Performance of NDA Techniques on a Vitrified Waste Form

Abstract

Vitrification is currently being investigated at Los Alamos National Laboratory (LANL) and Pacific Northwest National Laboratory (PNNL) as the technology for the treatment of several waste materials at Rocky Flats Environmental Technology Site (RFETS). One of the materials being considered for vitrification is low-fired ash. The project to develop a vitrification process for this ash includes determination of an appropriate non-destructive assay (NDA) technique for the analysis of the final vitrified waste form. Because little NDA data on vitrified waste forms have been acquired in the general NDA community to date, the following study was conducted at LANL to determine the feasibility of utilizing various NDA instrumentation available at RFETS to measure these particular waste forms. The technique of most interest to RFETS and most focused upon in this study was that of a segmented gamma scanner (SGS) with 1/2-inch collimation, although an SGS with 2-inch collimation was also included. The specific objective of this study was to determine if an SGS could meet an accuracy criterion of within 10% of the special nuclear material (SNM) content as measured by calorimetry and γ-ray isotopics. This study was performed on a full-scale vitrified (borosilicate-based glass) ash sample prepared similarly at LANL to that anticipated at RFETS. The study focused on several factors known to decrease the accuracy of the SGS technique: sample density and homogeneity, end effects, and lump (self-attenuation) effects. Tomographie gamma scanning (TGS) and real-time radiography were utilized in the determination of the waste form homogeneity. A drum-sized thermal neutron technique (TNC) was also included in the investigation to provide an alternative in the event the SGS failed to meet the required level of accuracy.KeywordsWaste FormPacific Northwest National LaboratoryTransmission SourceOpaque RegionOuter ContainerThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.