Chlorite: Geochemical Properties, Dissolution Kinetics and Ni(II) Sorption

Chlorite: Geochemical properties, Dissolution kinetics and Ni(II) sorption

Åsa Zazzi

Doctoral Thesis in Chemistry KTH Chemical Science and Engineering Stockholm, Sweden, 2009 ______AKADEMISK AVHANDLING Som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges till offentlig granskning för avläggande av Filosofie Doktorsexamen i Kemi fredagen den 24 april 2009, kl. 10.00 i D2, Lindstedsvägen 5, Entreplan, Stockholm. Fakultetsopponent är Ph. D. Peter Vilks, AECL, Whiteshell Laboratories, Canada. Avhandlingen försvaras på engelska.

Chlorite: Geochemical properties, Dissolution kinetics and Ni(II) sorption

Åsa Zazzi

Doctoral Thesis

KTH Chemical Science and Engineering Royal Institute of Technology Stockholm, Sweden, 2009 ISBN 978-91-7415-247-0 ISSN 1654-1081 TRITA-CHE Report 2009:9

Printed by E-PRINT AB, Stockholm 2009.

När man är en björn med en mycket liten hjärna och tänker ut saker, upptäcker man att en idé som verkade vara riktigt idéaktig inne i hjärnan, är annorlunda när den kommer ut i det fria och andra människor ser på.

Nalle Puh

When you are a Bear of Very Little Brain, and Think of Things, you find sometimes that a Thing which seemed very Thingish inside you, is quite different when it gets out into the open and has other people looking at it.

Winnie the Pooh

Abstract

In Sweden, among other countries, a deep multi-barrier geological repository, KBS-3, is planned for the burial of nuclear waste. One of the barriers is identified as the grantic bedrock itself and in this environment chlorite is present at surfaces in fracture zones. This thesis is focused on characterisation of chlorite samples and studies of their dissolution and sorption behaviour, in order to verify chlorites capacity to retard possible radionuclide migration in the case of leaking canisters. Chlorite dissolution of has been studied in the pH interval 2-12, and as expected the dissolution is highest at acidic pH and at most alkaline pH, whereas dissolution is lowest at near neutral pH values. Chemical and physical properties of chlorites clearly influence the dissolution rates, and at steady-state dissolution rates in the interval 10-12 - 10-13 mol g-1 s-1 was observed. Sorption studies were performed since Ni(II) is one of the important activation products in spent nuclear fuel and sorption data on minerals like chlorite are lacking. Ni(II) sorption onto chlorite was studied using batch technique as a function of; pH, concentration of Ni(II), ionic strength and solid concentrations. As expected, the sorption of Ni(II) onto chlorite was pH dependent, but not ionic strength dependent, with a sorption maximum at pH ~ 8, and with a Kd of ~ 103 cm3/g. This confirms that the Ni(II) sorption onto chlorite is primarily acting through surface complexation. The acid-base properties were determined by titrations and described by a non-electrostatical surface complexation model in FITEQL. Further, the sorption results were fit with a 2-pK NEM model and three surface complexes,

Chl_OHNi2+, Chl_OHNi(OH)+ and Chl_OHNi(OH)2, gave the best fit using FITEQL.

i ii Sammanfattning

Sverige är ett av de länder som planerar ett geologiskt slutförvar kallad KBS-3, bestående av ett antal barriärer, för placering utav det använda kärnbränslet. En av dessa barriärer är identifierad som själva berggrunden där det tilltänkta förvaret kommer att byggas och i denna miljö förekommer klorit på granitytor i sprickzoner. Denna doktorsavhandling karakteriserar kloriter och studerar deras upplösnings- och sorptionsbeetende, för att kunna bestämma huruvida kloriter är utav betydelse som naturlig barriär för eventuell radionuklidtransport från det använda kärnbränslet. Upplösning av klorit har undersökts i pH intervallet 2-12 och graden av upplösningen är som förväntat högst vid sura respektive mest basiska pH och lägst där pH är neutralt. Denna studie bekräftar att den kemiska sammansättning och de fysikaliska egenskaper hos kloriterna påverkar upplösningshastigheterna och vid steady-state har upplösningshastighet bestämts till 10-12 - 10-13 mol g-1 s-1. Sorptionsstudier genomfördes då Ni(II) är en viktig aktiveringsprodukt och data rörande Ni(II) sorption till klorit saknas. Ni(II) sorption till klorit har studerats i; varierande pH, olika initiala Ni(II) koncentrationen, olika jonstyrka och olika fastfas förhållanden där individuella satser i serie har nyttjats. Som förväntat är sorptionen av Ni(II) till klorit pH beroende men inte jonstyrkeberoende och ett sorpions maximum observerades vid pH ~ 8, med ett

Kd-värde på ~ 103 cm3/g. Från detta dras slutsatsen att sorptionen av Ni(II) till klorit sker mestadels genom ytkomplexering. Syra-bas egenskaperna hos kloriterna bestämdes genom titreringar och bekrevs med en icke-elektrostatisk modell i FITEQL. Vidare har passning av sorptionsresultaten utförts med en 2-pK NEM-modell och tre ytkomplex, Chl_OHNi2+,

Chl_OHNi(OH)+ och Chl_OHNi(OH)2, vilket gav den bästa passningen av data med FITEQL.

iii iv List of publications

This thesis is based on the following papers:

I. The effect of pH on chlorite dissolution rates at 25º C Åsa B. Gustafsson1 and Ignasi Puigdomenech In: Scientific Basis for Nuclear Waste Management, XXVI (R. J. Finch, D. B. Bullen, eds.), Material Research Society, Boston, MA, USA, 2002, vol. 757, p. 649-655.

II. Study of Ni(II) Sorption on Chlorite-A Fracture Filling Mineral In Granites Å. Gustafsson1, M. Molera, and I. Puigdomenech In: Scientific Basis for Nuclear Waste Management XXVIII (J.M. Hanchar, S. Stroes-Gascoyne, L. Browning, eds.), Material Reseach Society, San Fransisco, CA, USA, 2004, vol. 824 p. 373-379.

III. Structural Investigations of natural and synthetic chlorite minerals by X-ray diffraction, Mössbauer spectroscopy and Solid-state Nuclear Magnetic Resonance Åsa Zazzi, Tomas K. Hirsch, Ekaterina Leonova, Andrei Kaikkonen, Jekabs Grins, Hans Annersten, and Mattias Edén In: Clays and Clay Minerals; April 2006; v. 54; no. 2; p. 252-265

IV. Ni(II) sorption on natural Chlorite Åsa Zazzi, Anna-Maria Jakobsson and Susanna Wold Submitted to: Applied Geochemistry

V. Ni(II) sorption on the fracture filling mineral Chlorite Åsa Zazzi and Susanna Wold Accepted for publication in: Scientific Basis for Nuclear Waste Management XXXII (R.B. Bebak, N.C. Hyatt and D.A. Pickett, eds). Material Research Society, Boston, MA, USA, 2008, vol 1124.

VI. Dissolution rates and stoichiometry of two different chlorites as a function of pH Åsa Zazzi, Maria E. Malmström and Susanna Wold Manuscript

1 Maiden name, changed to Zazzi by way of matrimony. v vi Comment on my contribution to the publications

Paper I: I performed the experimental work, participated in evaluation of the data and wrote most part of the manuscript.

Paper II: I participated in the design of experiments and performed most of the experimental work. I performed parts of the simulations, participated in evaluation of the data and wrote the manuscript.

Paper III: I provided the chlorite samples, prepared the synthetic chlorite and wrote parts of the Introduction section of the manuscript.

Paper IV: I designed most of the experiments and performed most of the experimental work. I participated in discussions about the simulations as well as in evaluation of data and wrote most of the manuscript.

Paper V: I designed and performed the experiments, evaluated the data and wrote the manuscript.

Paper VI: I designed and performed the experiments, participated in data evaluation and prepared parts of the manuscript.

vii viii List of abbreviations

Abbreviations used in text.

AFM Atomic Force Microscopy BET Brunauer-Emmett-Teller, a method for measuring the surface area of powders CCM Constant Capacitance Model CEC Cation Exchange Capacity DLM Diffuse layer model ICP-MS Inductively Couple Plasma equipped with Mass Spectrometry ICP-OES Inductively Couple Plasma equipped with Atomic Emission Spectroscopy KBS-3 Kärnbränslesäkerhet-3, the Swedish concept for spent nuclear fuel, the abbreviation is always used. LSC Liquid Scintillation Counting MES 2-(N-morpholino)ethanesulfonic acid MUSIC Multi Site Complexation Model NEM Non Electrostatic Model SCM Surface Complexation Model SEM Scanning Electron Microscopy SEM-EDS Scanning Electron Microscopy-Energy Dispersive Spectroscopy SKB Swedish Nuclear Fuel and Waste Management Company TLM Triple Layer Model TOT Tetrahedral-Octahedral-Tetrahedral coordination TRIS 2-Amino-2-hydroxymethyl-1,3-propanediol UV-VIS Ultraviolet-Visible Spectroscopy XRD X-ray Diffraction

Table of contents

Abstract ...... i Sammanfattning...... iii List of publications ...... v Comment on my contribution to the publications...... vii List of abbreviations...... ix 1. Introduction...... 1 1.1 Background ...... 1 1.2 The mineral chlorite ...... 3 1.3 Theory of weathering processes...... 5 1.4 Sorption processes, surface complexation and surface complexation models...... 7 1.5 Literature survey of the area...... 12 1.6 Objectives of this work ...... 14 2. Experimental...... 15 2.1 Chlorites used...... 15 2.2 Reagents ...... 16 2.3 Experimental methods ...... 16 2.4 Treatment of data...... 20 3. Results and discussion...... 23 3.1 Characterisation of the used chlorites used ...... 23 3.2 Dissolution results...... 29 3.3 Sorption results...... 41 3.4 Titration results...... 53 3.5 Fitting the sorption and titration data ...... 55 4. Conclusions...... 59 5. Future work ...... 61 6. Acknowledgements...... 63 7. References...... 67 Appendix...... 73 Appendix A: Experimental details for sorption isotherm experiments...... 73 Appendix B: Summary of dissolution results...... 74 Appendix C: Summary of sorption results...... 78 Appendix D: Table of titration data...... 79 Appendix E: Literature data...... 81

xii 1. Introduction

1.1 Background

Swedish nuclear reactors will produce a total amount of 12 000 tons of spent nuclear fuel, if today’s nuclear power plants will be active for a period up to 60 years [1]. According to today’s Swedish regulations, spent nuclear fuel has to be taking care of within Swedish borders, and the company SKB was created for this purpose. In 1983 they proposed a design capable of storing spent nuclear fuel named KBS-3 [2]. The KBS-3 repositories are planned to be situated approximately 500 m below the surface in the Swedish granitic bedrock. The Swedish repository design is a multi-barrier concept developed for storage up to 100 000 years where the first 1000 years are the most critical for nuclear waste disposal, since a dramatic decrease in radioactivity occurs during that time [3]. KBS-3 consists of four different barriers, Figure 1, where each barrier will act independently of the other. The first barrier is the fuel itself, which has a low solubility in reducing groundwater. The second barrier is a corrosion- resistant canister made of copper and iron to contain the spent fuel. The canisters will be placed in deposition holes and surrounded by the third barrier, bentonite clay [4, 5]. The tunnels will be refilled with a mixture of bentonite clay and the original bedrock. The fourth and final barrier consists of the surrounding bedrock itself. In a worst case scenario any leakage will reach the bentonite clay and radionuclides, fission products or activation products will be transported through the bentonite clay and reach the surrounding geosphere. The fourth barrier should retard the migration of these products through sorption reactions in the near field and will concentrate them to the fracture systems and the wall rock adjacent to these, since the surrounding rock volume has a large sorption capacity coupled with the specific surface area [6]. The activation product 63Ni is estimated to be present in the surrounding area of the repository upon leakage of fuel 300- 1000 years after closure. During this time the maximum calculated activity released into the near field will be approximately 1000 Bq yr-1, peaking at 600 years after closure [7]. Nickel is a small surface-complexing divalent cation and is representative for cations such as Co2+, Mn2+, Cd2+ and Zn2+.

Figure 1.1. A schematic model of the KBS-3 structure designed for storing spent nuclear fuel.

Most of the Swedish bedrock is of crystalline granitic type formed during different intrusions of magma [8, 9] and is stable where mechanical and chemical changes occur extremely slowly [10]. The granitic bedrock consists of a number of closely related rocktypes2 e.g. different granitoids consist of different silicates, among those chlorite [11, 12]. Mapping of these fracture zones, with their minerals and adjacent rock walls, is important for example in choosing a building area with low frequency of fractures for the repository. Different areas of Swedish granitic bedrock have been investigated, the main sites of the geological investigations being Äspö, within the underground Hard Rock Laboratory (HRL) [13, 14] and the two areas of interest for future repositories, namely Forsmark and Laxemar/Simpevarp (divided into two sub-areas) [15, 16]. Mapping of drillcores shows that only 0.2% of Forsmark granite consists of chlorite, whereas Ävrö granite from Laxemar contains 4.4%, but chlorite is still one of the dominant minerals in fractures, comprising 30-70% of the fracture surface [15]. In the tunnel situated within the

2 A rocktype is defined by a number of different minerals. A mineral is a naturally occurring, homogenous solid with a defined chemical composition and highly ordered atomic arrangement (The new penguin dictionary of Geology, second edition). In a more general way minerals are the building blocks of the Earth formed by the history and show us the diversity of formation. 2 Äspö HRL, the fresh granite consists of 10% biotite and 1% magnetite, whereas altered granite consists of by 5-10% chlorite, 0.1% pyrite and 1-2% hematite [17]. The fracture coatings is composed of 35% chlorite, 13% epidote, 0.2% pyrite, 30% of calcite and 18% other clay minerals [17].

1.2 The mineral chlorite

Chlorite is a phyllosilicate arranged in a 2:1 structure type with an interlayer, Figure 1.2. The ideal structure of chlorite can be described as alternating talc-like layers (TOT) together with brucite-like layers (O), which give it a unit structure of 14 Å in the stacking direction [18].

Figure 1.2. The unit structure of chlorite.

Within the structure the TOT-layer has a negative permanent charge, whereas the brucite layer has a positive permanent charge [19]. Major contributors to surface charge are defects in the lattice and isomorphous substitution [19]. Isomorphous substitution occurs in the crystal lattice of the mineral, for example when Si4+ is replaced by Al3+ in the tetrahedral layer of clays. In addition, in the octahedral layer Al3+ may be replaced by divalent cations, such as Mg2+ [20]. In isomorphous substitution the replacement occurs between atoms of 3 similar sizes but with different charges, as described above. Meanwhile, the configuration of oxygen and hydroxide groups stays essentially unchanged. Chlorite may be a product resulting from hydrothermal alteration of pyroxenes, amphiboles and biotite. The overall alteration reaction of biotite to chlorite can be described as [21]:

Biotite + anorthite + H2O + O2 + H+ ≠ chlorite + sphene + epidote + muscovite + quartz + magnetite + K+ where the K+ from the biotite diffuses through the fluids, responsible for the alteration, and water within the bedrock provides H+, which diffuses into the interlayer of biotite and substitute for K+. When the K+ diffuses out from the structure the attached H+ weakens the Si-O or Al-O bonds and a brucite-like layer replaces the K interlayer while the talc-like layer is inherited directly from biotite [22, 23]. The process requires two biotites to become one chlorite with a volume loss of approximately 35% [21]. The colour of the chlorite varies from white to almost black or brown with a tint of green where these optical properties of chlorites are coupled to the chemical composition of chlorite. An increase in Fe/(Fe+Mg) ratio is followed by an increase in absorption, together with the Fe2O3 content, degree of oxidation, total iron in octahedral sites and Si/Al ratio in tetrahedral sites, different schemes describing the optical properties [24-26]. Using these schemes, Mg-Al rich varieties of chlorites are more or less colourless whereas Fe-rich members are of different green colours and Mn and Cr segments add orange-brown, pink or lavender colours to the chlorites. The family of chlorite minerals has the generic formula [19, 27, 28]:

(R2+6-y-zR3+yz)2(Si4-kR3+k)2O20(OH)16 (1.1) where the parameters y and k denote the degree of substitution of trivalent cations in the octahedral and tetrahedral sheets, respectively, and z accounts for vacancies [29]. In general, the cation distribution of divalent (R2+) and trivalent (R3+) ions is not a priori known either within each layer or between them. Nevertheless, the net formula, Equation 1.1, may schematically be decomposed into contributions from the two types of alternating layers. The composition of the brucite-like layer conforms to the formula:

2 + 3 + )(OH)12 (1.2) (R R y z1 6− y1−z1 1 whereas the generic composition of the 2:1 layer is

4 2 + 3 + 4-k 3+k 2 20 4 ( z2 )(Si R ) O (OH) , (1.3) R 6− y 2−z2 R y2

and the coefficients combine as y1+y2=y and z1+z2=z. R3+ in the tetrahedral sheets is usually Al3+. However, other cations are also often present. Expressing the general formula as Equation 1.1 implies that the chlorite is trioctahedral in both the 2:1 layer and the brucite-like layer, assuming eight tetrahedral cations and twelve octahedral cations per unit cell. A chlorite may be dioctahedral in the two sub-structures, as well as a combination of di-and trioctahedral [30]. The most common natural form of chlorite include clinochlore, chamosite and penninite [19], which differ in the nature of the dominant divalent octahedral cation; Mg2+ in clinochlore, Fe2+ in chamosite and Mn2+ in penninite.

1.3 Theory of weathering processes

The phenomenon of weathering can be divided into chemical weathering and physical weathering, where physical weathering is the result of wind, water and root abrasion, where part of the rock is worn down without any change in its elemental composition. Chemical weathering is a dissolution process through chemical reactions, where a release of ions gives a change in chemical composition. For most silicate minerals the dissolution and dissolution rates are strongly pH-dependent [31]. A schematic relationship between the dissolution rate for silicates and pH is shown in Figure 1.3. In the acid and basic pH regions, the dissolution rates increases while in the near neutral pH region the rate is independent of pH.

5 Log DissolutionLog Rate

0 246128 10

Figure 1.3. Schematic relationship between dissolution rate and pH for silicates.

Dissolution is a process undertaken in a number of different steps [32, 33]: 1. Migration of the reactants (H+, OH- or ligands) to the surface. 2. Surface adsorption of the reactants. 3. Formation of surface species. 4. Detachment of the surface species from the surface. 5. Transport of reactants from surface to solution. The rate-determining step is attachment of a reactant to the surface [32]. Rearrangement may occur where the coordination of the metal ions changes and bonds are weakened or broken, which facilitates the detachment of the surface metal species into the bulk solution. These steps indicate surface-controlled dissolution, Equation 1.4 and 1.5 [34]. Hydrolysed metal surface sites + reactants (H+, OH- or ligands) → surface species (1.4) Surface species → Me(aq) (1.5)

This surface-controlled model was developed by Stumm and co-workers and when the system is not in equilibrium, the back reactions can be excluded [35]. Silicates have a non-stoichiometric dissolution i.e. the different cations detach from the mineral surface at different rates [32]. Chemical weathering can be described as congruent or non-congruent dissolution. In congruent dissolution no new solid phases are formed after dissolution and the dissolution

6 is complete where the products formed are all soluble. In contrast, incongruent dissolution is an exchange process and gives rise to new solid phases. An example of congruent dissolution is the dissolution of quartz:

SiO2 (s) + 2H2O (l) → H4SiO4 (aq) (1.6) whereas an example of incongruent dissolution is the weathering of olivine, where quartz is formed as a precipitate [20]:

Mg2SiO4 (s) + 4H+ (aq) → 2Mg2+ (aq) + SiO2 (s) +2 H2O (l) (1.7)

A steady state is achieved when there is no concentration change over time. However, in experimental studies true steady-state may not be reached within a reasonable experimental period of time [36]. Therefore, steady state under experimental conditions is denoted as the time when the cation concentration within the outflow solution is nearly constant [37] [36].

1.4 Sorption processes, surface complexation and surface complexation models

There are a number of different processes taking place in the surface water interface region, such as adsorption, surface precipitation and ion exchange. In terms of surface complexation, adsorption reactions are described as chemical reactions between surface functional sites and dissolved chemical species, where the protonation of surface oxygen and deprotonation of hydroxyl groups at the surface take place. For ion exchange, and more specifically cation exchange, the reactions can be described as an interchange reaction between an ion in solution and another ion with electrostatic attraction. Cation exchange and surface complexation reactions may take place simultaneously, where sorption through cation exchange is generally pH independent and sorption through surface complexation is pH dependent [38]. Sorption may be used as a more general term instead of adsorption. The term sorption can be used when investigations are performed on the macroscopic level and not combined with microscopic studies [39].

7 A general cation exchange reaction can be written as: XH + Me+ XMe + H+ (1.8) where XH is a cation-exchange site [40].

≡SOH or ≡XOH are common symbols to denote a general neutral surface site. A cation exchange reaction at the surface can be written as: ≡SOH + Me+ ≡SOMe + H+ (1.9) and a general surface complexation reaction can be written as:

≡SOHrn + Mez+ + yH2O ≡SOHcMe(OH)yr+z-y +(n-c+y) H+ (1.10) where ≡SOHrn is a surface species with charge r, Mez+ is a sorbing cation, y is the number of protons released in the reaction and ≡SOHcMe(OH)yr+z-y is the surface complex formed.

For the surface precipitation the general reaction can be written using S(OH)3(s) as the hydrous oxide:

≡SOH + Me2+ + 2H2O S(OH)3(s) + =MeOH2 + H+ (1.11)

=MeOH2 + Me2+ + 2H2O MeOH2(s) + =MeOH2 + 2H+ (1.12)

The uptake and release of protons of a hydrolysed surface complex describes the acid-base reactions which take place at the surface. These reactions can be described either by two acidity constants called the 2-pK approach, which is more commonly used [41], or by one acidity constant, the 1-pK approach.

The reactions of the 2-pK model can be written as:

≡SOH + H+ ≡SOH2+ (1.13) ≡SOH ≡SO- + H+ (1.14)

The reaction with only one step can be written as:

≡SOH-0.5 + H+ ≡SOH2+0.5 (1.15)

The acid-base properties of the surface hydroxyl groups are affected by the ionic strength used and the permanent charge of the surface itself.

8 There are mathematical models such as Freundlich or Langmuir which applied to the sorption system describe the metal sorption as a function of the equilibrium concentration in solution [20, 42]. The Langmuir model assumes that the sorption takes place in a monolayer on a homogeneous surface, that the binding sites are evenly distributed over the surface, that all of them have the same adsorption affinity and that surface complexation is the sorption mechanism. The Freundlich model is an empirical model that can be applied to heterogeneous surfaces where all sites are not considered to be equal and no monolayer of adsorption is assumed. Instead accumulation of the adsorbate and no sorption maximum are achived with this model. Freundlich does not assume any specific sorption mechanism and is found to be most suitable in trace metal concentrations because surface precipitation possibly takes place [43]. The cation may attach to the surface in different ways, as an inner sphere, as an outer sphere or as a diffuse swarm of ions within the double layer [44, 45]. When an inner sphere is formed there is a bond between the cation and the electron-donating ions at the surface. When an outer sphere complex is formed, the ion pairs between the cation and the surface are separated by one or more water molecules. Furthermore, the outer sphere complexes are less stable than the inner sphere complexes since they involve electrostatic bonding whereas inner sphere complexes involve covalent bonds and some ionic bonds. To identify whether inner sphere or outer spheres complexes are formed different spectroscopic techniques are used, for example EXAFS, since this technique enables the atomic distances between adsorbed species and neighbouring cations to be seen [46]. A simpler way is to study the sorption of the cation as a function of ionic strength, since a dependence on the ionic strength is assumed to be the result of outer sphere complex formation [47].

To predict the experimental outcome involved and to understand the different reactions involved in surface complexation, several models have been developed. The Surface Complexation Model (SCM) describes the reactions in relation to the charging properties of mineral surfaces and adsorption mechanisms. Within the SCM it is assumed that the adsorbing ion forms surface complexes at the specific sorbing sites on the surface and different SCM use different descriptions for the distribution of the ions near and around the charged surface. Electrostatic effects and surface charge are incorporated and introduced in different ways in the different models. The surface charge can be determined with different

9 methods; titration with acid or base or applying an electrical field to the surface. If titration data are used for the determination, the surface charge can be calculated according to:

σ = F(ΓH+ - ΓOH-) (1.16) where F is Faraday constant (96490 C mol-1) and ΓH+ and ΓOH- are the adsorption densitiy of H+ and OH- (mol m-2).

The surface complexation models have been thoroughly discussed [46, 48-50] and here only a short summary is given with the major differences between the different models. The abbreviations for the different models are used in the text. The models summarised are; the Constant Capacitance model (CCM), the Diffuse Layer Model (DLM), the Triple Layer Model (TLM) and a Charge-Distribution Multi Site Complexation Model (CD_MUSIC). The differences are mainly coupled to how the charges and potential relationship are included and corrected for in the mass action equations for surface equilibrium see Figure 1.4. In DLM the protonation and deprotonation as well as ion adsorption, occur in only one plane in the interface region (the o-plane), and only these specifically sorbed ions contribute to the total surface charge density. CCM and DLM both assume only one layer of interface region but in CCM the charged surface is assumed to be separated from the bulk solution by a layer of constant capacitance. In TLM the different reactions take place in three different planes. The deprotonation and protonation naturally take place in the inner plane and specific sorption in the middle plane whereas the outer plane is described as the diffuse layer. CD-MUSIC is based on crystallographic knowledge and within this model the different crystal planes (the surface basal planes and the surface edges) are included. NEM does not include any charge or potential relationships instead the adsorption is treated purely as chemical reactions. The generalised two-layer model was introduced by Dzombak and Morel [49] and this model is based on the DLM model but includes two site types for cation binding. These different surface sites are classified as strong or weak sites, depending on the difference in the affinity to bind a sorbing ion or molecule. The weak sites, with a low binding affinity, correspond to the total number of reactive sites available for sorption of cations and are defined by their observed sorption maximum. The strong sites, with a high binding affinity, are defined by their sorption isotherms, meaning the point where the sorption density has a slope lower than 1.0 in a log-log isotherm plot [49]. The

10 concentration of strong and weak sites together corresponds approximately to the concentration of proton binding sites [49]

Figure 1.4. Summary of surface complexation models desctibed in the text.

A difficulty when applying SCM is the distinction between surface complexation and surface precipitation, reactions that may occur in parallel [42, 51]. The main differences are that surface complexation takes place in just one single layer, while surface precipitation can form multiple layers, and that ions complexed to the surface behave differently to precipitated ionss and thus surface precipitation requires a different model than surface complexation (SC). It is very difficult to determine the breakpoint where the SC monolayer formation changes to surface precipitation, in multi-layers [52]. However, Dzombak and Morel [49] introduced a rule of thumb for when surface precipitation should be considered,

11 namely if the dissolved sorbate concentration exceeds (i) one-tenth of its solubility or (ii) one-half of the total surface site concentration.

To compute the different models, different types of software such as FITEQL [53] and PHREEQC [54] are frequently used.

The pH where the net surface charge is zero is called zero point of charge (ZPC) or point of zero charge (PZC). If the change in surface charge is only dependent on the adsorption of H+ or OH-, the point of zero net proton charge (PZNPC) or the isoelctric point (IEP) is defined [48, 55]. Stumm and Morgan [44] reported a pHPZNPC for Mg-silicates in the pH range 9-12.

1.5 Literature survey of the area

Chemical weathering has been thoroughly studied in terms of dissolution rates and dissolution kinetics and the dissolution kinetics of sheet silicates have been reviewed [56]. The silicates studied include kaolinite [57], biotite [58] and smectite [59] and some studies even involve chlorite. The early publications regarding chlorite weathering, published in the mid-1950s, include studies of alteration of chlorite to vermiculite, using techniques such as XRD. It was found [60, 61] that one parameter has an impact on this alteration, the Fe(II) content, with a chlorite with high Fe(II) altering faster and directly to vermiculite. Recent studies of chlorite dissolution are listed in Table 1.1.

12 Table 1.1. Recent publications on dissolutions rates of chlorites. Reference Reference number pH Dissolution rates Comment (mol m-2s-1) Brandt et al. [62] 2-5 10-12 and 10-13 Also AFM Hamer et al. [63] 3.5 8.28·10-12 (nitric Organic and inorganic acid) acid solutions 25.53·10-12 (citric acid) Ross [61, 64, 65] < 1 10-9.29 Based on graphical data for Mg May et al. [66] 5 10-12.52 Malmström [67] 8.2 10-11.59 Flow through reactor et al. technique Rochelle et [68] 9.9 25°C 10-12 25°C and 70°C al. 10.2 70°C 1.5·10-11 Lowson et al. [69] 4 10-11.66 Using buffers and flow through cell

Like dissolution, sorption has been extensively studied using both macroscopic and microscopic techniques. Experimental studies of cation and anion sorption to a number of oxides and minerals have been reviewed [41, 49, 50]. Several studies regarding cation sorption on chlorite have been reported, but Ni(II) sorption on chlorite is scarcely studied, Table 1.2.

Table 1.2. Sorption studies onto chlorite. Reference Reference Cation Comment number studied Shawan and Erten [70] Co2+ and Cs+ Different temperatures,30 ,40 , 50 and 60°C . Data represented by Freundlich isotherm. Bond et al. [71] U(VI) and Pu Sorption maximum at pH ~ 8. TLM used. Zorn [72] U(VI) Sorption maximum at pH ~ 6.5. DLM used Krawczyk-Bärsch et [73] U(VI) Secondary Fe-oxyhydroxide al. phases formed during dissolution of chlorite Li et al. [74] Cs+and Yb3 Concentrations used are in the (article in Chinese) 10-6 and 10-3 M rangr. Data represented by Freundlich isotherm. Eylym et al. [75, 76] Ba2+ Using synthetic groundwater Koppelman et al. [77] Ni(II) Using XPS technique

13 Ni(II) sorption onto other surfaces has also been studied, as listed in Table 1.3.

Table 1.3. Ni-sorption onto different mineral surfaces. Reference Reference Surface studied Comment number Bradbury and [78, 79] Na-montmorillonite Sorption edge and isotherms Baeyens studies Scheidegger et al. [80] Pyrophyllite, X-ray absorption fine structure (XAFS) Dähn et al. [81] Montmorillonite P-EXAFS technique

Titration using suspensions of oxides or minerals is the traditional way to determine their acid-base properties and a large number of minerals, such as goethite, boehmite and hydrous ferric oxide, have been thoroughly investigated using this technique [41, 49, 50]. Chlorite, on the other hand, is scarcely studied and this is mainly due to the structural features of chlorite, primarily due to the high number of existing hydroxyl groups in the structure [72].

1.6 Objectives of this work

The aims of this work were to characterise chlorite samples and study their dissolution and sorption behaviour in order to verify the capacity of chlorites to retard radionuclide migration from the repositories for spent nuclear fuel. These aims were achived through the following specific objectives: Characterisation of chlorite - by its composition. Determination of the dissolution rate of different chlorites in a wide pH range. Determination of the sorption capacity of chlorite for Ni(II), a divalent metal cation, under various experimental conditions. Determination of the acid-base behaviour of chlorite by titrations.

2. Experimental

2.1 Chlorites used

Table 2.1 shows the origin and source of the natural chlorites used. The chlorites are named after their place of origin with the exception of KOV:01, which is the name of the drill core.

Table 2.1. Origin and source of the natural chlorites used. Sample Origin Source Catalog number Taberg Taberg, Värmland, Swedish Museum of 89530 Sweden Natural History, Stockholm, Sweden Karlsborg Karlsborg, Swedish Museum of 630491 Västergötland, Natural History, Sweden Stockholm, Sweden FlagStaff Hill FlagStaff Hill, USA Source Clay Repository, CCa-2 University of Missouri, Colombia, USA KOV:01 Oskarshamn, SKB, Sweden Drillcore KOV:01 Småland, Sweden

The chlorite pieces from the collections weighed approximately 30 grams and since experiments were performed with powder the chlorites were first mechanically crushed to pieces of ~1 cm2 and a couple of millimetres thick, then either treated with liquid N2 in cooling-heating cycles or further treated by mechanical crushing using a mill. The powder was then dry-sieved into different size fractions. During the mechanical treatment and the sieving, the finest fraction was repeatedly removed so that only the larger mineral particles were treated. In order to remove ultra-fine particles, the chlorite powder was ultrasonically washed in ethanol.

The synthetic chlorite was prepared from SiO2, γ-Al2O3 and MgO in different ratios placed inside an autoclave at 1.2 kb and 650°C for approximately 7 weeks [82].

The chlorite from the KOV:01 core pieces was scraped off the surfaces using a carbide blade since a iron free tool was preferable in order not to contaminate the samples with Fe.

15 2.2 Reagents

Milli Q water was used in all solutions and either NaClO4⋅H2O (Merck p.a.) or NaCl (Merck p.a.) was used in different concentrations as background electrolyte throughout the experiments. pH adjustments were made with small additions of HClO4, HCl, HNO3 or NaOH (analytical grade). For the sorption isotherms, the buffers TRIS and MES were used. The tracer 63Ni was used in the sorption experiments and was prepared from a stock solution of 63Ni in Ni(II)Cl2 with an activity of 740 MBq mL-1 (PerkinElmer LifeScience, Inc.). Tracer solutions were prepared by adding small amounts of the 63Ni solution to inactive solutions of Ni(NO3)2 at a concentration of 10-4, 10-5 or 10-6M, resulting in a total Ni concentration in the experimental tube of 10-6 M, 10-7 M or 10-8 M.

2.3 Experimental methods

2.3.1 Analytical techniques

The specific surface area was determined by the BET method [83] using a Micrometrics

Flow Sorb II with N2 as adsorbing gas. The relative error from the BET measurement has been evaluated as 10-15% [59].

The amounts of elements released i.e. Mg, Fe(II,III), Al and Si, were determined using ICP-AES with a iCAp 6500 Themo Fischer instrument (Nuclear Chemistry, Chalmers Technical University, Gothenburg) or with a Applied Research Laboratory (ARL) model 3520 B.

In the sorption experiments, the Si concentration was instead determined colorimetrically by the molybdate method [84] using UV-VIS spectrophotometry with a WPA lightwave S2000 and plastic cuvette.

In the CEC experiments, the Cu(II)-complex concentration [85] was determined by UV-VIS using a Varian Cary 300 at 620 nm in a 10 mm cuvette using Milli-Q water as blank.

The Fe2+/Fe3+ ratio was determined by 57Fe Mössbauer spectroscopy at room temperature at the Department of Earth Science, Solid Earth Geology, Uppsala University, Sweden. The data collection took 1-2 days depending on the iron content of the chlorites. The Fe2+/Fe3+ ratio of the chlorites investigated was calculated from the fitted areas under the absorption doublets assuming similar recoil-free fractions of the iron sites.

β-activity of 63Ni in aqueous phase was measured using Beta LSC on a Beckman LS1801 instrument or a PACKARD Tri-Carb Liquid Scintillation Analyser Model 1500.

The SEM analysis was performed either with XL 30 ESEM-FEG with a back-scattered electron detector situated at the Department of Geology and Geochemistry, Stockholm University (Stockholm, Sweden) or with Zeiss DSM 940 with EDS (Oxford Instruments link) or Hitachi S-3400N SEM equipped with an INCADryCool Energy Dispersive X-Ray Spectrometer (EDS) both at the Department of Earth Sciences Centre, University of Gothenburg, Sweden) or with a JEOL JSM 6490LV with SEI and EDS detector at Inorganic Chemistry, KTH, Stockholm, Sweden.

XRD was used for identification of chlorite as well as for identification of impurities and other phases and other minerals in the different samples prior to the experiments. A Rigaku powder diffractometer and a DANalytical Xpert pro with software packages were used.

The AFM images were taken using a Veeco DIMENSION 3100 system with a Nanoscope IV controller at INE, Karlsruhe, Germany, and images were recorded using contact mode. Grains of chlorite were attached to a sample holder with a thin layer of Tempfix® (Neubauer Chemicals, Germany), using a method described by Bickmore et al. [86]. The AFM images obtained were later processed using the WSxM software (Nanotech Electronica, Spain).

Solid state NMR was performed by one of the co-authors of Paper III. The instruments used were Varian/Chemagnetics Infinity-200 and 400 spectrometers operating at 4.7 T and 9.4 T respectively, and the work was carried out at the Department of Physical Chemistry, 17 Arrhenius Laboratory, Stockholm University, Stockholm, Sweden. The experimental details are given in Paper III and by Edén et al. [87] and Levitt [88] pH was measured using a Mettler-Toledo InLAB423 combined electrode, saturated with NaCl or with a Metrohm 713 Ph-meter and a Sentron Steam-Line IntelliProbe combined glass/reference electrode filled with 3 M LiCl or with an Hamilton slimtrode.

Further details concerning the analytical techniques are given in Papers I-VI.

2.3.2 Experimental set-up The weathering experiments were performed with a thin-film continuous flow-through technique [58, 89] in ambient atmosphere, Figure 2.1. The flow rate was adjusted with a peristaltic pump (ISAMETC IPC High Precision Multichannel Dispenser) to 5 or 2.7 mL h-1. The flow rate was determined for every specific sample using the sampling time and weight of the sample. Samples were regularly taken during a period up to 30 days. The pH of the inlet and outlet solutions was regularly measured during the experimental period.

Figure 2.1. The experimental set-up used for the weathering experiments.

The sorption experiments were performed as individual samples using a batch sorption technique at room temperature, Figure 2.2. The experiments were performed inside a glovebox in an Ar atmosphere. The sealed tubes were centrifugated in the ambient atmosphere and then passed into the glove-box for sampling. The pH measurements were performed after sampling, in ambient atmosphere.

18 Figure 2.2. The experimental set-up used for the sorption experiments.

The Ni(II) sorption was studied as a function of pH, concentration of sorbent, concentration of sorbate or concentration of background electrolyte, with a reaction time of 7 or 14 days.

The acid-base properties of the chlorite surface were studied using two types of titrations, conventional continuous titration using a suspension of the chlorite powder and batch titration on the supernatant with varying periods of contact with the chlorite surface.

All titrations were performed in a glove-box in an N2 atmosphere using a polypropylene titrations vessel, Figure 2.3. The time between the additions was controlled by an automatic titrator, and volumes of the additions were controlled by computer software.

More experimental details concerning the different techniques are given in Papers I-VI and in Appendix A.

19 Figure 2.3. The experimental set-up used for the titration experiments, inside a glove-box.

2.4 Treatment of data

The distribution of the sorbate between the surface and the aqueous phase can be calculated in different ways. One way is to use the distribution coefficient (Kd) which can be calculated according to the following formula:

3 (Ci − C f )V ⎡cm ⎤ K d = ⎢ ⎥ (2.1) C f m ⎣ g ⎦ where Ci and Cf are initial and final activity in solution, V is the volume of solution (cm3) and m is the mass of dry chlorite (g). The values of Kd obtained from this formula are expressed in cm3 g-1. Since Kd is system-specific [46], it is important to take the distribution between the solid phase and the solution into consideration. Percentage sorption can be calculated using the equation: 100(C − C ) sorbed(%) = i f (2.2) Ci The percentage of sorbed metal ion does not relate to the actual concentration of sorbed ion, since the percentage of sorbed ions is a comparison between the initial concentration and the concentration after reaction. The degree of reversibility of sorption is the relationship between sorption and desorption, where the fraction really sorbed can be described by: 20 A ⋅100 the fraction of cation sorbed(%) = d (2.3) As where Ad is the amount of desorbed cation and As is the amount of sorbed cation before the desorption experiments.

The Freundlich isotherm equation is :

n Q = K F Ce (2.4)

where Q is the equilibrium concentration of the adsorbed ion (mg g-1) and Ce is the equilibrium concentrations in the liquid phase (mg L-1). KF is the Freundlich adsorption coefficient and n is the Freundlich exponent. The Freundlich isotherm equation can be written in the linear form as :

logQ = log K F + nlogCe (2.5) where KF and n can be calculated from the slope and intercept of the linear plot where KF is related to the amount sorbed and n the sorption intensity [90]. If n = 1 KF = Kp, where Kp is the partition coefficient [42].

The dissolution rate for separate elements within the chlorite as well for the chlorite in general can be expressed in a number of equations. The release rate of an individual element, rj is calculated based on Equation 2.6 and as dissolution rate of the chlorite mineral, Rj, see Equations 2.7. The surface area normalised release rate of an element in the chlorite can be calculated according to:

Fi ci, j V dCi, j Fi ci, j ⎡moles of cation ⎤ r = − reactor ≈ i, j ⎢ 2 (2.6) m A mA dt m A ⎣ s m ⎦⎥

where Fi is the flow rate (L s-1); ci,j is the concentration (mol L-1) of element j in the sample i; m is the mass of the dry mineral powder (g) and A is the specific surface area (m2 g-1).

The dissolution rate of chlorite based on element j can be calculated according to:

ri, j ⎡moles of chlorite⎤ R = i, j ⎢ 2 ⎥ (2.7) Pj ⎣ s m ⎦

21 where Pj is the stoichiometric coefficient for element j (moles of element j per moles of chlorite).

The accumulated dissolution of chlorite based on element j up to the time ti from the start of the experiments can be described by:

t i ⎡moles of chlorite⎤ η (t ) = R (t) dt ≈ R ()t − t (2.8) j i ∫ j ∑ i , j i i−1 ⎢ 2 ⎥ 0 i ⎣ m ⎦ where Rj is the dissolution rate for each element and ti is the time elapsed from the start of the experiments.

The dependence of the dissolution rate as a function of pH, taking silica into consideration, can be described by the empirical rate law:

+ m − n R j = kH+ []H + k0 + kOH− [OH ] (2.9) where ki are the rate constants and m and n are empirically determined constants [31, 32, 91].

Depending on the model for fitting sorption data, a number of input parameters are needed; surface site density, surface acidity constants, hydrolysis constants and the concentration of sorbate. For the electrostatic models, values for capacitance and dielectric constant are also needed [42, 48]. Minor changes in some parameters, such as surface site density, have a large impact on the results in the fitting procedure, and therefore these parameters should be changed with caution.

22 3. Results and discussion

3.1 Characterisation of the chlorites used

The characterisation results are general and are not included in one specific publication. However, Papers III and V includes some of the characterisation results.

The sample from FlagStaff Hill was light green, Taberg sample darker green with black intrusions, KOV:01 was also dark green with black parts and the Karlsborg sample was predominantly black. The iron content affects the colour and for the chlorites studied the higher the Fe content the darker the colour [26].

The specific surface areas for Taberg and Karlsborg chlorites are listed in Tables 3.1 and 3.2.The limited amounts of KOV:01 chlorite made BET determinations impossible to perform for that sample. Table 3.1 lists the surface areas determined for different fractions obtained by liquid nitrogen crushing, whereas Table 3.2 shows the surface areas of mechanically crushed chlorite. These results show that the fractions used for the experiments were nearly identical, regardless of crushing method, which is particularly interesting since it has been implied that the mechanical crushing procedure influences the particles shape and increases reactivity, whereas liquid nitrogen will break along natural grain boundaries [92].

Table 3.1. Specific surface area of Taberg and Karlsborg, liquid N2 crushing. Fraction µm Taberg m2 g-1 Karlsborg m2 g-1 63-118 7.95 ± 0.46 0.59 ± 0.05 118-180 6.73 ± 0.65 0.51 ± 0.06 180-250 6.61 ± 0.79 0.51 ± 0.05 250-335 6.06 ± 0.62 0.48 ± 0.05 335-425 5.22 ± 0.50 0.37 ± 0.01

Table 3.2. Specific surface area of Taberg and Karlsborg, mechanical crushing. Fraction µm Taberg m2 g-1 Karlsborg m2 g-1 118-180 6.68 ± 0.73 0.50 ± 0.07 180-250 6.55 ± 0.59 0.49 ± 0.05

23 The range of relative error from BET measurements for the fractions investigated was estimated to be 4-13% which is in agreement with earlier evaluations on silicates [59].

Karlsborg Taberg Malmström et al 9.00 Brandt et al 8.00 Lowson et al Hamer et al 7.00 Rochelle et al Sverdrup 6.00 /g) 2

5.00

4.00

surface area (m 3.00

2.00

1.00

0.00 50 150 250 350 size (μm) Figure 3.1. Surface area as a function of particle size, for the actual chlorite samples together with literature values, references listed in Appendix E.

When the values for Taberg and Karlsborg were compared with other reported surface areas of chlorite, Taberg was the chlorite with the largest surface area. Malmström et al. [67] reported values half those of Taberg, but they used an another sorbing gas. The specific surface are was lower than the area reported by Brandt et al. [62] of 1.1 m2 g-1 but quite similar to that of the chlorite used by Rochelle et al. [68] of 0.89 m2 g-1. A strong relationship between surface area and size was found for Taberg, while Karlsborg showed a weaker dependence. Malmström et al. [67] reported surface areas for two different fractions for their chlorite, butr since they used a different sorbing gas for the BET determinations, comparison with their results is not possible. The specific gas molecules have unique cross-sections, which affects BET surface area determination [93, 94]. Furthermore, nitrogen has a permanent quadropolar moment whereas argon and krypton

24 are apolar monoatomic gases [95]. The surface area determination is affected by the surface area and the surface roughness [96]. Chlorites from Taberg and Karlsborg were dissimilar concerning their dependency on surface area, Figure 3.1. It would be interesting to analyse the surface area in relation to fraction size for the various chlorites described in the literture [31, 61, 62, 67-69], in order to compare them with the behaviour observed in our study. After crushing our samples were treated in the same way. SEM images were obtained in order to look at the particles and determine wheter the methods gave rise to any differences which did not affect the surface area, Figure 3.2.

Taberg untreated particle mill used Karlsborg untreated particle mill used

Taberg untreated particle liquid nitrogen Karlsborg untreated particle liquid nitrogen used used Figure 3.2. SEM images of untreated Taberg and Karlsborg samples, frozen or mechanically crushed. Note different scales.

The images in Figure 3.2 show that the shape of particles was in agreement and that ultra-fine particles were formed in both methods used. However, these small particles were much more frequent for the Taberg samples.

25 AFM was used to characterise topographical structures, on a nanometer scale, of the chlorite (001) surface. The images clearly reveal that the basal surface of the chlorite used in our experiments was not atomically flat but contained molecular scale steps. Such nano-topographic heterogeneity provided additional sorption sites, Figure 3.3.

100 nm 700 nm

Figure 3.3. Contact mode AFM image of a chlorite (001) basal surface (scan area: 700 x 700 nm). Molecular scale steps with a height down to one unit cell (14 Å) are clearly visible

Even though the chlorite underwent different degrees of isomorphous substitution, resulting in a negative charge on the surface, the CEC observed at pH 7.5 for both Taberg and Karlsborg was low. This is in agreement with earlier observations [20, 97]. Taberg has a CEC of 4.0 meq/100 g (=cmol/kg) and Karlsborg a CEC of 1.4 meq/100 g. Both these results are in the same range as published previously for chlorites [20, 97] and both results indicate a very low CEC.

Chemical composition of the chlorites Taberg, Karlsborg and FlagStaff Hill were determined by ICP-AES at Analytica AB, Luleå, Sweden, Table 3.3. SEM-EDS analysis confirmed the content of the major oxides (SiO2, Al2O3, Fe2O3, MgO). Values of the ratio between Fe(II) and Fe(III) obtained from Mössbauer spectroscopy, presented as percentage Fe(III) of the total amount Fe(II + III) for the chlorites used, are listed in Table 3.5. For Karlsborg, it was found that the Fe(III) was tetrahedrally coordinated whereas the other samples had octahedral coordination of both Fe(II) and

26 Fe(III). Octahedrally coordinated Fe(II) and Fe(II) is in accordance with other characterisations of chlorites [62, 66, 67, 69]. The FlagStaff Hill chlorite had too low iron content for Mössbauer determination.

Table 3.3. Chemical composition of chlorite samples, determined by Analytica AB, Luleå, Sweden. % TS Taberg Karlsborg FlagStaff Hill SiO2 33.6 30.0 30.7 Al2O3 13.7 19.3 22.9 MgO 32.9 19.8 31.1 Fe2O3 5.95 16.1 1.42 K2O 0.68 0.84 0.063 CaO <0.099 0.26 <0.090 MnO2 0.13 0.32 0.014 Na2O <0.05 <0.05 <0.05 P2O5 0.0073 0.176 0.011 TiO2 0.011 0.59 0.25

From the results of the major oxides, the stoichiometric coefficients were determined using least squares fitting based on the general formula given in Equation 1.1 and assuming that all chlorites studied were trioctahedral in both the 2:1 layer and the brucite-like layer, even through this was only confirmed for Taberg. Karlsborg XRD peaks were broader due to the higher iron content, which caused overlaps in the XRD pattern. The chemical formulae for Taberg and Karlsborg were corrected and refined after Papers I and II were published.

Taberg : (Mg9.45FeII0.60FeIII0.25Al1.560.14)(Si6.48Al1.52)O20(OH)16

Karlsborg : (Mg5.92 FeII2.04 Al3.12 Ca0.05 K0.50.37)(Si6.12 Al1.49 FeIII0.39)O20(OH)16

FlagStaff Hill : (Mg8.70 FeII0.19 FeIII0.01 Al2.810.29) (Si5.76 Al2.24) O20 (OH)16

XRD observations indicated that the Karlsborg sample contains intrusions of another phase possibly illite and the Taberg sample contains tracers of vermiculite.

The chemical composition of KOV:01 was based on the major oxide content gained from SEM-EDS analysis, shown in Table 3.4 and applying Equation 1.1 resulting in:

27 KOV:01: (Mg6.34FeII1.12FeIII0.78Al2.471.29)(Si6.91Al1.09)O20(OH)16

Table 3.4. Major oxides of KOV:01 sample, achivied from SEM-EDS determination. % oxide SiO2 Al2O3 Fe2O3 MgO CaO K2OMnO2 Na2O KOV:01 34.4 15.4 12.9 21.6 0.55 0.12 0.51 n.d.

The vacancies ( ) are not always presented within the chemical formulas since they may be a result of random isomorphous substitution, which will leads to less crystallised structures. For a tri-octahedral chlorite, all the tetrahedral and octahedral positions should be occupied, but when using the percentage of oxides and filling the eight tetrahedral positions there is a divergence within the octahedral position, which may be due to LOI (loss of ignition) or, as mentioned, isomorphous substitution.

Due to the limited amount of synthetic chlorite its chemical composition was not determined. The FlagStaff Hill sample was not used for sorption or weathering experiments due to its very low iron content, since iron was among the elements of major interest.

Table 3.5 Fe(III)/(Fe(II)+Fe(III)) data for the different chlorite samples. Fraction of Chlorite Fe(III)/(Fe(II)+Fe(III)) % Taberg 29 Karlsborg 16 FlagStaff Hill n.d. KOV:01 41

Taberg and Karlsborg chlorites were chosen based on differences in their chemical composition.

28 3.2 Dissolution results

This chapter summarises and discusses the results obtained in Paper I and Paper VI.

The release of the elements Al, Fe, Mg and Si was monitored over time throughout the flow-through experiments. Figure 3.4 shows the release rates of the different elements for the Taberg and Karlsborg samples. A high initial release rate declining to a stable value is the general dissolution behaviour for silicates. This behaviour was observed for the Taberg sample at pH 2, where the first 50 hours showed a drastic decrease in cation release, Figure 3.4. For the Karlsborg sample at pH 2, this behaviour was monitored for Mg, but when the other three elements were studied, a continuous release at low level was detected. Both chlorites displayed a minimum released concentration at neutral pH, Figure 3.5. The behaviour of the Taberg sample at pH 2 was in agreement with other mixed-flow reactor investigations, and it acted in the same manner as chlorite, kaolinite and biotite [57, 58, 62]. Experimental steady-state conditions were reached after approximately 15 days, but sampling continued for several days after this in order to ensure that no temporal plateau had been reached. In recent studies of chlorite dissolution, Lowson et al. [69] reported that the time needed to reach experimental steady state conditions varied between 10 to 50 days and Brandt et al. [62] reported that in their system the steady state was reached after 48 hours. The buffer solutions used by Lowson et al. [69] seem not to affect the time elapsed to steady state. However, comparison in the initial dissolution phase however is difficult, since few samples were collected during that period [69]. The high release rate in the beginning of the experiments depended on 1) fine particles at the surface not been removed during the washing processes 2) fresh surface being used for each experiments and 3) dissolution of reactive sites at the chlorite surface introduced during the crushing procedure. SEM images confirmed that there were some really small particles, ~ 2-5 μm, left after the washing procedure, independent of crushing method.

29 -08 Taberg pH 2

-09 s) 2

Mg -10 Fe (moles/m

j Al -11 Si Log r Log

-12

0 2 4 6 8 10121416182022 time (days) -10 Taberg pH 4 s)

2 -11

Mg -12

(moles/m Fe j Al -13 Si Log r Log

-14

0 2 4 6 8 10121416182022 time (days) -10 Taberg pH 12 s) 2 -11

Mg Fe (moles/m -12 j Al Si Log r Log

-13

0 2 4 6 8 10121416182022 time (days)

30 -08 Karlsborg pH 2 s)

2 -09

Mg Fe (moles/m -10 Al j Si Log r Log -11

0 2 4 6 8 10121416182022 time (days) Karlsborg pH 4 -10 s) 2

Mg Fe (moles/m -11 j Al Si Log r

0246810121416182022 time (days) Karlsborg pH 12 -10 s) 2

-11 Mg (moles/m Fe j Al

Log r Log -12 Si

0 2 4 6 8 10121416182022 time (days) Figure 3.4 Logarithm release rates of the different elements over a time-scale up to 22 days for Taberg and Karlsborg in mol m-2 s-1. Note that there are different scales in the plots.

For the dissolution rates at steady state conditions, the last measured values were used after approximately 22 days, and are tabulated in Table 3.6. In Paper I, Figure 3, the steady-state release rates of Si as a function of pH are published. Lowson et al. [69] have misinterpreted our release rates, since they automatically assumed that the data were normalised, which was not the case.

Table 3.6 Dissolution rates, normalised with respect to the stoichiometry and the final values after 22 days. -2 -1 -2 -1 -2 -1 -2 -1 Sample RAl mol m s RFe mol m s RMg mol m s RSi mol m s Taberg pH 2 1.00 ·10-12 1.24 ·10-12 6.86 ·10-13 1.29 ·10-12 Taberg pH 4 3.94 ·10-13 2.77 ·10-14 2.24 ·10-12 2.66·10-13 Taberg pH 10 2.64 ·10-13 4.91 ·10-14 1.10 ·10-11 4.74 ·10-14 Taberg pH 12 6.20 ·10-13 2.02 ·10-13 - 1.79 ·10-13

Karlsborg pH 2 6.75 ·10-11 5.07 ·10-11 3.91 ·10-13 4.20 ·10-12 Karlsborg pH 4 1.79 ·10-12 3.90 ·10-12 5.36 ·10-12 2.33 ·10-12 Karlsborg pH 10 2.82 ·10-12 8.50·10-14 2.74 ·10-13 8.99 ·10-13 Karlsborg pH 12 5.97·10-11 7.83·10-13 1.66·10-12 9.57 ·10-13

Since Table 3.1 and Table 3.2 together with Figure 3.1 and Appendix E show that the BET surfaces were quite different between the chlorites normalisation to mol g-1s-1 was used instead of the mol m-2 s-1 which are listed in Table 3.6.

Table 3.7 Dissolution rates, normalised with respect to the stoichiometry and the final values after 22 days. 1 1 1 1 1 1 1 1 Sample RAl mol g s RFe mol g s RMg mol g s RSi mol g s Taberg pH 2 6.70 ·10-12 8.31 ·10-12 4.60 ·10-13 8.64 ·10-12 Taberg pH 4 2.64 ·10-11 1.86 ·10-13 1.50 ·10-11 1.78·10-11 Taberg pH 10 1.77 ·10-12 3.29 ·10-13 6.70 ·10-11 3.18 ·10-13 Taberg pH 12 4.15 ·10-12 1.35 ·10-12 - 1.20 ·10-12

Karlsborg pH 2 3.44 ·10-11 2.59 ·10-11 1.99 ·10-13 2.14 ·10-12 Karlsborg pH 4 9.10 ·10-13 1.99 ·10-12 2.73 ·10-12 1.19 ·10-12 Karlsborg pH 10 1.44 ·10-12 4.34 ·10-14 1.40 ·10-13 4.58 ·10-13 Karlsborg pH 12 3.05 ·10-11 3.99 ·10-13 8.48 ·10-13 4.88 ·10-13

32 pH 02468101214 -10.00

-10.50

-11.00

-11.50

-12.00

-12.50 Taberg Log Log R(Si) mol/gs -13.00 Karlsborg Brandt et al -13.50 Lowson et al Rochelle et al -14.00 Malmström et al Hamer et al Figure 3.5 Logarithm steady state chlorite dissolution rates (based on Si data) as a function of pH.

The dissolution rates of chlorite for both Taberg and Karlsborg showed that the assumed pH dependency and the behaviour of chlorite followed the schematic picture for silicates, even though the specific dissolution rates diverged. For example, for pH 2 the dissolution rates calculated from Si data differed by almost one order of magnitude. The same difference was observed for rates calculated for Al, while Fe and Mg differed by a factor ~ 2. At pH 2 the dissolution rate for Karlsborg was similar to that obtained by Brandt et al. [62] and at pH 7.4 the dissolution rate for Taberg was in agreement with that reported by Rochelle et al. [68] and Lowson et al. [69]. Larger differences in dissolution rate were observed in the acidic pH range in comparison to basic pH l, which could be caused by the difference in composition between the chlorites studied and the higher degree of dissolution at that pH. The dissolution rate for chlorite reported by May et al. [66] increased with increasing acidity, which was also observed for Taberg and Karlsborg chlorites. Rochelle et al. [68] analysed Si, Mg, Al and Fe at different pH, in the neutral to basic range, varying the pH and

33 temperature. They observed that the dissolution rate increases with increasing pH, which is in line with our observations.

When the empirical rate law, Equation 2.9, was applied to the Si data in Table 3.7, the constants and reaction orders were calculated using a weighted non-linear least-square regression with kH = 1.90·10-11 mol g-1 s-1, kn ≤ 1.0·10-13 mol g-1 s-1 and kOH = 1.04·10-12 mol g-1 s-1 and the reaction orders in this study were m = 0.32 and n = 0.14. The vaules obtained in athis study were than the m= 0.5 obtained by May et al. [66], m = 0.49 and n = 0.43 determined by Lowson et al. [69] and n = 0.30 obtained by Rochelle et al. [68], although our value was in agreement with m= 0.29 ±0.05 obtained by Brandt et al. [62]. The reaction order of the overall rate law shows the release of H4SiO4 into solution, where Si from the surface is released due to proton-promoted dissolution, which is apparent as the pH dependency of the dissolution [98]. A higher number of reaction order m indicates that the dissolution requires a higher degree of protons present.

-10 Taberg Karlsborg ]

-1 -11 s -1

[mol g -12 Si

Log R -13

-14 135791113 pH

Figure 3.6. Logarithm steady-state dissolution rates calculated from Si data. The curve shows the calculated rate according to the empirical rate law in Equation 2.9.

The accumulated dissolution for the Taberg and the Karlsborg chlorites based on the elements studied as a function of time for some pH values investigated is for the Taberg presented in Figure 3.6. As in Figure 3.1, the fast reaction of chlorite in the beginning of the experiments indicates by a rapid increase in the accumulated dissolution in Figure 3.6.

34 Since no studies of eventual formation of secondary phases were performed in these dissolution studies, the present data were not interpreted in terms of congruent or non- congruent dissolution. Instead, the terms stoichiometric and non-stoichiometric dissolution were used. A release ratio lower than start ratio indicates that Si is preferentially released, while if the ratio is higher than the start ratio Mg, Fe or Al is preferentially released compared with Si. Kalinowski and Schweda [99] discuss the stoichiometry as ratios as in Table 3.8. Holdren and Speyer [100] and Stillings and Brantley [101] discuss stoichiometry as relative release ratios, RRRx. When RRRx = 1 the observed dissolution is stochiometric and when the ratios is < 1 Si is preferentially released compared with the other cation, while if the ratio is > 1 the cation x is preferentially released compared with Si.

RRRx = (x/Si)solution/(x/Si)starting material (3.2) where x = Al, Mg or Fe. For Taberg the following observations were made: At pH 2 in the beginning of the experiments there was preferential release of Mg, Fe and Al, while at steady-state Si was preferentially released. Both initial and final releases were non-stoichiometric, but Si and Al were close to stoichiometry at the end of the experiments. For pH 4 Si was preferentially released compared with Fe and Al throughout the whole experiments, while when Mg and Si were compared a different behaviour was observed, for the first 15 days Si was clearly preferentially released, while at the end of the experiments the reverse occured. At pH 12, Al and Si were close to stochiometric release in the beginning of the experiment but at steady state Al was preferentially released. When comparing Fe and Si, in the beginning of the experiments Si was preferentially released and at steady state the release was close to stoichiometry. For Karlsborg the following observations were made: At pH 2 in the beginning of the experiments Al, Mg and Fe were preferentially released compared with Si and at steady state Fe and Al were still preferentialyl released compared witho Si, while for Mg and Si the reversed occured. When pH was increased to 4, Mg and Al were still preferentially released in the beginning of the experiment. When comparing Fe and Si, Si was the element that was preferentially released. At steady state, on the other hand, Si was preferentially released compared with Al, Fe and Mg. At the most basic pH studied, pH 12, the Fe/Si ratio was more or less constant throughout the whole experiment, which could be explained by low release of Fe and possible formation of secondary phases during the dissolution experiment,

35 which PHREEQC calculations indicated. This has also been observed by AFM investigations [73]. The same possible formation of secondary phases during the dissolution was observed for Mg. However, Mg and Al were preferentially released in the beginning of the experiments and for Al and Si the same releationsship was observed at steady-state, while the reverse occured for Mg and Si. To summarise, Taberg and Karlsborg cation release was non-stochiometric for the elements studied, which is in agreement with the observations by Brandt et al. [62] for pH 2-4. Hamer et al. [63] reported non-stoichiometric dissolution with a preferential release of Si relative to Al and Fe and in some cases even to Mg. Ross [65] also reported that Mg, Fe and Al dissolved at the same rate in Si-saturated 2N HCl at 60°C and suggested a dissolution process where the hydroxide sheet and the 2:1 sheet within the chlorite structure were equally dissolved. Our observations contradict the rates reported by Lowson et al. [69] Ross [61, 64] since they stated that congruent dissolution takes place.

Table 3.8. Release rates ratios compared with ratio in starting material for Taberg and Karlsborg samples. Mg/Si Fetot/Si Al/Si TABERG Starting material 1.46 0.13 0.48 pH 2 1 day 2.62 0.22 0.59 22 days 0.88 0.08 0.43 pH 4 1 day 0.001 0.004 0.13 22 days 7.05 0.01 0.25 pH 12 1 day 0 0.05 0.51 22 days 0 0.14 2.69

KARLSBORG Starting material 0.97 0.40 0.75 pH 2 1 day 22.05 0.94 1.83 22 days 0.07 6.31 15.4 pH 4 1 day 1.34 0.06 0.96 22 days 0.68 0.24 0.40 pH 12 1 day 40.17 0.14 4.16 22 days 0.60 0.14 2.69 36

We let the accumulated release of Mg represent the octahedrally coordinated cations and Si represent the tetrahedrally coordinated cations as suggested by Lin and Clemency [102]. Using the accumulated releases at steady state, Figure 3.8, we observed that Mg amount was twice as high as the Si amount, which indicates a preferential release of the octahedrally coordinated cations at pH 2. At pH 4 and 12 the opposite behaviour was observed for Taberg. For Karlsborg, for the three pH presented in Figure 3.8, Mg was preferentially released when steady state was achived. At pH 4 for Karlsborg, the accumulated amount of Mg was six times higher than that of Si while at pH 12 Mg was approximately ten times higher than Si, which indicates that the octahedrally coordinated cations were preferentially released for this sample. Brandt el al. [62] observed preferential release of octahedrally coordinated cations up to pH 4, while Hwang [103] observed that Al, Mg and Fe are preferentially released compared with Si under acidic conditions. The Karlsborg sample and the chlorite used by Brandt et al. [62] behaved in the same way in the pH range 2-4 while Taberg exhibited different behaviour at pH.

SEM images were collected for the weathered samples after the flow-through experiments, Figure 3.7. At the surface of the weathered particles from pH 2 and 12 a pattern of channels was observed, which is in agreement with the observations at acidic pH for other chlorites following dissolution experiments [61, 62]. These channels have been denoted dissolution channels in previous work [61].

Figure 3.7. SEM imgages of weathered Taberg chlorite particles at pH 2.

37 9.00E-05 9.00E-05 Taberg pH 2 Karlsborg pH 2 8.00E-05 8.00E-05 Al Al 7.00E-05 Fe 7.00E-05 Si Si 6.00E-05 6.00E-05 Fe Mg Mg 5.00E-05 5.00E-05

4.00E-05 4.00E-05 Moles chlorite/g chlorite/g Moles Moles Moles chlorite/g chlorite/g Moles Moles 3.00E-05 3.00E-05

2.00E-05 2.00E-05

1.00E-05 1.00E-05

0 0 0246810121416182022 0 2 4 6 8 10 12 14 16 18 20 22 Time (days) Time (days) 1.00E-05 1.60E-04 Taberg pH 4 Karlsborg pH 4 9.00E-06 1.40E-04 Al 8.00E-06 Si Fe 1.20E-04 Fe 7.00E-06 Si Mg 6.00E-06 Mg 1.00E-04 Al

5.00E-06 8.00E-05 4.00E-06 Moles chlorite/g chlorite/g Moles Moles Moles chlorite/g chlorite/g Moles Moles 6.00E-05 3.00E-06 4.00E-05 2.00E-06

1.00E-06 2.00E-05

0 0.00E+00 0 2 4 6 8 10121416182022 0 2 4 6 8 10121416182022 Time (days) Time (days)

38 1.80E-05 1.80E-05 Taberg pH 12 Karlsborg pH 12

1.60E-05 Al 1.60E-05 Al 1.40E-05 Fe 1.40E-05 Si Si Fe 1.20E-05 1.20E-05 Mg Mg 1.00E-05 1.00E-05

8.00E-06 8.00E-06 Moles chlorite/g chlorite/g Moles Moles MolesMoles chlorite/g chlorite/g 6.00E-06 6.00E-06

4.00E-06 4.00E-06

2.00E-06 2.00E-06

0 0 0 2 4 6 8 10121416182022 0246810121416182022 Time (days) Time (days) Figure 3.8. Accumulated dissolution of chlorites, in moles chlorite g-1. Note: Difference in the vertical scale at pH 4 for Taberg and Karlsborg.

40 3.3 Sorption results

This chapter presents and discusses the sorption of Ni(II) on chlorites. The experiments were performed under a number of varying experimental conditions, see further in Papers II, IV and V for details and Appendix C for tabulated sorption results.

In Figure 3.9 the sorption is presented as Kd as a function of pH for three different background electrolyte concentrations and one of the initial Ni(II) concentrations, 10-6 M for

Karlsborg chlorite at a concentration of 5 g/L. The Kd values increased by three orders of magnitude in the pH range 4-9.5, with a Kd maximum of 1700 cm3/g at pH 8.4. Figure 3.10 shows the three different initial nickel concentrations together with the three different background electrolyte concentrations as a function of pH for Karlsborg chlorite with a solid to solution (s:s) ratio of 5g/L. The typical sorption edge appearance of the sorption curve was observed where the sorption increased over a limited pH range which is typical for pH- dependent sorption.

4.0 3.5 3.0 /g) 3 2.5

(cm 2.0 d

K 1.5 − 6 [Ni]initial = 10 M log 1.0 0.01 M [NaClO4] 0.5 0.1 M [NaClO4] 0.5 M [NaClO4] 0.0 4 5 6 7 8 9 10 11 pH Figure 3.9. Sorption as a function of pH and ionic strength for Ni concentration 10-6 M, Karlsborg chlorite.

4.0

3.5

3.0

2.5 /g) 3 [NaClO4] [Ni(II)] cm ( 0.01 M, 10-6 M

d 2.0 0.1 M, 10-6 M -6 log K 0.5 M, 10 M 1.5 0.01 M,10-8 M 0.1 M,10-8 M 1.0 0.5 M, 10-8 M 0.1 M,10-7 M 0.5 0.01 M 10-7 M 0.5 M, 10-7 M 0.0 45678910 pH Figure 3.10. Sorption as a function of pH and ionic strength for the three different Ni(II) concentrations onto Karlsborg chlorite.

Figure 3.11 shows the sorption data for Taberg and Karlsborg as percentage sorbed Ni(II) as a function of pH. One initial Ni(II)l concentrations and three different ionic strengths were analysed, using an s:s ratio of 5 g/L for both chlorites.. Both chlorites behaved in a similar way, showing a strong pH dependency and no dependence of background electrolyte concentration, which was the expected behaviour for Ni(II) sorption onto chlorite. Bradbury and Baeyens [78, 79] studied sorption as a function of pH and found that for Ni there were two main mechanisms that controlled sorption, one pH independent and one pH dependent. The pH independent mechanism was identified as ion exchange and the pH dependent mechanism as surface complexation. However, those authors studied the sorption onto montmorillonite which has a high CEC value. They observed that for the region where the sorption is more or less constant, the plateau-like area appears at different pH values for the different background electrolyte concentrations, but no such observation was found in the Ni(II) chlorite system.

100 90 80 70 ) [Ni(II)] [NaClO4] 60 Taberg 10-6 M, 0.01 M 50 Karlsborg 10-6 M, 0.01 M 40 Taberg 10-6 M, 0.1 M 30 % sorbed% Ni(II Karlsborg 10-6 M, 0.1 M 20 Taberg 10-6 M, 0.5 M 10 Karlsborg 10-6 M, 0.5 M 0 4 5 6 7 8 9 10 11 12 pH Figure 3.11. Percentage of Ni(II) sorbed onto Karlsborg and Taberg chlorites.

Lack of dependence on ionic strength confirms that Ni(II) sorbs onto chlorite by surface complexation and that the sorption occurs through inner-sphere complex rather than outer- sphere [47], still Koppelman et al. [77] in their work found that Ni(II) adsorbed to chlorite as a hexa-aquo ion, Ni(H2O)62+, due to the binding energies observed in the XPS spectrum. The observed binding energies are not similar to the binding energies of Ni(OH)2 or NiO. According to Lützenkirchen [47], Ni(II) could not been sorbed as an aquo ion since that would be in disagreement with the inner-sphere complexation theory. The mass of chlorite remained constant, i.e. the number of available sites did not change throughout the experiments, even though the surface area of the Taberg and Karlsborg chlorites differd, see Table 3.1. In Figures 3.9 – 3.15, the fraction used of the chlorite samples was 120 – 240 µm, for the KOV sample 63-118 μm. In order to compare data from Taberg and Karlsborg the small fraction was used for some sorption experiments. Our study showed no difference in Kd between the fractions used for Taberg and Karlsborg in spite of the observed difference in surface area was, Table 3.1. After normalisation of the distribution to surface area instead of to dry mass, Ka was used instead of Kd. Since dissolution rate was found to be dependent on some other factor than surface area, presenting sorption data as Kd values suffices for this type of batch experiment and the effect seen in Ka does not reflect sorption.

Table 3.9. Log Kd and Log Ka for Taberg and Karlsborg samples for the two different size fractions used in the sorption experiments. Chlorite Fraction pH Log Kd Log Kd Log Ka (m) (cm3 g-1) (m3 kg-1) Taberg 63-118 μm 8.0 3.13 ± 0.05 0.13 ± 0.05 -3.77 ± 0.05 Taberg 120-240 μm 8.0 3.11 ± 0.05 0.11 ± 0.05 -3.72 ± 0.05

Karlsborg 63-118 μm 8.0 3.14 ± 0.03 0.14 ± 0.03 -2.63 ± 0.05 Karlsborg 120-240 μm 8.0 3.16 ± 0.04 0.16 ± 0.04 -2.54 ± 0.05

4.0 /g) 3 3.0

(cm [Ni(II)] d 0.1 g chlorite 10-6 M 2.0 0.1 g chlorite 10-7 M log K 0.1 g chlorite 10-8 M 1.0 0.05 g chlorite 10-6 M 0.05 g chlorite 10-7 M 0.05 g chlorite 10-8 M 0.0 4 567891011 pH Figure 3.12. Sorption as a function of pH and solid to solution ratio for Karlsborg sample.

Two different solid to solution ratios (s:s), i.e. 5g/L and 10 g/L, were analysed, Figure 3.12. When the dry weight of chlorite is increrased surface area and thet surface sites present also increase. El Aamarani et al. [104] found that for a given pH the distribution factor increases with increased s:s and that the sorption maximium occurs at the same pH for different s:s- ratios. Their s:s ratio varied between 2.5 and 100 g/L and there was a factor of 40 between the lowest and highest ratio. Bors et al. [105] studied the variation in distribution ratio for different s:s-ratios at tracer concentrations and found that the distribution ratio increased

44 with the s:s-ratio over an s:s-ratio range of 1-200. In our study no difference between the two s:s-ratios was observed taking the distribution factors into account, which could be explained by the fact that the ratios were only separated by a factor of 2 and the sorption maximum were located at the same pH for both ratios studied.

4.00 3.50 3.00 /g) 3 2.50

(cm 2.00 [Ni(II)] [NaClO ] d 4 Taberg 10-6 M, 0.01 M 1.50 Karlsborg 10-6 M, 0.01 M Taberg 10-6 M, 0.1 M log K log Karlsborg 10-6 M, 0.1 M 1.00 Taberg 10-6 M, 0.5 M Karlsborg 10-6 M, 0.5 M 0.50 0.00 4 5 6 7 8 9 10 11 12

pH Figure 3.13. Sorption as a function of pH for Taberg and Karlsborg samples.

The decrease in sorption at high pH may be an effect of alteration of the surface or complexation with dissolved chlorite species in solution, since this pH is in the area where dissolution takes place, Figure 3.5.

The reaction times of 7 and 14 days were used in order to ensure that the sorption processes were in equilibrium. No differences in amount of sorbed Ni(II) were observed for the two different reaction times. Eylym et al. [75, 76] studied the reaction of Ba2+ onto chlorite and they showed that a equilibrium is reached within 6 days. Scheidegger and Sparks [106] observed that after 24 hours approximately 90% of the initial Ni(II) concentration was sorbed and after 72 hours Ni(II) sorption was nearly completed where after a plateau was reached. Shahwan and Erten [70] observed that one day of contact time was enough to reach equilibrium Co2+ sorption.

45 In our sorption experiments small amounts of desorbed Ni were observed. Applying Equation 2.3 it is indicates that Ni was preferably irreversibly sorbed to the chlorite surface. Desorption was studied in a couple of ways. When just adding fresh background electrolyte with corresponding pH above 7, no significant desorption of Ni could be detected. The desorption studies with 1mM Ca(ClO4)2 also showed that only small amounts of the Ni(II) had been desorbed in favour of Ca(II). When the pH was decreased to approximately 4.5 7±2% of the Ni(II) was desorbed. Scheidegger and Sparks [106] observed that at pH 4 approximately 8% of the sorbed Ni(II) was desorbed and at pH 6 approximately 3%, hence our desorption studies are in good agreement with their observations.

All calculations based on the Ni(II) activity within the aqueous phase were corrected for the amount of Ni(II) sorption onto the tube walls, Figure 3.14.

6.0 [Ni(II)]= Black: 10 −6 M 4.0 Grey: 10 −8 M 2.0 0.0 −2.0 −4.0

wall (%) adsorption −6.0 −8.0 246810

pH Figure 3.14. Percentage of Ni(II) sorbed to walls of reaction vessel as a function of pH. The legend describes the various nickel concentrations used.

KOV:01 sorption onto chlorite was performed with two different initial Ni(II) concentrations and one s:s ratio, 5 g/L. The mineral powder scraped off the core surfaces was in the size range of 63 -118 µm which is smaller than used for sorption experiments for Taberg and Karlsborg. To evaluate whether the size of the particles affects the degree of sorption, the smaller fraction was used in experiments for the Taberg and Karlsborg chlorites 46 using an initial Ni(II) concentration of 10-6 M. For all these experiments only one concentration of background electrolyte was used, 0.1 M NaClO4. Since sorption of Ni(II) onto the chlorite surface is independent of the ionic strength, only one was used. Figure 3.15 shows the distribution coefficient for the KOV:01 sorption together with the results from Taberg and Karlsborg. It is hard to compare our results with Co(II) sorption on a chlorite from Shahwan and Erten [70] since the Co(II) concentrations was orders of magnitude higher than our Ni(II) concentrations, and was also performed in a very narrow pH range. Kd values obtained on montmorillonite [79] are approximately one log-unit higher than ours obtained for chlorite, since montmorillonite is a sheet silicate with a high cation exchange capacity.

4.0 3.5 3.0 /g) 3 2.5 Sample [Ni(II)] Fraction -6 (cm 2.0 KOV 10 M, 63-118 μm d KOV 10-8 M, 63-118 μm 1.5 Karlsborg 10-6 M, 63-118 μm log K log 1.0 Karlsborg 10-6 M, 120-200 μm Taberg 10-6 M, 63-118 μm 0.5 Taberg 10-6 M, 120-200 μm 0.0 567891011

pH Figure 3.15. Sorption as a function of pH for the three chlorite samples, Taberg, Karlsborg and KOV:01.

The pH values given in the different diagrams as well as those tabulated in Appendix C are all final pH values measured after completed reaction time, whereas a pH drift (± 0.3 pH-units) was observed during the reaction time, with largest drift for the lowest and highest pH values. This can be explained by chlorite dissolution. To study the amount of dissolved chlorite during the sorption processes and the competition between Ni(II) and Si in the chlorite structure, the Si concentration was determined after the reaction time of 7 days for the Karlsborg chlorite using 10-6 M Ni(II), Figure 3.16. When comparing the Si release rate from sorption experiments with the Si release rate from the flow-through experiments, after

47 ~ 7 days of reaction time, it was found that in the near neutral pH range, Si could be detected in the sorption experiments but not in the flow-through experiments. This is due to the difference in methods since in the flow-through method, fresh solution is pumped through the system, whereas in the sorption method the same solution is in contact with the solid phase for the whole reaction time, Figure 3.16 – 3.17.

1.00E-04

8.00E-05

6.00E-05 0.01 M NaClO4

0.1 M NaClO4

[Si] (M) 4.00E-05 0.5 M NaClO4 2.00E-05

0 678910 pH Figure 3.16. Measured Si concentration after sorption experiments with 10-6 M Ni(II).

-9.60 0.01 M NaClO4 -9.80 0.1 M NaClO4 s)

2 -10.00 0.5 M NaClO4 -10.20 Dissolution experiments -10.40 (mol/m

Si -10.60 -10.80 log R -11.00 -11.20 4 5 6 7 8 9 10 11 12 pH Figure 3.17. Release rate of Si after ~7 days of reaction time, comparing batch and flow- through experiments.

The aim of the sorption isotherm experiment was to quantify the surface site density, since a pH drift was observed throughout the sorption experiments performed only with the background electrolyte. Even a small drift gives complex isotherms expressions, which is why buffers were introduced [107]. The buffers were tested and did not interfere with the sorption processes.

48 The isotherm data from the three pH values investigated, where the logarithm of sorbed Ni(II) concentration is presented as a function of the logarithm of the aqueous Ni(II) concentration at equilibrium, are shown in Figure 3.18.

-3

-4 pH 8.3 Karlsborg pH 7.4 Karlsborg -5 pH 6.0 Karlsborg pH 8.3 Taberg

] (M)] pH 7.4 Taberg -6 ph 6.0 Taberg sorbed

-7 log [Ni log

-8

-9

-10 -10-9-8-7-6-5-4 log [Niaq] (M) Figure 3.18. Soprtion isotherms for Taberg and Karlsborg for pH 6.0, 7.4 and 8.3.

The constants obtained when applying linear Freundlich isotherm to the experimental data, are listed in Table 3.8. The surface sites for surface complexation are not evenly distributed. XAFS studies by Dähn et al. [81] showed that Ni sorption takes place preferentially on the edges and therefore the Freundlich equation is preferable to the Langmuir equation since the degree of surface precipitation could not be defined in our sorption results and since the Freundlich isotherm does not assume that the sorption takes place with in a monolayer unlike Langmuir. The lower n value at the highest pH is a result of curve deviation at the highest Ni(II) concentration used, 10-4 M, probably due to surface co-precipitation in parallel with sorption reactions, since as Freundlich accumulates the sorbed Ni(Ii) no sorption maximum is reached. Bradbury and Baeyens [78, 79] observed at least two different regions in their Ni isotherm, which used the cross-point for the two main regions as the point where all strong surface sites are occupied at pH 8.2, together with a uncertainty of ± 50%. For Karlsborg and Taberg no such regions were found in the Freundlich isotherm plot, one reason being the large gap

49 between the concentrations studied. Further experiments should be performed in this narrow [Ni] range.

Table 3.10. Freundlich constants for Karlsborg chlorite. Chlorite pH KF n R2 Karlsborg 8.3 5.37 0.78 0.95 7.4 6.75 0.90 0.95 6.0 0.33 0.78 0.98

-5.0

-4.5

-4.0

-3.5 Karlsborg 8.3 -3.0 Karslborg 7.4 Karslborg 6.0 -2.5

log Q Linear (Karlsborg 8.3) -2.0 Linear (Karlsborg 7.4) -1.5 Linear (Karlsborg 6.0)

-1.0

-0.5

0 0 -1 -2 -3 -4 -5 -6 -7 log Ce Figure 3.19. Linear Freundlich isotherm of Karlsborg sorption data.

Zorn [72] used Langmuir sorption isotherms of U(VI) to determine the surface site density of 1.45 sites/nm2 for chlorite. When a linearised Langmuir isotherm equation was applied to experimental data at pH 8.3 for Karlsborg i.e. at sorption maximum and assuming monodentate binding of Ni(II) to the chlorite surface, the site density was found to be 0.67 sites/nm2 ([Chl_OH] = 5.56⋅10-6). This is 40 times lower than the site density obtained from the titrations. When applying the Langmuir isotherm, we assumed that surface complexation occurred within only a monolayer and that monodentate binding occured. The sorption experiments performed indicate that sorption of Ni(II) to the chlorite surface takes place through surface complexation, although the introduction of the Langmuir isotherm equation to our data introduced an element of uncertainty, Figure 3.20.

50 1.40E+07

1.20E+07 y = 0,0165x + 30793 R2 = 0,9982

1.00E+07

8.00E+06

6.00E+06 1/(g sorbed Ni/g chlorite) Ni/g sorbed 1/(g 4.00E+06

2.00E+06

0 0 1.00E+08 2.00E+08 3.00E+08 4.00E+08 5.00E+08 6.00E+08 7.00E+08 8.00E+08

1/[Ni]aq final Figure 3.20. Linear Langmuir isotherm of Karlsborg sorption data.

The degree of sorption was in agreement between Taberg and Karlsborg samples, even though the surface areas as determined by BET analysis differed by almost one order of magnitude. This phenomenon can partly be ascribed to the experimental method used, since batch technique is a macroscopic investigation method, and to the fact that chlorite surface contains few surface sites. Furthermore, SEM investigations of the two samples showed that the surface of the Taberg sample was much rougher and more irregular and included small outcrops on the basal plane, and this contributed to a larger basal area. The reactive surface area will not increase to the same magnitude and since chlorite does not contain many surface sites, a small increase does not affect the total number of reactive surface sites located predominantly on the edges [80]. Brandt et al. [62] used AFM for pH 2 dissolution and observed that the geometric reactive surface area is just 0.2 % of the BET area, which was 1.6 m2 g-1. However their measurements are performed on the basal plane, as in Figure 3.3.

51 52 3.4 Titration results

This chapter summarises and discusses the results obtained in Paper IV together with unpublished data for the Taberg sample.

The acid-base relationship of the chlorite surface was achieved from titrations. Suspensions and batch-back titrations of the supernatant were performed, but the results from the batch back-titrations technique were hard to evaluate since the chlorites dissolved under the contact time prior to titrations. Therefore the method of determining the constants presented by Bradbury and Baeyens [79] [108] could not be applied.

For the continuous titrations the chosen inverval between additions was 15 minutes which was the longer waiting time for these experiments but is short in terms of titrations. A minute scale between two titrant additions is found for fast speed titrations, but a waiting time of ~ 1 hour is not unusual [41]. For the chlorite titrations performed the stability criteria were potentially high according to Sjöberg and Lövgren, [109] (<0.05mV/90 min or <0.12 mV/60 min). For the chlorite these stability criteria were never reached, due to the ability to dissolve and the OH/O content in the structure. In all titrations of Taberg and Karlsborg chlorites, the contact time prior to experiments, the titration speed and the time between additions all affected the outcome of the titrations, i.e.. the shape of the titration curve. Above pH 10 and below pH 4 a small difference in the potential gives rise to a large difference in specific surface charge, σ, since it corresponds to a large difference in the H+/OH- concentration. This area corresponds to the area where the dissolution of the chlorites dominates, Figure 3.8.

Figure 3.21 shows the contact time prior to titration for two different contact times, using titration data with 3 minutes between additions performed as suspension titrations using Karlsborg chlorite.

53 ~2h of contact prior to titration with base

~2 h of contact prior to titration with acid

4.0 ~1 week of contact priot to titration with base 3.0 ~1 week of contact time prior to titration with acid Simulation 2.0

1.0

σ 0.0 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5 9.0 -1.0

-2.0

-3.0

-4.0 pH Figure 3.21. Summary of continuous titrations of Karlsborg. Effect of different contact time, with three minutes between additions.

Titration of the supernatant is presented in Figure 3.22 for Taberg chlorite with 15 minutes between additions and a contact time of one week. The shape and magnitude of this curve are in good agreement with the shape and magnitude of a curve presented by Zorn [72]. This consistency in curve shape indicates that it was only dissolved cations that have were titrated rather than titrated surface sites themselves in the work by Zorn [72]. Zorn [72] never reached a point where the surface was uncharged for the Grimsel chlorite, while for Taberg an interval was found for the isoelectric point in the pH range 6.5 to 8. The observed differences are probably due to different experimental conditions such as waiting time and contact time. Sondi et al. [110] observed an isoelectric point at pH 5.0 ± 0.2 for a chlorite with a BET of 4.8 m2/g and a CEC of 12 meq/100 g.

54 0,0007

0,0006

0,0005

0,0004

0,0003

0,0002 consumption solution by speciesM + H 0,0001

0 3456789 pH Figure 3.22. Titration curve of supernatant from batch-back titration of Tabeg chlorite.

3.5 Fitting the sorption and titration data

This chapter summarises and discusses the results obtained in Paper II and IV.

For the fitting of Ni(II) sorption data, two different models with some variety in the input parameters were used. The first model, Paper II, was simulated using DLM and some of the input parameters were literature data. The second model, Paper IV, was a NEM model using site density and acid base constants obtained from fitting of the titrations results. The Karlsborg chlorite was used in both models and the simulation of titration data was performed with results obtained from Karlsborg chlorite.

The following assumptions were made in the DLM modelling: Site density was set to 2.3 sites/nm2 [48] with a small proportion of strong sites of 0.0025 sites/nm2 [49] and the acid-base properties were assumed to follow the acid-base properties of the T:O:T in illite reported by Du et al. [111], listed in Table 3.11. During the fitting procedure, the whole surface was illustrated as one component, Chl_OH, since information on the different surface groups is unknown. 55

Table 3.11. Acidity constant of the TOT in illite (rom Du et al. [111]). Reaction pKa

Chl_OH Chl_O- +H+ -4.2

FITEQL, using inverse modelling applying a 2pK-model together with NEM to obtain the acidity constants for the chlorite surface as presented in Table 3.12, was used to fit the titration data.

Table 3.12. Results from inverse modelling of titration results using NEM in FITEQL. Reaction pKa

Chl_OH + H+ Chl_OH2+ 5.6

Chl_OH Chl_O- +H+ -8.2

The surface site density was determined, using titration data, to be 26.9 sites/nm2 ([Chl_OH] = 2.23⋅10-4), which is high in comparison with values reported in the literature 2.3 sites/nm2 [48] and 1.45 sites/nm2 [72].

Table 3.13. Surface complexation constants obtained from simulation of experimental data in PHREEQC (P) and FITEQL (F) using DLM.

Complex formed pKstrong (P) pKweak (P) Log Kweak (F) Chl_OH + Ni2+ Chl_ONi+ +H+ -0.5 -2.7 -3.5 2+ - + Chl_OH + 2H2O + Ni Chl_ONi(OH)2 +3H -14 -17 -16.3

Table 3.14. Surface complexation constants from the simulation of experimental data using a 2-pK model in FITEQL. Complex formed pK Chl_OH + Ni2+ Chl_OHNi2+ 3.63

Chl_OH + H20 + Ni2+ Chl_OHNi(OH)+ + H+ -2.65

Chl_OH + 2H20 + Ni2+ Chl_OHNi(OH)2 + 2H+ -11.94

56 5.00 simulation NEM 4.50 simulation DLM experimental 4.00 /g)

3 3.50 cm ( 3.00 d 2.50 log K 2.00

1.50

1.00

0.50

0.00 45678910 pH Figure 3.23. Experimental sorption results illustrated together with simulated results from NEM and DLM models.

Figure 3.23 shows simulations from DLM and NEM using together with experimental data using 0.1 M NaClO4 s:s ratio of 5g/L for Karlsborg chlorite.

The agreement between experimental data and models was good for the sorption edge part, but the models deviated when reaching the interval of more constant sorption. Even though,

Ni(OH)2 formation was introduced in the NEM as a parameter when establishing the model,

Ni(OH)2 precipitation could not be predicted in our PHREEQC simulation. Introducing weak and strong sites in the DLM model, improved the fit to the sorption data. This is in line with surface complexation models. Bradbury and Baeyens [78] also used a modified two-site surface complexation model for their pH dependent part of the Ni(II) sorption onto Na-montmorillonite, but they used the same acidity constants.

Bond et al. [71] applied the TLM sorption model to interpret the experimental results and the model and experimental data showed a generally good fit up to pH 10, where the model predicted that the sorption was predominantly onto the silica sites. This observation of the fit was confirmed by our observations for the DLM model, but the NEM model deviated at a lower pH, where the model overestimated the sorption.

58 4. Conclusions

Karlsborg and Taberg chlorite samples have been found to be representative for chlorites within the fracture zones in the granitic bedrock from Oskarshamn area, close to one of the candidate areas for the spent nuclear waste repository. The experimental results have proven that dissolution and sorption are capable of describing the behaviour of chlorites, as predicted from theory. One observed difference between Taberg and Karlsborg is the BET surface area. The difference can be ascribed to that the Taberg chlorite demonstrates a rougher surface area and that its surface contains thin flakes deviating from the basal plane. This roughness and protruding small flakes contributes to the total surface area but not to the reactive surface, which is where sorption takes place. That Ni(II) sorption occurs at the reactive surface sites is confirmed by the comparable degree of sorption between Taberg and Karlsborg. To further acquire detailed information for the surface complexation model, the surface site density was determined. This determination is strongly affected by the method used. From the titrations a value of 26.9 sites/nm2 was calculated and corresponding value from the sorption isotherm was 0.67 sites/nm2. The high presence of hydroxyl groups in the chlorite structure made it hard to study the acid-base properties of the chlorite surface using traditional titration methods. If Ni(II) and Co(II) leak from the bentonite barrier in the KBS-3 repository into the surrounding bedrock, chlorite present in the fracture zones will retard the migration of the cations through sorption, and will stay sorbed until condition drastically changes. One investigated condition where pH drastically decreased from 8 to 4.5, lead to a desorption of Ni(II) from the chlorite surface, however, ~80 – 85 % of the sorbed Ni(II) will not change its conditions at pH > 7. Site investigations and calculations have shown that the pH in areas around the future repositories will be in the interval 7-9 [112], if we take the variation of pH due to salinity intrusions, ice melting intrusions and dissolution of different mineral phases, into consideration. This pH interval corresponds to the stability window of chlorite dissolution and maximum sorption capacity of the chlorite surface. Thus, the experimentally determined Kd values for the chlorite Ni(II) system and the mineralogical composition of chlorites can be applicable into a larger modeling project, such as performance assessment calculations or predictive component additively approach modeling.

59 60 5. Future work

In order to create a more generalised surface complexation model for sorption of cations such as Ni(II), and not chlorite specific as in this work, some detailed experminets on the microscopic scale are needed. These experiments should include determination of the degree of surface precipitation during the sorption processes and identify where surface precipitation dominates instead of surface complexation. These investigations can be performed at the surface of chlorite using the AFM technique. Using the XAFS technique it would also be possible to observe how the binding of the cation into the surface takes place, whether it is monodente or bidentate and to which surface groups the binding occurs.

For the stability of the environment around a nuclear waste repository it is important that reducing conditions are present and some time after closure this environment will be achieved. If any leakages of fission products takes place it will take time before these products reach the fracture zones and the chlorite surface. In order to determine the behaviour of chlorite during those conditions which are different from the experimental conditions in this work, dissolution studies in an oxygen-free laboratory environment could be performed. These experiments should investigate the redox state of the Fe released from the chlorite, which in combination with investigations of secondary phases eventually formed would extend our knowledge of the dissolution behaviour of chlorite.

61 62 6. Acknowledgements

Throughout my years as a PhD-student, I have had the opportunity to collaborate with a number of skilful researches and I believe that this thesis never been the same without contributions from them, and they all are gratefully acknowledged.

First I would like to thank my three supervisors, Ignasi Puigdomenech, for accepting me as a PhD-student and I am glad that some of your many ideas could be fulfilled, Zoltán Szabó for continued as supervisor when Ignasi left for the big world outside KTH, for your patience and letting me working with my chlorites, Susanna Wold, for support and valuable comments on my different manuscripts, including this thesis, during my last years as a PhD-student.

Thank you: Anna-Maria Jakobsson, for our collaboration during the years especially the time-consuming titrations and for our never ending discussion about modelling. I will always remember Migration 2003 in South Korea☺

Maria Malmström for valuable discussions about weathering and dissolution, for squeezing meetings and discussions with me into your already tough agenda and for invaluable help presenting my experimental data in the best way.

Mireia Molera for introducing me to the world of radionuclides and for our discussion in the laboratory.

Lars Kloo and Andreas Fischer for your support and helpfulness during my years at KTH.

Mattias Edén and Hans Annersten for our collaboration, I have learned a lot from you!

Dan Holtstam and Swedish Museum of Natural History, Stockholm, for the chlorite samples.

Marianne Ahlbom, Björn Sandström and Alan Snedden for SEM analysis.

63 Ella Ekeroth, Oliva Roth, Anders Puranen and Kjell Svärdström for your assistance in the laboraotry at Nuclear Chemsitry.

Arvid Ödegaard-Jensen for ICP measurements when our instrument retired!

Siv Olsson for assistance with CEC determination and Clay Technology, Lund for their generosity, letting me used their lab for a couple of days.

Dirk Bosback, Volker Metz and Thuro Arnold for fruitful discussions about chlorite, weathering and sorption isotherms. For always have some time for answering my questions.

Eva-Lena Tullborg, for answering my mineralogy questions and help with the KOV:01 sample.

Isabelle Dubois for “attacking” my sorption data with PHREEQC!

Josefin Åkerstedt for XRD assistance.

Maggan, Eva-Stina and Helene for our struggling with teaching, nice lunches and dinners. KTH are not the same without you!

Former and present, colleagues at Inorganic Chemistry and Nuclear Chemsitry, especially you who have been working in room 756.

SKB for your financial support.

Per Persson and Krisnina Axe who encouraged me to apply for a PhD-position during my time as a diploma worker.

Kristina Andersson who introduced me to the world of Chemistry at Borgmästarhagsskolan, Nyköping.

64 My friends for your support and encourage, ingen nämnd ingen glömd ☺

Special thanks to my family and especially pappa, mamma and Anna, I have never reached this far without your love and support.

Uncle Tord, for being there as extra support during my stay in Umeå, I really appreciated your support and encourage.

Farmor, who taught me a number of important things in life, for example how to play cards and how to make a nice Sockerkaka. Nu farmor är jag äntligen klar med skolan!

Last, but not least, thanks to my beloved husband Henric for your patience, never ending support and love which encourage me during the years. To beloved Johanna, now mummy will have more time to read colourful books about Winnie the Pooh and friends instead of colour free articles and Chemistry books ♥ and thank you little one for endured my stress during the time that I finished this thesis.

65 66 7. References

1. SKB. Svensk Kärnbränslehantering AB, www.skb.se. [homepage] 2008. 2. SKBF/KBS, S.N.F.S.C.D.K., KBS 3 - Final storage of spent nuclear fuel - KBS-3 I-V. 1983, SKBF/KBS: Stockholm. 3. SKB, Deep repository for spent nuclear fuel. SR 97 - Post-closure safety. 1999, Swedish Nucl. Fuel Waste Manag. Co.: Stockholm, Sweden. 4. Karnland, O. and M. Birgersson, Montmorillonite stability. With special respect to KBS-3 conditions. 2006, SKB: Stockholm. 5. Karnland, O., S. Olsson, and U. Nilsson, Mineralogy and sealing properties of various bentonites and smectite-rich clay material. 2006, SKB: Stockholm. 6. Byegård, J., et al., The interaction of sorbing and non-sorbing tracers with different Äspö rock types - Sorption and diffusion experiments in the laboratory scale. 1998, SKB: Stockholm. 7. Hedin, A., 2002, Integrated analytic radionuclide transport model for a spent nuclear fuel repository in saturated fractured rock. Nucl. Technol. 138: p. 179-205. 8. Wahlgren, C.-H., et al., Geological description of rock domains and deformation zones in the Simpevarp and Laxemar subareas Preliminary site description Laxemar subarea - version 1.2. 2006, SKB: Stockholm. 9. Gaál, G. and R. Gorbatschev, 1987, An outline of the Precambrian Evolution of the Baltic Shield. Precambrian Research. 35: p. 15-52. 10. Stigh, J. and S.-Å. Larsson, Urberget - en bra miljö för ett säkert slutförvar av använt kärnbränsle. 2001. 11. Loberg, B., Geologi material, processer och Sveriges berggrund. sixth ed. 1999, Stockholm: Prisma. 515. 12. Kornfält, K.-A. and H. Wikman, The rocks of the Äspö island. Description of the detailed maps of solid rocks including maps of 3 uncovered trenches. 1988, SKB, Äspö Hard Rock Laboratory: Stockholm. 13. SKB, Äspö Hard Rock Laboratory - 10 years of research, M. Hammarström and O. Olsson, Editors. 1996, SKB: Stockholm. p. 88. 14. Rönnback, P., 2005, Äspö Hard Rock Laboratory: an overview of the geology, hydrology and hydrochemistry. the ESS Bulletin. 3: p. 32-47. 15. Drake, H., B. Sandström, and E.-L. Tullborg, Mineralogy and geochemistry of rocks and fracture fillings from Forsmark and Oskarshamn: Compilation of data for SR-Can. 2006, SKB: Stockholm. 16. SKB, Long-term safety for KBS-3 repositories at Forsmark and Laxemar - a first evaluation Main Report of the SR-Can project. 2006, Svensk Kärnbränslehantering AB: Stockholm. 17. Tullborg, E.-L., Mineralogical/Geochemical investigations in the fracture zone. 1995, SKB: Stockholm. 18. Klein, C., The 22nd edition of the Manual of Mineral Science. 2002: John Wiley & Sons, Inc. 19. Bailey, S.W., Chlorites: Structures and Crystal Chemistry, in Reviews in Mineralogy; Hydrous phyllosilicates (exclusive of micas), S.W. Bailey, Editor. 1988, Mineralogical Society of America. p. 725. 20. Langmuir, D., Aqueous Environmental Geochemistry. 1997, New Jersey: Prentice Hall. 21. Eggelton, R., A. and J. Banfield, F., 1985, The alteration of grantic biotite to chlorite. American Mineralogist. 70: p. 902-910. 22. Veblen, D.R. and J.M. Ferry, 1983, A TEM study of the biotite-chlorite reaction and comparison with petrologic observations. American Mineralogist. 68: p. 1160-1168.

67 23. Parry, W.T. and L.M. Downey, 1982, Geochemistry of hydrothermal chlorite replacing igneous biotite. Clays & Clay Minerals. 30: p. 81-90. 24. Albee, A.L., 1962, Relationships between the mineal association, chemical composition and physical properties of the chlorite series. American Mineralogist. 47(851-870). 25. Hey, M.H., 1954, A new review of chlorites. Mineralogical Magazine. 30: p. 277-292. 26. Saggerson, E.P. and L.M. Turner, 1982, General comments on the identification of chlorites in thin sections. Mineralogical Magazine. 46: p. 469-473. 27. Brindley, G.W. and G. Brown, Structures of Layer Silicates, in Crystal structures of clay minerals and their X-ray identification, G.W. Brindley and G. Brown, Editors. 1984, Mineralogical Society: London. p. 1-124. 28. Bailey, S.W., et al., 1971, Summary of national and international recommendations on clay mineral nomenclature by 1969-1970 Clay. Miner. Soc. Nomenclataire Comm. Clays & Clay Minerals. 64: p. 129-132. 29. Foster, M.D., 1962, Interpretation of the composition and a classification of chlorites. Prof. Pap. U.S. Geol. Surv. 414-A: p. 1-33. 30. Barbhisel, R. and P.M. Bertsch, Chlorites and Hydroxy-interlayed Vermiculite and Smectite, in Minerals in Soil Environments, J. Dixon and S. Weed, Editors. 1989, Soil. Sci. Soc. Am.: Madison, WI. p. 729-787. 31. Sverdrup, H., The Kinetics of Base Cation Release Due to Chemical Weathering. 1990, Lund: Lund University Press. 246. 32. Stumm, W. and R. Wollast, 1990, Coordination chemistry of weathering: Kinetics of the Surface- controlled dissolution of oxide minerals. Revs. Geoph. 28: p. 53-69. 33. Stumm, W., 1997, Reactivity at the mineral-water interface: dissolution and inhibition. Colloids and surfaces A: Physicochemical and Engineering Aspects. 120: p. 143-166. 34. Wieland, E., B. Wehrli, and W. Stumm, 1988, The coordination chemistry of weathering: III. A generalization on the dissolution rates of minerals. Geochim. Cosmochim. Acta. 52: p. 1969-1981. 35. Furrer, G. and W. Stumm, 1986, The coordination chemistry of weathering:I. Dissolution kinetics of &Al2O3 and BeO. Geochim. Cosmochim. Acta. 50?: p. 1847-1860. 36. Köhler, S.J., D. Bosbach, and E.H. Oelkers, 2005, Do clay mineral dissolution rates reach steady state? Geochim. Cosmochim Acta. 69: p. 1997-2006. 37. Kalinowski, B.E., C. Faith-Ell, and P. Schweda, 1998, Dissolution kinetics and alteration of epidote in acidic solutions at 25ºC. Chem. Geol. 151: p. 181-197. 38. Bradbury, M.H. and B. Baeyens, 2002, Sorption of Eu on Na- and Ca-montmorillonites: Experimental investigations and modelling with cation exchange and surface complexation. Geochimica et Cosmochimica Acta. 66(13): p. 2325-2334. 39. Sposito, G., Distinguishing Adsorption from Surface Precipitation, in Geochemical Processes at Mineral Surfaces, J.A. Davies and K.F. Hayes, Editors. 1986, American Chemical Society: Washington, USA. p. 683). 40. Wieland, E., et al., A surface chemical model of the bentonite-water interface and its implications for modelling the near field chemistry in a repository for spent fuel. 1994, SKB: Stockholm. 41. Lützenkirchen, J., Surface Complexation Models of Adsorption: A critical Survey in the context of Experimental Data, in Adsorption; theory, modelling and analysis, J. Toth, Editor. 2002, Marcel Dekker, Inc. 42. Sparks, D.L., Environmental Soil Chemistry. 2nd ed. 2003: Academic Press, An imprint of Elsevier. 352. 43. vanLoon, G.W. and S.J. Duffy, Environmental chemistry a global perspective. 2000: Oxford University Press. 492.

68 44. Stumm, W. and J.J. Morgan, Aquatic Chemistry. 3rd ed. 1996, New York: John Wiley & Sons. 1022. 45. Sposito, G., The Surface Chemistry of Soils. 1984, New York: Oxford Univ. Press. 46. Stumm, W., Chemistry of the Solid-Water Interface: Processes at the Mineral-Water and Particle- Water Interface in Natural Systems. 1992: John Wiley & Sons, Inc. 428. 47. Lützenkirchen, J., 1997, Ionic strength effects on cation sorption to oxides: Macroscopic observations and their significance in microscopic interpretation. J. Colloid Interface Sci. 195: p. 149-155. 48. Davis, J.A. and D.B. Kent, Surface complexation modeling in aqueous geochemistry, in Mineral- Water Interface Geochemistry, M.F. Hochella, Jr. and A.F. White, Editors. 1990, Mineral. Soc. Amer.: Washington, D.C. p. 177-260. 49. Dzombak, D.A. and F.M.M. Morel, Surface Complexation Modeling. Hydrous Ferric Oxide. 1990, USA: John Wiley & Sons, Inc. 393. 50. Lützenkirchen, J., ed. Surface Complexation Modelling. 2006, Elsevier Ltd. 51. Scheckel, K.G. and D.L. Sparks, 2001, Dissolution Kinetics of Nickel Surface Precipitates on Clay Mineral and Oxide Surfaces. Soil Sci Soc Am Journal. 65: p. 685-694. 52. Lützenkirchen, J. and P. Behra, 1996, On the Surface Precipitation Model for Cation Sorption at the (Hydr)oxide water interface. Aquatic Geochemistry. 1: p. 375-397. 53. Herbelin, A.L. and J.C. Westall, FITEQL A computer program for determination of chemical equilibrium constants from experimental data. 1999, Department of Chemistry, Oregon State University, USA: Corvallis. 54. Parkhurst, D.L. and C.A.J. Appelo, User's guide to PHREECQ (Version 2)- A computer program for speciation, batch reaction, one-dimensional transport and inverse geochemical calculations. 1999, U.S. Department of the Interior, U.S. Geological Survey: Denver, Colorado. 55. Sverjensky, D.A., 1994, Zero-point-of-charge prediction from crystal chemistry and solvation theory. Geochim. Cosmochim Acta. 58: p. 3123-3129. 56. Nagy, K.L., Dissolution and precipitation kinetics of sheet silicates, in Chemical Weathering Rates of Silicate Minerals, A.F. White and S.L. Brantley, Editors. 1995, Mineralogical Society of America: Washington DC. p. 173-233. 57. Metz, V. and J. Ganor, 2001, Stirring effect on kaolinite dissolution rate. Geochim. Cosmochim. Acta. 65: p. 3475-3490. 58. Malmström, M. and S. Banwart, 1997, Biotite dissolution at 25°C: The pH dependence of dissolution rate and stoichiometry. Geochim. Cosmochim. Acta. 61: p. 2779-2799. 59. Metz, V., Dissolution kinetics of smectite and kaolinite, in Department of Geological and Environmental Sciences. 2001, Ben-Gurion University of Negev: Beer Sheva. p. 168. 60. Ross, G.J., 1975, Experimental alteration of chlorites into vermiculites by chemical oxidation. Nature. 255: p. 133-134. 61. Ross, G.J., 1969, Acid dissolution of chlorites:Release of Magnesium, Iron and Aluminium and mode of acid attack. Clays & Clay Minerals. 17: p. 347-357. 62. Brandt, F., et al., 2003, Chlorite dissolution in the acid pH-range: a combined microscopic and macroscopic approach. Geochim. Cosmochim. Acta. 67: p. 1451-1461. 63. Hamer, M., et al., 2003, Dissolution of Ripidolite (Mg, Fe-chlorite in organic and inorganic acid solutions. Soil Sci Soc Am Journal. 67: p. 654-661. 64. Ross, G.J., 1968, Structural decomposition of an orthochlorite during its acid dissolution. Canadian Mineralogist. 9: p. 522-530. 65. Ross, G.J., 1967, Kinetics of acid dissolution of an orthochlorite mineral. Can. J. Chem. 45: p. 3031-3034. 66. May, H.M., et al., Aqueous dissolution of low-iron chlorite in dilute aid solutions at 25ºC, in Proc. 32nd Annual Meeting Clay Minerals Society. 1995: Baltimore, MD. p. 88 (abstract).

69 67. Malmström, M., et al., 1996, The dissolution of biotite and chlorite at 25°C in the near neutral pH region. J. Contaminant Hydrol. 21: p. 201-213. 68. Rochelle, C.A., et al. Experimental determination of chlorite dissolution rates. in Mater. Res Soc Proc. . 1995. Pittsburgh, PA. 69. Lowson, R., et al., 2005, The kinetics of the dissolution of chlorite as a function of pH and at 25°C. Geochim. Cosmochim. Acta. 69: p. 1687-1699. 70. Shahwan, T. and H.N. Erten, 1999, Radiochemical study of Co2+ sorption on chlorite and kaolinite. J. Radioanal. Nucl. Chem. 241: p. 151-155. 71. Bond, K.A., et al., Sorption of uranium(VI), plutonium and thorium onto aluminium oxide, muscovite and chlorite: An experimental and modelling study. 2001, AEA Technology plc.: Harwell, Oxfordshire. 72. Zorn, T., Untersuchungen der Sorption von U(VI) an das Gestein Phyllit zur Bestimmung von Oberflächenkomplexbildeungskonstanten. 2000, Technical University Dresden: Dresden. 73. Krawczyk-Bärsch, E., et al., 2004, Formation of secondary Fe-oxyhydroxide phases during the dissolution of chlorite - effects on uranium sorption. Applied Geochem. 19: p. 1403-1414. 74. Li, S.-H., et al., 2002, Study of adsorption behaviours of Cs+ abd Yb3+ on calcite, kaolinite, montmorillonite, chlorite and glauconite. J. Nucl. & Radiochem. 24(2): p. 70-76. 75. Eylem, C., H.N. Erten, and H. Göktürk, 1989, Sorption of barrium on kaolinite, montmorillonite and chlorite. Analyst. 114: p. 351-353. 76. Eylem, C., H.N. Erten, and H. Göktürk, 1990, Soprtion-desorption behaviour of barium on clays. Journal of Environmental Radioactivity. 11: p. 183-200. 77. Koppelman, M.H. and J.G. Dillard, 1977, A study of the adsorption of Ni(II) and Cu(II) by clay minerals. Clays & Clay Minerals. 25: p. 457-462. 78. Bradbury, M.H. and B. Baeyens, 1997, A mechanistic description of Ni and Zn sorption on Na-montmorillonite. Part II: Modelling. J. Contaminant Hydrol. 27: p. 223-248. 79. Baeyens, B. and M.H. Bradbury, 1997, A mechanistic description of Na and Zn sorption on Na-montmorillonite Part I: Titration and sorption measurements. J. Cont. Hydrol. 27: p. 199- 222. 80. Scheidegger, A.M., G.M. Lamble, and D.L. Sparks, 1996, Investigation of Ni sorption on pyrophyllite: an XAFS study. Environ. Sci. Technol. 30: p. 548-554. 81. Dähn, R., et al., 2003, Structural evidence for the sorption of Ni(II) atoms on the edges of montmorillonite clay minerals: A polarized X-ray absorption fine structure study. Geochim. Cosmochim. Acta. 67: p. 1-15. 82. Ferrow, E. and M. Roots, 1989, A preparation technique for TEM specimens; Application to synthetic Mg-chlorite. European Journal of Mineralogy. 1: p. 815-819. 83. Brunauer, S., P.H. Emmett, and E. Teller, 1938, Adsorption of gases in multimolecular layers. J. Am. Chem. Soc. 60: p. 309-319. 84. Scott, W.W., Scott's Standard Methods of Chemical Analysis. Fifth ed, ed. N.H. Furman. Vol. One. 1939: D. van Nostrand Co. Inc. 85. Meier, L.P. and G. Kahr, 1999, Determination of the cation exchange capacity of clay minerals based on the complexes of the Copper(II) ion with triethylenetetramine and tetraethylenepentamine. Clays and Clay Minerals. 47: p. 386-388. 86. Bickmore, B.R., et al., 1999, Methods for performing atomic force microscopy imaging of clay minerals in aqueous solutions. Clays & Clay Minerals. 47: p. 573-581. 87. Edén, M., H. Annersten, and Å. Zazzi, 2005, Pulse-assisted homonuclear dipolar recoupling of half-integer quadropolar spins in magic-angle spinning NMR. Chemical Physics Letters. 410: p. 24-30. 88. Levitt, M.H., 1997, The signs of frequencies and phases in NMR. Journal of Magnetic Resonance. 126: p. 164-182.

70 89. Bruno, J., I. Casas, and I. Puigdomenech, 1991, The kinetics of dissolution of UO2 under reducing conditions and the influence of an oxidized surface layer (UO2+X): Application of a continuous flow-through reactor. Geochim. Cosmochim. Acta. 55: p. 647-658. 90. Papageorgiou, S.K., et al., 2009, Prediction of binary adsorption isotherms of Cu2+, Cd2+ and Pb2+ on calcium alginate beads from single adsorption data. Journal of Hazardous Materials. 162: p. 1347-1354. 91. Blum, A. and A. Lasaga, 1988, Role of surface speciation in the low-temperature dissolution of minerals. Nature. 331: p. 431-433. 92. Brandt, F., Hetergogene Reaktionsmechanismen und deren Kinetik an Schichtsilikatoberflächen. 2003, Universität Karlsruhe: Karlsruhe, Germany. p. 91. 93. Kodera, K. and Y. Onishi, 1959, The Molecular Cross-sectional Areas for Determination of Specific Surface Areas of Solids. I. Carbon Black. Bulletin of the Chemical Society of Japan. 32: p. 356-361. 94. Kodera, K. and Y. Onishi, 1960, The Molecular Cross-sectional Areas for the Determination of Specific Surface Areas of Solids. II Titania and Silicia. Bulletin of the Chemical Society of Japan. 33: p. 338-343. 95. Trens, P., R. Denoyel, and J.C. Glez, 2004, Comparative adsorption of argon and nitrogen for the characterisation of hydrophobized surfaces. Colloids and surfaces A: Physiocochemical and Engineering Aspects. 245: p. 93-98. 96. Chow, K.Y. and D.J.W. Grant, 1988, Surface Analysis of Griseofulvin Powders by Krypton Adsorption: Evaluation of Specific Surface Area, BET Constant C and Polanyi Adsorption Potential. Powder Technology. 56: p. 209-223. 97. Henn, F., et al., 2007, DC conductivity, cationic exchange capacity, and specific surface area related to chemical composition of pore lining chlorites. Journal of Colloid and Interface Science. 311: p. 571-578. 98. Lasaga, A.C., Kinetic Theory in the Earth Sciences. Princetown series in geochemistry, ed. H.D. Holland. 1998, New Jersey: Princetown University Press. 99. Kalinowski, B.E. and P. Schweda, 1996, Kinetics of muscovite, phlogopite, and biotite dissolution and alteration at pH 1-4, room temperature. Geochimica et Cosmochimica Acta. 60: p. 367-385. 100. Holdren, G.R.J. and P.M. Spreyer, 1985, pH dependent changes in the rates and stochiometry of dissolution of an alkali feldspar at room temperature. American Journal of Science. 285: p. 994-1026. 101. Stillings, L.L. and S.L. Brantley, 1995, Feldspar dissolution at 25C and pH 3: Reaction stoihiometry and the effect of cations. Geochimica et Cosmochimica Acta. 59: p. 1483-1496. 102. Lin, F.-C. and C.V. Clemency, 1981, The dissolution kinetics of brucite, antigorite, talc and phlogopite at room temperature and pressure. American Mineralogist. 66: p. 801-806. 103. Hwang, J.-Y., 1988, Dissolution of chlorite and serictite by the treatment with dilute H2SO4 solution. Nendo Kagaku. 28: p. 119-125. 104. El Aamarani, F.Z., et al., 2002, Experimental study and modelling of the sorption of uranium (VI) onto olivine-rock. Applied Geochemistry. 17: p. 399-408. 105. Bors, J., H.N. Erten, and S. Assemi, Effect of the liquid-solid ratio on the distribution ratio (Rd) in sorption, in Scientific Basis for Nuclear Waste Management, XVIII, T. Murakami and R.C. Ewing, Editors. 1995, Materials Reserch Society: Kyoto, Japan. p. 1069-1076. 106. Scheidegger, A.M. and D.L. Sparks, 1996, Kinetics of the formation and the dissolution of nickel surface precipitates on pyrophyllite. Chem. Geol. 132: p. 157-164. 107. Perrin, P.D. and B. Dempsey, Buffers for pH and Metal Ion Control. Chapman and Hall Laboratory Manuals in Physical Chemistry and Biochemistry, ed. A. Albert. 1974, London: Chapman and Hall Ltd. 176.

71 108. Bradbury, M.H. and B. Baeyens, Experimental and Modelling Investigations on Na-Illite: Acid-Base Behaviour and the Sortption of Strontium, Nickel, Europium and Uranyl. 2005, Paul Scherrer Institut: Wettingen. 109. Sjöberg, S. and L. Lövgren, 1993, The application of potentiometric techniques to study complexation reactions at the mineral/water interface. Aquatic Sciences. 55: p. 324-335. 110. Sondi, I., J. Biscan, and V. Pradvdic, 1996, Electrokinetics of pure clay mineralas revisited. J Coll Interface Science. 178: p. 514-522. 111. Du, Q., et al., 1997, Adsorption of copper at the aqueous illite surfaces. J Coll Interface Science. 187: p. 232-242. 112. Laaksoharju, M., et al., 2008, Hydrogeochemical evaluation and modelling performed within the Swedish site investigation programme. Applied Geochemistry. 23: p. 1761-1795. 113. Gustafsson, Å., M. Molera, and I. Puigdomenech. Study of Ni(II) sorption on chlorite-A fracture filling mineral in granites. in Mater. Res. Soc. Proc. . 2004: Warrendale, PA. 114. Zazzi, Å., A.-M. Jakobsson, and S. Wold, 2009, Ni(II) sorption on natural Chlorite. p. unpublished manuscript.

72 Appendix

Appendix A: Experimental details for sorption isotherm experiments.

Sorption isotherms were performed at pH 6.0, 7.4 and pH 8.3 in 0.1 M NaClO4. These were chosen since they all are in the pH range where Ni(II) sorption occurs to the chlorite surface [113, 114] and the pH of 8.3 corresponds to sorption maximum 1 mM TRIS (2-amino-2-hydroxymethyl-1,3propan-dio)l was used at pH 8.3 and 7.4 and 1 mM MES) (2-Nmorpholino)ethane-sulphoric acid) was used at pH 6.0. Tests were performed to ensure that the buffers did not interact in the sorption process. The total Ni concentration was varied in the interval 10-4 to 10-8 M with no higher concentration to avoid Ni precipitation. 63Ni was used as a tracer in all Ni(II) solutions. After 7 days of equilibration, the samples were centrifuged (Hermle Z 252MK centrifuge) at 6000 rpm (3900g) for 10 minutes. The β-activity of 63Ni in the aqueous phase was measured using β-liquid scintillation counting (LSC) on a PACKARD Tri-Carb Liquid Scintillation Analyser Model 1500.

73 Appendix B: Summary of dissolution results. The tabulated values describi tbe the dissolution data, with just a few series are presented in this Appendix. Taberg chlorite, pH 2. Figure 3.4 and Figure 3.6 Days [Al] Acc [Fe] Acc [Mg] Acc [Si] mol/g Acc mol/g Dissolution mol/g Dissolution mol/g dissolution solution dissolution solution moles/m2 solution moles/m2 solution moles/m2 moles/m2 0.04 1.61E-04 5.50E-07 6.48E-05 8.70E-07 1.03E-03 8.35E-07 1.98E-04 2.41E-07 0.08 1.61E-04 6.76E-07 6.34E-05 1.03E-06 5.77E-04 1.04E-06 1.45E-04 3.06E-07 0.13 1.35E-04 7.93E-07 4.81E-05 1.17E-06 5.27E-04 1.18E-06 1.42E-04 3.61E-07 0.19 1.11E-04 9.31E-07 4.72E-05 1.28E-06 3.13E-04 1.34E-06 9.50E-05 4.25E-07 0.26 1.09E-04 1.08E-06 3.17E-05 2.25E-06 2.75E-04 1.47E-06 9.70E-05 4.87E-07 0.33 6.78E-05 1.19E-06 2.60E-05 2.42E-06 2.61E-04 1.58E-06 9.23E-05 5.42E-07 1.05 3.90E-05 1.90E-06 2.55E-05 3.13E-06 1.73E-04 2.53E-06 6.61E-05 1.05E-06 1.21 4.25E-05 2.02E-06 1.46E-05 3.33E-06 2.77E-04 2.74E-06 8.77E-05 1.16E-06 1.98 2.79E-05 2.51E-06 1.76E-05 3.75E-06 1.68E-04 3.77E-06 5.54E-05 1.64E-06 2.27 2.97E-05 2.66E-06 1.07E-05 3.87E-06 1.65E-04 4.06E-06 5.83E-05 1.78E-06 2.98 2.03E-05 2.99E-06 9.42E-06 4.34E-06 1.46E-04 4.73E-06 5.52E-05 2.14E-06 3.20 2.95E-05 3.09E-06 8.56E-06 5.33E-06 1.53E-04 4.92E-06 5.85E-05 2.25E-06 4.02 2.55E-05 3.50E-06 8.97E-06 5.70E-06 1.53E-04 5.68E-06 5.84E-05 2.67E-06 6.05 2.06E-05 4.35E-06 8.17E-06 5.96E-06 1.21E-04 7.35E-06 5.03E-05 3.63E-06 7.05 2.01E-05 4.72E-06 6.77E-06 6.42E-06 9.33E-05 7.99E-06 4.49E-05 4.05E-06 8.06 1.55E-05 5.05E-06 4.42E-06 7.03E-06 7.91E-05 8.52E-06 4.21E-05 4.44E-06 10.07 1.37E-05 5.59E-06 3.17E-06 7.19E-06 7.18E-05 9.43E-06 3.48E-05 5.12E-06 13.14 1.06E-05 6.27E-06 3.83E-06 7.31E-06 4.54E-05 1.05E-05 3.28E-05 6.03E-06 14.18 1.18E-05 6.48E-06 2.20E-06 7.39E-06 4.38E-05 1.08E-05 3.18E-05 6.32E-06 15.11 9.38E-06 6.66E-06 2.47E-06 7.50E-06 2.86E-05 1.10E-05 2.36E-05 6.53E-06 16.05 5.45E-06 6.78E-06 1.56E-06 7.50E-06 2.96E-05 1.11E-05 2.25E-05 6.73E-06 17.08 1.29E-05 6.96E-06 1.15E-06 7.50E-06 3.77E-05 1.14E-05 2.69E-05 6.95E-06 20.07 2.78E-06 7.38E-06 2.15E-06 7.50E-06 1.23E-05 1.18E-05 1.62E-05 7.51E-06 21.07 4.52E-06 7.45E-06 n.d. 7.50E-06 1.18E-05 1.19E-05 1.51E-05 7.65E-06 23.08 5.78E-06 7.64E-06 n.d. 7.50E-06 1.14E-05 1.20E-05 1.34E-05 7.90E-06 24.25 4.41E-06 7.75E-06 n.d. 7.50E-06 9.46E-06 1.21E-05 1.27E-05 8.04E-06 27.14 6.93E-06 8.04E-06 n.d. 7.50E-06 1.26E-05 1.21E-05 1.42E-05 8.37E-06 28.07 3.15E-06 8.13E-06 n.d. 7.50E-06 7.82E-06 1.23E-05 1.18E-05 8.48E-06 30.09 4.34E-06 8.27E-06 n.d. 7.50E-06 9.59E-06 1.24E-05 1.27E-05 8.69E-06

74 Dissolution rates for chlorites using normalised Si data, Figure 3.5. Data at steady-state, approx 22 days. Taberg Taberg Karlsborg Karlsborg pH R Log R Si Si 2 2 RSi mol/( m s) Log RSi mol/( m s) mol/( m2 s) mol/( m2 s) 2 1.29 ·10-12 -11.89 4.20 ·10-12 -11.38 4 2.66·10-13 -12.57 2.33 ·10-12 -11.63 6 3.32·10-14 -13.48 No data No data 7.4 4.23·10-14 -13.37 n.d. n.d. 10 4.74 ·10-14 -13.33 9.57 ·10-13 -12.02 12 1.79 ·10-13 -12.75 7.83 ·10-13 -12.11

Accumulated dissolution. Taberg chlorite pH 12. Figure 3.6. Acc Acc Acc Acc Days Dissolution Al Dissolution Fe Dissolution Mg Dissolution Si moles/m2 moles/m2 moles/m2 moles/m2 0.08 1.02E-08 1.25E-08 0 4.95E-09 0.13 1.25E-08 1.38E-08 0 8.38E-09 0.19 1.59E-08 1.54E-08 0 1.19E-08 0.26 1.98E-08 1.77E-08 0 1.53E-08 0.33 2.34E-08 1.94E-08 0 1.81E-08 1.05 6.20E-08 3.10E-08 0 5.11E-08 1.21 7.05E-08 3.61E-08 0 6.07E-08 1.98 1.11E-07 5.75E-08 0 1.00E-07 2.27 1.27E-07 6.09E-08 0 1.08E-07 2.98 1.65E-07 7.73E-08 0 1.28E-07 3.20 1.76E-07 8.16E-08 0 1.34E-07 4.02 2.20E-07 1.04E-07 0 1.42E-07 6.05 3.28E-07 2.37E-07 0 1.61E-07 7.05 3.81E-07 2.91E-07 0 1.71E-07 8.06 4.35E-07 3.06E-07 0 1.80E-07 10.07 5.42E-07 3.54E-07 0 1.99E-07 13.14 7.05E-07 4.25E-07 0 2.29E-07 14.06 7.54E-07 4.33E-07 0 2.37E-07 15.11 8.10E-07 4.50E-07 0 2.47E-07 16.05 8.61E-07 4.80E-07 0 2.56E-07 17.08 9.16E-07 5.06E-07 0 2.66E-07 20.07 1.07E-06 5.31E-07 0 2.94E-07 21.07 1.13E-06 5.45E-07 0 3.04E-07 23.08 1.23E-06 5.71E-07 0 3.23E-07 24.25 1.29E-06 5.81E-07 0 3.34E-07 27.14 1.45E-06 6.06E-07 0 3.61E-07 28.07 1.50E-06 6.15E-07 0 3.70E-07 29.09 1.55E-06 6.25E-07 0 3.79E-07

Accumulated dissolution. Karlsborg chlorite pH 2. Figure 3.6 Acc Acc Acc Acc Days Dissolution Al Dissolution Fe Dissolution Mg Dissolution Si moles/m2 moles/m2 moles/m2 moles/m2 0.22 4.20E-07 5.76E-07 3.95E-06 1.28E-07 0.37 8.26E-07 1.17E-06 7.78E-06 2.55E-07 1.03 2.67E-06 3.49E-06 2.51E-05 9.67E-07 1.19 3.10E-06 3.99E-06 2.81E-05 1.14E-06 2.04 5.46E-06 6.78E-06 3.53E-05 1.98E-06 2.20 5.97E-06 7.30E-06 3.64E-05 2.11E-06 3.02 8.63E-06 9.96E-06 4.12E-05 2.64E-06 3.20 9.21E-06 1.05E-05 4.20E-05 2.76E-06 4.01 1.23E-05 1.35E-05 4.41E-05 3.27E-06 4.17 1.30E-05 1.42E-05 4.43E-05 3.35E-06 5.09 1.68E-05 1.81E-05 4.50E-05 3.89E-06 7.03 2.68E-05 2.63E-05 4.58E-05 5.16E-06 8.08 3.29E-05 3.05E-05 4.61E-05 5.74E-06 9.03 3.87E-05 3.46E-05 4.64E-05 6.17E-06 10.09 4.47E-05 3.90E-05 4.67E-05 6.62E-06 11.04 4.96E-05 4.27E-05 4.69E-05 7.06E-06 14.07 6.66E-05 5.52E-05 4.74E-05 8.55E-06 16.07 7.78E-05 6.33E-05 4.76E-05 9.34E-06 18.07 8.75E-05 7.05E-05 4.76E-05 1.01E-05 21.19 1.07E-04 8.50E-05 4.77E-05 1.12E-05

Accumulated dissolution. Karlsborg chlorite pH 4. Figure 3.6 Acc Acc Acc Acc Days Dissolution Al Dissolution Fe Dissolution Mg Dissolution Si moles/m2 moles/m2 moles/m2 moles/m2 0.26 8.33E-07 4.72E-07 6.18E-06 9.41E-07 1.04 4.14E-06 8.04E-07 6.18E-06 3.89E-06 1.23 4.96E-06 8.93E-07 3.63E-05 4.49E-06 2.04 7.95E-06 1.20E-06 6.35E-05 6.06E-06 2.24 8.65E-06 1.27E-06 6.96E-05 6.39E-06 3.07 1.07E-05 1.58E-06 9.06E-05 8.06E-06 4.03 1.28E-05 1.93E-06 1.11E-04 9.80E-06 6.06 1.91E-05 2.62E-06 1.50E-04 1.42E-05 7.06 2.16E-05 2.94E-06 1.67E-04 1.66E-05 8.05 2.26E-05 3.20E-06 1.82E-04 1.90E-05 9.07 2.30E-05 3.39E-06 1.95E-04 2.18E-05 11.05 2.35E-05 3.74E-06 2.11E-04 2.70E-05 13.06 2.39E-05 4.26E-06 2.20E-04 3.16E-05 16.08 2.45E-05 5.31E-06 2.32E-04 3.62E-05 18.08 2.49E-05 6.01E-06 2.38E-04 3.80E-05 21.04 2.56E-05 6.99E-06 2.45E-04 3.97E-05 23.04 2.60E-05 7.66E-06 2.50E-04 4.05E-05 27.12 2.67E-05 9.04E-06 2.63E-04 4.15E-05 30.08 2.72E-05 9.99E-06 2.70E-04 4.19E-05

Accumulated dissolution. Karlsborg chlorite pH 12. Figure 3.6 Acc Acc Acc Acc Days Dissolution Al Dissolution Fe Dissolution Mg Dissolution Si moles/m2 moles/m2 moles/m2 moles/m2 0.22 1.25E-07 4.56E-09 4.85E-07 1.30E-08 0.37 2.32E-07 9.09E-09 1.38E-06 2.59E-08 1.03 5.00E-07 2.95E-08 3.58E-06 8.40E-08 1.19 5.84E-07 3.45E-08 4.19E-06 9.83E-08 2.04 9.54E-07 6.08E-08 7.39E-06 1.73E-07 2.20 1.01E-06 6.59E-08 7.90E-06 1.88E-07 3.02 1.27E-06 9.13E-08 1.03E-05 2.60E-07 3.20 1.32E-06 9.66E-08 1.08E-05 2.75E-07 4.01 1.58E-06 1.22E-07 1.31E-05 3.47E-07 4.17 1.63E-06 1.27E-07 1.35E-05 3.64E-07 5.09 1.92E-06 1.55E-07 1.62E-05 4.60E-07 7.03 2.56E-06 2.15E-07 2.26E-05 8.74E-07 8.05 3.07E-06 2.44E-07 2.54E-05 1.09E-06 9.03 3.54E-06 2.71E-07 2.70E-05 1.17E-06 10.09 3.88E-06 3.04E-07 2.81E-05 1.26E-06 11.04 4.18E-06 3.34E-07 2.86E-05 1.34E-06 14.07 5.15E-06 4.27E-07 2.94E-05 1.61E-06 16.07 5.70E-06 4.81E-07 2.97E-05 1.76E-06 18.07 6.17E-06 5.27E-07 2.99E-05 1.89E-06 21.19 6.95E-06 6.02E-07 3.01E-05 2.11E-06

Appendix C: Summary of sorption results. The tabulated values are describe the sorption data, with just a few series present in this Appendix. 5g/L Karlsborg [Ni]initial [NaClO4] pH Kd Log Kd % sorbed 1.00E-06 0.01 4.31 2.37 0.37 1.21 6.09 44.01 1.64 17.92 6.50 144.47 2.16 41.87 7.00 557.68 2.75 73.44 9.72 751.43 2.88 79.05 10.79 641.83 2.81 76.30 0.1 4.28 3.56 0.55 1.92 5.54 29.50 1.47 11.07 6.23 38.94 1.59 16.22 6.57 215.01 2.33 51.85 7.40 706.32 2.85 70.71 9.06 1688.39 3.23 89.42 10.41 2553.67 3.41 92.76 0.5 4.29 14.63 1.17 6.98 4.56 28.93 1.46 13.30 5.66 10.87 1.04 5.54 7.85 414.90 2.62 67.91 8.38 1017.79 3.01 83.63 8.91 1298.58 3.11 86.65 10.06 1783.79 3.25 89.97 10.16 1578.86 3.20 89.03

78 Appendix D: Table of titration data

The tabulated values describe the titration data, with just a few series present in this Appendix. Approximately one week of contact time prior to titrations with Karlsborg chlorite. pH Sigma pH Sigma Titration with base Titration with acid 7.521 0.081 7.299 0.046 7.651 -0.061 7.025 0.202 7.751 -0.204 6.846 0.372 7.826 -0.348 6.711 0.545 7.888 -0.493 6.607 0.720 7.947 -0.635 6.521 0.895 7.994 -0.780 6.450 1.070 8.031 -0.928 6.384 1.245 8.064 -1.076 6.329 1.420 8.099 -1.220 6.280 1.595 8.129 -1.367 6.233 1.769 8.154 -1.515 6.188 1.943 8.177 -1.664 6.145 2.117 8.201 -1.810 6.104 2.290 8.223 -1.957 6.061 2.463 8.242 -2.105 6.018 2.636 8.260 -2.253 5.972 2.807 8.277 -2.401 5.926 2.979 8.294 -2.548 5.878 3.149 8.310 -2.696 5.825 3.319 8.324 -2.844 5.766 3.487

Approximately two hours of contact time prior to titrations with Karlsborg chlorite. pH Sigma pH Sigma Titration with base Titration with acid 6.972 0.022 6.966 0.021 7.202 -0.135 6.729 0.200 7.442 -0.279 6.545 0.383 7.648 -0.409 6.400 0.567 7.809 -0.532 6.288 0.752 7.948 -0.643 6.192 0.937 8.049 -0.757 6.105 1.122 8.133 -0.869 6.021 1.306 8.202 -0.982 5.947 1.490 8.268 -1.087 5.874 1.672 8.319 -1.199 5.799 1.854 8.364 -1.313 5.720 2.033 8.401 -1.431 5.643 2.212 8.441 -1.541 5.568 2.389 8.473 -1.657 5.488 2.564 8.501 -1.777 5.399 2.734 8.525 -1.899 5.317 2.903 8.552 -2.016 5.236 3.067 8.574 -2.137 5.160 3.230 8.594 -2.261 5.083 3.388 8.611 -2.388 5.016 3.544

Simulation of titration data based on experimental data approx. 2 h of contact time and Karlsborg chlorite. pH sigma 5.016 3.19E+00 5.083 3.06E+00 5.16 2.89E+00 5.236 2.72E+00 5.317 2.53E+00 5.399 2.33E+00 5.488 2.11E+00 5.568 1.91E+00 5.643 1.73E+00 5.72 1.55E+00 5.799 1.37E+00 5.874 1.21E+00 5.947 1.07E+00 6.021 9.32E-01 6.105 7.94E-01 6.192 6.67E-01 6.288 5.43E-01 6.400 4.20E-01 6.545 2.88E-01 6.729 1.54E-01 6.966 1.21E-02 6.972 8.71E-03 7.202 -1.26E-01 7.442 -2.98E-01 7.648 -4.98E-01 7.809 -7.04E-01 7.948 -9.22E-01 8.049 -1.11E+00 8.133 -1.28E+00 8.202 -1.43E+00 8.268 -1.58E+00 8.319 -1.70E+00 8.364 -1.80E+00 8.401 -1.89E+00 8.441 -1.99E+00 8.473 -2.07E+00 8.501 -2.14E+00 8.525 -2.20E+00 8.552 -2.27E+00 8.574 -2.32E+00 8.594 -2.37E+00 8.611 -2.41E+00

80 Appendix E: Literature data

Brandt et al. [62]:

(Mg5.54FeII3.02 FeIII0.94Al2.48)(Si5.33 Al2.66)O20(OH)16 May et al. [66]

(Mg9.8FeII0.2 FeIII0.2Al1.8)(Si6.0 Al2.0)O20(OH)16 Rochelle et al. [68]

(Mg5.36Fe3.97Al2.62)(Si5.56 Al2.46)O20(OH)16 Malmström et al. [67]

(Mg9.8FeII0.2FeIII0.2Al1.4)(Si7.0 Al1.0)O20(OH)16 Lowson et al. [69]

(Mg5.52FeII3.80FeIII0.14Al1.94)(Si4.96 Al3.04)O20(OH)16 Hamer et al. [63]

(Mg5.60FeII3.90Al2.50)(Si5.50 Al2.50)O20(OH)16 Sverdrup et al. [31]

(Mg8.4Fe1.2Al2.0)(Si5.4 Al2.6)O20(OH)16 Ross [61]

(Mg8.68FeII0.54FeIII0.44Al2.10)(Si5.96 Al2.04)O20(OH)16

Reference BET Fraction used Comment Corresponding m2/g in numbers from dissolution reference list experiments Brandt et al. 1.1 ± 63 – 200 µm The Cca-2 chlorite [62] 0.1 Lowson et al. 1.41 38 - 75 µm [69] Hamer et al. 1.44 38 – 125 µm Petreated material [63]

Sverdrup 1.76 No information [31] Rochelle et al. 0.89 120 – 250 µm [68] Malmström et al. 4.38 75-125 µm Used He for BET [67] analysis Malmström et al. 3.36 125-300 µm Used He for BET [67] analysis