This topic comes up fairly frequently. I would like to throw out two issues I have with using gas to top up. Comments welcome.
1) One of the questions is how much gas is needed. Someone recently said that we need to remove all the air. Well that is not possible. We only need to get it down to the level that would be in the carboy normally. A normal head space would be 3 to 4 cu.in. of air. If you have one gallon of head space (231 cu.in.) then you need to dilute the air down to the equivalent of 3 to 4 cu.in. Even a heavy gas like Argon will mix when it is added, it will not go in as a blanket on the bottom. If we put in 1 gal of inert gas we will theoretically have diluted the air to half. Each additional volume will dilute by a half again. Continuing this, we need to put in 6 gal's of gas to get the equivalent air down to between 3 and 4 cu.in. Of course some of the gas will escape as you are doing this (even with argon but a lot with CO2) so you really need to put in a lot more than 6 times the head space. Any comments on how much?2) Another issue is protecting the wine with an airlock when topping up with inner gas. An airlock only has a little more that 0.5 cu.in. of liquid in it. As temperature and to a lesser extent pressure changes the volume of the contents of your carboy changes and the liquid in the airlock will swing back and fourth. If the volume in the carboy changes by more than the liquid volume in the airlock, gas will bubble in or out of the carboy. If temperature goes up, gas will bubble out of the airlock. If it goes down it will bubble in and that is not a good thing. The question here is: How much protection from temperature swings does the airlock give?
Let's say we have a normally topped up 6.5 gal. carboy. It will have 3 or 4 cu.in. of air in the head space and roughly 6.5 gal's of liquid (at 231 cu.in./gal) Air expands by about 0.0015 units/deg. F and water (sort of like wine I guess) expands at 0.0001 units/deg. F near room temp. Using these numbers you can calculate the volume swing per deg. change in temperature for the wine/air system as 0.16 cu.in./deg. If the airlock will accommodate a 0.6 cu.in. volume change, this suggests that the airlock will protect against about 4 deg. F. temperature swing.
If you do the same calculation on a 6.5 gal carboy with 1 gal of head space you will find that you are only protected from about 1.25 degrees of temperature change. And if you have 2 gal. head space you are protected against a 0.75 deg. F temperature swing.
Now it is true that a large carboy will protect wine from a fairly large temperature swing simply because of the time it takes to change the temperature of the wine, but I am not sure it will keep it below 1 deg.
This suggests that a large inert gas head space could cause the carboy to breath through the airlock, maybe on a daily basis. That would not be a good thing.
Have I missed something? Comments?