CO 2 hydrate formation in NaCl systems and undersaturated aqueous solutions

. A high-pressure experimental setup was used to obtain experimental data on the Hydrate Equilibrium (melting) Temperature (HET) of aqueous phases undersaturated with CO 2 , at different CO 2 saturation levels and with different aqueous phase compositions (sodium chloride and sodium chloride/calcium chloride). This paper details the experimental equipment, methods, test ﬂ uids, and results. To study the phase equilibria of CO 2 -brine systems and further validate the results, hydrate dissociation conditions of CO 2 in different concentrations of NaCl aqueous solutions were measured in the low-temperature region and over a wide range of pressure. Experimental results were compared against predictions of the simpli ﬁ ed Cubic Plus Association Equation of State (sCPA-EoS) coupled with the van der Waals and Platteeuw solid solution theory.

1 Introduction CO 2 injection in geological formations is a key aspect of the Carbon capture and Storage (CCS) strategy. The captured CO 2 will be injected into depleted oil and gas reservoirs and/or saline aquifers. The injection and potential CO 2 expansion can be associated with a large Joule-Thomson effect and may in turn pose a risk of hydrate or ice formation in the formation due to the cooling effect [1]. In addition, CO 2 can be injected/stored through Carbonated Water Injection technology [2], where CO 2 from offshore power production is captured from flue gas and compressed/injected into the geological formation as dissolved gas in the injection water. Hydrate could be formed in the injection water with dissolved CO 2 .
Although different thermodynamic models have been developed to predict the carbon dioxide hydrate stability in aqueous NaCl solutions, high-accuracy experimental data are always necessary to validate or improve these predictive tools. Experimental data for carbon dioxide hydrates in equilibrium with sodium chloride solutions have been measured and reported by various authors in different hydrate regions. Table 1 gives a list of these data, reporting the temperature range and source of the experimental data. Available experimental data in the literature for CO 2 in NaCl aqueous solutions have mostly been measured at low pressures below the saturation lines. As seen in the table, data for high salinity are also limited.
In this study, experimental investigations were conducted to measure the hydrate stability zone of CO 2 in NaCl solutions over a wide range of conditions, in particular, to fill the gaps above the liquid locus and at high salinity. Moreover, experimental data on the hydrate equilibrium temperature of aqueous phases undersaturated in CO 2 with different CO 2 saturations and different aqueous phase compositions were carried out in order to validate and complement the available literature data (Tab. 2).
The Cubic-Plus-Association (CPA-EoS) Equation of State combined with the solid solution theory of van der Waals and Platteeuw [27] as developed by Parrish and Prausnitz [28] was employed to model the fluid and hydrate phase equilibria as previously described by Chapoy et al. [29][30][31][32]. The predictions of the thermodynamic model were compared with the newly experimentally measured properties.

Experimental section 2.1 Materials
The CO 2 used was supplied by Air Products and has a stated purity of 99.995%. The NaCl and CaCl 2 were both supplied by Sigma-Aldrich and have a stated purity of 99.5%. Distilled water was used in all tests. Details of the used fluids are given in Table 3.

Experimental apparatus
The tests were conducted using a fixed-volume, temperature-controlled rocking cell as described below. The experimental set-up is comprised of an equilibrium cell (Fig. 1), cryostat, rocking/pivot mechanism, and temperature/ pressure recording equipment controlled by a PC. The equilibrium cell is a 300 mL titanium cylindrical pressure vessel with a mixing ball, mounted on a horizontal pivot with an associated stand for pneumatically controlled rocking through 180°. The rocking of the cell, and the subsequent movement of the mixing ball within it, ensures adequate mixing of the cell fluids.
The rig has a working temperature range of 203-323 K, with a maximum operating pressure of 69 MPa. Circulating coolant from a cryostat within a jacket surrounding the cell controls the cell temperature. The cryostat can maintain the cell temperature stability to within better than 0.05 K. To achieve good temperature stability, the jacket is insulated with polystyrene board while connecting pipe work is covered with plastic foam.
The temperature is measured and monitored using a PRT (Platinum Resistance Thermometer) located within the cooling jacket of the cell. The cell temperature is measured with an uncertainty of 0.1 K. A strain gauge pressure transducer with an uncertainty of 0.03 MPa is used  Table 2. Literature data of CO 2 solubility in water and brines at hydrate-liquid water (HL w ) equilibrium. References Aya et al. [17] 276.15-282.75 30 1.96-3.14 -Yang et al. [18] 279-281 5-6.1 2.32-2.69 -Servio and Englezos [19] 274.05-282.95 2-6 1.56-2.81 -Lee et al. [20] 273.15-280. 15 5-20 0.7-2.6 -Someya et al. [21] 275.8-282.3 7-10 1.6-3 -Kim et al. [22] 278.9-280. 4  to monitor pressure. Temperature and Pressure are monitored and recorded by a PC through a RS 232 serial port. The cryostat can be monitored and controlled via an interface connected to a serial port on the computer. The temperature probe and pressure transducer are regularly calibrated using NAMAS certified temperature probe and a Budenberg dead weight tester, respectively.

Experimental methods
Hydrate dissociation point measurements for saturated systems were conducted using a reliable isochoric step-heating method as described in our previous papers [33,34]. For undersaturated systems mixtures of CO 2 and aqueous phase, composed of either distilled water or water with salts were made in a separate 600 mL piston vessel. In the first instance, a measured amount of CO 2 was injected into the clean evacuated vessel. The amount of CO 2 was measured using a balance with a precision of 0.01 g. The aqueous phase was then pumped into the vessel using a Quizix high-pressure pump. The volume of injected liquid was determined according to the density of the aqueous phase and the desired CO 2 concentration. The weight of the aqueous phase injected into the vessel was checked using the balance. This method gave an uncertainty of ±0.05 weight % for the mixtures used in this study. The pressure of the mixture was increased by injecting nitrogen into the vessel on the opposite side of the piston to the test fluid. The pressure was increased at least 14 MPa above the expected bubble point of the mixture and was mixed for at least 30 min to ensure it was a single phase. The equilibrated mixture was then transferred to the rocking cell. During the transfer, the pressure of the mixture in the 600 mL cell and the test fluid pressure was maintained at a pressure at least 14 MPa above the expected bubble point to ensure it was always a single phase.
Once the test fluids had been loaded into the test cell the pressure and temperature were adjusted in order to make a HET measurement. The pressure was adjusted by injecting or withdrawing equilibrated fluid into the vessel. The cell was then rocked to achieve equilibrium before lowering the temperature stepwise below the expected HET temperature. The presence of hydrates was indicated by a change in the P versus T trend as shown in Figure 2.      The temperature was then increased stepwise allowing sufficient time for equilibrium to be achieved at each step. The equilibrium P/T data were then plotted to determine the point at which hydrates are no longer present as indicated by a change in the P versus T plot. An example is shown in Figure 2 below.

Thermodynamic modeling
A general phase equilibrium model based on the uniformity of component fugacities in all phases is used to predict phase equilibria, water activity, and the hydrate forming conditions. A description of the thermodynamic model and parameters can be found elsewhere [29][30][31][32]. The model has been validated against carbon dioxide hydrate dissociation data [35] and gas solubility in brines at temperatures above the carbon dioxide critical temperature [36]. The model can also predict the gas solubility at low temperatures (T < 283.15 K) with an excellent agreement as seen in Figure 3.

Results and discussions
The experimental hydrate dissociations for CO 2 in equilibrium with 5, 10, 15, 20, 23, and 25 wt% sodium chloride solutions are reported in Table 4 and plotted in Figures 4  and 5. In the vapor region (Fig. 4) the model is in excellent agreement with our new experimental data and most of the available literature data, with typical agreement better than 0.5 K. For the 25 wt% solutions, the problem is more complex as when the system is cooled down the system forms NaClÁ2H 2 O before clathrate hydrates and the concentration in salt will therefore decrease in the remaining aqueous phase until it reaches the eutectic concentration 23.2 wt%, nevertheless taking into account the salting-out effect the model can still predict the dissociation conditions. In the liquid region (Fig. 5), errors between models and experimental results are slightly higher, however not greater than 1 K. Table 5 gives a summary of the measurements made in this work for undersaturated systems. As can be seen, in all cases increasing the pressure leads to an increase in the Table 5. Hydrate equilibrium temperatures at HL w as function of CO 2 concentration (x CO2 ) in distilled water and saline systems (CO 2 concentration is on free-salt basis and mole fraction). Expanded uncertainties (k = 2): u(T) = 0.7 K, u(P) = 0.13 MPa, u(x CO2 Â 10 2 ) = 0.04. See Appendix.   HET for the saturated system. All of the data for distilled water are summarized in Figures 6-8 showing the effect of CO 2 saturation on HET for the tests with distilled water. The deviations between the measured HET and model predictions for distilled water are on average lower than 0.5 K and a maximum of 1 K for the lowest CO 2 concentration. For salt solutions, the results are plotted in Figure 9 where larger deviations between experimental data and model predictions are observed. For some CO 2 concentrations, the measured HET increases with increasing salinity, while the model predicts the opposite behavior. There is a need for further experimental measurements and model development for aqueous systems undersaturated with CO 2 .

Conclusion
In this work, new experimental data have been reported for hydrates formed from carbon dioxide in the presence of aqueous solutions of sodium chloride over a wide range of temperatures, pressure, and concentration. The predictions of the developed model are compared against independent experimental data and the data generated in this work over a wide range of temperature, pressure, and NaCl concentration. A good agreement between predictions and experimental data is observed, demonstrating the reliability of the developed model. The hydrate stability pressuretemperature zone of dissolved CO 2 in the presence of salt show that a decrease in the system pressure and/or an increase in salt concentration favors hydrate formation, as both factors reduce equilibrium gas solubility in the aqueous phase. This behavior is unlike that of the system including a gas phase, where a higher gas pressure corresponds to an increased gas solubility which favors hydrate formation.
To further validate the model, Hydrate Equilibrium (melting) Temperature (HET) of aqueous phases undersaturated with CO 2 , at different CO 2 saturation levels and with different aqueous phase compositions were measured. The model is in relatively good agreement with the experimental results for distilled water, but larger deviations are observed for saline systems highlighting the need for further work to improve the model.

HET for undersaturated systems
In this study, uncertainties of the measured hydrate equilibrium temperature were estimated based on the "Evaluation of Measurement Data -Guide to the Expression of Uncertainty in Measurement (GUM)" [37]. For HET measurements, considering the impact of uncertainties associated with calibration parameters, measured temperature, pressure, and sample preparation, the combined standard uncertainty of the measured density is calculated by: uðT HET Þ 2 ¼ uðT Þ 2 calibration þ uðT Þ 2 measurements : ðA:3Þ For the calibration uncertainties, the temperature probe was calibrated using a NAMAS-certified temperature probe. For the uncertainties of HET measurements, equation (A.4) was used to determine the standard uncertainties: oT ox was calculated using the thermodynamic model.
Considering the systematic uncertainties of measurement devices, the calibration procedures, and the reproducibility of the results, the standard uncertainties of the measured HET were found to be 0.74 K.

HET for saturated systems
For the uncertainties of HET measurements in presence of NaCl, the procedure described by Stringari et al. [38] was used, and the standard uncertainties of the measured HET were found to be 0.4 K.