Atomistic Molecular Dynamics Simulations of Shock Compressed Quartz

TL;DR

This study uses atomistic molecular dynamics simulations with a modified BKS potential to model shock compression of quartz, achieving good agreement with experimental data up to 25 GPa and revealing phase transition details.

Contribution

It introduces a geometry-optimized simulation scheme with a surface dipole relaxation and a simple modification to the BKS potential for shock wave modeling of quartz.

Findings

Hugoniot matches experimental data up to 25 GPa

Modified BKS potential is suitable for Earth's core conditions

Identifies a eta to lpha phase transition at 6 GPa

Abstract

Atomistic non-equilibrium molecular dynamics (NEMD) simulations of shock wave compression of quartz have been performed using the so-called BKS semi-empirical potential of van Beest, Kramer and van Santen to construct the Hugoniot of quartz. Our scheme mimics the real world experimental set up by using a flyer-plate impactor to initiate the shock wave and is the first shock wave simulation that uses a geom- etry optimised system of a polar slab in a 3-dimensional system employing periodic boundary conditions. Our scheme also includes the relaxation of the surface dipole in the polar quartz slab which is an essential pre-requisite to a stable simulation. The original BKS potential is unsuited to shock wave calculations and so we propose a simple modification. With this modification, we find that our calculated Hugoniot is in good agreement with experimental shock wave data up to 25 GPa, but significantly diverges beyond this point. We conclude that our modified BKS potential is suitable for quartz under representative pressure conditions of the Earth core, but unsuitable for high-pressure shock wave simulations. We also find that the BKS potential incorrectly prefers the {\beta}-quartz phase over the {\alpha}-quartz phase at zero-temperature, and that there is a {\beta} \rightarrow {\alpha} phase-transition at 6 GPa.