LiCondAI: A Machine Learning Model for Predicting Lithium-Ion Conductivity in Solid Electrolytes

Significant momentum is driving advancements in solid-state battery (SSB) technology, where the choice of electrolyte plays a key role in determining cell performance. In this study, an advanced machine learning ensemble model, called LiCondAI (lithium-ion conductivity using Artificial Intelligence), was developed to predict the ionic conductivity (IC) of solid electrolytes (SEs) in solid-state batteries (SSBs). LiCondAI combines the predictive capabilities of three models─XGBoost, Gradient Boosting, and Random Forest─using a stacking regressor mechanism, achieving 97% accuracy in predicting IC. The proposed model, LiCondAI, accurately predicts IC in SEs for lithium-ion and lithium metal SSBs. LiCondAI is a useful platform for examining the behavior of SEs in SSBs and serves as a tool for guiding the design of new electrolytes for SSBs.