On the frequency dependence of the bias temperature instability

An accurate understanding of the duty-factor and frequency dependence of the bias temperature instability (BTI) is essential since it can transfer to significant lifetime benefits. Contradictory reports regarding the frequency dependence have been published in literature, with particularly older studies and those interpreted in the framework of the reaction-diffusion model claiming BTI to be frequency-independent. Newer studies, on the other hand, clearly show a frequency-dependent component of BTI. Here we discuss this frequency dependence using the recently suggested capture/emission time map model, which explains BTI as a collection of a large number of independent first-order reactions of defects with two states. While this model can explain most features observed in DC and low-frequency AC stress experiments, it does not fully capture the experimentally observed frequency and duty-factor dependence. However, when a more accurate defect model is considered, which describes charging of the defect via an intermediate state, good accuracy with experimental data is obtained.