The evolution of clay rock/cement interfaces in a cementitious repository for low- and intermediate level radioactive waste

Abstract

In Switzerland, deep geological storage in clay rich host rocks is the preferred option for low- and intermediate-level radioactive waste. For these waste types cementitious materials are used for tunnel support and backfill, waste containers and waste matrixes. The different geochemical characteristics of clay and cementitious materials may induce mineralogical and pore water changes which might affect the barrier functionality of host rocks and concretes.

We present numerical reactive transport calculations that systematically compare the geochemical evolution at cement/clay interfaces for the proposed host rocks in Switzerland for different transport scenarios. We developed a consistent set of thermodynamic data, simultaneously valid for cementitious (concrete) and clay materials. With our setup we successfully reproduced mineralogies, water contents and pore water compositions of the proposed host rocks and of a reference concrete.

Our calculations show that the effects of geochemical gradients between concrete and clay materials are very similar for all investigated host rocks. The mineralogical changes at material interfaces are restricted to narrow zones for all host rocks. The extent of strong pH increase in the host rocks is limited, although a slight increase of pH over greater distances seems possible in advective transport scenarios. Our diffusive and partially also the advective calculations show massive porosity changes due to precipitation/dissolution of mineral phases near the interface, in line with many other reported transport calculations on cement/clay interactions. For all investigated transport scenarios the degradation of concrete materials in emplacement caverns due to diffusive and advective transport of clay pore water into the caverns is limited to narrow zones.

A specific effort has been made to improve the geochemical setup and the extensive use of solid solution phases demonstrated the successful application of a thermodynamically consistent union of very different materials like hydrated cement and clay phases. A reactive system utilizing a novel solid-solution approach is used, where cation exchange is an intrinsic property of the mineral phase definition. Although such features were not the primary aim of the study, they offer a large potential for studies where ion exchange and changing sorption properties are of interest.

Introduction

In Switzerland a site selection process for geological repositories for spent fuel (SF), vitrified high-level waste (HLW), long-lived intermediate-level waste (ILW) and for low- and intermediate-level radioactive waste (L/ILW) (“Sectoral Plan for Geological Repositories”) is under way. The safety of the SF/HLW and the L/ILW repositories shall be ensured by a multi-barrier concept. Barriers include waste matrices, steel containments (canisters, drums), emplacement containers, cavern backfills and the host rock (Nagra, 2002a).

A L/ILW repository is characterized by rather high waste volumes on one hand and by low radiotoxicities on the other hand. The stored wastes are characterized by a large variability and heterogeneity. Apart from the embedded radioactive wastes, the principal materials in the L/ILW mainly are hydrated cements, aggregate materials and steel. Usually, the wastes are solidified/embedded in a cement/concrete matrix, which itself is included in steel drums. These drums are mainly needed for interim storage, handling and transport prior to emplacement. Parts of the wastes might also be solidified in bituminized matrices. For the construction of the caverns a reinforced concrete or shotcrete wall support is used. Specifically designed concretes (e.g. “monocorn” concrete) may be used for backfilling the remaining cavities. The concrete and mortar aggregates are considered to consist of siliceous materials (primarily quartz), but calcareous materials (lime) are an option. In summary, the backfilled caverns contain, in addition to the wastes, chemically reactive hydrated cements, significant amounts of steel (drums, tunnel support, reinforcement bars and construction materials) and concrete aggregates (Nagra, 2008a).

In Switzerland, the proposed surrounding host rocks are clay-rich sediments (Opalinus Clay, “Brown Dogger”, Effingen Member, “Helvetic Marls”) (Nagra, 2008b). At first glance, all these rocks are characterized by similar solid phase assemblages, by similar pore solution compositions and by the presence of significant fractions of strongly sorbing clay materials.

The knowledge on the temporal and spatial evolution of alterations near interfaces between different materials is important for the performance assessment of deep geological repositories for radioactive waste. The chemical status of the repository, as well as its temporal changes, is driven by the interactions between hydrated cement, steel, aggregates and clays. Considering the chemical nature of the involved major materials, two basic processes take place: an acid–base reaction (aggregate/clays–alkaline cement) on one hand, and the corrosion of iron on the other hand. All chemical reactions solely proceed if a stagnant or mobile aqueous phase is present in sufficient quantity. Chemical gradients, here in particular emanating from the strongly alkaline cementitious materials, dictate the reaction courses, and the principles of chemical thermodynamics determine the fundamental reaction products of the interactions. In terms of transport a stagnant aqueous phase is dominated by diffusion of solutes, whereas for a mobile aqueous phase advective transport may dominate.

Information from literature reviews and mass balance calculations give indications on maximum possible changes of barrier materials. To investigate the spatial and temporal evolution of the overall system, the specific influence of single processes (e.g. pore space changes) and the potential feedback of different processes, numerical models are established research tools.

Research related to cement/clay interactions is, for more than 20 years, a focal point in the field of radioactive waste management.

Gaucher and Blanc (2006) give an overview on publications on experiments, natural analogs and modeling related to cement/clay interactions. Based on their literature review they conclude that it appears premature to conclude definitely on the consequences of an alkaline disturbance in a clay medium. They state that the mineralogical consequences of an alkaline disturbance are fairly well-known, although the knowledge of the thermodynamics of clays, zeolites and cement phases needs to be improved further (see also Savage et al. (2007) on this topic). Large uncertainties remain in the field of dissolution kinetics and mineral precipitation, and especially in understanding the dissolution kinetics of montmorillonites.

Very recently Wang et al. (2010) presented the results of a feasibility study for disposing radioactive waste in a cementitious repository in Boom Clay. They give an extensive overview on work done in Belgium, Switzerland and France on cement–clay host rock interactions, a summary of studies on bentonite–cement interactions and further present results of the ECOCLAY II project. They conclude that the nature of the alkaline plume disturbance in clay materials is relatively well understood.

1D reactive transport simulations typically show that the cement–clay interface most likely clogs and that mass transport across the interface is significantly slowed down. Reactive transport calculations for various clay–cement interfaces show porosity clogging at the interface (Fernández et al., 2009, Marty et al., 2009, Smellie, 1998, Trotignon et al., 2006, Trotignon et al., 2007, De Windt et al., 2004). Trotignon et al., 2006, Trotignon et al., 2007 did numerical modeling on the durability of concrete engineered barriers in contact with mudrock (clay host rock). They think that one key process is the progressive localized cementation of the altered mudrock. Their results suggest that a sharpening of the cementation front will occur and lead to low cementation lengths in the mudrock (∼0.2m). They showed that adding kinetic constraints to mineral dissolution/precipitation does not change the alteration distances significantly, but may influence the mineral assembly to a certain degree. Traber and Mäder (2012) investigated the evolution of an Opalinus Clay–concrete interface with different mineralogical and transport scenarios. They found a porosity reduction at the interface and a limited extent of mineralogical changes (less than 0.2 m).

However, several authors recognize a strong sensitivity of the time behavior of clogging processes on the numerical mesh size and kinetic parameters (Kosakowski et al., 2009, Marty et al., 2009, Traber and Mäder, 2012, Trotignon et al., 2006).

A natural analog for the alteration of clays by high-pH solutions was found in Maqarin (northern Jordan). A very detailed description of the site can be found in Smellie (1998). At the site, hyperalkaline waters, similar to cement waters with a high pH, circulated through fractures in a clayey biomicrite. The fracture edges show dissolution of different minerals, whereas within the fractures different opening/closing episodes may have taken place with complex mineralogical sequences. The extended summary of Smellie (1998) states, that evidence from Maqarin shows

• that sequences of minerals predicted by thermodynamic and coupled modeling are similar to those observed in hyperalkaline alteration zones,

• that the rock matrix may be accessible to diffusion of aqueous species even during the phase of on-going wallrock alteration,

• and that narrow aperture fractures will probably be self-healing.

• Although the investigated system is not purely diffusion controlled, the above findings are indications that clogging at clay–cement interfaces is possible and likely over long times.

Gaboreau et al. (2011) characterized the porosity at a concrete–clay–rock interface at the Underground Research Laboratory at Tournemire after 15 years of interaction and could for the first time quantitatively verify porosity clogging. The authors could show that the perturbations (in terms of mineralogical and porosity changes) are limited to 3.5 cm and 1.5–2 cm from the interface in cement and clay rock, respectively. Both, concrete and clay rock are not homogeneous and contain a fissure network. Alterations are observable near the material interface and near the fissures. The porosity in the concrete is increased due to portlandite dissolution, whereas the porosity in the clay rock is strongly reduced, which the authors primarily attribute to the precipitation of C–(A)–S–H phases. The corresponding mineralogical data are published in Techer et al. (2012).

The above mentioned studies show, that the evolution of cement/clay interfaces seem to follow a general pattern. Nevertheless, differences in host rock and concrete transport properties, mineralogy and pore water composition might significantly change the evolution of the interfaces or the extension of mineralogical and pH changes.

We present reactive transport calculations which are specifically tailored to the host rocks and concrete materials foreseen in the Swiss disposal design. Such reactive transport calculations do not only allow refined mass balance calculations, but also introduce a qualitative view onto the system evolution as several processes and their mutual impacts can be considered simultaneously. Some of the presented models include a full coupling between porosity changes due to precipitation/dissolution of minerals and the change of porosity dependent transport parameters (effective diffusion coefficients).

Setting up these coupled models in turn required the refinement of geochemical models for cementitious (concrete) and clay materials and, particularly, the development of a consistent set of thermodynamic data valid for both types of materials. This improved geochemical setup is also used to assess the solubility of radionuclides in the cementitious near field.