Groundwater travel time (GWTT) estimation at a potential high-level waste (HLW) repository is subject to various technical uncertainties. These uncertainties stem from model and data uncertainties and cannot be resolved with field tests because of the long time (> 1,000 yr) and large space (> 5,000 m) scales involved. Therefore, computational methods for demonstrating and determining compliance with the GWTT rule will be used. Stochastic theory based approaches constitute a natural framework for performing GWTT estimations under conditions of uncertain and/or limited data.

This study employs the generation of spatially correlated hydraulic conductivity fields by the Nearest Neighbor Model (NNM). Repeated (Monte Carlo) realizations of the statistically equivalent random fields are obtained, and the saturated steady-state groundwater flow equation is solved. These results are then used to estimate GWTT along particular paths by releasing a large number of water particles at various starting points. By doing so, path variability is sampled through the realization ensemble space and also through the independent particle “flights” within a specific flow field realization. The uncertainty in predicted GWTT due to parameter variability is assessed for a data set characteristic of the saturated zone at Yucca Mountain at three levels of parameter heterogeneity. However, since several parameters and the boundary conditions of the problem have been arbitrarily assumed, direct conclusions regarding the proposed Yucca Mountain site cannot be drawn from this study.

The UK current plans for a repository for the disposal of intermediate-level radioactive waste involve the use of a cementitious backfill within the repository vaults. Radionuclide transport could be concentrated within cracks that may form in the cement, or in fractures within the rock surrounding the repository. A version of CHEQMATE (CHemical EQuilibrium with Migration And Transport Equations) that includes a representation of transport through fractured media has been developed. This paper describes modifications of the code, to allow simulation of transport through either a single fracture, or, by adopting a continuum approach to represent fracture distribution, through multiple fractures. Successful verification of the modified program against analytical solutions for transport through fractured rock and cracked cementitious media is presented. An application of the modified program to study the evolution of the aqueous chemistry within a repository vault containing a single fracture through the cementitious backfill is also discussed.

One of the requirements for the performance of waste packages prescribed in 10CFR 60.113 is that the high level waste must be “substantially completely” contained for a minimum period of 300 to 1000 years. During this period, the radiation and thermal conditions in the engineered barrier system and the near-field environment are dominated by fission product decay. In the present U.S design of the engineered barrier system, the outer container plays a dominant role in maintaining radionuclide containment. A quantitative methodology for analyzing the performance of the container is described in this paper. This methodology enables prediction of the evolution of the waste package environment in terms of temperature fields, stability of liquid water on the container surface, and concentration of aggressive ions such as chloride. The initiation and propagation of localized corrosion is determined by the corrosion potential of the container material and critical potentials for localized corrosion. The coiTOsion potential is estimated from the kinetics of the anodic and cathodic reactions including oxygen diffusion through scale layers formed on the container surface. The methodology described is applicable to a wide range of metals, alloys and environmental conditions.

A mathematical model to predict the cumulative failure distribution for the engineered barrier system employed in a deep geologic disposal facility as a function of container design and near-field environmental conditions has been developed. The model employs Weibull and exponential distributions to describe cumulative container failures as a function of time. Parameter values employed in the model are based upon simple, time-dependent, mechanistic models and relevant corrosion data, which describe failure of individual components of the container as a function of environmental conditions.

Recent developments in container design for the Yucca Mountain site center on the possible deployment of Multi-Purpose Containers (MPC). These containers will be designed and constructed to serve as transport casks, interim storage containers, and disposal containers. The current container performance assessment model is applied to evaluate the long-term performance of various MPC designs under the areal power density and heat transfer regimes expected in the Yucca Mountain environment. This model has previously been employed to describe the performance of the container as one part of a risk-based performance assessment of the Yucca Mountain site.

The relative importance of container design, areal power density, and dominant heat transfer mode on predicted MPC performance is demonstrated through comparison of the cumulative container failure distributions for each MPC design when exposed to expected Yucca Mountain environmental conditions.

Previous work in our laboratory and elsewhere has suggested that MIC of stainless steels and nickel-base alloys occurs in locally anaerobic regions that support the growth of sulfate reducing bacteria (SRB). The cathodic reaction is provided by oxygen reduction at remote sites. Such a coupling between anode and cathode is difficult to reproduce in the laboratory, but can be simulated indirectly using a double electrochemical cell, as in previous work. A more realistic simulation using a single aerated electrochemical cell has now been developed, in which a second organism (P. aeruginosa) is used to provide an anoxic habitat for SRB growth and possibly a source of organic carbon, within a layer of silt. A bare alloy electrode is used as the oxygen cathode. Tests of this kind using rigorous microbiological procedures have generated pitting corrosion of several alloys in low chloride media simulating freshwater heat exchanger conditions. Similar test procedures are applicable to other environments of interest to this symposium.

The parameters that must be known if the extent of damage due to the crevice corrosion of Grade-2 titanium waste containers is to be predicted are discussed. These parameters include the kinetics of crevice propagation as a function of temperature and oxygen concentration, the critical values of temperature and oxygen concentration for which repassivation occurs, the rate of hydrogen absorption by the metal under crevice corrosion conditions, and the threshold value of absorbed hydrogen concentration above which the material may be susceptible to hydrogen-induced cracking. The results of experiments to measure these parameters are briefly summarized, and their significance as modelling parameters or criteria are discussed.

The corrosive environment around the containers in a Canadian nuclear fuel waste disposal vault will change over time from “warm and oxidizing” to “cool and anoxic”. As the conditions change, so too will the corrosion behaviour of the containers. For copper containers, uniform corrosion and, possibly, pitting will occur during the initial aggressive phase, to be replaced by slow uniform corrosion during the long-term anoxic period.

The corrosion behaviour of copper has been studied over a range of conditions representing all phases in the evolution of the vault environment. The results of these studies are summarized and used to illustrate how a model can be developed to predict the corrosion behaviour and container lifetimes over long periods of time. Lifetimes in excess of 106 a are predicted for 25-mm-thick. copper containers under Canadian disposal conditions.

This paper concludes that a 70 wt% salt/30 wt% bentonite mixture is preferable to pure crushed salt as backfill for disposal rooms in the Waste Isolation Pilot Plant. The performance of two backfill materials is examined with regard to various selection criteria related to compliance with the transuranic radioactive waste standard 40 CFR 191, Subpart B, such as the need for low liquid permeability after closure, chemical stability, strength, ease of emplacement, and sorption potential for brine and radionuclides. Both salt and salt/bentonite are expected to consolidate to a final state of permeability ≤ 10-18 m2, which is adequate for satisfying government regulations for nuclear repositories. The real advantage of the salt/bentonite backfill depends, therefore, on bentonite’s potential for sorbing brine and radionuclides. Estimates of the impact of these properties on backfill performance are presented.

In an earlier laboratory study1, ion exchange processes and smectite alteration were investigated through the interaction between compacted Na-bentonite (Volclay MX-80) and simulated granitic groundwater solutions. The change of montmorillonite from Na- to Ca-rich was found to be the major alteration process in the bentonite. In the water, a concentration decrease in Ca, Mg and K, and an increase in Na, HCO3 and SO4 were recorded. The total amount of Ca available in the water, however, was found insufficient to account for the recorded formation of Ca-smectite, and it is therefore assumed that the accessory Ca-bearing minerals in the bentonite provide the fundamental source of these cations.

X-ray powder diffraction analyses and microscope observations of the bentonite samples were re-conducted. Quartz, feldspars, pyrite, calcite and traces of gypsum were revealed as the primary accessories. In reacted samples, goethite and siderite are found as the secondary mineral products in association with corroded pyrite grains, while calcite and gypsum tend to disappear. From these results it is assumed that the oxygen present in the water and bentonite pore space promotes the oxidation of pyrite (dissolved) and the precipitation of goethite. The pore water pH decreases and calcite is partly dissolved. Through the dissolution, the bulk amount of Ca ions in addition to those arising by diffusion from the water is provided. Some of the reaction-released bicarbonate and Fe2+ re-precipitate in the bentonite as siderite, while the rest (as also SO4 ions) diffuse into the water.

Though the relative oxygen content in the experiment may be considered higher than that of the repository concept for nuclear fuel disposal (interaction in a semi-closed system with high water/bentonite ratio), the near field geochemistry predictions imply limited oxidizing conditions, which will be characterized by the above-described processes in sulphide-bearing bentonite and occur for some time after the sealing of the repository.

Several series of one dimensional tests were performed to examine the moisture and soil water potential distributions in the buffer material compacted to a dry density of 1.67 Mg/m3 and volumetric water content of 0.28. Diffusivity parameters were calculated using the measured moisture profiles combined with the finite difference method. Powell’s optimization was used to determine the material parameters.

It is experimentally demonstrated that the moisture distribution is highly dependent on the test boundary conditions. When volume change was allowed to take place during infiltration process, the buffer material adsorbs more water than in the case of no volume change. This is attributed to an increase in hydraulic conductivity by one order of magnitude. As the volumetric water content increases soil water diffusivity increases and soil water potential decreases.

The interaction of a saponitic clay from the Madrid Basin with synthetic groundwater (GW) simulating those from granitic areas is investigated. The experiments performed explore the reaction of clay with GW (1:3 solid to water ratio) at 45, 60 and 90°C. Water extracted from the clay slurries has been analyzed and water of this composition synthesized again to run a new reaction with the original clay (salt concentration process). The clay product of the first reaction was reacted again with GW (leaching process). These reactions were repeated in successive equilibrating steps.

Multicationic selectivity coefficients have been calculated. K > Ca > Mg > Na is the order of preference for adsorption by the clay. The selectivity of Ca is highly enhanced in both processes, the clay acting as a true complexing agent for Ca, as calcite controls the carbonate equilibrium. A clear dependence of the exchange constants on temperature with the range explored was not detected. The GW composition evolves from Ca,Na-HCO3 to Na,Mg-HCO3 (leaching) and NaCl (concentration), the solutions being buffered at pH 8.0. Mechanical properties, clay mineralogy and bulk adsorption properties remained virtually unchanged in the successive one month experiments although a slight decrease in BET specific surface area indicated some sepiolite dissolution at 90°C.

In a high-level waste repository, a carbon steel overpack will be corroded by consuming oxygen trapped in the repository after closure. This will create a reducing environment in the vicinity of repository. Reducing conditions are expected to retard the migration of redox-sensitive radionuclides such as uranium.

The apparent diffusivities of uranium were measured in compacted bentonite (Kunigel VI®, Japan) in contact with carbon steel under reducing conditions or without carbon steel under oxidizing conditions for comparison. The apparent diffusivities of uranium were 3.5 × 10-14 to 1.1 × 10-13 m2/s under reducing conditions and 9.0 × 10-13 to 1.4 × 10-12 m2/s under oxidizing conditions. There was no significant effect of dry density (1.6 to 1.8 g/cm3) and silica sand (0 or 40%) on the apparent diffusivities.

Since the bentonite pore water would be buffered at a pH between 8 and 9, uranium in the bentonite pore water would probably exist as a neutral hydroxide complex under reducing conditions and as an anioníc carbonate or hydroxide complex under oxidizing conditions. The anion exclusion theory cannot explain the difference of diffusivities between the two conditions. The uranium concentrations in bentonite under oxidizing conditions were one order of magnitude higher than those under the reducing conditions. The uranium concentration in the bentonite pore water under the reducing condition is estimated to be two orders of magnitude lower than that under the oxidizing conditions under the assumption of diffusion in porous media.

The reaction of unaltered olivine rock with groundwater and redox-sensitive elements (Tc, Re, and U) was investigated at elevated temperatures in the laboratory. Olivine was found to react easily with water taking up trace elements in alteration products. The alteration products were characterized using Rutherford backscattering, secondary ion mass spectrometry and electron microprobe analysis. The sample surfaces were studied by autoradiography and digital image processing techniques.

It has been proposed that smectite clays, the predominant minerals in bentonite, are metastable solid solutions whose compositional heterogeneity prevents strict adherence to fixed ion activities and activity ratios characterizing invariant equilibria among stoichiometric phases [1,2]. This is qualitatively confirmed in the present analysis of exceptionally well-constrained experimental data [3] using a solid-solution model defined by the phase rule and estimated ideal site-mixing parameters. The results suggest that equilibria are approached among smectite solid solutions and kaolinite or halloysite, and that these minerals coexist metastably with respect to asiO2(aq)· Irreducible uncertainties are generated in the model by analytical and conceptual deficiencies in understanding compositional variability in smectite. However, their estimated effects on smectite’s stability are relatively small, and are comparable to the effects of experimental uncertainty in standard Gibbs energies on the stabilities of stoichiometric minerals.

An alternate analysis of the data further confirms that smectite does not behave like a stoichiometric phase [3]. Ion-exchange models for this clay mineral may therefore be thermodynamically ill defined because stoichiometric behavior is assumed implicitly under limiting conditions of fixed proportions of cations on exchange sites. Ion exchange is a pragmatic simplification enabling empirical analysis of some experimental data. However, its empirical, rather than thermodynamic, basis should not be overextended to conditions that are beyond an experimentally calibrated range.

Static fatigue behavior is one of the important factors for life prediciton of ceramic materials. In this study, SCG (slow crack growth) parameters were measured under atmosphere conditions, and the static fatigue behavior of alumina, PSZ (partially stabilized zirconia), and titanium oxide was examined.

According to the results of the evaluatin of the static fatigue behavior, the destruction probability after 1,000 years would be less than 1/40,000 when tensile stresses occurring in the material were less than 79.4, 241.3, 8.0 MPa for alumina, PSZ, and titanium oxide, respectively. However titanium oxide could not be used because of the wall thickness that would be needed to accomplish this stress. The life prediction method includes only tht effect of preexisting flaws so the method to estimate the effect of localized corrosion is required for the future examination of the application of ceramic materials.

Synthesized glasses provide an opportunity to study natural corrosion processes which are intermediate in time span between geological examples of natural glasses, such as obsidians and tektites, and relatively short term laboratory tests lasting a few hours to several decades. Like natural glasses, the synthesized glasses are sufficiently abundant that laboratory studies can be conducted to study the processes, mechanisms and rates of corrosion. In addition, synthesized glasses can usually be tracked to particular archaeological find sites with known dates of production and often burial.