Applied Catalysis, 75 (1991) 321-330 Elsevier Science Publishers B.V., Amsterdam
Comparison of the oxidative dehydrogenation of ethane and oxidative coupling of methane over rare earth oxides Eric M. Kennedy and Noel W. Cant* School of Chemistry, Macquarie University, NSW 2109 (Australia), tel. (-I 61-2)8058313, fax. (+ 61-2)8058285 (Received 14 February 1991, revised manuscript received 22 April 1991)
Abstract The characteristics of the oxidative dehydrogenation of ethane and the oxidative coupling of methane have been compared over four high purity rare earth oxides-LaxO,, CeO*, SmxO:, and PrsOu. With each oxide ethane reacts approximately four times as fast as methane and the activity order per g is PqO,, > SmzOs t Laz03 > CeO,. In terms of unit area is LazO, > Sm,Os 2 Pr,O,, >> CeOz. The selectivity to hydrocarbons increases with temperature for both reactions but the absolute values are not only much higher for ethane than for methane but also much less catalyst dependent. For ethane at 750°C the order in selectivity to ethylene is La,O, (74%) > Sm,O, (68%) >CeO, (57%) > PraO,, (53%). Under identical conditions, (13% hydrocarbon, 3.5% oxygen, balance helium) the selectivity of the oxidative coupling of methane to ethane plus ethylene together is about 40% for La*O, and 33% for Smz03 but less than 3% for CeO, and PrsOll, the oxides with variable valence state. This selectivity pattern can be explained in terms of two interacting factors, the average lifetime of alkyl and alkylperoxy species in the gas phase and the tendency of the oxides to oxidise alkyl radicals. The higher oxidation power of CeO, and PrsOll relative to Sm,O, and La203, also manifests itself in a low yield of carbon monoxide relative to carbon dioxide and a lesser ratio of hydrogen to water. The water-gas shift reaction is not at equilibrium for the former pair of oxides. The yield of minor products, propene during methane coupling and methane and butene during ethane oxidation, increases with temperature. The more selective oxides, La,O, and SmlOs, also show a small maximum in methane production at about 650°C which may be due to carbon-carbon bond cleavage in an adsorbed ethoxy species.
Keywords: rare earth oxides, ethane dehydrogenation, methane oxidative coupling.
The oxidative coupling of methane to ethane and ethylene has been intensively studied since its initial discovery almost ten years ago. Whilst many oxides capable of catalysing the reaction are now known, the primary chemistry of ethane production appears to be the same with each. The initial step is detachment of a hydrogen from methane with liberation of a methyl radical into the gas phase [ 1,2]. Ethane is then formed by direct dimerisation of the methyl radicals [ 3,4]. Details of the chemistry which produces ethylene and by-product carbon oxides is less clear. Isotope tracing measurements show that
0 1991 Elsevier Science
All rights reserved.
at high oxygen conversions and temperatures greater than 700°C some of the carbon oxides are certainly produced by further oxidation of C, products [ 5,6]. However increasing amounts of carbon oxides are made with all catalysts as the temperature is reduced below 680’ C and little oxidation of C, compounds then occurs [ 51. Loss of hydrocarbon selectivity appears to be due to the homogeneous and/or surface catalysed reactions of species formed by the reaction of methyl radicals and oxygen molecules. Ethylene is predominantly if not exclusively made by the further reaction of ethane. At high temperatures ( > SOO’C) the conversion can be homogeneous but catalytic contribution seems likely at lower temperature. Various authors have made brief observations on the characteristics of the reaction of ethane in connection with studies focused on methane coupling but there have been very few investigations which concentrate on the oxidative dehydrogenation of ethane over a methane coupling catalyst. Morales and Lunsford [ 71 have recently studied the reaction over Li/MgO and found that it was highly selective with a reaction rate greater than that of methane coupling under the same conditions. They concluded that a considerable homogeneous contribution was possible since with one type of reactor at the highest temperature used ( 2 675 oC ) the conversion in the presence of catalyst was less than that when using the same reactor empty. The catalyst was thought to partially quench the gaseous reaction by removing chain carriers. Martin et al.  have also compared the homogenous and Li/MgO catalysed reactions of methane, ethane and ethylene and, more recently Bernal et al. [ 91 have reported on the oxidative dehydrogenation of ethane over lanthanum oxide. Of course Li/MgO is a rather inactive (although selective) catalyst for methane coupling [ 71. The purpose of the present study was to make a direct comparison of the oxidation of ethane relative to methane over particular rare earth oxides known to be very active for the methane reaction [ 10,111. In this way it was hoped to derive information under conditions where the gas-phase contribution is less significant. To enhance this possibility relatively dilute feed streams have been used since gas-phase contributions to the methane reaction can be made minimal under such conditions [12,13]. The measurements have concentrated on four particular rare earth oxides, in order to establish if those which are selective for methane coupling (e.g. LazO, and SmaOs) are similarly selective for the oxidation of ethane whilst those unselective with methane (e.g. CeO, and PreO,,) are correspondingly poor for ethane. EXPERIMENTAL
All catalyst testing was carried out in a conventional single pass flow system. The reactor comprised a fused alumina tube, 6 mm O.D. and 4 mm I.D., mounted vertically in a tube furnace, and fitted with top and bottom thermowells of 3 mm O.D. The distance between the tips of these thermowells was approxi-
mately 5 mm. The standard test conditions were as follows. The catalyst sample, 100 mg of particle size 0.6 to 1 mm, was supported on a l-mm layer of fused alumina particles placed on top of the lower thermowell. The test stream was down-flow and the standard composition was 13% hydrocarbon, 3.3% oxygen, the remainder helium to a total pressure of 1 atmosphere (101 kPa). The total flow-rate was 110 cm3 (STP)/ min giving a gas-phase residence time of less than 0.2 s in the heated region of the reactor tube. The product stream from the reactor was passed through a ten port Valco gas sampling valve and periodically analysed with a Shimadzu GC8A gas chromatograph fitted with a 3 metre Unibead 1s column. When maintained at 30’ C it provided baseline separation of hydrogen, oxygen, carbon monoxide, methane, ethane, carbon dioxide and ethylene in that order. Analysis of C, and C, hydrocarbons could be achieved by operating at 120°C. The sensitivity of the TCD detector of the chromatograph to the analysed compounds was determined by calibration against a standard gas mixture (CIG Limited) and the relative values agreed with reports in the literature. Water could not be determined directly and was calculated from the hydrogen, C, and C, product distribution after allowance for any C3 and C, products when they were observed. Trial testing showed that the selectivity using methane was noticeably dependent on the particular source of rare earth oxide. The results reported here were obtained using the following high purity materials-La,O, (99.99% ) and CeO, (99.999% ) from Aldrich Chemical Co. (Wisconsin) and, Smz03 (99.999%) and Pr,Oll (99.999% ) from Cerac (Wisconsin). Each oxide was pelleted, crushed and sieved to obtain the 0.6 to 1 mm fraction. After loading into the reactor the samples were heated to 815 ‘C in helium over an 8 h period. Reaction was commenced at 500 to 600” C and programmed in stages up to 815°C and then back down to 400°C with stabilisation and analysis every 50” C. The temperature gradient across the bed, as measured by the difference between two chromel-alumel thermocouples in the upper and lower thermowells, reached 20’ C for total oxygen conversion with the least selective catalyst but was usually less than 5 ‘C. The temperature of the lower thermocouple is reported. The surface areas of each catalyst sample was measured after use, using the BET method with krypton as the adsorbate. RESULTS
Catalyst activity The test conditions, low gas-phase residence time and dilute feed, were chosen so as to minimise contributions from homogeneous reactions. As shown in Figs. 1 and 2 the oxygen conversion in blank runs remained below 5% to 820’ C with methane and to 750°C with ethane. Under all conditions the blank conversions were less than one-tenth that obtained with the least active catalyst
Fig. 1. Oxygen conversion as a function of temperature during the oxidative coupling of methane under standard conditions: (0) Sm,O,, (0) La,O,, (a) CeOx, (0 ) Pr6011, (A) Blank. Fig. 2. Oxygen conversion as a function of temperature during the oxidative dehydrogenation of ethane under standard conditions. Symbols as in Fig. 1. TABLE 1 Comparison of the catalytic activity of rare earth oxides for the oxidation of methane and ethane at 510°C Catalyst
(m’/g) LasO, CeO, Pr8011 Sm,O,
1.2 2.7 6.0 3.7
Rate per gramb
Rate per m2 ’
Rate ( C2H6)
29 6 71 51
100 29 240 190
24 2.3 12 14
86 11 40 51
3.5 4.5 3.4 3.7
139 95 119 143
148 44 127 123
“After use for approximately 16 h. bpol Hydrocarbon converted to all products per gram of catalyst per minute. ‘pm01 Hydrocarbon converted to all products per square metre of catalyst per minute.
( CeO, ). For both reactions the rate of reaction of ethane is considerably greater than that of methane and the order of activity amongst the rare earth oxides is similar, with Pr6011 and Sm20S the most active and CeOz the least. A quantitative comparison is possible only at low temperature. Table 1 compares hydrocarbon reaction rates at 510°C (where conditions are approximately differential-oxygen conversion < 30% for all except the reaction of ethane over Pr,O,,) calculated according to rate [,umol (hydrocarbon) g-l min-‘1 =FX/ W, where F is the input molar flow-rate of hydrocarbon, X its fractional conversion and W the mass of catalyst. The rate of oxidation of ethane relative to that of methane does not vary significantly between the oxides tested and has an average ratio of 3.8. This higher oxidation rate of ethane versus methane presumably reflects the lower CH bond strength in the former (410 kJ/mol
versus 435 kJ/mol) . If this bond strength difference carried over to the transition state then, all else being equal, a rate ratio of fifty would be expected at 510°C. The factor of 3.8 implies a difference of about 9 kJ/mol in the transition state. It was not possible to obtain activation energies from Arrhenius plots with sufficient accuracy to detect such a small difference. As shown in Table 1 the measured values are 134 ? 15 kJ/mol for both reactions over La203, Sm,O, and Pr,O,,. The values for CeO,, particularly for ethane oxidation, are much lower. There was no apparent cause for this but the values were quite reproducible for the high purity CeO, used here. While the main purpose of the work here is to compare reactivities of the two hydrocarbons some comment on activity differences between the rare earth oxides is also warranted. For both reactions the activity spans a factor of twelve on a weight basis in the order CeO, < LazO, < Sm,Os < Pr,O1,. The range is preserved on an area basis but order of activity changes to CeO, < Pr6011 < SmzO, < LazO,. There is rough agreement with the results of Campbell et al.  who reported that CeO, has low activity and that La,O, was of equal or greater activity than SmzO,. However Otsuka et al. [lo] reported that Sm,03 was more than five times as active as La203. At the present time it is unclear if the observed differences are caused by differences in impurity levels or surface or bulk structure. Selectivity
to the major hydrocarbon products
Figs. 3 and 4 show the temperature dependence of selectivity, defined as the moles of reactant converted to a product (ethylene from ethane) or products (ethane plus ethylene from methane) as a percentage of the moles of alkane converted to all products. For the oxidation of ethane the behaviour is similar
Fig. 3. Selectivity to ethane plus ethylene as a function of temperature during methane coupling under standard conditions: ( l ) Sm,O,, (0 ) La,O,, (m) CeOz, (0 ) Pr,O,,. Fig. 4. Selectivity to ethylene as a function of temperature during ethane oxidation under standard conditions. Symbols as in Fig. 3.
for all four oxides. It rises steadily with temperature from < 10% at 500’ C to 60 to 75% at 800°C with La,O, and Sm,O, slightly more selective than CeO, and PrsO1,. The differences are much more pronounced when oxidising methane. The C, selectivity with CeOz and Pr6011 does not exceed 6% whilst that with La20, and Sm,O, peaks at approximately 43% and 38% respectively. Even the latter are fairly low but typical of literature values for the same oxides in undoped form when tested with a relatively low methane/oxygen ratio as here [ 141. Sodium doped oxides tested with high methane/oxygen ratios show selectivities of 60% or more [ lO,ll]. This overall selectively pattern for the two reactions can be explained as follows. The very low selectivity for methane oxidation over CeO, and Pr,O,, arises primarily because these oxides have very high activity for the oxidation of methyl radicals [ 151 presumably because of the existence of variable valence states. The activity of Sm,OB and La,O, for methyl radical oxidation is much less [ 151 thus allowing more time for their dimerisation to ethane. The selectivity increases with temperature since as the rate of production of methyl radicals increases, their concentration in the interparticulate space will increase and so will the probability of pairing (a second order process) relative to the further oxidation on the surface (which is at most first order). A further factor may operate. At the temperatures used, methyl radicals can react with oxygen to form methylperoxy radicals [ 161 which are likely to be oxidised more readily than methyl radicals. As the temperature rises the equilibrium between methyl and methylperoxy radicals moves in favour of methyl radicals [ 161 thus again enhancing the probability of dimerisation relative to the oxidation of methylperoxy radicals. The reverse effect will occur if oxygen pressure is raised since this favours methylperoxy radicals and reduces selectivity. When oxidising ethane the main route to ethylene is from ethyl radicals and oxygen. This can occur [ 171 either directly via an excited peroxy species C,H,+O,+C,H,O;-C,H,+HO,
or by reaction of collisionally deactivated ethylperoxy C2H502-C2H4 + HO,
These processes are quite fast so that radical lifetimes in the C,H,/C,H,O, system will be shorter than for the CH,/CH,O, system. As a consequence the probability of a C2H502 radical reaching the surface in the course of ethane oxidation will be less than that for a CH302 radical during methane coupling. The selectivity should be correspondingly higher and less catalyst dependent in the former case as is observed. The lifetime of either C2H502 or CH,O, radical intermediates affects the selectivity of the oxidation process. Both radicals are more stable at low temperatures, thus increasing the probability of surface reactions. These surface reactions result in non-selective product for-
mation for both C,H,O, and CH302, and thus low selectivity is observed for both ethane and methane oxidation at low ( < 650°C) reaction temperatures. Distribution between carbon oxides and between hydrogen and water As shown in Figs. 5 and 6, the ratio of carbon monoxide to total carbon oxides was very low for the oxidation of both methane and ethane over Ce,O, and Pr6011, the two catalysts with high oxidising power for methyl radicals. By contrast carbon monoxide comprised up to half the carbon oxides when using La20, and Sm,O,. The fraction passed through a maximum near 600’ C for both reactions. To some extent the correlation with catalyst oxidation power extended to the ratio of hydrogen to hydrogen plus water. As shown in Figs. 7 and 8 this was significantly lower for PrsO1, than for La,O, and Sm,O, but CeO, differs from the latter pair to a smaller extent and for the oxidation of ethane alone. In the case of CeO, and PrsOll it seems very unlikely that the water-gas shift reaction
plays a significant role in controlling the overall distribution between carbon monoxide, carbon dioxide and hydrogen, and water since at 1000 K Q was 55 for CeO, and 4.7 for Pr6011, both far above the equilibrium value of 1.4. At higher temperatures Q moved towards the equilibrium value but with the relative concentrations of carbon monoxide and hydrogen both increasing which
7w , ‘C
Fig. 5. Variation of carbon monoxide, as a fraction of total carbon oxides, with temperature during methane coupling under standard conditions. Symbols as in Fig. 3. Fig. 6. Variation of carbon monoxide, as a fraction of total carbon oxides, with temperature during ethane oxidation under standard conditions. Symbols as in Fig. 3.
; x I
a 0.0 5w
Fig. 7. Variation of hydrogen, as a fraction of hydrogen plus water, with temperature during methane coupling under standard conditions. Symbols as in Fig. 3. Fig. 8. Variation of hydrogen as a fraction of hydrogen plus water, with temperature during ethane oxidation under standard conditions. Symbols as in Fig. 3. is inconsistent with a shift of the above reaction to the left. The situation is not so clearcut with La,O, and Smz03. At all temperatures Q was within a factor of two of the equilibrium value for the water-gas shift reaction and some involvement of that reaction is possible.
Other hydrocarbon products Small amounts of C, and C, hydrocarbons were detected when reacting methane and ethane but only at high temperatures. With methane the selectivity to propene reached a maximum of 1.3% when using Sm,O, and 2.0% with La203. It can be attributed to the reaction of methyl radicals with ethylene [ 4,5]. Propene was not observable when reacting ethane but at 800’ C the selectivity to butenes reached 5% reflecting some dimerisation of ethyl radicals. Methane was also detected as a minor product when reacting ethane and its yield showed a significant dependence on the catalyst as shown in Fig. 9. All oxides produced slightly more methane above 700°C which probably reflects some decomposition of ethane in the gas phase. C,H,+CH,+ CH,; CH3 + C2Hs-+ CH4 + C,H,. The oxide pair which are selective for methane coupling, LazO, and Sm,O,, also exhibited small maxima for methane production from ethane near 600°C. Less is produced using CeOz and Pr6011 under the same conditions. The formation of methane requires breakage of a C-C bond. Given the catalyst specificity of the low-temperature methane yield, and the absence of a sufficiently fast homogeneous route for cleavage of ethane, ethyl or ethylperoxy under those conditions, a heterogeneous pathway seems likely. One possibility would be that the initial dissociation of ethane leads to a bound alkoxy species CzHG+ 20 (s) +CzH5 (s ) + OH (s), which reacts further according to C,H,O (a) + 0 (s) -+ CH, + HCO, (s), where HCO, represents a formate
1 0.04 3
Fig. 9. Selectivity to methane as a function of temperature during the oxidation of ethane under standard conditions. Symbols as in Fig. 3.
species of the type well known to form on oxide surfaces and to react further to carbon monoxide and hydrogen. The alkoxy species could also be formed by adsorption of an ethyl or ethylperoxy species. Alternatively the alkoxy species could be oxidised to acetate groups which are known to yield methane [ 18,191 perhaps by dissociation to form a methyl species [ 191. The absence of methane at the lower temperature using CeO, and Pr60,, would imply that they act on the alkoxy or acetate species by C-H rather than C-C scission. That would be in line with their greater oxidising power due to the presence of variable valence states. CONCLUSIONS 1. Under the conditions used here the oxidation of ethane is almost four times as fast as methane coupling for each of the four high purity rare earth oxides tested. Pr,O,, is the most active on a weight basis and La,O, on an area basis. CeOz is the least active on both counts. 2. The selectivity of the ethane to ethylene conversion is much higher than the selectivity of the methane to ethane plus ethylene reaction and is less dependent on the properties of the catalyst. 3. The higher selectivity for ethane oxidation is attributable to the shorter lifetime of gas phase alkyl and alkylperoxy species in that system. The lower selectivity of CeOz and PrGOll, particularly for methane oxidation, can be attributed to their high activity for the oxidation of methyl and ethyl radicals. 4. The strong oxidising power of CeO, and Pr6011 also manifests itself in very low ratios of carbon monoxide to carbon dioxide for both reactions. The hydrogen-to-water ratio is also somewhat lower for these oxides. The amounts of carbon dioxide and hydrogen are much in excess of that expected if the water-gas shift were at equilibrium. 5. Increasing.amounts of methane are made during ethane oxidation over all four catalysts for reaction temperatures above 700’ C. In addition, SmzO, and La,O, show a small peak in methane production at 600’ C and this may be due to carbon-carbon bond cleavage in adsorbed ethoxy or acetate species.
This work was supported by a grant from the Australian Research Council.
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
K.D. Campbell, E. Morales and J.H. Lunsford, J. Am. Chem. Sot., 109 (1987) 7900. K.D. Campbell, H. Zhang and J.H. Lunsford, J. Phys. Chem., 92 (1988) 750. P.F. Nelson, CA. Lukey and N.W. Cant, J. Phys. Chem., 92 (1988) 6176. CA. Mims, R.B. Hale, K.D. Rose and G.R. Meyers, Catal. Lett., 2 (1989) 361. P.F. Nelson and N.W. Cant, J. Phys. Chem., 94 (1990) 3756. A. Ekstrom and J.A. Lapszewicz, J. Phys. Chem., 93 (1989) 5230. E. Morales and J.H. Lunsford, J. Catal., 118 (1989) 255. G.A. Martin, A. Bates, V. Ducarme and C. Mirodatos, Appl. Catal., 47 (1989) 287. S. Bernal, G.A. Martin, P. Moral and V. Perrichon, Catal. Lett., 6 (1990) 231. K. Otsuka, K. Jinno and A. Morikawa, J. Catal., 100 (1986) 353. J.M. Deboy and R.F. Hicks, Ind. Eng. Chem. Res., 27 (1988) 1577. M. Hatano, P.G. Hinson, K.S. Vines and J.H. Lunsford, J. Catal., 124 (1990) 557. Z. Kalenik and E.E. Wolf, J. Catal., 124 (1990) 566. A. Kiennemann, R. Kieffer, A. Kaddouri, P. Poix and J.L. Rehspringer, Catal. Today, 6 (1990) 409. Y. Tong, M.P. Rosynek and J.H. Lunsford, J. Phys. Chem., 93 (1989) 2896. I.R. Slagle and D. Gutman, J. Am. Chem. Sot., 107 (1985) 5342. A.F. Wagner, I.R. Slagle, D. Sarzynski andD. Gutman, J. Phys. Chem., (1990) 1853. K. Aika, M. Tajima, M. Isobe and T. Onishi, Proc. Eighth International Congress on Catalysis, Berlin, Vol. 3, Verlag Chemie, Weinheim, 1984, p. 335. K.S. Kim and M.A. Barteau, Langmuir, 4 (1988) 945.