I'm seeing the other point of view now. However, in that video, they prove at the 2:20 mark that the heavier the gas, the longer it takes to mix with oxygen, thus proving that the bed of CO2 can exist, it is merely temporary. It took a
half of an hour for those gasses to mix. How long does it take to transfer to a carboy?
Two minutes? I am still 100% convinced that blowing CO2 into a carboy not only replaces almost all of the air with CO2, but the siphon as well, thus minimizing exposure to oxygen.
I misunderstood the gas/weight argument. Like most home brewers, I knew not the exact science of what I was doing, I simply knew the positive effects.
Sorry for sharing bad science.
Still good advice
As mentioned on some of the other threads (or possibly upthread here), the diffusion constant is a function of the gases in question (root of the sum of the reciprocals of the masses). That's why air diffuses slowly into a heavy gas like SF6, compared to the diffusion rate of oxygen into CO2.
There's also question of how much diffusion is needed to cause issues. 1 ppm of O2 is reportedly an issue, but would be pretty hard to detect by the methods shown in these videos.
Because I'm a physicist and I should be able to work this out
, let's do a not-so-quick calculation:
The mass diffusivity of a binary mixture of O2-CO2 is 0.159 cm²/s at STP (25C, 1 atm) (this varies as absolute temperature to the 3/2 power, so is about 0.14 at 25C). I can't find a direct measurement of the diffusivity of trace oxygen in CO2 but since the opposite value (trace CO2 in air) is almost identical to the binary mixture value, let's take it to be the same. The diffusion current is J = -D dn/dx where dn/dx is the concentration gradient.
Let's take the carboy neck as an example, and say it is 5 cm long, 8 cm² in area (1.25" diameter) and that after the neck, the headspace is free to mix. We are worried about small amounts of oxygen, so lets call the oxygen concentration gradient a constant 0.21*4.23e-5 mol/cm³ / 5 cm, = 1.766e-6 mol / cm^4. The molecular flux is then 0.159*1.766e-6*8 = 2.26e-6 mol / s through the carboy neck
Assuming that that oxygen eventually gets into the beer*, the effective rate of increase of eventual oxygen concentration in the beer is given by the flux over the mols of beer in the carboy. Since the molar mass of water is 18g/mol and the density is 1000g/litre, a 5 gallon (18.9L) batch of beer has just about 1000 mols of water plus some other substances. Let's call it a nice round 1000 mols - you can adjust the values hereafter if you really want. So the final effective concentration of oxygen in the beer will rise at about 2.25 ppb O2 for every second that the carboy neck is open. Getting to the 0.1 ppm that is reported to be a target will take about 45 seconds at 25C, about a minute at 0C. 1 ppm that might be noticeable would take about seven to ten minutes.
So my
very tentative conclusion (and assuming that I'm vaguely competent at setting up and doing these calculations) is that if there's no mixing or turbulence at the carboy neck (which is unlikely), then you might get away with opening the top of the carboy for up ten minutes before closing it and leaving the beer in place. But as soon as you put something in or out, that conclusion is invalid because it will cause mixing.
However, if you are transferring, then you are pulling air in through the carboy neck, and will have fairly complete mixing in the carboy. Then the critical number is how fast the surface of the beer can take up oxygen, which is a much lower rate than the flow of O2 through the neck of the carboy. In the destination vessel, then there's a flow of CO2 out of the purged container, which will prevent diffusion unless the flow of beer is very slow. Since you are removing the beer from the carboy though and putting it in the destination vessel, the O2 that is drawn in won't have much time to do much damage.
This all feels about right - people make pretty good beer in carboys without causing terrible oxidation. That doesn't mean it can't be better with more care and that that might be necessary for some styles.
By the way, in a bucket, the surface area will about a hundred times larger, with maybe double the distance for diffusion to happen to the surface of the liquid. So the rate of oxygen uptake will be about fifty times faster, and that's without considering how much easier it is to disturb the CO2/air interface and get mixing. Definitely don't rely on CO2 blankets when opening buckets and then closing up and leaving the beer in them. Opening for transfer is probably not so bad.
*As oxygen reacts to oxidise things in the beer, there will be a concentration gradient set up from the headspace to the beer. That will mean that the beer _gradually_ pulls in more and more oxygen.