Glass containers are usually used for gas samples with taps at either end, but stainless steel and Teflon containers can be equally effective. The sample tube is evacuated before beginning sampling. If the sample tube is not evacuated, the air initially in the sample tube, can be eluted by the sample gas itself in a logarithmic manner but at least 10 volumes of sample must be passed through the tube to remove all the air. If the sample tube is evacuated, the tap can be opened in the gas to be sampled, and the pressure in the sample tube allowed to come into equilibrium with the atmosphere and the taps then closed.
Solid particles can contaminate the chromatograph system but will not normally have any deleterious effect on the accuracy or precision of the analysis for major components. Consequently, a filter should be inserted between the sample tube and the sample inlet to remove any relatively large solid particles that may be present. If one or more of the sample components are significantly soluble in water and the gas is wet a drying tube should be inserted in line with the filter. Any drying agent that is chosen must not disproportionately adsorb any of the sample components as this can also introduce significant errors. Duplicate or triplicate samples should be taken and then each sample analyzed in duplicate.
On arrival in the laboratory, the sample can be sucked directly from the sample container into the valve sample loop by means of a pump. Alternatively, the sample can be displaced into the sample valve with mercury. This procedure is somewhat inconvenient, but permits the sample loop to be well purged with sample before analysis. It also allows the residual sample to be kept at atmospheric pressure for extended periods of time without risk of contamination.
Sampling for Minor Components of a Gas Mixture
Trace quantities of volatile substances in air or in other permanent gases require different techniques. For example the measurement of a toxic vapor (e.g. benzene vapor) in air is an environmental test of this type and the analysis of the 'spectrum' of essential oils generated by a flower to identify the optimum harvesting time is another. Yet another example is the analysis of the head-space over foods for quality control or to identify product deterioration.
