Paper No. 96-38
Presentation Time: 9:00 AM-6:30 PM
COMPARISON OF WATER WELL SAMPLING METHODS FOR DISSOLVED GASES
Currently there are three popular methods used to collect dissolved hydrocarbon and noble gas samples from water wells: exetainers, Isoflasks®, and copper tubes (standard for the USGS). There is a lack of consensus among the community as to the reliability and reproducibility of each method, as a comprehensive study on the viability of each for dissolved gas sampling has yet to be conducted. Major, hydrocarbon, and noble gas molecular, elemental, and isotopic analyses provide a valuable insight to many of the factors that are used to assess the overall quality of a water well, with various isotopic ratios being used to determine the genetic source of gas components, the residence time of groundwater, and the potential for contamination of groundwater systems. When sampling a water well for the purpose of major, hydrocarbon, or noble gas analyses, fractionation and atmospheric contamination of the sample are a serious concern, specifically for noble gases that are present in trace abundances. Using a combination of field samples and lab prepared controls, the relative effectiveness of each sampling method will be examined by testing for different hydrocarbon and noble gas abundances via a Membrane Inlet Mass Spectrometer (MIMS), headspace analyses, and traditional gas extraction/exsolution methods.
Preliminary data suggests that of the three sampling methods, the copper tube method is the least prone to atmospheric contamination or changes in gas composition due to the inherent lack of opportunity for air contamination as well as copper being an antimicrobial metal. Copper tubes and IsoFlasks have trade-offs when it comes to total hydrocarbon composition, and appear comparable for hydrocarbon isotopic work. Herein, we demonstrate the appropriate usage for each technique and develop a workflow chart that the geochemistry community can utilize to conduct cross-laboratory calibrations and to produce comparable datasets.