When immediate analysis of gas samples is not possible, we store the samples in pre-evacuated vials.  Most commonly, we use 10mL Wheaton vials with rubber septa, sealed with an aluminum crimp top (both available from Fisher).  We have tested many different septa for stability in gas concentration.  The best septa we have found are manufactured by Geo-Microbial Technologies.  The web site is http://gmtstoppers.com You can buy the septa directly from this web site; they are not distributed by any of the usual science product vendors.  Our tests show that sample air in these vials retain their concentrations for >60 days.  But we usually try to complete analysis within 14 days.

In practice, we fill the vials with 25mL of sample air, thus overpressurizing the vials.  This overpressurization allows us to remove sample air by syringe for manual injection into whatever analytical system we’re using.

The concentration of samples can change over time due to leakage and/or contamination.  To account for these effects, we fill vials with standard air at the same time and place as we fill the sample vials.  To account for both leakage and contamination, we use two levels of standards that bracket the range of sample concentrations.  For example, if we’re analyzing methane at 0.5 to 1.8ppm, we would fill standard vials with N2 (0 ppm) and with a 5ppm standard.  We can then correct the samples for any systematic change in the standards.  We bring standards to the field in weakly pressurized gas canisters, or in Tedlar gas bags (Sigma-Aldrich an other vendors).

We have not had good luck storing air samples in vacutainers.  These systems, although convenient, are not well evacuated and have high concentrations of contaminant gases (they’re designed for blood sampling).  Whatever vial system you choose, test it by putting low and high concentration standards in the vials, and comparing to the standard when it was not put in the vial.