3.3 Sample Collection and Analysis
Samples for [latex]_{}^{3}H[/latex] analysis should ideally be collected in glass bottles with minimal headspace and a tight‑fitting cap (e.g., a polyseal cap). However, when samples will be shipped to the laboratory, the use of high‑density polyethylene bottles—or other plastic that has a low permeability to water vapor—is generally acceptable as it avoids breakage issues associated with glass bottles. Typically, 1 L is sufficient volume, but readers are encouraged to contact their chosen laboratory for specific requirements.
Because the modern atmosphere may contain a [latex]_{}^{3}H[/latex] concentration much higher than the sample, it is important to avoid contact between the sample and the atmosphere. However, extreme care to avoid contact such as submerging the collection bottle in a larger container that is filled with sample water is generally not required. Historically, some watches had luminescent dials that utilized [latex]_{}^{3}H[/latex] gas and these should not be worn during sample collection. There are many different luminescent technologies used in modern watches, but, as [latex]_{}^{3}H[/latex] gas is still used by several manufacturers, it is best to avoid wearing a watch during sample collection. When [latex]_{}^{3}H[/latex] analyses are to be performed in the unsaturated zone, water must be first extracted using a suction lysimeter in situ or distillation from a core sample.
Tritium analyses are usually performed using liquid scintillation counting (LSC), gas proportional counting (GPC), or [latex]_{}^{3}{He}[/latex] in‑growth. In a recent comparison of 78 international [latex]_{}^{3}H[/latex] laboratories, 70 used LSC, seven used [latex]_{}^{3}{He}[/latex] in‑growth, and one used GPC (Copia et al., 2020).
A pre‑treatment using electrolytic enrichment of samples is commonly done for both the LSC and GPC methods for samples having [latex]_{}^{3}H[/latex] values less than several TU. The electrolytic enrichment process decomposes water to oxygen ([latex]O_{2}[/latex](g)) and hydrogen ([latex]H_{2}[/latex](g)) gas that is depleted in [latex]_{}^{3}H[/latex] relative to the remaining liquid phase. Typically, between 2,000 to 250 ml of sample is reduced to a volume of about 10 to 15 ml. The electrolysis process highly favors the formation of [latex]_{}^{1}H_{2}([/latex]g) at the cathode over [latex]_{}^{2}H_{2}[/latex](g) and [latex]_{}^{3}H_{2}[/latex](g), leaving the remaining solution enriched in [latex]_{}^{2}{H\ }[/latex](deuterium) and [latex]_{}^{3}{H\ }[/latex] (Copia et al., 2021). A [latex]_{}^{3}{H\ }[/latex] enrichment factor of about 16 to 32 is common, but a factor as great as 175 can be obtained (Morgenstern & Taylor, 2009).
The enrichment process involves five steps:
- predistillation to remove dissolved ions that might interfere with the electrolysis process,
- the addition of an electrolyte to facilitate electrolysis,
- electrolysis to enrich the sample in 2H and 3H,
- alkalinity neutralization to lower the pH of the solution after electrolysis, and
- final distillation to quantitatively recover the enriched water.
For the GPC method, methane is synthesized (e.g., reaction of water with aluminum carbide at 150 °C; Horvatinčić, 1980) from either the raw water sample or from the residual solution following electrolytic enrichment. The produced methane is then placed into a counting cell along with quenching gas at pressures that are typically 5 to 7 atm.
For the LSC method, about 10 ml of either the raw sample or the residual solution following electrolytic enrichment is added to a scintillation cocktail. The function of the scintillation cocktail is to absorb energy from radioactive decay and convert it into light pulses that can then be detected by a photomultiplier tube. Low background LSC are commercially available that employ both shielding and anti‑coincidence detection combined with sample blanks to reduce and correct for background radiation. The tritium detection limit for LSC of raw samples is typically between 2 to 10 TU and between 0.05 to 0.6 TU for enriched samples.
The [latex]_{}^{3}{He}[/latex] in‑growth method utilizes the child of [latex]_{}^{3}{H\ }[/latex] decay: [latex]_{}^{3}{He}[/latex]. A sample is first placed into a container having low [latex]He[/latex] permeability and degassed to remove dissolved [latex]He[/latex] from the sample. The container is then sealed and stored for one to 12 months during which time [latex]_{}^{3}{He}[/latex] is produced from [latex]_{}^{3}{H\ }[/latex] decay. After the in‑growth period, the evolved [latex]He[/latex] is let into a final clean‑up line and then a sector‑field mass spectrometer where [latex]_{}^{3}{He}[/latex] and [latex]_{}^{4}{He}[/latex] are measured using peak‑height manometry. The purpose of the [latex]_{}^{4}{He}[/latex] measurement is to correct for [latex]He[/latex] leakage, outgassing, or insufficient degassing of the sample. While the [latex]_{}^{3}{He}[/latex] in‑growth method is conceptually straight forward, several important factors affect the method’s implementation:
- The sample must be 99.9995 percent degassed of its dissolved [latex]He[/latex] content in a container that will not leak or outgas more than about 10,000 atoms of [latex]_{}^{3}{He}[/latex] during the in‑growth period. Special glass with a high aluminum content that is stored at −20 °C to reduce helium diffusion, or an all‑metal flask is used.
- The sector‑field mass spectrometer must have a resolving power greater than about one in 400 amu to separate the [latex]_{}^{3}{He}^{+}[/latex] peak from the hydrogen deuterium (HD) peak.
- The detection limit for the [latex]_{}^{3}{He}[/latex] in-growth method depends on the sample size, holding time, and residual helium from the degassing procedure or leakage during storage. For a sample volume of 500 ml and in‑growth period of six weeks, detection limits of 0.05 TU are common. For larger volumes (up to 3,000 ml) and/or longer holding times (up to one year) a detection limit of 0.003 TU has been reported (Jenkins, 1981). A significant advantage to the [latex]_{}^{3}{He}[/latex] in‑growth method is that time and sample amount are interchangeable in terms of the detection limit. For example, Marston and others (2012) measured the [latex]_{}^{3}{He}[/latex] content of structural waters in clay minerals on typical volumes of 20 to 30 ml with a detection limit of 0.5 TU by using an in‑growth time of 24 weeks compared with six weeks typically used for 500 ml samples.