table 15).
8.4. Health and Safety. Much of the adverse health issue publicity invol-
ving risks of exposures to carcinogens, such as benzo-a-pyrene, have been based on
recorded exposures of Scottish oil shale workers that took place nearly 100 years
ago. It is believed that the increase in cancer was due more to poor personal
hygiene than exposure to shale oil. Industrial hygiene monitoring and health sur-
veys indicate no significantly increased health risks among oil shale workers (62).
8.5. Socioeconomics. Impact from recurring boom-and-bust cycles
typified many of the earlier mining developments in the western United States
(65). However, this and the Oil Shale Trust Fund, established by the legislation
that set up the Federal prototype leases (C-a, C-b, U-a, and U-b), has provided
funds to local towns and counties to assist in the construction and upgrading
of infrastructures. These infrastructures are needed to accommodate the work
force needed to construct and operate the large oil shale processing facilities.
Further, companies such as Unocal have provided socioeconomic grants to
these counties (66).
9. Economic Aspects
As of 1995, there were only a few commercial oil shale facilities operating in the
world. These facilities are located in countries where the economic, political, and
environmental requirements for commercial oil shale development are met.
There are commercial oil shale facilities in Brazil, China, Estonia, and Israel.
No commercial oil shale facilities have existed in the United States because
the costs of shale oil processing exceed those associated with conventional petro-
leum crude processing.
In the United States, estimates of oil shale retorting have ranged from
$113/m3 ($18/bbl) to $567/m3 ($90/bbl). The lower estimate is based on using
the actual costs of constructing a commercial oil shale retorting facility and
using a high grade (>25 L=t ð>30 gal=short tonÞ) of oil shale (3). The higher
estimate is based on conservative estimates utilizing unproven or noncommercial
oil shale retorting technology.
The estimated costs for upgrading crude shale oil range from $38/m3 ($6/bbl)
to $63/m3 ($10/bbl). However, the resulting upgraded shale oil is superior to most
conventional crude petroleum and is more valuable as a refinery feedstock (67).
The costs for upgrading crude shale oil depend on the upgrading techniques, ie,
hydrotreating and coking or hydrotreating and fluid catalytic cracking. However,
the greatest economic factor in oil shale upgrading is the amount of hydrogen
required as reflected in the concentration of heteroatoms such as nitrogen,
sulfur, and oxygen.
The commercial production of shale oil as an alternative energy source has
not been economically feasible. As of 1995, all commercial oil shale operations in
the world (Petrobras, Brazil; PAMA, Israel; The Chinese Petroleum Corporation,
Fushun and Maoming, China; Kivioli Oil Shale Processing Plant, Kohtla-Jarve,
Estonia) receive some sort of economic incentives or assistance from the
countries in which they are operating.
The first stage of the Stuart oil shale project near Gladstone, Australia,
6000 t/d (6600 short tons/d), is scheduled to be constructed by Southern
Vol. 1 OIL SHALE 217
The number of commercial oil shale operations worldwide has decreased signifi-
cantly since the decade 1975–1985 and are producing only a fraction of the
world’s liquid fuels’ needs. Most commercial oil shale operations have been scaled
back.
10.1. PETROSIX Operations in Brazil. Petroleo Brasilerio (Petrobras)
has a dedicated facility to produce crude shale oil from the Irati formation in
southern Brazil. The facility is called the Oil Shale Industrialization Superinten-
dency (SIX) and uses the PETROSIX retorting technology (see Table 7 and
Fig. 3).
During its 40-year development, three different sizes of PETROSIX retorts
have been operated on a continuous basis: a 1.83-m (6-ft) diameter demonstra-
tion plant; a 5.49-m (18-ft) diameter Prototype Unit (UPI); and a 10.97-m
(36-ft) diameter Industrial Module (MI). Within the SIX facility are numerous
pilot plants available for retorting coarse-sized oil shale, fines utilization, and
oil shale upgrading (3,67).
The UPI and MI retorts are processing 7000 t/d (7700 short tons/d) of Irati
shale to produce 24,381 m3/d (3870 bbl) of shale as well as 80 t (55 short tons) of
LPG, 132 t (145 short tons) of clean fuel gas, and 98 t (108 short tons) of sulfur.
The SIX plant has reached its design rate (Table 16) in an energy efficient
manner with a high on-stream (operating) factor.
Pilot-plant studies have been conducted by Petrobras on many of the differ-
ent oil shales from around the world. Tests indicate that many oil shales could be
processed using the PETROSIX retorting technology (67) (Fig. 12).
218 OIL SHALE Vol. 1
10.2. Oil Shale Operations in Israel. Oil shale, the only fossil fuel
resource in Israel, is being used to generate electric power. The oil shale feed
stock, typical of the low grade Israeli oil shale (see Table 2), is situated in a
deposit overlying phosphate ore. The oil shale operations are being carried out
because the oil shale has to be mined to obtain the phosphate ore.
A circulating fluidized-bed boiler, using raw shale oil as a feedstock, is being
used to supply process heat for the phosphate operations and to operate a
100-MW power plant. Scale-up in the 1990s should increase the electric power
generation to 1000 MW (71).
10.3. United States. In 1980, Unocal began constructing the Parachute
Creek Project, designed to produce 1600 m3 (10,000 bbl) of upgraded shale oil per
day. The project included a conventional underground room-and-pillar mine, the
Unishale B (see Table 7) retort, and a special Unocal upgrading facility. Plant
startup occurred in 1986, and daily shale oil production reached 1100 m3/d
(7000 bbl/d). By 1991, total production exceeded 0:6 106 m3 (four million
barrels). However, the Parachute Creek Project was shut down in mid-1991 for
economic reasons.
The New Paraho Corporation has been conducting research on asphalt
derived from shale oil, SOMAT, at its pilot plant (Rifle, Colorado) (54,55). It is
the only active oil shale operation in the United States as of 1995. New Paraho
is continuing its pilot operations while designing a commercial facility to produce
SOMAT. The economics appear promising (56).
Utah (U.S.)
Wyoming (U.S.)
Colorado 1 (U.S.)
Colorado 2 (U.S.)
Indiana (U.S.)
Kentucky (U.S.)
China
Rundle (Australia)
Nagoorin (Australia)
Condor (Australia)
Estonia Oil content, wt % Abrasion index
Israel
Canada
10 20 30 40 50 60 70 80
Fig. 12. Shale oil characteristics compared to Irati oil shale (3).
Vol. 1 OIL SHALE 219
BIBLIOGRAPHY
‘‘Shale Oil’’ in ECT 1st ed., Vol. 12, pp. 207–221, by B. Guthrie and H. M. Thorne, Bureau
of Mines, U.S. Dept. of the Interior; in ECT 2nd ed., Vol. 18, pp. 1–20, by R. E. Gustafson,
Cameron and Jones, Inc.; ‘‘Oil Shale’’ in ECT 3rd ed., Vol. 16, pp. 333–357, by P. F. Dickson,
Colorado School of Mines.
CITED PUBLICATIONS
1. S. Siggia and P. Uden, eds., Analytical Chemistry Pertaining to Oil Shale and
Shale Oil, National Science Foundation Grant Number GP 43807, June 24–25,
1974, 11–13.
2. D. C. Duncan and V. E. Swanson, U.S. Geol. Surv. Circ. 523(7), 9 (1965).
3. D. L. Bachman and co-workers, 1993 Eastern Oil Shale Symposium, Lexington, Ky.,
Nov. 16–19, 1993.
4. H. M. Thorne and co-workers, U.S. Bur. Mines Inf. Circ. 8216, 6 (1964).
5. G. L. Baughman, ed., Synthetic Fuels Data Book, 2nd ed., Vol. 4, Cameron Engineers
(Division of The Pace Co.), Denver, Colo., 1978, 67–104.
6. Oil Shale and the Environment, EPA-600/9-77-033, U.S. Environmental Protection
Agency, Washington, D.C., Oct. 2, 1977.
7. ASTM D3904, American Society for Testing and Materials, Philadelphia,
Pa., 1980.
220 OIL SHALE Vol. 1
8. R. N. Heistand, Am. Chem. Soc. Div. Fuel Chemis. Preprint 21(6), (1976).
9. L. Goodfellow and M. T. Atwood, Colo. Sch. Mines Q. 69, 205 (Apr. 1974).
10. K. E. Stanfield, U.S. Bur. Mines Rep. Invest. 4825, (1951).
11. V. D. Allred, Colo. Sch. Mines Q. 62, 657 (1967).
12. J. J. Cummins and W. E. Robinson, U.S. Bur. Mines Rep. Invest. 7620, 1 (1972).
13. R. L. Braun and A. J. Rothman, Fuel 54, 129 (Apr. 1975).
14. J. H. Campbell and co-workers, paper presented at The 10th Oil Shale Symposium
Proceedings, Colorado School of Mines Press, July 1977, 148–165.
15. A. K. Burnham and co-workers, Comp. Stud. Various Oil Shales 34(1), 36–42
(Mar. 1989).
16. A. K. Burnham, Relationship Between Hydrous and Ordinary Pyrolysis, UCRL-
JC-114130, Lawrence Livermore National Laboratory, Livermore, Calif., June
1993.
17. S. S. Tihen, Ind. Eng. Chem. 47, 464 (1955).
18. G. U. Dineen, ‘‘Effect of Retorting Temperature on the Composition of Shale Oil,’’
AlChE Meeting, Denver, Colo., Aug. 1962.
19. J. H. Campbell and A. K. Burnham, paper presented at The 11th Oil Shale Sympo-
sium Proceedings, Colorado School of Mines Press, Golden, Colo., Nov. 1978,
242–259.
20. E. M. Piper, 6th IIASA Resource Conference, Luxenborg, Austria, Colorado School of
Mines Press, Golden, Colo., 1981.
21. J. B. Jones, Jr. and R. N. Heistand, Proceedings, 12th Oil Shale Symposium,
Colorado School of Mines Press, Golden, Colo., 1979, 184–195.
22. Energy Development Consultants, ‘‘Oil Shale in the United States, 1981,’’ EDC,
Golden, Colo., 1980.
23. A. Stevens and R. L. Zahradnik, Proceedings, 16th Oil Shale Symposium, Colorado
School of Mines Press, Golden, Colo., 1983, 267–268.
24. L. Dockter, Colo. Sch. Mines Q. 83(4), 120–122 (1988).
25. R. F. Crane, in Ref. 21, 1–16.
26. A. H. Pelofsky and co-workers, in Ref. 21, 32–42.
27. A. Long, Jr., N. W. Merriam, and C. G. Mones, in Ref. 15, 120–135.
28. N. D. South and co-workers, Colo. Sch. Mines Q. 71, 153 (Oct. 1976).
29. R. E. Poulson, C. M. Frost, and H. B. Jensen, ACS Div. Pet. Chem. Prepr. 17(2), 175
(1972).
30. M. T. Atwood, paper presented at The ACS Symposium: Fuels of the Future, Dallas,
Tex., Apr. 4–13, 1973.
31. C. M. Frost, R. E. Poulson, and H. B. Jensen, ACS Div. Pet. Chem. Prepr. 17(2), 156
(1972).
32. D. P. Montgomery, Ind. Eng. Chem. Prod. Res. Dev. 7, 274 (1968).
33. R. F. Sullivan, B. E. Stangeland, and H. A. Frumkin, ACS Div. Pet. Chem. Prepr. 22,
998 (1977).
34. V. F. Yesavage, C. F. Griswold, and P. F. Dickson, paper presented at The 180th
National ACS Meeting, San Francisco, Calif., Aug. 25–29, 1980.
35. H. Batrick and co-workers, Final Report—The Production and Refining of Crude
Shale Oil into Military Fuels, Applied Systems Corp. for Office of Naval Research,
Washington, D.C., Aug. 1975.
36. C. P. Reeg, A. C. Randle, and J. H. Duir, Proceedings, 23rd Oil Shale Symposium,
Colorado School of Mines Press, Golden, Colo., 1990, 68–95.
37. J. D. Baker and C. O. Hook, in Ref. 21, 26–31.
38. H. F. Feldman and co-workers, Inst. Gas Technol. Res. Bull. 36 (Aug. 1966).
39. S. A. Weil and co-workers, paper presented at The 167th National ACS Meeting, Los
Angeles, Calif., Apr. 1974.
Vol. 1 OIL SHALE 221
40. U.S. Pat. 3,703,052 (1972), H. R. Linden (to Institute of Gas Technology).
41. F. C. Schora and co-workers, Hydrocarbon Process. 56(4), 107 (1977).
42. F. C. Schora, ‘‘The Application of the IGT Hydroretorting Process to Eastern Shale,’’
Energy Topics—a Supplement to IGT Highlights, Institute of Gas Technology,
Chicago, Ill., May 9, 1977.
43. E. L. Burwell and I. A. Jacobson, Jr., ‘‘Concurrent Gasification and Retorting of Oil
Shale—a Dual Energy Source,’’ Rocky Mountain Regional Meeting, SPE 5535,
Society of Petroleum Engineers (SPE), AIME, Denver, Colo., Apr. 7–9, 1975.
44. E. L. Burwell and I. A. Jacobson, Jr., U.S. Bur. Mines Tech. Prog. Rep. 85, 1
(Nov. 1974).
45. E. L. Burwell and I. A. Jacobson, Colo. Sch. Mines Q. 71, 139 (Oct. 1976).
46. H. W. Sohns and co-workers, Ind. Eng. Chem. 47, 461 (1955).
47. C. F. Griswold, V. F. Yesavage, and P. F. Dickson, ACS Div. Fuel Chem. Prep. 21, 207
(1976).
48. E. A. Fritzler, V. F. Yesavage, and P. F. Dickson, Proceedings, The Second Pacific
Chemical Engineering Conference, American Institute of Chemical Engineers, New
York, 1977, p. 542.
49. P. D. Smith, P. F. Dickson, and V. F. Yesavage, ACS Div. Pet. Chem. Prepr. 23,
756 (1978).
50. C. F. Griswold, A. Ballut, H. R. Kavianian, P. F. Dickson, and V. F. Yesavage, Chem.
Eng. Prog. 75(9), 78 (1979).
51. C. F. Griswold, A. Ballut, H. R. Kavianian, P. F. Dickson, and V. F. Yesavage, Energy
Commun. 6(2), 153 (1980).
52. J. E. Sinor, ‘‘Niche Market Assessment for a Small Western Oil Shale Project,’’ Final
Report, DE-FC-86MC11076, U.S. Dept. of Energy, Washington, D.C., 1979.
53. J. E. Bunger and A. V. Deveni, Proceedings, 25th Oil Shale Symposium, Colorado
School of Mines Press, Golden, Colo., 1992, 281–294.
54. L. A. Lukens, Proceedings, 22nd Oil Shale Symposium, Colorado School of Mines
Press, Golden, Colo., 1989, 196–206.
55. L. A. Lukens and M. A. Plummer, Colo. Sch. Mines Q., 108–114 (1988).
56. L. A. Lukens, Col. Sch. Mines Q., 115–119 (1988).
57. An Environmental Impact Analysis for a Shale Oil Complex at Parachute Creek,
Colorado, Vol. I, II, III, Colony Development, Denver, Colo., 1974.
58. Detailed Development Plan, Vol. 1 and 2, Oil Shale Lease Tract C-b, C-b Shale Oil
Project, 1976.
59. Modifications to Detailed Development Plan, Oil Shale Tract C-b, C-b Shale Oil
Venture, 1977.
60. Project Independence Oil Shale Task Force Report, FEA, Washington, D.C., 1974,
p. 154.
61. T. L. Thoem and E. F. Harris, in Ref. 19, 1–9.
62. J. B. Benton, FUEL 71, 238–242 (Feb. 1992).
63. J. P. Fox, Proceedings of the 13th Oil Shale Symposium, Colorado School of Mines
Press, Golden, Colo., 1980, 131–139.
64. R. N. Heistand and W. G. Holtz, in Ref. 63, 140–150.
65. A. Gulliford, Boomtown Blues: Colorado Oil Shale, 1885–1985, University Press of
Colorado, Boulder, 1989.
66. J. Evans, Colo. Sch. Mines Q., 103–105 (1988).
67. E. M. Piper and co-workers, Proceedings of the 25th Oil Shale Symposium, Colorado
School of Mines Press, Golden, Colo., 1992, 221–242.
68. 1992 Annual Report, Southern Pacific Petroleum N.L., 1993.
69. 1984 Annual Report, United States Synthetic Fuels Corp., 1985.
70. Oil Gas J., 38 (Apr. 8, 1991).
222 TAR SANDS Vol. 1
71. J. Yerushalmi, Proceedings of the 1992 Eastern Oil Shale Symposium, IMMR Press,
Lexington, Ky., 1993, p. 367.
72. C. Du, Proceedings of the 18th Oil Shale Symposium, Colorado School of Mines
Press, Golden, Colo., 1985, 210–215.
73. B. C. Wright, Alternate Energy ’89 Proceedings, Council on Alternate Fuels,
Washington, D.C., 1989, 175–194.
EDWIN M. PIPER
Piper Designs LLC
ROBERT N. HEISTAND
Consultant
TAR SANDS
In addition to conventional petroleum (qv) and heavy crude oil, there remains
another subclass of petroleum, one that offers to provide some relief to poten-
tial shortfalls in the future supply of liquid fuels and other products. This
subclass is the bitumen found in tar sand deposits (1,2). Tar sands, also
known as oil sands and bituminous sands, are sand deposits impregnated
with dense, viscous petroleum. Tar sands are found throughout the world,
often in the same geographical areas as conventional petroleum.
Petroleum, and the equivalent term crude oil, cover a vast assortment of
materials consisting of gaseous, liquid, and solid hydrocarbon-type chemical com-
pounds that occur in sedimentary deposits throughout the world (3). When pet-
roleum occurs in a reservoir that allows the crude material to be recovered by
pumping operations as a free-flowing dark- to light-colored liquid, it is often
referred to as conventional petroleum.
Heavy oil is another type of petroleum, different from conventional petro-
leum insofar as the flow properties are reduced. A heavy oil is much more diffi-
cult to recover from the subsurface reservoir. These materials have a high
viscosity and low API gravity relative to the viscosity and API gravity of conven-
tional petroleum (Fig. 1) (3, 4), and recovery of heavy oil usually requires thermal
stimulation of the reservoir.
The definition of heavy oil is usually based on API gravity or viscosity, but
the definition is quite arbitrary. Although there have been attempts to rationa-
lize the definition based on viscosity, API gravity, and density (2,3), such defini-
tions, based on physical properties, are inadequate, and a more precise definition
would involve some reference to the recovery method.
In a general sense, however, the term heavy oil is often applied to a petro-
leum that has a gravity <208API. The term heavy oil has also been arbitrarily
used to describe both the heavy oil that requires thermal stimulation for recovery
from the reservoir and the bitumen in bituminous sand (also known as tar sand
or oil sand) formations, from which the heavy bituminous material is recovered
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.20011819160509.a01
Vol. 1 TAR SANDS 223
106
105
Athabasca
4
10
Viscosity, cps
102
Heavy oil
(Lloydminster)
10
1
50 100 150
Temperature, °C
Fig. 1. Relative viscosity data for conventional petroleum, heavy oil, and bitumen.
1. Origin of Bitumen
There are several general theories regarding the origin of the bitumen. One the-
ory is that the oil was formed locally and has neither migrated a great distance
nor been subjected to large overburden pressures. Because under these condi-
tions the oil cannot have been subjected to any thermal effects with the resulting
decomposition or molecular changes, it is geologically young and therefore dense
and viscous.
Another theory promotes the concept of a remote origin for the bitumen, or,
more likely, the bitumen precursor, both geographically and in geological time.
The bitumen precursor, originally resembling a conventional crude oil, is
assumed to have migrated into the sand deposit, which may originally have
been filled with water. After the oil migrated, the overburden pressures were
relieved, and the light portions of the crude evaporated, leaving behind a
dense, viscous residue.
Included in the remote origin theory is the postulate that the light hydro-
carbons were destroyed by bacteria carried into the petroleum reservoirs in oxy-
genated, meteoric waters. The remote origin theory would explain the water
layer surrounding sand grains in the Athabasca deposit. However, because the
metals and porphyrin contents of bitumen are similar to those of some conven-
tional Alberta crude oils of Lower Cretaceous age and because Athabasca bitu-
men has a relatively low coking temperature, the bitumen may be of Lower
Cretaceous age. This is the age of the McMurray formation (Canada), which is
geologically young. This evidence supports the theory that the oil was formed
in situ and is a precursor, rather than a residue of some other oil. The issue
remains unresolved as of this writing (ca 1997).
2. Occurrence
Many of the reserves of bitumen in tar sand formations are available only with
some difficulty, and optional refinery methods are necessary for future conver-
sion of these materials to liquid products, because of the substantial differences
in character between conventional petroleum and bitumen (Table 1).
Because of the diversity of available information and the continuing
attempts to delineate the various world oil sands deposits, it is virtually impos-
sible to reflect the extent of the reserves in terms of barrel units with a great
degree of accuracy. The potential reserves of hydrocarbon liquids that occur
Vol. 1 TAR SANDS 225
in tar sand deposits have, however, variously been estimated on a world basis to be
in excess of 477 109 m3 ð3 1012 bblÞ. Reserves that have been estimated for
the United States are believed to be in excess of 795 104 m3 ð50 106 bblÞ,
although estimates vary. Bitumen reserves throughout the world can compare
favorably with reserves of conventional crude oil.
Tar sand deposits are widely distributed throughout the world (Fig. 2)
(5,6) and the various deposits have been described as belonging to two types:
stratigraphic traps and structural traps (Table 2; Fig. 3) (7). However, there
are the inevitable gradations and combinations of these two types of deposits,
and thus a broad pattern of deposit entrapment is believed to exist. In general
terms, the entrapment character of the very large tar sand deposits involves a
combination of both stratigraphic and structural traps.
The largest tar sand deposits are in Alberta, Canada, and in Venezuela.
Smaller tar sand deposits occur in the United States (mainly in Utah), Peru,
Trinidad, Madagascar, the former Soviet Union, Balkan states, and the Philip-
pines. Tar sand deposits in northwestern China (Xinjiang Autonomous Region)
also are large; at some locations, the bitumen appears on the land surface around
226 TAR SANDS Vol. 1
Melville Island
Athabasca
Olenek
Selenizza
Derna
Whiterocks Asphalt Kobystan
Sunnyside Davis-Dismal
Burgan
Asphalt Ridge Creek Melang
Peor springs Kyrock
Guanoco Leyte
Santa Cruz La Brea
Edna Incarte Bemolanga
Sisquoc Bolivar
Santa Rosa
Oficina-Temblador
Fig. 2. Principal tar sand deposits of the world, where . represents > 2; 385; 000 m3
< 15 106 bblÞ bitumen; N, probably >159; 000 m3 ð < 1 106 bblÞ bitumen; and H,
reported occurrence information limited.
Karamay, China. The largest deposits are in the Athabasca area in the province
of Alberta, Canada, and in the Orinoco region of east central Venezuela.
The Athabasca deposit, along with the neighboring Wabasca, Peace River,
and Cold Lake heavy oil deposits, have together been estimated to contain
1:86 1011 m3 ð >1:17 1012 bblÞ of bitumen. The Venezuelan deposits may at
least contain >1:60 1011 m3 ð1:0 1012 bblÞ bitumen (2). Deposits of tar sand,
each containing >3 106 m3 ð20 106 bblÞ of bitumen, have also been located
in the United States, Albania, Italy, Madagascar, Peru, Romania, Trinidad,
Zaire, and the former Soviet Union, comprising a total of ca 450 109 m3
ð2:8 1012 bblÞ.
Melville Island
Athabasca
Sisquoc La Brea
Bemolanga
Santa Rosa
Oficina-Temblador
Fig. 3. Types of traps for tar sand deposits, where represents a stratigraphic trap, ,
an intermediate between stratigraphic and structural/stratigraphic traps; , a structural/
stratigraphic trap;., an intermediate between structural/stratigraphic and structural
traps; and 4, a structural trap.
The Alberta (Athabasca) tar sand deposits are located in the northeast part of
that Canadian province (Fig. 4). These are the only mineable tar sand deposits
undergoing large-scale commercial exploitation as of this writing (ca 1997).
The Athabasca deposits have been known since the early 1800s. The first
scientific interest in tar sands was taken by the Canadian government in
1890, and in 1897–1898, the sands were first drilled at Pelican Rapids on the
Athabasca River. Up until 1960, many small-scale commercial enterprises
were attempted but not sustained. Between 1957 and 1967, three extensive
pilot-plant operations were conducted in the Athabasca region, each leading to
a proposal for a commercial venture, eg, Suncor and Syncrude.
The Venezuelan tar sands are located in a 50–100-km belt extending east to
west for >700 km, immediately north of the Orinoco River. The precise limits of
the deposit are not well defined because exploration efforts in the past concen-
trated on light and medium crude accumulations.
The geological setting of the Orinoco deposit is complex, having evolved
through three cycles of sedimentation. The oil is contained by both structural
and stratigraphic traps, depending on location, age of sediment, and degree of
faulting. The tar sands are located along the southern flanks of the eastern Vene-
zuelan basin, where three distinct zones are apparent from north to south: a zone
of tertiary sedimentation, a central platform with transgressive overlapping
sediments, and a zone of erosional remnants covered by sediments. The deposit
also contains three systems of faulting. All the faults are normal and many are
concurrent with deposition.
Tar sands in the United States are contained in a variety of separate
deposits in various states (Fig. 5) but because many of these deposits are
228 TAR SANDS Vol. 1
Saskatchewan
Wabasca
Grand
Prairie
Cold Lake
Edmonton
Jasper
Calgary
0 100 200 km
Medicine
Hat
Canada
49°
United States 110°
3. Properties
Tar sand has been defined as sand saturated with a highly viscous crude hydro-
carbon material not recoverable in its natural state through a well by ordinary
production methods (2–8). Technically the material should perhaps be called
bituminous sand rather than tar sand because the hydrocarbon is bitumen, ie,
a carbon disulfide-soluble oil.
The data available are generally for the Athabasca materials, although
workers at the University of Utah (Salt Lake City) have carried out an intensive
program to determine the processibility of Utah bitumen and considerable
data have become available. Bulk properties of samples from several locations
(Table 3) (9) show that there is a wide range of properties. Substantial
differences exist between the tar sands in Canada and those in the United
Vol. 1 TAR SANDS 229
States; a difference often cited is that the former is water-wet and the latter,
oil-wet (10).
4. Recovery
Oil prices and operating costs are the key to economic development of tar sand
deposits. However, two technical conditions of vital concern for economic devel-
opment are the concentration of the resource (percent bitumen saturation) and
its accessibility, usually measured by the overburden thickness.
The remoteness of the U.S. tar sands is often cited as a deterrent to devel-
opment but topography of the site, overburden-to-ore body ratio, and richness of
the ore body are also important. In the 1990s context of mining tar sand deposits
in the United States, the Utah deposits (Tar Sand Triangle, P.R. Springs, Sunny-
side, and Hill Creek) generally have an overburden-to-net pay zone ratio above
232 TAR SANDS Vol. 1
Mining In situ
the 0.4–1.0 range, with a lean oil content. On the other hand, the Asphalt Ridge
deposit is loosely consolidated and could be mined using a ripper/front-end loader
(without drilling and blasting) at the near-surface location of the deposit.
Recovery methods are based either on mining combined with some further
processing or operation on the oil sands in situ (Fig. 6). The mining methods are
applicable to shallow deposits, characterized by an overburden ratio (ie, overbur-
den depth-to-thickness of tar sand deposit) of ca 1.0. Because Athabasca tar
sands have a maximum thickness of ca 90 m and average ca 45 m, there are indi-
cations that no more than 10% of the in-place deposit is mineable within 1990s
concepts of the economics and technology of open-pit mining.
The bitumen in the Athabasca deposit, which has a gravity on the API scale
of 88, is heavier than water and very viscous. Tar sand is a dense, solid material,
but it can be readily dug in the summer months; during the winter months when
the temperatures plunge to 45
C, tar sand assumes the consistency of concrete.
To maintain acceptable digging rates in winter, mining must proceed faster than
the rate of frost penetration; if not, supplemental measures such as blasting are
required.
4.1. Nonmining Methods. Nonmining (in situ) processes depend on
injecting a heating-and-driver substance into the ground through injection
wells and recovering bitumen through production wells. Such processes need a
relatively thick layer of overburden to contain the driver substance within the
formation between injection and production wells (2).
In principle, the nonmining recovery of bitumen from tar sand deposits is
an enhanced oil recovery technique and requires the injection of a fluid into
the formation through an injection well. This leads to the in situ displacement
of the bitumen from the reservoir and bitumen production at the surface through
an egress (production) well. There are, however, several serious constraints that
are particularly important and relate to the bulk properties of the tar sand and
the bitumen. In fact, both recovery by fluid injection and the serious constraints
on it must be considered in toto in the context of bitumen recovery by nonmining
techniques (see Petroleum, enhanced oil recovery).
Vol. 1 TAR SANDS 233
The Athabasca tar sands deposit in Canada is the site of the only commer-
cial tar sands mining operations. The Suncor operation (near Fort McMurray,
Alberta), started production in 1967. The Syncrude Canada project, located
8 km away, started production in 1978. In both projects, about half of the terrain
is covered with muskeg, an organic soil resembling peat moss, which ranges from
a few centimeters to 7 m in depth. The primary part of the overburden, however,
consists of Pleistocene glacial drift and Clearwater Formation sand and shale.
The total overburden varies from 7 to 40 m in thickness, and the underlying
tar sand strata averages about 45 m, although typically 5–10 m must be
discarded because of a bitumen content below the economic cut-off grade of ca
6% wt/wt.
Mining of the Athabasca tar sands presents two principal issues: in-place
tar sand requires very large cutting forces and is extremely abrasive to cutting
edges, and both the equipment and pit layouts must be designed to operate
during the long Canadian winters at temperatures as low as 40
C.
There are two approaches to open-pit mining of tar sands. The first uses a
few mining units of custom design, which are necessarily expensive, eg, bucket-
wheel excavators and large drag lines in conjunction with belt conveyors. In the
second approach, a multiplicity of smaller mining units of conventional design is
employed at relatively much lower unit costs. Scrapers and truck-and-shovel
operations have been considered. Each method has advantages and risks. The
first approach was originally adopted by Suncor and Syncrude Canada, Ltd.,
with Suncor converting to large-scale truck and shovel technology in 1993.
In the Suncor pit design, the ore body is divided into two layers (benches),
each nominally 23 m high. The pit floor and the dividing plane between the upper
and lower bench are roughly horizontal, and 7300-t/h bucket-wheel excavators
are employed as the primary mining equipment (Fig. 7). Tar sands loosened
from the face of each bench by the bucket-wheels are discharged onto a series
of conveyors. The overburden is stripped by an electric shovel that discharges
to trucks for removal of the overburden material. Syncrude utilizes a single-
bench design with four 60-m3 capacity draglines as the primary mining equip-
ment (Fig. 8). The draglines pile tar sands in windrows along the edge of the
pit; four 60,000-t/h bucket-wheels transfer the tar sands to a system of trunk con-
veyor belts that move the material to the extraction plant. The mining operations
Final Overburden
dyke
First mining
elevation
bench
90 m
(min)
Fig. 7. Mining with a bucket-wheel excavator, where () represents the ultimate size of
the dyke.
236 TAR SANDS Vol. 1
Overburden
Overburden
Orebody piles
Pit floor Worked-out area
at the two plants differ by choice of the primary mining equipment; the bucket-
wheel excavators sit on benches, whereas the draglines sit on the surface.
Bucket-wheel excavators use units having a 10 m dia digging wheel on the
end of a long boom. Each wheel has a theoretical capacity of 8700 t/h, but
the average output from digging is about 4500 t/h. At the rate of 122,000 t/d,
tar sand can be transferred from mine to plant by a system of 152-cm wide con-
veyor belts and 183-cm trunk conveyors, operating at 333 m/min. The bucket-
wheel excavators are supplemented by front-end loaders used to dig overburden
and load it through twin chutes onto ca 135 t capacity trucks. Additional equip-
ment is used for maintaining the haul roads and for spreading and compacting
the spoiled material. Overburden may be stripped with 14-m3 hydraulically oper-
ated shovels and a fleet of ca 135 t trucks.
Draglines are equipped with a 71-m3 bucket at the end of a 111-m boom and
can be employed to dig both a portion of the overburden, which is free-cast into
the mining pit, and the tar sand, which is piled in windrows behind the machine.
Bucket-wheel reclaimers, similar to bucket-wheel excavators, load the tar sand
from the windrows onto conveyor belts which transfer it to the plant.
5. Processing
Feed
bin Make up Make up
water water Final
Centrifuges
bitumen
Conditioning Vibr
drum at
scre ing Tailings
en
Settled froth
Oversize
Caustic
Steam
addition Froth
settler
Froth
237
Froth
Middlings
return Froth
Middlings
Scavenger
Middlings cells
Air
Sand
Separation cell
Tailings
Ore surge
re
do
s an
Tar
Crushed tar sand ore
Conditioning
Soda ash
Hot recycle water
Bitumen
Steam concentrate
Settling
239
Air
Solvent
Filtration
Steam Bitumen
clean up Bitumen– solvent Sand and water
Clean bitumen
Sand and water
3m
Tails
Clear water layer
30 m or more 6 m
Transition zone
Dyke
Dyke
Equilibrium
sludge volume
About 15cm
Consolidation
layer
valent bases. Polyvalent cations must be excluded because they tend to flocculate
clays and thus raise viscosity of the middlings in the separation cell.
One problem resulting from the hot-water process is disposal and control of
the tailings. Each ton of oil sand in place has a volume of ca 0.45 m3, which gen-
erates ca 0.6 m3 of tailings and gives a substantial volume gain. If the mine pro-
duces 200,000 t/d of oil sand, volume expansion represents a considerable solids
disposal problem.
Environmental regulations in Canada and the United States do not allow
the discharge of tailings streams into the river, onto the surface, or onto any
area where contamination of groundwater domains or the river may occur. The
tailings stream is essentially high in clays and contains some bitumen; hence the
need for tailings ponds, where some settling of the clay occurs (Fig. 11). In addi-
tion, an approach to acceptable reclamation of the tailings ponds must be accom-
modated at the time of site abandonment. Problems may be alleviated somewhat
by the development of process options that require considerably less water in the
sand–bitumen separation step. Such an option would allow a more gradual
removal of the tailings ponds.
5.2. Cold-Water Process. The cold-water bitumen separation process
has been developed to the point of small-scale continuous pilot plants. The pro-
cess uses a combination of cold water and solvent. The first step usually involves
disintegration of the tar sand charge, which is mixed with water, diluent, and
reagents. The diluent may be a petroleum distillate fraction such as kerosene
and is added in a ca 1:1 weight ratio to the bitumen in the feed. The pH is main-
tained at 9–9.5 by addition of wetting agents and ca 0.77 kg of soda ash per ton of
tar sand. The effluent is mixed with more water, and in a raked classifier the
sand is settled from the bulk of the remaining mixture. The water and oil over-
flow the classifier and are passed to thickeners, where the oil is concentrated.
Clay in the tar sand feed forms emulsions that are hard to break and are wasted
with the underflow from the thickeners.
The sand reduction process is a cold-water process without solvent. The
objective is removal of sand to provide a feed suitable for a fluid coking process.
In the first step, the tar sand feed is mixed with water at ca 208C in a screw
conveyor at a ratio of 0.75–3 t/t tar sand (lower range preferred). The mixed
pulp from the screw conveyor is discharged into a rotary-drum screen, which
Vol. 1 TAR SANDS 241
6. Bitumen Conversion
Flue gas
Oil product
Sand Spent
recycle sand By-product gas
(20–40%) is rejected and the remainder is recirculated to the coker to provide the
heat for the coking reaction. The products leave the coker as a vapor, which is
condensed in a receiver. Reaction off-gases from the receiver are recirculated
to fluidize the clean, hot sand, which is returned to the coker.
6.2. Conversion of Separated Bitumen. The overall upgrading pro-
cess by which bitumen is converted to liquid products is accomplished in two
steps (Fig. 13). The first step is the primary upgrading process, which improves
the hydrogen-to-carbon (H/C) ratio by either carbon removal or hydrogen addi-
tion, cracking the bitumen to lighter products which are more easily processed
Primary Hydrotreating
conversion Naptha
Pitch
or
coke
Collected gas
Sulfur
H2S Recovery
S Plant
Gas
H2 Plant
Gas
Health and safety factors associated with tar sand processing depend on
the nature of the process and products (Table 6; Fig. 14). Issues arising from
tar sand mining (Table 7) (25) are similar to other large-scale surface mining
Water or Cracked
solvent Bitumen products Hydrogen Conventional
245
Cracking treatment refinery
extraction
process feedstock
246
tailings disposal þ
bitumen in tailings or low grade þ þ
tar sand waste
fines in tailings
stripped waste
solubles or water transportation þ þ þ
particles in overburden
new surface þ
increases in surface slope from þ
waste disposal
rehandling of materials: back-filling, þ
grading, and recontouring
Vol. 1 TAR SANDS 247
Alberta Energy
Co., Ltd.
Cold Lake 13 immediate SAGD heavy 120 (1,000)
oil pilot plant heavy oil
Foster Creek 200 1997–1998 commercial 3,600 (30,000)
heavy oil heavy oil
Elan Energy, Ltd. 60 immediate SAGD heavy increased 1,680–
Lindbergh oil plant 2,640 (14,000–
22,000)
Elk Point
Gibson Petroleum, 7 1996–1997 SAGD bitumen increased 240–
Ltd., project 540 (2,000–
Utf 4,500)
Gulf Canada
Resources, Ltd.
Surmont 30 1996 phase 1 SAGD 360 (3,000)
heavy oil
1998–1999 phase 2 increased 2,400
(20,000)
Imperial Oil Limited
Cold Lake 250 immediate phases 9–10 in increased
situ bitumen 14,400;
production 3,000
(120,000;
25; 000þ)
Cold Lake 250 1997 phases 11–12 add 2,400 (20,000)
(proposed) in situ bitumen
production
following
completion
of phases 9–10
Imperial
Cold Lake 40 immediate heavy oil 1,128 (9,400)
development heavy oil
below current
Cold Lake lease
Koch Oil Co., Ltd.
Cold Lake 1997–1998 in situ bitumen 6,000 (50,000)
project
Reita Lake
Shell Canada Ltd.
Peace River 3 immediate SAGD heavy oil increased
project
Solvex
Bitumount 170 1997 bitumen mine/ 168 (14,000)
extraction bitumen, 100,000 t
complex, alumina, 200,000 t
minerals synthetic silica
extraction plant (tailings)
248 TAR SANDS Vol. 1
Table 8. (Continued)
Investment, Production,
Project $106 Timing Descriptiona m3/d (bbl/d)b
Suncor Inc.
Primrose 17 1996–1997 phase 1, 300 (2,500)
SAGD heavy oil
pilot plant
Burnt Lake 122 1998–1999 phase 2, increased 1,500
commercial (12,500)
plant heavy oil
Suncor Inc., Osg
Steepbank Mine 300 2001 new bitumen increased
mine production
10,800 (90,000þ)
light sweet crude,
custom blends
Fort McMurray 300 1998–2001 upgrader increased 9,600
improve- (80,000) light
ments, sweet crude,
plant custom blends
modifications
Fort McMurray 190 immediate environmental reduced SO2
measures emissions
Syncrude
Canada, Ltd.
North Mine 500 1998 new bitumen increased
Mildred Lake mine, 13 106 m3
debottlene-
cking plant
Syncrude Canada,
Ltd.
Aurora Mine 500 2001 new bitumen, first train, 2001;
remote increased 11.3
plant, (94 106 bbl=yr)
debottleneck
Total 3,452 $ 109
a
SAGD ¼ steam assisted gravity draining.
b
To convert m3/d to bbl/d, multiply by 6.29.
8. Future Outlook
BIBLIOGRAPHY
‘‘Tar Sands’’ in ECT 1st ed., Vol. 13, pp. 633–645, by K. A. Clark, Research Council of
Alberta, and G. B. Shea (U.S. Deposits), U.S. Department of the Interior; in ECT 2nd
ed., Vol. 16, pp. 682–732, by F. W. Camp, Sun Oil; in ECT 3rd ed., Vol. 22, pp. 601–
627, by D. Towson, Petro-Canada.
CITED PUBLICATIONS
JAMES G. SPEIGHT
Western Research Institute
DRILLING FLUIDS
1. Introduction
A range of specialty chemicals is used by the petroleum industry for oil and gas
exploration, drilling and production, and for enhanced oil recovery. Broadly
speaking, drilling fluids are used to lubricate the drill bit, to control formation
pressure and to remove formation cuttings. Workover and completion fluids
are utilized when operating in producing formations. Chemicals are used to
cement steel pipes or casing to the sides of the borehole and to encourage the
flow of crude oil to the well (stimulation). Chemicals find applications at all
stages from oil production at the well bore continuing on to the delivery of
crude oil to the refinery. During Enhanced Oil Recovery (EOR), chemicals
are used for the various techniques (ie, steam, carbon dioxide, polymer and
micellar-polymer flooding) that renew the oil flow from fields that have ceased
to produce by conventional methods.
Drilling fluids or muds are considered an essential component of the rotary
drilling process used in drilling for oil and gas on land and in offshore environ-
ments. This fluid performs a variety of functions that influence the drilling rate,
and the cost, efficiency, and safety of the operation (1–7). Some of the most
important functions of drilling fluids include (1) the transporting of the drilling
cuttings to the surface (2) balancing subsurface and formation pressures
required for the prevention of well blowouts (3) cooling, lubricating and support-
ing part of the weight of the drill and drill pipe. During the drilling of a well, the
drilling fluid mud is pumped from the mud tanks down the hollow drill pipe and
through nozzles in the drill bit. Typical diameter of a drill pipe is 4 in. (10 cm).
The borehole at the bottom of a deep well might be 8 in. (20 cm) in diameter. The
flowing mud sweeps the crushed rock cuttings from under the bit and carries
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0418091203120118.a01.pub2
Vol. 1 DRILLING FLUIDS 251
them back up the annular space between the drill pipe and the bore hole or cas-
ing to the surface. After reaching the surface, the drilling fluid is passed through
a series of vibrating screens, settling tanks or pits, hydrocyclones and centrifuges
to remove the cuttings brought to the surface. It is then treated with additives to
obtain a set of desired physical and chemical properties. Once treated, the fluid is
pumped back into the well and the cycle repeated.
Drilling fluids generally are composed of liquids: eg, water, petroleum oils,
and other organic liquids; dissolved inorganic and organic additives; and sus-
pended, finely divided solids of various types. The chemistry of the liquid
phase and the level of suspended solids determine the treatment strategy and
the efficiency of the mud-handling equipment. This chemistry also influences
the type and amount of materials needed to maintain or change the density,
viscosity, and other properties (8). Drilling fluids are dynamic systems. These
muds constantly change in response to the changing conditions while the well is
being drilled and downhole temperatures and pressures increase. Occasionally,
the downhole environment requires replacement of one fluid with another of a
different type.
Drilling fluid costs range from several thousand to several million dollars
per well and depend on the nature of the well being drilled. The length of the
drilling time may vary from a few days to more than a year. On the average,
6–8% of the total drilling cost arises directly from the drilling fluid and
additives. Additional fluid and total well costs may arise from improperly formu-
lated or treated fluids that can prolong the drilling time. Total worldwide sales of
drilling fluids in 2000 has been estimated to be $2.28 109 (excluding Eastern
Europe and China) (9). About $1.1 109 or 40% of the total was spent in North
America including the United States. It has been predicted that the worldwide
sales of drilling fluids will rise to $2.72 109 in 2005.
The leading three companies in the North American market that specialize
in the supply of certain aggregates of chemicals, equipment, and technology used
for finding oil and drilling wells are M-I L.L.C. (owned by Schlumberger and
Smith Industries), Haliburton Company (which owns Baroid), and Baker Hughes
Inteq (owned by Baker Hughes). These companies are commonly referred to as
drilling fluid (or drilling mud) companies.
The earliest fluid used for drilling oil wells was water. Various additives
controlled the loss of water to permeable formations exposed in the borehole.
As early as 1901, ground-formation material and surface clays were used to
impart viscosity and density in the Spindletop well drilled near Beaumont,
Texas. High density solids such as hematite and barite were applied in the
early 1920s to increase the drilling fluid density, prevent influx of formation
fluids, and control well blowouts. Drilling fluids were commercialized 1926.
Drilling fluid materials are also used as completion and workover fluids.
When the drilling reaches the oil producing zone, the subsequent operation is
referred to as completion. These fluids are formulated to enhance production
rates of crude oil and minimize permeability damage to the formation while dril-
ling through a production zone. Workover fluids are used on existing oil and gas
wells with the aim of improving or maintaining current production levels and
during remedial work on older wells. Some of the workover operations in oil
well drilling can include sand control measures, casing repair, reperforation,
252 DRILLING FLUIDS Vol. 1
Drilling fluids are classified as to the nature of the continuous phase: gas, water,
oil, or synthetic. Within each classification, there are divisions based on composi-
tion or chemistry of the fluid or the dispersed phase.
2.1. Gas-Based Muds. Gas-based drilling fluids are used mostly for
hard-rock drilling. These fluids range from compressed dry air or natural gas
to water-based mist and stable foams. Foam is considered gas based since the
gas comprises the bulk of the fluid volume. Bottomhole pressures imposed by a
gas-based fluid are low, and therefore, formation strengths must be relatively
high, having little or no influx or formation fluid. Air, gas, or mist drilling
requires a high annular velocity to remove drill cuttings. On the other hand,
large cuttings can be removed at low annular velocities when using stable foams.
Chemical additives for gas-based drilling fluids are generally biodegradable
mixtures of surfactants, certain polymers, and occasionally salts such as sodium
or potassium chloride. An aqueous solution of the additives is injected into the air
or gas flow to generate a mist or foam. No additives are used in dry air or gas
drilling operations. Gas-based fluids are not recirculated and materials are
added continuously. As the fluid exits the well, air or water vapor escapes to
the atmosphere, gas and oil are burned, and water and formation solids are col-
lected into a pit for later disposal. Stable foams must be destabilized to separate
the air from the liquid phase for disposal.
2.2. Water-Based Muds. The vast majority of all drilling fluids are
water-based systems. The types depend on the composition of the water phase
(pH, ionic content, etc), viscosity builders (clays or polymers), and rheological
control agents (deflocculants or dispersants).
Freshwater fluids can range from clear water having no additives to high
density muds containing clays, barite, and various organic additives. Onshore
wells typically use freshwater muds, as do some offshore wells where highly
weighted muds are needed. Freshwater muds may be operated at pH levels
ranging from 7 to 11. When drilling using clear water, small amounts of poly-
meric flocculants may be added to remove drill solids in a large settling pit in
order to maintain a clean fluid for fast drilling. When a viscous fluid is
required, clays or water-soluble polymers are added. Freshwater is ideal for
formulating stable drilling fluids as many mud additives are most effective
in a system of low ionic strength. Inorganic or organic additives control the
rheological behavior of the clays, particularly at elevated temperatures. An
organic polymer may be used for filtration control. Mud pH is generally alka-
line and, in fact, many viscosity control agents require an environment of
pH >9. Sodium hydroxide is by far the most widely used alkalinity control
Vol. 1 DRILLING FLUIDS 253
advantage of these systems is the high drilling rate that can be achieved because
of the lower colloidal solids content. These are normally applied in hard forma-
tions where increasing the penetration rate can reduce drilling costs significantly
and the tendency for solids buildup is minimal.
2.3. Oil-Based Muds. Oil-based drilling fluids have diesel or mineral oil
(14) as a continuous phase with both internal water and solid phases. Fluids hav-
ing no or very low water content are usually called oil-base muds or all oil muds;
fluids having higher water contents are called invert oil–emulsion muds, or sim-
ply inverts. Most oil muds maintain a fixed oil–water ratio depending on the
desired properties. Oil muds are employed for high angle wells where good lubri-
city is required, for high temperature wells where water-based systems may be
thermally unstable, for drilling water-sensitive shale formations, or where corro-
sive gases such as hydrogen sulfide and carbon dioxide may be encountered.
Environmental restrictions and cost often limit use, although higher drilling
rates achievable using oil muds and polycrystalline diamond compact (PDC)
bits can often offset the high fluid and disposal costs.
Oil-base muds have diesel or mineral oil as the continuous phase and either
are formulated using no internal aqueous phase or have only a minimal water
content (15). Organophilic clay or colloidal asphalt are used to control viscosity
and filtration rates. The internal water phase, either added as part of the formu-
lation or incorporated while drilling, is stabilized using emulsifiers. The desired
density is reached by adding a powdered high specific gravity solid. A wetting
agent ensures oil wetting of added or drilled formation solids.
Oil-based muds that incorporate an internal brine phase as an integral part
of the formulation are called inverts or invert oil–emulsion muds. The water con-
tent can be increased up to 50 vol% of the total liquid phase using the proper
emulsifiers and wetting agents. Suspension properties are achieved by the addi-
tion of organophilic clays. Colloidal solids such as oil-dispersible lignite, resins,
asphalt, or gilsonite impart filtration control. Increased density is achieved using
standard solid materials. The internal water phase is nearly always composed of
a concentrated sodium or calcium chloride brine to provide a sufficiently reduced
water activity to prevent osmotic uptake of drilling fluid water into water-
sensitive formations.
Oil muds are usually characterized by very low filtration rates and at one
time oil muds were not considered stable if other than minimal filtration rates
were observed in standard tests. This feature contributed to the low drilling
rates experienced for oil muds when using rock or natural diamond bits. Relax-
ing the normally tight filtration control specifications on oil muds can result in
higher drilling rates without loss of emulsion or mud stability (16). These relaxed
fluid-loss muds also are termed low colloid oil muds or relaxed filtration oil muds
because the higher filtration rates are achieved by omitting some of the colloidal
solids from the formulation and reducing the concentration of emulsifiers and
surfactants in the fluid. Most relaxed fluid-loss oil muds contain 5–15 vol%
brine, depending on the mud density. In most other respects these fluids are
similar to a conventional invert oil–emulsion mud and can easily be converted
to an invert. Use of relaxed filtration oil muds has dropped considerably owing
to improvements in bit design. These improvements allow high drilling rates
using the more conventional low filtration-rate oil muds.
Vol. 1 DRILLING FLUIDS 255
The synthetic-based muds (drilling fluids) are a relatively new class of dril-
ling muds that are particularly useful for deepwater and deviated hole drilling.
These muds were developed to provide an environmentally superior alternative
to oil-based drilling muds as well as offering an alternative to the high costs asso-
ciated with the disposal of drill cuttings generated when diesel or mineral oil-
based muds are used. The synthetic muds have a continuous phase that consists
of synthetic organic liquid (17). The synthetic-based liquids, because of their
similarity with the oil-based muds, are often called pseudo-oil muds outside
the United States. While the technical and economic benefits of using such fluids
in the deepwater Gulf of Mexico are recognized, the environmental impacts of the
discharged drill cuttings produced are not well understood. While the drill cut-
tings themselves are considered toxicologically inert, there is concern that any
adhering drilling fluid ingredients may be toxic, particularly if the cuttings are
produced during drilling with oil-based or synthetic-based fluids. The high cost
for the synthetic-based fluids can be offset by the enhanced drilling performance
for fast drilling achieved when using them, and the on-site discharge of the SBFs,
when this is permitted. The synthetic-based fluids (SBF) introduced at the start
of the 1990s may be classified into four general categories (1) Synthetic hydrocar-
bons, eg, polymerized olefins that include poly alpha olefins (PAO), linear alpha
olefins (LAO), and internal olefins (IO) (2) ethers (3) esters and (4) acetals. Dur-
ing the second half of the 1990s, the groups of synthetic drilling fluids that began
to achieve enhanced popularity were the LAOs, IOs, and linear paraffins (LPs),
which were often referred to as second generation SBFS (18). These second gen-
eration SBFs were found to have benefits over their predecessors in that they
had lower kinematic viscosities and were less expensive. The current concerns
with toxicity, biodegradation rates, environmental impacts and costs have vir-
tually eliminated the use of synthetic-based fluids except for olefins and esters
in offshore drilling. The environmental aspects of the SBFs are discussed more
fully in Section 7.
3. Properties
where P ¼ the hydrostatic pressure in kPa (psi); rm ¼ the drilling fluid density,
kg/m3 (lb/gal); and L ¼ the column length or well depth, m(ft). Wells with bottom-
hole pressures of 100 MPa (14,500 psi) at 5000 m (16,400 ft) are not unusual.
Such a well would require a mud density >2040 kg/m3 (17 lb/gal) to prevent for-
mation-fluid influx. At least 2 MPa (290 psi) above formation pressures normally
is recommended to ensure well control. Bottomhole pressures imposed by the
mud/clear brine column that are higher than necessary can reduce drilling
rates and induce fractures in exposed formations, allowing loss of drilling fluid
to the fractures (lost circulation). Operational safety can be adversely affected
if lost circulation occurs.
3.2. Flow Properties. The fluid viscosity and annular flow velocity must
be high enough to remove cuttings generated by the drill bit and other formation
material that may fall into the wellbore. These solids are carried up the annulus
to the surface where they are separated with varying degrees of efficiency and
then disposed of. In order to accomplish this, low viscosity drilling fluids are cir-
culated at high flow rates or high viscosity fluids at low flow rates. In addition,
for maximum drilling rate, a low effective viscosity is desired at shear rates
generated through the bit nozzles (10,000–100,000/s). The varying demands on
the flow properties are best met by fluids exhibiting non-Newtonian rheological
characteristics. Drilling fluids are normally shear thinning, having apparent
viscosity decreasing with increasing shear rate.
The gel strength (thixotropy) is a measure of the capability of a drilling fluid
to hold particles in suspension after the mud circulation ceases. It results from
the electrical charges on the individual clay platelets. The positively charged
edges of a platelet are attracted to the negatively charged flat surfaces of
adjacent platelets. In a bentonite mud in which the particles are completely
dispersed, the bonds between particles are broken while the mud is flowing.
When the mud pump is shut off and flow ceases, the attraction between the
Vol. 1 DRILLING FLUIDS 257
The pressure exerted by the column of drilling fluid in the well balances forma-
tion pressures to prevent uncontrolled influx of formation fluids that may result
in a blowout. The mud density must be controlled accurately by suitable weight-
ing materials that do not adversely affect the other properties. Most important is
the specific gravity of the weighting agent as well as its insolubility in water and
its chemical inertness. The weighting material should be ground to the preferred
particle-size distribution and be relatively nonabrasive. Various finely ground,
solid weighting materials that have found application in drilling are listed in
Table 1. As of 1995, little weighting material other than barite was used in the
United States.
Barite, which is predominately composed of BaSO4, meets the overall
requirements for weighting material better than other materials and is used
for increasing the density of drilling fluids throughout the world. Barite in its
pure form has a density of 4500 kg/m3. Commercial barite has a lower specific
gravity than pure barium sulfate owing to the presence of associated minerals,
such as silica. Barite is virtually insoluble in water and does not react with other
mud constituents. Most operators prefer barite that meets API specifications
(Table 2) (21). The barite content in mud depends on the desired density but it
can be as high as 2000 kg/km3 (700 lb/bbl).
Drilling technology favors systems having the lowest possible solids content
because of the resulting higher penetration rate, easier control of mud proper-
ties, and fewer problems experienced during drilling. Alternative weighting
materials having specific gravities higher than that of barite offer this advan-
tage. The only other high specific gravity material that is used to any degree
as of 1995 is hematite (iron oxide).
The specifications for drilling fluid hematite have been set by the API and
are listed in Table 2 (22). Hematite is used most frequently in high density oil-
based muds to minimize the total volume percent solids (23). The abrasivity of
hematite limits its utility in water-based muds.
Calcite and siderite (24) are used occasionally because of their solubility in
hydrochloric acid, which offers a method of removing mud filter cake deposited
on productive formations. The calcite (calcium carbonate) is supplied in the form
of ground marble, with the particle size adjusted to the porosity of the reservoir,
Calcite and siderite are used most frequently in workover or completion fluids
when a nondamaging fluid is required, ie, one that can be removed by acidizing
at a later time.
Solid salt, ground and packaged in several particle size grades, can be used
in saturated salt brines to increase the fluid density (25). However, sized salt is
also as a weighting agent, having specific use in sensitive reservoirs. In such
cases, an oversaturated solution is used to form a filter cake that can easily be
removed with fresh water flushing. This nondamaging has found increasing
use since the early 1990s. At one time the sized salt systems were used primarily
for completion or workover operations, but use has also increased as in drill-in
fluids for horizontal wells (26).
The use of solids-free fluids or clear brine fluids (CBF) is occasionally
resorted to for achieving high drilling rates, and also in workover and completion
operations, for avoiding damage arising from particulate invasion of the produc-
tive formations. Clear brine fluids and their additives are discussed more exten-
sively in the Section 5, Completion and Workover Fluids.
The chemical and mechanical dispersion of the drilled solids tends to
increase the percentage of small-sized solids in a mud as drilling progresses.
260 DRILLING FLUIDS Vol. 1
a viscosity builder and suspending agent and can be used in almost any type of
water (43). Although xanthan gum solutions support bacterial growth, a preser-
vative is usually not needed unless the solution is to be stored. Because of sus-
pending ability at low concentrations and in electrolyte solutions, xanthan gum
is widely used for drilling, workover, and completion fluids. Two other biopoly-
mers, succinoglucan gum [39464-87-4] (44) and welan gum [96949-22-3]
(42,45), are also finding some use in drilling fluids at concentrations similar to
xanthan gum.
Guar gum is a nonionic, branched-chain polysaccharide, a galactomannan
that is usually hydroxypropylated for use in drilling (46). It produces viscous
solutions in fresh or saltwater at concentrations of 3–6 kg/m3 (1–2 lb/bbl). It
is used in solids-free and low solids muds and degrades rapidly above 808C, limit-
ing its use to shallow wells.
High molecular weight polyacrylamides are used as viscosity builders in
freshwater muds (47) or as bentonite extenders. The ionic nature of the polyacryl-
amide may range from nonionic to anionic (30% hydrolyzed) depending on the
situation.
Occasionally polymers are used to increase the viscosity of oil-base and
synthetic-base muds. The polymers for this use are typically sulfonated polystyr-
enes or ethylene–propylene terpolymers (EPDM) (48,49). Such polymers are
usually used in conjunction with an organophilic clay.
4.2. Viscosity Reduction. Proper control of viscosity and gel strengths
is essential for efficient cleaning of the borehole, suspension of weight material
and cuttings when circulation is interrupted, and to minimize circulating
pressure losses and swab/surge pressures owing to axial movement of the
drill string. Viscosity may be increased as previously indicated, but there is
often the necessity of reducing the viscosity. A reduced viscosity can be achieved
by thinning or deflocculating clay–water suspensions. Thinning is measured as
a reduction of plastic viscosity, yield point, or gel strength, or a combination of
these properties. Viscosity is reduced by decreasing the solids content and the
number of particles per unit volume, or by neutralizing the attractive forces
between particles. Although mud thinners or viscosity-reducing chemicals are
added to reduce flow resistance and gel development, reduction of filtration
rates and filter cake thickness and stabilization of mud properties at elevated
temperatures often results. Typical mud-thinning chemicals are polyanionic
materials that are adsorbed on positive edge sites of the clay particles, thereby
reducing the attractive forces between the particles without affecting clay
hydration (27,50).
Vol. 1 DRILLING FLUIDS 265
organic polymers and lignites (see Table 5). Cost and viscosity-building charac-
teristics should be considered, particularly in weighted high solids muds.
Starches, used first in the late 1930s for filtration control, are still in use in
the 1990s. Cornstarch is most commonly used in the United States. Potato starch
is also used in the United States but primarily in Europe and elsewhere. Both
corn and potato starches are pregelatinized before dispersion in water (63).
The API specifications for drilling fluid starch are listed in Ref. 64.
Starch is subject to fermentation by many microorganisms and, unless the
mud is saturated with salt or the pH is 11.5, a preservative or biocide must be
added if the mud is to be used for an extended period of time. The most common
biocide until the mid-1980s was paraformaldehyde [9002-81-7]. This material
has been largely replaced by isothiazolones (at 5–10 ppm conc) (65), carbamates,
and glutaraldehyde [111-30-8]. Alternatively, the biocide may be incorporated
during the processing of the starch and is present in the commercial product.
Starch is degraded by heat and agitation. Under continued circulation at
958C, starch can breakdown rapidly. Its performance also is affected by pH
changes in the presence of calcium and magnesium ions. Starch coprecipitates
with calcium when caustic soda is added to mud containing dissolved calcium
salts. Starch molecules are hydrolyzed at very low or high pH in the presence
of calcium and magnesium >1108C. An acidic environment also may prevent
hydration and thus limit the effectiveness of starch as a filtration control agent.
Numerous modifications and derivatives of starch have been made for
application in drilling and workover fluids (66). Most modified starches are
cross-linked to some degree to improve thermal stability. Carboxymethyl and
hydroxypropyl starches are finding increasing application in drilling fluids.
There are some applications involving cationic starches as well, although toxicity
must be considered for offshore use.
Carboxymethyl cellulose (CMC) and polyanionic cellulose (PAC) are avail-
able in several viscosity grades and high and low purity grades. All grades can be
effective filtration control agents depending on the well conditions. The effective-
ness of CMC in reducing filtration decreases with increasing salt concentration.
A polyanionic cellulose (PAC) polymer is available that is designed particularly
for application in muds of high salt concentration (37,67). Most CMC and PAC
materials are sodium salts but there are also potassium versions to take advan-
tage of the shale stabilizing properties of this ion (68). While not designed to con-
trol filtration, HEC may be effective as a filtration control agent in combination
with other organic polymers in waters having salinities up to saturation.
Acrylate and acrylamide polymers have several uses in drilling fluids, one
of which is for filtration control. Sodium polyacrylates [9003-04-7] having mole-
cular weights near 250,000 are excellent temperature-stable filtration control
agents for both fresh- and saltwater muds, provided the concentration of
water-soluble calcium is <400 mg/L (69). The calcium ions are precipitated
using a carbonate such as soda ash, before adding the polyacrylate.
Lignite products, mined, ground, and possibly treated with sodium or potas-
sium hydroxide, are economical filtration control additives for some water-based
muds, in addition to improving flow properties (70). A sulfonated lignite
complexed with a sulfonated phenolic resin is an effective high temperature,
filtration control additive, for both fresh- and seawater muds that contain high
268 DRILLING FLUIDS Vol. 1
concentrations of soluble calcium, and does not affect viscosity (71,72). A high
molecular weight polyanionic lignin has also found application for high tempera-
ture muds with a high electrolyte content (73). These and similar products can
provide filtration control for high density, high solids muds above 1208C.
A number of synthetic polymers having the ability to control filtration rates
at high temperature and in the presence of calcium and magnesium have also
been developed (74). Such materials include vinyl sulfonate–vinyl amide copoly-
mers (75,76), a copolymer of AMPS and N,N-dialkyl (meth) acrylamide (77) and a
sulfonated hydroxylated polymer (78).
Filtration control in oil- and synthetic-base fluids is achieved by the emul-
sified aqueous phase, by the emulsifier package, and by additions of powdered
solid materials. The powdered solids used for this purpose consist of asphalt
[8052-42-4], gilsonite [12002-43-6], and amine-treated lignite. The choice of addi-
tive depends on the nature of the base fluid, the emulsifier package, and most
importantly the downhole temperature. Styrene–butadiene copolymers in the
form of aqueous dispersions have also found application as filtration control
agents in oil-base muds.
Cellulosic fibers, powdered limestone, gilsonite, and asphalt are frequently
added to both water and oil muds at levels of 10–25 kg/m3 (4–10 lb/bbl) when
high differential pressures are encountered to control seepage losses to the for-
mation. This treatment also is used to improve the quality of the mud filter cake
to reduce the chance of differential pressure sticking.
4.3. Alkalinity Control. Water-base drilling fluids are generally main-
tained at an alkaline pH. Most mud additives require a basic environment to
function properly and corrosion is reduced at elevated pH. The primary additive
for pH control is sodium hydroxide [1310-73-2].
The second most common alkalinity control agent is lime [1305-78-8], nor-
mally in the form of calcium hydroxide [1305-62-0], used in both water and oil
muds. In the latter, the lime reacts with added emulsifiers and fatty acids to sta-
bilize water-in-oil emulsions. Lime is used in brine systems containing substan-
tial quantities of soluble calcium and in high pH lime muds.
Potassium hydroxide [1310-58-3] is occasionally used for alkalinity control.
This is particularly true for some polymer and lime muds where a low sodium
level is desired. The potassium level of such muds is quite low but has been
attributed by some to provide stability to water-sensitive shale formations (61,79).
A fourth alkalinity control additive is magnesium oxide [1309-48-4], which
is used in clay-free polymer-base fluids (41). Magnesium oxide provides an alka-
line environment and, as it is only slightly soluble, also has a buffering effect. It
enhances the thermal stability of polymer solutions by preventing a pH decrease
to neutral or slightly acidic conditions at elevated temperatures. It is mainly
applied in completion or workover operations where clay-free acid-soluble fluids
are desired. There are occasions where the mud pH must be lowered such as
after drilling fresh cement or overtreatment by one of the alkaline materials dis-
cussed. Organic acids that have been used for this purpose include acetic acid
[64-19-7], citric acid [77-92-9], and oxalic acid [144-62-7], but these materials
have been employed only infrequently. Inorganic additives used to lower pH
levels include sodium bicarbonate [144-55-8] and sodium acid pyrophosphate
[7758-16-9] (SAPP). Of the two, sodium bicarbonate is used the most by far.
Vol. 1 DRILLING FLUIDS 269
Defoamers are frequently needed for salty muds, although in very small
quantities. A common defoamer is aluminum stearate [637-12-7]. Others include
tributyl phosphate [55612-35-6], alkyl aryl sulfonates, silicones, and alcohols
such as octanol [111-87-5].
Foaming agents maintain stable drilling foams in areas where minimal bot-
tomhole pressures are required. A large number of chemicals generate drilling
foams, including anionics, such as alkyl and alkylaryl sulfates and alkylarylsul-
fonates; nonionics, such as ethoxylated fatty alcohols and alkylaryl alcohols; and
cationics, such as imidazolines and tertiary amines. The foaming agent must be
chosen to handle a variety of possible contaminants (salt, crude oils, solids) and
downhole temperatures (113,114).
4.7. Lost Circulation Control. To function properly, a drilling fluid
must be circulated through the well and back to the surface. Occasionally, highly
permeable or cavernous formations and fractured zones, both natural and
induced by the mud pressure, are encountered and circulation is partially or
completely lost. Loss of drilling fluid, owing to openings in the formation, can
result in loss of hydrostatic pressure at the bottom of the hole and allow influx
of formation fluids and possibly loss of well control. It is essential that circulation
be regained for drilling to continue. A wide variety of materials can be added to
the drilling fluid to seal off the lost circulation zones (115,116). The particle sizes
of these materials are typically much larger than the particle sizes of solids nor-
mally suspended in the mud which are generally <150 mm in diameter. Some of
the same materials may be used for controlling filtration rates or for stabilizing
shale formations but in such cases would be much smaller in particle size than a
typical lost circulation material (LCM).
Lost circulation materials are flake, fiber, or granular-shaped particles.
Each type is sold individually, often in two or more size grades, or two or more
materials of different shapes may be sold as a blend (117–120). Materials of dif-
ferent shapes and sizes are often blended into the mud at the well site. Some
common flake-shaped LCMs consist of shredded cellophane and paper, mica,
rice hulls, cottonseed hulls, or laminated plastic. These materials lie flat across
the opening to be sealed or are wedged into an opening such as a fracture. Some
are sufficiently strong to withstand considerable differential pressure, whereas
others are weak and the seal may be broken easily. Weaker flake materials typi-
cally are used near the surface or in combination with fibrous or granular addi-
tives. Fibrous additives include a variety of cellulose fibers, sawdust, sugar-cane
bagasse, paper, straw, leather, and many others of similar size, shape, and avail-
ability. The larger fibers function by forming a brush-heap-type mat over the
opening. The seal so formed may require smaller fibrous particles to stop seepage
of mud through the mat. Fibers generally have little strength and cannot with-
stand high differential pressures. The brush-heap seal may extend far enough
into the wellbore to be dislodged by the drill string, or be rigid enough to interfere
with drill string movement.
Granular LCMs generally are much stronger than the other types and
include ground rubber, nylon, plastics, limestone, gilsonite, asphalt, and ground
nut shells, eg, walnut and pecan. Fine-, medium-, and coarse-size grades are
available. Granular-shaped particles enter the opening, bridge it, and form
a tight seal against further mud losses. Particle size and distribution are
Vol. 1 DRILLING FLUIDS 273
blended with a glycol-carrier are used to prevent pipe sticking and to act as lubri-
cants for water-base muds (128,129).
If the drill string becomes differentially stuck, mechanical methods or spot-
ting fluids can be applied, or the hydrostatic pressure can be reduced (130). In
general, penetration of water- or oil-based spotting fluids into the interface
between the filter cake and the pipe accompanied by dehydration and cracking
results in reduction of differential pressure across the drill string (130,131).
Spotting fluids are usually positioned in the open hole to completely cover the
problem area.
Recently, a number of water-base spotting fluids have been developed
(132) and consist of a variety of surfactants, glycols (133,134), organic acids
(135), glyceride esters (136), and vegetable oil derivatives. The newer water-
base spotting fluids have the distinct advantage that they can be incorporated
into the drilling fluid following use, without having an adverse effect on the
toxicity of the system (137).
Drill string and casing corrosion can present serious problems in some dril-
ling operations. Corrosion is generally caused by oxygen dissolved or entrained
in the mud as it is circulated through the well (138). Acid gases, such as carbon
dioxide and hydrogen sulfide, contribute to corrosion, particularly hydrogen sul-
fide which can produce catastrophic drill string failures through stress cracking.
Preventing entry of these corrosive gases into the well is the most effective
method of control, but is not always feasible. Therefore additives are employed
to counteract corrosive attack (139,140).
Oxygen corrosion is reduced significantly at pH >11 and is generally
lower at any alkaline pH than in an acidic environment. Maintaining a high
pH is the most common means of corrosion control. Other methods include oxy-
gen removal by scavengers such as sodium sulfite [7757-83-7] or ammonium
bisulfite [17026-44-7] (140,141), or protecting the pipe from attack by coating
with amines.
A high pH also affords protection against hydrogen sulfide and carbon diox-
ide. Zinc and iron compounds are used as sulfide scavengers (142). Zinc com-
pounds include the carbonate, oxide, hydroxide, sulfate, and water-soluble
organic chelates. High surface area iron oxides, particularly a synthetic magne-
tite, also are used (143). Scale inhibitors such as organic phosphonates and low
molecular weight polyacrylates may be added to prevent buildup of carbonate
scale on the pipe. Drilling rates can be severely reduced when drilling formations
that tend to stick to the surface of the drill bit, a situation known as bit balling. If
the bit cannot be kept clean and free of sticky material, drilling performance suf-
fers and the operational cost can increase substantially. The new generation syn-
thetic polycrystalline diamond compact (PDC) bits are capable of achieving high
penetration rates if the cutters are kept clean and the proper hydraulics and bit
rotary speeds can be maintained. Use of an oil- or a synthetic-base mud is nor-
mally preferred when drilling with a PDC bit as there is much less tendency for
bit balling to occur and high drilling rates can be realized (144). Water-base
muds are more prone to bit balling and PDC bit performance can be significantly
poorer than when a nonaqueous continuous-phase mud is used (145). Additives
being used to enhance the drilling performance of PDC bits in water-base muds
include water-soluble glycols (146,147) and various terpenes (144).
Vol. 1 DRILLING FLUIDS 275
Completion and workover fluids that are used during the completion phase of the
well or when performing a workover during the life of the well. Completion fluids
should (1) cause minimal formation damage, especially to reservoir permeability,
(2) maintain long-term stability at downhole temperatures, (3) be resistant to
contaminants, and (4) protect down-hole equipment against corrosion. These
fluid types can divided into two broad classifications: solids-free and solids-
laden. The solids-free completion/workover fluids are generally brines ranging
in composition from potassium chloride to zinc bromide, and their densities fall
into the range of 1.02 kg/L (8.5 lb/gal or ppg) to 2.46 kg/L (20.5 ppg). Weighted
solids-free fluids are obtained by the use of solutions of various water-soluble or
acid-soluble salts. In particular, the chlorides and bromides of sodium, potas-
sium, calcium, and zinc, and their mixtures are used to adjust the densities of
these brines. The solids-laden completion and workover fluids are generally com-
posed of a brine as a base fluid and solids, in the form of water or acid-soluble
particles, for density and fluid-loss control as well as polymers for suspension-
carrying capacity and filtrate control. In general, the solids-laden fluids cover
the same density range as the solids-free fluids.
Solids-free fluids are the predominant completion/workover fluids in use.
Use of these fluids isdictated in many completion operations, especially in gravel
pack or frac pack type completions. Use of a solids-laden fluid in such operations
can lead to failure. Solids-free fluids cause a minimum amount of formation
damage and allow tools to be run into and pulled out of the well with ease. Solids
can penetrate into the permeable media of the formation and become lodged,
thus causing reduced permeability and ultimately lower production. Solids-
laden fluids also have typically increased viscosities, which can inhibit tools
being run into and pulled out from the well. In order to keep solids-free fluids
free of solids, filtration, either cartridge or diatomaceous earth, is generally
used while these fluids are being employed.
There are a number of factors that determine the choice of fluid to be used.
If there is no milling or drilling operations to be carried out, then a solids-free
fluid can be used. The particular solids-free fluid/soluble salt to be employed
also depends on a number of factors that include:
(1) Density required: A compilation of some of the soluble salts and their max-
imum brine densities that have been used as completion and workover
fluids is given in Table 8. A specific density may be achieved by mixing to-
gether one or more brines listed in the table. The monovalent brines are
generally mixed with monovalent brines and divalent brines mixed with
divalent brines. Mixing of monovalent and divalent brines can cause the
formation of precipitates and turbidity. Density adjustment can also be
achieved by adding a salt to a specific brine.
(2) True crystallization temperature (TCT) required: It is important to know
expected ambient temperatures and expected sea floor temperatures. The
crystallization temperature can be adjusted by blending brines or water
to raise or lower the crystallization temperature.
276 DRILLING FLUIDS Vol. 1
Table 8. Soluble Salts that Increase the Density of Drilling Mud and Workover
and Completion Fluids
CAS Registry Specific gravity Maximum brine
Salt number Formula of solid density, g/L (ppg)
potassium chloride [7447-40-7] KCl 1.984 1.14 (9.5)
sodium chloride [7647-14-5] NaCl 2.165 1.20 (10.0)
calcium chloride [10043-52-4] CaCl2 2.150 1.39 (11.6)
sodium bromide [7647-15-6] NaBr 3.203 1.50 (12.5)
calcium bromide [7789-41-5] CaBr2 3.353 1.70 (14.2)
zinc bromide [7699-45-8] ZnBr2 4.201 2.46 (20.5)
potassium formate [590-29-4] KCHO2 3.48 1.60 (13.3)
sodium formate [141-53-7] NaCHO2 1.92 1.34 (11.2)
cesium formate [3495-36-1] CsCHO2 H2O 3.372 2.37 (19.7)
(3) Compatibility with the formation: Formation gases may not be compatible
with the completion/workover fluid. For example, calcium-based brines
should not be used in the presence of carbon dioxide as the formation of in-
soluble calcium carbonate can precipitate. Other problems can be that the
formation water may be incompatible with brine, or emulsions may form
with the crude oil.
(4) Cost: Generally speaking, the higher the density of the brine, the higher
its cost. Another rule of thumb is the lower the crystallization temperature
required, the higher will be its cost. Other important factors to weigh are
brine corrosivity (eg, metallurgy of the casing and/or tubing in the well),
brine stability, brine toxicity and brine availability (to remote and generally
international locations).
Brine sources may include seawater, natural brines and manufactured salt
brines. Densities up to 2.3 kg/L (19.2 ppg) can be obtained using mixtures of cal-
cium bromide and zinc bromide. Calcium bromide is commercially available as a
52 wt% solution in water, having a density of 1.704 kg/L (14.2 ppg). This brine
remains clear down to a temperature of 78C. Also available is a calcium bro-
mide/zinc bromide brine having a density of 2.304 kg/L (19.2 ppg), where the
brine contains approximately 52.8 wt% zinc bromide and 22.8 wt% calcium bro-
mide. Evey manufacturer of clear brines provides detailed blending tables show-
ing the relative amount of various brines, salts, and water in order to prepare a
brine with the desired specific gravity and crystallization point. Two examples
are given below. Table 9 provides the volume of water [fraction of a barrel
(bbl)] and weights of 95% calcium bromide solid that are needed to prepare 1 bar-
rel (bbl) of a calcium bromide brine having a specific brine density (ppg). The
crystallization point of the brine is also included in the table. Table 10 presents
the volumes of calcium bromide (1.7 kg/L /14.2 ppg) brine and zinc bromide/
calcium bromide (2.3 kg/L /19.2 ppg) that are needed for preparing calcium
bromide/zinc bromide brines of varying densities, together with their correspond-
ing crystallization points.
Slightly higher densities than the above-mentioned brines can be attained
using nearly saturated solutions of ZnBr2 having a density of 2.46 kg/L (20.5 ppg)
Vol. 1 DRILLING FLUIDS 277
or of cesium formate having a density of 2.367 kg/L (19.7 ppg) (148). The mono-
valent formate brines, which are highly soluble in water, were first introduced
into the oil industry in the early 1990s. The formates are organic and biodegrad-
able. Their chemistry stabilizes and protects the conventional polymers used in
viscosity modifiers and other additives, allowing such additives to be used at
much higher temperatures—often >3508F (1768C). Other relevant information
on formate brines is included in a series of papers that are cited in Refs. 149–154.
Table 10. Calcium Bromide/Zinc Bromide Solution Requirements Using 14.2 ppg
Calcium Bromide and 19.2 ppg Zinc Bromide/Calcium Bromidea
Drilling fluids made both with CaBr2/ZnBr2 brines and with formate-based
fluids generally need no solid weighting agents, since the density is an inherent
property of the fluid itself.
One inherent problem with solids-free brines is that since they have no
solids and in general have low viscosities, there is nothing to prevent the liquid
from leaking off into the formation. Therefore, an important part of planning in
the use of the solids-free completion and workover fluids is fluid loss control pills.
These pills are generally a blend of the brine being used and polymers for impart-
ing viscosity. The viscosifiers used in brine fluids should be stable at a tempera-
ture of 1508C and have good acid solubility. The polymers used are
hydroxyethylcellulose, xanthan gum and other specialty polymers. Guar gums
can be included in completion/workover fluids, but their use is more limited
because they are not completely soluble in acid; in addition, guar gum is also
only stable to 668C. For filtrate loss reduction, a blend of polymers and soluble
solids are usually incorporated into the fluid. These polymers are used to plug
the interparticle spaces within the bridging solids, thus reducing the amount
of filtrate entering the formation. The selection of the solids to be used will be
based on the pore throat openings in the formation in order that a bridge can
be rapidly formed on the outside of the formation, preventing solids and liquids
from entering the reservoir.
Some additional details about the additives used in workover/completion
fluids are listed in Table 11. Some solids-laden completion and workover fluids
are now being utilized as reservoir drill-in fluids. These are used once the well
6. Economic Aspects
Considerable changes have occurred in the makeup of drilling fluids during the
the last two decades and these changes are expected to continue during the
tweny first century. The price of the drilling fluid additives used in these fluids
varies according to company and location so that it is almost impossible to indi-
cate typical prices for the more common additives (used in drilling operations),
which are listed in Table 12. Price information on the full line of products nor-
mally is available from drilling fluids companies. The prices for those materials
listed in Table 12 would also depend on the nature of the contract between the
supplier and operating companies.
It is instructive to consider the price fluctuations that have occurred for bar-
ite, a densifier that constitutes the largest of the chemical expenses incurred in
drilling muds. Market competition has been significant in the barite segment, so
that its prices have been historically under pressure. The oil price crash and dril-
ling recession in 1998 resulted in the price of barites in Yarmouth (U.K.) falling
from an average of $100–120 per metric ton in 1997 to $75–90 per metric ton in
mid-1998. The price decrease continued in 1999 but started to increase in mid-
2000 to $80–100 per metric ton at the end of 2000.
8. Environmental Aspects
At present, the major challenge that must be met in formulating drilling fluids is
need to satisfy the increasingly demanding conditions of high temperature and
pressure that are found in some deep wells and horizontal wells, while avoiding
harm to the environment. Drilling fluid components should be selected in a way
such that the discharge of mud or cuttings will have a minimum possible envir-
onmental impact. In addition to environmental concerns being a major driving
force behind current drilling fluid R&D, the health of rig workers has become
an important concern in the development and use of new products.
While drilling fluids are essential for the success of an well drill, they can
also be one of messier parts of a drilling operation. Cuttings brought up from out
of a borehole have to be disposed of as well as any drilling fluid attached to them.
The degree of impact that drilling fluids can have on the environment depends on
the type of mud used and the existing environmental conditions. Offshore, water-
based muds are generally the least damaging when compared with oil-based
muds. The disposal of waste drilling fluids and drill cuttings in the United States
has long been regulated either by local authorities, the individual states, or by
the federal government. These regulations continue to change. In many
instances, it is environmental issues that have been the main driving force for
product choice in the drilling fluids business. It is no secret that the mud-
contaminated cuttings constitute a major portion of the waste that is generated
during drilling.
The offshore disposal of both diesel and mineral oil drilling fluids and asso-
ciated cuttings has always been prohibited in U.S. waters. However, offshore dis-
charge of mineral oil mud cuttings overboard has been permitted in the North
Sea and elsewhere as long as the oil content of the cuttings was below some reg-
ulatory limit. The regulatory oil-on-cuttings limit in some sectors of the North
Sea has been lowered significantly over the last decade. There is a definite
move toward alternative fluid systems many of which are used in U.S. offshore
areas, eg, synthetic-based drilling muds. Thus, in the 1990s, due to environmental
Vol. 1 DRILLING FLUIDS 281
metal carboxylate was selected from sodium, potassium, and cesium salts of
acetic and formic acids.
A 1997 patent (173) described a composition that is useful for viscosifying
aqueous drilling fluids that are utilized in subterranean drilling activities such
as drilling, drill-in, completion, hydraulic fracturing, work-over, packer and well
treating fluids. The composition is composed of a prehydrating alcohol having at
least two hydroxyl groups (with a molecular weight of 60 to 600), an inorganic
salt, and a polymer. The prehydrating alcohol reacts physically with the polymer
to cause a partial swelling effect, solvating effect or hydrating effect upon the
polymer.
A Great Lakes Chemical patent (174) presented aqueous mixed salt
systems that were viscosified with water-soluble or water-dispersable polymers.
These systems were comprised of water, a water-soluble or water-dispersable
polymers capable of viscosifying an aqueous media, and mixed salts comprised
of at least two cations and two anions. The cations cited were Li, Na, K, Cs,
Mg, Ca, Zn, or mixtures. The anions cited were Cl, Br, I, formate, nitrate, acet-
ate, cyanate, thiocyanate, or a Zn-complex anion. These compositions were pro-
posed as replacements for conventional drilling, drill-in, completion, hydraulic
fracturing, work-over, packer, well-treating, testing, spacer, or hole abandon-
ment fluids.
A patent (175) disclosed a low solid, high density fluid that was suggested as
being useful in drilling, workover and completion fluids in oil and gas well opera-
tions. The fluids comprised (1) one or more brines having a density of 1500 kg/m3
to 2400 kg/m3 (salt chosen from calcium bromide, zinc bromide or mixtures of
these), (2) a viscosifier that is a high surface area silica, preferably a fumed silica,
(3) a fluid loss control additive that was preferably selected from the group con-
sisting of starch derivatives, cellulose derivatives, lignocellulose derivatives or
mixtures of these; (4) one or more water soluble or acid soluble bridging agents,
and (5) optionally, an alkaline buffer and a polar additive.
A 1998 patent (176) described a drilling fluid comprised of brine and addi-
tives for maintaining rheological properties in a neutral or slightly acidic brine
and a method for using the drilling fluid to drill into a producing formation. The
drilling fluid was comprised of a viscosifying polymer and a fluid loss polymer
that function in the brine without substantially adversely affecting their proper-
ties. The brine was virtually free of insoluble solids.
A recent review (177) was published on the use of complex carbohydrates,
ie, polysaccharides as additives in the petroleum and natural gas industries,
especially as viscosifiers and fluid loss control additives for drilling fluids and
muds. Viscosifiers are typically xanthan gum, hydroxyethyl cellulose and modi-
fied guar gum, while fluid loss control additives are typically pregelatinized
starch, and polyanionic cellulose derivatives, especially CM-cellulose.
9.4. Densifiers. The invention (178) presented in another patent relates
to an additive which increases the density of the well bore fluids used during the
construction or repair of oil, gas, injection, water or thermal wells. The additive
is comprised of solid, colloidal particles of weight average particle diameter (D50)
of < 2 m. The particles have been deflocculated by the action of a dispersant, pre-
ferably incorporated during the grinding or comminution process. The colloidal
particles used include barium sulfate, calcium carbonate,dolomite, illmenite, or
286 DRILLING FLUIDS Vol. 1
BIBLIOGRAPHY
‘‘Drilling Fluids’’ in ECT 1st ed., Suppl. 1, pp. 246–258, by G. R. Gray and E. E. Huebotter,
Baroid Div., National Lead Co.; in ECT 2nd ed., Vol. 7, pp. 287–307, by E. E. Huebotter
and G. R. Gray, Baroid Div., National Lead Co.; ‘‘Drilling Fluids’’ under ‘‘Petroleum’’ in
ECT 3rd ed., Vol. 17, pp. 143–167, by R. K. Clark and J. J. Nahm, Shell Development
Co; in ECT 4th ed., Vol. 18, pp. 370–405, by R. K. Clark, Shell C&P Technology Company;
‘‘Drilling Fluids’’ in ECT (online), posting date: December 4, 2000, by R. K. Clark, Shell
E&P Technology Company.
CITED PUBLICATIONS
161. PCT Int. Appl. WO 99/23188 (May 14, 1999), D. P. Vollmer and R. L. Horton
(to Great Lakes Chemical).
162. PCT Int. Appl. WO 98/40446a (Sep. 17, 1998), W. S. Halliday, D. K. Clapper, and
M. R. Smalling (to Baker Hughes Inc.).
163. Eur. Pat. Appl. EP 1074598 (Feb. 7, 2001), J. W. Dobson, K. O. Tresco and J. Lay
(to Texas United Chemical Co.LLC).
164. PCT Int. Appl. WO 98/08914 (March 5, 1998), J. A. C. Van Ooyen (to Verdugt B.V.).
165. U.S. Pat. 5,547,925 (Aug. 20, 1996), W. M. Duncan (to Integrated Industries Inc.).
166. PCT Int. Appl. WO 99/48994 (Sep. 30, 1999), C.-A. Sjogreen and B. Sallmen (to
Perstorp AB).
167. Eur. Pat. Appl. EP 1103589 (Aug. 16, 2001), S. Carminati, A. Guarneri, and
M. Brignoli.
168. Eur. Pat. Appl. EP 0764709 (March 26, 1997), A. D. Patel (to M-I Drilling Fluids).
169. PCT Int. Appl. WO 97/10313 (March 20, 1997), L. Lee and co-workers (to M-I Drilling
Fluids).
170. PCT Int. Appl. WO 97/34963 (Sep. 25, 1997), R. J. Wittenbrink, C. J. Mart, and D. F.
Ryan (to Exxon Research and Engineering Co.).
171. U.S. Pat. 2003/203821 (Oct. 30, 2003), K. B. Fox.
172. PCT Int. Appl. WO 00/71635 (Nov. 30, 2000), E. E. Miller and W. J. Benton (to Cabot
Corp.).
173. PCT Int. Appl. WO 97/26310 (July 24, 1997), D. P. Vollmer, P. H. Javora, and R. L.
Horton (to Great Lakes Chemical).
174. PCT Int. Appl. WO 97/26311 (July 24, 1997), D. P. Vollmer, P. H. Javora and R. L.
Horton (to Great Lakes Chemical).
175. Eur. Pat. Appl. EP 0726302 (Feb. 9, 1996), J. W. Dobson, T. D. Robertson, and K. B.
Mondshine (to Texas United Chemical Co.).
176. PCT Int. Appl. WO 98/21291 (May 22, 1998), J. R. Murphey and co-workers (to Tetra
Technologies Inc.).
177. K. K. Saxena and co-workers, Trends in Carbohydrate Chemistry 5, 139–151 (1999).
178. PCT Int. Appl. WO 98/03609 (Jan. 29, 1998), C. A. Sawdon and A. J. Bradbury
(to Schlumberger Canada Ltd.).
SHMUEL D. UKELES
BARUCH GRINBAUM
IMI (TAMI) Institute for Research and
Development, DSBG (Israel)
PETROLEUM REFINERY
PROCESSES
1. Introduction
Petroleum (also called crude oil), in the unrefined or crude form, like many
industrial feedstocks has little or no direct use and its value as an industrial com-
modity is only realized after the production of salable products. Even then, the
market demand dictates the type of products that are needed. Therefore, the
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1805060919160509.a01.pub2
292 PETROLEUM REFINERY PROCESSES Vol. 1
value of petroleum is directly related to the yield of products and is subject to the
call of the market.
Petroleum refining, also called petroleum processing, is the recovery and/or
generation of usable or salable fractions and products from crude oil, either by
distillation or by chemical reaction of the crude oil constituents under the effects
of heat and pressure. Synthetic crude oil, produced from tar sand (oil sand) bitu-
men, is also used as feedstocks in some refineries. Heavy oil conversion (1), as
practiced in many refineries, does not fall into the category of synthetic fuels
(syncrude) production. In terms of liquid fuels from coal and other carbonaceous
feedstocks, such as oil shale, the concept of a synthetic fuels industry has dimin-
ished over the past several years as being uneconomical in light of current pet-
roleum prices.
As the basic elements of crude oil, hydrogen and carbon form the main input
into a refinery, combining into thousands of individual constituents and the eco-
nomic recovery of these constituents varies with the individual petroleum accord-
ing to its particular individual qualities, and the processing facilities of a
particular refinery. In general, crude oil, once refined, yields three basic groupings
of products that are produced when it is separated into a variety of different gen-
eric, but often overlapping fractions (Table 1). The amounts of these fractions pro-
duced by distillation depend on the origin and properties of crude petroleum (2).
The gas and gasoline cuts form the lower boiling products and are usually
more valuable than the higher boiling fractions and provide gas (liquefied petro-
leum gas), naphtha, aviation fuel, motor fuel, and feedstocks, for the petrochem-
ical industry. Naphtha, a precursor to gasoline and solvents, is extracted from
both the light and middle range of distillate cuts and is also used as a feedstock
for the petrochemical industry. The middle distillates refer to products from
the middle boiling range of petroleum and include kerosene, diesel fuel, distillate
fuel oil, and light gas oil; waxy distillate and lower boiling lubricating oils are
sometimes include in the middle distillates. The remainder of the crude oil
includes the higher boiling lubricating oils, gas oil, and residuum (the nonvolatile
fraction of the crude oil). The residuum can also produce heavy lubricating oils
and waxes, but is more often used for asphalt production. The complexity of
petroleum is emphasized insofar as the actual proportions of light, medium,
and heavy fractions vary significantly from one crude oil to another.
Vol. 1 PETROLEUM REFINERY PROCESSES 293
2. History
The use of petroleum or derived materials, such as asphalt, and the heavier non-
volatile crude oils is an old art (2). In fact, petroleum utilization has been docu-
mented for >5000 years. The earliest documented uses occurred in Mesopotamia
(ancient Iraq) when it was recognized that the nonvolatile derivatives (bitumen
or natural asphalt and manufactured asphalt) could be used for caulking and as
an adhesive for jewelry or as a mastic for construction purposes. There is also
documented use of bitumen for medicinal use.
Approximately 2000 years ago, Arabian scientists developed methods for the
distillation of petroleum, which were introduced into Europe by way of the Arabian
incursions into Spain. Petroleum, used in China since it was encountered when
drilling for salt, appears in documents of the third century. The Baku region of
294 PETROLEUM REFINERY PROCESSES Vol. 1
northern Persia was also reported by Marco Polo in 1271–1273 as having a com-
mercial petroleum industry.
Interest in naphtha (nafta) began with the discovery that petroleum could be
used as an illuminant and as a supplement to bituminous incendiaries, which were
becoming increasingly common in warfare. Greek fire was a naphtha–bitumen (or
naphtha–asphalt) mix; the naphtha provided the flame and the bitumen (or
asphalt) provided the adhesive properties that prolonged the incendiary effect.
Modern refining began in 1859 with the discovery of petroleum in
Pennsylvania. After completion of the first well, the surrounding areas were
immediately leased and extensive drilling took place. In the post-1945 era, Middle
Eastern countries continued to rise in importance because of new discoveries of
vast reserves. The United States, though continuing to be the biggest producer,
was also the principal consumer and thus was not an exporter of oil. At this time,
oil companies began to roam much farther in the search for oil, which has
resulted in significant discoveries in Europe, Africa, and Canada.
The impetus to develop the petroleum refining industry came from several
changes in life-styles. The increased needs for illuminants, for fuel to drive the fac-
tories of the industrial revolution, for gasoline to power the automobiles, as well as
the demand for aviation fuel, all contributed to the increased use of petroleum.
The product slate has also changed. The increased demand for gasoline and
lubricants brought about an emphasis on refining crude oil. This, in turn,
brought about changes in the way crude oil was refined and led to innovations
and developments in the refining industry, thereby giving birth to the integrated
petroleum refinery (Fig. 1).
3. Feedstock Evaluation
Specific gravity is the ratio of the weight of a given volume of oil to the
weight of the same volume of water at a standard temperature, usually 608F
(15.68C). This method of measuring density and gravity first arose as a result
of the need to define the character of products in more detail; it was natural to
extend the measure to crude oils in general.
Vol. 1 PETROLEUM REFINERY PROCESSES 295
LPG Petrochemical
H
fuel feedstock
Gas C4S
plant H2SO4
Catalytic Gasoline
alkylation
reformer
160°C Naphtha
hydrofiner Catalytic
hydrocracker
300– 340°C
Mid-distillate
Kerosene
hydrofiner
Catalytic
cracker
Heating oil
340°C
Fuel oil
Refinery heavy-
ends processing
(a)
Atmospheric Vacuum
residuum pipe still 470– 560°C
Vacuum Gasoline
gas– oil
hydrotreater
560°C H2
Coker Hydrocracker in
fuel processing
Coke
Solvent Precipitate
deasphalting Asphalt
Soluble
Solvent Aromatic oils
extraction
Lube, waxes,
greases
(b)
Fig. 1. General refinery operations: (a) light petroleum refining section; (b) heavy
feedstock refining section.
296 PETROLEUM REFINERY PROCESSES Vol. 1
The Watson characterization factor (K) has also been used as a measure of
the chemical character of a crude oil or its fractions:
where TB is the absolute boiling point in degrees Rankine (8R ¼ 9/5 K) and speci-
fic gravity is specific gravity compared to water at 608F (15.68C).
For a wide boiling range material such as crude oil, the boiling point is
taken as an average of the five temperatures at which 10, 30, 50, 70, and 90%
of the material is vaporized. A highly paraffinic crude oil can have a character-
ization factor as high as 13, whereas a highly naphthenic crude oil can be as low
as 10.5, and the breakpoint between the two types of crude oil is 12.
4. Refinery Processes
Thus, refining actually commences with the production of fluids from the well or
reservoir and is followed by pretreatment operations that are applied to the
crude oil either at the refinery or prior to transportation. Pipeline operators,
eg, are insistent upon the quality of the fluids put into the pipelines; therefore,
any crude oil to be shipped by pipeline or, for that matter, by any other form of
transportation must meet rigid specifications in regard to water and salt content.
In some instances, sulfur content, nitrogen content, and viscosity may also be
specified.
Field separation, which occurs at a field site near the recovery operation, is
the first attempt to remove the gases, water, and dirt that accompany crude oil
coming from the ground. The separator may be no more than a large vessel that
gives a quieting zone for gravity separation into three phases: gases, crude oil,
and water containing entrained dirt.
Desalting operations are necessary to remove salt from the brines that
are present with the crude oil after recovery. The salt or brine suspensions
may be removed from crude oil by heating (90–1508C, 200–3008F) under pres-
sure (50–250 psi) that is sufficient to prevent vapor loss and then allowing the
material to settle in a large vessel. Alternatively, coalescence is aided by passage
through a tower packed with sand, gravel, and the like.
Desalting is a water-washing or electrostatic operation performed at the
production field and at the refinery site for additional crude oil cleanup. If the
petroleum from the separators contains water and dirt, water washing can
remove much of the water-soluble minerals and entrained solids. If these crude
oil contaminants are not removed, they can cause operating problems during
refinery processing, such as equipment plugging and corrosion as well as catalyst
deactivation.
The usual practice is to blend crude oils of similar characteristics, although
fluctuations in the properties of the individual crude oils may cause significant
variations in the properties of the blend over a period of time. Blending several
crude oils prior to refining can eliminate the frequent need to change the proces-
sing conditions that may be required to process each of the crude oils individu-
ally. However, incompatibility of different crude oils, which can occur if, eg, a
paraffinic crude oil is blended with a heavy asphaltic oil, can cause sediment for-
mation in the unrefined feedstock or in the products, thereby complicating the
refinery process (7).
Emulsions may also be broken by addition of treating agents, such as soaps,
fatty acids, sulfonates, and long-chain alcohols. When a chemical is used for
emulsion breaking during desalting, it may be added at one or more of three
points in the system. First, it may be added to the crude oil before it is mixed
with fresh water. Second, it may be added to the fresh water before mixing
with the crude oil. Third, it may be added to the mixture of crude oil and water.
A high potential field across the settling vessel also aids coalescence and breaks
emulsions, in which case dissolved salts and impurities are removed with the
water.
4.2. Distillation. The first and most fundamental step in the refining
process (after the crude oil has been cleaned and any remnants of brine removed)
is distillation and which is often referred to as the primary refining process. Dis-
tillation involves the separation of the different hydrocarbon compounds that
298 PETROLEUM REFINERY PROCESSES Vol. 1
injection of steam that is added to the column for stripping the nonvolatile con-
stituents in the base of the column.
Fractions from the atmospheric and vacuum towers are often used as feed-
stocks to these second stage refinery processes that break down the fractions, or
bring about a basic chemical change in the nature of a particular hydrocarbon
compound to produce specific products.
Alternatively, the atmospheric residuum or the vacuum residuum may be
sent to a deasphalting unit to remove the very high molecular weight constitu-
ents (collectively called asphalt) and the deasphalted oil is used as feedstock to a
catalytic cracking unit. Solvent deasphalting is a unique separation process in
which the residue is separated by molecular weight (density), instead of by boil-
ing point as in the distillation process. The purpose of a deasphalting unit is to
produce asphalt as final product and to produce a soluble stream (deasphalted oil
that can be used in a catalytic cracking unit and have reduced coke deposition
on the catalyst. The solvents used in a deasphalting unit vary from propane to
pentane and the yields of asphalt and deasphalted oil vary with the hydrocarbon
used in the unit.
Azeotropic Distillation and Extractive Distillations. Effective as they are
for producing various liquid fractions, distillation units generally do not produce
specific fractions. In order to accommodate the demand for such products, refi-
neries have incorporated azeotropic distillation and extractive distillation in
their operations.
The principle of azeotropic distillation depends on the ability of a chemically
dissimilar compound to cause one or both components of a mixture to boil at a
temperature other than the one expected. Thus, the addition of a nonindigenous
component forms an azeotropic mixture with one of the components of the
mixture, thereby lowering the boiling point and facilitating separation by distil-
lation. The separation of components of similar volatility may become economical
if an entrainer can be found that effectively changes the relative volatility. It is
also desirable that the entrainer be reasonably cheap, stable, nontoxic, and read-
ily recoverable from the components. In practice, it is probably the ready recover-
ability that limits the application of extractive and azeotropic distillation.
The majority of successful processes are those in which the entrainer and
one of the components separate into two liquid phases on cooling if direct recov-
ery by distillation is not feasible. A further restriction in the selection of an azeo-
tropic entrainer is that the boiling point of the entrainer is 10–408C (50–1048F)
below that of the components.
4.3. Thermal Cracking. When kerosene (lamp oil) was the principal
desired product, gasoline was the portion of crude petroleum too volatile to be
included in kerosene. The refiners of the 1890s and early 1900s had no use for
gasoline and often discarded an accumulation of it. However, as the demand
for gasoline and aviation fuel increased with the onset of World War I and the
increased production of automobiles during the 1920s, more fuels had to be pro-
duced to meet the demand.
The problem of how to produce more of the lower boiling fractions
from crude oil was solved in 1913 when cracking units were incorporated into
refinery operations and fractions heavier than gasoline were converted into gaso-
line by thermal decomposition. The use of residua as feedstocks for thermal
300 PETROLEUM REFINERY PROCESSES Vol. 1
processes has become economically advantageous because, on the one hand, the
end result is the production of lower boiling salable materials, and, on the
other, the asphaltic materials in the residua are regarded as the unwanted
coke-forming constituents.
The basic processes introduced to bring about thermal decomposition of the
higher boiling streams are known as cracking. In these processes, the higher
boiling fractions are converted to lower boiling products. Catalytic cracking is
the most common cracking process, in which heavy feedstock or cuts are broken
down or changed by being heated, and reacted with catalysts.
The concept behind thermal cracking is the thermal decomposition of
higher molecular weight constituents of petroleum to produce lower molecular
weight, normally more valuable, products. The first commercial process was in
1913 and is known as the Burton process. Even though catalyst cracking gene-
rally replaced thermal cracking in 1940s, noncatalytic cracking processes using
high temperature to achieve the decomposition are still in operation. Catalytic
processes usually produce more gasoline having a higher octane number, but
with lesser yields of gases and heavy fuel oil. The gases produced by catalytic
cracking contain more olefins than those produced by thermal cracking.
Thermal Cracking. In the thermal cracking process, a feedstock (eg, gas
oil) is fed to the fractionator with their thermal reactivity to separate gasoline,
light, and heavy oil. The light oil is then fed to the heater at 540–5958C (1000–
11008F) and a pressure of 350–700 psi, the light oil transforms to the vapor
phase and is sent to the soaker. If the feedstock is heavy oil, temperatures on
the order of 400–4808C (750–9008F) are used and higher pressures (350–
700 psi) are used to maintain the feedstock in the liquid phase, then it is fed to
the soaker. The liquid- and vapor-phase mix in the soaker and sent to the separa-
tor, with the products coming out on the bottom as fuel oil and the light recycle
back to the fractionator. Coking in the reactor is the main problem when heavy
oil is heated at high temperatures.
Visbreaking. Viscosity breaking (reduction) is a mild cracking operation
used to reduce the viscosity of residual fuel oils and residua (8). The process,
evolved from the older and now obsolete thermal cracking processes, is classed
as mild because the thermal reactions are not allowed to proceed to completion.
Residua are sometimes blended with lighter heating oils to produce fuel oils
of acceptable viscosity. By reducing the viscosity of the nonvolatile fraction,
visbreaking reduces the amount of the more valuable light heating oil that is
required for blending to meet the fuel oil specifications. The process is also
used to reduce the pour point of a waxy residue. Visbreaking conditions range
from 455 to 5108C (851–9508F) and 345 to 2070 kPa (50 to 300 psi) at the heating
coil outlet. Liquid-phase cracking takes place under these low severity condi-
tions. In addition to the primary product, fuel oil, material in the gas oil and
gasoline boiling range is produced. Gas oil can be used as additional feed for cat-
alytic cracking units or as heating oil.
In the process (Fig. 2), a crude oil residuum is passed through a furnace
where it is heated to 4808C (896 8F) under an outlet pressure of 690 kPa
(100 psi). The heating coils in the furnace are arranged to provide a soaking
section of low heat density, where the charge remains until the visbreaking
reactions are completed; subsequently, the cracked products are passed into a
Vol. 1 PETROLEUM REFINERY PROCESSES 301
Crude Gasoline
oil
Heavy gas oil
Vacuum
Atmospheric
fractionator
fractionator
Reactor
Flasher
Tar
Gas
Gasoline (naphtha)
Gas oil
Fractionator
Crude
Operative Nonoperative
oil
coking drum coking drum
(soaker) (soaker)
Coke
Heater
other thermal processes are also used. The feedstock is introduced into the pro-
duct fractionator. The fractionator bottoms, including a recycle stream of heavy
product, are heated in a furnace whose outlet temperature varies from 480 to
5158C (896–9598F). The heated feedstock then enters one of a pair of coking
drums where the cracking reactions continue.
The cracked products leave as overhead materials, and coke deposits form
on the inner surface of the drum. To provide continuous operation, two drums are
used; while one drum is on-stream, the one off-stream is being cleaned, steamed,
water-cooled, and decoked in the same time interval. The temperature in the
coke drum is in the range of 415–4508C (779–8428F) with pressures in the
range of 103–621 kPa (15–90 psi). Overhead products go to the fractionator,
where naphtha and heating oil fractions are recovered. The nonvolatile material
is combined with preheated fresh feed and returned to the furnace. The coke
drum is usually on stream for 24 h before becoming filled with porous coke,
after which the coke is removed hydraulically.
Fluid coking (Fig. 4) is a continuous process that uses the fluidized solids
technique to convert atmospheric and vacuum residua to more valuable products
(12,13). The residuum is converted to coke and overhead products by being
Vol. 1 PETROLEUM REFINERY PROCESSES 303
Overhead
products Flue gas
Reactor Burner
Heavy
feedstock Coke
Excess coke
sprayed into a fluidized bed of hot, fine coke particles, which permits the coking
reactions to be conducted at higher temperatures and shorter contact times than
they can be in delayed coking. Moreover, these conditions result in decreased
yields of coke; greater quantities of more valuable liquid product are recovered
in the fluid coking process.
Fluid coking uses two vessels: a reactor and a burner; coke particles are cir-
culated between the two to transfer heat generated by burning a portion of the
coke to the reactor. The reactor holds a bed of fluidized coke particles, and steam
is introduced at the bottom of the reactor to fluidize the bed.
Flexicoking (Fig. 5), also a continuous process, uses the same configuration
as the fluid coker, but has a gasification section in which excess coke can be gasi-
fied to produce refinery fuel gas. Flexicoking is a process by which excess coke-
make is reduced in view of the gradual incursion of the heavier feedstocks into
refinery operations. Such feedstocks are notorious for producing high yields of
coke (>15% by weight) in thermal and catalytic operations.
4.4. Catalytic Cracking. Fluid catalytic cracking (FCC) (Fig. 6) was
first introduced in 1942 and uses a fluidized bed of catalyst with continuous feed-
stock flow. In the process, which has progressively supplanted thermal cracking,
the thermal decomposition of petroleum constituents occurs in the presence of a
catalyst (14,15). The acid catalysts first used in catalytic cracking were desig-
nated low alumina catalysts; amorphous solids composed of 87% silica, SiO2,
and 13% alumina, Al2O3. Later, high alumina catalysts containing 25% alumina
and 75% silica were used. However, this type of catalyst has largely been
replaced by catalysts containing crystalline aluminosilicates (zeolites) or molecu-
lar sieves (16–18).
The catalyst is employed in bead, pellet, or microspherical form and can be
used as a fixed, moving, or fluid bed. The fixed-bed process was the first process
used commercially and employs a static bed of catalyst in several reactors, which
allows a continuous flow of feedstock to be maintained. The cycle of operations
consists of (1) the flow of feedstock through the catalyst bed; (2) the discontinuance
Reactor products Coke gas to
to fractionator Steam sulfur removal
generation
Fines
Scrubber removal
Cooling
Recycle
Heater
Steam Reactor
304
Gasifier
Purge
coke
Air
Steam
Cold
water
Reactor
Gas to recovery
Waste Stripping steam
heat
Water
boiler Spent catalyst
Gasoline
305
Regenerator Air Light gas oil
catalyst
of feedstock flow and removal of coke from the catalyst by burning; and (3) the
insertion of the reactor back on-stream. The moving-bed process uses a reaction
vessel, in which cracking takes place, and a kiln, in which the spent catalyst is
regenerated and catalyst movement between the vessels is provided by various
means. The fluid-bed process differs from the fixed-bed and moving-bed
processes, insofar as the powdered catalyst is circulated essentially as a fluid
with the feedstock. The several fluid catalytic cracking processes in use differ pri-
marily in mechanical design. Side-by-side reactor-regenerator construction and
unitary vessel construction (the reactor either above or below the regenerator)
are the two main mechanical variations.
Catalytic cracking has progressively supplanted thermal cracking as the
most advantageous means of converting distillate oils into gasoline and a better
yield of higher octane gasoline can be obtained than by any known thermal
operation. At the same time the gas produced consists mostly of propane and
butane with less methane and ethane. The production of heavy oils and tars,
higher in molecular weight than the charge material, is also minimized, and
both the gasoline and the uncracked cycle oil are more saturated than the pro-
ducts of thermal cracking.
4.5. Hydroprocessing. Hydroprocesses use the principle that the pre-
sence of hydrogen during a thermal reaction of a petroleum feedstock will termi-
nate many of the coke-forming reactions and enhance the yields of the lower
boiling components such as gasoline, kerosene, and jet fuel. Thus, in this
group of refining processes, which includes hydrotreating and hydrocracking,
the feedstock is heated with hydrogen at high temperature and under pressure.
The outcome is the conversion of a variety of feedstocks to a range of products
(Table 2) (1,2,4,9,19).
The purpose of hydroprocessing is (1) to improve existing petroleum pro-
ducts or develop new products or uses; (2) to convert inferior or low grade mate-
rials into valuable products; and (3) to transform near-solid residua to liquid
fuels. Products are as follows: from naphtha, reformed feedstock and liquefied
petroleum gas (LPG); from atmospheric gas oil, diesel and jet fuel, petrochemical
feedstock, and naphtha; from vacuum gas oil, catalytic cracker feedstock, kero-
sene, diesel and jet fuel, naphtha, LPG, and lubricating oil; and from residuum,
catalytic cracker and coker feedstock, diesel fuel, etc.
Hydroprocesses for the conversion of petroleum and petroleum products can
be classified as destructive or nondestructive. The former (hydrogenolysis and
dibenzothiophenes + 2 H2 + H2S
S
C6 H5 CH3 þ H2 ! C6 H6 þ CH4
and mild hydrocracking. The purpose is to reduce the organic nitrogen and
sulfur compounds to low levels (<50 ppm). Typical conditions are 340–3908C
(644–7348F), 10,350–17,250 kPa (1500–2500 psi), and a catalyst contact time
of 0.5–1.5 h, up to 1.5 wt% hydrogen is absorbed, partly by conversion of the
nitrogen compounds, but chiefly by aromatic compounds that are hydrogenated.
This first stage is usually carried out with a bifunctional catalyst containing
hydrogenation promoters, eg, nickel and tungsten or molybdenum sulfides, on an
acidic support. The metal sulfides hydrogenate aromatics and nitrogen com-
pounds, and retard deposition of carbonaceous deposits; the acidic support accel-
erates nitrogen removal as ammonia by breaking carbon–nitrogen bonds. The
catalyst is generally used as 0.32 0.32-cm or 0.16 0.32-cm pellets, as well as
spheres or other shapes.
Most of the hydrocracking is accomplished in the second stage. Hydrogen
sulfide, ammonia, and low boiling products are usually removed from the first-
stage product; the remaining oil, which is low in nitrogen–sulfur compounds,
is passed over the second-stage catalyst. Some catalyst systems do not require
the removal of adsorbed compounds between stages. In the second stage, typical
conditions are 300–3708C (572–6988F), 7000–17,250 kPa (1000–2500 psi) hydro-
gen pressure, and 0.5–1.5 h contact time; 1–1.5 wt% hydrogen may be absorbed.
Conversion to gasoline or jet fuel is seldom complete in one contact with the cat-
alyst, so the lighter oils are removed by distillation of the products, and the hea-
vier, high boiling product is combined with fresh feed and recycled over the
catalyst until it is completely converted.
The catalyst for the second stage is also a bifunctional catalyst containing
hydrogenating and acidic components. Metals such as nickel, molybdenum, tung-
sten, or palladium are used in various combinations and dispersed on solid acidic
supports, such as synthetic amorphous or crystalline silica–alumina, eg, zeolites.
These supports contain strongly acidic sites and sometimes are enhanced by
the incorporation of a small amount of fluorine.
A long period of operation (eg, >3 years) between catalyst regeneration is
desirable; this is achieved by keeping a low nitrogen content in the feed and
avoiding high temperatures and high end-point feedstock, which leads to excess
cracking and consequent deposition of coke on the catalyst. Feedstock conversion
is the key insofar as the conversion dictates the temperature employed. When
activity of the catalyst has decreased, it can often be restored by controlled burn-
ing of the coke.
Hydrotreating. On the other hand, hydrotreating is a catalytic process
converts sulfur- and/or nitrogen-containing hydrocarbons into low sulfur low
nitrogen liquids, hydrogen sulfide, and ammonia (21). A wide variety of metals
are active hydrogenation catalysts; those of most interest are nickel, palladium,
platinum, cobalt, and iron. Special preparations of the first three are active at
room temperature and atmospheric pressure. The metallic catalysts are easily
poisoned by sulfur- or arsenic-containing compounds, and even by other metals.
To avoid such poisoning, less effective, but more resistant metal oxides or sulfides
are frequently employed, generally those of tungsten, cobalt, or molybdenum.
Alternatively, catalyst poisoning can be minimized by mild hydrogenation to
remove nitrogen, oxygen, and sulfur from feedstocks in the presence of more
resistant catalysts, such as cobalt–molybdenum–alumina, Co–Mo–Al2O3.
Vol. 1 PETROLEUM REFINERY PROCESSES 309
The process temperature affects the rate and the extent of hydrogenation as
it does any chemical reaction. Practically every hydrogenation reaction can be
reversed by increasing temperature. If a second functional group is present,
high temperatures often lead to the loss of selectivity and, therefore, loss of
desired product yield. As a practical measure, hydrogenation is carried out at
as low a temperature as possible which is still compatible with a satisfactory
reaction rate.
Hydrotreating is carried out by charging the feed to the reactor together
with hydrogen at 300–3458C (572–6538F); the hydrogen pressures are 3450–
6900 kPa (500–1000 psi). The reaction generally takes place in the vapor phase
but, depending on the application, can also be a mixed-phase reaction.
After passing through the reactor, the treated oil is cooled and separated
from the excess hydrogen recycled through the reactor. The treated oil is pumped
to a stripper tower where hydrogen sulfide, formed by the hydrogenation reac-
tion, is removed by steam or by hydrocarbon vapor via reboiling, and the finished
product leaves the bottom of the stripper tower. The catalyst can be regenerated
in situ and ultimately be replaced after several regenerations.
4.6. Reforming. Reforming processes are used to change the inherent
chemical structures of the hydrocarbons that exist in distillation fractions
crude oil into different compounds. Catalytic reforming is one of the most impor-
tant processes in a modern refinery, altering straight run fraction or fractions
from a catalytic cracker into new compounds through a combination of heat
and pressure in the presence of a catalyst. Reforming processes are particularly
important in producing high quality gasoline fuels. Reforming processes are clas-
sified as continuous, cyclic, or semiregenerative, depending upon the frequency
of catalyst regeneration.
When the demand for higher octane gasolines increased during the early
1930s, attention was directed to ways and means of improving the octane number
of fractions within the boiling range of gasoline. Straight-run (distilled) gasolines
frequently had low octane numbers, and any process that could improve the octane
numbers would aid in meeting the demand for higher octane number gasoline.
Such a process, called thermal reforming, was developed and used widely, but to
a much lesser extent than thermal cracking. Thermal reforming was a natural
development from older thermal cracking processes; cracking converts heavier
oils into gasoline whereas reforming converts (reforms) gasolines into higher
octane gasolines. The equipment for thermal reforming is essentially the same
as for thermal cracking, but higher temperatures are used in the former.
In carrying out thermal reforming, a feedstock such as 2058C (4008F)
end-point naphtha or a straight-run gasoline is heated to 510–5958C (950–
11008F) in a furnace, much the same as a cracking furnace, with pressures
from 400 to 1000 psi (27 to 68 atm). As the heated naphtha leaves the furnace,
it is cooled or quenched by the addition of cold naphtha. The material then enters
a fractional distillation tower where any heavy products are separated. The
remainder of the reformed material leaves the top of the tower to be separated
into gases and reformate. The higher octane of the reformate is due primarily to
the cracking of longer-chain paraffins into higher octane olefins. The products of
thermal reforming are gases, gasoline, and residual oil or tar, the latter being
formed in very small amounts (1%). The amount and quality of the gasoline,
310 PETROLEUM REFINERY PROCESSES Vol. 1
The molecular size of the product is limited insofar as the reaction is termi-
nated at the dimer or trimer stage. Thus the process is more properly termed oli-
gomerization. The 4- to 12-carbon compounds required as the constituents of
liquid fuels are the prime products.
Thermal polymerization is not as effective as catalytic polymerization, but
has the advantage that it can be used to polymerize saturated materials that can-
not be induced to react by catalysts. The process consists of the vapor-phase
cracking of, eg, propane and butane, followed by prolonged periods at high
temperature (510–5958C; 950–11038F) for the reactions to proceed to near com-
pletion. Olefins can also be conveniently polymerized by means of an acid
312 PETROLEUM REFINERY PROCESSES Vol. 1
catalyst. Thus, the treated olefin-rich feed stream is contacted with a catalyst,
such as sulfuric acid, copper pyrophosphate, or phosphoric acid, at 150–2208C
(302–4288F) and 1035–8275 kPa (150–1200 psi), depending on feedstock and
product requirement.
Phosphates are the principal catalysts used in polymerization units; the
commercially used catalysts are liquid phosphoric acid, phosphoric acid on kie-
selguhr, copper pyrophosphate pellets, and phosphoric acid film on quartz. The
last is the least active and has the disadvantage that carbonaceous deposits must
occasionally be burned off the support. Compared to other processes, the one
using liquid phosphoric acid catalyst is far more responsive to attempts to
raise production by increasing temperature.
5. Treating
Since the original crude oils contain some sulfur compounds, the resulting
products and gasolines also contain sulfur compounds, including hydrogen sul-
fide, mercaptans, sulfides, disulfides, and thiophenes. The processes used to
sweeten, ie, desulfurize, the products depend on the type and amount of the sul-
fur compounds present and the specifications of the finished gasoline or other
stocks.
Hydrotreating is the most widely practiced treating process for all types of
petroleum products. However, there are other treating processes suitable for the
removal of mercaptans and hydrogen sulfide; such processes are necessary and
are performed as part of the product improvement and finishing procedures. For
example, mercaptan, RSH, removal is achieved by using regenerative solution
processes, in which the treatment solutions are regenerated rather than dis-
carded. Mercaptan conversion is essentially a process of oxidation to disulfides,
RSSR, by lead sulfide treatment, copper chloride–oxygen treatment, sodium
hypochlorite treatment, or oxygen treatment with a chelated cobalt catalyst in
either a caustic solution or fixed bed.
Hydrogen sulfide, H2S, is removed by a variety of processes, of which one is
a regenerative solution process using aqueous solutions of sodium hydroxide,
NaOH; calcium hydroxide, Ca(OH)2; sodium phosphate, Na3PO4; and sodium
carbonate, Na2CO3.
5.1. Alkali Treatment. Caustic washing is the treatment of materials,
usually products from petroleum refining, with solutions of caustic soda. The
process consists of mixing a water solution of lye (sodium hydroxide or caustic
soda) with a petroleum fraction. The treatment is carried out as soon as possible
after the petroleum fraction is distilled, since contact with air forms free sulfur,
which is corrosive and difficult to remove. The lye reacts either with any hydro-
gen sulfide present to form sodium sulfide, which is soluble in water, or with mer-
captans, followed by oxidation, to form the less nocuous disulfides.
H2 S þ 2 NaOH ! Na2 S þ 2 H2 O
RSH þ NaOH ! NaSR þ H2 O
4 NaSR þ O2 þ 2 H2 O ! 2 RSSR þ 4 NaOH
Vol. 1 PETROLEUM REFINERY PROCESSES 313
many useful separations in the purification of gasoline, kerosene, diesel fuel, and
other oils.
In solvent dewaxing, the oil is diluted with a solvent that has a high affinity
for oil, chilled to precipitate the wax, filtered to remove the wax, stripped of
solvent, and dried. The solvents [principally propane, naphtha, methyl ethyl
ketone (MEK)] act as diluents for the high molecular weight oil fractions to
reduce the viscosity of the mixture and provide sufficient liquid volume to permit
pumping and filtering. Wax, produced by the solvent dewaxing process, is used to
make (1) paraffins for candle wax, (2) microwax for cosmetics, and (3) wax for
petroleum jelly.
Catalytic dewaxing is a process in which the chemical composition of the
feed is changed. The process involved catalytic cracking of long paraffin chains
into shorter chains to remove the wax and produced lower molecular weigh pro-
ducts suitable for other uses. As an example, the feedstock is contacted with
hydrogen at elevated temperature and pressure over a catalyst (such as a zeolite)
that selectively cracks the normal paraffins to methane, ethane, and propane.
This process also knows as hydrodewaxing. There are two types of catalytic
dewaxing: (1) single-catalyst process that is used for pour point reduction and
to improve the oxygen stability of the product, and (2) a two-catalyst process
that uses a fixed-bed reactor and essentially no methane or ethane is formed
in the reaction.
In addition, solvent extraction can replace fractionation in many separation
processes in the refinery. For example, propane deasphalting (Fig. 7) has
Propane
recovery
and recycle
Stripper
Heavy
feedstock Deasphalted
oil
Evaporator
Stripper
Propane
(liquid)
Flasher
Asphalt
6. Gas Processing
The gas streams produced during petroleum refining usually contain many nox-
ious constituents that have an adverse effect on the use of the gas for other pur-
poses, eg, as a fuel or as a petrochemical feedstock, and some degree of cleaning is
required (29).
Gas purification processes fall into three categories: the removal of gaseous
impurities, the removal of particulate impurities, and ultrafine cleaning. The
extra expense of the last process is only justified by the nature of the subsequent
operations or the need to produce a pure gas stream. Because there are many
variables in gas treating, several factors must be considered (1) the types and
concentrations of contaminants in the gas; (2) the degree of contaminant removal
desired; (3) the selectivity of acid gas removal required; (4) the temperature,
pressure, volume, and composition of the gas to be processed; (5) the carbon
dioxide/hydrogen sulfide ratio in the gas; and (6) the desirability of sulfur recov-
ery on account of process economics or environmental issues.
Process selectivity indicates the preference with which the process removes
one acid gas component relative to or in preference to another. For example,
some processes remove both hydrogen sulfide and carbon dioxide, whereas
other processes are designed to remove hydrogen sulfide only. Thus it is impor-
tant to consider the process selectivity for hydrogen sulfide removal compared to
carbon dioxide removal, ie, the carbon dioxide/hydrogen sulfide ratio in the nat-
ural gas, in order to ensure minimal concentrations of these components in the
product.
One of the principal aspects of refinery gas cleanup is the removal of acid
gas constituents, ie, carbon dioxide, CO2, and hydrogen sulfide, H2S. Treatment
of natural gas to remove the acid gas constituents is most often accomplished by
contacting the natural gas with an alkaline solution. The most commonly used
treating solutions are aqueous solutions of the ethanolamines or alkali carbo-
nates. There are several hydrogen sulfide removal processes, most of which are
followed by a Claus plant that produces elemental sulfur from the hydrogen
sulfide (29).
7. Petroleum Products
The constant demand for products, such as liquid fuels, is the main driving force
behind the petroleum industry (5,30). In fact, it is the changes in product
demand that have been largely responsible for the evolution of the industry.
In general, when the product is a fraction that has been produced from cov-
nentioal petroleum or from heavy oil or resid and includes a large number of indi-
vidual hydrocarbons, the fraction is classified as a refined product. Examples of
refined products are gasoline, diesel fuel, heating oils, lubricants, waxes, asphalt,
and coke. In contrast, when the product is limited to, perhaps, one or two specific
316 PETROLEUM REFINERY PROCESSES Vol. 1
used as fuel for the refinery itself and as an important feedstock for the petro-
chemical industry.
7.2. Gasoline. Gasoline (also known as petrol in many parts of the world)
is one of the more important refinery products. The final gasoline product as a
transport fuel is a carefully blended mixture having a predetermined composition
that is adequate to power requirements (octane number) and climatic conditions.
Octane number is a measure of a fuel’s ability to avoid knocking. The octane
number of a gasoline is determined in a special single-cylinder engine where various
combustion conditions can be controlled. The test engine is adjusted to give trace
knock from the fuel to be rated. Various mixtures of isooctane (2,2,4-trimethyl
pentane) and normal heptane are then used to find the ratio of the two reference
fuels that produce the same intensity of knock as that by the unknown fuel.
By defining isooctane as having an octane number of 100 and n-heptane
as having an octane number of 0, the volumetric percentage of isooctane in hep-
tane that matches the knock from the unknown fuel can be calculated as the
octane number of the fuel. For example, 90 vol% isooctane and 10 vol% normal
heptane produce a 90-octane-number reference fuel.
Two kinds of octane number ratings are specified, although other methods
are often used for engine and fuel development. Both methods use the same
reference fuels and essentially the same test engine, but engine operating condi-
tions are different. In one test, called the research method (hence the Research
Octane Number, RON), the spark advance is fixed, the air inlet temperature is
1258F (528C), and engine speed is 600 rpm. The other, called the motor method
(hence the Motor Octane Number, MON), uses variable spark timing, a higher
mixture temperature of 3008F (1498C), and a faster engine speed of 900 rpm.
The more severe conditions of the motor method have a greater influence on
commercial blends than they do on the reference fuels. Thus, the motor octane
number of a commercial blend tends to be lower than the research octane num-
ber. Common practice is to label gasoline with an arithmetic average of both rat-
ings, abbreviated as (RþM)/2, and often referred to as road octane number.
The naphtha fraction from crude oil distillation is ultimately used to make
gasoline. The two streams are isolated early in the refining scheme so that each
can be refined separately for optimum blending in order to achieve the required
specifications, including the octane number.
Gasoline is a complex mixture of hydrocarbons that boils <2008C (3928F).
The hydrocarbon constituents in this boiling range are those that have 4–12
carbon atoms in their molecular structure. Gasolines can vary widely in compo-
sition, even those having the same octane number can be quite different (30).
Because of the differences in composition of the various gasolines, gasoline blend-
ing is necessary. The physical process of blending the components is simple, but
determination of how much of each component to include in a blend is much more
difficult. The operation is carried out by simultaneously pumping all the compo-
nents of a gasoline blend into a pipeline that leads to the gasoline storage, and
the pumps must be set to deliver automatically the proper proportion of each
component. Sophisticated instrumentation is employed to achieve the desired
blends.
The objective of gasoline (in fact, product) blending is to allocate the
available blending components in such a way as to meet product demands and
318 PETROLEUM REFINERY PROCESSES Vol. 1
specifications. In the blending process, product streams from other units are col-
lected and blended to produce the desired product. For example, in gasoline
blending, product streams from hydrotreating units, reforming units, polymeri-
zation units and alkylation units are blended to produce specification gasoline.
Aviation gasoline, usually used in light aircraft and older civil aircraft, have
narrower boiling ranges (38–1708C; 100.4–3388F) than automobile gasoline
(0–2008C; 32–3928F). In winter, the extreme temperatures in the northern
climates allow a certain amount of butane to be dissolved in the gasoline to faci-
litate vaporization. The narrower boiling range ensures better distribution of the
vaporized fuel through the more complicated induction systems of aircraft
engines. Aircraft operates at altitudes where the prevailing pressure is less
than the pressure at the surface of the earth, eg, at 5334 m, 52 kPa (7.5 psi) com-
pared to 101.3 kPa (14.7 psi). As a result, the vapor pressure of aviation gaso-
lines, which is a function of the fuel’s boiling range, must be limited to reduce
boiling in the tanks, fuel lines, and carburetors.
Another condition to keep gasoline engines running smoothly is to allow the
fuel–air mixture to start burning at a precise time in the combustion cycle. An elec-
trical spark starts the ignition. The remainder of the fuel–air mixture should be con-
sumed by a flame front moving out from the initial spark. Under certain conditions,
a portion of the fuel–air mixture ignites spontaneously instead of waiting for the
flame front from the carefully timed spark. The extra pressure pulses resulting
from spontaneous combustion are usually audible above the normal sounds of a run-
ning engine and give rise to the phenomenon called knock, of which pinging and
rumbling are special attributes. However, knocks are undesirable because they
waste some of the available power of an otherwise smooth-running engine.
Reformulated gasoline is believed to be the answer to many environmental
issues that arise from the use of automobiles and there has been a serious effort
to produce reformulated gasoline components (Table 5) from a variety of
processes (25).
However, reformulation may increase gasoline consumption, when in fact
the converse is preferable (24).
The production of low sulfur gasoline (and diesel fuel) is a necessary step to
assure clean air and to reduce air pollution from the transportation sector. Sulfur
is a pollutant and prevents the adoption of all major pollution control technolo-
gies. No significant air pollution reduction strategy can work without reducing
burning. Although some aromatics may occur within the boiling range assigned
to kerosene, excessive amounts can be removed by extraction.
The significance of the total sulfur content of kerosene varies greatly with the
type of oil and the use to which it is put. Sulfur content is of great importance when
the kerosene to be burned produces sulfur oxides, which are of environmental con-
cern. The color of kerosene is of little significance, but a product darker than usual
may have resulted from contamination or aging; in fact, a color darker than speci-
fied may be considered by some users as unsatisfactory. Kerosene, because of its
use as a burning oil, must be free of aromatic and unsaturated hydrocarbons;
the desirable constituents of kerosene are saturated hydrocarbons.
Diesel fuel, jet fuel, kerosene (range oil), No. 1 fuel oil, No. 2 fuel oil, and
diesel fuel are all popular distillate products coming from the kerosene fraction
of petroleum. One grade of jet fuel uses the heavy naphtha fraction, but the ker-
osene fraction supplies the more popular and heavier grade of jet fuel, as well as
smaller amounts that are sold as burner fuel (range oil) or No. 1 heating oil.
Some heating oils (generally No. 2 heating oil) and diesel fuel are similar and
can sometimes substitute for each other. The home heating oil is intended to
be burned with a furnace for space heating. The diesel fuel is intended for
compression–ignition engines.
The cetane number of a diesel fuel is a number that indicates the ability of a
diesel engine fuel to ignite quickly, and burn smoothly, after being injected into the
cylinder. In high speed diesel engines, a fuel with a long ignition delay tends to pro-
duce rough operation. The cetane number should not be confused with the cetene
number, an obsolete designation for the starting and running quality of diesel
fuel that uses cetene, C16H30, as the reference fuel. The cetene number has been
replaced by the cetane number, a scale based on the ignition characteristics of
two well-defined hydrocarbons, cetane, (n-hexadecane), and 2,3,4,5,6,7,8-hepta-
methylnonane.
Cetane has a short delay period during ignition and is assigned a cetane num-
ber of 100; heptamethylnonane has a long delay period and has a cetane number of
15. Just as the octane number is meaningful for automobile fuels, the cetane number
is a means of determining the ignition quality of diesel fuels and is equivalent to the
percentage by volume of cetane, in the blend with heptamethylnonane, that
matches the ignition quality of the test fuel. The cetane number of diesel fuel usually
falls into the 30–60 range; a high cetane number is an indication of the potential for
easy starting and smooth operation of the engine.
Other methods are also available for the estimation of diesel fuel quality.
For example, the diesel index (DI) is defined by the relation:
DI ¼ ðA F APIÞ=100
where A8F is the aniline point in degrees Fahrenheit and 8API is the American
Petroleum Institute gravity. A high aniline point corresponds to a high pro-
portion of paraffins in the diesel fuel; such a fuel has a high diesel index and,
therefore, a high cetane number.
7.5. Fuel Oil. Fuel oil is classified in several ways, but generally into two
main types: distillate fuel oil and residual fuel oil. Distillate fuel oil is vaporized
and condensed during a distillation process; it has a definite boiling range and
Vol. 1 PETROLEUM REFINERY PROCESSES 321
does not contain high boiling oils or asphaltic components. A fuel oil that con-
tains any amount of the residue from crude distillation hydrocracking is a resi-
dual fuel oil. However, the terms distillate fuel oil and residual fuel oil are losing
their significance because fuel oils are made for specific uses and can be either
distillates, residuals, or mixtures of the two. The terms domestic fuel oil, diesel
fuel oil, and heavy fuel oil are more indicative of the uses of fuel oil.
Domestic fuel oils are those fuel oils used primarily in the home and include
kerosene, stove oil, and furnace fuel oil. Diesel fuel oils are also distillate fuel oils,
but residual oils have been successfully used to power marine diesel engines, and
mixtures of distillates and residuals have been used on locomotive diesels. Heavy
fuel oils include a variety of oils, ranging from distillates to residual oils, that
must be heated to 2608C or higher before they can be used. In general, heavy fuel
oil consists of residual oil blended with distillate to suit specific needs. Heavy fuel
oil includes various industrial oils and, when used to fuel ships, is called bunker oil.
Stove oil is a straight-run (distilled) fraction from crude oil, whereas other
fuel oils are usually blends of two or more fractions. The straight-run fractions
available for blending into fuel oils are heavy naphtha, light and heavy gas oils,
and residua. Cracked fractions such as light and heavy gas oils from catalytic
cracking, cracking coal tar, and fractionator bottoms from catalytic cracking
may also be used as blends to meet the specifications of different fuel oils.
Heavy fuel oil usually contains residuum that is mixed (cut back) to a spe-
cified viscosity with gas oils and fractionator bottoms. For some industrial pur-
poses in which flames or flue gases contact the product (eg, ceramics, glass, heat
treating, and open hearth furnaces), fuel oils must be blended to low sulfur spe-
cifications; low sulfur residues are preferable for these fuels.
The manufacture of fuel oils at one time largely involved using what was
left after removing desired products from crude petroleum. Now fuel oil manufac-
ture is a complex matter of selecting and blending various petroleum fractions to
meet definite specifications.
Fuel oil that is used for heating is graded from No. 1 Fuel Oil, to No. 6 Fuel
Oil, and cover light distillate oils, medium distillate, heavy distillate, a blend
of distillate and residue, and residue oil. For example No. 2 and No. 3 Fuel oils
refer to medium to light distillate grades used in domestic central heating fuel
oil refers to a medium heavy residual oil used for heating large commercial
premises.
7.6. Lubricating Oil. Lubricating oil is distinguished from other petro-
leum fractions by the high (>4008C) boiling point as well as their high viscosity.
Lubricating oil may be divided into many categories according to the types of ser-
vice; however, there are two main groups: oils used in intermittent service, such
as motor and aviation oils, and oils designed for continuous service, such as turbine
oils.
Lubricating oil used in intermittent service must show the least possible
variation in viscosity with respect to temperature and must be changed at fre-
quent intervals to remove the foreign matter collected during service. The stabi-
lity of such oil is therefore of less importance than the stability of oil used in
continuous service for prolonged periods without renewal. Lubricating oil for
continuous service must be extremely stable because the engines in which it is
used operate at fairly constant temperature without frequent shutdown.
322 PETROLEUM REFINERY PROCESSES Vol. 1
Lubricating oil may be divided into many categories according to the types
of service it is intended to perform. However, there are two main groups: (1) oils
used in intermittent service, such as motor and aviation oils and (2) oils designed
for continuous service such as turbine oils.
7.7. Wax. Petroleum waxes are of two general types: paraffin wax in
distillates and microcrystalline wax in residua. The melting point of wax is not
directly related to its boiling point because waxes contain hydrocarbons of dif-
ferent chemical structure. Nevertheless, waxes are graded according to their
melting point and oil content. Paraffin wax is a solid crystalline mixture of
straight-chain (normal) hydrocarbons ranging from mostly C20 –C30 and higher.
Wax constituents are solid at ordinary temperatures (258C; 778F) whereas petro-
latum (petroleum jelly) contains both solid and liquid hydrocarbons.
Wax production by wax sweating was originally used in Scotland to sepa-
rate wax fractions by employing various melting points from the wax obtained
from shale oils. Wax sweating is still used to some extent, but is being replaced
by the more convenient wax recrystallization process. In wax sweating, a cake of
slack wax, also known as crude or raw wax, is slowly warmed to a temperature at
which the oil in the wax and the lower melting waxes become fluid and drip (or
sweat) from the bottom of the cake, leaving a residue of higher melting wax.
Insofar as they are used to purify other products, several processes used in
the refinery fall under the classification of dewaxing processes; however, such
processes must also be classified as wax production processes (1,2). Most com-
mercial dewaxing processes utilize solvent dilution, chilling to crystallize the
wax, and filtration (28). The MEK process (MEK–toluene solvent) is widely
used. Wax crystals are formed by chilling through the walls of scraped surface
chillers, and wax is separated from the resultant wax–oil–solvent slurry by
using fully enclosed rotary vacuum filters.
Solvents used for dewaxing are naphtha, propane, sulfur dioxide, acetone–
benzene, trichloroethylene, ethylenedichloride-benzene (Barisol), MEK–benzene
(benzol), methyl n-butyl ketone, and methyl n-propyl ketone. Other solvents in com-
mercial use for dewaxing include N-methylpyrrolidinone, MEK–MIBK (methyl iso-
butyl ketone), dichloroethane-methylene dichloride, and propylene–acetone.
Solvent dewaxing can be applied to light, intermediate, and heavy lubricat-
ing oil distillates, but each distillate produces a different kind of wax, and each of
these waxes is actually a mixture of a number of waxes. For example, the wax
obtained from light paraffin distillate consists of a series of paraffin waxes that
have melting points in the range of 30–708C (86–1588F) and are characterized
by a tendency to harden into large crystals. However, heavy paraffin distillate
yields a wax composed of a series of waxes that have melting points in the
range of 60–908C (140–1948F) and that harden into small crystals from which
they derive the name microcrystalline waxes or microwaxes.
On the other hand, intermediate paraffin distillates contain paraffin waxes
and waxes intermediate in properties between paraffin and microwaxes. Thus,
the solvent dewaxing process produces three different slack waxes depending
on whether light, intermediate, or heavy paraffin distillate is processed. The
slack wax from heavy paraffin distillate may be sold as dark raw wax, the wax
from intermediate paraffin distillate as pale raw wax. The latter is treated with
lye and clay to remove odor and improve color.
Vol. 1 PETROLEUM REFINERY PROCESSES 323
Recycle gas
Hydrotreater
or fuel gas
Low
temperature
flash
Naphtha
Furnace
324
Distillation
Reactor
Waxy
raffinate
or
deasphalted
residuum
High
temperature
flash
Finished lube
base stock
Light
Partial hydrocarbons
Neutrals or Solvent Raffinate Catalytic
solvent
bright stock extract dewaxing Dewaxed lube
dewaxing
Foot oil
Extract
Deoil
Hardwax
date, but the use of asphaltic solids for paving goes back to the European prac-
tices of the early 1800s.
Cutback asphalts are mixtures in which hard asphalt has been diluted
with a lighter oil to permit application as a liquid without drastic heating.
They are classified as rapid, medium, and slow curing, depending on the vola-
tility of the diluent, which governs the rates of evaporation and consequent
hardening.
Asphalt can be emulsified with water to permit application without heat-
ing. Such emulsions are normally of the oil-in-water type. They reverse or
break on application to a stone or earth surface, so that the oil clings to the
stone and the water disappears. In addition to their usefulness in road and
soil stabilizations, they are useful for paper impregnation and waterproofing.
The emulsions are chiefly either the soap or alkaline type, or the neutral or clay
type. The former breaks readily on contact, but the latter is more stable and
probably loses water mainly by evaporation. Good emulsions must be stable dur-
ing storage or freezing, suitably fluid, and amenable to control for the speed of
breaking.
Recently, asphalt has grown to be a valuable refinery product. In the post-
1980 period, a shortage of good quality asphalt has developed. This is due in no
short measure to the tendency of refineries to produce as much liquid fuels (eg,
gasoline) as possible. Thus, residua that would have once been used for asphalt
manufacture are now being used to produce liquid fuels (and coke).
7.9. Coke. Coke is the residue left by the destructive distillation (coking)
of residua. The composition of coke varies with the source of the crude oil, but
in general, large amounts of high molecular weight complex hydrocarbons
(rich in carbon, but correspondingly poor in hydrogen) make up a high propor-
tion. The solubility of coke in carbon disulfide has been reported to be as high
as 50–80%, but this is, in fact, a misnomer, since the coke is an insoluble,
honeycomb-type material that is the end product of thermal processes.
Petroleum coke is employed for a number of purposes; its principal use is in
the manufacture of carbon electrodes for aluminum refining, which requires a
high purity carbon that is low in ash and free of sulfur. In addition, coke is
employed in the manufacture of carbon brushes, silicon carbide abrasives, struc-
tural carbon (eg, pipes and Rashig rings), as well as calcium carbide manufacture
from which acetylene is produced. Coke produced from low quality crude oil is
326 PETROLEUM REFINERY PROCESSES Vol. 1
mixed with coal and burned as a fuel. Flue gas scrubbing is required. Coke is
used in fluidized-bed combustors or gasifiers for power generation.
8. Petrochemicals
Hydrogen Steam
Steam or oxygen Hydrogen
Methane and carbon and carbon
monoxide Urea
dioxide
Oxygen Air
Hydrogen cyanide Carbon dioxide
327
Acetylene
Air
Methanol Formaldehyde
Acrylonitrile
Hydrogen chloride
Dimer Hydrogen
chloride Chlorine or Methyl chloride
or acetic alkali
Hydrogen acid
chloride
Methylene dichloride
Ethanolamines Diethylene
(Catalyst) and triethylene
Polyethylene glycols
Ammonia
Water
Oxygen
328
Hydrogen chloride
Bromine
Ethylene Ethylene dibromide
Hydrogen
chloride
Ethyl chloride
Water (catalyst)
Ethyl alcohol Acetaldehyde
Sulfuric
acid Sulfuric Water
Cracking
of ethane, esters
propane, Benzene
heavy Styrene
Ethylbenzene
liquids
Oxygen, acetic acid
(catalyst)
Vinyl acetate
from refinery gas streams. The various butylenes are more commonly obtained
from refinery gas streams. Butane dehydrogenation to butylene is known, but
is more complex than ethane or propane cracking, and its product distributions
are not always favorable. The production of gasoline and other liquid fuels con-
sumes large amounts of butane.
Hydrogen
Butanols
Carbon monoxide,
Butyraldehydes
hydrogen, catalyst Hydrogen
2-Ethylhexanol
(Catalyst)
Polypropylene
Water Dehydrogenation
Methyl ethyl ketone
(sulfuric acid)
Water
(sulfuric acid)
Copolymer with 2% isoprene
(butyl rubber)
Isobutylene
(Sulfuric acid)
Di- and triisobutylene
Polyisobutylene
Boron trifluoride (low temperature)
330
anhydride Diphenylmethane
Vinyl diisocyanate Urea
acetate Polyurethanes
Ethylene
Cellulose glycol
acetate Formic
Single-cell acid Acetal resins
protein Acetic Urea]
acid formaldehyde
Acrylates resins
Fuel
Methanol Formaldehyde
C4-diols Melamine- and
phenol-formaldehyde
Pentaerythritol resins
Ethylene diamine
Olefins Hexamethylene tetraacetic acid
tetramine
Gasoline Methacrylates
Air
Xylenes Phthalic anhydride
(Minor)
(Minor)
(Hydrogen) Nitric acid
Benzene Toluene Nitrotoluenes
Propylene trimer
Caprolactam
Epichlorhydrin
Formaldehyde,
urea
and o-xylene into phthalic anhydride. Both are involved in a wide variety of con-
sumer products.
Benzene, toluene, and xylene are made mostly from catalytic reforming of
naphthas with units similar to those already discussed. As a gross mixture, these
aromatics are the backbone of gasoline blending for high octane numbers.
However, there are many chemicals derived from these same aromatics; thus
many aromatic petrochemicals have their beginning by selective extraction
from naphtha or gas-oil reformate.
8.3. Inorganics. Of the inorganic chemicals, ammonia is by far the most
common. Ammonia is produced by the direct reaction of hydrogen with nitrogen;
air is the source of nitrogen: N2 þ 3 H2 ! 2 NH3. Refinery gases, steam reforming
of natural gas (methane) and naphtha streams, and partial oxidation of hydrocar-
bons or higher molecular weight refinery residual materials (residua, asphalts)
are the sources of hydrogen. Ammonia is used predominantly for the production
of ammonium nitrate, NH4NO3, as well as other ammonium salts and urea,
H2NCONH2, which are primary constituents of fertilizers.
Carbon black, also classed as an inorganic petrochemical, is made predomi-
nantly by the partial combustion of carbonaceous (organic) material in a limited
supply of air. Carbonaceous sources vary from methane to aromatic petroleum
oils to coal tar by-products. Carbon black is used primarily for the production of syn-
thetic rubber.
Sulfur, another inorganic petrochemical, is obtained by the oxidation of
hydrogen sulfide:
2 H2 S þ O 2 ! 2 H2 O þ 2 S
BIBLIOGRAPHY
‘‘Petroleum Refinery Processes’’ in ECT 1st ed., Vol. 10, pp. 109–161, by W. L. Nelson and
A. P. Buthod, University of Tulsa; in ECT 2nd ed., Vol. 15, pp. 1–77, by W. L. Nelson,
University of Tulsa; ‘‘Petroleum (Refinery Processes, Survey)’’ in ECT 3rd ed., Vol. 17,
pp. 183–256, by C. E. Jahnig, Consultant; in ECT 4th ed., Vol. 18, pp. 433–469, by
J. Speight, Western Research Institute; ‘‘Petroleum Refinery Processes’’ in ECT
(online), posting date: December 4, 2000, by J. Speight, Western Research Institute.
CITED PUBLICATIONS
1. J.G. Speight and B. Ozum. Petroleum Refining Processes, Marcel Dekker, New York,
2002.
Vol. 1 PETROLEUM REFINERY PROCESSES 333
2. J. G. Speight, The Chemistry and Technology of Petroleum, 3rd ed., Marcel Dekker,
New York, 1999.
3. J. G. Speight, The Chemistry and Technology of Coal, 2nd ed, Marcel Dekker,
New York, 1994.
4. J. G. Speight, The Desulfurization of Heavy Oils and Residua, Marcel Dekker,
New York, 1981.
5. M. S. Matar, in H. K. Abdel-Aal, B. A. Bakr, and M. A. Al-Sahlawi, eds., Petroleum
Economics and Engineering, 2nd ed., Marcel Dekker, New York, 1992, p. 33.
6. D. R. Burris and J. D. McKinney, in J. J. McKetta, ed., Petroleum Processing Hand-
book, Marcel Dekker, New York, 1992, p. 666.
7. G. W. Mushrush and J. G. Speight, Petroleum Products: Instability and Incompatibil-
ity, Taylor & Francis, Washington, D.C., 1995.
8. J. R. Salazar, in R. A. Meyers, ed., Handbook of Petroleum Refining Processes,
McGraw-Hill Book Co., New York, 1986, pp. 7–68.
9. J. G. Gary and G. E. Handwerk, Petroleum Refining: Technology and Economics, 3rd
ed, Marcel Dekker, New York, 1994.
10. J. R. Salazar, in Ref. 8, pp. 7–69.
11. H. M. Feintuch, J. A. Bonilla, and R. L. Godino, in Ref. 8, pp. 7–3.
12. J. D. McKinney, in Ref. 4, p. 245.
13. D. E. Blaser, in Ref. 4, p. 253.
14. D. G. Tajbl, in Ref. 8, pp. 2–9.
15. E. C. Luckenbach and co-workers, in Ref. 4, p. 349.
16. S. A. Bradley, M. J. Gattuso, and R. J. Bertolacini, eds., Characterization and Catalyst
Development: An Interactive Approach, American Chemical Society, Washington, D.C.,
1989.
17. R. T. K. Baker and L. L. Murrell, eds., Novel Materials in Heterogenous Catalysis,
American Chemical Society, Washington, D.C., 1990.
18. M. L. Occelli, ed., Fluid Catalytic Cracking II: Concepts in Catalyst Design, American
Chemical Society, Washington, D.C., 1991.
19. G. E. Weismantel, in Ref. 4, p. 592.
20. D. G. Tajbl, in Ref. 8, pp. 2–33.
21. C. N. Cabrera, in Ref. 8, pp. 6–17.
22. J. A. Weiszmann, in Ref. 8, pp. 3–3.
23. P. Greenough and J. R. K. Rolfe, in Ref. 8, pp. 7–69.
24. A. Douaud, Tomorrow’s Engines and Fuels, Report No. 94-2. Institut Francais du
Petrole, Paris, 1994.
25. R. Ragsdale, Oil Gas J., 51 (Mar. 21, 1994).
26. B. R. Shah, in Ref. 8, pp. 1–1.
27. T. Hutson, Jr. and W. C. McCarthy, in Ref. 8, pp. 1–23.
28. G. G. Scholten, in Ref. 4, p. 583.
29. J. G. Speight, Gas Processing: Environmental Aspects and Methods, Butterworth
Heinemann, Oxford, U.K., 1993.
30. H. L. Hoffman, in Ref. 4, p. 2.
31. J. D. Hargrove and J. D. McKinney, in Ref. 4, p. 558.
32. C. P.-C. Chang, J. R. Murphy, and J. D. McKinney, in Ref. 4, p. 527.
JAMES G. SPEIGHT
Consultant, CD & W Inc.
334 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS Vol. 1
OIL REFINING,
ENVIRONMENTAL
CONSIDERATIONS
1. Environmental Protection
1.1. Introduction 1–3. The natural product crude oil and the products
obtained from it are generally easy to handle and with careful treatment —
which is necessary because of their flammability — they cause little pollution
to the environment. (The CO2 emissions associated with all fossil fuels and
the corresponding effects on the earth’s climate cannot be considered in this
context.)
Environmental problems arise, however, in dealing with crude oil and its
products because the petroleum industry — like hardly any other branch of
industry — maintains exploration, transport, and refining installations for
crude oil that are scattered over the entire globe. The distribution and consump-
tion of burning and transportation fuels are even more scattered (Fig. 1). Since
the geographical location of crude oil fields does not coincide with that of the
large consuming and refining regions, enormous transportation distances exist.
The large amounts handled — the annual refining capacity in 1988 was ca.
3109 t in more than 600 refineries worldwide — means that the industry
needs to take precautions in all areas with regard to air emissions, water and
soil pollution and, to a lesser degree, noise. The problems of introducing uniform
standards are caused by legal standards which differ from country to country
depending on the local concentrations of industrial areas with their correspond-
ing environmental pollution and population density.
Because of the high cost of technical measures for environmental conserva-
tion, harmonization of legislation is absolutely necessary. As these costs amount
to more than 25 % of the total processing costs in the refineries equal competition
between different countries must be guaranteed (see Section Cost of Environ-
mental Conservation).
In the following sections, the environmental problems in crude oil refining,
storage and loading, and in the application of the products are discussed sepa-
rately, because the problems in the various fields differ.
1.2. Manufacturing Emissions. Although closed, gas-tight systems are
generally used in refinery units, emissions into air and water cannot be comple-
tely avoided even with careful handling during refining and storage of the crude
oil and its products. This is due to the management and control of the process
and the properties of the products concerned.
Hydrocarbons are discharged into the air because of their high vapor pres-
sure and they appear in refinery wastewater effluents because of their water
solubility, which is, however, small. The carcinogenic aromatic hydrocarbons
are particularly dangerous.
Further attention must be paid to the sulfur and nitrogen compounds
originating from the heteroatomic compounds in the crude oil, both because of
their smell and toxicity and because of the air pollution which arises in the
form of SO2 and NOx emissions during firing in process plants.
Fig. 1. Petroleum’s route from the well to the consumer shown as an example for
Germany.
1. Vapor recovery, where in a closed system the displaced gasoline vapors are
either returned to the product tank or collected in a gasometer and used for
the firing of process plants
2. Regenerative adsorption of the vapors on suitable adsorbents
3. Recovery of the products in liquid form after cooling or washing out the
vapors
Dispensing of gasoline from the road tanker to the service station is also
increasingly carried out with vapor recovery between the road tanker and
the installed storage tank of the service station. The possibilities of reducing
emissions in the final step (filling the customer’s motor vehicle) are described
in Section Transportation Fuels.
The measures described reduced the atmospheric emissions from proces-
sing, storage, and distribution of refinery products in the mid 1980s in Western
Europe to < 8 % of the total man-made HC emissions. The portion from the
refineries themselves was in turn only a quarter of this value. More recent inves-
tigations by Concawe show further improvements. Thus, relative to the crude oil
throughput, refinery emission losses of ca. 0.05 wt% are expected, depending on
the structure and size of the plant. For inland refineries in Germany, for exam-
ple, values of 0.02 – 0.03 wt% have been published.
Hydrocarbons in Wastewater. Hydrocarbon-containing wastewater is
unavoidably obtained at various points in the refinery because of the water con-
tent of the crude oil itself and because steam is employed in various processing
steps. The total amount of hydrocarbon-containing wastewater in a normal
refinery is of the order of 60 – 100 m3/h.
This wastewater must be removed from the process units after separation of
the oil phase, and led via a closed system to wastewater purification. Rainwater
from exposed plant areas and from tank yards, and possibly contaminated cool-
ing water from leaks or accidents must be treated in the same way. Considerable
buffer volumes must be made available for the latter amounts of water which are
formed discontinuously and sometimes in large quantities.
Treatment in the wastewater purification system is carried out stepwise by:
COD 60 – 100
BOD 5 – 15
Oil 0.5 – 2
Settleable solids 0.1 – 0.3
Phenols 0.1 – 0.2
338 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS Vol. 1
must be treated in situ. Depending on the nature of the soil and the
corresponding migration of the oil, this can be done by pumping off and purify-
ing the contaminated water, possibly by additional injection of fresh water in
adjacent wells.
Degradation of the oil by microorganisms is becoming increasingly
important. This can also be done in situ or by excavating the soil and treating
it externally (6).
Loading and storage of the products outside the refinery are subject to
similar regulations. However, these are still very different in individual
countries. As a result of the large number of external distribution depots for
transport fuels and heating oils and of service stations, and because of the
enormous number of oil-heated households, special care must be taken
against overfilling and escape of products due to leaks. Corrosion-resistant
and nonageing steels, plastic-lined steel tanks, and novel, glass-fiber-
reinforced, plastic tanks are widely used. These tanks must also be con-
structed such that the whole volume can be collected if leakage occurs (double-
walled tanks).
In some countries with a population density and dense housing and
industrial areas, the regulations which apply to pure hydrocarbons (i.e., pro-
ducts) are also applied to the handling of process water with much lower
HC-concentrations.
Sulfur and Nitrogen Compounds. As a natural product, crude oil also
contains heteroatomic compounds containing sulfur, nitrogen, and oxygen in
addition to hydrocarbons. Whereas the nitrogen and oxygen contents are in
the ppm range and play only a secondary role in atmospheric emissions, the
sulfur content of the crudes can be as high as several percent. The distribution
over the individual refinery fractions varies, but the content increases with
increasing molecular size.
Sulfur Compounds. Sulfur and its compounds are catalyst poisons and
adversely affect atmospheric emissions. Hydrogen sulfide, mercaptans, and dis-
ulfides are odor nuisances, and sulfur dioxide is formed during the combustion of
crude oil products. Therefore, sulfur and its compounds must be removed or their
contents reduced.
The light refinery products, liquefied petroleum gas (see LIQUEFIED PETRO-
LEUM GAS) and gasoline, must be almost completely sulfur-free; for diesel fuels
(see AUTOMOTIVE FUELS) and light heating oils (see HEATING OIL), a substantial
sulfur reduction to 0.1–0.5% is required by recent legislation.
Serious problems exist with heavy fuel oil, which is used almost
exclusively as fuel in large industrial furnaces and power stations and leads
to considerable SO2 emissions. Many countries have established a maximum
sulfur content in fuels of 1 – 2 wt%. This value can be reached without
additional treatment only with a few, low-sulfur crude oils, whose supplies
are limited.
Sulfur is generally removed from distillates by hydrodesulfurization,
whereby the chemically bound sulfur is converted to hydrogen sulfide. The
H2S is then removed in a gas scrubber, converted to elemental sulfur in
the downstream Claus process, and supplied to the chemical industry as a raw
material.
340 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS Vol. 1
The intense smell of the H2S-containing gases to be processed and the high
toxicity of H2S, even at high dilution, means that precautions must be taken
when handling. All sulfur-processing plants must be completely gastight: in
the hazard zones, instruments and alarm devices are installed which automati-
cally shut down the plants in case of danger.
Sulfur recovery in refineries has increased greatly as the result of the lower
permitted sulfur content of the products. In conventional Claus plants with two
reactors connected in series the conversion of the H2S feed is ca. 95 %, a third
reactor increases the conversion to 96 %. For a further reduction of the H2S con-
tent an aftertreatment step (Scot process, Sulfren, see NATURAL GAS) is required.
Almost complete removal that is required in various countries (eg, Germany
10 ppm H2S in off-gas) can only be achieved with an additional high-tempera-
ture combustion step. Figure 4 compares the sulfur removal from the processed
crude oil with the crude oil capacity in West German refineries for the period
1970–1988. The increase in sulfur removal resulted in a dramatic reduction of
overall SO2 emissions since most of the desulfurized hydrocarbons are used as
fuels.
Sulfur dioxide emissions in the refineries themselves are relatively small.
Depending on the refinery complexity, only 4–6% of the crude oil used is
required for fueling in the process plants. About half of this amount is covered
by the low-sulfur or sulfur-free gases obtained during processing, and the
remainder is heavy fuel oil. With more severe requirements to lower SO2 emis-
sions, the refiner might be forced to use as fuel all of the LPG that is produced in
the refinery.
Legislation differs greatly in the individual countries: the stipulated
limit values can be related to a single process unit, a stack, or the refinery
as a whole and can refer to the total sulfur discharged or the sulfur concentra-
tion in the flue gas. Combinations of several measures are also possible, and so-
called ‘‘bubble concepts’’ are used, particularly in the United States. (Bubble
Vol. 1 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS 341
concept means that for a given amount of emission from the total complex,
different limit values or amounts discharged are allowed for each individual
plant.)
The expensive flue gas desulfurization is normally uneconomic for firing
installations of the size of those employed in a normal refinery. Therefore it is
used only in special cases, eg, if the refinery is connected with a large power
station or if the sulfur content of the fuel is extremely high.
Nitrogen Compounds. The nitrogen content of most crude oils is rela-
tively low, and in the distillates it is reduced during hydroprocessing to a few
ppm. The residual nitrogen content causes no further difficulty in the use of
the products.
Ammonia, which is formed in the hydrogenation steps and added in many
refineries for process control in various process stages (pH adjustment), enters
the wastewater and can be removed in the biological stage of wastewater purifi-
cation (see Section Hydrocarbons in Watewater).
Noise. In the past, noise pollution in the neighborhood of petroleum refi-
neries played only a secondary role. Worldwide, pumps and compressors were
installed not in closed buildings but in the open air, because of the flammability
of petroleum and its products. However, because the plants were smaller at that
time, the noise emission of the units was relatively small. In addition, the sound
radiation from furnaces and their burners produced relatively low sound levels
in the refinery surroundings, because of the compact construction of the main
process plants that were simultaneously screened by lownoise auxiliary installa-
tions (tank farms) at the periphery.
Problems arose, though, in densely populated areas with insufficient dis-
tances between industrial and residential areas. Most noise problems were
caused by flare noise in the event of process disturbances.
Owing to growing environmental awareness in the public and a correspond-
ingly greater strictness of legislation, further measures will have to be taken in
the future to reduce sound emissions. This is also necessary for technical rea-
sons, because the complex conversion refinery maintains more and larger units
with greater furnaces and higher-powered gas compressors with high-speed
steam turbines etc.
Important measures for noise reduction could cover the following items:
one-third of the total output, corresponding to ca. 12% of the total man-made
emissions of organic substances.
Two solutions are under discussion (Fig. 5) (8):
been carried out with various types of vehicles. Installation in new cars pre-
sents no problems; retrofitting to old vehicles will, however, only be possible in
a few cases.
Owing to the wide use of motor fuels, particular attention must be
paid to benzene emissions. The limit for the benzene content of gasoline
(1 – 5 vol% is under discussion) presents the oil industry with problems,
particularly in the production of unleaded gasoline, because the benzene
content of various blending stocks can be considerably higher, depend-
ing on the quality or origin of the crude, up to 8 vol% for high severity refor-
mate and 18 – 40 vol% for pyrolysis gasoline. Benzene extraction from HC-
stocks does not solve this problem because of the cost and the question of
reuse of the benzene recovered. Therefore, the evaporative losses must be
minimized.
Diesel Fuel. The main exhaust gas problems of diesel vehicles con-
cern the emissions of soot particles and SO2. The NOx and CO components
are less important than in the spark ignition (Otto) engine. Demands are
increasingly being made for a reduction of these emissions, because of the
smell of the exhaust gases, and because the soot particles may have a carcino-
genic effect.
Burning Fuels. The largest proportion of refinery products worldwide is
used as a fuel for domestic or industrial heating as well as for energy produc-
tion. These are the fractions of LPG (see LIQUEFIED PETROLEUM GAS) the middle
distillates, and heavy fuel oil (see HEATING OIL). Emissions of SO2, NOx, and
CO that are associated with combustion, as well as emission of ash and
unburnt carbon (in the form of soot) can be considerably influenced by selection
and pretreatment of the products used and by the burner design. This is par-
ticularly true for the large heating installations operated with sulfur-rich
heavy fuel oils.
Liquefied Petroleum Gas. (see LIQUEFIED PETROLEUM GAS). The liquefied
petroleum gases used as fuels are almost sulfur-free as a result of pretreatment
in the refineries, and they burn without formation of soot. Their use presents no
problem for the user.
Light Heating Oil (No. 2 Fuel). Middle distillates contain up to 1.5 wt%
sulfur, depending on the origin of the crude. Most of the sulfur can be removed
in the refinery by hydrosulfurization.
The permissible sulfur limits for light heating oil in most countries are
< 0.5 wt%, because of its use in domestic heating. In the EU, the sulfur content
of domestic heating oil is now limited to 0.2 wt% (see HEATING OIL – ENVIROMENTAL
PROTECTION). If lower sulfur contents are to be achieved, desulfurization costs will
rise steeply with the current range of crudes, according to studies by Concawe (9)
(Fig. 6).
Heavy Fuel Oil (No. 6 Fuel). The residues from crude oil refining
are used as heavy fuel oil for industrial heating and energy production. Sulfur
dioxide, nitrogen oxides, CO, and particulate emissions must be considered for
heavy fuel oils; for environmental protection measures in this field.
1.4. Cost of Environmental Conservation. Because the understand-
ing of environmental problems, technical solutions, and legislation are still in
Vol. 1 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS 345
Table 2. Investment and operating costs for environmental conservation in the Federal
Republic of Germany in comparison with crude oil processing
1977 1979 1981 1983 1985 1987
6
Crude oil processed, 10 t/a 99.5 109.6 98.9 90.2 87.3 81.4
Investment costs, 106 DM 294 127 130 305 103 221
Operating costs, 106 DM 826 894 1129 1212 1097 900
346 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS Vol. 1
standards and regulations and those expected in future were taken into account.
The most important measures in the processing field are:
1. Gas washing systems and Claus plants for sulfur recovery from H2S
followed by fine purification.
2. Closed systems for the discharge of gaseous (via a flare) and liquid hydro-
carbons.
3. Floating roofs and floating covers in storage tanks for crude oil and
products.
4. Hydrocarbon vapor recovery systems for the storage and loading of volatile
products.
5. Collection spaces for escaping hydrocarbons and precipitation water in
production plants, tank farms and loading plants. Drainage to wastewater
purification plants.
6. Reduction of emissions (SO2, NOx) from the plant’s firing installations.
7. Closed sampling systems, laboratory analysis, and on-line instruments for
pollutant measurement in air and wastewater.
The following measures are for the production of less polluting products:
In determining the measures required and the resulting costs, the study
considered the environmental conservation laws applicable in each EC member
country at the time concerned or scheduled up to 1993. The results are given in
Table 3.
They show that: (1) the environmental costs rose in all EC countries;
(2) there are large differences in cost between the individual countries; (3)
for comparable costs arising from the EC directives, the additional national
Table 3. Environmental costs of petroleum processing in the EC member states (in ECU/t)
1993 1993
Country 1985 EC directives only ECþnational regulations
Sulfur
Fuel gas
Atmospheric
distillation
225-375°C
HDS Gas oil
Boiling 375°C
and higher
Fuel oil
Fig. 7. Hydroskimming.
Table 4. Yield structures of refinery conversion schemes for Arabian light crude
processing (11)
Product Hydroskimming a FCC –VBb HC – SDA –FCC c HC –Coking d
Intake, wt %
Crude oil 100.0 100.0 100.0 100.0
Hydrogen 0.4 0.5
Output, wt %
Fuel gas 1.8 3.4 3.1 4.7
LPG 1.6 3.0 2.9 2.9
Chemical 2.0 1.7
feed
Gasoline 15.9 33.0 32.9 24.4
Kerosene 9.6 9.6 9.6 9.6
Gas oil 25.3 26.3 40.2 50.9
Fuel oil 43.5 22.7 9.2
Coke 1.0 e 1.1 e 5.0 f
Sulfur 0.3 1.0 1.4 1.3
a
Hydroskimming = refinery configuration without residue conversion processes.
b
FCC – VB = conversion scheme with cat cracker and visbreaker.
c
HC – SDA – FCC = conversion scheme with hydrocracker, solvent deasphalting, and cat cracker.
d
HC – coking = conversion scheme with hydrocracker and coker.
e
Coke burnt in FCC unit.
f
Coke produced in coker unit.
Vol. 1 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS 349
Sulfur
Fuel gas
FCC
HDS
375-375°C alkylation
Vacuum
Boiling 375°C distillation Coke
and higher
Boiling 550°C Visbreaker Fuel oil
and higher
Sulfur
Fuel gas
Hydro-
375-550°C FCC and
cracking
alkylation
Steam Vacuum Coke
reformer distillation SDA HDS
Boiling 550°C
and higher Fuel oil
The chemical and physical properties of crude oils and their fractions depend on
the types and concentrations of their consituents. The typical properties of crude
oils and their products must be defined and determined by analytical methods, to
meet quality requirements for production, transport, storage and handling of
crude oils, and to assess specific quality criteria for further processing of products.
This quality testing is performed by regular laboratory analysis or by auto-
matic instruments.
4.1. Crude Oil and Product Properties. Density and API Gravity.
The density ranges of various crudes are as follows:
Sulfur
Fuel gas
Amines gas Propane
plant and
Claus plant Butane
H2
Hydro-
cracking
Steam
reformer Vacuum
distillation Fluid
coking Coke
141:5
API ¼ 131:5
Density½kg=L
Distillation. The boiling range of crude oils is one of the most important
quality criteria for determining the yield distribution of the different product
fractions. Boiling analyses under atmospheric conditions can be supplemented
by distillation under reduced pressure for oil fractions that boil above 360 8C.
More detailed and precise results are obtained by applying fractional or true
boiling point (TBP) distillation.
Viscosity. The viscosity of crude oils is determined by the ratio of low and
high molecular mass constituents in the oil. The kinematic viscosity of crude oils
(at 38 8C ¼ 100 8F) can vary from 4 mm2/s (0.04 St) for light (North Sea) crudes to
700 mm2/s (7 St) for heavy (Venezuelan) crudes.
Flash Point. The flash point depends on the concentration of light
(low-boiling) constituents. For crude oils, flash points between 20 8C and
þ 80 8C are observed.
Vol. 1 OIL REFINING, ENVIRONMENTAL CONSIDERATIONS 353
Pour Point. The pour point is the temperature at which an oil or oil pro-
duct ceases to flow. It is an important measure of its cold flow properties.
Other Properties. that are analytically tested include specific heat;
calorific value; contents of sulfur, nitrogen, oxygen, and inorganic compounds;
the C/H ratio; the hydrocarbon composition; and product-specific criteria (see
Section OIL PRODUCTIONS).
4.2. Testing Methods and Standards. In the oil industry a frame-
work of standard analytical procedures has been established, e.g., by ASTM,
DIN and the British Institute of Petroleum, (IP) standards. In addition,
various house standards have been introduced by most international oil
companies.
as long term (emergency) storage; caverns for gas are used as buffer storage
(see NATURAL GAS).
Subsea Storage. Crude oils (from offshore production) and hydrocarbon
condensates (from offshore separation) are stored in undersea tanks.
Transport of Crude Oils and Products. Crude oils and bulk (mass)
products are normally transported by large tankers and pipelines. Finished
products are usually transported by smaller tankers and river barges, product
pipelines, rail tank cars, and road tank trucks.
REFERENCES
1. The Oil Refining Industry and the Environment, CONCAWE, The Hague, Sept.
1987.
2. O. Neuwirth: ‘‘Umweltfragen bei der Nutzung von Mineralöl,’’ Erlanger Forsch.
Reihe B; Naturwiss. Med. 14 (1984) 119–134.
3. Mineralöl und Umweltschutz, 3rd ed., Publ. Mineralölwirtsch. Verband, Hamburg
1989.
4. A. H. Edwards et al.: ‘‘Volatile Organic Compound Emissions in Western Europe,’’
CONCAWE Rep. 1986, no. 2.
5. A. de Roocker et al.: ‘‘Trends in Oil Discharged with Aqueous Effluents from Oil
Refineries in Western Europe,’’ CONCAWE Rep. 1985, no. 61.
6. C. L. F. Webb: ‘‘Biodegradation of Contaminated Soil,’’ Erdoel Erdgas 106 (1990)
no. 11, 453–455.
7. J. S. McArragher et al.: ‘‘Motor Vehicle Emission Regulations and Fuel Specifica-
tions, CONCAWE Rep. 1990, no. 2.
8. International Workshop Berlin: ‘‘Emissionsarme Betankung von Kraftfahrzeugen,’’
Erdoel Kohle Erdgas Petrochem. 42 (1989) 468.
9. R. J. Ellis: ‘‘Cost Estimating Aspects of Environmental Control Technology — HDS
as a Case Example,’’ CONCAWE Rep. 1988, no. 51.
10. Commission of the European Communities: Study on the Costs Borne by the Refining
Industry in the Member States in Order to Comply with Environmental Legislation,
prepared by Chem. Systems Int., London 1986.
11. G. Heinrich, Y. Jacquin, A. Vidal, Proc. World Pet. Congr., 10th, 1979, vol. 4, paper
PD 18–1, Heyden & Sons, London 1979.
12. F. G. Turpin, C. R. Lipuma, A. M. Edelman, in (60), paper PD 18–2.
13. C. T. Douwes, J. van Klinken, J. B. Wyffels, W. C. van Zijll Langhout, in (60), paper
18–3.
WALTHER W. IRION
Deutsche Shell AG, Hamburg,
Federal Republic of Germany
OTTO S. NEUWIRTH
UK Wesseling, Wesseling,
Federal Republic of Germany
Vol. 1 NATURAL GAS 355
NATURAL GAS
1. Introduction
Ma ¼ Syi Mi (1)
Nonhydrocarbons
Nitrogen 0.00–0.15
Carbon dioxide 0.00–0.30
Hydrogen sulfide 0.00–0.30
Helium 0.00–0.05
pV ¼ znRT (2)
% ¼ pM=zRT (3)
% ¼ nM=V (4)
For a gas mixture such as natural gas, the molar mass M in Equation (3) is
replaced by the apparent molar mass, Ma, and z is the compressibility factor of
the gas mixture at the temperature and pressure conditions of interest.
Vol. 1 NATURAL GAS 357
For a gas mixture, the relative density can also be calculated from the
apparent molar mass Ma of the gas mixture divided by the molar mass of air
(28.963):
Ma
d¼ (6)
28:963
where Tc, pc, and Vc are the critical point conditions, and T, p, and V are the
values of the system, respectively.
358 NATURAL GAS Vol. 1
where the subscript pc designates the pseudocritical values of the gas mixture,
ci the critical values of the components, and yi is the molar fraction of compo-
nent i. The pseudocritical pressure and temperature are then used to determine
the pseudoreduced conditions:
ppr ¼ p= ppc
Tpr ¼ T=Tpc
Although they are convenient for correlating purposes the pseudocritical values
have no physical significance.
Viscosity. For an ideal gas, the pressure and temperature dependence of
viscosity is opposite to that of liquids; the viscosity of an ideal gas increases with
increasing temperature and is independent of pressure. Actual hydrocarbon
gases, however, deviate from ideal behavior and approach that of liquids; their
viscosity increases with increasing pressure, and decreases with increasing tem-
perature at intermediate or high pressure.
Figure 1 is employed for estimating the dynamic viscosity uga of natural
gases composed primarily of hydrocarbons at atmospheric pressure (6). The pre-
sence of carbon dioxide, hydrogen sulfide, and nitrogen increases the viscosity.
The viscosity ug at any desired temperature and pressure can be obtained from
Figure 2, where the ratio ug/uga is plotted as a function of the pseudoreduced
temperature and pressure.
Compressibility of Real Gases. At higher pressure the attractive
or repulsive forces between gas molecules become important because the
Vol. 1 NATURAL GAS 359
Fig. 1. Viscosity of paraffin hydrocarbon gases at 101.3 kPa as a function of molar mass
(6). [(8F32)/1.8 ¼ 8C] (Copyright Society of Petroleum Engineers, 1954).
equation of state can be solved directly for the compressibility factor z. The most
reliable methods for typical natural gases are those of ROBINSON and JACOBY (12),
and HALL and YARBOROUGH (13).
Well-known virial equations are the Benedict–Webb–Rubin equation (14)
and the Starling–Hon (15) extension of the Benedict–Webb–Rubin equation.
The European Gas Research Group (GERG) and the van der Waals Institute
at Amsterdam carried out an extensive research project (16) to develop a virial
equation of state for the accurate prediction of compressibility factors for natural
gas mixtures. They used the virial equation solved for the compressibility factor z
and truncated after the third term:
z ¼ 1 þ BðTÞ%M þ CðTÞ%2M
where %M is the molar density, %M ¼ 1/VM, and B and C are the second and third
virial coefficients, which depend on temperature and gas composition. With this
Vol. 1 NATURAL GAS 361
connecting the critical point C with the maximum temperature point CT. This
line separates the two-phase envelope in two regions. In the lower region the
gas mixture behaves as expected: At constant or decreasing temperature, the
liquid content in the gas mixture increases with increasing pressure. The beha-
vior of the gas mixture in the two-phase region above the connection line C–CT is
quite different. At constant temperature, the liquid content in the gas mixture
increases with decreasing pressure, and in the range between the critical pres-
sure and the maximum pressure, the liquid content increases at a pressure drop
even when the temperature increases.
This phenomenon, called retrograde condensation, may seriously impair the
transport of natural gas through pipelines because liquid slugs can be formed if
the higher hydrocarbons have not been removed sufficiently. The condensation
behavior of natural gas can be determined by employing equations of state or
by laboratory measurements (17).
Gas Hydrates. Gas hydrates are icelike solids composed of water and
hydrocarbons. Hydrocarbons ranging from methane to cyclopentane are known
to form hydrates. The formation of hydrates depends on pressure, temperature,
molecular size, and concentration of the component. Generally, hydrates form
at high pressure and low temperature, but they can occur as high as 308C and
below 0.7 MPa. Hydrate formation can be predicted from empirical vapor–solid
equilibrium ratios, empirical correlations, and laboratory measurements (18).
To avoid the formation of gas hydrates in pipelines, the water vapor content
of natural gas is commonly reduced by dehydration before transport. Lowering
the water dew point to 5 to 88C in relation to the maximum transmission pres-
sure, is a common stipulation in gas specifications.
1.3. Combustion Properties. Calorific Value. The calorific or heat-
ing value of natural gas is defined as the heat evolved by complete combustion
of a unit quantity of dehydrated gas with pure oxygen at 0.101325 MPa. The
initial temperature of gas and oxygen is 258C. The flue gas resulting from the
combustion process is cooled to the same temperature.
In determining the total calorific value, also called the gross calorific value
(higher heating value), all the water formed in the combustion process is con-
densed to the liquid state. For determination of the net calorific value (lower
heating value), all the water formed by the combustion reaction remains in the
vapor state: ie, the gross and net calorific values differ in the condensation
enthalpy of water. The calorific value is expressed in megajoules per cubic
meter or, occasionally, kilowatt-hours per cubic meter.
Calorific Value of Gas Mixtures. The calorific value of natural gas
can be calculated from the molar composition and the calorific values of the
components:
X
H¼ yi Hi
where yi is the molar fraction of component i, Hi is the gross or net calorific value
of component i. Table 4 shows calorific values for commonly occurring compo-
nents of natural gas.
Wobbe Index. When burned at the same pressure, natural gases with
different compositions generate the same amount of heat per time unit if their
Vol. 1 NATURAL GAS 363
calorific values divided by the square root of their relative densities (Wobbe
index) are equal. The Wobbe index can be referred to the gross or net calorific
value:
pffiffiffi
Wgross;net ¼ Hgross;net = d
Since gas burners tolerate only limited fluctuations in the heating value of fuel
gas, these fluctuations must be restricted. Therefore, natural gas specifications
generally stipulate ranges that are defined by Wobbe index rather than heating
values (20). Consequently, certain gas streams may require blending before
sales, or separate pipeline networks may be operated on different gas streams
with deviating Wobbe index levels.
2. Treating
364
Germany,
Süd-Oldenburgg
former states of 42 8.5 5.2 3.3 22 0.8 2.6 16 1000 ppm
USSR, Tenguizh
Canada, Bearberryi 4 1 5 90 80
a
Sweet gas with low calorific value, only adjustment of water dew point.
b
Natural gas with high calorific value, H2S conversion to COS at molecular sieve.
c
Mercury removal, C3þ separation by low temperature separation and adsorption.
d
Liquefaction, separation of gas condensate, CO2 – H2S – Hg removal.
e
Helium production.
f
Nitrogen separation, CO2 removal.
g
Sour gas, use of sulfur solvent, H2S, CO2 scrubber, sulfur production.
h
Condensate, LPG, sulfur recovery, removal of organic sulfur compounds.
i
Ultrasour gas, only for sulfur production.
Vol. 1 NATURAL GAS 365
Natural gas treating can take place both directly at the well and in
centralized plants. Treating at the well is always necessary when the gas cannot
be transported to a processing plant by pipeline without hazard. To avoid corro-
sion, sour gas is often dehydrated at the well so that free water can no longer
condense in the pipeline. The recovery of sulfur and the separation of nitrogen,
carbon dioxide, or helium are carried out in centralized plants, to which streams
of different wells are fed. The same holds true for liquefaction of natural gas,
recovery of LPG, and removal of higher hydrocarbons.
2.2. Natural Gas Wells and Field Installations. Installations at indi-
vidual gas wells must fulfill the following objectives:
Fig. 5. Treating of natural gas. —— Flow sheet for treatment of sweet gas.– – –
Additional steps required for treatment of sour gas.
Fig. 6. Schematic of a typical sour gas well completion. a) Well head; b) Anchor casing;
c) Cement; d) Intermediate casing; e) Injection annulus; f ) Tubing; g) Production packer;
h) Liner; i) Perforations
Fig. 7. Flow diagram of a sulfur solvent inhibitor system. a) Lean oil storage; b) Sour gas
well; c) Free water knockout; d) Liquid control valves; e) Flow control valve; f ) Cooler; g)
Separator; h) Stripper; i) Liquid storage; j) Phase separation; k) Extraction unit; l) Phase
separator; m) Ethylamine polysulfide; n) Ethylamine regeneration; o) Ethylamine storage.
Fig. 10. Attainable dew points of natural gas as a function of the (gas) contact tempera-
ture with triethylene glycol–water solutions.
Fig. 11. Hydrocarbon removal with adsorption and regeneration under pressure.
a)–c) Adsorbers; d) Heater; e) Cooler; f ) Chiller; g) Separator; h) Phase separator.
Fig. 12. Phase diagram for natural gas containing higher hydrocarbons and for pure
methane. a) Dew point curve; b) Boiling point curve; c) Critical point; d) Cricondentherm
point.
process is superfluous. At a water content of ca. 3–5% in the scrubbing agent (eg,
Selexol), it is also possible to regenerate without a strip gas compressor; the
stripping agent is then the water vapor present.
To reduce losses and prevent condensation of the scrubbing agent in the
pipeline, small posttreatment units are often installed (eg, with glycol). Washers,
activated carbon adsorbers, or refrigeration systems, are also possible. Because
of the good solubility of higher hydrocarbons in these physical solvents, these
processes also allow the recovery of LPG.
Amount of Scrubbing Agent. The circulating amount of scrubbing
agent required is almost independent of the concentration of the components
to be removed but does depend on pressure and raw gas throughput. Loading
of the physical scrubbing agent increases with the partial pressure of the compo-
nents to be removed: Thus, physical absorption is used advantageously at higher
concentration of these components and higher process pressure. A large amount
of coabsorbed methane can be released in the first depressurization step by flash-
ing (usually multistage flashing) and can be fed profitably into internal fuel-gas
systems.
A high selectivity for hydrogen sulfide compared to carbon dioxide leads
to process streams with high hydrogen sulfide concentration, which can be
used in the Claus process. A certain residual amount of carbon dioxide in the
treated gas may be desirable to adjust its calorific value to required specifica-
tions. To match quality specifications, physical scrubbing processes can be
operated with variable selectivity by adjusting circulation of the solvent and
temperature.
Physical scrubbing processes are noncorrosive, and regular carbon steel is
generally used as construction material. Complete removal of carbonyl sulfide
from the crude gas requires sophisticated equipment.
Chemical Absorption Processes (35,38–40). Alkanolamine scrubbing,
known since 1930, is currently one of the most frequently used processes
(! Gas Production). Aqueous solutions of monoethanolamine (MEA), diethanola-
mine (DEA), diisopropylamine (DIPA), diglycolamine (DGA), and methyldietha-
nolamine (MDEA), in particular, are used as absorbents. The choice of solvent is
determined by different factors.
Monoethanolamine allows a high final purity but requires a considerable
amount of energy for regeneration, and absorption of carbonyl sulfide leads to
degradation products that must be eliminated in a reclaimer (see Section 2.8).
Selective removal of hydrogen sulfide is not possible. Diglycolamine, like MEA
a primary amine, is suitable for the complete absorption of both hydrogen
sulfide and carbon dioxide and can also be used for scrubbing of carbonyl sul-
fide and mercaptans. The secondary amines DEA and DIPA show a lower heat
of reaction for absorption–desorption of sour gas components; as a result,
they require less regeneration steam than MEA. At higher carbon dioxide
levels, various degradation compounds are formed that cannot be regenerated,
and require use of a reclaimer (see Section 2.8). A certain selectivity for hydro-
gen sulfide is possible, particularly with DIPA, because its reaction with
carbon dioxide is slower. Tertiary amines such as MDEA have considerably
better selectivity for hydrogen sulfide. The heat of reaction with MDEA is
lower, degradation products are not formed with carbonyl sulfide and carbon
Vol. 1 NATURAL GAS 377
dioxide, and only a few corrosion problems arise. If gases at low pressure are
purified, the normal pipeline specification for hydrogen sulfide is not attained.
Comparative data for characterization of various alkanolamines are presented
in Table 7.
Addition of corrosion inhibitors and special chemically active components
leads to advanced processes that allow higher alkanolamine concentration and
higher loading, for example, or better selectivity for hydrogen sulfide and, there-
fore, lower circulation rates. A lower corrosion tendency and a reduction of
energy required for regeneration can thereby be obtained. Examples are the
Amine Guard, Econamine, activated MDEA (BASF), Gas/Spec, and Ucarsol
processes (41,42). Selectivity can be increased, and further savings in energy
can be obtained by the use of sterically hindered amines (Flexsorb, Ucarsol
Innovator Solvent 111). The energy consumption for the hydrogen sulfide/carbon
dioxide or carbon dioxide scrubbing system is also lower with these amines
(43,44).
Activated and inhibited potassium carbonate solution (Benfield, Catacarb)
is widely applied for purification of natural gas (45). The process can in principle
be used to remove carbon dioxide, hydrogen sulfide, carbonyl sulfide, and carbon
disulfide. However, it is used mostly in treating gases with high carbon dioxide
content. The steam requirement for regeneration can also be reduced by addition
of hindered amines.
Physical–Chemical Absorption Processes. The advantages of chemical
absorption (eg, removal of COS, low coabsorption of methane) are combined with
those of physically effective organic solvents (eg, high loading at high partial
pressure). As amines, DIPA and MDEA are preferentially employed. The physi-
cal absorbents are high-boiling solvents (eg, Sulfolan or Selexol). In addition,
5–30% water is present. The Sulfinol process has found widespread use (46,47)
and has proved useful in the United States, Canada, and the Federal Republic of
Germany for simultaneous removal of hydrogen sulfide, carbon dioxide and
organic sulfur compounds. However, the coabsorption of hydrocarbons is also
increased by a physical solvent, so that an intermediate flash step with reabsor-
ber for the recovery of sulfur-free fuel gas (shown as a dashed line in Fig. 14) is
indispensable. Selective hydrogen sulfide scrubbing can be attained with this
process by proper choice of the amine.
Other physical–chemical absorption processes include Selefining, Flexsorb
PS (44), and Amisol. With the Amisol process a mixture of methanol and
378 NATURAL GAS Vol. 1
5) Thiopaq
Thiopaq is a biological process for SO2 reduction, employing absorption
in caustic solution and a sulfur formation step by biological oxidation. It
requires a sulfate bleed stream (82,83).
regeneration gas may become so large that absorptive workup and recircula-
tion should be considered.
Adsorption of Hydrogen Sulfide. Various adsorptive processes are
available for the removal of hydrogen sulfide; these are differentiated by the
type of binding to the adsorbent and the method of regeneration. Activated
charcoal, iron oxide, and zinc oxide are generally used only for the removal of
smaller amounts of hydrogen sulfide because of the limited regeneration capabil-
ity of these adsorbents (53).
Molecular sieves display a high affinity for sulfur compounds. These
materials can be selected to permit simultaneous dehydration and desulfuriza-
tion without removal of carbon dioxide from the gas. Plants are in service which
can purify gas containing 10–6000 ppm of hydrogen sulfide and 0.1–50%
carbon dioxide. The valuable but hydrogen sulfide-rich regeneration gas is
usually treated further (54) and reused.
With special molecular sieves, hydrogen sulfide can be removed and, in the
presence of carbon dioxide, converted to carbonyl sulfide:
H2 S þ CO2 ! COS þ H2 O
For gas with low hydrogen sulfide content the permitted level of carbonyl
sulfide in marketable natural gas can be exploited by this means. Regeneration
and further treatment of the hydrogen sulfide-containing desorption gas are then
no longer necessary. A survey on adsorption processes is given in Table 8.
2.5. Membrane Processes. Gas separation with membranes has devel-
oped into an economically interesting process in the last decade. Membranes can
be used for almost all aspects of natural gas treating. The following are of parti-
cular interest:
Poreless, dense polymer films are used for the separation of gas mixtures.
Individual components of the gas dissolve in the polymer and move through it by
Table 9. Types of Membranes and Modules for Carbon Dioxide Separation and
Dehydration (76–80)
Selectivity Permeability
aaCO =CH , of COa2 , H2Ob
Type of membrane 2 4
m3 (STP)
and polymer Type of module Company abH
2 O=CH4 m –2 h –1 bar –1
CO2 Separation
Asymmetric spiral-wound Separex 25–30a 0.25–0.4a
cellulose acetate
Asymmetric hollow fiber Cynara, Dow 25–30a 0.25–0.4a
cellulose
triacetate
Asymmetric spiral-wound Grace 25–30a 0.25–0.4a
cellulose acetate
Composite hollow fiber Permea, 17–24a & 0.1a
polysulfone/ Monsanto
silicone (PRISM)
Polyaramide hollow fiber Medal
(Du Pont/Air
Liquide)
Asymmetric flat sheet Grace 19a 0.4a
polyimide (spiral-wound)
(6 FDA/DMB (79))
H2O Separation
Polysulfone, hollow fiber Permea
polyphenylene oxide
(PRISM CACTUS)
Composite flat sheet GKSS 500b 35b
polyetherimide– (envelope type)
ethylcellulose ether
Asymmetric cellulose spiral-wound Grace 300b 3.2b
acetate
2 H2 S þ O2 ! 2 S þ 2 H2 O (1)
Vol. 1 NATURAL GAS 385
The first exothermic reaction step (Eq. 2) takes place in the combustion
chamber at ca. 900–11008C:
2 H2 S þ 3 O2 ! 2 SO2 þ 2 H2 O (2)
2 H2 S þ 2 SO2 ! 4 S þ 2 H2 O (3)
Cooling of the process gas from the combustion chamber occurs in the
waste-heat boiler, which is combined with a brick-lined combustion chamber.
For each tonne of sulfur produced, more than 2 t of steam is coproduced. Conden-
sers are arranged between individual stages, in which the evaporation of water is
used to cool the reaction gases as much as possible; the aim is to attain as high a
degree of separation of elemental sulfur as possible.
Prior to the catalytic reaction, reheating of the gas occurs to the activation
temperature of the catalyst (ca. 180–2208C). Inactivation caused by condensa-
tion of sulfur must be excluded. In larger plants, both gas–gas heat exchangers
and in-line burners are used. The latter must be run exactly and substoichiome-
trically because the Claus reaction is very sensitive to disturbances and catalysts
are deactivated by reaction with oxygen. Attainable yields depend on the hydro-
gen sulfide content of the gas and the number of catalytic steps. The choice of
catalyst and the heater–separator concept are also important. The pressure is
predetermined: it is approximately the atmospheric regeneration pressure of
the upstream gas scrubbing units. The pressure drop in sulfur recovery plants
is 20–50 kPa.
In the combustion chamber and the waste-heat boiler, 50–70% of the hydro-
gen sulfide is converted to elemental sulfur. The combustion chamber conditions
and the construction of the burner—both multijet burners and single burners
have been built for plants generating sulfur at up to 1500 t/d—influence the
course of combustion and the formation of compounds that influence the process.
Significant amounts of carbonyl sulfide, carbon disulfide, carbon monoxide,
hydrogen, and sulfur trioxide are formed (59,60).
Catalysts for the Claus process are generally based on aluminum oxide,
which may also contain other metal oxides. Traces of sulfur trioxide, which
may be formed by reaction of sulfur dioxide and traces of oxygen during start-
up–shut-down, lead to rapid deactivation of the catalyst by formation of alumi-
num sulfate. Catalysts based on titanium dioxide are also available that are not
sensitive to oxygen and hydrolyze more than 90% of the COS–CS2 formed in the
combustion chamber, which would otherwise require additional measures in
downstream off-gas purification plants (61,62).
To achieve a high degree of sulfur recovery in a Claus plant, an exact
amount of combustion air is required. This amount cannot be calculated precisely
in advance because side reactions, such as the formation of hydrogen, carbon
monoxide, carbon disulfide, and carbonyl sulfide, also have some effect on the
optimal air quantity.
386 NATURAL GAS Vol. 1
Fig. 18. Sulfur recovery requirements for Alberta sour gas plants (August 1988).
All processes in this group can be designed so that sulfur recoveries of 99.5%
are attainable. With absorption processes, a recovery of > 99.7% is possible.
Processes with Catalytic Oxidation to Sulfur Dioxide. This group
encompasses processes in which all sulfur-containing components are oxidized
to sulfur dioxide. After cooling, sulfur dioxide is removed by absorption. The sul-
fur dioxide-rich gas arising from regeneration of the laden solvent is recycled to
the Claus plant. The Wellman–Lord and Solinox processes, in which a sulfur
recovery of >99.5% can be achieved, is commercially important.
Incineration. With only a few exceptions, the off-gas from Claus plants or
tail-gas cleanup plants must be fed to an incinerator in which the residual
sulfur-containing components are oxidized to sulfur dioxide. For thermal
incinerations 600–8008C is necessary, whereas catalytic plants require only
300–3508C. All these processes lead to an off-gas with <10 mg/m3 of hydrogen
sulfide. If the off-gas to be treated contains other sulfur compounds such as car-
bonyl sulfide or carbon disulfide, then a value of 10 mg/m3 can also be obtained
in a thermal incineration plant, provided the temperature is sufficiently high.
For the catalytic process in this case, however, up to 80 mg/m3 of sulfur must be
expected in the off-gas. In principle, energy recovery by waste-heat boilers or
gas–gas heat exchangers is possible for both processes. Nevertheless, traces
of sulfur trioxide can lead to condensation of sulfuric acid in the heat exchan-
ger. To avoid corrosion, heat exchange must be limited which means that less
energy can be saved (70).
2.8. Example of an Integrated Plant. A flow diagram of an operating
plant for the purification, dehydration, and conditioning of two different raw nat-
ural gases is shown in Figure 19. The design capacity amounts to a maximum of
14 106 m3/d for raw gas 1 and 5 106 m3/d for raw gas 2. Characteristic data for
the units processing raw gas are presented in Table 10.
388 NATURAL GAS Vol. 1
Production of sulfur is set at 2100 t/d. The sour gases are predehydrated in
the field (see Fig. 5) and are low in higher hydrocarbons. Treatment is thereby
reduced to the removal of sulfur components and the adjustment of the inert-
gas content. This special plant design has been chosen because of the following
criteria (71):
The high degree of carbonyl sulfide removal, which is necessary for under-
ground storage of the pure gas, can be achieved with the chemical–physical Sul-
finol scrubbing process. The latter is unselective but, depending on the ratio of
hydrogen sulfide to carbon dioxide in the crude gas, may provide a sufficient
Claus gas quality. The high energy consumption of solvent regeneration is
also taken into account. The high-calorific pure gas can be stored in a nearby
underground storage, but, after addition of nitrogen, can also be fed directly
to the sales pipeline as low-calorific gas. The gas must be dehydrated because
a water-containing scrubbing agent is used. The three-stage Claus plant attains
a sulfur recovery of 97.5%; the total recovery is increased to 99.8% in a down-
stream tail-gas cleanup plant using the Scot process. After degassing, the sulfur
produced is stored as a liquid and transported by rail. Sometimes the sulfur is
solidified to prills or pellets outside the plant. Steam generated in a Claus plant
is used in the reboiler for heating the Sulfinol solvent. Expansion turbines are
employed for depressurization of the rich Sulfinol, which produce surplus elec-
tricity. In many cases the energy recovered from expansion, heat exchangers,
and Claus waste-heat boilers is sufficient for regeneration and running the
pumps. The total energy requirement of the sour gas treatment described
here is ca. 3% of the feed gas. One of the greatest consumers of energy is the
tail gas purification plant.
Degradation of the DIPA content of the solvent to oxazolidone occurs
because of the high carbon dioxide content of crude gas. The oxazolidone is regen-
erated to reusable scrubbing agent in a secondary plant. In another secondary
plant a sulfur-dissolving agent, which prevents sulfur plugging of the wells, is
regenerated for further use.
Lean raw gas 2, which contains a low level of hydrogen sulfide, is trea-
ted in a Purisol scrubber. Carbon dioxide remaining in the raw gas helps in
Vol. 1 NATURAL GAS 389
adjusting the calorific value in the total pure gas stream from the Sulfinol
unit. The flash gases are so low in hydrogen sulfide that they can be burned in
a boiler.
Auxiliary Plants. Natural gas treating plants are connected to the public
energy distribution network. Reliable auxiliary plants are required to guarantee
a high level of availability and independence.
Steam Generation. Waste heat from the Claus process is used in the pro-
cessing plant to drive process pumps and turbine-driven generators, and to heat
solvents. The high-pressure steam generated is first expanded in turbines. The
saturated low-pressure steam obtained can be used reboiling of the solvent at
constant temperature. The reboiler steam pressure is 0.5–0.7 MPa and is
adjusted to match the thermal stability of the solvent. Waste-heat boilers in
the Claus plant operate at steam pressures of 1.6–4.0 MPa; the sulfur conden-
sers at 0.1–0.6 MPa. By proper choice of driving equipment and steam pressure,
energy requirements can be balanced so that only a little external energy is
needed. Small amounts of blow-down streams must be continuously removed
from boilers and condensers to avoid corrosion. These streams must be replen-
ished by water from a treatment plant and a boiler feedwater degassing plant.
Electrical Power. Large compressors, circulating pumps, and blowers are
often driven by steam turbines to achieve a better control of their speed. Fre-
quently, the plant’s own power station is only used for start-up and shutdown,
whereby both steam- and gas turbine-driven generators are employed.
Inert Gas. Occasionally, purified natural gas must be mixed with inert
natural gas to meet specified quality standards. When the admixing of air is
not possible because of limitations of the oxygen content, inert gas plants are
required. Inert gases produced from both catalytically purified flue gases and
air separation are used. The dry gases are generally produced at low pressure
and brought to pipeline pressure (ca. 10 MPa) by multistage compressors. Inert
gases are also employed for safe start-up and shutdown of plants and for regen-
eration of the Claus catalyst (72).
Tank Storage. Both higher hydrocarbons and sulfur are stored in tanks as
liquids. They are transported further by road, rail, and frequently, pipeline. With
special thermal insulation, liquid sulfur transportation through pipes of more
than 8 km in length is possible without additional heating (73).
Wastewater. Natural gas treating plants produce little noxious waste-
water. The boiler- and cooling tower blow-down water is fed to public sewage
systems. Waste waters from piping and equipment cleaning are collected and
treated in a reclaimer. Remaining residues are injected with the salt-containing
formation water in separate disposal wells.
Gas Mixing. Different field qualities are mixed in the so-called raw gas
manifold, so that the absorbers can be charged with gas of constant composition.
Special separators minimize the entrainment of liquids and impurities that
would damage the solvent or cause foaming. Pipelines from the field can be
charged from the manifold with pigs for cleaning purposes. Mixing of different
pure gases to adjust the Wobbe index and the net calorific value is possible in
the pure gas manifold.
Reclaimer, Filter. Through side reactions, chemical scrubbing solvents in
particular become enriched with degradation products, frequently of unknown
390 NATURAL GAS Vol. 1
Fig. 19. Flow diagram of a natural gas treating plant in the Federal Republic of
Germany. a) Sour gas manifold; b) Town gas manifold; c) Disposal well. – – –
Steam;– – – – Condensate.
nature. These substances are corrosive and may cause foam formation in the
absorbers. Solvent quality is stabilized by removing a small side stream and
addition of fresh solvent. This side stream is processed in a reclaimer, which is
usually an auxiliary distillation plant. Purified absorbent can be reused in the
process (74).
With physical scrubbing agents, removal of undesired substances (eg,
hydrocarbons) is frequently achieved by addition of water and phase separation.
Mechanical filters are used in a side stream.
Special Construction Features. High partial pressures of carbon dioxide
and hydrogen sulfide in a wet environment lead to serious corrosion. In particu-
lar, the stress corrosion cracking (SCC) caused by hydrogen sulfide corrodes
Vol. 1 NATURAL GAS 391
high-strength steel from inside: atomic hydrogen can be formed by acid corrosion
in connection with water, and can diffuse into the steel because of its small size of
the atom. The hydrogen atoms recombine in the area of microcracks and local
inhomogenities; the resulting pressure, which can reach 105 MPa, leads to cracks
and blistering (hydrogen-induced cracking).
High-alloy hard steels are particularly susceptible; the necessary construc-
tion and testing regulations must be adhered to. Consequently, the carbon
content of the steel is limited. Equipment and pipeline welding joints must be
annealed. A hardness HRC of <22 must be maintained. Excessive stresses
( particularly at plate edges) and thickness changes which are caused by the fab-
rication techniques must be avoided. Specially designed forged nozzles based on
finite element analysis are customary (75).
3. Liquefaction
Common to all natural gas liquefaction processes is the need to pretreat the
gas feed for removal of components that would freeze out (thus restricting the
flow of the process), cause corrosion, or lead to pollution upon combustion of
the revaporized LNG (see Chap. 2).
Baseload Liquefaction Processes. Large-capacity or baseload liquefac-
tion of natural gas is performed with emphasis on process efficiency. The scale
of operations means that production with lower installed power capacity and
lower fuel consumption is economically favored. Because the heat that must be
removed from natural gas to cool it to 1608C is eventually rejected to ambient
air or water, rather elaborate systems have resulted.
Cascade Refrigeration Processes. The earliest liquefaction processes
employed cascades of single-component refrigerants in series (see Fig. 20).
Each refrigerant is established as a separate closed-loop refrigerator that
supplies refrigeration at discrete temperature levels. Typically, propane, ethy-
lene, and methane are used to provide a wide, balanced range of refrigeration.
After compression (a) three temperature levels for each of the three refrige-
rants form a nine-stage cascade. Each of these temperature levels corresponds
to a preset pressure letdown (d) for evaporating the refrigerant in heat
exchange with the natural gas feed and a separate refrigerant stream that
requires cooling (e). In this fashion, heat is removed from the natural gas at
successively lower temperatures; ie, the refrigerant is boiled at successively
lower pressure. Heat is rejected to ambient air or water via the warmest
refrigerant, generally propane, and compressor aftercoolers (b). The methane
refrigerant loop is open in that it is combined with the natural gas feed and after
the final pressure letdown the liquid methane forms part of the LNG product.
Cascade processes allowed use of single-component refrigerants at a time
when thermodynamic correlations and thermophysical data bases were not as
well developed as they are at present. In addition, the processes could be made
very efficient (ie, the amount of irreversibility could be reduced) by increasing
both the number of refrigerants employed and the number of boiling stages
for each refrigerant. This improvement in efficiency, however, had significant
cost penalties. Each refrigerant requires its own compressor, compressor
driver, vessels, and heat exchangers, together with attendant piping, insulation,
and instrumentation. Each additional stage of boiling adds to the number
of heat exchangers and vessels and to the number of sidestreams entering the
compressor.
All-Mixed-Refrigerant Processes. The problems of complex layout,
relatively high capital cost, and limited train capacities inherent in cascade pro-
cesses were addressed by the mixed-refrigerant processes that succeeded them.
As equipment and analytical capabilities advanced, the refrigerants could be
combined in a single refrigeration loop (see Fig. 21). In such processes, a single
refrigerant mixture of pentane, butane, propane, ethane, methane, and nitrogen,
for example, is constituted to match the cooling curve of the natural gas to be
liquefied; ie, the temperature–enthalpy warming curve of the mixed refrige-
rant closely tracks the natural gas cooling curve in (g). This reduces the irrever-
sibility of the process. The same refrigerant is then condensed, a part against air
or water at (d) and the balance at lower temperature in heat exchange against
itself in (f). In the latter case, incoming gaseous refrigerant is cooled and con-
densed at elevated pressure in countercurrent flow against evaporating liquid
refrigerant fractions, each flashed (h) to a much lower pressure and, hence,
temperature. The all-mixed-refrigerant process has performed well in a number
of plants and represents a simplification compared to the cascade process. How-
ever, it is not thermodynamically efficient enough to be economical in the face of
rising energy prices. To match the cooling curve of the natural gas feed over the
wide range from the temperature of cooling water or ambient air down to the
liquefaction temperature, compromises in the mixed-refrigerant composition
are necessary. The wide range of boiling points for the refrigerant components
also means that some of the heavier components are compressed to higher pres-
sures than actually required for their condensation to ensure condensation of the
lighter, lower boiling components such as nitrogen and methane. Such a recom-
pression penalty cannot be avoided without a certain degree of separation of the
refrigerant components such as occurs in precooled mixed refrigerant processes.
Precooled Mixed-Refrigerant Processes. In the early 1970s, a third
generation of processes, precooled mixed-refrigerant processes, developed from
direct combination of the other two. The most widely used process employs two
separate refrigeration systems, a propane cascade refrigerant loop in series with
a mixed-refrigerant cycle that incorporates propane, ethane, methane, and nitro-
gen as components (89) (Fig. 22 shows the cooling curves for this process). The
propane loop precools the natural gas and serves as an intermediate refrigerant
Fig. 22. Cooling curve for a propane precooled mixed-refrigerant liquefaction process.
a) Cooling curve of natural gas; b) Propane cascade; c) Mixed refrigerant (stage 1);
d) Mixed refrigerant (stage 2).
396 NATURAL GAS Vol. 1
for heat rejection from the mixed-refrigerant section to cooling air or water. The
propane section can be fabricated from lower cost carbon steel, whereas the lower
temperature sections require aluminum or nickel steels. Thus breaking the cool-
ing process at the propane stage is compatible with economical material selec-
tion. Finally, by reducing the range of cooling that must be achieved by the
mixed refrigerant, its composition can be optimized and energy losses due to
recompression significantly reduced.
The increase in worldwide demand for LNG has prompted an increase in
the capacity of liquefaction trains in order to take advantage of the economies
of scale. This is being achieved by improvements in the liquefaction process as
well as increases in the equipment capacity. For the most widely used processes,
capacity increases of refrigerant compressors, gas turbines, and cryogenic heat
exchangers have led to train capacities of nominally 5 million metric tons
per year.
For larger train capacities, a hybrid process cycle (known commercially as
AP-XTM) has been developed (90,91). In this process, a cycle similar to the
precooled mixed-refrigerant process is used to precool and liquefy the LNG
(see Fig. 23A). However, the liquefied LNG is subcooled in a nitrogen refrigera-
tion closed-loop process cycle (Fig. 23B). Nitrogen gas is compressed, cooled to
near ambient conditions with cooling water or ambient air, and then further
cooled to cryogenic conditions by expansion to lower pressure. The gaseous nitro-
gen is then used to subcool LNG, after which it is returned to be recompressed,
completing the refrigeration cycle. By employing the nitrogen refrigeration cycle
to subcool the LNG, the mixed-refrigerant only has to cool the LNG to about
1158C (versus 1508C to 1608C for the typical precooled mixed-refrigerant
process). This allows for a larger production capacity without a substantial
increase in equipment size.
Additional process schemes have been developed that fall in the category of
precooled mixed-refrigerant processes. If ethane is added to the precooling refrig-
erant to form a dual mixed-refrigerant process, then the precooling temperature
can be lowered in accordance with the operating conditions of the loop and the
proportion of ethane added (92). With this process the refrigeration duty can
be shifted between the two refrigeration loops – a capability that may be useful
in the face of changing feed gas conditions or changing power availability. The
latter can occur because of ambient temperature changes that affect the power
output of gas turbine drivers. A binary precooling refrigerant is somewhat
more complicated to operate because simple pressure control is no longer ade-
quate, but it may be preferred in certain scenarios.
Another precooled process has been proposed for plants that employ gas
turbine drivers in the mixed-refrigerant section. In this configuration, waste
heat from the gas turbines is used to separate ammonia and water in an
ammonia absorption refrigerator (93). This system could provide precooling
duty while eliminating the need for costly compressors and drivers. This system
has not been employed in an operating facility.
Peak-Shaving Liquefaction Processes. Peaking-shaving LNG plants
differ from baseload facilities in several significant aspects that affect plant
design. Peak-shaving plants are much smaller, operate only seasonally, and
are often located near the point at which the pressure in gas trunk lines is
Vol. 1 NATURAL GAS 397
decreased and the lines branch into lower pressure local gas distribution
systems. The design emphasis for peak shavers is thus on capital cost minimiza-
tion rather than high thermodynamic efficiency. All-mixed-refrigerant liquefaction
cycles have therefore been employed in a large proportion of peak-shaving faci-
lities. If a local distribution pipeline is available at a pressure substantially
below that of the main gas trunk line, expander processes can be utilized to take
advantage of the pressure difference as shown in Figure 24. Several variations
exist, but the principle of these plants is to expand the inlet gas nearly isentropi-
cally through a turbine, thus rapidly lowering the temperature and partially lique-
fying the gas. The liquid is sent to storage, and the residual gas is compressed in a
compressor mechanically linked to, and driven by, the expander. This gas is then
sent to the lower pressure pipeline for distribution outside the plant.
3.4. Storage. Storage tanks for LNG are an important part of both base-
load and peak-shaving LNG facilities. In addition, they constitute the major por-
tion of the investment in import receiving terminals. Because of the high cost of
these units and their importance in the overall safety of LNG facilities, a great
deal of attention has focused on LNG tank design.
One of the earliest efforts in tank design was directed at proper material
selection. The Cleveland, Ohio tank failure in 1944 was attributable to the use
of 3.5% nickel steel which became embrittled at LNG temperature. Subsequent
to this, large-scale programs have proved the suitability of 9% nickel steel, stain-
less steels, and certain aluminum alloys (5000 series) for LNG tank fabrication.
Designs with the latter have been somewhat limited in size, because the coeffi-
cient of thermal expansion for aluminum is approximately double that for steel;
398 NATURAL GAS Vol. 1
in large tanks, such thermal movement during cooldown could result in tank
failure.
Tank designs have also evolved as more sophisticated safety analysis has
been applied to LNG installations (94). Early designs featured an internal cryo-
genic liquid tank enclosed by an outer tank that contained the insulation system
for the inner one. In some designs the outer tank contained gaseous nitrogen and
was, in turn, connected to a variable-volume or bladder tank, which compensated
for changes in the nitrogen volume due to changes in ambient temperature while
avoiding pressurization or depressurization of the outer tank. In other designs,
the roof of the inner tank was not gastight but merely supported insulation, and
the outer tank served as a natural gas holder (see Fig. 25). In both designs, the
outer tanks were fabricated of carbon steel and surrounded by a low dike to con-
tain any LNG spills.
Analytical studies indicated that the prime safety risk with an LNG spill is
the formation of a large product vapor cloud that can drift, ignite, and cause
widespread damage. Subsequent designs incorporated outer tank materials
less prone to failure at cryogenic temperatures and taller dikes built closer to
the tanks. These measures lead to less free surface for any LNG spill from a
tank failure and therefore a lower feed rate to the ensuing vapor cloud.
Additional studies reviewed the consequence of an external impact, such
as a crashing aircraft, leading to tank failure and the effect of a catastrophic
inner tank failure on outer tank containment. These studies all focused on the
need for much more secure outer containment. Resultant tank designs have
double-integrity tankage; ie, a liquid spill from a failure of the inner tank is con-
tained by a second concentric tank that is structurally independent of the first
(95). The outer tanks have been built of reinforced concrete with a stainless
steel liner. Finally, full-height earthen berms have been built to protect the tank-
age from external damage and to contain the postulated asymmetric loading on
the outer tank resulting from catastrophic failure of the inner tank. Where the
site is too small to permit full development of earthen berms, inground storage or
additional reinforced concrete outer containment is employed.
3.5. Transportation. Peak Shaving. Peak-shaving LNG facilities typi-
cally vaporize LNG from storage and send it out via a local distribution network.
In some installations, however, LNG is merely stored and vaporized for send-out.
These satellite facilities receive LNG from other production sites, which is trans-
ported in over-the-road trailers. The trailers are similar to those employed in
transporting liquid nitrogen and oxygen, and consist of an inner cryogenic vessel
and an outer protective vessel of noncryogenic material. The space between
vessels is typically filled with expanded perlite insulation and evacuated to sub-
atmospheric pressure to further decrease heat in-leak.
Baseload. Baseload LNG facilities are linked to their corresponding
receiving terminals by fleets of large oceangoing tankers. Modern tankers typi-
cally can hold 125 000 m3 of LNG. Each tanker employs one of several contain-
ment systems, which can be divided into three principal types. One design
employs free-standing spherical tanks supported on cylindrical skirts (see
Fig. 26). Each tank is insulated and located between bulkheads in a separate
section of the ship. These compartments are blanketed with inert nitrogen gas,
which is sampled periodically to detect any LNG leakage.
A second family of designs is the membrane tank. In ships using this type,
the insulation and a secondary barrier are applied to the hull of the ship (Fig. 27).
The insulation also transmits the load from the LNG cargo to the hull and its
Fig. 26. Spherical ship tank. a) Weather cover; b) Tank; c) Insulation; d) Support skirt;
e) Thermal break; f ) Inner hull; g) Ballast tank; h) Outer hull.
400 NATURAL GAS Vol. 1
Fig. 27. Membrane ship tank. a) Membrane liner; b) Insulation; c) Inner hull; d) Ballast
tank; e) Outer hull.
and initially form a dense, low-lying cloud. As the cloud mixes with air and is
warmed by its surroundings, it begins to rise. Wind serves to add a horizontal
component to the cloud’s motion. As air mixes with the natural gas, the mixture
becomes flammable (ie, local compositions are between the lower and upper
flammability limits of LNG). The flammability limits vary with LNG composi-
tion, particularly the proportion of propane. This cloud is then potentially hazar-
dous to areas beyond the battery limits of the LNG facility.
A number of large-scale tests have demonstrated certain characteristics of
LNG vapor cloud fires. These fires are deflagrations-rapid combustions in which
the flame front that moves through the cloud is preceded by a weak, decoupled
shock wave (98). An accompanying radiation hazard exists, but not the pressure
damage of a shock wave that occurs when mixtures of other hydrocarbons and air
are detonated.
More than 50% of world gas reserves are located in remote areas. For example,
part of the gas used in Western Europe is produced in the harsh permafrost
environment of Siberia or in the North Sea.
In most cases, producers ship the gas from the fields to the borders of the
countries in which it is used. Importers buy the gas at these points under long-
term contracts and resell the fuel to local utilities as well as to industrial users
and power stations connected directly to the transmission system. Residential
and commercial consumers are normally served by the local distribution
companies. The European natural gas transmission system is shown in Figure 28.
Natural gas is used chiefly for heating. Therefore, gas demand varies sub-
stantially between winter and summer, weekdays and week-ends, or day and
night. The ratio between summer and winter loads in Europe is between 1 : 5
and 1 : 10. Production, transmission, storage, and distribution facilities must be
designed and sized to handle these load variations.
Of the gas traded across international borders, 75% is shipped by pipelines,
and 25% is carried by LNG tankers.
Gas field development and the construction of transportation systems from
distant areas of production to gas-importing nations are highly capital-intensive
projects. Thus, the load factors at which gas is imported are nearly always very
high. For load equalization, gas is stored in underground storage facilities during
the off-peak season and produced from storage during periods of peak demand in
winter. Pipeline pressurization (line packing) in the transmission and distribu-
tion systems themselves, as well as peak-shaving installations, also help to han-
dle load variations.
4.1. Transportation. Pipeline Transmission. Transportation of large
gas volumes is best achieved by large-diameter pipelines operating at high pres-
sure. Currently, the pipeline diameter may be up to 1400 mm and working pres-
sure up to 8 MPa. Such pipelines carry gas at a flow rate of 2 106 m3/h (STP)
over a distance of ca. 1000 km. This capacity is insufficient, however, to ship
gas from distant fields to markets. Compressor stations must therefore be built
to increase the pressure of the gas. Gas compressors are normally driven by
402 NATURAL GAS Vol. 1
turbines or engines fueled by their own pipeline gas for reliability and lower cost.
The number of compressors required for a given carrying capacity depends on the
distance over which the gas must be shipped and on the p compression ratio,
which is given by
po
p¼ (1)
pi
where po is the maximum compressor outlet pressure (bar) and pi the minimum
compressor inlet pressure (bar).
As the compression ratio increases, the number of compressors needed to
operate a gas transmission line decreases. Hence, transmission engineers must
seek to optimize capital outlay and follow-up operating expenses for pipelines
and compressor stations.
For large-diameter pipelines, the optimum compression ratio is generally
relatively low; it is between p ¼ 1.25 and p ¼1.4 (99). Offshore pipelines are
currently built with diameters up to 1000 mm. Pipelines from the North Sea to
continental Europe are 1000-mm-diameter lines that are laid at a water depth of
Vol. 1 NATURAL GAS 403
150 m; Italy and North Africa are connected by 500-mm lines that cross 600-m-
deep water. The very high working pressures at which offshore lines can be oper-
ated partly make up for the loss in capacity due to the smaller diameters. If
necessary, pipeline bundles are laid (100).
Offshore compressor stations are extremely costly structures because they
must be built on platforms. For this reason, the optimum compression ratio of
offshore pipelines is higher than that of onshore pipelines. Depending on flow,
it may be as high as p ¼ 4.
LNG Transportation. Although the energy input required for liquefac-
tion of natural gas is substantial, the volume advantage makes liquefaction
economically viable. Cryogenic LNG is shipped by LNG tankers at atmospheric
pressure. The cargo tanks are insulated to minimize evaporation. Neverthe-
less, the boil-off rate is between 0.2 and 0.25% of the cargo per day. The gas
that evaporates is used to fuel the ships engine. Reliquefaction is presently
not viable.
The transportation of LNG is often the only way of bringing gas from dis-
tant gas-producing countries to market outlets. Any cost comparison between
LNG transportation and pipeline transmission must, of course, be geared to
the circumstances of the project. Generally speaking, an LNG chain is the only
answer if pipeline transmission is not feasible for technical or other reasons or if
the distance is sufficiently long. The cost of LNG transportation is lower than
that of offshore pipeline transmission even for distances of several hundred
kilometers, whereas onshore transmission is nearly always less expensive
than LNG transportation unless the distance is extremely long (99,101–103). A
comparison of costs of LNG transportation and pipeline transmission is given in
Figure 29.
Gas Transmission Pipeline Design. Pipelines for gas transmission must
be designed so that, under peak load conditions, gas is allowed to flow from the
point of delivery at the inlet of the system to the outlet of the system at a pres-
sure between the maximum and minimum allowable operating pressure. Gener-
ally, pipeline capacity is somewhat lower in summer than in winter because of
the following reason: The atmospheric temperature increases soil temperature
Fig. 29. Cost of LNG transportation and cost of pipeline transmission as a function of
transportation distance. a) Offshore pipeline (900-mm diameter); b) LNG transportation
with liquefaction onshore (125 800 m3); c) Onshore pipeline (1000-mm diameter).
404 NATURAL GAS Vol. 1
and hence gas temperatures, and power output from the compressor engine
decreases because of the higher temperature of the intake air.
According to DARCY and WEISBACH, the pressure drop in a gas transmission
pipeline is given by the following equation:
16 l T
p21 p22 ¼ f % q2 pn K (2)
p2 d5 n n Tn
l
p21 p22 ¼ 4:6401 1015 f Q2n %n T K (3)
d5
where Qn is the hourly flow rate under normal conditions (m3/h); p1 the absolute
pressure at the pipeline inlet (bar); and p2 the absolute pressure at pipeline
outlet (bar).
Equation (3) may be rewritten using the following terms
where
l
c ¼ 4:6401 1015 f % T
d5 n (5)
cðbar2 h2 m6 Þ
According to NIKURADSE, the pipe friction factor f for turbulent flow and
rough-surface pipes is given by (104).
1
f ¼ (6)
ð2 log dk þ 1:14Þ2
The limit between hydraulic roughness and the transition regime is given
by
2
200 d
fl ¼ (8)
Re k
Q2n %n
Re ¼ 353:678 (9)
d Z
The gas flow deviation factor K describing the real-gas behavior of natural
gases is another important term in flow equations. For most natural gases, K
may be approximated by
K ¼ Kð pm ; Tm Þ
Qn pffiffiffiffiffiffiffiffiffiffiffiffi
¼ n þ 1 100% (11)
Q0
0 100
1 50 21 141
2 33/67 13/31 173
3 25/50/75 9/21/38 200
4 20/40/60/80 7/16/26/42 224
QL 1
¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 100% (12)
Q0 13 L 4 L1
where QL is the capacity after addition of a loop line of length L1 (m3/h); Q0 is the
capacity without looping (m3/h); L is the length of pipeline between inlet and
outlet (m); and L1 the length of the loop line (m). The length of the loop line
required to provide a specified additional capacity depends critically on whether
or not gas is tapped from the pipeline system between the inlet and the outlet
points (105).
The increase in capacity as a function of loop line length is plotted in
Figure 30. If a loop line of the same diameter is assumed, 75% of the main line
must be looped to increase capacity by 50% if all gas flows from the inlet to the
outlet of the transmission system. If gas is taken off regularly between the inlet
and the outlet (as in the case of a regional line), only 40% of the original pipeline
has to be looped.
Transient Flow in Gas Transmission Systems. Flow rate and pressure
in pipelines are a function not only of distance but also of time. Therefore, line
packing effects play an important role. Simulation of transient flow, of course,
produce different results than steady-state simulation. In a typical gas system,
for instance, an hour of peak demand is usually preceded by a phase of low
demand. As pressure increases when demand is low, a large gas volume is
packed in the line when the demand peaks. During the period of peak load off-
takes, the pressure in the line decreases because of the pressure drop resulting
from the higher gas flow velocity through the line. Part of the gas packed in the
line is therefore discharged from the system, covering peak demand.
Figure 31 shows the flows through a pipeline system into which gas is deliv-
ered at location 1 and from which gas is taken off at location 2 at different load
factors. If the flow rate at the pipeline inlet (location 1) is constant (Fig. 31 A),
Vol. 1 NATURAL GAS 407
Fig. 30. Increase in carrying capacity as a function of loop length. a) Pipeline without
offtake; b) Pipeline with offtake.
inlet and outlet pressures will vary considerably and line packing will contribute
substantially to covering peak load demand. As gas is normally delivered into a
pipeline system at a continuous rate, the line will be repacked during periods of
low demand, for example, during the night. If delivery into the pipeline system is
at constant pressure (Fig. 31 B), the delivery rate will vary. However, the
response at the point where gas is delivered into a pipeline system is always
delayed with respect to the system behavior at the point where gas is taken off
from the system; for this reason, delivery into the system and delivery from the
system never coincide exactly.
4.2. Storage. Gas storage facilities are integrated into the pipeline sys-
tem to match the high load factor at which gas is produced and delivered by sup-
pliers to the pattern of end-user demand, which varies between winter and
summer, over the week, and from hour to hour.
Determination of Storage Requirements. Gas demand is influenced by
many different physical and economic parameters. Because substantial quanti-
ties of gas are used for space heating, ambient temperature is the most critical
one. Typical average daily deliveries Qdi in Northwestern Europe as a function of
temperature are shown in Figure 32. Gas demand increases as a linear function
of temperature decrease and is constant above a certain limit temperature. The
correlation is represented by the following equation:
where Qdi is the daily gas demand at the daily average temperature ti (kW h/d);
Qdg the daily baseload gas demand (kW h/d); tl the limit temperature for space
heating (8C); ti the average temperature on a given day (8C); and m the increase
in deliveries per 8C decrease in temperature (kW h d1 8C1). This equation
allows sufficiently accurate predictions of daily gas offtakes (100).
Transmission, storage, and distribution facilities must, of course, be sized
for the design day peak load on the system, determined by Equation (13). The
ratio between system capacity specified for peak load conditions and capacity uti-
lization is a crucial commercial and economic factor. Utilization time or system
load is expressed differently in various countries.
408 NATURAL GAS Vol. 1
Fig. 31. Transient flow in gas transmission systems. A) Constant flow rate at pipe-
line inlet; B) Constant pressure at pipeline inlet
In the United States and the United Kingdom, the load factor expresses the same
relationship. It correlates with the quantities given in Equations (14) and (15) as
follows:
Annual volume
Load factor ¼
8760 Peak hourly flow rate
Vol. 1 NATURAL GAS 409
Fig. 33. Daily load curve showing gas delivered from storage and gas added to storage.
1) depleted gas fields, oil fields, and aquifer storage facilities (porous rock storage);
2) salt cavern storage facilities; and
3) mined cavern storage facilities (only of minor importance).
Table 12. Suitability of Various Storage Systems for Different Peak Needs
Peaks
Fig. 34. Storage capacity required for seasonal peak shaving as a function of the supply
pattern.
Underground storage facilities hold both working gas and cushion gas. Working
gas is the gas that may be produced from storage for peak-shaving operations.
Cushion gas, sometimes referred to as base gas, is the gas required to
ensure sufficient reservoir pressure for recovery of the working gas. It provides
the pressure needed to cycle the normal working storage volume.
Porous Rock Storage. Gas can be stored below ground in rock pores.
The rock should have sufficient porosity and permeability, and the storage hor-
izon (which must be at sufficient depth) should be covered by tight caprock and
be a structure that entraps gas injected into the rock. Structures used for porous
rock storage include depleted oil and gas fields as well as aquifer structures.
Development of a gas storage facility in a depleted gas field does not pose a
major risk because data on rock properties and information on the tightness of
the caprock are available. The working gas capacity, the maximum reservoir
pressure, and the delivery rates as a function of reservoir pressure are known
from production history. For the storage of gas in depleted oil fields, gas–oil
separators are needed; however, secondary oil production by gas injection often
makes a depleted oil field a commercially attractive storage site (107).
In aquifer storage reservoirs, interstitial water in the pores is displaced by
the gas injected at high pressure. Considerable geological and reservoir engi-
neering effort is required to decide whether an aquifer structure is suitable for
underground storage of gas: ie, the structure of the potential storage horizon
must be evaluated; the thickness, porosity, and permeability of the storage
rock must be determined; and the thickness of the caprock must be verified.
Salt Cavern Storage. Gas may also be stored in caverns leached into
salt rock. Unlike the volume of a natural storage rock reservoir, the volume of
these cylinder-shaped caverns can be controlled within certain limitations. The
diameter of such a cavern may be as large as 80 m and its height 400 m. The
cavern volume can reach 500 000 m3.
The capacity of cavern storage facilities can be adapted to increased storage
needs by adding further caverns in a given formation as required. The cushion
gas volume required for operation of a storage cavern is much lower than the base
gas volume in a porous rock reservoir and may amount to only 20% of total capa-
city. Storage caverns allow the production and injection of gas at very high rates
that are limited chiefly by the capacity of the above-ground installations.
412 NATURAL GAS Vol. 1
such a nonlinear model, which originally consisted of very few nodes in view of the
limited performance of early computers. As data-processing technology progressed
during the 1950s, network analysis techniques become more widely used. They are
now state of the art and are employed as a matter of course in distribution system
design. Modern software models all the details of distribution systems.
Many of the input data required for network analysis can be determined
only by statistical methods and field tests. For example, statistics are necessary
to predict system loads controlled by the demand of numerous gas users whose
meters are usually read only once each year. Pipe roughness, which is crucial for
the hydraulic behavior of the grid, must be measured in many cases because pipe
properties depend on the year in which the distribution line was constructed.
Distribution networks must also be designed to handle peak loads that can
be 10 or 15 times higher than the baseload. To optimize the economics of gas dis-
tribution, distribution system capacity is regularly adapted to growing demand.
Methods of increasing capacity include the following:
Fig. 36. Optimum production of gas from porous rock storage and salt cavern storage for
peak shaving.
5. Economic Aspects
Natural gas is the world’s third most important energy source, behind oil and
coal, covering ca. 23% of the 1995 world energy demand.
Reserves. Proven world gas reserves at the end of 1995 totaled about
150 1012 m3(Vst ie, standard cubic meter of natural gas at 158C and 101.3 kPa
with a net calorific value of 35.2 MJ or 9.776 kW h) (equivalent to 5171 1018 J
or 134 109 t of oil), corresponding to ca. 95% of proven oil reserves (113). A geo-
graphical breakdown is shown in Table 13. These reserves equal about 55 years of
current annual gross production.
In addition, ca. 222 1012 m3(Vst ) (equivalent to ca. 7770 1018 J or
200 109 t of oil (114)) has been identified as probable gas reserves. Including
this figure, the total 1995 gas inventory equals approximately 135 years of cur-
rent annual gross production.
Production. The world’s 1995 gross gas production amounted to
2.75 1012 m3(Vst) (95.15 1018 J or 2.4 109 t of oil), of which about 80% was
marketed. The rest covers reinjection, flaring, losses, etc. The total 1995 mar-
keted volume of 2.21 1012 m3(Vst), broken down into major geographic zones,
is shown in Table 14.
Vol. 1 NATURAL GAS 415
Fig. 37. Pipeline diameter as a function of working pressure for constant pipeline
capacity (in relation to pipeline diameter for 2.5-kPa working pressure).
Table 13. World’s Proven Natural Gas Reserves as of January 1, 1996 (113)
Area 1012 m3(Vst) 1018 J %
North America 6.5 222 4.3
Latin America 8.0 277 5.4
Western Europe 6.3 220 4.3
Former Soviet Union, 59.2 2037 39.3
Eastern Europe
Africa 10.0 346 6.7
Middle East 45.3 1558 30.1
Asia, Oceania 14.8 512 9.9
Total 150.2 5171 100.0
416 NATURAL GAS Vol. 1
Table 14. World Natural Gas in 1995 [109 m3(V ST)] (113)
Marketed
Area production Consumption
Natural gas analysis can be divided into two areas: chemical analysis of the com-
position, and determination of the physical properties. Chemical analysis meth-
ods for natural gas are summarized in (115,116). Instructions for both the
chemical analysis and determination of physical properties for natural gases
are given in (116).
6.1. Analysis of Natural Gas Composition. The most important
method for chemical analysis of natural gas is gas chromatography (117–123).
Sampling must be carried out before the gas can be analyzed.
Sampling. Glass equipment is mainly used for sampling because glass is
not attacked by any component of natural gas and does not affect the composition
of the gas (124). Vessels with two taps (see Fig. 38) are used, which must be com-
pletely dry. The gas to be analyzed is first passed through the sampling vessel to
completely displace the air from it. Both taps are then closed so that the sample
is contained in the vessel. The sample should always be under slight overpres-
sure, so that no air is drawn into the vessel even if the taps are not closed
tightly.
Table 15. International Gas Trade in 1995 [109 m3 (V ST)]
Imports to
Supplies from Japan Germany United States France Italy United Kingdom Belgium Others Total
417
Malaysia 11 2 13
Brunei 7 1 8
United Arab Emirates 5 5
United States 2 2 4
Others 9 1 19 29
Total 57 69 80 34 35 2 12 106 395
418 NATURAL GAS Vol. 1
Fig. 39. Gas chromatograph. a) Carrier gas; b) Regulating valve; c) Manometer; d) Chart
recorder; e) Air thermostat; f ) Separating column; g) Thermal conductivity detector;
h) Flow meter; i) Three-way tap; j) Trap.
Vol. 1 NATURAL GAS 419
solution and the mercaptans in the basic solution. Hydrogen sulfide and mercap-
tans are then determined iodometrically.
Determination of Trace Components. Heavy metals, especially
mercury, are the most important trace components. There are at present no
standard determination methods; atomic absorption spectroscopy (AAS) is
usually used.
6.2. Determination of Physical Properties. Natural gases are charac-
terized and classified according to their physical properties. The Wobbe index
(W ) is the most important characteristic for the classification of natural gases
and they are divided into two groups: group L (low Wobbe index) and group H
(high Wobbe index). The Wobbe index is calculated according to
pffiffiffi
WS ¼ HS = dðMJ=m3 Þ;
where HS is the gross calorific value and d the relative density (based on air). In
addition to the more commonly used superior Wobbe index WS, the inferior Wobbe
index WI can be calculated, whereby the inferior calorific value HI is used instead
of HS. The Wobbe indices indicate gas quality for combustion.
Usually the Wobbe index is calculated from HS and d. Automated calori-
meters are available for its direct determination (Wobbe recorders) which pffiffiffi use
a gas stream that is adjusted so that the flow rate is proportional to 1= d.
Important physical properties of natural gas include the following:
Density
Calorific value
Molecular mass
Dew point
Density. Density can be measured with a gas density balance. The method
is based on the different buoyancy of a nitrogen-filled glass sphere suspended
from a balance in various test gases (131). The gas to be measured and nitrogen
Vol. 1 NATURAL GAS 421
should have the same temperature and pressure. Instruments of this type
also allow continuous measurement and recording of gas density in a gas
stream.
A particularly reliable method is to weigh a specific volume of gas in a glass
flask (eg, a Chancel flask according to IP 59). Either the density of the gas (mass
per unit volume) or the relative density d (the ratio of the density of the gas to
that of dry air) is determined. The density and relative density can also be calcu-
lated from the composition determined by gas analysis (132).
Calorific Value. The calorific value in MJ/m3 can be determined in three
ways: (1) directly by combustion of the gas at constant pressure in a gas calori-
meter, (2) by stoichiometric combustion and measurement of the gas : air ratio,
and (3) by calculation from the values obtained from gas analysis.
The principle of continuous recording gas calorimeters is combustion of a
controlled stream of gas, followed by heat exchange between the combustion
gases and a liquid, whose temperature increase is measured and recorded
(133). This method is the most widely used. Discontinuous gas calorimeters
involve combustion of a specific amount of gas and the resulting temperature
increase is measured directly.
Molecular Mass. The molecular mass of natural gas is calculated as an
average value from the concentrations determined by gas analysis and the mole-
cular masses of the components.
Dew Point. In order to avoid condensation of water, the gas tempe-
rature should not drop below the dew point of the water contained in the
gas. The dew point can be determined directly by passing a constant stream
of gas over a polished metal mirror that is slowly cooled. The dew point is the
temperature at which the first signs of condensation appear on the mirror.
These methods can be used for gases under pressures of up to 10 MPa.
Automated instruments using photocells are also employed (134). If the
gas stream contains easily condensable hydrocarbons, then these must be
washed out by passing the gas through high-boiling liquid paraffin prior to
determination.
7. Acknowledgement
REFERENCES
GENERAL REFERENCES
SPECIFIC REFERENCES
4. DIN-Norm 1871 : ‘‘Gasförmige Brennstoffe und sonstige Gase,’’ Beuth Verlag,
Berlin–Köln 1980.
5. VDI-Wärmeatlas, VDI-Verlag, Düsseldorf 1988.
6. N. L. Carr, R. Kobayashi, D. B. Burrows: ‘‘Viscosity of Hydrocarbon Gases Under
Pressure,’’ Trans. Am. Inst. Min. Metall. Pet. Eng. 201 (1954) 264–278.
7. M. B. Standing, D. L. Katz: ‘‘Density of Natural Gases,’’ Trans. Am. Inst. Min.
Metall. Pet. Eng. 146 (1942) 140–144.
8. J. D. van der Waals, Proc. Acad. Sci. Amsterdam 3 (1901) 515.
9. O. Redlich, J. N. S. Kwong: ‘‘On the Thermodynamics of Solutions. V—An Equation
of State. Fugacities of Gaseous Solutions,’’ Chem. Rev. 44 (1949) 233–244.
10. G. Soave: ‘‘Equilibrium Constants from a Modified Redlich–Kwong Equation of
State,’’ Chem. Eng. Sci. 27 (1972) 1197–1203.
11. O. Y. Peng, D. B. Robinson: ‘‘A New Two-Constant Equation of State,’’ Ind. Eng.
Chem. Fundam. 15 (1976) no. 1, 59–64.
12. R. L. Robinson, Jr., R. H. Jacoby: ‘‘Better Compressibility Factors,’’ Hydrocarbon
Process. 44 (1965) no. 4, 141–145.
13. K. R. Hall, L. Yarborough: ‘‘A New Equation of State for Z-Factor Calculations,’’ Oil
Gas J. 71 (1971) June 18, 82–91.
14. M. Benedict, G. B. Webb, L. C. Rubin: ‘‘An Empirical Equation for Thermodynamic
Properties of Light Hydrocarbons and Their Mixtures,’’ Chem. Eng. Prog. 47 (1951) 419.
15. K. E. Starling, M. S. Hon: ‘‘Thermo Data Refined for LPG,’’ Hydrocarbon Process. 52
(1972) no. 5, 129–133.
16. M. Jaeschke et al.: ‘‘Accurate Prediction of Compressibility Factors by the GERG
Virial Equation,’’ SPE Paper 17 766, SPE Gas Techn. Symp., Dallas, TX, 1988.
17. ISO 6570: ‘‘Natural Gas—Determination of Potential Hydrocarbon Liquid Content,’’
1983/1984.
18. D. L. Katz: ‘‘Prediction of Conditions for Hydrate Formation in Natural Gases,’’
Trans. Am. Inst. Min. Metall. Pet. Eng. 160 (1945) 140.
19. DIN 51 850: ‘‘Brennwerte und Heizwerte gasförmiger Brennstoffe,’’ Beuth Verlag,
Berlin-Köln 1980.
20. Deutscher Verein des Gas- und Wasserfaches: Techn. Regeln ‘‘Gasbeschaffenheit,’’
Arbeitsblatt G 260/I, Eschborn 1983.
21. D. L. Katz: Handbook of Natural Gas Engineering, McGraw-Hill, New York 1959.
22. G. Drewniok, D. Huber: ‘‘Neue Ergebnisse bei der Erdgasbehandlung im Feld,’’
Lecture at the 26th DGMK-Haupttagung, DGMK-Compendium 78/79, Berlin,
Oct. 4–6, 1978.
23. J. H. z. Siederdissen, F. Wundram, GWF Gas Wasserfach: Gas Erdgas 127 (1985)
no. 1, 3–11.
24. F. Brunner, W. Woll, JPT J. Petr. Technol. 10 (1980) 377–384.
25. O. Buchholz, S. Peter, Erdöl Kohle Erdgas Petrochem. 20 (1967) 268–272.
26. G. Wilken, SPE Paper 19 813, 64th SPE Technical Conference, San Antonio, Oct.
8–11, 1989.
27. W. Tunn, Erdöl Kohle Erdgas Petrochem. 26 (1973) no. 9, 498–500.
28. Dow Chemical Comp.: Gas Conditioning Fact Book, Midland, Michigan 1972.
29. NGPSA (Nat. Gas Proc. Suppl. Ass.): Engineering Data Book, Tulsa, Oklahoma, 1972.
30. A. Misra, Erdöl Erdgas Z. 91 (1975) Apr., 99–105.
31. D. Ballard, Hydrocarbon Process. 45 (1966) June, 171–180.
32. J. Kaliner et al.: ‘‘High Pressure Separation of Higher Hydrocarbons From Natural
Gas With a Continuous Regeneration of the Absorbent By Natural Gas,’’ 17th World
Gas Conference, Washington, D.C., June 5–9, 1988.
Vol. 1 NATURAL GAS 423
68. T. J. van Pol: ‘‘Claus Tailgas Treatment,’’ Part 2, Comprimo Gas Sweetening and
Sulphur Recovery Seminar, Amsterdam, Nov. 1985.
69. R. Kettner, N. Liermann, Oil Gas J. 86 (1988) Jan. 11, 63–66.
70. R. Kettner, T. Lübcke, VDI Ber. 730 (1989) 255–274.
71. R. Müller, Erdöl Ergas Kohle Petrochem. 28 (1975) 132–137.
72. R. Kettner, N. Liermann, Erdöl Erdgas 102 (1986), Oct., 448–453.
73. S. Lechler, A. Misra, A. Feizlmayr: New Technology Required for West German
Sulfur Line, Pipe Line Industry, June 1978.
74. T. Lübcke et al.: ‘‘The Mobil Alkanolamine Waste Recovery Process,’’ Comprimo
Sulphur Workshop, Amsterdam, Nov. 1987.
75. R. N. Tuttle, R. D. Kane, National Association of Corrosion Engineers Publication,
Houston 1981.
76. W. J. Schell, C. D. Houston, Chem. Eng. Prog. 78 (1982) Oct. 33–37.
77. S. S. Kulkarni, E. W. Funk, N. N. Li, R. L. Riley, AIChE Symp. Ser. No.229, 79
(1983) 172–178.
78. T. E. Sulpizio, B. S. Minhas, H. S. Meyer, AIChE National Meeting, New Orleans,
Louisiana, April 2, 1992.
79. L. S. White, T. A. Blinka, H. A. Kloczewski, I. Fan Wang, J. Membr. Sci. 103 (1995)
Oct., 73–82.
80. W. J. Koros, G. K. Fleming, J. Membr. Sci. 83 (1993) 1–80.
81. T. Shiratori, H. Sonta: ‘‘Application of Iron-Oxidizing Bacteria to Hydrometallurgi-
cal Flue Dust Treatment and Hydrogen Sulfide Desulfurization,’’ FESM Micro Biol.
Rev. 11 (1–3) 1993, 165–174.
82. C. J. N. Buisman, W. L. Prins: ‘‘New Process for Biological (Flue) Gas Desulphurisa-
tion,’’ Biologische Abgasreinigung, VDI-Bericht 1104, VDI-Verlag Düsseldorf, 1994.
83. H. Satoh, J. Yoshizava, S. Kametani: ‘‘Bacteria Help Desulfurise Gas,’’ Hydrocarbon
Process 67 (1988) no. 5, 76D–76F.
SPECIFIC REFERENCES
84. J. L. Birnbaum: ‘‘The Reemergence of LNG,’’ vols. 1 and 2, Energy Research Assoc.,
New York 1989.
85. ‘‘LNG World Overview,’’ Gotaas-Larsen Shipping Corp., New York, updated annually.
86. R. N. Davis, Y.-N. Liu: ‘‘Baseload LNG: An International Energy Source,’’ 1987
Spring National Meeting of the American Institute of Chemical Engineers, Houston,
USA, Mar. 29–Apr. 2, 1987.
SPECIFIC REFERENCES
87. R. N. Davis, L. E. Olson: ‘‘The Future of Natural Gas as a Fuel in Railroad Opera-
tions,’’ 16th Energy Technology Conference (ET ’89), Washington, D.C., USA, Feb.
28–Mar. 2, 1989.
88. D. Rooke: ‘‘LNG Industry—A Retrospective,’’ 9th International Conference on Lique-
fied Natural Gas (LNG-9), Nice, France, Oct. 17–20, 1989.
89. L. S. Gaumer, C. L. Newton, US 3 763 658, 1973.
90. M. J. Roberts, R. Agrawal, US Patent 6,308,531 B1, 2001.
91. M. J. Roberts, Y. Liu, J. C. Bronfenbrenner, J. M. Petrowski, ‘‘Reducing LNG Capital
Cost in Today’s Competitive Environment’’, PS2-6, The 14th International Conference
and Exhibition on Liquefied Natural Gas (LNG14), Doha, Qatar, March 21-24, 2004.
Vol. 1 NATURAL GAS 425
115. A. W. Drews (ed.): ASTM Manual on Hydrocarbon Analysis, ASTM Manual Series
MNL 3, 5th ed., Philadelphia, PA, 1992.
116. DVGW-Regelwerk, Arbeitsblatt G 261: Prüfung der Gasbeschaffenheit, ZfGW-
Verlag, Frankfurt 1985.
117. K. H. Altgelt, T. H. Gouw: Chromatography in Petroleum Analysis, Marcel Dekker,
New York 1979.
118. S. H. Kägler: Neue Mineralölanalyse, Hüthig Verlag, Heidelberg 1987.
119. T. A. Norris: ‘‘Gas Analysis: Analysis of C5 and Lighter Hydrocarbons,’’ in (115), pp. 1–5.
120. ASTM D 1945 - 91 in (115), pp. 326–337.
121. C. J. Cowper, A. J. De Rose: The Analysis of Gases by Chromatography, Pergamon
Press, Oxford 1984.
122. DIN 51 405, ed. May 1987.
123. E. Leibnitz, H. J. Struppe: Handbuch der Gaschromatographie, 3rd ed., Verlag
Chemie, Weinheim, Germany 1984.
124. H. Laurien: Taschenbuch Erdgas, R. Oldenburg Verlag, München–Wien 1970.
125. DIN 51 872 Part 3, ed. Feb. 1987.
126. ASTM D 2504–83 in (115), pp. 466–469.
127. DVGW-Regelwerk, Arbeitsblatt G 285: Hydratinhibierung in Erdgasen mit Metha-
nol, ZfGW-Verlag, Frankfurt 1974.
128. ASTM E 700–79, DIN 51 777, ed. March 1983.
129. M. A. Berliner: Feuchtemessung, VEB Verlag Technik, Berlin 1980.
130. ASTM 2385–81 in (115) , pp. 428–431.
131. ASTM D 1070–85.
132. ASTM D 3588–81, DIN 51 858, ed. Nov. 1982, ISO 6976–1983.
133. ASTM D 1826–88.
134. ASTM D 1142–90, ISO 6327–1981, DIN 51 871, ed. Dec. 1985.
GEORG HAMMER
TORSTEN LÜBCKE
ROLAND KETTNER
Mobil Erdgas – Erdöl GmbH, Celle,
Federal Republic of Germany
MARK R. PILLARELLA
Air Products and Chemicals, Inc.,
Allentown, PA, United States
HERTA RECKNAGEL
Ruhrgas AG, Essen, Federal Republic of
Germany
AXEL COMMICHAU
Mobil Europe Gas Inc., The Hague, Netherlands
HANS-JOACHIM NEUMANN
German Petroleum Institute,
Clausthal – Zellerfeld, Federal Republic
of Germany
BARBARA PACZYNSKA-LAHME
Consultant, Osterode, Federal Republic
of Germany
Vol. 1 BTX PROCESSING 427
BTX PROCESSING
Benzene (71-43-2) (B), toluene (108-88-3) (T), and the xylenes (X) are the lowest
molecular weight aromatic hydrocarbon homologues. They are each very large
scale chemical feedstocks. Since they are often produced together in the same
process, they can be considered as a group, ie, BTX. However, BTX as such is
not an article of commerce. It is either an important component of a crude mix-
ture such as reformate or pyrolysis gasoline, or it is separated and purified into
its individual components. This article mainly discusses the processes for making
those crude mixtures. Other articles cover the physical and chemical properties
of the individual BTX compounds themselves (see BENZENE; TOLUENE; and
XYLENES AND ETHYLBENZENE). This article focuses on technology involving the
group as a whole.
Originally BTX was obtained commercially by pyrolysis of coal (see COAL).
Since World War II, the production of BTX has been intimately connected with
the production of gasoline. BTX constitutes part of an important gasoline compo-
nent called reformate which is discussed below (see GASOLINE AND OTHER MOTOR
FUELS). Reformate is highly valued for gasoline because it has a very high octane
rating. This results from the high concentration of aromatic compounds, all of
which have very high octane values.
Any BTX needed for chemical use is separated from the reformate stream
before it is blended into the gasoline pool. Although at a given refinery the total
volume of gasoline production (eg, 16,000 m3/d) usually dwarfs the BTX volume
(eg, 800 m3/d) and may have a higher priority, BTX production is often important
enough to support its own reforming facilities and should not be considered sim-
ply as a gasoline by-product. This independence from gasoline may be even
further emphasized in the future because of restrictions on the allowed level
of BTX in gasoline and because new BTX processes may utilize light feeds or
natural gas.
The need for BTX in gasoline has varied considerably. In the 1970s and
1980s in the United States, more high octane reformate was needed as the use
of lead antiknock compounds was decreased for environmental reasons. More
reforming capacity was put into use. Now there is environmental pressure to
reduce the aromatic content (especially benzene) of gasoline (1). This may drama-
tically reduce production of gasoline reformate. Octane number requirements
would have to be met with other high octane components such as oxygenated
hydrocarbons (eg, methyl t-butyl ether (MTBE)). Therefore, the production of
BTX for gasoline probably will drop. However, despite possible dislocations in
supply, the demand for chemical uses will still be readily satisfied and the avail-
ability and price of BTX for chemical uses probably will not be greatly affected by
the change in gasoline composition.
The principal chemical uses of BTX are illustrated in Figure 1 and listed in
Table 1 (2). A very wide range of consumer products from solvents to fibers, films,
and plastics are based on BTX. The consumption of BTX is approximately in the
proportions of 67:5:28, respectively. However, no BTX process gives BTX in these
proportions. The economic value of benzene and xylenes (especially p-xylene) is
normally higher than that of toluene. Because of this, processes that convert
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.02202419230505.a01
428 BTX PROCESSING Vol. 1
Polystyrene
Ethylbenzene Styrene ABS resins
SBR elastomers
Cyclohexane
Adipic acid
Benzene
Toluene
Dinitrotoluene Toluene diisocyanate Polyurethanes
}
0-Xylene Phthalic anhydride
Plasticizers
Polyesters and alkyd resins
Xylenes m-Xylene Isophthalic acid
p-Xylene
Dimethyl terephthalate
Terephthalic acid } Polyester fibers and films
1. Reforming
100
Aromatics in reformate, liquid vol %
90
80
70
60
50
40
30
80 90 100 110
Octane, research
430
Brent (North 38.2 0.4 20.2 54.0 62.3 47 39 14
Sea, UK)
Mayan (Mexico) 21.5 3.4 12.0 56.9 45.3 62 26 12
Isthmus (Mexico) 33.3 1.2 19.2 56.5 65.9 61 26 13
Alaska North Slope 27.6 1.1 13.3 52.1 63.2 42 41 17
(Alaska, U.S.)
Huntington Beach 21.8 1.5 14.1 48.2 70.6 14 80 6
(California, U.S.)
West Texas 39.1 0.3 23.4 54.6 53.8 45 45 10
Intermediate
(Texas, U.S.)
Empire Mix 32.3 0.3 12.7 52.4 61.2 43 44 13
(Gulf Coast,
U.S.)
Vol. 1 BTX PROCESSING 431
Recycle compressor
Desulfurizer
Reformate
Straight run
naphtha
14 C
12
10
16
Toluene, liquid vol %
A B
14 C
12
10
8
Benzene, liquid vol %
6 A
B
5 C
2
0 10 20 30 40 50 60 70 80 90 100
C7 Paraffin conversion, liquid vol %
Fig. 4. Higher yields from lower pressures; reactor pressure: A, 791 kPa (100 psig);
B, 1480 kPa (200 psig); C, 2515 kPa (350 psig). Data for Arabian Naphtha, 54–1548C
fraction.
eg, 450–965 kPa (50–125 psig). Figure 4 shows the effects of pressure (mainly
partial pressure of hydrogen) and of conversion on BTX yield.
As coke fouling occurs, the catalyst loses activity. Temperature must be
raised to keep the conversion level constant. Typically only about a 20–308C tem-
perature rise is allowable. Then the very valuable platinum catalyst must be
regenerated. Three approaches are used. If the run length (fouling time) is
about six months or more, it is practical to shut down the unit and regenerate
the catalyst in place. This approach, called semiregenerative reforming,
is used in a number of commercial processes: Houdriforming, Magnaforming,
Platforming, and Rheniforming (6).
If, because of feed type, process severity, or catalyst composition, the run
length is short, the catalyst is regenerated in swing reactors or continuously.
Processes using four to six reactors, one or more of which may be undergoing
regeneration at any given time, have been developed, eg, Powerforming and
Ultraforming (6). The catalyst has a higher average activity in these pro-
cesses than in semiregenerative ones. Continuous regeneration is offered
by UOP in the Continuous Catalyst Regeneration (CCR) Platforming Pro-
cess, and by IFP in their Aromizing Process (9). A portion of the catalyst is
Vol. 1 BTX PROCESSING 433
Pt, acid
Isoparaffins Alkycyclopentanes
(ACP)
Pt, acid
Pt, acid
Pt
- Paraffins Alkylcyclohexanes Aromatics Lighter
(ACH) (Ar) aromatics
Hydrocracking
Paraffins Cycloparaffins Dealkylation
insomerization insomerization
Paraffins Alkylcyclohexane
dehydrocyclization dehydrocyclization
Fig. 5. Main reactions of catalytic reforming. Pt and acid refer to predominant active
catalytic sites.
434 BTX PROCESSING Vol. 1
2. Hydrocarbon Pyrolysis
A completely new approach for BTX production has emerged in recent years. It
converts C2 to C6 paraffins into aromatics using a modified ZSM-5 zeolite catalyst
which contains gallium (19). An example of this approach, the Cyclar process,
has been in commercial operation by British Petroleum at Grangemouth,
Scotland since August 1990 (20). It uses C3 –C4 feed and employs UOP’s CCR
technology to compensate for rapid catalyst coking.
The mechanistic steps are as follows: paraffins dehydrogenate to olefins; the
olefins oligomerize and cyclize; and the cyclics aromatize. Because the first step
is rate controlling, very little olefin is actually present. The BTX product is rela-
tively free of nonaromatics and therefore is very desirable as a chemical feed. As
in reforming, some C1 –C2 fuel gas is produced along with a valuable hydrogen
stream. From a C3 –C4 feed the BTX product is roughly 35:45:20, respectively.
5. BTX Recovery
The complexity of separating and purifying the individual BTX components from
crude BTX products depends on the amount of nonaromatic impurities present.
If the amount is small enough, simple distillation can suffice. If not, it is obvious
from Figure 6 that distillation alone will not be sufficient because the BTX
aromatic compounds are close in boiling point to some of the cycloparaffins of
the same carbon number or to paraffins of the next higher carbon number.
Because of this, extraction or extractive distillation with a polar solvent is
used to separate the slightly polar aromatic hydrocarbons from the nonpolar non-
aromatic hydrocarbons. When olefins are also present, as in pyrolysis gasoline,
these interfere with the extraction and must first be hydrogenated or removed
by adsorbents.
An option for avoiding the cost of extraction is to increase the severity of the
BTX formation step. This reduces the quantity of residual paraffins, and,
436 BTX PROCESSING Vol. 1
200
Temperature,°C 150
100
50
P C A P C A P C A P C A
C6 C7 C8 C9
Fig. 6. Boiling points of C6 –C9 hydrocarbons. P, iso and normal paraffins; C, C5- and
C6-cycloparaffins; and A, aromatics.
depending on the BTX formation process, may leave the BTX clean enough to
purify by distillation. The final impurity concentrations may still be too high
for merchant sale, but may be acceptable in some downstream operations. For
example, xylenes going into an isomerization/separation loop for p-xylene pro-
duction can contain some paraffins if the isomerization catalyst is capable of
decomposing them. Disadvantages of high severity processing are the increased
catalyst fouling rate and the potential increase in undesirable olefin impurities.
5.1. Extraction and Extractive Distillation. The choice of an extrac-
tion or extractive distillation solvent depends upon its boiling point, polarity,
thermal stability, selectivity, aromatics capacity, and upon the feed aromatic
content. Capacity, defined as the quantity of material that is extracted from
the feed by a given quantity of solvent, must be balanced against selectivity,
defined as the degree to which the solvent extracts the aromatics in the feed in
preference to paraffins and other materials. Most high capacity solvents have
low selectivity. The ultimate choice of solvent is determined by economics. The
most important extraction processes use either sulfolane or glycols as the polar
extraction solvent.
Sulfolane [126-33-0], used in UOP and Shell processes (25,26), offers good
thermal and hydrolytic stability, high density and boiling point, and a good bal-
ance of solvent properties. Its high density and boiling point make it easy to sepa-
rate from the hydrocarbon streams. A diagram of a sulfolane extraction unit is
shown in Figure 7. Fresh feed enters the extractor and flows countercurrent to
Vol. 1 BTX PROCESSING 437
Raffinate
Rich Extract to
solvent treating and
fractionation
E ED RC
Feed
Recycle
Lean solvent
Fig. 7. Shell sulfolane extraction process. E, extraction; ED, extractive distillation; RC,
recovery column. (Courtesy of UOP, Inc.)
the down flowing solvent. The raffinate is withdrawn at the top of the extractor
and leaves the system after water washing. The solvent, now rich in aromatics, is
sent to the top of the extractive stripper where the nonaromatic hydrocarbons
are removed. Aromatics and sulfolane are separated in the recovery column.
The lean solvent is recycled to the extractor and the aromatics are washed
with water and removed. The recovery of benzene and toluene is usually
99þ%; of C8 aromatics 97%; and of Cþ 9 aromatics 75–90%.
The widely employed UOP Udex Process uses a glycol solvent (27). Diethy-
lene glycol was used in early versions of the process; however, increased capacity
was obtained by adding dipropylene glycol or, in some cases, a change was made
to triethylene glycol. Further improvement was made by using tetraethylene gly-
col (28). The Union Carbide Tetra Process also employs tetraethylene glycol (29).
Other extraction processes are included in Table 4.
Extractive distillation, using similar solvents to those used in extraction,
may be employed to recover aromatics from reformates which have been prefrac-
tionated to a narrow boiling range. Extractive distillation is also used to recover
a mixed benzene–toluene stream from which high quality benzene can be pro-
duced by postfractionation; in this case, the toluene product is less pure, but is
still acceptable as a feedstock for dealkylation or gasoline blending. Extractive
distillation processes for aromatics recovery include those listed in Table 4.
5.2. Downstream Processing. In addition to extraction, various down-
stream operations are often carried out on the BTX product to produce products
in proportions to fit the market demand. A typical aromatics processing scheme
is shown in Figure 8 in which benzene, p-xylene, and o-xylene are the products.
438 BTX PROCESSING Vol. 1
After the crude BTX is formed, by reforming in this case, a heart cut is
sent to extraction. Actually, the xylenes and heavier components are often
sent to downstream processes without extraction. The toluene produced is con-
verted to benzene, a more valuable petrochemical, by running it through a
hydrodealkylation unit. This catalytic unit operates at 540–8108C with an
excess of hydrogen. Another option is to disproportionate toluene or toluene
plus C9 aromatics to a mixture of benzene and xylenes using a process such
as UOP’s Tatoray or Mobil’s Selective Toluene Disproportionation Process
(STDP) (36).
The o-xylene (95-47-6) in Figure 8 is recovered by a two-stage distillation.
First it is separated (or split) from m-xylene (108-38-3) and the other C8
aromatics in a superfractionating column, the xylene splitter, (Unit H). The
bottoms, a mixture of o-xylene and Cþ 9 aromatics, is redistilled (or rerun) in
Unit I to recover o-xylene of 96 þ% purity.
The distillate (overhead) from Unit H, containing mostly ethylbenzene
(100-41-4), p-xylene, and m-xylene, and some o-xylene becomes the feed for the
p-xylene separation process (Unit J).
p-Xylene (106-42-3) can be purified by crystallization or adsorption.
When a typical reformate-derived C8 aromatic mixture is cooled, p-xylene
crystallizes first. Most plants employing crystallization operate at 60 to
75%, depending on feed composition (37). The process is limited by a eutectic
temperature below which o- or m-xylene also crystallize. The solubility of
p-xylene in the remaining C8 aromatic mixture over the range of 60 to
758C is 9.6 to 6.2%.
UOP’s ParexProcess can be used to purify p-xylene by adsorption (38).
Toray has a similar process. These processes take advantage of the fact that
p-xylene is adsorbed more easily than the other C8 aromatics by a suitable mole-
cular sieve. The p-xylene is desorbed by either a lighter or heavier hydrocarbon
which is subsequently removed by distillation. p-Xylene is recovered in about
97% yield (see Adsorption).
The mother liquor obtained from the crystallization, or the raffinate after
removal by adsorption, is isomerized using an acidic catalyst to convert m-xylene
to the o- and p-isomers (Unit K in Fig. 8).
In the isomerization step, light and heavy impurities are generated,
including saturated hydrocarbons, benzene, toluene, and Cþ 9 aromatics.
These are removed by recycling the isomerate through the xylene splitter
and the light isomerate column (Unit L). If the light isomerate contains
much benzene or toluene, it can be recycled to Unit G. m-Xylene may be recov-
ered between the xylene splitter and the p-xylene separation plant by selective
sulfonation followed by regeneration or via complex formation with a mixture
of HF and BF3 (39).
To this point the presence of ethylbenzene in the mixed xylenes has been
ignored. The amount can vary widely, but normally about 15% is present.
The isomerization process must remove the ethylbenzene in some way to
ensure that it does not build up in the isomerization loop of Figure 8. The
ethylbenzene may be selectively cracked (40) or isomerized to xylenes (41)
using a platinum catalyst. In rare cases the ethylbenzene is recovered in high pur-
ity by superfractionation.
440 BTX PROCESSING Vol. 1
There are many variations of the basic processing loop shown in Figure 8.
Processing to produce only BT is common, often in conjunction with a toluene-
to-benzene dealkylation unit. If benzene and toluene are not to be recovered,
Column B may be used to remove toluene and lighter components. In
that case, Units E, F, and G would be eliminated. To recover p-xylene only,
the xylene splitter is reduced in size and is used to split between o-xylene and
the Cþ
9 aromatics. The o-xylene rerun still is then eliminated.
6. Environmental Considerations
BTX processing has come under steadily increasing pressure to reduce emissions
and workplace exposures (see Industrial hygiene). Reductions in the permissible
levels of both benzene and total aromatics (BTX) in gasoline have been legislated.
Whereas all BTX components are to be controlled, the main focus is on benzene
because it is considerably more toxic than the others and is classified as a known
carcinogen (42).
Workplace exposure limits for benzene have been regulated to levels as low
as 0.5 ppm (43). Industrial emissions affecting the public are now low enough
that the EPA considers that a greater hazard exists from mostly indoor sources
such as smoking, automobile exhausts, and consumer products (44).
The stringent controls over manufacturing and handling benzene may dis-
suade refiners from installing new equipment to extract benzene from gasoline to
satisfy gasoline requirements. Instead they will probably modify their reforming
operations to produce less benzene or to convert the remaining benzene to other
compounds. They will achieve the necessary octane rating for gasoline via
branched paraffins and oxygenates. Overall, this is likely to result in further
segregation of petrochemicals BTX production from gasoline production.
BIBLIOGRAPHY
‘‘BTX Processing’’ in ECT 3rd ed., Vol. 4, pp. 264–277, by D. L. Ransley, Chevron Research
Company.
CITED PUBLICATIONS
W. A. SWEENEY
P. F. BRYAN
Chevron Research and
Technology Company
442 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
FLUID CATALYTIC
CRACKING, UNITS,
REGENERATION
1. Introduction
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0612210921161915.a01.pub2
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 443
Fig. 1. (a) Modern UOP FCC unit with: Short contact time riser VSS disengager
combustor-style regenerator. (b) Mid-1970s UOP FCC unit with: Riser cracking Stacked
reactor bubbling-bed regenerator.
transport regime. By contrast, the typical 1970s style regenerator utilized a bub-
bling fluidized bed for air/catalyst contacting.
The heat released from coke burning is used to provide the heat require-
ments on the reactor side: to vaporize the feed, to compensate for the endother-
mic heat of cracking, to bring the reactor to the desired reactor temperature, and
to compensate for heat losses to the atmosphere. Thus the FCCU always operates
in complete heat balance at any desired hydrocarbon feed rate and reactor tem-
perature. In FCC units, such as the ones shown in Figure 1, riser temperature
control adjusts the catalyst circulation rate to achieve the heat balance. Catalyst
flow is controlled by a slide valve located in the catalyst transfer line from the
regenerator to the reactor and in the catalyst return line from the reactor to
the regenerator. In some older style units, where catalyst flow is controlled by
pressure balance between the reactor and regenerator, the heat-balance control
is more often achieved by changing the temperature of the hydrocarbon feed
entering the riser.
Prior to entering the regenerator, the spent catalyst from the reactor passes
through a steam stripper where the hydrocarbon vapors that are entrained by
the flowing catalyst particles are displaced by steam. The steam also acts to
desorb some of the lighter hydrocarbons that are adsorbed during the reaction
step. These stripped materials, which represent valuable reaction products,
are carried back into the reactor by the upflowing steam from the stripper and
go to the product-recovery section with the other reaction products. The removal
of these materials from the catalyst stream flowing to the regenerator also
reduces the amount of burnable hydrocarbon going to the regenerator, which
helps to decrease the regenerator temperature.
Fig. 2. Heat balances of the FCCU. Heat balance around the reactor A, heat balance
around the regenerator B, and the overall heat balance around the entire catalytic cracker
C must be considered.
Thus the amount of heat that must be produced by burning coke in the re-
generator is set by the heat balance requirements and not directly set by the
coke-making tendencies of the catalyst used in the catalytic cracker or by the
coking tendencies of the feed. Indirectly, these tendencies may cause the cracker
operator to change some of the heat-balance elements, such as the amount of
heat removed by a catalyst cooler or the amount put into the system with the
feed, which would then change the amount of heat needed from coke burning.
If FCCU operations are not changed to accommodate changes in feed or cat-
alyst quality, then the amount of heat required to satisfy the heat balance essen-
tially does not change. Thus the amount of coke burned in the regenerator
expressed as a percent of feed does not change. The consistency of the coke
yield, arising from its dependence on the FCCU heat balance, has been classified
as the second law of catalytic cracking (9).
3. Coke Formation
metals (principally nickel and vanadium) in the feed to the FCCU; (3) additive
coke is directly related to the high boiling, refractory components in the feed
to the FCCU. This coke correlates with the basic nitrogen and molecular weight
of the feed (11) and with the carbon residue content of the feed (10); and
(4) cat-to-oil coke is related to the catalyst circulation rate and derives
from the hydrocarbons leaving the stripper. These hydrocarbons are potentially
strippable but because of incomplete stripping are entrained by the flowing cat-
alyst into the regenerator. FCCU coke distributions for operations with zeolitic
catalysts are given in Table 1 (10,12).
The amount of catalytic coke that is formed depends on the type of catalyst
used in the FCCU, the coking tendency of the feed, the degree of conversion of
the feed, and the length of time the catalyst is exposed to the feed (eq. 2) (13).
conv k
conversion2nd order ¼ ¼ ð4Þ
100 conv sv
Space velocity is related to the C/O ratio through the catalyst residence time as
C 1
¼ ð5Þ
O ðsvÞt
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 447
C
CF ¼ conv2nd Order ð6Þ
kt
Equation 6 relates the catalytic coke yield (as a fraction of the feed) to the delta
coke, to the conversion, and to the catalyst residence time.
It has been shown that coke yield as a fraction of feed does give a linear
relationship with second-order conversion (15) indicating a positive coke yield
at zero conversion. This coke yield at zero conversion is the additive coke contri-
bution to the total coke yield and is related to feed properties, particularly
Conradson carbon content. The amount of this additive coke is significantly
less than the Conradson carbon value of the feed (16), probably in the range of
50% of the Conradson carbon.
Catalytic coke is also a function of the type of feed being processed. The
more aromatic the feed, the higher the coke (on feed) yield at a given conversion
(15). The type of cracking catalyst also influences the coke yield. Increasing
the H-transfer characteristics of a Y-type zeolite, by increasing the amount of
rare earth exchanged on the zeolite, results in an increase in the catalytic coke
yield (17). Modern catalysts for FCC applications frequently contain two dis-
tinctly different, active components; an active zeolite component for producing
high yields of gasoline while yielding minimum catalytic coke and an active,
nonzeolitic component, typically a catalytically active alumina located in the
so-called matrix portion of the catalyst surrounding the zeolite. The active alu-
mina matrix provides improved conversion of the heavy portion of the feed but
typically increases catalytic coke yield at constant conversion (18).
Contaminant coke can vary substantially because of large variations in the
metals content of the feed being processed and from variations in the manner
that cracking catalysts respond to these metals. Nickel is considered the most
active dehydrogenation agent, and hence the most troublesome of the contami-
nant metals. Copper is equally active as a dehydrogenation catalyst, but the
amount of copper present is usually too small to have an effect. Typically, copper
in the feed to the FCCU is only 2–5% of the nickel content. Vanadium is a less
active dehydrogenation agent than nickel, about one-fourth as effective as nickel
(10). The metals content of the equilibrium catalyst in the FCCU is frequently
referred to as the equivalent nickel content, which is defined as
V
equivalent nickel ¼ Ni þ ð7Þ
4
The equivalent nickel content of the feed to the FCCU can vary from
<0.05 ppm for a well-hydrotreated VGO to >20 ppm for a feed containing a high
resid content. The nickel and vanadium deposit essentially quantitatively on the
cracking catalyst and, depending on catalyst addition rates to the FCCU, result
in total metals concentrations on the equilibrium catalyst from 100 to 10,000 ppm.
The coking characteristics of a cracking catalyst increase as the equivalent
nickel level on the catalyst increases. At high equivalent nickel levels, contami-
nant coke can be the major coke component on the catalyst. The effect of the
448 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
contaminant nickel on the catalyst coke content can be mitigated through the use
of a metal passivator added to the feed or through the incorporation of an effec-
tive metals traps into the cracking catalyst. Coking problems due to contaminant
vanadium still remain a problem.
Feed additives were frequently used in the 1980s and early 1990s to reduce
the contaminant coke produced through nickel-catalyzed reactions. These inhibi-
tors were injected into the feed stream going to the catalytic cracker. The addi-
tive forms a nickel complex that deposits the nickel on the catalyst in a less
catalytically active state. The first such additive was an antimony compound
developed and first used in 1976 by Phillips Petroleum. During the 1980s, anti-
mony was widely used in FCCUs that had a problem with contaminant nickel.
The use of the antimony additive reportedly reduced coke yields by 15% in a com-
mercial trial (19). In the late 1980s, other additives were introduced to combat
contaminant nickel, eg, Chevron introduced a bismuth-based additive, which is
claimed to provide performance similar to antimony (20).
In the late 1980s, cracking catalysts were developed with metals traps that
appear to be so effective in containing the adverse effects of contaminant nickel
that additive-type inhibitors are no longer needed (21). These traps consisted of
reactive alumina. It has been reported that these aluminas promote a chemical
reaction with the nickel to form NiAl2O4 which then diffuses into the bulk alu-
mina phase, thereby minimizing any contact between nickel and the potential
coke-forming hydrocarbons (22). By the late 1990s, all FCC catalyst manufac-
turers incorporated such a trap into their metals resistant catalysts.
4. Coke Burning
The burning of coke in the regenerator provides the heat to satisfy the FCCU
heat balance requirements as shown in equation 1. The heat released from the
burning of coke comes from the reaction of carbon and hydrogen to form carbon
monoxide, carbon dioxide, and water. The heat generated from burning coke thus
depends on the hydrogen content of the coke and the relative amounts of carbon
that burn to CO and CO2, respectively.
The impact that variations in coke content and burning conditions can have
on the overall heat of coke combustion is shown in Table 2.
Because the heat balance dictates the amount of heat that is required from
burning coke, the heat of combustion then determines the amount of coke that
must be burned.
combustion heat required; kJ=h
coke required; kg=h ¼ ð8Þ
specific heat of combustion; kJ=kg
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 449
dcC
¼ kcC PO2 ð10Þ
dt
0.9
0.8
0.7
0.6
0.5
0.4
0.2
0.1
0.08
0.06
0.05
0.04
0.03
0.02
0.01
0.5 1.0 1.5 2.0 2.5 3.0
Regeneration time, h
Fig. 3. Burning at 5008C in flowing air of coke from H-Y FCC catalyst.
argued that coke burning follows first-order kinetics, but only after an initial
period of several minutes, when burning is not first order with carbon content
(31). The data in Figure 3 on coked FCC catalyst of the H-Y type agree with
this argument.
The activation energy for burning from a coked zeolite has been reported as
109 kJ/mol (32) and 125 kJ/mol (30 kcal/mol) has been found for coke burning
from a H-Y FCC catalyst. Activation energies of 167 kJ/mol (40 kcal/mol) (27)
and 159 kJ/mol (28) have been reported for the burning of carbon from a coked
amorphous silica–alumina catalyst.
The coke-burning rate in the regenerator is thus a function of the oxygen
partial pressure, the carbon content on the catalyst leaving the reactor, the tem-
perature in the regenerator, and the residence time in the regenerator. The last
three variables are all coupled together through the FCCU heat-balance require-
ments. For example, an increase in regenerator temperature normally results in
a decrease in catalyst circulation rate to maintain a constant temperature on the
reactor side. The decrease in catalyst circulation rate increases the residence
time in the regenerator and is accompanied by a change in the carbon content
of the catalyst coming into the regenerator. A typical interrelationship between
the O2 content in the regenerator flue gas, the regenerator temperature, and the
carbon content of the regenerated catalyst (CRC) is shown in Figure 4. At a con-
stant flue gas O2 content, a decrease in regenerator temperature of 208C results
in a CRC increase of 0.15–0.20 wt%.
The hydrocarbon feed rate to the reactor also affects the burning kinetics in
the regenerator. Increasing the reactor feed rate increases the coke production
rate, which in turn requires that the air rate to the regenerator increase.
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 451
1.0
0.9
0.8
0.7
0.6
0.5
Coke on regenerated catalyst, wt %
0.5%
0.4
1.0%
0.3
0.1% Excess O2
0.2
2.0%
0.1
620 640 660 680 700 720 740
Regenerator temperature, °C
Because the regenerator bed level is generally held constant, the air residence
time in the dense phase decreases. This decrease increases O2 content and con-
sequently CO combustion in the dilute phase, frequently referred to as after burn
(Fig. 5).
The coke-burning rates on silica–alumina catalysts have been found to
increase several orders of magnitude when transition-metal oxides are present
(28). The addition of 0.15 wt % chromium (as Cr2O3) to a silica–alumina catalyst
produces a nearly fivefold increase in carbon-burning rate. The cation exchanged
into a zeolite-containing catalyst affects the burning rate (32). A zeolite
exchanged with copper was found to have a burning-rate constant that was
18 times greater than a zeolite exchanged with chromium.
The hydrogen contained in coke burns at a higher rate than carbon.
Hydrogen-burning rates are four to five times greater than carbon-burning
rates (27,29).
452 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
+20
–10
Complete combustion
CO = 150 ppm
–20 O2 = 0.2 Vol %
50 60 70 80
Regenerator gas velocity, cm/s
Fig. 5. Effect of gas residence time on afterburning. The temperature change DT is the
difference in temperature between flue and dense phase (33).
importance was their discovery that at this low incorporation, the Groups 8–10
(VIII) metals did not catalyze undesirable dehydrogenation reactions during the
cracking reaction. The cracking-selectivity characteristics achieved by the crack-
ing catalyst in use in the FCCU are thus not altered when such a CO combustion
promoter is added to the catalyst inventory.
The use of 1–2 ppm platinum in FCCU circulating catalyst inventories is
now widely practiced throughout the world to catalyze the combustion of CO
to CO2 in the FCCU regenerator. The promoter is added to the FCC catalyst
inventory either as an integral part of the fresh FCC catalyst, where it is
included in the fresh catalyst at 1–2 wppm, or as a separate additive. The addi-
tive, which typically contains 500–1000 wt ppm of platinum, is added to the
FCCU by a small metering system that is independent of the fresh cracking-
catalyst addition system. In the United States, the separate additive approach
is generally used; in Europe, the additive is most commonly incorporated with
the fresh cracking catalyst.
Through the use of a combustion promoter, CO combustion occurs readily in
the dense phase at temperatures well <7008C. Guegan reported that promoted
CO combustion occurred as low as 6508C in a commercial FCCU operation (37).
Promoted combustion has a number of important benefits. First, the after-
burn problem, ie, burning in the dilute phase, is greatly reduced. By consuming
CO in the dense phase, the potential heat release from burning in the dilute
phase is significantly decreased. A commercial example of a 508C reduction in
regenerator afterburn by using a CO combustion promoter is as follows (38):
Conversion, wt %
.74
.73
.72
0.4
CRC, wt %
0.3
0.2
0.1
0
7
Cat:oil ratio
6
5
4
3
760
Regen. temp, °C
740
720
700
680
660
1 2 3 4 5 6 7 8
CO content in flue gas, vol %
70
60 10% promoted
30
20
10
0
0 1 2 3 4 5 6 7 8 9 10 11
CO in flue gas, vol %
5. Environmental Aspects
The FCCU regenerator is one of the principal sources of air pollutants from a
refinery. In the mid-1990s, a typical 23,850 m3/day (150,000 barrel per day), mod-
ern U.S. refinery with a 50,000 BPD FCCU processing a low sulfur, low metals
content VGO feed, and operating in complete CO combustion would typically
emit the following into the atmosphere from the FCCU regenerator flue gas
stack: fine catalyst particles, 2–3 t/day; sulfur oxides, 3–4 t/day; nitrogen oxides,
0.5 t/day; and CO, 1.5–2 t/day. Recently, governmental regulations, particularly
456 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
in the United States, have significantly reduced the allowable FCCU emissions
for particulates and sulfur oxides. A new 50 MBPD unit or an existing 50 MBPD
unit being significantly revamped in 2002 would be required to meet the
following contaminant levels in the regenerator flue gas: catalyst particulates
<0.5 t/day, sulfur oxides <0.7 t/day (41). Nitrogen oxide limits are not yet defined
but are expected to be similarly constrained. Carbon monoxide emissions would
be limited to <3 t/day. Much research has been done in the past and is ongoing to
develop technologies to meet these emissions levels and reduce them even
further.
5.1. Catalyst Emissions. High velocity air passing up through the
regenerator at velocities typically around 1 m/s carries large quantities of cata-
lyst into the vapor space above the catalyst bed. To contain these catalyst parti-
cles within the regenerator vessel, a highly efficient, generally two-stage cyclone
system is installed inside the regenerator. The flue gas and the entrained cata-
lyst particles pass through the cyclone system which is installed inside the regen-
erator. The flue gas and the entrained catalyst particles are typically returned by
the cyclones back into the regenerator. The particles that are not retained and
that leave with the exiting flue gas are small, typically <20 mm particle size.
These particles generally represent fines that have been produced as a
result of catalyst attrition. Two attrition mechanisms are generally recognized:
particle shattering and surface abrasion of particles. The abrasion mechanism
generates fines of much smaller average size. The manufacturing techniques
and chemical compositions of the various FCC catalysts influence the type and
severity of the attrition mechanism. A small amount of fines (<20 mm in size)
may also come from fresh catalyst additions.
Current particulate regulations vary from site to site depending on the local
regulatory agency’s standards. The New Source Performance Standards, NSPS,
set by the U.S. federal government state that a new unit must not exceed 1lb of
particulates per 1000 lbs of coke burned (1 kg/ 1000 kg). It is possible that an
existing unit, that undergoes a substantial revamp, can be reclassified as a
new source.
There are several ways to capture the particles remaining in the flue gas as
it exits the regenerator. The most common technique is through the use of elec-
trostatic precipitators (ESP), which can reduce particulate emissions to very low
levels (41). An ESP consists of rows of collector plates and voltage discharge elec-
trodes arranged in one or more gas tight chambers. The electrodes apply electri-
cal charges to particles in the FCCU regenerator flue gas. The charged particles
are attracted to the collector plates and are thus removed from the flue gas.
While this equipment is well known and has a long record of reliable perfor-
mance, it is expensive and can be dangerous. Explosions in ESPs have occured.
An alternative to the ESP is a third stage separator (TSS). The TSS is an
additional cyclonic separator, external to the regenerator, which supplements
the catalyst separation achieved by the two stages of cyclones typically present
in an FCCU regenerator.
In the TSS, catalyst is removed from the flue gas by passing the catalyst-
entrained gas stream through a set of cyclonic elements that separate out the
catalyst particles. Separation efficiency depends upon the particle-size distribu-
tion of the catalyst leaving the regenerator cyclones, being most efficient on large
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 457
The refiner does have several options to reduce SOx emissions, and gener-
ally will use a combination of these options to reach regulatory requirements.
These include
sulfur is separated from the liquid and the sodium bicarbonate solution returned
to the scrubber. The elemental sulfur is then either dried and sent to landfill as
nonhazardous waste or further purified for resale (50).
Calcium carbonate (limestone) slurry is another solution that is used for
scrubbing the FCCU flue gas. This material is less expensive to purchase, as it
is shipped as a solid. However, it can be difficult to get into solution and plugging
issues can be problematic. Its SO2 removal efficiencies are usually 95–98% (50).
The reaction chemistry produces calcium sulfite, CaSO3 , which for handling pur-
poses is oxidized to calcium sulfate (gypsum), CaSO4 . The gypsum is dried and
recycled as an additive for cement production. One drawback to this calcium car-
bonate scrubbing is the release of CO2 in the reaction. Current concerns over CO2
emissions as a greenhouse gas may make this scrubbing technique less desirable.
Sea water scrubbing to remove SO2 is being practiced by StatOil in Mong-
stat, Norway. The seawater contains natural calcium and sodium buffers that
absorb the SO2. The water is returned to the sea after dilution with sea water
cooling streams (52).
Although often lumped together in analysis results, SO2 and SO3 create
there own unique emission and control issues. The SO3 content, in FCCU flue
gas, can vary from 1 to 10% of the total SOx (53). The SO3 emissions, are espe-
cially troublesome in that SO3 condenses to form sulfuric acid mist that is not
readily removed by wet scrubbers. The light scattering properties of sulfuric
acid mist creates a visible plume that, depending on measurement technology
utilized, is often measured and treated as a particulate emission. SBS Injection
Technology, commercialized in the utility industry in 2002 has been proposed as
a solution to FCCU SO3 problem (53). Sodium bisulfite solution is injected into
the flue gas to react the SO3 to produce sodium sulfate and sodium bisulfate
particle (solids). These then can be readily removed by an electrostatic precipita-
tor or in an SO2 scrubber.
SOx Reduction Catalyst Additive. Currently, many refiners find that the
use of a SOx adsorbing additive is the most effective way to comply with the EPA
regulations. With such an additive, the following steps occur
Oxidation of SO2 to SO3 in the regenerator:
2 SO2 þ O2 ! 2 SO3
SO3 þ MO ! MSO4
MS þ H2 O ! MO þ H2 S
460 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
The H2S is carried out with the reactor products, goes through the product-
recovery system of the FCCU, and eventually to further processing for sulfur
recovery. The metal oxide adsorbent recirculates with the spent cracking catalyst
back to the regenerator for the next SOx adsorption cycle.
The first commercially effective metal oxide adsorbent consisted of a solid
solution of a pure magnesium aluminate spinel (MgAl2O4) with MgO (54).
Such a solid solution (Mg2AlO5) does not destroy the spinel framework. The
adsorption activity of the dispersed MgO in the spinel is much greater than
that of pure MgO itself (54).
Cerium is effective in oxidizing SO2 to SO3. Consequently, a cerium-impreg-
nated Mg2Al2O5 actively converts the SO2 to SO3, which is then strongly
adsorbed by the dispersed MgO as MgSO4. A completely cyclic SOx removal cat-
alyst contains, in addition to the above component, a fourth metal component
such as vanadium to catalyze the conversion of MgSO4 to MgO (55).
Multifunctional SOx removal catalyst systems have been in commercial use
since 1985 in the United States. Such systems have successfully reduced SOx
emissions in the FCCU regenerator by typically 20–60%, with reductions up to
90% reported (39,56,57). Additive levels in the circulating catalyst inventory
range from 1 to 10%. The additive level and the amount of SOx reduction depend
on conditions, such as feed quality, the presence or absence of a CO combustion
promoter, regenerator temperature, regenerator mixing efficency, and excess O2
content (39).
Hydrotalcite, a MgO/Al2O3 containing compound, has also been found to
have SOx adsorbent capability and is now being used in some SOx adsorbent
additives (39). In 2003, both spinel and hydrotalcite types of SOx reduction tech-
nology have been upgraded to higher performance standards, frequently by the
addition of more magnesium oxide component (39).
Because O2 is necessary to convert SO2 to SO3, decreasing O2 in the regen-
erator has been found to reduce the effectiveness of the SOx removal additive.
The SOx additives used in regenerators operating in a partial CO combustion
mode, where excess O2 is frequently limited to <0.2 vol% in the flue gas, are
less successful in reducing SOx. In such cases, SOx removal is typically 20–
30% less than for a full CO combustion (1þ% excess O2) case (56).
5.3. Nitrogen Oxide Emissions Source and Controls. The (NO)x
(nitrous oxides ) emissions are now recognized as a significant contributor to
photochemical smog and acid rain, and therefore have come under greater reg-
ulatory scrutiny than in the past. In terms of tons emitted per year, the FCCU
regenerator is one of the largest point source of NOx emissions. Only a small por-
tion of the NOx present in the flue gas is produced through the oxidation of N2 in
the regenerator air stream (40,58). The main source of the NOx comes from the
combustion of organic nitrogen, originating in the FCC feed of which 50% ulti-
mately reaches the regenerator section in the coke. However, only 5–20% of the
organic nitrogen compounds entering the regenerator end up as NOx, predomi-
nantly NO. The remainder is converted to N2. Since most of the organic nitrogen
ends up as N2, there is likely a secondary reaction with CO or coke to reduce NO
to N2 (59). Although feed nitrogen is the source of NOx, researchers have found that
final NOx concentrations are due more to the type of N in the feed and to the regen-
erator conditions, than they are to the absolute content of N in the feed (40).
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 461
The amount of excess oxygen may have a more direct correlation with the final
NOx concentration leaving the regenerator (39,40). Increasing the O2 in the
regenerator flue gas from 0.1 to 1.6% has been reported to double NOx emissions
when no CO combustion promoter is present. Adding a platinum-based combus-
tion promoter, which increases the available atomic O2, can increase the NOx
content much more (60).
Hydrotreating the reactor feed can reduce feed nitrogen by 20% (low sever-
ity), 90% (moderate Hydrotreating), and up to 99% (severe Hydrotreating) (61),
and thus can be a tool in reducing FCCU flue gas NOx emissions.
An emerging picture of the complex NOx chemistry in the FCCU regenera-
tor includes the following reaction networks:
Pyrolysis Oxidation
N (Coke) ! HCN ! N2, NO, N2O
N (Coke) ! NH3 ! N2, NO, N2O
NO Reduction
coke
2 NO þ 2 CO ! N2 þ 2 CO2
NO þ CFAS ! 1=2 N2 þ CO
FAS ¼ free active surface
Catalyst additives and regenerator design have both been found to be useful
in reducing flue gas NOx emissions. Uniform distribution of coked catalyst in the
regenerator improves the probability of NOx adsorption onto the carbon surface,
necessary for NOx reduction reactions to occur. Air grid design, spent catalyst
distribution, catalyst withdrawl location, and the fluidization regime can all
play major factors in the effectiveness of mixing and distribution. Poor design
and/or erosion of air grids and catalyst distributors can lead to maldistribution
and therefore higher NOx values. Regenerator designs that operate in a fast flui-
dized regime (combustor style) rather than a turbulent regime (bubbling bed
style) afford better catalyst-air mixing and therefore lower NOx emissions.
Recent research in NOx additives has taken two approaches: the develop-
ment of CO combustion promoters that do not catalyze NOx formation and addi-
tives that directly reduce NOx emissions. New NOx reduction additives have been
designed to catalyze the NO reduction by CO or to block the formation of NO by
oxidation of the reduced intermediates, HCN and NH3. The major FCC catalyst
vendors are now publishing the results of commercial trials of NOx reduction
additives in US FCC units. Variying degrees of successes have been reported
(39,57,62). This suggests that the complex interplay between FCCU regenerator
hardware, operating variables and additives for NOx reduction requires more
detailed study.
Post regenerator NOx Reduction Technology Two technologies have
been commercialized to reduce FCCU NOx emissions by treatment of the flue
gas after it leaves the regenerator.
Selective Catalytic Reduction (SCR), which is the most proven of these pro-
cesses, removes NOx by passing it over a vanadium pentoxide/titanium dioxide
462 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
catalyst bed in the presence of NH3 and O2. The reactions are
4 NH3 þ 4 NO þ O2 ! 4 N2 þ 6 H2 O
4 NH3 þ 2 NO2 þ O2 ! 3 N2 þ 6 H2 O
1. An SCR side reaction typically converts 1–5% of the SO2 to SO3. This sig-
nificant increase in SO3 can result in a visible plume in the flue gas leaving
the regenerator stack (53).
2. Unreacted NH3 can react with SO3 leaving the SCR to produce (NH3)2SO4
and NH3HSO4, which can increase particulate emissions and can also
contribute to plume formation (41).
Another dry process utilizes activated coke to remove both SOx and NOx
from the flue gas in a two stage, moving-bed process (51). The SOx is removed
in the first stage via adsorbtion on the activated carbon. NH3 is injected into
the flue gas, which then passes through the second bed, where the activated car-
bon acts to catalyze the reaction of NOx with NH3 to convert the NOx to N2 and
H2O. The activated carbon flows from the second stage to the first stage and then
to a regenerator to desorb the SOx. The desorption step releases SO2, which must
be recaptured and converted to elemental sulfur. This technology was commer-
cialized on a resid-processing FCCU in Japan in 1987 and achieved a reported
70% removal of both NOx and SOx (63).
Several emerging technologies, yet to be fully commercialized also appear to
have promise for reducing NOx from the flue gas leaving the regenerator. Scrub-
bing of NOx, as is done to remove SOx , is not an effective means of removing NOx,
since NO and NO2 are much less soluble than SO2. A process using ozone injected
into the flue gas stream converts NOx to N2O5, which is highly water soluble and
rapidly converts to nitric acid. Scrubbing of the flue gas with a NaOH solution is
then very effective in removing NOx and further converts the nitric acid to the
neutral salt (NaNO3) (53). A combination chemical–biological process utilizes
an iron chelate aqueous solution to convert NOx to a water soluble nitrosyl com-
plex. The solution is sent to a bioreactor, where the complex is reduced to N2,
CO2, and H2O (50). Finally, a high temperature (8708C) noncatalytic process,
suitable for use with FCCUs having a CO boiler, uses NH3 to convert NOx to
N2 and H2O (59).
5.4. CO2 Emissions Reduction Recently, the concern of CO2 emissions
as a contributor to green house gas levels has drawn attention to the FCCU as a
point source. In a FCCU based refinery, 20% of the CO2 emitted originates in the
FCCU (64). Carbon dioxide is the natural result of complete combustion of the
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 463
coke in the catalyst regeneration, the heat from which is also essential to heat
balance the unit. As a result of the Kyoto Protocol, the European Union has
already begun CO2 credit trading, and therefore interest among international
refiners is very strong for ways to lower FCCU based CO2 emissions. The United
States, having withdrawn from the Kyoto Protocol, has announced a separate
strategy for addressing climate change, and several climate change bills have
been introduced in the 108th Congress. The primary vehicle for climate change
legislation is comprehensive energy legislation (65).
To maximize the performance of an FCCU, most units run at one or more unit
constraints. Frequently, one of these constraints is the regenerator temperature,
which is set by metallurgical limits for safe operations. Process variables on both
the reactor and the regenerator side are thus manipulated to keep the regenera-
tor temperature as close as possible to this regenerator temperature limit.
Changes in feed quality, which will affect the coking tendency of the feed
and the coke-making characteristics of the circulating catalyst, influence the
regenerator temperature. The refiner can rapidly respond to these changes by
manipulating certain independent process variables, which include reactor
riser temperature, temperature of the hydrocarbon feed entering the riser,
fresh catalyst addition rate, and cooling duty if a catalyst cooler exists. If the
change in feed quality is of prolonged duration, the refiner can also change
catalyst type.
The heavy material in the reactor feed, which remains as a residue after
evaporation and pyrolysis of the oil, is designated as Conradson carbon (ASTM
test method D189-76). This residue is frequently used as an indicator of the
coking tendency of the feed. Incorporating heavy feedstocks such as vacuum
resids into the FCCU feed greatly increases the feed’s Conradson carbon
content. The influence of increasing resid content on regenerator temperature
has been determined and correlated against the feed Conradson carbon content
(66). A 2% increase in Conradson carbon can cause a regenerator temperature
increase of 288C. Other studies indicate that an increase in the resid content
of the feed that results in an increase in Conradson carbon from 1.5 to 2.9%
produces a regenerator temperature increase of 378C at constant riser tem-
perature (13).
The presence of contaminant metals on the equilibrium catalyst can signi-
ficantly increase the catalyst coking tendency, which in turn results in an
increase in regenerator temperature if all other factors remain unchanged. As
one example, if the metals on an FCCU equilibrium catalyst increased from an
equivalent-nickel value of 2000 to 3500 wt ppm, the catalyst coking tendency
would increase 30–50%. If all controllable parameters remained constant, the
regenerator temperature would be expected to increase 35–508C, resulting in a
decrease in catalyst circulation rate, and a subsequent drop in conversion.
To decrease the regenerator temperature, heat must be taken out of the
system either by
464 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
Since the first fluid-bed catalytic cracking unit was commissioned in 1942, the
process has seen considerable improvement in mechanical construction, reliabil-
ity, and process flow. A modern FCCU typically operates continuously for 3–
5 years between turnarounds, during which time quantities on the order of
1010 kg of feedstock are processed and quantities of catalyst in the range of
7 1010 kg are circulated through the regenerator. Early FCCU designs (69),
were complex compared with the configuration of more recent designs.
In the modern unit design, the main vessel elevations and catalyst transfer
lines are typically set to achieve optimum pressure differentials because the
process favors high regenerator pressure, to enhance coke-burning kinetics
and power recovery from the flue gas, and low reactor pressure to enhance pro-
duct yields and selectivities.
Comparing a modern design with a typically less desirable configuration of
the past (70), the principal improvement factors have been aimed toward:
reduced catalyst inventory and temperature limitations, enhanced coke-burning
kinetics, catalyst and air contacting; and catalyst residence time distribution.
New designs have eliminated horizontal transport, long standpipes, additional
standpipe aeration requirements, and catalyst fines recycle.
On the regenerator side, there have been several significant design changes
from the bed regenerators that were utilized in early FCC unit designs. In the
1970s, with an emphasis on more complete combustion of the coke deposits, a
fast-fluidized, combustor-style regenerator design was introduced. Some of the
characteristics of this new, widely applied design are described later in this section.
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 465
reactor unregenerated or partially coked, leading to poor activity and low reactor
conversion (52).
Turbulent-Bed Regime. The commercialization of high alumina catalysts
in 1955 was one of the chief reasons for a shift in regenerator design to the tur-
bulent-bed regime. The catalyst was much more active, but highly sensitive to
coke deactivation, so the poor regeneration of the bubbling-bed design was no
longer acceptable. The need to improve mixing of the combustion air and the cat-
alyst, led designs to change to the turbulent fluidization regime. At the higher
gas velocities of the turbulent-bed regime (0.3–1.0 m/s), the distinct bubble
phase disappears, and the bulk of the gas flows in a manner described as: ‘‘In
voids which continually coalesce and split tracing tortuous passages as they
rise through the bed’’ (75). The surface of the upper bed is considerably more dif-
fuse and has reduced pressure fluctuations and substantially higher entrain-
ment of solids into the freeboard region. Because of the higher coke-burning
capacity requirements and improved contacting efficiency, the vast majority of
commercial regenerators converted to the turbulent-bed regime. In this config-
uration, the ultimate regeneration capacity is set by the sharp increase in
solid entrainment that occurs as the linear velocity approaches 1.0 m/s and the
corresponding effect this air velocity has on the cyclone separation efficiency and
dipleg hydraulics (76).
Over the years, considerable efforts have been made to improve regene-
ration efficiency. These efforts focused on improving the radial mixing char-
acteristics and residence-time distributions (reduce short circuiting) of the
turbulent- and bubbling-bed systems. Some of these efforts include: optimizing
fluidized-bed length and diameter ratios, optimizing air distribution through
the proper grid pressure drop and grid plugging pattern, and optimizing catalyst
distribution via cyclone dipleg discharge orientation, and proper location of the
spent catalyst addition and withdrawal points into the regenerator. These addi-
tion and withdrawal devices include tangential inlets, swirl distributors, deflec-
tor plates, and baffles as well as the entry elevation into the fluidized bed (74).
Fast Fluidized Regime. In the late 1970s, the high efficiency combustor
style regenerator was commercialized. This type of regenerator operates in the
fast fluidized regime. The fast fluidized regime (1.0–3.0 m/s superficial velocity)
extends into the transport phase (74), where a sharp increase in the rate of solids
carryover occurs as the transport velocity is approached. In the absence of any
solid recycle, the bed would rapidly disappear. Beyond this velocity, catalyst
fed to the base of the regenerator travels upward in a fully entrained transport
flow. The voidage or density of the resulting suspension is dependent not only on
the velocity of the gas but also on the solid flow rate. If the solid rate is low, a
dilute-phase flow will result. If solids are fed to the regenerator at a sufficiently
high rate, eg, by recirculating the solid carried over back to the regenerator, then
maintaining a relatively large solid concentration is possible. This condition is
referred to as the fast-fluidized bed (73).
The high velocities in the combustor promote catalyst to air contact . As a
result, the catalyst residence time required for catalyst regeneration is greatly
reduced. The highly efficient mixing of the air and the catalyst improves the
overall conversion of CO to CO2, thus minimizing and in many cases eliminating
the need for CO combustion promoter (51).
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 467
Exit wt fraction/time
A
Time, s
8. Mechanical Hardware
From a mechanical point of view, the FCCU regenerator can be divided into two
main sections: the regenerator vessel and its internals, which include the air dis-
tributor, the cyclones, the plenum chamber, and catalyst coolers; and the flue gas
handling section, which includes power, heat, and emission control systems (SOx,
NOx, and particulates). Continuous advances in the mechanical design of both
these areas have led to significant improvements in overall FCCU reliability,
even though processing conditions, at the same time, have become more difficult
(harder and denser catalysts, higher regenerator temperatures).
8.1. Regenerator Vessel and Internals. The FCCU regenerator is one
of the largest vessels in the refinery (up to 18 m in diameter) and operates at tem-
peratures up to 7758C. The regenerator is usually a carbon steel vessel, intern-
ally lined with refractory insulation to maintain the wall temperature in the area
of 3408C, which is suitable for carbon steel.
Air Distributor Grid. The function of the FCCU regenerator air distributor
is to uniformly introduce O2 into the regenerator vessel to achieve even burning
of coke throughout the regenerator cross-section. Every air distributor must be
468 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
designed for three conditions: start-up, normal operation, and emergency. The
actual design temperatures and pressures may vary from unit to unit, depending
on the process requirements and configuration. However, these three design
conditions are common to all units and must be incorporated into the
design. The extreme temperature of the emergency case usually dictates the
selection of the distributor metallurgy.
Basically, two types of air distribution systems are used in FCCU regenera-
tors: those that distribute air into the catalyst bed through a perforated plate and
those that distribute air through a series of small pipes, known as a pipe grid
(77). The pipe grid design is the most commonly used. A typical pipe grid design
is shown in Figure 9. The pipe, which is typically constructed of 304 H stainless
steel or 1 14 Cr-12 Mo steel, distributes the air through three or four large laterals
into a number of small branches. The grid is designed for a total pressure drop in
the range of 3.5–7.0 kPa (0.5–1 psi). This pressure drop is set by regulating the
air leaving the branch arms through a series of nozzles that point downward.
Erosion patterns at the nozzle tip are minimized with the dual-diameter jets
that allow the air to develop a full-flow velocity profile across the entire nozzle
cross-sectional area prior to exiting. Frequently, some of these nozzles are
plugged to prevent jet impingement on adjacent pipes; increase pressure drop;
and enhance carbon and air distribution within the bubbling bed. The upper
operations limit of the grid is based on jet velocity limitations so as not to
cause catalyst attrition. Conversely, the downturn capacity is set by low pressure
drop, which can lead to aspiration of catalyst through the peripheral branches.
Such aspiration results in severe and accelerated erosion.
8.2. Cyclones. The high gas velocities inside the regenerator can result
in considerable catalyst entrainment into the vapor space above the catalyst bed.
Cyclones are used to separate these catalyst particles from the exiting regenera-
tor flue gas and return essentially all of these particles to the regenerator bed. In
a typical regenerator, the cyclones are generally installed in sets of two (Fig. 10).
In this two-stage arrangement, the gas outlet of the first cyclone is channeled
directly to the inlet of the second cyclone.
The cyclones are typically designed with diameters of 100–150 cm for ease
of maintenance. Cyclone inlet velocities are usually restricted to 18–22 m/s
in the first stage and to 21–26 m/s in the second stage to achieve satisfactory
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 469
Vapor outlet
Vapor/catalyst
inlet
P (Hopper)
Catalyst head
Dip legs
P (Dip leg
outlet)
efficiency, pressure drop, and erosion characteristics (57). The number of sets of
two-stage cyclones thus depends on the total gas flow. Finding room to house all
the necessary cyclones within the regenerator frequently requires considerable
ingenuity (78).
Catalyst leaving the cyclone flows down the cyclone discharge pipe, known
as the dipleg, and builds a layer of catalyst at the bottom of this pipe. Sufficient
catalyst head is developed in the dipleg to overcome the cyclone inlet and cyclone
cone outlet pressure drop and thus allow the entrained solids to return to the
regenerator bed. The second stage sets the required overall dip leg length, and
the catalyst seal is maintained with either the dipleg being submerged in the
bubbling bed or with a trickle-flapper valve arrangement at the catalyst dis-
charge end. The open bottom, submerged dipleg with only a splash baffle, is
used primarily on regenerator first-stage cyclones. When the diplegs are unsub-
merged in a bubbling bed, a flapper valve arrangement is utilized at the catalyst
discharge end of the dipleg.
470 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
Fig. 11. Catalyst cooler configurations where (a) is a backmix cooler and (b) is a flow-
through catalyst cooler (79).
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 471
flow. Typically, a single cooler can be operated and controlled anywhere within a
6–50 106 watt saturated-steam production range.
8.4. Flue Gas Handling. Waste Heat Boilers. In a conventional
FCCU flue gas system, the regenerator combustion gases generally pass through
two stages of cyclonic separators, a slide valve, orifice chamber, waste heat boi-
ler, and electrostatic precipitator or third stage separator. The slide valve and
orifice chamber act in combination to reduce the flue gas to essentially atmo-
spheric pressure.
If the CO is not completely combusted to CO2 in the regenerator, a
CO boiler is used to complete the combustion. The resulting heat of combus-
tion and the sensible heat of the flue gas along with any heat from auxiliary
fired fuel are recovered in the form of high pressure steam. When the regenerator
is operated in total CO burn, the CO boiler is replaced with either a shell and
tube exchanger or a box-type waste heat boiler.
The shell and tube exchanger, which is a single-purpose piece of equip-
ment, has the standard parts of a normal heat exchanger, but certain parts
have been modified to withstand the erosive service of the particulate-
laden flue gas stream. This design has the limitation of being able to produce
only saturated steam. The flue gas leaves the regenerator at temperatures up
to 7608C and exits the exchanger at 2908C. The flue gas then proceeds through
the pressure-control slide valve and orifice chamber, which reduces its pressure
to essentially atmospheric prior to entering the emission control system or stack.
The box cooler receives regenerator flue gas after it has been reduced to
essentially atmospheric pressure. This arrangement is not limited to the produc-
tion of saturated steam. Any number of coils can be installed in the box, and nor-
mally both steam-generating and superheater coils are present. The tube
temperatures within the box must be maintained above the SO3 dew point of
150–1758C to prevent sulfuric acid corrosion (80).
The principal differences between the shell and tube and the box designs
may be summarized as follows:
8.5. Power Recovery. All FCC Units emit a form of ‘‘zero cost’’ energy
as part of the flue gas discharge from the FCC regenerator. Typically, these
gasses are in the 7008C range at anywhere from 240 to 380 kPa. In many
cases, this hot, pressurized gas can economically be utilized by a Power Recov-
ery System to either drive the FCC main air blower, or produce electricity for
the refinery grid.
Types of Systems. There are several possible arrangements of Power
Recovery Systems available for incorporation into new or existing FCC Unit
operations. These systems can include 2, 3, 4, and 5 component systems (expan-
der, main air blower, steam turbine, gear, motor/generator), specifically designed
to meet a multitude of processing and economic requirements (81).
472 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
For new FCC unit applications, where plot space is generally not a primary
consideration, the power recovery expander will most often be directly connected
to the main air blower to drive the compressor in place of a steam turbine or
motor drive. This eliminates the need for either a large electric motor, or a
steam turbine to drive the main air blower requiring a surface condenser, utility
support system, and condensate pumps.
For revamp applications, where plot space is most often a primary consid-
eration, and the main air blower and driver are already in place, it is generally
unlikely that a new expander can economically be positioned, and directly con-
nected to the existing main air blower. In these situations, a stand-alone power
recovery expander is typically used to drive an electrical generator to produce
electricity for the refinery, rather than drive the main air blower. This installa-
tion can be a three or four component system. Generally, the three component
power recovery system (expander, gear, generator), protected by a third stage
separator, is the easiest configuration to add to an existing FCC unit. On
many occasions, the addition of a let-down steam turbine (four body system)
can be added to this train to generate additional power. In all cases, the installa-
tion of a power recovery system requires a large capital investment.
Process Flow. In a typical power recovery system, hot flue gas leaving
the regenerator goes to a third stage separator, where roughly 98% of the 10 m
and larger particulate material is removed to protect the expander blades in the
power recovery unit, as described earlier. Other particulate removal systems
(ESP or scrubber) are not appropriate for power recovery use, since they operate
at essentially atmospheric pressure.
Fig. 12. Regenerator Flue Gas System, Before and After Installation of a Power
Recovery System.
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 473
This ‘‘clean’’ flue gas leaving the separator is then sent through two butter-
fly control valves used to direct and throttle the flue gas, etc, and through the
Power Recovery Expander where the pressure in the flue gas is converted to
direct mechanical energy to drive a generator and produce electricity. The expan-
der acts as a restriction orifice and together with the inlet line butterfly valves,
maintains the regenerator operating pressure. A bypass valve is provided
upstream of the inlet butterfly valves to divert a portion of the flue gas around
the expander during periods when the potential recoverable energy exceeds the
capacity of the Power Recovery Train, or when the Power Recovery Train needs
to be bypassed for maintenance. Finally, the expander exhaust gas, generally
110–120 kPa at this point, is directed back to the flue gas system (downstream
of the orifice chamber) for further processing: steam production, particulate
removal, and atmospheric discharge, as seen in Figure 12.
Power Production. The power recovered from the flue gas is approxi-
mately equal to the main air blower requirements. On a BPSD basis, using a
50,000 BPSD FCC (2.5 MM t/year) operating at 310 kPa (30 PSIG), producing
approximately 200,000 kg/h (450,000 lb/h) of flue gas at 7208C (13258F), one
could expect this system to produce 11 MW of electrical power.
BIBLIOGRAPHY
‘‘Catalyst, Regeneration Fluid’’ in ECT 4th ed., Vol. 5, pp. 419–448, by L. L. Upson and
D. A. Lomas, UOP Research Center; ‘‘Catalyst Regeneration, FCC Units’’ in ECT (online),
posting date: December 4, 2000, by L. L. Upson, D. A. Lomas, UOP Research Center.
CITED PUBLICATIONS
13. A. Voorhies, Jr., Ind. Eng. Chem. 37, 318 (Apr. 1945).
14. V. W. Weekman, Jr., and D. M. Nace, AIChE J. 16, 397 (May 1970).
15. R. W. Mott, Oil Gas J., 73 (Jan. 26, 1987).
16. R. R. Dean, J. L. Mauleon, and W. S. Letzsch, Oil Gas J., 75 (Oct. 4, 1982).
17. J. S. Magee, W. E. Cormier, and G. M. Woltermann, Oil Gas J., 59 (May 27, 1985).
18. L. L. Upson, ‘‘How to Design Your Own FCC Catalyst,’’ presented at Katalistiks 8th
Annual FCC Symposium, Budapest, Hungary, June 1987.
19. G. H. Dale and D. L. McKay, Hydrocarbon Process. 56, 97 (1977).
20. P. Ramamoorthy, A. R. English, J. V. Kennedy, L. W. Jossens, and A. S. Krishna,
‘‘A New Metals Passivator in Fluid Catalytic Cracking,’’ presented at 1988 Annual
NPRA Meeting, San Antonio, Tex., Mar. 1988.
21. W. S. Letzsch, A. G. Ashton, and L. L. Upson, Hydrocarbon Proc. 70, 89 (June 1991).
22. C. C. Wear ‘‘Orion-LC Catalysts Set New Standards for Nickel Tolerance,’’ paper pre-
sented at Grace Davison FCC Technology Conference, Athens, Greece, September, 1994.
23. E. G. Wollastin, W. J. Haflin, W. D. Ford, and G. J. D’Souza, Hydrocarbon Proc., 93
(Sept. 1975).
24. P. L. Walker, F. Rusinko, and L. G. Austin, Advances in Catalysts XI, Academic Press,
New York, 1959, pp. 133–221.
25. J. R. Arthur, Trans. Faraday Soc. 47, 164 (1951).
26. P. B. Weisz, J. Catal. 6, 425 (1966).
27. F. E. Massoth, I. E.C. Proc. Des. Devel. 6, 200 (Apr. 1967).
28. P. B. Weisz and R. B. Goodwin, J. Catal. 6, 227 (1966).
29. J. C. Dart, R. T. Savage, and C. G. Kirkbride, Chem. Eng. Prog. 45, 102 (1949).
30. R. G. Haldeman and M. C. Botty, J. Phys. Chem. 63, 489 (1959).
31. T. B. Metcalfe, Brit. Chem. Eng. 12, 388 (Mar. 1967).
32. T. Hano, F. Nakashio, and K. Kusanoki, J. Chem. Eng. (Jpn) 8, 127 (1975).
33. L. L. Upson and H. van der Zwan, Oil Gas J., 65 (Nov. 23, 1987).
34. B. Lewis and G. Von Elbe, Combustions, Flames, and Explosions of Gases, 2nd ed.,
Academic Press, New York, 1961, p. 71.
35. U.S. Pat. 4,064,039 (Jan. 1978), J. E. Pennick.
36. U.S. Pats. 4,072,600 (Feb. 7, 1978) and 4,093,535 (June 6, 1978), A. B. Schwartz.
37. Y. Guegan, ‘‘Use of Combustion Promoter,’’ presented at Katalistiks 1st FCCU
Symposium, Bordeaux, France, Oct. 1980.
38. L. L. Upson, ‘‘Catalytically Promoted Combustion Improves FCC Operations,’’ paper No.
AM 79–39 presented at the 1979 NPRA Annual Meeting, San Antonio, Tex., Mar. 1979.
39. A. Vierhelig, R. Becker, and M. Evans, ‘‘The Role of Additives in Reducing FCC Emis-
sions to Meet Legislation,’’ paper No. AM-03-97 presented at the 2003 NPRA Annual
Meeting, San Antonio, Tex., Mar. 2003.
40. A. W. Peters, X. Zhao, and G. Weatherbee, ‘‘The Origin of NOx in the FCCU Regenerator,’’
Paper No. AM-95-59, presented at the 1995 NPRA Annual Meeting, San Francisco, Calif.
41. Phillip K. Niccum, Eusebius Gbordzoe, Stephen Lang, ‘‘FCC Flue Gas Emission
Control Options,’’ Paper No. AM-02-27 presented at the 2002 NPRA Annual Meeting,
San Antonio, Tex., Mar. 2002.
42. V. J. Memmott and B. Dodds ‘‘Innovative Technology Meets Processing and Environ-
mental Goals: Flying J Commissions New MSCC and TSS,’’ Paper AM-03-13
presented at the 2003 NPRA National Meeting, San Antonio, Tex., March, 2003.
43. L. L. Upson, R. J. Lawson, W. E. Cormier, and F. J. Baars, Oil Gas J., 64–74 (Oct. 1, 1990).
44. M. T. Smith and T. F. Petti, High Unit Retention Critical to Efficient Operation of
Power Recovery Trains, Grace Davison Catalagram, No. 88, 2001.
45. B. Poot, F. Fabiani, L. Reuvers, A. Selvaggi, L. White, Sulfur Dioxide Emissions from
Oil Refineries and Combustion of Oil Products in Western Europe and Hungary,
CONCAWE REPORT No. 10/02.
Vol. 1 FLUID CATALYTIC CRACKING, UNITS, REGENERATION 475
46. J. L. Mauleon and J. Marvillet, ‘‘Control and Monitoring of FCC Flue Gas SOx Emis-
sions,’’ presented at 83rd Annual Meeting Air and Waste Management Association,
Pittsburgh, Pa., June 1990.
47. R. H. Gilman, M. Y. Asim, and T. A. Reid, ‘‘Optimizing FCC Operations using
Pretreatment to Meet Future Market Challenges,’’ paper No. AM-91-37, presented
at the 1991 NPRA Annual meeting, San Antonio, Tex., March 1991.
48. L. D. Krenzke, K. Baron, ‘‘FCC Pretreating to meet New Environmental Regulations
on Gasoline,’’ Paper No. AM-95-67 presented at 1995 NPRA Annual Meeting, San
Francisco, Calif., March 1995.
49. T. A. Nguyen, and M. Skripek, ‘‘Reducing Sulfur in FCC Gasoline via Hydrotreating’’
Paper #54C, Presented at AICHE 1994 Spring meeting April 1994.
50. B. J. Arena, M. W. Schnaith, W. R. Abma, H. Dijkman, S. F. Meyer, and D. C. Nokes,
‘‘State of the Art Processes for Handling FCC Flue Gas Emissions,’’ Paper No. ENV-99-
186 presented at 1999 NPRA Environmental Conference, Dallas, Tex., September 1999.
51. CONCAWE Air and Water Quality Management Group, Special Task Force AQ/STF-
55 and WQ/STF-28, ‘‘Best Available Techniques to reduce emissions from Refineries,’’
Document No. 99/01, Brussels, May 1999.
52. J. W. Wilson, Fluid Catalytic Cracking; Technology and Operation, PennWell
Publishing Co. 1997.
53. E. H. Weaver, M. J. Barrasso, and J. B. Jarvis, ‘‘An Update of Wet Scrubbing Tech-
nology for FCCUS–Multiple Pollutant Control,’’ Paper No. AM-03-120 presented at
2003 NPRA Annual Meeting, San Antonio, Tex., March 2003.
54. A. A. Bhattacharyya, G. M. Woltermann, J. S. Yoo, J. A. Karch, and W. E. Cormier,
‘‘Catalytic SOx Abatement of FCC Flue Gases,’’ presented at 194th National Meeting
of the ACS, New Orleans, La., Aug. 31–Sept. 4, 1987.
55. U. S. Pat. 4,790,982 (Dec. 13, 1988), J. S. Yoo, C. A. Radlowski, J. A. Karch, and A.
Bhattacharyya.
56. J. W. Powell, W. S. Letzsch, R. M. Benslay, K. C. Chuang, and R. Bartek, ‘‘Advanced
FCC Flue Gas Desulfurization Technology,’’ paper No. AN-88-49 presented at the
1988 NPRA Annual Meeting, San Antonio, Tex., Mar. 1988.
57. T. J. Dougan, S. K. Purnell, and J. R. Riley, ‘‘FCC Catalyst Solutions for Gasoline
Sulfur, SOx and NOx,’’ paper presented at the 2003 KBR-Exxon Mobil FCC Licensee
Symposium.
58. A. W. Peters, K. L. Dishman, P. C. Doolin, and L. D. Turlock, Ind. Eng. Chem. Res. 37,
4631 (1998).
59. R. Sadeghbeigi, ‘‘NOx Control in FCC,’’ Paper No. AM-02-55 presented at the 2003
NPRA Annual Meeting, San Antonio, Tex., March 2002.
60. M. L. Batchelder, ed., 1989 NPRA Question and Answer Session on Refinery and
Petrochemical Technology, Gerald L. Farrar & Assoc., Tulsa, Okla., pp. 76–77.
61. R. Miller and M. Skripek, ‘‘Upgrading FCC Feeds with UNOCAL Hydroprocessing
Technology,’’ paper presented at UNOCAL Symposium, Vienna, Austria, June
1990.
62. C. Kuehler, ‘‘Meeting Clean Fuels and Clean Air Objectives with the FCC,’’ paper
presented at the 2003 Akzo Nobel Catalyst Technical Seminar, Banff, Canada,
July, 2003.
63. D. G. Olson, K. Tsuji, I. Shiraishi, and M. R. Miller ‘‘The Reduction of Gas Phase Air
Toxics from Combustion and Incineration Sources Using the MET-Mitsui-BF
Activated Coke Process,’’ paper presented at 1999 International Ash Utilization
Symposium, Lexington, N.Y., Oct. 1999.
64. G. Phillips, ‘‘CO2 Management in Refineries,’’ I. Chem. E. 5th International Gasifica-
tion Conference, Noordwijk, the Netherlands, April, 2002.
65. NPRA Web Site, www.NPRAdc.org, ‘‘Global Climate Change,’’ NPRA Issues.
476 FLUID CATALYTIC CRACKING, UNITS, REGENERATION Vol. 1
66. L. C. Yen, R. E. Wrench, and C. M. Kuo, ‘‘The Important Role of Regenerator Tem-
perature in Catalytic Cracking,’’ presented at Katalistiks 6th Annual FCC Sympo-
sium, Munich, Germany, May 1985.
67. J. R. Murphy and Y. L. Cheng, ‘‘The Interaction of Heat Balance and Operating
Variables in Zeolitic Catalyst Operations,’’ presented at Katalistiks 5th Annual
FCC Symposium, Vienna, Austria, May 1984.
68. G. Wilson, in G. L. Farrar, ed., transcript of 1977 NPRA Q and A Symposium, 1977, p. 54.
69. D. Kunii and O. Levenspiel, Fluidization Engineering, Krieger Publishing, New York,
1977, p. 18.
70. D. A. Lomas, C. A. Cabrera, D. M. Cepla, C. L. Hemler, and L. L. Upson, ‘‘Controlled
Catalytic Cracking,’’ presented at 1990 UOP Technology Conference, Chicago, Ill.,
Apr. 1990.
71. J. L. Mauleon, J. B. Sigaud, and G. Heinrich, ‘‘FCC Heat Balance Management with
Heavy Feeds, MTC Approach,’’ presented at JPI Petroleum Refinery Conference,
Tokyo, Japan, Oct. 1986.
72. D. A. Kauff and M. W. Schnaith, ‘‘FCC Regenerator Options to Process Residue
Feedstocks,’’ paper presented at the Rationalizing Capacity and Profitability in Asian
Oil Refining Symposium, Phuket, Thailand, February, 1997.
73. D. H. Turner, ‘‘Turbulent and Fast Fluidization (High Velocity Fluidization)’’ City
University of N.Y., Ph.D. dessertation, 1979.
74. L. L. Upson, C. L. Hemler, and D. A. Lomas, ‘‘Unit Design and Operational Control:
Impact on Product Yields and Product Quality, Chapter 11 of Fluid Catalytic Crack-
ing Science and Technology,’’ Elsevier, 1993.
75. J. Yerushalmi and N. T. Cankurt, Powder Technol. 24, 187 (1979).
76. Cyclone Separators, API Publication No. 931, May 1975.
77. P. E. Glascow and A. A. Murcia, ‘‘Process and Mechanical Design Considerations
for FCC Regeneration Air Distributors,’’ presented at Katalistiks 5th Annual FCC
Symposium, Vienna, Austria, May 1984.
78. N. L. Giuricich and B. Kalen, ‘‘Dominant Criteria in FCC Cyclone Design,’’ presented
at Katalistiks 3rd Annual FCC Symposium, Amsterdam, The Netherlands, May 1982.
79. C. A. Cabrera, C. L. Hemler, and S. P. Davis, ‘‘Improve Refinery Economics via
Enhanced FCC Operations,’’ presented at Katalistiks 8th Annual FCC Symposium,
Budapest, Hungary, June 1987.
80. R. R. Pierce, Chem. Eng. 84, 125 (Apr. 11, 1977).
81. L. A. Hissink and A. W. Drake ‘‘Power Recovery Expander Experience,’’ paper
presented at Katalistiks 5th Annual FCC Symposium, Vienna, Austria, May. 1984.
GENERAL REFERENCE
D. Selberg and J. R. Riley, Super DeSOx Provides Industry Leading Effectiveness, Grace
Davison Catalagram, No. 92, 2003.
L. L. UPSON
F. S. ROSSER
C. L. HEMLER
P. PALMAS
L. E. BELL
W. J. REAGAN
B. W. HEDRICK
UOP LLC
Vol. 1 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES 477
FLUID CATALYTIC
CRACKING, CATALYSTS
AND ADDITIVES
1. Introduction
FCC catalysts are generally comprised of zeolite, clay, active-matrix, and binder.
The primary source of cracking activity comes from the zeolite component.
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.fluidnee.a01
478 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES Vol. 1
Zeolite Y (faujasite) has been used in FCC catalysts since the early 1960s,
however, a tremendous amount of R&D effort over the years has led to the con-
tinuous introduction of zeolite Y derivatives with a wide variety of performance-
enhancing properties.
The principle modifications involve zeolite ultrastabilization and rare-earth
ion exchange. In the ultrastabilization process, steam calcination is used to
increase the framework Si/Al ratio of the zeolite Y. Not only does this process
result in an increase in the intrinsic stability of the zeolite, it also results in
the development of a secondary mesoporous network of pores that enhance the
bottoms upgrading activity of the zeolite.
Rare earth ion exchange increases the activity of the zeolite and retards
crystal destruction and dealumination in the hydrothermal environment of the
regenerator. The higher the degree of rare-earth exchange of the zeolite, the
higher the unit-cell size of the equilibrated zeolite in the equilibrium catalyst.
Increasing the degree of rare earth exchange increases the gasoline selectivity
of the FCC catalyst.
Most fluid cracking catalysts, especially those used in resid applications,
are manufactured with an active-matrix component usually based on specialty
aluminas. Active-matrix contributes a catalytically active surface that derives
its activity (and selectivity) predominantly from porous silica-alumina that
Vol. 1 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES 479
ZEOLITE
MATRIX CLAY
SYNTHESIS
SYNTHESIS PROCESSING
& PROCESSING
ACTIVE
MATRIX
ZEOLITE INERT
BINDER BINDER
SYNTHESIS
MIXING /
DISPERSION
SPRAY DRYING
FINISHING
Major Advantages:
Flexibility in formulation
Ni
(a) Highly Dispersed - Poor Ni Tolerance
Ni Ni Ni Ni Ni Ni Ni Ni
Ni
Good Ni Support Ni Ni Ni Ni
High Ni Dehydrogenation Activity Ni
Ni Ni
Ni Ni Ni Ni Ni
(b) Nickel Agglomeration - Better Ni Tolerance Ni Ni Ni
Chemical Reaction
Poor Ni Support 100 Å
Low Ni Dehydrogenation Activity
Al NiAl2O4
Al
Vanadium mobility varies from unit to unit, depending on, among other
things, the level of excess oxygen in the regenerator (10). Catalyst manufacturers
have developed more stable zeolites (eg, Z-17, Z-28, Z-30, CSSN, and CSX) and a
series of V traps to increase the ability of the zeolite to handle vanadium. These
traps are based on Ba, Ti, rare earth, and other elements. Some are more effec-
tive than others, but the basic idea is the same, ie, to keep the V away from the
zeolite by binding to the surface of an inactive particle (Fig. 3).
c
Destruction of Zeolite Particle to Particle Mobile H3VO4, V5+ + Steam
and Formation of Unstable
Migration of H3VO4, V5+ on Catalyst
Zeolite as in A and B
b a
RE + H3VO4 Forms
RE Vanadates Making
Zeolite Unstable
9.0
DCR Testing on Resid Feedstock
8.5 Max Coke Selectivity CPS Deactivation at 1450˚F:
Benchmark 4000 ppm Ni, 4000 ppm V
7.5
6.0
5.5
5.0
2 3 4 5 6 7 8
Coke, wt%
Feedstocks: Use low Fe, Na, and Ca feeds and reduce its acid content to lower
equipment corrosion.
FCC unit: Minimize the regeneration temperature.
FCC catalyst: Use Al2O3 bound catalysts instead of SiO2 bound ones.
Vol. 1 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES 483
reactions proceeding along the length of the riser. Contact time over the length of
the riser is 5–8 s. At the end of the riser, the catalyst and products are sepa-
rated in the disengager. It is here that product vapor can undergo nonselective
(essentially thermal) cracking as hydrocarbon vapors come into contact with hot,
coked catalyst.
In the early 1990s, a closed cyclone system was introduced that featured
cyclones directly coupled to the end of the riser (13), which allowed rapid separa-
tion of the catalyst and products, reducing post riser cracking in the disengager.
Today, all major process licensors offer advanced riser termination technologies
of one sort or another with one objective in common—reduction of nonselective
post-riser cracking.
New processes are emerging where the catalyst/feed contact time is reduced
to <2 s. The so-called millisecond catalytic cracking (MSCC) process was intro-
duced in the early 1990s in the United States (14).
4.1. FCC Catalyst Design Considerations for SCT Operations.
Catalysts for SCT operations must be designed taking into account the following
considerations (15):
F Gasoline 2–5%
Feed Sulfur
C Light Cycle Oil 10–25%
2–30%
Skripek (21) correlated sulfur in FCC gasoline with the sulfur species in FCC
feed. Alkylthiophenes and aromatic sulfides are believed to be the key contribu-
tors to gasoline range sulfur.
The concentration of sulfur in gasoline is only weakly dependent on con-
version as a result of opposing effects (18,21). Mercaptans and alkylthiophenes
are unstable with respect to desulfurization. The concentration of these species
decrease with increasing contact time or as gasoline overcracking occurs. How-
ever, benzothiophene, formed by dealkylation of higher molecular weight sul-
fur compounds, is relatively stable with respect to desulfurization. In general,
the concentration of benzothiophene increases with conversion. At constant
conversion, the concentration of sulfur in gasoline increases with riser tem-
perature due to increased dealkylation reactions of aromatic sulfur species.
Benzothiophene
2.8e5
2.7e5
Methylthiophenol
2.6e5
2.5e5 Methylthiophene
2.4e5 Thiophene
Benzenethiol
Me-THT
2.3e5
C3-thiophene
C2-Thiophene
2.2e5
C2-THT
Refinery B
60
55
50
Stack Opacity North, %
45
Si-Sol
Si-Sol Al-Sol
Al-Sol
40
35
30
25
20
15
10
–1 month 0 +1 month +3 months +5 months
Fig. 8. Al-Sol catalysts show outstanding unit retention in commercial operations (8).
FCC unit SOx control options include (1) feed hydrodesulfurization, (2)
‘end-of-pipe’ solutions such as flue gas scrubbing, and (3) the use of SOx-transfer
catalysts. Options (1) and (2) are capital intensive but do provide additional
benefits. The SOx-transfer catalysts or additives require little or no capital and
can be implemented quickly.
SOx Additives. SOx additives have been commercially available since
the early 1980s. These early additives were marginally effective at reducing
SOx. However, this changed with the introduction of DESOX. Its patented
spinel structure has proven to be the most effective material available for
SOx reduction. DESOX reduces (or more correctly transfers) SOx through a
three-step mechanism: (1) oxidation of SO2 to SO3, (2) capture of SO3 by pro-
prietary metal spinel, and (3) release of H2S in the FCC reactor (Fig. 9) (29).
DESOX technology has been used extensively throughout the refining
industry (>100 units worldwide) (27). Refiners have achieved SOx reductions
in excess of 90%, which has usually been accomplished at a cost of $500–
900/ton of SOx removed. There is no thermodynamic limit to the amount of
SOx that can be reduced using DESOX. Indeed commercial experience at several
locations shows that emissions can be <10 vppm.
Figure 10 shows the results of a test a refiner conducted to determine if
reaching 25 vppm SOx was feasible using a catalytic additive (30). Since the
only goal of the test was to determine if 25 vppm could be reached, the base
case was considered relatively unimportant, and was reduced to one day. Upon
base loading to 1% of inventory, the refiner saw SOx emissions drop to the 25 ppm
range, and was able to hold emissions at that level for an extended period with
DESOX additions of 3 wt% of fresh catalyst additions.
No yield debits have been reported with DESOX (30), even at addition rates
>10% of catalyst additions, and DESOX has not had any negative effects on equi-
librium catalyst properties (30).
490 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES Vol. 1
200
180
160
140
120
100
80
60
40
20
0
9/17/2001 9/19/2001 9/21/2001 9/23/2001 9/25/2001 9/27/2001 9/29/2001 10/1/2001
Fig. 10. DESOX is able to sustain SOx emissions below 25 vppm at refinery A . Average
SOx reduction is 85% with 71% of the SOx emissions below 25 ppm.
Vol. 1 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES 491
Pt Promoter
Pt Promoter
N2 HCN, NH 3, ...
CO2
A nitrogen balance around an FCC unit shows that 50% of the feed nitro-
gen ends up in the coke deposited on the catalyst (31). The burning of the nitro-
gen containing coke during the regeneration, generates primarily N2, but also
NO and occasionally small amounts of NO2 and N2O.
The chemistry of NOx formation has been studied in detail and a proposed
mechanism is summarized in Figure 11. According to this scheme, NO can be
reduced to N2 with the help of reductants such as coke, CO, and hydrocarbons.
Conventional combustion promoters (Pt/Al2O3) used to oxidize CO to CO2 have
been shown to increase NOx emissions. One likely mechanism is that they also
oxidize nitrogen intermediates (eg, NH3 and HCN). In 1996, a low-NOx combus-
tion promoter, XNOx, was introduced and shown to reduce NO emissions by
>50%. This NO reduction was achieved while maintaining good CO combustion
promotion. The results from the commercial use of a low-NOx promoter are
shown in Figure 12 (30). In this case, replacing conventional CO promoter
with XNOx resulted in a 60–70% NOx reduction while maintaining afterburn
and CO emissions control.
NOx reduction can also be achieved with NOx reducing additives that do not
have CO oxidation functionality. DENOX, was introduced commercially in 1997
and has been shown to decrease NOx by >50%.
In the United States, Consent Decree protocols typically call for replacing
conventional combustion promoters with low NOx promoters and conducting
optimization of NOx reducing additives at various levels. Figure 13 summarizes
Consent Decree test results with DENOX (30). In these trials, uncontrolled or
‘‘Base NOx’’ has been estimated using six sigma modeling techniques. Drivers
for NOx formation were found to be unit specific, with models containing between
4 and 17 input variables, including nonlinear and interaction terms. ‘‘Reduced
NOx’’ is the average NOx measured during the trial. In some instances, XNOx
additions started coincidentally with DENOX and it is not possible to separate
the relative contribution of each additive.
The amount of additive used in each trial varied greatly, as did the
results, with observed NOx reductions as high as 80%. The effectiveness
of an additive for controlling NOx emissions depends on unit design, mechan-
ical condition, operating parameters, feedstock, etc. The chemistry of NOx
formation and reduction is very complex. As a result the performance of NOx
additives is unit specific.
492 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES Vol. 1
120 50
160
>50% Reduction in Flue
140 Promoted Base
120
100
80
60
40
20
0
0 500 1000 1500 2000 2500 3000
CO Emissions (ppm)
300
50%
250
47%
ppm NO in flue gas
200
8%
150
17%
80%
67%
100
19%
0%
50
0
0.12 0.6 0.6 0.6 0.6 1.0 2.0 2.5
DENOX Addition Rate (% of fresh catalyst addition)
The next 2–3 years will see a significant growth in the demand for propylene
derivatives, particularly polypropylene, that will lead to structural tightness in
the global propylene market. Since steam cracker capacity expansion, which is
tied to a lower growth rate in the demand for ethylene, is forecast to lag pro-
pylene demand and much of the new planned capacity is in the Middle East
where light feedstocks will be processed, the petrochemicals industry is looking
to operators of FCC processes to fill the propylene supply gap (32).
In the short term, this will lead to opportunities for refiners to further
increase margins by using ZSM-5 based additive technologies to produce incre-
mental propylene for sale into the petrochemicals market. In the long term, a
number of refiners are considering either revamping their FCC units to operate
in ‘‘petrochemicals mode’’ or are considering investing in grass routes ‘‘on pur-
pose’’ propylene production facilities that utilize ZSM-5 based FCC catalyst
technologies.
7.1. ZSM-5 Technologies. The addition of ZSM-5 additives is the sin-
gle most important method to increase the yields of light olefins, particularly pro-
pylene, at the expense of gasoline (Fig. 14). Davison Catalysts was the first
catalyst manufacturer to supply ZSM-5 based additives for use in FCC units
back in 1984, where their primary role was to enhance gasoline octane. Since
then, the role of these additives has changed to that of optimizing the yields of
high value C3¼ and C4¼ olefins, while at the same time capturing the boost in
gasoline octane (33).
ZSM-5 additives utilize unique matrix technologies that are designed not
only to stabilize the ZSM-5 zeolite and provide the maximum activity per
pound of zeolite, but also to minimize unselective cracking that reduce the
yield of valuable light olefins. In addition, additives are designed with excellent
(a) (b)
ca.10wt% Reduction in
Fresh Catalyst Addition Rate
70 ECat Microactivity 75
Time
Fig. 15. OlefinsUltra enabled a european refiner to reduce fresh catalyst addition rate
while maintaining high propylene yield.
level to which the additive equilibrates will depend on how the ZSM-5 zeolite was
stabilized during manufacture (34).
Nuclear magnetic resonance (nmr) spectroscopy analysis indicates that pho-
sphorus-stabilized ZSM-5 zeolite is less dealuminated after steam deactivation
than standard nonstabilized additives (34). This observation implies that the
higher activity/stability of the ZSM-5 in these additives is due to a retardation
of the dealumination rate caused by the phosphorus stabilization process.
Two independent experimental results substantiate this conclusion. Figure 17
shows the 27Al nmr spectra of (a) a fresh (heat-treated nonstabilized) ZSM-5
and two hydrothermally deactivated ZSM-5 samples; (b) nonstabilized ZSM-5;
3.5
Delta Propylene Yield, wt%
2.5
1.5
0.5
0
1200 °F 1500 F/4 hrs/100% Stm
Unstabilized Stabilized
27
Fig. 17. Al nmr spectra of stabilized and unstabilized ZSM-5 crystal.
and (c) P-stabilized. As can be seen, the phosphorus stabilized ZSM-5 retains a
significant amount on aluminum in the framework compared to the unstabilized
version as evidenced by the large tetrahedral Al peak. Again, the activity reten-
tion of the additive in the actual unit will depend on the stabilization process
carried out during manufacture.
Differences in phosphorus stabilization of the ZSM-5 zeolite account for
much of the performance differences between additives from different suppli-
ers. Surface area and percent ZSM-5 crystal content are not good indicators
of catalytic performance especially between additives from different suppliers.
ZSM-5 additive performance can vary substantially as a result of differences
in ZSM-5 zeolite crystallization and stabilization. This has been seen commer-
cially (34).
8. Conclusions
The FCC unit remains the primary hydrocarbon conversion unit in the modern
petroleum refinery. To further improve profitability, refiners are increasing the
amount of lower cost, residual feedstocks processed. These feeds are not only
heavier, but also contain high levels of contaminant metals compared with con-
ventional vacuum gas oils. Zeolite and matrix technologies in the FCC catalyst
have evolved to maximize conversion of these difficult feeds to valuable transpor-
tation fuels and light olefins.
Increasingly stringent environmental regulations have also changed the
objective function of the modern refinery. Special gasoline sulfur reduction addi-
tive and catalyst technologies are now being widely applied in FCC units to help
Vol. 1 FLUID CATALYTIC CRACKING, CATALYSTS AND ADDITIVES 497
BIBLIOGRAPHY
CITED PUBLICATIONS
16. Auto/Oil Air Quality Improvement Research Program, Phase I Final Report, May,
1993.
17. S. K. Purnell, D. A. Hunt, and D. Leach, ‘‘Catalytic Reduction of Sulfur and Olefins
in the FCCU’’, NPRA Annual Meeting, Paper AM-02-37, San Antonio, Tex., March
2002.
18. R. F. Wormsbecher, G. D. Weatherbee, G. Kim, and T. J. Dougan, ‘‘Emerging Tech-
nology for the Reduction of Sulfur in FCC Fuels’’, NPRA Annual Meeting, Paper
AM-93-55, San Antonio, March 1993.
19. E. G. Wollaston, W. L. Forsythe, and I. A. Vasalos, Oil Gas J. 64 (Aug. 2, 1971).
20. G. P. Huling, J. P. McKinney, and T. C. Readal, Oil Gas J. 73 (May 19, 1975).
21. T. A. Nguyen and M. Skripek, AIChE Spring National Meeting, Atlanta, April,
1994.
22. R. W. Mott, T. G. Roberie, and X. Zhao, ‘‘Suppressing FCC Gasoline Olefinicity while
Managing Light Olefins Production’’, NPRA Annual Meeting, Paper AM-98-11, San
Francisco, March 1998.
23. J. W. Balko, D. Podratz, and J. A. Olesen, ‘‘Grace Davison FCC Catalytic Technolo-
gies for Managing New Fuels Regulations’’, NPRA Annual Meeting, Paper AM-00-14,
San Antonio, Tex., March 2000.
24. Petroleum Technol. Q. 3 (1) 39 (1998).
25. F. J. A. Geiger, H. W. A. Dries, and E. Bakker, Erosion, Fouling and Cracking of FCC
Cyclone Systems Resolved, VDI Berichte No. 1511, 1999.
26. H. von Uetz and K.-J. Grob, Stromungsmechanische Grundlagen der Erosion,
Literatur S. 235, pp. 228–235.
27. Davison KBR conference paper.
28. J. G. McDaniel and D. J. Neuman, NPRA Annual Meeting, Paper AM-92-60,
New Orleans, March 1992.
29. S. W. Davey, Environmental Fluid Catalytic Cracking Technology, in Proceedings of
the European Refining Technology Conference, Paris, France, Nov. 1999.
30. T. J. Dougan and J. R. Riley, ‘‘SOx, NOx and Consent Decrees; The FCCU Additive
Perspective’’, NPRA Environmental Conference, Paper ENV-02-198, New Orleans,
September 2002.
31. A. W. Peters, G. D. Weatherbee, and X. Zhao, Fuel Ref. 5(3), 45 (1995).
32. S. J. Zinger, ‘‘Propylene Supply for the Global Petrochemical Markets’’, Catalagram
No. 94, Special Edition Propylene, Davison Catalysts Publication, 2004.
33. J. R. D. Nee and M. Lesemann, ‘‘Propylene Production in the FCC Unit’’, Catalagram
No. 94, Special Edition Propylene, Davison Catalysts Publication, 2004.
34. M. Lesemann, J. R. D. Nee, and S. J. Zinger, ‘‘The Evolving Role of Fluid Catalytic
Cracking in the Petrochemicals Market’’, NPRA Annual Meeting, Paper AM-04-58,
San Antonio, Tex., March 2004.
J. R. D. NEE
R. H. HARDING
G. YALURIS
W. C. CHENG
X. ZHAO
T. J. DOUGAN
J. R. RILEY
W. R. Grace & Co.
Vol. 1 CATALYSIS 499
CATALYSIS
1. Introduction
C RC
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0301200107012005.a01.pub2
500 CATALYSIS Vol. 1
2. History
The science of catalysis is driven by technology, as it has been from the begin-
ning. Some of the earliest known examples of controlled chemical transforma-
tions are catalytic. Fermentation was used in ancient times to make alcoholic
beverages, and a number of the earliest examples of chemical technology were
also exploitations of catalysis (1). For example, before the sixteenth century
ether was made by distilling spirits in the presence of sulfuric acid. In 1746,
nitric oxide was used as a catalyst in the lead chamber process for oxidation of
sulfur dioxide to give sulfur trioxide in the manufacture of sulfuric acid. In 1781,
acids were used to catalyze the conversion of starch into sugar. In 1817, H. Davy
discovered that in the presence of platinum, mine gases were oxidized at low
temperatures; he designed a safety lamp for miners in which the platinum
glowed if the flame was extinguished (2). Many more practical discoveries
were made in the 1800s (1,2).
These examples existed prior to Ostwald’s 1895 definition (1), ie, before the
nature of catalysis was well understood. But in 1850 Wilhelmy (3) had already
made the first measurements of kinetics of catalytic reactions in an investigation
of sugar inversion catalyzed by mineral acids. In the years following Ostwald’s
definition, just as the principles of chemical equilibrium and kinetics were
502 CATALYSIS Vol. 1
becoming known, the field of catalysis became more quantitative and developed
rapidly. Kinetics of surface-catalyzed reactions were measured by Bodenstein
(4) just after the turn of the century. The defining work that set the stage for
modern catalytic technology was the development of the ammonia synthesis
process by Haber, Bosch, Mittasch, and co-workers, beginning 1908 (1,5).
These pioneers worked out methods of catalyst testing and process develop-
ment that are essentially the methods of choice today (6,7). They understood the
interplay between chemical equilibrium and reaction kinetics; indeed, Haber’s
research, motivated by the development of a commercial process, helped to
spur the development of the principles of physical chemistry that account for
the effects of temperature and pressure on chemical equilibrium and kinetics.
The ammonia synthesis reaction is strongly equilibrium limited. The equilibrium
conversion to ammonia is favored by high pressure and low temperature. Haber
therefore recognized that the key to a successful process for making ammonia
from hydrogen and nitrogen was a catalyst with a high activity to allow operation
at temperatures low enough for a relatively favorable equilibrium.
Bosch and co-workers (6) devised laboratory reactors to operate at high
pressure and temperature in a recycle mode. These test reactors had the essen-
tial characteristics of potential industrial reactors and were used by Mittasch
and co-workers (6) at BASF in Germany to screen some 20,000 samples as can-
didate catalysts. The results led to the identification of an iron-containing
mineral that is similar to catalysts used in industry today. The researchers
recognized the need for porous catalytic materials and materials with more
than one component, today identified as the support, the catalytically active com-
ponent, and the promoter. Today’s technology for catalyst testing has become
more efficient because much of the test equipment is automated and some is
miniaturized. The analysis of products and catalysts is also much faster and
more accurate today.
3. Homogeneous Catalysis
complexes, and must be recovered and recycled with high efficiency. The most
common separation devices are distillation columns; extraction is also applied.
Gas Handling. The reactants are often gaseous under ambient condi-
tions. To maximize the rate of the catalytic reaction, it is often necessary to mini-
mize the resistance to gas–liquid mass transfer, and the gases are therefore
introduced into the liquid containing the catalyst as swarms of bubbles into a
well-stirred mixture or into devices such as packed columns that facilitate
gas–liquid mixing and gas absorption.
Exothermicity. The catalytic reactions are often exothermic bond-forming
reactions of small molecules that give larger molecules. Consequently, the reac-
tors are designed for efficient heat removal. They may be jacketed or contain coils
for heat-transfer media, or the heat of reaction may be used to vaporize the pro-
ducts and aid in the downstream separation by distillation.
There are also a number of generalizations about the chemistry of these
processes. Often the reactants are small building blocks, many formed from
organic raw materials, namely, petroleum, natural gas, and coal. The reactants
include O2, low molecular weight olefins, and synthesis gas (CO and H2). In older
technology, acetylene was a common building block. Many reactions are cata-
lyzed by acids and bases, usually in aqueous solution. Many reactions are cata-
lyzed by transition-metal complexes, usually in nonaqueous organic solvents, but
sometimes in water and in ionic liquids with low melting points such as alkylam-
monium salts. The transition metal complex catalysts used in technology are
often highly selective.
3.2. Influence of Mass Transport on Reaction Rates. When a rela-
tively slow catalytic reaction takes place in a stirred solution, the reactants are
supplied to the catalyst from the immediately neighboring solution so readily
that virtually no concentration gradients exist. The intrinsic chemical kinetics
determines the rate of the reaction. However, when the intrinsic rate of the reac-
tion is relatively high and/or the transport of the reactant is relatively slow, as in
a viscous polymer solution, the concentration gradients become significant, and
the transport of reactants to the catalyst cannot keep the catalyst supplied suffi-
ciently for the rate of the reaction to be that corresponding to the intrinsic che-
mical kinetics. Assume that the transport of the reactant in solution is described
by Fick’s law of diffusion with a diffusion coefficient D, and the intrinsic chemical
kinetics is of the following form
r ¼ k½C½A ð1Þ
where C is catalyst and A the reactant. Then in the general case in which the
diffusion influences the rate (8),
1
Z ¼ r=rmax ¼ ð2Þ
1 þ RAkþRC
4p 2 D
where Z, the ratio of the reaction rate to the maximum value that could occur, ie,
the rate in the absence of a transport influence, is called the effectiveness factor,
and RC and RA are the radii of the catalyst and reactant molecules, respectively.
504 CATALYSIS Vol. 1
There are two limiting cases; if k 4pðRA þ2 RB ÞD, then there is no diffusion influ-
ence and the rate is described by equation 1, where the concentrations are the
bulk average concentrations in the solution. On the other hand, when the
inequality is inverted, the rate becomes
RA þ R B
r ¼ 4p D½C½A ð3Þ
2
In the former case, the rate is independent of the diffusion coefficient and is
determined by the intrinsic chemical kinetics; in the latter case, the rate is inde-
pendent of the rate constant k and depends on the diffusion coefficient; the reac-
tion is then diffusion controlled. This kind of mass transport influence is different
from that characteristic of dissolution of a gas into a liquid phase.
3.3. Examples of Solution Catalysis. Acid-Base Catalysis. Inex-
pensive mineral acids, eg, H2SO4, and bases, eg, KOH, in aqueous solution
are widely applied as catalysts in industrial organic synthesis. Catalytic reac-
tions include esterifications, hydrations, dehydrations, and condensations.
Much of the technology is old and well established, and the chemistry is well
understood. Reactions that are catalyzed by acids are also typically catalyzed
by bases. In some instances, the kinetics of the reaction has a form such as the
following (9):
r ¼ ðk0 ½H2 O þ kHþ ½Hþ þ kHA ½HA þ kOH ½OH þ kB ½BÞ½R ð4Þ
where r is the reaction rate, [R] the concentration of the reactant, and the terms
on the right-hand side represent catalysis by water, hydrogen ions, undisso-
ciated acid HA, hydroxide ions, and undissociated base. Often all the terms
on the right-hand side but one are negligible.
If the second term on the right-hand side in equation 4 is the only signifi-
cant one, then catalysis is by hydrogen ions (hydrated protons), and the catalysis
is called specific acid catalysis; the rate of the reaction is proportional to the con-
centration of the hydrogen ions and to the concentration of the reactant. Many
examples conform to this pattern. Similarly, if the third term on the right-hand
side of equation 4 is the only significant one, then the catalysis is by HA, and
general acid catalysis occurs. If the fourth term is the significant one, hydroxide
ions are the catalyst and specific base catalysis occurs. If the fifth term is the
significant one, general base catalysis occurs.
The chemistry of acid–base catalysis involves the transfer of hydrogen ions,
Hþ and sometimes H . In general or specific acid catalysis, the undissociated
acid molecules or the hydrated protons formed by dissociation of an acid, respec-
tively, donate protons to the reactant. The proton acceptor in an aqueous solution
is a good base, often an organic compound including O, N, or S. In catalysis by
bases, the reactant donates a proton to the catalyst. The proton donor in general
acid catalysis may be a weak or a strong acid, and often the activity of the cata-
lyst increases with its strength as an acid, measured by the acid dissociation
constant Ka.
Donation of a proton to the reactant often forms a carbenium ion or an oxo-
nium ion, which then reacts in the catalytic cycle. For example, a catalytic cycle
Vol. 1 CATALYSIS 505
suggested for the conversion of phenol and acetone into bisphenol A, which is an
important monomer used to manufacture epoxy resins and polycarbonates, in an
aqueous mineral acid solution is shown in Figure 1 (10).
The kinetics of the bisphenol A synthesis has been reported to be of the fol-
lowing form (11)
r ¼ k Hþ ½A½P ð5Þ
CH3
O
HO C OH
H3C C CH3
CH3
H OH OH
H3C C CH3 H3C C CH3
2 1
OH
CH3
OH HO C OH
CH3 H
CH3
H + HO C OH
CH3
CH3
C OH
CH3
H CH3
O C OH
H CH3
H2O
Fig. 1. Catalytic cycle for synthesis of bisphenol A from phenol and acetone in the
presence of a dissociated mineral acid (10).
506 CATALYSIS Vol. 1
where A is acetone and P is phenol. When the reaction is carried out in the pre-
sence of low concentrations of an added compound, thioglycolic acid,
HSCH2COOH, the rate is increased, with the kinetics reported to be of the fol-
lowing form (12)
r ¼ k0 Hþ ½A½Pðk1 þ k2 ½TÞ ð6Þ
C C
C
C C
C
C
C C C C
C
C
C C C C
C C
C C
C
less stable than the secondary carbenium ion formed by protonation of propy-
lene or straight-chain butenes. Ethylene is not used in the commercial alkyla-
tion processes.
Acid-catalyzed reactions of paraffins require much stronger acids than
those of olefins or aromatics, because the paraffins are much weaker
bases. The only acids that catalyze reactions of paraffins alone at relatively
low temperatures are combinations of Brønsted and Lewis acids; examples
include HCl þ AlCl3, HF þ SbF5, and FSO3H þ SbF5 (magic acid). Some of
these are called superacids (13), eg, magic acid. Superacids are stronger acids
than concentrated sulfuric acid and some have values of H0 of 20 and less,
whereas that of 100% sulfuric acid is only about 12. The Hammett acidity
function is logarithmic; thus some superacids are more than eight orders of
magnitude stronger proton donors than sulfuric acid. Hence, these superacids
can protonate paraffins, even methane. They catalyze isomerization and
carbon–carbon bond-forming and bond-breaking reactions at low tempera-
tures but are not found in commercial applications, in part because of their
extreme corrosiveness.
Metal Complex Catalysis. Most of the recent innovations in industrial
homogeneous catalysis have resulted from discoveries of transition-metal com-
plex catalysts. Thousands of transition-metal complexes (including organometal-
lics, ie, those with metal–carbon bonds), are known, and the rapid development
of organotransition metal chemistry in recent decades has been motivated
largely by the successes and opportunities in catalysis. The chemistry of metal
complex catalysis is explained by the bonding and reactivity of organic groups
(ligands) bonded to the metals. The important reactions in catalytic cycles are
those of ligands bonded in the coordination sphere of the same metal atom. Bond-
ing of ligands such as CO or olefin to a transition metal activates them and facil-
itates the catalysis. The following reactions of transition-metal complexes
account for almost all the steps in the catalytic cycles.
508 CATALYSIS Vol. 1
Ligand exchange:
where M is metal and L and L0 are ligands. This reaction may be thermally or
photochemically initiated; it may proceed by dissociation of a ligand to give a
coordinatively unsaturated complex, to which another ligand is then bonded,
but it may also proceed associatively, without prior dissociation of a ligand.
Ligand exchange is one way in which reactants become bonded to catalysts.
Oxidative addition and reductive elimination:
oxidative addition
MLx–2 + AB MLx–2 AB ð8Þ
reductive elimination
The forward reaction requires an electron-rich metal in a low oxidation state that
is coordinatively unsaturated. Oxidative addition provides another way for bond-
ing of reactant ligands to a metal, and reductive elminination provides a way for
products to leave a metal.
Migratory Insertion:
O
C O ð9Þ
Lx– 2M CH3 Lx– 2MC
CH3
In this reaction, one ligand migrates to a neighboring ligand that then becomes
formally inserted between the metal and the ligand that has migrated, creating a
site of coordinative unsaturation so that another reactant ligand can be asso-
ciated with the metal. The migratory insertion reaction accounts for chain-
growth steps of olefin polymerization reactions.
Wilkinson Hydrogenation. One of the best understood catalytic cycles is
that for olefin hydrogenation in the presence of phosphine complexes of rhodium,
the hydrogenation named after the Nobel-laureate Geoffroy Wilkinson (14,15).
The reactions of a number of olefins, eg, cyclohexene and styrene, are rapid, tak-
ing place even at room temperature and atmospheric pressure; but the reaction
of ethylene is extremely slow. Complexes of a number of transition metals in
addition to rhodium are active for the reaction.
The Wilkinson hydrogenation cycle shown in Figure 3 (16) was worked out
in experiments that included isolation and identification of individual rhodium
complexes, measurements of equilibria of individual steps, determination of
rates of individual steps under conditions of stoichiometric reaction with certain
reactants missing so that the catalytic cycle could not occur, and determination
of rates of the overall catalytic reaction. The cycle demonstrates some generally
important points about catalysis: the predominant species present in the react-
ing solution and the only ones that are easily observable by spectroscopic
methods, eg, RhCl[P(C6H5)3]3, RhCl[P(C6H5)3]2 (olefin), and RhCl2[P(C6H5)3]4,
are outside the cycle, possibly in virtual equilibrium with species in the cycle,
and not involved in the kinetically significant steps of the cycle. Often the only
C PPh3, C C H
Ph3P Ph3P PPh3 H2, k2 Ph3P H
C
Rh Rh Rh
H2, k– 2
Cl PPh3 C C PPh3 Cl PPh3 Cl PPh3
, PPh3
K5 C C PPh3
,
PPh3 PPh3 PPh3
PPh3
C C C C k1 k– 1 K4
Ph3P Cl PPh3
H C C H H
Rh Rh Ph3P H2 , k3 Ph3P H Ph3P
Ph3P PPh3 Rh Rh Rh
Cl H2 , k– 3
Cl PPh3 Cl PPh3 C C Cl PPh3
509
K7
H2 H2
C C
Fast
PPh3 K6
Ph3P Cl H C C
Rh Rh H H
H C C H
Ph3P
Ph3P H
Cl Rh
PPh3
Cl PPh3
Fig. 3. Catalytic cycle (within dashed lines) for the Wilkinson hydrogenation of olefin (14,16). Ph represents phenyl (C6H5). Values of rate
constants and equilibrium constants are as follows: k1 ¼ 0.68 s 1; k 1 7
104 L/(mols); K1 10 5 mol/L; k 1/k4 ffi 1; K5 ¼ 3
10 4; k2 ¼
4.8 L/(mols); k 2 ¼ 2.8
10 4 s 1; K2 ¼ 1.7
104 L/mol; k3 7
104 L/(mols); k6 ¼ 0.22 s 1.
510 CATALYSIS Vol. 1
species that can be observed during catalysis are those that are not involved in
the cycle. Consequently, elucidation of catalytic cycles is difficult, and inferences
about cycles that are based on easily measured compositions of reacting solutions
are usually less than well founded. The exceptions to this rule are usually the
cycles of least interest, those of slow catalytic reactions.
The Wilkinson hydrogenation illustrates the roles of the oxidative addition,
reductive elimination, and insertion reactions, all of which occur in the cycle
(Fig. 3). Some of the properties of the rhodium phosphine complexes that explain
their success as catalysts include the following: Rhodium exists in two stable oxi-
dation states, Rh(I) and Rh(III), separated by two units. Oxidative addition and
reductive elimination reactions formally require changes of two units in metal
oxidation state. There are no intermediates that are so stable that they form bot-
tlenecks in the cycle. There is an exquisite dynamic balance, and the cycle turns
over rapidly with only low concentrations of the intermediates being present in a
steady state. The phosphines are electron-donor ligands that are important in
controlling the reactivities of the intermediates and regulating the dynamic bal-
ance. Replacement of triphenylphosphine by other phosphines leads to signifi-
cant changes in the rate of the catalytic hydrogenation. The phosphines can
influence the reactivities of the intermediates, and thus the rate of catalysis,
by virtue of both electronic and steric effects.
Asymmetric Hydrogenation. Biological reactions are stereoselective, and
numerous drugs must be pure optical isomers. Metal complex catalysts have
been found that give very high yields of enantiomerically pure products, and
some have industrial application (17,18). The hydrogenation of the methyl
ester of acetamidocinnamic acid has been carried out to give a precusor of
L-dopa, ie, 3,4-dihydroxyphenylalanine, a drug used in the treatment of
Parkinson’s disease.
H COOCH3 H2
C C CH2CHCOOCH3
NHCOCH3 NHCOCH3
The strategy of the catalyst development was to use a rhodium complex similar
to those of the Wilkinson hydrogenation but containing bulky chiral ligands in an
attempt to direct the stereochemistry of the catalytic reaction to favor the desired
L isomer of the product (17). Active and stereoselective catalysts have been found
and used in commercial practice (although there is now a more economical route
to L-dopa than through hydrogenation of the prochiral precursor). The 2001
Nobel Prize in Chemistry was awarded to W. S. Knowles and R. Noyori, recogniz-
ing their work on these catalysts (and to K. B. Sharpless for work on asymmetric
oxidation catalysis).
Two pathways were found for the chiral hydrogenation, and they give pro-
ducts with different stereochemistries (19). One pathway involves the preferred
mode of initial binding of the reactant to the catalyst. The other pathway
involves an isomer of the reactant–catalyst complex that is formed in only
small amounts, but its conversion is energetically favorable and constitutes the
kinetically predominant pathway to products (9) (Fig. 4). Thus the chirality of
Vol. 1 CATALYSIS 511
⫹
CO2CH3
HN ‡ ‡
G°major
G°minor
P Minor
Rh
CH3 OPh *
P
⫹
Energy
CH3OOC NH
P Ph CH3OOC
* Rh CH3 NH
P O Major H P
*
Rh Ph
CH3
O
P H
CO2CH3
HN P H
*
Ph Rh
CH3
O
H P
Reaction coordinate
Fig. 4. Schematic representation of energy profiles for the pathways for the hydrogena-
tion of a prochiral precursor to make L-dopa (19). The chiral disphosphine ligand is shown
schematically as P * P. Ph represents phenyl (C6H5).
the product is determined not by the preferred mode of the initial binding, but
instead by the more favorable energetics of the pathway involving the minor
isomer of the reactant–catalyst complex.
Methanol Carbonylation. An important industrial process catalyzed by
rhodium (and more recently iridium) complexes in solution is methanol carbony-
lation to give acetic acid.
The catalytic cycle (Fig. 5) (20,21) is well established. The CH3I works as
a cocatalyst or promoter because it undergoes an oxidative addition with
[Rh(CO)2I2] , and the resulting product has the CO ligand bonded cis to the
CH3 ligand; these two ligands are then poised for a migratory insertion reac-
tion. When the catalyst is iridium instead of rhodium, a similar cycle pertains,
but the reaction of [H3CIr(CO)2I3] with CO to give [H3CCOIr(CO)2I3] is rate
determining (21). The migratory insertion in the iridium complex can be accel-
erated by the addition of either methanol or a Lewis acid (SnI2); each appears to
facilitate substitution of an iodide ligand by CO, resulting in faster methyl
migration. The greater stability of [H3CIr(CO)2I3] than of [H3CRh(CO)2I3]
accounts for the different characters of the reactions catalyzed by the two
metals (21).
Methanol carbonylation is one of only a few industrially important catalytic
reactions for which the quantitative reaction kinetics is known (22).
I CO
Rh
I CO
MeCOI MeI
H2O
COMe Me
OC I I CO
Rh HI Rh
I CO I CO
I I
MeCO2H MeOH
CO
COMe
I I
Rh
I CO
Scrubber To flare
system
Recycle
Product
to storage
Carbon
monoxide
Methanol Waste
acid
Reactor Light ends Drying Heavy ends Waste acid
column column column stripper
Fig. 6. Process flow diagram for methanol carbonylation catalyzed by a rhodium complex
to make acetic acid (23).
sufficiently stable. Recent work with more stable vinyl pyridine resins as sup-
ports has led to stable catalysts and apparently to a viable industrial process
(24); this is important as perhaps the first successful application of a supported
catalyst that is nearly analogous to a molecular metal complex catalyst.
Olefin Hydroformylation (The Oxo Process). One of the most important
industrial applications of transition-metal complex catalysis is the hydroformyla-
tion of olefins (25), illustrated for propylene:
CHO
CH3CH CH2 + CO + H2 CH3CHCH3 + CH3CH2CH2CHO
ð13Þ
The linear isomer is more valuable than the branched isomer (see BUTYRALDE-
HYDE). The product aldehydes are hydrogenated to give so-called oxo alcohols;
long-chain products are converted into sulfonates and used as detergents.
The oxo process was discovered in 1938 by Roelen and co-workers of
Ruhrchemie. The first catalysts were cobalt carbonyl complexes formed from
[HCo(CO)4]. They are still used on a large scale. The process is carried out at
temperatures of 1508C and pressures of >30 MPa (several hundred atm). The
high pressures are required to maintain the cobalt in the form of the soluble
metal carbonyl complex; at lower pressures, cobalt metal can form and deposit
514 CATALYSIS Vol. 1
on the reactor walls, possibly in a porous form that adds a significant resistance
to heat transfer. Use of cobalt complexes with phosphine ligands led to processes
with improved selectivities. The process flow diagram is complicated because it is
expensive to separate the catalyst from the products and recycle it in the high
pressure process. In the Kuhlmann process (25), the catalyst is extracted into
an aqueous stream and then regenerated with acid to give [HCo(CO)4], which
is absorbed into the reactant solution.
A significant advance in hydroformylation technology was made with the
discovery that phosphine complexes of rhodium, similar to those used in the
Wilkinson hydrogenation but incorporating H, CO, and olefin ligands as well
as triphenylphosphine, have catalytic activities several orders of magnitude
greater than those of cobalt; furthermore, the rhodium complexes are stable
enough to be used at low pressures, typically 1.5 MPa (15 atm) at 1508C, and
separated by distillation (25). Rhodium complex catalysts now are used in indus-
trial processes for hydroformylation of propylene. When a large excess of triphe-
nylphosphine is used as the solvent, the selectivity for the desired straight-chain
product is high, with >95% of the aldehyde being n-butyraldehyde.
The processing costs associated with separation and corrosion are still sig-
nificant in the low pressure process; for the process to be economical, the effi-
ciency of recovery and recycle of the rhodium must be very high.
Consequently, researchers have continued to seek new ways to facilitate the
separation and confine the corrosion. Extensive research was done with rhodium
phosphine complexes bonded to solid supports, but the resulting catalysts were
not sufficiently stable, as rhodium was leached into the product solution. A suc-
cessful solution to the engineering problem resulted from the application of
a two-phase liquid–liquid process (26) . The catalyst is synthesized with polar
–SO3Na groups on the phenyl rings of the triphenylphosphine
SO3Na
NaO3S SO3Na
P H
Rh P
SO3Na
NaO3S
P CO
NaO3S
SO3Na
SO3Na
SO3Na
Vol. 1 CATALYSIS 515
Olefin
Off-gas
CO + H2
M
2
4
3
Raw aldehyde
1
Water 5
Steam
Fig. 7. Process flow diagram for the two-phase hydroformylation of propylene where
1 ¼ reactor; 2 ¼ separator; 3 ¼ phase separator; 4 ¼ stripping column; and 5 ¼ heat exchan-
ger (26).
so that it is water soluble. In the reactor, the catalyst is largely confined to the
aqueous phase and the reactants are largely confined to the organic phase; reac-
tion likely takes place near the interface. A process flow diagram is shown in
Figure 7 (26). The two liquid phases are separated downstream of the stirred
reactor and the catalyst is recycled with a high efficiency. The process is applied
commercially and is competitive with the single-phase process. It is indicative of
a trend toward application of more homogeneous catalytic processes with envir-
onmentally friendly solvents such as water.
Attempts to anchor transition-metal complex catalysts to supports have
been extended to reactions far beyond just methanol carbonylation and hydro-
formylation, but almost all the resulting catalysts have failed to find practical
application. The principal difficulty is that the catalyst at each step of the cata-
lytic cycle must be stably bonded to the support, and this is often too stringent a
design criterion. Perhaps a more generally useful way to confine the catalyst in a
phase separate from that containing most of the reactants is to use a second
liquid phase. Water is one appealing choice, and promising work has also been
done with fluorous phases (27,28).
Olefin Metathesis. The olefin metathesis (dismutation) reaction (29,30)
converts olefins to lower and higher molecular weight olefins. For example, pro-
pylene is converted into ethylene and butene
¼CHR0 !RCH¼
2 RCH ¼ ¼CHR þ R0 CH¼
¼CHR0 ð14Þ
The olefins that undergo metathesis include most simple and substituted
olefins; cyclic olefins give linear high molecular weight polymers. The mechanism
of the reaction is believed to involve formation of carbene complexes that react
via cyclic intermediates, ie, metallacycles. An extension of the chemistry has
led to efficient ring-opening metathesis polymerization (32).
Industrial olefin metathesis processes are carried out with solid catalysts (29).
The Wacker Oxidation of Ethylene to Acetaldehyde. One of the early
industrial examples of organometallic catalysis in solution was described by
Smidt and co-workers in 1959 (33). This is the Wacker oxidation of ethylene to
give acetaldehyde. It had long been known that Pd(II) complexes in solution
would stoichiometrically oxidize ethylene to acetaldehyde. Smidt’s insight was
to make this reaction part of a catalytic cycle. The researchers created a process
by closing the cycle, using an excess of an oxidizing agent, Cu2þ, to rapidly reox-
idize the palladium that oxidized the ethylene and thereby keep palladium from
plating out on the reactor walls as metal. The cycle is completed as O2 reoxidizes
the Cuþ formed from Cu2þ. The sequence of reactions is the following:
CH2 ¼
¼CH2 þ H2 O þ PdCl2 ! CH3 CHO þ Pd þ 2 HCl ð15Þ
Pd þ 2 CuCl2 ! PdCl2 þ 2 CuCl ð16Þ
1
2 CuCl þ O2 þ 2 HCl ! 2 CuCl2 þ H2 O ð17Þ
2
These reactions constitute the cycle; their sum gives the stoichiometry of the
Wacker oxidation:
1
CH2 ¼
¼CH2 þ O2 ! CH3 CHO ð18Þ
2
O2
R• RH
C6H11OOH C6H12
H2O
Co⫹3
OH⫺ H⫹
Co⫹2
C6H11O• C6H11OO•
OH OOH
(R is C6H11), as shown in the upper right of the figure. One role of Co3þ is
suggested to be that of an electron-transfer agent undergoing a one-electron
change; the role of the metal is that of a redox initiator, as cobalt cycles between
the þ2 and þ3 oxidation states in a process that generates free radicals. As the
cobalt also reacts in a cycle, its role is catalytic, and only small amounts are
needed.
One of the most important oxidation processes is the conversion of p-xylene
into terephthalic acid or dimethyl terephthalate. There are numerous commer-
cial processes, generally using air as the oxidant and cobalt and manganese
salts as the catalysts (37). In one process, the oxidation takes place at 2258C
and 15 atm, and bromides are involved in the catalysis (37).
CO and H2 Reactions. Synthesis gas (CO and H2) can be converted
into numerous organic chemicals, including acetic anhydride formed in a com-
mercial process with a rhodium complex catalyst. Soluble rhodium catalysts
are also active for synthesis of ethylene glycol along with methanol and other
products (38). The process is not applied commercially because extremely
high pressures are required. But the process is conceptually important because
it is believed that the catalysts may be compounds with metal–metal
bonds, called metal clusters. A simple metal cluster is [Rh13(CO)24H3]2 ,
which has the following structure:
4. Heterogeneous Catalysis
The catalysts with the simplest compositions are pure metals, and the
metals that have the simplest and most uniform surface structures are single
crystals. Researchers have done many experiments with metal single crystals
in ultrahigh vacuum chambers so that unimpeded beams of particles and radia-
tion can be used to probe them. These surface science experiments have led to
fundamental understanding of the structures of simple adsorbed species, such
as CO, H, and small hydrocarbons, and the mechanisms of their reactions;
they indicate that catalytic activity is often sensitive to small changes in surface
structure. For example, paraffin hydrogenolysis reactions take place rapidly on
steps and kinks of platinum surfaces but only very slowly on flat planes; how-
ever, hydrogenation of olefins takes place at approximately the same rate on
each kind of surface site.
A few industrial catalysts have simple compositions, but the typical catalyst
is a complex composite made up of several components, illustrated schematically
in Figure 9 by a catalyst for ethylene oxidation. Often it consists largely of a por-
ous support or carrier, with the catalytically active components dispersed on the
support surface. For example, petroleum refining catalysts used for reforming of
naphtha have 1 wt% Pt and Re on the surface of a transition alumina such as
g-Al2O3 that has a surface area >100 square meters per gram. The expensive
metal is dispersed as minute particles or clusters so that a large fraction of the
atoms are exposed at the surface and accessible to reactants.
The typical industrial catalyst has both microscopic and macroscopic
regions with different compositions and structures; the surfaces of industrial cat-
alysts are much more complex than those of the single crystals of metal investi-
gated in ultrahigh vacuum experiments. Because surfaces of industrial catalysts
are very difficult to characterize precisely and catalytic properties are sensitive
Microparticle of a -Al2O3
Pore
Fig. 9. Schematic representation of a catalyst for ethylene oxide synthesis (not to scale).
The porous support particle consists of microparticles held together by a binder.
522 CATALYSIS Vol. 1
to small structural details, it is usually not possible to identify the specific com-
binations of atoms on a surface, called catalytic sites or active sites, that are
responsible for catalysis. Experiments with catalyst poisons, substances that
bond strongly with catalyst surfaces and deactivate them, have shown that the
catalytic sites are usually a small fraction of the catalyst surface. Although most
models of catalytic sites rest on rather weak foundations, some are well estab-
lished. For example, NO decomposition on ruthenium takes place at steps, as
shown by scanning tunneling microscopy images of individual reacting atoms
on the surface (42).
Important physical properties of catalysts include the particle size and
shape, surface area, pore volume, pore size distribution, and strength to resist
crushing and abrasion. Measurements of catalyst physical properties (43a)
are routine and often automated. Pores with diameters <2.0 nm are called
micropores; those with diameters between 2.0 and 5.0 nm are called meso-
pores; and those with diameters >5.0 nm are called macropores. Pore volumes
and pore size distributions are measured by mercury penetration and by N2
adsorption. Mercury is forced into the pores under pressure; entry into a
pore is opposed by surface tension. For example, a pressure of 71 MPa
(700 atm) is required to fill a pore with a diameter of 10 nm. The amount of
uptake as a function of pressure determines the pore size distribution of the
larger pores (44). In complementary experiments, the sizes of the smallest
pores (those 1 – 20 nm in diameter) are determined by measurements char-
acterizing desorption of N2 from the catalyst. The basis for the measure-
ment is the capillary condensation that occurs in small pores at pressures
less than the vapor pressure of the adsorbed nitrogen. The smaller the dia-
meter of the pore, the greater the lowering of the vapor pressure of the liquid
in it.
Surface areas are determined routinely and exactly from measurements of
the amount of physically adsorbed (physisorbed) nitrogen. Physical adsorption is
a process akin to condensation; the adsorbed molecules interact weakly with the
surface and multilayers form. The standard interpretation of nitrogen adsorption
data is based on the BET model (45), which accounts for multilayer adsorption.
From a measured adsorption isotherm and the known area of an adsorbed N2
molecule, taken to be 0.162 nm2, the surface area of the solid is calculated (see
ADSORPTION).
4.3. Influence of Mass Transport on Catalyst Performance. Reac-
tants must diffuse through the network of pores of a catalyst particle to reach
the internal area, and the products must diffuse back. The optimum porosity
of a catalyst particle is determined by tradeoffs: making the pores smaller
increases the surface area and thereby increases the activity of the catalyst,
but this gain is offset by the increased resistance to transport in the smaller
pores; increasing the pore volume to create larger pores for faster transport is
compensated by a loss of physical strength. A simple quantitative development
(46–48) follows for a first-order, isothermal, irreversible catalytic reaction in a
spherical, porous catalyst particle.
If there is a significant resistance to transport of the reactant in the
pores, a concentration gradient will exist at steady state, whereby the concen-
tration of the reactant is a maximum at the particle periphery and a minimum
Vol. 1 CATALYSIS 523
at the particle center. The product concentration will be higher at the particle
center than at the periphery. The concentration gradients provide the driving
force for the transport.
As a reactant molecule from the fluid phase surrounding the particle
enters the pore structure, it can either react on the surface or continue diffus-
ing toward the center of the particle. A quantitative model of the process is
developed by writing a differential equation for the conservation of mass of
the reactant diffusing into the particle. At steady state, the rate of diffusion
of the reactant into a shell of infinitesimal thickness minus the rate of diffusion
out of the shell is equal to the rate of consumption of the reactant in the shell
by chemical reaction. Solving the equation leads to a result that shows how the
rate of the catalytic reaction is influenced by the interplay of the transport,
which is characterized by the effective diffusion coefficient of the reactant in
the pores, Deff, and the reaction, which is characterized by the first-order reac-
tion rate constant.
The result is shown in Figure 10, which is a plot of the dimensionless effec-
tiveness factor as a function of the dimensionless Thiele modulus f, which is R(k/
Deff)1/2, where R is the radius of the catalyst particle and k is the reaction rate
constant. The effectiveness factor is defined as the ratio of the rate of the reaction
divided by the rate that would be observed in the absence of a mass transport
influence. The effectiveness factor would be unity if the catalyst were nonporous.
Therefore, the reaction rate is
r ¼ Zk½As ð19Þ
1.0
First order
0.8
0.6
Zero order
Effectiveness factor η
0.4
Second order
0.2
0.1
1 2 4 6 8 10 20 40
Thiele modulus φ = R k
D
Fig. 10. The Thiele plot accounting for the influence of intraparticle mass transport on
rates of catalytic reaction. The dimensionless terms Z and f are the effectiveness factor
and the Thiele modulus, respectively, and are defined in the text.
524 CATALYSIS Vol. 1
where Z is the effectiveness factor and [A]s the concentration of the reactant A at
the peripheral surface of the catalyst particle. The observed reaction rate con-
stant is not the intrinsic reaction rate constant k but the product Zk.
Figure 10 shows that Z is a unique function of the Thiele modulus. When
the modulus f is small (91), the effectiveness factor is unity, which means that
there is no effect of mass transport on the rate of the catalytic reaction. When f is
greater than 1, the effectiveness factor is less than unity and the reaction rate
is influenced by mass transport in the pores. When the modulus is large (010),
the effectiveness factor is inversely proportional to the modulus, and the reaction
rate (eq. 19) is proportional to k/f, which, from the definition of f, implies that
the rate and the observed reaction rate constant are proportional to (1/R)(Deffk)1/
2
[A]s. This result shows that both the rate constant, ie, a measure of the intrinsic
activity of the catalyst, and the effective diffusion coefficient, ie, a measure of the
resistance to transport of the reactant offered by the pore structure, influence the
rate. It is not appropriate to say that the reaction is diffusion controlled; it
depends on both the diffusion and the chemical kinetics. In contrast, as shown
by equation 3, a reaction in solution can be diffusion controlled, depending on
D but not on k.
In the case of a significant intraparticle diffusion influence, the tempera-
ture dependence of the observed rate constant, the apparent activation energy,
is not that of the intrinsic rate constant, ie, the true activation energy; rather, as
the Thiele modulus becomes large (>10), the temperature dependence
approaches that of (Deffk)1/2. The temperature dependence of a diffusion coeffi-
cient is usually small, and the measured activation energy is thus about half
the intrinsic activation energy. This is an important and general result: The
kinetics is disguised by the transport influence. When the effectiveness factor
is unity, there is no disguise, and the true activation energy is measured.
This development has been generalized. Results for zero- and second-
order irreversible reactions are shown in Figure 10. Results are given else-
where (48) for more complex kinetics, nonisothermal reactions, and particle
shapes other than spheres. For nonspherical particles, the equivalent spherical
radius, three times the particle volume/surface area, can be used for R to a
good approximation.
Even when there is a transport disguise, the reaction order remains one
for a first-order reaction. But for reactions that are not intrinsically first order,
the transport disguise changes the observed reaction order; for an intrinsically
zero-order reaction, the observed order becomes 1/2 and for an intrinsically
second-order reaction it becomes 3/2 when f010. For all reaction orders
the apparent activation energy is approximately one-half of the intrinsic
value in this limit.
The mass transport influence is easy to diagnose experimentally. One mea-
sures the rate at various values of the Thiele modulus; the modulus is easily
changed by variation of R, the particle size. Crushing and sieving the particles
provide catalyst samples for the experiments. If the rate is independent of the
particle size, the effectiveness factor is unity for all of them. If the rate is inver-
sely proportional to particle size, the effectiveness factor is less than unity and
f010. If the dependence is between these limits, then several experimental
points allow triangulation on the curve of Figure 10 and estimation of Z and
Vol. 1 CATALYSIS 525
Fig. 11. Structural model of the (111) face of g-alumina (53). The small solid circles
represent Al3þ, the large open circles OH groups, and the hatched circles oxygen. The
surface is 50% dehydroxylated.
clay mineral such as kaolinite. The clay is added to the mixture of microparticles
as they are formed into the desired particle shape, eg, by extrusion. Then the
support is heated to remove water and possibly burnout material and then sub-
jected to a high temperature, possibly 15008C, to cause vitrification of the clay;
this is a conversion of the clay into a glasslike form that spreads over the micro-
particles of the support and binds them together.
Promoters. Many industrial catalysts contain promoters, commonly che-
mical promoters. A chemical promoter is used in a small amount and influences
the surface chemistry. Alkali metal ions are often used as chemical promoters,
eg, in ammonia synthesis catalysts, ethylene oxide catalysts, and Fischer-
Tropsch catalysts (55). They may be used in as little as parts per million quan-
tities. The mechanisms of their action are often not well understood. In contrast,
seldom-used textural promoters, also called structural promoters, are used in
massive amounts in some catalysts and affect their physical properties. These
are used in ammonia synthesis catalysts.
4.5. Catalyst Treatments. Catalysts often require activation or regen-
eration and their disposal also requires special consideration (56).
Activation. Some catalysts, eg, Ziegler olefin polymerization catalysts
(57), are highly reactive in the presence of air, and some, eg, iron catalysts,
are even pyrophoric; these must be handled under a blanket of inert gas. The
surfaces of most catalysts are reactive and not easily maintained in an active
state in the form in which they are conveniently supplied. Thus, many catalysts
must be activated prior to use. The activation may be as simple as exposure to
reactants under processing conditions, but some catalysts require specialized
treatments. For example, catalysts used for hydroprocessing of fossil fuels are
usually delivered as supported metal oxides, but in the operating state they
are supported metal sulfides. Catalyst suppliers specify detailed procedures
for treatment, eg, with a mixture of hydrogen and oil containing organosulfur
528 CATALYSIS Vol. 1
compounds, to carry out the sulfiding properly. The details are often critical, and
the catalyst manufacturer’s guarantees may be void if the procedure is not
carried out properly.
Deactivation. Catalysts lose activity and selectivity in many ways, and
much effort in process development goes into measuring the deactivation and
finding means to minimize it. Some catalysts undergo physical changes during
normal operation; eg, a catalytically active phase may be transformed into an
inactive phase. Catalysts also undergo sintering, which is a coalescence of parti-
cles to give larger particles, accompanied by loss of surface area. Iron sinters
under conditions of ammonia synthesis, but when a textural promoter such as
alumina is present, the sintering is greatly reduced. Catalyst components may
be volatile and gradually vaporized during use; examples are silica used as a sup-
port and molybdenum oxides present in selective oxidation catalysts.
Catalysts commonly lose activity in operation as a result of accumulation of
materials from the reactant stream. Catalyst poisoning is a chemical phenom-
enon. A catalyst poison is a component such as a feed impurity that as a result
of chemisorption, even in small amounts, causes the catalyst to lose a substantial
fraction of its activity. For example, sulfur compounds in trace amounts poison
metal catalysts. Arsenic and phosphorus compounds are also poisons for a num-
ber of catalysts. Sometimes the catalyst surface has such a strong affinity for a
poison that it scavenges it with a high efficiency. The poison may then adsorb
strongly on the catalyst at the upstream end of a fixed-bed reactor at the begin-
ning of operation, with a wave then moving downstream through the reactor as
the upstream surface becomes saturated.
A selective poison is one that binds to the catalyst surface in such a way
that it blocks the catalytic sites for one kind of reaction but not those for another.
Selective poisons are used to control the selectivity of a catalyst. For example,
nickel catalysts supported on alumina are used for selective removal of acetylene
impurities in olefin streams (58). The catalyst is treated with a continuous feed
stream containing sulfur to poison it to an exactly controlled degree that does not
affect the activity for conversion of acetylene to ethylene but does poison the
activity for ethylene hydrogenation to ethane. Thus the acetylene is removed
and the valuable olefin is barely converted.
Because catalyst surfaces are reactive and often sensitive to their environ-
ments, they may be irreversibly changed by exposure to undesired reactants.
Upsets in plant operations can lead to catastrophic losses of whole catalyst
charges. A large catalyst charge that is ruined can cost hundreds of thousands
of dollars as well as the cost of lost operation.
Catalysts are also deactivated or fouled by physical deposition of materials
present in or formed from feeds. Sometimes massive deposits form on surfaces
and block access of reactants to the catalytic sites. Coke is carbonaceous material
of various compositions, often aromatic and with a high molecular weight and a
typical composition of approximately CH. Coke forms on every hydrocarbon pro-
cessing catalyst and on most catalysts used for organic chemical conversions.
Inorganic materials are also deposited on catalysts. For example, the organova-
nadium and organonickel compounds in petroleum residua react to form vana-
dium and nickel sulfides on the surfaces of hydroprocessing catalysts. The
solid deposits reduce activity by covering catalytic sites and by filling pores
Vol. 1 CATALYSIS 529
and restricting the entry of reactants. When the effectiveness factor for the
deposition reaction is small, the pore mouths can become blocked and catalysts
can suffer near catastrophic failure. Small particles of solid such as dust can also
foul catalysts. This may be a problem in processes for cleanup of NOx emissions
from coal-fired power plants, and catalysts are designed in the form of monoliths
(honeycombs) to minimize the effect.
Regeneration. Deactivated catalysts are treated to bring back the cataly-
tic activity in processes called regenerations. Coke deposits are removed by
controlled combustion. Often low partial pressures of oxygen are used to keep
the rate of combustion and the temperature rise from becoming too large and
leading to damage of the catalyst, such as by sintering. Periodic coke burn-off
can be carried out many times with little damage to many catalysts. Most cata-
lysts last for months or years between regenerations, but catalysts used for
cracking of petroleum are in contact with reactants for only a few seconds;
then they are separated and cycled to another reactor where they are regener-
ated (and then they are cycled back to the cracking reactor) (see CATALYSTS,
REGENERATION).
Redispersion. Expensive catalyst components such as precious metals
are used in high dispersions on supports. During operation, the small metal par-
ticles tend to sinter, ie, migrate and agglomerate, into larger particles with a loss
of metal surface area and thereby a loss of catalytic activity. The metals in such
catalysts may be redispersed as part of the catalyst regeneration (59). For exam-
ple, after the coke is burned off a supported platinum catalyst, the catalyst may
be treated with a reactive atmosphere containing chlorine and oxygen to form
volatile platinum oxychloride species that are transported through the gas
phase and deposited on the pore walls, where they are then treated in H2 and
reduced. The result is an increased dispersion of the platinum and a reactivated
catalyst.
Reclamation, Disposal, and Toxicity. Removal of poisons and inorganic
deposits from used catalysts is typically difficult and usually uneconomical.
Thus, some catalysts are used without regeneration, although they may be pro-
cessed to reclaim expensive metal components. Used precious metal catalysts,
including automobile exhaust conversion catalysts, are treated (often by the sup-
pliers) to extract the metals, and recovery efficiencies are high. Some spent
hydroprocessing catalysts may be used as sources of molybdenum and other
valuable metals.
Some catalysts are hazardous materials, or they react to form hazardous
materials. For example, catalysts used for hydrogenation of carbon monoxide
form volatile metal carbonyl compounds such as nickel carbonyl, which are
highly toxic. Many catalysts contain heavy metals and other hazardous compo-
nents, and environmentally safe disposal has become an increasing concern and
expense.
4.6. Catalyst Preparation. Catalyst preparation is more an art than a
science (60). Many reported catalyst preparations omit important details and
are difficult to reproduce exactly, and this has hindered the development of cat-
alysis as a quantitative science. However, the art is developing into a science and
there are now many examples of catalysts synthesized in various laboratories
that have nearly the same physical and catalytic properties.
530 CATALYSIS Vol. 1
Supports are often prepared first and the catalyst and promoter components
added later. Metal oxide supports are usually prepared by precipitation from aqu-
eous solutions. Nitrates are commonly used anions; alkalies and ammonium are
commonly used cations. Metal oxide supports, eg, silica and alumina, are prepared
in the form of hydrogels. Mixed oxides such as silica–alumina are made by cogela-
tion. Careful control of conditions such as pH is important to give uniform products.
Sol–gel processes are useful for the preparation of some high-area oxides (61).
Supports are washed at controlled values of pH to remove impurities. Ions
and impurities in the preparative solution are easily occluded in the solid and
difficult to remove by washing. Therefore, ions that might poison the catalyst,
eg, Cl , SO42 , or alkali metal ions, are avoided. Many of the materials are cation
exchangers, and washing does not remove cations from them. Metal ions can be
removed by exchange with ammonium ions, which on heating give off ammonia
and leave hydrogen ions in surface OH groups. Drying of precipitates and
hydrated gels leads to evolution of gases and may generate microporosity, as
described above for transition aluminas. Porosity can also be created by reduc-
tion of a nonporous oxide; porous iron can be made this way.
Catalyst components are usually added in the form of precursor metal salts
in aqueous solutions. In impregnation, the support may be dried, evacuated, and
brought in contact with an excess of an impregnating solution containing metal
salts. The processes are complex, possibly involving some dissolution of the sup-
port and reprecipitation of structures including mixed metal species. The solid is
then dried and calcined, ie, brought to a high temperature, usually in air. Alter-
natively, in the incipient wetness method, just enough of the impregnating
solution is used to fill the pores of the support. The chemistry of the interactions
of catalyst precursors with metal oxide supports is beginning to be understood.
Important parameters that control the adsorption of metal complex precursors
from aqueous solution are the isoelectric point of the metal oxide, the pH of
the solution, and the nature of the metal complex (62). Depending on the condi-
tions of the contact, cationic or anionic species may be adsorbed. Sometimes these
are simple mononuclear (single-metal atom) species, but sometimes they are
complicated polynuclear ions. The support may dissolve to some degree in the
preparation solution and be redeposited in some form (perhaps with the catalyst
precursor) (63). The nature of the initially adsorbed species may significantly
affect the structure of the catalytic species in the resultant catalyst. After
impregnation, the catalyst may be activated, for example, by drying and calcin-
ing. Promoters may be added at various stages, eg, as a final step in the prepara-
tion or just prior to operation.
Supported metal catalysts are reduced, eg, by treatment in hydrogen at
temperatures in the range of 300–5008C. The reduction temperature may influ-
ence the stability of the metal dispersion.
4.7. Examples of Surface Catalysis. Molecular Catalysis on
Supports. The term molecular catalysis is commonly applied only to reactions
in uniform fluid phases, but it applies nearly as well to some reactions taking
place on supports. Straightforward examples are reactions catalyzed by polymers
functionalized with groups that closely resemble catalytic groups in solution.
Industrial examples include reactions catalyzed by ion-exchange resins, usually
sulfonated poly(styrene-divinylbenzene) (64). This polymer is an industrial
Vol. 1 CATALYSIS 531
catalyst for synthesis of methyl tert-butyl ether (MTBE) from methanol and iso-
butylene and synthesis of bisphenol A from phenol and acetone, among others.
The former application grew rapidly as MTBE became a component of high-
octane-number gasoline (see ETHERS).
The polymer has the following structure, which is shown schematically.
CH CH2 CH CH2
SO3H
SO3H CH CH2
SO3H
Cl
Si
O O
This is an ion-exchanger like the sulfonated polymer. The silica surface can
also be functionalized with phosphine complexes; when combined with rhodium,
these give anchored complexes that behave like their soluble and polymer-
supported analogues as catalysts for olefin hydrogenation and other reactions:
CO
P Rh P
Cl
CH2 CH2
532 CATALYSIS Vol. 1
R Cz
CH2
M
P θP
The steric constraints imposed by the bulky ligands cause the propylene to bond
almost entirely with a single orientation with respect to the growing polymer
chain, CH2 P , which leads to the stereoregular product.
The explanation for the stereospecificity of the surface catalysis, which
preceded that for the solution catalysis, is based on the structure inferred for
Vol. 1 CATALYSIS 533
the a-TiCl3 crystal edges; the locations of the Cl ions at the anion vacancies cre-
ate an unsymmetrical environment whereby the growing polymer chain and the
adsorbed propylene are oriented predominantly in a single, energetically favored
way that leads to the stereoregular polymer as a result of a series of insertion
reactions. The explanation is simplified, but again there is a strong conceptual
link between molecular and surface catalysis. It is difficult to forge many such
links because of the lack of detailed understanding of most surface-catalyzed
reactions, which is a consequence of the complexity of the surface compositions
and structures of solid catalysts.
4.8. Catalysis by Metals. Metals are among the most important and
widely used industrial catalysts (72). They offer activities for a wide variety of
reactions (Table 1). Atoms at the surfaces of bulk metals have reactivities and
catalytic properties different from those of metals in metal complexes because
they have different ligand surroundings. The surrounding bulk stabilizes surface
metal atoms in a coordinatively unsaturated state that allows bonding of reac-
tants. Thus metal surfaces offer an advantage over metal complexes, in which
there is only restricted stabilization of coordinative unsaturation. Furthermore,
metal surfaces provide catalytically active sites that are stable at high
temperatures. For example, supported palladium catalysts have replaced soluble
palladium for vinyl acetate synthesis; the advantages of the solid include reduced
corrosion and reduced formation of by-products.
CO Oxidation Catalyzed by Palladium. One of the best understood cata-
lytic reactions occurring on a metal surface is the oxidation of carbon monoxide
on palladium:
2 CO þ O2 ! 2 CO2 ð20Þ
CO + 12_ O2
∆ H = 283
Energy, kJ/mol
260 Transition
state
Ea = 105
Reaction coordinate
Fig. 12. Schematic potential energy diagram illustrating the changes associated with
the individual reaction steps in CO oxidation on Pd (73). Eact ¼ 105 kJ/mol; DH ¼
283 kJ/mol. To convert kJ to kcal, divide by 4.184.
This depiction is vague because the exact nature of the sites S and their bonding
with reactants are not known. The experimental results have led to an approxi-
mate potential energy diagram characterizing these elementary steps on the
surface (Fig. 12) (73). This shows the role of the surface in providing an efficient
pathway for the reaction. Most of the energy is liberated as the reactants are
adsorbed; the activation energy for reaction of the adsorbed CO with the
adsorbed O is relatively small, and this step is only slightly exothermic.
Ammonia Synthesis. Another well-understood reaction is the ammonia
synthesis:
N2 þ 3 H2 ! 2 NH3 ð25Þ
N2 þ 2 S ! 2 Nads ð26Þ
H2 þ 2 S ! 2 Hads ð27Þ
2Nads þ Hads ! NHads þ S ð28Þ
NHads þ Hads ! NH2;ads þ S ð29Þ
NH2;ads þ Hads ! NH3 þ 2 S ð30Þ
Vol. 1 CATALYSIS 535
N + 3H
315
NH + 2 H
390
1130
NH2 + H
1400
960
N2Hads 450
540
~30 ~20 125
1
_
+ 3
_ H2
50 ∆ H = 45
N
2 2 2
260 42
1
_ 3
_
30 NH3, ads NH3, g
N
2 2, ads
+ 2
H2 120
NH2, ads + Hads
Nads + 3 Hads NHads + 2 Hads
Fig. 13. Schematic potential energy diagram for the catalytic synthesis and decomposi-
tion of ammonia on iron. The energies are given in kJ/mol; to convert kJ to kcal, divide by
4.184 (73).
4.9. Catalysis by Metal Oxides and Zeolites. Metal oxides are com-
mon catalyst supports and catalysts. Some metal oxides alone are industrial cat-
alysts; an example is the g-Al2O3 used for ethanol dehydration to give ethylene.
But these simple oxides are the exception; mixed-metal oxides are more common.
For example, silica–alumina was used earlier as a catalyst for cracking of petro-
leum and is still a component of such catalysts, and bismuth molybdates were
used for ammoxidation of propylene to give acrylonitrile. Metal oxides supported
on metal oxides are also commonly applied. For example, rhenium oxide, Re2O7,
supported on alumina is used for olefin metathesis, and complicated supported
oxides related to bismuth molybdates are used for ammoxidation.
Metal oxide surfaces are more complex in structure and composition than
metal surfaces, and they are not so easy to characterize with some ultrahigh
vacuum techniques, eg, electron spectroscopies, because they are poor electrical
conductors and build up electrical charge when subjected to streams of charged
particles. Consequently, understanding of catalysis on metal oxide surfaces is
less advanced than understanding of catalysis on metal surfaces, although sig-
nificant progress has been made (77,78).
Acid– base chemistry of metal oxide surfaces is important in catalysis and
is characterized by measurements such as infrared spectroscopy with adsor-
bed probe molecules, eg, the base pyridine. The surfaces have both basic and
acidic character (79). Both OH and O groups have base strengths ranging from
weak, eg, in silica gel, to moderate, eg, in g-Al2O3, to strong, eg, in highly dehy-
droxylated MgO. Surface OH groups are acids with proton donor strengths
ranging from weak, eg, in Al2O3, to strong, eg, in SiO2-Al2O3. Metal ions
exposed at surfaces are Lewis acids. Redox properties are also important in
some catalytic applications, as cations in some oxides, eg, V2O5, can change
oxidation state. Furthermore the principles of organometallic chemistry are
useful in describing the interactions of organic ligands with metal ions exposed
at surfaces.
Zeolites and Catalytic Cracking. Some of the best-understood catalysts
are zeolites, ie, crystalline aluminosilicates (80–82). They are well understood
because they have much more nearly uniform compositions and structures
than amorphous metal oxides such as silica and alumina. Here the usage of
amorphous refers to results of X-ray diffraction experiments; the crystallites of
a metal oxide such as g-Al2O3 that constitute the microparticles are usually so
small that sharp X-ray diffraction patterns are not measured; consequently the
solids are said to be X-ray amorphous or simply amorphous.
Zeolites contain Si, Al, and O ions and various other cations. The structures
are built up of linked SiO4 and AlO4 tetrahedra that share O ions. These tetra-
hedra are arranged in a number of ways to give the different zeolites. The struc-
tures are unique in that they incorporate pores as part of the regular crystalline
structures. The pores have dimensions of the order of molecular dimensions so
that some molecules fit into the pores and some do not. Hence, the zeolites are
molecular sieves, and they are applied in industrial separations processes to
take advantage of this property. Some zeolites and their pore dimensions are
listed in Table 2.
A catalytically important family of zeolites called faujasites (zeolites X
and Y) is represented in Figure 14. Here the points of intersection of the lines
Vol. 1 CATALYSIS 537
represent Si or Al ions; oxygen is present at the center of each line. This depiction
emphasizes the framework structure of the zeolite and shows the presence of the
intracrystalline pore structure in which there are spaces called supercages, each
with a diameter of 1.2 nm. The pore structure is three dimensional; the super-
cages are connected by apertures with diameters of 0.74 nm. Some rather large
molecules can fit through these apertures (Fig. 15) and undergo catalytic reac-
tion in the cages.
The zeolite frame is made up of SiO4 tetrahedra, which are neutral, and
AlO4 tetrahedra, which have a charge of 1. The charge of the AlO4 tetrahedra
is balanced by the charges of additional cations that exist at various crystallogra-
phically defined positions in the zeolite, many exposed at the internal surface.
Zeolites are thus ion exchangers. The cations may be catalytically active.
When the cations are Hþ, the zeolites are acidic. Acidic zeolite Y finds enormous
industrial application as a component of petroleum cracking catalysts (81). In
Fig. 14. Schematic representation of the structure of a faujasitic zeolite. I and II indicate
cation positions.
538 CATALYSIS Vol. 1
0.6
0.4 0.4 Linear paraffins
Zeolite A
0.7 Erionite
0.8
size, nm
0.9 2-Methyl-substituted
olefins
ZSM-5 Isoparaffins
1.0
o - Xylene
1.1
aromatics
0.6 0.6
Higher
1.2
Offretite
Cavity size, nm
1.3
Mordenite
1.4 0.7 0.7
Fig. 15. Pore dimensions of zeolites and critical dimensions of hydrocarbons (82).
the following simplified structure, the OH groups located near AlO4 tetrahedra
are moderately strong Brønsted acids and responsible for the catalytic activity
for many reactions:
H
H
O O O O O
O O
Si Al Si Si Al Si
O OO OO OO OO OO O
Zeolites are named to represent the exchangeable cations in them; eg, zeo-
lite NaY is zeolite Y with sodium ions in the cation exchange positions.
Another catalytically important zeolite is ZSM-5. There is a three-dimen-
sional network of pores in this zeolite, represented in Figure 16. A set of straight
parallel pores is intersected by a set of perpendicular zigzag pores. These pores
are smaller than those of the faujasites (Fig. 15). ZSM-5 is classified as a medium
pore zeolite, the faujasites are large pore zeolites, and zeolite A (Table 2) is a
small pore zeolite.
Both the faujasites and ZSM-5 in the acidic form catalyze many reactions
that are catalyzed by other soluble and solid acids. The zeolites are not very
strong acids at low temperatures, but at 5008C they are able to protonate paraf-
fins and initiate their cracking. Almost all the catalytic applications of zeolites
take advantage of their acidic properties. Activities of a family of HZSM-5
samples with different Si/Al ratios have been studied (84). When the Al
contents are low, the catalytic activity is proportional to the Al content of the
zeolite over a wide range of compositions. These results identify the proton-
donor sites associated with the Al cations as the catalytic sites for the cracking
Vol. 1 CATALYSIS 539
Chain
Pentasil
5-1 unit
Pair
a
b
(a) (b)
Fig. 16. Structure of the zeolite ZSM-5 (83): (a) framework of the zeolite; (b) schematic
representation of the pore structure.
þ
Rþ þ R0 H ! RH þ R0 ð31Þ
þ
R0 ! olefin þ Rþ ð32Þ
10 – 2 Molecular
Knudsen
10 – 4
10 – 6
D, cm2/s
10 – 8
Configurational
10–10
10–12
10–14
0.1 1 10 100 1000
Pore size, nm
are large in comparison with the dimensions of the diffusing molecules, then
molecular diffusion occurs, as it does in a fluid phase. When the pores become
smaller, the interactions of the molecules with the pore walls become dominant,
and Knudsen diffusion occurs. When the pores become so small that the mole-
cules barely fit through them, then configurational diffusion occurs. This diffu-
sion may be characterized by a substantial activation energy; the rate of
diffusion is a strong function of the pore and molecule sizes in this regime. In
the limit, a pore is too small for a molecule to fit in, and the diffusion coefficient
becomes zero.
Catalytic processes have been developed to take advantage of the unique
transport and molecular sieving properties of zeolites. The zeolite that has
found the most applications is the medium-pored HZSM-5 (81,91). The term
shape-selective catalysis is applied to describe the unique effects. There are
different kinds of shape selectivity. Mass transport shape selectivity is a conse-
quence of transport restrictions whereby some species diffuse more rapidly than
others in the zeolite pores. In the simplest kind of shape-selective catalysis, small
molecules in a mixture enter the pores and are catalytically converted whereas
large molecules pass through the reactor unconverted because they do not fit into
the pores where the catalytic sites are located. This statement is slightly over-
simplified as there are a few catalytic sites on the outer surfaces of zeolite crys-
tallites. Similarly, product molecules formed inside a zeolite may be so large that
their transport out of the zeolite may be very slow, and they may be converted
largely into other products that diffuse more rapidly into the product stream.
A different kind of shape selectivity is called restricted transition state selectivity
542 CATALYSIS Vol. 1
Hydrocarbon
O2–
H2O
e–
Partially
oxidized
hydrocarbon
Fig. 18. Schematic representation of the catalytic cycle for ammoxidation of propy-
lene and related reactions. M1 and M2 represent the two metals in a mixed-metal oxide
catalyst (10).
the Wacker oxidation, whereby the hydrocarbon reacts with an oxide, H2O, in a
step mediated by palladium, and a second metal, eg, copper, reacts with O2 and
then reoxidizes the palladium.
An important industrial partial oxidation process is the conversion of
propylene to acrylonitrile:
3
CH2 ¼
¼CHCH3 þ O2 þ NH3 ! CH2 ¼
¼CHCN þ 3 H2 O ð33Þ
2
The first catalysts used commercially to convert the propylene with high
selectivity were mixed oxides of bismuth and molybdenum, referred to as
bismuth molybdates. Improved catalysts consisting of a number of solid phases
have been developed, with each generation becoming more complicated
than its predecessor. Among the catalysts cited in a patent is the following:
Co2þ 2þ 3þ 3þ
6 Ni2 Fe3 Bi ðMoO4 Þ12 on 50 wt% SiO2 with some P and K (93). Silica is
the support. The other components provide new functions, perhaps making the
catalyst more stable. The catalysts are mechanically stable and used in fluidized-
bed reactors at low pressures 100 kPa (1 atm ), and temperatures of 400–
5008C. Years of development and improvement of catalysts for this process have
led to the recognition that lattice oxygen of certain multivalent transition-metal
oxides serves as a versatile, selective oxidizing agent, and these sites must be site
isolated on the catalyst surface (94).
544 CATALYSIS Vol. 1
O2 O
CH2 CH2 CH2 CH2
O2 O2
CO2, H2O
The selective oxidation is catalyzed by silver, which is the only good cat-
alyst. Other olefins are not converted selectively to the epoxides in the pre-
sence of silver. However, propylene epoxidation is applied commercially; the
catalysts are either molybdenum complexes in solution or solids containing
titanium.
The ethylene epoxidation catalysts (Fig. 9) are multicomponent mixtures
consisting of a support (a-Al2O3), the catalytically active component (silver par-
ticles), and chemical promoters (alkali metal ions such as Csþand perhaps anions
such as sulfate), and a binder; the older literature also describes textural promo-
ters. Furthermore, trace amounts of chlorine-containing compounds such as
ethylene dichloride are continuously added with the feed; these compounds,
like the alkali metal promoter, increase the selectivity of the catalyst for ethylene
oxide. The data of Figure 19 illustrate the role of an alkali metal promoter (97).
Extremely small amounts of the promoter markedly improve the selectivity of
the catalyst; selectivities as high as 80% are reported. Under conditions of the
catalytic oxidation, the silver surface is covered with a layer of oxygen and
may be more properly described as an oxide than as a metal. There is only an
incomplete understanding of the nature of the catalytic sites, the role of promo-
ters, and the bonding of the reactants to the catalyst surface.
The support needs to be inert, which explains the choice of a-Al2O3; most
metal oxides, including transition aluminas, catalyze unselective oxidation.
Vol. 1 CATALYSIS 545
80
78
76
Selectivity, % ethylene oxide
74
72
70
68
0 20 40 60 80 100 120 140 160
Potassium content, ppm
Fig. 19. Promotion of ethylene oxidation by potassium. The selectivity is the percentage
of ethylene converted to ethylene oxide (97).
The catalyst has a low surface area, 1 m2/g, and large pores to minimize the
influence of intraparticle diffusion, which would reduce the selectivity.
A new process for epoxidation of butadiene with a supported silver catalyst
has recently been announced (98).
Naphtha Reforming and Bifunctional Catalysis. In some supported
metal catalysts the support is not just an inert platform but plays an active
catalytic role. This point is illustrated by catalysts for reforming of naphtha
to make high octane number gasoline, a process that is a classic example of
bifunctional surface catalysis (99). The catalysts consist of metal, originally
platinum (100), but now largely rhenium – platinum or tin – platinum, on a
transition alumina, g-Al2O3 or Z-Al2O3. Platinum is chosen because it is
more active than any metal except iridium in a number of reactions that
increase the octane number of paraffins without substantially changing
their molecular weights. These reactions include dehydrogenation, eg, conver-
sion of methylcyclohexane to toluene, and dehydrocyclization, eg, conversion
of n-heptane to toluene. Skeletal isomerization is also desired, but platinum
has only a low activity for this reaction. Rather, the reaction is acid catalyzed.
Consequently, platinum supported on an amorphous solid acid is a good
catalyst.
The support must not be too strongly or too weakly acidic; chlorided
Al2O3 is optimal. The catalyst works well for reforming although the individual
functions, ie, the metal and the acid, alone do not. The metal alone does not
catalyze the branching reactions; they require an acidic function to gene-
rate carbenium ions, which undergo the desired isomerization. The acidic
function alone is not sufficient to generate the carbenium ions; if it were so
strongly acidic, the catalyst would be deactivated rapidly by carbonaceous
deposits. The metal in the catalyst catalyzes dehydrogenation of paraffins to
546 CATALYSIS Vol. 1
give olefins, which are easily protonated and thereby converted by carbenium
ion routes.
Some reforming processes operate with platinum on alumina catalysts at
5008C and high pressures, eg, 5 MPa (50 atm), with the reactant stream con-
taining predominantly hydrogen (33,99). It is unusual for a process in which
hydrogen is produced to be carried out in the presence of excess hydrogen, but
this is done to minimize catalyst deactivation. Coke causes deactivation, but
the high partial pressure of hydrogen retards coke formation as aromatic coke
precursors and coke are hydrogenated. Catalyst regeneration by controlled
coke burnoff may be needed only about twice a year.
However, there is a disadvantage to the high hydrogen partial pressures
that became especially apparent with the removal of tetraethyllead from gasoline
and an increased need for reforming processes as a source of high octane number
hydrocarbons. Aromatic compounds have high octane numbers, and the dehydro-
genation of naphthenes, ie, cyclic paraffins, to give aromatics is favored thermo-
dynamically at high temperatures and low hydrogen partial pressures.
Therefore, there was an incentive to find a way to operate reforming processes
economically at much lower pressures.
Low pressure operation became routine with the application of new cata-
lysts that are resistant to deactivation and withstand the low pressures. The cat-
alysts are bimetallic, eg, incorporating tin or rhenium as well as platinum
(101,102). The structures of these catalysts are still not well understood, but
they likely involve cationic forms of tin or rhenium, and the latter resists deac-
tivation better than the monometallic catalyst, especially in the presence of
sulfur-containing compounds.
There are now a number of applications of supported bimetallic catalysts. In
some instances, the role of the added metal is to isolate the atoms of the other
metal at the surface, thereby reducing the rates of reactions that require
ensembles of the atoms as the catalytic sites. These reactions are called struc-
ture-sensitive reactions (103). Some reactions, such as olefin hydrogenation,
are structure-insensitive reactions; they seem to be catalyzed by isolated metal
centers on surfaces, much as the Wilkinson hydrogenation is catalyzed by mono-
nuclear metal complexes. Iridium–platinum catalysts have been used for
naphtha reforming, and one of the roles of the platinum may be to combine
with the iridium in alloylike bimetallic clusters to suppress an undesired struc-
ture-sensitive reaction, viz, paraffin hydrogenolysis (103). This reaction takes
place in reforming and is largely undesired because it reduces the molecular
weight range.
4.11. Catalysis by Metal Sulfides. Metal sulfides such as MoS2, WS2,
and many others catalyze numerous reactions that are catalyzed by metals (104).
The metal sulfides are typically several orders of magnitude less active than the
metals, but they have the unique advantage of not being poisoned by sulfur com-
pounds. They are thus good catalysts for applications with sulfur-containing
feeds, including many fossil fuels.
Metal sulfide catalysts are widely applied in hydroprocessing of petroleum.
The reactions include hydrogenation, hydrodesulfurization, and hydrodenitro-
genation (105). Hydrodesulfurization is the reaction of organosulfur compounds
Vol. 1 CATALYSIS 547
Catalysts are discovered to meet processing needs and opportunities, but the dis-
covery of a catalytic application to take advantage of some newly discovered
material almost never occurs. Catalyst development is largely a matter of trial
and error testing. The methodology was defined by Mittasch in the development
of the ammonia synthesis process. Catalyst developers benefit from an extensive
548 CATALYSIS Vol. 1
and diverse literature and often can formulate good starting points in a search
for candidate catalysts by learning what has been used successfully for similar
reactions. Deeper insights, such as would arise from understanding of the
mechanistic details of a catalytic cycle, are usually not attained; the exceptions
to this rule largely pertain to molecular catalysis, usually reactions occurring in
solution. Fundamental insights were valuable in guiding the development of the
process for chiral hydrogenation and that for methanol carbonylation, among
others, but it would be inappropriate to infer that understanding of the funda-
mental chemistry led to straightforward design of the catalysts. Indeed, the
initial working hypothesis about the chiral hydrogenation turned out to be incor-
rect. The more complicated processes of surface catalysis are for the most part
only partially understood even when the processes are established and extensive
after-the-fact research has been done. Creative research in catalyst discovery
and development is usually the result of intuition and partial understanding
combined with efficient testing and serendipity. Researchers who are repeatedly
successful in finding new and improved catalysts seem to recognize needs and
opportunities and notice significant exceptions to expected patterns and reason
inductively by imperfect analogies.
Catalyst testing and evaluation have been revolutionized by computers,
automated and miniaturized test reactors, and analytical methods. Combinator-
ial methods are playing an increasing role in the preparation and testing with
modern equipment. Researchers can systematically prepare and screen many
catalysts in a short time and efficiently determine, not only the initial catalytic
activity and selectivity, but also the stability and the appearance of trace
products that may indicate some new catalytic properties worthy of further
development.
Catalyst design is in a primitive stage. There are hardly any examples of
true design of catalysts. However, development of improved catalysts has been
guided successfully in instances when the central issues were the interplay of
mass transport and reaction. An example is catalysts used for hydroprocessing
of heavy fossil fuels.
Almost all industrial catalysts are developed by researchers who are moti-
vated to improve processes or create new ones. Thus the organization that first
uses a new catalyst is usually the one that has discovered it. This organization,
however, only rarely becomes the manufacturer of the catalyst used on a large
scale. Catalysts are for the most part highly complex specialty chemicals, and
catalyst manufacturers tend to be more efficient than others in producing
them. Catalyst manufacturing is a competitive industry. Catalyst users often
develop close relations with catalyst manufacturers, and the two may work
together to develop and improve proprietary catalysts.
The literature consists of patents, books, journals, and trade literature. The
examples in patents may be especially valuable. The primary literature provides
much catalyst performance data, but there is a lack of quantitative results char-
acterizing the performance of industrial catalysts under industrially realistic
conditions. Characterizations of industrial catalysts are often restricted to phy-
sical characterizations and perhaps activity measurements with pure component
feeds, but it is rare to find data characterizing long-term catalyst performance
with impure, multicomponent industrial feedstocks. Catalyst regeneration proce-
dures are scarcely reported. Those who have proprietary technology are normally
reluctant to make it known. Readers should be critical in assessing published
work that claims a relevance to technology.
Often the catalysts described in the literature are not quite the same as
those used in industrial processes, and often the reported performance is for
pure single-component feeds. Sometimes the best quantitative approximations
that can be made from the available literature are those based on reported
kinetics of reactions with pure feeds and catalysts that are similar to but not
the same as those used in practice. As a first approximation, one may use the
published results and scale the activity on the basis of a few laboratory results
obtained with realistic feeds and commercially available catalysts.
Catalyst suppliers are valuable sources of information. They often have
extensive experience testing catalysts in long-term operation with real industrial
feedstocks and may be willing to share information to improve their chances of
selling a catalyst. Also they may work with a potential customer to develop
catalysts that they could then supply. There is an extensive literature produced
by catalyst manufacturers and organizations that license and market catalytic
technology. This trade literature should be read critically as its purpose is to gen-
erate sales, but it often contains valuable information. Catalyst manufacturers
and those who license and sell technology are motivated to demonstrate their
technical knowledge. Their success in marketing depends on their technical
reputations and their reliability in supplying catalysts and technology that con-
sistently meet specifications.
BIBLIOGRAPHY
‘‘Catalysis’’ in ECT 1st ed., Vol. 3, pp. 245–272 and Suppl. 1, pp. 144–150, by V. I.
Komarewsky, Illinois Institute of Technology, and J. B. Coley, Standard Oil Co. of Indi-
ana; in ECT 2nd ed., Vol. 4, pp. 534–586, by G. A. Mills, Houdry Process and Chemical
Co.; in ECT 3rd ed., Vol. 5, pp. 16–61, by G. A. Mills, U.S. Department of Energy, and
J. A. Cusumano, Catalytica Associates, Inc.; in ECT 4th ed,. Vol. 5 pp. 320–373, by Bruce
C. Gates, University of Delaware; ‘‘Catalysis’’ in ECT (online), posting date: December 4,
2000, by Bruce C. Gates, University of Delaware.
CITED PUBLICATIONS
2. A. Mittasch and E. Thies, Von Davy und Do€bereiner bis Deacon, ein Halbes
Jahrhundert Grenzflächenkatalyse, Verlag Chemie, Berlin, Germany, 1932.
3. L. Wilhelmy, Ann. Physik. Chem. (Poggendorf) 81, 413, 419 (1850).
4. M. Bodenstein, Bericht des V. Internationalen Kongresses für angewandte Chemie zu
Berlin 1903, Sektion X, Band IV, 1904, p. 561.
5. A. Mittasch, Adv. Catal. 2, 81 (1950).
6. S. A. Topham, in J. R. Anderson and M. Boudart, eds., Catalysis–Science and
Technology, Vol. 5, Springer-Verlag, Berlin, Germany, 1984, p. 119.
7. B. Timm, Proceeding of the 8th International Congress on Catalysis (Berlin) 1, 7
(1984).
8. M. Boudart, Kinetics of Chemical Processes, Prentice-Hall, New York, 1968.
9. R. P. Bell, The Proton in Chemistry, Cornell University Press, Ithaca, New York,
1973.
10. B. C. Gates, Catalytic Chemistry, John Wiley & Sons, Inc., New York, 1992.
11. P. K. Ghosh, T. Guha, and A. N. Saha, J. Appl. Chem. 17, 239 (1967).
12. J. I. de Jong and F. H. D. Dethmers, Rec. Trav. Chim. 84, 460 (1965).
13. G. Olah, G. K. S. Prakash, and J. Sommer, Superacids, John Wiley & Sons, Inc.,
New York, 1985.
14. J. P. Collman, L. S. Hegedus, J. R. Norton, and R. G. Finke, Principles and Applica-
tions of Organotransition Metal Chemistry, 2nd ed., University Science Books, Mill
Valley, Calif., 1987.
15. B. R. James, Homogeneous Hydrogenation, John Wiley & Sons, Inc., New York,
1974.
16. J. Halpern, Trans. Am. Crystallogr. Assoc. 14, 59 (1978); J. Halpern, T. Okamoto,
and A. Zakhariev, J. Mol. Catal. 2, 65 (1976); J. Halpern and C. S. Wong,
J. Chem. Soc., Chem. Commun., 629 (1973).
17. W. S. Knowles, M. J. Sabacky, and B. D. Vineyard, Adv. Chem. Ser. 132, 274 (1974).
18. R. Noyori, Asymmetric Catalysis in Organic Synthesis, John Wiley & Sons, Inc.,
New York, 1994.
19. J. Halpern, Science 217, 401 (1982).
20. D. Forster, J. Am. Chem. Soc. 98, 846 (1976).
21. P. M. Maitlis, A. Haynes, G. J. Sunley, and M. J. Howard, J. Chem. Soc. Dalton
Trans. 1996, 2187.
22. J. Hjorkjaer and V. W. Jensen, Ind. Eng. Chem. Prod. Res. Dev. 15, 46 (1976).
23. R. T. Eby and T. C. Singleton, in B. E. Leach, ed., Applied Industrial Catalysis,
Vol. 1, Academic Press, New York, 1983, p. 275.
24. N. Yoneda, S. Kusano, M. Yasui, P. Pujado, and S. Wilcher, Appl. Catal. A Gen. 221,
253 (2001).
25. H.-W. Bohnen and B. Cornils, Adv. Catal., in press.
26. H. Bach, W. Gick, W. Konkol, and E. Wiebus, Proceeding of the 9th International
Congress on Catalysis (Calgary) Vol. 1, 1988 p. 254.
27. I. T. Horvath, Acc. Chem. Res. 31, 641 (1998).
28. J. A. Gladaysz, Pure Appl. Chem. 73, 1319 (2001).
29. R. L. Banks, in Ref. 23, Vol. 3, p. 215.
30. H. S. Eleuterio, J. Mol. Catal. 65, 55 (1991).
31. (a) J. Kress, M. Wesolek, and J. A. Osborn, J. Chem. Soc., Chem. Commun., 514
(1982); (b) N. Calderon, E. A. Ofstead, J. P. Ward, W. A. Judy, and K. W. Scott,
J. Am. Chem. Soc. 90, 4133 (1968).
32. M. Scholl, S. Ding, C. W. Lee, and R. H. Grubbs, Org. Lett. 1, 953 (2000); C. W.
Bielawski and R. H. Grubbs, Angew. Chem. Int. Ed. Engl. 39, 2903 (2000).
33. B. C. Gates, J. R. Katzer, and G. C. A. Schuit, Chemistry of Catalytic Processes,
McGraw-Hill Book Co., Inc., New York, 1979.
Vol. 1 CATALYSIS 551
72. V. Ponec and G. C. Bond, Catalysis by Metals and Alloys, Elsevier, Amsterdam, The
Netherlands, 1995.
73. G. Ertl, Proceedings of the 7th International Congress on Catalysis, Part A, p. 21,
1981.
74. R. Jennings, ed., Ammonia Synthesis—Theory and Practice, Pergamon Press,
1991.
75. P. Stoltze and J. K. Nørskov, Phys. Rev. Lett. 55, 2502 (1985); P. Stoltze, Phys. Scr.
36, 824 (1987); B. Hammer and J. K. Nørskov, Adv. Catal. 45, 71 (2000).
76. J. S. Merriam and K. Atwood, in Ref. 22, Vol. 3, p. 113.
77. V. E. Henrich and P. A. Cox, The Surface Science of Metal Oxides, Cambridge
University Press, Cambridge, 1994.
78. H. Idriss and M. A. Barteau, Adv. Catal. 45, 261 (2000).
79. B. C. Gates, in Handbook of Heterogeneous Catalysis, John Wiley & Sons, Inc., in
press.
80. D. W. Breck, Zeolite Molecular Sieves, Wiley, New York, 1974.
81. N. Y. Chen, W. E. Garwood, and F. G. Dwyer, Shape Selective Catalysis in Industrial
Applications, Dekker, New York, 1996.
82. E. G. Derouane, in M. S. Whittington and A. J. Jacobson, eds., Intercalation
Chemistry, Academic Press, New York, 1982, p. 101.
83. E. M. Flanigen and co-workers, Nature (London) 271, 572 (1978).
84. W. O. Haag, R. M. Lago, and P. B. Weisz, Nature (London) 309, 590 (1984).
85. P. B. Venuto and E. T. Habib, Catal. Rev. Sci. Eng. 18, 1 (1978).
86. B. W. Wojciechowski and A. Corma, Catalytic Cracking: Catalysts, Chemistry, and
Kinetics, Dekker, New York, 1986.
87. I. E. Maxwell, CATTECH 1, 5 (1997).
88. M. E. Davis and R. F. Lobo, Chem. Mater. 4, 756 (1992).
89. B. Notari, Adv. Catal. 41, 253 (1996).
90. S. J. Miller, Stud. Surf. Sci. Catal. 84, 2319 (1994).
91. W. O. Haag and N. Y. Chen, in L. L. Hegedus, ed., Catalyst Design, Progress and
Perspectives, John Wiley & Sons, Inc., New York, 1987, p. 163.
92. P. B. Weisz, CHEMTECH 3., 498 (1973).
93. U.S. Pat. 3,414,631 (Dec. 3,1968), R. K. Grasselli, G. Heights, and J. L. Callahan
(to Standard Oil Co.); U.S. Pat. 3,642,930 (Feb. 15, 1972), R. K. Grasselli, G. Heights,
and H. F. Hardman (to Standard Oil Co.).
94. R. K. Grasselli, Top. Catal. 15, 93 (2001).
95. B. C. Gates, Adv. Chem. Eng. 27, 49 (2001).
96. J. M. Berty, in Ref. 23, Vol. 1, p. 207.
97. U.S. Pat. 4,356,312 (Oct. 26, 1982), R. P. Nielson and J. H. La Rochelle (to Shell Oil Co.).
98. J. R. Monnier, Appl. Catal. A Gen. 221, 73 (2001).
99. G. A. Mills, H. Heinemann, T. H. Milliken, and A. G. Oblad, Ind. Eng. Chem. 45, 134
(1953).
100. M. J. Sterba and V. Haensel, Ind. Eng. Chem. Prod. Res. Dev. 15, 2 (1976).
101. G. Antos, A. M. Aitani, and J. M. Parera, eds., Catalytic Naphtha Reforming: Science
and Technology, Dekker, New York, 1995.
102. U.S. Pat. 3,415,737 (Dec. 10, 1968), H. E. Klucksdahl (to Chevron Research Co.).
103. J. H. Sinfelt, Bimetallic Catalysts, Discoveries, Concepts, and Applications, John
Wiley & Sons, Inc., New York, 1983.
104. O. Weisser and S. Landa, Sulphide Catalysts, Their Properties and Applications,
Pergamon Press, Oxford, UK, 1973.
105. M. J. Girgis and B. C. Gates, Ind. Eng. Chem. Res. 30, 2021 (1991).
106. H. Topsøe, B. S. Clausen, and F. E. Massoth, Hydrotreating Catalysis, Science and
Technology, Springer, Berlin, 1996.
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 553
GENERAL REFERENCES
References 10, 14, 18, 23, 33, 37, 39, 40, 48–50, 56, 72, 77, and 80 are general references.
Also see B. Cornils, and W. A. Herrmann, eds., Applied Homogeneous Catalysis
with Organometallic Compounds, Wiley-VCH, Weinheim, Germany, 2002; G. Ertl,
H. Knözinger, and J. Weitkamp, eds., Handbook of Heterogeneous Catalysis, Wiley-
VCH, Weinheim, Germany, 1997. Advances in Catalysis, Academic Press, New York,
continuing series. Nomenclature used in catalysis is summarized in B. Cornils,
W. A. Herrmann, R. Schlögl, and C. H. Wong, eds., Catalysis from A to Z, Wiley-VCH,
Weinheim, Germany, 2000.
BRUCE C. GATES
University of California, Davis
CATALYST DEACTIVATION
AND REGENERATION
1. Introduction
Catalyst deactivation, the loss over time of catalytic activity and/or selectivity, is
a problem of great and continuing concern in the practice of industrial catalytic
processes. Costs to industry for catalyst replacement and process shutdown total
billions of dollars per year. Time scales for catalyst deactivation vary consider-
ably; for example, in the case of cracking catalysts, catalyst mortality may be
on the order of seconds, while in ammonia synthesis the iron catalyst may last
for 5–10 years. It is inevitable, however, that all catalysts will decay.
Typically, the loss of activity in a well-controlled process occurs slowly.
However, process upsets or poorly designed hardware can bring about cata-
strophic failure. For example, in steam reforming of methane or naphtha great
care must be taken to avoid reactor operation at excessively high temperatures
or at steam-to-hydrocarbon ratios below a critical value. Indeed, these conditions
can cause formation of large quantities of carbon filaments that plug catalyst
pores and voids, pulverize catalyst pellets, and bring about process shutdown,
all within a few hours.
While catalyst deactivation is inevitable for most processes, some of its imme-
diate, drastic consequences may be avoided, postponed, or even reversed. Thus, de-
activation issues (ie, extent, rate, and reactivation) greatly impact research,
development, design, and operation of commercial processes. Accordingly, there
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1415021218150209.a01.pub2
554 CATALYST DEACTIVATION AND REGENERATION Vol. 1
There are many paths for heterogeneous catalyst decay. For example, a catalyst
solid may be poisoned by any one of a dozen contaminants present in the feed. Its
surface, pores, and voids may be fouled by carbon or coke produced by cracking/
condensation reactions of hydrocarbon reactants, intermediates, and/or pro-
ducts. In the treatment of a power plant flue gas, the catalyst can be dusted or
eroded by and/or plugged with fly ash. Catalytic converters used to reduce emis-
sions from gasoline or diesel engines may be poisoned or fouled by fuel or lubri-
cant additives and/or engine corrosion products. If the catalytic reaction is
conducted at high temperatures, thermal degradation may occur in the form of
active phase crystallite growth, collapse of the carrier (support) pore structure,
and/or solid-state reactions of the active phase with the carrier or promoters. In
addition, the presence of oxygen or chlorine in the feed gas can lead to formation
of volatile oxides or chlorides of the active phase, followed by gas-phase transport
from the reactor. Similarly, changes in the oxidation state of the active catalytic
phase can be induced by the presence of reactive gases in the feed.
Thus, the mechanisms of solid catalyst deactivation are many; neverthe-
less, they can be grouped into six intrinsic mechanisms of catalyst decay: (1) poi-
soning, (2) fouling, (3) thermal degradation, (4) vapor compound formation and/
or leaching accompanied by transport from the catalyst surface or particle,
(5) vapor–solid and/or solid–solid reactions, and (6) attrition/crushing. As
mechanisms 1, 4, and 5 are chemical in nature while 2 and 5 are mechanical,
the causes of deactivation are basically threefold: chemical, mechanical, and
thermal. Each of the six basic mechanisms is defined briefly in Table 1. Mechan-
isms 4 and 5 are treated together, since 4 is a subset of 5.
2.1. Poisoning. Poisoning (3,16–22) is the strong chemisorption of reac-
tants, products, or impurities on sites otherwise available for catalysis. Thus,
poisoning has operational meaning; that is, whether a species acts as a poison
depends upon its adsorption strength relative to the other species competing
for catalytic sites. For example, oxygen can be a reactant in partial oxidation
of ethylene to ethylene oxide on a silver catalyst and a poison in hydrogenation
of ethylene on nickel. In addition to physically blocking adsorption sites,
adsorbed poisons may induce changes in the electronic or geometric structure
of the surface (17,21).
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 555
H H
S
H C C H H H
M M M M M M
M M M M M
Fig. 1. Conceptual model of poisoning by sulfur atoms of a metal surface during ethylene
hydrogenation.
556 CATALYST DEACTIVATION AND REGENERATION Vol. 1
four groups of catalyst poisons classified according to chemical origin and their
type of interaction with metals. It should be emphasized that interactions of
Group VA–VIIIA elements with catalytic metal phases depend on the oxidation
state of the former, ie, how many electron pairs are available for bonding and the
degree of shielding of the sulfur ion by ligands (16). Thus, the order of decreasing
toxicity for poisoning of a given metal by different sulfur species is H2S, SO2,
SO42 , ie, in the order of increased shielding by oxygen. Toxicity also increases
with increasing atomic or molecular size and electronegativity, but decreases if
the poison can be gasified by O2, H2O, or H2 present in the reactant stream (21);
for example, adsorbed carbon can be gasified by O2 to CO or CO2 or by H2 to CH4.
Table 3 lists a number of common poisons for selected catalysts in impor-
tant representative reactions. It is apparent that organic bases (eg, amines)
and ammonia are common poisons for acidic solids such as silica–aluminas
and zeolites in cracking and hydrocracking reactions, while sulfur- and
Ni
Fig. 2. Schematic view of sulfur adsorbed on a Ni(100) surface at a (a) S/Nis ¼ 0.25 in a
p(2
2) structure and (b) S/Nis ¼ 0.50 in a c(2
2) structure.
adsorbed sulfur, has been reported by a number of workers (27); for example,
McCarroll and co-workers (33,34) found that sulfur adsorbed at near saturation
coverage on a Ni(111) face was initially in a hexagonal pattern but upon heating
above 700 K reoriented to a distorted C(2
2) structure on a Ni(100) layer. In
another study (32), sulfur adsorbed on a Ni(810) caused decomposition to (100)
and (410) facets. On the basis of their review of the reconstruction studies,
Bartholomew and co-workers (27) concluded that at high temperatures and
near saturation coverages, restructuring by sulfur of different facets of Ni to
the more stable Ni(100) is probably a general phenomenon. If so, the S/Nis
ratio at saturation would in principle be 0.5 for the reconstructed surface. In
the first example above, restructuring would not affect the S/Nis ratio at satura-
tion, since it is 0.5 for both (100) and (111) planes; however, in the second exam-
ple, the S/Nis ratio at saturation would probably decrease, as rough planes
transform to smoother ones. Nevertheless, the possibility of increases in the S/
Nis ratio at saturation due to reconstruction cannot be ruled out.
The nature of reconstruction of a surface by a poison may depend on its pre-
treatment. For example, in a scanning tunneling microscopy (STM) study of
room temperature H2S adsorption on Ni(110), Ruan and co-workers (35) found
that the S/Ni structure at saturation varied with the initial state of the surface,
ie, whether clean or oxygen covered (see Fig. 3). This study showed that no recon-
struction occurs by direct exposure to H2S at room temperature, rather only in
Fig. 3. A series of in situ STM images recorded after exposure of Ni(110) to oxygen and
then progressively higher exposures of H2S: (a) (2
1)O overlayer; (b) white islands and
black troughs with a C(2
2)S structure after exposure to 3 and 8 L of H2S; (c) 25 L,
islands transform to low-coordinated rows in the [001] direction; and (d) 50 L, stable,
well-ordered (4
1)S (35).
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 561
the presence of O2 (or air). This emphasizes the complexities inherent in predict-
ing the structure and stability of a given poison adsorbed on a given catalyst dur-
ing a specified reaction as a function of different pretreatments or process
disruptions, eg, exposure to air.
It is evident that structure and stoichiometry of sulfur adsorbed on nickel
are complex functions of temperature, H2S concentration, sulfur coverage, and
pretreatment, phenomena that account at least in part for the complex nature
of nickel poisoning by sulfur (27,36). Could one expect similar complexities in
the poisoning of other metals? Probably, since poisoning of nickel is prototypical,
ie, similar principles operate and similar poisoning behaviors are observed in
other poison/metal systems, although none have been studied to the same
depth as sulfur/nickel.
Since one of the necessary steps in a catalytic reaction is the adsorption of
one or more reactants, investigation of the effects of adsorbed sulfur on the
adsorption of other molecules can provide useful insights into the poisoning pro-
cess (21,27). Previous investigations (27,37–43) indicate that both H2 and CO
adsorptions on nickel are poisoned by adsorbed sulfur. Sulfur poisoning can
affect reaction selectivity as well as activity (27).
Because sulfur adsorbs so strongly on metals and prevents or modifies the
further adsorption of reactant molecules, its presence on a catalyst surface
usually effects substantial or complete loss of activity in many important reac-
tions. The steady-state methanation activities of Ni, Co, Fe, and Ru are relative
to the fresh, unpoisoned surface activity as a function of gas phase H2S concen-
tration. Data indicate that Ni, Co, Fe, and Ru all suffer 3–4 orders of magnitude
loss in activity at 15–100 ppb of H2S, ie, their sulfur tolerances are extremely
low. Moreover, the sharp drop in activity with increasing H2S concentration sug-
gests highly selective poisoning. Nevertheless, the rate of sulfur poisoning and
hence sulfur resistance varies from catalyst to catalyst and is apparently a func-
tion of catalyst composition (27) and reaction conditions (44). Indeed, it is possi-
ble to significantly improve sulfur resistance of Ni, Co, and Fe with catalyst
additives such as Mo and B that selectively adsorb sulfur. Because the adsorp-
tion of sulfur compounds is generally rapid and irreversible, surface sulfur con-
centrations in catalyst particles and beds are nonuniform, eg, H2S adsorbs
selectively at the entrance to a packed bed and on the outer surface of catalyst
particles, making the experimental study and modeling of sulfur poisoning extre-
mely difficult.
There are other complications in the study of sulfur poisoning. For example,
the adsorption stoichiometry of sulfur in CO hydrogenation on Ni is apparently a
function of the temperature, H2/CO ratio, and water partial pressure (44). More-
over, at high CO partial pressures sulfur may be removed from the surface as
COS, which is not as strongly adsorbed as H2S. At low temperature conditions,
eg, those representative of Fischer–Tropsch synthesis or liquid phase hydro-
genations, the gas phase concentration of H2S in poisoning studies must be kept
very low, ie, below 0.1–5 ppm, to avoid formation of bulk metal sulfides —
a phenomenon that seriously compromises the validity of the results. Thus, the
importance of studying poisoning phenomena in situ under realistic reaction
conditions, at low process-relevant poison concentrations, and over a process-
representative range of temperature and concentration conditions is emphasized.
562 CATALYST DEACTIVATION AND REGENERATION Vol. 1
C in Ni Cv
Ni3C (s)
CH4 (g)
4 H (a)
Fig. 5. Formation and transformation of coke on metal surfaces (a, g, s refer to adsorbed,
gaseous, and solid states respectively); gas phase reactions are not considered (48).
564 CATALYST DEACTIVATION AND REGENERATION Vol. 1
type, metal crystallite size, promoter, and catalyst support. For example, sup-
ported Co, Fe, and Ni are active above 350–4008C for filamentous carbon forma-
tion from CO and hydrocarbons; the order of decreasing activity is reportedly Fe,
Co, Ni (48). Pt, Ru, and Rh catalysts, on the other hand, while equally or more
active than Ni, Co, or Fe in steam reforming produce little or no coke or carbon.
This is attributed to reduced mobility and/or solubility of carbon in the noble
metals, thus retarding the nucleation process. Thus, it is not surprising that
addition of noble metals to base metals retards carbon formation; for example,
addition of Pt in Ni lowers carbon deposition rate during methanation,
while addition of Cu or Au to Ni substantially lowers carbon formation in
steam reforming (48,56). In contrast to the moderating effects of noble metal
additives, addition of 0.5% Sn to cobalt substantially increases the rate of carbon
filament formation from ethylene (57), an effect desirable in the commercial
production of carbon filament fibers.
Since carbon formation and gasification rates are influenced differently by
modifications in metal crystallite surface chemistry, which are in turn a function
of catalyst structure, oxide additives or oxide supports may be used to moderate
the rate of undesirable carbon or coke accumulation. For example, Bartholomew
and Strasburg (58) found the specific rate (turnover frequency) of filamentous
carbon deposition on nickel during methanation at 3508C to decrease in the
order Ni/TiO2, NiAl2O3, Ni/SiO2, while Vance and Bartholomew (59) observed
Ca hydrogenation rates at 1708C to decrease in this same order (the same as
for methanation at 2258C). This behavior was explained in terms of promotional
or inhibiting effects due to decoration of metal crystallites by the support, for
example silica, inhibiting both CO dissociation and carbon hydrogenation. This
hypothesis is consistent with observations (60,61) that silica evaporated on metal
surfaces and supported metals inhibits formation of filamentous carbon. Simi-
larly Bitter and co-workers (62) observed rates of carbon formation in CO2/CH4
reforming to decrease in the order Pt/g-Al2O3 ! Pt/TiO2 > Pt/ZrO2; while 90% of
the carbon deposited on the support, the authors linked deactivation to carbon
accumulated on the metal owing to an imbalance between carbon formed by
methane dissociation and oxidation by chemisorbed CO2. The rate of formation
of coke in steam reforming is delayed and occurs at lower rates in nickel catalysts
promoted with alkali or supported on basic MgO (63).
Since formation of coke, graphite, or filamentous carbon involves the forma-
tion of C C bonds on multiple atoms sites, one might expect that coke or carbon
formation on metals is structure-sensitive, ie, sensitive to surface structure and
metal crystallite size. Indeed, Bitter and co-workers (62) found that catalysts
containing larger Pt crystallites deactivate more rapidly than those containing
small crystallites. Moreover, a crystallite size effect, observed in steam reforming
of methane on nickel (48,63), appears to operate in the same direction, ie, forma-
tion of filamentous carbon occurs at lower rates in catalysts containing smaller
metal crystallites.
In summary, deactivation of supported metals by carbon or coke may occur
chemically owing to chemisorption or carbide formation or physically and
mechanically owing to blocking of surface sites, metal crystallite encapsulation,
plugging of pores, and destruction of catalyst pellets by carbon filaments. Block-
ing of catalytic sites by chemisorbed hydrocarbons, surface carbides, or relatively
566 CATALYST DEACTIVATION AND REGENERATION Vol. 1
aspects of the chemistry are quite different. The principal chemical loss of activ-
ity in oxides and sulfides is due to the strong adsorption of coke molecules on
acidic sites. But as discussed earlier, strong acid sites also play an important
role in the formation of coke precursors, which subsequently undergo condensa-
tion reactions to produce large polynuclear aromatic molecules that physically
coat catalytic surfaces. Physical loss of activity also occurs as coke accumulates,
ultimately partially or completely blocking catalyst pores as in supported
metal catalysts. For example, in isomerization of cis-butene on SiO2/Al2O3 (68)
catalyst deactivation occurs by rapid, selective poisoning of strong acid sites;
coke evolved early in the reaction is soluble in dichloromethane and pyridine
and is slightly aromatic. Apparently, the blocking of active sites does not signifi-
cantly affect porosity or catalyst surface area, as SiO2/Al2O3 contains relatively
large mesopores.
In the case of supported bifunctional metal/metal oxide catalysts, different
kinds of coke are formed on the metal and the acidic oxide support, eg, soft coke
(high H/C ratio) on Pt or Pt–Re metals and hard coke (low H/C ratio) on the alu-
mina support in catalytic reforming (69). In this case coke precursors may be
formed on the metal via hydrogenolysis, following which they migrate to the sup-
port and undergo polymerization and cyclization reactions, after which the larger
molecules are dehydrogenated on the metal and finally accumulate on the sup-
port, causing loss of isomerization activity. Mild sulfiding of these catalysts
(especially Pt–Re/alumina) substantially reduces the rate of hydrogenolysis
and the overall formation of coke on both metal and support; it especially reduces
the hard coke, which is mainly responsible for deactivation.
Several studies (65,70–80) have focused on coke formation during hydro-
carbon reactions in zeolites including (1) the detailed chemistry of coke precur-
sors and coke molecules formed in zeolite pores and pore intersections (or
supercages) and (2) the relative importance of adsorption on acid sites versus
pore blockage. The principal conclusions from these studies can be summarized
as follows: (1) the formation of coke and the manner in which it deactivates a zeo-
lite catalyst are shape-selective processes, (2) deactivation is mainly due to the
formation and retention of heavy aromatic clusters in pores and pore intersec-
tions, and (3) while both acid-site poisoning and pore blockage participate in the
deactivation, the former dominates at low coking rates, low coke coverages (eg,
in Y-zeolite below 2 wt%), and high temperatures, while the latter process dom-
inates at high reaction rates, low temperatures, and high coke coverages. Thus,
pore size and pore structure are probably more important than acid strength and
density under typical commercial process conditions. Indeed, deactivation is typi-
cally more rapid in zeolites having small pores or apertures and/or a monodimen-
sional structure (78). Fig. 6 illustrates four possible modes of deactivation of
HZSM-5 by carbonaceous deposits with increasing severity of coking (78).
These conclusions (in the previous paragraph) are borne out, for example,
in the study by Cerqueira and co-workers (80) of USHY zeolite deactivation dur-
ing methylcyclohexane transformation at 4508C, showing the following:
Mode 1 Mode 2
Mode 3 Mode 4
That coke formed at large contact times not only blocks pores and/or pore
intersections inside the zeolite but also migrates to the outside of zeolite crystal-
lites where it blocks pore entrances has been observed in several studies
(74,76,77,80). However, the amount, structure, and location of coke in ZSM-5
depends strongly on the coke precursor, eg, coke formed from mesitylene is
deposited on the external zeolite surface whereas coking with isobutene leads
to largely paraffinic deposits inside pores; coke from toluene, on the other
hand, is polyaromatic and is deposited both on external and internal zeolite
surfaces (74).
2.3. Thermal Degradation and Sintering. Background. Thermally
induced deactivation of catalysts results from (1) loss of catalytic surface
area due to crystallite growth of the catalytic phase, (2) loss of support area
due to support collapse and of catalytic surface area due to pore collapse on
crystallites of the active phase, and/or (3) chemical transformations of catalytic
phases to noncatalytic phases. The first two processes are typically referred
to as ‘‘sintering.’’ Sintering processes generally take place at high reaction
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 569
edges and corner sites) dissociate and diffuse readily over the surface, while at
the Tamman temperature (0.5Tmp) atoms in the bulk become mobile. Accord-
ingly, sintering rates of a metal or metal oxide are significant above the Hüttig
temperature and very high near the Tamman temperature; thus, the relative
thermal stability of metals or metal oxides can be correlated in terms of the Hüt-
tig or Tamman temperatures (92). For example, sintering of copper catalysts for
methanol synthesis is promoted by traces of chlorine in the feed, which react at
about 2258C (500 K) with the active metal/metal oxide surface to produce a
highly mobile copper chloride phase having a Tamman temperature of only
79–1748C (352–447 K) relative to 405–5278C (678–800 K) for copper metal or
metal oxides (93).
Promoters or impurities affect sintering and redispersion by either increas-
ing (eg, chlorine and sulfur) or decreasing (eg, oxygen, calcium, cesium) metal
atom mobility on the support; in the latter case this is due to their high
resistance to dissociation and migration due to high melting points as well
as their hindering dissociation and surface diffusion of other atoms. Similarly,
support surface defects or pores impede surface migration of metal particles —
especially micropores and mesopores with pore diameters about the same size
as the metal crystallites.
Historically, sintering rate data were fitted to a simple power-law expres-
sion (SPLE) of the form
where ks is the sintering rate constant, D0 the initial dispersion, and n the sin-
tering order, which for typical catalyst systems may vary from 3 to 15; unfortu-
nately, the SPLE is in general not valid for sintering processes because it
assumes that surface area or dispersion ultimately reaches zero, given sufficient
time, when in fact, for a given temperature and atmosphere, a nonzero or limit-
ing dispersion is observed after long sintering times. Moreover, the use of the
SPLE is further questionable because variations in sintering order are observed
as a function of time and temperature for a given catalyst in a fixed atmosphere
(88–90); thus, data obtained for different samples and different reaction condi-
tions cannot be quantitatively compared. Nevertheless, it has been shown by
Fuentes (94) and Bartholomew (87–89) that the effects of temperature, atmo-
sphere, metal, promoter, and support can be quantitatively determined by fitting
sintering kinetic data to the general power-law expression (GPLE)
which adds a term Deq/D0 to account for the observed asymptotic approach of
the typical dispersion versus time curve to a limiting dispersion Deq at infinite
time; m, the order of sintering, is found to be either 1 or 2. A recently compiled,
comprehensive quantitative treatment of previous sintering rate data based on
the GPLE with an order m of 2 (87–89) quantitatively addresses the effects of
catalyst properties and reaction conditions on sintering rate (91,95–97).
Sintering studies of supported metals are generally of two types: (1) studies
of commercially relevant supported metal catalysts and (2) studies of model
572 CATALYST DEACTIVATION AND REGENERATION Vol. 1
metal–support systems. The former type provides useful rate data that can be
used to predict sintering rates, while the latter type provides insights into the
mechanisms of metal particle migration and sintering, although the results can-
not be quantitatively extrapolated to predict behavior of commercial catalysts.
There is direct evidence from the previous studies of model-supported catalysts
(87,90) for the occurrence of crystallite migration (mainly in well-dispersed sys-
tems early in the sintering process), atomic migration (mainly at longer sintering
times), and spreading of metal crystallites (mainly in oxygen atmosphere). There
is also evidence that under reaction conditions, the surface is dynamic, ie, adsor-
bates and other adatoms rapidly restructure the surface and slowly bring about
faceting; moreover, thermal treatments cause gradual changes in the distribu-
tion of coordination sites to minimize surface energy. There is a trend in increas-
ing sophistication of spectroscopic tools used to study sintering and redispersion.
Additional insights into atomic and molecular processes during reaction at the
atomic scale using STM, analytical HRTEM, and other such powerful surface
science tools are expected during the next decade.
Sintering of Catalyst Carriers. Sintering of carriers has been reviewed by
Baker and co-workers (86) and Trimm (98). Single-phase oxide carriers sinter by
one or more of the following processes: (1) surface diffusion, (2) solid-state diffu-
sion, (3) evaporation/condensation of volatile atoms or molecules, (4) grain
boundary diffusion, and (5) phase transformations. In oxidizing atmospheres,
g-alumina and silica are the most thermally stable carriers; in reducing atmo-
spheres, carbons are the most thermally stable carriers. Additives and impurities
affect the thermal properties of carriers by occupying defect sites or forming new
phases. Alkali metals, for example, accelerate sintering, while calcium, barium,
nickel, and lanthanum oxides form thermally stable spinel phases with alumina.
Steam accelerates support sintering by forming mobile surface hydroxyl groups
that are subsequently volatilized at higher temperatures. Chlorine also promotes
sintering and grain growth in magnesia and titania during high temperature cal-
cination (99). By contrast, sulfuric acid treatment of hydrated alumina (gibbsite)
followed by two-step calcination results in a very stable transitional alumina
with needle-like particle morphology (98). Dispersed metals in supported metal
catalysts can also accelerate support sintering; for example, dispersed nickel
accelerates the loss of Al2O3 surface area in Ni/Al2O3 catalysts.
Effects of Sintering on Catalyst Activity. Baker and co-workers (86) have
reviewed the effects of sintering on catalytic activity. Specific activity (based on
catalytic surface area) can either increase or decrease with increasing metal
crystallite size during sintering if the reaction is structure-sensitive, or it can
be independent of changes in metal crystallite size if the reaction is structure-
insensitive. Thus, for a structure-sensitive reaction, the impact of sintering
may be either magnified or moderated; while for a structure insensitive-reaction,
sintering has in principle no effect on specific activity (per unit surface area). In
the latter case, the decrease in mass-based activity is proportional to the
decrease in metal surface area. Ethane hydrogenolysis and ethane steam reform-
ing are examples of structure-sensitive reactions, while CO hydrogenation on
supported cobalt, nickel, iron, and ruthenium is structure-insensitive.
2.4. Gas/Vapor–Solid and Solid-State Reactions. In addition to poi-
soning, there are a number of chemical routes leading to catalyst deactivation:
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 573
(1) reactions of the vapor phase with the catalyst surface to produce (a) inactive
bulk and surface phases (rather than strongly adsorbed species) or (b) volatile
compounds that exit the catalyst and reactor in the vapor phase; (2) catalytic
solid-support or catalytic solid-promoter reactions, and (3) solid-state transfor-
mations of the catalytic phases during reaction.
Gas/Vapor–Solid Reactions. Reactions of Gas/Vapor with Solid to
Produce Inactive Phases. Dispersed metals, metal oxides, metal sulfides,
and metal carbides are typical catalytic phases, the surfaces of which are similar
in composition to the bulk phases. For a given reaction, one of these catalyst
types is generally substantially more active than the others, eg, only Fe and
Ru metals are active for ammonia synthesis, while the oxides, sulfides, and car-
bides are inactive. If, therefore, one of these metal catalysts is oxidized, sulfided,
or carbided, it will lose essentially all of its activity. While these chemical modi-
fications are closely related to poisoning, the distinction here is that rather than
losing activity owing to the presence of an adsorbed species, the loss of activity is
due to the formation of a new phase altogether.
Examples of vapor-induced chemical transformations of catalysts to inac-
tive phases are listed in Table 7.
Reactions of Gas/Vapor with Solid to Produce Volatile Compounds.
Metal loss through direct vaporization is generally an insignificant route to
catalyst deactivation. By contrast, metal loss through formation of volatile
compounds, eg, metal carbonyls, oxides, sulfides, and halides in CO, O2, H2S,
and halogen-containing environments, can be significant over a wide range of
conditions, including relatively mild conditions. Classes and examples of volatile
compounds are listed in Table 8.
While the chemical properties of volatile metal carbonyls, oxides, and
halides are well known, there is surprisingly little information available on
their rates of formation during catalytic reactions. There have been no reviews
on this subject and relatively few reported studies to define the effects of metal
loss on catalytic activity (27,108–121). Most of the previous work has focused on
volatilization of Ru in automotive converters (108–111); nickel carbonyl forma-
tion in nickel catalysts during methanation of CO (113,119) or during CO chemi-
sorption at 258C (27,115); formation of Ru carbonyls during Fischer–Tropsch
synthesis (116,117); volatilization of Pt during ammonia oxidation on Pt–Rh
gauze catalysts (120,121); and volatilization of Cu from methanol synthesis
and diesel soot oxidation catalysts, leading to sintering in the former and better
catalyst–soot contact but also metal loss in the latter case (92).
Results of selected studies are summarized in Table 9.
Loss of nickel metal during CO chemisorption on nickel catalysts at tem-
peratures above 08C is also a serious problem; moreover, this loss is catalyzed
by sulfur poisoning (27). In view of the toxicity of nickel tetracarbonyl, the
rapid loss of nickel metal, and the ill-defined adsorption stoichiometries,
researchers are advised to avoid using CO chemisorption for measuring nickel
surface areas; instead, hydrogen chemisorption, an accepted ASTM method
with a well-defined adsorption stoichiometry, is recommended (124).
Decomposition of volatile platinum oxide species formed during high tem-
perature reaction may (125–127) lead to formation of large Pt crystallites and/
or substantial restructuring of the metal surface. For example, Wu and Phillips
Table 7. Examples of Reactions of Gases/Vapors with Catalytic Solids to Produce Inactive Phases
Catalytic process Gas/vapor composition Catalytic solid Deactivating chemical reaction Ref.
auto emissions control N2, O2, HCs, CO, NO, Pt–Rh/Al2O3 2 Rh2O3 þ g-Al2O3 ! RhAl2O4 þ 0.5 100,101
H2O, SO2 O2
ammonia synthesis and H2, N2 Fe/K/Al2O3 Fe ! FeO at >50 ppm O2 8
regeneration
Traces O2, H2O Fe ! FeO at >0.16 ppm H2O/H2
catalytic cracking HCs, H2, H2O La-Y-zeolite H2O induced Al migration from 8
zeolite framework causing zeolite
destruction
CO oxidation, gas tur- N2, O2, 400 ppm CO, Pt/Al2O3 2 SO3 þ g-Al2O3 ! Al2(SO4)3 which 8
bine exhaust 100–400 ppm SO2 blocks catalyst pores
diesel HC/soot emissions N2, O2, HCs (gas and Pt/Al2O3 and b-zeolite; formation of Al2(SO4)3 or sulfates of 102–104
574
control liquid), CO, NO, oxides of CaCuFeVK Ca, Cu, Fe, or V which block cat-
H2O, soot, SO2 on TiO2 alysts pores and lower activity for
oxidation; Al2O3 stabilized by
BaO
Fischer–Tropsch CO, H2, H2O, CO2, HCs Fe/K/Cu/SiO2 Fe5C2 ! Fe3O4 due to oxidation at 105
high XCO by-product H2O, CO2
Fischer–Tropsch CO, H2, H2O, HCs Co/SiO2 Co þ SiO2 ! CoO SiO2 and collapse 106
of SiO2 by-product H2O
selective catalytic N2, O2, NO, PM,a H2O, V2O5/WO3/TiO2 formation of Al2(SO4)3 if Al2O3 is 107
reduction (SCR), SO2 used
stationary
steam reforming and CH4, H2O, CO, H2, CO2 Ni/Al2O3 Ni þ Al2O3 ! Ni2Al2O4 8
regeneration in H2O
a
Particulate matter.
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 575
Transport
Metal compound vapor Lost vapor
Decomposition
Vaporization of vapor
Formation
Metal + Volatile agent Volatile compound Metal
Decomposition
Fig. 7. Generalized mechanisms and kinetics for deactivation by metal loss (8).
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 577
process are often indicative of fouling, masking, or the fracturing and accumula-
tion of attritted catalyst in the reactor bed.
Commercial catalysts are vulnerable to mechanical failure in large part
because of the manner in which they are formed; that is, catalyst granules,
spheres, extrudates, and pellets ranging in diameter from 50 mm to several milli-
meters are in general prepared by agglomeration of 0.02–2 mm aggregates of
much smaller primary particles having diameters of 10–100 nm by means of pre-
cipitation or gel formation followed by spray drying, extrusion, or compaction.
These agglomerates have in general considerably lower strengths than the pri-
mary particles and aggregates of particles from which they are formed.
Two principal mechanisms are involved in mechanical failure of catalyst
agglomerates: (1) fracture of agglomerates into smaller agglomerates of approxi-
mately 0.2d0 –0.8d0 and (2) erosion (or abrasion) of aggregates of primary parti-
cles having diameters ranging from 0.1 to 10 mm from the surface of the agglomerate
(140). While erosion is caused by mechanical stresses, fracture may be due to
mechanical, thermal, and/or chemical stresses. Mechanical stresses leading to
fracture or erosion in fluidized or slurry beds may result from (1) collisions of par-
ticles with each other or with reactor walls or (2) shear forces created by turbu-
lent eddies or collapsing bubbles (cavitation) at high fluid velocities. Thermal
stresses occur as catalyst particles are heated and/or cooled rapidly; they are
magnified by temperature gradients across particles and by differences in ther-
mal expansion coefficients at the interface of two different materials, eg, catalyst
coating/monolith interfaces; in the latter case the heating or cooling process can
lead to fracture and separation of the catalyst coating. Chemical stresses occur as
phases of different density are formed within a catalyst particle via chemical
reaction; for example, carbiding of primary iron oxide particles increases their
specific volume and micromorphology leading to stresses that break up these
particles (141).
Role of Properties of Ceramic Agglomerates in Determining Strength
and Attrition Resistance. Factors Affecting the Magnitude of Stress
Required for Agglomerate Breakage and the Mechanisms by Which it
Occurs. The extent to which a mechanism, ie, fracture or erosion, participates
in agglomerate size reduction depends upon several factors: (1) the magnitude of
a stress, (2) the strength and fracture toughness of the agglomerate, (3) agglom-
erate size and surface area, and (4) crack size and radius. Erosion (abrasion)
occurs when the stress (eg, force per area due to collision or cavitation pressure)
exceeds the agglomerate strength, ie, the strength of bonding between primary
particles. Erosion rate is reportedly (140) proportional to the external surface
area of the catalyst; thus, erosion rate increases with decreasing agglomerate
size.
Most heterogeneous catalysts are complex, multiphase materials that
consist in large part of porous ceramic materials, ie, are typically oxides, sulfides,
or metals on an oxide carrier or support. When a tensile stress of a magnitude
close to the yield point is applied, ceramics almost always undergo brittle frac-
ture before plastic deformation can occur. Brittle fracture occurs through for-
mation and propagation of cracks through the cross section of a material in a
direction perpendicular to the applied stress. Agglomerate fracture due to a
tensile stress occurs by propagation of internal and surface flaws; these flaws
580 CATALYST DEACTIVATION AND REGENERATION Vol. 1
created by external stresses or inherent defects are stress multipliers, ie, the
stress is multiplied by 2(a/r)0.5, where a is the crack length and r is the radius
of curvature of the crack tip; since a/r can vary from 2 to 1000, the effective stress
at the tip of a crack can be 4–60 times the applied stress. Tensile stress multi-
pliers may be microcracks, internal pores, and grain corners.
The ability of a material to resist fracture is termed fracture toughness. The
plain strain fracture toughness KIc is defined as
Table 11. Mechanical Strengths and Attrition Rates of Catalyst Supports Compared
to Those of Sintered Ceramic Agglomerates
Catalyst Attrition
support or Preparation/pretreatment/ Strength, index,
ceramic properties MPa wt%/h Ref.
High surface area catalyst supports
g-Al2O3, 1.2– sol–gel granulation/dried 10 h at 11.6 1.9 0.033 144
4.25-mm spheres 408C, calcined 3 h at 4508C/
389 m2/g, dpore ¼ 3.5 nm
g-Al2O3, 4.25-mm Alcoa LD-350 0.7 0.177 144
spheres
g-Al2O3, 100 mm VISTA-B-965-500C 6.2 1.3 140
TiO2 (anatase), thermal hydrolysis/dried 1108C, 28a 145
2
30 mm calcined 2 h at 5008C/92 m /g,
<10-nm primary crystallites
TiO2 (anatase), basic precipitation/dried 1108C, 15a 145
90 mm calcined 2 h at 5008C/81 m2/g,
10–14-nm primary crystallites
TiO2 (75% anatase, Degussa P25, fumed/4-mm 0.9 146
25% rutile) extrudates/48 m2/g,
Vpore ¼ 0.34 cm3/g,
dpore ¼ 21 nm
TiO2 (anatase) Rhone-Poulenc DT51, ppt./4 mm 0.9 146
extrudates/92 m2/g,
Vpore ¼ 0.40 cm3/g,
dpore ¼ 8, 65 nm
Low surface area ceramics
Al2O3 spray dried with organic binder; 2.3 147
plastic deformation observed
Al2O3 heat treated (sintered), 99.9% 282–551 142
TiO2 (Rutile) partially sintered 194 147
ZrO2 (yttria commercial samples from three 0.035–0.43 148
additive) companies, spray-dried
ZrO2 (3% Y2O3) heat treated (sintered) 800–1500 142
a
Rough estimates from break points on relative density versus log[applied pressure] curves; data are
consistent with mass distribution versus pressure curves from ultrasonic tests.
stronger than commercial g-Al2O3 spheres (143). Moreover, 30- and 90-mm dia-
meter particles of TiO2 prepared by thermal hydrolysis or basic precipitation are
30 and 15 times stronger than commercially available 4-mm extrudates (146).
Catalyst attrition is a difficult problem in the operation of moving-bed,
slurry-bed, or fluidized-bed reactors. Generally, stronger materials have greater
attrition resistance; this conclusion is supported by representative data in
Table 11 for g-Al2O3, showing that the strength of the alumina prepared by
sol–gel granulation is 17 times higher, while its attrition rate is 5 times lower.
The mechanism by which attrition occurs (erosion or fracture) can vary
with catalyst or support preparation, crush strength, and with reactor environ-
ment; it can also vary with the mechanical test method. There is some evidence
in the attrition literature supporting the hypothesis that in the presence of a
large stress, weaker oxide materials are prone to failure by fracture, while
stronger materials tend to erode (149). However, there is also contrary evidence
582 CATALYST DEACTIVATION AND REGENERATION Vol. 1
(145), showing that fracture may be the preferred mechanism for strong TiO2
agglomerates, while abrasion is favored for weaker agglomerates. Supporting a
third trend, data (140) show that attrition mechanism and rate are independent
of agglomerate strength but depend instead on the type of material. 100-
mm-diameter agglomerates of precipitated Fe/Cu/K Fischer–Tropsch catalyst
[prepared by United Catalyst (UCI)] and having nearly the same strength
shown in Table 11 for Vista-B Al2O3 (6.3 vs. 6.2 MPa), were found to undergo
substantial fracture to 5–30-mm fragments (an increase from 45 to 85%) as
well as substantial erosion to 1 mm or less fragments (increase from 2 to 50%).
Under the same treatment conditions, 90-mm-diameter agglomerates of Vista-B
Al2O3 underwent by comparison much less attrition, mainly by erosion (20%
increase in 0.1–5-mm fragments). The very low attrition resistance of the Fe/
Cu/K UCI catalyst is further emphasized by the unsatisfactory outcome of a
test of this catalyst by the U.S. Department of Energy (DOE) in a pilot-scale
slurry-phase bubble-column reactor in LaPorte, Tex.; following one day of opera-
tion, the filter system was plugged with catalyst fines, preventing catalyst–wax
separation and forcing shutdown of the plant (150).
Thus, based on these three representative examples, it follows that which of
the two attrition mechanisms predominates depends much more on material
composition and type than on agglomerate strength. However, irrespective of
mechanism the rate of attrition is usually greater for the weaker material.
The catalyst preparation method can have a large effect on the attrition
resistance of an Fe/Cu Fischer–Tropsch catalyst (151). This catalyst, prepared
by precipitation, undergoes severe attrition during a 25-min treatment with
ultrasonic radiation; indeed the mass fraction finer than 0.1–5 mm increases
from 0 to 65%. However, after a spray drying treatment of the same catalyst,
less than a 10% increase in the same fractions is evident.
In their review of attrition and attrition test methods, Bemrose and Bridge-
water (152) discuss how attrition varies with reactor type, eg, involves mainly
particle–wall impacts in moving pellet bed reactors and particle–particle
impacts in fluidized-bed reactors of high fluid velocity. In fact, jet attrition of cat-
alyst particles in a gas fluidized bed involving principally abrasion due to colli-
sion of high-velocity particles has been modeled in some detail (149,153). Thus,
given such important differences in attrition mechanism, realistic attrition test
methods should attempt to model reactor operation as closely as possible. In
addition, the ideal test would require only a small catalyst sample, a simple,
inexpensive apparatus, and a few minutes to complete the test. Relatively
quick, inexpensive single-particle crushing tests have been devised (152); how-
ever, properties of a single particle are rarely representative of those for the
bed; moreover, it is difficult to relate the results of this crushing test to the actual
abrasion process. Realistic tests have been devised for two reactor types involv-
ing a moving catalyst, ie, an air-jet test for fluidized-bed catalysts (154,155), and
a rotating drum apparatus for moving-bed catalysts (156); however, the air-jet
test requires a large quantity (eg, 50 g) of catalyst, an expensive apparatus,
and about 20 h to run. In the past decade a new jet-cup test has been developed
for testing of fluidized-bed catalysts (154,155), which requires only a 5-g sample
and about 1 h to complete; comparisons of results for the jet-cup and air-jet tests
indicate that the two tests give comparable results (154,155). Nevertheless, the
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 583
mechanisms for the two tests are different, ie, the air-jet (fluid-bed) test is
abrasion- (erosion-) dominant, while the jet-cup test includes both abrasion
and fracture mechanisms (155). A 30-min, 10-g ultrasonic attrition test based
on cavitation has also been developed in the past decade (145,151,157); while
it likewise involves both abrasion and fracture mechanisms, the results appear
to correlate with other methods. For example, particle size distributions for the
same Co/silica catalyst after ultrasonic, jet-cup, and laboratory-scale, slurry-bed
column reactor (SBCR) tests are very similar, indicating that both fracture and
abrasion mechanisms operate in the small-scale SBCR. Moreover, the good
agreement among the three methods suggests that both the jet-cup and ultraso-
nic tests may provide data representative of the attrition process in laboratory-
scale SBCR reactors. It is evident that these two small-scale methods are espe-
cially useful for screening of a series of catalysts to determine relative strength.
Nevertheless, the more realistic large-scale tests are probably needed for
accurately determining design attrition rates of a commercial catalyst to be
used in a full-scale process. The observation that attrition of a fluid catalytic
cracking (FCC) catalyst initially involves fracture of weak agglomerates followed
by abrasion of strong agglomerates emphasizes the need to collect and analyze
the particle size distribution of attrited fines as a function of time in order to
define which mechanism (or mechanisms) operates at startup as well as in the
steady-state process. Because the mechanism may be time dependent, rapid,
small-scale tests may produce misleading results.
While realistic laboratory-scale tests have been developed for simulating
attrition in large moving-bed and fluidized-bed reactors, no such laboratory
test has been developed and demonstrated yet for simulation of large-scale
SBCR reactors, although recent research has focused on the development of
such tests. For example, in laboratory-scale, SBCR tests of supported cobalt cat-
alysts over several days (157), it was observed that the attrition resistance
decreases in the order Co/Al2O3, Co/SiO2, Co/TiO2 (especially the anatase form
underwent attrition at a high rate); attrition resistance was observed to increase
with increasing cobalt loading from 10 to 40 wt%.
Implications of Mechanistic Knowledge of Attrition for Catalyst Design.
The understanding of mechanisms important in attrition of catalyst supports
and catalysts, the relationship between strength and attrition rate for a given
material, and test data can be used to great advantage in the design of attrition
resistant catalysts. Several alternatives follow from the previous discussion for
increasing attrition resistance: (1) increasing aggregate/agglomerate strength
by means of advanced preparation methods, eg, sol–gel granulation, spray dry-
ing, and carefully controlled precipitation methods (see Table 11 for examples),
(2) adding binders to improve strength and toughness, eg, the addition of a poly-
vinylpyrrolidone binder to agglomerates of quartz sand increases agglomerate
strength from 0.1 to 3 MPa (158), (3) coating aggregates with a porous but
very strong material such as ZrO2, eg, embedding a fluidized-bed catalyst for par-
tial oxidation of n-butane to maleic anhydride in a strong, amorphous matrix of
zirconium hydrogen phosphate significantly improves its attrition resistance
(159), and (4) chemical or thermal tempering of agglomerates to introduce
compressive stresses that increase strength and attrition resistance, eg, heating
and cooling particles rapidly by passing them through a low-residence-time,
584 CATALYST DEACTIVATION AND REGENERATION Vol. 1
Loss of catalytic activity may be due to a decrease in any of the three factors in
equation 4, whose product leads to kobs. Thus, catalyst deactivation can be
caused by (1) a decrease in the site density s, (2) a decrease in intrinsic activity
(ie, decrease in kintr), and/or (3) lowered access of reactants to active sites
(decrease in Z). Poisoning, for example, leads to a loss of active sites, ie,
s ¼ s0(1 a), where a is the fraction of sites poisoned; sintering causes loss of
active sites through crystallite growth and reduction of active surface area.
Fouling can cause both loss of active sites due to blocking of surface sites as
well as plugging of pores causing a decrease in the effectiveness Z. Moreover,
poisoning, as discussed earlier, can also lead to a decrease in intrinsic activity
by influencing the electronic structure of neighboring atoms. Thus, each of the
deactivation mechanisms affects one or more of the factors comprising
observed activity (see Table 12); all of the mechanisms, however, can effect a
decrease in the number of catalytic sites.
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 585
Table 12. How Deactivation Mechanisms Affect the Rate of a Catalyzed Reaction and the
Rapidity and Reversibility of Deactivation Process
Effects on reaction rate Deactivation process
extracted from their biological source, purified, crystallized, and used in labora-
tory studies or industrial processes. Further, they can be attached to glass or
ceramic supports and used as heterogeneous catalysts. And their application in
industrial processes is rapidly increasing.
Enzymes are formed in living systems by condensation/dehydration of
amino acids to produce peptide (C–N) bonds that constitute the backbone of
long protein chains. The active conformation of an enzyme is produced by folding
of the protein chain into secondary (helical), tertiary (folded), and quaternary
(combined tertiary) structures. The folded layers are held in place by hydrogen
bonding and disulfide linkages. There are 20 naturally occurring amino acids,
each having the composition H2N–CHR–COOH, the R group (side chain) having
different molecular functions, eg, proton donation, proton removal, and bridge
formation; for example, amino acids include glycine, alanine, and serine with
side chains of H, CH3, and CH2OH respectively. Upon folding, side chains
become the functional groups of the active site or ligands for binding of metals
ions, which then become functional groups at the active site. Naturally occurring
metal ions in enzymes include Mg2þ, Zn2þ, Ca2þ, Ni2þ, Fe2þ, Fe3þ, Co3þ, and
Mo2þ.
Distinctive catalytic characteristics of enzymes (162) include (1) their flex-
ible structure, which facilitates an ‘‘induced fit’’ of the substrate, the making and
breaking of bonds, and the departure of products, and (2) their sensitivity to reac-
tion effectors (inhibitors or activators), which function similarly to promoters of
heterogeneous catalysts. Some enzymes require a cofactor that combines with
the enzyme to form a catalytic site; metal ions are examples of cofactors. Enzy-
matic reactions may also require a coenzyme that reacts with the reactant to pro-
duce an enzyme-compatible substrate. Living organisms control and optimize
biological processes using a variety of tools: (1) enzyme effectors, (2) regulation
of enzyme growth or activation rates, (3) compartmentalization of enzymes
within organs or organelles, and (4) destruction (editing) of undesired intermedi-
ates or products (162).
Deactivation of Enzymes. Enzymes generally function only under mild
conditions of temperature and pH observed in living organisms. Under typical
commercial reaction conditions (40–608C, 1 atm) enzymes otherwise stable in
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 589
solution may lose activity rapidly as a result of only slight changes in their
environment such as temperature, pressure, pH, and ionic strength that induce
small free energy changes from native to denatured states (168); moreover, their
separation from the product is generally difficult and may cause further dena-
turation and loss of catalytic activity. The modest, largely reversible losses of
activity resulting from small changes in reaction environment are largely due
to modest changes in conformation of the active site. More severe changes in
reaction conditions (eg, a 108C increase in temperature) typically bring about
the dissociation and unfolding of the quaternary and tertiary structures, respec-
tively, into primary chains that subsequently order into fibrous protein bundles;
in the process active sites are irreversibly destroyed. If further exposed to severe
conditions of temperature and pH, the principal chain structure of the protein
may undergo loss or modification of functional groups or amino acid residues.
The activity of a typical enzyme increases exponentially with temperature
in accordance with the Arrhenius law up to about 50–608C, passes through a
maximum and declines precipitously above about 60–708C. Thus, catalyst life
may be on the order of days to weeks at around 508C; however, the deactivation
rate is extremely high at only slightly higher temperatures, eg, 50% loss of activ-
ity in 5 min at 65–708C is typical. Nevertheless, a few enzymes are active
and stable at temperatures exceeding 1008C; for example, a-amylase catalyzes
starch liquifaction at 105–1158C. Because their deactivation rates are highly
temperature-dependent, enzymes are generally shipped and stored under refrig-
eration (0–48C); at these low temperatures they are generally stable for months.
Causes of deactivation can be classified (as in the case of heterogeneous cat-
alysts) as chemical, mechanical, and thermal. However, for enzymes these causes
are closely linked, since mechanically and thermally induced routes almost
always effect chemical changes. Thermally induced chemical change (at elevated
temperature) is the most likely scenario for enzyme deactivation.
Chemical deactivation mechanisms include (1) changes in stereo configura-
tion by protons or hydroxyl ions at or near active sites (169,170), (2) structural
modifications in aqueous or nonaqueous solvents (171–173), (3) poisoning of
active sites by inhibitors (162,174), including ‘‘Trojan-horse inhibitors’’ that are
activated by the target enzyme (162), (4) aggregation (175), (5) unfolding (6) frag-
mentation due to solvolysis, hydrolysis in water, or self-hydrolysis (autolysis) of
proteases, eg, trypsin (176), and (7) oxidation in air (177,178). Mechanisms 1–5
may be reversible, while mechanisms 6 and 7 are generally irreversible. Mechan-
ical deactivation may be caused by hydrodynamic shear forces, eg, by stirring or
gas sparging, sometimes leading to fragmentation and/or aggregation (175,178).
Thermal inactivation of enzymes is a well-studied phenomenon (179–185);
it may be either reversible or irreversible (180). Potentially reversible changes
(due to small, short excursions in temperature near the characteristic unfolding
temperature) include light aggregation, conformational changes, folding without
further chemical change, disulfide exchange, and/or breaking of hydrogen bonds.
Irreversible denaturation (due to prolonged, severe thermal treatment) may be
caused by cleavage of disulfide bonds and/or cystinyl cross-links; unfolding fol-
lowed by chemical change; chemical changes of the primary structure and/or
active site, eg, cleavage of the polypeptide chain by hydrolysis or destruction of
individual amino acid residues; strong aggregation of inactive unfolded forms;
590 CATALYST DEACTIVATION AND REGENERATION Vol. 1
and formation of rubbery, tough fibrous structures due to alignment and bund-
ling of unfolded primary chains (similar to that observed during the boiling of an
egg). Chemical bonding of unfolded primary chains to form fibers is thermodyna-
mically favorable because chemical bonding of hydrophobic functions exposed by
unfolding lowers the entropy and hence free energy of the system.
Table 14 summarizes representative examples of enzyme deactivation by
the various mechanisms.
Methods of enhancing enzyme stability have received considerable atten-
tion (168,173,178,180,183,185–187). Strategies to improve both chemical and
thermal stability include (1) use of soluble additives, (2) immobilization, (3) pro-
tein engineering, and (4) chemical modification. Chemical modification (183,
185–187) and immobilization (164,165,172,188–191) are probably the most suc-
cessful and widely used methods. As examples of the first kind, modification of
protein surfaces by chemical binding with polysaccharides can improve thermo-
stability, while polyol binding increases enzyme solubility in organic solvents
with little loss of activity (171,176,183). Enzyme stability can be greatly
enhanced and recovery problems obviated by immobilizing (heterogenizing)
enzymes (164,165,172,188–191) through (1) covalent binding to a support,
(2) cross-linking of enzymes using a bifunctional agent, (3) adsorption on a
solid surface, (4) entrapment in a gel, or (5) containment in a membrane. More-
over, immobilization enables the catalytic process to be run continuously using a
reactor of substantially lower volume, thereby substantially reducing capital and
operating costs. These important advantages have stimulated the development
of a significant number of commercial immobilized enzyme systems.
593
low temperature CO shift 200–2508C, copper on zinc and 2–4 slow poisoning and activity
CO þ H2O ! CO2 þ H2 10–30 atm aluminum oxides accelerated sintering
by poisons
hydrocarbon hydrode 300–4008C, cobalt and molybdenum 1–10 slow coking, poisoning activity, mass
sulfurization 30 atm sulfides on aluminum by metal deposits in transfer, and
R2S þ 2 H2 ! H2S þ R2 oxide residues pressure drop
high temperature CO shift 350–5008C, Fe3O4 on chromia 1–4 slow sintering, pellet activity and
CO þ H2O ! H2 þ CO2 20–30 atm breakage due to pressure drop
steam
steam reforming, natural gas 500–8508C, nickel on calcium aluminate 1–3 sintering, sulfur- activity and
CH4 þ H2O ! CO þ 3 H2 30 atm or a-alumina poisoning, carbon pressure drop
formation, and
pellet breakage due
to plant upsets
ethylene partial oxidation 200–2708C, silver on a-alumina with 1–3 slow sintering, activity and
2 C2H4 þ O2 ! 2 C2H4O 10–20 atm alkali metal promoters poisoning by Cl, S selectivity
Table 15. ðContinued Þ
Operating Typical life Process affecting life Catalyst property
Reaction conditions Catalyst (years) of catalyst charge affected
butane oxidation to maleic 400–5208C, vanadium phosphorus 1–2 loss of P; attrition or activity and
anhydride C4H10 þ 1–3 atm oxide with transition pellet breakage; S, selectivity
3.5 O2 ! C4H2O3 þ 4 H2O metal additives Cl poisoning
reduction of aldehydes 220–2708C, copper on zinc oxide 0.5–1 slow sintering, pellet activity or
to alcohols 100–300 atm breakage (depends pressure drop
RCHO þ H2 ! RCH2OH on feedstock)
ammonia oxidation 800–9008C, Pt–Rh alloy gauze 0.1–0.5 surface roughness, selectivity
2 NH3 þ 5/2 O2 ! 1–10 atm loss of platinum,
2 NO þ 3 H2O fouling by Fe
oxychlorination of ethylene 230–2708C, copper chlorides on 0.2–0.5 loss by attrition and fluidized state
to ethylene dichloride 1–10 atm alumina (fluidized other causes and activity
2 C2H4 þ 4 HCl þ O2 ! bed) resulting from
2 C2H4Cl2 þ 2 H2O plant upsets
catalytic hydrocarbon 460–5258C, platinum alloys on 0.01–0.5 coking, frequent activity and
594
reforming 8–50 atm treated alumina regeneration mass transfer
catalytic cracking of oils 500–5608C, synthetic zeolites 0.000002 very rapid coking activity and
2–3 atm (continuous mass transfer
(fluidized bed) regeneration)
a
Adapted from Ref. 9.
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 595
are based on one important fundamental principle, namely that carbon or coke
results from a balance between the reactions that produce atomic carbon or
coke precursors and the reactions of these species with H2, H2O, or O2 that remove
them from the surface. If the conditions favor formation over gasification, these
species accumulate on the surface and react further to less active forms of carbon
or coke, which either coat the surface with an inactive film or plug the pores,
causing loss of catalyst effectiveness, pore plugging, or even destruction of the
carrier matrix.
Methods to lower rates of formation of carbon or coke precursors relative to
their rates of gasification vary with the mechanism of formation (ie, gas, surface,
or bulk phase) and the nature of the active catalytic phase (eg, metal or oxide).
For example, gas phase formation can be minimized by choosing reaction condi-
tions that minimize the formation of free radicals, by using free-radical traps, by
introducing gasifying agents (eg, H2, H2O) or gas diluents, and by minimizing
the void space available for homogeneous reaction. Similarly, the formation
and growth of carbon or coke species on metal surfaces is minimized by choosing
reaction conditions that minimize the formation of atomic carbon or coke precur-
sors and by introducing gasifying agents. Selective membranes or supercritical
conditions can also be used to lower the gas-phase and surface concentrations
of coke precursors. Since carbon or coke formation on metals apparently requires
a critical ensemble of surface metal atoms and/or dissolution of carbon into the
bulk metal, introduction of modifiers that change ensemble sizes (eg, Cu or S in
Ni or Ru) or that lower the solubility of carbon (eg, Pt in Ni) can be effective in
minimizing these forms of deactivation.
596 CATALYST DEACTIVATION AND REGENERATION Vol. 1
Mo/alumina
[metals removal]
large pore
CoMo/alumina
[finish HDS]
small pore
Fig. 9. Staged reactor system with decreasing pore size strategy for HDM/HDS of
residues (224).
pyridines from the feed; ammonia in some feedstocks, for example, can be
removed by aqueous scrubbing. The poisoning of catalysts by metal impurities
can be moderated by selective poisoning of the unwanted metal. For example,
in catalytic cracking of nickel-containing petroleum feedstocks, nickel sites,
which would otherwise produce copious amounts of coke, are selectively poisoned
by antimony (249). The poisoning of hydrotreating catalysts by nickel and vana-
dium metals can be minimized by (1) using a guard bed of inexpensive Mo cata-
lyst or a graded catalyst bed with inexpensive, low-activity Mo at the top (bed
entrance) and expensive, high-activity catalyst at the bottom (see Fig. 9) and
(2) depositing coke prior to the metals since these metal deposits can be physi-
cally removed from the catalyst during regeneration (250).
It may be possible to lower the rate of poisoning through careful choice of
reaction conditions that lower the strength of poison adsorption (48) or by choos-
ing mass-transfer-limiting regimes that limit deposits to the outer shell of the
catalyst pellet, while the main reaction occurs uninterrupted on the interior of
the pellet (246). The manner in which the active catalytic material is deposited
on a pellet (eg, uniformly or in an eggshell or egg yolk pattern) can significantly
influence the life of the catalyst (17,251).
An example of reducing catalyst poisoning (and oxidation) through process
design has been reported in a process patent for staged hydrocarbon synthesis
via the Fischer–Tropsch reaction (252). While cobalt catalysts are favored
because of their high activities and while it is desirable to achieve high conver-
sions of CO in the process, the one-pass conversion for cobalt is limited by (1) its
tendency to be oxidized at high partial pressures of product water observed at
high CO conversions and (2) its tendency to form under these conditions the
oxygenated products (eg, alcohols and aldehydes) that poison or suppress its
synthesis activity. One alternative is to separate products and recycle the unused
CO and H2, but this requires costly recompression and separation of the oxyge-
nates. Costly separation and/or poisoning can be prevented by operating a first-
stage reactor containing a cobalt catalyst to a moderately high conversion
followed by reacting the remaining CO and H2 in a second stage to above 95%
600 CATALYST DEACTIVATION AND REGENERATION Vol. 1
rare-earth oxides (eg, La2O3 and CeO2) for stabilizing Pt and/or PdO on alumina
and ZrO2 as a thermal stabilizer for the CeO2, an oxygen storage material, and as
a noninteracting support for Rh in a separate layer or in a separate phase in a
composite layer.
4.6. Prevention of Mechanical Degradation. While relatively few
studies have focused on this topic, there are nevertheless principles that guide
the design of processes and catalysts in preventing or minimizing mechanical
degradation. In terms of catalyst design it is important to (1) choose supports,
support additives, and coatings that have high fracture toughness, (2) use pre-
paration methods that favor strong bonding of primary particles and agglo-
merates in pellets and monolith coatings, (3) minimize (or rather optimize)
porosity (thus maximizing density), and (4) use binders such as carbon to facil-
itate plastic deformation and thus protect against brittle fracture. Processes (and
to some extent preparation procedures) should be designed to minimize (1) highly
turbulent shear flows or cavitation that lead to fracture of particles or separation
of coatings, (2) large thermal gradients or thermal cycling leading to thermal
stresses, and (3) formation of chemical phases of substantially different densities
or formation of carbon filaments leading to fracture of primary particles and
agglomerates. Nevertheless, thermal or chemical tempering can be used in a con-
trolled fashion to strengthen catalyst particles or agglomerates.
Examples of catalyst design to minimize attrition can be found in the recent
scientific (239,240) and patent (241–244) literature focusing on the Fischer–
Tropsch synthesis in slurry reactors. These studies indicate that (1) spray drying
of particles improves their density and attrition resistance; (2) addition of silica
and/or alumina into titania improves its attrition resistance, while addition of
only 2000–3000 ppm of titania to g-alumina improves alumina’s attrition resis-
tance; and (3) preformed alumina spheres promoted with La2O3 provide greater
attrition resistance relative to silica. Increasing attrition resistance is apparently
correlated with increasing density (239,240,244). According to Singleton and co-
workers (244), attrition resistance of Co/Al2O3 is improved when the g-alumina
support is (1) formed from synthetic boehmite having a crystallite diameter of
4–5 nm and (2) pretreated in acidic solution having a pH of 1–3; moreover,
attrition resistance decreases in the order Co/Al2O3, Co/SiO2, Co/TiO2 and is
greater for catalyst prepared by aqueous versus nonaqueous impregnation.
Despite our best efforts to prevent it, the loss of catalytic activity in most pro-
cesses is inevitable. When the activity has declined to a critical level, a choice
must be made among four alternatives: (1) restore the activity of the catalyst,
(2) use it for another application, (3) reclaim and recycle the important and/or
expensive catalytic components, or (4) discard the catalyst. The first alternative
(regeneration and reuse) is almost always preferred; catalyst disposal is usually
the last resort especially in view of environmental considerations.
The ability to reactivate a catalyst depends upon the reversibility of the
deactivation process. For example, carbon and coke formation is relatively easily
reversed through gasification with hydrogen, water, or oxygen. Sintering on the
602 CATALYST DEACTIVATION AND REGENERATION Vol. 1
603
naphtha/Pt/Al2O3 pro- coke produced by conden- in a moving bed at 300–6008C, followed by oxychlorination
moted with Re, Sn, Ge, or Ir sation of aromatics and (350–5508C) (2) coke on Pt/zeolite is removed in halogen-free
olefins oxygen-containing gas at T < 4158C (3) sintering during
oxidation of coke on Pt–Ir/Al2O3 catalyst can be minimized at
low regeneration temperature (4) study of influence of heating
rate, temperature, and time on structural properties of regen-
erated Pt–Sn/Al2O3 (5) study of effects of Cl, Sn content, and
regeneration sequence on dispersion and selectivity of
Pt–Sn/Al2O3 (6) regenerated Pt–Re/Al2O3 is more stable that
the fresh catalyst in n-heptane conversion and more selective
for toluene
dehydrogenation of propane catalyst activity is low owing (1) temperatures gradients were measured during burn off of coke 275,275
and butane/Cr2O3/Al2O3, to equilibrium limitations formed on a chromia–alumina catalyst during butene dehy-
Cr2O3/ZrO2, FeO/K/MgO, and build-up of product drogenation; data were used in developing a mathematical
Pt/Al2O3, Pt–Sn/Al2O3, H2; rapid loss of activity model for predicting temperatures and coke profiles (2) coked
Pt–Sn/KL-zeolite occurs owing to coke for- supported palladium catalyst used in the dehydrogenation of
mation dimethyltertrahydronaphthalenes to dimethylnaphthalenes is
reactivated with an organic polar solvent at a temperature
below 2008C
Table 17. ðContinuedÞ
Deactivation mechanism/
reaction/catalyst Problem/cause Method(s) of regeneration/phenomena studied/conclusions Refs.
Fischer–Tropsch synthesis/ loss of activity due to block- (1) carbidic surface carbon deposited on cobalt can be largely 276,277
Co/Al2O3 ing of sites by carbon removed in hydrogen at 170–2008C and in steam at 300–4008C
overlayers and heavy (2) slurry-phase cobalt catalysts may lose 50% activity during
hydrocarbons synthesis over a period of a few days; the activity can be reju-
venated in situ by injecting H2 gas into vertical draft tubes
inside the reactor
fluid catalytic cracking (FCC) rapid loss of activity due to (1) process and apparatus for increasing the coke burning capa- 278,279
of heavy hydrocarbons/ poisoning of acid sites and city of FCC regenerators; auxiliary regenerator partially burns
USY or REO-Yb in silica blocking of small zeolite off the coke at turbulent or fast fluidized-bed conditions (2)
matrix pores by coke multistage fluidized-bed regeneration of spent FCC catalyst in
a single vessel by incorporating two relatively dense phase
fluidized beds beneath a common dilute phase region
hydrocracking of heavy loss of activity due to poi- (1) regeneration of noble metal/zeolite via progressive partial 280,281
naphtha/CoMo, NiW, MoW soning of acid sites and removal of carbonaceous deposits under controlled oxidizing
604
on Al2O3 or SiO2 –Al2O3; Pt blocking of small zeolite conditions to maximize sorption of a probe molecule while
or Pd on Y-zeolite, morde- pores by coke minimizing metal sintering (2) regeneration of noble metal/
nite, or ZSM-5 zeolite in air at about 6008C, followed by a mild treatment in
aqueous ammonia to improve catalytic activity
hydrotreating of gas oil loss of activity due to forma- (1) TPO studies of oxidative regeneration of CoMo and NiW HDS 282,283,
tion of types I, II, and III catalysts; sulfur is removed at 225–3258C, carbon at 375– 297–300
coke on metal sulfide and 5758C. Redispersion of NiW was observed by EXAFS (2) phy-
alumina surfaces and in sicochemical changes in CoMo and NiCoMo HDS catalysts
pores during oxidative regeneration, including redispersion of Co, Ni,
and Mo oxides and surface area loss, were examined (3) changes
in NiMo catalyst structure and coke composition during reac-
tion and regeneration were examined and correlated (4) prop-
erties of NiMo catalyst deactivated during shale oil
hydrogenation and regenerated in O2 or H2 were examined.
regeneration in 1.6% O2 was more effective than that in 5% H2.
Ni aluminate spinel was observed after burn off (5) hard and
soft cokes formed on CoMo catalysts during HDS of gas oil were
characterized. At low coke levels, hard coke was more easily
removed in H2 than in O2 (6) spent catalysts are washed with
solvent and contacted with steam at about 6008C
methanol to olefins or gaso- severe coking and deactiva- (1) kinetics of coke burnoff from a SAPO-34 used in converting 284,285
line/silica–alumina, Y-zeo- tion of silica–alumina and methanol to olefins were studied; kinetics are strongly depen-
lite, ZSM-5, other zeolites, Y-zeolite catalysts dent on the nature of the coke. Kinetics are slowed by strong
and aluminophosphate observed during high con- binding of coke to acid sites (2) ZSM-34 catalyst used in con-
molecular sieves versions of methanol, also version of methanol to light olefins is effectively regenerated in
substantial coking of H2-containing gas; this approach avoids the formation of cata-
ZSM-5, other lyst-damaging products such as steam that would be formed
zeolites, and aluminopho- during burn off in air
sphate molecular sieves
Poisoning
FCC of residues/USY or (1) poisoning of acid sites by (1) organometallic solutions of Sb and Bi are added to process 284,285,
REO-Y in silica matrix N- containing compounds. steam to passivate Ni by forming inactive Ni–Sb and Ni–Bi 301–304
(2) deposition of Ni and V species (2) V metal deposits are trapped by reaction with
metals on acid sites which magnesium orthosilicate to form an unreactive magnesium
change selectivity and vanadium silicate (3) spent metal-contaminated catalyst is
decrease activity demetallized by chlorinating and washing followed by con-
tacting with NH4F and one antimony compound (4) metal-
contaminated catalyst is contacted with an aqueous solution
of a carboxylic acid (eg, formic, acetic, citric, or lactic acid) (5)
605
metal-contaminated catalyst is contacted with HCl, HNO3, or
H2SO4 (6) metal contaminated catalyst is contacted with
reducing CO gas to form gaseous metal carbonyls that sepa-
rated from the catalyst
hydrogenation or poisoning of metal sites by (1) regeneration of Ni/SiO2 catalyst poisoned by thiophene using a 288,289
dechlorination arsenic, sulfur, and other sequence of oxidation–reduction treatments at low PO2 and 1
poisons atm H2 respectively (2) regeneration in dilute hypochlorite
solution of a Pd/Al2O3 catalyst deactivated during the aqueous-
phase dechlorination of trichloroethylene in the presence of
sulfite or HS ions present in ground water
hydrotreating of residues/ pore-mouth poisoning and (1) regeneration of catalysts containing V, Ni, or Fe by contacting 290,291
Al2O3-supported Mo and blockage by Ni, V, and Fe with H2O2 solution and organic acid (2) following removal of
CoMo sulfides present in feed as coke by air or solvent wash, catalyst is acid leached to remove
organometallics undesired metals
Table 17. ðContinuedÞ
Deactivation mechanism/
reaction/catalyst Problem/cause Method(s) of regeneration/phenomena studied/conclusions Refs.
Thermal degradation
Catalytic reforming of sintering of Pt causing (1) redispersion of Pt–Ir bimetallic catalysts using a wet HCl/air 270,273,
naphtha/Pt/Al2O3 pro- formation of large metal treatment, since the conventional oxychlorination is not effec- 292,293
moted with Re, Sn, Ge, or crystallites crystals and tive (2) redispersion of Pt/KL-zeolite using wet HCl/air treat-
Ir; Pt/KL-zeolite loss of active surface areament followed by brief calcination and reduction (3)
redispersion of Pt–Re/Al2O3 in Cl2 and O2 (4) redispersion of
supported Pt, other noble metals, and Ni in Cl2 and O2
hydrocracking of heavy sintering of noble metal redispersion of noble metals on molecular sieves including 294
naphtha/CoMo, NiW, MoW causing formation of large silica-aluminates, ALPOS, SAPOS
on Al2O3 or SiO2 –Al2O3; metal crystallites crystals
Pt or Pd on Y-zeolite, and loss of active surface
mordenite, or ZSM-5 area
hydrotreating of gas oil and sintering of Mo and Co sul- (1) oxidative regeneration of hydroprocessing catalyst at 6008C 295,296
residues/Al2O3-supported fides causing formation of optimizes surface area and Mo dispersion (2) oxidative regen-
606
Mo and CoMo large sulfide crystals and eration in several steps with a final oxidation at 500–6008C to
loss of active surface area restore residual catalyst activity
a
USY: ultrastable Y-zeolite.bREO-Y: rare-earth exchanged Y-zeolite.
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 607
Ni S þ H2 O ! NiO þ H2 S ð5Þ
H2 S þ 2 H2 O ! SO2 þ 3 H2 ð6Þ
coke burning some redispersion occurs, eg, D increases from 0.25 to 0.51, while dur-
ing oxychlorination the dispersion is further increased, eg, from 0.51 to 0.81 (262).
Some guidelines and principles regarding the redispersion process are
worth enumerating:
crystallite, which stresses and ultimately leads to splitting of the particle (310).
Dadyburjor hypothesizes that the crystallite splitting model is most applicable to
the behavior of large crystallites and to all particles at relatively small regenera-
tion times while the thermodynamic migration model is useful for small particles
and most particles after longer regeneration times.
6.1. Summary
These trends are also likely to continue. Moreover, the combination of more
sophisticated methods and models will hasten the practical application of models
for predicting catalyst/process life. This is already happening in selected compa-
nies. For example, for more than a decade now operators at Phillips Petroleum
Co. have been using deactivation models (developed at their corporate research)
in their refineries to predict when shutdown will be necessary. One of these
models enables them to predict accurately the lifetime of hydrotreating catalysts
on the basis of catalyst and feedstock properties.
6.3. Future Needs. Collection of Data. It is evident from careful
examination of the literature that few deactivation rate data are available for
even the most important large-scale catalytic systems. Accordingly, there is a cri-
tical need for collection of such data at the laboratory, bench, and plant scale.
There is much that could be done with good data. Sophisticated analytical
tools and well-designed reactors are available at most companies for collecting
and analyzing such data. The field is ripe and ready to harvest. The wise will
seize these opportunities.
Data Analysis and Model Development. Much of the previously col-
lected data were analyzed using outdated methods. There is much that could be
learned by reanalyzing some of these data using new approaches such as the
GPLE and microkinetic modeling. Critical reviews that include collections of
carefully selected rate and kinetic data would constitute important contributions
612 CATALYST DEACTIVATION AND REGENERATION Vol. 1
BIBLIOGRAPHY
‘‘Catalyst Regeneration, Metal Catalysts’’ in ECT (online), posting date: December 4, 2000,
by D. W. Robinson, UOP.
107. G. Busca, L. Lietti, G. Ramis, and F. Berti, Appl. Catal., B: Environ. 18, 1–36 (1998).
108. T. P. Kobylinski, B. W. Taylor, and J. E. Yong, in Proc. SAE, Detroit, 1974.
109. M. Shelef and H. S. Gandhi, Platinum Met. Rev. 18, 1 (1974).
110. H. S. Gandhi, H. K. Stepien, and M. Shelef, Mat. Res. Bull. 10, 837 (1975).
111. C. H. Bartholomew, Ind. Eng. Chem. Prod. Res. Dev. 14, 29 (1975).
112. R. W. Clark, J. K. Tien, and P. Wynblatt, J. Catal. 61, 15 (1980).
113. W. M. Shen, J. A. Dumesic, and C. G. Hill, J. Catal. 68, 152 (1981).
114. R. B. Pannell, K. S. Chung, and C. H. Bartholomew, J. Catal. 46, 340 (1977).
115. G. Lohrengel and M. Baerns, Appl. Catal. 1, 3 (1981).
116. I. Qamar and J. G. Goodwin, in Proc. 8th Am. Meeting Catal. Soc., Philadelphia,
1983 (Paper C-22).
117. J. G. Goodwin, D. O. Goa, S. Erdal, and F. H. Rogan, Appl. Catal. 24, 199 (1986).
118. O. Watzenberger, T. Haeberle, D. T. Lynch, and G. Emig, in Ref. 11, p. 441.
119. M. Agnelli, M. Kolb, and C. Mirodatos, J. Catal. 148, 9 (1994).
120. H. C. Lee and R. J. Farrauto, Ind. Eng. Chem. Res. 18, 1 (1989).
121. R. J. Farrauto and H. C. Lee, Ind. Eng. Chem. Res. 29, 1125 (1990).
122. F. Sperner and W. Hohmann, Platinum Met. Rev. 20, 12 (1976).
123. J. M. Hess and J. Phillips, J. Catal. 136, 149 (1992).
124. C. H. Bartholomew, Catalysis Lett. 7 (1990).
125. N. L. Wu and J. Phillips, J. Phys. Chem. 89, 591 (1985).
126. N. L. Wu and J. Phillips, Appl. Phys. 59, 769 (1986).
127. N. L. Wu and J. Phillips, J. Catal. 113, 129 (1988).
128. A. Bielanski and M. Najbar, in Ref. 9, p. 127.
129. N. Burriesci, F. Garbassi, M. Petrera, and G. Petrini, in Ref. 9, p. 115.
130. Y. L. Xiong, R. Castillo, C. Papadopoulou, L. Dada, J. Ladriere, P. Ruiz, and
B. Delmon, in Ref. 11, p. 425.
131. R. J. Farrauto, M. Hobson, T. Kennelly, and E. Waterman, Appl. Catal. 81, 227
(1992).
132. P. L. Gai-Boyes, Catal. Rev. Sci. Eng. 34, 1 (1992).
133. B. Delmon, in Ref. 12, p. 113.
134. K. M. Erickson, D. A. Karydis, S. Boghosian, and R. Fehrmann, J. Catal. 155, 32
(1995).
135. B. Delmon, in Ref. 13, p. 39.
136. N. B. Jackson, A. K. Datye, L. Mansker, R. J. O’Brien, and B. H. Davis, in Ref. 13,
p. 501.
137. S. A. Eliason and C. H. Bartholomew, in Ref. 13, p. 517.
138. A. Baranski, R. Dziembaj, A. Kotarba, A. Golebiowski, Z. Janecki, and J. B. C.
Pettersson, in Ref. 14, p. 229.
139. C. A. Querini, F. Ravelli, M. Ulla, L. Cornaglia, and E. Miro, in Ref. 14, p. 257.
140. H. N. Pham, J. Reardon, and A. K. Datye, Powder Technol. 103, 95 (1999).
141. D. S. Kalakkad, M. D. Shroff, S. Kohler, N. Jackson, and A. K. Datye, Appl. Catal.
133, 335 (1995).
142. W. D. Callister, Materials Science and Engineering: An Introduction, John Wiley &
Sons, Inc., New York, 2000.
143. R. L. Coble and W. D. Kingery, J. Am. Ceram. Soc. 39, 381 (1956).
144. S. G. Deng and Y. S. Lin, AIChE J. 43, 505 (1997).
145. S. G. Thoma, M. Ciftcioglu, and D. M. Smith, Powder Technol. 68, 53 (1991).
146. M. Bankmann, R. Brand, B. H. Engler, and J. Ohmer, Catal. Today 14, 225
(1992).
147. V. M. Kenkre and M. R. Endicott, J. Am. Ceram. Soc. 79, 3045 (1996).
148. H. Song and J. R. G. Evans, J. Am. Ceram. Soc. 77, 806 (1994).
149. J. Werther and W. Xi, Powder Technol. 76, 39 (1993).
616 CATALYST DEACTIVATION AND REGENERATION Vol. 1
187. M. G. Roig and J. F. Kennedy, Crit. Rev. Biotechnol. 12, 391 (1992).
188. A. M. Klibanov, Anal. Biochem. 93, 1 (1979).
189. P. Monsan and D. Combes, Methods Enzymol. 137, 584 (1988).
190. J. M. Guisan, P. Sabuquillo, R. Fernandez-Lafuent, G. Fernandez-Lorente, C. Ma-
teo, P. J. Halling, D. Kennedy, E. Miyata, and D. Re., J. Mol. Catal., B: Enzymatic
11, 817 (2001).
191. P. Wang, S. Dai, S. D. Waezsada, A. Y. Tsao, and B. H. Davison, Biotechnol. Bioeng.
74, 249 (2001).
192. R. G. Silver, J. C. Summers, and W. B. Williamson, Catalysis and Automotive Pollu-
tion Control II, Elsevier, Amsterdam, 1991, p. 167.
193. G. B. Fisher, M. G. Zammit, and J. LaBarge, SAE Report 920846, 1992.
194. R. J. Farrauto and R. M. Heck, Catal. Today 51, 351 (1999).
195. G. W. Huber, C. G. Guymon, B. C. Stephenson, and C. H. Bartholomew, Catalyst
Deactivation 2001 (Studies in Surface Science and Catalysis, Vol. 139), Elsevier,
Amsterdam, 2001, p. 423.
196. U.S. Pat. 6,169,120 (Jan. 2, 2001), G. L. Beer (to Syntroleum Corp.).
197. C. H. Bartholomew, M. W. Stoker, L. Mansker, and A. Datye, in Ref. 14, p. 265.
198. U.S. Pat. 5,728,894 (Mar. 17, 1998), O. Nagano and T. Watanabe (to Ashahi Kasei
Kogyo Kabushiki Kaisha).
199. T. Maillet, J. Barbier, and D. Duprez, Appl. Catal., B 9, 251 (1996).
200. WIPO Pat. 93/16020A3 (Sept. 16, 1993), G. Mathys, L. Martens, M. Baes,
J. Verduijn, D. Huybrechts, and C. Renata (to Exxon Chem.).
201. U.S. Pat. 5,672,800 (Sept. 30, 1997), G. Mathys, L. Martens, M. Baes, J. Verduijn,
and D. Huybrechts (to Exxon Chem.).
202. U.S. Pat. 6,080,903 (June 26, 2000), L. Stine, B. Muldoon, S. Gimre, and R. Frame (to
UOP).
203. B. Subramaniam, V. Arunajatesan, and C. J. Lyon, in Ref. 14, p. 63.
204. WIPO Pat. 99/33769 (July 8, 1999), D. Ginosar, R. Fox, and P. Kong (to Lockheed
Martin).
205. F. H. Ribeiro, A. L. Bonivardi, and C. Kim, J. Catal. 150, 186 (1994).
206. D. Ginosar and B. Subramaniam, in Ref. 12, p. 327.
207. E. E. Petersen, in Ref. 13, p. 87.
208. J. W. Gosselink and J. A. R. V. Veen, in Ref. 14, p. 3.
209. L. Lin, T. Zao, J. Zang, and Z. Xu, Appl. Catal. 67, 11 (1990).
210. D. E. Resasco and G. L. Haller, Catalysis (Spec. Period. Rept.) 11, 379 (1994).
211. R. D. Cortright and J. A. Dumesic, J. Catal. 148, 771 (1994).
212. H. Weyten, K. Keizer, A. Kinoo, J. Luyten, and R. Leysen, AIChE J. 43, 1819
(1997).
213. P. Praserthdam, T. Mongkhonsi, S. Kunatippapong, B. Jaikaew, and N. Lim, in
Ref. 13, p. 153.
214. WIPO Pat. 00/69993 (May 12, 2000), B. Rose and T. Kiliany (to Mobil Oil Corp.).
215. A. Guerrero-Ruiz, A. Sepulveda-Escribano, and I. Rodriguez-Ramos, Catal. Today
21, 545 (1994).
216. D. Qin and J. Lapszewicz, Catal. Today 21, 551 (1994).
217. S. Stagg and D. Resasco, in Ref. 13, p. 543.
218. K. Fujimoto, K. Tomishige, O. Yamazaki, Y. Chen, and X. Li, Res. Chem. Intermed.
24, 259 (1998).
219. U.S. Pat. 5,191,142 (Mar. 2, 1993), C. Marshall and J. Miller (to Amoco Corp.).
220. A. G. Gayubo, A. T. Aguayo, A. E. S. D. Campo, P. L. Benito, and J. Bilbao, in Ref. 14,
p. 129.
221. U.S. Pat. 5,248,647 (Sept. 28, 1993), P. T. Barger (to UOP).
222. WIPO Pat. 99/42202 (Feb. 20, 1998), J. Cox and J. Evans (to Johnson Matthey).
618 CATALYST DEACTIVATION AND REGENERATION Vol. 1
223. WIPO Pat. 98/50487 (May 2, 1997), S. Leviness, C. Mart, W. Behrmann, S. Hsia, and
D. Neskora (to Exxon Research and Engineering Co.).
224. C. H. Bartholomew, in M. Oballa and S. Shih, eds., Catalytic Hydroprocessing of
Petroleum and Distillates, Marcel Dekker, New York, 1993, p. 1.
225. J. Summers and W. B. Williamson, in J. Armor, ed., Environmental Catalysis 1993,
Vol. 552, American Chemical Society, Washington, D.C., 1993, p. 94.
226. J. Dettling, Z. Hu, Y. K. Lui, R. Smaling, C. Z. Wan, and A. Punke, in Studies in
Surface Science and Catalysis, Vol. 96, Elsevier, Amsterdam, 1995, p. 461.
227. U.S. Pat. 4,910,180 (Mar. 20, 1990), M. Berndt and D. Ksinsik (to Doduco).
228. U.S. Pat. 4,985,387 (Jan. 15, 1991), M. Prigent, G. Blanchard, and P. Phillippe (to
ProCatalyse).
229. U.S. Pat. 5,041,407 (Aug. 20, 1991), W. Williamson, D. Linden, and J. Summers (to
Allied-Signal Inc.).
230. U.S. Pat. 5,116,800 (May 26, 1992), W. Williamson, D. Linden, and J. Summers (to
Allied-Signal Inc.).
231. U.S. Pat. 5,234,881 (Aug. 10, 1993), C. Narula, W. Watkins, and M. Chattha (to Ford
Motor Co.).
232. U.S. Pat. 5,254,519 (Oct. 19, 1993), C. Wan, S. Tauster, and H. Rabinowitz (to
Engelhard Corp.).
233. T. Furuya, S. Yamanaka, T. Hayata, J. Koezuka, T. Yoshine, and A. Ohkoshi, in
Proc. Gas Turbine Conference and Exhibition, Anaheim, Calif., 1987.
234. T. Kawakami, T. Furuya, Y. Sasaki, T. Yoshine, Y. Furuse, and M. Hoshino, in Proc.
Gas Turbine and Aeroengine Congress and Exposition, Toronto, Ont., June 4–8,
1989.
235. U.S. Pat. 5,250,489 (1993), R. D. Betta, F. Ribeiro, T. Shoji, K. Tsurumi, N. Ezawa,
and S. Nickolas (to Catalytica, Inc.).
236. T. Fujii, Y. Ozawa, and S. Kikumoto, J. Eng. Gas Turbines Power 120, 509 (1998).
237. D. O. Borio and N. S. Schbib, Comput. Chem. Eng. 19, S345 (1995).
238. U.S. Pat. 5,028,634 (July 2, 1991), R. Fiato (to Exxon Research and Engineering
Co.).
239. R. Zhao and J. G. Goodwin, Jr., K. Jothimurugesan, S. K. Gangwal, and J. J. Spivey,
Ind. Eng. Chem. Res. 40, 1065–1075 (2001).
240. R. Zhao, J. G. Goodwin, Jr., K. Jothimurugesan, S. K. Gangwal, and J. J. Spivey,
Ind. Eng. Chem. Res. 40, 1076 (2001).
241. U.S. Pat. 5,939,350 (Aug. 17, 1999), A. H. Singleton, R. Oukaci, and J. G. Goodwin
(to Energy International Corp.).
242. U.S. Pat. 6,087,405 (July 11, 2000), S. Plecha, C. H. Mauldin, and L. E. Pedrick (to
Exxon Research and Engineering Co.).
243. U.S. Pat. 6,124,367 (Sept. 26, 2000), S. Plecha, C. H. Mauldin, and L. E. Pedrick (to
Exxon Research and Engineering Co.).
244. W.O. Pat. 00/71253 (Nov. 30, 2000), A. H. Singleton, R. Oukaci, and J. G. Goodwin
(to Energy International Corp.).
245. M. Seitz, E. Klemm and G. Emig, in Ref. 14, p. 211.
246. S. Masamune and J. M. Smith, AIChE J. 12, 384 (1966).
247. Y. Murakami, T. Kobayashi, T. Hattori, and M. Masuda, Ind. Eng. Chem. Fundam.
7, 599–605 (1968).
248. W. H. J. Stork, in G. F. Froment, B. Delmon, and P. Grange, eds., Hydrotreatment
and Hydrocracking of Oil Fractions, (Studies in Surface Science and Catalysis,
Vol. 106), Elsevier, New York, 1997, pp. 41–67.
249. G. D. Parks, A. M. Schaffer, M. J. Dreiling, and C. B. Shiblom, Prepr.—Am. Chem.
Soc., Div. Petr. Chem. 25, 335 (1980).
Vol. 1 CATALYST DEACTIVATION AND REGENERATION 619
CALVIN BARTHOLOMEW
Brigham Young University
Vol. 1 DISTILLATION 621
DISTILLATION
1. Introduction
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0409192006010918.a01.pub2
622 DISTILLATION Vol. 1
2. Vapor–Liquid Equilibria
1.0
Equilibrium line
Vapor dew yLA
point line 0.8 Point A
TE
e
lin
xLE yLE
y
=
Temperature
°x
0.6
45
yL
xLB xLF yLD
0.4
TA
0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
x L or y L xL
(a) (b)
Fig. 1. Isobaric VLE diagrams: (a) dew and bubble point; (b) vapor–liquid (y–x)
equilibrium.
Vol. 1 DISTILLATION 623
Vapor P
Liquid P
Fig. 2. Equilibrium between vapor and liquid. The conditions for equilibrium are
TV ¼ TL and PV ¼ PL . For a given T and P, phase fugacities are equal, ie, f V ¼ f L and
fi V ¼ fi L .
The upper line is the vapor dew-point line, representing points at which a vapor
on cooling forms the first drop of condensed liquid. The liquid and vapor compo-
sitions are conventionally plotted in terms of the low-boiling (more volatile) sub-
stance, L, in the mixture. The system point A has a vapor composition of yLA in
equilibrium with a liquid composition of xLA at a temperature of TA. Figure 1b is
a typical isobaric phase or y–x diagram. For further discussion, several textbooks
are available (15,16).
2.1. Thermodynamic Relationships. A closed container with vapor
and liquid phases at thermodynamic equilibrium may be depicted as in
Figure 2, where at least two mixture components are present in each phase.
The components distribute themselves between the phases according to their
relative volatilities. A distribution ratio for mixture component i may be defined
using mole fractions:
Ki ¼ yi =xi ð1Þ
K i y i x j y i ð 1 xi Þ
aij ¼ ¼ ¼ ð2Þ
K j x i y j xi ð 1 yi Þ
624 DISTILLATION Vol. 1
pi V ¼ xi Pi 0 ð3Þ
Ki ¼ gi L Pi 0 =P ð6Þ
aij ¼ gi L Pi 0 gj L Pj 0 ð7Þ
10
γ1∝
∝
γ ∝2
Activity coefficient (log scale), γ
1.0
γ ∝2
∝
γ ∝
∝
1
γ 1
1.0 0.2
0 1.0 0 1.0
x1 x1
(a) (b)
Fig. 3. Binary activity coefficients for two component systems having (a) positive and
(b) negative deviations from Raoult’s law. Conditions are either constant pressure or
constant temperature and terminal coefficients, g1
i , are noted.
Terminal activity coefficients, g1i , are noted in Figure 3. These are often
called infinite dilution coefficients and for some systems are given in Table 1.
The hexane–heptane mixture is included as an example of an ideal system. As
the molecular species become more dissimilar, they are prone to repel each other,
tend toward liquid immiscibility, and have large positive activity coefficients, as
in the case of hexane–water.
If the molecular species in the liquid tend to form complexes, the system
will have negative deviations and activity coefficients less than unity, eg, the
system chloroform–ethyl acetate. In azeotropic and extractive distillation (see
DISTILLATION, AZEOTROPIC AND EXTRACTIVE) and in liquid–liquid extraction, nonideal
liquid behavior is used to enhance component separation. An extensive discus-
sion on the selection of nonideal addition agents is available (17).
A great deal of study and research has gone into the development of work-
ing equations that can represent the curves of Figure 3. These equations are
based on solutions of the Gibbs-Duhem equation:
" # " #
a ln gi a ln gn
xi þ xn ¼0 ð8Þ
axi axi
T;P T;P
One of the simplest and often used equations, or models, is that of Van Laar
(18). For a binary system of components 1 and 2, these equations are
A12
lng ¼ " #2 ð9Þ
A12 x1
1
A21 x2
A21
lng2 ¼ " #2 ð10Þ
A21 x2
1
A12 x1
It should be noted that only two parameters are involved. They are directly
related to the terminal activity coefficients:
ln g1
1 ¼ A12 ð11Þ
ln g1
2 ¼ A21 ð12Þ
This, the Wilson model, is more complex than the Van Laar model, but it
does retain the two-parameter feature. The terminal activity coefficients are
related to the parameters:
ln g1
1 ¼ 1 ln 12 21 ð15Þ
ln g1
1 ¼ 1 ln 21 12 ð16Þ
Whereas the Wilson model has been found to represent a wide variety of nonideal
VLE, it cannot handle the case of partial immiscibility of the liquid phase; for
this purpose a three-parameter relationship, the nonrandom, two-liquid
(NRTL) model was developed (20).
The most recently developed model is called UNIQUAC (21). Comparisons
of measured VLE and predicted values from the Van Laar (22), Wilson, NRTL,
Vol. 1 DISTILLATION 627
and UNIQUAC models, as well as an older model, are available (3). Thousands
of comparisons have been made, and Reference 3, which covers the Dortmund
Data Base, available for purchase and use with standard computers, should be
consulted by anyone considering the measurement or prediction of VLE. The pre-
dictive VLE models can be accommodated to multicomponent systems through
the use of certain combining rules. These rules require the determination of
parameters for all possible binary pairs in the multicomponent mixture. It is
possible to use more than one model in determining binary pair data for a
given mixture (23).
To estimate VLE when no experimental data or model parameters are
available and the cost of special measurements cannot be justified, a group con-
tribution method based on the molecular structures involved called UNIFAC (8)
has been developed. Not all possible groups have been evaluated, but regular
progress reports are published. The UNIFAC method, as well as the other mod-
els, are critically important in extending limited data to conditions in distillation
columns that can cover wide ranges of temperatures, pressures, and composi-
tions. Handling of all these models by computer solution has been described in
some detail (24).
The vapor–liquid equilibria of dilute solutions are frequently expressed in
terms of Henry’s law:
p i V ¼ H i xi ð17Þ
Hi ¼ gi L Pi 0 ð18Þ
Henry’s law is useful for handling equilibria associated with gas absorption (qv)
and stripping problems. Henry’s law coefficients are useful for estimating term-
inal activity coefficients and have been tabulated for many compounds in dilute
aqueous solutions (25,26).
The foregoing discussion has dealt with nonidealities in the liquid phase
under conditions where the vapor phase mixes ideally and where pressure–
temperature effects do not result in deviations from the ideal gas law. Such con-
ditions are by far the most common in commercial distillation practice. However,
it is appropriate here to set forth the completely rigorous thermodynamic expres-
sion for the K value:
Z !
p
1
gLi f0i p0i exp vLi dP
R0 T p0i
Ki ¼ ð19Þ
^ iP
f
For nonideal vapor-phase behavior, the fugacity coefficient for component i in the
mixture must be determined:
Z P
^1 ¼ 1 v R0 T
ln f v i dP ð20Þ
R0 T 0 P
628 DISTILLATION Vol. 1
At pressures less than 2 MPa (20 bar) and temperatures greater than 273 K,
PC 1:0. When the vapor obeys the ideal gas law, z ¼ 1:0; then for ideal vapor
solutions and for conditions such that PC ¼ 1:0, Equation 19 reduces to
Equation 6.
The fugacity coefficient departure from nonideality in the vapor phase can
be evaluated from equations of state or, for approximate work, from fugacity/
compressibility estimation charts. References 11,14, and 25 provide valuable
insights into this matter.
Journals for the publication of VLE data are available as are comprehensive
tabulations of azeotropic data (27,28); if the composition and temperature of the
azeotrope are known (at a given pressure), then such information may be used to
calculate activity coefficients. At the azeotropic point, by definition, yi ¼ xi ; from
Equation 6,
gi L ¼ P=Pi 0 ð24Þ
The vapor pressure Pi0 can be obtained from any of many sources such as hand-
books.
The measurement of VLE can be carried out in several ways. A common
procedure is to use a recycle still, which is designed to ensure equilibrium
between the phases. Samples are then taken and analyzed by suitable methods.
It is possible in some cases to extract equilibrium data from chromatographic
procedures. Discussions of experimental methods are available (5,11). For the
more challenging measurements, eg, conditions where one or more components
in the mixture can decompose or polymerize, commercial laboratories can be
used.
2.2. Azeotropic Systems. An azeotropic mixture is one that vaporizes
without any change in composition. Figures 4 and 5 represent homogeneous
azeotropic systems. Figure 4 depicts a minimum boiling azeotropic system
such as ethanol–water; Figure 5 describes a maximum boiling azeotropic system
such as acetone–chloroform. The point Z defines the azeotropic composition; this
azeotropic mixture is also called the constant boiling mixture (CBM). Positive
activity coefficients tend to produce minimum boiling azeotropes, and negative
coefficients tend to produce maximum boiling azeotropes.
Vol. 1 DISTILLATION 629
1.0
0.8
C, D
0.6
yL
Temperature
Z
0.4
C D
A, B
A B
0.2
0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
xL or yL xL
(a) (b)
Fig. 4. Boiling point (a) and phase diagram (b) for a minimum boiling binary azeotropic
system at constant pressure. A,B and C,D are representative equilibrium points; Z is the
azeotropic point.
Z 1.0
C, D
D
C
A B 0.8
Temperature
0.6
yL
0.4
0.2
A, B
0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
xL or yL xL
(a) (b)
Fig. 5. Boiling point (a) and phase diagram (b) for a maximum boiling binary azeotropic
system at constant pressure. A,B and C,D are representative equilibrium points; Z is the
azeotropic point.
630 DISTILLATION Vol. 1
120
1.0
110 C, D
0.8 M
Temperature, °C
100 0.4 A, B
Z
D
0.2
C
x vs T M N
90 0
0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0
Mol fraction water in liquid or vapor Mol fraction water in liquid
(a) (b)
Fig. 6. Boiling point (a) and phase diagram (b) for the heterogeneous azeotropic system,
water/1-butanol at atmospheric pressure. A,B and C,D are representative equilibrium
points; Z is the azeotropic point; M and N are liquid miscibility limits.
the immiscible range, M–N, the equilibrium vapor is the heterogeneous azeo-
trope, Z, of constant composition and the equilibrium temperature is constant.
At liquid compositions lower in water than in the azeotrope, the relative volati-
lity of water/1-butanol is greater than one; at liquid compositions higher in water
than in the azeotrope, the relative volatility of water/1-butanol is less than one
(see DISTILLATION, AZEOTROPIC AND EXTRACTIVE).
3. Distillation Processes
F ¼DþB ð25Þ
the component L balance is
x L F F ¼ yL D D þ xL B B ð26Þ
−
qc
Cond
enser
T
V
Condensate
Vn+1 Distillate D
n+1 Reflux R
Ln + 1 Balance loop
Vn
n
V n−1 L n
n−1
f
V
Feed F Ln−1 −
f
Lf
V m+1
m+1
Vm
Lm + 1
m
V m−1
Lm
m−1
Vs
Lm−1
−
q− s
Ls
Calandria
Bottoms B
Fig. 7. Distillation column with stacked multiple equilibrium stages. Terms are defined
in text.
F ¼DþB ð27Þ
and for any component i,
HF F þ HS ¼ HD D þ HB B þ HC ð29Þ
A mass balance around plate n and the top of the column gives:
V n 1 ¼ Ln þ D ð30Þ
Vol. 1 DISTILLATION 633
V n 1 yi n 1 ¼ Ln xi n þ Dxi D ð31Þ
n
L D
yn 1
i ¼ x n
þ xD ð32Þ
V n 1 i V n 1 i
Below the feed, a similar balance around plate m and the bottom of the column
results in:
m
L B
ym 1
i ¼ x m
þ xB ð33Þ
V m 1 i V m 1 i
Equation 32 represents the upper (or rectifying) operating line equation, and
Equation 33 represents the lower (or stripping) operating line equation. The
slopes Ln =V n 1 and Lm =V m 1 can vary, depending on heat effects.
Graphical Method. The graphical McCabe–Thiele (30) design method
facilitates a visualization of distillation principles while providing a solution to
the material balance and equilibrium relationships. Here, the subscripts L and
H are not used, and x and y refer to the lower boiler, ie, more volatile component,
in the binary system. A McCabe–Thiele diagram is given in Figure 8 where P, Q,
and S are the xB, xF, and xD compositions on the y ¼ x, 458 construction line,
respectively. Line OP is the stripping operating line and line OS is the rectifying
operating line.
1.0
S
y n, x n
y n+1, x n+1
y n−1, x n
Mol fraction low boiler in vapor, y
y n, x n+1
O q line plate n
y m +1, x m +1
Q
y m, x m
plate m
y m, x m+1
y m−1, x m
P
0
xB xF x D 1.0
Mol fraction low boiler in liquid, x
The constant molal flows in each section are designated by L and V. The
McCabe–Thiele assumption of constant molal overflow implies that the molal
latent heats of the two components are identical, the sensible heat effects are
negligible, and the heat of mixing and the heat losses are zero. This simplified
situation is closely approximated for many distillations. Equation 34 now
represents the straight upper operating line OS, and Equation 35 represents
the straight lower operating line OP. The upper operating line has the slope
(L/V)R and the intercept at xD (¼ yD ) on the x ¼ y line. Note that this operating
line slope is less than one. Similarly, the lower operating line has a slope of
(L/V)S and the intercept is at xB on the y ¼ x line. This operating line has a
slope greater than one. The line QO from the feed intercept Q to the intersection
of the operating lines at O is called the q line.
The equilibrium curve gives the vapor–liquid relationships of yn and xn
above the feed and of ym and xm below the feed. The upper operating line gives
the relationship between yn 1 and xn and the lower operating line gives the rela-
tionship ym 1 and xm, ie, the streams passing each other. The graphical represen-
tation of theoretical equilibrium stages n and m is shown. The ym 1 , xm to ym,
xLþ1 represent the mass balance and phase equilibrium for theoretical stage m.
Similarly, yn 1 , xn to yn, xnþ1 represent theoretical stage n. The total number of
theoretical stages in the column can now be stepped off starting either at the
composition xB and stepping upward or starting at xD and stepping downward.
Condition of Feed (q Line). The q line, which marks the transition from
rectifying to stripping operating lines, is determined by mass and enthalpy bal-
ances around the feed plate. These balances are detailed in distillation texts (15).
The slope of the q line is q/(q 1) where:
e r
lin pe ine
riu
m Up ng l
ib i
uil era
t
Eq op e
E lin
y
D =
x
5 °,
C 4
B
Q
A
Lower
operating
lines
xF
Fig. 9. McCabe–Thiele q lines for various feed enthalpy conditions. Terms are defined in
text.
Reflux and Reflux Ratio. The liquid returned to the top of the column is
called reflux. The molar ratio R/D is the external reflux ratio. The ratio (L/V)R,
which is the slope of the rectifying operating line, is the rectifying internal reflux
ratio. Similarly, the ratio (L/V)S, which is the slope of the stripping operating
line, is the stripping internal reflux ratio. As the ratio R/D increases, the rectify-
ing internal reflux ratio increases and numerically approaches unity; similarly,
the stripping internal reflux ratio decreases and numerically approaches unity.
In the McCabe–Thiele plot the two operating lines move away from the equili-
brium line toward the y ¼ x diagonal as the reflux ratio increases, and the indi-
vidual theoretical stage steps become larger; accordingly, fewer theoretical
stages are required to make a given separation.
McCabe–Thiele Example. Assume a binary system L–H that has ideal
vapor–liquid equilibria and a relative volatility of 2.0. The feed is 100 mol of
xF ¼ 0:6; the required distillate is xD ¼ 0:95, and the bottoms xB ¼ 0:05, with
the compositions identified and the lighter component L. The feed is at the boil-
ing point. To calculate the minimum reflux ratio, the minimum number of theo-
retical stages, the operating reflux ratio, and the number of theoretical stages,
assume the operating reflux ratio is 1.5 times the minimum reflux ratio and
there is no subcooling of the reflux stream, then:
(1) Calculate the vapor composition in equilibrium with the liquid feed. From
Equation 2a and for x ¼ 0:60 mol fraction;
2ð0:6Þ
y ¼ ¼ 0:75 mol fraction ð37Þ
1 þ ð2:0 1Þ0:6
636 DISTILLATION Vol. 1
1.0 15
14
13
0.9 xD
12
11
0.8 10
9
8
0.7
7
Mol fraction L in vapor, y
0.6
6
xF
0.5
5
0.4
4
0.3
3
0.2
2
0.1 1
xB
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Mol fraction L in liquids, x
(4) Calculate reflux ratios. The minimum internal reflux ratio is a line
from the intercept of the q line with the equilibrium curve to the xD point
on the 458 line:
(9) Stripping section liquid and vapor flows, because the feed is at the boiling
point,
(11) Theoretical stages: the complete construction is shown in Figure 10. Stages
were stepped off starting at the base. Approximately 14.2 theoretical stages
are required. This includes the reboiler, which normally functions as an
equilibrium stage. Therefore, a capability of 13.2 theoretical stages in the
column is needed. If the condenser were to condense only reflux, with the
distillate product leaving the process as a vapor, it could be counted also as
an equilibrium stage, making 12.2 stages needed for the column.
xD
xD
Mol fraction L in vapor, y
xB xB
(a) (b)
Fig. 11. Limiting conditions in binary distillation. (a) Minimum reflux and infinite
number of theoretical stages; (b) total reflux and minimum number of theoretical
stages.
the separation. The minimum number of theoretical stages occurs when the
system is at total reflux: no feed, distillate, or bottoms. This is illustrated in
Figure 11b, where the operating lines are coincident with the 458, y ¼ x line.
For the McCabe– Thiele example presented above, the graphical proce-
dure would give slightly less than nine minimum theoretical stages including
the reboiler.
Simple analytical methods are available for determining minimum
stages and minimum reflux ratio. Although developed for binary mixtures,
they can often be applied to multicomponent mixtures if the two key com-
ponents are used. These are the components between which the specifica-
tion separation must be made; frequently the heavy key is the component
with a maximum allowable composition in the distillate and the light key
is the component with a maximum allowable specification in the bottoms.
On this basis, minimum stages may be calculated by means of the Fenske
relationship (33):
where i and j are the light and heavy components of a binary mixture, or the light
key and heavy key in a multicomponent mixture. The average relative volatility
is often taken as the geometric average of the relative volatilities at the top and
bottom of the column. For the McCabe–Thiele example,
ln½ð0:95=0:05Þ ð0:95=0:05Þ
Nmin ¼ ¼ 8:50 stages
ln 2:0
Vol. 1 DISTILLATION 639
For minimum reflux ratio, the following equations (35) may be used:
X ai xif
¼1 q ð39Þ
i
ai f
X ai ðxid Þ
¼ Rmin þ 1 ð40Þ
i
ai f
2:0ð0:6Þ 1:0ð0:4Þ
þ ¼ 1 q ¼ 0 ðbecause q ¼ 1Þ
2:0 f 1:0 f
solving f ¼ 1:25. Substituting in equation 40, for the given distillate composi-
tions,
2:0ð0:95Þ 1:0ð0:05Þ
þ ¼ Rmin þ 1
2:0 1:25 1:0 1:25
Theoretical stages
Minimum stages
Reflux ratio
Fig. 12. Representative plot of theoretical stages vs reflux ratio for a given separation.
Each curve is the locus of points for a given separation. Note the limiting conditions of
minimum reflux and minimum stages.
Total cost
Annual cost
Operating cost
Fixed cost
(R/D)min
Reflux ratio, R/D
Fig. 13. Fixed, operating, and total costs of a typical distillation, as a function of reflux
ratio.
Vol. 1 DISTILLATION 641
1.0
xD
P
Mol fraction of L in vapor, y
Q
q line
xF
xB
0 1.0
Mol fraction of L in liquids, x
Fig. 14. False minimum reflux for system of skewed equilibria. Minimum reflux occurs
at intersection P of operating line and equilibrium line, not at intersection of q line and
equilibrium line. Terms are defined in text.
in the distillate and in the bottoms is specified for the overall solution of a
binary distillation, the component balance around the column also is comple-
tely specified. In the multicomponent case, only a single high-boiling key
component can be specified in the distillate and a single low-boiling key com-
ponent in the bottoms; the split of other components can be determined
only by detailed calculations. These require a series of trial and error com-
putations to obtain the solution at any given reflux ratio and number of
stages. As the number of components and number of stages become large,
the mathematical problem becomes formidable. Two approaches may be fol-
lowed: use of approximate, ie, shortcut, methods, or use of a suitable computer
program that provides rigorous solutions. The former are used when approxi-
mate solutions are adequate or when a computer is not available. For the
latter, numerous commercial programs are available and may be used with
personal computers.
Most shortcut methods involve: (1) calculating the minimum number of
stages; (2) calculating the minimum reflux ratio; and (3) estimating, from empiri-
cal correlations, the actual number of stages at an operating reflux. For mini-
mum stages the Fenske relationship (eq. 38) is used, whereas for minimum
reflux ratio the Underwood relationships (eqs. 39 and 40) are used. The relation-
ship of operating to minimum reflux ratio and of operating to minimum number
of plates is then estimated from the Gilliland correlation (37), or from a more
recent correlation such as that of Reference 38.
642 DISTILLATION Vol. 1
The Gilliland correlation is in graphical form and the curve has been fitted
by several workers, an example of which is (39):
" #0:5668
Nt Nmin R Rmin
¼ 0:75 0:75 ð41Þ
Nt þ 1 Rþ1
from which Nt ¼ 15:7 stages. The original plot of Gilliland would give Nt ¼ 14:8,
closer to the McCabe–Thiele value of 14.2 stages.
Discussions of shortcut methods have appeared many times in the litera-
ture (16,35), accompanied by the usual admonition to use such methods only
for approximate designs or analyses. For multicomponent systems having signif-
icant nonidealities, the shortcut methods can be grossly in error.
Rigorous computer solutions are used for complex distillations involving
multiple stages, multiple components, nonideal phase equilibria, multiple feeds
and drawoffs, and heat addition or removal at intermediate stages. Most
calculations are made by computer and the algorithms are generally based on
the Thiele–Geddes model (40), which rates a given number of stages and reflux
ratio for separation capability. A detailed discussion of computer solutions,
including the handling of convergence problems, is available (41).
Computer solutions entail setting up component equilibrium and compo-
nent mass and enthalpy balances around each theoretical stage and specifying
the required design variables as well as solving the large number of simulta-
neous equations required. The explicit solution to these equations remains too
complex for present methods. Studies to solve the mathematical problem by algo-
rithm or iterational methods have been successful and, with a few exceptions, the
most complex distillation problems can be solved.
3.3. Multiple Products. If each component of a multicomponent distilla-
tion is to be essentially pure when recovered, the number of columns required for
the distillation system is N* 1, where N* is the number of components. Thus, in
a five-component system, recovery of all five components as essentially pure pro-
ducts requires four separate columns. However, those four columns can be
arranged in 14 different ways (42).
The number of columns in a multicomponent train can be reduced from the
N* 1 relationship if side-stream draw-offs are used for some of the component
cuts. The feasibility of multicomponent separation by such draw-offs depends
on side-stream purity requirements, feed compositions, and equilibrium relation-
ships. In most cases, side-stream draw-off distillations are economically feasible
only if component specifications for the side-stream are not tight. If a single com-
ponent is to be recovered in an essentially pure state from a mixture containing
both lower- and higher-boiling components, a minimum of two columns is
required, one column to separate the lower boilers from the desired component
and another column to separate the component from the higher boilers.
Vol. 1 DISTILLATION 643
The economics of the various methods that are employed to sequence multi-
component columns have been studied. For example, the separation of three-,
four-, and five-component mixtures has been considered (43) where the heuris-
tics (rules of thumb) developed by earlier investigators were examined and an
economic analysis of various methods of sequencing the columns was made.
The study of sequencing of multicomponent columns is part of a broader field,
process synthesis, which attempts to formalize and develop strategies for the
optimum overall process (44).
4. Distillation Columns
Downcomer
Liquid
deentraining
zone
Bubbling
area
Exit weir
Downcomer
Vapor
clearance
disengaging
from liquid
Degassed
liquid
from 0.3 to 15 m and plate spacings range from 0.15 to 1.2 m. The total pressure
drop per plate is often in the range of 0.25–1.6 kPa (2–12 mm Hg).
Three principal vapor–liquid contacting devices are used in current cross-
flow plate design: the sieve plate, the valve plate, and a hybrid of the tow in
which the ‘‘valves’’ are in a fixed, open position. In past years another type of
crossflow plate containing bubble caps was specified, and there are some existing
columns containing such vapor dispering devices. Details of the bubble cap tray
and its design features were covered by Bolles (45). All the devices mentioned
function to provide the needed intimate contacting of vapor and liquid, requisite
to maximizing transfer of mass across the vapor–liquid interfacial boundary.
Sieve Plates. The conventional sieve or perforated plate is inexpensive
and the simplest of the devices normally used. The contacting orifices in the con-
ventional sieve plate are holes that measure 1 to 12 mm diameter and exhibit
ratios of open area to active area ranging from 1:20 to 1:7. If the open area is
too small, the pressure drop across the plate is excessive; if the open area is
too large, the liquid weeps or dumps through the holes.
Valve Plates. Valve plates are categorized as proprietary and details of
design vary from one vendor to another. These represent a variation of the sieve
plate in which the holes are large and are fitted with liftable valve units such as
those shown in Figure 16. The principal advantage over sieve plates is the ability
to maintain efficient operation over a wider operating range through the use of
variable orifices (valves), which open or close depending on vapor rate. The most
common valve units consist of flat disks having attached legs that allow the valve
Vol. 1 DISTILLATION 645
w
flo
uid
Liq
Fig. 16. Examples of individual valve units used in valve plates: (a) Flexitray valve,
courtesy of Koch-Glitsch Inc.; (b) Float Valve, and (c) Fixed Valve (V-Grid), both courtesy
of Sulzer Chem-Tech, Inc.
to move upward or downward. Sometimes two weights of valves are used on a sin-
gle plate to extend operating range and improve vapor distribution. The valve
units usually have a tab or indentation that provides a minimum open area of
vapor flow, even when the valve is closed, and also prevents the valve from stick-
ing under corrosive or fouling conditions. Details on valve plate geometry, along
with methods for valve plate design, are available from valve plate vendors.
Fixed Valve Plates. For several years plates with a combination of sieve
holes and movable valves were marketed by several vendors, and can still be
obtained on special order. Of more current interest is the crossflow tray con-
taining small ‘‘valves’’ that are fixed in the full open position. Some designers
are specifing that the area under the downcomers be perforated with these fixed
valves, requiring the downcomers to be suspended as shown in Figure 17. This
Fig. 17. Flow pattern in a high-capacity crossflow plate with suspended downcomer.
646 DISTILLATION Vol. 1
arrangement allows an increase in vapor flow but requires careful design of the
downcomer to prevent vapor by-passing up the downcomer.
Multiple Liquid-Path Plates. As the liquid flow rate increases in large
diameter crossflow plates (ca 4 m or larger), the crest heads on the overflow weirs
and the hydraulic gradient of the liquid flowing across the plate become exces-
sive. To obtain improved overall plate performance, multiple liquid-flow-path
plates may be used, with multiple downcomers. These designs are illustrated
and discussed in detail in the literature (45).
Counterflow Plates. Counterflow plates are used less frequently than
crossflow plates. The liquid flows downward and the vapor upward through the
same orifices in a counterflow plate and the plate does not have downcomers.
The openings are round holes (dualflow tray) or slots (Turbogrid tray). A variation
of the dualflow tray is the Ripple tray in which the tray floor is shaped in a corru-
gated fashion (46). Counterflow plates are used advantageously in fouling services
because for each hole vapor and liquid flow alternately, providing a self-cleaning
action that is quite effective. The dualflow and Turbogrid plates have similar oper-
ating characteristics, and typical operating data have been published (47).
Another important plate which has characteristics similar to a counterflow
plate is the Multiple Downcomer (MD) plate (48). This is a plate where the active
area occupies the full column cross section but with a plurality of small downco-
mers interspersed among the perforations. The downcomers are specially sealed
to prevent upflow of vapor through them, as in the case of high-capacity sus-
pended down comer plates (Fig. 17). The plate has been used successfully in
many high-liquid-flow cases.
Vapor Capacity Parameters. The diameter of a distillation column is
determined by the capacity of the column to handle the required flows of vapor
and liquid. The vapor capacity parameter is
" #0:5
rg
Csb ¼ V ð42Þ
rL rg
, m/s
0.20
0.2
( ( ( 20
σ 0.10
900
600
0.08
450
0.5
0.06
300
ρL − ρg
225
ρg
0.04 150
( Csb = V ∗
0.02
Fig. 18. Flooding correlation for crossflow trays (sieve, valve, bubble-cap) where the
numbers represent tray spacing in mm. Also shown are approximate boundaries of the
spray zone, and mixed froth and emulsion flow regimes.
B
Floo
ding
Blowing
Vapor rate
r
t
come
inmen
ing
entra h
Stable operation
Hig
flood
Down
D
p ing
g, dum
Weepin
A
Liquid rate
liquid flow rates. Below AD, the vapor rate is too low to maintain the liquid on
the plate and, as a result, the liquid weeps excessively or dumps through the
plate orifices. Above line BC the column floods by entrainment. To the right of
CD the high liquid rate causes downcomer flooding. The area to the left of AB
represents high entrainment at low liquid flow rates, with vapor jets at the ori-
fices. Design procedures for sieve plates have been published (eg 49,50). Vendors
of valve trays make available their design methods.
Plate Efficiencies. Column requirements are calculated in terms of theo-
retical stages or plates. Actual plates must, however, be specified in the design.
Thus the effectiveness of the plate in approaching the equilibrium condition
must be predicted. This approach is called the plate efficiency, which is a mea-
sure of the rate of mass transfer on the actual plate. This efficiency, expressed
either as a fraction or as a percentage, depends on three principal factors: the
geometry of the plate (hole arrangement, valve design, etc); the loading of
vapor and liquid traffic on the plate; and the diffusional properties of the fluids.
The simplest efficiency is the overall column efficiency, which is the number
of theoretical plates in a column divided by the number of actual plates:
Eo ¼ Nt =Na ð44Þ
Thus the overall efficiency is an averaged efficiency of all the individual plates.
A more useful plate efficiency for theoretical prediction is the Murphree
plate efficiency:
yn yn 1
Emv ¼ ð45Þ
yn yn 1
where yn and yn 1 are the vapor compositions from plate n and n 1 (the plate
below n), and yn* is the vapor composition that would be in equilibrium with
Vol. 1 DISTILLATION 649
the liquid composition leaving plate n. Thus, for a given plate, Emv is a ratio of
the actual vapor composition change to the change that would occur if the plate
were effective enough to bring the vapor and liquid to thermodynamic equili-
brium. This definition is based on the outlet liquid composition, and says nothing
about the average liquid composition on the plate. In cases where a significant
concentration gradient exists in the liquid composition across the plate, it is pos-
sible for Emv to have a value greater than 1.0 (100%). Equation 45 is written in
terms of vapor composition. A similar equation can be written in terms of the
liquid compositions and is denoted as EmL.
Of still more theoretical importance is the efficiency at some point on the
plate:
" #
yn yn 1
Eog ¼ ð46Þ
yn yn 1
point
This parameter is called the point efficiency (or local efficiency). It cannot have a
value greater than 1.0, and it has a counterpart term for liquid compositions.
Prediction of Plate Efficiency. As of this writing, the most comprehen-
sive study of plate efficiency known was made in the mid-1950s, based on the
then-still-popular bubble cap plates (51). Unfortunately, the predictive model
developed has been shown to be inadequate for many industrial distillations.
An improvement of the model, more oriented toward the sieve plate, was pub-
lished in 1984 (52). There has been continuing research effort directed toward
a better understanding of the mechanisms that occur in the rather complex aer-
ated mass on the typical plate (53–55). A complicating factor is the lack of uni-
form liquid flow across the plate, and situations have been found where the liquid
actually stagnates in certain zones of larger-diameter plates. For larger columns
it is possible for the observed Murphree efficiency to exceed 100%. A satisfactory
method for predicting plate efficiency does not exist. Most recently there have
been studies of the various types of flow regimes that occur on operating plates
and of the effect of these regimes on tray performance, including plate efficiency.
Pursuit of the flow regime studies (56–59) may lead to improved plate efficiency
prediction methods. For example, a newer model (60) takes into account the
regime as well as the vapor bubble (froth flow) or liquid drop (spray flow) char-
acteristics in determining mass-transfer coefficients in the aerated zone on the
plate.
Empirical methods for predicting plate efficiency have been proposed. Prob-
ably the most widely used method correlates overall column efficiency as a func-
tion of feed viscosity and relative volatility (61).
General Comments on Plate Efficiency. The plate efficiencies of well-
designed commercial bubble cap, sieve, and valve plates are approximately the
same when the plates are operated within their normal design range. The
plate efficiency decreases both at the low end of the plate’s operating range,
where the liquid tends to leak through the plate, and at the high end of the oper-
ating range, where liquid entrainment becomes substantial.
Most distillation systems in commercial columns have Murphree plate effi-
ciencies of 70% or higher. Lower efficiencies are found under system conditions of
650 DISTILLATION Vol. 1
a high slope of the equilibrium curve (Fig. 1b), of high liquid viscosity, and of
large molecules having characteristically low diffusion coefficients. Finally,
most experimental efficiencies have been for binary systems where by definition
the efficiency of one component is equal to that of the other component. For mul-
ticomponent systems it is possible for each component to have a different effi-
ciency. Practice has been to use a pseudo-binary approach involving the two
key components. However, a theory for multicomponent efficiency prediction
have been developed (62,63) and are amenable to computational analysis.
4.2. Packed Columns. In packed columns, the vapor–liquid contacting
takes place in continuous beds of solid packing elements rather than in discrete
individual plates. The contacting can be visualized as occurring in differential
increments across the height of the packing; thus packings are known as coun-
terflow devices rather than stagewise devices. Mechanically, the packed column
is a relatively simple structure. In its simplest form the packed column comprises
a vertical shell having dumped or carefully arranged packing elements on an
open-type support, together with a suitable liquid distribution device above the
packed bed. A packed column having two packed beds and a midcolumn feed is
shown in Figure 20. The vapor enters the column below the bottom bed and flows
Packing
Packing support
Feed
Feed
distributor
Packing
support
Packing
Vapor Liquid
Fig. 20. Packed column shell and internals. Column shown has single packed beds above
and below the feed. For separations requiring a large number of stages, additional beds,
separated by redistribution devices, are likely to be needed.
Vol. 1 DISTILLATION 651
upward through the column. The liquid (reflux or other liquid stream) enters at
the top through the liquid distributor and flows downward through the packing
counter-currently to the rising vapor. The height of the individual packed beds is
limited to 2–9 m by the mechanical strength of the packing or by the need to
redistribute the liquid so that good mass-transfer efficiency can be maintained.
Packings. For many years packed columns consisted of randomly dum-
ped packings almost exclusively, with occasional applications of regularly
stacked packings or pads of woven or knitted wire. In the late 1960s a partial
trend away from random packings began when a special structured packing
made of wire gauze was introduced by Sulzer Brothers in Switzerland (64).
The indicated advantages of the structured packings were high mass-transfer
efficiency and very low pressure drop. These devices appeared to be ideal for
high-vacuum distillations. However, cost of fabrication was very high and they
were considered mainly for the vacuum distillation of specialty chemicals. In
1977, a lower-cost sheet metal version was introduced (65), and since that time
a large business in structured sheet metal packings has arisen. At the same time,
improved random packings have been developed and a comprehensive discussion
of their characteristics has been published (66). Some of the common random
packings are shown in Figure 21. The Raschig ring, one of the oldest of packings,
is an open cylinder of equal height and diameter. The Berl saddle and the cera-
mic Intalox saddle (Norton Co.) have a higher capacity and efficiency than the
Raschig ring. The Pall ring is a modification of the Raschig ring which allows
through-flow of liquid and vapor, with consequent lower pressure drop and better
efficiency. The newer Intalox metal saddle (IMTP) is an example of a random
packing having a very high void fraction and low resistance to the flowing
phases. Other newer random packings, not shown in Figure 21, include the
CMR ring (Koch-Glitsch, Inc.) and the Nutter ring (Sulzer Chemtech Inc.). The
random-type packings can generally be made from metal, plastic, or ceramic
materials; the approximate nominal size range for the individual elements is
12–75 mm.
Common structured packing geometries are shown in Figure 22. Flat plates
of gauze or sheet metal are perforated, or embossed or lanced, and corrugated.
Corrugated sheets are then stacked together such that adjacent sheets have
opposite corrugation directions. The corrugations have angles with the hori-
zontal of 45 to 60 degrees. Vapor and liquid contact each other in wetted-wall
Raschig
Fig. 21. Random packing elements for distillation columns: (a) Raschig ring (metal);
(b) Berl saddle (ceramic); (c) Intalox saddle (ceramic); (d) Pall ring (metal); and (e) Intalox
saddle (metal).
652 DISTILLATION Vol. 1
Flow channel
(a)
s
h
(b)
Fig. 22. (a) Flow channel arrangement; (b) flow channel triangular cross section where
for angles of 908, DEQ ¼ 4RH ¼ 4ð S2 Þ 2S
1
¼ S.
fashion, and the perforations plus other surface enhancements, eg, texturing,
serve to promote liquid spreading into thin films. Dimensions, performance char-
acteristics, and design procedures for the structured packings are summarized in
Reference 67.
Packed Column Internals. In order to ensure good packed column mass-
transfer efficiency, the liquid must be distributed uniformly over the surface of
the packing. As a general rule there should be at least 100 pour points per square
meter (10 points/ft2), although fewer points may be used for random packings of
the bluff-body type such as Raschig rings and Berl saddles. Although they have
capacity and pressure drop limitations, the bluff-body packing elements are able
to divide the downflowing liquid and thus improve on an initially marginal
distribution. On the other hand, the through-flow-type random packings, eg,
Pall rings and Intalox metal tower packings, as well as the structured packings,
are not able to correct the initial distribution and in fact may allow some dete-
rioration of distribution if the bed heights are greater than about 5 m.
Vol. 1 DISTILLATION 653
2.0
2.0
e
rat
0.5
rat
kPa/m
g
uid
0.5
uid
kin
iq
ac
liq
hl
yp
Low
Hig
Dr
0.1
0.1
Gas rate, mass per unit cross-sectional area per unit time
Fig. 23. Log–log plot of pressure drop per unit height of typical packing as a function of
gas rate at two liquid rates and for the unirrigated packing.
0.2
1.2
0.82
0.1
0.41
0.06
,(mm2 /s) 0.1
0.04 0.20
0.02
0.08
gρ g ( ρL − ρG )
G 2Fp v 0.1
0.01
∆ P = 0.04
0.006
0.004
0.002
0.001
0.0006
0.0004
0.004 0.04 0.4 4.0 8.0
0.006 0.02 0.06 0.2 0.6 2.0 6.0
0.01 0.1 1.0
ρg
( (
0.5
L
G ρL
Fig. 24. Generalized method using log scales for estimating packed column flooding and
pressure drop, DP, in kPa/m; g ¼ gravitational constant, 9.81 m/s2; v ¼ kinematic
viscosity in mm2 =s (¼ cSt); L, G have units of kg/(m2 s); rL, rG are in kg/m3; and the pack-
ing factor, Fp, in m 1 can be found in Table 2. To convert kPa/m to mm Hg/m, multiply
by 7.5 (66).
constant-pressure drop lines (75). The most recent revision in these correlations
is shown in Figure 24 (66). The idea of flooding has been eliminated from the
chart with the stipulation that the topmost curve represents the maximum capa-
city. Experimentally determined packing factors Fp, presented in Table 2, should
be used in the ordinate group. These factors distinguish between the various
shapes and sizes of the available packings. The curves are for constant pressure
drop and thus the chart enables estimation of both capacity and pressure drop.
Packing Mass-Transfer Characteristics. The contacting for mass trans-
fer in a packed column occurs differentially along the length of the column. The
separation calculations can thus be made on a differential basis along this
length, using mass-transfer coefficients or heights of transfer units. The calcula-
tions are somewhat imprecise because of the uncertainty in the fundamental
mass-transfer mechanisms in larger-scale columns. Useful models for predicting
the mass-transfer efficiency of randomly packed columns have been published
(76,77), using the same database of commercial-scale performance data. These
models cover the better known packings, eg, metal and ceramic Raschig rings,
ceramic Berl saddles, and metal Pall rings, in nominal sizes in the range of
12 to 50 mm. It has been found that to avoid excessive maldistribution of liquid
Vol. 1 DISTILLATION 655
near the wall, a ratio of column diameter to packing element size of at least eight
should be maintained. Thus if one wishes to conduct pilot-scale packed column
tests, a minimum column diameter of about 100 mm would be used together
with 12-mm packing elements. The models would then permit scale-up to large
columns containing 50-mm size elements of the same type, eg, Pall rings.
These models provide values of the height of a transfer unit for the liquid
phase HL and the vapor phase HV. These values are combined to form the height
of an overall transfer unit, Hov:
where V and L are molar flow rates of vapor and liquid and m0 is the slope of the
y–x equilibrium curve (Fig. 1b) in the concentration range of interest. The
656 DISTILLATION Vol. 1
required total height of the packed section is then obtained from the simple rela-
tionship,
Zp ¼ ðNov Þ ðHov Þ ð48Þ
0.50 20
HETP, in.
HETP, m
0.25 10
0 0
0 0.5 1.0 1.5 2.0
FS, m/s·(kg/m3)0.5
Fig. 25. Values of HETP as a function of throughput for ( ) 50-mm metal Pall rings
and ( ) No. 2 structured packing at 12-mm crimp height. Conditions are cyclohexane/
n-heptane system, 165 kPa (24 psia) operating pressure, total reflux, 0.43-m diameter
column. Courtesy of The University of Texas at Austin.
Vol. 1 DISTILLATION 657
columns prefer the use of HETP instead of Hov, but the latter is more fundamen-
tal and discriminates between liquid- and vapor-phase resistances. It should be
noted that terms such as Hov and Nov are based on vapor-phase concentrations;
equivalent terms based on liquid concentrations could be used.
For structured packings, methods for predicting Hv and HL are some-
what more reliable, partly because the fluid mechanics are less complicated
than those for random packings (they can be modeled by analogy to wetted
wall columns) and also because they have been extensively tested under distil-
lation conditions and for a variety of test mixture properties. The earliest
mechanistic model for structured packings appeared in 1985 (79), and has
since been expanded significantly for predicting hydraulics (80) and mass
transfer efficiency (81). An alternate model of Olujic (82) was developed
separately, and the two models have been compared (83). For these models
the key design parameters are: corrugation height, nature of metal surface,
angle of corrugation from the horizontal (45 vs 60 deg) in addition to the
usual physical property and flow rate considerations. Methods for predicting
pressure drop and flooding in beds of structured and random packings have
been reviewed (84).
4.3. Packed vs Plate Columns. Relative to plate towers, packed
towers are more useful for multipurpose distillations, usually in small
(under 0.5 m) towers or for the following specific applications: severe corro-
sion environment where some corrosion-resistant materials, such as plas-
tics, ceramics, and certain metallics, can easily be fabricated into packing
but may be difficult to fabricate into plates; vacuum operation where a low
pressure drop per theoretical plate is a critical requirement; high (eg, above
49,000 kg/(h m2 ) [10; 000 lb/(h ft2 )] liquid rates; foaming systems; or
debottlenecking plate towers having plate spacings that are relatively close,
under 0.3 m.
Plate columns have the advantage of lower fabrication cost, less depen-
dence on good liquid and gas distribution, and protection against vapor bypass-
ing the liquid in critical zones, eg, regions of extremely low impurities.
Further, methods for the design on plate columns are somewhat more reliable
than those for many of the packings, especially those packings of a proprietary
nature.
There are notable cases where plate columns have been converted to
packed columns to gain advantage of the low pressure drop exacted from the
vapor stream. More recently the packings have been largely of the structured
type. Illustrative of this is the trend toward the use of structured packing in
ethylbenzene –styrene fractionators, some of which have diameters of 10 m
or higher.
5. Steam Distillation
6. Molecular Distillation
Molecular distillation occurs where the vapor path is unobstructed and the
condenser is separated from the evaporator by a distance less than the
mean-free path of the evaporating molecules (86). This specialized branch
of distillation is carried out at extremely low pressures ranging from 13–
130 mPa (0.1 – 1.0 mm Hg). Molecular distillation is confined to applications
where it is necessary to minimize component degradation by distilling at
the lowest possible temperatures. Commercial usage includes the distillation
ofvitamins (qv) and fatty acid dimers (see DIMER ACIDS).
6.1. Distillation as a Separation Method. Distillation is the most
important industrial method of separation and purification of liquid compo-
nents. Liquid separation methods in less common use include liquid–liquid
extraction, membrane diffusion, ion exchange, and adsorption. However, distilla-
tion does not require a mass-separating agent such as a solvent, adsorbent, or
membrane, and distillation utilizes energy in a convenient heating medium
(often steam). Also, a wealth of experience with design and operations makes dis-
tillation column performance prediction more reliable than equivalent predic-
tions for other methods. At times distillation also competes indirectly with
methods involving solid–liquid separations such as crystallization. An
Vol. 1 DISTILLATION 659
Minimum no. of
Relative volatility Minimum reflux ratio theoretical plates
1.1 20 130
1.5 4.0 31
4 0.66 9
In the example, the minimum reflux ratio and minimum number of theore-
tical plates decreased 14- to 33-fold, respectively, when the relative volatility
increased from 1.1 to 4. Other distillation systems would have different specific
reflux ratios and numbers of theoretical plates, but the trend would be the same.
As the relative volatility approaches unity, distillation separations rapidly
become more costly in terms of both capital and operating costs. The relative
volatility can sometimes be improved through the use of an extraneous solvent
that modifies the VLE. Binary azeotropic systems are impossible to separate into
pure components in a single column, but the azeotrope can often be broken by an
extraneous entrainer (see DISTILLATION, AZEOTROPIC AND EXTRACTIVE).
6.3. Feed Composition. Feed composition has a substantial effect
on the economics of a distillation. Distillations tend to become uneconomical
as the feed becomes dilute. There are two types of dilute feed cases, one in
which the valuable recovered component is a low boiler and the second when
it is a high boiler. When the recovered component is the low boiler, the absolute
distillate rate is low but the reflux ratio and the number of plates is high.
An example is the recovery of methanol from a dilute solution in water.
When the valuable recovered component is a high boiler, the distillate rate,
the reflux relative to the high boiler, and the number of plates all are high.
660 DISTILLATION Vol. 1
An example for this case is the recovery of acetic acid from a dilute solution in
water. For the general case of dilute feeds, alternative recovery methods are
usually more economical than distillation.
6.4. Product Purity. Product purity requirements influence choice of
separation methods. For favorable equilibria, distillation energy requirements
do not increase significantly as purity specifications become tighter. For example,
in an ideal binary distillation of 60 mol % of A in the feed, the minimum and oper-
ating reflux ratios would be essentially the same whether the required purity of A
was 99 or 99.9999%. The number of plates would increase substantially, however,
as the purity requirements became more stringent. The shortcut methods of calcu-
lating minimum reflux ratio, minimum number of plates, operating reflux ratio,
and number of operating plates allow a rapid evaluation of the effect of changes
in purity requirements on the key economic factors in distillation.
6.5. Operating Pressure. The absolute pressure of the distillation may
have substantial economic impact. The temperature at which heat is supplied to
the reboiler and removed from the condenser determines the unit cost of the
energy. The cost of removing heat in the condenser increases rapidly as the con-
densing temperature drops below the range of air or water cooling capability; eg,
the cost of removing a unit quantity of heat at 25 C may be one hundred times
as high as removing it at 1008C. Similarly, the cost of the energy required for the
reboiler increases rapidly as the boiler temperature increases above some level
determined by local conditions. For example, at a particular site low-pressure
waste steam at 1108C may be essentially without cost, but if a temperature level
of 2008C is required, the unit cost of the heat is much higher. The relative
cost of the heat being removed and supplied is the controlling factor determining
the design of some distillations. The use of multiple interstage reboilers and con-
densers at different energy levels, as well as the use of other operational modes
used to optimize the overall economics, has been discussed (88).
The absolute pressure may have a significant effect on the vapor–liquid
equilibrium. Generally, the lower the absolute pressure, the more favorable
the equilibrium. This effect has been discussed for the styrene–ethylbenzene
system (29). In a given column, increasing the pressure can increase the column
capacity by increasing the capacity parameter (see eqs. 42 and 43). Selection of
the economic pressure can be facilitated by guidelines (89) that take into consid-
eration the pressure effects on capacity and relative volatility. Low pressures are
required for distillation involving heat-sensitive material.
6.6. Heat Sensitivity. The heat sensitivity or polymerization tendencies
of the materials being distilled influence the economics of distillation. Many
materials cannot be distilled at their atmospheric boiling points because of
high thermal degradation, polymerization, or other unfavorable reaction effects
that are functions of temperature. These systems are distilled under vacuum in
order to lower operating temperatures. For such systems, the pressure drop per
theoretical stage is frequently the controlling factor in contactor selection. An
excellent discussion of equipment requirements and characteristics of vacuum
distillation may be found in Ref. 90.
6.7. Corrosivity. Corrosivity is an important factor in the economics of
distillation. Corrosion rates increase rapidly with temperature, and in distilla-
tion the separation is made at boiling temperatures. The boiling temperatures
Vol. 1 DISTILLATION 661
7. Equipment Costs
8. Column Control
8.1. Nomenclature
Symbol Definition Units
A12, A21 constants in the Van Laar activity coefficient equation
B bottoms from column mol/s
Csb vapor capacity parameter m/s
D distillate from column mol/s
Eo overall column plate efficiency (eq. 44) fractional
Emv Murphree plate efficiency (eq. 45) fractional
Eog local, or point, efficiency based on vapor concentrations fractional
f fugacity kPa
F feed mol/s
F* F-factor (eq. 43) m=s ðkg=m3 Þ0:5
FA* F-factor based on active (bubbling) area m=s ðkg=m3 Þ0:5
FH* F-factor based on hole area m=s ðkg=m3 Þ0:5
Fp packing factor from Table 2 1/m
G gas mass rate kg/s
H; enthalpy per mole
H enthalpy per unit time
H* Henry’s law constant (eq. 17) kPa/mol fraction
HETP height equivalent of theoretical plate m
HL height of a liquid-phase transfer unit m
Hov height of an overall transfer unit, vapor concentrations m
Hv height of a vapor-phase transfer unit m
K y*/x, vapor–liquid equilibrium ratio (eq. 1)
L liquid rate mol/s
L average liquid rate for section mol/s
L; liquid mass rate kg/s
m an equilibrium stage below the feed
m0 slope of equilibrium line
n an equilibrium stage above the feed
N number of stages
N* number of components
Na number of actual stages
Nov number of transfer units, vapor concentration basis
Nt number of theoretical stages
P total pressure of system kPa
p partial pressure kPa
P0 vapor pressure kPa
q heat to vaporize 1 mol feed divided by molal latent
heat of feed (eq. 36)
q; heat removed or added at column auxiliaries
R reflux mol/s
R0 gas law constant
T temperature K
v vapor molar volume m3/mol
V vapor molar rate mol/s
V; average molar vapor rate for section mol/s
V* vapor velocity m/s
x mole fraction in liquid
y mole fraction in vapor
y* mole fraction vapor in equilibrium with x
z compressibility factor in gas law
Zp height of packed bed m
a relative volatility (eq. 2)
Vol. 1 DISTILLATION 663
BIBLIOGRAPHY
CITED PUBLICATIONS
32. R. Savarit, Arts Metiers 65, 145, 178, 266, 307 (1922).
33. M. R. Fenske, Ind. Eng. Chem. 24, 482 (1932).
34. A. J. V. Underwood, Chem. Eng. Progr. 44, 603 (1948).
35. J. R. Fair and W. L. Bolles, Chem. Eng. 75(9), 156 (Apr. 22, 1968).
36. G. G. Brown and co-workers, Unit Operations, John Wiley & Sons, Inc., New York,
1950.
37. E. R. Gilliland, Ind. Eng. Chem. 32, 918 (1940).
38. J. H. Erbar and R. N. Maddox, Petrol. Ref. 40(5), 183 (1961).
39. H. E. Eduljee, Hydrocarbon Proc. 54(9), 120 (1975).
40. E. W. Thiele and R. L. Geddes, Ind. Eng. Chem. 25, 290 (1933).
41. C. D. Holland, Fundamentals of Multicomponent Distillation, McGraw-Hill Book Co.,
Inc., New York, 1981.
42. R. N. S. Rathore, K. A. Van Wormer, and G. J. Powers, AIChE J. 20, 491 (1974).
43. D. C. Freshwater and B. D. Henry, Chem. Eng. (London) (301), 533 (1975).
44. J. E. Hendry, D. F. Rudd, and J. D. Seader, AIChE J. 19, 1 (1973).
45. W. L. Bolles, in B. D. Smith, ed., Design of Equilibrium Stage Processes, McGraw-Hill
Book Co., Inc., New York, 1963, Chapt. 14.
46. M. H. Hutchinson and R. F. Baddour, Chem. Eng. Progr. 52(12), 503 (1956).
47. F. Kastanek, M. V. Huml, and V. Braun, I. Chem. E. Symp. Ser. No. 32, 5(100),
(1969).
48. W. V. Delnicki and J. L. Wagner, Chem. Eng. Progr. 52(1), 28 (1956).
49. J. R. Fair, Chapt. 5 in Handbook of Separation Process Technology, R. W. Rousseau,
ed., John Wiley, New York, 1987.
50. J. R. Fair, in R. H. Perry and D. Green, eds., Perry’s Chemical Engineers’ Handbook,
7th ed., McGraw-Hill Book Co., Inc., New York, 1997, section 14.
51. Bubble-Tray Design Manual, American Institute of Chemical Engineers (AIChE),
New York, 1958.
52. H. Chan and J. R. Fair, Ind. Eng. Chem. Proc. Des. Devel. 23, 814, 820 (1984).
53. M. J. Lockett, Distillation Tray Fundamentals, Cambridge University Press,
Cambridge, Mass., 1986.
54. M. M. Dribika and M. W. Biddulph, Trans. I. Chem. E. 70, Part A, 142 (1992).
55. M. Prado and J. R. Fair, Ind. Eng. Chem. Res. 29, 1031 (1990).
56. K. E. Porter, M. J. Lockett and C. T. LimTrans. I. Chem. E. 50, 91 (1972).
57. W. V. Pincezewski, N. D. Benke, and C. J. D. Fell, AIChE J. 21, 1210 (1975).
58. K. E. Porter, A. Safekouri, and M. J. Lockett, Trans. I. Chem. E. 51, 265 (1973).
59. M. Prado, K. L. Johnson, and J. R. Fair, Chem. Eng. Progr. 83(3), 32 (1987).
60. J. A. Garcia and J. R. Fair, Ind. Eng. Chem. Res. 39, 1809, 1818 (2000).
61. H. E. O’Connell, Trans. AIChE 42, 741 (1946).
62. R. Krishna, H. F. MartinezR. Sreedhar, and G. L. Standart, Trans. I. Chem. E. 55,
178 (1977).
63. R. Taylor and R. Krishna Multicomponent Mass Transfer, John Wiley, New York,
1993.
64. A. Sperandio, M. Richard, and M. Huber, Chem.-Ing.-Tech. 37, 22 (1965).
65. W. D. Stoecker and B. Weinstein, Chem. Eng. Progr. 73(11), 71 (1977).
66. R. F. Strigle, Packed Tower Design and Applications, Gulf Publishing, Houston, Tex.,
1994.
67. J. R. Fair and J. L. Bravo, Chem. Eng. Progr. 86(1), 19 (1990).
68. P. J. Hoek, J. A. Wesselingh, and F. J. Zuiderweg, Chem. Eng. Res. Des. 64, 431
(1986).
69. R. E. Treybal, Mass Transfer Operations, 3rd ed., McGraw-Hill., Inc., New York,
1980.
70. H. Z. Kister, Distilation—Design, McGraw-Hill, Inc., New York, 1992.
666 DISTILLATION Vol. 1
JAMES R. FAIR
The University of Texas at Austin
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 667
DISTILLATION,
AZEOTROPIC AND
EXTRACTIVE
1. Introduction
Distillation (qv) is the most widely used separation technique in the chemical
and petroleum industries. Not all liquid mixtures are amenable to ordinary frac-
tional distillation, however. Close-boiling and low relative volatility mixtures are
difficult and often uneconomical to distill, and azeotropic mixtures are impossible
to separate by ordinary distillation. Yet such mixtures are quite common (1,2)
and many industrial processes depend on efficient methods for their separation
(see also SEPARATIONS PROCESS SYNTHESIS). This article describes several special
distillation techniques for economically separating low relative volatility and
azeotropic mixtures.
Vapor–liquid phase equilibrium (VLE) in a c-component mixture can be
represented by
Z P
ViL dP
yi fVi P ¼ xi gi Psat
i fi
sat
exp for i ¼ 1; :::; c ð1Þ
Psat
i
RT
where yi is the mole fraction of component i in the vapor phase; fiV is the vapor-
phase fugacity coefficient of component i; P is the total system pressure; xi is the
mole fraction of component i in the liquid phase; gi is the liquid-phase activity
coefficient of component i; Pisat is the vapor pressure of component i; fisat is
the saturation fugacity coefficient of component i; and the exponential term is
the Poynting correction factor (Poyi). At low to moderate pressures, the value
of the term fisat/fi*Poyi is typically close enough to unity that it can be ignored
and equation 1 reduces to
yi P ¼ xi gi Psat
i for i ¼ 1; :::; c ð2Þ
For simplicity, this is the VLE equation that will be used throughout the rest of
this article.
The activity coefficient (gi) is a measure of the liquid-phase nonideality of a
mixture and its value varies with both temperature and composition. When gi ¼ 1
the liquid phase is said to form an ideal solution and equation 2 reduces to
Raoult’s law. Nonideal mixtures (gi 6¼ 1) can exhibit either positive (gi > 1) or
negative (gi < 1) deviations from Raoult’s law. Positive deviations are more com-
mon and occur when the molecules of the different compounds in the solution are
dissimilar and have no preferential interactions between them. Negative devia-
tions occur when there are preferential attractive forces (hydrogen bonds, etc.)
between the molecules of the different species that do not occur in the absence
of the other species (3). If these deviations are large enough, the pressure-
composition (P-x,y) and temperature-composition (T-x,y) phase diagrams exhibit
a minimum or maximum point (Fig. 1). At these minima and maxima the liquid
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0409192004150805.a01.pub2
668 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Fig. 1. P-x,y and T-x,y phase diagrams showing binary, homogeneous, maximum, and
minimum azeotropes. (a) Chloroform (1)-tetrahydrofuran (2) at 308C; (b) chloroform (1)-
tetrahydrofuran (2) at 101 kPa; (c) ethanol (1)-toluene (2) at 658C; and (d) ethanol (1)-to-
luene (2) at 101 kPa (from Ref. 6). To convert kPa to atm, multiply by 9.87
10 3.
phase and its equilibrium vapor phase have the same composition, ie,
the mixture boils at constant temperature, and the dew-point (vapor) and bubble-
point (liquid) curves are tangent with zero slope. These are the defining condi-
tions for a homogeneous azeotrope where a single liquid phase is in equilibrium
with a vapor phase. Note that a maximum boiling azeotrope (Fig. 1b) is equi-
valent to a minimum pressure azeotrope (Fig. 1a) and a minimum boiling azeo-
trope (Fig. 1d) is also a maximum pressure azeotrope (Fig. 1c). Minimum boiling
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 669
azeotropes occur in mixtures with positive deviations from Raoult’s law and max-
imum boiling azeotropes occur in mixtures with negative deviations. The vast
majority of the known azeotropes are minimum boiling (see 1,2). Combining
equations 2 and 3 gives an alternative criterion for a homogeneous azeotrope,
in this case, a binary azeotrope:
g2 Psat
¼ 1 ð4Þ
g1 Psat
2
Equation 4 illustrates that, (1) only small deviations from Raoult’s law are neces-
sary for an azeotrope to exist in a close boiling mixture (ie, compounds with simi-
lar vapor pressures) and (2) the larger the difference in boiling points of the
compounds, the more nonideal the mixture must be to form an azeotrope and,
hence, the less likely that an azeotrope will exist. Generally, mixtures of com-
pounds with boiling points more than 308C apart do not exhibit azeotropes
(4). Exceptions to this heuristic do exist, however. For example, hydrogen chlor-
ide and water form a maximum boiling azeotrope and their normal boiling points
differ by 1858C. Martin (5) showed that a binary mixture forms a minimum boil-
ing azeotrope when the infinite dilution activity coefficient of the less volatile
component is greater than the ratio of the pure component vapor pressures, ie,
Psat
g1
2 >
1
ð5Þ
Psat
2
Partially miscible mixtures are more nonideal than completely miscible mixtures
and thus are more likely to form azeotropes. Usually, these will be heterogeneous
azeotropes where two (or more) liquid phases are in equilibrium with a vapor
phase, the mixture boils at constant temperature and at constant composition,
but the composition of the equilibrium phases are all different (see Fig. 17b).
The overall liquid-phase composition is, however, identical to the vapor-phase
composition. Since positive deviations from Raoult’s law are necessary for liquid
phase immiscibility, only minimum boiling heterogeneous azeotropes can exist.
When the immiscibility exists only over a limited composition range, the azeo-
trope can lie outside of the two-liquid-phase regions, giving rise to a homoge-
neous azeotrope. The methyl acetate–water and methyl ethyl ketone–water
mixtures are two such examples.
Separation by distillation depends on the vapor and liquid phases having
different compositions when a liquid (vapor) mixture is partially vaporized (con-
densed). The vapor phase becomes enriched in the more volatile components and
depleted in the less volatile components while the opposite occurs in the equili-
brium liquid phase. By successively repeating these partial vaporizations and
condensations in a countercurrent cascade, it is often possible to achieve the
desired degree of separation. A common measure of the degree of enrichment
or the ease of separation is the relative volatility defined as:
yi xj gi Psat
i
aij ¼ ¼ ð6Þ
xi yj gj Psat
j
670 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
The relative volatility of most mixtures changes with temperature, pressure, and
composition. The larger the value of aij, the easier it is to separate component i
from component j. From equation 3, at a c-component (ie, binary, ternary, etc.)
homogeneous azeotrope, yi ¼ xi for all c components in the mixture. Therefore,
aij ¼ 1 for all components i and j and no enrichment of the vapor takes place dur-
ing a partial vaporization. Thus, homogeneous azeotropes cannot be separated by
ordinary fractional distillation. Similarly, any mixture, be it ideal, nonideal,
close-boiling, etc. where the relative volatilities are close to unity will be difficult
to separate by ordinary distillation because little enrichment occurs with each
partial vaporization step. Although the actual minimum relative volatility neces-
sary for an economical separation is a function of a number of variables (most
notably, the required product purities, but also the value of the products and
the cost of the materials of construction and the heating and cooling mediums,
etc.), alternatives to ordinary fractional distillation should generally start to be
investigated when aij is less than 1.15–1.2.
Most methods for distilling azeotropic and low relative volatility mixtures
rely on the addition of specially chosen chemicals or ‘‘mass separating agents’’
to facilitate the separation. These separating agents can be divided into distinct
classes that define the principal distillation techniques used to separate such
mixtures. The five categories of distillation-based methods for separating low
relative volatility and/or azeotropic mixtures are (1) extractive distillation and
homogeneous azeotropic distillation where the liquid separating agent is comple-
tely miscible. In extractive distillation, the separating agents, variously known
as solvents, extractive agents, extractants, and sometimes entrainers, are high
boiling, non-azeotrope-forming, liquids that alter the relative volatility of the
mixture to be separated. (2) Heterogeneous azeotropic distillation or, more com-
monly, azeotropic distillation, where the liquid separating agent, called the
entrainer, forms one or more azeotropes with the other components in the mix-
ture and causes two liquid phases to exist over a broad range of compositions.
This immiscibility is the key to making the distillation sequence work. (3) Salt-
effect distillation, which is a variation of extractive distillation, where the separ-
ating agent is an ionic salt. The salt dissociates in the liquid mixture and alters
the relative volatilities to make the separation possible. (4) Pressure-swing or
pressure-sensitive distillation where a series of distillation columns operating
at different pressures is used to separate mixtures containing azeotropes
whose composition changes appreciably over a moderate pressure range.
(5) Reactive distillation where the separating agent reacts preferentially and
reversibly with one of the components in the mixture. The reaction product is
then distilled from the nonreacting components and the reaction is reversed to
recover the initial component. An example is the separation of isobutene from
butanes by etherification (7). This last method is not discussed in this article. All
five of these techniques can be used in batch as well as continuous distillations.
Distillation can also be combined with other separation methods such
as liquid – liquid extraction, absorption, adsorption/molecular sieves, melt
crystallization (qv), or membrane-permeation-based methods like pervapora-
tion (8 – 10) to separate azeotropic mixtures. Ordinary fractional distillation
is used for the separation over the composition range where the relative
volatility is appreciable and the separation is easy. One of the other
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 671
Intermediate
x2
Heavy x1 Light
Many different residue curve maps are possible when azeotropes are pre-
sent (see 13–15). For example, there are six possible residue curve maps for tern-
ary mixtures containing a single azeotrope. These six maps differ by the binary
pair forming the azeotrope and by whether the azeotrope is minimum or maxi-
mum boiling. Figure 3 represents the case where the intermediate and heaviest
components (A and B) form a minimum boiling binary azeotrope. Pure compo-
nent D is an unstable node, pure components A and B are stable nodes, the mini-
mum boiling binary azeotrope C is a saddle, and the boiling point order from low
x2
B C A
x1
Fig. 3. Residue curve map for a ternary mixture with a distillation boundary running
from pure component D to the binary azeotrope C.
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 673
dxi
¼ xi y i for i ¼ 1; :::; c ð7Þ
dx
where xi and yi are the liquid and vapor mole fractions, respectively, of compo-
nent i, x is a nonlinear time scale that is related to the fraction of liquid remain-
ing in the pot at any instant in time, and c is the number of components in the
mixture. Note that the pure components and azeotropes in the mixture corre-
spond to the steady-states or singular points of equation 7 (where the derivative
goes to zero) and they are all either nodes or saddles. A detailed mathematical
treatment of simple distillation residue curve maps and their properties can be
found in the literature (16,17). Given a method for calculating the VLE for the
mixture of interest, equation 7 can be numerically integrated forward and back-
ward in time from a number of different initial compositions to generate a resi-
due curve map. An alternative method for sketching residue curve maps that
only requires knowledge of the boiling points of the pure components and azeo-
tropes in the mixture has also been published (18–21). This latter method is well
suited to mixtures lacking detailed VLE data and thus can be useful in prelimin-
ary separation feasibility studies.
Even though the simple distillation process is not a practical method
for separating mixtures, simple distillation residue curve maps are extremely
useful. They can be used to test the consistency of experimental azeotropic
data (18,19,22); to predict the sometimes nonintuitive order and content of the
cuts from batch distillations of azeotropic mixtures (23–26); to determine
674 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
As mentioned above, simple distillation boundaries are the same as the dis-
tillation boundaries of a distillation column operated at total reflux. While the
two are not equivalent at finite reflux ratios, the simple distillation boundaries
remain a very good approximation to the finite-reflux distillation boundaries,
except in regions where the boundaries are extremely curved. Although the
composition profile of a distillation column operating at finite reflux can cross
a curved simple distillation (infinite reflux) boundary to a limited degree (28–
33,44,45) the resulting distillation sequences typically require large reflux(re-
boil) ratios and/or large recycle flowrates and, thus, are not normally economical.
Mixtures such as nitric acid–water–sulfuric acid, that have extremely curved
boundaries, can be exceptions. A good rule-of-thumb, therefore, is that the com-
positions of the distillate and bottom streams from a distillation column should
lie in the same simple distillation region in order for the separation to be feasible.
Crossing of simple distillation boundaries should only be attempted as a last
resort and only with caution, because the feasibility of such column designs
can be sensitive to small inaccuracies in the VLE model and, once built, the
column may be sensitive to disturbances.
An overall material balance for a continuously operated distillation column
requires that the feed, distillate, and bottoms compositions lie on a straight line
in the composition space (the composition triangle for ternary mixtures). Thus, a
feasible distillation sequence for separating a homogeneous azeotropic mixture
can be identified by determining whether or not the desired products lie in the
same distillation region and then can be synthesized by superimposing material
balance lines onto the distillation residue curve map. When setting the composi-
tions of the distillate and bottom streams leaving each column in the sequence:
(1) the distillate composition must have a lower boiling temperature than the
bottoms composition, however the component with the lowest (highest) boiling
point is not necessarily removed as the distillate (bottoms); (2) pure components
and azeotropes that are nodes on the residue curve map are easier to obtain
as pure products than saddles; and (3) a double-feed column is almost always
required in order for a saddle on the residue curve map to be the product from
a distillation column, eg, extractive distillations. For simplicity, only ‘‘sharp’’
splits, where a pure component or azeotrope is the product from at least one
end of the distillation column, are considered here. (This excludes situations
where a nonsharp or sloppy split is acceptable, such as when a small amount
of impurities are removed in a small distillate or bottoms stream. The complete
range of product compositions that can be reached in a single column for a given
feed can be determined by calculating the ‘‘bow-tie’’-shaped separation region
(4,21,29,31–34,44)).
As an example, consider the residue curve map for the nonazeotropic mix-
ture shown in Fig. 2. It has no distillation boundary so the mixture can be sepa-
rated into pure components by either the direct or indirect sequence (Fig. 4). In
the direct sequence, the unstable node (light component, L) is taken overhead in
the first column. The bottom stream composition is determined by following the
material balance line (the straight line through the distillate and feed composi-
tion) until a distillation boundary or the edge of the composition space (the com-
position triangle here) is reached. Thus, the bottom stream from the first column
is a binary mixture of the intermediate, I, and heavy, H, components (plus a trace
676 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Fig. 4. Column sequences and material balance lines for the (a) direct and (b) indirect
sequences for separating nonazeotropic mixtures.
amount of the light component). In the binary I–H mixture, I has the lowest boil-
ing temperature (unstable node) so it is recovered as the distillate in the second
column and the stable node, H, becomes the corresponding bottoms stream. The
indirect sequence removes the stable node (heavy component, H) from the bottom
of the first column. The composition of the distillate stream is determined by fol-
lowing the material balance line through the bottom and feed compositions until
a distillation boundary or the edge of the composition space is reached. Thus, the
overhead stream from the first column is a binary mixture of L and I (plus a trace
amount of H). In the second column, the unstable node, L, is taken overhead
and component I, the stable node in the binary mixture, is recovered as the
bottoms. The relative flowrate of any of the streams can be determined by the
well-known lever-arm rule. For example, in Fig. 4b, the flowrate of stream B1
is represented by the length of the line segment from F1 to D1 and the flowrate
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 677
azeotrope is an unstable node so it becomes the distillate and the stable node,
water, is removed from the bottom of the second column. If some up-stream mod-
ification were made which reduced the amount of water in the feed to the first
distillation column enough to move the feed composition into the other distilla-
tion region, the desired separation would no longer be possible. Acetone could
still be recovered as the distillate in the first column (assuming there are suffi-
cient stages and reflux), but it would no longer be possible to get pure water from
the base of the second column. Instead, 2-propanol would be come out of the bot-
tom of the second column. The binary azeotrope would continue to be distilled
overhead in the second column. This means there would no longer be a way to
remove pure water from the process.
The overwhelming majority of all ternary mixtures that can potentially
exist are represented by only 125 different residue curve maps (15, see also 13,
14). (The rare mixtures in which a pair of compounds form two binary azeotropes
(47–51) are exceptions.) For each type of separation objective, these 125 maps
can be subdivided into those that can potentially meet the objective (ie, the resi-
due curve maps where the desired pure component and/or azeotropic products lie
in the same distillation region) and those that cannot. Knowing the structure of
the residue curve map for a mixture is thus sufficient to determine if a given
separation objective is feasible. (There is, however, no guarantee that a feasible
separation will be economical).
The seven most favorable residue curve maps for the common task of dis-
tilling a mixture containing a binary minimum-boiling azeotrope into its two
constituent pure components by adding a suitable separating agent are
shown in Fig. 6. (Favorable residue curve maps for separating binary maxi-
mum-boiling azeotropes are given in Ref. 4). Other maps that exploit distilla-
tion boundary curvature, namely those with a low boiling separating agent
that forms a highly curved boundary ending at the azeotrope, have the poten-
tial to meet the stated objective (28 –33). However, as described earlier, distil-
lation sequences based on such maps have a number of potentially serious
drawbacks. Thus, for initial screening purposes, only compounds that yield
one of these seven maps need to be considered as possible separating agents
for separating binary minimum-boiling azeotropes. Five of the candidate
maps require the separating agent to form a maximum-boiling azeotrope
with one of the other components. Since maximum-boiling azeotropes are
much less common than minimum-boiling azeotropes, these five maps will
also be relatively rare. Two examples are adding formic acid to valeric acid–
water mixtures (4) and adding hydrazine to water–aniline mixtures (52).
The map shown in Fig. 6a requires the separating agent to have a boiling
point between the other two compounds, but not form any azeotropes. Such
separating agents will also be fairly uncommon. Because the likelihood that
two compounds will form an azeotrope increases the closer their boiling points
are and because the compounds of the original mixture already boil close
enough to form an azeotrope, many potential intermediate-boiling separating
agents will also form azeotropes with one or both of the original compounds.
Benzene is an example of a feasible intermediate- boiling separating agent for
separating acetone–heptane mixtures (29,30,32). Thus, of the seven favorable
residue curve maps for separating binary minimum-boiling azeotropes into
A A
A/B
azeo
B E
E B
(a)
A A B
A/B
azeo
E
B E
(b)
A A A
E B E B E B
(c)
E E A
A/B
azeo
B
A B A B
A/E azeo
(d)
Fig. 6. The seven most favorable residue curve maps and corresponding column se-
quences for the homogeneous azeotropic distillation of a minimum-boiling binary azeo-
trope using a mass separating agent E to recover the two constituent pure components
A and B (18). The symbol & represents an azeotrope. (a) Case I, where the separating
agent is intermediate boiling and does not introduce any new azeotropes: (b) Case II, ex-
tractive distillation with a heavy solvent that does not introduce any new azeotropes. In
some cases, B will come off the top of the first column instead of A; (c) Case III, where the
separating agent is intermediate boiling and forms a maximum boiling azeotrope with the
lighter of the two pure components (i.e., A). The agent may or may not form a minimum
boiling azeotrope with B, and there may or may not be a minimum boiling ternary azeo-
trope lean in A; and (d) the same column configuration as case III, but the separating
agent is lower boiling.
679
680 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Benzene
(80.1°C)
B1
F1
(77.3°C)
F2
(53.1°C)
(80.8°C) D1 (56.1°C)
Cyclohexane Acetone
(a)
H2O
CH H2O makeup
CH = cyclohexane solvent
Liquid-liquid
extraction
D1
Homogeneous
azeotropic CH-acetone
distillation azeotrope
F2
Feed F1
CH Acetone Recycle
Benzene Acetone
Makeup Distillation
Acetone
Recycle
B1 H2O
Benzene
(b)
Fig. 7. Hybrid process for separating the benzene-cyclohexane azeotrope using acetone
as the separating agent. (a) The residue curve map for the acetone–benzene–cyclohexane
mixture at 101 kPa with superimposed feed (----) and column ( ) material balance lines
for the homogeneous azeotropic distillation step. Azeotropes are denoted by &. (b) The
complete separation sequence (53).
(A wet methyl acetate stream is acceptable here because the methyl acetate is
sent to a hydrolysis reactor where more water is added). A mixture of metha-
nol and water leaves the bottom of the first column and becomes the feed to the
second column. Methanol is recovered as the distillate in the second column.
682 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Methanol
(64.5°C)
(53.5°C)
Fig. 8. Residue curve map for the methyl acetate–methanol–water mixture at 101 kPa,
where & represents an azeotrope. The immiscible region is not shown.
Water leaves the bottom of the second column and is recycled to the first
column, completing the sequence. This process and others with separating
agents that give rise to similarly structured four-sided distillation regions
are thus a variation of extractive distillation where only one of the azeotro-
pic-constituents is recovered as a pure product. Another example is using
methyl ethyl ketone (MEK) to separate the acetone–methanol azeotrope
(21,55). Methanol and MEK form a minimum boiling azeotrope and the tern-
ary mixture contains a distillation boundary running between the two azeo-
tropes, which puts acetone and methanol into different distillation regions,
ie, the residue curve map is analogous to Fig. 8. This extractive distillation
sequence distills the methanol–MEK azeotrope overhead in the first (double-
feed) column while acetone is recovered from the top of the second column.
Successful separating agents with residue curve maps similar to the one
shown in Fig. 8 will have the same attributes as those described later for
extractive distillation solvents, except for the prohibition against azeotrope
formation.
4. Extractive Distillation
mixtures. For separating binary mixtures into pure components, each category
is represented by a single, unique residue curve map.
4.1. Minimum Boiling Azeotropes. Every extractive distillation that
separates a binary minimum boiling azeotropic mixture into its pure components
has a residue curve map and column sequence like the one shown in Fig. 6b.
Typical tray-by-tray composition profiles are shown in Fig. 9. Notice that the
1.00
EtOH-H2O Ethylene
0.90 Feed point glycol
feed pt.
0.80
EtOH
0.70
Liquid mole fraction
H2O
EG
0.60
0.50
0.40
0.30
0.20
0.10
0.00
0 2 4 6 8 10 12 14 16 18 20 22
Reboiler Theoretical stage number Condenser
(a)
Water
1.0
0.8
0.6
x2
0.4
FL
0.2
B FU F D
0
0 0.2 0.4 0.6 0.8 1.0
Ethylene glycol x1 Ethanol
(b)
Fig. 9. Extractive distillation column profiles for separating ethanol (EtOH) and water
(H2O) using ethylene glycol (EG) as the extractive agent. (a) Liquid compositions versus
theoretical stage number where (-- -^-- -) represents ethanol, (-- -.-- -) water, and (-- ~ --)
ethylene glycol; and (b) liquid composition profiles in mole fraction coordinates where (*)
represents the feeds, (.) is the distillate, (X) is the bottoms; and (.) and (*) are the tray
compositions. (From Example 5.4 in Ref. 21.)
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 685
Fig. 10. Pseudo-binary (solvent-free) y–x phase diagrams for determining which compo-
nent will be the distillate. (a) No solvent present; (b) and (c) with sufficient solvent pre-
sent to eliminate the pseudo-azeotrope with component 1 and component 2, respectively,
becoming the distillate (96); and (d) experimental VLE data for cyclohexane–benzene
where A, B, C, and D represent 0, 30, 50, and 90 mol% aniline, respectively (97).
45
40
Number of stages
35
30
25
20
15
0 5 10 15 20 25 30
Reflux ratio
(a)
1.00
0.99
0.98
0.97
0.96
0.95
0 5 10 15 20 25 30
Reflux ratio
(b)
Fig. 11. Ethanol–water separation using ethylene glycol as the solvent (from Example
5.4 in Ref. 21). (a) Number of theoretical stages in the extractive column as a function of
reflux ratio. (b) Distillate mole fraction of ethanol as a function of reflux ratio.
688 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
45
40
Number of stages
35
30
25
20
15
0 5 10 15 20 25 30
Feed ratio
Fig. 12. Number of theoretical stages in the extractive column as a function of the
solvent-to-feed ratio for the extractive distillation of ethanol and water using ethylene
glycol as the solvent (29).
is a better solvent and should result in a lower cost design than a solvent with a
larger minimum feed ratio. The solvent composition required to just make the
pseudo-azeotrope disappear on a pseudo-binary y–x phase diagram, ie, to
‘‘break’’ the azeotrope, can be taken as a qualitative measure of the minimum sol-
vent flow. A simple method for estimating minimum solvent flows has been pub-
lished (95) as has an exact quantitative method (96,98). As a rough rule-of-
thumb, the economically optimal feed ratio is often between two and four
times the minimum feed ratio (96,98). Many industrial extractive distillations
operate with feed ratios between 1 and 5, though we know of commercial units
operating with feed ratios as low as 0.15 and as high as 20. Figure 12 shows how
the number of stages in the extractive column varies with feed ratio for the etha-
nol–water separation shown in Fig. 9. The relationship between stages required
and feed ratio is the same for all extractive distillations of azeotropic mixtures.
The number of stages required is infinite at the minimum feed ratio, rapidly
drops at feed ratios slightly above the minimum, and then continues to decrease
at a slowing rate at higher feed ratios until, eventually, an asymptotic region is
reached where further increases in solvent flow has little or no effect on the num-
ber of stages required. Sometimes the number of stages required in the extrac-
tive column will start to increase again at sufficiently high feed ratios. Fig. 13
shows how the feed ratio influences the total annualized cost (TAC) [defined in
Ref. 103] of an extractive distillation sequence for dehydrating ethanol using
ethylene glycol. Cost diagrams for all extractive distillations have this same dis-
tinctive shape. The cost is very high near the minimum feed ratio, but then
rapidly decreases to a minimum before starting to increase again at higher
feed ratios, as the costs associated with the increasing solvent recirculation
rate (increased heating and cooling loads and larger diameters in both
columns), begin to exceed the cost savings from the reduction in the number of
stages in the extractive column.
690 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
3.90
3.80
A
3.70
3.60
TAC, $ × 106/yr
3.50
1.70
1.60
1.50
1.20
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Feed ratio
Fig. 13. Extractive distillation sequence cost as a function of the solvent-to-feed ratio for
producing anhydrous ethanol from azeotropic ethanol using ethylene glycol at reflux
ratios of D ¼ 1.15 rmin, * ¼ 1.2 rmin, and & ¼ 1.3 rmin (77). Point A represents another pub-
lished design for the same mixture (75).
Maximum-boiling
azeotrope
A
Heavy
Solvent
Fig. 14. Residue curve map for separating a maximum boiling azeotrope using a high
boiling solvent where (- - - -) represents the distillation boundary and & the azeotrope.
and paraffins, diolefins and olefins, paraffins and napthalenes, styrenes and
aromatics, and a number of other hydrocarbon mixtures (109).
The extractive distillation of any close-boiling or other low relative volati-
lity, but non-azeotropic, mixture using a high boiling solvent is represented
by the residue curve map shown in Figure 2. Although this map is different
HNO3
D1 D1
F M Air
HNO3
F
Steam
D2
M Distillation
Air H 2O
boundary
B 1—
D2
B2
H2SO4 B2 B1 H 2O H2SO4–rich
(a) (b)
Fig. 15. Extractive distillation of nitric acid, HNO3, and water, H2O, using sulfuric acid,
H2SO4, as the solvent. (a) Residue curve map and material balance lines where & repre-
sents the maximum boiling azeotrope and (b) the column sequence (11).
692 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
than the one for the extractive distillation of minimum boiling azeotropes, the
distillation sequence is identical (see Fig. 6b) and the process works for the
same reason. The solvent alters the relative volatilities of the components to
be separated via liquid-phase interactions. Depending on the nature of these
interactions, either the low boiling or the intermediate boiling pure component
will be distilled overhead in the extractive column. Pseudo-binary y–x phase dia-
grams can again be used to determine which component will distill overhead in
the extractive column. When the composition of the lower boiling component is
plotted and increasing solvent concentrations cause the pseudo-binary VLE
curve to lie above the true binary VLE curve, then this solvent accentuates the
natural volatility difference and will cause the lower boiling component to distill
overhead. If the pseudo-binary VLE curve lies below the true binary VLE curve,
then this solvent reverses the natural volatility difference of the mixture and, at
sufficiently high solvent concentrations, the intermediate boiling component will
distill overhead.
Because there is no azeotrope, these mixtures could in theory be separated
without adding a solvent, though it would be a difficult and expensive separation.
Consequently, there is no minimum feed ratio (minimum solvent flow). Curves
relating the number of stages in the extractive column to the feed ratio (eg,
Fig. 12) end at a large but finite number of stages when the solvent flow rate
goes to zero. By using the heuristics provided earlier, optimization of extractive
distillation sequences for separating non-azeotropic mixtures usually reduces to
a one-variable optimization of the feed ratio. These extractive distillations can
also exhibit inverse response at sufficiently large reflux ratios, due to dilution
of the solvent concentration in the column. (See p. 447 of Ref. 110 for additional
information.)
References 109 and 112 provide additional information on the extractive
distillation of close-boiling, non-azeotropic mixtures and the use of mixed-solvent
systems.
4.5. Solvent Selection. The most important step in developing a
successful (ie, economical) extractive distillation sequence is identifying an
effective solvent for the separation. The solvent selection procedure is typi-
cally carried out in several stages. In the first step, simple qualitative methods
are used to identify general classes of compounds or functional groups that
may make effective solvents for a given separation. In the next step, individual
compounds are evaluated, typically by experimental methods, and then
ranked to identify the most promising candidates. In the final step, detailed
vapor– liquid equilibrium measurements are made for several of the top
candidates, the separation is simulated using chemical process simulation
software and/or tested in a lab-scale or mini-works column, and then the
final solvent selection is made. The final selection process should include an
economic comparison of the optimal separation sequence with each of the
final solvent candidates, because differences in the solvent recovery costs
can influence the decision. The remaining discussion focuses on the first two
stages.
As described earlier, extractive distillation works because the presence
of the solvent alters the relative volatilities of the components being separa-
ted. Consequently, an important measure of the effectiveness of an extractive
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 693
sat sat
ðaij Þs ðgi Pi =gj Pj Þs ðgi =gj Þs
Sij ¼ ¼ ffi ð8Þ
aij gi Psat
i =gj Pj
sat gi =gj
which is simply the ratio of the relative volatilities of the key components with
and without the solvent present. Solvents with selectivities farthest from unity
are generally preferred. Since interactions with the solvent predominantly occur
in the liquid phase—affecting the ratio of the liquid-phase activity coefficients
(gi/gj)—and the ratio of the pure component vapor pressures is normally not
significantly affected by changes in the boiling temperature caused by the pre-
sence of the solvent, equation 8 is often expressed as the ratio of the key compo-
nent activity coefficients with and without the solvent present. Alternatively,
since the denominator in equation 8 is the same for all solvents, the solvent
selectivity is sometimes written as just the ratio of the key component activity
coefficients in the presence of the solvent. While it is possible to find solvents
that increase or decrease the ratio of activity coefficients, it is usually preferable
to select a solvent that accentuates the natural difference in vapor pressures
between the components to be separated. In other words, a solvent that increa-
ses gi relative to gj when Psat i > Pj
sat
(Sij > 1) is favored over one that increases
gj relative to gi (Sij < 1). In the latter case, adding small amounts of the solvent
actually makes the separation more difficult, and relatively large quantities
may be required to completely overcome the natural volatility difference and
enhance the separability of the original mixture (105). Such solvents will cause
the intermediate boiling pure component to distill overhead in the extractive col-
umn. A heuristic, therefore, is to favor solvents that cause the more volatile
(lower boiling) component to distill overhead (ie, solvents with selectivities
greater than 1.0).
To force the naturally more volatile component i overhead, the solvent
should either behave essentially ideally with component j and cause positive
deviations from Raoult’s law for component i (gj1 and gi > 1), or behave essen-
tially ideally with component i and cause negative deviations from Raoult’s law
for component j (gi 1 and gj < 1). Compounds of similar type and size, eg,
pentane–hexane or methanol–ethanol, tend to behave ideally in the liquid
phase and thus have activity coefficients close to unity. Dissimilar molecules
tend to repel each other, causing positive deviations from Raoult’s law and, in
the extreme, resulting in liquid-phase immiscibilities. Compounds that tend to
associate in the liquid phase exhibit negative deviations from Raoult’s law.
Because systems showing positive deviations are more common, the usual
approach is to force the lower boiling component overhead by selecting a solvent
which is chemically similar to the higher boiling species and dissimilar to the
lower boiling species.
Homologues, polarity, and hydrogen-bonding tendencies can be used for a
crude screening of candidate solvents by qualitatively predicting the types of
deviations from ideality that can be expected. When the compounds being sepa-
rated are chemically dissimilar, suitable solvents may be found among the higher
694 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Carlson and Stewart’s (114) guidelines for choosing potential extractive dis-
tillation solvents based on polarity arguments are given below. Be aware, how-
ever, that exceptions do occur. (1) If the mixture to be separated is highly polar
(like acetic acid and water), choose a polar solvent or a nonpolar solvent (like a
hydrocarbon) that is more soluble with the higher boiling component. (2) For
nonpolar mixtures (like cyclohexane–benzene) choose a polar solvent (like ani-
line) that is more soluble with the higher boiling component. (3) For a mixture
where the more volatile component is polar and the less volatile component is
nonpolar, choose a nonpolar or a polar solvent. Be aware that polar solvents
may reverse the relative volatility. (4) For mixtures where the more volatile com-
ponent is nonpolar and the other component is polar, choose a polar solvent.
(5) For mixtures of compounds that are moderately polar, choose a polar solvent
or a nonpolar solvent that is more soluble with the higher boiling component.
For cases where either a polar or a nonpolar solvent is recommended, polar com-
pounds are often preferred because they typically cause the solution to be more
nonideal.
Strong deviations from ideality are often associated with hydrogen bonding
between molecules and many successful extractive distillation solvents are com-
pounds capable of forming strong hydrogen bonds such as phenols, aromatic
amines like aniline, higher alcohols, glycols, etc. (115). Several authors have
published tables based on the hydrogen-bonding tendency of different classes
of compounds and functional groups for qualitatively predicting the type of
deviation likely to occur in a mixture (34,115–118). Table 2 is representative.
To distill the more volatile component overhead, use Table 2 to select classes
of compounds that cause positive deviations with the more volatile component
or negative deviations with the less volatile component.
Once several potentially promising families of compounds have been iden-
tified using the above qualitative methods, individual solvent candidates need to
be identified and ranked. As a starting point, elminate all compounds with boil-
ing points lower than the compounds to be separated. This ensures that,
throughout the column, the solvent will be predominantly present in the liquid
phase where it can alter the activity coefficients of the mixture being separated.
The likelihood that a potential solvent will form azeotropes with any of the mix-
ture components can be minimized by selecting compounds that boil 30–508C or
more above the mixture. On the other hand, the solvent should not boil so high
that excessive temperatures or high vacuum are required in the solvent recovery
column.
Group contribution methods such as modified UNIFAC (Dortmund)
(119,120) can be used to predict activity coefficients in mixtures and thus provide
a nonexperimental method for estimating solvent selectivities. Bear in mind,
however, that all group contribution methods have weaknesses and they some-
times give, not only quantitatively inaccurate, but also qualitatively incorrect
results. COSMO–RS (121,122) is a promising, though relatively new, method
for a priori predicting activity coefficients and other thermophysical properties
based on unimolecular quantum chemical calculations. It has the advantage of
often being applicable to those problems where group contribution methods
fail. Currently, however, on mixtures for which they are well-suited, group con-
tribution methods give more accurate predictions than COSMO-RS (123).
696 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Table 2. Predicted Deviations from Raoult’s Law for Different Functional Groups Based
on Hydrogen-Bonding Interactionsa
Solvent classb
Solute
Class 1 2 3 4 5 6 7 8 9 10 11 12
H-Donor Groups
1 phenol 0 0 0 þ þ
2 acid, thiol 0 0 0 0 0 0 0 þ þ
3 alcohol, water 0 þ þ 0 þ þ þ þ
4 active H on multihalogen paraffin 0 0 þ 0 0 þ
H-Acceptor Groups
5 ketone, amide with no H on N, þ 0 þ þ þ þ þ þ þ
sulfone, phosphine oxide
6 tertiary amine 0 þ 0 þ þ 0 þ 0 0
7 secondary amine 0 þ þ 0 0 0 0 0 þ
8 primary amine, ammonia, amide with 2H 0 þ þ 0 0 þ þ þ þ
on N
9 ether, oxide, sulfoxide 0 þ þ 0 0 þ 0 þ 0 þ
10 ester, aldehyde, carbonate, phosphate, 0 þ þ þ 0 þ þ 0 þ þ
nitrate, nitrite, nitrile, intramolecular
bonding, e.g.,
o-nitrophenol
11 aromatic, olefin, halogen aromatic, multi- þ þ þ 0 þ 0 0 þ 0 þ 0 0
halogen paraffin without active H,
monohalogen paraffin
Non-H-Bonding Groups
12 paraffin, carbon disulfide þ þ þ þ þ 0 þ þ þ þ 0 0
a
From Reference 115.
b
Negative deviation. þ Positive deviation. 0 No deviation.
g1
S1
ij ¼
i
ð9Þ
g1
j
In most systems, the selectivity steadily, and often almost linearly, increases
with solvent concentration so that the solvent with the largest infinite-dilution
selectivity usually also has the largest selectivity at the lower solvent concen-
trations used in extractive distillation. In other words, solvents with the largest
infinite-dilution selectivities are generally preferred. Be aware, however, that
some systems exhibit a maximum in selectivity at finite solvent concentrations
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 697
(eg, n-butane and butene in furfural) while in other systems, typically those that
approach the point of immiscibility at high solvent concentrations, the selectivity
increases rapidly (ie, nonlinearly) with solvent concentration (eg, 2,4-dimethyl
pentane–benzene–aniline) (110,124). For such mixtures, the infinite-dilution
selectivity is not an accurate indicator of the solvent’s effectiveness in an extrac-
tive distillation. Fortunately, these systems are relatively rare.
Gas–liquid chromatography (GLC) is one of the most common experimental
methods used for screening potential extractive distillation solvents and/or for
determining infinite-dilution activity coefficients (125–127). The selectivity is
determined by measuring the retention time of each solute on an inert chromato-
graphic column that has been preloaded with enough of the candidate solvent to
approximate infinite dilution conditions. Other methods for measuring infinite-
dilution activity coefficients include ebulliometry (127,128), the dilutor technique
(129), and the dew point method (130). The dilutor technique is the only one that
can determine infinite-dilution activity coefficients in solvent mixtures.
Potential extractive distillation solvents can also be evaluated by experi-
mentally measuring the selectivity or relative volatility of a fixed composition
mixture of the components to be separated (often 50% each) in the presence of
a constant amount of each candidate solvent (114). Solvent-to-feed ratios of 1:1
to 3:1 are typically used. The preferred experimental apparatus is a modified
Othmer still operated to yield one theoretical stage of separation (114,131). As
before, the objective is to find the candidate solvents that cause the largest
increase in relative volatility (ie, the highest selectivities).
While good selectivity at reasonable solvent concentrations is essential, it is
not the only important property when selecting an extractive distillation solvent.
A solvent’s capacity (ie, its ability to solubilize the components in the mixture
being separated) is also important. If a solvent causes a second liquid phase to
form during the distillation, much of the selectivity enhancement can be lost.
Since higher selectivities are by achieved by increasing the nonideality between
the solvent and one of the key components, and mixtures with large nonidealities
often have limited miscibilities, solvents with the highest selectivities can have
lower capacities. Thus, a compromise between high selectivity and high capacity
sometimes has to be made. Another alternative is to use a mixture of solvents
where one of the solvents is chosen for its high selectivity and the other is chosen
to maintain high capacity (109,112). The solvent which provides the high selec-
tivity is usually the predominant component in the solvent mixture. Because of
its limited solubility with many compounds, the reverse is true when water is
used as the highly selective solvent. For example, aromatics are separated
from aliphatics using N-methyl-2-pyrrolidone (NMP) with a small amount of
water added to increase the selectivity.
In addition to having a high selectivity at reasonable concentrations and
a high capacity, desirable extractive distillation solvents should: (1) not form
azeotropes with the components in the mixture to be separated—this can almost
be guaranteed by picking solvents with boiling points 30–508C or more above
the other compounds, (2) be easily separated from the other components at
reasonable temperatures and pressures, (3) be thermally stable at the tempe-
ratures encountered in the distillation columns, (4) be nonreactive with the
other compounds, (5) be nontoxic, (6) inexpensive, (7) readily available (perhaps
698 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
5. Salt-Effect Distillation
needed). At least two industrial plants are mentioned in which ethylene glycol–
potassium acetate mixtures are used to dehydrate ethanol or tert-butanol.
See the articles by Furter (132,133,142–144) for additional information on
salt-effect distillation.
6. Pressure-Swing Distillation
It is well known that varying the system pressure can affect the composition of
an azeotrope. When the change in azeotropic composition is appreciable over a
moderate pressure range, this effect can be exploited to separate both minimum-
and maximum-boiling binary azeotropes using a two-column sequence with the
columns operated at different pressures. This process, called pressure-swing or
pressure-sensitive distillation, is illustrated by Fig. 16. For a binary mixture
forming a pressure-sensitive, minimum-boiling azeotrope, the fresh feed, F, is
mixed with the recycled stream from the second column to form the feed stream,
F1, to the first column, which operates at pressure P1. (Alternatively, these two
streams can be fed to separate points in the first column). Because F1 lies to the
right of the azeotrope at pressure P1 (Fig. 16a), pure component A is removed as
the bottom product, B1, and a mixture near the azeotropic composition at pres-
sure P1 is the distillate, D1. Stream D1 is changed to pressure P2 and fed to the
next column as stream F2. Because F2 now lies to the left of the azeotropic com-
position at pressure P2 (Fig. 16a), the other pure component, B, can be recovered
in the bottom stream, B2, and a near azeotropic mixture becomes the distillate,
D2, which is recycled to the first column. The greater the shift in azeotropic com-
position, the smaller the recycle flow rate will be relative to the product flow
rates, and thus, the smaller the column diameters and the more economical
the process will be. An analogous procedure is used for binary maximum boiling
azeotropes (110), except the pure component products are recovered as distillates
and the near azeotropic streams come out of the bottom of each column.
B2
Pressure P1 Pressure P2
P2
D1
F2 D2 D2
T
B1 F F1 F2
P1 1 2
D1
B1 B2
Pure B F F1 Pure A
Composition Pure A Pure B
(a) (b)
Fig. 16. Pressure-swing distillation of a minimum boiling binary azeotrope. (a) Tem-
perature–composition phase diagram showing the effect of pressure on the azeotropic
composition; (b) column sequence (145).
700 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
a binary mixture of ethanol and water. This batch process was patented in 1903
(157) and later converted to a continuous process (158). Good reviews of the early
development and widespread application of continuous azeotropic distillation in
the pre-World War II chemical industry are available (159).
Historically azeotropic distillation processes were developed on an indivi-
dual basis using experimentation to guide the design. The use of residue curve
maps as a vehicle to explain the behavior of entire sequences of heterogeneous
azeotropic distillation columns, as well as the individual columns that make
up the sequence, provides a unifying framework for design. This process can
be applied rapidly, and produces an excellent starting point for detailed simula-
tions and experiments.
7.1. Phase Diagrams. For binary mixtures, it is well known that when
a liquid–liquid envelope merges with a minimum boiling vapor–liquid phase
envelope the resulting azeotropic phase diagram has the form shown in
Fig. 17. When the liquid composition (x1) is equal to x1AZ, as in Fig. 17a, then
the vapor composition, y1, is also equal to x1AZ and the mixture boils at constant
temperature and at constant (and equal) composition in each phase. Thus a
homogeneous azeotrope is formed. When the overall liquid composition
x10 ¼ (x10)AZ, as in Fig. 17b, then y1 is also equal to (x10)AZ and again the mixture
boils at constant temperature and at constant composition in each phase. How-
ever, the liquid of composition (x10)AZ splits into two liquid phases, L1 and L2, so
that there are three coexisting equilibrium phases (two liquids and one vapor)
that have different compositions. That is, a heterogeneous azeotrope is formed.
Homogeneous and heterogeneous azeotropes share the common property that
the overall liquid composition is equal to the vapor composition, which provides
a means for identifying azeotropes experimentally and computationally.
The properties of ternary heterogeneous vapor–liquid–liquid equilibrium
(VLLE) phase diagrams are important for understanding azeotropic distillation.
The simplest VLLE phase diagram, where a liquid–liquid envelope merges with
T T
V
V
T AZ — L
L
AZ
T —
L1 L2
L–L L–L
0 x1AZ 1 0 (x01)AZ 1
(a) (b)
Fig. 17. Schematic isobaric phase diagrams for binary azeotropic mixtures (AZ). (a)
Homogeneous azeotrope; (b) heterogeneous azeotrope.
702 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
T T
Vapor line
Heterogeneous
B C B C liquid boiling
surface
Projection
A A of vapor
Heterogeneous line
liquid boiling
envelope
(a) (b)
Fig. 18. Schematic isobaric phase diagrams for ternary (A,B,C) azeotropic mixtures. (a)
Homogeneous liquid phase at all boiling points; (b) heterogeneous liquid phase for some
boiling points.
Ethanol
1.00
0.80
0.60
x2
A
0.40 8
6
5 7
6
4
0.20
3 2
2
1
1
0 0.20 0.40 0.60 0.80 1.00
x1
Water Benzene
(a)
Ethanol
(78.3°C)
1.00
78.17 C
0.80
0.60
x2
6
68.10°C
0.40
5
4
0.20 3
2
1
Fig. 19. Isobaric VLLE phase diagrams for the ethanol–water–benzene system at 101.3
kPa; (&-&) represent liquid–liquid tie-lines; (~—~), the vapor line; &, homogeneous azeo-
tropes; ~, heterogeneous azeotropes; * ~ , Horsley’s azeotropes. (a) Calculated, where A is
the end point of the vapor line and the numbers correspond to boiling temperatures in 8C
of 1, 70.50; 2, 68.55; 3, 67.46; 4, 66.88; 5, 66.59; 6, 66.46; 7, 66.47, and 8, the critical point,
66.48. (b) Experimental, where the numbers correspond to boiling temperatures in 8C of 1,
67.0; 2, 65.5; 3, 65.0; 4, 64.85; 5, 64.9; and 6, the critical point, 69.9.
704 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
ternary azeotrope. Experimental VLLE data for this mixture are shown in
Fig. 19b (160). Ternary heterogeneous saddle azeotropes are also possible, eg,
formic acid–water–m-xylene (161), water–acetone–chloroform (162), however,
maximum boiling heterogeneous azeotropes cannot exist (163). This differs
from homogeneous azeotropes where all three types are found in Nature.
The typical phase equilibrium problem encountered in distillation is to cal-
culate the boiling temperature and the vapor composition in equilibrium with a
liquid phase of specified composition at a given pressure. If the liquid separates
into two liquid phases, then the problem is to calculate the boiling temperature
and the compositions of the two equilibrium liquid phases plus the coexisting
vapor phase at the specified overall liquid composition. Robust and practical
numerical methods have been devised for solving this problem (164–171).
Thus, using these techniques and a nonideal solution model that is capable
of predicting multiple liquid phases, it is possible to produce phase diagrams
comparable to those of Fig. 19. These predictions are not, however, always quan-
titatively accurate (160,172–175).
7.2. Residue Curve Maps. Residue curve maps are useful for repre-
senting the infinite reflux behavior of continuous distillation columns and for
getting quick estimates of the feasibility of carrying out a desired separation.
In a heterogeneous simple distillation process, a multicomponent partially mis-
cible liquid mixture is vaporized in a still and the vapor that is boiled off is trea-
ted as being in phase equilibrium with the coexisting liquid phases. The vapor is
withdrawn from the still as distillate. The changing liquid composition is most
conveniently described by following the trajectory (or residue curve) of the over-
all composition of all the coexisting liquid phases (176). An extensive amount of
valuable experimental data for the water–acetone–chloroform mixture, includ-
ing binary and ternary LLE, VLE, and VLLE data, and both simple distillation
and batch distillation residue curves are available (162,177). Experimentally
determined simple distillation residue curves have also been reported for the het-
erogeneous system water–formic acid–1,2-dichloroethane (178).
Using the new generation of VLLE computation techniques, it is possible to
calculate residue curve maps for heterogeneous liquid systems (Fig. 20). In this
map, the heterogeneous liquid boiling envelope has been superimposed on the
residue curves in order to distinguish the homogeneous and heterogeneous
regions of the triangular diagram.
Systems Having a Ternary Heterogeneous Azeotrope. Binary or tern-
ary heterogeneous azeotropes are restricted to being either unstable nodes or
saddles in the residue curve map (163). An example is the computed residue
curve map for the ethanol–water–benzene mixture that exhibits a ternary
minimum boiling hetero-azeotrope (see Fig. 20). This map exhibits three dis-
tillation regions. The three distillation boundaries all begin at the minimum
boiling ternary heteroazeotrope and end at each of the binary azeotropes. If
ethanol is the desired product of the separation, the initial condition for the
simple distillation must lie in the upper region of Fig. 20, which is bounded
by the ternary azeotrope, the binary azeotropes between benzene þ ethanol
and water þ ethanol, and the pure ethanol vertex. The computed phase dia-
gram is shown in Fig. 19a. Other mixtures with a minimum-boiling ternary
heteroazeotrope include, ethanol–water–toluene, ethanol–water–cyclohexane,
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 705
Ethanol
(78.30°C)
1.0
0.8
78.06°C
0.6
67.45°C
x2 A
0.4
0.2
70.50°C
0
0 0.2 0.4 0.6 0.8 1.0
x1
Water Benzene
(100°C) (80.15°C)
Fig. 20. Residue curve map calculated for the ethanol–water–benzene mixture where A
is the end point of the vapor line; & represents a homogeneous azeotrope; * ~ , heteroge-
neous azeotropes; (~—~), the vapor line; (- . -), the heterogeneous liquid boiling envelope;
*, the critical point; and &, the end points of tie-lines (179).
Fig. 21. Column sequence for separating a binary heterogeneous azeotropic mixture, A
and B, where xF represents the process feed mole fraction. (a) Phase diagram; (b) column
sequence.
V, yN
xRich
FRecycle
L
xRecycle xN x 0R
N
D,x0D
xLean
V LT
F1
xF1
V LB
B,xB
(a)
Ethanol
B,xB
F1
F, x F
II
Distillation
boundary
I xN
xRich
xLean
D, xD V, yN
L, x0R
III
Water Benzene
(b)
with a homogeneous liquid. Therefore, the liquid composition leaving the top
tray, xN, lies in the homogeneous portion of region II as shown (Fig. 22). The
azeo-column is therefore designed so that the steady-state liquid composition
profile runs from xB to xN. This can normally be done in such a way that every
stage inside the column is homogeneous, ie, has only one liquid phase. A method
for doing this is given in Ref. 181. Not all azeo-columns, however, are designed to
have a single liquid phase on all the trays. Neither do columns that have been
designed in this manner necessarily remain homogeneous under the action of
disturbances (168,172–174,182–188). For some azeotropic distillations there
can be a significant savings in energy and/or stages by designing the column
so that the entrainer penetrates down the rectifying section causing some of
the stages in the rectifying section to exhibit multiple liquid phases
(180,189,190). One such case is the acetic acid þ water þ n-butyl acetate separa-
tion discussed at the end of this article.
The overhead vapor of composition yN is totally condensed and sent to a
decanter where it splits into two equilibrium liquid phases, an entrainer-rich
phase of composition xrich and an entrainer-lean phase of composition xlean.
The relative proportion of these two liquid phases in the decanter, f, is given
by the lever rule, where f represents the molar ratio of the entrainer-rich
phase to the entrainer-lean phase in the decanter. The overall composition of
the liquids in the decanter is the same as yN. The temperature of the liquid
in the decanter is normally held constant at a value below the VLLE boiling
temperature.
The two decanter liquid phases are used to provide the reflux and distillate
streams. Normally, the reflux ratio, r, is chosen so that r ¼ L/D f. This requires
that the reflux rate be greater than the condensation rate of entrainer-rich phase
and that the distillate rate be correspondingly less than the condensation rate of
entrainer-lean phase. This means that the reflux stream consists of all the
entrainer-rich phase plus some of the entrainer-lean phase, and the distillate
stream consists of the balance of the entrainer-lean phase, ie, xD ¼ xlean. Thus,
the overall composition of the reflux stream, xR0, lies on the decanter tie-line
between the entrainer-rich phase, xrich, and the overall decanter composition,
yN, as shown in Fig. 22.
Completing the Separation Sequence. To close the material balance
around the separation sequence, the distillate stream leaving the azeotropic
column (column 2 in Fig. 23a) is separated (in column 3) into a product stream,
B3, and an entrainer recycle stream, D3. The amount of benzene needed to move
the composition of stream F1 to F is equal to the amount of benzene lost in the
aqueous distillate D2 and recovered in the stream D3.
Kubierschky Three-Column Sequence. If only simple columns are used,
ie, no side-streams, side-rectifiers/strippers etc, then the separation sequence can
be completed by adding an entrainer recovery column, column 3 in Fig. 23a, to
recycle the entrainer, and a preconcentrator column (column 1) to bring the feed
to the azeotropic column up to the composition of the binary azeotrope.
The entrainer recovery column takes the distillate stream, D2, from the
azeo-column and separates it into a bottoms stream of pure water, B3, and a
ternary distillate stream, D3, for recycle to column 2. The overall material bal-
ance line for column 3 is shown in Fig. 23b. This sequence was one of two original
710 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
Entrainer make-up
V 2, y N
D3 , xD3
D3 L2
xRich
x0R2 xLean
D1
D2
2 xD2
F1 1 3
B1 B2 B3
(a)
Ethanol
1.0
B2
D1
0.8
D 1+ D 3
0.6
D3
Column 2 xN
x Lean
0.4
D2, xD2
x Rich
Column 3
0.2 yN
0
L2, xR2
B3
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Water Benzene
(b)
Fig. 23. Separation of ethanol and water from an ethanol–water–benzene mixture. Bot-
toms B1 and B3 are water, B2 is ethanol. (a) Kubierschky three-column sequence where
columns 1, 2, and 3 represent the preconcentration, azeotropic, and entrainer recovery col-
umns, respectively. (b) Material balance lines for the azeotropic and entrainer recovery
columns where & represents the overall vapor composition from the azeo-column, V2,
yN; & the liquid in equilibrium with overhead vapor composition from the azeo-column,
xN; ~, distillate composition from entrainer recovery column xD3; and ^, overall feed com-
position to the azeo-column, D1 þ D3.
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 711
continuous processes disclosed in 1915 (191). More recently, it has been applied
to other azeotropic separations (76,192,193) .
Extensive design and optimization studies have been carried out for this
sequence (193). The principal optimization variables, ie, the design variables
that have the largest impact on the economics of the process, are the reflux
ratio in the azeo-column; the position of the tie-line for the mixture in the decan-
ter, determined by the temperature and overall composition of the mixture in the
decanter; the position of the decanter composition on the decanter tie-line [see
Ref. 181 for a discussion of the importance of these variables]; and the distillate
composition from the entrainer recovery column.
Figure 24 shows material balance lines for three different decanter tie-
lines. The process feed to the preconcentrator in each sequence is a binary
mixture of 4.2 mol% ethanol and 95.8 mol% water. The product purity from
the azeo-column (column 2) is set at 99.9 mol% ethanol, and the water purity
leaving the entrainer recovery column is set at 99.5 mol% water. These specifica-
tions are essentially identical to those used for studying the optimal extractive
distillation sequences (77,78). For design 1, the decanter tie-line is set at the bub-
ble-point of the mixture leaving the top of the azeo-column, having composition
yN. This temperature is 64.48C. For design 2, the decanter composition is the
same as for design 1 but it is subcooled to 258C. For design 3 the decanter com-
position is placed closer to the ternary azeotrope than for designs 1 and 2, and
the decanter temperature is set at the bubble-point of the mixture. In each
Ethanol
1.0
D1
Design 1
0.8 Design 2
Design 3
0.6
0.4
0.2
F1
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Water Benzene
Fig. 24. Three sets of material balance lines for the Kubierschky three-column sequence
where design 1 corresponds to the upper tie-line having rmin ¼ 8.78; design 2, to the sub-
cooled upper tie-line having rmin ¼ 12.23; and design 3, to the lower tie-line having
rmin ¼ 17.31; & represents the overall decanter composition; ^, the overall feed composi-
tion to the azeo-column; }, the distillate composition from the entrainer recovery column;
and *, the end points of the liquid–liquid tie-lines.
712 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
design, the distillate composition from the entrainer recovery column is placed
close to the distillation boundary. The position of the tie-line has a significant
influence on the distillate composition from the azeo-column, which in turn
influences the position of the material balance line for the entrainer recovery
column. The position of the distillate composition from the entrainer recovery
column also influences the overall feed composition to the azeo-column (D1 þ
D3 in Fig. 23).
Optimization studies indicate that the distillate composition from the
entrainer recovery column has a strong influence on the process economics and
the optimum position is always close to the distillation boundary. By the lever
rule, this minimizes the amount of water in the entrainer recycle stream, or
equivalently, makes the overall feed to the azeo-column richer in ethanol. For
each tie-line, the optimal position of the decanter liquid composition is found
by the method proposed in Ref. 181. The minimum reflux ratio for a given tie-
line can be reduced by as much as 50% by making small changes in the decanter
composition. Calculations indicate that the optimal reflux ratio for the azeo-col-
umn is normally in the range 1.1–1.5 rmin and that the cost of the sequence is
insensitive to this factor. This leaves the position of the decanter tie-line as
the sole remaining optimization variable.
The intra-sequence flows, compositions, and reflux ratios are quite sensi-
tive to relatively small changes in the position of the decanter tie-line as evi-
denced by the rmin value for the azeo-column: 8.78 for design 1, 12.23 for
design 2, and 17.3 for design 3. Based on knowledge of homogeneous distilla-
tions, the vapor rate and total annualized cost for sequence 1 would appear
to be the lowest. For homogeneous distillations the feed and product flow
rates, as well as their compositions, can be held constant as the reflux ratio
is changed from one design to another. Thus there is a direct relationship
between increased minimum reflux ratio and increased costs. However, such
a relationship does not occur for heterogeneous distillations so we can’t con-
clude that design 1 will have the lowest cost.
A good approximation for the vapor rate leaving the reboiler, V, for any type
of distillation is
V ¼ ðr þ 1ÞD ð10Þ
where r is the reflux ratio and D is the distillate flow rate. For homogeneous
distillations, D is constant so that V increases as r increases. For azeotropic dis-
tillation, however, both r and D change from one tie-line to another. These effects
may tend either to reinforce or to cancel each other depending on the mixture.
There is no general rule and each mixture must be treated separately. In the
ethanol-water-benzene system, the reflux ratio increases from one design to
another, but the distillate flow rate, D2, decreases, as can be seen by application
of the lever rule to the material balance lines for the azeo-column in Fig. 24.
The net effect is that the vapor rate in the azeo-column hardly changes from
one design to another. All the sequences shown in Fig. 24 have approximately
the same cost. This fortuitous cancellation of effects does not occur in general.
It is always worthwhile exploring the economic impact of variations in the
position of the decanter tie-line.
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 713
Other Sequences. The Kubierschky sequence is not the only way to per-
form the separation. Alternatives include: (1) If the process feed already has a
composition at or near the composition of the binary azeotrope then the precon-
centrator is not needed. (2) The distillate stream from the entrainer recovery col-
umn can be recycled directly to the decanter, which is analogous to the binary
process shown in Fig. 21. This recycle alternative causes the reflux ratio in the
azeotropic column to be much larger than necessary (194) and should be avoided,
even though it has been studied extensively in the literature (172,182,195–199).
(3) Use of the Kubierschky two-column sequence (191), in which the preconcen-
tration and entrainer recovery steps are combined into a single column. For the
ethanol–water–benzene system, this alternative has lower capital costs but
higher operating costs than the Kubierschky three-column sequence so that
the total annualized cost is about the same for both sequences (193). (4) Use of
the Steffen three-column sequence, the basic layout of which is the same as the
Kubierschky three-column sequence. The essential new feature is to replace the
single decanter in the Kubierschky sequence by multiple decanters, adding fresh
or recycled water to each. The entrainer-rich phase from each decanter is
returned to the azeo-column as reflux and the aqueous-rich phase from each
decanter is fed to the next decanter in the train. This effectively replaces the
decanter by a liquid–liquid extraction step which pushes the ‘‘effective decanter
tie-line’’ deeper into the liquid–liquid region than it could possibly go by distilla-
tion alone. (5) Use of the Ricard-Allenet four-column sequence (3,159,191,
193,194,200). The first two columns in this sequence are the same as the first
two columns in the Kubierschky three-column sequence. The third column in
the Ricard-Allenet sequence takes the water-rich phase from the decanter and
splits it into a distillate stream containing the ternary azeotrope, and by material
balance, a bottoms stream consisting of water and ethanol. The ternary azeo-
trope is recycled to the azeo-column decanter, and the ethanol-water stream is
fed to a fourth column, which recovers pure water as bottoms and the binary
azeotrope as distillate. This azeotropic distillate stream is recycled to the feed
into the azeo-column. Ricard-Allenet proposed a variation on this sequence
(191) in which the overhead vapors from the entrainer recovery column (third
column in the sequence) are provided with a separate condenser–decanter
system, similar to the one provided for the overhead vapors from the azeo-col-
umn. (6) Use of the Ricard-Allenet three-column sequence in which the fourth
column is eliminated and the bottom stream from the third column (containing
only ethanol and water) is recycled to the preconcentrator column. This alter-
native has good economic possibilities, but there are no available studies of
this alternative.
In summary, for systems of the ethanol–water–benzene type, the three
most attractive sequences for carrying out azeotropic distillation are the
Kubierschky three-column sequence, the Kubierschky two-column sequence,
and the Ricard-Allenet three-column sequence. For each of these there is the
added possibility of replacing the decanter with a liquid–liquid extraction step.
Other Classes of Entrainers. Not all heterogeneous azeotropic mixtures
are of the ethanol–water–benzene type. The number of azeotropes in the mix-
ture may vary from system to system as may their character, ie, maximum or
minimum boiling, hetereogeneous or homogeneous. In addition, the size and
714 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
A A
B Entrainer B Entrainer
(a) (b)
A A
B Entrainer B Entrainer
(c) (d)
Fig. 25. A selection of feasible residue curve maps for ternary heterogeneous mixtures
where & represents homogeneous and * ~ heterogeneous azeotropes. See text.
shape of the liquid–liquid region varies greatly from system to system. The
feasibility and sequencing strategy for each new system is most conveniently
established using residue curve maps such as those shown in Fig. 25.
Any entrainer that does not divide the components to be separated into dif-
ferent distillation regions is normally a feasible entrainer, and remains feasible
even if it induces a region of partial miscibility. Examples are shown in Fig. 25a
and b and it is possible to construct many more such maps by strategically pla-
cing heterogeneous regions on the feasible residue curve maps given in Fig. 6.
Sequences based on these maps normally have multiple liquid phases on some
of the stages in the column, which may lower the mass-transfer efficiency on
those stages. In practice, mass-transfer efficiency and hydrodynamic perfor-
mance of three-phase columns does not present a problem. As early as 1938 it
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 715
was stated that ‘‘The efficiency of the plates was apparently undiminished by the
heterogeneity of the boiling liquid and that was undoubtedly due to the violent
agitation produced on the plates by the rapid bubbling of the vapours through
the liquid . . .’’ (159). A more recent study and assessment of the literature
from a practitioner’s viewpoint is available (201). However, at the conceptual
design stage of process development, mass-transfer efficiency is of secondary
importance to feasibility.
The map shown in Fig. 25b is relevant to the separation of acetic acid and
water which is of commercial significance. Although this binary mixture does not
form an azeotrope, it does have a severe tangent pinch at high water composi-
tions, which makes it difficult to produce an acid-free water stream. It is not eco-
nomical to separate this mixture into pure product streams without the aid of an
entrainer. All known commercial entrainers for this separation are heteroge-
neous and most produce residue curve maps similar to the one shown in Fig. 25b.
Clarke-Othmer Process for Acetic Acid–Water Separation. Large
amounts of dilute acetic acid are purified industrially. Feasible entrainers
include ethylene dichloride (nbp 83.78C), chloroform (npb 61.88C), ethyl ether
(nbp 34.58C), also called ether or diethyl ether), 1-propyl acetate (nbp 101.68C),
isopropyl acetate (nbp 88.48C), and 1-butyl acetate (nbp 125.98C). All except -
butyl acetate have residue curve maps similar to Fig. 25b. The azeo-column
has a process feed consisting of acetic acid (component A in Fig. 25b) and
water (component B in Fig. 25b). The entrainer is charged to the column, typi-
cally via the reflux stream, and remains captive within the column. The bottom
stream from the column is pure acetic acid, and the overhead vapor is close to the
composition of the minimum boiling heterogeneous binary azeotrope formed by
the entrainer and water (see Fig. 25b). The azeotropic vapors are condensed and
decanted into a water stream that leaves as distillate, and an entrainer-rich
stream that is returned to the column as reflux. In some process alternatives,
part of this stream is also recycled to the column feed (see p. 67 of Ref. 202). Addi-
tional reflux, if needed, is achieved by returning some of the water phase. It is
typical for these systems to have multiple liquid phases present on some stages
in the rectifying section. This process was invented by Clarke and Othmer (203);
more detailed operating information and entrainer comparison for this separa-
tion is available (204). Ethyl acetate (nbp 77.18C) is sometimes mentioned as a
feasible entrainer for separating acetic acid–water mixtures and, although it
does give a residue curve map similar to Fig. 25b, the liquid–liquid region is
so small that it is not nearly as attractive an entrainer as those cited above. It
is for this reason that many users of this system employed multicomponent
entrainers, adding for example cyclohexane and/or methylene chloride to ethyl
acetate so as to enlarge the liquid-liquid region while retaining as much as pos-
sible the favorable distribution coefficients (tie-line slopes) of the ethyl acetate
system. Another entrainer that is sometimes used industrially is n-butyl acetate
that has a normal boiling point of 1268C, which is 88C higher than acetic acid.
This is a feasible entrainer but the residue curve map is different than the one
shown in Fig. 25b (see the section below, Self-Entrained Systems, for more
details).
Wentworth Process for Ethanol–Water Separation. In the Wentworth
process ethyl ether is used as the entrainer for ethanol– water separation,
716 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
producing a residue curve map similar to the one shown in Fig. 25c (A and B
represent ethanol and water, respectively). Ethyl ether and water form a mini-
mum boiling heterogeneous azeotrope at 34.158C containing 98.75 wt% ether and
1.25 wt% water at atmospheric pressure. There is no azeotrope between ethyl
ether and ethanol and no ternary azeotrope. The dilute ethanol process feed is
first preconcentrated up to the composition of the ethanol–water azeotrope in a
preconcentrator column. This stream is fed to the azeo-column that produces
pure ethanol as bottoms product and an overhead vapor close to the composition
of the ethyl ether–water azeotrope. The overhead vapors are condensed and
decanted into an ether-rich layer that is returned to the column as reflux, and a
water-rich layer that leaves as distillate. The distillate contains no alcohol and
very little ether. Pure water may be obtained from this stream by sending it to
a stripping column where water is recovered as the bottom product and the over-
head vapor has a composition near the ethyl ether–water azeotrope. These vapors
are condensed and recycled to the decanter. The azeo-column is normally operated
at 700 kPa (7 atm) pressure because, (1) this increases the amount of water in
the ether–water azeotrope, thereby reducing the amount of ether needed in the
system, and (2) it allows for cooling water to be used in the condenser.
This process is described in the literature in more detail (205–207). Its
main advantage was a lower energy consumption ( 6000 kJ/L (20,000 Btu/
gal) ethanol product) relative to the benzene process (12,000 kJ/L (43,000 Btu/
gal) ethanol) (206). Since the 1940s the gap has been narrowed by better designs
for the benzene process that is now capable of producing 99.8 mol% ethanol at an
energy consumption of 8,400 kJ/L (30,000 Btu/gal) ethanol product (78,193), or
using thermally integrated columns for 5000 kJ/L (18,000 Btu/gal) ethanol pro-
duct (78,208). In recent years, homogeneous separating agents have shown great
promise for the ethanol–water separation, and extractive distillation processes
using ethylene glycol as the solvent have been designed with energy consump-
tions of 6000 kJ/L (22,000 Btu/gal) ethanol product (78), or using thermally
integrated columns for 2200–3300 kJ/L (8,000–12,000 Btu/gal) ethanol product
(78,209,210).
Rodebush Sequence for Ethanol–Water Separation. When ethyl acet-
ate is used as the entrainer to break the ethanol–water azeotrope the residue
curve map is similar to the one shown in Fig. 25d, ie, the ternary azeotrope is
homogeneous. Otherwise the map is the same as for ethanol–water–benzene.
In such cases the liquid leaving the azeo-column condenser does not separate
into two liquid phases, and the sequence is infeasible unless special tricks
are employed. In the Rodebush sequence, water is continuously added to the
decanter in order to shift the overall composition into the two-liquid phase
region. The water-rich phase from the decanter is fed to a distillation column
where pure water is recovered in the bottom stream (some of which is recycled
to the decanter) and the distillate, which lies near the ternary azeotrope, is
recycled to the azeo-column. The ethyl acetate-rich phase from the decanter
is returned to the azeo–column as reflux (191). Pure ethanol is recovered in
the bottom stream of the azeo-column. More recently, a clever variation on
this sequence was patented (211) for separating a ternary feed consisting of
ethanol, water, and diethoxymethane. This variation also has a residue curve
map similar to the one shown in Fig. 25d.
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 717
More Complex Mixtures. All the sequences discussed are type I liquid
systems, ie, mixtures in which only one of the binary pairs shows liquid–liquid
behavior. Many mixtures of commercial interest display liquid–liquid behavior
in two of the binary pairs (type II systems), eg, secondary butyl alcohol–
water–di-secondary butyl ether (SBA–water–DSBE), and water–formic acid-
meta-xylene (162). Sequences for these separations can be devised on the basis
of residue curve maps. The SBA–water–DSBE separation is practiced by
ARCO and is considered in detail in the literature (183,184,188,212).
Self-Entrained Systems. It is quite common to find that the feed mixture
to an azeotropic distillation column is not a binary azeotropic mixture but a tern-
ary mixture that already contains a third component that can act as an entrai-
ner. Such systems are called self-entrained. The presence of the third component
may be unavoidable on account of the reaction chemistry (eg, acetic acid–water–
vinyl acetate (213–215); ethanol–water–diethoxy methane, 211), or may be pre-
sent on purpose due to the choice of process technology. For example, in some
process alternatives for separating acetic acid–water mixtures by azeotropic dis-
tillation part of the entrainer-rich phase from the decanter is recycled back to the
column feed (202, p. 67), thereby creating a self-entrained feed to the column. In
other process alternatives for separating the same mixture, solvent extraction is
used to remove almost all the acetic acid from the water, creating an extract
stream that contains almost all of the acetic acid that entered with the process
feed, some water, and all of the solvent used in the extractor (202, p. 68; 216,217).
This extract stream is then fed (as a self-entrained ternary mixture) to an azeo-
tropic distillation column where glacial acetic acid (>99.7 wt%) is recovered as
the bottom product. In these kinds of separation sequences, the third component
is chosen to be both a good solvent in the extractor and a good entrainer in the
azeotropic distillation. Minotti et al. (218) give a detailed description of the pro-
cess alternatives and process economics of using diethyl ether to separate acetic
acid–water mixtures using a combination of extraction and azeotropic distilla-
tion flow sheets.
Two typical self-entrained azeotropic distillation systems are described
below.
Separation of Acetic–Water–Isopropyl Acetate Mixtures. At 1 atm
pressure, isopropyl acetate boils at 88.48C and forms a binary heterogeneous
azeotrope with water that boils at 77.28C and contains 40 mol% water. Iso-
propyl acetate is the lightest boiling component in the mixture and the residue
curve map, calculated using the NRTL activity coefficeint model, is shown in
Fig. 26. (Note that several other light-boiling entrainers, such as vinyl acetate,
diethyl ether, chloroform, etc, generate similar residue curve maps and have
qualitatively similar azeotropic distillations. Thus, the isopropyl acetate
entrainer is representative of other entrainers of the light-boiling type). For
illustration, a saturated liquid (q ¼ 1) feed composition of 20 mol% isopropyl
acetate, 40 mol% water and 40 mol% acetic acid is assumed. The desired
purity of the bottoms stream is glacial acetic acid containing 99.9 mol% acetic
acid. Two design alternatives exist. With the same feed composition, the
column can be designed to operate with either an entrainer-rich or an
entrainer-lean composition profile. In either case the column configuration is
the same as shown in Fig. 27, and the distillate composition (determined by
718 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
0.8
0.6
x2
0.4
0.2
xxxxxx
0.0
0.0 0.2 0.4 77.2 0.8 1.0
Water (3) x1 Isopropyl Acetate (1)
(100.0°C) (88.4°C)
Fig. 26. Residue curve map and liquid–liquid boiling envelope for acetic acid þ
water þ isopropyl acetate at 1-atm pressure.
material balance with the feed and bottoms compositions) is essentially the
same in both cases. (Note that the distillate composition of acetic acid deter-
mines which tie-line in the liquid – liquid region corresponds to the decanter
tie-line, but the relative proportion of phases in the decanter is determined
by the position of the overhead vapor composition on the decanter tie-line,
Fig. 28. Azeo-column liquid composition profile .—. (solid line connecting filled circles)
and design data for the acetic acid þ water þ isopropyl acetate separation. Entrainer-rich
design with low recovery of acetic acid (2.56 mol % acetic acid in the distillate).
see Fig. 28). The decanter composition, which is the same as the overhead
vapor composition, is, however, quite different in the two designs, as is the
reflux composition (ie, the proportion of the two decanter liquid phases
returned to the column is different), and the reflux ratios.
The first entrainer-rich design is shown in Fig. 28. The distillate is selected
so it contains some acetic acid in order to make the decanter tie-line visible on the
diagram and the relationship between the various points on it more obvious. The
bottom stream, the acetic acid product, contains more isopropyl acetate (0.001
mole fraction) than water (0.0001 mole fraction). This distribution of the trace
components forces the stripping section composition profile to move along the
acetic acid–isopropyl acetate edge of the diagram and intersect with the rectify-
ing section composition profile in the entrainer-rich region of the diagram. The
result is a column design where the liquid on all of the stages throughout the col-
umn contains more entrainer than water. The top of the column is designed so
that the overhead vapor composition (which becomes the feed to the decanter)
lies to the entainer-rich side of the azeotrope on the decanter tie-line. This causes
the liquid on the top tray of the column to be rich in entrainer (see Fig. 28). The
ratio, f, of entrainer-rich phase (L2) to entrainer-lean phase (L1) in the decanter
is 3:1 (this is determined by the design value of the decanter composition). Part of
the entrainer-rich decanter liquid is refluxed to the column, the balance is mixed
with the entrainer-lean decanter liquid to produce a distillate composition at the
required point along the decanter tie-line. There is a simple algebraic material
balance relationship between the reflux ratio and the reflux composition
720 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
(which also contains the distillate composition and the decanter composition as
known parameters). Therefore, a chosen value of r implies that the reflux com-
position is determined, and vice-versa. When we select the reflux composition to
be equal to the composition of the entrainer-rich phase the corresponding value
of the reflux ratio is r ¼ 1.9. This entrainer-rich design requires 17 theoretical
stages to perform the separation at a reflux ratio of 1.9. Sixteen stages are in
the stripping section and only 1 is in the rectifying section. A key measure of eco-
nomic performance (energy use) is the internal vapor rate relative to the feed
rate (V/F), which is 1.79 for this design. Generally in distillation, values of
V/F < 1 are considered very attractive, while values greater than 3 are unattrac-
tive. Values in between are normal.
A more realistic entrainer-rich design is shown in Fig. 29, where the distil-
late is selected to have only a small amount of acetic acid present (mole fraction
of 0.0001). The decanter tie-line lies almost on top of the isopropyl acetate-water
edge of the diagram. In this design, the feed and bottom composition are the
same as before, so is f ¼ L2/L1 ¼ 3. We again choose to reflux only the entrai-
ner-rich phase (which determines the reflux ratio to be r ¼ 1.88), giving a column
with a total of 19 theoretical stages (15 in the stripping section and 4 in the rec-
tifying section), and V/F ¼ 1.72. This design, therefore, produces high purity
glacial acetic acid at high recovery (ie, with very little acetic acid lost in the dis-
tillate) in a column with both a reasonable number of stages and a reasonable
boilup.
0.4
Feed
Feed stage
0.2
xxxxxx
0.0
0.0 0.2 0.4 0.6 yN 0.8 xN 1.0
Water (3) Distillate x1 Isopropyl acetate (1)
Fig. 29. Azeo-column liquid composition profile and design data for the acetic acid þ
water þ isopropyl acetate separation. Entrainer-rich design with high recovery of acetic
acid (100 ppm (molar) acetic acid in the distillate).
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 721
0.4
Feed
0.2
xN
xxxxxx
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Water(3) Distillate yN xreflux Isopropyl acetate (1)
x1
Fig. 30. Azeo-column liquid composition profile and design data for the acetic acid þ
water þ isopropyl acetate separation. Entrainer-lean design with low recovery of acetic
acid (2.47 mol% acetic acid in the distillate).
0.4
Feed
0.2
xN
xxxxxx
0.0
0.0 0.2 0.4 0.6 0.8 1.0
Water (3) Distillate yN xreflux Isopropyl acetate (1)
x1
Fig. 31. Azeo-column liquid composition profile and design data for the acetic acid þ
water þ isopropyl acetate separation. Entrainer-lean design with high recovery of acetic
acid (700 ppm (molar) acetic acid in the distillate).
0.8
0.6
x2
0.4
0.2
xxxxxxxxxxx
0.0
0.0 0.2 0.4 0.6 91.0 0.8 1.0
N-Butyl acetate (3) x1 Water (1)
(125.9°C) (100.0°C)
Fig. 32. Residue curve map and liquid–liquid boiling envelope for acetic acid þ
water þ n-butyl acetate at 1-atm pressure.
rather than around the outside of it. This may be advantageous for some systems
such as the n-butyl acetate case that follows.
Separation of Acetic Acid–Water–n–Butyl Acetate Mixtures. At 1-atm
pressure, n-butyl acetate boils at 1268C and forms a binary heterogeneous azeo-
trope with water that boils at 918C and contains 72 mol% water. N-Butyl acet-
ate is the highest boiling component in the mixture and the residue curve map,
calculated using the NRTL activity coefficient model, is shown in Fig. 32. This
map is quite different than the map shown in Fig. 26 for light-boiling entrainers,
and the feasible splits and azeotropic column designs are also different. Only
entrainer-lean designs are possible. This is due to the fact that acetic acid is
the stable node (high boiler) and the entrainer is a saddle (intermediate boiler)
in Fig. 26, while acetic acid is a saddle and the entrainer is a stable node in
Fig. 32.
Consider a saturated liquid feed containing 20 mol% n-butyl acetate,
40 mol% water and 40 mol% acetic acid. The desired purity of the bottoms stream
is again glacial acetic acid containing 99.9 mol% acetic acid. The column config-
uration is the same as in Fig. 27, and the distillate composition is again chosen to
have some acetic acid present in order to make the decanter tie-line visible on the
triangular diagram. A typical design is shown in Fig. 33. The distillate composi-
tion determines the position of the decanter tie-line, and the overhead vapor com-
position is selected to lie on the entrainer-lean side of the azeotrope (the ratio of
entrainer-rich phase (L2) to entrainer-lean phase (L1) in the decanter is 1:3),
which gives a liquid on the top tray of the column that is also lean in entrainer.
The rectifying profile starts at this point and moves toward the stable node (high
724 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
0.4
Feed
Feed stage
0.2
xN
xxxxxxxxxxx
xrich xlean
0.0
0.0 0.2 0.4 0.6 yN 0.8 1.0
N-Butyl acetate (3) x1 Distillate xreflux Water (1)
Fig. 33. Azeo-column liquid composition profile and design data for the acetic acid þ
water þ n-butyl acetate separation. Low recovery of acetic acid (2.53 mol % acetic acid
in the distillate).
boiler) at n-butyl acetate (ie, in the direction of increasing temperature down the
column). It intersects the stripping profile which moves from the acetic acid
corner of the diagram along the acetic acid–water edge before turning into the
diagram towards the azeotrope (ie, in the direction of decreasing temperature
up the column). The minimum reflux at which these profiles just intersect is
1. A reasonable trade-off between stages and reflux occurs at r ¼ 2 (V/
F ¼ 1.84), giving 30 theoretical stages in the column (23 in the stripping section
and 7 in the rectifying section), see Fig. 33. As can be seen in this Fig., the reflux
stream and the liquid composition on all of the stages are entrainer-lean. It is not
possible to design a column with an entrainer-rich profile because rectifying pro-
files that start near the entrainer-rich end of the decanter tie-line will move
directly to pure n-butyl acetate (in the direction of increasing temperature
down the column). The rectifying profile can never intersect the stripping profile,
which starts at acetic acid and moves toward the azeotrope (in the direction of
decreasing temperature up the column). Another way of saying this is that
heavy entrainers cannot be forced up the column to generate entrainer-rich
designs.
With n-butyl acetate as the entrainer, high recovery of acetic acid is possi-
ble only with larger reflux ratios (in the range of r ¼ 3–5, giving V/F in the range
of 2.4–3.6) and larger numbers of theoretical stages (40–34, respectively), eg, see
Fig. 34. Reflux ratios < 3 increase the number of stages dramatically (eg, r ¼ 2
requires 50 theoretical stages). This effect is due to the proximity of the column
composition profile to the tangent pinch between acetic acid and water located in
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 725
Feed Stage
0.2
xxxxxxxxxxx
0.0
0.0 0.2 0.4 0.6 yN 0.8 xreflux1.0
N-Butyl acetate (3) x1 Distillate
Water (1)
Fig. 34. Azeo-column liquid composition profile and design data for the acetic acid þ
water þ n-butyl acetate separation. Higher recovery of acetic acid (0.13 mol % acetic
acid in the distillate).
the vicinity of pure water. When the distillate contains very little acetic acid, the
decanter tie-line lies on the n-butyl acetate-water edge of the triangle, and this
puts the composition of the liquid on the top stage of the column right on top of
the tangent pinch where the separation per stage is very small, see Fig. 34.
For this system there is a clear advantage to having multiple liquid phases
on the top several trays of the rectifying section. Such designs put the top portion
of the rectifying profile inside of the liquid–liquid region of the diagram, well
away from the tangent pinch. This has the effect of simultaneously reducing
both the reflux ratio and the total number of theoretical stages, see Fig. 35.
Parten and Macpherson Ure (189) disclose a process for achieving this by only
refluxing the organic phase. Urdaneta and co-workers (180) and Wasylkiewicz
and co-workers (190) propose methodologies for systematically designing such
azeotropic distillations with multiple liquid phases on some of the stages inside
the column. Designs for the acetic acid-water-n-butyl acetate system are consid-
ered in detail in (190), and one is reproduced in Fig. 35.
Whereas there is extensive literature on design methods for azeotropic and
extractive distillation, much less has been published on operability and control.
It is, however, widely recognized that azeotropic distillation columns can be dif-
ficult to operate and control because some of these columns can exhibit complex
dynamic behavior, multiple steady-states, and parametric sensitivity (172–
174,182–187,219). In contrast, extractive distillations typically do not exhibit
such complex behavior and even highly optimized columns are typically no
726 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
0.8
0.6
0.4
0.2
xxxxxxxxxxxxxx
0.0
0.0 0.2 0.4 0.6 0.8 1.0
N-Butyl acetate (3) Water (1)
Fig. 35. Azeo-column liquid composition profile for the acetic acid þ water þ n-butyl acet-
ate separation with two liquid phases on the top stages in the rectifying section. Taken
from reference (190). The heterogeneous binary azeotrope and vapor-line are also shown.
more difficult to control than ordinary distillation columns producing high purity
products (220). Exceptions do exist, however. Multiple steady-states have been
observed in at least one industrial extractive distillation (221).
8. Nomenclature
BIBLIOGRAPHY
‘‘Azeotropes’’ in ECT 1st ed., under ‘‘Distillation,’’ Vol. 5, pp. 176–179, by E. G. Scheibel,
Hoffmann-LaRoche Inc.; ‘‘Azeotropy and Azeotropic Distillation’’ in ECT 2nd ed., Vol. 2,
pp. 839–858, by D. F. Othmer, Polytechnic Institute of Brooklyn; ‘‘Azeotropic and Extrac-
tive Distillation’’ in ECT 3rd ed., Vol. 3, pp. 352–377, by D. F. Othmer, Polytechnic Insti-
tute of New York; in ECT 3rd ed., Supl. Vol., pp. 145–158, by G. Prokopakis, Columbia
University.
CITED PUBLICATIONS
1. L. H. Horsley, Azeotropic Data III, Advances in Chemistry Series No. 116, American
Chemical Society, Washington, D.C., 1973.
2. J. Gmehling, J. Menke, J. Krafczyk, and K. Fischer, Azeotropic Data, Parts I & II,
VCH Publishers, Inc., New York (1994).
3. C. J. King, Separation Processes, 2nd ed., McGraw-Hill Publishing Co., New York,
1980.
4. J. D. Seader, J. J. Siirola, and S. D. Barnicki, in R. H. Perry, D. W. Green and J. O.
Maloney eds., Perry’s Chemial Engineer’s Handbook, 7th ed., McGraw-Hill, New
York (1997), Section 13, pp. 54–81.
5. G. Q. Martin, Hydroc. Proc., 241 (Nov. 1975).
6. J. M. Smith and H. C. Van Ness, Introduction to Chemical Engineering Thermody-
namics, 3rd ed., McGraw-Hill, New York, 1975.
7. U.S. Pat. 4,287,379 (Sept. 1, 1981), E. Brunner, E. Schubert, A. Lindner, F. Merger,
K. Volkamer, M. Strohmeyer, and G. Sandrock (to BASF AG).
8. A. Fahmy, D. Mewes, and K. Ebert, Separation Sci. and Technol., 36(15), 3287
(2001).
9. F. Lipnizki, R. W. Field, and P.-K. Ten, J. Membrane Sci., 153, 183 (1999).
10. G. Guerreri, Chem. Eng. Res. Des., 70(A5), 501 (1992).
11. J. G. Stichlmair, J. R. Fair, and J. L. Bravo, Chem. Eng. Prog. 85(1), 63 (1989).
12. S. Widagdo and W. D. Seider, AIChE J., 42(1) 96 (1996).
13. M. F. Doherty and G. A. Caldarola, Ind. Eng. Chem. Fund., 24, 474 (1985).
728 DISTILLATION, AZEOTROPIC AND EXTRACTIVE Vol. 1
14. H. Matsuyama and H. Nishimura, J. Chem. Eng. Jpn., 10, 181 (1977).
15. E. J. Peterson and L. R. Partin, Ind. Eng. Chem. Res., 36, 1799 (1997).
16. M. F. Doherty and J. D. Perkins, Chem. Eng. Sci., 33, 281 (1978).
17. M. F. Doherty and J. D. Perkins, Chem. Eng. Sci., 33, 569 (1978).
18. E. R. Foucher, M. F. Doherty, and M. F. Malone, Ind. Eng. Chem. Res., 30, 760
(1991).
19. M. F. Doherty, Chem. Eng. Sci., 40, 1885 (1985).
20. M. F. Doherty and J. D. Perkins, Chem. Eng. Sci., 34, 1401 (1979).
21. M. F. Doherty and M. F. Malone, Conceptual Design of Distillation Systems,
McGraw-Hill, New York (2001).
22. Y. Yamakita, J. Shiozaki, and H. Matsuyama, J. Chem. Eng. Jpn., 16, 145 (1983).
23. D. B. Van Dongen and M. F. Doherty, Chem. Eng. Sci., 40, 2087 (1985).
24. C. Bernot, M. F. Doherty, and M. F. Malone, Chem. Eng. Sci., 45, 1207 (1990).
25. C. Bernot, M. F. Doherty, and M. F. Malone, Chem. Eng. Sci., 46, 1311 (1991).
26. B. S. Ahmad, Y. Zhang, and P. I. Barton, AIChE J., 44(5), 1051 (1998).
27. M. Benedict and L. C. Rubin, Trans. AIChE, 41, 353 (1945).
28. E. Rev, Ind. Eng. Chem. Res., 31, 893 (1992).
29. L. Laroche, N. Bekiaris, H. W. Andersen, and M. Morari, Ind. Eng. Chem. Res., 31,
2190 (1992).
30. L. Laroche, N. Bekiaris, H. W. Andersen, and M. Morari, AIChE J., 38, 1309 (1992).
31. O. M. Wahnschafft, J. W. Koehler, E. Blass, and A. W. Westerberg, Ind. Eng. Chem.
Res., 31, 2345 (1992).
32. J. G. Stichlmair and J.-R. Herguijuela, AIChE J., 38, 1523 (1992).
33. Z. T. Fidkowski, M. F. Doherty, and M. F. Malone, AIChE J., 39(8), 1303 (1993).
34. G.-J. A. F. Fien and Y. A. Liu, Ind. Eng. Chem. Res., 33, 2505 (1994).
35. S. K. Wasylkiewicz, L. C. Kobylka, and M. A. Satyro, Chem. Eng., 80 (Aug. 1999).
36. D. Y.-C. Thong and M. Jobson, Chem. Eng. Sci., 56, 4369 (2001).
37. V. Julka, A Geometric Theory of Multicomponent Distillation, Ph.D. dissertation,
University of Massachusetts, Amherst, 1992.
38. Z. T. Fidkowski, M. F. Malone, and M. F. Doherty, Comput. Chem. Engng., 17(12),
1141 (1993).
39. S. T. Harding, C. D. Maranas, C. D. McDonald, and C. M. Floudas, Ind. Eng. Chem.
Res., 36, 160 (1997).
40. S. K. Wasylkiewicz, M. F. Doherty, and M. F. Malone, Ind. Eng. Chem. Res., 38, 4901
(1999).
41. B. T. Safrit and A. W. Westerberg, Ind. Eng. Chem. Res., 36, 1827 (1997).
42. R. E. Rooks, V. Julka, M. F. Doherty, and M. F. Malone, AIChE J., 44(6), 1382
(1998).
43. B. S. Ahmad and P. I. Barton, AIChE J., 42(12), 3419 (1996).
44. D. B. Van Dongen, Distillation of Azeotropic Mixtures: The Application of Simple-
Distillation Theory to Design of Continuous Processes, Ph.D. dissertation, University
of Massachusetts, Amherst, 1983.
45. S. G. Levy, Design of Homogeneous Azeotropic Distillations, Ph.D. dissertation,
University of Massachusetts, Amherst, 1985.
46. S. G. Levy, D. B. Van Dongen, and M. F. Doherty, Ind. Eng. Chem. Fund., 24, 463
(1985).
47. J. D. Olson, Fluid Phase Equil., 185(1–2), 209 (2001).
48. C.-P. Kao, R. N. Miller and J. F. Sturgis, J. Chem. Eng. Data, 46(2), 229 (2001).
49. C.-P. Kao, M. E. Paulaitus, and A. Yokozeki, Fluid Phase Equil., 127(1–2), 191 (1997).
50. S. P. Christensen and J. D. Olson, Fluid Phase Equil., 79, 187 (1992).
51. W. J. Gaw and F. L. Swinton, Nature, 212, 283 (1966).
52. R. Q. Wilson, W. H. Mink, H. P. Munger, and J. W. Clegg, AIChE J., 1, 220 (1955).
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 729
88. U.S. Pat. 6,175,045 (Jan. 16, 2001), S. Kohno and T. Shibanuma (to Daikin Ind. Ltd.).
89. U.S. Pat. 5,922,175 (July 13, 1999), R. Bertocchio, E. Deslandes, and E. Lacroix (to
Elf Atochem SA).
90. U.S. Pat. 6,039,845 (Mar. 21, 2000), R. Bertocchio, E. Lacroix, and S. Perdrieux (to
Elf Atochem SA).
91. U.S. Pat. 5,874,657 (Feb. 23, 1999), B. A. Mahler and R. N. Miller (to E. I. DuPont de
Nemours & Co.).
92. U.S. Pat. 4,428,798 (Jan. 31, 1984), D. Zudkevitch, S. E. Belsky, and P. D.
Krautheim (to Allied Corp.).
93. U.S. Pat. 4,455,198 (June 19, 1984), D. Zudkevitch, D. K. Preston, and S. E. Belsky
(to Allied Corp.).
94. E. G. Scheibel, Chem. Eng. Prog., 44(12), 927 (1948).
95. L. Laroche, N. Bekiaris, H. W. Andersen, and M. Morari, Can. J. Chem. Eng., 69,
1309 (1991).
96. J. P. Knapp, Exploiting Pressure Effects in the Distillation of Homogeneous Azeotro-
pic Mixtures, Ph.D. dissertation, University of Massachusetts, Amherst, 1991.
97. B. Kolbe, J. Gmehling, and U. Onken, I. Chem. E. Symp. Ser., (56), 1.3/23 (1979).
98. J. P. Knapp and M. F. Doherty, AIChE J., 40(2), 243 (1994).
99. V. Julka and M. F. Doherty, Chem. Eng. Sci., 48, 1367 (1993).
100. Z. T. Fidkowski, M. F. Malone, and M. F. Doherty, AIChE J., 37, 1761 (1991).
101. S. G. Levy and M. F. Doherty, Ind. Eng. Chem. Fund., 25, 269 (1986).
102. J. R. Knight, Synthesis and Design of Homogeneous Azeotropic Distillation
Sequences, Ph.D. dissertation, University of Massachusetts, Amherst, 1986.
103. J. M. Douglas, Conceptual Design of Chemical Processes, McGraw-Hill, New York,
1988.
104. U.S. Pat. 1,074,287 (Sept. 30, 1913), H. Pauling.
105. C. S. Robinson and E. R. Gilliland, Elements of Fractional Distillation, 4th ed.,
McGraw-Hill, New York, 1950.
106. U.S. Pat. 4,966,276 (Oct. 30, 1990), A. Guenkel.
107. T. Grewer, Chemie. Ing. Techn., 43, 655 (1971).
108. U.S. Pat. 4,076,594 (Feb. 28, 1978), H. Buelow, H. Hohenschutz, J. E. Schmidt, and
W. Sachaze (to BASF Aktiengesellschaft).
109. F.-M. Lee and J. C. Gentry, Chem. Eng. Prog., 93(10), 56 (1997).
110. M. Van Winkle, Distillation, McGraw-Hill, New York, 1967.
111. M. Benedict, C. A. Johnson, E. Solomon, and L. C. Rubin, Trans. AIChE, 41, 371
(1945).
112. F.-M. Lee, Chem. Eng., 105(12), 112 (1998).
113. J. M. Prausnitz and R. Anderson, AIChE J., 7(1), 96 (1961).
114. C. S. Carlson and J. Stewart, in E. S. Perry and A. Weissberger, eds., Techniques of
Organic Chemitry, Vol. IV, Distillation, Wiley-Interscience, New York, 1965.
115. L. Berg, Chem. Eng. Prog., 65(9), 52 (1969).
116. R. H. Ewell, J. M. Harrison, L. Berg., Ind. Eng. Chem., 36(10), 871 (1944).
117. L. A. Robbins, Chem. Eng. Prog., 76(10), 58 (1980).
118. L. A. Robbins, in R. H. Perry, D. W. Green and J. O. Maloney, eds., Perry’s Chemical
Engineer’s Handbook, 6th ed., McGraw-Hill, New York (1984), pp. 15–6 to 15–8.
119. J. Gmehling, R. Wittig, J. Lohmann, and R. Joh, Ind. Eng. Chem. Res., 41, 1678 (2002).
120. J. Gmehling, Fluid Phase Equilib., 144, 37 (1998).
121. F. Eckert and A. Klamt, Ind. Eng. Chem. Res., 40, 2371 (2001).
122. F. Eckert and A. Klamt, AIChE J., 48(2), 369 (2002).
123. M. Schiller and O. Muthu, ‘‘Is There Life after UNIFAC? Past, Present, and Future
of Mixture Phase Equilibria Prediction Methods,’’ Paper 4b, AIChE Spring National
Meeting, New Orleans, March, 2002.
Vol. 1 DISTILLATION, AZEOTROPIC AND EXTRACTIVE 731
GENERAL REFERENCES
MICHAEL F. DOHERTY
University of Massachusetts, Amherst
JEFFREY P. KNAPP
E. I. du Pont de Nemours & Co., Inc.
734 SEPARATIONS PROCESS SYNTHESIS Vol. 1
SEPARATIONS
PROCESS SYNTHESIS
1. Introduction
Chemical process and plant design involves a hierarchy of complex and creative
activities including both routine and innovative design. Routine design is largely
analytical and is primarily concerned with determining values for the specifica-
tion and operation of specific units, eg, reactors, extractors, and distillation col-
umns. By contrast, process synthesis, the generation of conceptual flow sheets
comprising such units, is a more open-ended activity characterized by a com-
binatorially large number of feasible alternatives. Finding better flow sheet
alternatives has a significant impact on overall process competitiveness. This
is particularly true for separations process synthesis, the selection of separation
methods, their interconnection, and their operating parameters. Virtually every
chemical process involves the recovery, isolation, or purification of products,
by-products, intermediates, wastes, or raw materials. The separation systems
to accomplish these tasks often dominate the capital and operating costs of che-
mical manufacturing processes.
Several approaches to the separations process synthesis problem have been
formulated including superstructure optimization, evolutionary modification, and
systematic generation. From a known feed composition, desired product composi-
tions, and a well-defined set of separation methods, superstructure optimization
approaches construct a hypothetical flow sheet that includes all applicable separa-
tion methods interconnected in every possible order so as to include all possible
separations scheme alternatives. Separations synthesis then becomes a problem
of systematically stripping away the less desirable parts of this superstructure
while simultaneously optimizing the design and operating parameters of the
remaining separation methods using mixed-integer nonlinear programming (1).
This purely mathematical approach is computationally intensive, but is becoming
well-established for certain classes of distillation synthesis problems.
Evolutionary modification starts with an existing flow sheet for a similar
separation to which adaptations are made as necessary to meet the objectives
of the specific case at hand. This approach is exemplified by standard flow
sheet patterns, eg, the schemes for breaking heterogeneous minimum-boiling
binary azeotropes or the sequences for extractive distillation (see DISTILLATION,
AZEOTROPIC AND EXTRACTIVE) (2–4). Although rarely resulting in novel designs, evo-
lutionary modification is a frequently used separations synthesis technique
because of the extensive existing repertoire of design heuristics, standard pat-
terns, and encyclopedias of complete flow sheets.
In the systematic generation approach, the separations flow sheet is con-
structed from a portfolio of basic components in a directed fashion so that a given
feed stream is progressively transformed into one or more target compositions.
This process can be viewed as the solution of integrated equipment selection, sequ-
encing, and specification problems. Resource constraints generally limit the num-
ber of separation process alternatives that may be generated and evaluated to a
small fraction of the total number of alternatives that are theoretically possible.
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1905160102011814.a01.pub2
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 735
Two Product Heuristic (No. 1). If there are only two products in a mixture, and
all of the components in one product are more volatile than all of the com-
ponents in the other product, then the next separation should divide the
components into two pure products.
Process Hazard Heuristic (No. 2). If a component is corrosive, unstable, reac-
tive, or otherwise hazardous, then the next separation should work toward
removing the component from the bulk process stream.
736 SEPARATIONS PROCESS SYNTHESIS Vol. 1
Direct Sequence Heuristic (No. 3). If the most volatile component (product) in
the mixture is greater than 20 mol% of the feed, and the most volatile
component (product) has the largest mole fraction in the feed, and the com-
ponent (product) with the second largest mole fraction is present in a ratio
of less than 90% of the most volatile component (product), and the proposed
separation is one of the easiest separations remaining, then the next separa-
tion should split off the most volatile component (product) as distillate.
Indirect Sequence Heuristic (No. 4). If the least volatile component (product)
in the mixture is greater than 20 mol% of the feed, and the least volatile
component (product) has the largest mole fraction in the feed, and the com-
ponent (product) with the second largest mole fraction is present in a ratio
of less than 90% of the least volatile component (product), and the pro-
posed separation is one of the easiest separations remaining, then the
next separation should split off the least volatile component (product) as
bottoms.
Distillate to Bottoms Ratio Heuristic (No. 5). If a separation has a distillate to
bottoms ratio (molar basis) of 40:60–60:40, and the proposed separation
is one of the easiest separations remaining, then the next separation should
be the one closest to a 50:50 distillate to bottoms ratio.
Easiest Next Heuristic (No. 6). If none of the other heuristics apply, then the
next separation to be done should be the easiest, that is the one with the
highest separation coefficient, S, defined as
Fig. 1. Four Component Separation Sequence: (a) noncoupled direct sequence, (b) ther-
mally coupled with 8 sections and 3 reboilers and condensers, and (c) thermally coupled
with 10 sections and 2 reboilers and condensers.
738 SEPARATIONS PROCESS SYNTHESIS Vol. 1
sequence (Fig. 1a) requires 3 columns with a total of 6 column sections and 6
reboilers and condensers. The partially coupled sequence (Fig. 1b) has halved
the number of reboilers and condensers, but increased the number of column sec-
tions to 8. The fully coupled sequence (Fig. 1c) requires 10 column sections with
the minimum 1 reboiler and 1 condenser. Each of these configurations and the
many other possible variations involve capital, energy usage, and operability
trade-offs that are difficult to know a priori. Often highly coupled sequences suf-
fer from difficult control issues and have seen limited application in industry.
Heuristics have been proposed to help generate practical distillation sequence
superstructures (14–16). The synthesis of thermally coupled distillation column
sequences is a significant step up in design complexity over conventional distilla-
tion sequencing. However, sophisticated heuristic/logic and MINLP techniques are
now able to solve robustly and reliably conventional to fully thermally coupled dis-
tillation sequencing problems in acceptable computation times (17,18).
MIPK MIPK
95.3°C 95.3°C
B2 B2
F2 F2
B1
A2 A2
80.9°C 80.9°C D2
B1
D2 F1 F1
D3
D3
B3 B3 F3
F3
D1
Fig. 3. Methyl ethyl ketone–MIPK–water system, where A1 and A2 represent two dif-
ferent azeotropes; F1, F2, and F3, different feed compositions; Bn and Dn the correspond-
ing bottoms and distillates, respectively; (—), the distillation boundary; and (X), the
reachable compositions for the various feeds: (a) approximate bow-tie and (b) exact reach-
able compositions.
is situated in a different distillation region than the desired product, pure MIPK
cannot be obtained by simple single-feed distillation at all.
Phase diagrams are particularly effective tools for separation synthesis
work because of their ability to combine various types of thermodynamic infor-
mation onto one representation including boiling points of pure components
and azeotropes, location of azeotropes and any VLE-based distillation bound-
aries, location of LLE binodal curves and tie lines, melting points, eutectic
points, and SLE phase compositions. Many of these thermodynamic features
have a significant impact on the separation system flowsheet. Figure 4 shows
examples of combined vapor–liquid and liquid–liquid phase diagrams for several
MIPK IPA
95.3°C 82.2°C
Node Node Distillation
Distillation region II
region I
Distillation
Distillation A2 region I
A2 region II 80.4°C
80.9°C Saddle
Saddle
(a) (b)
IPA
82.3°C Distillation
Node region II
Distillation
region I
A2
78.7°C
Saddle
A4
61.6°C
Saddle
100.0°C A3 68.8°C
Node 58.7°C Node Saddle
Water A1 Hexane
61.5°C
Saddle
(c)
Fig. 4. Combined residue curve and phase equilibria diagrams for the systems where
A1-4 represent azeotropes, (—) is the distillation boundary, the shaded areas represent
the region of two liquid phases, and (– – –) are tie lines: (a) MEK–MIPK–water;
(b) methylene chloride–IPA–water; and (c) hexane–IPA–water.
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 743
ternary systems. In each of these diagrams, a number of special points, lines, and
regions have been identified. Figure 4a shows the two-phase liquid region for the
MEK–MIPK–water system previously mentioned, while Fig. 4b illustrates the
methylene chloride–2-propanol–water system. Both systems have a distillation
boundary spanned by liquid–liquid tie lines. The n-hexane–isopropyl alcohol
(IPA)–water system is more complicated (Fig. 4c), exhibiting three distillation
regions as well as a two-phase liquid region (24).
Relatively few points or composition regions in the phase diagram are of
particular significance in separations process synthesis. These compositions of
interest include the feed and desired product compositions, as well as azeotropes,
eutectics, and selected points on liquid–liquid binodal curves. The choice of
composition for a mass separation agent, if required, is critical and usually is a
composition that can be conveniently regenerated in the process. Binary and tern-
ary azeotropes that are also high or low boiling nodes (as well as the corresponding
two liquid-phase compositions, if heterogeneous) are of particular interest in this
regard as these points are reachable compositions by distillation and decantation.
Pure components and azeotropes that are saddles are by contrast poor choices for
MSA recycle as they are not included in the set of reachable compositions, unless
they become nodes when another component in the system is deleted.
Although distillation is a favored separation method, quite a number of
situations prevent or interfere with the use of simple direct or indirect mode
distillation schemes. In general, the designer will be faced with avoiding, over-
coming, or exploiting a limited set of critical features in order to accomplish
the overall objective of the separation system. These critical features include
(1) distillation boundaries, ie, if product (or MSA) composition is in a different
distillation region from the feed mixture, the product cannot be obtained directly
by simple single-feed distillation; (2) saddle products, ie, if a product (or MSA) is
a saddle in a particular distillation region, that product cannot be obtained at
high purity directly by simple single-feed distillation; (3) pinched or close-boiling
regions, ie, if a feed and product are separated by a region of low relative vola-
tility, simple single-feed distillation is not precluded, but tends to require a
large number of stages and high reflux ratio; (4) overlapping melting–boiling
points, ie, some solutions may contain components with melting points that
are higher that the boiling points of other components and distillation of
such mixtures may result in solidification within the column; (5) temperature-
sensitive components, ie, some mixtures may contain components that degrade,
decompose, polymerize, or otherwise react in an undesirable manner at the tem-
perature conditions of distillation, thus some milder separation method must be
used (25).
3.2. Flowsheet Construction. A purely opportunistic approach to
separation synthesis for nonideal mixtures may again be considered. Guided
by the general sequencing heuristics given, any separation method may be
picked that is feasible and applicable to the state under consideration, and build-
ing the solution toward the desired products may proceed. However, the exis-
tence of the critical features mentioned previously can preclude the application
of simple distillation. In systems exhibiting critical features, it is quite possible
to reach an intermediate composition from which it is impossible to reduce
remaining composition differences by distillation or any other separation
744 SEPARATIONS PROCESS SYNTHESIS Vol. 1
Table 1. (Continued )
Separation
Strategy implementation Comments
exploit solid–liquid melt crystallization driving force for melt cryst. is
equilibrium to reach adsorption a melting point difference
saddle of impurities and desired
product. Feed must meet
certain melting point and
crystal characteristics for
feasibility (see also solu-
tion crystallization)
use existing component or extractive extractive agent modifies
add new MSA to reach distillation liquid-phase behavior
saddle in a single two-feed (activity coefficients) of key
column components. Residue curve
maps must be of appropriate
form for extractive
distillation to work
exploit solid–liquid solubility solution based on differences in
differences to reach saddle crystallization melting points and liquid
phase solubilities. Four
modes of operation
possible: drown out,
isothermal evaporation,
adiabatic evaporation,
cooling. Choice depends
on stream characteristics
reduce or expand the azeotropic a saddle in a three-component
dimensionality of problem distillation system becomes a node in a
to turn saddle into a node; distillation two-component
move to a face of the RCM system. Add-
ing the appropriate
component can alter the
dist. region boundaries
and turn a saddle into
a node
exploit kinetic phenomena adsorption requires the selection
to reach saddle membrane of solid-phase agent
permeation
pervaporation
react to turn saddle into node reaction reactions involve components
reactive already in mixture
distillation
Overcoming Distillation Boundaries
use existing liquid–liquid decanter uses LLE to avoid
region or add component liquid–liquid boundary in VLE
that causes new or altered extraction
liquid–liquid region
exploit solid–liquid melt crystallization driving force for melt cryst. is
equilibrium adsorption a melting point difference
of impurities and desired
product. Feed must meet
certain melting point and
crystal characteristics for
feasibility (see also
solution crystallization)
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 747
Table 1. (Continued )
Separation
Strategy implementation Comments
mix streams to mixer may generate infeasible
cross-boundary solutions due to mass
balance violations. Only
works if there is a way to
get a composition that is
purer than the desired
mixture product (farther
than the middle point of
the balance)
exploit existing convex distillation curvature is difficult to
curvature or add predict accurately
component that causes without experimental
favorable convex curvature data. Boundary must be
to cross boundary highly curved to avoid
large recycles
change system pressure to distillation feasibility highly dependent
alter–eliminate boundary on extent that azeotrope
composition is altered
by moderate change in
pressure
exploit solid–liquid solubility solution based on differences in
differences crystallization melting points and liquid
phase solubilities. Four
modes of operation
possible: drown out,
isothermal evaporation,
adiabatic evaporation,
cooling. Choice depends
on stream characteristics
exploit kinetic phenomena to molecular sieve requires the selection of
cross-boundary adsorption solid-phase agent
membrane
permeation
pervaporation
Vapor–Solid–Liquid Equilibria
use intermediate reflux of sidedraw-reflux occurs when boiling and
low melting point column melting points of
component to prevent components overlap
freezing of high melter
exploit solid–liquid melt crystallization driving force for melt cryst.
equilibrium to separate is a melting point
overlapping difference of impurities
melters–boilers and desired product. Feed
must meet certain
melting point and crystal
characteristics for
feasibility (see also
solution crystallization)
add a component to strip stripping
out lighter components
to separate overlapping
melters–boilers
748 SEPARATIONS PROCESS SYNTHESIS Vol. 1
Table 1. (Continued )
Separation
Strategy implementation Comments
exploit solid–liquid solubility solution based on differences in
differences to separate crystallization melting points and liquid
overlapping melters– phase solubilities. Four
boilers modes of operation
possible: drown out,
isothermal evaporation,
adiabatic evaporation,
cooling. Choice depends
on stream characteristics
Temperature Sensitive Mixtures
use existing liquid–liquid decanter perform separation at
region or add component liquid–liquid conditions below
that causes new or altered extraction reaction–degradation
liquid–liquid region temperature: usually
can be accomplished
at temperatures
between 20 and 458C
exploit solid–liquid melt crystallization driving force for melt cryst.
equilibrium adsorption is a melting point
difference of impurities
and desired product.
Feed must meet certain
melting point and crystal
characteristics for
feasibility (see also
solution crystallization)
alter system pressure vacuum distillation feasibility highly dependent
to decrease reboiler wiped film on pressure required to
temperature evaporation prevent degradation and
short-path stage requirements of
evaporation desired separation
exploit solid–liquid solution based on differences in
solubility differences crystallization melting points and liquid
phase solubilities. Four
modes of operation
possible: drown out,
isothermal evaporation,
adiabatic evaporation,
cooling. Choice depends
on stream characteristics
exploit kinetic phenomena molecular sieve requires the selection
adsorption mem- of solid-phase agent
brane permeation
DECISION POINTS:
Product specifications
Identity of MSA species
DECISION POINTS:
Selection of destination for current stream
Selection of recycle point for current stream
Selection of operation to be applied to current stream
DECISION POINTS:
NONE
DECISION POINTS:
All problem specs met?
(c) Very difficult when temperatures of high and low boiling nodes of
region differ by < 58C.
(d) Simple one-column distillation is generally unfavorable when
the bottoms to distillate ratio is either very high or very low
(eg, 95/5 D/B ratio 5/95). For low D/B ratio high reflux
required. For high D/B ratio large amount of feed must be
vaporized.
(e) Consider special multiple-effect distillation schemes if the bottoms
to distillate ratio is either very high or very low, or if the relative
volatility is low.
(f) Avoid attempts to recover simultaneously both high and low boiling
nodes in high purity from mixtures of three or more components,
particularly in columns that reflux compositions different from
the distillate composition (ie, reflux of one phase from a decanter),
as such operation may be difficult to control.
2. Azeotropic Distillation (2) (see DISTILLATION, AZEOTROPIC AND EXTRACTIVE)
(a) The more structurally, chemically similar components are, the
less likely that the separation will be improved by azeotropic dis-
tillation (if an MSA-key component azeotrope is being used to al-
ter the RCM). Any azeotropes formed between one component and
another similar component will tend to have similar boiling points,
compositions.
(b) MSA selection is critical.
(c) See 8a1-6 above.
3. Extractive Distillation (2) (see DISTILLATION, AZEOTROPIC AND EXTRACTIVE)
(a) Usually not favorable for separation of components that show
very similar liquid-phase behavior: steroisomers, homologous series,
isonormal–neoisomers. Components to be separated must have
some different functional group for MSA to affect liquid-phase beha-
vior differentially.
(b) Only certain residue curve maps are favorable for extractive
distillation.
(c) MSA selection is critical.
4. Pressure Swing Distillation (2,26) (see DISTILLATION, AZEOTROPIC AND
EXTRACTIVE)
(a) Azeotrope composition must change at least 5–10 mol% over mod-
erate pressure change for process to be economical. Generally lar-
ger change required for minimum boiling azeotrope than maximum
boiling azeotrope. Extent of pressure change somewhat dictated by
reboiler temperature in high pressure column, condensing tem-
perature in low pressure column (try to avoid refrigeration, special
heat exchange fluids).
(b) For minimum boiling azeotrope, generally less favorable be-
cause recycle composition between columns taken overhead,
pure products are bottoms: results in high energy usage, larger
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 755
(a) More curved the boundary, the less recycle required. Generally less
favorable for minimum boiling azeotropes due to large overhead
recycles.
(b) Consider other methods if recycle rates are high (ie, three to five
times feed rate or more).
(c) Boundaries involving maximum azeotropes often highly curved.
(d) Not applicable to binary systems; azeotrope is point in phase
diagram.
6. High Vacuum Distillation (28,29)
(a) Consider high vacuum distillation when material is temperature
sensitive at normal pressure ranges.
(b) Consider wiped-film evaporator (external condenser) when pres-
sure must be in 5–0.1-Torr range.
(c) Consider short-path evaporator (internal condenser) when pressure
must be in 0.1–0.001-Torr range. For larger scale production, 0.01–
0.005 Torr is practical lower pressure limit.
(d) At best, only one equilibrium stage achievable for each WFE or SPE
unit. Separation efficiency decreases as pressure decreases.
(e) Operate at a vacuum no lower than required. Short-path unit typi-
cally 25–35% higher cost than wiped-film unit.
(f) Separate all noncondensables before feeding to WFE or SPE unit.
Even a small amount of noncondensables will overload vacuum sys-
tem, especially at ultrahigh vacuum ranges. Remember that most
low molecular weight compounds will not condense at cooling water
temperatures under high vacuum.
II. General Separation Heuristics: Crystallization Methods (30–32)
1. Solution Crystallization: General
(a) Cost increases greatly with decreasing temperature. Try to keep at
or above 0–208C.
(b) Order of preference: adiabatic evaporation > evaporation > cooling
> drown-out.
(c) If system forms solid solution then multistage crystallization re-
quired. Likely to be expensive option.
2. Solution Crystallization: Adiabatic Evaporation (Vacuum Cooling)
(a) Favored for systems where temperature change has moderate
effect on solubility.
756 SEPARATIONS PROCESS SYNTHESIS Vol. 1
Table 5. (Continued )
Strategy Heuristic Status Comments
exploit kinetic III.3.(a) molecular sieve difference in kinetic
phenomena to adsorption diameters (water 2.65
cross-boundary feasible with Å; IPA 4.2 Å) acceptable
preconcentration for 3-A mol sieves.
Large amount of water
to adsorb; preconcen-
tration step probably
necessary to keep bed
size reasonable
III.4.(a) membrane dehydration of IPA
permeation known application
feasible with of pervaporation
preconcentration (eg, cellulose acetate
membranes). Large
amount of water to pass
through membrane,
preconcentration step
probably necessary to
keep membrane area
reasonable
Fig. 5. Distillation of aqueous 2-propanol (IPA), where A is the azeotrope, combined with
(a) pervaporation and (b) adsorption.
have an appreciable boiling point difference from both water and IPA. Further infor-
mation on solvent selection for liquid–liquid extraction is available (34,35).
Different MSAs may lead to completely different separation systems
designs. The systematic generation procedure is demonstrated for two potential
solvents, hexane and methylene chloride.
Water–IPA–Hexane System. The residue curve map for the ternary
system exhibits three distillation regions and a two-liquid-phase region (Fig. 6).
Each pure component is the high boiling node in a different distillation region.
Thus, from the feed mixture a distillation boundary must be crossed in order
to obtain pure IPA. The compositions of interest in this diagram include the
feed mixture (60% water, 40% IPA), the two products (in different distillation
regions), and potential MSA compositions, eg, pure hexane (high boiling node
in region III), the ternary azeotrope (low boiling node in regions I, II, III), and
points on the binodal curve, particularly tie-lines passing through azeotropes.
Since the binary azeotropes are saddle points (and thus hard to regenerate by
distillation), they are probably not good compositions to pick for an MSA. With
the limited information available at this point, pure hexane is a possible MSA.
This is a decision point. Another MSA composition may lead to a different flow
sheet. Moreover, if may be found that pure hexane is not required.
Without revisiting the kinetic-based methods already examined for the
binary case, several strategic operations can be used to cross the distillation
boundary between region I and II including Mixer 1, Extractor 2, and Decanter
3. The decantation strategy could be implemented either with a single feed,
which must be in the liquid–liquid region, or by adding an MSA, which puts
the overall mixture in the liquid–liquid region.
The feed compositions and products of each of these strategic separations
remain ill-defined. The unspecified IPA–water mixture, the input to each stra-
tegic separation, could be, but is not necessarily the original feed composition.
The MSA composition (pure hexane in this case) is such that one of the products
of the strategic separation is in region II (ie, the strategic separation crosses
the distillation boundary). Two opportunistic distillations from the original
feed mixture, Fractionators 4 and 5, are also possible. Since the feed is binary,
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 763
IPA
B9
Region
II
IPA
B9
Region
II
Azeo D5
D5 1 D9 O7
Feed 7
M1 Azeo Feed M1
D9 5 9 A7
Region
A7 Region I Azeo O7
III
B5 B5 B9
(c) (d)
Fig. 6. Isopropyl alcohol dehydration where Azeo is azeotrope, (– – –) are tie lines,
and the shaded areas represent the region of two liquid phases. Initial dehydration
(a) residue curve map and (b) flow sheet, and evolved IPA dehydration (c) residue curve
map and (d) flow sheet. The in b and d represent a two-phase decanter. (See text; see
Table 7.)
these separations are identical and are the same as the preconcentrator pre-
viously discussed. (See Table 7).
Following step 4 of the algorithm and consulting Table 2, the first opera-
tion should be Fractionator 5, an opportunistic separation that reaches a
product. None of the strategic separations is able to reach a product composi-
tion. Application of Fractionator 5 to the Feed produces the desired water
product as underflow and a water–IPA mixture (D5) as distillate. Stream D5
requires further processing to obtain the IPA product. At this point no other
opportunistic separations can be applied, but stream D5 is a possible input to
any of the three previously identified strategic operations. Arbitrarily picking
the mixing strategy, sufficient hexane must be added to D5 to produce a compo-
sition in region II.
764 SEPARATIONS PROCESS SYNTHESIS Vol. 1
The strategy for boundary crossing has now been implemented. However,
by the addition of the hexane another critical feature has been created. Hexane
must be regenerated, but it is in a different distillation region than the only
remaining unprocessed stream (M1). In this case, the possible boundary crossing
strategic operations are Mixer 6 and Decanter 7. Two opportunistic distillations,
Fractionators 8 and 9, can also be applied to M1 (note decantation is also a pos-
sible opportunistic separation).
Again referring to Table 2, the next separation should be Fractionator 9,
which produces the IPA product directly. If the distillation is driven far enough
stream D9 will be in the two-phase region and will be an appropriate input into
the strategic separation Decant 7. Decanter 7 yields phases in regions I and III.
Alternatively, D9 could be mixed with a sufficient amount of some stream
(unknown at this time) in region III to bring the overall mixture into region
III. Either strategy could be selected, but the decant is probably better as it
can be accomplished without the addition of another stream. The organic
phase of the decanter, O7, is in region III and needs to be further purified
to obtain pure hexane. Two opportunistic separations, Fractionators 10 and 11,
are possible.
Selection of Fractionator 11 gives pure hexane, which can be recycled
to Mixer 1. The distillate D11, however, is a problem. It cannot be distilled
because of its location next to a distillation boundary. It is outside of the two
phase region, so it cannot be decanted. In essence, no further separations are
possible. However, using the rules of Table 3, it can be mixed into the MSA
recycle stream without changing the operation of Mixer 1 appreciably. However,
as both outlet streams are mixed together, Fractionator 11 is not really needed.
The mixture of hexane and IPA, O7, could have been used as the MSA composi-
tion in the first place.
The final loose end in the process is the aqueous decanter product, A7. The
hexane must be removed before the mixture can be sent to waste water
treatment (ie, accepted as a water by-product). Two opportunistic separations,
Fractionators 12 and 13, are possible. Selection of Fractionator 13 gives pure
water underflow, and a distillate similar to D5. Distillate D13 can be recycled
back and mixed with D5 without affecting the operation of Mixer 1. All streams
have now been processed and the flowsheet produces both desired products
(Fig. 6b).
Although the flowsheet is complete, it does contain some redundant separa-
tion that adds to the capital cost. With the basic structure defined, the next step
is to consider some evolutionary improvements. First, Fractionator 13 performs
essentially the same job as Fractionator 5, ie, concentrating a mixture of water
and IPA. Fractionator 13 can be taken out and A7 can be recycled directly to the
feed of Fractionator 5. Although the flowrate to Fractionator 5 increases, its basic
separation function does not change. As the bubble point of A7 is somewhat
higher than that of Feed F, A7 is probably best introduced as a separate feed
lower in the column. Second, using pure hexane as the MSA was unnecessary.
Fractionator 11 can be eliminated and O7 used as the MSA composition. Finally,
considering the bubble point and amount of O7, operations Mixer 1, Fractionator
9, and Decanter 7 can be rearranged as a distillation-decanter combination with
identical overall material balance lines and function. The evolved flowsheet is
766 SEPARATIONS PROCESS SYNTHESIS Vol. 1
IPA
B2
D2 O3
Region 3
Azeo II
Feed M1 A3
2
Water
Feed Region I B2
IPA
M1 (b)
A3
Water Azeo D2, 03
Methylene chloride
(a)
IPA
B4
M2
D1 O3 D4
Region 3
Azeo D1 II
Feed M5 A3
1 4
A3
Region I
Feed M5 B1 B4
M2 O3 Water IPA
(d)
B1
Water Azeo D4
Methylene chloride
(c)
Fig. 7. Isopropyl alcohol dehydration using methylene chloride where Azeo = azeotrope,
the shaded areas represent the region of two liquid phases, (– – –) are tie lines, and
, two-phase decanters: (a) residue curve map and (b) flow sheet, and alternative methy-
lene chloride (c) residue curve map and (d) flow sheet. (See text.)
A similar equation can be defined for the transformed vapor fractions, Yi.
The reference components are a subset of the reacting species and are used to
indicate the extents of reaction in terms of the mole numbers of the reference
components. The number of reference components is equal to the number of inde-
pendent reactions. Selection of the proper reference component is system depen-
dent. Selection rules have been elucidated (38) and the transformed coordinate
approach has been applied to a variety of reactive separations and phase equili-
bria including reactive distillation, adsorption, and solid–liquid, liquid–liquid
equilibria (39–41).
Transformed compositions are particularly useful for visualization of
reactive distillation systems. It has become well established that phase and
distillation region diagrams (DRD) in transformed compositions are entirely
analogous to those for nonreactive systems, exhibiting the same material balance
properties. Thus, reactive DRD can be used to assist in design feasibility and
operability studies in much the same fashion as outlined above for non-reactive,
nonideal systems (3,4,42).
When an equilibrium reaction occurs in a vapor–liquid equilibrium system,
the phase compositions depend not only on the relative volatilities, but also on
the consumption or production of species. The condition for azeotrope formation
(vapor and liquid-phase compositions are equal for two or more components) no
longer holds true in a reactive system and must be modified to include reaction
stoichiometry:
y1 x1 yi xi X
n
¼ for all i ¼ 2 to n 1 and vT ¼ vi
v1 vT x1 vi vT xi i¼1
where xi, yi ¼ mole fractions in the liquid and vapor, respectively, and vi ¼
stoichiometric reaction coefficient (negative for reactants, positive for products).
A reactive azeotrope occurs at a temperature where the rate of mass
exchange between phases and the rate of reaction for each component are such
that no net change in composition occurs in either phase. Whether one assumes
equilibrium reaction behavior or kinetically controlled reaction behavior (very
much dependent on column holdup, catalysis, and other device implementation
parameters) affects the existence and location of reactive azeotropes. Reactive
azeotropes always lie on the chemical equilibrium line. Since the boiling points
of the mixture are affected by the column pressure, and the equilibrium constant
is influenced by temperature (and thus column pressure), the location of the reac-
tive azeotrope can be influenced by changing the column pressure.
A residue curve map in transformed coordinates for the reactive system
methanol–water–acetic acid–methyl acetate is shown in Fig. 8. The water–
methyl acetate azeotrope has disappeared (water and methyl acetate can back
react to methanol and acetic acid), while the methyl acetate–methanol azeotrope
Vol. 1 SEPARATIONS PROCESS SYNTHESIS 769
Fig. 8. Reactive residue curve map for a methyl acetate system. (Reprinted from Ref. 43,
with permission from Elsevier.)
remains. Only those azeotropes containing either all the required reactants or
products will be altered by the reaction. The system exhibits only one distillation
region in which the methanol–methyl acetate azeotrope is the low boiling node
and acetic acid is the high boiling node.
4.2. Flow Sheet Construction. Because of the analogous properties of
reactive and nonreactive residue curve maps, the opportunistic-strategic syn-
thesis methods described earlier can be extended to reactive system synthesis.
There has been limited progress in this direction (44,45). The elucidation of an
elegant reactive separation solution to a complex synthesis problem is often an
appealing prospect. However, there must be good reasons for combining reac-
tion and separation. Reactive separations introduce added design and operability
difficulties that must be weighed against any perceived benefits. In practice, the
operability space of a reactive separation is often quite limited compared to
sequential reaction and separation. For example, reactive distillation is only
applicable when the mixture boils under temperature and pressure conditions
favorable for the reaction. Other minimum compatibility criteria must also be
met. Other unfavorable situations for reactive distillation are the presence of
appreciable solids, operation at supercritical conditions, ie, where no separate
vapor and liquid phases exist, operation at very high or low temperature and
pressure conditions, ie, vapor and liquid phases both must be present and appro-
priate heating and cooling media available.
Once the minimum compatibility criteria have been met, then it is neces-
sary to evaluate whether simultaneous separation of components from the reac-
tion mixture is beneficial. Typical favorable situations include (1) when the
reaction must be carried out with a large excess of one or more of the reactants;
(2) when a reaction can be driven to completion by removal of one of the products
as formed, ie, equilibrium limited reactions; (3) when the desired product
770 SEPARATIONS PROCESS SYNTHESIS Vol. 1
(a) To have reflux ratio 100, then should have aCD > 1.1
(b) To have reflux ratio 10, then should have aCD > 1.25
(c) To have reflux ratio 5, then should have aCD > 1.45
(d) To have reflux ratio 2.5, then should have aCD > 2.1
7. For systems in which the reactants are all intermediate boiling and the products are
the highest and lowest boiling components, then a Keq of at least 105 is required for
potential economic viability
8. For systems in which the products are all intermediate boiling and the reactants are
the highest and lowest boiling components, then a Keq of at least 100 is required for
potential economic viability
9. For systems in which the desired product is the highest boiling component, then a Keq of
at least 0.1 is required for potential economic viability
10. For systems in which the desired product is the lowest boiling component, then a Keq of
at least 103 is required for potential economic viability
11. For systems in which the by-product is the lowest boiling component and the desired
product is intermediate boiling, then a Keq of at least 103 is required for potential
economic viability
12. For systems in which the by-product is the highest boiling component and the desired
product is intermediate boiling, then a Keq of at least 102 is required for potential
economic viability
methyl acetate and methanol are saddles. Examining the residue curve map for
this system, the strategic separation of extractive distillation can be used to draw
methanol down the column and reject methyl acetate out the top. Solvents, eg,
ethylene glycol, water, and acetic acid, are potential extractive distillation sol-
vents. This methyl acetate may be wet with water, so if desired an additional
opportunistic separation can be used to take the methyl acetate–water azeotrope
overhead (recycled to the extractive distillation column) and pure methyl acetate
out the bottom. Methanol for recycle is recovered by distilling the underflow of
the extractive distillation column.
The bottoms from the effluent splitter column containing the acetic acid can
be recovered by using a combination of strategic separation separations for
pinched systems. Adding an extraction agent, eg, ethyl acetate, to concentrate
the acetic acid, followed by heterogeneous azeotropic distillation to recover the
acetic acid and solvent finishes the flow sheet (Fig. 9).
772 SEPARATIONS PROCESS SYNTHESIS Vol. 1
Acetic Acid
Methanol
Catalyst
Methyl
Acetate
Solvent
Azeo Water
Water Heavies
Solvent/Entrainer
Water
Methyl Acetate
Distillation
Acetic Acid Task
Extractive
Distillation
Task
Catalyst
Reactive
Distillation
Task
Reaction
Task
Reactive
Distillation
Methanol Task
Combined
Distillation
Task
Water
BIBLIOGRAPHY
‘‘Separations Process Synthesis’’ in ECT 4th ed., Vol. 21, pp. 923–962, by S. D. Barnicki
and J. J. Siirola, Eastman Chemical Company; ‘‘Separations Process Synthesis’’ in ECT
(online), posting date: December 4, 2000, by S. D. Barnicki and J. J. Siirola, Eastman
Chemical Company.
CITED PUBLICATIONS
SCOTT D. BARNICKI
CRAIG A. HOYME
JEFFREY J. SIIROLA
Eastman Chemical Company
776 LIQUID–LIQUID EXTRACTION Vol. 1
LIQUID–LIQUID
EXTRACTION
1. Fundamentals and Fields of Application
Only a general outline is given in this article. For further details, the reader
is referred to books (1–7) and the Proceedings of the International Solvent
Extraction Conference (ISEC) (8–17). Data of liquid – liquid equilibria can be
found in (18,19) and (20).
2. Thermodynamic Fundamentals
Only ternary systems with miscibility gaps are suitable for extractions,
and the boundary line between the liquid single-phase region(s) and the
two-phase region is particularly important. This boundary line is called the
binodal curve (Fig. 3). Every point on the binodal curve (except the plait
point in systems with a complete binodal curve, see below) is in equilibrium
with another binodal point. The lines which connect points in equilibrium
with one another are called tie lines. All mixtures which correspond to a
point on the tie line separate into two phases. The composition of these
phases is given by the end points of the tie lines and their quantities are
given by the lever rule. The equilibrium data have to be determined experi-
mentally in each individual case. Lines parallel to the triangle sides CB and
AC respectively, drawn through related points of equilibrium on both
branches of the binodal curve, produce an auxiliary line, formed by the inter-
sections of the parallels, and enable to construct any tie line as demonstrated
in Figure 3 A.
Several types of triangular diagrams are shown in Figure 4. In every sys-
tem with a complete binodal curve (closed diagram, Type II in Fig. 4) the end
points of the tie lines draw closer and closer together until they finally coincide
at the plait point. The plait point on the binodal curve represents a single phase
which does not split into two phases.
Diagram I is called an open diagram in which B is not fully miscible with
the substance to be dissolved. The capacity of such a solvent is correspondingly
low and its selectivity is usually high (see Section 2.4).
780 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 4. Various types of phase equilibria in ternary systems. Type I: Open system,
miscibility gap between the key component of the solvent (B) and solute C. Type II: Closed
system, C is completely miscible with the key component of the feed (A) and the key
component of the extraction solvent (B). Type III: Open system, miscibility gap between
A and C Type IV: System with formation of three phases.
Fig. 5. Representation of phase relationship of the ternary system heptane (A), diethy-
lene glycol (B), benzene (C), at 20 8C. All data refers to mass fractions. I) Triangular
diagram; II) Equilibrium diagram; III) Distribution diagram; IV) Selectivity diagram.
782 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 7. Representation of phase relationship of the ternary system water (A), butyl
acetate (B), phenol (C), at 20 8C. All data refers to mass fractions. I) Triangular diagram;
II) Equilibrium diagram; III) Distribution diagram; IV) Selectivity diagram.
784 LIQUID–LIQUID EXTRACTION Vol. 1
To determine phase equilibrium and therewith the tie lines, ie, the points
which are in equilibrium with one another, a mixture from the two-phase region
is placed in a controlled-temperature, shaking cylinder. Equilibrium is estab-
lished by shaking. The phases are allowed to settle and are separated by draining
off into different vessels. Then their quantities are determined and the physical
variable measured during determination of the binodal curve is then measured
for both phases. This immediately gives the two tie line points. Furthermore,
both phases can also be analyzed by well known chemical methods, such as
gas or liquid chromatography and titration. The correctness of the measurement
can be checked by connecting the two tie line points in the triangular diagram.
The connecting line must pass through the point of the original mixture, and this
point must divide the length of the tie line in inverse ratio to the quantities of the
two phases.
If a system of more than three components has to be measured then several
components are combined into one group in order to arrive at a ternary system
again. Determination of a binodal curve would still be possible in such a case but
the two phases into which the test mixture separates are no longer identical with
the mixtures which were obtained during the determination of the binodal curve.
It is therefore better to allow the mixture to reach equilibrium conditions in a
controlled-temperature, shaking cylinder, separate it and then analyze the
phases using some preplanned method.
The time consuming step in the determination of equilibria is the prepara-
tion of the experiment, such as purification of the substances, setting up the cali-
bration curves, etc. Therefore it is senseless to restrict the number of measured
points and use interpolation methods. To increase the accuracy a material bal-
ance is set up for each component and the measured points are adjusted by draw-
ing the curves from Figures 5–7.
2.3. Empirical Correlations of Phase Equilibria. In ternary systems,
the use of empirical correlations to represent the binodal curves and distribution
equilibria of component C is of advantage (28). The mutual solubilities of
the liquid phases are usually determined by a laboratory turbidity titration.
The binodal curve is then determined by extrapolating the measured values
(see Section 2.2). In Type I systems with an open miscibility gap, the binodal
curve can be simply described by parabolic equations:
x ¼ D0 þ E0 x0A þ F 0 x02
A
where x0 A is the mass fraction of A in the A-rich phase, x00 B the mass fraction of
B in the B-rich phase and D0 , D00 , E 0 , E 00 , F 0 and F 00 are coefficients.
In systems with a closed miscibility gap (Type II), the parabolic equations
are only valid for the lower parts of the curve; they cannot be used to represent
the entire binodal curve. The best equation for the closed binodal curve is that
proposed by HLAVATY (29). The correlation is derived from a polynomial equation
and has the advantage that the binodal curve ends exactly at the binary solu-
bility points. Its disadvantage is, however, that the lower, in practice virtually
linear, regions of the binodal curves are represented in wave form as a result
786 LIQUID–LIQUID EXTRACTION Vol. 1
of the polynomial equation. The Hlavaty equation should, therefore, be used with
caution.
The equilibrium distribution of solute C in the A- and B-rich phases is gen-
erally determined independently from the binodal curve by means of laboratory
experiments. The mass fraction xC and yC of the two phases in thermodynamic
equilibrium are thus obtained and determine the positions of the end points of
the tie lines on the binodal curves (see Section 2.2). Equations for correlating
and interpolating the end point of the tie lines are described in the literature,
they are primarily exponential equations such as those proposed by OTHMER
and TOBIAS, and HAND (1). However, all exponential equations generally fail
when the concentration of C is very low. To allow a more accurate estimation
of distribution equilibria and extraction processes at the low raffinate concentra-
tions that are of industrial importance, BENDER and BLOCK (28) proposed the fol-
lowing correlation for the distribution ratio K:
K ¼ G1 þ G2 x0C þ G3 x02 03
C þ G 4 xC
The coefficients Gi must be fitted to the measured values of the tie lines x
and y. Although the equation does not generally permit extrapolation of the
curve above the measured values to the plait point, the distribution coefficient
K can be obtained accurately up to x and y ! 0.
The empirical correlations are simple and clear. They can be calculated
quickly with appropriate computer programs and give an accurate representa-
tion of the measured liquid – liquid equilibria. However, these advantages are
accompanied by several disadvantages. The correlations can only be used with
ternary systems and do not allow extrapolation to higher multicomponent sys-
tems. Extensive laboratory work is required to determine the mutual solubilities
of components A and B as a function of the concentration of C as well as the equi-
librium distribution of component C. Systematic measurement errors cannot be
detected because the experimental values are empirically correlated.
2.4. Selection of Solvent. The selection of the most suitable solvent is
one of the hardest problems of the extraction, and the decision made has a great
influence on the economy of the process. The solvent must fulfill two basic
requirements:
gaps with both water and hexane. Acetone, ethanol, and dioxan which have a
similar polarity are miscible with both water and hexane. The molecular form
and the size of the molecule also contribute to solvent behavior. The behavior
of solvents towards hydrocarbons has been investigated particularly intensively
(33,34). It is also possible to obtain effects from mixing two solvents which could
not be predicted from the properties of the mix components (35,36).
The most important property of a solvent is its selectivity. This indicates
how much better the solute is dissolved than the A-rich phase. As can be seen
from Figures 5–7 III the selectivity is strongly dependent on concentration.
In addition to selectivity the capacity of a solvent is also important. This is
defined by the distribution coefficient of C between the B-rich and A-rich phases
and determines the quantity of solvent which has to be recycled. When testing
different solvents for a specific extraction task it is often found that solvents
with higher selectivities have lower capacities and vice versa. The greater the
throughput of a plant the more important the capacity of the solvent is for the
economy of the process. If there are several possible solvents for a process then
their behavior with respect to selectivity and capacity can be tested with a gas
chromatograph without starting with time-consuming equilibrium measure-
ments (37).
In addition to these basic requirements there are also other criteria which
must be taken into consideration when selecting a solvent and which sharply
reduce the number of possible compounds. Most important is the availability of
the solvent in commercial quantities at an economically acceptable price. It is
then necessary to find data concerning boiling points, density, viscosity and —
where available — information about solubility (38,39).
The boiling point of the solvent is important because the extracted
component is usually separated from the solvent by distillation and the
boiling point difference determines the reflux ratio in this distillation. If
the solvent boils at a higher temperature than the solute, this solute is dis-
tilled overhead. If it is necessary to recycle a great deal of solvent then the
solvent has to be purified by distillation in a subsequent step; the specific
heat which is required for heating the solvent to its boiling temperature
must also be taken into account. Too great a temperature difference is then
a disadvantage.
The density of the solvent is important for obtaining a sufficiently large dif-
ference in densities between the two liquid phases for the separation process.
Chlorinated compounds are suitable if a solvent with a high density is to be
used, and particularly low-density compounds are found among the saturated
hydrocarbons and ethers.
The thermal stability of the solvent can be highly significant. The
researcher hardly ever finds the answer to this question in the literature and
is directed to carry out his own measurement. It is necessary to know (1) how
much solvent make-up will be needed in the process; (2) whether a regeneration
process for continuous removal of decomposition products needs to be installed.
Very careful tests must be carried out to find out whether decomposition pro-
ducts can contaminate the extract or have a corrosive effect on the construction
materials of the plant. Sometimes the toxicological properties and biological
degradation of a solvent have to be considered (40).
788 LIQUID–LIQUID EXTRACTION Vol. 1
The viscosity should be as low as possible as this favors mass transfer and
phase separation. However, this property is less significant than those men-
tioned above. If the possible solvents have been reduced to a manageable number
after testing all the aspects then the individual properties can be evaluated with
a point system to discover the best solvent.
The surface tension is another important parameter because it influences
the energy demand necessary to produce a suitable drop dispersion. High surface
tension gives rise to formation of stable drops and needs a high energy input. A
solvent with a high surface tension decreases the danger of producing an emul-
sion. The surface tension is dependent on the concentration of the transfer com-
ponent. Further properties such as suitability for regeneration, disposability,
costs, combustibility, toxicity and corrosiveness also have to be taken into
account to select a suitable solvent.
Bound up with solvent selection is the choice of the dispersed and the con-
tinuous phase, respectively. Direction of mass transfer and phase ratio have to be
taken into account. The direction of mass transfer influences the surface tension
and in this way can promote or hamper the coalescence of drops. The phase to be
dispersed should be the one which impedes the drop coalescence to produce
stable drops, and the one with the higher flow rate to obtain a mass transfer
area as large as possible.
2.5. Calculation of Theoretical Stages (! Absorption, Chap. 4.; !
Distillation and Rectification, Chap. 4.). Two possibilities exist to calculate
the mass transfer performance of extractors. The first one uses the concept of
theoretical stages and is based on the assumption that concentration equilibrium
is established in every section of an extractor. This assumption can be fulfilled
only in the mixer – settler with long contact times of the phases involved. In
this concept the total height of an extraction column is given by the product of
the number of theoretical stages (NTS) and the height equivalent of a theoretical
stage (HETS). NTS follows from a material balance in combination with the dis-
tribution equilibrium; HETS depends on fluid dynamics and mass transfer in the
columns and comprises all real conditions as for instance the conditions of phase
mixing. Therefore this concept is preferred only for rough estimations in practice
because it is easy to handle and sufficient for preparation of laboratory experi-
ments. In the following sections, the method will be presented by aid of equili-
brium diagrams.
The second possibility is a stream concept. The column height required for a
given throughput and concentration gradient of phases along the column can be
calculated from height integration of a set of differential equations, derived from
a differential material balance and the local mass transfer between the both
phases. The stream concept supplies the course of concentrations along the col-
umn and is suitable for analysis of the dynamic behavior of a column and for tak-
ing into account the drop population inside the column. It is preferred until now
for scientific investigations, eg, for stationarily operated pulsed sieve tray col-
umns (45,47–49,51,62) and for instationarily operated rotating extractors
(65,68).
Selection of Calculation Methods. The conditions for an applicable
method of calculation are set by the problem to be solved, by whether simpli-
fied assumptions are permissible and by the main influencing variables of an
Vol. 1 LIQUID–LIQUID EXTRACTION 789
extraction problem. They are, therefore, different. The following points should be
considered when selecting a method of calculation:
then the compositions of the phases can be expressed as loads. If a material bal-
ance is made as shown in Figure 8, the two loads Yc and Xc in a given column
cross section can be related as follows
m
_F
YC YC; S ¼ ðXC XC; R Þ
m_s
Fig. 9. Construction of steps in the operating diagram to obtain the number of theoreti-
cal stages (NTS).
Figure 9. The result of the use of the steps often means that the load obtained
from C in the raffinate (XC, R) does not correspond to the required load (XCR, req).
Figure 8 also shows the position of the balance line for the minimum extrac-
tant ratio Vmin
Vmin ¼ m
_ s; min =m
_F
The desired enrichment of the feed with C using the minimum amount of extrac-
tant ṁs, min can only be achieved with an infinite number of stages.
Method 2. Figure 10 shows the equilibrium diagram for the system
toluene – ethylene glycol – water as a double logarithmic plot. This type of plot
Fig. 10. Determination of the number of stages for a variable distribution ratio and a
high degree of depletion.
Vol. 1 LIQUID–LIQUID EXTRACTION 791
is useful if the distribution ratio varies considerably, despite low loading, and the
concentration range considerably exceeds the desired raffinate purity. In this
case the operating line is at an angle of 458.
m
_F
logðYC YC; S Þ ¼ log þ logðXC XC; R Þ
m
_S
The position of the straight lines is determined by the extractant ratio V. The
number of required theoretical stages can be determined by constructing a series
of steps.
Method 3. If the distribution ratio K is a constant, further simplification
is possible; in analogy with the Colburn equation (! Absorption, Chap. 4.2 ), the
number of theoretical stages NTS can be calculated:
h i
X Y
log 1 þ ðE 1Þ XC;C; RF þ XC;C; RS
NTS ¼ 1
logE
E ¼ Km
_ s =m
_F
If NTS is known and the degree of depletion XC, R/XC, F is desired, the follow-
ing equation is obtained:
YC; S
XC; R ðE 1Þ 1 þ XC; F
¼
XC; F EðNTSþ1Þ 1
This equation can only be solved iteratively; a graphical solution is shown in
Figure 11.
Method 4. If none of the above-mentioned conditions is satisfied, and the
extraction problem can be limited to consideration of the three key components,
the calculation can be carried out with the aid of the triangular mass fraction
diagram. This method is based on the mixing rules described in Section 2.1.
The specified compositions of the feed F and solvent S streams are plotted
on the phase equilibrium diagram for the ternary mixture (Fig. 12). For material
balance of the feed and product streams, the column can be assumed to be a
simple mixing vessel. The composition M of the mixture lies on the tie line FS;
the position of M on the tie line is determined by the ratio ṁS : ṁF = MF : MS. The
compositions of the raffinate R and extract E streams leaving the column are
obtained from the intercepts of the tie line with the binodal curve. The raffi-
nate composition is determined by the desired mass fraction of C in the raffinate
(xC, R). The composition of the extract E is given by the point at which the
extended line RM intersects the binodal curve. The ratio ṁR : ṁE is obtained
by applying the lever rule. The quantities and the compositions of the incoming
and outflowing streams are thus determined.
The required number of stages in a countercurrent column are obtained by
using the pole point procedure. The procedure can be explained by imaging a
792 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 11. Dependence of the theoretical number of stages, the degree of depletion, and the
yield on the extraction factor E.
stream P to maintain the material balance. Mixing of this stream with the raffi-
nate stream R gives the composition of the solvent stream S. To maintain the
material balance the extract on the other hand must be composed of the feed
stream F and the imaginary stream P. The imaginary stream P can thus be
regarded as a common mixing partner for F and R to maintain the compositions
of E and S. The composition of the imaginary stream P is represented by a point
outside the triangular diagram, the pole point. The imaginary mass fractions of
the three components in the pole point can be > 1 or negativ. If the material bal-
ance boundaries for the individual stages are taken as shown in Figure 13, the
above procedure gives the stage extracts E2, E3, etc. from the mixture of the ima-
ginary stream P and the individual stage raffinates in the proceeding stages R1,
R2, etc. The composition of the raffinates R1, R2, etc. are linked to the correspond-
ing extracts E1, E2, etc. via the distribution ratios (tie lines).
Evaluation of Stage-Number Calculation for Process Design. In order
to determine the operating conditions and the dimensions of the equipment
used for extraction, the results of several stage-number calculations must be
plotted in a diagram showing the relationship between the solvent circulation,
the number of stages, and the residual content of solute in the raffinate. Such
a diagram is shown in Figure 14 for mixture II (see also Fig. 6). It represents
the extraction of a 10 % solution of caprolactam in trichloroethylene with
water as the extraction solvent. The favorable operating region is usually located
in the region which shows the most pronounced curvature. Other variables
(eg, the effect of temperature and the residual loading of the solvent with solute)
may have to be taken into consideration.
3. Apparatus
designs use the principle to disperse one of the two liquids into the other in order
to enlarge the contact area for mass transfer and to influence the drop generation
by energy input such as mixing or pulsing. After the resulting droplets have been
contacted with the continuous phase long enough to allow the extractable com-
ponent(s) to migrate to the dispersed phase, both liquids have to be separated
again. Densities will define the flow directions of the phases: The heavier liquid
flows from top to bottom, the lighter one from bottom to top. Accordingly, the
fluids have to be transported to the apparatus and removed by pumps.
In the following chapter, a few typical designs have been selected from the
large number of equipment designs actually used in industry. A more detailed
treatment can be found in literature, especially in (5) and (6).
Columns without Energy Input. The simplest extractor is the spray col-
umn (Fig. 15 A). The spray column achieves only very small numbers of separa-
tion stages and is rarely used nowadays. If the column of Figure 15 A is filled
Vol. 1 LIQUID–LIQUID EXTRACTION 795
Fig. 15. Columns without energy input (LIC = level indicating controller). A) Spray
column (light phase dispersed); B) Packed column (heavy phase dispersed); C) Sieve plate
column (light phase dispersed).
mainly with the heavier liquid, the lighter liquid breaks at the feeding point into
drops that rise in the continuous liquid. Drops reaching the interface are kept at
this position by controlling the bottom discharge (LIC). There, they will then gra-
dually coalesce and form a continuous layer floating on top of the continuous
liquid.
The reverse procedure is possible as well, as shown in Figure 15 B. Here, the
distributor is located at the top and the water interface, called the principle
interface, is controlled at the bottom of the column by filling it mainly with the
lighter liquid. In this case, the heavier liquid is the dispersed one, broken into
droplets descending in the lighter liquid. The principle interface can be adjusted
at any height (between top and bottom) of the column as needed, thus the heavier
liquid being dispersed above and the lighter liquid below the principle interface.
Countercurrent columns are equipped with settling zones at the top and the
bottom, with coalescing aids provided if necessary. The liquids should never be
fed into the column in these settling zones because this would cause hydrody-
namic instability of the column thus reducing the operational performance.
The dispersed liquid often cannot be selected from theoretical considerations,
but only by aid of experiments in a pilot column using the real material system.
Packed columns. (structured and unstructured) and sieve plate columns
have been adopted from distillation technology (see Distillation). In packed
columns drop formation occurs on entry of the dispersed phase in the packed
bed (Fig. 15 B). The drops rise through the packed bed without significant
further dispersion. However, mass transfer takes place mainly during the
formation of a new interfacial area. This explains the low efficiency of packed
columns.
In a sieve plate column with downcomers, as shown in Figure 15 C, the
lighter dispersed phase coalesces beneath each plate to give a continuous
layer. Its buoyancy presses the liquid through the holes of the sieve plates, pro-
ducing jets which disintegrate into drops. The continuous liquid flows through
796 LIQUID–LIQUID EXTRACTION Vol. 1
the downcomers from stage to stage and then across the plates. In the case of low
interfacial tension the layers of coalesced phase are too small for a symmetrical
operation of the plates. Then it is advantageous to use dual flow plates without
downcomers (57).
Sieve plate columns have the advantage over packed columns that new
drops are formed at each tray. However, the drop load of the plate is often irre-
gularly distributed, especially if the column diameter is large (more than 1 m).
This is the reason why sieve plate columns can reach efficiencies of 10 – 15 %
only in rare favored cases. Moreover, since columns are generally built with a
plate spacing of 0.3 m, even simple separations require considerable overall
heights.
Mass transfer between the two liquids can be substantially improved by
application of additional mechanical energy in form of superimposed agitation
or pulsation, resulting in higher turbulence near the drops and enlarged mass
transfer area due to smaller drop sizes.
Pulsed Columns. The simplest design of pulsed column is the pulsed
sieve plate column (Fig. 16A). A piston pump with an electrical or an air-pressure
drive causes the entire liquid content of the column to vibrate. Both phases are
alternately forced through the holes in the rhythm of the pulsation. The down-
comers used in the conventional sieve plate column are thus superfluous. The
dispersed phase does not coalesce above or below the plates to give a continuous
layer. The pulsation shock flings the drops off the plate faster than would be
possible as a result of only the density difference. This leads to a fine, uniform
drop distribution and hence enlarges the mass transfer area. Applications are
limited by the physical properties of the substances concerned and high through-
puts. Systems which emulsify easily cannot be processed in pulsed columns just
as sticky and greasy liquids. For very large throughputs (column diameter
Fig. 16. Pulsed and reciprocating-plate columns (Light phase dispersed). A) Pulsed
sieve-plate column; B) Reciprocating-plate column (Karr).
Vol. 1 LIQUID–LIQUID EXTRACTION 797
> 0.6 m), the pulsed volume and pulsing power become very high; however, pul-
sing units for large column diameters have been developed.
The separation effect of the column decreases with increasing diameter,
this has not yet been fully investigated. The plate spacing is 50 – 100 mm. Larger
plate spacings favor backmixing and hence reduce the efficiency of the column.
Smaller plate spacings lead to a strong decrease of the throughput. Sieve plates
have cylindrical holes and a triangular pitch. The relationship between the frac-
tional free area j, the hole diameter di and the pitch p is given by
pffiffiffiffiffiffiffiffiffiffiffiffi
j ¼ pd2i = 12 p2
For usual liquid – liquid systems sieve plates with hole diameters from 2
to 4 mm resulting in free areas of 23 to 40 % are recommended (6, Chap. 10),
the lower value of 23 % is used for high and the higher value of 40 % for low
interfacial tensions. Free areas less than 20 % are used for liquids with higher
viscosities (58).
The plate material has a pronounced influence on drop formation and hence
on the separation efficiency. The plate material and phase selection must be
matched in such a way that the plates are wetted by the continuous phase but
not by the dispersed phase, because this would lead to strong coalescence effects.
If, for example, an aqueous phase is to be dispersed into drops, metal trays are
unsuitable, but plastic plates are satisfactory. The sharp edged side of the sieve
plate, which is due to manufacturing, should exactly face the flow direction of the
dispersed liquid so that drops will shear at the edges of the holes. Gaps between
the sieve plate edge and the column wall should be avoided by precise construc-
tion or by applying edge-sealing elements made of materials which are wetted by
the continuous liquid.
In most cases, piston pumps or diaphragm pumps are used as the pulsators.
The pulsing frequencies are between 30 and 150 min1, with pulsing amplitudes
from 5 – 15 mm (relative to the column cross section). (Pulsing amplitude is
defined as the distance between upper and lower center of the pulsing motion
in the active part of the column.) The dispersed phase has to be fed into the col-
umn via a distributor. The mean drop diameter of the predispersed swarm
should be larger than the mean drop size in the active part of the column, other-
wise the hydrodynamics of the column would mainly be affected by the distribu-
tor. The holes in the distributor should be very sharp-edged to avoid very fine
drops. Even in packed columns, the separation effect can be substantially
improved by pulsing of the continuous phase, at the expense of a reduced
throughput. The increased performance results from greater shear forces caus-
ing a reduction in drop size and an increase in interfacial area. As in distillation
columns, the phenomenon of maldistribution in randomly packed columns of lar-
ger diameter can be overcome by use of structured packings. These packings lead
to a good cross-sectional mixing and have a higher capacity, typically a total flow
rate of both phases of approximately 100 m3m2h1 compared to 30 m3m2h1 for
a randomly packed column. Structured packings are particularly suitable for low
interfacial tension systems or for small density differences; separation efficiency
is nearly independent of scale-up. Hydrodynamics and mass transfer perfor-
mance are well investigated (59).
798 LIQUID–LIQUID EXTRACTION Vol. 1
Pulsing Units. The use of pulsed columns is currently limited to small and
medium throughputs. The column diameter is limited by the pulsing unit. For
very large throughputs, the pulsed volume becomes so large that the use of a
piston pump as the pulsing unit causes difficulties. Consequently, pulsing
units have been developed which operate without a piston pump. The pulsator
developed by MISEK (Fig. 17) is based on a centrifugal pump. The pump (e) is con-
nected on both the suction side and the delivery side to pressure vessels which
are alternately connected to the column by means of a rotating valve (c). With
Fig. 17. Pulsator with rotating valve. A) Principle diagram of compression stroke;
B) Principle diagram of suction stroke; C) Technical design (Bronswerk) a) Suction
drum; b) Discharge drum; c) Rotating valve; d) Column; e) Centrifugal pump.
Vol. 1 LIQUID–LIQUID EXTRACTION 799
Fig. 18. Hydraulic pulsator. a) Pulsing piston; b) Working piston; c) Rinsing; d) Hydrau-
lic aggregate with radial piston pump.
the valve setting shown in Figure 17 A, the pump delivers the liquid into the
column via the discharge drum (b), taking the rate delivered from
the suction drum (a). After changeover of the valve (Fig. 17 B), the pump
delivers from the column via vessel (a) into drum (b). The pump thus always
delivers continuously in one direction. Reversal of the direction of flow in the
pulse line is effected by the change-over valve (c) in accordance with the
pulse frequency. A pulsator based on this principle has been developed by
Bronswerk/DSM (Fig. 17 C), and is now in operation for columns with a maxi-
mum diameter of 2.7 m.
For large pulsed volumes, the hydraulic pulsator from Wepuko (60) is also
suitable (Fig. 18). It is flanged directly to the lower bottom of the column end and
acts in the direction of the column axis. The pulsing piston is connected via a
common piston rod to a working piston, the hydraulic drive of the piston is a
rapidly reversible radial piston pump.
Reciprocating-Plate Columns. In another approach for solving the
problem of large throughputs, the plates, and not the liquid, are made to recipro-
cate, so that the pulsing pump is eliminated. This principle led to the develop-
ment of the reciprocating-plate column which can be regarded as the converse
of the pulsed sieve plate column. In the pulsed sieve plate column, the trays
are fixed and the entire liquid content of the column is caused to vibrate; in
the reciprocating-plate column, the plates are moved. In both cases, the liquid
undergoes an oscillating motion relative to the plate.
In the Karr column (Fig. 16 B), sieve plates with a free area of about 60 %, a
10 – 15 mm hole diameter, and a 25 –50 mm plate spacing, are mounted on a cen-
tral shaft. The complete pack is moved up and down, driven by an eccenter.
800 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 19. Rotating disk contactor (RDC). a) Rotating disk; b) Stator ring.
The relatively large free area of reciprocating plate columns allows much
greater throughputs than conventional pulsed columns. However, recent inves-
tigations showed that pulsed sieve plate columns can be equipped with larger
free areas, too. Both types of pulsed columns show very similar performances
and operational behavior if they are equipped with the same free areas (61).
Research on the established types of pulsed columns has been very intensive
since the 1970s.
Columns with Rotating Internals. Rotating Disk Contactor (RDC). In
the RDC (Fig. 19 A), stator rings are mounted at close intervals. Disks are fitted
between the rings to a rotating shaft. The rotating disks fling the drops of the
dispersed phase off centrifugally, onto the stator rings where they are braked
and deflected. The drop distribution is excellent and can reach the emulsification
limit as the rotational speed increases.
The RDC has a relatively simple construction with a large free cross-
sectional area. On the one hand, it reaches large throughputs, on the other
hand it favors backmixing of phases causing a decrease of efficiency. A large
range of throughputs is possible without remarkable decrease of efficiency.
The loading limit is generally lower than in pulsed columns. The RDC is unsui-
table for mixtures which tend to form emulsions and for products containing dirt.
The separation efficiency decreases sharply with increasing diameter. Neverthe-
less, successful RDCs of up to 3500 mm in diameter have been constructed. They
Vol. 1 LIQUID–LIQUID EXTRACTION 801
Fig. 20. Columns with rotating internals. A) Kühni extractor; B) QVF extractor.
are thus suitable for greater throughputs than pulsed columns and are prefer-
ably used to operate with large product flows when the demands on separation
performance are only low as for instance in the oil industry or in wastewater
treatment (42).
As in pulsed columns, the wetting behavior also influences drop formation
in the RDC, and should therefore be taken into account when selecting a mate-
rial for the column internals.
Kühni Extractor. The Kühni extractor (Fig. 20 A) uses a turbine as the
mixing element. Perforated partition disks separate the extraction stages and
reduce backmixing.
This apparatus is also used for performing reactions between two liquid
phases; the residence time can be adjusted via the free area of the perforated
disk.
QVF-Extractor. The QVF extractor RZE (Fig. 20 B) uses blade stirrer as
mixing elements. The partition disks have meander-formed metal stripes at the
inner circumference to reduce the phase dispersion by channeling of both phases.
The operational behavior is similar to that of the Kühni extractor (see Fig. 20 A).
Graesser Contactor. The Graesser contactor (Fig. 21) is another type
of rotating equipment. Here, not only the internals but the entire apparatus
rotate. Internal half-tubes are fitted as scooping elements in a horizontal
rotating cylinder. Due to the rotation, the half-tubes transport part of
the heavy phase into the overlying light phase, and vice versa. To reduce back-
mixing, the extractor length is divided into chambers by partition disks.
Because of its low dispersing effect, this equipment is particularly suitable
for systems which tend to emulsify. The throughput is low compared with
other extractors.
Mixer – Settlers. Mixer – settlers consist of single or several stages in
series (Fig. 22 A), each stage consisting of a stirred vessel and a settling vessel
(Fig. 22 B). Various arrangements are possible such as cocurrent, counter-
current, or crosswise flow of phases. The mixer – settler is the only extractor
802 LIQUID–LIQUID EXTRACTION Vol. 1
that reaches stage efficiencies of nearly 100 % and is almost insensitive for load
fluctuations.
This extractor has the advantage that backmixing can be completely pre-
cluded due to the complete phase separation in each stage. As a result, there
are no transfer problems in scale-up. The favorable efficiency is almost comple-
tely retained even for large throughputs. Any number of separation stages can be
built, and, moreover, throughput does not appear to be limited. These advan-
tages account for the widespread use of mixer – settlers in industrial processes
for many decades. However, early mixer – settler equipment was very complex.
Each separation stage consisted of a stirred vessel and a settling vessel. The two
phases had to be pumped between the individual stages. Each settler needed an
interface level control. The horizontal arrangement resulted in a large area
requirement.
In an effort to overcome these disadvantages of the mixer – settler batteries
without foregoing any of their advantages, a number of different designs were
developed. Mixer – settlers are used in two different areas. Industrial-scale
equipment is used to obtain extremely high throughputs when a small number
of stages is needed, eg, in the extraction of metal salts. Laboratory-scale devices
are employed in process development to check the number of calculated stages as
a function of the ratio solvent.
Large-scale mixer – settlers were originally developed to extract copper
from dilute aqueous solutions. The basic design is a box-type mixer – settler
according to Denver (Fig. 23). The mixing and separating zones are combined
in a single unit and are separated by an overflow weir for the light phase and
an underflow slit for the heavy phase. The individual stages are arranged so
that the outlet of the settler also acts as the inlet of the mixer. The agitator
Light Light
phase out phase in
Heavy Heavy
phase in phase out
Section A-B
Heavy Light
phase in phase in
A B
Fig. 23. Box-type mixer – settler (Denver). A) Side view; B) Ground plain.
804 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 24. Industrial mixer – settlers. A) Davy McKee mixer – settler; B) IMI turbine
pump-mix unit (cross section A – B – C – D shown in the lower part); C) Lurgi tower
extractor.
has two functions, it disperses and conveys the liquid phases. These functions
may conflict with one another. The rate of rotation necessary to overcome the
conveying height may be too high when unstable systems must be dispersed,
causing problems in phase separation, eg, a high degree of entrainment. Mixing
pumps with low shearing forces, and thus a low mixing intensity, have been
developed to avoid these difficulties. Examples are the Pumpmix impeller used
with the Davy McKee mixer – settler (Fig. 24 A) and the IMI turbine pump-mix
unit (Fig. 24 B).
A further development of the Davy McKee mixer – settler has been
described (43). Each stage consists solely of a vessel in which mixing and separa-
tion occur. The product inlet and the agitator are located in the middle and create
a dispersion zone. The light phase separates out at the top and the heavy phase
at the bottom.
In the Lurgi tower extractor, the battery of adjacent settlers is replaced by a
space-saving tower design in which the settlers are arranged on top of each
other. External centrifugal pumps are used for mixing (Fig. 24 C). Like all
mixer –settlers, this device is suitable for very large throughputs provided that
the liquid system can withstand the high energy input of the mixing pump. Back-
mixing to a certain degree is unavoidable since the stages have hydrodynamic
connections.
Multistage mixer – settlers are a valuable aid in process development on a
laboratory scale. They can be integrated in ‘‘mini plants’’; at this stage of devel-
opment no decision should be made which type of equipment is suitable for the
Vol. 1 LIQUID–LIQUID EXTRACTION 805
Fig. 25. Laboratory mixer – settlers. A) Normag mixer – settler; B) Schott/QVF mixer–
settler–>Schott/QVF mixer – settler.
necessary for the design of extraction apparatus. For this reason, a knowledge of
the operating characteristics of the various types of apparatus is of great interest
so that the number of experiments can be substantially reduced, and preselection
and preliminary sizing can be carried out at a relatively early stage of process
development.
The fluid-dynamic design of an apparatus is divided into the following
areas:
Fig. 27. Loading limit as a function of the pulsing frequency at different free area for
toluene – water (50). a ¼ 6 mm; vd/vc¼ 1; ~ j= 0.60; d ¼ 6.5 mm; &j = 0.40; d ¼ 4.0 mm;
*j ¼ 0.23; d ¼ 2.5 mm.
Fig. 29. Relative loading of a pulsed sieve plate column as a function of flow ratio (52)
* Toluene – water; & MIBK – water; ~ Butyl acetate – water; 3 n-Butanol – water .
810 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 30. Influence of mass transfer on loading limit and separation efficiency for the
toluene – acetone – water system (j = 0.60). ~ Mass transfer from the disperse to the con-
tinuous phase; & System without mass transfer; * Mass transfer from the continuous to
the disperse phase.
from the continuous phase to the dispersed phase (into the drop), smaller HETS
values result, ie, better separation efficiency is obtained (Fig. 30 B).
PILHOFER (56) has shown that the loading limit of extraction columns
depends only on the drop size. Accordingly, high loading limits can be obtained
only with relatively large drops.
Separation Efficiency. However to obtain a high separation efficiency a
large mass transfer area, eg, the greatest possible holdup of dispersed phase, is
required. The holdup rises with decreasing drop size. Any input of energy has the
purpose of producing the optimal drop dispersion. The interrelation of separation
efficiency and pulsing can be demonstrated by measurements on the toluene –
water – acetone system, which has a high interfacial tension (41). Figure 31
shows the measured height of a theoretical stage (HETS) for this system as a
function of the pulsation frequency, for the three different plate types (curves
a – c). In all cases, the HETS decreases as expected with increasing pulsing
intensity, eg, pulsing enhances the separation efficiency. The curves show an
Vol. 1 LIQUID–LIQUID EXTRACTION 811
important influence of the plate geometry: with increasing free area, the separa-
tion efficiency decreases sharply.
The butanol – succinic acid – water system shows very favorable HETS
values because of its extremely low interfacial tension (curve d). The effect of
free area is negligible. This means that, for systems of low interfacial tension,
the free area can be increased in favor of higher throughputs, without a signifi-
cant loss in separation efficiency.
A comparison of these two systems clearly shows the pronounced influence
of the interfacial tension on separation efficiency. This fluid-dynamic aspect
should also be taken into account when selecting solvents for extraction pro-
cesses. Hitherto, this selection has been made almost exclusively under thermo-
dynamic criteria.
The relationship between plate geometry, interfacial tension, and sepa-
ration efficiency, as described here for a pulsed sieve plate column, qualitatively
applies also to reciprocating-plate columns and pulsed packed columns.
Generally speaking, in systems with a high interfacial tension, particular atten-
tion must be paid to the choice of the internals, a higher energy input is also
necessary.
Fluid Dynamic Calculation Methods. The calculation methods for deter-
mining the loading limit can be classified in two categories. The first category
consists of empirical correlations which are based on physical and geometrical
data, and energy input. The second category is based on fluid-dynamic models
and presupposes knowledge of the drop size. Both method classes are applicable
only to two-component systems and thus neglect the effect of mass transfer, as
shown in Figure 30. The effects of wetting and of drop coalescence are likewise
disregarded. Moreover, the empirical correlations in particular are frequently
812 LIQUID–LIQUID EXTRACTION Vol. 1
Vc Vd
vc ¼ ; vd ¼ ð1Þ
S S
they have an effective velocity vc/(1 ed) or vd/ed. That yields the relative velocity
vr, also called slip velocity, between the two liquids
vd vc
vr ¼ þ ð2Þ
ed 1 ed
if
if
v p dp
Rep ¼ ð5Þ
nd
Archimedes number
dp 3 g %c %
Ar ¼ ð6Þ
Zc 2
%c 2 s3
KL ¼ ð7Þ
% gZc 4
density difference
% ¼ j%c %d j ð8Þ
Mathematical formulation implies the fact that an increase of the flow rate
at flooding cannot increase the drop holdup:
dvd
¼0 ð9aÞ
ded
when vc = constant,
dvc
¼0 ð9bÞ
ded
when vd = constant.
814 LIQUID–LIQUID EXTRACTION Vol. 1
With Equation (2) for the slip velocity vr the equations for calculating the
superficial velocities vc, vd at the flooding point are obtained in a dimensionless
form:
1=3
%c ed dRer
vc ¼ Ar1=3 ð1 ed Þ2 Rer þ ð10aÞ
% vc g ded
1=3
%c 1=3 2 dRer
vd ¼ Ar ed Rer ð1 ed Þ ð10bÞ
% vc g ded
Equations (10a) and (10b) together with Equation (3) for vr and Equation (4)
are the basis for flooding point diagrams according to Figure 32. Ordinate and
abscissa contain the superficial velocities vd and vc, ie, the phase flow rates, mod-
ified by physical data groups to a dimensionless form. The diagram shows the
calculated flooding curves of circulating drops as solid lines with the parameter
Ar. Dashed lines of constant phase ratio vd/vc and the total throughput as the
sum of vd and vc are also shown.
It can be seen from Equation (4b) that ReP is approximately proportional to
Ar0.3 and therefore to dP. That means that the drop velocity vP is nearly indepen-
dent of drop diameter dP in the region of oscillating drops. Therefore Figure 32
can be used also for oscillating drops if Ar is replaced by KL using Equation 4c.
Otherwise the diagram is valid for a certain KL value in the region of circulating
drops, in case of Figure 32, KL = 1011.
The dependence of drop diameter dP upon parameters of influence must be
known in order to use Figure 32 and to calculate flooding curves. Many scientific
investigations on drop disintegration in various extractors are found in the lite-
rature, but often without considering the influence of mass transfer and the
interaction of dispersing and coalescence. Furthermore, particularly for higher
Vol. 1 LIQUID–LIQUID EXTRACTION 815
where npi represent the numbers of drops of the class i of diameter dpi.
Finally, discussion will be restricted to empirical equations by KUMAR and
HARTLAND (6, Chap. 17) which give the best fit to experimental data and have
the broadest range of validity:
Pulsed sieve tray and reciprocating extractor
0:10
d32 0:74 h a f %1=4 a f %1=4
rffiffiffiffiffiffiffiffiffiffi ¼ C1 e exp 3:00 þ exp 28:65
s 0:05 m gs1=4 1=4 gs1=4 1=4
%g
ð12Þ
The best values of the parameter C1 are 1.51, 1.36, and 2.01 for no mass
transfer, c ! d, and d ! c direction of transfer, respectively. The product a f is
considered as the agitation variable in the equation, since the fit could not be
improved if a and f were treated separately. The average absolute value of
the relative deviation in the predicted values of d32 from the experimental points
is 16.3 %.
Rotating disk extractors
The optimized values of C1 are 0.63, 0.53, and 0.74 for no mass transfer,
c ! d and d ! c directions of transfer, respectively. The effect of holdup is ignored
due to lack of data. Equation (13) predicts the drop size with an average absolute
value of the relative deviation of 23 %.
Kühni columns
0:20 !0:24
d32 0:37 0:11 Zc D R 2 %c g
¼ C 1 e nS ½0:14 þ exp ð18:73FrR Þ
pffiffiffiffiffiffiffiffi
DR s%c DR s
ð14Þ
The values of the constant C1 are 9.81 102 for no mass transfer and c ! d
transfer and 0.31 for d ! c transfer. The stage number nS, which varies from 2 to
816 LIQUID–LIQUID EXTRACTION Vol. 1
17 in the present set of data, shows a rather weak effect on drop size. Equa-
tion (14) predicts the drop diameter with an average absolute value and relative
deviation of 17.6 %.
3.3. Apparatus Design. Extraction apparatus can be approximatively
designed purely by calculation using fluid-dynamic models (49,70). For reliable
design, however, pilot-scale tests are indispensable. If possible, the test should
be carried out with the original operating mixture in the type of apparatus
which is also envisaged for the large-scale unit. Desirably, the geometry of the
internals should be the same in the pilot-scale and industrial apparatus. How-
ever, this condition cannot always be satisfied. Generally, a column diameter
of at least 60 – 80 mm is necessary for design tests. Tests in smaller columns can-
not be used as the basis for equipment sizing (for criteria on equipment selection,
see Section 3.4). The number of tests can, however, be substantially reduced if
knowledge of the operating characteristics is utilized.
The physical data and the phase ratio are determined by the process used,
the following parameters are available for optimization: pulsing or rotation, geo-
metry of the internals, material of the internals, and phase selection, eg, the
direction of mass transfer. Energy input and the geometry of the internals are
selected on the basis of the physical data.
Internals and Operating Conditions. The following guidelines apply for
the selection of the free area in pulsed columns. An increase in the free area is
recommended in systems with a low interfacial tension where it leads to a great
reduction in the column volume. At moderate interfacial tension, the column dia-
meter can be reduced by a greater free area, but at the expense of the column
height. In the case of extremely high interfacial tension, there is no point in
increasing the free area.
The material of the internals should be selected so that the continuous
phase, but not the dispersed phase, wets the plates. As a rule of thumb, metal
plates can be used when the organic phase is dispersed; if the aqueous phase
is dispersed, the internals should be of plastic, both in pulsed columns and in
rotary-agitated columns.
In deciding which phase should be dispersed, the direction of mass transfer
must be taken into account. In accordance with Figure 30, mass transfer should
be from the continuous phase to the dispersed phase, if possible. In the final
phase selection, however, fluid-dynamic considerations frequently conflict with
operating and safety requirements. The position of the interface in the column
determines the phase which is dispersed. Figure 33 summarizes the possible
combinations.
The separation efficiency determines the column height and the load limit
determines the column diameter. For optimization, the two parameters are
related by the ratio y ¼ HETS=ðvc þ vdÞ which can be used to optimize the column
internals and energy input. A minimum value of this expression is equivalent to
a minimum column volume. Figure 34 shows that optimum pulsing is obtained
between the maximum of the flooding point curve and the minimum of the HETS
curve. An analogous procedure is used for rotary agitated columns. The design
point for industrial-scale columns is chosen to be 75 – 80 % of the load limit.
The operating point with respect to load and pulsing (rotation) is fixed by this
procedure.
Vol. 1 LIQUID–LIQUID EXTRACTION 817
ðvc þ vd Þ2 ¼ ðvc þ vd Þ1
with
ðvc þ vd Þ1 ¼ 0:75 ðvc þ vd Þ1 max
Index 1¼ pilot scale
Index 2¼ large scale
This gives the column diameter:
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u m _F m _S
u
u4 þ
t %F %S 2
D2 ¼
p ðvc þ vd Þ2
When the plate geometry and amplitude of the vibration are left unchanged
in the large-scale column, the frequency should be reduced, according to:
0:14
f2 D1
¼
f1 D2
The following applies to stirred columns (73):
0:33
HETS2 D2
¼
HETS1 D1
Robatel design provides more separation stages than, for example, the Podbiel-
niak extractor. Mixer – settlers are constructed in modules so that there is vir-
tually no limit to the number of separation stages.
A first estimate of the maximum possible throughputs in the individual
extractor types can be obtained by considering the largest diameters (Dmax)
which have hitherto been installed:
Dmax, m
Reciprocating-plate column 1.5
Pulsed column 2.7
Kühni extractor 2.5
QVF extractor 2.8
Rotating Disk Extractor 3.0
Mixer – settler tower 8.0
Fig. 36. Comparison of the number of theoretical stages NTS in dependence of total
load for Kühni, QVF, Karr and pulsed sieve plate extractors (PSE) (64) (system butyl
acetate – water – acetone, phase ratio vd/vc & 1.12, measurements of various authors cited
in (64)).
show higher NTS values than the pulsed columns but smaller ranges of load.
Generally speaking, all measures which improve the separation efficiency reduce
the load capacity.
A comparison between pulsed sieve trays and structured packings with the
test system toluene – water – acetone was published (74). Figure 37 shows the
diagram with the extractor efficiency in dependence on the load. It is obvious
that structured packings have the broadest range of load capacity and reach
the same or higher efficiencies as pulsed sieve trays. From the results for two
structured packings without pulsation the positive influence of pulsing can
clearly be seen: The pulsed unstructured packing, Pall rings 15 15 mm, has a
clearly smaller load range.
In Figure 38, the range of loads is compared for sieve trays and structured
packings with pulsation (75). According to that, packed columns have significant
advantages. As shown in (75), the investment costs of pulsed structured packed
columns are 30 % lower than the ones of pulsed sieve tray columns and 40 to 50 %
lower than pulsed unstructured packed columns.
In (76) a systematic study on extractor selection is given.
Adaptation of the Column Geometry to the Separation Problem. In
the early days of extraction, various types of apparatus were restricted to clearly
defined geometries and were thus sharply distinguished from one another.
Recent apparatus developments allow adaptation of the column geometry to phy-
sical data, both for pulsed columns and rotary-agitated columns. The boundaries
are thus no longer distinct and a general fluid-dynamic description of related
types of apparatus is feasible. The optimization of operating conditions and the
geometry of the internals is as equally important as the selection of a suitable
type of apparatus.
822 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 37. Comparison of efficiency versus load of packed columns with and without
pulsing and pulsed sieve tray extractors. a f ¼ pulsation intensity, Bf ¼ flooding load,
j ¼ relative free area.
Fig. 38. Range of load of pulsed structured packed columns in comparison to pulsed
sieve tray columns in dependence of the mass density difference for aqueous systems
and systems similar to water.
Vol. 1 LIQUID–LIQUID EXTRACTION 823
4. Phase-Separation Equipment
Fig. 39. Vertical (A) and horizontal (B) gravity settlers for drop separation from a
continuous liquid.
adjacent drops and at the principal interface. The coalesced drop phase leaves
the extractor at the top of the column.
In a horizontal settler, as shown in Figure 39 B, both phases flow in the
same direction. Drops of lighter dispersed phase rise to the interface where
they coalesce to form a homogeneous layer. In the lower part of the settler a
layer of the continuous phase remains, growing in flow direction due to decreas-
ing drop numbers. The wedge-shaped zone of dispersion should end before the
placement of phase outlets.
As the accumulation of impurities in technical liquid systems changes
important properties such as viscosity and surface tension, an exclusively theo-
retical modeling, as developed by many authors (see for instance (80)), is insuffi-
cient. However, the length of a horizontal settler can be estimated by a combined
procedure of calculations and simple and cheap batch settling experiments (79).
The calculation is based on a model of the physical processes in a settler, while
experiments supply one empirical parameter to include surface phenomena that
could not be modeled so far (see Figs. 40 and 41).
Fig. 42. Horizontal gravity settler with two inclined packages of plane, parallel plates.
The height of principal interface is best situated, when both phases have the same exit
velocity.
made during the 1970s and 1980s on the hydrodynamics of inclined films and the
coalescence of drops at the film surface, these are summarized in (83–86). The
main results are:
1) Thin laminar films should flow on the plates, because the coalescence
of drops is promoted by the shear stress of film flow near the plate;
the larger the volume flow of cocurrent continuous phase, the smaller
the film thickness and the better the coalescence.
2) The plate should be wettable by the drop phase (advancing contact angle
< 708, receding contact angle > 158).
3) The drop swarm has to be distributed equally over the width of plates.
4) Constructions that can thin the trickling film increase the settler
performance.
5) Short lengths of plates of about 400 mm are sufficient for coalescence of
drops riding on the film.
6) The width bp of a plate can be calculated by the empirical equation of the
Reynolds number of trickling film
0:19
Ref 80 KL;d
Vd % g sin a
bp ¼ 0:03
ðRef k vd Þ3=2 Zd
s3 %d %d
KL;d ¼
g Z4d %
where a is the plate inclination, Vd,1 = Vd/n the volume flow of dispersed phase on
a plate, and n the number of plates.
7) The optimum of plate inclination is within the range 108 < a < 158.
828 LIQUID–LIQUID EXTRACTION Vol. 1
All attempts to calculate the bed performance in advance via filter coeffi-
cients were fruitless (87–90). MAGIERA developed a mathematical model of physi-
cal mechanisms inside a fiber bed and was able to simulate the bed behavior
and drop enlargement by introduction of two experimental fitting parameters.
5.1. General. The quality of the raffinate in liquid – liquid extraction can
be controlled easily by changing the ratio of solvent to feed and the extraction
temperature. The quality of the extract is governed only by the conditions of
feed location and no direct control is possible.
This is especially true because most extractions are carried out in a closed
system to minimize solvent requirements. In such a system, a maximum extract
purity is obtainable and some components that should end up in the raffinate
leave with the extract.
Vol. 1 LIQUID–LIQUID EXTRACTION 831
Fig. 45. Use of cooling stage A) Process flow scheme: a) Extractor; b) Settler B)
Triangular diagram: a) Binodal curve at higher temperature; b) Binodal curve at lower
temperature.
Additional steps are, therefore, required to enhance extract purity and thus
raffinate yield. If the extract is the product desired, improving extract purity is
the most important — in fact, the only — objective. Different solutions to this pro-
blem are applied in commercial operation.
Change in Temperature. Extract purity can be increased by cooling the
extract from the main system and adding one more stage (see Fig. 45 A). In the
triangular diagram, this means that the binodal curve is changed from a to b
(Fig. 45 B).
Because the composition of the first extract is now inside binodal curve b, it
separates in a settler b in a purer final extract phase and a small amount of inter-
mediate raffinate phase. The latter is returned to the main extraction system.
The selectivity of separation in the additional stage is not very great
because the loading of the first extract phase is rather high and separation
takes place near the plait point. This system can be used when the main system
is operated at high temperature for other reasons or when an increase in extract
purity and the resultant increase in raffinate yield are desirable.
Addition of Water. The addition of water at or below the feed point gen-
erally changes the system from a closed to an open one (see Section 2.1). In this
way, extract purity can be increased to the required level. Water is soluble, at
least to some extent, in most polar solvents and therefore leaves the system
with the final extract phase. In this way, no unwanted influence is exerted on
the main extraction system.
However, because water must be separated from the solvent by distillation
before its reintroduction in the main extractor, the heat requirements are high.
Addition of water is, therefore, not regarded as a generally attractive solution.
Also, water can only be added when its boiling point is much lower than those
of the solvent and the feedstock.
832 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 47. Use of countersolvent with intermediate recovery. a) First extractor; b) Distilla-
tion column; c) Second extractor.
A disadvantage of this system is that the countersolvent enters the top sec-
tion. Consequently, the phase ratio of B-rich to A-rich phase is reduced, resulting
in a less pure raffinate (this also means a lower yield of extract). To restore raf-
finate purity, more B-rich phase must be applied. Such a system is thus costly
and sensitive to changes in feed composition, solvent-to-feed ratio, and ratio of
countersolvent to feed.
Some improvement can be obtained by carrying out the extraction in two
parts (Fig. 47).
In the first extractor (a), the feed is extracted with the primary solvent. The
intermediate extract phase is fed to a second extractor (c) and brought in contact
with the countersolvent. The mixture of countersolvent and intermediate raffi-
nate is fed to a column (b) in which the intermediate raffinate and the counter-
solvent are separated by distillation. The pure intermediate raffinate is then fed
to the first extractor, while the countersolvent is charged to the second extractor.
In this way, some of the disadvantages can be overcome.
5.2. Combined Processes of Extraction and Distillation. Azeotropic
Distillation. Usually it is necessary to combine extraction and distillation pro-
cesses for recovering the solvent from the extract phase and also from the raffinate
phase, if the solvent is soluble in the raffinate.
Figure 48 shows as an example the recovery of acetic acid from an aqueous
solution with methyl tert-butyl ether (MTBE) as solvent. The system of water
and acetic acid is a highly nonideal azeotropic mixture that could be separated
by simple distillation. However, for dilute feed compositions solvent extraction
834 LIQUID–LIQUID EXTRACTION Vol. 1
Fig. 48. Triangular diagram (A) and flow sheet (B) of a hybrid extraction – distillation
process for separation of water and acetic acid with MTBE as solvent (93) E = extraction;
C-1, C-2 = distillation; S = settling; H = condenser.
free raffinate R is fed into column C-2 where it is separated into the bottoms,
pure water B2 and water – MTBE azeotrope D2. Both distillates D1 and D2 lie
in the immiscible region between MTBE and water. This makes simple decanta-
tion (S) of the condensed phases possible. The solvent-rich fraction L1 is not
further purified since MTBE as solvent needs not to be pure. So L1 can be
recycled directly to the extractor E, a part of it is used as reflux for C-1. In column
C-2 a mixture of water and MTBE has been separated into pure water and
MTBE – water azeotrope. Therefore, L2, the water-rich phase from the settler,
can also be fed into column C-2.
The economy of the hybrid process depends on a careful optimization of
operational and constructional parameters, besides the careful selection of the
solvent. Recently, mixed integer nonlinear programming (MINLP) has been
successfully introduced in optimization of such processes (93,94). MINLP can
be used as a process synthesis tool. It is necessary to generate a so-called
superstructure of the hybrid process which contains a variety of alternative
flow sheets. Then MINLP is able to choose the best solution for given para-
meters and also from competitive unit operations like extractive distillation,
processes with more than one solvent or processes with reextraction for solvent
recovery.
Extractive Distillation. Another possibility to separate azeotropic
mixtures is extractive distillation. A solvent with a much higher boiling point
than the feed is fed at the top of a distillation column to entrain one component
of the distillation feed that enters the column at the bottom (see Distillation).
The combined system of liquid – liquid extraction and extractive distillation is
applied in nearly all processes used to manufacture very pure light aromatics.
N-Formylmorpholine is a suitable solvent to separate the light aromatics
benzene and toluene used as feedstock for the petrochemical industry
ions are replaced by the active solvent. The new complex will be insoluble in
water but soluble in the solvent.
Generally, the formation of a complex is a slow reaction. Therefore, in many
cases, it is not possible to use columns for this type of extraction. In many appli-
cations mixer – settler are used to make the extraction possible.
A reactive extraction will need a re-extraction to isolate the extracted sub-
stance from the solvent. The active solvents used in reactive extractions are
water insoluble compounds. They are used undiluted or in solution in an inert
solvent.
Extraction Mechanism of Different Types of Solvent. Classical
Complexing and/or Chelating Agents. The active solvents are usually
large, polyfunctional molecules. Most of them are designed to extract specific
cations. Dioximes, hydroxyoximes, quinolines, b-diketones, etc., are rather
expensive but widely used solvents to extract metal ions such as copper and
nickel from the aqueous phase.
Metal complexes are formed in the water phase in a substance-specific pH
range and extracted into an inert solvent. After extraction the metal ion is liber-
ated from the complex by changing the pH, and re-extracted into the water
phase. The complexing/chelating agents are designed such that they are chemi-
cally stable, nontoxic, and insoluble in water to minimize losses. The liquid ion
exchange (LIX) (5) agents used in the production of copper are typical for this
category.
Henkel, Hoechst, Shell, Sherex, and Acorga are the largest producers of
complexing/chelating extractants (see Table 1).
838 LIQUID–LIQUID EXTRACTION Vol. 1
The metal salts of these acids are soluble in most nonpolar solvents and,
depending on the pH of the extraction and the pKa of the acid, the cation will
migrate to the organic phase or stay in the water phase. The dependence of
the distributions coefficients on pH is shown in Figure 50.
The different curves give the variation of the distribution coefficient D in
dependence of pH for different cations. These solvents are not specific for certain
cations, but are very suitable to concentrate and recuperate trace elements from
effluent streams (109,110).
The major producers of this type of compounds are Bayer, Hoechst, Daiha-
chi, Shell, and Albright & Wilson.
Ethers and Thioethers
R2 O þ HX Ð R2 OHþ þ X
R2 S þ HX Ð R2 SHþ þ X
where HX stands for a strong acid, eg, H2SO4 or HCl, and X is the anionic
rest of the acid. Most metal ions that are capable of forming anionic complexes
are extractable with ethers and thioethers. Fe, Au, In, Ga, etc. can be extracted
with these solvents from an aqueous HCl solution (HCl concentration > 6 M)
and re-extracted from the organic phase with water. Most of these anionic com-
plexes form strong acids which protonate with the ether group to form oxonium
(thionium) salts. These salts, however, are soluble in ether or thioether. It is also
possible to extract Cl from water with (R2O)2 H2SO4, the sulfate ion being
replaced by the chloride. This type of solvent is normally used in undiluted
form (106).
Ketones. extract the same anionic complexes as ethers, but in most cases
at lower acid concentration. Complexes with weaker acids can also be extracted.
Methyl isobutyl ketone and cyclohexanone are industrially used in the separa-
tion of Nb – Ta as F complexes (112) (! Niobium and Niobium Compounds,
Chap. 4.3.1) and in the separation of Zr – Hf as CNS complexes (! Zirconium
and Zirconium Compounds, Chap. 1.4.2) (116). The solvents are used in undi-
luted form. A great disadvantage of this and of the former type of solvents is
that they are not completely insoluble in water. A distillation is needed to remove
the dissolved solvent from the water phase.
Trialkylphosphine oxides, Trialkylphosphinates, -phosphonates,
and -phosphates
R3 PO þ HX Ð R3 POHþ þ X
ROR2 PO þ HX Ð ROR2 POHþ þ X
ðROÞ2 RPO þ HX Ð ðROÞ2 RPOHþ þ X
ðROÞ3 PO þ HX Ð ðROÞ3 POHþ þ X
With this type of extractants most metals from group 3 to group 8 can be
extracted, even if only weak complexes are formed with the solvent. It is for
example possible to extract TiOSO4 (114) from a sulfuric acid solution with
dibutyl butylphosphonate (DBBP) as the solvated complex 2 DBBP TiOSO4
H2SO4 5 H2O. These organophosphorus solvents are, due to their outstanding
chemical stability and good physical properties, most widely used in reactive
extraction (107,108).
Free acids can also be extracted, the basicity of the solvents decreases from
phosphine oxide to trialkylphosphate. The distribution coefficients are rather
small so that large volumes of solvent are needed in an extraction process
840 LIQUID–LIQUID EXTRACTION Vol. 1
These onium compounds form ions both in acidic and basic medium. The
mechanism of extraction is practically always anion exchange (106). The
onium compounds are regenerated in the same way as anion exchanger resins
by re-extraction with a base. These solvents are used in the production of Cr,
V, Re, Tc, Mo, and W as a solution in a petroleum fraction.
Producers are Henkel, Sherex, Hoechst, and BASF.
Amine and Phosphine Salts
R3 N þ HX ! R3 NHþ þ X
R3 P þ HX ! R3 PHþ þ X
½R3 NHþ A
Solvent þ ½B
Raffinate Ð ½R3 NHþ B
Solvent þ ½A
Raffinate
This series is dependent on the type of amine, the diluent, and the tempera-
ture and must be determined for any given system (5). This series can be
enlarged for organic compounds, capable of forming ion pairs with the amine.
Phenols and substituted phenols can be extracted with better distribution coeffi-
cients than chlorides. In this series thiophenols and some mercaptans are situ-
ated between bromide and nitrate and all sulfonic acids are better extracted than
Vol. 1 LIQUID–LIQUID EXTRACTION 841
phosphate. This makes the sulfate salts of amines a very good choice for new
extraction systems.
(R3N)2 H2SO4 (R = octyl) is capable of extracting Cl and/or NO3 (113) from
a mixture of both anions with SO2 4 . The value of the distribution coefficient of
the anions is surprisingly high.
A disadvantage of these solvents is the fact that per mole amine one mole of
salt is formed after regeneration. However, most of the time, the anions to be
removed are present in the solution only as trace impurities, so that this draw-
back is not significant. Another disadvantage of amines is the poor solubility of
ammonium salts in aliphatic hydrocarbons. To avoid third phase formation, the
use of aromatic solvents (smell) or aliphatics with modifiers is necessary. Amines
can also act as solvating agents in an acid medium. Iron, zinc, and uranium salts
can be extracted in this way (106).
Uses
1) A bleed from a (NH4)2SO4 crystallization unit in a caprolactam plant con-
tains 40 % (NH4)2SO4, 10 % NO
3 , 50 ppm Cl and 3 – 4 % organic sulfonic
acids. The solution is extracted by trioctylammonium sulfate, the added
amount is stoichiometric to [NO
3 ] + [Cl ] + [sulfonic acid]. In three stages
> 99 % of the nitrate, 90 % of the chloride, and 85 % of the sulfonic acids
are extracted by the trioctylamine in solution (40 %) in Shellsol AB (aro-
matic hydrocarbon). The raffinate is returned to the crystallization plant
for (NH4)2SO4 recovery. The re-extract contains 35 % of nitrate, 0.016 %
of chloride, the sulfonic acids, and only 0.1 % of sulfate and is treated
separately.
Compared with a conventional stripping, the use of chemicals in this
system is strongly reduced. 80 % less NaOH is consumed because the
base is only used to liberate the amine and not to neutralize the NH3
from (NH4)2SO4 (113).
2) Traces of phenol (20 – 200 ppm) and bisphenol (20 – 100 ppm) can be ex-
tracted from the wastewater of a polycarbonate plant to values < 1 ppm,
thus avoiding biological treatment of the water. The re-extract containing
sodium phenolate and bisphenolate (300 – 500 times more concentrated)
can be returned to the polycarbonate plant. The solvent used is 5 % triocty-
lamine in Shellsol AB (117).
3) From a plant for the production of vulcanization accelerators 2-mercapto-
benzothiazol (100 – 300 ppm) and its oxidation products are completely ex-
tracted in two stages and re-extracted as a 20 – 30 % solution. This
concentrated waste can be easily removed. The solvent is 20 % trioctyla-
mine in Shellsol AB (113).
4) Bromide ions can be extracted selectively from sea water (concentration
in the Dead Sea & 4 %) with amine hydrochlorides and concentrated in a
re-extraction step, thus avoiding the energetically unfavorable stripping
842 LIQUID–LIQUID EXTRACTION Vol. 1
of sea water. The extraction can be carried out with various types of amines
as HCl or H2SO4 salt (106).
5) Mg2+ can be removed from a saturated solution of (NH4)2SO4 by extraction
with the ammonium salt of diethylhexyl phosphonic acid (DEHPA) at pH
> 5. It can be recovered by lowering the pH to < 3 (115).
6) Traces of uranium are extracted from phosphate ores. The uranium content
in these ores is only 30 – 300 ppm. The reserves of those ores are so huge
that the quantity of uranium from phosphates many times exceeds that
of uranium found in other uranium ores.
7) Lanthanides and actinides have only recently become available in higher
quantity and high purity. This was enabled by the development of the ex-
traction technology.
8) Worldwide about 40 % of the copper is produced by reactive extraction. The
plants are an order of magnitude larger then any other type of solvent
extraction presently in operation. With an aqueous feed of 200 – 3000 m3/h
a whole new technology had to be developed. Copper can be extracted after
acid or ammoniacal leaching. The aqueous stripping phase is used directly
in the electrowinning of copper. To avoid the build-up of impurities in the
electrolysis of copper, Fe, Zn, and other metals are removed from this phase
by extraction with other solvents (5,102,111).
9) More than 90 % of the world production of cobalt is performed by reactive
extraction. Separation of Co and Ni can be achieved from alkaline, neutral,
or from acidic media. Most of the impurities (Sn, Bi, Fe, Pb, Cu, Cd, and Zn)
have to be removed from the solution by precipitation or extraction. From a
medium concentrated aqueous chloride solution (100 – 200 g/L), Co can be
extracted by amines together with only traces of Ni. After scrubbing with
4 – 5 M HCl, most of the Ni is removed from the solvent. The cobalt is
stripped by water.
From neutral or weakly acidic media (pH 4.8 – 5.5) Co can be extracted
by the Naþ or NHþ 4 salt of di-2-ethylhexylphosphoric acid (D2EHPA)
(to avoid acidification of the aqueous solution). Some Ni is co-extracted
and is removed by scrubbing with a Co solution thus replacing any ex-
tracted Ni by Co. Co is finally stripped by HCl, H2SO4, or HNO3 depending
on the use.
The use of monooctyl octylphosphonic acid (MOOP) to replace
D2EHPA (101) permits the same separation, but at lower pH and with en-
hanced distribution constants for Co.
In alkaline solution Co and Ni can be extracted by LIX agents from
ammoniacal solution. Prior to the extraction, cobalt is oxidized to the Co3þ
state. Nickel is stripped first with dilute acid, Co remaining in the solvent.
Cobalt is then reduced to the Co2þ state and stripped with diluted acid
(5,102,104,105).
All these examples are used in industry or in pilot plants. Elaborate pro-
cess and flow control are essential to use this technology in a economical and
meaningful way.
Vol. 1 LIQUID–LIQUID EXTRACTION 843
E ¼ D S=R:
Z ¼ ðEnþ1 EÞ=ðEnþ1 1Þ
over n stages. For this equation the number of stages can easily be calculated
To remove metal B from the solvent the solvent has to be washed with
water or acid in such a way that most of metal A stays in the solvent, while
metal B is washed out as completely as possible. This is done in zone 3 of Fig-
ure 51, the scrubbing zone. The amount of water to be used is given again by
844 LIQUID–LIQUID EXTRACTION Vol. 1
the distribution coefficients of both acid and metal B. If too much acid is removed,
metal A will also be washed out. Therefore it is more adequate to use an
acid solution. The concentration of the dilute acid should be such that the acid
solution can only take up the amount of acid to be removed from the solvent.
By scrubbing with pure water only a limited amount of scrubbing fluid can be
used because of the low distribution coefficient of the acid. By adding a calculated
amount of acid the amount of scrubbing fluid can be increased and more compo-
nent B can be removed without removing component A. The raffinate, leaving
the scrubbing zone contains too much of metal A to be discarded and is sent
back to one of the steps in the extraction. The number of steps can again be cal-
culated with the Kremser formula.
In zone 2 of Figure 51, the stripping zone, metal A is washed out of the sol-
vent with water or a dilute acid solution. The amount of water is chosen such as
to obtain the desired concentration of metal A in the re-extract. With the given
distribution coefficient, it is not possible to remove metal A completely form the
solvent. At any acid concentration, traces of metal A will remain in the solvent.
In zone 1 of Figure 51, the washing zone, the acid and the rest of metal A are
washed out of the solvent. After this step, the solvent is completely regenerated
and can be reused.
General Flow Sheet. Figure 52 shows a general flow sheet of a reactive
extraction system.
Solvent Make-Up. Due to the poor distribution coefficient of the acid seve-
ral extraction steps are needed to load the solvent with sufficient acid to make
extraction of the metals possible. In the given example the acid comes from
the initial solution. To avoid losses of metal A, the solvent must be completely
free of A. At the low acid concentration in the first steps of the solvent make-
up any metal A present in the solvent will be re-extracted in the raffinate.
Reactive Extraction. If sufficient acid is present in the solvent, both metals
will be extracted from the feed solution. To limit the number of steps in the
extraction, concentrated acid is added. The high concentration is necessary for
the metal extraction and in the solvent make-up.
Vol. 1 LIQUID–LIQUID EXTRACTION 845
Solvent Scrubbing. To make sure that most of metal B is washed out from
the solvent, the acid solution containing metal A and acid from the solvent wash-
ing step is used, so that we stay in the extraction zone for metal A. The second
component, with a lower distribution coefficient, is already in its stripping zone
and will be washed out of the solvent. To avoid losses, this stream is sent back to
one of the steps in the extraction.
Solvent Stripping. After scrubbing, the solvent contains only metal A
and acid. Using water or a dilute acid solution, the purified metal A can now
be re-extracted into a water phase.
Solvent Washing. The solvent must be completely regenerated before it
can be recycled to the solvent make-up. This is done by adding some supplemen-
tary washing steps to the system. The amount of water is chosen according to the
requirements in the scrubber and the number of steps is calculated as a function
of that amount.
Diluents and Modifiers. Diluents are used to reduce the solvent viscosity,
improve phase separation, and most of all, when properly chosen, to reduce the
loss of solvent. The solubility of tri-n-butyl phosphate (TBP, one of the most fre-
quently used extractants) in pure water is 420 mg/L at room temperature; a 50 %
dilution of TBP in n-hexane lowers this solubility to 175 mg/L. The choice of
diluent is very important. Higher amines are completely soluble in aliphatic dilu-
ents. Their salts, however, are only slightly soluble in the same solvents. There-
fore it is essential to test the solutions in all steps of the process. If third-phase
formation occurs in one of the steps, the diluent must be changed or a modifier
is to be added. These modifiers are normally long chain alcohols, monoesters of
polyhydric alcohols, or phosphate esters. These modifiers may considerably influ-
ence the kinetics of the extraction. Especially in the extraction of copper the
choice of modifier is very important. The different types of solvent are sold
mixed with a suitable modifier which greatly enhances the extraction perfor-
mance of the solvent.
6. Symbols
2a Pulsing amplitude, m
A key component of the phase from which the solute is ex-
tracted
B key component of the phase into which the solute is ex-
tracted
ci constants
C transferred component
di hole or packing diameter, m
dp particle diameter, m
d32 Sauter mean drop diameter, m
D column diameter, m
DR rotor diameter, m
D distribution coefficient
E = K ṁs/ṁF extraction factor
f pulsing frequency, s1
846 LIQUID–LIQUID EXTRACTION Vol. 1
Greek symbles
Dimensionless Groups
Subscripts
REFERENCES
92. Abwassertechnischer Verein e.V. (ed.): Industrieabwässer, Verlag Ernst & Sohn,
Berlin 1999.
93. S. Glanz, Ph. D. Thesis, Technische Universität München, 1998.
94. S. Glanz, J. Stichlmair in R. Darton (ed.): ‘‘Distillation and Absorption ’97,’’ Ichemt,
Symp. Ser. No. 142, 1997.
95. F. Trefny, Erdöl Kohle Erdgas Petrochem. 23 (1970) 337 – 340.
96. E. Cinelli et al., Hydrocarbon Process. 51 (1972) no. 4, 337 – 340.
97. Hydrocarbon Process. 51 (1972) no. 4, 141 – 144.
98. G. Precesser et al., Oil Gas J. July 16 (1973) 114 – 118.
99. Krupp Koppers, Hydrocarbon Process. 52 (1973) no. 4, 139 – 141.
100. Hydrocarbon Process. (1982) 182.
101. G. M. Ritcey, A. W. Ashbrook: Solvent Extraction, Part 1, Elsevier, Amsterdam 1984.
102. G. M. Ritcey, A. W. Ashbrook: Solvent Extraction, Part 2, Elsevier, Amsterdam 1979.
103. Solvent Extraction in Metallurgical Processes, Internat. Symp. Antwerpen 1972.
104. W. W. Schulz, J. D. Navratil: Science and Technology of Tributyl Phosphate, vol. I,
CRC Press, Boca Raton, FL 1984.
105. W. W. Schulz, J. D. Navratil, A. S. Kertes: Science and Technology of Tributyl Phos-
phate, vol. VI, CRC Press, Boca Raton, FL 1991.
106. A. K. De, S. M. Khopar, R. A. Chalmers: Solvent Extraction of Metals, Van Nostrand
Reinhold, London 1970.
107. Bayer AG: Geschäftsbereich Anorganische Chemikalien, Baysolvex – TBP für die
Flüssig-Flüssig Extraktion 1993.
108. Bayer AG: Geschäftsbereich Anorganische Chemikalien, Baysolvex – Dialkan-phos-
phate für die Flüssig-Flüssig Extraktion 1993.
109. Bayer AG: Geschäftsbereich Anorganische Chemikalien, Baysolvex – D2EHPA für
die Flüssig-Flüssig Extraktion 1993.
110. Bayer AG: Geschäftsbereich Anorganische Chemikallien, Baysolvex – D2EHTPA für
die Flüssig-Flüssig Extraktion 1993
111. Bayer AG: Geschäftsbereich Anorganische Chemikalien, Baysolvent for your
Solvent-extraction 1993.
112. F. Fairbrother: The Chemistry of Niobium and Tantalum, Elsevier Amstrad, London –
New York 1967, pp. 8 – 10.
113. Bayer Antwerpen N.V., EP 0 396 790 A1, 1989.
114. Bayer Antwerpen N.V., DE 3 643 711 A1, 1986.
115. Bayer Antwerpen N.V., EP 638 515-A2, 1993.
116. W. D. Jamrack: Rare Metal Extraction by Chemical Engineering Technology, Perga-
mon Press, Oxford 1963.
117. M. L. Wang, B. L. Liu: ‘‘Extraction Equilibrium of Phenol by Sulphuric Acid Salts of
Triisooctylamine,’’ Chem. Eng. Comm. 156 (1996) 131 – 146.
118. Bayer Antwerpen N.V., DE 3 524 053 A1, 1985.
ECKART MÜLLER
Frankfurt a. Main Germany
ROSEMARIE BERGER
Ludwigshafen Germany
ECKHART BLASS
Technische Universität München
München Germany
DOMIEN SLUYTS
Bayer AG Antwerpen Belgium
Vol. 1 PIPELINES 851
PIPELINES
Pipelines or pipe lines are continuous large-diameter piping systems, usually
buried underground where feasible, through which gases, liquids, or solids sus-
pended in fluids are transported over considerable distances. They are used to
move water, wastes, minerals, chemicals, and industrial gases, but primarily
crude oil, petroleum products, and natural gas. In the oil and gas business, a
pipeline system consists of a trunkline, ie, the large-diameter, high pressure,
long-distance portion of the piping system through which crude oil is shipped
to refineries, or natural gas and oil products, respectively, are transported to dis-
tribution points, and smaller low pressure gathering lines that transport oil or
gas from wells to the trunkline. Smaller lines used by natural gas distributors
are not considered part of a gas pipeline system (see NATURAL GAS).
Pipeline transport involves the application of force to the material being
moved, either through the use of pumps to transport liquids, compressors to
move gases, or flowing water to move solids. In some applications, vacuum
may create the pressure differential.
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1609160523080112.a01
852 PIPELINES Vol. 1
prices during the late 1980s and early 1990s and resulted in the formation of a
spot market for natural gas, a gas futures market, and a need for more sophisti-
cated gas technology. The common carrier status of natural gas pipelines allows
many small producers to enter the pipelines, if pipeline capacity is available, and
make deliveries on a short-term (spot) basis. Flow computers at entry and exit
points along the pipeline provide an instantaneous and accurate account of the
gas that enters or leaves the pipeline. The flow computers tie into a large com-
puter at pipeline headquarters to record an accurate account of gas from produ-
cers entering the pipeline as well as gas exiting to a gas user or distributor. In
addition, FERC Order 636 mandated the establishment of an electronic bulletin
board (EBB) for the gas industry, by which anyone with a personal computer can
obtain the latest information needed to purchase gas and transport it by pipeline,
such as suppliers’ prices, available pipeline capacity, etc (1). The EBB will also
carry information about capacity release in pipeline operations and permit
trading of capacity release options so that elements of pipeline capacity will
become a commodity, just as the natural gas itself. The objective is to optimize
the efficiency of the national gas pipeline grid (2).
1.2. Methane (Natural Gas). Although the first natural gas pipeline
was constructed in 1870, most of the gas consumed at that time was manufac-
tured from coal (qv) and used locally rather than transported by long-distance
pipeline. In the 1990s, natural gas is conveyed in strong, thin-walled, long-dis-
tance pipelines in virtually all principal countries of the world. In the United
States, the total length of gas pipelines is ca 5:82 105 km, which includes
ca 4:24 105 km of transmission pipelines, ca 1:49 105 km of field (gathering)
lines, and ca 0:09 105 km of lines in storage areas (3,4). This represents ca 73%
transmission pipelines, 25.5% gathering lines, and ca 1.5% storage lines. Collec-
tively, long-distance gas pipelines account for only about one-third the natural
gas transport system in the United States; the other two-thirds are gas distribu-
tion mains to consumers.
Polyethylene (PE), poly(vinyl chloride) (PVC), or polypropylene plastic pipe
is being used in increasingly greater amounts throughout the world in both gas
gathering systems and gas distribution systems. U.S. gas distribution companies
have installed 644,000 km of polyethylene piping since its introduction in 1968
and are adding over 40,000 km each year (5). Plastic pipe is used in sizes greater
than 400-mm dia for certain nonpressurized nongas applications, but is not used
in sizes of greater than 400-mm dia for natural gas and not at all for truly long-
distance natural gas pipelines. When plastic pipe is installed under roads, cas-
ings are often required. Railroads do not allow polyethylene pipe to be used at
crossings; the additional installation costs for crossings can reduce the economic
advantages of using plastic pipe for long pipelines (6).
Transport of natural gas starts with small-diameter gathering lines that
convey the raw gas from individual wells to a collection point and from there
to a gas treating plant, where heavier hydrocarbons (qv), particulates, and
water are separated from the gas. The hydrocarbons known as natural gas
liquids (NGL) are transported as raw mix in liquid pipelines to chemical plants,
where they are fractionated into lighter and heavier fractions that are trans-
ported to market areas in products pipelines. The methane-rich gas, which
usually contains some ethane, is processed to remove particulates, water, and
Vol. 1 PIPELINES 853
gaseous impurities such as hydrogen sulfide, carbon dioxide, etc, and then com-
pressed to the appropriate pressure for transmission by large-diameter pipeline.
Pressure is maintained by large compressors at compressor stations along the
length of the pipeline, usually spaced 80–160 km apart (7). Maximum operating
pressures are 3.45 MPa (500 psi) for older lines and 9.93 MPa (1440 psi) for
newer lines.
Of the 125 U.S. gas pipelines listed in 1991, 47 were rated as principal
pipelines, ie, 1:42 109 m3 of gas transported or stored for three consecutive
years (4). The longest gas pipelines originate near the largest U.S. gas producing
areas, primarily the Gulf of Mexico and the mid-continent area of Oklahoma,
Texas, Kansas, New Mexico, etc, and move gas to the more populated areas of
the country, ie, Los Angeles, Chicago, New York, or New Jersey. Maps showing
the locations of natural gas pipelines in the United States (8) and Europe (9) are
available.
Some U.S. natural gas pipeline companies are subsidiaries of gas holding
companies. The largest U.S. natural gas pipeline companies, in terms of overall
length of transmission systems are Northern Natural Gas Company, 26,539 km;
Tennessee Gas Pipeline Corporation, 23,567 km; Columbia Gas Transmission
Company, 18,481 km; Natural Gas Pipeline Company of America, 17,200 km;
and Transcontinental Gas Pipe Line Corporation, 17,071 km. For gas moved in
1994, the four largest pipelines were ANR Pipeline Company, 95; 278 106 m3
(3,363,275 MMcf), of which 40.8% was gas moved for others; Transcontinental
Gas Pipe Line Corporation, 87; 050 106 m3 (3,073,801 MMcf), of which 99.7%
was moved for others; Natural Gas Pipeline Company of America,
83; 089 106 m3 (2,933,940 MMcf), of which 87.1% was moved for others; and
Northern Natural Gas Company, 56; 523 106 m3 (1,995,861 MMcf), with
100% moved for others.
In Europe, Russia’s huge natural gas resources are being moved to Western
Europe by pipelines from Siberia to the consuming countries. Gas from Algeria’s
large resources is moved by 2500-km, 1200-mm dia pipeline from the Hassi R’Mel
gas field across the Mediterranean Sea and the Strait of Messina to Italy in water
depths reaching to 610 m. A 1448-km pipeline will carry Algerian gas to Spain
and Portugal via Morocco. Norway’s Statoil will build a 1095-km pipeline system
(Zeepipe) from the North Sea to Europe. In South America, a 2252-km gas pipe-
line will transport natural gas from producing fields in Bolivia to Sao Paulo, Bra-
zil, and hook up with a line between Sao Paulo and Rio de Janiero. A gas pipeline
that would cover 27,000 km has been proposed by the National Pipeline Research
Society of Japan to transport gas from Siberia and Sakhalin Island to Australia
by way of Japan, China, Thailand, Malaysia, and Indonesia (9).
Ammonia Pipelines. Ammonia [7664-41-7] is a commodity produced from
natural gas and may be produced in plants located near its market areas or
plants may be located near gas-producing regions and the ammonia transported
to the market areas by tank car, barge, or pipeline. The primary ammonia pipe-
lines in the United States are the Mid-America Pipeline Company (MAPCO) and
Gulf Central, both connecting ammonia-producing areas in Texas, Oklahoma,
and/or the Gulf Coast with farming districts throughout the Midwest. The pipe-
lines are made of high strength pipeline steel, 200–250-mm dia, and operate
with an ammonia pressure of 2.07–10 MPa (300–1450 psi) to maintain the
854 PIPELINES Vol. 1
ammonia in the liquid state. A limit of 0.2% H2O and 0% CO2 in the ammonia is
specified to minimize stress–corrosion problems. Stress–corrosion cracking of
the high strength pipeline steel can occur if the ammonia is contaminated by
atmospheric carbon dioxide. Copper-base alloys are corroded by ammonia in
the presence of air and water, and equipment containing these metals can be cor-
roded by even small ammonia leaks. The MAPCO ammonia pipeline has resolved
these problems by using a special steel containing 0.25% C, 1.0–1.7% Mn, and
0.15% Cu; eliminating copper alloys from valves, fittings, and pumps; stress-
relieving fabricated valves or using cast valves and stress-relieving pipe fabrica-
tions containing branch connections; and using neoprene for valve trim (10).
A Russian ammonia pipeline of nearly 2400 km extends from Togliatti
on the Volga River to the Port of Odessa on the Black Sea, and a 2200-km,
250-mm dia branch line extends from Gorlovka in the Ukraine to Panioutino.
The pipeline is constructed of electric-resistance welded steel pipe with 7.9-mm
thick walls but uses seamless pipe with 12.7-mm thick walls for river crossings.
The pipeline is primed and taped with two layers of polyethylene tape and
supplied with a cathodic protection system for the entire pipeline. Mainline
operating pressure is 8.15 MPa (1182 psi) and branch-line operating pressure
is 9.7 MPa (1406 psi) (11).
Hydrogen Pipelines. The manufactured gas distributed by early gas dis-
tribution systems contained up to 50% hydrogen, and at least two large-scale
operations in the 1990s have evaluated 10–20% hydrogen mixtures with natural
gas, but actual experience with long-distance hydrogen pipelines is rather lim-
ited. The oldest hydrogen pipeline (started in 1938) is the Chemische Werke
Hüls AG 220-km, 150–300-mm dia system in the German Ruhr Valley that
transports 100 106 m3 of hydrogen annually to multiple users at a nominal
pressure of 1.55 MPa (225 psi). Following its expansion after 1954, some fires
have occurred but no hydrogen embrittlement or explosions (12,13). Other
shorter H2 pipelines include a 340-km network in France and Belgium, an
80-km pipeline in South Africa, and two short pipelines in Texas that supply
hydrogen to industrial users (14). NASA has piped H2 through short pipelines
at their space centers for several years.
Natural gas is roughly eight times more dense than hydrogen (0.72 g/L vs
0.09 g/L). Because the pipeline capacity of a pipeline depends on the square root
of the gas density, the pipeline capacity of hydrogen is nearly three times greater
than for natural gas. The heating value of hydrogen, however, is only about one-
third of the 37.2 MJ/m3 heating value for an equivalent volume of natural gas.
Thus, the energy carrying capacity of a given size pipeline is approximately the
same when it is carrying either hydrogen or natural gas, provided that it is oper-
ating in turbulent flow with the same pressure drop along its length and at the
same operating pressure. At higher pressures, the ratio of heating values of nat-
ural gas and hydrogen increases and at 5.17 MPa (750 psi); the ratio is 3.8:1,
compared with about 3:1 at atmospheric pressure. This also means that hydro-
gen compressors must handle 3.8 times more gas than natural gas compressors
for the same energy throughput, thus indicating higher pipeline transmission
costs for hydrogen than for natural gas. One study suggests that regional trans-
port costs for hydrogen may be as much as five times greater than for natural gas
at the higher pressures but that long-distance pipeline transport should not show
Vol. 1 PIPELINES 855
such a large difference (15). Some components currently used for natural gas
may be adequate for hydrogen service but not compressors and meters. In
view of the different compressor requirements for hydrogen and natural gas,
the design of rotary compressors used for natural gas is probably not satisfactory
for hydrogen use (12,16).
The question of whether hydrogen embrittlement of pipeline steel is a pro-
blem has not been completely resolved. No hydrogen embrittlement problems
were reported for hydrogen pipeline transmission at 5.17 MPa (750 psi), and
an American Petroleum Institute (API) study of steels operating in hydrogen
service indicates no problems at temperatures below 2048C and pressures
of 69 MPa (10,000 psi). Other tests of pipeline steels at hydrogen pressures of
3.5–6.9 MPa (500–1000 psi) have shown a loss of tensile ductility and acceler-
ated fatigue-crack growth. Embrittlement is a potential problem in high pressure
pipeline transmission of hydrogen (13).
1.3. Industrial Gases. Industrial gas (oxygen, nitrogen, etc) pipelines
are short compared with long-distance pipelines that transport crude oil, natural
gas, or petroleum products; however, more than 80% of the oxygen and more
than 60% of the nitrogen produced in the United States by air separation is
transported by pipeline. Air-separation plants that are built adjacent to the oxy-
gen or nitrogen users’ facilities are usually owned and operated by the user or a
second party. However, plants located at more distant locations are usually
owned and operated by the oxygen producer and the gases transported by pipe-
line to multiple users in the area, eg, the 150-km, 406-mm dia Houston
Ship Channel pipeline system that supplies oxygen and nitrogen from air-
separation plants to chemical plants and oil refineries from Houston to Texas
City, Texas (17). Carbon dioxide and hydrogen are also moved by pipeline in
the Houston area. Other U.S. industrial gas pipeline systems are located near
Gary, Indiana, and along the Mississippi River near the Gulf Coast (18).
Multiple users of oxygen, nitrogen, hydrogen, and carbon monoxide near
Rotterdam, the Netherlands, are supplied by pipeline from a nearby industrial
gas complex. Oxygen has been transported for many years by a French pipeline
connecting Metz with Nancy and extending to Luxembourg and Saarbrucken in
Germany.
The application that has led to increased interest in carbon dioxide pipeline
transport is enhanced oil recovery (see PETROLEUM). Carbon dioxide flooding is
used to liberate oil remaining in nearly depleted petroleum formations and trans-
fer it to the gathering system. An early carbon dioxide pipeline carried by-pro-
duct CO2 96 km from a chemical plant in Louisiana to a field in Arkansas, and
two other pipelines have shipped CO2 from Colorado to western Texas since the
1980s. Feasibility depends on crude oil prices.
Helium is extracted from natural gas in the southwestern United States
and moved by a 685-km, 50-mm dia pipeline to storage in a partially depleted
gas field near Amarillo, Texas, as part of the U.S. government’s helium conserva-
tion program.
1.4. Cryogenic Gases. Some of the most sophisticated pipeline technol-
ogy deals with the transport of liquefied gases with very low boiling points,
ie, cryogenic gases such as oxygen (182:96 C), argon (185:7 C), nitrogen
(195:8 C), hydrogen (252:8 C), and helium (268:9 C). These gases are
856 PIPELINES Vol. 1
Shield
spacers-location
staggered from
liquid helium
line spacers
Aluminum–copper
liquid N2 / gaseous He Stainless steel
transfer line jacket pipe
Inner
line
spacers
Aluminum
shielding
extrusion
Liquid helium
transfer line SST
systems based on similar technology are also designed for transfer of liquid
oxygen, liquid nitrogen, and liquid hydrogen; liquefied natural gas has also
been transferred in vacuum-jacketed piping. Thus far, transfer lines are short
by comparison with the longer-range pipelines for other applications, but many
thousands of vacuum-jacketed piping installations in the 30–150-m range have
been installed, with one system over 8 km in length.
The main technical difference between liquid and gas pipeline transport is the
compressibility of the fluid being moved and the use of pumps, rather than com-
pressors, to supply the pressure needed for transport. The primary use for liquids
pipelines is the transport of crude oil and petroleum products.
2.1. Crude Oil and Products Pipelines. The true pioneering in long-
range modern oil pipelines was the construction of two pipelines to move crude
oil from Texas to New Jersey and Pennsylvania during World War II. These were
named the Big Inch (2156-km, 600-mm dia) and the Little Big Inch (2373-km,
500-mm dia) pipelines.
The crude oil delivery system starts with relatively small-diameter
gathering lines from individual producing wells to a main-line pump station,
from where it is pumped through a larger transmission trunkline to a refinery
or other destination. At the refinery, the crude oil is separated into gasoline,
jet fuel, kerosene, distillate fuel oil, etc, and the refined products are
transported by products pipelines to markets, storage, shipping terminals,
etc. In modern lines, all inputs and outputs are metered, monitored, and remo-
tely controlled by supervisory control and data acquisition (SCADA) computer
systems.
Several different refined products are shipped in the same pipeline
(batching) by using control methods to minimize intermingling at different inter-
faces. This is achieved by maintaining high turbulent flow in the pipe. Synthetic
rubber spheres have been used as separators at interfaces but are not used often
in the 1990s. Intermingling occurs at the interfaces, but it can be estimated and
minimized. At the terminal, the comingled fraction is separated from other pro-
ducts and either blended or returned to the refinery. Interfaces between products
are delineated with densitometers. This technique is used to ship different types
of crude oil in the same pipeline.
The liquids pipeline subsidiaries of principal oil companies with both
production and refining facilities generally operate all three types of liquids pipe-
lines: gathering, transmission, and products lines. However, the independent
9000-km, 900-mm dia Colonial Pipeline transports only products. It carries
about 105 106 m3 of products annually. Of the nearly 276,000 km of liquids
pipelines reported in 1991 by FERC, 47% were products lines, 34% were trans-
mission lines, and only 19% gathering lines. This compares with 35, 34, and 31%,
respectively, in the 1970s. The total length of the U.S. operating interstate pipe-
line system has remained relatively constant since 1982 at about 720,000 km,
with roughly 32% represented by liquids pipelines and 62% by natural gas
pipelines (4).
858 PIPELINES Vol. 1
At the beginning of 1992, the largest liquids pipelines in the United States,
based on pipeline length, were Amoco Pipeline Co., 19,096 km; Mobil Pipe
Line Co., 15,026 km; Exxon Pipeline Co., 14,983 km; and Conoco Pipe Line
Co., 12,980 km. Distances do not include 1316 km of the Trans-Alaska Pipeline
with multiple ownership. In both 1991 and 1992, the product pipeline company
with the most product deliveries was Colonial Pipeline with 104,990,000 m3,
more than double the amount delivered by Santa Fe Pacific Pipelines, Inc.
The top pipeline in terms of crude oil deliveries was the Alyeska Pipeline Service
Co., operator of the Trans-Alaska Pipeline System, with movement of
105,735,000 m3 (3).
2.2. Sulfur and Chlorine Pipelines. Underground sulfur is melted by
superheated water and then piped as liquid to the surface with compressed air.
At the surface, molten sulfur is transported by heated pipeline to a storage or
shipping terminal. One such pipeline, located under 15 m of water in the Gulf
of Mexico, is insulated and surrounded by steel casing to which are strapped
two 130-mm dia pipelines that carry return water from the deposit. The super-
heated water is carried from shore to the deposit in a 63.5-mm dia pipe inside the
pipeline that carries the molten sulfur (21).
Chlorine is shipped by pipeline in either gaseous or liquid form; however,
great care must be taken to ensure that liquid pipelines are operated only in
the liquid phase and gaseous pipelines operated only in the gaseous phase.
Liquid chlorine has a high coefficient of thermal expansion and high pressure
develops with increasing temperature; liquid chlorine trapped between two
valves can lead to hydrostatic rupture of the pipeline. Because chlorine is corro-
sive and has a high vapor pressure, it and the pipeline must be dry before the
chlorine enters the pipeline. Operating pressure should not exceed 217 MPa
(31,465 psi) and temperature should be no higher than 1218C (22).
2.3. Trans-Alaska Pipeline. The design and construction of the Trans-
Alaska Pipeline was one of the most difficult and ambitious pipeline projects ever
attempted. Its 1316-km length extends from the producing fields at Prudhoe Bay
across three mountain ranges, where up to 12 m of winter snow and wind chills
of 73 C can be expected, to the Valdez Marine Terminal on the southern coast
of Alaska. Thirteen bridges were built to carry the pipeline across rivers and dif-
ficult terrain. To avoid localized melting of the permafrost by the 32–498C crude
oil and possible instability of the soil, a significant portion of the pipeline was
installed above ground, supported by 78,000 vertical support members embedded
in permafrost to depths ranging from 4.5 to 18 m. The pipeline is insulated and
jacketed to reduce heat losses, and passive refrigeration is used to keep the soil
frozen around the vertical support members which can move up or down, thus
allowing a 0.6-m vertical movement of the pipeline. The pipeline is laid out in
a zigzag fashion to allow a 3.6-m horizontal movement for pipe expansion and
contraction and possible seismic disturbances. The buried portion of the pipeline
is 611-km long, with 6.4 km of it refrigerated. The aboveground portion of the
pipe is at least 1.5 m above the surface to permit caribou migration underneath
and at least 3 m above the surface at the 554 special animal crossings. The pipe-
line is constructed of 1219-mm dia high tensile strength carbon steel, with wall
thickness of either 11.7 or 14.3 mm. Although the Prudhoe Bay oil discovery was
announced in 1968, the initial pipe of the pipeline was not laid until early 1975
Vol. 1 PIPELINES 859
because of many delays. Final weld was made May 31, 1977 (23, 24). Final cost,
including construction interest, was roughly $11 billion.
Pipelines to transport solids are called freight pipelines, of which three different
types exist: pneumatic pipelines, the use of which is known as pneumotransport
or pneumatic conveying; slurry pipelines, which may also be called hydrotran-
sport or hydraulic conveying; and capsule pipelines. When air or inert gas is
used to move the solids in the pipeline, the system is called a pneumatic pipeline
and often involves a wheeled vehicle inside the pipeline, propelled by air moving
through the pipe (25). Slurry pipelines involve the transport of solid particles
suspended in water or another inert liquid. Hydraulic capsule pipelines trans-
port solid material within cylindrical containers, using water flow through the
pipeline for propulsion.
3.1. Pneumatic Pipelines. Pneumatic pipe systems are used to move
blood samples, medicine, and supplies between buildings in hospital complexes;
cash and receipts in drive-up banks; parts and materials in factories; refuse from
apartment complexes; and grain, cement, and many other materials. Most of
these are small diameter and usually short; however, a 17-km, 1220-mm dia
pneumatic pipeline has been used to transport rock in the former Soviet Union
since 1981, and a 3.2-km, 1000-mm dia line has moved limestone from the mine
to a cement plant in Japan since 1983 (22).
3.2. Slurry Pipelines. Finely divided solids can be transported in pipe-
lines as slurries, using water or another stable liquid as the suspending medium.
Flow characteristics of slurries in pipelines depend on the state of subdivision of
the solids and their distribution within the fluid system. Although slurry flow
and conventional liquid flow are both divided into laminar, transitional, and
turbulent flow, the contribution of the solids-to-slurry flow results in a further
characterization as either homogeneous or heterogeneous flow. Homogeneous
flow of slurries occurs when very finely divided solids are distributed uniformly
throughout the suspending medium, eg, bentonite slurries as drilling muds.
These slurries are usually quite concentrated and often have viscosities much
higher than those of the suspending medium itself. Heterogeneous flow occurs
when larger, irregular solid particles are slurried and distribution is not uniform
throughout the suspension; these slurries are usually of lower solids concentra-
tion and viscosities are comparable to that of the suspending liquid.
The difference between homogeneous and heterogeneous flow is due to the
deposition velocity, below which solid particles start to separate from the slurry
and build up in the pipeline, and is related to the degree of turbulence and the
particle fall rate. Homogeneous slurries behave more like single-fluid systems;
deposition velocity is primarily a property of heterogeneous slurries. Deposition
velocity increases with increasing particle density, solids concentration, particle
size, and pipeline diameter. In the transport of heterogeneous slurries, turbulent
flow is an important requirement to prevent solids from building up in the pipe-
line. Slurries may be classified according to particle size, eg, colloidal (<1 mm),
structured (1–50 mm), and finely dispersed (50–150 mm, mostly produced by
860 PIPELINES Vol. 1
The coal log pipeline involves forming logs of coal by compressing and
extruding finely divided coal into cylindrical shapes having a diameter 90–
95% the diameter of the pipeline. The coal logs are then injected continuously
into a water-filled pipeline for transport by the flowing water. The log length
is between 1.5 and 3 times the log diameter and buoyancy is achieved at a
water flow rate of 2.4–3.0 m/s. This reduces the energy required for pumping,
as well as the friction between coal log and pipeline walls. The coal log pipeline
is said to use only one-third as much water for coal transport as coal slurry pipe-
lines, in addition to eliminating the dewatering step. No full-scale coal log pipe-
line has been built, but short test lines and research activities continue, and a
Coal Log Pipeline Consortium has been formed, with both private companies
and governmental agencies participating (34).
4. Pipeline Technology
P ¼ ð2St=DÞ F E T
where P ¼ design pressure in kPa (psig), and may be limited per federal stan-
dard 192-105 to 75% of the pressure if steel pipe has been subjected to cold
expansion to meet the specified minimum yield strength (SMYS) and then
heated to 4828C or held above 3168C for more than 1 h, other than by welding;
S ¼ yield strength in kPa (psig); t ¼ nominal wall thickness in mm (in.);
Vol. 1 PIPELINES 863
F ¼ design factor, which is 0.72 in rural areas and 0.40 in urban areas, with
intermediate values of 0.60 and 0.50 for gas lines and 0.72 for liquid lines;
E ¼ longitudinal joint factor, which depends on the type of pipe weld seam and
is 1 if seam strength equals pipe strength; T ¼ temperature derating factor,
which is 1 for operating temperatures of 1218C or less but <1 for higher tempera-
tures, such as at compressor discharge; and D ¼ nominal outside diameter in
mm (in.).
This formula determines only the primary pipe (hoop) stress caused by the
internal fluid pressure perpendicular to the pipe wall; however, it also recognizes
secondary hoop stress through the design factor, F, which represents a safety
factor. The maximum design factor of 0.72 recognizes the secondary stresses
caused by the load of the overburden against an imperfect trench bottom,
those produced by thermal expansion or contraction against earth constraint
on buried pipe, and others. The design factor is lowered in populous areas to
provide additional safety and to recognize additional stresses in those areas
compared with rural areas; it is also lowered for installations at river crossings,
highways, etc. The maximum allowable operating pressure (MAOP) of the
pipeline can be affected by the pipe mill test pressure level and/or the hydro-
static strength test after construction is complete; either could underrate the
pipeline. Longitudinal primary tensile stresses are half of the primary hoop
stress. Specifications for valves, fittings, etc, must be written as part of the
design phase. Shut-off valves are required at specific intervals to isolate any
pipeline damage; powered valve-closing devices are commonly installed to
shut off valves automatically in the event of a line break and isolate the
damaged portion of the line. Valves on cross-country pipelines should be of a
full round-opening design to permit passage of devices, called pigs, for line
cleaning, separating water from gas during hydrostatic testing, corrosion
detection, etc. Specifications are made for exterior coatings and cathodic pro-
tection to minimize external corrosion, and for interior coatings or additives
to minimize interior corrosion, as well as decrease flow resistance and improve
transmission of odorants. Interior pipe coatings are usually epoxy paints
applied before pipeline assembly.
Spacing between compressor/pumping stations is usually 80–160 km along
the pipeline to boost pressure lost to internal friction by the gas or liquid being
transported. The compressors/pumps are driven by either reciprocating internal
combustion gas engines of up to 4.5 MW (6000 hp), centrifugal gas turbines
of up to 11.2 MW (15,000 hp), or electric motors of up to 7.5 MW (10,000 hp).
A compressor/pumping station usually consists of several compressors/pumps
operating in parallel.
4.2. Construction. Pipeline construction is a continuous activity
required by supply changes due to decreasing production from older fields, open-
ing up new fields, or bringing in new supplies from Canada; obsolete pipelines
are decommissioned, additions are made to existing pipelines, or new pipelines
are designed and laid to take advantage of new supply sources. Construction
of a cross-country pipeline may be divided into segments (spreads) and proceeds
by the following steps: right-of-way clearing and grading, trenching, pipe string-
ing along the right-of-way, pipe bending to fit the trench bottom, welding, clean-
ing and priming the exterior surface for coating, coating and wrapping for
864 PIPELINES Vol. 1
cathodic protection unless it is mill-coated and only the weld areas need coating,
lowering pipeline into trench, and backfill of earth over the pipeline.
Alternative construction methods are used for crossing obstacles. For river
crossings, floats may be used to carry the pipe across the river with concrete coat-
ings or weights, flooding the floats to sink the pipe to the river bottom where the
weights hold it in place. Land under highways is bored with an earth auger fol-
lowed closely by the pipe or casing, if required, advanced by the same machine.
With advances in horizontal directional drilling, bored river or wetlands cross-
ings of up to 762 m of bore have been made.
When construction is complete, the pipeline must be tested for leaks and
strength before being put into service; industry code specifies the test proce-
dures. Water is the test fluid of choice for natural gas pipelines, and hydrostatic
testing is often carried out beyond the yield strength in order to relieve secondary
stresses added during construction or to ensure that all defects are found. Indus-
try code limits on the hoop stress control the test pressures, which are also lim-
ited by location classification based on population. Hoop stress is calculated from
the formula, S ¼ PD=2t, where S is the hoop stress in kPa (psig); P is the internal
pressure in kPa (psig), and D and T are the outside pipe diameter and nominal
wall thickness, respectively, in mm (in.).
Tests of cross-country pipelines are recommended at 90% of pipe yield
strength in order to avoid having insufficient field-test pressure, which can
limit the MAOP of the pipeline to less than that calculated by the design formula.
According to industry code, this can be achieved only by hydrostatic testing with
nonflammable liquid, such as water. Filtered water is used to displace air left in
the line during construction; a pipeline pig separates the two phases and elimi-
nates air pockets. The water is pressurized to produce the hoop stress desired
and is recommended to be held for a minimum of 8 h to test pipe strength.
When the test is finished, the water is displaced by the pressurized fluid for
which the line was built, ie, natural gas, crude oil, etc. A pig is used to separate
the two phases and remove accumulated water at low points; for natural gas
pipelines, multiple pigs also remove water retained on pipe walls that could
cause operational problems. The pipeline is tied in to distribution and supply
systems, if it passes the tests, and put into service. The final step is a cleanup
and posting of required location markers.
Construction of underwater (submarine) pipelines does not take place
under water. Pipelines are welded onshore and dragged into position by powerful
winches on ships floating on the water surface (for short lines), welded on a
specially constructed lay barge, and lowered to the ocean floor by a stinger
from one end of the barge or welded onshore, floated on pontoons, and towed
to the offshore area where they are lowered into position. For smaller size
pipelines, the lines can be welded onshore and spooled onto large reels, placed
on special ships, and spooled into the offshore trench. In shallow water, or
where endangered by anchors or wave action, submarine pipelines are laid in
trenches in the sea bottom (35). Underwater pipelines are concrete-coated or
weighted to overcome the buoyancy effect; concrete coatings applied over the pri-
mary coating provide additional protection against damage during laying and
against corrosion. Submarine pipelines are being used regularly to transport
oil, natural gas, and other commodities to shore from offshore locations, such
Vol. 1 PIPELINES 865
as the Gulf of Mexico, the North Sea, and the Arabian Gulf. One of the longest
and technically challenging submarine pipeline systems is a 2599-km, 1200-mm
dia pipeline for transporting natural gas across the Mediterranean Sea and the
Strait of Messina from gas fields in Algeria to Italy; pipes are laid in water depths
to 610 m. Offshore and onshore pipelines require different design factors (36).
Trenchless Construction. Increased emphasis on environmental and
ecological factors has presented considerable difficulties for normal pipeline
trenching, even in unpopulated areas, if considered harmful to wildlife in their
native habitats. In densely populated areas, trenching is costly and complex
because of airports, commercial buildings, factories, etc. To circumvent them,
trenchless procedures are being developed for installing fuel lines, water lines,
sewer pipe, etc, or in some cases for steel casings that can serve as conduits
for several smaller diameter pipe systems. Horizontal directional drilling is a
trenchless construction technique that drills under an obstacle, such as a street,
airport runway, or river, and the welded continuous steel pipeline segment is
pulled through the opening after horizontal drilling is complete. Another proce-
dure that was used in Berlin to put a large-diameter casing under a railway
embankment between a train station and the River Spree (37) involves the use
of pneumatic pipe ramming to ram the steel pipe through the soil. Trenchless
construction is used commonly for small gas distribution lines.
Microtunneling involves sinking of two vertical shafts, large enough to
lower a horizontal drilling machine to the level from which to dig a tunnel
between the two vertical shafts to accommodate the pipeline. The technique
has proved useful for installing or rehabilitating sewer pipe systems and often
includes the sinking of additional shafts to provide access to the moling (horizon-
tal drilling) machines. Microtunneling was originally and principally developed
in Europe, but variations are being developed in both the United States and
Japan, and costs are expected to drop rapidly with expanded usage (38).Houston
and Dallas are among the U.S. cities that have applied this technique to revita-
lize underground utilities infrastructures (39–41).
4.3. Maintenance. Pipeline Pigging. After construction, a pipeline
must be tested, inspected, cleaned, mapped for bends, dents, or ovalties, main-
tained during operation, and monitored for leaks or corrosion to ensure safety
of operation. These operations involve pigging, moving devices called pigs
through the pipeline that can carry out these missions. These may be sophisti-
cated electronic devices or something as simple as a rubber ball pumped through
the pipeline to displace fluids (42).
Modern pipeline pigging applications begin with the commissioning of a
pipeline when pigs are first used to remove air prior to hydrostatic testing and
then used to remove any water left after the testing. A gauging (caliper) pig is
used to locate any dents or buckles resulting from the laying or backfilling opera-
tions. Special cleaning pigs are used to remove scale, dirt, etc, from interior
walls. If the pipeline transports multiple products, pigs can be used to separate
them at the liquid interfaces. To prevent loss of pipeline efficiency or increased
corrosion due to deposits of wax, scale, or bacteria, the deposits are removed with
brush pigs or even more sophisticated pin-wheel pigs that remove deposits to a
precise depth (43,44). Smart or intelligent pigs involving electronic sensor
interfaces with computers and miniature video cameras (45), ultrasonics,
866 PIPELINES Vol. 1
magnetic-flux leakage, etc, have been developed to inspect pipeline interiors for
integrity, metal loss, corrosion, and other defects; however, some pipelines have
physical limitations that do not accommodate smart pigs. The potential of smart
pigs to improve pipeline safety is considered to be so great that the Office of
Pipelines Safety published a ruling that would require modification of pipelines
to permit their use where feasible or practical; however, many petitions for
reconsideration were filed and a limited suspension of compliance has been
granted while the situation is under study (46). Several pigging operations can
be put into a single train (47).
Corrosion. Anticorrosion measures have become standard in pipeline
design, construction, and maintenance in the oil and gas industries; the prin-
cipal measures are application of corrosion-preventive coatings and cathodic
protection for exterior protection and chemical additives for interior protection.
Pipe for pipelines may be bought with a variety of coatings, such as tar, fiber
glass, felt and heavy paper, epoxy, polyethylene, etc, either pre-applied or
coated and wrapped on the job with special machines as the pipe is lowered
into the trench. An electric detector is used to determine if a coating gap
(holiday) exists; bare spots are coated before the pipe is laid (see CORROSION
AND CORROSION CONTROL).
Cathodic protection is provided by inducing an electric current on the pipe-
line to ensure that the electric potential of the metallic pipe is less than the earth
surrounding the pipeline after it is laid, by using a sacrificial anode bed with a
more electropositive metal than iron, or by thermoelectric protection, using heat
to generate a direct current to lower the potential of the pipe. Inducing an
electric current involves the use of a rectified alternating current to lower the
potential of the pipe relative to the ground, which also provides a way to monitor
the system by an aerial pipeline survey. Pulsating the direct current produces a
magnetic field with a strength proportional to the current flowing in the pipe;
using field coils to measure the magnetic field strength provides a method of
monitoring the current, even for underground pipe. If the survey indicates
that the magnetic strength is reduced over any segment of the pipeline, it
would indicate change of potential and current loss, such as would occur with
a hole in the pipeline coating, an electric casing short, a separated coupling, or
other event that would permit corrosive action (48).
U.S. Department of Transportation (DOT) statistics on liquids pipelines
operated under the Code of Federal Regulations (49) indicate that corrosion
was the second largest contributor to accidents and failures for the period from
1982 to 1991. These statistics covered an average of 344,575 km of liquids pipe-
lines and were derived from required reports to DOT on all pipeline accidents
involving loss of at least 7.95 m3 of liquid, death or bodily harm to any person,
fire or explosion, loss of at least 0.8 m3 of highly volatile liquid, or property
damage of $5000 or more (50). Similar results were also reported for 1991 in
the 1992 DOT/OPS report on both oil and gas pipeline incidents; 62 out of
210 oil pipeline incidents were due to corrosion, of which 74% were due to
external corrosion (43). For gas pipelines, 16 of all 71 reported incidents were
due to corrosion, of which 63% were reported as due to internal corrosion;
however, internal corrosion of gas pipelines is likely only if CO2 and H2O
and/or H2S are present, as with unprocessed gas in gathering lines.
Vol. 1 PIPELINES 867
5. Economic Aspects
Pipelines usually represent the least costly way to move fluid products, including
solid slurries, wherever they compete with truck or rail transportation. Compar-
ing the transport costs for removing municipal sludge in the former Soviet Union
shows a surprising similarity to a similar comparison of transport costs for mov-
ing U.S. crude oil and petroleum (qv) products. The Russian study (26) showed
that pipeline transport of sludge is equal to 25–38% the cost of barge transporta-
tion, 6–13% the cost of rail transport, and 4–5% the cost of truck hauling. By
comparison, pipeline cost/km of moving crude oil and petroleum products in
the United States has been only about 17% of rail costs and 3.5% of truck
transport costs (26,51).
The cost of building a pipeline is usually divided into four categories: mate-
rials (line pipe, fittings, coatings, cathodic protection, etc), right-of-way and
damages, labor, and miscellaneous (surveying, engineering, supervision, admin-
istration and overhead, interest, contingencies, afudc, and FERC filing fees). The
relative contribution of each category depends on the size of pipe, the difficulty of
terrain, and delays and added costs due to environmental factors, but generally
the costs for materials and labor account for about 75% of the entire cost of the
pipeline, with each accounting for about half this amount. The remainder is
engineering, overhead, fees, interest, etc. The relative contribution of each
category can vary widely in difficult terrain or climate, however, as illustrated
868 PIPELINES Vol. 1
1860 3000
1802 2900
682 1100
620 1000
Costs in $1000 / km
Costs in $1000 / mi
558 900
496 800
434 700
372 600
310 500
248 400
186 300
124 200
62 100
102
152
203
254
305
356
401
457
508
559
610
660
711
762
813
864
914
965
1016
1067
Nominal pipe diameter, mm
by the high labor costs of the Trans-Alaska Pipeline, which were 69.3% of the
total, and ca $1:3 109 interest charges during construction or roughly 15% of
the total pipeline cost (52).
The variation of average pipeline construction costs with increasing size of
line pipe is shown in Figure 2, based on data taken from FERC construction
permit applications from July 1991 to July 1992. The cost of a common carrier
pipeline project must be reported to the FERC no later than six months after
successful hydrostatic testing.
The 10-yr construction costs for land-based gas pipelines show no obvious
trend either in the four individual categories or the high–low range, on a $/km
basis; however, pipeline land acquisition costs rose steadily, with average $/km
cost increasing from $370,881 to $528,355 for proposed U.S. gas pipeline projects
over the 1990–1992 period. The $/km costs for 1991–1992 were material,
$190,379; labor, $200,127; right-of-way (ROW) and damages, $21,938; and mis-
cellaneous, $115,930 (4). The cost of compressor stations for the same period
averaged $1894/kW.
When a natural gas pipeline has been commissioned and is operating,
its transmission costs are the operating expenses plus maintenance expenses.
In 1989, these two cost elements for gas pipelines were $13,573/106 m3 and
$2872/106 m3, respectively; for 1990, the corresponding values were $17,669/
106m3 and $3851/106m3, respectively. The primary operating expense for
natural gas pipelines is the cost of compression and gas transmission by
others, which accounted for ca 40% of the 1989–1992 expenses. The next largest
Vol. 1 PIPELINES 869
operating expenses were compressor station labor and expenses, and gas for com-
pressor fuel, accounting for 30% of the total (4). Pigging and testing costs depend
on pig type, pipeline cleanliness and diameter, extent of corrosion, etc. The
Government Accounting Office survey shows a range of $404–$1492/km for
the cost of smart pig inspections; a similar survey of three foreign smart pig
manufacturers indicated a range of $746–2486/km for a 1609-km, 610-mm
dia pipeline. Hydrostatic test costs are even higher; estimates by the U.S.
DOT are $2890/km for a 508-mm dia hazardous liquids pipeline; this cost
rises to $5523–8794/km if costs of transporting, testing, and disposing of the
test water and the revenue loss due to the pipeline being out of service are
added. The cost of pressure testing and transportation or disposal of test
water is about $6563/km of pipeline, depending on the pipeline diameter, the
terrain crossed, and water scarcity in the area (46).
Data compiled by the U.S. DOT indicate that pipeline transport is the safest
materials transport mode, particularly over long distances. In 1990, fatalities
attributed to oil and gas pipelines were significantly lower (8 out of 4679) for
all materials transporters in the United States, compared with 599 for rail trans-
port and 3281 for motor transport (trucks) (53).
To ensure the safety of gas pipelines, the Natural Gas Safety Act of 1968
was created to mandate federal regulation of gas storage facilities and pipeline
transport of natural gas, with the U.S. Department of Transportation given
exclusive authority to regulate safe operation of natural gas pipeline systems
that fall within the jurisdiction of the Federal Power Commission under the Nat-
ural Gas Act. The Hazardous Liquid Pipeline Safety Act of 1979 and the Hazar-
dous Materials Transportation Act of 1979, and their respective amendments,
gave authority for the U.S. DOT to regulate safety for all pipelines operating
under the Code of Federal Regulations (49). The U.S. DOT continues to regulate
safety on pipelines and has established legal standards for all pipeline phases,
from design to construction, operation, and maintenance. DOT also issues statis-
tics on pipeline accidents from required reports on accidents involving release of
gas or LNG and death or hospitalization or losses in excess of $5000.
Between 1985 and 1991, 1726 natural gas pipeline ruptures and leakages
were reported in the United States. These incidents resulted in 634 injuries
and 131 fatalities. Third-party damage was the most common cause of these inci-
dents, followed by corrosion. The GAO believes that the corrosion-related inci-
dents can be reduced with the use of smart pigs (46). U.S. DOT 1992 accident
statistics showed that 52.5% of U.S. oil spills involving loss of at least 1590 m3
came from pipeline accidents, comparable to the worldwide statistic of 51.5%.
The U.S. DOT regulated 344,575 km of liquids pipelines during the 10-yr
study period and received reports on 1901 accidents during that time; thus the
number of failures per year per 1000 miles was 0.888, of which 27% was due to
corrosion and 31% to outside forces (48).
Most ecological issues involve the pipeline’s construction phase and the
effect that it will have on vegetation, wildlife, topography, population density,
870 PIPELINES Vol. 1
and land use. All pipeline design and construction programs must include an
environmental impact statement which must be submitted and evaluated before
permission to proceed with construction is granted. Construction through ecolo-
gically sensitive obstacles, such as rivers and wildlife areas, is often suspended
during certain times of the year, such as fish spawning and wildlife migration
seasons, unless acceptable alternatives can be found. Archaeological concerns
must also be addressed.
Greater natural gas use for power generation and vehicular fuel use, to help
meet more stringent environmental standards, is expected to expand natural gas
pipeline usage. High technology will influence the role of post-Order 636 gas
pipelines (54). Pipeline transport of coal could become a reality in the United
States, as well as other coal-rich countries, to reduce dependence on imported
oil. Greater use of freight pipelines to reduce pollution from trucks and other
transportation modes may arise (25). Pipeline research activities are expected
to expand through organizations such as Japan’s National Pipeline Research
Society and the Capsule Pipeline Research Center and Gas Research Institute
in the United States.
The concept of transporting heat over great distances with chemical energy
pipelines has been considered (55).
BIBLIOGRAPHY
‘‘Pipelines’’ inECT 3rd ed., Vol. 17, pp. 906–929, by T. P. Whaley and G. M. Long, Institute
of Gas Technology.
CITED PUBLICATIONS
THOMAS P. WHALEY
Consultant
GEORGE M. LONG
Institute of Gas Technology
The previous decade has seen a rapid shift in the mode of sulfur production
with native sulfur recovery by the Frasch process being eliminated in the
United States and in steep decline elsewhere (1). The increasing reliance on
sour natural gas (CH4, H2S, and CO2) throughout the World and the pro-
duction of low sulfur fuels in refineries has resulted in nearly all of the World’s
elemental sulfur coming from conversion of H2S. This trend will continue as
very large reserves of oil with associated sour gas and new nonassociated
sour natural gas reservoirs are developed in Kazakhstan and in the Middle
East region (1). In addition, high sulfur-content bitumen recovered from
Canada and Venezuela is expected to become a major source of synthetic
crude oil. The accessible bitumen in Alberta, equivalent in volume to all
crude oil reserves in Saudi Arabia, is upgraded to yield synthetic crude oil
with associated production of large amounts of sulfur. By 2010, it is expected
that production of 3 million bbl/day of synthetic crude oil in northern Alberta
will yield 5 million tons of sulfur per annum (1). With these factors in mind,
the bulk of this article will describe processes for removal of H2S from large
volume gas streams that originate from oil and sour natural gas recovery and
refining.
2.1. General Comments. Removal of H2S and CO2 from industrial gas
streams is practiced on a very large scale in the natural gas, refining, and
hydrogen-production industries. Most processes are based on a closed recircu-
lation system that includes absorption and chemical reaction to remove the H2S
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1921120603011615.a01.pub2
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 873
and CO2, then regeneration of the treating agent. Chemical solvents, eg, aqu-
eous ethanolamines and a variety of physical and hybrid solvents, are used to
perform this function (2). For chemical solvents, the key aspect of the technol-
ogy (Fig. 1) is that H2S and CO2 are removed from high pressure gas streams in
a contactor at moderate temperatures (15–358C) and the chemistry is reversed
at higher temperature (1208C) in a regenerator yielding low pressure acid gas,
a mixture of H2S and CO2. As may be seen from the general scheme depicted in
Figure 1, heat exchangers are used to heat the rich solvent and to cool the aqu-
eous solvent to ambient temperatures such that residual levels of H2S do not
exceed 4 ppmv in the treated gas. The amount of CO2 removal depends on
the type of solvent used, operating conditions, and physical arrangement of
the contactor. In some processes, selective separation of H2S from CO2 is
achieved (3,4).
2.2. The Chemistry of H2S and CO2 Removal by Alkanolamines. All
amines are weak bases and produce aqueous solutions with pH 10–12, the
actual pH being dependent on the type of amine and the amount in solution
(Fig. 2). Absorption of H2S occurs by solubility and reaction with the free
amine and the protonated amine to yield soluble hydrosulfide species (Fig. 2).
The rates of these reactions are rapid with the result that removal of H2S from
a gas stream depends on mass transfer rates of the H2S from the gas-to-liquid
phases. Thus, the internal structure of an amine contactor is a critical design fea-
ture that along with countercurrent downflow of regenerated amine through the
contactor, engenders highly efficient H2S removal.
874 SULFUR AND HYDROGEN SULFIDE RECOVERY Vol. 1
Fig. 2. The uptake of H2S by aqueous amines. The efficiency of H2S removal depends on
the concentration of [RNH3þ][OH] (gas–liquid contact requires countercurrent flow of
gas and liquids).
O ‘bicarbonate’
RNH2
(b)
[RNHCO–][RNH+3]
=
O
‘carbamate’
(a)
Fig. 3. The uptake of CO2 by aqueous amines. The carbamate pathway is favored for
primary and secondary amines only; tertiary amines take up CO2 by bicarbonate route
only.
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 875
H H
‘ethanol’ ‘diglycol’
H CH 3 H CH 3
‘isopropanol’
Note that rich loading values are a function of pressure. Values given in Table 1
are good for absorber pressures > 2700 kPa. Rich loading values should be
checked against equilibrium values available in open literature.
2.3. Hybrid and Physical Solvents. As is implied by the name, physi-
cal solvents, eg, glycols and NFM (Fig. 5) remove H2S and CO2 from industrial
gas streams without chemical reaction (5,6). Advantages of these solvents are
they can accept higher amounts of H2S and SO2 per volume of solvent; less
energy is required to regenerate the solvent; and these solvents possess
increased efficiency over alkanolamines for removal of COS and mercaptans,
two minor components sometimes present in industrial gases. A hybrid solvent,
a mixture of an ethanolamine and a physical solvent, is used when a combination
of the properties of two solvents is required (5). Overall, nearly all application
needs can be met by integration of design and solvent choice.
One limitation of both amine and physical solvent processes is that the
loaded solvent dissolves some CH4 or other hydrocarbons. Thus, the regenerated
Hybrid:
Sulfinol - Sulfolane / H2O / amine (DIPA, MDEA )
Physical:
Selexol - dimethylethers of polyethylene glycols
CH3 O(CH 2 CH2 O)n CH3 (n = 3 – 9)
acid gas contains H2S, CO2, and CH4, and aromatic and larger aliphatic hydro-
carbons if these materials are present in the raw gas. Usually, the hydrocarbon
content of the acid gas does not exceed 1 vol.%, but even this small amount can
have a significant impact on the operation of downstream sulfur recovery
units. Physical solvents typically absorb significantly more hydrocarbons than
chemical solvents.
1 1
H2 S þ O2 ! S8 þ Heat
2 8
4 S2 Ð S8 þ Heat
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 877
2H2S: SO2
Fig. 6. Simplified process scheme for the modified-Claus sulfur recovery process.
PRACTICAL 100 98 %
LIMIT 95 %
CONV 3
CATALYTIC
205°C THERMAL STAGE
CONV 2 STAGE
225°C
90 89 %
CONV 1
305°C
% RECOVERY
80 R
R
TGCU c
R
70 c
67 %
98 % c
95 %
c WHB F
99 % + 89 %
60 67 %
900° – 1200°
LIQUID SULFUR
50
200 300 400 600 800 1000 1200
TEMPERATURE (°C)
OXIDATION
original –2 0
CAT.
Claus H 2S + 1/2 O 2 1/8 S8 + H 2O + heat (209 kJ)
process
SINGLE STEP ‘PARTIAL’ OXIDATION
MODIFIED - CLAUS:
SPLIT THE PROCESS INTO TWO STEPS
–2 +4 CAT. 0
2 H 2S + SO 2 low T
3/8 S 8 + 2 H 2O + heat (108 kJ)
3
2 H2 S þ SO2 Ð S8 þ 2 H2 O þ Heat
8
flame zone
acid gas Anoxic zone
(H2S/CO2/H2O/CH4)
2 H 2S + O 2 S 2 + 2 H 2O (3) 2.0
Carbon dioxide can be a major constituent of acid gas stream, along with CH4
and other hydrocarbons that are present as a result of their solubility in rich
amine systems. The highly reducing conditions of the anoxic region of the com-
bustion chamber lead to the formation of COS and CO from CO2, as follows:
CO2 þ H2 ! CO þ H2 O
CO2 þ H2 S ! COS þ H2 O
2 CO þ S2 ! 2 COS
2 CO2 þ S2 ! CO þ COS þ SO2
Products
3/2 S2 + 2 H 2O 2 H 2S + SO 2
S 2 + 2 H 2O H 2S + H 2 + SO 2 (favoured)
2 H 2S 2 H 2 + S2
2 H 2 + SO 2 2 H 2O + 1/2 S 2
CH 4 C + 2 H2
C + H 2O CO + H 2
Endothermic reactions,
CO + H 2O CO 2 + H 2 a source of hydrogen
Overall : CH 4 + 2 H 2O CO 2 + 4 H 2
Also: CH 4 + 2 S 2 CS 2 + 2 H2S
C + S2 CS 2
CS2 þ 2 H2 O ! 2 H2 S þ CO2
COS þ H2 O ! H2 S þ CO2
CS2 þ SO2 ! CO2 þ 3=2 S2
2 COS þ SO2 ! 2 CO2 þ 3=2 S2
But, the large quantity of CO2 in the acid gas from natural gas operations
results in too low a temperature for complete CS2 and COS destruction. This
situation may have to be rectified in the downstream catalytic stages to avoid
low total sulfur recovery and excessive emissions.
Due to the presence of nitrogen compounds in nearly all crude oils, refinery
Claus plants must be configured to deal with an NH3-containing stream originat-
ing from the refinery hydrotreating units. In major conurbations, transport of
ammonia is highly regulated and banned in some jurisdictions. Therefore, tech-
nology has been developed for conversion of the NH3 to N2 in the partially oxidiz-
ing regime of the Claus furnace typically with temperatures of 12008C
required. In most cases, specialized burners must be employed with injection
of the NH3-containing stream into the furnace chamber so as to avoid conversion
to NOx. The chemistry of this process (Figs. 13 and 14) has come under intense
882 SULFUR AND HYDROGEN SULFIDE RECOVERY Vol. 1
2 NH 3 + 3/2 O 2 N2 + 3 H 2O
2 NH 3 + O2 N2 + H 2 + 2 H 2O Exothermic
2 NH 3 + O2 N2 + 2 H 2 + H 2O
2 NH 3 N2 + 3 H 2 Endothermic
Consequences: H 2S + 3/2 O 2 SO 2 + H 2O
(excess)
4 NH 3 + 3 SO 2 2 N2 + 6 H 2O + 3/2 S 2
Rapid
reactions
2 NH 3 + SO 2 N2 + 2 H 2O + 1/2 S 2 + H 2
this adsorption is that a deficit of SO2 may be observed in the period shortly
after start-up of a new catalyst bed, resulting in the tail gas analyzer ‘‘overad-
justing’’ air or O2 supply to the main burner. It is generally accepted that the
catalyst bed must be preheated to the operating temperature (300–3408C for
the first converter) before introducing process gas to the catalyst. Preheat of
process gas leaving the upstream sulfur condenser is achieved by either O2-
fired combustion systems, heat exchange units or by electric heaters, as pre-
viously described.
The equilibrium limitation of the Claus reaction is illustrated by the Kp
values (Fig. 16) calculated for the 120–5008C range. The actual conversion to
PS83/8 PH2O2
Kp = K varies with temperature
PH2S2 PSO2
Temperature Kp
120°C 1.26 × 10 8
CATALYST
Al2O 3
2 H 2S + SO 2 2 H 2O + 3/8 S8 + heat
200 – 315 oC
EXOTHERMIC
EQUILIBRIUM vs KINETICS
%
CONV. RATE
TEMP TEMP
sulfur is dependent on the amount of water already in the process gas from the
upstream amine unit and that produced in the furnace stage. Clearly, it would be
very beneficial to configure a Claus process that includes both water removal and
sulfur condensation after all stages but, as yet, no technology has been developed
that would allow this improvement to be implemented. The fact that a catalyst
has to be used at all illustrates that the gas-phase Claus reaction is too slow
for practical purposes at 120–3508C. Moreover, as depicted in Figure 17,
the practical operation of a Claus converter is actually an interplay between
three factors: equilibrium limitations (Fig. 7); reaction kinetics; and sulfur
dew-point control; with these last two factors being the overriding issues for
commercial operation.
A valuable insight into the catalytic chemistry of the Claus process is
obtained from the observation that sulfate and thiosulfate species can be
found on alumina and titania that have been used in Claus service (14). More-
over, as a general rule of thumb, the amount of sulfate found on the catalyst
increases as the temperature of operation decreases (Fig. 18). In the past, it
was generally accepted that sulfate was an artifact of O2 ingress into the
system either as a result of poor combustion in the furnace stage or as a conse-
quence of direct fired reheaters upstream of the converter. However, sulfate-
contaminated alumina is found in plants in which the furnace operates
efficiently and which use indirect (no O2) reheat of the gas. Of course, unless
the main burner is operating so inefficiently as to allow air/O2 to stream
through the furnace and the WHB, one may well ask the question as to how
O2 could possibly survive a 12008C reaction chamber that always contains
excess fuel (H2S and S2).
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 885
Al2 O3 Al2 O3
A mechanistic scheme for the reaction of H2S and SO2 (Fig. 19) reveals
that sulfate forms as part of the catalytic cycle and should reach a steady-state
concentration on the catalyst surface that is dependent on the temperature and
the amount of H2S in the process gas. Laboratory studies confirmed that sulfate
SO 2 + H2O [ H 2SO3] 2H + + SO 3 2–
S2 O3 2– + SO 2 S3O 52–
S3O5 2– + H 2 S H S4O4 + OH
H S4 O4 + H 2S [ S3 ] + H 2O + H S 2 O3
H S2 O3 + OH H 2O + S 2 O3 2–
OR H S2 O3 + OH SO 4 2– + H 2 S
SO4 2– + H2S H2 O + S 2 O3 2–
S8 Formation ~ S6
2 [ S3 ] S6 Sx
x = 6, 7, 8, ………26
predominant species
was always produced during the Claus reaction on alumina and titania under
conditions that rigorously excluded O2.
CS2 þ 2 H2 O ! 2 H2 S þ CO2
COS þ H2 O ! H2 S þ CO2
Fig. 20. (a) Average bed weight % SO42 as a function of H2S:SO2 ratio in the sulfating
environment at 3208C. (b) Steady-state CS2 conversion as a function of H2S:SO2 ratio in
the sulfating environment at 3208C.
operation of Claus plants. The first is that plants operated at deliberately high
H2S/SO2 ratios, as is the case for those with direct oxidation or reducing tail
gas systems, will experience superior CS2 conversions (usually > 85%). Second,
it would seem that the practice of catalyst rejuvenations for sulfate removal at
high ratio for plants that normally operate at a standard 2:1 ratio has no
merit since the sulfate level will simply readjust to its natural sulfate level
almost immediately on reestablishment of the working H2S/SO2 ratio. This
conclusion has been justified by observations in the laboratory (13) that reveal
‘‘normal’’ sulfate levels are reestablished within 2 h of restart of a converter at
the new H2S/SO2 ratio. This comment should not be taken to infer that high
ratio H2S/SO2 soaks are of no value as it has been shown that ‘‘carsul’’ can be
removed by this process.
The level of CS2 conversion is then determined by the H2S/SO2 ratio, the
temperature of operation of the converter, and the type of catalyst. As already
discussed, high ratio operation increases conversion of CS2 and COS, but the con-
verter must also be operated at appropriate temperatures and space velocities in
order to achieve optimum CS2 conversion. In practice, a high quality alumina
will enable 90% CS2 and total COS conversion at 3308C and a 1000 h1 space
velocity. If CS2 is anticipated to be a significant problem, the first converter
888 SULFUR AND HYDROGEN SULFIDE RECOVERY Vol. 1
Fig. 21. The CS2 conversion by direct SO2 reaction. Titania more active than alumina.
could be designed to operate at a lower space velocity. The more usual strategy
has been to utilize titania as all or part of the catalyst in the first converter as it
is considerably more active than alumina for CS2 conversion. Although titania is
very efficient in catalyzing the hydrolytic conversion of CS2 and COS, it has been
established that it also promotes CS2 conversion by reaction with SO2. As may
be seen from the data plotted in Figure 21, titania very effectively promotes
the CS2/SO2 conversion reaction, but alumina is, in comparison, only a weak
catalyst for this process (15). Most often, when titania is used as a composite
catalyst bed with alumina, it is used as the lower layer of catalyst in order to
protect the more expensive catalyst from particulate matter entering with the
process gas.
4.3. Catalyst Degradation by BTX. The formation of sulfate as part of
the Claus reaction catalytic cycle is also an important factor in controlling cata-
lyst degradation by BTX. The abbreviation BTX is the acronym given to a collec-
tion of aromatic hydrocarbons that are sometimes found in Claus process gas
(Fig. 22). These hydrocarbons may be present in the acid gas in refineries and
CH 3 CH 3
CH 3
CH 3
CH 3 CH 3 CH 3
B T X
1st Converter Conditions: 320° – 350°C, Al2 O3 catalyst
“cracking” Soot(Carbon)
BTX +
+ S
Carsul (C-S polymer)
Fig. 23. A mechanism for catalyst degradation by BTX. *Immature carsul; substitution
pattern changed for artistic convenience.
in sour gas processing plants, although their presence tends to have a much
greater consequence in the gas plants where CO2 is present in high concentra-
tions. In these cases, low furnace temperatures or some acid gas by-pass to the
first converter results in some BTX entering the catalytic units. The overall pro-
blem of having BTX in the gas is that chemical interaction with sulfur results in
polymeric carbon–sulfur junk (carsul) accumulating in the pore structure of the
catalyst. Xylenes are, by far, the most troublesome aromatic, but toluene also
results in significant degradation of the catalyst. The options of removing BTX
from an acid gas or operating the reaction furnace at a high enough temperature
to completely destroy these aromatics are technically difficult and expensive (16).
Designers are advised to consider measures in the gas treating unit to reduce
BTX pickup.
Sulfate on the catalyst appears to be the driving force for BTX degradation
because the countercation to sulfate on the surface is a proton (Hþ) (it could also
be an aluminum cation) and this Brønsted acidity (Lewis acidity in the case of
an aluminum cation) initiates reactions with the electron-rich aromatics, BTX
(Fig. 23). Since the positively charged aromatic species are surrounded by a
sea of sulfur, rapid C S bond formation proceeds with production of carsul
that occludes the catalyst surface. Laboratory experiments confirmed that
sulfate was a key player in this chemistry as it was seen that alumina catalyst
that had not been exposed to Claus conditions (no sulfate) did not promote
reaction with xylene or other aromatics (17). Thus, any oxide that forms a stable
sulfate at Claus reactions conditions will promote both H2S/SO2 conversion and
BTX degradation.
4.4. Tail Gas Processing. A standard Claus sulfur recovery system
consists of a furnace and two or three catalytic converters, which, as is evident
for the theoretical sulfur recovery (Fig. 7), limits total sulfur recovery to
95–98%. Environmental considerations have lead to legislated limits that sti-
pulate >99% recovery in most regions of the World, although the total sulfur
recovery required is often linked to the capacity of the plant. Many regions of
the world now require a minimum 99.8% recovery of all sulfur contained in
the SRU feed gas. The regulations have lead to the development of a variety
of ‘‘tail gas’’ treatment options to achieve recoveries > 99% (2).
890 SULFUR AND HYDROGEN SULFIDE RECOVERY Vol. 1
Fig. 24. The chemistry of subdew point Claus tail gas operation.
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 891
liquid S 8
Another major tail gas process for increasing the total sulfur recovery beyond 99%
are technologies based on direct oxidation of H2S using catalytic methods (Fig. 26)
(18). It is very important to remove as much SO2 as possible from the tail gas before
the direct oxidation step as this species is not treated and, consequently, will
contribute to emissions in the final incinerator step. As in the reducing tail gas
technologies, the plant is operated to achieve a high H2S/SO2 ratio (4–8) before
the tail gas enters the direct oxidation unit. This is accomplished by restricting
the air supply to the main burner of the Claus furnace and, in an adaptation
recently introduced to commercial operation, by placing a layer of hydrogenation
catalyst at the bottom of the catalyst bed preceding the direct oxidation reactor.
This new strategy then decreases the quantity of SO2 entering the oxidation reac-
tor, possibly allowing a lower amount of SO2 leaving the converter in comparison to
equilibrium prediction. The chemistry of the hydrogenation on the reducing cata-
lyst in the second converter may be similar to that described for the low tempera-
ture reduction described in the previous section.
The direct oxidation catalyst is based on iron species deposited on silica
although various modifiers are used to limit secondary reactions that result in
SO2 formation. Silica is chosen because it is an oxide that does not form a stable
sulfate and, thus, cannot promote the Claus reaction. This facet of catalyst
behavior is important because use of a Claus active support would result in
reconversion of sulfur to H2S and SO2 by the reverse Claus reaction. The
catalytic cycle of the direct oxidation reaction formally involves an Fe(III)/Fe(II)
couple that is cycled by reaction of Fe(II) species with O2 added in controlled
amounts upstream of the direct oxidation reactor. Apparently, the catalyst can
only be kept in an active state if excess O2 is present in the tail gas. However,
this feature of the process imposes a small limitation on overall recovery as sul-
fur vapor produced in the direct oxidation reaction can be converted to SO2 in
homogeneous or surface-catalyzed chemistry. Overall, direct oxidation tail gas
treatment can result in 98.8–99.5% total recovery dependent on the original
acid gas under treatment.
Another popular tail gas process that enables >99.8% total sulfur recovery is
reduction of all sulfur species to H2S with recycle of H2S to the front end of
the plant. A key feature of the process is that the plant is controlled to result
in a high H2S/SO2 ratio (4–8) to limit the amount of SO2 entering the tail gas
unit. This technology (Fig. 27) was introduced in the late 1960s at a time when
there was little concern about CO2 emissions. Operation of the reducing gas
generator/reheating unit at the front end of the plant, handling sour water
from the quench unit and recycle of the H2S, utilizing an amine unit, all
have associated energy demands that result in considerable CO2 emission.
Note that the amine unit, not shown in Figure 27, processes the gas from the
top of the quench column. The alumina-CoS/MoS2 catalyst employed in the
hydrogenation unit converts all sulfur species to H2S although in the case of
CS2 and COS, it is thought that conversion occurs via hydrolysis reactions
(Fig. 28). In original versions of the process, the H2 and CO produced in the
Claus furnace was supplemented by operation of a partial combustion system
(labeled reheater in Fig. 27) in which CH4 is burnt with a restricted air supply.
This unit also serves to reheat the gas to 3408C, the temperature required for
operation of the catalytic unit.
In a drive to make the process more efficient, low temperature catalysts
have been introduced that operate at 2408C (19). This innovation has the
very important effect of replacement of the partial oxidation system with an
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 893
Indirect re-heater
Hydrogenation Quench
reactor column
Gas cooler
condensate
to SWS
∗ Catalyst active at lower temperatures
Fig. 27. Low temperature tail gas reduction.
The Claus reaction of H2S and SO2 can be carried out in liquid solvent, eg, poly-
ethylene glycols, when a basic catalyst in present is the medium (2). The only
894 SULFUR AND HYDROGEN SULFIDE RECOVERY Vol. 1
Oxidation of Claus tail gas converts all sulfur compounds to SO2. In many
cases, this simply represents the tail gas incineration step before final release
of the tail gas to the atmosphere, but it is possible to capture that SO2 in
various forms (2). Such technologies are not used widely, but may be applied
more extensively in the future. One adaptation involves conversion of the
SO2 to useful inorganic derivative taking the SO2 directly from the tail gas
while another utilizes a diamine solvent to capture the SO2 for recycle to the
front end of the Claus plant. These technologies may be limited by the markets
for the inorganic derivatives in the former case and by issues of economics and
process consequences of returning large volumes of SO2 to the furnace in the
diamine technology.
11. Acknowledgments
The author would like to thank his colleagues at ASRL, particularly, Dr. Norman
Dowling and Dr. Minming Huang, for support over many years. Also thanks
Vol. 1 SULFUR AND HYDROGEN SULFIDE RECOVERY 895
are due to the reviewers of this article for their improvements to the original
manuscript.
BIBLIOGRAPHY
‘‘Sulfur Removal’’ in ECT 3rd ed., Vol. 22, pp. 267–292, by J. R. West, Texagulf Inc.;
‘‘Sulfur Removal and Recovery’’ in ECT 4th ed., Vol. 23, pp. 432–452, by M. Capone,
Exxon Engineering; ‘‘Sulfur Removal and Recovery’’ in ECT (online), posting date:
December 4, 2000, by M. Capone, Exxon Engineering.
CITED REFERENCES
P. D. CLARK
University of Calgary
CORROSION AND
CORROSION CONTROL
1. Introduction
Ores are mined and are then refined in an energy intensive process to produce
pure metals, which in turn are combined to make alloys. Corrosion occurs
because of the tendency of these refined materials to return to a more thermody-
namically stable state (1–3). The key reaction in corrosion is the oxidation or ano-
dic dissolution of the metal to produce metal ions and electrons
M k!
1
Mnþ þ n e ð1Þ
The ions, Mnþ, formed by this reaction at a rate, k1, may be carried into a bulk
solution in contact with the metal, or may form insoluble salts or oxides. In order
for this anodic reaction to proceed, a second reaction which uses the electrons
produced, ie, a reduction reaction, must take place. This second reaction, the
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0315181813151801.a01.pub2
Vol. 1 CORROSION AND CORROSION CONTROL 897
cathodic reaction, occurs at the same rate because the electrons produced by the
anodic reaction must be consumed by the cathodic reaction to maintain electro-
neutrality. Therefore, Ic ¼ Ia, where Ic and Ia are the cathodic and anodic cur-
rents, respectively. The cathodic reaction, in most cases, is hydrogen evolution
or oxygen reduction.
The four elements necessary for corrosion are an aggressive environ-
ment, an anodic and a cathodic reaction, and an electron conducting path
between the anode and the cathode. Other factors such as a mechanical
stress also play a role. The thermodynamic and kinetic aspects of corrosion
determine, respectively, if corrosion can occur, and the rate at which it does
occur.
3. Manifestations of Corrosion
The most common form of corrosion is uniform corrosion, in which the entire
metal surface degrades at a near uniform rate (1 – 4). Often the surface is
covered by the corrosion products. The rusting of iron (qv) in a humid atmo-
sphere or the tarnishing of copper (qv) or silver alloys in sulfur-containing
environments are examples. High temperature, or dry, oxidation, is also
usually uniform in character. Uniform corrosion, the most visible form of cor-
rosion, is the least insidious because the weight lost by metal dissolution can
be monitored and predicted.
An especially insidious type of corrosion is localized corrosion (1–3,5,6),
which occurs at distinct sites on the surface of a metal while the remainder
of the metal is either not attacked or attacked much more slowly. Localized
corrosion is usually seen on metals that are passivated, ie, protected from corro-
sion by oxide films, and occurs as a result of the breakdown of the oxide film.
Generally the oxide film breakdown requires the presence of an aggressive
anion, the most common of which is chloride. Localized corrosion can cause con-
siderable damage to a metal structure without the metal exhibiting any appreci-
able loss in weight. Localized corrosion occurs on a number of technologically
important materials such as stainless steels, nickel-base alloys, aluminum, tita-
nium, and copper.
Two types of localized corrosion are pitting and crevice corrosion. Pitting
corrosion occurs on exposed metal surfaces, whereas crevice corrosion occurs
within occluded areas on the surfaces of metals such as the areas under rivets
or gaskets, or beneath silt or dirt deposits. Crevice corrosion is usually associated
with stagnant conditions within the crevices. A common example of pitting cor-
rosion is evident on household storm window frames made from aluminum
alloys.
Another type of corrosion is dealloying which has also been called parting
or selective leaching. Dealloying (1–4) is the preferential removal of one of
the alloying elements from an alloy resulting in the enrichment of the other
alloying element(s). Common examples are the loss of zinc from brasses (dezin-
cification) and the loss of iron from cast irons (graphitization).
Corrosion may also appear in the form of intergranular attack, ie, preferen-
tial attack of the boundaries between the crystals (grains) in metals and alloys
898 CORROSION AND CORROSION CONTROL Vol. 1
(1–5). Intergranular attack generally occurs because the grain boundary and the
grain have different corrosion tendencies, ie, different potentials. Intergranular
corrosion often leads to a loss in strength or ductility of the metal.
Corrosion also occurs as a result of the conjoint action of physical processes
and chemical or electrochemical reactions (1–5). The specific manifestation of
corrosion is determined by the physical processes involved. Environmentally
induced cracking (EIC) is the failure of a metal in a corrosive environment and
under a mechanical stress. The observed cracking and subsequent failure would
not occur from either the mechanical stress or the corrosive environment alone.
Specific chemical agents cause particular metals to undergo EIC, and mechanical
failure occurs below the normal strength (yield stress) of the metal. Examples are
the failure of brasses in ammonia environments and stainless steels in chloride
or caustic environments.
When a stress is cyclic rather than constant, the failure is termed corrosion
fatigue. Fretting corrosion results from the relative motion of two bodies in con-
tact, one or both being a metal. The motion is small such as a vibration. Erosion
corrosion results from the action of a high velocity fluid impinging on a metal
surface. Metals and alloys can also experience cracking in liquid metal environ-
ments. This form of corrosion is referred to as liquid metal cracking (LMC)
(1,2,7). For example, mercury promotes cracking in highly stressed copper alloys
and high strength aluminum alloys, and liquid copper promotes cracking of
steels and stainless steels. Titanium and nickel alloys are also susceptible to
LMC in specific environments.
Galvanic corrosion (1–4) occurs as a result of the electrical contact of differ-
ent metals in an aggressive environment. The driving force is the electrode
potential difference between the two metals. One metal acts principally as a
cathode and the other metal as the anode. Galvanic corrosion can result from
the presence of a second phase in a metal. An example is manganese sulfide
[18820-29-6], MnS, inclusions in steels. Galvanic corrosion is also an important
consideration for the environmental stability of metal matrix composites (qv)
such as graphite reinforced aluminum. Galvanic corrosion can accelerate many
of the other types of corrosion.
Microbiologically influenced corrosion results from the interaction of micro-
organisms and a metal (3,8–11). The action of microorganisms is at least one of
the reasons why natural seawater is more corrosive than either artificial sea-
water or sodium chloride solutions. Microorganisms attach to the surfaces of
metals and can, for example, act as diffusion barriers; produce metabolites
that enhance or initiate corrosion; act as sinks or sources for species involved
in cathodic reactions, such as oxygen and hydrogen; increase the pH at the sur-
face as a result of photosynthesis; or decrease the pH by production of acid meta-
bolites. A more detailed discussion of the various forms of corrosion may be found
in the literature (1–4).
4. Origin of Corrosion
In the presence of oxygen and water the oxides of most metals are more thermo-
dynamically stable than the elemental form of the metal. Therefore, with the
Vol. 1 CORROSION AND CORROSION CONTROL 899
The rates of these two reactions must be the same to preserve electroneutrality,
and additionally, some of the zinc surface must be exposed to the solution. That
is, if a continuous coating of copper were to be electroplated on the zinc surface
such that none of the zinc were exposed, further oxidation would be prevented.
Corrosion occurs even if the two reactants involved are not at standard con-
ditions. In this case, the nonstandard equilibrium potential for each reaction,
often referred to as the reversible potential, can be calculated from the Nernst
equation. Additional information on thermodynamic aspects of corrosion can be
found in the literature (1–4,12,13).
The larger the potential difference between the equilibrium potentials of
the anodic and cathodic reactions involved in corrosion, the greater the driving
force which is more traditionally represented as the change in Gibbs free energy,
DG. The free-energy change for a reaction is related to the difference in the equi-
librium potentials of the two reactions, cathodic minus anodic by
G ¼ nFE ð2Þ
where n is the number of electrons involved in the overall reaction, F is the Fara-
day constant, and E, the cell voltage or the difference in the equilibrium potential
values of the cathode and the anode, respectively, given in volts.
The standard reduction potential of oxygen (taken at pH = 0)
O2 þ 4 Hþ þ 4 e ! 2 H2 O ð3Þ
is 1.229 V, a much more positive number than the metal reactions given in Table 1.
Therefore, the free-energy change for a coupling of any of the metal reactions,
considered as the anode, and the oxygen reaction, considered as the cathode,
would be a large negative number; ie, the reaction is thermodynamically highly
favored to proceed. Reduction of oxygen is the most common cathodic reaction
in the corrosion of metals in neutral and alkaline media. In acidic media the
common cathodic reaction is hydrogen evolution
2 Hþ þ 2 e ! H2 ð4Þ
The standard potential for this reaction is 0.000 V. Most of the metal reac-
tions have standard potentials below this value.
1.6 0
–2
Fe3+ –4 2 –?
(B) –6 FeO4
FeOH2+
0.8
Fe (OH)2+
E, V
(A) Fe (OH)3
0 Fe2+
Fe (OH)2
– 0.8
Fe HFeO2–
–1.6
0 2 4 6 8 10 12 14 16
pH
Fig. 1. Pourbaix diagram for the iron–water system at 258C, considering as solid
substances only Fe, Fe(OH)2, and Fe(OH)3, where (), line A, represents the hydrogen
reaction line; ( ), line B, the oxygen line(13). The values 0,2, 4, 6 are molar
concentrations of the ions involved in each reaction.
If the potential of a metal surface is moved below line A, the hydrogen reac-
tion line, cathodic hydrogen evolution is favored on the surface. Similarly a
potential below line B, the oxygen reaction line, favors the cathodic oxygen
reduction reaction. A potential above the oxygen reaction line favors oxygen evo-
lution by the anodic oxidation of water. In between these two lines is the region
where water is thermodynamically stable.
The Pourbaix diagram is constructed from data for a variety of reactions.
The expressions for these reactions can be (1) potential dependent and pH inde-
pendent, (2) pH dependent and potential independent, (3) both pH and potential
dependent, or (4) both pH and potential independent. The dependence is deter-
mined by whether the reaction of interest is electrochemical or chemical and
whether or not it involves the components of water. From this diagram it is pos-
sible to determine if there is a driving force for corrosion at any given potential–
pH combination and what the corresponding cathodic reaction is likely to be.
Regions where insoluble corrosion products form are indicated. These are
regions where passivity is possible. Passivity generally results from the forma-
tion of an oxide film on a metal surface. The film can dramatically reduce
further corrosion of the metal. Regions of immunity, where there is no thermo-
dynamic driving force for the metal to corrode, can also be determined. Figure 2
presents a simplified picture of the Pourbaix diagram of Figure 1 showing the
902 CORROSION AND CORROSION CONTROL Vol. 1
1.6
0.8
E,V
Passivation
0
Corrosion
– 0.8
Corrosion
Immunity
–1.6
0 4 8 12 16
pH
Even in uniform corrosion, a corroding metal surface has numerous local anodes
and cathodes. The sites of these local reactions may be fixed by microstructural
features or may change as corrosion proceeds. The oxidation reaction at anodic
sites on the metal surface can be represented as in equation 1. A corresponding
reduction reaction must be occurring at cathodic sites. In acidic solutions this
would likely be hydrogen evolution as shown in equation 4. The potentials of
these two reactions would be moved toward each other, away from the respective
equilibrium potentials, and the metal surface would assume an overall uniform
potential. The dotted lines in Figure 3 illustrate the relationships of the indivi-
dual electrochemical reaction kinetics as influenced by electrochemical potential.
Vol. 1 CORROSION AND CORROSION CONTROL 903
e)
–
A
2
+ +
H
2
(H 2
10 100 1000
Log(current density)
Fig. 3. Hypothetical Evans diagram and polarization curve for a metal corroding in
an acidic solution, where point A represents the exchange current density, i0, for the
hydrogen electrode at equilibrium; point B, the exchange current density at the reversible
2þ
or equilibrium potential, Erev, for M ! M þ 2e ; and point C, the corrosion current
density, icorr, at the open-circuit corrosion potential, Ecorr. See also discussion in text.
These lines also illustrate the interaction of the two reactions on the metal sur-
face. Diagrams of this sort are referred to as Evans or mixed potential diagrams.
The terms are used interchangeable in the corrosion community.
The two dashed lines in the upper left hand corner of the Evans diagram
represent the electrochemical potential vs electrochemical reaction rate
(expressed as current density, ie, current divided by surface area) for the oxida-
tion and the reduction forms of the hydrogen reaction. At point A the two are
equal, ie, at equilibrium, and the potential is therefore the equilibrium potential,
E0, for the hydrogen reaction for the specific conditions involved. Note that the
reaction kinetics are linear on these axes. The change in potential for each dec-
ade of log current density is referred to as the Tafel slope (14). The Tafel slope is
related to the activation energy for the reaction. A more detailed treatment of
Tafel slopes can be found elsewhere (1–4,14,15).
The value of the current density where the oxidation and reduction forms of
a reaction cross, ie, at point A or point B in Figure 3, is referred to as the
exchange current density, i0. For any equilibrium point, the exchange current
density is the rate of the two forms of the reaction, ie, cathodic and anodic, occur-
ring at equilibrium. The greater the exchange current density the more kineti-
cally favored is the reaction. The exchange current density for the reduction of
hydrogen on platinized platinum is 103A/cm2, whereas on mercury it is
1013A/cm2, indicating that platinum is a good catalyst for hydrogen evolution.
The exchange current densities for most electrochemical reactions fall in
between these two values. Exchange current density and Tafel slope values for
904 CORROSION AND CORROSION CONTROL Vol. 1
Magnesium
Zinc
Beryllium
Aluminum alloys
Cadmium
Mild steel cast iron
Low alloy steel
Austenitic nickel cast iron
Aluminum bronze
Naval brass, yellow brass, red brass
Tin
Copper
Pb–Sn solder (50/50)
Admiralty brass, aluminum brass
Manganese bronze
Silicon bronze
Tin bronzes (G & M)
Stainless steel-types 410, 416
Nickel silver
90 –10 Copper–nickel
80–20 Copper–nickel
Stainless steel-types 430
Lead
70 –30 Copper–nickel
Nickel–aluminum bronze
Nickel–chromium alloy 600
Silver bronze alloys
Nickel 200
Silver
Stainless steel-types 302, 304, 321, 347
Nickel–copper alloys 400,K-500
Stainless steel-types 316, 317
Alloy 2C stainless steels cast and wrought
Nickel–iron–chromium alloy 825
Ni–Cr–Mo–Cu–Si alloy C
Titanium
Ni–Cr–Mo alloy C
Platinum
Graphite
Fig. 4. Galvanic series in flowing seawater. Certain alloys may become more active in
low velocity or poorly aerated seawater and the potentials exhibited under these condi-
tions are indicated as darkened potential ranges (1).
potential to intentionally pull the potential of the higher potential metal down
and thus decrease the corrosion rate, it is called cathodic protection. This method
of corrosion mitigation is common for underground pipelines (qv), residential hot
water heaters, and hulls and tanks of ships. The lowering of potential can also
be achieved by the application of external electrical current, ie, impressed cur-
rent cathodic protection. Electrons are pumped into the structure to lower the
potential and thus discourage anodic (corrosion) reactions. A second electrode
(anode) is needed for this procedure and this electrode is often a highly inert
material such as platinum to prevent it from being corroded away. However,
scrap iron or steel is sometimes used as the impressed current anode. This
approach remains functional until the scrap metal anode corrodes away.
The lower potential metal in a galvanic couple does not always have its cor-
rosion rate accelerated. For metals that form a passive film, coupling with
another metal of higher potential can cause the film-forming metal’s potential
to shift from a value at which it corrodes to one at which it passivates and there-
fore corrodes less. When this is done intentionally the procedure is referred to as
anodic protection, ie, achieving protection by intentionally shifting the potential
in the positive direction. Anodic protection is generally achieved by adding oxidi-
zers to the electrolyte or by an external electrical circuit.
The higher potential metal in a galvanic couple is not always rendered less
corrodible. A passive metal can sometimes be pulled out of its passive region and
into a more corrosive region. A metal can also have its potential pulled down to
the point where hydrogen reaction is possible and this can lead to one of the
forms of EIC called hydrogen induced cracking (HIC). Another example involves
organic coatings on metal structures. These coatings can be damaged by having
the potential of the structure shifted too far in the negative direction. This is one
of the consequences of excessive cathodic protection.
8. Environmental Effects
The environment plays several roles in corrosion. It acts to complete the electri-
cal circuit, ie, supplies the ionic conduction path; provides reactants for the
cathodic process; removes soluble reaction products from the metal surface;
and/or destabilizes or breaks down protective reaction products such as oxide
films that are formed on the metal. Some important environmental factors
include: the oxygen concentration; the pH of the electrolyte; the temperature;
and the concentration of anions.
Reduction of oxygen is one of the predominant cathodic reactions contribut-
ing to corrosion. Awareness of the importance of the role of oxygen was developed
in the 1920s (18). In classical liquid drop experiments, the corrosion of iron or
steel by drops of electrolytes was shown to depend on electrochemical action
between the central relatively unaerated area, which becomes anodic and suffers
attack, and the peripheral aerated portion, which becomes cathodic and remains
unattacked. In 1945, the linear relationship between rate of iron corrosion and
oxygen pressure from 0 to 2.5 MPa (0–25 atm) was shown (19).
The concentration dependence of iron corrosion in potassium chloride
[7447-40-7], sodium chloride [7647-14-5], andlithium chloride [7447-44-8]
Vol. 1 CORROSION AND CORROSION CONTROL 907
16.66
13.33
Loss in weight, mg
10.00
6.66
3.33
0 0.5 1.0 1.5 2.0
Normality of solution
Fig. 5. Corrosion–concentration curves for alkali chlorides where denotes LiCl; &,
NaCl; and ~, KCl (20).
solutions is shown in Figure 5 (20). In all three cases there is a maximum in cor-
rosion rate. For NaCl this maximum is at 0.5 N (3 wt%). Oxygen solubility
decreases with increasing salt concentration, thus the lower corrosion rate at
higher salt concentrations. The initial increase in the iron corrosion rate is
related to the action of the chloride ion in concert with oxygen. The corrosion
rate of iron reaches a maximum at 708C. As for salt concentration, the
increased rate of chemical reaction achieved with increased temperature is
balanced by a decrease in oxygen solubility.
The corrosion rate of iron in aerated water is also a function of pH and gen-
erally follows the pattern in Figure 6 (21). At pH 4–10 , the rate is controlled by
the availability of oxygen. In more acidic solutions (lower pH) the corrosion rate
is accelerated, and the reduction of hydrogen ion replaces the reduction of oxygen
1.00
0.75
Corrosion rate, mm/yr
0.50
A +
+
+
0.25 + + + +
+++
+
+ +
+
14 13 12 11 10 9 8 7 6 5 4 3 2
pH
Fig. 6. Effect of pH on corrosion of iron in aerated water at room temperature (21). Point A
is where H2 evolution begins. To convert mm/yr to mils per year (mpy), multiply by 39.37.
908 CORROSION AND CORROSION CONTROL Vol. 1
as the rate controlling cathodic reaction. Many metals follow approximately the
same behavior with the exception of those metals that dissolve to form ampho-
teric ions. Zinc forms the zincate ion, ZnO2
2 , which causes zinc to corrode exces-
sively above a pH 12; whereas Al forms the aluminate ion [11098-82-1], AlO2,
which increases the dissolution rate above a pH of 8.
Chlorides, which are ubiquitous in nature, play an important role in the
corrosion of metals. Chlorides and other anions also play an important role in
localized corrosion, ie, the breakdown of the insoluble protective reaction product
films, eg, passive films, that prevent corrosion of the underlying metal. A variety
of mechanisms attempting to explain the role of chloride in general and in loca-
lized corrosion have been proposed (6,22–25).
Very often the environment is reflected in the composition of corrosion
products, eg, the composition of the green patina formed on copper roofs over a
period of years. The determination of the chemical composition of this green
patina was one of the first systematic corrosion studies ever made (see COPPER).
The composition varied considerably depending on the location of the structure
as shown in Table 2 (26,27).
Fig. 7. Compositional and property linkages for the stainless steels (29).
a type 304 stainless steel containing 0.039% carbon for 10 h at 7008C will reduce
the chromium level from 19% to <13% in the region next to the carbide pre-
cipitate (29). The depleted areas are anodic to the grain interior and tend to dis-
solve thereby causing intergranular corrosion. Sensitization can also make
stainless steel more prone to stress–corrosion cracking (SCC). There are several
measures available to mitigate sensitization. Low carbon grades such as AISI
304L and 316L are available that have much less tendency toward sensitization.
Also, alloying additions of titanium or niobium and tantalum can be used to tie
up carbon. Another area of concern with stainless steel alloys is EIC and this
topic is treated in more detail later in this article for stainless steels as well as
for other selected alloys. A more detailed discussion of stainless steels may be
found in the literature (5,29).
9.2. Copper Alloys. Copper and its alloys have good corrosion resis-
tance in many nonoxidizing aqueous environments and atmospheric conditions.
Several copper alloys are exceptionally resistant to certain atmospheres. The
copper–nickel alloys are the most corrosion resistant of the commercial copper
alloys (3). The two most common Cu Ni alloys are 90% Cu 10% Ni and 70%
Cu 30% Ni. They have good resistance to corrosion in both fresh and salt
water (4).
The most known of the copper alloys is brass, which is produced by the addi-
tion of zinc. The zinc improves the mechanical properties by solid solution
strengthening. However, the zinc reduces the corrosion resistance. A variety of
names are associated with brasses with a particular zinc content: 40% Zn Cu
Vol. 1 CORROSION AND CORROSION CONTROL 911
alloys are referred to as Muntz metal; 30% Zn Cu alloys are referred to as yel-
low brass; and 15% Zn Cu alloys are termed red brass (2). Brasses are suscep-
tible to dealloying in the form of dezincification, ie, the preferential loss of zinc
from the alloy. Brasses having zinc concentrations of 15% or greater are prone to
dezincification and dezincification is generally more severe in brasses that have
two metallurgical phases. For the two-phase alloys, greater than 38 wt% Zn,
dezincification usually starts in the high zinc content b-phase and is followed
by dezincification of the lower zinc a-phase (4). Naval brass and admiralty
brass are similar in composition to Muntz metal and yellow brass, respectively,
but have 1 wt% tin added for improved resistance to dezincification. Red brass is
relatively resistant to dezincification but is more susceptible to impingement
attack than, for example, yellow brass (2).
Conditions that favor dezincification include stagnant solutions, espe-
cially acidic ones, high temperatures, and porous scale formation (2). Tin
tends to inhibit dealloying especially in cast alloys (4). Additions of small
amounts of arsenic, antimony, or phosphorus can increase the resistance to
dezincification. These elements are, however, not entirely effective in prevent-
ing the dezincification of the two-phase (a-b) brasses because dezincification of
the b-phase is not prevented (4). Another area of corrosion concern involves
applied or residual stresses from fabrication that can lead to EIC of brasses
in the form of SCC.
9.3. Aluminum Alloys. The corrosion resistance of aluminum, like
stainless steels, is provided by a passive film that protects the surface from degra-
dation by the environment. However, in environments that contain aggressive
anions such as chloride, the passive film can be degraded locally causing film break-
down and, depending on the geometry of the sample, pitting or crevice corrosion.
Copper, silicon, magnesium , zinc, and manganese are some common alloy-
ing additions to aluminum. Most of the alloying elements are added to aluminum
to produce alloys having improved mechanical properties. However, the
strengthening phases that result from the alloying can disrupt the passive
oxide layer on aluminum and lead to localized corrosion. Also, the second
phase constituents can produce local galvanic cells as discussed below. Common
alloying elements that are added to steels and many other metals to improve cor-
rosion resistance, such as Cr and Mo, have a very low solubility in aluminum and
result in second phase precipitates. The presence of the second phase constitu-
ents, in turn, lowers the corrosion resistance. A number of surface modification
techniques that produce single phase, metastable aluminum alloys have been
shown to increase the localized corrosion resistance as compared to pure alumi-
num (30–34). These techniques include ion implantation, laser alloying, and
magnatron sputtering. However, due to capital costs and the sizes of the sample
that can be treated, these techniques have not gained widespread use.
Aluminum alloys are susceptible to intergranular corrosion when precipi-
tates form. Intergranular corrosion in aluminum alloys is caused by a potential
difference, ie, galvanic couple. The formation of the anodic and cathodic areas in
the couple varies depending on the alloying additions (3). Reference (35) is a com-
pilation of potentials reported for intermetalic phases in aluminum alloys.
Stress–corrosion cracking is characteristically intergranular for aluminum
alloys. Wrought high strength aluminum alloys, whether the products are rolled,
912 CORROSION AND CORROSION CONTROL Vol. 1
the brass surface. The premature failure of yellow brass brackets in the humidi-
fier chamber of an air-conditioning system was traced to this latter cause (2). In
another case, SCC of 12% Ni 23% Zn Cu alloy, known as nickel brass, parts of
the Central Office Telephone equipment in Los Angeles occurred within 2 years
of installation for similar reasons. The source of nitrogen was the air in the Los
Angeles area which has high concentrations of nitrogen oxides and suspended
nitrates (see AIR POLLUTION). The nitrates settle as dust on the brass parts (2).
The susceptibility to SCC can be minimized by (1) proper alloy selection; (2) ther-
mal stress relief; (3) avoiding contact with ammonia and ammonia compounds;
and (4) using an inhibitor (2,4).
10.2. Aluminum Alloys. Both the 2000 and the 7000 series aluminum
alloys have experienced significant SCC in service. These high strength alloys in
the wrought form are highly textured, the grains are flattened and elongated into
specific crystallographic directions due to the processing. SCC behavior differs
greatly according to the direction of tensile stress with respect to the texture direc-
tions. The alloys are most vulnerable to SCC if stressed parallel to the short trans-
verse grain direction, ie, parallel to the thinnest dimension of the grain and most
resistant if stressed only parallel to the longest grain dimension. In practice, it is
the short transverse-direction stresses that cause SCC problems. Hence prudent
practice is to avoid designs in which high sustained stresses are imposed across
the short transverse direction. For example, one avoids an interference-fit
fastener, or a taper pin fastener oriented to stress a part across the vulnerable
texture unless the alloy is inherently of low susceptibility to SCC. Table 3 sum-
marizes alloys and tempers in various categories of susceptibility as judged from
specimens having short transverse, ie, most vulnerable, orientations.
In addition to the possibility of selecting alloys having minimum SCC sus-
ceptibility while retaining other properties as needed, there are other steps pos-
sible to reduce the SCC probability: (1) avoid designs that permit water to
accumulate; (2) avoid conditions in which salts, especially chlorides, can concen-
trate; and (3) where available and otherwise acceptable, use a clad alloy (2) (see
METAL SURFACE TREATMENTS).
10.3. High Strength Steels. Steels that owe their strength to heat treat-
ment, whether martensitic, precipitation hardened, or maraging, and whether
stainless or not, are susceptible to SCC in aqueous environments, including
water vapor. The primary factor in determining the degree of SCC susceptibility
of a given steel is its strength. There is no sharply defined threshold strength that
defines a threshold susceptibility, but above 1200 MPa (174,000 psi) yield strength
the problem becomes of increasing concern, until at 1400 MPa (200,000 psi)
susceptibility is generally very high. This does not mean that steels cannot be
used at such strength levels or even higher, but for steels to be used at these
strengths, great care must be exercised in design such that any tensile or bending
stresses are small, or that moisture is excluded from the surface of the steel.
High strength steels are also susceptible to HIC. Cadmium electroplating
is a useful protection measure for steels, but hydrogen either must not be
codeposited with cadmium, or if it is codeposited, as in the cyanide plating
bath, hydrogen must be safely redistributed by thermal treatment before the
high strength steel component is stressed. Otherwise the steel may experience
HIC.
914 CORROSION AND CORROSION CONTROL Vol. 1
11. Inhibitors
Corrosion inhibitors are substances that slow or prevent corrosion when added to
an environment in which a metal usually corrodes. Corrosion inhibitors are
usually added to a system in small amounts either continuously or intermittently.
The effectiveness of corrosion inhibitors is partly dependent on the metals or alloys
to be protected as well as the severity of the environment. The main factors, which
must be considered before application of a corrosion inhibitor to an aqueous sys-
tem, are the compatibility of the inhibitor and the metal(s), the salt concentration,
the pH, the dissolved oxygen concentration, and the concentration of interfering
species such as chlorides or metal cations. In addition, many inhibitors, most nota-
bly chromates, are toxic and environmental regulations limit use. (38).
Inhibitors act and are classified in a variety of ways (1,3,38,39). The classi-
fications used herein closely follow the discussion in (38). Types of inhibitors
discussed below include (1) anodic, (2) cathodic, (3) organic, (4) precipitation,
and (5) vapor-phase inhibitors.
11.1. Anodic Inhibitors. Passivating or anodic inhibitors produce a
large positive shift in the corrosion potential of a metal. There are two classes
of anodic inhibitors which are used for metals and alloys where the anodic
shift in potential promotes passivation, ie, anodic protection. The first class
includes oxidizing anions that can passivate a metal in the absence of oxygen.
Chromate is a classical example of an oxidizing anodic inhibitor for the passiva-
tion of steels. The second class of anodic inhibitors contains ions which need oxy-
gen to passivate a metal. Tungstate and molybdate, eg, require the presence of
oxygen to passivate a steel. The concentration of the anodic inhibitor is critical
for corrosion protection. Insufficient concentrations can lead to pitting corrosion
or an increase in the corrosion rate. The use of anodic inhibitors is more difficult
at higher salt concentrations, higher temperatures, lower pH values, and in some
cases, at lower oxygen concentrations (38).
11.2. Cathodic Inhibitors. Cathodic inhibitors act to retard or poison
the cathodic reaction or selectively precipitate onto cathodic areas producing dif-
fusion barriers to cathodic reactants, thereby reducing the rate of the cathodic
reaction. There are three types of cathodic inhibitors: (1) hydrogen poisons,
(2) oxygen scavengers, and (3) cathodic precipitates. Hydrogen poisons are chemical
species such as arsenic or antimony that retard the hydrogen reduction reaction.
Because the hydrogen poison slows the cathodic reaction, and because the catho-
dic and anodic reactions must proceed at the same rate, the whole of the corro-
sion process is slowed. A potentially serious drawback upon use of hydrogen
poisons is that the hydrogen on the surface can be more easily absorbed
into the metal or alloy and can lead to HIC in susceptible materials. Oxygen
916 CORROSION AND CORROSION CONTROL Vol. 1
metal substrate and the environment. However, the coating can also act to pro-
vide cathodic protection or by serving as holding reservoirs for inhibitors. Coat-
ings may be divided into organic, inorganic, and metallic coatings (1–3,42). For a
more in-depth coverage of this topic see (1–3) (see also COATINGS, MARINE; METAL
SURFACE TREATMENTS).
12.1. Coating Types. Organic coatings afford protection by providing a
physical barrier and are often used as holding reservoirs for corrosion inhibitors.
Organic coatings include paints, resins, lacquers, and varnishes (see also BARRIER
POLYMERS; PAINT). These coatings usually comprise four basic constituents: binder,
pigments and fillers, additives and solvents. The properties of the coating
strongly depend on the formulation of the coating (43). The effective application
of an organic coating requires (1) proper surface preparation of the metal
substrate, (2) selection of the proper primer or priming coat, and (3) selection
of the appropriate top coat(s). Poor performance of organic coatings most often
results from improper surface preparation or poor application (1). Maintenance
programs involving periodic inspection and repair are necessary for reliable cor-
rosion protection by organic coatings. Organic coatings probably protect more
metal on a tonnage basis than any other means of corrosion protection. As a
rule, however, organic coatings should not be used in environments that would
rapidly corrode the metal if the coating were compromised. For example, paint
would not be used to protect the inside of a tank car used for shipping hydrochlo-
ric acid because one small coating defect would result in the rapid perforation of
the tank car wall.
Inorganic coatings are also used to provide a barrier between the environ-
ment and the metal (1,2,42). Inorganic coatings include chemical conversion
coatings, glass (qv) linings, enamels (see ENAMELS, PORCELAIN OR VITREOUS),
and cement (qv). Chemical conversion coatings are produced by intentionally
corroding the metal surface in a controlled manner. This is done so as to pro-
duce an adherent corrosion product that protects the metal from further corro-
sion. Anodization of aluminum, one of the more commonly used conversion
coatings techniques, produces a protective aluminum oxide film on the alumi-
num metal. Another example of a chemical conversion coating is phosphatizing
for the protection of automobile bodies. Porcelain enamel coatings, that
are inert in water and resistant to most weather are routinely applied to
steel, cast iron, and aluminum (4). They are commonly seen on appliances
and plumbing fixtures. Glass-lined metals are used in process industries
where there is concern over corrosion or contamination of the product.
For example, glass-lined metals are used in the pharmaceutical industry (4).
Portland cement coatings have been used to protect steel and cast-iron water
pipes (2).
Often metallic coatings, in addition to providing a barrier between the
metal substrate and the environment, provide cathodic protection when the
coating is compromised (2,3). Metallic coatings and other inorganic coatings
are produced using a variety of techniques including hot dipping, electroplat-
ing, cladding, thermal spray techniques, chemical vapor deposition, or by sur-
face modification using directed energy (laser or ion) beams. Two classic
metallic coating techniques and a relatively recent technique are described
below.
918 CORROSION AND CORROSION CONTROL Vol. 1
Hot dipping is one of the oldest methods used for coating metals. In hot dipping,
coatings are applied by dipping the metal in a molten bath commonly of zinc, tin,
lead, or aluminum (2,4). The best known hot dipping procedure is that of coating
steel with zinc to produce galvanized steel. In addition to providing a barrier
between the steel substrate and the environment, the zinc acts as a sacrificial
anode and provides cathodic protection to the underlying steel when the coating
is breached.
Electroplating consists of immersing a metal in a plating bath and
electrochemically plating the solution species onto the metal substrate (1,4).
Electroplating is used to provide coatings for a variety of reasons including
corrosion protection, wear resistance, decoration, and to build up the dimen-
sions of a substrate. Additives to the plating bath can improve coating
properties such as grain size, strength, uniformity, and brightness. The elec-
troplate can be a single metal, an alloy, or a sequence of layers of different
composition. Coating thicknesses can range from on the order of thousandths
of a mil to 20 mils (1). Zinc, nickel, tin, and cadmium are the most commonly
plated materials on a tonnage basis. Electroplated tin is used as a protective
coating for food cans. The electroplated tin provides a physical barrier and
galvanic protection (in most food products) if the coating is compromised.
However, the tin coating cannot provide galvanic protection in the presence
of dissolved oxygen so that food should not remain in tin plated cans after
opening (3).
The ability to modify metal surfaces using directed energy beams is a
new approach being used to improve the corrosion resistance of metals
(31,44). Techniques such as ion implantation (qv), ion beam mixing, and ion
beam assisted deposition (IBAD), as well as laser-surface alloying and proces-
sing have been shown to improve the corrosion behavior of metals (see LASERS).
These techniques are versatile and have been used to modify the cathodic
or anodic reactions, improve the nature of passive films, and produce barrier
coatings (31). Ion implantation has been used to improve the corrosion
resistance of titanium in hot acids and aluminum and steels in aqueous
chloride environments (44). However, due to capital costs and the sizes of
the sample that can be treated, these techniques have not yet gained wide
spread use.
BIBLIOGRAPHY
‘‘Corrosion Inhibitors’’, in ECT 2nd ed., Vol. 6, pp. 317–346, by C. C. Nathan; ‘‘Corrosion
and Corrosion Inhibitors’’, in ECT 3rd ed., Vol. 7, pp. 113–142, by R. T. Foley and B. F.
Brown, The American University; in ECT 4th ed., Vol. 7, pp. 548–572, by Patrick &
Moran, United States Naval Academy, and Faul M. Natishan, Naval Research Labora-
tory; ‘‘Corrosion and Corrosion Control’’ in ECT (online), posting date: December 4,
2000, by Patrick J. Moran, United States Naval Academy and Paul M. Natishan, Naval
Research Laboratory.
Vol. 1 CORROSION AND CORROSION CONTROL 919
CITED PUBLICATIONS
PAUL NATISHAN
Naval Research Laboratory
PATRICK MORAN
U.S. Naval Academy
BIOREMEDIATION
Bioremediation is the process of judiciously exploiting biological processes
to minimize an unwanted environmental impact; usually it is the removal of a
contaminant from the biosphere. Like most definitions in biology, that of biore-
mediation is the subject of some debate. A narrow definition might focus on the
conversion of contaminating organic molecules to carbon dioxide, water, and
inorganic ions, and the oxidation or reduction of contaminating inorganic ions.
A broader definition would include biological processes for ameliorating extremes
of pH, concentrating contaminants so that they can be more easily removed by
physical techniques, converting toxic species to less toxic or less bioavailable
forms that pose less of a threat to the environment, and restoring functional eco-
systems to contaminated or disturbed sites when the contaminants or distur-
bance cannot be removed.
The concept of ‘‘judiciously exploiting biological approaches’’ is also a sub-
ject of debate:
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0209151816180914.a01
Vol. 1 BIOREMEDIATION 921
Table 2. (Continued)
Technology Description
3. Organic Contaminants
S
Octane
Dibenzothiophene
Isooctane N
2,2,4-trimethylpentane H
Carbazole
HO
Toluene
A naphthenic acid
Phenanthrene
Tetralin
1,2,3,4-tetrahydronaphthalene
A putative asphaltene
30:30:30:10. Most crude oils contain hydrocarbons ranging in size from methane
to molecules with hundreds of carbons, although the lightest molecules are
usually absent in oils that have been partially biodegraded in their reservoir.
When crude oils reach the surface environment the lighter molecules evaporate,
and are either destroyed by atmospheric photooxidation or are washed out of the
atmosphere in rain, and are biodegraded. Some molecules, such as the smaller
aromatics (benzene, toluene, etc) have significant solubilities, and can be washed
out of floating slicks, whether these are at sea, or on terrestrial water tables. For-
tunately the majority of molecules in crude oils, and refined products made from
them, are biodegradable, at least under aerobic conditions.
Biodegradation. Methane and the volatile plant terpenes are fully biode-
gradable by aerobic organisms, and most refined petroleum products are essen-
tially completely biodegradable under aerobic conditions. Estimates for crude oil
biodegradability range up to 90% (8), and the least biodegradable material, prin-
cipally polar molecules and asphaltenes, lacks the ‘‘oily’’ feel and properties that
are associated with oil. These are essentially impossible to distinguish from more
928 BIOREMEDIATION Vol. 1
HO O –O
O
R R R R
Linear alkanes
O–
O
O O
O
–O
Cycloalkanes
recent organic material in soils and sediments, such as the humic and fulvic
acids, and appear to be biologically inert.
Numerous bacterial and fungal genera have species able to degrade hydro-
carbons aerobically and the pathways of degradation of representative aliphatic,
naphthenic and aromatic molecules have been well characterized in at least some
species (8). Other organisms, such as algae and plants, do not seem to play a very
important role in the biodegradation of hydrocarbons. It is a truism that the hall-
mark of an oil-degrading organism is its ability to insert oxygen atoms into the
hydrocarbon, and there are many ways in which this is achieved. Figures 2 and 3
show the most well-studied. Once a hydrocarbon possesses a carboxylate or alco-
hol functionality it is almost invariably a readily degradable compound. A simple
example at the human level is the difference between oleic acid, a high calorie
food, and octadecane, present in mineral oil, which is so inert that it serves as
an intestinal lubricant!
For many years it was assumed that oil biodegradation was an exclusively
aerobic process, since any degradation must involve oxidation. Indeed the very
existence of oil reservoirs indicates that anaerobic degradative processes in
such environments must be very slow. Nevertheless, in recent years it has
become clear that at least some hydrocarbons are oxidized by bacteria under
completely anaerobic conditions, where the oxygen is probably coming from
water. Limited hydrocarbon biodegradation has now been shown under
sulfate-, nitrate-, carbon dioxide- and ferric iron-reducing conditions (Table 3).
The phenomenon is still poorly understood, however, and at present the largest
molecules demonstrated to undergo biodegradation under these conditions are
hexadecane, heptadecene, and phenanthrene. The pathways of degradation are
only beginning to be addressed. Figure 4 shows the intermediates identified in
anaerobic toluene degradation in different organisms. It is noteworthy that
Vol. 1 BIOREMEDIATION 929
OH OH
OH OH OH
O
O⫺
O
Naphthalene
HO OH
OH
OH
HO
HO
Toluene
HO
Benzyl alcohol
OH
para-Cresol
CoA S
β-PhenylpropionylCoA
requirement is usually that 2% (w/w) of the fuel be oxygen, which requires that
5–15% (v/v) of the gasoline be an oxygenated additive (eg, methanol, ethanol,
methyl tert-butyl ether (MTBE), etc). Although methanol and ethanol are readily
degraded under aerobic conditions, the degradability of MTBE remains some-
thing of an open question. The compound was previously very rare in the envir-
onment, but now it is one of the major chemicals in commerce. At first it seemed
that the compound was completely resistant to biodegradation, but complete
mineralization has now been reported (9). Whether biodegradation can be opti-
mized for effective bioremediation remains to be seen.
Bioremediation. Crude oil and refined products are readily biodegradable
under aerobic conditions, but they are only incomplete foods since they lack any
significant nitrogen, phosphorus, and essential trace elements. Bioremediation
strategies for removing large quantities of hydrocarbon must therefore include
the addition of fertilizers to provide these elements in a bioavailable form.
Air. Hydrocarbon vapors in air are readily treated with biofilters. These
are typically rather large devices with a very large surface area provided by
bulky material such as a bark or straw compost. The contaminated air, perhaps
from a soil vapor-extraction treatment, or from a factory using hydrocarbon sol-
vents, is blown through the filter, and organisms, usually indigenous to the filter
material or provided by a soil or commercial inoculum, grow and consume the
hydrocarbons. Adequate moisture must be maintained for effective operation.
Alternatively, trickling biofilters with recycled water are also in use. Both bac-
teria and fungi readily colonize such filters, and they can be very effective.
Nevertheless, biofilters are usually equipped with a small granulated activated
carbon ‘‘backup’’ filter to handle any sudden pulse loads that might overwhelm
the biological capacity of the filter. Biofilters compete with granulated activated
carbon filters, and are often cheaper because they minimize the cost of the granu-
lated activated carbon, and the energy required to destroy the contaminant and
the granulated activated carbon when the latter is saturated (10). Potential
Vol. 1 BIOREMEDIATION 931
problems include plugging and uneven air or water flow, but successful designs
work for many years with minimal maintenance except the occasional addition of
nutrients and stirring of the bed.
Sea. Crude oil spills at sea are perhaps the most widely covered environ-
mental incidents in the national and international media. Despite their notori-
ety, catastrophic tanker spills and well blow-outs are fortunately rather rare,
and their total input into the world’s oceans is approximately equivalent to
that from natural seeps; significantly more oil reaches the world’s oceans
from municipal sewers (11). Physical collection of the spilled oil is the preferred
remediation option, but if skimming is unable to collect the oil, biodegradation
and perhaps combustion or photooxidation are the only routes for elimination of
the spill. One approach to stimulating biodegradation is to disperse the oil with
chemical dispersants. Early dispersants had undesirable toxicity, but modern
dispersants and application protocols can stimulate biodegradation by increas-
ing the surface area of the oil available for microbial attachment, and perhaps
providing nutrients to stimulate microbial growth (12). Patents have been
issued for dispersant formulations that specifically include nitrogen and
phosphorus nutrients (13), but the products are not currently commercially
available.
Bioremediation by the addition of oil-degrading microbes is often promoted
as a treatment option for floating spills, but this approach has not yet met with
any documented success (13).
Shorelines. The successful bioremediation of shorelines affected by the
spill from the Exxon Valdez in Prince William Sound, Alaska, was perhaps the
largest bioremediation project to date (14,15). More than 73 miles of shoreline
were treated in 1989 and similar amounts of fertilizer were used in 1990. Oil
had typically penetrated into the surface gravel on these shorelines, occasionally
getting as deep as 30 cm into the sediment. Since the gravel was typically very
permeable, oxygen availability was unlikely to be the limiting factor for biodegra-
dation, and indeed this was subsequently shown to be correct. Bioremediation
thus focused on the addition of nitrogen and phosphorus fertilizers to partially
remove the nutrient-limitation on oil degradation. Of course the addition of fer-
tilizers was complicated by the fact that oiled shorelines were washed by tides
twice a day. These tides would have rapidly removed any soluble fertilizer, so
a strategy was sought that would provide nutrients for a significant length of
time. Various approaches to applying fertilizers were tried, including both stan-
dard and slow release nutrients, oleophilic nutrients and solutions of liquid fer-
tilizers. Two fertilizers were used in the full-scale applications; one, an oleophilic
product known as Inipol EAP22 (trademark of CECA, Paris, France), was a
microemulsion of a concentrated solution of urea in an oil phase of oleic acid
and trilaurethphosphate, with butoxyethanol as a cosolvent. This product was
designed to adhere to oil, and to release its nutrients to bacteria growing at
the oil-water interface. The other fertilizer was a slow-release formulation of
inorganic nutrients, primarily ammonium nitrate and ammonium phosphate,
in a polymerized vegetable oil skin. This product, known as Customblen (trade-
mark of Grace-Sierra, Milpitas, California), released nutrients with every tide,
and these were distributed throughout the oiled zone as the tide fell. Fertilizer
application rates were carefully monitored so that the nutrients would cause no
932 BIOREMEDIATION Vol. 1
harm, and the rate of oil biodegradation was stimulated between two- and five-
fold (14,15).
A wide range of fertilizers, including agricultural and horticultural fertili-
zers, and bone and fish meals have been tried at the pilot scale, usually with
at least modest success (13). Some current work is aimed at addressing whether
providing a readily degradable substrate with the fertilizer nutrients helps
immobilize the nutrients in biomass at the oiled site. Of course nutrient-
supplementation is only likely to markedly stimulate the rate of biodegradation
where nutrient levels are naturally low. It is unlikely that fertilizers will have a
dramatic effect in situations where agricultural or municipal run-off maintains
elevated levels of nutrients, such as happens in some estuaries and bays. Here
aeration is likely to be most effective.
Bioremediation by the addition of oil-degrading microbes has been pro-
moted as a treatment option for oiled shorelines, but this approach has not yet
met with any documented success (13).
Areas where there are currently few remediation options include oiled
marshes, mangroves, and coral reefs. These environments are generally easily
damaged by human intrusion and physical cleaning options may not provide
any net environmental benefit. Bioremediation may provide some attractive
options, and some success has been claimed, on a small scale, with fertilizer
applications (13). Marshes and mangroves offer the additional complication
that they are typically anoxic. Perhaps the anaerobic degradation of oil could
be stimulated by inoculation with anaerobic hydrocarbon degrading microbes,
or perhaps gentle aeration or the addition of slow release oxygen compounds,
such as some inorganic peroxides, might stimulate aerobic degradation without
significantly changing the redox balance of these environments. This is an area
where research is very much in its infancy and there are no well-documented
success stories to date.
Bioremediation also offers options for dealing with oiled material, such as
seaweed, that gets stranded on shorelines; composting has been shown to be
effective.
Groundwater. Spills of refined petroleum product on land, and leaking
underground storage tanks, sometimes contaminate groundwater. Bioremedia-
tion is becoming an increasingly popular treatment for such situations.
Hydrocarbons typically have a specific gravity of less than 1, and refined
products usually float on the water table if they penetrate soil that deeply. In
the parlance of the remediation industry, such floating spills are often called
NAPLs (nonaqueous phase liquids). Indeed they are sometimes known as
LNAPLs for light nonaqueous phase liquids, to distinguish them from more
dense materials, such as halogenated compounds, which are more likely to
sink in groundwater. Stand-alone bioremediation is an option for these situa-
tions, but ‘‘pump and treat’’ is the more usual treatment. Contaminated water
is brought to the surface, free product is removed by flotation, and the cleaned
water re-injected into the aquifer or discarded. Adding a bioremediation compo-
nent to the treatment, typically by adding oxygen and low levels of nutrients, is
an appealing and cost-effective way of stimulating the degradation of the resi-
dual hydrocarbon not extracted by the pumping. This approach is becoming
widely used.
Vol. 1 BIOREMEDIATION 933
Hydrocarbons are not very soluble in water, but the most soluble compo-
nents leach out of a spill if there is continual flushing. Typically only small aro-
matic molecules, the infamous BTEX (benzene, toluene, ethylbenzene, and
xylenes, Fig. 5), are soluble enough to contaminate groundwater. Although
with the advent of oxygenated gasolines, it is expected that these oxygenates
(ethanol, methanol, MTBE (methyl-tert-butyl ether) etc) will also be found in
groundwater. In the past, remediation of such situations has usually used
pump-and-treat methodologies. These methods are slow and may leave reser-
voirs of contaminants in pockets that are poorly connected to the main water
body. Of course the contaminant is biodegradable, and some biodegradation is
probably already occurring when the contamination is discovered. The cheapest
approach to remediation is, thus, to allow this intrinsic process to continue.
Evidence that it is indeed occurring can be found in the selective disappearance
of the most biodegradable compounds in the contaminant mixture, and the con-
comitant disappearance of electron acceptors from the groundwater. Thus oxy-
gen is depleted as the preferential terminal electron acceptor for metabolism,
followed by nitrate, ferric iron, sulfate, and finally CO2 for methanogenesis (16).
Intrinsic bioremediation is becoming an acceptable option in locations
where the contaminated groundwater poses little threat to environmental
health. Nevertheless, although intrinsic bioremediation is appealingly simple,
it may not be the lowest cost option if there are extensive monitoring and docu-
mentation costs involved for several years. In such cases it may well be more cost
effective to optimize conditions for biodegradation.
One approach is to optimize the levels of electron acceptors. Oxygen can be
pumped in as the pure gas, or as air, although this is relatively energy intensive
since oxygen is so poorly soluble. Hydrogen peroxide has been used in some situa-
tions, but there have been problems with biomass plugging near the injection
wells. Slow release formulations of inorganic peroxides, such as magnesium per-
oxide, have recently been used with success (17). Nitrate may be added, although
there are sometimes regulatory limitations on the amount of this material that
may be added to groundwater (18). Ferric iron availability may be manipulated
by adding ligands (19).
934 BIOREMEDIATION Vol. 1
deleterious. Not only is salt toxic to most plants, and can inhibit many soil
bacteria, but it also can have a major effect on the soil structure by altering
the physical properties of clays. Successful bioremediation strategies must there-
fore include remediating the brine. In wet regions the salt is eventually diluted
by rainfall, but in arid regions, and to speed the process in wetter regions,
gypsum is often added to restore soil porosity.
Many hydrocarbons bind quite tightly to soil components, and are thereby
less available to microbial degradation. The kinetics of binding seem to be com-
plex, and the process of ‘‘aging’’ is only poorly understood. Nevertheless, it seems
clear that hydrocarbons that have been in contact with soil for a long time are not
as available for biodegradation as fresh spills. Several groups of researchers have
suggested the addition of surfactants to overcome this limitation, but this
approach is not yet widely used. A significant potential concern is that the sur-
factant will be degraded in preference to the contaminant of concern.
Intrinsic biodegradation occurs, but it usually only removes the lightest
refined products, such as gasoline, diesel and jet fuel. Active intervention is
typically required. Usually the least expensive approach is in situ remediation,
typically with the addition of nutrients, and the attempted optimization of
moisture and oxygen by tilling. Various approaches to applying fertilizers have
been tried, including both standard and slow-release nutrients, oleophilic nutri-
ents and solutions of liquid fertilizers. Oxygen is a likely limiting nutrient in
many cases, and soil tilling is widely practiced. This in situ bioremediation of
hydrocarbon-contaminated soils is akin to the old practice of ‘‘land-farming’’,
wherein sludges and other refinery wastes were deliberately spread onto soil
and tilled and fertilized to stimulate biodegradation. Although this practice is
now discontinued in the United States, it was quite widely used.
Deeper contamination may be remedied with bioventing, where air is
injected through some wells, and extracted through others to both strip volatiles
and provide oxygen to indigenous organisms. Fertilizer nutrients may also be
added. This is usually only a viable option with lighter refined products.
A recent suggestion has been to use plants to stimulate the microbial degra-
dation of the hydrocarbon (hydrocarbon phytoremediation). This has yet to
receive clear experimental verification, but the plants are proposed to help deli-
ver air to the soil microbes, and to stimulate microbial growth in the rhizosphere
by the release of nutrients from the roots. The esthetic appeal of an active phy-
toremediation project can be very great.
When soil contamination extends to some depth it may be preferable to
excavate the contaminated soil and put it into ‘‘biopiles’’ where oxygen, nutrient
and moisture levels are more easily controlled. Biopiles can also be kept warm
during winter months, increasing the amount of time available for biodegrada-
tion in colder climates. Since the soil is well mixed during the construction of
the pile, there is an opportunity to add selected microbial and fungal strains in
an additional attempt to maximize biodegradation.
Composting by the addition of readily degradable bulking agents is also a
useful option for relatively small volumes of excavated contaminated soil.
Since efficient composting invariably involves self-heating as biodegradation pro-
ceeds, this also offers an option for extending the bioremediation season into the
winter months in cold climates. A potential drawback of composting is that it
936 BIOREMEDIATION Vol. 1
usually increases the volume of contaminated material, but if fully successful the
finished compost can be returned to the site as a positive contribution to soil
quality.
Slurry bioreactors offer the most aggressive approach to maximizing con-
tact between the contaminated soil and the degrading organisms. Both lagoons
and reactor vessels have been used, but the former are often not optimally
designed for all the soil to be partially suspended by the mixing impellers. Con-
tained reactor designs include mixing tank, airlift, and fluidized-bed aeration. A
major advantage of contained slurry bioreactors is the potential ability to opti-
mize nutrients, aeration and degradative inocula as fresh soil is added, and
the control of waste materials, including gases continually. Slurry bioreactors
are usually the most expensive bioremediation option because of the large
power requirements, but under some conditions this cost is offset by the rapid
biodegradation that can occur.
In all these cases it is important to bear in mind that although the majority
of hydrocarbons are readily biodegraded, some, such as the steranes and
hopanes, are very resistant to microbial attack. Estimates of oil biodegradation
range from 60–95% for different crude oils, so fresh spills of crude oils are readily
treated by bioremediation (21). Refined products, such as gasoline, diesel, jet
fuels, and heating oils are usually more biodegradable than typical whole oils,
but the various heavy fractions of crude oils, such as the asphalts, are far less
biodegradable, and are not such attractive targets for bioremediation. Some
crude oils have already been extensively biodegraded in their reservoirs, and
these are also poor targets for bioremediation. An example is Orimulsion, a
heavy oil in water emulsion (70% bitumen) stabilized by low levels of surfactants,
used as a fuel for electricity generation. Similarly, old spills may have already
undergone significant biodegradation and the residue may be relatively biologi-
cally inert. It is thus important to run laboratory studies to ensure that the con-
taminant is sufficiently biodegradable that clean-up targets can be met.
3.2. Halogenated Organic Solvents. Constituents. Halogenated
organic solvents are widely used in metal processing, electronics, dry cleaning
and paint, paper and textile manufacturing, and some representative examples
are shown in Figure 6. These solvents have been used for more than fifty years,
and unfortunately they are fairly widespread contaminants. Unlike the hydro-
carbons, which usually float on water, the halogenated solvents typically have
specific gravities greater than 1, and they generally sink to the bottom of any
groundwater, and float on the bedrock. For this reason they are sometimes
known as DNAPLs for dense nonaqueous phase liquids.
Biodegradation. Halogenated solvents are degraded under aerobic and
anaerobic conditions. The anaerobic process is typically a reductive dechlorina-
tion that progressively removes one halide at a time (Fig. 7). For example, under
methanogenic conditions, carbon tetrachloride is sequentially dechlorinated to
chloroform, dichloromethane, methyl chloride and methane, while trichloroethy-
lene is sequentially reduced to ethylene. Some of these compounds are also deha-
logenated under sulfate-reducing conditions, and under denitrifying conditions
there are reports that the final product can be CO2 (22). Chloromethane and
dichloromethane have been shown to be the sole carbon source for several anae-
robic organisms (23), and it seems there is much to be learned about the
Vol. 1 BIOREMEDIATION 937
Cl
Cl Cl
Cl Cl
Cl Cl Cl
Cl
Cl Cl
Cl
Cl Cl
1,1,1-Trichloroethane Trichloroethylene
Cl
Vinyl chloride
Cl Cl Cl Cl Cl Cl Cl
CH4
Cl Cl Cl
Carbon tetrachloride
Cl Cl Cl Cl Cl Cl Cl
Cl Cl Cl
Tetrachloroethylene
Cl Cl Cl O Cl
Cl Cl
Monooxygenase
Cl Cl OH OH
Cl Cl
Cl Cl Cl
Dioxygenase
atom into their substrate, and reduce the other to water. Dioxygenases insert
both atoms into their substrate. The effect of these two types of enzyme is
illustrated in Figure 8. In either case, biodegradation proceeds to complete
mineralization (25).
The biodegradation of trichloroethylene is the most studied since this
is probably the most widespread halogenated solvent contaminant. Several
substrates drive trichlorethylene co-oxidation, including methane, propane, pro-
pylene, toluene, isopropylbenzene, and ammonia (25). The enzymes that metabo-
lize these substrates have subtly different selectivities with regard to the
halogenated solvents, and to date none are capable of co-oxidizing carbon tetra-
chloride or tetrachloroethylene. Complete mineralization of these compounds
can, however, be achieved by sequential anaerobic and aerobic process.
Bioremediation. Air. Biofilters are an effective way of dealing with air
from industrial processes that use halogenated solvents such chloromethane,
dichloromethane, chloroethane, 1,2-dichloroethane and vinyl chloride, that sup-
port aerobic growth (26). Both compost-based dry systems and trickling filter wet
systems are in use. Similar filters could be incorporated into pump-and-treat
operations.
Groundwater and Soil. Halogenated solvents have contaminated soils
and groundwater throughout the industrialized world, and remediation has a
high priority. Since the solvents are so dense, they are typically found on the bed-
rock underlying aquifers. Pumping out the liquid phase is an obvious first step if
the contaminant is likely to be mobile, but in situ bioremediation is a promising
option. Thus, the U.S. Department of Energy is investigating the use of anaero-
bic in situ degradation of carbon tetrachloride in an aquifer some 76 m below
the Hanford, Washington site, with nitrate as electron acceptor, and acetate as
electron donor (22).
Trichloroethylene is the most frequent target of remediation, and as dis-
cussed above, this is only metabolized cometabolically. Remediation operations
thus incorporate the addition of the cometabolized substrate. Methane was
used successfully at the U.S. Department of Energy site at Savannah River,
near Aiken, South Carolina (27) which had both an air-stripping and a biological
component. Horizontal wells were used to pump methane and air below the
contaminant, while an upper horizontal well in the vadose zone was used to
withdraw these gases through the contaminated zone. Optimum biodegradation
performance seemed to come from alternate injection of air and methane in air,
Vol. 1 BIOREMEDIATION 939
and the inclusion of nitrous oxide and triethylphosphate, both gases, to give a
C:N:P ratio of 100:10:1.
Plants may have a role to play in enhancing microbial biodegradation
of halogenated solvents, for it has recently been shown that mineralization of
radiolabelled trichlorothylene is substantially greater in vegetated rather than
unvegetated soils (28), indicating that the rhizosphere provides a favorable
environment for microbial degradation of organic compounds.
Methane has also been used in aerobic bioreactors that are part of a pump-
and-treat operation, and toluene and phenol have also been used as co-substrates
at the pilot scale (29). Anaerobic reactors have also been developed for treating
trichloroethylene. For example, Wu and co-workers (30) have developed a suc-
cessful upflow anaerobic methanogenic bioreactor that converts trichloroethy-
lene and several other halogenated compounds to ethylene.
Groundwater contaminated with other halogenated solvents can also be
treated in aboveground reactors. Aerobic reactors are useful for those compounds
that can support growth. For example, a membrane reactor has been designed
for treating 1,2-dichloroethane (31), and bubble columns and packed bed reactors
have been developed for the aerobic degradation of 2-chloroethanol (32). As men-
tioned above, sequential anaerobic and aerobic reactors are capable of mineraliz-
ing tetrachloroethylene (33).
3.3. Halogenated Organic Compounds. Constituents. Complex
halogenated organic compounds have been widely used in commerce in the last
fifty years. A few representative examples are shown in Figure 9; pentachlorophe-
nol has been widely used as a wood preservative, and also for termite control.
OH
Cl Cl Cln Cln
Cl Cl
Cl
Polychlorinated biphenyl
Pentachlorophenol
O OH OH
O Cl Cl
O– Cl
Cl Cl
Cl Cl
Cl
Hexachlorophene
(2,4-Dichlorophenoxy)acetate
Cl
Cl Cl
Cl Cl
DDT
on a molecule, the slower its biodegradation, and this is borne out with the
related herbicide 2,4,5-T ((2,4,5-trichlorophenoxy)acetate). Nevertheless, bacter-
ial degradation has been seen under both aerobic and anaerobic conditions, the
latter involving reductive dechlorination via 2,4-D. Aerobic degradation removes
acetate from (2,4-dichlorophenoxy)acetate to yield 2,4-dichlorophenol, which is
subsequently hydroxylated to 3,5-dichlorocatechol, followed by ring cleavage
and complete mineralization (40). Genetic engineering has been used to con-
struct strains that are particularly adept at consuming these compounds, but
whether these will overcome the regulatory hurdles to allow their use outside
the laboratory remains to be seen.
DDT (1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane) is a remarkably resis-
tant molecule, which explains both its efficacy as an insecticide, and its accumu-
lation at the top of the food-chain. Nevertheless, there are indications that it can
be biodegraded, both anaerobically with initial dechlorination and aerobically
with initial ring hydroxylation (40). DDT and its partially degraded congeners
are very hydrophobic, and biodegradation seems to be stimulated by adding sur-
factants. White-rot fungi also degrade DDT under lignolytic conditions, although
there is little mineralization to CO2 (5).
The polychlorinated biphenyls are quite recalcitrant. Some lightly chlori-
nated biphenyls are readily mineralized under aerobic conditions (41), and
indeed the structure of an enzyme that catalyzes the key ring-cleavage oxygena-
tion has recently been determined by x-ray crystallography (42). More chlori-
nated congeners are resistant to aerobic degradation, but they are reductively
dechlorinated under anaerobic conditions. Complete degradation of the commer-
cial mixtures, thus, generally requires an anaerobic process followed by an aero-
bic one. A major issue for the oxidative process is that biphenyl seems to be
required for significant expression of the biodegradative system (41); the chlori-
nated compounds do not induce the enzymes that would degrade them. The bio-
chemistry of the biodegradative process is only beginning to be unraveled, but
already there are suggestions that there is considerable diversity in the enzyme
systems able to degrade these compounds. There is a lot of effort aimed at engi-
neering organisms to degrade polychlorinated biphenyls, and in finding ways to
make these very hydrophobic compounds more bioavailable by the use of chemi-
cal oxidants, such as Fenton’s reagent (43), or surfactants.
Bioremediation. Soil. Pentachlorophenol has been the target of biore-
mediation at a number of wood-treatment facilities, and good success has
been achieved in several applications. It is rarely the sole contaminant, and
is often present with polynuclear hydrocarbons from coal creosote. In situ
degradation has been stimulated by bioventing, where air is injected through
some wells, and extracted through others to both strip volatiles and provide
oxygen to indigenous organisms (44). Just as with the halogenated solvents,
it seems that plants stimulate microbial degradation of pentachlorophenol in
the rhizosphere (45).
The kinetics of such in situ degradation are rather slow, however, and more
active bioremediation is usually attempted. For example, contaminated soil at
the Champion Superfund site in Libby, Montana, was placed into 1-acre land
treatment units in 6-in. layers, and irrigated, tilled, and fertilized. Under
these conditions, the half-lives of pentachlorophenol, pyrene, and several other
942 BIOREMEDIATION Vol. 1
compounds are mineralized under aerobic conditions, but the more chlorinated
species seem completely resistant to aerobic degradation, even by white rot
fungi (54). On the other hand, there is extensive dechlorination of highly
chlorinated forms under anaerobic conditions, particularly methanogenic
conditions. Bioremediation thus requires anaerobic and aerobic regimes.
Intrinsic biodegradation of polychlorinated biphenyls can be recognized by
the changing ‘‘fingerprint’’ of the individual isomers as biodegradation pro-
ceeds (41, 55). The anaerobic dechlorination of the most recalcitrant congeners
can apparently be primed by adding a readily dehalogenated congener, such as
2,5,30 ,40 -tetrachlorobiphenyl (56), but whether this is a realistic approach for in
situ bioremediation remains to be seen. Harkness and co-workers (57) have
successfully stimulated aerobic biodegradation in large caissons in the Hudson
River by adding inorganic nutrients, biphenyl, and hydrogen peroxide, but
found that repeated addition of a polychlorinated-biphenyl degrading bacter-
ium (Alcaligenes eutrophus H850) had no beneficial effect. Essentially no bio-
degradation occurred in the stirred control caissons, but losses on the order of
40% were seen in the caissons that received nutrients and peroxide, regardless
of whether the stirring was aggressive or rather gentle. Whether this approach
can be scaled-up for large-scale use, with a net environmental benefit, remains
to be seen.
3.4. Nonchlorinated Pesticides and Herbicides. Constituents.
Pesticides A vast number of compounds are used as herbicides, pesticides, fungi-
cides, etc, and a few are shown in Figure 10. In order to be effective they must
have their effect before they are degraded in the environment, but on the other
hand they must not be so resistant to degradation that they accumulate where
they are used, or accumulate in food chains. It is unusual for these compounds
to become contaminants where they are applied correctly, but manufacturing
facilities, storage depots and rural airfields where crop-dusters are based have
had spills that can lead to long lasting contamination. Bioremediation is a
promising technology to remediate such sites. There are also some locations
were groundwater has become contaminated by these chemicals, and again
bioremediation may be a cost-effective remediation strategy.
Biodegradation. The vast majority of pesticides, herbicides, fungicides,
and insecticides in use today are biodegradable, although the intrinsic biode-
gradability of individual compounds is one of the variables used in deciding
which compound to use for which task. Some herbicides are acutely toxic to
plants that absorb them, but are so readily degraded in soil that seeds can be
planted at the same time as the herbicide application. Other herbicides are
known to be effective at preventing plant growth for many months, and are
used in situations where this is the desired goal. Very few degradation pathways
of these compounds have been worked out in detail, but some generalizations can
be made.
Compounds with organophosphate moieties, such as Diazinon, Methyl
Parathion, Coumaphos and Glyphosate are usually hydrolyzed at the phos-
phorus atom (40,58). Indeed several Flavobacterium isolates are able to grow
using parathion and diazinon as sole sources of carbon.
Triazines pose rather more of a problem, probably because the carbons are
in an effectively oxidized state so that no metabolic energy is obtained by
944 BIOREMEDIATION Vol. 1
N
NH N Cl
N H
N
N N N
O
O O N
NH P NH2
S
Atrazine Diazinon Amitrole
NO2 NO2 O O2
NH
O2N HO P OH
O
O O OH O
P
S
Methyl Parathion Dinoseb Glyphosate
O
S N O N
H
Aldicarb
their metabolism. Very few pure cultures of microorganisms are able to degrade
triazines such as Atrazine, although some Pseudomonads are able to use the
compound as sole source of nitrogen in the presence of citrate or other simple
carbon substrates. The initial reactions seem to be the removal of the ethyl or
isopropyl substituents on the ring (41), followed by complete mineralization of
the triazine ring.
Nitroaromatic compounds, such as Dinoseb, are degraded under aerobic
and anaerobic conditions (59). The nitro group may be cleaved from the molecule
as nitrite, or reduced to an amino group under either aerobic or anaerobic condi-
tions. Alternatively, the ring may be the subject of reductive attack. Thus, while
these molecules are sometimes quite long-lived in the environment, they can be
completely mineralized under appropriate conditions (59). Recent work has iso-
lated a Clostridium bifermentans able to anaerobically degrade dinoseb cometa-
bolically in the presence of a fermentable substrate. The dinoseb was degraded to
below detectable levels, although only a small fraction was actually mineralized
to CO2 (60).
Carbamates such as Aldicarb undergo degradation under both aerobic and
anaerobic conditions. Indeed the oxidation of the sulfur moiety to the sulfoxide
and sulfone is part of the activation of the compound to its most potent form. Sub-
sequent aerobic metabolism can completely mineralize the compound, although
this process is usually relatively slow so that it is an effective insecticide,
Vol. 1 BIOREMEDIATION 945
NO2
O2N NO2
N
N N
O2N NO2
NO2
TNT RDX
O2N
N
NO2
N
N NH2
N
O2N
N
NO2
HMX N,N-Dimethylhydrazine
some sites by anaerobic degradation. Research into whether this can be stimu-
lated with a net environmental benefit is in its very earliest stages, and no
clear evidence for success has been presented.
A commercial technology (69), the SABRE process, treats contaminated
water and soil in a two-stage process by adding a readily degradable carbon
and an inoculum of anaerobic bacteria able to degrade the contaminant. An
initial aerobic fermentation removes oxygen so that the subsequent reduction
of the contaminant is not accompanied by oxidative polymerization.
Soil. Composting of soils contaminated by high explosives is being carried
out at the Umatilla Army Depot near Hermiston, Oregon (70). Soil from muni-
tions washout lagoons is being treated indoors in compost rows of 2,000m3,
and the estimated cost is less than one-third the estimated cost of incineration.
If this is successful, there are 30 similar sites on the National Priority List that
could be treated in a similar way.
3.6. Other Organic Compounds. The majority of organic compounds
in commerce are biodegradable, so bioremediation is a potential option for
cleaning up after industrial and transportation accidents. For example, Ref.
71 reports the successful bioremediation of 23,000 liters of vinyl acetate spilled
from a railroad tank car in Albany, New York, at a cost approximately half that
of excavation and disposal of the 1,100 m3 of contaminated soil. They also
report that in situ biological treatment has been used to remediate spills of
other organics, including acrylonitrile, styrene, 2-butoxyethanol, and ethacry-
late, at other sites where there had been railroad accidents. Bioremediation
is thus already an important tool in remediating accidental spills of organic
compounds.
Bioremediation is also an option when spills of such compounds contami-
nate groundwater. For example, bioremediation seems a feasible treatment for
aquifers contaminated with alkylpyridines (72) and phenol (73).
4. Inorganic Contaminants
and bacterial growth, so one option is to use these organisms to take up and use
the contaminants.
Bioremediation. Surface Water. One example of exploiting biology to
handle excess nitrogen in a surface water is the case of the Venice Lagoon in
Italy. About a million tons of sea lettuce (Ulva) grows in the lagoon annually
because of the high levels of nitrogenous nutrients in this relatively landlocked
bay. This material is harvested, composted and sold as a low-cost remediation of
this problem (74). In areas where water quality is sufficiently high, growing even
more valuable crops, such as nori (Porphyra) may be an attractive bioremedia-
tion option.
A more constrained opportunity for nitrate bioremediation arose at the US-
DoE Weldon Spring Site near St. Louis, Missouri. This site had been a uranium
and thorium processing facility, and treatment of the metal had involved nitric
acid. The wastestream, known as raffinate, was discharged to surface inpound-
ments and neutralized with lime to precipitate the metals. Two pits had nitrate
levels that required treatment before discharge, but heavy rains in 1993 threa-
tened to cause the pits to overflow. Bioremediation by the addition of calcium
acetate as a carbon source successfully treated more than 19 million liters of
water at a reasonable cost (75).
Groundwater. One approach to minimizing the environmental impact of
excess nitrogen in groundwater migrating into rivers and aquifers is to intercept
the water with rapidly growing trees, such as poplars, that will use the contami-
nant as a fertilizer.
An alternative approach is to add a readily degradable substrate to the con-
taminated aquifer, in the absence of oxygen, to stimulate bacterial denitrifica-
tion. Soluble substrates such as ethanol have been suggested, but a recent
alternative suggestion is to provide vegetable oil, either down the well itself, or
on a gravel trickling filter. Since the oil is insoluble in water, it becomes immo-
bilized on the aquifer material, allowing denitrification as the water flows
through (76).
4.3. Metals and Metalloids. A wide range of metallic and nonmetallic
elements are present as contaminants at industrial and agricultural sites
throughout the world, both in ground- and surface water, and in soils. They
pose a quite different problem from that of organic contaminants, since they can-
not be degraded so that they disappear. Some metal and metalloid elements have
radically different bioavailabilities and toxicities depending on their redox state,
so one option is stabilization by converting them to their least toxic form. This
can be a very effective way of minimizing the environmental impact of a contami-
nant, but if the contaminant is not removed from the environment, there is
always the possibility that natural processes, biological or abiological, may
reverse the process. Removing the contaminants from water phases is relatively
straightforward, and the wastewater treatment industry practices this on an
enormous scale. Pump-and-treat systems that mimic waste water treatments
are already being used for several contaminants, and less complex systems invol-
ving biological mats are a promising solution for less demanding situations.
Microbial leaching of metals from ores is a promising adjunct to more
aggressive metal recovery technologies (77), but is generally achieved by oxida-
tive processes that generate very acidic waters. It seems unlikely that similar
Vol. 1 BIOREMEDIATION 949
that this part of the constructed wetland be kept anaerobic, to prevent oxidation
and remobilization of the precipitated metals. For this reason, this part of the
wetland is typically kept flooded and free of aquatic plants that might introduce
oxygen through their roots (100). Finally, an aerobic algal mat can act as a pol-
ishing step to complete the removal of contaminants, particularly manganese
(101,102). Mine drainage that is not acidic may not need such complicated sys-
tems and individual parts of the treatment train described above may suffice in
some situations.
Soil. The first reported field trial of the use of hyperaccumulating plants
to remove metals from a soil contaminated by sludge applications has been
reported (103). The results were positive, but the rates of metal uptake suggest
a time scale of decades for complete cleanup. Trials with higher biomass plants,
such as B. juncea, are underway at several chromium and lead contaminated
sites (88), but data are not yet available.
The bacterial reduction of Cr(VI) to Cr(III) discussed above is also being
used to reduce the hazards of chromium in soils and water (104).
5. Conclusions
organisms. For example, modern molecular biology may offer opportunities for
moving effective degradation pathways into organisms native to a contaminated
site, improving biodegradation pathways by broadening the substrate range of
degradative enzymes, or removing regulatory constraints to maximize degrada-
tive activities. There are, however, many technical and regulatory hurdles to be
surmounted before this potential can be tested.
Another area where there is controversy is in the role of surfactants. Many
bacteria, particularly those that degrade hydrocarbons, produce surfactants as
they grow. Some release them into the medium, others incorporate them into
their cell exterior, and there have been elegant experiments to show that inhibit-
ing the production of these compounds inhibits the ability of the bacteria to
degrade oil. There have thus been many suggestions to add surfactants, either
bacterial or synthetic, to stimulate biodegradation. This seems to be beneficial
in the case of some oil dispersants at sea (12), but there have not yet been any
clear demonstrations of efficacy on a large scale in soils. The role of surfactants in
bioremediation is an area that requires further study.
Bioremediation has many advantages over other technologies, both in cost
and in effectively destroying or extracting the pollutant. An important issue is
thus when to consider it, and a series of questions may lead to the appropriate
answer (see Table 4).
H
H
BIBLIOGRAPHY
CITED PUBLICATIONS
GENERAL REFERENCES
ROGER C.PRINCE
Exxon Research and Engineering Company
PART II: REFINED
PRODUCTS AND FUELS
LIQUEFIED PETROLEUM
GAS
Liquefied petroleum gas (LPG) is a subcategory of a versatile class of petroleum
products known as natural gas liquids (NGLs) that are produced along with and
extracted from natural gas (see NATURAL GAS). LPG is also produced from the refin-
ing of crude oil (see PETROLEUM). Although LPG is commercially defined as propane
[74-98-6], butane [106-97-8], and butane–propane mixtures, commercial availabil-
ity is primarily limited to propane (see HYDROCARBONS). There are two grades of spe-
cification propane, propane HD-5 and special-duty propane. The primary difference
in the two grades is that the Propylene [115-07-1] content of propane HD-5 is
restricted to a maximum of 5 vol%. Propylene (qv) is found only in refinery-
produced propane. The principal uses of LPGs are as fuels and feedstocks (qv)
for the production of motor gasoline and a wide variety of chemicals (see GASOLINE
AND OTHER MOTOR FUELS).
Other natural gas liquids include natural gasoline [8006-61-9], which is com-
posed of the pentanes and heavier components of the natural gas stream, and ethane
[74-84-0]. Most recently ethane has become the principal product of natural gas proces-
sing plants.
1. Properties
1
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1209172120012512.a01
2 LIQUEFIED PETROLEUM GAS Vol. 2
wide limits. Physical properties of the principal components of LPG are sum-
marized in Table 1 (1).
Recompressor
Residue gas
to pipeline
Turbo
Compressor expander
Hot-gas Cold-gas
exchanger
Condensate
exchanger
separator
Dehydrated
gas inlet
Control Propane
valve refrigerant Demethanizer
Chiller
Side heater
Product
exchanger Reboiler
Fig. 1. One-expand cycle with external refrigeration for high ethane recovery in the
hydrocarbon liquid product.
bottoms product, the reboiler, and the side heater. A significant amount of low
level refrigeration from the demethanizer liquids and the cold residue gas stream
is recovered in the inlet gas stream.
The two portions of the feed stream recombine and flow into the high pres-
sure separator where the liquid is separated from the vapor and is fed into an
intermediate section of the demethanizer with liquid level control. The decrease
in pressure across the level-control valve causes some of the liquid to flash which
results in a decrease in the stream temperature. The pressure of the vapor
stream is decreased by the way of a turboexpander to recover power, thus to
achieve more cooling than would be possible by Joule-Thompson expansion.
The outlet of the turboexpander then is fed into the top of the demethanizer
where the separation of liquid and vapor occurs. The vapor is passed as cold resi-
due to the heat exchanger and the liquid is distributed to the demethanizer top
tray as reflux.
Essentially all of the methane [74-82-8] is removed in the demethanizer
overhead gas product. High recovery of ethane and heavier components as
demethanizer bottoms products is commonplace. The work that is generated
by expanding the gas in the turboexpander is utilized to compress the residue
gas from the demethanizer after it is warmed by heat exchange with the inlet
gas. Recompression and delivery to a natural gas pipeline is performed down-
stream of the plant. A propane recovery of 99% can be expected when ethane
recoveries are in excess of 65%.
4 LIQUEFIED PETROLEUM GAS Vol. 2
Recoveries of 90–95% ethane have been achieved using the expander pro-
cesses. The liquid product from the demethanizer may contain 50 liquid vol %
ethane and usually is delivered by a pipeline to a central fractionation facility
for separation into LPG products, chemical feedstocks, and gasoline-blending
stocks.
2.2. Absorption. Oil absorption is another process used for recovery of
LPG and natural gas liquids from natural gas. Recovery is enhanced by lowering
the absorption temperature to –458C and by keeping the molecular weight of the
absorption oil down to 100. Heat used to separate the product from the absorp-
tion oil contributes to the cost of recovery. Therefore, this process has become less
competitive as the cost of energy has increased. A simplified flow diagram of a
typical oil-absorption process is shown in Figure 2.
The natural gas feed exchanges heat with the residue gas from the absorber
overhead. Ethylene glycol [107-21-1] is injected as an antifreeze and the stream
is cooled further by refrigeration to –378C. The gas from the cooler enters a glycol
separator where the glycol that contains water is separated from the natural gas
Compressor
Cooler
Refrigerating
Oil absorber
Exchanger cooler
separator
Glycol
inlet gas
deethanizer
Rich-oil
Rich-oil
flash
Cooler
Hydraulic
Glycol turbine
injection Exchanger Exchanger
Exchanger
Glycol
regeneration
Butane splitter
deethanizer
Lean-oil
Isobutane
Debutanizer
n-Butane
Booster
Residue
natural gas Propane
Natural gasoline
liquids
Fig. 2. Refrigerated absorption process for the production of LPG and natural gasoline
liquids.
Vol. 2 LIQUEFIED PETROLEUM GAS 5
as a liquid phase. A dry glycol is recovered for recycling to the injection point by
distilling the water. The gas and any higher boiling hydrocarbons that are pre-
sent pass to the base of the absorber where the gas comes into contact with
absorption oil which enters at the top of the absorber at 37 C. Approximately
85% of the propane and essentially all of the higher boiling hydrocarbons are
absorbed in the oil. The overhead residue gas from the absorber (34 C) is heat
exchanged with the inlet gas; at 48C and 4.24 MPa (600 psig), the gas flows to
the booster where the pressure is increased to that of the natural gas line.
The rich oil from the absorber is expanded through a hydraulic turbine for
power recovery. The fluid from the turbine is flashed in the rich-oil flash tank to
2.1 MPa (300 psi) and –328C. The flash vapor is compressed until it equals the
inlet pressure before it is recycled to the inlet. The oil phase from the flash passes
through another heat exchanger and to the rich-oil deethanizer. The ethane-rich
overhead gas produced from the deethanizer is compressed and used for produ-
cing petrochemicals or is added to the residue-gas stream.
The bottoms, consisting of absorption oil, absorbed propane, and higher
boiling hydrocarbons, are fed to the lean-oil fractionator. The LPG and the nat-
ural gas liquids are removed as the overhead product from the absorption oil
which is removed as a kettle-bottom product.
The lean oil from the lean-oil fractionator passes through several heat
exchangers and then through a refrigerator where the temperature is lowered
to 37 C. Part of the lean oil is used as a reflux to the lower section of the
rich-oil deethanizer. Most of the lean oil is presaturated in the top section of
the deethanizer, is cooled again to 37 C, and is returned to the top of the absor-
ber, thus completing the oil cycle.
The overhead product from the lean-oil fractionator, consisting of propane
and heavier hydrocarbons, enters the depropanizer. The depropanizer overhead
product is treated to remove sulfur and water to provide specification propane.
The depropanizer bottoms, containing butane and higher boiling hydrocarbons,
enters the debutanizer. Natural gasoline is produced as a bottom product from
the debutanizer. The debutanizer overhead product is mixed butanes, which
are treated for removal of sulfur and water, then fed into the butane splitter.
Isobutane is produced as an overhead product from the splitter and n-butane
is produced as a bottoms product.
2.3. Compression. Compression is the simplest and the least effective
of the four recovery methods. It was the first process used for the recovery of
hydrocarbon liquids from natural gas but is used only in isolated cases. The
most significant application of the compression process is for gas-cycling plants
where the natural gas liquids are removed and the remaining gas is returned to
the production formation. Figure 3 is a schematic of a typical gas-cycle plant.
The pressure used in producing gas wells often ranges from 690– 10,300 kPa
(100–1500 psi). The temperature of the inlet gas is reduced by heat-exchange
cooling with the gas after the expansion. As a result of the cooling, a liquid
phase of natural gas liquids that contains some of the LPG components is formed.
The liquid is passed to a set of simple distillation columns in which the most vola-
tile components are removed overhead and the residue is natural gasoline. The
gas phase from the condensate flash tank is compressed and recycled to the gas
producing formation.
6 LIQUEFIED PETROLEUM GAS Vol. 2
Compressor
Gas
Expander
well Compressor
Butane Propane
Condensate
separator
Debutanizer
Debutanizer
Condensate
flash tank
Natural gasoline
Condensable liquids also are recovered from high pressure gas reservoirs by
retrograde condensation. In this process, the high pressure fluid from the reser-
voir produces a liquid phase on isothermal expansion. As the pressure decreases
isothermally the quantity of the liquid phase increases to a maximum and then
decreases to disappearance. In the production of natural gas liquids from these
high pressure wells, the well fluids are expanded to produce the optimum
amount of liquid. The liquid phase then is separated from the gas for further pro-
cessing. The gas phase is used as a raw material for one of the other recovery
processes, as fuel, or is recompressed and returned to the formation.
2.4. Adsorption. Adsorption processes have been used to recover hydro-
carbons that are heavier than ethane from natural gas. Although the adsorption
process has applications for the recovery of pentane and heavier hydrocarbons
from lean gas, the percentage recovery of LPG components in these plants
usually is low compared to the normal recovery of LPG in modern turboexpander
or oil-absorption plants.
A simplified flow diagram for the fast-cycle adsorption plant is shown in
Figure 4. Activated carbon, alumina gel, and silica gel are used as adsorbents.
Use of internal insulation in the adsorption towers affords less cycle time. The
complete process cycle consists of three phases: regeneration, cooling, and
adsorption. In Figure 4, the inlet gas is divided. The larger portion flows directly
to the bed in the adsorption cycle; the smaller portion flows first through the bed
in the cooling phase of the cycle, then through the regeneration cycle. The efflu-
ent gas from the cooling cycle is heated further to 260–3158C in a separate fired
Vol. 2 LIQUEFIED PETROLEUM GAS 7
Booster Compressor
Inlet Adsorption
gas towers
Low pressure
High pressure
Regeneration
Adsorption
separator
Cooling
flash
cycle
cycle
cycle
Exchanger
Regeneration
furnace Hydrocarbon liquids:
Cooler LPG and natural
gasoline
Dry gas
furnace before it is injected into the bed in the regeneration cycle. The effluent
from the regeneration cycle contains the condensable hydrocarbons that have
been stripped from the adsorbent and that are removed in the high pressure
separator after the gas that is leaving the regeneration cycle is cooled. The liquid
from the high pressure separator is flashed through to remove light hydrocar-
bons. The flash vapor is compressed and is mixed with the vapor from the high
pressure separator. The pressure on the compressed gas stream is boosted to the
inlet gas pressure and the gas is recycled to the feed to the adsorption cycle. The
liquid from the low pressure flash, which contains LPG and natural gas liquid
components, is processed further in a series of distillation towers to produce pro-
pane, butanes, and natural gasoline.
Less propane and butanes are produced compared to natural gas liquids by
the adsorption process than are obtained normally for the same gas by the oil-
absorption process. Because adsorption efficiency increases with a decrease in
temperature, the adsorption cycle should operate at the lowest temperature
that is economically feasible.
2.5. Purification. The LPG generally requires treatment for removal of
hydrogen sulfide [7783-06-4], H2S, organic sulfur compounds, and water in order
to meet specifications. Several methods are used.
Amine Treatment. The LPG is brought into contact with a 15– 20 wt%
solution of ethanolamine [141-43-5] in water, which removes H2S to specifica-
tion levels. The foul amine solution that contains the H2S is regenerated in
a stripper at low pressure using indirect stream stripping, and the stripped
amine solution is returned to the LPG contactor. The amine solution gene-
rally has little affect on organic sulfur compounds, although diethanolamine
8 LIQUEFIED PETROLEUM GAS Vol. 2
[111-42-2] (DEA) and diglycolamine [929-06-6] (DGA) have been used to obtain
acceptable carbonyl sulfide [463-58-1], COS, levels in propane.
Caustic Treatment. Amine treatment may be followed by a caustic treat-
ment step in which the LPG is brought into contact with 10–20 wt % caustic
solution to remove any residual H2S and to remove mercaptans. The mercaptans
may be stream-stripped from the caustic solution, after which the stripped caus-
tic is recycled to the LPG caustic contactor. Caustics that contain H2S must be
discarded. Various promoters may be added to the caustic to improve the effi-
ciency of the mercaptan removal.
Coalescing. Sand towers or cartridge-type coalescers may be used to
separate any undissolved water from the LPG. Removal of the undissolved
water meets the specification moisture limit for butanes. However, this step
does not produce specification propane.
Solid-Bed Dehydration. Silica gel, bauxite, activated alumina, or molecu-
lar sieves can be used for removing dissolved water to meet propane specifica-
tions. The solid-bed dehydrators are used in a cyclic adsorption process. After
an adsorption cycle has completed, the bed is heated with a purge gas or a vapor-
ized liquid-product stream for regeneration. If the latter is used, the liquid pro-
duct is condensed, separated from the free water, and returned to the process.
After the beds are regenerated, they are cooled and returned to the adsorption
cycle.
Molecular Sieve Treatment. Molecular sieve treaters can be designed to
remove H2S, organic sulfur compounds (including carbonyl sulfide), and water in
one step. Solid-bed units are utilized and regeneration occurs in the same man-
ner as simple, solid-bed dehydrators.
Solid-Bed Caustic Treatment. Solid-bed caustic units utilizing methanol
[67-56-1] injection into the LPG feed stream can be used for carbonyl sulfide
removal. The methanol–caustic solution must be drained periodically from the
beds and discarded. When the solid bed is exhausted, the spent caustic must
be discarded and replaced. The LPG from the treater has a low enough water
content to meet the propane specification.
Fractionation. Direct fractionation also can be used to remove dissolved
water from LPG. The water-rich overhead vapor from the dryer fractionator is
returned to the fractionator as reflux and the water phase is discarded. A dry
LPG product that meets either propane or butane water specifications is pro-
duced as a kettle product from the fractionator.
Historically, about two-thirds of the LPG produced in the United States came
from natural gas processing and one-third was produced from refinery opera-
tions (2). In 1991, this ratio was 61% from natural gas processing and 39%
from refinery operations. Total production of LPG in 1991 was 76:85 106 m3
ð294:19 106 bblÞ from natural gas processing and 30:08 106 m3 ð189:23
106 bblÞ produced from refinery operations.
Ethane production for 1991 was 30:74 106 m3 ð193:32 106 bblÞ from natu-
ral gas operations and 1:49 106 m3 ð9:34 106 bblÞ from refinery operations for
Vol. 2 LIQUEFIED PETROLEUM GAS 9
a total production of 32:21 106 m3 ð202:66 106 bblÞ. A summary of total nat-
ural gas liquids (NGL) including LPG supply and demand in the United States is
shown in Table 2.
The progress of LPG utilization has been closely related to progress in
transportation and storage of this fuel. Large volumes of LPG must be trans-
ported from the producing plants to centers of consumption, and transportation
costs are a principal factor in the cost of LPG to the consumer. Large volumes of
LPG usually are transported by high pressure pipelines (qv). As of this writing,
this use of pipelines is increasing rapidly. Large quantities of LPG are trans-
ported in railroad tank cars which have an average capacity of 113:5 m3 =car
(3 104 gal=car), although this use is decreasing. Tank-truck transports having
capacities of about 17.8 m3 (104 gal) also move large quantities of product from
producing plants to distribution centers and from the pipeline terminals to
points of distribution or consumption. Delivery trucks having capacities from
3.8–11.4 m3 (1000–3000 gal) generally are used for the final delivery from the
distribution bulk plant to the storage plants of the larger consumers. Smaller
quantities are shipped in metal cylinders having capacities from 9–45 kg LPG.
These cylinders are filled at the distribution bulk plant and are delivered by
truck to the consumer or to the cylinder dealer or cylinders may be filled for
the customer at small distribution stations.
Tankers and barges are also used for transporting LPG. Tankers and
barges are designed for both high pressure ambient temperature and for low
pressure refrigerated transportation. For larger volumes and long distances,
low pressure refrigerated tankers almost always are used (2).
10 LIQUEFIED PETROLEUM GAS Vol. 2
4. Economic Aspects
The production and consumption of LPG in the United States increased drama-
tically from its early beginnings in the 1930s until the international energy crises
of the 1970s when rising prices and regulatory restraints resulted in reduced
domestic production. However, total consumption, including imports, resumed
a modest growth characteristic after that time. In 1984, total LPG consumption
in the United States was 76:58 106 m ð481:71 106 bblÞ; by 1991, total con-
sumption was 84:98 106 m3 ð534:50 106 bblÞ. The principal growth segment
has been the increasing use of LPG for petrochemical feedstocks as can be
seen from Table 3.
6. Storage
Large volumes of LPG are stored to meet peak demand during cold seasons.
LPGs are both volatile and flammable and must be stored and handled in special
equipment. Standards for storing and handling LPG are published by the
National Fire Protection Association (5) and API (6).
Four main types of storage are used: high pressure storage above ground,
low pressure refrigerated storage above ground, frozen earth storage, and under-
ground cavern storage. The capacities of the storage unit vary from 500 mL pres-
sure cylinders to 1:9 106 m3 ð500 106 galÞ underground storage caverns.
Economic factors determine the proper storage for any given requirement.
Above ground pressure-storage tanks usually are designed for a 1720 kPa
(250 psi) working pressure for propane and 860 kPa (125 psi) for butane. Refri-
gerated, aboveground storage tanks usually are designed for a few kilopascals of
pressure. These tanks must be coupled with refrigeration systems to cool the pro-
duct that is to be stored to a temperature equal to the product’s boiling point at
the operating pressure of the tanks. Vapors generally are recondensed by refrig-
eration and returned to the tanks.
Table 3. Sales of LPG 1984–1991, 106 m3a,b
Residential and Motor fuel
Year Chemical commercial heating Engine fuel Industrial Utility gas blending Farm Export Other Total
1984 47.49 15.05 2.80 10.06 1.45 35.71 3.77 2.93 11.82 2,115.08
1985 51.57 16.97 2.78 7.66 0.91 37.09 4.38 3.72 8.90 2,118.98
1986 48.72 16.62 2.48 8.23 2.73 36.36 4.29 2.59 2.90 2,110.92
1987 54.45 18.39 2.39 8.35 0.54 37.69 4.08 2.33 2.96 2,118.17
1988 60.24 18.25 2.21 8.19 1.23 35.15 4.03 2.96 3.97 2,124.23
11
1989 55.52 20.41 2.20 9.90 1.14 33.26 4.44 2.37 6.74 2,123.99
1990 57.57 18.84 2.01 5.80 1.06 32.71 4.30 2.16 3.78 2,118.23
1991 66.49 20.16 2.06 6.36 0.89 29.50 4.29 2.45 5.76 2,128.95
a
Ref. 2.
b
To convert m3 to bbl, divide by 0.159.
12 LIQUEFIED PETROLEUM GAS Vol. 2
In frozen earth storage of propane, the walls and bottom of a pit in the ground
are frozen and a dome is constructed over the pit. The pressure in the storage
cavern is maintained at nearly atmospheric pressure by refrigeration systems
that cool the product to its boiling point at storage pressure. Heat leaks into the
cavity and vaporizes some of the propane. The vapor that is formed is compressed,
cooled, and returned to the pit as a liquid by the refrigeration system. Because this
storage must operate at temperatures considerably below the freezing point of wet
earth and at atmospheric pressure, it cannot be used for butane storage.
Underground storage caverns, which operate at approximately formation
temperatures and at the corresponding LPG vapor pressure, may be either
mined underground storage caverns or cavities that have been produced in a
salt formation by solution mining. The underground caverns must be of sufficient
depth to develop an overburden pressure greater than the vapor pressure of the
stored liquid. Mined storage caverns are 60–152 m deep, whereas salt formation
caverns may be from 106–1524 m deep. Underground as compared to above-
ground storage is much more economical for storage of large volumes, ie, more
than 2785 m3 (107 gal) of LPG. A washed-out salt cavern costs only 10–50% as
much to develop as typical mined cavern storage. In 1987, underground storage
capacity for LPG in the United States was 79 106 m3 ð493 106 bblÞ (2) and
90% of this was in salt formations.
7. Uses
About 35% of total U.S. LPG consumption is as chemical feedstock for petrochem-
icals and polymer intermediates. The manufacture of polyethylene, Polyethylene
and poly(vinyl chloride) requires huge volumes of ethylene (qv) and propylene
which, in the United States, are produced by thermal cracking/dehydrogenation
of propane, butane, and ethane.
Vol. 2 LIQUEFIED PETROLEUM GAS 13
Residential and commercial fuel demands represent about 24% of total U.S.
LPG consumption. Although this market demand is weather dependent, it has
assumed the characteristics of a mature market. Growth is related to the general
economic trends.
Nearly two-thirds of total butane supply, about 20% of total LPG, is con-
sumed in the manufacture of motor gasoline. However, the environmental man-
dates of the early 1990s have had a negative impact on this market segment.
These mandates have reduced gasoline volatility requirements, effectively redu-
cing the value of the butanes as blending stocks. However, normal butane can be
used as feedstock for production of isobutylene, a key ingredient of ether blend-
stocks, such as methyl tert-butyl ether [1634-04-4] (MTBE) for motor gasoline.
Shifts in U.S. use patterns can be seen in Table 3.
BIBLIOGRAPHY
‘‘Liquefied Petroleum Gas’’ in ECT 2nd ed., Vol. 12, pp. 470–480, by L. Pollack, Phillips
Petroleum Co.; in ECT 3rd ed., Vol. 14, pp. 383–394, by F. E. Selim, Phillips Petroleum
Co.
CITED PUBLICATIONS
1. Engineering Data Book, 10th ed., Gas Processors Supplier’s Association, Tulsa,
Okla., 1987.
2. LP-Gas Market Facts, National LP-Gas Association, Oak Brook, Ill., 1977.
3. Liquefied Petroleum Gas Specifications and Test Methods, Gas Processors Associa-
tion, GPA Publication 2140-92, Tulsa, Okla.
4. ASTM Standard D1835-91, American Society for Testing and Materials,
Philadelphia, Pa., 1992.
5. Storage and Handling of Liquefied Petroleum Gases, National Fire Protection
Association, NFPA 58, Boston, Mass., 1989.
6. Design and Construction of LP-Gas Installations at Marine Terminals, Natural Gas
Processing Plants, Refineries, Petrochemical Plants, and Tank Farms, API Standard
2510, 4th ed., American Petroleum Institute, Washington, D.C., Dec. 1978.
GENERAL REFERENCES
Petroleum Products Handbook, McGraw-Hill Book Co., Inc., New York, 1960.
C. C. McKee, ‘‘The Supply/Demand Outlook for LP-Gas,’’ Proceedings of the 58th Annual
Convention 1979, Gas Processors Association, Tulsa, Okla.
Magic Formula, LP-Gas, Duluth, Minn., Jan. 1980.
Stayin’ Alive, LP-Gas, Duluth, Minn., Jan. 1980.
R. RAY TAYLOR
Phillips Petroleum Company
14 LUBRICATION AND LUBRICANTS Vol. 2
LUBRICATION
AND LUBRICANTS
1. Introduction
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1221021802151519.a01.pub2
Vol. 2 LUBRICATION AND LUBRICANTS 15
in the cost of lubrication, in manpower and material (4%), account for the
remainder.
Tribology is a multidisciplinary science that embraces lubrication, friction,
wear, properties of lubricants, surface characterization, bearing materials, and
the selection and design of lubricating systems. The lubrication engineer
would add to this list lubricant and coolant selection, plant lubrication and
maintenance programs, and machine condition monitoring.
2. Fundamentals of Lubrication
Ra ¼ ðj y1 j þ j y2 j þ þ j yn jÞ=n ð1Þ
Where Ra is the center line or arithmetic average of the absolute distances, yi,
from the imaginary surface (mean line) for a given sampling length (usually
0.80 mm).
Roughness (Ra) values of machined surfaces range from 0.025 mm for ball
bearing surfaces to 25 mm clearance surfaces on rough machine parts (4–6).
The roughness of computer hard disk surfaces is measured in angstroms (Å) or
nanometers (nm) (7).
The total profile of a surface consists of a ‘‘waviness’’ and a roughness com-
ponent. The parameter Ra, although it is the most common measure of surface
roughness, is insensitive to the shape or waviness of the profile. A more useful
parameter is the root-mean-square (rms) roughness, Rq.
Where Rq is the rms deviation of yi from the mean line for a given sampling
length (4). The rms roughness of computer hard disk surfaces is <2 nm (7).
16 LUBRICATION AND LUBRICANTS Vol. 2
Where Rz is the average distance between the five highest peaks (Pi) and
the five deepest valleys (Vi) within the sampling length. It is also linked
with the machining parameter S2/8r, where S is the feed rate and r is the
tool radius (4).
Another index of roughness is Rt, the maximum peak-to-valley height.
Rt ¼ Rp þ Rv ð4Þ
Where Rp is the maximum peak height and Rv is the maximum valley depth
within the sampling length (4).
The description of the test disk used in the ASTM D 6425-02 test method for
measuring friction and wear illustrates the impact of surface texture on these
quantities (8) and is shown below and used with ASTM’s permission.
Test Disk, AISI 52100 Steel, 62 1 HRC Hardness. The surfaces of the disk
are lapped and free of lapping raw materials. The topography of the disk will be
determined by four values: 0.005 mm hRzi 0.65 mm; 0.035 mm hRai 0.050 mm;
0.020 mm hRpi 0.035 mm; 0.050 mm hRvi 0.075 mm
Four different measures of surface topography are specified for the disk in
order to get acceptable reproducibility of the test method (9).
Approximate roughness indexes obtained with various metal-working
processes are
admitted to a pressure of 0.133 Pa (13). The oxide and adsorbed layers on metal
surfaces can, therefore, be considered as lubricating films.
2.2. Friction. When two of these surfaces are brought together, they
initially touch at the highest asperities. The load, N, normal to the surfaces at
the contact points causes the asperities to deform until the pressure in the result-
ing contact areas just equals the yield pressure, p, of the asperities. The sum of
these contact areas is the real contact area, Ar. The yield pressure is equivalent
to the Brinell Hardness Number (BHN), in consistent units, measured at the
surface of the material (4).
The real area of contact is a minute fraction of the total surface area. For
example, with a typical bearing contact stress of 3 MPa and a bronze bearing
asperity yield pressure of 500 MPa, <1.0% of the nominal area would involve
asperity contact.
As the load on the surfaces increases, the asperities continue to deform,
the softer surface more than the other. More of the asperities come in contact
and the real area of contact grows. The opposing surfaces also tend to adhere
or bond to each other in the contact area. The shear strength of these bonds
depends on the time of contact and the difference in composition of the two
surfaces.
Sliding of one of these nonlubricated surfaces across the other requires a
friction force, F, to displace the contacting asperities. This force includes several
components, among them: a shear or adhesion component arising from bonding
of the contacting asperities; a plowing or deformation component, arising from
the interlocking of asperities; a lifting component to raise asperities over the
roughness of the mating surfaces.
The shearing component, Fs, may account for 90% or more of the total fric-
tion force. This component is proportional to the shear strength, s, of the asperity
junctions:
Fs ¼ Ar s ð6Þ
f ¼ sf =pm ð8Þ
18 LUBRICATION AND LUBRICANTS Vol. 2
V ¼ kNx=p ð9Þ
The fluid in contact with the surface of the moving plate has the same velo-
city, U, as the plate. The fluid in contact with the stationary plate has zero velo-
city. There is, therefore, a shear stress, T, on the fluid equal to the force, Fs,
required to keep the plate moving divided by the area, As, of the moving plate,
and is proportional to the velocity gradient, dU/dh, of the fluid. The parameter
Fs is a frictional force and the coefficient, m, is the dynamic viscosity of the fluid
(29). The unit of dynamic viscosity in the System International (SI) is pascal-
second (Pas). The customary unit is centipoise (cP), which is Pas 103.
Schoff and Kamarchik (30) describe methods of measuring the viscosity of a
wide variety of materials. Standard methods for measuring the dynamic viscos-
ity of lubricating oil can be found in Refs. 31–34.
Generally, however, it is easier to measure the kinematic viscosity, v, of a
lubricating oil using a capillary viscometer. Kinematic viscosity is the dynamic
viscosity divided by the density, d, of the fluid at the same temperature.
v ¼ m=d ð11Þ
Where Z ¼ (v þ 0.7) for 2.00 cSt v 2 107 cSt; A and B ¼ constants, and
T ¼ absolute temperature, 8K ¼ temperature,8C þ 273.2.
20 LUBRICATION AND LUBRICANTS Vol. 2
where YN ¼ H/U and Y ¼ kinematic viscosity at 1008C of the oil whose viscosity
index is to be calculated.
Viscosity index is sometimes used as a measure of the quality of lubricating
oil, especially for selecting base stocks for automotive engine oils and automatic
transmission fluids. This is not always applicable, however. The role of viscosity
index in base stock selection is described later.
Several aromatic type fluids, polyphenyl ethers, aryl phosphate esters, and
halogenated aromatic hydrocarbons have negative VIs.
Effect of Pressure on Viscosity. The lubricant film pressure in the con-
centrated contact areas of rolling element bearings, gears, cams, etc, can be as
high as 2000–3000 MPa. The viscosity of the film at these pressures could be
a million times higher than that of the lubricating oil at atmospheric pressure,
as illustrated in Fig. 1, or the lubricant may have solidified (38,40). The rate of
viscosity increase with pressure of a liquid lubricant varies with its composition
and chemical structure and with temperature and pressure.
Traction drives, found in some industrial machinery and the toroidal drives
in some continuously variable transmissions (CVTs) depend on this property of
traction fluids to transfer power.
Vol. 2 LUBRICATION AND LUBRICANTS 21
106
0°C
105
°C
33
Absolute viscosity, mPa·s(=cP)
104
°C
99
103
°C
218
102
10
1
0 200 400 600 800 1000
Pressure, MPa
Fig. 1. Viscosity pressure curve for typical petroleum oils, where —— ¼ Paraffinic, - - - -
¼ Alicyclic, " ¼ Solid.
and the load on the surfaces. In a journal bearing, eg, the lubricant film thick-
ness and the coefficient of friction are functions of the dimensionless Sommerfeld
bearing characteristic number S.
where P is the average pressure on the bearing surface, W/2RL; Z is the dynamic
viscosity, m, of the fluid; and N ¼ is the rotational speed of the journal, U/2pR.
The relationship between the kinetic coefficient of friction and the dimen-
sionless quantity, ZN/P, is illustrated in Fig. 2 (10). This is known as a
Stribeck curve.
At values of ZN/P greater than 30, lubrication is in the full fluid film
regime. Friction increases with increasing ZN/P because of increasing resis-
tance to flow. As ZN/P is reduced, by reductions in the speed or viscosity, or
by increases in load, the coefficient will reach a minimum value. Further reduc-
tion in ZN/P leads to partial breakdown in the fluid film and lubrication is in the
mixed-film regime. Friction increases as ZN/P decreases in this regime, as more
and more of the load is carried by asperity contact and less by fluid-film pressure.
Finally, a point is reached, as ZN/P gets smaller, where there is no fluid pressure,
the rate of increase in the coefficient of friction starts to level out, and lubrication
is in the thin-film boundary regime (43). Ideally, a bearing is designed to operate
where the coefficient of friction is at its minimum.
Hydrodynamic Lubrication Regime. Rohde (44) and Dowson (45) remind
us that the basic mechanism of fluid-film lubrication was explained by Reynolds
(46) in 1886, based on the earlier work of Petrov (47), and Tower (48).
The fundamental requirements of lubrication in the hydrodynamic regime
are the formation of a wedge-shaped film; and generation of pressure in the film,
sufficient to keep the surfaces apart, by the motion of the surfaces themselves.
Vol. 2 LUBRICATION AND LUBRICANTS 23
The pressure distribution in the lubricating film of the sliding bearing illu-
strated in Fig. 3 (29), moving with velocity U relative to a slanted stationary pad
in a fluid with viscosity m, and assuming no flow out of the sides of the bearing,
is (49)
The total force P that the bearing will support per unit width is
The analysis of hydrodynamic fluid films assumes laminar flow in the film, ie, a
Reynolds number, Re, <1000.
l ¼ h=s ð22Þ
where h is the calculated EHD film thickness, and s is the composite surface
roughness.
Vol. 2 LUBRICATION AND LUBRICANTS 25
where a and b are dimensions of the ellipse, h0 is the original film thickness
(for small values of h0 relative to a or b), m is the dynamic viscosity of the fluid,
and W is the load that produces a film of thickness h after time t.
Boundary Film Lubrication Regimes. At lower values of the Sommerfeld
number (eq. 15), ie, when unit loads, P, are very high or when sliding speed, N, or
viscosity, Z, are low, a full-fluid film cannot be generated. The film thickness is
smaller than the height of asperities on the surfaces and the surfaces come in
contact with each other. When there is no fluid pressure, and only a thin film
of lubricant on the surfaces, lubrication is said to be in the boundary regime.
Except in the case of starved lubrication, ie, inadequate or blocked lubricant sup-
ply, lubricating oils and greases usually provide lubrication in the mixed or par-
tial film regime, where the load is supported by a combination of surface asperity
contact and fluid film pressure, at lower Summerfeld numbers.
Lubricants designed to operate in the mixed- or partial-film regime may
contain additives that form lubricating films on the bearing surface by adsorp-
tion, chemisorption, or tribochemical reaction. The additives may be friction
modifiers, antiwear additives, EP additives, or multifunctional additives. These
films are discussed later under lubricant films.
2.7. Micro- or Nanotribology. Micro- or nanotribology began with the
ability to study tribological phenomena at the molecular or atomic level. A key
tool is the atomic force microscope. With this tool a probe, the tip of which has
a radius <10 nm (<109 m), can be moved over a surface to measure frictional
resistance and other phenomena. At this level, the apparent surface interactions
do not obey the same ‘‘laws’’ as those observed on a macroscopic scale (57,58).
Topics in micro- or nanotribology are discussed in Refs. 59 and 60.
A principal application for micro- or nanotribology is computer hard disk
drives. The rms roughness, Rq, of the disk surface is <2 nm. The surface is
26 LUBRICATION AND LUBRICANTS Vol. 2
The most important component of a lubricating oil or grease is its base oil.
Although often supplemented by additives, the base oil determines the flow char-
acteristics of the lubricant, its oxidation stability (sludge and deposit-forming
tendency), its volatility, and its corrosion potential.
Most lubricant base oils are mixtures of paraffins (straight- or branched-
chain hydrocarbons), naphthenes (cycloparaffins), and aromatics (alkyl benzenes
and multiring aromatics), typically containing 20–40 carbon atoms per molecule.
If paraffins predominate, the base stock is paraffinic. If naphthenes predominate,
it is a naphthenic base stock. Paraffins and naphthenes are saturates, meaning
that all of the carbon atoms in the hydrocarbon molecule are singly bonded to
another carbon atom or to a hydrogen atom. These lubricant base stocks are
manufactured by the distillation of selected crude oils, followed by further refin-
ing of the lube oil distillates by conventional separation or modern conversion
processes. When refined by conventional separation processes, the type of base
stock is crude specific, meaning that paraffinic base stocks come from paraffinic
crude oils, and naphthenic oils from naphthenic crudes. When refined by modern
conversion processes, base stocks are less crude specific, since these processes
are capable of converting naphthenic and aromatic compounds to paraffins
(see also HYDROCARBONS).
3.1. Base Stock Classification. The American Petroleum Institute
(API) defines a base stock as ‘‘. . . a lubricant component that is produced by a sin-
gle manufacturer to the same specifications (independent of feed source or man-
ufacturer location); that meets the same manufacturer’s specification; and that is
identified by a unique formula, product identification number, or both. Base
Stocks may be manufactured using a variety of different processes including
but not limited to distillation, solvent refining, hydrogen processing, oligomeriza-
tion, esterification, and rerefining. Rerefined stock shall be substantially free
from materials introduced through manufacturing, contamination, or previous
use’’ (62).
‘‘A base stock slate is a product line of base stocks that have different
viscosities but are in the same base stock grouping and from the same manufac-
turer’’ (62).
‘‘A base oil is the base stock or blend of base stocks used in an API-licensed
oil’’ (62).
API has also established five base stock categories, classified according
to saturates content, sulfur content, and VI. The classification is shown in
Table 2 (62).
Base oils with higher saturates content are generally more resistant to,
and easier to protect against, oxidation. They also have higher VIs. Paraffinic
oils, at the same saturates level, have higher VIs than naphthenic oils. Sulfur
compounds produce corrosive material when oxidized. Base stocks with a wide
28 LUBRICATION AND LUBRICANTS Vol. 2
molecular weight range tend to be more volatile than those with a narrow
range.
Base Stocks in API groups I–IV are paraffinic hydrocarbons and are manu-
factured in several viscosity grades. Naphthenic base stocks are in group V
(Refs. 63–70). Typically, the saturates content of group II and III base stocks
is >99%, and the sulfur content is <15 ppm (<0.0015%). Group IV stocks are
100% saturates and contain no sulfur. The viscosity indexes of group II stocks
are typically 100–115, and those of groups III and IV are 120–140. Group IV
base stocks have the best low temperature flow characteristics because they
contain no wax. Naphthenic stocks, with minimal wax contents, also have good
low temperature flow properties. They also have the lowest viscosity indexes.
Oxidation stability of the paraffinic stocks improves with each group number,
as do volatility and deposit, sludge, and soot control.
3.2. Synthetic Base Stocks. The word ‘‘synthetic’’ is not part of the
API classification. It is a marketing term, not a technical term. The application
of the word synthetic is the subject of controversy because of the difficulty of
defining it. It was originally used to differentiate between base stocks made by
conventional crude oil refining processes and those synthesized from other che-
micals. Polyalphaolefins in group IV are still called synthetic, as are the diesters,
polyol esters, polyglycols, etc, in group V. This definition has been blurred, how-
ever, by the availability of base stocks derived from petroleum feed stocks by
modern conversion processes (71). ‘‘Synthetic’’ is an accepted description, eg,
for group III oils.
3.3. Base Stock Manufacturing Processes. Figure 4 lists some of the
refining and conversion processes used to manufacture groups I, II, and III base
stocks (72,73). Crude oil is first fractionated in an atmospheric distillation tower
to produce light gases and fuel products. The residue, or bottoms, from the
atmospheric tower are then fractionated in a vacuum distillation tower to pro-
duce gas oil and lube oil fractions. The vacuum residue is separated with propane
to produce asphalt and deasphalted cylinder oil.
Solvent Extraction. In the solvent extraction process, lube oil feedstock
from the vacuum tower flows upward through a treating tower, countercurrent
to a stream of solvent. The solvent preferentially removes undesirable tars,
resins, asphaltic compounds, and polycyclic naphthenes and aromatics. Solvents
used include phenol, furfural,and N-methyl-pyrrolidone (NMP). Solvent is then
Vol. 2 LUBRICATION AND LUBRICANTS 29
Atmos-
pheric
Crude oil Fuel products Vacuum gas oil
pipe-
still
Vacume
Atmospheric Lube oil fractions
pipe-
residuum
still
Vacuum residume
Raffinate
Lube oil Solvent
hydro- Dewax Group II base stocks
fractions extraction
conversion
Raffinate Wax
Lube oil Solvent hydro- Groups II and III
hydro-
fractions extraction isomerization base stocks
conversion
Lube Wax
Lube oil hydro- hydro- Groups II and III
fractions cracking isomerization base stocks
stripped from both the aromatic extract and raffinate streams. Solvent extraction
increases the VI and the stability of the raffinate (71,73).
Hydrofinishing. Hydrofinishing is catalytic hydrogenation process that
converts unstable compounds remaining after solvent extraction to stable ones.
It converts some aromatics to naphthenes and removes some sulfur compounds
and other trace materials. It is a relatively mild process compared to the modern
hydrocracking process that causes major molecular changes (74). It is severe
enough, however, to produce naphthenic base oils from unextracted distillates
that do not require labeling as carcinogenic under OSHA’s Hazardous Sub-
stances Communication Standard. This makes them suitable for use as base
oils for metalworking fluids, greases, and other industrial lubricating products.
Dewaxing. Paraffinic distillates contain some high molecular weight,
high melting point paraffin waxes. At temperatures below their melting point,
these waxes crystallize and cause the oil to gel. Several methods are used to
reduce the wax content of lubricating oil base stocks. In one method, the hydro-
fined raffinate is mixed at low temperature with a chilled solvent, in which the
heaviest waxes are insoluble. The wax crystallizes and is then separated from
the mixture by cold filtration. Solvents used include propane, methyl ethyl
ketone (MEK), methyl isobutyl ketone, and a mixture of MEK and toluene.
Another method for removing wax from the raffinate is catalytic dewaxing.
This method uses a shape-selective dewaxing catalyst to crack straight-chain or
slightly branched waxes into naphtha and gas (73).
Isomerization technology is used in the manufacture of group II and III
base stocks. In this process, high melting point straight-chain paraffins are
30 LUBRICATION AND LUBRICANTS Vol. 2
3.11. Rerefined Base Stock. In its definition of a base stock, the API
states ‘‘Rerefined stock shall be substantially free from materials introduced
through manufacturing, contamination, or previous use’’ (62).
Products made from rerefined stock are subject to the same stringent refin-
ing, compounding and performance standards applied to virgin oil products.
Rerefined oil may, in fact, have superior oxidation stability than virgin stocks
because the easily oxidized compounds will have been reacted during previous
use and then removed during reprocessing (71,94).
In the rerefining process, used oil is preferably segregated by type, col-
lected, and delivered to the reprocessing facility. The oils are then screened
and inappropriate feed stocks rejected. Solid and other gross contaminants are
then separated by, eg, propane precipitation, alcohol–ketone precipitation,
acid–clay filtration, etc. The filtrate is dehydrated and then refractionated by
vacuum distillation. The distillates are then finished by hydrotreating or clay
filtration or both, and then vacuum distilled to obtain the desired viscosity
grades (71,94).
OH O
ROO ROOH
R R
O O
ROO
R ROO R
pumps. Concentrated contacts have a very small contact area which results in
extremely high pressure on the contact area. Some friction modifying additives
work with ZDDP to reduce both friction and wear of the rubbing surfaces.
The ZDDP compounds are multifunctional. They are secondary oxidation
inhibitors, decomposing hydroperoxides in the branching step of hydrocarbon
oxidation. They are also effective bearing corrosion inhibitors.
This compound is manufactured by first reacting an alkyl or aryl alcohol
with phosphorous pentasulfide (P2S5), and then neutralizing the resultant acid
with zinc oxide. The alkyl alcohol may be primary alkyl, branched chain primary
alkyl, secondary alkyl or, rarely, tertiary alkyl. As a rule, the secondary and ter-
tiary alkyl derivatives are more effective as antiwear additives, but are the least
stable. The aryl derivatives generally have the best thermal stability and the
least potent antiwear activity (108,109).
The ZDDP compounds, particularly those with lower thermal stability, can
decompose at temperatures below those of rubbing steel surfaces and become cor-
rosive to copper and copper alloys. This is generally not a problem in internal
combustion engines, but it can be a problem in machinery using, eg, bronze bush-
ings, pistons, or clutch plates. However, by the judicious selection of alcohols, and
careful control of the manufacturing process, ZDDP compounds are made that
perform well in machines such as automatic transmissions and piston type
hydraulic pumps that contain parts made with copper alloys.
Despite its effectiveness as a wear, oxidation and corrosion inhibitor, and
its relatively low cost, there are serious efforts to reduce or eliminate the use
of ZDDP additives in engine oils. When the oil gets in the combustion chamber,
the additive decomposes at high temperature and, because a metal is present,
leaves ash, a solid residue, and other deposits in the chamber. These deposits
interfere with engine efficiency and emissions control. There is also evidence
that ZDDP decomposition products poison the catalysts used to convert CO,
NOx, unburned hydrocarbons and volatile organic compounds to less harmful
emissions. Zinc oxide, phosphorus pentasulfide, and a zinc pyrophosphate
glaze have been found on catalyst surfaces that prevent passage of exhaust
gases into the porous structure of the catalyst (110). (see also EMISSION CONTROL,
AUTOMOTIVE). Lacking an acceptable emission systems protection test, the
International Lubricant Standardization and Approval Committee (ILSAC)
places maximum phosphorus and sulfur limits on the GF-4 engine oil specifica-
tion. These limits are expected to be reduced further in the next generation
specification.
4.6. EP and Antiwear Additives. There is very little distinction
between so-called EP and antiwear additives. Both react with rubbing metal
surfaces to form lubricating films, and they both inhibit wear. ZDDP additives
have relatively high activation energy and are not effective until the surface
temperature gets >1508C. The EP additives are used in relatively high load,
slow speed applications to prevent catastrophic failure of the surface. Mild EP
additives used in automotive and industrial gear oils have lower activation
temperatures than ZDDP. ‘‘Active’’ EP additives used in metal cutting oils, like
elemental sulfur, chlorinated hydrocarbons and some sulfurized esters, can react
at room temperature. Table 4 lists some of the many compounds used as EP
and antiwear additives in lubricating oils (103,111,112).
Vol. 2 LUBRICATION AND LUBRICANTS 37
5. Lubricating Oils
Lubricating oils are specifically formulated for virtually every type of machine
and manufacturing process. The base stocks and the type and concentration of
additives used for these oils are selected based on the requirements of the
machinery or process being lubricated, by the quality required by the builders
and the users of the machinery, and by government regulation. Table 5 is a
partial list of the principal applications for which lubricating oils are specifically
formulated:
Each of these oils has a unique set of performance requirements. In addition
to proper lubrication of the machine or process, these requirements include
maintenance of the quality of the lubricant itself, as well as the effect of the lubri-
cant’s use and disposal on energy use, the quality of the environment and on the
health of the user. The performance requirements of a few of these lubricating
oils are discussed below.
40 LUBRICATION AND LUBRICANTS Vol. 2
5.1. Gasoline Engine Oils for Passenger Cars and Light Trucks.
Internal combustion engines provide the best example of performance require-
ments dictating the formulation and quality of the engine oil. These engines
have lubrication requirements in nearly every regime: hydrodynamic, elastohy-
drodynamic, squeeze film, EP, boundary, and mixed film. They operate at high
oil and surface temperatures, must start up at very low temperatures, are sub-
ject to contamination (by fuels, combustion products, oxidation products, and
dirt), and they must meet stringent fuel economy and exhaust emission stan-
dards. Their lubricating oils contain a wide variety of additives and their base
oil requirements determine the type and quality of base stocks produced in
lube refineries and chemical manufacturing plants. The performance require-
ments of gasoline engine oils for use in automobiles are a good example of the
relationship between the performance requirements and the formulation of
lubricating oils for a specific application.
Gasoline engine oils for use in automobiles in the United States are licensed
and certified to assure that they meet the minimum performance standards
established by industry technical specifications. The Engine Oil Licensing and
Certification System (EOLCS), administered by the API, is a voluntary program
that authorizes oil marketers to use the API Engine Oil Quality Marks–the
API Service Symbol ‘‘Donut’’ and Certification Mark ‘‘Starburst’’ (127). This
Vol. 2 LUBRICATION AND LUBRICANTS 41
program is a cooperative effort between the oil industry and vehicle and engine
manufacturers–Ford, General Motors and Daimler Chrysler; the Japan Auto-
mobile Manufacturers Association (JAMA); and the Engine Manufacturers
Association (EMA).
Performance requirements, test methods, and limits are cooperatively estab-
lished by vehicle and engine manufacturers, technical societies, [eg, the Society of
Automotive Engineers (SAE) and the American Society of Testing and Materials
(ASTM)], and industry associations (such as the American Chemistry Council and
API). An ongoing monitoring and enforcement program to ensure licensees adhere
to industry technical specifications backs APIs Engine Oil Program (128).
The API Service Symbol ‘‘Donut’’ is divided into three parts: (1) The top of
the Donut shows the oil’s performance level for gasoline and/or diesel engines.
The letter ‘‘S’’ followed by another letter (eg, ‘‘SM’’) refers to oil suitable for gaso-
line engines. The letter ‘‘C’’ followed by another letter and/or number (eg, CI-4)
refers to oil suitable for diesel engines. (2) The center identifies the oil’s viscosity
characteristics. The numbers and the letter ‘‘W’’ indicate the SAE viscosity
grade. These grades are defined in SAE J300, May 2004 (129). The low tempera-
ture ‘‘W’’ grades indicate how quickly an engine will crank and how well the oil
will flow to lubricate critical engine parts at low temperatures. The grades with-
out the ‘‘W’’ indicate the oil’s operating temperature flow properties. A multi-
grade oil (eg, 5W-30) provides good flow capability for low temperatures and
adequate viscosity characteristics for high temperature lubrication. It meets
the viscosity requirements of both the ‘‘W’’ grade and the other grade indicated
(129). (3) The bottom of the Donut tells whether the oil has demonstrated energy
conserving properties in a standard test in comparison with reference oil. It may
also include a supplemental category description (eg, CI-4 PLUS).
The API Certification Mark ‘‘Starburst’’ (Fig. 6) is designed to identify
engine oils recommended for a specific application (such as gasoline engine
service). An oil may be licensed to display the Starburst only if the oil satisfies
the most current requirements of the International Lubricant Standardization
and Approval Committee (ILSAC) minimum performance standard for this appli-
cation. Members of ILSAC include representatives of the Japanese Automobile
Manufacturers Association, DaimlerChrysler Corporation, Ford Motor Company
and General Motors Corporation.
An engine oil meeting the latest API service category may or may not meet
all of the requirements of the latest ILSAC category. When it does meet both
requirements, its container may display both symbols.
The current ILSAC minimum performance standard, suitable for use in
the current and all earlier model year automobiles, is ILSAC GF-4 Standard
for Passenger Car Motor Oils (130). The corresponding API Service Category is
API Service SM. Both were issued in 2004 and superceded the 2001 standards,
ILSAC GF-3 (131) and API Service SL. Those, in turn, superceded ILSAC GF-2
and API Service SJ, issued in 1996 and API Service SH, issued in 1992 and
now obsolete. After April 2005, the Starburst can no longer be used for GF-3 oils.
The performance required by ILSAC for GF-4 Engine Lubricating Oils
is measured in a battery of full scale engine tests and laboratory bench tests
conducted by qualified independent laboratories. Tables 6–11 describe these
engine and bench test requirements for ILSAC GF- 4, GF-3, and GF-2 and for
API Service SM, SL, SJ, and SH (130,131).
Engine test requirements in the ILSAC GF-4 Standard include (1) The
ASTM Sequence IIIG test measures viscosity increase, piston deposits at high
Table 6. Engine Test Requirements, ASTM Sequence IIIG, IIIF, IIIE for Wear, Piston
Deposits, and High Temperature Oxidation
ILSAC ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
CLASSIFICATION:
API SERVICE Proposed API SM API SL API SJ API SH
CATEGORY:
ISSUE DATE: 2004 2001 1996 1992
ASTM Engine SEQUENCE IIIG SEQUENCE SEQUENCE SEQUENCE
Test Sequence IIIF IIIE IIIE
ASTM Test Procedure: ASTM D 5533 ASTM D 5533
kinematic viscosity 150 max 275 max 375 max 375 max
increase at 408C, %
average weighted piston 3.5 min 4
deposits, merits
hot stuck rings none none none none
cam plus lifter wear,
average, mm 60 max 20 max 30 max 30 max
maximum per 64 64
position, mm
piston skirt varnish 9.0 min 8.9 min 8.9 min
oil consumption, L 5.2 max 5.1 max 5.1 max
avg. oil ring land deposits 3.5 min 3.5 min
avg. engine sludge rating 9.2 min 9.2 min
lifter sticking none none
cam þ lifter scuffing none none
low temperature stay in grade or rate and
pumping viscosity at next higher report
end of test by ASTM D grade
4684 (MRV TP-1)
Vol. 2 LUBRICATION AND LUBRICANTS 43
Table 7. Engine Test Requirements, ASTM Sequence VG and VE for Wear, Sludge
and Varnish
ILSAC CLASSIFICATION: ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
API SERVICE CATEGORY: Proposed API SM API SL API SJ API SH
ISSUE DATE: 2004 2001 1996 1992
ASTM ENGINE TEST SEQUENCE VG SEQUENCE SEQUENCE SEQUENCE
SEQUENCE VG VE VE
ASTM Test Procedure: ASTM D 6593 ASTM D 6593 ASTM D 5302 ASTM D 5302
average engine sludge, 7.8 min 7.8 min 9.0 min 9.0 min
merits
average rocker cover 8.0 min 8.0 min 7.0 min 7.0 min
sludge, merits
average engine varnish, 8.9 min 8.9 min 5.0 min 5.0 min
merits
average piston skirt 7.5 min 7.5 min 6.5 min 6.5 min
varnish, merits
oil screen sludge 20 max 20 max 20 max 20 max
(clogging), % area
oil screen debris, rate and rate and
% area report report
hot stuck compression none none none none
rings
cold stuck rings rate and rate and
report report
oil ring clogging, rate and rate and rate and 15
% area report report report
cam lobe wear,
average, mm 127 max 130 max
maximum per 380 max 380 max
position, mm
cam follower pin rate and rate and
wear, cyl #8, report report
avg, mm
ring gap increase, rate and
cyl #1 & #8, avg, mm report
cylinder bore wear rate and
report
ring wear rate and
report
operating temperature and valve train wear. At the end of the test, the low tem-
perature pumping viscosity of the oil is measured. It is run in a 1996/1997 3800 cc
series II General Motors V-6 fuel-injected gasoline engine. After a 25-min
break-in, it operates at 125 bhp, 3600 rpm, 1508C oil temperature for 100
hours (130,132). (2) The ASTM Sequence VG test (ASTM D 6593-04) evaluates
the oil’s ability to prevent sludge and varnish formation. It is run in a 1994
Ford V-8, 4.6 liter fuel injected gasoline engine. The procedure involves 54 cycles
in 216 h, each cycle consisting of three different sets of operating conditions
(130,132,133). (3) The ASTM Sequence IVA test (ASTM D 6891-03) measures
the average camshaft lobe wear in a 100 hour test in a KA24E Nissan 2.4
liter, 4-cylinder engine (130,132,134). (4) The ASTM Sequence VIII test (ASTM
44 LUBRICATION AND LUBRICANTS Vol. 2
Table 8. Engine Test Requirements, ASTM Sequence IVA, VIII, IID and CRC-L-38
For Valve Train Wear, Bearing Corrosion, Shear Stability and Engine Rusting
ILSAC CLASSIFICATION: ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
API SERVICE Proposed API API SL
CATEGORY: SM
ISSUE DATE: 2004 2001
ASTM ENGINE TEST SEQUENCE SEQUENCE API SJ API SH
SEQUENCE IVA IVA
ASTM Test Procedure: ASTM D 6891 ASTM D 6891 1996 1992
average cam wear 90 max 120 max
(7 position average) mm
ENGINE TEST SEQUENCE SEQUENCE CRC L-38 CRC L-38
SEQUENCE VIII VIII
ASTM Test Procedure: ASTM D 6709 ASTM D 6709 ASTM D 5119 ASTM D 5119
bearing weight loss, mg 26 max 26.4 max 40 max 40 max
shear stability (10 hour stay in grade stay in grade stay in stay in grade
stripped viscosity) grade
piston skirt varnish 9 min
ASTM ENGINE TEST (See Ball rust SEQUENCE SEQUENCE
SEQUENCE: bench test) IID IID
ASTM Procedure: ASTM D ASTM D
5844 5844
avg. engine rust rating 8.5 min 8.5 min
lifter sticking none none
Table 9. Engine Test Requirements, ASTM Sequence VIB and VIA for Fuel Economy
ILSAC CLASSIFICATION: ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
API SERVICE CATEGORY: Proposed API SM API SL API SJ
ISSUE DATE: 2004 2001 1996
ASTM ENGINE TEST SEQUENCE VIB SEQUENCE VIB SEQUENCE VIA
SEQUENCE
ASTM Test Procedure: ASTM D 6837a ASTM D 6837a ASTM D 6202a
SAE 0W-20 and 5W-20
viscosity grades
% fuel economy improvement, 2.3 min 2.0 min
Phase 1
% fuel economy 2.0 min 1.7 min 1.4 min
improvement, Phase 2
SAE 0W-30 and 5W-30
viscosity grades
% fuel economy 1.8 min 1.6 min
improvement, Phase 1
% fuel economy 1.5 min 1.3 min 1.1 min
improvement, Phase 2
% total fuel economy improvement 3.0 min
SAE 10W-30 and all other
grades not listed above
% fuel economy improvement, 1.1 min 0.9 min
Phase 1
% fuel economy improvement, 0.8 min 0.6 min 0.5 min
Phase 2
% total fuel economy improvement 1.6 min
a
All fuel economy values, Phases 1 and 2, are determined relative to ASTM Reference Oil BC.
Vol. 2 LUBRICATION AND LUBRICANTS 45
Table 10. Bench Test Requirements for Catalyst Compatibility, Wear and Volatility
ILSAC CLASSIFICATION: ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
API SERVICE CATEGORY: Proposed API SM API SL API SJ API SH
ISSUE DATE: 2004 2001 1996 1992
Catalyst compatibility
phosphorus content, mass %, 0.08 max 0.1 max
ASTM D 4951
SAE 0W-20, 5W-20, 5W-30, 0.1 max
and 10W-30 only
SAE 5W-30 and 10W-30 only 0.12 max
sulfur content, mass %, ASTM
D 4951 or D 2622
SAE 0W and 5W multigrades 0.5 max
SAE 10W multigrades 0.7 max
Wear
phosphorus content, mass %, 0.06 min
ASTM D 4951
NOACK Volatility, ASTM D 5800
evaporation Loss after 1 h 15 max 15 max
at 2508C, %
SAE 0W-XX 22 max NR
SAE 5W-XX 22 max 25 max
SAE 10W-30 22 max 20 max
SAE 15W-40 20 max 18 max
Volatility, Simulated
Distillation, ASTM D 6417
% volatilized to 3718C 10 max 10 max
SAE 0W-XX 17 max NR
SAE 5W-XX 17 max 20 max
SAE 10W-30 17 max 17 max
SAE 15W-40 15 max 15 max
Table 11. Bench Test Requirements for High Temperature Deposits, Filterability, Foam
Tendency and Stability, Engine Rusting and Flash Point
ILSAC CLASSIFICATION: ILSAC GF-4 ILSAC GF-3 ILSAC GF-2
API SERVICE CATEGORY: Proposed API SM API SL API SJ API SH
ISSUE DATE: 2004 2001 1996 1992
High temperature deposits,
TEOST, ASTM D 6335
Total Deposits, mg 35 max 45 max
Filterability
Engine Oil Water Tolerance
Test, ASTM D 6794
% flow reduction with 50 max 50 max
0.6% water
% flow reduction with 50 max 50 max
1.0% water
% flow reduction with 50 max 50 max
2.0% water
% flow reduction with 50 max 50 max
3.0% water
Engine Oil Filterability Test,
ASTM D 6795
% flow reduction 50 max 50 max 50 max 50 max
Foam tendency/stability,
ASTM D 892 (Option A)
sequence I Oct-00 Oct-00 Oct-00 Oct-00
sequence II 50/0 50/0 50/0 50/0
sequence III Oct-00 Oct-00 Oct-00 Oct-00
high temperature 100/0 100/0 200/50 NR
Engine rusting, Ball Rust Test, (See Engine Test
ASTM D 6557 Sequence IID)
Avg Gray Value 100 min 100 min
Flash point, ASTM D 92, 8C
0W-XX 200 min NR
5W-XX 200 min 200 min
10W-30 200 min 200 min
15W-40 215 min NR
are engine durability, reduction of NOx, and soot emissions and compatibility
with exhaust aftertreatment systems. As with gasoline engine oils, the perfor-
mance of diesel engine oils is continually being upgraded to meet engine manu-
facturers’ needs, consumer requirements and government regulations. Diesel
engine oils are licensed and certified by the API in its Engine Oil Licensing
and Certification System (EOLCS) (130). Performance Requirements, test meth-
ods, and limits are cooperatively established by the ASTM Heavy Duty Engine
Oil Classification Panel (HDEOCP) and the Diesel Engine Oil Advisory Panel
(DEOAP). Both of these groups are made up of members of the Engine Manufac-
turers Association (EMA) and oil and additive company members of API (148).
Table 12 is an example of the performance criteria and the engine and
bench tests used to measure them, for the API ‘‘C’’ category diesel engine oils.
This table illustrates the increasingly rapid growth of new performance require-
ments (categories), the number of tests and the number of new engines required
Vol. 2 LUBRICATION AND LUBRICANTS 47
CRC L-37 24-h dynamometer test, operated at high torque, low speed condi-
tions, to measure gear scoring and wear (162).
CRC L-42 2-h dynamometer test, operated at high speed with occasional shock
loading, to measure gear scoring (162).
CRC L-60-1 48-hour oxidation test with motored gears to measure viscosity in-
crease and insolubles (162).
CRC L-33 7 days in humidity cabinet after initial motoring phase to rate rust-
ing (162).
ASTM D 130 to measure copper strip corrosion (163).
ASTM D 892 to measure foam tendency (164).
50 LUBRICATION AND LUBRICANTS Vol. 2
Gear Oil viscosity grades are defined by the SAE J306 Automotive Gear Oil
Viscosity Classification (160,161,165).
The MT-1 requirements include additional CRC L-60-1 ratings, ASTM D 5662-
99 seal compatibility tests (166), ASTM D 5579-04 high temperature cyclic dur-
ability test (167), a more severe copper corrosion requirement, and ASTM D
4998-95(2003)e FZG tractor hydraulic fluid wear test (161,162,168).
SAE J2360 and MIL-PRF-2105E, which have identical performance require-
ments, combine the requirements of GL-5 and MT-1 gear oils (160).
Table 14. Suggested Viscosity Grade of Oil for Plain Bearings Operating between
15 and 608C
Loading on projected bearing area
Table 15. ISO Viscosity Grade Guide Recommendations for Enclosed Gears
Ambient temperature
Type of gear 10 to 108C 10 to 508C
parallel shaft units: input speeds to 3600 rpm
single reduction: shaft center distancea
up to 20 cm 68, 100 100, 150
20–50 cm 68, 100 150, 220
> 50 cm 100, 150 150, 220
double reduction: shaft center distance
up to 20 cm 68, 100 100, 150
> 20 cm 100, 150 150, 220
triple reduction: shaft center distance
up to 20 cm 68, 100 100, 150
20–50 cm 100, 150 150, 220
>50 cm 150, 220 220, 320
input speeds > 3600 rpm: all types and sizes 46 68
planetary units: diameter of housing
up to 40 cm 68, 100 100, 150
> 40 cm 100, 150 150, 220
bevel gears, spiral or straight: cone distanceb
up to 30 cm 68, 100 150, 220
> 30 cm 100, 150 220, 320
gear motors and shaft mounted units 68, 100 150, 220
a
Shaft center distance ¼ distance between centers of largest gear and its pinion.
b
Cone distance ¼ one-half the outside diameter of the largest gear.
52 LUBRICATION AND LUBRICANTS Vol. 2
5.7. Turbine Oils for Industrial Steam and Gas Turbines. Steam
and gas turbines are prime movers for electric power generation, for large
ocean-going vessels, for pumps and compressors, for processing sugar cane, for
rolling steel, and in many other industrial processes. Lubricating oils for the
thrust and radial shaft bearings on these turbines may also be used for the gen-
erators, reduction gears, compressors, and other machines that the turbines are
driving.
Compared to internal combustion engine oils, turbine oils have a relatively
small number and concentration of additives. Their lives, however, are measured
not in miles or hours, but in decades. A popular advertisement several years ago
showed a young engineer next to a large turbine with the caption The oil in this
turbine was put there before this engineer was born. Steam and gas turbine
bearings are lubricated by an oil circulation system, which includes the oil, a
pump, driven by the turbine shaft, an oil reservoir, a filter, an oil cooler, the bear-
ings, and auxiliary equipment.
Turbine oils must have excellent thermal and oxidation stability, protect
bearing surfaces from corrosion, separate easily from water, and they may not
entrain air or foam excessively. Some may also require load carrying capacity
beyond that afforded by a hydrodynamic lubricating film.
Typical specifications for an ISO 32 grade steam turbine oil and an ISO 32
grade gas and steam turbine oil are shown in Table 16. Specifications for oils
requiring additional load carrying capacity have a minimum ASTM D 5182
FZG failure load stage of 8 or 9 (177–181).
Premium turbine oils can be formulated for >10,000 h D 943 oxidation lives,
RPVOT lives of >1000 min, D 1401 demulsibility of 41/39/0 in 3 min and an
FZG failure load stage of 12 (178).
6. Metalworking Fluids
Rake face
Cutting
tool
Clearance face
Direction
of motion
Built
up
edge
metal, the mix of cutting operations, recovery and disposal of the spent fluid,
plant size, water quality, type of filtration, and cost (183).
6.2. Metal Cutting and Grinding Applications. Lubrication in metal
cutting and grinding operations differs from lubrication in bearings and gears in
several important ways: (1) The relative motion of the chip and workpiece sur-
faces on the rake face and clearance face of the cutting tool (Fig. 7) tends to
move fluid away from the contact zone rather than into it. Lubrication, therefore,
is always in the boundary regime. Cutting fluid enters the contact zone through a
labyrinth of capillaries formed by asperities on the surfaces. It is drawn in by the
pressure difference between the contact zone and the atmosphere. (2) As the chip
shears off the workpiece, nascent metal surfaces are formed, which rub on the
rake and clearance faces of the cutting tool. Without an oxide coating, these
surfaces are very reactive. Pieces of this active metal weld together to form a
‘‘built-up edge’’, which causes poor quality of the machined surface and wear of
the clearance face. Wear of the rake face occurs behind the built-up edge and
shows up as a crater some distance from the tool tip. Lubricating additives in
the cutting fluid must react with the nascent metal surfaces, at the temperature
on the surfaces, to minimize formation of the built-up edge and prevent wear of
the cutting tool. (3) The additives must also reduce friction force on the chip. The
higher the friction force on the chip, the smaller the shear angle, the longer the
shear plane and the more force it takes to shear the metal and form the chip.
About one-quarter of the work in metal cutting is done to overcome friction
and about three-quarters is done to form the chip. (4) Finally, cooling is extre-
mely important to prevent distortion of the machined surface and to extend
the life of the tool. Tool life decreases exponentially as the temperature at the
tool–chip interface increases (188):
Tyn ¼ K ð24Þ
where T is the tool life (min); y is temperature at the tool–chip Interface (8C); and
n is an exponent dependent on type of tool, with value usually between 20 and 30;
and K is a constant dependent on tool and workpiece material.
The other function of the cutting fluid is to flush chips and metal fines,
called swarf, from the cutting area. This prevents the swarf from damaging
the finished surface.
Vol. 2 LUBRICATION AND LUBRICANTS 55
7. Lubricating Grease
100
80
60
40
Soda
base
20
Aluminum
10 base
Apparent viscosity, Pa·s
8
6 Calcium
base
4
1
0.8
Viscosity of mineral oil used
0.6
0.4
0.2
0.1
10 20 40 60 100 200 400 600 1000 2000 4000 10,000
Rate of shear, s−1
Fig. 8. Effect of shear rate on the apparent viscosity of three NLGI grade 1 Greases at
258C.
8. Solid-Film Lubrication
There are several applications where liquids do not provide adequate lubrication.
In aerospace applications, eg, wide temperature ranges (240 to 9008C) are
encountered accompanied by high vacuum. The manufacture of glass products
requires lubricants that work at the temperature of molten glass. Lubricants
are also needed that can withstand chemical attack in a corrosive atmosphere
and are resistant to acids, aggressive gases, liquid oxygen, fuels, and solvents.
Electrical contacts and precision machinery that require low start-up friction
Table 17. Grease Application Guidea,b
Calcium, Calcium, Aluminum Calcium Lithium
Properties Aluminum Sodium conventional anhydrous Lithium complex complex complex Polyurea Organo clay
dropping 110 163–177 96–104 135–143 177–204 260 þ 260 þ 260 þ 243 260 þ
point, 8c
usable tempe- 79 max 121 max 93 max 110 max 135 max 177 max 177 max 177 max 177 max 177 max
rature, 8c
water good– poor–fair good– excellent good good– fair– good– good– fair–
resistance excellent excellent excellent excellent excellent excellent excellent
work stability poor fair fair–good good– good– good– fair–good good– poor– fair–good
excellent excellent excellent excellent good
oxidation excellent poor– poor– fair– fair– fair– poor–good fair– good– good
stability good excellent excellent excellent excellent excellent excellent
rust good– good– poor– poor– poor– good– fair– fair– fair– poor–
protection excellent excellent excellent excellent excellent excellent excellent excellent excellent excellent
pumpability poor poor–fair good– fair– fair– fair– poor– good– good– good
(in centra- excellent excellent excellent good fair excellent excellent
58
lized
systems)
oil separation good to fair–good poor– good good– good– good– good– good– good–
good excellent excellent excellent excellent excellent excellent
appearance smooth smooth– smooth smooth smooth smooth smooth smooth smooth and smooth
and clear fibrous and and and and and and buttery and
buttery buttery buttery buttery buttery buttery buttery
other adhesive EP grades EP grades EP grades EP grades inherent EP grades EP grades
properties and available available available available EP and available available
cohesive reversible reversible antiwear
principal thread rolling general military multiservice multiservice multiservice multiservice multiservice high
uses lubricants contact uses for multi- automotive industrial automotive automotive automotive temperature
bearings economy service and and industrial and and (frequent
industrial industrial industrial relube)
a
Multiservice includes rolling contact bearings, plain bearings and others. Reversibility is the ability of a grease to return to its normal grease-like
consistency after temporary exposure to temperatures near or above its dropping point.
b
Reprinted with permission.
Vol. 2 LUBRICATION AND LUBRICANTS 59
and cannot tolerate contamination by lubricating oils and greases are further
examples (201). For these applications, solid lubricating films are applied to
the rubbing surfaces.
The wide range of solid lubricants can generally be classified as either
inorganic compounds or organic polymers (both commonly used in a bonded
coating on a matching substrate), chemical conversion coatings, and metal
films. Since solid-film lubricants often suffer from poor wear resistance and
inability to self-heal any breaks in the film, the search continues for improved
compositions.
8.1. Inorganic Compounds. The most important inorganic mater-
ials are layer – lattice solids in which the bonding between atoms in an indi-
vidual layer is by strong covalent or ionic forces and those between layers are
relatively weak van der Waal’s forces. Because of their high melting points,
high thermal stabilities, low evaporation rates, good radiation resistance,
and effective friction lowering ability, molybdenum disulfide (MoS2)
[1317-33-5], and graphite [7782-42-5] are the preferred choices in this
group. Hexagonal boron nitride and boric acid also provide excellent lubri-
cation (202). Among other layer – lattice solids that find occasional use are
tungsten disulfide (WS2) [12138-09-9], tungsten diselenide (WeS2) [12067-46-8],
niobium diselenide (NbSe2) [12034-77-4], calcium chloride [10108-64-2];
cadmium iodide (CdI) [7790-80-9], and graphite fluoride [11113-63-6]
(Table 18) (203).
Graphite is widely used as a dry powder or as a colloidal dispersion in
water, petroleum oil, castor oil, mineral spirits, or other solvents. The water dis-
persions are used for lubricating dies, tools, metalworking molds, oxygen equip-
ment, and wire drawing. Graphite dispersed in solvents is used for drawing,
extruding, and forming aluminum and magnesium, as a high temperature lubri-
cant for conveyors, and for a variety of industrial applications. Graphite alone is
ineffective in vacuum since adsorbed water normally plays a decisive role in its
lubricating ability. Its film-forming ability can be restored, however, by mixing
with cadmium oxide or MoS2 and most organic materials, so that graphite may
offer effective lubricating action when bonded to the surface with organics. Air
oxidation commonly sets a use limit of 5508C, and high friction may occur in
air with water desorption >1008C (205).
Molybdenum disulfide has increasingly supplanted graphite for three rea-
sons: consistent properties in rigid specifications, independence from need for
adsorbed vapors in providing lubrication, and superior load capacity (205).
Like graphite, MoS2 has a layer–lattice structure in which weak sulfur–sulfur
bonds allow easy sliding between each sulfur–molybdenum–sulfur layer.
Molybdenum disulfide covered by MIL-M-7866 is the most common lubricant
grade: It is purified from molybdenite ore and is essentially free of abrasive
constituents (203).
Petroleum oil and grease dispersions of MoS2 are used extensively in
automotive and truck chassis lubrication and in general industrial use. Dis-
persions are also made with 2-propanol, polyalkylene glycols, other syn-
thetic oils, and water for airframe lubrication, in wire drawing, and for
splines, fastenings, gears, and fittings. Above 4008C, the MoS2 is oxidized to
Table 18. Common Solid Lubricantsa
Acceptable usage temperature, 8C
Minimum Maximum Avg friction coefficient, f
In N2 or In N2 or In N2 or
Material In air vacuum In air vacuum In air vacuum Remarks
molybdenum 240 240 370 820 0.10–0.25 0.05–0.10 low f, carries high load, good overall
disulfide (MoS2) lubricant, Can promote metal
corrosion
polytetrafluoro- 70 –70 290 290 0.02–0.15 0.02–0.15 lowest f of solid lubricants, load
ethylene (PTFE) capacity moderate and decreases
at elevated temperature
fluoroethylene- 70 70 200 200 0.02–0.15 0.02 low f, lower load capacity than PTFE
propylene
copolymer (FEP)
graphite 240 540 unstable in 0.10–0.30 0.02–0.45 low f and high load capacity in air,
vacuum high f and wear in vacuum,
60
conducts electricity
niobium diselenide 370 1,320 0.12–0.40 0.07 low f, high load capacity, conducts
(NbSe2) electricity (in air or vaccum)
tungsten disulfide 240 240 430 820 0.10–0.20 f not as low as MoS2, temp capability
(WS2) in air a little higher
tungsten diselenide 370 1320 same as for Ws2
(WSe2)
lead sulfide (PbS) 480 0.10–0.30 very high load capacity, used primarily
as additive with other solid
lunricants
lead oxide (PbO) 650 0.10–0.30 same as for PbS
calcium fluoride– 430 430 820 820 0.10–0.25 can be used at higher temp than other
barium fluoride above 5408C solid lubricants, high f < 5408C
eutectic 0.25–0.40
(CaF2 –BaF2) above 5408C
antimony trioxide high load capacity, used as corrosion
(Sb2O2) inhibitor in MoS2 lubricants
a
See Ref. 203.
Vol. 2 LUBRICATION AND LUBRICANTS 61
62
radiation: Fair N/A N/A N/A N/A V.G. V.G. V.G.
operating temp.
air (high) 2608C 2608C 3718C 2608C 1768C 2048C 6498C þ10938C 4008C
air (low) 2208C 2208C 1578F 2208C 2208C 1858C 2408C 2408C 2208C
vacuum 104Pa 107Pa 107Pa N/A N/A N/A 103Pa 107Pa 107Pa
load capacityc
force test 2500 2500 2500 150 2500 2500
lb gage lb gage lb gage lb lb gage lb gage
Falex Falex Falex LFW-1 Falex Falex
wear lifec
load 1000 1000 1000 150 1000 1000 50 50
lb gage lb gage lb gage lb lb gage lb gage
test Falex Falex Falex LFW-1 Falex Falex Falex Falex
time 60 min 450 min 60 min 120,000 120 min 70 min 2 min 5 min
coefficient of friction <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1 <0.1
corrosion resistance G. V.G. F E G F
a
See Ref. 210.
b
X ¼ Compatible; L ¼ Low; N/A ¼ Not Applicable.
c
Falex tests are all designed for English units and are, therefore, reported as such.
Vol. 2 LUBRICATION AND LUBRICANTS 63
150–3008C, baked coatings are commonly used with thermosetting resins, eg,
phenolics, epoxies, alkyds, silicones, polyimides, and urethanes. Of these, the
MIL-L-8937 phenolic type is widely used (204).
Inorganic binders are used, usually with graphite or MoS2, for extreme
conditions such as high vacuum, liquid oxygen, radiation resistance, and
high temperatures (208). The most common binder systems are silicates, phos-
phates, and aluminates. Some silicon and titanate metalloorganics used for
high temperature binders become inorganic on curing. An emerging class of
ceramic bonded materials for aerospace applications use either graphite, a
CaF2BaF2 eutectic, or proprietary systems, often with a glass frit binder
that is fused into a continuous film (203,208). Plasma spray coating avoids
overheating damage to the substrate metal while achieving the melting
point of at least one component in high temperature film compositions
(203,205,210).
The solid lubricant/binder ratio is a principal performance factor. High
lubricant content usually gives minimum friction, while high binder content
tends to give better corrosion resistance, hardness, durability, and a glossy finish
(210). With commonly used MoS2 –graphite and organic resin binders, the
optimum lubricant/binder ratio usually is from 1:1–4:1. With inorganic bind-
ing agents, the ratio is from 4:1 to as high as 20:1 and increases with high
temperatures.
Substrate Properties. It is clear from equation 8 that higher hardness of
the substrate lowers friction. Wear rate of the film also is generally lower. Phos-
phate undercoats on steel considerably improve wear life of bonded coatings by
providing a porous surface that holds reserve lubricant. The same is true for sur-
faces that are vapor- or sandblasted prior to application of the solid-film lubri-
cant. A number of typical surface pretreatments are given in Table 20 to
prepare a surface for solid-film bonding (208).
Optimum surface roughness usually is 0.05–0.5 mm; a very smooth surface
contains very little lubricant within its depressions, whereas rough peaks pene-
trate the lubricant to promote wear. Improved corrosion resistance may be
when applied in a vacuum with AISI 440C stainless steel and with ceramics
(qv) such as boron carbide or aluminum oxide, which can be applied as
undercoats (217).
A number of metal films are used industrially. Copper and silver are elec-
troplated on the threads for bolt lubrication. Slurries of powders of nickel, cop-
per, lead, and silver are also used in commercial bolt lubricants. Tin, zinc,
copper, and silver coatings are used as lubricants in metalworking, where toxi-
city has virtually eliminated lead as a lubricating coating (218). Gold and silver
find limited use on more expensive workpiece materials such as titanium. Silver
films are useful in a variety of other sliding and rolling contacts, in vacuum and
at high temperatures, since silver forms no alloys with steels and is soft at high
temperatures.
Under severe conditions and at high temperatures, noble metal films may
fail by oxidation of the substrate base metal through pores in the film. Improved
life may be achieved by first imposing a harder noble metal film, eg, rhodium or
platinum–iridium, on the substrate metal. For maximum adhesion, the metal
of the intermediate film should alloy both with the substrate metal and the
soft noble-metal lubricating film. This sometimes requires more than one
intermediate layer. For example, silver does not alloy to steel and tends to lack
adhesion. A flash of hard nickel bonds well to the steel, but the nickel tends to
oxidize and should be coated with rhodium before applying silver of 1–5-mm
thickness. This triplex film then provides better adhesion and greatly increased
corrosion protection.
17.0
Turbine oil
5.0
1.8
1.6
1.4
Viscosity, mPa·s(=cP)
Mercury
1.2
1.0
0.8
0.6
Rubidium
0.4
Air
0
0 100 200 300 400 500 600 700
Temperature, °C
and fuel handling devices, and, of course, are exposed to varying amounts of
ionizing (224).
Degree of damage suffered by a lubricant depends primarily on the total
radioactive energy absorbed, whether it is from neutron bombardment or from
gamma radiation. The common energy unit for absorbed dosage, the gray (Gy),
is equal to 105 J (100 ergs) absorbed per gram of material, or 0.01 Gy ¼ 1 rad.
The first changes observed with petroleum oils (at 104 Gy dosage) is evolution
of hydrogen and light hydrocarbon gas as fragments from the original molecule.
68 LUBRICATION AND LUBRICANTS Vol. 2
Lubricant demand in 2003, without marine oils, was 36 million metric tons.
Lubricating oil base stock production in North America was 71 million bar-
rels in 2003. These numbers are broken down by region and product type in
Table 23 (231,232).
The sales of automotive oils and greases have been fairly level in the period
1998–2002. Sales of industrial oils have declined in this period. The number of
base stock refineries in the United States and Canada has fallen from 40 in
1989–23 in 2004. United State and Canadian base stock capacity has dropped
from 250,000 barrels/day in 2001 to 217,000 barrels/day in 2004. Production of
groups II and III base stocks has grown over the past 10 years, however, from
4 to 42% (231).
Used oil
(automotive or industrial)
Reclaim
Heat: 40–90°C
Gravity settle Solids
Surfactants Water
Screening Sludge
Centrifuge
Reprocess
Rerefine
Dehydration
Chemical test Water
Distillation/clay 340°C Sludge
Filtration Bottoms
Spent clay
Older rerefining units used 2–5 kg/L of activated clay at 40–708C and
higher temperatures in place of TFE to clean the oil (238). More elaborate che-
mical and hydrotreating of used engine oils without a distillation step has been
developed by Phillips Petroleum for processing 40,000 m3/year (10 106 gal/
year). Establishment of a reliable feedstock supply is a critical consideration
for larger rerefining plants.
11.5. Toxic and Hazardous Constituents. Questionable constituents
of lubricating oils are polycyclic aromatics in the base oil plus various additives
(239). Of refining steps used in preparing lubricating oil base stocks from toxic
distillates, only effective solvent extraction, severe hydrogenation, or exhaus-
tive fuming sulfuric acid treatment appear adequate to eliminate carcinogeni-
city. Mild hydrofinishing or light solvent extraction reduces carcinogenic
potential, but does not necessarily eliminate it. Increasing the severity of
hydrofinishing can eliminate carcinogenic compounds. Group IV base stocks,
polyalphaolefins, are not expected to be carcinogenic since no polycyclic aro-
matics are present.
Most additives for lubricants present little risk, but the following involve
significant hazards: lead compounds, phenyl 2-naphthylamine, sodium nitrite
plus amines, tricresylphosphate high in the o-cresol isomer, and chlorinated
naphthalenes. A number of sulfur compounds used as additives cause skin irri-
tation; however, properly refined base oils containing usual concentrations of
these additives have a low degree of toxicity. Used motor oil has displayed
increased carcinogenic activity over its new counterpart (234,239). Users should
also avoid contact with lubricants, metalworking oils, and quench oils that are
highly degraded, were in service at extremely high temperatures, or are contami-
nated with toxic metals or bacteria.
The latest government regulations set forth under the Toxic Substan-
ces Control Act and in Public Health Service publications should be
checked before formulating new lubricants. Users of lubricants should request
Material Safety Data Sheets for each substance involved plus certification of
compliance from vendors. Lubricant compounders should insist on similar
information from their suppliers for any additive packages. Manufacturers of
both additives and lubricants commonly make toxicity checks on commercial
products.
11.6. Food Processing. Prior to September, 1998, the U.S. Depart-
ment of Agriculture (USDA), Food Safety Inspection Service (FSIS),
working with the Food and Drug Administration (FDA) maintained a system
to assure minimal risk to the consumer from incidental or unintended
contact with lubricants used in the food and beverage industries (see FOOD
PROCESSING) (240).
Upon satisfactory determination of nontoxicity, the USDA issued one of
three ratings (241): H1 Lubricants: Food-grade lubricants used in the food pro-
cessing environment, where there is a possibility of incidental food contact, eg, as
by splashing or dripping from machinery above an edible product. H2 Lubri-
cants: Nonfood-grade lubricants used on equipment and machine parts in loca-
tions where there is no possibility of food contact, eg, as in sealed gear boxes
or machinery below a product line. H3 Lubricants: Food-grade lubricants,
74 LUBRICATION AND LUBRICANTS Vol. 2
typically edible oils, used to prevent rust on hooks, trolleys and similar equip-
ment. These classes include a number of petroleum and nonpetroleum oils and
greases.
In September, 1998, the USDA issued the Hazard Analysis and Critical
Control Point (HACCP) protocol and eliminated the approval program. This
action shifted the burden of assessing the risk and approval of lubricants to
the food processor or manufacturer. Under the HACCP system, food processing
facilities must show the FSIS inspectors ‘‘documentation substantiating the
safety of a chemical’s (lubricant’s) use in a food processing environment’’
(240,241).
In order to minimize disclosure of proprietary formulation details, lubricant
manufacturers, with the approval of food processors, equipment builders and the
USDA, have developed a system whereby third party certifiers assume the role
previously handled by USDA. This includes preauthorization of nonfood com-
pounds, screening of product formulations, product registration, and the issu-
ance of approval documents (242,243).
The rating system used by USDA for lubricants will continue to be used.
Food-grade lubricants rated H1 or H3 for incidental contact must be generally
regarded as safe (GRAS); used in accordance with the provisions of a prior sanc-
tion or approval; or listed under 21 CFR 178.3570. Lubricants rated H2 for no
allowable contact must not contain poisons, toxins, or other materials that can
cause a health risk.
BIBLIOGRAPHY
‘‘Lubrication and Lubricants’’ in ECT 1st ed., Vol. 8, pp. 495–540, by R. G. Larsen and A.
Bondi, Shell Development center; in ECT 2nd ed., Vol. 12, pp. 557–616, by R. E. Lee, Jr.,
and E. R. Booser, General Electric Co.; in ECT 3rd ed., Vol. 14, pp. 477–526, by E. R.
Booser, General Electric Co.; in ECT 4th ed., Vol. 15, pp. 463–517, by E. R. Booser,
Consultant; ‘‘Lubrication and Lubricants’’ in ECT (online), posting date: December 4,
2000, by E. R. Booser, Consultant.
CITED PUBLICATIONS
31. ASTM D 2983 Standard Test Method for Low-Temperature Viscosity of Automotive
Fluid Lubricants Measured by Brookfield Viscometer, in Annual Book of ASTM
Standards, Section Five, ASTM International, Issued Annually.
32. ASTM D 5293 Standard Test Method for Apparent Viscosity of Engine Oils between
5 and 308C using the Cold-Cranking Simulator, in Annual Book of ASTM Stan-
dards, Section Five, ASTM International, Issued Annually.
33. ASTM D 4684 Standard Test Method for Determination of Yield Stress and Appar-
ent Viscosity of Engine Oils at Low Temperature, in Annual Book of ASTM Stan-
dards, Section Five, ASTM International, Issued Annually.
34. ASTM D 4683 Standard Test Method for Measuring Viscosity at High Shear Rate
and High Temperature by Tapered Bearing Simulator, in Annual Book of ASTM
Standards, Section Five, ASTM International, Issued Annually.
35. ASTM D 445 Standard Test Method for Kinematic Viscosity of Transparent and
Opaque Liquids (the Calculation of Dynamic Viscosity), in Annual Book of ASTM
Standards, Section Five, ASTM International, Issued Annually.
36. ASTM D 2161-04, Standard Practice for Conversion of Kinematic Viscosity to
Saybolt Universal Viscosity or Saybolt Furol Viscosity, in Annual Book of ASTM
Standards, Section Five, ASTM International, Issued Annually.
37. ASTM D 341 Standard Viscosity-Temperature Charts for Liquid Petroleum
Products, in Annual Book of ASTM Standards, Section Five, ASTM International,
Issued Annually.
38. E. E. Klaus and E. J. Tewksbury, in E. R. Booser, ed., Handbook of Lubrication:
Theory and Practice of Tribology, Vol. II, CRC Press, Boca Raton, Fla., 1984,
pp. 229–254.
39. ASTM D 2270, in E. R. Booser, ed., Handbook of Lubrication: Theory and Practice of
Tribology, Vol. II, CRC Press, Boca Raton, Fla., 1984.
40. ASME Pressure-Viscosity Report, American Society of Mechanical Engineers,
New York, 1953.
41. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, 1984, p. 19.
42. R. S. Fein, ASM Handbook, Vol. 18, ASM International, Metals Park, Ohio, 1992,
pp. 81–88.
43. D. D. Fuller, in E. A. Avallone, and T. B. Baumeister III, eds., Marks’ Standard
Handbook for Mechanical Engineers, 9th ed., McGraw-Hill, New York, 1987,
pp. 8-114–8-131.
44. M. Rhode, Fluid Film Lubrication: A Century of Progress, The American Society of
Mechanical Engineers, 1983 (Preface).
45. D. Dowson, in M. Rhode, ed., Fluid Film Lubrication: A Century of Progress, The
American Society of Mechanical Engineers, 1983, pp. 1–51.
46. O. Reynolds, in M. Rhode, Fluid Film Lubrication: A Century of Progress, The
American Society of Mechanical Engineers, 1983, pp. 135–217.
47. N. Petrov, in M. Rhode, Fluid Film Lubrication: A Century of Progress, The
American Society of Mechanical Engineers, 1983, pp. 107–133.
48. B. Tower, in M. Rhode, Fluid Film Lubrication: A Century of Progress, The American
Society of Mechanical Engineers, 1983, pp. 63–93.
49. V. L. Streeter, Fluid Mechanics, McGraw-Hill, New York, 1958, pp. 194–197.
50. A. Harnoy, Bearing Design in Machinery: Engineering Tribology and Lubrication,
Marcel Dekker, New York, 2003.
51. J. Boyd and A. A. Raimondi, in J. J. O’Connor, J. Boyd, and E. A. Avallone, eds.,
Standard Handbook of Lubrication Engineering, McGraw-Hill, New York, 1968,
Chapt. 3, pp. 3-1–3-10.
Vol. 2 LUBRICATION AND LUBRICANTS 77
107. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, pp. 69–77.
108. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, pp. 77–80.
109. R. A. McDonald, in L. R. Rudnick, Lubricant Additives: Chemistry and Applications,
Marcel Dekker, New York, 2003, pp. 29–43.
110. K. Nice, How Catalytic Converters Work, http://auto.howstuffworks.com.
111. W. D. Phillips, ‘‘Ashless Phosphorus-Containing Lubricating Oil Additives’’ in L. R.
Rudnick, ed., Lubricant Additives: Chemistry and Applications, Marcel Dekker,
New York, 2003, pp. 45–111.
112. L. O. Farng, ‘‘Ashless Antiwear and Extreme-Pressure Additives’’ in L. R. Rudnick,
ed., Lubricant Additives: Chemistry and Applications, Marcel Dekker, New York,
2003, pp. 223–257.
113. H. P. Bloch, Practical Lubrication for Industrial Facilities, The Fairmont Press,
Lilburn, Ga. and Marcel Dekker, New York, London, 2000, p. 20.
114. H. E. Dean, Handbook of Lubrication: Theory and Practice of Tribology, Vol. I, CRC
Press, Boca Raton, Fla., 1983, pp. 45–53.
115. R. F. Watts, 10th International Colloquium on Tribology, Technical Academy Esslin-
gen, Ostfilern, Germany, 1996.
116. D. Kenbeek and T. F. Buenemann, in L. R. Rudnick, Lubricant Additives: Chemistry
and Applications, Marcel Dekker, New York, 2003, pp. 203–221.
117. B. G. Kinker, in L. R. Rudnick, Lubricant Additives: Chemistry and Applications,
Marcel Dekker, New York, 2003, pp. 329–353.
118. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, pp. 63–64.
119. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, 1984, pp. 193–195.
120. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, pp. 65–68.
121. L. I. Osipow, Surface Chemistry, Rheinhold Publishing Corporation, New York,
1962, pp. 344–376.
122. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, 1984, pp. 210–211.
123. E. A. Bardasz and G. D. Lamb, in L. R. Rudnick, Lubricant Additives: Chemistry and
Applications, Marcel Dekker, New York, 2003, pp. 387–428.
124. F. L. Lee, and J. J. W. Harris, L. R. Rudnick, Lubricant Additives: Chemistry and
Applications, Marcel Dekker, New York, 2003, pp. 557–586.
125. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, 1984, p. 212.
126. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, 1984, pp. 215–216.
127. API 1509, API Engine Oil Licensing and Certification System, 15th ed., April 2002.
128. American Petroleum Institute, Engine Oil Licensing and Certification System,
http://api-ep.api.org.
129. SAE J300 Engine Oil Viscosity Classification, May 2004, in SAE Handbook, Society
of Automotive Engineers, Warrendale, Pa., Issued Annually.
130. International Lubricant Standardization and Approval Committee, ILSAC GF-4
Standard for Passenger Car Engine Oils, January 14, 2004.
131. International Lubricant Standardization and Approval Committee, ILSAC GF-3
Standard for Passenger Car Engine Oils, July, 2001.
132. Southwest Research Institute, Abstracts of Test Procedures Performed By Fuels and
Lubricants Research Division, www.swri.org.
80 LUBRICATION AND LUBRICANTS Vol. 2
133. ASTM D6593-04 Standard Test Method for Evaluation of Automotive Engine Oils
for Inhibition of Deposit Formation in a Spark-Ignition Internal Combustion Engine
Fueled with Gasoline and Operated Under Low-Temperature, Light-Duty Condi-
tions, in Annual Book of ASTM Standards, Section 5, ASTM International, West
Conshohocken, Pa., Issued Annually.
134. ASTM D6891-03 Standard Test Method for Evaluation of Automotive Engine Oils in
the Sequence IVA Spark-Ignition Engine, in Annual Book of ASTM Standards,
Section 5, ASTM International, West Conshohocken, Pa., Issued Annually.
135. ASTM D6709-03a Standard Test Method for Evaluation of Automotive Engine Oils
in the Sequence VIII Spark-Ignition Engine (CLR Oil Test Engine), in Annual Book
of ASTM Standards, Section 5, ASTM International, West Conshohocken, Pa.,
Issued Annually.
136. ASTM D6837-04 Standard Test Method for Measurement of the Effects of Automotive
Engine Oils on the Fuel Economy of Passenger Cars and Light-Duty Trucks in the
Sequence VIB Spark Ignition Engine, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
137. ASTM D4951-02 Standard Test Method for Determination of Additive Elements in
Lubricating Oils by Inductively Coupled Plasma Atomic Emission Spectrometry, in
Annual Book of ASTM Standards, Section 5, ASTM International, West Consho-
hocken, Pa., Issued Annually.
138. ASTM D2622-03 Standard Test Method for Sulfur in Petroleum Products by
Wavelength Dispersive X-ray Fluorescence Spectrometry, in Annual Book of
ASTM Standards, Section 5, ASTM International, West Conshohocken, Pa., Issued
Annually.
139. ASTM D5800-04 Standard Test Method for Evaporation Loss of Lubricating Oils by
the Noack Method, in Annual Book of ASTM Standards, Section 5, ASTM Interna-
tional, West Conshohocken, Pa., Issued Annually.
140. ASTM D6417-03 Standard Test Method for Estimation of Engine Oil Volatility by
Capillary Gas Chromatography, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
141. ASTM D6335-03b Standard Test Method for Determination of High Temperature
Deposits by Thermo-Oxidation Engine Oil Simulation Test, in Annual Book of
ASTM Standards, Section 5, ASTM International, West Conshohocken Pa., Issued
Annually.
142. ASTM D6794-02 Standard Test Method for Measuring the Effect on Filterability of
Engine Oils After Treatment with Various Amounts of Water and a Long (6-h) Heat-
ing Time, in Annual Book of ASTM Standards, Section 5, ASTM International, West
Conshohocken, Pa., Issued Annually.
143. ASTM D892-03 Standard Test Method for Foaming Characteristics of Lubricating
Oils, in Annual Book of ASTM Standards, Section 5, ASTM International, West Con-
shohocken, Pa., Issued Annually.
144. ASTM D6557-04 Standard Test Method for Evaluation of Rust Preventive Charac-
teristics of Automotive Engine Oils, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
145. ASTM D92-02b Standard Test Method for Flash and Fire Points by Cleveland Open
Cup Tester, in Annual Book of ASTM Standards, Section 5, ASTM International,
West Conshohocken, Pa., Issued Annually.
146. CRC Deposit Rating Manual, Manual 20, Cordinating Research Council, Alpharetta,
Ga., 2002.
147. CRC Distress Rating Manual, Manual 21, Cordinating Research Council, Alpharetta,
Ga., 2002.
148. D. McFall, Lubes-n-Greases 10, 2 (2004).
Vol. 2 LUBRICATION AND LUBRICANTS 81
149. J. A. McGeehan, Diesel Engines Have a Future and That Future is Clean, SAE
Paper 2004-01-1956, Engine Lubricants, Effects of Fuels & Lubricants on Automo-
tive Devices, and Lubricant Applications & New Test Methods, SAE SP-1885,
2004.
150. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, Chapt. 6, 219–262.
151. ASTM D4485-04, in Annual Book of ASTM Standards, Section 5, ASTM Interna-
tional, West Conshohocken, Pa., Issued Annually.
152. D. McFall, Lubes-n-Greases 10(7), 6 (2004).
153. D. McFall, Lubes-n-Greases, 10(8), 31 (2004).
154. MIL-PRF-7808L Lubricating Oil, Aircraft Turbine Engine, Synthetic Base, NATO
Code Number O-148.
155. MIL-PRF-23699F Lubricating Oil, Aircraft Turbine Engine, Synthetic Base, NATO
Code Number O-156, 1997.
156. A. G. Papay, Lub. Eng. 46(8), 511 (1989).
157. Trends 2003, Infineum USA, L.P.
158. P. A. Willermet, Lub. Eng. 55(2), 39 (1999).
159. T. Matsuoko, et al., Development of Toroidal Traction Drive CVTF for Automobile,
SAE Paper 2002-01-1696, International Spring Fuels & Lubricants Meeting & Exhi-
bition, Reno, Nev., May 6–9, 2002.
160. Lubrizol Ready Reference for Lubricant and Fuel Performance Driveline Lubricants
– Automotive Gear Lubricants, www.lubrizol.com 2005.
161. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, Chapt. 7, pp. 263–277.
162. Southwest Research Institute Gear oil Testing, www.swri.org May 2004.
163. ASTM D130-04 Standard Test Method for Corrosiveness to Copper from Petroleum
Products by Copper Strip Test, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
164. ASTM D892-03 Standard Test Method for Foaming Characteristics of Lubricating
Oils, in Annual Book of ASTM Standards, Section 5, ASTM International, West
Conshohocken, Pa., Issued Annually.
165. SAE J306 Automotive Gear Lubricant Viscosity Classification Jul 1998 in SAE
Handbook: Society of Automotive Engineers, Warrendale, Pa., Issued Annually.
166. ASTM D 5662-99 Standard Test Method for Determining Automotive Gear Oil Com-
patibility with Typical Oil Seal Elastomers, in Annual Book of ASTM Standards,
Section 5, ASTM International, West Conshohocken, Pa., Issued Annually.
167. ASTM D5579-04 Standard Test Method for Evaluating the Thermal Stability of
Manual Transmission Lubricants in a Cyclic Durability test, in Annual Book of
ASTM Standards, Section 5, ASTM International, West Conshohocken, Pa., Issued
Annually.
168. ASTM D 4498-95(2003)e1 Standard Test Method for Evaluating Wear Characteris-
tics of Tractor Hydraulic Fluids, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
169. ASTM D 2422-97 (2002) Standard Classification of Industrial Fluid Lubricants by
Viscosity System, in Annual Book of ASTM Standards, Section 5, ASTM Interna-
tional, West Conshohocken, Pa., Issued Annually.
170. ISO Standard 3448 Industrial Liquid Lubricants - ISO Viscosity Classification (1992).
171. ASTM D 6080 Standard Practice for Defining the Viscosity Characteristics of
Hydraulic Fluids, in Annual Book of ASTM Standards, Section 5, ASTM Interna-
tional, West Conshohocken, Pa., Issued Annually.
172. ANSI/AGMA 9005–E02, Industrial Gear Lubrication, American Gear Manufac-
turers Association, Alexandria, Va.
82 LUBRICATION AND LUBRICANTS Vol. 2
173. Exxon Application Data Sheet DG-5D, Lubrication of Plain Bearings, Exxon
Company, Houston, Tex., 1997.
174. Exxon Application Data Sheet DG-5G, Lubrication of Enclosed Gears, Exxon
Company, Houston, Tex., 1990.
175. Exxon Application Data Sheet DG-5D, Lubrication of Rolling-Contact Bearings,
Exxon Company, Houston, Tex., 1999.
176. H. P. Bloch, Practical Lubrication for Industrial Facilities, The Fairmont Press,
Lilburn, Ga. and Marcel Dekker, New York, London, 2000, pp. 281–394.
177. I. Macpherson and D. McCoy, Hydraulic and Turbine Oil Specifications, in E. R. Booser,
ed., Tribology Data Handbook, CRC Press, Boca Raton, Fla., 1997, pp. 274–281.
178. V. Bajpai, M. Fletschinger, N. Hoeck, and P. Rohrbach, Additive Technology for
Formulating Modern Turbine Oils, Ciba Specialty Chemicals, Tarrytown, N.Y.,
2004.
179. ASTM D 943-04 Standard Test Method for Oxidation Characteristics of Inhibited
Mineral Oils, in Annual Book of ASTM Standards, Section 5, ASTM International,
West Conshohocken, Pa., Issued Annually.
180. ASTM D 2272-02 Standard Test Method for Oxidation Stability of Steam Turbine
Oils by Rotating Pressure Vessel, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
181. ASTM D 5182-97(2002)e1 Standard Test Method for Evaluating the Scuffing Load
Capacity of Oils (FZG Visual Method), in Annual Book of ASTM Standards, Section
5, ASTM International, West Conshohocken, Pa., Issued Annually.
182. G. J. Folz and H. J. Noble, Metal Removal: Fluid Selection and Application Guide,
in E. R. Booser, ed., Tribology Data Handbook, CRC Press, Boca Raton, Fla., 1997,
pp. 831–839.
183. J. D. Silliman and R. Perich, eds., Cutting and Grinding Fluids: Selection and
Application, 2nd ed., Society of Manufacturing Engineers, Dearborn Mich., 1992.
184. T. G. McClure, in E. R. Booser, ed., Tribology Data Handbook, CRC Press, Boca
Raton, Fla., 1997, pp. 840–853.
185. N. M. Canter, in E. R. Booser, ed., Tribology Data Handbook, CRC Press, Boca
Raton, Fla., 1997, pp. 862–871.
186. A. C. Eachus, in E. R. Booser, ed., Tribology Data Handbook, CRC Press, Boca
Raton, Fla., 1997, pp. 854–861.
187. F. J. Passman, Controlling Microbial Contamination in Metalworking Fluids, SME
Paper MF92-127, Society of Manufacturing Engineers, Dearborn, Mich., www.biode-
terioration-control.com 1992.
188. T. J. Drozda and C. Wick, eds., Tool and Manufacturing Engineers Handbook,
Vol. 1 Machining, Society of Manufacturing Engineers, Dearborn, Mich., 1983,
pp. 4-1–4-60.
189. L. B. Sargent Jr., in Handbook of Lubrication: Theory and Practice of Tribology,
Vol. 1, CRC Press, Boca Raton, Fla., 1983, pp. 321–339.
190. J. O. McLean, M. C. Hawkins, and R. H. Hendrick, Lubrication in the Non-
ferrous Industries; in Standard Handbook of Lubrication Engineering, J. J.
O’Connor, J. Boyd, and E. A. Avallone, eds., McGraw-Hill, New York, 1968,
pp. 42-1 – 42-14.
191. C. Wick, J. T. Benedict, and R. F. Veilleux, Tool and Manufacturing Engineers
Handbook, Vol. 2 Forming, Society of Manufacturing Engineers, Dearborn, Mich.,
1984, pp. 3-1–3-22.
192. W. J. Landry, Use of Normal Paraffins in Cold Rolling of Aluminum, Technigram
D239-J2, Exxon Company U.S.A., 1995.
193. Lubricating Grease Guide, 3rd ed., National Lubricating Grease Institute (NLGI),
Kansas City, Mo., 1994.
Vol. 2 LUBRICATION AND LUBRICANTS 83
194. R. F. Haycock and J. E. Hillier, Automotive Lubricants Reference Book, 2nd ed., SAE
International, Warrendale, Pa., 2004, pp. 295–313.
195. ASTM D 217-02 Standard Test Methods for Cone Penetration of Lubricating Grease,
in Annual Book of ASTM Standards, Section 5, ASTM International, West Consho-
hocken, Pa., Issued Annually.
196. Exxon Booklet DG-400, Encyclopedia for the User of Petroleum Products, Exxon
Company, Houston, Tex., 1993.
197. ASTM D566-02, Standard Test Method for Dropping Point of Grease, in Annual
Book of ASTM Standards, Section 5, ASTM International, West Conshohocken,
Pa., Issued Annually.
198. ASTM D2265-00, Standard Test Method for Dropping Point of Lubricating Grease
Over Wide Temperature Range, in Annual Book of ASTM Standards, Section 5,
ASTM International, West Conshohocken, Pa., Issued Annually.
199. C. J. Boner, Manufacture and Application of Lubricating Greases, Rheinhold Pub-
lishing Corporation, New York, 1954, p. 397.
200. ASTM D4950-01(2004)e1 Standard Classification and Specification of Automotive
Service Greases, in Annual Book of ASTM Standards, Section 5, ASTM Interna-
tional, West Conshohocken, Pa., Issued Annually.
201. D. Klamann, Lubricants and Related Products, Verlag Chemie, Weinheim,
Germany, Deerfield Beach, GL, Basel, 1984, pp. 154–169.
202. A. Erdemir, Solid Lubricants and Self-Lubricating Films, in B. Bhushan, ed.,
Modern Tribology Handbook, Vol. 2, pp. 787–825 (2001).
203. L. C. Lipp, Lub. Eng. 32(11), 574 (1976).
204. J. K. Lancaster, in Handbook of Lubrication: Theory and Practice of Tribology,
Vol. II, CRC Press, Boca Raton, Fla., 1983, 269–290.
205. H. F. Sliney, ASM Handbook, Vol. 18, ASM International, Metals Park, Ohio, 1992,
pp. 113–122.
206. A. Erdemir and co-workers, Lub. Eng. 47(3), 179 (1991).
207. R. M. Gresham, in E. R. Booser, ed., Tribology Data Handbook, CRC Press, Boca
Raton, Fla., 1997, pp. 600–607.
208. E. R. Booser, ed., Handbook of Lubrication: Theory and Practice of Tribology,
Vol. III, CRC Press, Boca Raton, Fla., 1983, pp. 167–181.
209. W. E. Campbell, in F. F. Ling, E. E. Klaus, and R. S. Fein, eds., Boundary Lubrica-
tion: An Appraisal of World Literature, American Society of Mechanical Engineers,
New York, 1969.
210. R. M. Gresham, Lub. Eng. 44(2), 143 (1988).
211. C. DellaCorte, Tribol. Trans. 46(3), 361 (2003).
212. K. Carnes, Tribol. Lub. Technol. 59(11), 14 (2004).
213. A. Erdemir and C. Donnet, ‘‘Tribology of Diamond, Diamond-Like Carbon, and
Related Films’’ in B. Bhushan, ed., Modern Tribology Handbook, Vol. 2, CRC Press
LLC, Boca Raton, Fla., 2001, pp. 871–908.
214. M. F. Alzoubi, O. O. Ajayi, J. B. Woodford, A. Erdemir, and G. R. Fenske, Tribol.
Trans. 44(4), 591 (2001).
215. Near-Frictionless Coating Nears Commercial Applications, www.spacedaily.com,
Argonne, Sept. 3, 2002.
216. C. Donnet and A. Erdemir, Tribol. Lett. 17(3), 389 (2004).
217. D. H. Buckley and R. L. Johnson, ASLE Trans. 6, 1 (1963).
218. J. A. Schey, Tribology in Metalworking: Friction, Lubrication and Wear, ASM Inter-
national, Metals Park, Ohio, 1983.
219. D. F. Wilcock, in E. R. Booser, ed., Handbook of Lubrication: Theory and Practice of
Tribology, Vol. II, CRC Press, Boca Raton, Fla., 1983, pp. 291–300.
220. W. A. Gross, ed., Fluid Film Lubrication, John Wiley & Sons, Inc., New York, 1980.
84 LUBRICATION AND LUBRICANTS Vol. 2
221. M. M. Khonsari, L. A. Matsch, and W. Shapiro, Gas Bearings in GR 14, Vol. III,
pp. 553–575.
222. A. J. Baumgartner, in R. A. Burton, ed., Bearing and Seal Design in Nuclear Power
Machinery, American Society of Mechanical Engineers, New York, 1967.
223. J. Gerkema, ASLE Trans. 28(1), 47 (1985).
224. D. M. Pirro and A. A. Wessol, Lubrication Fundamentals, 2nd ed., Marcel Dekker,
New York, 2001.
225. J. G. Carroll and S. R. Calish, Lub. Eng. 13(7), 388 (1957).
226. E. R. Booser, in J. J. O’Conner, J. Boyd, and E. A. Avallone, eds., Standard Hand-
book of Lubrication Engineering, McGraw-Hill, New York, 1968, pp. 44-1–44-13.
227. R. O. Bolt, in E. R. Booser, ed., Handbook of Lubrication: Theory and Practice of
Tribology, Vol. I, CRC Press, Boca Raton, Fla., 1983, pp. 209–223.
228. J. E. Kelley, T. D. Roberts, and H. M. Harris, A Penetrometer for Measuring the
Absolute Viscosity of Glass, report 6358, U. S. Bureau of Mines, Washington, D.
C., 1964.
229. R. A. Reich, P. A. Stewart, J. Bohaychick, and J. A. Urbanski, Lub. Eng. 59(7),
16 (2003).
230. Solid and Liquid Lubricants for Extreme Environments ASLE Special Publication
SP-15, 1984.
231. 2004–2005 Lubricants Industry Sourcebook, LNG Publishing Company, Inc., Falls
Church, Va., 2004.
232. M. Khonsari and E. R. Booser, Mach. Lub., May 2004.
233. Toxic Substances Control Act, Public Law No. 469, 94th U. S. Congress Chemical
Substance Inventory, U. S. Environmental Protection Agency, Office of Toxic Sub-
stances, Washington, D. C., 1975.
234. J. M. Perez and D. J. Hoke, in E. R. Booser, ed., Tribology Data Handbook, CRC
Press, Boca Raton, Fla., 1997, pp. 1031–1045.
235. J. W. Swain Jr., in E. R. Booser, ed., Handbook of Lubrication: Theory and Practice
of Tribology, Vol. I, CRC Press, Boca Raton, Fla., 1983, pp. 533–549.
236. S. Napier and K. Rich, ‘‘Waste Treatability of Aqueous-Based Synthetic Metalwork-
ing Fluids’’, Lub. Eng. 41(6), 361 (1985).
237. J. W. Swain Jr., Lub. Eng. 39(9), 551 (1983).
238. D. W. Brinkman, Lub. Eng. 43(5), 324 (1987).
239. T. M. Warne and C. A. Halder, Lub. Eng. 42(2), 97 (1986).
240. M. J. Raab, Lub. Eng. 58(2), 16 (2002).
241. Draft Standard NSF 116-2001 Nonfood Compounds Used in Food Processing
Facilities - Food-Grade Lubricants, NSF International, Ann Arbor, Mich., 2001.
242. M. J. Raab, Food-Grade Lubricants: Current Issues presented to New York Section,
STLE, December, 2004.
243. K. Yano, Non-Food Compounds Registration: Past, Present and Future, NLGI Tech-
nical Paper 0308, Presented at 70th NLGI Annual Meeting, NLGI, Kansas City,
Mo., 2003.
JAMES R. DICKEY
Lubricants Consultant
Vol. 2 WAXES 85
WAXES
1. Introduction
Waxes are among the oldest worked materials used by humans. Their value as
versatile construction materials (‘‘man’s first plastic’’) was discovered very early.
Today, waxes are used mostly as additives and active substances. The use of
waxes is expected to increase in the future because of their generally favorable
toxicological and ecological properties.
The historic prototype of all waxes is beeswax. Since it could be obtained
with relatively little effort, it was popular in antiquity, and even now the
term wax is occasionally used in everyday speech as a synonym for beeswax.
However, the scientific and commercial definition of wax covers a much wider
area.
1.1. History of Waxes and Their Applications. Utilization of waxes
was probably begun in prehistoric times, but because of their transitory nature,
no definite archaeological evidence exists. Thus, the utilization of wax and
related substances in mummification and as protective coverings in ancient
Egypt (from ca. 3000 B.C.) represents the earliest scientific proof of the use of
waxes. Ancient written documents contain many indications that waxes found
many different applications. The most well known is the story of Daedalus and
Icarus, who used wax as an adhesive to make wings by attaching feathers to each
other. In antiquity, wax was used as a raw material for modeling, in the produc-
tion of casting molds, as a pigment carrier, and for surface protection.
In colonial times, hitherto unknown waxes, such as carnaúba, candelilla,
and Chinese insect wax, were introduced in Europe. From the occurrence of
wax, COLUMBUS inferred the riches of the Caribbean islands: ‘‘Where there is
wax, there are also thousands of other things’’ (1).
For a long time, not much was known about the chemical nature of wax.
Only in the 18th century did the discovery occur that beeswax is an animal
secretion and not a plant product collected by bees. Wax research was estab-
lished as a scientific discipline in 1823 (2). It became part of the new research
area of soaps, oils, fats, and waxes. The real breakthrough of wax as an impor-
tant raw material, in terms of quantity as well, occurred at the beginning of the
Industrial Revolution. Ozocerite (fossil wax) was mined and refined to give cer-
esin (1875), montan wax was obtained from Eocene lignite (1897), and paraffin
waxes were obtained from crude petroleum. In 1935 the first fully synthetic
waxes were produced by the Fischer–Tropsch process. Polyethylene wax
has been synthesized by the high pressure process since 1939, and became
available by the low-pressure Ziegler process after 1953. On a laboratory
scale polyolefin waxes can also be synthesized by using modern metallocene
catalysts.
Degradation processes for the production of waxes, which start from high
molar mass plastics (mainly polypropylene), have also achieved a certain degree
of importance. Finally, low molar mass substances, which would otherwise have
to be disposed of in plastics production, can be processed or refined to give indust-
rially utilizable waxes.
A large and still increasing number of natural and synthetic waxes and
related substances exists, as well as applications for these materials. No decrease
in overall demand is apparent, so waxes are likely to remain important
during the coming decades. The history of waxes has been described in great
detail in (3).
1.2. Definition. No generally accepted definition exists for the term wax.
All attempts to formulate a precise, comprehensive, and scientifically verifiable
definition of wax must take the large number of waxlike products and the chemi-
cal complexity of individual types into account. Nevertheless, use of the term wax
for different chemical species with common properties is still reasonable.
Typically waxes do not consist of a single chemical compound, but are often
very complex mixtures. Being oligomers or polymers in many cases, the compo-
nents differ in their molar mass, molar mass distribution, or in the degree of side-
chain branching. Functional groups (eg, carboxyl, alcohol, ester, keto, and amide
groups) can be detected in waxes, sometimes several different groups.
The academic definition still quoted in chemistry text books — that waxes
are esters of long-chain carboxylic acids with long-chain alcohols — is no longer
useful. It applies fairly well only to some classical waxes, such as beeswax; others
(eg, petroleum waxes) do not fall in this category. Today, physical and technical
definitions are preferred. Several attempts were made to differentiate between
waxes and other classes of substance, particularly fats, resins, and high molar
mass polymers, by using several criteria. These primarily physical definitions
are to some extent arbitrary and are not generalled accepted. Waxes can also
be classified according to their applications.
Probably the most conclusive definition has been drawn up in Europe by
the Deutsche Gesellschaft für Fettwissenschaft (DGF, German Association for
Fat Science). It was used in modified form in the customs tariff of the European
Union (4,5). According to this definition, waxes must have:
waxes are seldom used industrially in their original form. They are generally
converted into modified natural waxes by refining, eg, by distillation, or extrac-
tion. Chemical processes, such as hydrogenation, bleaching, and oxidation, can
also be used. Physical and chemical treatment can be combined; the objective
of all these processes is to obtain a wax that is as pure as possible.
Partially Synthetic Waxes. If natural waxes or waxlike materials are
modified by chemical reactions such as esterification, amidation, or neutraliza-
tion of acidic waxes (with aqueous alkali or alkaline-earth metal hydroxides),
partially synthetic waxes are obtained. These can be tailored to particular appli-
cations. In esterified waxes the relationship to natural waxes is still easily recog-
nizable. For example, natural montan waxes, which consist mainly of esters of
long-chain acids with long-chain alcohols, can be converted into acid waxes by
cleavage of the esters and oxidation of the alcohols. Subsequent reaction with
dihydric alcohols gives esters that have chemical structures comparable to
those of the starting materials. Amide waxes do not occur naturally to any signif-
icant extent. This group consists of reaction products of fatty acids with ammo-
nia, amines, and diamines. An industrially important amide wax is distearoyl
ethylenediamine.
Alcohol waxes (lanette waxes) also belong to the partially synthetic waxes.
They are mixtures of long-chain alcohols that are optimized for their main area of
use, in ointments and creams for pharmaceuticals and cosmetics, by the addition
of fatty acid esters and emulsifiers. Alcohol waxes can also be used for industrial
emulsions with good long-term stability.
Wool wax (lanolin) is obtained in crude form from sheep’s wool. It is refined
by chemical and adsorptive bleaching, and sometimes by subsequent hydro-
genation or separation of the acidic fractions after hydrolysis. Wool wax is a
mixture of esters, some being derived from branched or cyclic waxy or fatty
acids and alcohols.
Fully synthetic waxes were developed only in the 20th century. Low
molar mass compounds are used as starting materials. The products can
be waxes in the narrower sense of the definition (see Section 1.2), or sub-
stances with only partial wax character. The two main groups of fully syn-
thetic waxes are the Fischer – Tropsch and the polyolefin waxes. These
waxes can be classified according to the starting material used for production
(C1 or C2 chemistry).
C1 Building Blocks. The Fischer– Tropsch (FT) synthesis is described
elsewhere in detail. The oldest FT variant is the ARGE process (Arbeitsge-
meinschaft Ruhrchemie– Lurgi). Synthesis gas is obtained by reaction of
steam with carbon and is then converted catalytically into a broad spectrum
of saturated and unsaturated hydrocarbons. Among these products are the
Fischer– Tropsch waxes. They are characterized by low branching of the hydro-
carbon chain and a narrow molar mass distribution. They are relatively hard
products. The ARGE process has been used commercially by SASOL in the
Republic of South Africa.
Besides carbon, natural gas can also be used as a raw material for C1 syn-
theses. The Shell Middle Distillate Synthesis (SMDS) process utilizes syn-
thesis gas produced by steam reforming of natural gas (6). Fischer–Tropsch
wax fractions are produced and marketed at the new SMDS plant in Bintulu
Vol. 2 WAXES 89
When a suitable emulsifier is added, polar waxes can form fine stable dis-
persions in water. Mixed aqueous–organic emulsions can also be produced.
These mostly stable dispersions and emulsions are used for surface protection.
After application, the wax-containing phase dries and a protective, sometimes
even glossy, coating is formed.
In addition to surface protection, hydrophobing, and gloss-giving, other
functions of waxes include surface protection, hydrophobing, gloss-giving,
matting, mold releasing, lubricating, binding, imparting compatibility or flexibil-
ity, regulating viscosity, adjusting consistency, emulsifying, dispersing, and
adjusting drop points. Waxes are also used as combustible and illuminating
materials.
For special applications only a few – frequently only one – of these functions
are important. Some occasionally opposite effects are noteworthy: wax gives pol-
ish and gloss to various surfaces but has a matting effect as an additive in paints
and coatings. Waxes are used as binders in master batches but as release agents
in molding plastics. Table 1 gives some idea of the versatility of waxes but does
not claim to be complete. For the specific suitability of waxes to individual appli-
cations, monographs, company brochures, and prospectuses should be consulted.
of esters derived from acids and alcohols with an uneven number of carbon atoms
as well as bifunctional components such as diols, hydroxycarboxylic, and dicar-
boxylic acids. Natural waxes contain very small proportions of branched car-
boxylic acids, mainly those with o-methyl branching. Aromatic acids, such as
cinnamic, 4-hydroxycinnamic, and ferulic (4-hydroxy-3-methoxycinnamic)
acids, have also been found in some waxes. Because of the presence of bifunc-
tional components, natural waxes can contain not only the simple, classical
wax esters, but also oligoesters with two or more ester functions and other func-
tional groups in the molecule.
Free acids and alcohols are present in natural waxes in widely varying
quantities. They are generally identical to those present as ester components.
Homologous n-alkanes in the ca. C15 –C37 range, mainly those with an
uneven number of carbon atoms, are also components of natural waxes in propor-
tions of 1–50 %. Branched-chain and olefinic hydrocarbons are present in insig-
nificant quantities. Glycerides (fats), phytosterols, terpenes, resins, long-chain
carbonyl compounds, and flower pigments are also present in very small propor-
tions. Because of the complex mixtures involved and the relatively simple analy-
sis techniques of earlier years, the inconsistency of literature data on the
composition of natural waxes is understandable.
Modern analytical instruments are required for complete analysis of a nat-
ural wax. High-temperature capillary GC, possibly coupled to a mass spectro-
meter, is a suitable the best analytical method.
Beeswax and carnaúba and candelilla waxes have a total annual market
volume of ca. 28 000 t worldwide. These are economically the most important nat-
ural waxes (8).
Among the large number of other known waxes, ouricury, esparto, bamboo,
and Japan wax (actually a special vegetable tallow) are of regional or local impor-
tance. Their availability is estimated to be 3000–4000 t/a (8,9).
In all areas of application for waxes, recent natural waxes are now mostly
used together with modified fossil and synthetic waxes. Applications include pre-
parations for cleaning, polishing, and preserving floors, furniture, and car body-
work surfaces; the candle industry; craft work; production of pharmaceuticals,
cosmetics, and confectionery (observing appropriate legal approval); metal and
ceramic sintering technology; coating techniques; the paint and coating indus-
tries; and the production of carbon paper compounds. The latter, which was
important in terms of quantity, has been replaced almost completely by new
duplicating processes. Despite the market recession this has caused (10), an
annual growth of ca. 4 % is expected in the U.S. market for natural and synthetic
waxes (11).
Special, ion-free or low-ion natural wax raffinates have also been used in
electronics.
In his standard book The Chemistry and Technology of Waxes, A. J.
WARTH included a comprehensive chapter on all known recent natural waxes.
The second edition of the book was published in 1956 and is currently out of
print (12).
2.2. Vegetable Waxes. In contact with the environment, plants form a
compressed epidermis on their surfaces by depositing secondary layers of cellu-
lose, except in the case of inflorescence (13). This mechanical reinforcement is
Vol. 2 WAXES 93
The central ribs are often removed from the leaves before drying. The dried
leaves are then chopped and the chips separated by air classification to give a
mixture of ca. 60 % wax and 40 % chopped plant, which is stirred with water
and oxalic acid to form a paste. The latter is heated to its boiling point and
pressed through a filter while hot. Centrifugation of the filtrate gives virtually
anhydrous wax. The remaining wax is extracted from the dried filter cake
with aliphatic hydrocarbons (eg, heptane).
The crude product is normally supplied in lumps that are obtained by
breaking up the wax blocks into which the melt was cast. Carnaúba wax is
also sold now in other forms.
Types. Internationally recognized classifications and specifications for
genuine, unfalsified types of carnaúba wax were worked out by the American
Wax Importers and Refiners Association according to established methods of
determination (20).
The classification used in Brazil, shown in Table 2, has been valid world-
wide since the beginning of the 1980s.
The highest-value type, T 1, is a prime yellow wax purified by filtration. To
obtain T 3, suitable crude waxes (colored) are filtered and then bleached with
hydrogen peroxide. The T 4 wax is exported predominantly in the form of centri-
fuged or filtered variants.
Chemically refined or derivatized carnaúba waxes are no longer available
commercially because of lack of demand (8). Chemical refining processes, such
as heating with alkali in the presence of paraffins, gave carnaúba wax residues
in addition to the corresponding raffinates (21–27). Where these processes are
used, they account for <1 % of the total volume of carnaúba wax.
Properties. Carnaúba wax is one of the hardest and highest-melting nat-
ural waxes. Because of its fine crystalline structure, its fracture edges are smooth
*Values based on 100 % total acid, total alcohols, and total paraffins in each
case (deviations from 100 % ¼ sum of all other acids and alcohols).
**Approximate.
the fact that Type 1 is harvested from genetically younger leaves, whereas the
others come from older leaves that have been exposed to more sunshine.
Uses. T 4 Variants. The wide use of carnaúba wax in the production of
preserving and cleansing agents for floors (floor polishes, ionogenic and noniono-
genic day-bright emulsions, wiping waxes and cleaners, sealing emulsions, spray
cleaners, etc.); and polishes for furniture, cars, and shoes (solvent–water or
straight solvent products) is based on its ability to form pastes and glossy and
repolishable films and its ability to be dispersed. In this classical area a large pro-
portion of T 4 variants of carnaúba wax are used.
Type 1 carnaúba wax and some of Type 3 (after absorptive purification to
remove peroxide components arising from hydrogen peroxide bleaching) are
approved for use in the pharmaceutical industry (eg, as polishes for pills) and
the cosmetics industry (as lipsticks and lip salves).
In the food sector, T 1 and T 3 waxes are used as release agents for bakery
and confectionery products and as additives in chewing gum production.
In polymer processing, carnaúba wax is used in release agent preparations
and, to a small extent, as a lubricant. The varnish industry uses carnaúba wax as
an additive in the interior coating of food containers (8).
In the leather industry, carnaúba wax is used for cutting waxes, cleaning
waxes, and edging inks. It is also approved for coating citrus fruit.
Vol. 2 WAXES 97
mp 68.5–72.5 8C
Refractive index 1.4550–1.4611
d15
4 0.950–0.990
Acid number 12–22 mg KOH/g
Saponification number 43–65 mg KOH/g
Hydrocarbon content 30.6–45.6 %
Total acid content 20.6–29.0 %
Color amber
Flash point 235.4–248.4 8C
Components insoluble in xylene/toluene (1 : 1) 0.0–0.1 %
The data agree with those previously laid down by the American Wax
Importers and Refiners Association for prime crude candelilla wax (20).
The average composition of candelilla wax gives below is derived from the
large number of values found in the literature.
Candelilla wax differs fundamentally from carnaúba wax in its high hydro-
carbon content of 45% (max.) and resin content of 20%. Whereas carnaúba wax
contains >80 % wax according to the classical chemical definition, candelilla wax
consists of <35 % of these components.
As in all natural waxes, in candelilla wax the paraffins and wax acids or
alcohols with even-numbered carbon chains have typical and distinctive chain-
length distributions (Table 4).
Vol. 2 WAXES 99
Sugarcane wax is generally extracted from the dried filter cake (ca. 1 % based
on sugarcane) (12) formed during sugarcane processing by using alcohols, carbon
disulfide, carbon tetrachloride, or aromatic and aliphatic hydrocarbons (39).
The crude wax content of the filter cake (in weight percent) varies
between the following limits, depending on sugarcane origin and method of
processing:
India 8.0–18.0 wt %
Louisiana 4.4–17.9 wt %
Puerto Rico 12.0–14.0 wt %
Hawaii 9.6–11.0 wt %
Philippines 10.5–11.0 wt %
Cuba 12.4–22.0 wt %
Java 5.0–15.0 wt %
Republic of South Africa 6.9–14.6 wt %
Argentina 11.9–15.5 wt %
Brazil 8.9–17.8 wt %
Crude wax can be freed from most glycerides (fats and oils) by treatment
with acetone, methyl ethyl ketone, methanol, ethanol, 1-propanol and isopropa-
nol, 1-butanol, or ethyl acetate. In a subsequent recrystallization step, preferably
in isopropanol, resinous and colored components can be removed. Further treat-
ment with bleaching earths or oxidative bleaching of the wax melt with air gives
waxes ready for use in a variety of applications. Bleaching processes for prepur-
ified wax using potassium chlorate–sulfuric acid or chromic acid–sulfuric acid
would be effective but are not used for cost reasons.
Properties and Composition. Crude sugarcane wax is a black-brown to
green-brown semisolid with an unpleasant, rancid odor.
Semi- and completely purified products are brown- to yellow-colored hard
waxes, whose characteristics vary considerably. Melting points of 68–81 8C,
acid numbers of 7–22 mg KOH/g saponification numbers of 32–65 mg KOH/g
and proportions of unsaponifiable material of 52–62 % have been published in
the literature. The following values are typical for a semiraffinate (12):
In the past, many attempts have been made to produce crude sugar-
cane wax on an industrial scale. As early as 1841, isolation of the wax from
filter cakes was begun, and in 1909 a corresponding extraction process was
patented.
Around 1916, crude sugarcane wax was extracted as a commercial product
in South Africa and Java. In 1958 the wax was produced in two Cuban and one
Australian plant. However, in 1960, production in Australia was halted and only
one plant was still operating in Cuba (39). In India, an extraction plant is
thought to have been operating continuously since 1950 (39). This plant is
attached to a sugar refinery with a daily throughput of 1200 t of sugarcane.
Difficulties resulting from the necessary centralization of production loca-
tions, the necessarily rapid and complete drying of large quantities of filter
cake material (which has a tendency to ferment), and last, but not least, the
increasing costs of extraction have made sugarcane wax raffinate incapable of
competing with carnaúba wax on the world market in terms of price. Conse-
quently, sugarcane wax raffinates have only regional importance.
In producer countries, crude sugarcane wax can be used in floor and shoe
polishes, carbon paper production, impregnating agents, and production of wax
compounds.
Retamo Wax. Occurrence and Isolation. In arid areas of Argentina,
Bulnesia retama grows wild as trees and shrubs. Branches and twigs are coated
with a wax deposit that can be collected mechanically after drying the plant
material harvested in summer. The crude wax is purified by melting with dilute
sulfuric acid and subsequent filtration through cloths. It is sold exclusively on
the home market as retamo wax (12).
Properties and Composition. Retamo wax is a light to medium brown,
odorless, hard wax. It is very similar to other vegetable waxes, such as carnaúba
wax, with regard to its solubility in polar and nonpolar organic solvents.
Data found in the literature, such as mp (76–78 8C), acid number (ca.
49 mg KOH/g), saponification number (ca. 87 mg KOH/g), proportion of hydro-
carbons (15–25 %), and content of unsaponifiable components (38–42 %), char-
acterize the wax less precisely than the typical chain-length distributions of
even-numbered wax acids and alcohols and of paraffins (40).
Uses and Importance. With a decreasing annual production of several
hundred tonnes (1975, ca. 500 t/a), retamo wax is of limited importance on the
Argentinian market. It is used for floor, car, and shoe polishes and in the produc-
tion of colored polishing inks.
Jojoba Oil. According to the DGF definition, jojoba oil is not a wax. How-
ever, its chemical composition allows it to be classified as a ‘‘liquid wax’’ in the
sense of the classical definition of waxes.
Occurrence and Isolation. The jojoba bush (Simmondsia chinensis ¼
Simmondsia californica = Buxus chinensis) grows wild over an area of ca.
250 000 km2, most of which lies in the Sonora desert. The population density
varies from a few to 500 plants per hectare. In certain areas, millions of jojoba
bushes can be found.
The bushes carry fruit capsules that can contain one to three nuts of differ-
ing sizes. The yield from the harvest of wild jojoba bushes can range from a few
nuts to 14 kg of nuts per plant. The average yield is ca. 1.8 kg per bush.
Vol. 2 WAXES 103
The nuts contain 50–60 wt % oil, which can be extracted almost completely
from the deshelled, granulated fruit. Around 35–40 % of a particularly high-
value oil is obtained by cold pressing and a further 15–20 % of a lower-value
oil by extraction of the pressed cake. The oil is purified by simple filtration (41).
Properties and Composition. Jojoba oil is a virtually colorless to
golden yellow, odorless, unsaturated oil, for which the most important character-
istics are as follows:
mp 6.8–7.0 8C
Acid number 2 mg KOH/g
Saponification number 92 mg KOH/g
Iodine number 82 g I2/100 g
Unlike other vegetable waxes, the acids and alcohols in jojoba oil are mono-
unsaturated, straight-chain representatives of the chain-length range C16 –C24.
Among the acids the C18 (ca. 12.5 % of the total acids), C20 (ca. 67 % of the total),
and C22 components (ca. 14.5 % of the total) predominate, and among the alco-
hols, the C20 (ca. 48 % of the unsaponifiable components) and C22 components
(ca. 41 % of the unsaponifiable components). With this unusual composition,
jojoba oil is very similar to sperm oil contained in the skull a bone of the
sperm whale.
Uses and Importance. An extraordinarily high market demand exists
for a sperm oil substitute, which can be satisfied to a great extent by jojoba oil.
This demand is a result of the worldwide prohibition of whaling, even though the
latter is not obeyed by all countries.
The first experimental jojoba plantations were begun in the United States
(Arizona, Texas, and California) to effect infrastructural improvements. Many
other jojoba oil plantations were started in Mexico, Argentina, Australia, Brazil,
Africa, Egypt, Israel, and countries whose climatic conditions allowed desert
farming.
Because management of these plantations is difficult and the time required
to achieve economical harvests is relatively long (eight to ten years), jojoba oil
is available in insufficient quantities at a very high price. Its use is therefore
limited to cosmetics and pharmaceuticals. As soon as reliable quantities can be
supplied at acceptable prices, jojoba oil may attract extraordinarily great interest
as a raw material, particularly in the production of high-performance lubricants
and in modified form in plasticizers and stabilizers. Hydrogenated jojoba oil,
which corresponds to fatty esters such as stearyl stearate in its properties, will
remain noncompetitive with these esters in terms of price and will therefore not
achieve market importance.
104 WAXES Vol. 2
Other Vegetable Waxes. Many other vegetable waxes are known, and
the composition of some of them has been investigated (12). However, they are
either of very limited local importance or of purely academic interest. These
waxes include:
Caranday wax
Raffia wax
Columbia wax
Esparto wax
Alfalfa wax
Bamboo wax
Hemp wax
Douglas fir wax
Cork wax
Sisal wax
Flax wax
Cotton wax
Dammar wax
Cereal wax
Tea wax
Coffee wax
Rice wax
Ocatilla wax
Oleander wax
* Palm waxes
mp 62–65 8C
Acid number 17–24 mg KOH/g
Ester number 72–79 mg KOH/g
Ester number/acid number 3.3–4.2
Saponification turbidity point 65 8C
Extensive investigations show that properties vary even among waxes from the
same source (44).
Composition. The following values are given in the literature for the
approximate composition of beeswax with regard to wax classes:
Hydrocarbons 13–16 %
Fatty acids 30–35 %
Palmitic acid 15–20 %
Hydroxy fatty acids 13–17 %
Fatty alcohols 30–35 %
Other substances 4–8 %
Uses. Beeswax, mainly the yellow variety, is one of the raw materials
used in the production of household candles, expensive decorative candles, and
candles for religious purposes. For the latter, the Roman Catholic Church pre-
scribes a minimum beeswax content of 10 %. Beeswax is also used for artistic
and handicraft purposes (wax flowers, wax sculptures, painting, and batik).
Most bleached and purified beeswax is used in cosmetics and pharmaceuticals
(to regulate the consistency of lipsticks, creams, ointments, and suppositories)
and in the food sector (as a release agent and a polish in confectionery produc-
tion). Its use in these areas is favored by corresponding licensing regulations.
A significant proportion of beeswax, ca. 40 % of the total market volume, is
sold back to beekeepers for the production of synthetic honeycombs, in which
blends with paraffins and ceresin, which are not allowed for pharmaceuticals,
are often used.
Economic Aspects. From the worldwide honey production figures for
1976, production of 11 000–19 000 t of beeswax was estimated. Whether the
quantities used in the candle and beekeeping industries were included in the
worldwide consumption figure of ca. 10 000 t for 1981, is not known. At that
time, total imports for the Federal Republic of Germany amounted to 1000 t.
At present (1994), ca. 15 000 t of beeswax is available on the world market.
Ca. 6000 t is used in industry, and ca. 4000 t is sold back to beekeepers (8).
The annual availability of beeswax depends strongly on climatic vari-
ations in exporter countries and bee diseases that break out occasionally. In
1993, for example, a marked drop in exports from the United States occurred
because of catastrophic weather conditions in the Mississippi area (8).
Other Insect Waxes. The following waxes are not market products and
only in some cases have limited regional importance in the countries of origin.
Ghedda Wax. Ghedda wax is produced by some types of wild bee bound
in Asia (Apis indica, A. florea, A. dorsata). It was formerly exported as such but
more frequently now is blended with beeswax from the East Indies.
Ghedda wax has a yellow-brown color and varying characteristic data (eg,
acid number 4–11 mg KOH/g, saponification number 86–130 mg KOH/g, iodine
number 5–11 g I2/100 g, mp 60–66 8C). It has a consistency similar to beeswax
with a fatty feel, which can be attributed to its glyceride content. Ghedda wax
consists of ca. 78–80 % wax esters, 4 % glycerides (fats), 5–6 % free acids,
8–9 % hydrocarbons, and 1–2 % partly resinous impurities (12).
Among the acids are 7- and 16-hydroxypalmitic acids and 7,16-dihydroxy-
margaric acid, which can be used for identification of ghedda wax or other bees-
waxes blended with ghedda wax in modern umpire assays.
Shellac Wax. The resinous exudate of the scale insect Laccifer lacca
(formerly Tachardia lacca) of the Coccidae family is an important commercial
product, known as shellac. Shellac contains 4–5 % wax, which can be obtained
by treatment with dilute soda solution or 90–95 % ethanol (spirit extraction)
as an insoluble component.
Shellac wax is a hard, yellow to brown product. Very light, bleached types
are also known. The widely varying data found in the literature (eg, acid number
12–24 mg KOH/g, saponification number 63–126 mg KOH/g, iodine number
6–9 g I2/100 g, solidifying point 58–80 8C) may be attributed to the different
sources of shellac and different methods of extraction and posttreatment of the
Vol. 2 WAXES 109
wax component. Thus, the composition with regard to substance classes also var-
ies as follows (12):
Investigations (55) of bleached shellac waxes show that the total acid con-
tent of the wax, similar to that of beeswax, consists essentially of a mixture of
even-numbered fatty acids (C12 –C18, 21–26 % of the total acid) and waxy acids
(C28 –C34, mainly C32 and C34). A significant component is trihydroxypalmitic
acid (aleuritic acid), even though it occurs only in quantities of 0.5–1.0 %.
Among the alcohols (C28 –C32) the C28 component predominates (62–65 % of
the total alcohol). The main hydrocarbon components are paraffins with 27, 29,
and 31 carbon atoms.
Chinese Insect Wax. The East Asian wax scale insect (Ericerus pela,
subgenus Lecaniinae, family Coccidae) secretes a wax on the twigs and leaves
of trees and shrubs on which it is a parasite. The wax can be extracted by melting
with hot water. It is hard, brittle, crystalline, and odorless, with a light color, and
consists of 95–97 % wax esters, 0.5–1.0 % free acids, and 2–3 % hydrocarbons
(12). The high proportion of wax esters is significant. The East Indian wax
scale insect (Ceroplastes ceriferus) is said to produce a wax similar to beeswax
in composition (12).
Wool Wax. Sheep secrete a substance from their sebaceous glands to pro-
tect the epidermis and the wool from adverse weather conditions and from their
own acidic perspiration. This substance is incorrectly referred to in the literature
as wool fat. Because of its chemical composition it should definitely be classified
as a wax.
Wool wax is obtained in emulsified form from raw wool by machine washing
with soap liquor. Extraction of wool with suitable organic solvents is not
employed because it would lead to complete degreasing of the wool and thus
make it nonelastic.
Many processes exist for obtaining crude wool wax from the emulsion, some
of them patented (36). Wax is first precipitated along with the acids of the soap
solution by addition of sulfuric acid. The fatty acids remain in the crude product
regardless of whether the precipitates are pressed out, centrifuged, or extracted
with solvents. They can be removed almost completely by dissolving the crude
product (eg, in hexane) and subsequent extraction with a mixture of alcohol
and dilute aqueous sodium hydroxide. Further purification steps (eg, bleaching
with hydrogen peroxide in the presence of phosphoric acid (56)) give neutral wool
wax, which is usually treated subsequently with adsorbents such as bleaching
earth or activated carbon and sold in a highly pure form as Adeps lanae.
Crude wool wax is a greasy, glutinuous, brown-yellow to brown-black sub-
stance with a penetrating goatlike odor (mp 34–38 8C). Neutral wool wax is yel-
low to light brown in color (mp 38–42 8C), with a milder odor, whereas Adeps
lanae is a pale yellow, almost odorless substance (mp 40–42 8C).
110 WAXES Vol. 2
The composition of wool wax varies depending on the origin of the wool,
breed of sheep, climatic conditions, farming method, and type of pasture.
Wool wax usually consists of a mixture of ca. 48 % wax esters, 33 % sterol
esters, 6 % free sterols, 3.5 % free acids, 6 % lactones, and 1–2 % hydrocarbons
(12). The low mp and appearance of wool wax may be attributed to the high pro-
portion of branched acids.
Uses. Wool wax is used mainly as the crude product, as neutral wool wax,
or as hydrogenated neutral wool wax and in the form of wool wax alcohols
(obtainable by hydrolysis) in leather polishes, cosmetics (eg, creams, baby-care
products, and toilet soap), pharmaceuticals (eg, plasters, ointments, and suppo-
sitories), and lubricants. The purified, commercial product is known as wool wax
or lanolin.
3. Montan Wax
Wax acids 35 %
Wax alcohols 20 %
Hydroxycarboxylic acids 10 %
Dicarboxylic acids 3%
Wax ketones 1%
Hydrocarbons 2%
Resin acids 15 %
Sterols 10 %
Dark residues 4%
112 WAXES Vol. 2
hydrolyzed; the wax alcohols formed are oxidized to wax acids; and the hydroxycar-
boxylic acids are oxidized to dicarboxylic acids. Some esters remain unaltered.
CH3
ðCH2 Þn
COOH þ HOCH2
CH2 OH ! CH3
ðCH2 Þn
COO
CH2
CH2
OR
R
H or CH3
ðCH2 Þn
CO
CH3
ðCH2 Þn
COOH þ CaðOHÞ2 ! ðCH3 COOÞ2 Ca2þ
ðCH2 Þn
The most important products in this series are the esters of ethylene glycol
(eg, Hoechst Wax E or Hoechst Wax KSL) and mixtures of esters of 1,3-butane-
diol with calcium montanate (eg, Hoechst Wax OP). Broad variation in properties
can be achieved by using tri- and polyhydric alcohols, aromatic alcohols, aminoal-
cohols, and ethoxylated or fluorinated alcohols. Reactive waxes can be produced
from partial esters of polyols by treatment with dicarboxylic acids, isocyanates,
epoxides, or unsaturated carboxylic acids.
These waxes exhibit interesting properties and have been produced and
used for special applications. If the reaction is not stopped at the reactive wax
stage and a suitable stoichiometry is chosen, oligomeric waxes are produced.
3.5. Uses and Economic Aspects. Montan wax acid and its deriva-
tives are versatile materials (64,65,69–71). They are used in the form of flakes,
powders, pastes with solvents, or aqueous emulsions. The main areas of use are
in polishes for floors, cars, and leather; in plastics processing as lubricants,
release agents, and nucleating agents; in the paper and building industries;
and in wood and metal processing. Montan wax derivatives are also used in
small quantities in pharmaceutical formulations; cosmetics; and the production
of adhesives, resins, and office equipment. In addition to their technical proper-
ties, the low toxicity of montan wax derivatives and the fact that BGA or FDA
approval has already been granted for some derivatives are important criteria
favoring their use.
Production figure of montan wax for 1994 are given below, together with a
breakdown of montan wax consumption in different application areas:
4. Petroleum Waxes
Table 8. Variation in Physical Data with Degree of Refining of Paraffin Wax (Starting
Material : Slack Wax from a Medium Machine Oil)
Slack Deoiled Filtered Fully
Characteristics wax wax (decolorized) wax refined wax
Congealing point, 8C 59 62 62 62
Needle penetration, 0.1 mm
at 25 8C 57 18 18 17
at 30 8C 80 26 26 25
at 35 8C 105 36 36 36
Oil content, % 9.9 0.5 0.5 0.4
Viscosity (at 100 8C), mm2/s 6.8 6.1 6.0 5.8
Color brown brown whitish white
Fluorescence very strong weak none
Table 9. Composition of Slack Waxes from Various Lubricating Oil Distillates of the
Same origin (80 % Arabian, 10 % German, and 10 % Norwegian Crude Oil)
Slack waxes from
Heavy Light Medium Heavy
Characteristic spindle oil machine oil machine oil machine oil
Congealing point, 8C 47 50 58 62
Oil content, % 9.5 9.0 9.8 8.0
n-Paraffin content, % 73 58 52 39
C-number range 16–36 19–43 21–48 22–58
C-number max. 24 29 34 41
Cmax content, % 17 14 10 8
Table 10. Composition of Deoiled Slack Waxes from Two Lubricating Oil Distillates
(Origin – See Table)
Characteristic Slack wax * Deoiled Slack wax ** Deoiled
Congealing point, 8C 47 50 58 67
Oil content, % 9.5 0.3 9.8 0.4
n-Paraffin content, % 73 84 52 64
C-number range 16–36 19–36 21–48 23–48
C-number max. 24 24 34 34
Cmax content, % 17 19 10 11
* From heavy spindle oil.
** From medium machine oil.
120 WAXES Vol. 2
Fig. 4. Variation of needle penetration with temperature for deoiled slack waxes of dif-
ferent origins. a) From heavy spindle oil (North Sea); b) From light machine oil (North
Sea); c) From heavy machine oil (North Sea); d) Wax from Indonesia.
darken on heating. The congealing points are between 48 and 72 8C; the drop
points between 48 and 75 8C; and the needle penetrations at 25 8C between 10
and 30. At the same degree of deoiling, the course of the penetration – temperature
curve (increase in penetration in the penetrogram (83) essentially depends on the
origin of the waxes (Fig. 4).
Crude soft waxes [64742-67-2] are sometimes known as foot oils (see
Section 4.2.3.2) and are formed in the deoiling of slack waxes. They consist pre-
dominantly of low molar mass n-paraffins and strongly branched isoparaffins;
higher molar mass, weakly branched isoparaffins; naphthenes; and alkylaro-
matics. The n-paraffin content of soft waxes is generally <25 wt %.
Soft waxes are very soft, unctuous to inhomogeneous materials at room
temperature because of their high oil content, which can be 30 wt % or greater.
The oil is included in voids of the crystal lattice of the solid hydrocarbons. Soft
waxes have a light- to dark-brown color and congealing points <40 8C.
Their viscosities at 100 8C are between 3 and 12 mm2/s; their flash points,
between 180 and 220 8C.
Semi-refined, filtered (decolorized), and fully refined waxes are
similar to deoiled slack waxes in the composition of their main components.
Through refining they are freed from more highly condensed alkylaromatics
and naphthenes, olefins, and sulfur and nitrogen compounds. Depending on
the origin, their content of n-paraffins varies between 92 wt % (East-Asian
waxes) and 45 wt % (waxes from heavy machine oil distillates of German origin).
The composition of the refined waxes is determined not only by the origin of the
crude (Table 11), but also by the boiling range of the starting distillate. As the
boiling ranges of the lubricating oils rise, the molar masses of the hydrocarbons
Vol. 2 WAXES 121
Because of their higher molar masses and thus higher boiling ranges,
waxes become enriched in the residues from atmospheric distillation (long resi-
dues). In processing the residues to lubricating oils, waxes are obtained in all
fractions of the vacuum distillation. Because of their poor solubility, they give
rise to the high pour points of the basic oils and must therefore be removed. The
yields of wax vary depending on the origin and quality requirements of the oils.
For a wax content of crude petroleum of ca. 5 %, 8–18 % solid waxes in the indi-
vidual lubricating oil fractions can be expected.
Slack waxes are produced in the dewaxing of lubricating oil distillates. The
type of process used and process parameters used are according to the desired
quality of the lubricating oils.
The process steps necessary for the production of refined waxes are shown
in Figure 5.
Dewaxing Lubricating Oil Distillates. The principle of dewaxing lubri-
cating oil distillates is based on the different crystallization temperatures of
straight-chain and weakly branched paraffins and the oil phase. The wax-
containing solvent-neutral oils (ie, the raffinates from solvent refining) are
mixed with suitable solvents such as propane, naphtha, chlorohydrocarbons,
ketones, or (most frequently) a toluene–methyl ethyl ketone mixture. The oil–
solvent mixture is subsequently warmed to obtain a homogeneous solution and
then cooled continuously in scraping chillers to obtain the wax crystals in a loose
a higher or lower proportion of soft waxes, are known as foot oils or soft waxes,
depending on the consistency and the process used.
Solvent deoiling is the most commonly used process, involving one of
three possible techniques: pulping, crystallization, or spray deoiling. Starting
materials for all deoiling processes are either molten slack waxes (if deoiling is
carried out in a location other than that of slack wax production) or slack wax–
solvent mixtures, formed in dewaxing lubricating oil distillates (see Table 12).
Slack waxes from spindle oils to heavy machine oils, and to some extent pet-
roleum as well (see Section 4.3), can be deoiled by the three solvent deoiling
methods. The yields of deoiled slack waxes depend on their origin, the process
used, and the degree of deoiling. Yields are between 80 % for spindle oil distil-
lates and ca. 60 % for heavy machine oil distillates, based on the slack wax used.
Pulping Process. The inhomogeneous mixture of crystallized wax, oil,
and solvent is diluted or repulped by addition of the same solvent or a solvent
mixture (normally by using part of the wash filtrates produced later). The oils,
soft waxes, and part of the low molar mass hard waxes thus dissolve. The pulp
is cooled with stirring in scraping chillers, and the temperature is adjusted to
that necessary for the desired degree of deoiling. Hard waxes, which are partly
undissolved and have partly recrystallized, are removed on vacuum rotary filters
and washed with solvent; the solvent is subsequently distilled off.
The quality of the hard waxes produced is determined by the deoiling tem-
perature and the quantity of solvent used. With a 1,2-dichloroethane–methylene
chloride mixture as solvent (Edeleanu Process), for example, typical deoiling
temperatures are 10 8C for slack waxes from heavy spindle oil, 5 8C for
those from a light machine oil, and þ15 8C for those from medium to heavy
machine oil.
Crystallization Process. The molten slack wax or solvent-containing
wax crystallizate from the dewaxing process is dissolved completely, or almost
Vol. 2 WAXES 125
Sulfuric acid refining can only be carried out batchwise because of the slow
separation of the wax and acid sludge phases. Molten wax is mixed with sulfuric
acid at 80–140 8C for some time (between 0.5 and 2 h, depending on the origin of
the wax). After separation of the heavier acid sludge, the wax is washed with
alkali, treated with decolorizing clay, and stripped with steam to remove the
last residues of impurities and odorous substances.
All compounds that react chemically with the aggressive acid, such as unsa-
turated aliphatic and aromatic hydrocarbons, metal compounds, and those con-
taining hetero atoms, are removed by sulfuric acid refining. Labile quaternary
carbon atoms are also attacked, with bond cleavage and formation of aliphatic
sulfonic acids.
The process can be optimized by variation of the number of reaction steps,
degree of acidity, reaction temperature, ratio of wax to acid used, reaction time,
and intensity of mixing. The higher the average molar mass of the wax fraction,
the slower is the separation between the wax and the acid sludge phase.
The process has several disadvantages such as the necessity for batch
operations, poor yields of raffinates, formation of polluting byproducts, and
waste gas and corrosion problems. For these reasons, almost all new plants
are based on hydrotreating.
Hydrotreating (Fig. 7). Deoiled slack wax is heated to the required tem-
perature together with fresh and recycyled hydrogen in the preheater (b) and
passed over a sulfur-resistant fixed-bed catalyst (c). After cooling, the reaction
mixture and hydrogen are separated in the high-pressure gas separator (d),
and hydrogen is recycled into the process. After depressurizing in the low-
pressure gas separator (e), the wax passes into the stripper (f), where all the
light crack products, odorous substances, and reaction gases are stripped off com-
pletely with compressed steam under vacuum. The wax is subsequently dried
with nitrogen (g). A combination of metals from groups 6 and 10 on an inert
Uses. Paraffin waxes are used in very different fields of industry. They
not only are used as such, but are also important components of wax compounds
(ie, mixtures of paraffin waxes with other waxes, plastics, resins, etc.). The con-
sumption figures listed in Table 16 cover the most important areas of use.
Slack waxes that are not processed further to refined waxes or sub-
jected to dehydrogenation or cracking processes – for example, to give a-olefins
(see Hydrocarbons) – are used mostly in the wood and mineral oil processing
industries.
Wood processing requires slack waxes, either molten or in the form of emul-
sions, mainly for the production of particle board and fiberboard. In the United
States the production of artificial logs for barbecues and fires is also important.
In the mineral oil processing industry, slack waxes are used to produce
petroleum jellies, wire-drawing agents, lubricants, spray emulsions, and
anticorrosives.
In addition to wax industry products with a wide area of applications,
slack waxes are also used to produce chlorinated paraffins, papers and card-
boards for technical uses, cable coverings, solid anticorrosives, fire lighters,
and wax torches.
Soft waxes have only a limited market in the production of industrial
petroleum jellies and emulsions (wood and building industry), torches, and fire
lighters and for conditioning fertilizers. The deoiling of soft waxes gives highly
plastic and extendable products (isowaxes).
Refined Waxes [Scale Waxes, Filtered (Decolorized), Semi- and
Fully Refined Waxes]. In Germany the producers of candles and wax goods
(pictures, sculptures, ornaments) are the main users of refined waxes, which
can be processed by all common production methods for candles (molding, draw-
ing, extrusion molding, and powder compressing). In terms of quantity, paraffin
wax is the most important raw material for the candle industry.
The second most important consumer of refined waxes is the wax industry,
which requires paraffin waxes as an important raw material for compounding
wax products for the most widely varying applications.
In the paper, cardboard, and packaging industries, refined waxes are used
in the form of emulsions for paper sizing as antifoams in the production of paper,
cardboard, and corrugated cardboard; and as impregnating and coating waxes
for paper and packaging materials (112).
132 WAXES Vol. 2
enriched compared with the petrolatum from which they are obtained by deoil-
ing. Molar masses are between 450 and 1200 g/mol. The n-paraffin content can be
as high as 40 %. Depending on the degree of refining the impurities present in
petrolatum are removed completely or to a great extent. The oil content is
usually between 1.5 and 5 %.
The plastic microwaxes, although relatively hard, are definitely plastic,
readily moldable, and kneadable. The plasticity is due to the presence of
very long carbon chains in the n-paraffins and terminally branched isopar-
affins. The waxes feel dry to slightly adhesive, and are sold as white to
brown products. Congealing points are 65 –80 8C, and the hardness (needle
penetration at 25 8C) is 10 to 60. Highly refined microwaxes are odorless and
tasteless.
Hard microwaxes also essentially consist of high molar mass alkanes
(predominantly n-alkanes and isoalkanes with short side chains statistically dis-
tributed over the main carbon chain) and naphthenes. Average molar masses are
between ca. 500 and 1200 g/mol. Thus, for example, a refined, hard microwax of
North American origin contains 89 % nonpolar hydrocarbons (n-, iso-, and cyclo-
paraffins) and has an average C-number of 48 (85). The n-paraffin content can be
as high as 70 %.
The extremely microcrystalline to almost amorphous structure of hard
microwaxes is associated with the formation of supercooled melts during the set-
ting process.
The waxes are tough and hard, barely ductile and weakly adhesive.
Depending on the degree of refining, hard white, yellow and brown microwaxes
are on the market, the white to pale yellow products being of highest purity.
They solidify between 80 and 94 8C. The needle penetration at 25 8C is <10.
The hardness of hard waxes decreases only slightly with temperature until
close to their softening point (Fig. 8).
Occurrence of Raw Materials and Processing. Like paraffin waxes,
microwaxes are components of crude petroleum. Because of their high molar
masses, and thus high boiling range and poor solubility, they either are enriched
in the vacuum residues (short residues) from lubricating oil distillation (residual
waxes) or separate during the transportation and storage of crude oils (settling
waxes). Settling waxes form pasty to solid deposits on the walls of pump tubes
and pipelines and on the bottom of storage tanks (pipe and tank bottom
waxes). Vacuum residues and pipe and tank bottom waxes are the raw materials
for recovery of microcrystalline waxes.
Figure 9 shows schematically the process steps involved in the production
of microcrystalline waxes.
In the United States, the most important producer country for microwaxes,
the vacuum residues are subjected to deasphalting, refining, and dewaxing to
produce high-value bright stock oils. Bright stock slack waxes (petrolatum) are
produced in the process as economically important byproducts. However, only in
the rarest cases they are processed further to microwaxes in the lube oil refi-
neries themselves. Petrolatum is either marketed directly or used, for example,
for the production of petroleum jellies. For further processing it is sold to compa-
nies specializing in the production of microwaxes. The same companies also pro-
cess settling waxes purchased from oil companies.
Vol. 2 WAXES 135
The crude waxes (deasphalted settling waxes) and bright stock slack waxes
are mixed either molten or as a pulp (if the deoiling is to follow immediately after
dewaxing the brightstock oils), without isolation of the slack wax with fresh,
warm solvents and warmed until only the less soluble (high molar mass)
waxes remain undissolved. These waxes are then filtered, washed, and freed
from solvent (hard microwax fraction). To obtain plastic microwaxes the filtrate
is cooled further. The properties of reprecipitated and refiltered waxes are deter-
mined by the filter temperature. The low molar mass, very plastic, and soft
microwax foot oils [64742-67-2] remains in the filtrate. If the hard microwaxes
are not extracted, a one-step crystallization process is sufficient for obtaining
plastic microwaxes.
Deoiled microwaxes are still dark yellow to dark brown and not completely
odorless, so they must subsequently be refined to satisfy a broad application range.
Refining. Refining is carried out either with adsorbents (decolorizing) or
with hydrogen (hydrotreating). Refining with concentrated sulfuric acid, which is
possible in principle, has not become important economically because of the high
loss of material.
For refining microwaxes with adsorbents, the same process principles apply
as for decolorizing paraffin wax (see page 33). Since microwaxes are more diffi-
cult to decolorize than paraffin waxes, countercurrent percolation is used predo-
minantly. Here, the wax melt continually comes in contact with fresh
decolorizing clay (115). Bauxite is also frequently used as a decolorizing agent.
Hydrotreating is carried out as described for paraffin waxes, in one or two
steps in reactors packed with sulfur-resistant, heavy-metal-doped mixed catalysts
(see page 34). Reaction conditions are ca. 350 8C and 150 bar, depending on the
properties of the catalyst, at a space velocity of 0.3 L of wax per liter of catalyst
and hour. Hydrogenated products are treated further by adsorptive decolorizing
or stripping to completely remove odorous materials and volatile components.
In the United States, a whole range of microwaxes is produced according to
the processed described. Products differ in their congealing point, drop point,
hardness, viscosity, and color. They are sold in the liquid (molten) form and as
solids in the form of slabs, pellets, and powder.
Table 18. Producers of Microwaxes in the United States and Commercial Products
Producer Production site Commercial products
Bareco Div. of Kilgore, Tex. Petrolite Crown, Be Square,
Petrolite Corp. Barnsdale, Okla. Mekon, Stawax, Cardis,
Victory, Ultraflex
Mobil Oil Co. Beaumont, Tex. Wax Rex, Mobilwax
Paulsboro, N.J.
Quaker State Oil Emlenton, Pa. Microwax, Superflex
Ref. Corp. Farmers Valley,
Pa. Newell, W.V.
Shell Oil Co. Deer Park, Tex. Shellwax
Wood River, Ill.
Sun Oil Co. Tulsa, Okla. Sunoco Wax, Sunolite
Witco Chem. Corp. Bradford, Pa. Multiwax
Vol. 2 WAXES 137
The EWF specifies microwaxes for use in packaging materials for foods, as
shown in Table 19.
In the use of microwaxes to produce food packaging materials, and cheese
and chewing gum waxes, purity and physiological nontoxicity are particularly
important. For microwaxes the absence of specific polycyclic aromatics is required.
These substances can occur as impurities in insufficiently refined waxes, and some
of them have been found to be carcinogenic. Polycyclic aromatics can be detected
by UV extinction after a special enrichment process (detection limit in micro-
waxes 107 g per gram of wax substance). Microwaxes have been shown to be
toxicologically and physiologically harmless in extensive animal experiments,
when UV adsorption values specified according to test regulations of the FDA
(101) and the BGA (103) are not exceeded.
The physical characteristics of microwaxes are determined according to the
same standardized methods used for paraffin waxes (see Table 14). In addition,
some special analytical methods have been introduced, for example:
1. Paper and cardboard coating for packaging meat and sausage products,
cereals, bread, candy, ice cream, cookies, yogurt, and frozen food
2. External lubricant (sliding) for extrusion of plastics
3. Formulation of cheese waxes and chewing gum bases
4. Decorative for candles, production of wax pictures and sculptures
5. Polishes and grinding agents
6. Lost wax casting and dental compounds
7. Foam regulators for detergents
8. Protectives against crack formation in tires and rubber articles (sun-proof
waxes) (116)
9. Additives for explosives and propellant materials.
Scotland, and Northern Ireland (117). Only specified microwaxes may still be
added to foods for a limited period.
In the United States, not indication has been given of any intention to ban
paraffins in food.
4.5. Ecology. Paraffin and microcrystalline waxes are insoluble in
water and can be separated as solids, together with sludge, in oil separators
and sewage treatment plants. In the German catalogue of substances that are
hazardous to water, they are therefore listed in water hazard class 0 (Wasserge-
fährdungsklasse, WGK 0). Since paraffins have a water solubility of <1 mg/L,
they have no toxic effects on aquatic organisms. Used paraffin-containing packa-
ging materials can be combusted without any problem in waste incineration
plants. No polluting substances are formed, and the high calorific value of the
paraffins improves the energy balance of the plants.
The biodegradability of petroleum waxes satisfies OECD guideline 301 B
(1981) (123). With the exception of microwaxes, biodegradabilities of more than
60 % after 28 d were measured. Bacterial degradation mechanisms are described
in (124). According to this, paraffin-containing packaging materials can also be
composted (125–127). No substances that are toxic or hazardous to groundwater
are formed.
The biodegradation of waxes under anaerobic conditions (covered landfills)
proceeds much more slowly.
A large number of investigations have been carried out on the recycling of
used wax-containing packaging materials. The water-insoluble, hydrophobic pet-
roleum wax tends to form particles in the paper mills and give rise to blockages
in the sizing equipment (128).
In Europe, ca. 20106 t of waste paper was recycled as a raw material for
paper production in 1991. This figure includes 0.35106 t of wax-containing
paper and cardboard with average wax content of 12 %. Assuming an even dis-
tribution, this gives a wax content of 0.21 % wax in wastepaper (129). The prac-
tice has shown that in modern wastepaper processing plants, wax-containing
packaging materials do not interfere with the dilution process,but nearly
increase the quantity of waste products. These waste products can be used to pro-
duce water-resistant cardbord.
4.6. Economic Aspects. In 1992, 9.5105 t of paraffin and microcrys-
talline waxes were produced in the United States, 8105 t in Asia, 6.5105 t in
Western Europe, 2.6105 t in Germany alone, 2.0105 t in Eastern Europe. and
4.0105 t in the rest of the world.
In industrialized countries, a stagnation (probably a decrease) in petroleum
wax production is anticipated for the coming years, independent of the growth in
demand. This is because of the further development of car engine and tribology
technology (reduced specific lubricant consumption and longer periods between
oil changes), an increase in the proportion of partially or fully synthetic and
biodegradable lubricants, the recycling of used oils, and the closing down of
excess capacity in distillation plants despite an increase in the number of
motor vehicles and the distances traveled (130).
However, in Asia, particularly China and Malaysia, an increase in wax pro-
duction to 1.5106 t in 2000 is predicted, because these waxes can be inexpen-
sively produced from indigenous wax-rich crude petroleum by direct dewaxing.
Vol. 2 WAXES 141
5. Fischer–Tropsch Paraffins
Oxidized waxes containing fatty acids and fatty acid esters are produced
from synthetic paraffin by oxidation and partial saponification. Major areas of
application are as mold-release agents in plastics processing, in polishes and
cleaning agents, and as auxiliaries in the textile and paper industries.
Trade Names. Fischer–Tropsch waxes are marketed, eg, as Sasol Wax
(Sasol Marketing Co., Johannesburg, South Africa) and Vestowax SH/SP (Hüls
AG, Marl, Germany).
6. Polyolefin Waxes
the density and molar mass of the homopolymers and functionalizing polyethy-
lene by copolymerization with various monomers (eg, vinyl acetate or acrylic
acid) or by melt oxidation.
Because PE waxes are rather inexpensive, have improved applicability and
consistency in quality, and are in constant supply they have displaced expensive
natural waxes (eg, carnaúba and montan waxes) in many areas.
The high-pressure process is not applicable for production of polypropy-
lene waxes, because only soft or oily products can be obtained. These waxes
are therefore produced by the Ziegler process or by depolymerization of high-
density polypropylene.
Production. The technology involved in the production of high-pressure
PE waxes (see Fig. 10) is analogous to that used for high-pressure polyethylene.
The only differences are in the product finishing process. Unlike viscous plastic
melts, which must be granulated underwater in granulators, for example, the
mobile wax melt can be converted into powder by spraying or into granules by
using dicers.
The reaction vessels are stirred autoclaves or tubular reactors, the design
of which can vary considerably.
Reaction Mechanism. The high-pressure polymerization of ethylene
is a highly exothermic radical chain reaction, which is initiated by reaction
of ethylene with radicals formed by decomposition of an initiator (136).
Chain propagation occurs through addition of further ethylene units. Chain
growth is terminated, for example, by reaction with a regulator molecule
such as propene:
Table 22. Physical and Chemical Data for Copolymeric High-pressure Polyethylene
Waxes
Ethylene – vinyl Ethylene – acrylic
acetate copolymer acid copolymer
Characteristic wax wax
Weight-average molar mass M w, g/mol ca. 6800 ca. 6100
Number-average molar mass M n, g/mol 3000 3000
Melt viscosity (120 8C), mm2/s (DGF-M-III 8) 2000 1300
Solidification point, 8C (DGF-M-III 4a) 85 93
Drop point, 8C (DGF-M-III 3) 96 102
Ball indentation hardness (120 8C), 120 410
bar (DGF-M-III 9a)
Acid number, mg KOH/g 45
Vinyl acetate content, % 10
Table 23. Structure and Properties of Typical Commercial Ziegler Polyolefin Waxes
A B C D E
Ethylene Ethylene Ethylene Ethylene Propylene
homopolymer copolymer homopolymer homopolymer homopolymer
wax wax wax wax wax
Number-average 1600 2200 850 5000 3000
molar mass M n,
g/mol (GPC)
Weight-average 4800 6300 920 17 000 18 000
molar mass
M w, g/mol (GPC)
Molar mass 3.0 2.9 1.1 3.4 6.0
distribution
(M w/M n)
Density (23 8C), 0.97 0.93 0.96 0.97 0.88
g/cm3 (DGF-M-
III 2a)
Heat of fusion, 240 120 230 200 60
J/g (DSC)
Viscosity (140 8C), 400 600 10 25 000 1700 *
mPa s
Drop point, 8C 125 118 107 135 ** 158 **
(DGF-M-III 3)
Needle penetration, <1 3 2 <1 <1
0.1 mm
(DGF-M-III 9b)
* Temperature of measurement 170 8C.
** Softening point ring/ball (DGF-M-III 13).
linear molecular structure. They are brittle and hard (needle penetration num-
bers <1) and have high densities and melting points (up to ca. 127 8C). These
products can be obtained only by Ziegler polymerization. Radical polymerization
always gives branched waxes.
By using 1-alkenes (generally propene) as copolymers, defects can be intro-
duced into the polymer chain in the form of short-chain branches, which lead to a
decrease in crystallinity. The density, hardness, drop point, and heat of fusion
are also decreased, while flexibility and plasticity increase.
In Table 23, some important physical properties of typical commercially
available Ziegler polyolefin waxes are compared.
Ziegler waxes are colorless and odorless and form clear, transparent melts
with high thermostability. Like most other waxes, they are soluble on warming
in nonpolar aliphatic and aromatic solvents without leaving a residue, and on
cooling often crystallize out as pastes. Commercial products have a melt viscosity
range from ca. 10 to 30 000 mPa s (at 140 8C). Their densities are between 0.92
and 0.97 g/cm3.
Degradation Polyolefin Waxes. By heating in the absence of air, the
molecular chain of high molar mass polyolefins can be cleaved into smaller mole-
cules with waxlike character. The starting materials are usually low- and high-
density polyethylene, isotactic polypropylene, and polybutene. The production
capacities of degradation polyolefin waxes have diminished since ca. 1980 in
152 WAXES Vol. 2
favor of the built products. While the proportion of degradation PE waxes has
diminished, that of PP waxes has increased. Despite the relatively unfavorable
energy balance of the degradation process, it is still used because of its technolo-
gical simplicity and, for PE waxes, occupies third place with regard to the pro-
duction quantity after the high-pressure and low-pressure PE waxes. The
degradation process was developed approximately in parallel to high-pressure
polymerization (176–178).
Production. Polyethylene is a thermally stable polymeric hydrocarbon.
In the absence of oxyen it can withstand temperatures up to ca. 290 8C (179).
At higher temperature, thermal degradation begins with a decrease in the
molar mass; from 400 8C, degradation occurs with high rate and yield (179).
The cleavage products consist almost exclusively of a wax fraction with a
molar mass distribution that is barely wider than that of high-pressure PE
waxes. Low-boiling and gaseous crack products are formed only in small quanti-
ties. Individual process steps are shown in Figure 12.
The starting materials for thermal degradation to polyolefin waxes are
LDPE and HDPE (180) and HD polypropylene (181–184). Degradation of copo-
lymeric polyolefins (185) and polybutenes (186) has also been described but has
not achieved great economic importance.
In the industrial process (180), polyolefin pellets are melted under nitrogen
in a single-screw extruder and extruded into a steel tube heated to ca. 400 8C.
The residence time is 15–30 min. The product is then cooled by evaporative
cooling in ca. 5 min to <250 8C. It is freed from low molar mass components by
degassing in a separator. The product can then be finished as described for
high-pressure PE wax.
The degree of branching of degradation waxes formed, and thus all
structure-dependent properties such as crystallinity, density, hardness, and
Fig. 12. Process schematic for the production of degradation polyolefin waxes.
Vol. 2 WAXES 153
mp, are controlled by the choice of starting material (LDPE or HDPE, stabilizer
free). The average molar mass depends on residence time in the hot-tube reactor
and the average molar mass of the starting material. The degradation tempera-
ture affects mainly the reaction time and thus the throughput rate.
Batchwise, thermal degradation in autoclaves has also been described
(182,185). It is performed in certain cases in the presence of hydrogen and hydro-
genation catalysts to saturate double bonds formed on chain cleavage (176).
Like other polyethylene waxes, degradation waxes can be modified by air
oxidation and thus rendered emulsifiable. Preliminary hydrogenation of double
bonds can be necessary to avoid cross-linking reactions during oxidation.
Because of their double bonds, degradation waxes are particularly
suited to grafting with unsaturated carboxylic acids (eg, with maleic anhy-
dride (187,188)).
Little information is available in the literature on the mechanism of ther-
mal degradation of polyolefins to form waxes (179). Presumably, radicals form
from the few oxygen-containing groups in the polyolefin and initiate a radical
chain reaction, in which the chain is cleaved through radical transfer and a dou-
ble bond is formed:
mp (DSC) 130–160 8C
Ball indentation hardness (23 8C) 500–1900 bar
(DGF-M-III 9a)
Melt viscosity (200 8C) (DGF-M-III 8) 800–1900 m2/s
Weight-average molar mass (GPC, 17 000–37 000 g/mol
standard: PP)
For polypropylene wax (211–213), this is the only method of introducing oxygen
functions that is used in practice, since air oxidation would lead to low molar
mass, colored, soft degradation products.
6.2. Uses. Nonpolar Polyolefin Waxes. Polyolefin waxes are used in
widely differing applications, mostly as auxiliaries in products and produc-
tion processes. The areas of use of high-pressure, Ziegler, and degradation
waxes overlap extensively. More details can be found in leaflets and bro-
chures of various production companies. Important areas of use include the
following:
7. Toxicology
On the basis of experience accumulated over thousands of years, waxes are gene-
rally considered harmless, nontoxic, environmentally friendly products. The
results of scientific toxicological and ecotoxicological tests on natural waxes, mon-
tan waxes, polyethylene waxes, and other wax products generally support this.
Natural Waxes. From the group of animal and vegetable waxes, carnaúba
wax was subjected to toxicological tests. In a two-generation feeding experiment,
1 % wax (max.) based on the feed, was administered first to pregnant rats and then
to their first-generation offspring over a period of 90 d. The results gave no indica-
tion of dose-dependent effects in the rats treated compared to the untreated con-
trol group (218).
Montan Waxes and Their Derivatives. The toxicological properties of
acid, ester, and partially saponified ester waxes (partially neutralized with
calcium hydroxide) derived from montan wax are determined mainly by the indi-
gestibility of the products, as in the case of natural waxes. Besides determining
acute oral toxicity, extensive feeding experiments on rats and dogs over periods
of 90 d, 140 d, and 2 a have been carried out with montan wax derivatives (219).
The LD50 values are >15 000 and 20 000 mg per kilogram of body weight, respec-
tively. According to the usual classification, these products are considered prac-
tically nontoxic and harmless (220).
Long-term feeding experiments led to no macroscopically or microscopically
recognizable toxicological damage to the animals, even at very high dosages of
5 % in the feed (Table 26).
Dermatological examination of montan wax derivatives in the percutaneous
test and the mucous membrane test on rabbits (Hoechst waxes S, E, and OP) did
not lead to any irritation of the skin or mucous membranes.
In the Ames mutagenicity test with and without external metabolic activa-
tion, a montanic acid–ethylene glycol ester wax (Hoechst wax E) did not show
any mutagenicity (tested dosages 4–5000 mg per plate) (221).
The ecotoxicological behavior of montan waxes is attributed to their water
insolubility. Mixtures with water are attacked very slowly by microorganisms
and cause no perceptible biological oxygen demand (BOD) in water bodies and
sewage treatment plants. The waxes are separated from wastewater as ballast
materials together with sludge. In the determination of biodegradability accord-
ing to OECD test method 302 B (static test/Zahn–Wellens test), the finely dis-
persed montan wax derivatives in wastewater show good elimination of
chemical oxygen demand (! Ecology and Exotoxicology). They are separated
with sludge and are nontoxic to fish and bacteria.
In the case of wax emulsions, when the usual test methods for determina-
tion of COD and BOD are used, a very slow oxygen consumption can be detected,
which is higher than the oxygen demand of the emulsifiers (219). In wastewater,
these emulsions break up as a result of great dilution, altered pH, and the pre-
sence of calcium and magnesium ions. The findings described for pure wax–
water mixtures also apply to the agglomerates formed.
Polyolefin Waxes. Polyethylene waxes, oxidized polyethylene waxes, and
polypropylene waxes are similar to those based on montan wax and natural waxes
with regard to toxicology. The LD50 values for rats are >5000 and 15 000 mg per
Vol. 2 WAXES 159
Table 26. Acute Oral, Subchronic, and Chronic Toxicity of Montan Wax Derivatives
Subchronic
toxicity, Chronic
Acute oral Subchronic 140-d toxicity, 2-a
toxicity, toxicity, 90-d no-effect no-effect
Chemical LD50, no-effect level *, level *, level *,
Wax composition mg/kg mg/kg mg/kg mg/kg
Acid wax montanic acid >15.000
(Hoechst C16 – C36 (rat)
wax S)
Ester wax esters of >20 000 50 000 50 000
(Hoechst montanic (mouse) (dog) (rat)
wax E) acid with
ethanediol
Ester wax esters of >15 000 50 000 (dog) 50 000
(Hoechst montanic (rat) (rat)
wax KPS) acid with
ethanediol
and 1,3-
butanediol
Ester wax esters of 50 000 (rat)
(Hostalub montanic
WE 4) acid with
glycerol
Partially esters of >20 000 50 000 50 000
saponified montanic (mouse) (dog) (rat)
ester wax acid with 1,
(Hoechst 3-butanediol
wax OP) and calcium
salts of
montanic
acid
* Proportion by weight in the feed.
Table 27. Acute oral, subchronic, and chronic toxicity of polyolefin waxes
Subchronic Subchronic
Acute oral toxicity, 90-d toxicity, 9-month
toxicity, LD50, (rat) no-effect (rat) no-effect
Wax mg/kg level *, mg/kg level *, mg/kg
Polyethylene wax (Hoechst >15 000 85 000
wax PE 520)
Polyethylene wax oxidates
(Hoechst wax PED 121) 16 000
(Hoechst wax PED 522/ >15 000 50 000
PED 136)
Polypropylene wax (Hoechst >5000
wax PP 230)
* Proportion by weight in the feed.
In a study of the toxicity of petroleum waxes, rats were fed with several
paraffin products for 2 years. At a dosage of 10 % wax in the feed, no toxicolo-
gical or carcinogenic effects were observed. Even on repeated application of
petroleum waxes to the skin of mice and rabbits, no carcinogenic effects
were detected (222).
Toxicological tests have been carried out with paraffins and microwaxes are
mainly 90-d feeding experiments on rats. In these, varying toxicological effects
were observed. For low molar mass paraffins, accumulation in the tissues and
increased organ weights of the liver, kidney, spleen and lymph glands were
observed. The most important histopathological findings were granulomata in
the liver and histiocytosis in the lymph glands. These effects were not found
with high molar mass waxes.
No toxic effects were found at a dose of 0.6 mg per kilogram of body weight
and day for low molar mass paraffins (C-number maximum 25) and 1.8 g per
kilogram of body weight per day for high molar mass paraffin (microwaxes,
C-number maximum 42).
The latest results are published in BIBRA reports (223,224), CONCAWE
reports (225), and reports of the Toxicology Forum (226,227).
Amide Waxes. The most well known amide wax, bis(stearoyl)- bis-
(palmitoyl)ethylenediamide, has an LD50 of >20 000 mg/kg (mouse). In determi-
nation of the chronic toxicity in a 2-a feeding experiment on rats, a non-effect
level of 5 % in the feed was established (218).
Approval for Food Contact. On the basis of their low toxicity, a range
of natural waxes, montan wax derivatives, amide waxes, polyethylene waxes,
paraffins, and microcrystalline waxes have been approved for food contact by
the health authorities of many countries [eg, the German BgVV (formerly
BGA), the U.S. FDA, and in Australia and Japan]. The products are also con-
tained in the draft version of the EC additives list (Synoptic Document no. 7 of
May, 15, 1994). Some of the waxes are approved in the European Directives
95/2/EC and 95/3/EC for food contact.
Registration According to EINECS/TSCA/AICS/DSL. Waxes are
registered under the CAS numbers given in Table 28 in the European Inventory
Vol. 2 WAXES 161
OSHA-PEL 2 mg/m3
ACGIH-TLV 2 mg/m3
REFERENCES
1. Chr. Columbus, Bordbuch, November, 29, 1492. Reprint: edition Erdman, Thienemann,
Stuttgart 1983.
2. M. E. Chevreul: ‘‘Recherches chimiques sur les corps gras d’origine animale,’’ Frankreich
1823.
3. Hoechst AG: ‘‘Vom Wachs,’’ Hoechster Beiträge zur Kenntnis der Wachse, Frankfurt
1959 ff.
4. DGF-Einheitsmethoden, Abteilung M: Wachse und Wachsprodukte, Stuttgart
1975.
5. Erläuterungen des Rates für die Zusammenarbeit auf dem Gebiet des Zollwesens
3404/1 (1992) und 3404/2 (1988): International Customs Tariff, Harmonized System
HS 3404: ‘‘Artificial Waxes (Including Water-Soluble Waxes); Prepared Waxes, not
Emulsified or Containing Solvents’’.
6. J. Haggin: ‘‘Asian Natural Gas Pipeline Proposed for Easing Energy Feedstocks
Strains,’’ Chem. Eng. News (1994) no. 6, 17–18.
7. ‘‘Wax Data,’’ Newsletter for Marketers and Consumers of Petroleum, Natural and
Synthetic Waxes, Rauch Associates Inc., Bridgewater, New Jersey (published 13
times annually).
8. Kahl & Co., Otto-Hahn-Str. 2, 22946 Trittau, personal communication, 1994.
9. ‘‘Animal and Vegetable Waxes–A Review,’’ Wax Data 1993, Rauch Associates Inc.,
P.O. Box 6802, Bridgewater, N.J. 08807.
10. Chem. Mark. Rep. 244, (1993) no. 12, 7, 11.
11. ‘‘Study Forecasts Growth for Natural and Synthetic Waxes,’’ Wax Data 1994, Rauch
Associates Inc., P.O. Box 6802, Bridgewater, N.J. 08807.
12. A. H. Warth: The Chemistry and Technology of Waxes, Reinhold Publ. Corp.,
New York 1956.
13. R. Büll et al.: ‘‘Vom Wachs,’’ vol. 1; Hoechster Beiträge zur Kenntnis der Wachse, vol. 1
(contributions 1–8); vol. 2 (contributions 1–10), Hoechst AG, Werk Gersthofen,
Augsburg 1958–1970.
14. E. B. Kurtz, Jr., Fette Seifen Anstrichmittel 61 (1959) no. 6, 505.
15. E. J. Fischer, W. Presting: Laboratoriumshandbuch für die Untersuchung tech-
nischer Wachs-, Harz- und Ölgemenge, VEB Wilhelm Knapp Verlag, Halle/Saale
1958.
16. L. Ivanovszki: Wachs-Enzyklopädie, Verlag für Chem. Industrie H. Ziolkowsky,
vol. 1 Augsburg, 1954; vol. 2, 1960.
17. W. Piso, Historia Naturalis Brasiliae (1648).
18. J. B. De Moraes: Ensaios sobre à Carnanbeira, Nat. Ölinstitut, Rio de Janeiro 1942.
19. J. B. Nogueira, R. D. Machado: Glossario da Palmeiras, Nat. Ölinstitut, Rio de
Janeiro 1950.
20. Am. Wax Importers and Refiners Assoc.: Blue Book, New York 1973.
21. Chem. Fabrik Dessau, DE 581 130, 1931.
22. Dtsch. Hydrierwerke, DE 658 749, 1932.
23. Degussa, DE 938 501, 1942.
24. E. Eisenring, CH 271 936, 1950.
25. W. A. Sprague, US 1 658 062, 1926.
26. I.G. Farbenind., US 1 989 626, 1932.
27. S. C. Johnson & Son, US 2 527 481, US 2 531 785, 1948.
28. G. Kronbach, Seifen Öle Fette Wachse 99 (1973) 97.
29. L. E. Vandenburg, E. A. Wilder, J. Am. Oil Chem. Soc. 44 (1967) 659.
30. S. Markley, Econ. Bot. 9 (1955) 39.
Vol. 2 WAXES 163
66. VEB Braunkohlekombinat Gustav Sobotka, DD 135 089, 1979 (H. J. Grauen et al.).
67. Hoechst AG, DE 954 903, 1953 (R. Schirmer); DE 957 508, 1955 (R. Schirmer);
DE 1 007 918, 1955 (R. Schirmer); DE 2 855 263, 1979 (A. Riedel); DE 2 915 764,
1979 (H. Stetter); DE 2 915 802, 1979 (H. Stetter); DE 2 915 755, 1979 (H. Stetter).
68. BASF, DE 3 726 514, 1987 (L. Schuster).
69. O. Malitschek: Hoechst Wachse, Ein Leitfaden für Wachsanwender, Hoechst AG,
SBU Wachse und Additive, Werk Gersthofen 1990.
70. R. Gächter, H. Müller: Kunststoffadditive, Hanser Verlag, München 1993.
71. P. Piesold: ‘‘Montanwachs als Kunststoffadditiv,’’ in J. Edenbaum (ed.): Plastics,
Additives and Modifiers Handbook, Van Nostrand Reinhold, New York 1992,
pp. 757–772.
GENERAL REFERENCES
SPECIFIC REFERENCES
80. European Wax Federation (EWF): Glossary of Wax Terminology, Brussels 1981.
81. C. Kajdas, Seifen Öle Fette Wachse 96 (1970) 251–260.
82. C. Kajdas, Seifen Öle Fette Wachse 112 (1986) 471–474.
83. G. Matscholl, Coating 14 (1981) 117–118.
84. I. A. Michailov, Chem. Tech. (Leipzig) 20 (1968) 351–353.
85. ‘‘Assessment and Comparison of the Composition of Food-Grade White Oils and
Waxes Manufactured from Petroleum by Catalytic Hydrogenation Versus Conven-
tional Treatment,’’ CONCAWE Rep. 84/60 (1984) 18–26, 33–46.
86. European Wax Federation (EWF): Specifications of Petroleum Waxes, Brussels 1991.
87. F. V. Synder: Presentation at National Petroleum Refiners Assoc., Houston, Texas
1979.
88. Röhm GmbH, Spektrum 39, company brochure, Darmstadt 1992.
89. CONCAWE 1991: Report no. 91/ 33, 58–62 CONCAWE, Brussels.
90. R. J. Taylor, Oil Gas J. 91 (1993) no. 4, 48–50.
91. D. A. Gudelis et al., Hydrocarbon Process. 52 (1973) no. 9, 141.
92. M. Soudek, Hydrocarbon Process. 53 (1974) no. 12, 64, 66.
Vol. 2 WAXES 165
GENERAL REFERENCES
SPECIFIC REFERENCES
172. Hoechst, EP 602 509, 1993 (R. Hess, H. Voigt, L. Böhm, H.-F. Herrmann).
173. Hoechst, DE-OS 3 743 320, 1987 (A. Winter, M. Antberg, J. Rohrmann).
174. Hoechst, DE-OS 3 904468, 1989 (A. Winter et al.).
175. Mitsui Toatsu Chemicals, EP-A 563 834, 1993 (Y. Sonobe, M. Mizutani, T. Shiomura,
N. Hirayama).
176. BASF, DE 922 618, 1952 (H. Hopff, H. Eilbracht).
177. Eastman Kodak, FR 1 289 766, 1960 (H. W. Coover, Jr., M. A. McCall, J. E. Guillet).
178. W. G. Oakes, R. B. Richards, J. Chem. Soc. 1949, 2929–2935.
179. A. Schwarz, G. Cramer in Polyethylene, 2nd ed., Iliffe & Sons, London; Interscience
Publ., New York 1960, pp. 232–239.
180. BASF, DE-AS 1 940 686, 1969 (H.-G. Trieschmann et al.).
181. Allied Chemical, US 3 345 352, 1964 (J. J. Baron, J. P. Rahus).
182. Eastman Kodak, US 2 835 659, 1957 (J. E. Guillet).
183. Eastman Kodak, US 3 519 609, 1967 (R. L. McConnell, D. A. Weemes).
184. J. E. Guillet, H. W. Coover, SPE J. 16 (1960) 311–314.
185. Eastman Kodak, US 3 562 788, 1967 (D. A. Weemes, R. L. McConnell).
186. H. W. Coover, R. L. Combs, J. E. Guillet, D. A. Weemes, J. Appl. Polym. Sci. 11
(1967) 1271–1281.
187. BASF, DE 1 364 747, 1962.
188. Eastman Kodak, US T 869 011, 1966 (M. O. Bronson).
189. A. Thalhofer in R. Vieweg, A. Schley, A. Schwarz (eds.): Kunststoff-Handbuch,
vol. VI, Polyolefine, Hanser Verlag, München 1969, p. 161.
190. VEB Leuna-Werke, DD 283 730, 1988 (I. Klimek et al.).
191. Mitsui Petrochemical, DE-OS 2 126 725, 1971 (A. Mori et al.).
192. N. M. Emanuel, E. T. Denisov, Z. K. Maizus: Liquid-Phase Oxidation of Hydrocar-
bons, Plenum Press, New York 1967.
193. Hoechst, DE-OS 2 035 706, 1970 (H. Korbanka, K.-H. Stetter).
194. Hoechst, EP 28 384, 1980 (H. Korbanka et al.).
195. Allied Chemical Corp., DE-OS 1 495 938, 1963 (R. W. Bush, C. L. Kehr).
196. Allied Chemical Corp., DE-OS 1 520 008, 1964 (C. L. Kehr).
197. Allied Chemical Corp., DE-OS 1 570 652, 1965 (G. L. Braude, S. T. Phoa).
198. BASF, DE-OS 3 112 163, 1981 (A. Hettche et al.).
199. BASF, DE-OS 3 720 952, 1987 (L. Schuster et al.).
200. BASF, DE-OS 3 720 953, 1987 (L. Schuster et al.).
201. BASF, DE-OS 3 238 652, 1982 (R. W. Reffert et al.).
202. Hoechst, DE-OS 2 201 862, 1972 (H. Korbanka, K.-H. Stetter).
203. Th. Goldschmidt, EP 0 257 506, 1987 (E. Ruf).
204. Th. Goldschmidt, EP 0 258 724, 1987 (E. Ruf).
205. Th. Goldschmidt, EP 0 259 652, 1987 (E. Ruf).
206. Hüls, EP-A 0 451 352, 1990 (A. Kühnle, V. Schrenk, M. Vey, E. Wolf).
207. Mitsui Petrochemical, DE-OS 285 3862, 1978 (A. Tachi, T. Tomoshige, H. Furuta, N.
Matsuzawa).
208. Petrolite Corp., US 3 590 076, 1967 (W. J. Heintzelmann, M. I. Naiman).
209. Eastman Kodak Co., US 4 028 436, 1975 (R. T. Bogan, C. M. Shelton).
210. Mitsui Petrochemical, EP 0 024 034, 1980 (T. Tomoshige, H. Furuta, A. Tachi, N.
Kawamoto).
211. Eastman Kodak Co., US 3 480 580, 1965 (F. B. Joyner, R. L. McConnell).
212. Eastman Kodak Co., US 3 481 910, 1965 (M. O. Brunson).
213. Eastman Kodak Co., US 5 290 954, 1992 (T. D. Roberts, T. W. Smith, C. A. Galvin).
214. K. Rieger, Fette Seifen Anstrichmittel 73 (1971) 231.
215. K. Rieger, G. Grimm, Adhäsion 11 (1977) 311.
168 WAXES Vol. 2
UWE WOLFMEIER
HANS SCHMIDT
FRANZ-LEO HEINRICHS
JOSEF WILDGRUBER
GERD HOHNER
Hoechst Aktiengesellschaft, Werk Gersthofen,
Augsburg, Federal Republic of Germany
GEORG MICHALCZYK
DEA Mineralöl AG,
Hamburg, Federal Republic of Germany
WOLFGANG PAYER
KLAUS BOEHLKE
Hoechst Aktiengesellschaft, Werk Ruhrchemie,
Oberhausen Federal Republic of Germany
WOLFRAM DIETSCHE
BASF Aktiengesellschaft,
Ludwigshafen, Federal Republic of Germany
Vol. 2 ASPHALT 169
ASPHALT
Asphalt [8052-42-4] is defined by the American Society for Testing and Materials
(ASTM) (1) as a dark brown to black cementitious material in which the predo-
minating constituents are bitumens that occur in nature or are obtained in pet-
roleum processing. Bitumen is a generic term defined by ASTM as a class of black
or dark-colored (solid, semisolid, or viscous) cementitious substances, natural or
manufactured, composed principally of high molecular weight hydrocarbons, of
which asphalts, tars, pitches, and asphaltites are typical.
Canadian and European practice (2–4) and geologists and archaeologists in
the United States use bitumen or asphaltic bitumen as a synonym for asphalt,
and apply asphalt to the mixture of bitumen and inorganic matter that is used
for paving purposes. On the other hand, pitches and tars are derived from the
destructive distillation of coal, crude oils, and other organic materials.
ASTM (1) further classifies asphalts or bituminous materials as solids, semi-
solids, or liquids. Semisolid and liquid asphalts predominate in commercial prac-
tice today. Prior to 1907, the principal portion of asphalt used occurred naturally;
since that time most asphalts have been produced from the refining of petroleum.
Air-blown petroleum asphalts (5) of diverse hardness became available in the early
1900s, and as their use increased, the demand for native products diminished.
More recently, asphalts have been defined as the manufactured materials
that are produced during petroleum processing whereas bitumens have been
defined as the naturally occurring materials (5,6). That is, petroleum and related
materials can be divided into various class subgroups:
The term tar sands is a misnomer; tar is a product of coal processing. Oil
sands is also a misnomer but equivalent to usage of ‘‘oil shale.’’ Bituminous
sands is more correct; bitumen is a naturally occurring asphalt. Asphalt is a
product of a refinery operation, usually made from a residuum. Residuum is
the nonvolatile portion of petroleum and often further defined as atmospheric
(bp > 350 C) or vacuum (bp > 565 C). For convenience, the terms ‘‘asphalt’’
and ‘‘bitumen’’ will be used interchangeably in this article.
Asphalts characteristically contain very high molecular weight molecular
polar species, called asphaltenes, which are soluble in carbon disulfide, pyridine,
aromatic hydrocarbons, chlorinated hydrocarbons, and tetrahydrofuran.
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0119160819160509.a01
170 ASPHALT Vol. 2
2. Manufactured Materials
Since the early 1900s most of the asphalts produced from the refining of petro-
leum have been used primarily in paving and roofing applications. The advent of
motorized transportation led to increased asphalt manufacture from petroleum
in order to provide binders for hard-surfaced pavements.
Petroleum asphalts, compared to native asphalts, are organic with only
trace amounts of inorganic materials. They derive their characteristics from
the nature of their crude origins with some variation possible by choice of
manufacturing process. Although there are a number of refineries or refinery
units whose prime function is to produce asphalt, petroleum asphalt is primar-
ily a product of integrated refineries (Fig. 1). Crudes may be selected for
these refineries for a variety of other product requirements and the asphalt
(or residuum) produced may vary somewhat in characteristics from one
172 ASPHALT Vol. 2
Louisiana Bachaquero
Gravity, API 13.1 2.8
Sulfur, wt % 0.9 3.7
Nitrogen, wt % 0.4 0.6
Carbon, wt % 15.8 27.5
Nickel, ppm 20 100
Vanadium, ppm 8 900
Pour point, 8C 55
Fig. 1. Residua are obtained by removal of the volatile constituents of the feedstock at
atmospheric pressure or at reduced pressure and can be converted to asphalt by various
methods.
Table 2. Properties of Various Residua Produced from Tia Juana, Venezuela Light Crude Oil
Property Values
boiling range, 8C whole crude >220 >295 >345 >400 >455 >505
yield on crude, vol % 100.0 70.2 57.4 48.9 39.7 31.2 17.9
gravity, API 31.6 22.5 19.4 17.3 15.1 12.6 7.1
specific gravity 0.8676 0.9188 0.9377 0.9509 0.9652 0.9820 1.0209
sulfur, wt % 1.08 1.42 1.64 1.78 1.93 2.12 2.59
carbon residue 6.8 8.1 9.3 11.2 13.8 21.6
(Conradson), wt %
nitrogen, wt% 0.33 0.39 0.45 0.60
pour point, 8C 20 9 1 7 16 24 49
viscosity
173
kinematic,
mm2 =sð ¼ cStÞ
388C 10.2 83.0 315 890 3100
998C 9.6 19.6 35.0 77.0 220 7959
Furol (SFS), s
508C 70.6 172 528
998C 37.6 106 3760
Universal (SUS), 57.8 96.8 165 359 1025
s at 998C
metals
vanadium, ppm 185 450
nickel, ppm 25 64
iron, ppm 28 48
174 ASPHALT Vol. 2
Fig. 2. Different crude oils contain different amounts of residua and the properties of
specific ‘‘cut point’’ residua may be different, eg, for the 5658C residua:
2.1. Straight Run Asphalt. In crude-oil refining, the crude oil at 340–
4008C is injected into a fractionating column (5,6,19,20). The lighter fractions are
separated as overhead products and the residuum is straight-reduced asphalt.
Crude oil containing about 30% asphalt can be refined completely in an
atmospheric unit to an asphalt product. However, most crude oil cannot be
distilled satisfactorily to an asphalt product at atmospheric pressure because
of the presence of substantial proportions of high boiling gas oil fractions.
Thus, as a supplement to the atmospheric process, a second fractionating
tower (a vacuum tower) is added (Fig. 1).
This two-stage process is particularly applicable to crude oils containing
15–30% asphalt. Vacuum distillation units are continuous flow-through units
and include a heating unit, the vapor–liquid flash separation zone, the fractio-
nating zone, and the vacuum producing equipment. The heated feedstock is
pumped into a flash zone where the more volatile constituents vaporize.
The asphalt is the nonvolatile residual fraction which is removed continuously.
Manufacturing asphalt by straight reduction does not change the chemical
nature of asphalt, other than its viscosity.
2.2. Propane Asphalt. As noted above, crude oils contain different
quantities of residuum (Fig. 2) and, hence, asphalt. Asphalt is also a product of
the propane deasphalting and fractionation process (5,6,21,22) which involves
the precipitation of asphalt from a residuum stock by treatment with propane
under controlled conditions. The petroleum charge stock is usually atmo-
spheric-reduced residue from a primary distillation tower.
Propane is usually used in this process although propane–butane mixtures
and pentane have been used with some variation in process conditions and
Vol. 2 ASPHALT 175
hardness of the product. Propane deasphalting is used primarily for crude oils of
relatively low asphalt content, generally <15%. Asphalt produced from this pro-
cess is normally blended with other asphaltic residua for making paving asphalt.
The process (Fig. 3) is a countercurrent liquid–liquid extraction. The feed-
stock is introduced near the top of an extraction tower and the liquid propane
near the bottom, using solvent-to-oil ratios from 4:1 to 10:1. The deasphalted
oil–propane solution is withdrawn overhead and the asphalt from the bottom,
and each is subsequently stripped of propane.
Temperature, solvent ratio, and pressure each have an effect upon the split
point or yield of the oil and asphalt components (Table 3). Contrary to straight
reduction which is a high temperature and low pressure process, propane deas-
phalting is a low temperature and high pressure process.
There are small differences in the properties of asphalts prepared by
propane deasphalting and those prepared by vacuum distillation from the
same stock (Fig. 4). Propane deasphalting also has the ability to reduce a resi-
duum even further and produce an asphalt product having a lower viscosity,
higher ductility, and higher temperature susceptibility than other asphalts
(23). However, such properties might be anticipated to be very much crude
oil dependent. Propane deasphalting is conventionally applied to low asphalt-
content crudes, which are generally different in type and source from those
processed by distillation of higher yield crudes.
2.3. Air-Blown Asphalt. Air-blowing is an exothermic process that is
dependent upon several variable process parameters (24). In the process, an
asphalt (flux) is converted to a harder product by air contact at 200–2758C.
Dehydrogenation and polymerization are involved, and oxygen is not retained
in the asphalt product except in a very minor amount (25). In theory, the oxygen
in the air combines with the hydrogen in the asphalt to evolve water vapor. This
leaves unsaturation for cross-linking. However, the chemistry may be much
176 ASPHALT Vol. 2
Table 3. Typical Yields and Characteristics of Products Obtained from the Propane
Deasphalting Process
Crude source
Process Pa. Okla. Ill. Le Duc, Canada Iraq
reduced crude
gravity, 8APIa 25.1 19.3 13.7 12.1 7.1
kinematic viscosity at 98.98C, 47.2 82.8 378 441 1050
mm2 =sð¼ cStÞ
Saybolt viscosity, s 220 385 1760 2050 4880
crude, vol% 24.0 26.1 11.3 15.2 16.4
deasphalted oil
yield, vol % 83.7 75.6 50.7 47.1 34.1
gravity, 8APIa 26.6 23.6 23.4 22.7 20.2
kinematic viscosity at 98.98C, 30.4 29.8 40.7 42.6 38.2
(mm2 =s ð¼ cStÞ)
Saybolt viscosity, s 144 141 190 199 179
precipitated asphalt
yield, vol% 16.3 24.4 49.3 52.9 65.9
specific gravity at 15.68C 0.953 1.023 1.030 1.054 1.064
softening point 62.2 62.8 80 62.8
(ring and ball), 8C
penetration of 100 g 23 21 0 16
at 258C for 5 s, mm/10
operating conditions
throughput, kL/db 800
charge stock reduced crude
gravity charge, 8APIa 20.0
temperature of deasphalting 65
tower, 8C
solvent-to-oil ratio, by volume 6:1
pressure, kPac 2170 (300)
API gravity at 15:6 C ¼ sp gr at141:5
a
15:6=15:6 C 131:5.
b
To convert kL/d to barrels/d, multiply by 6.25.
c
To convert kPa to psi, multiply by 0.145.
more complex than the simple theory predicts (24). For example, asphaltenes are
produced during the air-blowing process (Table 4) and the evidence suggests
that those constituent species of the asphalt which have higher proportions of
heteroatoms (ie, nitrogen, oxygen, and sulfur) react preferentially with oxygen
leaving less polar substances unreacted (26,27). This also raises the specter of
incompatibility of the resulting products. For example, if the more polar species,
ie, the resins, are preferentially oxidized to produce more asphaltenes, then the
means by which the asphaltenes are dispersed in the oily medium is removed.
Incompatibility should be anticipated.
A variety of other substances can provide the same reaction: sulfur yields
hydrogen sulfide; chlorine yields hydrogen chloride. In some cases, some of the
bonds created are quite weak resulting in, after an induction period, a phenom-
enon termed ‘‘fallback.’’ When fallback occurs, usually at a time when the har-
dened or oxidized asphalt is stored at or near the original processing or
reaction temperature, softening of the asphalt is the result (28–31).
Vol. 2 ASPHALT 177
Table 5. Air-Blowing of Fluxes With and Without Ferric Chloride and Phosphorus
Pentoxide Catalysts
California stock Arkansas stock
No 0.8% 0.5% No 2.2%
Property catalyst FeCl3 6 H2 O P2O5 catalyst FeCl3 6 H2 O 1.0% P2O5
blowing 221.1 221.1 221.1 254.4 237.8 237.8
temperature, 8C
blowing time, min 135 44 103 105 46 128
softening pointa , 8C 87.8 86.1 86.1 86.7 83.3 87.8
penetration of 100 g 27 51 53 27 62 59
at 258C for 5 s,
mm/10
penetration of 200 g 13 30 28 15 38 38
at 08C for 60 s,
mm/10
penetration of 50 g at 74 48 109 110
468C for 5 s, mm/10
ductility at 258C, cm 3.5 3.4 4.8 3.3 3.9 5.3
a
Ring and ball method.
high specific gravity, low viscosity, and high temperature susceptibility, and they
contain cokelike bodies (carbenes) as indicated by the Oliensis test (38). Thermal
asphalts have significant application as saturants for cellulosic building products
such as insulation boards, brick-finish siding, and fiber soil pipes.
The thermal cracking process (Fig. 7) involves preheating of the charge
to 480–6008C and then discharging into a reaction vessel or chamber under
Table 6. Three Types of Liquid Asphalt Made by Cutting Back with Petroleum Diluents
Type asphalt Diluent type Diluent, % Viscosity rangea , mm2 =s ð¼ cStÞ
slow curing (SC) gas oil 0–50 70–6000
medium curing (MC) kerosene 15–45 30–6000
rapid curing (RC) naphtha 15–45 70–6000
a
The liquid cutback asphalts are prepared in a number of viscosity grades, ranging generally from 70
to 6000. The grade number indicates the viscosity at 608C.
pressures up to 1480 kPa (200 psig). Cracking decomposition takes place to form
both lighter and heavier products, including a certain amount of fixed carbon.
Distillation is used to separate the overall products into gas, gasoline, middle
distillate, and an asphaltic residuum.
2.5. Blended Asphalt. Any particular, refinery may stock two grades of
asphalt, one at each end of the viscosity spectrum of the entire product grade
requirements. Intermediate grades are prepared by blending (proportioning)
the extremes (39). The preparation of asphalts in liquid form by blending (cutting
back) an asphalt with a petroleum distillate fraction is customary. There are
three general types of cutback asphalt which differ mainly in the diluent used
(Table 6). The slow-curing type is often called road oil, Road oil which is made
by direct reduction, but most are commonly blended. The medium curing and
rapid curing products usually have four grades within a given type and differ
mainly in the amount of diluent used and in the kinematic viscosity of the
blended product.
Preparation is accomplished by simple blending of the diluent into the hot
base asphalt. This is generally accomplished in tanks equipped with coils for air
agitation or with a mechanical stirrer or a vortex mixer. Line blending in a batch
circulation system or in a continuous fashion (40) is used where the volume
produced justifies the extra facilities. A continuous, line-blending system is
applicable to the manufacture of cutback asphalts and asphalt cements (Fig. 8).
2.6. Asphalt Emulsions. An emulsion is one immiscible liquid dis-
persed in the other in the form of very fine droplets from about 1–25 mm and
an average of 5 mm diameter. In the most common asphalt emulsion, the oil-in-
water type, the asphalt is the dispersed (internal) phase, and water is the contin-
uous (external) phase. Most of the important properties of an emulsion are
dependent upon the amount and type of emulsifying agent used. The more pop-
ular cationic emulsions (41,42) are made with cation-active agents which provide
emulsion droplets with positive charges. The anionic agent is usually the sodium
or potassium salt of a fatty acid (43). Nonionic cellulose derivatives are also used
to increase the viscosity of the emulsion if needed. The acid number of asphalt is
an indicator of its emulsifiability and reflects the presence of high molecular
weight asphaltogenic or naphthenic acids. Diamines, frequently used as cationic
agents, are made from the reaction of tallow acid amines with acrylonitrile,
followed by hydrogenation.
Emulsion processes (Fig. 9) vary and the emulsions are made in rapid-,
medium-, and slow-setting types for diverse application techniques in the road-
building industry. Emulsions are most commonly formulated with rapid-setting
properties and are designated as a spraying type, ie, for road seal-coating
182 ASPHALT Vol. 2
Fig. 10. The relationship between a crude oil, two residua from the crude oil, and the
propane asphalt.
184 ASPHALT Vol. 2
Fig. 11. Simplified representation of the separation of petroleum into six principal
fractions.
Vol. 2 ASPHALT 185
As this ratio increases so does the gel character of the asphalt cement, and
its colloidal stability increases. The properties of the asphalt (softening point,
ductility, embrittlement temperature, thermal susceptibility, elastic recovery,
shearing susceptibility, etc) vary significantly with the colloidal instability
index and hence with composition.
Many investigations of relationships between composition and properties
take into account only the concentration of the asphaltenes, independently of
any quality criterion. However, a distinction should be made between the asphal-
tenes which occur in straight run asphalts and those which occur in blown
asphalts. Remembering that asphaltenes are a solubility class rather than a
distinct chemical class means that vast differences occur in the make-up of
this fraction when it is produced by different processes.
186 ASPHALT Vol. 2
ratio). The significance of metals in the behavior of asphalts is not yet well under-
stood or defined.
3.3. Molecular Weight. The molecular weights of the individual frac-
tions of asphalt have received more attention, and have been considered to be
of greater importance, than the molecular weight of the asphalt itself. Asphal-
tenes display a wide range of molecular weights, from 500 to at least 2500,
depending upon the method of measurement. Asphaltenes associate in dilute
solution in nonpolar solvents (69) giving higher molecular weights than is actu-
ally the case on an individual molecule basis. The molecular weights of the resins
are somewhat lower than those of the asphaltenes and usually fall within the
range 500 to 1000 (70). This is not only because of the absence of association
but also because of a lower absolute molecular size. The molecular weights of
the oil fractions (ie, the asphalt minus the asphaltenes and minus the resins)
is usually less than 500, often 300 to 400.
3.4. Acid Number. Asphalt contains a small amount of organic acids
and saponifiable material and the acid values of asphalt, 0.1–2.8 mg potassium
hydroxide per gram of asphalt, are related to emulsification conditions and ease
of dispersion. The acid content is largely determined by the percentage of
naphthenic (cycloparaffinic) acids of higher molecular weight that are originally
present in the crude oil. With increased hardness, asphalt from a particular
crude oil normally decreases in acid number as more of the naphthenic acids
are removed during the distillation process.
3.5. Rheology. Asphalt is a viscoelastic material whose rheological
properties reflect crude type and, to a lesser extent, processing. The ability of
the asphalt to perform under many conditions depends on flow behavior
(71,72). Asphalt films or coatings showing no appreciable change from original
conditions are usually desired, ie, they should allow some structural movement
without permanent deformation. Some procedures for measuring viscoelastic
properties of asphalt have been described (2,73–75). Liquid asphalts, in effect,
substitute diluent (or water emulsification) for heat to allow pumping, spraying,
or application of the material at lower temperature.
The viscosity of hydrocarbons and temperature are related by the Walther
equation (76):
where Z0 is the limiting viscosity at low shear rates, T is the absolute tempera-
ture, a and m are constants reflecting the intercept level and slope (or measure of
susceptibility of viscosity to temperature, respectively). On the other, hand, the
general relationship for viscosity, temperature, and shear rate is
log log 100Z0 1 CDN ¼ a m log T
materials which are usually graded by softening point) can be generated by the
use of the equation (Fig. 12). At lower temperatures (608C and lower) and/or
higher shear rates, which are typical of asphalt service conditions after incor-
poration in a roof or pavement, semisolid and solid asphalts display an increas-
ing elastic component which relates viscosity with shear rate (73,77,78). The
constant high viscosity at lower shear rates is the limiting viscosity. Viscosities
in the area where viscosity changes with shear rate are generally termed appar-
ent viscosities.
Curves for the viscosity data, when displayed as a function of shear rate
with temperature, show the same general shape with limiting viscosities at
low shear rates and limiting slopes at high shear rates. These curves can be com-
bined in a single master curve (for each asphalt) employing vertical and horizon-
tal shift factors (77–79). Such data relate reduced viscosity (from the vertical
shift) and reduced shear rate (from the horizontal shift).
A number of viscometers have been developed for securing viscosity data at
temperatures as low as 08C (58,59). The most popular instruments in current use
are the cone plate (ASTM D3205), parallel plate, and capillary instruments
(ASTM D2171 and ASTM D2170). The cone plate can be used for the determina-
tion of viscosities in the range of 10 to over 109 Pa s (1010 P) at temperatures
of 0–708C and at shear rates from 103 to 102 s1 . Capillary viscometers are com-
monly used for the determination of viscosities at 60–1358C.
Tests developed for measurement of viscoelastic properties are directly
usable in engineering relations. Properties can be related to the inherent struc-
ture of bituminous materials. The fraction of highest molecular weight, the
asphaltenes, is dispersed within the asphalt and is dependent upon the content
and nature of the resin and oil fractions. Higher aromaticity of the oil fractions
or higher temperatures leads to viscous (sol) conditions. A more elastic (gel)
Vol. 2 ASPHALT 189
condition results from a more paraffinic nature and is indicated by large elastic
moduli or, empirically, by a relatively high penetration at a given softening
point. Empirically, the penetration index PI(2) and penetration temperature sus-
ceptibility PTS have been used to measure the degree of dispersion.
30
PI ¼ 10
1 þ 90 PTS
This equation is based on the approximation that the penetration is 800 at the
softening point, but the approximation fails appreciably when a complex flow is
present (80,81). However, the penetration index has been, and continues to be,
used for the general characteristics of asphalt; for example asphalts with a PI
less than 2 are considered to be the pitch type, from 2 to þ2, the sol type,
and above þ2, the gel or blown type (2). Other empirical relations that have
been used to express the rheological-temperature relation are fluidity factor; a
Furol viscosity V, at 1358C and penetration P, at 258C, relation of (V–P)P/100;
and penetration viscosity number PVN again relating the penetration at 258C
and kinematic viscosity at 1358C (82,83).
Asphalts develop an internal structure with age, steric hardening (3), in
which viscosity can increase upon aging without any loss of volatile material
(73, 83). Those with a particularly high degree of gel structure exhibit thixotropy.
3.6. Durability. The term ‘‘durable’’ has several meanings, but in the
present context it is used to describe an asphalt that possesses the necessary che-
mical and physical properties required for the specified pavement performance,
being resistant to change during the in-service conditions that are prevalent dur-
ing the life of the pavement.
Asphalts are used as protective films, adhesives, and binders because of
their waterproof and weather-resistant properties. One particularly valuable
property, the ability of the asphalt to undergo movement without fracture, can
occur because of the viscous (sol) nature. In addition, asphalts have long and con-
tinuous satisfactory service because of their slow rate of hardening from heat,
oxidation, fatigue, and weathering (84–87). Exposed asphalt films harden par-
tially from a loss of volatile oils and to a greater extent from the formation of
additional asphaltene material, with the concurrent loss of petrolene-maltene
constituents through oxidation. Such chemical change undoubtedly is catalyzed
by uv irradiation (86–90). The oxidation may form water-soluble degradation
products which are removed from the asphalt film (89,90), but the conversion
of the petrolene–maltene constituents to lower molecular weight asphaltene
material (27) cannot be ignored.
Failure of an asphalt film is reflected in the appearance of typical crack
patterns because of a decrease in the plasticity of the asphalt through insuffi-
cient lower molecular weight constituents remaining in the continuous phase.
Other manifestations of failure are loosening of the bond, such as loosening of
190 ASPHALT Vol. 2
granules in roofing shingles to actual loss of the film, or the binder in pavement
mixtures might become friable and crack. A stiffer asphalt, under uniform
loading conditions, could reduce pavement deflection, extend fatigue life, and
allow less flow deformation. A softer material would normally allow a longer
weathering life before the maltene–asphaltene composition becomes critical
in service. Usually the softest material allowed by initial service needs is
selected.
The water resistance of asphalt films is also a manifestation of durability.
Asphalts that have a low content of soluble salts show a low water absorption.
The pickup of water is primarily a surface phenomenon that causes the film to
soften leading to blistering. Even with a high rate of absorption, asphalt films
show little loss of bond to surfaces on continued immersion in water, and con-
tinue to protect metals from corrosion for long periods of time. Bacteria and
fungi can attack the very low molecular weight portion of bituminous materials.
However, deterioration of the harder bituminous films as a result of such agents
seems insignificant (91,92).
Mineral fillers are often added to asphalts to influence their flow properties,
reduce costs, and are commonly used as stabilizers in roofing coatings at concen-
trations up to 60 wt% (93). Mineral-filled films show improved resistance to
flow at elevated temperatures, improved impact resistance, and better flame-
spread resistance. Fillers may increase the water absorption of asphalts. Fillers
commonly used are ground limestone, slate flours, finely divided silicas, trap
rocks, and mica; they are often produced as by-products in rock-crushing opera-
tions. Opaque fillers offer protection from weathering. Asbestos (qv) filler has
special properties because of its fibrous structure, high resistance to flow, and
toughness. It has been used in asphalt paving mixes to increase the resistance
to movement under traffic (94) and in roofing materials for fire-retardant
purposes.
Fig. 13. Relation of physical properties to the yield of residuum from Tia Juana,
Venezuela light crude oil.
During the 1920s the following three national specifications for asphalt
cements were published: (1) Federal specifications, adopted in 1925; (2) the
American Association of State Highway Transportation Officials (AASHTO)
specifications, adopted in 1924, revised in 1926; and (3) ASTM specifications,
adopted in 1922 to 1926, withdrawn in 1939, and re-issued in 1947.
The Federal specifications stipulated that only those asphalts that had been
demonstrated by service tests as satisfactory for the intended use would be
accepted. The specifications also indicated the type and location of construction
and the relative amount of traffic for each of the grades. The AASHTO specifica-
tions indicated that the use of each grade depended on the type of road, climate,
and traffic. The ASTM suggested the type of construction for which each grade
would be used.
With minor exceptions the requirements for the physical and chemical
properties of asphalt were essentially the same for the three national specifica-
tions and included: penetration and ductility at 258C; flash point; % loss at
1638C; penetration of residue as a % of original; solubility in carbon disulfide;
solubility in carbon tetrachloride; specific gravity at 258C; and softening point.
Remembering that the properties of residua vary with cut-point (Table 2),
ie, the vol% of the crude oil (Fig. 13) helps the refiner produce asphalt of a
192 ASPHALT Vol. 2
specific type or property. There are several properties that are usually controlled
in asphalt specifications:
Property Test
safety flash point
purity solubility, ash, water content
composition naphtha insolubles, various separation techniques, distillation
tests, homogeneity, wax content
rheology penetration, viscosity, softening point, float, ductility, flow,
impact or shock, break point, temperature susceptibility ratios
durability thin-film oven, aging index, weathering
density or specific specific gravity
gravity
special properties bond or adhesion, compatibility, stain, storage stability,
chemical resistance, etc
Asphalts are usually specified in several grades for the same industry, dif-
fering in hardness or viscosity (95). However, with the changing nature and com-
position of crude oil feedstocks over the past two decades, performance and
supply factors are also an important consideration (95).
Specifications for paving asphalt cements usually include five grades differ-
ing in either viscosity or penetration level at 608C (Table 8) (ASTM D496).
Susceptibility of viscosity to temperature is usually controlled in asphalt cement
whether they are likely to bleed, show strike-through, or disbond under stress at
the coating felt interface.
Distillation (ASTM D402). Approximate amounts of volatile constituents
are determined by this test which is particularly applicable to cutback asphalt
and road oils.
Ductility (ASTM D113). The ductility of an asphalt is expressed as the
distance in cm which a standard briquet can be elongated before breaking.
Ductility is a combination of flow properties and reflects both cohesion and shear
susceptibility.
Emulsified Asphalts (ASTM D244). This standard covers a variety of
tests for the composition, handling, nature and classification, storage, use, and
specifying of asphalt emulsions used primarily for paving purposes.
Flash Point (ASTM D92). The Cleveland open cup method is most com-
monly used although the Tag open cup (ASTM D3143) is applicable to cutbacks.
Flash point is an indication of fire hazard and the test is frequently used to
indicate whether a given product has been contaminated with materials of
lower flash point.
Float Test (ASTM D139). The consistency properties of an asphalt at a
very low applied force are indicated by this test. The test is normally used for
those asphalts that are too soft for the penetration test.
Homogeneity. Visual tests for homogeneity are generally used to judge
uniformity.
Penetration (ASTM D5). This is a commonly used consistency test. It
involves the determination of the extent to which a standard needle penetrates
a properly prepared sample of asphalt under definitely specified conditions of
temperature, load, and time. The distance that the needle penetrates in units
of mm/10 measured from 0 to 300, is the penetration value. Soft asphalts have
high penetration values.
Sampling (ASTM D140). This standard provides guidance for the sam-
pling of asphalts, liquid and semisolid, at point of manufacture, storage, or delivery.
Softening Point (Ring and Ball Method, ASTM D36). The softening
point of an asphalt may be defined as that temperature at which an asphalt
attains a particular degree of softness under specified conditions of test. It is com-
monly used to classify industrial and roofing asphalt grades. A method (ASTM
D2398) has been developed to provide a single bath medium for the full range
covered by the two media in ASTM D36.
Specific Gravity (ASTM D70). For solid and semisolid asphalts a
pycnometer is generally used and a hydrometer is applicable to liquid asphalts
(ASTM D3142).
Spot Test (AASHTO T102). The test distinguishes asphalts that contain
bodies poorly tolerated by the asphalt system.
Stain (ASTM D1328). This test measures the amount of stain on paper
or other cellulosic materials by asphalt. Variations of the cigarette paper stain
procedure include the Barber stain, usually conducted at 54.48C and 3.9 MN
(400 g-force). Talc stain tests are also used.
Temperature Susceptibility. This is the term used to designate the
change in consistency of an asphalt with changes in temperature. The Walther
equation slope constant m can be used as a fundamental measure.
Vol. 2 ASPHALT 195
5. Uses
6. Modified Asphalts
The deterioration of asphalt pavements under heavy traffic loads and/or under
hostile weather conditions, coupled with the inability of many modern crude
oil feedstocks to meet the challenge of producing specification-grade asphalts,
has resulted in the search for asphalt modifiers. These are materials that can
be added to the asphaltic binder to improve service life through enhancement
of a particular property or providing a stronger bond between the binder and
the aggregate thereby reducing the susceptibility of the mix to, for example,
water damage (112,113).
A variety of materials has been proposed to modify the properties of
asphaltic binders to enhance the properties of the mix (112), including fillers
and fibers to reinforce the asphalt–aggregate mixture (114), sulfur to strengthen
or harden the binder (115,116), polymers (98,117–121), rubber (122), epoxy–
resin composites (123), antistripping agents (124), metal complexes (125,126),
and lime (127,128). All of these additives serve to improve the properties
of the binder and, ultimately, the properties of the asphalt–aggregate mix.
The addition of fossil fuel derived materials has also been investigated. Of
particular interest are the materials which contain functional groups that are
capable of enhancing the asphalt–aggregate bond. Modifiers such as those iso-
lated from the products of coal liquefaction (129) and the components of shale
oil and even shale oil residua (130) have been used. The latter is especially inter-
esting since the concept focuses on the use of basic nitrogen functions in the shale
oil which enhance the adhesion of the asphalt to the aggregate. Test strips
Vol. 2 ASPHALT 199
of shale oil modified asphalt have been laid on highways in various parts of the
United States.
7. Aggregates
Although much of the focus tends to be on the composition and behavioral char-
acteristics of the asphalt (ie, the binder or cement), the properties of the aggre-
gate also play a significant role in determining the ultimate properties of the
asphalt–aggregate mix. There are many factors to consider and they include:
200 ASPHALT Vol. 2
the functionalities in the asphalt (131), absorption of the asphalt into the
pores of the aggregate (132), the aging properties of the asphalt when it is
on the aggregate (133), as well as methods for coating the asphalt on to the
aggregate (134).
Attempts have also been made to carry out surface modifications of the
aggregate to enhance interactions with the asphalt (135) and other workers
have made attempts to measure or predict the strength and type of asphalt–
aggregate bonds (136,137). However, it must also be remembered that mix
design parameters play an important role in determining the performance of
asphalt–aggregate mixes (138–142).
The word asphalt has been carelessly used in that it is not adequately differen-
tiated from thermally degraded materials, especially coal tar and derivatives. It
is essential to differentiate asphalt from these materials which contain known
carcinogens and health hazards. For this reason, the use of cracked asphalts
must be treated with caution.
There are a variety of tests for asphalt emissions (Table 11). No signifi-
cant air pollution problems are associated with emissions from hot paving
operations using several asphalt cements (143). In fact, the concentrations of
gaseous substances and emissions from paving asphalt cement have been
found to be in very low concentration and within existing EPA and OSHA
standards, even when the ambient air sampling was done under confined
conditions (144–146). In addition, the high molecular weight polynuclear aro-
matic constituents that occur in asphalt have been studied as health hazards
(147–151). The general conclusion from these studies is that surveillance
should be continued although asphalt has not been shown to be a material of
significant hazard.
Steps to minimize potential safety hazards in the handling of asphalt are set
forth by the American Petroleum Institute (152). These include: (1) sudden pressure
increases from hot asphalt in contact with moisture in enclosed tanks or transports;
(2) exposure to air at 1508C or above; (3) local overheating above heating coils;
flashing of asphalt volatiles in the presence of an ignition source or possible auto-
ignition; and (4) hydrogen sulfide from high temperature operations.
BIBLIOGRAPHY
‘‘Asphalt’’ in ECT 1st ed., Vol. 2, pp. 164–199, by A. P. Anderson, Shell Oil Company, and
V. A. Endersby, Shell Development Company; in ECT 2nd ed., Vol. 2, pp. 762–806, by A. J.
Hoiberg, The Flintkote Company, L. W. Corbett, Esso Research & Engineering Company,
and R. B. Lewis, Shell Oil Company; in ECT 3rd ed., Vol. 3, pp. 284–327, by J. V. Evans,
Amoco Oil Company.
CITED PUBLICATIONS
1. Terms Relating to Materials for Roads and Pavements, ASTM Designation D8,
American Society for Testing and Materials, Philadelphia, Pa., 1976.
2. P. H. Pfeiffer, The Properties of Asphaltic Bitumen, Elsevier Publishing Company,
Amsterdam, The Netherlands, 1950.
3. H. Abraham, Asphalts and Allied Substances, 6th ed., Vol. 1, Van Nostrand Co. Inc.,
Princeton, N.J., 1960, Chapt. 2.
4. A Glossary of Petroleum Terms, The Institute of Petroleum, London, UK, 1961.
5. J. G. Speight, Fuel Science and Technology Handbook, Marcel Dekker, Inc.,
New York, 1990.
6. J. G. Speight, The Chemistry and Technology of Petroleum, 2nd ed., Marcel Dekker,
Inc., New York, 1991.
7. J. S. Miller, Jr., in E. A. Dunstan, ed., The Science of Petroleum, Vol. 4, Oxford
University Press, New York, 1938.
8. B. Redwood, Treatise on Petroleum, 5th ed., Vol. 1, Griffin, London, 1976, plate 7,
pp. 53, 208–212.
9. Herodotus, Persian Wars, Vol. 6, Modern Library, New York, p. 119.
10. A. Sequin, Proc. 2nd World Pet. Congr. 4, 56 (1937).
11. R. J. Forbes, Bitumen 5, 9, 41, 63 (1935).
12. R. J. Forbes, Bitumen and Petroleum in Antiquity, E. J. Brill, Leiden. The
Netherlands, 1936.
13. W. A. Attwooll and D. C. Broome, Trinidad Lake Asphalt, 2nd ed., Baynard Press,
London, UK, 1954.
14. Proc. Assoc. Asphalt Paving Technol. 43A, (1974).
15. Statistical Abstracts of the United States, U.S. Department of Commerce, Washington,
D.C., 1961, p. 716.
16. W. R. Jillson, Pan Am. Geol. 41, 251 (1924).
17. N. Berkowitz and J. G. Speight, Fuel 54, 138 (1975).
18. J. C. Petersen, Fuel Sci. Technol. Int. 6, 255 (1988).
19. W. L. Nelson, Petroleum Refining Engineering, 4th ed., McGraw-Hill Book Co., Inc.,
New York, 1958.
20. C. H. Nielson, Distillation in Practice, Chapman and Hill Ltd., London, UK, 1956.
21. H. M. Chelton and R. N. Traxler, Proc. 5th World Pet. Congr., New York, 1959, p. 247.
22. A. E. Leybourne, III and H. E. Schweyer, Ind. Eng. Chem. Process Des. Dev. 1, 127
(1962).
23. I. Ishai and Y. A. Tuffour, Proc. Assoc. Paving Technol. 56, 599 (1987).
24. M. A. Quddus and F. Khan, Fuel 69, 509 (1990).
25. L. W. Corbett and R. E. Swarbrick, Proc. Assoc. Asphalt Paving Technol. 29, 104 (1960).
26. A. J. Hoiberg and W. E. Garris, Jr., Ind. Eng. Chem. Anal. Ed. 16, 294 (1944).
27. S. E. Moschopedis and J. G. Speight, Fuel 57, 235 (1978).
28. W. L. Nelson, Oil Gas J. 53(16), 158 (Aug. 23, 1954).
202 ASPHALT Vol. 2
109. Mix Design Methods for Asphaltic Concrete, Manual Series No. 2, The Asphalt
Institute, College Park, Md., 1974.
110. R. N. Traxler, Asphalt, Reinhold Publishing Co., New York, 1961.
111. Roofing and Shingles; Asphalt-Prepared, Mineral Surfaced, Federal Specification
SS-R-521, General Services Administration, Washington, D.C.
112. W. A. Higgins, Trans. Res. Rec. 1115, 89 (1987).
113. J. A. Scherocman, K. A. Mesch, and J. J. Proctor, in Ref. 46, p. 213.
114. U.S. Pat. 4,634,622 (1987), K. L. Jenkins and co-workers (to Manville Corp.).
115. U.S. Pat. 4,769,288 (1988), D. Saylak (to Texas A & M University).
116. W. Akili and G. J. Courval, in Ref. 71.
117. U.S. Pat. 4,639,273 (1987), D. W. Gilmore and T. G. Kugele (to Morton Thiokol Inc.).
118. L. D. Coyne, Proc. Assoc. Asphalt Paving Technol. 57, 545 (1988).
119. C. Lenoble, Am. Chem. Soc. Div. Pet. Prepr. 35(3), 541 (1990).
120. T. S. Shuler, J. H. Collins, and J. P. Kirkpatrick, in Ref. 71.
121. H. W. Murray, G. N. King, and J. B. Prudhomme, in Ref. 71.
122. H. B. Takallon, R. G. Hicks, and D. C. Esch, in Ref. 99, p. 68.
123. C. H. Lhymn, J. Mater. Sci. 8, 1019 (1989).
124. H. W. Busching and co-workers, in Ref. 46, p. 120.
125. U.S. Pat. 4,721,529 (1988), T. E. Mullins (to Unichem International Inc.).
126. J. Epps and co-workers, in Ref. 99, p. 106.
127. T. W. Kennedy, Bulletin No. 325, National Lime Association, Arlington, Va.
128. M. Stroup-Gardiner and J. Epps, in Ref. 112, p. 12.
129. H. H. Yoon and A. R. Tarrer, Am. Chem. Soc. Div. Fuel Chem. Prepr. 33(4), 975
(1988).
130. U.S. Pat. 4,325,738 (1982), H. Plancher and J. C. Petersen (to U.S. Department of
Energy).
131. C. W. Curtis, Y. W. Jeon, and D. J. Clapp, in Ref. 45, p. 1225.
132. J. A. Guin, H. S. Chang, and S. C. Yen, Am. Chem. Soc. Div. Fuel Chem. Prepr. 35(3),
370 (1990).
133. J. F. McKay, in Ref. 132, p. 496.
134. U.S. Pat. 4,692,350 (1987), A. R. Clarke, C. L. Martin, and R. H. Bowering (to Mobil
Oil Corp.).
135. R-F. Ju, Y-L. Lee, S. H. Lin, and W. J. Burke, Colloids and Surfaces 49, 395
(1990).
136. R. T. Podoll and K. C. Irwin, in Ref. 132, p. 346.
137. A. R. Tarrer and V. P. Wagh, in Ref. 132, p. 361.
138. F. M. Bayomy and J. A. Naji, in Ref. 47, p. 362.
139. R. L. Davis, in Ref. 71.
140. R. Roque, M. Tia, and B. E. Ruth, in Ref. 71.
141. N. W. McLeod, in Ref. 71.
142. P. S. Kandhal and W. C. Koehler, in Ref. 71.
143. Asphalt Hot-Mix Emission Study, Research Report RR-75-1, The Asphalt Institute,
College Park, Md., 1975.
144. R. P. Hangebranck, D. J. von Lehmden, and J. E. Meaker, Sources of Polynuclear
Hydrocarbons in the Atmosphere, Publication No. 999-AP-33, U.S. Public Health
Service, Washington, D.C., 1967.
145. Particulate Polycyclic Organic Matter, National Academy of Sciences, Washington,
D.C., 1972.
146. J. M. Colucci and C. R. Begeman, Environ. Sci. Technol. 5, 145 (Feb. 1971).
147. Petroleum Asphalt and Health, API Medical Research Rep. EA 7103, American
Petroleum Institute, Washington, D.C., Dec. 1, 1971.
148. Health Lab. Sci. 11, 218 (July 1974).
Vol. 2 PETROLEUM COKE 205
149. C. H. Baylor and N. K. Weaver, Arch. Environ. Health 17, (Aug. 1968).
150. W. Luinsky, Anal. Chem. 32, 684 (1960).
151. W. C. Hueper, Occupational Tumors and Allied Diseases, Charles C Thomas,
Springfield, Ill., 1942, p. 69.
152. Guide for the Safe Storage and Handling of Heated Petroleum-Derived Asphalt
Products and Crude Oil Residue, Publication 2023, American Petroleum Institute,
Washington, D.C., Mar. 1977.
JAMES G. SPEIGHT
Western Research Institute
PETROLEUM COKE
1. Introduction
– Shot coke
– Sponge coke (fuel coke)
– Sponge/honeycomb coke (regular or anode-grade coke)
– Needle coke (graphite or premium coke)
is not defined precisely (1). A differentiation with regard to isotropy and crystal
structure is logical. There is a graded transition from:
to
Directive) and CONCAWE, product dossier no. 93/105 (3), petroleum coke is
considered a nondangerous product. The level of polycyclic aromatics in petro-
leum coke is about 250 mg/kg. Benzo[a]pyrene makes up about 25 mg/kg. The
amount of heavy metals apart from V and Ni is insignificant (< 1 mg/kg). In
general the quality of the feed components and the production conditions
2
determine the quality of the product.1
3. Production
While one coking chamber is being filled in the cycle time, the other cham-
ber is emptied. A typical volume of a modern coke drum is about 1000 m3 with a
range in diameter of 5 to 8 m and in height of 25 to 35 m. After steaming and cool-
ing of the coke chambers, the coke is removed by drilling and cutting with high
pressure water (up to 300 bar pressure). The green coke obtained is conveyed to
its application after the draining step: either as fuel of feed for calcining. Table 2
shows the different green coke qualities in relation to use.
Vol. 2 PETROLEUM COKE 209
Fig. 2. Photograph of a refining delayed coker complex: furnace, coke drums, fractiona-
tor, and dewatering bins.
210 PETROLEUM COKE Vol. 2
The following components can be used as feed for delayed cokers (7):
1. Crude residues (atmospheric or vacuum),
2. Crack components (visbroken tar, cycle oil, decant oil, or thermal crack tar),
3. Deasphalting residues,
4. Coal oils, and
5. Used plastic materials (recycling)
The products of delayed coking processes are (basis: vacuum residue or crack
components):
1. Gas/LPG (ca. 13 wt %)
2. Naphtha (ca. 11 wt %)
3. Middle distillate (ca. 45 wt %)
4. ‘‘Green coke‘‘ petroleum coke (ca. 31 wt %)
Delayed coking is carried out in order to free refinery from the product
‘‘heavy fuel oil‘‘ which is not easily marketable. In terms of investment delayed
coking is a much cheaper technology than flexicoking or hydrocracking (factor 3
to 5). A further argument in favor of delayed coking can be the revalorization
potential for the carbon products ‘‘regular calcinate‘‘ or ‘‘needle coke‘‘ produced.
Moreover, delayed coking shows good product results, which, however, like all
thermal conversions, have ‘‘middle distillates‘‘ as main product.
Figure 3 shows the product yields for the North Sea crude ‘‘Brent‘‘. All ther-
mal conversion processes reduce the crude residues higher than middle distil-
lates, especially the delayed coker unit.Goal of nearly all delayed coker units is
to minimize coke yield and obtain as many distillate fractions as possible. How-
ever, when the objective is to produce green coke that shall serve as raw material
for higher quality coke products (regular calcinate and needle coke) a specific
feed optimum must be found that by no means results in a minimum coke yield.
The distillate yields and thus the capacities of delayed coking can be
increased by (7,8):
1. Preconversion of feed components (6)
2. Addition of light middle distillate to the feed stream
3. High coke drum outlet temperature/high coker furnace outlet temperature
(COT)
4. Low operating pressure
5. Low fresh feed: recycle ratio
6. Addition of coil steam
Vol. 2 PETROLEUM COKE 211
1. Onstream decoking (allows decoking of one of the four furnace coils during
current production)
Fig. 4. Relationship between plant operating conditions and plant yield; parameters:
coke drum pressure = 2.5 bar; feed/recycle ratio = 1 : 1.3; coke drum outlet tempera-
ture = 440 8C. (9).
212 PETROLEUM COKE Vol. 2
Table 3. Time Allotted for Different Steps in Typical 12-h and 24-h Coking cycles
Time, h
Operation step 24-h cycle 12-h cycle
a
Switch drum 0.5
b
Steam out to fractionator 0.5
Steam out to blowdown 1.0
b
Slow water cooling 1.0
b
Full water rate 5.0
Drain water and remove top head 3.5 1.5
Drop bottom head and feed line 0.5 0.5
Hydraulic decoking 4.0 2.0
Replace heads and feed line 1.0 0.5
Steam purge and pressure test 1.0 0.75
Drum heat up 4.0 2.25
Slippage allowance 2.0 1.0
Total 24 h 12 h
a
Total time for both operation steps: 0.5.
b
Total time for both operation steps: 3.0.
With these technical measures the cycle time for delayed coking can be reduced
from about 24 to 12 h (14–17). The time required for different steps in typical 12-
h and 24-h coking cycles is listed in Table 3.
Fluid Coking. The continuous fluid coking process utilizes fluidized solids
technique developed for fluid catalytic cracking (! Oil Refining, Chap. 3.2),
except that no catalyst is used. A flow scheme is shown in Figure 5.Residue is
feed to the reactor (a). The cracking reactions occur at 500 – 550 8C in a fluidized
bed of coke particles. The coke formed in the reactor flows continuously to the
heater, where it is heated up to 600 – 650 8C by partial combustion in a second
fluidized bed. Part of the heated coke particles are returned to the reactor to sup-
ply the energy to the endothermic crack reactions and to maintain the reactor
temperature. The remainder of the coke after cooling is removed as a stream
of fine particle ‘‘petroleum coke’’. Fluid coke is only occasionally used for
production of anodes (18).
Flexicoking. The flexicoking process combines fluid coking with gasifica-
tion. A flow scheme of flexicoking is shown in Figure 6. Flexicoking is a residue
Vol. 2 PETROLEUM COKE 213
Fig. 5. Flow sheet of fluid coking. a) Reactor; b) Scrubber; c) Burner; d) Air blower.
Fig. 7. Schematic of the rotary kiln calciner. a) Screen; b) Crusher; c) Weigh feeder;
d) Green coke sampling; e) Rotary kiln; f) Cooler; g) Sampling calcined coke; h) Incinera-
tor; k) Stack.
cooler (f). The flue gases are cooled in a waste-heat boiler (i) and cleaned in an
electrostatic precipitator (j).
Rotary Hearth Calciner. The rotary hearth calciner (Fig. 8) was originally
developed by the Wise Coal and Coke company. Green coke charged at the
Fig. 8. Schematic of the rotary hearth calciner. a) Feed bin; b) Rotary hearth;
c) Calcinate cooler; d) Waste-heat boiler; e) Electrostatic precipitator; f) Stack.
Vol. 2 PETROLEUM COKE 217
periphery of the hearth travels around the calciner (b) in a series of concentric
paths. At the completion of each revolution the coke is moved by a rabble arm
set in the roof into the next concentric path. This process is repeated, until
at the final rabble, the coke is discharged into the soaking pit. From the
soaking pit the coke can be discharged into a rotary cooler (c). During
normal operation the retention time for the coke in the hearth and soaking pit
is approximately 2 h (21).
4. Uses
5. Quality Aspects
The quality of coke is chiefly influenced by the feed material used in the petro-
leum refining. Of second and third importance are the coking and the calcining
conditions. Structure, density and VCM contents can vary significantly as can
the sulfur, metals, and ash contents.
Since the mid 1980s an international standardization of testing methods for
carbon materials has been developed. ISO, DIN or ASTM specifications or drafts
exist for nearly all parameters. Standardization has almost been achieved even
in the case of coefficient of thermal expansion measurement (CTE), for a long
time tested by proprietary methods of electrode manufacturers. The situation
that national standards (ASTM, DIN) are incorporated in the international
ISO specifications makes itself positively noticeable.
5.1. Green Coke. The quality of normal green coke mainly used as fuel
is defined by the sulfur content. Additional parameters are the water content (for
the quantity calculation), the ash, and VCM content. In the case of green coke
used for higher grade products the level of metals and the grain size distribution
are further quality criteria.
5.2. Regular Calcinate. The quality of calcined petroleum coke pro-
duced in Europe is superior to that produced in other continents. Due to author-
ity requirements and improvements in efficiency, regular calcinate in Europe
has undergone a quality improvement towards needle coke: the sulfur contents,
reactivity results, and CTE figure are lower, and the vibrated bulk density
is higher. Some non-European manufacturers are already producing similar
Vol. 2 PETROLEUM COKE 219
qualities. Typical European quality characteristics are listed below. The figures
in brackets refer to quality requirements outside Europe (23–26).
Sulfur content, wt % max. 1.7 (3)
Ni/ V content, ppm max. 120 – 150 (200)
Na/ Ca/ Si content, ppm max. 100
Vibrated bulk density (1 – 2mm), g/cm3 0.80 – 0.85 (0.75 – 0.80)
Air reactivity (1 – 1.4 mm), wt %/min max. 0.15
CO2 reactivity (1 – 1.4 mm), wt % max. 12
CTE factor (carbon anodes), 10-6 /8C 3.5 – 4.5 (5.5)
Hydrogen content, wt % 0.05 – 0.09
Real density, g/cm3 2.06 – 2.10
Ash content, wt % 0.05 – 2.20
Crystallite size Lc, nm 2.5 – 3.2
Specific electric resistivity, mO m 460 – 520
Grain stability, wt % min. 70
Dust forming factor, wt % max. 0.02
Water/oil content, wt % max. 0.3
(from antidust additive)
Except for the SO2 emission the sulfur content only has slight side effects on
aluminum electrolysis.
The goal in aluminum electrolysis is to have a small consumption of anodes
with a good electricity yield. Good production results are 95 % electricity yield
with an anode consumption of < 410 kg per tonne aluminum. These figures can
be reached if the anodes meet the specifications for:
With the capacity increase of aluminum smelters, larger anodes have come
into use. Cracking of the anodes due to thermal shock has been observed occasion-
ally in these cases. Good thermal shock resistance of the anodes is therefore desir-
able. Improvements in thermal shock behavior will occur if CTE, the Young’s
modulus, and the flexural strength are decreased, and the thermal conductivity
is increased.
5.3. Needle Coke. In the production of steel by the arc furnace process
about 1 106 t/a graphite electrodes are needed. Charging stock for the electro-
des are needle coke (about 85 %), and pitch binder for backing and impregnation.
Needle coke is the highest grade product of calcinated petroleum coke. Because of
the massive thermal and mechanical stresses in the arc furnaces, the highest
demands are placed on the electrode properties. Moreover, subsequent modified
graphitization due to the baking of the electrodes result in additional stricter
quality requirements.
Modern lengthwise graphitization is considerably more economic and
only needs a third of the time required by conventional Acheson graphi-
tization, but the more drastic reaction conditions demand superior quali-
ties from the needle coke. For this reason needle coke specifications have
220 PETROLEUM COKE Vol. 2
greatly improved over the last 15 years. Room for further improvement seems
only marginal.
For needle coke the chemical properties (low contents of S, Ni, V, Na, Ca, Si)
and the following physical characteristics must be optimum:
1. Feed selection for delayed coking (low sulfur and low metal residues with
high contents of 3, 4, and 5-ring aromatics and low asphaltene content)
2. Delayed coking conditions (longer cycle, higher recycle ratio, right furnace
coil out temperature ramp)
3. Calcination conditions (heating rate and final calcination temperature)
The quality analysis of needle coke is special because two parameters are specified
that can only be analyzed after graphitization: The CTE and the puffing factor; ie,
the irreversible lengthwise expansion during graphitization (requirement: < 2 %).
The quality comparison has been more complicated up to now by the fact that
the preparation, graphitization, and measurement of CTE were proprietary com-
pany methods. Standardization however is nearly concluded (last step: proposal of
CTE–DIN method, an English translation exists as ISO draft) (27).
The quality requirements of electrodes depend on the electrode diameter
and also on the type of arc furnace (d.c. or a.c. operation). That is why different
types of needle coke are available. Some properties of ‘‘super premium‘‘ and
‘‘intermediate premium‘‘ qualities are listed in Table 6. The CTE values refer
to the measuring range 20 – 200 8C after preparation and graphitization accord-
ing to the DIN method.
There are ongoing tests to produce ‘‘ultra premium‘‘ needle coke with CTE
levels of 0.25 106/ 8C. Production constancy, possible production quantity and
processing characteristics of this quality still need to be evaluated.
Green coke is normally stored in open piles with side walls as protection
against the wind. In order to reduce the dust discharge, green coke piles are
sprinkled with water. The potential dust emission is visibly improved by
means of dosing antidust additives (polyacrylate types) to the sprinkling water.
The general explosive boundary value at work of 6 mg/m3 for fine dust con-
centrations is to be complied with.
6.2. Calcined Petroleum Coke. All hydrocarbons in green coke are
eliminated by calcination. The remaining hydrogen content is < 0.1 %. Therefore,
no polycyclic aromatics are traceable. The content of S, N, and metals is virtually
unchanged. With water, only part of the salt content of the cutting water can be
elutriated. No classification and labeling is required.
Due to the dryness of the calcinate, irritation of eyes and throat by the dust
is possible. The general boundary value at work (max. 6 mg/m3 for fine dust con-
centration) is to be complied with.
Calcined coke is usually stored in closed silos or bins to keep the material
dry. To reduce dust formation, the calcinate is sucked out during transport and
loading and the dust held back by filter stations. Sprinkling the calcinate
with residue parts of 0.2 – 0.4 wt % (old process) or an aqueous emulsion of
222 PETROLEUM COKE Vol. 2
dust binder prevents environmental problems from dust. The dust forming fac-
tor (‘‘R+D-Carbon‘‘-method) must be < 0.02 wt %. The combustion and explosive
behavior of calcined coke presents no problem:
REFERENCES
21. H. Predel, ‘‘Steps for Green Coke Calcination-Mathematic Model and Practical Tests
and Experiences,’’ Light Metals 1996, TMS, p. 499–507.
22. W. K. Fischer, G. Collin, J. W. Stadelhofer (1990), ‘‘Technischer Kohlenstoff—
Herstellung und Anwendung,’’ DGMK-Haupttagung, Sept. 1921, p. 423.
23. H. Blumberg, ‘‘Fibers for Composites Status Quo and Trends,’’ Chemical Fibers
Internat. 47 (1997) Feb., 36–41.
24. C. Jonville, J. C. Thomas,C. Dreyer, ‘‘The Influence of Coke Source on Anode
Performance,’’ JOM 47 (1995) Aug. 23–24.
25. S. M. Hume, W. K. Fischer, R. C. Perruchoud, B. Welch, ‘‘A Model for Petroleum
Coke Reactivity,’’ Light Metals 1993, TMS, p. 525–534.
26. F. Keller, U. Mannweiler, ‘‘Anode Performance: The Link between Coke Properties
and Aluminum Production Cost,’’ PACE, Petroleum Coke Quarterly 1995.
27. M. W. Meier: Cracking Behavior of Anodes, R+D Carbon Ltd., Sierre 1996.
28. J. Mochida, K. Fujimoto, T. Oyama in P. A. Thore (ed.): ‘‘Chemistry in the Production
and Utilization of Needle Coke,’’ Chemistry and Physics of Carbon, vol. 25, Marcel
Dekker, New York – Basel – Hongkong 1997, p. 111–212.
HEINRICH PREDEL
MiRO GmbH & Co. KG, Karlsruhe
Federal Republic of Germany
Gasoline and other motor fuels comprise the largest single use of energy in the Uni-
ted States, and in 2002 accounted for 22% of all energy usage (1). The cost of this
energy has been and is expected to continue to be a primary factor in the national
economy. Moreover, the fraction of resources from which these fuels come that is
provided by foreign sources is a matter of political concern. The fraction of total
crude oil produced domestically shrunk from 75% in 1970 to 39% in 2002 (2) and
is predicted to continue to drop (see NOMENCLATURE IN THE PETROLEUM INDUSTRY).
Imported products also continue to grow. In 2002, the United States imported
130 106 L/day of gasoline and gasoline components. In the 1970s, two Organization
of Petroleum Exporting Countries (OPEC) embargoes resulted in rapid increases
in the price of crude oil and therefore motor fuels. These increases triggered pro-
grams designed to develop alternative sources of fuels such as coal (qv), oil shale
(qv), and natural gas (see NATURAL GAS). In 2001, as a result of lower price volatility
and improved energy efficiencies, the inflation adjusted cost of driving an average
passenger car is about 40% lower than it was in 1960 (3). Today, alternative trans-
portation fuels are being considered as replacements for traditional petroleum
based fuels based on their perceived potential to improve the security of energy
supply and to reduce emissions of greenhouse gases. Among alternative fuels
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0701191508150308.a01.pub2
224 GASOLINE AND OTHER MOTOR FUELS Vol. 2
under consideration are hydrogen, natural gas and biofuels such as ethanol and
esterified edible oils. (see also AIR POLLUTION; EXHAUST CONTROL, AUTOMOTIVE).
1.1. General Aspects of Manufacture of Motor Fuels. All motor fuel
in the United States is manufactured by private companies. Many of these are
vertically integrated. That is, the same company finds the crude oil or buys it
from a producing government, refines it into finished products, and then sells
these products to independent retailers who specialize in that company’s blended
products or sells them at company operated service stations. There are also a
significant number of companies that participate in only some aspects of the
business cycle such as refining or marketing.
Four groups are involved in the production or use of motor fuels in the
United States: (1) manufacturers of the vehicles; (2) manufacturers and/or mar-
keters of the fuels; (3) purchasers and users of fuels and vehicles; and (4) federal,
state, and local regulatory agencies. Each has a different role.
Vehicle manufacturers must build cars and trucks that operate well on
available fuels. They also specify the fuel requirements of their vehicles. Fuel
marketers must produce fuels that operate in both new and old vehicles. The con-
sumer, the purchaser of the fuel and the vehicle, wants the fuel to be affordable,
readily available, and able to provide a high level of performance. Regulators
check that fuels are labeled properly and meaningfully and that no unwarranted
claims are made. They also regulate emissions from vehicles and set fuel speci-
fications based on environmental considerations.
Appropriate specifications are set when the vehicle and the fuel are viewed
as a system, and when all four groups, or stakeholders, work together to define
cost effective and efficient solutions that satisfy the needs of all stakeholders.
The American Society of Testing and Materials (ASTM), now known
as ASTM International was founded in 1902 to promote just such a process.
ASTM Committee D-2 provides a forum for regulators, vehicle manufacturers,
fuel producers, and consumers to develop and recommend nonbinding standards
for petroleum products. Among the types of ASTM standards are Specifications
and Test methods. Specifications define a precise set of requirements to be satis-
fied by a material. Test methods define procedures for measuring qualities
and characteristics of a material. Specifications for gasoline are contained in
ASTM D 4814 (4) and for automotive diesel fuel in ASTM D 975 (5). These stan-
dards are updated periodically.
ASTM committees must be balanced in that the number of voting produ-
cers must not be greater than the number of voting nonproducers. For petroleum
products, nonproducers are regulators, consumers, and equipment manufac-
turers. Committee chairs must be nonproducers. Although standards must be
approved by a majority, all negative votes must be carefully considered and a
response made. In practice, because all issues are fully discussed and the discus-
sions are based on hard data, very few negative votes are cast when standards
are submitted for final approval. Although ASTM specifications do not have the
force of law, many states have adopted them as state regulation, thus forcing a
minimum quality level in the field.
ASTM specifications cover those aspects of fuel properties that ensure good
performance in gasoline and diesel powered vehicles. Other fuel properties are con-
trolled for environmental reasons. These environmental related specifications are
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 225
2. Gasoline
Gasoline, the preferred fuel for the Otto engine, evolved from an unwanted by-
product to an indispensable mainstay of modern life in only a few short years. In
1900, gasoline was little more than a disposal problem for the infant petroleum
industry in the production of stove and lamp oils, lubricating oils, and greases.
The first prototype four-stroke, internal combustion, spark-ignited engine was
built by Nicolaus Otto in 1876 in Germany. This engine, which ran on gasoline, pro-
vided enormous advantages over earlier internal combustion designs in terms of
weight, power, and efficiency. By 1890, almost 50,000 of these engines had been
sold. As the popularity of these engines grew at an exponential rate, the demand
for gasoline grew accordingly. Gasoline demand grew by a factor of five between
1907 and 1915.
Figure 1 shows the growth in gasoline demand in the United States since
1950 (6). Demand is largely determined by the growth in the number of cars,
the kilometers of paved roads available for driving, population, and economic
growth. Figure 1 also shows the number of passenger vehicles and the average
yearly distance driven per vehicle (7). Most Americans drive to work, to shop,
and for recreation. In 2002, American consumers burned 503 109L of gasoline.
The average passenger vehicle is driven about 19,000 km/yr, up from about
15,000 km/yr in 1950. Despite stated goals of reducing the number of kilometers
that people drive to work through increased use of mass transit and car-pooling,
the public’s preference for personal driving has remained strong.
A primary factor in moderating the demand for gasoline has been the dra-
matic improvement in automotive fuel economy from the mid-1970s through the
mid-1980s. Figure 2 shows the trend in fuel economy for new cars since 1955
and for light duty trucks from 1980 (8). Improvement is the result of governmental
226 GASOLINE AND OTHER MOTOR FUELS Vol. 2
500 25
200 10
100 5
0 0
1950 1960 1970 1980 1990 2000
Year
Fig. 1. Growth in annual gasoline demand, in GL/yr; millions of passenger cars; and
travel per vehicle, in km/yr. Passenger vehicles, 106, gasoline demand, GL/yr,
and travel per vehicle, 1000 km/yr.
regulations that were a direct outcome of the oil shortages of the 1970s.
Average fuel economy has remained virtually constant since the mid-1980s, as
no new regulations have been imposed. This clearly shows that the current cost
structure of gasoline is not a significant impediment to people’s desire to drive,
and that, in general, the consuming pubic would rather have larger, more powerful
family vehicles than smaller more efficient ones.
Energy outlooks are developed annually by the Department of Energy. In
their most recent forecast for the years 2002–2025, energy demand for light
duty vehicles is projected to grow about 1.9% per year between 2002 and 2025,
14.0
12.0
Fuel economy, km/L
10.0
8.0
6.0
4.0
2.0
0.0
1955 1965 1975 1985 1995 2005
Year
Fig. 2. Fleet average fuel economy for new cars and light duty trucks. Passenger cars
light trucks.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 227
1.9% per year for light commercial trucks and 2.2% per year for freight trucks (9).
These projections include growth in travel and modest improvements in fuel
economy. The recent trend of using technology to improve performance rather
than efficiency will continue; average horsepower is projected to grow by 24%
while fuel economy will improve by only 6%. It is expected that by 2025, 2.1%
of light duty vehicles will be fueled by alternative fuels. It should be recognized
that these projections are sensitive to many factors, including overall economic
growth and fuel costs.
2.1. Requirements of Good Gasoline. To satisfy high performance
automotive engines, gasoline must meet exacting specifications, some of which
are varied according to location and based on temperatures or altitudes. The fuel
must evaporate easily and burn completely when the spark plug fires in each
cylinder. Early detonation of the fuel in the cylinder can cause destructive engine
knock. The fuel must be chemically stable. It should not form gums or other poly-
meric deposit precursors. There should be no particulate contaminants or entra-
ined water. Contaminants must be prevented from the point of manufacture in
the refinery all the way through the distribution system until the fuel is metered
from the vehicle tank into the engine.
Octane. Octane is probably the single most recognized measure of gaso-
line quality. The octane value of gasolines are posted on service station dispen-
sers, and most drivers recognize a fuel in which the octane is too low. Broadly
speaking, octane is a measure of the combustion characteristics of gasoline. In
an Otto cycle engine, gasoline vapor in the combustion chamber must start to
burn only when the spark plug ignites after the charge has been compressed.
Then it must burn with a well-defined flame front that travels all the way across
the combustion chamber. Low octane gasoline has a tendency to preignite. As the
flame front sweeps across the combustion chamber, the unburned portion of the
fuel (end gas) heats up under rapidly rising temperature and pressure condi-
tions. The fuel–air mixture undergoes chemical reactions that may cause it to
autoignite and detonate the entire remaining mixture. Instead of being pushed
down smoothly on the power stroke, the piston is given a hard instantaneous rap
to which it cannot respond because of the large mechanical inertia present in the
crankshaft and other pistons. This rapid energy release causes pressure fluctua-
tions in the cylinder which result in a loud metallic noise commonly called knock.
In addition to producing an objectionable sound, knock can reduce the amount of
useful work that is extracted from the engine. Power is dissipated in the pressure
waves, and increased heat is radiated to the cylinder walls and into the cooling
water. Under extreme conditions of prolonged knock, overheating and even
engine damage can occur. The damage is typically caused by catastrophic melt-
ing of piston crowns or head gaskets.
Knock is caused by unwanted chemical reactions in the combustion cham-
ber. These reactions are a function of the specific chemical species which make
up the fuel and the environmental conditions to which the fuel is subjected
during the compression and power stroke in the engine. Therefore, both the
chemical makeup of the fuel and the engine design parameters must be consi-
dered when trying to understand knock.
Chemical Factors. Because knock is caused by chemical reactions in the
engine, it is reasonable to assume that chemical structure plays an important
228 GASOLINE AND OTHER MOTOR FUELS Vol. 2
external factors such as the wall temperature or by the amount of heat released
in the combustion process itself, which is directly related to the density of the
fuel–air mixture. A vehicle factor which increases charge density, combustion
chamber temperatures, or available reaction time promotes the tendency to
knock. Engine operating and design factors which affect the tendency to produce
knocking are
number of zero. Octane numbers of primary reference fuels between zero and 100
are formed by volumetric mixtures of isooctane and n-heptane. For example, a
50:50 mixture of the two has an octane number of 50. Primary reference fuels
for measuring octane numbers above 100 contain specified amounts of tetra-
ethyllead. For instance, isooctane containing 0.238 g TEL/L has a defined octane
number of 108.
The octane value of an unknown fuel sample is determined by comparing its
knocking tendency to various primary reference fuels. Its measured octane is
equal to the octane of the PRF which has the same knocking intensity. Knock
intensity is controlled to an average value by varying the compression ratio of
the CFR engine. In practice, the exact value of a fuel’s octane number is deter-
mined to the nearest 0.1 octane number by interpolation from two PRFs that are
no more than two octane numbers apart.
The CFR engine is operated at two conditions to simulate typical on-road
driving conditions. The less severe condition measures research octane number
(RON); the more severe one measures motor octane number (MON). Table 2
summarizes the operating conditions for each test.
The difference between RON and MON for a particular fuel is called the
sensitivity. By definition, the RON and MON of the primary reference fuels are
the same and the sensitivity is zero. For all other fuels, the sensitivity is almost
always greater than zero. Generally, paraffins have low sensitivities whereas
olefins and aromatics have sensitivities ranging up to 10 and higher.
The octane numbers of many pure compounds have been measured and
reported in the literature. Probably the most comprehensive project was carried
out under the auspices of the American Petroleum Institute (13). Table 3 lists
RON and MON values for a number of representative compounds. Some aromatic
compounds cannot be tested neat in the knock engine, so these are evaluated at
levels of 20%, and the equivalent octane number is calculated. The values for
oxygenates in Table 2 have been reported elsewhere (14).
These two test methods and the octane numbers of the fuels measured
are ultimately used to evaluate the performance of vehicles. The two methods
provide useful information about how fuels perform in cars. The RON corre-
sponds to light load, low speed conditions, whereas the MON corresponds to
heavier loads, and high speed severe driving conditions.
The octane number requirement (ONR) of a car is the octane number which
causes barely audible, ie, trace knock when driven by a trained rater. The Coor-
dinating Research Council (CRC), a research organization funded jointly by the
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 231
American Petroleum Institute (API) and six members of the automotive industry
(General Motors, Ford, Daimler-Chrysler, Toyota, Volkswagon and Honda), has
defined test procedures for measuring ONR. Each car is driven under a set of
light and heavy accelerations until the most sensitive driving mode is deter-
mined. Then a series of fuels is run in the car until trace knock is determined.
Auto manufacturers have developed knock detecting and limiting devices
called knock sensors which are installed in most engines today. The sensor is a
piezoelectric crystal which is tuned to the natural frequency of engine knock
sound (see SENSORS). When knock is detected, a signal is sent to the on-board
computer which retards the spark timing. In this way, knocking noise is elimi-
nated until the maximum spark retard is reached. The tradeoff, however, is
reduced power. Full throttle acceleration times can be increased by up to 10%
when these cars use low octane fuel. Measuring octane number requirements
in cars with knock sensors involves either determining when octane is sensed
and spark retard is initiated, or sensing when the maximum level of retard
has been reached. CRC has defined a test procedure for measuring the octane
response of cars equipped with knock sensors (15).
It has long been known that the design of a car affects the way in which it
responds to RON and MON. The response of a car to RON and MON can be char-
acterized by a parameter called severity, which is a measure of its relative res-
ponse to RON and MON. In the 1930s, most cars responded only to RON. That is,
it did not really matter what the MON of a fuel was. As long as the RON was
232 GASOLINE AND OTHER MOTOR FUELS Vol. 2
above a minimum value, the car would not knock. Cars of the 1990s generally
have a severity of about 1, meaning that a 1 number change in RON can be offset
by a corresponding and opposite change in MON to provide equivalent knocking
protection. Octane numbers which are posted on gasoline dispensers in all states
show the antiknock index which is the average of the RON and MON values of
the gasoline being sold. Recent data suggest that modern cars respond mostly to
RON and that consideration should be given to changing the way that octane is
posted on gasoline dispensers (16,17).
In designing engines, automotive manufacturers must decide what kind
of fuel that engine should use, from low to high octane. Compression ratio is
the primary determining factor in a vehicle’s octane requirements. High perfor-
mance, high compression ratio engines need high octane or premium fuel. It is
also known that the octane requirement of a vehicle increases as a vehicle ages.
This phenomena, known as octane requirement increase (ORI), is caused by the
buildup of fuel and lubricant based deposits in the combustion chamber. These
deposits are highly insulating and increase the combustion wall temperature,
thereby increasing the tendency to knock. Generally, ORI reaches an equili-
brium value of 6–9 octane numbers over the first 24–32,000 kilometers of opera-
tion, although higher increases are not uncommon. In some cases, as cars age
and start to burn more oil, octane requirements can increase again.
Over the years, engine compression ratios have undergone a number of
changes. These increased rapidly through the 1950s and 1960s. The advent of
unleaded gasoline and the demise of lead as a cheap source of octane, brought
about a drop in fuel octane numbers and compression ratios. In the early
1980s, both compression ratios and average octane numbers requirements
increased as a result of customer demand for more powerful cars.
The particular fuel octane number which a gasoline marketer chooses to
produce and sell is determined by a number of factors including the cost of
producing the various levels of octane, the number of customers requiring the
various octane levels, and competitive offerings. Determining customer require-
ments is the most difficult. The surveys conducted by CRC measure vehicle ONR
using trained raters. Trained raters are much more sensitive to knocking noise
than the average customer. Consequently, octane requirements percieved by
customers are generally lower than trained raters. This customer/rater delta
has been the subject of a number of studies (18,19) and the average difference
has been estimated to be between 3–5 octane numbers. Typically, a marketer de-
cides that the premium grade should satisfy a certain fraction of customers,
usually well over 90%. The octane values for the regular and/or intermediate
grades are determined in a similar manner.
The optimum level of gasoline octane is a complex function of a number of
factors involving the customer as well as the automotive and petroleum indus-
tries. Increasing octane values may require significant investment and higher
operating costs on the part of the oil industry that could raise the price of gaso-
line to the consumer. On the other hand, vehicle manufacturers could take
advantage of the availability of higher octane gasolines by raising the compres-
sion ratios of their engines which could in turn increase fuel economy due to
higher thermal efficiency. Higher compression ratio engines may cost more to
build, but the consumer would save in fuel costs over the life of the car.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 233
requirements, like octane, are a strong function of vehicle design. A great deal of
data on the volatility requirements of new cars is collected in cooperative pro-
grams run by CRC. Trained raters drive cars under strictly controlled conditions
and evaluate the driveability performance on a range of fuels. Customers are
generally less sensitive to problems than the raters. Each company decides which
level of protection to provide. ASTM also uses these data to set target levels
for the various fuel parameters.
Startability. In order to achieve combustion in an Otto cycle engine, the
air/fuel ratio in the combustion chamber must be near the stoichiometric ratio.
Unfortunately, when the engine is first started, the walls of the combustion
chamber and the intake manifold are not hot enough to vaporize much fuel.
Therefore, the vehicle is designed to meter extra fuel and less air to the engine
upon start up so that there is adequate vapor in the engine to support combus-
tion. From the fuel’s perspective, a fuel which is easily vaporized and which con-
tains a large percentage of light compounds such as butane is desirable to achieve
good starting. The ability of a fuel to achieve good starting can be correlated with
RVP and a measure of the front end of the distillation curve, either E70 or T10.
Usually, minimum levels of RVP such as 60 kPa (0.6 bar) and E70 minimum of
10% (or a T10 maximum of 608C) are satisfactory for good startability in most
winter locations (20). Obviously, fuel specifications change with ambient tem-
peratures. At higher temperatures, lower RVP and front-end volatilities are
adequate to provide good starting characteristics.
Vapor Lock. At the other end of the spectrum from starting is vapor lock,
a problem of too much volatility. Vapor lock occurs when too much of the fuel
evaporates and either starves the engine for fuel or provides too much fuel to
the engine. It occurs on days that are warmer than usual and when the car
has reached full operating temperatures. A typical situation would be when a
car is stuck in heavy traffic; the engine reaches high temperatures and the
fuel reaches its boiling point. Most cars today are fuel injected, have fuel
pumps located in the tank and fuel lines which are under positive pressure.
They are therefore less prone to vapor lock than older carbureted vehicles.
Vehicle manufacturers minimize vapor lock by keeping the fuel system
cool and under positive pressure. Fuel manufacturers minimize the problems
by seasonal volatility blending. As might be expected, the front end of the gaso-
line controls vapor lock. The temperature to obtain a V/L ratio of 20 is referred
to in ASTM D 4814 as a way to control the vapor locking tendencies of gasoline.
Warm-Up. Warm-up refers to that period of operation beginning imme-
diately after the car has started and continuing until the engine has reached
normal operating temperatures, usually after 10 minutes or so of operation.
During this period, the vehicle designer wants to get the vehicle equivalence
ratio to stoichiometric as soon as possible to minimize emissions. On the other
hand, if the mixture is leaned out too soon, the car experiences poor driveability
during the warm-up period. Fuel system design is critical during this period.
With older, single point fuel metering systems such as carburetors and throttle
body injectors, there is generally liquid fuel on the walls of the intake manifold
during warm-up. If this liquid only reaches the cylinders in bursts the vehicle
may experience unwanted surges. Multipoint fuel injectors minimize many of
these problems but makes the scheduling of the fuel pulses critical.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 235
ASTM has adopted DI as one of the standard requirements for good quality
gasoline. DI should be less than 5978C (1250 8F) during warm weather and less
than 5698C (12008F) in cold weather. It has been shown that gasolines blended
with ethanol do not follow this relationship. For fuels having high values of DI,
the presence of ethanol degrades performance relative to non-oxygenated fuels,
even at equal levels of T10, T50 and T90 and an offset is required (21). The exact
magnitude of the offset and the form of the equation have not been resolved, and
ASTM still recommends using the equation shown above.
Icing. At temperatures within 58C of freezing and under conditions of
high humidity, ice can form in the intake system of vehicles with carburetors
or throttle body fuel injectors. When the gasoline in these systems evaporates,
it cools the incoming air because of the latent heat of vaporization. If the air is
near its saturation point, ice crystals can form on the throttle plate, on the choke
plate, or in the venturi throat. The ice can cause rough idle if it accumulates
around the edges of the throttle plate when it is nearly closed. In the extreme,
ice can clog the carburetor jets and stall the car completely. After the car is fully
warmed-up there is generally enough heat in the intake system to prevent any
ice buildup. The universal adoption of port fuel injection has eliminated this
concern from today’s cars.
Back-End Volatility. The portion of the gasoline that boils above 1508C
is referred to as the back end. Molecules in this region have high energy density
and provide a significant contribution to fuel economy. Too much material
in this boiling range, however, can cause problems. This material is hard to
volatilize and when the engine is cold, tends to accumulate on the walls of
the cylinder. From there it can be washed into the oil sump and dilute the
oil. Generally, as the engine heats up this material evaporates. However, if
there are too many back ends in the gasoline, then not all may boil off and
236 GASOLINE AND OTHER MOTOR FUELS Vol. 2
the performance of the lubricant may be degraded. Very heavy molecules, such
as those having more than 12 carbon atoms, may contribute to combustion
chamber deposits. Condensed ring aromatics are particularly strong contribu-
tors to these deposits (22).
Vehicle volatility requirements are a strong function of ambient tempe-
ratures. As described above, at any given ambient temperature, volatility
should be within a certain range. Too high and vehicles may experience
vapor lock. Too low and vehicles may experience hard starting and poor drive-
ability. ASTM has defined six volatility classes for good driveability and six
classes for vapor lock protection. These specifications are shown in Tables 4
and 5 respectively.
Each month, each state is assigned a class from Table 4 depending on the
low temperatures experienced, and a class from Table 5 depending on historical
high temperatures. Some states, because of widely varying ambient conditions
may contain two or more areas. Together, these two ratings define the volatility
range for good gasolines over the course of a year.
Cleanliness. Good gasoline must be both chemically and physically clean.
Chemical cleanliness means that it does not contain nor react under conditions
of storage and use to form unwanted by-products such as gums, sludge, and
deposits. Chemical cleanliness is assured by controlling the hydrocarbon compo-
sition and by appropriate additives. Physical cleanliness means that there are
no undissolved solids or large amounts of free water in the gasoline.
After production in the refinery, gasoline may undergo two types of oxi-
dative degradation. The first occurs under ambient temperature conditions and
long periods of time, as long as six months, and is reflected in storage stability.
The second has an impact on high temperature stability and is a phenomenon
of the high temperatures and short residence times encountered as the fuel
makes its way through the vehicle intake system. Both mechanisms share some
similar chemical features in that they involve oxidation of fuel compounds and
free-radical polymerization to form high molecular weight gums. The gums and
gum precursors formed under storage conditions can lead to damaging deposits
in fuel injectors or intake valves.
Compounds associated with poor stability are olefins in general and conju-
gated diolefins in particular. A compound such as 1,3-cyclopentadiene is known
to be particularly detrimental. Compounds having poor stability can be produced
in cracking processes such as thermal cracking, including visbreaking and
steam cracking, and catalytic cracking. The presence of compounds such
as sulfur and nitrogen can also degrade stability and promote gum formation.
Vehicle manufacturers know that copper promotes oxidation of the fuel and are
careful to minimize the use of copper in the fuel system materials.
A number of laboratory tests are used to predict chemical stability. The
amount of existent gum in a gasoline is determined by ASTM D 381. This method
involves evaporating a sample by a jet of heated air. The residue is weighed,
solubles are extracted with n-heptane, and the sample is reweighed. The total
is called unwashed gum and the insoluble portion is called solvent-washed
gum. ASTM D 4814 specifies that gasoline should contain less than 5 mg/100 mL
of solvent-washed gum. Most gasolines contain less than 3 mg/100 mL.
Although ASTM D 381 measures the gum existing in gasoline at a parti-
cular point in time, it does not indicate how much more might be formed during
storage at the refinery, during various modes of transportation, in service sta-
tion tanks, or in vehicle tanks. The oxidative stability test (ASTM D 525) was
developed to provide a rapid means of predicting future gum formation. This
method consists of placing a sample of the gasoline in a bomb, pressurizing it
to 690 kPa (100 psi) with oxygen, and maintaining it at a temperature of
1008C. The pressure inside the bomb is monitored continuously. The oxygen
reacts with the gasoline slowly at first, but eventually the pressure drops sharply,
14 kPa (2 psi) or more within a 15 minute interval, indicating a breakpoint. The
time from the start of the test to the breakpoint is the induction period and is a
measure of the oxidative stability of the gasoline. ASTM D 4814 specifies that
gasoline induction period must exceed 240 minutes, although most good gaso-
lines have induction periods in excess of 960 minutes, the duration of the test.
Another ASTM test method, Potential Gum (D 873), combines the existent
gum and the oxidation stability tests to measure potential gum. A sample of
gasoline is subjected to the oxidation stability test for 960 min, filtered to remove
particulates, and then subjected to an existent gum test. The potential gum is
expressed as the total (unwashed) gum in this test.
Other tests to predict stability of gasoline have been developed and reported
in the literature. One, developed by the U.S. military, stores gasoline at elevated
(438C) temperatures for up to 12 weeks and measures existent gum at the end of
that period (23). Another measures existent gum in the presence of copper. The
copper catalyzes oxidation and may be a better estimator of the stability of
gasoline at high temperature/low residence time conditions.
238 GASOLINE AND OTHER MOTOR FUELS Vol. 2
(78 RON/75 MON). Whereas the potential yield of gasoline directly from crude oil
is less than 20%, the demand is about 50%. The heavy material must be convert-
ed to lighter material and the octane of many of the existing streams improved by
changing chemical composition.
Catalytic Cracking. Although it has long been known that heating crude
oil fractions could break or crack the compounds into smaller molecules, the
development of suitable catalysts and processing designs has made catalytic
cracking the premier refinery process for changing the molecular structure of
the crude (see CATALYSIS). Catalytic cracking generates higher yields than ther-
mal cracking as well as superior quality products. Over 50% of the gasoline in
the United States is obtained by catalytic cracking which uses a fluidized bed
of powdered or small diameter catalysts that are continuously regenerated
in an adjacent vessel called a regenerator (see FLUID CATALYTIC CRACKING (FCC)
UNITS, REGENERATION). The fluidized-bed catalytic cracking (FCC) process was
first commercialized in 1942 by Standard Oil Co. (New Jersey) at its Baton
Rouge refinery (26) using powdered silica/alumina as a catalyst. Fluidized beds
offered the ability to use powdered, high surface area catalysts, to operate con-
tinuously, to regenerate easily, and to use short contact times which increased
yield and selectivity. Since 1942, many improvements have been made in process
design, catalyst formulation, and the ability to handle heavier feeds (27–29).
Today’s catalysts are zeolites having highly controlled pore size and surface
area.
The principal class of reactions in the FCC process converts high boiling,
low octane normal paraffins to lower boiling, higher octane olefins, naphthenes
(cycloparaffins), and aromatics. FCC naphtha is almost always fractionated into
two or three streams. Typical properties are shown in Table 7. Properties of
specific streams depend on the catalyst, design and operating conditions of the
unit, and the crude properties.
Two undesirable aspects of FCC naphtha quality are that it may contain
unacceptably high amounts of foul smelling mercaptans, and that its thermal
stability may be too low. Mercaptans are usually found in the light FCC naphtha
and may be removed or converted to sulfides and disulfides by a sweetening pro-
cess such as Merox, developed by UOP. Thermal stability is improved in sweet-
ening processes through removal of cresylic and naphthenic acids. It may be
further improved by clay treating and by addition of oxidation inhibitors such
as phenylene diamine.
Thermal Cracking. Certain cracking conversion processes are carried out
without catalysts. Heavy residuum streams in the refinery can be cracked ther-
mally to produce coke and a mixture of lighter products. If naphtha is produced,
it may require extensive treating before it can be used directly in gasoline or used
as feed to other processes. Depending on the specific process conditions
and design these processes are known as delayed coking, visbreaking , Fluid
Coking, and Flexicoking (the last two developed by Exxon). Heavy distillate
may also be thermally cracked using high (8008C) temperature steam. This pro-
cess, steam cracking, is used to generate olefins for use in chemicals plants, but
also generates material in the naphtha range which, if the quality is appropriate,
may be used in gasoline.
Reforming. Catalytic reforming is a process to increase the octane of
gasoline components. The feed to a reforming process is naphtha (usually virgin
naphtha) boiling in the 80– 2108C range. The catalysts are platinum on alu-
mina, normally with small amounts of other metals such as rhenium. Because
the catalysts contain noble metals, the feed must have very low levels of sulfur,
nitrogen, and heavy metals such as lead and arsenic. Reactors may operate
in continuous or semicontinuous mode, the difference being the method in
which the catalyst is regenerated. Various designs have been developed by
companies such as UOP (Platforming and CCR), Exxon (Powerforming), and
Amoco (Ultraforming).
Depending on the catalysts and operating conditions, the following types
of reactions occur to a greater or lesser extent: (1) heavy paraffins lose hydrogen
and form aromatic rings; (2) cycloparaffins lose hydrogen to form corresponding
aromatics; (3) straight-chain paraffins rearrange to form isomers; and (4) heavy
paraffins are hydrocracked to form lighter paraffins.
Reformers generate highly aromatic, high octane product streams, and a
great deal of hydrogen (qv). Reformate can have RON values over 100 and
MON values of 90. The hydrogen, which can be used to improve the quality of
many other refinery streams, is an extremely valuable product of the reformer.
Some of the negative aspects of reformate are production of benzene, poly-
nuclear or multiring aromatics (PNAs), and light gas (C1 –C4) (see also BTX
PROCESSING). Benzene is a recognized carcinogen and its concentration in gaso-
line is regulated, whereas PNAs can contribute to combustion chamber deposits.
PNAs may be removed by distilling the whole reformate and discarding the
heaviest fractions.
Alkylation. Alkylation (qv) is the chemical combination of two light
hydrocarbon molecules to form a heavier one and involves the reaction of butenes
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 241
and butanes in the presence of a strong acid catalyst such as sulfuric or hydro-
fluoric acid. The product is a heavier multibranched isoparaffin. Propene and
the various pentenes may also be used, to produce C7 or C9 isoparaffins, respec-
tively. Process designers need to be careful not to let three molecules combine to
form C12 molecules because these have too high a molecular weight and probably
would contribute to poor combustion and higher emissions. Octane numbers for
alkylate are usually above 95 for both RON and MON, and the reaction has the
effect of lowering the RVP of the available pool, a plus when refiners are trying to
reduce it. The principal competitors to alkylation are the various etherification
processes which also use C4 and C5 olefins as feeds. As a result of environmental
concerns associated with use of ethers in gasoline, alkylation is becoming a more
desirable process.
There are environmental concerns over the use of HF catalyst. Although the
refining industry has taken steps to reduce the likelihood of an accidental release
and to minimize the environmental impact in the event of a release, most new
units use sulfuric acid catalysts.
Isomerization. Isomerization is a catalytic process which converts normal
paraffins to isoparaffins. The feed is usually light virgin naphtha and the cata-
lyst platinum on an alumina or zeolite base. Octanes may be increased by over
30 numbers when normal pentane and normal hexane are isomerized. Another
beneficial reaction that occurs is that any benzene in the feed is converted to
cyclohexane. Although isomerization produces high quality blendstocks, it is
also used to produce feeds for alkylation and etherification processes. Normal
butane, which is generally in excess in the refinery slate because of RVP con-
cerns, can be isomerized and then converted to alkylate or to methyl tert-butyl
ether (MTBE) with a small increase in octane and a large decrease in RVP.
Hydrogen Processing. Hydrogen is probably the most valuable refinery
chemical in terms of its ability to improve the quality of refinery streams. It can
be used to remove unwanted species such as sulfur and nitrogen from gasoline
and diesel. Hydrogen treating improves the ability to resist oxidative degrada-
tion and reduces the carbon-forming potential of diesel fuels. Under more severe
conditions and with the proper catalysts, hydrogen can saturate olefinic and
aromatic bonds. Modern desulfurization process which rely on hydrogen are
designed to remove sulfur and avoid unwanted reactions which increase hydro-
gen consumption and can lower octane by saturating olefins and aromatics. The
largest producer of hydrogen in a typical refinery is the naphtha reformer, but
hydrogen can also be produced by reforming methane into CO2, CO, and H2.
Blending Agents. Blending agents are components of gasoline that are
used at levels up to 20% and which are not natural components of crude oil.
As of this writing, all blending agents are oxygenated compounds such as ethers
(qv) and alcohols and are used for one or more of a variety of reasons, including to
increase the octane of the fuel, to reduce vehicle emissions, to reduce dependency
on imported crude oil, and/or to reduce emissions of greenhouse gases by using
renewable resources.
Ethers, such as MTBE and methyl tert-amyl ether (TAME), are made by a
catalytic process from methanol (qv) and the corresponding isomeric olefin.
These ethers have excellent octane values and compete on an economic basis
with alkylation for inclusion in gasoline. Another ether, ethyl tert-butyl ether
242 GASOLINE AND OTHER MOTOR FUELS Vol. 2
(ETBE) is made from ethanol (qv) and isobutylene (see BUTYLENES). The cost and
economic driving forces to use ETBE vs MTBE or TAME are a function of the raw
material costs and any tax incentives that may be provided because of the etha-
nol that is used to produce it. In the United States, MTBE has been found in
groundwater and drinking water supplies, where it imparts a noticeable taste
and smell at levels as low as 10 ppb. For this and other reasons, MTBE has
been banned in California, New York and Connecticut. Consequently, the use
of MTBE in gasoline is expected to decline.
Alcohols such as methanol, ethanol, and tert-butyl alcohol [75-64-9] (TBA)
C4H10O, have also been used as gasoline blending agents. The only alcohol used
in large volumes is biomass-derived ethanol. When ethanol is used as a gasoline
blending agent at a level of 10 vol%, the mixture is known as gasohol. Gasohol
actually contains only 9.5 vol% pure ethanol because the ethanol must be dena-
tured using 5% unleaded gasoline before it can legally be shipped anywhere. As of
this writing, gasohol receives a 5.2 ¢/gal (1.4 ¢/L) federal tax subsidy in the
United States. Some states, mostly farm states, also provide additional tax
rebates for blenders of ethanol. These incentives significantly improve the eco-
nomics of ethanol use.
Ethanol gasoline mixtures have a tendency to separate into two phases in
the presence of water and ethanol is corrosive in its own right. Therefore, pipe-
line companies do not allow ethanol blends to be transported through the normal
pipeline distribution system (see PIPELINES). Denatured ethanol is transported
in trucks, barges or tank cars and blended at terminals. Ethanol also has signi-
ficant nonlinearity when blended with hydrocarbons, especially for RVP and
volatility. Adding only a few percent ethanol can raise RVP of a blend by approxi-
mately 1 psi. In order to achieve a final blend with the desired properties,
refiners must prepare a special mixture called an RBOB (Reformulated gasoline
Blendstock for Oxygenated Blending). Due to ethanols nonlinear blending char-
acteristics, RBOBs must be prepared for specific ethanol contents (ie, 5.7%, 7.7%
or 10%). If used for Federal RFG, refiners must make a hand blend of RBOB and
ethanol and measure its properties before shipping the RBOB. For California
RFG, refiners may use a mathematical model to predict the properties of fully
formulated gasoline containing ethanol.
Despite these limitations, gasohol has achieved a significant market share.
In 2002, ethanol production accounted for 1.6% of the total U.S. gasoline supply
(30). In large farm states, such as Iowa and Nebraska, gasohol sales were 60%
and 38%, respectively, of the total gasoline market (31,32).
According to Section 211 of the Clean Air Act, EPA must approve use of any
new blending agent in gasoline. ‘‘New’’ means that the agent is not ‘‘substantially
similar’’ to unleaded gasoline used to certify 1974 vehicles. In 1981, EPA ruled
that gasolines containing up to 2 wt% oxygen were substantially similar (33).
All aliphatic ethers and alcohols could be used, except that methanol had to
be blended with equal parts of higher alcohols. In 1991, EPA increased the
approved level of oxygenates to 2.7 wt% oxygen for aliphatic ethers and alcohols,
except methanol (34). A complete list of allowable oxygenates may be found in
the literature (35).
Another factor which must be considered when blending oxygenates is the
effect of a large amount of a single component on the volatility and therefore
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 243
190
Temperature, ˚C
140
90
40
0 20 40 60 80 100
Percent evaporated
separate tenders, and so that service stations can easily check that the correct
grade was placed in the underground tanks (aq). Dyes are usually based on
azo chemistry and are added in concentrations below 10 ppm. Dyes are rarely
used in gasoline today and pipelines rely on other measures such as density to
separate tenders.
Antioxidants. Antioxidants (qv) are used in gasoline to improve storage
stability. Poor storage stability may sometimes be traced to high concentrations
of olefins, especially conjugated diolefins. In many cases, an antioxidant is added
to the FCC naphtha in the sweetening process. The antioxidant may be added in
large enough amounts so that the blended gasoline contains the proper concen-
tration. Antioxidants are almost always one of two chemical types: phenylenedia-
mines or hindered phenols (qv). These were first used in gasoline around 1930
(36) by Universal Oil Products.
Metal Deactivators. Small amounts of metals in gasoline can contribute
to enhanced rates of oxidation. Copper is especially potent in this regard. In order
to prevent metal catalyzed oxidation, chelating agents (qv) are often added to
gasoline. The most commonly used additive is N,N0 -disalicylidene-1,2-propane-
diamine [94-91-7], C17H18N2O2 (37) at concentrations up to 10 ppm. Special
care must be taken if any of the refinery streams are treated with a copper-
based process. Automobile manufacturers sometimes use higher levels of metal
deactivators in factory fill gasoline, because this gasoline may sit in vehicle tanks
for a long period of time until the vehicle is sold.
Corrosion Inhibitors. Corrosion-causing free water is almost always
present in at least small amounts in all parts of the gasoline distribution system.
It can come about when hot, water saturated gasoline cools in storage tanks or
pipelines. Salt water is found in ship holds, and rainwater is found in under-
ground tanks. Acidic or caustic water may be carried over inadvertently from
various processes in the refinery. Careful monitoring of service station storage
tanks is an important part of good operating practice to prevent unwanted
water from contaminating customers’ vehicle tanks (see CORROSION AND CORROSION
CONTROL).
All corrosion inhibitors in use as of this writing are oil-soluble surfactants
(qv) which consist of a hydrophobic hydrocarbon backbone and a hydrophilic
functional group. Oil-soluble surfactant-type additives were first used in 1946
by the Sinclair Oil Co. (38). Most corrosion inhibitors are carboxylic acids (qv),
amines, or amine salts (39), depending on the types of water bottoms encoun-
tered in the whole distribution system. The wrong choice of inhibitors can lead
to unwanted reactions. For instance, use of an acidic corrosion inhibitor when the
water bottoms are caustic can result in the formation of insoluble salts that can
plug filters in the distribution system or in customers’ vehicles. Because these
additives form a strongly adsorbed impervious film at the metal liquid interface,
low liquid concentrations are usually adequate. Concentrations typically range
up to 5 ppm. In many situations, pipeline companies add their own corrosion
inhibitors on top of that added by refiners.
Rust protection is measured using any one of a number of tests: National
Association of Corrosion Engineers (NACE) rust test (40), ASTM D 665, and
MIL-I-25017 B/C (41). These all involve immersing a steel spindle in a mixture
of gasoline and water for a specified time and evaluating the resulting corrosion.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 245
ASTM D 4914 lists the Copper Strip Corrosion Test which protects against the
corrosive effects of fuel sulfur on copper.
Antiicing Additives. Additives to control icing are rarely used in the
United States because vehicle design changes have made icing problems almost
nonexistent.
Detergent Additives. Deposits may build up in those parts of an engine
wetted by fuel: fuel injector, intake valves, and the combustion chamber itself.
Both fuel and vehicle design contribute to deposit formation. Proper choice of
additive or additives can prevent deposit buildup and in some cases can clean
up existing deposits. The exact mechanisms of deposit formation are poorly
understood. Moreover, the fuel undergoes a different time/temperature history
in each location leading to different types of deposits. An additive that is effective
in controlling one type of deposit may not work at all, or may even contribute to
increased deposits, in another engine location. For instance, solvent oils can be
particularly effective in preventing the buildup of intake valve deposits, but too
much solvent oil causes excessive combustion chamber deposits and leads to high
ORI. Detergents were first developed for carburetors, and the first was intro-
duced by Standard Oil Co. Of California in 1954 (42). Few if any carbureted
vehicles have been made since the mid-1980s, and today’s vehicles all use fuel
injectors to meter the fuel.
Fuel Injector Detergents. Port fuel injectors generally experience tempe-
ratures around 1008C. Deposits can accumulate in the small annulus in the tip
of the injector through which fuel must flow. As the annulus become plugged,
fuel flow becomes erratic, atomization degrades, and control of stoichiometry
becomes more difficult. Although well-designed injectors can tolerate plugging
levels of up to 30%, variability among injectors can cause problems even at
much lower levels. This is because the vehicle computer controls stoichiometry
for all cylinders on average. If one cylinder is running rich, then the other cylin-
ders run lean so that average stoichiometry is correct. Even partially plugged
injectors lead to poor driveability and higher emissions as well as loss of fuel
economy and engine power.
Although the exact chemical mechanism which leads to injector deposits
has not been identified, it is generally believed that heavy hydrocarbons oxidize
and polymerize during vehicle hot soak. The resulting gums and resins trap
other particles present in the crankcase blowby and in the EGR. Olefins and
diolefins are known to be significant contributors, whereas the role of sulfur is
still unresolved (43–45). Polar compounds have been shown to contribute to
injector deposits. In one experiment, a dirty deposit forming fuel was clay filtered
and the resulting filtrate had very low deposit-forming tendencies (46).
A number of additives have been developed which not only prevent injec-
tor deposits from forming, but which also remove deposits already in place. The
best additives can unclog clogged injectors in less than one tankful of gasoline.
Polybutene succinimides, polyisobutylene amines, polyether amines, and certain
lower molecular weight amines have been shown to be effective injector deter-
gents (43). Phenylenediamine, which is known to improve the oxidative storage
stability of gasoline, does not prevent the reactions which lead to fuel injector
deposits. Performance of additives in controlling fuel injector deposits is mea-
sured using ASTM D 5598. This is a vehicle based test and involves driving a
246 GASOLINE AND OTHER MOTOR FUELS Vol. 2
94 LBV1 94
92 92
Octane number
90 90
A ON1
88 88
LBV2
B 86
86
ON2
84 84
0 0.2 0.4 0.6 0.8 1
Fraction of component 1
Fig. 4. Octane blending behavior where (—) represents the measured octane response
curve, ( ) the tangent to the curve, and (– – –) the linear connection between the
octane number of components 1 and 2. The line AB corresponds to the octane blending
bonus of this composition.
raising pool octane often entails significant investment and increased operating
costs. Also, it is possible to meet targets for the different grades by properly choos-
ing blend stocks to take advantage of octane bonuses available from the non-
linear blending characteristics.
Blending behavior of a binary mixture may be characterized by a linear
blending value (LBV). Figure 4 shows the response curve of a hypothetical
two-component mixture. The LBV for each of the components at any composition
fc is defined by the tangent at that point according to the formula.
LBV 1 ¼ ON c þ ð1 f1 Þ ðdON=df1 Þc
LBV 2 ¼ ON c f1 ðdON=df1 Þc
which assumes that the lack of linearity results from the interaction of pairs of
components. An approach which focuses on the difference between the weighted
linear average of the components and the actual octane number of the blend
(bonus or debit) has also been developed (58). The independent variables in
this type of model are statistical functions (averages, variances, etc) of blend
properties such as octane, olefins, aromatics, and sulfur. The general statistical
problem has been analyzed (59) and the two approaches have been shown to be
theoretically similar though computationally different.
Most of the octane blending values reported in the literature use a slight
variation on this theoretically sound approach. The composition and octane of
the base fuel are assumed to be fixed and the second component is assumed to
be added. Using the same nomenclature, the blending octane number (BON) of
component 2 is defined as
The advantage of this definition is that it does not depend on measuring the
tangent of the response curve, although the variation in the value of the blending
octane number is greater. Typically, BONs are measured at an 80/20 mixture.
This technique is also useful when trying to measure the octane of a compound
such as butane or methanol that is difficult or impossible to measure in its pure
state.
Most refineries develop individual octane blending equations which do a
good job of predicting that refinery’s blending behavior. In order to use these
equations in refinery planning and operations, these may be linearized in a
piecewise fashion.
Until the 1960s, most gasoline blending was done in batches. All the com-
ponents were stored in large tanks, a recipe was defined by a master blender, and
an empty tank was filled according to the recipe. After the blend was finished, a
sample was sent to the laboratory for testing. If all properties were within spe-
cifications, the blend was released. If not, certain components would be added
to fix the blend properties. In the extreme, a blend could be downgraded from
premium to regular, it could be totally reblended, or it could be sold at a discount
to a wholesale customer willing to use off-spec material.
Today, many, if not all of the steps in blending and certifying gasoline have
been automated. Computers define preliminary recipes. Although ASTM certi-
fication tests are typically laboratory based, some have been adapted and auto-
mated for use in the refinery. As a blend is being made in the refinery, samples
may be taken automatically as often as every 20 minutes. These samples are fed
to special instruments which measure critical properties such as octane and vola-
tility. If the integrated values of the blend are not within specifications, then the
automated blenders adjust the recipe.
In a large refinery, an average blend might be 15 106 L , having an aver-
age flow rate though the blender of 106 L/h. When in-line testing is used, a large
number of repeat tests conducted on the blend as it is made, and the properties of
the blends may be known to a high degree of statistical confidence. The final
blends may be pumped directly into a pipeline or a barge.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 251
9.3
9.2
9.1
Fuel economy, km/L
8.9
8.8
8.7
8.6
8.5
30 30.5 31 31.5 32 32.5 33
Net heating value, MJ/L
Fig. 5. Fuel economy vs energy content for various fuels. Each data point represents the
average of twenty 1989 vehicles. To convert MJ/L to Btu/gal, multiply by 2:78 108 .
The Clean Air Act directs EPA to set National Ambient Air Quality Stan-
dards (NAAQS) for a number of pollutnats such as ozone, CO and NOx, and
sets or directs EPA to set vehicle emissions standards and fuel composition
standards to help meet these NAAQSs. Because of its especially serious air
pollution problems, the state of California is allowed to impose its own strict
set of standards under the Clean Air Act.
The first of these standards set by EPA went into effect in 1968 and man-
dated that the vapors from the vehicle crankcase be routed back through the
engine and burned. Since then, the standards have continued to grow stricter.
Table 8 shows the federal exhaust emission standards and Table 9 shows the
more stringent California standards. A novel feature of the latest federal and
California standards is that they must be met by each manufacturer on their
sales weighted average. This means that certain vehicles may emit more than
others, as long as the average meets the standard. For the federal program, man-
ufacturers must certify vehicles in various bins, which define standards for non-
methane hydrocarbons (NMHC), CO, NOx and PM, and the sales weighted fleet
must meet the NOx average shown in Table 8. In a similar way, the California
program defines various emission classes, and sales weighted averages must
meet a standard defined in terms of nonmethane organic gases (NMOG).
NMOG is very similar to the EPA’s NMHC.
The emissions standards have forced changes in vehicle hardware as shown
in Table 7. Many of these changes in hardware have resulted in changes in fuel
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 253
equation, and because very little data were available, the three domestic auto
companies and fourteen petroleum companies formed a joint research program
in late 1989 called the Auto/Oil Air Quality Improvement Research Program
(AQIRP). This four-year, $40 million program had the objective of ‘‘conducting
a research and testing program to develop data on potential improvements in
vehicle emissions and air quality, primarily ozone, from reformulated gasoline,
various other alternative fuels, and developments in vehicle technology’’ (63).
The program developed a great deal of information on the effects of fuel pro-
perties on exhaust and evaporative emissions (64–66). The data collected by
AQIRP represent the largest single body of information collected on this subject.
Over two thousand individual emission tests were conducted between 1990 and
1992. EPA used this and other data to generate an equation to predict emissions
as a function of fuel properties.
Final regulations were published after much consultation with the oil and
automotive industries, and with consumer and environmental groups. Specific
reduction targets were set for RFG, and gasoline manufacturers were given the
option of complying with the regulations on every gallon or on average. In gene-
ral, if averaging is used, the standard is more severe (1.5% for VOC, NOx, toxics)
and the performance of any one blend may not be too far below the regulatory
level (within 2.5% for VOC, NOx).
Federal RFG and wintertime oxygenated fuels must contain compounds
that add oxygen to the fuel (2% and 2.7% by weight respectively). These require-
ments have been met for the most part by using MTBE and to a lesser extent
ethanol. MTBE use has caused concerns because of underground tank leakage
and groundwater contamination. Some states have banned MTBE, and where
oxygenates are required by federal mandate, ethanol has been used exclusively.
There is considerable political pressure to repeal the portion of the CAA and asso-
ciated regulations that require oxygenate use. Technically, oxygenates have little
or no impact on emissions from modern vehicles, and do not reduce either ozone
or CO levels in the atmosphere (67).
The state of California has taken a different conceptual approach to redu-
cing emissions through control of gasoline composition. Instead of defining a
performance target, ie, 25% reduction, California has defined composition targets
which are aimed at achieving emissions reductions. The third round of regu-
lations, known as Phase 3, are shown in Table 10 and took effect on January 1,
2004. Gasoline producers must meet these targets on every liter of gasoline. If
desired, a company may choose to meet yearly average targets, but these are
slightly more severe than the per liter specifications, and the ranges which
may be used for averagers are restricted.
In order to provide refiners some flexibility in meeting requirements,
California developed the Predictive Model, which defines trade-offs among the
specifications, while maintaining equivalent emissions performance. Refiners
use the Predictive Model to define alternative specifications for their operation.
These alternative specifications, which can be changed as often as desired, are
used in place of the values in Table 9. For instance a refiner may choose a lower
sulfur level, say 10 ppm, to offset higher olefins, say 9%.
California regulations have been deemed equivalent to the Federal RFG
rules and supersede them except for the use of oxygenates which must be used
in areas that would otherwise require Federal RFG (Los Angeles, San Diego, and
the San Joaquin Valley). Since California has banned use of MTBE and other
ethers, ethanol is the only oxygenate used in the state.
In recent years, other government agencies have regulated gasoline speci-
fications and composition. Some examples are:
Area Controls
Detroit, Mich. Low RVP
Pittsburgh, Pa. Low RVP
Atlanta, Ga. Low RVP, low Sulfur
Birmingham, Ala. Low RVP, low Sulfur
Phoenix, Ariz. CA or Federal RFG, oxygen
not required
New York State MTBE banned
Connecticut MTBE banned
to full reformulations such as those adopted in the United States and Europe.
The choice of appropriate regulations should be based on many factors, including
vehicle technology, the status of air quality, and cost effectiveness relative to
other emission control steps.
3. Diesel Fuel
As a fuel for internal combustion engines, diesel fuel ranks second only to gaso-
line. In 2001, total U.S. demand for diesel fuel for transportation was about 150
billion L, about 29% of the demand for gasoline (68). Use of diesel fuel for high-
way vehicles accounted for 81% of the total, off-highway vehicles for 6%, ships for
5%, locomotives for 7%, and military use for 1% (69). Diesel cars, thought at one
time to be very promising, have encountered significant customer resistance in
the United States. There are significant obstacles to overcome. Consumers have
perceived diesels as noisy, smoky, and hard to start in winter. The fuel economy
benefit that diesel engines can offer is not as attractive when gasoline prices are
relatively low. Unites States automotive manufacturers are considering diesels
as a way to meet corporate average fuel economy standards and still be able to
sell large, powerful SUVs. Increased penetration of diesel engines in the light
duty market means that they will have to meet exceedingly strict emission stan-
dards and will have to compete with gasoline-battery hybrids for those interested
in higher fuel economy.
In many countries, diesels engines do not have the negative image that
they do in the United States. Furthermore, tax differentials on fuels and vehi-
cles, as well as overall high fuel prices can make diesel economically attractive.
In France, for example, diesel engines constituted 56% of new car passenger and
light duty truck sales in 2002, vs about 0.1% in the United States.
3.1. Combustion in Diesel Engines. Unlike the spark-ignited gaso-
line engine, the diesel engine, first used by Rudolf Diesel in the 1890s to burn
finely powdered coal dust, employs compression ignition. Liquid fuel was employ-
ed soon after. Through the action of the pistons, air alone is drawn into the cylin-
ders and compressed. Near the end of the compression stroke, fuel is injected into
the cylinder and after a short delay, is ignited by the high temperature gene-
rated during compression. The fuel must be finely atomized and well mixed
with the air for complete combustion to take place. The ignition delay is useful
because it allows the fuel to be injected and mixed before combustion starts. Its
length depends largely on the composition of the fuel, and is described by cetane
number.
Diesel engines offer a number of advantages over gasoline engines. Diesel
engines are able to operate at high (16–22) compression ratios which improves
energy efficiency. Because combustion starts at many different sites and there is
no flame progressing across the combustion chamber, there are no knock con-
cerns. The engine operates with unthrottled air flow, which also improves effi-
ciency. Because the air flow is unthrottled, changing fuel flow results in a wide
range of stoichiometry, always in the lean region. Lean operation also means
that HC and CO emissions are generally low for diesels. If too much fuel is fed
to the engine to boost power, visible black smoke and soot are emitted.
258 GASOLINE AND OTHER MOTOR FUELS Vol. 2
viscous and is used in low and medium speed engines having sustained loads at
substantially constant speed. Most cars and trucks use 2-D, a general-purpose
grade. 1-D, a more volatile, lower density, lower aromatic fuel, used in cold
weather and in municipal buses. Starting in 1993, EPA has lowered the regulat-
ed sulfur level of diesel fuel twice and ASTM has defined two new sets of grades-
low sulfur (LSD) and ultra low sulfur (ULSD). These are described in more detail
below.
Ignition Quality. The ability of diesel fuel to burn with the proper charac-
teristics is described by its cetane number, a measure of ignition delay. Exces-
sively long ignition delays (low cetane number) cause rough engine operation,
misfiring, incomplete combustion, and poor startability. Because the fuel starts
to burn later in the cycle, pressure rise is more rapid without increasing the net
work from the cycle. Power may be reduced and combustion may be incomplete
leading to high emissions if the fuel does not have time to burn in the expansion
stroke.
The procedure for measuring the cetane number of diesel fuel (ASTM D 613)
is similar to that used for measuring gasoline octane number. Cetane [544-76-3]
(n-hexadecane), C16H34, is defined as having a cetane number of 100; a-methyl-
naphthalene [90-12-0], C11H10, is defined as having a cetane number of 0. 2,2,4,-
4,6,8,8-Heptamethylnonane [4390-04-9] (HMN), C16H34, which can be produced
in high purity, is used as the low reference fuel and has a cetane number of 15.
Blends of cetane and HMN represent intermediate ignition qualities according to
the formula:
Flash Point. The flash point of a fuel indicates the temperature below
which the fuel can be handled without danger of fire. This is the temperature to
which the fuel must be heated to create sufficient vapors above the surface of the
liquid that they can be ignited in the presence of an ignition source. The flash
point of diesel fuel is measured by ASTM D 93, using a closed-cup Penske-
Martens tester. The fuel is heated at a rate of 5.68C/min and a test flame is intro-
duced into the test chamber at 30 s intervals. The fuel temperature reached
when the flame ignites the vapors is called the flash point. Specifications for
flash point vary with grade; the lowest value is 388C for grade 1-D. Controlling
flash point is important in order to prevent the vapor space in storage and vehicle
tanks from being in the explosive range. Setting the flash point at 388C protects
most storage vessels from exploding.
Carbon Residue. The tendency of a diesel fuel to form carbon deposits in
an engine can be roughly predicted by the Ramsbottom Coking Method (ASTM D
524), which determines the amount of carbon residue left after evaporation and
chemical decomposition of the fuel at elevated temperatures for a specified length
of time. The Ramsbottom method involves heating the last 10% residue of the
ASTM distillation at 5498C for 20 minutes. The result is reported as percent
carbon on 10% residuum. For use in high speed diesel engines operating over
a range of loads and speeds, ASTM specifications call for no more than 0.15%
Ramsbottom carbon residue. Because ignition control additives may interfere
with the tests without actually leading to deposit formation in the engine, these
laboratory tests should be made on fuel before any addition of ignition control
agent.
Sulfur. Sulfur in diesel fuel should be kept below set limits for both en-
vironmental and operational reasons. Operationally, high levels of sulfur can
262 GASOLINE AND OTHER MOTOR FUELS Vol. 2
lead to high levels of corrosion and engine wear owing to emissions of SO3 that
can react with condensed water during start-up to form sulfuric acid. From an
environmental perspective, sulfur burns to SO2 and SO3, the exact split being
a function of temperature and time in the combustion chamber. SO3 can react
with water vapor in the exhaust which can further react to form sulfates, leading
to high levels of particulates. Although the SO3 fraction may be small, diesel fuel
has traditionally contained sulfur concentrations as high as 5000 ppm. As parti-
culate emission standards for diesels became more stringent, engine manufac-
turers made the case that they could not meet the standards unless sulfur
levels were reduced. Through negotiations between the Engine Manufacturers
Association and API, it was agreed that sulfur content of highway diesel could
be reduced, and EPA specified a maximum level of 500 ppm starting in October,
1993. In December 2000, EPA promulgated new emissions standards for diesel
engines and concluded that these standards could only be met if diesel fuel for
on-road engines were also reduced. Starting in 2006 and phasing in fully by
2010, diesel fuel for on-road engines must have a sulfur level below 15 ppm.
Starting in 2010, diesel fuel for nonroad engines, such as farm and construction
equipment, must also be below 15 ppm. Manufacturers are expected to
target refinery release limits of approximately 8 ppm to take into account the
sulfur picked up in pipelines and other parts of the distribution system. This margin
may change over time as experience is gained in shipping low sulfur products.
Ash Content. The fuel injectors of diesel engines are designed to very
close tolerances and are sensitive to any abrasive material in the fuel. Therefore,
the maximum permissible ash content of the fuel is specified. This is measured
by ASTM D 482, which consists of burning a small sample of the fuel in a weighed
container until all the carbonaceous matter has been consumed. The permissible
amount of ash is between 0.01 and 0.1 wt% depending on the grade of diesel.
Low speed engines operating at constant speed and load can tolerate higher
levels of ash.
Aromatics Content. Aromatic compounds have very poor ignition quality
and, although they are not specifically limited in ASTM D 975, there are practi-
cal limitations to using high aromatic levels in highway diesel fuel. In the United
States, where gasoline demand represents about one-half of the crude barrel, and
where heating oil demand is relatively low, there is great pressure to be able to
blend aromatic FCC streams into diesel fuel. Average aromatic levels in the
United States are about 33%. In addition to having poor fuel quality, aromatics
also contribute to exhaust emissions. The federal government began effectively
limiting aromatic content to below 40% starting in October, 1993 by specifying
a minimum cetane index of 40. California limits aromatic levels below 10%
beginning in the same time period, also because of emissions concerns. It is
expected that average aromatic levels in ULSD will be lower than current levels
because the refinery processes used to reduce sulfur will also saturate some
aromatics. The magnitude of this change is difficult to predict.
Stability. Diesel fuel can undergo unwanted oxidation reactions leading to
insoluble gums and also to highly colored by-products. Discoloration is believed
to be caused by oxidation of pyrroles, phenols, and thiophenols to form quinoid
structures (71). Eventually, these colored bodies may increase in molecular
weight to form insoluble sludge.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 263
Gums can lead to deposits in critical injector orifices that can degrade atomi-
zation and combustion performance of the engine. The chemistry involved in the
formation of these gums appears to be similar to that which occurs in gasolines.
Stability is measured using ASTM D 2274. In this test, a sample is heated at
958C for 16 hours while oxygen is bubbled through the liquid. After cooling,
the insoluble material is filtered and washed. The amount remaining is reported
as gums.
Lubricity. Diesel fuel has traditionally had adequate lubricity to protect
fuel pumps and injectors from excessive wear. Changes in equipment and fuel
processing have caused concern that lubricity may no longer be adequate. Pres-
sures in modern unit fuel injectors may be as high as 170 MPa (25 kpsi) and
tolerances for direct injection equipment have become much tighter. Lower visco-
sity fuels have poorer lubricity, and processing steps that lower sulfur may also
remove trace levels of surface active compounds that provide natural lubricity.
Lubricity is measured in two different tests: ASTM D 6078, the scuffing
load ball-on-cylinder evaluator (SLBOCLE); and ASTM D 6079, the high fre-
quency reciprocating rig (HFRR). In response to concerns voiced by engine and
equipment manufacturers, ASTM will likely adopt a maximum specification of
520 mm average wear scar in the HFRR test. Fuels which do not meet this
specification will have to be additized, either at the refinery or at the terminal.
ASTM specifications for different grades of diesel fuel are summarized
in Table 12.
3.3. Diesel Fuel Manufacture. The biggest factors in determining how
diesel fuel is blended in a given refinery are the availability of high cetane stocks.
In order of decreasing ignition quality, the hydrocarbon types rank in the follow-
ing order: normal paraffins, olefins, cycloparaffins, branched paraffins, and aro-
matics. Because straight-run distillates contain the greatest amount of normal
paraffins and cycloparaffins and the least amounts of branched paraffins and
aromatics, these are the preferred stocks for diesel blending. Cracked stocks,
which are relatively rich in aromatics, are less desirable from the standpoint
of ignition quality. However, these have high energy density and good cold
temperature properties. The amount of cracked material allowable in diesel
fuel depends largely on the cetane number specification. In the United States,
where a high level of cracking is necessary to meet gasoline demand, the large
supply of cracked fractions and the relatively small supply of straight-run distil-
lates make substantial use of cracked stocks economically necessary. This has
been made possible through the use of cetane improvers to improve cetane and
through the use of hydrogenation to improve stability.
Ignition Improvers. In order to meet the increasing demand for diesel fuel
and to allow use of blendstocks having good low temperature properties but
low cetane number, diesel manufacturers frequently use cetane improvers in
the fuel. Cetane improvers work in just the opposite way that antiknock addi-
tives do. During the preignition period, the improvers generate free-radical
species that promote faster onset of combustion and reduce the ignition delay
properties of the base fuel. Various types of chemicals have been shown to be
effective, including nitrates and nitrites, nitro and nitroso compounds, and per-
oxides (71). Alkyl nitrates are by far the most common and are used in concen-
trations up to 0.3%. Although the effectiveness varies with base fuel composition,
0.1% in a typical fuel would give an increase of five numbers (72).
Stability Improvers. Diesel fuels that contain high amounts of cracked
stocks generally have poorer stability than virgin diesel. If the fuel is hydro-
treated to remove sulfur, the stability is vastly improved. Traditional antioxi-
dants such as hindered phenols are not particularly effective in preserving color
although these still prevent other oxidation reactions. Stabilizers are amines
or other nitrogen-containing compounds that prevent sediment formation by
interfering with the oxidation reactions that occur between heteroatoms
and available oxygen. These act as radical traps and/or peroxide decomposers.
Corrosion Inhibitors. The corrosion inhibitors used in diesel fuel are gene-
rally similar to those used in gasoline and, like the latter, produce an effect pri-
marily by surface action. If amine additives are used for detergency, these may
provide some corrosion protection as well.
Detergent Additives. Diesel engine deposits are most troublesome in the
fuel delivery system, ie, the fuel pump and both fuel side and combustion side of
the injectors. Small clearances and high pressures mean that even small
amounts of deposits have the potential to cause maldistribution and poor atomi-
zation in the combustion chamber. The same types of additives used in gasoline
are used in diesel fuel. Low molecular weight amines can also provide some cor-
rosion inhibition as well as some color stabilization. Whereas detergents have
been shown to be effective in certain tests, the benefit in widespread use is not
fully agreed upon (73).
Cold Flow Improvers. The cold flow properties of diesel fuel may be
improved by the use of additives or blending agents. As is the case for gasoline,
additives are materials present in low concentrations (ppm), whereas blending
agents are present in levels of a few percent or more. Blending agents work
by diluting the concentration of wax crystal forming paraffins to a point
below which these paraffins are no longer a problem. Kerosene and grade 1-D
diesel fuel may be used as effective diluents. Gasoline, which also improves
cold flow properties, is unsafe and dangerous to use as a blending agent
because it can drastically lower the flash point of the mixture and lead to poor
lubricity.
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 265
Additives are often the most economical way to improve the cold flow pro-
perties of middle distillates. These additives have been called by a number of
different names including wax crystal modifiers, mid-distillate flow improvers,
and wax antisettling additives. All additives work by modifying the crystalliza-
tion process in some way. Some change the shape of the wax crystals so that the
crystals are less likely to plug fuel system filters. Others work by promoting
nucleation of many small crystals instead of the growth of relatively few large
crystals (74). Various types of chemistry have been shown to be effective in im-
proving low temperature flow properties of diesel fuels. Examples are ethylene–
vinyl ester copolymers, chlorinated hydrocarbons, and polyolefins. The exact
nature of most commercially available additives is proprietary. Treatment rates
and the specific additive formulation used are a strong function of the fuel
composition and depend in large part on the distribution of molecular weights
of the normal paraffins in the fuel and the concentration of high molecular
weight paraffins (75). Fuels having a wide distribution of molecular weights
are the most responsive to flow improver additives.
Three types of cold flow additives have been described (76). The first are
pour point depressants, which are primarily low molecular weight ethylene–
vinyl ester copolymers. These work by cocrystallizing with the paraffin wax and
preventing further paraffin addition to existing crystals. The second type of addi-
tive is the cloud point depressant (CPD). CPDs are low molecular weight poly-
mers where the paraffin segments have been designed to interact with the
paraffins in the fuel to delay the onset of crystallization. The third type of addi-
tive is called an operability additive (OA). OAs are multicomponent additives
optimized for given fuel properties and show an improvement in the cold filter
plugging point and low temperature flow tests as well as lowering the pour point.
Lubricity Improvers. Lubricity performance of modern low sulfur, low aro-
matic diesel fuel in modern diesel engines is a concern. Some of these concerns
surfqaced in Sweden with the introduction of Swedish Class 1 and Class 2 diesel
in the early 1990s. A number of additives are available which can improve
lubricity as measured in ASTM D 7078 mand D 6079 at concentrations of
100–200 ppm. Use of fatty acid methyl esters has also been shown to improve
lubricity.
Dyes. Azo type dyes are fairly common in diesel fuel, and red dye is used
to identify fuel for nonroad applications, which is not subject to highway taxes.
This practice makes it easy to identify cheating and to enforce tax regulations.
Any red diesel found in nonroad equipment is evidence of cheating. Clear diesel
found in nonroad equipment usually means that on-road fuel was downgraded at
some point in the distribution system.
Antifoamants. When diesel fuel is splashed into tanks, it can foam up to
such an extent that the foam can splash onto the customer or onto the ground.
Antifoamants are silicon based additives that reduce foaming tendnecies so that
a high fueling rate can be maintained. Treat rates are usually below 30 ppm.
Oxygenates. Oxygenated materials have been considered for addition to
diesel fuels for the same reasons these compounds are added to gasoline. Put-
ting oxygen in the fuel should lead to better combustion and reduced emissions
of CO and, more importantly, particulates. However, use of oxygenates has
also been shown to increase emissions of nitrogen oxides. A number of different
266 GASOLINE AND OTHER MOTOR FUELS Vol. 2
oxygenates have been tested, including esters; ethers (qv); and alcohols. The
most common oxygenates used or advocated for use in diesel fuel are fatty
acid methyl esters (FAME). FAME can be made from vegetable oils such as
soy (SOME), or rapeseed oil (RSME), or from animal fat, such as cooking
grease. The term biodiesel generally represents FAME made from a renewable
oil. If pure biodiesel is used, it is called B100. Other concentrations, such as B5
or B20, can be used too.
FAME has a number of advantages. It has a high cetane number and good
lubricity. Emissions of hydrocarbons, and particulates are significantly lower
than standard diesel fuel. Finally, use of biodiesel can reduce emissions of green-
house gases.
FAME also has a number of disadvantages. Emissions of NOx are higher
than commercial diesel. Cold flow and stability may also be worse.
Generally, engine manufacturers have supported FAME use at levels below
5%, if the material meets the requirements of ASTM D 975 and if the ester meets
the requirements of ASTM D 6751. Use of biodiesel has been limited primarily
because of its high cost.
3.4. Diesel Environmental Regulations. Emission standards have
been set for heavy-duty vehicles in much the same manner as they have been
set for gasoline engines. Because heavy-duty vehicles are primarily diesels, the
focus is on diesel engine emissions. Standards have been written in units of
grams per brake-horsepower-hour (g/bhph ¼ g/kW h 1.34), which normalize
the emissions according to the total energy output of an engine over the specified
driving cycle. In contrast to light-duty vehicle testing where testing is carried
out on the total vehicle, heavy-duty engines are certified in tests on an engine
dynamometer. A series of accelerations is carried out and the emissions are
measured. Table 13 shows U.S. emissions standards. For heavy-duty engines,
the most difficult emissions standards to meet are total particulates and NOx.
When standards were relatively high, they were met by engine redesign without
fuel changes. Since 1994, fuel changes have also been specified. Many studies
have been published on the interactions between engine design, fuel composition,
and emissions, and the conclusions are not always in agreement. It is generally
agreed that sulfur contributes to sulfate particulate emissions and that lowering
sulfur helps to meet the particulate standards. Higher aromatics, especially
PAHs (polycyclic aromatic hydrocarbons) are known to contribute to organic
carbon based particulates. Particulate emissions have also been shown to be
affected by fuel density (77). The impact of cetane has been shown to be positive
and negative. In older heavy duty engines, higher cetane was generally cor-
related with lower levels of NOx and particulates. In modern heavy duty engines,
this effect seems to be negligible. In modern European light duty diesels, higher
cetane can actually increase particulate emissions.
It is also generally agreed that in terms of engine design parameters, there
is an inverse relationship between NOx and particulates. Rapid, complete combu-
stion reduces particulate emissions but also promotes the formation of NOx.
Exhaust gas recycle (EGR), which lowers NOx emissions, generally causes
particulate emissions to increase.
It is expected that diesel engine manufacturers will use exhaust aftertreat-
ment devices to meet the NOx and particulate emissions standards that take
effect in 2007–2010. Particulates will likely be controlled by catalyzed traps; NOx
will be controlled either by NOx reduction catalysts that operate under lean con-
ditions or by catalysts that use a reductant such as urea. In anticipation of these
devices, and in anticipation that the devices would only work with extremely low
levels of sulfur, EPA enacted the ULSD regulations described above.
California has taken a slightly different approach to diesel fuel composition
standards from that of the federal government. In October, 1993, California limi-
ted diesel fuel to no more than 10% aromatics and 500 ppm sulfur. Alternative
formulations are possible if these are shown to have equivalent NOx emissions to
a base reference fuel. In addition to the specifications that apply to the commer-
cial fuel, other aspects of the reference fuel composition are tightly controlled.
The fuel must also have a minimum cetane number of 48 without cetane impro-
vers. Fuel manufacturers test the emissions from proposed alternatives and com-
pare them to the emissions from the reference fuel. For instance, manufacturers
may be able to produce higher aromatic fuel if the fuel also has higher cetane
number. In 2003, most diesel fuel in California used an alternative formulation.
Based on a small sample, average aromatics were 20%, average cetane was 51
and average sulfur was 40 ppm in the Los Angeles area in the Summer of 2003
(78). Starting in 2006, California is adopting the federal sulfur limit of 15 ppm
in addition to its other requirements.
In Europe, the European Union has defined diesel composition standards as
shown in Table 14.
4. Alternative Fuels
Alternative fuels fall into two general categories. The first class consists of fuels
that are made from sources other than crude oil but that have properties the
same as or similar to conventional motor fuels. In this category are fuels made
from coal, shale and natural gas through Fischer-Tropsch synthesis. (see FUELS,
SYNTHETIC, GASEOUS FUELS). In the second category are fuels that are different from
gasoline and diesel fuel and which use redesigned or modified engines. These
include compressed natural gas (CNG), liquefied petroleum gas (LPG) and
hydrogen. The most probable use of hydrogen would be in fuel cells vehicles,
both of which will be discussed elsewhere.
During World War II, Germany developed large-scale production of fuel
from coal using Fischer-Tropsch synthesis. In the United States, in the 1970s,
interest in producing alternative fuels was generated by upheavals in the price
and availability of crude oil. In addition to coal, a great deal of research was
carried out to develop processes to develop fuels from oil shale and natural gas.
Production of synthetic fuels from coal, oil shale, and methane involves
changing the chemical structure of the raw material, especially the hydrogen–
carbon ratio. In coal, the hydrogen–carbon ratio is too low and the molecular
weight of the raw material is too high to burn cleanly. Coal can be converted
to syngas (CO/H2) and then converted to liquids via Fischer-Tropsch synthesis.
Alternatively, it may be liquefied directly with the addition of hydrogen. Oil
shale is retorted in large reactors or in situ to convert the solid organic material
(kerogen) into a synthetic crude oil. The oil is processed to resemble crude frac-
tions. Methane may be converted to hydrocarbon liquids or to methanol via gasi-
fication to syngas (CO/H2) followed by Fischer-Tropsch reactions. A principal
factor in all of these processes is the amount of energy needed to run the process
and the disposition of the waste products. In the case of shale oil, the spent rock
takes up more volume than the original shale and has to be disposed of in an
environmentally safe manner.
Decisions on commercialization depend on a number of factors, including
the price of the new energy source relative to the price of crude oil, need for
energy self-sufficiency, and environmental considerations. The factors are differ-
ent for each of the potential sources. In the case of natural gas, conversion to
Fischer-Tropsch liquids must compete with the production of liquefied natural
gas (LNG). Environmental considerations include pollution from burning the
fuel itself, pollution from the whole production cycle, and production of gases
that are involved in the enhanced greenhouse effect (see AIR POLLUTION). With
the dramatic improvements in the low emissions capability of gasoline vehi-
cles, the incremental improvement offered by some alternative fuels can be very
expensive and has tended to limit their introduction. Today, most alternative
fuels are being considered for their ability to reduce emissions of greenhouse
gases and the desire to reduce dependence on petroleum based fuels.
Fischer-Tropsch liquids made from natural gas (GTL) seem to be the most
likely fuel in this category to reach the market. GTL plants can produce a very
high quality fuel in the diesel boiling range. Sulfur and aromatics are essentially
zero, and cetane numbers can be as high as 75. The most efficient use of this
Vol. 2 GASOLINE AND OTHER MOTOR FUELS 269
BIBLIOGRAPHY
‘‘Gasoline and Other Motor Fuels’’ in ECT 3rd ed., Vol. 11, pp. 652–695, by J. C. Lane,
Ethyl Corp.; ‘‘Refinery Processes, Survey’’ under ‘‘Petroleum’’ in ECT 3rd ed., Vol. 17,
pp. 183–256, by C. E. Jahnig, Consultant; in ECT 4th ed., Vol. 12, pp. 34–388, by Albert
M. Hochhauser, Exxon Research and Engineering Co.; ‘‘Gasoline and Other Motor Fuels’’
in ECT (online), posting date: December 4, 2000, by Albert M. Hochhauser, Exxon
Research and Engineering Co.
CITED PUBLICATIONS
GENERAL REFERENCES
ALBERT M. HOCHHAUSER
Exxon Mobil Research and Engineering Co.
Vol. 2 AVIATION TURBINE FUELS 273
Turbine-powered aircraft were used by both British and German air forces for
military operations during the latter part of World War II. The British engines
were operated on illuminating kerosine containing 1 % light lubricating oil (jet
engine oil) (1). The German engines burned a wider boiling mixture of gasoline
and heavier distillate fractions. Fuel development in Germany stopped at war’s
end. Continuing British research has tightened some requirements over the
years, yet today’s primary British specifications still describe kerosine-type
fuels for both commercial and military aircraft, the main difference being in
the additive content.
The initial United States military fuel was a low freezing point kerosine
called JP-1, but its limited availability forced the development of a wider boiling
fuel that included gasoline as well as kerosine fractions (2). This ‘‘wide-cut‘‘ fuel
called JP-4 was, until recently, the major military grade. However, safety concerns
associated with its flammability led to its phase-out in favor of the kerosine based
grade JP-8. The switch from JP-4 to JP-8 was largely completed by the end of 1996.
Commercial aircraft have been burning mostly kerosine-type fuels since
1953 when the early turboprop aircraft came into service, and this has remained
the major civil fuel in the turbojets and later the turbofans which power most of
the world’s civil aircraft fleet. The principal grades are Jet A in the United
States, TS-1 in Russia and Jet A-1 in the rest of the world. The wide-cut fuel
Jet B, analogous to the military grade JP-4, is now rarely used in civil aviation.
The most widely used civil aviation turbine fuel grades are defined by a small
number of key specifications as listed in Table 1. These specifications are similar
in most respects although a small number of significant differences remain.
The US Jet A fuel, defined by specification ASTM D1655 (3), differs from Jet
A-1 only in freezing point. The British DEF STAN 91-91 specification (formerly
well known as DERD 2494 whilst under the jurisdiction of the UK Ministry of
Defence) (4) is recognized as a civil jet fuel specification and is used as such in
many countries around the world. The Russian GOST 10227 specification also
defines a grade designated RT which is very similar to western Jet A-1, as is
the Chinese no.3 Jet Fuel.
In addition to these prime specifications, the International Air Transport
Association (IATA) maintains fuel ‘‘Guidance Material’’ (5) for Jet A-1 and
Jet B grades for the use of its member airlines. Because of the worldwide nature
of commercial aviation, the close correspondence between specifications is a
necessity to discourage the proliferation of many different national or local
specifications. The major jet fuel suppliers subscribe to a document entitled
‘‘Aviation Fuel Requirements for Jointly Operated Systems’’ (AFQRJOS) which
defines a ‘‘Check List’’ which comprises the most restrictive requirements of the
ASTM D 1655, DEF STAN 91-91 fuel specifications and IATA Guidance Mate-
rial. Thus fuel meeting the Check List automatically meets all the requirements
of the parent specifications. This document was developed to cater for the situa-
tion where many different fuel suppliers deliver fuel into the common storage
facilities which predominate at most international airports.
Further efforts are underway to minimize the differences between the
major fuel specifications with the eventual aim of developing a single, interna-
tional specification to define aviation turbine fuel. An International Civil Jet
Fuel Specification Liaison Group was formed in 1989 to promote this activity
and a growing number of national specification bodies are contributing to its
progress.
2.1. Specification Requirements. Details of the major fuel specifi-
cations for Jet A-1 type fuel are shown in Table 2. Although test methods in
most western countries are either identical or very similar, no international
standardization exists and direct comparisons between limits may be misleading
(refer to the notes applying to Table 2). Test method designations published by
the principal standards organizations are given in Table 3.
2.2. Fuel Properties. Unlike most chemicals but like other petroleum
fuels, aviation turbine fuels are not defined by composition, molecular structure,
and purity. Instead their characteristics are limited by boundary conditions
imposed by inspection tests. Some of these tests are fundamental, others empiri-
cal; many reflect a solution to past engine or aircraft problems. However, any fuel
meeting all requirements is acceptable for aircraft operation. The specifications
are dynamic, but changes are conservative because of the potential risk to avia-
tion operations. Long equipment life (
20 years) and certification restrictions
(see Chap. 5) also mitigate against drastic specification changes.
The individual specification properties or requirements can be divided
rather arbitrarily into bulk and trace properties. Bulk properties are those
requiring significant composition changes, involving more than 5 vol% of the
total fuel, to effect a change in the property. Trace properties tend to respond
to changes < 1 vol% or even 1 ppm or less. Changes in bulk properties affect pro-
duct yield, whereas changes in trace properties do not. The grouping has no
direct bearing on fuel performance. Instead it helps assess effects of crude com-
position and manufacturing steps on product yield (8).
Bulk Properties. Density. The mass – volume relationship is important
because the quantity of fuel loaded onboard an aircraft is metered volumetri-
cally, but fuel consumption is measured by mass. This relationship is also used
Table 2. Major Fuel Specifications for Civil Jet A-1 Type Fuel
United States United Kingdom International Russiaa Chinab
Specification ASTM D1655 DEF STAN 91-91 Checklist GOST 10227-86 GB 6537-94
Issue 96b 2 16 Amendment 4 effective June 95
Date Jul. 96 May 96 Jun. 96 Apr. 96 Aug. 94
Grade Jet A-1 Jet A-1 Jet A-1 TS-1 No.3 Jet Fuel
Composition
Acidity, total, mg KOH/g (max) 0.10 0.015 0.015 0.009c 0.015
Aromatics, vol % (max) 25 22.0 b 22.0 b 22 mass % 20.0
Olefins, vol % (max) NRe NRe NRe iodine no.2.5 max 5.0
Sulfur, total, mass % (max) 0.30 0.30 0.30 0.25 0.20
Sulfur, mercaptan (thiol), 0.003 0.003 0.003 0.005 0.002
mass % (max)
or
Doctor test negative negative negative NRe negative
Volatility
275
Distillation:
Initial boiling point (IBP), 8C (max) NRe report report 150 report
10 % vol recovered at, 8C (max) 205 205 205 165 205
20 % vol recovered at, 8C (max) NRe NR report NRe report
50 % vol recovered at, 8C (max) report report report 195 232
90 % vol recovered at, 8C (max) report report report 230 report
Final boiling point (FBP), 8C (max) 300 300 300 98 % (250 8C) 300
Residue, vol % (max) 1.5 1.5 1.5 NRe 1.5
Loss, vol % (max) 1.5 1.5 1.5 NRe 1.5
Flash pointf , 8C (min) 38 40 (38) 40 (38) 33 (28) 38
Density at 15 8C , kg/m3 775–840 775–840 775–840 775 (min) (20 8C) 775–830 (20 8C)
Fluidity
Freezing point, 8C (max) 47 47 47 50 g 47
Viscosity at 20 8C , mm2/s (min) NR NR NR 1.25 1.25
Viscosity at 20 8C, mm2/s (max) 8.0 8.0 8.0 8 ( 40 8C ) 8.0
Table 2. (Continued)
United States United Kingdom International Russiaa Chinab
Combustion
Specific energy (net), MJ/kg (min) 42.8 42.8 42.8 42.9 42.8
Smoke point, mm (min) 25 25 25 25 25
or
Smoke point, mm (min) 18 19 19 NRe 20
and naphthalenes, vol % (max) 3.0 3.0 3.0 NRe 3.0
or
Luminometer no. (min) 45 45 45 NRe 45
Corrosion
Copper corrosion (max) 1 1 1 passes 1
Silver corrosion (max) NRe 2h 2 NRe 1
Stability
Thermal stability: not comparable
276
JFTOT test temp., 8C 260i 260 260 260
Filter pressure drop, mm Hg (max) 25 25 25 3.3 kPa
Tube deposit (visual) rating LT 3 LT 3 LT 3 LT 3
Peacock/abnormal deposits fail fail fail fail
Contaminants
Copper content ** (max) NRe NRe NRe NRe 150 mg/kg
Existent gum, mg/100mL (max) 7 7 7 5 7
Water reaction
Interface rating (max) 1b 1b 1b 1 1b
Separation rating NRe NRe NRe NRe report
Microseparometer (MSEP) (min) NRe 85 (þ SDA ¼ 70) 85 (þ SDA ¼ 70) NRe NRe
Particulates, mg/L (max) NRe NRe NRe none report
Ash content, mass % (max) NRe NRe NRe 0.003 NRe
Electrical conductivityk *, pS/m 50–450 50–450 50–450 50–600 50–450 (20 8C)
Additives
Antioxidant (hydroprocessed fuels) optional mandatory mandatory mandatory optional
Antioxidant (other fuels) optional optional optional optional optional
Metal deactivator optional optional optional NRe NR e
Static dissipator optional mandatory mandatory optional optional
Fuel system icing inhibitor permitted permitted permitted permitted permitted
Lubricity improver permitted permitted permitted mandatory in HP permitted
a
Specification limits are in terms of Russian GOST test methods and may not be directly comparable with limits obtained using ASTM test methods.
b
Specification limits are in terms of Chinese GB/T or SH/T test methods and may not be directly comparable with limits obtained using ASTM test methods.
c
Value in terms of mg KOH/g calculated from GOST limit which is in terms of mg KOH/100 mL.
d
25.0 % max aromatics are permitted provided that hydrogen content is reported also.
e
NR = no requirement under this heading.
f
Flash point limits are in terms of ASTM D56 (TAG) test method; limits in parantheses are in terms of other, local, test methods.
g
The GOST test method records wax appearence point on cooling which is 3 8C below the wax disappearance point on warming, as recorded in the ASTM
method. Thus the specification limits for TS-1 and Jet A-1 are equivalent.
h
It is planned to delete the silver corrosion clasification requirement by June 1997.
i
If the test at 260 8C produces a failure, a pass at a test temperature of 245 8C is acceptable.
j
Applies only if fuel produced by copper-sweetening process.
k
277
Conductivity limits apply only if fuel contains static dissipator additive.
278 AVIATION TURBINE FUELS Vol. 2
extensively in commercial sales of fuel. Although in the past there have been sev-
eral different ways of defining this relationship, such as specific gravity, relative
density, or API gravity, all the major jet fuel specifications now define density at
15 8C in units of kg/m3. Changes of density with temperature are given in (9).
Density is commonly measured using hydrometers or digital density meters.
Vol. 2 AVIATION TURBINE FUELS 279
In-line density measurement is also possible and offers the opportunity for fuel
delivery by mass instead of volume.
Specific Energy. Specific energy or heat of combustion is directly related
to density. The fuel’s specific energy, its density, the aircraft’s performance char-
acteristics, and the predicted flight conditions are employed to calculate the
amount of fuel required for a given flight or conversely to estimate the aircraft’s
range with maximum fuel aboard. Specific energy is determined by measuring
the total heat released by a unit mass of fuel and subtracting from this value
the heat of condensation of the water vapor formed during the combustion. How-
ever, this measurement is time consuming, so several empirical relationships
have been developed to calculate specific energy from combinations of other
inspection tests. The most common is based on the product of aniline point and
density, ASTM D 4529 (3); others use the density, boiling range, and aromatic
content, ASTM D 3338 (3), or the hydrogen content and density, IP 355 (6). All
three methods include corrections for fuel sulfur content. Although these calcu-
lation methods are used routinely, a method of measurement, ASTM D 4809 (3),
is still specified as the referee method for use in cases of dispute.
Volatility. A fuel’s volatility or its tendency to vaporize affects evapora-
tion rate from tanks, including aircraft fuel tanks. Volatility can cause vapor lock
in fuel systems when the combined fuel and vapor volume exceeds pump capa-
city; cavitation can result when the fuel vapor pressure exceeds the liquid system
pressure between gear teeth or in plain bearings. Volatility also directly affects
flammability and causes safety problems both inside and outside of tankage (10).
Because of the complex composition of turbine fuel, its volatility cannot be
described by a single fundamental property. Therefore, volatility is normally
expressed using a combination of boiling range, flash point, or vapor pressure.
The volatility of kerosine type fuels is controlled by flash point and distilla-
tion range (Table 2). Wide-cut fuel is regulated by a combination of distillation
range and vapor pressure.
Combustion. Combustion in a modern jet engine is a highly turbulent,
high-pressure,and high-intensity mixing process. Engine designers carefully
control air flow, fuel flow, and fuel atomization to maximize combustion efficiency
and to minimize smoke, carbon deposits, and the formation of partially oxidized
gases, such as carbon monoxide. Fuel affects all these as well as the intensity of
radiation emitted by the flame [11–13]. No single fuel property requires a greater
degree of empiricism than the estimation of fuel combustion characteristics.
Smoke point, and to a lesser extent the luminometer, have been favored because
as actual combustion tests, they respond to a combination of other fuel properties
including volatility, composition, and viscosity. However, a direct comparison of
these laminar, wick-fed flames with the high pressure, turbulent combustor
flame is not possible and thus other alternative controls also appear in the spe-
cifications. Naphthalenes content was added because multi-ring aromatics have
shown adverse combustion effects, in terms of flame radiation and smoke forma-
tion, out of proportion to their concentration (14). During the 1980s, the use of
hydrogen content, which can be measured very accurately by NMR (ASTM
D 3701), gained favor as a possible replacement for the less precise measure-
ments of aromatics and naphthalenes contents. However, it was found that
the total hydrogen content of the fuel could not predict reliably the combustion
280 AVIATION TURBINE FUELS Vol. 2
behavior of all fuels(15) and it is now unlikely that it will be adopted for
civil fuels, although it continues to find limited use in certain military fuel
specifications.
Total aromatic content (ASTM D1319, (3) ), covered under ‘‘Elastomer Com-
patibility’’, does not correlate particularly well with engine combustion, possibly
because of the particular separation and classification of hydrocarbons by the
method (16).
In general, correlations between engine combustion performance and fuel
inspection tests will always be difficult to establish because of the differing
performance of the same fuel in different engines, the extreme expense of
controlled engine tests, and the long time required to obtain engine experience
in the commercial world, where only a limited quantity of minimum quality
fuel is burned (16).
Elastomer Compatibility. A major requirement of jet fuels is their com-
patibility with materials of construction, particularly with the elastomers used in
diaphragms, fuel tanks, fuel tank coatings and sealants, linings, and a wide vari-
ety of static and dynamic seals in all parts of the fuel system. The fuel property
most indicative of elastomer compatibility has been total aromatic content.
Low-Temperature Operability. The extremely low temperatures at
high altitudes can cause several different fuel problems. During fuel cooling,
hydrocarbon molecules precipitate as wax crystals. n-Hydrocarbons are the
most insoluble, and in fact measurement of their type and concentration can
indicate the insoluble fraction expected at low temperature (17).
The freezing point of the fuel is defined as the temperature at which the
last wax crystal melts on warming the fuel following a specified cooling process.
Traditionally, this has been determined visually but now several automated
methods exist which employ variously, light scattering, phase transition mea-
surement or flow through a fine filter (ASTM D 4305) (3).
With the increasing use of twin-engined aircraft on long over-water routes,
the so-called extended operations (ETOPS) flights, and with the introduction of
very long range aircraft such as the Boeing 747–400, there has been increasing
focus on fuel low-temperature properties. Studies underway may question the
usefulness of ‘‘freezing point’’ as currently measured and could lead to its repla-
cement by a more meaningful, flow-based test.
Low fuel temperature can also affect engine starting, which depends
heavily on fuel atomization. Adequate atomization is determined largely by
fuel viscosity. Typical viscosities versus temperature are plotted in Figure 1.
Because fuels within the same grade have the same viscosity-temperature
slope, a single maximum viscosity at a standard temperature adequately controls
this property.
Trace Properties. Instead of describing trace properties individually,
a better overview is obtained by grouping these properties into the following
performance categories.
Oxidative Stability. Jet fuel’s exposure to oxidative conditions over a
wide temperature and time range can result in a wide variety of problems (19).
Ambient temperature oxidation can form peroxides that degrade certain
elastomers, eg Buna N, which are used in aircraft fuel systems. Long-term
storage can cause soluble and insoluble oxidation products labeled ‘‘gums’’.
Vol. 2 AVIATION TURBINE FUELS 281
Peroxide formation is usually seen in fuels which have been produced in the
refinery by catalytic hydrogen processes (see Chap. 3) which remove certain
trace compounds containing sulfur or nitrogen which are natural anti-oxidants.
Anti-oxidancy is restored by the addition of oxidation inhibitor additives
(Section 3.2). Gum formation is rare but is monitored and controlled by the
Existent Gum test.
High-temperature oxidation is of major concern. Oxidation products–gums,
lacquers and insoluble deposits – can form in the hot areas of an aircraft’s fuel
system causing a number of problems (20). The heat exchanger of the fuel-cooled
oil-cooler can lose efficiency; flow control valves in the fuel control unit can stick
or jam, coke deposits can form in the burner feed arms which can, in the worst
case, cause plugging of the fuel injector nozzles. These deposits represent such a
small fraction of the fuel that the oxygen dissolved in the fuel is sufficient to
cause their formation. Although deoxygenation is effective on a laboratory
scale, in large scale practice it is impossible to reduce the oxygen content of
the fuel to a level low enough to prevent oxidation. The problem is attacked by
designing aircraft fuel systems to minimize hot spots; indeed, thermal manage-
ment is one of the most important criteria in modern engine and aircraft design.
In addition, the thermal stability of fuel supplies is controlled by a test in the fuel
specification (Table 2) using an instrument called the Jet Fuel Thermal Oxida-
tion Tester (JFTOT), which is a small-scale simulation of a heat exchanger in
which formation of lacquer and deposit are measured (ASTM D 3241) (3). Cer-
tain metals, notably copper, are oxidation catalysts and when dissolved in the
fuel at extremely low levels, eg, 20 mg/kg (ppb), can promote thermal instability
and deposit formation. Traces of copper can be picked up by the fuel from copper-
containing alloys sometimes encountered in the refinery or during transporta-
tion. The deleterious effects of this dissolved copper can be alleviated by dosing
the fuel with a metal deactivator additive (Section 3.2).
Corrosivity. Corrosion of fuel system metals can be a problem. Copper or
bronze in early fuel pump bearings and other components needed protection (1).
282 AVIATION TURBINE FUELS Vol. 2
opposite charge is imposed on that surface. One set of charges is swept along
with the fuel while the opposite charges, which accumulate along the wall sur-
face, usually leak to ground (earth). This charge separation results in a rise in
voltage in the moving fuel (26). High speed pumping of aviation fuel through
pipes and filters can generate very high charges and when the charged fuel
enters a tank, a substantial voltage difference may be produced between the sur-
face of the liquid and the tank walls and this may result in a static discharge (27).
If the vapor space above the liquid is in the flammable range, such a spark dis-
charge can cause a fire or explosion.
Charge build-up can be reduced by reducing the flow-rates of the fuel or
designing and constructing facilities — pipework, filters, valves, etc. – to mini-
mize charge generation and/or provide time for the charge to relax to ground
before entering a tank (28). Another approach is to add a static dissipater addi-
tive to the fuel (Section 3.2) to increase the electrical conductivity of the fuel to
facilitate charge relaxation.
2.3. Nonspecification Properties. Some aviation fuel properties are
not limited by specifications but are useful in equipment design or performance.
Absolute Vapor Pressure. Reid vapor pressure, specified for Jet B fuel,
is not an absolute vapor pressure because it is measured with a vapor–liquid
ratio of 10 : 1; it is determined only at 38.8 8C, and also it includes the partial
pressures of water vapor and air. With low-vapor-pressure products, such as
Jet A or A-1, absolute or true vapor pressure is used instead. Figure 2 shows typi-
cal true vapor pressure-temperature curves for different jet fuel grades (18).
Water Miscibility. Even the very low concentration of dissolved water
in aviation turbine fuels is hazardous if this water separates and freezes in a
critical aircraft location. Because dissolved water cannot be avoided in vented sys-
tems and is not removed by airport fuel handling equipment, aircraft fuel systems
either heat the fuel above 0 8C or depend on anti-icing additives (see Section 3.2).
Typical water saturation levels for commercial jet fuels are listed in Table 4 (18).
Lubricity. Various aircraft fuel system components – fuel pumps, servo
valves, fuel control units – rely on the fuel to lubricate their moving parts. The
effectiveness of a jet fuel as a lubricant in such equipment is referred to as its
‘‘lubricity’’. Differences in component design and materials result in varying
degrees of equipment sensitivity to fuel lubricity. Similarly, jet fuels vary in
their level of lubricity. The chemical and physical properties of jet fuel cause
it to be a relatively poor lubricating material under high temperature and load
conditions where lubrication may depend on adsorbed films of trace polar com-
pounds present in the fuel, so-called ‘‘boundary lubrication’’. Severe hydroproces-
sing (see Chap. 3) removes trace components, resulting in fuels which tend to
have lower lubricity than straight-run or wet-treated fuels. This lack of lubricity
can be remediated by the addition of corrosion inhibitor/lubricity improver addi-
tives (Section 3.2). Such additives are mandatory in most military jet fuel speci-
fications but they have not found widespread use in civil aviation. It is generally
acknowledged that the ultimate solution to the fuel lubricity ‘‘problem’’ is to
design fuel system equipment such that it is insensitive to fuel lubricity. Civil
fuel specifications do not contain a requirement for a minimum level of lubricity,
not least because there is no method of test for this property which is universally
accepted. One method, the Ball-on-Cylinder Lubricity Evaluator (BOCLE) ASTM
D 5001 (3), can be used for identifying possible problem fuels and is used by
equipment manufacturers to monitor special low-lubricity fluids used in develop-
ment testing of fuel system equipment, and by the USAF to monitor the lubricity
additive concentration in their fuel supplies.
Other Properties. Detailed data on other properties, including gas
solubility, flammability, surface tension, heat of vaporization, enthalpy, specific
heat, thermal conductivity, and dielectric constant, are contained in the CRC
Vol. 2 AVIATION TURBINE FUELS 285
Handbook of Aviation Fuel Properties (18). Some typical values for several prop-
erties are given in Table 4.
3. Production
3.1. Fuel. Since its earliest days virtually all aviation turbine fuel has
been refined from selected petroleum crudes by atmospheric distillation followed
by limited processing to reduce thiol or total sulfur. Crude selection is vital for
such simple processing because the only other variable available for controlling
bulk properties is adjusting the product’s boiling range. Desirable crudes are
low in sulfur (sweet crudes), low in n-hydrocarbons (nonwaxy crudes), low in aro-
matic content, and high in product boiling between 160 and 290 8C.
Adjusting either end of the boiling range has predictable results. Lowering
the initial boiling point improves combustion and low-temperature properties
but decreases handling safety (see Table 5). Raising the final boiling point
degrades combustion and low-temperature performance and has no effect on
handling safety. Because the crude also yields both lighter and heavier fractions,
the demand for these products affects distillation cut points and therefore the
volume of jet fuel available (29).
Total sulfur can be reduced by relatively low-pressure hydrotreating. Thiol
sulfur is reduced (sweetened) by low-pressure hydrotreating or wet chemical
treating, such as Merox, Linde, Doctor, or Bender treating. Flash point may be
adjusted so that the same product can be sold as either aviation turbine fuel or
commercial illuminating or heating kerosine.
Some heavier fractions are hydrocracked to make jet fuel, but the 160 –
290 8C fraction from catalytic cracking is not suitable for jet fuel unless cracking
is followed by high-pressure hydrotreating. A limited amount of kerosine is
hydrotreated to reduce aromatic content and increase smoke point by aromatic
saturation. In the United States the finished jet fuel is often filtered through
towers of activated Attapulgus clay to adsorb surfactants and to remove particu-
late fines.
3.2. Additives. All aviation fuel additives must be tested and approved.
The extent of testing is at the discretion of the aircraft and engine manufacturers
who issue the approval. Additive composition plays a major part in approval pro-
cedures, with ash-forming materials receiving the greatest scrutiny. Recom-
mended Practice D 4054 (ASTM) (3) covers the evaluation of compatibility
between additives and aircraft fuel system materials. Approved additives and
their permitted levels are normally listed in the fuel specifications. Unapproved
additives cannot legally be used in aircraft operation.
Oxidation Inhibitors. Oxidation inhibitors prevent peroxide formation
and improve storage stability in certain fuels (! Antioxidants). They are gener-
ally secondary and tertiary hindered phenols and are listed in the fuel specifica-
tion by chemical composition. They are effective primarily during storage, not
during high-temperature use. Antioxidants are mandatory for hydroprocessed
fuels in the major military fuel specifications and in civil Jet A-1 covered by
the DEF STAN 91-91 (DERD 2494) specification and IATA and Joint System
Checklist documents, but only optional in the ASTM D 1655 specification. Dosage
rate is 17 – 24 mg/L. Annual consumption of oxidation inhibitors in aviation tur-
bine fuel is difficult to estimate because the same additives are widely used in
other petroleum products.
Metal Deactivators. Metal deactivators are intended to passivate trace
metals by chelation (! Antioxidants). The only approved deactivator is N,N0 -dis-
alicylidene-1,2-diaminopropane [94-91-7]. It is permitted up to 5.7 mg/L, but is
rarely used.
Corrosion Inhibitors and Lubricity Improvers. Unlike the preceding
additives, corrosion inhibitor formulations are proprietary. They usually contain
dibasic acids that function as liquid phase rust inhibitors. The polar ends of the
molecule adsorb onto metal surfaces and in ground handling and distribution
systems, e.g., steel pipelines, they act to protect surfaces from corrosion. In air-
craft fuel systems, the tendency to adsorb on to metal surfaces is utilized for
lubricity improvement rather than corrosion inhibition. Because these additives
are surface-active (commonly referred to as ‘‘surfactant’’), they can display the
adverse side-effect of impairing the water separation properties of the fuel.
Furthermore, their surfactancy means that they can be lost from the fuel during
distribution (as, of course, they should be if they are to protect metal surfaces
from corrosion); consequently, if added for lubricity enhancement, these addi-
tives are often injected at the airport, just prior to aircraft fueling. Lists of
approved additives are published in the Qualified Products Lists (QPLs) asso-
ciated with the U.S. MIL-I-25017 and UK DEF STAN 68-251 specifications for
this type of additive.
Vol. 2 AVIATION TURBINE FUELS 287
JP-8, rather than developing a special fuel. It was calculated that an increase of
100 8F (56 8C) in the thermal stability level of the fuel would be required and this
led to the new fuel being referred to as ‘‘JP-8 þ 100’’. The extensive research pro-
gram to develop and test such additives was referred to as ‘‘HiTTS’’, standing for
High Temperature Thermal Stability (30). The main problems at high tempera-
ture are fouling of the injector feed arm and coking in the fuel supply rings to the
augmenter (afterburner or reheat) section. An additive package comprising a
detergent/dispersant, an antioxidant, and metal deactivator proved to be very
effective in preventing deposit formation in the hot areas of existing aircraft
fuel systems. Already the benefits in terms of reduced maintenance costs are
being appreciated. However, because the dispersant/detergent element of the
additive package is very surface active, it cannot be added to the fuel in storage
because it would encourage water and dirt pick-up and also disarm the filter/
water separators in the airport fueling system. Instead, the additive package
is injected into the fuel as the aircraft is being fueled.
5. Legal Aspects
All aircraft, including all flight components, are certified for flight by a govern-
ment agency which promulgates and enforces the certification rules. These rules
have the force of law. Aircraft and engines use fuel meeting certain specifications
during certification. These specifications become part of the aircraft’s operating
certificate. The fuel must therefore completely meet the certification specification
even though a deviation may not affect operation adversely. Normally aircraft
and engine manufacturers designate industry fuel specifications (e.g., ASTM
D 1655 (3) ), but manufacturers can certify to their own proprietary specification.
In such a case any points of difference from the industry norms must be specified.
Once certified, aircraft must be operated by the rules of the country in which
they are registered.
6. Environmental Aspects
7. Economic Aspects
Compared with gasoline and diesel fuel production, aviation turbine fuel repre-
sents a small proportion of the crude barrel. In 1995, worldwide production of
kerosine was 285 106 t of which ca.190 106 t was for jet fuel with the remain-
der for non-aviation uses (heating, lighting, cooking, etc.). This total kerosine
volume represents ca. 9 % of world oil consumption (37). World air travel is pre-
dicted to grow, on average, by 5.5 % per year over the next decade and 4.9 % over
the coming twenty years (38–40). Highest growth rates are in the ‘‘Pacific Rim’’
290 AVIATION TURBINE FUELS Vol. 2
countries. By 2006, air travel will rise by nearly 75 %; by 2016, it will reach two-
and-a-half times current levels. Fuel volumes will also grow significantly but at a
lower rate due to improvements in aircraft fuel efficiency. The United States is
the largest single consumer of jet fuel with about 50 % of the total world volume;
this share is expected to decrease in the future as the Far Eastern economies con-
tinue to grow.
A forecast of world aviation fuel demand in 106 t is given below (37):
8. Future Trends
aircraft and engine manufacturers for its use as jet fuel. However, like SMDS
kerosine, many of the coal-derived fractions contain no aromatics and will need
to be blended with other streams before being acceptable. Although this produc-
tion route has been confined to South Africa, at some time in the future it may be
applied in other countries with large coal reserves, e.g., China. Oxygenated
liquids like alcohols or esters have been examined as potential aviation fuels or
fuel constituents. The pure materials penalize aircraft operation by their much
lower specific energy content, whereas blends with petroleum fuels may suffer
from water extractability and system material compatibility difficulties (29).
Thus the widespread adoption of oxygenated compounds, particularly methanol,
as commercial aviation fuel appears unlikely.
It is often said that the long-term future may lie with cryogenic fuels such as
liquid methane or liquid hydrogen. Successful flights have been made using
methane in Russia, while the ‘‘Cryoplane’’ project (44) is aimed at proving the
feasibility of a liquid-hydrogen powered aircraft. However, a major barrier to
these cryogenic fuels is that they cannot be used in conventional aircraft, and
a new generation of aircraft specifically designed for them would need to be
developed. Furthermore, completely new production, storage, distribution, and
handling systems would have to be constructed all around the world to ensure
supply at all major airports. Proponents of cryogenic fuels tend to underestimate
the capital investment involved in establishing a supply infrastructure and it is
questionable whether these fuels will ever become commercially viable (45).
Methane may be more advantageously used as the starting material in the man-
ufacture of conventional liquid hydrocarbon fuels rather than as a liquid fuel
itself (46). Liquid hydrogen has excellent combustion qualities and produces no
CO2 and, at first sight, appears to be more environmentally friendly than conven-
tional kerosine-based jet fuel. However, unless it is produced via nuclear or
hydroelectric power, its production from petroleum resources would continue
to consume energy and produce CO2. Moreover, combustion of hydrogen in air-
craft engines would still not avoid the production of NOx and, furthermore, pro-
duces 2.5 times as much water vapor as kerosine, which will also contribute to
the greenhouse effect.
The massive investments required for new fuels together with the long lead
time in design and certification will encourage the use of fuels similar to those
currently used for the longest possible time.
REFERENCES
34. L. J. Archer, Aircraft Emissions and the Environment: COX, SOX, HOX and NOX,
Oxford Institute for Energy Studies, 1993.
35. Air Transport and the Environment, IATA Seminar, Washington, March 1993.
36. Air Transport and Environmental Protection, Airbus Industrie, March 1993.
37. Aviation Fuels Business Worldwide 1996, Petroleum Economist, London April 1996.
38. ‘‘1997 Current Market Outlook,’’ Boeing Commercial Airplane Group, March 1997.
39. Outlook for Commercial Aircraft 1995–2014, McDonnell-Douglas, August 1996.
40. Global Market Forecast 1995–2014, Airbus Industrie, March 1995.
41. J. P. Longwell, J. Grobman, J. Eng. Power 101 (1979) 161.
42. The Shell Middle Distillate Synthesis Process, Shell Selected Paper, November 1989.
43. P. J. A. Tijm, J. M. Marriott, The Markets for Shell Middle Distillate Synthesis
Products, Alternative Energy ’95, Vancouver, May 1995.
44. ‘‘The Light Stuff,’’ Chemistry in Britain 1997 29–31.
45. A. M. Momenthy, Aviation Fuel in a Changing World, Boeing, January 1994.
46. O. Hadaller, A. M. Momenthy, Fuel Reformulation and Jet Fuel, Boeing, November
1994.
GEOFF J. BISHOP
Shell Aviation Ltd., London, United Kingdom
HEATING OIL
1. Domestic Heating Oil
Domestic heating oil (domestic fuel oil, industrial gas oil, Heizöl EL) is an
extra-light liquid distillate fuel produced from crude oil in refineries (see ! Oil
Refining). In some countries the product can contain a certain percentage of fatty
acid methyl ester (FAME). This fuel is used for commercial and industrial pro-
cesses, and mainly for heating purposes in working areas and domestic accom-
modations. The fuel is burned in atomizing or vaporizing burners.
The first industrial applications of heating oil were reported at the begin-
ning of the 20th century. In the United States, industrial gas oil was being
used to produce hot water and steam for residential heating as early as the
1930s. The major expansion of domestic heating oil and oil firing in Europe
has occurred since the end of the 1950s. Today domestic heating oil and natural
gas (in some countries electricity) are the most common energy carriers in the
domestic heating market of many European countries, in Northern America,
and in Japan. Due to emission legislation and technical requirements different
products are available in different countries. Basically heating oils can be divided
into three classes:
1. Kerosene
2. Standard heating oil
3. Low-sulfur heating oil
The large volume and share of industrial gas oil in the total refinery stream
(20 to 30 %) makes this product economically significant.
1.1. Composition. Domestic heating oil is an extra-light liquid fuel
composed almost entirely of paraffinic, naphthenic, and aromatic hydrocarbons.
In addition to these hydrocarbons, it contains small quantities of sulfur com-
pounds and still smaller (ppm) amounts of nitrogen and oxygen compounds.
Since the beginning of the 21st century special heating oil qualities have been
developed, which are almost free of sulfur and nitrogen. The chemical character-
ization of the fuel is limited to group analysis by classes of compound. The pro-
duct structure depends on the origin of the crude oil and the processing
employed. The elemental composition, however, does not vary significantly.
The boiling range is between 160 and 390 8C (kerosene: 300 8C). A typical domes-
tic heating oil has the elemental composition given in Table 1.
1.2. Quality Aspects. Quality aspects are mostly described by physical
parameters or special test methods for mineral oil products. Basically for all
three different qualities (kerosene, standard heating oil, low-sulfur heating oil)
the requirements can be classified as follows.
Density. The density depends on temperature and for heating oil is spe-
cified at 15 8C. Density alone is of little significance as an indication of the burn-
ing characteristics of heating oil in modern atomizing type burners. However, in
vaporizing-type burners soot and CO emissions frequently increase when fuels
are used whose density and viscosity are close to the specified limits. That is
the reason why in some countries (UK and Japan) kerosene with its lower den-
sity and boiling range is used as heating fuel for vaporizing-type burners. Den-
sity limits vary in a wide range from country to country. When used in
conjunction with other properties the density is of value in mass–volume rela-
tionships and in calculating the specific energy (heating value, net calorific
value) of an oil.
Vol. 2 HEATING OIL 295
Net Calorific Value. The net or inferior calorific value Hi (MJ/kg; synon-
ymous with lower heating value, see ! Gas Production, Chap. 8.2.1.1) does not
vary significantly between different heating oils. Hi can be calculated from the
density at 15 8C (d15, kg/m3) and the sulfur content w(S) (mass %):
d15
Hi ¼ 52:92 11:93 0:29 wðSÞ
1000
Viscosity. Viscosity limits are specified to help maintain uniform fuel
flow in appliances with gravity flow, and to provide satisfactory atomization
and constant flow rate through the small nozzles of domestic burners. A viscosity
of 2.5 to 3 mm2/s is needed for good spray formation in conventional atomizing
burners. This can be achieved by preheating of the oil to a temperature of
60 8C maximum. At the same time, the oil throughput in a conventional nozzle
is reduced by preheating because of thinner film formation in the nozzle itself.
The effect of temperature on viscosity is shown in Figure 1.
Distillation. The distillation test shows the volatility of a fuel and the
ease with which it can be vaporized. It is of greater significance for oils that
are to be burnt in vaporizing type burners than for those burnt in the atomizing
type. For example, a low 10 % value assures easy starting of a vaporizing type
burner. Nevertheless, a homogeneous distillation curve is necessary to reduce
emissions of CO and soot. A typical distillation curve of a domestic heating oil
is depicted in Figure 2. Kerosene type fuels however vaporize at significantly
lower temperatures.
Flash Point. The flash point of a fuel is an indication of the maximum
temperature at which it can be stored and handled without serious fire hazard.
The minimum permissible flash point is usually regulated by law and can there-
fore differ from country to country.
Sulfur Content. The sulfur content is usually limited by national legis-
lations. The sulfur content in domestic heating oil has been reduced significantly
over the years. In Europe an EC directive allows sulfur content to be as high as
0.2 %. This value is reduced to 0.1 % starting from 2008. The amount of low-
temperature corrosion products formed in modern energy efficient boilers can
296 HEATING OIL Vol. 2
be influenced by the sulfur content of the oil. In condensing boilers which is the
most efficient burner technology but not very common for domestic heating oil in
contrast to natural gas equipment — the pH value of the condensate depends on
the sulfur content. To promote these energy saving appliances in many countries
a second quality with significantly lower sulfur content has been introduced to
the market and standardized.
Low-Temperature Behavior. In heating oil crystallization of high
molecular mass n-paraffins (waxes) can occur at low ambient temperature. The
wax crystals can block fuel pipes and especially the fuel filter. Three test meth-
ods are used to assess the low-temperature behavior.
Pour Point. The pour point gives the lowest temperature at which a fuel
oil can be stored and still is capable to flow under very weak forces. The determi-
nation of the pour point alone is not sufficient to assess the cold flow behavior in
practical application.
Cloud Point. The cloud point is the temperature at which heating oil
begins to become turbid at a defined cooling rate because of the formation of
small wax crystals.
Cold Filter Plugging Point (CFPP). The CFPP is the temperature at
which, under defined conditions, a heating oil sample can no longer pass a
defined filter within a specified time. The CFPP can be improved by additives
(see Section 1.3.2). Neither cloud point nor cold filter plugging point alone can
sufficiently describe the cold flow behavior in real installations but the combina-
tion of both does.
Carbon Residue. The carbon residue of a fuel is a measure of the carbo-
naceous material that is left after all volatile fuel components are vaporized in
the absence of air. It is a rough approximation of the tendency of a fuel to
form deposits in vaporizing burners where the fuel is vaporized in an air-defi-
cient atmosphere. To obtain measurable values of carbon residue in the lighter
distillate heating oils, first 90 % of the oil must be removed by distillation, then
the carbon residue concentrated in the remaining 10 % bottoms is determined.
Ash. The amount of ash is the quantity of noncombustible material in an
oil. The amount of ash in domestic heating oil is almost zero if heating oil directly
from the refinery is investigated. However, traces of metals can be found in heat-
ing oil, as steel equipment is mainly used for transportation and storage (1).
Vol. 2 HEATING OIL 297
Heat
Light
Oxygen
Catalytic effects of metals (copper)
The necessity and extent of blending domestic heating oil from different
components depends on the crude oil or the mixture of crude oils employed, on
the refinery facilities and the product mix that has to be achieved, and on the
specifications which have to be met.
Upgrading with Additives. Product quality and application requirements
can quite often only be met by employing special purpose additives, which are
added to the final product in the required concentrations. Some types of additive
are usually introduced into the product in the refinery. The use of additives for
quality improvement of heating oil is approved according to many heating oil
standards. These additives may be applied in suitable concentrations, provided
they are not known to exhibit any harmful side effects. Recommendations, which
should help to select suited additives, are published in (4).
Flow Improvers. Flow improver additives improve the cold flow proper-
ties of domestic heating oil. These additives are added to the product stream in
many refineries. Ash-free polymers are used as flow improvers. At a temperature
slightly higher than the cloud point, the additive forms crystallization nuclei,
to which the wax crystals become absorbed. Further growth of the wax crystals
is reduced by adsorption of the additive on the crystal surface. Thus, both the
crystal size and the tendency to agglomerate are lowered. The small crystals
Vol. 2 HEATING OIL 299
formed pass through the filters of the heating system. Therefore, the CFPP can
be lowered significantly by adding flow improvers. The dosage depends on the
base fuel and on the CFPP value which is to be achieved. Typical concentrations
are 50–500 ppm.
Flow improvers have to be added to the fuel at temperatures above the
cloud point; addition after crystal formation is not effective.
Metal Deactivators. In special cases, metal deactivators are also added
to industrial gas oil to attain long-term storage stability of more than three
years. The negative influence of copper (5,6) on storage stability can be reduced
by these additives.
Other types of additives are sometimes introduced into the customer’s sto-
rage tank but not into the fuel in the refinery.
Combustion Improvers. The purpose of these additives is to suppress
carbon formation during combustion. Combustion improvers are predominantly
organometallic compounds, ferrocenes have proved particularly effective in
domestic heating oil. Although burner technology has advanced considerably
with respect to near-stoichiometric combustion with low carbon formation, com-
bustion improvers can still effect and stabilize optimal combustion over long per-
iods. The application of combustion improver additives is not necessary in
modern blue flame burners with reduced NOx emissions.
Additives for Increased Storage Stability. The application of metal
deactivators and antioxidants can improve the storage stability of the fuel. The
importance of these types of additives is increasing because the average storage
time of heating oil in domestic storage tanks increases due to reduced oil con-
sumption (higher efficiency equipment, better insulation of houses) while the
size of the storage tanks has been kept constant.
Reodorants. Some additive packages contain reodorants to improve the
odor of heating oil. The unpleasant smell is caused particularly by certain sulfur-
containing and unsaturated components. If there is no leakage in the oil heating
system, the application of reodorants is not necessary.
Biocides. In some cases water can be found in the bottom of storage
tanks. Conditions at the interface between water and oil allow the growth of
microorganisms. The result is very often sediment formation. The addition of sui-
table biocides stops the growth of microorganisms but the best way is to remove
water regularly from the bottom of the tank.
1.4. Uses and Heating Efficiency. Domestic heating oil is mainly
burnt for heating purposes and for the heating of drinking water in burner/boiler
systems. The substantial rise in energy prices, the growing awareness of air pol-
lution and the international discussion on greenhouse gases have changed the
assessment of heating systems. It is possible to save a lot of energy in the domes-
tic heating sector if modern heating systems replace the old-fashioned systems.
A modern heating system has to fulfill the following demands:
calculated from the difference in temperature between flue gas and ambient air
(DT) and the concentration of CO2 or O2 in the flue gas:
0:5
flue gas loss ¼ DT þ 0:007
½CO2
or
0:68
flue gas loss ¼ DT þ 0:007
ð21 ½O2 Þ
These formulae cannot be used for condensing boiler technologies because the
latent heat of condensation of the water (produced in combustion) is utilized in
addition to the sensible heat. Therefore, the flue gas loss and the efficiency of a
boiler should be calculated with the gross or superior calorific value, Hs
(synonymous with higher heating value, see ! Gas Production, Chap. 8.2.1.1).
The flue gas loss can be taken from Figures 4 and 5.
Domestic heating oil is used in small power stations (with Diesel engines)
for simultaneous production of electricity and hot water or steam for heating
purposes (10). In research projects and pilot applications domestic heating oil has
been used to produce hydrogen and carbon monoxide as feedstock for high tem-
perature fuel cells (11).
1.5. Product Specifications. There is no international or European
standard for domestic heating oil. The minimum requirements of domestic heat-
ing oil in some national standards are given in Table 2 and 3. Some countries
have introduced standards which differ mainly in sulfur content. Other countries
have set different limits for cold flow properties in winter and summer. Generally
the lighter qualities (eg, kerosene type fuels, Grade 1 in USA) are specially suited
for vaporizing burners.
In 2003 a new European Standard (EN 14213) has been published. This
standard specifies requirements and test methods for marketed and delivered
fatty acid methyl esters (FAME) to be used either as a heating fuel at 100 % con-
centration, or as a blending component for the production of heating fuel. How-
ever, mixtures have not been standardized yet. An increasing demand for
bio-fuels as components of automotive fuels can be observed especially in Europe.
As in some countries the logistics of diesel and heating oil are not completely
separated, it has to be specified to what extent FAME is allowed to be a compo-
nent of heating oil.
1.6. Storage and Transportation. A particular characteristic of
domestic heating oil is the flash point, established at >55 8C. This value leads
to classification of the fuel in Hazard Class III A (GefStoffV). There are no special
safety restrictions concerning storage or transportation of the product.
During transportation and storage domestic heating oil has to be kept at
such temperatures that the sedimentation of wax crystals is avoided. Therefore
the cold flow behavior of the fuel has to be considered for storage and transporta-
tion. This can be a problem especially in underground storage tanks, when the
pipes to the burner are not sufficiently insulated. In Germany the storage tem-
perature must be above 0 8C. Lower temperatures are allowed if a special fuel
with satisfactory low temperature-properties is used.
1.7. Environmental Protection. The importance of the sulfur content
of domestic heating oil for the consumer lies in the emission of sulfur dioxide
and thus the problem of environmental pollution. In the EU, the sulfur content
of domestic heating oil is now limited to 0.2 mass % (Austria 0.1 mass %). The EU-
sulfur directives (1999/32/EC) requires to reduce that limit to 0.1 mass % by
2008. In some countries or regions there are special fuels available with sulfur
content <0.05 mass % or even <50 mg/kg (0.005 mass %, low-sulfur heating
oil). In general the average sulfur content has been decreasing continuously.
Further reduction of the sulfur content of the fuel and thus reduced
SO2 emissions when burning the fuel can be accomplished with desulfurization
processes, if these facilities are available in the refinery. However, removal of
sulfur requires hydrogen and is performed at pressures of 2.5 to 7 MPa with
the result of increased energy consumption and CO2 emissions. On the other
hand this is more than compensated by higher annual efficiencies of low tem-
perature or condensing boilers, because fouling of the heat exchanger surfaces
is reduced.
The nitrogen content of domestic heating oil is not specified in national
standards. It depends on the crude oil and the processing employed. Some
Table 2. Product Specifications of Domestic Heating Oil in Various EU Countries (Minimum Requirements)
Austria Belgium France Germany Great Britain Italy Sweden
ÖNORM C1109 NBN T NBN EN 590 M 15-008 DIN 51603-1 BS 2869-2 UNI 6579 SS 15 54 10
52-716
Issue 2003 2000 2005 1997 2003 1998 1998 1998
Remarks Low sulfur Standard Low sulfur Low sulfur Standard Class C1 Class D Summer/Winter
Remarks Diesel EN 590 Kerosene
Property Requirements Requirements Requirements Requirements Requirements Requirements Requirements Requirements Requirements
Density at max. 0.860 0.820– 0.820– 0.830 – max. 0.860 — 0.815– stated
158C. g/cm3 0.880 0.845 0.880 0.875
Gross calorific — min. 44.4 — — — — —
value, MJ/kg
Net calorific min. 42.0 — — min. 42.6 — — —
value, MJ/kg
Flash point > 55 min. 55 >55 min.55 > 55 min. 43.0 min. min. 55 min. 56
Pensky- 56
Martens, 8C
Viscosity 2.8 – 6.0 max .6.0 max 9.5 max. 6.00 — — —
kinematic at
20 8C, mm2/s
Viscosity — — 2.0 – 4.5 — — 1.50 – 5.50 2.0–7.4 2.0–4.0/1.5–3.0
kinematic at
303
40 8C, mm2/s
Viscosity — — — — — — —
kinematic at
38 8C, mm2/s
Distillation — — — — — max. 2 —
recovery at
1508C, vol%
Distillation — — — — min. 15, max. 60 — — —
recovery at
2008C, vol%
Distillation — — — — — — min. 10
recovery at
2308C, vol%
Distillation — max. 65 max. 65 < 65 < 65 max. 65 < 65 —
recovery at
2508C, vol%
Distillation — — — — — — —
recovery at
3008C, vol%
Distillation min. 90 min. 85 min. 85 min. 85 min. 85 min. 85 min. 85 —
recovery at
3508C, vol%
Table 2. (Continued)
Austria Belgium France Germany Great Britain Italy Sweden
ÖNORM C1109 NBN T 52-716 NBN EN 590 M 15-008 DIN 51603-1 BS 2869-2 UNI 6579 SS 15 54 10
Issue 2003 2000 2005 1997 2003 1998 1998 1998
Remarks Low sulfur Standard Low sulfur Low sulfur Standard Class C1 Class D Summer/Winter
Remarks Diesel EN 590 Kerosene
Property Requirements Requirements Requirements Requirements Requirements Requirements Requirements Requirements Requirements
Distillation — — min. 95 — — — — —
recovery at
3608C, vol%
Distillation — — — — — — min. 95
recovery at
3708C, vol%
Distillation
temperature, 10 %
recovered, 8C
Distillation
temperature,
90 %
recovered, 8C
Distillation
temperature,
304
95 %
recovered, 8C
Pour point, 8C — — max. 8 — — max. 6 —
Cloud point, 8C — — max. 2 max. 3 — — max. 0/22
Cold filter max. 8 max. 10 max. 0/15 — max. 12 max. 4 / 12 max. 0 max.10/32
plugging point (CP 3)/11 (CP 2)/
(CFPP), 8C 10 (CP 1)
Carbon residue max. 0.15 max. 0.10 max. 0.30 max. 0.3 max. 0.30 max. 0.2
(on 10%
residue),
%(m/m)
Sulfur content, — max. 0.2 max. 0.20 0.0050 – 0.20 max. 0.04 max. 0.20 max. 0.20 < 0.10
%(m/m)
Sulfur content, max. 50 50–1000 max. 50 max. 50
mg/kg
Water content, max. 200 max. 200 max. 200 max. 200 max. 200 max. 200 — max. 200
mg/kg
Particulate — — — — — — —
contamination,
mg/L
Total max. 24 max. 24 max. 24 — max. 24 max. 24 — max. 24
contamination,
mg/kg
Ash, %(m/m) max. 0.01 — max. 0.01 — max. 0.01 max. 0.01 max.0.01 max. 0.01
Neutralisation — — — — — —
number, mg
KOH/100 mL
Copper — max. 1 class 1 — — class 1 class 1 — max. 1
corrosion
Sediment, %(m/m) — — — — max. 0.01 — max. 0.05
Mechanical — — — — — —
contamination,
% (m/m)
Water and — — max. 0.10 — — max. 0.05 —
sediment, vol%
Color red — red — — — —
Color ASTM — — — — — —
D 1500
Oxidation — — max. 2.5 — — — — —
stability,
mg/100 mL
305
Cetane index — — — — — — —
Organic — — — — — — max. 6
chlorine, mg/kg
PCB, mg/kg — — — — — — max. 2
Nitrogen, mg/kg — — — — — — stated
* See
Table 3. Product Specifications of Domestic Heating Oil in Some Non-Eu Countries
Japan Norway Switzerland USA
Issue JIS Gasoil SN ASTM
No.3 181160-2 D396
Remarks 2002 1999
Remarks Kerosene Production CH- Euro- Grade 1 Grade 2
Standard Quality Quality
Property Requirements Requirements Requirements Requirements Requirements Requirements
Density at 158C. g/cm3 0.820–0.860 0.820–0.855 0.820–0.860 max. 0.850 max. 0.876
Gross calorific value, MJ/kg — min. 45 —
Net calorific value, MJ/kg — min. 42.6 min. 42.6
Flash point Pensky-Martens, 8C min. 40 min. 60 min. 55 min. 55 min. 38 min. 38
Viscosity kinematic at 20 8C, mm2/s — — —
Viscosity kinematic at 40 8C, mm2/s 2.0–4.0 2–4 max. 4 1.3–2.1 1.9–3.4
Viscosity kinematic at 38 8C, mm2/s — — —
Distillation recovery at 1508C, vol% — — —
306
Distillation recovery at 2008C, vol% — — —
Distillation recovery at 2308C, vol% — — —
Distillation recovery at 2508C, vol% max.65 — max. 65
Distillation recovery at 3008C, vol% — — —
Distillation recovery at 3508C, vol% min. 85 — min. 85
Distillation recovery at 3608C, vol% — min. 90 —
Distillation recovery at 3708C, vol% min. 95 — —
Distillation temperature, 10 % recovered, 8C max. 215
Distillation temperature, 90 % recovered, 8C max. 288 282–338
Distillation temperature, 95 % recovered, 8C max. 270
Pour point, 8C — — — max. 18 max. 6
Cloud point, 8C max. 0 max. 2 max. 3
Cold filter plugging point (CFPP), 8C max. 11 max. 10 12/11/10
Carbon residue (on 10% residue), %(m/m) max. 0.15 max. 0.05 max. 0.05
Sulfur content, %(m/m) max. max. 0.05 max. 0.2
0.0499
Sulfur content, mg/kg max. 80
Water content, mg/kg max. max. 150 max. 200
100
Particulate contamination, mg/L max.8.0 — —
Total contamination, mg/kg — max. 12 max. 24
Ash, %(m/m) max.0.01 max. 0.01 max. 0.01
Neutralisation number, mg KOH/100 mL max.0.5 — —
Copper corrosion 1 — —
Sediment, %(m/m) — — —
Mechanical contamination, % (m/m) — — —
Water and sediment, vol% — — —
Color — — —
Color ASTM D 1500 <=1.5 — —
Oxidation stability, mg/100 mL max.2.5 — —
Cetane index min. 45 — —
Organic chlorine, mg/kg — — —
PCB, mg/kg — — —
307
Nitrogen, mg/kg — ÖKO-oil max. 100
308 HEATING OIL Vol. 2
desulfurization processes may also reduce the nitrogen content of the fuel. Typi-
cally, the nitrogen content is between 50 and 200 mg/kg, occasionally it can be as
high as 400 mg/kg. For low-sulfur heating oil or kerosene the nitrogen content
often is below 50 mg/kg. The fuel nitrogen contributes to the emissions of nitro-
gen oxides (NOx) during combustion. 50 to 100 % of the fuel nitrogen are con-
verted to NOx depending on the type of burner (12).
Typical emission levels for oil heating (Table 4) are dependent on the tech-
nology employed (13). As could be expected, modern boiler and burner technology
results in lower emission values of CO, NOx, and particulate matter (PM10) than
older technologies.
The emissions of components that are traditionally not determined have
been investigated (14). The study encompasses emissions of N2O, methane, ben-
zene, toluene, xylene, and PCDD/PCDF (polychlorinated dibenzo-p-dioxin/diben-
zofuran). The emission levels of these components are very low. Oil-fired systems
do not contribute significantly to total emissions of these components.
The emissions of NOx and SO2 could be reduced by secondary treatment of
the flue gas. The technology for that flue gas treatment is available on the mar-
ket but is usually not used for domestic installations, because the emission limits
can be met by modern combustion systems alone.
Recently particulate matter emissions have become more important. It is
well accepted that with modern combustion technology in which the fuel is eva-
porated before being combusted the particulate matter emissions are similar to
those of gas appliances.
1.8. Economic Aspects. The consumption of heating oil in selected
countries is given in Table 5 (17). The sizes of the US market and the Japanese
market are similar to the market in Germany. Market data are available on
www.euro-fuel.net. Data for the US market can be taken from (15).
The consumption of heating oil has been decreasing in many countries. This
is caused by better insulation of houses and more efficient boiler systems and by
the fact that in some countries oil heating systems have been switched to other
systems. In many countries natural gas is the main competitor, in other coun-
tries electrical heat-pumps are gaining market share. In contrast, in some coun-
tries the number of oil installations has increased by replacing direct electrical
heating.
Vol. 2 HEATING OIL 309
Residual fuel oil (RFO) is a refinery product derived from petroleum crudes.
Since the introduction of oil conversion technologies in the mid-1960s most
RFO has been produced by blending vacuum distillation residues with residues
from thermal crackers (e.g., visbreakers). Material with a marketable viscosity
range results from further addition of catcracked components as well as gas oil
fractions of varying density derived from thermal processing. Heavy oils derived
from lignite and coal will not be discussed here because these are only of local
interest.
2.1. Composition and Properties. Residual fuel oil is an extremely
complex material as a consequence of its origin in the cracking process. In addi-
tion to paraffinic, naphthenic, and aromatic hydrocarbons (some with molecular
masses approaching 20 000), RFO contains nitrogen and oxygen compounds as
well as organometallic substances that contribute to its ash content. A typical
elemental analysis shows a distribution (in wt %) of 87.5 % C, 10.6 % H, 1.8 % S,
and 0.1 % ash (C : H ratio 8.25); these values may be assumed valid for perform-
ing combustion calculations.
RFOs are colloidal systems of highly dispersed macromolecular components
(asphaltenes) in solution equilibrium with a continuous oil phase (the maltenes).
Just like the maltenes, asphaltenes may be regarded as natural fuel components,
although with a much higher C : H ratio. Most of the metallic ash components are
present in oil-soluble form, either as porphyrins or as part of the asphaltene
phase. Metal atoms are in most cases bonded to the heteroelements sulfur,
oxygen, and nitrogen.
310 HEATING OIL Vol. 2
Table 6. Metal Content of Residual Fuel Oil (in mg Per Kilogram Atmo-
spheric Residue) as a Function of Source
Crude oil source Vanadium Sodium Nickel
Nigeria/Forcados 1 8 10
North Sea/Brent 6 1 3
Libya/Brega 6 2 10
Middle East/Oman 30 5 10
Middle East/Arab Heavy 75 15 27
Venezuela/Tia Juana Pesado 426 22 56
The choice of burner type depends on the design of the combustion facility.
Key factors include adequate residence time for combustion of the asphaltenes
and low specific heat load, together with near-stoichiometric combustion to mini-
mize NOx emission.
Some marine users of RFO continue to employ steam boiler–turbine com-
binations, although diesel engines now account for 96 % of the market world-
wide, and their use continues to increase. Diesel engines are clearly more
economical to operate than turbines, especially under partial load conditions.
With the aid of turbocharging and waste heat recovery, a maximum thermal effi-
ciency of ca. 50 % can now be achieved at a specific fuel consumption as low as
150 g/kWh. Today’s RFO can be utilized without difficulty in both 4-stroke trunk
piston engines (minimum output: 160 kW per cylinder at 750 min1) and
2-stroke cross head engines (maximum output: 3500 kW/cylinder at 80 min1).
Pretreatment of RFO for use as a motor fuel entails centrifugal separation (puri-
fier/clarifier) and fine filtering (10–20 mm). In addition, effective preheating is
necessary (to 160 8C with an injection viscosity of 9–12 mm2/s and an injection
pressure ahead of the jet of as much as 140 MPa). Taken together, these
measures assure optimal combustion as well as adequate protection against
abrasive wear.
2.4. Quality Specifications, Testing, and Analysis. For technologi-
cal and economic reasons reliable standards must be available for RFO classifi-
cation, and product specifications must be clear and acceptable to all parties
involved. To date, standards have been developed for residual fuels for industrial
uses only on a national basis. The most important distinctions between various
product grades relate to viscosity, density, flash point, and sulfur and ash con-
tent. Standards applicable within the United States are covered by ASTM-D
396–84; for the United Kingdom, BS 2869; for the Federal Republic of Germany,
DIN 51063 (part 3 and, in the future, part 5, which specifies a 1 % sulfur content);
for Switzerland, SN 181160/2–86; for Austria, öNORM C 1108–84; and for
Japan, JIS K K 2205.
Standards for RFO intended for marine use are given in ! Marine Fuels.
312 HEATING OIL Vol. 2
REFERENCES
GENERAL REFERENCES
CHRISTIAN KÜCHEN
Institut für Wirtschaftliche Oelheizung
Hamburg, Germany
KNUT SPITZMÜLLER
Hamburg, Germany
Vol. 2 NAPHTHENIC ACIDS 317
NAPHTHENIC ACIDS
The term naphthenic acid, as commonly used in the petroleum industry, refers
collectively to all of the carboxylic acids present in crude oil. Naphthenic acids
[1338-24-5] are classified as monobasic carboxylic acids of the general formula
RCOOH, where R represents the naphthene moiety consisting of cyclopentane
and cyclohexane derivatives. Naphthenic acids are composed predominantly
of alkyl-substituted cycloaliphatic carboxylic acids, with smaller amounts of
acyclic aliphatic (paraffinic or fatty) acids. Aromatic, olefinic, hydroxy, and diba-
sic acids are considered to be minor components. Commercial naphthenic acids
also contain varying amounts of unsaponifiable hydrocarbons, phenolic com-
pounds, sulfur compounds, and water. The complex mixture of acids is derived
from straight-run distillates of petroleum, mostly from kerosene and diesel frac-
tions (see PETROLEUM).
Naphthenic acids have been the topic of numerous studies extending over
many years. Originally recovered from the petroleum distillates to minimize
corrosion of refinery equipment, they have found wide use as articles of com-
merce in metal naphthenates and other derivatives. A comprehensive overview
of the uses of naphthenic acid and its derivatives can be found in References 1
and 2. A review of the extensive research on carboxylic acids in petroleum
conducted up to 1955 is available (3), as is a more recent review of purification,
identification, and uses of naphthenic acid (4).
1. Chemical Structure
The name naphthenic acid is derived from the early discovery of monobasic
carboxylic acids in petroleum, with these acids being based on a saturated single-
ring structure. The low molecular weight naphthenic acids contain alkylated
cyclopentane carboxylic acids, with smaller amounts of cyclohexane derivatives
occurring. The carboxyl group is usually attached to a side chain rather than
directly attached to the cycloalkane. The simplest naphthenic acid is cyclopen-
tane acetic acid [1123-00-8], (1, n ¼ 1).
CH2
H2C CH (CH2)n COOH
H2C CH2
(1)
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1401160802180905.a01
318 NAPHTHENIC ACIDS Vol. 2
COOH COOH
R COOH R COOH R R
COOH COOH
R COOH R COOH R R
COOH
R COOH R
Fig. 1. Typical naphthenic acid structures, where R ¼ alkyl. For the acyclic case z ¼ 0
and the structure is simply R COOH.
Naphthenic acids are viscous liquids, with phenolic and sulfur impurities present
that are largely responsible for their characteristic odor. Their colors range from
pale yellow to dark amber. An odor develops upon storage of the refined
acids. Naphthenic acids have wide boiling point ranges at high temperatures
(250–3508C). They are completely soluble in organic solvents and oils but are
insoluble ð <50 mg=LÞ in water. Commercial naphthenic acids are available
in various grades and are marketed by acid number, impurity level, and color.
Vol. 2 NAPHTHENIC ACIDS 319
3. Occurrence
Naphthenic acids are normal constituents of nearly all crude oils, but not all
crudes contain sufficient quantities of usable acids to make recovery an economic
process. Heavy crudes from geologically young formations have the highest acid
content, and paraffinic crudes usually have low acid content. The acid content
320 NAPHTHENIC ACIDS Vol. 2
of crude petroleum varies from 0–3%, with crudes from California, Venezuela,
Russia, and Romania having the highest content. Smaller amounts are found
in U.S. Gulf Coast crudes, whereas little or no naphthenic acids are found in
Pennsylvania, Iraq, or Saudi Arabia crudes. Typical concentrations are shown
in Table 2. Minor amounts of naphthenic acids are also found in bituminous
oil sands, but these are not economically recoverable. Identification of naphthe-
nic acids in water from oil-bearing strata is being examined as a potential
method of petroleum exploration (18).
Naphthenic acids occur in a wide boiling range of crude oil fractions, with
acid content increasing with boiling point to a maximum in the gas oil fraction
(ca 3258C). Jet fuel, kerosene, and diesel fractions are the source of most
commercial naphthenic acid. The acid number of the naphthenic acids
decreases as heavier petroleum fractions are isolated, ranging from 255 mg
KOH/g for acids recovered from kerosene and 170 from diesel, to 108 from
heavy fuel oil (19). The amount of unsaturation as indicated by iodine number
also increases in the high molecular weight acids recovered from heavier
distillation cuts.
4. Manufacture
5. Production
6. Economic Aspects
7. Analysis
Naphthenic acids are only slightly toxic to mammals but are toxic to fish,
bacteria, and wood-destroying insects. The lethal oral dose for humans is
approximately 1 L (44), and the oral LD50 is 3.0–5.2 g/kg in rats (45). The deaths
appeared to result from gastrointestinal disturbances. Naphthenic acid is non-
mutagenic by the Ames mutagenicity test (46), and is not listed as a carcinogen
by the International Agency for Research on Cancer, the National Toxicology
Program, or the U.S. Occupational Safety and Health Act.
9. Commercial Uses
More than two-thirds of the naphthenic acid produced is used to make metal
salts, with the largest volume being used for copper naphthenate, consumed
in the wood preservative industry (see WOOD). Metal salts used as paint driers
accounted for only 16% of the naphthenic acid market in 1993. This is a
dramatic contrast with 1977 usage, when 75% of the naphthenates went into
324 NAPHTHENIC ACIDS Vol. 2
the paint drier market. An overall view of the 1993 naphthenic acid market in
North America shows the following uses:
Use Percentage
wood 40.1
oil field 27.7
paint 15.8
tires 8.5
miscellaneous 7.9
Oil field uses are primarily imidazolines for surfactant and corrosion inhi-
bition (see PETROLEUM). Besides the lubrication market for metal salts, the mis-
cellaneous market is comprised of free acids used in concrete additives, motor oil
lubricants, and asphalt-paving applications (47) (see ASPHALT; LUBRICATION AND
LUBRICANTS). Naphthenic acid has also been studied in ore flotation for recovery
of rare-earth metals (48).
Naphthenic acid is ideal for synthesizing metal carboxylates that re-
quire a ligand with some oxidative stability, solubility in hydrocarbons
and oils, and insolubility in water. The general chemical structure of the
naphthenate salts is Mxþ(OOCR)x where x ¼ metal charge and molar stoichio-
metry of naphthenic acid to the metal M. Table 3 lists commercially available
metal naphthenates and their uses. Metal carboxylates are used in the oil-
based (alkyd) paint industry as catalysts to accelerate the drying of coatings;
hence the name driers was given to them. The metal naphthenates were origi-
nally used to replace the corresponding linoleates, resinates, and tallates.
These naphthenate salts have the advantages of high metal content, low viscos-
ity, better solubility in hydrocarbons and oils, and greater stability to oxidation.
BIBLIOGRAPHY
‘‘Naphthenic Acids’’ in ECT 1st ed., Vol. 9, pp. 241–246, by V. L. Shipp, Socony-Vacuum
Oil Co; in ECT 2nd ed., Vol. 13, pp. 727–734, by S. E. Jolly, Sun Oil Co.; in ECT 3rd ed.,
Vol. 15, pp. 749–753, by W. E. Sisco, W. E. Bastian, and E. G. Weierich, CPS
Chemical Co.
CITED PUBLICATIONS
1. E. S. Lower, Specialty Chem. 7, 76 (1987); 7, 282 (1987); 8, 174 (1988); 9, 135 (1989);
9, 267 (1989).
2. E. S. Lower, Pigment Resin Technol. 15(5), 4 (1986); 15(6), 4 (1986); 15(7), 11 (1986);
15(8), 6 (1986); 15(9), 8 (1986); 15(10), 9 (1986); 15(11), 7 (1986); 15(12), 6 (1986).
3. H. L. Lochte and E. R. Littmann, The Petroleum Acids and Bases, Chemical Publish-
ing Co. Inc., New York, 1955.
4. G. Narmetova, B. Khamidov, N. Ryabova, and E. Aripov, Purification, Identification,
and Use of Naphthenic Acid, Fan, Tashkent, former USSR, 1983.
5. N. Zelinsky and A. Chukasanov, Ber. 57, 42 (1924).
6. J. Von Braun, The Science of Petroleum, Vol. II, Oxford University Press, London,
1938, 1007–1015.
Vol. 2 NAPHTHENIC ACIDS 327
50. W. van Ooij, W. Weening, and P. Murray, Rubber Chem. Technol. 54, 227 (1981).
51. U.S. Pat. 3,897,583 (July 29, 1975), C. Bellamy (to Uniroyal SA).
52. J. Donalson, S. Clark, and S. Grimes, eds., Cobalt in Catalysts, Cobalt Development
Institute, London, 1990, Chapt. 8.
53. Eur. Pat. 512,778 (Nov. 11, 1992), J. Harrison, A. Bhattacharya, M. Patel, and D.
Rao (to Texaco Development Corp.).
54. P. LePerchec and co-workers, Prepr.-Amer. Chem. Soc., Div. Pet. Chem. 38, 401
(1993).
55. H. Chen and co-workers, AOSTRA J. Res. 5, 33 (1989).
56. U.S. Pat. 3,861,467 (Jan. 21, 1975), B. Harnsberger (to Texaco Inc.).
57. U.S. Pat. 3,997,469 (Dec. 14, 1976), C. Howle (to Nalco Chemical Co.).
58. U.S. Pat. 3,788,989 (Jan. 29, 1974), P. Adams and A. Petrocci (to Millmaster Onyx
Corp.).
59. Ger. Pat. 1,103,223 (Feb. 15, 1955), M. Lhorty (to Bataasfe Petroleum).
60. USSR Pat. 1,054,322 (Nov. 15, 1983), V. Shestakov and co-workers.
61. L. Schramm and co-workers, AOSTRA J. Res. 1, 5 (1984).
62. B. Zeinalov and co-workers, J. Polymer Sci. Part C 33, 353 (1971).
63. A. Ismailov and B. Salimova, Khim. Tekhnol. 9, 79 (1967).
64. Rom. Pat. 51,410 (Nov. 7, 1968), F. Popescu and co-workers (to Romania Institute for
Drilling and Extraction Research).
65. U.S. Pat. 2,512,255 (June 20, 1950), W. Gilbert, R. Kline, and C. Montgomery (to
Gulf Research and Development).
66. B. Kochetygov and P. Levenko, Kozh. Obuv. Prom. 10, 42 (1983).
67. J. Carmine and R. Ryntz, J. Coatings Technol. 66(836), 93–98 (1994).
GENERAL REFERENCES
JAMES A. BRIENT
PETER J. WESSNER
Merichem Company
MARY NN DOYLE
Shepherd Chemical Company
Vol. 2 FUELS, SYNTHETIC, LIQUID 329
The creation of liquids to be used as fuels from sources other than natural
crude petroleum (qv) broadly defines synthetic liquid fuels. Hence, fuel
liquids prepared from naturally occurring bitumen .deposits qualify as syn-
thetics, even though these sources are natural liquids. Synthetic liquid
fuels have characteristics approaching those of the liquid fuels in commerce,
specifically gasoline, kerosene, jet fuel, and fuel oil (see AVIATION TURBINE
FUELS; G ASOLINE AND OTHER MOTOR FUELS). For much of the twentieth century,
the synthetic fuels emphasis was on liquid products derived from coal upgrad-
ing or by extraction or hydrogenation of organic matter in coke liquids, coal
tars, tar sands, or bitumen deposits. More recently, however, much of the
direction involving synthetic fuels technology has changed. There are two
reasons.
The potential of natural gas, which typically has 85–95% methane, has
been recognized as a plentiful and clean alternative feedstock to crude oil (see
GAS, NATURAL). Estimates (1–3) place worldwide natural gas reserves at ca 1
1014 m3 ð3:5 1015 ft3 Þ corresponding to the energy equivalent of ca 11011 m3
(637 109bbl) of oil. As of this writing, the rate of discovery of proven natural
gas reserves is increasing faster than the rate of natural gas production. Many
of the large natural gas deposits are located in areas where abundant crude
oil resources lie such as in the Middle East and Russia. However, huge reserves
of natural gas are also found in many other regions of the world, providing oil-
deficient countries access to a plentiful energy source. The gas is frequently
located in remote areas far from centers of consumption, and pipeline costs can
account for as much as one-third of the total natural gas cost (1) (see PIPELINES).
Thus tremendous strategic and economic incentives exist for on-site gas conver-
sion to liquids.
In general, the proven technology to upgrade methane is via steam reform-
ing to produce synthesis gas, CO þ H2. Such a gas mixture is clean and when
converted to liquids produces fuels substantially free of heteroatoms such as
sulfur and nitrogen. Two commercial units utilizing the synthesis gas from
natural gas technology in combination with novel downstream conversion pro-
cesses have been commercialized.
The direct methane conversion technology, which has received the most
research attention, involves the oxidative coupling of methane to produce higher
hydrocarbons (qv) such as ethylene (qv). These olefinic products may be
upgraded to liquid fuels via catalytic oligomerization processes.
A second trend in synthetic fuels is increased attention to oxygenates as
alternative fuels (4) as a result of the growing environmental concern about
burning fossil-based fuels. The environmental impact of the oxygenates,
such as methanol (qv), ethanol (qv), and methyl tert-butyl ether (MTBE) (see
OCTANE IMPROVERS) is still under debate, but these alternative liquid fuels
are gaining new prominence. The U.S. Alternative Fuels Act of 1988, .and
the endorsement of oxygenate fuels that act contains, clearly underscore the
idea that economics is no longer the sole consideration with regard to alterna-
tive fuels production (5).
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.12091721080114.a01
330 FUELS, SYNTHETIC, LIQUID Vol. 2
Combustion
C þ O2 Ð CO2 H298k ¼ 394 kJ=molð94:2 kcal=molÞ ð1Þ
Gasification
C þ 1=2O2 Ð CO H298k ¼ 111kJ=molð26:5Þkcal=mol ð2Þ
C þ H2 OðgÞ Ð CO þ H2 H298k ¼ þ131kJ=molð31:3Þkcal=mol ð3Þ
C þ CO2 Ð 2CO H298k ¼ þ172kJ=molð41:1Þkcal=mol ð4Þ
and the water gas shift reaction (eq. 5), used to increase the H2/CO ratio of the
product synthesis gas.
Coal gasification technology dates to the early nineteenth century but has
been largely replaced by natural gas and oil. A more hydrogen-rich synthesis gas
is produced at a lower capital investment. Steam reforming of natural gas is
332 FUELS, SYNTHETIC, LIQUID Vol. 2
Methanol
synthesis
Fig. 1. Routes to liquid fuels from natural gas and coal via synthesis gas. F-T is the
Fischer-Tropsch process.
applied widely on an industrial scale (9,10) and in particular for the production of
hydrogen.
The conversion of coal and natural gas to liquid fuels via indirect technology
can be achieved by the routes shown in Figure 1. Two pathways from synthesis
gas can be taken. Both have been commercialized. One pathway involves cou-
pling with Fischer-Tropsch technology to produce fuel range hydrocarbons
directly or upon further processing. Using coal feedstock, this route has been
commercialized in South Africa since the 1950s and a process using natural
gas was commercialized in Malaysia by Shell Oil Co. in 1993. An alternative
route relies on the production of methanol from synthesis gas and subsequent
transformation of the methanol to fuels using zeolite catalyst technology intro-
duced by Mobil Oil Corp. This route was commercialized using indigenous natu-
ral gas in New Zealand in 1985.
1.1. Coal Upgrading via Fischer-Tropsch. The synthesis of methane
by the catalytic reduction of carbon monoxide and hydrogen over nickel .and
cobalt .catalysts at atmospheric pressure was reported in 1902 (11).
In the early 1920s Badische Anilin- und Soda-Fabrik announced the specific cat-
alytic conversion of carbon monoxide and hydrogen at 20–30 MPa (200–300 atm)
and 300–4008C to methanol (12,13), a process subsequently widely industria-
lized. At the same time Fischer and Tropsch announced the Synthine process
(14,15), in which an iron catalyst effects the reaction of carbon monoxide and
hydrogen to produce a mixture of alcohols, aldehydes, ketones, and fatty acids
at atmospheric pressure.
Vol. 2 FUELS, SYNTHETIC, LIQUID 333
(940 bbl) of liquid hydrocarbons per day. In Japan, two full-scale plants were also
operated under Ruhrchemie license. Combined capacity was ca 400 m3 (2500 bbl)
of liquids per day.
In the mid-1930s Universal Oil Products reported (33,34) that gasoline of
improved quality could be produced by cracking the high boiling fractions of
Fischer liquids, and a consortium, the Hydrocarbon Synthesis, Inc., entered into
an agreement with Ruhrchemie to license the Fischer synthesis outside Germany.
In 1955 the South African Coal, Oil, and Gas Corp. (Sasol) commercialized
the production of liquid fuels utilizing Fischer-Tropsch technology (35). This
Sasol One complex has evolved into the streaming of second-generation plants,
known as Sasol Two and Three. The Sasol One process, shown in Figure 2a (36),
combines fixed-bed Ruhrchemie-Lurgi Arge reactor units with fluidized-bed
Coal
Mine Power station Air separation
Steam Oxygen
NH3
H2 + CO
Sasol
Fluid
Fixed Fischer-Tropsch
bed
bed hydrocarbon
Synthol
synthesis processes
Fluid bed
Synthol
Tail gas
Product recovery Methane CH4 Product
to pipeline
reforming recovery
20 MJ/m3 STP
LPG
LPG
Gasoline
Ethylene
Diesel O2 Steam
Product refinery Gasoline
Fuel oils Product
refinery Diesel
Waxes
Fuel oils
Chemicals
Jet fuel
Chemicals
(a) (b)
Fig. 2. Process flow sheet of (a) Sasol One and (b) the Sasol synfuel process for Sasol Two
and Three (36). LPG is liquefied petroleum gas; other terms are defined in the text. To
convert J to cal, divide by 4.184.
Vol. 2 FUELS, SYNTHETIC, LIQUID 335
Synthol process technology (37). For Sasol One, 16,000 t/d of coal is crushed and
gasified with steam and oxygen. After a number of gas purification steps in which
by-products and gas impurities are removed, the pure gas is processed in both
fixed- and fluidized-bed units simultaneously. Table 1 gives product selectivity
comparisons of fixed-bed and Synthol operations. Conversion to hydrocarbons
is higher in the Synthol unit and the H2/CO ratio is also higher. Because
the fixed-bed Arge reactor favors the formation of straight-chain paraffins,
there is greater production of diesel and wax fractions than the Synthol unit.
The Arge reactor products have lower gasoline octane number but higher diesel
cetane number relative to Synthol. The high wax production using the Arge reac-
tor was disadvantageous at the time owing to market limitations of wax fuels.
Sasol One produced a vast array of chemical and fuel products, including
gasoline at 1:5 106 t=yr.
The 1973 oil crisis resulted in the Sasol Two unit which started up in early
1980 followed by the nearly identical Sasol Three unit two years later. Figure 2b
gives the schematic flow diagrams for the Sasol Two and Three processes. Sasol
Two uses 36 Lurgi gasifiers in parallel to process ca 31,000 t/d of sized coal. By-
product effluents and gas impurities are removed in Rectisol (sulfur compounds
and CO2 removal), Phenosolvan .(oxygen compounds and ammonia removal),
and tar separation units. Synthol fluid-bed units were employed because of the
product distribution and ease of design scale-up. Approximately 80,000 t/d of coal
are needed for the two plants. Composition and manufacturing information for
Sasol Fischer-Tropsch catalysts are trade secrets, but the catalyst is widely
accepted as being an alkali-metal promoted iron-based material.
More recently, Sasol commercialized a new type of fluidized-bed reactor and
was also operating a higher pressure commercial fixed-bed reactor (38). In 1989,
a commercial scale fixed fluid-bed reactor was commissioned having a capacity
similar to existing commercial reactors at Sasol One (39). This effort is aimed
at expanded production of higher value chemicals, in particular waxes (qv) and
linear olefins.
Properties. Fischer-Tropsch liquid obtained using cobalt catalysts is
roughly equivalent to a very paraffinic natural petroleum oil but is not so com-
plex a mixture. Straight-chain, saturated aliphatic molecules predominate but
336 FUELS, SYNTHETIC, LIQUID Vol. 2
Fuel gas
Including LPG
Syngas
HPS Flash
Tops / naphtha
Distil- Kerosene
H2 HPC
lation
Gas oil
Fig. 3. The Shell middle distillate synthesis (SMDS) process. HPS ¼ heavy paraffin
synthesis. HPC ¼ heavy paraffin conversion.
Vol. 2 FUELS, SYNTHETIC, LIQUID 337
10 10
Product composition, wt %
Product composition, wt %
8 8
6 6
4 4
2 2
0 0
0 10 20 30 40 0 10 20 30 40
Number of carbons Number of carbons
(a) (b)
Fig. 4. Product compositions as a function of carbon number for the Shell middle distil-
late synthesis process: (a) the Fischer-Tropsch product following HPS, and (b) the final
hydrocracking product following HPC. See text (45).
n=2 ½2 CH3 OH !
CH3 OCH3 þ H2 O
!Cn H2n !n½CH2
ð12Þ
70
60
Product distribution, wt %
50
40
30
20
10
0
10 –4 10 –3 10 –2 10 –1 1 10
Space time (LHSV)–1, h
P
r
o Methanol Heavy
Natural gas BFW MTG MTG HGT
c synthesis gasoline Blending
e 155,000 m3/h treatment 4,400 t/d reactors distillation units
two trains storage
s
340
s
Light Gasoline
gasoline
Light gasoline storage
E
f
f
l Sanitary Effluent
u sewage treatment
e
n
t
Discharge
144 m3/h
Fig. 6. Simplified block flow diagram for the New Zealand gas-to-gasoline (GTG) plant (61). To convert m3/h to gal/min, multiply by 4.40.
HGT¼heavy gasoline treatment facility; MTG ¼ methanol-to-gasoline; BFW ¼ boiler feed water.
Vol. 2 FUELS, SYNTHETIC, LIQUID 341
Crude
CH3OH
To duplicate
reactor/effluent
exchanger trains
DME
reactor
Recycle
compressor
342
Hot excess ZSM-5 reactors
effluent four on line
from ZSM-5 one in regeneration
Purge gas to reactors
fractionation
Effluent/recycle
exchanger
Cooled effluent
from ZMS-5 reactor
Gasoline to
fractionation
Cooler
Water
Fig. 7. A methanol-to-gasoline (MTG) fixed-bed process flow diagram. DME ¼ dimethyl ether.
Vol. 2 FUELS, SYNTHETIC, LIQUID 343
BPR
Vent
xxxxx
Methanol
Condenser
Two-stage
scrubber
Water
Regenerator
Flare
Reactor
preparation
Feed
Air
+ Filter
Cooler
N2
C4 gas
N2 TC
Light gas
compressor
N2
Debutanizer
Vaporizer
Separator
Superheater
N2
Gasoline
Water
Fig. 8. Fluid-bed MTG demonstration plant schematic diagram. BPR ¼ Back pressure
regulator; TC ¼ temperature controller.
MTO process was first scaled up in a 0.64 m3/d (4 bbl/d) pilot plant and a success-
ful 15.9 m3/d (100 bbl/d) demonstration plant was operated in Germany with U.S.
and German government support.
The MOGD process oligomerizes light olefins to gasoline and distillate
products over ZSM-5 zeolite catalyst. Gasoline and distillate selectivity is >95%
of the feed olefins and gasoline/distillate product ratios can vary, depending on
Methanol
MTO
MOGD
Hydrogen
Distillate
hydrotreater Fractionation
Fuel gas
process conditions, from 0.2 to >100. High octane MTO gasoline is separated
before the MOGD section and blended with MOGD gasoline. Some MOGD gasoline
may be recycled. The distillate product requires hydrofinishing. Generally, the
process scheme uses four fixed-bed reactors, three on-line and one in regeneration.
A large-scale MOGD refinery test run was conducted by Mobil in 1981.
Properties. The gasoline product from the integrated MTO/MOGD pro-
cess is predominately olefinic and aromatic. The gasoline quality (ca 89 octane)
is comparable to FCC gasoline. Typical distillate product properties are given in
Table 5. After hydrofinishing, the distillate product is mostly isoparaffinic and
has high cetane index, low pour point, and negligible sulfur content. MOGD
diesel fuel has somewhat lower density than typical conventional fuels (0.8 vs
0.86). Low aromatics levels contribute to a stable jet fuel with very little smoke
emission during combustion. MOGD diesel and jet fuels meet or exceed all
conventional specifications.
The capital costs associated with indirect natural gas upgrading technology
are high, thus research and development has focused on direct conversion of
natural gas to liquid fuels. Direct conversion is defined as upgrading methane
to the desired liquid fuels products while bypassing the synthesis gas step, ie,
direct transformation to oxygenates or higher hydrocarbons. Direct upgrading
routes which have been extensively studied include direct partial oxidation to
oxygenates, oxidative coupling to higher hydrocarbons, and pyrolysis to higher
hydrocarbons. Owing to the inert nature of methane, the technology is limited
by the yields of desired products which in turn affects the process economics.
Only one direct oxidative methane conversion process has been commer-
cialized. A plant at Copsa Mica (Romania) in the 1940s (70) produced formal-
dehyde directly from methane and air by partial oxidation. This plant is
no longer in operation. Plants to produce acetylene .from methane by high tem-
perature pyrolysis routes have been commercialized.
Generally, the most developed processes involve oxidative coupling of
methane to higher hydrocarbons. Oxidative coupling converts methane to ethane
and ethylene by
The process can be operated in two modes: co-fed and redox. The co-fed
mode employs addition of O2 to the methane/natural gas feed and subse-
quent conversion over a metal oxide catalyst. The redox mode requires the
oxidant to be from the lattice oxygen of a reducible metal oxide in the reac-
tor bed. After methane oxidation has consumed nearly all the lattice oxy-
gen, the reduced metal oxide is reoxidized using an air stream. Both
methods have processing advantages and disadvantages. In all cases, how-
ever, the process is run to maximize production of the more desired ethylene
product.
Direct conversion of natural gas to liquids has been actively researched.
Process economics are highly variable and it is unclear whether direct natural
gas conversion technologies are competitive with the established indirect pro-
cesses. Some emerging technologies in this area are presented herein.
2.1. ARCO Gas-to-Gasoline Process. A two-step process using oxida-
tive coupling to upgrade natural gas to liquid fuels has been proposed by
ARCO (Atlantic Richfield Co.) (71,72). A simplified process scheme is given in
Figure 10. Methane is passed through a redox-mode oxidative coupling reactor
which generates C2 þ hydrocarbons such as ethylene. The olefinic products are
346 FUELS, SYNTHETIC, LIQUID Vol. 2
CO2 to atmosphere
Propane
Untreated gas Feed gas Butane
fractionation Condensate
Recycle gas
Air
Redox Water
CO2 to atmosphere
Olefin Product
conversion fractionation
High octane gasoline
Fig. 10. Simplified flow diagram depicting the ARCO gas-to-gasoline process for a con-
ceptual gasoline production plant (72).
3. Oxygenate Fuels
Alcohols and ethers, especially methanol, ethanol, and methyl tert-butyl ether
[1634-04-4] (MTBE), have been widely used separately or in blends with gaso-
lines (reformulated gasoline) and other hydrocarbons to fuel internal combustion
CO2 H2O
Gas
treatment
Heat Compression
recovery
Alkane
C2+ Alkanes C 2H6 Unsaturates
pyrolysis Hydrocarbon Olefin
recovery oligomerization
347
Oxidative
CO2, H2O coupling Diesel
CH4/CO/H2
Sulfur compounds Gasoline
800–850 C
O2
Air
separation
Fig. 11. The OXCO process for natural gas conversion (74).
348 FUELS, SYNTHETIC, LIQUID Vol. 2
engines. Fuel properties of key oxygenates are presented in Table 6 (5). These
compounds, as a class, may be considered to be partially oxidized, ie, each has
a mole of oxidized hydrogen. They differ from the hydrocarbons that make up
gasoline principally in lower heating values and in higher vaporization heat
requirements. This constitutes a serious disadvantage to the substitution
of oxygenates, especially lower alcohols, for motor gasoline. For example,
the heating value of methanol is about half that of gasoline on an equivalent
volume basis. Other properties which greatly influence the potential of oxyge-
nates as fuels include octane performance, solubility in gasoline, effect on gaso-
line vapor pressure, sensitivity to water, and evaporative/exhaust emissions.
Oxygenate fuels tests are often debated because the tests employed were devel-
oped for conventional gasolines.
The addition of small percentages of oxygenates to gasoline can produce
large gains in octane. Thus, as blending components in gasoline, oxygenates
improve octane quality. As neat fuels for spark-ignition engines, octane values
for oxygenates are not useful in determining knock-limited compression ratios
for vehicles because of the lean carburetor settings relative to gasoline. Neither
do these values represent the octane performance of oxygenates when blended
with gasoline.
In part because neat alcohols are insufficiently volatile to enable a cold
engine to start, even at moderate temperatures, the use of neat alcohols for auto-
motive motor fuel is problematic. Manufacturers have, however, reported that
alcohol-powered cars, after being started and warmed up, can have the same
or better driveability as gasoline cars (75–77). As for gasoline vehicles, port
fuel injector fouling has occurred in some methanol vehicles and has affected dri-
veability and emissions. Other problems related to high alcohol content gasoline
in conventional engines include vapor lock and corrosion. Flexible-fuel vehicles
(FFV), which can operate on either neat methanol or gasoline, or mixtures
thereof, are being evaluated.
Vol. 2 FUELS, SYNTHETIC, LIQUID 349
Direct liquefaction, the production of liquids from feed coal in a single processing
scheme without a synthesis gas intermediate step, includes two routes for the
Vol. 2 FUELS, SYNTHETIC, LIQUID 353
compared with typical petroleum fractions, they are high in oxygen and nitrogen
levels. No residual oil products (bp > 540 C) are formed.
Solvent-Refined Coal Process. In the 1920s the anthracene oil fraction
recovered from pyrolysis, or coking, of coal was utilized to extract 35–40% of
bituminous coals at low pressures for the purpose of manufacturing low cost
newspaper inks (113). Tetralin .was found to have higher solvent power for
coals, and the I. G. Farben Pott-Broche process (114) was developed, wherein a
mixture of cresol and tetralin was used to dissolve ca 75% of brown coals at 13.8
MPa (2000 psi) and 4278C. The extract was filtered, and the filtrate vacuum dis-
tilled. The overhead was distilled a second time at atmospheric pressure to sepa-
rate solvent, which was recycled to extraction, and a heavier liquid, which was
sent to hydrogenation. The bottoms product from vacuum distillation, or solvent-
extracted coal, was carbonized to produce electrode carbon. Filter cake from the
filters was coked in rotary kilns for tar and oil recovery. A variety of liquid pro-
ducts were obtained from the solvent extraction-hydrogenation system (113). A
similar process was employed in Japan during World War II to produce electrode
coke, asphalt (qv), and carbonized fuel briquettes (115).
In the United States there was little interest in solvent processing of coals.
A method to reduce the sulfur content of coal extracts by heating with sodium
hydroxide and zinc oxide was, however, patented in 1940 (116). In the 1960s
the technical feasibility of a coal deashing process was studied (117), and a
pilot plant able to process ca 45 t/d was completed in late 1974 (118).
A flow diagram of the solvent-refined coal or SRC process is shown in
Figure 12. Coal is pulverized and mixed with a solvent to form a slurry contain-
ing 25–35 wt% coal. The slurry is pressurized to ca 7 MPa (1000 psig), mixed
with hydrogen, and heated to ca 4258C. The solution reactions are completed
in ca 20 min and the reaction product flashed to separate gases. The liquid is fil-
tered to remove the mineral residue (ash and undissolved coal) and fractionated
to recover the solvent, which is recycled.
The liquid remaining after the solvent has been recovered is a heavy resi-
dual fuel called solvent-refined coal, containing less than 0.8 wt% sulfur and
0.1 wt% ash. It melts at ca 1778C and has a heating value of ca 37 MJ/kg
Sulfur
Coal Dissolver
Recycle
Slurry hydrogen
preheater
Raw gas
Slurry
mix tank
Coal slurry
Pulverizer
Absorber Sulfur
plant
356
Hydrogen
Solvent
recovery
unit
Product and solvent
Filter
Shipment
Solvent–refined coal
(16,000 Btu/lb), regardless of the quality of the coal feedstock. The activity of the
solvent is apparently more important than the action of gaseous hydrogen in this
type of uncatalyzed hydrogenation. Research has been directed to the use of pet-
roleum-derived aromatic oils as start-up solvents (118).
In the early 1970s production of low sulfur, ashless (solid) boiler fuel was
the preferred commercial application (119). This basic process (SRC-I) yielded
small amounts of liquid oil products with additional processing. Liquid
output was significantly increased by the coal-oil-gas (COG) refinery concept
(120–122) which incorporated high degrees of hydroconversion and hydrotreat-
ing. A SCR-II process has been developed, in which hydrocracking occurs in the
solution (hydrogenation) vessels (123). A low viscosity fuel oil is the primary dis-
tillate product in this case, although naphtha and LPG are also recovered.
Two pilot plants have been built and operated to demonstrate the feasibility
of the SRC process. These included a 6 t/d plant at Wilsonville, Alabama (vide
infra) and a 50 t/d plant at Ft. Lewis, Washington which was operated from
1974 to 1981.
In an effort to obtain higher value products from SRC processes, a hydro-
cracking step was added to convert resid to distillate liquids. The addition
of a hydrocracker to the SRC-I process was called nonintegrated two-stage
liquefaction (NTSL). The NTSL process was essentially two separate processes
in series: coal liquefaction and resid upgrading. NTSL processes were inefficient
owing to the inherent limitations of the SRC-I process and the high hydrocracker
severities required.
Properties. The properties of the liquid fuel oil produced by the SRC-II
process are influenced by the particular processing configuration. However, in
general, it is an oil boiling between 177 and 4878C, having a specific gravity of
0.99–1.00, and a viscosity at 388C of 40 SUs (123). Pipeline gas, propane and
butane (LPG), and naphtha .are also recovered from an SRC-II complex.
Exxon Donor Solvent-Coal Liquefaction Process. The EDS process
.from Exxon is a hydrogenation process using a donor solvent for the direct con-
version of a broad range of coals to liquid hydrocarbons (124). In the process
sequence, shown in Figure 13, the feed coal is crushed, dried, and slurried
with hydrogenated recycle solvent (the donor solvent) and fed to the reactor
with hydrogen. The reactor is an upward plug-flow design operating at
430–4808C and at ca 14 MPa (2000 psi) total pressure.
The reactor effluent is separated by conventional distillation into recycle
solvent, light gases, C4 to 5378C bp distillate, and a heavy vacuum bottoms
stream containing unconverted coal and ash. The recycle solvent is hydrogenated
in a separate reactor and sent back to the liquefaction reactor.
The heavy vacuum bottoms stream is fed to a Flexicoking unit. This is a
commercial (125, 126) petroleum process that employs circulating fluidized
beds at low (0.3 MPa (50 psi)) pressures and intermediate temperatures, ie,
480–6508C in the coker and 815–9808C in the gasifier, to produce high yields
of liquids or gases from organic material present in the feed. Residual carbon
is rejected with the ash from the gasifier fluidized bed. The total liquid product
is a blend of streams from liquefaction and the Flexicoker.
The EDS process was developed starting from 1976 in a 10-year joint under-
taking between DOE and private industry (127). Under the direction of Exxon
358 FUELS, SYNTHETIC, LIQUID Vol. 2
Catalytic
Coal H2 H2O
Liquid
H2 Liquefaction Distillation
products
Heavy
bottoms
slurry
H2O
Flexicoking
Air Fuel gas
Ash
residue
Co. USA, a 250 t/d pilot plant was operated at Baytown, Texas. Operation of this
unit began in 1980 and was completed by late 1982.
Properties. Pilot-unit data indicate the EDS process may accommodate
a wide variety of coal types. Overall process yields from bituminous, subbitumi-
nous, and lignite coals, which include liquids from both liquefaction and
Flexicoking, are shown in Figure 14. The liquids produced have higher nitrogen
contents than are found in similar petroleum fractions. Sulfur contents reflect
the sulfur levels of the starting coals: ca 4.0 wt% sulfur in the dry bituminous
coal; 0.5 wt% in the subbituminous; and 1.2 wt% sulfur in the dry lignite.
Table 8 shows that the naphthas produced by the EDS process have
higher concentrations of cycloparaffins and phenols than do petroleum-derived
naphthas, whereas the normal paraffins are present in much lower concentra-
tions. The sulfur and nitrogen concentrations in coal naphthas are high com-
pared to those in petroleum naphthas.
Gas oil fractions (204–5658C) from coal liquefaction show even greater
differences in composition compared to petroleum-derived counterparts than
do the naphtha fractions (128). The coal-gas oils consist mostly of aromatics
(60%), polar heteroaromatics (25%), asphaltenes (8–15%), and saturated
compounds (< 10%). Petroleum-gas oils, on the other hand, contain more than
50% saturated compounds, less than 5% polar heteroaromatics, and no asphal-
tenes. Furthermore, the aromatics of petroleum-gas oils have longer side chains.
Coal Liquefaction at Wilsonville. Starting in 1974 the Advanced Coal
Liquefaction R&D Facility at Wilsonville, Alabama operated a 6 t/d pilot plant
and studied various coal liquefaction processing schemes. The facility, cospon-
sored by the DOE, the Electric Power Research Institute (EPRI) and Amoco
Oil Co, was shut down in early 1992.
Vol. 2 FUELS, SYNTHETIC, LIQUID 359
50
40
Liquid yields, wt % dry coal
30
20
A B C D E
10
0
Bituminous Subbituminous Lignite
Coal type
Fig. 14. Preferred liquefaction-coking liquid yields in the EDS process for various coals
where represents Flexicoking liquids and &, liquefaction liquids (124). A, Ireland (West
Virginia); B, Monterey (Illinois); C, Burning Star (Illinois); D, Wyodak (Wyoming); and E,
Big Brown (Texas).
Initial operation at the Wilsonville pilot plant was in SRC-I mode and later
evolved into a two-stage process (129) by operation in NTSL mode. NTSL limita-
tions described previously combined with high hydrogen consumptions resulted
in subsequent focus on a staged integrated approach, which was to be the basis
for all further studies at Wilsonville.
The integrated two-stage process (ITSL) combined short contact time lique-
faction in one reactor with ebullated-bed hydrocracking in a second stage (130).
The short contact time conditions permitted better hydrogen transfer from the
solvent rather than from the gas phase. The hydrocracking step operated at
lower severity resulting in lowered gas make and improved hydrogen efficiency.
Recycle solvent was generated from the hydrocracked distillates and coupled the
two reaction stages. Results of ITSL processing of Illinois No. 6 coal at Wilson-
ville are given in Table 9. Distillate yields and coal throughput for ITSL were
higher than those obtained by NTSL.
Further developments of the ITSL process resulted in incremental gains in
distillate yields (131). Reconfigured integrated two-stage liquefaction (RITSL)
involved placing the solvent deasher after the hydrocracker thus producing
a recycle solvent consisting of deashed resid and distillate. This resulted in
reduction of feed to the deasher and reduced organic rejection. Close coupled
Table 8. Composition of Naphthasa from Various Sourcesb
EDS coal liquefaction Petroleum naphthas
Component, wt% Illinois coal Wyodak coal Cycloparaffinicc Paraffinicd
saturated compounds 69.9 60.8 77.9 81.3
paraffins 13.4 19.4 36.0 58.2
cycloparaffins 56.5 41.4 41.9 23.1
olefinse 0.5 3.7
aromatics 17.0 28.6 22.1 18.7
benzenes 11.7 25.1 20.8 17.2
indans, tetralins 5.1 3.5 1.0 1.2
indenes 0.2 0.0
naphthalenes 0.01 0.0 0.3 0.3
phenols 12.6 6.9 traces traces
Total 100.0 100.0 100.0 100.0
sulfur 0.57 0.10 0.035 0.049
nitrogen 0.15 0.18 0.0001 0.0001
oxygen 1.60 2.49
a
C5 to 2048C bp.
b
Ref. 128.
c
Prudhoe Bay.
d
Arab Light.
e
Values are approximate.
Table 9. Wilsonville Plant Operating Conditions and Yields for ITSL and CTSL Modesa
Mode of operationb
Parameter ITSL CTSL
Operating conditions
run number 7242BC; 243JK/244B 253A
catalyst Shell 32M Shell 317
First stagec
average reactor temperature, 8C 460; 432 432
space velocity 690; 450d 4.8e
f
pressure, MPa 17; 10–17 17.9
Second stage
average reactor temperature, 8C 382 404
space velocity, feed/catalyste 1.0 4.3
catalyst age, resid/catalyst 278–441; 380–850 100–250
Yieldsg
C1 –C3 gas 4; 6 6
C4 þ distillate 54; 59 70
resid 8; 6 1
hydrogen consumption 4.9; 5.1 6.8
Other
hydrogen efficiency, C4 þ distillate=H2 consumed 11; 11.5 10.3
distillate selectivity, C1 –C3/C4 þ distillate 0.07; 0.10 0.08
energy content of feed coal reject 24; 20–23 20
to ash concentrate, %
a
Feed is Illinois No. 6 coal.
b
CTSL ¼ catalytic two-stage liquefaction; ITSL ¼ integrated two-stage liquefaction.
c
First stage is thermal for ITSL.
d
Value given is coal space velocity at temp >371 C in kg/m3.
e
Value given is in h1 .
f
To convert MPa to psia, multiply by 145.
g
Wt % on a moisture- and ash-free (MAF) coal basis.
360
Vol. 2 FUELS, SYNTHETIC, LIQUID 361
integrated two-stage liquefaction (CC-ITSL) linked the two reactors and removed
several operations between the two stages. A deleterious effect of these two pro-
cessing modes was increased hydrogen consumption over ITSL.
From 1985 to 1992, development activity at Wilsonville was on catalytic
two-stage liquefaction (CTSL). CTSL, initiated by HRI (132), consists of catalytic
processing in two ebullated-bed reactors which lower reaction temperatures and
increase distillate yields, up to 78% yield. CTSL results from Wilsonville for
Illinois No. 6 coal are also given in Table 9. Distillate yields were shown to be
significantly higher for CTSL over ITSL; however, hydrogen consumption was
somewhat increased.
Properties. CTSL distillates have qualities comparable to or better than
No. 2 fuel oil and have good hydrogen content and low heteroatom contents. Dis-
tillates having a higher boiling point distribution from Wilsonville CTSL opera-
tion (131) showed 26.88 API gravity with heteroatom levels of 0.11 wt% sulfur,
<1 wt% oxygen, and 0.16 wt% nitrogen.
4.2. Coal Pyrolysis. Pyrolysis is the destructive distillation of coal in
the absence of oxygen typically at temperatures between 400 and 5008C (133).
As the temperature of carbonaceous matter is increased, decomposition ulti-
mately occurs. Melting and dehydration may also occur. Coals exhibit more or
less definite decomposition temperatures, as indicated by melting and rapid evo-
lution of volatile components, including potential fuel liquids, during destructive
distillation (134). Table 10 summarizes an extensive survey of North American
coals subjected to laboratory pyrolysis. The yields of light oils so derived average
no more than ca 8.3 L/t (2 gal/short ton), and tar yields of ca 125 L/t (30 gal/short
ton) are optimum for high volatile bituminous coals (135).
Coal pyrolysis has been studied at both reduced and elevated pressures
(136), and in the presence of a variety of agents and atmospheres (137). Although
important to the study of coal structure and reactions, coal pyrolysis, as a means
to generate liquids, has proved to have limited commercial value.
COED Process. Sponsored by the Office of Coal Research of the U.S.
Department of the Interior, the COED process was developed by FMC Corp. as
Project Char-Oil-Energy Development (COED) through 1975 (138–140). Bench-
scale experiments led the way to construction in 1965 of a process development
unit employing multistage, fluidized-bed pyrolysis to process 45 kg/h (141). Cor-
related studies included hydrotreating of COED oil (142), high temperature
hydrodesulfurization of COED char, and investigations of char-oil and char-
water slurry pipelining economics (143). A pilot plant capable of processing up
to 33 t/d and hydrotreating 4.7 m3/d (30 bbl/d) was started up in 1970 (144),
and was operated successfully for a number of years (145).
The COED concept (139), designed to recover liquid, gaseous, and solid fuel
components, consists of four stages. Heat is generated by the reaction of oxygen
with a portion of the char in the last pyrolysis stage and is also introduced by the
air combustion of gas to dry feed coal. The number of stages in the pyrolysis, and
the operating temperatures in each, may be varied to accommodate high volatile
bituminous and subbituminous feed coals with widely ranging caking or agglom-
erating properties.
Oil condensed from the released volatiles from the second stage is filtered
and catalytically hydrotreated at high pressure to produce a synthetic crude oil.
Table 10. Average Yields and Range of Yields of Fischer Assay of Various Coalsa
Coke, % Tar, L/tc Light oil, L/tc Gas, m3/td Water, %
Rank of coal Average Range Average Range Average Range Average Range Average Range
semianthracite 3.2 0.14
low volatile bituminous 89.7 85.8–93.3 39.6 29.0–58.4 4.69 3.36–7.41 59.8 54.4–66.6 3.2 1.1–6.6
medium volatile bituminous 83.3 77.4–90.4 86.9 44.6–117.8 7.68 4.92–10.58 66.0 47.3–76.2 4.1 2.8–7.0
high volatile A bituminous 75.5 68.8–81.4 142.1 105.3–187.2 10.53 6.81–15.09 67.0 57.5–80.2 6.0 3.0–9.2
high volatile B bituminous 70.4 66.0–73.2 139.4 111.8–198.3 10.03 7.13–15.82 68.3 56.4–82.3 11.1 10.2–13.1
high volatile C bituminous 67.1 65.4–68.6 124.2 85.1–178.5 8.65 5.93–12.47 61.2 53.0–70.4 15.9 12.0–19.1
high volatile C bituminous 59.1 94.3 84.6–112.2 7.59 6.26–8.88 90.4 23.4
362
or subbituminous A
subbituminous A 81.9 81.0–82.8 6.21 6.12–6.26
subbituminous B 57.6 54.8–59.9 70.8 60.7–76.8 6.12 5.24–7.13 90.4 62.2–93.8 27.8 23.3–30.4
lignite 36.5 69.9 30.8–124.2 5.47 2.90–8.69 71.4 44.0
cannel 58.8 44.1–69.0 338.1 247.0–498.2 23.28 16.84–34.13 61.5 51.0–72.1 3.7 2.0–4.8
a
Ref. 135.
b
As-received basis; maximum temperature, 5008C.
c
To convert L/t to gal/short ton, divide by 4.6.
d
To convert m3/t to ft3/short ton, multiply by 29.4.
Vol. 2 FUELS, SYNTHETIC, LIQUID 363
Medium heat-content gas produced after the removal of H2S and CO2 is suitable
as clean fuel. The pyrolysis gas produced, however, is insufficient to provide the
fuel requirement for the total plant. Residual char, 50–60% of the feed coal, has a
heating value and sulfur content about the same as feed coal, and its utilization
may thus largely dictate process utility.
Properties. The properties of char products from two possible coal feeds,
a low sulfur Western coal, and a high sulfur Midwestern coal, are shown in
Table 11. The char derived from the low sulfur Western coal may be directly sui-
table as plant fuel, with only minor addition of clean process gas to stabilize its
combustion. Flue gas desulfurization may not be required. Flue gas from the
combustion of the char derived from the high sulfur Illinois coal, however,
requires desulfurization before it may be discharged into the atmosphere.
Typical COED syncrude properties are shown in Table 12. The properties of
the oil products depend heavily on the severity of hydroprocessing. The degree of
severity also markedly affects costs associated with hydrogen production and
compression. Syncrudes derived from Western coals have much higher paraffin
and lower aromatic content than those produced from Illinois coal. In general,
properties of COED products have been found compatible with expected indus-
trial requirements.
Occidental Petroleum Coal Conversion Process. Garrett R&D Co.
(now the Occidental Research Co.) developed the Oxy Coal Conversion process
based on mathematical simulation for heating coal particles in the pyrolysis
unit. It was estimated that coal particles of 100-mm diameter could be heated
throughout their volumes to decomposition temperature (450–5408C) within
0.1 s. A large pilot facility was constructed at LaVerne, California, in 1971. This
unit was reported to operate successfully at feed rates up to 136 kg/h (3.2 t/d).
Hot product char carries heat into the entrained bed to obtain the high
heat-transfer rates required. Feed coal must be dried and pulverized. A portion
of the char recovered from the reactor product stream is cooled and discharged as
product. The remainder is reheated to 650–8708C in a char heater blown with
air. Gases from the reactor are cooled and scrubbed free of product tar. Hydrogen
364 FUELS, SYNTHETIC, LIQUID Vol. 2
sulfide is removed from the gas, and a portion is recycled to serve as the entrain-
ment medium.
Properties. A high volatile western Kentucky bituminous coal, the tar
yield of which by Fischer assay was ca 16%, gave a tar yield of ca 26% at a pyr-
olysis temperature of 5378C (146–148). Tar yield peaked at ca 35% at 5778C and
dropped off to 22% at 6178C. The char heating value is essentially equal to that of
the starting coal, and the tar has a lower hydrogen content than other pyrolysis
tars. The product char is not suitable for direct combustion because of its 2.6%
sulfur content.
The TOSCOAL Process. The Oil Shale Corp. (TOSCO) piloted the low
temperature carbonization of Wyoming subbituminous coals over a two-year per-
iod in its 23 t/d pilot plant at Rocky Falls, Colorado (149). The principal objective
was the upgrading of the heating value in order to reduce transportation costs on
a heating value basis. Hence, the solid char product from the process represented
50 wt% of the starting coal but had 80% of its heating value.
Furthermore, 60–100 L (14–24 gal) oil, having sulfur content below 0.4 wt%,
could be recovered per metric ton coal from pyrolysis at 427–5178C. The
recovered oil was suitable as low sulfur fuel. Figure 15 is a flow sheet of the
Rocky Flats pilot plant. Coal is fed from hoppers to a dilute-phase, fluid-bed pre-
heater and transported to a pyrolysis drum, where it is contacted by hot ceramic
Vol. 2 FUELS, SYNTHETIC, LIQUID 365
Flue gas to
atmosphere
Balls
Scrubber
Coal
Gas
Separator
Ball
Surge heater Air Naphtha
hopper
Accumulator
Pre Gas oil
hea
ted
Ball elevator
coa
l Hot
ball
Lift pipe
s Resid
Pyrolysis
drum Trommel
as
eg
t flu
Ho
W
ar
m
ba
lls
Hot
char Char
cooler
Char to
sales
balls. Pyrolysis drum effluent is passed over a trommel screen that permits
char product to fall through. Product char is thereafter cooled and sent to
storage. The ceramic balls are recycled and pyrolysis vapors are condensed
and fractionated.
Properties. Results for the operation using subbituminous coal from
the Wyodad mine near Gillette, Wyoming, are shown in Table 13. Char yields
decreased with increasing temperature, and oil yields increased. The Fischer
assay laboratory method closely approximated the yields and product assays
that were obtained with the TOSCOAL process.
The volatiles contents of product chars decreased from ca 25–16% with tem-
perature. Char (lower) heating values, on the other hand, increased from ca
26.75 MJ/kg (11,500 Btu/lb) to 29.5 MJ/kg (12,700 Btu/lb) with temperature.
Chars in this range of heating values are suitable for boiler fuel application
and the low sulfur content (about equal to that of the starting coal) permits direct
combustion. These char products, however, are pyrophoric and require special
handling in storage and transportation systems.
Properties of the tar oil products are given in Table 14. The oils change only
slightly with change in the retorting temperature; sulfur levels are low. The
fraction boiling up to 2308C contains 65 wt% of phenols, cresols, and cresylic
acids.
366 FUELS, SYNTHETIC, LIQUID Vol. 2
Hydrogen Naphthalene
–252.5°C 218°C
Coal gas
Hydrogen 3,4–Dimethylphenol
sulfide 225°C
–61.8°C
Light oil
Benzene Picene
80.09°C 519°C
Coal tar
TOSCO tar oils have high viscosity and may not be transported by conven-
tional pipelines. Heating values of product gas on a dry, acid gas-free basis are in
the natural gas range if butanes and heavier components are included.
Coal Carbonization. In the by-product recovery of a modern coke oven,
coal tar is removed first by cooling the gases emanating, and light oil is removed
last by scrubbing the gas with solvents. Other products, including ammonia, phe-
nols, pyridine, or naphthalene, may be recovered between these operations. The
constituents of coal tar, light oil, and gas usually overlap considerably, ie, the
fractional condensation does not effectively separate individual components.
Assuming the lowest boiling coal tar constituent to be benzene (bp 80.098C),
and the highest boiling to be naphthalene, the overlapping compositions of gas,
light oil, and tar may be as shown graphically in Figure 16. Many chemical
compounds have been identified (8) in these substances. Included are most of
the significant constituents of petroleum-derived fuel liquids, although only a
few components are present in sufficient quantity to make commercial recovery
feasible.
The precise compositions of the light oil and coal tar recovered from coke-
oven gas is a distinct function of the design of the recovery system, as well as of
the properties of the starting coal. In general, 12.5–16.7 L/t (3–4 gal/m light oil
per short ton) of coal carbonized is recovered from high temperature coke-oven
operations. Light oil may contain 55–70% benzene, 12–20% toluene, and
4–7% xylene. Unrecovered light oil appearing in the effluent coal gas may
comprise ca 1 vol% and contribute ca 5% of the gas’s heating value. Refining of
light oil consists mainly of sulfuric acid washing, followed by fractional
distillation.
Large-scale recovery of light oil was commercialized in England, Germany,
and the United States toward the end of the nineteenth century (151). Industrial
coal-tar production dates from the earliest operation of coal-gas facilities. The
principal bulk commodities derived from coal tar are wood-preserving oils,
368 FUELS, SYNTHETIC, LIQUID Vol. 2
road tars, industrial pitches, and coke. Naphthalene is obtained from tar oils by
crystallization, tar acids are derived by extraction of tar oils with caustic, and tar
bases by extraction with sulfuric acid. Coal tars generally contain less than 1%
benzene and toluene, and may contain up to 1% xylene. The total U.S. production
of BTX from coke-oven operations is insignificant compared to petroleum product
consumptions.
5. Other Processes
5.1. Shale Oil. In the United States, shale oil, or oil derivable from oil
shale, represents the largest potential source of liquid hydrocarbons that can
be readily processed to fuel liquids similar to those derived from natural petro-
leum. Some countries produce liquid fuels from oil shale. There is no such indus-
try in the United States although more than 50 companies were producing oil
from coal and shale in the United States in 1860 (152,153), and after the oil
embargo of 1973 several companies reactivated shale-oil process development
programs (154,155). Petroleum supply and price stability has since severely cur-
tailed shale oil development. In addition, complex environmental issues (156)
further prohibit demonstration of commercial designs.
5.2. Heavy Oil. The definitions used to distinguish among naturally
occurring heavy petroleum oils, bitumens, asphalts, and tars, are subject to
broad variations. More than 10% of the world’s current crude oil production
has an API gravity below 208, or a specific gravity greater than 0:93415 15 . Oils
having sp gr1515 > 0:904 0:934 (20–258API) are considered heavy oils in most
classifications. However, Safaniyah crude oil produced in Saudi Arabia
having sp gr1515 of 0.893 (278API) carries the designation Arabian heavy, and
in petroleum parlance is generally referred to as heavy crude oil. Yet its pro-
duction method does not differ from that of Arabian medium or Arabian light
crude oils.
Energy in the form of injected water or CO2 may be supplied to increase the
rate of production of light crude oils. Application of heat to the reservoirs, eg,
using hot water, steam, heated CO2, fireflood, or in situ combustion, however,
is generally associated with the production of heavier, viscid crudes.
Heavy crude oil is widely distributed, and it is difficult to estimate reserves
separate from normal crude oil reserves or from tar sands deposits. Estimates of
petroleum reserves frequently include a large heavy oil component, which can
only be produced at significantly higher cost than light oil.
Most heavy oil production is concentrated in California, Canada, and Vene-
zuela. There is significant production of heavy oil in California from the Kern
River field near Bakersfield and in Canada from the Cold Lake deposit in
Alberta. Production generally involves steam drives, or the injection of steam
into reservoirs through special wells in prescribed sequences. Oil–water
mixtures are recovered, and often separated water is treated and reinjected.
Heavy oil may be upgraded through two main routes: coking and hydropro-
cessing. Virtually all established upgrading schemes involve some variant of
those two routes. The challenges in upgrading and refining are from the low
Vol. 2 FUELS, SYNTHETIC, LIQUID 369
hydrogen content and specific gravity and high sulfur, nitrogen, and metals
content of the heavy oil.
5.3. Tar Sands. Tar sands are considered to be sedimentary rocks
having natural porosity where the pore volume is occupied by viscous, petro-
leum-like hydrocarbons. The terms oil sands, rock asphalts, asphaltic
sandstones, and malthas or malthites have all been applied to the same
resource. The hydrocarbon component of tar sands is properly termed bitumen.
Distinctions between tar sands’ bitumens and heavy oils are based largely
on differences in viscosities. The bitumen in oil sand has a specific gravity of less
than 0.986 g/mL (128API), and thus oil sands may be regarded as a source of
extremely heavy crude oil. Whereas heavy oils might be produced by the same
techniques used for the lighter crude oils, the bitumens in tar sands are too vis-
cous for these techniques. Consequently these oil-bearing stones have to be
mined and specially processed to recover contained hydrocarbon.
Tar sands have been reported on every continent except Australia and
Antarctica. The best known deposits are the Athabasca of Canada, where
almost 60,000 km2 in northeastern Alberta is underlain with an estimated
138 109 m3 ð870 109 bblÞ of recoverable bitumen (157). The Alberta deposits
may contain up to 215 109 m3 ð1350 109 bblÞ of bitumen reserves. Venezuela
may have the largest accumulations in the world; the Orinoco heavy-oil belt has
been estimated by some (157) to contain as high as 636 109 m3 ð4000 109 bblÞ.
The Olenek reserves in the former USSR may contain ca 95 109 m3 ð600
109 bblÞ. The United States is estimated to have deposits of about 4:5
109 m3 ð28 109 bblÞ.
The Great Canadian Oil Sands, Ltd. (GCO) (Sun Oil Co.) has been operat-
ing a plant at Fort McMurray, Alberta, Canada, since 1967. Initially, some 8050
t/d (55,000 bbl/d) of synthetic crude oil was produced from coking (158) with the
project expanding to 9220 t/d (63,000 bbl/d). Since 1978, Syncrude Canada has
been producing ca 22,000 m3/d (140,000 bbl) synthetic crude oil by fluid coking
from their plant at Cold Lake, Alberta, Canada (159) with expansion planned
for ca 35,000 m3/d (225,000 bbl/d).
BIBLIOGRAPHY
‘‘Fuels, Synthetic Liquid’’ in ECT 1st ed., Vol. 6, pp. 960–983, by J. H. Arnold and H. Pich-
ler, Hydrocarbon Research, Inc.; ‘‘Carbon Monoxide–Hydrogen Reactions’’ in ECT 2nd
ed., Vol. 4, pp. 446–489, by H. Pichler and A. Hector, Carl-Engler und Hans-Bunte-Insti-
tut für Mineralöl und Kohleforschung der Technischen Hochschule Karlsruhe; ‘‘Fuels,
Synthetic (Liquid)’’ inECT 3rd ed., Vol. 11, pp. 447–489, by C. D. Kalfadelis and H.
Shaw, Exxon Research and Engineering Co.
CITED PUBLICATIONS
90. ARCO Chemical Co., Testimony to the Colorado Air Quality Control Commission on Pro-
posed Regulation No. 13 (Oxygenate Mandate Program), Denver, Colo., June 4, 1987.
91. Chem. Eng. News, 28 (Apr. 12, 1993).
92. U.S. Pat. 1,251,954 (Jan. 1, 1918), F. Bergius and J. Billwiller.
93. F. Fischer and H. Tropsch, Ges. Abhandl. Kenntis Kohle 2, 154 (1918).
94. F. Bergius, Pet. Z. 22, 1275 (1926).
95. Gas World 104, 421 (1936).
96. Pet. Times 42, 641 (1939).
97. H. H. Storch and co-workers, U.S. Bur. Mines. Tech. Pap. 622 (1941).
98. A. C. Fieldner and co-workers, U.S. Bur Mines Tech. Pap. 666 (1944).
99. M. L. Kastens and co-workers, Ind. Eng. Chem. 41, 870 (1949).
100. J. L. Wiley and H. C. Anderson, U.S. Bur. Mines Bull. 485, I J. L. Wiley and H. C.
Anderson, U.S. Bur. Mines Bull. 485, (1950), II J. L. Wiley and H. C. Anderson,
U.S. Bur. Mines Bull. 485, (1951), III J. L. Wiley and H. C. Anderson, U.S. Bur.
Mines Bull. 485, (1952).
101. C. C. Chaffee and L. L. Hirst, Ind. Eng. Chem. 45, 822 (1953).
102. Bituminous Coal Staff, U.S. Bur. Mines Rep. Invest. 5506 (1959).
103. E. L. Clark and co-workers, Ind. Eng. Chem. 42, 861 (1950).
104. L. L. Newman and A. P. Pipilen, Gas Age 119(10), 16 (1957); 119(11), 18 (1957).
105. U.S. Pat. 2,860,101 (Nov. 11, 1958), M. G. Pelipetz (to the United States of America).
106. S. Akhtar, S. Friedman, and P. M. Yavorsky, U.S. Bur. Mines Tech. Prog. Rep. 35 (1971).
107. S. Akhtar and co-workers, ‘‘Process for Hydrodesulfurization of Coal,’’ paper
presented at 71st National AIChE Meeting, Dallas, Tex., Feb. 20, 1972.
108. B. Linville and J. D. Spencer, U.S. Bur. Mines Inf. Cir. 8612 (1973).
109. U.S. Pat. 3,321,393 (May 23, 1967), S. C. Schuman, R. H. Wolk, and M. C. Chervenak
(to Hydrocarbon Research, Inc.).
110. Coal Age, 101 (May 1976).
111. G. A. Johnson and co-workers, ‘‘Present Status of the H-Coal Process,’’ paper 30, IGT
Coal Symposium, Chicago, 1973.
112. J. E. Papso, in Ref. 35, pp. 1-47–1-63.
113. ‘‘High Pressure Hydrogenation at Ludwigshafen-Heidelberg,’’ Vol. IA, FIAT Final
Report No. 1317, ATI No. 92,762, Central Air Documents Office, Dayton, Ohio, 1951.
114. H. H. Lowry and H. J. Rose, U.S. Bur. Mines Inf. Cir. 7420 (1947).
115. A. Baba and co-workers, Rep. Resources Res. Inst. Jpn., (22) (1955).
116. U.S. Pat. 2,221,866 (Nov. 19, 1940), H. Dreyfus.
117. D. L. Kloepper and co-workers, Solvent Processing of Coal to Produce a De-ashed
Product, Contract 14-01-0001-275, OCR Report No. 9, U.S. Government Printing
Office, Washington, D.C., 1965.
118. V. L. Brant and B. K. Schmid, Chem. Eng. Prog. 68(12), 55 (1969).
119. B. K. Schmid and W. C. Bull, ‘‘Production of Ashless, Low-Sulfur Boiler Fuels From
Coal,’’ paper presented at ACS Division of Fuel Chem. Symposium on Pollution
Control, New York, Sept. 12, 1971.
120. Demonstration Plant, Clean Boiler Fuels From Coal, OCR R&D report no. 82,
Interim report no. 1, Vols. 1–3, Ralph M. Parsons Co., Los Angeles, Calif., 1973–1975.
121. M. E. Frank and B. K. Schmid, ‘‘Economic Evaluation and Process Design of a Coal–
Oil–Gas (COG) Refinery,’’ paper presented at Symposium on Conceptual Plants
for the Production of Synthetic Fuels From Coal, AIChE 65th Annual Meeting,
New York, Nov. 26, 1972.
122. U.S. Pat. 3,341,447 (Sept. 12, 1967), W. C. Bull and co-workers (to the United States
of America and Gulf Oil Corp.).
123. B. K. Schmid and D. M. Jackson, ‘‘The SRC-II Process,’’ paper presented at Third
Annual International Conference on Coal Gasification and Liquefaction, University
of Pittsburgh, Aug. 3–5, 1976; D. M. Jackson and B. K. Schmid, ‘‘Production of
Vol. 2 FUELS, SYNTHETIC, LIQUID 375
Distillate Fuels by SRC-II,’’ paper presented at ACS Div. of Ind. and Eng. Chem.
Symposium, Colorado Springs, Col., Feb. 12, 1979.
124. W. R. Epperly and J. W. Taunton, ‘‘Exxon Donor Solvent Coal Liquefaction Process
Development,’’ paper presented at Coal Dilemma II ACS Meeting, Colorado Springs,
Colo., Feb. 12, 1979.
125. D. E. Blaser and A. M. Edelman, ‘‘Flexicoking for Improved Utilization of Hydrocar-
bon Resources,’’ paper presented at API 43rd Mid-Year Meeting, Toronto, Canada,
May 8, 1978.
126. S. F. Massenzio, in Ref. 35, pp. 6-3–6-18.
127. T. A. Cavanaugh, W. R. Epperly, and D. T. Wade, in Ref. 35, pp. 1-3–1-46.
128. L. E. Swabb, Jr., G. K. Vick, and T. Aczel, ‘‘The Liquefaction of Solid Carbonaceous
Materials,’’ paper presented at The World Conference on Future Sources of Organic
Raw Materials, Toronto, Can., July 10, 1978.
129. E. L. Huffman, Proceedings of the Third Annual International Conference on Coal
Gasification and Liquefaction, Pittsburgh, Pa., 1976.
130. H. D. Schindler, J. M. Chen, and J. D. Potts, Final Technical Report on DOE Con-
tract No. DE-AC22-79ET14804, Department of Energy, Washington, D.C., 1983.
131. H. D. Schindler, Final Technical Report on DOE Contract No. D-AC01-87ER30110,
Vol. 2, Department of Energy, Washington, D.C., 1989.
132. A. G. Comolli and co-workers, Proceedings of the DOE Direct Liquefaction Contrac-
tors’ Review Meeting, Pittsburgh, Pa., 1986.
133. M. G. Thomas, in B. R. Cooper and W. A. Ellingson, eds., The Science and Technol-
ogy of Coal and Coal Liquefaction, Plenum Press, New York, 1984, 231–261.
134. M. J. Burges and R. V. Wheeler, Fuel 5, 65 (1926).
135. W. A. Selvig and W. H. Ode, U.S. Bur. Mines. Bull. 571, (1957).
136. H. C. Howard, in Ref. 7, Vol. 1, 761–773.
137. Ibid., Suppl. Vol., 340–394.
138. J. F. Jones and co-workers, Chem. Eng. Prog. 62(2), 73 (1966).
139. J. A. Hamshar, H. D. Terzian, and L. J. Scotti, ‘‘Clean Fuels From Coal by the COED
Process,’’ paper presented at EPA Symposium on Environmental Aspects of Fuel
Conversion Technology, St. Louis, Mo., May 1974.
140. C. D. Kalfadelis and E. M. Magee, Evaluation of Pollution Control in Fossil Fuel
Conversion Processes, Liquefaction, Section 1, COED Process, EPA-650/2-74-009e,
Environmental Protection Agency, Washington, D.C., 1975.
141. R. T. Eddinger and co-workers, Char Oil Energy Development, Office of Coal
Research R&D Report No. 11, Vol. 1 (PB 169,562) and Vol. 2 (PB 169,563), U.S.
Government Printing Office, Washington, D.C., Mar. 1966.
142. J. F. Jones and co-workers, Char Oil Energy Development, Office of Coal Research
R&D Report No. 11, Vol. 1 (PB 173,916) and Vol. 2 (PB 173,917), U.S. Government
Printing Office, Washington, D.C., Feb. 1967.
143. M. E. Sacks and co-workers, Char Oil Energy Development, Office of Coal Research
Report 56, Interim Report No. 2, GPO Cat. No. 163.10:56/Int.2, U.S. Government
Printing Office, Washington, D.C., Jan. 1971.
144. J. F. Jones and co-workers, Char Oil Energy Development, Office of Coal Research
R&D Report No. 56, Final Report, GPO Cat. No. 163.10:56, U.S. Government Print-
ing Office, Washington, D.C., May 1972.
145. J. F. Jones and co-workers, Char Oil Energy Development, Office of Coal Research
R&D Report No. 73, Interim Report No. 1, GPO Cat. No. 163.10:73/Int 1, U.S.
Government Printing Office, Washington, D. C., Dec. 1972.
146. A. Sass, ‘‘The Garrett Research and Development Company Process for the Conver-
sion of Coal into Liquid Fuels,’’ paper presented at 65th Annual AIChE Meeting,
New York, Nov. 29, 1972.
147. Oil Gas J., 78 (Aug. 26, 1974).
376 FUELS, SYNTHETIC, LIQUID Vol. 2
GENERAL REFERENCE
SCOTT HAN
CLARENCE D. CHANG
Mobil Research and Development
Corporation
Vol. 2 FUELS, SYNTHETIC, GASEOUS 377
FUELS, SYNTHETIC,
GASEOUS
Substitute or synthetic natural gas (SNG) has been known for several centu-
ries. When SNG was first discovered, natural gas was largely unknown as a
fuel and was more a religious phenomenon (see NATURAL GAS) (1). Coal (qv)
was the first significant source of substitute natural gas and in the early
stages of SNG production the product was more commonly known under varia-
tions of the name coal gas (2,3). Whereas coal continues to be a principal source
of substitute natural gas (4) a more recently recognized source is petroleum
(qv) (5).
Coal can be converted to gas by several routes (2,6 – 11), but often a parti-
cular process is a combination of options chosen on the basis of the pro-
duct desired, ie, low, medium, or high heat-value gas. In a very general
sense, coal gas is the term applied to the mixture of gaseous constituents
that are produced during the thermal decomposition of coal at tempera-
tures in excess of 5008C ( > 930 F), often in the absence of oxygen (air) (3).
A solid residue (coke, char), tars, and other liquids are also produced in
the process:
The tars and other liquids (liquor) are removed by condensation leaving
principally hydrogen, carbon monoxide, and carbon dioxide in the gas phase.
This gaseous product also contains low boiling hydrocarbons (qv), sulfur-con-
taining gases, and nitrogen-containing gases including ammonia (qv) and
hydrogen cyanide. The solid residue is then treated under a variety of condi-
tions to produce other fuels which vary from a purified char to different
types of gaseous mixtures. The amounts of gas, coke, tar, and other liquid pro-
ducts vary according to the method used for the carbonization (especially the
retort configuration), and process temperature, as well as the nature (rank) of
the coal (3,11).
The recorded chronology of the coal-to-gas conversion technology beganin
1670 when a clergyman, John Clayton, in Wakefield, Yorkshire, produced in
the laboratory a luminous gas by destructive distillation of coal (12). At the
same time, experiments were also underway elsewhere to carbonize coal to pro-
duce coke, but the process was not practical on any significant scale until 1730
(12). In 1792, coal was distilled in an iron retort by a Scottish engineer, who used
the by-product gas to illuminate his home (13).
The conversion of coal to gas on an industrial scale dates to the early nine-
teenth century (14). The gas, often referred to as manufactured gas, was pro-
duced in coke ovens or similar types of retorts by simply heating coal to
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0701190519160509.a01
378 FUELS, SYNTHETIC, GASEOUS Vol. 2
vaporize the volatile constituents. Estimates based on modern data indicate that
the gas mixture probably contained hydrogen (qv) (ca 50%), methane
(ca 30%), carbon monoxide (qv) and carbon dioxide (qv) (ca 15%), and some
inert material, such as nitrogen (qv), from which a heating value of approxi-
mately 20.5 MJ/m3 (550 Btu/ft3) can be estimated (6).
Blue gas, or blue-water gas, so-called because of the color of the flame upon
burning (10), was discovered in 1780 when steam was passed over incandescent
carbon (qv), and the blue-water gas process was developed over the period 1859–
1875. Successful commercial application of the process came about in 1875 with
the introduction of the carburetted gas jet. The heating value of the gas was low,
ca 10.2 MJ/m3 (275 Btu/ft3), and on occasion oil was added to the gas to enhance
the heating value. The new product was given the name carburetted water gas
and the technique satisfied part of the original aim by adding luminosity to gas
lights (10).
Coke-oven gas is a by-product fuel gas derived from coking coals by the pro-
cess of carbonization. The first by-product coke ovens were constructed in France
in 1856. Since then they have gradually replaced the old and primitive method of
beehive coking for the production of metallurgical coke. Coke-oven gas is pro-
duced in an analogous manner to retort coal gas, with operating conditions,
mainly temperature, set for maximum carbon yield. The resulting gas is, conse-
quently, poor in illuminants, but excellent as a fuel. Typical analyses and heat
content of common fuel gases vary (Table 1) and depend on the source as well
as the method of production.
In Germany, large-scale production of synthetic fuels from coal began in
1910 and necessitated the conversion of coal to carbon monoxide and hydrogen.
The mixture of carbon monoxide and hydrogen is enriched with hydrogen from
the water gas catalytic (Bosch) process, ie, water gas shift reaction, and passed
cobalt
n CO þ ð2n þ 1Þ H2 ! C H
catalyst n 2nþ2
þ n H2 O
iron
2n CO þ ðn þ 1Þ H2 ! Cn H2nþ2 þ n CO2
catalyst
cobalt
n CO þ 2n H2 ! Cn H2n þ n H2 O
catalyst
iron
CO þ n H2 ! Cn H2n þ n CO2
catalyst
cobalt
n CO þ 2n H2 ! Cn H2nþ1 OH þ ðn 1Þ H2 O
catalyst
iron
ð2n 1Þ CO þ ðn þ 1Þ H2 ! Cn H2nþ1 OH þ ðn 1Þ CO2
catalyst
2 Ccoal þ O2 !2 CO
Ccoal þ H2 O!CO þ H2
CO þ H2 O!CO2 þ H2
380 FUELS, SYNTHETIC, GASEOUS Vol. 2
CO þ 3 H2 !CH4 þ H2 O
Prior to methanation, the gas product from the gasifier must be thoroughly pur-
ified, especially from sulfur compounds the precursors of which are widespread
throughout coal (23) (see SULFUR AND HYDROGEN SULFIDE RECOVERY). Moreover, the
composition of the gas must be adjusted, if required, to contain three parts hydro-
gen to one part carbon monoxide to fit the stoichiometry of methane production.
This is accomplished by application of a catalytic water gas shift reaction.
CO þ H2 O !
CO2 þ H2
The ratio of hydrogen to carbon monoxide is controlled by shifting only part
of the gas stream. After the shift, the carbon dioxide, which is formed in the gasi-
fier and in the water gas reaction, and the sulfur compounds formed during gasi-
fication, are removed from the gas.
Vol. 2 FUELS, SYNTHETIC, GASEOUS 381
The processes that have been developed for the production of synthetic nat-
ural gas are often configured to produce as much methane in the gasification step
as possible thereby minimizing the need for a methanation step. In addition,
methane formation is highly exothermic which contributes to process efficiency
by the production of heat in the gasifier, where the heat can be used for the
endothermic steam–carbon reaction to produce carbon monoxide and hydrogen.
C þ H2 O!CO þ H2
Cchar þ H2 O ! CO þ H2
The gaseous product from a gasifier generally contains large amounts of carbon
monoxide and hydrogen, plus lesser amounts of other gases and may be of low,
medium, or high heat value depending on the defined use (Table 3) (39,40).
The importance of coal gasification as a means of producing fuel gas(es) for
industrial use cannot be underplayed. But coal gasification systems also have
384 FUELS, SYNTHETIC, GASEOUS Vol. 2
Table 3. Gas Composition Requirements for Substitute Natural Gas and for Power
Generation
Product requirement
Characteristic Synthetic natural gas Power generation
methane content high, less synthesis required low, probably means no tars
H2/CO ratio high, less shifting required low, CO more efficient fuel
moisture content high, steam required for shift low, lower condensate treatment
costs
outlet low, maximizes methane high, precludes tar formation
temperature minimizes sensible heat loss provides for steam generation
leads to high cold gas efficiency reduces cold gas efficiency
gasifier oxidant O2 only, cost of N2 removal air or O2, low heating gas value
excessive acceptable fuel
Ccoal þ 2 H2 !CH4
although the reactions are more numerous and more complex as can be seen in
Table 4.
If air is used as a combustant, the product gas contains nitrogen and,
depending on design characteristics, has a heating value of approximately
5:6 11:2 103 MJ=m3 (150–300 Btu/ft3). The use of pure oxygen, although
expensive, results in a product gas having a heating value of 11–15 MJ/m3
(300–400 Btu/ft3) and carbon dioxide and hydrogen sulfide as by-products.
If a high heat-value gas (33.5–37.3 MJ/m3 (900–1000 Btu/ft3)) ie, SNG, is
the desired product, efforts must be made to increase the methane content.
Shift conversion reaction
CO þ H2 O!CO2 þ H2
CO þ 3 H2 !CH4 þ H2 O
2 CO þ 2 H2 !CH4 þ CO2
CO þ 4 H2 !CH4 þ 2 H2 O
The gasification is performed using oxygen and steam (qv), usually at elevated
pressures. The steam–oxygen ratio along with reaction temperature and
Vol. 2 FUELS, SYNTHETIC, GASEOUS 385
pressure determine the equilibrium gas composition. The reaction rates for these
reactions are relatively slow and heats of formation are negative. Catalysts may
be necessary for complete reaction (2,3,24,42,43).
Process Parameters. The most notable effects in gasifiers are those of
pressure (Fig. 1) and coal character. Some initial processing of the coal feedstock
60
50 H2
Composition, vol %
40 20
Heating value, MJ/m3
30
CO
20
CH4 15
10
0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Gasification pressure, MPa
Fig. 1. Variation of (—) gas composition and () heating value with gasifier pressure.
To convert MJ/m3 to Btu/ft3, multiply by 26.86. To convert MPa to psi, multiply by 145.
386 FUELS, SYNTHETIC, GASEOUS Vol. 2
C + H2O CO + H2
Coal Gas
CO + H2O CO2 + H2 + Heat
(a)
Coal
C + 2 H2 CH4 Gas
Steam and
H2-rich gas
(b)
Hydrogen
Coal + H2 CH4+ C Gas
Coal
(c)
Fig. 2. Chemistry of (a) gasifier; (b) hydrogasifier; and (c) devolatization processes. The
gaseous product of (a) is of low heating value; that of (b) and (c) is of intermediate heating
value.
Steam,
Coal oxygen,
Gas Coal or air
Coal
Gas
Steam, Steam,
oxygen, Ash oxygen, Ash Slag
or air or air
Fig. 3. Gasifier systems: (a), moving bed (dry ash); (b), fluidized bed; and (c), entrained
flow.
388 FUELS, SYNTHETIC, GASEOUS Vol. 2
The fluidized-bed system (Fig. 3) uses finely sized coal particles and the bed
exhibits liquid-like characteristics when a gas flows upward through the bed.
Gas flowing through the coal produces turbulent lifting and separation of
particles and the result is an expanded bed having greater coal surface area to
promote the chemical reaction. These systems, however, have only a limited abil-
ity to handle caking coals (see FLUIDIZATION).
An entrainment system (Fig. 3) uses finely sized coal particles blown into
the gas steam prior to entry into the reactor and combustion occurs with the
coal particles suspended in the gas phase; the entrained system is suitable for
both caking and noncaking coals.
Gasifier Options. The standard Wellman-Galusha unit, used for noncak-
ing coals, is an atmospheric pressure, air-blown gasifier (Fig. 4) fed by a two-com-
partment lockhopper system. The upper coal-storage compartment feeds coal
intermittently into the feeding compartment, which then feeds coal into the gasi-
fier section almost continuously except for brief periods when the feeding com-
partment is being loaded from storage. In small units up to 1.5 m internal
diameter ash is removed by shaker grates. In larger units, ash is removed con-
tinuously from the bottom of the gasifier into an ash-hopper section by a revol-
ving grate.
The grate is constructed of flat, circular steel plates set one above the other
with edges overlapping. The grate, revolving eccentrically within the gasifier,
causes ash to fall from the coal bed as the space between the grate and the
shell increases, and then pushes the ash down into the ash-hopper as the
space decreases. The smaller size units are brick-lined, although the larger
sizes are unlined and water-jacketed. Combustion air, provided by a fan, passes
over the warm (828C) jacket water causing the water to vaporize to provide the
necessary steam for gasification. Gas leaves from above the fuel bed at 428–
5388C for bituminous coal.
On the other hand, the agitator type of Wellman-Galusha unit, used for
gasification of any type of coal, uses a slowly revolving horizontal arm which
also spirals vertically below the surface of the fuel bed to retard channeling
and maintain a uniform bed. Use of the agitator not only allows operation with
caking coals but also can increase the capacity of the gasifier by about 25% for
use with other coals.
The Winkler gasifier (Fig. 5) is an example of a medium heat-value gas pro-
ducer which, when oxygen is employed, yields a gas product composed mainly of
carbon monoxide and hydrogen (43).
In the process, finely crushed coal is gasified at atmospheric pressure in a
fluidized state; oxygen and steam are introduced at the base of the gasifier. The
coal is fed by lockhoppers and a screw feeder into the bottom of the fuel bed. Sin-
tered ash particles settle on a grate, where they are cooled by the incoming oxy-
gen and steam; a rotating, cooled rabble moves the ash toward a discharge port.
The ash is then conveyed pneumatically to a disposal hopper.
The gas, along with entrained ash and char particles, which are subjec-
ted to further gasification in the large space above the fluid bed, exit the
gasifier at 954– 10108C. The hot gas is passed through a waste-heat boiler to
recover the sensible heat, and then through a dry cyclone. Solid particles are
removed in both units. The gas is further cooled and cleaned by wet scrubbing,
Vol. 2 FUELS, SYNTHETIC, GASEOUS 389
Green coal
Process steam
Saturated steam
Feed water
Process steam
O2 or air
Process steam
O2 or air
O2 or air
Process steam
Ash
Gas
flow
Boiler
Low
feed
pressure
water
steam
Four-headed
gasifier
Coal
Steam
Oxygen
Boiler feed
water
Ash discharge
Feed coal
Recycle tar
Coal lock
Drive
Steam
Scrubbing
cooler
Distributor Gas
Grate
Steam +
oxygen
Ash lock
The operating conditions in the gasifier (temperature and pressure) and the
reaction kinetics (residence time, concentration of the constituents, and rate con-
stants) determine the extent of conversion or approach to equilibrium.
The coal is fed through a lockhopper mounted on the top of the gasifier
where a rotating distributor provides uniform coal feed across the bed. When
processing caking coals, blades attached to the distributor rotate within
the bed to break up agglomerates. A revolving grate supports the bed at the
bottom and serves as a distributor for steam and oxygen. Solid residue is
removed at the bottom of the gasifier through an ash lockhopper. The entire
gasifier vessel is surrounded by a water jacket in which high pressure steam
is generated.
Vol. 2 FUELS, SYNTHETIC, GASEOUS 393
Crude gas leaves from the top of the gasifier at 288–5938C depending on the
type of coal used. The composition of gas also depends on the type of coal and
is notable for the relatively high methane content when contrasted to gases
produced at lower pressures or higher temperatures. These gas products can
be used as produced for electric power production or can be treated to remove
carbon dioxide and hydrocarbons to provide synthesis gas for ammonia, metha-
nol, and synthetic oil production. The gas is made suitable for methanation, to
produce synthetic natural gas, by a partial shift and carbon dioxide and sulfur
removal.
As in most of the high heat-value processes, the raw gas is in the medium
heat-value gas range and can be employed directly in that form. Removing the
carbon dioxide raises the heating value, but not enough to render the product
worthwhile as a high heat-value gas without methanation.
Methanation of the clean desulfurized main gas (less than 1 ppm total sul-
fur) is accomplished in the presence of a nickel catalyst at temperatures from
260–4008C and pressure range of 2–2.8 MPa (300–400 psi). Equations and reac-
tion enthalpies are given in Table 4.
Hydrogenation of the oxides of carbon to methane according to the above
reactions is sometimes referred to as the Sabatier reactions. Because of the
high exothermicity of the methanization reactions, adequate and precise cooling
is necessary in order to avoid catalyst deactivation, sintering, and carbon deposi-
tion by thermal cracking.
Catalytic methanation processes include (1) fixed or fluidized catalyst-bed
reactors where temperature rise is controlled by heat exchange or by direct cool-
ing using product gas recycle; (2) through wall-cooled reactor where temperature
is controlled by heat removal through the walls of catalyst-filled tubes; (3) tube-
wall reactors where a nickel–aluminum alloy is flame-sprayed and treated to
form a Raney-nickel catalyst bonded to the reactor tube heat-exchange surface;
and (4) slurry or liquid-phase (oil) methanation.
To enable interchangeability of the SNG with natural gas, on a calorific,
flame, and toxicity basis, the synthetically produced gas consists of a minimum
of 89 vol% methane, a maximum of 0.1% carbon monoxide, and up to 10% hydro-
gen. The specified minimum acceptable gross heating value is approximately
34.6 MJ/m3 (930 Btu/ft3).
In a combined power cycle operation, clean (sulfur- and particulate-free) gas
is burned with air in the combustor at elevated pressure. The gas is either low or
medium heat-value, depending on the method of gasification.
The hot gases from the combustor, temperature controlled to 9808C by excess
air, are expanded through the gas turbine, driving the air compressor and gener-
ating electricity. Sensible heat in the gas turbine exhaust is recovered in a waste
heat boiler by generating steam for additional electrical power production.
The use of hot gas clean-up methods to remove the sulfur and particulates
from the gasified fuel increases turbine performance by a few percentage
points over the cold clean-up systems. Hot gas clean-up permits use of the sen-
sible heat and enables retention of the carbon dioxide and water vapor in the
gasified fuel, thus enhancing turbine performance. Further, additional power
may be generated, prior to combustion with air, in an expansion turbine as the
394 FUELS, SYNTHETIC, GASEOUS Vol. 2
Gasifier
Oxygen
Steam Raw gas
Coal
Mill Water
Waste
Gasification heat
zone system
Purification
Mixer
Wastewater
Quench
zone
Coal Settler
slurry
Water
Ash lock
Slag
hot product fuel gas is expanded to optimum pressure level for the combined
cycle. Future advances in gas turbine technology (turbine inlet temperature
above 16508C) are expected to boost the overall combined cycle efficiency
substantially.
More recently, advanced generation gasifiers have been under develop-
ment, and commercialization of some of the systems has become a reality (36,
41). In these newer developments, the emphasis has shifted to a greater through-
put, relevant to the older gasifiers, and also to high carbon conversion levels and,
thus, higher efficiency units.
For example, the Texaco entrained system features coal–water slurry
feeding a pressurized oxygen-blown gasifier with a quench zone for slag cooling
(Fig. 8). In fact, the coal is partially oxidized to provide the heat for the gasifica-
tion reactions. The Dow gasifier also utilizes a coal–water slurry fed system
whereas the Shell gasifier features a dry-feed entrained gasification system
which operates at elevated temperature and pressure. The Kellogg Rust Wes-
tinghouse system and the Institute of Gas Technology U-Gas system are repre-
sentative of ash agglomerating fluidized-bed systems.
In response to the disadvantage that the dry ash Lurgi gasifier requires
that temperatures have to be below the ash melting point to prevent clinkering,
improvements have been sought in the unit; as a result the British Gas-Lurgi
GmbH gasifier came into being. This unit is basically similar to the dry ash
Lurgi unit insofar as the top of the unit is identical but the bottom has been mod-
ified to include a slag quench vessel (Fig. 9). Thus the ash melts at the high tem-
peratures in the combustion zone (up to 20008C) and forms a slag that runs into
Vol. 2 FUELS, SYNTHETIC, GASEOUS 395
Lock hopper
Distributor
Gas outlet
Steam/
Gasifier oxygen
(refractory inlet
lined, water
cooled)
Quench
Slag water
quench inlet
vessel
Slag
Quench
lock
water
hopper
outlet
the quench chamber, which is in reality a water bath where the slag forms gran-
ules of solid ash. Temperatures and reaction rates are high in the gasification
zone so that coal residence time is markedly reduced over that of the dry
ash unit.
In summary, these second-generation gasifiers offer promise for the future
in terms of increased efficiency as well as for use of other feedstocks, such as
biomass. The older, first-generation gasifiers, however, continue to be used.
396 FUELS, SYNTHETIC, GASEOUS Vol. 2
2 Hcoal þ O2 !H2 O
After burning, the sensible heat in the products of combustion can then be con-
verted into steam that can be used for external work or can be converted directly
into energy to drive a shaft, eg, in a gas turbine. In fact, the combustion process
actually represents a means of achieving the complete oxidation of coal.
The combustion of coal may be simply represented as the staged oxidation
of coal carbon to carbon dioxide
2 Ccoal þ O2 !2 CO
2 CO þ O2 !2 CO2
with any reactions of the hydrogen in the coal being considered to be of secondary
importance. Other types of combustion systems may be rate-controlled as a
result of the onset of the Boudouard reaction (see Table 4).
Vol. 2 FUELS, SYNTHETIC, GASEOUS 397
Scoal þ O2 !SO2
2 SO2 þ O2 !2 SO3
2 Ncoal þ O2 !2 NO
2 NO þ O2 !2 NO2
Ncoal þ O2 !NO2
The sulfur and nitrogen oxides that escape into the atmosphere can be converted
to acids by reaction with moisture in the atmosphere (see also AIR POLLUTION;
ATMOSPHERIC MODELS)
accelerated by contact with hot surfaces and, whereas the reaction away
from the walls may proceed slowly, reaction at the surface proceeds much
more rapidly.
Fluidized Beds. Fluidized-bed combustion occurs in expanded beds
(Fig. 3) and at relatively lower (9258C) temperatures; high convective transfer
rates exist resulting from the bed motion. Fluidized systems can operate under
substantial pressures thereby allowing more efficient gas clean-up. Fluidized-bed
combustion is a means for providing high heat-transfer rates, controlling sulfur,
and reducing nitrogen oxide emissions from the low temperatures in the combus-
tion zone.
There are, however, problems associated with pollution control. Whereas
the sulfur may be removed downstream using suitable ancillary controls, the sul-
fur may also be captured in the bed, thereby adding to the separations and
recycle problems. Capture during combustion, however, is recognized as the
ideal and is a source of optimism for fluidized combustion.
A fluidized bed is an excellent medium for contacting gases with solids, and
this can be exploited in a combustor because sulfur dioxide emissions can be
reduced by adding limestone, CaCO3, or dolomite, CaCO3 MgCO3, to the bed.
or
2 SO2 þ O2 þ 2 CaCO3 !2 CaSO4 þ 2 CO2
The spent sorbent from fluidized-bed combustion may be taken directly to dispo-
sal and is much easier than the disposal of salts produced by wet limestone
scrubbing. Alternatively, the spent sorbent may be regenerated using synthesis
gas, CO/H2.
CaSO4 þ H2 !CaO þ H2 O þ SO2
CaSO4 þ CO!CaO þ CO2 þ SO2
The calcium oxide product is supplemented with fresh limestone and returned to
the fluidized bed. Two undesirable side reactions can occur in the regeneration of
spent lime leading to the production of calcium sulfide:
CaSO4 þ 4 H2 !CaS þ 4 H2 O
CaSO4 þ 4 CO!CaS þ 4 CO2
the injection of coal–oil slurries into blast furnaces or for the burning of
coal–water slurries. The cyclone furnace, developed in the 1940s to burn coal
having low ash-fusion temperatures, is a horizontally inclined, water-cooled, tub-
ular furnace in which crushed coal is burned with air entering the furnace tan-
gentially. Temperatures may be on the order of 17008C and the ash in the coal is
converted to a molten slag that is removed from the base of the unit. Coal fines
burn in suspension; the larger pieces are captured by the molten slag and burn
rapidly.
As of this writing natural gas is a plentiful resource, and there has been a
marked tendency not to use other fossil fuels as SNG sources. However, pet-
roleum and oil shale (qv) have been the subject of extensive research efforts.
These represent other sources of gaseous fuels and are worthy of mention
here.
2.1. Petroleum. Thermal cracking (pyrolysis) of petroleum or fractions
thereof was an important method for producing gas in the years following its
use for increasing the heat content of water gas. Many water gas sets operations
were converted into oil-gasification units (55). Some of these have been used for
base-load city gas supply, but most find use for peak-load situations in the
winter.
In the 1940s, the hydrogasification of oil was investigated as a follow-up to
the work on the hydrogasification of coal (56,57). In the ensuing years, further
work was carried out as a supplement to the work on thermal cracking (58,59),
and during the early 1960s it became evident that light distillates having end
boiling points < 182 C and containing no sulfur could be catalytically reformed
by an autothermic process to pure methane (60). This method was extensively
used in England until natural (North Sea) gas came into use.
Prior to the discovery of plentiful supplies of natural gas, and depending on
the definition of the resources (1), there were plans to accommodate any short-
falls in gas supply from solid fossil fuels and from gaseous resources by the con-
version of hydrocarbon (petroleum) liquids to lower molecular weight gaseous
products.
CHpetroleum !CH4
but often to an equal or lesser extent on the feedstock type (5,62,63). In general
terms, gas heating values are on the order of 30–50 MJ/m3 (950–1350 Btu/ft3).
Catalytic Processes. A second group of refining operations which contri-
bute to gas production are the catalytic cracking processes, such as fluid-bed cat-
alytic cracking, and other variants, in which heavy gas oils are converted into
gas, naphthas, fuel oil, and coke (5).
The catalysts promote steam reforming reactions that lead to a product gas
containing more hydrogen and carbon monoxide and fewer unsaturated hydro-
carbon products than the gas product from a noncatalytic process (5). Cracking
severities are higher than those from thermal cracking, and the resulting gas is
more suitable for use as a medium heat-value gas than the rich gas produced by
straight thermal cracking. The catalyst also influences the reactions rates in the
thermal cracking reactions, which can lead to higher gas yields and lower tar and
carbon yields (5).
The basic chemical premise involved in making synthetic natural gas from
heavier feedstocks is the addition of hydrogen to the oil:
As in the case of coal, synthetic natural gas can be produced from heavy oil
by partially oxidizing the oil to a mixture of carbon monoxide and hydrogen
2 CHpetroleum þ O2 !2 CO þ H2
Whereas biomass has not received the same attention as coal as a source of gas-
eous fuels, questions about the security of fossil energy supplies related to the
availability of natural and substitute gas have led to a search for more reliable
and less expensive energy sources (75). Biomass resources are variable, but it has
been estimated that substantial amounts (up to 20 106 mJð20 1015 BtuÞ) of
energy, representing ca 19% of the annual energy consumption in the United
States. In addition, environmental issues associated with the use of coal have
led some energy producers to question the use of large central energy generating
plants. However, biomass may be a gaseous fuel source whose time is approach-
ing (see FUELS FROM BIOMASS).
The means by which synthetic gaseous fuels could be produced from a vari-
ety of biomass sources are variable and many of the known gasification technol-
ogies can be applied to the problem (70,71,76–82). For example, the Lurgi
circulatory fluidized-bed gasifier is available for the production of gaseous pro-
ducts from biomass feedstocks as well as from coal (83,84).
3. Gas Treating
The reducing conditions in gasification reactors effect the conversion of the sul-
fur and nitrogen in the feed coal to hydrogen sulfide, H2S, and ammonia, NH3.
Some carbonyl sulfide, COS, carbon disulfide, CS2, mercaptans, RSH, and hydro-
gen cyanide, HCN, are also formed in the gasifier. These compounds, along with
carbon dioxide, are removed simultaneously, either selectively or nonselectively,
from the gas stream in the clean-up stages of the process using commercially
available physical or chemical solvents and scrubbing agents (1,5,85–88).
Solvents used for hydrogen sulfide absorption include aqueous solutions of
ethanolamine (monoethanolamine, MEA), diethanolamine (DEA), and diisopro-
panolamine (DIPA) among others:
2 RNH2 þ H2 S!ðRNH3 Þ2 S
ðRNH3 Þ2 S þ H2 S!2 RNH3 HS
2 RNH2 þ CO2 þ H2 O!ðRNH3 Þ2 CO3
ðRNH3 Þ2 CO3 þ CO2 þ H2 O!2 RNH3 HCO3
2 RNH2 þ CO2 !RNHCOONH3 R
These solvents differ in volatility and selectivity for the removal of H2S,
mercaptans, and CO2 from gases of different composition. Other alkaline sol-
vents used for the absorption of acidic components in gases include potassium
carbonate, K2CO3, solutions combined with a variety of activators and solubili-
zers to improve gas–liquid contacting.
Vol. 2 FUELS, SYNTHETIC, GASEOUS 403
BIBLIOGRAPHY
‘‘Manufactured Gas’’ in ECT 1st ed., Vol. 8, pp. 765–800, by W. H. Fulweiler, Consultant;
‘‘Gas, Manufactured’’ in ECT 2nd ed., Vol. 10, pp. 353–442, by M. A. Elliot, Illinois Insti-
tute of Technology, and H. R. Linden, Institute of Gas Technology; ‘‘Carbon Monoxide–
Hydrogen Reactions’’ in ECT 2nd ed., Vol. 4, pp. 446–489, by H. Pichler and A. Hector,
Carl-Engler-und Hans-Bunte-Institut für Mineralöl und Kohleforschung der Technischen
Hochschule Karlsruhe; ‘‘Gaseous’’ under ‘‘Fuels, Synthetic’’ in ECT 3rd ed., Vol. 11,
pp. 410–446, by J. Huebler and J. C. Janka, Institute of Gas Technology.
CITED PUBLICATIONS
1. J. G. Speight, ed., Fuel Science and Technology Handbook, Marcel Dekker, Inc.,
New York, 1990.
2. J. G. Speight, The Chemistry and Technology of Coal, Marcel Dekker, Inc.,
New York, 1983.
3. R. A. Hessley, in Ref. 1, 645–734.
4. E. J. Parente and A. Thumann, eds., The Emerging Synthetic Fuel Industry,
Fairmont Press, Atlanta, Ga., 1981.
5. J. G. Speight, The Chemistry and Technology of Petroleum, 2nd ed., Marcel Dekker,
Inc., New York, 1991.
6. A. Kasem, Three Clean Fuels from Coal: Technology and Economics, Marcel Dekker,
Inc., New York, 1979.
7. L. L. Anderson and D. A. Tillman, Synthetic Fuels from Coal, John Wiley & Sons,
Inc., New York, 1979.
8. A. D. Dainton, in G. J. Pitt and G. R. Millward, eds., Coal and Modern Coal Proces-
sing: An Introduction, Academic Press, Inc., New York, 1979.
9. D. M. Considine, Energy Technology Handbook, McGraw-Hill Book Co., Inc.,
New York, 1977.
10. A. Francis and M. C. Peters, Fuels and Fuel Technology, Pergamon Press, Inc.,
New York, 1980.
404 FUELS, SYNTHETIC, GASEOUS Vol. 2
11. J. L. Johnson, in M. A. Elliott, ed., Chemistry of Coal Utilization, 2nd suppl. vol.
John Wiley & Sons, Inc., New York, 1981, Chapt. 23.
12. A. Elton, in C. Singer and co-workers, eds., A History of Technology, Vol. 4, Oxford
University Press, Oxford, U.K., 1958, Chapt. 9.
13. L. Shnidman, in H. H. Lowry, ed., Chemistry of Coal Utilization, John Wiley & Sons,
Inc., New York, 1945, Chapt. 30.
14. C. M. Jarvis, in Ref. 12, Vol. 5, Chapt. 10.
15. P. F. Mako and W. A. Samuel, in R. A. Meyers, ed., Handbook of Synfuels Technol-
ogy, McGraw-Hill Book Co., Inc., New York, 1984, Chapt. "2-1".
16. Oil Gas J. 90(3), 53 (1992).
17. K. T. Klasson and co-workers, C. Akin and J. Smith, eds., Gas, Oil, Coal, and
Environmental Biotechnology II, Institute of Gas Technology, Chicago, 1990, p. 408.
18. B. D. Faison, Crit. Revs. Biotechnol. 11, 347 (1991).
19. S. R. Bull, Energy Sources 13, 433 (1991).
20. R. C. Reid, J. M. Prausnitz, and T. K. Sherwood, The Properties of Gases and
Liquids, McGraw-Hill Book Co., Inc., New York, 1977.
21. Wellman Galusha Gas Producers, research bulletin no. 576A, McDowell-Wellman
Engineering Co., Cleveland, Ohio, May 1976.
22. R. F. Probstein and R. E. Hicks, Synthetic Fuels, pH Press, Cambridge, Mass.,
1990.
23. J. S. Sinninghe Damste and J. W. de Leeuw, Fuel Processing Technol. 30, 109 (1992).
24. N. Berkowitz, Introduction to Coal Technology, Academic Press, Inc., New York,
1979.
25. R. K. Hessley, J. W. Reasoner, and J. T. Riley, Coal Science, John Wiley & Sons, Inc.,
New York, 1986.
26. M. O. Holowaty and co-workers, in R. A. Meyers, ed., Coal Handbook, Marcel
Dekker, Inc., New York, 1981, Chapt. 9.
27. W. Eisenhut, in Ref. 11, Chapt. 14.
28. F. W. Gibbs, in Ref. 12, Vol. 3, Chapt. 25.
29. G. J. Pitt and G. R. Millward, eds., in Ref. 8, p. 52.
30. L. Seglin and S. A. Bresler, in Ref. 11, Chapt. 13.
31. E. Aristoff, R. W. Rieve, and H. Shalit, in Ref. 11, Chapt. 16.
32. D. McNeil, in Ref. 11, Chapt. 17.
33. C. Y. Cha and co-workers, Report No. DOE/MC/24268-2700 (DE89000967), United
States Department of Energy, Washington, D.C., 1988.
34. F. M. Fess, History of Coke Oven Technology, Gluckauf Verlag, Essen, Germany,
1957.
35. M. A. Elliott, ed., in Ref. 11.
36. D. Hebden and H. J. F. Stroud, in Ref. 11, Chapt. 24.
37. R. A. Meyers, ed., in Ref. 15.
38. D. R. Simbeck, R. L. Dickenson, and A. J. Moll, Energy Prog. 2, 42 (1982).
39. W. W. Bodle and J. Huebler, in Ref. 26, Chapt. 10.
40. L. K. Rath and J. R. Longanbach, Energy Sources 13, 443 (1991).
41. S. Alpert and M. J. Gluckman, Ann. Rev. Energy 11, 315 (1986).
42. R. A. Meyers, ed., in Ref. 26.
43. J. H. Martin, I. N. Banchik, and T. K. Suhramaniam, Report of the Committee on
Production of Manufactured Gases, report no. 1GU/B-76, London, 1976.
44. J. F. Farnsworth, Ind. Heat. 41(11), 38 (1974).
45. J. F. Farnsworth, Proceedings Coal Gas Fundamentals Symposium, Institute of Gas
Technology, Chicago, 1979.
46. J. C. Hoogendoorn, Proceedings of the Ninth Pipeline Gas Symposium, Chicago,
Oct. 31–Nov. 2, 1977.
Vol. 2 FUELS, SYNTHETIC, GASEOUS 405
JAMES G. SPEIGHT
Consultant
OCTANE ENHANCERS
1. Introduction
Historically, the quality of motor gasoline and the technical development of the
spark ignition Otto engine were closely connected. Since its invention in 1876,
the Otto engine technology has been continuously improved, a process which is
still going on.
Of particular importance for this development was the parallel improve-
ment in gasoline knock resistance. Knock resistance is one of the most important
gasoline properties and its improvement allowed the engine manufacturers to
increase the compression ratio. This in turn resulted in higher engine efficiency
and improved exhaust emission performance. The knock resistance is measured
in terms of gasoline octane numbers in accordance with worldwide standardized
test procedures.
Figure 1 (see next page) shows, for Germany as an example, the close rela-
tionship between octane number and engine compression ratio from the turn of
the century until now. Over this period a variety of measures were applied to
improve refinery technologies to meet the increasing engine octane demand.
These measures incorporated the integration of complex processing plants,
from which a range of gasoline component streams with different octane levels
became available. In 1929, alkyllead compounds were introduced in the United
States as efficient octane enhancers. Thereafter these were used generally to
improve the gasoline octane quality until environmental concerns led to their
use being stepwise restricted worldwide. The resulting shortfall of octane was
compensated by improvements in refinery processes, as well as by the use of
alternative octane enhancers such as fuel oxygenates. Starting in the 1970s a
number of fuel oxygenates became available and were increasingly used for
gasoline blending.
More recently, efforts to reduce air pollution led to the enforcement of more
restrictive exhaust emission limits for vehicles. In this context, passenger cars
with catalytic converters were introduced. Since the addition of lead to gasoline
results in the deactivation of the catalytic converters, unleaded gasoline of high
octane quality was consequently made available.
1
2. Technical Information
Fig. 2. Pressure profile versus crank angle during combustion. A) Normal combustion;
B) Knocking combustion.
3. History
Until the mid-1920s there was only one gasoline grade available in most coun-
tries, predominantly produced from straight-run crude oil products which were
selected according to their composition of aliphatic and aromatic components. In
order to meet the increasing octane demand, eg, indicated by the introduction of
a premium grade in 1924, several refinery processes were developed. Today, a
Vol. 2 OCTANE ENHANCERS 409
typical refinery produces up to twenty individual product streams used for gaso-
line blending (! Automotive Fuels, Chap. 4.1). The octane enhancement of tetra-
ethyllead (TEL) was discovered in 1922 in the United States by T. MIDGLEY, JR.
(2). It was produced by the Ethyl Gasoline Corporation and first introduced onto
the U.S. market in 1929. In Germany it became available shortly before World
War II as an octane enhancer. TEL and TML (tetramethyllead) are so-called
‘‘lead fluids’’ and are still used in some parts of the world. However, leadalkyl
compounds have disappeared from gasoline in most countries due to environ-
mental concerns and because lead deactivates catalytic converters.
Other octane enhancers–except fuel oxygenates–never achieved any signif-
icant degree of application. The current dominance of unleaded gasolines in the
market, however, causes octane shortfalls. This situation has led to renewed
attention to additional octane sources, including both new and known oxyge-
nates and alternative octane enhancers. Therefore, octane enhancers may play
a more pronounced role in the future provided that environmental aspects are
satisfied, such as toxicological properties and full compatibility with modern
exhaust emission control systems.
ONM ONBG ð1 xÞ
B ON ¼
x
where B–ON is the blend octane number of the component; ONM the octane
number of the mixture; ONBG the octane number of the base gasoline; and x is
the volume fraction of the component in the blend. All parameters on the right-
handside of this equation must be determined experimentally. This equation
attributes the entire change in octane number to the component. Graphically
(Fig. 3), it reflects a linear extrapolation from the base gasoline ONBG, via the
mixture ONM, to a fictive B–ON value of the neat component. If the calculated B–
ON is compared with the measured octane number of the neat component (ONC),
two alternatives are possible. If the B–ON is larger than the measured ONC, the
mixture has a ‘‘synergistic’’ property, if the B–ON is smaller than the ONC, the
mixture shows an ‘‘antagonistic’’ behavior.
Whereas organic nitrates and aliphatic amines were generally found to pro-
mote knocking, aromatic amines achieve relatively high octane enhancements.
Examples of B-ON values determined for aniline and some aniline derivatives
are shown in Table 1. All measurements were performed in the same base gaso-
line with a component addition of 2 vol%. The octane enhancement efficiency var-
ies with molecular structure. Relative to aniline, differences are small when
hydrogen atoms of the aromatic ring are substituted by methyl or ethyl groups.
Substitution with longer aliphatic chains results in reduced efficiency. Whereas
primary and secondary amines have good octane enhancing properties, tertiary
amines eg, N,N-dimethylaniline are ineffective.
A serious disadvantage of nitrogen-containing components results from the
fact that nitrogen is partially oxidized in the combustion process. This increases
nitrogen oxide emissions in the exhaust gases. There is also experimental evi-
dence that some of the components may cause deposit formations in fuel intake
manifolds and/or combustion chambers, when used at the required concentra-
tions. Of the listed components, only N-methylaniline was used as an octane
enhancer for a short period after World War II in Germany.
5.3. Oxygen-Containing Components. Organic oxygen compounds
such as alcohols, ketones, ethers, esters, and phenols are commonly characterized
as ‘‘oxygenates’’ or ‘‘fuel oxygenates’’. The majority of the oxygenates act as octane
enhancers; their efficiency depends, however, on the functional group and on the
individual structure of the molecule. Some derivatives of esters, ketones, and phe-
nols achieve relatively high B-ON values, as indicated in Table 2. Because of other
inappropriate properties eg, high boiling points, high acidity, low material com-
patibility, and high toxicity they are not used as octane enhancers.
Of the oxygenates, only alcohols and ethers are used to any significant
extent. They are preferred over other octane enhancers because of lower costs,
availability, handling, and their physical and chemical properties, which are
similar to gasoline components. Names, abbreviations, formulae, and molecular
mass of the most important alcohols and ethers suitable for fuel blending are
listed in Table 3.
EC Directive on Fuel Oxygenates. The use of oxygenates in gasolines
was established during the 1970s and has developed rapidly to a large-scale
Vol. 2 OCTANE ENHANCERS 413
business. Prior to any official regulation, the German automobile and oil indus-
try reached an interindustrial agreement to restrict the methanol content of fuel
to 3 vol% with the obligation to use an adequate cosolvent component (eg, TBA) to
ensure appropiate water tolerance of methanol-containing fuels. For the EC the
European Council passed a directive in December 1985, regulating the use of
oxygenates in gasolines after several years of intensive negotiations of the min-
istries of commerce supported by experts (4). Meanwhile, the technical annex of
the directive has been implemented in the gasoline standards of all EC member
states.
The annex lays down definitions and blend compositions of oxygenates. The
most important information is summarized in Table 4, providing the maximum
volumetric additions of individual oxygenates and the maximum gravimetric
oxygen content of the blends.
Blend Octane Numbers of Oxygenates. The octane efficiency of oxyge-
nates cannot be characterized in a simple way because it depends on a variety of
factors. Therefore, B-ON ranges are usually defined to account for their possible
variation. This makes it difficult to compare octane efficiencies of different
oxygenates.
To overcome this problem, Veba Oil conducted a series of investigations
with the objective to characterize the octane blend behavior of some oxygenates.
In the progress of this work, the gasoline sensitivity (SEN, ie, the difference
between RON and MON) was found to be a suitable correlation parameter (5).
Empirically it may be considered as an approximate gauge for the gasoline com-
position; a high paraffinic content corresponds to low SEN, high aromatic and/or
olefinic contents result in high SEN.
The measurements were performed in a variety of different unleaded gaso-
lines, the compositions of which were widely varied. Oxygenate concentrations
ranged from 10 to 20 vol%. Examples of the graphic correlations based on linear
regression of B-ON data versus gasoline SEN are shown in Figure 5 for methanol
and in Figure 6 for PTBE. Both B-RON and B-MON decrease with increasing
gasoline SEN. Similar blend behavior was found for the majority of the oxyge-
nates investigated. Typically, lower alcohols such as methanol and ethanol
show a strong dependence of B-ON on SEN. For higher alcohols and ethers
Vol. 2 OCTANE ENHANCERS 415
this dependence is less pronounced; in a few cases, eg, for TBA, no significant
dependence was observed.
Other Properties of Oxygenates. Beside their octane efficiency, there are
other properties of oxygenates which also affect the gasoline quality. Again their
blend behavior may not be linear so that blend properties must be determined
analoguously to the B-ON procedure. Selected properties of alcohols and ethers
used as oxygenates are listed in Table 5.
Density. Densities of alcohols are generally higher than the specified
density region for gasolines, whereas those of ethers range within it. Blend den-
sities of alcohols and ethers show only moderate differences when compared to
the densities of the neat components.
Vapor Pressure. Due to their polarity, lower alcohols such as methanol
and ethanol possess blend vapor pressures (B-VP) which exceed the vapor pres-
sure of the neat components, because association of alcohol molecules is hindered
in the mixture. Therefore, their B-VP values strongly depend on the component
concentration. In contrast, B-VP data of ether components hardly show any
dependence on concentration. They are significantly lower than those of lower
alcohols and are similar to the vapor pressure of the neat components.
Volatility. The addition of oxygenates affects the boiling characteristics of
the gasolines. When ethers are blended, only moderate changes in the distillation
416 OCTANE ENHANCERS Vol. 2
gasoline pool (10). More than 40 % of the total onstream capacity of MTBE is used
as merchant MTBE, which is available to supply refiners in need of additional
octane quality (11). For production of MTBE, see ! Methyl Tert-Butyl Ether,
Chap. 4.
Reaction of ethanol with isobutene produces ethyl tert-butyl ether (ETBE),
which has similar properties to MTBE. This process offers an outlet of bioethanol
(12,13). Reaction of 2-propanol with isobutene yields PTBE (14,15), an oxygenate
with excellent blend properties such as low vapor pressure and low oxygen con-
tent in combination with high octane quality.
Based on a similar technology methylbutenes (amylenes) of C4 to C6
streams can be etherified with methanol to TAME (16,17), a blending ether
with a high octane number. The BP Etherol process (18) uses unsaturated C5
compounds of cracker fractions. Branched olefins react with methanol to form
higher ethers. However, most of the branched olefins have terminal double
bonds with low reactivity. To increase the yield of 3-methyl-2-butene (isoamy-
lene) the double bond is shifted in the presence of a palladiumimpregnated
cation-exchanger resin. The isomerization proceeds according to:
fuels, depending on the system and oxygen content (Fig. 8). The gasolines were
blended with MTBE and ethanol. Emissions of unburned hydrocarbons and
nitrogen oxides were virtually unaffected.
Under European conditions, studies were made with ethers and alcohols on
the influence of oxygen content on the limited pollutants such as carbon monox-
ide, hydrocarbons, and nitrogen oxide (5). Blend components were methanol with
IPA – TBA as cosolvents and the ethers MTBE, PTBE, and BTBE. Another
investigation was conducted by the European Fuel Oxygenates Association
(EFOA) using mixtures of ethers and alcohols to study the dependence on oxyge-
nates systematically (22). Oxygen content was varied up to 3.6 wt %. The results
showed that oxygenates reduce CO emissions to a large degree. The reduction
increases with oxygen content but is not a function of the type of oxygenate.
Toxicology and Occupational Health. MTBE. Similar investigations of
toxicity and other potential health effects of MTBE were performed. Animal test-
ing to date has indicated that after ingestion, inhalation, or even injection of
MTBE, it is mostly excreted unchanged (4).
PTBE. The Ames test does not show any mutagenity. According to an
acute toxicity study in rats, PTBE is considered to be nontoxic up to a dose of
2000 mg per kilogram body weight.
Storage and Transportation. The higher ethers MTBE, PTBE, and
TAME are stable in the presence of air under atmospheric conditions. Formation
of peroxides is not observed during contact with oxygen. Containers, sealing
materials, and also elastomers in automobiles have to meet the same require-
ments as for contact with gasoline.
6. Metal-Containing Additives
slightly less effective. For practical use, most of them must be excluded because
of high costs, instability, toxicity, impact on catalytic converters, or other unde-
sired properties.
Their use in gasolines also requires an official approval in most European
countries and in the United States. In Germany, the application of metal-
containing additives–other than lead–is restricted by the gasoline lead act
(Benzin – Blei – Gesetz).
The octane enhancement of organometallic additives is explained by a cat-
alytic effect, because during combustion the metal compound is oxidized to form
very fine solid particles of metal oxide. These particles have large surface areas
with active centers which can react with free radicals formed by partial oxidation
of the gasoline. Therefore the spontaneous combustion is delayed and knocking
does not occur. The consumed centers are reactivated by oxygen and can react
again with free radicals. Due to this cycle only small concentrations of additive
are necessary to cause high octane enhancement.
A metal-free component is unable to delay the spontaneous combustion cat-
alytically. The organic compounds are consumed irreversibly by the intermediate
reactants which are responsible for knocking. Therefore the effectivity of anti-
knocking components is limited and high concentrations are needed to get a
higher octane quality.
A disadvantage of metal-containing additives results from the formation of
solid combustion products that lead to deposit accumulation in the combustion
chambers. For the lead fluids this problem was solved by addition of lead scaven-
gers. However, this is unfeasible for most other organometallic additives, thus
their use must be restricted to very low concentrations. Additives that achieve
sufficient octane improvements in the ppm concentration range and that meet
other technical and economical requirements are methylcyclopentadienylmanga-
nese tricarbonyl (MMT) and dicyclopentadienyliron (ferrocene).
6.1. Alkyllead Compounds. The strong octane-boosting effect of tetra-
ethyllead (TEL) was discovered in 1921 in the United States (23). Its wide use
began in the United States in 1929 and in Germany shortly before World
War II. Curiously, investigations at that time regarded lead as relatively envir-
onmentally friendly.
The effect of lead is based on an intervention in the mechanism of the com-
bustion process. In precombustion reactions, organic peroxides are formed,
which decompose to form free radicals. The latter initiate knocking reactions.
These highly reactive radicals can be deactivated by very finely divided lead com-
pounds formed from the alkyllead compounds in the combustion chamber
(24,25).
In addition to tetraethyllead, which was used exclusively at first, tetra-
methyllead (TML) came into use from 1960 onward. Mixed alkyllead compounds,
such as diethyldimethyllead or ethyltrimethyllead, and mixtures of TEL and
TML were also employed, because a broad-spectrum effect can be achieved
through a combination of different characteristics (Table 6). Thus, the compara-
tively low boiling point of TML is good for increasing the front octane number.
(Fig. 9).
The increase in octane number that can be achieved with alkyllead com-
pounds depends strongly on fuel composition. The lead susceptibility, a measure
Vol. 2 OCTANE ENHANCERS 421
of the response behavior to the addition of lead, decreases in the following order:
paraffins cycloparaffins > aromatics > olefins > diolefins (& 0). Accordingly,
fuel components rich in paraffins, such as straight-run gasoline, alkylate, and
isomerate, react very well to lead addition, whereas olefin-rich catalytically
cracked gasoline has a poor response (Fig. 10).
In connection with the addition of alkyllead compounds to gasoline, the
term ‘‘lead bonus’’ was sometimes used. It meant that highly leaded fuels, in com-
parison with unleaded or low-lead ones, have a disproportionately high road
octane number, although the RON and MON determined in test engines are
equally high (26–30). Investigations carried out with the introduction of
unleaded fuels showed, however, that low-lead and unleaded fuels behave almost
identically with regard to the road octane number (31) in most cases.
Fig. 9. Effect of tetraethyllead and tetramethyllead on the gasoline front octane number
(FON).
422 OCTANE ENHANCERS Vol. 2
Fig. 10. Effect of tetraethyllead on some base fuels (octane quality of the base fuel types
are widely varying in practice). a) Olefin-rich fuel; b) Aromatic-rich fuel; c) Paraffin-rich
fuel.
Mr 186.04
Density 1.49 g/cm3
mp 173 8C
bp (at 102.3 kPa) 249 8C
Triple point 183 8C
Vapor pressure
40 8C 0.0066 kPa
100 8C 0.34 kPa
Heat of fusion (at 175 8C) 17.8 kJ/mol
Heat of vaporization (at 175 8C) 47.2 kJ/mol
Heat of sublimation (at 25 8C) 70.2 kJ/mol
Start of decomposition 465 8C
Fe þ 2 C2 H5 OH!FeðOC2 H5 Þ2 þ H2
2 Chemical conversion
FeðOC2 H5 Þ2 þ 2 C5 H6 ! FeðC5 H5 Þ2 þ 2 C2 H5 OH
Total:
Fe þ 2 C5 H6 ! FeðC5 H5 Þ2 þ H2
Europe. Generally it was found that the lower the octane number of the base
gasoline, the higher the octane gain. Several measurements of octane response
versus concentration have shown that response is more pronounced in the low
ferrocene concentration region and that levelling-off effects are obtained at
higher concentrations.
Toxicology. The toxicological data of ferrocene are listed in Table 8.
REFERENCES
KUNO SCHÄDLICH
PETER SCHUG
VEBA OEL AG, Gelsenkirchen Federal
Republic of Germany
WERNER DABELSTEIN
ARNO REGLITZKY
Deutsche Shell AG, Hamburg Federal
Republic of Germany
METHYL TERT-BUTYL
ETHER
1. Introduction
and 3 (3). Methyl tert-butyl ether has unlimited miscibility with all ordinary
organic solvents and all hydrocarbons.
Chemical Properties. Methyl tert-butyl ether is very stable under alka-
line, neutral, and weakly acidic conditions. In the presence of strong acids, it is
cleaved to methanol and isobutene. Depending on reaction conditions the latter
can form isobutene oligomers.
4. Production
the spent C4 fraction. Usually, if the feedstock is a raffinate 1 stream, the spent
C4 fraction is called raffinate 2.
This high conversion can also be obtained by using highly sulfonated acidic
resins in the reaction section followed by an additional catalytic distillation col-
umn in the refining section.
Refining Section. Catalytic distillation (CD) or reactive distillation (RD)
refers to a process, where both catalytic reaction and distillation are carried
out simultaneously in the debutanizer column. From the viewpoint of reaction
engineering, this column acts as a two-phase countercurrent flow, fixed-bed cat-
alytic reactor. Usually CD is defined as a process, in which a heterogeneous cat-
alyst is located in a distillation column. A more common term for this operation is
RD. This term does not distinguish between a homogeneous or heterogeneous
catalyzed reaction in distillation columns.
The most important advantage of using CD for MTBE synthesis lies
in the elimination of equilibrium limitation of isobutene conversion as a
result of continuous removal of the reaction product MTBE from the reaction
mixture.
The partly reacted mixture from the reaction section – which is usually in
chemical equilibrium – enters the CD column below the catalyst packing zone to
ensure the separation of the high-boiling component MTBE from the feed
stream. The catalyst packing is installed in the upper mid portion of the column
with normal distillation sections above and below.
The CD-based MTBE process was first developed by CR & L using a
patented catalyst support system (commercialized as CD MTBE or CD
ETHEROL) (16,61). The catalyst used in the CD columns is essentially the
434 METHYL TERT-BUTYL ETHER Vol. 2
5. Environmental Protection
The small amounts of byproducts in MTBE need not be removed if the product
is to be used for gasoline. The catalyst is regenerable, and no environmental pro-
blems arise from its disposal.
Because MTBE is soluble in water, suitable precautions against
groundwater contamination must be taken in loading and storage areas.
Recently published complains about groundwater contamination and the
reactions taken by public authorities in parts of the United States arose
from careless storage of the product. For the environmental significance of
MTBE, see Section 11. For emission control in the storage and handling of
MTBE.
6. Quality Specifications
MTBE 98–99 wt %
Alcohols (methanol, tert-butanol) 0.5–1.5 wt %
Hydrocarbons (C5, C6-hydrocarbons, 0.1–1.0 wt %
diisobutenes)
Water 50–1500 ppm by weight
Total sulfur max. 10 ppm by weight
Residue on evaporation max. 10 ppm by weight
For special applications (see Section 10.2), a high-purity MTBE (ether con-
tent > 99.98 % wt) is marketed under the trade name Driveron S by Oxeno. High
purity MTBE is also available from Shell and ARCO.
436 METHYL TERT-BUTYL ETHER Vol. 2
7. Chemical Analysis
Being noncorrosive, MTBE can be handled in the same way as fuel, and the exist-
ing distribution system can also be used for fuels containing MTBE. The ether
has an indefinite storage life even in the presence of air. Because water is mis-
cible with MTBE (though only slightly), dry storage is required. The product can
be stored in nonpressurized containers. Carbon steel can be used as container
material, as well as aluminum, brass, copper, polyethylene, or polypropylene.
Teflon, Buna-N, and other fuel-resistant plastics and rubbers can be used for
seals; the use of vinylidene fluoride-hexafluoropropene copolymer (Viton) is not
recommended.
Methyl tert-butyl ether has a vapor pressure of 61 kPa at 40 8C. Emission
from storage facilities can be controlled or prevented by ordinary measures.
The usual safety precautions for highly inflammable liquids must be employed.
Extinguishing agents for fire fighting are powder, carbon dioxide, and
alcohol-resistant foam such as Tutogen L. With alcohol-resistant foams, a higher
application rate is necessary for MTBE than for pure hydrocarbons.
Transport classifications are as follows:
9. Legal Aspects
The principal reason for the current importance of MTBE is environmental leg-
islation (especially in the United States, Japan, and Western Europe). Although
laws do not require the use of MTBE, it is often the best way to satisfy their
provisions.
In the European Community, the Council Directive 85/536/EEC, December 5,
1985 (33), does not require special labeling at filling stations for fuels containing
up to 15 vol % MTBE.
Vol. 2 METHYL TERT-BUTYL ETHER 437
10. Uses
12. Ecotoxicology
pipelines, minor on-going spill incidents, and larger volume surface spills. In
many cases, gasoline retail stations and production terminals are sources of
MTBE in groundwater. The concentration of MTBE in groundwater and the
extent of groundwater contamination depend largely on the volume and duration
of a gasoline release and whether remediation has been initiated at the site. As
with any remedial effort for most contaminants, source removal is critical for the
success of remediation systems at MTBE-impacted sites. A quick response fol-
lowing an accidental release of an MTBE-blended gasoline is of great importance
regardless of the selected remediation method since it decreases the size of the
area requiring characterization and the volume of groundwater needing treat-
ment thereby minimizing site characterization and remediation costs.
13.1. Overview of MTBE Remediation Technologies. The selection
of a remedial strategy at an MTBE-impacted site depends on the volume of con-
taminated soil and groundwater, contaminant concentrations and mass, site-spe-
cific hydrogeological characteristics, and risk-based or regulatory-driven cleanup
objectives. Of particular interest at MTBE-impacted sites is the behavior of
MTBE relative to other gasoline components in subsurface environments.
The properties of MTBE greatly influence its fate in the environment, gov-
ern transport mechanisms, phase partitioning and biological interactions, and
impact the success of remedial technologies. Key physiochemical properties of
MTBE were presented in Tables 2 and 3. As compared with toluene, a common
gasoline component and groundwater contaminant (! Toluene, Chap. 2.1.),
MTBE has a lower density and boiling point, but a higher vapor pressure and
much higher water solubility. The properties of MTBE suggest the following
regarding its fate in subsurface environments: (1) MTBE does not readily sorb
to soil particles in aquifers; (2) MTBE volatilizes readily from non-aqueous
phase liquid (NAPL) sources; and (3) MTBE strongly partitions into the aqueous
phase and preferentially remains in water. As a consequence, MTBE can be
highly mobile in subsurface environments relative to other gasoline constituents.
Contarary to early reports, most of the technologies used historically at
gasoline-contaminated sites are effective at removing MTBE from groundwater.
Successful technologies include groundwater extraction followed by air
stripping, advanced oxidation or granular activated carbon systems. In-situ tech-
nology options include air sparging, multi-phase extraction, soil vapor extrac-
tion, bioremediation and phytoremediation. In addition to single technologies
traditionally used at gasoline-contaminated sites, emerging technologies, modifi-
cations to existing technologies or the use of a treatment train involving more
than one technology can greatly improve MTBE removal efficiency and/or reduce
life cycle remediation costs.
13.2. Phase Transfer Technologies. Phase transfer technologies
move MTBE from the water phase to either the gas phase (eg, air stripping) or
the solid phase (eg, granular activated carbon) without chemically transforming
it. Consequently, additional treatment steps to destroy MTBE (eg, off-gas treat-
ment from air strippers) or waste disposal (eg, spent granular activated carbon)
may be required.
Ex-Situ Phase Transfer Technologies. Air Stripping. Air stripping
is a physical process for removing volatile organic compounds from water. In
general, the effectiveness of air stripping increases with a compound’s Henry’s
442 METHYL TERT-BUTYL ETHER Vol. 2
constant and the operating air-to-water ratio. MTBE’s Henry’s constant is much
lower than that of common groundwater contaminants such as toluene (0.018–
0.026 vs. 0.27, (66)). Thus, high air-to-water ratios (100 to 250) are needed to
remove MTBE from water. Air stripping has been used to successfully remove
MTBE at concentrations that are typically associated with groundwater contam-
ination from LUFTs (67). Evaluation of MTBE performance data from eight case
studies of low profile and packed tower air strippers showed that air stripping
has the potential to be more widely used in both drinking water and remediation
applications (68). For each of the studies evaluated, MTBE removal efficiencies
exceeded 90%. Because state and/or local air quality regulations sometimes
required stripper off-gas treatment (granular activated carbon, thermal and cat-
alytic oxidation, advanced oxidation), process costs can increase significantly.
Granular Activated Carbon (GAC). Although MTBE has a low affinity
for sorption to solids, several types of activated carbon including coconut shell or
coal-based GAC can be used to remove low concentrations of MTBE in water
(200 mg/L) cost-effectively (67,69). GAC can be used to treat low flows of water
from private wells with low MTBE concentrations. For treatment of higher con-
centrations of MTBE, GAC can be used as a polishing step in a treatment train
following air stripping or chemical oxidation (69). Background water quality and
co-contaminant concentrations influence MTBE removal efficiencies. High con-
centrations of natural organic matter (NOM) and other gasoline constituents
in influent water streams compete with MTBE for GAC sorption sites thereby
increasing GAC usage rates and process costs.
Other Ex-Situ Phase Transfer Technologies. Results from laboratory
and pilot scale studies suggest that synthetic resins have the potential for suc-
cessful remediation of MTBE-impacted sites (67). One of the advantages
of using synthetic resins is the ease of regeneration relative to GAC. One of
the limitations of resins is the experimentally observed reduction in MTBE sorp-
tion due to the presence of other gasoline constituents, such as benzene, toluene,
ethylbenzene, and xylene (BTEX) compounds and tert-butanol (TBA, tert-
butylalcohol), in contaminated water. This is analogous to the effects of NOM
and BTEX on GAC sorption as discussed above. The use of synthetic resins is
costly relative to GAC but resins can be designed to achieve a higher degree of
selectivity and trend to have a longer lifetime than GAC.
In addition to synthetic resins, other sorbents such as high silica zeolites
have been shown to be effective in removing MTBE from contaminated water (70).
Other technologies, such as membranes and solvent extraction, can be used
to remove MTBE from contaminated groundwater (71,72,73) but only a few
bench- and pilot-scale studies have been conducted.
In-Situ Phase Transfer Technologies. Demonstrated in-situ phase
transfer technologies include soil vapor extraction (SVE), multi-phase extraction
(MPE), in-situ air sparging (IAS) and phytroremediation. Emerging phase
transfer technologies include steam injection/vacuum extraction and six-phase
heating.
Soil Vapor Extraction (SVE) and Multi-Phase Extraction (MPE).
Soil vapor extraction (SVE) is commonly used for the remediation of vadose
zones at gasoline-contaminated sites. In order to remove MTBE from both
saturated and unsaturated zones, other remediation technologies can be used
Vol. 2 METHYL TERT-BUTYL ETHER 443
in combination with SVE. The success of SVE has been documented at MTBE-
impacted sites when implemented either as the sole technology or in conjunction
with other groundwater treatment technologies (74–76) The combination of SVE
and groundwater extraction, known as multi-phase extraction, is effective in
removing MTBE from contaminated aquifers (75) and is routinely used at
many sites. Effectiveness with MTBE can be enhanced if the remediation system
is operated under high vacuum.
In-Situ Air Sparging (IAS). In-situ air sparging involves the injection
of air below the groundwater table to transfer contaminants from the aqueous
phase into the gaseous phase. The injection of air has the additional advantage
of introducing oxygen to subsurface environments thereby potentially enhancing
aerobic contaminant biodegradation rates. Laboratory air sparging column stu-
dies in different soils showed that the rate of MTBE volatilization is very rapid in
sandy soils (77). Results from another laboratory study showed that 85% of
MTBE in water equilibrated with gasoline was successfully removed using air
sparging (78). MTBE recovery rates were shown to increase with increased air
injection rates and water saturation levels. Results from a number of field stu-
dies revealed that MTBE groundwater concentrations at gasoline-contaminated
sites were reduced by 88 to 99% after two years of air sparging (79). As a general
rule, if the site hydrogeology is conducive to successful use of IAS, e.g., soil por-
osity is satisfactory, this technology could be very successful at removing MTBE
from groundwater.
Phytoremediation. Phytoremediation involves the use of plants or trees
to either remove or degrade groundwater contaminants. Phytoremediation
occurs by plant uptake of the contaminants followed by translocation and accu-
mulation into plant tissues, or transpiration through leaves. In some cases, the
breakdown of contaminants can take place either within plant tissues or in the
rhizosphere (80). Phytoremediation is most effective for contaminants with log
Kow values in the range of 0.5 to 3.0 (81). Since MTBE has a log Kow of 1.2, phy-
toremediation is expected to be an effective technology at MTBE-impacted sites.
Several plant and tree species have been recently reported to uptake MTBE
including alfalfa plants (82), poplar trees (80,83,84) and Eucalyptus trees
(84,85). In laboratory studies, poplar trees were shown to remove 30% of the
MTBE in test solutions at concentrations of 300 to 1600 mg/L in a week (81).
Results of several laboratory studies suggest that phytoremediation has the
potential to be effective for the removal of MTBE from groundwater, especially
at sites where the contaminated groundwater is shallow.
13.3. Transformation or Destruction Technologies. The underlying
principle of transformation or destruction technologies is the transformation of
MTBE in water to another chemical form either partially (eg, acetone) or comple-
tely (carbon dioxide). Transformation or destruction technologies can either be
implemented ex-situ following groundwater extraction (eg, chemical oxidation,
biological treatment) or in-situ (eg, in-situ chemical oxidation, in-situ
bioremediation).
Ex-Situ Transformation Technologies. Advanced Oxidation Processes
(AOPs). Advanced oxidation processes involve the use of ozone-peroxide,
ozone – UV, peroxide – UV, Fenton’s reaction (hydrogen peroxide in com-
bination with iron(II)salts), high-energy electron beam irradiation (E-beam)
444 METHYL TERT-BUTYL ETHER Vol. 2
REFERENCES
17. E. Trubac: ‘‘Oxygenate Removal from MTBE Raffinate’’, lecture presented at the
AIChE Meeting, Minneapolis, Minn., Aug. 16–19, 1987.
18. Union Carbide Corp., EP-A 0 229 994, 1987 (M. Nagji).
19. Hüls AG, DE 2 629 769, 1976 (F. Obenaus et al.).
20. W. Schneider, J.-Ch. Frohne, H. Bruderreck, J. Chromatogr. 245 (1982) 71–83.
21. J. Winskowski, Erdöl Kohle Erdgas Petrochem. 41 (1988) no. 1, 30–36.
22. J. M. Levy, J. A. Yancey, J. High Resolut. Chromatogr. Commun. HRC CC J. High
Resolut. Chromatogr. 9 (1986) no. 7, 383–387.
23. AFNOR (Association Française de Normalisation) M 07–084, Dec. 1984, Norme
experimentale.
24. DIN 51 413, Parts 4 and 5.
25. BSI (British Standards Institution) proposed method JP-STG-6 C (The Institute of
Petroleum).
26. EEC Council Directive 85/536/EEC, Dec. 5, 1985 (Column 3).
27. Chem. Eng. News, Oct. 24, (1988) 12.
28. G. H. Unzelmann, Oil Gas J. (1989, April 10) 33–37.
29. F. Obenaus, W. Droste, Erdöl Kohle Erdgas Petrochem. 33 (1980) no. 6, 271–275.
30. M. Denham, M. Mays, K. Schug, M. Sposini, lecture presented at the 8th Interna-
tional Symposium on Alcohol Fuels, Nov. 13–16, Tokyo 1988.
31. W. J. Piel, Energ. Prog. 8 (1988) no. 4, 201–204.
32. Fattore et al., Hydrocarbon Process. 60 (1981) Aug., 101. August (1981) 101.
33. Chem. Eng. N.Y. 90 (1983) no. 25, 60.
34. H. Kröper, K. Schlömer, H. M. Weitz, Hydrocarbon Process 48 (1969) no. 9, 195–198.
35. Nippon Shokubai Kagaku Kogyo Co., EP-A 0 304 867, 1988 (K. Kinumi et al.).
36. Kuraray Co, JP-Kokai 60 193 932 <85 193 932>, 1985 (Y. Ninagawa et al.). Sumi-
tomo Chem., US 3 574 780, 1971 (Y. Watanabe et al.).
37. Degussa AG, EP 0 050 869, 1981 (B. Lehmann).
38. Anic S.p.A., DE-OS 3 130 428, 1982 (F. Ancillotti et al.). BASF AG, DE-OS
27 40 590, 1979 (F. Merger et al.).
39. BASF AG, DE-OS 29 11 466, 1980 (H. Naarmann et al.).
40. BASF AG, DE-OS 29 28 098, 1981 (W. Hoffmann).
41. Bayer AG, DE-OS 30 02 203, 1981 (G. Bonse et al.).
42. H. Grangette, C. Bernasconi, S. Randriamahefa, R. Gallo, Inf. Chim. 246 (1984)
Jan./Feb., 135–140.
43. Exxon Research and Engineering Co., EP-A 00 88 603, 1983 (T. H. West).
44. Chem. Week (1997) Dec. 3, 60.
45. M. L. Morgan, DGMK Conference ‘‘C4 Chemistry–Manufacture and Use of
C4-Hydrocarbons’’, Oct 6–8 (1997) Aachen.
46. K. D. Miller, 23rd Annual Dewitt Petrochemical Review, March 31–April 2 (1998)
Houston.
47. P. R. Sarathy, G. S. Suffride, Hydrocarbon Process. (1993) Feb, 43–51.
48. Chem. Week (1997) Oct. 29, 44.
49. P. Chu, G. H. Kühl, Ind. Eng. Chem. Res. 26 (1987) 366.
50. A. Kogelbauer et al., J. Catal. 152 (1995) 122–129.
51. K. Sundmacher, U. Hoffmann, Chem. Eng. Sci. 49 (18) (1994) 3077–3089.
52. T. Zhang, R. Datta, Ind. Eng. Chem. Res. 34 (1995) 730–740.
53. D. Parra et al., Chem. Eng. Sci. 49 (24A) (1994) 4563–4578.
54. A. Rehfinger, U. Hoffmann, Chem. Eng. Sci. 45 (1990) no. 6, 1605–1617.
55. F. Obenaus, Oil Gas J. 77 (1979) 76–77.
56. A. Rehfinger, U. Hoffmann, Chem. Eng. Technol. 13 (1990) 150–156.
57. S. Rossini et al., DGMK-Conference ‘‘C4 Chemistry–Manufacture and Use of
C4-Hydrocarbons’’, Oct. 6–8 (1997), Aachen.
Vol. 2 METHYL TERT-BUTYL ETHER 447
UDO PETERS
FRANZ NIERLICH
EKKEHARD SCHULTE-KÖRNE
Oxeno Olefinchemie GmbH, Marl,
Federal Republic of Germany
MICHAEL SAKUTH
Hüls Infracor GmbH, Marl,
Federal Republic of Germany
RULA DEEB
MARYLINE LAUGIER
MICHAEL KAVANAUGH
Malcolm Pirnie, Inc., Emeryville, USA
MARTTI SUOMINEN
Fortum Oil and Gas Oy, Finland
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 449
COMBUSTION SCIENCE
AND TECHNOLOGY
1. Introduction
2. Combustion Science
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0315130219080118.a01
450 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
Vapor pressure
curve Nonflammable
Φ
Rich
Φ=1 Explosion
Mist combustion Flammable
possible
Lean
Nonflammable
TL TU AIT
Temperature
Fig. 1. Effect of temperature on limits of flammability of a pure liquid fuel in air, where
TL ¼ lean (or lower) flash point; TU ¼ rich (or upper) flash point; and AIT ¼ auto ignition
temperature (4,5).
mf
m
f ¼ o ð1Þ
mf
mo stoich
If f < 1, the mixture is said to be lean (in fuel) and the products of combustion
contain unreacted or excess O2. If f > 1 the mixture is said to be rich (in fuel) and
the products of combustion contain CO and possibly H2 because of incomplete
combustion caused by the oxygen deficiency (1–3).
Flammability Limits. Any given mixture of fuel and oxidant is flam-
mable (explosive) within two limits referred to as the upper (rich) and lower
(lean) limits of flammability (Fig. 1). Most mixtures are flammable when the
fuel to air volume ratio lies between 50 and 300% of the stoichiometric ratio
(2). The stoichiometric ratio is the exact theoretical ratio of fuel to air required
for complete combustion. The flammability region widens with increasing
temperature, and usually with increasing pressure, although the effect of
pressure is less predictable. The region within the flammability limits of a
fuel – air mixture can be divided into two subregions, the slow oxidation region
and the explosion region, separated by the spontaneous ignition tempe-
rature. To determine this temperature for liquid fuels, standard tests are
used in which the liquid fuel is dropped into an open-air container heated to
a known temperature. The lowest temperature at which visible or audible evi-
dence of combustion is observed is defined as the spontaneous ignition
temperature (1).
Flash Point. As fuel oil is heated, vapors are produced which at a certain
temperature ‘‘flash’’ when ignited by an external ignition source. The flash point
is the lowest temperature at which vapor, given off from a liquid, is in sufficient
quantity to enable ignition to take place. The flash point is in effect a measure
of the volatility of the fuel. The measurement of flash point for pure liquids is
relatively straightforward. However, the measured value may depend slightly
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 451
on the method used, especially for liquid mixtures, since the composition of the
vapor evolved can vary with the heating rate. Special problems arise when
the liquid contains a mixture of fuels and/or fuels and inhibitors. In these
cases the vapor composition will be different from the liquid composition. Conse-
quently, successive tests with the same sample can lead to erroneous results
because the composition of the liquid in the sample holder changes with time
as a result of fractional distillation (4–7).
Ignition. To understand the phenomenon of ignition it is necessary to con-
sider the following concepts: ignition source, gas temperature, flame volume, and
presence of quench wall surfaces. In general, there are two main methods of
igniting a flammable mixture. In the self-ignition method, the mixture is heated
slowly so that the vapor released as the temperature is raised ignites sponta-
neously at a particular temperature. In the forced ignition method, a small quan-
tity of combustible mixture is heated by an external source and the heat released
during the combustion of this portion results in propagation of a flame. The
external ignition source can be an electric spark, pilot flame, shock wave, etc.
For ignition to take place the following conditions should be satisfied: (1) the
amount of energy supplied by the ignition source should be large enough to over-
come the activation energy barrier; (2) the energy released in the gas volume
should exceed the minimum critical energy for ignition; and (3) the duration of
the spark or other ignition source should be long enough to initiate flame propa-
gation, but not too long to affect the rate of propagation. Ignition models fall into
two categories: the thermal model explains the ignition as resulting from supply-
ing the mixture with the amount of heat sufficient to initiate reaction. In the che-
mical diffusion model the main role of the ignition source is attributed to the
formation of a large number of free radicals in the preheat zone, where their dif-
fusion to the surrounding region initiates the combustion process. The thermal
model is applied more widely in the literature and shows better agreement with
experimental data.
At ambient temperatures, reaction rates for gaseous mixtures of fuel and
oxidant are extremely slow. As temperature is increased gradually, slow oxida-
tion begins and as a result of the exothermic reactions that occur, the tempera-
ture keeps increasing. With further temperature increase, the reaction rate
suddenly increases causing rapid combustion reactions to occur. This is provid-
ing, of course, that the rate of heat release is greater than the rate of heat loss
through the container walls. The temperature at which the heat released by the
reaction exceeds the heat loss is commonly referred to as the ignition tempera-
ture. The spontaneous ignition temperature (SIT), on the other hand, is the low-
est temperature at which ignition occurs. Increasing the pressure results in a
decrease in the spontaneous ignition temperature (1) (Fig. 2 and Table 1). This
temperature is highly dependent on the material of construction, apparatus con-
figuration, and test procedure, therefore reported test values vary widely (2,6,7).
From the chemical diffusion model standpoint, the usual values of ignition
temperatures of mixtures often do not correlate well with other flame properties.
Nevertheless, the following highly simplified, even speculative, qualitative des-
cription may be useful in thinking about flames. Early in the ignition or induc-
tion period, the attack of fuel by O2 is initially slow, but generates free radicals
452 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
Cool flames
Ignition delay
10 3 10 2 10 1 10 –1
Pressure
T1 T2 T3
Explosion
Slow oxidation
P3
P2
Long delay
explosion
P1
Temperature,°C
Fig. 2. Slow oxidation, spontaneous ignition, and explosion as a function of pressure and
temperature variations in hydrocarbon mixtures (1).
(OH , H, O, HO2, hydrocarbons) and other intermediate species (CO, H2, and
partial oxidation and decomposition products of hydrocarbons, if present). For
some time, there may be little or no temperature rise, the energy essentially
being stored in the free radicals. In this stage the reactions may be similar to
those that occur in very slow and nearly isothermal oxidation and cool flames
observed in very rich hydrocarbon–O2 mixtures, usually at lower temperatures
and pressures. These reactions often stop at the production of stable oxidation
and decomposition products, such as aldehydes, peroxides, lower hydrocarbons,
etc, but under some conditions may lead to explosions. In a variety of chain
reactions, the fuel and any intermediates are rapidly attacked by radicals,
some of which are also undergoing very fast chain-branching reactions. Evi-
dence for the behavior of such free radicals is provided by studies of the
mechanism of H2/O2 and similar explosions. These free radicals can multiply
rapidly and attain high transient concentrations, initiating chain reactions
that can eventually lead to the explosion of the flammable mixture. They are
eventually consumed to form stable species by three-body recombination reac-
tions in which most of the heat release occurs. An example of such a reaction,
where M is any atom or molecule in the gas, is represented by
The electrical energy is rapidly transformed into thermal energy, and because
the temperature of the ionized gas is generally above 300 K, the ignition
delay time is short compared with the spark duration, . Ignition only takes
place if the electrical energy exceeds the critical value, Ec, and if this energy is
liberated within a critical volume, vc (1,8,10,11).
Ignition can also be produced by a heated surface. During the process of
heat transfer from a hot surface to a flammable mixture, reactions are initiated
as the temperature rises and the combination of additional heat transfer from
the surface and heat release by chemical reactions can lead to ignition of the
mixture.
In experiments in which a heated cylinder was introduced into a propane–
air mixture, it was found that approximately 11 ms were required for ignition.
During this time, heat released by chemical reactions is negligible. This time
interval before ignition is referred to as the ignition delay or induction time
and is a function of the physical properties of the fuel, the rate of heat conduc-
tion, and the pressure of the combustion chamber (1,3,8,10,11). Ignition is char-
acterized by a rapid change in temperature (eq. 3), where tignition ¼ ignition delay
time, s; Ea ¼ activation energy, J/mol; R ¼ universal gas constant, J/(K mol);
and Ti ¼ ignition temperature, K.
Ea
tignition a exp ð3Þ
RTi
For liquid fuels, ignition delay times are of the order 50 ms at 700 K and 10 ms
at 800 K. At low temperatures most of the ignition delay is the result of slow,
free-radical reactions, and a distinction between the initiation and explosion
periods within the ignition delay time can be made. With increasing ignition
temperature for a given mixture, these times become comparable and at
temperatures as high as 1500 K, both times may be of the order of 104 s. Con-
sequently, the reaction zone in the flame of a mixture is observed to be one
continuous event (12–14).
Another important concept is that of the critical ignition volume. During
the propagation of the combustion wave, the flame volume cannot conti-
nually grow beyond a critical value without an additional supply of energy.
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 455
The condition that controls the critical volume for ignition is reached when the
rate of increase of flame volume is less than the rate of increase of volume of the
combustion products. In this condition a positive exchange of heat between the
flame and the fresh mixture is achieved.
For a point spark source, the flame volume is initially spherical and the cri-
tical ignition volume is determined by calculating the rate of change of flame
volume with respect to radius compared to the rate of change of volume of the
combustion products (eq. 4)
d 4 d 4
p ðr þ eÞ3 r3 p r3 ð4Þ
dr 3 dr 3
2 er þ e2 r2 ð5Þ
pffiffiffi
rc ¼ e þ 2 ð6Þ
For a line spark source, the flame volume is initially cylindrical with the
cylinder length equal to the separation distance between the electrodes. Thus,
for a cylindrical flame, rc ¼ e, and the critical ignition volumes are equation 7
for a spherical flame and equation 8 for a cylindrical flame where vc ¼ critical
ignition volume, m3/kg; e ¼ thickness of flame front, m; and d ¼ flame height, m.
pffiffiffi3
4pe3 1 þ 2
c ¼ ð7Þ
3
vc ¼ pe2 d ð8Þ
Because of heat being transferred to the exterior of the critical volume, only
a fraction of the total spark energy, E, is utilized as ignition energy.
Minimum ignition energies are determined experimentally by means of a
combustion bomb with two electrodes. Both the ignition energy and the quen-
ching distance can be varied in this device. By using high performance (usually
456 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
air-gap) condensers, the majority of the stored energy will appear in the spark
gap. The stored energy can be calculated using equation 10 (1,2,9), where
E ¼ total spark energy released/stored, J; C ¼ capacity of the condenser; and
V ¼ voltage just before the spark is passed through the gas, V.
CV 2
E¼ ð10Þ
2
3.5
3
Quenching distance, mm
2.5
1.5
0.5 1 1.5 2
Fuel/air equivalence ratio
temperature of the mixture outside the cool flame can produce normal sponta-
neous ignition (1).
Flame Temperature. The adiabatic flame temperature, or theoretical
flame temperature, is the maximum temperature attained by the products
when the reaction goes to completion and the heat liberated during the reaction
is used to raise the temperature of the products. Flame temperatures, as a func-
tion of the equivalence ratio, are usually calculated from thermodynamic data
when a fuel is burned adiabatically with air. To calculate the adiabatic flame
temperature (AFT) without dissociation, for lean to stoichiometric mixtures,
complete combustion is assumed. This implies that the products of combustion
contain only carbon dioxide , water, nitrogen, oxygen, and sulfur dioxide.
Actual temperatures in practical flames are lower than calculated values as
a result of the heat losses by radiation, thermal conduction, and diffusion. At
high temperatures, dissociation of products of combustion into species such as
OH, O, and H reduces the theoretical flame temperature (7). Increasing the pres-
sure tends to suppress dissociation of the products and thus generally raises the
adiabatic flame temperature (4).
Adiabatic flame temperatures agree with values measured by optical tech-
niques, when the combustion is essentially complete and when losses are known
to be relatively small. Calculated temperatures and gas compositions are thus
extremely useful and essential for assessing the combustion process and predict-
ing the effects of variations in process parameters (4). Advances in computational
techniques have made flame temperature and equilibrium gas composition calcu-
lations, and the prediction of thermodynamic properties, routine for any fuel-
oxidizer system for which the enthalpies and heats of formation are available
or can be estimated.
Flame Types and Their Characteristics. There are two main types of
flames: diffusion and premixed. In diffusion flames, the fuel and oxidant are
separately introduced and the rate of the overall process is determined by the
mixing rate. Examples of diffusion flames include the flames associated with can-
dles, matches, gaseous fuel jets, oil sprays, and large fires, whether accidental or
otherwise. In premixed flames, fuel and oxidant are mixed thoroughly prior to
combustion. A fundamental understanding of both flame types and their struc-
ture involves the determination of the dimensions of the various zones in the
flame and the temperature, velocity, and species concentrations throughout
the system.
The development of combustion theory has led to the appearance of several
specialized asymptotic concepts and mathematical methods. An extremely strong
temperature dependence for the reaction rate is typical of the theory. This makes
direct numerical solution of the equations difficult but at the same time accurate.
The basic concept of combustion theory, the idea of a flame moving at a constant
velocity independent of the ignition conditions and determined solely by the
properties and state of the fuel mixture, is the product of the asymptotic
approach (18,19). Theoretical understanding of turbulent combustion involves
combining the theory of turbulence and the kinetics of chemical reactions (19–23).
Laminar Premixed Flames. The structure of a one-dimensional premixed
flame is well understood (1). By coupling the rate of heat release from chemical
reaction and the rate of heat transfer by conduction with the flow of the
458 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
unburned mixture, an observer moving with the wave would see a steady lami-
nar flow of unburned gas at a uniform velocity, Su, into the stationary wave or
flame. Hence Su is defined as the burning velocity of the mixture based on the
conditions of the unburned gas. The thickness of the preheating zone and the
equivalent reaction zone is found to be inversely proportional to the burning velo-
city. By considering the heat release from the chemical reaction, it is possible to
calculate the thickness of the effective reaction zone. For example, at atmo-
spheric pressure, for hydrocarbon flames, the thicknesses of the preheat and
reaction zones are found to be 0.7 and 0.2 mm, respectively (1–3). A propagating,
premixed flame is in essence a thin wave in which the temperature rise is con-
tinuous and rapid. The thicknesses of the wave depend on f and on pressure; for
a typical hydrocarbon–air mixture at f ¼ 1 at 101.3 kPa (1 atm), the entire rise,
from the initial temperature to the adiabatic temperature, occurs in 0.01 mm,
and the concentration and temperature gradients are accordingly very steep.
Su has been approximated for flames stabilized by a steady uniform flow of
unburned gas from porous metal diaphragms or other flow straighteners. How-
ever, in practice, Su is usually determined less directly from the speed and area
of transient flames in tubes, closed vessels, soap bubbles blown with the mixture,
and, most commonly, from the shape of steady Bunsen burner flames. The
observed speed of a transient flame usually differs markedly from Su. For exam-
ple, it can be calculated that a flame spreads from a central ignition point in an
unconfined explosive mixture such as a soap bubble at a speed of (ro/rb)Su,
in which the density ratio across the flame is typically 5–10. Usually, the
expansion of the burning gas imparts a considerable velocity to the unburned
mixture, and the observed speed will be the sum of this velocity and Su.
By applying the conservation equations of mass and energy and by neglect-
ing the small pressure changes across the flame, the thickness of the preheating
and reaction zones can be calculated for a one-dimensional flame (1).
There are a number of sources of instability in premixed combustion
systems (23,24).
rud
RE ¼ ð11Þ
m
V
h¼ ð12Þ
2pD
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 461
a a
f a f a a f a a f a a f a
a a
f a a f a a f a
a
a a
Top views
Fig. 4. Various configurations (a–e) used to obtain gaseous diffusion flames where
a ¼ air and f ¼ fuel (33).
In addition to the Burke and Schumann model (34) and the Displacement
Distance theory, a comprehensive laminar diffusion flame theory can be written
using the equations of conservation of species, energy, and momentum, including
diffusion, heat transfer, and chemical reaction.
Evaporation and burning of liquid droplets are of particular interest in
furnace and propulsion applications and by applying a part of the Burke and
Schumann approach it is possible to obtain a simple model for diffusion flames.
Combustion chemistry in diffusion flames is not as simple as is assumed in
most theoretical models. Evidence obtained by adsorption and emission spectro-
scopy (37) and by sampling (38) shows that hydrocarbon fuels undergo appreci-
able pyrolysis in the fuel jet before oxidation occurs. Further evidence for the
existence of pyrolysis is provided by sampling of diffusion flames (39). In general,
the preflame pyrolysis reactions may not be very important in terms of the gross
features of the flame, particularly flame height, but they may account for the for-
mation of carbon while the presence of OH radicals may provide a path for NOx
formation, particularly on the oxidant side of the flame (39).
Combustion chemistry in diffusion flames also accounts for the smoke
formation process. Characteristic behavior of smoking diffusion flames includes:
(1) the first appearance of smoke is generally at the flame tip, the width of the
smoke trail increasing as the amount of smoke increases; (2) the appearance and
quantity of smoke increases with increasing flame height; (3) the smoking ten-
dency changes with oxidant flow rate and the oxygen content of the oxidant
stream; (4) the smoking tendency is a function of fuel type; and (5) the smoking
tendency generally increases with increasing pressure (39).
Diffusion Flames in the Transition Region. As the velocity of the fuel jet
increases in the laminar to turbulent transition region, an instability develops at
the top of the flame and spreads down to its base. This is caused by the shear
forces at the boundaries of the fuel jet. The flame length in the transition region
462 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
is usually calculated by means of empirical formulas of the form (eq. 13): where
l ¼ length of the flame, m; r ¼ radius of the fuel jet, m; v ¼ fuel flow velocity, m/s;
and C1 and C2 are empirical constants.
r
l¼ ð13Þ
C2
C
Envelope of
flame length
Flame height, hF
Envelope of
break-point
0
Increasing nozzle velocity
(a)
60 24
50 20
Total height, on-part flames
Distance from nozzle, cm
40 16
Flame separation begins
in.
30 Total height, 12
off-part flames
20 8
Break-point height
10 Bottom 4
of off-part
flames
(b)
Fig. 5. Effects of nozzle velocity on flame appearance in laminar and turbulent flow:
(a), flame appearance; (b), flame height and break-point height (40).
repeated until the ignition of the sample is observed. It has been a matter of con-
troversy as to whether the value of the critical temperature determined using
this method for these low temperature tests represents the ignition tendency
or the combustibility of the coal during the high heating rates encountered in
a pulverized coal flame.
The structure of residual char particles after devolatilization depends
on the nature of the coal and the pyrolysis conditions such as heating rate,
peak temperature, soak time at the peak temperature, gaseous environment,
and the pressure of the system (72). The oxidation rate of the char is primarily
influenced by the physical and chemical nature of the char, the rate of diffusion
and the nature of the reactant and product gases, and the temperature and pres-
sure of the operating system. The physical and chemical characteristics that
influence the rate of oxidation are chemical structural variations, such as the
concentration of oxygen and hydrogen atoms (73–75), the nature and amount
of mineral matter (76–79), and physical characteristics such as porosity, particle
size, and accessible surface area. The rate of diffusion of the reactant gas is
governed by the temperature and pressure of the operating system.
The rate limiting step in the combustion of char is either the chemical
kinetics (adsorption of oxygen, reaction, and desorption of products) or diffusion
of oxygen (bulk and pore diffusion). Variations in the reaction rate with tempera-
ture for gas–carbon reactions have been grouped into three main regions or
zones depending on the rate limiting resistance (80) (Figs. 6 and 7).
The overall reaction rate is determined by equation 17,
1
q¼ pg ð17Þ
1 1
þ
kdiff ks
where q ¼ rate of removal of carbon atoms per unit external surface area, kg/
m2s; pg ¼ partial pressure of oxygen in the free stream, Pa; kdiff ¼ diffusional
Fig. 6. The three ideal zones (I–III) representing the rate of change of reaction for a por-
ous carbon with increasing temperature where a and b are intermediate zones, Ea is acti-
vation energy, and Et is true activation energy. The effectiveness factor, Z, is a ratio of
experimental reaction rate to reaction rate which would be found if the gas concentration
were equal to the atmospheric gas concentration (80).
468 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
Boundary layer
Zone 3
Zone 2
0 R
Distance from the center of
the solid reactant
Fig. 7. The concentration of the reacting fluid as a function of the distance from the
center of the solid reactant in various reaction zones.
reaction rate coefficient, kg/(m2 Pa); and ks ¼ surface reaction rate coefficient,
kg/(m2 Pa).
In Zone 1, the low temperature zone, the reaction rate is slow and the
concentration of the reactant gas is uniform throughout the interior of the solid.
The overall rate is controlled by the chemical reactivity of the solid, the kinetics
are not influenced by mass transfer, and the kinetics are the intrinsic kinetics.
The measured activation energy is equal to the true activation energy. In a
higher temperature zone, Zone 2, the concentration of the gaseous reactant
becomes zero at a point somewhere between the surface and the center of the
particle, and the reaction rate is controlled by both the chemical kinetics and
the diffusion rate of the gaseous reactant. The measured activation energy in
this zone is one-half of the true activation energy. At still higher temperatures,
in Zone 3, the chemical reactivity is so high that oxygen is consumed as soon as it
reaches the surface of the solid and the concentration of the gaseous reactant
approaches zero at the surface, indicating that the reaction rate is controlled
by bulk diffusion of the reacting gas. The measured activation energy approaches
a value of zero in this zone.
Burning times for coal particles are obtained from integrated reaction rates.
For larger particles (>100 mm) and at practical combustion temperatures, there
is a good correlation between theory and experiment for char burnout. Experi-
mental data are found to obey the Nusselt ‘‘square law’’ which states that the
burning time varies with the square of the initial particle diameter (tc D2:0 o ).
However, for particle sizes smaller than 100 mm, the Nusselt relationship
seems to predict higher burning times than those observed and it has been
noted that burning times are proportional to the initial diameter of the particles
raised to the power of 1.4 (tc D1:4
o ) (54).
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 469
3. Combustion Technology
V_ mix ¼ Af Su ð18Þ
where V_ mix ¼ volumetric flow rate, m3/s; Af ¼ area of the steady flame, m2; and
Su ¼ burning velocity, m/s.
The volumetric flow rate of the mixture is, in turn, proportional to the rate
of heat input (eq. 19):
where V_ mix ¼ volumetric flow rate, m3/s; HHV ¼ higher heating value of the
fuel, J/kg; and Q_ ¼ rate of heat input, J/s.
In the simple Bunsen flame on a tube of circular cross-section, the stabiliza-
tion depends on the velocity variation in the flow emerging from the tube. For
laminar flow (parabolic velocity profile) in a tube, the velocity at a radius r is
given by equation 20:
v ¼ const R2 r2 ð20Þ
The maximum velocity at the axis is twice the average, whereas the velocity
at the wall is zero. The effect of the burner wall is to cool the flame locally and
decrease the burning velocity of the mixture. This results in flame stabilization.
However, if the heat-transfer processes (conduction, convection, and radiation)
involved in cooling the flame are somehow impeded, the rate of heat loss is
decreased and the local reduction in burning velocity may no longer take
place. This could result in upstream propagation of the flame.
To make the flame stable against the flow in a thin annular region near the
rim, the flow velocity vr should be made equal to the burning velocity at some
radius r. This annulus serves as a pilot and ignites the main flow of the mixture,
ie, the flame gradually spreads toward the center. In most of the mixture flow,
vr > Su which results in a stable flame. With increasing mixture flow, the height
and area of the flame increase. Measurement of the area of a stable Bunsen flame
is the basis for the method most commonly used to determine Su (81–83).
By feeding the mixture through a converging nozzle, the velocity profile
may be made nearly flat or uniform. A Bunsen flame in such a flow has a smaller
range of stability but the mechanism is essentially the same and the flame very
closely approximates a cone. If the apex angle of the flame is , then Su can be
obtained from equation 21
Su ¼ r sin ð21Þ
2
where Su ¼ burning velocity, m/s; vr ¼ mixture velocity at the nozzle exit, m/s;
and ¼ the apex angle of the flame in degrees.
If the tube diameter is appreciably larger than the quenching distance, Su
will exceed vr in some parts of the flowing mixture due to a lack of quenching,
and the flame will then propagate down the tube as far as there is mixture to
consume. This undesirable condition is referred to as flashback. If, on the
other hand, vr exceeds Su in the mixture flow, the flame lifts from the port and
blows off. This condition is referred to as blowoff and like flashback should be
avoided (Fig. 8). The velocity gradient at the wall, gw, is defined as
dv
gw ¼
dr r¼R
3200
A
3000
B
2600
C
2400
2200
Fig. 8. Flashback and blowoff critical velocity gradients for a natural gas flame (10),
where A is N2 =O2 ¼ 2, p ¼ 1:01 MPa (10 atm); B, N2 =O2 ¼ 2, p ¼ 101:3 kPa (1 atm); C,
air, p ¼ 1:01 MPa; and D, air, p ¼ 101:3 kPa.
gas, and of the shape or height of the flame cone. Often this must be done pre-
cisely, and with no changes in orifices or adjustments in the feed system.
If the substitute fuel is of the same general type, eg, propane for methane,
the problem reduces to control of the primary equivalence ratio. For nonaspiring
burners, ie, those in which the air and fuel supplies are essentially independent,
it is further reduced to control of the fuel flow, since the air flow usually consti-
tutes most of the mass flow and this is fixed. For a given fuel supply pressure and
fixed flow resistance of the feed system, the volume flow rate of the fuel is inver-
pffiffiffiffiffi
sely proportional to rf . The same total heat input rate or enthalpy flow to the
flame simply requires satisfactory reproduction of the product of the lower heat-
pffiffiffiffiffi
ing value of the fuel and its flow rate, so that WI Qp = rf remains the same. WI
is the Wobbe Index of the fuel gas, and is a commonly used criterion for inter-
changeability in adjusting the composition of a substitute fuel. The units of WI
are variously given, but, if used consistently, are unimportant since only ratios of
the Wobbe Indices are ordinarily of interest. Sometimes rf is taken as the specific
gravity relative to some reference gas, eg, air, or average molecular weights may
be used.
The Wobbe Index criterion also applies to substitution with aspirating or
atmospheric pressure burners in which the volume flow of primary air induced
pffiffiffiffiffi
by momentum exchange with the fuel increases with rf . Because the volu-
metric air requirement for a given f is nearly proportional to the heating
value of the fuel, an adjustment of Qp and ff to the same WI results in about
the same stoichiometry of the primary flame. For example, if propane is an avail-
able substitute for methane or natural gas, it is common practice to prepare a
mixture of approximately 60% propane–40% air (which of course is well above
the upper flammability limit)pffiffiffiffiffiffiffiffiffi
to use
ffi as the fuel supply; though the heating
value of the mixture is 1:53= 2:36 ¼ 1:0 so that its Wobbe Index is the same.
Though there would be slight differences in the stoichiometry of the flame, aris-
ing from the air mixed with the propane, and in Su of the final mixture, substitu-
tion with the same supply conditions would be quite satisfactory. On the other
hand, if a mixture of the same heating value as that of methane were used
(39.2% propane in air) at the same supply pressure, the flame would be much
leaner and generally unsatisfactory (84,85).
There are direct substitutions of possible interest that would not be feasible
without drastic changes in the feed system or pressure. Thus if the available sub-
stitute for natural gas is, eg, a manufactured gas containing much CO, there
would almost always be a mismatch of the WIs unless the fuel could be further
modified by mixing with some other gaseous fuel of high volumetric heating
value (propane, butane, vaporized fuel oil, etc). Moreover, if there are substantial
differences in Su, eg, as a result of the presence of considerable H2 as well as CO
in the substitute gas, the variation in flame height and flashback tendency can
also make the substitution unsatisfactory for some purposes, even if the WI is
reproduced. Refinements and additional criteria are occasionally applied to mea-
sure these and other effects in more complex substitution problems (10,85).
Turbulence in the flow of a premixture flattens the velocity profile and
increases the effective burning velocity of the mixture; eg, at a pipe-Reynolds
number of 40,000 the turbulent burning velocity is several times the laminar
burning velocity and it can be perhaps fifty times larger at very high Reynolds
474 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
numbers. A turbulent flame is always somewhat noisy, the apparent flame sur-
face becomes diffuse owing to the fluctuations in the actual or flame surface
about its average position, and its stability tends to be less predictable. The
instantaneous flame surface may be thought of as wrinkled by velocity variations
in turbulent flow, or by the average distribution over a greater thickness (or
time). Although the resulting enhancement of the mixture consumption rate
may be considerable, turbulence is often considered undesirable in Bunsen-
type flames. For this and other reasons, a large number of burner ports of
small characteristic dimension, rather than a single large port, are frequently
used to assure laminar flow to the individual flames. However, turbulence has
an essential role in facilitating the mixing of fuel, oxidizer, and flame products,
and serves an important function in the various types of flame-stabilizers of prac-
tical importance.
3.2. Combustion of Liquid Fuels. There are several important liquid
fuels, ranging from volatile fuels for internal combustion engines to heavy hydro-
carbon fractions, sold commercially as fuel oils. The technology for the combus-
tion of liquid fuels for spark-ignition and compression-ignition internal
combustion engines is not described here. The emphasis here is primarily on
the combustion of fuel oils for domestic and industrial applications.
In general, the combustion of a liquid fuel takes place in a series of stages:
atomization, vaporization, mixing of the vapor with air, and ignition and main-
tenance of combustion (flame stabilization). Recent advances have shown the
atomization step to be one of the most important stages of liquid fuels combus-
tion. The main purpose of atomization is to increase the surface area to volume
ratio of the liquid. This is achieved by producing a fine spray. The finer the ato-
mization spray the greater the subsequent benefits are in terms of mixing, eva-
poration, and ignition. The function of an atomizer is twofold: atomizing the
liquid and matching the momentum of the issuing jet with the aerodynamic
flows in the furnace (86–88). Some of the more common designs of atomizers
are shown in Figure 9.
Atomizers for large boiler burners are usually of the swirl pressure jet or
internally mixed twin-fluid types, producing hollow conical sprays. Less common
are the externally mixed twin-fluid types (89,90).
The principal considerations in selecting an atomizer for a given application
are turn-down performance and auxiliary costs. An ideal atomizer would possess
all of the following characteristics: ability to provide good atomization over a
wide range of liquid flow rates; rapid response to change in liquid flow rate; free-
dom from flow instabilities; low power requirements; capability for scaling (up or
down) to provide desired flexibility; low cost, light weight, easily maintained, and
easily removed for servicing; and low susceptibility to damage during manufac-
ture and installation (89,90). There are differences in the structures of the sprays
between atomizer types which may affect the rate of mixing of the fuel droplets
with the combustion air and, hence, the initial development of a flame.
For distillate fuels of moderate viscosity 30 mm2/s (¼ cSt), at ordinary
temperatures, simple pressure atomization with some type of spray nozzle is
most commonly used. Operating typically with fuel pressure of 700–1000 kPa
(7–10 atm) such a nozzle produces a distribution of droplet diameters from
10–150 mm. They range in design capacity of 0.5–10 cm3/s or more and the
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 475
Liquid Plain-orifice
Liquid Liquid
Simplex Internal-mixing
air assist
Air
Primary
Duplex
Secondary
Liquid
External-mixing
Air air assist
Secondary Dual-orifice
Primary
Liquid Liquid
Liquid Plain-jet
Liquid Fan spray airblast
Air
Fig. 9. (a) Pressure atomizers; (b) rotary atomizer; and (c) twin-fluid atomizers (89,90).
Dry steam, if available, may also be used in a similar way, as is common practice
in the furnaces of power plant boilers using residual oil (89,90).
Air atomization with low pressure and relatively low velocity air is also
used in some burners for low viscosity distillate oils. In most aircraft gas turbines
some, or even a large part, of the atomization is done in this way by a small
fraction of the warm, compressed combustion air supplied in swirling flow
around the fuel nozzles. Imparting swirl to at least some of the air flow around
fuel injectors of all types is a common feature of many burners and combustors;
in some, swirl is introduced on a larger scale in all of the primary combustion air.
The velocity gradients or shear in the resulting vortexlike flow promote mass
transfer or mixing, including the recirculation of hot products of combustion to
the rich mixture or suspension in the low pressure core that contributes to sta-
bilization of the primary combustion zone (25). The angular or swirl velocity
imparted to the air and the strength of such flows are of course limited by the
available pressure drop; eg, in gas turbine combustors the allowable pressure
loss is usually <4% of the absolute pressure.
Combustion of fuel oil takes place through a series of steps, namely vapor-
ization, devolatilization, ignition, and dissociation, which finally lead to attaining
the flame temperature. Vaporization and devolatilization of the fine spray of fuel
droplets take place as physiochemical processes in the combustion chamber. The
vaporization temperature for fuel oil is in the range of 311–533 K, depending on
the grade of the fuel. Devolatilization takes place at about 700 K. The final flame
temperature attained is between 1366–1918 K. Complete combustion of an oil
droplet can occur in 2–20 ms depending on the size of the droplet. A typical char-
acteristic of an oil flame is its bright luminous nature which is the result of incan-
descent carbon particles in the fuel rich zone.
The study of the combustion of sprays of liquid fuels can be divided into two
primary areas for research purposes: single-droplet combustion mechanisms and
the interaction between different droplets in the spray during combustion with
regard to droplet size and distribution in space (91–94). The wide variety of ato-
mization methods used and the interaction of various physical parameters have
made it difficult to give general expressions for the prediction of droplet size and
distribution in sprays. The main fuel parameters affecting the quality of a spray
are surface tension, viscosity, and density, with fuel viscosity being by far the
most influential parameter (95).
The following general expression (eq. 22) is commonly used to describe the
droplet size distribution in a spray:
dn ¼ ar a expðbr b Þ dr ð22Þ
where dn ¼ number of droplets with radii between r and r þ dr. The constants a,
b, a, and b are independent of r and are usually determined empirically. The best-
known special case of this general equation is the Rosin-Rammler distribution
(95–97).
Theoretical modeling of single-droplet combustion has provided expressions
for evaporation and burning times of the droplets and the subsequent coke par-
ticles. A more thorough treatment of this topic is available (88,91–93,98).
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 477
obtaining stable mixtures that can be successfully atomized and burned. Much
research has been carried out in the area of CWSF atomization but this phenom-
enon is far from being well understood. On the combustion front, novel techni-
ques such as the coupling of a high intensity acoustic field have been employed
to enhance the convective processes occurring during the combustion of CWSF
(116,117).
Fluidized-Bed Technology. In fluidized-bed combustion of coal, air is fed
into the bed at a sufficiently high velocity to levitate the particles. This velocity is
referred to as the minimum fluidizing velocity, umf. At this velocity, the volume
occupied by the bed increases abruptly and the bed exhibits some of the charac-
teristics of a fluid. The two predominant designs of fluidized beds are bubbling
and recirculating, with most theories of fluidization being based on the simpler
bubbling bed concept.
Fluidized combustion of coal entails the burning of coal particles in a hot
fluidized bed of noncombustible particles, usually a mixture of ash and limestone.
Once the coal is fed into the bed it is rapidly dispersed throughout the bed as it
burns. The bed temperature is controlled by means of heat exchanger tubes.
Elutriation is responsible for the removal of the smallest solid particles and
the larger solid particles are removed through bed drain pipes. To increase com-
bustion efficiency the particles elutriated from the bed are collected in a cyclone
and are either re-injected into the main bed or burned in a separate bed operated
at lower fluidizing velocity and higher temperature.
Fluidized beds are ideal for the combustion of high sulfur coals since the
sulfur dioxide produced by combustion reacts with the introduced calcined lime-
stone to produce calcium sulfate. The chemistry involved can be simplified and
reduced to two steps, calcination and sulfation.
Calcination
CaCO3 ! CO2 þ CaO
Sulfation
SO2 þ CaO þ 12 O2 ! CaSO4
Oxidant flow
Jets near distributor
Bubble formation u – umf
and coalescence
Exchange
Emulsion phase umf
Solid circulation
Entrainment in bubble wake
Elutriation by bursting of bubbles
SO2
sorbent
Fines
recycle
Spent
Coal
sorbent
Air
Fig. 10. The main processes taking place in a fluidized bed (92). Heat transfer to
immersed tubes is 30% by radiation.
Plan Plan
Plan Plan area 1.26W area 1.29W
Plan
area 1.08W area 1.16W = 1.56 = 1.63
area W
= 1.15 = 1.25
= 1.00
1.52h 2.1h
1.17h
1.15h
h
1.45H
1.30H
1.07H
1.05H
H
A B C D E
Fig. 11. Effect of coal rank on furnace sizing (constant heat output) (82), where
W ¼ width, D ¼ depth, and h and H are the heights indicated. A represents medium
volatile bituminous; B, high volatile bituminous or subbituminous; C, low sodium lignite;
D, medium sodium lignite; and E, high sodium lignite.
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 483
Jf hf
Iv ¼ ð23Þ
Vc P
where Iv ¼ volumetric combustion intensity, kJ/(m3 h); Jh ¼ fuel feed rate, kg/h;
hf ¼ heating value of the fuel, kJ/kg; and Vc ¼ volume of the combustion cham-
ber, m3.
The area firing density is defined by equation 24,
J f hf
Ia ¼ ð24Þ
Ac
where, Ia ¼ area firing density, kJ/(m2 h); Jh ¼ fuel feed rate, kg/h; hf ¼ heating
value of the fuel, kJ/kg; and Ac ¼ crosssection of a fluid-bed distributor plate or
grate area of a mechanical stoker or plan area in pulverized-coal combustors.
Table 3 provides some design parameters for fossil fuel burners.
3.5. Environmental Considerations. Atmospheric pollutants released
by combustion of fossil fuels fall into two main categories: those emitted di-
rectly into the atmosphere as a result of combustion and the secondary pollu-
tants that arise from the chemical and photochemical reactions of the primary
pollutants.
The main combustion pollutants are nitrogen oxides, sulfur oxides, carbon
monoxide, unburned hydrocarbons, and soot. Combustion pollutants can be
reduced by three main methods depending on the location of their application:
before, after, or during the combustion. Techniques employed before and after
combustion deal with the fuel or the burned gases. A third alternative is to mod-
ify the combustion process in order to minimize the emissions.
Nitrogen Oxides. From the combustion of fuels containing only C, H, and
O, the usual air pollutants or emissions of interest are carbon monoxide,
unburned hydrocarbons, and oxides of nitrogen (NOx). The interaction of the
last two in the atmosphere produces photochemical smog. NOx, the sum of NO
and NO2, is formed almost entirely as NO in the products of flames; typically 5
or 10% of it is subsequently converted to NO2 at low temperatures. Occasionally,
conditions in a combustion system may lead to a much larger fraction of NO2 and
the undesirable visibility thereof, ie, a very large exhaust plume.
NO is formed to some extent from N2 and O2 in flame products when N
atoms are produced at a significant rate. Above 1700 K, the important step in
the much studied Zeldovitch (thermal) mechanism is the production of N
atoms by:
O þ N2 ! NO þ N
N þ O2 ! NO þ O
Here, xO2 is the mole fraction of O2 in the products at temperature T, and the rate
is given in ppm/ms. The exponential implies a large effective activation energy of
570 kJ/mol, the sum of that for the O–N2 reaction and half the dissociation
energy of O2. In typical hydrocarbon–air flames, the rate of NO formation by
the thermal mechanism can be shown to be about 8 ppm/ms, or in a 10 ms resi-
dence time the thermal NO would be about 80 ppm. If preheating the mixture
were to raise the gas temperature by 100 K, the rate of NO production would
be nearly tripled, making the NO concentration unacceptable. Conversely, the
rate can be reduced by the same amount by a 100 K reduction in temperature
by precooling or heat abstraction from the flame itself, or by dilution of the mix-
ture with excess air, steam, or other inert gas such as recirculated, relatively cool
exhaust gas. Control of thermal NOx thus involves reduction of the maximum
Vol. 2 COMBUSTION SCIENCE AND TECHNOLOGY 485
will usually also tend to produce unburned hydrocarbons with their potential
contribution to smog formation. Both may be objectionable, though for different
reasons, at concentrations equivalent to only 0.01–0.1% of the initial fuel.
Although their effect on combustion efficiency would be negligible at these levels,
it is nevertheless important to reduce such emissions.
Neither soot nor unburned hydrocarbons are found in the products of a lean
or stoichiometric premixed flame with air or O2, although hydrocarbons may be
formed or survive unburned in lean flames partially quenched, eg, by the cold
wall of a combustion chamber. A moderately rich flame also should yield only
the water-gas equilibrium products (and N2); but with increasing equivalence
ratio, at some f still well below the upper flammability limit, the burned gas
becomes faintly luminous with precipitated soot particles that increase in num-
ber density and luminous intensity with further increase in f. The appearance of
the condensed phase (soot) is connected with appreciable nonequilibrium concen-
trations of fairly stable low molecular-weight hydrocarbons, notably acetylene,
from the decomposition of unoxidized fuel. These polymerize (condense with
elimination of H2) to form high molecular-weight products, mostly with ring
structures. If these intermediates are not consumed, eg, by OH, in simultaneous
and competing oxidation reactions, they grow until nucleation, and eventually
precipitation occurs as visibly radiating soot, typically over a period of several
milliseconds after the main flame reaction zone. The particle growth and compet-
ing oxidation then continue in the burned gas.
The composition, properties, and size of soot particles collected from flame
products vary considerably with flame conditions and growth time. Typically the
C–H atomic ratio ranges from two to five and the particles consist of irregular
chains or clusters of tiny spheres 10–40 nm in diameter with overall dimensions
of perhaps 200 nm, although some may agglomerate further to much larger sizes.
Whether soot particles form at all and grow depends on f, the fuel type, and
other variables, eg, the growth is easier and faster at a given f with fuels of
higher C–H ratio and at elevated pressure. Given sufficient residence time to
attain equilibrium at the burned gas condition, the soot and hydrocarbons
would eventually be consumed. In practice, their rapid oxidation occurs in a sec-
ondary flame in which the hot primary products are burned with the required
excess air, which is added by diffusion or by more intensive mixing. However,
if a large excess of air is added too rapidly, the cooling can, in effect, quench
the oxidation of both unburned hydrocarbons and the accompanying soot,
which would then persist as visible smoke. The blackness of the smoke depends
on the size and number density of the particles when quenched, their further
aggregation, etc. Some may also survive as much smaller invisible particles, or
condensation nuclei. For thorough reviews on the mechanisms of pollutant for-
mation see References (123–125).
Diffusion Flame Chemistry. Since most combustion systems employ mix-
ing-controlled diffusion flames, which are characterized by very high pollutant
emissions, it is imperative to look into the chemistry occurring in diffusion
flames. In a typical diffusion flame the mixture composition in the reaction
zone is close to the stoichiometric proportion and the temperature is at a maxi-
mum resulting from the large volume of this zone, thus NOx production is
favored. If, however, the surrounding gas cools the combustion products rapidly,
488 COMBUSTION SCIENCE AND TECHNOLOGY Vol. 2
BIBLIOGRAPHY
‘‘Fuels (Combustion Calculation),’’ in ECT 1st ed., Vol. 6, pp. 913–935, by H. R. Linden,
Institute of Gas Technology; in ECT 2nd ed., Vol. 10, pp. 191–220, by D. M. Himmelblau,
The University of Texas; ‘‘Burner Technology,’’ in ECT 3rd ed., Vol. 4, pp. 278–312, by
G. E. Moore, Consultant; in ECT 4th ed., Vol. 6, pp. 1049 –1092, by R. Sharifi, S. A.
Pisupati, and A. W. Scaroni, Pennsylvania State University.
CITED PUBLICATIONS
99. T. F. Wall, The Combustion of Coal as Pulverized Fuel Through Swirl Burners, in
Ref. 100.
100. M. A. Field, D. W. Gill, B. B. Morgan, and P. J. W. Hawksley, Combustion of Pulver-
ized Coal, BCURA, UK, 1967.
101. H. H. Lowry, ed., Chemistry of Coal Utilization, Suppl. Vol., John Wiley & Sons, Inc.,
New York, 1963.
102. L. D. Smoot and D. T. Pratt, eds., Pulverized Coal Combustion and Gassification,
Plenum Press, New York, 1979.
103. L. D. Smoot, Coal and Char Combustion, in Ref. 2.
104. B. R. Cooper and W. A. Ellingson, The Science and Technology of Coal Utilization,
Plenum Press, New York, 1984.
105. M. F. R. Mulcahy and I. W. Smith, Rev. Pure. Appl. Chem. 19, 81 (1969).
106. N. M. Laurendeau, Prog. Energy Comb. Sci. 4, 221 (1978).
107. Modern Power Station Practice, 2nd ed., Pergamon Press, Oxford, UK, 1971.
108. L. D. Smoot and P. J. Smith, Coal Combustion and Gasification, Plenum Press, New
York, 1985.
109. A. Stambuleanu, Flame Combustion Process in Industry, Abacus Press, Tunbridge
Wells, UK, 1979.
110. Prog. Energy Comb. Sci., Special Issue 10, 81–293 (1984).
111. J. B. Howard, Fundamentals of Coal Pyrolysis and Hydropyrolysis, in Ref. 56.
112. C. Y. Wen and E. Stanley Lee, eds., Coal Conversion Technology, Addison-Wesley
Publishing Co., Reading, Mass. 1979.
113. R. H. Essenhigh, in Ref. 61, p. 372.
114. A. W. Scaroni, P. L. Walker, and R. G. Jenkins, Fuel 60, 70–76 (1981).
115. P. Ramachandran, A. W. Scaroni, and R. G. Jenkins, I. Chem., E. Symp. Ser. 107,
128–219 (1987).
116. P. Ramachandran, A. W. Scaroni, G. Reethof, and S. Yavuzkurt, 13th International Con-
ference on Coal and Slurry Technology, Slurry Technology Association, 1988, 241–247.
117. G. Huang and A. W. Scaroni, Fuel 71, 159–164 (1992).
118. J. R. Howard, ed., Fluidized Beds: Combustion and Applications, Elsevier, New
York, 1983.
119. R. Schweiger, Fluidized Bed Combustion and Application Technology, The 1st Inter-
national Symposium, Hemisphere Publishing Corp., New York, 1987.
120. H. R. Hoy and D. W. Jill, The Combustion of Coal in Fluidized Beds, in Ref. 100.
121. H. B. Palmer and C. F. Cullis, The Formation of Carbon from Gases, in P. L. Walker,
ed., Chemistry and Physics of Carbon, Marcel Dekker, New York, 1976.
122. S. Kaliaguine and A. Mahay, eds., Catalysts on the Energy Scene, Elsevier,
New York, 1984.
123. J. M. Beer, in Ref. 59, 439–460.
124. A. F. Sarofim and J. M. Beer, in Ref. 30, 189–204.
125. K. C. Taylor, Automobile Catalytic Converters, Springer-Verlag, New York, 1984.
126. H. G. Wagner, in Ref. 30, p. 3.
127. K. H. Homann, in Ref. 70, p. 857.
128. P. Kesselring, Catalytic Combustion, in F. Weinberg, ed., Advanced Combustion
Methods, Academic Press, Inc., New York, 1986.
EMISSION CONTROL,
AUTOMOTIVE
1. Introduction
There are over 210 million U.S. automobiles and trucks on the road, consuming
large amounts of gasoline and producing commensurately large amounts of gas-
eous exhaust. The exhaust consists primarily of carbon dioxide [124-38-9], water,
unburned hydrocarbons (qv), carbon monoxide [630-08-0], oxides of nitrogen
(NOx), remaining oxygen, and nitrogen (see GASOLINE AND OTHER MOTOR FUELS).
The latter three atmospheric pollutants have been regulated since the 1970s
by the U.S. government and more stringently by the State of California. Automo-
bile companies have developed fuel metering and exhaust systems using the cat-
alytic converter to meet emission regulations. Carbon dioxide emissions are
indirectly controlled by corporate average fuel economy (CAFÉ) standards for
passenger cars and small trucks.
Prior to emission control, passenger car and truck emissions together were
the largest contributors to atmospheric pollution in the United States (1). By
2004, the Tier 2 exhaust emission standards mandated by the Clean Air Act
Amendments of 1990 required automobiles and small trucks to reduce emissions
of hydrocarbons by 98%, carbon monoxide by over 90% and oxides of nitrogen by
98% as compared to 1970 pre-control emissions. California, having more severe
pollution events, has established a system of the most stringent emission control
regulations in the world. The California standards can be adopted by other
states.
The key components of the spark-ignited engine emission control systems
are the three-way catalytic converter and closed loop oxygen sensor (qv) (See
SENSORS) controlled fuel metering system. Exhaust gas recirculation is used to
reduce formation of engine NOx and provides antiknock values. Over 600 million
vehicles have been equipped with the catalytic converter which, has been rated
among the top 10 engineering breakthroughs of the twentieth century (2).
Emission control is achieved without negatively affecting fuel economy or
performance.
Diesel engine emission control technology is not discussed in detail as very
few light-duty diesel engines are now in service, and although the number
is expected to rise in this coming decade, the emissions control technology
has not been selected. Light-duty and medium-duty diesel engine powered vehi-
cles have to meet the same Tier 2 standards (Table 1) as gasoline-fueled engines
including a strict particle emission standard. Heavy-duty diesel engine trucks
must meet very stringent standards starting with the 2007 model year.
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0524080113151514.a01.pub2
494 EMISSION CONTROL, AUTOMOTIVE Vol. 2
of temporary Bins 9 and 10 but eventually have to meet Bin 6, 7, 8 or lower. Any
vehicles sold by a manufacturer in one year in the higher bins must be offset by
an equivalent number of vehicles sold by a manufacturer in that same year that
meet the cleaner emissions of Bins 1, 2, 3, and 4, so as to counterbalance and meet an
overall average 0.07 g/mile NOx emission standard of Bin 5. California LEV2
standards are similar to those of Tier 2 with more emphasis on corporate average
HC emissions. For California automobile manufacturers can choose to comply
with voluntary emissions system durability of 15 years or 240,000 kilometers
at slightly relaxed standards. Extended emission system durability requirements
reflect much the improved and longer engine life now found in actual vehicle use
as a result of clean unleaded gasoline. Highly corrosive leaded gasoline has been
completely banned in the United States since 1996.
The U.S. Tier 2 and California LEV-II program will dominate advances in
emission control in this next decade.
2.2. Test Procedure. Originally vehicle manufacturers had to actually
run test track vehicles for 80,000 km as durability vehicles (3). Periodic exhaust
emissions tests using the U.S. Federal Test Procedure (FTP) determined emis-
sions at several mileages between 6,400 km and 80,000 km and a deterioration
factor (DF) was computed. For example, a calculated DF factor of 1.1 for the dur-
ability vehicle could be applied to the emissions found on any number of 6,400 km
emission data vehicles that represented various vehicle models within that
engine family to predict compliance with the in-use standard at 80,000 km.
Thus not all versions of vehicles had to be run for 80,000 km. Nevertheless,
this original certification procedure was laborious and expensive. A simplified
modern verification has been proposed for new emission durability procedures
using a new standard road cycle (SRC) for whole vehicle durability or an alter-
native method of new standard engine bench cycle (SBC) for accelerated aging of
the catalytic converter and oxygen sensor and other exhaust system components
(4). The objective is to use durability procedures that are at least as severe as
actual vehicle on-road use. The SRC would be used for whole vehicle durability
using 7 laps (each different) in a 3.7 mile test track course or on a mileage accu-
mulation dynamometer for the full useful life mileage. The SBC would be used to
age catalytic converter, oxygen sensor and exhaust system under accelerated test
conditions on a slave engine installed on an engine test bench. The aged compo-
nents are then installed and tested on an actual new vehicle model undergoing
compliance testing. The SBC uses a mathematical approach based on the
Arrehenius equation to determine aging time and exhaust gas temperature to
achieve catalyst and O2 sensor aging effects equivalent to that found with the
SRC whole vehicle durability cycle that are at least as severe as actual in-use
vehicle aging. Auto manufacturers are permitted to develop their own SRC
and SBC and have them approved. Tier 2 emission regulations require field
surveys by automobile manufacturers to assure that vehicles in actual use are
in compliance with the emission standards.
The Federal Test Procedure (FTP) specifies that a test vehicle is fueled with
commercial unleaded gasoline and be stored in a temperature environment of
between 20 and 298C for at least 12 hours immediately prior to the emission
test. The vehicle is placed on a chassis dynamometer which is calibrated for
the vehicle weight and road load. The vehicle is started cold and driven for
496 EMISSION CONTROL, AUTOMOTIVE Vol. 2
Length 1874 seconds - Distance = 11.04 miles - Average Speed = 21.2 mph
70
Cold start phase Transient phase Hot start phase
60 505 seconds 864 seconds 505 seconds
Vehicle speed, mph
50
40
30
20
10
0
0
132
198
264
330
396
462
528
594
660
726
792
858
924
990
66
1056
1122
1188
1254
1320
1386
1452
1518
1584
1650
1716
1782
1848
Test time, secs
(a)
Sample Period = 596 seconds - Distance = 8.01 miles - Average Speed = 48.37
90
80
70
Vehicle speed, mph
60
50
40
30
20
10
0
115
138
161
184
207
230
253
276
299
322
345
368
391
414
437
460
483
506
529
552
575
46
69
92
0
23
(b)
Fig. 1. (a) EPA federal test procedure. (b) US 06 or supplemental FTP driving schedule.
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 497
analyzed and computed with a constant volume sampling system (CVS). All
engine exhaust gases from test start to test end are fed into the (CVS) system
where it is diluted with large amount of air drawn in by the CVS constant
flow pump. Since CVS flow rate is constant throughout the test the air drawn
in modulates inverse to vehicle exhaust flow. Total CVS system flow is recorded
for each test period. Diluted exhaust samples are continuously collected from the
well-mixed diluted exhaust CVS stream for each of the following portions of
the entire test: The first cold start bag contains the continuous diluted gas
sample from the first 505 seconds of the test. The second transient portion bag
is the diluted sample from 506 to 1372 seconds. The engine ignition is turned
off for a 10-min hot soak. The third hot start bag is the diluted gas sample
after the 10-min hot soak from the point of engine ignition of the hot restart
for the last 505 seconds of the test. Each of the three bags are measured for con-
centration of hydrocarbons (HC), carbon monoxide (CO) and oxides of nitrogen
(NOx) using, respectively, a flame-ionization detector (FID) HC analyzer, an in-
frared CO analyzer, and a chemiluminescent NOx analyzer. A CO2 nondispersive
infrared (ndir) is used as input to calculate dilution. The concentration measure-
ments are converted into mass emissions per unit of distance utilizing the
known inputs of distance traveled and total measured CVS system flow. FTP
weighting factors are assigned for each phase to give a composite mass per
unit of distance traveled values for each pollutant for the total test—for example:
0.25 g/km HC; 2.0 g/km CO and 0.25 g/km NOx.
It is interesting to analyze the accumulation of emissions from a vehicle
under test using the US FTP. Fig. 2 shows the typical cumulative grams (g) of
HC emissions with respect to time from a Tier 2 vehicle undergoing the FTP.
Fig. 2 shows the catalyst bed temperature for a typical vehicle meeting the
U.S. Tier 2 emission standards (5). HC accumulation rate is high during the
first 50 seconds from engine start before the catalyst has started to function.
Very little accumulation of HC occurs after the engine and catalyst is up to
operating temperature as can be seen in Fig. 2. NOx emissions occur during cold
and hot starts and coincident with high speed modes. It can be clearly understood
Tier 2 engines will have to be designed to heat up the catalytic unit in about 10
to 15 seconds. In fact, Tier 2 compliant engine design factors being examined to
Fig. 2. Automotive emissions importance of cold start Tailpipe HC emissions (FTP bag 1)
(19). Courtesy Ford Motor Company.
498 EMISSION CONTROL, AUTOMOTIVE Vol. 2
produce hotter exhaust gas are: (1) reduced exhaust system thermal mass;
(2) lean air/fuel cold start; (3) retarded ignition timing; and (4) in some cases vari-
able valve timing. The catalytic converter is located close to the exhaust manifold
(close coupled) to heat up faster. New catalytic units have been developed with:
(5) increased geometric surface area per unit of volume; (6) reduced thermal
mass; and (7) improved thermal/physical properties of the catalyst layer. Clearer
fuels with very low sulfur content are also needed.
These factors are explained in Section 5.
0.8 16 3200
0.6 12 2400
NO
0.5 10 2000
Emissions, mol %
Emissions, ppm
kg/kWh
BSFC
0.4 8 1600
BSFC
0.3 6 1200
CO O2
0.2 4 800
0.1 2 HC 400
H2
0 0 0
10 12 14 16 18 20
A/F mass ratio
Fig. 3. Effect of mixture strength on exhaust gas composition (dry basis) and brake
specific fuel consumption (BSFC) for an unsupercharged automotive-type engine using
indolene fuel, H/C ¼ 1.86, where the ignition is tuned to achieve maximum best torque
(MBT), the brake mean effective pressure (BNEP) is 386 kPa at 1200 rpm (7).
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 499
20
H 2O
18
4.0
16
3.0
1.0
14
CO2
Combustion product, mol %
2.0
12 2.0
10 3.0
4.0
8 1.0
1.0
6
2.0
CO
4.0 3.0
3.0 4.0 1.0
4
2.0 4.0
O2
1.0
2 H2
fuel where the hydrogen to carbon ratio (H/C) is 1.86. The values for a stoichio-
metric air/fuel mixture are given at l ¼ 1.0. The exhaust gas composition will
change substantially as the H/C ratio changes (8), as shown in Fig. 4 (8–10).
Sources of exhaust unburned hydrocarbons are crevices in the combustion
chamber, such as gaps between the piston and cylinder wall, where the com-
bustion flame cannot burn. Unburned hydrocarbon composition is dictated by
fuel composition (11). Carbon monoxide results from areas of insufficient oxygen.
Oxides of nitrogen are produced in the high temperature flame zones during
combustion by reaction of nitrogen molecules and oxygen atoms thermally pro-
duced from oxygen and oxygen-containing species, according to the Zeldovich
mechanism (6,12).
Hydrocarbons and carbon monoxide emission can be minimized by lean air/
fuel mixtures (Fig. 3), but lean air/fuel mixtures maximize NOx emissions. Very
lean mixtures (>20 air/fuel) result in reduced CO and NOx, but encounter
increased HC due to unstable combustion at the lean limit. Improvements in
500 EMISSION CONTROL, AUTOMOTIVE Vol. 2
lean-burn engines have extended the lean limit. Rich mixtures, containing excess
fuel and insufficient air, produce high HC, CO and H2 emissions in the exhaust.
Very rich mixtures have been used in some engine applications at high engine
speeds and loads where the cooling effect of gasoline as it is vaporized in the
cylinder is used to limit high cylinder temperature. A recent supplementary
US FTP procedure called US FTP 06 Fig. 1b is designed to limit such excessive
emissions during high speed driving modes. Rich engine calibrations result in
toxic CO exhaust concentrations of 4 to 5% or more.
The best engine power is achieved slightly rich and the best fuel economy is
achieved slightly lean whereas the stoichiometric air/fuel is the optimal point for
best power, fuel economy and emissions.
Over 150 hydrogen and carbon species are present in the exhaust mix of a
gasoline fueled engine (13–17) including methane, various paraffins, olefins,
aldehydes (qv), aromatics, and polycyclic hydrocarbons as well as unburned gaso-
line. Competitive catalytic oxidative reactions favor oxidation of the latter four
classes before paraffins and methane. Analyses of older catalysts after long in-
use on vehicles reveal that loss of catalytic performance is higher for the methane
and paraffin hydrocarbons than for more reactive species.
Sulfur dioxide is present from the combustion of sulfur contained in the
gasoline. The average U.S. gasoline sulfur content had been 350-ppm (20-ppm
SO2 in exhaust) ranging upwards to 1000-ppm in some sections of the U.S.
From 2004 to 2006 the gasoline fuel sulfur specification will gradually change
to an average of 30-ppm with 80-ppm maximum cap.
Exhaust gas also contains small amounts of hydrogen cyanide and ammo-
nia depending on the air/fuel ratio.
Fig. 5a is the schematic of the major components of the TWC Closed-loop Fuel
Metering System. The catalytic converter is located at the engine exhaust mani-
fold (in a close coupled position for Tier 2 compliant engines). An O2 sensor is
located ahead of the catalyst unit inlet and can be positioned in the converter
inlet shell or in the exhaust manifold. A second O2 sensor is located at the cata-
lytic converter outlet. The inlet O2 sensor generates the operating control signal
to the electronic control unit (ECU). The outlet O2 sensor provides a means for
trimming the control point to correct for system aging changes over the life of the
engine. Both the inlet and outlet O2 sensors also perform a required emission con-
trol quality on-board diagnostic (OBD) function which is explained in Section 8.
The operating O2 sensor is exposed to raw exhaust emitted from the engine and
generates an electrical signal indicative of the concentration of oxygen in the
exhaust. The signal is sent to the ECU. The ECU also obtains an inlet airflow
measurement signal from an engine inlet combustion air measurement meter.
Inlet air flow varies in response to the throttle position and engine load (manifold
vacuum). In addition, the ECU (1) obtains information from the throttle position
switch and engine temperature sensor, (2) has a stored library of engine speed
and load maps, (3) calculates the stoichiometric amount of fuel to be injected
into the measured amount of inlet air, and (4) activates the fuel injector opening
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 501
Fig. 5. (a) Basic components of the TWC closed-loop fuel metering system. (b) The oxy-
gen sensor controls the air/fuel mixture to an average mixture. (Courtesy of Robert Bosch).
time to deliver the proper amount of fuel to a location just ahead of the inlet
valve. After combustion in the cylinder, the exhaust gas is measured by the O2
sensor (3 millisecond response time) producing a millivolt signal indicating
rich (insufficient oxygen) or lean (excess oxygen) combustion products. The ECU
responds to deviations from the stoichiometric control point by correcting the
quantity of fuel injected in the next fueling sequence. Thus the air/fuel ratio is
constantly being adjusted slightly rich and slightly lean of the stoichiometric
mixture (Fig. 5b). The three-way conversion (TWC) catalyst receives exhaust
gas that reflects this constant change back and forth in intake air/fuel mixture
that is stoichiometric centered, but because of the nature of the control system
502 EMISSION CONTROL, AUTOMOTIVE Vol. 2
deviates widely in periodic fashion. The TWC is designed to operate under these
conditions to convert NOx by reduction and HC and CO by oxidation by over 95%.
designed to provide sufficient geometric surface so that all exhaust gas pollu-
tants contact hot catalyst surfaces as they pass through its channels (20–23).
In order to rapidly heat to operating temperature upon engine start the catalytic
unit has low thermal mass and good heat-exchange and mass-transfer properties
to extract heat and efficiently exchange pollutants to the catalyst surfaces and
products back to the exhaust gas stream.
Activated Catalyst Layer. All catalytic chemical reactions occur on cata-
lyst sites within the activated catalyst layer. The layer contains all the cataly-
tically active ingredients consisting of precious metals, base metal promoter
additives, stabilized alumina support particles, and oxygen storage compo-
nents. The activated catalyst layer of a TWC is applied uniformly to all channel
surfaces of the high geometric surface area substrate (ceramic or metallic). The
activated catalyst layer consists of finely divided high BET (see explanation
below) surface area particles that are porous and thermally resistant. The com-
monly used process of applying the catalyst layer to the substrate is one of pre-
paring a water slurry of the particles and components and applying the slurry
uniformly in a thin layer to the all surfaces of the catalyst substrate by: (1) dip-
ping the substrate into a slurry bath; (2) pouring the slurry through the chan-
nels of the substrate; and (3) drawing the slurry up through the substrate by
vacuum. The unit is dried and calcined. Small 2 to 5 nanometer precious
metal crystals are uniformly dispersed on and within the porous particles of
the catalyst layer. The thickness of the catalytic layer is usually between 25
and 100 mm. The substrate or monolith honeycomb body typically has a geometric
surface area (the surface of all channels contained in the substrate) of between
35.1 and 43.7 cm2/cm3 so that l.6 L catalytic unit volume would contain 5.6 to
7.0 m2 of geometric surface area. The activated catalyst layer increases the
total surface to 7000 m2/L, ie, about three orders of magnitude. The total surface
area consists of the combined internal and external surface of all the minute
particles contained in the activated catalyst layer. Total surface area is mea-
sured by the Brunauer-Emmet-Teller (BET) technique, which involves N2
adsorption at low temperature (24).
The minute porous particles are typically aluminum oxide [1344-28-1],
Al2O3. These particles can have BET surface area of 100 to 300 m2/g. The ther-
mal and physical properties of alumina crystalline phases vary according to the
starting phase (aluminum hydroxide or hydrate) and thermal treatment (see
ALUMINUM COMPOUNDS, ALUMINUM OXIDE).
Alumina is used because it is relatively inert and provides the high surface
needed to efficiently and uniformly disperse the expensive active catalytic
components in a thin effective catalyst layer throughout the entire catalytic
unit. However, no one alumina phase possess the thermal, physical, and chemi-
cal properties ideal for the perfect activated catalyst layer. A great deal of
research has been carried out in search of modifications that can make one or
more of the alumina crystalline phases more suitable. For instance, components
such as ceria, baria, lanthana, stronthana, or zirconia are added to enhance the
thermal characteristics of the alumina. Most recent thermal resistant improve-
ments, dictated by high temperature exposure of catalysts directly coupled to
the engine (close coupled), include improved thermal properties of ceria/zirconia
combinations (25–27).
504 EMISSION CONTROL, AUTOMOTIVE Vol. 2
0.115/62
Items Unit 0.17/62 [hexagonal] 0.11/93 0.09/93 0.064/140 0.05/180
material porosity % 35 35 35 35 27 35
wall thickness [nominal] mm 0.17 0.115 0.11 0.09 0.064 0.05
cell pitch [nominal] mm 1.27 1.36 1.04 1.04 0.85 0.73
open frontal area % 75.0 83.8 80.0 83.4 85.6 86.8
506
hydraulic diameter mm 1.10 1.25 0.93 0.95 0.79 0.68
geometric surface area cm2/cm3 27.3 26.9 34.4 35.1 43.5 50.8
pressure dropa % 100 72 130 119 168 217
bulk density g/cm3 0.43 0.29 0.35 0.29 0.28 0.24
bulk density ratio %b 100 67 81 67 65 55
specific heat cal/g/8K <— 0.2 —>
thermal conductivity cal/g/sec/8K <– 3
103 —>
a
Pressure drop values are calculated with the following conditions: Substrate size ¼ 101.6 D
101.6 L; Exhaust gas temp. ¼ 8508C; Gas flow ¼ 5 Nm3/min.
b
Ratio against 0.17mm wall thickness / 62 cell/cm3 substrate. Courtesy of NGK Ceramics, USA, Inc.
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 507
DaimlerChrysler Viper and others (42). The advantage here is lower exhaust gas
pressure drop can result in increased horse-power (1 or 2 hp at high engine
speed) and performance. Other special niche applications exist.
5.3. Heat and Mass Transfer. The heat and mass exchange processes
of exhaust gas to the catalytic surfaces takes place under hot pulsating exhaust
flow conditions. Turbulent flow conditions (Reynolds numbers above 2000) exist
in the exhaust manifold and inlet cone of the catalytic converter in response to
the exhaust stroke of each cylinder (about 6 to 25 times per second) times the
number of cylinders under fluctuating gas pressure. Upon entering the cell pas-
sages of the catalytic unit laminar flow conditions (Reynolds number below 2000)
are developed in a distance of 0.6 cm (45–48).
Heat and mass transfer are driven by temperature and concentration dif-
ferences, respectively, between catalyst surfaces and the bulk gas. Effective
mass transfer of reactants and products commence at operating temperatures
over 2508C as catalyst reactions take place with the catalytic layer.
The process of catalyst heating and initiation of the catalytic functions
involves three or four distinct regions as shown in Fig. 6. Depending on the loca-
tion of the catalytic unit in the exhaust system, total thermal mass before and
including the catalytic unit, and exhaust flow and temperature, it can take
from 10 to 120 seconds for the catalyst unit to reach ignition temperature of
approximately 250–3008C (Region I) where heat exchange processes dominate.
At temperatures above the ignition point, catalyst function increases directly
with temperature increase as catalyst activity increases sharply (Region II)
involving both heat exchange and reactant/product mass transfer processes.
Exothermic reactions contribute to temperature rise within the catalyst layer.
Some catalysts reach a point where the sharp increase in catalytic activity
with inlet temperature abruptly takes on an undesirable mild positive slope
(Region III). Then a point is reached at which catalytic performance improves
only slightly in response to further increases in temperature (Region IV) where
reactant mass transfer processes dominate.
100 CO
HC
80
Conversion, %
60
40
20
In Region I (below the ignition point) the process is one of heat exchange.
Catalytic activity is prevented by lack of sufficient activation energy and because
of chemisorption of other species. Within Region II, the specific activity of the
catalyst is the rate-limiting step; this is the kinetically controlled region. Highly
active catalysts have a lower ignition point and exhibit large increases in
catalytic performance associated with small increases in catalyst temperature.
The behavior of a catalyst in Region II is important to the selection of an
automotive emission catalyst because catalyst light-off is a prime metric in
achieving adequate emission control early in the FTP as shown in Fig. 7.
Region III is not present in most catalysts. It depends on the porous struc-
ture design of the catalyst layer and is often only found in used or spent catalyst.
If present, the rate-limiting step is pore diffusion control. Catalysts clogged with
masking agents such as lube oil ash and/or accumulations of metal oxides from
gasoline organic metallic additives, ie, iron or manganese, would be expected
to show Region III behavior. Region IV is a mass-transfer limited region, ie,
catalytic reactions occur so fast that the rate-limiting factor is the rate of transfer
of reactants and products to and from the surface of the catalyst. Mathematical
models describing the entire process are found in several references (45–48).
and catalyst performance. Base metals deactivate due to reactions with sulfur
compounds at the low temperature end of auto exhaust environments. Base
metal catalysts would need to be considerably larger. Consequently, the larger
volume would take more time to heat up quickly enough to achieve the catalytic
performance demanded by stringent emission standards. Base metal catalytic
components that provide positive synergistic effects on catalyst performance
when combined with precious metal are called promoters. Reducible base
metal oxides are very important to three-way catalyst designs as explained in
a following section.
6.3. Diffusion of Pollutants to Active Catalyst Sites. A representa-
tion of gas transport from bulk gas to a catalyst reaction site is shown in Fig. 8. A
gaseous pollutant such as CO diffuses from the main gas stream into the porous
catalyst matrix. The rate of diffusion depends on molecular weight of the mole-
cule—heavy gaseous hydrocarbons diffuse slower and hydrogen has the fastest
rate of diffusion. There is a net CO flow from the main flow stream to the catalytic
surface because CO is at a higher concentration in the exhaust stream and,
once catalyst reactions take place, of lower concentration within the catalyst
layer. The CO is attracted to a precious metal catalyst site that is temporarily
unoccupied. For the same reason, O2 molecules are attracted to the site. The
O2 molecule is chemisorbed on the precious metal surface where oxygen molecule
bond stretching occurs, leading to dissociation into oxygen atoms and electron
sharing with the clean precious metal catalyst site surface. The reaction of CO
with adsorbed O atoms proceed to form CO2, which in the product state is
released and reenters the exhaust stream as there is less CO2 in the exhaust
stream and more within catalytic layer so there is a net flow of CO2 away from
the catalyst surface.
CO CO2
Macropore
Alumina
Micropores
Fig. 8. Catalyst pore and reaction. The CO diffuses into a precious metal site &; reacts
with O2; and leaves as CO2.
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 511
H2 þ 1=2 O2 ! H2 O
CO þ 1=2 O2 ! CO2
CO þ NO ! 1=2 N2 þ CO2
H2 þ NO ! 1=2 N2 þ H2 O
2 NO þ 5 H2 ! 2 NH3 þ 2 H2 O
2 NO þ CO ! N2 O þ CO2
at 200 8C ðbelow exhaust gas temperatureÞ
2 NO þ H2 ! N2 O þ H2 O
Fuel sulfur reactions
S þ O2 ! SO2
>6008C
Al2 ðSO4 Þ3
3 SO3 þ Al2 O3 !
SO2 þ 3 H2 ! PM catalyst H2 S þ 2 H2 O
The cerium þ3 species captures O when in excess that would otherwise exit the
tailpipe and cerium þ4 releases O for oxidization reactions when in short supply.
These reactions provide a counterbalance chemical force toward maintaining
stoichiometric conditions within the catalyst layer even though the dynamic of
the fuel metering system provide cycles ‘‘rich – lean – rich – lean’’ in response
to the O2 sensor feedback control signal.
Nevertheless, the fuel metering system dynamic (or lag time) has a pro-
found effect on any three-way catalyst formulation. Fig. 9a shows a single-
point fuel injection system with the injector located in the main tube of the
inlet manifold and Fig. 9b show a multi-point fuel injection system with a fuel
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 513
Fig. 9. (a) Single-point fuel injection (TBI). 1 Fuel, 2 air, 3 throttle valve, 4 intake
manifold, 5 injector, 6 engine. (b) Multipoint fuel injection. 1 Fuel, 2 air, 3 throttle valve,
4 intake manifold, 5 injectors, 6 engine. (Courtesy of Robert Bosch).
514 EMISSION CONTROL, AUTOMOTIVE Vol. 2
Fig. 10. The quality of the closed-loop system is very important to catalyst function.
(a) Closed loop 0.3 A/F at 1.5 Hz (multi-point fuel injection). (b) Closed loop 0.5 A/F
at 1.0 Hz (single-point fuel injection). (c) Closed loop 1.0 A/F at 1.0 Hz (closed-loop
carburetor) (51). For (b) and (c) Simulated Utilizing Frequency Generator.
injector located at each inlet valve. Fig. 10 is the catalytic performance evalua-
tion of the same TWC when evaluated on an engine sweep test equipped with a
variable closed-loop feed back O2 sensor controlled fuel metering system so that
different fuel metering systems can be simulated: (a) Actual multi-point fuel
injection. (b) Simulated single-point fuel injection. (c) Simulated closed loop car-
buretor. It is proven here that advanced fuel metering is needed to extract the
optimal catalytic performance of a three-way catalyst (54).
The CO oxidation reactions steps proceed as factors. An O2 molecule is
chemisorbed on Pt metal surface as well as two CO (absorbed) species, resulting
in CO2 formation that enters the exhaust stream as mentioned above.
7.1. Catalyst Durability. The actual in-use overall durability of the three-
way catalyst has proven to be very good even under extreme operating condi-
tions. A durability test of an advanced three-way catalytic converter that was
aged on the MVEG-Cycle (European automobile manufacturers aging cycle
that mimics autobahn driving) for 160,000 km. The decline in HC, CO and
NOx emissions control performance is minor over 20,000 to 160,000 km. The
result shows that very robust three-way catalysts have been developed.
Automobile catalysts last for the life of the vehicle and still function well at
the time the vehicle is scrapped. However, there is potential for decline in total
catalytic performance from exposure to very high temperatures, accumulation of
catalyst poisons, or loss of the active layer (52–57).
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 515
Fuels. The most serious catalyst poison resulted from combustion of a tet-
raethyl lead (TEL) [78-00-2] and ethylene dibromide or chloride (EDB,EDC),
cocktail mixture additive to gasoline to improve octane. Octane enhancement
of gasoline is the only positive feature of TEL with many negative consequences.
Exhaust of leaded gasoline produces ultra-fine air borne lead perfect for inhala-
tion deep into the lung directly contributing to mental development impairment
of children (unrecoverable loss of up to 10 IQ points in urban areas) and serious
detriments to classroom learning processes. Serious life-shortening health pro-
blems for adults and children are also documented (62,63). Combustion of the
TEL cocktail produces acid components that saturate lube oil resulting in severe
engine corrosion and wear resulting in shortened engine life. Continuous dete-
rioration of spark plugs resulted in decline of ignition and combustion quality
and shortened spark plug life. The entire exhaust system was corroded from
the inside out resulting in reduced exhaust system life.
Leaded gasoline (about 10% of world gasoline) is still used in many parts of
the world—Africa, former Soviet Union countries and Middle East countries.
Leaded gasoline use in North America is banned. The only remaining U.S. source
of leaded gasoline is aviation grade gasoline and thus mechanics and a sector of
the population near airports are still exposed to this poison. Also, tetraethyl lead
in containers is reported available in auto supply stores for individuals for their
own vehicles. Catalytic converters are permanently poisoned and rendered inef-
fective from small amounts due to contamination of leaded gasoline in unleaded
gasoline—30 mg Pb/gal caused rapid decline in catalyst performance in a short
period. The mechanism of lead poisoning of automobile exhaust catalysts has
been identified by Ford researchers (64).
Other metal-based fuel additives (Fe, Mn) also create engine debris and accu-
mulate within the engine and exhaust system. Catalysts have been found clogged
with inlet face surface debris/deposits. See reference 5 for a comprehensive analy-
sis of the engine, oxygen sensor and catalyst component functions where one set of
engine components was exposed to clear unleaded gasoline and an identical set
exposed to a Mn-additive containing gasoline. The study is complete with reveal-
ing photographs. The components were interchanged between the respective
engines used for clear unleaded and Mn-additive in a sound scientific study to
reveal: (1) engine out emissions effects; (2) negative effects of O2 sensor perfor-
mance; and (3) declines in catalyst performance. The negative effects appear to
be a direct consequence of accumulation of Mn deposits on and within the O2
sensor and catalyst (5). In recognition of the negative effects of fuel-borne metal-
organic additives, the World-Wide Fuel Charter (WWFC), a document produced in
cooperation with all world engine manufacturers and automobile manufacturers,
specify metal content of unleaded gasoline (Fe, Mn, Pb and other) be non-detectible
for all world in-use vehicle markets (65).
7.4. Fuel Sulfur. Fuel sulfur is responsible for inhibition of TWC
catalyst performance. Fuel sulfur combustion produces sulfur oxides–(SO2/
SO3). Sulfur compounds react with catalyst layer materials and catalyst sites
in a complex manner. Fuel sulfur has been found to contribute to a decline
in HC, CO, and NOx catalyst performance directly associated with the level of
fuel sulfur in two separate studies (66). Sulfur is stored on catalyst sites and
within the catalyst layer in various forms and is released under certain
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 517
conditions. One mechanism is the release of H2S yielding a rotten egg odor to the
exhaust. A small amount of nickel oxide incorporated into the TWC captures H2S
and releases it inoffensively later as SO2 under higher temperature operating
conditions (50,67). The U.S. EPA Tier 2 specification for gasoline sulfur content
is 30-ppm average with a maximum cap of 80-ppm to start phasing-in from 2004
to 2006. When the lower sulfur gasoline phase-in is complete it is expected that
all current on- and off-road in-use vehicles will improve emissions control perfor-
mance by up to 30%. The World Wide Fuel Charter (65) recommends similar fuel
sulfur specifications. It is noted that TWC catalysts avoided a major fuel sulfur
problem with respect to the in the U.S. light-duty emission control program. Oxi-
dation catalysts, originally used from 1975 through the early 1980s, oxidized SO2
to SO3 resulting in H2SO4 emissions in the order of 0.2 g/km—100 times higher
than TWC catalysts. Had oxidation catalysts continued to be needed for emis-
sions control fuel sulfur content would have had to be addressed in the early
1980s.
7.5. Lubricating Oil. Masking agents deactivate catalysts by plugging
the open porous surface of the catalytic layer. Oil ash compounds resulting from
lubricating oil combustion accumulate on catalyst layer surfaces, or lubricating oil
emitted from the exhaust valve stems deposit on the surfaces of the catalyst where
the organic portion is oxidized leaving inorganic ash compounds of zinc, phosphor-
ous, calcium and barium (see LUBRICATION AND LUBRICANTS) on the catalyst surface.
Significant quantities can accumulate depending on lube oil consumption. The ash
compounds tend to partially or fully block some of the surface pores of the catalyst
layer called pore mouth plugging. It only takes a small amount of material on the
surface of the catalyst to restrict the free passage of exhaust gases into and out of
the porous catalyst layer. A zinc pyrophosphate glaze had been found to form on
the catalyst surface under certain temperature conditions (68,69) masking large
areas of catalyst surface and preventing passage of gases into and out of the cat-
alyst porous structure. Materials such as calcium inhibit the formation of this
glaze. Because modern engines using clean unleaded fuel have a much longer
life, masking mechanisms are becoming a more important in catalyst performance
decline for high mileage engines. Metal contents of lubricating oils are currently
being studied and it is expected they will be reduced to match engine needs fueled
with Tier 2 low sulfur unleaded gasoline (70).
7.6. Catalyst Layer Loss. Partial loss of the catalyst layer via attrition,
erosion, or loss of adhesion and exfoliation could result in some loss of catalyst
performance. The monolithic catalyst is designed to resist all of these mechan-
isms. Nevertheless, catalyst materials have been found in roadside dust extend-
ing a few meters from the roadway indicating some loss is occurring. Analyses of
the leading inlet face edge of catalytic units have revealed evidence of small loss
via a gas erosion process at the leading edge. If this becomes an issue it can be
addressed by a hardening procedure.
The first commercial application of the TWC catalyst and closed loop fuel meter-
ing system using an oxygen sensor was on the 1977 Volvo (four models) sold in
518 EMISSION CONTROL, AUTOMOTIVE Vol. 2
the California market (71–74). The catalyst was invented by Engelhard (50).
Volvo, Bosch, and Engelhard cooperated to bring the system to production over
a 2.5 year period. Several other car companies adopted the system the following
year, and by the 1990s almost all cars sold in the United States and Canada
utilized it. The fully developed system is described in the literature (see SENSORS).
It is now used on almost all light-duty, medium-duty and heavy-duty SI engine
vehicles produced in the world.
A brief description of the system follows. The function of the oxygen sensor,
electronic control unit (ECU), and fuel metering system is to deliver a stoichio-
metric air/fuel mixture to the engine cylinder combustion process, ie, there
should be neither excess air nor excess fuel. The main purpose is to provide opti-
mal exhaust conditions for the TWC catalyst to reduce NOx while at the same
time to oxidize HC and CO within the same catalyst bed forming nonpollutants
N2, H2O, and CO2. In actual operation, the O2 sensor continuously senses
whether there is an excess or lack of O2 present in the exhaust products. The
O2 sensor signal is continuously monitored by the ECU. The ECU also receives
a continuous air flow measurement signal from a mass-flow inlet airflow
measurement device (usually a hot wire anemometer) (75,76); calculates the
proper amount of fuel for the measured amount of air, and controls each fuel
injector opening time to deliver the calculated amount of fuel. The ECU uses
the O2 sensor signal as a correction signal to the computed amount of fuel to
be injected into a measured flow of input air. This control approach is known
as a feed-back system and by its nature is always correcting after the cylinder
combustion event has taken place. The result is a correction cycle above and
below the stoichiometric air/fuel mixture centering at stoichiometric. The TWC
is designed to operate effectively in such an environment. It is found that mini-
mizing the deviation from the stoichiometric control point and minimizing the
frequency with which it occurs is very beneficial to the performance of any
TWC catalyst (51).
8.1. Oxygen Sensor. The O2 sensor is also referred to as lambda sonde
or lambda sensor (lambda is the Greek letter used to denote the air/fuel ratio),
and as an exhaust gas oxygen sensor (EGO). The O2 sensor is a concentration
cell that responds to a changes in exhaust oxygen content in 3 milliseconds yield-
ing a millivolt output signal, ie, 50 mV at i ¼ 1.05 (lean) to 900 mV at i ¼ 0.99
(rich) as shown in Fig. 11. The voltage changes sharply in the immediate vicinity
of the stoichiometric ratio (i ¼ 1.0). The sensor consists of (qv) ceramic porous
solid electrolyte that is ionically conductive at operating temperatures >3508C.
The outside of the ceramic is coated with platinum electrodes: one electrode
exposed to the exhaust gas, the second to ambient air. Between these electrodes
is a zirconia solid electrolyte. The voltage generated depends on the oxygen con-
centration difference between each electrode. The exhaust porous platinum elec-
trode is a catalyst at the exhaust gas surface and equilibrates the mixture,
consuming, by catalytic action, unreacted oxygen and CO, H2,and HC, yielding
a net amount of oxygen present. Near the stoichiometric point, ie, i ¼ 1.0, the
difference between atmospheric O2 and exhaust O2, and thus the voltage gener-
ated, approach a maximum. Thus a control point sensor signal of about 450 to
500 mv is used to detect whether the exhaust is above or below the stoichiometric
control point (75–77). In order for the sensor to function properly, it needs to be
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 519
800
600
Voltage, Vm
400
200
0
0.6 0.8 1.0 1.2 1.4 1.6
A/F ratio, λ
Fig. 11. The A/F ratio ad the lambda sensor’s voltage signal. Courtesy of Robert Bosch.
heated to above 3508C. To accommodate the need for quick O2 sensor catalyst
function, a heated version of the O2 sensor is used.
8.2. Computer Controller. The computer controller also known as the
electronic control unit (ECU), is used for various engine management and
other vehicle functions. It is rightly claimed that the TWC closed-loop system
brought the ECU into dominant use. The ECU has a stored library of engine
speed and load maps. The input signals of mass inlet air measurement,
engine temperature, throttle position and O2 sensor in conjunction with stored
data are employed to assist the engine in cold starting and to calculate the
proper amount of fuel to be injected by each fuel injector for its current oper-
ating condition. The ECU control system uses the O2 sensor signal as a switch-
ing device using the logic ‘‘rich! Go lean!’’ or ‘‘lean! Go rich!’’ Advanced closed
loop systems use the input of a second O2 sensor located downstream of the
catalyst for minor correction to the system due to long term system drift. An
advanced Honda system using an observer O2 sensor to anticipate rate of
change (rate adjustment) has had a dramatic effect in maximizing the perfor-
mance of the TWC. Utilizing these improvements was a key to a Honda emis-
sion control system able to meet the California super ultra-low emission
vehicle standard (SULEV) and partial zero emission vehicle standard
(PZEV) with a single bed TWC catalyst.
The oxygen sensor closed loop system automatically compensates for
changes in fuel composition or air density. For instance, the system corrects for
stoichiometry differences in winter and summer grades of gasoline or those
fuels containing various amounts of oxygenated hydrocarbons such as MTBE
or ethyl alcohol. Also the stoichiometric air/fuel mixture is maintained even
when the vehicle climbs from sea level to high altitudes where the air density
is lower.
520 EMISSION CONTROL, AUTOMOTIVE Vol. 2
The oxygen sensor signal indication for the stoichiometric point can be a bit
different for certain gases, for instance H2 has a high rate of diffusion. Natural
gas and methanol have high H2 in the exhaust (see Fig. 4). This has to be taken
into account for flexible fuel vehicles that operate on two different fuels; ie
methanol and gasoline.
8.3. On-Board Diagnostics. An on-board diagnostic system (OBD-II) is
required on all new vehicles since 1994. The object is to assure and monitor con-
tinued functional performance of combustion quality (engine misfire), the perfor-
mance of the catalytic converter, and the evaporative emission control system.
The ECU would record incidents of failure by displaying a notice to the vehicle
driver/owner and for use in annual vehicle inspection purposes.
An OBD-II system, during an assessment event, compares the output signal
of two O2 sensors: (1) the standard operating sensor is located in front of the
TWC catalyst; and (2) a second OBD-II O2 sensor located downstream of the cat-
alytic converter (or behind the first converter when more than one is used). The
assessment event is conducted by the ECU which schedules a series of distinct
air/fuel ratio step changes from rich to lean and back to rich air/fuel at a defined
frequency of 2 Hz. During an assessment event, a fully functional TWC catalyst
will produce a characteristic OBD-II O2 sensor signal indicating proper catalyst
performance. An impaired or failed TWC catalyst would produce a different
characteristic OBD-II O2 sensor signal indicating catalyst performance below a
desired rate. The component of the TWC being measured is the ceria function as
described earlier in Section 6 and in Refs. 49,78. A fully functional TWC catalyst
containing activated ceria would capture O2 when in excess and release O2 when
in short supply. Thus the OBD-II O2 sensor located downstream of the catalyst
would sense a stoichiometric A/F mixture and produce a steady 450 to 500 milli-
volt signal indicating that the catalyst function is normal. However, if the OBD-II
O2sensor produced a millivolt signal of 50 to 700 at 2 Hz that mirrored or
approached that of the operating O2 sensor signal, it would indicate that the
TWC catalyst ceria function was damaged or no longer able to perform the oxy-
gen storage and release reactions and by inference had lost catalyst functional
performance (79–82).
9.1. Exhaust Gas Recirculation. NOx control within the engine can be
achieved by recycling a fraction of the exhaust gases into the engine air inlet
stream. The resultant mixture upon combustion results in lower peak combus-
tion temperature and lowers the amount of thermal NOx produced within the
combustion flame. The reaction of N2 þ O ! NOx is strongly dependent on the
combustion flame temperature (75,83). The degree of NOx suppression is depen-
dent on the amount of exhaust gas recirculation (EGR). EGR reduces flame speed
of the mixture, and thus also provides a certain antiknock quality to the combus-
tion process. The impact of EGR on engine parameters has been detailed (84).
EGR in modern engines are quite transparent to the driver/operator with
respect to engine power and performance. In the 1970s vehicle engine perfor-
mance had been noticeably degraded due to early EGR systems; especially at
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 521
idle, under load at low speed, and during cold start. This negative public memory
lingers but is only of historic importance.
9.2. Evaporative Emissions. Gasoline fumes from the on-board fuel
tank, from fuel left in the fuel delivery system, and from running losses are regu-
lated by U.S. EPA standards. Gasoline is a wide mixture of hydrocarbon species
from high volatility butane (C-4) to lower volatility C-8 to C-10 hydrocarbon com-
pounds. High volatility fractions are needed for choked engine cold start. Winter
and summer grade fuels vary the volatile fractions accordingly.
Evaporative sources are (1) Diurnal emissions due to daily fluctuations in
ambient temperature where evaporative emissions depend on the combined
vapor pressure of the fuel hydrocarbon components and the fuel tank tempera-
ture; (2) Hot soak emissions due to fuel left in the fuel metering system or the
inlet manifold after the engine is turned off. Engine and exhaust system thermal
energy heats the left over fuel causing evaporation and emission loss; (3) Running
loss emissions which occur during vehicle operating modes such as low speeds on
a hot day when the rate of vapor generation in the entire fuel system exceeds
engine fueling demand; and (4) filling vapor losses during fuel tank filling
where the liquid gasoline displaces the fuel tank vapor. Fuel tank filling
losses are controlled by a fuel filling pump device at the gasoline filling station
(85,86).
The current EPA evaporative emission regulation prohibits more than 0.95
to 1.2 grams per test to escape to the atmosphere depending on the vehicle class
(3). The test is a 3-day diurnal temperature cycle and an immediate hot soak in a
shed enclosure following the dynamometer run.
The vehicle evaporative control system typically consists of a charcoal can-
ister that connects to the intake manifold. At engine shut off a valve permits
hydrocarbon fumes to be absorbed and stored by a carbon canister. When the
engine is operating, the stored hydrocarbons are purged into the inlet manifold
where they enter the combustion process.
9.3. Crankcase Emissions. Exhaust gases enter the crankcase by
what is known as blow-by gases that escape the combustion chamber past the
piston rings or inlet and outlet valve stem seals. The exhaust gases would
enter the atmosphere via the crankcase breather tube. As the engine wears,
blow-by increases. A positive crankcase ventilation (PCV) valve is required by
regulation (with specified test procedure) to assure these blow-by gases are fed
into the inlet manifold so that they pass through the engine and exhaust emis-
sion control system (87,88).
The National Energy Policy Act of 1992 requires that nonpetroleum-based trans-
portation fuels steadily replace petroleum fuels in the United States. Such fuels
include natural gas (see NATURAL GAS), liquid petroleum gas (qv) (LPG), methanol,
ethanol and hydrogen (qv).
10.1. Natural Gas. Natural gas fueled passenger cars and light-duty
trucks are produced by U.S. automobile manufacturers (89). Emission control
systems are developed for natural gas fueled engines. The preferred emission
522 EMISSION CONTROL, AUTOMOTIVE Vol. 2
control TWC catalyst is based on Pd and Rh rather than Pt and Rh. Engine out
emissions are shown in Fig. 4 for 4/1 H/C ratio. Natural gas combustion produces
larger amounts of H2 and H2O, lesser amounts of CO and CO2, and CH4 is the
predominant HC emission species with formaldehyde as minor. Formaldehyde is
carcinogenic and has it own regulated emission standard (Table 1). TWC cata-
lysts are quite effective for formaldehyde emission control. The U.S EPA and
California emission standards exclude CH4 emissions in the definitions of
NMOG (nonmethane organic gas) and NMHC (nonmethane hydrocarbons) emis-
sion standards. CH4 has low reactivity in the atmosphere and low ozone (qv)
formation potential (90). Pd-based TWC catalysts require exhaust gas tempera-
ture over 4508C to oxidize CH4. The symmetrical methane molecule is the most
refractory of the hydrocarbon family and resists catalytic reactions on traditional
TWC catalysts below 4508C. The CH4 oxidation performance of Pd and Pd-Rh
TWC catalysts are quickly impeded by very small amounts of sulfur in natural
gas. Interesting catalyst work in this area can be found in work by R. Hicks (91).
Natural gas fuel metering systems must assure sufficient CO be present to per-
mit efficient reduction of NOx with CO. For this purpose a 2/1 ratio of CO/NOx
appears sufficient.
10.2. Liquid Petroleum Gas (LPG). LPG is a transportation fuel used
for special vehicle and equipment applications. Emission control systems are
developed and typically use Pt-based (Pt-Rh) TWC catalysts which exhibit
lower light off temperature for propane (C3H8), the major HC species in LPG.
10.3. Alcohol Fuels. Emission control systems have been developed for
methanol and ethanol transportation fuels which normally contain 15% gasoline
and are known respectively as M85 and E85 alcohol fuels. Engines have been
designed for these fuels and emission control systems have been modified for effi-
cient emission control (92–94). M85 fuel exhaust formaldehyde emissions are the
order of 0.60 g/m. Pd-Rh TWC catalysts are found very effective in controlling
formaldehyde emission. E85 fuel exhaust emissions contain acetaldehyde and
formaldehyde. Standard TWC catalyst formulations effectively control aldehyde
emissions. Ethanol is added to gasoline as an oxygenated additive to meet fuel
composition requirements for certain non-attainment or non-compliance regions
in the United States. The closed loop TWC catalyst fuel metering system is able
to accommodate the change in air/fuel stoichiometry induced by the addition of
alcohols without loss in its ability to control exhaust emission.
Light-duty, high speed, diesel compression ignition (CI) engine powered vehicles
must meet the same emission standards as gasoline spark ignited (SI) engines
(Table 1). The CI engine has been greatly improved over that last decade with
respect to fuel efficiency and engine performance. Diesel engines are found in
light and medium trucks but in very few passenger cars. Penetration into
the light-duty vehicle market is expected to increase sharply over the next dec-
ade. Light-duty diesel vehicles have reached 50% of sales in France and a sub-
stantial percentage in most European countries because of improvements in
fuel efficiency and dramatic improvements in performance due to excellent
Vol. 2 EMISSION CONTROL, AUTOMOTIVE 523
engine torque at low speeds. Diesel fuel has more energy per liter than gasoline
(35,785 vs 31,845 kJ/L). Fuel efficiency for light-duty diesel vehicles can range
from 13% to 24% better than for similar gasoline vehicles in similar service
(39). Higher gasoline prices and growing concern for global warming could
increase attractiveness of diesel engines for the North American market.
11.1. Emission Standards. Emission standards for light- and medium
engines are shown in Table 1. Diesel engine has to meet all standards without
exception. The most difficult standards for diesel engines are the particulate
matter (PM) and NOx standards whereas HC and CO are easily removed with
diesel oxidation catalysts.
11.2. Nature of Diesel Exhaust. Modern high speed compression igni-
tion engines used for passenger cars and LDV are equipped with turbochargers
and electronically controlled fuel injection utilizing a common rail fuel feed line,
individual electronically actuated injectors and very high pressure fuel injection.
The individual injectors provide full flexibility of fueling timing and fueling rates
within the start of injection and finish of injection event; ie, (1) a small amount of
fuel can be added early in the compression stroke referred to as pilot injection;
(2) the main amount of fuel can be injected in one squirt, in several squirts and
with modulated fuel rates or shapes that are adjustable during the fueling event;
and (3) post ignition injection can be scheduled to activate downstream emis-
sion control regeneration when necessary. Base engine emissions from these
diesel engine have lower NOx and PM mass. However, ultra-fine particles (30
to 100 nanometer diameter in size) have not been reduced with any of these
improvements.
Combustion air is inducted into the engine in response to engine RPM and
fueling rate without a throttle. Excess air is present for every condition of engine
speed and load. Mass emissions of carbon monoxide and hydrocarbons are low
compared to a SI gasoline engine. NOx generation is high even though overall
air/fuel is very lean because most combustion is diffusion flame combustion of
fuel droplets rather than homogeneous combustion.
11.3. Emission Control Technology. The light-duty vehicle emission
control system that will conform to Tier 2 emission light-duty vehicle emission
standards is currently under development. In 2006 the vehicles would have to
meet at lease Bin 8 NOx 0.124 and PM 0.0124 g/km (Table 1). There two major
challenges: 1) control of NOx and 2) control of particle emissions:
NOx Control. Two approaches under consideration are: (1) Use of large
fractions of EGR (see Section 9) to prevent formation of thermal NOx; or (2)
Catalytic approaches such as NOx adsorbers. Catalytic reduction of NOx in
the oxidizing conditions of diesel exhaust is difficult at the 90% effective reduc-
tion rates needed. One effective approach that is under intense development
is one that stores and releases NO2. In this respect the NOx adsorber is much
like a TWC catalyst the latter stores and releases O2 whereas the NOx adsorber
stores and releases NO2. The feat is accomplished by oxidizing NO ! NO2 on a
Pt component of the NOx Adsorber and storing it chemically as Ba(NO3)2 for
instance. Periodically, the engine is calibrated slightly rich of stoichiometric
for a few seconds and NO2 is released and catalytically reduced by the Rh com-
ponent in a TWC catalytic reaction. This technology requires ultra-low sulfur
diesel fuel.
524 EMISSION CONTROL, AUTOMOTIVE Vol. 2
BIBLIOGRAPHY
‘‘Automotive Exhaust Control’’ in ECT 2nd ed., Vol. 2, pp. 814–839, by W. B. Innes and
K. Tsu, American Cyanamid Co.; ‘‘Exhaust Control, Automotive’’ in ECT 3rd ed., Vol. 9,
pp. 494–510, by C. M. Heinen, Chrysler Corp.; in ECT 4th ed, Vol. 9, pp. 982–1022, by
John J. Mooney, Engelhard Corp.
CITED REFERENCES
52. K. Otto, W. B. Williamson, and H. Gandhi, Ceram. Eng. Sci. Proc. 2(6), (May/June,
1981). Good review of various catalyst deactivation processes.
53. H. S. Gandhi, W. B. Williamson, and J. L. Bomback, Appl. Catal. 3, 79–88 (1982).
54. E. Jobson and co-workers, Deterioration of Three-Way Automotive Catalysts, Part I—
Steady State and Transient Emission of Aged Catalyst, SAE 930937, Society of Auto-
motive Engineers, Warrendale, Pa., 1993.
55. S. Lundgren and E. Jobson, Deterioration of Three-Way Automotive Catalysts,
Part II—Oxygen Storage Capacity at Exhaust Conditions, SAE 930944, Society of
Automotive Engineers, Warrendale, Pa., 1993.
56. L. A. Carol, N. E. Newman, and G. S. Mann, High Temperature Deactivation of
Three-Way Catalyst, SAE 892040, Society of Automotive Engineers, Warrendale,
Pa., 1989.
57. N. Miyoshi and co-workers, Development of Thermal Resistant Three-Way Catalyst,
SAE 891970, Society of Automotive Engineers, Warrendale, Pa., 1989.
58. J. K. Hockmuth and co-workers, Hydrocarbon Traps for Controlling Cold
Start Emissions, SAE 930739, Society of Automotive Engineers, Warrendale,
Pa., 1993.
59. C. D. Tyree, Emission Level in Catalyst Temperature as a Function of Ignition-
Induced Misfire, SAE 920298, Society of Automotive Engineers, Warrendale, Pa.,
1992.
60. J. E. Kubsh, J. S. Rieck, and N. D. Spencer, in A. Crucq, Ed., Catalysis and Automo-
tive Pollution Control II, Elsevier, New York, 1991, p. 125.
61. N. Miyoshi and co-workers, Development of Thermal Resistant Three-Way Catalysts,
SAE 891970, Society of Automotive Engineers, Warrendale, Pa., 1989.
62. W. Rogan and J. Ware, Exposure to Lead in Children—How Low is Low Enough?
New Engl. J. of Med 348:16, 17 Apr 2003.
63. Needleman and co-workers, Deficits in Psychologic and Classroom Performance of
Children with Elevated Dentine Lead Levels, New Engl. J. Med. 300(13), (Mar. 29,
1979).
64. H. S. Gandhi and co-workers, ‘‘Affinity of Lead for Moble Metals on Different
Supports,’’ Surface Interface Analy. 6(4), (1984).
65. World-Wide Fuel Charter, Alliance of Automobile Manufacturers, 1401 H St.,
Washington, DC 20005. www.autoalliance.org, December 2002.
66. AAMA / AIAM Study on te Effects of Fuel Sulfur on Low Emission Vehicle Criteria
Pollutants, American Automobile Manufacturers Association and Association of
International Automobile Manufacturers, Decemberr 1997. CRC Sulfur/LEV
Program, CRC Project No. E-42, Coordinating Research Council Inc., December
27, 1997.
67. U.S. Pat. 5196390 (Mar. 23, 1993), S. Tauster, J. Dettling, and J. Mooney (to
Engelhard Corp).
68. W. B. Williamson and co-workers, Catalyst Deactivation Due to Glaze Formation
from Oil-Derived Phosphorus and Zinc, SAE 841406, Society of Automotive Engi-
neers, Warrendale, Pa., 1984.
69. K. Inoue, T. Kurahashi, and T. Negishi, ‘‘Effect of Phosphorus and Ash Con-
tent of Engine Oil on Deterioration of Monolithic Three-Way Catalysts and
Oxygen Sensors’’ SAE 92054, Society of Automotive Engineers, Warrendale, Pa.,
1992.
70. Tier 2/LEV II Emission Control Technologies For Light-Duty Gasoline Vehicles,
Manufacturers of Emission Controls association, 1660 L St NW, Washington DC
20036. August 2003 www.meca.org
71. G. T. Engh and S. Wallman, Development of the Volvo Lambda-Sond System, SAE
770295, Society of Automotive Engineers, Warrendale, Pa., 1977.
528 EMISSION CONTROL, AUTOMOTIVE Vol. 2
GENERAL REFERENCES
R. B. Bird and co-workers, Transport Phenomena, John Wiley & Sons, Inc., New York, 1965.
Code of Federal Regulations, Title 40, Part 86, Washington, D.C., July 1, 1992.
G. C. Bond and G. Webb, eds, ‘‘Catalysis,’’ Royal Soc. Chem., London 6 (1984).
D. D. Eley, H. Pines, and P. B. Wiez, eds., Advances in Catalysis, Vol. 33, Academic Press,
Inc., New York, 1985.
H. Heinman and J. J. Carberry, eds., Catalysis Reviews—Science and Engineering,
Vol. 26, Marcel-Dekker, New York, 1984.
C. N. Satterfield, Heterogeous Catalysis in Practice, 1st ed., McGraw-Hill Book Co., Inc.,
New York, 1980.
C. D. Falk and J. J. Mooney, Three-Way Conversion Catalysts—Effect of Closed Loop Feed-
back Control and Other Parameters on Catalyst Efficiency, SAE 800462, Society of
Automotive Engineers, Warrendale, PA., 1980.
Tier2/LEV Emission Control Technologies For Light-Duty Gasoline Vehicles, Manufac-
turers of Emission Controls Association, 1660 L St., NW, Washington DC 20035
www.meca.org. August 2003.
JOHN J. MOONEY
Environmental and
Energy Technology and
Policy Institute
EMISSION CONTROL,
INDUSTRIAL
1. Introduction
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.0524080102051807.a01.pub2
530 EMISSION CONTROL, INDUSTRIAL Vol. 2
There are two broad strategies for reducing volatile organic compound (VOC) emis-
sions from a production facility: (1) altering the design, operation, maintenance, or
manufacturing strategy so as to reduce the quantity or toxicity of air emissions pro-
duced, or (2) installing after-treatment controls to destroy the pollutants in the gen-
erated air emission stream (5). Whether the exhaust stream contains a specific
hazardous air pollutant, a VOC, a nitrogen oxide, or carbon monoxide, the best
way to control the pollutant is to prevent its formation in the first place. Many tech-
nologies are being developed that seek to minimize the generation of undesirable
by-products by modifying specific process materials or operating conditions.
Whereas process economics or product quality may restrict the general applicabil-
ity of these approaches, an increased understanding of the mechanisms and condi-
tions by which a pollutant is created is leading to significant breakthroughs in
burner design and operation (for nitrogen oxide control), equipment design, main-
tenance, and operation for fugitive and vent VOC emission control, and product
and waste storage and handling design and operation (for VOC emission control).
One source category of increasing importance is that associated with fugitive emis-
sions from equipment leaks.
Vol. 2 EMISSION CONTROL, INDUSTRIAL 531
The approaches used to reduce equipment leak emissions fall into two
categories: (1) modifying or replacing existing equipment with a more effective
Table 1. Fugitive Emissions—Sources and Control Effectivenessa
LDAR Program control
effectiveness, %
532
dual mechanical seal 100
with barrier fluid
valves 150–600 0.32 4.1 sealless design, 100 87 67 92
bellows
compressors 1–3 12 85 closed-vent system 90 93
dual mechanical seal 100
with barrier fluid
open-ended 7–30 0.11 0.63 blind, cap, plug, or 100
lines second valve
flanges and 700–3000 0.10 0.60 weld or replace 100 33
connectors
a
Based upon information provided in Refs. 9 and 10.
Vol. 2 EMISSION CONTROL, INDUSTRIAL 533
design, and (2) implementing a leak detection and repair (LDAR) program. The
focus of such a program will vary depending on component type, the service (gas,
light liquid or heavy liquid) and previous leak history (10,11) (see Table 1).
3.1. Pumps. Chemicals transferred by pumps can leak at the point of
contact between the moving shaft and the stationary casing. Thus, pumps
generally require a seal at the point where the shaft penetrates the housing in
order to isolate the pumped fluid from the environment.
Equipment modifications for pumps include (1) routing leaking vapors to a
closed-vent system, (2) installing a dual mechanical seal containing a barrier
fluid, or (3) replacing the existing pump with a sealless type (canned-motor, mag-
netic drive, and diaphragm pumps). The control efficiency of sealless pumps and
a dual mechanical seal is essentially 100%, assuming both the inner and outer
seal do not fail simultaneously.
3.2. Valves. Most common valve designs contain a valve stem that oper-
ates to restrict or limit fluid flow. Typically, the stem is sealed by a packing gland
or O-ring to prevent leakage of process fluid to the atmosphere. Emissions from
valves occur at the stem or gland area of the valve body.
Emissions from process valves can be eliminated when the valve stem can
be isolated from the process fluid, (ie, using sealless valves). Two types of sealless
valves, diaphragm valves and sealed bellows, have been developed. The control
efficiency of both types is essentially 100%.
3.3. Compressors. Compressors transport gases through a process unit
in much the same way that pumps transport liquids. Compressors are typically
driven with rotating or reciprocating shafts. The sealing mechanisms for com-
pressors are similar to those for pumps.
Emissions from compressors may be reduced by collecting and controlling
the emissions from the seal using a closed-vent system or by improving seal
performance by using a dual mechanical seal system similar to pumps. The
dual mechanical seal system has an emissions control efficiency of 100%, assum-
ing both the inner and outer seal do not fail simultaneously.
3.4. Pressure Relief Devices. Pressure relief devices are safety devices
commonly used in chemical facilities to prevent operating pressures from exceed-
ing the maximum allowable working pressures of the process equipment. Equip-
ment leaks from pressure relief devices occur when material escapes from the
pressure relief device during normal operation, either because the device does
not properly reseat after a release, from general deterioration of the seal, or
when the process pressure is operating too close to the set pressure of the PRV.
There are two primary equipment modifications that can be used for control-
ling equipment leaks from pressure relief devices: (1) a closed-vent system, or
(2) use of a rupture disk in conjunction with the PRV. The control efficiency of
the closed vent system is dependent upon a number of operational factors.
When Rupture disks can be installed upstream of a PRV they can prevent fugitive
emissions through the PRV seat. The control efficiency of a rupture disk/PRV
combination is essentially 100% when operated and maintained properly. Rup-
ture disks cannot be used on processes where solids or polymers might collect
around the disk, reducing its ability to release at the design pressure.
3.5. Flanges and Connectors. Connectors and flanges are the most
common pieces of equipment in a chemical process unit. They are used to join
534 EMISSION CONTROL, INDUSTRIAL Vol. 2
535
temperatures, and
maintenance
catalytic 95–99% heat recovery spent catalyst simple systems, lower fouling of catalysts,
oxidation regeneration T than thermal temperature limits
acids and economical operation
alkalines
biofiltration 65–99% none some contaminated low installation and large area required for
water and inert operating cost. Minimal installation. Low
solids waste and air pollutant removals for heavy
generation. Particularly organics. Some
suited for odor reduction. pretreatment required.
potential for upsets with
variable exhaust stream.
536 EMISSION CONTROL, INDUSTRIAL Vol. 2
4. Biofiltration Systems
Pretreatment
no
Dilution and / or
additional
monitoring needed
no
Auxiliary heat needed
Calculate temperature
rise through bed
Develop
system
Fig. 1. Process flow diagram for the selection of an exhaust control system.
538 EMISSION CONTROL, INDUSTRIAL Vol. 2
<50 ppmv. Unfortunately, not all organics are degraded at the same rate. For
that reason, the destruction efficiency of a bioreactor depends heavily on the
composition of the organic feed stream. Large quantities of heavier organics
can depress the average VOC destruction efficiency even while the oxidation
efficiency of simpler compounds remains very high (18).
The EPA has indicated that where applicable, both the capital cost of a bio-
filter installation and its operating costs can be just a fraction of the cost of a
traditional emission control devices.
An efficient Biofiltration system should only use small amounts of electrical
power and some micronutrients in order to operate, and will generate water and
CO2 as by-products. Material that is indigestible becomes residue. However,
because microbes are sensitive to environmental conditions (temperature,
humidity, pH) as well as to the pollutant concentration and type, a biofiltration
system is only suitable for selected applications (15).
Biofilters, usually use beds (media on which microbes live) made from natu-
rally occurring organic materials (yard cuttings, peat, bark, wood chips or
compost) that are slowly consumed by the microbes. Sometimes oyster shells
(for neutralizing acid build-up) and fertilizer (for macronutrients) are mixed
with bed media (see Fig. 2). These organic beds usually can supply most of the
macronutrients needed to sustain the microbial activity. Under normal operating
conditions, the beds must be replaced every 2–5 years (16), depending on the
choice of bed material. However, in handling excessively moist exhaust streams,
sloughing of the beds may clog the discharge system and require replacement of
the beds more often (19).
A fan is used to collect contaminated air from a building or process, but if
the air is too hot, too cold, too dry, or too dirty, it may be necessary to pretreat the
contaminated air stream to obtain optimum conditions before introducing it into
a biofiltration system.
Variables that effect the operation and efficiency of a biofiltration sys-
tem include: temperature, pH, moisture, pollutant mix, pollutant concen-
tration, macronutrient feeding, residence time, compacted-bed media, and
gas channeling. These are crucial variables for which optimum conditions
must be determined, controlled and maintained. Below are variables and lim-
itations that affect the performance of all biofiltration systems, regardless of
process type.
Biofilter Disadvantages:
5. Oxidization Devices
explosive limit (LEL), the gas stream’s explosive properties and the potential for
catalyst overheating may require the addition of dilution air to the waste gas
stream (27).
Whereas the catalytic converter has been used in automobiles to control air
pollutants only since 1975 (5), catalytic oxidation of industrial exhaust emissions
began in the late 1940s, and is a reasonably mature technology (29). Initially it
was used only in circumstances where an extremely serious odor problem was
associated with an industrial system, or where the concentration of organic sol-
vents in the gases to be discharged to the air was high enough that these could
be burned and the heat utilized in the process (23). By the mid-1950s there
were several dozen catalytic incinerators in California, primarily in Los Angeles
county, the first sizable area within the United States to experience a serious air
pollution problem. Early applications of this technology involved some serious
odor, eye irritation, or visible organic emission problems resulting from halogen
poisoning and catalyst fouling (30).
The chemical industry was the first to utilize catalytic oxidation exten-
sively for emission control, building units capable of treating up to 50 m3/s
(100,000 scfm) of exhaust gas containing VOCs. Catalytic systems accounted
for roughly one-fourth of the $200 million market for VOC control systems in
1992, and over one thousand catalytic oxidization devices were in place by the
end of that year (5).
Catalysts. A catalyst has been defined as a substance that increases
the rate at which a chemical reaction approaches equilibrium without becoming
permanently involved in the reaction (31). Thus a catalyst accelerates the
kinetics of the reaction by lowering the reaction’s activation energy (5), ie, by
introducing a less difficult path for the reactants to follow. For VOC oxidation,
a catalyst decreases the temperature, or time required for oxidation, and hence
also decreases the capital, maintenance, and operating costs of the system (see
CATALYSIS).
A key feature of a catalyst is that the catalytic material is not consumed by
the chemical oxidation reactions, rather it remains unaltered by the reactions
that occur generally on its surface and thus remains available for an infinite
number of successive oxidation reactions.
Many chemical elements exhibit catalytic activity (5) which, within limits,
is inversely related to the strength of chemisorption of the VOCs and oxygen,
provided that adsorption is sufficiently strong to achieve a high surface coverage
(32). If the chemisorption is too strong, the catalyst is quickly deactivated as
the active sites become irreversibly covered. If the chemisorption is too weak,
only a small fraction of the surface is covered and the activity is very low (32)
(Fig. 3).
Catalysts vary both in terms of compositional material and physical struc-
ture (33). The catalyst basically consists of the catalyst itself, which is a finely
divided metal (29,32,34); a high surface area carrier; and a support structure.
Three types of conventional metal catalysts are used for oxidation reactions:
single- or mixed-metal oxides, noble (precious) metals, or a combination of the
two (34).
The precious metal or metal oxide imparts high intrinsic activity, the
carrier provides a stable, high surface area for catalyst dispersion, and the
Vol. 2 EMISSION CONTROL, INDUSTRIAL 543
Adsorption Adsorption
too weak too strong
Catalytic activity
Strength of adsorption
mechanical support gives a high geometric surface area for physical support and
engineering design features (35). Only the correct combination of these compo-
nents provides suitable performance and long catalyst life of a properly designed
catalytic system (36).
Metal Oxides. The metal oxides are defined as oxides of the metals occur-
ring in Groups 3–12 (IIIB–IIB) of the periodic table. These oxides, characterized
by high electron mobility and the positive oxidation state of the metal, are gen-
erally less active as catalysts than are the supported nobel metals, but the oxides
are somewhat more resistant to poisoning. The most active single-metal oxide
catalysts for complete oxidation of a variety of oxidation reactions are usually
found to be the oxides of the first-row transition metals, V, Cr, Mn, Fe, Co, Ni,
and Cu.
Noble Metals. Noble or precious metals, ie, Pt, Pd, Ag, and Au, are fre-
quently alloyed with the closely related metals, Ru, Rh, Os, and Ir. These are
usually supported on a metal oxide such as a-alumina, a-Al2O3, or silica, SiO2.
The most frequently used precious metal components are platinum [7440-06-
4], Pt, palladium [7440-05-3], Pd, and rhodium [7440-16-6], Rh. The precious
metals are more commonly used because of the ability to operate at lower tem-
peratures. As a general rule, platinum is more active for the oxidation of paraf-
finic hydrocarbons; palladium is more active for the oxidation of unsaturated
hydrocarbons and CO (34).
Each precious metal or base metal oxide has unique characteristics, and
the correct metal or combination of metals must be selected for each exhaust
control application. The metal loading of the supported metal oxide catalysts
is typically much greater than for nobel metals, because of the lower inher-
ent activity per exposed atom of catalyst. This higher overall metal loading,
however, can make the system more tolerant of catalyst poisons. Some com-
pounds can quickly poison the limited sites available on the noble metal
catalysts (34).
544 EMISSION CONTROL, INDUSTRIAL Vol. 2
pellets are assembled into a packed bed containing large numbers of these pellets
through which the exhaust passes. The honeycomb supports are monolithic
structures having numerous parallel channels through which the exhaust
passes, the channel sizes ranging from 8–50 cells/cm2 of catalyst frontal
area. Each cell has a width opening ranging from 0.29 to 0.13 cm, respectively.
Some commercial honeycombs are available from 1.6 to 100 cells/cm2 (35). The
shape in the individual honeycomb channel is unlimited, eg, circle, square, and
triangle.
Although more expensive to fabricate than the pelleted catalyst, and us-
ually more difficult to replace or regenerate, the honeycomb catalyst is more
widely used because it affords lower pressure losses from gas flow; it is less likely
to collect particulates (fixed bed) or has no losses of catalyst through attrition,
compared to fluidized bed; and it allows a more versatile catalyst bed design
(33), having a well-defined flow pattern (no channeling) and a reactor that can
be oriented in any direction.
The honeycomb structure is either fabricated of ceramic or stainless steel.
The high surface area carrier and catalytic precious metal crystallites are coated
onto the walls of the channels in the honeycomb. The honeycomb catalyst blocks
generally range form 15 to 30-cm square at depths from 5 to 10 cm. These blocks
are packed into larger modules containing many catalyst blocks. Flow through a
honeycomb catalyst structure is shown in Figure 4 (33,37). A typical 30 cells/cm2
honeycomb structure has about 4600 m2 of geometric wall area per cubic meter of
catalyst volume (33). The actual shape of the individual honeycomb channel is
unlimited, eg, it may be circular, square, triangular, etc. In addition, the channel
density can be varied. Commercial honeycombs are available that range from 1.5
to 100 cells/cm2 (10 to 600 cells(in.2)) (35).
Only by using the carrier can the catalyst be sufficiently active because
the majority of applications require 10–100 m2/g of surface area (35). Surface
areas for a typical monolith support structure and a carrier are given (35).
High surface
area
Wall of honeycomb
Exhaust
flow
(a) (b)
Fig. 4. Schematic of the flow through (a) a honeycomb catalyst structure and (b) a cross-
section of a honeycomb channel.
546 EMISSION CONTROL, INDUSTRIAL Vol. 2
Catalytically supported
homogenous reaction
Reaction rate
Mass diffusion
controlled
Surface kinetics
controlled
Temperature
temperature is reached, then the rate of conversion increases rapidly with rising
temperature. This is referred to as the kinetically limited region. Conversion
increases in this region because catalytic reaction rates increase with tempera-
ture, until the catalyst’s normal operating temperature is achieved. Then the
conversion rate increases only slightly with further temperature rise in the
mass-transfer limited region. At some advanced temperature, the conditions
reach a point where thermal oxidation begins to play a role, and the rate of con-
version again increases rapidly.
In the mass-transfer limited region, conversion is most commonly increased
by using more catalyst volume or by increasing cell density, which increases
the catalytic wall area per volume of catalyst. When the temperature reaches
a point where thermal oxidation begins to play a role, catalyst deactivation
may become a concern.
Reaction Rate. The kinetics for a single catalytic reaction can be
modeled as
where rm is the rate of the main reaction; k(T) is the rate constant, a function of
temperature, T; f(C) is the function of reactant and product concentration, C;
and n is the effectiveness factor, which accounts for pore-diffusional resistance
(39). The form of the terms k(T) and f(C) depends on the kinetic model for the
system. Kinetic models for the catalytic oxidation can either be empirical or
mechanistic.
Empirical Models. In the case of an empirical equation, the model is a
power law rate equation that expresses the rate as a product of a rate constant
and the reactant concentrations raised to a power (32), such as
variation with t is predictable from kinetic theory and follows the Arrhenius
equation,
k ¼ AexpðE=RTÞ ð5Þ
Halogenated hydrocarbons
Nitrogenated hydrocarbons
Oxygenated hydrocarbons
1
C4 H8 O þ 5 O2 ! 4 CO2 þ 4 H2 O ð9Þ
2
Vol. 2 EMISSION CONTROL, INDUSTRIAL 549
Process
exhaust
Clean Clean
Burner air Burner air
Preheat section Preheat section
Catalyst Catalyst
Heat Process
exchanger exhaust
(a) (b)
To dryer
Secondary heat
exchanger
Clean
Burner air
Preheat section
Catalyst bed
Process Air
Primary heat exhaust
exchanger
(c)
Fig. 6. Catalytic system designs (26) of (a) basic VOC catalytic converter containing a
preheater section, a reactor housing the catalyst, and essential controls, ducting, instru-
mentation, and other elements; (b) a heat exchanger using the cleaned air exiting the
reactor to raise the temperature of the incoming process exhaust; and (c) extracting
additional heat from the exit gases by a secondary heat exchanger.
only a catalyst bed over which the gas stream passes, usually after some indirect
preheating. The third (Fig. 5c) involves more extensive indirect preheating and
heat exchange. The difference in the three configurations is the method for pre-
heating the gas.
There are two general temperature policies: increasing the temperature
over time to compensate for loss of catalyst activity, or operating at the maxi-
mum allowable temperature. These temperature approaches tend to maximize
destruction, yet may also lead to loss of product selectivity. Selectivity typically
decreases with increasing temperature; faster deactivation; and increased costs
for reactor materials, fabrication, and temperature controls.
Reactor Design. The catalytic reactor is designed to be operated in the
masstransfer controlled catalytic region. The prime design parameter is the geo-
metric surface area. The honeycomb catalyst shows substantial advantage over
other forms because of the high geometric surface areas obtainable with low
pressure drop (35).
Catalyst Selection. The choice of catalyst is one of the most important
design decisions. Selection is usually based on activity, selectivity, stability,
552 EMISSION CONTROL, INDUSTRIAL Vol. 2
mechanical strength, and cost (46). Stability and mechanical strength, which
make for steady, long-term performance, are the key characteristics. The basic
strategy in process design is to minimize catalyst deactivation, while optimizing
pollutant destruction.
Both catalyst space velocity and bed geometry play a role. The gas hourly
space velocity (GHSV) is used to relate the volumetric flow rate to the catalyst
volume. GHSV has units of inverse hour and is defined as the volume flow
rate per catalyst volume.
The size of the catalyst bed depends mainly on the degree of VOC reduction
required (29). VOC destruction efficiencies up to 95% can usually be attained
using reasonable space velocities (29). However, the low GHSVs, and subse-
quently high catalyst volumes required to achieve extremely high (eg, 99%) con-
versions, can sometimes make catalytic oxidation uneconomical. Conventional
bed geometries may be found in the literature (29).
Process Conditions. To effectively design a catalytic control system, the
Manufacturers of Emissions Controls Association recommends the following data
be obtained (5): list of all VOCs present and range of concentration of each, flow
rate of exhaust and expected variability, oxygen concentration in exhaust and
expected variability, temperature of exhaust and expected variability, static
pressure, potential uses for heat recovery, particular performance criteria and/
or regulations to be met, capture efficiency, ie, fraction of all organic vapors
generated by the processes that are directed to the control device, presence of
hydrocarbon aerosols in the effluent exhaust, identity and quantity of all inor-
ganic and organic particulate, amount of noncombustibles, presence of possible
catalyst deactivators, and anticipated start-up/shutdown frequency of the
system.
Pilot Studies. Applications requiring the reduction of VOC emissions
have increased dramatically. On-site pilot tests are beneficial in providing useful
information regarding VOC emission reduction applications. Information that
can be obtained includes optimum catalyst operating conditions, the presence
of contaminants in the gas stream, and the effects of these contaminants.
Catalyst Inhibition. A number of potential applications for catalytic oxida-
tion of organic materials have resulted in serious odor or eye irritation, or visible
emission problems (23). Some of these failures are a result of fouling of the cat-
alyst surface. Others occur because materials such as halogens in the gas stream
interfere with or suppress the activity of the catalyst, or because the substances
react with the precious metals, rendering them permanently inactive. Finally, all
catalysts eventually deteriorate by aging or thermal processes (23).
Many of the exhaust streams that must be purified contain significant
amounts of halogenated organics, such as polychlorinated ethanes and ethylenes
vented in the manufacture of vinyl chloride monomer or released in usage sol-
vents (47). However, the catalysts used in the conventional catalytic oxidation
are severely inhibited by the halogen atoms in these compounds (47). Other
trace contaminants of concern in air streams may include phosphorus-, nitro-
gen-, and sulfur-containing compounds. Whereas gases containing chlorine, sul-
fur, and other atoms can deactivate supported noble metal catalysts such as
platinum, chlorinated VOC can be treated by certain supported metal oxide
catalysts (22).
Vol. 2 EMISSION CONTROL, INDUSTRIAL 553
Coating of
catalyst site Plugging of Inactive compound
catalyst pores Exhaust
Exhaust Active site
gas
gas
Wall of support
(a) (b)
Sintering of Sintering of
alumina precious metal
Oxide/oxide
Exhaust
gas Ceramic oxide
Roughened surface
Ceramic oxide
monolith Strong chemical and
physical bond
Thermal expansion
(C)
Corrosion at interface
Oxide/metal
Exhaust
gas Ceramic oxide
Smooth surface
Metal substrate
monolith Weaker bonds
Thermal expansion
(d)
Fig. 7. Catalyst inhibition mechanisms where (.) are active catalyst sites; , the cata-
lyst carrier; and the catalytic support: (a) masking of catalyst; (b) poisoning of catalyst;
(c) thermal aging of catalyst; and (d) attrition of ceramic oxide metal substrate monolith
system, which causes the loss of active catalytic material resulting in less catalyst in the
reactor unit and eventual loss in performance.
The four basic mechanisms of catalyst decay are shown in Figure 7 (5,33,39).
These are fouling or masking, poisoning, thermal degradation through aging or
sintering, and loss of catalyst material through formation and escape of vapors.
Poisoning and vapor transport are basically chemical phenomena, whereas fouling
is mechanical. Table 5 lists substances that inhibit catalyst activity (5).
554 EMISSION CONTROL, INDUSTRIAL Vol. 2
indicate that H2S adsorbs strongly and dissociates on nickel surfaces. The sulfur
adsorbs essentially irreversibly and over most of the catalyst–metal surface. It
has been observed that SO2 and SO3 also poison catalysts differently (28); SO2
selectively adsorbs on Pt or Pd in an oxidation catalyst, whereas SO3 reacts
with the Al2O3 carrier, forming Al2(SO4)3, which destroys the structure of the
catalyst. The latter can be prevented by using a more inert support such as
SiO2 or TiO2. The former requires a change in operating conditions such as a
higher temperature. When No. 2 fuel oil is used in some gas turbines, the sulfur
compound in the fuel oil can be converted to SO2 at levels of 40–150 ppm in the
exhaust. For such applications, the presence of 100–200 ppm SO2 can require
150–2008C higher temperature for the catalyst to give the same CO conversion
as without SO2.
Toxic heavy metals and ions, eg, Pb, Hg, Bi, Sn, Zn, Cd, Cu, and Fe, may
form alloys with catalytic metals (39). Materials such as metallic lead, zinc,
and arsenic react irreversibly with precious metals and make the surface una-
vailable for catalytic reactions. Poisoning by heavy metals ordinarily destroys
the activity of a precious-metal catalyst (23).
Molecules having unsaturated bonds, eg, CO, NO, HCN, and benzene, may
chemisorb through multiple bonds (39).
Catalysts having improved poison resistance have been developed. Cata-
lysts are available that can destroy chlorine-, fluorine-, or bromine-containing
organic compounds (5).
Thermal Degradation and Sintering. Thermally induced deactivation
of catalysts may result from redispersion, ie, loss of catalytic surface area
because of crystal growth in the catalyst phase (36,39,48) or from sintering, ie,
loss of catalyst-support area because of support collapse (39). Sintering processes
generally take place at high (>5008C) temperatures and are generally acceler-
ated by the presence of water vapor (see Fig. 6c). Another thermal effect is the
transformation of catalytic phases to noncatalytic ones, eg, the reaction of nickel
and alumina to form nickel aluminate (39). Each catalyst has a recommended
temperature window of operation. At temperatures above this window (usually
7608C), sintering can occur.
Loss of Catalyst by Vapor Transport. The direct volatilization of cata-
lytic metals is generally not a factor in catalytic processes, but catalytic metal
can be lost through formation of metal carbonyl oxides, sulfides, and halides in
environments containing CO, NO, O2, and H2S, and halogens (39).
The ceramic oxide carrier is bonded to the monolith by both chemical and
physical means. The bonding differs for a ceramic monolith and a metallic mono-
lith. Attrition is a physical loss of the carrier from the monolith from the surface
shear effects caused by the exhaust gas, a sudden start-up or shutdown causing a
thermal shock as a result of different coefficients of thermal expansion at the
boundary between the carrier and the monolith, physical vibration of the cata-
lyzed honeycomb, or abrasion from particulates in the exhaust air (36)
(see Fig. 6d).
Avoiding Catalyst Deactivation. Catalyst deactivation is more easily
prevented than cured. Poisoning by impurities may be prevented by removing
impurities from the reactants. Carbon deposition and coking may be prevented
by minimizing formation of precursors and manipulating mass-transfer regimes
Vol. 2 EMISSION CONTROL, INDUSTRIAL 557
base metal catalyst, which showed rapid deactivation as a result of sulfur in the
SLA, presumably introduced in the fuel used to fire the heatset dryer. It was
found that the P concentration in the SLA might be as high as 0.16 ppm. The
phosphorus concentration on the deactivated catalyst was found to be 1.4% of
the catalyst. The printer was urged by the supplier to find and eliminate the
cause of the phosphorus contamination of the waste gas entering the catalyst
bed. However, after it was determined that use of phosphorus-containing addi-
tives was crucial to many of the high quality printing jobs, attention was directed
at the catalyst bed itself.
A catalyst with a substantially improved resistance to poisoning by
phosphorus in catalytic oxidation applications was developed. In part, the cata-
lyst in this program permitted printers to use photolighographic technology
without paying an unreasonable cost in terms of frequent replacement of oxida-
tion catalysts.
Catalyst Reactivation. Some catalytic systems are reported to have
operted continuously for >10 years with little or no loss in control efficiency
(5). In most processes catalysts inevitably lose activity, and when the activity
has declined to a critical level, the catalyst needs to be discarded or regenerated
(39). Regeneration is only possible when the deactivation is reversible by chemi-
cal washing or heat treatment or oxidation (36,39).
Thermal Treatment. A thermal treatment for catalyst regeneration
is usually effective when deactivation is a result of coking or masking of the
catalyst surface. Thermal treatment can usually be done on-site, by elevating
the temperature of the catalyst bed by 50–1008C above the normal operating
point and running at this oxidizing condition for a specified limited period of
time (35). The elevated temperature vaporizes or oxidizes the organic compounds
or char that may be masking the catalyst surface.
Physical Treatment. If inspection of the catalyst indicates deposits
again, or if an excessive pressure drop across the catalyst is noted, then the cat-
alytic bed may be lanced, on-site, using compressed air or water until the depos-
its are removed. Abrasion by contact with excessively high pressure from the
compressed air should be avoided (35). If this treatment is combined with heat-
ing, hot spots or overtemperatures that could further deactivate the catalyst
should be avoided (39). In many cases, periodic maintenance, removing the
catalyst bed and blowing or washing off residues, has restored catalyst to original
or near-original activity levels (5).
Chemical Treatment. The most involved regeneration technique is
chemical treatment (35) which often follows thermal or physical treatment,
after the char and particulate matter has been removed. Acid solution soaks,
glacial acetic acid, and oxalic acid are often used. The bed is then rinsed with
water, lanced with air, and dried in air. More involved is use of an alkaline solu-
tion such as potassium hydroxide, or the combination of acid washes and alkaline
washes. The most complex treatment is a combination of water, alkaline, and
acid washes followed by air lancing and drying. The catalyst should not be appre-
ciably degraded by the particular chemical treatment used.
Analyses of a catalyst used in a process involving cloaning products and pig-
ments and achieving a hydrocarbon destruction capacity of only 13% showed
deposition of P, Sn, Pb, and Na contaminants (35). Initial acid treatment
Vol. 2 EMISSION CONTROL, INDUSTRIAL 559
4 NO þ 4 NH3 þ O2 ! 4 N2 þ 6 H2 O ð10Þ
2 NO2 þ 4 NH3 þ O2 ! 3 N2 þ 6 H2 O ð11Þ
The NO reduction is the most important because NO2 accounts for only
5–10% of the NOx in most exhaust gases.
Ammonia oxidation reactions
4 NH3 þ 5 O2 ! 4 NO þ 6 H2 O ð12Þ
4 NH3 þ 3 O2 ! 4 N2 þ 6 H2 O ð13Þ
When sulfur dioxide is also present there are important side reactions in which
SO2 is oxidized to SO3. The main side reaction in the SCR catalyst is the conver-
sion of SO2 to SO3, thus facilitating the reaction above. The SO3 in turn reacts
with ammonia to form ammonium sulfates.
These various reactions should be minimized to avoid plugging the catalyst and
to prevent fouling of the downstream air preheaters, when these components
condense from the gas at the lower temperatures.
The SCR Process. The first step in the SCR reaction is the adsorption of
the ammonia on the catalyst. SCR catalysts can adsorb considerable amounts of
ammonia (65). However, the adsorption must be selective and high enough to
yield reasonable cycle times for typical industrial catalyst loadings, ie, uptakes
in excess of 0.1% by weight. The rate of adsorption must be comparable to the
rate of reaction to ensure that suitable fronts are formed. The rate of desorption
must be slow. Ideally the adsorption isotherm is rectangular. For optimum per-
formance, the reaction must be irreversible and free of side reactions.
It has been suggested that the first step, weak coadsorption of NO and O2 on
a reduced vanadium site, may represent the slow step in the mechanism. Subse-
quent formation of a N2O3-like intermediate could be a fast step, because it is
known that in the gas phase the equilibrium of NONO2 and N2O3 is established
within microseconds (52).
At low temperatures the SCR reactions dominate and nitrogen oxide con-
version increases with increasing temperature. But as temperature increases,
the ammonia oxidation reactions become relatively more important. As the tem-
perature increases further, the destruction of ammonia and generation of nitro-
gen oxides by the oxidation reactions causes overall nitrogen oxide conversion to
reach a plateau then decreases with increasing temperatures. Examples are
shown in Figure 8 (64).
In the SCR process, ammonia, usually diluted with air or steam, is injected
through a grid system into the flue/exhaust stream upstream of a catalyst bed
(55). The effectiveness of the SCR process is also dependent on the NH3 to NOx
ratio. The ammonia injection rate and distribution must be controlled to yield an
1:1 molar ratio. At a given temperature and space velocity, as the molar ratio
increases to approximately 1:1, the NOx reduction increases. At operations above
1:1, however, the amount of ammonia passing through the system increases (56).
This ammonia slip can be caused by catalyst deterioration, by poor velocity dis-
tribution, or inhomogeneous ammonia distribution in the bed.
Types of SCR Catalysts. The catalysts used in the SCR were initially
formed into spherical shapes that were placed either in fixed-bed reactors for
562 EMISSION CONTROL, INDUSTRIAL Vol. 2
100
80
NOx Conversion, %
60
40
20
0
100 200 300 400 500 600
Temperature, °C
Fig. 8. NOx reduction versus temperature for SCR catalysts: ( ) CAT LT-1, ( ) CAT
LT-2, ( ) CAT VNX, and ( ) CAT 2NX (59).
clean gas applications or moving-bed reactors where dust was present. The
moving-bed reactors added complexity to the design and in some applications
resulted in unacceptable catalyst abrasion. As of 1993 most SCR catalysts are
either supported on a ceramic or metallic honeycomb or are directly extruded
as a honeycomb (1). A typical honeycomb block has face dimensions of 150 by
150 mm and can be as long as one meter. The number of cells per block varies
from 20 by 20 up to 45 by 45 (57).
No SCR catalyst can operate economically over the whole temperature
range possible for combustion systems. As a result, three general classes of cat-
alysts have evolved for commercial SCR systems (44): precious-metal catalysts
for operation at low temperatures, base metals for operation at medium tempera-
tures, and zeolites for operation at higher temperatures.
The precious-metal platinum catalysts were primarily developed in the 1960s
for operation at temperatures between 200 and 3008C (1,56,64). However,
because of sensitivity to poisons, these catalysts are unsuitable for many combus-
tion applications. Variations in sulfur levels of as little as 0.4 ppm can shift the
catalyst required temperature window completely out of a system’s operating tem-
perature range (64). Additionally, operation with liquid fuels is further compli-
cated by the potential for deposition of ammonium sulfate salts within the pores
of the catalyst (64). These low temperature catalysts exhibit NOx conversion that
rises with increasing temperature, then rapidly drops off, as oxidation of ammonia
to nitrogen oxides begins to dominate the reaction (see Fig. 7).
The most popular SCR catalyst formulations are those that were developed
in Japan in the late 1970s comprised of base metal oxides such as vanadium
pentoxide [1314-62-1], V2O5, supported on titanium dioxide [13463-67-7], TiO2
(1). As for low temperature catalysts, NOx conversion rises with increasing tem-
peratures to a plateau and then falls as ammonia oxidation begins to dominate
the SCR reaction. However, peak conversion occurs in the temperature range
between 300 and 4508C, and the fall-off in NOx conversion is more gradual
than for low temperature catalysis (64).
A family of zeolite catalysts has been developed, and is being increasingly
used in the United States in SCR applications. Zeolites that can function at
Vol. 2 EMISSION CONTROL, INDUSTRIAL 563
Other problems that can be associated with the high dust plant can include
alkali deterioration from sodium or potassium in the stack gas deposition on the
bed, calcium deposition, when calcium in the flue gas reacts with sulfur trioxide,
or formation and deposition of ammonium bisulfate. In addition, plugging of the
air preheater as well as contamination of flyash and FGD wastewater discharges
by ammonia are avoided if the SCR system is located after the FGD (38).
A significant problem area for initial SCR systems has been the continuous
emission monitoring (CEM) systems. In power plants, all sites equipped with
CEM systems report the highest failure frequency. The CEM systems are the
most labor intensive component, requiring as much as full-time attention from
one technician. At one power plant CEM systems were responsible for 100% of
73 reported SCR system shutdowns (56). As CEM systems continue to improve,
these concerns may disappear.
Nonselective Catalytic Reduction. Hydrocarbons, hydrogen, or carbon
monoxide can be used as reducing agents for NOx in applications where the
exhaust oxygen concentration is low, as it is in fuel rich-burn reciprocating
engines, where it is <1%, and in nitric acid plants, when it is from 2 to 3%.
This approach is called nonselective catalytic reduction (NSCR). In some applica-
tions, the oxygen must be removed from the feed stream prior to the catalyst (52).
An oxygen sensor in the exhaust stream signals the air–fuel delivery system to
adjust the air/fuel ratio so it is just slightly fuel-rich, having enough reducing
agent present to react with all the oxygen and nitrogen oxides (1).
Nonselective catalytic reduction systems are often referred to as three-way
conversions. These systems reduce NOx, unburned hydrocarbon, and CO simul-
taneously. In the presence of the catalyst, the NOx are reduced by the CO result-
ing in N2 and CO2 (55). A mixture of platinum and rhodium has been generally
used to promote this reaction . It has also been reported that a catalyst using pal-
ladium has been used in this application (1). The catalyst operation temperature
limits are 350–8008C, and 425–6508C are the most desirable. Temperatures
>8008C result in catalyst sintering (55). Automotive exhaust control systems
are generally NSCR systems, often shortened to NCR.
Typically NOx conversion ranges from 80 to 95% and there are correspond-
ing decreases in CO and hydrocarbon concentrations. Potential problems asso-
ciated with NSCR applications include catalyst poisoning by oil additives, such
as phosphorus and zinc, and inadequate control systems (37).
6.2. Carbon Monoxide. Carbon monoxide is emitted by gas turbine
power plants, reciprocating engines, and coal-fired boilers and heaters. Carbon
monoxide can be controlled by a precious-metal oxidation catalyst on a ceramic
or metal honeycomb. The catalyst promotes reaction of the gas with oxygen to
form CO2 at efficiencies that can exceed 95%. Carbon monoxide oxidation cata-
lyst technology is broadening to applications requiring better catalyst durability,
such as the combustion of heavy oil, coal (qv), municipal solid waste (qv), and
wood (qv). Research is underway to help cope with particulates and contami-
nants, such as flyash and lubricating oil, in gases generated by these fuels (1,66).
Carbon monoxide conversion is a function of both temperature and catalyst
volume, and increases rapidly beginning at just under 1008C until it reaches a
plateau at 1508C. But, unlike NOx catalysts, >1508C there is little benefit to
further increasing the temperature (64). Above 1508C, the CO conversion is
Vol. 2 EMISSION CONTROL, INDUSTRIAL 565
controlled by the bulk phase gas mass transfer of CO to the honeycomb surface.
That is, the catalyst is highly active, and its intrinsic CO removal rate is exceedingly
greater than the actual gas transport rate (36). When the activity falls to such an
extent that the conversion is no longer controlled by gas mass transfer, a decline of
CO conversion occurs, and a suitable regeneration technique is needed (36).
It has been reported that below 3708C, sulfur oxides reversibly inhibit CO
conversion activity. This inhibition is greater at lower temperatures. CO conver-
sion activity returns to normal shortly after removal of the sulfur from the
exhaust (64). Above 3158C, sulfur oxides react with the high surface area oxides
to disperse the precious-metal catalytic agents and irreversibly poison CO
conversion activity.
Catalyst contamination from sources such as turbine lubricant and boiler
feed water additives is usually much more severe than deactivation by sulfur
compounds in the turbine exhaust. Catalyst formulation can be adjusted to
improve poison tolerance, but no catalyst is immune to a contaminant that
coats its surface and prevents access of CO to the active sites. Between 1986
and 1990 >25 commercial CO oxidation catalyst systems operated on gas turbine
cogeneration systems, meeting both CO conversion (40–90%) and pressure drop
requirements.
6.3. Halogenated Hydrocarbons. Destruction of halogenated hydro-
carbons presents unique challenges to a catalytic oxidation system (65–72).
The first step in any control strategy for halogenated hydrocarbons is recovery
and recycling (65). However, even upon full implementation of economic recovery
steps, significant halocarbon emissions can remain. In other cases, halogenated
hydrocarbons are present as impurities in exhaust streams (65). Impurity
sources are often intermittent and dispersed.
The principal advantage of a catalytic oxidation system for halogenated
hydrocarbons is in operating cost savings. Catalytically stabilized combusters im-
prove the incineration conditions, but still must employ very high temperatures
as compared to VOC combustors; eg, carbon tetrachloride [56-23-5], CCl4, has a
40-fold lower heat of combustion than a typical organic vapor such as toluene
[108-88-3], thus CCl4 requires much more supplemental fuel to burn than do
typical organics (65). Alternatively, the low temperature catalytic oxidation pro-
cess is typically designed for a maximum adiabatic temperature rise of only
2008C. This would correspond to only 1500 ppm of an organic compound in
the exhaust stream. But, with the lower heat of combustion, up to 40,000 ppm
of carbon tetrachloride could be treated in the same temperature rise, or with
less dilution air.
By-Product Formation. The presence of halogenated hydrocarbons dra-
matically increases the yield of aldehydes from the oxidation process (65). For
example, in the partial oxidation of methane on a PdO sponge catalyst (34),
methylene chloride, CH2Cl2, was added in pulses to the inlet gas, which also
contained the methane. Methane oxidation was strongly inhibited and formal-
dehyde was formed. The formaldehyde production continued even after the
methylene chloride addition was stopped, suggesting a strong interaction of
chlorine with the catalyst. However, pulses of pure CH 4 plus oxygen gradually
restored the original activity to the catalyst, indicating that the effect of this
interaction was reversible.
566 EMISSION CONTROL, INDUSTRIAL Vol. 2
It is important to produce HCl rather than elemental chlorine, Cl2, because HCl
can be easily scrubbed out of the exhaust stream, whereas Cl2 is very difficult to
scrub from the reactor off-gas. If the halogenated hydrocarbon is deficient in
hydrogen relative to that needed to produce HCl, low levels of water vapor
may be needed in the entering stream (65) and an optional water injector may
be utilized. For example, trichloroethylene [79-01-6], C2HCl3, and carbon tetra-
chloride require some water vapor as a source of hydrogen (65).
7. Uses
100
90
Conversion of chlorocarbons, %
80 CCl4 CHCl3 CH2Cl2
70
60
50
CH3Cl
40
30
20
10
0
100 150 200 250 300 350 400 450 500 550
Inlet temperature, °C
Fig. 9. Destruction of C1 chlorocarbons over HDC in the presence of 1.5% H2O in air at
15,000 h1 GHSV at STP. Chlorocarbon concentrations are CCl4, 900 ppmv; CHCl3,
500 ppmv; CH2Cl2, 800 ppmv; and CH3Cl, 600 ppmv.
Solvent usage: adhesives, disk manufacture, dry cleaning, fiber glass manu-
facture, food tobacco manufacture, metal cleaning, pharmaceutical, photo
finishing labs, semiconductor manufacture.
Chemical and petroleum processes: cumene manufacture, ethylene oxide
manufacture, acrylonitrile manufacture, caprolactam manufacture, maleic
anhydride manufacture, monomer venting, phthalic anhydride manufac-
ture, paint and ink manufacture, petroleum product refining, petroleum
marketing, resin manufacture, textile processing.
Industrial/commercial processes: aircraft manufacture, asceptic packaging,
asphalt blowing, automotive parts manufacture, breweries/wineries, carbon
fiber manufacture, catalyst regeneration, coffee roasting, commercial char-
broiling, electronics manufacture, film coating, filter paper processing, food
deep frying, gas purification, glove manufacture, hospital sterilizers, pea-
nut and coffee roasting, plywood manufacture, rubber processing, spray
painting, tire manufacture, wood treating.
Engines: diesel engines, lean burn internal combustion, natural gas compres-
sors, oil field steam generation, rich burn internal combustion, gas turbine
power generation.
Cross media transfer: air stripping, groundwater cleanup, soil remediation
(landfills), hazardous waste treatment, odor removal from sewage gases.
of both fuel and solvent fractions (13). The catalytic oxidation of any given com-
pound is generally negatively impacted by the presence of others in mixtures,
and higher catalyst bed operating temperatures are required to achieve adequate
destruction.
Some catalysts exposed to air stripping off-gas were subject to deactivation.
However, using a catalytic oxidizer at a U.S. Coast Guard facility (Traverse City,
Mich.) for the destruction of benzene, toluene, and xylene stripped from the
groundwater, the catalytic oxidization unit operated at 260–3158C, and was
able to achieve 90% destruction efficiency.
7.2. Printing and Graphic Arts. In the graphic arts industry, the cata-
lyst in the oxidizer needs to be monitored regularly because it is susceptible
to contamination by phosphorus from fountain solutions, silica from silicone
gloss enhancer sprays, and chlorides from chlorinated solvents or blanket
wash solutions. Phosphorus and silica accumulate most rapidly on the leading
edge of the catalyst bed, deactivating the catalyst by masking the precious
metals. In a fluidized-bed configuration, the catalyst surface is continually
renewed by abrasion and the problem of masking the catalyst surface with sili-
cones is avoided.
7.3. Chemical Processing. Terephthalic Acid Production. The con-
trol of exhaust from production of pure terephthalic acid (PTA) has been a
challenge (see CARBOXYLIC ACIDS SURVEY) (73). Eight million metric tons of PTA
are produced annually worldwide for use primarily in high grade polyester
fiber production. Based on a para-xylene feedstock, vent gases from the process
contain such by-products as methyl acetate, organic acids, and often methyl bro-
mide. These exhausts have been estimated to total 34,000-m3 worldwide. Histori-
cally, the presence of the methyl bromide limited the use of fixed-bed catalytic
oxidation as a control technology using precious-metal catalysts. Thus base
metal catalysts in fluidized-bed reactors have been the primary catalytic technol-
ogy of choice. In this application, the continuous abrasion of the outer layer of the
catalyst particles exposes a fresh surface of unpoisoned material to the reactants,
allowing the catalyst to effectively treat the exhaust stream.
In the late 1980s, however, the discovery of a noble metal catalyst that
could tolerate and destroy halogenated hydrocarbons such as methyl bromide
in a fixed-bed system was reported (73,74). The products of the reaction were
water, carbon dioxide, hydrogen bromide, and bromine. Generally, a scrubber
would be needed to prevent downstream equipment corrosion. However, if the
focus of the control is the VOCs and the CO rather than the methyl bromide, a
modified catalyst formulation can be used that is able to tolerate the methyl bro-
mide, but not destroy it. In this case the methyl bromide passes through the bed
unaffected, and designing the system to avoid downstream effects is not neces-
sary. Destruction efficiencies of hydrocarbons and CO of better than 95% have
been reported, and methyl bromide destructions between 0 and 85% (73).
Latex Monomer Production. ARI Technologies, Inc. has introduced a cat-
alyst system which, it is claimed, can operate at an average bed temperature of
3708C while achieving conversion efficiency in excess of 99.99% on exhaust
streams from latex monomer production.
Acrylonitrile Manufacture. In the manufacture of acrylonitrile (qv), off-
gases containing from 1–3% of CO plus various hydrocarbons are emitted.
Vol. 2 EMISSION CONTROL, INDUSTRIAL 569
Catalytic beds of platinum-group metals are used to reduce the regulated com-
pounds to acceptable levels. Close attention to bed design is required to prevent
the formation of appreciable quantities of NOx caused by the fixation of combus-
tion-air nitrogen. Some NOx also is produced from fuel nitrogen by oxidation.
Because of the high thermal energy content of the off-gases, considerable heat
recovery is possible in abating acrylonitrile plant emissions (see HEAT-EXCHANGE
TECHNOLOGY).
Vinyl Monomer Manufacturing. Process vent gases containing small
quantities of halogenated hydrocarbons and substantial quantities of nonhaloge-
nated hydrocarbons have been successfully reduced to comply with regulatory
objectives in large-scale laboratory–pilot catalytic fume abaters having satisfac-
tory long-term catalyst performance. The design freedoms offered by precious
metals on ceramic honeycomb support catalysts have been demonstrated in
equipment that utilizes the heat energy resulting from the substantial exotherm
of the nonhalogenated hydrocarbon oxidation to preheat the exhaust gases. Fuel
consumption is thereby minimized.
7.4. Coatings Industries. Surface coating processes produce similar
air pollution problems in a number of different industries.
Can Manufacturing. An internal coating is necessary to protect the
purity and flavor of can contents for beverages or any edible product that
might react with the container metal. Both the exterior decorative and interior
sanitary coatings are applied to the metal surface by rolls or spray guns using a
solvent vehicle. Catalytic oxidation systems are used by the principal can
manufacturers to treat coatings exhaust streams. The can manufacturers’ indus-
try is estimated to utilize more catalysts than any of the other surface coating
industries.
A large number of diverse solvents are used in exterior and interior coatings
in plants for manufacturing both three-and two-piece cans. Most of the organic
solvents are found in the cure-oven exhausts at concentrations of 2–16% of the
lower explosive limit (LEL). The oven exhaust volumes are usually 1–35 m3/s.
When burned, these concentrations of combustibles provide an exotherm of
30–2208C. The heat that is released is used for preheating the incoming effluent
and/or heating the cure oven by recycling the hot, cleaned gases to the supply
blowers or by heating makeup air by heat exchange. A few plants use the heat
of the cleaned exhaust to produce hot water for the two-piece can line washers,
hot air for dry-off ovens, or building space heating. For example, one large can
company utilizes the heat energy contained in the stream, leaving some of
their catalytic fume abaters to supply all the heat energy required by the
oven’s heating zones, which have no burners. The fuel energy supplied to the catalytic
fume abater is less than would be needed to heat the oven if the solvent fumes were
exhausted directly to the atmosphere without use of the fume abater. The exhaust rate
of the oven is adjusted to maintain a solvent concentration of at least 8% of the LEL,
equivalent to a 1108C temperature differential.
The various reaction rate properties of the different solvents influence the
design of a catalytic reactor. For example, for a specific catalyst bed design, an
effluent stream containing a preponderance of monohydric alcohols, aromatic
hydrocarbons, or propylene requires a lower catalyst operating temperature
than that required for solvents such as isophorone and short-chain acetates.
570 EMISSION CONTROL, INDUSTRIAL Vol. 2
BIBLIOGRAPHY
‘‘Exhaust Control, Industrial’’ in ECT 3rd ed., Vol. 9, pp. 511–541, by J. H. Conklin,
D. M. Sowards, and J. H. Kroehling, E. I. du Pont de Nemours & Co., Inc.; in ECT
4th ed., Vol. 9, pp. 1022–1060, by Ronald L. Berglund, The M. W. Kellogg Company;
‘‘Exhaust Control, Industrial’’ in ECT (online), posting date: December 4, 2000, by Ronald
L. Berglund, The M. W. Kellogg Company.
CITED PUBLICATIONS
37. R. E. Kenson, ‘‘Control of Volatile Organic Emissions’’, Bulletin 1015, Series 1000,
Met-Pro Corp., 1981.
38. W. L. Prins and Z. L. Nuninga, Catal. Today 16, 187–105 (1993).
39. C. H. Bartholomew, Chem. Eng. 96 (Nov. 12, 1984).
40. A. C. Frost and co-workers, Environ. Sci. Technol. 25(12), 2065 (1991).
41. D. Ciccilella and B. Holt Environ. Protec., 41 (Sept. 1992).
42. K. J. Herbert, ‘‘Catalysts for Volatile Organic Control in the 1990’s’’, presented at the
1990 Incineration Conference, San Diego, Calif., May 14–18, 1990.
43. M. A. Palazzolo, J. I. Steinmetz, D. L. Lewis, and J. F. Beltz ‘‘Parametric Evaluation
of VOC/HAP Destruction via Catalytic Incineration’’, U.S. Environmental Protection
Agency, Report no. EPA/600/S2-85/041, Research Triangle Park, N.C., July 1985.
44. S. K. Gangwal, M. E. Mulling, J. J. Spivey, and P. R. Caffrey, Appl. Catal. 36, 231 (1988).
45. M. A. Palazzolo, and C. L. Jamgonhian, ‘‘Destruction of Chlorinated Hydrocarbons by
Catalytic oxidation’’, U.S. Environmental Protection Agency, Report EPA/600/82-86/
079, Research Triangle Park, N.C., Jan. 1987.
46. Difford and Spensor, Chem. Eng. Prog. 71, 31 (1975).
47. G. R. Lester, ‘‘Catalytic Destruction of Hazardous Halogenated Organic Chemicals’’,
presented at the 82nd Annual AWMA Meeting & Exhibition, Anaheim, Calif., June
25–30, 1992.
48. S. E. Wanke and P. C. Flynn, Cat. Rev. Sci. Eng. 12, 93 (1975).
49. G. R. Lester and J. C. Summers ‘‘Poison-Resistant Catalyst for Purification of Web
Offset Press Exhaust’’, presented at the Air Pollution Control Association, 81st
Annual Meeting, Dallas, Tex., June 19–24, 1988.
50. R. G. Nicolli, ‘‘Overview of NOx Chemistry and Emissions Control Technology’’, pre-
sented at 16th AESF/EPA Pollution Prevention and Control Conference, Feb 1995.
51. U.S. Environmental Protection Agency, Alternative Control Techniques Document—
NOx Emisisons from Industrial/Commercial/Institutional (ICI) Boilers, Office of Air
Quality Planning and Standards, March 1994 (EPA-453/R-94-022).
52. F. Luck and J. Roiron, Catal. Today 4, 205 (1989).
53. E. Cichanowicz, Power Eng., 36 (Aug. 1988).
54. T. A. Melnick, R. Sommerlad, R. Payne, T. Sommer, ‘‘Low NOx Burner Modifications
for Cost Effective NOx Control’’, Forum 2000 on Cutting NOx Emissions, Institute of
Clean Air Companies, March 2000 (Arlington, VA).
55. L. M. Campbell, D. K. Stone, and G. S. Shareef, Sourcebook: NOx Control Technology
Data, EPA Report NO. EPA600/S2-91/029., Washington, D.C., Aug. 1991.
56. G. S. Shareef, D. K. Stone, K. R. Ferry, K. L. Johnson, and K. S. Locke ‘‘Selective
Catalytic Reduction NOx Control for Small Natural Gas-Fired Prime Movers’’, paper
92-136.06, in Ref. 16.
57. J. R. Donnelly and B. Brown ‘‘Joy/Kawasaki Selective Reduction DE-NOX Technol-
ogy’’, paper No 89-96B.6, in Ref. 32.
58. H. Gutberlet and B. Schallert, Catal. Today 16, 207 (1993).
59. R. D. Walloch, Oil Gas J., 39 (June 13, 1988).
60. R. Craig, G. Robinson, and P. Hatfield, ‘‘Performance of High Temperature SCR
Catalyst System at Unocal Science and Technology Division’’, paper 92-109-08, in
Ref. 15.
61. D. Borsare, R. Hanson, E. Angelini, R. Peterson, and J. Buschmann, ‘‘Recent Experi-
ences in SCR System Design’’, Forum 2000 on Cutting NOx Emissions, Institute of
Clean Air Companies, March, 2000 (Arlington, VA).
62. D. L. Champagne, Gas Turb. World, 20 (Nov.–Dec. 1987).
63. R. Rodenhausen, ‘‘Case Study: Choosing Selective Catalytic Reduction as a Preferred
Technology for the Destruction of NOx’’, Environmental Progress, Winter 1999,
pp. 260–266.
Vol. 2 EMISSION CONTROL, INDUSTRIAL 575
RONALD L. BERGLUND
Terracon
PART III:
PETROCHEMICALS
PETROCHEMICAL
FEEDSTOCKS
1. Introduction
576
Kirk-Othmer Encyclopedia of Chemical Technology. Copyright John Wiley & Sons, Inc. All rights reserved.
10.1002/0471238961.1605201801210218.a01.pub2
Vol. 2 PETROCHEMICAL FEEDSTOCKS 577
fuel, also impact petrochemical feedstock availability, relative cost, and choice
of alternative feedstocks.
It is convenient to divide the petrochemical industry into two general sec-
tors: (1) olefins and (2) aromatics and their respective derivatives. Olefins are
straight- or branched-chain unsaturated hydrocarbons, the most important
being ethylene (qv), [74-85-1] propylene (qv) [115-07-1], and butadiene (qv)
[106-99-0]. Aromatics are cyclic unsaturated hydrocarbons, the most important
being benzene (qv) [71-43-2], toluene (qv) [108-88-3], p-xylene [106-42-3], and
o-xylene [95-47-5] (see XYLENES). There are two other large-volume petro-
chemicals that do not fall easily into either of these two categories: ammonia
[7664-41-7] and methanol [67-56-1]. These two products are derived primarily
from methane [74-82-8] (natural gas).
Olefins are produced primarily by thermal cracking of a hydrocarbon feed-
stock that takes place at low residence time in the presence of steam in the tubes
of a furnace. In the United States, natural gas liquids derived from natural gas
processing, primarily ethane [74-84-0] and propane [74-98-6], have been the
dominant feedstock for olefins plants, accounting for 60–70% of ethylene pro-
duction. Most of the remainder has been based on cracking naphtha or gas oil
hydrocarbon streams that are derived from crude oil. Naphtha is a hydrocarbon
fraction boiling between 40 and 1708C, whereas the gas oil fraction boils
between 310 and 4908C. These feedstocks, which have been used primarily
by producers with refinery affiliations, account for most of the remainder of
olefins production. In addition, a substantial amount of propylene and a small
amount of ethylene are recovered from gas streams produced in petroleum
refinery cracking units.
In most of the rest of the world, the olefins industry was originally based on
naphtha feedstocks. Naphtha is the dominant olefins feedstock in Europe and
Asia. In the mid 1980s, several large olefins complexes were built outside the
United States primarily based on ethane feedstocks, most notable in western
Canada, Saudi Arabia, and Scotland. During the 1990s, new large olefins com-
plexes based on ethane feedstocks were also built in Kuwait, Qatar, and Malaysia.
In each case, the driving force was the production of natural gas, perhaps asso-
ciated with crude oil production, which was in excess of energy demands.
Recently, several large olefins complexes have been designed based on the
use of condensates, a heavier natural gas liquid. Condensates are mixtures of
liquid hydrocarbons, typically mixtures of pentane (C-5) and heavier compo-
nents, which are recovered from natural gas either at the wellhead or in gas pro-
cessing plants. Condensates usually are classified as either ‘‘light’’ or ‘‘heavy,’’
according to their boiling range, and light condensates are sometimes classified
as ‘‘natural gasolines’’. The typical properties of light condensates— >65% par-
affins content and a boiling range of 30–1808C—make them desirable feedstocks
for steam cracking to produce olefins.
Heavy condensates have a boiling range between 30 and 4008C and have
characteristics of light crude oils. Like light crudes, heavy condensates contain
naphtha, kerosene, gas oil, and minor quantities of heavier fractions, which
makes them attractive feedstocks for refinery primary distillation, as well as
olefins and aromatics production. A large U.S. Gulf Coast olefins complex
based on naphtha feedstocks derived from heavy condensate streams started
578 PETROCHEMICAL FEEDSTOCKS Vol. 2
up in 2001 (1) and another world-scale olefins complex based on heavy conden-
sate feedstocks is now under construction in China (2).
Since the early 1980s many olefin plants in the United States were designed
to have substantial flexibility to consume a wide range of feedstocks. Most of the
flexibility to use various feedstocks is found in plants with associated refineries,
where integrated olefins plants can optimize feedstocks using either gas liquids
or heavier refinery streams. Companies whose primary business is the produc-
tion of ethylene derivatives, such as thermoplastics, tend to use ethane and pro-
pane feedstocks that minimize by-product streams and maximize ethylene
production for their derivative plants.
Flexibility allows the operator to pick and choose the most attractive feed-
stock available at a given point in time. The steam-cracking process produces not
only ethylene, but other products as well, such as propylene, butadiene, buty-
lenes (a mixture of monounsaturated C-4 hydrocarbons), aromatics, etc. With
ethane feedstock, only minimal quantities of other products are produced. As
the feedstocks become heavier (ie, as measured by higher molecular weights
and boiling points), increasing quantities of other products are produced. The
values of these other coproduced products affect the economic attractiveness,
and hence the choice of feedstock.
Aromatics are produced primarily from two sources. The most important in
the United States is refinery catalytic reformer operations. Catalytic reformers
are high temperature catalytic dehydrogenation processes that convert
naphthenes into aromatics in order to increase the octane level, and hence the
gasoline blending quality of the stream being processed. Reformers produce
large quantities of the primary aromatic chemicals. Benzene, toluene, and a
mixed-xylene stream are subsequently recovered by extraction or extractive
distillation using a solvent (see BTX PROCESSING). Recovery of p-xylene from a
mixed-xylene stream requires a further processing step of crystallization and fil-
tration or adsorption followed by desorption on beds of molecular sieves.
The second source of aromatics is pyrolysis gasoline from olefins plants.
Pyrolysis gasoline is one of the by-product streams produced in olefin plants
when cracking heavier feedstocks such as naphtha or gas oil. Pyrolysis gasoline
contains hydrocarbon fractions from C-5 olefins through C-9 aromatics. Recovery
of aromatics from this stream is by extraction and fractionation, just as in recov-
ery from reformate. Pyrolysis gasoline is a more important aromatics feedstock
in Europe and Asia than in the United States. A reason is that naphtha is the
dominant feedstock for olefins production in these regions of the world, and
less gasoline is produced in the refineries of these regions; hence, there is less
reforming capacity. Otherwise there are only minimal differences in aromatics
production in various regions of the world.
2. Olefin Feedstocks
olefins since that time (3). Olefin plants are complex and costly, with both com-
plexity and cost increasing as the design feedstock increases in heaviness (mole-
cular weight) from the lightest feed (ethane) to the heaviest feed (gas oil).
Heavier feeds produce a wider array and larger quantities of products that
require more complex separation and recovery processes. A new world-scale
ethylene plant with a capacity of 1000
106 kg/year based on naphtha feedstock
will cost $750 million (4). The total plant cost, including off-site facilities, is
$1.1 billion. Originally olefins plants were designed to consume a specific feed-
stock, such as ethane or naphtha, or a narrow range of feedstocks, most com-
monly ethane–propane mixtures. Starting 1980, producers in the United
States that were based on heavier feedstocks such as naphtha or gas oil began
to introduce feedstock flexibility to consume a wide range of feedstocks in their
olefin plants. This was done to allow the operator to pick and choose the most
attractive feedstock at a given point in time in order to maximize profitability.
Olefin plants in the United States have incorporated more feed flexibility than
those in Europe, and are therefore able to take advantage of a wider array of
potential feedstocks.
Most of the flexibility to use various feedstocks is found in plants with asso-
ciated refineries, where integrated olefins plants can optimize feedstocks using
either gas liquids or heavier refinery streams. Figure 1 illustrates the many lin-
kages between an olefin plant and a refinery that is typical of integrated com-
plexes in the United States.
Companies whose primary business is the production of ethylene derivatives,
such as thermoplastics, tend to use ethane and propane feedstocks that minimize
by-product streams and maximize ethylene production for their derivative plants.
Table 1 provides a summary of the 2003 production quantity and value of primary
olefins petrochemicals and olefin feedstocks in the United States.
Refinery–olefin plant
Cat product linkages
cracker
Propylene
Kerosene Butadiene
Crude Naphtha
Benzene
Gas oil Olefin plant
Condensate–NGL Fuels
Gasoline
Purchased Pitch
feeds
Condensate Heavy Gas oil Jet fuel
NGL’s gas oil Hydrocracker
Refinery return
The cracking process produces not only ethylene, but other products as well.
With ethane feedstock, only minimal quantities of other products are produced.
As the feedstocks become heavier, increasing quantities of other products are
produced. Most operators characterize these other products as either coproducts
or by-products. In this article, coproducts are defined as other primary olefins
(propylene and butadiene), and primary aromatics (benzene, toluene, and
xylenes). By-products are everything else ranging from C-4 monoolefins to
tars. The principal distinction between coproducts and by-products is that copro-
ducts are the more desired and valuable products where operators often adjust
feedstocks and operating conditions to maximize or minimize production,
depending on the economics at the point in time. The values of these coproducts
and by-products affect the economic attractiveness, and hence the choice of feed-
stock. Table 2 presents typical yields for a new plant employing the latest tech-
nology from various feedstocks as well as a characterization of the feedstock
density and boiling range. Yields from older existing plants differ somewhat
from these yields, generally being lower in ethylene production (see ETHYLENE).
It can be seen that in the case of ethane, the yields of ethylene are the great-
est and only minimal quantities of other products are produced. In the case of gas
oil, the yields of ethylene are the lowest and large quantities of heavier copro-
ducts and by-products are produced. When cracking heavier feedstocks, eg,
naphtha, the severity of the cracking can be varied by adjusting the cracking
temperature. Yields of ethylene can be maximized at the expense of propylene
582 PETROCHEMICAL FEEDSTOCKS Vol. 2
400
350
CURRENT DOLLARS PER TON
300
250
200
150
100
50
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
ETHANE
PROPANE
LIGHT NAPHTHA
Fig. 2. Quarterly olefin feedstock prices, 1985–2003, for (*) ethane (^) propane, and
(~) light naphtha.
Vol. 2 PETROCHEMICAL FEEDSTOCKS 583
1.8
1.6
PRICE RATIO BY WEIGHT
1.4
1.2
0.8
0.6
0.4
0.2
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
PROPANE
LIGHT NAPHTHA
Fig. 3. Price ratio of various feeds to ethane, 1985–2003; (^) propane and, (~) light
naphtha.
East leading up to the war in Iraq. This situation more typically occurs during
summer months when natural gas prices are weaker due to lower seasonal
demand and prices for gas liquids are seasonally lower. Conversely, gasoline
demand and prices are seasonally higher in the summer due to the peak driving
season, and the cost of naphtha reflects this strength.
During the January–March 2003 period, the cash costs of producing ethy-
lene from naphtha were 12¢ per kg of ethylene higher than the cash costs from
ethane and about nine cents per kg of ethylene higher than the cash costs from
propane. Therefore, gas liquids were the most economical source of feedstock
during that period. Table 4 also presents the cash costs of ethylene from the
same feedstocks during the April–June 2003 period when ethane became the
least economical feedstock and naphtha was the preferred feedstock. Very low
inventories of natural gas resulted in sharply higher natural gas prices, also for-
cing up ethane prices in order to provide an incentive to recover it from natural
gas. Thus an ehtylene producer with feedstock flexibility would vary the feed-
slate over time in order to achieve the lowest possible costs. The historical feed-
stock sources for ethylene production in the United States are presented in
Table 5. The shift to heavier feedstocks due to the high natural gas prices during
2003 is also shown.
The corresponding historical feedslate for ethylene production in the United
States is presented in Table 6.
Most new plants built in the United States during the 1990s were based pri-
marily on ethane and propane feeds because the companies constructing these
facilities are predominantly thermoplastic producers requiring ethylene, but
most industry observers expect that olefin feedslates in the United States will
become progressively heavier in the future. Additional supplies of ethane will be
limited, barring significant new discoveries of natural gas, and higher natural gas
prices relative to prices for liquid fuels are expected to reduce the future attractive-
ness of ethane as a feedstock. Some additional propane will probably be imported
from foreign sources. Adequate supplies of naphtha and gas oil are available to
meet the remainder of growth demands for ethylene. On a worldwide basis, gas
liquids accounted for 40% of ethylene production in 2003 and are expected to con-
tinue at that level in the future because of additional recovery of gas liquids asso-
ciated with crude oil and natural gas production outside of the United States. As
the United States olefins feedslate becomes somewhat heavier, the feedslate in the
rest of the world will become somewhat lighter.
Refinery gas streams are a significant source of propylene and a minor source
of ethylene in the United States. About one-half of the propylene produced in the
United States is recovered directly from refinery gas streams, which are largely
produced in fluid catalytic cracking (FCC) units (FLUID CATALYTIC CRACKING (FCC)
UNITS, REGENERATION). FCC units are designed to convert heavy gas oil fractions
into gasoline blend stocks and are used more extensively in the United States
than in Europe and Asia. When used for producing a fuel, this propylene stream
would be converted into heavier gasoline blend stocks by alkylation or dimerization
processes or blended into liquefied petroleum gas (LPG). Refinery-grade propylene
is usually in the 60–70% purity range, with the principal impurity being propane.
Propylene prices are normally such that it is more attractive for refiners in close
proximity to propylene consumers to sell or transfer the refinery-grade propylene
to operators for the purpose of purification to chemical-grade quality (minimum
92% propylene) or polymer-grade quality (99þ% propylene). Other refiners also
have facilities to perform this purification and thus are able to sell one of the higher
purity grades of propylene. Purification consists of fractionation and removal of
trace impurities. There are a few chemical processes that can utilize refinery-
grade propylene without purification; cumene [98-82-8], 2-propanol [67-63-0], and
higher olefins such as nonene (C-9 monoolefins) and tetramer (C-12 monoolefins).
3. Aromatics Feedstocks
Aromatics are produced primarily from two sources, refinery catalytic reformer
operations and recovery from the pyrolysis gasoline fraction from olefin plants.
The most important source in the United States is refinery catalytic reformer
operations. Catalytic reformers are high temperature catalytic dehydrogenation
processes that convert naphthenes contained in virgin naphthas into aromatics
in order to increase the octane value, and hence the gasoline blend value of
the stream being processed. Reformers produce large quantities of the primary
aromatics; benzene, toluene, and isomers of xylene (see BTX PROCESSING).
The preferred feedstocks for aromatics production are naphthas containing
high concentrations of the naphthene precursors to benzene, toluene, and xylenes.
586 PETROCHEMICAL FEEDSTOCKS Vol. 2
the composition of reformer feed and reformer severity to produce less benzene. In
addition, reformulated gasoline accounts for about one-third of U.S. sales since its
use is only required in metropolitan areas with the most severe summertime
ozone levels. Therefore, benzene levels >1% are still permissible in the other
two-thirds of gasoline sold in the United States, making it possible for refiners
to preferentially use low benzene streams for reformulated gasoline blends.
4. Other Feedstocks
5
DOLLARS PER GJ
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
CRUDE OIL
NATURAL GAS
Fig. 4. Prices of natural gas (~) and crude oil (^), 1985–2003. To convert ¢/GJ to
¢/BTU, multiply by 1.055
106.
588 PETROCHEMICAL FEEDSTOCKS Vol. 2
1.4
NATURAL GAS TO CRUDE OIL RATIO
1.2
0.8
0.6
0.4
0.2
0
1985 1987 1989 1991 1993 1995 1997 1999 2001 2003
Fig. 5. Fuel price ratio of natural gas at Henry Hub to WTI crude oil, 1985–2003.
5. Alternative Feedstocks
Alternative feedstocks for petrochemicals have been the subject of much research
and study over the past several decades, but have not yet become economically
attractive. Chemical producers are expected to continue to use fossil fuels for
energy and feedstock needs for many decades in the future. The most promising
sources that have received attention include coal, tar sands, oil shale, and
biomass. Near-term advances in coal-gasification technology offer the potential
to replace oil- and gas-based feedstocks in selected applications (10) (see FEED-
STOCK, COAL CHEMICALS). Processes for the production of olefins from alternative
feedstocks are in the development phase with no commercial plants operating.
Several processes using methanol as a feedstock are considered to be the most
promising and are economically attractive since the methanol could be produced
from natural gas in locations having cheap, abundant natural gas reserves.
Lurgi’s MTP technology has been demonstrated in relatively small scale.
This technology converts methanol into propylene at high selectivity. The
UOP/HYDRO MTO process converts crude methanol into an olefinic gas rich
in ethylene and propylene. The ratio of propylene to ethylene produced can be
adjusted within the MTO process. The process yield performance for the UOP/
HYDRO MTO process as presented in public domain literature is presented in
Table 7 (10). The effect of operational severity enables the ethylene/propylene
ratio to be changed. UOP has reported that this ratio can be varied from 1.5 to
0.64 by changing reactor conditions.
Because oil and gas are not renewable resources, at some point in time alter-
native feedstocks will become attractive; and as natural gas and petroleum prices
rise, the timetable for commercialization of alternative feedstocks will accelerate.
Of the alternatives, only agricultural-based materials are a renewable resource
(see BIOMASS ENERGY). Currently, the only chemical produced from a natural pro-
duct in large-volume commercial quantities is ethanol via fermentation of natural
sugar and starches. The vast majority of fermentation ethanol is used for fuel
Vol. 2 PETROCHEMICAL FEEDSTOCKS 589
Cargill also has other programs in place to produce industrial chemicals from
renewable agricultural resources. Some of these projects include conversion of glu-
cose to 3-hydroxypropionic acid (3-HP). 3-HP, using conventional industrial chem-
istry, can be converted to acrylic acid, PDO, and other specialty polyesters;
production of urethane polyols from vegetable oils and using metathesis chemistry
to cleave and modify unsaturated natural oils to olefins and difunctional olefins.
Another promising bioconversion process that has the potential to compete
with petrochemical processes is the conversion of glucose to succinic acid. Several
U.S. Government sponsored laboratories (Argonne National Laboratory, Pacific
Northwest National Laboratory and National Renewable Energy Laboratory)
have collaborated to develop fermentation technology to convert sugars to succinic
acid. Succinic acid can undergo hydrogenolysis to 1,4-butanediol (BDO). In 2003,
the global market for BDO is estimated to be over 800,000 metric tons per year.
Besides the U.S. based development in this area, Mitsubishi Chemical, working
together with Ajinomoto in Japan, is also investigating a similar approach.
BIBLIOGRAPHY
‘‘Petroleum Products’’ in ECT 1st ed., Vol. 10, pp. 161–177, by V. L. Shipp, Socony-
Vacuum Oil Co., Inc.; ‘‘Petroleum Products’’ under ‘‘Petroleum’’ in ECT 2nd ed., Vol. 15,
pp. 77–92, by G. B. Gibbs and H. L. Hoffman, Hydrocarbon Processing; ‘‘Petroleum
(Products)’’ in ECT 3rd ed., Vol. 17, pp. 257–271, by H. L. Hoffman, Gulf Publishing
Co.; ‘‘Feedstocks’’ in ECT 3rd ed., Vol. 9, pp. 831–845, by F. A. M. Buck, King, Buck, &
Associates, and M. G. Marbach, Shell Chemical Co.; ‘‘Feedstocks, Petrochemicals’’ in
ECT 4th ed., Vol. 10, pp. 349–360, by Darryl C. Aubrey, Sacred Heart University Chem
Systems, Inc.; ‘‘Feedstocks, Petrochemicals’’ in ECT (online), posting date: December 4,
2000, by Darryl C. Aubrey, Sacred Heart University Chem Systems, Inc.
CITED PUBLICATIONS
JOHN T. BOEPPLE
Nexant ChemSystems
Vol. 2 HYDROCARBONS 591
HYDROCARBONS
1. Saturated Hydrocarbons
Table 3. Enthalpy of Formation, Entropy, and free Energy of Formation for Saturated
Hydrocarbons as Gases (25 8C) (3)
DH0, kJ/mol S0, J mol1 K1 DG0, kJ/mol
Methane 74.898 186.313 50.828
Ethane 84.724 229.646 32.908
Propane 103.916 270.090 23.505
Cyclopropane þ53.382 237.810 104.503
n-Butane 126.232 310.326 17.166
Cyclobutane þ27.256 265.569 110.741
2-Methylpropane 134.606 294.834 20.934
n-Pentane 146.538 349.179 8.374
Cyclopentane 77.079 293.076 38.895
2-Methylbutane 154.577 343.820 14.654
2,2-Dimethylpropane 166.090 306.599 15.240
n-Hexane 167.305 388.661 0.293
2-Methylpentane 174.422 380.789 5.024
3-Methylpentane 171.743 380.036 2.135
2,2-Dimethylbutane 185.685 358.474 9.923
2,3-Dimethylbutane 177.897 366.010 4.103
n-Heptane 187.945 428.058 þ8.039
n-Octane 208.586 467.038 16.412
n-Dodecane 291.066 622.912 50.074
n-Hexadecane 373.588 778.829 83.820
n-Icosane 456.068 934.745 117.398
594 HYDROCARBONS Vol. 2
Table 4. Critical Data and Molar heat Capacities of Saturated Hydrocarbons (2)
Compound Tc, 8C pc, MPa cp, * J mol1 K1
Methane 82.6 4.60 35.74
Ethane 32.28 4.88 52.67
Propane 96.7 4.25 73.56
Cyclopropane 125.1 5.57 56.27
n-Butane 152.0 3.80 97.51
Cyclobutane 187 4.99 72.26
2-Methylpropane 135.0 3.65 96.88
n-Pentane 196.5 3.37 120.3
Cyclopentane 238.5 4.50 82.98
2-Methylbutane 187.2 3.38 118.9
2,2-Dimethylpropane 160.6 3.20 121.7
n-Hexane 234.2 3.01 143.2
2-Methylpentane 224.3 3.01 144.3
3-Methylpentane 231.2 3.12 143.2
2,2-Dimethylbutane 215.6 3.08 142.0
2,3-Dimethylbutane 226.8 3.13 140.6
n-Heptane 267.0 2.74 166.1
n-Octane 295.6 2.49 189.0
n-Nonane 321.4 2.29 211.8
n-Decane 344.4 2.10 234.7
n-Undecane 365.6 1.97 257.6
n-Dodecane 385.1 1.82 280.5
n-Tridecane 402.6 1.72 303.4
n-Tetradecane 418.7 1.62 326.3
n-Pentadecane 433.6 1.52 349.2
n-Hexadecane 447.4 1.42 372.0
n-Heptadecane 460.2 1.32 394.9
n-Octadecane 472.1 1.22 417.8
n-Nonadecane 483 1.12 440.7
n-Icosane 494 1.12 463.6
*Based on 25 8C and ideal gas state.
carbon and hydrogen. They are, however, kinetically stable at ambient tempera-
ture. Thermal decomposition of saturated hydrocarbons proceeds stepwise by
loss of hydrogen or hydrocarbon fragments with concomitant formation of indust-
rially useful unsaturated cracking products, such as acetylene, olefins, or aromatic
hydrocarbons.
On ignition, mixtures of saturated hydrocarbons and oxygen or air may
lead to combustion or explosion, depending on the ratio of hydrocarbon and
oxygen. Such reactions can be initiated either by imposed ignition or, if the
ignition temperature of the mixture is exceeded, by self-ignition. These reac-
tions are the basis for the use of hydrocarbons as heating- and engine fuel.
Pertinent fuel properties of individual saturated hydrocarbons are summarized
in Table 4.
1.3. Production. By far the largest amount of saturated hydrocarbons is
obtained from the natural sources natural gas and petroleum, either by isolation
or by suitable conversion reactions. Additional sources include various products
derived from coal processing. A number of saturated hydrocarbons, unavailable
from natural sources, are produced by special synthesis or by conversion
processes.
Vol. 2 HYDROCARBONS 595
gas see, Gas Production, acetylene, olefins, aromatics), or for the recovery of
industrially important alkanes and cycloalkanes.
From the liquefied petroleum gas fraction, propane, n-butane, and i-butane
are isolated by distillation. n-Pentane and i-pentane can be isolated from the
liquid components of natural gas or from light gasoline (naphtha) either by mole-
cular sieve separation or by superfractionation (10). Steam-cracking of light
gasoline provides, among other products, cyclopentadiene (see Cyclopentadiene
and Decyclopentaden), benzene (see Benzene), and 1,3-butadiene (see Buta-
diene); these can, in turn, be converted into the C5-, C6-, C8-, and C12-cycloalk-
anes. Catalytic reforming of heavy gasoline provides C6- to C8-aromatics,
which can subsequently be converted into cyclohexane (see Cyclohexane), e.g,
the C7- and C8-aromatics are hydrodealkylated, and the ensuing benzene subse-
quently hydrogenated. From the higher boiling petroleum, gas-oil, and wax dis-
tillate fractions, mixtures of homologous n-alkanes are isolated by techniques,
such as molecular sieve separation or urea extractive crystallization. Paraffin
waxes can also be isolated by precipitation with suitable solvents.
Higher boiling petroleum distillates and distillation residues can be con-
verted into mixtures of lower molecular mass hydrocarbons by hydrocracking
(11,12). From such mixtures, saturated C3 – C5-hydrocarbons and a broad
spectrum of higher n-alkanes can be recovered.
Vol. 2 HYDROCARBONS 597
Cyclizations that lead to small (C3-, C4-) rings are not industrially impor-
tant, although various laboratory methods are available (14). Examples include
the coupling of terminal carbon atoms by elimination of suitable a,o-substituents
from acyclic compounds, e.g, a,o-dihalides, o-halogenated ketones, esters, or
nitriles (Eq. 1), addition of carbenes, to olefins to give cyclopropanes (Eq. 2) and
[2 + 2] -cycloadditions of suitable unsaturated compounds, such as olefins,
allenes, or acetylenes, to give cyclobutanes (Eq. 3) (15):
ð1Þ
ð2Þ
ð3Þ
feed materials can vary from methane to crude oil. Medium temperature
pyrolysis at 750–900 8C in the presence of inert diluents such as steam
(steam-cracking) or hydrogen (hydropyrolysis) yields ethylene (see Ethylene).
A wide range of hydrocarbons, e.g, ethane, liquefied petroleum gas, naphtha,
and gas oil may serve as raw material, and increasing proportions of C3- to
C5-olefins, C4- and C5-dienes, and C6- to C8-aromatics are coproduced with
increasing molecular mass of the feed material. Pyrolysis of petroleum waxes
at 500–600 8C affords mixtures of homologous olefins with predominantly
terminal double bonds.
Catalytic Dehydrogenation. Catalytic dehydrogenation of acyclic satu-
rated hydrocarbon atoms yields mono- and/or diolefins with the same number of
carbon atoms. Industrial processes include the production of butenes and buta-
diene from n-butane, of isoprene from i-pentane, and C6- to C19-monoolefins,
mainly with internal double bonds, from the corresponding n-alkanes (22).
More recently, improved processes for the production of ethylene, propene, and
600 HYDROCARBONS Vol. 2
O2
CH4 ! CH3 CH3 þ CH2 ¼ CH2 ð4Þ
H2 O
O2
CH4 þ CH3 X ! CH2 ¼ CH X ð5Þ
H2 O
hydrogen and carbon monoxide. n-Alkenes in the range C10 to C19 are used
for the production of biodegradable detergents: acid-catalyzed reaction of
C10- to C13-n-alkenes with benzene, followed by sulfonation and neutralization
of the ensuing alkylbenzene sulfonic acids, affords linear alkylbenzenesulfo-
nate (LABS) detergents. Oxo-reaction of C10- to C19-n-alkenes yields C11- to
C20-alcohols, which are converted into nonionic detergents by reaction
with ethylene oxide, or into anionic detergents by reaction with sulfur
trioxide or chlorosulfonic acid. Liquid-phase oxidation of n-alkanes with
oxygen in the presence of boric acid yields predominantly secondary alcohols.
The latter can be dehydrated to the corresponding n-alkenes. The C12- to
C19-alcohols are also used for the production of both ionic and nonionic
detergents.
Chlorination of n-alkanes yields mono- or polychlorinated alkanes, depend-
ing on the conditions. Mixtures of C10- to C13-monochloro-alkanes are used for
the production of linear alkylbenzenes as raw materials for LABS-detergents.
For this, the chloroalkanes are either directly reacted with benzene or they are
dehydrochlorinated to the corresponding n-alkenes, which are then used to
alkylate benzene. Polychlorinated n-alkanes are used in a number of special
applications.
n-Alkanes in the range of C10 to C21 are suitable substrates for the pro-
duction of single cell proteins by fermentation. The microorganisms used selec-
tively attack n-alkanes, so that industrial processes have been developed in
which either isolated n-alkanes or the appropriate petroleum distillates
containing n-alkanes (petroleum-, gas oil-, or lube oil fractions) are used as
starting materials.
Solid n-alkanes (paraffin waxes) are used in a variety of applications, both
as such and as feed materials in cracking, oxidation and chlorination.
Cycloalkanes. Of the various accessible cycloalkanes, only cyclohexane
and cyclododecane are used industrially on a large scale; cyclooctane is used
on a relatively small scale. The strained C3- and C4-cycloalkanes are not
produced industrially. Cyclopentane is present in natural hydrocarbon sources
and is formed as a byproduct in the processing of cracking products, e.g,
by hydrogenation of cyclopentadiene or cyclopentene in the C5-fraction
of steamcracked products. No important industrial use exists for cyclopentane.
Cyclooctane is accessible by catalytic hydrogenation of 1,5-cyclooctadiene.
Liquid-phase oxidation of the former with air yields 1,8-octanedicarboxylic acid,
which is used on a moderate scale for the production of nylon 68.
Cyclododecane is produced by liquid-phase hydrogenation of 1,5,9-cyclo-
dodecatriene over nickel catalysts. Liquid-phase oxidation of cyclododecane with
air in the presence of boric acid yields mixtures of cyclododecanol and cyclodode-
canone in high selectivity. Further oxidation of such alcohol – ketone mixtures
with nitric acid gives 1,12-dodecanedicarboxylic acid, used for the production
of polyamides, polyesters, and synthetic lubricating oils. Catalyzed dehydrogena-
tion of alcohol-ketone mixtures in the liquid phase yields cyclododecanone, which
is then converted into laurinlactam and nylon 12.
Bicyclo [4.4.0] decane (decalin) is obtained by catalytic hydrogenation
of naphthalene. Liquid-phase oxidation of decalin with oxygen gives the
604 HYDROCARBONS Vol. 2
2. Olefins
dissolve well in most organic solvents, e.g, alcohols, ethers, and aromatic
hydrocarbons. Some physical properties of important monoolefins are listed in
Table 6.
Chemical Properties. The chemistry of monoolefins is dominated by the
reactive double bond. Some important reactions of olefins are summarized in
Table 7.
Most addition reactions proceed by an electrophilic mechanism via carbo-
nium ions. Chlorination and bromination yields the trans adducts. Hydrohalo-
genation generally follows Markovnikov’s rule to give a product containing
halogen linked to secondary carbon. Addition of sulfuric acid to olefins affords
alkyl sulfates, which are readily hydrolyzed to alcohols; the reaction corresponds
to the acid-catalyzed addition of water. Ethanol, isopropylalcohol, and tert-butyl
alcohol are produced industrially by this method from ethylene, propene, and iso-
butene, respectively.
Hydrogenation of olefins to alkanes requires a catalyst: Raney nickel is
most frequently used industrially, although platinum-group metals are also
used.
Epoxidation of ethylene may be carried out with oxygen in the presence of a
silver catalyst. Higher olefins are converted to epoxides using hydroperoxides or
peracids. A related reaction is the hydroxylation of olefins with hydrogen perox-
ide in glacial acetic acid: an epoxide is initially formed, which subsequently
reacts with water in the acidic medium to form a glycol.
Another olefin oxidation is the Wacker oxidation of ethylene to acetalde-
hyde. The stoichiometric reaction of ethylene with water and palladium(II) chlor-
ide yields acetaldehyde and palladium metal. Addition of copper(II) chloride
converts the stoichiometric reaction into a catalytic process, wherein the
copper(II) chloride reoxidizes the palladium to palladium(II) chloride the result-
ing copper(I) chloride is reoxidized to copper(II) chloride by atmospheric oxygen
in hydrochloric acid solution.
Ozonolysis is used mainly for structure determination of olefins, but may
also be used preparatively. Reaction of olefins with carbon monoxide and addi-
tional reagents (e.g, hydrogen, water, alcohols) leads to a plethora of valuable
products (e.g, aldehydes, alcohols, carboxylic acids, esters). These reactions pro-
ceed under catalysis by transition metals, e.g, hydroformylations (Oxo process)
with rhodium- or cobalt carbonyls.
Sulfonation of olefins leads to alkenesulfonic acids, which are used as deter-
gents. Hydroboration is used to generate primary alcohols via trialkylboranes. In
contrast to the direct addition of water, this pathway is regioselective to yield
the anti-Markovnikov products.
Olefins may undergo a number of reactions resulting in new C-C bonds.
With aromatic compounds alkylaromatics are formed by the Friedel – Crafts
reaction; with isoalkanes chain-lengthened, branched hydrocarbons, which are
used in motor fuels, are formed. Olefins may also react with themselves or
each other to form oligomers and polymers.
The industrial use of the higher monoolefins is discussed in Section uses of
Higher Olefins.
Production of Higher Olefins. Olefins are synthesized on a laboratory
scale by a variety of routes: e.g, selective hydrogenation of alkynes; dehydration
Table 6. Physical Properties of Monoolefins
CAS registry
Olefin number mp, 8C bp, 13 8C d20 n20
Ethylene [74-85-1] 169.5 103.8 0.5699 1.363
(103.8 8C) (100 8C)
Propene [115-07-1] 185.3 47.7 0.6095 1.3640
(47.4 8C) (50 8C)
0.5139
1-Butene [106-98-9] 185.3 6.26 0.6356 1.3791
(15.1 8C) (25 8C)
cis-2-Butene [590-18-1] 138.9 3.7 0.6447 (0 8C) 1.3783
(0 8C)
trans-2-Butene [624-64-6] 105.5 0.88 0.6269 (0 8C) 1.3696
(20 8C)
Isobutene [115-11-7] 140.4 6.9 0.6398 (20 8C) 1.3776
1-Pentene [109-67-1] 167.2 29.97 0.6405 1.3718
cis-2-Pentene [627-20-3] 151.3 37.1 0.6554 1.3828
trans-2-Pentene [646-04-8] 140.2 36.4 0.6482 1.3798
2-Methyl-1-butene [58718-78-8] 137.5 31.2 0.6504 1.3778
3-Methyl-1-butene 168.5 20.1 0.6332 1.3640
2-Methyl-2-butene [26760-64-5] 133.6 38.6 0.6625 1.3874
1-Hexene [592-41-6] 140.0 63.5 0.6732 1.3878
cis-2-Hexene [7688-21-3] 141.8 68.8 0.6969 1.3976
trans-2-Hexene [4050-45-7] 133.5 67.5 0.6780 1.3935
cis-3-Hexene [7642-09-3] 137.8 66.5 0.6794 1.3951
trans-3-Hexene [13269-52-8] 113.4 68.0 0.6778 1.3937
1-Heptene [592-76-7] 119.6 93.8 0.6993 1.3998
cis,trans-2-Heptene 109.5 98.2 0.7034 1.4041
cis-3-Heptene [7642-10-6] 136.6 95.8 0.7030 1.4059
trans-3-Heptene [14686-14-7] 136.3 95.7 0.6981 1.4043
1-Octene [111-66-0] 102.7 121.3 0.7160 1.4086
cis-2-Octene [7642-04-8] 100.5 125.6 0.7243 1.4146
trans-2-Octene [13389-42-9] 87.8 124.5 0.7184 1.4132
cis-3-Octene [14850-22-7] 126 122.7 0.7223 1.4140
trans-3-Octene [14919-01-8] 110.0 123.3 0.7149 1.4129
cis-4-Octene [7642-15-1] 118.1 122.5 0.7212 1.4150
trans-4-Octene [14850-23-8] 93.8 122.3 0.7136 1.4120
1-Nonene [124-11-8] 81.6 146.8 0.7292 1.4157
2-Nonene [2216-38-8] 149.4 0.7407 1.4201
3-Nonene [20063-77-8] 148.2 (102 kPa) 0.7334 1.4191
4-Nonene [2198-23-4] 148.0 0.7313 1.4200
1-Decene [872-05-9] 66.6 170.6 0.7408 1.4215
2-Decene [6816-17-7]
3-Decene [19348-37-9] 173–173.5 0.7447 1.4221
4-Decene [19689-18-0] 170.6 0.7404 1.4243
cis-5-Decene [7433-78-5] 112 169.5 (98.5 kPa) 0.7445 1.4252
trans-5-Decene [7433-56-9] 73 170.2 0.7412 1.4235
1-Undecene [821-95-4] 49.1 192.67 0.75033 1.4261
5-Undecene [4941-53-1] 192.2 0.7511 1.4289
1-Dodecene [25378-22-7] 33.6 213.3 0.7583 1.4300
6-Dodecene [7206-29-3] 212.9 (99.3 kPa) 0.7597 1.4328
1-Tridecene [2437-56-1] 22.2 104 (1.33 kPa) 0.7654 1.4335
1-Tetradecene [1120-36-1] 12 125 (1.85 kPa) 0.7726 1.4367
1-Pentadecene [13360-61-7] 3.8 133.5 (1.33 kPa) 0.7763 1.4389
1-Hexadecene [629-73-2] 4.0 155 (2.0 kPa) 0.7811 1.4412
1-Heptadecene [6765-39-5] 10.7 159 (1.33 kPa) 0.7860 1.4436
1-Octadecene [112-88-9] 18.0 176 – 178 (1.85 kPa) 0.7888 1.4450
1-Icosene [3452-07-1] 28.5 151 (0.2 kPa) 1.4440
(30 8C)
1-Heneicosene [1599-68-4] 35.5 134 (0.005 kPa) 0.7985 1.4510
1-Triacontene [18435-53-5] 62–63 218 (0.07 kPa) 0.7620 (100 8C)
a
At 101.3 kPa, unless otherwise noted.
Vol. 2 HYDROCARBONS 607
Fig. 3. Simplified flow diagram for a combination of the UOP PACOL process and
benzene alkylation. a) Pacol reactor; b) Separator; c) Stripper; d) Alkylation reactor.
610 HYDROCARBONS Vol. 2
The growth step occurs at ca. 100 8C under 10 MPa ethylene pressure. In
the second stage, the high temperature displacement reaction, the alpha-olefins
are displaced by ethylene at ca. 300 8C and 1 MPa. The composition of the product
mixture corresponds to a Poisson distribution. This process suffers from the dis-
advantage of requiring stoichiometric quantities of reaction components, i.e,
industrial plants would require large quantities of aluminum alkyls. Conse-
quently, Gulf Oil and Ethyl Corporation developed two process variants that
require less triethylaluminum.
The Gulf Oil Process. In the Gulf process, only catalytic quantities of
triethylaluminum are used and growth step and elimination reactions occur
simultaneously in the same reactor (49). The reaction is carried out at <200 8C
and ca. 25 MPa. The product mixture consists of a Schulz – Flory distribution of
linear olefins. A typical composition of individual fractions from C6 – C18 is listed
in Table 9.
Vol. 2 HYDROCARBONS 611
The Gulf process produces markedly purer a-olefins than can be obtained
with paraffin cracking (see Table 8). The major impurities in the a-olefins
(trade name: Gulftenes) obtained by the Gulf process are ca. 1.4% paraffins;
with increasing carbon number, an increasing amount of b-branched a-olefin is
also formed. The occurrence of these vinylidene and 2-ethyl compounds may be
rationalized by consecutive reactions of a-olefins with the aluminum alkyls, as
follows:
The Gulf Oil Co. uses the Gulf process to produce linear a-olefins in the
C4 – C30 range at their plant in Cedar Bajou, Texas. Mitsubishi Chemical Indus-
tries uses the same process at Mizushima.
Ethyl Process. The Ethyl process combines a stoichiometric and a catalytic
stage (50). This modified Ziegler process is capable of controlling the chain length
of the resulting a-olefins much more precisely, because the shorter chain olefins
may be subjected to further chain growth.
The Ethyl process is illustrated schematically in Figure 4. Ethylene is first
oligomerized using catalytic quantities of triethylaluminum in analogy to the
Gulf process. The resulting products are fractionated (b) in Fig. 4) yielding frac-
tions containing C4, C6 – C10, and C12 – C18. The higher boiling fractions may be
used directly, whereas the shorter a-olefins, especially C4, are subjected to trans-
alkylation with long chain aluminum alkyls (c) in Fig. 4). This reaction releases
the desired long-chain a-olefins; short-chain trialkylaluminum compounds are
simultaneously formed. The aluminum trialkyls are separated in the second dis-
tillation step (Fig. 4, (d)) and transformed in another reactor (e) in Fig. 4) into
longer chain alkyls by stoichiometric reaction with ethylene. The latter are
then recycled to the transalkylation.
The Ethyl Corporation has operated a plant in Pasadena, Texas using this
process since 1970; the initial capacity was 110 000 t/a. The process permits 95%
conversion of ethylene into higher a-olefins, a conversion efficiency not achiev-
able by the single-step catalytic Gulf Oil process. This advantage is however,
counterbalanced by a much lower product quality of the long-chain a-olefins.
The C16 – C18 cut consists of only 63% linear a-olefins (Table 10), the remainder
being mainly b-branched a-olefins and internal olefins.
612 HYDROCARBONS Vol. 2
Fig. 4. Typical flow diagram of Ethyl process. A) Catalytic step; B) Stoichiometric step;.
a) Catalytic reactor; b) First distillation; c) Transalkylation reactor; d) Second distillation;
e) Stoichiometricreactor.
Table 10. Composition of Fractions (wt%) Obtained From the Ethyl Process
C6 C8 C10 C12 C12 – C14 C14 – C16 C16 – C18
Linear a-olefins 97.5 96.5 96.2 93.5 87.0 76.0 62.7
Internal olefins 0.6 1.2 1.6 1.5 4.2 5.0 8.2
b-Branched a-olefins 1.9 2.3 2.2 5.0 8.8 19.0 29.1
Alkanes 0.1 0.6 0.3 0.4 0.4 0.4 0.8
Vol. 2 HYDROCARBONS 613
Table 11. Composition of a-Olefin Fractions (wt%) Obtained From the SHOP Process
C6 C8 C10 C12 C14 C16 C18
All monoolefins 99.9 99.9 99.9 99.9 99.9 99.9 99.9
Linear a-olefins 97.0 96.5 97.5 96.5 96.0 96.0 96.0
Branched olefins 1.0 1.0 1.0 2.0 2.5 2.5 2.5
Internal olefins 2.0 2.4 1.0 1.5 1.5 1.5 1.5
Dienes <0.1 <0.1 <0.1 <0.05 <0.05 <0.05 <0.05
Aromatics <0.1 <0.1 <0.1 <0.05 <0.05 <0.05 <0.05
Alkanes <0.1 <0.1 <0.1 <0.05 <0.05 <0.05 <0.05
nðCnþ2 olefinsÞ
K¼
nðCn olefinsÞ
(see Fig. 5) the mixture is separated into the desired C12 – C18 a-olefins and into low
(C4 – C10) and high-boiling (>C20) fractions. The C4 – C10 fraction can then be
further fractionated, if desired, into individual compounds, for example, those
used as comonomers for the manufacture of linear low density polyethylene
(LLDPE). However, the C4 – C10 fraction and the C 20 fractions are generally
combined for isomerization and subsequent metathesis. These two steps require
only moderate reaction conditions: 80 – 140 8C, 0.3 – 2 MPa. The heat of reaction
of both steps is quite low; because in isomerization only intramolecular migration
of double bonds occurs, in metathesis carbon – carbon bonds are simultaneously
cleaved and reformed. Isomerization takes place in the liquid phase in the presence
of magnesium oxide catalyst (58). About 90% of the terminal olefins are converted
to internal olefins in this reaction.
The subsequent joint metathesis of the internal, low, and high molecular
mass olefins, mainly over heterogeneous rhenium or molybdenum catalysts,
yields a mixture of internal olefins having a new chain length distribution, i.e,
a random mixture of olefins of even and odd numbers of carbon atoms.
The products of metathesis are distilled and the C11 – C14 fraction, repre-
senting ca. 10 – 15% of the mixture, is separated for production of alcohols
used in the detergent industry (see Section Uses of Higher Olefins).
The composition of the internal C11 – C14 olefins produced by the SHOP pro-
cess is given in Table 12; more than 99.5% is monoolefins, with only small
amounts of dienes, aromatics, and paraffins.
Although the double bonds of the olefins at the beginning of metathesis are
distributed almost at random over the carbon backbone, they are shifted in the
Vol. 2 HYDROCARBONS 615
end product toward the chain end, a result of the high concentration of
short-chain olefins (2-butene, 2-hexene, 3-hexene) in the metathesis feed.
Distillation of the metathesis products also yields low- and high-boiling
components in addition to the C11 – C14 fraction. One of the great advantages
of the SHOP process is that these fractions do not have to be discarded but
can be recycled. The fraction consisting of compounds with less than 11 carbon
atoms is returned directly to the metathesis reactor; the fraction containing ole-
fins with more than 14 carbon atoms is again subjected to isomerization before it
reaches the metathesis stage. By combining oligomerization, isomerization, and
metathesis, the SHOP process is capable of converting almost all the ethylene
feed into olefins having the desired number of carbon atoms.
An interesting alternative of the SHOP process described above is the
metathesis of the isomerized C 20 fraction together with a ten-fold molar excess
of ethylene (58). Olefins, predominantly a-olefins, within the desired carbon
range may be prepared in the presence of a rhenium catalyst (20% Re2O7 on alu-
minum oxide).
The first SHOP processing plant was built in 1977 by Shell Oil Corporation
at Geismar, Louisiana.
The mechanism of the highly selective ethylene oligomerization with nickel
chelate catalysts has been largely elucidated by the research group of W. KEIM
(52,59–63). A chelated nickel hydride complex is formed initially. This complex
reacts with ethylene (Fig. 7) to form alkyl nickel complexes. Linear a-olefins
are eliminated from the alkyl nickel complexes; the nickel hydride species is
formed simultaneously to reenter the catalytic cycle.
The SHOP process is by far the most important method for oligomerization
of ethylene involving transition-metal catalysts. To complete the picture, two
other processes will be described below, although they have not yet reached com-
mercial importance.
Exxon Process. The Exxon Corporation has developed a process in
which ethylene is converted to high molecular mass olefins in the presence of a
catalyst consisting of a transition-metal component and a soluble alkyl alumi-
num chloride (64). Titanium chloride or an alkoxy titanium chloride are favored
616 HYDROCARBONS Vol. 2
The reaction takes place in the liquid phase without solvent. The pressure
in the reactor is adjusted so that the C3 compounds remain liquid at the operat-
ing temperature. The product mixture leaving the reactor head is fed to a vessel
where the catalyst is neutralized with alkali. The product (trade name: Dimate)
is then washed with water and freed of any remaining propane and propene on a
stabilizing column.
The Dimersol X process is similar, although the starting materials are
butenes or a mixture of propene and butenes. The products formed are, accord-
ingly, isooctenes or mixtures of C6 – C8 olefins. The first Dimersol X plant was
started in 1980 by Nissan Chem. in Kashima, Japan. A typical starting material
for the Dimersol X process is the raffinate II obtained from the C4 cut of the
steam cracker, after removal of butadiene and isobutene. The resulting C8 olefins
can be converted by Oxo-synthesis into nonanols, which are used for the synth-
esis of plasticizers (dinonyl phthalate).
A more recent development is the Dimersol E process, which employs the
exhaust gases of fluid catalytic crackers (FCC) that contain ethylene and pro-
pene. A gasoline fraction can be obtained from this product after hydrogenation.
The first experiments with cracker exhaust gases from the Elf cracker near Fey-
zin (France) began in 1984.
By 1982, IFP had licensed 16 Dimersol plants, eight of which were in pro-
duction. In addition to IFP, other industrial and academic groups have investi-
gated the transition metal-catalyzed oligomerization of lower olefins (72–80). The
observation that in the propene dimerization with nickel-based Ziegler catalysts
the addition of phosphines may significantly affect the isohexene product distri-
bution is of interest. With the catalyst [NiBr(Z3-C3H5)(PCy3)] – EtAlCl2 (Cy = cy-
clohexyl) a remarkable turnover of 800 000 mol propene per mol nickel h1 could
be achieved. A turnover of 60 000 s1 was calculated for operation at room tem-
perature. This value is of an order of magnitude usually achieved only by enzyme
reactions.
Oligomerization of Propene and Butene by Acid Catalysis. Mineral
acids such as phosphoric or sulfuric acid can also be used as catalysts for the
oligomerization of olefins; the major products are branched olefins. Alkylben-
zenes were once prepared from tetramers of propenes (tetrapropylene). This
reaction is now rarely used in Europe because of the low biodegradability of
the alkylbenzenesulfonates obtained.
Oligomerization of isobutene leads to the formation of diisobutene, triisobu-
tene, and tetraisobutene; after hydrogenation, these may be added to gasoline to
improve the octane rating (polymer gasoline, oligomer gasoline). In the cold acid
process, isobutene-containing C4 fractions are treated with 60 – 70% sulfuric acid
at room temperature. Isobutene dissolves in the sulfuric acid as tert-butyl sul-
fate. This solution is then oligomerized at 100 8C. The products, mainly diisobu-
tenes, separate as a second phase. In the hot acid process, the n-butenes of the C4
fraction are also converted, forming i-C4 – n-C4 cooligomers.
The oligomerization can also be performed with poly(styrene sulfonic
acid) resins as catalysts. This process is used by Erdölchemie – Dormagen
using the raffinate I (C4 fraction after extraction of butadiene). The reaction
is performed in loop reactors at 100 8C and 1.5 – 2.0 MPa with residence
times of 20 – 60 min. The product is composed of 36% diisobutene, 22% C8
618 HYDROCARBONS Vol. 2
In Europe, Asia, South Africa, and South America linear alkylbenzenes are
produced from C10 – C15 alpha-olefins. In 1982, 420 000 t of LAB were synthe-
sized in Western Europe from ca. 80 000 t of linear olefins.
Vol. 2 HYDROCARBONS 619
The double bonds and the hydroxyl groups contribute to the excellent water
solubility of the AOS compounds. They are effective laundry agents even in hard
water and at low concentration and are used increasingly in liquid soaps and cos-
metics. In order to be used in such products the AOS must be of high quality,
especially with regard to color; this is only possible with a-olefins that contain
no more than traces of dienes.
Although AOS compounds have been known since the 1930s, they gained
commercial importance only in the late 1960s. In AOS represents an interesting
alternative to alkyl sulfates and alkyl ether sulfates in cosmetic products.
Synthesis of Bromoalkanes and Derived Products. An important
reaction of a-olefins is radical hydrobromation to give primary bromoalkanes.
These compounds are valuable intermediates for, e.g, thiols, amines, amine
oxides, and ammonium compounds:
Thiols are produced by this route in Europe and in the United States by
Pennwalt Corp. and Phillips Petroleum Co. They are used in the synthesis of her-
bicides, pesticides, pharmaceuticals, in the textile industry, or as polymer regu-
lators.
In the U.S. alkyldimethylamine (ADA) from linear a-olefins were directly
converted to fatty amine oxides (FAO) or quaternary ammonium salts (Quats).
Fatty amine oxides are nonionic surfactants that, because they are biodegrad-
able and do not irritate the skin, are incorporated into numerous products,
e.g, into rinsing agents. Quats are produced by alkylation of amines, e.g, with
620 HYDROCARBONS Vol. 2
a-Olefin epoxides have also been used in polymer chemistry, e.g, for the
modification of epoxy resins. United States producers of a-olefin epoxides
include: Union Carbide, Viking, and Dow Chemical.
Another use of a-olefins is conversion into secondary monoalkylsulfates,
which are readily degradable surfactants useful in wetting agents and household
detergents. They are synthesized by reaction with concentrated sulfuric acid at
low temperature (10 – 20 8C) and with short reaction times (5 min):
Whereas the cyclic 1 : 1 product vinylcyclobutane and the C10 oligomers cyclode-
cadiene and 1,4,9-decatriene are not yet used commercially, the linear oligomer,
1,4-hexadiene, is produced as a monomer for the production of ethylene – pro-
pene – diene (EPDM) elastomers. Homogeneous catalysis with rhodium or palla-
dium complexes leads to trans-1,4-hexadiene; cobalt and iron catalysts yield
predominantly the cis-1,4-product. Rhodium catalysts are used in commercial
production.
Ethylidenenorbornene, also a comonomer for EPDM polymers, is also acces-
sible via cooligomerization. One of the routes employs the Diels – Alder reaction
starting with butadiene and cyclopentadiene, the other couples norbornadiene
and ethylene in the presence of a nickel catalyst. Vinylnorbornene is formed as
an intermediate in both routes.
3. Alkylbenzenes
mp, 8C 96.04
bp, 8C 152.39
d20. g/cm3 0.86179
n20
D 1.49145
Specific heat (25 8C), J g1 K1 1.96
Heat of vaporization, kJ/mol
at 25 8C 44.23
at 152.4 8C 37.74
Tcrit, K 631
pcrit, MPa 3.21
Vcrit, cm3/mol 428
Flash point, 8C 43
Ignition temperature, 8C 425
For more thermodynamic data, see (111). Explosion limits for the system
cumene vapor–oxygen–nitrogen at standard pressure and at 0.56 MPa are
630 HYDROCARBONS Vol. 2
given in (112). For cumene–air mixtures the explosion limits are 0.8–6.0 vol%
cumene, corresponding to a cumene concentration of 40–300 g/m3 (113). Cumene
forms azeotropes with water, aliphatic carboxylic acids, ethylene glycol ethers,
cyclohexanol, and cyclohexanone.
Production. Cumene is produced from benzene [71-43-2] and propene
[115-07-1] in the presence of an acidic catalyst:
Two processes are employed industrially (see also Acylation and Alkylation).
In the UOP process a fixed bed consisting of polyphosphoric acid on a silica sup-
port is used as the catalyst (114). The reaction takes place mainly in the gas
phase at 200 8C and 2 MPa. Pure propene or propene with a propane content
5% is used as raw material, together with benzene (min. purity 99.5%) in up to
12-fold molar excess; this leads to high distillation costs in the product work-up.
The yield of cumene is 97% based on benzene and 90% based on propene (115).
The losses are caused by overalkylation and oligomerization of the propene. The
high-boiling residue contains 50–60% diisopropylbenzenes (o- : m- : p- ratio
1 : 2 : 2–3) and numerous other byproducts. It is not possible to recycle these pro-
ducts to the alkylation reactor, because polyphosphoric acid has no transalkyla-
tion properties.
The second industrial process works in the liquid phase at 50–80 8C; alumi-
num chloride or the liquid double compound formed from aluminum chloride,
(isopropyl-)benzene, and hydrogen chloride is employed as the catalyst. This
complex is separated from the product stream, as a heavy phase and is returned
to the alkylation reactor. Except for a small purge stream higher alkylated
byproducts are recycled to the alkylation step, whereby essentially complete
use of both benzene and propene bound in di- and triisopropylbenzenes is
achieved by transalkylation. The yield of cumene, referred to benzene and pro-
pene, is 99% and 98%, respectively. The UOP process is used predominantly.
Higher selectivity and the lower energy requirements, however, are now making
the aluminum chloride process more attractive (115).
Uses. The production capacity for cumene in Western Europe, the United
States, and Japan amounts to 2106, 2106, and 4105 t/a, respectively. Cumene
is used almost exclusively for the production of phenol [108-95-2] and acetone
[67-64-1]. This process is based on the discovery in 1944 that cumene hydroper-
oxide [80-15-9], obtained by oxidation of cumene with oxygen, can be cleaved
under the catalytic action of strong acids:
Both reactions are exothermic; the enthalpy of reaction amounts to 121 and
255 kJ/mol, respectively.
Pure a-methylstyrene can be obtained as a byproduct of the phenol process.
Cumene hydroperoxide is used as a radical initiator for the copolymerization of
styrene with butadiene and acrylates, and also for the radical cross-linking of unsa-
turated polyester resins. Addition of cumene hydroperoxide to a-methylstyrene
Vol. 2 HYDROCARBONS 631
mp 26–27 8C
bp
101.3 kPa 261–262 8C
1.33 kPa 125.5 8C
d20
4 1.006
Flash point ca. 130 8C
Enthalpy of formation
(room temperature) 109.6 kJ/mol
Enthalpy of combustion
(20 8C) 41.2103 kJ/kg
Biphenyl is obtained as the pot residue, following removal of gas and distil-
lation of benzene and toluene; about 1 kg of biphenyl is obtained per 100 kg ben-
zene. The biphenyl can be enriched to a purity of 93–97% by distillation.
Thermal Dehydrocondensation of Benzene. Until the early 1970s, biphenyl
was produced exclusively by specific thermal dehydrocondensation of benzene
(155). This process yields a higher quality product than the toluene route, e.g,
distilled material with a purity >99.5%; varying quantities of terphenyl isomers
and several higher homologues are also obtained.
When, in the early 1960s, terphenyl was proposed to be used as coolant and
moderator in nuclear reactors (207), numerous reports were published of
attempts to enhance the yield of terphenyl in dehydrocondensation reactions.
Increasing the temperature or residence time was ruled out because of increased
carbonization. These procedures also enhanced the biphenyl fraction, so that the
latter had to be recycled to the reactor (162,166), (208). Recycling a concentration
of up to 30% biphenyl has been described (162). A practical level appears to be
using benzene with 10% biphenyl (177). Attempts have also been made to recycle
the extremely high melting p-terphenyl, after isolation from the reaction
0.015 ppm pseudocumene; see (241) for more details. An analysis of the drinking
water of Cincinnati, Ohio, in 1980, revealed a concentration of 45 ng/L hemimel-
litene, 127 ng/L pseudocumene, and 36 ng/L mesitylene (242). Trimethylben-
zenes are toxic to aquatic organisms; LC50 values between 10 and 100 ppm
have been reported (243).
The TLV (TWA) level for trimethylbenzene is 25 ppm (244). The MAK for Tri-
methylbenzenes (all isomers (1,2,3–;1,2,4–;1,3,5–) has been set at 20 ppm. (245).
The United States toxicity assigment ranges from low to moderately toxic for pseu-
documene, to highly toxic for the isomer mixture; disturbances in the central ner-
vous system and abnormal blood pictures have been reported (243).
Tetramethylbenzenes. The oral toxicity of tetramethylbenzenes is low;
LD50 >5000 mg/kg (rat, oral). Durene is additionally classified as intravenously
highly toxic; LD50 180 mg/kg (i.v., mice). Tetramethylbenzenes cause mild skin
reactions (243). Two values are cited for the odor threshold concentration of dur-
ene: 0.083 and 0.087 mg/m3 (241).
Hexamethylbenzene. Hexamethylbenzene has low oral toxicity; LDLo
5000 mg/kg (rat). It is, however, suspected of causing neoplastic effects.
Addition of nitromethane to hexamethylbenzene can lead to an explosive
reaction (243).
Diethylbenzenes. Diethylbenzene is more toxic than monoethylbenzene
(246). DEB is emitted into the environment from engine fuel and heating oil
(242). Diethylbenzenes are toxic to aquatic organisms; the LC50 is between 10
and 100 mg/L (243). In the cell multiplication inhibition test with protozoa, the
toxicity threshold is <10 mg/L (241). In the Federal Republic of Germany DEB
is classified as an aquatic hazard; aquatic hazard class 2. DEB causes mild to
moderate irritation of the eye and mucous membranes. The oral toxicity is low:
LDLo 5000 mg/kg (rat) (243). The vapor possesses an anesthetic action and may
cause headache, vertigo, or vomiting; the odor threshold concentration is
<10 ppm (247).
Triethylbenzene. The LC50 of triethylbenzene (aquatic organisms) is
between 100 and 1000 mg/L. The oral toxicity is low: LDLo 5000 mg/kg (rat).
Implant experiments with hexaethylbenzene in mice have raised the suspicion
that the compound may be carcinogenic (243).
Ethyltoluene. Ethyltoluene can, as with aromatic hydrocarbons, be
detected in the environment (241). The oral toxicity of 2- and 4-ethyltoluene is
low: LDLo 5000 mg/kg (rat) (243).
Cumene. Cumene appears to be orally as nontoxic as propylbenzene, it is
even less so by inhalation (239,248,249). Cumene has a prolonged depressant
effect on the central nervous system (250).
The MAK for cumene has been set at 250 mg/m3, corresponding to a cumene
vapor concentration in air of 50 mL/m3. The vapor pressure of cumene at 20 8C is
0.5 kPa. There is a risk of cumene being absorbed through the skin (251). Inhalation
of cumene vapor leads to the delayed appearance of a long lasting narcotic effect (252).
5.4. Biphenyls and Polyphenyls. Biphenyls. Biphenyl dust or
vapor is irritating to the eye and mucous membrane at a concentration as low
as 3–4 ppm (246), and to the skin after extended exposure. A concentration of
>5 mg/m3 for long periods is considered a health hazard; systemic toxic effects
were elicited in humans by a concentration with inhalatory exposure maxima
644 HYDROCARBONS Vol. 2
of 128 ppm (253,254). The olfactory threshold is ca. 0.06–0.3 mg/m3. Some toxi-
city data are listed below (255–257):
REFERENCES
150. M. McEwen, E. Wiederhold, Nucl. Eng. Part IV Symp. Ser. 5 (1959) no. 22, 9–15.
151. M. J. Foole, Chem. Eng. Prog. 58 (1962) no. 4, 37–40.
152. Hydrocarbon Process. 46 (1967) no. 11, 187.
153. S. Feigelman, C. B. O’Connor, Hydrocarbon Process. 45 (1966) no. 5, 140–144.
154. Hydrocarbon Process. Pet. Refin. 42 (1963) no. 11, 204.
155. Bayer, BIOS Final Rep. Nr. 1787.
156. R. Egbert, Trans. AICHE vol. 34 (1938) 435–488.
157. R. Dasgupta, B. R. Maiti, Ind. Eng. Chem. 25 (1986) 381–386.
158. Otkrytiya/Itobret, SU 1 165 677, 1985.
159. BIOS Final Rep. Nr. 893.
160. Soc. Progil, US 3 227 525, 1964 (M. E. Degeorges, M. Jaymond).
161. Soc. Progil, US 3 112 349, 1961 (M. E. Degeorges, M. Jaymond).
162. Soc. Progil, DE-AS 1 249 244, 1961 (M. Jaymond).
163. Dow Chemical Co., US 1 938 609, 1933 (J. H. Reilly).
164. Monsanto Co., US 3 228 994, 1962 (W. D. Robinson, W. Groves).
165. Hawker Siddeley Nuclear Power Co., FR 1 272 068, 1961.
166. Montecatini, BE 586 826, GB 865 302, IT Anm. 1144/59, 1960, IT Prior. 1959 (C.
Ferry, B. Ponzini).
167. Monsanto Co., US 2 143 509, 1939 (C. Conover, A. E. Huff).
168. Montecatini, IT 586 113, 1962.
169. Kambaror YU G; SU 6 79 565, 1977, Chem. Abstr. 92 (1980) 58407 d.
170. Monsanto Co., US 3 228 994, 1962 (W. D. Robinson).
171. N. M. Indyukov, R. I. Gasanova, Khim. prom-st. (Moscow) 1975, no. 7, 555.
172. Farbenfabr. Bayer AG, DE-AS 1 012 595, 1955 (H. Holzrichter, O. Tegtmeyer).
173. F. Ullmann, Justus Liebigs Ann. Chem. 332 (1904) 48.
174. L. Gattermann, R. Ehrhardt, Ber. Dtsch. Chem. Ges. 23 (1890) 1226.
175. I. H. Gardner, P. Bergstrom, J. Am. Chem. Soc. 51 (1962) 3376.
176. Asahi Kasei Kogyo K. K., DE-AS 2 039 904, 1970.
177. Shell Oil Co., US 3 145 237, 1962 (R. van Helden, B. Balder).
178. R. van Helden, G. Verberg, Rec. Trav. Chim. Pays-Bas 84 (1965) 1263.
179. Gulf Research & Developm. Co., US 3 401 207, 1966 (C. M. Selwitz).
180. E. S. Rudakov, A. I. Lutsyk, R. I. Rudakova, Kinet. Katal. 18 (1977) 525; Kinet.
Catal. (Engl. Transl.) 18 (1977) 441.
181. M. K. Starchevskii, M. N. Vargaftik, I. I. Moiseev, Izv. Akad. Nauk SSSR 1979,
242;(Engl. Transl.) 227.
182. Teijin Ltd., US 3 963 787, 1973 (Y. Ichikawa, T. Yamaji).
183. Ube Industries, JP 55 079 324, 1978.
184. L. N. Rachkovskaya et al., Kinet. Katal. 18 (1977) 792; Kinet. Catal. (Engl. Transl.)
18 (1977) 660.
185. Ube Industries, JP 80 141 417, 1979, Chem. Abstr. 94 (1981) 156 561 k.
186. British Petroleum, DE-OS 1 909 045, 1969 (D. Fiolka, C. R. Jentsch).
187. L. H. Slaugh, Tetrahedron 24 (1968) 4523.
188. L. H. Slaugh, J. A. Leonard, J. Catal. 13 (1969) 385.
189. G. N. Koshel et al., Dokl. Akad. Nauk SSSR 237 (1977) no. 1, 164 (Chem. Tech.);
Chem. Abstr. 88 (1978) 50 394 u.
190. Phillips Petroleum, US 4 123 470, 1977 (T. P. Murths).
191. Farberov M I, SU 416 339, 1972; Chem. Abstr. 88 (1978) 62 124 u.
192. Standard Oil, US 4 218 572, 1978 (S. R. Dolhyj, L. J. Velenyi).
193. Y. Yamazaki, T. Kawai, A. Yagishita, Sekiyu Gakkaishi 21 (1978) no. 4, 229.
194. Koppers Co., US 3 059 987 1962 (H. P. Baumann).
195. N. Hendrich, Melliand Textilber. 43 (1962) 258.
196. W. E. Bachmann, H. T. Carke, J. Am. Chem. Soc. 49 (1927) 2089–98.
Vol. 2 HYDROCARBONS 649
235. H. E. Swann Jr., B. K. Kwon, G. H. Hogan, W. M. Snellings, Am. Ind. Hyg. Assoc.
J. 35 (1974) no. 9, 311.
236. W. R. F. Notten, P. W. Henderson, Biochem. Pharm. 24 (1975) 1093.
237. C. F. Reinhardt et al., Arch. Environ. Health 22 (1971) 265.
238. A. Y. H. Lu, H. W. Strobel, M. J. Coon, Mol. Pharamcol. 6 (1970) 213.
239. N. W. Lazarew, Arch. Exp. Pathol. Pharmakol. 143 (1929) 223.
240. Criteria for a Recommended Standard–Occupational Exposure to Xylene, U.S. Dept.
Health, Education and Welfare, Natl. Inst. Occup. Safety and Health, Washington
D.C. (1975).
241. O. A. Anunziata et al., Appl. Catal. 15 (1985) 235–245. K. Verschueren: Handbook
of Environmental Data on Organic Chemicals, 2nd ed., Van Nostrand Reinhold, New
York 1983.
242. W. E. Coleman et al., Arch. Environ. Contam. Toxicol. 13 (1984) 171–178.
243. N. I. Sax: Dangerous Properties of Industrial Materials, 6th ed., Van Nostrand Re-
inhold, New York 1984.
244. Threshold Limit Values and Biological Exposure Indices for 1987–1988, American
Conference of Governmental Industrial Hygienists. Cincinnati, Ohio 1987.
245. DFG, MAK- und BAT-Werte Liste 2001, Mitteilung 37, Wiley-VCH, Weinheim,
Germany 2001.
246. H. W. Gerarde: Toxicology and Biochemistry of Aromatic Hydrocarbons, Elsevier
Publishing Co., London 1960.
247. G. Hommel: Handbuch der gefährlichen Güter, Springer Verlag, Berlin Heidelberg
1987.
248. H. W. Gerarde, Am. Med. Assoc. Arch. Ind. Health 19 (1959) 403.
249. M. A. Wolf et al., Am. Med. Assoc. Arch. Ind. Health 14 (1956) 387.
250. H. W. Werner, R. C. Dunn, W. F. von Oettingen, Ind. Health Toxicol. 26 (1974) 264.
251. Maximale Arbeitsplatzkonzentrationen und biologische Arbeitsstofftoleranzwerte,
Dt. Forschungsgemeinschaft, Verlag Chemie, Weinheim 1984.
252. N. I. Sax: Dangerous Properties of Industrial Materials, 6th ed., van Nostrand Rein-
hold. New York 1984, p. 819.
253. I. Hakkinen et al., Arch. Environ. Health 26 (1973) 70.
254. A. M. Seppalainen, I. Hakkinen, J. Neurol. Neurosurg. Psychiatry 38 (1975) 248.
255. H. E. Christensen, T. T. Luginbyhl: Registry of Toxic Effects of Chemical Substances,
U.S. Dept. of H. E. W., U.S. Government Printing Office, Washington, D.C. 1975.
256. J. Opdyke, Food Cosmet. Toxicol. 12 (1974) 707.
257. J. M. Haas, H. W. Earhart, A. S. Todd: Book of Papers, 1974 AATCC National Tech-
nical Conference, American Association of Textile Chemists and Colorists, Research
Triangle Park, N. C., p. 427.
258. Patty’s Industrial Hygiene and Toxicology, vol. 2 B, Toxicology, 3rd ed.
259. Z. F. Khromenko, Gig. Tr. Prof. Zabol. 1976, 44.
260. Y. R. Adamson, J. L. Weeks, Arch. Environ. Health 27 (1973) no. 8, 68.
261. Y. R. Adamson, J. M. Furlong, Arch. Environ. Health 28 (1974) no. 3, 155.
262. R. J. Hawkins, J. L. Weeks, At. Energy Can. Ltd. (Rep.) 4439 (1973); Chem. Abstr.
80 (1974) 65 963.
263. J. S. Henderson, J. L. Weeks, Ind. Med. Surg. 40 (1973) 10.
GENERAL REFERENCES
KARL GRIESBAUM
Universität Karlsruhe (TH) Karlsruhe
Federal Republic of Germany
ARNO BEHR
Henkel KGaA, Düsseldorf Federal Republic
of Germany
DIETER BIEDENKAPP
BASF Aktiengesellschaft Ludwigshafen,
Federal Republic of Germany
HEINZ-WERNER VOGES
Hüls Aktiengesellschaft, Marl, Federal
Republic of Germany
DOROTHEA GARBE
Haarmann & Reimer GmbH Holzminden
Federal Republic of Germany
CHRISTIAN PAETZ
Bayer AG Leverkusen, Federal Republic of
Germany
GERD COLLIN
Rüttgerswerke AG, Duisburg, Federal
Republic of Germany
DIETER MAYER
Hoechst Aktiengesellschaft, Frankfurt
Federal Republic of Germany
HARTMUT HÖKE
Rüttgerswerke AG, Castrop-Rauxel,
Federal Republic of Germany
652 ACETYLENE Vol. 2
ACETYLENE
1. Introduction
Acetylene [74-86-2] is the simplest hydrocarbon with a triple bond. In the days
before oil gained widespread acceptance as the main feedstock of chemical
industry, acetylene was the predominant building block of industrial organic
chemistry. The calcium carbide process was the sole route for acetylene produc-
tion until 1940, when thermal cracking processes using methane and other
hydrocarbons were introduced. At first, these processes used an electric
arc; then, in the 1950s, partial oxidation and regenerative processes were
developed.
However, along with the expansion of the petroleum industry there was a
changeover from coal chemistry to petrochemistry, in the 1940s in the United
States and in the 1950s in Europe. As a consequence, acetylene lost its compe-
titive position to the much cheaper and more readily available naphtha-derived
ethylene and other olefins. This competition between acetylene and ethylene as
feedstocks for chemical industry has been much discussed over the last
20 years (1,2). The few hopes, such as BASF’s contribution to the submerged
flame process, Hoechst’s crude oil cracking (HTP), or Hüls’ plasma process,
have not halted the clear trend toward ethylene as a basic chemical. With
the first oil price explosion in 1973, the development of crude cracking pro-
cesses suffered a setback, and the new processes, such as the Kureha/Union
Carbide process, DOW’s PCC process (PCC ¼ partial combustion cracking), or
the Kureha/Chiyoda/Union Carbide ACR process (ACR ¼ advanced cracking
reactor), raise little hope for a comeback of acetylene chemistry. Acetylene pro-
duction peaked in the United States at 480000 t in the 1960s, and in Germany
at 350000 t in the early 1970s (3). Since then, acetylene production has
decreased steadily. In both countries the losses were principally in carbide-
derived acetylene; in fact, Germany has produced acetylene for chemical
purposes almost exclusively from natural gas and petrochemical sources
since 1975.
All acetylene processes, including carbide processes, are high-temperature
processes, requiring a large amount of energy. They differ essentially only in
the manner in which the necessary energy is generated and transferred. They
can be classified into three groups: partial combustion processes, electrothermic
processes, and processes using heat carriers. Finally, the use of byproduct acet-
ylene from olefin plants is economically viable in many cases. For each group of
acetylene processes several variants have been developed using various feed-
stocks and techniques. Today, only three processes remain for the commercial
production of acetylene: the calcium carbide route, in which the carbide is pro-
duced electrically, the arc process, and the partial oxidation of natural gas.
Other once popular processes have become uneconomical as the price of naphtha
has increased.
Some processes were shelved in the experimental or pilot-plant stage as the
importance of acetylene declined. However, other new processes involving the
use of coal, sulfur-containing crude oil, or residues as feedstocks for acetylene
production are in the pilot-plant stage.
2. Physical Properties
Due to the carbon – carbon triple bond and the high positive energy of formation,
acetylene is an unstable, highly reactive unsaturated hydrocarbon. The C – C
triple bond and C – H s bond lengths are 0.1205 and 0.1059 nm, respectively.
For the electronic structure of acetylene and a molecular orbital description,
see (6). The acidity of acetylene (pKa ¼ 25) permits the formation of acetylides
(see Section 3.2).
Under normal conditions acetylene is a colorless, nontoxic but narcotic gas;
it is slightly lighter than air. The main physical properties are listed in Table 1.
The critical temperature and pressure are 308.32 K and 6.139 MPa. The triple
point at 128.3 kPa is 192.4 K. The vapor pressure curve for acetylene is shown
in Figure 1. The formation of acetylene is strongly endothermic (DHf ¼ þ 227.5
kJ/mol at 298.15 K).
Self-decomposition can be initiated when certain pressure limits above
atmospheric pressure are exceeded (for details see Section 5.1).
The crystalline structure of solid acetylene changes at 140.15 8C from a
cubic to an orthorhombic phase. The heat of reaction for this phase change is
2.54 kJ/mol (8); two different values for the enthalpy of fusion are reported in
the literature (6,7). Figure 2 shows the density of liquid and gaseous acetylene.
Details about flame properties, decomposition, and safety measures are
given in Section 5.
Table 2. Solubility Coefficients of C2H2 in Various Solvents (in mol kg1 bar1) (9)
Solvent C2H2 pressure, bar 208C 258C
Methanol 0.98 1.979 0.569
Ethanol 0.98 0.851 0.318
n-Butanol 0.245 – 0.657 0.237
1,2-Dichlorethane 0.4–1.05 0.569 0.218
Carbon tetrachloride 0.98 0.164 0.075
n-Hexane 6.90 0.523 0.264
n-Octane 0.196 – 14.71 0.205 0.146 (08C)
Benzene 0.98 0.225
Toluene 0.98 0.619 0.214
Xylene (tech.) 0.98 0.528 0.189
4-Methyl-1,3-dioxalan-2-one 0.98 1.137 0.350
Tri-n-butylphosphate 0–0.4 2.366 0.614
Methyl acetate 0.98 2.912 0.878
Triethylene glycol 0.98 0.205
Acetone 0.98 4.231 1.069
N-Methyl-2-pyrrolidone 0.98 5.687 1.319
N,N-Dimethylformamide 0.98 5.096 1.501
Dimethyl sulfoxide 0.98 1.001
Ammonia 0.98 7.052 2.229
3. Chemical Properties
Because of its strongly unsaturated character and high positive free energy of for-
mation, acetylene reacts readily with many elements and compounds. As a result
acetylene is used as raw material for a great variety of substances. Important are
addition reactions, hydrogen replacements, polymerization, and cyclization.
Acetylene is more susceptible to nucleophilic attack than, for instance, ethy-
lene. In addition, the polarized C-H bond makes acetylene acidic (pKa ¼ 25) (16).
Because of this acidity, acetylene is very soluble in basic solvents (17,18), forming
hydrogen bonds with them (19). Therefore, the vapor pressures of such solutions
cannot be described by Raoult’s law (20).
The development of the acetylene pressure reactions by W. REPPE (1892 –
1969), BASF Ludwigshafen (Federal Republic of Germany) (21–23) began mod-
ern acetylene chemistry. The most interesting groups of reactions are vinylation,
ethynylation, carbonylation, and cyclic and linear polymerization.
HC CH þ H2 O ! CH3 CHO
HC CH þ RSH ! CH2 ¼ CH S R
Vol. 2 ACETYLENE 659
R1 R2 NH þ HC CH ! R1 R2 N CH ¼ CH2
HC CH þ HCN H2 C ¼ CH CN
HC CH þ RCOR0 ! HC C CðOHÞRR0
CH CH þ HCHO ! HC CCH2 OH
HC CH þ CO þ H2 O ! CH2 ¼ CH COOH
The reaction of acetylene with water or alcohols and carbon monoxide using
Ni(CO)4 catalyst was first reported by W. REPPE (26). If water is replaced by
thiols, amines, or carboxylic acids, then thioesters of acrylic acid, acrylic amides,
or carboxylic acid anhydrides are obtained.
Ethyl acrylate [140-88-5]:
HC CH þ HC CH ! H2 C ¼ CH C CH
662 ACETYLENE Vol. 2
The direct reaction of the acetylene with a molten metal, such as sodium, or
with a finely divided metal in an inert solvent, such as xylene, tetrahydrofuran,
or dioxane, at a temperature of about 408C, is also possible:
2M þ C2 H2 ! M2 C2 þ H2
The very explosive copper acetylides, e.g, Cu2C2 H2O, can be obtained by
reaction of copper(I) salts with acetylene in liquid ammonia or by reaction of
copper(II) salts with acetylene in basic solution in the presence of a reducing
agent such as hydroxylamine. Copper acetylides can also form from copper oxides
and other copper salts. For this reason copper plumbing should be avoided in
acetylene systems.
Vol. 2 ACETYLENE 663
4. Production
requires very high temperatures and very short reaction times. The
main reasons for the extreme conditions are the temperature dependence of
the thermodynamic properties (molar enthalpy of formation, DHf , and molar
664 ACETYLENE Vol. 2
Table 3. Standard Molar Enthalpies of Formation and Gibbs Free Energy of Formation at
298 K
DHf (kJ/mol) DGf (kJ/mol)
C (s) 0 0
H2 (g) 0 0
CH4 (g) 74.81 50.82
C2H2 (g) +226.90 +209.30
C2H4 (g) + 52.30 + 68.15
C2H6 (g) 84.64 32.90
C3H6 (g) + 20.43 + 62.75
C3H8 (g) 103.90 23.48
n-C4H10 (g) 126.11 17.10
free energy of formation, DGf) of the hydrocarbons; the position of the chemical
equilibria under the reaction conditions; and the kinetics of the reaction.
Thermodynamic data relevant to the hydrocarbon–acetylene system are
shown in Table 3 and Figure 7. These data show clearly that at normal tempera-
tures acetylene is highly unstable compared to the other hydrocarbons. However,
Figure 7 also shows that the free energy of acetylene decreases as temperature
increases, whereas the free energies of the other hydrocarbons increase. Above
about 12308C, acetylene is more stable than the other hydrocarbons. The tem-
perature at which the acetylene line intersects an other line in Figure 7 is higher
the shorter the chain length of the hydrocarbons. Acetylene production from
methane requires higher reaction temperatures than production from heavier
hydrocarbons.
Fig. 7. Gibbs free energy of formation per carbon atom of several hydrocarbons as a
function of temperature.
Vol. 2 ACETYLENE 665
Fig. 8. Equilibrium curve for the methane cracking reaction, 2 CH4 Ð C2H2 þ 3 H2.
plant, however, these may be converted to synthesis gas, pure hydrogen, and
pure CO and can actually improve the economics of acetylene production.
BASF Process (Sachsse-Bartholome). The BASF process for the pro-
duction of acetylene from natural gas has been known since 1950 (43). World-
wide, some 13 plants used this process in 1983, a total capacity of about
400000 t/a. All use a water quench, except the plant in Ludwigshafen (Federal
Republic of Germany) operated with an oil quench (44).
The basic idea of partial combustion involves a flame reaction on a premixed
feed of hydrocarbon and oxygen. In this way the rate of hydrocarbon conversion
is made independent of the gas-mixing rate, which is governed by diffusion. Only
then can the residence time in the reaction zone be made much smaller than the
average decay time of acetylene. The separate preheating of the reactants to the
highest temperature possible before introduction into the burner reduces the
consumption of oxygen and the hydrocarbon within the burner. It also causes
a higher flame propagation speed and therefore a higher mass flow within the
acetylene burner.
The smallest, but most important, part of a partial oxidation acetylene
plant is the burner, Figure 9. Its design is nearly identical in the two process var-
iants (i.e, oil and water quench).
At the top of the burner, the preheated reactants, (6008C in the case of
methane) must be mixed (c) so rapidly that there are no domains with a high oxy-
gen concentration. Such domains cause preignition before the reactants are
introduced into the reaction zone (g). In fact, the reaction mixture ignites after
an induction time depending on the hydrocarbon used as feed and on the preheat
temperature, on the order of a few tenths of a second. The maximum preheat
temperature is lower for higher hydrocarbons than for methane. Backmixing
of the gas between the mixing and the reaction zones is avoided by the diffuser
(e), a tube which connects the mixing zone and the burner block (f). Because of its
smooth surface and the small opening angle the reaction feed is decelerated
gently and backmixing does not occur.
The burner block (f) consists of a water-cooled steel plate with a large number
of small channels. The flow velocity through these channels is substantially higher
than the flame propagation speed, so that the flame below the burner block cannot
backfire into the diffuser. The lower side of the burner block has small openings
between the channels through which additional oxygen is fed into the reaction
mixture. At these openings small flames form and initiate the flame reaction.
The strong turbulence below the burner block stabilizes the flame.
Under unfavorable conditions the flame may appear above the burner
block. In this case the oxygen feed must be shut off immediately and replaced
by nitrogen. This extinguishes the preignition before it can cause any damage
to the equipment. Such preignitions can result from a momentary shift in the
oxygen : hydrocarbon ratio or the entrainment of small particles of pyrophoric
iron formed from rust in the preheaters.
As mentioned above, the hot gas leaves the reaction chamber within a few
milliseconds and passes through sprays of water or oil, which cool the gas almost
instantaneously, to about 808C in the case of water or 200–2508C in the case of
oil. The quench system consists of a set of nozzles that are fed by three annular
tubes below the reaction chamber.
668 ACETYLENE Vol. 2
Fig. 11. Acetylene water quench process (AWP). a) Preheaters; b) Acetylene burner;
c) Cooling column; d) Electrofilter; e) Soot decanter; f ) Cooling tower.
200–2508C. The oil absorbs the heat from the gas and then passes through waste
heat boilers before returning to the quench. The sensible heat of the cracked gas
represents more than 15% of the heating value of the feedstock. The pressure of
the generated steam depends on the process configuration and can reach 15 bar
(1.5 MPa).
Unlike the water quench process, where the scraped coke deposits sink to
the bottom of the quench chamber and are easily removed, in the oil quench the
coke deposits do not settle immediately. In order to prevent plugs in the quench
Fig. 12. Acetylene oil quench process (AOP). a) Preheaters; b) Acetylene burner;
c) Burner column; d) Mill pump; e) Coker; f) Decanter; g) Final cooler.
Vol. 2 ACETYLENE 671
nozzles a mill pump (d) is installed immediately underneath the burner col-
umn. the coke and soot content in the quench circuit is kept near 25% by send-
ing a fraction of the coke-containing oil to externally heated, stirred kettles
(coker (e)). In the kettles the volatile matter evaporates very quickly, leading
to fluidization of the coke bed. The vapor is returned to the burner column,
while the soot is agglomerated. A fine-grained coke is withdrawn from the
bottom of the coker.
Because of the cracking losses in the quench a certain amount of quench oil
has to be added continuously to the process. This makeup oil is at least 0.15 to
0.3 t per ton of acetylene, depending on the stability of the oil used. When resi-
dual oil from steam crackers is used, it can be desirable to add up to 1 t of oil per
ton of acetylene, because the excess oil is partially converted to light aromatic
hydrocarbons.
The cracked gas leaving the quench is cooled in a burner column (c), where
there are additional oil circuits for the production of 3-bar steam and for boiler
feedwater preheat. At the top of the column a small amount of a low-boiling oil
(BTX ¼ benzene, toluene, and xylene) is added to prevent deposit-forming aro-
matics (mainly naphthalene) from passing downstream into other parts of the
plant. The cracked gas, which has to be compressed before separation, is cooled
further (g) by water. At this stage most of the BTX condenses and is separated
from the water in a large decanter (f).
Table 4 shows the cracked gas compositions for the BASF acetylene oil
quench process when natural gas, liquid petroleum gas (LPG), or naphtha is
used as feedstock. The water quench process gives very similar compositions.
The relative amounts of hydrogen and carbon monoxide formed depend on the
hydrogen : carbon ratio of the feedstock used. Even when naphtha is used, almost
no ethylene forms. This is because the reaction takes place above 12008C where
the formation of ethylene is thermodynamically impossible. Only a prequench
with additional naphtha or LPG produces additional acetylene and ethylene at
intermediate temperatures, as in the case of two-step processes. The higher
hydrocarbons require a somewhat lower reaction temperature than methane
Table 4. BASF Acetylene Oil Quench Process, Cracked Gas Composition (vol %)
Raw material (DH, kJ/mol)
Methane LPG Naphtha
Component * (400) (325) (230)
H2 56.5 46.4 42.7
CH4 5.2 5.0 4.9
C2H4 0.3 0.4 0.5
C2H2 7.5 8.2 8.8
C3þ** 0.5 0.6 0.7
CO 25.8 35.0 37.9
CO2 3.2 3.4 3.5
O2 0.2 0.2 0.2
Inerts balance
*Dry gas, water, and aromatic compounds condensed out.
**Hydrocarbons with three or more carbon atoms.
672 ACETYLENE Vol. 2
and have a less endothermic heat of reaction: oxygen consumption per ton of
acetylene is lower for the higher hydrocarbons in spite of the lower preheating
temperature.
Comparison of Oil Quench and Water Quench Processes. The
advantage of the oil quench process is obvious: the heat recovery in the form
of steam makes the overall thermal efficiency in relation to primary energy
input rather high. If the thermal efficiency for the production of electricity is
33%, over 70% of the net heating value of the overall primary energy input is
recovered in the form of products and steam. A comparison between the oil
quench and water quench (see Table 6) shows that the oil quench requires a
net heating value input of 300–330 GJ per ton of acetylene, of which 82 GJ
(27–25%) is lost, whereas the water quench requires a 288 GJ input, of which
113 GJ (39%) is lost.
Acetylene Recovery. Liquid acetylene is a dangerous product, even at
low temperatures. Separation of the cracked gas by cryogenic processes such
as those used in olefin production is clearly ruled out. One exception to this
rule is the acetylene recovery unit of the submerged flame process (Section
4.2.3) (45), in which all hydrocarbons except methane are condensed
at –1658C. Otherwise, acetylene is recovered by selective absorption into a sol-
vent. This procedure is economical only when the cracked gas is compressed.
The upper limit for the pressure is determined by the danger of explosions,
and as a rule the partial pressure of acetylene should be kept below 1.4 bar
(0.14 MPa).
The solubility of acetylene in the solvents used is between 15 and 35 m3
(STP) per m3 of solvent under process conditions. The dissolved gas is recov-
ered by depressurizing the solvent and by vapor stripping at higher tempera-
tures. All solvents used commercially, N-methylpyrrolidone (NMP), methanol,
ammonia, and dimethylformamide (DMF), are miscible with water. They are
recovered from the gas streams leaving the plant by water scrubbing and
distillation.
The kinetics of acetylene formation always lead to the formation of higher
homologues of acetylene as byproducts (46), mainly diacetylene, but also methy-
lacetylene, vinylacetylene, and others. These compounds polymerize very easily
and must be removed from the cracked gas as soon as possible. Because they are
much more soluble in the solvents than acetylene, scrubbing the cracked gas
with a small amount of solvent before it enters the acetylene recovery stages is
sufficient.
Absorption Section (Fig. 13). Acetylene recovery is illustrated here by the
BASF process. N-Methylpyrrolidone is used to separate the cracked gas into
three streams:
1. Higher homologues of acetylene and aromatics, the most soluble part of the
cracked gas. (This is a small stream of gas, which is diluted with crude
synthesis gas for safety reasons and is used as fuel.)
2. Product acetylene, less soluble than the higher acetylenes, but much more
soluble than the remainder of the gas
3. Crude synthesis gas (off-gas), mainly hydrogen and carbon monoxide
Vol. 2 ACETYLENE 673
In the prescrubber (b) the cracked gas is brought into contact with a small
amount of solvent for removal of nearly all the aromatic compounds and C4 and
higher acetylenes except vinylacetylene. This is done after the compression of the
gas if screw compressors are used but before compression if turbo compressors
are used because turbo compressors cannot tolerate deposits on their rotors. In
the main scrubber (d) the gas is brought into contact with a much larger amount
of N-methylpyrrolidone (NMP), which dissolves all the acetylene, the remaining
homologues, and some carbon dioxide. Crude synthesis gas (off-gas) leaves at the
top of the column.
The NMP solution is degassed in several steps in which the pressure is
reduced and the temperature increased. The stripper (e) operates at pressures
and temperatures slightly above ambient. In this tower, the solution is put in
contact with a countercurrent gas stream from the subsequent degassing step
(f). This leads to the evolution of carbon dioxide, the least soluble of the dis-
solved gases, at the top of the stripper. The carbon dioxide is recycled to the suc-
tion side of the compression and thereby is shifted into the crude synthesis gas.
The acetylene product is withdrawn as a side stream from the stripper. The N-
methylpyrrolidone solution is then completely degassed (f) in two further steps
at 110–1208C, first at atmospheric, then at reduced pressure. Vinylacetylene,
methylacetylene, and excess process water are withdrawn as bleed streams
from the vacuum column (f). The water content of the solvent is controlled by
the reboiling rate in the vacuum column. At the bottom of the vacuum column,
degassing is completed, and the solvent is cooled and returned to the main
scrubber (d).
The small amount of solvent from the prescrubber (b) is stripped with crude
synthesis gas for recovery of the dissolved acetylene, the overhead gas being
674 ACETYLENE Vol. 2
recycled to the suction side of the compressor. The solvent is then degassed com-
pletely in the vacuum stripper (g), a column which also accepts the bleed stream
from the vacuum column (f) containing the excess process water together with
some higher acetylenes. The overhead vapor of the vacuum stripper contains
the higher acetylenes, water, and some NMP vapor. In a side column (h) the
NMP is recovered by scrubbing with a small amount of water, which is recycled
to the main solvent stream. The gas is cooled (i) by direct contact with water
from a cooling circuit to condense most of the water vapor. The higher
acetylenes are diluted with crude synthesis gas before they enter and after
they leave the vacuum pump (j). The diluted higher acetylenes, which are now
at a pressure slightly above atmospheric, can be used as fuel gas, e.g, for soot
incineration.
In order to minimize the polymer content of the solvent, about 2% of the cir-
culating flow is withdrawn continuously from the vacuum stripper circuit and
distilled under reduced pressure, leaving the polymers as a practically dry
cake for disposal.
The acetylene product from the process as described above has a purity of
about 98.4%, the remainder consisting mainly of propadiene, methylacetylene,
and nitrogen. For most applications the purity is increased to 99.7% by scrubbing
with sulfuric acid and sodium hydroxide solutions. Table 5 compares the compo-
sitions of crude and purified acetylene. Table 6 compares the consumption and
product yields per ton of acetylene for the oil quench process with those for the
water quench process.
Other Partial Combustion Processes. The main features of the BASF
process described in detail above are common to all partial oxidation processes.
Therefore only the differences between the BASF acetylene burner and burners
used in the Montecatini and the SBA processes (41,47) are described. These two
processes have also attained some importance. The details of the acetylene recov-
ery process depend on the properties of the solvent, but here too the basic prin-
ciples are the same for all processes.
Montecatini Process. The Montecatini burner (48) has the same main
components as the BASF burner: mixing unit, gas distributor, reaction chamber,
and quench. The essential difference is the pressure for acetylene synthesis,
which can be as high as several bar. This saves compression energy, improves
heat recovery from the quench water, which is obtained at 1258C, and is claimed
Vol. 2 ACETYLENE 675
Table 6. BASF Acetylene Process, Consumption and Product Yields Per Ton
of Acetylene
Consumption and product yields Oil quench Water quench
Feed and energy requirements
Natural gas, 36000 kJ/m3(STP) 5833 m3 = 210 GJ 5694 m3 = 205 GJ
(LHV) *
Oxygen, 0.55 kWh/m3(STP) ** 3400 m3 = 20.4 GJ 3400 m3 = 20.4 GJ
Fuel gas = 12.0 GJ = 18.0 GJ
Residue oil minimum (surplus) 0.3 (1.0) t = 12.0 (40.0) GJ
Sulfuric acid 160 kg 160 kg
Sodium hydroxide 5 kg 5 kg
N-Methylpyrrolidone 5 kg 5 kg
Electric energy ** 3200 kWh = 34.9 GJ 3100 kWh = 33.8 GJ
Steam, 4 bar 5.0 t = 11.7 GJ 4.5 t = 10.5 GJ
to make soot removal easier because the cracked gas is scrubbed with water
above 1008C. Although it is well known (8) that acetylene decomposition is accel-
erated under pressure at high temperatures (>10008C), the acetylene yield is
comparable to that obtained at atmospheric pressure because of the short resi-
dence time in the reactor. Methanol is used at cryogenic temperatures for acet-
ylene recovery. The main steps of the gas separation are absorption of higher
acetylenes and of aromatics, absorption of acetylene, stripping of coabsorbed
impurities, and desorption of acetylene.
SBA Process (of the Société Belge de l’Azote). The SBA burner (49)
has the same main components as the other processes. However, it has a tele-
scope-like reaction chamber and a device for shifting the quench up and down.
Thus it is possible to adjust the length of the reaction zone for optimum
676 ACETYLENE Vol. 2
residence time at any throughput. The walls of the reaction chamber are
sprayed with demineralized water to prevent coke deposits. This eliminates the
need to scrape the reaction chamber periodically. Acetylene recovery is carried
out with several scrubbing liquids — kerosene, aqueous ammonia, caustic soda,
and liquids ammonia, each with its own circuit. After soot is separated from the
gas in an electrofilter, higher hydrocarbons are absorbed in kerosene or gas oil.
Carbon dioxide is scrubbed in two steps, first with aqueous ammonia and then
with caustic soda solution. The acetylene product is absorbed into anhydrous
ammonia and must be scrubbed with water after desorption. All the ammonia–
water mixtures are separated in a common distillation unit. This recovery scheme
leads to exact separation of the various cracked gas components.
Additional Remarks. The Montecatini and SBA processes can also be
operated with two-stage burners. A prequench with light hydrocarbons cools
the cracked gas to about 8008C. After a residence time at this intermediate tem-
perature the gas is cooled down with water. In this way the heat content of the
hot gases is used for further cracking of hydrocarbons to yield extra acetylene
and olefins. The presence of additional components in the cracked gas requires
more process steps in the gas separation units.
Submerged Flame Process. The submerged flame process (SFP) of
BASF attracted considerable interest up to 1973 as a partial combustion pro-
cess for the production of acetylene, ethylene, C3, and C4 hydrocarbons, and
synthesis gas from feedstock of crude oil and residues, such as Bunker C oil
and vacuum residue (44,45). Although it was abandoned at the end of 1973,
the need to make the most economic use of raw materials has renewed interest
in this process (50).
Oxygen compressed to 16 bar (1.6 MPa) feeds a flame that is submerged in
the oil. The oil surrounding the flame is partially burnt to obtain the necessary
reaction temperature and also acts as the quenching medium. This process dif-
fers from the partial oxidation processes using natural gas and lighter hydrocar-
bons in five main respects:
1. Crude oil can be gasified without the formation of residues, and the process
can be operated under certain conditions with heavy fuel oil.
2. All the soot formed is consumed when crude oil feedstock is used, eliminat-
ing all the problems associated with the storage, disposal, or utilization of
acetylene soot.
3. The heat of reaction is removed by steam generation at 8 bar (0.8 MPa).
4. The process is operated at 9 bar (0.9 MPa) so that the oxygen is the only
compressed stream. The cracked gas is formed at a pressure sufficient for
economic separation.
5. The design of the cracking unit is greatly simplified because the reaction
feed, fuel, and quenching medium are identical.
The process is described in detail in the literature cited; therefore, only gen-
eral overviews of the cracking unit (Fig. 14) and the separation unit (Fig. 15) are
shown here. The capacities of a submerged flame burner for acetylene and ethy-
lene are 1 t/h and 1.15 t/h, respectively. To produce these quantities, 5000 m3
Vol. 2 ACETYLENE 677
Fig. 14. Submerged flame process (SFP) — cracking unit. a) Reactor; b) Oil cooler;
c) Steam generator; d) Oil recycle pump; e) Scrubber; f) Naphtha cooler; g) Naphtha se-
parator; h) Naphtha pump; i) Spray cooler; j) Separating vessel; k) Recycle-water pump;
l) Recycle-water cooler.
(STP) of oxygen and 8–10 t of oil are required per hour. The cracked gas shows
the following average composition (vol%, the components grouped as streams
leaving the separation unit):
Main products
Acetylene 6.2
Ethylene 6.5
Crude synthesis gas
Carbon monoxide 42.0
Hydrogen 29.0
Methane 4.0
Inerts 0.6
Other hydrocarbons
Ethane 0.5
Propane 0.1
Propene 1.2
Propadiene, propyne 0.7
1,3-Butadiene 0.5
Other C4 and C5+ 14 hydrocarbons 1.5
Remainder
Carbon dioxide 7.0
Hydrogen sulfide 0.05–0.5
Carbon oxide sulfide 0.03–0.3
*C5+ , five or more carbons
Unlike all other processes the submerged flame process uses low tempera-
tures (–1658C) to separate the off-gas, consisting of carbon monoxide, hydro-
gen, and methane, from the C2 and higher hydrocarbons. On account of the
acetylene in the condensed phase, extensive decomposition tests have been
678 ACETYLENE Vol. 2
carried out. Whereas the cracking unit (Fig. 14) and the amine scrubbing unit
have been tested by Soc. Ital. Serie Acetica Sintetica, Milan, on a commercial
scale, the remaining purification units (Fig. 15) have not. However, the experi-
ence obtained with a pilot plant indicates that major difficulties are not to be
expected.
The submerged flame process may become competitive because of its ability
to use crude oil and especially residues and because of its low losses on the pri-
mary energy input.
Partial Combustion Carbide Process. Calcium carbide production from
lime and coal requires a large high-temperature heat input (see Section 4.3.4).
In the thermal process some of the coal must be burnt to attain the necessary
reaction temperature and supply the heat of reaction. The thermal carbide pro-
cess was developed by BASF (8,51) from 1950 to 1958 to eliminate the input of
electrical energy necessary in the classic carbide process. Starting in 1954, a
large pilot plant, with a nominal carbide capacity of 70 t/d, was operated, but
in 1958 the more economical petrochemical acetylene production halted further
development. Carbide production is just one way of converting coal chemically;
other methods include pyrolysis, hydrogenation, and gasification. The question
arises as to the conditions under which a thermal carbide process using oxygen
can compete with the electric carbide process. The biggest drawback of carbide
production in a shaft furnace (Fig. 16) compared to the electric carbide process
is the lack of commercial-scale operational experience. Specific disadvantages
are greater susceptibility to disruption because of plugging of the furnace feed,
Vol. 2 ACETYLENE 679
Fig. 16. Partial combustion carbide process. a) Carbide furnace; b) Refractory brick lin-
ing; c) Charging hopper; d) Gas outlet; e) Oxygen jet; f) Tapping burner; g) Tapping chute;
h) Bogey; i) Cyclone; j) Washing column; k) Desintegrator; l) Compressor.
more stringent specifications for the raw materials, more handling of solids, and
the large amount of byproduct. There are two main advantages:
1. A thermal efficiency of about 50% versus about 30% for the electrothermal
process if the thermal efficiency of electricity production is 33%
2. Carbon monoxide production, which is desirable because carbon monoxide
can be converted to synthesis gas by the water-gas shift reaction (!Gas
Production)
Table 7. Partial Combustion Carbide Process, Consumption and Product Yields Per Ton
Acetylene
Raw materials
Coke, dry (88% C) 5700 kg
Lime (92% CaO) 3140 kg
Oxygen (98%) 3560 m3 (STP) 5090 kg
Total consumption 13930 kg
Products
Carbide (80.5%) 2850 kg ¼
^ 1000 kg acetylene
Carbon monoxide 7980 m3 (STP) 9975 kg
(CO 95.5, H2 2.0, N2 2.0, CO2 0.5 vol %)
Dust 900 kg
Losses 205 kg
Total products 13930 kg
680 ACETYLENE Vol. 2
The energy requirement for the formation of acetylene can be made indepen-
dent of the hydrocarbons used as feedstock.
Hydrocarbon consumption can be reduced by 50%.
Provided that electrical energy is available under favorable conditions
(nuclear power, hydroelectric power, cheap coal) and/or the availability of
hydrocarbons is limited, electrothermic processes are more economical.
Fig. 17. Hüls electric-arc furnaces for gaseous, liquid, and solid feed. A) One-step
process; B) Two-step process.
wall of the electrode are only 6008C. Thermal losses are therefore limited to less
than 10% of the electrical power input of 8.5 MW.
The residence time of the gas in the arc furnace is a few milliseconds. In this
interval, the hydrocarbons are cracked, mainly into acetylene, ethylene, hydro-
gen, and soot. At the end of the arc furnace, the gases are still at a temperature of
about 18008C. The high heat content of this gas can be exploited for additional
ethylene production by means of a prequench with liquid hydrocarbons. This
lowers the temperature to about 12008C. Because acetylene rapidly decomposes
into soot and hydrogen at these temperatures, the gases must be quenched
immediately with water to about 2008C, i.e, a quench rate of 1068C/s must be
achieved.
The specific energy requirement (SER) and the acetylene yield depend on
the geometry and dimensions of the furnace and electrodes, the velocity distribu-
tion of the gas, and the kind of hydrocarbon to be cracked. Once the furnace has
been designed, only the hydrocarbons can be varied.
Process Without Prequench. Figure 18 shows acetylene and ethylene
yields and the specific energy requirement (SER) of various saturated hydrocar-
bons under constant conditions without prequench. Methane shows the highest
SER and acetylene yield, but the lowest ethylene yield. As the chain length is
increased, both acetylene yield and SER decline, corresponding to the declining
heat of acetylene formation from the various hydrocarbons.
Normally, pure hydrocarbons are not available. The results obtained from
mixtures of hydrocarbons can be expressed as a function of the carbon number,
682 ACETYLENE Vol. 2
Fig. 18. Acetylene yield, ethylene yield, and energy consumption for various hydrocar-
bons in the Hüls arc process.
which is the number of moles of carbon atoms bound in hydrocarbons per mole of
the gaseous mixture. Figure 19 shows specific amounts of acetylene, ethylene,
and hydrogen formed and of hydrocarbon consumed as a function of carbon num-
ber. This function enables the Hüls process to be optimized within certain limits,
for example, for hydrogen output in relation to acetylene production.
Process with Prequench. Cracking in the prequench section is essen-
tially an ultrasevere steam cracking process. The kind and amount of hydrocar-
bons used for the prequench can be varied. Figure 20 shows the specific product
yield for different prequench rates for feeding methane to the arc furnace and
propane to the prequench. Acetylene and hydrogen yields are unaffected,
whereas ethylene shows a slight maximum and declines when the temperature is
not sufficient at a given residence time. Propene shows a steady increase, and the
C3 : C2 ratio is below 0.25. The relative ethylene yield from various hydrocarbons
is as follows: ethane 100, propane 75, n-butane 72, isobutane 24, 1-butene 53.
Oil quench. Because the gas temperature of the furnace gas after pre-
quench is on the order of 12008C, an oil quench system has been developed to
regain about 80% of the sensible heat content of the furnace gas as steam by
heat exchange. The soot–oil mixture formed can be upgraded to a sulfur- and
ash-free high-grade petroleum coke. Figure 21 shows the newly designed Hüls
system with oil quench.
The Purification System. The process of purification depends on the
type of the quench system. In the case of water quenching, 80% of the carbon
black is removed by cyclones as dry carbon black, the remaining 20% as soot
Fig. 21. Process with oil quench system. a) Heat recovery; b) Arc furnace; c) Oil recovery;
d) Separation of medium-boiling compounds; e) Separation of low-boiling compounds;
f) Oil regeneration.
684 ACETYLENE Vol. 2
Fig. 22. Principal separation and purification steps for the furnace gas of the Hüls arc
process.
Fig. 23. Acetylene and ethylene yields as a function of the low-boiling components.
Consumption
Crude oil consumed 3.5 t
Power (d.c.) 10500 kW h
Byproducts
Ethylene 0.46 t
Hydrogen (99.9%) 1100 m3 (STP)
Fuel gas 0.74 t
Soot–oil mixture (20% soot) 1.2 t
Vol. 2 ACETYLENE 687
Table 9. Consumption and Yield Per Ton of Acetylene for the Hoechst Arc Process and
Naphtha Feed for Different Reactor Designs
Low ethylene yield High ethylene yield
Consumption
Naphtha 1.92 t 2.50 t
Quench oil 0.53 t 0.63 t
Energy(2-phase a.c.) 9300 kW h 10500 kW h
Byproducts
Ethylene 0.5 t 0.95 t
Hydrogen 1450 m3 (STP) 1500 m3 (STP)
Soot – oilmixture(20% soot) 0.75 t 1.00 t
Hoechst used a high-intensity three-phase a.c. arc furnace at 1.4 kV and 4.2 kA,
giving a power input of 10 MW (54). The thermal efficiency was 90%. Because of
the high amperage the graphite electrodes had to be replenished continually. The
generator was lined with graphite. Different reactor designs for ethylene: acety-
lene ratios of 0.5 and 1.0 were developed by varying the mixing intensity of the
hydrogen plasma jet with the liquid hydrocarbon. The tests were carried out with
naphtha feed (see Table 9). The cracked gas was quenched with residue oil, in a
manner similar to that described in the Hüls process.
The acetylene concentration in the Hüls process and the Hoechst process
was ca. 14 vol% so that in principle the same acetylene separation process can
be used as described above for the arc process.
Production from Coal (Arc Coal Process). Numerous laboratory tests for
the conversion of coal to acetylene using the arc or plasma processes have been
carried out since the early 1960s (55). The results can be summarized as
follows:
The AVCO arc furnace (Fig. 24) consists of a water-cooled tungsten-tip cath-
ode and a water-cooled anode (56). The arc is stabilized by a magnetic field sur-
rounding the anode, forcing the anode striking point of the arc to rotate rapidly
and so avoiding burnthrough. The dried and finely ground coal is injected by
means of a hydrogen gas flow around the cathode. Additional gas without coal
is introduced around the cathode and at the anode as a sheath. On passing the
arc zone the coal particles are heated up rapidly. The volatiles are released and
are cracked to acetylene and byproducts, leaving a residue of fine coke particles
covered with soot. After a residence time of some milliseconds the gas–coke mix-
ture is quenched rapidly with water or gases. The use of a prequench system
688 ACETYLENE Vol. 2
Fig. 24. Principal scheme of the AVCO plasma furnace for the pyrolysis of coal.
similar to that of the Hüls arc process was also tested. The system pressure can
be varied between 0.2 and 1.0 bar (20 and 100 kPa).
Hüls’ pilot plant uses the same plasma furnace as for the crude oil cracking,
but with 500 kW of power (57). The dried and ground coal is injected into the
plasma jet, and the coal is cracked to acetylene and byproducts in the reactor.
The reactor effluent can be prequenched with hydrocarbons for ethylene produc-
tion or is directly quenched with water or oil. Char and higher boiling compo-
nents are separated by cyclones and scrubbers, respectively. The problem in
the reactor design is to achieve thorough and rapid mixing of the coal with the
plasma jet and to avoid forming carbonaceous deposits on the wall. Smaller
amounts of deposits can be removed by periodic wash cycles with water. Opera-
tion times of 2.5 h by AVCO and 5 h by Hüls have been reported.
Experiments published by Hüls and AVCO show that at the optimal resi-
dence time the energy density of the plasma jet, the specific power, and the pres-
sure all greatly affect the acetylene yield (Fig. 25 and Fig. 26). Other parameters
affecting the yield are the amounts of volatiles in the coal and the particle size.
The lowest figures for the specific energy consumption published by AVCO are of
the order of 27–37% based on water-free coal.
In addition to acetylene, the exit gas contains considerable amounts of CO,
depending on the oxygen content of the coal. Because nitrogen and sulfur are
present in the coal, other byproducts are HCN, CS2, COS, and mercaptans.
The gas separation system is therefore designed accordingly (58). Depending
on the hydrogen content of the coal, the process is either self-sufficient in hydro-
gen or has a slight surplus. The total gas yield of the coal based on a volatile con-
tent in the coal of 33% is up to 50%. Thus 50% of the coal remains as char. Tests
with a view to using this char in the rubber industry have been unsuccessful so
far. Thus the char can be used only for gasification or as a fuel.
Vol. 2 ACETYLENE 689
Fig. 25. Effect of the energy density of the plasma on the cracking of coal (AVCO).
Fig. 26. Acetylene yield and specific energy requirement as a function of pressure
(Hüls).
690 ACETYLENE Vol. 2
and the use of special processes are two common reasons for continuing to use
acetylene generated from carbide in the chemical industry.
The reaction of calcium carbide and water to form acetylene and calcium
hydroxide is highly exothermic:
Pure calcium carbide has a yield number of 372.66. This means that
the gasification of 1 kg of carbide yields 372.66 L acetylene at 158C and
1013 mbar (101.3 kPa). Commercially available carbide has a yield number of
260–300.
A distinction is made between two groups of acetylene generators (with con-
tinuous rates of production greater than 10 m3 acetylene per hour): the wet type
and the dry type.
In wet generators, the acetylene is converted with a large water excess. In
most cases, a lime slurry containing 10–20 wt% calcium hydroxide is obtained.
The heat of reaction increases the temperature of the generator water and is
removed from the reactor with the lime slurry.
In dry generators, the water mixed with the carbide is just sufficient for che-
mical reaction and for dissipating the heat of reaction. The calcium hydroxide is
obtained in the form of a dry, easily pourable powder having a residual moisture
content of 1–6%. The heat of reaction is dissipated by evaporation of part of the
generator water.
Generators are classified according to their working pressure as
either low or medium pressure. This classification is governed by the regula-
tions concerning acetylene plants and calcium carbide storage facilities
(Acetylenverordnung) (59) issued by Deutscher Acetylenausschuß (German
acetylene committee) and the associated technical rules for acetylene plants
and calcium carbide storage facilities TRAC (Technische Regeln für Acetyle-
nanlagen und Calciumcarbidlager) (60). The Acetylenverordnung and TRAC
constitute a comprehensive set of rules for handling acetylene. Recommenda-
tions in other countries (e.g, United States) deal only with some aspects,
such as safety precautions (61).
Low-pressure generators are designed for a maximum allowable working
pressure of 0.2 bar. They must be rated for an internal pressure of at least
1 bar. Lower pressure ratings are possible if proof is given in each particular
case that the generator can withstand the expected stress (maximum working
pressure, water filling, agitator, etc.; TRAC 201).
Medium-pressure generators have a maximum allowable working pressure
of 1.5 bar. They must be rated for an internal pressure of 24 bar. A design pres-
sure of 5 bar suffices whenever the generators are equipped with rupture disks of
a defined size and specified response pressure (3–4.5 bar, TRAC 201).
Wet Generators. Wet generators are used primarily for the production of
small amounts of acetylene, e.g, for welding purposes. Wet generators work by
one of three different principles (62):
Fig. 28. Medium-pressure wet generator. a) Carbide skip; b) Hopper; c) Feeding system;
d) Gasification chamber; e) Agitator; f) Slurry valve, g) Safety device.
that the water drifts off the basket as a result of the gas pressure at low gas
withdrawal rates and, conversely, returns to the basket when gas withdra-
wal rates increase.
Fig. 29. Knapsack dry generator. a) Chain conveyor; b) Feed bin; c) Star wheel; d) Car-
bide feed screw; e) Generator; f) Lime lockhopper; g) Lime discharge screw; h) Lime scra-
per; i) First scrubbing tower; k) Second scrubbing tower; l) Dip seal.
shell side and at the central agitator shaft. The agitator shaft moves stirrer pad-
dles across the trays.
The carbide first reaches the uppermost tray where the generator water is
also admitted. The reaction mixture consisting of carbide, water, and calcium
hydroxide is pushed by the stirrer paddles towards the outer edge, drops on to
the second tray, returns towards the center, etc. When it reaches the last tray,
the carbide has been fully gasified. The calcium hydroxide, which still contains
up to 6% water, drops into the lime lockhopper (f). Here, a lime layer two meters
deep serves as the gas closure between generator and lime conveying system.
The lime is withdrawn continuously. the gas leaving the generator through the
lime scraper consists of 25% acetylene and ca. 75% water vapor. The water vapor
is the result of dissipating the major portion of the reaction heat. Depending on
the generator load, up to several hundred kilograms of lime hydrate dust are car-
ried along with the acetylene. The lime scraper (h) retains the major portion of
this dust and returns it to the generator. The remainder is sent together with the
gas into the first scrubbing tower (i). Here, lime slurry is sprayed into the hot
acetylene gas (ca. 908C) to scrub out the lime dust; part of the water vapor con-
denses because of the simultaneous cooling. In the second scrubbing tower (k),
the acetylene is sprayed with atomized water to cool the gas below 408C; addi-
tional water vapor condenses here. Any ammonia still present in the gas is
also removed. the acetylene leaves the generator via the dip seal (l). It still con-
tains certain impurities in the form of sulfur and phosphorus compounds. the
Knapsack dry generator is suitable for a carbide throughput of 15 t/h, corre-
sponding to an acetylene quantity of 3750 m3/h. During this process, about
Vol. 2 ACETYLENE 695
17.5 t of calcium hydroxide per hour are obtained. The pressure in this low-pres-
sure generator amounts to approximately 1.15 bar (115 kPa).
The dry generator of Shawinigan Chemicals, Montreal (66), also processes
finely ground carbide and has a variety of applications. It consists of several
superimposed troughs. Carbide and water are fed into the uppermost trough.
The reacting mixture, which is constantly kept in motion by blades, flows over a
weir onto the trough below, etc. At the uppermost trough, water is admitted at
such a rate that carbide-free calcium hydroxide can be withdrawn at the lowermost
trough. The generated acetylene is purified in two scrubbing towers and cooled.
The calcium hydroxide formed (carbide lime) has a wide range of applications,
e.g, in the building industry (for preparing mortar, cement, etc.), in the chemical
industry (for neutralization and for recycling to the carbide furnace), in agriculture
(as fertilizer), and for water purification and waste water treatment (63).
Acetylene Purification. During the gasification of carbide with water,
gaseous compounds become mixed with the acetylene, and these must be
removed because they have a harmful effect on the downstream chemical synth-
esis processes. The impurities are mainly sulfur and phosphorus compounds.
They can be removed by one of the following purification processes.
In the first process, dilute chlorinated water is used as the oxidizing agent.
The chlorine concentration of the water is limited to 1.5 g/L to prevent the forma-
tion of unstable chlorine compounds, which present an explosion hazard. The
chlorine scrubbing step is followed by a caustic soda scrubber to remove the hydro-
gen chloride formed during the oxidation process. The disadvantage of this purifi-
cation process is that considerable quantities of scrubbing water are produced.
The second process uses 98% sulfuric acid as the oxidizing agent (67).
Because very small quantities of sulfuric acid are admitted, it is difficult to dis-
sipate the heats of absorption and reaction. Heating the acetylene results in
increased formation and settling of polymerization products in the purification
stage. For this reason the gas requires additional cooling in the event that the
acetylene contains appreciable quantities of impurities. Moreover, it is recom-
mended that a second scrubbing tower be kept on standby if a high onstream fac-
tor is desired (e.g, 91% ¼ ^ 8000 h/a operating time).
The sulfuric acid scrubber is followed by a caustic soda scrubber, in which
the sulfur dioxide formed during oxidation is removed.
The main advantage of this purification method is that virtually no waste
water is obtained. The small amount of polluted, highly concentrated sulfuric
acid can be used, for example, in fertilizer plants.
These two purification processes yield the following acetylene purities
(by volume):
in the burner by the combustion of residual cracked gas from the acetylene recov-
ery section (CO, H2, CH4) with oxygen. Immediately after combustion, the tem-
perature is about 27008C; this is moderated to about 23008C by the injection of
steam before the reactor is entered. In the second stage, the feedstock naphtha is
injected, and the adiabatic cracking reaction takes place. A final temperature of
about 13008C is reached: This determines the cracked gas composition. By vary-
ing the feed rate of naphtha the acetylene–ethylene ratio can be altered from
30 : 70 to 70 : 30. However, thermodynamic and economic considerations show
that the optimum ratio is 40 : 60.
After a reaction time of a few milliseconds the cracked gas is quenched to
approximately 2508C by the injection of cracked oil from the process. The oil
absorbs heat from the cracked gas and is passed through waste heat boilers, rais-
ing the steam pressure. No soot is formed in this process, even when crude oil is
used as feed because of the high steam content of the carrier gas.
After the oil crisis of 1973 the process became uneconomical in spite of its
high thermal efficiency, and in 1976 it was shut down after 15 years of operation.
However, one unit is still running in Czechoslovakia.
Kureha, Chiyoda, Union Carbide Advanced Cracking Reactor
(ACR) Process (71). To avoid dependence on oil refineries or gas processors
for the supply of feedstocks, processes for directly cracking crude oil have been
developed by various companies for the production of olefins. Some of these pro-
cesses operate at reaction temperatures intermediate between those of the usual
crack processes for olefins and those for acetylene. The ACR process (Fig. 31) uses
a multi-port burner to produce a heat carrier gas of 20008C by the combustion of
H2 – CH4 mixtures with oxygen in the presence of steam preheated to 8008C.
The oil to be cracked is introduced through nozzles into the stream of car-
rier gas and passes into an advanced cracking reactor, where the reaction takes
place adiabatically at 5 bar (0.5 MPa). The initial temperature is 16008C; the
Fig. 31. Advanced cracking reactor process (ACR). a) Crude distillation column; b) Bur-
ner; c) Advanced cracking reactor; d) Ozaki quench cooler; e) Oil gasoline fractionator;
f) Compressor; g) Acid gas removal column; h) Gas separator.
698 ACETYLENE Vol. 2
final temperature at the exit of the reactor is 700–9008C after a residence time of
10–30 ms. The cracked gas is quenched by oil in an Ozaki quench cooler (Fig. 32),
where steam production up to 120 bar (12 MPa) is possible. This particular boiler
design was developed for a high heat transfer rate without coke formation on the
exchanger surfaces. Yields reported for Arabian light crude oil are 11.2 wt%
hydrogen and methane, 40.7 wt% olefins, and 4.2 wt% acetylene. The acetylene
yield is about ten times higher than in usual olefin processes.
Dow Partial Combustion Cracking (PCC) Process (71). The basic
idea of this process is to reduce coking and soot formation considerably when
heavy feeds are cracked and when hydrogen is present in the reaction mixture.
The PCC process (Fig. 33), which accepts crude oil and heavy residue as feed-
stock, attains a high partial pressure of hydrogen in the reaction zone by recy-
cling the quench oil (produced in the process) to the burner where it is partially
oxidized to yield synthesis gas. Thus there is no need to find a use for the quench
oil as in the case of the ACR process. Starting from residual oil boiling above 3438C,
yields are given as 12.4 wt% methane, about 38 wt% alkenes, and 2.5 wt% acety-
lene. This is seven to eight times more acetylene than that obtained from a steam
cracker, but less than the acetylene yield of the ACR process, because of a resi-
dence time in the reaction zone which is three to ten times longer.
Acetylene as a Byproduct of Steam Cracking. In a steam cracker satu-
rated hydrocarbons are converted to olefinic products such as ethylene and pro-
pylene. Besides these desired components, acetylene and many other products
are formed in the cracking process (Fig. 36). The concentration of acetylene
depends on the type of feed, the residence time, and temperature (cracking
Vol. 2 ACETYLENE 699
Fig. 33. Dow partial combustion cracking process (PCC). a) Reactor; b) Quench boiler;
c) Quench column; d) Stripper; e) Decanter.
General aspects of the process and the catalyst requirements are reviewed
in (73,74).
Table 10. Yields of Unsaturated Components (wt%) in Raw Gas From Steam Cracking
Feedstock Cracking severity Acetylene Propyne Propadiene
Ethane 65% convers. 0.4–0.50 0.04 0.02
LPG 90% convers. 0.65–1.20 0.63 0.35
Full-range naphtha P/E: 0.4 0.9–1.05 0.81 0.54
Full-range naphtha P/E: 0.53 0.5–0.70 0.68 0.50
Full-range naphtha P/E: 0.65 0.25–0.42 0.46 0.38
Atmospheric gas oil P/E: 0.55 0.40 0.34 0.29
Hydrocracker residue P/E: 0.55 0.50 0.36 0.31
700 ACETYLENE Vol. 2
Melting point lower than the dew point of the feed gas
High solubility of acetylene at a temperature near the dew point of the C2
fraction
High acetylene selectivity (9,11)
High chemical and thermal stability
No foaming tendency due to traces of hydrocarbons
Low toxicity
Low vapor pressure at the operating temperature
Table 11. Material Balance (mol %) for an Acetylene Recovery Process Operating on the
C2 Fraction from a Plant Producing 400 000 t/a C2H4 (Linde) (76)
Gas to absorber Purified gas Recycle gas Product C2H2
Methane trace trace
Acetylene 2 1 ppm 4 99.8
Ethylene 82 83.5 85.7 0.2
Ethane 16 16.5 10 trace
C3 trace trace 0.3 trace
DMF 1 ppm trace trace
Temperature, K 252 249 255 258
Pressure, MPa 2 1.98 0.11 0.12
Flow rate, kmol/h 2186 2126 17.5 52.5
After cooling and heat recovery, the acetylene-free solvent is recycled to the
absorber and ethylene stripper. The acetylene product has a purity of > 99.8%
and a DMF content of less than 50 ppm and is available at 10 kPa and ambient
temperature.
The material balance and the utilities consumption of an acetylene recovery
unit are listed in Tables 11 and 12. At present, more than 112 000 t/a of petro-
chemical acetylene from twelve olefin plants worldwide is recovered by this tech-
nology; three other facilities are under construction (1998). Figure 35 shows an
industrial plant with a design capacity of 14 400 t/a of high-purity acetylene.
A material balance for ethylene plant outputs including acetylene extrac-
tion or hydrogenation is shown in Figure 36 (72). The economic evaluation
shows that petrochemical acetylene remains attractive even if the price of ethy-
lene is doubled. It is economical to retrofit acetylene absorption in an existing
olefin plant equipped with a catalytic hydrogenation.
A similar process is available for propyne (see Chap. 7)
Table 12. Consumption of Utilities for an Acetylene Recovery Process Operating on the
C2 Fraction from a Plant Producing 400 000 t/a C2H4 (Linde) (76)
DMF, kg/h 1.3
Heating steam, t/h 3.9
Cooling water, m3/h 100
Electrical energy, kW 125
Refrigerant, GJ/h 6.3
Quench water, GJ/h 3.1
Plot area, m m 15 40
702 ACETYLENE Vol. 2
Fig. 35. Acetylene recovery plant (name plate capacity: 14 400 t/a of high-purity
acetylene).
of about 31008C, but due to formation of other products, the temperature reached
adiabatically is 2800–29008C. The decomposition can be initiated by heat of reac-
tion, by contact with a hot body, by an electrostatic spark, by compression heat-
ing, or by a shock wave. The decomposition induced by heating the wall of the
container or pipe is very sensitive to the pressure, the size and shape of the
container or the diameter of the pipe, the material of the container, and traces
of impurities or other components. Solid particles such as rust, charcoal,
alumina, and silica can lower the ignition temperature compared to clean
steel pipe.
Decomposition gives rise to different scenarios:
Fig. 36. Material balances of a 300 000 t/a ethylene plant equipped with either C2 hydro-
genation or acetylene extraction (all rates in kg/h, the numbers in parentheses are for the
solvent extraction process) (72). *Chemical grade.
of deflagration pressure on the flow in pipes (85) and the decomposition of high-
pressure acetylene in branched piping (86). Solid acetylene is not critical with
regard to decomposition, provided it is the only material involved (8). In liquid
oxygen, solid acetylene can readily ignite on mechanical impact and react vio-
lently (87). Recommendations for equipment used in gas welding and cutting
technology, such as rubber hoses, safety devices, and flame arresters, are
given in (88,89,90).
Combustion of Acetylene in Oxygen (Air). The reaction of acetylene
and oxygen at 258C and 1 bar to form water and CO2 generates 1255.6 kJ/mol.
Table 13. Stability Pressure (bar) of Acetylene and Acetylene Mixtures for Two Methods
of Ignition and Pressure Increase (81,82)
Ignition method pex/pa
Mixture Reppe BAM (BAM)
100% C2H2 1.4 0.8
90% C2H2/10% N2 1.8 1.0 8.6
90% C2H2/10% CH4 2.1 1.0 7.2
90% C2H2/10% H2 1.6 0.9 6.9
50 % C2H2/50% N2 9.0 3.6 6.3
50% C2H2/50% CH4 14.7 12.9
50% C2H2/50% H2 4.7 2.5 5.5
*pex/pa: ratio of maximum pressure to pressure before ignition.
Vol. 2 ACETYLENE 705
Fig. 38. Decomposition pressure versus ignition energy for unsaturated hydrocarbons.
Fig. 39. Flame temperatures and ignition velocities of acetylene – oxygen mixtures and
mixtures of other hydrocarbons with oxygen.
706 ACETYLENE Vol. 2
Fig. 40. Chemical composition of an oxygen – acetylene flame at its tip versus mixing
ratio.
formulate general safety instructions for the great variety of chemical processes
with acetylene as reaction component under diverse reaction conditions.
Both handling acetylene and experiments with it necessitate critical exam-
ination of sources of possible danger. The literature cited can only serve as an aid
to decisions on precautions. The safety regulations mentioned above have been
determined in experiments with well-defined apparatus dimensions (length, dia-
meter, geometry). For other dimensions they can only serve as an indication of
explosive behavior and should not be considered as rigid limits. The development
of economical chemical processes involving acetylene at elevated pressures or
under other hazardous conditions calls for decomposition tests for the crucial
stages where decomposition could occur. This must be done in close cooperation
with official testing institutions, such as the Bundesanstalt für Materialprüfung
(BAM) in the Federal Republic of Germany.
In general, the following rules should be observed in handling acetylene:
Table 14. Fundamental Safety Data for Acetylene – Air and Acetylene – Oxygen Mixtures
Air Oxygen
Lower flammability limit, vol% 2.5 2.4
Upper flammability limit, vol% 82 93
Flame temperature*, K 2863 3343
Flame front velocity, m/s 1.46 7.6
Increase of pressure (deflagration) 11 50
Detonation velocity, m/s 2300 2900
*Stoichometric mixture.
Vol. 2 ACETYLENE 707
Fig. 41. Hydraulic flame trap for acetylene lines (Union Carbide), (97).
Fig. 42. The 8 km acetylene pipeline from the Marathon refinery, Burghausen, to Farb-
werke Hoechst, Gendorf, Federal Republic of Germany (98). a) Compressor; b) Control
points; c) Automatic quick-closing valves; d) Rupture disks to atmosphere; e) Flame traps;
f) Pipeline.
Vol. 2 ACETYLENE 709
Table 15. Permitted Acetylene and Acetone Filling of Seamless Gas Cylinders
(Satisfying German Standards and Safety Rules) (112)
Gas cylinder Acetone filling, kg
pressure, and ignition and impact testing. Table 16 lists three porous materials
approved for use in Germany. Figure 45 shows a photograph of ‘‘Linde M1’’ mag-
nified 1 : 10000, clearly revealing the porous structure of such materials.
Methods for examining the materials have been standardized by CPI
(Commission Permanente Internat. de l’Acétylène, de la Soudure Autogène et
des Industries qui s’y rattachent, Paris) and ISO (International Organization
for Standardization) (117).
Discharging acetylene from a gas cylinder leads to acetone losses because
the partial pressure of acetone at 158C ranges from 0.14 bar at 15 bar total
Fig. 45. Porous silica material Linde M1 for acetylene cylinders (magnification
1 : 10000).
Vol. 2 ACETYLENE 713
pressure to 0.18 bar at 1 bar total pressure. Solvent loss has to be replaced when
an acetylene cylinder is reloaded.
Fig. 47. Heating efficiency of acetylene – oxygen mixtures and mixtures of other hydro-
carbons with oxygen. *based on area of primary flame cone
714 ACETYLENE Vol. 2
main products from acetylene, are produced today mainly from ethylene and
propene (3).
6.3. Competitive Position of Acetylene as Chemical Feedstock.
Today, acetylene plays an important role only in the production of the acetylenic
chemicals. The fact that acetylene production has not decreased further seems to
indicate that the competition from the olefins is no longer as strong as it was. The
main reason for this is that European olefin chemistry depends on refinery pro-
ducts, which have become more expensive than natural gas, the main feedstock
for acetylene. Another contributing factor is that acetylene is produced only in
old plants, which have low capital costs.
In addition, process improvements, such as an increase in thermal effi-
ciency and optimum use of byproducts by other plants, can make acetylene
more competitive. The position of acetylene in chemical industry may be
advanced because of the variety of valuable products to which it can be converted
in high yields with known technology. Acetylene must compete with ethylene for
the production of vinyl chloride and vinyl acetate, and for the production of
acrylic acid and its esters it must compete with propene. The Stanford Research
Institute (3) has investigated this question in detail. The results, which take into
account both capital investment (25% return) and specific consumption figures
for the various processes, show that the prices at which the alternate routes
are competitive can be expressed by the following equations:
where A is the acetylene price, E is the ethylene price, and P is the propene price,
all in $/kg. For example, if ethylene costs 0.65 $/kg, acetylene can cost 1.15 $/kg
for the production of vinyl chloride and 1.21 $/kg for the production of vinyl
acetate. If propene costs 0.49 $/kg, acrylic acid can be profitably produced from
acetylene if it costs 1.08 $/kg or less.
Acetylene prices of 1.08 to 1.21 $/kg or less can be reached in a new plant
only with optimal integration of energy and byproducts within an integrated
chemical plant. Figure 49 shows the flow sheet of a 100000 t/a gas-based partial
oxidation acetylene plant in a chemical complex. The required oxygen facility
and methanol plant based on acetylene synthesis gas (off-gas) are included.
The main products of the complex are acetylene and methanol. The acetylene
process is operated with improved quenching technology, allowing a high
proportion of energy to be regained in the form of steam. The aromatic
residue oil from a steam cracker is converted into high-purity coke and light
aromatics.
A production cost estimate based on power, feedstock, and product prices
roughly corresponding to market prices in 1982 shows that low production
costs for acetylene are possible under the conditions described, in spite of the
relatively high natural gas price of 5.5 $/106 BTU (0.021 $/kW h), which in heat-
ing value terms comes very close to that of crude oil (0.022 $/kW h). The differ-
ence in production costs between acetylene from natural gas and ethylene
from naphtha (price 333 $/t, corresponding to 0.028 $/kW h) is so slight that acet-
ylene from new plants can once again compete with ethylene for certain synth-
eses, provided that there is a difference between the costs of natural gas and
naphtha. This was nearly 0.007 $/kW h in 1982 in Germany. A similar calcula-
tion for the acetylene production by the Hüls arc process based on an ethylene
price of 600 $/t and a hydrogen price 40% above the heating value results in
the same acetylene product value at 0.038 $/kW h for electrical power. Lower
electrical energy costs favor the arc process.
Vol. 2 ACETYLENE 717
7. Propyne
Table 19. Solubility Coefficients of Propyne in Various Solvents (mol kg1 bar1)(9)
Solvent C3H4 pressure, bar 20 8C 25 8C
Methanol 0.1 9.099 1.865
1,2-Dichlorethane 0.098 – 0.196 3.276 (0 8C) 1.546
Carbon tetrachloride 0.25 4.732 0.842
n-Octane 0.098 3.412
Toluene 0.25/1.0 8.644 2.047
Xylene (technical). 0.49/0.98 9.782 1.683
4-Methyl-1,3-dioxalan-2-one 0.49/1.0 8.644 1.183
Triethylene glycol 1.0 0.400 (30 8C)
Acetone 0.6/1.0 35.03 4.186
N-Methyl-2-pyrrolidone 0.78 6.597 (0 8C) 2.502
DMF 0.15 12.512 3.003
Water 1.0 0.071
Ammonia 0.04 20.018 4.436
Vol. 2 ACETYLENE 719
Fig. 53. Propyne recovery process (Linde, Shell (75)). MA: methylacetylene; PD:
propadiene.
Trade names are Tetrene or MAPP gas. These mixtures are stabilized by propa-
ne,propene, and/or butane and are used for metal cutting, welding, hardening,
and brazing. The flame properties are closer to those of propane–propene mix-
tures. Therefore stabilized C3H4 mixtures have not yet won a large-scale market.
Further information is available in (131).
REFERENCES
6. DIPPR Database, STN International, Design Institute for Physical Property Data
c/o DIPPR Project Staff, Pennsylvania State University 167 Fenske Lab, University
Park, Pa 16802 U.S.A American Institute of Chemical Engineers, 3345 E. 47th
Street, New York, NY 10017, U.S.A., basis February 1998.
7. Beilstein Database, STN International, Beilstein Informationssysteme Carl-Bosch-
Haus, Varrentrappstr. 40–42, D-60486 Frankfurt am Main, Germany, basis Febru-
ary 1998.
8. W. Wiechmann, Amts- und Mitteilungsblatt der Bundesanstalt für Material-
forschung und -prüfung (BAM) (1987) 3, 505.
9. A. Kruis: ‘‘Gleichgewicht der Absorption von Gasen in Flüssigkeiten,’’ in H. Hausen
(ed.): Landolt-Börnstein Zahlenwerte und Funktionen, 4. Teil, Bestandteil c,
Springer Verlag, Berlin–Heidelberg–New York 1976.
10. S. A. Miller: Acetylene, Its Properties, Manufacture and Uses, Ernest Benn, London
1965.
11. F. Rottmayr, H. Reimann, U. Lorber, Linde Ber. Tech. Wiss. 30 (1971) 3.
12. H. Schmidt, D. Forney: Oxygen Technology Survey, vol. 9, NASA SP-3090, ASRDI,
1975.
13. E. Lassmann, Linde Reports on Science and Technology 57 (1996) 16.
14. UCC isotherm data sheet No. 44, Union Carbide Corp., New York 1980.
15. D. P. Velenuela, A. L. Myers: Adsorption Equilibrium Data Handbook, Department
of Chem. Engineering, University of Pennsylvania, Philadelphia, Prentice Hall,
Englewood Cliffs, NJ 1989.
16. T. F. Rutledge: Acetylenic Compounds–Preparation and Substitution Reactions,
Reinhold Publ. Co., New York 1968.
17. A. C. Mc Kinnis, Ind. Eng. Chem. 27 (1962) 2928.
18. S. A. Miller: Acetylene, Its Properties, Manufacture and Uses, Ernest Benn, London
1965.
19. R. C. West, C. S. Kraihanzel, J. Am. Chem. Soc. 83 (1961) 765.
20. H. J. Copley, C. E. Holley Jr., J. Am. Chem. Soc. 61 (1939) 1599.
21. W. Reppe: Neue Entwicklungen auf dem Gebiet der Chemie des Acetylens und
Kohlenoxids, Springer Verlag, Berlin-Göttingen-Heidelberg 1949.
22. W. Reppe: Chemie und Technik der Acetylen-Druckreaktionen, 2nd ed., Verlag
Chemie, Weinheim, Germany 1952.
23. N. von Kutepow in: Ullmann, 4th ed., vol. 7, p. 44.
24. W. Reppe, Justus Liebigs Ann. Chem. 601 (1956) 81.
25. W. Reppe et al., Justus Liebigs Ann. Chem. 569 (1955) 1.
26. W. Reppe et al., Justus Liebigs Ann. Chem. 582 (1953) 1.
27. W. Reppe, N. von Kutepow, A. Magin, Angew. Chem. 81 (1969) 717.
28. Lonza, DE 1251329, 1964 (P. Pino, G. Braca, G. Sbrana).
29. BASF, DE-AS 1071077, 1965 (W. Reppe, A. Magin). Du Pont, DE 1054086, 1955
(J. C. Sauer). J. C. Sauer, R. D. Cramer, V. A. Engelhardt, T. A. Ford, H. E. Holmquist,
B. W. Howk, J. Am. Chem. Soc. 81 (1959) 3677. G. Albanesi, M. Toraglieri, Chim. Ind.
(Milan) 41 (1959) 189.
30. G. Palyi, G. Varadi, J. T. Horvath, J. Mol. Catal. 13 (1981) 61.
31. W. Reppe et al., Justus Liebigs Ann. Chem. 560 (1948) 1.
32. G. Schröder: Cyclooctatetraen, Verlag Chemie, Weinheim, Germany 1965.
33. J. A. Nieuwland, W. S. Calcott, F. B. Downing, A. S. Carter, J. Am. Chem. Soc.
53 (1931) 4197.
34. J. C. W. Chien, Polym. News 6 (1979) 52. D. Bloor, New Sci. 93 (1982) no. 1295, 577.
H. Shirakawa, Kotai Butsuri 16 (1981) no. 7, 402; Chem. Abstr. 96 (1982) 172568.
A. J. Heger, A. G. Macdiarmid, Int. J. Quantum Chem. Quantum Chem. Symp. 15
(1981) 243. H. Shirakawa, Kagaku (Kyoto) 37 (1982) no. 3, 181; Chem. Abstr. 96
Vol. 2 ACETYLENE 723
(1982) 153398. H. Shirakawa, Kagaku, Zokan (Kyoto) 1980, no. 87, 165; Chem.
Abstr. 94 (1981) 139105 x.
35. G. Wegner, Angew. Chem. 93 (1981) 352.
36. T. Mole, J. R. Suertes, Chem. Ind. (London) 1963, 1727.
37. P. N. Rylander, ‘‘Platinum Metal Catalysts in Organosilicon Chemistry,’’ Engelhard
Ind. Tech. Bull. 10 (1970) no. 4, 130.
38. J. L. Speier, Adv. Organomet. Chem. 17 (1979) 407.
39. E. Ott et al., Ber. Dtsch. Chem. Ges. 76 (1943) 80–91.
40. I. G. Farbenind., DE 495787, 1927.
41. Chem. Week 98 (1966) Apr. 16, 90.
42. Eur. Chem. News 25 (1974) Feb. 1, 5.
43. H. Sachsse, Chem. Ing. Tech. 26 (1954) 245. E. Bartholome, Chem Ing. Tech.
26 (1954) 253. T. P. Forbath, B. I. Gaffney, Pet. Refiner 33 (1954) 160.
44. H. Friz, Chem. Ing. Tech. 40 (1968) 999.
45. K. G. Baur, K. Taglieber, Chem. Ing. Tech. 47 (1975) 385.
46. M. J. Zundel: Proble`mes de Fabrication d’Acetylène et d’Ethylène à partir d’Hydro-
carbures, Societe de Chimie Industrielle, Paris 1962.
47. Hydrocarbon Process. Pet. Refiner 44 (1965) no. 11, 163.
48. Montecatini, GB 1000480, 1962.
49. J. L. Petton et al., Pet. Refiner 37 (1958) 180.
50. Hydrocarbon Process. 54 (1975) no. 11, 104. Stanford Research Report no. 109,
September 1976. L. Verde, R. Riccardi, S. Moreno, Hydrocarbon Process. 57 (1978)
no. 1, 159.
51. G. Hamprecht, M. Gettert: ‘‘Sauerstoff-Thermisches Calziumcarbid,’’ Festschrift für
Carl Wurster, BASF, Ludwigshafen 1960, p. 43.
52. H. Gladisch, Hydrocarbon Process. Pet. Refiner 41 (1962) 159–164.
53. H. Gladisch, Chem. Ing. Tech 61 (1969) 204–208.
54. K. Gehrmann, H. Schmidt, Water Air Conserv. Pet. Ind, 1971, 379.
55. R. L. Bond, W. R. Ladner, G. I. McCommet, Fuel 45 (1966) 381–395. R. E. Gannon,
V. J. Krukonis, Th. Schoenberg, Ind. Eng. Chem. Prod. Res. Dev. 9 (1970) 343–347.
R. E. Gannon, S. K. Ubhayakar, Fuel 56 (1977) 281. AVCO Corp., US 179144,
1980 (Ch. Kim). D. Bittner, H. Baumann, C. Peuckert, J. Klein, H. Jüntgen, Erdöl
Kohle Erdgas Petrochem. 34 (1981) no. 6, 237–242. R. Müller, C. Peuckert,
Int. Symp. on Plasma Chem., 5th, Edinburgh 1981, p. 197–202.
56. Eur. Chem. News 36 (1981) Feb. 2, 19.
57. R. Müller, World Hydrogen Energy Conf., Conf. Proc., 4th, Pasadena, 1982,
p. 885–900.
58. GAF, US 219756, 1980 (M. Katz, F. Carluccio).
59. Verordnung über Acetylenanlagen und Calcium-carbidlager (Acetylenverordnung–
Acet V) vom 27. Februar 1980, Carl Heymanns Verlag, Köln 1980.
60. TRAC 201–Acetylenentwickler, Carl Heymanns Verlag, Köln.
61. Acetylene Transmission for Chemical Synthesis (Recommended Minimum
Safe Practices for Piping Systems) International Acetylene Association, New York
1980.
62. C. Hase, W. Reitze: Fachkunde des Autogenschweißens, 7th ed., Girardet, Essen
1965.
63. Carbidkalk, Hinweise für seine Verwendung, Fachbuchreihe Schweißtechnik,
9th ed., Deutscher Verlag für Schweißtechnik (DVS), Düsseldorf 1968, p. 16.
64. Shawinigan Chemicals, US 1872741, 1931 (R. S. Jane).
65. Bayerische Stickstoff-Werke, DE 714 323, 1938 (R. Wendlandt, R. Neubner).
66. Shawinigan Chemicals, US 1343185, 1942 (A. C. Holm, E. Poirier).
67. Linde, DE 906 005, 1951 (F. Rottmayr). Linde, DE 2 549 399, 1975 (E. Laßmann).
724 ACETYLENE Vol. 2
68. Hydrocarbon Process. 46 (1967) no. 11, 139. T. Wett, Oil Gas J. 70 (1972)
101–110.
69. Ullmann, 4th ed., vol. 7.
70. H. K. Kamptner, Erdöl Kohle Erdgas Petrochem. 16 (1963) 547. H. K. Kamptner,
W. R. Krause, H. P. Schilken, Hydrocarbon Process. Pet. Refiner 45 (1966) no. 4,
187. K. Lissa, Chem. Anlagen Verfahren 1970, no. 6, 83.
71. Y. C. Hu, Hydrocarbon Process. 61 (1982) no. 11, 109.
72. D. Sohns, Linde Reports Science and Technology 30 (1979) 21.
73. F. Mey, H. D. Neubauer, R. Schubert, Petrochemicals and Gas Processing Petroleum
Technology quaterly (1997) Autumn, 119.
74. K. J. Sasaki, Petrochemicals and Gas Processing Petroleum Technology quaterly
(1997) Autumn, 113.
75. P. Cl. Haehn, Dr. E. Haidegger, Dr. N. Schödel, Hydrocarbon Engineering (1997)
January/February, 41.
76. Linde München folder C/3.3.e/93
77. G. Marcks, Schadenprisma 3 (1982) 37. D. Lietze, Amts- und Mitteilungsblatt der
Bundesanstalt für Materialprüfung (BAM) 16 (1986) 1, 23.
78. W. Reppe, Chem. Ing. Tech. 22 (1950) 273.
79. CEN/TC 121/SC 7/WG3 German Proposal: Pipelines in Acetylene Systems,November
1996, NAS-DIN Deutsches Institut für Normung e.V., D 10772 Berlin or ‘‘Empfeh-
lungen für Acetylenleitungen auf der Grundlage von Arbeitsbereichen’’, Industrial
Gases Committee IGC Document 9/78/D, F 75880 Paris, CEDEX 18
80. H. B. Sargent, Chem. Eng. 64 (1957) 250.
81. Th. Schendler, H.-P. Schulze, Chem.-Ing.-Tech. 62 (1990) 1, 41. W. Reppe: Chemie
und Technik der Acetylen-Druck-Reaktionen, Verlag Chemie, Weinheim, Germany
1951.
82. D. Lietze, H. Pinkofsky, T. Schendler, H.-P. Schulze, Chem.-Ing.-Tech. 61 (1989) 9,
736. H. Große-Wortmann, N. Kalkert, H.-G. Schecker, Chem.-Ing.-Tech. 53 (1981)
461.
83. A. Williams, D. B. Smith, Chem. Rev. 70 (1970) 267. B. A. Ivanov, S. M. Kogarko,
Int. Chem. Eng. 4 (1964) 4, 670. C. M. Detz, Combust. Flame 34 (1979) 187. M. A.
Glikin et al., Sov. Chem. Ind. (Engl. Transl.) 7 (1975) 1373. A. Baumeier, D.
Conrad, S. Dietlen, W. Pezold, T. Schendler und H.-P. Schulze, Chem.-Ing.-Tech.
64 (1992) 3, 260.
84. R. Grätz, M. Wagenknecht, vfdb, Zeitschrift für Forschung, Technik und Manage-
ment im Brandschutz (1994) 3, 103.
85. D. Lietze, Chem. Ing.-Tech. 63 (1991) 11, 1148.
86. D. Lietze, Chem.-Ing.-Tech. 62 (1990) 3, 238.
87. E. Karwat, Chem Eng. Progr. 54 (1958) 10, 96.
88. D. Lietze, Journal of Hazardous Materials. 54 (1997) 227.
89. D. Lietze, J. Loss Prev. Process Ind. 8 (1995) 6, 319.
90. D. Lietze, J. Loss. Process Ind. 8 (1995) 6, 325.
91. A. D. Hewitt: Technology of Oxy-Fuel Gas Processes, Welding and Metal Fabrica-
tion, Part 2, 1972 (November) 382. H. Meinass, H. Manhard, J. Schlander, L. Fruh-
storfer, Linde Reports on Science and Technology 24 (1976) 32. L. Kögel, Linde
Reports on Science and Technology 32 (1981) 36.
92. Technische Regeln für Acetylenanlagen und Calciumcarbidlager, TRAC 207, 208,
Deutscher Acetylenausschuß, Bundesanstalt für Arbeitsschutz und Unfall-
forschung, Dortmund 1969.
93. K. H. Roch, Amts Mitteilungsbl. Bundesanst. Materialprüf. Berlin 12 (1982) 283.
94. K. H. Roch, Schweissen + Schneiden 25 (1973) 94.
95. D. Lietze, Amts Mitteilungsbl. Bundesanst. Materialprüf. Berlin 2 (1972) 9.
Vol. 2 ACETYLENE 725
126. Shell Internationale, EP 0 392 601, 1990 (M. J. Doyle, J. Van Gogh, J. Van Ravens-
waay Claasen)
127. D. A. Frank-Kamenetzki, V. G. Markovich, Acta Physicochim, (U.R.S.S.) 17 (1942)
308. H. Zeise (ed.): Thermodynamik, vol. III/1, Verlag Hirzel, Leipzig 1954.
128. T. Redeker, G. Schön, Sicherheitstechnische Kennzahlen brennbarer Gase und
Dämpfe, 6. Nachtrag Deutscher Eichverlag GmbH, Braunschweig 1990.
129. Linde AG, US 4 701 190, 1986 (P. C. Haehn), EP 0 224 748, 1986 (P. C. Haehn).
130. J. Happel, S. Umemura, Y. Sakakibara, H. Blanck, S. Kunichika, Ind. Eng. Chem.
Process Des. Develop. 14 (1975) 1, 44. E. Drent, P. Arnoldy, P. H. M. Budzelaar,
Journal of Organometallic Chemistry, 455 (1993) 247. Shell Internationale, EP
0 539 628, 1993 (J. Hengeveld, P. B. de Blank).
131. R. F. Huston, C. A. Barrios, R. A. Holleman, J. Chem. Eng. Data 15 (1970) 168. R. D.
Green, Met. Prog. 108 (1975) 2, 71. A. Farwer, Gas Aktuell (1982 Nov.) 24.
132. L. T. Fairhall: Industrial Toxicology, 2nd ed., The Williams and Wilkins Co.,
Baltimore 1957, p. 270.
133. The International Technical Information Institute: Toxic and Hazardous Industrial
Chemicals Safety Manual, Tokyo, 1975, p. 10.
134. T. Sollmann: A Manual of Pharmacology and its Application to Therapeutics, 7th
ed., W. B. Saunders Co., Philadelphia-London 1949, p. 662.
135. F. Flury, F. Zernik: Schädliche Gase, Springer Verlag, Berlin 1931, p. 270.
136. NIOSH, U.S. Dept. of Health and Human Services: Registry of Toxic Effects of
Chemical Substances, vol. 1, Washington, DC, 1980, p. 70.
137. E. R. Plunkett: Handbook of Industrial Toxicology, Heyden, Barberton, Ohio, 1976,
p. 8.
138. F. A. Patty: Industrial Hygiene and Toxicology, 2nd ed., Interscience Publ.,
New York 1962, p. 1205.
PETER PÄSSLER
WERNER HEFNER
BASF Aktiengesellschaft, Ludwigshafen,
Federal Republic of Germany
KLAUS BUCKL
HELMUT MEINASS
HANS-JÜRGEN WERNICKE
Linde AG, Höllriegelskreuth, Federal
Republic of Germany
GÜNTER EBERSBERG
RICHARD MÜLLER
Degussa-Hüls AG, Marl, Federal Republic
of Germany
JÜRGEN BÄSSLER
Uhde GmbH, Dortmund, Federal Republic
of Germany
HARTMUT BEHRINGER
DIETER MAYER
Hoechst Aktiengesellschaft, Pharma,
Forschung, Toxikologie, Frankfurt,
Federal Republic of Germany
Vol. 2 ETHYLENE 727
ETHYLENE
1. Introduction
2. Physical Properties
Ethylene is a colorless flammable gas with a sweet odor. The physical properties
of ethylene are as follows:
mp 169.158C
bp 103.718C
Critical temperature, Tc 9.908C
Critical pressure, Pc 5.117 MPa
Critical density 0.21 g/cm3
Density
at bp 0.57 g/cm3
at 08C 0.34 g/cm3
Gas density at STP 1.2603 g/L
Density relative to air 0.9686
Molar volume at STP 22.258 L
Surface tension
at bp 16.5 mN/m
at 08C 1.1 mN/m
Heat of fusion 119.5 kJ/kg
Heat of combustion 47.183 MJ/kg
Heat of vaporization
at bp 488 kJ/kg
at 08C 191 kJ/kg
Specific heat
of liquid at bp 2.63 kJ kg1 K1
of gas at Tc 1.55 kJ kg1 K1
Enthalpy of formation 52.32 kJ/mol
Entropy 0.220 kJ mol1 K1
Thermal conductivity
at 08C 177104 W m1 K1
at 1008C 294104 W m1 K1
at 4008C 805104 W m1 K1
3. Chemical Properties
The chemical properties of ethylene result from the carbon – carbon double
bond, with a bond length of 0.134 nm and a planar structure. Ethylene is
a very reactive intermediate, which can undergo all typical reactions of
a short-chain olefin. Due to its reactivity ethylene gained importance as a
chemical building block. The complex product mixtures that have to be
separated during the production of ethylene are also due to the reactivity of
ethylene.
Ethylene can be converted to saturated hydrocarbons, oligomers, polymers,
and derivatives thereof. Chemical reactions of ethylene with commercial impor-
tance are: addition, alkylation, halogenation, hydroformylation, hydration, oligo-
merization, oxidation, and polymerization.
The following industrial processes are listed in order of their 1993
worldwide ethylene consumption (6):
4. Raw Materials
Table 1 lists the percentage of ethylene produced worldwide from various feed-
stocks for 1981 and 1992 (7). In Western Europe and Japan, over 80% of ethylene
is produced from naphthas – the principal ethylene raw materials.
A shift in feedstocks occurred for the period from 1980 to 1991. In the
United States and Europe larger amounts of light feedstocks (LPG: propane þ
butane) and NGL (ethane, propane, butane) are used for ethylene production,
whereas in Japan more naphtha was used in 1991 compared to 1981. The use
of gas oils for ethylene production decreased slightly during the 1980s.
Ethane [74-84-0] is obtained from wet natural gases and refinery waste
gases. It may be cracked alone or as a mixture with propane. Propane [74-98-6]
is obtained from wet natural gases, natural gasolines, and refinery waste gases.
Butanes are obtained from natural gasolines and refinery waste gases. A mixture
of light hydrocarbons such as propane, isobutane [75-28-5], and n-butane [106-97-81],
commonly called liquefied petroleum gas (LPG) and obtained from natural gasolines
and refinery gases, is also used as a feedstock.
Naphthas, which are the most important feedstocks for ethylene pro-
duction, are mixtures of hydrocarbons in the boiling range of 30 –2008C.
Processing of light naphthas (boiling range 30 –908C, full range naphthas
(30 –2008C) and special cuts (C6 – C8 raffinates) as feedstocks is typical for
naphtha crackers.
A natural-cut full-range naphtha contains more than 100 individual compo-
nents, which can be detected individually by gas chromatography (GC). Depend-
ing on the origin naphtha quality can vary over a wide range, which necessitates
quality control of the complex feed mixtures. Characterization is typically based
on boiling range; density; and content of paraffins (n-alkanes), isoalkanes, ole-
fins, naphthenes, and aromatics (PIONA analysis) by carbon number. This char-
acterization can be carried out by GC analysis or by a newly developed infrared
method (8). Full characterization of feedstocks is even more important when pro-
duction is based on varying feedstocks, eg, feedstocks of different origins pur-
chased on spot markets.
The quality of a feedstock is depending on the potential to produce the
target products (ethylene and propylene). Simple yield correlations for these
products can be used to express the quality of a feedstock in a simple figure,
Table 1. Raw Materials for Ethylene Production (as a Percentage of Total Ethylene
produced)
USA W. Europe Japan World
Raw
materials 1979 1991 1981 1991 1981 1991 1981 1991
Refinery 1 3 2 17
gas
LPG, NGL 65 73 4* 14 10* 2* 31* 27
Naphtha 14 18 80 72 90 98 58 48
Gas oil 20 6 16 12 0 0 11 8
*Including refinery gas.
730 ETHYLENE Vol. 2
the quality factor, which indicates wether yields of the target products are high
or low, with aromatic feedstocks being poor and saturated feedstocks being good
feedstocks.
Quality characterization factors for naphthas have been developed, which
indicate the aromatics content by empirical correlation. Since aromatics contri-
bute little to ethylene yields in naphtha cracking, a rough quality estimate can be
made for naphthas with a typical weight ratio of n- to isoparaffins of 1–1.1. The
K factor is defined as (9):
ð1:8Tk Þ1=3
K¼
d
where T is the molal average boiling point in K and d is the relative density
d15:6
15:6 . A high value of BMCI indicates a highly aromatic naphtha; a low value,
a highly saturated naphtha. Gas oils are feedstocks that are gaining import-
ance in several areas of the world. Gas oils used for ethylene production are
crude oil fractions in the boiling range of 180 –3508C (atmospheric gas oils,
AGO) and 350– 6008C (vacuum gas oils, VGO). In contrast to naphtha and
lighter gas feeds, these feedstocks can not be characterized by individual
components.
Gas chromatography coupled with mass spectrometry (GC – MS) or high
performance liquid chromatography (HPLC) allow the analysis of structural
groups, ie, the percentage of paraffins, naphthenes, olefins, monoaromatics,
and polyaromatics in the gas oil, and can be used to determine the quality of
the hydrocarbon fraction. If this information is used together with data such
as hydrogen content, boiling range, refractive index, etc., the quality can be
determined quite accurately. A rough estimate of feed quality can be made by
using the BMCI or the calculated cetane number of a gas oil. The cetane number,
normally used to calculate the performance of diesel fuels, is an excellent quality
measure, since it is very sensitive to the n-paraffin content, which is one of the
key parameters for the ethylene yield. The cetane number CN is calculated as
follows (11):
5. Production
C3 H8 ! a H2 þ b CH4 þ c C2 H4 þ d C3 H6 þ e C4 H8 þ f C5þ
C2 H6 ! C2 H4 þ H2 (1)
If this were the only reaction, the product at 100% conversion would consist
solely of ethylene and hydrogen; at lower conversion, ethylene, hydrogen and
ethane would be present. In fact, the cracked gas also contains methane, acety-
lene, propene, propane, butanes, butenes, benzene, toluene, and heavier compo-
nents. This reaction (Eq. 1) is clearly not the only reaction occurring.
In the 1930s, the free-radical mechanism for the decomposition of hydrocar-
bons was established (20). Although the free-radical treatment does not explain
the complete product distribution, even for a compound as simple as ethane, it
has been extremely useful. Ethane cracking represents the simplest application
of the free-radical mechanism. Ethane is split into two methyl radicals in the
chain initiation step (Eq. 2). The methyl radical reacts with an ethane molecule
to produce an ethyl radical (Eq. 3), which decomposes to ethylene and a hydrogen
atom (Eq. 4). The hydrogen atom reacts with another ethane molecule to give a
molecule of hydrogen and a new ethyl radical (Eq. 5).
Initiation
C2 H6 ! CH 3 þ CH 3 (2)
Propagation
CH 3 þ C2 H6 ! CH4 þ C2 H 5 (3)
C2 H 5 ! C2 H4 þ H (4)
H þ C2 H6 ! H2 þ C2 H 5 (5)
H þ H ! H2 (6)
CH 3 þ H ! CH4 (7)
H þ C2 H 5 ! C2 H6 (8)
C2 H 5 þ CH 3 ! C3 H8 (9)
C2 H 5 þ C2 H 5 ! C4 H10 (10)
This b-scission reaction produces a shorter radical (RCH 2 ) and ethylene. Radi-
cals normally decompose in the b-position, where the C C bond is weaker due
to electronic effects. Large radicals are more stable than smaller ones and can
therefore undergo isomerization.
Radical isomerization
n-C4 H 9 ! C2 H4 þ C2 H 5 (15)
C2 H 5 ! C2 H4 þ H (16)
Reactions like (1) and (15) are highly endothermic. Reported values of DH at
8278C are þ144.53 kJ/mol for Equation (1) and þ232.244 kJ/mol for Equation (15).
The mathematical description of these complex systems requires special
integration algorithms (22). Based on the pseudo steady state approximation,
the chemical reactions can be integrated and the concentration of all components
at each location of the reactor (cracking coil) can be computed (23,24).
In a generalized and very simplified form the complex kinetics of cracking of
hydrocarbons (ethane to gas oil) in steam crackers can be summarized as follows:
The fundamentals of furnace design and the main influences of the different
parameters can be understood even with this simplified mechanism:
– Residence time: From the above scheme it is clear that a long residence time
favors the secondary reactions, whereas a short residence time increases
the yields of the primary products, such as ethylene and propylene.
– Partial pressure: Since most of the secondary products result from reactions
in which the number of molecules decreases, increasing the pressure favors
the secondary products. One function of the steam present in the system is
to reduce the hydrocarbon partial pressure and thus favor the formation of
primary products.
– Temperature and temperature profiles: The oligomerization reactions
involved in the formation of secondary products are favored by lower
temperatures; therefore, special temperature profiles are applied along the
cracking coil to avoid long residence times at low temperatures.
of the heat transferred simply raises the reactant from the inlet temperature to
the necessary reaction temperature. In the middle of the coil, the rate of tem-
perature rise is lower, but cracking rates are appreciable. In this section, the
endothermic reaction absorbs most of the heat transferred to the mixture. At
the coil outlet, the rate of temperature rise again increases but never becomes
as rapid as at the inlet.
The designers of cracking coils try to optimize the temperature and pres-
sure profiles along the radiant coils to maximize the yield of valuable products
yields by special coil design that allows rapid temperature increase in the inlet
section and low pressure drops in the outlet section of the cracking coils.
Typical process gas temperature profiles along the radiant coil of modern
ethylene furnaces are shown in Figure 2 for ethane, propane, butane, and
naphtha cracking.
The quantity of steam used, generally expressed as steam ratio (kilograms
of steam per kilogram of hydrocarbon), varies with feedstock, cracking severity,
and design of the cracking coil. Typical steam ratios used at a coil outlet pressure
of 165–225 kPa (1.65–2.25 bar) for various feedstocks are:
Ethane 0.25–0.35
Propane 0.30–0.40
Naphthas 0.4–0.50
Atmospheric gas oils (cut: 180–3508C) 0.60–0.70
Hydrocracker bottoms (cut: 350–6008C) 0.70–0.85
The sulfur content of the feed is important, since sulfur passivates active Ni
sites of the cracking coil material by forming nickel sulfides, which do not cata-
lyze coke gasification, in contrast to nickel metal and nickel oxides. This explains
why CO production is high (up to 1% in the cracked gas) if the feedstock is free of
sulfur. To prevent this effect, ca. 20 ppm of sulfur (eg, as dimethyl sulfide) are
added to sulfur-free feedstocks.
Feedstock composition is another important consideration in commercial
production of ethylene. Ethylene-plant feedstocks generally contain straight-
and branched-chain alkanes, olefins, naphthenes, and aromatics. Ethylene and
other olefins are formed primarily from alkanes and naphthenes in the feed.
n-Alkanes are the preferred component for high ethylene yields, and those con-
taining an even number of carbon atoms give slightly better ethylene yields than
odd-numbered ones. n-Alkanes also produce propene, whose yield decreases with
increasing chain length.
Isoalkanes, in general, produce much smaller yields of ethylene and pro-
pene than n-alkanes. They give higher yields of hydrogen and methane, and of
C4 and higher olefins. Isoalkanes tend to produce more propene than ethylene
compared to n-alkanes.
Olefin and diolefin yields from cyclopentane, methylcyclopentane, and
cyclohexane have been reported (21); methylcyclopentane and cyclohexane give
substantial amounts of butadiene.
Simple and condensed ring aromatics produce no ethylene. Benzene
is stable under normal cracking conditions. It is formed during cracking
and remains unchanged in the feed. Other aromatics yield primarily higher
molecular mass components. Aromatics with long side chains, such as
dodecylbenzene, yield ethylene and other olefins by side-chain cracking.
The aromatic nucleus still, however, generally produces pyrolysis fuel oil
and tar.
In summary, maximum ethylene production requires:
These conditions maximize the yield of olefins and minimize the yield of
methane and high molecular mass aromatic components.
Cracking conditions and yield structure are often optimized for economic
reasons in plant design and operation; even small changes in cracked gas compo-
sition influence the overall plant economics dramatically.
The dilution steam ratio is in most cases at the minimum of the range for
the specified feedstock, due to the contribution of the dilution steam to the pro-
duction costs.
Heat Requirements for Hydrocarbon Pyrolysis. Primary hydrocarbon
pyrolysis reactions are endothermic; heat requirements are determined by feed-
stock and cracking conditions. The temperature required to attain a given reac-
tion rate decreases as the carbon chain length in the feedstock increases. Thus,
738 ETHYLENE Vol. 2
1. The enthalpy required to heat the feedstock, including the latent heat of
vaporization of liquids
2. The endothermic heat of cracking
3. The enthalpy required to heat the cracked gas from the radiant-coil inlet
temperature to the radiant-coil outlet temperature
The first is accomplished in the convection section of the furnace; the last two, in
the radiant section. Enthalpies (1) and (3) can be calculated from standard heat-
capacity data; they represent 25–60% of total process heat input in the furnace
as the feedstock changes from ethane to gas oils (21).
A typical heat balance for a modern naphtha furnace including the TLE is
shown in Figure 3.
Hydrogen 0.0
Methane 5646.3
Ethane 3548.1
Ethylene 1345.3
Acetylene 8540.7
Propane 2949.9
Propene 17.9
Methylacetylene 4317.1
Propadiene 4496.8
n-Butane 2694.4
Isobutane 2827.5
1-Butene 459.4
cis-2-Butene 649.5
trans-2-Butene 668.2
Isobutylene 746.5
Ethylacetylene 2720.4
1,3-Butadiene 1721.8
Benzene 783.6
Toluene 243.4
Cyclopentane 1611.3
Methylcyclopentane 1722.6
Cyclohexane 1846.4
Methylcyclohexane 1890.3
742
Vol. 2 ETHYLENE 743
The properties of the naphtha feed stock used to generate the yield patterns
given in Tables 5, 6, and 7 are as follows:
The yield increases with decreasing residence time. The maximum benefit
of a selective cracking coil is achieved in high-severity cracking, in which a 5%
feedstock saving can be achieved by using short residence times. However, care-
ful analysis is required since, similar to gas cracking, shorter run lengths are
typical for short-residence-time cracking. Maximum annual furnace productivity
is achieved with a combination of short residence time and long run length.
Gas Oil (AGO; Boiling Range 180–3508C). Compositions and boiling
ranges of gas oils vary greatly with source and refinery conditions. Saturated
gas oils may have BMCI values as low as 20, whereas aromatic gas-oil BMCI
values can exceed 45. Typical yield distributions are shown in Table 8 for an
AGO with properties as listed in Table 9.
Gas oils produce much more C5+ material than the above feedstocks because
they contain large amounts of condensed polynuclear aromatics. These compo-
nents resist cracking and may remain unchanged or condense to higher molecu-
lar mass materials.
times requiring many more individual coils than longer residence times for the
same production capacity. This is due to the shorter lengths of the short-resi-
dence-time coils, which can be as short as 10–16 m per coil. Long-residence-
time coils can have lengths of 60–100 m per coil. The number of coils required
for a given ethylene capacity is determined by the radiant coil surface, which
is in the range of 10–15 m2 per tonne of feedstock for liquid feedstocks.
Production rate for each coil is determined by its length, diameter, and
charge rate, which translates into a certain heat flux on the radiant coil. Values
of 85 kW/m2 for the average heat flux of a coil should be the maximum.
Radiant coils are usually hung in a single plane down the center of the
firebox. They have also been nested in a single plane or placed parallel in a stag-
gered, double-row tube arrangement. Staggered tubes have nonuniform heat-
flux distributions because of the shadowing effect of one coil on its companion.
If maximum capacity and full utilization of the firebox are desired, tube stagger-
ing can be adjusted so that the shadowing effect is only marginally greater than
with a single-row arrangement.
748 ETHYLENE Vol. 2
Heat transfer to the radiant tubes occurs largely by radiation, with only
a small contribution from convection. Firebox temperature is typically
1000–12008C.
The radiant coil of an ethylene furnace is a fired tubular chemical reactor.
Inlet conditions must meet the required temperature, pressure, and flow rate.
Hydrocarbon feed is heated from its convection-section entry temperature to
the temperature needed for entry into the radiant coils. Dilution steam is intro-
duced into the hydrocarbon stream in the convection section. The combined
hydrocarbon and steam stream is then superheated to the target temperature
at the radiant coil inlet (crossover temperature, XOT). Gaseous feeds require
only the energy necessary to heat the feed and steam to the radiant coil inlet tem-
perature. Liquid feeds require additional energy to heat the liquid to its vapor-
ization temperature plus its latent heat of vaporization.
Energy is saved in the convection section by preheating the cracking
charge; by using utility preheating, ie, boiler-feedwater preheating, dilution-
steam superheating, high-pressure steam superheating or by preheating the
combustion air.
A typical arrangement of furnace elements is shown in Figure 6. In this
case, the convection section contains six separate zones. As an example, the
feed could be preheated in the upper zone, the boiler feedwater in the next,
and mixed feed and dilution steam in the third zone. The process steam is super-
heated in the fourth zone, the HP steam is superheated in the fifth zone and the
final feed and dilution steam are preheated in the bottom zone. The number of
zones and their functions can differ according to furnace design. For example,
the upper zone could be used to recover heat for air preheating.
With the trend to higher capacities per individual furnace a new furnace
set-up has been introduced to the industry with great commercial success. In
this type two fireboxes are connected to one convection section. This design
allows liquid feedstock cracking furnaces with capacities of 130 000–150 000 t/a
to be installed. This twin radiant cell concept has been applied more than 60
industrial furnaces since the late 1980s.
The induced-draft fans allow complex convection sections to be designed
with very high gas velocities to give optimum heat transfer in the convection sec-
tion banks. Modern designs include integrated catalyst systems for NOx removal
by SCR (selective catalytic reduction) technology, whereby the NOx present in
the flue gas reacts with ammonia:
4 NOx þ 4 NH3 þ O2 ! 4 N2 þ 6 H2 O
The temperature of the cracked gas leaving the radiant coils can range from 750
to 9008C. Rapid reduction of gas temperature to 500–6508C, depending on the
feedstock, is necessary to avoid losses of valuable products by secondary reac-
tions. This is accomplished by a transfer-line exchanger (TLE) (see Fig. 6),
which cools the cracked gases and recovers much of the heat contained in the
cracked gas as high-pressure steam.
In all modern furnaces the radiant coils exit the firebox at the top; the
TLEs are top-mounted. In older designs coils also leave the bottom of the firebox,
with TLEs mounted vertically along the side of the furnace or horizontally
Vol. 2 ETHYLENE 749
one or more times along the complete coil; two or more individual tubes may be
combined in parallel.
An example of a so-called split radiant coil arrangement is shown in
Figure 7; here parallel small-diameter coils are combined into a larger diameter
outlet coil. This arrangement allows advantageous temperature profiles to be
applied, with a rapid temperature increase in the inlet section, and offers slightly
higher yields than a uniform diameter coil of the same residence time.
Coil lengths and diameters vary widely as can be seen from the following
design characteristics: