[en] In Belgium, the possibility to dispose of high-level radioactive waste in clay formations is studied since 1976. In the PAGIS report, which was the first performance assessment of the disposal of vitrified high-level waste in a clay formation and which was published in 1988, the most important contributors to the total dose via a water well pathway were ²³⁷Np, ¹³⁵Cs and ⁹⁹Tc. Since 1988, several elements that strongly influence the calculated doses have evolved:?the inventory of long-lived mobile fission and activation products in vitrified high-level waste has been improved; the half-life of ⁷⁹Se has been re-estimated; substantial progress has been made in the determination of migration parameters of the main fission and activation products and actinides. In recent performance assessments, the actinides and ¹³⁵Cs do not significantly contribute to the total dose, as they remain confined in the host clay formation during several millions of years due to sorption on clay minerals. Consequently, the total dose resulting from the disposal of vitrified high-level waste or spent fuel is essentially due to releases of mobile fission and activation products. On the basis of recent waste inventory data and parameter values, the most important contributors to the total dose via a water well are: in the case of disposal of spent fuel: ⁷⁹Se, ¹²⁹I, ¹²⁶Sn, ³⁶Cl, and ⁹⁹Tc; in the case of disposal of vitrified HLW: ⁷⁹Se, ¹²⁶Sn, ³⁶Cl, ¹²⁹I, and ⁹⁹Tc. Important remaining uncertainties are the transfer factors of volatile fission and activation products into the vitrified waste during reprocessing and migration parameters of Se. (author)

[en] Especially for geological disposal options in clay, the safety of the repository relies chiefly on the performance of the host formation as the main barrier. Understandably, scenarios in which this clay barrier is somehow bypassed earn great concern in PA (Performance Assessment) studies. The Poor Sealing Scenario is one of those scenarios that have been recently studied by the PA section of the Waste and Disposal department in the framework of the Belgian programme on deep disposal of high-level radwaste in Boom Clay. This scenario hypothesises that at least one disposal gallery and an access shaft have been poorly sealed off, providing a preferential pathway for RNs (radionuclides). The scenario further assumes a severe climate change, which would invert the presently downward hydraulic gradient, such that the potential impact would be maximal. The main objective is assessing the contribution from two transport processes to the overall radionuclide migration from a spent fuel repository towards the Neogene aquifer. The processes considered are advective transport through the poorly sealed repository and diffusive transport through the host formation. In addition, we would like to identify the most influential parameters with respect to repository design and performance

[en] Document available in extended abstract form only. The demonstration of the long-term safety of a nuclear waste repository relies heavily on earth science models, integrated in a total repository system safety assessment. It is important to demonstrate the validity of those models so that various stakeholders can have confidence in the safety assessment outcome. The models are therefore subject to quality assurance procedures of which model validation, qualification and verification are essential elements. However, it is acknowledged that in the context of repository safety assessment, model validation is often limited because of two reasons: i) the extreme timescale covered by the assessment and ii) the use of natural barriers, for which complete characterisation is impossible. Nowadays, validation is considered a confidence building process, aimed at demonstrating that the model is consistent with the scientific understanding and that it adequately represents the considered phenomena and interactions relevant to the assessment case. In this study, the validity of model and parameters is demonstrated for non-retarded radionuclide transport in Boom Clay, currently considered by Ondraf/Niras as a potential host formation for geological disposal of radioactive waste in Belgium. Performance assessment studies so far have shown that in the integrated system of engineered and natural barriers, the Boom Clay is the dominant barrier contributing to long-term radiological safety. Indeed, the very low hydraulic conductivity and absence of significant hydraulic gradients make diffusion the dominant radionuclide transport mechanism and most radionuclides are even further retarded in their transport due to various retardation processes on clay particles. Between 1980 and 1983, the HADES underground research facility was constructed at a depth of 223 m in the Boom Clay layer underneath the nuclear site in Mol (Belgium), in order to assess the feasibility of repository construction in this plastic clay formation, and to do in situ testing. In 1988, the in situ migration test 'Concrete Plug 1' (CP1) was started to validate radionuclide migration behaviour in Boom Clay at large scale. The principle of this migration test consists of installing a piezometer tube equipped with filters at 1 m from each other, in which a known quantity of radioactive tracer is injected in a central filter and breakthroughs at the other filters on the same tube are measured. The tracer used is tritiated water (HTO), which is a small, neutral and non-sorbing radioactive tracer with a half-life of 12.3 years. This tracer is ideally suited for studying the transport characteristics inherent to the clay formation itself, and serves as a reference molecule for migration parameters of a whole range of solutes, both neutral and cationic. The CP1 test, now on-going for 24 years, gives well developed breakthrough curves up to a distance of three metres from the injection filter. Intermediate results with measurements up to one and two metres from the source filter were previously reported. The concentration measurements were compared to blind predictions using the MICOF code, which solves the advection-dispersion transport equation analytically in 3D. More recently, results were updated to include the tracer breakthrough at three metres distance from the injection filter. The model prediction is based on HTO diffusion properties obtained from centimetre-scale migration tests, and includes anisotropy and a small advective component. Indeed, although the permeability of the Boom Clay is very low advection can not be neglected due to drainage towards the open gallery infrastructure. Therefore, a Darcy velocity profile as function of the radial distance to the gallery was implemented, based on detailed modelling of the hydraulic evolution around the HADES URF. Including this refinement leads to an even further increase of the goodness-of-fit and could be regarded as an independent validation of the advective component of the transport model, although it is noted that the results are not very sensitive to the flow properties in this diffusion-dominant system. Overall, validation of the HTO transport model and parameters is achieved using input parameters obtained from centimetre-scale migration experiments performed in the laboratory. The results show that up-scaling to metre-scale of the solute transport problem does not invoke parameter uncertainty for as far as the composition of the clay samples remain representative for the considered thickness. The results further showed that this system is robust: a simplified base model with only diffusion and isotropic parameters in the direction of interest is already sufficient to give a good correspondence with the observed breakthrough curves. (authors)

[en] Document available in extended abstract form only. The Belgian agency for radioactive waste and enriched fissile materials Ondraf/Niras presently considers Boom Clay as a potential host formation for the disposal of high-level and long-lived radioactive waste. The production of gas is unavoidable within a geological repository. Gas is produced by different mechanisms: anaerobic corrosion of metals in waste and packaging, radiolysis of water and organic materials in the waste and engineered barriers and microbial degradation of various organic wastes. Corrosion and radiolysis yield mainly hydrogen while microbial degradation leads to methane and carbon dioxide. The gas generated in the near field of a geological repository will dissolve in the pore water and is transported away from the repository by diffusion as dissolved species. If the gas generation rate is larger than the diffusive flux, the pore water will become over-saturated and a free gas phase will form. Initially, isolated gas bubbles will accumulate until a continuous gas phase is formed. As gas pressure continues to increase, discrete gas pathways may be formed by tensile fractures within the rock fabric. Consequently, this entire process may locally and at least temporarily alter the hydraulic and mechanical properties of the engineered barriers and the clay and, perhaps, their performance. Therefore it is important to assess whether or not gas production rates might exceed the diffusive gas flux. The currently available gas diffusion parameters (D_eff: effective diffusion coefficient) for hydrogen in Boom Clay, obtained from the MEGAS project, and re-evaluated after lead to an estimated D_eff between 1.9 10^-12 and 1.5 10^-10 m²/s. Sensitivity calculations showed that this uncertainty on the diffusion coefficient, combined with that on the gas source term, made it impossible to exclude the formation of a free gas phase. To reduce the uncertainty, an experimental method was developed to determine more precisely the gas diffusion coefficient for dissolved gases in Boom Clay. The methodology was demonstrated for He and CH₄ and the estimated D_eff values were resp. 4.51 10^-10 and 0.90 10^-10 m²/s. The method is based on the counter-diffusion of two dissolved gasses under (quasi-)constant concentration gradients, yielding two gas diffusion coefficients in a single test. In practice a clay core is sealed in a stainless steel cell and connected at both sides with water vessels that are pressurized with 2 different gasses at the same pressure. Gas dissolves into the water and because of the gas concentration gradient, these dissolved gasses diffuse through the clay. The dissolved gas that diffuses into the other water vessel will equilibrate with the gas atmosphere and the changes in the gas composition is determined by gas chromatography. In the current experimental phase, we focussed first on reproducibility, then on the effect of the bedding planes in the clay sample to determine the anisotropy in the gas diffusion parameters. Furthermore we noticed that the size of the diffusing gas molecule leads to different formation factor values. The latter is further investigated by performing gas diffusion experiments with gasses of different size (He, Xe, CH₄, C₂H₆). Based on observed relationships D₀ vs. the gas kinetic diameter, and formation factor vs. gas kinetic diameter, we investigate the possibility to estimate D_eff for other gasses as H₂ and Xe. The tests are interpreted with a simple diffusive transport model. The model solves the diffusive transport equation in a 1D geometry and is based on the second law of Fick for diffusive transport in porous media. As output, fluxes at both faces are calculated with the first law of Fick, as well as concentration profiles at regular time intervals and concentration. All calculations are performed with COMSOL Multiphysics 3.5a, Earth science module. The diffusion coefficients are obtained by using a least squares fitting procedure to the experimental data with the MATLAB Optimization Toolbox. the reproducibility is very good as performing tests in another setup with a new cored clay sample provided very similar results. The anisotropy ratio for diffusion was determined using He and CH₄ on a clay core with a stratification parallel to the bedding plane, and for both He and CH₄ the same anisotropy ratio of 1.59 was obtained. Based on D_eff and D₀, we calculated the formation factor for He, CH₄ and C₂H₆. We noticed that this ratio is not constant and it increases with the size of the molecule. Also the D₀ decreases with the size of the molecule. If a straightforward relationship exists between these parameters and the gas molecule size, it could be used to interpolate D_eff values for other gases. When plotting D₀ vs kinetic diameter and the calculated formation factor values vs kinetic diameter, exponential relationships can be fitted. Comparing measured and calculated D_eff values show a good fit. With the obtained relationship, D_eff values for H₂ and Xe are estimated, resp. 3.24 10^-10 and 0.66 10^-10 m²/s. Currently, experiments with H₂ and Xe are on-going to verify this relationship, and also in the future we plan to perform tests with other gases having different sizes

[en] PRACLAY aims to demonstrate the suitability of the Boom Clay host rock, in terms of performance of the disposal, to undergo the thermal load induced by the vitrified HI. PRACLAY represents an important milestone for the Safety and Feasibility Cases 1 (2013) and II (2020). PRACLAY is developed to be design-independent to overcome possible future changes in the design. The temperature criterion is: The maximum temperature in clay-based backfill materials, used as engineered barriers, for heat producing radioactive waste, must be kept below 100 C. PRACLAY regroups a set of four experiments. The PRACLAY Crossing consists in the intersection of the connecting gallery and the PRACLAY gallery, and aims to demonstrate that it is possible to construct a crossing between an access gallery and a disposal gallery at an acceptable cost and limiting the perturbation in the host-rock. The PRACLAY Heater Test has to demonstrate that Boom Clay will behave as predicted under a thermal load. The PRACLAY Plug Test is aimed at demonstrating that it is possible to cut-off hydraulically the Excavation Damaged Zone (EDZ) and the engineered barriers of the disposal galleries with a horizontal plug. 4. The PRACLAY backfill test aims to test the installation and the performance of different types of backfill or buffer material (cement, slaked lime, bentonite, Boom Clay...) that could be considered in the design of disposal galleries. The paper will present the objectives, the preliminary model predictions, and, as a result of these, the design of the test, including the monitoring plan and the choices regarding the boundary and initial conditions. (authors)

[en] Document available in extended abstract form only. The main mechanisms by which gas will be generated in deep geological repositories are: anaerobic corrosion of metals in wastes and packaging; radiolysis of water and organic materials in the packages, and microbial degradation of various organic wastes. Corrosion and radiolysis yield mainly hydrogen while microbial degradation leads to methane and carbon dioxide. The gas generated in the near field of a geological repository in clay will dissolve in the ground water and be transported away from the repository by diffusion as dissolved species. However if the gas generation rate is larger than the diffusive flux, the pore water will get over-saturated and a free gas phase will be formed. This will lead to a gas pressure build-up and finally to an advective gas flux. The latter might influence the performance of the repository. Therefore it is important to assess whether or not gas production rates can exceed the capacity of the near field to store and dissipate these gases by dissolution and diffusion only. The current available gas diffusion parameters for hydrogen in Boom Clay, obtained from the MEGAS project, suffer from an uncertainty of 1 to 2 orders of magnitude. Sensitivity calculations performed by Weetjens et al. (2006) for the disposal of vitrified high-level waste showed that with this uncertainty on the diffusion coefficient, the formation of a free gas phase cannot be excluded. Furthermore, recent re-evaluations of the MEGAS experiments by Krooss (2008) and Aertsens (2008) showed that the applied technique does not allow precise determination of the diffusion coefficient. Therefore a new method was developed to determine more precisely the gas diffusion coefficient for dissolved gases (especially dissolved hydrogen) in Boom Clay. This should allow for a more realistic assessment of the gas flux evolution of a repository as function of the estimated gas generation rates. The basic idea is to perform a kind of a through diffusion test with dissolved gasses. A clay core is sealed in a stainless steel cell and connected at both sides with water vessels that are pressurized with 2 different gasses at the same pressure. In this way no advective gas flux can occur and the clay sample remains fully water saturated. The water at both sides is pumped around. Gas dissolves into the water and because of the gas concentration gradient, these dissolved gasses diffuse through the clay. The dissolved gas that diffuses into the other water vessel will equilibrate with the gas atmosphere and the changes in the gas composition can be determined through e.g. gas chromatography. As such, the diffusion parameters of two gasses can be determined in a single experiment. Tests will be conducted with the following gas combinations: He and Ar, He and CH₄, Ar and Hytec (5% H₂ in Ar). These tests are interpreted with a simple diffusive transport model. The model solves the diffusive transport equation in a 1D geometry. The porosity of the clay core is set to 30%. At one end, a fixed gas concentration was imposed as boundary condition, corresponding to the dissolved fraction at the prevalent pressure in the vessel following Henry's law. At the other end, the concentration of the considered gas is assumed to be 0. As model output, fluxes at both faces, as well as concentration profiles at regular time intervals are obtained. For both He and CH₄ the evolution of the measured gas concentrations in the downstream vessel for each gas fall within simulation curves using rather narrow ranges of the Dapp: 1.4 - 1.6 10^-10 m²/s for He and 2.5 - 3 10^-10 m²/s for CH₄. It is clear that the method is very sensitive and that quite accurate measurements of the diffusion coefficients can be obtained, far better than the previous results obtained in the MEGAS project. Currently an experiment to determine the diffusion coefficient for hydrogen is on-going. This method allows for an accurate determination of diffusion parameters of dissolved gasses in the Boom Clay and can be applied to other low permeability porous medium. (authors)

[en] For the assessment of the long-term safety of a geological disposal of high- and intermediate-level radioactive waste and/or spent fuel in the Boom Clay, a better understanding of the migration behaviour of Natural Organic Matter (NOM) is needed because it can act as a carrier molecule for radionuclides. Therefore, an in-situ migration experiment with ¹⁴C-labelled NOM was performed to study the NOM migration behaviour on a large scale (m), on the long-term (> 10 a) and in directions parallel and perpendicular to the bedding plane (transport anisotropy). The numerical modelling tool HYDRUS2D/3D was used to interpret the results. The model was built stepwise, testing the influence of advection, (non-)linear equilibrium sorption, colloid attachment/detachment and anisotropy. The up scaling of previously determined parameters from small scale lab tests was also tested. A classic diffusion-advection-retardation description, using parameter values in the range of those obtained in the lab tests, provided already reasonable results. Inclusion of a colloid filtration term in the model significantly improved the simulation. Finally, the model was successfully tested against a second dataset and the anisotropy of the Boom clay was brought into account. (orig.)

[en] ONDRAF/NIRAS is developing and evaluating a surface disposal concept for low and intermediate level short-lived radioactive waste (LILW-SL) at Dessel (Belgium)). In support of ONDRAF/NIRAS's assignment, SCK/CEN carried out long-term performance assessment calculations for the inorganic non-radioactive components that are present in LILW-SL. This paper summarizes the results obtained from calculations that were done for a heavily engineered surface disposal facility at the nuclear zone of Mol/Dessel. The calculations address the migration of chemo-toxic elements from the disposed waste to groundwater. Screening calculations were performed first to decide which non-radioactive components could potentially increase concentrations in groundwater to levels above the groundwater standards. On the basis of very conservative calculations, only 6 out of 41 chemical elements could not be classified as having a negligible impact on man and environment. For each of these six elements (B, Be, Cd, Pb, Sb, and Zn), the source term was characterized in terms of its chemical form (i.e., metal, oxide, or salt), and a macroscopic transport model built that would capture the small-scale dissolution processes relevant to element release from a cementitious waste container. Furthermore, reliable transport parameters in support of the convection dispersion-retardation (CDR) transport calculations were determined. This included derivation of (1) solubility for a cementitious near field environment based on thermodynamic equilibrium calculations with The Geo-chemist's Workbench, and (2) distribution coefficients based on a compilation of literature values. Scoping calculations illustrated the effects of transport parameter uncertainty on the rates at which inorganic components in LILW-SL leach to groundwater. (authors)

[en] The Eurobitum bituminised waste produced by the former EUROCHEMIC reprocessing plant in Mol-Dessel (Belgium) contains a lot of salts, mainly NaNO₃. This NaNO₃ may possibly affect significantly the evolution of a radwaste repository through redox and ion exchange reactions with Boom Clay, which in Belgium is studied as a reference host formation. Yet, the leaching behaviour of NaNO₃ from the Eurobitum and its migration in the Boom Clay is rather unclear. Therefore, some scoping numerical calculations have been performed. In a first phase, we calculated the water flux towards the disposal gallery to assess the timescale of the saturation process of the disposal galleries. In a second phase, the leaching rate of NaNO₃ out of the drums is compared with its diffusive removal through the Boom Clay. The water flux to the gallery was calculated at approximately 100 ml/day per metre gallery, based on the natural pressure gradient observed around the HADES Underground Research Laboratory. Different approaches were elaborated to find a NaNO₃ source term for the diffusion calculations. For various ways of characterising the nitrate release, the numerical calculations generally show that the NaNO₃ concentrations in the first decimetres of Boom Clay will not be much higher than 1 M. Moreover, a sensitivity analysis showed that if the nitrate is released from the bituminised waste product within 10 000 years, the Boom Clay controls the concentration profiles; if the release is slower, the source term has a more pronounced effect on the near field concentrations. (author)

[en] Document available in extended abstract form only. In the frame of the Belgian research program on long term management of high-level and/or long-lived radioactive wastes coordinated by ONDRAF/NIRAS, plastic clays (i.e., Boom Clay and Ypresian clays) are investigated for their potential to host a deep geological disposal repository for radioactive waste because of, among others, their ability to significantly retard radionuclide releases to the biosphere. The Boom Clay is characterised by the presence of a relatively high amount of dissolved organic matter (DOM, humic substances) which show a strong interaction with a suite of radionuclides (RN) like lanthanides, actinides and transition metals. This interaction with DOM leads in general to an increased mobility of the radionuclide as the OM can act as a colloidal carrier for the radionuclide. Therefore, the quantification and the understanding of the underlying processes are needed for the demonstration of confidence in the host formation to act as a suitable barrier. However, this is not an easy task, given the multitude of processes involved: complexation/colloid formation, sorption, kinetics, filtration, -. In this presentation, we will provide an overview of the research work that leads to a straightforward reactive transport model capturing fairly well the experimental observations. The model can be considered as an intermediate model providing a good basis for further safety abstraction on the one hand and the way to a more detailed phenomenological understanding on the other hand. The research is focussed on the underlying processes that govern speciation, sorption and transport. These underlying processes are investigated in a bottom-up approach, from simple systems to more complex systems. Interpretation is done using thermodynamic based models. Whereas the contribution of Bruggeman et al. focusses mainly on (batch) sorption studies (of trivalent RN), this presentation will provide more details on the conducted transport studies, performed both under controlled conditions in the lab and in in-situ environment. Transport experiments were conducted to study on one hand the behaviour of DOM itself using natural DOM as well as ¹⁴C- labelled DOM fractions of different sizes separated from Boom Clay pore water. Transport of DOM is investigated at large time- and spatial scale in an in-situ experiment in the HADES URL. These experiments enable us to obtain general migration parameters for DOM as well as some information on filtration processes. The behaviour of RN, from di- to pentavalent, with the affinity to from DOM complexes is investigated in long-term running column migration experiments either as single tracer or complexed with ¹⁴C- labelled DOM, in so-called double-tracer experiments. The use of two radionuclide labels allows the migration behaviour of both the DOM and the RN to be investigated. From these experiments we were able to obtain information on the kinetic dissociation behaviour that influences the transport of the RN. Based on these and other detailed studies, a consistent phenomenological model was put forward and tested which describes the transport behaviour of the RN-DOM linked species. The model considers the radionuclide to be transported as an organic matter complex/colloid that slowly dissociates, and both the RN-DOM and the RN sorb to the solid phase. It was observed that the transport of a suite of RN with different chemical behaviour, but which all show strong affinity to DOM, (tri-, tetra-, pentavalent La/Ac as well as some fission products) can be described by this model and even with parameter ranges that are quite narrow. This model shows several advantages as a first step towards PA abstraction use i) it is process based, ii) it is easy to implement without oversimplification (limited number of parameters which are determined or verified independently), iii) seems applicable for all RN that associate with DOM with a rather narrow range of input parameters. This conceptual model is an important step forward in the description of DOM linked RN transport in Boom Clay. (authors)