) or not is resolved in the following.
It was already shown in Chapter 2 that the experimental access to BTI is extremely challenging due to the rapid recovery of the degradation, setting in as soon as the stress is removed. In particular, it has been observed that when after NBTI stress the device is positively biased, a considerable part of the recoverable component is lost [94, 6, 62, 77]. Until recently, this has been explained by the detrapping of holes [6, 62], while interface states have been assumed to only change their occupancy but do not recover. Unfortunately, this makes any experimental accessment of the defects contributing to BTI very complex.
A quite striking result obtained with on-the-fly charge pumping (OFIT) measurements is that in contradiction to charge pumping (CP) measurements, OFIT data suggest a considerable amount of fast initial recovery of interface states. It has to be noted that this fast initial recovery is not explicitly measured, but is only inferred from the differences between the last stress and the first recovery measurement. The correctness of this assumption heavily affects the understanding of the short-term behavior of interface states. It is hence a major topic to clarify this subject as it is the prime requisite for the development of a reliable model.
By performing CP and OFIT measurements on different technologies as
decribed in Chapters 2.4 and 2.5, the issue whether interface states do recover
quickly () or not is resolved in the following.