Quantification of the Dissolved Inorganic Carbon Species and of the

Article pubs.acs.org/ac

Quantification of the Dissolved Inorganic Carbon Species and of the pH of Alkaline Solutions Exposed to CO2 under Pressure: A Novel Approach by Raman Scattering Thomas Beuvier,*,† Brice Calvignac,‡ Jean-François Bardeau,† Alain Bulou,† Frank Boury,‡ and Alain Gibaud*,† †

LUNAM, Université du Maine, Institut des Molécules et Matériaux du Mans, UMR CNRS 6283, Avenue Olivier Messiaen, 72085 Le Mans Cedex 9, France ‡ LUNAM, Université d’Angers, INSERM U1066-Micro-Nanomedecines Biomimétiques, 49933 Angers Cedex 9, France S Supporting Information *

ABSTRACT: Dissolved inorganic carbon (DIC) content of aqueous systems is a key function of the pH, of the total alkanility (TA), and of the partial pressure of CO2. However, common analytical techniques used to determine the DIC content in water are unable to operate under high CO2 pressure. Here, we propose to use Raman spectroscopy as a novel alternative to discriminate and quantitatively monitor the three dissolved inorganic carbon species CO2(aq), HCO3−, and CO32− of alkaline solutions under high CO2 pressure (from P = 0 to 250 bar at T = 40 °C). In addition, we demonstrate that the pH values can be extracted from the molalities of CO2(aq) and HCO3−. The results are in very good agreement with those obtained from direct spectrophotometric measurements using colored indicators. This novel method presents the great advantage over high pressure conventional techniques of not using breakable electrodes or reference additives and appears of great interest especially in marine biogeochemistry, in carbon capture and storage and in material engineering under high CO2 pressure.

I

dissolution of marine carbonates or the change in blood pH during hyperventilation caused by the decrease of CO2(aq) concentration.6 However, the changes in pH in these two examples are rather weak in comparison with those obtained at high CO2 pressure. For instance for water exposed at 200 bar of CO2 at T = 40 °C, the molality of CO2 in water reaches 1.37− 1.40 mol·kg−1 water7,8 which yields a pH close to 3.9,10 At such a low pH, the major species in the solution is CO2(aq). The concentrations of HCO3− and CO32− are therefore quite small ([HCO3−] ≈ 1 mmol/kg and [CO32−] ≈ 5.10−11 mol/kg). In alkaline solutions, they can be drastically enhanced either at low11−13 and high14,15 pressure. If the molalities of these species can be determined by titration at ambient pressure, it is no more the case at high pressure where the only key variable

n situ chemical speciation together with pH determination of solutions is a very important issue in environmental science. Of particular interest is the specific case of aqueous solutions in contact with carbon dioxide at ambient or elevated pressure for which solving this issue is a major challenge. Many scientists are directly concerned as geophysicists, who work in carbon capture and sequestration (CCS),1,2 material engineers,3 who prepare materials using supercritical CO2 (sc-CO2), or marine biologists, who are worried about the long-term fate of anthropogenic CO2 in the atmosphere and oceans.4,5 One of the key parameters to monitor is the dissolved inorganic carbon (DIC = [CO2(aq)] + [HCO3−] + [CO32−]) of water in contact with CO2. The DIC is intimately correlated to the pH, to the partial pressure of CO2 and to the total alkalinity (TA) of the solution. For a given TA, the higher the CO2 solubility, the lower the pH. Remarkable examples of this correlation are the acidification of oceans induced by the uptake of anthropogenic carbon dioxide from atmosphere4,5 which in turn provokes the © 2014 American Chemical Society

Received: July 10, 2014 Accepted: August 28, 2014 Published: August 28, 2014 9895

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900

Analytical Chemistry

Article

among the pH, TA, DIC that can be determined is the pH. Indeed the pH can be measured by two different available techniques: potentiometry and spectrophotometry. Potentiometry is largely utilized but suffers from problems that may arise with pH electrodes such as potential drift, asymmetric potential variations and/or damage without mentioning the difficulty to operate them.9,16 The alternative method (spectrophotometry) that consists in adding a colored indicator in the aqueous solution and to measure in situ its absorption spectrum10,17−19 has also some limitations.18 The major drawback of these two techniques is that apart from giving the pH, they become useless to identify the chemical composition of the solution. In this Article, we report a novel approach to measure the dissolved inorganic carbon speciation of an alkaline aqueous solution in contact with CO2 under pressure up to 250 bar. This method based on Raman spectroscopy does not necessitate the use of any delicate electrode nor the addition of any colored indicator. It provides access to the molalities of the DIC species which in turn allows to determining the pH. The basic principle of this method can be understood by considering the following chemical equations CO2 (g) ↔ CO2 (aq)

(1)

Figure 1. Schematic experimental setup for in situ micro-Raman measurements of aqueous solutions in contact with CO2 at elevated pressure. The pathway of the laser beam is shown in green.

CO2 (aq) + H 2O ↔ H 2CO3 ↔ HCO3− + H+ ↔ CO32 − + 2H+

situ Raman analysis in a backscattering geometry have been recently reported on microreactors.20 We present here a new experimental setup which also works in a backscattering geometry but which uses a large cell in which it is possible to stir fluids under any pressurized gas. This setup based on the use of an inverted microscope is of great interest for measuring in situ liquid phase compositions or for analyzing microscopic samples or microscopic areas of larger samples. As shown in Figure 1, the laser beam is shined inside the cell after reflection on two mirrors fixed at 45° on the Raman stage of the BX41 microscope (Olympus Corporation, Japan). The laser beam thus penetrates inside the bottom of the cell through a 12.0 mm thick sapphire window. A ×10 objective having a focal distance of 10.6 mm was used to focus the beam at 1.35 mm (assuming 1.33 for refractive index in the liquid) away from the sapphire/ water interface. The scattered beam was also collected using the same setup with light traveling in the reverse direction. Raman spectra were collected on a T64000 Jobin-Yvon (Horiba, Japan) spectrometer and excited by the 514 nm wavelength radiation of an Ar/Kr laser with a 470 mW output power. Spectra were typically measured with an integration time of 10 s and averaged 10 times in the 100−2300 cm−1 range, and two times in the 2300−3820 cm−1 range. Deconvolution of the Raman peaks were carried out with the Origin 7.0 software. The stainless steel cell with a capacity of 60 mL (Separex, Champigneulles, France) was heated at 40.0 ± 0.5 °C. Liquid CO2 (Air Liquid, N45) is pumped by a cabestan pump (Separex) before feeding the autoclave equipped with a stirring mechanical device (Top-Industrie, Vaux le Penil, France) mounted on a vertical axis at the top of the cell. The axis of the magnetic stirrer was equipped with an anchor and the stirring speed could reach 3000 rpm. Twenty-five milliliters of aqueous solution was placed at the bottom of the reactor. Below 5 bar, the solution was slightly stirred before each measurements. Above 5 bar, to accelerate the equilibrium with CO2 in the vapor phase, the solution was stirred at each pressure at 1500 rpm for 10 min. During the measurements, the stirring was

(2)

These two equations which govern the chemical equilibriums arising from the interaction of water with CO2 dictate the value of the pH according to the classical relationship ⎛ (HCO3−) ⎞ pH = pK a0 + log⎜ ⎟ ⎝ (CO2 (aq) ⎠

(3)

where (HCO3−) and (CO2(aq)) are the activities of the 2 species and K0a is the equilibrium constant of the CO2(aq)/ HCO3− couple. The activity being the product of the molality by the activity coefficient, eq 3 proves that the pH can determined if the molalities of CO2(aq) and of the hydrogenocarbonate species are known. The objective of this paper is to demonstrate that under high CO2 pressure DIC speciation and pH can be determined by Raman spectroscopy. This technique uses usual optical probes and gives access to vibrational modes of molecules in solution and especially CO2(aq), HCO3−, and CO32− species. To validate this approach we first start with the description of the high pressure setup that we have designed to perform the measurements.

■

EXPERIMENTAL SECTION All experiments were done with an alkaline NaOH aqueous solution having a molal concentration equal to 0.377 mol·kg−1 of water (TA = 0.377 mol·kg−1) and an initial pH around 13.5 at pCO2 = 0 bar. The choice of this concentration was dictated by considerations that are explained in the Supporting Information (S1−3). Experiments were carried out in a backscattering geometry in an innovative dedicated cell shown in Figure 1. So far most of the Raman scattering experiments made in supercritical conditions were carried out by collecting the scattered light at 90° from the incident radiation. However, in 9896

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900

Analytical Chemistry

Article

cm−1, and the COO symmetric stretching mode νs(COO) at ∼1362 cm−1 (Supporting Information Figure S4b). The assignments are approximate because the modes are considerably coupled. Above 5 bar, CO2(aq) which coexists with HCO3− has a concentration high enough to be detectable. It is characterized by the Fermi diad Raman lines with frequencies ν− ∼ 1275 cm−1 and ν+ ∼ 1382 cm−1.22−24 As expected,7,8 their intensities increase almost linearly with pressure before leveling off circa the critical pressure of CO2 (Pc = 73.9 bar). In addition to ν− and ν+, a weak and broad peak centered at 1311 cm−1 labeled Σ was clearly found in the present work (see Supporting Information S5). Note that the position of these lines can slightly shift with increasing pressure or depending on the ionic strength of the solution. However, the shift remains barely visible up to 250 bar. From the above measurements, a major issue for the quantitative speciation and the pH determination is to probe the molality of each species. This can be achieved provided that the Raman line intensity Ii(ν) of a specific species i with a wavenumber ν can be determined in a absolute way. For this it is necessary to normalize Ii(ν) by a reference which is independent of the CO2 pressure. This reference could be either the ν(OH) stretching located in the 3000−3700 cm−1 range or the δ(HOH) bending modes of water at ∼1639 cm−1. We choose the intensity IH2O(δ(HOH)) of the δ(HOH) mode as the reference (see Supporting Information S1) as the vibration energy of this mode is close to the major vibration modes of CO2(aq), HCO3−, and CO32−. The use of this δ(H2O) internal standard is an advantage compared to spectrophotometry for which the addition of a color indicator is necessary. The normalized intensity is thus given by

stopped. Above 130 bar, the speed was reduced at 450 rpm to avoid the formation of CO2 bubbles on the sapphire window and the duration of stirring was therefore increased to 30 min. The pressure sensor accuracy is about 1 bar.

■

RESULTS AND DISCUSSION Figure 2a shows the Raman spectra of this aqueous solution with mNaOH = 0.377 mol·kg−1 as a function of CO2 pressure. At

Iinor(ν) = Ii(ν)/IH2O(δ(HOH))

(4)

Once normalized, the intensity Inor i (ν) is proportional to the molality mi of the species i (in mol.kg−1 water) and reads as

Figure 2. (a) In situ Raman spectra of the H2O−CO2 system with mNaOH = 0.377 mol·kg−1 water as a function of CO2 pressure at T = 40 °C. (b) Example of deconvolution of the a Raman spectrum at P = 31.5 bar with 3 components for HCO3− (in black) and 3 components for CO2(aq) (in gray). At this pressure, the amount of CO32− is negligible. The Raman spectrum is fitted with a background characterized by 2 broad bands (green) centered at 1040 (76 cm−1 fwhm) and 1320 cm−1 (158 cm−1 fwhm), respectively. (This last peak is not visible in this figure; see Supporting Information S4 for details.) (c) Molality of CO32− (open circles), HCO3− (full circles), and CO2(aq) (stars) deduced from the Raman spectra analysis. For a given species, the molality for each mode was found to vary linearly with the peak intensity whatever the mode. For sake of comparison, the molality of CO2(aq) in the H2O−CO2 system is plotted as a continuous line (see Supporting Information Figure S5 for details). Only coarse scale is given below 5 bar because of the poor accuracy in this pressure range.

Iinor(ν) = miJi (ν)

(5)

where Ji(ν) is the molal scattering coefficient of the Raman scattering signal. The determination of Ji(ν) for each vibration mode ν of each species is achieved by a calibration procedure with solutions of known concentrations illustrated in Figure 3. In Figure 3a, the Raman spectra of a solution with a molality of mCO32− = 0.394 mol·kg−1 is shown. This solution was obtained by dissolving Na2CO3 solid in water. A fit to the Raman spectrum using pseudo-Voigt functions gives access to the intensity of the ν1 mode of CO32− (only one mode for this species). This intensity is then divided by the integrated intensity of the bending mode of water δ(H2O) to get the normalized intensity Inor CO32− (ν1). This procedure was repeated for several known concentrations of CO32−. As shown in the 2− inset of Figure 3a, Inor CO32− (ν1) scales linearly with the CO3 molality and the slope gives the value of the molal coefficient − nor JCO 2− (ν1). The same approach was followed for HCO3 3 (Figure 3b) and CO2(aq) (Figure 3c). More details are given in Supporting Information S4 and S5. Table 1 summarizes the results of molal scattering coefficients for each mode of CO2(aq), HCO3−, and CO32− with δ(HOH) as the reference intensity. By using these molal coefficients, normalized Raman intensities of all detected species can be converted into molalities. Note that we assume here that the molal coefficient does not depend on pressure.

low CO2 pressure (P < 5 bar), the dissolution of CO2(gas) into the solution first induces the formation of CO32− characterized by its major stretching Raman line at 1067 cm−1. In agreement with its pseudo-D3h symmetry,21,22 it is assigned to the symmetric stretching C−O mode (Supporting Information Figure S4a). Because of its weak intensity, the broad nonsymmetric double mode22 of CO32− at ∼1378 and ∼1436 cm−1 can be ignored. At increasing pressure, the dissolution of CO2 acidifies the solution. As a consequence CO32− decreases and vanishes at the benefit of HCO3−. Assuming for this ion the C1 symmetry,22 three main Raman lines are expected as observed in Figure 2b: the C−OH stretching mode νs(C−OH) at ∼1015 cm−1, the COH bending mode δ(COH) at ∼1315 9897

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900

Analytical Chemistry

Article

half of the total alkalinity TA. This is in perfect agreement with the experimental value of the maximum of CO32−. By knowing the TA and the dominant species at a given pressure, it is thus possible to check from eq 6 if the molalities of CO32− and of HCO3− are correct. When the CO2 pressure increases, the molality of HCO3− grows at the expense of CO32−which disappears. It reaches a molality of 0.39(2) mol· kg−1 water close to TA fulfilling again eq 6 (region II in Figure 2c). This value is then constant at higher presssure. For pressure higher than 5 bar, the signature of CO2(aq) is then observed (region III in Figure 2c). Its molality (stars) is found to be about 13% below the one of the H2O−CO2 system (full line, see Supporting Information S5). This decrement in the molality of CO2(aq) results from the increment of the activity coefficient of CO2(aq) above unity (γCO2(aq) >1). For a constant activity of CO2, the higher the activity coefficient, the lower the molality.7 By considering Pitzer’s model25,26 with only the interactions of CO2(aq) with Na+ and HCO3−, γCO2(aq) is found to be about 1.08−1.09 (see Supporting Information S11). As a consequence, the molality of CO2(aq) in the H2O−CO2 alkaline system is reduced by about 9% which is close to our observation (−13% see region III in Figure 2c). On the contrary, the activity coefficient of HCO3− is below 1 because of ionic strength. Using the Debye−Hückel equation corrected by the B-dot equation,27,28 γHCO3− is about 0.71 and is almost constant with pressure (see Supporting Information S11). Knowing the molalities mHCO3− and mCO2(aq) and the activity coefficients γHCO3− and γCO2(aq) allows to determining the pH of the solution according to eq 3 in which the pKa0 of the CO2(aq)/HCO3− couple is taken to be equal to 6.30 at 40 °C and is assumed to be independent of the pressure (dash line in Figure 4a).29 pH values deduced from Raman measurements are shown in Figure 4 in triangles. In order to validate our novel approach of pH determination, these values were compared with those obtained by spectrophotometric absorption using color indicators: BTB and MR pH indicators (bromothymol blue, pKa0 = 6.99, methyl red, pKa0 = 5.34) respectively in the

Figure 3. (a) Fit of the Raman spectra of a 0.394 mol CO32− per kg water solution issued from the dissolution of Na2CO3 at ambient pressure and 40 °C. There is no Raman signal for the HCO3− species, as expected with the fact that the pH is high (pH ≈ 11.7). In inset is plotted the normalized intensity of the ν1 line as a function of the molality of CO32−. (b) Fit of the Raman spectra of a 0.376 mol HCO3− per kg water solution issued from the dissolution of NaHCO3 at ambient pressure and 40 °C. There is no Raman signal for the CO32− species, as expected with the fact that the pH is low (pH ≈ 8.3). In inset is plotted the intensity of the 3 modes of HCO3− as a function of the molality of HCO3−. (c) Fit of the Raman spectra of a 0.80(1) mol CO2(aq) per kg water solution at P = 41.8(5) bar and 40 °C. In inset is plotted the evolution of the normalized intensity of each peak (ν−, ν+, and Σ) as a function of molalities extracted from the works of Duan et al.7 and of Diamond et al.8 Numbers in parentheses stand for the uncertainties. The Raman spectrum contains a bumpy background characterized by 2 broad bands (green) centered at 1040 (76 cm−1 fwhm) and 1320 cm−1 (158 cm−1 fwhm), respectively. For panels a−c, the fit curves are in red.

Table 1. Vibrational Assignments and Spectral Parameters of CO2(aq), HCO3−, and CO32− Species in Aqueous Solution species

νmax (cm−1)

fwhm (cm−1)

assignment

Ji (kg mol−1)

CO2(aq)

1275 1382 1311 1067 1015 1315 1362 1639

22.6(5) 15.5(5) 73.3(0) 13.1(5) 29.2(5) 79.0(0) 36.7(0) 105(0)

ν− ν+ Σ ν1: νσ(CO) ν5: νs(C−OH) ν4: δ(COH) ν3: νs(CO2) δ(HOH)

0.468(8) 1.242(15) 0.129(4) 1.71(3) 1.09(3) 0.49(3) 0.89(3) 0.018

CO32− HCO3−

H2O

Molalities of all species in our alkaline solution exposed to CO2 pressure are plotted as a function of the CO2 pressure in Figure 2c. This figure shows at once what are the dominant species at a given pressure. To fully understand the meaning of the values shown in this graph, let us recall that the molalities of the dissolved inorganic carbon species are constrained by the total alkalinity TA. In our system, the total alkalinity TA is defined by TA = mOH− + mHCO3− + 2mCO32− − m H+ = m Na+

(6)

Figure 4. (a) pH of the H2O−CO2 alkaline system with mNaOH = 0.377 mol·kg−1 water as a function of CO2 pressure, obtained by Raman spectroscopy (triangles), color indicator with Bromothymol Blue (BTB, circles), and color indicator with methyl red (MR, stars). (b, c) Spectrophotometry measurements on the solution with mNaOH = 0.377 mol.kg−1 water with CO2 pressure using BTB (b) and MR (c). See Supporting Information S6−S10 for details. In panel b, photos of the solution in the cell at 0 bar (left) and 48.1 bar (right). In panel c, photos at 16.4 and 81.5 bar.

(mi is the molality of the i species) and is equal to 0.377 mol· kg−1 whatever the CO2 pressure. At P < 5 bar, CO32− dominates with a maximum molality of 0.18(1) mol·kg−1 water (region I in Figure 2c). When CO32− molality reaches its maximum, the molalities of OH−, H+, and HCO3− species are negligible in comparison to the one of CO32−. The molality of CO32− defined by eq 6 is hence equal to 9898

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900

Analytical Chemistry

■

1−50 and 30−100 CO2 pressure ranges. Combining these 2 color indicators was necessary to cover a pH range from 4.5 to 7.8 (Figure 4b, c). See Supporting Information S6−S10 for details. pH values determined by the two methods are clearly in good agreement. This ascertains the usefulness of Raman spectroscopy to identify both the dissolved inorganic carbon speciation and the pH values.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. Phone: +33 (0)2 43 83 32 67. *E-mail: [email protected]. Phone: +33 (0)2 43 83 32 62. Notes

The authors declare no competing financial interest.

■

■

CONCLUSION In this work, it has been evidenced for the first time to the best of our knowledge that Raman spectroscopy can be used to measure in situ both the molalities of CO2(aq), CO32−, and HCO3− and the pH of an alkaline solution as a function of the CO2 pressure inside an innovative cell specifically designed for this purpose. In a solution with a constant alkalinity of 0.377 mol.kg−1, we show that at low pressure, the dissolution of CO2(g) into the solution induces first the formation of CO32− with a maximum molality equal to half the alkalinity (mCO32− ≈ TA/2). Then the further acidification of the solution reduces the CO32− molality and increases the one of HCO3− which becomes nearly constant and equal to the alkalinity (mHCO3− ≈ TA). Above 5 bar, CO2(aq) species get concentration high enough to be detectable. As expected, its molality increases almost linearly with pressure before leveling off circa the critical pressure of CO2 (Pc = 73.9 bar). The presence of CO2(aq) and HCO3− in combination with the knowledge of the activity coefficients allows to determining in an unprecedented way the pH of the solution inside the pressurized cell. It is shown that the pH values obtained by Raman are quite similar to those determined by spectrophotometric measurements obtained with color indicators (methyl red and bromothymol blue). In addition to measuring the pH of an aqueous solution under CO2 pressure, this method is opening tremendous opportunities to examine the thermodynamics of carbonate species in various ionic media as a function of temperature and ionic strength. Among the many examples of application are the in situ monitoring of the interaction of CO2 with materials as for instance brines, the peering into the in situ synthesis of CaCO3 under pressurized CO2 conditions or the study of the conversion of CO2 into valuable chemicals30 via its catalytical reduction under pressure in aqueous solution.

■

Article

ACKNOWLEDGMENTS The authors wish to thank the ANR for funding through the Calcomed (ANR-09-PIRI-0004) and CGSμLab programs and the Region des Pays de la Loire through the Bioregos program.

■

REFERENCES

(1) Herzog, H. J. Environ. Sci. Technol. 2001, 35, 148A−153A. (2) Barbero, R.; Carnelli, L.; Simon, A.; Kao, A.; Monforte, A.; Riccò, M.; Bianchi, D.; Belcher, A. Energy Environ. Sci. 2013, 6, 660−674. (3) Beuvier, T.; Calvignac, B.; Delcroix, G.; Tran, M.-K.; Kodjikian, S.; Delorme, N.; Bardeau, J.-F.; Gibaud, A.; Boury, F. J. Mater. Chem. 2011, 21, 9757−9761. (4) Hofmann, M.; Schellnhuber, H. J. Energy Environ. Sci. 2010, 3, 1883−1896. (5) Hönisch, B.; Ridgwell, A.; Schmidt, D. N.; Thomas, E.; Gibbs, S. J.; Sluijs, A.; Zeebe, R.; Kump, L.; Martindale, R. C.; Greene, S. E.; Kiessling, W.; Ries, J.; Zachos, J. C.; Royer, D. L.; Barker, S.; Marchitto, T. M., Jr.; Moyer, R.; Pelejero, C.; Ziveri, P.; Foster, G. L.; Williams, B. Science 2012, 335 (6072), 1058−1063. (6) Schuchmann, S.; Schmitz, D.; Rivera, C.; Vanhatalo, S.; Salmen, B.; Mackie, K.; Sipilä, S.; Voipio, J.; Kaila, K. Nat. Med. 2006, 12 (7), 817−823. (7) Duan, Z. H.; Sun, R. Chem. Geol. 2003, 193, 257−271. (8) Diamond, L. W.; Akinfiev, N. N. Fluid Phase Equilib. 2003, 208, 265−290. (9) Shao, H.; Ray, J. R.; Jun, Y.-S. Environ. Sci. Technol. 2010, 44 (15), 5999−6005. (10) Toews, K. L.; Shroll, R. M.; Wai, C. M.; Smart, N. G. Anal. Chem. 1995, 67 (22), 4040. (11) Egleston, E. S.; Sabine, C.; Morel, F. M. M. Global Biogeochem. Cycles 2010, 24, GB1002. (12) Keller, D. P.; Feng, E. Y.; Oschlies, A. Nat. Commun. 2014, 5 (3304), 1−11. (13) Kirchofer, A.; Brandt, A.; Krevor, S.; Prigiobbe, V.; Wilcox, J. Energy Environ. Sci. 2012, 5, 8631−8641. (14) Mitchell, A. C.; Phillips, A.; Schultz, L.; Parks, S.; Spangler, L.; Cunningham, A. B.; Gerlach, R. Int. J. Greenhouse Gas Control 2013, 15, 86−96. (15) Shao, H.; Thompson, C. J.; Cantrell, K. J. Chem. Geol. 2013, 359, 116−124. (16) Peng, C.; Crawshaw, J. P.; Maitland, G. C.; Trusler, J. P. M.; Vega-Maza, D. J. Supercrit. Fluids 2013, 82, 129−137. (17) Liu, X. Z.; Wang, Z. A.; Byrne, R. H.; Kaltenbacher, E. A.; Bernstein, R. E. Environ. Sci. Technol. 2006, 40 (16), 5036−5044. (18) Shao, H.; Thompson, C. J.; Qafoku, O.; Cantrell, K. J. Environ. Sci. Technol. 2013, 47, 63−70. (19) Hopkins, A. E.; Sell, K. S.; Soli, A. L.; Byrne, R. H. Mar. Chem. 2000, 71, 103−109. (20) Liu, N.; Aymonier, C.; Lecoutre, C.; Garrabos, Y.; Marre, S. Chem. Phys. Lett. 2012, 551, 139−143. (21) Davis, A. R.; Olivier, B. G. J. Solution Chem. 1972, 1 (4), 329− 338. (22) Rudolph, W. W.; Fischer, D.; Irmer, G. Appl. Spectrosc. 2006, 60, 130−144. (23) Fermi, E. Z. Phys. 1931, 71, 250−259. (24) Anderson, G. R. J. Phys. Chem. 1977, 81 (3), 273−276. (25) Pitzer, K. S. J. Phys. Chem. 1973, 77, 268−277. (26) Harvie, C. E.; Møller, N.; Weare, J. H. Geochim. Cosmochim. Acta 1984, 48, 723−751.

ASSOCIATED CONTENT

S Supporting Information *

Overview of the Raman spectrum of H2O−CO2 system showing the contribution of the sapphire window together with the dominant mode of water at 3200 cm−1; what is happening for the H2O−CO2 alkaline system at very high concentration of NaOH; overview of the Raman spectrum of the H2O−CO2 alkaline system with mNaOH = 3.8 mol·kg−1 showing the contribution of NaHCO3 solid; on the calibration of intensities measured on the Raman signals; the vibrational modes of CO2(aq); on the way to determine the pH of a NaOH solution by spectrophotometry using BTB indicator; on the way to determine the pH of the NaOH solution using methyl red indicator; relation between pH and the absorbance ratio R; determination of the pH under pressure via color indicators; uncertainty of the pH under pressure; on the determination of the activity coefficients of HCO3− and CO2(aq). This material is available free of charge via the Internet at http://pubs.acs.org. 9899

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900

Analytical Chemistry

Article

(27) Helgeson, H. C.; Kirkham, D. H. Am. J. Sci. 1974, 274, 1199− 1261. (28) Truesdell, A. H.; Jones, B. F. U. S. Geol. Survey J. Res. 1974, 2, 233−248. (29) Millero, F. J.; Graham, T. B.; Huang, F.; Busto-Serrano, H.; Pierrot, D. Mar. Chem. 2006, 100, 80−94. (30) Aresta, M.; Dibenedetto, A.; Angelini, A. Chem. Rev. 2014, 114, 1709−1742.

9900

dx.doi.org/10.1021/ac5025446 | Anal. Chem. 2014, 86, 9895−9900