Researchers have harnessed artificial intelligence (AI) to create innovative nanomaterials that combine the strength of carbon steel with the lightweight qualities of styrofoam. These newly developed nanomaterials, fabricated through machine learning techniques and 3D printing, have more than doubled the strength of previously existing designs. The scientists involved in this study indicated that these materials could lead to the development of stronger, lighter, and more fuel-efficient components in automobiles and aircraft. Their findings were published on January 23 in the journal Advanced Materials. Co-author Tobin Filleter, a professor of engineering at the University of Toronto, expressed optimism, stating, "We hope that these new material designs will ultimately result in ultra-lightweight components for aerospace applications like planes, helicopters, and spacecraft, thereby reducing fuel consumption during flight while ensuring safety and performance. This could significantly lower the environmental impact of aviation. " In many materials, achieving strength often compromises toughness. For instance, a ceramic dinner plate is typically strong enough to support heavy items, but its rigidity often leads to fragility, making it prone to shattering with minimal force. This challenge also applies to nano-architectured materials, which are composed of countless tiny, repetitive building blocks that are a hundredth of the width of a human hair. Although these materials offer remarkable strength relative to their weight, they can also develop stress points that result in sudden failures.

This propensity for fracture has thus far restricted their practical applications. First author Peter Serles, an engineering researcher at Caltech, noted, "As I reflected on this issue, it became clear that it was an ideal problem for machine learning to address. " To explore improved designs for nanomaterials, the researchers simulated various geometries before analyzing them using a machine learning algorithm. The algorithm, informed by the designs produced, could predict optimal shapes that would effectively distribute applied stresses while bearing substantial loads. After finalizing these designs, the researchers employed a 3D printer to manufacture the new nanolattices. They discovered that these structures could endure a stress of 2. 03 megapascals for every cubic meter per kilogram—a strength five times that of titanium. "This marks the first instance of applying machine learning to optimize nano-architected materials, and we were amazed by the advancements made, " Serles shared. "The algorithm did not merely replicate successful designs from the training set; it learned from the modifications that worked and those that didn't, allowing it to propose entirely new lattice structures. " Moving forward, the researchers plan to focus on scaling up these materials for the production of larger components while continuing to explore even better designs through their method. The overarching goal is to engineer lighter and stronger components for future vehicles. Serles added, "For instance, if aircraft components made of titanium were replaced with this new material, the result could yield annual fuel savings of 80 liters for every kilogram of material substituted. "