B. Neerdael, G. Volckaert, A. Sneyers
SCK•CEN, Mol,
Belgium
ABSTRACT
Disposal of properly conditioned long-lived radioactive waste in geological formations is the most promising and realistic option in this respect. The R&D program on geological disposal for high-level and long lived waste was initiated in Belgium in 1974 by the Research Centre for Nuclear Energy (SCK•CEN). A deep Tertiary clay formation, the "Boom clay", present under the Mol-Dessel nuclear site, was selected as a candidate host formation for experimental purposes; the research program and assessment studies were therefore developed according to a site- or at least formation-specific approach, without however anticipating the conclusions of a more detailed site selection procedure.
The current R&D program is mainly focused on site characterization (mechanical, physico-chemical, and hydrogeologic properties) and waste packages (characterization and compatibility in clay). Performance assessment studies lead to the identification of R&D priorities for the running research programs on the different components of the disposal system.
Rather early in the program (1980) an underground laboratory was constructed in the clay at the Mol site at a depth of 223 metres. The underground laboratory was intended to evaluate the technical feasibility of such a construction and to become an in situ facility for performing tests in close to real conditions within the HADES (High Activity Disposal Experimental Site) project.
Understanding the basic phenomena which control the retention and/or mobility of radionuclides in the clay is a key issue in this research program. According to the reference multi-barrier disposal concept, sufficient knowledge and reliable information are required on items as various as the definition of the source term, the performance of the near field, the geological environment, etc. This paper will review the main results of previous research, the priorities granted to current research activities and the evolution of present program issues. The same methodology is now considered for studies on the direct disposal of spent fuel.
BELGIAN FRAMEWORK FOR WASTE DISPOSAL
Safe disposal of radioactive waste is one of the key issues in the consolidation process of the nuclear industry. For long-lived waste, specifically spent fuel and vitrified high-level waste, disposal in deep geological formations is at present the most promising option. In our country, investigations were concentrated on argillaceous formations because of their abundance at various depths in large parts of Belgium, their expected favorable properties and the absence in the country of salt- and/or suitable hard rock formations.
In Belgium, reprocessing is now the reference policy. The volume of long-lived waste to be produced by a total installed nuclear power of 5.5 GWe during a 40 year period can be estimated to 4.000 m3 of vitrified heat emitting waste and about 35.000 m3 of alpha- contaminated waste, conditioned in concrete or bitumen matrix.
It is assumed in the present reference scenario that the operation of a repository for underground waste disposal could start in 2035 to last until 2070/2080 (closure phase). This requires, in case of vitrified waste, a temporary surface storage of about 50 years allowing the heat output to be significantly reduced. This limited heat output in the repository results in temperatures not exceeding 100°C in the host clay.
In order to assess the performance as well of individual components of the disposal system as of the integrated system, the current R&D program covers mainly research on basic phenomena, demonstration tests and safety studies. Scientific and/or financial support was provided with time by the European Commission (EC) and the Radioactive Waste Management Agency (ONDRAF/ NIRAS) which is in charge of the implementation of the disposal program. If most of the research performed at SCK*CEN has become contractual, both the extent of international partnerships and the broadening of collaborations with universities allow to complement the contractual work with exploratory and more basic studies initiated at the SCK•CEN.
With regard to the disposal of low-level and short-lived waste, Shallow Land Burial (SLB) is considered in Belgium as a possible option but will not be examined in the present paper.
METHODOLOGY AND CONCEPTS
The geological disposal of long-lived waste is based on a "multi-barrier" concept; several "engineered" barriers (overpack, backfill, gallery lining) are installed between the waste matrix itself (primary containment) and the host rock. Their purpose is to delay (e.g. by 500 to 1000 years) release of activity from the repository structure into the geological environment. The latter time period corresponds approximately to the thermal transient phase of the repository. Consequently, several coupling effects with temperature can be avoided in the host clay, provided the required performance of these engineered barriers can be guaranteed.
For argillaceous formations and particularly for the Tertiary "Boom clay" layer, tentatively selected for site characterization studies, the host "rock" is the primary natural barrier against radionuclide migration and provides good sorption capacities, very low permeability and favorable geochemical conditions. The backfill has to fulfill geomechanical and geochemical stability and compatibility. Phenomena relevant to repository performance are generally subdivided into those occurring in the near field and those occurring in the far field (geosphere). By near field, one generally considers the engineered repository structure (including the waste packages and engineered barriers) and those parts of the surrounding rock whose characteristics have been or could be altered by the repository or its content (excavation, temperature, radiation).
Since 1978 different repository concepts have been considered for high-level and long-lived radioactive waste in the Boom clay formation. In the reference concept now considered by NIRAS/ONDRAF, these waste would be emplaced in separate disposal galleries. Approximate dimensions (expressed in inner diameter) for the different components are 6 m for the shafts, 4 m for the primary galleries and 2 m for the HLW disposal galleries. The length of the disposal galleries would be 800 m (4 sections of 200 m). Concrete would be used for lining of shaft and galleries. The lining is only designed to ensure mechanical stability during the operational phase of the repository. For the non heat-producing long-lived MLW, the whole section of the galleries, 3.5 m in diameter, would be filled with MLW canisters and remaining voids possibly with concrete.
The design of the HLW disposal galleries is more advanced and will, during the next 10 years, be the main subject of the PRACLAY project, requiring a substantial extension of the present Underground Research Facility (URF). According to this concept, the HLW canisters, simulated for the purpose of the test by heating elements, are assumed to be placed in long metallic tubes in the centre of the disposal galleries. The thickness of this "overpack" will have to be sufficiently high to guarantee its role of reliable barrier during the thermal phase of the disposal and sufficiently low to limit the potential risk of gas generation by anaerobic corrosion. The void between the metallic tube and the concrete gallery lining would, according to present views, be backfilled with pre-compacted calcium bentonite blocks.
INTRODUCTION TO THE HADES PROJECT
In the first decade of the HADES project (1974-1984), a preliminary R&D program consisting of laboratory studies on sampled clay cores and observations on the geology and hydrology of the area was performed. The initial results were very promising with regard to sorption properties of the considered host rock, its chemical stability and permeability.
Early in the program (1980), an underground laboratory (HADES) was constructed in the clay at the Mol site; a tunnel, 35 m long and 3.5 m in diameter provides a large number of accesses to the clay formation, at a depth of 223 m. The first objectives were related to the technical feasibility of gallery digging in plastic clay under high lithostatic pressures and to the evaluation of the compatibility of vitrified HLW and candidate container materials with the clay, in close to real conditions. In 1987, an additional 60 m long drift was built for integrated tests, e.g. the impact of temperature and radiation.
In the meantime, much experience has been gained on sampling of fluids and solids, in characterization of clay and waste forms; an important scientific data base is now available at SCK•CEN, specifically for vitrified waste and Boom clay. Some observations could be made or confirmed in the 80's, e.g: the feasibility, using conventional civil engineering technology, to drive horizontal galleries in plastic clay, at a depth of more than 200 m, with an internal diameter of at least 3.5m, without special ground conditioning; the nature of stabilizing factors in the Boom clay, such as its cationic exchange capacity mainly due to the presence of smectite and smectite-illite interstratified clay, the strong reducing conditions which are due to finely dispersed pyrite and humic materials intimately associated with the clay minerals and a slightly alkaline environment caused by the presence of carbonates.
Preliminary safety assessments were carried out for many scenarios ranging from natural evolution to important disruption. For the reference mix of reprocessing waste (HLW, MLW, hulls) from a 40 year nuclear program in Belgium, the highest calculated dose to population, for a normal evolution scenario and water well pathway, are of the order of 3 10-7 Sv/a, which can be compared with a dose constraint 10-4 Sv/a, as suggested by international bodies; these doses would occur after a period of about 60.000 years. The methodology for risk assessment was established and tested in the framework of international exercise in cooperation with the Nuclear Energy Agency of the OECD and the E.C ("PAGIS", "PACOMA", more recently "EVEREST"). The two first exercises were updated in the early 90's and the results published in the "Updating 1990", Ref. 1. Due to the limitations of this paper and in spite of their essential support in defining the R&D program, the performance assessments realized in the framework of the Belgian program will not be further detailed hereafter.
RECENT AND FUTURE R&D PROGRAM
The formation- and site specific approach allowed SCK•CEN to collect valuable results and input data required for modeling work. Beside the continuity of the research program for more than 20 years now and the availability of both underground and surface laboratories, the structure of the research team ensures the integrated character of the HADES program, providing a good collaboration between experimentalists and modelers and the necessary relationship between the performance assessments and the definition of R&D priorities.
The understanding of the migration processes in the host rock, together with the hydrogeological context of the host and surrounding formations are of prime importance for the long term safety of a reference geological disposal concept; of course, all phenomena and processes involved in and around a deep geological repository (decompression, heat emission, radiolysis, gas production, etc) and their evolution with time are also dependent on the waste types considered.
Research on the basic phenomena expected to take place in the disposal system are a prerequisite to demonstration tests. We describe hereafter the main research topics required for studying the phenomena which control the release of radionuclides from the waste packages to the geosphere through the various successive barriers; modeling aspects and model validation are included.
-Characterization of Waste Packages and their Compatibility with Clay
For about 15 years, the performance of various candidate waste-package
components (overpack, canister, waste forms) has been studied under realistic
geological repository conditions.
A large part of the research program is related to the qualification and verification of waste forms produced by Cogema, to be returned to Belgium and waste forms resulting from the reprocessing of spent fuel by Eurochemic. The main objective is to determine the important characteristics of simulated or real waste packages and to compare these data with the specifications of the producer. The testing program includes the following items:
The evaluation of the compatibility of waste packages with the disposal environment is a larger task aimed at defining acceptance criteria for disposal in the selected formation.
The main waste forms under consideration in this program are produced by Cogema and Eurochemic. The research program on Cogema waste forms is focused on the glasses R7T7 and SON68 and STE bitumen. The research program on Eurochemic waste involves studies on the Pamela high-level glasses SM513 (low-enriched waste concentrates), SM527 and SM513 (high-enriched waste concentrates) and medium-level bituminized waste (Eurobitum). In particular, the corrosion mechanisms influencing the release of radionuclides from the conditioned waste product are studied. The results of in-situ experiments, Ref. 2, are in agreement with those of laboratory tests confirming that glass is an efficient first barrier for the radionuclides. Dissolution rates lower than 0.1 µm per year were measured at ambient clay temperature (16°C) in direct contact with Boom clay. However, the dissolution of glass is strongly enhanced by increasing temperature; clay is acting as a sink for many glass constituents such as Si, rare earths and actinides. The long-lived actinides (Pu, Am, Np) in general leach slower than the bulk glass, and the fraction leached is for more than 90% sorbed onto the clay.
The corrosion of vitrified waste is studied by geochemical and mathematical models. In particular, a geochemical code (PHREEQE) was applied to corrosion of glasses at 90°C in pure water or synthetic claywater. A good agreement with results from leach experiments was found for B, Na, Li, and Si. For the mathematical modeling of the glass corrosion, stochastic models are applied. The mathematical model takes account of the receding of the glass surface with the progress of corrosion. Flow-through migration tests were carried out to determine the diffusion coefficient of silica in Boom clay. Although not expected, a slight silica sorption was observed in the Boom clay; this can be explained by the formation of surface complexes between the silicate anion and the aluminium hydroxide groups.
So far, in-situ tests on waste forms were restricted to small, inactive or doped samples. A new experiment including glass samples doped with large amounts of alpha emitters has been recently designed (CORALUS) to determine the dissolution of the glass in simulated disposal conditions (presence of container and backfill materials) and to assess the migration of the radionuclides through the interacting media in a radiation field.
Leach experiments on medium-level bituminized waste (Eurobitum) have confirmed the rather high leach rate of NaNO3 . Eurobitum is a reference Belgian bitumen containing on the average 35 wt% of reprocessed sludge. This would imply that a full size drum would be depleted in NaNO3 (25%) within a few thousand years leading to the release of important amounts of nitrate into the near field, Ref. 3. Lower leaching rates were measured for the actinides; their leaching appears to be controlled by solubility and is not enhanced by the lithostatic pressure. Preliminary studies on the microbial degradation of bitumen in repository conditions stressed the importance of long term effects,Ref. 4.
In addition to the studies on high-level vitrified waste forms, the compatibility of other waste forms of interest such as SYNROC, bituminized and cemented waste with the Boom clay is investigated. In particular, the influence of organic degradation products from cellulose containing cemented waste on the mobility of Pu and Am in Boom clay is under study; these degradation products may form at high pH in a cementitious environment. Moreover, a similar study investigating the influence of organic degradation products formed by the radiolytic degradation of Eurobitum will be performed.
New programs on the performance of spent fuel in case of direct disposal in the Boom clay were started, Ref. 5. The considered spent fuel types are uranium oxide fuels with burn-ups of 33 and 45 GWd/tHM and MOX fuel with a burn-up of 45 GWd/tHM. The lack of experimental data describing the behaviour of spent fuel in realistic disposal conditions remains however a serious problem. The present program therefore includes an experimental study aiming at quantifying the processes controlling the dissolution/alteration of the spent fuel under realistic disposal conditions.
The performance of various candidate waste-package components (overpack, canister, waste forms) has been studied under realistic geological repository conditions by in-situ corrosion experiments. For carbon steel, it was found that the corrosion process during the aerobic and anaerobic phase is dominated by pitting corrosion. However, the surface morphology is quite different: during the aerobic phase, small and deep pits are formed while during the subsequent anaerobic phase, the pit edges are abraded, resulting in pit coalescence. Pits with a depth up to 240µm have been detected (at 90°C after 2 years). The main shortcoming of such tests is that the different phenomena in aerobic and anaerobic conditions cannot be studied or quantified separately. Therefore, a new experimental laboratory program is started which includes immersion tests on the stainless steel AISI316L hMo and electrochemical tests on C-steel, stainless steel (AISI316L, AISI316Ti, UHB904L, Cronifer1925hMo), Ni-alloy (Hastelloy C4) and Ti-alloy (Ti/0.2Pd). The corrosion rate and mechanisms will be studied as a function of temperature (16°C<temp<140°C), oxygen concentration (aerobic, anaerobic), and the content of chlorine and thiosulphate.
-Gas Generation and Transport
For the option of a deep repository for HLW, several potential sources of
gases were identified: i.e. the anaerobic corrosion of iron, the degradation of
organic materials, the generation by gamma radiolysis. Of those gases, hydrogen
is certainly the gas which can be released in the potentially largest amounts.
The gas release may increase the rates of flow of potentially contaminated
groundwaters to the surface and two-phase flow may entrain radionuclides in
solution at the gas-water interface, affecting the long term safety of the whole
disposal system.
In the surface laboratory special oedometers were built to saturate clay cores, to measure their hydraulic conductivity and to perform gas injection; at the end of the gas flow experiment, the cores were still more than 95% saturated, as confirmed by X-ray tomographic analysis.
The formation of preferential pathways and a low breakthrough pressure were observed during the field tests. These paths are unstable and undergo successive opening and closure cycles depending on local stress and gas pressure evolution. The self-healing capacity of the plastic clay probably reduce any long-term gas flow effect. Modeling work carried out in the framework of the MEGAS project, Ref. 6, indicates that the capillary bundle approach is promising. Further investigations are required to study in depth the phenomenology involved. These research topics are included in two new EC projects (PROGRESS, RESEAL). They will focus on the following items:
More specifically within the RESEAL project aimed at demonstrating on a representative scale the performance of a low permeability seal, to be installed in an experimental shaft in the URF, the gas tightness of this seal plug will be tested in close to real conditions.
-Radionuclide Migration
The research on radionuclide migration through the Boom clay has provided
data needed by the performance assessors. Work was concentrated on
methodological aspects, codes for flow and transport, theoretical studies,
laboratory and field experiments on geochemical transport phenomena in order to
develop and to test conceptual and numerical models.
The transport of radionuclides in porous media is controlled by diffusion, advection, retardation processes and radioactive decay. In the Boom clay, the radionuclides migration is dominated by diffusion; the advection plays a secondary role due to the very low hydraulic conductivity and the absence of preferential paths for the water in this clay formation.
This research item will be more detailed in another paper to be presented at this conference where the R&D program on migration is illustrating the interaction between host rock characterization and performance assessment, Ref. 7. The latter precisely led to the selection of groups of critical radionuclides, of interest in the Belgian program; the most critical ones are 129I, 14C and some other fission products. We will restrict ourselves within this paper to some important results:
-Thermal, Radiation and Chemical Effects
Radiolysis of water by gamma rays can lead to the production of oxidants
which might seriously alter the near field chemistry and retention properties of
the clay. The radiation effects need to be combined with heat effects as they
occur during the same period.
Chemical effects have not yet been studied in detail; however some potential problems have already been identified, as mentioned above. The presence of large quantities of degradable organic material and soluble salts such as nitrates/sulphates can have an important effect on the global near-field chemistry, the solubility of the radionuclides and the sorption properties of clay.
The study of the combined effect of heat and radiation is one of the main objectives of the CERBERUS in situ experiment, ref. 8. The CERBERUS experiment simulates the in-floor configuration for the disposal of a Cogéma HLW canister after 50 years cooling time. In its periphery, instruments or test specimens have been installed to obtain data for several years on water chemistry, radionuclide migration, corrosion, radiolysis, water pressure and temperature distribution.
Seven years of follow-up and measurements in the clay around this experiment have shown a decrease of the pH with about one pH unit and a very slight increase in redox potential so that the clay remained reducing. The presence of thiosulphates, which can seriously increase metal corrosion, and oxalate, which can act as a complexant, was observed. In-house codes for the calculation of the thermal, hydraulical and radiation field were validated. Long term survey is continued. Other aspects need further development and/or evaluation after exposition for more than 5 years to realistic temperature and radiation fields: the performance of a potential backfill material, the near-field effects (mineralogy, geochemistry, migration) on the surrounding clay and on selected waste forms (test corrosion tubes).
-Geomechanical/geochemical Behavior of Clay-based Materials
The mechanical and chemical stability of the near field provide a set of
controlling factors for groundwater movement, radionuclide transport and heat
dissipation through the interface between near and far field. Therefore, the
excavation disturbed zone in clay and the backfill/sealing material might
substantially affect the subsequent migration behaviour of radionuclides. These
aspects are quantified by modeling coupled processes of the
thermo-hydro-mechanical (THM) and geochemical behaviour of these materials.
The measurements of pore water pressure and stress distribution around boreholes and galleries during and after construction, Ref. 9, have shown that the extent of the disturbance depends on the excavation diameter, the excavation rate and the oversizing with respect to the lining diameter; the magnitude of the disturbance drops shortly with the distance and decreases slowly with time thanks to the plasticity of the clay. The quantification of the disturbed zone around a disposal gallery successively submitted to decompression and thermal loading is one of the main objectives of the PRACLAY demonstration test, simulating the thermal output of a 30 m long HLW disposal gallery, 2 m in diameter.
A clay-based material is generally chosen as backfill material for the same reasons as put forward when dealing with the geological barrier. Boom clay has been characterized with a view to re-use the excavated clay spoils for backfill, since oxidation (pyrite) can be prevented easily by a fast drying. The following indications have been found for highly compacted Boom clay with regard to compacted bentonite: slightly higher but sufficiently low hydraulic conductivity, lower swelling but still convenient regarding the lithostatic pressures at the Mol site, better retention properties for some radionuclides and steam effect less dramatic than for some bentonite.
When dry compacted clay is used as a backfilling around HLW packages, its behaviour during saturation is complicated by the thermal transient caused by the heat emission of the waste. It must be shown that during this transient period, no (negative) influence on the long term behaviour of the repository can occur due to e.g. shrinkage or collapse of the clay backfill. Typically hydro-mechanical and thermo-hydraulical models have to be coupled to describe stress/strain behaviour, moisture migration and heat transfer. The thermo-hydraulic model has also to be coupled to a geochemical code to describe the migration and formation of chemical species. A fairly extended database concerning THM properties of unsaturated clays is a prerequisite to allow for physical model testing, model calibration and validation. Impressive work has been done on this subject during the last 4 years in the framework of three E.C research contracts, Ref. 10. The next step is the development for validation purposes of the large in-situ demonstration RESEAL test.
Beside the hydraulical aspects here involved, the regional hydrogeology of the aquifers surrounding the host clay layer is one of the key issues in performance assessment studies. After updating of the regional aquifer model and data, additional hydrogeological investigations focusing on the spatial variability of the parameters and a better characterization of the aquifers are launched, including a hydrochemical survey.
CONCLUSIONS
The successive research activities, hereabove shortly described, confirmed the selection of the clay option in Belgium and the continuation of the current characterization program directed to the Boom clay layer at the Mol-Dessel site. The HADES underground research facility is presently the only one installed in a deep clay layer. This unique infrastructure offers important opportunities for the participation of foreign organizations willing to demonstrate disposal concepts, to provide reliable data on the long term performance of repository barrier components and to allow model testing and validation for representative conditions. With a view to the installation of the PRACLAY project, the capacity of this facility will be extended; this extension will be managed by an Economic Interest Grouping (E.I.G. PRACLAY).
In general, selecting appropriate concept, cooling time and overpack design allows to avoid the study of high temperature and steam effects. However, the impact of radiation and heat on retention/mobility of radionuclides in clay and on conditioned waste will have to be further investigated. Also the influence or significance of non-radioactive components in the anaerobic environment of the repository, affecting acceptance criteria for packaged waste will be considered and could even lead to the consideration of alternative conditioning processes.
With respect to waste packages, the discussion on both characterization and compatibility of waste forms clearly identifies further shortcomings in running actions; on one hand, the development, application and standardization of quality checking for waste packages must allow to identify R&D requirements and to coordinate the development of new test methods. On the other hand, more investigations is e.g. required for HLW glass on the secondary phase formation affecting the dissolution mechanisms for the long term.
A recent R&D project started on the study of neptunium complexes formed upon interaction of HLW glass and Boom clay where the characterization of the complexes using rather sophisticated techniques (laser spectroscopy) would provide information on the solubility of the neptunium to be compared with the predictions of chemical codes based on experimentally determined complexation constants. Similar research work is needed on the soluble concentrations of the actinides in clay media in presence of bitumen degradation products, on the ageing behaviour of bituminized waste and on the microbiological degradation of bitumen.
Finally, more practical studies with regard to natural analogues, to be really used in performance assessments to increase confidence building, are required. In this context, the geochemistry of the Boom clay formation will be studied in detail in view of obtaining more information on the long term behaviour of some key-elements and radionuclides in the natural environment. By this approach, we consider the Boom clay formation itself as a natural analogue, allowing to validate geomechanical model and to better assess the safety over geologic time scales.
REFERENCES