Argillaceous rocks are considered as potential host rocks for radioactive waste repositories. The design of an underground disposal often includes large amounts of concrete which are proposed as waste matrix, backfill material and structural components. The leaching of concrete by groundwater produces alkaline solutions that will react with the host rock inducing modifications of the geological barrier performances. In order to examine the effects of an alkaline plume on the properties of claystone, an experimental program is being developed in our institute involving batch, diffusion and percolation experiments. This paper presents the results obtained with first batch studies performed at 25°C and 70°C over 3 months. The effect of high pH solutions (corresponding to the early period of leaching) on the composition of a claystone (originating from Tournemire - France) was examined. The changes in solutions composition were monitored as a function of temperature and V/m ratio. In spite of the short contact time between solution and solid, some modifications in these two phases were noticed. The interactions between solids and solutions were more pronounced at 70°C than at 25°C, resulting in pH and potassium concentration decrease, while the sodium concentration remained constant. Moreover, the dissolution of mineral phases such as dolomite and kaolinite was proposed.

The reactive transport model HYTEC was used to simulate the migration over 100,000 years of cesium, americium and uranium released from spent fuel packages in the near-field components of an underground stiff clay disposal site. A global equilibrium thermodynamic approach including kinetic control of the spent fuel pellets was used with instantaneous release fractions and congruent dissolutions of the rim and the core zones. A failure scenario of the waste package after 10,000 years was considered with magnetite as the main corrosion product. The retention properties of magnetite and the different effects of bentonite and cementitious backfill materials were specifically analysed.

In this paper, we are concerned with a specific scenario where a large fracture intersects, at its center, a canister that contains spent nuclear fuel. Assuming that a nuclide is free to release from the canister into groundwater flowing through the fracture, a detailed formulation of the volumetric flow rate and the equivalent flow rate are made for the parallel plate model. The formulas proposed have been validated by numerical examinations; they are not only simple in forms but also universal in applications where the flow may be taken normal, inclined or parallel to the axis of the canister. Of great importance, they provide a convenient way to predict the average properties of fluid flow and solute transport through a single fracture with spatially variable apertures.

Models of groundwater flow and transport are routinely used as part of the safety assessment of a site, and increasingly to aid the site investigation strategy and design of a proposed repository. This emphasis puts greater demands on those models to reflect a higher level of realism to incorporate additional information in the vicinity of the repository. This paper describes a successful approach to incorporate these different scales, from canister- to regional-scale, within one model. Such an approach is able to resolve the details of groundwater flow around the waste packages in the overall hydrological setting of the site.

Chemical retention of radionuclides in the near field and in the host rock is an important safety pillar within the Swiss concept for a radioactive waste repository. Geochemical databases for solubility limits in cement and bentonite porewater and sorption / diffusion in cement, compacted bentonite and Opalinus Clay were compiled for use in safety assessment calculations. The databases are presented for selected safety-relevant radioelements. In contrast to earlier safety assessment studies Kd values were not truncated at 5 m3 kg−1 and values well above 10 m3 kg−1 (e.g. for actinides) are reported.

The interpretation of geochemical and isotopic data as indicators of past hydrodynamic stability of groundwaters in fractured rock is discussed. Knowledge of the conditions and timing of past groundwater stability supports scenarios for safety assessments. Stratification of water masses according to densities is an important factor affecting stability. Chloride variations and concentration gradients are therefore diagnostic of mixing and the duration of stable conditions. Further evidence of timescales is provided by carbon-14 and tritium data although sampling difficulties and other uncertainties may allow only qualitative interpretations. Data from Äspö and other Swedish sites, Sellafield (UK) and Olkiluoto (Finland) are used to illustrate the approach. The palaeohydrogeology of such sites has been strongly influenced by distributions of saline waters. Scenarios for groundwater stability should consider these as well as glacial melt waters.

The need to decode the complex groundwater information in terms of origin, mixing (transport) and reactions at site scale, necessitated the development of a new modelling tool. This new modelling concept was named M3 (Multivariate Mixing and Mass-balance calculations). Initially, the method quantifies the contribution from the flow system. Subsequently, contributions from reactions are calculated. The M3 code has been used for the following types of modelling: to calculate the mixing portions at different sites and to quantify the contribution from inorganic and organic reactions. For instance, the groundwaters at many sites have been calculated to consist of complex mixtures of different water types such as meteoric water, biogenetically modified water, modern and ancient sea water, glacial meltwater and brine water. Examples from modelling of data from Äspö in Sweden, Oklo in Gabon and the URL in Canada are given.

In this work we have developed a model for the release of radionuclides from the spent fuel coupled with their transport through the near field. A compartmental approach has been used, as this methodology is well suited to model integrated systems. Several processes have been taken into account: oxidative dissolution of the spent fuel matrix, radioactive decay and chains, diffusive and advective transport, retardation by sorption and secondary phase precipitation. Results illustrate the complex evolution of the radionuclide concentrations in the gap and the near field. Hence, the main conclusion from this study is the requirement to model this coupled system using a compartmental integrated approach.

A new and simple approach for modelling solute transport in fractured rock is presented, based on our previously used Channel Network Model (CNM). In our underlying concept, solute transport in fractured rocks takes place through undulating flat channels in the rock fractures. The channels form a 3-dimensional network. For non-interacting solutes the volume of the channels determines the residence time of a stream of water. For sorbing radionuclides the residence time is mainly determined by the ratio between the Flow Wetted Surface (FWS) and the flow rate in the channel (Q), in addition to the parameters that determine the interaction between the rock and the radionuclides. The main question explored in this paper is whether it is necessary to solve the flow problem for each studied case. From previous investigations, we have found that the flow rates in the channels where the solutes travel are closely related to the transmissivities of the channels. Simulations are presented in which streams of water are allowed to randomly select channels and trace out the paths where the path volume and its FWS can be assessed. The results obtained in this manner for the RTD (Residence Time Distribution) show rather good agreement with those obtained by solving the full flow problem and tracing particles in the network.

Transport of radionuclides from a repository to the biosphere is calculated in the presence of varying boundary conditions. This may occur in Sweden, for example, during post-glacial rebound. To calculate the transport of radionuclides, the flow field has to be calculated at each time step. An alternative approach is to determine the flow field at certain time intervals; i.e., modelling the dynamic flow system as a sequence of punctuated steady states. This can be done in a rather simple way by using the Channel Network model (CNM), in which radionuclide transport is calculated by particle tracking. Intuitively one would expect such a simple procedure to be permissible and lead to small errors if the time intervals are so short that only small changes in flow rates and flow directions take place. In this paper, the technique is applied to a study case. This describes a repository located at a sub sea level location outside the coastline; where, owing to the land lift that is taking place in Scandinavia the coastline is advancing towards the repository. An important issue to be determined is the minimum number of time intervals that need to be used in order to obtain a reliable solution to radionuclide transport. It is found that the number of time intervals needed is strongly dependent on the variation in the boundary conditions. For the changing conditions used in the example tested, a number on the order of 5–9 time intervals allows a sufficiently good representation of the transport.

The use of nuclear reactors produces electrical power but also high-level radioactive waste (HLW). In Sweden it has been decided that the HLW shall be disposed of in a deep geological repository 500 meters underground. The HLW will be encapsulated in copper canisters, surrounded by bentonite and then the access tunnel will be backfilled with a bentonite and crushed rock mixture.

We have investigated activity of sulphate reducing bacteria (SRB) in compacted bentonite with densities of 1.5, 1.8 and 2.0 g·cm−3 using stainless steel oedometers with 35SO42- as the main sulphur source for the SRB. The investigations were performed under in situ conditions at a depth of 450 m at the Äspö Hard Rock Laboratory. Canister type copper plates were placed in the oedometers, with the top either in contact with the added groundwater or with 3–4 mm bentonite between the plate and the added groundwater. Radioactive hydrogen sulphide formed by the SRB reacted with the copper and formed radioactive copper sulphide, which was measured by electronic radiography using an Instant Image Micro autoradiography instrument (Packard). A similar set-up with filter-sterilized groundwater (0.2 μm) was used as control. A second experiment was set up as the first experiment but using only filtered groundwater and adding heat treatment of the bentonite in the control. Here, we show that SRB were active and produced hydrogen sulphide during the initial phase of bentonite swelling. The activity of the SRB was inversely correlated with the final density of the bentonite as the highest density had the lowest overall SRB activity. We also found that SRB were present in a dormant state in the commercial MX-80 bentonite. By addition of water, these dormant SRB became active and started to produce hydrogen sulphide.

Hydrodynamic modelling on the site- and repository scale was aimed at testing the performance of both, geological and engineered barriers as part of the Swiss program for disposal of high- and intermediate-level radioactive waste. Conceptual and parameter uncertainties were addressed by sensitivity studies to explore the robustness of the overall disposal concept. Consistency was demonstrated between modelling results and hydrogeological and geochemical evidence from the Benken borehole, increasing confidence in the comprehension of the key hydrogeological features of the candidate repository site. Assessment of a wide spectrum of repository closure conditions highlighted the function of the waste confinement concept. Visualisation and animation of the modelling results has proven to be of great help in communicating the level of comprehension of groundwater movement in the vicinity of a potential repository in the Zürcher Weinland.

A model has been developed to predict the impact of microbiological processes on the long-term corrosion behaviour of copper containers in a deep geologic repository. The model accounts for a range of aerobic and anaerobic microbial processes. Various factors expected to limit the extent of microbial activity in the repository, such as the lack of water, evolving redox conditions, and the nutrient-poor environment, are taken into account in the model. Amongst other effects, the model predicts that microbial activity will not occur close to the container in the presence of highly compacted bentonite buffer material.

Biofilms were grown in situ 296 metres below sea level in the Äspö Hard Rock Laboratory. The prominent organism in these biofilms was Gallionella ferruginea, which is an iron oxidising chemolithotrophic microorganism that grows in low oxygen conditions. This organism grows an organic stalk structure capable of binding and concentrating trace metals. This stalk structure also allows amorphous ferric iron oxyhydroxides, or BIOS, (bacteriogenic iron oxides) to precipitate. The pH of the groundwater within the system was between 7.4 and 7.6, with Eh potential between 150 and 190 mV and oxygen saturation between 3 and 15%. Biofilms developed within two weeks and were sampled every two weeks for three months. Cell number and stalk length was recorded for each sample. The concentration of Cr, Ni, Cu, Zn, Mo, REE (rare earth elements), U-238 and Th-232 was measured by ICP-MS. Early results suggested that Gallionella biofilms and associated BIOS could potentially concentrate trace metals up to 1000 fold higher than levels within the host rock and over 1 000 000 times the levels in the groundwater over a period of years. These new experiments indicate that Gallionella biofilms and BIOS can rapidly attenuate metals to levels over 1000 fold higher than the levels in the groundwater. This process can occur anywhere where reduced groundwater enters the waste repository tunnel, open cavities or where groundwater extrudes at the surface. Because of this, there is huge potential to use biofilms and BIOS for retention of radionuclides and pollution control.

Microbial processes can influence retention of radionuclides directly or indirectly in several different ways. Many microorganisms produce various kinds of ligands or chelating compounds to increase the bioavailability of essential elements needed for metabolism. These ligands are not always highly specific, and several of them will also mobilise other elements such as heavy metals and radionuclides. Three bacterial species (Shewanella putrefaciens, Pseudomonas fluorescens and Pseudomonas stutzerii), isolated from the deep subsurface, and four radionuclides, 59Fe(III), 147Pm(III), 234Th(IV) and 241Am(III), were selected for this study. The microbes were cultured in the laboratory, separated from their exudates by centrifugation, and the supernatants were collected. The supernatants were mixed with radionuclide and solid phase (TiO2 or SiO2). The pH ranged from 7.5–9.0. All three bacterial species produced ligands that were able to complex up to 90% of the radionuclides in competition with the solid phases. High performance liquid chromatography analysis detected four Fe-complexing substances in the supernatant from P. fluorescens and two peaks and one peak, respectively, from P. stutzerii and S. putrefaciens. All substances eluted from the column varied in retention times, indicating that the microbes studied produced several metabolites that have different chelating abilities.

There are several oxygen consuming processes occurring in deep subsurface fractured rocks. The importance of each process depends on the prevailing conditions and what function it has in the overall oxygen scavenging process. In the present work, some of these processes are studied. The aim is to build a frame of maximum rates for each parallel process. Therefore, several rate-limiting factors are not included. Rate expressions were obtained from the literature and compared for different cases to address under which conditions each process is important. Some of the processes may be considered parallel and independent while others occur in series. The main reactions accounted for are abiotic and biotic pyrite and ferrous iron oxidation and the non-oxidative dissolution of matrix ferrous minerals. In addition, transport processes of oxygen and other substrates within the rock matrix are considered. The main conclusions are that initially, when the diffusion resistance in the matrix is nonexistent or small, reaction kinetics are important. However, in a longer time perspective, diffusion processes limit the oxygen scavenging process.

Experimental and Modelling studies have been used to investigate the biogeochemical processes occurring at the interface zone between the near-field and far-field of the Drigg Low-Level radioactive Waste (LLW) trenches. These have led to a conceptual model of interface biogeochemistry, which has subsequently been modelled by the BNFL code known as the Generalised Repository Model (GRM). GRM simulations suggest that as organic rich leachate migrates into the far-field, iron III minerals such as iron hydroxide will dissolve, and reduced iron minerals such as siderite will precipitate due to microbial processes. Batch and column experimental studies have verified this conceptual understanding. Some unexpected processes, such as the presence of nitrate enhancing sulphate reduction, were also observed

Experiments were conducted to identify the rock-water and microbial interactions influencing accelerated smectite-clay formation. Packed columns and stirred batch reactors contained Äspö granodiorite, artificial groundwater mimicking that from Äspö and combinations of three types of subsurface chemolithotrophic bacteria, two of which were indigenous to the Äspö rocks. Results showed evidence that, within 5 days under anaerobic reducing conditions, all three of the bacterial types produced copious biofilamentous ‘meshes’ across porespaces, apparently using the larger grains as anchor points. The biofilaments quickly became encrusted with fine grained material and surrounded with neoformed clay-like deposits. In contrast, the abiotic controls showed little or no evidence of clay formation suggesting that this process is biologically induced or controlled. A second series of abiotic experiments to determine the effects of increased acidity showed evidence of mineral pitting and dissolution along with an increase in concentration of soluble species thought to be important in smectite formation (i.e. Si, Al, Mg, Fe, Ca, Na). However, there was no evidence of clay formation, and the biotic experiments showed no signs of bulk scale pH change, suggesting that either the bacteria are actively concentrating relevant chemical species at a local level or they are acting as templates or nucleation points for clay formation.

Bentonite clay has an important buffer function in several of the proposed systems for nuclear waste repositories. Some proposed repository locations have significant concentration of sodium and calcium chloride in the ground-water, and others may be under sea-level for part of their life-time. A crucial question is consequently if there will be any long term effects on the stability of clay minerals due to exposure to saline conditions and thereby also on the desired physical properties.

In the present study we used material from a small bentonite habitat at San José, southeastern Spain, as a natural analogue to a bentonite buffer exposed to high saline ground-water. The composition of the bentonite material was determined by use of XRD, ICP/AES, and cation exchange capacity (CEC) analyses both of bulk material and of fine fraction. The important physical properties swelling pressure and hydraulic conductivity were measured for a range of densities and NaCl concentrations. The observed smectite structure and the measured properties of the San José bentonite were similar to those found in Wyoming bentonite and no obvious effects of the harsh saline environment were observed.

A geochemical modelling of one of the experimental parcels of the LOT experiment (Äspö, Sweden) is presented. The model consists on a 2D reactive transport model of a perpendicular section of the experimental borehole with a stationary temperature gradient ranging from 130 to 15°C. The results of the model show that temperature has an important effect on the geochemical evolution of the system. Due to the temperature effect, anhydrite replaces gypsum and hematite precipitates instead of Fe(OH)3(am). Calcite and siderite dissolve in the warmer part of bentonite; whereas a major replacement of Na by Ca in the bentonite takes place close to the bentonite-granite boundary. pH is buffered by the equilibrium with calcite, although the effect of temperature leads to a decrease in pH.