AI database points to rare earth alternatives for electric vehicles

Scientists at the University of New Hampshire are using AI to accelerate the search for advanced magnetic materials that could reduce reliance on rare earth elements. Their Northeast Materials Database contains 67,573 magnetic compounds, including 25 newly recognized materials that can remain magnetic at high temperatures.

The resource is designed to help researchers identify cheaper and more sustainable alternatives to today’s strongest magnets, which are widely used in smartphones, medical devices, power generators, electric vehicles, and other systems. Many of those magnets depend on rare earth elements that are costly, largely imported, and increasingly difficult to secure.

The study, published in Nature Communications, describes an AI system that reads scientific papers and extracts experimental data. Researchers used the information to train computer models that determine whether a material is magnetic and calculate the temperature at which it loses magnetism, then organized the results into a searchable database.

The team said the approach could make a difficult materials science challenge more achievable by narrowing the search among possible candidates that could reach into the millions. The project was supported by the Office of Basic Energy Sciences, Division of Materials Sciences and Engineering, at the U.S. Department of Energy.