Is it finally possible that the effect cannot be measured because it is compensated by the
impact-ionization current? The results shown in Fig. 4.7 could suggest this. One
problem is that the device characteristics depend sensitively on impact-ionization. Furthermore, the
kink-effect occurs at higher drain voltages, such that a region with negative differential
conductance can still remain (Fig. 4.5). Therefore using just impact-ionization in
the simulation cannot be treated as a solution to the SOI problem.