A Novel Model of the Aging Effect on the ON-State Resistance of SiC Power MOSFETs for High-Accuracy Package-Related Aging Evaluation

Package-related aging (PRA) defects can cause severe failures in SiC <sc>mosfet</sc>s, so an aging evaluation method is essential. <sc>on</sc>-state resistance (<italic>R</italic>_ON) is the most widely used aging indicator for SiC <sc>mosfet</sc>s. However, its performance is limited by the <italic>R</italic>_ON swing as aging levels develops. This phenomenon is contradictory to the conventional model of aging effect on <italic>R</italic>_ON. In this article, a novel aging effect model is proposed, where <italic>R</italic>_ON is divided into two components named <italic>R</italic>_AG and <italic>R</italic>_SW. <italic>R</italic>_AG is related to the PRA and <italic>R</italic>_SW is related to the chip-related aging (CRA). In this way, the swing phenomenon can be eliminated. A decoupling method based on the east square method is proposed to calculate <italic>R</italic>_AG, <italic>R</italic>_SW, and <italic>V</italic>_TH. This method can decouple the effect of PRA and CRA on <italic>R</italic>_ON. The proposed aging effect model and decoupling method are verified by experiments based on a thermal cycle accelerated aging test. Experimental results indicate that swing in <italic>R</italic>_SW is the root cause of the swing in <italic>R</italic>_ON. Finally, an online aging evaluation strategy is proposed. The accuracy of the PRA evaluation is improved by five times compared with conventional methods, and the CRA evaluation can be conducted simultaneously.