Txf DGA sampling question

[Latango]

There's been a question raised about DGA sampling methods in our organization. The procedure calls for filling both a syringe and a 0.5L glass bottle. The question is, how much airspace should be left in the top of the 0.5L bottle to allow for expansion? The opinions range from "a little bit" to "sample it while it's hot and leave no airspace at all". The latter is a possibility if the txf is either online or recently isolated, but units in storage are probably going to be a little under ambient temp.

The vendor specifies an aluminium flask, and leaving a 10-15mm gap, however we do not have access to any aluminium flasks.