IN-SITU SENSING & CONTROLLING OF HIGH TEMPERATURE MANUFACTURING PROCESSES...

METAL-ORGANIC CHEMICAL VAPOR DEPOSITION (MOCVD) IS ONE OF THE LEADING TECHNOLOGIES FOR HGCDTE WAFER PRODUCTION, THE MATERIAL OF CHOICE FOR IR FOCAL PLANE ARRAY FABRICATION. CURRENTLY, THERE ARE NO CLOSED FORM MATHEMATICAL MODELS FOR HGCDTE WAFER GROWTH USING MOCVD. MOCVD GROWTH OF HGCDTE INVOLVES NON-LINEAR RELATIONSHIPS BETWEEN REACTOR CONTROL PARAMETERS AND MATERIAL GROWTH OUTPUT PARAMETERS (VIA OPTICAL SENSORS). THIS IS PARTLY DUE TO THE FACT THAT CHANGES IN MERCURY GAS FLOW RATES HAVE NON-LINEAR EFFECTS ON REACTION RATES. IT WOULD BE BENEFICIAL IF A CONTROLLER COULD LEARN THE NON-LINEAR RELATIONSHIP BETWEEN CONTROL PARAMETERS AND OUTPUT SENSORS FROM THE GROWTH PROCESS. NEURAL NETWORK LEARNING MODELS OFFER THE TECHNOLOGY FOR THE DEVELOPMENT OF CONTROLLERS BY INTERACTIVELY OBSERVING THE BEHAVIOR OF THE GROWTH PROCESS AND BUILDING AN INTERNAL MODEL OF MOCVD WAFER FABRICATION. THIS PROPOSAL ADDRESSES THE DEVELOPMENT OF A NEURAL NETWORK ARCHITECTURE FOR IN-SITU CONTROL OF THE METAL-ORGANIC CHEMICAL VAPOR DEPOSITION (MOCVD) PROCESS FOR WAFER GROWTH. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - THE RESULT OF FINER CONTROL OF WAFER GROWTH IS THE FABRICATION OF HIGHER QUALITY WAFERS. CURRENTLY, MOCVD IS ONE OF THE LEADING TECHNOLOGIES FOR HGCDTE WAFER PRODUCTION. THE HGCDTE MATERIAL IS OF GREAT IMPORTANCE IN THE DEVELOPMENT OF FOCAL PLANE ARRAYS, WHICH ARE CRITICAL COMPONENTS OF FLIRS FOR GROUND, AIRBORNE, ANDCOMBAT WEAPON SYSTEMS. THE FABRICATION OF HIGHER QUALITY HGCDTE DETECTORS WILL RESULT IN THE DEVELOPMENT OF MORE SOPHISTICATED FOCAL PLANE ARRAYS AND, IN TURN, THE CONSTRUCTION OF FLIR-BASED WEAPONS CAPABLE OFFAR GREATER STANDOFF RANGES.