Revisiting the minimum image locus method for calculating the radial distribution functions

Abstract

The minimum image locus method first introduced by Theodorou and Suter (1985) allows us to extend the effective ranges of radial distribution functions obtained from molecular simulations at no extra cost. In this paper, we present a few extensions of the minimum image locus method to improve the accuracy and applicability of the original formulation, followed by performance evaluation with sufficient statistical accuracy to provide guidelines for use in real-world problems. Our results on liquid water and molten lithium fluoride indicate that the effective ranges of radial distribution functions can be safely increased by ≃ 25% for a simple cubic lattice, while the gain is much lower for body centered cubic and face centered cubic lattices. Therefore, as far as the radial distribution functions are concerned, the use of a simple cubic lattice is recommended for modeling homogeneously disordered systems.