Accelerating Discovery of Ternary Chiral Materials via Large-Scale Random Crystal Structure Prediction

TL;DR

This paper introduces a scalable method combining machine learning potentials and random structure search to predict and identify new stable ternary chiral inorganic crystals with potential topological and quantum properties.

Contribution

It presents a novel high-throughput prediction framework that significantly expands the discovery space for chiral materials using large-scale structure generation and screening.

Findings

Identified over 260 potentially stable chiral inorganic crystals.

Confirmed materials exhibit quantum phenomena like nonlinear Hall effect and topological points.

Screened from over 20 million structures, demonstrating scalability and effectiveness.

Abstract

Chiral inorganic crystals, particularly semiconductors with Weyl points near the band edges or semimetals hosting Weyl points at the Fermi level, have attracted considerable interest, yet they remain scarce in existing materials databases. This study presents a prediction pathway by combining universal machine learning interatomic potentials (uMLIPs) for high-throughput structure optimization with the broad exploration capability of random structure search (RSS), enabling large-scale crystal structure prediction in ternary systems with variable compositions, followed by targeted screening for chiral space groups. Through uMLIP-based high-throughput optimization and stability assessment, a large number of potentially stable phases were identified from over 20 million randomly generated chiral structures. First-principles validation further confirmed more than 260 chiral inorganic crystals with potential applications in topological properties, nonlinear optics, and superconductivity. Some of these materials exhibit notable quantum phenomena, such as the nonlinear Hall effect driven by Berry curvature dipole, quantum metric and symmetry-protected sixfold degenerate topological points, long Fermi arcs, and large magnetoresistance. This work substantially expands the pool of candidate chiral functional materials and offers a scalable strategy for predicting ternary material systems.