![]() |
The kink in the drain current does not appear because both contributing effects are suppressed, namely the body effect and the amplification of the impact-ionization current through the bipolar effect. As expected, a positive output conductance is obtained.
The strong influence of impact-ionization can be seen in the corresponding bulk currents
(Fig. 4.7). With impact-ionization included the expected result of a body current
flowing out of the transistor is obtained (
). But if in contrast impact-ionization is
neglected there is a body current flowing into the device (
), which is
several orders of magnitude smaller. It is to note that the real substrate current due to
impact-ionization has the opposite sign. The situation of a positive substrate current shows that even in
this bulk MOSFET hot electron diffusion into the p-body occurs. Note that this is a
prediction of the energy transport model only, and is not confirmed by experimental data.
![]() |
To estimate if the resulting drain current obtained by the drift-diffusion simulation using impact-ionization shown
in Fig. 4.2 is really caused by the increased body potential, simulations using
the same transistor but with a body contact applied (Device 2) were made. The results are
shown in Fig. 4.8 where the curve from Fig. 4.2 which used
impact-ionization is depicted again--this time the full
range is shown.
![]() |
From Fig. 4.3 it can be seen that the body potential is shifted from
at
to
at
resulting in a total shift of
. This
voltage is now applied at the body contact of Device 2. In this case the source-body diode
(and at small
even the drain-body diode) is biased in forward direction
yielding a negative drain current of
at
. Accounting for this negative current offset total agreement with the curve
using impact-ionization is obtained at
.
M. Gritsch: Numerical Modeling of Silicon-on-Insulator MOSFETs PDF