The detection and analysis of multiple types of radiation is of paramount importance in nuclear safe-guarding and as such is a primary concern of sensor development. In the past, many nuclear safe-guarding have employed a combination of plastic scintillator and He-3 filled proportional gas counters for neutron and gamma detection, but since the increased use of He-3 for homeland security and science applications the world has been facing a shortage of this Nobel gas. Efforts have been made to find a suitable replacement for these detectors, most solutions to this problem have involved the use of Lithium or Boron for thermal neutrons and liquid or plastic scintillator for fast neutron solutions.
In this work two potential replacement materials were investigated, namely CLYC(Ce) and EJ299-33A coupled to SiPMs(Silicon photomultipliers) with the aim to produce a compact, low power, robust detection system for both neutron and gamma radiation. The response of the materials to neutron and gamma radiation was investigated to probe the pulse shape discrimination potential of the materials. Simulations were then performed in order to inform the design of a light guide to optimize light collection for coupling to arrays of SiPMs. The signals produced by the system were then digitized, and algorithms applied in order to discriminate between neutron and gamma radiation, and the possibility of using machine learning methods such as deep learning were also investigated.
It was found that both time domain and frequency domain algorithms had their strengths and weaknesses, with time domain methods such as charge comparison methods producing reasonable separation of spices and having very fast execution time, they were more susceptible to noise than methods which operated in the frequancy domain. Methods which take advantage of the ability to remove much of the noise in the signal tend to be more computationally expensive than that of the charge comparison method, but the results did benefit from the reduction in noise, which reduced the dispersion of the spieces distrabuitons, which can sometimes be seen in other methods. This results in the PSD performance being improved in some applications in comparison to other time domain methods such as charge comparison or pulse gradient analysis.
In this work, it has been shown that neutron-gamma PSD detectors can be constructed from the combination of CLYC and EJ299-33A scintillators and SiPM arrays which have the potential to replace He-3 detection systems in many applications while giving the benefit of reduction in detector size as well as an increase in the robustness of the detector system for in field deployment. It has also been shown that multiple approaches to the PSD analysis can yield favorable results and each have benefits which make it well suited to certain applications.