chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Research and Design

journal homepage: www.elsevier.com/locate/cherd

Catalytic and kinetic study of methanol dehydration to

dimethyl ether

S. Hosseininejad, A. Afacan, R.E. Hayes ∗

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

a b s t r a c t

Dimethyl ether (DME), as a solution to environmental pollution and diminishing energy supplies, can be synthesized
more efficiently, compared to conventional methods, using a catalytic distillation column for methanol dehydration
to DME over an active and selective catalyst. In this work, using an autoclave batch reactor, a variety of commercial
catalysts are investigated to find a proper catalyst for this reaction at moderate temperature and pressure (110–135 ◦ C
and 900 kPa). Among the ␥-alumina, zeolites (HY, HZSM-5 and HM) and ion exchange resins (Amberlyst 15, Amberlyst
35, Amberlyst 36 and Amberlyst 70), Amberlysts 35 and 36 demonstrate good activity for the studied reaction at
the desired temperature and pressure. Then, the kinetics of the reaction over Amberlyst 35 is determined. The
experimental data are described well by Langmuir–Hinshelwood kinetic expression, for which the surface reaction
is the rate determining step. The calculated apparent activation energy for this study is 98 kJ/mol.
© 2011 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Keywords: Dimethyl ether; Kinetics; Rate equation; Zeolites; Amberlyst

1. Introduction In the past decade, the use of dimethyl ether (DME) as an

Two major challenges faced today are climate change induced
by global warming and the threat of the diminution of cheap
and readily available energy supplies, especially liquid trans-
portation fuels. Both of these challenges are driving the search
towards alternative fuels. With a more varied fuel mix, atten-
tion must be paid to the reduction of emissions, especially
particulate matter (PM) and NOX . Although these emissions
are not an urgent issue for stoichiometric gasoline fuelled
engines equipped with three way catalytic converters, with
lean burn compression ignition direct injection (CIDI) engines
that use diesel fuel, both of these emissions are significant
problems. As Europe and North America move to a diesel
based economy as an aid to improving fuel economy and
hence a reduction in greenhouse gas (GHG) emissions, these
issues are rising to the top of the priority list. Current regula-
tions have led to a major review of engine design with strict
forthcoming limits on emissions. Reformulated diesel fuel will
likely play a key role in this future. Reformulation includes,
for example, bio-diesel, reduced sulphur content, adding oxy-
genates and using alternate fuels.
alternative fuel has been researched, and many car companies
have been developing DME engines and related technology.
DME has a cetane number and ignition temperature close
to that of diesel fuel and gives low NOX , low smoke and
low engine noise, compared to conventional diesel engines
(Semelsberger et al., 2006). It ranks near the top in well-to-
wheel (WTW) efficiency among alternative fuels, regardless
of vehicle technology, and when coupled to hybrid engines
can have higher WTW efficiency compared to fuel cells with
similar overall levels of GHG emissions.
DME can also be used as a chemical feedstock to make
many products, such as short olefins (ethylene and propylene),
gasoline, hydrogen, acetic acid and dimethyl sulfate. DME can
be easily transported to the areas far from oil and gas sources.
DME can be produced by the dehydration of methanol,
which in turn is made from synthesis gas. Synthesis gas (syn-
gas) can be produced from, for example, natural gas, coal
and biomass. Traditionally, DME has been produced from syn-
gas in a two step process, in which methanol is produced
from syngas, purified, and then converted to DME in another
reactor.

∗
Corresponding author.
E-mail address: bob.hayes@ualberta.ca (R.E. Hayes).
Received 10 March 2011; Received in revised form 26 July 2011; Accepted 8 October 2011
0263-8762/$ – see front matter © 2011 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
doi:10.1016/j.cherd.2011.10.007
826 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833

Nomenclature

C concentration (mol m−3 )

E activation energy (J mol−1 )
−H enthalpy change of reaction (J mol−1 )
kS rate constant
K adsorption equilibrium constant
Q enthalpy of adsorption
R gas constant (J mol−1 K−1 )
(rDME ) rate of formation of dimethyl ether
(mol m−3 s−1 )
T temperature (K)

Subscripts
D dimethyl ether
M methanol
W water
0 initial value

In the conventional method, DME synthesized in a fixed-

bed reactor is purified using at least two distillation columns.
To reduce both capital and operating costs, and to increase
energy efficiency, process integration can be considered. Cat-
alytic distillation (CD) is an integrated process where the
reactor and distillation column are combined into a single
unit. The advantages of using CD for methanol dehydration
include a higher selectivity of products to DME synthesis,
higher conversion compared to a single reactor and lower
operational cost. However, the CD requires operation at mod-
erate temperature and pressure (40–180 ◦ C and 800–1200 kPa).
Most of the catalysts previously studied for this reaction are
solid-acid catalysts (e.g. zeolites) which tend to be active at
high temperature (250 ◦ C), and less research has been done at
the milder conditions required for CD.
DME is produced by the conventional bimolecular catalytic
dehydration of methanol using solid acids (Spivey, 1991). Both
Brønsted and Lewis acid sites can catalyze the methanol DME
reaction. Silica alumina, ␥-alumina and different kinds of zeo-
lites, namely, Mordenite, ZSM-5 and Y show good methanol
conversion and selectivity to DME at high temperature and
pressure. Ion exchange resins have shown activity at lower
temperature, and cannot be used at high temperature. A good
catalyst for methanol dehydration reaction should work at
as low a temperature as possible to avoid subsequent dehy-
dration of DME to olefins or hydrocarbons. Because catalytic
distillation of DME takes place at relatively low pressure
(800–1200 kPa) and temperatures in the range of 50–180 ◦ C (Di
Stanislao et al., 2007), it is not clear from the literature as to
the best choice of catalyst. The final choice will have a combi-
nation of strongest acidic strength and the highest number of
active sites and resistance to water inhibition and side product
formation. Although the acidity of the catalyst plays a cru-
cial role in its performance, other factors such as thermal and
mechanical stability, pore size and distribution as well as cost
will determine the final choice.
The purpose of the present study was to find a suit-
able commercial catalyst for methanol dehydration to DME
reaction at moderate temperature (110–135 ◦ C) and pres-
sure (900 kPa). The activity of commercial solid-acid catalysts
including ␥-alumina, HY, HZSM-5, HM zeolites and ion
exchange resins (Amberlyst 15, Amberlyst 35, Amberlyst 36,
Amberlyst 70) were investigated using an autoclave batch
Fig. 1 – Diagram of the batch reactor used in this study.
reactor. Reaction kinetics over the selected commercial cata-
lyst was then measured to determine a reaction kinetic model.
2. Experimental
The reactions were carried out in a 480 cm3 stainless steel
batch autoclave equipped with a variable speed stirrer (four
blade glass impeller) and a heating jacket. A diagram of the
reactor is given in Fig. 1. Reactor temperature was controlled
using a Parr 4841 proportional controller. The temperature was
measured with a J-type thermocouple. The liquid was sam-
pled through a 1.6 mm diameter tube fitted with a sintered
316 stainless steel filter with pore size of 300 mesh to prevent
it from being plugged by catalyst. A Swagelok needle valve
was used to control the liquid flow rate. Vapour samples were
collected through the gas vent, which was also fitted with a
Swagelok needle valve. A condenser was connected to the gas
vent to prevent methanol and water vapour escaping through
the gas vent during gas sampling.
The liquid samples were analyzed using a Hewlett–Packard
5710A series gas chromatograph (GC) equipped with a thermal
conductivity detector (TCD) and a 3 m long, 1.6 mm diameter
Supelco Co. stainless steel HayeSep D column with mesh size
80/100. The carrier gas was UHP helium with the flow rate of
35 cm3 /min. The detector and injection port temperatures are
set to 200 ◦ C and 250 ◦ C, respectively. The oven temperature
was 165 ◦ C.
A Hewlett–Packard 5970 series GC/MS equipped with a
DB-5MS capillary column with 30 m in length and 0.25 mm
in diameter was used to determine if the liquid samples
contained any product other than that DME. This GC was
able to detect product with molecular weight of 15–550. The
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833
Table 1 – Properties of three different zeolite catalysts.
Zeolyst product Zeolite
CBV 28014 ZSM-5
CBV 8014 ZSM-5
CBV 21 Mordenite
injector and detector temperatures were set to 280 ◦ C. The
oven temperature was kept constant at 35 ◦ C for 5 min and
then increased to 280 ◦ C at a rate of 10 ◦ C/min. One microlitre
of liquid sample was injected with a split ratio is 100:1 and
gas sample injection is splitless 10 ␮L of sample. The result
of the GC–MS analysis showed that there was no detectable
by-product produced at this operating condition.
The three liquids used as starting materials were methanol,
water and tetrahydrofuran, the latter being used as a diluent.
Research grade (99.9%) methanol was obtained from Fisher
Scientific. The water was obtained from a reverse osmosis
system. Tetrahydrofuran (THF) was obtained from Fisher Sci-
entific.
The catalysts were ␥-alumina, zeolites (Y, ZSM-5 and Mor-
denite) and ion exchange resins (Amberlyst 15, Amberlyst 35,
Amberlyst 36, Amberlyst 70 and Amberlite IR-120).
2.1. Zeolites
ZSM-5 and Mordenite, namely, CBV21, CBV8014 and CBV28014
were obtained from the Zeolyst International Company
(USA). These catalysts have different SiO2 /Al2 O3 ratios, which
indicate different acidity strengths. Increase in SiO2 /Al2 O3
decreases the acidity strength but the amount of acidity
remains almost the same (Khandan et al., 2008). Table 1
shows the zeolites’ properties, as provided by the Zeolyst Com-
pany. Zeolites were received in NH4 + form, which is inactive
for methanol dehydration. They were calcined in a Ther-
molyne 79400 tube furnace to convert the ammonium cations
to hydrogen by removing the ammonia. The catalysts were
heated at 6 ◦ C/min from 25 to 500 ◦ C and then held at 500 ◦ C
for 4 h. After removal from the furnace, the calcined zeolite
were moved to a vacuum chamber to prevent adsorption of
water from the air.
2.2. Amberlyst
Amberlyst 15, 35, 36 and 70, and Amberlite IR-120 were
obtained from the Rohm and Haas Company (USA). They were
received in wet form, and were dried prior to use using a vac-
uum dryer. Amberlyst series catalysts’ properties are shown
in Table 2, which were provided by the company. The acid-
ity of the catalyst was measured according to the procedure
suggested by Rohm and Haas Co. We used one tenth of the
Table 2 – Amberlyst series catalysts properties.
Name Amberlyst 15
Acidity (equiv./kg)a 4.6
Surface area (m2 /g) 53
Average pore diameter (A) 300
Mean size (mm) 0.6–0.85
Pore volume (ml/g) 0.4
Swelling (water to dry) 37
Operating temperature limit 120
a
Calculated by Rohm and Haas Co. procedure.
827
SiO2 /Al2 O3 (mole ratio) Surface area (m2 /g)
280 400
80 425
20 500
values suggested by the procedure. For instance, 1.5 g of each
Amberlyst is ion exchanged with 100 cm3 of sodium nitrate
and 100 cm3 of HCl in regeneration. The procedure includes
passing sodium nitrate through the catalyst bed where the
cation exchange happens. After exchanging the hydrogen ions
in catalyst with sodium, the catalysts were washed and regen-
erated by HCl to ion exchange the sodium ions by hydrogen.
The regenerated catalyst is again ion exchanged by sodium
nitrate, and exactly 100 cm3 of solution is collected, and
titrated by standard NaOH solution.
3. Catalyst screening tests
A set of experiments was conducted using all of the com-
mercial solid-acid catalysts. The reaction was conducted at
110 ± 1 ◦ C. For each run, 4 ± 0.005 g catalyst and 120 ± 0.2 g
solution were charged into the reactor. The reactor was run
for 3.5 h, and the DME synthesis rate and methanol conversion
of the reaction were compared. It was found that ␥-alumina,
zeolites (Y) and Amberlite IR-120 did not have any detectible
conversion of methanol. ZSM-5 and Mordenite on the other
hand had small but quantifiable conversions of less than 3%
methanol conversion at temperatures up to 130 ◦ C.
ZSM-5, HM and Amberlyst 70 were also tested at 150 ◦ C
and 1.7 MPa. The DME moles produced per gram catalyst and
methanol conversion as a function of reaction time is shown
in Figs. 2 and 3. It can be seen that Mordenite has about
half of activity of the Amberlyst 70. Although both catalysts
showed some methanol conversion, the reaction temperature
and pressure were higher than those desired.
Amberlyst 15, Amberlyst 35, Amberlyst 36, Amberlyst 70
catalyst performance was studied at 110 ◦ C and 900 kPa for 8 h
using 6 ± 0.005 g catalyst and 120 ± 0.2 g solution. Fig. 4 shows
the DME produced per gram of catalyst, while Fig. 5 shows the
methanol conversion as a function of reaction time. Both fig-
ures show that the DME production and methanol conversion
for Amberlysts 35 and 36 are higher than Amberlysts 15 and
70. This was expected because both Amberlysts 35 and 36 have
higher acidity than that Amberlysts 15 and 70.
The dehydration of methanol over Amberlyst 15, Amberlyst
35, Amberlyst 36 and Amberlyst 70 at 130 ◦ C and pressure of
900 kPa was also carried out to examine the effect of temper-
ature on performance of these catalysts. The trend in activity
Amberlyst 35 Amberlyst 36 Amberlyst 70
5.13 5.38 2.7
50 33 36
300 240 220
0.7–0.95 0.6–0.85 0.5
0.35 0.2 NA
40 54 NA
150 150 190
828 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833
Fig. 2 – Cumulative moles of DME produced over Amberlyst
70 and H-ZSM-5 at 150 ◦ C and 1.7 MPa. Starting
composition pure methanol.
Fig. 3 – Methanol conversion over Amberlyst 70 and
H-ZSM-5 at 150 ◦ C and 1.7 MPa. The same experiments as
shown in Fig. 2.
was the same as observed at 110 ◦ C, with all catalysts showing
an increase in conversion.
Water inhibits catalytic methanol dehydration to DME over
either solid-acids or ion exchange resins. Water and methanol
molecules compete for adsorption at catalytic active sites on
Fig. 4 – Cumulative moles of DME produced over four
Amberlyst catalysts at 110 ◦ C and 900 kPa with pure
methanol initial composition.
Fig. 5 – Methanol conversion over four Amberlyst catalysts
at 110 ◦ C and 900 kPa using pure methanol as feed. The
same experiments shown in Fig. 4.
Fig. 6 – Cumulative moles of DME produced over four
Amberlyst catalysts at 130 ◦ C and 900 kPa with 2.5 mol/L
water in methanol solution as initial composition.
the surface of acid catalyst. Fig. 6 shows the reaction over
Amberlysts 15, 35, 36 and 70 at 130 ◦ C and pressure of 900 kPa
for an initial water concentrations of 2.5 M and 4 ± 0.005 g
catalyst. The same trend was observed with 3.5 M water in
methanol. The figure shows that Amberlyst 35 and Amberlyst
36 have the same activity and much higher than Amberlysts
15 and 70. The amount of the DME formation was observed to
decrease slightly by increasing the initial water concentration
from 2. 5 M to 3.5 M.
Fig. 7 shows the initial rates of reaction for Amberlysts 15,
35, 36 and 70 for pure methanol, 2.5 M and 3.5 M water con-
centrations in methanol. The initial rate of reaction for each
catalyst was obtained using nonlinear regression between the
DME moles produced and the reaction time data. The initial
rate of the reaction is equal to the value of the derivative
at time 0. This figure also shows that Amberlysts 35 and 36
have higher initial rate and show more activity than that
Amberlysts 15 and 70 for pure methanol and for both water
concentrations.
Fig. 8 shows the correlation between the initial rate of
the reaction at 110 ◦ C and the acidity capacity of Amberlyst
catalysts. It can be seen that there is a direct relation-
ship between the initial reaction rate and the acidity of
the catalyst. As the acidity of the catalyst increases, the
rate of reaction increases. These preliminary investigations
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833 829

4. Kinetic study of Amberlyst 35

Many investigations on the kinetics of the synthesis of DME

Fig. 7 – Initial reaction rate as a function of initial water
concentration over four Amberlyst catalysts at 130 ◦ C and
900 kPa.
by dehydration of methanol on solid-acid catalysts have been
published. The majority agree that the mechanism follows
either Langmuir–Hinshelwood (Gates and Johanson, 1971) or
Eley–Rideal kinetic models (Kiviranta-Paakkonen et al., 1998),
with water and DME both acting as reaction inhibitors. A sum-
mary of some of published kinetic models for DME synthesis
by catalytic dehydration of methanol is given in Table 3. Some
studies have proposed a mechanism for this reaction. Lu et al.
(2004) developed a detailed intrinsic mechanism containing
seven elementary reactions. This mechanism was used by
Mollavali et al. (2008) to derive kinetic global reaction equa-
tions, as shown in Table 3. The two groups used different rate
determining steps and arrived at different form rate equa-
tions. In the mechanism introduced by Gates and Johanson
(1969), shown in Fig. 9, it is assumed that two methanol
molecules occupy two adjacent acid sites. On the other
hand, in the Eley–Rideal (ER) model proposed by Kiviranta-
Paakkonen et al. (1998), only one methanol molecule adsorbs
on the acid site which reacts with a second molecule from the
liquid bulk phase (see Fig. 10). The models developed by these
groups can be represented by the generic equation:

kS KM
2 C2
M
rDME = n m (1)
(1 + KM CM + (KW CW ) + KD CD )

kS is the surface reaction rate constant, and KM , KW , and

Fig. 8 – The initial reaction rate of DME production as a
function of catalyst acidity for four Amberlyst catalysts at
110 ◦ C using pure methanol as initial composition.
show that Amberlysts 35 and 36 have higher DME pro-
duction and consequently higher initial rate of reaction
at lower reaction temperatures. Although, both Amberlysts
35 and 36 shown very similar activity Amberlyst 35 has
more crosslinks and less swelling than that Amberlyst 36.
Amberlyst 35 has better catalytic properties and physical
stability. Thus, we chose Amberlyst 35 for further kinetics
studies.
KD , and CM , CW , and CD are the adsorption equilibrium
constants and concentration of methanol, water and DME,
respectively. The power m takes the value of two for the
Langmuir–Hinschelwood model and a value of one for the
Eley–Rideal model. The value of n is 0.5, 1 or 2.
Most models presented in the literature show that water
formed during the reaction inhibits the reaction, and that
the inhibition by dimethyl ether is very small compared to
water. In addition, Gogate et al. (1990) indicates that because
of higher vapour pressure of DME compared to methanol and
water, it can be assumed that the mol fraction of DME in the
liquid-phase will be much less than that of water or methanol.
Hence, the extent of the reverse reaction is decreased and the
equilibrium conversion is close to 100%. The generic equation
represented by Eq. (1) can be simplified to:
kS KM
2 C2
rDME = M
(2)
n m
(1 + KM CM + (KW CW ) )

Table 3 – Kinetic models studied for methanol dehydration to DME.

Reaction kinetic equation Catalyst used Reference
kK2 P2
1. rDME M M Ion exchange resin Gates and Johanson (1971)
(1+KM PM +KW PW +KD PD )2
1/2
kKM P
2. rDME = 1/2
M ␥-Al2 O3 Bercic and Levec (1992)
1+KM P K P
M W W
kK2 (P2 −(PW PD /Keq ))
3. rDME = M M
√ 4 ␥-Al2 O3 Bercic and Levec (1992)
(1+2 KM PM +KW PW )
(P2 /PW )−(PD /Keq )
4. rDME = M ␥-Al2 O3 Lu et al. (2004)
(1+KM PM +KW PW )2

kKM C2
5. rDME = M
(1+KM CM +KW CW +KD CD )
Ion exchange resin An et al. (2004)

kPM −(k/Keq )(PD PW /PM )

6. rDME = 1+KM PM +(PW /KW )
␥-Al2 O3 Mollavali et al. (2008)
830 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833

Fig. 9 – Gates and Johanson (1969) mechanism for methanol dehydration reaction.

Fig. 10 – Kiviranta-Paakkonen et al. (1998) mechanism for methanol dehydration reaction.

All of the calculations performed for the kinetic modelling We can linearize Eqs. (3) and (4) with respect to methanol
study were based on the calculation of the initial rate of reac- concentration:
tion, as discussed earlier.
 0.5
To test the importance of external diffusion on the reac- C2M 1 K
Model 1 (LH) =  + M CM (5)
tion kinetics, the effect of stirring speed was examined. It was (rDME )0 kS kS
found that for 750 rpm and 650 rpm the initial reaction rates
were 0.047 and 0.046 mol/g cat h, thus the external diffusion is  
C2M 1 KM
not a limiting factor. Model 2 (ER) = + CM (6)
(rDME ) kS kS
The effect of internal diffusion limitation on overall kinet-
ics of dehydration of methanol to DME was investigated
To determine which model fits the experimental data in
by comparing two catalyst sizes, 0.2–0.6 mm. For the two
the absence of water, the left hand side of Eqs. (5) and (6)
tests, the reactor was charged with 4 g Amberlyst 35 and
was plotted versus methanol concentration, CM , as shown in
120 g methanol, pressurized to 900 kPa and heated to 130 ◦ C.
Figs. 12 and 13. Comparison of these figures indicates that
The stirrer speed was set to 750 rpm. There was no change
the experimental data fits better with Langmuir–Hinshelwood
observed in the rate of production of DME, therefore, we can
(model 1) for the temperature range studied. Hence, the
conclude that internal diffusion was not significant.
Langmuir–Hinshelwood was selected as the best model for our
study.
4.1. The effect of methanol concentration on reaction Using a linear regression for the data shown in Fig. 13, the
rate surface reaction rate constant, kS and the adsorption equi-
librium constants, KM were determined. For higher methanol
The first set of tests was performed to examine the effect of concentrations, the value obtained for KM CM is significantly
methanol concentration on the rate. The methanol concentra-
tion was varied between 5 M and 24.6 M using tetrahydrofuran
(THF) as an inert diluent. For temperatures of 110, 120, and
130 ◦ C, the reactor was charged with 10, 6 or 4 g Amberlyst
35, respectively, and 120 g of methanol/THF solution. Fig. 11
shows the effect of methanol concentration on the initial
reaction rate for three temperatures. The initial rate stayed
relatively constant in the range of methanol concentrations
investigated.
The observed effect of methanol concentration can be used
to discriminate the models. In the absence of water, Eq. (2) can
be written to express the initial reaction rate as:

kS KM
2 C2
M
Model 1 (LH) (rDME )0 = 2
(3)
(1 + KM CM )

Fig. 11 – Effect of methanol concentration on initial reaction

kS KM
2 C2
M rate at different temperatures and 900 kPa for different
Model 2 (rDME )0 = (4)
1 + KM CM concentrations of methanol/THF solutions.
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833 831

Fig. 14 – Arrhenius plot for methanol dehydration reaction

Fig. 12 – Left hand side of Eq. (6) versus methanol
at temperature range of 110–135 ◦ C and 900 kPa using pure
concentration.
methanol as initial composition.

higher than 1 (25  1); thus, 1/ kS term in Eq. (5) is negligi-

ble compared to KM CM / kS term. If we ignore the 1 in Eq. (5),
then the model can be simplified further:

kS KM
2 C2
M
Model 1 (LH) rDME = = kS (7)
(KM CM )

Because it was found that the methanol concentration had

negligible effect on initial reaction rate, two more tests
were conducted at temperatures of 115 and 135 ◦ C. Linear
regression of the Arrhenius plot (see Fig. 14) was used to
determine parameters k0 and Ea in Arrhenius equation (Eq.
(8)) were determined. The calculated values for k0 and Ea
are 6.12 × 107 kmol/s kg cat and 98 kJ/mol, respectively. The
value of apparent activation energy is similar to the activa- Fig. 15 – Effect of water concentration on initial reaction
tion energy calculated by Kiviranta-Paakkonen et al. (1998) and rate at 130 ◦ C and 900 kPa using different concentrations of
Di Stanislao et al. (2007) which are 95 kJ/mol and 98 kJ/mol, methanol/water solutions as initial composition.
respectively.
charged with 4 g Amberlyst 35 catalyst and 120 g of 1.5, 2.5
 −E 
kS = k0 exp (8) and 3.5 M water/methanol solutions. Then the reactor was
RT pressurized to 900 kPa and heated to 130 ◦ C. Fig. 15 shows
that water concentration has significant effect on the initial
4.2. The effect of initial water concentration on
reaction rate, consistent with previous work (An et al., 2004;
reaction rate
Kiviranta-Paakkonen et al., 1998). In the presence of water, Eq.
(2) becomes:
In this set of tests, the initial water concentration in the reac-
tor was varied from 0 to 3.5 M to determine the effect of water kS KM
2 C2
M
concentration on the initial reaction rate. The reactor was rDME = (9)
n 2
(KM CM + (KW CW ) )

Eq. (9) can be rearranged to the linear form:


n
kS (KW CW )
√ =1+ (10)
rDME KM CM

To determine the best value of n in Eq. (10) from the suggested

values (i.e. 0.5, 1 and 2), the left hand side of Eq. (10) was plotted
versus (C0.5
W /CM ) and (CW /CM ) as shown in Figs. 16 and 17. It
can be seen from these two figures, when n = 1.0, the linear
regression line fits the experimental data better. Thus Eq. (10)
can be written as:

kS KW CW
√ =1+ (11)
rDME KM CM

Fig. 13 – Left hand side of Eq. (5) versus methanol The value for the surface reaction rate constant kS calculated
concentration. by linear regression of Eq. (12) is 1.19 × 10−5 mol/kg cat s at
832 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833

0.5
Fig. 16 – Left hand side of Eq. (11) versus (CW /CM ). Fig. 18 – Plot of ln(KW /KM ) versus 1/T in temperature range
110–135 ◦ C and 900 kPa using 3.5 mol/L water/methanol
solution as initial composition.
130 ◦ C, which is in good agreement with the value we obtained
from the initial rates for the set of experiments conducted in
the absence of water (i.e. 1.21 × 10−5 mol/kg cat s)

1 1 1 KW CW
√ =  +  (12)
rDME kS kS KM CM

In Eq. (11), KW and KM are temperature dependence adsorp-

tion equilibrium constants of water and methanol and can be
defined using Van’t Hoff relationship

 −H 
W
KW = KW0 exp (13)
RT

 −H 
M
KM = KM0 exp (14)
RT

Fig. 19 – Comparison of model predictions for typical

The ratio of KW /KM can be written as
experiments used to obtain the rate parameters. Note that
KW
Q the initial rates were used in the analysis. The lines are the
= K exp (15) model predictions.
KM RT

where KW0 /KM0 = K and Q = (HM − HW ). To determine K and Q
in Eq. (15), set of experiments was conducted using a constant
water concentration (i.e. 3.5 mol/L) with reactor tempera-
tures of 110, 115, 120, 130 and 135 ◦ C. Using linear regression
between ln(KW /KM ) versus 1/T shown in Fig. 18, K and Q val-
ues found to be 1.57 × 10−3 and 24.6 kJ/mol, respectively. The
temperature dependence of the ratio of adsorption equilib-
rium constants of water and methanol can be calculated with
following equation.
KW
 2964

= exp −6.46 + (16)
KM T
Finally, the predictive ability of the model is tested against
the experimental data. The initial rates being used to calculate
the rate constants, the resulting model was used to predict the
results from the full experiments. Fig. 19 shows three exper-
iments with the predictions. Two are for pure methanol as
a starting composition, with two temperatures selected. The
third shows the experiment with added water. It is seen from
the figure that the fits are reasonable.

5. Conclusions

The activity of a series commercial solid-acid catalysts for

Fig. 17 – Left hand side of Eq. (11) versus (CW /CM ).
the dehydration of methanol to dimethyl ether was tested.
The catalysts were ␥-alumina, HY, HZSM-5, HM zeolites and
ion exchange resins (Amberlyst 15, Amberlyst 35, Amberlyst
36, and Amberlyst 70). It was found that ␥-alumina and HY,
HZSM-5, HM zeolites did not have a promising activity in the
temperature range of 110–135 ◦ C. Ion exchange catalysts had
 chemical engineering research and design 9 0 ( 2 0 1 2 ) 825–833
significant activity, with Amberlysts 35 and 36 having similar
activities.
The kinetics of dehydration of methanol to DME over
Amberlyst 35 was studied in the absence of mass transfer lim-
itations to determine a reaction kinetic model. The methanol
concentration did not have any effect on the reaction rate,
which is in accordance with the mechanism proposed by Gates
and Johanson (1971). In this mechanism, the two molecules of
methanol, occupy two adjacent acid sites, and the reaction
happens between those molecules. It was also found that the
presence of water had inhibiting effect on the reaction rate
by competing with methanol molecules over acid sites. It was
found that the Langmuir–Hinshelwood model is the best fit for
data.
Acknowledgements
This work was sponsored by a grant from Alberta Agricultural
Research Institute. The authors would also like to acknowl-
edge helpful discussions with Dr. D. Bressler and Dr. K.T.
Chuang.
References
An, W., Chuang, K.T., Sanger, A.R., 2004. Dehydration of methanol
to dimethyl ether by catalytic distillation. Canadian Journal of
Chemical Engineering 82, 948–955.
Bercic, G., Levec, J., 1992. Intrinsic and global reaction rate of
methanol dehydration over gamma-alumina pellets.
Industrial and Engineering Chemistry Research 31, 1035–1040.
833
Di Stanislao, M., Malandrino, A., Patrini, R., Viva, A., Brunazzi, E.,
2007. Green Fuel Synthesis via Reactive Distillation, Récents
Progrès en Génie des Procédés, Numéro 94. SPFG, Paris,
France, pp. 1–8.
Gates, B., Johanson, L., 1969. The dehydration of methanol and
ethanol catalyzed by polystyrene sulfonate resins. Journal of
Catalysis 14 (1), 69–76.
Gates, B., Johanson, L., 1971. Langmuir–Hinshelwood kinetics of
the dehydration of methanol catalyzed by cation exchange
resins. AIChE Journal 17 (4), 981–983.
Gogate, M.R., Lee, B.G., Lee, S., Kulik, C.J., 1990. Kinetics of liquid
phase catalytic dehydration of methanol to dimethyl ether.
Petroleum Science and Technology 8 (6), 637–671.
Khandan, N., Kazemeini, M., Aghaziarati, M., 2008. Determining
an optimum catalyst for liquid phase dehydration of methanol
to dimethyl ether. Applied Catalysis A: General 349, 6–12.
Kiviranta-Paakkonen, P.K., Struckmann, L.K., Linnekoski, J.A.,
Krause, A.O.I., 1998. Dehydration of the alcohol in the
etherification of isoamylenes with methanol and ethanol.
Industrial and Engineering Chemistry Research 37, 18–24.
Lu, W., Teng, L., Xiao, W., 2004. Simulation and experimental
study of dimethyl ether synthesis from syngas in a fluidized
bed reactor. Chemical Engineering Science 59 (22–23),
5455–5464.
Mollavali, M., Yaripour, F., Atashi, H., Sahebdelfar, S., 2008.
Intrinsic kinetics study of dimethyl ether synthesis on
(–Al2 O3 ). Industrial and Engineering Chemistry Research 47,
3265–3273.
Semelsberger, T.A., Borup, R.L., Greene, H.L., 2006. Dimethyl ether
(DME) as an alternative fuel. Journal of Power Sources 156,
497–511.
Spivey, J., 1991. Review: Dehydration catalysts for the
methanol/dimethyl ether reaction. Chemical Engineering
Communications 110, 123–142.