Optimization and uncertainty evaluation of electrolytical enrichment for tritium tracer experiments in rivers

Uncertainty analyses in the collection and processing of environmental data are often not adequately taken into account and, where technically possible, minimised. This can be particularly critical when the data is used to calibrate transport models, as is the case in dispersion experiments employing tritium - a commonly used tracer - to predict hydrological processes. In order to estimate the uncertainties of electrolytical enrichment for determining tritium activities, an electrolytical system was set up to test different cell designs and operating settings. In addition to considering the uncertainties of the cells used, the efficiency calibration, the net count rate, the enrichment factor and the loss of tritium during the oxy-hydrogen reaction were determined. The system was designed so that a high throughput of water samples was achieved, as is the case with conducting dispersion tests in the course of collecting data for hydrological modelling and risk assessment. An optimal enrichment factor Z of about 10-15 was identified, balancing improved counting statistics with increasing uncertainty of the enrichment system. At Z = 14, a minimum of relative uncertainty of 5.2-6.1 % was achieved under high throughput conditions. All uncertainties refer to 20 ml LSC vials containing 10 ml of transferred electrolyte after electrolysis and distillation and 10 ml scintillation cocktail, an LSC counting time of 43,200 s with a background count rate of 8.33E-3 cps. The uncertainties, especially in the range of low activity concentrations in the order of magnitude typical of surface water, are significantly lower than for direct measurements and once again underscore the robustness and efficiency of electrolytical enrichment, especially with regard to hydrological investigations using tritium as a tracer.