Dr. Murat Barisik (photo by Angela Foster)
For the third time over the past year, a research paper co-authored by University of Tennessee at Chattanooga Associate Professor of Mechanical Engineering Murat Barisik has been featured on the cover of an international journal.
Barisik is a co-author of the cover article in Nanomaterials, a journal focused on materials science research. The paper, titled “Machine Learning for Thermal Transport Prediction in Nanoporous Materials: Progress, Challenges, and Opportunities,” examines how machine learning can be used to better predict how heat moves through nanoporous materials used in energy storage, insulation and advanced manufacturing.
The Nanomaterials cover follows recent cover features of Barisik’s research in Nanoscale—a flagship journal of the Royal Society of Chemistry—and Small, a nanoscience and nanotechnology journal.
For Barisik, the recognition reflects both the direction of the work and the standards he sets for his research group.
“I’m grateful for the recognition,” Barisik said. “But what matters most to me is the bar we set for ourselves. We try to take on the most pressing challenges in science and engineering and share results that bring something new and useful to others.
“The cover is encouraging, but I’m proudest of the quality of the work and the students and collaborators who made it happen.”
The Nanomaterials paper tackles a long-standing challenge in materials engineering: predicting how heat moves through complex, porous structures at the nanoscale. Because of their intricate geometry, these materials are difficult to model accurately, and traditional approaches often struggle with the limited high-quality data available.
The paper was co-authored by Amirehsan Ghasemi, a Ph.D. student in the Oak Ridge Innovation Institute/Bredesen Center for Interdisciplinary Research and Graduate Education program. Together, they review how machine learning techniques—ranging from convolutional and graph-based neural networks to physics-informed models—can address those challenges.
This work is supported by federal funding agencies, including the National Science Foundation and the Department of Energy, in collaboration with Oak Ridge National Laboratory. Barisik is a recipient of an NSF CAREER Award, which supports early-career faculty who advance high-impact research while strengthening education.
Dr. Murat Barisik’s research has been featured in recent months on the front cover of Nanomaterials, left, the back cover of Nanoscale and the front cover of Small.
The focus of the Nanomaterials cover, Barisik explained, illustrates a closed loop in which simulation, artificial intelligence, and manufacturing inform one another.
The process begins with detailed simulations that model materials at nanoscales, generating data about how a material’s structure influences thermal behavior. That information serves as the foundation for machine learning models that can make reliable predictions without requiring new, expensive simulations each time.
Barisik said those predictions then guide manufacturing decisions. By understanding how changes at the nanoscale affect material properties—such as thermal conductivity—researchers can tailor how a material is made, adjust its structure and refine the process. That new information loops back into the models, improving their accuracy over time.
The goal, he said, is a predictive framework that helps engineers design materials more efficiently rather than relying on trial and error.
To help illustrate the concept, he displayed a quarter-sized piece of aerogel—materials sometimes described as “frozen smoke.”
“This is 99% air,” he said. “You can see through it, it’s incredibly light and in your hand it almost feels like there’s nothing there.”
Because that air is trapped in nanoscale pockets, aerogels can have even lower thermal conductivity than air itself—an example of how nanoscale structure can dramatically change physical behavior.
That focus on physical behavior also shapes Barisik’s approach to artificial intelligence.
“We don’t treat AI as something to believe in blindly; it’s a tool,” Barisik said. “The word ‘intelligence’ is used too loosely, and that can create blind trust in the output. So we don’t just give it data and hope for the best; we guide it with the physics of what we’re studying. Based on that, we feed it the multiscale modeling data we generate so it learns how a material’s properties change with its structure.
“AI can give an answer, but I want it to explain its decision. If it says, ‘That’s a cat,’ I want it to say why—‘Because it has whiskers, pointy ears, a fluffy tail.’ Then I can trust it. We want to move away from black-box answers toward explainable AI guided by physics.”
Taken together, the approach reflects a broader goal of Barisik’s work: using machine learning not as a shortcut, but as a tool that helps engineers better understand and design complex materials.
“Dr. Barisik’s work reflects the kind of research we value in the College of Engineering and Computer Science—research that is rigorous, relevant and grounded in fundamentals,” said Dr. Kumar Yelamarthi, dean of the UTC College of Engineering and Computer Science. “The repeated journal cover features speak to the care and purpose behind the work and the academic satisfaction that comes from doing it well.”
Learn more
College of Engineering and Computer Science
UTC mechanical engineering professor Murat Barisik receives prestigious NSF CAREER Award
