Measurement of Solubility of CO 2 in NaCl, CaCl 2 , MgCl 2 and MgCl 2 + CaCl 2 Brines at Temperatures from 298 to 373 K and Pressures up to 20 MPa Using the Potentiometric Titration Method

: Understanding the carbon dioxide (CO 2 ) solubility in formation brines is of great importance to several industrial applications, including CO 2 sequestration and some CO 2 capture technologies, as well as CO 2 -based enhanced hydrocarbon recovery methods. Despite years of study, there are few literature data on CO 2 solubility for the low salinity range. Thus, in this study, the solubility of CO 2 in distilled water and aqueous ionic solutions of NaCl, MgCl 2 , CaCl 2 and MgCl 2 + CaCl 2 were obtained in a low salinity range (0–15,000 ppm) at temperatures from 298–373 K and pressures up to 20 MPa using an accurate and unconventional method called potentiometric titration. An experimental data set of 553 data points was collected using this method. The results of the experiments demonstrate that increasing pressure increases the solubility of CO 2 in various brines, whereas increasing temperature and salinity reduces the solubility. The role of different ions in changing the solubility is elaborated through a detailed discussion on the salting-out effect of different ionic solutions. To verify the experimental results of this research, the solubility points obtained by the potentiometric titration method were compared to some of the well-established experimental and analytical data from the literature and a very good agreement with those was obtained.


Introduction
It is unanimously accepted that global warming and its dire consequences have become a serious problem for the whole world. Carbon dioxide (CO 2 ), which accounts for over 62% of all greenhouse gases, has a significant impact on global warming [1]. The primary source of CO 2 emissions is anthropogenic activities, as well as deforestation due to land clearing and a number of production and resource extraction processes [2]. The use of fossil fuels as the main source of energy increases the level of CO 2 in the atmosphere. According to Apadula et al. [3], CO 2 concentration in the atmosphere gradually increases with the growth rate of 2.05 ± 0.03 ppm/year. The key mechanism for mitigating the greenhouse effect is to considerably lower CO 2 emissions into the atmosphere. Various brine, CO 2 solubility data are shown at 14 points of state, 36 points in CaCl 2 and 38 points in MgCl 2 . The findings greatly broaden the range of conditions in which CO 2 solubility in these brines may be determined (temperature, pressure and molality). Furthermore, the results demonstrate that CO 2 solubility in CaCl 2 and MgCl 2 brines of the same molarity are, in fact, very close. Drummond [28] experimentally measured numerous CO 2 solubilities in NaCl-saturated brines. The latter is among the most complete experimental database of solubility; however, since several presumptions were used to calculate the solubility, the accuracy of the data is dubious.
The majority of previous studies on CO 2 solubility in literature assumed NaCl is the only constituent of formation brines. However, there are many brine formations around the world in which a considerable number of other salts, such as MgCl 2 and CaCl 2 , can be found [22,27]. Furthermore, despite years of prior research on CO 2 solubility in ionic liquids over a wide range of pressures, temperatures and salinity (see Table 1), there are still research gaps that need to be filled. While a wide range of salinity has been explored in the literature, evidence on solubility in the low salinity region is scarce, for example, in the range from 0 ppm to 15,000 ppm (from 0 to 0.258 mol.kg −1 ) brine salinity. The Sabah basin, off the coast of Peninsular Malaysia, contains such geological formations, with a mean salinity of roughly 10,000 ppm (0.17 mol.kg −1 ) [29]. As a result, one of the goals of this study is to develop a study on solubility applicable to the injection of CO 2 into low salinity subsurface formations.
This study uses an unconventional solubility measurement method, i.e., potentiometric titration, to determine the solubility of CO 2 in brine. The technique described above is typical in chemical engineering, although it is rarely employed in research on CO 2 sequestration/injection into subsurface formations. Furthermore, because there is a scarcity of data in the literature on CO 2 solubility in brines with low salinity, in this study, the solubility of CO 2 was computed in brines saturated with NaCl, CaCl 2 , MgCl 2 and MgCl 2 + CaCl 2 in the low salinity range of 0-1.5 wt.%. The salting-out effect, which is a measure of decreasing solubility by increasing salinity, is studied in depth. Although the findings of the current study are mainly aimed at CO 2 sequestration projects, such as CO 2 injection into subsurface formation, the results could be of importance for other applications, such as CO 2 capture technologies [30], CO 2 mineral carbonation [31] and food industry [32].

Materials
SIGTM provided CO 2 (purity > 99.9%), which was used in all of the experiments. The brines of salinity and composition were made with distilled and deionized water (Milli-Q filter) with a resistance of 18.20 ohms. Systerm TM provided NaCl, MgCl 2 and CaCl 2 with a mass fraction purity of 0.99. No further purification or alternation was performed on the chemicals. NaOH and HCL were used with a purity of 99.7 and were purchased from Emsure TM . The reactor was made of stainless steel with a pressure rating of 45 MPa and a temperature rating of 400 K. It also had sufficient resistivity towards corrosive materials it might have come in contact with during the experiments. Figure 1 depicts a schematic of the experimental setup used in this investigation. A Teledyne ISCO pump, a CO 2 bottle and a 0.1 L autoclave reactor with a magnetic stirrer were the key components. An electric heater was used to keep the reactor warm. A dip tube was attached to a floating piston sampler made locally and powered by a med TM syringe pump. An immersion tube was used to sample the CO 2 -saturated saline solution. An ISCO pump regulated the pressure in each experiment. The reactor was equipped with a temperature jacket with an accuracy of 273.25 K according to the instructions for use of the device. Despite the 0.85 cm thickness of the reactor's base, a proper "connection" between the magnetic stirrer and the stirrer ball was accomplished, resulting in a well-mixed solution. The brine was poured into the reactor and warmed to the desired temperatures. CO 2 was then supplied at the desired pressure into a warmed reactor containing 70 mL of brine. After that, the reactor's inlet and outlet valves were closed and the solution was stirred for 3 h until it achieved equilibrium. Equilibration times have been observed to range from 10 min to 24 h in previous studies [41]. Thereafter, the bottom valve in the reactor was gently opened to minimize the pressure change in the reactor. An immersion tube was used to transport a sample of CO 2 -saturated brine from the reactor to the sampling chamber, which had a floating piston. As soon as the sample entered the chamber, it reacted with the 0.5 M NaOH solution, filling half of the chamber.

Experimental Methods
The sampling cylinder was half-filled with brine that was later removed to let the CO 2 saturated sample from the reactor be mixed with the NaOH solution via a syringe pump. Because there was an abundance of NaOH in the solution, it dissolved all types of carbon particles and converted them to carbonates, resulting in no bubble gas [41]. The sample was then titrated with hydrochloric acid (HCl) until it reached the equivalency points. Once the endpoint of the reaction was reached, the volume of titrant was measured and the solubility of CO 2 was calculated using Equation (1): where C a is the analyte concentration (solubility of CO 2 ) in the brine (in mol.kg −1 ), C t is the titrant concentration (in mol.L −1 ), V t is the volume of titrant (mL), N is the molar ratio of analyte and reactant from a balanced chemical equation and M a is the mass of the sample to be titrated (grams). The advantage of utilizing the solution mass rather than the solution volume V a , as has been performed in many earlier research studies, is that the mass of the solution is not affected by temperature or pressure. As a result, the calculation of solubility is less ambiguous. The titrant was (0.5 M) HCl, which was utilized to react in a 5 mL sample. The pH of the sample was determined as a function of the titrant volume given and the titration was maintained until the pH reached values below 2; plotting the derivative of HCl volume versus pH yielded equivalence points. In addition to the technical simplicity, another benefit of this method compared to previous methods by which the solubility is measured is that the preservation of the samples inhibits the loss of CO 2 because of the degassing during the depressurization phase. Moreover, in contrast to previous research studies, the potentiometric method, unlike a number of previous methods, does not depend on any additional parameters, such as fugacity, density, or volatility, to be able to estimate solubility accurately.
To ensure repeatability and accuracy of the results, several experiments were repeated 3 times. The errors in the measurements were found to be 0.9-7.8%. As expected, the highest error occurred with the measurements near atmospheric pressure (0.1 MPa), regardless of the temperature, salinity and type of brine. The errors were markedly lower when the pressure in the experiments exceeded 0.2 MPa. The reason for this phenomenon was that the very low values of CO 2 solubilities close to atmospheric pressure were in the order of a thousandth of mol/kg, as compared to solubility values at higher pressures. Therefore, these low values could not be accurately detected using the experimental method used in this study. It is noteworthy that the focus of our study was the solubility of CO 2 in subsurface formations; in those scenarios, CO 2 is often injected at high pressures, hence it is in liquid or supercritical fluid state [42]. Therefore, near-atmospheric measurements are of a little significance for the aforementioned applications.

Results
Despite the fact that there have been several investigations on CO 2 solubility in various liquids, evidence in the low salinity range is limited. As a result, the impact of pressure change on CO 2 solubility was investigated under a variety of conditions in the current study. The experiments were carried out at pressures ranging from 1 to 20 MPa and temperatures ranging from 298 to 373 K. Furthermore, the experiments were carried out in a saline solution of different values of salinity (0-15,000 ppm) to confirm the reliability of the results under conditions more representative of CO 2 injection to subsurface formations (aquifers and hydrocarbon reservoirs). Likewise, the solubility of CO 2 in distilled water was tested under identical temperature and pressure conditions (1-20 MPa, 298-373 K). The outliers in the solubility databank, i.e., data with unusually high or low values, were identified through analyzing the z-score of the data points. The data points with unusually high or low values were treated as outliers and hence removed from the database using the z-score method that was applied with SPSS 18TM. The abnormal solubility data points were mostly caused by the rapid opening of the sampling valve, which resulted in a significant pressure decrease in the solution and subsequent supercritical CO 2 breakthrough.

Impacts of Pressure and Temperature on the Solubility
CO 2 solubility in distilled water, NaCl, MgCl 2 , CaCl 2 and MgCl 2 + CaCl 2 is shown in Figure 2 in four different temperature series, namely, 298, 333, 353 and 373 K, respectively. It can be seen, from the figures, that increasing the pressure increased the solubility of CO 2 in the brine, irrespective of the type of ionic solution and temperature. In addition, it is apparent, from the figures, that the pressure dependence of carbon dioxide decreased with increasing the pressure in all temperature series. However, in this study, the points at which the solubility would become entirely unresponsive to pressure were not observed. The same effect (pressure insensitivity at higher pressures) was reported for pressure around 30 MPa by previous researchers who used different solubility measurement methods at different temperatures, pressure and salinities [11,27]; however, as the highest point of solubility measurement in this study was 20 MPa, the point of pressure insensitivity was not observed.
The effect of pressure on solubility can be expressed using Henry's law of solubility (partial pressures) [43]. According to Henry's Law, the partial pressure of the gas above the solution determines the solubility of the gas in the water. Because the concentration of molecules in the gas phase increases as pressure increases, the concentration of dissolved gas molecules in the solution at equilibrium also increases. When a gas is introduced to a system that is primarily made up of brine (solvent), some of the gas molecules collide with the liquid's surface and dissolve. When the concentration of dissolved gas molecules rises to the point where the rate at which gas molecules escape into the gas phase equals the rate at which it dissolves, dynamic equilibrium is reached. As the gas pressure rises, the amount of gas molecules per unit volume rises, increasing the rate at which gas molecules collide with the liquid's surface and dissolve. The concentration of dissolved gas rises as more gas molecules dissolve at higher pressures, until a new dynamic equilibrium is reached [43].
When it comes to the effect of temperature, it can be seen that, as the temperature rose, the solubility decreased. CO 2 solubility in NaCl is 1.421 mol kg −1 at 10.33 MPa and 298 K, whereas it is 1.037 mol kg −1 , 0.877 mol kg −1 and 0.788 mol kg −1 at the same pressure and 353 K, 353 K and 373 K, respectively. In other words, there is a decrease in solubility of 27%, 38% and 44.54% as the temperature rises to 333 K, 353 K and 373 K from the initial value of 298 K. Previous researchers have also reported a decrease in the solubility with the increase in the temperature [12,44,45]. Le Chatelier's law could explain the reduction in solubility at higher temperatures. CO 2 dissolves in brine due to the interactions of the molecules of solute with those of solvents. The process of dissolving CO 2 in brine is exothermic (∆H reaction < 0), which implies that heat is produced as new attractive contacts arise as a result of the dissolution process [35]. According to the principle of Le Chatelier, if the system is heated, since this is an exothermic reaction, the system shifts towards the reactant's side to neutralize the applied stress, which is the rise in the temperature. The kinetic energy of a system increases as the temperature rises. As the temperature rises, this causes a more rapid motion among the molecules and the breakage of intermolecular bonds, allowing molecules to escape to the gas phase from the solution [46]. As a result, independently from pressure, type of ion in the brine, or brine salinity, increasing the temperature decreased the solubility in this set of experiments.

Effects of Salinity on CO 2 Solubility
The solubility of CO 2 versus pressures at 0, 1000, 10,000 and 15,000 ppm in NaCl, MgCl 2 , CaCl 2 and MgCl 2 + CaCl 2 solutions at 298 K is shown in Figure 3. Figure 3 shows the reduction in CO 2 solubility in formation brine as the salinity increased at various pressures and temperatures for all types of aqueous solutions employed in the current study. The solubility of CO 2 decreased by 1% with an increase in the salinity from 0 to 1000 ppm for brine solutions in the experiments conducted in this work, while a decrease in the solubility of 3-6% was found with a factor of 10 increase in brine concentration (from 1000 to 10,000 ppm). Moreover, increasing the brine concentration from 10,000 to 15,000 ppm resulted in a 4-5% decrease in CO 2 solubility. The range of reduction in solubility as the salinity increased is in line with those reported in the literature when pressure and temperature were set in the same range of this study [6,25,47]. The decrease in the solubility can be explained by the fact that when salt ions such as NaCl, MgCl 2 and CaCl 2 are added to water, they bind water molecules to "solvates", leaving less water for CO 2 to adhere to. In other words, the presence of water molecules in the solvation of ions significantly lowers CO 2 molecules' weak attraction to water/brine and displaces dissolved CO 2 from polar water. When solutes such as NaCl, MgCl 2 and CaCl 2 (or any combination of them) are present, the solubility of CO 2 in brine is greatly impacted. In reality, because of the enhanced salting-out effect, the solubility reduces as the salinity rises (the salting-out effect is discussed in detail in the next part of the results). Similar results can be observed in the previous studies in which different solubility measurement methods (such as depressurization) were used to measure CO 2 solubility in brines which were significantly more saline than the brine used in this study [37,45].

Salting-Out Effect
As the concentration of dissolved solids in the brine rises, the salting-out effect reduces CO 2 solubility in aqueous solutions (in this case, brine). The effect is significant because it aids in quantifying the decrease in CO 2 solubility as salinity rises. Studies on the hydration of ions and the interaction of ions with water molecules have shown that, at a high density, smaller ions tend to bind the molecules of water more effectively, while larger ions with a low charge density bind the water molecules weakly [48,49]. Therefore, high charge density ions have a robust impact on the structure of the water, which governs the ability of the brine to dissolve higher amounts of CO 2 . The experiments show that, if two single-salt aqueous solutions have the same electrolyte type and share the same anion (e.g., Cl − ), the cation with a higher charge density (smaller radius and grater charge) has a greater saltingout effect on dissolved CO 2 than the cation with a lower charge density (larger radius and lower charge). For instance, Mg 2+ has a charge density that is a little higher than that of Ca 2+ (they have the same charge number, but Mg 2+ has a smaller radius than Ca 2+ ) [50]; hence, the amount of CO 2 dissolved in aqueous MgCl 2 is less than that in aqueous CaCl 2 at the same ionic strength. The solubility of CO 2 in an aqueous solution of NaCl follows a similar pattern. Because Na + has a lower charge density than Mg 2+ , CO 2 is substantially more soluble in aqueous NaCl solutions than in aqueous MgCl 2 solutions [21]. Figure 4 depicts the salting out effect in NaCl, MgCl 2 , CaCl 2 and MgCl 2 + CaCl 2 solutions with concentrations ranging from 1000 to 15,000 ppm, at pressures ranging from 1 to 20 MPa and temperatures of 298 K. Equation (2) was used to calculate the percentage of salting-out effect (S-O%).
where S-O (%) is the salting out percentage, x Dw is CO 2 solubility in distilled water and x b is CO 2 solubility in any brine. Figure 4 shows that, at a 1000 ppm salinity, there was no substantial change in solubility. As a result, in low-salinity brine, the salting-out effect is insignificant. The salting-out effect, on the other hand, increased as the concentration of solids in the brine rose. The maximum percentage of S-O is found in the 15,000-ppm data series, where the effect reached 6-9% for all solutions at lower pressures, whereas the lowest percentage of S-O is found in brine with a concentration of 1000 ppm, where S-O fluctuated between 0.25% and 0.65%. This result is in line with the findings of Tong et al. [27], who found that increasing the temperature enhances the salting-out effect, while increasing the pressure tends to diminish it.

Comparison of Experimental Results with Previous Studies
The CO 2 solubility data points obtained using the potentiometric titration method were compared with literature data obtained well-established methods under similar conditions to examine the accuracy of the findings of the experimental method used in this study, even though the data points obtained under the same conditions of pressure, temperature and, specifically, salinity (low salinity range) were very limited. As shown in Figure 5, the experimental results of this study are in good agreement with those obtained in previous studies using more common solubility measurement methods, such as depressurization or combinations [37] and depressurization [39] methods. In Figure 5, the solid lines indicate a regression line drawn based on the data obtained in the current study and error bars indicate a 5% difference from the experimental data of this study.
The data points obtained by Duan and Sun [12] and Li et al. [39] were measured at 323 K and 332 K, respectively. The difference among the results obtained in this study, Duan and Sun's and Lee et al.'s, is most likely due to the difference in the experimental temperature in our study, which was 333 K vs. 323 K and 332 K in Duan and Sun and Li et al., respectively. Furthermore, as shown in Figure 5, comparing the measurement data from this investigation with the data from Cruz et al. [6] supports the accuracy of the solubility points obtained in this work. As a result, the potentiometric titration method produces reliable results and could be used as an alternative to some of the previous complex and often expensive methods required to accurately measure CO 2 solubility in a condition representative of CO 2 injection/sequestration to subsurface geological formations, such as aquifers and mature hydrocarbon fields.
There is a scarcity of experimental data in a similar range of pressure, temperature, salinity and composition, making it impossible to compare the findings of this work with other studies. For two MgCl 2 and CaCl 2 brines, Figure 6 shows a comparison of the solubility results from this study with solubility estimates from Duan and Sun's (2003) theoretical model. Figure 7 also illustrates the parity-plot at the same conditions of pressure, temperature, salinity and brine type. The solubility values between the two models are almost perfectly in agreement in both brines, with a coefficient of correlation of more than 99% (R 2 > 0.99).

Field Implications and Recommendations for Further Studies
The impacts of pressure, temperature, salinity and brine composition are elaborated in the previous sections of this study. In geological formations, the salinity increases with depth. Since the solubility reduces with the increase in the salinity (irrespective of the type of salt), it can therefore be concluded that, in deeper geological formations, the contribution of the solubility mechanism weakens. Therefore, from point of view of CO2 solubility mechanisms, shallower formations, or depositories with low salinities (such as hydrocarbon fields in the Sabah basin, offshore Sabah, Malaysia) would be more suitable for CO2 sequestration, as larger amounts of CO2 can be rendered immobile using the solubility mechanism. With regards to temperature, as shown in this study, the increase in temperature reduces the solubility of CO2 in brine. Hence, it could be concluded, although only from the point of view of pressure, that deeper formations are more suitable, as the effects of temperature and salinity render the solubility mechanism less effective. The effect of pressure, on the other hand, is favorable on the solubility mechanisms. Hence, quantification of the impact of each parameter on the solubility needs to be conducted to find the optimum depth for sequestration. On the other hand, for

Field Implications and Recommendations for Further Studies
The impacts of pressure, temperature, salinity and brine composition are elaborated in the previous sections of this study. In geological formations, the salinity increases with depth. Since the solubility reduces with the increase in the salinity (irrespective of the type of salt), it can therefore be concluded that, in deeper geological formations, the contribution of the solubility mechanism weakens. Therefore, from point of view of CO 2 solubility mechanisms, shallower formations, or depositories with low salinities (such as hydrocarbon fields in the Sabah basin, offshore Sabah, Malaysia) would be more suitable for CO 2 sequestration, as larger amounts of CO 2 can be rendered immobile using the solubility mechanism. With regards to temperature, as shown in this study, the increase in temperature reduces the solubility of CO 2 in brine. Hence, it could be concluded, although only from the point of view of pressure, that deeper formations are more suitable, as the effects of temperature and salinity render the solubility mechanism less effective. The effect of pressure, on the other hand, is favorable on the solubility mechanisms. Hence, quantification of the impact of each parameter on the solubility needs to be conducted to find the optimum depth for sequestration. On the other hand, for enhanced oil recovery (EOR) applications, lower solubility of CO 2 in formation brines is more favorable for CO 2 -based EOR methods, as the amount of available CO 2 to interact with and eventually mobilize the residual oil is higher. Hence, it is recommended, for future studies, to address this issue by proposing a selection criterion for CO 2 sequestration, CO 2 injection and a combination of the two.
Moreover, in most studies focusing on CO 2 sequestration, CO 2 is considered to be 100% pure. This, however, is not often the case for practical scenarios in which CO 2 could be stemmed from different industries (such as gasification, post-combustion CO 2 capture, sour gas processing, or even recycled CO 2 from EOR operations) [51]. Moreover, it is not uncommon for natural CO 2 from subsurface formation to have associated gases. The CO 2 stream may contain several impurities, such as H 2 S, N 2 , Ar, etc. [51], and it might be economically and technically viable to consider CO 2 injection with no further purification. Therefore, it would be of great industrial importance to be able to study the solubility of CO 2 stream containing impurities, so that a more realistic estimation of the solubility mechanisms and, ultimately, sequestration efficiency can be made.

Conclusions
Using potentiometric titration, 553 data points of CO 2 solubility in various brines (NaCl, MgCl 2 , CaCl 2 and MgCl 2 + CaCl 2 ) were obtained at temperatures ranging from 298 to 373 K and pressures up to 20 MPa. In comparison with earlier traditional procedures, the new method is shown to be reproducible and accurate. The findings are in good accord with those of prior studies at the same pressure, temperature, salinity and brine composition ranges (salt type). In terms of pressure, independently from the salinity of the brine composition, it is obvious that increasing the pressure increases CO 2 solubility in aqueous solutions. However, the pressure dependence of solubility decreased with the increase in the pressure, although, in this study, the point at which the solubility becomes completely independent from the pressure was not detected. The temperature has a reverse effect on the solubility; as the temperature increased, the solubility was significantly reduced, regardless of the salinity and the composition of the brine. Finally, increasing salinity also negatively affects solubility, though, at a low salinity, the effect was not so noticeable. However, the solubility decreased by over 6% as the salinity of brine increased from 0 to 15,000 ppm. Moreover, the effect of the presence of divalent and monovalent ions in brine is here discussed in detail with reference to the salting-out effect. Under the same conditions of pressure and temperature, both ion charge and ion radius were found to be influential factors in the solubility of CO 2 in brine. Lastly, the comparison of the experimental results obtained from the potentiometric titration method, as an unconventional method, with those of more conventional and well-established methods from the literature proved the method to be accurate and reliable.