In the Spanish concept of a high level radioactive waste (HLRW) repository, the canister with the spent fuel would be surrounded by a layer of FEBEX bentonite (Ca-Mg smectite clay) compacted to a dry density of 1.65 g/cm3. Since the permeability of the compacted bentonite is very low, the main mechanism for radionuclide transport in the clay barrier is expected to be diffusion, retarded by the sorption on the solid. Since the mayor component of the radioactive waste is uranium it is very important to study its diffusion behaviour in the compacted clay.

Generally, effective (De) and apparent (Da) diffusion coefficients are obtained from "through-diffusion" and "in-diffusion" experiments, respectively. The obtention of both parameters in the same sample is a very interesting issue because the former includes the information on the porous medium structure (connectivity, constrictivity, tortuosity) and the latter takes into account the sorption on the solid phase. In this work, through-diffusion experiments were carried out for studying uranium diffusion and both effective and apparent diffusion coefficients were estimated, from the same experiment, by using different theoretical approaches.

Chemical behavior of selenium (Se), which was one of the important elements for performance assessment of geological disposal of high-level radioactive waste, was investigated under reducing and iron-containing conditions. A washing method for an iron diselenide (FeSe2(cr)) reagent with acidic and basic solutions was carried out for the purification of FeSe2(cr) reagent, which was considered to be a solubility limiting solid for Se under the geological disposal conditions. Furthermore, solubility of FeSe2(cr) was measured in alkaline solutions under reducing conditions, and thermodynamic data on equilibrium reactions between Se in solution and Se precipitate were obtained. It was found that the dependencies of solubility values on pH and redox potential (Eh: vs. standard hydrogen electrode) were best interpreted assuming that the solubility limiting solid was not FeSe2(cr) but Se(cr) and the aqueous species was SeO32- in the present experimental conditions. The equilibrium constant between Se(cr) and SeO32- at zero ionic strength was determined and compared with literature values. The chemical behavior of Se under geological disposal conditions was discussed.

In the proposed Swedish nuclear fuel repository, intended to be placed some 500 meters down in granitic bedrock, the aquatic conditions will be reducing. In order to investigate the plutonium dissolution properties under near repository conditions Pu(III) was prepared using H2 bubbling in the presence of a platinum wire. The plutonium oxidation state was checked by spectrophotometry and a blue Pu(OH)3 precipitate was produced using deoxygenated NaOH. During drying the precipitate turned brownish and was then put in 0.1 M NaCl in pressure vessels. X-ray powder diffraction (XPD) showed no crystallinity of the precipitate. Solubility experiments were performed in pressure vessels that have an insert of PEEK (polyether-etherketone) and two valves, one for gas application and one with a PEEK tube reaching down into the vessel content for sample withdrawal. A small piece of platinum wire was put in each vessel and 50 bar H2 was applied. The pH was adjusted to range from about 3 to 10. All handling was performed in an inert gas (N2) glove box with active oxygen removal yielding oxygen concentrations of <10 ppm. The results after about one year showed concentrations from 10−8 M to 10−5 M at pH 10 and pH 3 respectively.

Rock matrix diffusion is one of the possible mechanisms for radionuclide retardation in a deep geological high-level radioactive waste repository, and it is usually considered that radionuclides diffuse as solutes through the rock. Nonetheless, the potential effects that clay, from the bentonite barrier, may induce on the radionuclides migration should be taken into account. Furthermore, transport models generally assume that the whole mineral surface is accessible to transport, whereas transport is highly conditioned by the heterogeneous mineral distribution, since different minerals may act as preferential pathways, while others may present higher sorption capability. It is therefore necessary to determine the actual surface area accessible to transport.

The aim of the present work is the identification of the uranium preferential pathways to the granite, both in presence or absence of bentonite clay. Results showed that uranium as solute diffused in specific mineral areas, indicating that the actual surface area accessible to matrix diffusion, and/or sorption on the surface, is significantly lower than the whole mineral surface. By the other hand, the uranium in presence of the clay was randomly distributed on the surface, and penetrated into the granite mainly through “defects” (as fractures or grain boundaries); its migration being enhanced on specially fractured or disturbed areas.

The role of excavation damaged zone (EDZ) in a KBS-3 type repository for spent nuclear fuel depends significantly on the design of the engineered barrier systems such as tunnel backfill since one obvious function of the backfill is to prevent the growth of EDZ. Main results of EDZ studies based on using 14C-PMMA method carried out at Äspö Hard Rock Laboratory in Sweden, and at Research Tunnel at Olkiluoto in Finland by Posiva and SKB in co-operation are presented. The significance of the results is evaluated quantitatively and qualitatively. The EDZ around the deposition hole on migration of radionuclides diffusing out of a waste canister was found unlikely to be a significant migration route in the absence of large hydraulic gradients. According to the study the EDZ caused by state-of-the-art drill and blast excavation adjacent to walls and roof of deposition tunnels was found negligible, however the EDZ adjacent to deposition tunnel floor is evidently more extensive and more significant.

The potential effect of water-soluble degradation products (DPs) formed from cellulose-containing cemented radioactive waste on the behaviour of Pu(IV) and Am(III) in Alkaline-Plume-Affected (APA) and in unaffected Boom clay was assessed. After solubility measurements in APA Real Clay Water, sorption experiments of Pu(IV) and Am(III) onto APA Boom clay were performed in the presence or not of DPs or pure α-isosaccharinic acid (α-ISA). Moreover, the sorption of α-ISA onto the unaffected clay was studied.

The effects of organic cement additives (totally 7) on the sorption of Eu on cement and TiO2 were measured with variation of the concentration of the additive (0.001–10% of the water phase) and reaction time (up to 420 d). There was a significant reduction of the sorption of Eu on cement as well as on TiO2 after 24 h of exposure time in the presence of additives at levels that would be obtained in a cement/water system if the total load of the additive in the cement (up to 3%) would be released into the water phase. Sorption increased with time and would be expected to reach constant levels after 100–200 d. Storage of the additives at 60 and 120 °C, pH 12.5, for up to 110 d (in order to degrade them) did not lead to significantly altered sorption behaviour.

A two-dimensional approach to understanding the heterogeneity in retardation effect governing the solute transport was carried out. To express the simplified spatial distribution of retardation effect, this study prepared parallel flat boards packed with two kinds of solid materials, where these materials were not mixed, but arranged as two kinds of the layers in a direction parallel to the flow. The ratio of the width of the constituent layer to the total transport distance was assumed to be one of key parameters describing the solute release profile transported through the media. As a tracer, Eu3+ in HNO3 solution was used, and its breakthrough curve was monitored in the experiments. In order to determine the retardation coefficient of the each layer, the retardation coefficients was obtained from the separate column experiment. Moreover, the permeabilities of both layers were confirmed to be almost identical from the one-dimensional, column experiments. A mathematical model was constructed to express the two-dimensional advection and dispersion of the solute through the media with the distribution of retardation effect. The proposed mathematical model and the experimental results showed good agreements, and indicated that the solute release profile strongly depended on the ratio of the width of the constituent layers to the total transport distance and the standard deviation of the retardation.

Colloids have the potential to enhance transport of contaminants in the subsurface. Kersting et. al.,[1] have shown that plutonium, as well as europium, cobalt and cesium, were associated with groundwater colloids at the Nevada Test Site. 240Pu/239Pu isotopic evidence confirmed that the plutonium was from the Benham nuclear test located 1.3 km away. This observation has lead to an increased concern that colloids may be potential vectors for radionuclide transport from an underground geologic repository [1].

The Anaylsis/Model Reports (AMRs) describe the characteristics of colloids for sparingly soluble radionuclides. The types of colloids formed, stability of the colloids as a function of pH and ionic strength, and the attachment of radionuclides to the colloids are modeled by the AMRs. The AMRs attempt to present bounding and conservative estimates for key parameters; however, waste form corrosion data, natural system studies, and experimental data do not always support the assumptions made in the AMRs. Most specifically, all colloids at Yucca Mountain, whether generated by the dissolution of waste glass or spent nuclear fuel or occurring naturally in groundwater, are assumed to be either smectite or iron-oxy-hydroxides. Data from waste form corrosion experiments do not support this assumption: numerous other phases have been identified including weeksite, clay particles with brockite inclusions, schoepite, soddyite, uranophane, boltwoodite, calcite, dolomite, Na-silicates, Al-oxides, absolone, birnessite, analcime, hydrated silicates, and apatite. Furthermore, the response of many of these phases to changes in pH or ionic strength is not known. If fundamental data for these phases, such as the point of zero charge (PZC), is different than that of either smectite or iron oxy-hydroxides, the TSPA may not accurately predict colloid transport of radionuclides.

We have analyzed colloid samples from groundwater at the Nevada Test Site using high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) and energy dispersive spectroscopic (EDS) elemental mapping, and found that light rare-earth elements (LREE) and Th are associated with monazite. Cesium and uranium are associated with an unidentified phase that is rich in Si and Fe. In addition, Co-60, a fissiogenic element, is associated with a complex metal colloid that includes Fe, Ni, Mo, and Cr. None of these phases are accounted for in the AMRs and their response to changing geochemical conditions is unknown. In addition, the PZC of several Mn-oxide phases was determined and was found to extend beyond the range of PZCs accounted for in the AMRs by smectite and iron oxy-hydroxides. Pu has been shown to preferentially adsorb onto Mn-oxides in the presence of Fe-oxides [2, 3], and the AMRs may not accurately predict Mn-oxide colloid stability and subsequent transport of radionuclides of interest.

We have studied the surface interaction between tetravalent plutonium and two different steel corrosion products: magnetite, which is the final product of the anoxic corrosion of steel, and goethite, which exemplifies a further oxidation of the steel surface due to the presence of active oxidants in the system. The pH of the experiments has been varied from 2 to 10 and the ionic strength from 0.001 to 0.1 M NaClO4. All the sorption experiments were carried out under N2 atmosphere. No significant effect of ionic strength was observed under the conditions studied.

The pHpcz of the solids (7.78 for goethite and 6.95 for magnetite) was determined by modelling potentiometric titration data.

The results of the experiments show that the sorption edge of plutonium occurs between pH 3 and 4 when using goethite as a sorbing surface and between pH 4 and 5 when magnetite is used.

We have modelled the sorption data by using a simple surface complexation approach with no electrostatic term. The model used involves a reduction process of Pu(IV) to Pu(III) in the presence of magnetite, which can be explained by the interaction of the actinide with the ferrous iron present in the solid. In the case of the experiments conducted with goethite, this reduction process is not possible and, therefore, in the model we have included the sorption of tetravalent Pu.

Within the scope of a bilateral cooperation between Svensk Kärnbränslehantering AB (SKB) and Forschungszentrum Karlsruhe, Institut für Nukleare Entsorgung (FZK-INE), actinide migration experiments with Pu, Am, and Np are conducted at the Äspö Hard Rock Laboratory. Migration experiments are complemented by batch experiments providing detailed information on the relevant retention processes for actinides onto granite and altered fracture material. The same kinds of experiments are performed with well defined minerals. α-autoradiography and XPS were used to quantify local sorption properties of the rock samples. It is shown that Np is retained by reduction to Np(IV) in the presence of Fe(II) minerals. The sorption coefficient for Pu is significantly higher compared to Np or U. Pu retention takes place on a multitude of minerals. Sorption of U is also strongly correlated with the occurrence of Fe oxide phases.

The effects of key geochemical parameters on Kd values for radionuclides in the host rock (pumice, sandstone) of a LLW repository were elucidated through a sensitivity analysis, using a thermodynamic speciation/sorption model for the elements Sr and Ni. The complex mineral assemblage of the rock was approximated by a component-additivity approach. Using published ion exchange and surface complexation parameters, Kd for both Sr and Ni could be well explained by the same model mineralogy and surface chemistry. Model results suggest that pCO2 can have a significant effect on Kd, and that a correct approximation of groundwater chemistry is a critical component of sorption modeling.

Interaction of Cs on commercial magnetite has been studied under different experimental conditions at high solid/liquid ratios. At pH 8.2 the Freundlich isotherm was able to describe an important range of the experimental data and at high Cs concentration surface precipitation was found likely to occur. Special attention has been given to silica, present in the system as an impurity of magnetite, since it plays an important role on the sorption mechanism in which a ternary complex magnetite-Si-Cs is formed.

Sorption coefficients are traditionally obtained in batch experiments if the sorbent is strongly or intermediately sorbing. In a batch experiment the rock is crushed and this could increase the surface area as well as induce new and fresh surfaces. Therefore there is some concern whether sorption coefficients obtained in batch experiments represents those of intact rock. Performing sorption experiment by diffusion in intact rock with intermediately and strongly sorbing species seems impossible in practice due to extremely long experimental times. In this paper the possibility of increasing the migration rate in the rock by two of three orders, thus enabling KD measurements of intermediately sorbing species in intact rock, is discussed. The increase in migration ratehas already been achieved successfully in so called through electromigration experiments using non-sorbing species. Here a potential gradient acts as the main driving force. In our experiments the migration rate was increased 320 times by using a potential drop of only 9 volts.

In performance assessments for nuclear wastes repositories, Kd values are often used to describe sorption of radionuclides on fracture surfaces. The weakness of the Kd concept is that values determined under certain conditions can only be used when the chemistry is similar to those. Alternatives include surface complexation models. The problem with such models is that the parameters required are not easily determined, and thus are missing for several systems.

Another alternative is to use surface phase thermodynamics. Such a theory has been developed, in which the surface of a solid phase and the species adsorbed upon it are treated as a separate phase of variable composition. The composition is described in terms of pure end members. For each end member, the free enthalpy is described by a series expansion. To calculate coefficients, the theory makes use of the fact that at equilibrium, the chemical potential of a component has the same value in all phases. Thus, speciation of the bulk aqueous phase can be used. Making use of only the linear term in the expansion, one measured Kd value is enough to determine the corresponding coefficient.

The theory has been applied to sorption of U(VI) on alumina with the aid of data from the literature. The coefficient was determined from the measured Kd values in the interval pH = 4 – 12. The resulting model curve was found to agree reasonably well with observed values, although the Kd varies by a factor of 200.

GASNET is a “thematic network” project supported by the European Commission. Its objectives are to evaluate the present treatment in safety assessments of issues arising from the generation of gas in deep geological repositories for nuclear waste, and to improve the translation of scientific information on these issues into safety arguments for such repositories.

An overview is provided of the work that has been undertaken within the project and of the preliminary observations and conclusions that have been assembled.

Use of compacted bentonite buffers is frequently planned in radioactive waste repositories to isolate waste canisters from the geological environment. When gas is generated from the wastes or their containers, it will need to migrate through the bentonite if pressure rises local to the canisters are to be avoided. Assessment of the effect of gas produced from waste canisters in bentonite buffers requires both experimental data on and models of gas migration through initially water-saturated bentonite. Several different approaches to modelling gas migration through bentonite are described. Some example comparison of model results with experimental data are provided, and some general discussion about the modelling approaches is offered.

This paper describes a long-term laboratory test designed to examine the sensitivity of gas flow in Mx80 buffer bentonite subject to a constant volume boundary condition. A constant volume and radial flow (CVRF) apparatus was designed to enable gas flow from a centrally located injection filter to be independently monitored at three sink-filter arrays mounted around the circumference of the clay specimen. Axial and radial total stresses and internal porewater pressure were continuously monitored. Gas entry, breakthrough and peak gas pressures were found to be systematically higher under constant volume boundary conditions than under previously reported constant stress and radially-constrained test conditions [6, 9, 10]. The observation that gas pressures are sensitive to test boundary conditions supports the hypothesis that gas entry is accompanied by dilation of the bentonite fabric. Gas penetration of the clay caused a substantial increase in total stress and internal porewater pressure. Abrupt drops in gas pressure, accompanied by similar drops in total stress, were interpreted as fracture propagation events. The outflow of gas was always non-uniformly distributed between the sinks. Furthermore, the distribution of flow between sinks often changed abruptly during the course of an experiment indicating that gas pathways were very unstable. When gas injection stopped, the gas pressure and rate of outflow spontaneously declined with time. Under constant volume conditions, the gas pressure at the asymptote exceeded the internal porewater pressure by an amount equal to the capillary pressure. In constant volume tests on clay with high water saturation, capillary pressure has a value close to the measured swelling pressure of the clay.

To determine the initial rates and effects of silica in solution on the dissolution kinetics of smectite, short- and long-term batch experiments (0.5 hour to 30 days) were completed at three temperatures (T = 25, 50, and 75°C) using stock solutions pH adjusted by NaOH (pH = 12, 13, and 13.5) with varying initial silica concentrations (0, 30, 60, and 100 ppm). The following important characteristics were observed at pH = 12: (1) The concentrations of Al, Si, Mg, Fe, and Ti in solutions increase rapidly during the first ∼2 hours and reach steady state (equilibrium) within ∼5 days. (2) The concentration ratios of Al, Si, Fe, Mg, and Ti in solutions during the early (<2 hours) reaction phase differ significantly from those of smectite, indicating initial dissolution proceeds non-stoichiometrically; Al dissolves much faster than Si, Mg, Fe, and Ti. (3) Further dissolution of smectite proceeds nearly stoichiometrically, including Fe and Ti. (4) The high solubility of Ti in highly alkaline solutions may be due to the formation of aqueous complexes, such as TiO(OH)3− and TiO2(OH)22−, similar to aqueous silica species. (5) The initial rate of smectite dissolution increases with increasing pH, T, and initial silica content of solution. (6) The silica in solution acts as a promoter and a catalyst, rather than an inhibitor, of smectite dissolution in high-alkaline solutions. This role is easily recognizable when the solubility of smectite and amorphous silica are very high, i.e., at pH >∼9.

The analyses of layer structure in Na-montmorillonite have been performed using 27Al MAS and 27Al Multi Quantum (MQ) MAS NMR techniques. In 27Al 5QMAS NMR spectra measured at higher magnetic field of 16.4 T, two peaks were observed at 74.2 and 62.9 ppm. Hence, it has been suggested that the 4-coordinated Al site in Na-montmorillonite has two different structures. For detail assignment of the two peaks due to such 4-coordinated Al sites in Na-montmorillonite, 27Al MAS NMR measurements have been carried out for Na-montmorillonites with different water contents. The ratio of two 4-coordinated Al sites was found to be affected by water in the interlayer space of Na-montmorillonite, because the area ratio of peaks corresponding to two 4-coordinated Al sites changes with the water content, i.e., the peak area at around 69 ppm is increased sensitively with an increase in water content. Moreover, Cross Polarization/MAS (CP/MAS) NMR measurements have been carried out. As a result, the enhancement of peak at around 69 ppm was observed. Therefore, the two peaks observed by 27Al 5QMAS methods are assigned to AlO4 in the vicinity of water bound to tetrahedral sheets and the remote AlO4 from such water, respectively.