Understanding Structure-Composition-Property Relationships of Ni-P Bulk Metallic Glasses

Abstract

Bulk metallic glasses (BMGs) are a unique class of materials characterized by their disordered atomic structure, which imparts exceptional mechanical strength, corrosion resistance, and catalytic activity. Yet, optimizing the composition of BMGs for desired properties typically relies on empirical trial-and-error, at most guided by qualitative computational models. Here, we combined machine learning with electronic-structure theory to quantitatively map the structure-composition-property relationships of Ni-P-based BMGs. Our simulations using a neural-network-based machine learning interatomic potential predict that the glass transition temperature of the BMG decreases with the phosphorus content, in quantitative agreement with experimental observations. We find that this trend is correlated with medium-range order in the material that emerges when the phosphorus content is sufficiently high. On the atomic scale, we find P-centered cluster motifs that vary in structure with the composition and temperature and impact the atomic mobility in the Ni-P BMG. This atomic-scale insight explains the composition-dependent stability of the Ni-P BMG and demonstrates how machine-learning interatomic potentials can guide the design and optimization of glassy/amorphous materials such as BMGs.