A quantitative Monte Carlo modelling of the uranium and plutonium X-ray fluorescence (XRF) response from a hybrid K-edge/K-XRF densitometer

A mathematical simulation approach based on the general purpose Monte Carlo N-particle transport code MCNP was developed to predict the response of the XRF branch of the hybrid K-edge/K-XRF densitometer (HKED). The respective MCNP models for two different versions of HKED instruments currently in use were set up and experimentally validated. The setting up of the models involved comprehensive simulations of a bremsstrahlung photon source, the examination of different particle transport models, as well as the examination of different photon attenuation and X-ray fluorescence data libraries. The computation speed was significantly increased through the extensive use of the variance reduction techniques. The models were validated through the series of benchmarking experiments performed with a representative set of uranium, plutonium and mixed U/Pu reference solutions. The models and simulation approach developed are intended for: (i) establishing a consistent mathematical calibration approach for the XRF branch of the HKED instruments, which will require minimum calibration effort and time, (ii) extending the applicability of the HKED method to non-standard samples (e.g. U/Pu mixtures with unusual element ratios) and non-standard sample matrices (e.g. HM matrices from the pyro-processing of irradiated nuclear fuel) without investing a great deal of extra calibration work, and (iii) improving the accuracy of the measurements through the modelling of special measurement effects (e.g. the secondary excitation effect, the interference with X-ray escape peaks, the inconsistent unfolding of the overlapping peaks and peak background delineation in the measured XRF spectrum), which are difficult or sometimes impossible to account for experimentally.