The client had a simulation of the equipment and chemical/thermal process in part of a silicon wafer manufacturing system. The simulation was sufficiently accurate to be useful but required manual tuning to match test data for specific machines and materials. The run time was also too slow to be used for making regular adjustments or for tasks like real-time process control.
Using a combination of SmartUQ’s advanced spatial-temporal emulation and model calibration tools, the simulation response was emulated using simulation runs as the training sample. The emulator predictions were then validated and used for model calibration.
The emulation process and tools demonstrated in this project enable the customer to automatically tune the simulation model to new process variants and individual pieces of equipment. The emulators were also fast enough to allow real time process optimization.
