Uncovering the microstructural origins of shear transformation events in metallic glasses: Insights from combining knowledge and data

Abstract

In this work, an innovative dimensionless static structural parameter called the “Composite flexibility factor (CFF)” was proposed based on a knowledge-driven research paradigm, aiming to establish a relationship between the static structure of metallic glasses (MGs) and shear transformation (ST) events. The simulation results demonstrated that a significant correlation between the CFF parameter and the ST events in the metal systems, metal–metalloid systems and metal-nonmetal systems MGs, as well as under different quenching conditions. An interpretable ST events prediction model based on a deep neural network and the Shapley Additive exPlanations method was constructed using the atomic structure features extracted from the knowledge-driven approach. This work not only enhances the correlation between the static structure of MGs and ST events but also provides new research pathway and perspective for revealing the microscopic mechanisms of plastic deformation in MGs.