Comparison of On-The-Fly, DC $I_{d}$–$V_{g}$, and Single-Pulse Methods for Evaluating Threshold Voltage Instability in High-$\kappa$ nMOSFETs

Stress-induced degradation of the threshold voltage of high-kappa nMOSFETs measured by on-the-fly and single-pulse methods is investigated. It was found that the relaxation of the stress-induced threshold voltage (V_th) shift during stress interruption (sensing) time is primarily governed by the fast detrapping of charges trapped through the fast transient charging (FTC) process. Subtraction of the FTC contribution from the total stress time-dependent V_th shift provides a practical and convenient method for eliminating a major source of V_th relaxation, which results in identical lifetime estimations by the on-the-fly and pulsed methods

I.

A SIMPLE power law relationship is usually adopted to describe the stress time dependence of the threshold voltage shift and is commonly used to predict device lifetime. However, it has been reported that in high- MOSFETs under bias-temperature-instability (BTI) stress exhibits significant relaxation during stress interruption for sense measurements [1]. This strong relaxation phenomenon was attributed to the fast charge detrapping that occurs within microseconds once the stress bias is removed [1]–[4]. Such a fast detrapping process raises questions about the validity of the conventional methodology for monitoring during BTI stress and, subsequently, the accuracy of lifetime predictions for high- MOSFETs.