Accurate prediction of hot-carrier effects for a deep sub-μm CMOS technology based on inverse modeling and full band Monte Carlo device simulation

The ultimate goal of device modeling is the accurate prediction of device characteristics before the technological realisation. Due to insufficiencies of process simulation and to a lesser extent of device simulation this goal has not yet been reached. The aim of this work is to reduce the number of wafers in split lots used to investigate the effects of device parameter variation by predicting these effects with device modeling. Our approach is based on a device model (geometry and doping profile) which is extracted for one wafer by inverse modeling. This model is then used to predict the effects of parameter variation by device simulation with Galene III and our full band Monte Carlo (FB-MC) program Falcon. In this work we apply the new method to a state of the art 0.25μm-CMOS technology and validate the approach by comparison with experiment.