Analysis of NBTI Degradation- and Recovery-Behavior Based on Ultra Fast VT-Measurements

Abstract

We present a new direct V <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</sub> measurement technique with arbitrary choice of drain voltage and a mus delay (factor of 1000 improvement) after stress. As shown this technique enables a meaningful comparison of data to theory (e.g. DeltaV <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">T</sub> measurable as response to stress times from 100mus over 10 decades in time and an analysis of recovery over 11 decades in time) which may lead to a better understanding of NBTI. A fast precursor due to bulk trapping was found to significantly influence degradation slopes for all times. Based on a physical model - standard reaction/diffusion model plus fast bulk trapping -with just 3 fit parameters experimental degradation can be well modelled and degradation slopes <frac14 as well as > frac14 (both reported in literature) can be explained. A lifetime extrapolation based on this physical model is superior to the common straight line fit. There is no satisfying agreement of recovery data with reaction/diffusion theory. Further experimental and theoretical work is going to be needed