A new numerical correction method for gamma spectra based on the system transformation theory of random signals.

Research paper by Chongjie C Wang, Qiuyan Q Zhang, Yue Y Sun, Jingnan J Liu, Yingying Y Zhou, Min M Zhang

Indexed on: 13 Mar '21Published on: 12 Mar '21Published in: Applied Radiation and Isotopes


The conventional correction techniques for gamma spectra usually use approximate algorithms, the calculation procedures are tedious and complicated, the obtained corrected spectra are usually the channel domain spectra even if they are calibrated by energy. Moreover, the spectra are still dependent upon system parameters themselves such as the gain of amplifier and the offset of ADC etc. In this paper, a new numerical correction method of gamma spectra was developed on the basis of the system transformation theory of random signals. By transforming the measured spectrum from the channel domain to the energy domain, the theoretical deposition energy spectrum which is energy domain spectrum, rather than channel domain spectrum can be obtained without any algorithm approximation processing. The presented method makes it possible to compare and exchange the gamma-ray spectrum data that are measured under different conditions or by different spectrometer systems equipped with the same type of detectors. To validate the method, a set of NaI (Tl) gamma spectra and three sets of HPGe gamma spectra were collected. The results show that serious spectral drifts and intensity variation in NaI (Tl) gamma spectra were effectively eliminated. The maximum relative drift of peak position, the maximum relative variation of peak height and that of FWHM before correction were 65.03%, 40.37% and 61.65%, respectively, and after correction, they became 0.19%, 2.81% and 0.85%, respectively, indicating that the method has strong ability to correct gamma spectra. HPGe gamma spectra were also well corrected using the presented method, and were identified as gamma radiation fingerprints of nuclear materials. The results show that the method can also improve the confidence degree of identifying nuclear materials significantly. Copyright © 2021 Elsevier Ltd. All rights reserved.