The Uranium concentration leaching behaviour for UO2 and SIMFUEL in granite groundwater under oxidizing conditions is modelled. The model contemplates two terms: a source term and a loss term. The source term represents the bulk dissolution process and its upper limit is controlled by the solubility of the thermodynamically stable solid phase. The loss term emulates the solute removed due to groundwater flow into the repository. It can forecast the uranium steady state concentration when groundwater flow is considered. The model has been tested with three different types of leaching experiments, performed with UO2 and SIMFUEL. Good agreement between experimental results and model predictions are found.

Performance assessment calculations that support geologic disposal of spent nuclear fuel in a potential repository at Yucca Mountain, Nevada, are based in part on the assumption that 2% of the total inventories of 135Cs, 129I, and 99Tc are located in the gap and grain-boundary regions where they could dissolve rapidly if the spent fuel were to be contacted by groundwater. Actual measured values reported here for a few light-water reactor (LWR) spent fuels show that the combined gap and grain-boundary inventories of 129I approximately equaled the fission-gas release fractions. For 137Cs, the combined gap and grain-boundary inventories were approximately one third of the fission-gas release fractions. These measured values can be used to replace the 2% estimate and thus reduce the uncertainties in the calculations.

The melt-dilute treatment technology program is focused on the development and implementation of a treatment technology for diluting highly enriched (>20% 235U) aluminum spent nuclear fuel to low enriched levels (<20% 235U and qualifying the LEU Al-SNF form for geologic repository storage. In order to reduce the enrichment of these assemblies prior to ultimate geologic repository disposal, the melt-dilute technology proposes to melt these SNF assemblies and then dilute with additions of depleted uranium. The benefits accrued from this treatment process include the potential for significant volume reduction, reduced criticality potential, and the potential for enhanced SNF form characteristics. The emphasis within the development program to date has been on determining the process metallurgy and off-gas system design for the treatment of all types of Al SNF (UAIx, Al-U3O8, and AI-U3Si2). In determining the process metallurgy a wide range of alloys, representative of those expected in the Al-SNF form, have been fabricated and their product characteristics, namely microstructure, homogeneity, phase composition, and “ternary” constituent effects have been analyzed. As a result of the presence of species within the melt which will possess significant vapor pressures in the desired operating temperature range an off-gas system is necessary. Of the volatile species the one of greatest concern is 137Cs.

Specimens of aluminum-uranium alloys at 10 and 18 weight percent uranium were exposed to a saturated water vapor condition at 200 °C up to about 12 weeks and compared to previous results for aluminum 1100. The aluminum-uranium materials exhibited a range of initial corrosion rates and approached similar rates with the formation of a passive film of boehmite (Al2O3·H2O). The initial corrosion rates of the aluminum-uranium materials were one to four times higher than that for aluminum 1100. It is postulated that a micro-galvanic coupling between the large UA14 particles and the aluminum matrix has caused this difference. Sectioning the exposed specimens shows different characteristics of the oxide layers. In the oxide on the aluminum-10% uranium alloy (Al-10%U), small uranium aluminide particles can be seen in a boehmite matrix and do not seem to be corroded. The oxide film on the aluminum- 18% uranium alloy (Al-18%U) appears to have two distinct oxide layers. The outer layer has mass aggregates in a boehmite matrix, while the inner layer contains UA14 particles as in the case of Al-10%U.

The Viability Assessment (VA) of the potential repository for high-level nuclear waste at Yucca Mountain, Nevada was completed [1]. In the VA design concept, a two-layer waste container provides the primary component of the engineered barrier system (EBS). The VA reference design specifies a 100-mm thick carbon steel as the outer barrier and a 20-mm thick Alloy 22 as the inner barrier. A stochastic simulation model was developed to analyze long-term performance of the waste package in the potential repository. The model was developed by incorporating the latest corrosion data and models for the candidate waste package materials and the information developed from the Waste Package Degradation Expert Elicitation (WPDEE) [2].

The reference case results showed that only a small fraction of waste packages fail by localized corrosion (i.e., pit penetrations). The analysis also indicated the degradation mode that is most important to long-term waste package degradation is general corrosion (or passive dissolution) of the inner barrier under dripping conditions. However, the general corrosion rates for dripping conditions used in the analysis have a range over three to five orders of magnitude. This is due mainly to a lack of information on local chemical and electrochemical conditions on the inner barrier after the outer barrier breach. This paper discusses further the areas and issues that need improvement to reduce uncertainty in the waste package degradation analysis. It also discusses additional waste package degradation modes and associated processes that need to be included in the future analysis for the potential repository to enhance the confidence of the analysis.

The performance of Zircaloy nuclear fuel cladding in the environment of the proposed Yucca Mountain (YM) high-level waste (HLW) repository is evaluated. Because the uniform aqueous corrosion/oxidation rate is extremely slow, this evaluation focuses on mechanical failure, localized corrosion, stress corrosion cracking (SCC), hydrogen embrittlement, and initial failure. Mechanical failure is expected to result from (1) disruptive events such as rockfalls from seismicity, faulting, and igneous activities, (2) creep, and (3) splitting by oxidation of the spent fuel (SF) matrix. Effects of chloride ions and radiolysis are evaluated in localized corrosion and SCC. Embrittlement can be caused by delayed-hydride cracking (DHC) and hydride reorientation. Among these cladding failure modes, rockfalls, other disruptive events and initial defects can be important to performance prior to the container failure by corrosion. Confirmatory tests are required to evaluate the susceptibility to splitting by secondary mineral formation, localized corrosion, SCC, and hydride reorientation. Reliable temperature calculations with backfilling are also required. After breach, cladding may still assure slow release of radionuclides through perforations because localized failures may limit the exposure of the SF matrix and may provide high mass-transfer resistance. Failures prior to the repository emplacement arising from reactor operation, pool storage, dry storage, and transportation are also considered.

This paper summarizes some of the most recent total system performance assessment (TSPA) work supported by EPRI for the proposed repository at Yucca Mountain, Nevada. In Part I of the paper, (standard) TSPA analyses are presented. In Part II examples of two types of analyses that augment standard TSPAs are provided that suggest: (1) many components of the Yucca Mountain system contribute to overall hazard reduction; and (2) the biosphere dose conversions used, based on reasonably maximizing assumptions about future human behavior, provide a reasonable upper bound. These additional analyses should provide further confidence when considering more (standard) TSPA analyses.

To decide the final fate of high-level radioactive wastes that are in Krasnoyarsk Land, Russia, and, most importantly the wastes from a weapons complex at the Mining-Chemical Combine (MCC, Krasnoyarsk-26), suitable sites for geological disposal of the wastes are being sought in Middle Siberia. The ultimate goal is to select one site for development of a repository for these wastes. As a result of complex geological-geophysical work completed to date, the two most promising sites are in the northern part of the Nizhnekanskiy granitoid massif (the Yenisei Ridge). These sites are approximately 25 km from the MCC. Based on geotectonic analysis and electrical surveying data, these sites are composed of low-permeability biotite granite that extends to relatively great depth. At present, both sites are undergoing more detailed studies to define the most suitable one for detailed geological exploration. Radiogeochemical studies showed that granitoid rock samples from the massif exhibited high retention for major long-lived radionuclides in the wastes. The general results of the investigations are presented.

Performance assessment calculations for the proposed high level radioactive waste repository at Yucca Mountain, Nevada, were conducted using the Nuclear Regulatory Commission Total-System Performance Assessment (TPA 3.2) code to test conceptual models and parameter values for the source term based on data from the Peña Blanca, Mexico, natural analog site and based on a model for coprecipitation and solubility of secondary schoepite. In previous studies the value for the maximum constant oxidative alteration rate of uraninite at the Nopal I uranium body at Peña Blanca was estimated. Scaling this rate to the mass of uranium for the proposed Yucca Mountain repository yields an oxidative alteration rate of 22 kg yr−1, which was assumed to be an upper limit on the release rate from the proposed repository. A second model was developed assuming releases of radionuclides are based on the solubility of secondary schoepite as a function of temperature and solution chemistry. Releases of uranium are given by the product of uranium concentrations at equilibrium with schoepite and the flow of water through the waste packages. For both models, radionuclides other than uranium and those in the cladding and gap fraction were modeled to be released at a rate proportional to the uranium release rate, with additional elemental solubility limits applied. Performance assessment results using the Peña Blanca oxidation rate and schoepite solubility models for Yucca Mountain were compared to the TPA 3.2 base case model, in which release was based on laboratory studies of spent fuel dissolution, cladding and gap release, and solubility limits. Doses calculated using the release rate based on natural analog data and the schoepite solubility models were smaller than doses generated using the base case model. These results provide a degree of confidence in safety predictions using the base case model and an indication of how conservatism in the base case model may be reduced in future analyses.

Four researchers were asked to provide independent modeled estimates of the solubility of a radionuclide solid phase, specifically Pu(OH)4, under five specified sets of conditions. The objectives of the study were to assess the variability in the results obtained and to determine the primary causes for this variability.

In the exercise, modelers were supplied with the composition, pH and redox properties of the water and with a description of the mineralogy of the surrounding fracture system A standard thermodynamic data base was provided to all modelers. Each modeler was encouraged to use other data bases in addition to the standard data base and to try different approaches to solving the problem.

In all, about fifty approaches were used, some of which included a large number of solubility calculations. For each of the five test cases, the calculated solubilities from different approaches covered several orders of magnitude. The variability resulting firom the use of different thermodynamic data bases was in most cases, far smaller than that resulting from the use of different approaches to solving the problem.

Sensitivity analyses were conducted to evaluate the impact of a number of waste package and engineered barrier system (EBS) design options on long-term waste package degradation in the potential repository at Yucca Mountain, Nevada. The evaluations were performed by analyzing the waste package degradation in terms of the first breach and first pit-breach profiles with time. Results for each design option were assessed by comparing to results for the reference-case design. An important finding from the analyses is that the waste package lifetime could be extended substantially by limiting and delaying the contact of dripping water with the Alloy 22 corrosion resistant material (CRM) barrier. In the current model, this would keep the CRM barrier outside the conditions that makes the alloy susceptible to localized corrosion. Uncertainty in waste package degradation analysis would be reduced by improved understanding and characterization of the processes that create the local exposure condition on waste package.

The chemical composition of ambient groundwater for a geological, high level radioactive waste repository is of crucial significance to issues such as radioelement solubility limits, sorption, corrosion of the overpack, behavior of compacted clay buffers, and many other factors involved in repository safety assessment. At this time, there are no candidate repository sites established in Japan for the geological disposal of high-level radioactive waste, and only generic rock formations are under consideration. It is important that a small, but representative set of groundwater types be identified so that defensible models and data for generic repository performance assessment can be established. Over 15,000 separate analyses of Japanese groundwaters have been compiled into a data set for the purpose of evaluating the range of geochemical conditions for waste repositories in Japan. This paper demonstrates the use of a multivariate statistical analysis technique, principal component analysis (PCA), to derive a set of statistically based, representative groundwater categories from the multiple chemical components and temperature that characterize the deep Japanese groundwater analyses. PCA also can be used to guide the selection of groundwaters that could be used in scenario analyses of future geological events in Japan.

An approach is developed for including aspects of mechanistic models of radionuclide sorption into performance assessment (PA) calculations. Water chemistry data from the vicinity of Yucca Mountain (YM), Nevada are screened and used to calculate the ranges in key parameters that could exert control on radionuclide sorption behavior. Using a diffuse-layer surface complexation model, sorption parameters for Np(V) and U(VI) are calculated based on the chemistry of each water sample. Model results suggest that lognormal probability distribution functions (PDFs) of sorption parameters are appropriate for most of the samples; the total calculated range is almost five orders of magnitude for Np(V) sorption and nine orders of magnitude for U(VI) sorption, but most samples fall in a narrower range. Finally, statistical correlation between the calculated Np(V) and U(VI) sorption parameters can be included as input into PA sampling routines, so that the value selected for one radionuclide sorption parameter is conditioned by its statistical relationship to the others. The approaches outlined here can be adapted readily to current PA efforts, using site-specific information to provide geochemical constraints on PDFs for radionuclide transport parameters.

To investigate the effect of geochemical conditions of the repository on the release rate of radionuclides from an initially defective canister, a mathematical model is proposed. The model is based on the concept that ligands diffuse through the damage into the canister and form complexes with uranium, increasing its solubility. The complexes then diffuse out to the surroundings. Accordingly, the model combines a chemical equilibrium model for the solubility of uranium under strongly oxidizing conditions, and a transport model for the release rate of uranium. The transport model takes into account the diffusion resistance in the amage as well as in the surrounding compacted bentonite clay and the water flow in the rock. Using this model, the sensitivity of the release rate of aqueous uranium species to the concentrations of various groundwater component species, to the diffusion coefficients and to other parameters is systematically investigated. The results show that, the concentrations of carbonate, phosphate, silicate and calcium species in deep groundwaters play the most important role in determining the release rate of radionuclides. The results help identify the key diffusion coefficients that are needed to obtain more reliable estimates of the release rate.

The geochemical evolution of a clay engineered barrier subjected simultaneously to granitic groundwater on its external boundary and to a steady flux of iron released by canister corrosion on the other boundary is simulated for time periods up to 1000 y. A multisite ion-exchange model combined with dissolution/precipitation of accessory minerals (calcite, gypsum, quartz, goethite, pyrite, pyrrhotite, iron silicate, magnetite) is used to describe the geochemical evolution of the clay barrier. Diffusive transport of aqueous species and dissolved gases is simulated in a one dimension space. Retroaction of chemical reactions on the corrosion process of the canister is not included in the model. The temperature is assumed to be constant (25°C). Calculations are performed with the IMPACT geochemical coupled code. Two main scenarios are considered: (i) the reduction of sulfates (gypsum) is kinetically hindered, (ii) the reduction of sulfates is possible and leads to the precipitation of sulfides. Main results obtained are dealing with H2(gas) generation and control, pH transient and buffering effect of the clay barrier system. These simulations show that the hydrogen partial pressure in the clay barrier could possibly stay for long periods of time at a very low level due to interactions with mineral buffers or sinks. In the sulfate reduction scenario, a transient pH, xcursion is predicted at the vicinity of the canister. Future developpements of this work include the taking into account of a temperature gradient. Improvement of the geochemical model and reaction pathway is also necessary and will be done with the guidance of on going experimental work.

The degradation of concrete structures in a storage vault will release an alkaline plume influencing the geochemical evolution of clay engineered barriers or plugs. The studied configuration has a representative scale and is composed of a rectangular clay barrier interfaced on one side with a cement block and on the other side with an external granitic field. A clay model including major and accessory minerals is used; cement material is supposed to be made of portlandite and CSH minerals. The clay barrier and the cement block are assumed to be initially saturated and in equilibrium with their respective interstitial water. Diffusive transport of aqueous species coupled to chemistry is simulated in a one dimension space up to 10000 y. The feed-back of the porosity evolution on the transport properties is not taken into account. The temperature is supposed to be constant (25°C). Calculations are performed with the TRIO-EF coupled transport-chemistry code.

The main results deal with the evolution of the pH profile across the clay barrier, the dissolution of clay minerals and the precipitation of new solid phases in the clay such as albite, gibbsite and CSH. These simulations show that the pH excursion includes only a minor part of the clay barrier at 10 000 y and that the CSH precipitation retards the alkaline plume progression. The appearance of local accumulations of newly precipitated minerals might reduce strongly the porosity. Future developpements of this work include: (i) the completion of the geochemical model by introducing other minerals in concrete (other CSH phases and sulfo-aluminates) and ion exchange sites within the clay, (ii) the implementation of a precipitation/dissolution model in order to describe on a realistic way the porosity profile evolution and possible clogging phenomena.

A multi-scale, thermohydrologic (TH) modeling methodology has been developed that integrates the results from 1-, 2-, and 3-D drift-scale models and a 3-D mountain-scale model to calculate the near-field TH variables affecting the performance of the engineered barrier system (EBS) of the potential repository at Yucca Mountain. This information was used by Total System Performance Assessment—Viability Assessment (TSPA-VA) and is being used by the ongoing TSPA, supporting the License Application Design Selection, to assess waste-package (WP) corrosion, waste-form dissolution, and radionuclide transport in the EBS. Line-load WP spacing, which places WPs nearly end to end in widely spaced drifts, results in more locally intensive and uniform heating along drifts, causing hotter, drier, and more uniform conditions on WPs than point-load spacing, which is used in the VA design. Backfilling drifts with a granular material with coarse, well-sorted, nonporous grains (e.g., a coarse quartz sand) results in a large, persistent reduction in RH on WPs; point-load spacing allows only the medium-to-high-heat-output WPs to benefit from RH reduction, but line-load spacing enables all WPs to benefit.

DOE has proposed to add MgO to backfill in the WIPP repository to stabilize pH and reduce gas pressures and spallation. Experiments show that MgO dissolution rates are proportional to surface area, nearly independent of pH, and decrease with increasing dissolved NaCl. The probable repository MgO dissolution rate will be ≤ 10 to 20 μmol m2 min−1. Simulations indicate that MgO dissolution will buffer pH and provide stabilizing cement; however, under highly reducing conditions MgO will not control total gas pressures. Adding oxidizing agents or materials containing oxidized species, for example anhydrite, to the backfill, might control gas pressures.

Silica additives have been proposed to protect nuclear waste glass packages in deep geological repositories by presaturating silicon adsorption sites* in the media surrounding the package (engineered barrier, canister and overpack corrosion products). The durability of such additives was estimated using a one-dimensional dissolution-transport model developed at the CEA's Rhône Valley Research Center for the French reference glass in typical disposal sites. The silica additive would be placed between the wasteform and the overpack and thus dissolved before the waste glass. Its lifetime is proportional to M add 2(where M add is the added silica mass) up to a critical value M crit at which it increases considerably (from 103 or 104 to 108 years) due to saturation of the adsorption sites. Excess silica consumption is then limited by the typically very low (3 × 10−3 l·m−2y−1 ) groundwater flow in the geological site. After consumption of the added silica, the evolution of the altered waste glass fraction is comparable to the evolution without added silica if M add « M crit , but much lower if M add ≈ M crit or M add > M crit . These findings were confirmed by some experimental evidence, and establish that a few kilograms of added silica for a 400 kg package can considerably enhance the waste glass durability.

Matrix diffusion in rocks has been extensively studied experimentally with various methods including tracer experiments. Comparing the rock-capacity factors (α) for Cl-36 to corresponding values for H-3, an anion-exclusion tendency has been observed in many cases. This study uses a well-characterized reference material: porous silica (Geltech Inc.), with pore diameters of 2.5, 5, 7.5 and 20 nanometers. The equilibration-leaching method and tracers H-3 (H-3) and Cl-36 (Cl−) were used to determine the capacity factors of the samples in three molarities of NaCl solution: 0.0026, 0.01 and 0. 1 mol/liter. The ratio αcl/αH-3 gives the anion exclusion for each solution/pore diameter combination. A clear anion-exclusion tendency was seen both as a function of the decreasing pore diameter and liquid molarity.

The modeling was based on the assumption that the silica surface acts as a proton donor according to SiOH = SiO− + H+. The pore dimensions are also included in the calculations. In the case of the silica samples, good agreement was found when comparing the experimental and calculated values of chloride exclusion. This suggests that the surface complexation model for matrix diffusion can explain the observed exclusion phenomena in rocks although the parameters needed for the application of the model are difficult to obtain in the case of rocks.