7.2  Electric Field Dependence

Experimentally it was confirmed that hot-carrier degradation in MOSFETs is highly electric field dependent (cf. Figure 7.3)  [172164170], for both short- and long-channel devices. Thus the influence of the electric field and channel length on HCD is uncorrelated. This is due to the constant field scaling applied by the industry  [173] and can be attributed to the fact that even a few nanometers of silicon are sufficient for charge carriers to be accelerated or otherwise high drain currents in scaled devices could not be sustained.


PICT
Figure 7.3: Hot-carrier stress assessed as ΔId at room temperature for various drain voltages. The data has been recorded for n-channel MOSFETs with an EOT of 16nm at a gate voltage of 2.0V. With increasing drain voltage and thus channel electric field, the hot-carrier induced damage is increasing (arrow)  [174].

The significant lateral field dependence inspired the first lucky electron model  [164], which assumed that the lateral electric field is the driving force of hot-carrier degradation (field driven paradigm)  [164]. Systematic investigations based on Fowler-Nordheim  [175] tunneling stress, hot substrate and channel charge carriers revealed that the carrier energy rather than the electric field is the driving force behind hot-carrier degradation (energy driven paradigm)  [176177]. However, charge pumping experiments revealed that neither the peak of the electric field nor the peak of average kinetic carrier energy coincides with the experimentally found maximum of the created interface defect density Nit  [178] (cf. Figure 7.4). Thus, the information on the electric field and average carrier energy is not enough to adequately evaluate the the hot-carrier induced Nit profile.


PIC

Figure 7.4: The peak positions of various quantities used to model hot-carrier degradation in comparison to the extracted Nit peak at the drain side of a MOSFET channel. The figure shows that neither the electric field nor the average carrier energy maximum coincide with the Nit peak and thus cannot be used as the driving force behind HCD. Adopted from  [170].