Department of Energy's Genesis Mission selects Tulane team using AI to discover next-generation superconductors

Tulane was among the teams selected by the U.S. Department of Energy to receive Phase I proof-of-concept funding through the Genesis Mission, a national initiative to harness AI for breakthroughs in energy dominance, discovery science and national security. The announcement was made July 22 at the Genesis Mission Summit in Washington, D.C.

During the approximately nine-month Phase I period, the selected teams will build their AI-driven research frameworks, generate initial results and demonstrate the feasibility of their approaches. DOE will then select a smaller group of awardees for Phase II deployment projects, which could provide multiyear awards ranging from $6 million to $15 million.

Tulane's interdisciplinary team will develop a workflow combining artificial intelligence, high-fidelity quantum-mechanical simulations and laboratory experiments to accelerate the discovery of new superconductors and quantum materials for future energy and quantum technologies.

The AI-driven workflow will learn from quantum-mechanical calculations and experimental measurements to identify the most promising materials and determine which calculations and experiments to perform next.

"Discovering a new quantum material is extraordinarily difficult because the number of possible compositions and atomic structures is almost limitless," said Jianwei Sun, the project's principal investigator and a professor of physics and engineering physics at Tulane University School of Science and Engineering. "By combining high-fidelity quantum-mechanical calculations, physics-aware AI and experimental measurements, we want to create a scientific workflow that learns continuously and directs researchers toward the most promising possibilities."

While superconductors can carry electrical current with no resistance, they must be cooled to extremely low temperatures, making them costly and difficult to use outside of specialized settings.

A superconductor that operates at substantially higher temperatures could contribute to a more efficient and resilient electrical grid, improved medical imaging systems, more powerful magnets, energy-efficient data centers, advanced transportation technologies, quantum computing and fusion energy.

Unlike generative AI systems trained on language, the Tulane team's AI will be built on scientific data from high-fidelity quantum-mechanical calculations and experimental measurements, with the laws of physics incorporated directly into the model.

"Our AI is not a chatbot," said Aron Culotta, a Tulane professor whose work focuses on artificial intelligence and machine learning. "Instead of learning patterns in language, it will learn patterns in atoms, electrons, magnetism and atomic vibrations that can help predict how a material will behave."

"There is still no general theory of high-temperature superconductivity," said Ruiqi Zhang, a research assistant professor in Tulane University's Department of Physics and Engineering Physics, who helped develop the project's computational framework. "Combining high-fidelity quantum-mechanical simulations, experimental data and physics-informed AI could provide a new pathway for discovering superconductors and quantum materials."

The computational predictions will be tested through laboratory experiments. Daniel B. Straus, an assistant professor of chemistry at Tulane School of Science and Engineering, whose research focuses on materials chemistry, will help connect the team's predictions with materials that can be synthesized and studied at Tulane.

The Tulane researchers are also collaborating with Huibo Cao of Oak Ridge National Laboratory, where neutron-scattering experiments will help determine the crystal and magnetic structures of selected materials and test the team's predictions.

The project brings together researchers in theoretical and computational physics, artificial intelligence, materials chemistry, data science, high-performance computing and experimental science.

"What makes this project especially exciting is its interdisciplinary team," said Sun, who earned his master's degree and PhD from Tulane, where he has conducted research for more than two decades. "Physicists, chemists, AI researchers and experimental scientists are working together on a problem that none of us could solve alone."