Dawsonite synthesis and reevaluation of its thermodynamic properties from solubility measurements: Implications for mineral trapping of CO2

Geochimica et Cosmochimica Acta 71 (2007) 4438–4455 www.elsevier.com/locate/gca

Dawsonite synthesis and reevaluation of its thermodynamic properties from solubility measurements: Implications for mineral trapping of CO2 Pascale Be´ne´zeth a

a,*

, Donald A. Palmer b, Lawrence M. Anovitz

b,c

, Juske Horita

b

L.M.T.G. UMR-5563 CNRS/IRD/Universite´ Toulouse 3, 14 Avenue Edouard Belin 31400, Toulouse, France b Chemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831-6110, USA c Department of Earth and Planetary Sciences, University of Tennessee, Knoxville, TN 37996, USA Received 28 November 2006; accepted in revised form 25 June 2007; available online 18 July 2007

Abstract Over the last decade, a significant research effort has focused on determining the feasibility of sequestering large amounts of CO2 in deep, permeable geologic formations to reduce carbon dioxide emissions to the atmosphere. Most models indicate that injection of CO2 into deep sedimentary formations will lead to the formation of various carbonate minerals, including the common phases calcite (CaCO3), dolomite (CaMg(CO3)2), magnesite (MgCO3), siderite (FeCO3), as well as the far less common mineral, dawsonite (NaAlCO3(OH)2). Nevertheless, the equilibrium and kinetics that control the precipitation of stable carbonate minerals are poorly understood and few experiments have been performed to validate computer codes that model CO2 sequestration. In order to reduce this uncertainty we measured the solubility of synthetic dawsonite according to the equilibrium: NaAlCO3 ðOHÞ2ðcrÞ þ 2H2 OðlÞ  AlðOHÞ4  þ HCO3  þ Naþ þ Hþ , from under- and oversaturated solutions at 50–200 C in basic media at 1.0 mol Æ kg1 NaCl. The solubility products (Qs) obtained were extrapolated to infinite dilution to obtain the solubility constants (K s o Þ. Combining the fit of these logK s o values and fixing DC p;r o at  185:5 J  mol1  K 1 at 25 C, which was derived from the calorimetric data of Ferrante et al. [Ferrante, M.J., Stuve, J.M., and Richardson, D.W., 1976. Thermodynamic data for synthetic dawsonite. U.S. Bureau of Mines Report Investigation, 8129, Washington, D.C., 13p.], the following thermodynamic parameters for the dissolution of dawsonite were calculated at 25 C: DGr o ¼ 102:1 kJ  mol1 , DH r o ¼ 97:0 kJ  mol1 and DSr o ¼ 17:1 J  mol1  K 1 . Subsequently, we were able to derive values for the Gibbs energy of formation (Df Go298:15 ¼ 1782  2 kJ  mol1 Þ, enthalpy of formation (Df H o298:15 ¼ 1960  7 kJ  mol1 Þ and entropy (S o298:15 ¼ 131  2 J  mol1  K 1 Þ of dawsonite. These results are within the combined experimental uncertainties of the values reported by Ferrante et al. (1976). Predominance diagrams are presented for the dawsonite/boehmite and dawsonite/bayerite equilibria at 100 C in the presence of a saline solution with and without silica-containing minerals.  2007 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Evidence of global warming has been observed since the beginning of the industrial revolution (Marchetti, 1977), and is largely being attributed to the accumulation of

*

Corresponding author. Fax: +33 5 61 33 25 60. E-mail address: [email protected] (P. Be´ne´zeth).

0016-7037/$ - see front matter  2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2007.07.003

anthropogenic carbon dioxide in the Earth’s atmosphere. One of the many solutions proposed to reduce further emissions, and remediate the existing problem, involves the capture and storage of CO2 in deep subsurface sedimentary formations. Among these formations, deep saline aquifers are attractive both because they have a large potential storage capacity for carbon dioxide (estimates are on the order of several hundred years of total anthropogenic CO2 emissions, cf., Bergman and Winter, 1995) and, in many parts of

Dawsonite synthesis and solubility measurements

the world, they are located near CO2 emission sources (Bachu et al., 1994), thereby reducing the cost of CO2 transportation. Pilot projects are underway in Norway where approximately one million metric tons of CO2 have been injected annually since 1996 into the saline Utsira formation at the Sleipner natural gas production field in the North Sea (cf., Holloway, 2005). In addition, the Frio Pilot Project in Texas has provided a smaller-scale, scientifically-monitored test of the effects of CO2 injection into a saline aquifer (cf., Kharaka et al., 2006). If CO2 is to be safely and permanently stored by this method, two processes need to be understood: ‘‘solubility trapping’’—dissolution and long-term retention of CO2 in saline groundwater; and ‘‘mineral trapping’’—reaction of CO2 with non-carbonate minerals to form stable carbonates (cf., Gunter et al., 1997; Wawersik et al., 2001; Johnson et al., 2001, 2002). However, a number of reactions may occur involving supercritical CO2, aquifer fluids and formation rocks, depending on the chemical composition of the fluid, pH, temperature, pressure and the mineralogy of the host rock (e.g., Gaus et al., 2005; Moore et al., 2005). The present study focuses on mineral trapping, in particular the formation of dawsonite (NaAlCO3(OH)2). Johnson et al. (2001, 2002) showed, on the basis of reactive transport modeling of CO2 injection into the Sleipner aquifers, that reaction among Al-bearing silicates, groundwater, and injected CO2 will produce not only calcite-group carbonates but also substantial quantities of dawsonite in the ‘‘intraaquifer plume’’. This reaction is enhanced by high ambient Na+ concentrations, plume-induced CO2 aqueous solubility, and acid-induced K-feldspar (orthoclase) dissolution, according to the following reaction: KAlSi3 O8ðsÞ þ Naþ þ CO2ðaqÞ þ H2 OðlÞ  NaAlCO3 ðOHÞ2ðsÞ þ 3SiO2ðsÞ þ Kþ

ð1Þ

The overall reaction (1) can be broken down into two separate processes, namely (i) the dissolution of feldspar (reactions involving Na- or Ca-rich feldspars can be substituted) by exposure to CO2-induced acidic conditions: KAlSi3 O8ðsÞ þ 4Hþ  Kþ þ Al3þ þ 3SiO2ðsÞ þ 2H2 OðlÞ ð2Þ followed by (ii) precipitation of dawsonite by interaction of the released aluminum with Na+ and CO2(aq)-rich plume fluids: Naþ þ Al3þ þ CO2ðaqÞ þ 3H2 OðlÞ  NaAlCO3 ð OHÞ2ðsÞ þ 4Hþ ð3Þ

Similarly, simulations of CO2 sequestration in Ohio’s deepsaline aquifers (The Rose Run Sandstone aquifer) performed by Zerai et al. (2006) indicate that, under certain conditions (e.g., fCO2), dissolution of albite (NaAlSi3O8, an abundant feldspar in that aquifer) can lead to mineral trapping of CO2 in dawsonite through a reaction analogous to reaction (1). Some kinetic data are available for reaction (2) and its analogs under acidic (cf., Blum and Stillings, 1995) and basic conditions (e.g., Gautier et al., 1994; Oelkers et al., 1994; Hellmann and Tisserand, 2006). Recently, research efforts have focused on the effect of CO2 on the dissolution rates

4439

of silicates (e.g., Carroll and Knauss, 2005; Golubev et al., 2005) and carbonates (Pokrovsky et al., 2005). These data show no evidence of a direct effect of CO2 other than increasing the acidity. Surprisingly few experiments (cf., Kaszuba et al., 2003, 2005) have been performed to validate computer codes that model mineral trapping of CO2. In particular, there are currently no data available to adequately assess the rate constants for dawsonite formation, which geochemical models have either set to those of K-feldspar (Xu et al., 2003 from Blum and Stillings, 1995); siderite (Xu et al., 2005); calcite (White et al., 2005) or estimated as intermediate between calcite and magnesite (e.g., Johnson et al., 2001, 2002; Gaus et al., 2003; Knauss et al., 2005). The goal of this study was to perform solubility experiments based on in situ pH measurements that can be used to (re)evaluate its thermodynamic properties and better elucidate the factors that control dawsonite formation. Literature data suggest (cf., Herold, 1992; Zhang et al., 2004; see below) that the optimal experimental conditions to synthesize dawsonite require moderately basic solutions, such that, þ  Naþ þ AlðOHÞ 4 þ HCO3 þ H ! NaAlCO3 ðOHÞ2ðsÞ þ 2H2 OðlÞ

ð4Þ because much higher total carbonate concentrations can be obtained. Although the fluid pH does not reach high values in most simulations, this reaction is less complicated to study than its acidic analog because, over the wide range of conditions found in deep sedimentary formations (before injection of CO2), the aluminate anion, AlðOHÞ4 - , is predominant and the presence of multiple hydrolyzed Al species can be ignored (Be´ne´zeth et al., 1997, 2001; Palmer et al., 2001). We have, therefore, measured the solubility quotient of this reaction from underand oversaturated conditions as a function of temperature (50–200 C) in either NaOH, a mixture of sodium carbonate, or sodium bicarbonate/NaOH solutions, at 1.0 mol Æ kg1 NaCl. In addition, in order to perform these measurements, it was first necessary to synthesize hydrothermally the starting material, as sufficiently large samples of pure natural dawsonite were unavailable, and such materials were likely to be contaminated with impurities. The processes and results of these syntheses are also described herein. 2. THE MINERAL DAWSONITE 2.1. Natural dawsonite Dawsonite is a naturally occurring mineral with various habits (e.g., prismatic, striated crystals, acicular or fibrous, spheres, rosettes and random aggregates. . .). It was first discovered in a dike cutting the Trenton limestone on the campus of McGill University (Montreal, Quebec), and was named by Harrington (1874) after Sir John William Dawson (Canadian geologist and Principal of McGill University). A description of dawsonite occurrence in the feldspathic dike at the Montreal locality was given by Stevenson and Stevenson (1965). The crystal structure of that dawsonite, determined by Frueh and Golightly (1967),

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

4440

is orthorhombic-disphenoidal consisting of distorted AlO2(OH)4 and NaO4(OH)2 octahedra, and CO3 groups. The structure is consistent with the space group Imma with a = ˚ , and z = 4 (Corraza et al., 1977). 6.77, b = 10.40 and c = 5.58 A Dawsonite occurs at several localities worldwide: as an hydrothermal alteration product of igneous rocks; in sedimentary rocks derived from volcanics or affected by CO2rich alkaline, thermal solutions, and in a wide variety of sedimentary rocks (e.g., Baker et al., 1995; Sirbescu and Nabelek, 2003). For example, the presence of pervasive dawsonite cement in the Bowen-Gunnedah-Sydney Basin, Eastern Australia, was interpreted to reflect magmatic CO2 seepage on a continental scale (Baker et al., 1995). Traces of dawsonite are common as a secondary product in hydrothermal fluid inclusions (e.g., Coveney and Kelly, 1971; Sirbescu and Nabelek, 2003). Pore-filling dawsonite has also been found in core samples from the Springerville-St. Johns area, in New Mexico and Arizona, USA (Moore et al., 2005). From petrographic relationships and results of geochemical simulations, these authors suggested that deposition of dawsonite occurred when the fugacity of CO2 reached 20 bars, but they could not explain its limited distribution. Conversely, substantial quantities exist in two other locations, which suggest that dawsonite requires a highly alkaline environment in which to form: (i) in oil shales of the Eocene Green River Formation in the Piceance Creek Basin of northwestern Colorado (Smith and Milton, 1966) associated with analcime (NaAlSi2O6 Æ H2O) and (ii) in the weathered nepheline syenite tuffs of Olduvai Gorge, Tanzania, East Africa (Hay, 1963) associated with natrolite (Na2Al2Si3O10 Æ 2 H2O). The first occurrence generated interest in dawsonite as a potential ore of aluminum as it was estimated to contain 26 billion metric tons of dawsonite (e.g., Hite and Dyni, 1967), which is sufficient to produce 42 million tons of alumina (compared to

30 million tons from bauxite deposits of the US). This quantity of dawsonite corresponds to 7950 Gt of sequestered CO2. 2.2. Synthetic dawsonite Dawsonite was first synthesized by Bader (1938), and later by Jackson et al. (1972), Besson et al. (1973), and Furmakova (1981). Bader (1938) reported, on the basis of X-ray diffraction data, that crystallization was incomplete, but synthesized dawsonite again in 1944 (Bader and Esch, 1944) and reported that the crystallinity of this material was nearly equal to that of natural dawsonite from Albania. Jackson et al. (1972) synthesized dawsonite with a crystallinity similar to that found in oil shale deposits of the Piceance Creek area of Colorado. Their optimum experimental conditions were an atomic ratio of 43 for Na/Al, temperatures between 175 and 200 C and 1 bar CO2 for 5 h. More recently, Zhang et al. (2004) investigated the effects of hydrothermal conditions (temperature, pH, MHCO3/Al ratio) on the synthesis of dawsonite-type compounds (MAl(OH)2CO3, with M = Na, K, NH4). They concluded that increasing the MHCO3/Al ratio and temperature favors the growth of large dawsonite crystals and that the most favorable pH value to obtain sodium dawsonite is 10.3. The same range of conditions (high pH, Na/Al ratio, temperature. . .) can also be found in U.S. patents issued on dawsonite syntheses as it is widely used as an antacid (with sugar added), as a flame retardant and as an absorbent for industrial flue gases (e.g., Van Der Heem, 1980; Misra, 1980; Altman, 1982; Kaufman, 1984; Herold, 1992). The patents of Van Der Heem (1980) and Herold (1992) will be described in more detail below, as their methods were used and optimized to synthesize our dawsonite.

Table 1 Stoichiometric molal compositions of starting solutions and solid phase used in each run Run No.

t C

Reference cell 3

1 2 2 3 3 4 5 5 6 6 7 8 8 8 8 8

100.1 100.1a 150.1 100.1a 150.1 99.8 99.8a 200.7 99.8 99.8b 100.0c 50.0d 75.0 100.0 150.0 200.0 a b c d

Test cell 3

4

3

10 m(HCl)

m (NaCl)

10 m(NaOH)

10 m(HCl)

10 m(Na2CO3)

10 m(NaHCO3)

m(NaCl)

2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 2.000 1.999 1.999 1.999 1.999 1.999

0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998 0.998

0.0996 1.035 1.035 1.001 1.001 0.999 2.000 2.000 0 0 0 1.000 1.000 1.000 1.000 1.000

0 0 0 0 0 0 0 0 3.000 3.000 2.999 0 0 0 0 0

0 9.74 9.74 9.99 9.99 10.01 0 0 0 0 0 10.00 10.00 10.00 10.00 10.00

0 0 0 0 0 0 0.04002 0.04002 0 0 0 0 0 0 0 0

1.004 0.973 0.973 0.974 0.974 0.969 0.878 0.878 1.000 1.000 1.000 0.980 0.980 0.980 0.980 0.980

Run carried out by increasing the temperature from 100 to 150 C (runs 2 and 3) or from 100 to 200 C (run 5). Two injections of CO2 gas (10 bar) at this temperature. Injection of CO2 gas (10 bar) at 25 C. In this case the temperature was ramped from 50 to 200 C.

Dawsonite used (see text) Daws-A Daws-B Daws-B Daws-B Daws-B Daws-B Daws-B Daws-B Daws-C Daws-C Daws-C Daws-D Daws-D Daws-D Daws-D Daws-D

Dawsonite synthesis and solubility measurements

2.3. Thermodynamic properties Ferrante et al. (1976) determined the thermodynamic properties of synthetic dawsonite, precipitated according to the procedure of Jackson et al. (1972), from a series of calorimetric studies. Low-temperature heat capacities were determined from 6 to 307 K using an adiabatic calorimeter. A copper-block drop calorimeter was used to obtain superambient enthalpies at temperatures to 477 K. Enthalpy of formation measurements were made using a solution calorimeter with HCl. The thermodynamic data tabulated by these authors are: Df Go298:15 ¼ 1786  4 kJ  mol1 ; Df H o298:15 ¼ 1964  4 kJ  mol1 S o298:15 ¼ 132  2 J  mol1  K1 and C op298:15 ¼ 142:6  0:4 J  mol1  K 1 . The uncertainties given above were calculated by us as 3r of the experimental uncertainties reported by Ferrante et al. (1976). Note that the C op298:15 value and some of the

4441

uncertainties quoted in Robie and Hemingway (1995) are incorrect. 3. MATERIALS AND EXPERIMENTAL METHODS 3.1. Starting solutions All solutions were prepared from reagent-grade chemicals and distilled deionized water (resistivity 0.18 MX Æ m). Concentrated stock solutions of NaOH, HCl, NaCl, Na2CO3 and NaHCO3 were used to make up the desired experimental solutions. Preparation of these stock solutions, as well as their storage and handling, followed standard procedures developed in our laboratory and are described in detail elsewhere (e.g., Palmer and Wesolowski, 1993). The compositions of the starting reference and test solutions for each experimental run are given in Table 1.

Fig. 1. XRD diffractogram (a) and SEM photomicrograph (b) of the synthesized dawsonite: Daws-A.

4442

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

3.2. Hydrothermal synthesis and characterization of dawsonite Most of the syntheses were carried out at saturated vapor pressure in an Autoclave Engineers 300 mL Hastelloy C bolt-closure pressure vessel fitted with a Teflon liner and a Teflon-coated magnet for agitation. The pressure vessel was placed in a vertical Marshall furnace housed inside an aluminum liner. Temperatures were maintained using a Honeywell digital controller (±0.5 C). A permanent horseshoe magnet mounted below the pressure vessel was used to drive the Teflon-coated stirring magnet in the pressure vessel. Most of our synthesis experiments were designed on an optimization of the procedure patented by Herold (1992). This procedure consists of either an homogeneous nucleation in which an aqueous sodium aluminate solution, obtained by dissolution of gibbsite, Al(OH)3, in a caustic solution of sodium bicarbonate, is reacted with urea and citric acid, or an heterogeneous process in which gibbsite is reacted directly with a sodium carbonate, bicarbonate

and citric acid solution. Urea is used in this method to buffer the pH of the solution whereas the hydroxycarboxylic acid, in our case citric acid, is added to produce crystals with a short, prismatic isometric habit (length of 0.1– 2 lm) instead of an acicular habit. The former is preferred for industrial preparations of materials incorporated into plastics, in order to avoid production of fibers for toxicological considerations and was preferred for our experiments in order to study a single crystalline habit. The aqueous sodium aluminate solution for the homogeneous reaction was prepared by dissolving ca. 4 g of gibbsite (treated, acid-washed gibbsite; see Wesolowski, 1992) in 164 g of H2O and 28 g of NaOH (50 %wt) at 50 C for several hours in a Savilex container. Subsequently, 25 g of NaHCO3, 2.4 g of urea and 0.08 g of citric acid were added to the sodium aluminate solution and the mixture was transferred to a Teflon-lined pressure vessel. The vessel was then sealed, heated and stirred continuously. Four syntheses were carried out this way at various temperatures (130–180 C) and/or reaction times (4 h to three days) in order to optimize the dawsonite yield. The opti-

Fig. 2. XRD diffractogram (a) and SEM photomicrographs (b and c) of the synthesized dawsonite: Daws-B and C.

Dawsonite synthesis and solubility measurements

mal condition, following this procedure, required the reaction to proceed at 150 C for at least 4 h. Note that, as re´ lvarez-Ayuso and Nugteren (2005), a constant ported by A XRD pattern was obtained after 24 h. Powder X-ray diffraction (XRD) and Scanning Electron Microscopy (SEM) examinations (Figs. 1a and Fig. 1b, respectively) of the run products revealed that well-crystallized dawsonite (Daws-A, used in run 1, see Table 1) was obtained but, in all cases, various amounts of boehmite, AlOOH, were present (Fig. 1a). A better yield of dawsonite (containing less boehmite than above) was obtained using the heterogeneous reaction pathway of Herold (1992). In this approach 2.9 g of gibbsite, 10 g of NaHCO3, 13 g of Na2CO3 and 0.05 g of citric acid were mixed in 120 g of H2O in the pressure vessel for 4 h at 175 C. As can be seen from Figs. 2a and b (XRD and SEM photographs, respectively), this procedure also produced well-crystallized dawsonite as striated prismatic

4443

crystals. This material was used for runs 2–5 (Daws-B, Table 1). In one case, however, we obtained dawsonite with an acicular morphology whereby the small needles formed balls (Fig. 2c), probably because the temperature reached 200 C during the synthesis. This sample (Daws-C) was used in runs 6 and 7. A final run (8) was carried out by using a dawsonite synthesized by optimizing (temperature and length of reaction time) the method described in the patent of Van Der Heem (1980). XRD analysis showed that very pure dawsonite without any other detectable aluminum phases (see Fig. 3a) was obtained by mixing 15 g of gibbsite (from Riedel-De Hae¨n), 45 g of Na2CO3, 18 g of NaHCO3 and 281 g of H2O at 175 C for 24 h in a 500 mL titanium pressure vessel (without a Teflon liner). The pressure vessel was placed in a rocking furnace to provide efficient mixing in this larger capacity vessel and this procedure produced a better yield of dawsonite compared to the previous

Fig. 3. XRD diffractogram (a) and SEM photomicrograph (b) of the synthesized dawsonite: Daws-D.

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

4444

procedures. An SEM photograph of this dawsonite (DawsD) is shown in Fig. 3b. 3.3. Experimental procedure In previous studies, we demonstrated our ability to make precise in situ measurements of pH during homoge-

neous and heterogeneous aqueous reactions over long periods of time (days to weeks) at temperatures from 0 to 300 C and salinities ranging up to NaCl saturation, using a hydrogen-electrode concentration cell (HECC) (cf., Mesmer et al., 1995; Wesolowski et al., 1995; Be´ne´zeth et al., 1997, 1999; Palmer et al., 2001). In this study, a series of dawsonite solubility measurements was carried out from

Table 2 Experimental results for dawsonite solubility experiments in NaCl media t C

Run No.

log½H þ meas a

log½RAlmeas a

½RNameas a

½Rcarbonatemeas a

Time (h) of samplingf

100.1 100.1 100.1 100.1 100.1b 100.1b 150.1 150.1 100.1b 100.1b 150.1 150.1 150.1 150.1 99.8 99.8 99.8b 99.8b 99.8b 99.8b 200.7 200.7 99.8 99.8c 99.8 99.8 99.8c 99.9d 99.9 99.9 99.9 99.9 99.9 99.9 99.9 50.1e 50.1 50.1 50.2 75.1 75.1 75.2 100.3 150.1 150.0

1 1 1 1 2 2 2 2 3 3 3 3 3 3 4 4 5 5 5 5 5 5 6 6 6 6 6 7 7 7 7 7 7 7 7 8 8 8 8 8 8 8 8 8 8

8.708 8.694 8.664 8.411 8.996 8.939 8.215 7.989 9.025 9.005 8.232 8.217 7.943 6.772 9.003 8.950 8.098 8.035 8.003 7.991 7.713 7.294 7.723 5.630 5.678 5.702 5.298 4.215 4.618 4.747 4.419 4.489 4.535 4.620 4.664 9.849 9.831 9.802 9.757 9.318 9.309 9.282 8.888 8.148 8.138

2.76 2.83 2.95 3.28 2.57 2.74 2.85 3.19 2.52 2.65 2.77 2.96 3.04 3.06 2.49 2.83 3.81 3.84 3.83 3.81 3.11 3.16 4.00 5.92 5.89 5.99 6.11 4.89 5.52 6.15 6.15 6.17 5.97 6.33 6.59 3.04 3.29 3.33 3.39 3.11 3.17 3.20 2.93 2.97 3.12

1.08 1.09 1.10 1.09 1.01 1.01 1.05 1.11 1.01 1.01 1.07 1.07 1.08 1.14 1.00 1.02 0.926 0.965 0.938 0.949 1.17 1.20 1.08 1.08 1.09 — 1.09 — — — — — — — — 1.05 1.04 1.03 1.01 1.01 1.18 1.06 1.06 1.17 1.16

0.00690 0.00853 0.00835 0.0106 0.0139 0.0161 0.0770 0.0639 0.0129 0.0148 0.0572 0.0669 0.0668 0.0690 0.0128 0.0182 0.0340 0.0343 0.0349 0.0325 0.152 0.117 0.0236 0.0311 0.0309 — 0.0472 0.0310 0.0308 0.0314 0.0412 — — — — 0.0102 0.0115 0.0120 0.0127 0.0149 0.0160 0.0163 0.0222 0.0962 0.100

24 48 68 188 17 41 6 95 6 22 5 22 46 73 2 24 4 100 120 144 41 137 16 24 120 125 20 5 29 50 97 169 220 388 504 17 89 136 183 64 112 160 72 96 168

a

(3.84) (3.83)

(4.00) (6.04) (5.99) (6.34)

Molal concentrations in the experimental solutions; the aluminum molalities in parentheses represent the corresponding calculated aluminate concentrations (see text). b Run carried out by increasing the temperature from 100 to 150 C (runs 2 and 3) or from 100 to 200 C (run 5). c Two injections of CO2 gas (10 bar) at this point. d Injection of CO2 gas (10 bar) at 25 C; these values were not included in the solubility quotients calculation. e In this case the temperature was ramped from 50 to 200 C. f Time of sampling after reaching thermal equilibrium.

Dawsonite synthesis and solubility measurements

H2, Pt | HCl, NaCl | | NaX, NaCl, dawsonite | Pt, H2 Reference

Test

where NaX represents NaOH or a mixture of either NaOH/ Na2CO3 or NaOH/NaHCO3. Two runs (6 and 7) were performed in acidic media with injection of CO2 where the starting test solution contained HCl, NaCl and dawsonite (see Table 1). After equilibration at temperature, the cell responded in an perfectly Nernstian manner to the half-cell reaction, H2,g  2 H+ + 2e. Since the cell compartments share a common hydrogen fugacity, the potential between the electrodes is defined as: DE ¼ ðRT =F Þ lnðaHþtest =aHþref Þ þ Elj

ð5Þ

where DE is the measured cell potential, Elj, is the liquidjunction potential, and F, R, and T represent the Faraday constant, the universal gas constant and the temperature in kelvin, respectively. A solution of known pHm (”log mH+, where mH+ is the molality of hydrogen ion) is used in the reference cup and an equivalent amount of inert salt is added to the test cup, such that differences in the activity coefficients and liquid-junction potentials (Elj) between the solutions are minimized. Consequently, the molal hydrogen ion concentration in NaCl–NaOH–Na2CO3– NaHCO3 solutions can be calculated very accurately (to 0.001 pHm units) over wide ranges of temperature and ionic strength from the dissociation constants of water in NaCl media measured by Busey and Mesmer (1978) and the firstand second-ionization constants of carbonic acid determined by Patterson et al. (1982) and Patterson et al. (1984), respectively. Note that the convention used in our laboratory is that H+ is not associated with the medium ions and any ion pairing is treated implicitly by the activity coefficient model employed. In all experiments, ca. 100 g of test solution was first allowed to equilibrate thermally with ca. 1.0–1.5 g of synthetic dawsonite (see Table 1). Subsequently, the cell potential, pressure and temperature were recorded, and up to 10 aliquots of the test solution were withdrawn over a period of 2 weeks. During each sampling episode, the stirring motor was turned off (ca. 30 min) to allow the solid to settle, whereupon approximately 1 mL of test solution was withdrawn through a platinum dip tube gold-welded to a platinum frit (to prevent particles from entering the platinum tube that could act as seeds for precipitation during the sampling process) then through a 13 mm, 0.2 lm PVDF filter (Acrodisc LC13) and discarded. Two additional samples were then collected into sterilized, polypropylene/polyethylene syringes. The first sample of 5–10 g was transferred into 0.5–1 g of 1 mol Æ L1 HCl for subsequent analyses of aluminum by spectrophotometry (Dougan and Wilson, 1974) and sodium by either flame-emission AA (Perkin-Elmer 3110) or ICPAES (Thermo Jarrell Ash, IRIS). In duplicate analyses of

the samples and standards, a precision better than 2% was achieved routinely. Detection limits for the spectrophotometric analyses of aluminum were 50–300 ppb and 1– 10 ppm for Na analyses by ICP-AES. The second sample (1–2 g) was collected and immediately injected into a septum-capped Vacutainer. Prior to sample injection, the Vacutainer was first evacuated and filled with approximately 0.1 g of anhydrous H3PO4. Once the sample was acidified, dissolved CO2 was released into the head-space of the Vacutainer. The head-space CO2 was later transferred quantitatively into a liquid-nitrogen cold trap in a vacuum line via a hypodermic needle. The Vacutainer was agitated for a few minutes during the transfer to ensure complete degassing and collection of CO2. The CO2 collected was separated cryogenically from water and transferred into a calibrated volume where the quantity of gas was measured manometrically to a precision of 1–2%. The overall accuracy of the analysis of dissolved CO2 in the solution samples is estimated to be ±10%. Using the same technique, 19.4 mg of Dawsonite-A was dissolved into anhydrous phosphoric acid. The reaction yielded 1.35 · 104 moles of CO2, in agreement with the theoretical value. 4. RESULTS Results of the solubility measurements carried out from 50 to 200 C in basic solutions, either in dilute NaOH, a mixture of sodium carbonate, or sodium bicarbonate/ NaOH, are reported in Table 2. These data reveal that dawsonite dissolution is incongruent because the measured dissolved aluminum to total inorganic carbon concentration ratio is far lower than the 1:1 stoichiometry expected from Eq. (4). Note that the sodium concentration is fixed in these

1

run 1 in NaOH run 2 in Na2CO3/NaOH

0

run-4 in Na2CO3/NaOH

run-3 in Na2CO3/NaOH run-5 in NaHCO3/NaOH

-1

Bayerite-run B1 in NaOH Bayerite-run B2 (in NaHCO3)

-2

log[ΣAl]

50 to 200 C in basic media, either pure NaOH solutions or a mixture of sodium carbonate or sodium bicarbonate/ NaOH and 1.0 mol Æ kg1 NaCl. The initial cell configuration in a typical experiment was as follows:

4445

-3 -4 boehmite solubility (Palmer et al., 2001)

-5 -6

bayerite solubility

-7 -8

t = 100ºC

-9 2

3

4

5

6

7

8

-log[H]+

9

10

11

12

Fig. 4. Solubility of dawsonite at 100 C, I = 1.1 mol Æ kg1, where the solubility curve for bayerite was derived from the logQs4bayerite determined in this study (see Table 4) and the aluminum hydrolysis constants derived from Palmer et al. (2001).

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

4446

experiments by addition of NaCl to control the ionic strength and maintain a known excess of sodium. The logarithm of the measured Al concentration for runs 1–5, carried out at 100 C, is shown in Fig. 4 as a function of the measured pHm. The measurements were reproducible, even though different batches of dawsonite were used (see Table 1) and the solutions remained oversaturated with respect to boehmite (Palmer et al., 2001), which is a relatively stable aluminum phase at temperatures >90 C. Although diaspore is reported to be the most stable aluminum hydroxide under our experimental conditions (e.g., Essene et al., 1994), it is seldom observed, presumably due to kinetic constraints. It will not be considered further in this paper. In our experiments the solution chemistry is, in fact, controlled by the equilibrium between dawsonite and bayerite, Al(OH)3(cr), according to the reaction: NaAlCO3 ðOHÞ2ðcrÞ þ H2 OðlÞ  AlðOHÞ3ðcrÞ þ HCO3  þ Naþ

ð6Þ

Equilibrium is obtained very rapidly with the Emf values stabilizing within the first few hours of reaching thermal equilibrium. The presence of bayerite in these experiments

was confirmed by XRD (Fig. 5a) and SEM (Fig. 5b). The latter is an image of the solid recovered after a run carried out at 100 C, showing that the well-crystallized bayerite exists as small rhombic-shaped crystals. In order to extract the solubility product of dawsonite from these data, according to the reaction: NaAlCO3 ðOHÞ2ðcrÞ þ 2H2 OðlÞ  AlðOHÞ4  þ HCO3  þ Naþ þ Hþ

ð7Þ

knowledge of the solubility of bayerite is required. Unfortunately, such data were unavailable at the conditions of our experiments, ca. 100 C, so that it became necessary to measure the solubility of bayerite as part of this effort. Bayerite, synthesized according to the method described in Palmer et al. (2003), was equilibrated at 100 C in 1 · 104 mol Æ kg1 NaOH and 1.0 mol Æ kg1 NaCl solutions (the same apparatus, reference solution, see Table 1, and analytical methods were used as in run 1 for dawsonite). The results are reported as the logarithm of the measured Al concentration in Fig. 4 (star symbols, run B1) and as a function of the equilibration times in Table 3. From these measurements we calculated the solubility quotient of the following reaction:

Fig. 5. XRD diffractogram (a) and SEM photomicrograph (b) of recovered solids showing the bayerite phase.

Dawsonite synthesis and solubility measurements Table 3 Experimental results for the bayerite solubility experiments in NaOH (run B1) and in 0.1 mol Æ kg1 NaHCO3 (run B2) at 100 C Run No.

log½Hþ meas a

log½RAlmeas a

Time (h) of sampling

B1 B1 B1 B1 B1 B1 B1 B2 B2 B2 B2

7.679 7.665 7.683 7.718 7.740 7.733 7.938 7.938 7.921 7.853 7.825

3.96 4.04 4.24 4.13 4.15 4.09 4.10 3.83 3.83 3.84 3.84

2b 30 37 54 82 181 251 64 93 142 160

4447

1 bayerite solubility

run-5 in NaHCO3/NaOH (0.04/0.001)

0

run-6 in HCl (injection of CO2 at 100ºC) run-7 in HCl (injection of CO2 at 25ºC)

-1

run-8 in Na2CO3/NaOH (Daw-D)

log[ΣAl]

-2 -3 -4 boehmite solubility

-5 -6 injection of CO2

-7

a

Molal concentrations in the experimental solutions. Not included in calculations of the solubility product of bayerite due to the short equilibration time. b

-8

t = 100ºC

-9 2



3

4

5

6

7

+

8

9

10

AlðOHÞ3ðcrÞ þ H2 OðlÞ  AlðOHÞ4 þ H

ð8Þ

The dashed line in Fig. 4 was subsequently generated by combining the bayerite solubility quotient at 100 C, reported in Table 4, with the aluminum hydrolyses constants quoted by Palmer et al. (2001). By combining data for reactions (6) and (8) it was then possible to calculate the solubility quotient of reaction (7) at 100 C. The results of an additional experiment performed with bayerite in 0.1 mol Æ kg1 NaHCO3 (run B2) are reported in Table 3 and Fig. 4. This experiment was our first attempt to reverse the approach to equilibrium from oversaturated conditions with respect to dawsonite. The analyses (SEM and XRD) of the solid recovered after seven days did not reveal the presence of dawsonite and the aluminum concentrations remained stable during the experiments, in good agreement with the independent bayerite solubility experiments (run B1, Table 3). Attempts were also made to measure the solubility of dawsonite in acidic media at 100 C, as shown in Fig. 6 (runs 6 and 7 for which the starting concentrations are shown in Table 1). In run 6, the pH shifted immediately to ca. 7.7. Presumably, reaction (6) is responsible for this

Table 4 Solubility quotients of bayerite and boehmite at 1.1 mol Æ kg1 ionic strength t C

logQs4-bayerite reaction (8)

log Qs4bayerite reaction (8)a

log Qs4boehmite reaction (11)b

50 75 100 150 200

13.59c 12.77c 11.8 ± 0.3d — —

13.57 ± 0.08 12.70 ± 0.08

— — — 11.47 10.62

a

12

Fig. 6. Solubility of dawsonite at 100 C, I = 1.1 mol Æ kg1, showing the solubility reversal following CO2 injection (run 6) and the run performed in acidic media with injection of CO2 at 25 C (run 7).

pH buffering effect through the dissolution of dawsonite to produce bayerite. We took advantage of this system response and made a second attempt to precipitate dawsonite by injecting CO2(g) twice (10 bars, each time) into the cell (noting that dissolution of CO2 should have no measurable effect on the pHm of the acidic reference solution). This resulted in a decrease in pHm to a minimum value of 5.3 (see arrows in Fig. 6), corresponding to a reversal of the equilibrium solubility (viz., approach to equilibrium from oversaturation). As can be seen from Fig. 6, the resulting aluminum concentration follows the solubility curve of bayerite. In this run where pHm reached lower values, it was necessary to calculate the contribution of AlðOHÞ3 0 to the total aluminum concentration. Therefore, the hydrolysis constant (Palmer et al., 2001; Be´ne´zeth et al., 2001) of the following reaction AlðOHÞ3 0 þ OH  AlðOHÞ4 

— —

11

-log[H ]

þ

Values calculated with the Pitzer equation given by Palmer et al. (2003) (which corresponds to a fit of all their data as a function of temperature and ionic strength). b Palmer et al. (2001). c From Eq. (12). d Mean value from data in Table 3 excluding the first equilibration.

ð9Þ

was considered in the speciation calculation. The sum of the concentrations of the two aluminium species are shown in Table 2, but the contribution of AlðOHÞ3 0 is small in all cases. In a separate experiment (run 7), CO2 was injected at the beginning of the run at 25 C (12 bar) and the vessel was then heated to 100 C. In this case, the dissolved aluminum, reported in Fig. 6 (hexagonal symbols), was first oversaturated with respect to bayerite but then decreased with time. Note that for this particular run, no solubility quotients were calculated due to the lack of sufficient analytical data. At 150 and 200 C, dawsonite was no longer in metastable equilibrium with bayerite, instead it was in equilibrium with boehmite, according to reaction (10), as shown for example at 150 C in Fig. 7 (runs 2, 3 and 8) and confirmed by XRD of the recovered solid (cf., Fig. 8). NaAlCO3 ðOHÞ2ðcrÞ  AlOOHðcrÞ þ HCO3  þ Naþ

ð10Þ

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

4448 1

run 2 in Na2CO3/NaOH 0

run-3 in Na2CO3/NaOH run-8 in Na2CO3/NaOH

-1

log[ΣAl]

-2 -3 bayerite solubility

-4 -5 -6 boehmite solubility

-7 -8

Fig. 9. SEM photomicrograph of recovered solid after a 150 C run showing boehmite pseudomorphing dawsonite.

t = 150 ºC

-9 2

3

4

5

6

7

8

9

10

11

12

+

-log[H ] Fig. 7. Solubility of dawsonite at 150 C, I = 1.1 mol Æ kg1.

An SEM micrograph of this solid is reproduced in Fig. 9 showing that boehmite forms as a pseudomorph of dawsonite. Therefore, at these temperatures, the dissolution equilibrium for boehmite AlOOHðcrÞ þ 2H2 OðlÞ  AlðOHÞ4  þ Hþ

ð11Þ

was taken into account using the appropriate solubility quotients reported by Palmer et al. (2001) and given in Table 4, in order to derive the solubility quotient for reaction (7). An additional experiment (run 8) was performed at lower temperatures, starting at 50 C, and then increasing

stepwise, after equilibrium was attained, to 75, 100 and 150 C. The solubility products of bayerite (reaction (8)) are available from the results of our previous measurements of the system Na–NO3–OH–Al(OH)4 at 30, 63 and 90 C at ionic strengths from 0.1 to 5.1 mol Æ kg1 NaNO3 (Palmer et al., 2003). The bayerite solubility data obtained at ionic strengths between 1.0 and 1.1 mol Æ kg1 were fitted as a function of reciprocal temperature (see Fig. 10), from which we obtained the following equation: log Qs4bayerite ¼ ð1:45  0:08Þ  ð3923  27Þ=T ðKÞ

ð12Þ

The mean value obtained in the current study at 100 C is also presented in Fig. 10 and is in good agreement with the regression. We assume that the solubility of gibbsite in sodium chloride and nitrate solutions will be the same

Fig. 8. XRD diffractogram of the recovered solid after a 150 C run.

Dawsonite synthesis and solubility measurements

-11

25

50

100 t ºC

Palmer et al. (2003) at I=1.0-1.1 linear regression (Equation 12) this work

-12

logQs4-bayerite

75

-13

-14

-15 3.4

3.3

3.2

3.1

3.0

2.9

2.8

2.7

2.6

1000/ T(K) 1

Fig. 10. logQs4 of bayerite at I = 1.1 mol Æ kg function of the reciprocal of temperature.

ionic strength as a

4449

with only slight differences in the activity coefficients of the dissolved aluminum species at equivalent ionic strengths. The values obtained from Eq. (12) at 50 and 75 C are also reported in Table 4 together with values calculated with the Pitzer equation given by Palmer et al. (2003) and the solubility products of boehmite given by Palmer et al. (2001) for t>100 C, all at 1.1 mol Æ kg1 ionic strength. The solubility products derived for reactions (6) and (11), as well as Qs4 (= ½AlðOHÞ4  ]½HCO3  ½Naþ ½Hþ ), reaction (7), are reported in Table 5. The solubility products (Qs4) for dawsonite can then be expressed as solubility constants (Ks4) by consideration of the appropriate activity coefficients and water activities such that: K s4 ¼ Qs4 cAlðOHÞ4  cHCO3  cNa þ cH þ =a2w

ð13Þ

where cn are the activity coefficients of the aqueous species n, and aw is the activity of water. In order to treat the

Table 5 Dawsonite solubility quotients, Qs4, ionic strength, I, activity coefficients, c, activity of water, aw, and dawsonite solubility constants, Ks4, at each temperature investigated in this study Run no.

t C

logQs reaction (6) dawsbayerite

logQs reaction (10) dawsboehmite

logQs4 reaction (7) ± 0.11a

I mol Æ kg1

c±(NaCl)

aw

logKs4 reaction (7) ± 0.11a

1 1 1 1 2 2 2 2 3 3 3 3 3 3 4 4 5 5 5 5 5 5 6b 6c 6c 6d 8 8 8 8 8 8 8 8 8 8

100.1 100.1 100.1 100.1 100.1 100.1 150.1 150.1 100.1 100.1 150.1 150.1 150.1 150.1 99.8 99.8 99.8 99.8 99.8 99.8 200.7 200.7 99.8 99.8 99.8 99.8 50.1 50.1 50.1 50.2 75.1 75.1 75.2 100.3 150.1 150.0

2.30 2.20 2.19 2.04 2.14 2.05 — — 2.18 2.11 — — — — 2.18 1.99 1.55 1.52 1.53 1.55 — — 1.62 1.93 1.89 1.98 2.67 2.61 2.56 2.52 2.24 2.15 2.16 1.87 — —

— — — — — — 1.17 1.20 — — 1.29 1.22 1.19 1.19 — — — — — — 0.79 0.91 — — — — — — — — — — — — 1.02 1.01

14.15 14.05 14.04 13.88 13.98 13.89 12.64 12.67 14.03 13.96 12.77 12.69 12.66 12.65 14.03 13.84 13.40 13.37 13.37 13.40 11.41 11.53 13.47 13.77 13.74 13.83 16.24 16.18 16.14 16.09 14.94 14.85 14.86 13.72 12.48 12.47

1.09 1.09 1.10 1.09 1.02 1.02 1.06 1.12 1.02 1.02 1.08 1.08 1.09 1.14 1.01 1.03 0.929 0.968 0.941 0.951 1.17 1.20 1.08 1.08 1.09 1.09 1.06 1.05 1.04 1.02 1.02 1.19 1.07 1.07 1.18 1.17

0.623 0.623 0.622 0.622 0.623 0.623 0.561 0.560 0.623 0.623 0.561 0.561 0.560 0.559 0.623 0.623 0.624 0.624 0.624 0.624 0.476 0.473 0.623 0.622 0.622 0.622 0.657 0.657 0.657 0.657 0.644 0.644 0.644 0.622 0.559 0.559

0.964 0.964 0.963 0.964 0.966 0.966 0.966 0.964 0.966 0.966 0.965 0.965 0.965 0.963 0.967 0.966 0.969 0.968 0.969 0.969 0.964 0.963 0.964 0.964 0.964 0.964 0.965 0.965 0.965 0.966 0.966 0.960 0.964 0.965 0.962 0.962

14.94 14.84 14.84 14.68 14.78 14.69 13.61 13.64 14.83 14.75 13.74 13.66 13.63 13.63 14.82 14.63 14.19 14.16 14.17 14.19 12.69 12.79 14.26 14.56 14.53 14.62 16.94 16.88 16.84 16.79 15.67 15.57 15.59 14.51 13.46 13.45

a b c d

Calculated from the combined experimental uncertainties (3r). Before injection of CO2 (see Table 2). First injection of CO2. Second injection of CO2.

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

4450

function of reciprocal T. In order to create a model that is consistent with the heat capacity of dawsonite (C op 298:15 ¼ 142:6  0:3 J  mol1  K1 Þ measured by Ferrante et al. (1976), we first combined this value with the heat capacities of the aqueous species involved in Eq. (7) as reported in Table 6, to derive a DC p;r o for reaction (7) of 185.5 J Æ mol 1 Æ K1. Of all the temperature functions tested, the simplest equation giving a reliable fit of our data has the form shown in Eq. (15)

experimental data for the dissolution of dawsonite according to reaction (7), it was necessary to adopt an iterative procedure to resolve the pHm, ionic strength and carbonate speciation based on the measured total aluminum, carbonate and sodium concentrations. The first- and second-ionization constants of carbonic acid required in the iterative calculation were taken from Patterson et al. (1982) and Patterson et al. (1984), respectively. The mean stoichiometric activity coefficients were derived from the Meissner equation (Lindsay, 1989) with the implicit assumption that for an ion of charge, z: 2

cjzj ¼ czðNaClÞ

log K s4 ¼ a þ b=T ðKÞ þ cT ðKÞ

where c is fixed and set equal to 185.5/(2RT ln(10))= 0.01625. A regression then yielded the two coefficients: a = 8.797 and b = 6510.1 corresponding to the solid curve shown in Fig. 11. This equation yields DGr o ¼ ðR lnð10ÞfaT þ b þ cT 2 gÞ=1000 ¼ 102:1 kJ  mol1 ; DH or ¼ 1 2 ðR lnð10Þfb þ cT gÞ=1000 ¼ 97:0 kJ  mol , and consequently DSr o ¼ 17:1J  mol1  K 1 at 25 C. Finally, taking these values and those of the aqueous species in Table 6, we obtained Df Go298:15 ¼ 1782  2 kJ  mol1 , Df H o298:15 ¼ 1960  7 kJ  mol1 , and S o298:15 ¼ 131  2 J  mol1  K 1 for dawsonite (uncertainties are 3r) that are within the experimental uncertainties of the values reported by Ferrante et al. (1976). Note that the large uncertainty calculated for the enthalpy of formation (± 7 kJ Æ mol1) is mainly due to the uncertainty assigned to the Df H o298:15 of the aluminate ion by Be´ne´zeth et al. (1997). Having established an improved set of thermodynamic data for dawsonite one can now estimate its relative stability with respect to other Al- and Na-bearing phases in geological environments where CO2 injection could be considered. Of the multiple sets of phases and boundary conditions that could be considered, we chose to test cases at 100 C involving either bayerite or boehmite, as well as quartz and albite for their ubiquity in sandstone aquifers. This temperature is not only the upper limit for the detection of bayerite in our study, but is also the condition at which boehmite forms. Other phases whose stability relationships could also affect the formation of dawsonite (e.g., analcime, NaAlSi2O6 Æ H2O, Na-montmorillonite or kaolinite, etc.) were not considered, as a full analysis of the potential reactions in these systems is beyond the scope of this study. We also investigated the effect of total inorganic carbon on the stability fields of the respective solid

ð14Þ

where c±(NaCl) is the mean molal stoichiometric activity coefficient of NaCl. The latter values and the corresponding activities of water were taken from Archer (1992) and are reported in Table 5 together with the ionic strengths and the calculated solubility constants (Ks4). The validity of the above assumption is supported by the fact that the corresponding boehmite solubility quotients (reaction 11, Palmer et al., 2001) yielded infinite dilution solubility constants that differed by only 0.05–0.15 log units (150–200 C) when applying the activity coefficients and activities of water calculated from Archer (1992) up to 1.1 mol Æ kg1 ionic strength. The logarithms of the solubility constants taken from Table 5 are reported in Fig. 11 (open triangles) as a -11

300 250

200

150

100

-12

50

25 t ºC

experimental results Equation (15)

logKs4-dawsonite

-13 -14 -15 -16 -17 -18 -19 1.6

1.8

2.0

2.2

2.4 2.6 2.8 1000/T (K)

3.0

3.2

3.4

ð15Þ

3.6

Fig. 11. logKs4dawsonite as a function of the reciprocal of temperature with the experimental uncertainty, 3r, shown as dotted lines.

Table 6 Standard state properties of dawsonite and other aqueous species involved in reaction (7) at 25 C and 1 bar Species

a

DfGo

b o

a

NaAlCO3(OH)2(cr) NaAlCO3(OH)2(cr) AlðOHÞ4 

1786 ± 4 1782 ± 2 (1) 1305.7

132 ± 2 131 ± 2 (1) 103.7

Na+ HCO3 

261.91 (1) 586.77

59.0 (1) 91.2

H2O(l)

237.13

69.95

S 298:15

DfHo

b

References

1964 ± 4 1960 ± 7 (1) 1503.0

142.6 ± 0.3 142.6 ± 0.3 (2) 96.5

(1)

240.12 691.99

46.4 (2) 35.40

285.83

75.19

Ferrante et al. (1976) This study (1) Be´ne´zeth et al. (2001) and (2) Hovey et al. (1988) Wagman et al. (1982) (1) Wagman et al. (1982) (2) Shock and Helgeson (1988) Wagman et al. (1982)

C op298:15

All of the uncertainties are quoted to 3r [In order to be conservative we used the maximum uncertainty recommended by Wagman et al. for the thermodynamic quantities in their tables]. a In kJ Æ mol1. b In J Æ K1 Æ mol1.

Dawsonite synthesis and solubility measurements

bonate species’ boundaries. For clarity, they are not repeated in Figs. 12 and 13b where, instead, the domains of the aqueous aluminum species are illustrated. Note that the AlðOHÞ2 þ species is not displayed in these figures because its stability domain is very narrow under these conditions (Palmer et al., 2001). Several conclusions can be drawn from these diagrams. First, in the absence of silica, Figs. 12 and 13a illustrate that dawsonite is stable over a wide pH range, flanked by the a and f boundaries above log[C] >2 and 2.6, respectively. Similarly, with respect to the aluminum concentration,

phases with respect to pH as well as at varying total aluminum concentrations at a given inorganic carbon concentration (fixed here arbitrarily at 0.1 mol Æ kg1). The total sodium concentration was also fixed arbitrarily at 1.0 mol Æ kg1 in each calculation such that these conditions, albeit in the absence of silica, approximate those prevailing in our experiments. The phase or predominance diagrams, based on the thermodynamic data derived in this study, are shown in Figs. 12a and b for boehmite and Figs. 13a and b for bayerite. In Figs. 12 and 13a, the vertical short-dashed lines represent the aqueous car-

0

boehmite quartz

a

dawsonite albite quartz c

-1

b

log[carbon]

100 C [Na+] = 1.0 m

dawsonite e

boehmite

-3

-4

f

d

o

-2

HCO3

CO2(aq)

2

3

4451

4

5

6

7

8

-

CO3

9

10

2-

11

12

pH 0 o

-2

log[aluminum]

boehmite

100 C [Na+] = 1.0 m [C] = 0.1 m a

-1

-3

f

e

dawsonite

-4

boehmite

-5

d

-6 -7 Al(OH)

-8

3+

Al

2+

b c -

Al(OH)4

Al(OH)3(aq)

-9 2

3

4

5

6

7

8

9

10

11

12

pH Fig. 12. Stability diagram (a) log[C] versus pH and (b) log[Al] versus pH in the presence of boehmite at an arbitrarily fixed [Na+] of 1 mol Æ kg1 and for (b) [C] = 0.1 mol Æ kg1. Minerals in italics represent those present when silica is in the system. The phase boundaries correspond to the following equilibria: for 12(a): a, AlOOH(cr) + CO2(aq)+Na+ + H2O(l)  NaAlCO3(OH)2(cr) +H+; b, NaAlSi3O8(cr) + CO2(aq) + H2O(l)  NaAlCO3(OH)2(cr) + 3SiO2(cr); c, NaAlSi3O8(cr) + HCO3  + H+  NaAlCO3(OH)2(cr) + 3SiO2(cr); d, AlOOH(cr) + Na+ + 3SiO2(cr)  NaAlSi3O8(cr) + H+; e, AlOOH(cr) + HCO Na+  NaAlCO3(OH)2(cr); f, AlOOH(cr) + CO3 2 + Na+ + 3 + H+  NaAlCO3(OH)2(cr). For 12(b): a, AlOOH(cr) + CO2(aq) + Na+ + H2O(l)  NaAlCO3(OH)2(cr) + H+; b, Al(OH)3(aq) + Na+ + CO2(aq)  NaAlCO3(OH)2(cr) + H+; c, AlðOHÞ4  + Na+ + CO2(aq)  NaAlCO3(OH)2(cr) + H2O(l); d, AlðOHÞ4  + Na+ + HCO3  + H+  NaAlCO3(OH)2(cr) + 2H2O(l); e, AlðOHÞ4  + Na+ + CO3 2 + 2H+  NaAlCO3(OH)2(cr) + 2H2O(l); f, AlOOH(cr) + Na+ + CO3 2 + H+  NaAlCO3(OH)2(cr).

4452

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

0

albite

dawsonite quartz

bayerite quartz

c

a

o

log[carbon]

-1

100 C + [Na ] = 1.0 m

b

f -2

d

dawsonite bayerite

-3

-4

3

4

CO32-

HCO3-

CO2(aq)

2

e

5

6

7

8

9

10

11

12

pH 0

f

-1

bayerite

dawsonite

bayerite

-2

log[aluminum]

e

a

-3 -4 -5

d

-6 -7 Al(OH)

-8 Al

3+

b

2+

c Al(OH)4

Al(OH)3(aq)

-

100 ºC + [Na ] = 1.0 m [C] = 0.1 m

-9 2

3

4

5

6

7

8

9

10

11

12

pH Fig. 13. Stability diagram (a) log[C] versus pH and (b) log[Al] versus pH in the presence of bayerite a at an arbitrarily fixed [Na+] concentration of 1 mol Æ kg1 and for (b) [C] = 0.1 mol Æ kg1. Minerals in italics represent those present when silica is in the system. The phase boundaries correspond to the following equilibria: for 13(a): a, Al(OH)3(cr) + CO2(aq) + Na+  NaAlCO3(OH)2(cr) + H+; b, NaAlSi3O8(cr) + CO2(aq) + H2O(l)  NaAlCO3(OH)2(cr) + 3SiO2(cr); c, NaAlSi3O8(cr) + HCO3  + H+  NaAlCO3(OH)2(cr) + 3SiO2(cr); d, Al(OH)3(cr) + Na+ + 3SiO2(cr)  NaAlSi3O8(cr) + H+ + H2O(l); e, Al(OH)3(cr) + HCO3  + Na+  NaAlCO3(OH)2(cr) + H2O(l); + + f, Al(OH)3(cr) + CO2  NaAlCO3(OH)2(cr) + H2O(l). for 13(b): a, Al(OH)3(cr) + CO2(aq) + Na+  NaAlCO3(OH)2(cr) + H+; 3 + Na + H b, Al(OH)3(aq) + Na+ + CO2(aq)  NaAlCO3(OH)2(cr) + H+; c, AlðOHÞ4  + Na+ + CO2(aq)  NaAlCO3(OH)2(cr) + H2O(l); + d, AlðOHÞ4  + Na+ + HCO3  + H+  NaAlCO3(OH)2(cr) + 2H2O(l); e, AlðOHÞ4  + Na+ + CO2  NaAlCO3(OH)2(cr) + 3 + 2H + + H  NaAlCO (OH) + H O . 2H2O(l); f, Al(OH)3(cr) + Na+ + CO2 3 2(cr) 2 (l) 3

Figs. 12 and 13b show that dawsonite has a lower limit of stability relative to boehmite at a pH of approximately 5.5 (Fig. 12b) compared to 4.8 relative to bayerite (Fig. 13b); the latter being metastable with respect to the former. In other words, at total carbon concentrations of P0.1 mol Æ kg1, dawsonite is stable relative to either boehmite or bayerite above boundary a and below boundary f. Second, in albite-bearing rocks in contact with saline fluids at 100 C (Figs. 12 and 13a), the acid-induced dissolu-

tion of albite (after addition of large quantities of CO2) will buffer the pH and drive it to the region where dawsonite dominates over either boehmite or bayerite. In other words, the reaction: Albite þ HCO3  þ Hþ  Dawsonite þ 3 Quartz

ð16Þ

(c boundary in Figs. 12 and 13a) sets the lower pH limit for the stability of albite with respect to dawsonite, whereas dawsonite is stable to significantly lower pH above log[C] 1. Therefore, as long as the total dissolved carbon

Dawsonite synthesis and solubility measurements

concentration is above approximately 0.1 mol Æ kg1, albite should buffer the pH in a range in which dawsonite provides an effective trap for CO2 sequestration albeit in very narrow domains of stability delimitated in Figs. 12 and 13a by the boundaries abc. It must be remembered that this assumes that fluid–rock interactions remain at equilibrium. If, on the other hand, CO2 injection rates are faster than the fluid–rock reaction kinetics, lower pH values will persist at least during and for some time after injection (depending on the distance from the injection well, flow rates, mixing efficiencies, permeability, etc.) but will ultimately return to some equilibrium leading to an increase in pH. This is consistent with observations reported by Kharaka et al. (2006) in the case of the small scale Frio Pilot Project in Texas, where a significant pH increase (from ca. 5.9 to 6.5) occurred after a pause in CO2 injection due to technical problems (booster pump failure). 5. CONCLUSIONS AND REMARKS In this study we synthesized dawsonite by optimizing the method of Van Der Heem (1980), and measured its solubility to 200 C in 1.0–1.1 mol Æ kg1 ionic strength solutions. Dawsonite dissolution occurs via an incongruent pathway, favoring bayerite formation at temperatures 6100 C and boehmite formation at higher temperatures. Additional bayerite solubility experiments were performed that confirmed quantitatively the more comprehensive results of our previous study (Palmer et al., 2003). An empirical equation for the temperature dependence of our dawsonite solubility constants (Eq. (15)) was derived, constrained by the value of the heat capacity of the solid at 25 C reported by Ferrante et al. (1976). The three parameters obtained from this regression allowed us to calculate the thermodynamic properties of dawsonite, providing refinements of the values derived by Ferrante et al. (1976) from calorimetric measurements. These data were used to construct simplified predominance diagrams that suggest that dawsonite can be stable down to slightly acidic pH values whereas the addition of pH buffering assemblages, such as silica-containing minerals (Figs. 12 and 13a), will impose tight constraints on its formation. Our synthesis protocols (and others from the literature) strongly suggest that an alkaline environment (pH >9) is required for the optimal formation of dawsonite. On the other hand, whereas dawsonite is by no means a common mineral, it is found in a number of localities (cf., www.mindat.org), and is widespread in others (e.g., the Piceance Creek Basin nahcolite and dawsonite deposits, Smith and Milton, 1966; Beard et al., 1974). These facts may imply that, in addition to simple thermodynamic considerations, nucleation and other kinetic processes may impose limitations on dawsonite precipitation. Therefore, repressed nucleation may be an additional factor of importance in considering whether dawsonite will play a role in the geological sequestration of carbon dioxide. Likewise, Kaszuba et al. (2005) did not observe the formation of dawsonite during their brine–rock–CO2 experiments under conditions relevant to CO2 storage (200 C, 200 bars), but rather detected the precipitation of analcime, indicating the complexity of such multi-component systems that could not be addressed in our study.

4453

The results of our study provide a framework for future research by defining the conditions under which dawsonite will form. Follow-up work could focus on the rate of dawsonite precipitation as a function of pH at various carbonate (CO2), sodium and aluminum concentrations, conditions suitable to CO2 sequestration, and the kinetics of dawsonite formation from albite or other Na- and Albearing phases relative to the rates of conversion of supercritical to dissolved CO2. ACKNOWLEDGMENTS This research was sponsored by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Sciences, U.S. Department of Energy, under contract DE-AC0500OR22725 with Oak Ridge National Laboratory, managed and operated by UT-Battelle, LLC. Thanks to J.C. Harrichoury from the Laboratoire des Me´canismes de Transfert en Ge´ologie for his help in synthesizing dawsonite (Daws-D), and Michel Thibaut for some of the XRD analyses. This manuscript was greatly improved by thorough reviews by A. Mucci (Associate Editor), J.P. Kaszuba, K.G. Knauss and an anonymous reviewer.

REFERENCES Altman, R.L. (1982). Synthesis of dawsonites. U.S. Patent 4,356,157. ´ lvarez-Ayuso E. and Nugteren H. W. (2005) Synthesis of A dawsonite: A method to treat the etching waste streams of the aluminium anodising industry. Water Res. 39, 2096–2104. Archer D. G. (1992) Thermodynamic properties of the NaCl + H2O System. II. Thermodynamic properties of NaCl(aq), NaCl Æ 2H2O(cr), and phase equilibria. J. Phys. Ref. Data 21, 793–829. Bachu S., Gunter W. D. and Perkins E. H. (1994) Aquifer disposal of CO2: hydrodynamic and mineral trapping. Energ Convers Manage 35, 269–279. Bader E. (1938) Formation, constitution and synthetic preparation of dawsonite. Neues Jahrb. Miner. Geol. Paleont. B74A, 449– 465. Bader E. and Esch U. (1944) Experiments on the pressure synthesis of dawsonite. Zeitschr. Elektrochemie 50, 266–268. Baker J. C., Bai G. P., Hamilton P. J., Golding S. D. and Keene J. B. (1995) Continental-scale magmatic carbon dioxide seepage recorded by dawsonite in the Bowen-Gunnedah-Sydney Basis system, eastern Australia. J. Sediment Res. 65, 522–530. Beard T. N., Tait D. B. and Smith J. W. (1974) Nacholite and dawsonite resources in the Green River Formation, Piceance Creek Basin, Colorado. Rocky Mountain Association of Geologists Guidebook. Be´ne´zeth P., Palmer D. A. and Wesolowski D. J. (1997) The aqueous chemistry of aluminum. A new approach to hightemperature solubility measurements. Geothermics 26, 465–481. Be´ne´zeth P., Palmer D. A. and Wesolowski D. J. (1999) The solubility of zinc oxide at 0.03m NaTr as a function of temperature, with in situ pH measurement. Geochim. Cosmochim. Acta 63, 1571–1586. Be´ne´zeth P., Palmer D. A. and Wesolowski D. J. (2001) Aqueous high temperature solubility studies. II. The solubility of boehmite at 0.03 ionic strength as a function of temperature and pH as determined by ‘‘in situ’’ measurements. Geochim. Cosmochim. Acta 65, 2097–2111. Bergman P. D. and Winter E. M. (1995) Disposal of CO2 in aquifers in the U.S.. Energ Convers Manage 36, 523–526.

4454

P. Be´ne´zeth et al. / Geochimica et Cosmochimica Acta 71 (2007) 4438–4455

Besson H., Caillere S., Henin S. and Prost R. (1973) Formation expe´rimentale et conditions de gisement de la dawsonite. C.R. Acad. Sci. Ser. D277, 261. Blum A. E. and Stillings L. L. (1995) Feldspar dissolution kinetics. In Chemical Weathering Rates of Silicate Minerals, 31 (eds. A. F. White and S. L. Brantley). Rev. Min. Soc. Amer., Washington, DC, pp. 291–351. Busey R. H. and Mesmer R. E. (1978) Thermodynamic quantities for the ionization of water in sodium chloride media to 300 C. J. Chem. Eng. Data 23, 175–176. Carroll S. A. and Knauss K. G. (2005) Dependence of labradorite dissolution kinetics on CO2(aq), Al(aq), and temperature. Chem. Geol. 217, 213–225. Corraza E., Sabelli C. and Vannucci S. (1977) Dawsonite: new mineralogical data and structure refinement. Neues Jahrbuch fu¨r Mineralogie Monatshefte 9, 381–397. Coveney R. M. and Kelly W. C. (1971) Dawsonite as a daughter mineral in hydrothermal fluid inclusions. Contr. Mineral. Petrol. 32, 334–342. Dougan W. K. and Wilson A. L. (1974) The absorptiometric determination of aluminum in water. A comparison of some chromogenic reagents and the development of an improved method. Analyst 99, 413–430. Essene E. J., Anovitz L. M. and Perkins D. (1994) Mineral metastability in the system Al2O3-SiO2- H2O: A reply. Clays and Clay Miner. 42, 98–105. Ferrante M. J., Stuve J. M. and Richardson D. W. (1976). Thermodynamic data for synthetic dawsonite. U.S. Bureau of Mines Report Investigation, 8129, Washington, D.C., 13p. Frueh A. J. and Golightly J. P. (1967) The crystal structure of dawsonite NaAl(CO3)(OH)2. Can. Miner. 9(Part 1), 51–56. Furmakova L. N. (1981) The question of formation conditions for dawsonite. Lithol. and Miner. Res. 15, 565–572. Gaus I., Azaroual M. and Czernichowski-Lauriol I. (2003). Reactive transport modeling of dissolved CO2 in the Cap Rock Base during CO2 sequestration (Sleipner Site, North Sea). Second Annual Conference on Carbon Sequestration, May 5–8, 2003, Alexandria, Virginia, USA, p. 1–13. Gaus I., Azaroual M. and Czernichowski-Lauriol I. (2005) Reactive transport modeling of the impact of CO2 injection on the clayey cap rock at Sleipner (North Sea). Chem. Geol. 217, 319–337. Gautier J. M., Oelkers E. H. and 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, 4549–4560. Golubev S. V., Pokrovsky O. S. and 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. Gunter W. D., Wiwchar B. and Perkins E. H. (1997) Aquifer disposal of CO2-rich greenhouse gases: extension of the time scale of experiment for CO2-sequesttering reactions by geochemical modeling. Mineral. Petrol. 59, 121–140. Harrington B. J. (1874) Notes on dawsonite, a new carbonate. Canadian Naturalist New Ser. 7, 305–309. Hay R. L. (1963) Zeolitic weathering in Olduvai Gorge, Tanganyika. Geol. Soc. Am. Bull. 74, 1281–1286. Hellmann R. and Tisserand D. (2006) Dissolution kinetics as a function of the Gibbs free energy of reaction: An experimental study based on albite feldspar. Geochim. Cosmochim. Acta 70, 364–383. Herold G. (1992) Process for producing dawsonite. US Patent 5,078,983. Hite R. J. and Dyni J. R. (1967) Potential resources of dawsonite and Nahcolite in the Piceance Creek Basin, Northwest Colorado. Q. Colorado School of Mines 62, 25–38.

Holloway S. (2005) Underground sequestration of carbon dioxide – a viable greenhouse gas mitigate option. Energy 30, 2318– 2333. Hovey J. K., Hepler L. G. and Tremaine P. R. (1988) Thermodynamics of aqueous aluminate ion: Standard partial molar heat capacities of AlðOHÞ4  from 10 to 55 C. J. Phys. Chem. 92, 1323–1332. Jackson J. Jr., Huggins C. W. and Ampina S. G. (1972) Synthesis and characterization of dawsonite. U.S. Bureau of Mines Report Investigation, 8129, Washington, D.C., 19p. Johnson J. W., Nitao J. J., Steefel C. I. and Knauss K. G. (2001). Reactive transport modeling of geologic CO2 sequestration in saline aquifers: the influence of intra-aquifer shales and the relative effectiveness of structural, solubility, and mineral trapping during prograde and retrograde sequestration. NETL Publications 2001 Conference Proceedings, First National Conference on Carbon Sequestration. Johnson J. W., Nitao J. J. and Steefel C. I. (2002) Fundamental elements of geologic CO2 sequestration in saline aquifers. ACS Fuel Chemistry Division Symposia Preprints 47, 41–42. Kaszuba J. P., Janecky D. R. and Snow M. G. (2003) Carbon dioxide reaction processes in a model brine aquifer at 200 C and 200 bars: implications for geologic sequestration of carbon. Appl. Geochem. 18, 1065–1080. Kaszuba J. P., Janecky D. R. and Snow M. G. (2005) Experimental evaluation of mixed fluid reactions between supercritical carbon dioxide and NaCl brine: Relevance to the integrity of a geologic carbon repository. Chem. Geol. 217, 277–293. Kaufman J. L. (1984). Preparation of dihydroxylaluminium sodium carbonate. U.S. Patent 4,438,085. Kharaka Y. K., Cole D. R., Thordsen J. J., Kakouros E. and Nance H. S. (2006) Gas–water–rock interactions in sedimentary basins: CO2 sequestration in the Frio formation, Texas, USA. J. Geochem. Explor. 89, 183–186. Knauss K. G., Johnson J. W. and 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. Lindsay, Jr., W. T. (1989) Chemistry of steam cycle solutions: Principles. In The ASME Handbook on Water Technology for Thermal Power Plants (ed. P. Cohen). The American Society of Mechanical Engineers, p. 483 (Chapter 7). Marchetti C. (1977) On Geoengineering and the CO2 Problem. Climate Change 1, 59–68. Mesmer R. E., Palmer D. A. and Wesolowski D. J. (1995) Potentiometric studies at ORNL with hydrogen-electrode concentration cells. In Proc. 12th Int. Conf. Properties of Water and Steam (eds. H. J. White, J. V. Sengers, D. B. Neumann and J. C. Bellows). Begell House, New York, pp. 423–4331. Misra C. (1980) Process for the production of a basic carbonate of aluminum and an alkali or alkali earth metal. US Patent 4,238,458. Moore J., Adams M., Allis R., Lutz S. and Rauzi S. (2005) Mineralogical and geochemical consequences of the long-term presence of CO2 in natural reservoirs: An example from Springerville-St. Johns Field, Arizona, and New Mexico, USA. Chem. Geol. 217, 365–385. Oelkers E. H., Schott J. and Devidal J. L. (1994) The effect of aluminum, pH, and chemical affinity on the rates of aluminosilicate dissolution reactions. Geochim. Cosmochim. Acta 58, 2011–2024. Palmer D. A. and Wesolowski D. J. (1993) Aluminium speciation and equilibria in aqueous solution: III. Potentiometric determination of the first hydrolysis constant of aluminum(III) in sodium chloride solutions to 125 C. Geochim. Cosmochim. Acta 57, 2929–2938.

Dawsonite synthesis and solubility measurements Palmer D. A., Be´ne´zeth P. and Wesolowski D. J. (2001) Aqueous high temperature solubility studies. I. The solubility of boehmite at 150 C as a function of ionic strength and pH as determined by ‘‘in situ’’ measurements. Geochim. Cosmochim. Acta 65, 2081–2095. Palmer D. A., Be´ne´zeth P., Wesolowski D. J. and Hilic S. (2003) Experimental study of the dissolution of aluminum phases as a function of temperature, caustic concentration and additives. Light Metals, 5–10. Patterson C. S., Slocum G. H., Busey R. H. and Mesmer R. E. (1982) Carbonate equilibria in hydrothermal systems: First ionization of carbonic acid in NaCl media to 300 C. Geochim. Cosmochim. Acta 46, 1653–1663. Patterson C. S., Busey R. H. and Mesmer R. E. (1984) Second ionization of carbonic acid in NaCl media to 250 C. J. Solution Chem. 13, 647–661. Pokrovsky O. S., Golubev S. V. and Schott J. (2005) Dissolution kinetics of calcite, dolomite and magnesite at 25 C and 0 to 50 atm pCO2. Chem. Geol. 217, 239–255. Robie, R. A. and Hemingway, B. S. (1995). Thermodynamic properties of minerals and related substances at 298.15 K and 1 bar (105 Pascals) pressure and at higher temperatures, US Geological Survey Bulletin 2131, Washington. Shock E. L. and Helgeson H. C. (1988) Calculation of the thermodynamic and transport properties of aqueous species at high pressures and temperatures: Correlation algorithms for ionic species and equation of state predictions to 5kb and 100 C. Geochim. Cosmochim. Acta 52, 2009–2036. Sirbescu M. L. C. and Nabelek P. I. (2003) Dawsonite: An inclusion mineral in quartz from the Tin Mountain pegmatite, Black Hills, South Dakota. Am. Miner. 88, 1055–1060. Smith J. W. and Milton C. (1966) Dawsonite in the Green River formation of Colorado. Econ. Geol. 61, 1029–1042. Stevenson J. S. and Stevenson L. S. (1965) The petrology of dawsonite at the type locality, Montreal. Can. Mineral. 8, 249– 252. Van Der Heem P. (1980) Dawsonite. U.S. Patent 4221771.

4455

Wagman D. D., Evans W. H., Parker V. B., Schumm R. H., Halow I., Bailey S. M., Churney K. L. and Nuttal R. L. (1982) The NBS tables of chemical thermodynamic properties. Selected values for inorganic C1 and C2 organic substances in SI units. J. Phys. Chem. Ref. Data 11, 1–392. Wawersik W. R., Rudnicki J. W., Dove P., Harris J., Logan J. M., Pyrak-Nolte L., Orr, Jr., F. M., Ortoleva P. J., Richter F., Warpinski N. R., Wilson J. L. and Wong T. F. (2001) Terrestrial sequestration of CO2: An assessment of research needs. Adv. Geophys. 43, 97–177. Wesolowski D. J. (1992) Aluminum speciation and equilibria in aqueous solution: I. The solubility of gibbsite in the system Na– K–Cl–OH–Al(OH)4 from 0 to 100 C. Geochim. Cosmochim. Acta 56, 1065–1091. Wesolowski D. J., Palmer D. A. and Mesmer R. E. (1995) Measurement and control of pH in hydrothermal solutions. In Proc. 8th Int. Symp. Water–Rock Interaction (eds. Y. K. Kharaka and O. V. Chudaev). A.A. Balkema, Rotterdam, pp. 1–55. White S. P., Allis R. G., Moore J., Chidsey T., Morgan C., Gwynn W. and Adams M. (2005) Simulation of reactive transport of injected CO2 on the Colorado Plateau, Utah, USA. Chem. Geol. 217, 387–405. Xu T., Apps J. A. and Pruess K. (2003) Reactive geochemical transport simulation to study mineral trapping for CO2 disposal in deep arenaceous formations. J. Geophys. Res. 108(B2), 2071. Xu T., Apps J. A. and Pruess K. (2005) Mineral sequestration of carbon dioxide in a sandstone-shale system. Chem. Geol. 217, 295–318. Zerai B., Saylor B. Z. and Matisoff G. (2006) Computer simulation of CO2 trapped through mineral precipitation in the Rose Run Sandstone, Ohio. Appl. Geochem. 21, 223–240. Zhang X., Wen Z., Gu Z., Xu X. and Lin Z. (2004) Hydrothermal synthesis and thermodynamic analysis of dawsonite-type compounds. J. Solid State Chem. 177, 849–855. Associate editor: Alfonso Mucci