Statistical Modeling of GaN HEMTs by Direct Transfer of Variations to Model Parameters

A statistical physics-based model for GaN HEMTs with a direct transfer of variations to model parameters is proposed in this paper. This statistical approach includes the collection of available data from process control monitor (PCM) on-wafer measurements and their transfer to the respective physical model parameters. Three main variables are analyzed and used in the simulations. Those are the semiconductor sheet resistance R sh , the ohmic contact resistance R c and the threshold voltage V th . The model’s performance is validated by comparing Monte Carlo large-signal simulations and measurements of several devices on the same wafer.