In Europe, formations become more and more important as candidate geological formations (e.g. Boom Clay, Callovo-Oxfordian, Opalinus...) for the deep disposal of High-Level radioactive Waste (HLW). In demonstrating the suitability of a geological site for the disposal of radioactive waste, it is essential to consider the potential mobility of critical radionuclides through the relevant types. This project addresses the migration behaviour of radionuclides, identified as important for the long term safety (U, Se, Pu, Am), in a reducing environment, with special emphasis on the role of the Natural Organic Matter (NOM). In such reducing environments, the solubility limit is considered as the most important mechanism to lower the mobile concentration of these radionuclides and the speciation is most likely neutral or negatively charged so a low retardation is expected. However, the presence of NOM may jeopardise the expected low concentration and sorption: by solubility enhancement due to complexation/colloid formation with NOM or by influencing the sorption behaviour. The objective is to develop and demonstrate a conceptual model for the description of the migration of radionuclides in a reducing, NOM rich environment that can be implemented in performance assessment (PA) models. The main questions are: (i) does the NOM increase the radionuclide concentration in solution due to complex formation and (ii) what is the possible role of mobile NOM as radionuclide carrier? To answer these questions, following research strategy was adopted taking the Boom Clay Formation as a Case Study. Prior to any experimental work, speciation and solubility's for the considered radionuclides were calculated not accounting for NOM. The mechanisms determining the overall behaviour of the radionuclides in the environment are then studied in solution in presence of NOM (solubility, complexation) and with respect to the solid phase (retention by immobilisation, sorption). Modelling is used to interpret and derive interaction constants which can be implemented in a geochemical clay database. Understanding and quantifying the interaction mechanisms in the environment should allow to model and interpret migration experiments with mixtures of the radionuclides and 1 4 C-labelled NOM. Separately, methodologies were developed to identify the source of mobile organic matter. The obtained results are translated into conceptual models which can be used to evaluate the performance of Boom Clay as potential host rock Batch experiments revealed that the solubility of amorphous UO 2 is ∼10 - 8 mol dm - 3 with no effect of NOM complexation. However, it was evidenced that the presence of NOM facilitated the formation of uranium colloids upon dissolution of UO 2 . The formation of colloids, with molecular size between 2 nm and 0.45 μm, accounts for a total uranium concentration, three orders of magnitude higher than the solubility of the amorphous UO 2 . Uranium colloids were also found being dominant in leaching experiments of natural uranium from Boom Clay samples. These colloids, only evidenced in batch experiments, are unlikely to be mobile in the compact natural Boom Clay. Electromigration experiments showed that U(VI) reduces to U(IV) and precipitates. This U(IV) precipitated phase constantly releases positive or neutral charged U-species. Furthermore, classical migration experiments with U(IV) mixed with 1 4 C-labelled NOM showed that U(IV) migrates independently of the NOM through the Boom Clay. The migration of uranium in the Boom Clay is governed by strong retention due to precipitation (solubility limit) and sorption and is not enhanced by the mobile organic matter.

Pore water has been extracted from Boom Clay by mechanical squeezing. Clay cores were obtained from various boreholes, all drilled at the SCK.CEN domain (Mol, Belgium).In contrast to pore water collected from piezometers, high sulphate concentrations are measured in the squeezed pore water. The lowest sulphate concentrations (<60 mg/l) were measured in pore waters squeezed immediately after drilling. Higher sulphate concentrations were often measured in the pore water when the clay cores were preserved for some time (generally <500 mg/l SO42-, but sometimes up to 20,000 mg/l SO42-). Nevertheless, a relation between preservation time and sulphate content could not be retrieved. However, major ion concentrations were obviously correlated with the sulphate content in the squeezed waters. The observed evolution in chemical composition were explained by water-rock interactions considering the pyrite oxidation and the subsequent ion exchange and mineral dissolution reactions. (C) 2004 Elsevier Ltd. All rights reserved.

This paper presents an analytical model to describe pulse injection experiments. This model solves the advection-diffusion equation while taking into account back diffusion from the clay core to the inlet and from the outlet to the clay core. In most analytical models, back diffusion is neglected. For sufficiently high Péclet numbers, this is a good approximation. However, in experiments where the Péclet number is low, back diffusion is important and must be taken into account. An additional advantage of the present model is that both concentration and flux are conserved at the inlet and at the outlet of the clay core. This model is used to fit pulse injection experiments with iodide and tritiated water (HTO) in clay cores. The (new) model is required for fitting the experimental results since in clay layers advection is very slow leading to a low Péclet number. The experiments are performed on clay cores taken from different depths from the Boom Clay and the Ypres Clay layer under the site of the nuclear power plant of Doel (Belgium). The quality of all fits is excellent and the obtained parameter values are coherent. For HTO, the fitted value for the diffusion accessible porosity is consistent with measurements of the water content in Ypres Clay cores. In both types of clays, the apparent diffusion coefficient at zero flow is between 10(-10) and 2 x 10(-10) m(2)/s for iodide and between 2 x 10(-10) and 3 x 10(-10) m(2)/s for HTO. The dispersion length is in the order of 10(-3) m. The average value for the diffusion accessible porosity is between 0.35 and 0.4 for HTO and between 0.2 and 0.25 for iodide.

Natural uranium concentration has been measured in the pore water of Boom Clay extracted by three different techniques: (1) batch leaching of clay suspensions; (2) squeezing of clay cores; (3) collection by piezometers. The results show that the batch and the squeezing techniques extract higher uranium concentrations (up to three orders of magnitudes) than measured in the pore water collected by piezometers. Batch techniques (leaching and squeezing) tend to remobilise natural uranium from clay surfaces as colloids into aqueous phase. As a consequence, the batch distribution coefficient (K-d) may underestimate the retention capability of the clay. The in situ retention mechanisms for natural uranium may involve colloid filtration and/or dissolution of uranium-bearing minerals.

Summary Under the geochemical conditions prevailing in situ in the Boom Clay Formation (pH, Eh, …), calculations predict that U(OH)4 is the dominant uranium species present in the interstitial water and the concentration is solubility limited. However the boundary of the domain where the non solubility limited UO2(CO3)3 4− species dominates is very close. It is therefore of prime interest to know the correct speciation of uranium during the migration process. Electromigration was used as technique with the advantage that it can provide information on the speciation because the movement of the species towards the electrodes depends on its charge and speciation. Electromigration experiments have been performed with preconditioned 233UO2(CO3)3 4− sources, starting from the hypothesis that this species should migrate without retardation towards the anode. Despite relatively long electromigration times, sufficient to displace strong retarded tracers, no displacement of the migration profile towards the anode was observed. All 233U remained near the source position, but the electropherograms clearly showed the presence of species moving towards the cathode. This indicates the presence of neutral or positively charged uranium species. These electropherograms are interpreted as a change in uranium valence state: reduction of UO2(CO3)3 4− and precipitation of U(IV) oxy-hydroxides near the source position. The solubility limited species, U(OH)4(aq), are carried with the pore water towards the cathode. The electromigration experiments indicate, in support of the speciation calculations, that the dominant migrating U-species is probably the solubility limited U(OH)4.

The Boom Clay is studied as potential host rock for the deep disposal of radioactive waste. To fill the lack between information on radionuclide behaviour from the short term laboratory or in situ experiments and the longterm need for performance assessment, SCKCEN studies the Boom Clay (Rupelian) as natural analogue since 1996. The natural analogue study on Boom Clay focuses on the behaviour of REE, U, Th and their natural isotopes in the clay considered as chemical analogues to critical radionuclides. XRD mineralogical analyses and XRF, ICP-MS, INAA chemical analyses were performed on solid samples from the Mol-1 drilling core in the Boom Clay. The uranium content varies from 3 to 7 ppm in the solid fraction. The thorium content varies from 8 to 15 ppm. Multivariate statistical analyses allow to divide the Boom Clay in three groups according to their influence on the chemical analogue behaviour: (1) quartz, muscovite and feldspars have a trace element dilution effect; (2) clay minerals, such as chlorite, illite, smectite and intersratified illite-smectite, sorb thorium, caesium and LREE's; (3) carbonates, pyrite and organic matter concentrate HREE's and redox sensitive elements such as U, Cu and Ni. Radiochemical analyses (α- and γ-spectrometry methods) were performed on the same samples. Most of the daughter/parent activity ratios measured in bulk Boom Clay samples are equal to unity within the 2σ standard deviation uncertainties. One important exception is observed in the Double Band which is a silty layer in the Boom Clay, and therefore, expected to be a potential zone of higher permeability and pore water mobility. In this layer, the 2 3 0 Th/ 2 3 4 U and 2 2 6 Ra/ 2 3 0 Th activity ratios are significantly different from unity, indicating that radionuclide mobility/immobility has been occurred. Furthermore, radionuclide mobility in the surrounding aquifers is suggested by the daughter/parent activity ratios which are different from unity within the 2σ standard deviation uncertainties. Apart of the Double Band, the current distribution of the chemical analogues (U, Th and REE) is completely explained by their behaviour during deposition and early diagenetic processes. No recent geochemical process seems to have influenced their distribution. Future work concerns the improvement of the squeezing technique to extract pore water and the study of uranium and thorium distribution in various fractions obtained by mineralogical separations and chemical sequential leaching.

The thermochemical databases of the NEA (TDB) are used in support of the determination of solubility values for performance assessment (PA). PA uses solubility to define the concentration limit at which radionuclides migrate from the near-field to the far-field of the repository. Being different from a thermodynamically defined solubility, a concentration limit can be equal to the solubility of pure phases but can also be imposed by other processes, e.g. formation of colloids or solid solutions. This paper illustrates a way that the TDB is utilised to interpret some migration experiments through geochemical and solute transport modelling. As a result, some meaningful migration parameters are derived based on both experimental observations and the application of the TDB.

For the performance assessment study of a geological disposal of High Level Waste (HLW) in clayey formations, migration studies are essential. For low permeability soils (clays), classical diffusion studies take a very long time. In order to reduce the experimental time, we propose an electrical field as driving force to accelerate the migration of ionic species. This paper reports the assessment of the electromigration technique as a powerful new and fast technique for migration studies. The apparent molecular diffusion coefficient can be derived by two independent methods using the migration parameters obtained from an electromigration experiment, namely the apparent dispersion coefficient and the apparent convection velocity. First, it can be calculated from the velocity of the migrating species by the Einstein relation. But, corrections are necessary for electroosmotic flow. The apparent electroosmotic mobility is experimentally determined as 2.2·10−9 m2/Vs. Second, it can be calculated from the relation between the apparent dispersion coefficient and the total apparent convection velocity. But it is necessary to know the dispersion length of the medium. The dispersion length for Boom Clay is experimentally determined as 8·10−5 m. Because of the serious reduction in time, it becomes possible to run series of experiments at different electrical fields to obtain averaged values for the apparent molecular diffusion coefficient according to the two methods. Experiments at different electrical fields have another advantage: the intercept of the linear relationship between the total apparent convection velocity and the apparent dispersion coefficient gives the apparent molecular diffusion coefficient. The apparent molecular diffusion coefficients obtained for 85Sr, 131I and HTO are respectively 0.8·10−11, 15·10−11, and 24·10−11 m2/s. These values are confirmed by pure diffusion experiments. The excellent agreement with the apparent molecular diffusion coefficients obtained by classical diffusion tests clearly demonstrates the feasibility of the electromigration technique for the determination of diffusion coefficients.

We present an analytical model to describe impulse injection experiments. This model solves the advection-diffusion equation, taking into account back diffusion at the inlet and at the outlet of the sample. In most analytical models, back diffusion is neglected. This is a good approximation for a sufficiently high Peclet number. However, in experiments with a low Peclet number, back diffusion is important and must be taken into account. An additional advantage of the present model is that at the inlet and at the outlet, both concentration and flux are conserved.Our model leads to excellent fits of impulse injection experiments with iodide and tritium in Boom Clay cores, taken at different depths in the Boom Clay layer. Besides, all parameter values are consistent with one another and lie in the expected range. Averaged over all clay cores, we find a diffusion accessible porosity eta = 0.16 +/- 0.02 for iodide and eta = 0.37 +/- 0.03 for HTO. The hydrodynamic dispersion coefficient D is fitted reasonably well as a function of the apparent velocity V with a dispersion length alpha of the order of 10(-3) m. At zero flow, the diffusion coefficient is of the order 10(-10) m(2)/sec.

Performance assessment studies are expected to predict the enhancement of the migration of trivalent lanthanides and actinides due to their complexation with organic matter, which play a role as a transport agent [1]. Therefore, the mobility of the dissolved organic matter in the interstitial Boom Clay water is studied. For the first time, the mobile fraction present in the clay water is concentrated and labelled with a radioisotope to study the mobility of the organic matter in clay and the interaction of the mobile with the non-mobile. The isotopes tested as label are I-125 and C-14. The I-125 label proved to be unstable and hence discarded. The labelled organic matter is then diluted for migration experiments on Boom Clay cores under anaerobic conditions. The influence of the molecular size on its mobility is studied by the separation of the labelled organic matter in different size fractions.

The candidate host formation for the disposal of radioactive waste in Belgium is Boom Clay which may contain up to 4% organic matter (OM) [1], A limited fraction (less than 0.05%) of this OM is mobile. OM can complex radionuclides and so influence their migration. The migration behaviour of the OM itself has been extensively studied but to date such studies have used absorbancy measurements to quantify the OM. Unfortunately various problems accompany the use of absorbancy measurements. The particular problems may be overcome by using radiolabelled OM. Accordingly as a precursor to planned in situ migration experiments in Boom Clay (BC) using radiolabelled OM, stability studies on I and C labelled materials have been conducted. The I containing solutions were analysed using Gel Permeation Chromatography (GPC) and the C solutions using High Performance Size Exclusion Chromatography (HPSEC). Dissappointingly at the relevant pH of 8.5, even in the absence of the clay, the I label was found to be unstable. However the C labelled OM (C-BC-OM) was stable under the mild conditions employed in the test, so its stability was investigated in the presence of Boom Clay. The results were compared with that of C labelled humic acids (C-HA), treated similarly. Unexpectedly the C labelled material was found to be partially unstable in the presence of Boom Clay. However the instability has not hampered the laboratory column experiments and should not hamper the proposed in situ experiments with this material. allow the possible exchange between the mobile and immobile OM fractions to be distinguished neither is it possible to make a mass balance. Consequently it is now proposed to use radiotracers to monitor OM behaviour. Incorporation of a radioactive label into the OM will allow migration experiments to be performed in-situ without further disturbance of the clay core. Clearly there are two pre-requisites for any label. Firstly the label must not alter the characteristics of the OM and secondly the labelled material must be sufficiently stable to permit unambiguous interpretation of the results. Accordingly the stabilities of I and C labelled OM have been investigated under conditions similar to those known to exist in Boom Clay. For comparative purposes C labelled Aldrich humic acids (HA) have also been prepared and tested. The investigations and findings especially in relation to the planned in situ migration experiments are detailed below.

The candidate host formation for the disposal of radioactive waste in Belgium is Boom Clay which may contain up to 4% organic matter (OM) [1]. A limited fraction (less than 0.05%) of this OM is mobile. OM can complex radionuclides and so influence their migration. The migration behaviour of the OM itself has been extensively studied but to date such studies have used absorbancy measurements to quantify the OM. Unfortunately various problems accompany the use of absorbancy measurements. The particular problems may be overcome by using radiolabelled OM. Accordingly as a precursor to planned in situ migration experiments in Boom Clay (BC) using radiolabelled OM, stability studies on I-125 and C-14 labelled materials have been conducted. The I-125 containing solutions were analysed using Gel Permeation Chromatography (GPC) and the C-14 solutions using High Performance Size Exclusion Chromatography (HPSEC). Dissappointingly at the relevant pH of 8.5, even in the absence of the clay, the I-125 label was found to be unstable. However the C-14 labelled OM (C-14-BC-OM) was stable under the mild conditions employed in the test, so its stability was investigated in the presence of Boom Clay. The results were compared with that of C-14 labelled humic acids (C-14-HA), treated similarly. Unexpectedly the C-14 labelled material was found to be partially unstable in the presence of Boom Clay. However the instability has not hampered the laboratory column experiments and should not hamper the proposed in situ experiments with this material.

The long term corrosion rate of nuclear waste glass in a repository might be controlled by the steady-state diffusion of dissolved silica and would be enhanced by the sorption of silica onto clay minerals. Irreversible sorption and moderate retardation have been observed for dissolved silica in Boom Clay. Values of eta/R between 10 and 20 have been determined by means of four Flow-Through migration experiments, while K-d in the range 20 to 100 cm(3).g(-1) have been measured by batch sorption tests with Si-32 on fresh and slightly oxidized Boom Clay. As non specific interactions cannot explain the sorption of neutral Si(OH)(4), or of negative silicate species (Donnan exclusion), onto negatively charged clay minerals other mechanisms must be invoked: i.a., the chemisorption of dissolved silica Si(OH)(4) onto specific Lewis acid sites (Al3+, Fe3+) present at the clay surface. The suggested mechanism could be similar to this explaining the irreversible chemisorption of oxy-anions of weak acids (as phosphate, or berate) onto aluminum and iron hydroxides in soils. Ligand exchange of aqueous silica with a hydroxyl group of Al(OH)(3) may form a hydroxy-aluminosilicate (HAS) surface complex. Gibbsite layers accessible on the basal plane of kaolinite and on the edges of illite and smectite are possible sorption sites for the dissolved silica in Boom Clay. Moreover, hydrous ferric oxide produced by pyrite oxidation significantly increases the extent of silica sorption.

Migration studies are necessary for the performance assessment of a geological disposal of HLW. Classical diffusion experiments on Boom Clay take long time because of the excellent retention characteristics. To accelerate the migration of ionic species, an electrical field is used to reduce the experimental time drastically. The electromigration technique has also the advantage that the apparent molecular diffusion coefficient can be derived by two independent methods. First, it can be calculated from the dispersion coefficient, knowing the dispersion length and convection velocity The dispersion length for Boom Clay was experimentally determined as 8.5 10(-5) m. Secondly, from the Einstein relation between the velocity of the migrating species and the apparent molecular diffusion coefficient. But corrections are necessary for electro-osmotic flow. The electro-osmotic mobility was experimentally determined as 1.4 10(-9) m(2)/Vs. Both methods were used to determine the apparent molecular diffusion coefficient of Sr-85. Both methods result in an apparent molecular diffusion coefficient of 6.3-7.8 10(-12) m(2)/s which is validated by pure diffusion experiments. The feasibility of the technique is further demonstrated for Na-22, I-131 and I-ITO. The excellent agreement with the apparent molecular diffusion coefficients obtained by classical diffusion tests clearly demonstrates the feasibility of the electromigration technique for the determination of diffusion coefficients.

Performance assessment studies for the deep geological disposal of radioactive waste in the Boom clay indicate the importance of the radionuclides C and I. The migration properties of these radionuclides in Boom clay are studied by Flow-Through type diffusion experiments. In the diffusion tests a mixture of C labelled bicarbonate, tritiated water (HTO) and carrier free I sodium iodide (Nal) is used. The isotope I is used as an equivalent for I. The mixture allows for a good comparison of the migration behaviour of the three different species. The mean value of the diffusion constant for bicarbonate is 1.8X10, for iodide 4.2X10" and for tritiated water 6.2X10" cm· s. For the interpretation of the experiments we consider sorption (isotopic exchange), diffusion and first order chemical reaction. The results of the diffusion tests are consistent with the concept of the diffusion accessible porosity and proves the importance of this concept. The diffusion accessible porosity for both bicarbonate and iodide is 0.11. The results are important for the safety assessment of a radioactive waste repository in the clay formation.