Alas the relatively easy to obtain lambda sensors used for oxygen measuring in car engine control applications only work at very high temperatures... There are ppO2 sensors used by divers apparently, but I know little about them. For the mice cages wouldn't a CO2 sensor be a better method?