David Savage

Modelling the interaction of bentonite with hyperalkaline fluids

Many designs for geological disposal facilities for radioactive and toxic wastes envisage the use of cement together with bentonite clay as engineered barriers. However, there are concerns that the mineralogical composition of the bentonite will not be stable under the hyperalkaline pore fluid conditions (pH > 12) typical of cement and its properties will degrade over long time periods. The possible extent of reaction between bentonite and cement pore fluids was simulated using the reaction-transport model, PRECIP. Key minerals in the bentonite (Na-montmorillonite, analcite, chalcedony, quartz, calcite) were allowed to dissolve and precipitate using kinetic (time-dependent) reaction mechanisms. Simulations were carried out with different model variants investigating the effects of: temperature (25 and 70 � C); cement pore fluid composition; dissolution mechanism of montmorillonite; rates of growth of product minerals; solubilities of product minerals; and aqueous speciation of Si at high pH. Simulations were run for a maximum of 3.2ka. The results of all simulations showed complex fronts of mineral dissolution and growth, driven by the relative rates of these processes for different minerals. Calcium silicate hydrate (CSH) minerals formed closest to the cement-bentonite boundary, whereas zeolites and sheet silicates formed further away. Some growth of primary bentonite minerals (analcite, chalcedony, calcite and montmorillonite) was observed under certain conditions. Most alteration was associated with the fluid of highest pH, which showed total removal of primary bentonite minerals up to 60 cm from the contact with cement after � 1 ka. The maximum porosity increase observed was up to 80–90% over a narrow zone 1–2cm wide, close to the cement pore fluid- bentonite contact. All simulations (except that with alternative aqueous speciation data for Si) showed total filling of porosity a few cms beyond this interface with the cement, which occurred after a maximum of 3.2ka. Porosity occlusion was principally a function of the growth of CSH minerals such as tobermorite. There was very little difference in the alteration attained using different model variants, suggesting that bentonite alteration was not sensitive to the changes in parameters under the conditions studied, so that transport of pore fluid through the bentonite governed the amount of alteration predicted. Principal remaining uncertainties associated with the modelling relate to assumptions concerning the evolution of surface areas of minerals with time, and the synergy between changing porosity and fluid flow/diffusion. # 2002 Elsevier Science Ltd. All rights reserved.

The potential for aquifer disposal of carbon dioxide in the UK

Natural analogues indicate that it is possible to dispose of CO2 underground in closed structures on deep aquifers. Disposal into depleted or exhausted hydrocarbon fields has many advantages, e.g. proven seal, known storage capacity, no exploration costs. Unfortunately there are very few hydrocarbon fields in the UK onshore area, and their total CO2 storage capacity is very low compared to annual UK CO2 production from power generation. The best aquifers for CO2 disposal onshore are the widespread Permo-Triassic sandstones. Further onshore potential exists in younger Mesozoic reservoirs. Offshore, disposal into depleted oil fields (where cost credits from enhanced oil recovery could be beneficial) or the Perno-Triassic gas fields of the southern North Sea, and nearby associated closures of the Triassic Sherwood Sandstone aquifer, appear to provide the best prospects.

Rate and mechanism of the reaction of silicates with cement pore fluids

Abstract The reaction mechanisms and rates of reaction of a number of the common rock-forming silicates with synthetic cement pore fluids have been evaluated in a series of laboratory experiments at 70°C. Mass transfer is dominated by the dissolution of the primary silicate and the precipitation of a range of Na-K-Al substituted calcium silicate hydrates, and a possible zeolite. Calcium was lost from, and silicon gained by, the fluid phase as a result of the reactions. Secondary solids formed thick layers on primary silicates, but dissolution of the silicates was not diffusion-limited. The rate of dissolution of the silicates was determined to be 2–3 orders of magnitude greater at pH 12–13 than at neutral pH, and confirm measurements by other authors. The rate of growth of calcium silicate hydrates was limited by the rate of supply of silicon from the primary silicates. Although the results of the laboratory experiments were dominated by the loss of calcium from the fluid and the precipitation of calcium silicate hydrates, thermodynamic modelling suggests that these may be replaced by zeolites and/or feldspars when groundwater residence times are considered.

High pressure metamorphism in the Scourian of NW Scotland: Evidence from garnet granulites

Ultramafic and mafic granulites from Archaean gneisses in N.W. Scotland (the Scourian) show evidence of two periods of granulite facies mineral growth. The first produced a high pressure clinopyroxene +garnet±plagioclase assemblage at an estimatedP-T of 12–15 kb and 1,000° C. Uplift of the complex caused partial breakdown of the garnet by reaction with clinopyroxene to produce orthopyroxene +plagioclase ±spinel±amphibole symplectites, at an estimatedP-T of 10–14 kb and 800°–900° C. Garnet stability is shown to depend on both whole-rock Fe/Mg ratios and onP-T conditions. The pressures imply crustal thicknesses in the Archaean of least 35–45 km.

Hydrothermal alteration of granite by meteoric fluid: an example from the Carnmenellis Granite, United Kingdom

The interaction of granitic rock with meteoric fluid is instrumental in determining the chemistry of pore fluids and alteration mineralogy in downflow portions of convective groundwater circulation cells associated with many hydrothermal systems in the continental crust. Hydrothermal experiments and a detailed mineralogical study have been carried out to investigate the hydrothermal alteration of the Carnmenellis Granite, Cornwall, UK. Samples of drill chippings from a borehole 2 km deep in the Carnmenellis Granite have been reacted with a dilute Na-HCO3-Cl fluid in hydrothermal solution equipment at temperatures of 80°, 150° and 250° C and a pressure of 50 MPa, with a water/rock mass ratio of 10, for experiment durations up to 200 days. Fluid samples were analysed for seventeen different chemical components, and solids were examined prior to, and after reaction using SEM, electron microprobe and conventional light optic techniques. Experimental fluids were mildly alkaline (pH 7–8.5) and of low salinity (TDS <800 mgl−1). Mineral-fluid reaction was dominated by the dissolution of plagioclase and the growth of smectite, calcite (at all temperatures), laumontite (at 150° C), wairakite and anhydrite (at 250° C). Final fluids were saturated with respect to quartz and fluorite. Certain trace elements (Li, B, Sr) were either incorporated into solids precipitated during the experiments or sorbed onto mineral surfaces and cannot be considered as ‘conservative’ (partitioned into the fluid phase) elements. Concentrations of all analysed chemical components showed net increases during the experiments except for Ca (at 250° C) and Mg (at all temperatures). A comparison of the alteration mineralogy observed in the experiments with that present as natural fracture infills in drillcore from the Carnmenellis Granite reveals that the solid products from the experiments correspond closely to mineral assemblages identified as occurring during the later stages of hydrothermal circulation associated with the emplacement of the granite.

Layered ultramafic-gabbro bodies in the Lewisian of northwest Scotland: geochemistry and petrogenesis

Abstract Layered ultramafic-gabbro bodies occur widely in the Archaean of northwest Scotland. They were metamorphosed at granulite or high amphibolite facies and were tectonically thinned and broken up during deformation. They comprise repeated ultramafic-gabbro layers, locally with Ni-poor sulphide-rich tops, each rhythmic unit showing decreasing MgO, Ni and normative anorthite with stratigraphic height. Major, trace and rare earth element data are presented for the range of rock types. In ultramafic rocks, MgO varies from 22 to 37 wt.%, Ni from 1000 to 2500 ppm and TiO 2 from 0.08 to 0.40 wt.%, while the MgO content of the gabbros ranges from 14 to 6 wt.%. The REE patterns are flat to LREE enriched with no significant Eu anomalies. In ultramafic rocks REE are from 4 to 10 times chondrite, and in the gabbros LREE range from 8 to 30 times chondrite and HREE from 6 to 15 times chondrite. Study of incompatible elements (Ti, Zr, Y) which are relatively immobile during metamorphism shows that neither garnet nor hornblende were involved in fractionation. Trace element modelling shows it is improbable that the ultramafic rocks represent primary MgO-rich liquids even though their incompatible element contents are quite high. The chemical trends are interpreted in terms of olivine and pyroxene settling from a tholeiitic high-Mg magma with 15–20 wt.% MgO derived by 30–40% partial melting of an undepleted mantle. The ultramafic rocks are the cumulates and the gabbros the derived liquids.

Modelling reactions between cement pore fluids and rock: implications for porosity change

Abstract The migration of groundwater equilibrated with cement from a deep geological disposal facility for radioactive wastes will perturb the chemical, minerological and physical properties of the geosphere in advance of the migration of radionuclides. Preliminary modelling of a simplified scenario has been conducted to assess these changes using appropriate data for mineral dissolution kinetics, the chemical composition of cement pore fluids, and the hydrogeological characteristics of fractured crystalline rock. Chemical exchanges between rock immediately adjacent to the engineered barriers of a waste disposal facility and pore fluids were evaluated using the speciation-reaction path code. EQ 3 6 which revealed rapid loss of Ca, and gains in Na and Si of the evolved fluids, with little change in pH. Secondary minerals show a sequence of calcium silicate hydrates, and zeolites. Precise definition of the overall mass balance is uncertain due to the absence of both thermodynamic data for many zeolites and kinetic data for the precipitation of feldspars. Modelling has demonstrated that reaction kinetics will be important in governing chemical exchanges for length scales up to 20 m. Radionuclide retardation will be enhanced by the growth of zeolites and calcium silicate hydrates.

Performance assessments for the geological storage of carbon dioxide: Learning from the radioactive waste disposal experience

The geological storage of carbon dioxide is currently being considered as a possible technology for reducing emissions to atmosphere. Although there are several operational sites where carbon dioxide is stored in this way, methods for assessing the long-term performance and safety of geological storage are at an early stage of development. In this paper the similarities and differences between this field and the geological disposal of radioactive wastes are considered. Priorities are suggested for the development of performance assessment methods for carbon dioxide storage based on areas where experience from radioactive waste disposal can be usefully applied. These include, inter alia, dealing with the various types of uncertainty, using systematic methodologies to ensure an auditable and transparent assessment process, developing whole system models and gaining confidence to model the long-term system evolution by considering information from natural systems. An important area of data shortage remains the potential impacts on humans and ecosystems.

Granite-water reactions in an experimental Hot Dry Rock geothermal reservoir, Rosemanowes test site, Cornwall, U. K

Abstract Fresh meteoric water was circulated through a jointed granite rockmass beneath the Rosemanowes test site at a depth of 1.6–2.6 km, and at initial rock temperatures of 58–100°C, with the objective of testing techniques for heat extraction from Hot Dry Rock (HDR). Geochemical data from selected circulation experiments over a period of 8 a are presented. The differences between injection and production water compositions changed with time. Early circulation tests yielded production waters depleted in K, Ca and Mg and enriched in Na, SiO2, Cl and alkalinity. Later production waters were depleted in Mg and enriched in Na, Ca, SiO2, Cl and alkalinity. Various processes that might have given rise to these changes are considered, including diffusion from saline pore fluid, cation exchange, mineral dissolution and precipitation reactions, and bacterially mediated reactions. All of these are inferred to have been important, to different degrees at different times and in different parts of the underground system. The major salinity-generating processes in the reservoir are inferred to have been the diffusion of Cl salts from saline pore fluids into the circulation waters and the generation of HCO3 by bacterial oxidation of dissolved and particulate organic matter in the injection water. Chemical transient experiments were performed to demonstrate the role of cation exchange processes. The cation exchange sites are inferred to be on natural clay minerals coating the joint surfaces and possibly additional clay minerals resulting from plagioclase dissolution during circulation. In the early circulation tests, the principal evidence of dissolution reactions was the release of SiO2. Other solutes are inferred to have been controlled mainly by cation exchange. Late in the circulation, it is inferred that dissolution reactions were an important source of Na. Using estimates of joint apperture (∼0.5 mm) calculated from Rn data and mean residence time (∼200 h) calculated from tracer data, an apparent bulk dissolution rate of 2 × 10−11 (kg/m2)/s has been calculated. This implies a mean joint surface corrosion rate of 0.25 μm/a, which is too small to have affected the hydro-mechanical properties of the underground system during its lifetime.

Granite-water interactions at 100°C, 50 MPa: An experimental study

Abstract A monzogranite was reacted with water in hydrothermal solution equipment at 100°C, 50 MPa for 203 days, during which time six fluid samples were extracted at run temperature and pressure for chemical analysis of 18 chemical species. Solids were examined using XRD and SEM and the high-temperature speciation of the fluid phase chemistry was investigated using the geochemical software package EQ3/6. The generated fluid phase was mildly alkaline [ pH (100° C ) = 7.5–8] and low in total dissolved solids ( mg 1 −1 ). Chemical components apparently buffered in solution by the solubility of a solid phase (in parentheses) were: SiO 2 (quartz); K (muscovite—K-feldspar); Ca (calcite); Sr (strontianite); and Ba (witherite). The final fluid phase was under-saturated with respect to most Na-aluminosilicates. Chlorite was the only mineral totally consumed during the experiment, whereas the only identifiable new phase was a complex smectite, the rate of precipitation of which apparently limited Al concentration and pH of the fluid. The implications of the results of this experiment to modelling the geochemical environment of a high-level radioactive waste repository in granite are discussed.