Development of a deep neural network-based pulse shape discrimination for organic scintillators using GEANT4 generated pulses

Abstract

Since nuclear fission is a very complex and not fully understood physical process, it is an interesting field for theoretical and experimental investigation. This study analyses the method of pulse shape discrimination for a measurement related to nuclear fission. Organic scintillators are often used to measure prompt neutrons and gammas from nuclear fission. It is investigated how the neutrons and gammas can be distinguished from their different pulse shapes. A deep neural network is used for this purpose. This network is trained with pulse shapes generated by Geant4 and later tested with experimental data. This work shows that simulated Geant4 data can be successfully used to train a deep neural network that is able to distinguish experimental pulse shapes. This offers a promising approach for future measurements in nuclear fission studies.