INTELLIGENT CONTROL OF SEMICONDUCTOR MANUFACTURING PROCESSES

AN INTELLIGENT EQUIPMENT ARCHITECTURE FOR CLUSTER AND IN-SITU SEMICONDUCTOR MANUFACTURING PROCESSES IS PROPOSED. IT HAS TWO COMPONENTS: (1) MANUFACTURING EQUIPMENT - EQUIPMENT HARDWARE AND SENSORS, AND (2) INTELLIGENT EXPERT WORKSTATION - A HYBRID AI/NEURAL NETWORK EXPERT SYSTEM. THE INTELLIGENT CONTROL SCHEME IN THE INTELLIGENT EXPERT SYSTEM HAS THE FOLLOWING FEATURES: (1) EQUIPMENT/PROCESS MODELING BY NEURAL NETWORKS, (2) NEURAL NETWORK HARDWARE EMULATOR AS CONTROLLER, (3) SENSOR-BASED CONTROL, (4) LOCAL SIMULATION BY NEURAL NETWORKS FOR CONTROL DESIGN, (5) HEURISTICS FOR CONTROL DESIGN BY RULE-BASED SUBSYSTEM, (6) LEARNING, MONITORING, AND CONTROL BY THE SAME MODELS, (7) SENSOR DRIFTS AND NOISES MAY BE MONITORED BY NEURAL NETWORKS, (8) TIGHT COUPLING OF MANUFACTURING LINE AND ITS SIMULATOR. ANTICIPATED BENEFITS/POTENTIAL COMMERCIAL APPLICATIONS - A UNIQUE CONTROL SCHEME WITH REAL-TIME, ADAPTIVE, AND MODELING CAPABILITIES IS CRUCIAL TO VLSI AND ULSI DEVICES AND ADVANCED ELECTRONIC MATERIAL PROCESSING. THIS INTELLIGENT EQUIPMENT ARCHITECTURE FOR CLUSTER PROCESSING CAN ACHIEVE: (1) PROCESS UNIFORMITY, (2) HIGHER YIELD, (3) INTEGRATION OF PROCESS MEASUREMENT AND CONTROL, (4) PROCESS PROGRAMMABILITY, (5) MIXING TECHNOLOGIES, AND (6) EFFICIENT FABRICATION. KEY WORDS - INTELLIGENT CONTROL, EQUIPMENT/PROCESS MODELS, SENSOR-BASED CONTROL, INTELLIGENT EQUIPMENT ARCHITECTURE, MODEL-BASED CONTROL, CLUSTER PROCESSING, NEURAL NETWORK CONTROLLER, HYBRID AI/NEURAL NETWORK EXPERT SYSTEM.