Accelerating global search of adsorbate molecule position using machine-learning interatomic potentials with active learning

TL;DR

This paper introduces a machine-learning-based method that accelerates the global search for optimal adsorbate molecule positions on surfaces by combining interatomic potentials with active learning, validated across multiple catalytic systems.

Contribution

It develops an active learning algorithm with machine-learning interatomic potentials to efficiently find adsorption sites, improving search speed and accuracy over traditional methods.

Findings

Results agree with literature across various catalytic systems.

The method reduces computational cost of adsorption site searches.

Validated on multiple surface and adsorbate configurations.

Abstract

We present an algorithm for accelerating the search of molecule's adsorption site based on global optimization of surface adsorbate geometries. Our approach uses a machine-learning interatomic potential (moment tensor potential) to approximate the potential energy surface and an active learning algorithm for the automatic construction of an optimal training dataset. To validate our methodology, we compare the results across various well-known catalytic systems with surfaces of different crystallographic orientations and adsorbate geometries, including CO/Pd(111), NO/Pd(100), NH$_3$/Cu(100), C$_6$H$_6$/Ag(111), and CH$_2$CO/Rh(211). In the all cases, we observed an agreement of our results with the literature.