Elsevier

Applied Geochemistry

Volume 19, Issue 9, September 2004, Pages 1431-1451
Applied Geochemistry

Modelling the interaction of hyperalkaline fluids with simplified rock mineral assemblages

https://doi.org/10.1016/j.apgeochem.2004.01.018Get rights and content

Abstract

Designs for geological disposal facilities for radioactive wastes often envisage the extensive use of cementitious materials. After closure, the repository will saturate with groundwater and cement porewater will migrate into the geosphere to form an alkaline (pH 12.5–13) plume. This plume will react with the components of the surrounding host rock leading to mineralogical, chemical and physical changes which will be complex in nature. Coupled computer models of geochemistry and hydrogeological transport will be needed to scope these changes. In the recent past, a series of laboratory column experiments were carried out by the British Geological Survey in order to test the capabilities of coupled models to predict the evolution of outflow fluid compositions and product solids. These experiments reacted single minerals (i.e., quartz, albite, calcite and muscovite/quartz) and potential host rock lithologies (i.e., Borrowdale Volcanic Group fault rock, Äspö granite and Wellenberg marl) with cement pore fluids. The objectives of the present work were to develop the understanding of these experiments and to improve simulations of the data by computer models. To this end, a systematic approach was adopted in which dissolution rates for individual minerals were calibrated to the results of the single mineral columns and the resulting parameter values applied to the more complex columns. It was found that for the systems studied, reasonable agreement between the experimental data and the calculated results could be obtained. However, the calibrated dissolution rate constants used in the models first vary between experiments by up to half an order of magnitude and secondly differ from literature values by up to one order of magnitude, with the calibrated values being generally slower than those found in the literature.

Introduction

There are a range of options for the long-term management of radioactive waste. One of these is to place the wastes in a repository excavated in stable rock formations, deep underground (deep geological disposal). Current concepts for the geological disposal of radioactive wastes envisage the use of multiple barriers (i.e., both engineered and geological barriers) to contain the radioactive materials in the waste. The barriers include the waste form itself, a container (or canister) surrounded by a backfill and finally the geosphere. Cementitious materials, for example concretes and grouts, may be used extensively in the construction of the disposal facility and as a backfill. After closure, the repository will saturate with groundwater and become part of a modified regional groundwater flow system. One possible scenario then is that groundwater equilibrated with the cement porewater will migrate into the geosphere and as a result form an alkaline plume (initially with a pH of 12.5–13). The alkaline plume will react with the components of the rock and that may lead to changes in the mineralogical, chemical and physical characteristics of the surrounding rock. The interactions of the alkaline plume with the surrounding rock will be complex, and therefore the use of coupled geochemistry and hydrogeological transport computer models to scope the changes is indicated. The extent and timescale of leaching of alkali from the cementitious materials will depend on the relative importance of different groundwater transport processes that could operate in the repository and on the groundwater chemistry. For example, self-sealing of a repository could occur as a result of calcite precipitation under some conditions [e.g., Pfingsten (2002)], reducing the formation of an alkaline plume. Nevertheless, it is important to understand the consequences that could result for scenarios in which a significant alkaline plume is formed downstream from a cementitious repository.

In the recent past, a series of laboratory column experiments were carried out by the British Geological Survey in order to test the capabilities of coupled models to predict the evolution of outflow fluid compositions and product solids. These experiments reacted single minerals (i.e., quartz, albite, calcite and muscovite/quartz) (Bateman et al., 2001a) and potential host rock lithologies (i.e., Borrowdale Volcanic Group fault rock, Äspö granite and Wellenberg marl) (Bateman et al., 2001b) with simplified cement pore fluids.

An overview of the experimental procedures and preliminary predictive modelling have been given in Bateman et al. (1999). The objectives of the present study were to develop a better understanding of these experiments and to improve the correspondence between the data and the predictions of the computer models.

Section snippets

Description of experiments

A schematic diagram of the experimental set-up is shown in Fig. 1. The equipment consisted of two reservoirs, which were kept at room temperature and which contained the reactant fluids under a protective N2 atmosphere (1 bar) to prevent ingress of CO2. The reservoirs were connected to a series of heated conditioning vessels, held at the experimental temperature (70 °C), which in turn were connected via peristaltic pumps to the columns. Each column (internal diameter 7.5 mm and length 0.3 m)

Modelling codes

A number of reactive transport codes have been developed over recent years by various authors for a variety of purposes. The rigour with which these computer programs address each of the coupled processes relevant to the near-field environment of an underground repository varies depending on the application envisioned by the developer. That is, there are significant differences in the conceptual and mathematical models, the numerical implementations and the solution methods. The simulations

Models of single minerals

Coupled geochemistry and hydrological transport models of the experiments on columns packed with a single mineral (i.e., quartz, albite, calcite or a mixture of muscovite and quartz) are developed systematically in this section.

The features of the experimental results that can be used to assess the quality of a model include:

  • (a)

    The chemistry of the outflow fluid.

  • (b)

    The extent of primary mineral dissolution.

  • (c)

    The type and location of the secondary minerals formed.


The first of these is the most

Models of synthetic host rock lithologies

Coupled geochemistry and hydrological transport models of the experiments on columns packed with a synthetic representation of a potential host rock (i.e., Borrowdale Volcanic Group fault rock, Äspö granite or Wellenberg marl) are developed systematically in this section. Synthetic representations of the potential host rocks are considered (rather than the actual rocks) because it should be simpler to relate these to the experiments on columns packed with a single mineral (Section 4). In

Discussion

In this work, an attempt has been made to develop systematically reactive transport models of the column experiments carried out by the British Geological Survey. The results of this study are summarised below.

Conclusions

The following general conclusions derive from this study:

  • (a)

    Some care is required in the employment of “standard” thermodynamic databases (e.g., see Section 3.2.1) to ensure that the data they contain is appropriate for the intended application (e.g., the data is satisfactorily complete).

  • (b)

    In simulating the percolation, at rather high flow rates, of highly alkaline fluids through columns packed with either single minerals or simple mixtures, the model parameters that have the most significant effect

Acknowledgements

The work described here was carried out as part of the ECOCLAY II project, which was undertaken within the framework of the European Commission’s R&D programme on Nuclear Fission Safety (1999–2002), with funding by the European Commission and United Kingdom Nirex Limited.

References (25)

  • Bateman, K., Coombs, P., Pearce, J.M., Noy, D.J., Wetton, P.D., 2001b. Fluid rock interactions in the disturbed zone:...
  • Bond, K.A., Heath T.G., Tweed, C.J., 1997. HATCHES: a referenced thermodynamic database for chemical equilibrium...
  • Cited by (11)

    • Two-dimensional reactive transport modeling of the alteration of a fractured limestone by hyperalkaline solutions at Maqarin (Jordan)

      2016, Applied Geochemistry
      Citation Excerpt :

      The interaction between groundwater and cement causes the generation of hyperalkaline solutions (pH 12.5–13.5), which may react with the rocks hosting the repositories and change their physical and chemical properties. Experimental and modeling studies of such interactions have been common in the last years (e.g. Adler, 2001; Read et al., 2001; Savage et al., 2002, 2011; Soler, 2003, 2013; Gaucher et al., 2004; Hoch et al., 2004; Mäder et al., 2005; Soler and Mäder, 2005, 2007, 2010; Sánchez et al., 2006; Marty et al., 2009, 2014; Honty et al., 2010; Soler et al., 2011; Kosakowski and Berner, 2013; Moyce et al., 2014). A common finding of these studies has been a reduction of porosity near the cement–rock interface due to the precipitation of secondary phases.

    • High-pH plume from low-alkali-cement fracture grouting: Reactive transport modeling and comparison with pH monitoring at ONKALO (Finland)

      2012, Applied Geochemistry
      Citation Excerpt :

      In the context of the safety assessment of deep geological repositories for radioactive waste, it has been common practice to assume long-term interaction between host-rocks or engineered-barrier materials with hyperalkaline solutions (pH > 12) originating from the degradation of hydrated cement and concrete. Experimental and modeling studies of such interactions have been common in recent years (e.g. Adler, 2001; Savage et al., 2002; Soler, 2003; Gaucher et al., 2004; Hoch et al., 2004; Mäder et al., 2005; Sánchez et al., 2006; Soler and Mäder, 2005, 2007, 2010; Honty et al., 2010). Additionally, low-alkali cements have been recently developed to minimize the potential impact of hyperalkaline solutions (e.g. Arenius et al., 2008; Lothenbach and Wieland, 2010).

    • A comparative study of the modelling of cement hydration and cement-rock laboratory experiments

      2011, Applied Geochemistry
      Citation Excerpt :

      Many of these studies have highlighted difficulties in scaling reactive surface areas, not only as applied to mineral dissolution (Hoch et al., 2004; Soler et al., 2006; Soler and Mäder, 2007), but also to mineral growth (Watson et al., 2009). The incorporation of geometric or BET surface areas in simulations to model mineral dissolution has led to the overestimation of reaction rates by up to two orders of magnitude (Hoch et al., 2004; Soler et al., 2006; Soler and Mäder, 2007). The issues associated with the up-scaling of laboratory-derived kinetic data to model at the natural system scale has previously been encountered during the evaluation of weathering in groundwater catchments in granitic rocks (e.g. White and Brantley, 1995, 2003).

    View all citing articles on Scopus
    1

    Present address: ANSTO Environment, PMB 1 Menai, NSW 2234, Australia.

    View full text