This paper presents studies of the influence of the coincidence resolving time on the activity calculated by the Triple-to-Double Coincidences Ratio (TDCR) method in Liquid Scintillation (LS) counting. Recently, published methods for the correction for accidental coincidences in TDCR counting open the possibility to use resolving times up to several microseconds, long enough not to miss true coincidences and to study the effects of delayed fluorescence.
H-3, C-14, Fe-55 and Ni-63 LS-sources in UltimaGold (UG), UG LLT and Toluene+PPO cocktails were measured using a TDCR counter connected to a digitizer working in list-mode. The necessary resolving time to include 99.9% of the logical sum of double (D) coincidences was found to be 1.2 microseconds for H-3, 1 microseconds for Fe-55 and 500 ns for Ni-63 in UG. The activity of all LS-sources was calculated using the TDCR method for resolving times from 10 ns to 2 microseconds and a significant dependence between the calculated activity and resolving time was observed. A dedicated Monte Carlo (MC) code was used to simulate list-mode data from TDCR measurements. The simulation results suggest that the H-3 activity calculated by the TDCR method is overestimated regardless of the used resolving time if delayed fluorescence is present which is not described by the used ionization quenching function.
Efficiency variation measurements of H-3 in UG LLT show a strong dependence of the optimal $kB$ parameter on the used resolving time: 85 um/MeV at 40 ns and 110 um/MeV at 200 ns, leading to 2.5% difference in calculated activity. In the framework of this study the efficiency variation methods by chemical quenching and gray filters were compared and a difference of 60 um/MeV between the two was observed.
The results from this article demonstrate that regardless of the available corrections for accidental coincidences, it is not advisable to increase the resolving time beyond what is necessary to register all prompt fluorescence events. Moreover, even for short coincidence resolving times, delayed fluorescence could have a significant influence on the activities calculated by the TDCR method.