The effect of impact ionization induced bipolar action on n-channel hot-electron degradation

The relationship between the total impact ionization rate and the measured substrate current is analyzed, using short-channel NMOS devices. It is shown that holes that are injected into the source and turn on the parasitic source-bulk-drain bipolar may actually be a significant portion of the total impact ionization current. The authors explain how the commonly used model, which ignores this bipolar effect, can lead to incorrect predictions regarding hot-electron degradation. A related criterion for maximum source-drain voltage during accelerated stress is discussed and justified.<
>