Sei sulla pagina 1di 2

Fresenius J Anal Chem (2000) 366 : 438443 Springer-Verlag 2000

S P E C I A L I S S U E PA P E R

N. S. Bloom

Analysis and stability of mercury speciation in petroleum hydrocarbons

Received: 4 June 1999 / Revised: 18 October 1999 / Accepted: 22 October 1999

Abstract Raw petroleum and natural gas often contain gies for the removal of Hg from both crude oils and sell-
high concentrations of mercury, which can be damaging able products. The most dramatic impact of Hg for the pe-
to the metal components of production facilities, as well troleum industry relates to condenser failures in natural
as to the environment. Various Hg species have different gas plants. Costly damage from long-term exposure of
properties in terms of mobility, reactivity and bioavail- metal components especially cryogenic condenser tubes
ability. Thus, for cost-effective decisions regarding plant to condensed elemental mercury (Hgo), results when it
design, Hg extraction, and pollution control, speciation amalgamates with and weakens those components [2].
information must be available at the production facility. In Volatilized Hg has also been responsible for damaging
this paper, a simple, wet chemical speciation method, valuable refinery catalysts, and for contaminating both
which provides data on Hgo, dissolved and particulate to- plant equipment and products. As the marketplace be-
tal Hg, Hg(II), and methyl Hg is presented. The method comes more aware of the negative environmental conse-
incorporates species-specific extraction and separation quences of Hg [3], the inability to minimize the Hg con-
procedures, followed by cold vapor atomic fluorescence tent of commercial and consumer products diminishes
spectrometry (CVAFS). For each species, detection limits their relative value.
of approximately 0.1 ng/g were obtained. Storage experi- To date, little research has been published on the deter-
ments in various containers showed that organo-mercury mination of total mercury (THg) in petroleum products,
species were stable for at least 30 days in all containers and even less on its chemical speciation. THg is routinely
except those made of polyethylene; and Hgo was stable in determined by the oxygen bomb combustion plus cold va-
all containers except those made of stainless steel or poly- por atomic absorption spectroscopy (CVAAS) technique
ethylene. Hg(II) was rapidly lost from all containers ex- [4], although the limits of detection and precision are gen-
cept those made of aluminum, which rapidly converted it erally too poor to use as anything except a screening tool
to Hgo, which was stable. In general, most of the total Hg [5]. Neutron activation analysis [6] can provide superla-
in petroleum products was particulate Hg, followed by tive detection limits and accuracy, but few laboratories or
dissolved Hg(II) and Hgo. Sub-ng/g concentrations of refineries have access to this sophisticated methodology.
methyl-Hg were observed in most samples. Recently, several new methods have been reported, in-
cluding perchloric acid bomb digestion with cold vapor
atomic fluorescence (CVAFS) detection [7], and volatiliza-
Introduction tion directly onto gold traps, followed by microwave
plasma emission (MPE), or CVAAS detection [810]. Ex-
Mercury is found in a variety of chemical and physical cellent limits of detection using these techniques have
forms (species). Each has different characteristics in terms been reported, but they are too specific, or even dangerous
of solubility, volatility, toxicity, and reactivity [1]. Under- (perchloric acid) for common use. In the few cases where
standing the chemical speciation of Hg in petroleum and Hg speciation has been investigated, the emphasis has
refinery products is critical to predicting potential adverse been on quantifying the organo-mercurial components of
effects in the refinery itself, and to optimizing technolo- the oil; generally using derivatizaion, followed by gas
chromatography (GC) interfaced with a sensitive, ele-
ment-specific detector such as MIP or CVAFS [11, 8, 12].
These methods have shown the presence of small quanti-
ties of various dialkyl Hg compounds, as well as Hgo and
N. S. Bloom
Frontier Geosciences Inc., 414 Pontius N., alkylatable Hg(II), but are not capable of resolving the
Seattle, WA 98109, USA concentration of other inorganic forms, such as organo-
439

complexed Hg(II), or colloidal HgS, which may account Table 1 Measured mercury concentrations in solvents saturated
for the majority of Hg found in many petroleum products. with various pure Hg compounds
This paper presents the development of a sequence of Solvent C Saturation [Hg], ng/g (ppb)
relatively simple wet chemical extractions, which, when
combined with a robust and sensitive detector such as Hgo HgCl2 CH3HgCl
CVAFS, allows accurate quantification of THg, dissolved/ Hexane 27 1,230 11,600
particulate Hg, Hg(II), Hgo, and monomethyl Hg (CH3Hg) Octane 27 1,150
at detection limits of less than 1 ng/g. Results of species- Dodecane 24 1,590 15,800 1,080,000
specific storage experiments, and a selection of complete Paraffin oil 22 1,470
field data are also presented. Methylene chloride 27 4,200 2,540,000
Freon 27 450

Analytical methods
General comments urated solution were made and calibrated for THg on an as-needed
basis. Table 1 shows the saturation concentrations obtained in this
Samples were processed using ultra-clean sample handling [13], process. Dimethyl mercury ((CH3)2Hg) is miscible in most hydro-
including storage in a laboratory environment where the atmo- carbons.
sphere was kept at less than 25 ng/m3 total gaseous Hg. In addi-
tion, samples were stored and extracted in glass containers with Extraction vials. Extractions were carried out in disposable 40 mL
Teflon-lined caps, to minimize exchange of volatile Hg with the glass vials with Teflon-lined caps. Dilutions filled the vials up to
container walls or surrounding atmosphere. Hydrocarbon samples the base of the neck. This procedure is somewhat less accurate
with initial boiling points of < 40 C were refrigerated prior to fil- ( 0.5 mL, or 12%) than using volumetric glassware, but it speeds
tration and extraction. Samples with initial boiling points between and economizes processing large numbers of samples, while con-
40 C and 100 C were processed at room temperature, while those tributing minimally to the overall method analytical variability of
with boiling points above 100 C were heated to 40 C before pro- 510%.
cessing. Very viscous bunker oils and asphalts were first diluted by
a factor of 10 with hexane.
All of the extractions were ultimately analyzed using cold va- Extraction procedures
por atomic fluorescence spectrometry (CVAFS) in the detection
step. With all analytical data sets, the following quality assurance Total Hg. Into a 40 mL vial filled with 0.07 M BrCl solution,
samples (at minimum) were collected: matrix spike, matrix spike 0.51.5 g of oil were weighed to the nearest milligram. For sam-
duplicate, sample duplicate, and three method blanks. Final results ples that contained high concentrations of sulfur or aromatic com-
were corrected using the mean method blank obtained from the pounds, smaller aliquots (0.10.3 g) were digested, to avoid con-
blanks co-processed with the particular sample set. suming the oxidizing power of the BrCl. Vials were capped
tightly, and placed in an oven maintained at 80 C ( 5 C). To
avoid over-pressurization, samples with boiling points less than
Reagents and materials 100 C were extracted at room temperature. The samples were
vigorously shaken for 2 min intervals, at least 4 times over the
Common reagents. Laboratory reagents, acids, solvents, gases, and following 1218 h extraction period. After digestion, the samples
deionized water were reagent-grade or better, and pre-analyzed for were allowed to settle, or were centrifuged for 15 min at 3000
Hg. RPM to separate organic and aqueous layers. If the aqueous phase
changed from orange to colorless or pale yellow indicating that
BrCl extracting solution. 0.2 M BrCl solution [14] was prepared all of the oxidizing power of the BrCl was consumed a smaller
by adding 27 g of potassium bromide (KBr) to a 2.5 L bottle of sample aliquot was re-digested. Before analyzing, the oil layer was
HCl, and stirring vigorously. Next, 38 g of potassium bromate carefully pipetted off and discarded. THg was determined on the
(KBrO3) was slowly added to the acid mixture as it was rapidly aqueous fraction by analyzing a 0.015.00 mL aliquot using SnCl2
stirred in a fume hood. For petroleum extractions, a 0.07 M solu- reduction, purge-and-trap on gold, and CVAFS detection [7, 15].
tion of BrCl was prepared by 2 : 1 dilution of the 0.2 M solution
with deionized water. Dissolved Hg. Except when otherwise stated, all dissolved results
were obtained by vacuum filtering the sample through a 47 mm
Saturated KCl extracting solution. 400 g of KCl were placed into diameter 0.8 nitrocellulose membrane filter (Nalge disposable
a 1 L bottle, and filled with deionized water plus 5 mL of concen- 125 mL sterilization units). In a few cases, multiple filtrations were
trated HCl. To the solution, 0.1 g of NH2OH.HCl (hydroxylamine performed to obtain sufficient volume, and in other cases, pre-di-
hydrochloride) was added to destroy free chlorine, which may be lution of the sample with hexane was required to successfully ob-
present in the reagents or in the water. The solution was shaken pe- tain a filtration. The filtrate was transferred to 20 mL glass vials
riodically until no more KCl dissolved, and then was allowed to with Teflon-lined caps for storage. Aliquots of the filtered samples
settle. were analyzed for THg exactly as described above, and the value
was reported as total dissolved Hg. Because up to 30% of Hgo may
Oxidant-free hexane. 900 mL of reagent-grade hexane was shaken be lost during vacuum filtration, in cases where volatile Hg con-
for 5 min with 100 mL of deionized water containing 0.1 g of centrations were a significant fraction of the total (i.e., > 5%),
NH2OH.HCl, and the phases were then allowed to separate. a correction for the loss was performed. This correction was made
by measuring the Hgo in both the unfiltered and filtered samples,
Pure Hg species in hydrocarbons. Solutions of Hg species in vari- then adding the difference between these back to the measured dis-
ous hydrocarbons were first prepared as saturated solutions by solved THg concentration.
shaking several grams of each compound (CH3HgCl, HgCl2, Hgo)
with 100 mL of liquid hydrocarbon for several days. The solutions Ionic and weakly complexed Hg(II). Into a 40-mL vial containing
were allowed to settle, and then dilutions were made with paraffin 10.0 mL of saturated KCl solution, 10.0 mL of the homogenized,
oil or other solvent. The concentrations of the inorganic species in filtered petroleum sample was accurately weighed. The Hg(II) ex-
hydrocarbons are not stable over time, so fresh dilutions of the sat- traction was performed only on filtered samples, to avoid unre-

Potrebbero piacerti anche