Integrated AI-driven framework for precise prediction of electrolyte additive oxidation potentials in lithium-ion batteries

Electrolyte engineering focuses on optimizing the physicochemical properties of electrolytes to enhance the performance of lithium-ion batteries (LIBs). The Introduction of electrolyte additives has proven to be a highly effective strategy in the field of electrolyte engineering, as even small quantities of the additives significantly enhance battery performance. Notably, additives with high oxidation potential are known to be advantageous for improving battery performance. However, despite extensive research on predicting oxidation potentials, discrepancies between theoretical calculations and experimental results remain a challenge. In this work, we propose a pioneering computational framework that integrates the principles of Density Functional Theory (DFT) with advanced machine learning and deep learning techniques, achieving superior consistency of experimental data with predicted oxidation potential compared to existing methodologies. Using this prediction model, novel electrolyte additives with high oxidation potentials were identified and experimentally validated through coin cell tests, demonstrating improved battery performance when incorporated into the electrolyte. This proof-of-concept study establishes a transformative framework for accelerating battery additive discovery and provides a foundation for future investigations into long-term cycling stability and interfacial characterization.