Leveraging Machine Learning to Understand Energy Storage Mechanisms in Nanoscale Electrode Materials

Designing high-performance batteries necessitates a mechanistic understanding of the intricate processes governing energy storage mechanisms. This entails providing atomistic-level details to decipher the chemical processes limiting nanoscale electrode materials' performance. These processes include ion adsorption, ion intercalation, and pore filling, all requiring scalable systems and accurate models to be effectively studied and optimized.<br/><br/>Therefore, in this study, we develop machine learning interatomic potentials (MLIP) to understand the energy storage mechanisms of Li-ion in MnO₂ and Na-ion in Hard Carbon. These contrasting systems provide a fundamental baseline to compare charge carriers and nanoscale materials. We couple them with enhanced sampling simulations to decode the phase transformation in MnO₂ and identify the rate-limiting steps in Na-ion storage in Hard Carbon. Understanding these properties through MLIP and enhanced sampling simulations allows us to design better electrode materials, ultimately leading to more efficient and higher-performing batteries.<br/><br/>This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under contract DE-AC52-07NA27344.