Gamma spectrometry method for evaluation of low-activity radioactive waste

Abstract

The growing demand for high-purity radionuclides in nuclear medicine leads to increasingly complex radioactive waste streams in production facilities. These wastes often consist of heterogeneous, low-density materials containing long-lived radionuclide impurities at activity levels near regulatory clearance thresholds. Their reliable characterization requires accurate, low-level, and non-destructive measurement methods that can account for non-uniform source and matrix conditions.

This dissertation presents the development and experimental validation of a gamma spectrometry method tailored to unconditioned, low-activity radioactive waste from the production of medical radionuclides. The method is specifically designed for waste packaged in box-shaped containers and targets low-Z, low-density materials such as disposable protective clothing, laboratory items, and auxiliary equipment.

A stationary multidetector array was constructed using eight CeBr3 scintillation detectors. Each detector was individually calibrated for energy and resolution and arranged in a geometry optimized for maximum efficiency uniformity. The array operates without mechanical scanning and is integrated into a shielded measurement chamber. Compared to NaI(Tl), CeBr3 detectors offer significantly better energy resolution (3.9% at 662 keV), enabling the resolution of overlapping gamma peaks and robust identification of radionuclide impurities.

MCNP6.2 simulations were used to construct detailed detector models, refined using radiographic imaging. Simulated full-energy peak efficiencies deviate by less than 2.5% from experimental measurements. A probabilistic model of the spatial source distribution was introduced, and the most probable efficiency values were determined, including relative uncertainties caused by source heterogeneity. Fifth-order logarithmic polynomial was fitted to efficiency values across the 100–1400 keV range.

The newly developed optimization method determined the optimal detector spacing based on simulated spatial efficiency grids and a figure of merit quantifying system response variability. The resulting configuration reduced efficiency variation by 17% and was validated experimentally using randomized source positions.

Self-absorption correction factors were derived from simulations that included randomly generated material densities with relative uncertainties up to 5%, caused by heterogeneous distribution of the waste matrix. A linear correlation was found between correction factors and average matrix density. The model was experimentally validated, with deviations remaining below 2% across relevant scenarios.

A semi-automated algorithm was developed to define regions of interest (ROIs) for 13–14 relevant radionuclides expected in radioactive waste. The superior resolution of CeBr3 detectors enabled reliable ROI assignment for all target nuclides, whereas NaI(Tl) could only accommodate six due to peak overlaps. Validation on real radiopharmaceutical waste samples showed that the developed method reduces measurement deviations and improves regulatory compliance compared to conventional gross counting techniques. Detection limits below 0.1 Bq/g were achieved within 5 minutes of measurement time.