Automatic identification and quantification of radionuclides in gamma spectra using numerical optimization

We propose a novel method for gamma-ray spectral analysis that populates and solves an arbitrarily large system of radionuclides in order to explicitly model the production and loss of chains of radionuclides during counting. This solution is embedded into an optimization problem in order to resolve spectral interferences and find the nuclide mixture that best explains the measured spectrum. We demonstrate the efficacy of this method using a synthetic example with complex spectral interferences, short half lives, and nuclide transmutation, as well as a spectrum collected of a multi-gamma standard typically used for efficiency calibration.