Microbranching in mode-I fracture using large scale simulations of amorphous and perturbed lattice models

TL;DR

This study uses large-scale GPU simulations to investigate microbranching in mode-I fracture, revealing convergence patterns, growth behavior, and branching angles consistent with experimental observations.

Contribution

It provides the first large-scale, GPU-accelerated simulations of microbranching in amorphous and perturbed lattice models, achieving more realistic results.

Findings

Microbranching pattern converges with lattice width

Microbranch size increases with crack velocity

Branching angles match experimental data

Abstract

We study the high-velocity regime mode-I fracture instability when small microbranches start to appear near the main crack, using large scale simulations. Some of the features of those microbranches have been reproduced qualitatively in smaller scale studies (using ${\cal O}(10^4)$ atoms) on both a model of an amorphous materials (via the continuous random network model) and using perturbed lattice models. In this study, larger scale simulations (${\cal O}(10^6)$ atoms) were performed using multi-threading computing on a GPU device, in order to achieve more physically realistic results. First, we find that the microbranching pattern appears to be converging with the lattice width. Second, the simulations reproduce the growth of the size of a microbranch as a function of the crack velocity, as well as the increase of the amplitude of the derivative of the electrical resistance RMS with respect to the time as a function of the crack velocity. In addition, the simulations yield the correct branching angle of the microbranches, and the power law governing the shape of the microbranches seems to be lower than one, so that the side cracks turn over in the direction of propagation of the main crack as seen in experiment.