Mean-Field Ring Polymer Rates Using a Population Dividing Surface

Abstract

Mean-field Ring Polymer Molecular Dynamics (MF-RPMD) offers a computationally efficient method for the simulation of reaction rates in multi-level systems. Previous work has established that, to model a nonadiabatic state-to-state reaction accurately, the dividing surface must be chosen to explicitly sample kinked ring polymer configurations where at least one bead is in a different electronic state than the others. Building on this, we introduce a population difference coordinate and a kink-constrained dividing surface, and we test the accuracy of the resulting mean-field rate theory on a series of linear vibronic coupling model systems as well as spin-boson models. We demonstrate that this new MF-RPMD rate approach is efficient to implement and quantitatively accurate for models over a wide range of driving forces, coupling strengths, and temperatures.