During the incubation period preceding pitting, current fluctuations indicate the beginning of the breakdown of the passive film. The charac- teristics of these current transients are being examined as a possible way to predict pitting. Stochastic models applied to the breakdown process have been proposed, and in order to test how well they account for the experi- mental results, various forms of processing of the current vs. time records are necessary. This paper describes the experimental data-taking methods, the processing routines so far developed for the statistical analysis of the data, and compares the experimental results with computer simulations based on a stochastic model.

The Canadian Nuclear Fuel Waste Management program is evaluating two basic supported-shell container concepts for used-fuel immobilization.

One concept uses a compacted particulate material within the container to sunport a thin, corroston-resistant shell. The other concept, called “the metal-matrix concept”, uses a cast-in-place metal matrix for shell support.

This paper describes corrosion studies on the candidate matrix materials, lead, zinc and aluminum-7 wt.7 Si (A1-7 wt.7 Si), in the temperature range 293 to 423 K and the dissolved oxygen content range 8 to 410 mg/L. Test specimens were analyzed for general and localized corrosion and the corrosion products were identified.

The results showed that, unlike lead, zinc and Al-7 wt.% Si are susceptible to localized corrosion. Lead showed only low general corrosion rates of 2 to 83 pm/a compared with zinc (36 to 15800 μm/a) and Al-7 wt.% Si (6 to 209 μm/a) in the above temperature and oxygen content ranges. Maximum pit depths of about 0.9 mm on Al-7 wt.% Si and 0.6 mm on zinc were observed. These results indicate that, of the candidate matrix metals studied, lead has the greatest potential to retard radionuclide release from used-fuel disposal containers.

Theory of a mechanistic model for estimating the maximum rate of uniform corrosion for a steel or copper container in a saturated repository environment is described. Four oxidants that are considered are oxygen, sulfate, radiolysis products, and water. Rates of corrosion from oxygen and sulfate reduction are assumed limited by the rate of inward diffusion of the oxidant through the rock, packing, and corrosion product layers surrounding the container. Reduction of water is limited by kinetic processes as influenced by mass transfer of products and reactants, temperature, and pH. Substantial changes in the system resulting from the corrosion process are predicted to occur over time.

Under repository conditions, it is likely that radioactive waste canisters will be subjected to both uniform and localised corrosion. Localised corrosion can take many forms depending on the precise physical and chemical environment of the metal at the time of attack, but generally the rates of penetration are much greater than those associated with uniform corrosion. The most likely forms under aerated repository conditions are pitting and crevice corrosion. To ensure adequate radionuclide containment an allowance for these rates must be included in the final canister dimensions. It is considered the best way to predict long term penetration rates is to develop mathematical models, which include all the physical and chemical processes necessary to describe the system and use experimentally determined input parameters relating to these processes. These models must then draw on further experimental data for validation over short timescales. In this paper, we discuss several techniques of modelling long term pit propagation in waste canisters. The complexity of the problem has lead to a number of physical and chemical approximations in the modelling. We investigate the applicability and ranges of validity of several of the more common approximations, both in our own models and in the literature, and compare the predictions with experimental pit growth rates. An investigatation of the sensitivity of the models to the various empirical input parameters indicates which need to be determined most accurately.

Three candidate waste package container materials were tested for 5 mo at 200 °C in Hanford Site Grande Ronde Basalt groundwater (9.75 pH) under anoxic conditions (<0.1 mg/L dissolved oxygen in water). The materials were cast carbon steel (American Society for Testing and Materials (ASTM) A27), wrought carbon steel (American Iron and Steel Institute (AISI) 1020), cupro- nickel 90-10, and Fe9CrlMo steel. Testing was performed in 1-L titanium autoclaves at a pressure of 6.9 MPa and a flow rate of ∼O.02 mL/min. Anoxic conditions were achieved by constantly sparging the synthetic groundwater in the reservoir with argon. In addition, the groundwater was conditioned by placing a 2.54-cm layer of crushed basalt (∼O.635 cm average size) at the bottom of each autoclave. The average corrosion rates at 200 °C were 0.9 μm/yr for cupronickel 90-10 and Fe9CrlMo steel, 1.1 μm/yr for cast carbon steel, and 1.4 μm/yr for wrought carbon steel. Pitting was not detected in any of the specimens. Posttest analysis of the corrosion speci- mens indicated the formation of a thin film of iron smectite clay on the surface of all specimens. In addition, magnetite formation immediately adjacent to the specimen surface was observed in the iron-base alloys, consistent with previous investigations. Based on the present investiga- tions, all the materials exhibited significantly lower corrosion rates than were used in the Environmental Assessment (∼5 μm/yr) [1] for calculating acceptable container lifetimes. However, longer term general corrosion tests and tests to study the container materials susceptibility to other degradation modes are necessary prior to making a final evaluation of the suitability of these materials for use in fabricating high-level nuclear waste containers.

The relationship of pore solution composition to observed leach behavior for cesium-doped, cement-based waste form has been demonstrated.

A leaching model is proposed, based on early diffusion and homogenization of the pore solution and leachant followed by re-equilibration of the waste form with the greater pore solution (pore solution plus leachant) over the long term.

Extensive employment of concrete is foreseen in radioactive waste repositories. A prerequisite for modelling the interactions between concrete and formation waters is characterisation of the concrete system. Available experimental data from high pressure squeezing of cement pore-water indicate that, besides the high pH due to alkali hydroxide dissolution, cement composition itself influences the solubility determining solid phases. A model which simulates the hydration of Portland cement assuming complete hydration of the main clinker minerals is presented. The model also includes parameters describing the reactions between the cement and blending agents.

Comparsion with measured pore-water data generally gives a consistent picture and, as expected, the model gives correct predictions for pure Portland cements. For blended cements, the required additional parameters can, to some extent, be derived from pore-water analysis.

The suitability of cement matrices for the long-term immobilisation of radionuclides cannot be predicted from short-term experimental data alone, because the chemical properties of cementitious systems change continuously over the repository lifetime. To model such changes is complex, so a stepwise approach has been adopted. A chemically simplified model for the solubility and compositional properties of calcium silicate hydrate gels in the system CaO-SiO2 -H20 was previously developed but has now been extended and improved and is applicable to gels in the wider composition range 0.8 < Ca/Si < 1.7. The effects of silicate speciation in aqueous solution on the formation of the solid phase have been more fully considered and the dissolution equilibrium has been revised. Solubility products and free energies of C-S-H formation have been evaluated and predictive applications of the model are discussed.

The UK Department of the Environment (DOE) is developing an independent capability to assess proposals made by the nuclear industry for the underground disposal of low and intermediate level radioactive waste. Modelling chemical processes forms an integral part of the underlying DOE research programme which includes the development of probabilistic risk assessment codes, more detailed deterministic models and supporting field and laboratory studies, in a co-ordinated, hierarchical approach. This paper summarises the results of a survey on the current status of this and related research, with emphasis on the assessment of a safety case for shallow land disposal of low level waste in about two years time. Shortcomings in existing work are identified, and a structured programme for future research described.

Sorption experiments with 85Sr, 137CS, 239pu and 241Am on four types of concrete (normal density, normal density with fly ash, high density and high density with silica fume) were carried out in a predominantly CaCl2 groundwater, in a predominantly NaCl groundwater and in nitrate solutions. Subsequently, autoradiography was used to identify the sorption pattern on the concrete surfaces. In order to distinguish the effects of chemical sorption from that of variable surface area, the effective surface area was used to derive reactive- surface-related distribution coefficients instead of the geometric- surface-related distribution coefficients.

The purpose of this work was to estimate the near-field radioactivity release rates associated with the high-activity, low-level radioactive wastes currently emplaced in Solid Waste Storage Area 6 (SWSA-6) at Oak Ridge National Laboratory (ORNL). Examination of the existing records [1] showed that about 84% of the radioactivity emplaced in SWSA-6 could apparently be accounted for by neutron-activated metal parts discharged from the High Flux Isotope Reactor (HFIR); therefore, this analysis was limited to consideration of those HFIR wastes. The radio- nuclide inventory contained in the HFIR wastes was first calculated, and then realistic or conservative assumptions as well as existing data were employed to calculate near-field radionuclide release rates. The dominant uncertainties in the calculations were summarized, and the research which would be needed to reduce these uncertainties was described. A more complete description of this work can be found in reference [2].

In Finland, utilities producing nuclear energy are responsible for the management of the radioactive waste, including final disposal. As regards low and intermediate level waste, the approach has been adopted to employ the power plant sites for locations of repositories. The repositories will be excavated at the depth of about 50 to 125 m in the bedrock of the two Finnish nuclear power plant sites, Loviisa and Olkiluoto.

The performance analysis presented in this paper has been carried out for the Preliminary Safety Analysis Report (PSAR) of the Olkiluoto reposi- tory. A flexible model has been developed to estimate the release of radio- nuclides from waste packages and their subsequent transport through the engi- neered barriers in the repository. Gradual degradation of the engineered barriers is accounted for by altering parameters at fixed time points. Safety margins of the disposal concept have been evaluated by including disturbed evolution scenarios in the analysis.

A new solidification technique using cement-glass, which is a mixture of sodium silicate, cement, additives, and initiator of the solidification reaction, was developed for sodium borate liquid waste generated from pressurized water reactor (PWR) plants. The cement-glass could solidify eight times as much sodium borate as cement could, because the solidifying reaction of the cement-glass is not hindered by borate ions.

The reaction mechanism of sodium silicate and phosphoric silicate (initiator), the main components of cement-glass, was studied through X-ray diffraction and compressive strength measurements. It was found that three- dimensionally bonded silicon dioxide was produced by polymerization of the two silicates. The leaching ratio of cesium from the cement-glass package was one-tenth that of the cement one. This low value was attributed to a high cesium adsorption ability of the cement-glass and it could be theoretically predicted accordingly.

An estimate of the rate of biodegradation of bituminous material is necessary to predict the long-term stability of low- and intermediate- level waste solidified using bitumen. Data from a series of scoping experiments have been analyzed to determine the rate of degradation of blown bitumen samples under a variety of conditions. Among the variables investigated were the effect of soil type, moisture, sample surface area and microbial strain. The rate of degradation was measured by monitoring metabolic CO2release. Using this data it was found that, for degradation in soil, a mean rate of 5.5 × 10−4 cm/yr represented all data to within a factor of about two. This mean rate is nearly that for distilled bitumen samples measured by other workers.

Experiments were performed at 200 °C and 8.5 MPa on used fuel and a borosilicate waste glass in contact with clay, container metals, groundwater and granite selected as reference materials in the Canadian Nuclear Fuel Waste Management Program. The experimental design considered five parameters that determine a fractional factorial design 25-1. The parameters varied were: waste form (used fuel or Moly 99“W” glass), cation exchange capacity of the clay (calcium bentonite, 1360 meq/kg or illite, 160 meq/kg), ionic strength of the groundwater (Standard Canadian Shield Saline Solution, SCSSS, 1.4 mol/L or granitic groundwater GGW, 1 × 10−3 mol/L), surface area to volume ratio (SA/V) (120 or 12 m−1), and container material 3Inconel 625 or Ticode 12). The results show that (i) maximum release of 137Cs from used fuel and waste glass occurs in the presence of illite and GGW, (ii) maximum releases of 90Sr and the actinides from used fuel occur in the presence of bentonite and SCSSS, and (iii) maximum release of 90Sr from the waste glass occurs in the presence of illite and GGW.

A tracer-injection system has been developed for use in characterizing reactions of waste package materials under hydrothermal conditions. High-pressure liquid chromatographic instrumentation has been coupled with Dickson-type rocking autoclaves to allow injection of selected components into the hydrothermal fluid while maintaining run temperature and pressure. Hydrothermal experiments conducted using this system included the interactions of depleted uranium oxide and Zircaloy-4 metal alloy discs with trace levels of 99Tc and non-radioactive Cs and I in a simulated groundwater matrix. After waste-package components and simulated waste forms were pre-conditioned in the autoclave systems (usually 4 to 6 weeks), known quantities of tracer-doped fluids were injected into the autoclaves' gold reaction bag at run conditions. Time-sequenced sampling of the hydrothermal fluid provided kinetic data on the reactions of tracers with waste package materials. The injection system facilitates the design of experiments that will better define “steady-state” fluid compositions in hydrothermal reactions. The injection system will also allow for the formation of tracer-bearing solid phases in detectable quantities.

Integrated testing of the important components of a glass waste form waste package has been performed in order to gain a better understanding of the processes of radionuclide release and transport in the near field environment. Based upon an interpretation of the depth of penetration of hydrogen in reacted SRL-165 glass we have modeled the radionuclide release from the glass as a combined process of (1) the diffusive exchange of alkalis and boron in the glass for hydrogen species in the solution (D=10−16 cm2/s) and (2) surface dissolution. Surface dissolution controls the release of components not exchanged by diffusion and takes place at a rate of 1.5-3.0 μm/yr. Subsequent to release the radionuclides may remain in the leach solution, diffuse into the tuff, or precipitate as secondary phases. Precipitation is particularly important for plutonium and americium. Diffusive transport of radionuclides through the tuff takes place at an extremely slow rate, D=10−16 cm2/s. As such, the mass of radionuclides incorporated in the tuff by diffusion during the tests is inconsequential relative to that in the leach solution (with the exception of plutonium) and can be ignored in mass balance calculations. Mass balance calculations based upon the release of radionuclides by surface dissolution of the glass waste form are in good agreement with observed solution chemistry when allowances are made for a pulse of dissolution early in the tests. This pulse may be due to either the rapid dissolution of high-energy surface features early in the inLegrated tests, or an initially high surface dissolution rate that decreases with time as silica saturation is approached [1], or a combination of the two.

A laboratory leaching test has been performed as part of a project to evaluate the suitability of tuff rocks at Yucca Mountain, Nevada, as a site for a high-level nuclear waste repository. Glass samples were placed in water inside tuff vessels, and then the tuff vessels were placed in water inside Teflon containers. Glass-component leach rates and migration through the tuff were measured for samples of the ATM- 9actinide glass, which is a PNL 76-68 based glass doped with low levels of 99Tc, 237Np, 238U and 239Pu to simulate wastes. Disc samples of this glass were leached at 90°C for 30, 90, and 183 days inside tuff vessels using a natural groundwater (J-13 well-water) as the leachant. Some samples were held by 304L stainless steel supports to evaluate the effect of this metal on the release rate of glass constituents. At the end of each leaching interval, the J-13 water present inside and outside the rock vessel was analyzed for glass components in solution.

On the basis of these analyses, boron, molybdenum, and technetium appear to migrate through the rock at rates that depend on the porosity of each vessel and the time of reaction. The actinide elements (uranium, neptunium, and plutonium) were found only in the inner leachate. Sodium, silicon, and strontium are present in the rock as well as in the J-13 water, and the addition of these elements from the glass could not be determined. Normalized elemental mass loss values for boron, molybdenum, and technetium were calculated using the combined concentrations of the inner and outer leachates and assuming a negligible retention on the rock. The maximum normalized release was 2.3 g/m for technetium. Boron, molybdenum, technetium, and neptunium were released linearly with respect to each other, with boron and molybdenum released at about 85% of the technetium rate, and neptunium at 5-10% of the technetium rate. Plutonium was found at low levels in the inner leachate but was strongly sorbed on the steel and Teflon supports. Neptunium was sorbed to a lesser extent. Future analysis of the tuff vessels will determine whether the actinides were strongly sorbed on the surface of the tuff rock.

Since 1982, the corrosion behaviour of radioactive high level waste glass and of inactive glasses of the same composition has been studied in the framework of a Japanese-Swedish-Swiss (JSS) project. In addition to granite and backfilling material, the influence of steel corrosion products on glass corrosion has been investigated in experiments of up to one year at 90°C with a sample surface area to leachant volume ratio of 10 m−1. Leachant and steel corrosion products were separated from each other by filtration and analyzed by atomic absorption spectroscopy and γ-spectroscopy. The glass corrosion layers were characterized by optical and scanning electron microscopy and by secondary ion mass spectrometry.

It could be shown, that glass corrosion is enhanced by steel corrosion products. The quantity of corroded glass seems to be dependent on the surface area and not on the total amount of steel corrosion products in the system, suggesting a mechanism controlled by silica sorption on the steel corrosion product. If lead oxide is added to the system, glass corrosion is almost to- tally suppressed.

Experiments [1], performed in the frame of the JSS-project (joint Japanese [CRIEPI], Swiss [NAGRA], Swedish [SKB] project), demonstrated that the effect of iron corrosion products on the long-term stability of HLW-glass depends on the effective surface area of the corrosion products. The experimental results were described quantitatively by thermodynamic and kinetic modeling with the computer codes GLASSOL and PHREEQE. The model interprets the effect in terms of surface sorption of a monolayer of silica molecules from solution. Silica saturation occurs in solution only, when all surface sites are occupied. The effect of bentonite is also considered. The agreement between calculations and experiment is about as good as in the case of modeling the glass-water reaction only.