Predictions of adsorption energies of methane-related species on Cu-based alloys through machine learning
Yun Zhang, Xiaojie Xu
Abstract
Recent studies show that the adsorption energy can be used as a descriptor of the catalytic activity in methane direct conversion. We develop Gaussian process regression models to predict DFT-calculated adsorption energies of CH4 related species – CH3 , CH2, CH, C, and H – on Cu-based alloys from elements’ readily available physical properties. As compared to conventional first-principle-based methods, the models are simple and fast to implement. They produce predictions with root mean squared errors of below 0.15 eV. The models also present numerical and statistical relationships between fundamental physiochemical parameters of doped elements and adsorption energies. Hence, they might be considered as efficient alternatives to the DFT approach for adsorption energy calculations, which allow for further assessments of certain solid catalysts’ catalytic performance.