J. Chem. SOC., Faraday Trans. I, 1984,80, 1645-1650 Mechanistic Study of the Catalytic Hydrogenolysis of Ethane BY SANDOR KRISTYAN Institute of Isotopes of the Hungarian Academy of Sciences, P.O. Box 77, H-1525 Budapest, Hungary AND JANOS SZAMOSI* Department of Chemistry, The University of Texas at Arlington, P.O. Box 19065, Arlington, Texas 76019, U.S.A. Received 17th October, 1983 A mechanistic investigation of the hydrogenolysis of ethane is presented. The chemisorptions of ethane and hydrogen produce a common surface species, adsorbed hydrogen, and the coverages of the two-carbon-atom surface compound and the adsorbed hydrogen are inter- dependent through the partial pressures of ethane and hydrogen. The kinetically slow rupture of the C-C bond takes place in an interaction with a free site, adsorbed hydrogen or molecular hydrogen.Based on a theoretical analysis and previous experimental results, our conclusion is that molecular hydrogen is the most probable agent in the bond splitting. The reaction equation catalyst C2H6 +H, -2CH, illustrates the catalytic hydrogenolysis of ethane to methane over a metal catalyst (Ni, Pd or Co). Experirnent~~-~ have indicated that the reaction rate is a function of the partial pressures of hydrogen and ethane. (It is possible to vary the partial pressures individually by adding an inert gas such as He to the gas mixture.) The rate is proportional to the partial pressure of ethane, up to a limit; beyond that the rate can be enhanced further only by altering the partial pressure of hydrogen.When the dependence of the reaction rate on the partial pressure of hydrogen is investigated, it is found that the rate goes through a maximum: the reaction is very slow if too much or too little hydrogen is present. Even though most experiments have had the same outcome, none of the proposed mechanisms can fully account for the observed phenomena. Sinfelt and Taylor1* assume that the initial chemisorption of ethane gives an adsorbed ethyl radical which is not in equilibrium with ethane gas. The adsorbed C,H,* then undergoes further dehydrogenation to C2HZ, such that an equilibrium is maintained between C,H,* and C,HZ (rn is defined later). The rate is given by r = kpE/(l + bp,), where pE and pH are the partial pressures of ethane and hydrogen, respectively, and b is a positive number arising from the equilibrium constant and rate constants.The function r does not account for the facts that the reaction is slow at low H, pressure and that the rate cannot be enhanced limitlessly by simply increasing the partial pressure of ethane. In Martin’P8 blocked coadsorption mechanism, adsorption of ethane is the rate-determining step. This can be ruled out, since experimental results clearly indicate1* 5 9 that the chemisorption step reaches equilibrium almost instantly. 16451646 CATALYTIC HYDROGENOLYSIS OF ETHANE Most of the assumptions about the reaction mechanism made by Forster and Ottos cannot be disputed; however, it is not necessarily true that C-C bond-splitting occurs via a Langmuir-Hinshelwood mechanism with adsorbed hydrogen.Their proposed rate is inversely proportional to the partial pressure of hydrogen. We have tried to include all possible factors and made our judgement by analysing the obtained rate functions. Both earlier and recent5 experimental results show good agreement with our model. DISCUSSION Hydrogen-deuterium exchange experiments3y have proved that both ethane and hydrogen are adsorbed dissociatively but that the C-C bond is preserved in this step. Furthermore, it is probably a good assumption that only one C atom of the C,HZ surface compound participates in surface bonding.O* lo We must take into account that the hydrocarbon coverage, 8, (of C,H3, and the hydrogen coverage, OH, are interrelated. If the partial pressures of H, and C,H, are varied, the coverages could change dramatically (8, the coverage, is the fraction of all sites occupied by a species). We note that the adsorbed hydrogen comes partially from the chemisorption of ethane, further complicating the mechanism.The proposed mechanism is as follows. The first step is the adsorption of the ethane + hydrogen gaseous mixture. Step I (fast): KE KH C2H6+(7-m)* +C2H;+(6-m)H* H,+2* t 2H* where * indicates surface site and m is the number of hydrogen atoms remaining on the two-carbon-atom surface compound. The two equilibrium constants are given by where 8, and 8, are the coverages of C,Hk and hydrogen, respectively, 8, is the fraction of free sites, and pE and pH are the partial pressures of ethane and hydrogen, respectively. Step I1 (slow): k C,Hk + B + CHG + CHt.The reaction is irreversible; B is either a free site (*), adsorbed hydrogen (H*) or gas-phase molecular hydrogen (H,). The last step is the formation of methane by hydrogenation of the one-carbon-atom surface compounds with adsorbed or molecular hydrogen. Step 111 (fast): CHtS. KRISTYAN AND J. SZAMOSI 1647 Table 1. Comparison between calculated and measured (Ni catalyst, 250 "C) rate values ratelpmol g(catalyst)-' s-l 9.4 4.1 0.30 0.013 3.9 3.48 9.4 1.8 0.13 0.019 2.5 2.83 4.7 3.3 0.32 0.066 2.9 2.45 2.9 4.1 0.28 0.064 2.1 1.79 1.2 1.8 0.34 0.050 1.5 1.18 Assuming that d[O,]/dt = d[O,]/dt = 0, the rate is given by r = k0,B. (3) Since the third step is fast, coverages of the one-carbon surface compounds are negligibly small; the sites are either free or occupied by adsorbed hydrogen or C,Hg : o o + o ~ + e m x 1 .We now introduce the following new variables : Y = KEpE x = (KHPHP G6 = ~ 6 - m G7 = ~ 7 - m . Using eqn (l), (2) and (4) the coverages of interest are: ' 0 = G6/(Y + G6 + G 7 ) Om = Y / ( Y + G6 + G7) O H = G7/(y+ G6 + G7)' Depending on the identity of B, the possible rate equations are as follows: if B = B0; if B = 8, ; r = kyG,/(y+G6+G7)2 Y = kyG7/( y + G, + Gp)2 if B = p H ; r = k y p ~ / ( y + G , + G , ) . RESULTS Analyses of eqn (8)-( 10) show that ifp. is kept constant, r goes through a maximum in all three cases if rn = 0-5; however, if pE changes, the rates given by eqn (8) and (9) do not show the experienced proportionality between the rate and the partial pressure of ethane. At the same time, eqn (10) clearly indicates that the rate grows with pE and that the increase is not limitless.In a recent experiment5 with Ni catalyst at 250 "C the partial pressures of hydrogen and ethane were varied between 0.5 and 10 kPa. The following results were obtained. (i) The rate constant is k z 1.56 pmol g(catalyst)-l kPa-l s-l. (ii) The equilibrium constants are KH = 1.3 x dm3 kPa-'. (iii) The rate is never inversely proportional to the partial pressure of ethane, even if only a dm3 kPa-l and KE = 5.8 xv1 - I n Y w - Y m 00 W CATALYTIC HYDROGENOLYSIS OF ETHANE 0.15 r 40 50 60 30 P H /kPa \ 30 * I v1 - h I U w - cd Y m M W I I I I I J 10 20 30 0 40 50 60 -;i 0 I I - 40 50 60 0 10 20 30 Fig. 1. Simulated reaction rate, r, plotted against partial pressure of hydrogen, p H , with p E , the partial pressure of ethane, kept constant at 10, 20 and 30 kPa.(a) Eqn (8), (b) eqn (9) and (4 eqn (10).S. KRISTYAN AND J. SZAMOSI 1649 2.0 0 ' I I J 0 5 10 15 PElkPa Fig. 2. Simulated reaction rate, r, plotted against partial pressure of ethane, p E , with the partial pressure of hydrogen kept constant at 5 kPa. very small amount of hydrogen is present; i.e. the curves never cross each other. (iv) The maxima in plots of rate against pH occur at higher pH values as the ethane pressure grows. (v) Rate measurements with Pd catalyst at 350 "C exhibit essentially the same feature^,^ although the parameters differ: k = 1.18 pmol g(catalyst)-l kPa-l s-l, KH = 0.74 dm3 kPa-l and K , = 0.73 dm3 kPa-l. The results of other experiment+ 3 9 are consistent with the above findings.Using Simplex curve-fitting Kristyanll has confirmed (iii) and (iv) above. In addition, rn has been determined for both catalysts: it is probably 2 in the case of Ni catalyst and 3 in the case of Pd. We have calculated the three possible rates with the parameters given by (i), (ii) and (v) at different pressures and compared them with experimental data.5 Table 1 shows that eqn (10) correlates very well with the measured values; at the same time the rates obtained from eqn (8) and (9) are always substantially lower than the corresponding experimental results. Comparison between the computed and the measured values in the case of Pd catalyst has also proved that only eqn (10) describes the rate satisfactorily. Our opinion is that in the rate-determining step it is the gas-phase (more exactly non-chemisorbed) hydrogen whose interaction with the two-carbon-atom surface compound brings about the rupture of the C-C bond.Beside the good quantitative correlation between the experimental results and eqn (10) the proposed mechanism is supported by a closely related experiment. Galway's research12 on the hydrogenation of nickel carbide has shown an almost linear dependence of the reaction rate upon the partial pressure of hydrogen, indicating that in a system similar to ours the mechanism is of the Rideal type. Fig. 1 (a) [eqn (S)], (b) [eqn (9)] and ( c ) [eqn (lo)] illustrate the three possibilities. The rate curves have been simulated by using k = 1.6 pmol g(catalyst)-l kPa-l s-l, KH = 1.5 x dm3 kPa-l, i.e.the rounded values of the parameters obtained from the experiments with the Ni catalyst. Thep, varies from 0 to 60 kPa, and each figure has three curves with p , taking the values 10, 20 and 30 kPa. Only fig. 1 ( c ) exhibits the characteristics described under (iii) and (iv) above. Fig. 2 shows the rate dependence on p , , using eqn (10) with a constant hydrogen pressure of 5 kPa. It is clear that the rate increases with p,, up to a limit. dm3 kPa-l and K , = 6 x1650 CATALYTIC HYDROGENOLYSIS OF ETHANE CONCLUSION In this mechanistic study we have focused our attention on the determination of the C-C bond-splitting agent, which is the most crucial problem of the whole mechanism. A comparison between the measured and the calculated rate values and an analysis of the possible rate functions indicate that the kinetically slow rupture of the C-C bond takes place in an interaction with molecular hydrogen. We gratefully acknowledge the support of the R. A. Welch Foundation. J. H. Sinfelt and W. F. Taylor, Trans. Faraday Soc., 1968, 64, 3086. J. H. Sinfelt and W. F. Taylor, J. Catal., 1977, 27, 468. L. Guczi, A. Sarkany and P. Tetenyi, J. Chem. Soc., Faraday Trans. 1, 1974, 70, 1971. L. Guczi, B. S. Gudkov and P. Tetenyi, J. Catal., 1972, 24, 187. A. Sarkany, P. Tetenyi and S. Kristyan, to be published. G. A. Martin, C.R. Acad. Sci., Ser. C, 1977, 284,479. ' G. A. Martin and B. Imelik, Surf. Sci., 1974, 42, 157. G. A. Martin, J. Catal., 1974, 60, 345. H. Forster and H. J. Otto, 2. Phys. Chem., N.F., 1980, 120, 223. 1977). lo J. R. Anderson, Chemisorption and Reaction on Metallic Films (Academic Press, London, New York, l1 S. Kristyan, DSc. (Tech.) Dissertation (Technical University of Budapest, 1982). l2 A. K. Galway, J. Catal., 1962, 1, 227. (PAPER 3/ 1844)