Growth Of Gan Films Smart High-precision Temperature Controller

GaN has become one of the most attractive materials in wide bandgap semiconductor material. It is now the hot point of global semiconductor research. Metal organic chemical vapor deposition (MOCVD) is widely used by researchers among all methods today. Temperature is a key influencing factor of material growth by MOCVD. In this paper, temperature control system of GaN thin film deposition process is discussed in details. For control objective of reactor chamber of MOCVD with the properties of nonlinear, large heat inertia, a large temperature lag, a big time delay, multi-variable, and time-varying parameter, a hybrid Proportional (P), Fuzzy and PID temperature control system with fuzzy switch rules is presented to get high control precision. Different control modes are used to realize piecewise control strategy. The piecewise control strategy integrates merits of P, Fuzzy and PID controller, and achieves dominance complementation and performance improvement. When the temperature deviation is big, P controller works as a main controller to accelerate response speed; temperature deviation medium, Fuzzy controller works chiefly to improve damping performance and reduce overshoot; temperature deviation small, PID controller starts to make the system static performance great and meet the needs of precision. Bumpless switch between the P, Fuzzy and PID controller are guaranteed by using fuzzy inference, which solves the disturbance problem occurred in conventional threshold switchover. Then, smooth actions of three controllers are assured. Using the definition of 'fuzzy transfer function', the approximate corresponding relationship between well-tuned parameter of PID liner controller and scaling gains of fuzzy controller is established. So, the optimal parameters of fuzzy controller can be obtained by less trial-and-error and the design of fuzzy controller can be simplified. Genetic Algorithms can get globally optimal solution without any initial information. It is a high efficient optimization assembly algorithm and used to set parameters of PID controller. After further adjusting, satisfied results of the hybrid controller could be attained.