I remember discussing a similar subject before on another board. The main topic was if the CO2 content that should be considered for bottling needs to be calculated with the temp at which the beer stopped fermenting or the current beer temp. My argument was that cooling the beer will allow more CO2 to be absorbed which will come from the headspace. Another user mentioned that be barely sees the airlock water level move when he crashes his beers.
Aside from that I still think that the beer will suck in CO2 from the head space if it is done fermenting and cooled considerably. There is more room for CO2 in the beer at colder temps and it needs to be filled. I don’t know how long it would take though but am concerned that a considerable amount of air would be pulled in.
So I ran a quick calculation of the problem. I assume that there are 20 l of beer with a 2 l head space. It is sitting at 20 C (68 F) and is at equilibrium with the head space CO2. That headspace CO2 is at atmospheric pressure (~100 kPa). No more CO2 is produced by the fermentation. Now the beer is chilled to 0 C (32 F). Since the beer can absorb more CO2 it will do that. As I mentioned earlier it will pull in air and dilute the CO2 which lowers the CO2 head pressure. Once that CO2 head pressure has fallen to about 55 kPa, which corresponds to a head space CO2 content of 55%, the 0 C beer CO2 content will be at equilibrium with the head space again. As a result about 65% of 2 l = 1.3 l air must have been pulled in through the airlock.
2 l head space might be a bit generous for a secondary. If you have only a pint (500 ml) you will pull in only 325 ml ( ~0.3 qt) of air. I’m not sure if that is enough to cause considerable oxidation since a lot of brewers are doing exactly that w/o apparent stability problems. If I go a bit further and make the assumption that the 20 C beer may be oversaturated with CO2 (i.e. holds more CO2 that what it should be able to hold at its pressure and temp) it would have to hold almost twice as much CO2 in order to prevent air from being sucked into the head space while and after chilling to 0 F.
If the beer is only chilled to 10 C (50F) only 25% of the head space will be replaced with air.
These calculations neglect the temperature dependent contraction of the head space volume which only makes the problem worse.
In Keith’s case the airlock ran dry which allowed much more air to diffuse into the head space.
For those who cold crash beer in a secondary, do you see negative pressure on the airlock to the extend that air is pulled in? Or do you see CO2 escape from the beer. The latter can be evident by the formation of bubbles on the surface.