jim rozen writes:

Interestingly the solubility of CO2 seems to *reduce* if
the temperature is dropped enough. Somebody once attempted
to cool his warm bottle of coke by dipping it into an
open topped dewar of liquid nitrogen, when I was working
in boston.

All the CO2 comes out at once, the gyser of coke hit the
ten foot ceiling in the lab.

Despite what you observed, the solubility did not reduce. Le Chatelier's
principle states that solubility of CO2 in water increases as
temperature decreases.

However, that assumes a freezing phase-change is not involved. What you
observed sounds like the LN2 started to freeze the beverage at the
interface to the bottle. CO2 does come out of solution when carbonated
water freezes. And that will happen quite dramatically with the rapid
heat transfer of immersion in LN2. Brrr!

It is also the case that a warm soda has several volumes of CO2 more
than its equilibrium saturation. So if you had a warm soda, with the
sides suddenly frozen, you would get CO2 abruptly effervescing from the
freezing area, which in turn would trigger the warm area (supersaturated
with CO2 if recently uncapped) to violently effervesce.

Capped soda is charge with 4 volumes of CO2, while warm soda is flat
soda (when in equilibrium). So if you uncap a warm soda, you have an
unstable system that can be triggered to turn into a foam geyser. Thus
the locker-room champagne celebrations.