Large-Scale Computational Modeling of [Rh(DuPHOS)]+-Catalyzed Hydrogenation of Prochiral Enamides: Reaction Pathways and the Origin of Enantioselection
journal contributionposted on 2000-12-27, 00:00 authored by Steven Feldgus, Clark R. Landis
The potential energy surface for the [Rh((R,R)-Me-DuPHOS)]+-catalyzed asymmetric hydrogenation of a prochiral enamide, α-formamidoacrylonitrile, has been computed using a three-layer hybrid quantum mechanics/molecular mechanics method (ONIOM). The bond-breaking and bond-forming region is modeled using a nonlocal density functional method (B3LYP), whereas HF theory and molecular mechanics (UFF) are used to describe the electronic and steric impact of the outer coordination sphere of the catalyst. Intermediates and transition states were calculated along four isomeric pathways of two diastereomeric manifolds. The starting point for each manifold is a square planar catalyst−enamide complex. Binding of the re enantioface of the enamide to the catalyst generates the more stable, major diastereomer, favored by 3.6 kcal/mol over the minor diastereomer, which has the si face bound. However, the net free energy barrier for the reaction is 4.4 kcal/mol lower for the minor diastereomer than for the major diastereomer, making the minor diastereomer considerably more reactive and reproducing the “anti-lock-and-key” behavior observed experimentally in rhodium-catalyzed asymmetric hydrogenations. The difference in transition-state energies corresponds to an enantiomeric excess of 99.9% (R), within the range of experimental enantioselectivities of [Rh((R,R)-Me-DuPHOS)]+ hydrogenations. The stability and reactivity differences of the two diastereomers are explained using simple steric and electronic arguments. The sequence of elementary steps, as well as the relative orderings of intermediates and transition states, is very similar to that found in our previous work on the achiral model system [Rh(PH3)2(α-formamidoacrylonitrile)]+. We find oxidative addition to be the turnover-limiting step of the catalytic cycle. Our results are consistent with available empirical data for rhodium-catalyzed asymmetric hydrogenations, although more detailed experimental studies are needed on the specific model system studied herein.