Gamma-gamma coincidence spectra and background subtraction

The raw gamma-gamma coincidence spectra were dominated by 511 keV counts and radioactivity [1]. Two procedures were investigated for subtracting this background. The first procedure [1] was an extrapolation of the method employed to obtain the singles spectrum of figure 1(d). This relied on the beam pulsing to separate in-beam and out-of-beam events but made no further use of fast timing. The resultant coincidence spectra were virtually free of background [1]. The second procedure [2] made no use of the beam pulsing, but utilized a fast coincidence requirement ($\pm 10$ns) between the gamma-rays. It was necessary to exclude events in which the opposite pair of detectors fired, as these 511 keV events accounted for 68% of all coincidence events on tape. The resultant spectrum included a strong component due to radioactivity in the target (see figure 2(a)). When, in addition, a charged particle coincidence was required, the spectrum in figure 2(b) was obtained, which is almost as clean as the spectrum obtained using the subtraction procedure based on beam pulsing (see discussion in [2]).

 

Figure: Gamma-ray spectra for events where two detectors fired within a $\pm$10ns time gate and one of them recorded >511 keV: (a) all such events, with prominent $\beta$-delayed transitions labelled and (b) events in which at least one charged particle was also recorded. Events where the pair of opposite detectors fired have been excluded [2].

The timing spectra for gamma-rays relative to the beam pulses demonstrate dramatically the difference between using stable and radioactive beams [1]. This is shown schematically in figure 3(a), which is for singles events; with a stable beam, the flat background is virtually absent. The equivalent spectrum for coincidence events, shown schematically in figure 3(b), can be misleading without further interpretation. Whilst the peak to background is improved relative to singles events, only a small fraction of events in the peak is useful. With fast timing, the situation is much improved, as shown in figure 3(c).

 

Figure: Origins of the gamma-gamma singles and coincidence counts recorded in the `TaLL' experiment. These schematic spectra show the time of each recorded gamma ray, relative to the beam pulses, and may be compared with actual spectra in ref. [1]. The relative areas within each spectrum are accurate for the TaLL experiment.