Dissolution kinetics of selected natural minerals relevant to potential CO2-injection sites − Part 1: A review

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS Chemie der Erde xxx (2016) xxx–xxx

Contents lists available at ScienceDirect

Chemie der Erde journal homepage: www.elsevier.de/chemer

Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review Astrid Holzheid Institut für Geowissenschaften, Universität Kiel, Ludewig-Meyn-Straße 10, 24118 Kiel, Germany

a r t i c l e

i n f o

Article history: Received 31 March 2016 Received in revised form 9 July 2016 Accepted 23 September 2016 Keywords: Dissolution Kinetics Carbonates Feldspar group minerals CO2 –sequestration

a b s t r a c t This publication provides a literature review on experimental studies of dissolution kinetics of mainly carbonates and feldspar group minerals, i.e. most common minerals at potential CO2 -injection and/or storage sites. Geochemical interaction processes between injected CO2 and coexisting phases, namely reservoir and cap rock minerals and formation fluids close to the CO2 -injection site can be simulated by flow-through or mixed flow reactors, while processes far from the injection site and long-term processes after termination actual CO2 -injection can be mimicked by batch reactors. At sufficient small stirring rates or fluid flow rates as well as low solute concentrations flow-through reactors are also able to simulate processes far from the injection site. The experimental parameter temperature not only intensifies the dissolution process, the dominant dissolution mechanisms are also influenced by temperature. The dissolution mechanisms change from incongruent and surface controlled mechanisms at lower temperatures to congruent and transport controlled mechanisms at higher temperatures. The CO2 partial pressure has only a second order influence on dissolution behavior compared to the influence of pH-value and ionic strength of the CO2 -bearing brine. Minerals exposed to CO2 -bearing brines at elevated temperatures and pressures are subject of alteration, leading to severe changes of reactive surfaces and potential precipitation of secondary minerals. Computational simulations of mineral reactions at potential CO2 storage sites have therefore to include not only the time-resolved changes of dissolution behavior and hence kinetics of mineral dissolution, but also the influence of secondary minerals on the interaction of the minerals with CO2 -enriched brines. © 2016 Elsevier GmbH. All rights reserved.

1. Introduction The CO2 emission of fossil fuel burning power plants is a widespread problem due to the increasing global energy consumption. As one of the most important greenhouse gases CO2 contributes significantly to the climate change. For several years scientists as well as the energy producing industries are concerned with the topic of CO2 sequestration and work on methods and techniques to minimize the CO2 emission that is directly related to fossil fuel burning power plants. In general, CO2 sequestration includes the separation of CO2 from power plants, the transport and the storage of CO2 in capable reservoirs. To make the storage technique as safe as possible it is important to have profound knowledge of the processes in the reservoir and the surrounding rock formations. This requires research activities in quite diverse fields of scientific research. One example is the modeling of different CO2 -injection and expansion scenarios in the reservoir. Modeling

E-mail address: [email protected]

software is partly adapted from existing software or newly generated with regard to the topic of CO2 sequestration. Geomechanical and geochemical interaction processes between the injected CO2 and coexisting phases in the reservoir are part of the required input data and have to be studied at first either in laboratory experiments or with the help of field/pilot projects. The aim of all these research activities is to better estimate the dimension, the risk, and the sustainability of CO2 sequestration. The knowledge of geochemical interactions, e.g., dissolution and/or precipitation processes, between injected CO2 and coexisting phases (e.g., reservoir and cap rock minerals and formation fluids) as well as the particle displacement of less consolidated mineral grains due to the superimposed pressure during and/or after the CO2 -injection is crucial to evaluate the safeness of potential reservoir sites. Potential injection sites generally consist of reservoir rocks (e.g., sandstones, coalbeds) and cap rocks (e.g., mudstones, limestones, shales), which include silicates like quartz (SiO2 ), feldspar (CaAl2 Si2 O8 − NaAlSi3 O8 − KAlSi3 O8 ), pyroxene ((Mg,Fe)2 Si2 O6 -Ca(Mg,Fe)Si2 O6 ) or olivine ((Mg,Fe)2 SiO4 ), carbonates for instance calcite (CaCO3 ), dolomite (CaMg(CO3 )2 ), or

http://dx.doi.org/10.1016/j.chemer.2016.09.007 0009-2819/© 2016 Elsevier GmbH. All rights reserved.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 2

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

magnesite (MgCO3 ), and/or clay minerals like kaolinite (Al2 Si2 O5 (OH)4 ), illite (K0.65 Al2 Al0.65 Si3.35 O10 (OH)2 ), or smectite group minerals (e.g., montmorillonite (Na,Ca)0.3 (Al,Mg)2 Si4 O10 (OH)2 ·nH2 O) to name a few. All these natural mineral phases exhibit different reactions in contact with CO2 -bearing fluids. Therefore a reliable statement about geochemical interactions in a capable reservoir requires detailed information about kinetic processes of the individual mineral and possible reaction partners. In the past numerous dissolution experiments on natural monomineralic materials like the mineral phases mentioned above were accomplished by different research groups. A complex literature survey reveals a high diversity of scientific questions. The following represents a selection of various research topics that only focus on studies regarding carbonate dissolution processes: Earlier studies mainly dealt with (i) exchange and/or replacement reactions on carbonates (Chou et al., 1989), (ii) relationships between active sites of strained calcites and the dissolution rate (Schott et al., 1989), or (iii) understanding of simple CaCO3 –CO2 –H2 O system processes to resolve various questions in the aqueous modeling (Plummer and Busenberg, 1982). The aim of these studies was to understand fundamental mineral dissolution processes at simple experimental conditions (e.g., atmospheric pressure, deionized water or low ionized salt solutions). In contrast numerous more recent publications deal with the subject of CO2 sequestration. The research group of Oleg Pokrovsky (Toulouse, France) for example studies the characterization of reactivity of carbonates under elevated pressures (Pokrovsky et al., 2005) or the influence of inorganic and organic ligands on the calcite and/or magnesite dissolution processes (Pokrovsky et al., 2009). Other authors describe dissolution processes on carbonates as functions of (i) fluid composition and/or ionic strength (Finneran and Morse, 2009; Gautelier et al., 2007; Gledhill and Morse, 2006a), (ii) mineral impurities (Eisenlohr et al., 1999) or (iii) various experimental parameters like pH-value, temperature or pressure (Druckenmiller et al., 2006; Yadav et al., 2008). These experiments were adapted to more natural conditions (i.e. prevailed pressure, temperature, and/or fluid composition as well as using sophisticated apparatuses). Furthermore molecular simulations of reactions between injected CO2 and coexisting phases as well as computational modeling of complex nano- and macro-scaled processes during the CO2 sequestration play an important role in recent publications (e.g., Kang et al., 2010; Bauer et al., 2012; Crawshaw and Boek, 2013; Hamm et al., 2013; Hellevang et al., 2013; Power et al., 2013; Steefel et al., 2013). The study of interaction processes, exchange and/or replacement reactions requires a large number of various experimental parameters − reactor type, experimental conditions, starting materials to name a few − and raises the following questions: Which experimental setups and/or conditions describe most precisely certain natural circumstances? Which parameters influence the kinetic data and how strong are these effects? Which literature data is suitable to compare to own data? Is it meaningful to describe data compilations of e.g., dissolution rates by using multi parameter regressions neglecting interconnected influences of parameters among themselves? Recent publications address these questions in the context of dissolution of mainly phyllosilicates in CO2 -bearing brines (Marty et al., 2015: montmorillonite and other clay minerals; Black et al., 2015: chlorite), while our main focus is on carbonates and feldspar group minerals. This publication is part 1 of a series of 2. Part 1 describes the differences of raised scientific questions and consequential changes of required experimental conditions and parameters. The influence of each single experimental parameter as well as the mutual interaction of various parameters on dissolution processes is also addressed.

Part 2 of the series focuses on own experiments to study dissolution behavior of mainly carbonate but also feldspar group minerals in CO2 -bearing brines that are placed in the context of CO2 sequestration as well as on time-resolved changes of reactive surfaces of calcite during exposure to CO2 -bearing brines at elevated temperatures and pressures (Holzheid, 2016 − same issue). In the present study first of all suitable experimental setups to simulate reactions during CO2 sequestration are summarized. These setups are put in the context of fluid regimes that prevail at CO2 sequestration sites and resultant dissolution mechanisms of the minerals that are exposed to the CO2 -bearing brines. This is followed by a detailed discussion on the influence of various parameters on dissolution (e.g., variables as temperature, total pressure, CO2 partial pressure or material properties like pH-value and ionic strength of the CO2 -bearing brine or reactive surfaces of the minerals) and concluding remarks. 2. Experimental setups and variables and their impact on kinetic data To illustrate the high diversity of scientific questions and to estimate the influence of individual experimental parameters on the dissolution behavior, a detailed literature study was performed. According to possible reservoir and cap rock materials the literature survey includes mineral dissolution reactions of carbonates, quartz, various feldspar group members, and clay minerals. Comparisons of dissolution rates are based on 30 representative publications, which provide an overview of the experimental work of the last forty years. All used references of dissolution rates are listed in Table 1. At first an overview will be given regarding the two mostly used experimental setups and the prevailed fluid regimes as well as mineral dissolution mechanisms during the experiments. Following that the focus will be on the impact of experimental variables (e.g., temperature and pressure) and material properties like pH-values and ionic strengths of the brine and exposed reactive surfaces of the mineral of interest on the dissolution rates of various minerals. 2.1. Experimental setups Experimental setups have to mimic natural procedures and/or storage environments as precise as possible to evaluate the risk of CO2 sequestration. Fig. 1 shows a schematic representation of one possible virtual geological storage scenario, namely the geological trapping in oil/gas deposits or saline aquifers. Porous rock formations as for example sandstones build the reservoir when dense CO2 from industrial separation processes will be injected. The sandstone layers are surrounded by cap rock formations like mudstones, which trap the injected CO2 in the reservoir. Based on Fig. 1 it can be noticed that different locations in the reservoir exhibit different geomechanical, geochemical, and flow regimes. At the point of injection, dense CO2 pushes away most of the formation water and/or gas/oil, respectively, and causes a disturbed multiphase fluid flow. In contrast, phases like dissolved CO2 , CO2 bubbles, and formation water or gas/oil dominate the reservoir far from the injection site, where a natural fluid flow is restored or undisturbed. In addition, the lithologic pressure of the reservoir depends on the depth of the geological formation that serves as reservoir. Table 2 shows exemplary three different injection projects and their reservoir conditions. The wide range of injection site parameters becomes apparent already on these three examples. Whereas the Qinshui Basin (China), i.e. the CO2 storage depth, is situated at about 471 m depth, the Triassic sandstone formation of Ketzin (Germany) exists in 630–650 m depth, and the carboniferous sandstone formation in Algeria is located in a depth of more than 1800 m (Michael

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

3

Fig. 1. Schematic scenario of CO2 -sequestration (top) and various reactor types and their possible capabilities for experimental studies on CO2 -sequestration aspects (bottom): bottom left − flow-through column, bottom center − mixed flow reactor (after Gautier et al., 2000), bottom right − closed batch reactor (after Beier, 2012). Fig. is taken from Beier (2012).

et al., 2010; Canadian International Development Agency, 2007). Temperature and pressure variations (25–90 ◦ C, 12–179 bar) are directly related to the reservoir’s depth. Hence, not only different locations in the reservoir exhibit various geomechanical, geochemical, and flow-through regimes, also the temperature and pressure conditions are highly variable and depend on the reservoir’s depth. The range of these various regimes and conditions hints at diverse experimental setups to simulate natural procedures at the storage environment. Despite the various scientific questions mainly two different reactor types are described in the literature: flow-through reactors (Fig. 1 left bottom box, e.g., after Gautier et al., 2000) and batch reactors (Fig. 1 right bottom box, e.g., Beier, 2012). Depending on the regulated stirring rate and fluid flow, respectively, both reactor types are able to reproduce laminar as well as turbulent fluid regimes. Reactions close to the CO2 -injection site can be simulated by flow-through reactors (also called mixed flow reactors in some publications). This experimental setup permits a fluid flow (e.g., brine, gas phase or brine-gas mixture) through a rock body (e.g., sandstone drill core, selected bulk solids), see for more details the most recent publications by Huq et al. (2015) and Tutolo et al. (2015). Flow-through reactors are open systems, where unreacted fluid is continuously injected and reacted fluid is successively extracted during the entire run duration. The accomplished fluid flow rates as well as stirring rates mentioned in the literature range from 2.5 × 10−4 to 20 ml/min and 100 to 1000 rpm. The fluid flow is

induced by various pump systems and the stirring is induced by impeller systems, magnetic stirrer or the rotating disk method. For the latter method monomineralic solid phases or pressed powder pellets are installed on a holder and rotated continuously. Fluid volumes of the used reactors vary between 75 and 550 ml. At sufficient small stirring rates or fluid flow rates, flow-through reactors are also able to simulate processes far from the injection site, where natural, i.e. undisturbed, fluid flow regimes prevail (e.g., at hydraulic conductivities of sandstones of 3 × 10−10 to 6 × 10−6 m/s, Domenico and Schwartz, 1990). In contrast to flow-through reactors, batch reactors are able to reproduce processes far from the injection site as well as long-term processes in the reservoir after termination of actual CO2 -injection. Only limited exchange between fluid and solid in a natural reservoir with undisturbed flow regime can be simulated due to the static experimental conditions of batch reactors. In the literature batch reactors with magnetic stirrer, impeller systems or the rotating disk method are described, where the accomplished stirring rates range from 10 to 2630 rpm. Occasionally an internal circulation of the fluid is mentioned. Volume data for the used batch reactors range from 30 to 550 ml. Both reactor types (flow-through reactors and batch reactors) are able to simulate high temperature and high pressure conditions. Temperatures mentioned in the literature range from 2.5 ◦ C to 300 ◦ C and the total pressure ranges from 1 to 124 bar. In consequence, possible CO2 sequestration scenarios (e.g., as listed in Table 2) could be simulated by these two experimental setups. But attention must be paid to the fluid to solid

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model

ARTICLE IN PRESS

CHEMER-25406; No. of Pages 21

A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

4

Table 1 Complete list of mineral dissolution experiments and corresponding references regarding compared dissolution rates. mineral

#

references

aragonite CaCO3 calcite CaCO3

2 14

dolomite CaMg(CO3 )2

6

magnesite MgCO3

2

witherite BaCO3 (with minor amounts of Ca) albite NaAlSi3 O8

1

Chou et al. (1989), Cubillas et al. (2005) Berner and Morse (1974), Plummer et al. (1978), Schott et al. (1989), Chou et al. (1989), Alkattan et al. (1998), Alkattan et al. (2002), Gledhill and Morse (2004), Palandri and Kharaka (2004), Cubillas et al. (2005), Pokrovsky et al. (2005), Gledhill and Morse (2006a), Yadav et al. (2008), Finneran and Morse (2009), Pokrovsky et al. (2009) Gautelier et al. (1999), Lund et al. (1973), Pokrovsky et al. (2005), Gautelier et al. (2007), Yadav et al. (2008), Zhang et al. (2007) Pokrovsky et al. (2005), Chou et al. (1989) Chou et al. (1989)

5

anorthite CaAl2 Si2 O8

2

orthoclase KAlSi3 O8 plagioclase (bytownite) (Na,Ca)(Al,Si)Si3 O8 plagioclase (labradorite) (Na,Ca)(Al,Si)Si3 O8 plagioclase (oligoclase) (Ca,Na)(Al,Si)Si3 O8 K-rich feldspar K∼0.8 Na∼0.2 (Al,Si)4 O8

1 1

quartz SiO2

3

2 1 3

kaolinite Al2 Si2 O5 (OH)4 1 1 montmorillonite (Na,Ca)0.3 (Al,Mg)2 Si4 O10 (OH)2 ·4H2 O

Oelkers et al. (1994), Stillings and Brantley (1995), Chen and Brantley (1997), Blake and Walter (1999), Harouiya and Oelkers (2004) Oelkers and Schott (1995), Sorai et al. (2007) Blake and Walter (1999) Stillings and Brantley (1995) Stillings and Brantley (1995), Blake and Walter (1999) Stillings and Brantley (1995) Gautier et al. (1994), Stillings and Brantley (1995), Harouiya and Oelkers (2004) Dove and Crerar (1990), Blake and Walter (1999), Harouiya and Oelkers (2004) Oelkers et al. (1994) Golubev et al. (2006)

Table 2 Reservoir information, parameters, and conditions of three ongoing injection projects. Ketzin (Germany)a Qinshui (China)b In Salah (Algeria)a reservoir material

Triassic sandstone (Stuttgart Fm.) reservoir depth 630–650 m 80 m reservoir thickness reservoir temperature 34 ◦ C 73 bar pressure 86 t/day injection rate pilot project utilization a b

Permian coal

471 m 6m 25 ◦ C 12.96 bar 40 t/day pilot project

Carboniferous sandstone (Krechba Fm.) 1850 m 29 m 90 ◦ C 179 bar 3500 t/day commercial

Michael et al. (2010). Canadian International Development Agency (2007).

ratios during the experimental simulations to be not too far from the reservoir condition. These fluid to solid ratios strongly depend on the geological storage parameters of the different locations like porosity and permeability of the rock formations as well as the formation water’s flow regimes. 2.1.1. Fluid regimes and predominant dissolution mechanisms To compare literature data, the knowledge of dissolution mechanisms during different flow regimes is crucial. Various authors like Plummer et al. (1978), Dreybrodt and Buhmann (1991), Raines and Dewers (1997), and Pokrovsky et al.

(2005) describe dissolution mechanisms for various experimental setups (closed undisturbed systems, closed systems with laminar fluid flow, closed systems with turbulent fluid flow). The starting point (t = 0) of each experiment − regardless of the reactor type − is illustrated in Fig. 2 A: A solid sample (e.g., individual mineral crystal, mineral powder or pressed mineral powder pellet) with known surface and a fluid with known composition are in direct contact to each other. At the beginning of the experiment the fluid is present in its original composition and has no solutes from mineral dissolution processes. In addition, the initial concentrations of possible solutes within the original fluid are far from the saturation concentrations of these solutes and therefore allow dissolution processes. Furthermore p-T conditions of the experiment trigger mineral dissolution processes. With the initiation of dissolution (t = 1, Fig. 2B) in closed undisturbed systems, like batch reactors without stirring and/or fluid flow, transport controlled processes will start (see also Plummer et al., 1978; Raines and Dewers, 1997; Pokrovsky et al., 2005). These transport controlled mechanisms are controlled by a multiplestep process, where chemical and hydrodynamical interactions between mineral and fluid take place (Raines and Dewers, 1997). At first the reactants of the initial fluid have to reach the mineral surface and these reactants have to be adsorbed on the mineral surface to trigger a chemical reaction between fluid and solid. As final step the reaction products have to be transported and incorporated into the initial solution. A diffusion layer forms during this multiple-step process, i.e. a gradient of solute concentration develops within the fluid. Close to the reactive surface the solute concentration is the highest. With increasing distance to the reactive surface, the solute concentration decreases. With proceeding run duration (t = 2, Fig. 2C) the concentration gradient between the reactive surface and the fluid becomes steeper due to the ongoing release of solutes. Along with these phenomena the diffusion layer also expands. Dreybrodt and Buhmann (1991) ascribe two inhibiting effects to the expansion of the diffusion layer: (i) thicker diffusion layers cause slower diffusion velocities of reactants and reaction products and (ii) heterogeneity of chemical reactions on the reactive surface slows down the dissolution. Due to these limiting effects, reactions within undisturbed closed reactor systems are termed “transport controlled” processes. Even at slow diffusion velocities in the presence of thick diffusion layers, an exchange of reactants and reaction products takes place and results in an increase of solutes in the total fluid volume. During appropriate run duration it is possible to achieve equilibrium conditions between solid and fluid, i.e. the dissolution rate approaches zero. Starting a movement process, i.e. stirring the fluid, rotating the sample or passing fluid through, the hydrodynamic conditions in the described closed batch reactor systems change extremely in some cases. In the case of a laminar flow regime (Re ≤12.050 after Raines and Dewers, 1997) the following will happen (Fig. 2D): as a consequence of the initiating movement, the transport of both, reactants and reaction products, is not only determined by diffusion processes, but is also determined by hydrodynamic processes. Raines and Dewers (1997) describe the developing interface between fluid and solid as a hydrodynamic boundary layer compared to a diffusion layer in transport controlled regimes. The mass transport of reactants/reaction products accelerates due to the movement, which results in a decrease of the hydrodynamic boundary layer thickness. As an accompanying fact the dissolution rate increases. Furthermore the concentration gradient of solutes between mineral surface and fluid phase becomes steeper. The transport of reactants/reaction products is still determined by diffusion processes within the hydrodynamic boundary layer, whereas a perfect mixing of reactants/reaction products beyond the hydrodynamic boundary layer takes place by convective fluid flow (Dreybrodt

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

5

Fig. 2. Evolution of fluid chemistry next to solid-fluid boundary in closed reactors and predominant dissolution mechanisms. A − Starting point for a dissolution process (t = 0): solid and fluid are in direct contact with each other. The mineral surface and the fluid composition are well-defined. The fluid is present in its original composition and does not include solutes from mineral dissolution. The initial concentrations of possible solutes within the original fluid are far from the saturation concentrations of these solutes. Pressure and temperature conditions in the reactor allow dissolution processes. B − Beginning of the dissolution process (t = 1) in a closed reactor system without stirring or fluid flow: caused by a five stage dissolution process a diffusion layer builds up whereat a concentration gradient of solutes arises (see text for more details). The overall concentration of solutes in the fluid increases as result of the dissolution process. The slowest dissolution step determines the rate of the overall reaction progress. C − Ongoing dissolution process in a closed reactor system without stirring or fluid flow (t = 2): proceeding dissolution processes result in an increase of the diffusion layer thickness. The concentration of solutes near the solid surface increases with increasing run duration and the concentration gradient gets steeper. With ongoing run duration

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

6

and Buhmann, 1991). With adequate dissolution of the mineral phase an equilibrium state will be achieved faster in experiments including stirring/fluid flow than in completely static experiments. This is due to higher dissolution rates within stirred/flow-through systems. Various authors not only describe changes in dissolution rates between static and stirred/flow-through systems, they also describe dependencies of the dissolution rate regarding the stirring rate/disk rotation speed/fluid flow rate. Dreybrodt and Buhmann (1991), Raines and Dewers (1997) as well as Pokrovsky et al. (2005) for example describe a linear increase of the dissolution rate with increasing stirring rate/disk rotation speed/fluid flow under laminar fluid flow conditions (Fig. 3A). In addition Plummer et al. (1978) as well as Alkattan et al. (1998) describe an influence of the dissolution rate by the position of the sample or stirring propeller within the reactor. Similar configurations resulted in linear dependencies of the dissolution rate with regard to the stirring rate/flow rate. However, small changes in the configuration have severe impacts on the dissolution behavior (Fig. 3B). A further increase in the stirring rate/fluid flow within closed batch reactor systems induces a turbulent flow regime (Fig. 2E). This turbulent fluid flow could cause a complete breakdown of the hydrodynamic boundary layer which results in higher dissolution rates as observed by Raines and Dewers (1997). They describe a change from linear trend of the dissolution rate at laminar fluid regimes to parabolic trend at turbulent fluid regimes. In the above-mentioned scenarios the transport of reactants/reaction products through the diffusion layer and the hydrodynamic boundary layer, respectively, was the limiting factor of the dissolution processes. As the mass transport of reactants/reaction products is significant due to high flow speed of the fluid, chemical reactions on the mineral surface are the limiting factor and significantly influence the dissolution rate (see e.g., for more detail Dreybrodt and Buhmann, 1991). Therefore the occurring processes are no longer transport controlled processes but rather surface controlled processes (Raines and Dewers, 1997). Based on the movement process all reaction products which are released during the dissolution process will disperse evenly in the fluid (Fig. 2E). A solute concentration gradient between mineral surface and fluid cannot be formed. In regard to the above mentioned processes of closed batch reactor systems − with and without stirring/fluid flow − well defined differences of the dissolution rate can be observed. Within a completely static reactor a diffusion layer builds up and consequently the dissolution process slows down. Once a movement mechanism initiates, the diffusion layer/hydrodynamic boundary layer shrinks. At laminar flow regimes the shrinking diffusion layer/hydrodynamic boundary layer results in a linearly increasing dissolution rate with increasing stirring rate/disk rotation speed/fluid flow. During experiments with turbulent flow regimes the hydrodynamic boundary layer disappears completely with the result of a parabolic increase of dissolution rates with increasing stirring rate/disk rotation speed/fluid flow. The initial transport controlled

dissolution processes change into surface controlled dissolution processes. It is obvious that different experimental setups exhibit various levels of difficulty of the characterization of the dissolution processes. However, Dreybrodt and Buhmann (1991) conclude that rotating disk experiments are most suitable for determination of kinetic data because of defined hydrodynamic conditions and simple calculations. But the sample volumes have to be taken into account as well: if the fluid − to − solid ratio is large, longer experimental run duration is required due to slow dissolution processes. Less suitable to determine kinetic data are experiments where solid and fluid form a suspension and both, solid and fluid, are exposed to continuous stirring. In this case hydrodynamic processes are not well defined. In contrast, Raines and Dewers (1997) believe in so called mixedflow reactors as the method of choice to determine kinetic data. If the system has achieved steady state conditions, dissolution rates can be directly determined by changes of influent and effluent solute concentrations. In addition to Dreybrodt and Buhmann (1991), Raines and Dewers (1997) see advantages of using rotating disk setups due to the simple geometry as well as the known hydrodynamic conditions. To conclude, every scientist who uses an experimental setup − regardless of the individual type of reactor − should be aware of differences within the systems and should adapt their own setup to the desired scientific topic or questions. 2.2. Experimental variables In addition to the above mentioned experimental techniques and their associated dissolution mechanisms, the knowledge of experimental variables is equally important to estimate their influences on the mineral dissolution and/or precipitation process. The following two subsections will describe the parameters temperature and CO2 partial pressure (pCO2 ) as variables that influence dissolution and/or precipitation mechanisms. 2.2.1. Temperature Studies including experiments with carbonate minerals were chosen for a first comparison. In the past carbonate minerals were the preferred mineral of interest because of their distinct solubility behavior and fast kinetics of the dissolution reactions. Thus numerous experimental studies result in a substantiated data base of dissolution rates of carbonate minerals which allows a detailed comparison of various influencing parameters. Furthermore, in respect to CO2 sequestration processes, carbonates are able to represent mineral phases of reservoir materials both, as carbonates in grain assemblages and as part of the bonding cement. In Fig. 4 dissolution rates of carbonates (aragonite (CaCO3 ), calcite (CaCO3 ), dolomite (CaMg(CO3 )2 ), magnesite (MgCO3 − with minor amounts of Ca), witherite (BaCO3 − with minor amounts

the overall solute concentration within the fluid increases because of diffusion effects. The saturation concentration can be reached if the dissolution process continues. D − Influence of a movement process in a closed reactor system (t = 3): stirring and/or fluid flow induces laminar fluid flow (Re ≤ 10.000). Merging of the fluids results in a decrease of the hydrodynamic boundary layer thickness. The solute concentration gradient within the hydrodynamic boundary layer is steep at the beginning of the movement process. The ongoing movement effects an increase of the overall solute concentration and therefore a flatter solute concentration gradient develops. Since the reaction products are not removed, a saturation concentration and consequently equilibrium between mineral and fluid can be reached. E − Influence of a faster movement process in a closed reactor system (t = 4): stirring and/or fluid flow induces turbulent fluid flow (Re ≥ 10.000). Due to the fast velocities a complete breakdown of the hydrodynamic boundary layer is possible. Dissolution processes change from transport controlled into surface controlled processes. The increasing rate of transporting reaction products results in a faster achievement of the solute saturation concentration within the fluid. Fig. is modified after Beier (2012).

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

7

Fig. 3. Influence of fluid movement (rotation speed, stirring rate) on dissolution rate RCa of calcite. All experiments are performed in batch reactors. A − data of Alkattan et al. (1998), experimental parameters: pH = 1; ionic strength <0.5 M; p = 1 bar, pCO2 <1 bar. Dashed lines are least square fits of the data at the individual temperatures. B − data of Plummer et al. (1978), experimental parameters: pH = 4; ionic strength <0.5 M; p = 1 bar, pCO2 <1 bar. Solid line is a least square fit of the data, dashed lines symbolize the area of RCa values.

of Ca)), expressed as dissolution rate RCa of the ion calcium in the carbonates, are plotted as function of temperature. The experimental temperatures range between 3 ◦ C and 300 ◦ C. The

majority of experiments is performed within the temperature range from 3 ◦ C to 150 ◦ C, at which most experiments were at standard (room) temperature (25 ◦ C) followed by temperatures of 50 ◦ C

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 8

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 4. Influence of temperature on carbonate dissolution rate RCa. Data consider all compared publications on carbonate experiments (aragonite, calcite, dolomite, magnesite, witherite) regardless of other influencing parameters like pH-value, ionic strength or total pressure. No obvious dependence of RCa on temperature seems to exist.

and 80 ◦ C. The large variation of dissolution rates at these three temperatures is obvious. For example the dissolution rates at 25 ◦ C vary by about 12 orders of magnitude. No profound dependence of dissolution rates on temperature seems to exist. Influences of other parameters like pH-value or ionic strength are not taken into account and these parameters vary from 1 to 11 (pH-value) and 0 to 5.5 M (ionic strength). A subset of the data with similar experimental conditions (batch reactor, pH-value: 3–5.5, ionic strength: <0.5 M, total

pressure (p): 1 bar (i.e. atmospheric pressure), CO2 partial pressure (pCO2 ): <1 bar, stirring rate: <427 rpm) is plotted in Fig. 5 and a slight temperature dependency of calcite dissolution rates with increasing dissolution rate at higher temperature can be observed. Although similar experimental conditions were chosen, RCa values of Schott et al. (1989) are consistently lower compared to the data of Alkattan et al. (1998), Alkattan et al. (2002), and Pokrovsky et al. (2005), while the temperature dependencies are identical. The reason of the offset remains unclear.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

9

Fig. 5. Influence of temperature on calcite dissolution rate RCa. Compared data are subset of data plotted in Fig. 4. All data are based on batch reactor experiments at low acid pH-values (3–5.5), low ionic strength (<0.5 M), p = 1 bar, pCO2 < 1 bar, and stirring rates <427 rpm. Dashed lines are least square fits of the data. The data of Alkattan et al. (1998), Alkattan et al. (2002), and Pokrovsky et al. (2005) are grouped together (log RCa = 0.0045·T [◦ C] − 4.241). RCa values of Schott et al. (1989; log RCa = 0.0051·T (◦ C) − 5.499) are lower by a constant factor compared to the other data, but temperature dependencies are almost identical.

Feldspar dissolution rates, i.e. dissolution rate RSi of the ion silicon in the feldspar group minerals, show a more distinct temperature dependency (Fig. 6). The dissolution rate RSi of feldspar (namely binary Ab (NaAlSi3 O8 ) − An (CaAl2 Si2 O8 ), binary Ab − Or (KAlSi3 O8 ) and ternary Ab-An-Or solid solutions) increases at higher temperatures, despite the fact of influences of other parameters (p, pCO2 , pH-value, ionic strength, etc.) on the dissolution rates and severe variations in dissolution rates at single temperatures. For example, the dissolution rate of feldspar group minerals varies by four orders of magnitude at 25 ◦ C. A more precise comparison of dissolution rates with similar experimental conditions (flow-through reactor, pH-value: 3–5.5, ionic strength: <0.1 M, p: <90 bar, pCO2 : <1 bar, fluid flow rate: <9 ml/min) verifies the observation of increasing dissolution rates with increasing temperature (Fig. 7). Dissolution rates of calcite and feldspar do not only show differently pronounced temperature dependences. Dissolution of, generalized speaking, carbonates is faster compared to feldspar dissolution with comparatively slow kinetics, i.e. the dissolution rates of carbonates are larger than of feldspar. Various authors attribute the temperature dependent dissolution rates of minerals − regardless of the used minerals − to changes of the mineral’s dissolution behavior from surface controlled processes to transport controlled processes (e.g., Casey and Sposito, 1992; Gautelier et al., 1999; Gledhill and Morse, 2006a; Finneran and Morse, 2009; Lasaga and Lüttge, 2004; Lund et al., 1973; Morse and Arvidson, 2002, or Raines and Dewers, 1997): whereas at low temperatures surface controlled processes dominate the mineral dissolution, at high temperatures transport controlled processes are assumed. For example, Gautelier et al. (1999) and Lund et al. (1973) describe transport controlled processes at temperatures of 100 ◦ C as dominant reaction behavior. In connection with that

Lund et al. (1973) note that increasing temperatures result also in increasing influences of the stirring rate on the dissolution rates, implying that at low temperatures the dissolution of minerals is limited by chemical reactions on the mineral surface, whereas at high temperatures the transport of reactants/reaction products is the limiting factor. Furthermore temperature not only influences dissolution mechanisms, but also influences the dissolution of complex minerals regarding their incongruent or congruent dissolution behavior. If minerals dissolve in stoichiometric proportions of the constituent elements, then dissolution is congruent. If some constituent elements dissolve preferentially to others, then dissolution of the mineral is incongruent. Experiments with kaolinite (Al2 Si2 O5 (OH)4 ; Carroll and Walther, 1990) as well as labradorite ((Ca,Na)Al(Si,Al)3 O8 ; Carroll and Knauss, 2005) show a change of incongruent dissolution behaviors at lower temperatures to congruent dissolution behaviors at higher temperatures. Although incongruent dissolution is not strongly pronounced for kaolinite, aluminum dissolution rates are slower than silicon dissolution rates at pH-values from 2 to 9 and a temperature of 25 ◦ C (Carroll and Walther, 1990). Incongruent dissolution behavior of labradorite can be observed at temperatures up to 60 ◦ C where an earlier and faster release of sodium, calcium, and aluminum takes place compared to the silicon release. At temperatures higher than 100 ◦ C congruent dissolution mechanisms of labradorite are reported by Carroll and Knauss (2005). However, Zhang et al. (2007) describe incongruent dissolution behavior regarding calcium and magnesium even at temperatures above 100 ◦ C for dolomite (CaMg(CO3 )2 ) dissolution experiments. Activation energy values of the dissolution reactions can be derived based on experiments with varying temperatures. Minerals with perfect surface areas exhibit the highest activation energies

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 10

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 6. Influence of temperature on feldspar dissolution rates RSi. Data consider all compared publications on feldspar experiments (binary Ab-An, binary Ab-Or and ternary Ab-An-Or solid solutions) regardless of other influencing parameters like pH-value, ionic strength or total pressure. A temperature dependence of RSi seems to exist.

and the activation energy decreases with increasing surface defects like steps, kinks or holes (Lasaga and Lüttge, 2004). Furthermore the complexity of the mineral composition influences the value of the activation energy. The more complex a mineral composition is, the larger is the activation energy, because more bonds have to be broken (Lasaga and Lüttge, 2004). The activation energy also hints to the various types of dissolution processes. Low activation energies of up to 10 kJ/mol (in some cases 20 kJ/mol) indicate transport

controlled processes. Activation energies higher than 10 kJ/mol (or 20 kJ/mol) represent surface controlled dissolution rates (as described by e.g., Carroll and Walther, 1990; Casey and Sposito, 1992; Finneran and Morse, 2009; Gledhill and Morse, 2006a). Various authors describe the influence of temperature in combination with other experimental parameters. The most common combination is temperature and pHvalue. Casey and Sposito (1992; dissolution of various silicate

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

11

Fig. 7. Influence of temperature on feldspar dissolution rate RSi. Compared data are subset of data plotted in Fig. 6. All data are based on flow-through reactor experiments at low acid pH-values (3–5.5), low ionic strength (< 0.1 M), total pressures up to 90 bar, pCO2 < 1 bar, and flow rates < 9 ml/min. Solid line represents least square fit of all data (log RSi = 0.0374·T [◦ C] − 11.97). Dashed lines symbolize the area of RSi values. The temperature dependence of RSi is more pronounced than of RCa . But feldspar dissolution exhibits slower kinetics than dissolution of carbonates (see Figs. 4 and 5 for comparison of dissolution rates).

minerals, i.e. andalusite (Al2 SiO5 ), kaolinite (Al2 Si2 O5 (OH)4 ), nepheline (Na3 KAl4 Si4 O16 ), olivine group minerals, quartz, Sranorthite ((Sr,Ca)Al2 Si2 O8 )) and Gautelier et al. (1999; dissolution of dolomite) mention this combination without clarifying the influence of temperature and pH-value on dissolution rates. Carroll and Walther (1990; dissolution of kaolinite) and Chen and Brantley (1997; dissolution of albite) describe an increasing influence of the pH-value on the dissolution process with increasing temperature. But other authors like Oelkers and Schott (1995; dissolution of anorthite) did not observed a change of pH-values dependency on dissolution rates with temperature. Zhang et al. (2007) proposed a correlation of temperature and mineral grain size on dissolution rates of dolomite. Experiments with dolomite (grain size fractions between 20 ␮m and 40 ␮m) and deionized water led to highest dissolution rates at 200 ◦ C. But dolomites with larger grain sizes (40–60 ␮m and even 60–80 ␮m) had maximum dissolution rates at lower temperatures (100–150 ◦ C) and the dissolution rates significantly decreased at temperatures higher than 150 ◦ C. A third combination of parameters is described by Finneran and Morse (2009). They observed a decreasing temperature influence on the dissolution rate of calcite with increasing ionic strength of the fluid used in the experiments. In conclusion, the experimental parameter temperature has a positive effect on the mineral dissolution process. With increasing temperature the dissolution process intensifies. Furthermore correlations exist between temperature and parameters like pH-value, grain size, and ionic strength.

2.2.2. Pressure A second severe influencing − and in terms of CO2 sequestration fundamental − parameter on dissolution rates is the CO2 partial pressure (pCO2 ).

The compared publications are not as beneficial regarding pCO2 and total pressure dependences on dissolution rates as they are for the influence of temperature on dissolution rates: in eight publications neither information of prevailed pCO2 nor of total pressure is provided, thus a detailed comparison is not possible. It is striking that this lack of data is generally observed for feldspar, olivine or pyroxene experiments, indicating that the influence of pressure is either not well investigated until now or negligible for these minerals as they do not contain carbonate ions. Twelve publications include both, total pressure and pCO2 information. The total pressure ranges between atmospheric pressure and 124 bar, whereas the pCO2 ranges between zero and 100 bar. Most experiments with carbonates (e.g., calcite, dolomite, magnesite or siderite (FeCO3 )) were performed at atmospheric pressure and associated pCO2 . Different observations in connection with a CO2 atmosphere and/or pCO2 are described in the literature, whereat a well-defined, distinctive pCO2 dependency is not reported in most of the cases. Increasing dissolution rates of carbonates with increasing pCO2 are described by, e.g., Pokrovsky et al. (2005), Gledhill and Morse (2006a), Finneran and Morse (2009), Golubev et al. (2009) or Pokrovsky et al. (2009). The largest influence of pCO2 on dissolution rates is observed for calcite. Gledhill and Morse (2006a) and Finneran and Morse (2009) describe a similar influence of pCO2 on calcite dissolution: a three- to four-fold (Finneran and Morse (2009)) and a five-fold (Gledhill and Morse (2006a)) increase in dissolution rates within a pCO2 range from 0.1 to 1 bar. Pokrovsky et al. (2005, 2009) even describes a five- to ten-fold as well as an eight-fold increase in dissolution rates within a pCO2 range from 1 to 50 bar and 1 to 25 bar, respectively. Calcite dissolution rates as function of pCO2 are compared in Fig. 8. Only studies performed in batch reactors at 25 ◦ C, acid pHvalues (3–5.5), low ionic strengths (<0.5 M), and stirring rates up to 2000 rpm are plotted. It is obvious that the change of dissolution rate with pCO2 is more pronounced at lower pCO2 than at higher

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 12

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 8. Influence of CO2 partial pressure on calcite dissolution rates RCa. All data are based on batch reactor experiments at room temperature (25 ◦ C), low acid pH-values (3–5.5), low ionic strengths (< 0.5 M), and stirring rates below 2000 rpm. Dissolution rate RCa is more sensitive to pCO2 at lower pCO2 than at higher pCO2 . Dotted line is an optical fit to all data to graphically illustrate the change of pCO2 influence on RCa .

pCO2 : the dissolution rate increases by two orders of magnitude within the first ten bars, but only by 0.5 orders of magnitude from 10 to 50 bar. Dissolution rates of dolomite and magnesite increase by factors of 2 to 3 (Pokrovsky et al., 2005, 2009) within a pCO2 range from 0 to 10 bar. In contrast to the three carbonates calcite, dolomite, and magnesite, the carbonate siderite exhibits no influence of pCO2 on its dissolution rate (Golubev et al., 2009). All authors describe changes of the pCO2 effect on dissolution rates of calcite, dolomite, and magnesite in connection with other parameters, at which varying pH-values influence the pCO2 effect on dissolution rates at most. In general, the addition of CO2 triggers a decrease in pH-value caused by the dissolution of CO2 in the fluid. At neutral to basic pH-values the addition of CO2 causes carbonate (CO3 2− ) and bicarbonate (HCO3 − ) formations (Gledhill and Morse, 2006a; Pokrovsky et al., 2009). As a consequence the change of the dissolution rate of carbonate minerals is not caused by pCO2 per se, but caused by the decreasing pH-value (acid pH: advanced dissolution) and the formation of carbonate complexes (neutral/basic pH: inhibited dissolution), respectively. Pokrovsky et al. (2005), Gledhill and Morse (2006a), Golubev et al. (2009), and Pokrovsky et al. (2009) state that the effect of pCO2 on the dissolution rates of calcite, dolomite, and magnesite is overprinted by the pH effect at low pH-values. Pokrovsky et al. (2005) recalculated dissolution rates of carbonate minerals for experiments with pH < 4 to disentangle the interrelation of pH and pCO2 . The influence of pCO2 on the recalculated dissolution rates lessened. The dissolution rates within the same pCO2 range change by a factor of three for the recalculated rates instead of a factor of eight without disentanglement of the interrelation of pH and pCO2 . Another parameter that influences the pCO2 effect is temperature. With increasing temperature the pCO2 effect on dissolution rates of calcite, dolomite, and magnesite decreases. Both Golubev

et al. (2009) and Pokrovsky et al. (2009) describe negligible pCO2 effects on dissolution rates of calcite, dolomite, and magnesite at 60 to 100 ◦ C and 80 to 100 ◦ C, respectively. In addition to the influence of pH-values and temperature on the pCO2 effect, Gledhill and Morse (2006a) and Finneran and Morse (2009) describe also changes of the pCO2 effect on dissolution rates of calcite, dolomite, and magnesite by varying ionic strengths as well as chemical compositions of the used fluid. At low pCO2 values the influence of the ionic strength and of the fluid’s chemical composition on the pCO2 effect of the dissolution rate is lesser pronounced than at high pCO2 values. Other minerals show a much smaller influence of pCO2 on their dissolution rates compared to carbonate minerals. The primary effect of pCO2 is either very small (e.g., forsteritic olivine ((Mg,Fe)2 SiO4 ): Hänchen et al., 2006) or even negligible in the range of analytical error (augite ((Ca,Fe)(Mg,Fe)Si2 O6 ): Brady and Carroll, 1994; anorthite: Berg and Banwart, 2000; labradorite ((Ca,Na)Al(Si,Al)3 O8 ): Carroll and Knauss, 2005; forsterite (Mg2 SiO4 ), diopside (MgCaSi2 O6 ), wollastonite (CaSiO3 ), hornblende ((Ca,Na,K)2-3 (Mg,Fe, Al)5 (OH,F)2 (Si,Al)2 Si6 O22 ): Golubev et al., 2005). Most of the authors state that the addition of CO2 triggers a pH-value decrease caused by the dissolution of CO2 in the fluid. Consequently, the pH-value has a far stronger influence on the dissolution rate than pCO2 on the dissolution rate of the minerals. Furthermore the authors describe an insufficient disentanglement of both effects during the experiment. A robust statement about the pCO2 influence on dissolution rates is therefore not possible. Influences of temperature or ionic strength on the dependence of pCO2 on dissolution rates of e.g., feldspar, olivine, and pyroxene minerals are not described in detail in the literature. To sum up, an overall well-defined, distinctive dissolution rate dependency regarding CO2 partial pressures cannot be observed. Only the carbonate mineral calcite and − to a lesser amount − the carbonate minerals dolomite and magnesite show a

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

dependence of pCO2 on the dissolution rates with a more pronounced influence of pCO2 at lower pCO2 values as pCO2 influences the pH-value (see 2.3.1.). Influences of pCO2 on dissolution rates of the carbonate mineral siderite as well as of silicates like e.g., feldspar group minerals, olivine, or pyroxenes are not reported in the literature. Experimental parameters like pH-value and ionic strength of the brine or temperature additionally affect the pCO2 influence of the dissolution rates of the carbonate minerals calcite, dolomite, and magnesite. They might be even the dominating influencing parameter on the dissolution rate as e.g., the pCO2 influence on the dissolution rate is overprinted by the pH-value influence on the dissolution rates as a consequence of CO2 dissolution processes in the fluid at low pH-values. The effect of pCO2 on the dissolution rates of carbonate minerals has therefore a second order influence compared to influences of the pH-value and/or the ionic strength. 2.3. Material properties The choice of appropriate starting materials (solids and fluids) is as important as choosing the proper experimental setup and experimental conditions. The solid material results in a high variability of dissolution behaviors and consequently kinetic data due to differences in the chemical composition and/or the crystal structure of the used minerals. Furthermore the treatment of solid starting materials is also important for dissolution experiments, i.e. how the materials were prepared: were they polished or mechanically fractured and subsequently washed with different fluids? Which grain size fractions were used in the experiments? Are the fractionated grains pressed into pellets or are the grains available as unconsolidated components within a fluid? However, there are no universal answers to the raised questions as they strongly depend on the aimed individual simulation scenarios. Another significant criterion is the selection of the sample, e.g., use of a monomineralic sample or a mineral paragenesis. Varying reaction agents enable to influence the reaction behavior of identical mineral phases in different ways. Not only solid materials, also the used fluid influences the dissolution process. In the literature a variety of fluids is described: deionized water, varying organic and inorganic acids and bases or brines of different composition and ionic strength, to name a few. For dissolution processes the bulk composition as well as the relative abundances of single components of the fluids are important and dictate the reaction progress in different ways. In addition buffered fluids exhibit other effects on the mineral dissolution than non-buffered fluids. Furthermore, the ratio of solid to fluid also influences the dissolution behavior of the studied minerals. The following paragraphs exemplary describe the influence of the fluid’s properties pH-value and ionic strength on dissolution behavior of carbonates, feldspar group minerals, and clay minerals. The chapter will be concluded by information regarding the alteration of reactive surfaces of these minerals and formation of secondary minerals while exposed to CO2 -bearing brines. 2.3.1. pH − value Numerous authors describe strong dependencies of mineral dissolution rates on the pH-value, whereat the degree of the influences varies between different mineral phases. Carbonates, for example, exhibit well defined linear dependencies on dissolution rate and pH-value. In all 30 compared publications dissolution rates decrease with increasing pH-values (Fig. 9). Even without disentanglement of other parameters such as temperature or ionic strength on mineral dissolution rates, the significant dependence of pH-values on dissolution rate is

13

obvious. The dissolution rate of calcite decreases by about 12 orders of magnitude within a pH range from −1.1 to 9.8 with most of the experiments performed at low acidic pH-values (3 ≥ pH ≤ 6). The subset of data (at T = 25 ◦ C; I = ≤5.6 M; p = 1 bar; pCO2 ≤1 bar; stirring rate <2000 rpm) indicates a non-linear dependence of dissolution rates on pH-value (Fig. 10). Within this subset the dissolution rates vary by six orders of magnitude between the lowest and the highest pH-value with a slight level off of the dependence at higher values. The decreasing dissolution rate with increasing pH-value is most likely caused by the proton induced dissolution of carbonates. During the dissolution process in a H2 O-CO2 -CaCO3 system multiple consecutive and parallel reactions take place, for example the dissociation of H2 O and CO2 in water or saline brine that causes acidification of the aqueous solution: H2 O + CO2 = H+ (aq) + HCO3 − (aq) Subsequently released protons of the aqueous solution react with carbonate ions CO3 2− (present in the reservoir rocks, e.g., bonded as calcium carbonate CaCO3 ) with the potential to remove some of the acidity of the aqueous solution H+ (aq) + CaCO3 = Ca2+ (aq) + HCO3 − (aq) to finally form calcium bicarbonate (Ca(HCO3 )2 ) in the aqueous solution, if there is sufficient carbonate present in the reservoir rocks CaCO3 + H2 O + CO2 = Ca(HCO3 )2 In addition, the recombination of released divalent cations like Ca, Mg, and Fe with dissolved CO2 will influence the pH-value of the aqueous solution again while forming secondary solid carbonate minerals, e.g., calcite (CaCO3 ) will form in the presence of Ca2+ : Ca2+ (aq) + HCO3 − (aq) = CaCO3 + H+ (aq) Agreement exists that the actual dissolution of carbonates is induced by the adsorption of hydrogen ions (H + ) on the mineral surface (Alkattan et al., 1998; Gautelier et al., 1999; Golubev et al., 2009; Sjöberg and Rickard, 1984a, 1984b). The effect of H+ adsorption depends on the pH-values and is most pronounced at pH-values between two and five (e.g., Golubev et al., 2009). But diversity exists on the mechanisms that control the dissolution: earlier studies (e.g., Sjöberg and Rickard, 1984a, 1984b; Berner and Morse, 1974) describe transport controlled mechanisms, where the diffusion of H+ through the diffusion boundary layer limits the overall dissolution process. In contrast authors like Gautelier et al. (1999) state that the dissolution processes of carbonates are governed by surface controlled mechanisms. The authors justify their statements with different rates of H+ diffusion and reaction progresses on the mineral surface. Whereas the diffusional transport of H+ increases proportional with increasing concentration, the adsorption of H+ on the mineral surface increases by a fractional power of the activity of H+ (aH+ ) and indicates surface controlled mechanisms. Morse and Arvidson (2002) partly confirmed this observation and stated that the diffusion coefficient of H+ is about three times faster than that of other diffusion species. As a consequence surface controlled mechanisms as well as transport controlled mechanisms might take place: the fast diffusion of H+ will be limited by H+ adsorptions on the mineral surface (surface controlled mechanisms) and the slower diffusion coefficients of other species (fluid reactants or released ions from the mineral surface) also limit the reaction progress (transport controlled mechanisms).

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 14

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 9. Influence of pH-value on carbonate dissolution rates RCa. Data consider all compared publications of experimental studies at a temperature range from 3 to 150 ◦ C and ionic strength values from 0 to 5.5 M. Dissolution rates RCa decrease as function of pH-values − even without disentanglement of other experimental parameters that influence RCa values.

Fig. 10. Influence of pH-value on calcite dissolution rates RCa. Compared data are subset of data plotted in Fig. 9. All data are based on batch reactor experiments at room temperature (25 ◦ C), ionic strengths from 0 to 5.6 M, p = 1 bar, pCO2 < 1 bar, and stirring rates <2000 rpm. A non-linear dependence of dissolution rates RCa on pH-value is obvious. Dotted line is an optical fit to all data to graphically illustrate the change of pH influence on RCa .

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

The influence of pH-value on dissolution rates abates at neutral to alkaline pH-environments (pH >5). Gautelier et al. (1999) link this observation with the predominant transport controlled mechanisms at neutral to alkaline pH-values. Kehew (2001) describes decreasing H+ activities as a plausible reason for slower dissolution kinetics at these pH-environments. However, Sjöberg and Rickard (1984a) distinguished between a transitional regime and neutral to alkaline regimes: at the transitional regime H+ -dependent reactions with not entirely transport controlled mechanisms become more and more negligible and change into H+ -independent reactions with mixed kinetics at neutral to alkaline regimes. The influence of other parameters like temperature or pCO2 on the pH-effect of the dissolution rate of carbonates is addressed to some extent in the literature. But these influences are easily overprinted by the more dominant pH-effect on the dissolution of carbonates. Nevertheless Pokrovsky et al. (2005), Pokrovsky et al. (2009) and Gautelier et al. (1999) described a more pronounced pHdependence of carbonate dissolution rates at low temperatures (25 ◦ C) than at high temperatures (150 ◦ C), whereat changing dissolution mechanisms with increasing temperature have to be kept in mind. The influence of pCO2 on the pH-effect is already addressed (see 2.2.2. pressure). In contrast to the dependence of calcite dissolution, dissolution rates of feldspar group and clay minerals show an u-shaped dependence on pH-value with a minimum of dissolution rates at neutral pH-values (see, e.g., albite: Brady and Walther, 1992; Chen and Brantley, 1997; Hellmann, 1994; kaolinite: Brady and Walther, 1992; Carroll and Walther, 1990). This pH-dependence is caused by adsorption processes of protons or hydroxyl ions on Al3+ or Si4+ lattice sites, whereat the dissolution process at acid pH-values (pH <5) is proton-influenced and the dissolution process at alkaline pH-values (pH >8) is hydroxyl-influenced. These changes of dissolution mechanisms cause the u-shaped dependence of the dissolution rate on the pH-value with a slowest dissolution rate at neutral pH-regimes and faster dissolution rates at acidic and alkaline pH-values due to congruent dissolution behavior with surface controlled mechanisms (Chen and Brantley, 1997; Carroll and Walther, 1990). The influence of other parameters (e.g., temperature) on the pH-effect of dissolution rates of feldspar group and clay minerals is controversially discussed in the literature. Whereas Chen and Brantley (1997) and Carroll and Walther (1990) observed no temperature effect on the pH-dependence, Brady and Walther (1992) and Hellmann (1994) predicted a temperature effect on the pHdependence. In summary most of mineral dissolution processes are significantly influenced by the pH-value of the fluid and are generally induced by first order reactions. The characteristics of pH-dependencies differ for various minerals and dissolution mechanisms are controversially discussed in the literature. Furthermore the influence of other parameters (e.g., temperature) on the pH-effect of mineral dissolution rates remains vague. 2.3.2. Ionic strength Understanding interactions between injected CO2 , fluids with high salt contents and therefore high ionic strengths as well as mineral phases is of great importance to shed light on chemical reactions at CO2 sequestration in deep saline aquifers. The influence of the chemical composition of fluids on dissolution rates is satisfactorily addressed in the literature. Both, synthetic and natural fluids, are used as starting fluids in experimental studies. Deionized water as the simplest and at the same time also most non-natural fluid is used by Eisenlohr et al. (1999) or Zhang et al. (2007), but most studies used more complex

15

fluids, especially brines. The brines differ in regard to the components themselves, the amount of each component within the overall fluid as well as the total salt content. Commonly used components are NaCl, KCl, MgCl2 or CaCl2 (see studies by e.g., Dove and Crerar, 1990; Finneran and Morse, 2009; Gautelier et al., 2007; Gledhill and Morse, 2006a; Golubev et al., 2006; Liteanu and Spiers, 2009; Pokrovsky et al., 2009). In some studies components like NaPO3 , NaHCO3 , AlCl3 or H4 SiO4 are also used (e.g., Alkattan et al., 2002; Carroll and Knauss, 2005; Gautelier et al., 2007; Gautier et al., 1994; Pokrovsky et al., 2009). Furthermore in many studies acids or bases are used as fluids to keep the solution at constant pHvalues or even pure acids and bases − for example HCl, KCl, NaOH or succinic acid − were used (e.g., Alkattan et al., 1998; Carroll and Walther, 1990; Chou et al., 1989; Cubillas et al., 2005). The ionic strengths of all fluids used in the literature range from 0 to 5.57 M. Fig. 11 illustrates the influence of ionic strength on dissolution rates of calcite. The range of dissolution rates varies by about 12 orders of magnitude, at which most experiments were accomplished at ionic strengths between 0 and 0.1 M. The large variability in dissolution rates of calcite even within the narrow range from 0 to 0.1 M might be caused by the varying compositions of the fluids. Whereas a fluid with an ionic strength of 0 M is generally described by deionized water, fluids with slightly higher ionic strength can include various acids and bases, salts or organic ligands. These varying compositions result in different dissolution rates. A subset of data at fixed temperature (T = 25 ◦ C), pressure (atmospheric pressure), and stirring rate (< 2000 rpm), and a narrow pH range (3 ≥ pH ≤ 5.5) still shows no systematic dependence of the ionic strength on the dissolution rate (Fig. 12). Even more, dissolution rates vary by ∼ two orders of magnitude at fixed ionic strength values of 0 to 0.1 and 0.7 M. This points to the fact that not only the ionic strength influences the mineral dissolution process, but also the single components and their individual quantities in the fluid. Although the compared dissolution rates of calcite vary by orders of magnitude at single ionic strength values (Fig. 11: all compared data; Fig. 12: subset of data) and no clearly pronounced decrease in dissolution rate with increasing ionic strengths is obvious, several authors describe an effect of the ionic strength on the dissolution processes of minerals: Finneran and Morse (2009), Gledhill and Morse (2006a), and Raines and Dewers (1997) link the dependence of mineral dissolution on the ionic strength with the water activity (aH2O ) in high concentrated fluids, as changes in dissolution rates are not proportional to changes of aH2O and dissolution rates are therefore also affected by different dissolved components. These authors observed (i) a three times faster decrease in dissolution rates compared to the decrease of aH2O and (ii) a more pronounced influence on the dissolution behavior by double charged ions (e.g., Mg2+ , Ca2+ ) compared to the influence of single charged ions (e.g., K+ , Na+ ). In addition, Finneran and Morse (2009) and Gledhill and Morse (2006a) correlate changes in mineral dissolution behavior to different hydration processes on the mineral surfaces. As mentioned earlier different flow regimes cause different dissolution mechanisms of minerals. These mechanisms are additionally influenced by the saturation state of the used fluid phase and therefore of the ionic strength of the fluid. Morse and Arvidson (2002) describe transport controlled dissolution behaviors in under-saturated fluids. In approximation to the chemical equilibrium and hence, with ongoing run duration, the transport controlled processes change to surface controlled dissolution processes. After reaching the chemical equilibrium solely surface controlled reactions take place. Other authors describe effects that depend on the ionic strength in combination with the CO2 partial pressure. For example, Gledhill and Morse (2006a) observed ionic strength independent calcite dissolution behavior for experiments performed at low CO2

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 16

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 11. Influence of ionic strength on calcite dissolution rates RCa. Data consider all compared publications on carbonate experiments regardless of other influencing parameters. The large range of RCa values at ionic strengths of 0 and 0.1 M is striking. For more details see text.

Fig. 12. Influence of ionic strength on calcite dissolution rates RCa. Compared data are subset of data plotted in Fig. 11. All data are based on batch reactor experiments at room temperature (25 ◦ C), low acid pH-values (3–5.5), p = 1 bar, pCO2 < 1 bar, and stirring rates <2000 rpm. No well-defined dependence of RCa on ionic strength of the brines exists, pointing towards more complex influences of the brine chemistry. For more details see text.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

partial pressures (pCO2 = 0.1 bar). Similar observations are made by Pokrovsky et al. (2005). They mention ionic strength independent calcite dissolution rates at CO2 partial pressures of 2 to 4 bar. Even though the mentioned observations describe various mineral dissolution behaviors as an effect of the ionic strength itself as well as an effect of combined experimental parameters (e.g., ionic strength, pCO2 ), the influence of the ionic strength on the dissolution rates is less pronounced when compared to other parameters like temperature, pH-value or pCO2 . 2.3.3. Reactive surface Information on reactive surfaces of minerals of interest, their changes during experiments and the impact of reactive surfaces on dissolution rates is only provided in some of the compared studies. In the following information regarding carbonates, feldspar group and clay minerals are summarized and − if available − illustrated by SEM images of our own experimental post-run samples. In part 2 of our series on dissolution kinetics of minerals relevant for potential injection sites, time-resolved changes of reactive surfaces of calcite during exposure to CO2 -bearing brines at elevated temperatures and pressures will be discussed in more detail (Holzheid, 2016 − same issue). 2.3.3.1. Carbonate minerals. Fig. 13A–C displays dissolution features of calcite and Fig. 13D–F of dolomite. All carbonate grains were exposed to CO2 -bearing brines at 150 ◦ C and ∼85 bar total pressure, while the run duration varied (1 day run duration: Fig. 13A–C; 2 days: Fig. 13D–E; 30 days: Fig. 13F). The initial rhombohedral shape of the carbonate grains maintains during the exposure to CO2 -bearing brines, including the curved, saddle-like faces of dolomite (calcite: Fig. 13A; dolomite: Fig. 13D). For calcite, dissolution features along all crystallographic directions develop, indicating dissolution of calcite not only on the mineral surface but also on kinks and steps on the mineral edges. The kinks and steps are dislocation sites where atoms preferentially detach from the surface due to their lower binding energy (Arvidson et al., 2003). Yadav et al. (2008) describe an increased Ca release of a factor of 2.25 on these dislocation sites compared to flat smooth surfaces. Considerable structural features of post-run calcite grains are needles formed on the overall mineral surface with all needles exhibiting the same orientation on single crystal faces (Fig. 13B; close-up: Fig. 13C). Pokrovsky et al. (2009) characterize these needles as remnants of the initial mineral surface: at the beginning of the dissolution process etch pits develop along the crystallographic directions. At progressing dissolution the etch pits enlarge and overlap with each other. In consequence of that, needles remain as remnants of old etch pit walls which have overlapped with other etch pits. During ongoing dissolution experiments the ‘etch pit-dominated’ mineral surface morphology converts into a ‘needle-dominated’ mineral surface with significantly increased reactive surface area. With ongoing run duration structural changes on the mineral surfaces continue and the former pointy needles change to stub point needles (see Fig. 13C for the entire variety of needle shapes). The development of the needles therefore represents different dissolution iterations with stub point needles characterizing an almost completely dissolved calcite layer. Pointy needles in close vicinity to stub point needles might just reflect remnants of still less advanced pointy needles. In comparison to the observed surface changes of calcite grains, dolomite exhibits a lesser pronounced alteration of its surface. Irregular formed etch pits (Fig. 13E) develop to more stepped dissolution features along the crystallographic directions (Fig. 13F). Needles do not form. This is most likely caused by the atomic structure of dolomite crystals, in which Ca2+ and Mg2+ ions are arranged in alternate layers contrary to the more layered atomic structure

17

of Ca2+ ions in calcite (in analogy to the framework structure of halite). Due to the dense atomic packing of dolomite a destruction of framework bonds is hampered and complete dissolution of these alternate layers does not happen as assumed for calcite. Rather, Ca2+ and Mg2+ ions are released from the mineral surface resulting in more irregular shaped surfaces of dolomite compared to calcite. Regarding the carbonate mineral siderite, Golubev et al. (2009) describe only minor changes of the post-run mineral’s surface. However, it has to be mentioned that changes in reactive surface are not observed in all studies on dissolution of carbonate minerals. For example, Gledhill and Morse (2006b) observed no significant changes in reactive surface of calcite during similar run condition as of Pokrovsky et al. (2005). Both, Pokrovsky et al. (2005) and Cubillas et al. (2005) reported that dissolution of calcite and aragonite in HCl solutions leads to 30% and 100% increased BET surface areas compared to the initial BET surfaces. These changes of reactive surfaces directly influence the dissolution rate (e.g., Zhang et al., 2007) and doubling of the reactive surface of calcite increases the dissolution rate by ∼10% (Pokrovsky et al., 2005). However, the interlinkage between variable dissolution kinetics during progressing dissolution of carbonate minerals and ever-changing reactive surface areas are still a subject of debate (e.g., Truesdale, 2015; Arvidson et al., 2015).

2.3.3.2. Feldspar group and clay minerals. First of all the observed dissolution rates of anorthite and orthoclase exposed to CO2 bearing brines as determined in experiments of the accomplished study (see also Beier, 2012) will be described. This is followed by a comparison with literature findings. Fig. 14A–C and D–F displays SEM images of post-run anorthite and orthoclase. The initial irregular shape with even as well as conchoidal/uneven surfaces of the starting minerals (anorthite: Fig. 14A; orthoclase: Fig. 14D) is maintained even after long run durations (30 days) and elevated temperatures and total pressures (150 ◦ C, ∼85 bar). All post-run anorthite grains are coated with secondary mineral phases (Fig. 14A,B). Up to 40% of the entire surface is coated. The secondary minerals’ shape is platy/planar and in some cases radiating rose-like aggregates of grain sizes ≤1 ␮m formed (Fig. 14C). This habitus is typical for many phyllosilicates like clay minerals or mica. Hence, the minerals were characterized as phyllosilicates like illite, smectite or kaolinite. This is in agreement with literature studies. Hangx and Spiers (2009) describe their post-run plagioclases as minerals coated with precipitations of fine, platy, lath-like, honeycomb-like, and/or rose-like minerals with sizes ≤1 ␮m or with agglomerations of hexagonal shaped plates. In addition Hangx and Spiers (2009) observed occasional etch pits. Etch pits are also described by Fischer et al. (2010), Hodson (2006), and Oelkers and Schott (1995), whereat these authors describe no secondary mineral phases. Chen and Brantley (1997) were able to correlate surface changes with pH-values with the largest and fastest changes of albite surfaces at low pH-values. Secondary minerals on post-run orthoclase are less frequent (Fig. 14D) and have star-like habitus (Fig. 14E). Precipitation of secondary minerals on post-run orthoclase is also not well documented in the literature. Although Blake and Walter (1999) observed Al-containing precipitations and characterized these mineral phases as smectites, Gautier et al. (1994) report no precipitates at all. The much more significant change on the orthoclase surfaces is the extensive development of etch pits and steps. The etch pits exhibit irregular shapes and are cord-like placed (Fig. 14F). These phenomena are also described by Gautier et al. (1994), who combine the location of etch pits with the presence of feldspar exsolution features. Furthermore Harouiya and Oelkers (2004) describe etch pits as indicator for heterogeneous dissolution processes.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 18

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Fig. 13. SEM images of post-run calcite (A-C) and dolomite (D-F) grains. A-C: calcite grains after 1-day run duration and at 150 ◦ C, ∼85 bar total pressure; B and C are close-up images of A and B, respectively D- F: dolomite grains after 2-days (D, E) and 30-days (F) run duration and at 150 ◦ C, ∼85 bar total pressure; E is a close-up image of D The initial rhombohedral shape of the calcite starting grain is preserved (A) and dissolution features developed with pointy needles on the mineral surface (B). At progressing dissolution the pointy needles change to more stub point needles (C). The initial shape − including the curved, saddle-like faces − of dolomite starting grain is preserved (D). Needles do not form, instead etch pits (E) develop to more stepped dissolution features along the crystallographic directions (F) with longer exposure time to CO2 -bearing brines. SEM images: SEM laboratory, Institute of Geosciences, CAU Kiel

Fig. 14. SEM images of post-run anorthite (A-C) and orthoclase (D-F) after 30-days run duration, at 150 ◦ C, and ∼85 bar total pressure. The initial shape of the feldspar grains is preserved (partly visible at A and D). Surfaces of anorthite as well as to a lesser extent orthoclase are coated with secondary mineral phases. The secondary minerals’ shape is platy/planar and in some cases radiating rose-like (B and close-up image C for anorthite) or the secondary minerals have a star-like habitus (E for orthoclase). In addition, etch pits and steps developed on orthoclase surfaces (F). SEM images: SEM laboratory, Institute of Geosciences, CAU Kiel

In general, surface changes of anorthite are more pronounced and extensive than those of orthoclase as also remarked by Hangx and Spiers (2009).

Only one literature study comments on possible changes of reactive surfaces of clay minerals: Carroll and Walther (1990) report no observed changes of the reactive surface of kaolinite while exposed to CO2 -bearing brines.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

3. Concluding remarks placed in the context of CO2 sequestration Dissolution rates of minerals that are exposed to CO2 -bearing brines at elevated temperature and pressure are not solely controlled by a single parameter and dissolution experiments include a lot of variables. Hence, no general statement about the influence of one specific parameter on dissolution rates can be made as long as the influence of others parameters is not disentangled or all other parameters are not fixed. It is therefore important to describe and disentangle the influences of experimental parameters like temperature, total pressure as well as CO2 partial pressure, and of material properties like pH-value and ionic strength of the brine as well as the extent of alteration of minerals’ reactive surfaces and formation of secondary minerals on dissolution behavior of minerals. In addition, attention has to be paid to the differences between the various experimental setups and their impact on the dissolution behavior of the minerals. Experimental setups have to be adapted to the desired scientific questions and have to be able to simulate proceeding processes at CO2 sequestration sites as precise as possible. With regard to possible numerical modeling scenarios it has to be noted that increasing temperatures positively influence the dissolution process. Predominant dissolution mechanisms also change with temperature: at low temperatures incongruent and surface controlled mechanisms dominate, and at high temperatures congruent and transport controlled mechanisms prevail. Casey and Sposito (1992) already described incorrect results of modeled activation energies related to neglected influences of varying dissolution mechanisms. In addition, Raines and Dewers (1997) explicitly pointed out that numerical models, which were developed for low temperatures, cannot be easily extrapolated to higher temperatures with the help of the Arrhenius relation because of changing dissolution mechanisms with increasing temperatures. The effect of CO2 partial pressure on the dissolution rates of carbonate minerals has a second order influence compared to the pH effect and/or the ionic strength effect of the CO2 -bearing brine. Although characteristics of pH-dependencies on mineral dissolution behavior differ for various minerals and dissolution mechanisms are controversially discussed in the literature, agreement exists that mineral dissolution processes are significantly influenced by the pH-value of the fluid and are generally induced by first order reactions. The pH-effect in combination with the CO2 partial pressure plays a superior role regarding reactions related to CO2 sequestration sites: during the CO2 -injection the pCO2 increases close to the injection point and therefore results in a decreasing pH-value within the reservoir. In the case of carbonate bonded sandstones the brine is buffered by the dissolution of the carbonatic components at the beginning of the injection, but the pH-value starts to decrease after complete dissolution of the carbonatic components. Consequently the decrease in pH-value triggers additional dissolution processes of the other non-carbonate components (Knauss et al., 2005). At non-carbonate bonded sandstones the buffering impact is missing and the dissolution processes are severe already at the beginning of CO2 -injection action. The chemical composition of the prevailing formation water has to be taken into account, but is not the dominating factor which influences chemical processes within the respective CO2 storing reservoir when compared to other parameters like temperature, pH-value or even pCO2 . Nevertheless, the influence of the formation water composition and, hence, the ionic strength should not to be underestimated because the high variability of formation water compositions − in combination with the injected gas-phase

19

mixture − results in a large number of chemical reactions which have to be taken into account. Changes of reactive surfaces and habitus of carbonates during the cause of interaction with CO2 -bearing brines at elevated temperature and pressure are limited to dissolution features like etch pits and needles. Severe changes of reactive surfaces are reported in the literature (see also the most recent publication by Black et al., 2015). These changes of reactive surfaces directly influence the dissolution rate (e.g., Zhang et al., 2007). Doubling of the reactive surface of calcite increases the dissolution rate by ∼10% (Pokrovsky et al., 2005). Changes of feldspar surfaces are characterized by both, dissolution features like etch pits and dissolution steps as well as formation of secondary minerals, namely the formation of phyllosilicates like illite, smectite or kaolinite. Hence, computational simulations of mineral reactions at potential CO2 storage sites have to include both, the changes of reactive surfaces and therefore kinetics of mineral dissolution and the influence of secondary minerals on the interaction of the minerals with CO2 -enriched brines. Acknowledgments The literature review up to 2010 is part of the PhD thesis of Katja Beier. As Katja Beier moved on, she unfortunately refrained from being co-author. Some sentences of the present publication are literal citations from Beier (2012) but not all are marked individually. The study is part of the CO2 -MoPa project (Modeling and Parameterization of CO2 storage in deep saline formations for dimension and risk analyses). The overall aim of CO2 -MoPa is to investigate dimension and risk analyses for subterrestrial CO2 sequestration on virtual scenarios. CO2 -MoPa is funded by the German Federal Ministry of Education and Research (BMBF), EnBW Energie BadenWürttemberg AG, E.ON Energie AG, E.ON Ruhrgas AG, RWE Dea AG, Vattenfall Europe Technology Research GmbH, Wintershall Holding AG and Stadtwerke Kiel AG within the framework of the special program GEOTECHNOLOGIEN. The author is grateful to all actively participating scientists of the CO2 -MoPa project. Special thanks go to Karin Kissling, Karen Bremer, Petra Kluge, Bredan Ledwig, and Ute Schuldt (all Institute of Geosciences, Kiel) for help and assistance in performing the experiments, analyses, and SEM images. The scientific input from Wolf-Achim Kahl (formerly Institute of Geosciences, Kiel, now at University Bremen, Germany) is highly appreciated. References Alkattan, M., Oelkers, E.H., Dandurand, J.-L., Schott, J., 1998. An experimental study of calcite and limestone dissolution rates as function of pH from −1 to 3 and temperature from 25 to 80 ◦ C. Chem. Geol. 151, 199–214. Alkattan, M., Oelkers, E.H., Dandurand, J.-L., Schott, J., 2002. An experimental study of calcite dissolution rates at acidic conditions and 25 ◦ C in the presence of NaPO3 and MgCl2. Chem. Geol. 190, 291–302. Arvidson, R.S., Ertan, I.E., Amonette, J.E., Lüttge, A., 2003. Variation in calcite dissolution rates: a fundamental problem? Geochim. Cosmochim. Acta 67 (9), 1623–1634. Arvidson, R.S., Fischer, C., Lüttge, A., 2015. Calcite dissolution kinetics − a comment upon “Evidence and potential implications of exponential tails to concentration versus time plots for the batch dissolution of calcite by V.W. Truesdale”. Aquat. Geochem. 21, 415–422. Bauer, S., Class, H., Ebert, M., Feeser, V., Götze, H., Holzheid, A., Kolditz, O., Rosenbaum, S., Rabbel, W., Schäfer, D., Dahmke, A., 2012. Modeling: parameterization and evaluation of monitoring methods for CO2 storage in deep saline formations: the CO2 -MoPa project. Environ. Earth Sci. 67, 351–367. Beier, K., 2012. CO2 -sequestration on Laboratory Scale: Geochemical Interactions Between Injected CO2 , Saline Fluid Phases, and Potential Reservoir Materials PhD Thesis. University Kiel, Germany, 180 p. Berg, A., Banwart, S.A., 2000. Carbon dioxide mediated dissolution of Ca-feldspar: implications for silicate weathering. Chem. Geol. 163, 25–42. Berner, R.A., Morse, J.W., 1974. Dissolution kinetics of calcium carbonate in sea water IV. Theory of calcite dissolution. Am. J. Sci. 274, 108–134.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21 20

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Black, J.R., Carroll, S.A., Haese, R.R., 2015. Rates of mineral dissolution under CO2 storage conditions. Chem. Geol. 399, 134–144. Blake, R.E., Walter, L.M., 1999. Kinetics of feldspar and quartz dissolution at 70–80 ◦ C and near-neutral pH: Effects of organic acids and NaCl. Geochim. Cosmochim. Acta 63 (13/14), 2043–2059. Brady, P.V., Carroll, S.A., 1994. Direct effects of CO2 and temperature on silicate weathering: possible implications for climate control. Geochim. Cosmochim. Acta 58 (8), 1853–1856. Brady, P.V., Walther, J.V., 1992. Surface chemistry and silicate dissolution rates at elevated temperatures. Am. J. Sci. 292, 639–658. Canadian International Development Agency, 2007. Development of china’s coalbed methane technology/CO2 sequestration project. Technical report, Canadian International Development Agency, Alberta Research Council, Final project report, Project No. A-030841. Carroll, S.A., Knauss, K.G., 2005. Dependence of labradorite dissolution kinetics on CO2 (aq) Al(aq), and temperature. Chem. Geol. 217, 213–225. Carroll, S.A., Walther, J.V., 1990. Kaolinite dissolution at 25 ◦ C 60 ◦ C, and 80 ◦ C. Am. J. Sci. 290, 797–810. Casey, W.H., Sposito, G., 1992. On the temperature dependence of mineral dissolution rates. Geochim. Cosmochim. Acta 56, 3825–3830. Chen, Y., Brantley, S.L., 1997. Temperature- and pH-dependence of albite dissolution rate at acid pH. Chem. Geol. 135, 275–290. Chou, L., Garrels, R.M., Wollast, R., 1989. Comparative study of the kinetics and mechanisms of dissolution of carbonate minerals. Chem. Geol. 78, 269–282. Crawshaw, J.P., Boek, E.S., 2013. Multi-scale imaging and simulation of structure: flow and reactive transport for CO2 storage and EOR in Carbonate reservoirs. Rev. Mineral. Geochem. 77, 431–458. Cubillas, P., Köhler, S., Prieto, M., Chaïrat, C., Oelkers, E.H., 2005. Experimental determination of dissolution rates of calcite aragonite, and bivalves. Chem. Geol. 216, 59–77. Domenico, P.A., Schwartz, F.W., 1990. Physical and Chemical Hydrogeology. John Wiley & Sons, New York, 824 p. Dove, P.M., Crerar, D.A., 1990. Kinetics of quartz dissolution in electrolyte solutions using a hydrothermal mixed flow reactor. Geochim. Cosmochim. Acta 54, 955–969. Dreybrodt, W., Buhmann, D., 1991. A mass transfer model for dissolution and precipitation of calcite from solutions in turbulent motion. Chem. Geol. 90 (1–2), 107–122. Druckenmiller, M.L., Maroto-Valer, M.M., Hill, M., 2006. Investigation of carbon sequestration via induced calcite formation in natural gas well brine. Energy Fuels 20, 172–179. Eisenlohr, L., Meteva, K., Gabrovˇsek, F., Dreybrodt, W., 1999. The inhibition action of intrinsic impurities in natural calcium carbonate minerals to their dissolution kinetics in aqueous H2 O − CO2 solutions. Geochim. Cosmochim. Acta 63 (7/8), 989–1002. Finneran, D.W., Morse, J.W., 2009. Calcite dissolution kinetics in saline waters. Chem. Geol. 268, 137–146. the CO2 SINK GroupFischer, S., Liebscher, A., Wandrey, M., 2010. CO2 -brine-rock interaction − first results of long-term exposure experiments at in situ P-T conditions of the Ketzin CO2 reservoir. Chemie der Erde 70, 155–164. Gautelier, M., Oelkers, E.H., Schott, J., 1999. An experimental study of dolomite dissolution rates as a function of pH from −0: 5 to 5 and temperature from 25 to 80 ◦ C. Chem. Geol. 157, 13–26. Gautelier, M., Schott, J., Oelkers, E.H., 2007. An experimental study of dolomite dissolution rates at 80 ◦ C as a function of chemical affinity and solution composition. Chem. Geol. 242, 509–517. Gautier, J.-M., Oelkers, E.H., Schott, J., 1994. Experimental study of K-feldspar dissolution rates as a function of chemical affinity at 150 ◦ C and pH 9. Geochim. Cosmochim. Acta 58 (21), 4549–4560. Gautier, J.-M., Oelkers, E.H., Schott, J., 2000. Are quartz dissolution rates proportional to B.E.T. surface areas? Geochim. Cosmochim. Acta 65 (7), 1059–1070. Gledhill, D.K., Morse, J.W., 2004. Dissolution kinetics of calcite in NaCl-CaCl2 -MgCl2 brines at 25 ◦ C and 1 bar pCO2 . Aquat. Geochem. 10, 171–190. Gledhill, D.K., Morse, J.W., 2006a. Calcite dissolution kinetics in Na-Ca-Mg-Cl brines. Geochim. Cosmochim. Acta 70, 5802–5813. Gledhill, D.K., Morse, J.W., 2006b. Calcite dissolution kinetics in Na-Ca-Mg-Cl brines. Chem. Geol. 233, 249–256. Golubev, S.V., Pokrovsky, O.S., Schott, J., 2005. Experimental determination of the effect of dissolved CO2 on the dissolution kinetics of Mg and Ca silicates at 25 ◦ C. Chem. Geol. 217, 227–238. Golubev, S.V., Bauer, A., Pokrovsky, O.S., 2006. Effect of pH and organic ligands on the kinetics of smectite dissolution at 25 ◦ C. Geochim. Cosmochim. Acta 70, 4436–4451. Golubev, S.V., Bénézeth, P., Schott, J., Dandurand, J.L., Castillo, A., 2009. Siderite dissolution kinetics in acidic aqueous solutions from 25 ◦ C to 100 ◦ C and 0 to 50 atm pCO2 . Chem. Geol. 265, 13–19. Hänchen, M., Prigiobbe, V., Storti, G., Seward, T.M., Mazzotti, M., 2006. Dissolution kinetics of fosteritic olivine at 90–150 ◦ C including effects of the presence of CO2 . Geochim. Cosmochim. Acta 70, 4403–4416. Hamm, L.M., Bourg, I.C., Wallace, A.F., Rotenberg, B., 2013. Molecular simulation of CO2 - and CO3 -brine-mineral systems. Rev. Mineral. Geochem. 77, 189–228. Hangx, S.J.T., Spiers, C.J., 2009. Reaction of plagioclase feldspars with CO2 under hydrothermal conditions. Chem. Geol. 265, 88–98.

Harouiya, N., Oelkers, E.H., 2004. An experimental study of the effect of aqueous fluoride on quartz and alkali-feldspar dissolution rates. Chem. Geol. 205, 155–167. Hellevang, H., VanPham, T.H., Aagaard, P., 2013. Kinetic modelling of CO2 -water-rock interactions. Int. J. Greenh. Gas Control 15, 3–15. Hellmann, R., 1994. The albite-water system part I. The kinetics of dissolution as a function of pH at 100, 200, and 300 ◦ C. Geochim. Cosmochim. Acta 58, 595–611. Hodson, M.E., 2006. Does reactive surface area depend on the grain size? Results from pH 3, 25 ◦ C far-from-equilibrium flow-through dissolution experiments on anorthite and biotite. Geochim. Cosmochim. Acta 70, 1655–1667. Holzheid, A., 2016. Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 2: dissolution and alteration of carbonates and feldspars in CO2 -bearing brines, submitted to Chemie der Erde − Geochemistry on 31/03/2016. doi: 10.1016/j.chemer.2016.09.008. Huq, F., Haderlein, S.B., Cirpka, O.A., Nowak, M., Blum, P., Grathwohl, P., 2015. Flow-through experiments on water-rock interactions in a sandstone caused by CO2 -injection at pressures and temperatures mimicking reservoir conditions. Appl. Geochem. 58, 136–146. Kang, Q., Lichtner, P.C., Viswanathan, H.S., Abdel-Fattah, A.I., 2010. Pore scale modeling of reactive transport involved in geologic CO2 sequestration. Transp. Porous Media 82, 197–213. Kehew, A.E., 2001. Applied Chemical Hydrogeology. Prentice-Hall, Inc., New Jersey, 368 p. Knauss, K.G., Johnson, J.W., Steefel, C.I., 2005. Evaluation of the impact of CO2 co-contaminant gas, aqueous fluid and reservoir rock interactions on the geologic sequestration of CO2 . Chem. Geol. 217, 339–350. Lasaga, A.C., Lüttge, A., 2004. Mineralogical approaches to fundamental crystal dissolution kinetics. Am. Miner. 89, 527–540. Liteanu, E., Spiers, C.J., 2009. Influence of pore fluid salt content on compaction creep of calcite aggregates in the presence of supercritical CO2 . Chem. Geol. 265, 134–147. Lund, K., Fogler, H.S., McCune, C.C., 1973. Acidization − II: The dissolution of dolomite in hydrochloric acid. Chem. Eng. Sci. 28, 691–700. Marty, N.C.M., Claret, F., Lassin, A., Tremosa, J., Blanc, P., Madé, B., Giffaut, E., Cochepin, B., Tournassat, C., 2015. A database of dissolution and precipitation rates for clay-rocks minerals. Appl. Geochem. 55, 108–118. Michael, K., Golab, A., Shulakova, V., Ennis-King, J., Allinson, G., Sharma, S., Aiken, T., 2010. Geological storage of CO2 in saline aquifers − A review of the experience from existing storage operations. Int. J. Greenh. Gas Control 4, 659–667. Morse, J.W., Arvidson, R.S., 2002. The dissolution kinetics of major sedimentary carbonate minerals. Earth Sci. Rev. 58, 51–84. Oelkers, E.H., Schott, J., 1995. Experimental study of anorthite dissolution and the relative mechanism of feldspar hydrolysis. Geochim. Cosmochim. Acta 59 (24), 5039–5053. Oelkers, E.H., Schott, J., Devidal, J.-L., 1994. The effect of aluminum, pH, and chemical affinity on the rates of aluminosilicate dissolution reactions. Geochim. Cosmochim. Acta 58 (9), 2011–2024. J.L. Palandri, Y.K. Kharaka, 2004. A compilation of rate parameters of water-mineral interaction kinetics for application to geochemical modeling. Technical report, U.S. Geological survey, Open File Report 2004–1068. Plummer, L.N., Busenberg, E., 1982. The solubilities of calcite, aragonite and vaterite in CO2 -H2 O solutions between 0 and 90 ◦ C, and an evaluation of the aqueous model for the system CaCO3 -CO2 -H2 O. Geochim. Cosmochim. Acta 46, 1011–1040. Plummer, L.N., Wigley, T.M.L., Parkhurst, D.L., 1978. The kinetics of calcite dissolution in CO2 -water systems at 5 ◦ to 60 ◦ C and 0.0 to 1.0 atm CO2 . Am. J. Sci. 278, 179–216. Pokrovsky, O.S., Golubev, S.V., Schott, J., 2005. Dissolution kinetics of calcite, dolomite and magnesite at 25 ◦ C and 0 to 50 atm pCO2 . Chem. Geol. 217, 239–255. Pokrovsky, O.S., Golubev, S.V., Jordan, G., 2009. Effect of organic and inorganic ligands on calcite and magnesite dissolution rates at 60 ◦ C and 30 atm pCO2 . Chem. Geol. 265, 33–43. Power, I.M., Harrison, A.L., Dipple, G.M., Wilson, S.A., Keleman, P.B., Hitch, M., Southam, G., 2013. Carbon mineralization: from natural analogue to engineered systems. Rev. Mineral. Geochem. 77, 305–360. Raines, M., Dewers, T., 1997. Mixed kinetics control of fluid rock interaction in reservoir production scenarios. J. Petrol. Sci. Eng. 17 (1–2), 139–155. Schott, J., Brantley, S.L., Crerar, D., Guy, C., Borcsik, M., Willaime, C., 1989. Dissolution kinetics of strained calcite. Geochim. Cosmochim. Acta 53, 373–382. Sjöberg, E.L., Rickard, D.T., 1984a. Calcite dissolution kinetics: surface speciation and the origin of the variable pH dependence. Chem. Geol. 42, 119–136. Sjöberg, E.L., Rickard, D.T., 1984b. Temperature dependence of calcite dissolution kinetics between 1 and 62 ◦ C at pH 2.7 to 8.4 in aqueous solutions. Geochim. Cosmochim. Acta 48, 485–493. Sorai, M., Ohsumi, T., Ishikawa, M., Tsukamoto, K., 2007. Feldspar dissolution rates measured using phase-shift interferometry: implications to CO2 underground sequestration. Appl. Geochem. 22, 2795–2809. Steefel, C.I., Molins, S., Trebotich, D., 2013. Pore scale processes associated with subsurface CO2 -injection and sequestration. Rev. Mineral. Geochem. 77, 259–303. Stillings, L.L., Brantley, S.L., 1995. Feldspar dissolution at 25 ◦ C and pH 3: Reaction stoichiometry and the effect of cations. Geochim. Cosmochim. Acta 59 (8), 1483–14965.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007

G Model CHEMER-25406; No. of Pages 21

ARTICLE IN PRESS A. Holzheid / Chemie der Erde xxx (2016) xxx–xxx

Truesdale, V.W., 2015. Evidence and potential implications of exponential tails to concentration versus time plots for the batch dissolution of calcite. Aquat. Geochem. 21, 365–396. Tutolo, B.M., Luhmann, A.J., Kong, X.-Z., Saar, M.O., Seyfried, W.E., 2015. CO2 sequestration in feldspar-rich sandstone: coupled evolution of fluid chemistry mineral reaction rates, and hydrogeochemical properties. Geochim. Cosmochim. Acta 160, 132–154.

21

Yadav, S.K., Chakrapani, G.J., Gupta, M.K., 2008. An experimental study of dissolution kinetics of calcite dolomite, leucogranite and gneiss in buffered solutions at temperature 25 and 5 ◦ C. Environ. Geol. 53, 1683–1694. Zhang, R., Hu, S., Zhang, X., Yu, W., 2007. Dissolution kinetics of dolomite in water at elevated temperatures. Aquat. Geochem. 13, 309–338.

Please cite this article in press as: Holzheid, A., Dissolution kinetics of selected natural minerals relevant to potential CO2 -injection sites − Part 1: A review. Chemie Erde - Geochemistry (2016), http://dx.doi.org/10.1016/j.chemer.2016.09.007