The beneficial role of nitrogen as a blocking barrier against impurity penetration was the initial driving the interest in incorporating nitrogen into dielectrics. Besides this, nitrogen reduces the leakage current due to an increase in the dielectric constant and thus prolongs the lifetime of silicon dioxide-based dielectrics. However, the detrimental effect of Negative Bias Temperature Instability (NBTI) has been reported to dramatically increase with the incorporation of nitrogen. Recent investigations have revealed that NBTI is a consequence of complex trapping and detrapping behavior. The underlying trapping mechanism is based on bistable defects that feature one stable as well as one metastable configurations in both the neutral and the positive charge state. Furthermore, they have certain energy differences between their metastable and equilibrium configurations that are small enough to allow for high transition rates. These are restrictions to the defect properties, which allow the identification of defects on the basis of atomistic simulations.
In literature, no defect has been discovered that fulfills all of the above requirements. For instance, the oxygen vacancy, the precursor of the E' center, is found to have a trap level close to the oxide valence band edge. On the search for a defect with suited properties, a promising candidate, consisting of a nitrogen and a silicon dangling bond, has been encountered. This Nitrogen-Related Complex (NRC) also involves a threefold coordinated silicon atom that resides inbetween these dangling bonds and saturates either of them. The hole capture levels of the neutral variant of the NRC is situated far below the substrate bandgap so that no appreciable amount of charge can be injected into the dielectric. However, there exists hole capture levels for a negatively charged NRC that are located in a region close to the substrate bandgap. Futhermore, their stable configurations differ by only a small energy that allows for transitions over thermal barriers within experimental times. For this reason, the NRC can be viewed as one promising candidate explaining NBTI. When calculating its ESR signal, it might be identified in real dielectrics that would then help to improve the processing of future devices.