The Role of the Observation Window in the Determination of BTI Defect Time Constants and their Distribution

We propose guidelines to achieve reliable estimate of Bias Temperature Instability (BTI) signal parameters. One of the key parameters in BTI modeling is the defect time constant and its statistical distribution. However, limited attention has been paid to measurement procedures and their implications on data analysis, which may have significant impact on the construction of the model and its components, particularly the one that pertains to the time constant distribution. In the BTI measurement and characterization procedure, the observation window is finite and discrete in time, hence imposing limitations on the extraction of trap time constants and their distribution. This work discusses the role of the observation window on the extraction of time constant distribution, as well as on the determination of the time constant of individual traps. In the case of BTI, there may be a delay between the end of the stress phase and the effective start of the measurement in the recovery phase. Furthermore, sampling is discrete in time. These practical limitations of the measurement procedure imply in limitations on the time constant characterization of individual defects. Furthermore, emission events may be missed, what impacts the characterization of time constant distribution.