Waste-glass castings generated from engineering-scale developmental processes at the Pacific Northwest Laboratory are generally found to have significant levels of cracks. This report discusses the causes and extent of fracturing in full-scale canisters of waste glass as a result of cooling and accidental impact. The work was performed at the Pacific Northwest Laboratory under the High-Level Waste Container Development Project, part of the U.S. Department of Energy's Long-Term High-Level Waste Technology Program technically coordinated by Savannah River Laboratory. Although the effects of cracking on waste-form performance in a repository are not well understood, cracks in waste forms can potentially increase leaching surface area. If cracks are minimized or absent in the waste-glass canisters, the potential for radionuclide release from the canister package can be reduced. Additional work on the effects of cracks on leaching of glass is needed. In addition to investigating the extent of fracturing of glass in waste-glass canisters, methods to reduce cracking by controlling cooling conditions were explored. Thermal fracturing can be controlled by using a fixed amount of insulation for filling and cooling of canisters. In order to maintain production rates, a vi small amount of additional facility space is needed to accomodate slow-cooling canisters. Alternatively, faster cooling can be achieved using the multistaged approach. Additional development is needed before this approach can be used on full-scale (50-em) canisters. vii. .

This report presents technical data and performance characteristics of a high-level waste glass and canister intended for use in the design of a complete waste encapsulation package suitable for disposal in a geologic repository. The borosilicate glass contained in the stainless steel canister represents the probable type of high-level waste product that will be produced in a commercial nuclear-fuel reprocessing plant. Development history is summarized for high-level liquid waste compositions, waste glass composition and characteristics, and canister design. The decay histories of the fission products and actinides (plus daughters) calculated by the ORIGEN-II code are presented • iii

The Pacific Northwest Laboratory (PNL) is conducting research and development studies on technology for immobilization of nuclear wastes. In this report, an overview of stress analysis methods, structural design procedures, and design data is presented for canisters used to package solidified wastes, particularly borosilicate glass. In addition, waste processing, canister materials, fabrication and inspection methods, and performance testing are summarized. Sources of stress in canisters are lifting and handling loads, internal pressure, hightemperature filling operations, transient heating and cooling, differential thermal expansions of canisters and glass, and impact loadings from low-probability accidents. Results of case studies that illustrate applicable methods of stress analyses are presented for these sources of stress. Existing sections of the American Society of Mechanical Engineers (ASME) Boiler and Pressure Vessel Code are reviewed and found to be applicable to canister fabrication, but the code does not cover many aspects of canister service loadings. Specialized criteria for minimum wall thicknesses to sustain filling stresses are proposed in this report. Results of a test program to measure the creep strength of candidate canister materials are described. Methods to predict residual stresses in the walls of waste canisters are described; predicted residual stress levels agree with measured stress levels. The consequences of these residual stresses are reviewed, and stress-corrosion cracking is identified as the mode of canister failure affected by residual stresses. Canister-closure design is covered in detail, particularly the welding and inspection of the final closure seal-weld. It is shown that the methods of fracture mechanics and fatiguecrack-growth analyses are valuable tools for evaluating the performance of closure welds in the presence of crack-like defects. Canister performance in process trials at PNL shows the ability of canisters to survive high temperatures and loadings during processing. The results of impact tests at PNL show that a suitably designed canister can sustain severe impacts without loss of integrity • iii

DESCRIPTION OF WORK REPORTED This report presents the results of a test run that produced glass from simulated radioactive wastes. The simulated wastes approximated wastes that would result from the radionuclide removal process. The wastes were vitrified in an electric melter, and the glass product was cast in metal canisters. These glass castings were subsequently evaluated to determine chemical composition, density, leachability, and surface area both for the "as-cast" condition and after a 7.6-m (25-f) drop onto a massive pad. The objective of the week-long run was to demonstrate the technical feasibility of vitrifying Hanford wastes in a joule-heated ceramic melter and of producing a large prototypic casting in metal containers. APPROACH USED The technical feasibility of incorporating Hanford wastes into borosilicate glass was explored by converting 8180 kg of simulated waste and glass formers into glass in a joule-heated ceramic melter. The waste feed was made by mixing powdered chemicals into a chemical blend or batch(a) whose composition was typical of that for a Hanford waste depleted of sodium salts. Glass produced was collected in four canisters that had 0.41-, 0.61-, or 0.91-m (16-, 24-, or 36-in.) diameters. The glass in the canisters was subjected to various tests to determine its characteristics. CONCLUSIONS Experience gained in the week-long vitrification test and characterization of the glass produced in the run support the following conclusions: (a) "Batch" is the common word used in the glass industry to describe a chemical blend of oxides, salts and/or minerals. Batch should not be confused with an operational mode-i.e., batch vs. continuous. iii Summary of Porosity Characteristics of Simulated Waste Glass Specimens, Test Series RHO-2H-3• • 45 Relative Peak Heights from X-Ray Fluorescence Analyzer Results for Samples from Canister N o. 3 .-47 X-Ray Fluorescence Analysis of Samples from Canister No.1.

The Pacific Northwest Laboratory performed experiments on canisters of simulated high-level waste (HLW) glass to estimate the effects of radial temperature gradients on the degree of glass fracture in actual HLW canisters. Glass cracking may result in an increased amount of fine glass particles that are of concern in transportation and repository safety analyses. A test matrix was developed to examine the effect of air cooling, water quenching, and internal fins on canister cooling rates, radial temperature gradients, glass cracking levels, and particle size distributions. The data shows the reduction of radial temperature gradients to be important in reducing the amount of waste glass fracture. Cracking profiles show that internal carbon steel fins significantly decrease the degree of fracture. Different cooling rates only affect the amount of particles above a certain size. The amount of particles below that size depends on the interaction between the glass and canister.