Elastic consequences of a single plastic event: towards a realistic account of structural disorder and shear wave propagation in models of flowing amorphous solids

TL;DR

This paper introduces a simple finite-element method to model shear transformations in amorphous solids, accurately capturing elastic heterogeneity and shear wave propagation, validated against molecular dynamics simulations.

Contribution

The authors develop and validate a finite-element approach that incorporates elastic heterogeneity and inertial effects for modeling shear transformations in disordered solids.

Findings

FE simulations match MD results for displacement fields

Elastic heterogeneity influences fluctuation magnitudes

Elastic anisotropy is not critical at mesoscale

Abstract

Shear transformations (i.e., localised rearrangements of particles resulting in the shear deformation of a small region of the sample) are the building blocks of mesoscale models for the flow of disordered solids. In order to compute the time-dependent response of the solid material to such a shear transformation, with a proper account of elastic heterogeneity and shear wave propagation, we propose and implement a very simple Finite-Element (FE) -based method. Molecular Dynamics (MD) simulations of a binary Lennard-Jones glass are used as a benchmark for comparison, and information about the microscopic viscosity and the local elastic constants is directly extracted from the MD system and used as input in FE. We find very good agreement between FE and MD regarding the temporal evolution of the disorder-averaged displacement field induced by a shear transformation, which turns out to coincide with the response of a uniform elastic medium. However, fluctuations are relatively large, and their magnitude is satisfactorily captured by the FE simulations of an elastically heterogeneous system. Besides, accounting for elastic anisotropy on the mesoscale is not crucial in this respect. The proposed method thus paves the way for models of the rheology of amorphous solids which are both computationally efficient and realistic, in that structural disorder and inertial effects are accounted for.