Generic parameters of trajectory-extending kinetic Monte Carlo for calculating diffusion coefficients

WJ Tien and CC Chiu, AIP ADVANCES, 8, 065311 (2018).

DOI: 10.1063/1.5035553

One of the challenging applications of molecular dynamics (MD) simulations is to determine the dynamic properties such as the diffusion coefficient of the molecule of interest, particularly slow dynamic systems such as hydrogels and polymer melts. Recently, Neyertz et al. proposed a trajectory-extending kinetic Monte Carlo (TEKMC) algorithm combining both MD and kinetic Monte Carlo to probe the penetrant diffusion within the glassy polymer systems (S. Neyertz and D. Brown, Macromolecules 43, 9210, 2010). Yet, the original TEKMC relies on the manual adjustments of the key parameters of the sampling interval tau and the discretizing grid size r(grid), which limits its applicability to systems with unknown kinetic properties. Here, we reviewed the theoretical background of kinetic Monte Carlo to establish the generic criteria for selecting TEKMC parameters. Also, we modified and expanded the TEKMC algorithm for bulk fluid systems. The modified TEKMC algorithm were applied to systems with various kinetic properties, including Lennard Jones liquid, bulk water, Li+ liquid electrolyte, and Li+ polymer electrolyte. The diffusion coefficients obtained from the modified TEKMC and the generic parameter selections were promising and robust compared with the conventional MD results. With the proposed TEKMC approach, one can extend the MD trajectories to unambiguously characterize the diffusion behavior in the long-time diffusive regime. (C) 2018 Author(s).

Return to Publications page