Accelerating Structure Prediction of Molecular Crystals using Actively Trained Moment Tensor Potential

TL;DR

This paper introduces an active learning approach using Moment Tensor Potentials to speed up molecular crystal structure prediction, demonstrated on benzene and glycine, revealing new structural insights.

Contribution

It presents a novel methodology combining active learning with Moment Tensor Potentials for efficient molecular crystal structure prediction.

Findings

Successfully applied to benzene and glycine.

Discovered a new polymeric benzene structure.

Accelerated structure prediction process.

Abstract

Inspired by the recent success of machine-learned interatomic potentials for crystal structure prediction of the inorganic crystals, we present a methodology that exploits Moment Tensor Potentials and active learning (based on maxvol algorithm) to accelerate structure prediction of molecular crystals. Benzene and glycine are used as test systems. Interestingly, among obtained low energy structures of benzene we have found a peculiar polymeric benzene structure.