Neural network control of variable-frequency microwave processing of polymer dielectric curing

Variable-frequency microwave (VFM) curing can perform the same processing steps as conventional thermal processing in minutes, without compromising intrinsic material properties. With increasing demand for novel dielectrics, there is a corresponding demand for new processing techniques that lead to comparable or better properties than conventional methods. VFM processing could be a viable alternative to conventional thermal techniques. However, current limitations include uncertain process characterization methods, a lack of reliable temperature measuring techniques, and limited control methodologies. This research focuses on the development of a neural network controller for curing low-k polymer dielectrics on silicon wafers in a VFM furnace. The neural network controller exhibits temperature set point control with percent error of 7% when compared with the target trajectory.