I have a question about SEU (single-event-upset) in CMOS, hope it's relavent to this group.
I've been working on security/attacks on microprocessors, and is currently working on the laser illumination effect on CMOS circuits, which causes ionization in silicon and may change the state of the struck nodes, just as in a SEU caused by heavy ions.
We got some experiment results (like reliably setting a register value to 0, etc), which can be explained by the assumption that the radiation attack is more effective at turning on n-type transistors than their p-type counterparts. However I can't find any references in SEU to prove this assumption, except in some destructive effects (e.g. as for single event burnout (SEB), NMOS are more sensitive to SEB than PMOS. bacause the impact ionization rate is larger for electrons than for holes) .
I know bulk silicon has some different properties from epitaxial design in aspect of SEU. Dopant concentrations may also affect the ionization rate. Does it mean that there is no solid rule about turning on n-type and p-type with laser illumination?