Studies in Surface Science and Catalysis 130 A. Corma, F.V. Melo, S. Mendioroz and J.L.G. Fierro (Editors) 9 2000 Elsevier Science B.V. All fights reserved.
Oxidative Dehydrogenation of Propane Over Alkali-Mo Catalysts Supported on Sol-Gel Silica-Titania Mixed Oxides Rick B. Watson and Umit S. Ozkan* Department of Chemical Engineering, The Ohio State University, Columbus, OH 43210, USA A group of novel molybdena catalysts for propane oxidative dehydrogenation (ODH) is presented. These catalysts, which were prepared using a sol-gel/co-precipitation technique, were supported on silica-titania mixed oxides and promoted by alkali metals (Li, Na, K, Cs). While reaction results show no specific trend with the type of alkali promoter, comparisons indicate that promotion with potassium gives a superior catalyst. Propylene yields around 30% were obtained with K/Mo catalysts at 550~ in dilute feed stream experiments. Differences in reaction performance are found to be closely related to surface acidity, oxidation- reduction behavior, and dispersion of the active components. 1.
Olefin feedstocks, such as propylene, find wide use in industrial processes. The strong industry demand for propylene has led to numerous research efforts on the use of a catalytic oxidative dehydrogenation (ODH) process to produce propylene from propane. This process offers several advantages compared to conventional cracking and dehydrogenation processes because ODH is not thermodynamically limited at lower temperatures and usually does not lead to the formation of coke and smaller hydrocarbons. Research efforts on ODH primarily focus on the development of catalysts active at low temperatures (<550~ that show low selectivities to carbon oxides. The most active and selective catalysts studied in recent literature include vanadium-magnesium, vanadia supported on niobium, and nickel molybdates. In particular, promising results have been obtained when molybdate-based catalysts are promoted or supported. For example, Ni-CoMo  V-Nb-Mo/TiO2, K-MnMoO4, and K2MoO414] have shown promise in ODH and other partial oxidation reactions. Furthermore, the increase in activity and/or selectivity of alkali (Li, Na, K, Rb, and Cs) doped metal oxide catalysts has been widely studied [5-7]. These positive effects arise from the alkali's ability to alter oxidation/reduction behavior, affect surface acidity, and/or cause a synergism between alkali and transition metal oxide phases. *To whom correspondence should be addressed Phone: (614)292-6623 Fax: (614)292-3769 E-mail: [email protected]
1884 Our initial work focused on bulk MoO3 and K2MoO4 catalysts. Bulk, precipitated potassium molybdate catalysts were up to 60% selective to propylene formation but at a conversion less than 6%. Furthermore, these samples showed considerable selectivity to cracking products, i.e. methane, ethane, and ethylene. Un-supported potassium molybdate catalysts had low surface areas because of the precipitation and calcination procedures. In an attempt to form highly dispersed molybdate catalysts and possibly create metal-support interactions, molybdate catalysts supported over the mixed oxides of silica and titania were developed. Study of silica-titania mixed oxides has gained much attention because of their high activity for epoxidation reactions of olefins with hydroperoxides . Silica-titania mixed oxides have been studied extensively [9-14] for attributes such as acidity, porosity, TiO-Si bond connectivity, and phase separations. However, few studies focus on their use for transition metal oxide supports. In this work, alkali-promoted molybdenum catalysts supported over silica-tiania mixed oxides have been studied in regard to their activity for the oxidative dehydrogenation (ODH) of propane. The effect of alkali doping on the catalyst surface characteristics and, in turn, on the catalytic performance has been examined. The catalysts used in this study have been synthesized by a "one-pot" sol gel/co-precipitation technique. The main focus of this work was to investigate the use of different alkali dopants and the effect of alkali/Mo molar ratio. 2.
Catalyst preparation Catalysts were prepared using a modified sol-gel/co-precipitation technique. Sol-gel chemistry is extensive  and cannot be covered here completely. The method used here involves the reactions of metal alkoxide precursors in an alcohol solvent when contacted with water. Ammonium heptamolybdate (AHM) and alkali (Li, Na, K, Cs) hydroxides were used as molybdenum and alkali precursors, respectively. For silica-titania mixed oxides, tetraethylorthosilicate (TEOS) and titanium(IV)isopropoxide (TIPO) were used. The solvent was isopropyl alcohol. This method is referred to as sol-gel/co-precipitation because as the silica and titania precursors are hydrolyzed and precipitate out of solution, active metal species present in the aqueous solution, that are insoluble in alcohol, also precipitate. A more detailed description of the preparation method has been given elsewhere [ 16]. The catalysts reported in this paper are listed in Table 1. Our previous results have shown that a silica-titania molar ratio of 1:1 support performed the best in the ODH reaction. Therefore, all catalysts were supported over Si:Ti 1:1. These include a series of alkali-doped molybdate catalysts with alkali/Mo molar ratio of 0.1 at constant 10% weight loading of Mo. Potassium-promoted catalysts with different alkali/Mo ratios were also synthesized at a constant Mo loading of 10% and examined in propane ODH. 2.1
Oxidative dehydrogenation of propane Steady-state reaction experiments were carried out in a fixed-bed, quartz reactor, operated at ambient pressure. Catalyst samples, ranging from 0.1 g to 1.5g, were held in place by a quartz frit. To minimize effects from any homogeneous reaction or surface-initiated gas phase reaction and to provide a short residence time for propylene formed, the dead volume of the quartz microreactor was filled with quartz wool and/or ceramic beads.
1885 Reaction temperatures ranged from 450 to 550~ The feed consisted of propane, oxygen, and nitrogen usually at a flowrate of 25 cm3/min. Two different feed conditions were evaluated, dilute (5% propane) and concentrated (26% propane). However, the propane/oxygen molar ratio was held constant at 2. The product distributions maintained a carbon balance of 100% (+/- 5%).
Table 1 Catalyst Compositions Composition
BET Surface Area (m2/g)
Si:Ti 1:1 10%Mo/Si:Ti 1:1 10% Li/Mo=0.1)/Si:Ti 1:1 10%, Na/Mo=0.1)/Si:Ti 1:1 10% K/Mo=0.07)/Si:Ti 1:1 10%, K/Mo=0.1)/Si:Ti 1:1 10% K/Mo=0.3)/Si:Ti 1:1 10%, Cs/Mo=0.1)/Si:Ti 1:1
320 229 170 323 136 191 166 188
RESULTS AND DISCUSSION
Effect of preparation method on physical properties
Alkali-doped molybdate catalysts show no specific trend in BET surface area with differing alkalis or amounts of potassium added. All catalysts containing alkali exhibited lower surface area than the "molybdenum only" catalyst with the exception of the sodiumdoped catalyst. The nitrogen adsorption-desorption isotherm of the Si:Ti 1:1 support indicated a micro to meso-porous structure. The pore size distribution was calculated using the desorption isotherm. This yielded an average pore diameter of 2.1nm and a pore volume of 0.34cma/g. X-ray diffraction of the Si:Ti 1"1 support yielded a pattem typical of a silicatitania sample with one broad peak located at a d spacing of 3.59 A, which is the most intense diffraction line from the anatase structure. Molybdenum species were not detected in the 10%(alkali/Mo) loaded Si:Ti 1:1 catalysts. This suggests that molybdena species on the mixed oxide supports are more finely dispersed than on a silica or titania support alone. Raman spectra of the Si:Ti 1:1 support showed that the bands associated with the anatase structure were shifted to lower wavenumbers than those of pure anatase . Evidence for SiO-Ti connectivity also was observed in the Raman spectra. Bands associated with terminal Mo=O stretches around 950cm -~ were observed on all catalysts studied except for the 10%(Na/Mo=0.1)/Si:Ti 1:1 catalyst. Considering the higher surface area of this catalyst, surface molybdate species may be undetectable. Broad bands arising from Mo-O-Mo vibrations were observed around 850cm -1. From the Raman spectra, there is no evidence of crystalline MoO3. Molybdenum in these catalysts is in the state of surface coordinated molybdena species. Details of further characterization studies are presented elsewhere [ 16].
Equal surface area reaction experiments Alkali/Mo catalysts were tested in the ODH reaction using equal surface area loading (65m 2) in the reactor and at temperatures of 450~ and 550~ The feed percentages for these experiments were Nz/C3/Oz: 61/26/13. The results are presented in Table 2. Table 2 Reaction Results for Alkali/Mo Catalysts 10% (Mo)/Si:Ti 1:1 T (~ T (~ 10%(Li/Mo =0.1)/Si:Ti 1:1 T (~ T (~ 10%(Na/Mo=0.1)/Si:Ti 1:1 T (~ T (~ 10% (K/Mo=0.1)/Si:Ti 1:1 T (~ T (~ lO% (Cs/Mo=O.l)/Si:Ti 1:1 T (~ T (~
C3H 6 11.4 20.0
Yield(%) CO 2 CO 2.3 4.9 3.1 6.7
C2H 4 0.0 0.4
CH 4 0.1 0.8
Conditions: equal surface area (65m2), %N2/C3/O2:61%/26%/13%, 25cc/min. Results indicate that the alkali promoted catalysts behave rather similarly under these conditions. However, the Cs-promoted catalyst shows a considerably lower activity at 450~ whereas the potassium-promoted catalyst exhibits the highest yield of propylene. The propylene yield of these catalysts increases in the order Cs, Li, Na, K at 450~ and Li, Na, Cs, K at 550~ There appears to be no correlation between catalyst activity and atomic weight of the alkali promoter. However, the differences in performance may be due to not only the size of the alkali but other parameters such as dispersion and interaction of the support with alkali and molybdena species. While the differences were small, the potassiumpromoted catalyst gave a better ODH performance, warranting its further investigation. 3.3
Selectivity comparisons at equal conversions
To compare catalysts with different K/Mo ratios, a series of equal conversion experiments were run at 450~ Equal conversions were obtained by changing the mass of catalyst loaded inside the reactor. The feed concentrations were %N2/C3/O2: 61/26/13. The results are presented in Table 3. It is seen that alkali-promoted catalysts give rise to significantly higher selectivities to propylene and lower selectivities to carbon oxides.
1887 The selectivities to cracking products do not seem to vary much with the incorporation of the alkali promoter. Another important feature of these comparison experiments is that the product distribution obtained over the bare support heavily favors carbon oxides. Ethylene production over the support is also substantial.
Table 3 Equal Conversion Comparison for K/Mo Catalysts Selectivity (%) C3H6 CO2 CO C2H4 CH4
"~5% Calls Conversion Si:Ti 1"1 Support
32.9 86.0 92.3 96.1
10%(Mo)/Si:Ti 1:1 10%(K/Mo =0.07)/Si:Ti 1:1 10%(K/Mo=0.3)/Si:Ti 1:1
40.8 5.6 5.5 3.2
22.5 8.4 2.2 0.5
2.9 0.1 0.1 0.1
0.6 0.0 0.0 0.1
C2H6 0.3 0.0 0.0 0.0
Selectivity (%) -10% C3Hs Conversion Si:Ti 1"1 Support 10%(Mo)/Si:Ti 1:1 10%(K/Mo =0.07)/Si:Ti 1:1 10%(K/Mo=0.3)/Si:Ti 1:1 Conditions: 450~ ~
23.7 79.3 88.5 91.7
44.1 13.6 6.4 3.3
27.2 7.0 5.0 4.9
C2H4 CH4 4.1 0.1 0.1 0.1
0.9 0.0 0.0 0.0
C2H6 0.0 0.0 0.0 0.0
Dilute feed stream experiments
A set of experiments was performed with more dilute propane concentrations, NJCaH8/O2: 92.5/5/2.5, which exhibited a higher molar yield of propylene. The results are presented in Table 4. Under these conditions, the catalyst with a K/Mo ratio of 0.07 gave the highest yield of propylene yield (around 30 %). It was also seen that the propylene yield starts decreasing when alkali loading is further increased. The bare support showed a very small propylene yield for this set of experiments. The potassium-containing catalysts showed Table 4
Reaction Results for Dilute Feed Stream Experiments Si:Ti 1:1 Support
C3H8 Conversion (%) C3H6 CO2 11.2
C3H6 CO2 39.7
10%(K/Mo=0.07)/Si:Ti 1:1 48.4 10%(K/Mo=0.3)/Si:Ti 1:1 35.0 Conditions: 5s feed residence time, 550~
Yield (%) CO C2H4 CH4 C2H6 3.4
C2H4 CH4 C2H6 2.1
C2H4 CH4 C2H6 2.0
C2H4 CH4 C2H6 0.4
1888 improved yields to propylene and lower yields of CO, CO2, and cracking products when compared to the un-promoted catalyst. It appears that molybdena catalysts supported over Si:Ti mixed oxides has ODH activity, but the incorporation of small quantities of an alkali promoter significantly improves the product distribution in favor of propylene. Our characterization results [ 16], as well as some previous reports in the literature [17,18], suggest that this effect is achieved by suppression of the acidity and reducibility of the molybdenum surface sites. These changes in the surface characteristics, in turn, affect the alkane activation performance of the catalyst and the adsorption/desorption characteristics for the olefins. It appears that there is an optimum alkali/Mo ratio, which provides the best balance between the alkane activation and the olefin combustion rates to give the highest propylene yields. It should also be kept in mind that the differences observed in the ODH performance of these catalysts cannot be explained solely in terms of the direct modification of the surface by the alkali promoters. Other structural and surface characteristics, which may be affected by the sol-gel parameters, can also change the catalytic behavior. Studies are underway to examine the oxygen mobility, dispersion, support structure and the adsorption/desorption behavior of these catalysts. These studies will help correlate the surface and structural characteristics of this group of novel catalysts with their catalytic properties for propane oxidative dehydrogenation. 4. R E F E R E N C E S 1.
2. 3. 4. 5. 6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
Stem, D. L., Grasselli, R., J. Catal., 167 (1997) 550. Meunier, F. C., Yasmeen, A., Ross, J.R.H., Catal. Today 37 (1997) 33. Driscoll, S. A., Gardner, D. K., Ozkan, U. S., J. Catal., 147 (1994) 379. Erdohelyi, A., Mate, F., Solymosi, F., J. Catal., 135 (1992) 563. Mross, W.D., Catal. Rev.-Sci. Eng., 25(4) (1983) 591. Abello, M.C., Gomez, M.F., Cadus, L.E., Catal. Letters, 53 (1998) 53. O'Young, C.L., J. Phys. Chem., 93 (1989) 2016. Notari, B., Adv. Catal., 41 (1996) 253. Gao, X., Wachs, I. E., Catal. Today, 51 (1999) 233. Stakheev, A.Y., Shipiro, E.S., Apijok, J., J. Phys. Chem., 97 (1993) 5668. Lassaletta, G., Femandez,A., Espinos, J.P., Gonzalez-Elipe, A.R., J. Phys. Chem., 99 (1995) 1484. Waiters, J.K., Rigden, J.S., Dirken, P.J., Smith, M.E., Howells, W.S., Newport, R.J., Chem. Phys. Letters, 264 (1997) 539. Kumar, S.R., Suresh, C.,. Vasudevan, A.K, Suja, N.R., Mukundan, P., Warrier, K.G.K., Mat. Letters, 38 (1999) 161. Klein, S., Thorimbert, S., Maier, W.F., J. Catal., 163 (1996) 476. Brinker, C.J, Scherer, D.W., "Sol-Gel Science", Academic Press, New York, 1990. Watson, R.B., Ozkan, U.S., J. Catal, Accepted for publication. Feng, Z., Liu, L., Anthony, R.G., J. Catal., 136 (1993) 423. Grabowski, R., Grzybowska, B., Samson, K., Sloczynski, J., Stoch, J., Wcislo, K, Appl. Catal., 125 (1995) 129.
Financial support provided by the National Science Foundation (Grant# CTS-9412544) is gratefully acknowledged. The authors would also like to thank Dr. Gurkan Karakas for his technical assistance at the early stages of the project.