FuSe2: Topic 3: AI-Enhanced Material-Device Codesign of Boron Arsenide as the Next-Generation Semiconductor

Nontechnical description High-quality semiconducting materials have been the driving force for the information and computing revolution in the past century. The rapidly increasing integration density, operational speed and power of electronic devices require highly efficient dissipation of heat from operating devices in order to reduce the risk of overheating and thermal failure. This project aims to develop a next-generation semiconducting material – cubic boron arsenide (BAs) – with intrinsically high electrical and thermal conduction properties. Preliminary studies have suggested that BAs can conduct heat at least 10-times better than silicon, in addition to having superior electrical conduction and optoelectronic properties. Despite its promising properties, currently, high-quality BAs can only be made in very small crystals using an inefficient growth method. The research team plans to develop new methods to grow BAs in form factors that are relevant for practical applications, such as thin films and large single crystals, and understand how defects can modify its properties. In addition, the research team focuses on developing a co-design platform to optimize material property and device design simultaneously using artificial intelligence. This study aims to demonstrate prototypal BAs devices building upon these fundamental advancements. This project also supports educational and research activities to train the next-generation semiconductor industry workforce with combined skills in theory and experiment. The team plans to achieve this goal by directly training graduate student researchers, incorporating research progress into new hands-on courses, hosting undergraduate researchers with a diverse background, and engaging industrial partners. Technical description The overarching goal of this collaborative project is to develop BAs as the next-generation semiconductor that combines an ultrahigh thermal conductivity, high bipolar charge mobilities, and a long hot photocarrier lifetime for microelectronic and optoelectronic applications. The project aims to bridge the knowledge gap between fundamental electron and phonon interaction properties in BAs and their impact on coupled electrical and thermal transport and practical device applications. The core strategy to achieve this goal is rooted in the principle of material/device codesign enabled by a "digital twin" of BAs-based devices that is powered by physics-integrated deep learning. To complement this platform, the research team plans to develop a mesoscopic modeling framework to establish theoretical understanding of electrical and thermal transport properties in BAs, synthesize and characterize high-quality bulk crystals and thin films of BAs with state-of-the-art techniques including high-pressure flux growth and molecular beam epitaxy, and obtain experimental data and knowledge that can feed back and refine the digital twin. Furthermore, the research team aims to develop novel experimental methods capable of directly probing electron and phonon interaction with point and extended defects at the atomic level and systematically examine doping strategies and heterostructures to enable practical device applications of BAs. The project paves the way for BAs to become a practical new semiconductor material and provides fundamental insights into transport and defect physics in emerging semiconductors with unusual electron and phonon structures. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.