A new dynamic Monte Carlo method satisfying n-particle diffusion equation with position-dependent diffusion coefficient, free energy, and intermolecular interactions

Abstract

A dynamic Monte Carlo (MC) method recently proposed by us [Nagai et al., J. Chem. Phys. 156, 154506 (2022)] to describe single-particle diffusion of a molecule in a heterogeneous space with position-dependent diffusion coefficient and free energy is generalized here to n-particle dynamics, where n molecules diffuse in heterogeneous media interacting via their intermolecular potential. Starting from the master equation, we give an algebraic proof that the dynamic MC transition probabilities proposed here produce particle trajectories that satisfy the n-particle diffusion equation with position-dependent diffusion coefficient D0i(ri), free energy F1i(ri), and intermolecular interactions Vij(ri, rj). The MC calculations based on this method are compared to molecular dynamics (MD) calculations for two-dimensional heterogeneous Lennard-Jones test systems, showing excellent agreement of the long-distance global diffusion coefficient between the two cases. Thus, the particle trajectories produced by the present MC transition probabilities satisfy the n-particle diffusion equation, and the diffusion equation well describes the long-distance trajectories produced by the MD calculations. The method is also an extension of the conventional equilibrium Metropolis MC calculation for homogeneous systems with a constant diffusion coefficient to the dynamics in heterogeneous systems with a position-dependent diffusion coefficient and potential. In the present method, interactions and dynamics of the real systems are coarse-grained such that the calculation cost is drastically reduced. This provides an approach for the investigation of particle dynamics in very complex and large systems, where the diffusing length is of sub-micrometer order and the diffusion time is of the order of milliseconds or more.