Online NBTI-induced partially depleted (PD) SOI degradation and recovery prediction utilizing long short-term memory (LSTM)

Abstract

Negative bias temperature instability (NBTI) is one of the major aging effects that account for degradation or even failure of SOI pMOSFETs. However, several modeling methods for aging prediction show limitations in some conditions. Therefore, combined with the recent development of online diagnose and prediction application, a long short-term memory (LSTM) time series data-driven model is proposed for NBTI-induced partially depleted (PD) SOI devices aging prediction, including both stress and recovery phases. The results demonstrate improved accuracy and universality while holding a reasonable time consumption.