Machine learning prediction of monatomic adsorption energies with non-first-principles calculated quantities

Abstract We established adsorption energy-property models using the properties of adsorbate atom and substrate metal atom that do not need first-principles or quantum mechanics calculations (non-QM based parameters) with machine learning (ML) methods of SVR, RFR and MLPR. The models produce adsorption energies close to the density functional theory calculated results for the 12 types of atoms on the 38 metal surfaces. Application of the models to predict the adsorption energies on binary alloys is also successful. The present work provides a means to evaluate adsorption energies with non-QM quantities and is helpful for catalyst screening.

Paper

Full text

PDF

Machine learning prediction of monatomic adsorption energies with non-first-principles calculated quantities

Semantic Scholar · Materials Science · 2020

Abstract

Abstract We established adsorption energy-property models using the properties of adsorbate atom and substrate metal atom that do not need first-principles or quantum mechanics calculations (non-QM based parameters) with machine learning (ML) methods of SVR, RFR and MLPR. The models produce adsorption energies close to the density functional theory calculated results for the 12 types of atoms on the 38 metal surfaces. Application of the models to predict the adsorption energies on binary alloys is also successful. The present work provides a means to evaluate adsorption energies with non-QM quantities and is helpful for catalyst screening.

Similar papers

© 2026 NYSGPT2525 LLC