The German waste glasses VG 98/12.2 and GP 98/12.2, a Li-doped version GP 98/12.2 Li and a French waste glass SON 68–18–17 Li were produced as cylinders with different cooling rates and with different additions of simulated high level waste (HLW). The glass cylinders were drop-tested with speeds of up to 125 m/s. No measurable influences of glass compositions or cooling time could be identified. The dominant parameter in size distribution of fracture pieces and fines was the drop velocity (i.e. impact energy) whereas the effects of different glass temperatures during the drop tests and drop ground properties were negligible. Vickers indentation and short rod fractometry tests on pieces surviving the drop tests showed a small but scientifically interesting effect: the fracture toughness Kic increased in the order: base glass, glass containing fission products without noble metals, and glass containing all fission products, and there was a small increase in K[c for faster cooling rates. The effect of the noble metal precipitates was particularly pronounced and was due to crack branching or arrest at the metal precipitates. All these effects, however, were overcompensated by the influence of decreasing K[c in the presence of humidity and by effects of radiation damage: damaged Cm-244 doped glasses and all glasses stored and measured in a dry N2 atmosphere showed a larger increase in Kic than that due to the above differences in composition and cooling rates.

The fracture and release mechanism of radioactive aerosols of HLW glass and HLW canisters are studied experimentally by laboratory scale and full scale drop tests. The experimental conditions model the conditions of accidental drops in a deep salt repository. The laboratory scale drop tests have a scaling factor of 1:10. Accelerated probes of simulated HLW glass impact on a ground plate and the size distributions of broken fines and released aerosols are measured by sieving and scanning electron microscopy (SEM) of aerosol samples.

The impact velocity is determined as the dominating impact parameter. Further parameters tested, such as waste glass composition, cooling time (residual thermal stresses), probe temperature at impact, and ground characteristics, show no measurable influence. Source terms of released respirable aerosols are evaluated for two reference cases, borehole drop (impact velocity v = 80 m/s) and reloading hall drop (v = 14 m/s), the values being 0.1 % and to 2.10-4 % respectively of the glass probe mass. The full scale drop tests are performed with European Standard HLW canisters. The canisters keep their integrity in all tests up to drop heights of 14 m. On opening the canisters, the broken fines are analyzed by sieving. The results are in good agreement with the small scale tests and confirm their acceptability for use in a safety analysis.

Various cooling procedures have been applied to canisters filled with inactive simulated HLW glass and the measured temperature distributions have been compared with numerically computed data. Stress computations of the cooling process were carried out with a finite element method. Only those volume elements having temperatures below the transformation temperature Tg were assumed to contribute thermoelastically to the developing stresses. Model calculations were extended to include real HLW glass canisters with inherent thermal power. The development of stress as a function of variations of heat flow conditions and of the radioactive decay was studied.

Technetium is a potentially hazardous radionuclide in the waste vitrification process as its oxides are volatile. Electrochemical measurements, i.e. potentiometry and voltammetry are applied to investigate the redox reactions of Tc in the glass melt. Re is studied for comparison. Under normal oxidizing conditions Tc is in the -T-state, and substancial losses of Tc occur as a result of compound volatility. Reducing conditions lead to the formation of TcO2, which precipitates from the melt. Stronger reducing conditions lead to the formation of metallic Tc

Technetium is a hazardous fission product with a long half-life. In vitrification of nuclear waste, technetium tends to be lost substantially by evaporation [1], and the formation of gaseous Tc oxides is assumed to be the reason. Reliable thermochemical treatment of the problem is difficult, since data on the Tc-O system are surprisingly scarce [2]. Therefore, the system Re-O is treated for comparison. Key thermodynamic data for the condensed rhenium oxides exist [3,4,5] as well as measurements on the sublimation and the evaporation of the oxides [6–11]. Consistency of the different data is investigated by assessing the sublimation data of solid Re2O7, using them to calculate other sublimation equilibria and to compare the results with published measurements. Then a predominance area diagram is constructed and discussed with respect to the pressures of the gaseous oxides and their dependence on the temperature, oxygen partial pressure, and condensed oxide phase present. Predominance area diagrams of the Tc-O system are calculated and critically discussed. Owing to the small amout of available data, the possible existence of solid TcO3 is discussed. Comparison with the system Re-O is used to clarify, where further investigations need to be done.

Property models have been developed for the major properties that need to be controlled in the production of borosilicate glasses for West Valley high-level nuclear waste immobilization. The chemical durability is the most important parameter for product performance, while melt viscosity is the most critical parameter in assuring the processability of the glass. Simple models for these properties are described that are based on data from numerous glasses which were prepared with compositions in the region around the West Valley reference glass. A scheme for optimization of the target glass and for predicting the acceptability of glasses resulting from natural process variations is illustrated. This involves integration of the product models with a process model that was described previously. This approach has guided the present placement of the West Valley reference glass.

Synroc containing mixed fission products and four actinides (237Np, 239Pu, 241Am and 244Cm) was fabricated by hot pressing on a 75 g-scale in separate hot-cell and glove-box facilities. Samples were physically characterised by density, which ranged from 98.2 to 99.1% of theoretical, by scanning electron microscopy (SEM) and by the cc-track etch technique to determine the distribution of actinide elements.

The chemical durability of specimens prepared from the hot-pressed Synroc was determined in 460 leach tests carried out under static conditions at 70°C. Release of 137Cs dominated the fission-product source term while 237Np was the most leachable actinide element. Short-term release is dominated by the initial spilce from Cs and other readily soluble species at grain boundaries and in non-equilibrium phases. Long-term release is probably controlled by matrix solubility. A phenomenological model is developed which accounts for both short- and long-term release.

The perovskite CaTiO3 is one of the major phases of the Synroc titanate mineral assemblage. Its chemical durability in an aqueous environment, which is relevant to the Synroc concept, has been investigated by solution analysis, surface analysis and electron microscopy. In general it has been found that dissolution due to base catalyzed hydrolysis is the most significant mechanism of attack; the results suggest that an ion exchange mechanism is confined to the first monolayer. Below 90°C the extent of attack, and release of Ca into solution, is limited by the formation of a titanaceous amorphous layer (” 100A thickness) which imposes a reaction constraint at the film-solid interface. Dissolution may also be constrained by the pH and Ca2+concentration in the bulk liquid. Above 90°C dissolution is relatively less constrained due to instability of the amorphous layer which is replaced by nucleation and epitaxial growth of TiO2 on the dissolving substrate. Thermodynamic stability of CaTiO3 for T< 90°C can easily be engineered into the waste repository, while hydrothermal stability (T > 90°C) is more difficult to achieve.

The leaching performance of pressureless sintered SYNROC crystalline ceramic waste forms containing simulated high-level radioactive waste has been evaluated under different leach conditions, i.e. in deionized water and in a simulated brine (Q-brine) at different leach temperatures. The leach solutions were analyzed using inductively coupled plasma and atomic absorption spectrometry (ICPS and AAS) to determine the concentrations of the simulated waste elements and the matrix components released into the leachants. The leach rates of Ba and Ca were in the range of 0.1 to 0.02 g/(m2.d) after one week of leaching in deionized water at 150 °C. No significant difference in the leach behaviour of SYNROC was observed in distilled water and in simulated brine. The leached surfaces of SYNROC were characterized using Rutherford backscattering spectrometry (RBS), secondary ion mass spectrometry (SIMS), scanning electron microscopy (SEM) and elastic recoil detection analysis (ERDA). RBS and SIMS analyses revealed the existence of near-surface layers enriched in Ti and depleted in Ba and Ca. Elastic recoil detection analysis showed the existance of hydrated surface layers on the leached SYNROC which were much thinner than those observed in parallel work with borosilicate waste glasses. Radiation damage produced by ion implantation increased the otherwise very-small leach rates of SYNROC. At high damage levels, SYNROC was shown to become amorphous (metamict).

Minerals of the crandallite group (Ca, Sr. Ba, Pb, REE) H Al3((po4) 2/(OH)6) exhibit manifold possibilities for element substitutions. They occur as hydrothermal precipitates together with tin tungsten mineralization [1], as placer minerals [2], neo-formed minerals in lateritic profiles [3, 4] and lake sediments [5].

An easy and cheap method of synthesizing an artificial crandallite-type material was developed. By adding 4.2 g of crandallite to an aqueous solution containing 300 g/l sodium nitrate, 0.6 g/l europium nitrate and 0.6 g/l thorium nitrate, europium and thorium were effectively separated from the solution. This was also demonstrated for Ce. La, Nd, Gd, Zr, Ba and Sr.

Artificial crandallite-type material laden with strontium was subjected to a Soxlet test. A dissolution of strontium of 0.06 to 0.08 wt.% with reference to the total strontium content resulted. As shown by Herold [6] the substitution of Ca by larger bivalent or higher charged ions results in an increase in the thermodynamic and thermal stability of artificial crandallites.

Cubic pyrochlore structure types, A2-mB2O6(O, OH, F) i-n*pH2O, and their derivatives (e.g., monoclinic zirconolite) are important actinide-bearing phases in polyphase, ceramic waste forms (e.g., SYNROC). These waste form phases may typically accumulate alpha-decay doses of 1025 alpha-events/m3 in 1, 000 years or 1026alpha-events/m3 in one million years (i.e., for SYNROC with 20 wt. % HLW). Natural pyrochlores have calculated doses ranging from 1024 to 1027 alpha-events/m3 (= 0.02 to 50 dpa) which have accumulated over ten to a thousand million years. Actinide doping experiments typically reach doses of 1025 alpha-events/m3over periods of several years. Detailed x-ray diffraction analysis of natural samples reveals that the alpha-decay dose at which there is an initial loss of crystallinity (i.e., transition from crystalline to the aperiodic, metamict state as a result of alpha-decay damage) increases as a function of the geologic age of the sample. The increase in the calculated alpha-decay dose which is associated with a specific degree of damage (e.g., loss of x-ray diffraction intensity) is attributed to annealing of isolated alpha-recoil tracks back to the original, crystalline structure. Based on a model of gradual track fading, the alpha-recoil tracks in natural pyrochlores have mean lives on the order of 108 years. In contrast, minerals which remain crystalline (e.g., uraninite, UO2) despite doses of over 1027 alpha-events/m3 have mean alpha-recoil track lives of approximately 104 years. This demonstrates that the microstructure of alpha-decay damaged materials depends not only on the total alpha-event dose, but also on the annealing kinetics of alpha-recoil track fading. Therefore, the prediction of the long-term performance and final state of crystalline phases in ceramic nuclear waste forms requires the determination of alpha-recoil damage annealing as a function of time and temperature.

X-ray absorption spectroscopy has been used to investigate the Nb B-site in pyrochlores (A1.2B2O6Y0–1, Fd3m, Z=8) and samarskites (A3B5O16) in both metamict and annealed condition. The XANES and EXAFS measurements indicate significant changes in pyrochlore and smaller changes in samarskite as a result of radiation damage. In the metamict state the Nb site in both pyrochlores and samarskites is similar to Nb in Nb2O5. Short Nb-O (1.65Å) bonds are not disrupted by alpha-decay/recoil-nuclei events as much as longer bonds (2.00Å). This increases the asymmetry and static disorder at the local Nb site while long range order is greatly diminished resulting in considerable distribution in Nb-M distances and bond angles.

The remarkable geological stability of the actinide-containing rare-earth orthophosphate mineral monazite has led to consideration of synthetic analogs of this material as the wasteform for high-level nuclear wastes (1). Monazite is resistant to metamictization in spite of the alpha radiation from decaying actinides to which it has been subjected during its lifetime. This material is also very insoluble in most groundwaters. Consequently, dissolution of the mineral and leaching of its constituent actinides are very small. However, leaching experiments on natural monazite specimens, presented here, show that radioactive decay and radiation damage produce significant differences in the release rates of several isotopes into aqueous solutions. Within the duration of the present experiments, 6.3 to 6.8 years, these effects are particularly prominent in the nuclides in the decay chain of 232-Th. This series includes the alpha-recoil 228-Ra atom (a 5.8-year fT emitter formed directly by 232-Th decay), and its daughter 228-Th (a 1.9-year a emitter, produced by 3 decay). The present experiments exhibit preferential release into solution of 228-Th relative to 232-Th by a factor of 1.3 to 21. This highly enhanced 228-Th release is explained by four different isotope incongruency effects that operate during leaching. One mechanism involves the preferential etching of alpha-recoil tracks. The three other mechanisms depend upon the behaviour of 228-Ra. In particular, dissolution-released, recoil-released and diffusion-released 228-Ra atoms (the latter process is applicable to a high-temperature anneal) are observed to remain with the solid as an insoluble precipitate, possibly loosely attached to the mineral surface. Subsequent decay of this released insoluble 228-Ra immediately leads to the formation of readily dissolved 226-Th. The phenomena may be quite general, so potential wasteforms should be examined for these isotope fractionation effects in order to insure satisfactory prediction of the long-term leaching of actinides from the material.

Due to the long containment periods required for radionuclides in a nuclear waste disposal vault, the justification that a particular containment system is acceptable will be based on relatively short-term experimental data used to support predictive models. To justify this approach, we must possess a sound mechanistic understanding of processes such as fuel dissolution, radionuclide release and container corrosion. Since these processes are driven by oxidants in the vault, it is natural to study them by electrochemical techniques. In this paper, we have reviewed a number of electrochemical methods used in the study of waste vault reactions. More detailed descriptions are given of the development of predictive models for the dissolution of UO2, the crevice corrosion of titanium alloys and the uniform dissolution of copper.

Analytical transmission electron microscopy (AEM) was performed on light-water reactor (LWR) spent fuel samples from different radial locations in several low-gas release fuels. The examinations revealed high densities of sub-micrometer gas and solid aggregates. Precipitate coarsening and a change in the nature of the fission gas precipitates was observed near the centers of the fuel pellets. The fission-product phases were: 1) ε-ruthenium (Mo-Ru-Tc-Rh-Pd solid-solution alloy); 2) unidentified gases in intra- and intergranular bubbles from fuel edge to mid-radius locations; 3) dense, highly pressurized xenon/krypton “particles” associated with intragranular °-phase near the fuel centers. All other fission products apparently remained in solution. The nature and distribution of these phases are likely to affect fuel behavior in waste repositories and may be important for understanding fission gas behavior.

In the context of the Canadian Nuclear Fuel Waste Management Program, experiments were performed in which whole, intentionally defected used CANDU™ fuel elements were exposed to deoxygenated saline solutions at 100°C for 484 days. Results suggest that both 90Sr and 99Tc are preferentially released from grain boundaries and that 85Kr release is a good measure of the extent of grain-boundary dissolution. Severe grain-boundary attack in particular areas of the fuel is also evident from scanning electron microscope observations. It is hypothesized that such localized attack is related to fuel porosity. Because of its poor corrosion performance in saline solutions, the Zircaloy sheath cannot be considered as a radionuclide release barrier.

Dissolution experiments have been performed at 95 or 100°C on used fuel in the presence of various components of a disposal system: granite rock, clay (Na-montmorillonite, illite), container (titanium), and two groundwaters with different ionic strengths (0.0015, 0.28). The results showed that Cs and Sr release to solution is highest in the high-ionic-strength groundwater and that clay type is less important in determining short-term release from the fuel. It also appears that after an initial release period lasting about 10 to 20 days, radionuclide release from the fuel to solution slows considerably. Cs was found to sorb more effectively than Sr on illite and montmoril-lonite, while Sr sorbed more effectively than Cs on the granite.

The effects of groundwater radiolysis on the oxidation and dissolution of UO2 has been studied. The production of a specific radical was maximized by using suitable solution chemistry, and the effect of a particular radical was investigated. Experiments were performed to study the effects of OH, O2, and CO, radicals. An electrochemical cell and a UO2 electrode were designed to do ijri situ electrochemical measurements in gamma fields, 300 -20 Gy/h. During radiolysis, the oxidation of UO2 was monitored by recording the corrosion potential of the UO2 electrode as a function of time. The fastest increase and highest corrosion potentials were observed in solutions favoring the formation of O2 radicals. XPS analysis of the UO2 surface showed that oxidation has proceeded beyond the UO2.33 stage. Computer calculations of water radiolysis showed that steady-state concentrations of the radicals are very low, about 10−9 mol-dm−3, in most cases, and even lower if the reactions with the UO2 surface are included. The radicals are produced at a much faster rate than the observed changes in the corrosion potential, suggesting thereby that another slower reaction is corrosion rate controlling.

Characterizations are being conducted on light-water reactor (LWR) spent fuel with peak burnups from ∼31 to 46 MWd/kgM and rod-average fission gas releases of ∼0.3 to 11%. Measured concentrations of nuclides agreed within ∼10 to 15% of the predicted amounts. Radiochemistry of deposits on the cladding interior surface, fuel ceramography, beta/gamma autoradiog-raphy, and gamma scan data were consistent with fuel burnups and apparent operating temperatures. Microstructural observations also indicated increased porosity at the fuel center in high-release fuels and at the edge in high-burnup fuels. Cladding oxide thickness increased with exterior cladding temperatures as did the trend in hydriding.

The solubilities of uranium measured in deionized water, in sodium bicarbonate solutions as a function of carbonate concentration and in two types of synthetic groundwaters have been compared with those predicted using the reaction pathway program, PHREEQE. All the measurements were carried out under both air-saturated, oxidizing and anoxic, reducing conditions.

The experimental solubility values of uranium under oxidizing conditions are, in general, lower when compared to the corresponding theoretical ones calculated by PHREEQE. A critical factor is the choice of solubility constant for the solid phase. The reason for the lower solubility values may also be the mechanism of dissolution leading for example either to a situation where low dissolution rate is a limiting factor, or to formation of some solid phase of uranium with lower solubility. The experimental solubility values under reducing conditions appear to be in good agreement with the theoretical solubility values.