AI plus quantum physics identifies two new kagome-lattice superconductors, unlocking a faster search method
An international team paired machine learning with quantum physics to confirm two new superconductors, YRu3B2 and LuRu3B2, at 0.81 K and 0.95 K. Both rely on flat electron bands in a kagome lattice. The reusable output is the screening method. Source: Physical Review Research.
An international research team has used artificial intelligence together with quantum physics to identify two new superconductors, and the method they built may matter more than the materials themselves. The two compounds are yttrium-ruthenium-boride (YRu3B2) and lutetium-ruthenium-boride (LuRu3B2), with measured critical temperatures of 0.81 K and 0.95 K. Both become superconducting because their electrons form flat bands inside a kagome lattice, a repeating geometric structure. To find them, the team first used machine learning to screen a large number of possible elemental combinations, then applied quantum-physics calculations to confirm the most promising candidates. The materials are far too cold to be practical, but the screening approach is designed to be reused across thousands of other candidate combinations. The work involved Aalto University, Rice University, Princeton University, Ruhr University Bochum and the Donostia International Physics Center, and was published in Physical Review Research on June 17, 2026.