Dynamic solvent effects on activated chemical reactions. I. Classical effects of reaction path curvature

Journal Article
The Journal of Chemical Physics, vol. 97, iss. 12, pp. 9116-9137, 1992
Authors
Gregory K. Schenter, Robin P. McRae, Bruce C. Garrett
Abstract
In gas phase reactions, dynamical recrossings across a phase space dividing surface induced by nonlinear reaction path curvature coupling leads to the breakdown of the fundamental dynamical approximation of classical transition state theory (TST). In the following study, we examine the nature of this breakdown for chemical reaction dynamics occurring in solution. As a model system, we consider the collinear A+BC reaction where reaction path curvature increases as the mass of B becomes small compared to the mass of A and C. We use a London–Eyring–Polanyi–Sato (LEPS) potential to describe the solute interaction and model the influence of the solvent by using a generalized Langevin equation that is further represented by a system of coupled harmonic oscillators. Exact classical rate constants are compared to those obtained from conventional TST and canonical variational transition state theory (CVT) as a function of solvent friction coupling. A harmonic TST analysis at the saddle point of the full system (solute plus solvent) with an optimum dividing surface containing both solute and solvent degrees of freedom returns the Grote–Hynes expression for the rate. For the case of no solvent coupling, both TST and CVT are identical and fail to account for the dynamical recrossings induced by reaction path curvature. At intermediate couplings, CVT provides an improvement to the TST estimate and agrees with dynamical simulation results. All estimates of the rate constant approach each other asymptotically at large couplings. The results are interpreted in terms of recrossings in the extended system (solute and solvent coordinates).
English