Automated atomistic simulations of dissociated dislocations with ab initio accuracy

TL;DR

This paper extends an automated machine-learning interatomic potential training algorithm to simulate dissociated dislocations in face-centered cubic aluminum, achieving DFT-level accuracy and enabling large-scale dislocation simulations.

Contribution

It introduces a universal algorithm for simulating dissociated dislocations in FCC metals, validated on aluminum, with potential applications to other materials and alloys.

Findings

Excellent agreement with DFT reference results.

Potential to simulate dislocations in various FCC and HCP materials.

Framework for constructing training sets for MLIPs in large-scale dislocation simulations.

Abstract

In (M Hodapp and A Shapeev 2020 Mach. Learn.: Sci. Technol. 1 045005), we have proposed an algorithm that fully automatically trains machine-learning interatomic potentials (MLIPs) during large-scale simulations, and successfully applied it to simulate screw dislocation motion in body-centered cubic tungsten. The algorithm identifies local subregions of the large-scale simulation region where the potential extrapolates, and then constructs periodic configurations of 100--200 atoms out of these non-periodic subregions that can be efficiently computed with plane-wave Density Functional Theory (DFT) codes. In this work, we extend this algorithm to dissociated dislocations with arbitrary character angles and apply it to partial dislocations in face-centered cubic aluminum. Given the excellent agreement with available DFT reference results, we argue that our algorithm has the potential to become a universal way of simulating dissociated dislocations in face-centered cubic and possibly also other materials, such as hexagonal closed-packed magnesium, and their alloys. Moreover, it can be used to construct reliable training sets for MLIPs to be used in large-scale simulations of curved dislocations.