Generalized stacking fault energies and Peierls stresses in refractory body-centered cubic metals from machine learning-based interatomic potentials

XW Wang and SZ Xu and WR Jian and XG Li and YQ Su and IJ Beyerlein, COMPUTATIONAL MATERIALS SCIENCE, 192, 110364 (2021).

DOI: 10.1016/j.commatsci.2021.110364

The generalized stacking fault energies (GSFE) and Peierls stresses are strongly related to the mechanical properties of refractory metals. In this work, the GSFE curves and Peierls stresses of screw and edge dislocations in four body-centered cubic refractory metals (Mo, Nb, Ta, and W) on the 110, 112, and 123 slip planes are calculated using molecular statics simulations. A recently developed machine learning (ML)-based interatomic potential, called the spectral neighbor analysis potential (SNAP), is employed in all simulations. The computed GSFE curves achieve reasonable agreement with those from ab initio calculations and predict the asymmetry with respect to sense of glide direction on the 112 and 123 planes better than non-ML interatomic potentials. In general, SNAP provides screw dislocation Peierls stresses close to those of density functional theory, closer than those achieved by non-ML potentials. The screw dislocation Peierls stress values confirm slip symmetry on the 110 plane and exhibit pronounced slip asymmetry on the 112 and 123 planes. For all metals, the edge dislocation Peierls stress are the lowest on the 110 plane and the highest on the 112 plane. For screw dislocations, glide on either the 110 or the 123 plane is the easiest.

Return to Publications page