Effective desorption of tritium from diverse solid matrices and its application to routine analysis of decommissioning materials

Abstract

Tritium extraction from materials is most commonly carried out using oxidative thermal desorption in purpose-built furnace systems and typically involves trapping the product in a water bubbler which is sampled for measurement using liquid scintillation counting (LSC). The performance of perhaps the most widely used commercial system, the Raddec Pyrolyser, has been evaluated for a broad range of sample types. Several parameters that were expected to affect tritium desorption and recovery were systematically studied. These included sample heating rates and end-point temperatures, carrier/oxidant gas flows, catalyst temperature, bubbler trapping and carry-over/memory effects. A catalyst such as platinised-alumina is used to ensure the quantitative oxidation of volatile combustion products to HTO and CO 2. This also ensures that the trapped decomposition products do not colourise the bubbler solutions that are subsequently sampled for LSC. Tritium evolution profiles were determined for a range of sample types and were obtained by systematically changing bubblers at a set of progressively increasing temperatures. These experiments showed the maximum heating temperature and total combustion time required for the complete recovery of tritium from samples was dependent on the sample matrix types and the 3H form. These evolution profiles need only be determined once and are readily transferable to other Pyrolyser systems. For example tritiated water is rapidly liberated from samples at temperature around 100 °C whereas 3H substituted for structural H in organic species can require a temperature in excess of 300 °C to be released. Higher temperatures (up to 900 °C) are needed to liberate 3H originating from neutron capture reactions on trace Li or B within a material (e.g. reactor graphite or concrete). The furnace system investigated is highly effective at extracting tritium and 14C from all sample types studied (soil, sediment, biota, wood, metal, plastic, concrete, graphite, etc.) and overall it demonstrates high and reproducible recoveries.