Tunneling in an MOS transistor, as shown in the left part of
Fig. 5.1, basically can be separated into a path between the gate
and the channel, and a path between the gate and the the source and drain
extension areas [249]. Tunneling in the source and drain extension
areas can exceed tunneling in the channel by orders of magnitude. This is
related to two effects: First, instead of n-p or p-n tunneling, n-n or p-p
tunneling prevails. Second, the potential difference and thus the bending of
the energy barrier is high. This increased tunneling current in the source
and drain extension areas can be a serious problem if measurements are
performed on long-channel MOSFETs to characterize their short-channel
pendants, because the edge tunneling currents exceed the channel tunneling
current by orders of magnitude. Furthermore, there is a fundamental
difference between tunneling in MOS transistors and MOS
capacitors [250,96]. In contrast to MOS transistors, MOS
capacitors which are biased in strong inversion cannot supply the amount of
carriers as predicted by the tunneling model. This effect is termed substrate-limited tunneling, because the tunneling current is limited by the
generation rate in the substrate. In the channel of an inverted MOS
transistor, on the other hand, carriers can always be supplied by the source
and drain contacts. This effect is depicted in the right part of
Fig. 5.1.
Figure 5.1:
The different tunneling paths (channel tunneling,
source and drain extension tunneling) in an MOS transistor (left). In an MOS
transistor biased in inversion (right), tunneling electrons are supplied from
the source and drain reservoirs, which is not possible in an MOS capacitor.