Distance metrics for digital pulse-shape discrimination of scintillator detectors. (March 2019)

• Record Type: Journal Article
• Authors: Alharbi, T.
• Is Part Of: Radiation physics and chemistry. Volume 156(2019:Mar.)
• Page Start: 205
• Page End: 209
• Publication Date: 2019-03
• Subjects: Pulse-shape discrimination -- Scintillator detectors -- Distance metrics -- Pattern recognition; Radiation chemistry -- Periodicals; Radiometry -- Periodicals; Radiation -- Periodicals; Chimie sous rayonnement -- Périodiques; 539.2
• DOI: 10.1016/j.radphyschem.2018.11.014
• Languages: English
• ISSNs: 0969-806X

Abstract

Abstract: Digital pulse-shape discrimination (PSD) of scintillator detectors has gained popularity as an efficient technique for the discrimination between different types of particles. In this paper, we have investigated the use of simple distance metrics for digital PSD of neutron and gamma-rays in liquid scintillator detectors. The distance metrics are commonly used in the field of pattern recognition and offer the advantage of a simple and effective performance. Our method is based on quantifying the difference between a sampled detector pulse and an ideal step pulse. For this purpose, some of the common distance metrics including Euclidean, City-block, Canberra, root mean square (RMS) and Hellinger were experimentally examined through the discrimination of neutrons and gamma-rays in a liquid scintillator detector and their performances were compared to that of the conventional charge-comparison method. It was found that the City-Block distance metric exhibits a very simple and effective PSD approach by producing a figure-of-merit (FOM) of 0.72 against a FOM = 0.54 for the charge-comparison method in the very low light output range of below 160 keVee (electron equivalent energy) and a mean FOM value of 1.23 over the whole examined light output range of 60–1000 keVee. Highlights: Distance metrics are proposed for digital discrimination of neutrons and gamma-rays. Five metrics of Euclidean, City-block, Canberra, RMS and Hellinger were experimentally examined. The City-Block distance metric exhibits a very simple and effective PSD approach.