MCC-1 static leaching experiments were carried out for a simulated waste glass in deionized water for up to 364 days at 90°C in order to examine the relationship between alteration layer thickness and elemental mass losses. The thickness of the layer increased linearly at a rate of 0.63,μm/day for the first 91 days, and thereafter it remained constant (about 60 μm) independent of time. The change in normalized elemental mass losses of B, Na, and Si with time shows the same trend as that of the change in the layer thickness. It is also demonstrated that the growth rate of the layer can be calculated from the released boron in the leachates. The shrinkage of the alteration layer during air-drying is not significant in spite of the porous structure of the layer, which results from the depletion of cations and the recrystallization reported previously. Although the apparent release of elements almost ceases after a 91-day leaching experiment, reactions in the layer such as crystal growth continue.

We report data on the relationship between nuclear waste glass durability, as measured by leachate concentrations and leach rates, and the ratio of glass surface area to solution volume, S/V. The study includes West Valley non-radioactive production glasses (SF6 and SF10), West Valley glasses containing U and Th designed at CUA (WVCM47, WVCM50, and WVCM59), Savannah River SRLTDS-131 glass, and Hanford waste glass HW-39 (for which existing literature data is used). While some of these glasses show departures of leachate concentrations from simple (S/V)t scaling others conform well. The departures are, in all cases, most evident at high values of S/V. It is therefore important, therefore to understand how glass composition determines which corrosion mechanism is dominant in order to assess the region of validity of extrapolations based on (S/V)t scaling. While leach rates show a general tendency to decrease with S/V and time, exceptions are evident for the less durable glasses which show minima and maxima in the S/V-dependence at fixed time.

Samples of West Valley reference glass WVCM-59 were subjected to isothermal heat treatment according to a systematic schedule of time-temperature combinations. Phases crystallizing during heat treatment were analyzed and quantified using an SEM equipped with an energy dispersive x-ray detector and image processing and analyzing capabilities. Only small amounts of noble metals (e.g. Rh, Pd) and RuO2 were present in the as-melted glass, but these frequently serve as nucleation sites during heat treatment. Iron-group spinels containing Fe, Ni, Cr, and Mn in variable proportions were the most common phases observed, appearing in quantities up to 3.5 vol% in heat treated glass. The formation of a thoria-ceria phase occurred at temperatures of 900°C and below. Acmites formed at temperatures of 800°C and below. A lithium phosphate phase with a particularly interesting morphology developed after prolonged heat treatment at temperatures between 600–750°C. The same phase appeared in a range of glass compositions. However, canister cooling curves preclude such a phase in production glass. Furthermore, no such phase was found in actual canister-cooled glass.

Fully radioactive waste glasses produced in a hot vitrification pilot plant were characterized in the hot cells at the JRC-Karlsruhe. The work includes microstructural analysis before and after leaching by SOXHLET extractors (MCC-5S), as well as leachate and gel-layer chemical analysis. Partial devitrification, with phases enriched in metals and rare earths, was observed in one of the products. In the gel layers formed during the dissolution tests, enrichment in some elements (Fe,Ti,Gd,Nd,Ce) and depletion in glass constituents (Si,B,K,Ba,Ca,Na) as well as in Mo and Cs, were determined. This was in good qualitative agreement with the leachate chemical analysis.

In vitrification of nuclear waste, technetium tends to be lost substantially by evaporation [1,2,3,4,5]. If Cs is not present, 63% of the Tc volatilized from the molten borolilicate glass at 1420K under oxidizing conditions and about 40% under reducing conditions achieved by adding 0.5 wt% graphite to the dried frit mixture[5]. This behaviour could be correlated to thermochemistry in the simple Tc-O system considering the predominance areas of the different oxides and the partial pressures of Tc2O7 above them[6]. If Cs is present(0.544 g Cs/l together with 1.61 g Na/l in the synthetic HAW solution), Tc was found to start a genuine volatilization at about 800K which proceeded rapidly in the range 875 to 950K[2]. At 1300K between 60 and 80% of the Tc and 10% of the Cs had volatilized. Because of the ratio Tc:Cs≉7:l volatility by formation of gaseous CsTcO4 was excluded[2]. On the other hand, for the same glass and HAW solution as reported in Ref.[2], Tc was found to produce a synergistic effect on the volatility of Cs[3,4]: between 2 and 5% Cs volatilized without Tc but about 22% in the presence of Tc and temperatures up to 1400K. The amount of volatilized Tc was about 50%. Furthermore[3,4], Tc was identified together with Cs in crystals deposited in the aerosolfilters after continuous heating to about 1170K. Since the aerosol crystals showed a constant ratio of Tc:Cs =1:1 and since the valency of Tc in the crystals was found to be seven the formation and the condensation of gaseous CsTcO4 was assumed [3,4]. Gaseous CsTcO4 or other gaseous pertechnetates of the alkaline metals have not yet been identified experimentally. Its existence may be concluded, however, from the existence of the gaseous perrhenates which form with every alkaline metal. Solid CsTcO4 and other solid alkalipertechnetates are well known [7]. By using the thermochemical data of the alkaliperrhenates, free energies of condensed as well as of gaseous CsTcO4 are estimated. Together with the results of the thermochemical treatment of the Tc-O system[6] and experimental data of the Cs-O system[8] they are used to calculate tentative isothermal predominance area diagrams of the ternary Cs-Tc-O system. Isobars of the gases Cs2O,CsO,Tc2O7, and CsTcO4 above the different condensed phases are also cal culated and used to understand the experimental results[2,3,4] on the simultaneous evaporation of Cs and Tc during vitrification.

The oxidation-reduction equilibria of selected multivalent elements in an alkali borosilicate glass melt (Savannah River Laboratory frit #131) were measured as a function of the imposed oxygen fugacity over the temperature range from 950°C to 1350°C. Redox constraints on the processing of high-level nuclear waste into the glass melt require that the prevailing oxygen fugacity be about 10−5 to 10−12 Zatm at 950°C, about 10−2 to 10−9 atm at 1150°C, and about 100 to 10−7 atm at 1350°C. Such conditions circumvent foaming under oxidizing situations and metal/sulfide precipitation if the system becomes too reducing. The defined oxygen fugacity ranges correspond to the previously prescribed range of 0.1 to 0.5 for the [Fe2+]/[Fe3+] ratio in the resulting glass, independent of the processing temperature from 950°C to 1350°C.

Viscosity is one of the most important processing parameters and one of the most difficult to calculate theoretically, particularly for multicomponent systems like nuclear waste glasses. Here, we propose a semi-empirical approach based on the Fulcher equation, involving identification of key variables, for which coefficients are then determined by regression analysis. Results are presented for two glass systems, and compared to results of previous workers and to experiment. We also sketch a first-order statistical mechanical perturbation theory calculation for the effects on viscosity of a change in composition of the melt.

Waste forms have been developed and characterized at PNC (Power Reactor and Nuclear Fuel Development Corporation)to immobilize high-level liquid waste generated from the reprocessing of nuclear spent fuel.

Mechanical strength tests were excecuted on simulated solidified highlevel waste forms which were borosilicate glass and diopside glass-ceramic. Commercial glass was tested for comparison. Measured strengths were three-point bending strength,uniaxial compressive strength,impact strength by falling weight method,and Vickers hardness. Fracture toughness and fracture surface energy were also measured by both notch-beam and indentation technique.

The results show that mechanical strengths of waste glass form are similar and that the glass ceramic form has the higher fracture toughness.

Diffusion of cesium in borosilicate glass for high-level radioactive waste is discussed. For this purpose model glasses with non-radioactive elements are being made, in accordance with the specifications of the reprocessing plants, from which concentration couples are composed. A concentration couple consists of two cylinders of borosilicate glass which contain different amounts of cesium. After heat treatment the couples are studied by means of scanning electron microscopy and X-Ray microanalysis. The model study will provide a basis for predictions of the containment achieved over a longer period of time.

The uranium and plutonium chemistry relevant for spent fuel leach studies is discussed. Model calculations are evaluated by comparison with the results of spent UO2 fuel leaching obtained in the Swedish SKB program. The thermodynamic data were found to agree sufficiently with the measured solution concentrations that it can be assumed that in oxic natural waters, leachate uranium from spent fuel would be uranyl carbonate complexes in the solution (or hydroxo complexes in the absence of carbonate) while schoepite would be the solubility limiting solid. For plutonium, PuO2+ would be the solution species and Pu(OH)4 the solubility limiting solid.

The effects of radicals, formed during radiolysis of water, on the oxidation and dissolution of U02 were investigated as a function of dose rate and total dose. The U02 oxidation rate during radiolysis was monitored by recording the corrosion potential of a U02 electrode as a function of time. Changes in the surface of U02 were determined using cathodic stripping voltammetry (CSV) and X-ray photoelectron spectroscopy (XPS). Our studies suggest that the oxidation of U02, in irradiated de-oxygenated solutions consists of two stages. The first stage consists of the growth of a surface film of composition close to U02.33 and of thickness similar to that obtained (over longer exposure periods) in unirradiated oxygenated solutions. The rate of growth of this film appears to be proportional to the square root of the dose rate. The second stage consists of oxidative dissolution of this film (as U02+2). This step mainly occurs at higher doses. Relative rates of oxidation of U02 to U02.33, the first stage, were obtained in four different solutions favouring the formation of selected radicals.

Comparison of spent fuel corrosion data from nuclear waste management projects in Canada, Sweden and the USA strongly suggests that the release of 90Sr to the leachant can be used as a measure of the degradation (oxidation/dissolution) of the fuel matrix. A surprisingly quantitative similarity in the 90 Sr release data for fuel of various types (BWR, PWR, Candu), linear power ratings and burnups leached under oxic conditions was observed in the comparison. After 1000 days of leachant contact, static or sequential, the fractional release rates for 90Sr (and for cesium nuclides) were of the order of 10−7/d.

The rate of spent fuel degradation (alteration) under oxic conditions can be considered to be controlled either by the growth rates of secondary alteration products, by oxygen diffusion through a product layer, by the rate of formation of radiolytic oxidants or by solubility-controlled dissolution of the matrix. These processes are discussed. Methods for determining upper limits for long-term 90Sr release, and hence fuel degradation, have been derived from the experimental data and consideration of radiolytic oxidant production.

The Yucca Mountain Project of the U.S. Department of Energy is investigating the suitability of a site in the unsaturated zone at Yucca Mountain, NV, for a high-level nuclear waste repository. Most of the waste will consist of UO2 spent fuel in Zircaloy-clad rods from nuclear reactors. If failure of both the waste containers and the cladding occurs within the lifetime of the repository, then the UO2 will be exposed to oxygen in the air and higher oxides of uranium may form. The oxidation state of the spent fuel may affect its dissolution behavior if later contacted by water. A model for the kinetics of spent fuel oxidation under repository-relevant conditions is thus necessary to predict the behavior of the waste form for assessing the performance of the repository with respect to the containment of radionuclides. In spent fuel experiments, the UO2 oxidation front initially propagates along grain boundaries followed by propagation into grain volumes. Thus, the oxidation kinetics is controlled by two processes and the oxidation of spent fuel fragments will depend on the density and physical attributes of grain boundaries. With this in mind, concepts from statistical mechanics are used to define a density function for grain boundaries perunit volume per unit species in a spent fuel fragment. Combining the integral forms of mass conservation and this grain boundary density function, a model for the global rate of oxidation for a spent fuel fragment is obtained. For rapid grain boundary oxidation compared to grain volume oxidation, equations of the model are solved and results compared to existing data.

Gaining a better understanding of the potential release behavior of water-soluble radionuclides is the focus of new laboratory spent fuel dissolution studies being planned in support of the Yucca Mountain Project. Previous studies have suggested that maximum release rates for actinide nuclides, which account for most of the long-term radioactivity in spent fuel, should be solubility-limited and should not depend on the characteristics or durability of the spent fuel waste form. Maximum actinide concentrations should be sufficiently low to meet the NRC annual release limits. Potential release rates for soluble nuclides such as 99Tc, 135Cs, 14C and 129I, which account for about 1-2% of the activity in spent fuel at 1000 years, are less certain and may depend on processes such as oxidation of the fuel in the repository air environment.

Dissolution rates for several soluble nuclides have been measured from spent fuel specimens using static and semi-static methods. However, such tests do not provide a direct measurement of fuel matrix dissolution rates that may ultimately control soluble-nuclide release rates. Flow-through tests are being developed as a potential supplemental method for determining the matrix component of soluble-nuclide dissolution. Advantages and disadvantages of both semi-static and flow-through methods are discussed. Tests with fuel specimens representing a range of potential fuel states that may occur in the repository, including oxidized fuel, are proposed. Preliminary results from flow-through tests with unirradiated UO2 suggesting that matrix dissolution rates are very sensitive to water composition are also presented.

A set of experiments, wherein UO2 has been contacted by dripping water, has been conducted over a period of 182.5 weeks. The experiments are being conducted to develop procedures to study spent fuel reaction under unsaturated conditions that are expected to exist over the lifetime of the proposed Yucca Mountain repository site. One half of the experiments have been terminated, while one half are ongoing. Analyses of solutions that have dripped from the reacted UO2 have been performed for all experiments, while reacted UO2 surfaces have been examined for the terminated experiments. A pulse of uranium release from the UO2 solid, combined with the formation of schoepite on the surface of the UO2, was observed between 39 and 96 weeks of reaction. Thereafter, the uranium release decreased and a second set of secondary phases was observed. The latter phases incorporated cations from the EJ-13 water and include boltwoodite, uranophane, sklodowskite, compreignacite, and schoepite. The experiments are continuing to monitor whether additional changes in solution chemistry or secondary phase formation occurs.

A mathematical and numerical model for evaluation of crevice and pitting corrosion in radioactive waste containers is presented. The model considers mass transport, mass transfer at the metal/solution interface, and chemical speciation in the corrosion cavity. The model is compared against experimental data obtained in artificial crevices. Excellent agreement is found between modeled and experimental values for the potential. The importance of full consideration of complex ion formation in the aqueous solution is emphasized and illustrated.

Previously reported work by others indicates that dicopper trihydroxide nitrate, Cu2NO3(OH)3, forms on copper and copper alloys subjected to irradiated moist air near room temperature. We have performed experiments over a range of temperature and humidity, and have found that this species is formed at temperatures up to at least 150° C if low to intermediate relative humidities are present. At 150°C and 100% relative humidity, only Cu2O and CuO were observed. The relative general corrosion rates of the copper materials tested in 1-month experiments at dose rates of 0.7 and 2.0 kGy/h were Cu > 70/30 Cu-Ni > Al-bronze. High-nickel alloy 825 showed no observable corrosion.

This paper describes a study of corrosion behavior of a mild steel as a candidate of the high-level nuclear waste package in the geological disposal situations, conducted to establish a model for estimating the corrosion allowance requirement to achieve the 1,000 year lifetime for the package. In several series of galvanostatic tests, the maximum penetration depth and the depth distribution were measured for each specimen with a sophisticated ultrasonic inspection technique. The Gumbel distribution model was successfully used in analyzing each set of data for the maximum penetration depth. Relations among the average penetration depth, the maximum penetration depth, and the corrosion allowance requirement were discussed.

Cortest Columbus is investigating the long-term performance of container materials used for high-level waste packages as part of the information needed by the Nuclear Regulatory Commission to assess the Department of Energy's application to construct a geologic repository for high-level radioactive waste. In one task of the program, a cyclic potentiodynamic polarization (CPP) technique was used to evaluate the corrosion behavior of the candidate container materials. In order to evaluate the expected range of environmental variables, a statistical experimental design approach was used. A Resolution IV experimental design for 15 variables was selected. The variables included temperature, pH, species present in the groundwater, and those generated by radiolysis. Complete matrices of CPP tests, which consist of tests in 33 environments, were performed on two candidate container materials; Type 304L Stainless Steel and Incoloy Alloy 825. In these tests, both alloys exhibited a wide range of behavior; including passive behavior, pitting and active corrosion. For each alloy, the environmental variables that affected corrosion behavior were identified.

The conceptual design and fabrication test of a full-scale ceramic overpack were performed from the viewpoint of structural barriers as a part of program to evaluate their potential use as overpack under conditions of deep geological disposal.

Materials investigated were porcelain (used for insulators) and A12O3 with high purity of 99.7 %. The selected design consisted of a cylindrical body with hemispherical heads at each end. The design thickness of overpack is the sum of the structural thickness and corrosion allowance. The thickness required to resist the lithostatic pressure was estimated by the basic cylinder buckling formulas and finite element stress analyses in both case of uniform and non-unifom external pressure conditions. These analyses showed that structural thickness of 119 mm was necessary for overpack of porcelain and 40 - for A12O3 under the predicted maximum uniform pressure. In addition, fracture probability of delayed failure, one of significant degradation mode, was estimated for overpack of porcelain.

A full-scale overpack of porcelain, of dimensions 800 mm outer diameter x 2200 mm length x 150mm wall thickness, was fabricated under the ordinary level of fabrication technology.