A high-fidelity neural network-based force field, NN-F3, is developed to cover the strain states up to material failure and the non-equilibrium, intermediate nature of fracture. Simulations of fracture in 2D crystals using NN-F3 reveal spatial complexities from lattice-scale kinks to sample-scale patterns. We find that the fracture resistance cannot be quantified by the energy densities of relaxed edges as in the literature. Instead, the fracture patterns, critical stress intensity factors at the kinks, and energy densities of edges in the intermediate, unrelaxed states offer reasonable measures for the fracture toughness and its anisotropy.
@article{arxiv.2307.16126,
title = {Non-Equilibrium Nature of Fracture Determines the Crack Paths},
author = {Pengjie Shi and Shizhe Feng and Zhiping Xu},
journal= {arXiv preprint arXiv:2307.16126},
year = {2024}
}