I worked at Moxtek in Orem Utah for a period of time in the lat 90's, they build low noise JFETs, assemblies containing the JFETs and other items for XRF (X-ray Fluorescence) equipment. When the JFET assemblies are operating they're cooled to a temperature close to liquid nitrogen the only reason they're not at the same temperature as the liquid nitrogen is they have a tiny heater that keeps them at their optimal operating temperature. When not operating & turned off their temperature gets very close to the tempo of the liquid nitrogen. One of the tests I had to run was on some Honeywell Mil-Spec IR LEDs, was simulating the cold /hot cycles the equipment would go through. The diodes had to with stand being dunked into liquid nitrogen for a specified number of times. I looked at putting together a mechanical dunker for the tests, and decided I could get it done faster by doing it by hand. While the Diodes passed with flying colors any temperature color shift didn't make them visible.