Catalysis Communications 6 (2005) 147–152

www.elsevier.com/locate/catcom

Catalytic dehydration of methanol to dimethyl ether (DME)

over solid-acid catalysts
a,* *,a,
F. Yaripour , F. Baghaei I. Schmidt b, J. Perregaard b

a
Catalysis Research Group, Arak Center, Petrochemical Research & Technology Company NPC, P.O. Box 1435, Arak, Iran
b
Haldor Topsøe A/S Research Laboratories, Nymøllevej 55, DK-2800, Lyngby, Denmark

Received 14 June 2004; accepted 30 November 2004

Available online 5 January 2005

Abstract

A series of solid-acid catalysts with different components contents were prepared by coprecipitation (sol–gel) method. These sam-
ples comprised c-alumina and modified c-Al2O3 with silica. The effects of silica various contents have been investigated on purpose
to determine an optimum one. Dehydration of methanol to dimethyl ether (DME) on solid-acid catalysts was studied in a fixed-bed
flow reactor at a temperature of 300 °C under atmospheric pressure and a GHSV of 15,600 h
1. The catalysts have been character-
ized using BET, TGA, XRD, NH3–TPD and elemental analysis techniques and also the results are reported. According to the exper-
imental results, the pure c-Al2O3 catalyst shown a good catalytic activity, but this sample undergoes a fairly rapid and irreversible
deactivation. Silica-modified catalysts have shown better performance compared to the untreated c-Al2O3. It was found that surface
areas increase with increasing in the silica loading at aluminosilicate catalysts. Also, by modifying alumina with silica, it was
observed that the surface acidity of aluminosilicate catalysts increased with increasing in silica loading. The sample with 6 wt% silica
loading has exhibited the best conversion without any by-product.
Ó 2004 Elsevier B.V. All rights reserved.

Keywords: Methanol; Dehydration; Dimethyl ether; Solid-acid catalysts

1. Introduction rofluorocarbons (CFC), has received attention as an

LPG alternative, transportation fuel in substitution
Air pollution is one of the most serious environ- to diesel and power generation [2–5].
mental problems all over the world. Since, onboard Catalytic dehydration of methanol over solid-acid
diesel engines of buses and trucks exhaust a huge catalysts offers a potential process for dimethyl ether
amount of NOx and particulates, a clean alternative synthesis. Several catalysts having activity and selectiv-
fuel is desired [1]. In recent years, dimethyl ether ity for the catalytic conversion of methanol into DME
(DME), which is now used as a raw material for mak- are known, the so called acidic dehydration catalysts
ing chemicals and aerosol propellant such as hair [6,7]. From the literature, it can be concluded that the
spray, shaving cream to replace ozone-destroying chlo- reaction takes place on different solid-acid catalysts such
as c-alumina, modified-alumina with silica and phos-
*
phorus, Al2O3–B2O3 and molecular sieves materials
Corresponding author. Tel.: +98 861 228 7701 4; fax: +98 861 228
7705.
(chabazites, mordenites, SAPOs, . . .), in a temperature
E-mail addresses: [email protected], [email protected] range of 250–400 °C and pressures up to 18 bar [6–10].
(F. Yaripour). Extensive research has been made especially in order

1566-7367/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.catcom.2004.11.012
148 F. Yaripour et al. / Catalysis Communications 6 (2005) 147–152

to find catalysts having higher selectivity for the ether 2 °C/min to remove the residual water, ammonia,
formation and less tendency to coke formation [7]. It ammonium nitrate and organic substances. In more
has recent been found that if the reaction of methanol detail aluminosilicate catalysts with 1, 3, 6, 9 and
dehydration take place on the modified-alumina cata- 15 wt% silica loading, were applied DME-AlS1,
lysts with silica, can be caused to reduce the amounts DME-AlS2, DME-AlS3, DME-AlS4 and DME-AlS5,
of coking and by-products [5,7]. respectively.
In the present work, the catalytic dehydration of
methanol to DME has been studied over c-Al2O3 and 2.3. Characterization of the catalysts
modified-alumina with silica catalysts. The effect of sil-
ica on the catalytic activity of c-Al2O3 have been inves- BET surface area, pore volume and pore diameter
tigated. In addition, we have investigated the correlation were measured by N2 adsorption–desorption isotherm
of the activity results with surface acid properties of at liquid nitrogen temperature using Autosorb-3B
these catalysts. So, we have attempted to characterize (Quantachrome, USA). Prior to the adsorption–desorp-
the properties of catalysts by surface area measure- tion measurements, all the samples were degassed at
ments, XRD, etc. This work is part of a long-range ef- 300 °C in N2 flow for 16 h.
fort to develop a new know–how for the conversion of The XRD pattern of all the calcined samples were
methanol to DME at Haldor Topsøe A/S. recorded on an X-ray diffractometer (X
Pert PRO,
Model PW3040/60 console) using Cu Ka monochroma-
tized radiation source and Ni filter in the range
2. Experimental 2h = 5–90°.
Ammonia temperature programmed desorption
2.1. Material (NH3–TPD) experiments in the temperature range 90–
850 °C were conducted using a 0.2 g of catalyst (in a
The starting materials were aluminium nitrate nano- fixed-bed quartz microreactor: i.d., 4 mm and length,
hydrate [ANN; Al(NO3)3 Æ 9H2O], tetraethyl orthosili- 200 mm) under 2% NH3/He gas mixture stream
cate [TEOS; Si(OC2H5)4], ethanol [C2H5OH] and (100 N ml/min) from room temperature up to 500 °C
ammonia solution [max. 33% NH3]. at a heating rate of 10 °C/min. Before all experiments,
catalyst sample (0.2 g) were treated in situ for 2 h in a
2.2. Catalysts preparation He flow rate of 100 N ml/min while the temperature
was raised from room temperature up to 500 °C at a
2.2.1. c-Al2O3 heating rate of 10 °C/min.
c-Al2O3 (DME-SCAT2) was obtained by the thermal TGA pattern was detected by a Haldor Topsøe TGA-
decomposition of the boehmite (c-AlOOH, DME- HT thermogravimetric analyzer in order to determine
SCAT1) precursor at 550 °C for 6 h at heating rate of proper calcination and phase transformation conditions.
2 °C/min. A commercially available catalyst (it has been The sample was heated from 25 to 950 °C at heating rate
developed by Haldor Topsøe A/S, DME-FCAT) was 5 °C/min, in flow of 100 N ml/min of helium.
used as a reference solid-acid catalyst for the methanol The chemical compositions of catalysts (Al, Si) were
dehydration catalyst. determined by an inductively coupled plasma-optical
emission spectrometer (ICP-OES, Perkin–Elmer, model
2.2.2. Aluminosilicates Optima 3000) and using Ar as plasmogene.
A series of aluminosilicate catalysts were prepared
by varying amount of silica loading (1, 3, 6, 9 and 2.4. Catalytic activity
15 wt%) by co-precipitation (sol–gel) method. All of
the samples were prepared at 50 °C. According to pre- Catalytic activity of all samples for dehydration of
vious study on the aluminosilicate catalysts [11,12], in methanol to DME were studied at a steady state in a
all these experiments the standard preparation condi- fixed-bed quartz reactor (i.d., 10 mm and length
tions were as follows: co-precipitation by the ammonia 650 mm) with an on-line GC. All of the experiments
solution, the precipitant solution was added slowly at were performed at atmospheric pressure, and tempera-
a rate of 4 ml/min, the concentration of ANN and ture of 300 °C. Prior to catalytic testing the samples
TEOS in ethanol is 2 mole/l, H2O/TEOS = 10, mixing were crushed and then sieved. For each experiment,
time 1440 min and NH4OH/C2H5OH ratio = 1/3. The 0.5 g of the catalyst (grain size 25–50 mesh or 300–
mixture was allowed to react in a glass beaker using 710 lm) was loaded to the reactor by packing quartz
a magnetic stirrer (IKA, cylindrical shape, 30 mm wool pads at both of ends of catalyst bed. Nitrogen
length) at about 500 rpm at 50 °C for 24 h. After pre- saturated by pure methanol (11% MeOH in N2) was
cipitation, the product dried at 110 °C overnight and used as feed, and the gas hourly space velocity
calcined in air at 650 °C for 6 h at heating rate of (GHSV) was monitored to 15,600 h
1 (feed flow rate
 F. Yaripour et al. / Catalysis Communications 6 (2005) 147–152 149

of 130 ml/min). The temperature is used for saturating 3. Results and discussion
the N2 stream with methanol was 15 to 18 °C. Before
the test, the catalyst sample was activated in situ at 3.1. Catalyst characterization
atmospheric pressure from 25 to 200 °C at heating
rate 5 °C/min, in a flow of nitrogen 150 ml/min. Then Suitable temperature of calcination for phase trans-
the feed was sent into fixed-bed reactor and the reac- formation of boehmite to c-alumina was determined
tor temperature increase up to 300 °C simultaneously. by the TGA method. Fig. 1 shows the TGA profile of
In order to survey the activity of the various solid- boehmite. According to this figure, it is clear that
acid catalysts, feed was sent for 70 h. After 70 h, anal- boehmite transformed to c-Al2O3 at higher temperature
ysis of the outlet gas stream was performed using GC. of 425 °C. Calcination of boehmite under conditions
(550 °C, 6 h) has shown good results because boehmite
does not show any weight loss at higher temperature
2.5. Product analysis of 550 °C.
Table 1 summarizes the specific BET surface areas,
The reaction products were analysed by an on-line pore volumes and pore diameters of the catalysts em-
gas chromatograph (Hewlett–Packard, Model: HP- ployed in this study. The phase transformations of
5890 series II) using CP-PorPlot Q-HT column (con- Al2O3 (boehmite) to c-Al2O3 become significant at high-
nected to FIDA) for separating DME and MeOH and er temperature of 425 °C and remarkably decrease the
Al2O3/NaOH column (connected to FIDB) for separat- surface area of the c-Al2O3 catalyst. This variation in
ing C1 –Cþ6 hydrocarbons. The product lines were heated the porous properties; is likely due to change in the pri-
electrically where necessary in order to avoid unwanted mary particle size and/or crystal growth of either the
condensation of methanol and water. Several types of bulk material or the active phase (sintering phenomena).
different methanol dehydration catalysts were prepared Surface area of the aluminosilicate samples monoto-
and tested. As measures of the catalytic activity, and nously increased with increasing the silica content, and
for purpose of quantitative comparison with commercial all the samples had surface areas higher than 250 m2/g,
ones, the conversion of the feed methanol (X%) and indicating that they had well-developed pore structures.
turnover frequency (TOF) were used. These were de- The chemical compositions of the solid-acid catalysts
fined as follows: used in the present study are also given in Table 1. In all
cases, the real content was very similar to the nominal one.
  The XRD patterns (not shown) of all the calcined
Outlet partial pressure of MeOH
X ¼ 1
 100%; catalysts exhibited different diffraction lines. The XRD
Inlet partial pressure of MeOH
  analyses of the DME-SCAT1 catalyst determined the
2  Number of mole of DME produced per second
TOF ¼ presence of the orthorhombic (JCPDS No. 83-2384)
Number of mole of acid site on the catalyst

Fig. 1. TGA curves of boehmite sample (in a flow of 100 N ml/min of helium, heating rate 5 °C/min).
150 F. Yaripour et al. / Catalysis Communications 6 (2005) 147–152

Table 1
Compositions, acidity contents and some characterization data of the calcined catalysts
Properties Surface Pore Pore Surface area of Amount of Acidity Components
area (m2/g) volume (cm3/g) diameter the used catalyst NH3 (mmole) (mmole/g dry sample)
Al (%) Si (%) P (%)
(nm) (m2/g)
Catalyst
DME-SCAT1 338.7 0.65 7.69 – 0.0397 0.437 44.9 – –
DME-SCAT2 268.2 0.68 10.18 295 0.0402 0.442 52.8 – –
DME-FCAT 168.1 0.50 11.65 184 0.0261 0.287 29.6 6.91 4.26
DME-AlS1 252.4 0.43 6.75 224 0.0276 0.301 43.1 0.767 –
DME-AlS2 286.7 0.46 6.48 274 0.0269 0.295 41.4 2.044 –
DME-AlS3 333.8 0.56 6.75 305 0.0283 0.311 37.6 3.72 –
DME-AlS4 363.5 0.48 5.32 294 0.0324 0.362 35.6 5.83 –
DME-AlS5 383.2 0.54 5.63 359 0.0347 0.398 32.6 9.28 –

boehmite (c-AlOOH) crystal phase. However, XRD Also, by modifying alumina with silica, it is observed
analyses of DME-SCAT2 confirmed presence of two that the surface acidity of aluminosilicate catalysts in-
crystalline phase, namely cubic c-Al2O3 (JCPDS No. creased with increasing in silica loading. The NH3–
79–1558) and some amorphous phase. It was confirmed TPD profiles of DME-FCAT, DME-SCAT2 and
that the DME-FCAT sample has an c-Al2O3 crystalline DME-AlS3 (DME-S) catalysts are shown in Fig. 2,
phase on an amorphous background. All the aluminosil- while the temperature of desorption maxima (Tdi) and
icate calcined catalysts exhibited diffraction line of c- the amount of desorbed ammonia and water are sum-
Al2O3. However, aluminosilicate containing more than marized in Table 2. The peak temperatures for all the
6 wt% silica are almost amorphous c-Al2O3 and show catalyst samples were approximately identical (260–
a shoulder by 2h = 28° could be Al2(SiO4)O/kyanite tri- 275 °C), indicating that acid strength was not changed
clinic crystalline phase. by the silica content. TPD results shown that all the acid
The results of ammonia TPD-titration (Boric acid sites on the surface catalysts are a medium type. The
procedure) that contain the amount of desorbed ammo- amounts of desorbed water decreased with addition
nia and acidity content of the prepared catalysts are the silica to pure c-Al2O3 catalyst.
summarized in Table 1. It is well known that the acidic
sites on the surface solid-acid catalysts are two type; 3.2. Catalytic activity
Brønsted and Lewis acid sites. So, it is well known that
Brønsted acid sites are able to catalyze the dehydration Table 3 show the catalytic performance at steady
of methanol [13]. The prepared c-Al2O3 has higher acid- state in methanol dehydration over reference and pre-
ity than commercial one. By modifying c-Al2O3 with sil- pared catalysts at the atmospheric pressure. As indicated
ica it was observed that the surface acidity improved. by the results of Table 3, c-Al2O3 catalyst is active and

Fig. 2. NH3–TPD profiles of: (a) DME-SCAT2, (b) DME-FCAT and (c) DME-S catalysts.
 F. Yaripour et al. / Catalysis Communications 6 (2005) 147–152 151

Table 2
NH3–TPD: temperature of desorption maxima (Tdi) and amounts desorbed
Catalysts DME-SCAT2 DME-FCAT DME-S
Parameters
Amount of NH3 (mmole) 0.0402 0.0261 0.0283
Amount of H2O (mmole) 0.538 0.400 0.514
Acidity (mmole/g dry sample) 0.442 0.287 0.311
Tdi (°C), ammonia peak 261 270 262
Tdi (°C), water peak 657.5 705 660
Temperature range of NH3 desorbed (°C) 171–793 176–729 158–770
Temperature range of H2O desorbed (°C) 385–854 437–853 434–854

Table 3
Methanol dehydration over reference and modified c-alumina catalysts
Catalysts DME-AlS1 DME-AlS2 DME-AlS3 DME-AlS4 DME-AlS5 DME-SCAT2 (c-Al2O3) DME-FCAT
Parameters
DME (wt%) 75.6 73.4 71.3 64.7 58.6 63.23 67.24
MeOH (wt%) 24.4 26.6 28.7 35.3 41.6 38.77 32.76
CH4 (wt%) 0.0326 0.0128 – – – – –
Conversion (X%) 86.4 84.5 85.0 83.5 73.6 77.15 82.13
TOF (103/s) 121.9 158.2 164.5 178.9 457.4 131.7 333.7
Rate (10
7 mol g
1 s
1) 2.20 2.84 1.88 3.22 8.25 3.77 3.79
Rate/surface area 0.87 0.99 0.56 0.89 21.5 2.24 1.41
(10
6 kmole/h m2)

selective for DME formation. In the case of commercial similar performances in comparison with unmodified
catalyst, methanol conversion slightly increased. The re- c-Al2O3. This subject can be explained the surface of so-
sults shown that c-Al2O3 catalyst have high initial rate lid acid catalysts contain the Lewis and Brønsted acid
of the DME formation but slowly deactivates with time sites which it has now been found that the dehydration
on stream in methanol dehydration to DME. reaction took place mainly on Brønsted acid sites. It is
To improve the catalytic performance, c-Al2O3 cata- also known that Brønsted acid sites on c-Al2O3 are very
lyst was modified with Si. By modifying c-Al2O3 with sil- weak. Therefore, since the acidity of the above-men-
ica, it has been observed that catalytic performance is tioned catalysts is almost similar with c-Al2O3, it can
positively affected (Table 3). The rates of DME forma- concluded that the amount of Brønsted acid sites is
tion do not change on all the prepared catalysts. But, low. So, it can be concluded that the aluminosilicate cat-
the formation rate of dimethyl ether increased signifi- alysts with mild acidity will be effective on DME 10 syn-
cantly in DME-AlS5, in comparison with that of the thesis in the long term reaction. The low conversion over
pure c-Al2O3. The significant acceleration of the dehy- the samples having more than 6 wt% silica suggest that
dration rates may be due to an enhancement of acid the excess amount of silica creates the hindrance to the
strength by the intercalation and/or an increase in alu- existence of acid sites due to the accumulation on the
mina surface area on intercalation [14]. The methanol surface of the catalyst, which is already evidenced from
partial pressure and the amount of the inlet methanol XRD and SEM spectra (not shown). Under the experi-
depended to the cooling bath temperature. It should mental conditions, methanol was converted into DME
be pointed out that the DME selectivity (SDME%) is al- and methane was formed as the main by-product in cat-
most 100% on the different catalysts. The variations of alysts which contains the SiO2 amount of lower than
the DME selectivity can be attributed to the methanol 3 wt%. It can be concluded that the number of strong
partial pressure. SiO2-modified catalysts showed better acidic centers is high on these catalysts surfaces. This
performance than the untreated c-Al2O3. By modifying is due to DME formation is related to sites with weak
alumina with silica, it is observed that the surface acidity and medium acidity whereas, selectivity to hydrocar-
of aluminosilicate catalysts increased with increasing in bons such as methane is due to the number of strong
silica loading. Also, it is understood that the catalytic acidic centers.
activity decreases after 6 wt% of silica loading. Further-
more, the surface acidity and catalytic activity of cata-
lysts with 3 wt% silica loading is low. The activity,
selectivity and stability studies showed that the 4. Conclusions
c-Al2O3 modified with 6 wt% silica is more active, noth-
ing by-product and less deactivated by the reaction con- A series of solid-acid catalysts were prepared by co-
ditions. 15 wt% SiO2-modified catalyst showed almost precipitation method. c-Al2O3 and modified-alumina
152 F. Yaripour et al. / Catalysis Communications 6 (2005) 147–152

with silica catalysts were used for the synthesis of DME discussions and their collaboration at catalyst character-
through dehydration of methanol. ization by different methods.
Methanol is dehydrated to form DME over acid sites
of solid catalysts. DME formation is mainly related to
sites with weak and medium acidity. Under the experi- References
mental conditions used, methanol was converted into
DME as a main product and methane was formed as [1] T. Takeguchi, K. Yanagisawa, T. Inui, M. Inoue, Appl. Catal. A:
the main by-product on the some catalysts. General 192 (2000) 201–209.
[2] Ki-Won Jun, Hye-Soon Lee, Hyun-Seog Roh, Sang-Eon Park,
SiO2-modified catalysts showed better performance Bull. Kor. Chem. Soc. 24 (2003) 106–108.
than the untreated c-Al2O3. The effect of SiO2 contents [3] A.M. Duarte de Farias, A.M. Lavogade, M.A. Fraga, M.I. Pais
were studied to find out the optimum SiO2 contents. da Silva, L.G. Apple, DME Production from Methanol Dehy-
Among the SiO2-modified c-Al2O3 catalysts, 6 wt% dration over Boria Modified Alumina, LACAT/Instituto Nac-
ional de Technologia, Rio de Janeiro, Brazil, 2001.
SiO2/c-Al2O3 (DME-AlS3) exhibited the best conversion
[4] M. Xu, D.W. Goodman, A. Bhattacharyya, Appl. Catal. A 149
without any by-product. DME was selectively formed (1997) 303–309.
over DME-AlS3 catalyst in the catalytic dehydration [5] Ki-Won Jun, Hye-Soon Lee, Hyun-Seog Roh, Sang-Eon Park,
of methanol at 300 °C and atmospheric pressure in com- Bull. Kor. Chem. Soc. 23 (2002) 803–806.
parison with other catalysts. [6] A.W. Wang, S. Weigel, G. Muraro, Topical Report Molecular
Sieves as Catalyst Methanol Dehydration, Air Products and
Chemicals Inc., 2002.
[7] Jorgen Topp-Jorgensen, US Patent 4536485.
[8] G. Bercic, J. Levec, Ind. Eng. Chem. Res. 31 (1992) 1035–1040.
Acknowledgements [9] J. Bandiera, C. Naccache, Appl. Catal. 69 (1991) 139–148.
[10] L.D. Brake, D.C. Woodcraft, EP0099676 A1.
The financial support to this work by the National [11] F.A. Twaiq, A.R. Mohamed, S. Bhatia, Micropor. Mesopor.
Petrochemical Company (NPC), Iran and Haldor Top- Mater. 64 (2003) 95–107.
søe A/S (HTAS), Denmark as joint research activity [12] K. Okada et al., Micropor. Mesopor. Mater. 37 (2000) 355–364.
[13] J.M. Camplo, A. Garcia, J.F. Herencia, D. Luna, J.M. Marinas,
whose aid is gratefully acknowledged. The authors A.A. Romero, J. Catal. 151 (1999) 307–314.
would like to thank HTAS employees of Syngas/Chem- [14] K. Hashimoto, Y. Hanada, Y. Minami, Y. Kera, Appl. Catal. A:
icals Technology and Analysis Divisions for valuable General 141 (1996) 57–69.