LLNL uses AI and simulations to improve sodium-ion and lithium-ion battery materials for better energy storage
miragenews.comResearchers at Lawrence Livermore National Laboratory are combining molecular dynamics simulations with physics-informed machine learning to study complex battery materials. Their work focuses on two areas: hard carbon anodes for sodium-ion batteries and liquid electrolytes for lithium-ion batteries. By simulating atomic behavior and training AI models on that data, they can predict how ions move, cluster, or get trapped, and identify design changes that improve performance and safety. Sodium-ion batteries use abundant domestic materials, making them important for supply chain resilience. The team created a quantitative map linking hard carbon microstructure to sodium ion transport, which could help engineers design better anodes. For lithium-ion electrolytes, their 3D molecular models predicted stability windows more accurately than traditional text-based approaches, revealing how salt choice and concentration affect performance. This framework could accelerate discovery across multiple battery chemistries and replace trial-and-error design with physics-guided screening. The work was supported by LLNL's Laboratory Directed Research and Development program and the U.S. Department of Energy.
