Self-diffusion of nanoscale particles with hard and soft sphere models

Abstract

The diffusion of particles in suspension is investigated by a thermostat based on fluctuating hydrodynamics for dynamic simulations of implicit-solvent coarse-grained model which can take into account both hydrodynamic and Brownian effects. Particles with cut-and-shifted Lennard-Jones and Gaussian-core potential are studied. The results show that their diffusion process can be characterized by three regimes: ballistic motion, short-time diffusion and long-time diffusion. We observe that the mean square displacement (MSD) of regime I, ballistic motion, is proportional to t2. For the other two regimes, its MSD is proportional to t with different slopes. Furthermore, we study the diffusion coefficients of spherical particles from MSD at different volume fractions. For the cut-and-shifted Lennard-Jones potential model, we observe the diffusion coefficients decrease monotonously with the increase of volume fraction (0.02–0.3), consistent with the results of the experiment. However, for the Gauss-core potential model, the curve of long-time self-diffusion coefficient as a function of dimensional density (0.001 to 1) appears to be nonmonotonic. It shows that the long-time self-diffusion coefficient decreases monotonically when the dimensional density is below 0.3, and then increases anomalously when the dimensionless density passes through 0.3.