Evidence that imidazolium-based ionic ligands can be metal(0)/nanocluster catalyst poisons in at least the test case of iridium(0)-catalyzed acetone hydrogenation.

This study begins with the question of whether ionic liquids (ILs), such as 1-butyl-3-methylimidazolium hexafluorophosphate, [bmim][PF6], can be catalyst poisons for transition-metal catalysts rather than a preferred stabilizing media as typically assumed in the literature. The test case of acetone (propanone) hydrogenation is picked for two reasons: (i) acetone hydrogenation is important for its applications in heat pumps, H2 storage schemes, and fuel cells and for the commercial value of the resultant product, propan-2-ol, and (ii) two recent, independent studies have reported putative Ir(0)n nanocluster-catalyzed hydrogenations of acetone beginning in each case with the identical precursor, [{(COD)IrCl}2] (where COD=1,5-cyclooctadiene) (1). A close comparison of the results of those two literature studies and their related, but different, experimental conditions (vide infra) suggests the hypothesis that the IL is actually a catalyst poison. Indeed, the investigations herein (i) find that 1.0 equiv of added IL, [bmim][PF6], completely inhibits the formation of Ir(0)n nanoclusters under conditions 1 in Table 1 in the main text (namely, 3.6 mM precatalyst 1, 22 degrees C, and 2.76 bar H2) and (ii) demonstrate that 0.1 and 1.0 equivs of this same IL, [bmim][PF6], poisons 74 and 90%, respectively, of the acetone hydrogenation activity of premade, previously catalytically active nanoclusters. The above results in turn compelled a reinvestigation of the claim that Ir(0)n nanoclusters are the catalyst in what was reported as a colloidal suspension prepared under conditions 2 in Table 1 of the main text (namely, 52 mM precursor 1, 92 equiv of IL, 75 degrees C, and 4.05 bar of H2). We further (iii) find that the colloidal suspension prepared under conditions 2 is a mixture of unreacted precursor, 1, some nanoclusters, and isolable bulk metal, and we also (iv) find, somewhat surprisingly, in light of the IL-poisoning results found under conditions 1, that the Ir(0) catalyst prepared under conditions 2 is active, precisely as reported, for acetone hydrogenation. This, in turn, further demanded that we go on to (v) investigate the nature of the true catalyst under conditions 2, the results of which we are able to interpret only by the hypothesis that bulk metal is the dominant, true catalyst under conditions 2. Overall, the results provide strong evidence that ILs can be potent inhibitors of metal(0)/nanocluster catalysis, rather than the often-assumed superior solvent for nanocluster catalysis. The results also fortify our recent report that, under conditions where stoichiometrically high amounts of coordinating ligands are present (vs the amount of surface metal atoms), bulk-metal catalysts can actually be superior to nanocluster catalysts of the same metal, a seemingly heretical finding prior to our recent experimental evidence for this (Besson, C.; Finney, E. E.; Finke, R. G. J. Am. Chem. Soc. 2005, 127, 8179; Besson, C.; Finney, E. E.; Finke, R. G. Chem. Mater. 2005, 17, 4925).