No abstract

An extensive programme of research has been conducted for scintillation liquids and plastics capable of neutron-gamma discrimination for deployment in future passive and active Homeland Security systems to provide protection against radiological and nuclear threats. The more established detection materials such as EJ-301 and EJ-309 are compared with novel materials such as EJ-299-33 and p-terphenyl. This research also explores the benefits that can be gained from improvements in the analogue-to-digital sampling rate and sample bit resolution. Results are presented on the Pulse Shape Discrimination performance of various detector and data acquisition combinations and how optimum configurations from these studies have been developed into fieldready detector arrays. Early results from application-specific experimental configurations of multielement detector arrays are presented.

Certain scintillating materials are sensitive to both gamma and neutron radiation and can give information about the type of interacting radiation due to differences in the light output response. By collecting the light pulses and converting them to electrical signals the nature of the radiation can be determined by measuring the amount of electrical charge in the pulse tail -for neutrons, the pulses are longer, with more charge in the tail than for the shorter gamma pulses. This determination called Pulse Shape Discrimination (PSD) can nowadays be performed in real-time onboard digitisers during data collection. In this work several detectors (EJ301, EJ309 liquids; EJ299-33 plastic and p-terphenyl scintillators) of various shapes and sizes were connected to several digital Data Acquisition (DAQ) systems as well as the established digital / analogue hybrid Mesytec MPD8 / MADC-32 set up in a comparative study. The aim of the campaign was to produce a Figures of Merit (FOM) for the PSD performance of the various detector / DAQ combinations to give relative performance estimates of the CAEN V1751 10-bit 1 GSample/s digitiser in comparison with other DAQ solutions within a near-standardised experimental environment. It islikely that the DAQ set ups were not equivalent as significant differences in the matching of the detector outputs to the dynamic range of the digitisers were observed -however, with the configurations used in this campaign the CAEN V1751 digitiser showed superior FOM values to the Struck SiS3320, Bridgeport usbBase and Mesytec MPD-8 DAQ systems tested. Furthermore, there seemed little difference between the FOM from the faster but lower voltage resolution (1 GSample/s with 10 bits) CAEN V1751 compared to the slower but higher resolution (250 MSample/s with 12 bits) CAEN N6720 digitiser for this application.

In a liquid scintillator detector such as EJ-309, neutrons and gamma-rays give rise to different light pulse shapes due to the way they give up their energy but, in addition to this, there are fluctuations in the shape of the light pulse that varies with the point of interaction in the detector volume. The way in which light is dispersed and reflected throughout a detector volume introduces small fluctuations in the shape of the light pulse. This has an impact on the spread of pulses from neutrons and gamma rays and hence impacts the ability to perform discrimination of the two radiation types by pulse shape. This paper investigates, by experimentation, how the spread in the pulses produced by neutrons and gamma rays varies as the detector volume varies from 350 cubic centimeters to 6,570 cubic centimeters. The effect that this has on pulse shape discrimination performance is described. Also investigated are the effects of right-cylindrical geometries versus cuboid geometries. Pulse-shape discrimination of digitally captured neutron and gamma-ray event pulses was performed using the charge integration technique. A figure of merit at 100 keVee is seen to degrade by 25.2 ± 1.2% as the detector volume is increased from 350 to 6,570 cubic centimeters. At 1450 keVee this degradation increases to 30.9 ± 3.8%. In these experiments, cubic geometries appear to have similar pulse shape discrimination performance to right-cylindrical geometries. A 102 mm cube is calculated to have a figure of merit of 4.1 ± 0.7% better than a 102 x 102 mm cylinder at 100 keVee and on average about 4% better across the energy range 100 to 1450 keVee.

No abstract