Abstract
The observed output from a liquid scintillation counter is compared with that predicted theoretically for a wide range of beta-ray emitters. It is demonstrated that the extrapolated integral bias curve has an intercept which is always less than the absolute activity but this intercept can be calculated for any isotope including tritium. The probability that no output will be observed from any event (zero probability) can be determined as a function of the combined efficiencies of the scintillator and photomultiplier (figure of merit). Methods of measuring the figure of merit are presented for both single photomultiplier tube and coincidence counters. Corrections for high specific energy loss (dE/dx) and losses at the boundaries of the scintillator are calculated and the corrected figure of merit is found to be independent of electron energy from 5 keV to 1 MeV. Absolute activities were determined for three isotopes, tritium, carbon-14 and chlorine-36. The results were in good agreement with the standards of Radio-chemical Centre, Amersham, and the National Bureau of Standards, USA. The zero probabilities are largest for tritium, for example 35% and 58% for single and coincidence counters respectively. The tritium integral bias curve is shown to be essentially exponential over a wide range of bias voltage and a logarithmic extrapolation to zero bias corresponds to the absolute activity with an accuracy to ±10%.