Berkeley Lab AI model predicts solid-state reaction pathways and impurities in minutes

A Department of Energy Lawrence Berkeley National Laboratory team demonstrated an AI modeling approach that predicts how reactions between solid materials unfold over time, Phys.org reported. The research is published in Nature Materials (DOI: 10.1038/s41563-026-02596-5). The model is described as the first predictive approach that accounts for how atoms travel through materials during solid-state reactions, combining thermodynamics with machine learning estimates of atom-transport kinetics. Inputs include starting materials, their ratios and the temperature ramp-up. In just minutes, the model simulates the full reaction pathway from start to finish, revealing intermediate and final products as well as impurities. Kristin Persson, a study author, said the model enables materials scientists and industry stakeholders to make promising new materials dramatically faster and with higher purity and yield. Solid-state reactions often require heating starting powders to temperatures as high as 800°C (1,470°F) to make atoms more mobile; prior thermodynamics-only tools struggled because they ignored kinetics. Phys.org said predictions for barium-titanium oxides closely matched experimental synthesis data across compositions and temperatures.