Improved sample preparation and counting techniques for enhanced tritium measurement sensitivity

Abstract:

Tritium (T) measurements offer insight to a wealth of environmental applications including hydrologic tracking, discerning ocean circulation patterns, and aging ice formations. However, the relatively short half-life of T (12.3 years) limits its effective age dating range. Compounding this limitation is the decrease in atmospheric T content by over two orders of magnitude (from 1000-2000 TU in 1962 to < 10 TU currently) since the cessation of above ground nuclear testing in the 1960’s. We are developing sample preparation methods coupled to direct counting of T via ultra-low background proportional counters which, when combined, offer improved T measurement sensitivity (~4.5 mmoles of H₂ equivalent) and will help expand the application of T age dating to smaller sample sizes linked to persistent environmental questions despite the limitations above. For instance, this approach can be used to T date ~ 2.2 mmoles of CH₄ collected from sample-limited systems including microbial communities, soils, or subsurface aquifers and can be combined with radiocarbon dating to distinguish the methane’s formation age from C age in a system. This approach can also expand investigations into soil organic C where the improved sensitivity will permit resolution of soil C into more descriptive fractions and provide direct assessments of the stability of specific classes of organic matter in soils environments.

We are employing a multiple step sample preparation system whereby organic samples are first combusted with resulting CO₂ and H₂O being used as a feedstock to synthesize CH₄. This CH₄ is mixed with Ar and loaded directly into an ultra-low background proportional counter for measurement of T β decay in a shallow underground laboratory. Analysis of water samples requires only the addition of geologic CO₂ feedstock with the sample for methane synthesis. The chemical nature of the preparation techniques enable high sample throughput with only the final measurement requiring T decay with total sample analysis time ranging from 2 -5 weeks depending on T content.