APPLIED CATALYSIS A:GENERAL
Applied Catalysis A: General 149 (1997) 303-309
Catalytic dehydration of methanol to dimethyl ether (DME) over Pd/Cab-O-Sil catalysts M i n g t i n g X u a, D. W a y n e G o o d m a n a,*, A l a k B h a t t a c h a r y y a b a Department of Chemistry, Texas A & M Unitersi~', College Station, TX 77843-3255, USA b Amoco Research Center, Naperville, IL 60566, USA
Received 8 March 1996; revised 2 July 1996; accepted 6 July 1996
Abstract Ten wt.-% Pd/Cab-O-Sil reduced at 300°C has been found to be an effective catalyst for the catalytic dehydration of methanol to dimethyl ether (DME). The presence of hydrogen in the reagent stream inhibited the catalytic activity, but increased the stability of the catalyst. High reaction temperature and low methanol partial pressure did not favor DME formation. Keywords: Palladium; Methanol; Dimethyl ether; Dehydration
1. Introduction Dimethyl ether (DME) has received recent attention as an alternative diesel fuel due to its low NO x emission, near-zero smoke, and less engine noise compared with traditional diesel fuels . Recently, researchers at Haldor Topsoe  have developed a hybrid catalyst which includes a methanol synthesis catalyst, i.e., C u / Z n O / A I 2 0 3 , and an ammonia-treated solid-acid catalyst, i.e., H-ZSM-5, to make DME from synthesis gas in one step. DME production from synthesis gas is thermodynamically more favorable than methanol, thus, in principle, the cost for DME production should be less than that of methanol provided a proper catalyst is available. The role of the specially treated H-ZSM-5 catalyst in the Topsoe process is to convert methanol in situ to DME. As an alternative to C u / Z n O / A I 2 0 3 , Pd/Silica has also been shown to be an active and selective methanol-synthesis catalyst. Poutsma et al.  demon* Corresponding author. [email protected]
( + 1-409)
( + 1-409)
0926-860X/97/$17.00 Copyright © 1997 Published by Elsevier Science B.V. All rights reserved. Pll S0926- 8 6 0 X ( 9 6 ) 0 0 2 7 6 - 1
M. Xu et al. /Applied Catalysis A: General 149 (1997) 303-309
strated that methanol can be produced from syngas highly selectively over P d / S i O 2 catalysts at 260-350°C and 150-1600 psig. Lunsford and co-workers [4,5] found that the selectivity and the activity of these Pd catalysts are strongly dependent on the nature of the support. For example, methanol was formed on Pd/SiO2(57) and Pd/Cab-O-Sil catalysts but was not observed on Pd/SiO2(01). The SIO2(57) support is neutral, whereas SiO2(01) and Cab-O-Sil are acidic silicas with similar amounts of reactive acidic groups per gram of catalyst. The authors believed that the metal particle size plays a dominant role, with the efficient catalysts being those containing small palladium crystallites onto which CO is weakly adsorbed. A small amount of DME was formed over a 3.8% Pd/Cab-O-Sil catalyst at 290°C. The formation of DME has also been reported by Ryndin et al.  over palladium supported on T-A1203, TiO 2, and ZrO 2. These authors explained the formation of DME by the acidity of the support, which catalyzes the dehydration of methanol to DME. In pulsed experiments, Hensel and Pines , found that reduced non-supported iridium and palladium oxides gave better ether selectivities (77% and 54%, respectively) than the transition metals iron, cobalt, platinum and copper, which showed no conversion of alcohol to ether in the catalytic dehydration of neopentyl alcohol. Palladium oxide on Cab-O-Sil (10% Pd) reduced at 300°C was an effective catalyst for the conversion of primary alcohols to the corresponding ethers . For example, in a flow type microreactor with a hydrogento-alcohol ratio of 3.1, the selectivity toward ether formation at 160°C was ca. 93% in the case of 1-propanol and butanol. Cab-O-Sil itself was catalytically inert. The substitution of hydrogen by helium reduced the alcohol conversion and ether selectivity. Licht et al. , explained the formation of ether by assuming intrinsic acid and basic sites. A similar study [9,10] over reduced nickel oxide supported on Cab-O-Sil suggested that these intrinsic acidic and basic sites were generated at the nickel/nickel oxide interface. The metal part of the catalyst acts as an electron acceptor (Lewis acid), while the oxide, an electron donor (Lewis base). The role of the intrinsic acidic sites has been demonstrated by the inclusion of 0.05 wt.-% sodium ions into the nickel catalyst containing 1.5 mol-% nickel oxide, which decreases the yield of ether from 79% to ca. 6%. Licht et al.  observed that in the dehydration of 2-propanol over Pd/Cab-O-Sil, the injection of air (ca. 50 m l / m i n ) into the reagent stream caused a pronounced increase in the selectivity to ether, with a simultaneous decrease in the initial activity of the catalyst. The authors believed that the addition of air resulted in an increase of new basic sites through metal oxidation. Since the formation of ether depends on the cooperative action of both acidic and basic sites, the proper balance of the two sites is essential for the high yield of ether. In the present work the catalytic dehydration of methanol to DME has been studied over a 10 wt.-% Pd/Cab-O-Sil catalyst. The effects of hydrogen, helium and oxygen on the catalytic activity and selectivity were investigated. This work
M. Xu et al. / Applied Catalysis A: General 149 (1997) 303-309
is part of a long-range effort to develop an integrated catalyst for the direct conversion of syngas to DME.
2. Experimental 2.1. Catalyst Cab-O-Sil (Grade M5), (fumed silica), whose physical properties have been reported by Fajula et al.  was obtained from Cabot Corporation. A 10 wt.-% Pd/Cab-O-Sil catalyst was prepared by impregnation of the Cab-O-Sil (10 g) with 2.7 g of Pd(NO3) 2 • 2H20 (99.9%, Johnson Matthey) dissolved in 20 ml doubly deionized water [7,8]. The resulting material was stirred and heated until a paste was formed. The paste was then left at room temperature for several days before drying at ca. 90°C overnight.
2.2. Catalytic reactor The reaction was carried out in a plug-flow reactor constructed of fused quartz. The upper section of the reactor, which contained the catalyst, was 10 mm i.d., while the lower section consisted of a 1 mm i.d. capillary tube. The smaller diameter allowed the products to pass rapidly out of the heated zone. Quartz chips (20-42 mesh) were placed on the inlet side of the catalyst bed to aid in preheating the reactants. A thermocouple inside a capillary well was located at the level of the catalyst on the outside of the reactor. Hydrogen and helium flows were regulated by mass flow controllers (Brooks, Model 5850).
2.3. Products and analysis In a typical experiment, 0.51 g of 10 wt.-% Pd/Cab-O-Sil (20-42 mesh), supported between two layers of quartz wool, was reduced in flowing hydrogen (ca. 30 m l / m i n ) at 190°C for 0.5 h, and then at 300°C for 16 h to avoid local overheating. The sample temperature was then lowered to 225°C in flowing hydrogen. At this temperature, methanol (Aldrich, A.C.S. HPLC grade) was fed (at approx. 0.15 atm) by bubbling either hydrogen or helium through two saturators connected in series at 22.5 _+ 0.5°C. The total pressure was 1.0 atm and total flow rate, 60 ml/min. In the experiment to study the effect of methanol partial pressure on the rate of methanol consumption and DME formation, methanol was introduced by using a syringe pump (Model 341A, Sage Instrument). The exit of the pump line was maintained in contact with the inner wall of the preheated gas inlet tube, ensuring the steady flow of methanol into reactor. The total gas flow through the catalyst was maintained constant at all methanol partial pressures. The effluent
M. Xu et al. / A p p l i e d Catalysis A: General 149 (1997) 303-309
gas was analyzed on-line by means of a Varian Model 3400CX gas chromatograph that was coupled with a Varian Model 4270 integrator. The product stream was separated on a super Q column in series with a Porapak Q column. A thermal conductivity detector (TCD) was used to detect the reaction products.
3. Results and discussion
As pointed out previously , the catalytic dehydration of methanol over solid-acid catalysts is characterized by favorable methanol conversion, high DME selectivity, and long-term stability. As indicated by the results of Fig. l, a 10 wt.-% Pd/Cab-O-Sil catalyst is also active and selective for DME formation. For example, with 10 min on stream at 225°C, a level of methanol conversion of 27.9% with 78.5% DME selectivity was achieved over the catalyst prereduced with hydrogen at 300°C for 2 h. The reaction side products were CO and C H 4. Over a period of ca. 4.5 h, methanol conversion decreased from 27.9% to 16.5%, while the DME selectivity decreased somewhat. The decrease in activity may be attributed to the formation of carbonaceous species on the palladium surface, which has been observed in the synthesis of methanol from syngas over palladium . As depicted in Fig. 1, regeneration of the deactivated catalyst with hydrogen at 300°C for 16 h resulted in an even more active catalyst. This enhancement of activity following the hydrogen treatment likely arise due to the fact that the palladium surface is cleaned by the hydrogen treatment of carbonaceous species yielding hydrocarbon products. This explanation is consistent with the fact that the length of reduction has a positive effect on the activity. After 2
° ~ 100
0 c e-
Time on Stream, h Fig. 1. Methanol conversion and D M E selectivity versus time on stream. Open and filled circles represent conversion and selectivity, respectively. ((3, O ) , Reduced in H 2 at 300°C for 2 h; (V, • ), reduced in H 2 at 300°C for 16 h. PT = 1.0 atm; PH2 = 0.85 atm; PMeOn = 0.14 atm.
M. Xu et al. / A p p l i e d Catalysis A." General 149 (1997) 303-309 50
tO O --
e-' t¢ e~
O ~ Ill
Time on Stream, min Fig. 2. Effect o f diluent gases on m e t h a n o l c o n s u m p t i o n a n d D M E selectivity. O p e n a n d filled circles represent methanol conversion with H 2 and He as diluent gases, respectively. Filled triangles show D M E selectivity.
h of reduction at 300°C, the level of methanol conversion is 28%; however, with a ca. 16 h reduction, the conversion increased to 32%. A similar result was reported by Licht et al. . It should be pointed out that Cab-O-Sil itself is inert toward the methanol dehydration reaction. As pointed out by Licht et al. , hydrogen is an integral part of the catalytic system. The authors reported that both the conversion and the yield of ether dropped considerably when 1-butanol was passed over a Pd/Cab-O-Sil catalyst in flowing helium rather than hydrogen. However, as depicted in Fig. 2, this behavior was not observed for the catalytic dehydration of methanol to DME. Instead, the methanol conversion increased from 32% to 37% upon the replacement of hydrogen by helium. That is, the presence of hydrogen has a negative effect on the activity. One possible explanation of the negative effect of hydrogen on activity is the formation of palladium hydride under reaction
i" > t- 101 O
o c ¢-
1000/T, K Fig. 3. Arrhenius plot o f m e t h a n o l conversion with different diluent gases. ( 0 ) , He, E a = 16 k c a l / m o l ; ( O ) , H 2, E a = 27 k c a l / m o l .
M. Xu et al. /Applied Catalysis A: General 149 (1997) 303-309
conditions [13,14]. Since the hydride phase is not stable at this temperature in the absence of hydrogen, purging the catalyst with helium should regenerate hydride-free palladium. With either hydrogen or helium as a diluent gas, DME was only slightly affected. Fig. 3 shows an Arrhenius plot of methanol conversion with respect to reaction temperature. The data were obtained in the order of decreasing temperature. After the lowest temperature point, the datum at the highest temperature was repeated to make sure that the catalyst was stable over the test period. The apparent activation energy for methanol conversion in the presence of hydrogen is 27 kcal/mol, which is consistent with the result obtained over solid-acid catalysts. Surprisingly, the E a is only 16 k c a l / m o l with helium as the diluent gas. The difference may result from the poisoning effect of H 2 through the formation of palladium hydride. The dependence of the rates of methanol conversion and DME formation on the partial pressure of methanol are shown in Fig. 4. The data in this figure were used to determine the power-law rate expression for rcn~o n and rDME. It can be seen that the rates for both methanol consumption and DME formation are half-order with respect to methanol partial pressures. It should be pointed out that the DME selectivity decreased as the partial pressure of methanol decreased. For example, the DME selectivity changed from 84% to 63% with a decrease in the methanol partial pressure from 229 to 27 Torr. At even lower methanol partial pressures, one might expect the main reaction products to be CO and C H 4. As over solid-acid catalysts, the surface methoxide species might be responsible for DME formation over Pd/Cab-O-Sil. In a FT-IR study over P d / S i O 2, Rasko et al.  have identified methoxy species, formed via the cleavage of an O - H bond over Pd. At low partial pressures, the surface
o "6 e-
pO.S (CHaOH),Torr Fig. 4. Rates of methanol consumption and DME formation with respect to the partial pressure of methanol. ( 0 ) , CH3OH; (O), DME.
M. Xu et al. /Applied Catalysis A: General 149 (1997) 303 309
coverage of the methoxide species is low, and therefore the chance for reaction between surface species to form DME is small. As the reaction temperature was increased from 225°C to 280°C, methanol conversion increased from 38% to 77%, while DME selectivity decreased from 78% to 47%. The other reaction products are CO and C H 4. Within the methanol-synthesis temperature, i.e., 270-340°C, over the Pd/Cab-O-Sii catalyst, the production of DME from methanol is not favorable.
4. Conclusions DME was selectively formed over Pd/Cab-O-Sil catalyst in the catalytic dehydration of methanol at 225°C. As the reaction temperature was increased, DME selectivity decreased. Although hydrogen had a negative effect on the catalytic activity, its presence reduced the rate of surface carbonaceous species formation, and therefore enhanced the catalytic stability. The rates of both methanol consumption and DME formation were half-order with respect to the methanol partial pressure. DME selectivity decreased with decreasing methanol partial pressure.
Acknowledgements We acknowledge with pleasure the financial support of this work by the Exploration and Production Technology Group of the Amoco Corporation.
References               
A.M. Rouhi, Chem. Eng. News, May 29 (1990). Haldor Topsoe, U.S. Patent 4 536 485 (1993). M.L. Poutsma, L.F. Elek, P.A. Ibarbia, A.P. Risch and J.A. Rabo, J. Catal., 52 (1978) 157. F. Fajula, R.G. Anthony and J.H. Lunsford, J. Catal., 73 (1982) 237. K.P. Kelly, T. Tatsumi, T. Uematsu, D.J. Driscoll and J.H. Lunsford, J. Catal., 101 (1986) 396. Y.A. Ryndin, R.F. Hicks and A.T. Bell, J. Catal., 70 (1987) 287. J. Hensel and H. Pines, J. Catal., 24 (1972) 197. E. Licht, Y. Schachter and H. Pines, J. Catal., 55 (1978) 91. H. Pines and T.P. Kobylinski, J. Catal., 17 (197(I) 375. H. Pines, J. Hensel and J. Simonik, J. Catal., 24 (1972) 206. M. Xu, J.H. Lunsford, D.W. Goodman and A. Bhattacharyya, Appl. Catal. A, 149 (1997) 289. P.J. Berlowitz and D.W. Goodman, J. Catal., 108 (1987) 364. F.A. Lewis, The Palladium Hydrogen System, Academic Press, New York, 1967. R.K. Nandi, R. Pitchai, S.S. Wang, J.B. Cohen, R.L. Burwell, Jr. and J.B. Butt, J. Catal,, 70 (1981) 298. J. Rasko, J. Bontovics and F. Solymosi, J. Catal., 146 (1994) 22.