Superconductors, capable of conducting electricity without loss, are essential for quantum computers, MRIs, and maglev trains. But these materials only work at extremely low temperatures, which hinders their use. Finding a superconductor that operates at room temperature is considered the holy grail of modern physics.
An international team of scientists has just taken a decisive step in this quest. By combining artificial intelligence with advanced quantum physics calculations, they have developed a method that can screen billions of material combinations to identify the most promising ones.
Their approach has already borne fruit: two new superconductors have been discovered, named YRu3B2 and LuRu3B2. These materials owe their properties to a particular geometric structure called the kagome lattice, named after a Japanese basket-weaving pattern.
YRu3B2 and LuRu3B2 owe their superconductivity to electrons forming flat bands in a kagome lattice, named after a hexagonal Japanese basket-weaving pattern.
Credit: Esa Kapila The major difficulty in the search for superconductors lies in the almost infinite number of possible alloys. Of the roughly 7,000 materials already identified as superconductors, most were discovered by chance. Theoretical calculations to predict whether a compound can become superconducting are so heavy that they have only been performed for about twenty of them. The new strategy of the SuperC consortium changes the game: a machine learning algorithm preselects the most interesting candidates before performing detailed quantum calculations, significantly reducing the time and resources required.
Professor Päivi Törmä, who leads the collaboration, explains that this method could dramatically accelerate the discovery of new materials. With machine learning, it becomes possible to process billions of chemical combinations, whereas classical approaches maxed out at a few hundred. The researchers were thus able to focus on the most promising structures, such as the kagome lattice, where electrons form "flat bands" that favor the emergence of superconductivity.
Once the theoretical predictions were established, the team at Rice University, led by Professor Emilia Morosan, synthesized the two new materials in the laboratory. Tests confirmed that they were indeed superconducting, validating the method. This proof of concept, published in the journal
Physical Review Research, paves the way for a much more systematic and rapid search for superconductors.
The ultimate goal remains the discovery of a superconductor operating at room temperature. Such a material could transform the storage and transport of electricity, significantly reducing global energy consumption by eliminating losses. The SuperC consortium, founded in 2023, has set itself the goal of achieving this by 2033. Thanks to the alliance of artificial intelligence and quantum physics, this quest may well succeed sooner than expected.