JOURNAL OF CATALYSIS 71, 373-380 (1981) CO Hydrogenation Catalyzed by Magnesia-Supported Osmium Derived from Os,(CO),, M. DEEBA,' J. P. SCOTT, R. BARTH, AND B.C. GATES* Center for Catalytic Science and Technology. Department of Chemical Engineering, University of Delaware, Newark. Delawure 1971 I Received March 19, 1981; revised July 2, 1981 A catalyst prepared by the reaction of [Os,(CO),,] with the surface of magnesia exhibited infrared spectra in the carbonyl region suggestive of the formation of the supported cluster [HOs,(CO),,OMg~]. After oxidation in air, the sample exhibited a spectrum indicative of a mononuclear osmium carbonyl complex. The catalyst, initially present in either the original or oxidized form, was active for CO hydrogenation; the products at low conversions were C,-C, pa&ins formed at a rate of about 10e3 molecules/(Os atom s) at 300°C and 31.8 atm with a 4: 1 H, : CO ratio. The infrared spectrum of the used catalyst is suggestive of a trinuclear osmium carbonyl cluster, and we propose that a cluster is the catalytically active species. INTRODUCTION The need for selective catalysts for syn- thesis gas conversion has focused attention on metal clusters (I, 2). Triosmium (I, 3), triruthenium (3), and tetrairidium (I, 3) carbonyl clusters in solution have been implicated in catalysis of CO hydrogenation to give methane, but the reaction rates were so low and the numbers of turnovers so few that the catalytic species could not be de- termined. Rhodium clusters have been identified in solutions catalyzing CO hydro- genation at pressures of hundreds of atmo- spheres to give ethylene glycol and metha- nol, but mononuclear rhodium complexes were also present and the catalytic species remain undetermined (4). There is still no evidence of a structurally defined metal cluster acting as a synthesis gas conversion catalyst. Attempts have been made to stabilize metal-cluster catalysts by bonding them to solid supports, thereby minimizing the in- teractions which occur in solution and often lead to loss of the metal framework integ- i Present address: Air Products and Chemicals, Inc., Allentown, Pa. 18105 * To whom correspondence should be addressed. rity. Most of the reported supports are polymers or silica with pendant electron- donor phosphine ligands (5). Only a few authors have reported the stabilization of metal clusters on typical oxide supports such as silica and alumina; most have found that the metal aggregated to give crystal- lites. Robertson and Webb (6) brought solu- tions of [RuQ(CO)r2] in contact with silica and inferred that at low temperatures (~127°C) the metal framework structures remained intact on the surface and pro- vided catalytic sites for olefin isomeriza- tion; at high temperatures, the clusters decomposed, giving mononuclear Ru complexes. Kuznetsov et ul. (7) prepared silica- and alumina-supported ruthenium using [RK~CO)~~I, V-LRU~(W~~I, or [Ru,C(CO),,] as a starting material. Infra- red spectra showed that all the clusters were converted into mononuclear com- plexes, apparently including the same spe- cies observed by Robertson and Webb (6). The alumina- and silica-supported Ru (upon reduction in H2 at temperatures >4Oo”C) aggregated and formed crystallites (7). Several groups have done experiments with [Rb(CO),,] brought in contact with alumina; the surface chemistry is complex, 373 0021-9Sl7/81/100373-08$0200/O Copyright @ 1981 by Academic Press, Inc. AU rights of reproduction in any form reserved.