AI-enabled discovery of rare earth elements

Physics-Informed AI for Accelerated Discovery of Ion-Adsorption-Type Rare Earth Element Resources in Granitic Systems

Heavy rare earth elements (REEs) are essential for modern industry, yet the United States relies heavily on imported supplies. Most of the world's heavy REEs come from deposits in Asia that formed as granite slowly broke down over time. In these deposits, the REEs are loosely attached to secondary minerals formed during weathering, making them easier to recover. Scientists still do not fully understand why these deposits become enriched with REEs, making it difficult to identify similar resources in the United States. This project will investigate how natural processes, including rock weathering, water infiltration, and landscape erosion, can move and concentrate REEs. The team will develop a physics-informed AI framework that integrates field observations, laboratory experiments, and computer models to improve predictions of where these deposits may occur. Initial work will focus on three granite watersheds in North Carolina. The resulting approach will speed the discovery of new domestic sources of heavy REEs across the southeastern United States.