AI Discovers New Materials for Next-Generation Batteries, Moving Beyond Lithium

AI Discovers New Materials for Next-Generation Batteries, Moving Beyond Lithium
—————————
Researchers at the New Jersey Institute of Technology (NJIT) have used a new dual-AI system to discover five promising new materials for multivalent-ion batteries, a potential alternative to today’s standard lithium-ion technology.

The research, which was recently published in SciTech Daily, utilized a Crystal Diffusion Variational Autoencoder (CDVAE) and a Large Language Model (LLM) to rapidly screen thousands of potential crystal structures. This approach led to the identification of new porous transition metal oxide structures. These materials are well-suited for multivalent-ion batteries, which use elements like magnesium and zinc. These batteries could potentially store more energy because their ions carry a higher charge. The AI-designed materials feature large, open channels to accommodate these larger, high-charge ions, overcoming a key challenge in this field.

This AI-driven method, researchers believe, could be a scalable approach for discovering advanced materials in other fields, from electronics to clean energy.