FMRG: Cyber: Diamond growth Informed by Autonomous Manufacturing for Opportunities in Next generation workforce Development (DIAMOND)

About This Grant

As manufacturing in the USA becomes increasingly automated by artificial intelligence (AI), semiconductor chip manufacturing hasn’t yet seen widespread AI automation. Currently, semiconductor manufacturing research requires human operators to perform tasks, both in the lab and at the factory. Adding to the problem with chip manufacturing, defect formation in some types of semiconductor chips can ruin their performance, but defects can typically be detected only after the process has been completed. At that point, the chip must be scrapped or repurposed. This problem is especially challenging with promising materials such as diamond, which could transform power transmission and other advanced applications due to its unparalleled thermal and electrical properties. This research endeavors to use AI monitoring together with innovative defect detection and process analysis to automate these processes. This work could lead to an entirely automated cybermanufacturing approach to semiconductor chips, using diamond as the flagship material. The work also includes objectives to assist the existing and future manufacturing workforce to understand, learn, and adapt to AI processes, so as to translate existing jobs into next-generation Industry 4.0 careers. The research will be integrated into a forward-thinking training program in semiconductor manufacturing for graduate students and senior-level undergraduates at MSU and local community colleges, as well as events and opportunities for the existing manufacturing workforce in Michigan. This work’s impact extends beyond chip manufacturing into sectors such as energy, electronics, healthcare monitoring, and defense. Manufacturing ultrawide bandgap (UWBG) single crystals requires a leap in scientific understanding, because growing these crystals with sufficiently high quality for power electronics is at a plateau of development for some of the most promising materials -- namely diamond and aluminum nitride (AlN). This work addresses several of the major scientific and technical challenges associated with building a full understanding of single crystal epitaxial growth of diamond and other UWBG materials. First, the proposed work combines in-situ hyperspectral and infrared spectroscopic imaging, ex-situ and comprehensive 3D material characterization, and physics modeling of microwave plasma assisted chemical vapor deposition (MPACVD) diamond growth to map between reactor telemetry, process, and resulting crystal quality. This creates an industry-accessible diagnostic solution to understanding the processes in a sealed reactor environment. Second, incorporation of this diagnostic data into a deep learning – based AI model will allow fundamental research into how to accommodate the vast variability and rapidly shifting spatial and temporal aspects of crystal growth. Third, the work will include research on strategies to improve the robustness of the AI model in the materials science context, which is a novel endeavor. Finally, the work will uncover key findings related to translation of the AI model for diamond growth to other epitaxial crystal manufacturing systems, namely for AlN. Beyond these technical objectives, the research will address education and workforce development via a robust plan that incorporates outreach, curricular updates, and training across a spectrum of participation levels from students through existing professionals. This effort emphasizes AI-enabled manufacturing roles, particularly in semiconductors, through collaborations with educational institutions and industry. This Future Manufacturing award is co-funded by the Division of Materials Research (DMR) in the Directorate for Mathematical and Physical Sciences (MPS), and the divisions of Electrical, Communication and Cyber Systems (ECCS) and Engineering Education Centers (EEC) in the Engineering Directorate (ENG). 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.