During this optimization the values of the optimization parameters were chosen and entered into the process flow specification by the user. Then the simulations were performed under control of the SFC. Using the VISTA optimization module [39] the values of the optimization parameters can be determined and entered into the process flow automatically.
At each optimization step the derivatives of the optimized quantity with respect to the optimization parameters have to be calculated to determine the parameter set for the next optimization step. For each of these derivatives is calculated by the forward difference method after a complete simulation run including process and device simulation.
For this particular example the demand on computational resources required for performing a simulation of the complete process flow and the device simulations for determining the on-resistance and the maximum electric field is very high. Therefore with the presently available computational hardware an automatic optimization using the VISTA framework does not seem feasible for this example.
Instead an appropriately chosen design of experiment (DOE) [40] should be used to create a response surface model (RSM). The response surface model is then used in the optimization loop instead of the process and device simulation. After identifying the optimization parameters which have the strongest influence on the optimized quantity and limiting the variation of the optimization parameters to an appropriate interval an automatic optimization could be performed to find the exact values of the optimization parameters.