Performance assessments of geologic repositories for high-level nuclear waste will be used to determine regulatory compliance. The determination, that with a “reasonable expectation” regulatory limits are met, is based on the presumption that all of the relevant physical, chemical and biological processes have been modeled with enough accuracy to insure that a confident judgment of safety may be made. For the geologic disposal of high-level nuclear waste, this generally means that models must be capable of calculating radiation exposures to a specified population at distances of tens of kilometers for periods of tens to hundreds of thousands of years. A total system performance assessment will consist of a series of cascading models that are meant in toto to capture repository performance. There are numerous sources of uncertainty in these models: scenario uncertainty, conceptual model uncertainty and data uncertainty. These uncertainties will propagate through the analysis, and the uncertainty in the total system analysis must necessarily increase with time. For the highly-coupled, non-linear systems that are characteristic of many of the physical and chemical processes, one may anticipate emergent properties that cannot, in fact, be predicted. For all of these reasons, a performance assessment is not in and of itself a sufficient basis for determining the safety of a repository, but it remains a necessary part of the effort to develop a substantive understanding of a repository site.

The devitrification products of two simulated high level radioactive waste (HLW) glasses have been investigated. Magnox waste glass contained RuO2 as its primary crystal phase, Pd-Te inclusions, a Cr-, Fe- and Ni-rich spinel phase, a Si- and lanthanide-rich phase and also a Zr-rich phase which incorporated Ce and Gd. Upon heat treatment the glass developed a zektzerite-type phase, CeO2, a strontium molybdate (containing Nd and La) as well as a Si-rich phase. 75/25 glass, comprising a blend of reprocessing waste derived from UO2 and Magnox fuels, contained RuO2 as its primary crystalline phase, CeO2 and a Si- and Ru-rich phase. Heat treatment of this glass resulted in the growth of the CeO2 crystals, the development of a strontium molybdate (containing Nd and La) and a Si- and lanthanide-containing phase. A sodium lanthanum molybdate with a powellite-type structure formed on the surface of both glasses after heat treatment in air.

Carbon steel is one of the candidate overpack materials for high-level waste disposal and is expected to assure complete containment of vitrified waste glass during an initial period of 1000 years in Japan. The lifetime of the carbon steel overpack will depend on its corrosion rate. The corrosion rate of carbon steel is reduced by the presence of buffer material such as bentonite. Buffer material will delay the supply of corrosive materials and discharge of corrosion products through it. Carbon steeloverpack will be corroded by consuming oxygen introduced by repository construction after closure of repository and then will keep the reducing environment in the vicinity of repository. Therefore, it is important to study the migration of iron corrosion products through the buffer material because it may affect the corrosion rate of overpack, migration of redox-sensitive radionuclides, and the properties of the buffer material. Electromigration experiments have been carried out with source of iron ions supplied byanode corrosion of iron coupon in compacted bentonite. The carbon steel coupon was connected as the working electrode to the potentiostat and was held at a constant applied potential between - 200 to 1000 mV vs. Ag/AgCl electrode for 48 hours. Corrosion currents were 0.5 to 2mA initially and depended on the supplied electrical potential, then decreased to approximately 0.1 mA in a few hours. The final corrosion current was independent of supplied electrical potential. It is expected that iron ion could migrate as ferrous ion through interlayer of montmorillonite replacing exchangeable sodium ions in the interlayer. The rate-determining process of this experimental configuration could be infiltration rate of ferrousioninto bentonite. Infiltration rate of ferrous ion into bentonite was increasing with dry density of bentonite.

The Sverjensky-Molling linear free energy relationship was originally developed to correlate the Gibbs free energies of formation of an isostrutural family of solid phases to the thermodynamic properties of aqueous cations. In this paper, we demonstrate that the similar relationship also exists between metal complexes and simple metal cations in aqueous solutions. We extend the Sverjensky-Molling relationship to predict the Gibbs free energies of formation or dissociation constants for a family of metal complexes with a given complexing ligand. The discrepancies between the predicted and experimental data are generally less than 1.5 kcal/mol (or one log unit for stability constants). The use of this linear free energy correlation can significantly enhance our ability to predict the speciation, mobility, and toxicity of heavy metals in natural environments. According the obtained results, Gibbs free energies of formation of cations (δG0f, Mn+) can be used as an indicator for the hardness/softness of a metal cation (acid). The higher negative value of a metal cation, the harder acid it will be. It is logical to postulate that the coefficient a*ML characterizes the softness of a complexing ligand (base).

Diffuse reflectance spectroscopy measurements at ambient temperatures have been made over the near infrared-visible range (5000-25000 cm−1) on polycrystalline ceramic zirconolite (CaZrTi2O7) and perovskite (CaTiO3) samples doped with Pu4+. The Pu concentrations were varied between 0.001 and 0.1 formula units. The Pu ions gave rise to a number of unresolved intraconfigurational f-f electronic absorption bands of a few hundred cm−1 bandwidth. Pu ions were targeted to substitute in the Ca sites as either trivalent or tetravalent species and as tetravalent species in the Zr site of zirconolite by the appropriate choice of charge compensation and firing atmosphere. There was approximate agreement of the Kubelka-Munk absorption intensities with Beer's Law for the different Pu4+ substitution schemes, apart from some “new” bands, attributed to impurities, observed in the most dilute zirconolite sample. No clear spectral differences were evident when Pu4+ was targeted to Ca or Zr sites in zirconolite. Samples prepared in reducing atmospheres with a view to producing Pu3+ were strongly absorbing, leading to suppression of Pu transitions.

As a result of the development of a reference glass formulation for the immobilization of West-Valley nuclear waste, a large number of glass compositions have been tested under a variety of leaching conditions for extended durations. In this work, data from the standard PCT leaching procedure (deionized water, 90°C, S/V = 2000 m-1) are evaluated for a subset of 98 of these glasses, for which the alteration times extend up to fifteen years. Analysis of the leachate data over this extended period has led to a classification of the glass compositions according to their PCT behavior, as characterized by the boron release in solution, the decrease in the rate of alteration, or the occurrence of a resumption of alteration. Three types of behavior are associated with a normalized mass loss of boron after 100 days of around 1, 2, or 5 g m-2 respectively, and a ratio of the rate of alteration at 7 days to the rate at 1000 days of about 30. A fourth type is characterized by a much larger decrease between the rates at 7 and 1000 days with a ratio exceeding 120. Four further categories are associated with late-stage resumption of glass alteration. A statistical analysis of the data has been used to model the leachate data as a function of glass composition. The effects of seven major elements (Si, B, Al, Li, Na, K, Fe) on the long-term glass durability have been studied. As a result of this analysis, the different types of leaching behaviors can be predicted with a good reliability from the composition of the pristine glass.

A rapid and reliable method for quantification of transuranium elements in environmental samples is required from the aspect of emergency preparedness. Determination of these elements in e.g. sediment samples includes three steps: sample dissolution, chemical separation and measurement. Different acids were used for leaching of plutonium, americium and neptunium from contaminated sediment (during heating or microwave digestion). Chemical separation was performed by co-precipitation (with Fe(OH)3 or NdF3), by anion exchange or by liquid-liquid extraction (with tributylphosphate, trioctylamine and thenoyl trifluoroacetone, respectively). The following radioisotopes were quantified by quadrupole ICP-MS equipped with an ultrasonic nebulizer (USN): 237Np, 239Pu, 240Pu, 241Pu, 242Pu and 241Am. Data are only given for 239Pu and 240Pu in this paper. Plutonium was also measured by alpha spectrometry in some samples, as well as 137Cs and 241Am by gamma spectrometry. The agreement between concentrations of plutonium obtained by ICP-MS and alpha spectrometry was good, which demonstrates that a simple and rapid procedure for analysis of plutonium by quadrupole ICP-MS is feasible.

This paper presents the results obtained from coprecipitation experiments of uranyl solutions in the presence of metallic iron and/or its alteration phases in synthetic bentonitic-granitic composition water. Experiments were done under oxidising conditions at room temperature. The pH range covered was 7.4 – 8.8. Changes in the uranium concentrations and the characterisation of the secondary phases formed in the experiments were done using XRD and SEM-EDS and are presented herein.

Final uranium concentration values were in the range of 2·10−5 – 5·10−4 mol (kg of H2O)−1. In all cases, results from these experiments did not show evidence of a clear effect due to the presence of iron (metallic or previously corroded) on the uranium concentration. These data were similar to those obtained by coprecipitation in similar conditions but in absence of iron material. Boltwoodite was observed [K2(UO2)2(SiO3)2(OH)2·3H2O] in iron surface materials and characterized in all experiments. Based on the experimental data obtained (uranium concentration in solution and bulk solid phase characterisation) we propose the following surface-mediated reaction:

2 K+ + 2 H4SiO4 + 2 UO2+2 + 3 H2O ⇔ K2(UO2)2(SiO3)2(OH)2·3H2O + 6 H+

The groundwater at the Nevada Test Site (NTS) contains many long-lived radionuclides, including 99Tc (technetium) and 129I (iodine), as a result of 828 underground nuclear weapons tests conducted between 1951 and 1992. We synthesized a body of data collected on the distribution of 99Tc and 129I in groundwater to assess their migration at NTS, at field scales over distances of hundreds of meters and for durations up to forty years and under hydrogeologic conditions very similar to the proposed geological repository at Yucca Mountain. The results of our study show that Tc does not necessarily exist as a mobile and conservative species TcO4−, as has been commonly assumed. This conclusion is corroborated by recent in situ redox potential measurements, which show that groundwaters at multiple locations of the NTS are not oxidizing, and mobility of reduced Tc species (TcO2·nH2O) is greatly decreased. Speciation of iodine and its associated reactivity is also complex in the groundwater at the NTS, and its effect on the mobility of iodine should be the subject of future studies.

An assessment of long-term performance of low level waste-enclosing cement grouts requires diffusivity data for radionuclide species such as, 129I and 99Tc. The diffusivity of radionuclides in soil and concrete media was collected by conducting soil-soil and concrete-soil half-cell experiments. The soil diffusivity coefficients for iodide were 7.03 × 10−8 cm2/s and 2.42 × 10−7cm2/s for soils at 4% and 7% moisture contents, respectively. Iodide diffusivity in soil is a function of moisture content and is about an order of magnitude slower at lower moisture content. The soil diffusivity coefficients for 99Tc were 5.89±0.80 × 10−8 cm2/s (4% moisture content) and 2.04±0.57 × 10−7 cm2/s (7% moisture content), respectively. The soil diffusivity of iodide and 99Tc were similar in magnitude at both water contents, indicating that these ions have similar diffusion mechanisms in unsaturated coarse-textured Hanford soil. The diffusivity of iodide in concrete ranged from 2.07 × 10−14 cm2/s (4% soil moisture content) to 1.31 × 10−12 cm2/s (7% soil moisture content), indicating that under unsaturated soil moisture conditions, iodide diffusivity is highly sensitive to changing soil moisture conditions. Depending on the soil moisture content, the diffusivity of 99Tc in concrete ranged from 4.54 × 10−13 cm2/s to 8.02 × 10−12 cm2/s. At 4% soil moisture content, iodide diffused about 20 times more slowly than 99Tc, and at 7% soil moisture content, iodide in concrete diffused about 6 times slower than 99Tc.

An engineered ion exchange material has been prepared for the specific purpose of removing radioactive cesium from acidic waste. Separating the fission product 137Cs from the bulk of the nuclear waste stream is often advantageous because, after typical cooling times, this isotope is usually the primary source of gamma radiation dose. The engineered ion exchanger was prepared using ammonium molybdophosphate impregnated into hollow glass crystalline microspheres. The microspheres or cenospheres, are refractory compounds of silica and alumina that are derived from the fly ash produced in coal combustion. This paper describes equilibrium experiments that were conducted with the engineered ion exchanger and a simulated acidic waste solution. These tests indicate that the new material has a high capacity and selectivity for cesium in these matrices.

Licensing a geologic repository for nuclear waste will be based primarily upon the results of a performance assessment (PA), which evaluates compliance of the repository with radiation exposure limits by modeling the engineered and natural barrier systems over extended periods (104 to 106 years). The properly completed PA must include an analysis of the sources of uncertainty, as well as the range of input parameter values, in order to determine whether the analysis is “robust”, “realistic” or “conservative”. This paper examines the sources of uncertainty that may be encountered in source-term and near-field models in a performance assessment of a repository for spent nuclear fuel by examining the uncertainties in modeling the geochemical behavior of the Bangombé natural reactor in Gabon.

The sorption of radionuclides onto mineral-water interfaces is an important mechanism of radionuclide retardation in subsurface environments. Although a large body of data on metal sorption has been accumulated, the interaction of radionuclides at the mineral-water interface is still not well quantified. In this paper, we establish a linear free energy relationship that can correlate intrinsic metal sorption constants to the known thermodynamic properties of metal cations and mineral phases. This relationship can be used to predict the intrinsic sorption constants of metal cations adsorption on various mineral-water interfaces. Based on a surface complexation model, an adsorption constant of a cation Mn+ onto a mineral surface can be represented by the sum of “intrinsic” and coulombic terms, that is logKads, Mn+ = logKint, Mn+ - (nF/2.303RT)0. To construct our linear free energy relationship, we further decompose an intrinsic sorption constant into the solvation contribution, which characterizes the work to be done to move a cation from one dielectric medium (solution) to another (interface), and the chemical contribution, which characterizes the chemical bonding abilities of cations to mineral surface sites. The overall intrinsic sorption constant can be calculated by the equation of:

2.303RTlogKint, Mn+ = -ΩMn+ (1/εk) - a* δG0n, Mn+ - b* - β*rMn+ + δG0f, Mn+, or

2.303 RT logKint, Mn+ = -ΩMn+ (1/εk) + a* δG0s, Mn+ + (1 - a*)δG0f, Mn+ - b* - β* rMn+,

where, ΩMn+ is interfacial Born solvation coefficient of cation M, εk is the dielectric constant of a mineral; δG0n, Mn+, δG0s, Mn+ and δG0f, Mn+ are the non-solvation energy, solvation energy and the Gibbs free energy formation of a cation, respectively; rMn+ is the ionic radius of a cation; and a*, β*, and b* are constants, which can be determined by fitting the equation to the existing experimental data. We have applied the equation to the sorption of divalent cations on oxide, hydroxide, and silicate minerals. In particular, based on additivity of molecular polarizabilities, we have used the equation to calculate the intrinsic sorption constants of divalent metals and radionuclides on interstratified clay minerals that are commonly found in the nature or are proposed as backfills for nuclear waste geologic repositories. The discrepancies between calculated and measured values are generally within 0.8 log unit.

The elastic and plastic stress distribution in strained Si1−xGex heterostructure (x=0.2 and x=0.5), was investigated during the growth process using a nonlinear transient finite element modeling. The material plastic behavior is described by the von Mises yield criteria coupled with isotropic work hardening conditions. The calculated stress distribution during growth of the SiGe cap layer shows that the von Mises stress fluctuates strongly within the layers and at the interfaces. The surface of constant composition Si1−xGex layer is found under compressive stress in both cases. Within that layer the normal stress in the growth direction remains compressive for x=0.2, while it changes from compressive to tensile for x=0.5. In the graded layer the stress goes from tensile to compressive for x=0.5 and in the opposite way for x=0.2. High plastic deformation is observed in the layers, with the maximum von Mises plastic stress being higher for x=0.5 and localized in the SiGe graded region. The plastic strain vanishes monotonically up to 8 μ deep into the Si bulk substrate, in agreement with TEM images that revealed dislocation loops penetrating into the substrate. The time dependent analysis shows that elastic and plastic deformation appear almost instantaneously in the sublayers, while in the Si substrate it is delayed up to 300 s.

Technetium-99 99Tc) is an important radionuclide in repository models owing to its relatively long half-life and the high aqueous solubility of compounds where it is inthe heptavalent state. The vast majority of the 99Tc inventory presently slatedfor disposal is contained in oxide commercial spent nuclear fuel (CSNF), where it is divided (along with its cohort, molybdenum) betweenexsolved, intermetallic “epsilon” particles and isolated, likely oxidized, atoms distributed in the uranium dioxide matrix. We present recent evidence from synchrotron x-ray fluorescence spectroscopy, electron microscopy, and fuel dissolution testing on the likely oxidation state, coordination environment, and physical disposition of technetium and molybdenum in CSNF. Effects of the relative proportioningof technetium and molybdenum among the metallic and oxidized states in CSNF, andtheir distribution in or near grain boundaries and gaps on release during CSNF corrosion testing are discussed.

In an effort to enhance safety for long time disposal of waste nuclear fuel, friction stir welding has been developed as one alternative to seal copper canisters. To avoid the formation of voids and cracks during the welding process, an understanding of the heat and material flow andthereby the evolution of the microstructure, is of great importance. Finite element modelling has been used to simulate the heat and material flow as well as thermal expansion during the friction stir welding process. A model involving heat transfer, material flow, and continuum mechanics has been developed. The steady state solutions have been compared with experimental temperature observations as well as analytical solutions, showing good agreement. Temperature distribution is affected by the welding speed. For a given reference pointperpendicular to the welding direction, a lower welding speed corresponds to a higher peak temperature. The plunging position of welding tool influences the temperature distribution and therefore the displacement distribution of the weldment.

We have characterized considerable amounts of the two known uranyl peroxide phases that formed under static immersion conditions on commercial spent nuclear fuel (CSNF) samples. Milligrams of corroded fuel aggregates were observed at the air-water interface in each sample. The bulk fuel and the interfacial solids were examined by SEM, EDX, and XRD and were found to contain studtite and metastudtite, respectively. The reason for the partitioning of the two phases is not clear at this time. The unique occurrence of the floating phase prompted a radiochemical analysis of these solids. The analysis indicated that 90Sr, 137Cs, 99Tc, and to a lesser extent of 238,239Pu and 237Np, had partitioned with the air-water interface aggregates. The concentration of 241Am in the interfacial solids was two orders of magnitude lower than the inventory in the fuel prior to contact with water. The radiochemical analyses of two fuel leachate samples are compared to reported leaching data of a similar fuel which did not result in the formation of studtite.

In performance assessments, sorption of radionuclides dissolved in groundwater is mostly handled by the use of fixed Kd values. It has been well known that this approach is unsatisfying. Only during the last few years, however, tools have become available that make it possible to predict the actual Kd value in an aqueous solution that differs from the one in which the sorption properties were measured.

One such approach is surface complexation (SC) that gives a detailed knowledge of the sorption properties. In SC, one tries to find what kinds of sorbed species are available on the surface and the thermodynamics for their formation from species in the bulk aqueous solution. Recently, a different approach, surface phase method (SP), has been developed. In SP, a thin layer including the surface is treated as a separate phase. In the bulk aqueous solution, the surface phase is treated as a virtual component, and from the chemical potential of this component, the sorption properties can be found.

In the paper, we compare advantages and disadvantages of the two kinds of models. We also investigate the differences in predicted sorption properties of a number of radionuclides (Co, Np, Th and U). Furthermore, we discuss under which circumstances, one approach or the other is preferable.

A Monte Carlo-based approach has been adopted for development of a chemical thermodynamic model to describe the goethite surface in contact with sodium nitrate solutions. The technique involves the calculation of the goethite surface properties for the chemical conditions corresponding to each experimental data point. The representation of the surface was based on a set of model parameters, each of which was either fixed or was randomly sampled from a specified range of values. Thousands of such model representations were generated for different selected sets of parameter values with the use of the standard geochemical speciation computer program, HARPHRQ. The method allowed many combinations of parameter values to be sampled that might not be achieved with a simple least-squares fitting approach. It also allowed the dependence of the quality of fit on each parameter to be analysed. The Monte Carlo approach is most appropriate in the development of complex models involving the fitting of several datasets with several fitting parameters.

Introduction of selenate surface complexes allowed the model to be extended to represent selenate ion sorption, selenium being an important radioelement in evaluation of the long-term safety of ILW disposal. The sorption model gave good agreement with a wide range of experimental sorption datasets for selenate.

Researchers at the Pacific Northwest National Laboratory (PNNL) investigated the effects of gadolinium oxide concentration on the air oxidization of gadolinium oxide-doped uranium dioxide using thermogravimetry and differential scanning calorimetry to determine if such doping could improve uraniumdioxide's stability as a nuclear fuel during potential accident scenarios in a nuclear reactor or during long-term disposal. We undertook this study to determine whether the resistance of the uranium dioxide to oxidation to the orthorhombic U3O8 with its attendant crystal expansion could be prevented by addition of gadolinium oxide. Our studies found that gadolinium has little effect on the thermal initiation of the first step of the reported two-step air oxidation of UO2; however, increasing gadolinium oxide content does stabilize the initial tetragonal or cubic product allowing significant oxidation before the second expansive step to U3O8 begins.