In-Situ Epitaxial Defect Reduction with Augmented Instrumentation and Intelligent Control

Mainstream Engineering and Penn State University’s Applied Research Lab (PSU-ARL) will develop a novel in-situ Machine Vision system coupled to an artificial intelligence/neural network (AI/NN) that controls epitaxy process conditions and mitigate defects in real time. Enabled in part by Mainstream’s novel patented Machine Vision system, “Mantis Eye”, in Phase I Mainstream plans to establish a correlative database of epilayer defects in SiC, ex-situ, with our Machine Vision system and relate those defects to time-series classification events from real-time process monitoring and control with novel epitaxy process hardware. The database will serve as set of causal relations and form the kernel space for an AI/NN algorithm that will be conceptually developed in Phase I and implemented in Phase II. The ultimate goal of this program is to produce a new sensory/control/feedback hardware/software system for epitaxial growth tools. The system will be able to learn from a dataset generated by both ex-situ characterization and correlation to process condition deviations, and real-time, in-situ Machine Vision diagnostics of newly encountered process condition deviations that are observed to produce, propagate, or eliminate epitaxial defects.