Composition-dependent dislocation mobility in FeNiCrCoCu high-entropy alloys based on atomistic simulations and machine learning analysis

Abstract Solid solution strengthening is a key mechanism for enhancing the strength of high-entropy alloys (HEAs). However, conventional strengthening theories fail to capture the complex environments in HEAs. Here, we present a data-driven framework to investigate the composition-dependent intrinsic strength of FCC HEAs. Using large-scale molecular dynamics simulations, we compute dislocation mobility under various temperatures and compositions, revealing jerky and wavy glide behavior due to fluctuating local pinning. The critical resolved shear stress (CRSS) at 0 K is extracted from these data, and a linear correlation is revealed between CRSS and the standard deviation of atomic pinning strength. Then, we propose atomic features describing local structural and compositional fluctuations and construct a symbolic model to predict the atomic pinning strength variability from these features, using the Sure Independence Screening and Sparsifying Operator method. This framework provides both mechanistic insight and predictive capability for the design of strong, compositionally complex alloys.