Speaker
Description
This work presents an optimized photon-neutron discrimination procedure developed during the commissioning of the neutron generator at the IFUSP Linear Accelerator Laboratory. The experimental setup employs an EJ-276D plastic scintillator for neutron detection and a silicon detector to tag $^3$He nuclei from deuterium fusion reactions. Time-coincident signals between the two subsystems are used to suppress accidental backgrounds from photons and scattered neutrons.
Pulse-shape discrimination was optimized through the $f_{\rm prompt}$ parameter, defined as the ratio of pulse integrals over prompt and total time windows. A data set of approximately $4.36\times10^5$ pulses was analyzed while scanning integration intervals to maximize the photon–neutron separation, quantified by the figure of merit. Gaussian fits to the $f_{\rm prompt}$ distributions and a dedicated cost function combining neutron loss and photon contamination were used to determine the optimal parameters in bins of deposited energy.
The resulting optimization yields photon contamination below 1% for energies above 75keV$_{ee}$ and provides a safe neutron selection threshold. The method is readily applicable to ReD, ReD+, and DarkSide-20k analyses.
| I read the instructions above | Yes |
|---|