CHAPTER 58 A General Method for Determining Sample Efficiencies in Liquid Scintillation Counting

Many modern liquid scintillation counters use two coincident photomultiplier (PM) tubes to reduce background noise. These counters detect most charged particle emitting nuclides with an efficiency near 1; however, they are noticeably less efficient for detecting lower energy emissions from nuclides such as tritium, because each decay produces so few photons that any coincident detection is less probable. For quenched samples, the photon production is further reduced, yielding a corresponding contraction of the observed energy spectra. The probability (or efficiency) for detecting a charged particle increases from 0 to 1 as its photon production (or its energy) increases from 0 to higher values. A model for this probability was developed and tested against experimental measurements. A trinomial probability distribution modeled the efficiency as a function of observed spectral energy. The three probabilities of the trinomial include two for photon detection by the two PM tubes and one for nondetection. For a TRI.CARB 2000 CNLL liquid scintillation analyzer, reasonable efficiency models resulted when using the trinomial formalism as a basis for empiri-cally selecting the efficiencies. Using measurements of low-energy Auger and conversion electrons of quenched 57Co standards as an experimental guide, four efficiency curves were modeled and tested. Scintillation spectra of 3H and 14C quenched standards were convoluted with the efficiency curves to test each model. The dpm predicted by the best model agreed with that known for the standards, provided that cpm/dpm> 0.25. The method may be extended down to cpm/dpm 0.1 by incorporating a "bootstrap " correction technique