Calibration of a liquid scintillation counter for alpha, beta and Cerenkov counting

Abstract

Calibration data are presented for 25 radionuclides that were individually measured in a Packard Tri-Carb 2250CA liquid scintillation (LS) counter by both conventional and Cerenkov detection techniques. The relationships and regression data between the quench indicating parameters and the LS counting efficiencies were determined using microliter amounts of tracer added to low {sup 40}K borosilicate glass vials containing 15 mL of Insta-Gel XF scintillation cocktail. Using {sup 40}K, the detection efficiencies were linear over a three order of magnitude range (10 - 10,000 mBq) in beta activity for both LS and Cerenkov counting. The Cerenkov counting efficiency (CCE) increased linearly (42% per MeV) from 0.30 to 2.0 MeV, whereas the LS efficiency was >90% for betas with energy in excess of 0.30 MeV. The CCE was 20 - 50% less than the LS counting efficiency for beta particles with maximum energies in excess of 1 MeV. Based on replicate background measurements, the lower limit of detection (LLD) for a 1-h count at the 95% confidence level, using water as a solvent, was 0.024 counts sec-{sup -1} and 0.028 counts sec-1 for plastic and glass vials, respectively. The LLD for a 1-h-count ranged from 46 to 56 mBq (2.8 - 3.4 dpm) for both Cerenkov and conventional LS counting. This assumes: (1) a 100% counting efficiency, (2) a 50% yield of the nuclide of interest, (3) a 1-h measurement time using low background plastic vials, and (4) a 0-50 keV region of interest. The LLD is reduced an order of magnitude when the yield recovery exceeds 90% and a lower background region is used (i.e., 100 - 500 keV alpha region of interest). Examples and applications of both Cerenkov and LS counting techniques are given in the text and appendices.