For a "safe" final bedrock deposition of high-level nuclear waste, we need guarantees for the immence time period of "many hundreds of thousands of years". No one can, of course, give such guarantees over such long time periods; on the contrary, the one who claims such a thing "has disqualified himself". The Fennoscandian crystaline bedrock is, by no means, as stable and reliable as claimed by the nuclear agencies. Only some 10,000 years back in time, the seismic activity was tremendously high; in amplitude as well as in frequency. In such an environment - to be repeated at the next future ice age (estimated to occur 5000 AP, 23,000 AP and 70,000 AP) - we can, of course, give no guarantees for a waste disposal in the bedrock; on the contrary, most facts suggest that it would be seriously damaged. In the absence of a true long-term safity, we can only propose that the waste produced: (1) is kept at a minimum, (2) is stored under constant control, and (3) is kept accessable for reparations as well as possible future innovations of how to rander the waste harmless. Therefore, we recommend the DRD-method.
Mörner, 1988, Striae, 31, 55-58, (1988).
Mörner, 1993, Zeitschr. Geomorph. , N. F. , Supl. -Bd. , N. F. Suppl. Bd, 102, 107-117, 159-166, (1993).
Mörner, 1996, Quaternary Science Review, 15, 939-948, (1996).
Mörner et al, 1997, Glaciation and Hydrology, SKI Report, 97:13, 89-94, A49-52, A67, A73-74, (1997).
Mörner et al, 1998, International Symposium on Radioactive Waste. SEI active Waste Disposal, Abstracts, S6, 7, (1998).
The Geological Survey of Sweden (SGU) in collaboration with the Swedish Nuclear Fuel and Waste Management Company (SKB) has recently completed a multidisciplinary, geological feasibility study of the Precambrian bedrock of Sweden. This study has been carried out to help identify areas which are of interest for more detailed investigations in order to locate a site for the disposal of highly radioactive nuclear waste. The results of the study have been presented in a systematic manner over the country in the form of twenty, separate county reports and form part of SKB's regional overview programme for site selection.
Geological factors upon which attention has been addressed in each county include bedrock composition and homogeneity, mineral and bedrock resources, identification of areas which are potentially of interest for mineral exploration, composition and thickness of Quaternary deposits, regional hydrogeological characteristics, regional ductile and brittle high-strain zones including late- and post-glacial faults, land uplift and registered earthquake activity. Special attention has been focussed on regional high-strain zones and the embryo to a new tectonic synthesis of the Precambrian shield area in Sweden has emerged under the auspices of this study. A summary assessment of data availability in the different counties has also been completed. Geological and geophysical information from various data bases, which are predominantly compiled and archived at SGU, as well as from published and unpublished maps and reports have been utilized.
This paper aims to present the method of approach and some of the results of this regional-scale study. Attention will be focussed on the factors which have dominated when selection of areas of interest for future, more detailed investigations has been carried out, viz. bedrock composition, exploration aspects, regional high-strain zones and registered earthquake activity.
The safety assessment of a repository for radioactive waste is based on the results of model simulations and scenario analyses. To obtain a well founded geological basis for this, extensive geological and structural studies have been carried out for the Morsleben repository. The Morsleben repository for radioactive waste was constructed in Zechstein strata. The Werra Formation (z1) has been located in boreholes and the Stassfurt (z2), Leine (z3) and Aller (z4) Formations are exposed in the repository mine. The Zechstein Salt has a total thickness in the western part of the salt body of 580 m. In the eastern part it is only 380 m thick. The Morsleben repository is located in the Allertal zone salt structure, the root zone of a deeply eroded salt body. This is evidenced by the relatively small thickness of the Zechstein Salt, as well as by the many enclaves of the Hauptanhydrit (z3HA) in the salt, which have been concentrated in the synclinal hinge zones, i.e. in the deeper parts of the structure, by a process of tectonic separation of ductile salt from competent anhydrite. The tectonics of the structure are documented by polyphase folds, characteristic of deformation of ductile material. The general trend of the fold axes is horizontal, NW-SE, parallel to the Allertal zone. The western part of the salt structure is characterised by high, NE-vergent, isoclinal folds. z2 to z4 strata involved in the folding extend from the salt table (top of the salt body) down to the lowest synclinal trough.In the eastern part the folds are open and symmetrical with considerably smaller amplitude. Correspondingly, z4 strata mostly occur just under the salt table, z3 at the repository level, and z2 below that. The different styles of folding in the east and west is due to the structure of the rocks beneath the Zechstein Salt. The boundary between the two fold types lies above a major fault in the rocks below the Zechstein Salt, separating a downthrown block in the west from an upthrown block in the east. The NE-vergent, isoclinal folds occur above the downthrown block, where the major uprise of salt occurred. In contrast, the folding on the upthrown eastern block has hardly been affected by the rising salt; hence the wide, open, symmetrical folds. Thus, the structure of the rocks beneath the Zechstein determined the folding style in the overlying salt. The present-day fold structures chiefly result from Late Cretaceous compression that affected the North German basin. The salt body is therefore primarily a tectonic structure and is not halokinetic.
During site investigation for the Morsleben repository for radioactive waste, the tectonic development of the strongly eroded Upper Allertal salt structure was reconstructed. Since the results of surface and underground work on the Morsleben salt dome were insufficient for this reconstruction, the geological history of the Altmark and Subhercynian (area north of the Harz Mts.) region was incorporated into the study. The stratigraphic succession from Upper Rotliegendes to Quaternary is well known from recent boreholes. The sequence up to Middle Lias is complete. From here to the top of the Palaeogene, the succession is incomplete. The relatively complete Quaternary succession comprises Lower Pleistocene deposits and products of the three glaciations of northern Germany, as well as Holocene deposits.Analysis of the sequence of movements that accompanied the rise of the Allertal structure demonstrated that the first strong tectonic activity took place in the Middle Keuper. This took the form of extensional movements that affected entire northern Germany. In the study area it led to the formation of a fracture called the Allertal fault zone, which became wider with time and separated the cover rocks into blocks. These rafts underwent gravitational sliding towards the deeper parts of the subsiding basin and thus initiated the flow of Zechstein salt into the fault zone. Stress reversal associated with the Late Cretaceous inversion caused compression of the Allertal fault zone, forcing the Zechstein salt still remaining beneath the Lappwald syncline to migrate into the fault zone and into higher levels. The stress reversal also led to the formation of compression structures at depth. The normal faults originally present at the base of the salt structure were reactivated during Late Cretaceous wrench tectonics into asymmetrical positive flower structures. From middle Santonian onwards, regional uplift and erosion began and removed some of the sedimentary succession. Altogether, the sediment pile was reduced in thickness by about 1500 m. Consequently, the salt body reached a level in which it was subject to underground solution (subrosion). Contemporaneously a crestal graben formed. Later a subrosion trough 500-100 m wide and up to 240 m deep developed and was filled with Campanian to Palaeogene clastic sediments. Temporary interruption of subrosion during Early - Middle Eocene was followed by removal of the Lower Tertiary sediments during the Late Tertiary. In the Quaternary, subrosion recommenced and still continues. Calculation of the amount of salt present today compared with the original thickness of the salt shows that 73% of the mobilized salt that flowed laterally into the salt structure has been eroded or dissolved away. The present-day salt structure averages 490 m in height; this is remarkably less than the thickness of the original Zechstein salt formation. The Allertal structure must therefore be interpreted as the remains of a salt diapir in a fracture zone. The present-day caprock represents the insoluble residue from only a third of the salt that has been removed. The missing two-thirds is probably explained by the salt being eroded during the period of strongest uplift.
For the licence application of the ERA Morsleben radioactive waste repository in Germany, a long-term prediction was made of the behaviour of the geological barrier system. The Morsleben repository is sited in rock salt of the Allertal salt structure.
The duration of the period predicted was about 150,000 years and contained one full glacial-interglacial cycle. Climatic change was thought to be the most important driving force for natural processes that may have affected the geological barriers on this time scale. The time dependence of climatic change is based on the astronomical or Milankovitch-cyclicity.
The following processes, which may directly influence the geological barriers, were analysed: subsurface solution of salt (subrosion), fluvial erosion and subglacial erosion. The process of fluvial erosion connected with epirogenic uplift was evaluated for the Morsleben region. Computer models were developed to analyse the effects of subrosion (Bremmer et al., 1999) and of fluvial erosion (Van den Berg et al., 1999). Depending on the process, the model was constructed on two or three superposed spatial scales: local scale (repository), regional scale (Subhercynian Basin) and supraregional scale (N.W. Europe). A palaeogeographical and -hydrological reconstruction of the region during the late Quaternary was performed to provide the models with the necessary boundary condition at the top-boundary of the model (Van Gijssel, 1999).
Subrosion appeared to be most intense under permafrost conditions. Under these circumstances infiltration of water takes place in the Aller valley, because the valley bottom itself is inferred to have remained unfrozen. In most other situations the Aller valley is an exfiltrating system and associated with lower subrosion rates. In the most extreme case a rock-salt layer with a thickness of several metres could be subroded within the next 150,000 years. Fluvial erosion is driven rather by tectonic uplift than by climatic forces. Depending on the tectonic uplift scenario, up to about 10 m of surface sediments or rocks could be eroded within the next 150,000 years.
Bremmer CN, Kreft E, Rijkers RHB, Simmelink HJ & Van Weert FHA, Conf. Abs., 4, (1999).
Van den Berg MW, Veldkamp A & Van Dijke JJ, Conf. Abs., 4, (1999).
Van Gijssel K, Conf. Abs., 4, (1999).
A "Skura" is a canyon-like valley, generally interpreted to be formed by glacifluvial erosion. In southeastern Sweden, there are a number of such canyons in Precambrian granites and vulcanites. Sizes of the canyons vary considerably from just a few meters deep and some hundreds of meters long to 30-40 m deep, several tens of meters wide and up to seven kilometers long. The general morphology is characterized by steep-sided rock walls with no obvious sign of glacial erosion. A majority of these types of canyons strikes approximately N-S, which is parallel to the dominant fracture direction and the orientation of dolerite dikes.
There appears to be no difference in the fracture systems in the lowland compared to the area above the highest marine level. Although most occurrences of skura valleys are found in the area above the highest marine level, valleys in the lowlands exist. An example of such a canyon, striking NW-SE and about 50 m deep, is located in the sea NE of Moredalen.
Structures of this kind indicate certain characteristics of the bedrock that need to be considered during safety analysis of repositories for nuclear waste. Selective weathering of the distinct weakness zones creates prominent transport paths for local to regional groundwater flow. Furthermore, the zones may be potential hazards from a rock mechanical point of view.
This paper discusses a combined geological and geophysical investigation carried out in the area with the aim to investigate possible reasons for the forming of this unusual feature. The studied object, Moredalen, is a marked, approximately 7 km long, E-W striking valley that cuts through an elevated plateau (140-150 m.a.s.l.) coinciding with the sub-Cambrian peneplain. During late glacial time the More river discharged into the Baltic Ice Lake. To the west, eskers and glacifluvial sediments can be found where the canyon widens. Just east of the valley is a larger delta. To the east of, and in line with, Moredalen is an esker.
Lineament studies (relief maps and aerial photos) show that Moredalen follows one of several accentuated E-W trending structures, parallel to the contact between two rock complexes. The shape of the valley, with erosion due to crosscutting structures, indicates that the formation of the valley predates the latest glaciation. Geophysical investigations, comprising electrical and electromagnetic surveys, helped determine the depth to the bedrock but also to investigate sedimentary sequences of the material in the canyon and the delta. The delta formed by water running through Moredalen appears to be deposited on top of the sediments (at present farmland) carried by the river Emån or its precursor. The formation of the Moredalen canyon by a single catastrophic event, a discharge of a subglacial reservoir, is questionable.
The geotechnical stability analysis is a critical part of any safety assessment. Engi-neering geology study of the site, laboratory and in-situ experiments, geomechanical modelling, and numerical static calculations comprise such an analysis.
Central to the calculations for the safety analysis is the construction of a model of the rock that simulates as closely as possible the actual conditions of the rock (i.e., the geology, stress conditions and constitutive laws).
According to the multi-barrier principle, in waste repositories the geological setting must be able to contribute significantly to the waste isolation over long periods. The assessment of the integrity of the geological barrier again can only be performed by making calculations with geomechanical and hydrogeological models. The proper idealisation of the host rock in a computational model is the basis of a realistic calcu-lation of stress distribution and excavation damage effects.
In the paper some important processes for the stability analysis are described:- creep and fracturing- permeability and infiltration- halokinese and subrosion.For the future, the role and contributions of geoscientific and rock mechanics work within the safety assessment issues (e.g. geomechanical safety indicators) must be identified in greater detail, e.g. considerations of geomechanical natural analogy for calibration of constitutive laws.
Permeability of undisturbed rock salt or rock salt which has been sufficiently compacted during laboratory testing is less than 10-20m2. Rock salt deforms plastically without any crack formation and dilatant crack propagation as long as the state of stresses is within the non-dilatant stress regime. Results of deformation experiments on rock salt (axial compression and axial extension) will be presented describing the transition from compressive to dilatant deformation behaviour at the dilatancy boundary. In addition to the measurement of the volumetric strain, combined measurements of the permeability and the ultrasonic wave velocities (Vp, Vs) have been used to determine the state of stresses at the dilatancy boundary. For different types of rock salt this boundary can be formulated by a function which solely depends on the principle stresses, although the rocks can possess rather different mineralogical composition and creep ductility. Load geometry and type of rock salt seem to have nearly no influence on the dilatancy boundary.
During loading in the dilatant stress regime cracks are propagating and damage increases with strain (or time resp.). This can be continuously monitored by the combined permeability-ultrasonic technique. The development of damage and permeability depends on several parameters. Most important are load geometry, minimal principle stress and ductility of the type of rock salt. The experimental results will be discussed with respect to well known equations which describe permeability as a function of damage (i.e. dilatancy, porosity). At the present state of knowledge permeability-porosity relations have to be formulated separately for each type of rock salt, because path ways for fluids are partly controlled by the type.
Salt rocks reveal a large variety of geomechanical behaviour. Rock salt is capable of creep deformation without the formation of micro- and macrofractures in a wide range of the conditions of state. Because of its ductility, rock salt is considered to be a favourable host rock for the underground disposal of hazardous waste. On the contrary, anhydrite and salt clay are examples for elastic-brittle behaviour. So strong contrasts in the geomechanical properties occur at the boundaries of rock salt and anhydrite or salt clay. Differences, which are less pronounced, result in case of rocks with different degree of ductility like rock salt and potash or - in special cases - rock salt of different stratigraphy. In a thick homogeneous layer of rock salt fracturing in form of microcracks will be limited to thin zones around the cavities. If rocks with different geomechanical properties are present, the redistribution of stresses due to mining will extend deeper into the rock and fracturing might occur at the boundaries. The investigation of the micro- and macrofracturing processes extending into the rock is important for the evaluation of long-term safety of underground disposal.
In order to study crack formation in the rock, acoustic emission (AE) measurements are conducted in case of various geological situations in salt rock. The acoustic energy radiated during the crack process is recorded by piecoelectric transducers in the frequency range fom 1 to 100 kHz. Then the locations of the AE events can be determined using the travel times of the P and S waves and their magnitudes can be determined using the maximum amplitudes. The method is very sensitive so that even small cracks with linear dimensions of millimeters can be detected. In addition, geological and geotechnical investigations are carried out. As AE measurements are a relative new method to obtain statements on the geomechanical situation, drilling into areas of AE activity und following inspections are performed in examplary cases to derive the geomechanical implications of the AE results.
Results to be presented have been obtained from a boundary between rock salt and anhydrite which has a distance of about 20 m from the roof and the wall of large rooms in rock salt. The spatial distributions of AE locations approximately show the form of rings. It was assumed that this kind of microcrack activity outlines the growth of more or less circular fracture planes of some meters extent. Drilling and coring confirmed this conclusion. Other results to be presented can be attributed to the boundary of rock salt and potash in one case and to thin anhydrite layers in a thick rock salt layer in another case.
A three-dimensional finite-element model was constructed to examine the role of groundwater flow on the advective transport in a porous silty medium, in connection with the suitability of an underground laboratory for studies relating to the disposal of radioactive wastes in southern France. Emphasis was directed to precisely reproducing the complex geological description of this multilayer aquifer-aquitard system, including a large valley of Pliocene erosion that formed the Rias in the area. The finite-element mesh, comprising three-dimensional prisms and tetrahedra, was created on the basis of a detailed geological description including maps of isobaths and isopachs.
A thorough sensitivity analysis on the hydrodynamics was performed including the effects of tectonics (inclusion of faults), geological structures (messinnienne erosion) and lithofacies (hydrogeological heterogeneities). The sensitivity to specific heterogeneities (channelling, lithofacies distribution) were characterised using geostatistical simulations of the hydraulic conductivity conditioned on hydrogeological information.
Results of several direct and inverse steady-state solutions showed the relative importance of the Pliocene erosion, the presence of faults, the surface boundary conditions, and the hydraulic connections to the deeper hydrogeological formations. The preliminary results also suggest that, at the local scale, the deep aquifers are isolated from the hydrodynamics of the more shallow formations located directly above the proposed site. The main model uncertainties are related to the surface boundary conditions and the Rias heterogeneity. This local-scale model is intended to be used for the optimisation of field investigations around the site for the purpose of more specific hydrogeological simulations in the future.
As part of a preliminary geological characterization programme to assess the feasibility of an underground laboratory in granitic rock, a series of deep boreholes (down to a maximum depth of 900 m) have been drilled by ANDRA in the Vienne district. Sr- and B-isotope ratios were measured in the fluids encountered during the coring operations in order to trace the degree of water-rock interaction and mixing processes in the local groundwaters, contribute to the understanding of present and past groundwater fluxes, and develop a conceptual model of the groundwater flow system. The groundwater samples from the deep boreholes, including deep saline groundwaters as well as the shallow aquifers from the overlying sedimentary cover, were also analysed and compared to CO2-rich thermal waters of the Massif Central granitic basement. The chemical analyses of the groundwaters were performed by ICP-MS for B, Rb and Sr concentrations and by mass spectrometry for the 87Sr/86Sr and 11B/10B ratios.
Plotting the data on a 87Sr/86Sr vs. 1/Sr diagram reveals three distinct groups of waters (Négrel et al., 1997). The deep borehole fluids, with salinities up to 10 g/l, constitute the first group. These saline waters are characterized by a homogeneous low Sr isotope ratio (87Sr/86Sr = 0.70831 ± 0.0004) and display the largest 11B -enrichment (24.91< 11B <36.08‰). The second group comes from the Infra-Toarcien aquifer and has the highest Sr isotope ratio (87Sr/86Sr = 0.71196 ± 0.00063) and an intermediate B isotopic ratio (11B = 13.02-13.15‰). The third group, from the Dogger aquifer, has an intermediate Sr isotope ratio (87Sr/86Sr = 0.70993 ± 0.00033); its negative 11B value (11B = -2.38‰) compares with the depleted B-isotopic field of the deep granitic waters from Massif Central (range -0.07 - 7.3‰; Mossadik, 1997). This clear crustal boron signature of the third group is, however, carried by a boron concentration that is smaller by an order of magnitude in the Dogger. The chloride-bromide relationships in the deep saline and shallow groundwaters and in the surface waters from the Vienne district show a Cl/Br range similar to that of seawater. The data plot on a dilution curve between marine and dilute meteoric water end-members. The Cl-Br investigation agrees with a marine origin for the saline waters without evolution from seawater. The combined use of 11B, Cl, B, Br, Sr contents and 87Sr/86Sr ratios makes it possible to define and quantify a mixing model between marine and crustal end-members in order to explain the origin of the deep saline groundwaters in the Vienne granitoids.
Négrel Ph, Casanova J & Aranyossy JF, Goldschimdt Conference 1997, Tucson, J. of Conference Abstracts, 1, 149, (1997).
Mossadik H, Ph. D. Thesis, Univ. Orleans, 224p, (1997).
The Gorleben salt dome, a potential repository site for all kinds of radioactive waste, is located near the community of Gorleben in the northeastern part of Lower Saxony. Hydrogeological investigations are being conducted in an area of more than 300 km2 around the salt dome to study the aquifer system in the sediments above it. During the extensive field investigations and data evaluation, the existence of heterogeneities on various scales has been verified. These heterogeneities predominantly reflect the differences in the lithology. At a large scale, ranging from decameters to kilometers, the site is characterized by a predominance of aquifers with intercalated aquitards and aquicludes. The salt dome is crossed by a subglacial erosion channel in which the lowermost aquifer is partly in contact with the caprock or the salt itself. The density of the groundwater, which is a function of salt concentration, varies spatially. It generally increases with depth, often up to that of saturated brine at considerable depth. The change in groundwater density has to be considered together with spatially variable hydrogeological parameters because it has a pronounced influence on groundwater movement and contaminant transport.
Three-dimensional, porous medium, freshwater models represent the overall hydrogeological structure and neglect the small-scale heterogeneities. These models are used to evaluate the regional groundwater movement. They provide characteristic features of the main flow paths under freshwater conditions. Preliminary uncertainty and sensitivity analyses have been performed using such freshwater models. Two-dimensional, porous medium, variable salinity models are used to calculate changes in the flow field, density distribution, flow paths and travel times. They have demonstrated the sensitivity of the salt/fresh water system to major changes in the hydrogeological structure, as well as other important factors like initial density distribution and period of time being modeled. Comparison of measured and calculated hydrochemical data shows the importance and validity of the incorporation of the hydrogeological structures in these flow and transport models. But calculations using the two-dimensional and simple three-dimensional models demonstrated that it was not possible to model such complex transient density-dependent flow system in three dimensions with the existing computer programs owing to limited computing time and memory. A program that is being developed using advanced numerical techniques will soon enable more sophisticated three-dimensional sensitivity studies that take into account the spatial variability of the hydrogeological parameters and the variability in groundwater density.
Safety assessments of proposed repositories for subsurface disposal of long-lived radioactive wastes involve predictions of groundwater movements over timescales into the future during which there could be major climate changes, e.g. ice sheet cover and permafrost over northern Europe. The susceptibility of groundwaters to these transient disturbances, and the overall stability of deep groundwaters for long timescales, can be investigated by palaeohydrogeology of the Quaternary period. The duration of Quaternary glaciation is comparable to the timescale into the future over which radioactive waste should be safely contained.
Project 'EQUIP' is investigating the records of past hydrogeochemical conditions (i.e. the chemical reactions between groundwater and rock) that are preserved in secondary minerals in low permeability fractured rocks. EQUIP is part of the EC programme of research for radioactive waste management and disposal, and involves collaboration between the following radioactive waste management agencies: Posiva (Finland), Andra (France), SKB (Sweden), Nirex (UK) and Enresa (Spain).
The general approach involves:· considering palaeohydrogeological models (hydraulic and chemical) for study sites,· identifying the most recent episodes of fracture mineralisation,
* characterising the mineralisation (phases, morphology, amount and distribution),
* analysing minerals by electron beam and other techniques for compositional zoning,
* detecting and analysing fluid inclusions trapped in recently-formed minerals,
* obtaining isotopic evidence of age of mineralisation where possible,
* interpreting mineral compositions with respect to present and past groundwater compositions.
Fracture-infilling minerals in crystalline rock are being studied at four locations, three of which have been affected by ice sheets and permafrost for various periods during the Quaternary. A previous study of drillcores from basement rocks in northwest England has shown that the crystal morphology and zoned composition of the most 'recent' calcite fracture-infill are related to spatial variations in groundwater salinity (Metcalfe et al., 1998).
This paper presents some preliminary results from the mineralogical studies. Fracture-filling calcite is found at all the studied sites, though the quantities of the youngest calcites are generally low. Other recently-formed minerals have been observed, including iron oxyhydroxide at some sites. Microanalytical methods are being used to characterise compositions and growth zoning, and to identify fluid inclusions. The most useful methods for analysing zoning are cathodoluminescence, backscattered electron microscopy, and electron microprobe. Zones are characterised by their concentrations of Ca, Mg, Sr, Ba, Fe, Mn and other trace elements.
Metcalfe R, Hooker PJ, Darling WG & Milodowski AE, Dating and Duration of Fluid Flow and Fluid Rock Interaction, Parnell J (ed), Special Publication, Geological Society, London, 144, 233-260, (1998).
Reconstructing the chemistry of ancient groundwaters for comparison with modern groundwaters is a crucial factor in the risk assessment of underground, nuclear waste repositories. Thermodynamic modelling of mineral-fluid interactions, mineral chemistry and mineral isotope analyses provide valuable indirect estimates of the overall hydrogeochemical environment. Direct information on solute chemistry is generally lacking however. In theory, this information is preserved in the aqueous fluid inclusions trapped in cements and minerals deposited from the palaeogroundwaters. Laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) is now an established technique for the elemental analysis of high salinity (>100,000 ppm TDS) hydrothermal fluid inclusions (Shepherd et al., 1995, 1998; Moissette et al., 1996). In this study, we have extended the technique to analyse low salinity, aqueous inclusions in calcites deposited from ancient groundwaters at the Sellafield site, Cumbria, UK.
Carbonate deposition at Sellafield is predominantly fracture-controlled and can be related to mineralization (calcite-sulphide-sulphate) that occurred during the Quaternary and is ongoing; being closely associated with present-day flow-zones in the site boreholes (Milodowski et al., 1998). Individual calcite crystals range from <5 µm to 2 cm in size and often overgrow earlier Permo-Triassic calcite-dolomite-fluorite-hematite-barite-anhydrite mineral assemblages. Aqueous fluid inclusion within the Quaternary calcites are exclusively monophase (i.e. liquid only) and typically <20 µm in diameter. Microthermometric analysis of these inclusions yields final ice melting temperature equivalent to TDS values of <3000 ppm.
To achieve the detection limits required for the analysis of such dilute fluids it was necessary to develop a cold plasma LA-ICP-MS methodology. This works extremely well for the measurement of the alkali and alkaline earth elements but cannot be used for the measurement of chlorine. A description of the technique will be discussed and data presented to illustrate the hydrogeochemical characteristics of Quaternary groundwaters at Sellafield. Comparisons will be made with older Permo-Triassic brines and present-day groundwaters.
Shepherd TJ & Chenery SR, Geochimica Cosmochimica Acta, 59, 3997-4007, (1995).
Shepherd TJ, Ayora C, Cendon DI, Chenery SR & Moissette A, European Journal Mineralogy, 10, in press, (1998).
Moissette A, Shepherd TJ & Chenery SR, Journal Analytical Atomic Spectrometry, 11, 177-185, (1996).
Milodowski AE, Gillespie MR, Naden J, Fortey NJ, Shepherd TJ, Pearce JM & Metcalfe R, Proceedings Yorkshire Geological Society, 52, in press, (1998).
The natural nuclear fission reactors of Oklo and Bagombé are the only natural analogues for nuclear waste management where actinides and fission products (FP) have been trapped in a geological environment for a very long period of time (2.0 b.y). For this reason, they have been intensively studied in an European program: Oklo-Natural Analogue Phase II.
The fission reaction took place in a very high grade uranium ore (30-60%U), namely the core of the reactor. The heat produced during criticallity induced intense hydrothermal circulations resulting in the crystallisation of new minerals: mainly clays and to a less extend phosphates and various oxides (Fe, Pb, U). This allows to study the role of the clays and other secondary mineral phases for U and FP retention during a very long period of time and the effect of the radiolysis of the water in and around the core of the reactor.
Most of the previous studies were made on the deepest reactors (450 m deep) where it has been shown that the actinides and lot of FP were retained in the uraninite and to a lesser extend in secondary minerals such as clays and phosphates.
More recently, studies were focused on the shallowest (11 m deep) reactor, namely the reactor of Bagombé, which has been submitted to a recent weathering under a wet equatorial climate. At Bagombé, the reactor is located 3 meters below a lateritic profile resulting from the recent weathering of ancient black shales. Because U and LREE from the core of the reactor have «abnormal» isotopic compositions relative to the common natural values, they can be used as good tracers to study their behaviour in a geological environment. In order to determine the various processes involved in the migration/retention of U and REE in such open system, both the water phase and the solid phases have been studied. This allows to discuss the water-rock interactions in various geochemical environments (reduced and oxidised) and to better understand the importance of the different retention processes (adsorption, complexation, incorporation in mineral structures) of elements on various minerals (oxides, phosphates and clays). Finally, analytical and experimental data allow to propose a model for uranyl sorption on the complex natural samples from Bagombé.
The aim of this presentation it to give the geological, mineralogical and hydrological background on the different reactors and to formulate the questions we can answer in the various projects actually running at the CNRS in Strasbourg.
The natural nuclear fission reactor of Bangombé is located at only 12 m depth and is therefore subjected to weathering and to chemical exchange with groundwater. Water samples have been taken from the reactor and from different drill holes lined up with the direction of the groundwater flow. All waters have been filtered. On both suspended loads and water ICP-MS and TIMS measurements have been performed in order to determine the REE concentrations and Sm-Nd isotopic compositions. Only the water taken directly in the reactor shows anomalous 143Nd/146Nd, 145Nd/146Nd and 149Sm/147Sm isotope ratios of 0.7235, 0.4933 and 0.843, respectively (normal: 0.7094, 0.4828 and 0.9173). This suggests that only 2.3% of the dissolved Nd is of fissiogenic origin. The water collected in the well closest to but outside the reactor shows normal isotopic compositions. The leachates of the corresponding suspended loads show similar isotopic anomalies with a fissiogenic contribution of 2.8 to 3.2%. Both dissolved and suspended load samples from the more distant wells have normal isotope ratios. Thus, the prevailing physico-chemical conditions within the water table did not allow the migration of the fissiogenic REE over large distances (<4 m). The fissiogenic component of 2 to 3% is rather low considering the huge REE availability in the reactor zone. Also the REE concentrations in the water from the reactor zone (pH 5) are very low (e.g. 1.9 ppt Nd). This is even more puzzling considering the huge amounts of leachable P (8.9%wt.) but also rather low REE contents (e.g. 1 ppm Nd) in the suspended load of the water from the reactor zone. This suggests on one hand corrosion of large quantities of phosphates and on the other hand the retention of especially REE in the reactor zone by solid phases like secondary uraninites, coffinites, phosphates and organic matter.
The natural fission reaction at Bagombé and Oklo was detected by the depletion of 235U/238U ratio in the reactor core. The presence of fission products and their anomalous isotope composition in the reactor core provided further unique and definite proof (Bodu et al, 1972; Neuilly et al, 1972). The fissiogenic isotope composition is strongly pronounced in REEs and is used for tracing migration of REE from the reactor core e.g. due to weathering or hydrothermal events. The actual isotope composition of REEs in the reactor core is depending on fission yield and neutron capture reactions (nuclear parameters). The isotope composition is also depending on pre-reaction REE- and U-concentrations (Ruffenach et al., 1980). This has taken into account when defining a source term for mixing fissiogenic REEs and normal REE by migration. In some cases it may be difficult to distinguish an in-situ fission reaction, a neutron capture reaction and migration of fission products when additional information is lacking e.g. concentration of fission products (e.g. SHRIMP measurements). ICP-MS and TIMS were used to obtain concentration and the precise isotope composition of Nd, Sm, Eu, Gd and Dy on reactor clay (BAX 08, 10.80 m) and on organic matter (BAX 08, 12.63 m) located 2 m below the reactor. The reactor clay has a strong depletion in 155Gd/160Gd (0.58) and 157Gd/160Gd (0.62) whereas the organic matter has a small depletion in 155Gd/160Gd (0.675) and 157Gd/160Gd (0.7148). The normal value of 155Gd/160Gd is 0.677 respectively 157Gd/160Gd is 0.7158. (2 <sigma> of the mean is smaller than last digit). These differences from a normal isotope composition can be explained by neutron capture reaction easily although migration is likely. Further large (reactor clay) and small (organic matter) differences from normal isotope composition are detected in Nd, Sm, Eu, Gd and Dy. Therefore some considerations are done in order to detect and estimate mixing due to migration of fission products using the isotope composition only. This provides a tool for the use of fissiogenic isotope ratios in mixing plots and is applied on selected samples.
Neuilly M, Bussac J, Fréjacques C, Neif G, Vendryes G, Yvon J, C. R. Acad. Sci. Paris, 275, 1847-1849, (1972).
Bodu R, Bouzigous H, Morin N, Pfiffelmann JP, C. R. Acad. Sci. Paris, 275, 1731-1732, (1972).
Ruffenach JC, Hagemann R, Z. Naturforsch, 35a, 171-179, (1980).
This study deals with the behaviour of the rare earth elements (REE) during the corrosion of glassy basalts in a salt environment and with REE mobility in salt rocks. Our study objects are Tertiary basaltic dykes from central Germany which crosscut potash salts of Upper Permian age (Zechstein). The aim of the study is to characterize the mechanisms which control REE mobility during basalt alteration and REE migration in salts. This is a natural analogue study for radionuclide behaviour in a high level radioactive waste (HLW) salt repository, because the trivalent rare earths are chemical analogues for the trivalent actinides. The study aims to contribute to a better knowledge of radionuclide behaviour during (HLW) glass corrosion and radionuclide migration in salt formations. In addition to element concentration data, we also use Nd and Sr isotopic compositions in order to determine the origin of the rare earths and to quantify fluid-rock interaction, respectively.
Our data from the basalt dykes demonstrate mobility and fractionation of the REE during a post-intrusion circulation of salt brines, during which the light rare earths (LREE, La-Nd) were more mobile than the heavy ones (HREE, Dy-Lu). The processes controlling this behaviour of the REE were dissolution and reprecipitation of phosphates. The refixation of the mobilized REE into secondary phosphates shows that the scavenging capacity of the precipitating apatite was stronger than the complexation of the rare earths in the circulating brine. This suggests that outgoing radionuclides from a leaking HLW salt repository would probably also be retained when phosphate minerals are present in the backfill material.
Although the basalt data indicate no substantial migration of the rare earths into the adjacent salts, basalt-derived REE could be detected in a salt profile next to a basaltic dyke. The amount of this REE transfer can unfortunately not be quantified because of the open system conditions. The REE distributions analyzed in the salts document a strong depletion of Ce, Pr, Nd, Sm and Eu with increasing distance from the contact and with increasing degree of fluid-rock interaction. This indicates together with the lack of a correlation with the mineralogical composition of the salt that the observed depletion is related to differential REE solubilities in the transporting fluid. A comparison of the fractionated REE patterns with available REE complex stability data suggests that the REE have been transported as chloride complexes. This finding implies that for a leaking HLW salt repository radionuclides could also under certain conditions (elevated T and P, high Cl concentrations) be transported as chloride complexes.
Some minerals are known to contain defects produced by natural radiations. They can record past occurrence of radionuclides in geosystems, providing unique information which is not accessible by petrographical, geochemical, isotopic analyses of present-day distribution of radioelements. Clay minerals as kaolinite and dickite are good candidates for such tracing, owing to their high contact area with alteration solutions and their sensitivity to natural irradiations. Several types of radiation-induced defects (RID) have been described in natural kaolinites by Electron Paramagnetic Resonance (EPR) (Angel et al., 1974; Muller et al., 1992). They mainly consist of hole-centers associated to oxygen atoms of the structure and are differenciated by their crystallographical location, the kinetics of their formation and their stability (Clozel et al., 1994). One RID, namely the A-center (Si-O- center), is stable at the scale of geological periods. A quantitative approach is proposed to the reconstruction of the past migration of uranium in the geosphere through A-centers dosimetry (Allard et al., 1994; Allard and Muller, 1998). Two sets of data are required for derivation of mean past U-concentration from the cumulative dose of natural kaolinites: (i) experimentally-determined parameters for A-center production by ionizing radiations (alpha particles, gamma rays...) in variably ordered kaolinites (ii) geochemical parameters of the studied site (time range of alteration periods, present-day concentration and distribution of uranium, relative contributions of natural radio-sources). This methodology was applied to the study of uranium deposits considered as natural analogues of nuclear waste repositories. Conclusions and prospectives will be derived from the study of weathered (soil) and hydrothermally altered geosystems (volcanic tuff, granite).
Allard Th, Muller JP, Dran JC & Menager MT, Phys. Chem. Miner, 21, 85-96, (1994).
Allard Th & Muller JP, Appl. Geochem, 13, 751-765, (1998).
Angel BR, Jones JPE & Hall PL, Clay Miner, 10, 247-255, (1974).
Clozel B, Allard Th & Muller JP, Clays Clay Miner, 42, 657-666, (1994).
Muller JP, Clozel B, Ildefonse Ph & Calas G, Appl. Geochem, 1, 205-216, (1992).
Potential matrix for waste radiative material disposal in deep geological sites, are present in the nature. They are mainly studied for their properties to trapped lanthanides and some actinides (U, Th) by substitution, to have a low rate of solubility and to resist to metamictisation under irradiation processes.
Among them, mainly phosphates (apatite, francolite), orthophosphates (monazite, xenotime), silicates (zircon) and tungstates (scheelite) are studied. Such matrix doped by REE activators are well know by solid-state physics for their photoluminescence properties and studied as luminophors, scintillators, dosimeters or laser materials.
The luminescence of activators centers in natural and synthetic samples is analysed by equipment enabling non destructive time-resolving spectra acquisition (ICCD coupled with delay generator). The system is combined with optical microscope enabling spectra acquisition at a micrometer scale. Such experiments allow to identified precisely the valence and the substituting sites of REE, homologues of actinides. Further more, their diffusion and their fixation in function of time under thermal energy (figure) or irradiation, can be follow-up in synthetically doped-samples.
Monazite is a rare-earth element (REE) phosphate commonly occuring as an accessory phase in granitic and pegmatite rocks, as well as in metamorphic and detrital sedimentary rocks. Thorium is commonly present up to 10 wt% ThO2 but may reach 31.4 wt%. Two principal substitutions were clearly evidenced from microprobe analysis of natural monazites:- the cheralitic substitution is the replacement of two REE3+ by one Th4+ and one alkaline earth element (e.g. Ca2+) according to the equation (1):
2 REE3+ = Th4+ + Ca2+ (1)
mainly observed in peraluminous leucogranites,- the huttonite substitution is characterized by the simultaneous replacement of one REE3+ by Th4+ and P5+ by Si4+ according to the equation (2):
REE3+ + P5+ = Th4+ + Si4+ (2)
mainly observed in alkaline and high K-calcalkaline granites.
For the cheralitic substitution, experimental work were reported in the two systems:-CePO4-(Ca0.5Th0.5)PO4 at 1400°C, 0.1 MPa (Hikichi et al., 1978) and-LaPO4-CaO-ThO2-H3PO4-H2O at 700°C, 100 MPa for which the complete solid solution between the LaPO4 and (Ca0.5Th0.5)PO4 end-members was synthesized (Podor et al., 1997). For the huttonitic substitution, only one experiment appeared to be available in the dry system CePO4-ThSiO4 with a small ratio x=Th/La <=0.5 at.% at 1400°C, 0.1 MPa.
The present experimental work is a contribution to the better knowledge of the huttonitic substitution. The two systems ThO2-SiO2-LaPO4-H3PO4-H2O and UO2-SiO2-LaPO4-H3PO4-H2O were systematically investigated. Solid compounds of the type La(1-x)ThxP(1-x)SixO4 (x = 0-1) were hydrothermally synthesized at 780°C, 200 MPa. A linear correlation between the crystallographic parameters (a0, b0, c0, ß0, V0) and the substitution ratio x (atomic%) was established. For the UO2 bearing system in the same T, P conditions and a fO2 control by the Ni-NiO buffer, the substitution is limited to x=0.13.
Results obtained have direct implications for the understanding of the substitution mechanisms in natural monazite crystals and for the use of compounds of the monazite type as matrixes for the storage of nuclear waste. Compared to other matrixes, phospho-silicate compounds have two major advantages: (i) low solubility in aqueous solution and (ii) high thermal stability.
Hikichi Y, Hukuo Kand Shiokawa J, Nippon Kagaku Kaishi, 12, 1635-1640, (1978).
Podor R, Cuney Mand Nguyen-Trung C, Amer. Mineral, 82, 765-771, (1997).
A multi-barrier system for secure radionuclides localization widely adopted now includes preparation of stable matrixes in which radionuclides are safely immobilized. Along with known glass methods an alternative materials were proposed: ceramic matrixes (with Ti, Zr, Ba etc), clays, basalts, zeolites. As was shown by the authors earlier high-temperature annealing of Cs-contaminated zeolitic tuff at 1000-1400°C results in essential immobilization of Cs. Resulting material contained amorphous phase and various quantities of Cs-bearing crystalline phases.The goal of the present work was to study Cs localization in the presence of other cations: Na, K, Ca, which are usually present in liquid nuclear waste of intermediate (ILW) and high (HLW) level activity.
Samples of various cationic composition: 1 meq Cs and from 2.0 to 7.0 meq of accompanying cations per 1 g of rock, were prepared from powdered (0.25-1.0 mm) clinoptilolite-bearing tuff. Concentrated (1N-5N) solutions of Cs, Na, K, and Ca salts or dry salts were mixed with the tuff in various combinations and dried for 4 hours. Then mixture was ground to powder and small cylinders (weight about 1.5 g and surface 5 cm2) were prepared in a press-form. Cylinders were annealed at 1000°C for 60 min and cooled with the furnace. Resulting phases were determined by X-ray diffractometry. Acid resistance was measured by the amount of cations went from the annealed cylindrical sample to the solution with the pH 1 (150 ml) during 24 hours at 25°C.
It was shown that addition of up to 7.0 meq of Na, K, Ca per 1 g of rock did not decrease the acid resistance of resulting annealed glass-crystalline composite. Even at pH 1 amount of cations went to the solution was less than 10-5g/cm3/day. Presence of K and Na favored formation of well crystallized pollucite-type phase.
Thus, natural zeolite-rich tuffs can be used for immobilization of Cs even in presence of large enough quantities (up to 2-3 meq) of Na, K, and Ca cations as is the case in many ILW and HLW.
Zirconolite is a major constituent of Synroc type nuclear waste forms and a principal host for actinides and certain fission products. Previous studies have shown that zirconolite is highly resistant to leaching in pure water at 70-150°C. The aim of this study is to investigate the stability of zirconolite at deep borehole conditions.
A 100% single phase, polycrystalline Nd-doped zirconolite-2 M (Ca0.8Nd0.2Zr1.0Ti1.8Al0.2O7, Nd as actinide analogue) was loaded together with 25 µl fluid in small electrolytically pure gold capsules. The ratio of zirconolite surface to fluid volume is 4-6 cm-1 at loading conditions. The gold capsules were welded to obtain a closed system. The leaching experiments were carried out in externally heated pressure vessels for 21, 63 and 189 days at elevated temperature (150-700°C) and pressure (50 MPa) for various fluid compositions (deionized water, NaCl, HCl, H3PO4, SiO2 and CO2) with various concentrations (0.1-0.001 M). Solid starting material and solid run products were both characterized by ESEM/EDX, EPMA, XRD and RAMAN-spectroscopy. The fluid run products were analyzed by three different methods, all using ICP-MS.
The degree of reaction is strongly related to temperature, molality of the fluid and time. Corrosion of zirconolite starts at crystal edges and grain boundaries, progressively embraces the entire surface and finally starts to replace primary zirconolite by secondary crystals. With increasing temperature and molality, secondary crystals (< 1 µm to 10 µm) form on both zirconolite surface and capsule wall, first as isolated crystals, then as aggregates, and finally covering the entire surface. Larger secondary crystals are well characterized, but were only observed at advanced corrosion stages; the smaller crystals require further study using TEM. In acidic environment, rutile and anatase are the predominant alteration products, whereas under basic conditions the secondary phases include orthorhombic Na-bearing perovskite, a calzirtite-related phase, and secondary zirconolite. Perovskite and "calzirtite" are Al-free and contain Nd in concentrations that are slightly higher than those in primary zirconolite. Preliminary data from the leachate indicate elemental fractionation between the fluid and the solid secondary phases. Future investigations will focus on determining the solubility of zirconolite under specific conditions. The solubility will be calculated from the elemental concentrations determined for the leachate, the secondary phases, and the starting material.
The results of this study demonstrate that zirconolite is corroded by dilute acidic and basic fluids at 50 MPa and at temperatures above 250°C. In a closed system, saturation of the fluid phase and steady state dissolution-precipitation of crystals is attained after a few days. Our data can be used to assess the long-term behavior of zirconolite-based nuclear waste forms in final disposal sites.
Calcium Silicate Hydrates (C-S-H) are major components of cement pastes, which may be used for various purposes in nuclear waste disposal sites. The structure of C-S-H presents some analogy with that of smectites, which are swelling clays with a great cationic retention capacity. These minerals both have a silicate layer structure, with the possibility to exchange cationic species in their interlayer space. Multinuclear solid-state NMR was used as the main characterisation technique to elucidate the structure of C-S-H in the presence of several alkaline cations. Cesium was mainly studied, as its two radioactive isotopes (135Cs and 137Cs) have long half-lives and could migrate easily in the natural environment. C-S-H were synthesised from silica, lime, and water, and were then added with cesium chloride. NMR allowed to distinguish several sites for the exchange of cations in the C-S-H, in relation with the Ca/Si ratio and the humidity yield of the sample. X-ray diffraction and elementary analyses were also used to help to interpret NMR results and to propose a model for the interaction between C-S-H and cesium cations. This model is based on the structural analogy with smectites and on cesium hydration considerations. At high humidity, cesium cations are totally hydrated and may be exchanging fast between the diffuse ion swarm and the surface sites of the C-S-H. When the sample is dry, cesium is partially or totally dehydrated in the C-S-H structure and is likely to enter the interlayer space, dilating the layers. Its capacity to be retained in the C-S-H structure depends on the Ca/Si ratio, since at high Ca/Si, the interlayer space is filled with calcium cations. This study leads to the conclusion that cesium can only be retained in the C-S-H structure at low Ca/Si ratio. The absence of isomorphic substitutions in the C-S-H layers may prevent cesium to be greatly retained in the C-S-H structure, as it is the case in smectites. Further studies are underway to increase the retention capacity of C-S-H towards cesium, in order for cement to be more effective in nuclear waste disposal.
High resolution images of the borehole wall can be obtained with different borehole scanner tools. New digital, optical core scanner systems provide colour images of the full circumference of cores with a resolution of 0.1 x 0.1 mm. The correlation of the main structural features - like bedding planes, fractures, foliation, and shistosity - both from the borehole wall image and from the core image is the most reliable and cost-effective method of core orientation. Once the core is orientated all detailed structural evaluation of core data, and all laboratory measurements on cores to determine physical, hydrogeological and engineering properties can be assigned to true directions. Taking into account at least two components of the tensor of properties is essential for fluid flow calculations and rock stability assessments, as most rock masses are anisotropic. Also, anisotropy has to be considered when measurements on rock samples are used to calibrate log responses and to derive lithological, hydrogeological and engineering properties from geophysical logs. The existence of anisotropy and their effects are best studied on core samples, as anisotropy is normally not revealed by logs. Compared to detailed but sparsely sampled core analysis, geophysical logs complementary offer the advantage of continuous information, determination of true in-situ properties, filling the gaps of core loss intervals, and providing information from cheap percussion boreholes. Moreover, careful log interpretation can help to detect thin, weak layers, which are often not present in the core data. But these thin layers can have a major impact on rock mass stability. Automatic log processing using Fuzzy classification schemes is a well proven, objective method to derive lithologies from geophysical log data. A major advantage of this approach is that with each classification a level of confidence is given. Combining these results with those from core data helps to assess the lithological interpretation: in all intervals where the confidence level is high the interpreter is freed from tedious work but rather he can concentrate on those intervals of ambiguity. Through the lithological classification from logs each lithology class is automatically assigned to a number of physical properties. Within a certain lithology the scatter of the data around the prototype of that lithology can vary in a typical way. Looking out for such characteristic pattern in the physical property logs is often the best way of a lithological correlation between two boreholes. In a final step, hydrogeological and engineering properties can be derived from logs, if the results from laboratory measurements, the in-situ measurements of physical properties and the lithology related corrections are adequately considered. The benefits of combined core and log interpretation is demonstrated by several field examples.
The installation of high-level nuclear wastes will give rise to heat and fluid flow, leading to a change of the ambient relative humidity in the gas phase environment of the waste container. The relative humidity is a key parameter controlling waste isolation because high relative humidity enhances the rate of corrosion of container materials. Furthermore, the relative humidity is related to the activity of liquid water, which is an important parameter controlling the water-rock interactions. Assuming a constant composition for the gas phase, it has been suggested that a strong thermal gradient between a hot waste container and cooler drift wall should lead to a reduction in the relative humidity at the waste container surface compared to that in the backfill material (because saturated water vapor pressure increases with temperature). The assumption that the composition of the air plus water vapor gas mixture is the same at the waste container surface and at the drift wall results from the assumption that the air-vapor binary diffusion can be described by Fick's law. However, diffusion also occurs in response to a thermal gradient. The coupled thermal diffusion effect contributes to the establishment of concentration profiles in the gas phase. This effect is generally ignored relative to Fickian diffusion. Estimates of the thermal diffusion constant for air-vapor mixtures permit the magnitude of the thermal diffusion effect to be estimated using conventional Onsager relations. For the thermal gradients and geometries approximating an air gap adjacent to a high level waste container in a hypothetical geological repository, the thermal convection of the gas phase (thermogravitation), and optimized geometries are considered, a significant vertical separation of water vapor and air can occur. For geometries including a vertical container surface and an air gap on the order of 2 cm wide and 600 cm tall, the relative humidity at the top of the convective system is up to 30 percent higher than when the gas mixture between a waste container and nearby backfill is assumed to be homogeneous. Such a difference may have important consequences, in particular for the rate of corrosion of the waste container. Maintaining geometries to preserve an air gap of 5 cm or more eliminates significant thermal diffusion; or geometries (such as no air gap) that prohibit free convection over relatively long vertical distances (e.g., meters) would eliminate thermogravitational separation of water vapor and air.
Extended geochemical and mineralogical studies of the Oklo reaction zones (Franceville basin, Gabon) have shown that the functioning of the natural nuclear reactors have induced strong disturbances in host sandstones. Hydrothermal circulations around nuclear reaction zones have locally transfered fissiogenic products such as Nd and Cs (Sère, 1996) or Pu (Bros, 1993). To better assess the migration related specifically to the functioning of the reactors zones, the behaviour of monazite and zircon was studied in the near field of OK84bis (Okélobondo mine), RZ 7-8-9 (Oklo open-pit) and Bangombé reactors zones and far-field (BA.2-900 m deep bore hole).
Several stages of monazite alteration were observed: (1) non altered monazites (80 to 96% LREE[PO4] with 4 to 20% huttonitic substitution), (2) corroded monazites mixed with Th-silicate alteration phases (60-40% LREE[PO4] - 40-60% (U,Th)[SiO4]), (3) totally altered monazites, the residual phase corresponding to a poorly crystallized Th-silicate (< 20% LREE[PO4] and > 80% (U,Th)[SiO4]). The Th/U ratio increase from monazite (Th/U=20) to the Th-silicate phase (Th/U>50) is interpreted as an alteration by oxidising fluids with leaching of U together with LREE. The alteration is probably synchronous with the siliceous diagenesis. More than 90% of monazite crystals present alteration features. Galena was observed in cracks and as inclusions mainly in In altered zones of monazite crystals but sometime also in non altered monazites.
In zircon crystals, the cores are generally weakly fractured, without galena inclusions. Their composition correspond to the pure end-member (Zr,Hf)[SiO4]), poor in Th and U (Th/U ~ 1). Whereas, the rims, with several growth zones, are enriched in REE, P, Si Th and U with higher Th/U ratios (5-10). The alteration of the zircon lead to the new-formation of a Zr/LREE-Si/P phase in cracks fillings and/or along growth zones. Galena inclusions are also common in such external growth zones of zircon crystals.
The volume of newly formed galena (2 to 14 vol.%) largely exceed the amount of lead (< 0.1 wt.%) which was produced by radioactive decay of U and Th in the accessories. Most of the lead has been introduced during the diagenetic event from external radiogenic (208Pb/206Pb = 7000 for the near-field OK84bis sample) and/or non radiogenic sources.
Bros R, Mem. Thesis, Un. Strasbourg, 154, (1993).
Sère V, Mem. Thesis, Un. Paris VI, 226, (1996).
Brabantite CaTh(PO4)2 is one of the end-member of the monazite mineral. It is one of the most efficient way to incorporate actinides in monazite (LREEPO4, monoclinic P-2-1/n, four formula per cell). We studied various compounds similar to brabantite in order to determine the possible substitutions in that structure. For synthesis, the best results were obtained by hydrothermal reaction of phosphates (750°C, 2.5 kbars), by dry reaction of a gel at 1 atm. and high temperature (1200°C) or by Mo-fluxes growing at moderate temperature (900°C). In additions to brabantite s.s., we obtained similar structure by replacing Ca by Cd, Sr, Pb, Ba. On the contrary it was not possible to replace Th by other 4+ ions Zr and Hf. For the smallest (Cd) and biggest ions (Ba) we also obtained other structures suggesting that we are here at the limit of the monazite structure. The cell parameters of the monazite structures increase regularly with the increasing size of the divalent cation. For example, the volume increases from 287 Å3 (CdTh(PO4)2) to 324 Å3 (BaTh(PO4)2). We are currently studying the solid solutions between LaPO4 (monazite s.s.) and the various brabantite. The first results indicate that the solid solution is total between LaPO4 and SrTh(PO4)2 but that there exist a miscibility gap at 1 atm. beween LaPO4 and BaTh(PO4). This is expected because at 1 atm. BaTh(PO4)2 does not display a monazite structure. We show by this study that the monazite is a ''soft'' structure, accepting ions with large difference in ionic radius (1.10 A for Cd to 1.42 A for Ba). Therefore monazite can accept ions with different charges (2+: Ca, 3+: REE, 4+ Th, 5+ P) and different size. Because monazite is also resistant to weathering and to radiation damages, we think that it should be studied as a possible matrix for radioactive waste.
Experimental studies have been undertaken in order to provide new insights on the relative efficiency of the different mineral phases and sorption processes for the control of U retention in the weathered zones surrounding the natural nuclear reactor at Bangombé (Oklo, Gabon). Clayey and Fe-oxihydroxides rich samples from the oxidizing weathered zones located above the reactor were examined. An experimental study of uranyl adsorption/desorption processes in these samples was carried out using an uranium isotope exchange technique in order to estimate the proportion of uranium adsorbed on mineral surfaces. The major U-containing minerals in the samples were also identified using a sequential extraction technique. It is found that, in the U-rich iron crust rocks close to the reactor, uranium is mainly associated to iron oxihydroxides and ferric phosphate phases, whereas manganese oxides play a minor role in U retention. For clayey samples in the weathering profile, it may be inferred that several processes account for U retention: ion exchange on clays, adsorption processes occuring mainly at clay surfaces, incorporation into mica and amorphous iron oxihydroxides structures. A large proportion of total U is adsorbed at mineral surfaces and is also in exchangeable position on clay minerals like illite, and is thereby easily accessible to weathering solutions.
In a second part of this work, 233U sorption data obtained on an illitic Bangombé sample were modeled using a surface complexation modeling approach. As a first approximation, it was assumed in modeling that uranyl binding occurs at aluminol edge sites of the illite component in the sample. The binding constant required for modeling was firstly determined for the non-electrostatic model (NEM) from experimental work on the U(VI)/hydrargilite (<alpha>-Al(OH)3) system. The fractional uptake of U(VI) on hydrargilite was measured as a function of pH, for different ionic strength values (0.1 and 0.01 M) under CO2 - free conditions, and in solutions with a total concentration of carbonate of 10-3 M. In the surface complexation reaction pK value equal to -0.8±0.2 has been used. For the NEM (I.S.:0.01 M) it provided an accurate description of the pH-dependent sorption of U(VI) on both a well-characterized hydrargilite and on the clayey sample of Bangombé. These results highlight the potential use of simple surface complexation models to the description of uranyl sorption on complex natural samples.
The SHRIMP (Sensitive High Resolution Ion Microprobe) is a powerful tool to obtain in-situ isotopic information on Pb, U, Nd and Sm in different mineral phases (e.g. galena, uraninite, pyrite, hematite, manganese oxides, organic matter and fracture fillings). The isotope composition of these elements can be measured with good precision (Hidaka, 1998). This technique has been applied on minerals in selected thin sections (drill core BAX 08) of the natural reactor site of Bagombé for dating and tracing of fission products. Lead-lead dating gave ages of aprox. 2.2 Ga for most samples (galena) of the FB- (weathered clays) and FA-Formation (sandstone) reflecting the age of source rock and mineralisation. Organic matter and uraninite minerals in the FA-formation below the reactor zone were depleted in 235U having a ratio of 235U/238U as low as 0.00664±.00014 (2 <sigma>). They were accompanied with uraninite minerals with a normal 235U/238U ratio of 0.00725. A small enrichment of 235U/238U (0.00744±00014) was observed in some uraninite minerals of the FA-formation, if fractionation is not considered. The FB-formation above the reactor zone was not depleted or enriched in 235U/238U. Fissiogenic Nd and Sm were detected in the reactor clays surrounding the reactor, in zircons and in a U/REE-mineral located in a fracture filling 3 m below the reactor zone. The 149Sm/147Sm ratio was 0.874±006, which indicates fissiogenic and neutron capture products (normal: 0.921). Using in-situ isotopic analytical technique it was possible to detect isotopic anomalies outside of the reactor core and to trace fission products. The presence of depleted and normal uranium in the same thin section indicates two generations of U - mineralisation. The small enrichment in 235U/238U has to be confirmed. The detection of a mineral containing fission products in a fracture filling of hematite below the reactor indicates migration and formation of minerals after the fission reaction had ceased. This shows the importance of fractures for redistribution of fission products in a "natural waste deposit".
Hidaka H, Radiochimica Acta, in, press, (1998).
Uranium ores can be used as analogues of nuclear waste in a geological environment. Oklo, Gabon, is a unique type of natural analogue due to the occurrence of natural, fossil fission reactors in a highly concentrated uranium ore. The reactor zones can be seen as fuel that was used ca 2000 Ma ago when the natural reactors operated. Thus there are possibilities to study the behaviour of this fuel, and the efficiency of the barriers, on a geological time scale. Comparing the barriers in the model of a nuclear waste deposit to the Oklo reactors, there are analogues to the fuel matrix, the clay halo and the surrounding bedrock found in Oklo. In this study the focus lies on the first barrier, the uraninite matrix. How has the geological environment affected the Oklo uraninite during the last 2000 Ma? To resolve that, uraninite from reactor zone 16 has been studied with a Cameca 1270 ion probe at the NORDSIM facility at the Swedish Museum of Natural History in Stockholm.
The episodes of element mobility in the Oklo area can be studied using the U-Pb system of the reactors, since Pb produced from the radioactive decay of uranium is easily lost from the uraninite lattice. All distinct episodes which could cause episodic Pb loss from the uraninite lattice are likely to have been episodes of migration of other elements found in the fuel, such as fission products and actinides. In situ studies of the U-Pb system on uraninite show there are slight within grain variations of 235U/238U and 207Pb/206Pb. The 207Pb/206Pb ratios appear to be normally distributed around a maximum. This maximum could be said to represent the time of one episode of lead loss. However, to date this event, one has to know the extent of the lead loss. More complications have to be considered if there were several episodes of lead loss. Studies of Pb isotopes on galena indicate the more complicated scenario.
A better way to study the U-Pb system in uraninite is to use not only the Pb isotopes but to measure daughter/parent ratios of the both uranium isotopes and plot it on a concordia diagram. These studies would put some light on the lead loss history of the uraninite. However, inter-element ratios measured on an ion probe requires a standard. Uraninite, often discordant and inhomogenous, is a difficult mineral for standardisation on the ion probe. A possible preliminary standard for uraninite analyses at the NORDSIM facility has been studied for further analyses on Oklo uraninite.
The high grade uranium ore (up to 50% U) of the core of the 1.98 Ga old natural fission reactors of Oklo is made of uraninite. During the first stage of the fission reaction operation, the heat due to the reactions induced an intense hydrothermal circulation of fluids in and around the reactors. This has been responsible for the migration of major elements in the hosted sandstones: major loss of silica and gain of U and other elements such as Mg and to a less extend Fe and Al. During this hydrothermal event, the reactors acted as «open system» resulting in uranium enrichment which conducted to the formation of the rector core and to the crystallisation of newly formed clays which forms the gangue of the core of the reactors, mainly made of pure chlorite. When the clayey gangue reaches some importance, then such system forms a nearly «closed system» for actinides and most fission products trapped in the uraninite. Oxygen isotopes studies of both uraninite and chlorite are used to depict the conditions of water/rock interaction during criticality. Investigations were carried out on reactor 10 which is known as remaining preserved from any late fluid alterations. The range of 18O of chlorites is very wide (about 15‰). In most cases, chlorites are in isotopic equilibrium with the diagenetic fluids which percolated through the reactor at the time of the fission reaction operation, during the «open stage». Some chlorites record the high temperature which occurred during the fission reaction whereas others crystallised at lower temperature, during the cooling stage. However, strongly 18O depleted chlorites yield unrealistic high temperatures. Such strong depletion probably reflects equilibrium with fluid which has reacted with the strongly 18O depleted uraninites of the core of the reactor. We assume that the alteration of uraninite may be related to water radiolysis process (Le Bricon et al., 1999). After the closure of the reactor core by the impermeable clayey gangue, radiolysis of the water at the contact of uraninites grains has allowed: (1) the oxidation of uraninites in grain boundaries due to oxygen formation and (2) local new crystallisation of uraninite in H2 rich zones, in isotopic equilibrium with 18O depleted water. Oxygen isotopes investigations allow to distinguish three different steps during the natural reactors genesis: (1) «open system» during reactor operating (2) progressive closure of the reactor by clay crystallisation, at the end of the reactor operating and during the cooling stage (3) isotopic exchanges in the «closed system» between minerals and 18O depleted fluid having reacted with uraninite due to water radiolysis.
Le Bricon et al., J., Conf., Abs., 4, (1999)
The uranium deposits of Gabon (Oklo and Bagombé) host high grade ores in which nuclear fission reactions occurred 2.0 billion years ago. The fossil reactors represent natural analogues for modern nuclear waste and therefore offer the unique opportunity to study the processes controlling the containment or migration of actinides and fission products in a geological system and to establish analogies with nuclear waste repositories. The Bagombé reactor zone is located at a depth of 10.80 m and characterized by a slight depletion in U-235 (235U/238U = 0.006) and a high U-content (54 wt%). Due to the shallow location, the reactor zone has been affected by supergene weathering. In this study, a cross-section of four drill-cores from Bagombé has been investigated. Special attention has been paid to the mineralogy and the U and REE geochemistry. In order to determine the redistribution and mobilization of major elements, REE and U in the lateritic profiles, mass balance calculations have been performed. The results indicate that weathering has a great influence on the mineralogy as well as on the elemental distributions. Different physico-chemical environments could be identified in the profiles. To obtain information about the mechanisms responsible for the migration/retention of U and REE under different geochemical conditions, sequential extraction experiments were carried out with special emphasis on the role of secondary mineral phases (e.g. iron oxihydroxides, manganese oxides, clays and phosphates). The results of the experiments suggest that the mobility of U is controlled by sorption/desorption processes under oxidizing conditions, and by dissolution/precipitation reactions under reducing conditions (Del Nero et al., 1999). Concerning REE, distinct behaviours are encountered within the different zones of the profiles. Most REE are leached in the upper part of the profile (soil) and immobilized in lower units by different mechanisms (retention in primary minerals resistant to weathering, incorporation in secondary amorphous or well crystallized phases and/or adsorption by clays). In addition, a «fractionation» between HREE and LREE is observed in the rocks. Whereas in general the LREE are enriched (normalized to the unweathered rocks), the HREE are depleted. This depletion in HREE is explained by the preferential complexation and mobilization of the HREE by weathering solutions (Stille et al., 1999). Finally, isotope analyses allowed us to determine the contribution of fissiogenic REE to total REE. Significant deviations from natural values are observed in the vicinity of the reactor. Anomalous ratios of 149Sm/147Sm near the reactor zone reflect the presence and migration of fissiogenic REE.
Del Nero et al., J. Conf. Abs., 4, (1999)
Stille et al., J. Conf. Abs., 4, (1999)
The Bagombé natural nuclear reactor (Gabon) is located at 11 meters depth. It undergoes surface weathering conditions which occur under an equatorial climate. The reactor is located at the top of a fractured sandstones formation. It is overlain by a 50 cm thick oxidised layer, made of a red pelite and Fe-Mn oxides rich crusts, which only occur in the bore-holes cross cutting the reactor. This oxidised layer is overlain by 3 meters thick of more or less weathered black shales. Then, the top of the profile is a 7 m thick lateritic profile characterised by a succession of alterites.
This study is performed to find out the origin of the oxidised layer (red pelite containing fission products and oxides extracted from the crusts) which is located between the reactor and the weathered black shales. Is the oxidised layer related to the lateritic profile or does it correspond to an ancient crust? Therefore U-Pb system and oxygen isotopes are used respectively for dating and for determination of geochemical conditions of crust genesis. That is to constrain migration of the actinides and fission products from the reactor core to the surrounding host rocks.
The oxides extracted from the crust yield a 18OSMOW of 5.7‰. Such isotopic ratios support crystallisation of iron oxides in equilibrium with a diagenetic fluid (between 5 and 8‰), at temperatures ranging between 60 and 130°C. In a U-Pb diagram, the data points of the red pelite and the oxides extracted from the crusts plot above the concordia. This shows a loss of uranium and/or a lead gain. The upper intercept of the discordia gives a formation age of 1.97 Ga, which corresponds to the age of the fission reactions. The data of an uraninite from the reactor fit on this discordia, but plot beneath the concordia. The lower intercept agrees with an event that affected the uraninites and the oxides at the same time. The apparent 207Pb/206Pb ages of the red pelite (1.97 Ga) and of the oxides of the crusts (between 2.06 and 2.12 Ga) are similar and correspond to the uranium mineralisation event and to the fission reactions.
These results indicate that the oxidised layer overlying the reactor was not formed by recent weathering. Instead the oxidised layer may have been formed during the fission reactions by radiolysis of the water. The oxidised layer underwent a second event later leading to an uranium loss or lead gain. Other mineralogical and chemical evidences support this conclusion. Furthermore, this isotopic study shows that iron oxides are stable in the U-Pb and oxygen isotopic systems for a long period of time and that they can therefore be used for dating old oxidations events.
Subrosion is the process of subsurface dissolution of rock salt due to groundwater flow. Part of the research programme on the Morsleben repository was directed to the long-term effects of subrosion on the host-rock rock salt. The repository is located in the Allertal salt structure.
To estimate the variation in subrosion at the Allertal salt structure, a study was performed aimed at coupling the variation in groundwater flow in the past 200,000 years with variations in subrosion rate at the salt structure. For this purpose a local 2D-groundwater model at a scale of 3 km, capable of simulating density-driven groundwater flow, was constructed. One of the boundary conditions is the dissolution of rock salt at the interface of the salt structure and the groundwater flow domain. The other boundary conditions are given by a regional 3D-groundwater model of the Subhercynian Basin (on a scale of 100 km), which in turn is constrained by conditions derived from a study of supraregional groundwater flow for North-western Europe. This telescopic approach is very suitable for simulating small-scale processes (subrosion) influenced by processes taking place on a much larger scale (climate).
The influence of six climate periods on the subrosion rate was analysed. The conditions in these periods are: warm-temperate, discontinuous permafrost, continuous permafrost, advancing ice sheet, complete glaciation and retreating ice sheet. The conditions during these periods were derived from reconstructions of palaeo-conditions during the Saalian glacial (Van Gijssel, 1999). Using the local groundwater model it could be derived that subrosion is lowest during interglacial, warm-temperate periods (average 0.28 µmm/year) and highest during periods with continuous permafrost (average 5.3 µmm/year). High subrosion rates coincide with a high level of infiltration above the salt structure, resulting in the transport of large quantities of brine. It is therefore concluded that the variation in subrosion rate at this site is mainly influenced by the capability of the groundwater system to discharge the brine at the interface of the salt structure and the flow domain.
Van Gijssel K, J. Conf. Abs., 4, (1999)
Within the scope of the research programme on the Morsleben repository a prediction of possible effects of fluvial erosion and sedimentation in the future 150,000 years was made. The prediction was based on numerical modelling of the erosion process. An elaborate reconstruction of past fluvial dynamics was carried out using all available geological and geomorphological data. Climate reconstructions were based on regional data from the Schöningen excavation and European correlations. Tectonic reconstructions were based on subsurface data and morphological interpretations. Tectonic uplift or subsidence rate data were very limited and vary within a considerable range of possibilities.
With the help of the FLUVER2 package a 2D erosion model for the simulation of the longitudinal dynamics of the Aller, a tributary of the Weser, was constructed. Calibration of the model was based on the palaeohydrological reconstruction of the past 200,000 years, which overlaps the post-Saalian period (from oxygen isotope stage 6 onwards). The calibrated model was subsequently used to simulate possible future dynamics using an extrapolated SPECMAP-curve and linear extrapolated tectonic rates for the next 150,000 years. Depending on the tectonic scenario chosen, the simulations yielded net erosion depths of 3 m with a slow uplift rate up to 11 m with a fast uplift rate near the Morsleben repository. Regional semi-3D-simulations of valley development suggest that the fluvial incision at the repository site itself will be probably less deep than predicted with the 2D-model. The large range of uncertainties involved in the semi-3D-modelling do not allow for exact quantification of this model.
Within the scope of the research programme on the Morsleben repository indicative palaeoclimatic and -environmental boundary conditions for geohydrological model studies (see Bremmer et al., 1999) were inferred from regional Quaternary litho- and biostratigraphic information.
For this purpose a genetic stratigraphic framework of the Subhercynian Basin was reconstructed on the basis of superposition, correlation and dating, including:- sedimentary units generated in glacial depositional environments (Saalian and Elsterian tills and glaciofluvial deposits);- sedimentary units generated in subaerial periglacial depositional environments (loess and solifluidal deposits);- alluvial deposits produced by the river systems draining the southern uplands (Harz Mts.);- lacustrine and organogenic deposits preserved in small-scale basins.
The sedimentary sequences and their unconformities then were interpreted as products of (cyclic) palaeoclimatic events which are related to ice-sheet expansion, permafrost distribution, periglacial deserts with loess deposition, vegetation climaxes and fluvial response (aggradation/incision).
The compiled sequence of events and palaeoclimatic conditions was assumed to provide a reasonably reliable index, from which palaeogeographical and palaeo(geo)hydrological conditions in the study area could be reconstructed for modelling purposes. Major hydrodynamic changes occurred during the twofold glaciation maxima in the Subhercynian Basin when subglacial and ice-marginal environmental conditions prevailed.
Because of the fragmentary nature of the local successions and the restricted chronometric control, trends in the (orbitally- and polarity-tuned) marine oxygen isotope (OI) record were used as a guide for the timing and patterning of the European glaciation events. This global correlative framework allowed to confine further the (chrono-)stratigraphic position of the regional (peri-)glacial sequences and intermediate non-glacial sequences. Locally-controlled lake and mire deposits found in several subrosion depressions are immediately post-dating the OI-fixed glaciations implying that subrosion probably was related to glacial environments.
Bremmer CN, Kreft E, Rijkers RHB, Simmelink HJ & Van Weert FHA, J. Conf. Abs., 4, (1999)
The Palmottu U/Th ore is located in mica gneisses and pegmatitic granites in south-west Finland. We studied 238U-230Th disequilibrium in calcium carbonate veins within fractures related to this ore, as a contribution to the EEC Palmottu natural analogue project, on nuclear waste disposal in cristalline bedrocks. Four drill cores have been sampled along an east-west profile across the ore. Fractures filled with calcite with low amount of clay minerals were selected both in hydraulic conductive sections and in apparently dry sections of the cores. Due to the very small size of the fractures (0.05 to 1 mm), only small amounts of sample were available for analysis (10 to 100 mg). SEM and EPMA analyses of the samples reveal a complex mineralogy, consisting of calcite associated with clays, sulfide inclusions and diffuse uranium silicates (coffinite). U/Th and lead isotope measurements were performed by TIMS on bulk vein samples, on a VG 54-30 at CEREGE (Aix en Provence). U and Th concentrations measured in these samples range from 0.2 to 2500 ppm and from 0.08 to 34 ppm respectively. 234U/238U activity ratios range from 0.9 to 1.8 and (230Th/234U) from 0.002 to 1.32. Uranium is mainly located in the uranium silicate inclusions, while the concentration is below EPMA detection limit in the calcite.High uranium concentrations are found only close (i.e. within 10 meters) to the uranium orebody itself. A Pb/Pb pseudo-isochron at 1882-1925 Ma suggest the hydrothermal origin of the fracture fillings, simultaneous to the formation of the ore. 234U and 230Th excess measured in most samples reveal uranium accumulation in the matrix of the fracture fillings during the last 600 000 years. On four samples out of twelve, U/Th ages can be calculated without detrital 230Th correction (i.e. 230Th/232Th > 1 10-4). The youngest age (2417 ± 222 years) is found in one thin silicate layer isolated in central position within the calcite vein. These results suggest a recent crystallization of coffinite within the veins, related to uranium mobility in percolating fluids. This conclusion is also sustained by the relatively low 206Pb/204Pb ratios (48<206Pb/204Pb<68) measured by SIMS in these minerals (Pomiès et al. 98), much lower than expected if the coffinites were contemporeneous to the ore formation.
Pomiès C, Hamelin B, Lancelot J & Casanova J, Mineralogical Magazine, 62A, 1200-1201, (1998).
In nature, fault patterns exhibit a very broad distribution of fault length that is generally modeled by a power-law distribution, n(l)=<alpha>l-a. Such a distribution has no characteristic length scale and does not allow to define any a priori pertinent scale of observation for hydraulic field experiments in fractured media. To investigate the relative effects of faults depending on their length, we undertake in the spirit of percolation theory a theoretical and numerical study of the connectivity of three-dimensional fault networks following power-law length distributions.
We first formulated an analytical expression of the percolation threshold, based on the distinction between small faults and macro-structures whose lengths are larger than the system size. We numerically checked this expression by means of simulations at different system sizes, and showed that the correct expression of the percolation threshold depends on the third moment of the length distribution. Different regimes of connectivity were identified depending on a, the exponent of the fault length distribution. For a>4, faults smaller than the system size rule the network connectivity, and classical laws of percolation theory apply. For 1<a<4, both small and large faults control the connectivity of the system, in a ratio which depends on a. In that case, the probability to observe large faults increases with the system size, so that the connectivity increases with scale whatever a given scale-invariant fault density. Therefore, the percolation threshold should be always reached for a sufficiently large system size. The scale at which the percolation threshold is reached depends on the fault density and the exponent of the power-law length distribution. The geometrical properties of the infinite cluster, which partly control transport properties, are also characterized at the percolation threshold. A theoretical relationship between the fractal dimension of the infinite cluster, its length distribution, and the basic exponent a, has been established and numerically verified. The definition of the length scales governing the connectivity properties of three-dimensional faults networks may be of great help to understand and model experimental results of fluid flow in fractured media.
The trapping of toxic elements is of great interest because of the new kinds of wastes to be treated (liquid effluents or alteration of solid wastes). One challenge for future is to try and clear of pollution liquid phases in an irreversible way. So this work is devoted to the study of the trapping of such elements by some mineral phases (apatite).
The alteration tests were conducted on 2 glassy matrixes in an oven at 90°C between 1 to 56 days. The glass composition is close to one of a bioglass (CaO-SiO2-P2O5-Na2O). Two reacting solutions were chosen: one with strontium and the other with lanthanum.Solution aliquots were periodically sampled to determine elements concentrations by ICP. Surface samples were analysed by DRX and electron microscopy (SEM and TEM)
The concentrations of Sr and La in the solutions decreased with time in an important way, up to 99% of the initial concentration for Sr and to 95% for La. After 1 month of alteration, the samples surfaces were covered by an alteration layer whose external part was covered by well crystallised apatite, composed of Ca and P and Sr or La (up to 40 wt% of Sr and 35 wt% of La).From these results, it can be deduced that incorporation of Sr and La is not related to a cationic exchange between solutions and neo-formed crystals, but rather to an incorporation of these elements in the crystalline network of the apatite during the growth phase.
The Chelyabinsk - 65, which is a unique and one of the world's largest depositories of low-level liquid radioactive waste (LLW), has been operated for decades. During the creation of armament plutonium at the Mayak plant, the acute problem of utilization and storage of low-level liquid nuclear wastes arises. The technology allows to discharge them into the open storage reservoirs without hydroisolation of reservoirs, whereas in the USA, the bottom and sides of reservoirs are cemented or paved with asphalt. During the exploitation of storage reservoirs the danger of their overfilling or out-break arises, as well as of the filtration through their body and foundation. The problem will be less acute, if the ways of rational reservoir control are found. The alternative economically acceptable and environmentally substantiated ways and technologies for the storage of low-level wastes are still absent in Russia. The reservoir cascade at Techa River is created because of the necessity to store the low- and mediumlevel industrial wastes of Mayak plant. During the decades Russian scientists have carried out long-term observations of the interaction of large open LLW reservoirs with surface and ground waters. Influence of meteorological factors on the water balance of the reservoirs has also been studied. The hydrological regime of the reservoirs region was studied. The mathematical model of the Techa River radioactive water reservoirs was constructed. Laws of the long-term changes of the balance elements also were analysed. Vari-statistical model for reservoirs was developed and forecast of water levels change in them was carried out.The analysis of the Techa River Radioactive Reservoirs water balance shows that the information on volumes of groundwater receiption into the reservoirs and value of filtration losses from them is the most poorly investigated. At the same time this part of the water balance can be large and sometimes close to surpass values of radioactive waste leakages. Hydrological-filtration model can serve as independent tool for fulfilment of expert valuations connected with the question on interrelation between researched reservoirs and groundwater. The created model helps to explain the dynamics and directions of the processes and allows to specify balance accounts.
A study of calcite veins induced by compressional tectonic activities in the West gallery of the Tournemire tunnel (Aveyron, France) (Mathieu et al., 1997) has concluded that microfissuration most probably favored local and limited fluid migrations. A later opened gallery in the East side has revealed different calcite veins, with especially abundance of geodic facies. A combined petrological, mineralogical and geochemical study shows that in the fluid responsible for the formation of the geodic calcite there is a component which was not observed in the preceeding study. SEM-BSE-EDS studies indicate that there are several generations of calcite with significant differences in Fe-Mg and Mn contents. The 13C>PDB values < -2.5‰ for geodic calcite, in the East gallery, are comparable to values obtained on Aalenian calcite whereas the 13C>PDB values for vein calcites, in the West side, have values similar to those of carbonates in the shale. Geodic calcites also contain less REE, have a slightly higher 87Sr/86Sr ratio and show much more chemical variations than the calcite veins in the West gallery. These data suggest that fluids issued from the above Aalenian formation have circulated in the Toarcian shale fractures. It is noted that these shales were highly indurated and that the fracture network has probably controlled the paleocirculations.
Mathieu R, Pagel M, Clauer, De Windt L, Cabrera Jand Boisson JY, Terra Nova, 9, 298, (1997).
Much interest is presently devoted to the study of radionuclides in geological formations. In this note, we report results obtained on the mobility of helium and argon in sedimentary rocks. Some of their isotopes, for example 4He and 40Ar, are produced in-situ by the radioactive decay of long-lived natural nuclides (238, 235U, 232Th and 40K) and the retention of natural radiogenic isotopes over geological time give insights into the mobility of elements having similar dimension. Therefore, we compare 4He and 40Ar abundances measured directly in samples of sandstones, shales and limestones from the Paris Basin with amounts which should have accumulated in these rocks. For all lithologies, we observe that 40Ar abundances (0.6 - 16 x 10-10 mol/g) are very close to those produced by the decay of 40K (0.1 - 7,6 x 10-10 mol/g). It is clear that Ar is retained in the rocks. In detail, 40Ar concentrations are sometimes higher (< 2) than those derived from the production rate. Such excesses may correspond to inherited radiogenic 40Ar in the detritic minerals. In contrast, comparison between observed 4He abundances (0.08 - 3.2 x 10-10 mol/g) and those which have been produced by radioactivity (11 - 66 x 10-10 mol/g) show that sedimentray rocks lost more than 95% of their in-situ produced 40He.
The extreme mobility of He lead us to study helium diffusion in shales. The experiment consisted in measuring He amounts liberated at room temperature first. Shale samples recovered from a borehole where loaded in the site in a stainless steel container and immediately pumped. After 1 year, the amount of helium released at room tempreature was analysed. Then a 8 mm-sphere of the sample was heated sequentially between 200°C to 1000°C. Different helium components are identified. 50 to 65% He is very mobile and is liberated at 25°C. The rest is recovered during stepheating experiments, which allow to diffusion coefficients from an Arrhenius. 3 types of diffusion profiles are observed : the first profile between 200 and 400°C leads to E = 10.4 kcal/mol and could correspond to the diffusion of He located in boundaries of the rock. Such a He component is probably associated with the porosity water and, possibily, with hydroxyl group in the sheets of the shale. The second profile between 400 and 600 with E = 27.4 kcal/mol and the third profile between 600-1000°C with E = 4.9 kcal/mol could result from diffusion of He from different minerals or from different grain sizes in the rock.