Machine Learning Assisted Exploration of High Entropy Alloy-Based Catalysts for Selective CO2 Reduction to Methanol

Abstract

Catalytic conversion of CO2 to carbon neutral fuels can be ecofriendly and allow for economic replacement of fossil fuels. Here, we have investigated high-throughput screening of high entropy alloy (Cu, Co, Ni, Zn, and Sn) based catalysts through machine learning (ML) for CO2 hydrogenation to methanol. Stability and catalytic activity studies of these catalysts have been performed for all possible combinations, where different elemental, compositional, and surface microstructural features were used as input parameters. Adsorption energy values of CO2 reduction intermediates on the CuCoNiZnMg- and CuCoNiZnSn-based catalysts have been used to train the ML models. Successful prediction of adsorption energies of the adsorbates using CuCoNiZnMg-based training data is achieved except for two intermediates. Hence, we show that activity and selectivity of these catalysts can be successfully predicted for CO2 hydrogenation to methanol and have screened a series of high entropy-based catalysts (from 36750 considered catalysts) which could be promising for methanol synthesis.