Application of multiobjective optimizer algorithms to the design of SiC devices

This work evaluates the feasibility of optimizing SiC power MOSFETs through multiobjective optimizers (MOOs) based on modern elitist genetic algorithms. This technique is validated on a 0D problem of high practical interest — the optimization of the epitaxial drift layer, using the ionization coefficients from [1] and [2] in addition to those from [3], extensively used in literature. Then the results are confirmed with 1D TCAD, removing approximations introduced by [4] and [5]. Finally MOOs are used with 2D TCAD simulations to evaluate the tradeoff of breakdown voltage (V B ), threshold voltage (V T ), saturation current (I dsat ) and on-state resistance (r on ) as the channel profile is varied.