Selective Recovery of Light and Heavy Rare-Earth Element Pairs via Functionalized Magnetic Sorbents with Induction-Assisted Regeneration
The United States lacks a sustainable, domestic, chemically efficient process for separating critical rare earth elements (REEs), creating a major bottleneck in the supply chain for clean energy technologies and advanced manufacturing.
This project, led by the University of Miami, will develop an AI-driven, acid-free Magnetic Induction Swing Adsorption (MISA) platform to replace conventional solvent extraction for REE separation. Using phosphonate-functionalized magnetic nanoparticles, the technology will selectively recover and separate neodymium (Nd), praseodymium (Pr), dysprosium (Dy), and terbium (Tb) from low-grade domestic scrap and unconventional ores. Artificial intelligence will accelerate process development through multi-fidelity Bayesian optimization informed by Department of Energy databases and physics-informed neural networks. Phase I will establish the physicochemical limits of field-driven separation, provide validated surrogate models and sorbent performance data for production of ≥95% pure light (Nd/Pr) and heavy (Dy/Tb) rare earth fractions while demonstrating sorbent durability over more than 50 adsorption–desorption cycles.
