Department of Energy Genesis Mission grant to UF-led team will speed up development of stronger materials

A cross-disciplinary research team led by the University of Florida won a prestigious federal grant on Wednesday to develop an autonomous artificial intelligence platform that could fundamentally change how new materials are discovered.  

Selected from more than 5,000 applications, the project is one of 278 chosen under the U.S. Department of Energy's highly competitive Genesis Mission Request for Applications program. It brings together universities, national laboratories and industry to accelerate scientific discovery through next-generation artificial intelligence.  

“It is going to be a huge collaborative effort,” said Ravit Silverstein, Ph.D., a project lead and assistant professor with UF’s Department of Materials Science & Engineering. For phase 1, UF is collaborating with the University of California Santa Barbara and industry partner Thermo-Fisher Scientific. 

The project will integrate electron microscopy, AI and physics-based modeling to create a platform that will accelerate the discovery of stronger, tougher and more heat-resistant structural materials for aerospace, energy and nuclear applications. 

“If you can operate your jet engine at a higher temperature, you can get more fuel efficiency and better aviation performance. We need new materials to operate at that temperature,” said Doug Spearot, Ph.D., chair of the UF’s Department of Mechanical and Aerospace Engineering and the project’s principal investigator.  

Advanced characterization methods, from electron microscopy to high-energy X-ray techniques, generate enormous datasets that reveal different aspectsof how materials evolve under extreme conditions. This platform is expected to speed up data interpretation more than 100 times. 

The Genesis Mission grants are designed to accelerate scientific discovery and development using novel AI models and frameworks. Teams are selected to help the DOE address critical energy, environmental and nuclear challenges through science and technology. 

The UF-led team plans to develop a feedback-driven AI framework that will quickly analyze massive amounts of data gleaned from materials experiments led by Silverstein. Physical experiments on the materials produce terabytes of data, which typically take weeks to analyze. This is a significant barrier to reliably modeling and designing the next generation of high‑strength alloys.  

“This is where the AI comes in,” Spearot explained. “The AI tool is meant to give a more real-time interpretation of diffraction patterns and nanoscale chemical information that are collected during in situ experiments.” 

UF team members will tap into UF’s supercomputer, HiPerGator, to help process that data. 

“We call it a closed-loop (AI framework), which can take the data, run it on the experimental side, on the modeling side, and try to get some sort of connection between the two,” Silverstein added. 

By drastically reducing development timelines, this project supports the rapid development and deployment of metallic materials with improved strength, toughness and reliability. The framework will be designed for scalability and future deployment at DOE user facilities, enabling broader access to advanced data analysis tools. 

“UF is honored to contribute to the DOE’s Genesis Mission,” said David Norton, Ph.D., UF’s vice president of research.  “Artificial intelligence promises to accelerate the understanding, design and implementation of new materials for a host of applications. UF is excited to be at the cutting edge of the AI revolution.”  

Spearot, Silverstein and MSE assistant professor Yijia Gu are working with UCSB’s Arda Genc, Ph.D., and Professor Daniel Gianola, Ph.D. Industry partner Thermo-Fisher Scientific is a large life sciences and clinical research company based in Massachusetts. It is a major manufacturer of electron microscopes and has a great interest in advanced materials. 

“What excites us most,” Genc said, “is the opportunity to transform electron microscopy from a tool that generates vast amounts of complex data into one that can rapidly extract high-impact scientific insights.”  

The project is timely, Gianola contends, as current electron microscopy data streams overwhelm a scientist’s ability to make real-time and reasoned decisions. 

“This project aims to transform the current paradigm into one where the use of AI-acceleration gives instant access to answers of the most vexing materials characterization questions,” Gianola said.