One-to-one reactant–modifier interactions in enantio- and diastereoselective hydrogenation of chiral α-hydroxyketones on Pt(111)

Abstract

Hydrogen-bonding interactions between α-hydroxyketones (i.e., ( R)- and ( S)-1-hydroxy-1-phenyl-2-propanones and ( R)- and ( S)-2-hydroxy-1-phenyl-1-propanones) and protonated cinchonidine in Open(3) conformation relevant to enantioselective hydrogenation over Pt were studied computationally at the B3LYP/TZVP level. The density functional theory (DFT)-optimized structures were reoptimized on a flat Pt(111) surface with molecular mechanics using the condensed phase-optimized molecular potentials for atomistic simulation studies (COMPASS) force field. Two possible interaction modes—the so-called bifurcated and cyclic hydrogen bonded complexes—were studied. In the former, both oxygens of the reactant interact with the proton attached to the modifier's quinuclidine nitrogen; in the latter, one modifier's hydroxyl group also interacts with the substrate's oxygen. Bifurcated complexes were found to be 3–8 kJ mol −1 more stable than the cyclic complexes by DFT calculations. By the force field calculations, only three cyclic complexes relevant to the reaction were found to be stable on the Pt(111) surface, and they were less stable than the bifurcated complexes. Thus, the relevance of the cyclic complexes between α-hydroxyketones and cinchonidine to enantiodifferentiation can be considered negligible. Furthermore, no bifurcated but single hydrogen bonded complexes were found to be stable on the Pt(111) surface for 1-hydroxy-1-phenyl-2-propanones with an sp 3-hybridized carbon next to the phenyl ring. DFT calculations indicated that complexes leading to ( R)-stereoisomers were more thermodynamically stable than the complexes leading to ( S)-stereoisomers. In general, orbital analysis of the reacting C O keto carbonyl orbitals indicated that formation of ( R)-stereoisomers was kinetically preferred as well. However, DFT calculations of isolated complexes cannot qualitatively predict the enantiomeric excess, requiring that the steric restriction of the Pt surface be taken into account. By combining the results of DFT and force field calculations, a reasonable explanation for experimentally observed product distribution was obtained.