A comprehensive and critical re-assessment of 2-stage energy level NBTI model

The two stage a.k.a. four energy level NBTI model has been comprehensively re-evaluated by investigating its predictive capabilities beyond the ultra-short stress duration for which it was originally validated. It is found that the model, with its default parameters, can indeed reproduce short time (~1s) stress and subsequent recovery with good accuracy. The default model however does not anticipate well-known experimental results for longer stress duration. Other combination of parameters may be used to improve prediction of long time stress, but at the cost of inaccurate recovery estimation. The capability of this model to predict AC NBTI is also explored. We conclude that the model predictions are highly sensitive to input parameters, and we find no combination of parameters that reproduces the observed DC, AC, and duty-cycle dependent NBTI phenomena.