(image) (image) [Previous] [Next]

Modeling of Defect Related Reliability Phenomena
in SiC Power-MOSFETs

6.2 Outlook

Irrespective of the progress made in the field for describing BTI and TAT based on the defect centric NMP modeling approaches, there is still a number of potential improvements to the presented methods and models to fully model SiC MOSFETs reliability issues and the most important considered are outlined in the following:

  • • The post-stress recovery of (math image) for NBTI characterization is commonly studied on p-channel MOSFETs. However, these devices are rarely needed in power switch applications and therefore such studies are rare for SiC MOSFETs. Thus, extending NBTI studies to SiC pMOS would allow to capture a larger fraction of defects in the lower half of the SiC bandgap and in the vicinity of the valence band edge due to the significantly slower charge transfer kinetics obtainable.

  • • A previously reported switching cycle dependence of BTI in SiC trench MOSFETs [91, 93] cannot be captured by a two-state NMP model, and requires further studies to reveal the underlying mechanism, e.g. as presented in [95].

  • • Recently, optical emission following bias switches from the accumulation to the inversion regime has been correlated with BTI in SiC trench MOSFETs [286, CSJ1]. The characterization of radiative transitions contributing to BTI allows for detailed studies of the parameters of the involved defects, e.g. charge transition levels.

  • • A single defect study, employing TDDS or RTN measurements, could reveal the stochastic properties of charge transfer kinetics and allow for a more detailed investigation of the defects NMP parameters. As in Si-based MOSFETs an extension to a four-state NMP model could be used to explain the underlying reactions. Although there is no obvious obstacle for an experimental scaling of the channel geometry, such studies have not been presented for SiC MOSFETs.

  • • The new TAT modeling approach allows studying BTI and TAT in parallel within one technology. However, for this appropriate test structures have to be designed with large enough gate area to measure small leakage currents and smaller structures that allow for detailed BTI studies.

  • • Additionally, the TAT model may be applied to explain leakage currents in other technologies, such as thick inter-layer dielectrics. An extension of the model to include defect generation, i.e. time evolution of defect densities, would extend its applicability to SILC and to impact ionization and allow to explain TDDB.