ADVANCED MATERIALS & PROCESSES | MARCH 2026 10 MACHINE LEARNING | AI In 2011, the DOE’s Lawrence Berkeley National Laboratory launched what would become the world’s mostcited materials database— the Materials Project. The platform is continuing to evolve its machine learning capabilities, with plans for enhanced computational methods and improved handling of complex materials behavior. The database now serves over 650,000 users. lbl.gov. BRIEF FOAM FLOW PHYSICS MIRRORS MACHINE LEARNING Scientists at the University of Pennsylvania recently discovered that foams flow ceaselessly inside while holding their external shape. This is a stark contrast with what scientists believed for decades—that foams behave like glass with their microscopic components trapped in static, disordered configurations. Even more unusual is that from a mathematical perspective, this internal motion resembles the process of deep learning, the method typically used to train AI systems. The finding could hint that learning, in a broad mathematical sense, may be a common organizing prin- ciple across physical, biological, and computational systems, and provide a conceptual foundation for future efforts to design adaptive materials. The insight could also shed new light on biological structures that continuously rearrange themselves, like the scaffolding in living cells. In their study, the team used computer simulations to track the movement of bubbles in a wet foam. Rather than eventually staying put, the bubbles continued to meander through possible configurations. Mathematically speaking, the process mirrors how deep learning involves continually adjusting an AI system’s parameters during training. “Foams constantly reorganize themselves,” says researcher John C. Crocker. “It’s striking that foams and modern AI systems appear to follow the same mathematical principles. Understanding why that happens is still an open question, but it could reshape how we think about adaptive materials and even living systems.” upenn.edu. AI SPEEDS ORGANIC BATTERY RESEARCH Researchers at the DOE’s Argonne National Laboratory used robotics, automation, and AI to conduct more than 6000 experiments in just five From left, John C. Crocker and Robert Riggleman spent years investigating the math that describes how bubbles in foam move and found that it mirrors how AI systems learn. months on chemicals used in organic redox flow batteries (RFBs). Such an effort would have taken five to eight years with traditional experimentation. Organic RFBs use organic molecules instead of traditional metal ions. During their study, the team made a crucial finding about these batteries: A fundamental barrier at the molecular level limits their stability. This insight is expected to inspire new directions in battery chemical research. With human-driven experiments, tackling this ambitious research would require years of coordinated global effort to investigate the vast space of organic solvents. Laboratory automation offered a key opportunity to address the challenge in a much shorter time and with significantly fewer resources. Machine learning algorithms guided the test iterations based on analysis of experimental data. This allowed the team to characterize 540 solvents by sampling just a third of them. The study’s insights could spur development of innovative use cases to make organic RFBs commercially viable. For example, organic materials could be used in grid-scale batteries for a limited time and then repurposed for other applications such as agricultural herbi- cides and materials for the chemical industry, say researchers. anl.gov. Scientists used high-throughput experiments and AI to reveal stability limits in redox flow batteries. Courtesy of Argonne National Laboratory.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==