Machine learning and density functional theory-based analysis of the surface reactivity of high entropy alloys: The case of H atom adsorption on CoCuFeMnNi

Abstract

This study examines the adsorption of H atom on CoCuFeMnNi(111) high entropy alloy (HEA) surface using a combination of density functional theory (DFT) and machine learning (ML) techniques. Hume-Rothery rule, thermodynamic parameters, and electronic structure analysis were utilized to elucidate the stability and reactivity of the CoCuFeMnNi surface. We found that CoCuFeMnNi is a stable solid solution with a fcc structure. By integrating surface microstructure-based input features into our ML model, we accurately predicted H adsorption energies on the hollow sites of CoCuFeMnNi surfaces. Our electronic properties analysis of CoCuFeMnNi revealed that there is an evident interaction among the elements, contributing to a broad range of adsorption energies. During adsorption, the nearest neighbor surface atoms to H directly engage with the adsorbate by transferring charge significantly. The atoms in other regions of the surface contribute through charge redistribution among the surface atoms, influencing overall charge transfer process during H adsorption. We also observed that the average of the d-band centers of the nearest neighbor surface atoms to H influence the adsorption energy, supporting the direct participation of these surface atoms toward adsorption. Our study contributes to a deeper understanding of the influence of surface microstructures on H adsorption on HEAs.