Phase diagram of silica from computer simulation

TL;DR

This study uses molecular dynamics simulations to map the phase diagram of the BKS silica model, revealing qualitative agreement with real silica but quantitative differences in phase boundary locations.

Contribution

It provides a detailed computational phase diagram of the BKS silica model, assessing its accuracy and limitations in representing real silica phases.

Findings

Quantitative discrepancies in phase boundary locations compared to real silica.

Qualitative reproduction of the topology of the silica phase diagram.

Correlation between phase boundaries and liquid-state anomalies.

Abstract

We evaluate the phase diagram of the ``BKS'' potential [Van Beest, Kramer and van Santen, Phys. Rev. Lett. 64, 1955 (1990)], a model of silica widely used in molecular dynamics (MD) simulations. We conduct MD simulations of the liquid, and three crystals (beta-quartz, coesite and stishovite) over wide ranges of temperature and density, and evaluate the total Gibbs free energy of each phase. The phase boundaries are determined by the intersection of these free energy surfaces. Not unexpectedly for a classical pair potential, our results reveal quantitative discrepancies between the locations of the BKS and real silica phase boundaries. At the same time, we find that the topology of the real phase diagram is reproduced, confirming that the BKS model provides a satisfactory qualitative description of a silica-like material. We also compare the phase boundaries with the locations of liquid-state thermodynamic anomalies identified in previous studies of the BKS model.