This document summarizes a study on the catalytic dehydration of methanol to dimethyl ether over γ-alumina. The researchers investigated the effects of temperature and feed composition on the conversion of methanol and deactivation of the catalyst. They found that methanol conversion strongly depended on the reactor operating temperature, increasing with higher temperatures. Using pure methanol as a feed resulted in slow catalyst deactivation, while adding water to the feed increased deactivation significantly. A temperature-dependent model was developed to predict methanol conversion and reasonably correlated with experimental data.

1. Effects of temperature and feed composition on catalytic dehydration of methanol
to dimethyl ether over c-alumina
Freshteh Raoof a
, Majid Taghizadeh a
, Ali Eliassi b,*, Fereydoon Yaripour c
a
Chemical Engineering Department, Babol University of Technology, Babol, Iran
b
Chemical Industries Research Department, Iranian Research Organization for Science and Technology (IROST), No. 71, Forsat Street, Enghelab Avenue, Tehran, Iran
c
Catalysis Research Group, Petrochemical Research & Technology Company NPC, Tehran, Iran
a r t i c l e i n f o
Article history:
Received 29 September 2007
Received in revised form 17 March 2008
Accepted 18 March 2008
Available online 22 April 2008
Keywords:
Clean fuel
Dimethyl ether
Methanol dehydration
Catalyst deactivation
a b s t r a c t
Catalytic dehydration of methanol to dimethyl ether (DME) is performed in an adiabatic fixed bed heter-
ogeneous reactor by using acidic c-alumina. By changing the mean average temperature of the catalyst
bed (or operating temperature of the reactor) from 233 up to 303 °C, changes in methanol conversion
were monitored. The results showed that the conversion of methanol strongly depended on the reactor
operating temperature. Also, conversion of pure methanol and mixture of methanol and water versus
time were studied and the effect of water on deactivation of the catalyst was investigated. The results
revealed that when pure methanol was used as the process feed, the catalyst deactivation occurred very
slowly. But, by adding water to the feed methanol, the deactivation of the c-alumina was increased very
rapidly; so much that, by increasing water content to 20 weight percent by weight, the catalyst lost its
activity by about 12.5 folds more than in the process with pure methanol. Finally, a temperature depen-
dent model developed to predict pure methanol conversion to DME correlates reasonably well with
experimental data.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
Dimethyl ether (DME) with the chemical formula of CH3–O–CH3
is the simplest ether for preparation of some chemicals such as di-
methyl sulfate and high-value oxygenated compounds. In addition,
it has been used as an aerosol propellant to replace chlorofluoro
carbons which can destroy the ozone layer of the upper atmo-
sphere. It is a colorless gaseous with an ethereal smell. Unlike
methane, DME does not require an odorant because it has a sweet
ether-like odor. Dimethyl ether is also a clean fuel alternative to
liquified petroleum gas (LPG), liquified natural gas (LNG), diesel
and gasoline [1,2].
DME can be made from natural gas, coal, or biomass. This fuel
burns with a visible blue flame and is non-peroxide forming in
the pure state or in aerosol formulations. DME is a volatile organic
compound, but is non-carcinogenic, non-teratogenic, non-muta-
genic, and non-toxic [1]. Its physical and chemical properties in
comparison with diesel can be summarized as follows:
À The low heat value of DME is only 64.7% of that of diesel, there-
fore a larger amount of fuel supply is needed to deliver the same
power output for the engine.
À Cetane number of DME is higher and its auto ignition tempera-
ture is lower than that of diesel.
À DME has only got C–H and C–O bonds, but no C–C bond and it
contains about 34.8% oxygen, therefore the combustion prod-
ucts such as carbon monoxide and unburned hydrocarbon emis-
sions are lower than those of natural gas.
À The latent heat of evaporation of DME is much higher than that
of diesel, so it will be beneficial to the NOx reduction due to the
larger temperature drop of the mixture in the cylinder.
À DME’s boiling point is À24.9 °C and it must be pressurized to
keep it in liquid state under ambient conditions.
Two processes are used for DME production, indirect [3–8] and
direct processes [9–11]. In indirect process, methanol is converted
to DME in a catalytic dehydration reactor over a solid-acid catalyst
by the following reaction
2CH3OH $ CH3OCH3 þ H2O ð1Þ
In the second process (direct process), a synthesis gas (a mix-
ture of H2 and CO gases) is used as the feed of the process. In this
process, the synthesis gas is primarily converted to methanol and
then it is followed by methanol dehydration to DME. The net reac-
tion is as follows:
3CO þ 3H2 $ CH3OCH3 þ CO2 ð2Þ
0016-2361/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.fuel.2008.03.025
* Corresponding author. Tel./fax: +98 21 88838324.
E-mail address: alieliassi@yahoo.com (A. Eliassi).
Fuel 87 (2008) 2967–2971
Contents lists available at ScienceDirect
Fuel
journal homepage: www.fuelfirst.com

2. In this work, we have considered the conversion of methanol to
DME by dehydrating process. This process is moderately exother-
mic and usually is conducted in an adiabatic fixed bed reactor.
One of the most important problems related to the operation of
heterogeneous catalysts is the loss of catalyst activity with time-
on-stream. In the indirect process to produce DME using acidic
c-alumina, water has the most important effect on catalyst deacti-
vation [11].
In this work, acidic c-alumina has been used as the catalyst for
the dehydration of methanol to DME. Methanol conversion to DME
and the deactivation of the catalyst have been studied in a labora-
tory-scale system at various operating temperatures. A tempera-
ture dependent model has been developed to predict methanol
conversion to DME at various temperatures.
2. Experimental
2.1. Apparatus
A schematic diagram of the laboratory-scaled system employed
in this study is shown in Fig. 1. Pure methanol was pumped from
methanol storage tank at a rate of 0.121 L/h to an evaporator and
then to a supperheater before entering the reactor. The super-
heated methanol was sent to an adiabatic fixed bed reactor. The
length of catalyst bed could be adjusted for a given experiment.
The axial reactor temperature at any point of the catalyst bed
was measurable via a thermo-well using a thermocouple. The reac-
tor outlet products were passed through an air cooler and a double
pipe heat exchanger to cool down to the ambient temperature.
Cooled products were sent to a gas–liquid separator. A back pres-
sure regulator (BP-LF690, pressure Tech2000, England) was placed
on this separator to regulate the system pressure. Before any
experiments all of set up was swept by using nitrogen gas. Reaction
products were analyzed by a gas chromatograph (Varian CP-3800)
equipped with TCD and FID detectors. Also, the remaining metha-
nol in the exit reactor products was measured and with compari-
son to the entrance methanol, the methanol conversion was
estimated. BET surface area, pore volume and pore radius of the
catalysts were measured by N2 adsorption–desorption isotherm
at liquid nitrogen temperature using Autosorb-3B (Quantachrome,
USA). Experiments were carried out at average reactor bed temper-
ature (operating temperatures) ranging from 233 to 303 °C at con-
stant atmospheric pressure. The operating conditions as well as
some of the characteristic of the system are reported in Table 1.
For study of the adiabatic status of the applied reactor, an
experiment was performed as a blank run (without any reaction)
and the temperature profile of the bed was measured by using
thermocouple and thermowell. The result of this experiment is
shown in Fig. 2. According to this figure, the performance of the
reactor as an adiabatic reactor is reasonable. Therefore, the adia-
batic assumption for the used reactor is reasonable. This experi-
ment was performed by passing air through to the catalyst bed.
2.2. Chemicals
Acidic c-alumina, 1–2 mm in particle size, was obtained from
BASF (Kat.D10-10 S4). Physical and chemical properties of the
catalyst are reported in Table 2. Methanol was obtained from Iran
Fig. 1. A schematic diagram of the experimental apparatus for catalytic production of DME from methanol: (1) nitrogen cylinder, (2) methanol feed tank, (3) dosing pump, (4)
flow meter (5) mixer, (6) evaporator, (7) preheater, (8) adiabatic fixed bed reactor, (9) air cooler, (10) condenser, (11) liquid–gas separator, (12) back pressure regulator.
Table 1
Operating conditions and some characteristics of the laboratory-scale apparatus and
the catalyst used in this study
Set up characteristics
Absolute pressure (atm) 1
Average temperature of the catalyst bed (°C) 233–303
Liquid methanol flow rate (L/h) 0.121
Reactor inside diameter (mm) 18
Thermowell outside diameter (mm) 6.35
Catalyst
Catalyst weight (g) 8.2
Catalyst volume (cm3
) 14.2
Liquid hourly space velocity (hÀ1
) 8.5
2968 F. Raoof et al. / Fuel 87 (2008) 2967–2971

3. Petrochemical Company (IPC) as the feed to the reactor. Its compo-
sition is given in Table 2.
3. Results and discussion
3.1. The temperature effect on methanol conversion
The first series of experiments were conducted to investigate
the effect of operating temperature of the reactor on methanol con-
version. Since the reactor was adiabatic and the reaction exother-
mic, the temperature of catalyst bed increased from the feed
inlet temperature to a maximum value, as illustrated in Fig. 3.
The change in catalyst bed average temperature, or reactor operat-
ing temperature, versus inlet feed temperature is shown in Fig. 4.
According to this Figure, there seem to be a relatively linear
correlation between the change in the reactor operating tempera-
ture and the feed inlet temperature.
Fig. 5 shows the variations of conversion at different inlet feed
temperatures versus time on stream. This Figure demonstrates that
methanol conversion was not substantial below 230 °C and in-
creased with temperature to the equilibrium amount of 85% at
250 °C. Beyond this temperature the conversion remained nearly
constant. It should be noted that methanol conversion did not
change substantially with time at a set feed inlet temperature
meaning that the activity of the catalyst used remained nearly con-
stant throughout the 30 h of experiment.
3.2. Catalyst deactivation
Fig. 6 shows the change in methanol conversion with time for
pure methanol feed (solid circles) and methanol–water mixture
feed (solid triangles; 80 wt% methanol and 20 wt% water) at the in-
let feed temperature of 250 °C. The slops of the trend lines in Fig. 6
show the catalyst activity loss with time on stream. This Figure
indicates that conversion of methanol rapidly decreased with time
when the mixture of methanol–water was used as the reactor feed.
For the latter feed, the catalyst activity loss was about 12.5 times
larger than that of the pure methanol feed. This phenomenon is
due to the negative effect of water on c-Al2O3 acidic sites. A drop
in methanol conversion due to the presence of water decreased
the heat of reaction released, hence, lowering the reactor temper-
ature as can be seen in Fig. 7.
200
220
240
260
280
300
0 5 10 15 20
Reactor length (cm)
Temperature(°C)
Fig. 2. Performance of the adiabatic reactor in a blank experiment (without any
reaction).
Table 2
Physical and chemical properties of the catalyst and methanol composition
Catalyst
Catalyst particle sizes (a) 1 mm < a < 2 mm
Knife edge hardness (KG) Min 4
Inner surface (m2
/g) 200
Tapped density (g/L) 650
L. o. attrition (%) Max 3.5
L. o. Ignition (%) Max 6
Main component Al2O3
Na2O (%) Max 0.10
Fe2O3 (%) Max 0.15
Methanol
Purity (%) 99.85
H2O (%) 0.1
Acetone (%) 0.003
190
210
230
250
270
290
310
330
0 1 2 3 4 5 6
Catalyst bed length (cm)
Temperature(°C)
217 °C 230 °C 240 °C 242 °C 245 °C
250 °C 259 °C 280 °C
Fig. 3. Temperature profile along the catalyst bed at various inlet feed tempera-
tures. Solid lines show the trend of these variations.
210
230
250
270
290
310
210 230 250 270 290
Inlet temperature (°C)
Bedaveragetemperature(°C)
Fig. 4. Catalyst bed average temperature (or operating temperature) versus inlet
feed temperature. Solid circles are experimental data and solid line shows the trend
of changes.
F. Raoof et al. / Fuel 87 (2008) 2967–2971 2969

4. Table 3 shows some characterization data of the used catalysts
for two different experiments, namely pure methanol and metha-
nol with water as the process feed. According to this table, the
catalyst characterization data had no considerable changes after
the experiments. As a result, decreasing of catalyst activity in
methanol + water feed, is not related to the changes of catalyst sur-
face area. Water is believed to block the active sites of the catalyst
for methanol consumption through competitive adsorption with
methanol on the catalyst surface [6,11].
3.3. The mathematical model for methanol conversion
A mathematical model has been developed to relate pure meth-
anol conversion to the reactor operating temperature. This model
is presented by the following equation:
ln
x
1 À x
¼
À13669
T
À 0:1523T þ 93:171
where, x is fractional methanol conversion and T is the operating
reactor temperature (°C). Fig. 8 presents a comparison between
the proposed model results and the experimental data. This Figure
demonstrates a good consistency between the calculated and
experimental results. The average relative error of this model in
predicting the results of this study is less than 3%. The cross plot
of estimated versus experimental values for methanol conversion
is presented in Fig. 9. As expected, this Figure demonstrates a good
Table 3
Some characterization data of the used catalysts
Catalyst sample Surface area (m2
/g) Pore volume (cm3
/g) Pore radius (nm)
1 191.55 0.450 470.14
2 184.93 0.433 468.81
Catalyst 1 is used in the process with pure methanol and catalyst 2 is used with
methanol + water as feed process.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
190 210 230 250 270 290 310 330
Operational temperature (°C)
Methanolconversion
Fig. 8. A comparison between the calculated conversions of pure methanol using
the proposed model (dashed lines) and the experimental data (circles).
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 10 20 30
Time on stream (hr)
Methanolconversion
217 °C
230 °C
240 °C
245 °C
250 °C
259 °C
280 °C
Fig. 5. Methanol conversions versus time on stream at different inlet feed
temperatures.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30 35
Time on stream (hr)
Methanolconversion
Fig. 6. Methanol conversion variations with catalyst time on stream for pure me-
thanol (solid circle) and methanol–water mixture (80 wt% methanol and 20 wt%
H2O; solid triangle) at 250 °C inlet feed temperature. Solid lines show the trends of
the results.
240
250
260
270
280
290
300
0 2 4 6 8
Catalyst bed length (cm)
Operationaltemperature(°C)
Fig. 7. Catalyst bed temperature profile for pure methanol (solid circle) and met-
hanol–water mixture (80 wt% methanol and 20 wt% H2O; solid triangle) at 250 °C
inlet feed temperature. Solid lines show the trends of the results.
2970 F. Raoof et al. / Fuel 87 (2008) 2967–2971

5. consistency between the proposed model and the experimental
results.
4. Conclusion
Methanol dehydration over acidic c-alumina in an adiabatic
packed bed reactor to produce DME has been studied over inlet
feed temperature ranging from 217 to 280 °C. This study showed
that the methanol conversion to DME was not substantial at feed
temperatures below 230 °C and increased to the limit of about
85% at 250 °C where it remained unchanged with further increase
in feed temperature. A mathematical model was developed and
successfully tested for pure methanol conversion to DME at
different reactor operating temperatures. The obtained results also
showed that water presence in the feed led to a visible reduction in
methanol conversion due to catalyst deactivation.
Acknowledgments
This work was partially supported by Iran Petrochemical
Research and Technology Company. The authors specially thank
to the Chemical Industries Research Department of Iranian
Research Organization for Science and Technology (IROST) for their
generous assistance with this project.
References
[1] Semelsberger TA, Borup RL, Green HL. Dimethyl ether (DME) as an alternative
fuel. J Power Source 2006;156:497–511.
[2] Semelsberger TA, Ott KC, Borup R. Generating hydrogen-rich fuel-cell feeds
from dimethyl ether (DME) using physical mixtures of a commercial Cu/Zn/
Al2O3 catalyst and several solid-acid catalysts. Appl catal B 2006;65:291–300.
[3] Bercic G, Levec J. Catalytic dehydration of methanol to dimethyl ether – kinetic
investigation and reactor simulation. Ind Eng Chem Res 1993;32:2478–84.
[4] Xu M, Goodman DW, Bhattacharyya A. Catalytic dehydration of methanol to
dimethyl ether (DME) over Pd/Cab–O–Sil catalysts. Appl Catal A 1997;149:
303–9.
[5] Yaripour F, Baghaei F, Schmidt Ib, Perregaard J. Catalytic dehydration of
methanol to dimethyl ether (DME) over solid-acid catalysts. Catal Commun
2005;6:147–52.
[6] Xu M, Lunsford JH, Goodman WD, Bhattacharyya A. Synthesis of dimethyl
ether (DME) from methanol over solid-acid catalysts. Appl Catal A 1997;149:
289–301.
[7] Eliassi A, Savadkoohi L, Kargari A. Effects of various process parameters on
temperature profile of adiabatic fixed bed reactor for production of dimethyl
ether (DME) from methanol. Chem Eng Commun 2007;194:1495–502.
[8] Voss B, Joensen F, Hansen JB. Preparation of fuel grade dimethyl ether, US
patent 5908963, 1999.
[9] Takeguchi T, Yanagisawa KI, Inui T, Inoue M. Effect of the property of solid acid
upon syngas-to-dimethyl ether conversion on the hybrid catalysts composed
of Cu–Zn–Ga and solid acids. Appl Catal A 2000;192:201–9.
[10] Ramos FS, Duarte de Farias AM, Borges LEP, Monteiro JL, Fraga MA, Sousa-
Aguiar EF, et al. Role of dehydration catalyst acid properties on one-step DME
synthesis over physical mixtures. Catal Today 2005;101:39–44.
[11] Jun K-W, Lee H-S, Roh H-S, Park S-E. Highly Water-Enhanced H-ZSM-5
Catalysts for dehydration of methanol to dimethyl ether. Bull Korean Chem Soc
2003;24:106–10.
0
0.2
0.4
0.6
0.8
1
0 0.2 0.4 0.6 0.8 1
xexp
xcal
Fig. 9. A comparison between calculated conversions (xcal) using the proposed
model and the experimental conversions (xexp).
F. Raoof et al. / Fuel 87 (2008) 2967–2971 2971