Faculty from ECE and the University of Notre Dame have received a $1.7 million grant from the National Science Foundation to explore 2D heterostructures that demonstrate perfect light absorption, and giant piezoelectricity, using machine learning tools. The research team will be led by Prof. Tony Low, who will collaborate with ECE colleagues, professors Vladimir Cherkassky and Steve Koester, and Prof. Chris Hinkle from University of Notre Dame. The award comes from the NSF’s Division of Materials Research that supports materials discovery, and their design, synthesis, and characterization.
The isolation of atomically thin 2D graphene ten years ago introduced us to an entire family of truly 2D atomic layered materials with a wide range of electrical and optical properties, such as transition metal dichalcogenides, black phosphorus, and boron nitride, among many others. In recent years, we have learned that the mere stacking of just two such atomic layers can drastically alter its electronic properties, from turning on and off its piezoelectricity and nonlinear optical response, to emerging superconductivity, and functional properties such as magnetism and ferroelectricity. Indicating the enormous potential of such stacking, Prof. Tony Low says, “The opportunities for materials science and engineering are huge with 2D materials heterostructures.
However, stacking these 2D layers to assess their specific properties is no small matter. There are about 1000 different 2D materials that can be accessed experimentally, and simply the lowest energy configuration stacking of these 2D atomic layers into bilayers will lead to a million possible heterostructures. The permutations would become incomprehensible if we consider trilayers, teralayers, and beyond. State-of-the-art materials computation tools allow for the evaluation of the properties of a given 2D stack typically in a matter of hours. However, scanning through all these different permutations would be near impossible even with the most powerful computers.
THE ROLE OF MACHINE LEARNING
To bring the goal of evaluating these 2D layers within reach, the research team has turned to machine learning. Much like how artificial intelligence can one day potentially drive the car to its destination, machine learning tools can be employed to guide progressive materials computation towards a specific goal, such as perfect light absorption in the visible spectrum, and giant piezoelectricity.
To put the team’s endeavor in perspective, 50 years ago at an American Physical Society meeting at Caltech, physicist and Nobel laureate Richard Feynman posed : What could we do with layered structures with just the right layers? What would the properties of materials be if we could really arrange the atoms the way we want? […] I can hardly doubt that when we have some control of the arrangement of things on a small scale we will get an enormously greater range of possible properties that substances can have, and of different things that we can do.” With this NSF grant, the researchers are closer to bringing Feynman’s vision to reality.
To reach the goal, the team comprises experts in data science and the application of machine learning, materials modeling of 2D materials, molecular beam epitaxial growth of 2D heterostructures, and materials and device characterization. The successful demonstration of these new designer 2D heterostructures will usher in a new era of efficient and purposeful materials design methodology.
The Division of Materials Research (DMR) awards are designed to expand the understanding of “electronic, atomic, and molecular mechanisms and processes that govern nanoscale to macroscale properties; manipulation and control of these properties; discovery of emerging phenomena of matter and materials; and creation of novel design, synthesis, and processing strategies that lead to new materials with unique characteristics. These discoveries and advancements transcend traditional scientific and engineering disciplines. The Division supports research and education activities in the United States through funding of individual investigators, teams, centers, facilities, and instrumentation. Projects supported by DMR are essential for the development of future technologies and industries that meet societal needs, as well preparation of the next generation of materials researchers.”