Technical Report NTB 14-11

Geochemical Evolution of the L/ILW Near-field

The deep geological repository for low- and intermediate-level radioactive waste (L/ILW) and the emplacement rooms for long-lived intermediate-level radioactive waste (ILW) contain large amounts of cement based materials (concrete, mortar) used for waste conditioning, tunnel support (shotcrete) and the backfill of cavities. The waste inventory is composed of a wide range of organic and inorganic materials. This study describes the spatial and temporal geochemical evolution of the cementitious near-field, and the interactions with the technical barriers and the surrounding host rock.

The spatial and temporal evolution of the cementitious near-field is governed by several coupled externally and internally induced processes. The development of saturation by groundwater ingress from the host rock is of vital importance for many processes. Full or partial desaturation of the emplacement caverns, of the adjacent host rock or of the technical barriers in the access tunnels will reduce or even prevent the transport of dissolved species and therefore significantly impede chemical reactions. Saturation of the near-field is controlled by the inflow of water from the host rock, by the transport of partly dissolved gases from the near-field into the host rock and in the engineered gas transport system, and by the transport of humidity in the gas phase. The production of gas by anoxic corrosion of metals and by microbial degradation of organic wastes also consumes water. In addition, the mineral reactions which give rise to concrete degradation, such as carbonation or alkali-silica-aggregate reactions may also consume or produce water, respectively. However, these reactions (as well as the microbially driven gas production) require only a minimum humidity in the gas phase in the near-field so that a liquid film forms along mineral surfaces in which dissolved substances may react with each other.

In this study the description of cement degradation is consistent with that in previous studies. It is expected that the degradation of the cementitious near-field will occur in several phases. The first phase of cement degradation is related to the hydration of cement minerals.

In this phase the pore water has a pH of 13 or even higher as a consequence of the high content of dissolved alkali hydroxides. This phase will persist only for a short period of time, especially in situations where advective transport dominates.

A constant pH of 12.5, buffered by the equilibrium with portlandite, determines the second phase of the cement degradation. The alkali concentration is reduced by mineral reactions and/or solute transport. This phase will persist for a long time, especially when clogging of the cement-clay interface occurs and/or the transport is governed by diffusion.

In the third phase of the cement degradation the portlandite is completely dissolved due to the reaction with silicates/aluminates present in the near-field and carbonate in the groundwater of the host rock or associated with reactive waste materials. The pore water is in equilibrium with calcium-silicate-hydrates (C-S-H) which gives rise to a pH value near 11 or even lower. The Ca/Si ratio of C-S-H will change towards lower values (Ca/Si < 0.84).

In a very late phase of the cement degradation the formation of carbonates, clays or zeolites will cause the pH to drop to near neutral values. The geochemical evolution of the cementitious near-field is influenced by various processes. The most important ones are considered to be the interactions with the host rock, the interactions with waste, the degradation of cement minerals by alkali-silica-aggregate reactions and by cement carbonation.

The diffusive and advective exchange of pore water between the cementitious near-field and the host rocks gives rise to mineral reactions and changes of the pore water pH. Mineral reactions at the interface were investigated with the help of numerical models. The clay minerals of the host rock are dissolved and transformed into secondary minerals (e.g. zeolites) up to a distance of a few dm in 100'000 years (period considered in safety analysis for the L/ILW repository). For the case of higher water fluxes in the host rock this zone extends possibly up to 1 m over the same time scale. The sorption capacity of host rocks with low water fluxes and where transport is diffusion-dominated is not affected by these mineral changes. For host rocks where transport occurs mainly through fissures (Marl), the effect of the mineral formation on the sorption capacity of the host rock is taken into account in the provisional safety analyses for Stage 2 of the Sectoral Plan.

Beyond this transition zone and further into the host rock, a zone will develop with an elevated pH of 8 – 9, but without significant mineralogical changes. This zone extends a few meters into the host rock in the case of a diffusive transport regime, whereas in the downstream direction it may reach more than thousand meters in the case of higher water fluxes and very low host rock porosities. For a diffusion dominated regime the portlandite in the cementitious near-field will be dissolved up to a distance of 2 m from the near-field – host rock interface. In the region where portlandite dissolution occurred, the concrete pore water pH will drop to values corresponding to the third phase of the cement degradation. In the case of higher water fluxes from single facial units of the host rock into the caverns the portlandite dissolution front may extend a further few meters downstream in 100'000 years, while in the portlandite-depleted region all the cement minerals will completely degrade with time.

If the concrete aggregate contains SiO2, the cement minerals may degrade due to an alkali-silica-aggregate reaction. Silicon dioxide from concrete aggregates will react with portlandite and form C-S-H phases. This causes a decrease of the pore water pH and results in a complete dissolution of the cement minerals in the long term. The temporal progression of this reaction is poorly known. From available kinetic constants of mineral dissolution reactions it is inferred that the alkali-silica-aggregate reaction could be relatively fast, i.e. concrete degradation might occur within some hundreds to a thousand years.

The degradation of organic waste in a cementitious repository will presumably happen predominantly by methanogenesis which will produce CH4 and CO2. Due to the unfavorable conditions (high pH), the microbial activity driven degradation process will be low. The released CO2 will degrade the concrete surrounding the organic waste by carbonation. A prediction of the degradation rates of the organic waste is difficult since the rates depend not only on pH but also on the availability of water and the presence of certain nutrients such as phosphor and nitrogen. All of these factors limit the microbially mediated degradation of organic compounds. Based on our current knowledge it can be assumed that low-molecular weight organic materials will degrade within about 1'500 years. The degradation rates of the high-molecular-weight organic substances (polymeric materials) are highly uncertain. It is expected that the degradation may take at least some thousands of years, but probably up to several tens or hundreds of thousands of years, if they degrade at all. The effects of radionuclide complexation in a cementitious environment are taken into account in preparing the corresponding sorption database used in the provisional safety analyses for Stage 2 of the Sectoral Plan.

Some organic waste materials will react with the surrounding materials, assuming that water availability is not a limiting factor. Particularly important is the anoxic corrosion of metals (incl. steel) which can produce large amounts of H2. Corrosion products (e.g. rust) could contribute to the degradation of the surrounding cement minerals. Note that this again requires a connected water phase in the pores which allows dissolved substances to be transported. The general description of concrete degradation has not changed in this work compared to previous studies. A consequence is that the current concepts for radionuclide transport (retention) and radionuclide solubilities are not changed. It is therefore recommended to retain previously used descriptions for radionuclide solubility and radionuclide sorption (Kd-concept). In particular, the occurrence of colloid facilitated radionuclide transport is considered to be very unlikely since the host rock and near-field pore water chemistries give rise to very low colloid stabilities which will significantly reduce colloid concentrations.