The University of Tennessee, Knoxville, has received $20 million from the National Science Foundation to establish ATHENA (Advanced Testbed for High-throughput Experimentation in Nano- and Atomic Science), one of 20 research hubs in the U.S. that will form a national network of AI-powered laboratories to accelerate scientific discovery.
Led by Sergei Kalinin, Weston Fulton Professor in UT’s Tickle College of Engineering; Mahshid Ahmadi, associate professor of materials science and engineering; and Hairong Qi, Gonzalez Family Professor of Electrical Engineering and Computer Science, ATHENA brings together a multidisciplinary team of 12 UT faculty members along with key collaborators at Northwestern University and Johns Hopkins University. The ATHENA initiative aims to revolutionize materials discovery by dramatically accelerating the design, synthesis, characterization and autonomous optimization of advanced materials, enabling the rapid identification of materials with the greatest potential for real-world applications.
While artificial intelligence has dramatically improved scientists’ ability to predict promising new materials, testing those predictions takes time and remains labor intensive.
“Predictions alone don’t create new technologies,” Kalinin said. “At some point, those materials have to be synthesized, measured and understood in the real world. ATHENA is about dramatically accelerating that process.”
ATHENA is part of NSF’s Platform for Cloud Laboratories initiative, a nationwide effort to transform scientific research by connecting remotely accessible AI-enabled laboratories that can autonomously plan, conduct and refine experiments. The initiative is designed to accelerate discoveries in materials science and biotechnology while expanding researchers’ access to advanced scientific instrumentation across the country.
Building the next generation of scientific discovery
Today’s materials research often requires scientists to manually prepare samples, operate sophisticated microscopes and analyze data — a process that can take hours or days for a single experiment.
ATHENA seeks to dramatically improve that workflow with AI-enabled “self-driving laboratories” that can perform experiments, interpret results and determine next steps with minimal human intervention.
The team expects the platform to increase the speed of some materials characterization experiments by as much as 10 to 30 times, dramatically reducing one of the biggest bottlenecks in materials discovery. By studying materials at the atomic scale and nanoscale, researchers can rapidly evaluate thousands of material combinations before determining the most promising candidates for real-world applications.
A national resource built at UT
ATHENA builds on decades of work by UT researchers in autonomous microscopy, automated materials synthesis and artificial intelligence.
Long before AI became a major focus in scientific research, Kalinin and colleagues across the university were developing machine learning tools that allow microscopes and other scientific instruments to communicate directly with intelligent software. Those foundational technologies now position UT to define how autonomous laboratories operate nationwide.
“This proposal reflects years of investment by the university and an exceptional team of researchers,” Kalinin said. “Many of the technologies that make self-driving laboratories possible were developed here at UT long before this became a national priority.”
The project will also create open software standards, cloud-accessible laboratory workflows and digital tools that allow researchers across the country to remotely access advanced scientific instruments. As one of the NSF’s PCL nodes, ATHENA will work with other national laboratories to develop common standards, expand access to cutting-edge research infrastructure and train the next generation of scientists working at the intersection of AI and experimental science.
“We believe the future isn’t simply about computing faster,” Kalinin said. “It’s about discovering, making and understanding new materials faster. If we can dramatically shorten the time between an idea and a new material, we can accelerate innovation across nearly every industry.”