An update of the status of the joint SKB/Studsvik fuel leaching programme is given. High burnup BWR fuel has been leached for up to 5.5 years cumulated contact time. Data for leaching of PWR fuel are also presented. The present results confirm earlier results from the programme, i.e., in groundwater and deionized water under oxidizing conditions uranium rapidly attains saturation at about 1 mg/l and 10 μg/l, respectively.

Under reducing conditions, the uranium saturation levels are below 1μg/l. The plutonium saturation level under oxidizing conditions is at about 0.3 μg/l. Under reducing conditions, the plutonium concentrations in solution appear to be controlled by the solubility of the fuel matrix.

Most of the detected fission products appear to be released through selective leaching of fuel/cladding gap, fuel cracks and fissures, and grain boundaries.

The Basalt Waste Isolation Project has conducted a series of hydrothermal experiments to characterize waste/barrier/rock interactions as a part of its study of the Columbia River basalts as a potential medium for a nuclear waste repository. Hydrothermal tests of 3–15 months duration were performed with light water reactor spent fuel and simulated groundwater, in combination with candidate container materials (low-carbon steel or copper) and/or basalt, in order to evaluate the effect of waste package materials on spent fuel radionuclide release behavior. Solutions were filtered through 400 and 1.8 nm filters to distinguish colloidal from dissolved species. In all experiments, 14C, 129I, and 137Cs occurred only as dissolved species, whereas the actinides occurred in 400 nm filtrates primarily as spent fuel particles. Actinide concentrations in 1.8 nm filtrates were below detection in steel-bearing experiments. In the system spent fuel + copper, apparent time-invariant concentrations of 14C and 137Cs were obtained, but in the spent fuel + steel system, the concentrations of 14C and 137Cs increased gradually throughout the experiments. In experiments containing basalt or steel + basalt, 137Cs concentrations decreased with time. In tests with copper + basalt, 14C and 129I concentrations attained time-invariant values and 137Cs concentrations decreased. Concentrations for the actinides and fission products measured in these experiments were below those calculated from Federal regulations governing radionuclide release.

Samples of uranium ore from the 1,300 million year old Cigar Lake deposits in northern Saskatchewan have been investigated using X ray Photoelectron Spectroscopy (XPS) and X ray Diffraction (XRD). The XPS spectra showed the peaks for U, Pb, 0, C, Si, Ca and Mg. Analysis of highresolution XPS spectra showed that the oxidation state of uranium in these minerals is equal to, or lower than, that of U3O7. The XRD data confirm the presence of an oxidized phase, most likely α U3O7., in addition to UO2. The low U(VI)/U(IV) ratio in these minerals is consistent with the low uranium concentration found in groundwater in contact with these uranium minerals. The relevance of these results to the disposal concept of used nuclear fuel in the Canadian Nuclear Fuel Waste Management Program is discussed.

The Nevada Nuclear Waste Storage Investigations (NNWSI) Project is studying dissolution and radionuclide release behavior of spent nuclear fuel in Nevada Test Site groundwater. Specimens prepared from pressurized water reactor (PWR) fuel rod segments were tested for multiple cycles in J-13 well water. The Series 2 tests were run in unsealed silica vessels under ambient hot cell air (25°C) for five cycles for a total of 34 months. The Series 3 tests were run in sealed stainless steel vessels at 25°C and 85°C for three cycles for a total of 15 months. Selected summary results from Series 2 and Series 3 tests with bare fuel specimens are reported.

Actinide concentrations tended to saturate and then often decreased during test cycles. Uranium concentrations in later test cycles ranged from 1 to 2 μg/ml in the Series 2 Tests versus about 0.1 to 0.4 μg/ml in Series 3 with the lowest concentrations occurring in the 85°C tests. Formation of a calciumuranium-silicate phase identified as uranophane in the 85°C Series 3 Tests is thought to have limited uranium concentration in these tests. Americium-241, Pu-239 and Pu-240 activities measured in filtered solution decreased to less than 1 pCi/ml in the 85°C tests. Preferential release of fission products Cs, I, Sr and Tc, and activation product C-14, was indicated relative to the actinides. Tc-99 and Cs-137 activities measured in solution after Cycle 1 increased linearly with time, with the rate of increase greater at 85°C than at 25°C. Continuous preferential release of soluble fission products is thought to result primarily from the dissolution of fine particles of fission product phases concentrated on grain boundaries.

Geochemical simulations of the degradation of spent fuel waste form in the presence of groundwater at the candidate Yucca Mountain, Nevada repository have been carried out to attempt to predict elemental concentrations in solution and to identify potential radionuclide-bearing precipitates. Spent fuel was assumed to dissolve congruently into a static mass of J-13 groundwater at 25°C and 90°C. No inhibitions to the precipitation and dissolution of secondary phases were assumed to exist. The elements Ac, Zr, Nb, Pd, Sm, Mo, Sb and Cm were not considered in the simulations because of a lack of thermodynamic data.

Simulation results indicate that haiweeite, soddyite, Na2U2O7(c) and schoepite are potential U-bearing precipitates. Na2U2O7(c) is only predicted to occur at 90°C. U concentrations in solution and the identity of the U-bearing precipitate depend on the activity of SiO2(aq) in solution. U concentrations are limited to < 1 mg/kg when sufficient SiO2(aq) exists in solution to precipitate uranyl silicates. Depletion of SiO2(aq) in solution by the precipitation of silicates results in predicted increases of U concentrations to 87 and 619 mg/kg at 25°C and 90°C, respectively. Subsequent reaction and precipitation of schoepite cause U concentrations to decrease.

Radionuclides other than U commonly precipitate as oxides in the simulations. The precipitation of solid phases appears to be extremely effective in limiting the concentrations of some radionuclides, such as Pu and Th, in solution. Concentrations of other elements are held constant (Sn) or are alternately held constant and then increase (Am, Ni, Np) as various solid phases precipitate and pH decreases from 8.5 to 6.5 at 25°C and 8.7 to 8 at 90°C. No solid phases containing Cs or Tc are predicted to form. Increasing the temperature from 25°C to 90°C does not impact greatly the identity of precipitated phases or solution composition, except in the case of U.

A technique involving isotope dilution measurements may allow determination of the rates of spent fuel dissolution in future experiments.

Isochronal annealing (100°–1200°C) of alpha-decay damage was performed on UO2 and ThO2 and isothermal annealing (400°C and 750°C) of UO2 was done to determine the change in unit cell parameter (UO2: ao = 5.501Å to 5.458Å ± 0.001; ThO2: ao = 5.548Å to 5.523Å ± 0.001), strain, and crystallite size (data inconclusive). Two recovery stages (UO2: Stage I at ≈225°C, Stage II at ≈675°C; ThO2: Stage I at ≈250°C, Stage II at ≈550°C) were observed for each. Strain decreased with temperature for both, but never completely recovered in either (UO2 final value: ≈2.37 × 10−3; ThO2 final value: ≈2.06 × 10−3).

The Basalt Waste Isolation Project has performed initial characterization of the release behavior of radionuclides from ATM-101 spent fuel reacted under hydrothermal conditions. Results to date indicate that 1) 14C, 90Sr, 129I, and 137Cs exhibit preferential release to solution relative to the actinides, 2) 239+240Pu, 241Am, and 244Cm ratios in solution are consistent with stoichiometric dissolution of the spent fuel matrix, whereas uranium concentrations are higher than those predicted for stoichiometric dissolution, 3) 137Cs release varies as a function of the fluid to solid mass ratio, and 4) 14C, 129I, and 137Cs release is affected by fuel particle size and/or fuel washing.

Dissolution of nuclear waste glass occurs by corrosion mechanisms similar to those of metallurgical and mineralogic systems albeit on different time scales. The effects of imposed pH and oxidation potential (Eh) conditions existing in natural environments on metals and minerals have been quantatively and phenomenologically described in compendiums of Pourbaix (pH-potential) diagrams. Construction of Pourbaix diagrams to quantify the response of nuclear waste glasses to repository specific pH and Eh conditions is demonstrated. The expected long-term effects of groundwater contact on the durability of nuclear waste glasses can then be unified.

Performance assessment of high-level nuclear waste forms in a repository must cover periods of time on the order of hundreds of thousands of years. There has rarely been a need to extrapolate the properties of a material over such long periods of time (except in the geologic sciences, e.g., the evaluation of long-term weathering effects based on results from laboratory experiments) and there has never been a more important need to make accurate and -verifiable “predictions” of the long-term materials performance. The results of such extrapolations could affect decisions concerning the selection of a repository and the number and types of barriers that are required to insure isolation over long periods of time. In typical safety analyses [1,2] waste form corrosion is described by a single constant rate, and no effort is made to evluate whther the modal of the “source term” can be substantiated or verified by comparison similar or analogous processes [3].

A model based upon structural thermodynamic considerations has been proposed for both chemical durability and viscosity of nuclear waste glasses. This model assumes the bond strength between atoms in the glass to be the predominant factor in controlling the composition dependence of physical properties such as viscosity and chemical properties like durability. These characteristic bond strengths in the glass are obtained from known heats of formation of the constituent oxides with a few simple structural rules deduced by distinguishing the structural roles of the various oxides, i.e., separating those components that are clearly “network-formers” (e.g. SiO2, Al2O3, ZrO2) from those that are clearly “network-breakers” (e.g. alkali oxides). This model has been successfully applied to the correlation of glass composition with modified MCC-3 leach test results on nuclear waste glasses measured at various laboratories on one hand and with the viscosities of numerous glasses over a temperature range of 850°C to 1600°C with remarkable accuracy on the other hand.

The French LWR reference glass R7T7 was leached at temperatures between 100°C and 300°C at a pressure of 10 MPa.

An activation energy of 25 kJ/mole was found between 100°C and 250°C which is different from the 60 kJ/mole previously found between 25°C and 70°C. It is suggested that a potential glass hydration mechanism becomes predominant when dissolution is significantly slowed by high concentrations of silica in solution. At 300°C, a complete change in the corrosion mechanisms is found as the glass undergoes a rapid hydrothermal alteration

The release of Am-241 during corrosion of the radioactive waste glass, JSS-A, has been studied in the presence of corrosion products and/or uncom-pacted bentonite. The corrosion behaviour of Am-241 has been analyzed using gamma spectrometry. Adsorption of Am-241 on bentonite is observed in all cases. The contents of Am-241 in centrifuged leachates are in most cases less than 1/100 of total values. The normalized elemental mass loss of Am increases initially with corrosion time, and the values in the presence of bentonite and corrosion products are larger than those in the presence of bentonite alone. This tendency is in agreement with results previously found for other elements. The release of Am is low, only about 10–20 % of the corresponding total mass loss.

In the Federal Republic of Germany, the salt dome formation at Gorleben has been selected as a possible site for final disposal of HLW. The waste form will be borosilicate glass, e. g. the French Cogema glass R7T7. The corrosion of this glass has been investigated in three reference salt brines (brine 1: high Mg, brine 2: Ca and Mg, and brine 3: high Na) at three different temperatures, 110°, 150°, and 190°C and three S/V ratios, 10, 100, and 1000m−1. Results are presented for brine 1 and are compared qualitatively with those obtained for the other brines. As in deionized water and pure NaCl, the corrosion rate decreases when silica saturation is approached, but the reaction does not cease. The data after silica saturation were used to calculate upper limiting, i. e. time-independent corrosion rates. The long-term data indicate that a transport process through the growing surface layer may become rate controlling, and that the rate may further decrease after silica saturation.

Results of leach tests with Tc-99 doped SRP borosilicate waste glass are presented. The glass was prepared by melting a mixture of SRP 165 powdered frit doped with a carrier free solution of Tc-99 at 1150°C. Dissolution of portions of the resulting glass indicated that the Tc-99 was distributed homogeneously throughout the glass. Static leach tests up to 90 days were performed at 90°C in J-13 tuff groundwater or WIPP brine A at a SA/V of 100m−1. Normalized mass losses were calculated for Tc-99 as well as all the major elements in the glass. Results indicated that under ambient oxidizing conditions Tc-99 leached no faster than the glass-forming elements of the glass. In J-13 water, Tc-99 leached congruently with B. In WIPP brine A, it leached congruently with Si. Leach rates for Li were higher in both groundwaters, probably due to a contribution from an ion exchange mechanism. Leach tests were performed under reducing conditions in J-13 water by adding Zn/Hg amalgam to the leachate. In these tests the pH increased significantly, probably because of the reaction of the amalgam with the water. In a 21-day test, the pH Increased to 13 and leach rates for the glass were very high. Even though there was significant dissolutiop of the glass, the normalized mass loss based on Tc-99 was only 0.02g/m2. This result and the fact that reducing conditions at normal pH values do not significantly affect the dissolution of the glass, indicate that the low concentrations for Tc-99 obtained under reducing conditions are due to its solubility and not due to an increased durability of the glass.

A mechanistic model describing a dynamic mass balance between the production and consumption of dissolved silica was found to describe the dissolution behavior of SRL-165 defense waste glass in a high-magnesium brine (PBB3) at a temperature of 90°C. The synergistic effect of the waste package container on the glass dissolution rate was found to depend on a precipitation reaction for a ferrous silicate mineral.

The model predicted that the ferrous silicate precipitate should be variable in composition where the iron/silica stoichiometry depended on the metal/glass surface area ratio used in the experiment. This prediction was confirmed experimentally by the variable iron/silica ratios observed in filtered leachates. However, the interaction between dissolved silica and iron corrosion products needs to be much better understood before the model can be used with confidence in predicting radionuclide release rates for a salt repository.

Dissolution of West Valley and Defense Waste Product Facility (DWPF) glasses in J-13 water at 90°C at the candidate Yucca Mountain, Nevada repository was simulated using the EQ316 computer code package. The objectives of the study were to attempt to predict the concentrations of radionuclides and other glass components in solution resulting from glass dissolution, and to identify potential precipitates that sequester glass components.

Modified projected inventories of 10,000 year-old Nest Valley and DNPF SRL-165 frit glasses were used as starting glass compositions. J-13 water was considered to be representative of groundwater at Yucca Mountain. A total of 10 grams of each glass was assumed to dissolve congruently into a kilogram of J-13 water in a closed system. No inhibitions to precipitation, except for crystalline SiO2 polymorphs, were assumed to exist. Radiolysis and materials interactions were not considered.

Simulation results predict that radionuclides and other glass components precipitate predominantly in the form of oxides and hydroxides, together with carbonates, silicates and phosphates. Precipitates appear to be effective in limiting the concentrations of radionuclides and other elements in solution. The general compositional trends in precipitates and solution chemistry are the same in the West Valley and DMPF simulations, except for variations arising from differences in glass chemistry.

Concentrations of elements released from glass increase until the solution reaches saturation with respect to solids that contain these elements. Elemental concentrations are then predicted to remain constant, increase or decrease depending on: 1) whether the reaction between the dominant aqueous species of the element in solution and its precipitate is pH and/or Eh-dependent; 2) whether the species distribution of the element in solution changes significantly in response to changes in pH, Eh, or other factors; and 3) the competition with other phases for elements required to form the precipitate. pH increases from 7.3 to 9.8 and from 7.2 to 10 in the West Valley and DWPF simulations, respectively. Eh decreases abruptly from about 0.5 to 0.3 volts after dissolution of 3.4 and 5.8 grams of glass in the Nest Valley and DMPF simulations, respectively, because of depletion of dissolved oxygen in solution. Complexing of aqueous species has a significant impact on radionuclide concentrations in solution; predicted concentrations of U in solution, for example, are controlled by the presence or absence of P in solution because H2PO4 is an extremely effective complexing agent for U.

Flow-leaching studies were conducted using borosilicate glass beads containing 241-Am, 237-Np and 238-Pu. The leachates were a bicarbonate groundwater and a brine. Tests were carried out unger oxidizing and reducing conditions. In bicarbonate groundwater (2.6 × 10−3 M NaHCO3, pH 8.4) under oxic conditions, steady-state leach-rates were found to increase in the order Pu∼Am<Np, with Np species leached congruently with sodium. Direct identification of Np carbonate complexes in the leachate was carried out by. electromigration. A reversed order of leachability was obtained in a reducing environment. In this case, Pu was found to be associated with particulate matter. In a salt brine leachate (5.4 M NaCl, pH∼6) a periodic cycle of increasing-decreasing concentrations was observed under reducing conditions. This was found to be dependent upon the redox state of actinides. Radionuclide adsorption into an unstable gel layer was postulated.

The corrosion behaviour in distilled water of two simulated candidate high level waste borosilicate glasses (SAN602519L3C2 and SM58LW11) his been investigated at 90°C for different SA/V condition's 10, 100, 7800 m−1). The experimental data were modelled using the PHREEQE and GLASSOL computer codes. The model is quite successful for describing the corrosion behaviour, using experimentally derived values for the forward rate, silica saturation and the final rate. GWss SAN60 is more stable than glass SM58 at SA/V values of 10 and 100 m−1, but in the long term the relative performance is inverse. Indeed, the high Al content of SAN60 induces the creation of analcime crystals after SiO2 has reached its saturation concentration in solution, which cause an enhancement of the final rate of dissolution of the glass; for SM58 on the contrary the SiO2 solution is a stable condition.

This study provides experimental data on the corrosion of lead-iron phosphate (LIP) glass (6.4 wt% LWR waste loading) and the first comparison with the dissolution kinetics of borosilicate (BS) nuclear waste glass (all experiments at 90°C). Based on previous experiments the hypothesis was made that the alteration phenomena and the corrosion mechanism are analogous to what is known for BS glasses. The corrosion rate was found to be constant, 0.06gm−2d−1, at high flow rates (MCC 5 type test), based on total mass loss and on ion conductivity measurements in the leachate. In an MCC 1 type experiment, doped glasses (32p, 210Pb and 137Cs, respectively) were corroded and the release of activity into solution monitored. The release rates were 0.05gm−2 d−1 in the beginning, but decreased drastically when the solution became saturated with respect to Pb(OH)2 and Pb3 (PO4)2. Glass corrosion continued at a low rate as indicated by Cs activity measurements. The higher chemical durability of the LIP glass vs. BS glasses in DI water is a result of a smaller initial corrosion rate (10 to 10O×) and the fact that the saturation concentration is lower for LIP glass, i. e. it takes less dissolved glass (10 to 50×) to reach saturation than in the case of BS glass.

As part of the effort by the Nevada Nuclear Waste Storage Investigations (NNWSI) Project to evaluate the volcanic tuff beds of Yucca Mountain, Nevada, as a repository for the permanent storage of high-level nuclear waste, the interaction of actinide-doped Savannah River Laboratory (SRL) 165 type glass with the unsaturated repository environment has been studied. The NNWSI Unsaturated Test method has been used, and the results from batch and continuous tests completed through 18 months demonstrate that several interactions are important for controlling both the reaction of the glass and the release of radionuclides. These interactions include (1) the reaction between the glass and moist air with interludes of liquid water contact, which results in the release of alkali metals from the glass; and (2) the reaction between standing water, glass, and presensitized 304 L type stainless steel which results in breakdown of the glass matrix and the release of radionuclides from the glass-metal assemblage. A comparison of the results of the Unsaturated Test with those of parametric experiments illustrates the importance of presensitized steel in enhancing the glass reaction, and demonstrates the applicability of the Unsaturated Test to those conditions anticipated to exist in the NNWSI repository horizon.