An AI platform scans published research to identify magnetic compounds that may reduce dependence on constrained rare earth elements.
Researchers at the University of New Hampshire have developed an AI driven platform to accelerate the discovery of new magnetic compounds. The project resulted in the Northeast Materials Database, which compiles 67,573 magnetic material entries, including 25 compounds identified as remaining magnetic at elevated temperatures.
Permanent magnets are essential for electric vehicles, renewable energy systems, medical devices and consumer electronics. Many rely on rare earth elements that are costly and subject to supply constraints. Identifying alternative materials has been challenging because testing millions of possible chemical combinations in the laboratory is impractical. By automating the search process, the AI system is intended to reduce the time and cost associated with early stage materials screening and help researchers focus on the most promising candidates.
The platform’s core features include automated reading of scientific literature using a large language model, extraction of experimental parameters such as composition and magnetic transition temperatures, and organization of this data into a structured, searchable database. Machine learning models analyse the compiled data to predict whether a material is magnetic and estimate the temperature at which it may lose magnetism. The database framework also allows continuous updates as new research is published.
Suman Itani, a Ph.D. student in physics and the lead author at UNH, says, “By accelerating the discovery of sustainable magnetic materials, we can reduce dependence on rare-earth elements, lower the cost of electric vehicles and renewable-energy systems, and strengthen the U.S. manufacturing base.”
Yibo Zhang, a postdoctoral researcher in both physics and chemistry notes that going forward these large LLMs behind this project can have widespread applications beyond this database, particularly in higher education fields like converting images to modern rich text format could also be used to modernize library holdings.
