Sei sulla pagina 1di 8

Dehydration of Methanol to Dimethyl Ether by

Catalytic Distillation
Weizhu An, Karl T. Chuang* and Alan R. Sanger

Department of Chemical and Materials Engineering, University of Alberta, Edmonton, AB Canada T6G 2G6

D
imethyl ether (DME) is a key intermediate in methanol to
gasoline (MTG) and methanol to olen (MTO) processes (Chang The kinetics of liquid catalytic dehydration of methanol
1983, 2000). DME has several properties that make it attractive over an ion exchange resin (Amberlyst 35) has been
as a fuel or fuel additive for automobile engines (Fleisch et al., 1997; determined for the temperature range 343 to 403
Kikkawa and Aoki, 1998). The physical properties of DME are similar to K using a batch reactor. The experimental data
those of LPG. The weight-based heating value of DME is higher than are described well by an Eley-Rideal type kinetic
that of methanol. The gaseous volume-based heating value is higher expression, for which the surface reaction is the rate-
determining step. A catalytic distillation process for
than that of methane. The cetane number is between 55 and 60, and
methanol dehydration to dimethyl ether (DME) has
is higher than those of diesel fuels. Moreover, DME is non-toxic and been modeled using the experimentally determined
reacts quickly in the atmosphere to form CO2 and water. DME is an kinetic data. The results were incorporated into the
ultra-clean alternative fuel that contains no sulfur or nitrogen (Fleisch rate-controlled reaction mode for RadFrac, a part of
et al., 1997; Adachi et al., 2000; Muller et al., 2000). Presently, DME the commercial simulation program Aspen Plus. It
is also used as feed for the synthesis of chemicals including acetic acid was shown that synthesis of high purity DME can be
and its esters and other chemicals, as a hydrogen source for fuel cells, achieved using a single catalytic distillation column.
as a refrigerant, and as an aerosol propellant in the cosmetic industry Thus there is signicant potential for reduction of
to replace environmentally harmful CFC propellants (Kikkawa and Aoki, overall capital cost for a plant for methanol dehydra-
1998). tion to DME when compared to conventional produc-
tion facilities that involve separate reaction and distil-
The potential use of DME as fuel for local power generation,
lation processes.
transportation and household gas requires the production of DME on a
large scale. Presently, the major process for the production of DME on La cintique de la dshydratation catalytique liquide du
an industrial scale comprises dehydration of methanol using an acidic mthanol sur une rsine changeuse dions (Amberlyst
dehydration catalyst in a xed bed reactor in the temperature range 523 35) a t dtermine pour la gamme de tempratures
to 673 K, followed by rectication of the product stream to recover high de 343-403 K laide dun racteur discontinu.
purity DME (Bercic and Levec, 1993; Spivey, 1991). Les donnes exprimentales sont bien dcrites par
Zeolites, silica/aluminas, alumina, metal phosphates and une expression cintique de type Eley-Rideal, pour
sulfates, and ion exchange resins have been investigated as acidic laquelle la raction de surface est ltape dterminant
catalysts for alcohol dehydration. Among these solid acid catalysts, la vitesse. Un procd de distillation catalytique pour
la dshydratation du mthanol en dimthylther
ion exchange resins have the advantages that they require relative
(DME) a t modlis laide de donnes de cintique
low operation temperatures (303423 K) and show high selectivity dtermines exprimentalement. Les rsultats ont t
to DME (Spivey, 1991). introduits dans le modle de raction contrle par la
Direct synthesis of DME from synthesis gas, with co-production vitesse de RadFrac, qui fait partie du programme de
of methanol, uses a physical combination of a methanol synthesis simulation commercial Aspen Plus. On montre que
catalyst, most commonly Cu, Zn and Al, or Cr catalyst, and a methanol la synthse du DME puret leve peut tre russie
dehydration catalyst, such as -Al2O3 or other acidic catalysts (Haugaard avec une seule colonne de distillation catalytique.
and Voss, 2001; Voss et al., 1999; Lewnard et al., 1990; Armbruster, Il a donc un potentiel signicatif pour la rduction
2000). However, the concentration of DME in the product stream des cots globaux en capitaux pour une usine de
from a single reactor is not sufciently high for use as fuel grade DME, dshydratation de mthanol en DME, comparative-
ment des installations de production convention-
and so further distillation of the liquid product is required to recover a
nelles, qui impliquent des procds de raction et de
concentrated DME stream (Voss et al., 1999). distillation distincts.
High purity DME, or a product mixture with a high DME
concentration, has higher commercial value than mixtures having a Keywords: Dimethyl ether; catalytic distillation;
low concentration of DME. Direct catalytic dehydration of methanol methanol dehydration to DME; ion exchange resin
catalyst; Eley-Rideal kinetics.
* Author to whom correspondence may be addressed. E-mail address:
karlt.chuang@ualberta.ca

948 The Canadian Journal of Chemical Engineering, Volume 82, August 2004
to DME offers the prospect of a straightforward process for Materials
production of DME. The foundations for this approach include A commercial cation ion exchange resin catalyst (DuPont,
the availability of mature industrial processes for production Amberlyst 35 wet; Table 1) was dried and then used for the
of methanol and known catalysts having high selectivity kinetics experiments. Amberlyst series catalysts are divinyl
for dehydration of methanol to DME. However, a separate benzene/styrene copolymers with sulfonic acid functional
distillation apparatus is required to produce high purity DME groups situated on the aromatic rings, which act as acidic
using present processes, which increases the overall capital and catalytic sites for the methanol dehydration reaction. The
operating costs. catalyst was dried under vacuum at 358 to 363 K for 24 hours
Application of catalytic distillation (CD) methodology to to remove adsorbed water so as to determine the dry weight
methanol dehydration offers advantages in operation, costs, before use. Methanol (Aldrich, 99.99%) and o-xylene (Aldrich,
and quality of product when compared with conventional anhydrous, 99%) were used as received.
processes (Nemphos and Hearn, 1997). The principles behind
application of CD are well established, and CD has been applied Gas and Liquid Sample Analysis
successfully to several industrial processes, including production An on-line GC equipped with a Porapak T column and TCD
of methyl acetate and synthesis of fuel ethers (Nemphos and (HP 5896 Series II) was used to analyze the gas phase and
Hearn, 1997; Malone and Doherty, 2000; Taylor and Krishna, liquid samples from reactors. A gas sampling port on the batch
2000). The catalytic reaction and separation processes typically reactor was connected to the GC using a short line heated with
are performed in a single distillation column. The limitations heating tape to avoid condensation of sample vapours.
imposed by thermodynamic equilibrium of the reaction are
overcome by continuous removal of product from the reaction Catalytic Reactive Distillation Simulation
zone. Combining reaction and separation processes within one
Experimentally determined reaction kinetic data were used to
vessel reduces the overall capital cost compared with use of
develop the model for a CD process for methanol dehydration
separate vessels for each process.
to DME (Table 2). In this study RadFrac, the rigorous equilibrium
A CD system for dehydration of methanol to DME described
stage model that is a part of the commercial simulation package
by Nemphos and Hearn (1997) comprises contacting the stream
Aspen Plus (version 10.10) (Aspen Technology Inc., 1999, and
with a xed bed zeolite catalyst. However, ion exchanged resins
Simulation Sciences Inc., 1994), was used to simulate the
are preferred over acidic zeolites as catalysts for the methanol
catalytic distillation process. Vapour-liquid equilibrium was
to DME process because they require relative low operation
reached at every stage of the column. In this simulation, the
temperatures (303423 K) and show high selectivity to DME
operating pressure was specied via the condenser pressure,
(Spivey, 1991). In this paper we will describe the kinetics of
and it was assumed that there was no pressure drop across the
catalytic dehydration of methanol over the ion exchange resin
column (Sneesby et al., 1997).
Amberlyst 35. Amberlyst 35 is preferred over the related resin
The catalyst was predominantly in contact with liquid,
Amberlyst 15 because it has a higher tolerance to elevated
and so the kinetic equation developed from our experimental
temperatures. The inhibition effects of reaction products such as
results was used with a user kinetics subroutine for the program
water on catalytic activity of acidic resin catalysts are alleviated
REAC-DIST to calculate the liquid phase generation rate for
by separating them from the reaction zone in a CD column
each component at each stage in the reaction zone. The
(Aiouache and Goto, 2002a). The kinetic data then are used to
location of the reaction zone within the column was selected
design and to model a CD process for production of DME. We
by determining the stage having a reaction temperature that
have determined the dependence of DME yield and purity on
gives maximum product yield. The composition at the reaction
several key design parameters, including operating pressure,
zone optimally comprised a high concentration of methanol to
reux ratio, location of feed stage, total number of theoretical
ensure a high rate of production of DME.
stages, and catalyst loading. We will show that dehydration
of methanol to DME by catalytic distillation is more efcient
than conventional methanol dehydration processes, and that Results and Discussion
Amberlyst 35 is superior to zeolites as catalyst in the CD process. Kinetics of Liquid Phase Catalytic Dehydration of
Synthesis of high purity DME can be achieved in a single CD Methanol to DME
column, without the need for an additional separation unit, The kinetics of catalytic dehydration of methanol (Eq. 1) has
resulting in signicantly reduced capital cost. been studied extensively using either acidic oxide catalysts or
acidic ionic exchange resins other than Amberlyst 35. However,
Experimental Amberlyst 35 has advantages over alternative catalysts. Ion
exchanged resins have higher activity and required milder
Kinetics
operating conditions than zeolite catalysts (Table 3) (Nemphos
Liquid phase catalytic dehydration of methanol was performed
and Hearn, 1997). Amberlyst 35 is preferred because it has
in a 5 104 m3 stainless steel autoclave batch reactor
better temperature tolerance than Amberlyst 15, for example.
equipped with four-bladed Teon impellers. The stirrer speed
To date, all alcohol dehydration mechanisms have been found
was 800 rpm in all experiments, at which speed the rate of
to follow Langmuir-Hinshelwood or Eley-Rideal kinetics. Both
reaction was not inuenced by external mass transfer. Argon
the Langmuir-Hinshelwood (Gates and Johanson, 1969) and
was used as inert gas to purge the system and to provide an
Eley-Rideal (Kiviranta-Paakkonen et al., 1998) mechanisms have
initial reaction pressure of 0.82 MPa. To determine the effect
been proposed to describe the dehydration of methanol over
of concentration, o-xylene was used as inert solvent. For each
acidic resin catalysts, and included the inhibiting effect of water.
run, 4g catalyst and 120g solution were charged into the
It was unknown which mechanism would apply for reaction over
reactor, which was then pressurized to 0.82 MPa and heated
Amberlyst 35. Consequently, the two possible kinetic models
to a selected temperature in the range 343403 K.

The Canadian Journal of Chemical Engineering, Volume 82, October 2004 949
Table 1. Physiochemical properties of Amberlyst 35.
methanol to DME over Amberlyst 35 (Eq. 1) was determined
using a batch slurry reactor in the temperature range of 343
Ionic form Hydrogen to 403 K and initial reaction pressure of 0.82 MPa. Adsorption
of the more polar components (alcohol and water) is much
Concentration of acid sites $ 5.2 eq/kg stronger than the adsorption of the less polar component
(ethers) (Linnekoski et al., 1997; Zhang and Datta, 1995), due to
Surface area 45,000 m2/kg
the signicant difference in dielectric constants (e.g. methanol:
33; DME: 5) (CRC Handbook of Chemistry and Physics, 2002
Average pore diameter 25 m
2003). Equations (2) and (3) can be rearranged by dividing both
Maximum operating temperature 413 K denominators and numerators by KM. The resulting terms
1 KC
and E E are both relatively very small when compared
Particle size 0.700 0.950 mm KM KM
with the other components of the denominator. Consequently,
Table 2. Optimum input parameters for Aspen Plus simulation of Equations (2) and (3) can be simplied to Equations (4) and
catalytic distillation of methanol to DME, low catalyst loading. (5) respectively.

Feed Stream CD Column 2


ksCM
rDME = 2
Temperature: 298 K Total stages: 30 KW (4)
Pressure: 0.9 MPa Rectication stages: 17 K CW + CM
M
Flow rate: 2.5 mol/s Reaction stages: 820
Feed composition: Stripping stages: 2130
Methanol: 100 mol% Feed stage: 8
DME: 0 Catalyst loading: 9.23 kg/stage 2
ksCM
rDME =
KW (5)
Water: 0 Column pressure: 0.9 MPa CW + CM
KM
Reux ratio: 9 (mol)

Distillate to feed ratio 0.5 To determine which model best describes the kinetics
of dehydration of methanol to DME, a series of experiments
1 of Kpatented was performed to examine the effects of methanol and water
Table 3. Comparison EC E CD conditions with this study.
concentration separately on the initial reaction rate. First, the
KM KM
Nemphos and initial methanol concentration was varied (5.3524.57M) for
anhydrous solutions in an inert solvent, o-xylene, to determine
This Study Hearn, 1997 the dependence of initial surface reaction rate on initial methanol
concentration in the absence of water. Second, the initial water
Catalyst Amberlyst 35 Zeolite
concentration was varied (03.56M) while keeping the initial
Catalyst Zone Temperature (K) 399411 623673 methanol concentration constant, to determine the effect of
water on reaction rate. DME and water were the only products
Column Pressure (MPa) 0.82 4.1 under these reaction conditions.

Effect of Methanol Concentration on Initial


have been evaluated for modeling reaction 1 over this catalyst. Reaction Rate
Each of the Langmuir-Hinshelwood (Eq. 2) and Eley-Rideal The initial rate of formation of DME varied linearly with
(Eq. 3) models is based on the surface reaction as the rate- initial methanol concentration (Fig. 1). Therefore, catalytic
determining step, and incorporates the effect of competitive dehydration of methanol over Amberlyst 35 in the absence of
chemisorption of water and methanol: water was rst order with respect to methanol concentration
over the range of initial methanol concentration. The results
2CH3OH CH3OCH3 + H2O (1) were consistent only with Equation (5). Equation (4) requires
that the initial reaction rate should be zero order with respect to
2 2
ksK MCM methanol in the absence of initial water, which is contradicted
rDME = (2) by the experimental data. Thus the initial rates for the present
(1+ KWCW + KMCM + KECE )2 process in the absence of water can be described by the Eley-
2 Rideal model and not by the Langmuir-Hinshelwood model.
ksK MCM
rDME = (3) The kinetic equation for dehydration of methanol to DME
1+ KW CW + K MCM + K ECE over Amberlyst 35 in the absence of water reduces to:
rDME = ksCM (6)
where ks is the surface reaction rate constant, and KM, KW and
KE, and CM, CW and CE are the adsorption equilibrium constants
and concentration of methanol, water, and DME, respectively. where
The global kinetics of liquid phase catalytic dehydration of

950 The Canadian Journal of Chemical Engineering, Volume 82, October 2004
Figure 1. Effect of initial MeOH concentration on DME formation
rate. Figure 2. Arrhenius plot and ratio of adsorption equilibrium constants
of water and methanol as a function of temperature.

ks = k0 exp( Ea RT ) (7)

Linear regression of the Arrhenius plot for the initial reaction


rates (Fig. 2) gave values for k0 of 4.72 m3/kgcat.s and Ea of
51.7 kJ/mol

Effect of Water Concentration on Initial Reaction Rate


Water inhibits catalytic methanol dehydration to DME over
either acidic oxides or acidic ion exchange resins (Spivey,
1991; Xu et al., 1997). Desorption of water from acidic active
sites was identied as a component of the global reaction
kinetics (Tanabe et al., 1989). Water and methanol compete
for adsorption at catalytic active sites on the surface of acid
catalysts. Hence, there is an inverse relationship of reaction rate
and concentration of water in the reaction mixture.
Figure 3 shows the inhibiting effect of water by comparing
reaction proles for DME formation as a function of reaction time
at 393 K and 0.82MPa for reaction mixtures having different
initial concentrations of water. The non-linear regression curve Figure 3. Effect of initial water concentration on DME formation.
based on the E-R model (Eq. 5) tted the experimental results
very well (Fig. 4), thus supporting attribution of the Eley-Rideal
mechanism for methanol dehydration to DME over Amberlyst KW
35 under the present reaction conditions. The value for the = K exp(Q RT ) (10)
KM
surface reaction rate constant kS calculated from the non-
linear regression equation (Fig. 4) was 4.72 107m3/kgcat.
s at 393 K, which was in very good agreement with the value where K = KW 0 and Q = (HM HW) (Eq. 10). The linear
KM 0
we obtained from the initial rates for the set of experiments
conducted in the absence of water (5.6 107 m3/kgcat. regression of ln(KW/KM) versus 1/T (Fig. 2) gave the temperature
s). The temperature dependence of the ratio of adsorption dependence of the ratio of adsorption equilibrium constants of
equilibrium constants of water and methanol was derived water and methanol (Eq. 11).
from the dependence of adsorption constants, KW and KM on KW 11138
temperatures, van,t Hoff Equations (8) and (9): = exp 25.75 + (11)
KM T
KW = KW 0 exp( RT )
HW R (8)

The ratio of adsorption equilibrium constant of water to


K M = K M 0 exp( RT )
HM R (9) methanol was 13.3 at 393 K, calculated from Equation (8),
compared to 9.95 for the related acidic ionic exchange resins
Amberlyst 15 (Aiouache and Goto, 2002b) and Dowex 50 (Al-

The Canadian Journal of Chemical Engineering, Volume 82, October 2004 951
Figure 4. Effect of water concentration on initial formation rate of
DME at 393 K.

Jarallan et al., 1988) at that temperature. Thus water is more


strongly adsorped at acid sites of Amberlyst 35, and so will
have a stronger inhibiting effect on forward reaction 1.

Mass Transfer Resistance


The stirring speed for the mixture was set sufciently high that
external mass transfer was not a limiting factor.
The resin particle size was not varied in this study.
However, we have shown in a parallel study (Meng et al.,
2003) that varying the particle size of Amberlyst 35 in the
range 0.32 103 to 1.2 103 m had little effect on the Figure 5. Schematics of the CD column for dehydration of methanol
overall kinetics of hydration/dehydration and etherication to DME.
reactions of butene-water-butanol-ethanol-ETBE systems
under similar conditions. It is therefore anticipated that column was determined by comparing the relative volatilities
internal mass transfer resistance will have little impact on the of reactant and product, and the corresponding temperature
kinetics of the present system. within the zone. An optimized set of input specications for
ASPEN PLUS simulation of catalytic distillation of methanol to
Simulation of Catalytic Distillation Process for DME was determined by sequentially varying the parameters
Methanol Dehydration to DME according to the following steps (Table 2).
A catalytic distillation process for kinetics-controlled methanol
dehydration to DME has been modeled. The process for Total Number of Vapour-Liquid Equilibrium Stages
reaction 1 is based on Equations (5), (7) and (11). The CD A set of simulations was performed in which the total number
column used for the simulation (Fig. 5) comprised three of stages was 15, 20 or 30, and the size of the reaction
sections: an upper rectication zone, a central reaction zone, zone was always 13 stages. A second set of simulations was
and a lower stripping zone. The total number of vapour/liquid performed using the same total numbers of stages, in which
(V/L) stages, the number and location of the V/L stages for the corresponding size of the reaction zone was 6, 11 or
the reaction zone, the location for feeding liquid methanol, 13 stages, to accommodate the limitations imposed by the
the column operating pressure, and the reux ratio were each operating reaction temperature range of the catalyst. In each
varied independently. The temperature at each stage was case the feed location was positioned closely above the top
determined by the other parameters. stage of the reaction zone. The purity of both products, DME
For determination of the optimum size of the column, and water, was improved as the total number of stages was
and the optimum size and location reaction zone, the location increased, up to 30 stages, and the reaction zone was 13
for liquid feed (pure methanol) was positioned just above the stages (Fig. 6). There was no signicant further improvement
reaction zone. The reaction zone was located in the middle from use of a larger CD column or reaction zone. The purity
section of the CD column, where the concentration of reactant of the DME that was recoverable from the top condenser
had a maximum value due to the difference of volatilities was 98.4 mol% when using a 15-stage CD column with a 6-
between methanol and DME, and the corresponding column stage reaction zone (condenser, stage 1; rectication, stages
temperature range was suitable for the highest achievable 24; reaction zone, stages 510; stripping, stages 1114; and
reaction rates. The positioning of the reaction zone within the reboiler: stage 15). The purity of the DME and water products

952 The Canadian Journal of Chemical Engineering, Volume 82, October 2004
Figure 6. Dependence of DME purity on number of stages.

each was increased to almost 100 mol% when using a 30- Figure 7. Dependence of DME purity on location of methanol feed
stage CD column having a 13-stage reaction zone (condenser, stage.
stage 1; rectication, stages 27; reaction zone, stages 820;
stripping stages 2129; and reboiler, stage 30). Thus, it is
necessary to use the larger column to obtain substantially pure
DME as product.
Feed Stage
The optimum location for the feed was determined by varying
that location between stages 2 and 21 within the 13 stage
reaction zone (stages 820) (Fig. 7) . The concentration of
methanol in the reaction zone, and therefore the reaction rate
in the reaction zone, was maximized by feeding methanol
closely above the reaction stage, between stages 5 and 8.
When methanol was fed at the top of the reaction zone (stage
8), the column temperature was high enough to ensure a
high reaction rate and methanol was consumed to produce a
substantially pure product stream. In contrast, when methanol
was injected at a higher stage, some methanol was entrained in
the product stream and the purity of product DME decreased.
When methanol was injected at a lower stage, either within
or below the reaction zone, conversion of methanol was
compromised, and the purity of the DME product was again
Figure 8. Relationship between column pressure and column
reduced. Therefore the optimum position for methanol feed
temperature.
was at the top of the reaction zone.

Reaction Zone the range 0.9 to 1.0 MPa provided the optimum combination
The reaction zone comprised an alternating sequence of of reaction rate and high purity of DME product.
catalyst beds and vapour-liquid equilibrium stages. The catalyst
bed at each reaction zone effectively acted as a small liquid- Reux Ratio
phase reactor. The packed bed was designed to ensure that Reux ratio inuences both reaction rate and separation
the liquid reaction mixture made good contact with the performance in CD columns (Subawalla and Fair, 1999). In
catalyst in each small reactor. Based on our kinetic data, the the present model, a higher reux ratio increased the reactant
total amount of catalyst required in the reactors is 120 kg for methanol recycle across the reaction zone, thereby increasing the
a production scale of 1500 tons DME/year. The amount of local methanol concentration and so increased the driving force
catalyst was an optimum value determined by considering for the reaction (Fig. 10). The residence time of methanol in the
catalyst requirement and reux ratio. reaction zone, and hence the required catalyst loading, was also
dependent on the reux ratio. It was found that high purity of
Column Pressure DME (99.997% as top liquid) can be achieved at a reux ratio
Operating pressure in a CD column strongly affects reaction of 7 for catalyst loading of 150.3 kg, compared with 99.97%
rates by changing the column temperature prole (Taylor and purity at the higher reux ratio of 9 for the corresponding
Krishna, 2000). The dependence of reaction zone temperature catalyst loading of 120 kg. Consequently, the selection of reux
on column pressure in the range 0.7 to1.0 MPa is shown in ratio, and corresponding catalyst amount, to be used in practical
Figure 8, and the resulting dependence of purity of product applications of the CD process for methanol dehydration to DME
DME on pressure is shown in Figure 9. An operating pressure in will be dictated by the compromise between required purity of

The Canadian Journal of Chemical Engineering, Volume 82, October 2004 953
Figure 12. Proles of vapour composition and temperature within
the column.
water or methanol, and was collected as the overhead product.
Figure 9. Relationship of DME product purity to column pressure.
The least volatile component, water, was collected at the
bottom. Methanol, the intermediate component in terms of
boiling point, was retained inside the column and converted
to DME and water. Methanol dehydration to DME is a mildly
endothermic reaction. The DME concentration in the liquid
phase within the reaction zone was very low due to its high
volatility. Removal of DME as vapour from the reaction zone
overcame equilibrium limitations of reaction 1, and drove the
reaction to the right. The concentrations of methanol and water
in the rectication zone were each very low. Consequently,
substantially pure DME can be obtained using the present CD
process. Compared to the patented CD process based on zeolite
catalyst (Table 3) (Nemphos and Hearn, 1997), the advantages
found for the current CD process for DME production were
signicantly lower reaction zone temperatures and pressures.
The simulation also showed advantages to using a single CD
column and process to produce DME, when compared with
the conventional two-step approach having separate reaction
and distillation units, including reduced capital costs and
Figure 10. Dependence of DME product purity on reux ratio.
capability to produce very pure DME.

Conclusions
Kinetics of dehydration of methanol to DME over Amberlyst 35
as catalyst has been modeled using a Eley-Rideal mechanism
in which the surface reaction is the rate-determining step. The
reaction under anhydrous conditions is rst order in methanol.
Water inhibits the reaction. The activation energy over this
catalyst is 51.7 kJ/mol.
Acidic ion exchanged resin catalysts have higher activity than
zeolite catalysts, and operate under milder conditions. Amberlyst
35 is preferred as it has a higher temperature tolerance than
alternative resin catalysts, including Amberlyst 15.
A process using a catalytic distillation process for DME
production has been simulated, based on the experimental
kinetic data. High purity DME can be produced using a 30-
stage catalytic distillation column having a 13-stage reaction
Figure 11. Proles of liquid composition and temperature within the zone, operating in the pressure range 0.9 to 1.0 MPa. Use of a
column. single CD column, instead of the separate reactor, distillation
product and economic considerations. and/or rectication units used in a conventional methanol to
Figures 11 and 12 show proles for the compositions of DME process, offers economic advantages for production of
the liquid and vapour phases respectively within the CD column very pure DME.
when operated at the optimal conditions as determined from
the simulation (Table 2). DME is more volatile than either

954 The Canadian Journal of Chemical Engineering, Volume 82, October 2004
References Haugaard, J. and B. Voss, Process for the Synthesis of a Methanol/
Dimethyl Ether Mixture from Synthesis Gas, U.S. Patent
Adachi, Y., M. Komoto, I. Watanabe, Y. Ohno and K. Fujimoto,
6,191,175, (2001).
Effective Utilization of Remote Coal Through Dimethyl Ether
Kikkawa, Y., Aoki, I., Dimethyl Ether Fuel Proposed as an Alternative
Synthesis, Fuel, 79, 229234 (2000).
to LNG, Oil & Gas Journal, April 6, 5559 (1998).
Aiouache, F. and S. Goto, Rate Acceleration of 2-Methyl-1-Butanol
Kiviranta-Paakkonen, P.K., L.K. Struckmann, J.A. Linnekoski and
Dehydration in a Reactive Distillation Column, J. Chem. Eng.
A.O.I. Krause, Dehydration of the Alcohol in the Etherication of
Japan 35, 443449 (2002a).
Isoamylenes with Methanol and Ethanol, Ind. Eng. Chem. Res.
Aiouache, F. and S. Goto, Kinetic Study on 2-Methyl-1-Butanol
37, 1824 (1998).
Dehydration Catalysed by Ion Exchange Resin, J. Chem. Eng.
Lewnard, J.J., T.H. Hsiung, J.F. White and D.M. Brown, Single-Step
Japan 35, 436442 (2002b).
Synthesis of Dimethyl Ether in a Slurry Reactor, Chem. Eng. Sci.
Al-Jarallan, A.M., M.A. Siddiqui, B. and A.K.K. Lee, Kinetics of Methyl
45, 27352741 (1990).
Tertiary Butyl Ether Synthesis Catalyzed by Ion Exchange Resin,
Linnekoski, J.A., A.O. Krause and L.K., Rihko, Kinetics of the
Can. J. Chem. Eng. 66, 802807 (1988).
Heterogeneously Catalyzed Formation of tert-Amyl Ethyl Ether,
Armbruster, H., Catalytic Dehydration of Methanol (MeOH) to
Ind. Eng. Chem. Res. 36, 310316 (1997).
Dimethyl Ether (DME) as an Alternative Fuel for Diesel Engines,
Malone, M.F. and M.F Doherty, Reactive Distillation, Ind. Eng.
Report, Laboratory for Energy and Materials Cycles, Paul Scherrer
Chem. Res. 39, 39533957 (2000).
Institute, Department of General Energy Research, Switzerland,
Meng, N., W. An, K.T. Chuang and A.R. Sanger, unpublished results
April (2000).
(2003).
Aspen Technology Inc., Aspen Plus, Version 10.10, Cambridge; MA,
Muller, J.T., P.M. Urban, W.F. Holderich, K.M. Colbow, J. Zhang and
(1999).
D.P. Wilkinson, Electro-Oxidation of Dimethyl Ether in a Polymer-
Bercic, G. and J. Levec, Catalytic Dehydration of Methanol to
Electrolyte-Membrane Fuel Cell, J. Electrochem. Soc. 147, 4058
Dimethyl Ether. Kinetic Investigation and Reactor Simulation, Ind.
4060 (2000).
Eng. Chem. Res. 32, 24782484 (1993).
Nemphos, S.P. and D. Hearn, Method for the Preparation of Dialkyl
Chang, C.D., The Methanol-to-Hydrocarbons Reaction: A
Ethers, US Patent 5,684,213 (1997).
Mechanistic Perspective, ACS Symposium Series, 738, 96114
Simulation Sciences, Inc., PROvision, Simulation Sciences, Inc, CA,
(2000).
(1994).
Chang, C.D., Hydrocarbons from Methanol, Catal. Rev.-Sci. Eng.
Sneesby, M.G., M.O. Tade, R. Datta and T.N. Smith, ETBE Synthesis
25(1), 1118 (1983).
via Reactive Distillation. 1. Steady-State Simulation and Design
CRC Handbook of Chemistry and Physics, 83rd ed.; CRC Press Inc.,
Aspects, Ind. Eng. Chem. Res. 36, 18551869 (1997).
Boca Taton, FL (20022003), pp. 945 and 6153.
Spivey, J.J., Review: Dehydration Catalysts for the Methanol/
Fleisch, T.H., A. Basu, M.J. Gradassi, J.G. Masin, Dimethyl Ether: A
Dimethyl Ether Reaction, Chem. Eng. Comm. 110, 123142
Fuel for the 21st Century, Stud. Surf. Sci. Catal. 107, 117125
(1991).
(1997).
Subawalla, H. and J.R. Fair, Design Guidelines for Solid-Catalyzed
Gates, B.C. and L.N. Johanson, The Dehydration of Methanol and
Reactive Distillation Systems, Ind. Eng. Chem. Res. 38, 3696
Ethanol Catalyzed by Polystyrene Sulfonate Resins, J. Catal. 14,
3709 (1999).
6976 (1969).
Tanabe, K., M. Misono, Y. Ono and H. Hattori, New Solid Acids and
Bases, Their Catalytic Properties, Stud. Surf. Sci. Catal. 51, 175
(1989).
Taylor, R. and R. Krishna, Modelling Reactive Distillation, Chem.
Eng. Sci. 55, 51835229 (2000).
Voss, B., F. Joensen and J.B. Hansen, Preparation of Fuel Grade
Dimethyl Ether, U.S. Patent 5,908,963, (1999).
Xu, M., J.H. Lunsford, D.W. Goodman and A. Bhatacharyya,
Synthesis of Dimethly Ether (DME) from Methanol over Solid-Acid
Catalysts, Appl. Catal. A: General. 149, 289301 (1997).
Zhang, T. and R. Datta, Ethers from Ethanol. 4. Kinetics of the
Liquid-Phase Synthesis of Two tert-Hexyl Ethyl Ethers, Ind. Eng.
Chem. Res. 34, 22472257 (1995).

Manuscript received July 22, 2003; revised manuscript received March


11, 2004; accepted for publication April 23, 2004.

The Canadian Journal of Chemical Engineering, Volume 82, October 2004 955

Potrebbero piacerti anche