In Belgium the possibilities to dispose of high-level waste in the plastic Boom Clay formation has been studied since 1975. Consequently many results of the site characterisation programme are already available. Various performance assessments have been carried out dealing with the disposal of high-level waste arising from reprocessing and with direct disposal of spent fuel. The performance assessment consists of two main steps: scenario development and consequence analyses. The scenario analysis is based on a catalogue of features, events and processes (FEPs) having the potential to influence the behaviour of the repository system. Two main groups of scenarios are distinguished. The normal evolution scenario, including a number of variants, treats the FEPs that are fairly sure to take place. Altered evolution scenarios focus on FEPs having a probability of occurrence lower than one but that might influence the performance of the repository system. For the impact analyses, a robust concept approach was introduced, which focused the analyses on a limited number of well-characterised barriers and processes. The impact analyses are complemented with sensitivity and uncertainty analyses based on deterministic and probabilistic approaches.

Recent advances in the development of reactive chemical transport simulators have made it possible to use these tools in performance assessments (PAs) for nuclear waste disposal. Reactive transport codes were used to evaluate the impacts of design modifications on the performance of two shallow subsurface disposal systems for low-level radioactive waste. The first disposal system, located at the Hanford site in Richland, Washington, is for disposal of lowlevel waste glass. Glass waste blocks will be disposed in subsurface trenches, surrounded by backfill material. Using different waste package sizes and layering had a small impact on technetium release rates to the vadose zone. The second disposal system involves a hypothetical repository for low-level waste in Italy. A model of uranium release from a grout waste form was developed using the STORM reactive transport code. Uranium is predicted to be relatively insoluble for several hundred years under the high-pH environment of the cement pore water. The effect of using different filler materials between the waste packages on uranium flux to the vadose zone proved to have a negligible impact on release rates.

We demonstrate the use of transmission electron microscopy (TEM) to study the effects of beta-decay of radioactive 137Cs to 137Ba in crystalline pollucite (CsAlSi2O6). Most prior work on radiation effects in materials has focused on structural damage from alpha radiation. Beta radiation, on the other hand, causes little atomic displacement, but the decay transmutation, that is, the radioactive decay of a radioisotope to an isotope of another element, results in progeny with different the valence and ionic radius. Cesium-137, a fission product of uranium, is a major contaminant at U.S. Department of Energy production facilities. Pollucite is an aluminosilicate ceramic with potential use for long-term storage of 137Cs. We focused on one of several available 137Cs sources originally fabricated in the 1970s and 1980s. These sources were small, sealed, stainless steel capsules containing pollucite in which varying amounts of the natural Cs had been replaced by radioactive 137Cs (t1/2 = 30.13 years). The sample chosen for TEM examination, aged for nearly 20 years, contained the most radiogenic barium and was expected to show the largest radiation effects. Bright field transmission images revealed a homogeneous crystalline matrix, with no evidence of distinct Ba phases or ex-solution phenomena resulting from the 137Cs transmutation. Electron diffraction patterns obtained from several portions of the sample were consistent with literature values for pollucite. These data suggest that little substantial damage was done to the crystal structure of this sample, despite the transmutation of nearly 1.5% of the total cesium to barium over the elapsed 20 years. Although our observations are limited, to our knowledge these are the only available data in which transmutation effects have been isolated from other radiation damage phenomena.

As a first step toward understanding the effects of radiation damage on element release rates from Ti-bearing waste forms, we performed single-pass flow-through (SPFT) experiments with 239Pu- and 238U-bearing ceramics over a solution pH-interval of 2 to 10 at 90°C. The ceramics tested are chemically complex and are dominated volumetrically by betafite (Ti-pyrochlore) (ABTi2O7) (A = Ca2+, Gd3+, B = Gd3+, Hf4+, Pu4+, U4/6+). The 239Pu-bearing specimens contained 11.9 mass% PuO2 and 23.7 mass% UO2. In addition, a 238Pu-bearing (11.8 and 23.9 mass% PuO2 and UO2, respectively) specimen was tested at pH = 2, 90°C. The 239Pu-bearing specimens slowly released Pu to solution (rate = 7.6x10-6 g m-2 d-1), even at pH = 2. Release of elements across the pH interval investigated exhibits a weak amphoteric behavior. Compared to results from the 239Pu-bearing materials, the 238Pu-bearing specimen released Pu >1000X faster (rate = 9.3x10-3 g m-2 d-1) at pH = 2. Release rates of U, Gd, and Hf are also faster from the 238Pu-bearing ceramic compared to the specimen containing 239Pu. Although preliminary, the data can be interpreted to indicate that accumulation of radiation damage may result in faster release of Pu and U to solution than previously suspected.

Ceramics rich in pyrochlore-structured titanate and fluorite-structured zirconate phases designed for surplus Pu immobilisation, with and without process impurities, have been leach tested at 90°C in deionised water. The zirconates consist mainly of a defect-fluorite with secondary impurity-containing phase-powellite/scheelite (when sintered in Ar but not when sintered in air), a spinel or magnetoplumbite type phase, a glass forming silicate and a secondary U-rich phase (when sintered in air with added impurities). The pyrochlore-rich baseline titanate ceramic consists of pyrochlore, brannerite and Hf-rutile. When impurities are added zirconolite and a silicate glass are also present. The pyrochlore-rich titanate with 5 wt% of impurities sintered at 1300°C is highly durable. A well-densified zirconate version without impurities has comparable elemental releases to those of the titanate ceramic but a zirconate with 5 wt% of impurities sintered at 1400°C in air or Ar shows much higher U and Ca releases than the titanate ceramic. Sintering atmospheres, changing from Ar to air, can influence Pu and U release rates up to an order of magnitude. High Ga releases from zirconates with impurities show that the secondary phase containing Ga is not durable. The higher processing temperature and the apparent inability to incorporate many impurity elements suggest that zirconates are not as flexible as titanates in respect of processing conditions and aqueous durability.

Radiation damage experiments are being performed with pyrochlore and zirconolite in support of the disposition of surplus weapons-ready Pu. Pyrochlore becomes amorphous in approximately 1 year from the alpha recoil damage of ∼ 1018 alphas/g from the decay of 238Pu. The dissolution rate of 238Pu-bearing ceramics increases with increasing radiation damage as measured in a 3 d MCC-1 test at 90°C. Over the same period, zirconolite retains substantial crystallinity albeit with broadened diffraction peaks. The dissolution rate also increases with increasing radiation damage.

Pyrochlore-rich ceramics containing additional zirconolite, brannerite and rutile have been developed for immobilizing excess weapons plutonium. This study reports data for a ceramic containing small amounts of Al and Mo in addition to the major oxide components of Ti, U, Ca, Hf, Gd and Ce. Hafnium and Gd are added as neutron absorbers, Al and Mo represent impurities. Quantitative electron microprobe data demonstrate that UO2 is strongly partitioned into brannerite (45 wt%), which is present as euhedral crystals with inclusions of unreacted UO2. Pyrochlore forms the groundmass and has an average UO2 content of 28 wt%. Zirconolite contains only 15 wt% UO2, but is significantly more effective in accommodating Hf and Gd than both brannerite and pyrochlore. Zirconolite incorporates U together with Hf and Ce in one structural site, whereas brannerite and pyrochlore accommodate U with Gd in a site that is distinct from that occupied by Hf. Incorporation of Gd into zirconolite takes place via a coupled substitution involving Al, thus explaining the high Al2O3 contents (3 wt%). Molybdenum was not detected in the major oxides, and it might be present in the accessory rutile. Although the studied waste form was designed to incorporate Pu, the present dataset is also valuable because immobilization of highly fissile U (e.g., U-233) might need to be considered in the future.

Radionuclide adsorption onto microbes, microbial retention in the engineered barrier system (EBS), and their potential release from the EBS as microbial colloids have been investigated. The microbial source term for these calculations was derived using MING V 1.0 software code [1]. Multiple model calculations from MING representing variations in possible microbial communities in the EBS were abstracted into two equations representing one meter segments of potential repository drift containing either commercial spent nuclear fuel (CSNF) or defense high level waste (HLW) packages. These two equations (Equations 1 and 2) represent the average cumulative microbial biomass generated in the EBS at any given time. A distribution for uranium uptake onto microbes (162.88 ± 133.05 mg U/gm dry cell) was applied to the microbial source term. The distribution was derived from the data set in Suzuki and Banfield [2] representing 45 different species of bacteria and fungus, covering uranium uptake at optimum pH values of 1 to 7. The mass of uranium sorbed onto the biomass was either sequestered in the EBS or transported as a microbial colloid based on a regression of data from Jewett et al. [3] representing microbial sorption onto air-water interfaces in unsaturated column experiments. Over one million years, it is estimated that EBS microbes may adsorb from 77 to 2302 kg of uranium [2302 kg U > 100% of the uranium available in a one meter segment of a CSNF waste package] per meter of waste package depending on the saturation of the invert and type of waste package. Over the same time, microbial colloids may transport from 8 to 1250 kg of adsorbed uranium per meter of waste package from the EBS.

In glass-bonded sodalite, which is the ceramic waste form (CWF) to immobilize radioactive electrorefiner salt from spent metallic reactor fuel, uranium and plutonium are found as 20-50 nm (U,Pu)O2 particles encapsulated in glass near glass-sodalite phase boundaries. In order to determine whether the (U,Pu)O2 affects the durability of the CWF, and to determine release behavior of uranium and plutonium during CWF corrosion, tests were conducted to measure the release of matrix and radioactive elements from crushed CWF samples into water and the properties of released plutonium. Released colloids have been characterized by sequential filtration of test solutions followed by elemental analysis, dynamic light scattering, transmission electron microscopy (TEM), and X-ray absorption spectroscopy. This paper reports the composition, size, and agglomeration of these colloids. Significant amounts of colloidal, amorphous aluminosilicates and smaller amounts of colloidal crystalline (U,Pu)O2 were identified in test solutions. The normalized releases of uranium and plutonium were significantly less than the normalized releases of matrix elements, i.e., the CWF retains these radionuclides well.