A Bayesian inference approach for the quantitative analysis of high-resolution gamma-ray spectra is proposed. The resulting posterior distribution for the sought isotope activities allows to calculate estimates, as well as corresponding uncertainties and credible intervals. The approach is based on a full-spectrum-analysis methodology by utilizing a set of well characterized template spectra for the different isotopes. A statistical model for the analyzed gamma-ray spectra is developed which includes all relevant uncertainty sources such as random fluctuations in the data and the uncertainty of the isotope activities underlying the template spectra. The approach includes a novel correction methodology which accounts for possible deviations between the energy scales of the analyzed spectrum and the template spectra, which is required for a precise analysis of the high-resolution spectra. Furthermore, the approach enables the inclusion of multiple template spectra for a single isotope recorded at different activity levels, which can be advantageous in the presence of a non-linear detector response. Calculation of the posterior distribution is carried out using Markov chain Monte Carlo sampling techniques. Application of the approach is demonstrated using a set of spectra from samples containing different isotope mixtures. A comparison of the activities obtained by the proposed Bayesian analysis with reference values determined by an established peak evaluation method reveals a high agreement with a root mean square error of only 1.2%. The determined activities are shown to be statistically consistent with the reference procedure, which emphasizes the reliability of the uncertainties derived by the proposed methodology. Software to ease the uptake of the proposed methodology is made available.
Kostenfreier Zugang
Open Access Hybrid
Rechteinformation
CC BY 4.0 ; Creative Commons Attribution 4.0 License
Themenbereich der Metrologie
Ionisierende Strahlung
Zitierung
Wübbeler, G., Stein, M., Fleischhack, H., Röttger, S., Honig, A., & Elster, C. (2025). Bayesian full-spectrum analysis of high-resolution gamma-ray spectra with energy scale correction and MCMC uncertainty quantification. Measurement Science and Technology, 36, 1–14. https://doi.org/10.1088/1361-6501/adf138
