The dissertation focuses on the development and experimental realization of optimization- and model-based-control methods for thermal semiconductor processe s. The new method proposed in this work uses a split approach of the underlying control problem into a detailed simulation-based analysis of the process and an optimization using only a small set of measurements. To reduce the burden of the control problem improvements on the process chamber are employed using equipment simulation in an early stage. In a second step a numerically reduced model is developed which contains both -- simulation- and experiment-based -- parameters. Optimization strategies are used then after the parameter-extraction to achieve a homogenous configuration of the target process parameters. Of equal importance is the development of control techniques with and without feedback in thermal, single-wafer process steps. Fields of application are the analysis of controller performance, the feedback control of wafers with varying optical properties, and process control under the constraint of the thermal budget. Using this approach one of the achievements is the optimization of temperature uniformity in an rapid thermal oxidation chamber.
«