Next: 3.4.2 Analysis of WCC within Simplified Transport Schemes Up: 3.4 Impact of the Carrier Distribution Function on HCD Modeling Previous: 3.4 Impact of the Carrier Distribution Function on HCD Modeling
(a)
(b)
(c)
|
The electron AIs computed within the MC-, HD- and DD-based models are plotted in Figure 3.13 (note that all the findings are similar for the hole acceleration integral). The driving force of the degradation in the DD-based model is the electric field. However, the DF follows the electric field with a certain delay [93]. As a result, the maximum of the MC-based AI is shifted towards the drain with respect to the result of the drift-diffusion scheme. As demonstrated in Section 3.1 starting from source to drain first the maximum of the carrier average energy appears, followed by the electric field peak, and finally the maximum AI when calculated with the Monte-Carlo method. In Figure 3.13 the maxima of corresponding acceleration integrals are in this order.
(a)
(b)
(c)
|
The Nit profiles calculated employing different transport schemes also confirm this behavior, see Figure 3.14. The interface state density evaluated with the HD-based model spuriously overestimates the damage as compared to DD and MC schemes. Such a trend was expected based on hot-carrier tunneling studies [170] where the tunneling process was also overestimated when DF was simulated employing the HD scheme. As a result, the linear drain current degradation predicted by the HD-based model is much stronger than those obtained employing the MC and DD approach (Figure 3.15). Finally, the DD-based model predicts ΔIdlin close to the result obtained by the MC-based model for Lch = 1.2 and 2.0um but totally fails for Lch = 0.5um.
(a)
(b)
(c)
|