Efficient Machine-Learning-Aided Screening of Hydrogen Adsorption on Bimetallic Nanoclusters

Area of Science:

• Materials Science
• Computational Chemistry
• Catalysis

Background:

• Catalytic activity is influenced by material properties at the nanoscale.
• Computational screening of nanoclusters is challenging due to the vast search space.

Purpose of the Study:

• To develop an automated workflow for efficient computational screening of nanocluster catalysts.
• To explore bimetallic icosahedral clusters for the hydrogen evolution reaction.

Main Methods:

• Automated simulation management from nanocluster generation to adsorption energy prediction.
• Application of machine learning for efficient exploration and screening.
• Focus on bimetallic icosahedral clusters and the hydrogen evolution reaction.

Main Results:

• Demonstrated efficient exploration of nanocluster configurations.
• Successfully screened adsorption energies using machine learning.
• Identified a strong correlation between the d-band Hilbert-transform and adsorption energies.

Conclusions:

• The developed workflow enables systematic and efficient screening of nanocluster catalysts.
• The d-band Hilbert-transform is a promising descriptor for predicting catalytic activity at the nanocluster level.