Speaker
Description
In the field of radiation detection using Ge detectors, pulse pile-up rejection processing (PUR) is widely used. PUR reduces the distortion of the spectrum shape but it causes further increase of dead time. Many papers have been written on the dead time losses with PUR and methods to correct the influence [1],[2]. The proposed correction methods presuppose that the timing of pulses is random. However, in neutron time-of-flight (TOF) experiments, event rates dramatically change depending on TOF and the presupposed randomness is not satisfied. Then a new dead time correction method applicable for neutron TOF experiments is required.
The dead-time influence in TOF is simulated by using a CAEN v1724 digitizer, a BNC DB-2 random pulse generator, a BNC PB-5 pulse generator and a clock generator. 1 kHz triggers were produced by the clock generator and fed into the digitizer as a reset signal of a clock counter (dummy of triggers from an accelerator). With the rate of 70% of the reset signals, trigger signals delayed by 12 $\mu$s were fed into the external trigger input of PB-5 to produce input signals. The random pulse generator was used to produce dummy constant background signals. The produced input and background signals were fed together into the digitizer. The time interval from the reset signal and the pulse height of the signal for each event were recorded as a list-formatted data.
An example of the obtained time dependence spectrum is shown in Figure. The constant background rate decreases not only after the event, but also before it as well. In the time range from 11.5 $\mu$s to 12.25 $\mu$s, the constant background rate is recovered. This is because the digitizer is incapable of discriminating two events in this range and reports as a single event. The reported signal height indicates the sum of the deposited energy in the detector of the two events. In the present work, we refer the time range from 9.5 $\mu$s to 11.5 $\mu$s and 12.25 $\mu$s to 15.2 $\mu$s as “dead state” and that from 11.5 $\mu$s to 12.25 $\mu$s as “accidental sum state”.
The dead time of the digitizer can be calculated in offline data analysis using the list format data. In a normal gamma-ray spectroscopy, the accidental sum state is not acceptable, because single gamma-ray energy is needed. In this case, the rate of the dead time, $P_{dt}(i)$, is calculated at each TOF channel as follows:
$P_{dt} (i)=(T_{ds} (i)+T_{as} (i))/(N_{shot} \delta_t )$
where $T_{ds}(i)$ is the time length of the digitizer being in the dead state, $T_{as}(i)$ is that being in the accidental sum state, $N_{shot}$ is the number of the reset signals and $\delta_t$ is the TOF bin width. On the other hand, in neutron TOF experiments with the pulse height weighting technique, the accidental sum state is acceptable if a weighting function of a detector is close to be a linear. (Of course, its uncertainty should be estimated.) In this case, the rate of the dead time is calculated by following equation.
$P_{dt}(i)=(T_{ds}(i))/(N_{shot} \delta_t )$
This dead time correction has been applied to the analysis of the data measured with CAEN V1724 digitizers in Beam Line #04 (ANNRI) at J-PARC to deduce neutron capture cross section.