A detailed evaluation of model defects as candidates for the bias temperature instability

Despite its long research history, the bias temperature instability (BTI) is still not fully understood. Recent advances on both the experimental and theoretical side have deepened our understanding of the phenomenon, but the microscopic origin is still unknown. We report on a detailed evaluation of atomistic models of the oxygen vacancy and the hydrogen bridge defects in SiO₂ as candidates for the defect responsible for the BTI. For this purpose, time constants are calculated using a combination of atomistic and semiconductor device modeling. These time constants are then compared to electrical measurement data obtained from BTI experiments on individual defects in small-area MOS transistors. The inherent uncertainty in the energetic position of the energy levels in the density functional calculation with respect to the device simulation is accounted for using an empirical energy shift. Very good agreement with the experimental data is found for the hydrogen bridge defect, while for the oxygen vacancy severe discrepancies between the predicted behavior and the experimental observation arise.