Thermodynamic Modelling of Dolomite Behavior in Aqueous Media

Hindawi Publishing Corporation Journal of Thermodynamics Volume 2012, Article ID 723052, 12 pages doi:10.1155/2012/723052 Research Article Thermodynamic Modelling of Dolomite Behavior in Aqueous Media Tadeusz Michałowski1and Agustin G. Asuero2 1Faculty of Engineering and Chemical Technology, Technical University of Cracow, 31-155 Krak´ ow, Poland 2Department of Analytical Chemistry, The University of Seville, 41012 Seville, Spain Correspondence should be addressed to Tadeusz Michałowski, [email protected] Received 20 August 2011; Revised 19 October 2011; Accepted 27 October 2011 Academic Editor: Jaime Wisniak Copyright © 2012 T. Michałowski and A. G. Asuero. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. The compact thermodynamic approach to the systems containing calcium, magnesium, and carbonate species is referred to dissolution of dolomite, as an example of nonequilibrium ternary salt when introduced into aqueous media. The study of dolomite is based on all attainable physicochemical knowledge, involved in expressions for equilibrium constants, where the species of the system are interrelated. The species are also involved in charge and concentration balances, considered as constraints put on a closed system, separated from the environment by diathermal walls. The inferences are gained from calculations performed with use of an iterative computer program. The simulated quasistatic processes occurred under isothermal conditions, started at a preassumed pH0value of the solution where dolomite was introduced, and are usually involved with formation of other solid phases. None simplifying assumptions in the calculations were made. 1. Introduction Dolomite (CaMg(CO3)2, abbr. pr1) is perceived as an unusual, metastable mineral [1,2], and its behavior is considered as one of the most exciting topics in geology [3]. Its chemical properties should be put in context with other, most important carbonate minerals: calcite (CaCO3,pr2) and magnesite (MgCO3,pr3)[4].Thetrigonalstructureof calcite is composed of alternate layers of calcium and carbonate ions. The crystal structure of magnesite is the same as one for calcite, and then magnesite properties are similar to those of calcite. However, a significant difference in Pauling’s ionic radii: 99 pm (for Ca+2), 65 pm (for Mg+2), [5] causes an incompatibility of the cations in the same layer of dolomite structure. Crystal lattice of ideal dolomite (M1= 184.4g/mol,ρ1=2.899 g/cm3) consists of alternating octahedral layers of Ca+2 and Mg+2 ions, separated by layers of CO3−2ions [6]. Complete ordering is energetically favourable at lower temperatures. Presumably, it is the principal crystallographic constraint securing nearly ideal (Ca : Mg = 1 : 1) dolomite stoichiometry [6]. Ideal dolomite, CaMg(CO3)2, is a particular case (x= 0) of the compound Ca1−xMg1+x(CO3)2[7–10], considered as a result of partial disarrangement of Ca+2 and Mg+2 in structure of real dolomite. This disarrangement is evidenced by weak or diffusespectrallinesinthereflectedX-ray diffraction patterns [11]. The composition of sedimentary dolomites is not exactly stoichiometric; it ranges from Ca1.16Mg0.84(CO3)2to Ca0.96Mg1.04(CO3)2(i.e., −0.16 < x<0.04) [1]. The mineral Ca0.5Mg1.5(CO3)2(x=0.5) is named huntite. Sedimentary dolomites with x>0are named calcium dolomites, and the ones with x<0are named magnesian dolomites. The carbonate dominated by calcite (ρ2=2.71 g/cm3) is called limestone. Other carbonates: ankerite Ca(Fe, Mg, Mn)(CO3)2and kutnohorite Ca(Mn,Mg,Fe)(CO3)2, with bivalent cations: Fe+2 and Mn+2, substituting a part of Mg+2 ions in the layered structure, enter also the dolomite group. A rock formed by the replacement of dolomite with calcite [12] CaMg(CO3)2+Ca +2 =2CaCO3+Mg +2 (1) in dedolomitisation (calcitisation) process is named dedolomite [13]. Calcium ions in this process are provided, for example, by calcium-rich water, and magnesium ions 2Journal of Thermodynamics are liberated. Molar volume of dolomite (184.4/2.899 = 63.608 cm3/mol) is lower than two molar volumes of calcite (2 ·100.09/2.71 =73.867 cm3/mol), and, therefore, reaction (1) causes an increase in the solid volume, equal (73.867 − 63.608)/63.608, that is, ca. 16% [4]. Dedolomitisation is a particular case of a diagenesis [14], where an alteration of sediments into sedimentary rocks occurs. This process is accompanied by an increase in porosity, expressed by a percentage pore space. The pores form a void space or are filled with a fluid. Dolomite is thermodynamically unstable, and dedolomitisation occurs under different conditions. The dedolomitisation in alkaline media is represented by reaction [15] CaMg(CO3)2+Ca (OH)2=2CaCO3+Mg (OH)2.(2) The reverse process (i.e., dolomitisation) occurs, for example, during evaporation of seawater. High temperatures enhance the dolomitisation process [16]. In dolomitisation, magnesium ions from seawater replace calcium ions in calcite, and dolomite is formed [17]. Dolomite growth is favoured by high Mg/Ca ratios and high carbonate contents; this fact is predictable from Le Chatelier-Brown principle [18]. The main objective of this paper is to provide the way for better understanding the dolomite dissolution at different conditions [19]affected by pH, initial concentration of dolomite (C0) in the system ([pr1]t=0=C0), and concentration of CO2. The knowledge of dissolution in acidic media is essential in aspect of improved recovery of oil and gas from sedimentary basins at low temperatures [20], whereas dedolomitisation in alkaline media plays a significant role in deteriorating the concrete structure [15,21]. As will be stated in what follows, the solubility product value (Ksp1) of dolomite is a critical factor in quantitative description of its behavior in the systems where dolomite is put in contact with aqueous solutions, containing dissolved CO2and, moreover, a strong acid (e.g., HCl) or a strong base (e.g., KOH). This paper follows the one concerning struvite [22,23], as another representative of a group of the nonequilibrium precipitates formed by ternary salts. 2. Kinetics of Dolomite Dissolution and Stoichiometry In aqueous systems, ideal dolomite can be considered as an equimolar mixture of two carbonate components: calcite and magnesite, that is, pr1 =pr2 + pr3. This simplifying assumption deserves some reservation, concerning relative rates of dissolution of the dolomite components; the literature provides ambiguous data in this respect, however. The results obtained according to AAS method by Lund et al. [24] exhibited stoichiometric dissolution of dolomite in relation to Ca and Mg, whereas other experiments [25] showed that pure dolomite dissolves more slowly than pure calcite. Dolomite exists in a variety of morphological forms [26]. Minerals with greater defect densities dissolve faster since their effective surface areas are greater than more perfect specimens of the same compound. The rate of surfacecontrolled dolomite dissolution is significantly less than one of calcite [27]. Dissolution rate increases with decreasing grain size [28]. The experiments done for kinetic purposes showed that the mass loss of single dolomite crystals [29] (in response to pH and pCCO2) or one for finely dispersed dolomite particles [27] (in response to pH) was measured. The dissolution of ionic crystals is a complex process, involving some surface and transportation phenomena. Ions are transferred from the surface of the solid material to an unsaturated solution [30]. The surface phenomena depend on the morphology (microstructure) of the crystals. The rate of any dissolution process is effected by surface and transport phenomena. Kinetics of dolomite dissolution has been tested at different pH and temperatures [20]. According to a model by Busenberg and Plummer [29], the dissolution of dolomite is an effect of simultaneous action of H+,H 2CO3,andH 2O. The dissolution studies were usually carried out with suspensions or powdered materials employed, and the resulting concentration changes of Ca and Mg species in the bulk solution were measured [31]. For this purpose, in situ (e.g., conductometric, pH-metric, pH-static [32]) or ex situ (e.g., titrimetry, AAS [33]) methods of analysis were employed [30]. As an option, a rotating disc (RD) formed of dissolving dolomite attached at the end of rotating disc shaft was applied [34]. A loss in mass of the solid material was also measured [35]. It should be noted that repeated trials to precipitate ideal dolomite under laboratory conditions at room temperature were unsuccessful [36,37]; dolomite was precipitated at elevated temperatures (150–300◦C) [38], for example, by heating calcite with Mg salt in aqueous media, at elevated CO2pressures [39]. Dolomite is formed as a result of complex, not well-understood physicochemical phenomena [40,41], because of the difficulties arising in preparation of stoichiometric dolomites [42]. These difficulties caused, among others, that the solubility product (Ksp1)ofdolomite in water Ksp1 =Ca+2Mg+2CO3−22(3) measured according to different methods yielded inconsistent and unreliable results. The mechanism of dolomite formation in sedimentary environment (the so-named dolomite problem [43]) is not well understood, as hitherto [44]. 3. Principles of Simulation of Dolomite Dissolution 3.1. General Remarks. Simulations are needed to check the models used. In modelling of chemical systems, different computer programs were developed. Among others, the Joint Expert Speciation System (JESS) computer program [45–48] is sometimes applied, for example, in [49]. A new approach, called Generalized Approach to Electrolytic Systems (GATES), was elaborated by Michałowski in 1992 and presented lately in some review papers [50–52] and in Journal of Thermodynamics 3 5 6 7 8 9 10 11 pH 2 3 4 5 m m m m −5 −4 −3 −2 −6 s Ca s Ca s Ca s Ca ∞ log ΣMe (a) 78910 pH 2 3 4 5 m m m m −5 −4 −3 −6 s Ca s Ca s Ca s Ca ∞ log ΣMe (b) 789 pH 234m −5 −4 −3 −6 s Ca log ΣMe 6543 −7 5, ∞ (c) 67891011 pH −5 −4 −3 −6 log ΣMe −7 2 3 4 5 m mm s Ca s Ca s Ca ∞ (d) Figure 1: The log Me versus pH relationships ((9), (10)) plotted for different sets of (pC0,pCCO2,pH 0)values:(a)(2,pCCO2, 5); (b) (2, pCCO2, 7); (c) (3, pCCO2, 3); (d) (3, pCCO2,6).ThepCCO2values are indicated at the corresponding curves. For further details, see text. textbooks [53,54]. For example, the systems with struvite [22,23] were elaborated according to GATES. The equilibria in the system with solid carbonates are affected by total concentration (CCO2) of carbonate species, introduced by CO2dissolved in aqueous media, and by presence of NaOH (Cb)orHCl(Ca), used to moderate pH of the solution. At Cb=Ca, the effect is practically tantamount with absence of the related species in the solution, provided that Caand Cbvalues are small; that is, the effect of ionic strength is negligible. 3.2. Equilibrium Constants. The set of equilibrium constants [1,55–59] referred to the two-phase system in question is involved with solubility products (Kspi) of precipitates, (pri, i=1, ..., 5), expressed by (3)and Ca+2CO3−2=Ksp2pKsp2 =8.48, Mg+2CO3−2=Ksp3pKsp3 =7.46, 4Journal of Thermodynamics Mg+2OH−12=Ksp4pKsp4 =10.74, Ca+2OH−12=Ksp5pKsp5 =5.03, (4) and soluble species (complexes, protonated forms): MgHCO3+1=101.16 ·Mg+2HCO3−1, MgCO3=103.4·Mg+2CO3−2, CaHCO3+1=101.11 ·Ca+2HCO3−1, [CaCO3]=103.22 ·Ca+2CO3−2, HCO3−1=1010.33−pH ·CO3−2, [H2CO3]=106.38−pH ·HCO3−1, MgOH+1=102.57 ·Mg+2OH−1, CaOH+1=101.3·Ca+2OH−1, H+1OH−1=KwpKw=14.0. (5) One should be noted that soluble complex species: MgCO3and CaCO3, characterized by stability constants, are different from the precipitates: pr2 =CaCO3and pr3 = MgCO3,characterized by the solubility products (Kspi,i= 2, 3) values. The pKsp1 values for dolomite reported in the literature range from ca. 16.5 to 19.35 [37,55,56,60,61], that is, within ca. 3 units. Some inferences [37] lead to the conclusion that the most probable pKsp1 value is 17.2±0.2. In this context, the value 19.35 taken in [55] seems to be excessively high. Such diversity in pKsp1 value may reflect the difficulties involved with obtaining stoichiometric dolomite. The inequality Ksp1 <K sp2 ·Ksp3, that is, pKsp1 > pKsp2 +pKsp3,validforallKsp1 values quoted above, expresses a kind of synergistic effect securing almost perfect arrangement of Mg+2 and Ca+2 ions in area of the corresponding planes of dolomite crystallographic structure. The values for stability constants of soluble complexes: Mg(OH)2and Ca(OH)2, introduced for calculations in [55], are highly controversial and then were omitted in the related balances formulated below. In this context, the complexes MgOH+1 and CaOH+1 and their stability constants are well established. The dissolution of dolomite (pr1) proceeds up to the saturation of the solution against the corresponding precipitate. Provided that the solution is unsaturated against the corresponding solid phase (pri), the expression for the related solubility product (Kspi) is not valid, under such conditions. For this purpose, the expressions q1=Ca+2Mg+2CO3−22 Ksp1 , q2=Ca+2CO3−2 Ksp2 , q3=Mg+2CO3−2 Ksp3 , q4=Mg+2OH−12 Ksp4 , q5=Ca+2OH−12 Ksp5 (6) related to all possible precipitates (pri, i=1, ...,5) were considered. 3.3. Formulation of the Dolomite Dissolution Model. We refer to the system obtained after introducing m1gofdolomite into VmL of aqueous solution with dissolved CO2(CCO2), NaOH (Cb), and/or HCl (Ca); NaOH and/or HCl are used to moderate pH value of the solution. Assuming that the volume change resulting from addition of pr1 is negligible, and denoting current (t≥0) concentration of pr1 by [pr1], [pr1]t=0=C0=(103·m1/M1)/V, we get the concentration and charge balances F1(x)=pr1+Mg −C0=0, F2(x)=pr1+Ca −C0=0, F3(x)=2pr1+MgHCO3 +1+MgCO3 +CaHCO3+1+[CaCO3]+[H2CO3] +HCO3−1+CO3−2−2C0+CCO2=0, F4(x)=H+1−OH−1+Δ+2 Mg+2+MgOH+1 +2 Ca+2+CaOH+1+MgHCO3 +1 +CaHCO3+1−HCO3−1−2CO3−2=0, (7) where Δ=Cb−Ca(8) the expressions: Mg=Mg+2+MgOH+1+MgHCO3 +1+MgCO3, (9) Ca=Ca+2+CaOH+1+CaHCO3+1+[CaCO3] (10) Journal of Thermodynamics 5 pH 1 3 7 7 6 4 2 5 4 3 8 8 7 7.5 8 8.5 9 9.5 10 10.5 9.98 −13 −15 −11 −9 −7 −5 −3 5, 6 1, 2 log [Xi] (a) pH 1 3 7 6 4 2 5 4 3 8 7 7.5 8 8.5 9 9.5 10 10.5 9.99 −13 −15 −11 −9 −7 −5 −3 5, 6 1, 2 7, 8 log [Xi] (b) pH 1 3 7 6 4 2 5 4 3 8 7 7.4 8.2 9 9.4 8.73 −12 −10 −8 −6 −4 −2 5, 6 1, 2 7 7.8 8.6 8 log [Xi] (c) pH 1 3 7 7 6 4 2 5 4 3 8 8 7 7.5 8 8.5 9 9.5 8.74 −13 −11 −9 −7 −5 −3 5, 6 1, 2 log [Xi] (d) Figure 2: The log[Xi] versus pH relationships plotted for different sets of (pC0,pCCO2,pH 0)values:(a)(2,∞, 7); (b) (2, 3, 7); (c) (3, ∞, 7); (d) (3, 3, 7) for different soluble species Xiindicated at the corresponding curves: 1—Ca+2,2—Mg +2,3—CaOH +1, 4—MgOH+1, 5—CaHCO3+1, 6—MgHCO3+1,7—CaCO3,8—MgCO3. involve all soluble magnesium and calcium species; x= [x1,x2,x3,x4]is the vector of independent variables specified below. The number of variables forming the vector xis equal to the number of equations Fi(x)=0(i=1, ...,4 )(11) specified above. 3.4. Calculation Procedure. The relations (11) are valid, if xis chosen properly. For any other vector x,x/ =x, at least one of the functions (11)differs from zero, and then F(x )= 4  i=1 [Fi(x )]2>0.(12) The calculation procedure is based on minimizing principle, quite analogous to one presented in [22,23]. According to 6Journal of Thermodynamics this approach, the sum of squares (12) is taken as the minimized function. In a dynamic process, as the dissolution of pr1 is, a choice of ppr1 =−log[pr1] as the steering variable is advised; the value of [pr1] changes during the pr1 dissolution. The next step is the choice of independent variables, xi=xi(ppr1), i=1, ...,4. The xivalues are involved with concentrations of some independent species. In our case, the best choice is x=[x1,x2,x3,x4]=[pH, pHCO3,pMg,pCa] , where pH =−log[H+1], pHCO3=−log[HCO3−1], pMg = −log[Mg+2], pCa =−log[Ca+2] were considered as independent variables. A choice of pXindices, not concentrations [X], resulted from the fact that 10−pXi>0foranyreal pXi=−log[Xi]value,pXi∈. The third step is the choice of numerical values for components of the starting vector x=x∗;ifx∗/ =x, one can expect that F(x∗)>0(12). The x∗value is referred to particular value of the steering variable, x∗=x∗(ppr1), for example, to ppr1 =−log C0. The step Δppr1 is needed, and initial steps ΔpXifor pXiand lower (pXi,inf )andupper(pXi,sup) limits for expected pXi values are also required by the iterative computer programs applied in the minimization procedure. The searching of x(ppr1) vectors, where F(x(ppr1)) is closetozerofordifferent ppr1 values, can be made according to MATLAB iterative computer program. The searching procedure satisfies the requirements put on optimal x(ppr1) values—provided that the F(x(ppr1)) value (12), considered as optimal one, is lower than a preassumed, sufficiently small δ-value Fxppr1 <δ. (13) Then we consider that the equality x(ppr1) =x(ppr1) is fulfilled. It means that Fi(x(ppr1)) ∼ =0foralli=1, ...,4, that is, (7) are fulfilled simultaneously, within tolerable degree of proximity assumed for all ppr1 values taken from defined ppr1 interval. If pr1 dissolves wholly, the ppr1 covers the interval from pC0up to the value resulting from graphical needs, that is, from the scale for ppr1 assumed to plot the related (e.g., speciation) curves. If the solubility product for pr1 is attained at defined point of the dissolution process, then upper value assumed for ppr1 is the value corresponding to this point. The iterative computer programs are usually designed for the curve-fitting procedures where the degree of fitting of a curve to experimental points is finite. In particular case, the criterion of optimization is based on differences F(x(ppr1(N+ 1))) −F(x(ppr1(N))) between two successive (Nth and N+1th) approximations of the F-value, and the optimisation is terminated if the inequality  Fxppr1(N+1 )−Fxppr1(N) <δ (14) is valid for any preassumed, sufficiently low δ-value, for example, δ=10−15. However, one may happen that the condition (14)can be fulfilled at local minimum different from the global minimum. It occurs when the starting values x∗(ppr1) are too distant from the true value x(ppr1), where the equality F(x(ppr1)) =0 is fulfilled. In this case, one should repeat the calculations for a new vector x∗(ppr1), guessed at a particular ppr1 value chosen at the start for minimisation. All vectors x=x(ppr1) obtained for different ppr1 values are the basis for plotting the speciation curves for all species (Xi) in the system. The curves are usually plotted in the logarithmic diagrams, on 2D plane, with ppr1 on the abscissa and log[Xi] on the ordinate. Other variables, for example, pH on the abscissa, can also be applied. 4. Simulated Dedolomitisation 4.1. Preliminary Data. We refer first to aqueous solutions obtained before introducing pr1 into it. Any particular solution is characterised by pCCO2=−log CCO2and pH = pH0values, where pCCO2equals 2, 3, 4, 5, or ∞(the latter value refers to CCO2=0) and pH0values cover the set of natural numbers within the interval 3, 12.ThepH 0value corresponds to the presence of a strong acid (HB, Camol/L) or strong base (MOH, Cbmol/L), see (8). The solution is then characterised by a defined pair of (pCCO2,pH 0)values. After introducing pr1 into the solution, its initial (t= 0) concentration in the two-phase system equals [pr1]t=0= C0mol/L. Two values for pC0=−log C0,equal2and3,were assumed. The set of parameters (pC0,pCCO2,pH 0)assumed involves a multitude of different phenomena occurred in the system considered. The examples presented in the follwing concern only some particular cases. The volume change of the system, affected by addition of pr1, can be neglected. The volume changes, involved with further dissolution and formation of precipitates, are neglected too. In order to neglect the diffusion effects, the systems were (virtually) mixed (homogenised). The dissolution has been considered as a quasistatic process, carried out under isothermal conditions. 4.2. Discussion on Particular Systems. At the first stage of the process that occurred after pr1 addition, the solution is unsaturated against any particular precipitate, that is, qi<1(i= 1, ...,5)in(6). At defined point, it saturates against another precipitate. As will be stated in the following, two different precipitates: pr2 =CaCO3or pr5 =Mg(OH)2,areformedin the systems considered. Then the solution saturates against pr1 or transforms wholly to the second precipitate; in the latter case, the saturation towards pr1 is not attained. At a relatively high CCO2value, pr1 dissolves wholly and no other precipitate is formed (qi<1). The curves expressing the relations (9)and(10) are referred to unsaturated solutions in nonequilibrium systems and termed dissolution curves. The curves expressing the relations (9)and(10), when referred to the solutions saturated against a particular precipitate, are termed the solubility curves (Me ≤sMe,Me=Ca, Mg). The dissolution curves presented below are then terminated: (a) by the bifurcation point, where the solubility product (Kspi) for the corresponding precipitate (pr2 or pr5) is attained (q2=1orq5=1), or (b) by the point, where [pr1] =0atqi<1(i=1, ...,5), that is, pr1 dissolves wholly. Journal of Thermodynamics 7 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca log s Me (a) 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca log s Me (b) 891011 12 pH 3 4 12 11 1 −5.5 −5 −4.5 −4 −3.5 −3 −6 s Mg s Ca 7–9 log s Me 5-6 (c) 8 9 10 11 12 pH 3 4 1 −5.5 −5 −4.5 −4 −3.5 −3 −6 s Mg s Ca 5–9 1 1 log s Me (d) Figure 3: The log sMe versus pH relationships (Me =Ca, Mg) plotted at pKsp1 =16.54 for pr1, Me =sMe ((9), (10)) at different (pC0, pCCO2)values:(a)(2,∞); (b) (3, ∞); (c) (2, 4); (d) (3, 4); the numbers at the corresponding points correspond to pH0values. The dissolution/solubility curves, obtained at pC0=2 and 3, are plotted in Figures 1(a)–1(d),fordifferent pH0and pCCO2values. The curves plotted in Figures 1(c) and 1(d) refer to different sets of (pC0,pCCO2,pH 0)=(3, pCCO2,pH 0) values. When CCO2exceeds distinctly the C0value (pC0=3), pr1 dissolves wholly before the solubility product for pr2 is attained. The plots for soluble species consisting the expressions for Ca and Mg ((9), (10)) and referred to different sets of (pC0,pCCO2,pH 0) values are presented in Figures 2(a) and 2(b) (for pC0=2) and Figures 2(c) and 2(d) (for pC0=3). As results from the course of speciation curves plotted in Figure 2, the first (dissolution) step can be represented by reactions: pr1 + 2H2O=Ca+2 +Mg +2 + 2HCO3−1+2OH −1,(15a) pr1 + H2O=Ca+2 +Mg +2 +HCO 3−1+CO 3−2+OH −1, (15b) pr1 =Ca+2 +Mg +2 +2CO 3−2,(15c) 8Journal of Thermodynamics 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca 7 log s Me (a) 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca 7 log s Me (b) 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca 7 log sMe (c) 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca 7 log sMe (d) 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 s Mg s Ca 7 log sMe (e) 8 9 10 11 12 pH −4.5 −4 −3.5 −3 −5 7 17.2 17.4 16.7 16.5 17 log sMg (f) Figure 4: The log sMe versus pH relationships (Me =sMe in (9), (10)at(pC0,pCCO2)=(2, 5) for different pKsp1 values assumed for pr1: (a) 16.54; (b) 16.7; (c) 17.0; (d) 17.2; (e) 17.4; (f) collected effect of Ksp1 values (indicated at the corresponding curves) on log sMg value. Journal of Thermodynamics 9 8 9 10 11 12 pH 3 4 12 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 sMg sCa 7 log sMe m (a) 8 9 10 11 12 pH 3 4 1 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 sMg sCa 7 log sMe (b) 8 9 10 11 12 pH 3 4 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 sMg sCa 7 12 log sMe (c) 8 9 10 11 12 pH 3 4 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 sMg sCa 7 12 log sMe (d) 8 9 10 11 12 pH 3 4 11 10 −5.5 −5 −4.5 −4 −3.5 −3 5–9 −6 sMg sCa 7 12 log sMe (e) 891011 12 pH −5 −4.5 −4 −3.5 −3 7 17.2 17.4 16.7 16.5 17 log sMe (f) Figure 5: The log sMe versus pH relationships (Me =sMe in (9), (10)) at (pC0,pCCO2)=(3, 5) for different pKsp1 values assumed for pr1: (a) 16.54; (b) 16.7; (c) 17.0; (d) 17.2; (e) 17.4; (f) collected effect of pKsp1 values (indicated at the corresponding curves) on logsMg value.