A new liquid scintillation counting technique to resolve mixtures of two pure beta-emitting radionuclides

ABSTRACT A new liquid scintillation counting technique to resolve mixtures of two pure beta-emitting radionuclides Winifred Margaret van Wyngaardt Submitted November 2007 Methods currently available for the accurate activity resolution of dual-label solutions of pure β-emitting radionuclides are mostly time consuming and involve much effort. The goal of this thesis was therefore to devise a simpler method to achieve the same objective. The technique developed is based on elements of two liquid scintillation counting techniques that are widely used to measure single-radionuclide solutions, namely the triple-to-double coincidence ratio (TDCR) efficiency calculation technique and the CIEMAT/NIST efficiency tracing method. Double- and triple-coincidence count rates, together with the figure-of-merit P determined from an external tracer, provide sufficient information to extract the component activities of a source in a simple manner. The mathematical basis for the method is demonstrated and equations are derived to estimate statistical uncertainties, taking correlation effects into account. A simulation based on counting statistics was performed to assess the practicality of the method under normal counting conditions and to validate the derived uncertainty formulae. A critical overview of the underlying photon statistics of liquid scintillation counting is also given, together with clarification of various descriptions published in the literature. Further insight into the applicability of the new method was achieved by measuring a range of activity compositions of various radionuclide pairs, in particular 14C-63Ni, 33P-35S, 32P-33P and 32P-35S. In all of these experiments, the best results were obtained when the tracer standard used was the same as the low energy component of the mixture. The accuracy of the results for the mixtures of low-energy radionuclides, 14C-63Ni and 33P-35S, was excellent and mostly matched the best achievable by existing methods. The results obtained for the 32Pcontaining mixtures (with large energy differences) were not as good, showing deviations from the gravimetrically prepared activities that were larger than the estimated uncertainties. The successful demonstration of this new technique provides a useful extension to the application of the TDCR detector. In addition, owing to the method being particularly sensitive to incorrectly determined efficiencies, it shows potential as a tool to gauge the accuracy of various aspects relating to the efficiency model used.