Explanation of the striking non-linear effect of glass composition on the aqueous dissolution represents an important challenge to existing dissolution mechanisms. Surface layers that are formed during glass dissolution may play an important role in this effect. One chemically reactive and one less-reactive nuclear waste glass (leachate concentrations differ by about a factor of 10) were reacted in deionized water. Two types of glass powders were used: Type A powders were pristine glass powders; Type B powders were the glass powders which had been reacted for 120 days to develop the surface layers. Both the solution concentrations and the surface layers were investigated. The experimental observations indicate that: (i) There is a range of glass compositions over which small differences in composition lead to large changes in both reaction rates and surface layer thickness; and (ii) The reaction rate is strongly affected by the formation of the surface layer (the layer appears to be protective) and cannot be explained in terms of saturation effects alone. The findings are contrary to the conclusion of a previous study and serve to highlight the inadequacy of existing dissolution models predicated on an overly simplistic mechanism, especially with regard to glass composition effects.

In our previous studies on the optimization of glass compositions for high-level nuclear waste vitrification it was found that, over certain composition ranges, PCT leachate concentrations increased dramatically with very small changes in glass composition. The large differences that are observed between the leachate pH values for the “durable” and the “less-durable” glasses is one possible cause for this strongly non-linear glass composition effect; conversely, the pH difference may be merely another symptom. In this study, four simulated nuclear waste glasses (two of the less-durable and two of the durable types), were leached in both zwitterionic and inorganic buffer solutions, at fixed pH-values in the ranges of 7 to 12. The very different leaching behaviors of the two types of glasses persisted and, furthermore, different pH-dependence was found despite their very similar glass composition. This study suggests that the leachate pH difference observed between the less-durable and the durable glasses under uncontrolled pH conditions is not the major cause of the large difference of leaching behavior between those glasses. The normalized release ratios of soluble components (B, Li, Na) to Si show significant differences for the two types of glasses.

A glass-based waste form used for the immobilization of high-level nuclear wastes should exhibit good resistance to aqueous corrosion since typically this is the primary process by which radionucleides could be released into the environment upon failure of other barriers. In the USA, the Waste Acceptance Product Specifications (WAPS) provides a set of requirements to ensure the consistency of the waste forms produced and specifies the Product Consistency Test (PCT) as a measure of relative chemical durability. While the PCT procedure permits usage of both Teflon and stainless steel vessels for testing of simulated development glasses, Teflon is not permitted for testing of production glasses due to radiative degradation. The results presented in this paper indicate that there are very significant differences between tests conducted in the two types of vessels due to the well-known permeability of Teflon to atmospheric carbon dioxide which results in lowering of the solution pH and a consequent reduction in the leach rate of silicate glasses. A wide range of nuclear waste glass compositions was subjected to the PCT procedure using both Teflon and stainless steel vessels. The magnitude of the effect (up to a factor of four for B, Na, Li concentrations) depends strongly on glass composition, therefore the isolated checks performed previously were inconclusive. The permeability to CO2 of two types of Teflon vessels specified in the PCT procedure was directly measured using buffer solutions: ingress of CO2 is linear in time, strongly pH-dependent, and was as high as 100 ppm after 7 days. In actual PCT tests in Teflon vessels, the total CO2 content was 560 ppm after 87 days and 1930 ppm after one year.

The dissolution kinetics of the French “R7T7” nonradioactive LWR reference glass in solutions containing dissolved humic acids were investigated at 90°C during static tests with imposed or free pH. Experiments conducted in highly dilute media, with a glass-surface-area-to-solution-volume (SA/V) ratio of 5 m-1, showed that the glass dissolution surface reaction is catalyzed by humic acids. With higher degrees of reaction progress (SA/V = 100 m-1 and free pH) the humic acids impose pH modifications on the system compared with inorganic media; moreover, they directly or indirectly enhance the dissolution of certain alkali metals and transition elements, forming aqueous complexes with the latter. During experiments with an imposed pH of 8.5 (SA/V = 1300 and 5300 m-1), the humic acids appear to cause increased silica solubility that cannot be accounted for by the formation of silica complexes. A residual corrosion rate in the humic acid media exceeding the rate measured in inorganic media suggests that, in addition to silica, one or more element complexes formed by humic acids may be a kinetically limiting factor. This hypothesis must be confirmed, however, as the quantity of humic acids per unit glass surface area was too small in this experiment to allow unambiguous characterization of the phenomenon.

A corrosion test of a model waste glass was carried out with and without carbon steel, and the leaching behavior of the glass was studied under reducing condition in a glove box purged with N2. The redox potential and pH were monitored, and the concentration of Fe, Na, Si, and Mo in the leachates were measured to understand the leaching behavior of the glass. The redox potential of the leachates obtained in the corrosion test coexisting with carbon steel under reducing conditions showed initially that it was based on an Fe0/Fe2+ redox potential, and gradually it increased to that based on Fe2+/Fe3+ one. The corrosion rate of the glass under oxidizing conditions was ten times greater than that obtained with carbon steel, and that under reducing condition with carbon steel was negligible small.

A corrosion test of model waste glass was carried out with and without a carbon steel under reducing condition in a glove box purged with nitrogen gas, and then the characteristics of alteration layers of the glass were examined.

On the glass surfaces corroded under reducing condition with carbon steel, a thick precipitated layer was observed by Scanning Electron Microscopy (SEM). In this layer concentrated iron was observed by Particle Induced X-ray Emission Spectrometry (PIXE) and Rutherford Backseattering Spectrometry (RBS). Such a thick precipitated layer was not observed on the corroded glass under oxidizing conditions with and without carbon steel and reducing condition without carbon steel.

Alteration phases, found on the leached surfaces and present as colloids in the leachates of 200-based frit (fully active and simulated) nuclear waste glass, reacted under static test conditions, at a surface area to leachate volume ratio of 20,000 m-1 for 15 days to 728 days, have been examined by analytical electron microscopy. The compositions of the secondary phases were determined using X-ray energy dispersive spectroscopy and electron energy loss spectroscopy, and structural analysis was accomplished by electron diffraction. Long-term samples of simulated glass, which had undergone an acceleration of reaction after 182 days, possessed a number of silicate secondary phases, including; smectite (iron silicate and potassium iron alumino-silicate), weeksite (uranium silicate), zeolite (calcium potassium alumino-silicate), tobermorite (calcium silicate), and a pure silica phase. However, uranium silicates and smectite have also been observed in tests which have not undergone the acceleration of reaction, in both the leachate and leached layer, suggesting that these phases are not responsible for the acceleration of reaction.

The experimentally determined initial dissolution rate R0 of nuclear glass was correlated with thermodynamic parameters and structural parameters. The initial corrosion rates of six “R7T7” glass samples measured at 100°C in a Soxhlet device were correlated with the glass free hydration energy and the glass formation enthalpy. These correlations were then tested with a group of 26 SON glasses selected for their wide diversity of compositions. The thermodynamic models provided a satisfactory approximation of the initial dissolution rate determined under Soxhlet conditions for SON glass samples that include up to 15 wt% of boron and some alumina. Conversely, these models are inaccurate if the boron concentration exceeds 15 wt% and the glass contains no alumina. Possible correlations between R0 and structural parameters, such as the boron coordination number and the number of nonbridging oxygen atoms, were also investigated. The authors show that R0 varies inversely with the number of 4-coordinate boron atoms; conversely, the results do not substantiate published reports of a correlation between R0 and the number of nonbridging oxygen atoms.

Obsidian and basaltic glass are opposite end-members of natural volcanic glass compositions. Syngenetic and diagenetic tensile failure in basaltic glass (low silica glass) is pervasive and provides abundant alteration fronts deep into the glass structure. Perlitic fracturing in obsidian (high silica glass) limits the alteration zones to an “onion skin” geometry. Borosilicate waste glass behaves similarly to the natural analog of basaltic glass (sideromelane).

During geologic time, established and tensile fracture networks form glass cells (a three-dimensional reticulated pattern) where the production of new fracture surfaces increases through time by geometric progression. This suggests that borosilicate glass monoliths will eventually become rubble. Rates of reaction appear to double for every 12C° of temperature increase. Published leach rates suggest that the entire inventory of certain radionuclides may be released during the 10,000 year regulatory time period. Steam alteration prior to liquid attack combined with pervasive deep tensile failure behavior may suggest that the glass waste form is not license defensible without a metallic- and/or ceramic-type composite barrier as an overpack.

Hydrothermal systems in the Taupo Volcanic Zone, North Island, New Zealand are being used as field-based modeling exercises for the EQ3/6 geochemical modeling code package. Comparisons of the observed state and evolution of the hydrothermal systems with predictions of fluid-solid equilibria made using geochemical modeling codes will determine how the codes can be used to predict the chemical and mineralogical response of the environment to nuclear waste emplacement. Field-based exercises allow us to test the models on time scales unattainable in the laboratory.

Preliminary predictions of mineral assemblages in equilibrium with fluids sampled from wells in the Wairakei and Kawerau geothermal field suggest that affinity-temperature diagrams must be used in conjunction with EQ6 to minimize the effect of uncertainties in thermodynamic and kinetic data on code predictions.

Ianthinite, UO2.86·1.5H2O, is a weathering product of uraninite, UO2+x, and has been reported as a corrosion product of UO2. Ianthinite alters to schoepite, UO3·2H2O, or studtite, UO4·2-4H2O, in water. The structures of ianthinite, schoepite and studtite are unknown, but their unit cell parameters differ significantly, and transformations are likely reconstructive.

XRD powder data for ianthinite and becquerelite are similar, and they are probably iso-structural; U4+, as U(OH)22+, replaces Ca2+ in the interlayer. The proposed formula is

U(OH)22+[(UO2)6O4(OH)6]·6H2O.

Oxidation in air, has been reported to form schoepite; however, in the absence of water, the oxidation of ianthinite,

U(OH)22+[(UO2)6O4(OH)6]·6H2O + 1/2 O2 => (UO2)2+[(UO2)6O4(OH)6]·7H2O ,

shows that epi-ianthinite is structurally distinct from schoepite. The oxidation of fine-grained material is rapid, and the color change — from dark purple to yellow — is dramatic. The acicular habit of ianthinite is preserved during oxidation, and acicular “schoepite” is observed in nature, as well as in leaching experiments on spent fuel. Possible substitution of Pu4+ for U4+ suggests that Pu-ianthinite may control Pu solubility in oxidizing, U-saturated solutions.

Mineral samples from the natural fission reactors 10 and 13 in the Oklo uranium deposits were studied using X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD) to gain information about the long-term behaviour of UO2 fuel in a geological disposal vault. Two samples from reactor zone 10 (samples # D81N-190292 and D73-88) and one sample from reactor zone 13 (sample # SD37-S2/CD) were analysed. Low-resolution XPS spectra were recorded to determine the major elements present in the ore. High-resolution spectra were recorded to gain information about the chemical state of the elements present in the mineral samples. The samples show low values for the U6+/U4+ ratio. The oxidation state of uranium in these samples is even lower than that in U409. The binding energies of the Pb 4f bands indicate most of the Pb is in the +2 oxidation state in these samples. The C Is band indicates the presence of organic carbon. XRD analysis shows that the main uranium-bearing phase is uraninite and lead is present mainly as galena. The significance of the results for nuclear fuel waste management is discussed.

Calculation of the movement of chemical fronts over long timescales could be important in underpinning performance assessments for radioactive waste disposal. A quasi-stationary state model, MARQUISS (Mineral Alteration Reactions using the QUasI-Statίonary State approximation), has been developed to achieve this objective by avoiding many of the problems encountered using more conventional approaches to coupled chemistry and transport calculations. MARQUISS simulates advective, dispersive and diffusive transport through a one-dimensional porous medium coupled with the chemical kinetics of mineral precipitation and dissolution. A description of its development and verification for simple systems is provided, together with its application in a study of the migration of mineral alteration zones at a natural analogue for a cementitious radioactive waste repository located at Maqarin in northern Jordan.

In order to clarify the effect of mineral alteration on nuclide migration, we examined the processes, mechanisms, and kinetics of chlorite weathering, and the uranium concentrations in minerals and rocks at Koongarra, Australia. The observed concentrations of uranium in rocks were compared to those calculated. The sequence of chlorite weathering may be simply expressed as a chlorite → vermiculite → kaolinite conversion. These minerals occur as a function of depth, which corresponds well to uranium concentrations on the meter scale. Iron minerals, closely related to the uranium redistribution, are released during the weathering. The first-order kinetic model of the weathering process suggests that the weathering rate is not constant but time-dependent. The uranium concentrations are qualitatively proportional to the extent of the weathering; the weathered part having higher uranium concentration. Uranium mainly occurs with iron minerals, and sub micron sized saléeite, a uranyl phosphate, is one of the most probable uranyl phases associated with the iron minerals. The uranium fixation mechanisms are probably saléeite microcrystal coprecipitation and sorption to the iron minerals. Our model, which describes uranium concentrations in rocks as a function of time, shows that the transition zone (a vermiculite dominant area) plays an important role in the uranium migration. We have established that weathering of chlorite has affected the redistribution of uranium for more than one million years. The present study demonstrates the significance of mineral alteration when we estimate nuclide migration for geologic time.

Secondary uranium minerals from the Koongarra deposit, Northern Territory of Australia, were examined in order to understand the formation and alteration processes of the uranium minerals and their relevance to the migration behavior of uranium, lead, calcium and rare earth elements in the weathered zone. In most of the secondary ore zone, the only stable uranium mineral was saléeite (Mg(UO2)2(PO4)2·10H20), occurring as euhedral platy crystals up to 1 mm in length in veins and at surfaces. Apatite (Ca5(PO4)3F), an accessory mineral of the host rock, has saléeite reaction rims, suggesting formation at the expense of apatite. Ca-uranyl phosphates, such as autunite (Ca(UO2)2(PO4)2·10H2O), were not identified, and Ca-rich uranyl silicates are also absent in the primary ore zone. Pb-bearing uranyl phosphates were found only in the graphite layer cross-cutting the secondary ore zone. In the graphite layer, the local low oxidation condition and high hydrocarbonate content of ground water have affected the formation of uranium minerals and the migration behavior of uranium.

Many reactions involving inorganic minerals at water-rock interfaces have now been recognized to be bacterially mediated; these reactions could have a significant effect in the excavation of vaults for toxic and radioactive waste disposal. To investigate the role that bacteria play in the natural aqueous environment of crystalline rock the microbial growth factors of nutrition, energy and environment are described. Microbial activity has been investigated in Atomic Energy of Canada's Underground Research Laboratory (URL), situated in the Archean granitic Lac du Bonnet Batholith, Winnipeg, Manitoba. Faults, initiated in the Early Proterozoic, and later-formed fractures, provide ground-water pathways. Planktonic bacteria, free-swimming in the groundwater, have been observed in over 100 underground borehole samples. The number of bacteria varied from 103 to 105 mL-1, and appeared to decrease with depth and with increased salinity of the water. However, in the natural environment of deep (100–500 m) crystalline rocks, where nutrition is limited, formation of biofilms by sessile bacteria is a successful survival strategy. Natural biofilms at the URL and biofilms grown in bioreactors have been studied. The biofilms can accumulate different elements, depending upon the local environment. Precipitates of iron have been found in all the biofilms studied, where they are either passively accumulated or utilized as an energy source. Within the biofilm active and extensive biogeochemical immobilization of dissolved elements is controlled by distinct bacterial activities which are sufficiently discrete for hematite and siderite to be precipitated in close proximity.

We conducted in situ experiments with simulated, i.e., nonradioactive, contact-handled trans-uranic (CH TRU) waste drums at the Waste Isolation Pilot Plant (WIPP) facility for a period of about four years. We performed these tests in two rooms in rock salt, at WIPP, with drums surrounded by crushed salt or 70 wt. % salt/30 wt. % bentonite clay backfills, or partially submerged in a NaCl brine pool. Air and brine temperatures were maintained at ~40°C. These full-scale (210-L drum) experiments provided in situ data on: backfill material moisture-sorption and physical properties in the presence of brine; waste container corrosion adequacy; and, migration of chemical tracers (nonradioactive actinide and fission product simulants) in the near-field vicinity, all as a function of time. Individual drums, backfill, and brine samples were removed periodically for laboratory evaluations. Waste container testing in the presence of brine and brine-moistened backfill materials served as a severe overtest of long-term conditions that could be anticipated in an actual salt waste repository. We also obtained relevant operational-test emplacement and retrieval experience. All test results are intended to support both the acceptance of actual TRU wastes at the WIPP and performance assessment data needs. We provide an overview and technical data summary focusing on the WIPP CH TRU environmental overtests involving ] 74 waste drums in the presence of backfill materials and the brine pool, with posttest laboratory materials analyses of backfill sorbed-moisture content, CH TRU drum corrosion, tracer migration, and associated test observations.

This paper presents three case studies of corrosion of waste drums at the Los Alamos National Laboratory (LANL). Corrosion was not anticipated by the waste generators, but occurred because of subtle chemical or physical mechanisms. In one case, drums of a cemented transuranic (TRU) sludge experienced general and pitting corrosion. In the second instance, a chemical from a commercial paint stripper migrated from its primary containment drums to chemically attack overpack drums made of mild carbon steel. In the third case, drums of mixed low level waste (MLLW) soil corroded drum packaging even though the waste appeared to be dry when it was placed in the drums.

These case studies are jointly discussed as “lessons learned” to enhance awareness of subtle mechanisms that can contribute to the corrosion of radioactive waste drums during interim storage.

A number of groundvater parameters have been studied at AECL’s Underground Research Laboratory (URL) in support of the Canadian Nuclear Fuel Waste Management Program. The concentration of microbes in groundvater is of interest as they may modify the transport of dissolved radionuclides. Preliminary results from an earlier study suggested that the microbe concentrations may be affected by the extent of borehole flushing prior to sampling. A study was therefore carried out in which packer-isolated intervals of two boreholes intersecting a fracture zone at 250-m depth in the URL were flushed and sampled on two occasions at various flow rates. High initial microbial concentrations (most likely due to leaching of nutrients from sample tubes) decreased rapidly as flushing progressed, suggesting enhanced microbial growth near the top of the borehole zone. Also, a tenfold increase in flow rate during flushing caused an increase in microbial concentrations in the groundwater of one of the boreholes, concurrent with an increase in total particle count. This suggests that particulate and biofilm material may be flushed out of the fracture zone at this particular location.

Exposure of carbon and HY-80 high strength low alloy (HSLA) steels to anaerobic Postgate cultures containing sulfate reducing bacteria leads to production of mackinawite, Fe9S8, and vivianite, Fe3(PO4)2.8H2O, with minor amounts of other minerals, the vivianite being sometimes accompanied by siderite, FeCO3. The vivianite and mackinawite persist on exposure to air or oxygenated water. Green rust is a significant alteration product on some steels. The implications of these findings are discussed in light of present and possible future nuclear waste containers.