The UK Committee on Radioactive Waste Management (CoRWM) recommended, in 2006, that geological disposal coupled with safe and secure interim storage should have been the way forward for the long-term management of the UK’s higher activity wastes. The design of the underground repository contemplates the presence of bentonite plugs to seal access galleries and deposition boreholes and hence the interaction between the clay-based backfill material and the underground water. Remote monitoring of the fluid saturation of the barrier, the waste canisters and of the surrounding subsurface Geological Disposal Facility environment assumes a relevant importance to guarantee the safety of the repository and to increase the confidence and the reliance of the communities living in areas potentially affected by the repository over time. This remote monitoring of the Engineered Barrier System represents a technical challenge due to the unsuitability of some of the traditional geotechnical techniques or to the intrinsic unreliability of many geophysical prospecting techniques in providing information about the evolution of the Thermo-Hydro-Mechanical-Chemical coupling of the system over long timescales up to and including post-closure evolution. In this project, I offer an initial approach to an innovative way of using mineral magnetism, and, in particular, I analyse the possible exploitation of corrosion-induced variations of the magnetic properties of several magnetic materials to monitor water saturation in the Engineered Barrier System and its evolution through time. Initially the reactivity of several natural and synthetic materials is tested under different “extreme” conditions to analyse the feasibility of the research concept and identify the materials more adapt to carry out the job. The effects that the corrosion of the magnetic materials has on the clay matrix is also analysed in detail throughout all the thesis work. The initial tests lead to the identification of specific transitions in the hysteretic behaviour of three of the initial candidates (Nd-Fe-B, AlNiCo and SmCo alloys). These three materials are subsequently tested under conditions closer to a real “evolved” Barrier System, where the groundwater interacts, with cementiferous grout producing hyperalkaline leachates. The final tests consider the temporal evolution (after 4, 8 and 12 months) of the magnetic properties of these materials in a dysoxic environment under imposed fluid-flow. The results show a clear change in the hysteretic properties of the three materials analysed and the feasibility of the monitoring of the Barrier fluid saturation in the short-term. Furthermore, the corrosion of the magnets, under the conditions applied, did not cause formation of non-swelling clays.

Materials representative of those found in a Geological Disposal Facility (GDF) for the long-term storage of nuclear waste have been investigated for their ability to retard the movement of ionic species found in nuclear waste. Fe1-xO, Fe2O3, Fe3O4 (from steel corrosion) and sandstone (bedrock) used as physical barriers in the GDF have been treated using solutions of pH 7-13 which are representative of the leachate expected from concrete encapsulation of waste in contact with ground water. A mimic of portlandite cement, Ca(OH)2 was also prepared carbonate-free via a sacharate method for use in these leachate experiments. Materials have been characterised using a mixture of techniques such as Powder X-ray Diffraction (PXRD) and Infra-red Spectroscopy which focus on the bulk, short range techniques such as Extended X-ray Absorption Fine Structure (EXAFS), Scanning Electron Microscopy(SEM) and Nuclear Magnetic Resonance(NMR) and physical measurements such as diffusion experiments and fluorescence spectroscopy. Characterisation of the bulk materials before and after treatment using PXRD and SEM indicates that high purity iron oxides are affected differently by the solutions of varying pH. While not detectable by bulk techniques, SEM analysis evidence of the surface of the materials showed that Fe1-xO was deleteriously affected by solutions with pH > 7 more than the more oxidised materials. Initially needle-like crystals formed on the surface of Fe1-xO that are characteristic of goethite which at long aging times up to 168 h, showed transformation to crystal morphologies characteristic of Fe2O3. As the alkalinity increased, the transformation of Fe1-xO to Fe2O3 slowed. Dissolution of the iron surfaces in the solutions of pH 7-13 were determined by measuring the concentration of dissolved iron using ICPMS. While Fe1-xO and Fe3O4 followed first order kinetics, the dissolution kinetics for Fe2O3 appeared more complex. As the alkalinity increased, the rate constant for dissolution decreased in all cases indicating that higher pH is better for containment due to the formation of a passivated surface layer evidenced by SEM. The sorption of uranium to the iron oxide surfaces was investigated as a function of pH (7-13). In all cases, there was evidence of uranium sorption. The greatest sorption was evidenced when Ca(OH)2 was present which is most likely due to the precipitation of the known phase, calcium uranate. In the absence of calcium hydroxide, the sorption of uranium to the iron oxide surfaces decreased as the pH increased, reflecting the increase in formation of the anionic uranium species. In the presence of carbonate, the sorption of uranium onto the surfaces also decreases reflecting the formation of the soluble uranyl carbonate species. NMR spectroscopy of uranyl species in solution indicates that the chemical shift is strongly affected by pH shifting from 163 ppm to 175 ppm as the pH changes from 7 to 13 and allowing the uranium speciation to be used as a pH probe. A much -2- smaller shift in respect of temperature of less than 0.5 ppm was observed in the temperature range studied between 25 and 50°C. The quality of fluorescence spectra has been shown to be strongly affected by complexing species present in solution, the best spectra achieved with non-complexing species such as perchlorate. Migration experiments of the radionuclides uranium, thorium and technetium has been investigated by placing sandstone cores in alkaline solution and analysing both the water itself and the core to examine retention and transport. The results determined that the technetium diffused readily through the sandstone matrix. The uranium and throrium did not achieve breakthrough. This was attributed to the low solubilities and the formation of stable precipitates.

The generation of nuclear energy leads to the generation of contaminated, radioactive wastes. The current strategy in the UK is to dispose of high and intermediate level wastes to a geological disposal facility with no possibility for retrieval. The waste is contained in an encapsulation matrix, which is usually cement, however cement is unsuitable for certain waste types, for which epoxy resins have been proposed as an alternative. The radiation resistance of two candidate epoxy/amine resin formulations under repository conditions were tested with regards to the degradation of the backbone structure and the release of potential organic ligands from the polymer. The difference in the polymers was the choice of amine curing agent. Analysis of the polymer by infra-red spectroscopy and nuclear magnetic resonance spectroscopy revealed that the carbon nitrogen bonds are the most susceptible to radiation damage, regardless of the atmospheric and aqueous environment. The presence of an aqueous phase greatly reduces the availability of oxygen and reduces the rate of degradation when irradiated under an atmosphere of air. The properties of the aqueous phase has little effect on the degradation of the polymer. Thermal analysis revealed that the effects of the environment are limited to a thin surface layer of the polymer. Leachate analysis revealed that both organic and nitrogen containing compounds are leached from the polymer when irradiated in pure water. Under repository conditions however very little carbon and nitrogen is observed, suggesting that the calcium hydroxide present in repositories is capable of removing the leached species from solution. The generation of nitrate ions from air radiolysis over water is suppressed in the presence of the polymers, suggesting that nitrate is removed from solution by leached species or reaction with the polymer.

Geological disposal is the preferred option for the long term management of British intermediate level radioactive waste. The disposal site is currently being identified, with possible geological environments including fractured crystalline rocks and low permeability rocks such as clay. The selection of the host rock will have an impact on the design of the waste repository. This thesis investigates the ways the behaviour of repository borne gas can be affected by the repository design and the selection of the host rock. Commercially available TOUGH2 package is used to model the resaturation of the disposal facility, along with gas migration out of the repository and towards the ground surface in a generic geology. A facility located in fractured rock is estimated to resaturate within 6.5 years of its closure. The resaturation time is found to be strongly dependent on the presence and properties of a low permeability liner around the disposal vaults. The inflowing water starts gas generation processes within the repository; gas initially accumulates within the facility, but it is estimated to find its way into the host rock approximately 450 years after the facility has been closed. A maximum outflow rate is reached after approximately 1,000 years. The flow of gas migrating through the host rock is strongly affected by site-specific features. In the case of a uniform crystalline rock, gas is found to break through at the surface after 29,000 years. For a disposal site with a very slow groundwater flow rate, the resaturation phase may take several decades and gas outflow will occur much later. It is estimated that, in very low permeability environments, gas breakthrough may not occur before 100,000 years.

The UK has a large and growing inventory of higher activity radioactive waste awaiting safe long term disposal. The international consensus is to dispose of this radioactive and toxic waste within a deep geological repository, situated 200-1,000 metres beneath the ground surface. The deep geological disposal facility is designed to be a series of engineered and natural barriers. Groundwater forms an integral component of the natural barrier because it 1) controls the flux of reactive components towards the engineered repository, and 2) forms one of the primary transport mechanism through which released radionuclides can be transported away from the repository. The timescale of protection provided by the natural barrier exceeds those provided by the engineered barriers. Knowledge of the regional hydrogeology is a vital step towards predicting the long term performance of any potential repository site. Topically, a UK government decision in 2017 to re-open a nation-wide repository location search has now created a renewed mandate for site exploration. This research aims to determine the regional groundwater characteristics of three UK settings, selected to be hydrogeologically distinct, in order to determine which, if any, offers natural long term hydrogeological containment potential. The settings selected for analysis include Sellafield in West Cumbria, the Tynwald Basin within the East Irish Sea Basin, and Thetford within East Anglia. Site selection is based on diverse groundwater characteristics, and on previous research suggesting potential hydrogeological suitability at these locations. This research is novel in that it provides, for the first time, a direct comparison between the characteristics and qualities of different regional groundwater settings to contain and isolate radioactive waste, based on UK site specific data. Large and detailed numerical models for the three sites, covering areas of 30 km length by 2- 4 km depth have been developed using the open source finite element code 'OpenGeoSys'. The models couple the physical processes of liquid flow and heat transport, in order to replicate regional scale groundwater flow patterns. Models are calibrated to measured rock properties, and predict groundwater behaviour 10,000 years into the future. Uncertain parameter ranges of lithological and fault permeabilities, and peak repository temperatures are tested to determine the possible range of groundwater outcomes. Geochemical retention is assessed separately and validated using the finite difference modelling software 'GoldSim'. Worst case groundwater characteristics for containment and isolation at each site are compared to an 'ideal' benchmark far-field hydrogeological outflow scenario, and scored accordingly using a newly proposed method of assessment. Results show that the Tynwald Basin offers the best potential of the three sites for natural radionuclide containment, performing between 3.5 and 4 times better than Sellafield, and between 1.7 and 4 times better than Thetford. The Tynwald Basin is characterised by 1) long and deep groundwater pathways, and 2) slow local and regional groundwater movement. Furthermore, the Tynwald Basin is located at a feasible tunnelling distance from the coast, adjacent to the UK's current nuclear stockpile at Sellafield, and thus could provide a simple solution to the current waste legacy problem. Results from the Sellafield model indicate that this location cannot be considered to exhibit beneficial characteristics due to short and predictable groundwater pathways which ascend, from the repository, towards surface aquifers. Finally, Thetford within East Anglia has never been drilled to depth so that sub-surface rock properties of basement, located beneath layered sediments, are based on evidence inferred from around the UK. Uncertainties in rock properties has produced a wide range of groundwater characteristic possibilities, with results indicting prospective performance to range from 0 to 2.4 times better than Sellafield. As such, the hydrogeological suitability to host a potential deep geological repository is promising when modelled with most-likely permeability values, but cannot be accurately determined at present. Consideration of decaying heat from the heat emitting waste packages at the three sites reveal that the natural groundwater flow patterns can be distorted up to as much as 7 km away from the theoretical repository, depending on setting. This thus changes the use of the term 'near-field' for safety assessments, as implying an area within the immediate vicinity of the excavated repository site. The overarching findings from this research are that: 1) some locations have greater long term radionuclide containment and isolation prospectivity than others, due to variable quality far-field geological and hydrogeological characteristics; 2) the effect of radiogenic heat emission on the natural groundwater flow pattern is dependent on the site specific geological and hydrogeological characteristics, and therefore so is the area defined as the 'near-field'; and 3) a simple method of site comparison is possible for regional groundwater system under steadystate conditions. Recommendations are for scoping models of regional groundwater settings to be used as a comparative tool, such as undertaken as part of this research, to differentiate between potential sites at an early stage of the current UK site selection programme.

The safety case for a proposed geological disposal facility (GDF) for radioactive wastes relies upon a series of engineered and natural barrier systems to limit the migration of harmful radionuclides into the geosphere over geological timescales. Natural minerals, dominantly phyllosilicates, are expected to be the most reactive components of both the host rock and the clay-based backfill surrounding the highly radioactive waste canisters for as long as 100,000 years. Upon eventual canister degradation, alpha-emitting radionuclides will leach into the backfill material (and eventually beyond) and the constituent mineral systems will accumulate radiation damage upon radionuclide uptake and/or surface precipitation. The following study is an assessment of the structural and chemical effects caused by alpha-particle bombardment of silicate minerals, as proxies for the radiation stability of natural materials present in the near and far field of a GDF.Microscopy and spectroscopy studies from naturally occurring radiation damage accumulated in silicates over geological timescales (forming distinct 'radiohaloes') have shown that both alpha-particle and alpha-recoil bombardment results in altered unit cell dimensions caused by the accumulation of point (Frenkel) defects. In the example of highly damaged biotite, structural breakdown through the reorientation of discrete lattice crystallites was observed; the variability of the interlayer spacing within these regions reveal the potential for damaged mica to adopt the structure of phyllosilicate breakdown products over geological time. Controlled alpha-particle irradiation using the Dalton Cumbrian Facility's 5 MV tandem pelletron ion accelerator, combined with microfocus spectroscopy analysis has revealed the mechanisms of high fluence alpha-radiation damage across 2:1 phyllosilicate minerals (biotite and chlorite); reducing the layered structures into a series of loosely connected domains of alternating lattice expansion and collapse. Radiation induced Fe redox changes have been revealed, with Fe reduction apparent at relatively low alpha-particle doses, giving way to Fe oxidation at high doses. A 'redox gradient', based on alpha-particle energy deposition through a silicate structure has therefore been proposed. In addition, the increase in 'edge' sites generated by structural deformation has been shown to be favourable for the adsorption of the Se(IV) oxyanion to the mica surface. Comprising a body of additional work, a core sample has been extracted from a spent nuclear fuel pond wall at the decommissioned Hunterston A nuclear power station and the radioactive contamination on the painted core surface has been analysed by microfocus spectroscopy. The contaminant radiostrontium has been shown to be associated with the Ti rich pigment in the surface paint, resulting in a 'patchy' accumulation of radioactivity at the core surface. In addition, inert Cs reactivity experiments using the underlying concrete have shown that Cs is preferentially uptaken by phyllosilicates within the altered mafic clasts used in the concrete aggregate.

Geological disposal of radioactive wastes currently stored at Earth's surface is now the favoured management pathway for these materials. Typically, intermediate level wastes (ILW) are grouted and emplaced in a geological disposal facility (GDF) which will be backfilled, possibly with cementitious materials. Post-closure leaching of the cementitious materials in a GDF is expected to create hyperalkaline conditions in and around the repository, resulting in mineral alteration and crystallisation, both within the engineered barrier and host rock; creating a persistent chemically disturbed zone (CDZ). Iron derived from within the host rock as a result of alkaline breakdown of Fe-bearing silicate minerals (e.g. biotite, chlorite); corrosion products formed within the repository; or iron contained within the waste; will form secondary iron (oxyhydr)oxide minerals. The formation and re-crystallisation of these reactive mineral phases may sequester radionuclides through a host of processes: surface-mediated reduction to less soluble forms; adsorption onto, and/or incorporation into stable secondary or tertiary iron oxide phases. Therefore iron (oxyhydr)oxides will be key to the fate of radionuclides potentially released from within radioactive wastes disposed of in a GDF.In this study, the fate of U(VI) and Tc(VII) was considered during crystallisation of ferrihydrite to more stable iron oxide phases (e.g. hematite and magnetite) and, in three synthetic cement leachates (pH 13.1, 12.5, 10.5) designed to reflect the early-, middle- and late-stage evolution of the CDZ. XRD and SEM/TEM have been used to characterise the mineralogy during crystallisation. Partitioning of U(VI) and Tc(VII) between the solid and solution has been followed throughout, with chemical extractions used to determine the distribution of the radionuclides adsorbed to, and incorporated within the solid. Synchrotron-based XAS techniques have been utilised to probe the oxidation state and molecular scale bonding environment of the radionuclides associated with the solids. The data suggest that: U(VI) is incorporated into the hematite structure in place of Fe(III), in a distorted octahedral environment with elongation of the uranyl bond; Tc(VII) is reduced to Tc(IV) and incorporated into the octahedral site within the magnetite structure in place of Fe(III), and is retained in the same environment even after extensive oxidation of the magnetite to maghemite; and that U(VI) may also be incorporated as U(V) or U(VI) into the magnetite structure, with similar recalcitrant behaviour during oxidation. These results highlight the importance of mineral reactions within the CDZ as potentially significant pathways for immobilising radionuclides released from a GDF.

The geological repository for final storage of spent nuclear fuel, envisioned by the Swedish Nuclear Fuel and Management Company (SKB), relies on several barriers: copper canisters deposited in holes in the floor of underground tunnels in deep bedrock, embedded in a buffer of compacted bentonite. The initially unsaturated buffer would take up water from the surrounding rock mass and swell to seal any potential gap. This initial two-phase (gas and liquid) regime with two components (air and water) may impact the final density, swelling pressure and biogeochemical conditions in the buffer. A main objective of this work is to identify factors and mechanisms that govern deposition hole inflow and bentonite wetting under the prevailing two-phase flow conditions in sparsely fractured bedrock. For this purpose, we use the numerical code TOUGH2 to perform two-phase flow simulations, conditioned by a companion field experiment (the Bentonite Rock Interaction Experiment or BRIE) performed in a 417 m deep tunnel of the Äspö Hard Rock Laboratory in southeastern Sweden. The models predict a significant de-saturation of the rock wall, which was confirmed by field data. To predict the early buffer wetting rates and patterns, the position of local flowing fractures and estimates of local rock matrix permeability appear more important than the total open hole groundwater inflow. A global sensitivity analysis showed that the buffer wetting time and the persistence of unsaturated conditions over extended periods of time in the rock depend primarily on the local fracture positions, rock matrix permeability, ventilation conditions in the tunnel and pressure far in the rock. Dismantling photographs from BRIE were used to reconstruct a fine-scale snapshot of saturation at the bentonite/rock interface, showing tremendous spatial variability. The high level of heterogeneity in the rock generates complex two-phase flow phenomena (air trapping, dissolution), which need to be accounted for in buffer design and rock suitability criteria. In particular, results suggest that uncertainties regarding two-phase flow behavior are relatively high close to residual air saturation, which may also have important implications for other applications involving two-phase flows, such as geological storage of carbon dioxide. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 1: Manuscript. Paper 4: Manuscript.</p>

In Sweden the encapsulated nuclear waste will be surrounded by compacted bentonite in the granitic host rock. In contact with water-bearing fractures the bentonite barrier may release montmorillonite colloids that may be further transported in groundwater. If large amounts of material are eroded from the barrier, the buffer functionality can be compromised. Furthermore, in the scenario of a leaking canister, strongly sorbing radionuclides, can be transported by montmorillonite colloids towards the biosphere. This thesis addresses the effects of groundwater chemistry on the generation, stability, sorption and transport of montmorillonite colloids in water bearing rock fractures. To be able to predict quantities of montmorillonite colloids released from the bentonite barrier in contact with groundwater of varying salinity, generation and sedimentation test were performed. The aim is first to gain understanding on the processes involved in colloid generation from the bentonite barrier. Secondly it is to test if concentration gradients of montmorillonite colloids outside the barrier determined by simple sedimentation experiments are comparable to generation tests. Identical final concentrations and colloid size distributions were achieved in both types of tests. Colloid stability is strongly correlated to the groundwater chemistry. The impact of pH, ionic strength and temperature was studied. Aggregation kinetics experiments revealed that for colloid aggregation rate increased with increasing ionic strength. The aggregation rate decreased with increasing pH. The temperature effect on montmorillonite colloid stability is pH-dependent. At pH≤4, the rate constant for colloid aggregation increased with increasing temperature, regardless of ionic strength. At pH≥10, the aggregation rate constant decreased with increasing temperature. In the intermediate pH interval, the aggregation rate constant decreased with increasing temperature except at the highest ionic strength, where it increased. The relationship between the rate constant and the ionic strength allowed the critical coagulation concentration (CCC) for Na- and Ca-montmorillonite to be determined. In order to distinguish the contribution of physical filtration and sorption to colloid retention in transport, the different retention mechanisms were quantified. Sorption on different representative minerals in granite fractures was measured for latex colloids (50, 100, 200 nm) and montmorillonite colloids as a function of ionic strength and pH. Despite of the negative charge in mineral surfaces and colloids, sorption was detected. The sorption is correlated to the mineral point of zero charge and the zeta potential of the colloids, and increases with increasing ionic strength and decreasing pH. In transport experiments with latex colloids in columns packed with fracture filling material, the retention by sorption could clearly be seen. In particular at low flow rates, when the contact time for colloids with the mineral surfaces were the longest, sorption contributed to retention of the transport significantly. The retention of latex colloids appeared to be irreversible in contrary to the reversible montmorillonite colloid retention. Generation, stability and sorption of the montmorillonite colloids are controlled by electrostatic forces; hence, the results were in qualitative agreement with DLVO.

L’objectif principal de cette étude est d’évaluer le rôle de l’activité de bactéries hydrogénotrophes et ferri-réductrices sur le processus de corrosion anoxique en utilisant des indicateurs géochimiques. Il est considéré que le couple redox H2/Fe(III) est un moteur important pour les activités bactériennes qui peuvent ainsi affecter les vitesses de corrosion par la déstabilisation des couches de passivation (i.e. magnétite, Fe3O4). Les résultats indiquent que la magnétite de synthèse est déstabilisée en présence de bactéries hydrogénotrophes et ferri-réductrices due à la réduction du Fe(III) structural couplée à l’oxydation de l’H2. La quantité de Fe(III) bioréduit est augmentée en présence de concentrations croissantes en H2 dans le système: 4% H2 < 10% H2 < 60% H2. De plus, les résultats indiquent que la réaction de corrosion est différente selon la composition de la solution et la surface de contact de l’échantillon métallique (poudre de fer ou coupon en acier au carbone). Les produits de corrosion solides sont différents pour chaque échantillon étudié: vivianite, sidérite et chukanovite sont les principales phases minérales identifiées dans les expériences avec de la poudre de fer, tandis que vivianite et magnétite sont identifiées en présence de coupons en acier au carbone. Les résultats montrent que la vitesse de corrosion est quasiment deux fois plus importante en présence de bactéries après 5 mois de réaction. Cette étude apporte une nouvelle approche sur la compréhension des phénomènes de biocorrosion, l’identification des mécanismes physico-chimiques et la détermination des paramètres contrôlant la vitesse de corrosion. / The main objective of this study is to evaluate the role of hydrogenotrophic and IRB activities on anoxic corrosion process by using geochemical indicators. It is assumed that the redox couple H2/Fe(III) is an important driver for bacterial activities potentially affecting the corrosion rate by destabilization of passive layers (i.e. magnetite, Fe3O4). Our results indicate that synthetized Fe3O4 is destabilized in the presence of hydrogenotrophic IRB due to structural Fe(III) reduction coupled to H2 oxidation. The extent of Fe(III) bioreduction is notably enhanced with the increase in the H2 concentration in the system: 4% H2 < 10% H2 < 60% H2. Moreover, the results indicate that corrosion extent changes according to the solution composition and the surface of metallic sample (iron powder and carbon steel coupon). The solid corrosion products are different for each sample: vivianite, siderite and chukanovite are the main mineral phases identified in the experiments with iron powder, while vivianite and magnetite are identified with carbon steel coupons. Our results demonstrate that corrosion rate is increased almost two-fold in the presence of bacteria after 5 months of reaction. This study gives new insights regarding the understanding of biocorrosion phenomena, identification of physicochemical mechanisms, and determination of key parameters controlling the corrosion rate.