Topological Quantum Cathode Materials for Fast Charging Li‐Ion Battery Identified by Machine Learning and First Principles Calculation

Wei Wu,Shuo Wang,Qiang Sun

DOI: https://doi.org/10.1002/adts.202100350

2022-01-11

Advanced Theory and Simulations

Abstract:Fast charging electrode materials require excellent ion conductivity as well as high and stable electrical conductivity. However, the main commercial cathode materials are semiconducting transition metal oxides suffering from low electrical conductivity, so cathode materials with good electrical conductivity are highly desirable. It is well‐known that topological quantum materials (TQMs) can exhibit robust electrical conductivity, thus providing a promising solution to this problem. The key question becomes which TQMs can have high ionic conductivity. Herein, such topological quantum cathode materials for fast charging Li‐ion battery are identified by using machine learning combined with first‐principle calculation, where the supervised regression models are trained for diffusion energy barrier prediction. Among 7385 TQMs, 20 materials are found to have a diffusion energy barrier less than 1.0 eV. Especially, LiMnAs exhibits a reversible capacity of 195.9 mAh g−1, higher than that of commercial cathodes while with comparable diffusion energy barrier. Furthermore, LiMnAs shows high interface stability with selected solid lithium metal oxides electrolytes. This study not only opens a new path to developing novel cathode materials, but also expands the applications of TQMs. High performance cathode material LiMnAs is identified by screening more than 7000 topological materials with machine learning, which exhibit intrinsic high electric conductivity, low diffusion barrier, high reversible capacity, and high interface stability with selected solid lithium metal oxides electrolytes.

multidisciplinary sciences