Molecular Dynamics and Langevin Simulations of One-Dimensional Interfaces.

Abstract

By making use of the techniques of molecular dynamics simulation and Langevin simulation, we study one-dimensional interfaces which separate two distinct fluids. In particular, we address the potential capabilities of the two techniques for simulating the canonical ensemble averages of one-dimensional interfaces. In the molecular dynamics simulations, a microcanonical ensemble average is simulated by introducing the canonical momentums of the molecules which make up the interface. In the Langevin simulations, we solve a discretized Langevin equation of the molecules, which are considered to be Brownian particles, by a Euler algorithm. We find that the two techniques and the Monte Carlo give the same results for internal energy and specific heat, and that there are systematic discrepancies between the molecular dynamics and the MC for the average square size and fractal dimensions. Our conclusion is that the Langevin technique can successfully simulate the canonical ensemble average of one-dimensional interfaces ; however, the molecular dynamics in this study cannot. We also examine a hybrid technique of molecular dynamics and Langevin simulations, and find that it yields the same results that the MC and Langevin simulations yield.