Machine Learned Feature Identification for Predicting Phase and Young's Modulus of Low-, Medium- and High-Entropy Alloys

Abstract

Machine learning (ML) has emerged as a potential tool to rapidly accelerate the search for novel high entropy alloys (HEAs) due to its reasonably accurate property predictions. Here, we implement ML tools, to predict the crystallographic phase and Young's modulus of 26 Mo-Ta-Ti-W-Zr based HEAs. Our results, with experimental validation, reveal that mean melting point, electronegativity difference and the enthalpy of mixing are key features impacting the phase and Young's modulus of HEAs. Contrarily, entropy of mixing negligibly influences phase or the Young's modulus predictions, reigniting the issue of its actual impact on the phase and properties of HEAs.

Keywords: High-entropy alloys, machine learning, gradient boost algorithm, crystallographic phase, Young's modulus