Estimation of an empirical formula for efficiency of a BEGe type detector using machine learning based algorithm. (May 2023)

• Record Type: Journal Article
• Authors: Dey, Riya; Chinnaesakki, S.; Dhumale, M.R.; Bara, S.V.; Singh, Kapil Deo; Jha, S.K.; Kulkarni, M.S.; Anand, S.
• Is Part Of: Radiation physics and chemistry. Volume 206(2023)
• Publication Date: 2023-05
• Subjects: BEGe detector -- Machine learning -- Nonlinear regression -- FEPE -- FLUKA; Radiation chemistry -- Periodicals; Radiometry -- Periodicals; Radiation -- Periodicals; Chimie sous rayonnement -- Périodiques; 539.2
• DOI: 10.1016/j.radphyschem.2023.110761
• Languages: English
• ISSNs: 0969-806X

Abstract

Abstract: The objective of this study is to establish an empirical formula of full-energy peak efficiency (FEPE), for a Broad Energy High Purity Germanium (BEGe) type detector. FLUKA Monte Carlo-based code is used in this work to simulate radiation transport in the BEGe detector, and FEPE values have been estimated for various gamma energies and wide range of commercially available detector configurations. Then, using machine learning-based non-linear regression models, an empirical formula relating FEPE values with detector configuration (active volume and diameter-to-length ratio) and incident gamma radiation (energy >122 keV) is obtained. The results show that the effect of diameter-to-length ratio on FEPE values cannot be ignored in the case of a BEGe detector, which is not observed in case of coaxial type detectors. At lower energies, the efficiency of the BEGe detector is higher than the coaxial detector due to its large diameter-to-length ratio. This empirical formula can be used to quickly estimate FEPE values and verify nominal geometric parameters provided by the manufacturer. Highlights: Empirical formula for FEPE is developed by machine learning based regression models. FEPE depends on detector volume and dia-to-length ratio apart from gamma energy. This formula can be used to verify nominal geometric parameters of a BEGe detector.