As the demand for safer and higher-performance batteries grows with the rise of electric vehicles (EVs) and energy storage systems (ESSs), solid-state batteries have emerged as a promising alternative to conventional lithium-ion batteries. However, low lithium-ion conductivity remains a major obstacle to commercialization.

To address these challenges, researchers at the Korea Institute of Science and Technology (KIST), in collaboration with POSCO Holdings, have developed an AI-driven method to optimize lithium argyrodite (Li₆PS₅X, X = Cl and Br) solid electrolytes. Using machine learning interatomic potential (MLIP) models, they accelerated simulations to conventional methods, enabling large-scale material exploration. Their study revealed that halogen doping (Cl and Br) and rearrangement of sulfur (S) significantly enhances ion conductivity—up to 100 times—by modifying the lithium migration pathways.

Dr. Byungju Lee from KIST stated, “This research integrates AI with computational chemistry, accelerating next-generation battery development and setting a foundation for AI-driven materials science.” The team plans to further validate their findings through experimental studies and refine their AI model for broader material applications.

Schematics of material research on AI-based simulation

[Reference] Lee et al., (2024) “Lithium Localization by Anions in Argyrodite Solid Electrolytes from Machine-Learning-based Simulations”, Advanced Energy Materials, 10.1002/aenm.202402396

[Main Author] Hyun-Jae Lee(Korea Institute of Science and Technology), Byungju Lee(Korea Institute of Science and Technology),

* Contact : Dr. Byungju Lee blee89@kist.re.kr