Selectivity of a Heterogeneous Rhodium Catalyst for theCarbonylation of Monohydric AlcoholsBY BJARNE CHRISTENSEN AND MICHAEL S. SCURRELL*Instituttet for Kemiindustri, Technical University of Denmark,2800 Lyngby, DenmarkReceived 22nd April, 1977The carbonylation activity of a heterogeneous rhodium-zeolite catalyst has been examined forthe reactions of methanol, ethanol and propan-2-01. A marked contrast in the behaviour of thesethree alcohols is seen. Selectivity for the carbonylation of methanol is high (>go%) at all tempera-tures, whereas the sole reaction product with propan-2-01 is propene. Ethanol shows intermediatebehaviour, exhibiting high selectivity (approaching 100 %) for carbonylation at low temperatures, butvery poor selectivity at higher temperatures.The results are consistent with the relative ease withwhich dehydration of the reactants occurs on polar catalysts.Recently tremendous interest has arisen concerning the development of hetero-genised forms of rhodium based catalysts for the carbonylation of methanol toacetic acid.The high activity and selectivity of homogeneous catalysts containing rhodium forthis reaction is well known 2-4 and has had a considerable impact on the large-scalemanufacture of acetic acid.4 The catalysts are apparently useful in carbonylatinghigher alcohol^,^ although no detailed studies of this aspect have so far appeared.Catalytically active rhodium complexes have been supported on materials such ascarbon,6* ' alumina,8 modified poly(styrene-divinylbenzene) and molecular sievezeolites lo and the resulting materials are successful heterogeneous versions of thehomogeneous catalysts.Results concern methanol exclusively and there seem to beno readily available reports dealing with the behaviour of these heterogenised cata-lysts with ethanol and other alcohols.Oxide supports such as alumina and zeolites are known to efficiently dehydratealcohols, with ethanol being in general more reactive than methanol and with second-ary alcohols being dehydrated faster than primary alcohols. Methanol shows onlya very slight tendency to produce dimethyl ether as a side product during carbonyla-t i ~ n , ~ ' ~ but it was felt that the effect might be more dramatic for alcohols of highermolecular weight.Accordingly we undertook to investigate the carbonylation ofmethanol, ethanol and propan-2-01 on one supported rhodium catalyst. A zeolitebased catalyst, similar in composition to that employed by other workers lo wasselected. Of the heterogenised rhodium catalysts so far investigated the zeolitebased material has a particularly high activity and selectivity for methanol carbonyla-tion.EXPERIMENTALThe catalyst was prepared using a Linde molecular sieve zeolite Type 13X and rhodiumtrichloride. The zeolite was immersed in an aqueous solution of the rhodium salt and heldat 80°C for 15 h with constant stirring. The solid material after filtration, washing anddrying was compressed into tablets (diameter 13 mm) using a static applied pressure of203B .CHRISTENSEN AND M. S. SCURRELL 2037383 MNm-Z. The tablets were crushed and the fraction of powder with particle size60-100 mesh (140-25Opm) isolated for use by sieving. The rhodium content of the finalcatalyst was - 1 % by weight.Rhodium-oncarbon (Engelhard Industries) contained 5 % rhodium by weight. Particlesize was > 170 mesh (< 88 pm).The rhodium-zeolite catalyst ( - 0.2 g) was placed in a Pyrex glass reactor and activated byheating in a stream of dry air at 723 K for 15 h. Carbon monoxide was passed throughthermostatted saturators containing liquid alcohol and liquid alkyl iodide. The separatestreams were joined and passed over the catalyst at a total pressure of 1 atmosphere. Thetotal flow rate of carbon monoxide was 80 an3 min-l and the molar ratios alochol : carbonmonoxide and alcohol : alkyl iodide varied by adjustment of the relative flow rates of gasthrough the saturators and by the temperature at which saturation took place.Alcohol :CO molar ratios were typically -0.05.Reactant and product streams were analysed using gas chromatographic separation witha column of Chromosorb 101 (Perkin Elmer). Column length was from 1.5 to 5.5 m andthe temperature in the range 418-433 K, depending upon the components present in thesamples. Flame ionization detection (Perkin Elmer F1 1) was utilised with nitrogen carriergas.Further experimental details are given in table 1.TABLE RA RATES AND SELECTIVITIES FOR REACTIONS STUDIEDmolar ratio rate of carbonylationl 'selectivity/ %reactants RI/ROH temperature/K mol ester (gRh)-1 h-10.047 513 0.97 > 90MeOH+ Me1 0.1430.0755134731.010.10>w>900.075 433 0.03 > 900.147 5230.100 473EtOH+ EtI 0.235 4730.282 4330.250 383Pr-2-OH + Pr-2-1 0.1300.130473433EtOH+ EtI 0.360 503(Rh-on-carbon)0.130.070.170.030.02000.036506085>990013] x 100 RCOOR+RCOOH[2R20 + RCOOR + RCOOH + alkene a calculated as molUnder the experimental conditions employed conversion of the product acid to thecorresponding ester took place readily and free acid was not detected.In order to minimisethe number of possible products obtained, it was ensured that the identity of the alkylgroups of the reactant alcohol and alkyl iodide were the same.RESULTS AND DISCUSSIONAll results are summarized in table 1.Methanol was converted to methyl acetateat temperatures in the range 433-513 K. The selectivity for carbonylation wastypically >90 %, even at the highest reaction temperatures. The side product wasdimethyl ether. A slight ( N 10 %) increase in the rate of carbonylation resulted froman increase in the methyl iodide : methanol molar ratio in the reactant stream from0.047 to 0.143. LittIe or no loss of the iodide promoter was seen. In the absenc2038 SELECTIVITY OF A RHODIUM CATALYSTof methyl iodide in the reactant mixture the rate of carbonylation fell rapidly but thedehydration reaction was virtually unaffected.Ethanol carbonylation occurred at 383-523 K and selectivity was determined forthe most part by the choice of temperature.At 523 K the overall conversion ofethanol was slightly greater than for methanol at 513 K but the selectivity was verylow at - 6 %. At lower temperatures the selectivity was higher, reaching a value of99 % at 383 K. Low selectivity was associated with the formation of both ethyleneand diethyl ether, alkene production predominating at temperatures above - 470 K.No significant consumption of ethyl iodide occurred.The carbonylation of propan-2-01 was attempted at temperatures of 433 and 473 K,but the sole product in each case was propene, conversion to the alkene being com-plete at the higher temperature under the experimental conditions employed.The results indicate the striking differences in behaviour of the three alcohols,selectivity for methanol and propan-2-01 approaching 100 and 0 % respectively, andfor ethanol ranging approximately between these limits depending largely on reactiontemperature and to a small extent on the molar ratio of ethyl iodide : ethanol.These observations can be reconciled with the relative ease of dehydration of thealcohols commonly found for reaction on polar cata1ysts.lThe dehydration sites in the rhodium zeolite might be expected to be provided bythe carrier and to be physically distinct from the carbonylation centres, but theresults of a series of experiments using rhodium supported on carbon gave rise todoubts about this conclusion.Thus, at 503 K, ethanol carbonylation took place onthe latter catalyst, but the selectivity was only - 13 %, due to the extensive formationof ethylene.Since high dehydration activity is not normally associated with carbonitself, it may well be that the carbonylation centres provide the catalytic activity forthe elimination reaction. This may also be true at least to some degree for the zeolitebased catalyst, although further experimentation is required for clarification.Since molecular sieve zeolites are known to catalyse the formation of alkenes fromhalogenoalkanes by elimination at temperatures close to those used in this work,l2' l 3it is possible that some ethylene and propene are produced from the respective alkyliodide and not from the alcohol alone. Overall conversion of the iodide would benegligible if the hydrogen halide so produced reacted sufficiently rapidly with thealcohol present, e.g.for ethyl iodideC2H51 -+ C2H,+HIfollowed byHI + CZHSOH + CzHSI + H20.Reaction (2) is also likely to take place readily in the presence of a zeolite catalyst.12A comparison with the data provided by other workers on rhodium zeolite car-bonylation catalysts lo shows that the rate found by us for methanol conversion at513 K [ N 1 mol acetate (gRh)-l h-'1 is higher than theirs [ -0.14 mol acetate (gRh)-'h-l] for reaction under comparable conditions, and is also higher than the ratesreported for other heterogenised rhodium catalysts.'This work confirms that catalysts produced from rhodium trichloride and a zeolitepossess high activity and selectivity for the carbonylation of methanol, but demon-strates that this behaviour is not encountered in the attempts to carbonylate ethanoland propan-2-01.M. S .Scurrell, Platinum Metals Rev., 1977, 21, 92.F. E. Paulik and J. F. Roth, Chem. Comm., 1968, 1578.J. Hjortkjaer and V. W. Jensen, Ind. and Eng. Chem. (Product Res. and Development), 1976,15,46€3. CHRISTENSEN AND M. S . SCURRELL 2039J. F. Roth, J. H. Craddock, A. Hershman and F. E. Paulik, Chem. Tech., 1971, 600.F. E.. Paulik, A. Hershman, J. F. Roth, J. H. Craddock, W. R. Knox and R. G. Schultz,South African Patent 682174 (1968) to Monsanto Co.R. G. Schultz and P. D. Montgomery, Amer. Chem. Soc., Div. Petrol. Chern., Preprints,1972,17, B 13.A. Kryzwicki and G. Pannetier, Bull. SOC. chim. France, 1975, 1093.M. S. Jarrell and B. C. Gates, J. Catalysis, 1975, 40, 225.Nauk. S.S.S.R., Ser. khim., 1976, 582.ti R. G. Schultz and P. D. Montgomery, J. Catalysis, 1969, 13, 105.lo B. K. Nefedov, N. S. Sergeeva, T. V. Zheva, E. M. Shutkina and Ya. T. Eidus, Zzuest. Akad.l 1 H. Noller and W. Kladnig, CataZysis Reu. (-Sci. Eng.), 1976, 13, 149.l2 P. B. Venuto, E. N. Givens, L. A. Hamilton and P. S. Landis, J. Catalysis, 1966,6,253.l3 W. Kladnig and H. Noller, J. Catalysis, 1973, 29, 385.(PAPER 71676