A Versatile Catalyst for Reductive Amination by Transfer Hydrogenation

Reductive amination (RA), the coupling of ketones or aldehydes with amines in the presence of a reducing reagent, is one of the most studied and useful reactions in synthetic chemistry. Applications of the reaction are widespread in the pharmaceutical, agrochemical, and chemical industries, materials science, and biotechnology. In both academic research and commercial-scale preparations, RA is effected mainly with stoichiometric boron hydrides and heterogeneous hydrogenation. Apart from generating copious waste, the use of boron hydrides is associated with other problems, such as the high toxicity of NaBH3CN and the inability to aminate aromatic ketones with NaBH(OAc)3, two most widely used hydrides in RA. RA by heterogeneous hydrogenation has long been practiced, but its application is limited by its relatively poor chemoselectivity, for example, reduction of C= O, C=C, and -NO2 over C=N bonds. [1, 2b] In the past decade or so, a small number of homogeneous catalysts and enzymes have been developed, allowing for enantioselective RA. However, the substrate scope remains to be improved. Herein we disclose a class of air-stable cyclometalated imido Ir complexes that catalyze transfer hydrogenative RA with safe, inexpensive formate, providing high chemoselectivity and activity along with wide substrate scope. Transfer hydrogenation has enjoyed a huge success in the reduction of ketones. However, its application in the reduction of imines, the key intermediate in RA, is less developed; examples of transfer hydrogenative RA are even rarer. In our effort to develop an asymmetric transfer hydrogenation system for RA, we initially examined reaction conditions for the transfer hydrogenation of a model imine prepared from acetophenone and p-anisidine, which was not reduced under asymmetric transfer hydrogenation conditions. With a previously prepared [Cp*IrCl(Tsdpen-H)] catalyst [Cp* = pentamethylcyclopentadienyl; Tsdpen = (1R,2R)-N-(p-toluenesulfonyl)-1,2-diphenylethylenediamine] (0.5 mol%), 30 % conversion to the corresponding amine in 13 % ee was observed by using an azeotropic mixture of formic acid/triethylamine (F/T) in MeOH at 40 8C in 3 h. Surprisingly, when the catalyst was prepared in situ by treating [{Cp*IrCl2}2] with Tsdpen, the conversion rose to 95%, but the resulting amine was racemic. Of yet further interest is that when [{Cp*IrCl2}2] (0.25 mol %) was used without any added ligand, the reaction proceeded to afford the same result as that obtained with the in situ catalyst. These observations prompted us to ponder what the real catalytic species might be. One possibility is a cyclometalated iridium complex with the substrate ketimine acting as ligand through C H activation. This hypothesis was quickly supported by the fact that by simply stirring [{Cp*IrCl2}2] and the imine substrate in MeOH at 40 8C for 1 h, the iridium dimer was converted into a cyclometalated imido complex in greater than 99% conversion. Encouraged by the results above, we then explored cyclometalated Ir complexes of aromatic ketimines for the reduction of an aliphatic ketimine, which itself is difficult to undergo cyclometalation and indeed showed no activity in the [{Cp*IrCl2}2]-catalyzed reduction (Figure 1). Dramatic accel-