Introduction
Since the 1700s, atmospheric carbon dioxide (CO2) concentration has increased by 49% because of land-use change, and use of coal and fossil fuels as a form of energy (Friedlingstein et al., Reference Friedlingstein, O’Sullivan, Jones, Andrew, Hauck, Olsen, Peters, Peters, Pongratz, Sitch, Le Quéré, Canadell, Ciais, Jackson, Alin, Aragão, Arneth, Arora, Bates, Becker, Benoit-Cattin, Bittig, Bopp, Bultan, Chandra, Chevallier, Chini, Evans, Florentie, Forster, Gasser, Gehlen, Gilfillan, Gkritzalis, Gregor, Gruber, Harris, Hartung, Haverd, Houghton, Ilyina, Jain, Joetzjer, Kadono, Kato, Kitidis, Korsbakken, Landschützer, Lefèvre, Lenton, Lienert, Liu, Lombardozzi, Marland, Metzl, Munro, Nabel, Nakaoka, Niwa, O’Brien, Ono, Palmer, Pierrot, Poulter, Resplandy, Robertson, Rödenbeck, Schwinger, Séférian, Skjelvan, Smith, Sutton, Tanhua, Tans, Tian, Tilbrook, van der Werf, Vuichard, Walker, Wanninkhof, Watson, Willis, Wiltshire, Yuan, Yue and Zaehle2019). To mitigate the anthropogenic increase of CO2, one proposed strategy is to capture and store CO2 in deep saline aquifers (IPCC, Reference Pachauri and Reisinger2007). Deep saline aquifers are excellent candidates to contain CO2 because of their high storage capacity (to 10,000 Gt of CO2), easy access near CO2 capture sites, and high porosity and permeability (Davidson et al., Reference Davidson, Freund and Smith2001). When supercritical CO2 (scCO2) is injected, the scCO2 migrates upwards, displacing the brine. The saline aquifer requires an impermeable caprock, which may be composed of evaporites (anhydrites or halites), carbonates (limestones or dolostones), or argillaceous rocks (clay-rich shales and mudstones) to effectively store the injected scCO2 (Song and Zhang, Reference Song and Zhang2013).
Most current and future CO2 storage sites are sandstone aquifers sealed by clay-rich shale caprocks (Michael et al., Reference Michael, Golab, Shulakova, Ennis-King, Allinson, Sharma and Aiken2010). Both sandstone and shale may contain smectite, a swelling clay mineral. Smectite often occurs in sandstones as a detrital sand-grain coating (Baker et al., Reference Baker, Uwins and Mackinnon1993), whereas in shales, smectite is a main component. The interlayer distance (d 001) of the smectite changes as the activity of H2O (a(H2O)) varies, which is affected by the brine chemistry, pressure, and temperature. Because smectite is often found in pore spaces of sandstones, changes in d 001 of smectite may affect the porosity and permeability in the sandstone reservoir, impacting its CO2 storage capacity, and causing shrinkage and microfracturing in shales.
Laboratory experiments were performed previously to investigate the effect of changes in brine concentration, pressure, and/or temperature on the d 001 of smectite. Norrish (Reference Norrish1954) first showed that increasing NaCl brine concentration decreases the d 001 of smectite, and this result was further supported by Slade et al. (Reference Slade, Quirk and Norrish1991) using varied brine compositions and smectite samples. The effect of CO2 pressure and temperature on Na-rich smectite (SWy-2) was studied by Giesting et al. (Reference Giesting, Guggenheim, Koster van Groos and Busch2012a), and on K- and Ca-exchanged smectite by Giesting et al. (Reference Giesting, Guggenheim, Koster van Groos and Busch2012b). They found that smectite with 0 to 1 plane of H2O in the interlayer (0W to 1W) allowed CO2 to migrate in the smectite interlayer to cause swelling. By investigating smectite with a different interlayer H2O content, Loring et al. (Reference Loring, Schaef, Thompson, Turcu, Miller, Chen, Hu, Hoyt, Martin, Ilton, Felmy and Rosso2013) found that CO2 adsorption is at a maximum with 1W, and decreases with increasing interlayer H2O content. However, these previous studies were performed under dry to limited H2O conditions, i.e. relative humidity <100%.
In situ X-ray measurements of smectite under water-saturated conditions present several experimental difficulties. These difficulties include smectite flocculation when in contact with brine and a low signal-to-noise ratio from the effect of the liquid dispersing X-rays. In addition, reactions between smectite and its environment are rapid and non-quenchable. Using a high-pressure environmental chamber (HPEC) developed by Guggenheim and Koster van Groos (Reference Guggenheim and Koster van Groos2014), Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) were able to resolve these problems. They investigated the molar volume, i.e., of Na-rich smectite by simulating environmental conditions relevant to CO2 storage: CO2 pressure (P(CO2) = ambient to 500 bars), temperature (T = ~33 to 150°C), and NaCl brine concentration (0.17 M to saturation). They found that the d 001 of Na-rich smectite is significantly affected by brine concentration and T compared with P(CO2). Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) also found that Na-rich smectite has hydration levels of more than two interlayer H2O planes (>2W), and CO2 does not migrate into the interlayer, which supports the study by Loring et al. (Reference Loring, Schaef, Thompson, Turcu, Miller, Chen, Hu, Hoyt, Martin, Ilton, Felmy and Rosso2013).
Studies with smectite–scCO2/CO2–brine systems under water-saturated conditions were carried out using NaCl and KCl brines (e.g. Benavides et al., Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020), but CO2 storage-site brines may also contain other cations and anions. Smectite in such storage sites may undergo cation exchange, thereby changing the interlayer cation composition, e.g. Na in the smectite may be partially replaced by Ca. The interlayer cation composition affects the interlayer distance of smectite because cations have different hydration energies. The hydration energy of a cation is a function of its charge and size, where a smaller cation with a higher charge (e.g. Mg) attracts H2O more strongly compared with a larger cation with a lower charge (e.g. Na). Ferrage et al. (Reference Ferrage, Lanson, Sakharov and Drits2005) examined the hydration of smectites with different interlayer cation compositions. They found that hydration of smectites may be affected by the ionic potential (charge to size ratio) of the cation, which is consistent with similar studies (e.g. Sato et al., Reference Sato, Watanabe and Otsuka1992). Furthermore, Ferrage et al. (Reference Ferrage, Lanson, Sakharov and Drits2005) observed that there is heterogeneity in the hydration levels of smectite. With increasing relative humidity, smectites with greater interlayer cation ionic potential (i.e. more cations with small size and greater charge) have greater proportions of 2W hydration than 1W and 0W compared with smectite with smaller interlayer cation ionic potential. Slade et al. (Reference Slade, Quirk and Norrish1991) investigated smectites with different interlayer cation compositions in brines, e.g. Ca-exchanged smectite with CaCl2 brine. They found that osmotic pressure, directly related to a(H2O), required to decrease the hydration level from 3W to 2W of smectite is related to the size and charge of the cation in both the brine and the interlayer. Hence, including the effect of the interlayer cation composition would add a better understanding of smectite–scCO2/CO2–brine systems.
When CO2 is injected in saline aquifers, environmental conditions change which affects the interlayer (i.e. layer-to-layer) distance of smectite and the stratigraphic- and structural-trapping CO2 storage mechanism. These changes in smectite can affect the porosity and permeability of the aquifer and caprock. Stratigraphic and structural trapping are the main CO2 storage mechanism for the first 10 years after CO2 injection stops. Over time, the trapping mechanism becomes dominated by geochemical trapping by which the injected CO2 dissolves in the brines in the formation (solubility trapping) or reacts with other dissolved ions and mineral phases to form carbonate minerals (mineral trapping) (IPCC, Reference Pachauri and Reisinger2007). Based on a solubility model by Duan and Sun (Reference Duan and Sun2003), the CO2 solubility in brine decreases with temperature and brine concentration, and increases with pressure. This model is supported by Shao et al. (Reference Shao, Thompson, Qafoku and Cantrell2013), whose in situ pH measurements are comparable to the pH values they calculated with modeling programs (e.g. PHREEQC) using the CO2 solubility model by Duan and Sun (Reference Duan and Sun2003).
Dissolution of CO2 in the brines in the formation forms carbonate (CO32–) and hydrogen ions (H+), and results in an increase in the activity of H+ (a(H+)), causing a decrease in pH. The measured and calculated pH values in CO2 storage site conditions range from 2.9 to 3.7 (Shao et al., Reference Shao, Thompson, Qafoku and Cantrell2013). Zysset and Schindler (Reference Zysset and Schindler1996) performed dissolution experiments with K-exchanged smectite in varying KCl solutions within a pH range of 1.0–6.0, and found that the edge sites of the smectite become protonated, i.e. H+ ion adsorption, which promotes the dissolution of K-rich smectite. Stadler and Schindler (Reference Stadler and Schindler1993) showed that the density of protonated edges increases with a(H+), i.e. decrease in pH. In studies with CO2 storage sites, the decrease in pH associated with the CO2 injection promotes the dissolution of K-rich feldspars and plagioclase feldspars, which contributes to the precipitation of carbonates and clay minerals, such as smectite. This was observed using in situ laboratory (e.g. Wigand et al., Reference Wigand, Carey, Schütt, Spangenberg and Erzinger2008) and geochemical modeling data (e.g. Ilgen and Cygan, Reference Ilgen and Cygan2016).
The purpose of the present study expands the work of Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) on the effects of changes in pressure, temperature, and brine concentration on the d 001 of smectite using the HPEC by including the effect of divalent cations, Ca and Mg, in the interlayer. These cations are in saline aquifer brines, e.g. in the Utsira Formation at the Sleipner CO2 storage site in the North Sea (Gregersen et al., Reference Gregersen, Johannessen, Møller, Kristensen, Christensen, Holloway, Chadwick, Kirby, Lindberg and Zweigel1998). Furthermore, the long-term stability of smectite in deep saline aquifers under a wide range of environmental conditions is investigated using geochemical modeling. Although K is also a common cation in saline aquifer brines, K-exchanged smectite in KCl brines is not studied further because K-exchanged smectite lacks periodicity owing to random interstratification of layers with varying hydration levels and no diagnostic d 001 X-ray peak (Benavides et al., Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020).
Materials and methods
Starting material
The starting material was a natural, Na-rich smectite (montmorillonite), SWy-2, obtained from the Source Clays Repository of The Clay Minerals Society. Because SWy-2 contains soluble salts and larger-grain minerals, 50 mg of SWy-2 was purified by sonification in 50 mL of distilled water followed by centrifugation. The process was repeated five times until the supernatant no longer reacted with AgNO3 with visible AgCl precipitates. The <2 μm size fraction of the purified SWy-2 was obtained by following the procedure of Moore and Reynolds (Reference Moore and Reynolds1997; pp. 211–213). The Ca- and Mg-exchanged SWy-2 (hereafter CaSWy-2 and MgSWy-2, respectively) were obtained by saturating and agitating via sonification of the <2 μm SWy-2 with 0.10 M CaCl2 or 0.10 M MgCl2 made using high-purity (99.99%) salts, respectively, and letting the mixture sit at room temperature for 24 h. This process was repeated two more times to ensure complete cation exchange. To remove excess salt, the cation-exchanged smectite was purified using the same procedure as with the starting material, and then dried under ambient conditions and stored in a glass vial.
Equipment
In situ X-ray diffraction (XRD) experiments of a smectite–brine–CO2 system were performed using a high pressure experimental chamber (HPEC) made with Ti-V-Al (6AL 4V ELI, ASTM Grade 5) alloy (Guggenheim and Koster van Groos, Reference Guggenheim and Koster van Groos2014). The HPEC interior is interconnected by two horizontal and two vertical channels. One of the vertical channels allows X-rays to pass through via a pair of opposing diamond windows, 1 mm apart. The other vertical channel contains an internal pump that maintains a suspension of the smectite–brine–CO2 mixture while X-ray data were collected. The suspension is produced by the rapid flow rate caused by the narrow space between the diamond windows. The rapid flow rate also allowed rapid mixing and equilibration. The HPEC allows study of smectite in suspension while XRD data are being collected. The HPEC resolves problems on studying smectites that flocculate under brine solutions, as well as problems on low signal-to-noise ratio caused by X-ray dispersion from the liquid.
Experimental procedure
CaSWy-2 or MgSWy-2 was loaded in the HPEC with brine solutions corresponding to the interlayer cation (e.g. CaCl2 brine with CaSWy-2). This procedure avoids cation exchange if brines of a different or mixed composition were used. CaCl2 or MgCl2 brine solutions at varying concentrations (0.17 M, 0.34 M, 0.68 M, 1.34 M, 1.71 M, 2.05 M, 3.42 M, and saturated) were made using high-purity (99.99%) salts. The brine solutions were prepared on the day of the experiment. In variable-pressure experiments, X-ray data were initially collected under ambient conditions before increasing the CO2 partial pressures (P(CO2)) to 30, 70, 100, 200, 300, 400, and 500 bars at a constant temperature (T) of approximately 33°C. In temperature experiments, X-ray data were also collected under ambient conditions before increasing the P(CO2) to 250 bars at T ~33°C. The temperature was then increased to 50, 75, 100, 125, and 150°C, and P(CO2) was allowed to vary. The experimental P(CO2) and T range were chosen to include in situ formation P(CO2) and T of naturally occurring CO2 reservoirs (Miocic et al., Reference Miocic, Gilfillan, Roberts, Edlmann, McDermott and Haszeldine2016). Prior to collecting the X-ray data for each experiment, the clay suspension was circulated for 30 min so that the system could reach equilibrium.
Data collection
The HPEC loaded with 200 mg of sample and 2–3 mL of brine was mounted on a Bruker D8 3-circle transmission mode X-ray diffractometer; the diffractometer is equipped with a PHOTON 100 CMOS area detector (1024×1024 pixels) at a distance of 120 mm from the sample center, a Monocap collimator with a 0.3° divergence, and a graphite monochromator. The diffractometer was operated at 45 kV and 25 mA with a Mo X-ray tube.
The pressure in the HPEC was monitored with a transducer, and the temperature with a thermocouple, external from the brine but within the HPEC body, located 1 mm from the diamond windows (see Guggenheim and Koster van Groos, Reference Guggenheim and Koster van Groos2014). Data frames were collected for 1200 s (20 min) using Bruker APEX3 (version 2019.11-0) software. To generate the intensity vs 2θ diffraction plot, the frames were processed using the Bruker application GADDS (version 4.1.60, 2017). The diffraction plot was calibrated using the (003), (004), and (005) peaks of Ag(I)-behenate. The Ag(I)-behenate was loaded in the HPEC with distilled H2O and was analyzed similar to the samples. The d 001 value at full-width half-maximum was obtained from the diffraction plot imported to Materials Data Inc. JADE+ (version 9.6.0, 2015).
Geochemical modeling
Extent of dissolution
The reaction of SWy-2, brine, and CO2 was modeled using the numerical model code PHREEQC v.3.7.3 (Parkhurst and Appelo, Reference Parkhurst and Appelo2013) with the database from the Lawrence Livermore National Laboratory dataset (llnl.dat). The phase Montmor-X [Xy0.33/yMg0.33Al1.67Si4O10(OH)2], where X is the interlayer cation Na+, Ca2+ or Mg2+ and y is the charge of the ion in the database, was used to model SWy-2. The rate equation used to model the dissolution of SWy-2 was derived from the transition state theory (TST) rate law (Aagaard and Helgeson, Reference Aagaard and Helgeson1982; Hellevang et al., Reference Hellevang, Pham and Aagaard2013):
$$ r= Sk\prod_i{a}_i^{v_{+}}\left(1-\Omega \right), $$
where r is the reaction rate (mol s–1), S is the reactive surface area (m2), k is the far-from-equilibrium dissolution rate coefficient (mol m–2 s–1), a is the activity of species i, v+ is the reaction order, and Ω is the saturation state expressed as the exponential of the Gibbs free energy of the reaction over the gas constant and absolute temperature [exp(ΔG/RT)]. The dissolution rate coefficient parameters for SWy-2 were obtained from Palandri and Kharaka (Reference Palandri and Kharaka2004) using the acid mechanism parameters for montmorillonite. Because the parameters from Palandri and Kharaka (Reference Palandri and Kharaka2004) are pH dependent, the activity product term in Eqn 1 is defined by the pH of the solution. The reactive surface area was approximated by:
$$ S= nM\unicode{x03B2}, $$
where n is the number of moles, M is the molecular weight (g mol–1) and β is the specific surface area (m2 g–1). The value of β was obtained from Golubev et al. (Reference Golubev, Bauer and Pokrovsky2006).
PHREEQC uses the ideal gas law to compute the solubility of gases in solution. To model the solubility of CO2, the input P(CO2) values were corrected using a fugacity coefficient computed using the Soave–Redlich–Kwong (SRK) equation of state (Soave, Reference Soave1972) and a Poynting correction term. The resulting solubility of CO2 with these corrections is in good agreement with the solubility computed using the model of Duan and Sun (Reference Duan and Sun2003) (Hellevang and Kvamme, Reference Hellevang and Kvamme2007). The equilibrium constant (K) of dissociation for all reactions performed in PHREEQC was based on the following equation:
$$ \log \mathrm{K}=\mathrm{a}+\mathrm{b}T+\frac{\mathrm{c}}{T}+\mathrm{dlog}T+\frac{\mathrm{e}}{T^2} $$
where K is the equilibrium constant, and a, b, c, d, and e are constants. The constants for each phase are included in the llnl.dat database.
Prior to computing the extent of dissolution of SWy-2, pH values were calculated using the corrected CO2 solubility at 3.00 M NaCl brine under different CO2 pressures [P(CO2)] at 40 and 75°C. The calculated pH values are in good agreement with experimental values from Shao et al. (Reference Shao, Thompson, Qafoku and Cantrell2013). The extent of dissolution of SWy-2 was modelled using the XRD experimental P(CO2), T and brine composition and concentration as the initial conditions. The model simulated the length of the one experiment, 1 h, after which the SWy-2 was no longer reacting.
Swelling pressure
The swelling pressures were calculated based on the extended Derjaguin–Landau–Verwey–Overbeek (DLVO) model by Liu (Reference Liu2013). The DLVO model approximates σ as the sum of the osmotic repulsive pressure (P osm), the van der Waals attractive pressure (P vdW), and the hydration pressure (P hyd):
where P osm, P vdW, and P hyd are calculated using Eqn (5), (6), and (7), respectively:
and
$$ {P}_{\mathrm{hyd}}= Kexp\left(-\frac{h}{\unicode{x03BB}}\right) $$
where R is the gas constant, T is the temperature in Kelvin,
$ {V}_{{\mathrm{H}}_2\mathrm{O}} $
is the molar volume of water, A is the Hamaker constant, h is the thickness of the interlayer, and K and λ are constants. The values of A, K, and λ were obtained from Liu (Reference Liu2013), and h is calculated by the difference of the d
001 values of SWy-2 and its dehydrated thickness, 9.6 Å (Ferrage et al., Reference Ferrage, Lanson, Sakharov, Geoffroy, Jacquot and Drits2007).
Results
Effect of brine composition and concentration
The effect of varying CaCl2 and MgCl2 brine concentrations from 0.17 M to saturation on the interlayer distance of CaSWy-2 and MgSWy-2, respectively, were examined and compared with the effect of NaCl brines on Na-rich SWy-2 (hereafter Na-SWy-2) (Benavides et al., Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) (Fig. 1). Both CaSWy-2 and MgSWy-2 results showed trends similar to the Na-SWy-2 results. In general, as brine concentration increases, the d 001 value decreases along with noticeable sharp decreases. In the CaSWy-2 experiments, a sharp d 001 decrease from 18.8 to 15.6 Å (18%) is observed when the brine concentration was increased from 0.68 to 1.71 M. In contrast, for MgSWy-2 the sharp d 001 decrease occurred at a higher concentration. The largest d 001 value decrease from 19.0 to 15.6 Å (18%) was observed when the brine concentration was increased from 1.37 to 3.42 M for MgSWy-2, which is similar to the decrease for Na-SWy-2. The corresponding d 001 values of CaSWy-2 are approximately 0.4 Å lower compared with MgSWy-2 and Na-SWy-2. At lower brine concentrations (<0.34 M), d 001 values for both CaSWy-2 and MgSWy-2 do not exceed 20.1 Å, which is the greatest d 001 value observed for Na-SWy-2. The d 001 values in CaSWy-2 and MgSWy-2 reach a value of 15.1 Å at saturation.
Figure 1. Effect of brine composition and concentration on the d 001 value of SWy-2 at P(CO2) of 30 bars and T of approximately 33°C. The brine composition corresponds to the interlayer cation (e.g. CaCl2 brine with CaSWy-2). The d 001 value error bars are +0.2 Å.
Effect of CO2 pressure
Figure 2 shows that the effect of increasing CaCl2 brine concentration with increasing P(CO2) to 500 bars at T of approximately 33°C on the d 001 value of CaSWy-2. The gap between 0.68 and 1.71 M corresponds to the steep slope in Fig. 1. At lower brine concentrations (<0.38 M), the d 001 values remain the same within the standard error when P(CO2) is increased from ambient pressure to 500 bars. In general, when the P(CO2) is increased from ambient pressures to 500 bars, there is no change in the d 001 values observed for CaSWy-2 in CaCl2 brines.
Figure 2. Effect of CaCl2 brine concentration and CO2 pressure (P(CO2)) on the d 001 of CaSWy-2 at T of approximately 33°C. The d 001 error bars are +0.2 Å.
The d 001 values of MgSWy-2 change with increasing P(CO2), as shown in Fig. 3. The gap between 1.37 and 3.42 M corresponds to the steep slope in Fig. 1, similar to the CaSWy-2 experiments.
Figure 3. Effect of MgCl2 brine concentration and CO2 pressure (P(CO2)) on the d 001 of MgSWy-2 at T of approximately 33°C. The d 001 error bars are +0.2 Å.
For brine concentrations <1.37 M, the d 001 values of MgSWy-2 decrease from 19.2 to 18.9 Å (2%) when P(CO2) is increased to 500 bars. Between 1.37 and 3.42 M, an increase in the d 001 value is observed when P(CO2) is increased to approximately 30 bars, then the d 001 values remain the same when P(CO2) is increased to 500 bars. For example, at a brine concentration of 2.05 M at ambient conditions, the d 001 value of MgSWy-2 is 16.4 Å. When P(CO2) is increased to 30 bars, the d 001 values increased to 17.9 Å and remained the same within the standard error when P(CO2) is increased to 500 bars. At brine concentration >3.42 M, the d 001 values remain within the standard error when the P(CO2) is increased from ambient conditions to 500 bars.
Effect of temperature
The effect of varying CaCl2 brine concentrations and increasing T to 150°C on the d 001 value of CaSWy-2 is shown in Fig. 4. The starting P(CO2) prior to increasing T is 250 bars. At brine concentrations of <2.05 M, a decreasing trend in the d 001 values is observed as T increases to 150°C. The greatest decrease is observed at a brine concentration of 1.37 M. As T increases from 50 to 105°C, the d 001 value decreases from 17.1 to 15.4 Å (11%). As the brine concentration increases, the sharp decrease in d 001 values occurs at a lower T. For example, the d 001 value starts to decrease at 125°C at 0.68 M, in contrast to the 1.37 M concentration where the d 001 starts to decrease at 50°C. At brine concentrations of <0.68 M, the T where the d 001 value sharply decreases probably occurs beyond the experimental T, 150°C. At brine concentrations of >1.71 M, d 001 values remain within the standard error when T is increased to 150°C.
Figure 4. Effect of CaCl2 brine concentration and temperature (T) on the d 001 of CaSWy-2. The observed P(CO2) recorded for each T is shown below the T axis. The P(CO2) is a dependent variable. The d 001 error bars are +0.2 Å.
The d 001 values of MgSWy-2 at varying MgCl2 brine concentrations show similar trends to those of CaSWy-2 when T is increased to 150°C (Fig. 5). However, at brine concentrations of <3.42 M, the sharp decrease in d 001 values for MgSWy-2 as T is increased to 150°C is not as apparent as those of CaSWy-2. At brine concentrations of <2.05 M, the sharp decrease in the d 001 values probably occurs at T beyond the experimental temperature of 150°C. For example, at brine concentration of 2.05 M, the d 001 is beginning to decrease at T of 100°C, and might decrease sharply at a T greater than 150°C. At saturation, the d 001 values are within the standard error when T is increased to 150°C.
Figure 5. Effect of MgCl2 brine concentration and temperature (T) on the d 001 of MgSWy-2. The observed P(CO2) recorded for each T is shown below the T axis. The P(CO2) is a dependent variable. The d 001 error bars are +0.2 Å.
Swelling pressures
Figure 6 shows the effect of the swelling pressures (σ) on the d 001 of Na-SWy-2 in NaCl brines with data from Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020), and CaSWy-2 and MgSWy-2 in CaCl2 and MgCl2 brines, respectively, from this study. Regression coefficients for the linear relationships are summarized in Table 1. The hydration states with three planes of H2O [3W, d 001=18.5–19.5 Å] and two planes of H2O [2W, d 001=13.9–15.8 Å] and their corresponding d 001 value ranges are highlighted in gray. The d 001 value range for the 2W hydration state in Fig. 6 is only from 14.5 to 15.8 Å. The d 001 value range for the 2W hydration state is based on best-fit modeling on SWy-2 obtained by Ferrage et al. (Reference Ferrage, Lanson, Sakharov and Drits2005), whereas the range for the 3W is based on this study.
Figure 6. Effect of the calculated log swelling pressure on the d 001 values of SWy-2. The circles represent Na-SWy-2 in NaCl brine experiments with data from Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020). The squares and triangles represent CaSWy-2 and MgSWy-2 in CaCl2 and MgCl2 brine experiments in this study, respectively. The continuous, dotted, and dashed lines represent the best linear fit for Na-SWy-2, CaSWy-2, and MgSWy-2 experiments, respectively. The d 001 value ranges for the 3W and 2W hydration states are highlighted in gray, and the d 001 value range for 2W does not show values from 13.9 to 14.5 Å. Errors are not given because of the data point density.
Table 1. Linear regression coefficients for the Na-SWy-2, CaSWy-2, and MgSWy-2 plots in Fig. 6
In all experiments, the d 001 values decrease as σ increases. The transition from a 3W to a 2W hydration state occurs at different swelling pressures for the three sets of experiments. The σ where the 3W–2W transition begins was estimated from the d 001 value range midpoint for the hydration state, e.g. d 001=19.0 Å for 3W. The transition for Na-SWy-2 experiments occurs at approximately σ=7.04 MPa, whereas the transition for CaSWy-2 and MgSWy-2 experiments occurs at approximately σ=7.68 and σ=7.74 MPa, respectively. Only Na-SWy-2 has d 001 values greater than that for a 3W hydration state, and the 3W to a higher hydration state begins at approximately σ=7.04 MPa.
Extent of dissolution of SWy-2 in experiments
The calculated pH and percentage of dissolved SWy-2 for each P(CO2) (Fig. 7) and temperature (Fig. 8) are presented. The data include different brine compositions and concentration experiments from Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) for Na-SWy-2 in NaCl brines, and from this study for CaSWy-2 in CaCl2 brines and MgSWy-2 in MgCl2 brines. The percentage of dissolved SWy-2 is the percentage difference of the initial mass of Na-SWy-2, CaSWy-2, or MgSWy-2, and their mass at steady state.
Figure 7. Calculated pH and percentage of dissolved SWy-2 under different CO2 pressure (P(CO2)) and brine composition and concentration experimental conditions for Na-SWy-2 in NaCl brines (A,B) with data from Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020), and CaSWy-2 in CaCl2 brines (C,D) and MgSWy-2 in MgCl2 brines (E,F) in this study. The percentage of dissolved SWy-2 values are the percentage difference of the initial and final amount of SWy-2 at steady state.
Figure 8. Calculated pH and percentage of dissolved SWy-2 under different temperature (T) and brine composition and concentration experimental conditions for Na-SWy-2 in NaCl brines (A,B) with data from Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020), and CaSWy-2 in CaCl2 brines (C,D) and MgSWy-2 in MgCl2 brines (E,F) in this study. The percentage of dissolved SWy-2 values are the percentage difference of the initial and final amount of SWy-2 at steady state.
Pressure
In all brine compositions and at all concentrations, as P(CO2) increases, pH rapidly decreases from 30 to 70 bars and slowly decreases from 70 to 500 bars. Also, pH decreases as the brine concentration increases from 0.17 to 3.42 M (Fig. 7A,C,E). An opposite trend is observed for the percentage of dissolved SWy-2. The percentage of dissolved SWy-2 rapidly increases as P(CO2) increases from 30 to 70 bars, and the percentage of dissolved SWy-2 slowly increases from 70 to 500 bars. As brine concentration increases from 0.17 to 3.42 M, the percentage of dissolved SWy-2 also increases (Fig. 7B,D,F).
The MgSWy-2 in MgCl2 brine experiments exhibit the lowest pH range values, of 3.07–3.51 (Fig. 7E). However, the pH ranges for CaSWy-2 in CaCl2 brines (Figure 7C) and Na-SWy-2 in NaCl brines (Fig. 7A) are similar with range values of 3.20–3.60 and 3.25–3.67, respectively. A different trend is observed for the percentage of dissolved SWy-2. The highest percentage of dissolved SWy-2 range is observed with CaSWy-2 in CaCl2 brine experiments from 0.38 to 1.04% (Fig. 7D), followed by MgSWy-2 in MgCl2 brine experiments from 0.30 to 0.82% (Fig. 7F). The lowest percentage of dissolved SWy-2 range is observed with Na-SWy-2 in NaCl brine experiments with values ranging from 0.30 to 0.61% (Fig. 7B).
Temperature
As the temperature is increased from 50 to 150°C, pH decreases for all sets of experiments (Fig. 8A,C,E). However, for Na-SWy-2 in NaCl brine experiments, the pH decreases when the temperature is increased from 100 to 150°C, with the exception of 3.42 M. At these concentrations, the pH values start increasing again when temperature is increased from 100 to 150°C (Fig. 8A). In contrast to P(CO2), the percentage of dissolved SWy-2 shows a similar trend with pH. As temperature is increased from 50 to 150°C, the percentage of dissolved SWy-2 decreases for all sets of experiments (Fig. 8B,D,F). In all experiments, pH decreases when the brine concentration increases, whereas the percentage of dissolved SWy-2 increases.
Discussion
Effects of P(CO2), temperature, and brine concentration
Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020) examined the effect of NaCl brines on the d 001 of Na-SWy-2 and found that as brine concentration increases, the d 001 value decreases, with the greatest decrease observed at concentrations between 1.37 and 2.05 M. Furthermore, those authors observed that the transition from osmotic swelling to intracrystalline swelling of Na-SWy-2 occurred when the brine concentration was increased from 0.17 to 0.34 M. Osmotic swelling results from the chemical potential difference between the brine and the interlayer, whereas intracrystalline swelling is a stepwise change in hydration state in the interlayer. The present study also shows that for CaSWy-2 and MgSWy-2, the d 001value decreases as the brine concentration increases. However, in CaSWy-2 and MgSWy-2 experiments, only intracrystalline swelling is observed, where the hydration state in the interlayer changed from 3W to 2W with increasing brine concentration. The concentration of cations in the clay silicate layer is fixed to satisfy the layer charge, whereas the concentration in solution depends on the brine concentration. When the brine concentration is increased, the ions in solution compete for the H2O molecules around the interlayer cation. Thus, the d 001 value is lowered from that corresponding to a 3W hydration state as the H2O molecules leave the interlayer. Furthermore, the d 001 values at a 3W hydration state are greater in comparison with the d 001 values found in Ferrage et al. (Reference Ferrage, Lanson, Sakharov and Drits2005) (e.g. 18.0–18.5 Å). These higher d 001 values in the present study are attributed to the experimental set-up, where the SWy-2 particles were allowed to flow freely in the suspension, thereby allowing the SWy-2, brine, and CO2 to achieved equilibrium rapidly. Intermediate d 001 values between 2W and 3W hydration states are also observed for both CaSWy-2 and MgSWy-2 experiments. Intermediate d 001 values result from increased ion–ion interactions in the interlayer as brine concentration increases. This would be expected to affect the position of the H2O molecules around the cation in the interlayer.
The d 001 values sharply decreased from 18.8 to 15.6 Å at 1.25 M for CaSWy-2 in CaCl2 brines, and from 19.0 to 15.6 Å at 1.50 M for MgSWy-2 in MgCl2 brines (Fig. 1). Slade and Quirk (Reference Slade and Quirk1991) investigated the effect of varying CaCl2 and MgCl2 solutions on the intracrystalline swelling of different smectites, including SWy-2. The results in the present study are consistent with the results obtained by Slade and Quirk (Reference Slade and Quirk1991). The sharp decrease in the d 001 value occurs at 1.37 M for MgSWy-2 in MgCl2 brine experiments, which is in good agreement with Slade and Quirk (Reference Slade and Quirk1991) at 1.50 M. However, in CaSWy-2 in CaCl2 brine experiments, the sharp decrease in the d 001 value occurs at 0.68 M, which is lower than Slade and Quirk (Reference Slade and Quirk1991) at 1.25 M. This difference is attributed to the gentler slope in the initial d 001 decrease between 0.68 and 1.37 M. The sharp d 001 decrease may occur closer to 1.37 M if the CaSWy-2 experiments follow the sharp decrease trend observed in the Na-SWy-2 and MgSWy-2 experiments.
Effect of interlayer cation
The interlayer cation affects the d 001 value (Fig. 1) and swelling pressure (Fig. 6) of SWy-2. In experiments with divalent cations (Ca2+ and Mg2+), only intracrystalline swelling is observed with the stepwise change in hydration level at the transition brine concentration, whereas with Na, osmotic swelling is observed at concentrations less than 0.34 M (Benavides et al., Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020). This swelling is consistent with literature data for experiments in which clays were oriented on slides (e.g. Ferrage et al., Reference Ferrage, Lanson, Sakharov and Drits2005), or prepared as thin films (e.g. Slade et al., Reference Slade, Quirk and Norrish1991). Furthermore, in MgSWy-2 experiments done at brine concentrations of 1.71 and 2.05 M, the d 001 value increases when P(CO2) is increased from ambient conditions to 30 bars at T approximately 33°C (Fig. 3). The d 001 value at ambient conditions is probably metastable. The initial low d 001 value is probably a result of the temperature increase caused by the exothermic reaction when the MgCl2 brine was prepared, which effectively shift this experimental point to a higher temperature (Fig. 4).
The difference in the d 001 values for the different cations under similar P(CO2), T, and brine concentration (e.g. 19.3, 19.2, and 18.8 Å for Na-SWy-2, CaSWy-2, and MgSWy-2, respectively, at P(CO2) = 30 bars, T = approximately 33°C, and brine concentration = 0.68 M) at a 3W hydration level is a result of the hydration energy of the cations, and how the cations are coordinated with the silicate layer (outer-sphere or inner-sphere complex). The higher d 001 values for Na-SWy-2 probably occur because of the tendency of Na cations near the silicate layer to form outer-sphere complexes, where the coordinated H2O molecules around the cation are retained, whereas the cations for CaSWy-2 and MgSWy-2 form inner-sphere complexes, where the cations are directly coordinated with the silicate layer (Planková and Lísal, Reference Planková and Lísal2020). Furthermore, the charge density of Ca and Mg is greater than that of Na, and so the show a greater attraction for the negatively charged silicate layers. Hence, CaSWy-2 and MgSWy-2, in general, have smaller d 001 values compared with Na-SWy-2 for the same brine concentration.
At 3W and 2W hydration levels, the d 001 values are approximately 0.4 Å higher for MgSWy-2 than for CaSWy-2 (Figs 1, 2, and 3) because of the difference in the hydration energy of the interlayer cations. The hydration energy, which depends on the charge and size of the cation, influences the ability of a cation to attract H2O molecules. The more negative hydration energy of Mg (–1920 kJ mol–1) is smaller compared with Ca (–1560 kJ mol–1) and Na (–410 kJ mol–1) (Smith, Reference Smith1977). Hence, MgSWy-2 is likely to attract more H2O molecules into the interlayer, causing greater expansion, i.e. higher d 001 values, compared with CaSWy-2. The hydration energy is also consistent with the range of swelling pressures over which the 3W to 2W hydration level transition occurs (Fig. 6). Among the three experiments, MgSWy-2 transitions with the highest swelling pressure at 7.74 MPa followed by CaSWy-2 at 7.68 MPa, then Na-SWy-2 at 7.04 MPa.
Calculated pH
The calculated pH for all experiments decreases when P(CO2) is increased to 500 bars (Fig. 7A,C,E) or T is increased to 150°C (Fig. 8A,C,E). The decrease in pH results from dissolved CO2 reacting with H2O to form H2CO3 that can dissociate to produce H+ ions. The concentration of dissolved CO2 increases as P(CO2) increases, and a decrease is observed when temperature and brine concentration are increased (Duan and Sun, Reference Duan and Sun2003). The decrease in the calculated pH when P(CO2) is increased is a result of an increase in dissolved CO2 in the solution. The rapid decrease in pH occurs when P(CO2) is increased to 70 bars because CO2 is in the gas phase. However, the slow decrease in pH (up to 0.10 pH units) when P(CO2) is increased from 70 to 500 bars occurs because the CO2 is the supercritical phase. This trend is observed at all brine compositions and concentrations and is consistent with the trend observed by Shao et al. (Reference Shao, Thompson, Qafoku and Cantrell2013) with their pH values obtained experimentally.
The increase in temperature in closed-system experiments is also accompanied by an increase in P(CO2), which allows more CO2 to dissolve in the brine to result in a decrease in pH. However, the trend of the pH decrease changes when temperature is increased from 100 to 150°C in NaCl brine experiments (Fig. 8A). This occurs because less H+ is produced, i.e. the pH is higher, although the brine composition and concentration also affects the pH values. Hence, changes in P(CO2), temperature, and brine composition and concentration affect the calculated pH of the solution.
Dissolution of SWy-2
When P(CO2) is increased to 500 bars (Fig. 7B,D,F), the observed increase in percentage of dissolved SWy-2 is a result of the decreased pH of the brine. A decrease in pH corresponds to an increase in H+ activity a(H+). The complete dissolution reaction of SWy-2 can be expressed as:
$$ {\displaystyle \begin{array}{l}{X^y}_{0.33/ y}{\mathrm{Mg}}_{0.33}{\mathrm{Al}}_{1.67}{\mathrm{Si}}_4{\mathrm{O}}_{10}{\left(\mathrm{OH}\right)}_2+6.00\;{\mathrm{H}}^{+}\to \\ {}\hskip1em 0.33/ y\;{X}^y+0.33\;{\mathrm{Mg}}^{+2}+1.67\;{\mathrm{Al}}^{+3}+4.00\;{\mathrm{H}}_2\mathrm{O}+4.00\;{\mathrm{Si}\mathrm{O}}_2,\end{array}} $$
where X is the interlayer cation (Na+, Ca+2, Mg+2) and y is the charge of the ion. The increase in the a(H+) will cause the reaction to proceed to the right. However, the percentage of dissolved SWy-2 decreases for all experiments despite the decrease in pH when T is increased to 150°C (Fig. 8). This result is related to the decrease in solubility of SWy-2 when temperature is increased, i.e. retrograde solubility. The equilibrium constant for the complete dissolution of SWy-2 (Eqn 8) decreases when temperature is increased and results in the reaction favoring the reactants, i.e. decrease in solubility of SWy-2.
The percentage of dissolved SWy-2 from the models for all experiments ranges from 0.2 to 1.1% for a pH range of 2.74–3.67 (Figs 7 and 8). The low percentage of dissolved SWy-2 in acidic conditions is consistent with kinetic experiments of K-exchanged montmorillonite dissolution in KCl solutions at T=23+1°C obtained by Zysset and Schindler (Reference Zysset and Schindler1996). The total amount of K-exchanged montmorillonite dissolved in their experiments ranged from 0.46% (pH 4.0) to 6.09% (pH 1.0). Zysset and Schindler (Reference Zysset and Schindler1996) attributed this low value to dissolution that occurred mostly at edge sites (Si-O-Al or Al-OH-Al) and not at basal sites. The mechanism of dissolution occurs by protonation of the edge sites. Under acidic conditions, the density of protonated edge sites (>AlOH2+) increases (Stadler and Schindler, Reference Stadler and Schindler1993). The increase in protonated edge sites leads to the detachment of an Al3+‑proton complex followed by hydrolysis of the tetrahedral sheet. Furthermore, the detached transition-state complex acts as a limiting step for the dissolution reaction (Huertas et al., Reference Huertas, Chou and Wollast1999).
Conclusions and implications
The present study shows that changes in environmental conditions affect the interlayer distance, i.e. the d 001 value, of CaSWy-2-CaCl2 and MgSWy-2-MgCl2 brine systems, which is consistent with results from the study of Na-SWy-2 in NaCl brines by Benavides et al. (Reference Benavides, Kowalik, Guggenheim and Koster van Groos2020). Overall, the d 001 value of SWy-2 changes significantly when brine concentration and T are varied, whereas little to no change was observed when P(CO2) is varied. When brine concentration is increased from 0.17 M to saturation, the d 001 value of CaSWy-2-CaCl2 and MgSWy-2-MgCl2 brines systems can decrease by 18% at 0.68 and 1.37 M, respectively. Furthermore, both brine systems show that the d 001 value can decrease by up to 11% when T is increased from approximately 33 to 150°C; however, the d 001 value decrease in MgSWy-2-MgCl2 brine systems is less apparent. The difference in results in both brine systems occurs because of the difference in hydration energies of the interlayer cation, where Mg has a greater hydration energy than Ca. Thus, a greater brine concentration and T are needed to decrease the d 001 values of MgSWy-2. This observation is consistent with the swelling pressures (σ) needed to change the hydration state of SWy-2 from 3W to 2W.
Geochemical modeling was used to simulate how SWy-2 will react under XRD experimental conditions until SWy-2 no longer reacts. When scCO2 is injected into the HPEC, the pH of the brine decreases as a result of the dissolution of CO2 in the solution. The decrease in pH produces more H+, which causes SWy-2 to dissolve. The decrease in pH is associated with an increase in percentage of dissolved SWy-2 because there are more H+ ions reacting with SWy-2. However, the inverse is observed with a pH decrease when T increases to 150°C. The inverse is attributed to the retrograde solubility of SWy-2, i.e. SWy-2 is less soluble at higher T. In general, the modeling shows that the pH decrease only results in a low dissolution of SWy-2 of up to 1.1% dissolved SWy-2.
The range of P-T conditions in the present study included in situ P-T conditions of naturally occurring saline aquifers (Miocic et al., Reference Miocic, Gilfillan, Roberts, Edlmann, McDermott and Haszeldine2016). Overall, under the short-term stratigraphic and structural trapping mechanism, the interlayer distance, d 001 value, changes in smectite, such as SWy-2, can result in changes in the porosity and/or permeability of a saline aquifer and a caprock. When the d 001 value of smectite increases, the saline aquifer becomes less porous and less permeable and results in a decrease in its capacity to store CO2. In contrast, this will enhance the sealing capacity of a caprock. However, a decrease in the d 001 value of smectite might result in an increase in porosity and permeability. Thus, the saline aquifer benefits because an increase in the CO2 storage capacity of the aquifer occurs, whereas for a caprock, the same result may create conduits for CO2 to escape to the surface. However, when CO2 injection stops, geochemical trapping becomes the dominant mechanism of sequestering CO2. The stored CO2 in the aquifers will decrease the brine pH, which results in the dissolution of smectite until it reaches a steady state. Smectite will probably continue to be present in the reservoir rock because of its low solubility.
Smectite has a wide range of silicate-layer compositions affecting its layer charge location and density, which also affect its d 001 value with changes in P(CO2), T, and brine concentration, and its extent of dissolution. Experimental work coupled with modeling of smectite with different silicate-layer compositions and appropriate brines and other minerals present are essential for full assessment of the stability of smectites in saline aquifer-caprock systems. In addition, the parameter of ‘time’ (i.e. reaction kinetics) needs to be considered in the modeling because CO2 containment (e.g. geochemical trapping) may require long-term storage, perhaps hundreds or thousands of years.
Author contributions
Paolo Andre Benavides: Conceptualization, Formal Analysis, Funding Acquisition, Investigation, Methodology, Visualization, and Writing; Stephen Guggenheim: Conceptualization, Methodology, Project Administration, Supervision, Resources, and Writing.
Acknowledgements
The authors thank August Koster Van Groos and Kathryn Nagy for reviewing the manuscript, and Richard Dojutrek for laboratory assistance.
Financial support
The Authors acknowledge the funding provided by the University of Illinois at Chicago Chancellor’s Graduate Research Award, The Clay Minerals Society Student Research Grant, and the Department of Earth and Environmental Sciences, University of Illinois at Chicago.
Competing interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data availability statement
The data used to support the findings of this study are available from the corresponding author, PAB, upon request.
