Hallo Steve,

when I studied physics, we had to develope radian hard microstrip detectors for ATLAS fundamental particle accelerator at CERN.

Major effect of radiation is the change of doping concentration. This can even lead to inversion, where n-doped material can mutate into a p-doped one, and vice versa.
Another effect is drastical increase of leakage current across pn-junctions.

There are many ways to harden chips: Other doping ions are used, bigger junctions with more volume, adding of buried layers, which capture injected charge otherwise causing leakage through pn-junction, etc.

Another remedy against radiation damage is to use semiconductors with much higher Z (numbers of nucleons in atomic core). So, today many radiation hard applications use GaAs-semiconductors. We also used this material for microstrip particle detectors.

Even digital logic families totally built in GaAs are available today. I think company producing these is called 'Gigabit Logic'. Another advantage of GaAs is their extreme fast switching. If you need propagation delay times in the 10picosecond range...

Unfortunately, power supply current is universes higher compared to advanced CMOS process: While a flip-flop in CMOS only needs some µW, GaAs needs more than thousand times higher supply current. When I left university, no low supply current GaAs process was available to build circuits like transimpedance amplifiers and other radiation hard front-end stuff.

At ATLAS project radiation is so high, that a silicon microstrip detector would fail after only ONE day, equally how radiation hardend chips were made!! GaAs provides an extreme advantage over silicon detectors. But GaAs front end signal conditioning circuits draw much too much supply current and couldn't be used either.

Bye,
Kai