Researchers at the New Jersey Institute of Technology (NJIT) and Rensselaer Polytechnic Institute (RPI) have identified five previously unreported oxide compositions that could serve as electrode hosts for multivalent-ion batteries.
The team, led by Joy Datta under adviser Dibakar Datta at NJIT, working with Nikhil Koratkar at RPI, combined a crystal diffusion variational autoencoder (CDVAE) with a fine-tuned large language model to generate roughly 20,000 candidate transition metal oxide structures. The models were trained on more than 44,000 known inorganic crystal structures from the Materials Project database.
The researchers filtered the generated structures using formation energy, thermodynamic stability and electronic band gap criteria, narrowing the field to 55 compositions. Of those, the CDVAE model produced the five structures, featuring open-tunnel frameworks designed to accommodate magnesium, calcium, aluminum or zinc ions – alternatives to lithium that the authors say are more abundant and lower-cost.
One candidate, Ca4In2O2, showed an energy above the convex hull of 0.36 eV/atom, a level the researchers describe as metastable rather than fully stable. A phonon dispersion analysis of the structure showed no unstable vibrational modes across its entire Brillouin zone, which the scientists say indicates it could potentially be synthesized under non-equilibrium conditions despite its thermodynamic metastability.
The research team also developed a retrieval-augmented generation system that searches scientific literature and proposes potential synthesis conditions – reaction temperature, atmosphere and safety protocols – for compositions with no experimental precedent. Applied to a related compound, K2Cu4F10, the system proposed a solid-state reaction route and identified nine feasible elemental substitutions.
None of the AI-generated compositions has yet been synthesized or tested experimentally, and the authors describe the work as a computational framework rather than a validated battery material. Funding for related earlier work in the same research program came from a National Science Foundation award supporting niobium tungsten oxide anode research at NJIT.
The researchers discussed their findings in “Generative AI for Discovering Porous Oxide Materials for Next-Generation Energy Storage,” which was recently published in Cell Reports Physical Science.
The post AI framework identifies five novel oxide candidates for battery hosts appeared first on Energy Storage.
