As part of a programme to identify and quantify the mechanisms controlling the aqueous corrosion of spent fuel,a short fuel rod was refabricated from a full-length rod from the Swedish Ringhals 1 BWR (burnup 44-48 MWd/kgU). After a conditioning period at 25 kW/m,the rod was power-bumped in Studsvik Nuclear´s R2 test reactor up to a maximum linear heat rating of 43 kW/m,with a hold time at power of 3 hours,imposing steeper radial temperature gradients on the fuel pellets. A few hours after termination of irradiation,the rod was shipped to Studsvik Nuclear´s Hot Cell Laboratory,where the pellet inventories of gamma-emitting fission products were determined non-destructively.Two 20mm long fuel/clad sections were cut from the rod at positions corresponding to power-bump LHR values of 43 and 33 kW/m.Two similar sections were cut from the original rod as references,together with a shorter section for destructive inventory analysis.

The fuel/clad sections were leached in 200ml of synthetic groundwater (l23ppm HCO3) under oxic conditions for consecutive periods of 7,7,14 and 14 days.The samples arising from the tests were analysed by conventional radiometric and radiochemical methods,and also using an ICP-MS (Inductively Coupled Plasma Mass Spectrometer) instrument commissioned for active use.

Fractional release data for both radioactive and stable fission products has been obtained. Since the main aim of the test was the study of the early stages of corrosion,expected to be dominated by migrational effects in the fuel during irradiation,particular attention has been paid to Cs,I,Rb,Sr, Ba,Mo and Tc. Although the migration and selective leaching of Cs and I are well-established,it is shown that there is also appreciable leaching of Rb nuclides,whereas the results show no significant Sr migration.This represents support for the hypothesis that the apparent enhanced leach rate of Sr-90 observed experimentally can be due to migration of Rb-90 precursors.

Flowthrough dissolution tests have been conducted on two different light-water-reactor spent fuels oxidized to U4O9+x or U3O8. Oxidation had a bigger impact on the dissolution of U and a smaller impact on the dissolution of Tc from the fuel with higher burnup and higher fission gas release. Possible reasons for the observed differences in test results are discussed, but clarification awaits results from tests on other fuels, which are in progress.

Preliminary results for the composition of the leachate from unsaturated tests at 90°C with spent fuel for two successive periods of ~60 days each with pretreated J-13 groundwater are reported. The pH of the leachate solutions ranged from 4 to 7. The americium concentration was 104 to 105 greater than that reported for saturated spent fuel tests in which the leachate pH was 8. The major fraction of material in the leachate was present as colloids containing both americium and curium. The presence of actinides in a form not currently directly included in repository radionuclide transport models provides information that can be used in spent fuel reaction modeling, the performance assessment of the repository and the design of the engineering barrier system.

Previously it was not possible to predict reliably the rate at which spent fuel would react with groundwater because of conflicting data in the literature. The dissolution of the UO2 spent fuel matrix is a necessary step for aqueous release of radioactive fission products. Statistical experimental design was used to plan a set of UO2 dissolution experiments to examine systematically the effects of temperature (25-75°C), dissolved oxygen (0.002-0.2 atm overpressure), pH (8-10) and carbonate (2-200x10-4 molar) concentrations on UO2 dissolution. The average dissolution rate was 4.3 mg/m2/day. The regression fit of the data indicate an Arrhenius type activation energy of -8750 cal/mol·K and a half-power dependence on dissolved oxygen in the simulated groundwater.

High-burnup (<50 MWd/kgU) spent fuel samples of various sizes were exposed to NaCl solutions under static, anaerobic and reducing conditions. The accumulated corrosion time was about 200 days. Gas phase and leach solutions were analyzed. By dissolving mm sized fragments in large volumes of solution, saturation effects were avoided and upper limits for intrinsic dissolution rates of about 5-20 mg/(m2d) were measured. Surface area normalized reaction rates were significantly lower when using fine grained fuel powder (estimated sample surface area to solution volume ratio S/V ca. 3000 m-1), indicating saturation effects. The maximum concentrations of Pu and Am in the tests are close to reported solubility limited concentrations in pure 5m NaCl solutions in the absence of radiolysis effects. The presence of iron effectively reduces the solution concentration of all measured radionuclides (except Cs).

The separate effects of alpha- and gamma-radiolysis on UO2 dissolution can be studied with unirradiated UO2, whereas studies vith used nuclear fuel necessarily always include both alpha- and gamma-radiolysis effects. This paper attempts to separate these effects by comparing the leaching behaviour in saline solution of a number of UO2 samples (each vith a particular radiation characteristic or chemical property inherent to used fuel) vith the leaching behaviour of used fuel. Data from leaching experiments vith lov- and high-burnup CANDU (CANada Deuterium Uranium) fuels are also compared. The results indicate that the presence of an alpha field at 100°C under reducing conditions does not increase UO2 dissolution but suggest that the combined effects of the beta and gamma fields in used CANDU fuel may enhance UO2 dissolution.

Photothermal Deflection Spectroscopy (PDS) has been applied to the study of uranium oxide electrochemistry. PDS measures the optical absorption of the sample surface and concentration gradients formed in the electrolyte. Both of these measurements are performed in situ under dynamic conditions. The combination of these two measurements provides information that can be used to infer the mechanism of the UO2 surface chemistry. These studies of the uranium dissolution mechanism are performed in pH 10.5 sodium sulfate electrolytes at 22°C. The electrolytes are free from oxygen, and complexing species. Our results suggest that dissolution of UO2 can occur at oxidizing potentials as low as -300 mV vs. saturated calomel electrode (SCE). The optical absorption and concentration gradient results provide evidence for a substantial surface change that occurs at an oxidation potential of +300 mV. The results show that the surface layer formed by this change dissolves slowly by a non-electrochemical reaction.

Electron beam techniques have been used to characterize uranium-contaminated soils at the Fernald Site, Ohio. Uranium particulates have been deposited on the soil through chemical spills and from the operation of an incinerator plant on the site. The major uranium phases have been identified by electron microscopy as uraninite, autunite, and uranium phosphite [U(PO3)4]. Some of the uranium has undergone weathering resulting in the redistribution of uranium within the soil.

Studies of the spent fuel waste package have been conducted through the use of a Monte-Carlo neutron simulation program to determine the ability of the fuel to sustain a chain reaction. These studies have included fuel burnup and the effect of water mists on criticality. Results were compared with previous studies.

In many criticality studies of spent fuel waste packages, fresh fuel with an enrichment as high as 4.5% is used as the conservative (worst) case. The actual spent fuel has a certain amount of “burnup” that decreases the concentration of fissile uranium and increases the amount of radionuclides present. The LWR Radiological Data Base from OCRWM has been used to determine the relative radionuclide ratios and KENO 5.a was used to calculate values of the effective multiplication factor, keff.1

Spent fuel is not capable of sustaining a chain reaction unless a suitable moderator, such as water, is present. A completely flooded container has been treated as the worst case for criticality. Results of a previous report that demonstrated that keff actually peaked at a water-to-mixture ratio of 13% were analyzed for validity. In the present study, these results did not occur in the SCP waste package container.

Magnetic measurements constitute a promising method for the characterization of nuclear waste glasses in view of their simplicity and small sample weight requirements.

Initial studies of simulated high-level waste glasses show that the Curie constant is generally a useful indicator of the Fe2+:Fe3+ ratio. Glasses produced by air-cooling in large vessels show systematic deviations between experimental and calculated values, which are indicative of the presence of small amounts of crystalline iron-containing phases. Most of the iron in these phases becomes dissolved in the glass upon re-heating and more rapid quenching. The studies further show that upon leaching the glass in water some of the iron in the surface regions of the glass is converted to a form which has high temperature-independent magnetic susceptibility.

The Materials Characterization Center (MCC) at Pacific Northwest Laboratory(a) conducts Nuclear Waste Analytical Round Robins on simulated waste glasses with and without radionuclides. The analytical round robins are used to provide analytical laboratories with a forum for improving the accuracy and precision of analytical techniques that will be used for characterizing nuclear waste glasses. Chemical analysis data on glass waste forms produced by the nuclear waste community may need to withstand legal challenges for waste material storage qualification and for licensing repositories. Determining the quality of analytical data is important for meeting these challenges, making process operation decisions, and predicting long-term waste behavior.

The MCC has conducted six round robins for the waste management, research, and development community from 1987 to the present time. The laboratories that have participated regularly are Ames, Argonne, Catholic University, Lawrence Livermore, Pacific Northwest Laboratory, Savannah River, and West Valley Nuclear. The glass types analyzed in these round robins all have been simulated nuclear waste compositions expected from the vitrification of high-level nuclear waste. A wide range of analytical procedures have been used by the participating laboratories, including Atomic Absorption spectroscopy, inductively coupled plasma-atomic emission spectro-scopy, direct current plasma-emission spectroscopy, and inductively coupled plasma-mass spectroscopy techniques. The consensus average relative error for Round Robins 1 through 6 is 5.4%, with values ranging from 9.4 to 1.1%. The trend on the average improved with each round robin. When the laboratories analyzed samples over longer periods of time, the intralaboratory variability increased. Lab-to-lab variation accounts for most of the total variability found in all the round robins. To date, participation in the radiochemistry portion has been minimal and the analytical results poor when compared to the nonradiochemistry portion of the round robins. Additional radiochemical work is needed in future round robins.

Glasses containing fixed mutual ratios of Al, B, Ca, Fe, Mg, Na, and Si were melted while varying the fluoride content. The effect of fluoride was to suppress the formation of non-fluoride-containing crystals in the composition range from 0 to 11 wt. % F; above 11 wt. % F, the formation of fluoride containing crystals began to define the liquidus temperature. The effect of fluoride on the liquidus temperature was quantified and used to predict the liquidus temperature of melts made from real radioactive wastes. Although the liquidus temperatures of the surrogate and radioactive glasses agreed well, the types of crystals formed differed due to slight compositional differences introduced by the waste stream.

Previous study of crystallization behavior in heat-treated Fernald waste glasses has produced an extensive data base of crystal phases likely to appear in various composition ranges and their corresponding liquidus temperatures. In addition, we have frequently observed amorphous phase separation in these glasses and, occasionally, evidence of crystallization originating from such phase separation. These glasses contain more than 10 components. The composition ranges for the major components are: MgF2 10–26 wt%; CaO 4–27 wt%; Al2O3 3–15 wt%; SiO2 25–40 wt%. The morphology of the phase separation as observed in the Scanning Electron Microscope (SEM) is dark, spherical globules dispersed in a continuous matrix. Globules are depleted in Mg, Ca and F, and enriched in Al and Si compared to the matrix. Phase separation occurs more frequently in melts relatively higher in Si and F. A more systematic study on a simplified and simulated seven component system (Al2O3, B2O3, CaO, Fe2O3, SiO2, Na2O and MgF2) has been undertaken to determine the subliquidus miscibility gap and liquidus curve data. Glasses were formulated by varying the concentrations of MgF2, CaO, Al2O3 and SiO2 within the ranges specified above at fixed levels of Fe2O3, B2O3 and Na2O. The miscibility gap and liquidus curve were obtained by heat-treating the glass samples at different temperatures and observing any phase separation and crystallization in the SEM and the Transmission Electron Microscope (TEM). We report here the results of this study to enhance the understanding of the thermodynamic properties of multi-component silicate systems which are usually the basis of nuclear waste glasses.

Results are presented from treatability studies on the vitrification of hazardous and low-level radioactive soils and sludges from the Fernald site. The majority of the contaminants of concern are contained in large volumes of sludges accumulated from uranium and thorium production operations that are presently stored in pits at Fernald. These sludges present particular challenges to the vitrification process in that they contain significant amounts of fluorides (mostly magnesium and calcium). In this study, glass formulations used sludges, soils and a minimum amount of additives resulting in waste loadings as high as 94%. Several of these formulations were evaluated by continuous feeding of slurries into a joule-heated vitrification system. Melter off-gases were treated in a two-stage NaOH scrubbing system. HF acid fumes were neutralized, resulting in the formation of a NaF solid precipitate which is removed from the scrubber and recycled to subsequent melter feeds. Elemental and isotope compositions of the glasses, scrubber solutions, NaF precipitate and off-gases show that greater than 99.7% of uranium and thorium, 95% of Pb, Cd, Ni, Sb, Cr, and Hg were retained in the glass. Fluoride and other gaseous emissions from the scrubber system were below regulatory limits, even before HEPA filtration.

The Hanford Waste Vitrification Plant (HWVP) is being designed for the Department of Energy to immobilize high-level and transuranic wastes as glass for permanent disposal. Pacific Northwest Laboratory is supporting the HWVP design activities by conducting laboratory-scale studies using a HWVP simulated waste slurry. Conditions which affect the slurry processing chemistry were evaluated in terms of offgas composition and peak generation rate and changes in slurry composition. A standard offgas profile defined in terms of three reaction phases, decomposition of H2CO3, destruction of NO2-1, and production of H2 and NH3 was used as a baseline against v/hich changes were evaluated. The test variables include nitrite concentration, acid neutralization capacity, temperature, and formic acid addition rate. Results to date indicate that pH is an important parameter influencing the N2O/NOx generation ratio; nitrite can both inhibit and activate rhodium as a catalyst for formic acid decomposition to CO2 and H2; and a separate reduced metal phase forms in the reducing environment. These data are being compiled to provide a basis for predicting the HWVP feed processing chemistry as a function of feed composition and operation variables, recommending criteria for chemical adjustments, and providing guidelines with respect to important control parameters to consider during routine and upset plant operation.

Predictions of long-term glass dissolution rates are highly dependent on the form of the affinity term in the rate expression. Analysis of the quantitative effect of saturation state on glass dissolution rate for CSG glass (a simple analog of SRL-165 glass), shows that a simple (1-Q/K) affinity term does not match experimental results. Our data at 100°C show that the data is better fit by an affinity term having the form where σ =10.

The R7T7 and synthetic basaltic glasses were submitted to corrosion in a saline MgCl2 dominated solution at 190°C. For both glasses, the early alteration product is a hydrotalcite-like compound in which HPO42-, SO4-2 and Cl- substitutes to CO32. The measured d003 spacing is 7.68 Å for the hydrotalcite formed from R7T7 glass and 7.62 Å for the hydrotalcite formed from basaltic glass which reflect the high aluminium content. Chemical microanalyses show that the hydrotalcite is subsequently covered by a silica-rich gel which evolves into saponite after few months.

Leaching studies were performed on powdered PNL 76-68 glass in de-ionized water in the presence of Mg solute. The results showed that the presence of Mg in the leachant greatly reduced the rate of glass dissolution. The equation Q=krα was used to express the experimental data. In the absence of Mg, α was about 1, i.e., the amount of glass dissolved was linear with time. In the presence of Mg, α was close to 0.5, i.e. the extracted amount was proportional to the square root of time. Therefore, the reduction of the dissolution rate of PNL 76-68 glass in the presence of Mg solute could be explained as a result of a change in the glass dissolution mechanism.

Comparative leaching studies on bulk soda-lime silicate glass in a sodium borate buffered system (pH 8.1) showed the same results. The presence of Mg in the leachant reduced the rate of glass dissolution. In the absence of Mg, α was about 1, while in the presence of Mg, α was 0.5. This change in α was not caused by changes in pH, and it represents a real change in the glass dissolution mechanism.

Naturally corroded, poorly durable glasses are compared with results from MCC-4 leach tests on similar glasses. It is suggested that different corrosion products are formed at different temperatures; therefore, care should be taken when interpreting experimental leach rates. Solution composition, pH and flow around the sample can also have a profound effect on corrosion rates.

Preliminary results show that glass durability is dependent on reactions occurring at the glass-solution interface. CSG glass (18.2 wt. % Na2O,5.97 wt. % CaO, 11.68 wt. % Al2O3, 8.43 wt. % B2O3, and 55.73 wt. % SiO2) dissolution and net surface H+ and OH- adsorption are minimal at near neutral pH. In the acid and alkaline pH regions, CSG glass dissolution rates are proportional to 2 to , respectively. In contrast, silica gel dissolution and net H+ and OH" adsorption are minimal and independent of pH in acid to neutral solutions. In the alkaline pH region, silica gel dissolution is proportional 0.9 to . Although Na adsorption is significant for CSG glass and silica gel in the alkaline pH regions, it is not clear if it enhances dissolution, or is an artifact of depolymerization of the framework bonds.