The expansion and contraction of the materials is directly related to the process utilized, and how precise that process is.
Hmmm..Alan, Alan, Alan,,, how can you say that??? tisk tisk tisk...

What you meant was the "
differential expansion rates between the different materials". The process cannot control the expansion rates per se, nor the absolute differences in expansion, but it can prevent failure as a result of a cooling gradient caused by excessive rates and the thermal mass.
Hey there, long time no talk Alan.....hope all is well with you.
The glass to metal hermetic seal can indeed be compromised by the thermal expansion differences between the metal and the glass. This is why most military hybrids utilize kovar as a feedthrough, as it has a better match to the glass TCE. One drawback to kovar is it is magnetic.
Using a computer controlled ramp rate will reduce (if not remove) the possibility of breaching the hermeticity during cooldown or warmup.
Opaqueice: if you look up "martensite finish temperature", you will find some information on this type of process. In steels, there is a diffusionless transformation of austenite to ferrite, this being a FCC to BCC transformation. At room temperature, the conversion is incomplete. There will be a low temperature called the "martensite finish temperature, where the transformation is complete. This is the "type" of transformation utilized in cryogenic treatments.
Barrett, Nix, and Tetelman discuss this on page 311 of "The principals of engineering materials", 1973, Prentice-Hall.
Cheers, John