Predicting temperature-dependent failure and transformation zones in 2D silica glass through quasistatic Gaussian Phase Packets

TL;DR

This paper extends the quasistatic Gaussian Phase Packets method to finite temperatures to predict failure zones and thermal effects in 2D silica glass, offering an efficient alternative to molecular dynamics simulations.

Contribution

It introduces a temperature-dependent GPP framework for quasistatic analysis of disordered materials, capturing thermal expansion and failure zones accurately.

Findings

Predicts temperature effects on fracture onset

Accurately captures thermal expansion

Identifies zones prone to atomic rearrangements

Abstract

The athermal quasistatic (AQS) method is a powerful technique to study the mechanical behavior of disordered systems. However, its applicability is limited to temperatures near zero, where thermal activation is unlikely. In this work, we extend the AQS method to finite temperatures, based on a formulation that describes atoms as temperature-dependent Gaussian packets (GPPs) in phase space under quasistatic conditions, thus equivalent to minimum free energy conditions. This framework is used to study the effect of temperature on the onset of inelasticity and fracture in amorphous two-dimensional silica glass approaching quasistatic conditions under uniaxial tensile loading. An important characteristic of this formulation is the directional dependence of the variance of each Gaussian packet in configuration space, making this formulation an inexpensive and accurate predictor of zones prone to atomic-scale rearrangements, both in the undeformed state and (with increasing accuracy) as the deformation progresses. This method is also shown to accurately capture the thermal expansion of the disordered material. Furthermore, combining the GPP description with Metropolis sampling predicts the effect of temperature on the onset of fracture of the material, which is validated through MD simulations at strain rates as low as $10^{4}$s$^{-1}$. The presented framework therefore provides a valuable technique for studying the nonlinear mechanics of disordered materials at finite temperature and for predicting local rearrangement zones in disordered solids efficiently without the need for expensive MD simulations.