RESUMO
Hazardous contaminants buried within vadose zones can accumulate in soil gas. The concentrations and spatial extent of these contaminants are measured to evaluate potential transport to groundwater for public risk evaluation. Tritium is an important contaminant found and monitored for in vadose zones across numerous sites within the US nuclear weapons complex, including Los Alamos National Laboratory. The extraction, collection, and laboratory analysis of tritium from subterranean soil gas presents numerous technical challenges that have not been fully studied. Particularly, the lack of moisture in the soil gas in the vadose zone makes it difficult to obtain enough sample (e.g., > 5 g) to provide for the required measurement sensitivity, and often, only small amounts of moisture can be collected. Further, although silica gel has high affinity for water vapor and is prebaked prior to sampling, there is still sufficient residual moisture in the prebaked gel to dilute the relatively small amount of sampled moisture; thereby, significantly lowering the "true" tritium concentration in the soil gas. This paper provides an evaluation of the magnitude of the bias from dilution, provides methods to correct past measurements by applying a correction factor (CF), and evaluates the uncertainty of the CF values. For this, 10,000 Monte Carlo calculations were performed, and distribution parameters of CF values were determined and evaluated. The mean and standard deviation of the distribution of CF values were 1.53 ± 0.36, and the minimum, median, and maximum values were 1.14, 1.43, and 5.27, respectively.