Improved Algorithm for Close Geometry Characterization of Waste using Gamma-Ray Spectroscopy

Proper characterization of waste is important to reduce the costs of disposal. The requirement for short data collection time often means putting the detector as close to the waste container as possible. Nondestructive assay of gamma ray emitting nuclear waste requires modeling because preparing a standard to match the physical and radiological properties of any waste item is not possible. Several assumptions must be made about the waste. Many models use simplified efficiency determinations and attenuation corrections. These models work well for medium sized items when the detector-to-container distance is at least half of the largest dimension of the object. Other models use a hybrid Monte Carlo approach, still using the container assumptions, but using a complete set of measured detector efficiency parameters. A new algorithm for calculating the detector efficiency has been developed to allow close-in (that is, high efficiency) detector placement without extensive detector characterization, while retaining acceptable accuracy of the analysis results. This algorithm uses a simple mixed-nuclide gamma calibration, the HPGe detector description (crystal diameter and length, crystal type (n-type or p-type), thickness of germanium dead layer, and distance from the top of the end cap to the crystal) to compute the intrinsic detector efficiency. The intrinsic detector efficiency for the front and side of the detector are related to the detector diameter and length. The efficiency and corrections for the gamma rays from the item being measured are calculated in a voxel-by-voxel manner. A new collimator correction algorithm has also been developed for close geometry measurements. These new algorithms have been implemented in the ISOTOPIC software. Results show significant improvement for large boxes (B-25 box or larger) and items counted as close as 10 cm. Results from low-level drum counter measurements show useful improvements.