CMOS NBTI degradation and recovery behaviors in a wide temperature range

Difference analysis method for negative bias temperature instability lifetime prediction in deeply scaled pMOSFETs

“…We found that the fluctuation-related component is not similar to the conventional NBTI degradation, which follows a power law against stress time. 1,27,28) Here, the time exponent (n) of ΔV th,noise (t) is approximately zero and ΔV th,noise (t) exhibits a parallel increase for different V gs values, which are well within the voltage stress range typically used for Si MOSFETs. The explanation for the time-independent behavior of ΔV th,noise is that there is no new fluctuation-related trap generation under the constant stress V gs and temperature.…”

“…The possible fluctuation source may originate from the as-grown defects with the energy level near the Fermi level of the Si substrate. On the basis of earlier works and understanding, [27][28][29][30] at a short stress time, I d 2) obtaining ΔV th,noise (t); 3) extracting ΔV th,average (t); 4) measuring several samples; 5) measuring the voltage and temperature dependences of ΔV th,noise (t) and ΔV th,average (t); and 6) lifetime prediction. degradation under NBTI is dominated by as-grown oxide defects near the Femi level of the substrate.…”

“…The negligible temperature dependence of ΔV th,noise shows good consistency with the physical properties of as-grown defects. 1,27) Previous works demonstrated that the temperature activation energy (E a ) of as-grown defect-induced NBTI degradation is quite small with a value of about 0.03 eV. 27) A TAT model based on hole tunneling via the oxide traps can explain the weak temperature dependence of ΔV th,noise .…”

“…1,27) Previous works demonstrated that the temperature activation energy (E a ) of as-grown defect-induced NBTI degradation is quite small with a value of about 0.03 eV. 27) A TAT model based on hole tunneling via the oxide traps can explain the weak temperature dependence of ΔV th,noise . On the basis of these results, the as-grown defects are responsible for the fluctuation-related component extracted by the difference method.…”

Physical understanding of hot carrier injection variability in deeply scaled nMOSFETs

“…BTI variability has been widely investigated. [11][12][13][14][15][16][17][18] It is reported that NBTI variability could be accurately described by the number of Poisson-distributed defects in the oxide layer and the distribution of each individual defect. 19,20) Several studies on HCI variability [21][22][23][24][25][26][27][28] have proposed that the model for HCI variability could be similar as that for BTI variability.…”