J . Phys. Chem. 1984, 88, 2426
2426
COMMENTS Isotopic Exchange In the Ethanol Synthesis from CO
+ H,.
Comments on Papers by A. Takeuchi and J. R. Katzer Sir: Takeuchi and Katzer' provided convincing evidence that the methanol synthesis with supported Rh catalysts takes place without dissociation of the CO molecule into its atomic components. A mixture of I3Cl6Oand 12C180(molar ratio = 1:l) was hydrogenated, and the resulting methanol had an isotopic distribution very close to that of the CO feed. In a second paper,2 the same authors investigated the ethanol formed under very similar experimental conditions except for higher conversion (48.2 and 98.7%2 vs. 17.8 and 27.8%'). The isotopic distribution of unreacted CO did not change drastically. The isotopic distribution of the ethanol, however, was close to that expected for a statistical mixing of 12C, 13C, l60,and '*C, Le., as if the CO molecules had been totally dissociated before reacting to form ethanol. In order to reconcile these findings with their evidence for the nondissociative methanol synthesis, the authors suggested a carbene-ketene-oxirene mechanism for the formation of ethanol which would permit scrambling of C and 0 without having to assume dissociation of the second CO molecule prior to generating the alcoholic group, and without altering too much the isotopic distribution of CO in the gas phase. The purpose of these comments is not to discuss that mechanism step by step but to suggest an alternative pathway for the oxygen scrambling at the Cz level, and also to interpret the small, but significant changes in the isotopic distribution of CO in the gas phase. We have to consider that the reactions were carried out in a batch reactor, at very long residence times (25 h for 17.8% conversion;' no time given for 48 and 98% conversion'). Moreover, it was reported' that 93.9% of the reacted C O was converted to hydrocarbons and water, and 6.1% to alcohols; the molar ratio ethanol/methanol was 2.66. At the apparently very long reaction times, the presence of water which can take part in a number of equilibria has to be considered. In particular, ether formation cannot be excluded. Thus, for instance, ethers have been observed during the hydrocarbon synthesis from methanol on zeolite^.^ Dombek and Warren4 found traces of dimethyl and methyl ethyl ether during the synthesis of ethanol with homogeneous ruthenium catalysts. Wilkinson et aL5 reported the formation of dimethyl and diethyl ether during (1) A. Takeuchi and J. R. Katzer, J . Phys. Chem., 85, 937 71981). (2) A. Takeuchi and J. R. Katzer, J . Phys. Chem., 86, 2438 (1982). (3) See, e.g., G. Perot, F. X.Cormerais, and M. Guisnet, J . Mol. Catal., 17, 255 (1982). (4) B, K. Warren and B. D. Dombek, J. Catal., 79, 334 (1983). ( 5 ) R. J. Daroda, J. R. Blackborow, and G. Wilkinson, J . Chem. Soc., Chem. Commun., 1098 (1980).
0022-3654/84/208S-2426$01,50/0
the homogeneous hydrogenation of C O with Fe catalysts. It is evident that the equilibrium 2ROH ROR + H2O (1) will eventually lead to a complete scrambling of the isotopes. The same is to be expected from equilibria involving methyl and ethyl formates, which have been detected in most of the heterogeneous and homogeneous ethanol producing system^:^,^ ROC(=O)H + H20 F? ROH HC(=O)OH (2)
+
Probably acetal equilibrium is also possible and may additionally contribute to the scrambling of the isotopes since acetaldehyde is one of the major byproducts found by Ichikawa7 during the ethanol synthesis with the same catalyst (Rh/TiO,) as used by Takeuchi and Katzer: CH3CH(OCHZCH3)2 + H2O 2CH3CH20H CH3CHO
+
(3) Similar acetal equilibria have also been observed frequently in homogeneous systems.' If we assume that equilibria 1-3 are responsible for isotope scrambling, the same effects should be observable for methanol, at long reaction times. In fact, the trend can be seen in Table 111, ref 1 (compare the data for 17.8 and 27.8% conversion); unfortunately, no methanol data are available for the higher conversions investigated in ref 2. Finally, the water-gas shift equilibrium CO + H2O F? C02 + H2 although shifted to the left under the experimental conditions (H2/C0 = 22), can easily account for small but significant isotope exchange between 12C'80and 13C160in the gas phase. Since the ratio H2'60:H2180 N 1, similar increases of the concentrations of l2CI6Oand 13C180 should be expected and were actually found (Table 12). Summarizing, the results of ref 2 do not necessarily invalidate the reaction mechanism for the Fischer-Tropsch synthesis (hydrocarbons and alcohols) involving CO insertion into alkyl-metal bonds as the growth r e a ~ t i o n . ~ Registry No. CO, 630-08-0; CH,CH20H, 64-17-5. (6) See, e.g., ref 4 and references therein. (7) M. Ichikawa, Bull. Chem. SOC.Jpn., 51, 2273 (1978). (8) D. R. Fahey, J. Am. Chem. Soc., 103, 136 (1981). (9) G. Henrici-Olive and S. Olivt, Angew. Chem., 88, 144 (1976); Angew, Chem., In?. Edit. Engl. 15, 136 (1976).
Chemistry Department University of California at San Diego La Jolla, California 92093 Received: July 19, 1983
0 1984 American Chemical Society
G . Henrici-Olive S . Olive*