Batch Discovery of New Metal Superhydrides via Chemical Template Theory and Machine Learning

Abstract

Metal superhydrides, known for their high hydrogen content and polyhedral hydrogen cages, are promising candidates for high-temperature superconductivity. Recent research has emphasized "chemical pre-compression," enabling hydrogen metallization at lower pressures and facilitating the discovery of superconductors with near-room temperature transitions. Despite extensive studies on binary metal superhydrides, there remains a vast, unexplored chemical space, particularly regarding non-integer hydrogen-to-metal ratios. By integrating the "chemical template effect" with machine learning algorithms, we developed a specialized structure discovery workflow that significantly enhances the efficiency of predicting stable superhydrides. Our method led to the identification of 13 new structural prototypes and 31 stable metal superhydrides, representing a 25% increase in discoveries. The 3D hydrogen clathrates in these compounds are strongly correlated with high superconducting transition temperatures, and our approach achieves a remarkable 70% increase. Most of these structures contain over 50 atoms per primitive cell, with the I4/m M10H84 prototype having the largest unit cell, containing 94 atoms. Additionally, 19 of the newly identified superhydrides exhibit superconducting transition temperatures (Tc) exceeding 100 K, highlighting the potential for higher Tc materials within the 3D hydrogen clathrate structures. The method also shows good potential to search for ternary superhydrides on a large scale.