CO2 Storage in Deep Saline Aquifers

C H A P T E R 10

CO2 Storage in Deep Saline Aquifers Xiaoyan Ji1, Chen Zhu2 1 2

Division of Energy Science/Energy Engineering, Lulea University of Technology, Lulea, Sweden; Department of Geological Sciences, Indiana University, Bloomington, IN, USA

1. Introduction Global warming is now widely recognized as one of the biggest global issues facing mankind, and anthropogenic CO2 generated mainly from the combustion of fossil fuels for power generation is believed the main cause of global warming. To mitigate CO2 emissions from fossil-fueled power plants, CO2 capture from flue gases is an important option. The captured CO2 is further transported into storage sites in which the deep saline aquifers appear to hold the largest potential capacity. The injection of large-scale CO2 into deep saline aquifers will induce multiphase fluid flow, solute transport, and chemical reactions between fluids and formation minerals. The estimation of the storage capacity, the understanding of the storage mechanism, and the prediction of the fate of the injected CO2, as well as their environmental impacts call for the study of the properties of CO2-storage-related systems, such as the density, viscosity, surface tension, and gas solubility, as well as the effects of impurities in CO2 on these properties and storage processes. Due to the high pressure in transportation and storage sites, the existence of multicomponents, as well as the duration of long time scale for storage-related processes, experimental measurements are generally difficult and time consuming. It is thus desirable to develop a predictive model to represent the properties of storage-related systems based on the easily assessable experimental data. The developed model can be further incorporated into a process model to simulate the storage process and then predict the fate of the injected CO2 and the environmental effects. The general information of CO2 capture and storage has been described in detail in the IPCC report.1 To make it easy to understand the significance of the modeling work and follow the content of this chapter, a brief introduction of CO2 storage is described in this section including CO2 emission, CO2 storage and its mechanism, safety and risk, as well as the importance of the model development in the CO2-storage area. Novel Materials for Carbon Dioxide Mitigation Technology. http://dx.doi.org/10.1016/B978-0-444-63259-3.00010-0 Copyright © 2015 Elsevier B.V. All rights reserved.

299

300 Chapter 10

1.1 CO2 Emissions and Storage 1.1.1 CO2 Emissions Global warming is now widely recognized as one of the biggest global problems facing humanity. It has been stated that the global average temperature has increased 0.74
0.18  C during the last 100 years ending in 2005, and the rate of warming over the last 50 years is almost double of that over the last 100 years.1 According to global climate scenarios, the temperature may rise even further, in the range of 1.5e6.4  C during the twenty-first century.1 The increasing global temperature will lead to severe consequences, such as prolonged droughts, crop failure, changes in cropping patterns, increased desertification, and polar ice melting that will result in ocean flooding and submergence of major portions of low-lying islands and coastal areas.1 Due to the studies of the past five decades, particularly the past 15 years, it is believed that the increased greenhouse gas concentrations in the atmosphere are the main cause of global warming. The most significant anthropogenic greenhouse gas is CO2, which arises mainly from the use of fossil fuels and counts for around 80% of global emissions.1 To avoid potential climate-related disasters, an effective framework of mitigating CO2 emissions urgently needs to be established. Several methods have been suggested for mitigating CO2 emissions; for example: (1) reducing fossil-fuel consumption by improving energy conversion efficiency and enhancing less-energy-intensive economic activities; (2) switching to less-carbonintensive fuels for which suitable supplies of natural gas are available; and (3) using renewable energy sources (nuclear, wind, solar, biomass, etc.).1 However, the options of reducing fossil-fuel consumption and switching to less-carbon-intensive fuels are not enough to mitigate global warming in the near future. Renewable energy sources have a high establishment cost, are location-dependent, and their price has not yet been sufficiently competitive. Subsequently, it is infeasible to dramatically cut the utilization of fossil fuels if the standard of living is to be maintained. Therefore, CO2 capture and storage (CCS) is needed as a short-term solution to mitigate environment impacts and allow humans to continue using fossil-fuel energy until renewable energy technologies are ready for application. Following the recommendations of the IPCC and the EU directive2 for the abatement of climate change, 9.4 Gt/year of CO2 need to be captured and safely transported and stored underground by 2050 according to the Blue Map Scenario.3 1.1.2 CO2 Storage CCS provides an effective way to prevent CO2 emissions into the atmosphere by capturing CO2 from major stationary sources, transporting the captured CO2, usually by pipeline, and injecting it into suitable areas for storage. These areas can be underground deep rock

CO2 Storage in Deep Saline Aquifers 301 formations (geological storage), the deep sea (ocean storage), or minerals (mineral storage). Geological storage can be undertaken in a variety of geological settings: in sedimentary basins, oil fields, depleted gas fields, deep coal seams, and saline formations, all possible geological storage formations. Injection of CO2 into depleted oil and gas reservoirs is a particularly reasonable approach to storing CO2 because the infrastructure is largely already in place. The estimated storage capacity is around 900 Gt CO2,1 which may not be sufficient to meet the long-term needs, but is sufficient for the present time. Abandoned or uneconomic coal seams could also be used as CO2-storage sites. In addition, the injection of CO2 into unminable methane-rich coal deposits could enhance the production of methane, which can be used for energy (termed as enhanced coal-bed methane). This method is promising due to the combination of CO2 emission mitigation with a valuable energy resource production. It was estimated that the worldwide storage potential of coal seams is up to 200 Gt CO2.1 Deep saline aquifers have the largest potential storage capacity among all the geological CO2-storage options, although they provide noneconomic return for CO2 injection. Deep saline aquifers are widespread, and potentially have CO2-storage capacities sufficient for holding many decades worth of CO2 emissions with a global capacity varying from 1000 to 10,000 Gt CO2,1 which makes it crucial to study the possibility of CO2 storage in deep saline aquifers. 1.1.2.1 CO2-Storage Mechanism

The CO2-storage mechanism directly relates to the storage sites. In this section, the CO2-storage mechanism in deep saline aquifers is briefly summarized on the basis of the IPCC report.1 Deep saline formations are porous rock formations that are typically several kilometers below the surface and contain enormous quantities of unusable water with high salt and/or mineral content. This saltwater brine is around 10 times saltier than the oceans and has been trapped by impermeable rock, called a “caprock,” for millions of years. The brine (saltwater) is called formation fluid. Once CO2 is injected into a deep saline aquifer, as CO2 is less dense than the formation fluid, the supercritical CO2 rises buoyantly as a separated phase through the porous rocks until it reaches the top of the formation where it meets and is trapped by an impermeable layer of caprock. This is called structural/stratigraphic trapping. Meanwhile, the injected supercritical CO2 displaces fluid as it moves through the porous rock. As the CO2 continues to move, brine replaces it again, but some of the CO2 will be left behind as disconnected (residual) droplets in the pore spaces and are immobile. This is called

302 Chapter 10 residual trapping. When CO2 encounters the brine present in the porous rock, CO2 will dissolve into the brine. As the brine containing dissolved CO2 is denser than the surrounding fluids (i.e., brine without dissolved CO2), it will sink to the bottom of the rock formation over time. This process is defined as solubility trapping, and it will trap the CO2 even more securely with less risk of leakage. When CO2 dissolves in water, it will form a weak carbonic acid. Over a long period, this weak acid can react with the minerals in the surrounding rock to form solid carbonate minerals. This process is called mineral trapping. Depending on the mineralogy of the rock and water in a specific storage site, the mineral trapping process can be rapid or very slow, but it effectively binds CO2 to the rock.1 These trapping processes take place at different rates ranging from days to years to thousands of years. In general, geologically stored CO2 becomes more securely trapped with time due to the higher density of fluid that contains dissolved CO2 as well as the fixation of CO2 through its reaction with minerals. CO2 dissolution occurs over hundreds to thousands of years, whereas the conversion of the injected CO2 to solid carbonate minerals takes over millions of years.

1.2 CO2-Storage Safety and Risk Escape of the buoyant CO2 to the surface is a significant concern because (1) escape of the buoyant CO2 results in the failure to permanently retain CO2 in the storage aquifer, and (2) buoyant CO2 may contaminate shallow potable aquifers if released from the storage formation. Safe and long-term storage requires that the CO2 is immobilized or prevented from moving upward. For deep saline formations, the closed structures will immobilize the buoyancy-driven flow (supercritical CO2) in oil and gas reservoirs. In fact, they have retained oil, natural gas, and naturally occurring CO2 in the subsurface for millions of years, which demonstrates the low leakage risk. In addition, based on the CO2storage mechanism, the leakage risk is expected to decrease over time due to the dissolution of CO2 into saltwater and its sinking to the bottom of the geologic formation, as well as the further reaction of the dissolved CO2 with the minerals in the surrounding rock to form solid carbonate minerals. A number of pilot and commercial projects of CCS are underway or proposed and have been summarized in IPCC reports.1 However, the numbers are far away from those that are needed to significantly reduce atmospheric CO2 emissions. The feasibility of CCS depends on a number of factors such as cost, storage capacity, safety, and risk, as well as environmental effects. These all relate to the properties of CO2 and deep saline formations, the physical-chemical process involved, and the impurities in the CO2.

CO2 Storage in Deep Saline Aquifers 303

1.3 Thermodynamic Model Development and CO2 Storage Storage of CO2 in deep saline aquifers is one of the promising approaches for the reduction of greenhouse gases. Large-scale injection of CO2 into saline aquifers will induce a variety of coupled physical and chemical processes including multiphase fluid flow, solute transport, and chemical reactions between fluids and formation minerals. To predict the sequestration capacity, understand the consequences of gas injection (the fate of the injected gases), and satisfy the requirements for the geochemical applications, it is critical to study the phase equilibrium and other thermodynamic properties for CO2ebrineerock related-systems at temperatures and pressures of interest. In general, for CO2 deep saline storage, the temperature can be up to 200  C and the pressure can be up to 600 bars. Experimental measurement is a direct method for studying the related phase equilibrium, properties, and processes. However, the demanding conditions (high temperature, pressure, and multicomponent systems) make the experimental measurements very difficult and time consuming. Furthermore, the long time scale of the CO2-storage process, which can take up to hundreds or thousands of years, makes it impossible to run experiments. For such a process, it is very crucial to develop a prediction model to simulate the process. The prerequisite of a reliable process simulation is a reliable representation of phase equilibrium and properties. Due to the extreme conditions, again, it would be very advantageous to develop a predictive model to represent phase equilibrium and properties in a wide temperature and pressure range based on the experimental data measured at mild conditions. This is a challenging task, especially for such a system with the existence of ions. In addition, the captured CO2 streams may contain impurities. The types and concentrations of impurities vary significantly depending on the fuel type and capture process, and a CO2 stream may represent mixtures of CO2 from several capture plants. For example, streams from power plants depend on the type of feedstock used (coal, liquid, and gas fuel), the process by which the fuel is converted into energy, the capture process, and postprocessing. CO2 streams from important industrial processes (cement, steel, ammonia and hydrogen production, and gas processing) vary in both impurity concentration and types of impurities. Impurities, if permissible in CO2 streams, can lead to great potential for capital and energy cost savings for CO2 capture and separation, but they also increase costs for transportation and risk management in transport and storage. Although the benefits of permitting some impurities in CO2 streams have been a subject of interest to industries and governmental regulatory bodies, and dozens of injection projects are already in operation, notably in Canada,4,5 the scientific and technical knowledge base needed to evaluate the impurities on CCS is still lacking. For example, the research work on the influence of impurities on the phase equilibrium and thermodynamic properties such as density is still very limited.

304 Chapter 10

2. Modeling of Properties and Phase Equilibrium It is crucial to develop a predictive model to represent phase equilibrium and other thermodynamic properties for CO2ebrineerock systems at temperatures and pressures of interest as well as the effect of impurities on their phase equilibrium and properties for CO2 storage. The brine can contain ions of Naþ, Kþ, Ca2þ, Mg2þ, Cl, and SO4 2 in which Naþ and Cl are usually the main ionic species. The impurities can be N2, Ar, O2, H2, SOx, NOx, CO, H2S, and CH4. Among these impurities, H2S is one of the most common components in natural gas and products from oil processing and production.4 Because surface desulfurization and surface storage is not economical and carries significant liability, it was proposed to co-inject CO2eH2S into depleted oil and gas reservoirs and deep saline formations.6 The first co-injection operation started in 1989, and according to the most recent count, 48 injection sites across the Alberta Basin and British Columbia are operating.4,5,7 Therefore, H2S will be the main focus in impurities in this section. Experimental data play a significant role in model development. The model parameters can be obtained from the fitting of experimental data, and the model prediction can be verified by the available experimental data. For CO2-storage-related systems, experimental data are available for the ternary systems of CO2eH2OeNaCl (see review in Ref. 8 and new measurements in Refs. 9e11) and H2SeH2OeNaCl (see review in Ref. 12 and CO2eN2eH2O13). Recently, experimental data were also measured for N2eCO2eH2OeNaCleKCleCaCl2.13 Although the co-injection of CO2 and H2S has been carried out in operation, to the best of our knowledge, no experimental data for the quaternary CO2eH2SeH2OeNaCl system are publicly available. In addition, experiments for systems containing H2S are difficult, mainly because they are expensive and time consuming due to the corrosiveness of H2S. In theory, two approaches can be used in modeling phase equilibrium, the g-f approach or the f-f approach. In the g-f approach, an equation of state (EOS) is used to describe the nonideality in the light/vapor phase, and Henry’s law or an activity model is used to describe the nonideality in the liquid phase. The inherent disadvantage of this approach is that it does not allow for estimating the density of the liquid phase. This approach has been used to develop models for the CO2eH2O/CO2eH2OeNaCl,14e19 H2SeH2O/H2SeH2Oesalt,12,15 CH4eCO2eH2Sebrine,20 as well as H2OeCO2eNaCleCaCO3eCaSO4.21 In the f-f approach, an EOS is used for both phases, and the density and other thermodynamic properties can be estimated for both phases. As density is one of the more important properties for CO2 storage, f-f is preferable. The cubic EOS is the most commonly used EOS with the ideal gas as a reference. Due to this, this series of EOSs is preferable for representing the properties for the gas phase or for the fluid at low pressure

CO2 Storage in Deep Saline Aquifers 305 when the state is not so far away from an ideal gas. For example, Perfetti and coworkers22,23 and Li and Firoozabadi24 used cubic plus association EOS for the CO2eH2O and H2SeH2O systems; Pappa et al.25 used the PengeRobinson EOS for the CO2eH2O system. More than 20 years ago, a new EOSdstatistical associating fluid theory (SAFT) was proposed. It has been successfully used to describe the thermodynamic properties and phase behavior of fluids including associating systems, electrolytes, and polymers up to high pressures for which the popular cubic EOS fails to provide an adequate description.26 The key of this model is its firm statistical-mechanics basis, which allows for a rigorous physical interpretation of the contributions due to the various intermolecular interactions, such as hard-sphere, dispersion, chain, association, polar, and ionic interactions.26 This provides a framework from which the effects on the thermodynamic properties of the various molecular features can be separated and quantified.26 This is very important for CO2-storage-related systems because CO2 is a molecule with quadrupolar interactions, the impurities of SO2 and H2S are polar molecules, H2O is a molecule with a hydrogen bond, and ions are charged species. SAFT-based models have been developed to represent phase equilibrium and/or thermodynamic properties for CO2-storage-related systems. Sun and Dubessy27,28 used SAFT-LennardeJones (LJ) for CO2eH2O and CO2eH2OeNaCl; dos Ramos and McCabe29 used the SAFT-variable range (VR)-D for H2SeH2O; Tang and Gross30 used perturbed chain polar (PCP)-SAFT for the H2SeCO2 system; and Tan et al.31 used perturbed chain (PC)-SAFT/primitive mean spherical approximation (PMSA) for SO2eCO2ebrine. In addition, a systematic study has been carried out based on SAFT1-restrictive primitive model (RPM)/SAFT2 to describe the phase equilibria and density for H2SeH2O,32 CO2eH2OeNaCl,33H2SeCO2eH2OeNaCl,34,35 as well as the properties of the aqueous electrolyte solutions with the ions of Naþ, Kþ, Ca2þ, Mg2þ, Cl, and SO4 2 .36e42 For the first time, SAFT2 was developed to represent the properties of brine at high pressures41,42 and was extended to H2SeCO2eH2OeNaCl.34,35 Therefore, this model is one of the focuses of the chapter.

2.1 Statistical Associating Fluid Theory (SAFT)2 In SAFT2,33,38 square-well (SW) fluid is used as a reference. The SW fluid has a steep repulsion and a short-range attraction. Three parameters, the hard-corn diameter (s), the well depth (u), and the reduced range of the potential well (l), were used to characterize the SW potential model, and the intermolecular potential f is defined as: 8 > : 0 r>s

306 Chapter 10 In SAFT2, the model accounts for molecular interactions in terms of Helmholtz free energy, and the dimensionless residual Helmholtz energy is defined as: a~res ¼ a~seg þ a~assoc þ a~chain þ a~ion

(2)

in which the superscripts refer to terms accounting for the residual, segment, association, chain, and ionic interactions, respectively. 2.1.1 Hard-Sphere Term a ~hs 43e45 The hard-sphere contribution a~hs in heterosegmented SAFT is given by " ! # 3 2 3 6 ðz Þ þ 3z z z  3z z ðz Þ ðz Þ 2 1 2 3 1 2 3 a~hs ¼  z0  2 2 lnð1  z3 Þ pNAv rm z3 ð1  z3 Þ2 ðz3 Þ

(3)

in which NAv is the Avogadro number, rm is the molar density, and zk ¼

X X p NAv rm Xi mi xa ðsa Þk 6 a i

ðk ¼ 0; 1; 2; 3Þ

(4)

in which Xi is the mol fraction of component i, mi is the number of segments of component i, sa is the diameter of segment a, and xa is the segment fraction defined as xa ¼

number of moles of segments a number of moles of all segements

(5)

2.1.1.1 Dispersion Term a~disp 38,43e45

The dispersion term is calculated from a~disp ¼

X i

"

#

1 disp 1 Xi mi adisp þ a~t a1 þ 2 2 kB T ðkB TÞ

in which kB is the Boltzmann constant, T is the temperature in Kelvin, and XX ¼ xa xb adisp adisp 1 1;ab a

(6)

(7)

b

adisp 1;ab

is the first-order binary term for the in which a and b are the segment types, and aeb segment interaction given by p      3 3 hs adisp s ¼ 4 N r  1 g ; 2 s u l Av ab ab 3;eff m ab ab ab 1;ab 6 In Eqn (8), sab is the distance between centers of segment a and b at contact, uab is the well depth of square-well potential for the aeb interaction, and lab is the reduced

(8)

CO2 Storage in Deep Saline Aquifers 307 range of the potential well for the aeb interaction. The combining rules used for sab and uab are,  1 sa þ sb 2  pffiffiffiffiffiffiffiffiffiffi ¼ ua ub 1  kab

sab ¼ uab ¼ uba

(9) (10)

in which ua is the segment energy of segment a, and kab is the binary interaction parameter, which can be temperature-dependent or temperature-independent. A simple arithmetic-mean combining rule is used for lab, analogous to that for the segment diameters:  1 la þ lb 2 in which la is the reduced range of the potential well of segment a. lab ¼

(11)

The radial distribution function for a mixture of hard spheres in Eqn (8) is calculated using CarnahaneStarling’s equation but evaluated at the effective reduced variable zkeff, 2      z2;eff 3sa sb sa sb 2 z2;eff 1 hs gab sab ; z3;eff ¼ þ (12)   þ2  3 1  z3;eff sa þ sb 1  z3;eff 2 sa þ sb 1  z3;eff In the range of 1.0 < lab  2.5, the effective reduced variable z3,eff is approximated from    

(13) z3;eff lab ; z3 ¼ z3 1 þ d lab ; z3 in which 6 X 6 X     z3i1  1 ð1  0:59z Þ c l d lab ; z3 ¼ l5:397 ij ab 3 ab j1 lab i¼1 j¼1

(14)

and cij’s are universal constants listed in Table 1. In Eqn (12), the effective reduced variable z2,eff is calculated from z2;eff ¼

z2 z z3 3;eff

in Eqn (6) has the same form as the term adisp The term adisp 2 1 , X X adisp ¼ xa xb adisp 2 2;ab a

b

(15)

(16)

308 Chapter 10 Table 1: Universal coefficients cij in Eqn (14)38 i\ j

1

2

1 2 3 4 5 6

0.010 348 412 43 0.034 371 512 70 0.046 688 853 54 0.286 193 974 80 0.679 398 501 40 1.380 935 033

5.012 940 585 45.703 391 47 161.837 063 5 276.070 506 3 224.732 718 6 69.715 059 55

i\j

4

5

1 2 3 4 5 6

271.226 970 3 2,296.265 301 7,368.693 459 10,912.042 77 7,025.297 671 1,282.236 283

645.515 037 9 5,430.859 895 17,288.510 83 25,296.495 02 15,911.317 81 2,690.356 186

disp in which adisp 2;ab is related to a1;ab as follows: disp

adisp 2;ab

va1;ab 1 z0 ð1  z3 Þ4 ¼ uab rm 2 vrm z0 ð1  z3 Þ2 þ 6z1 z2 ð1  z3 Þ þ 9z32

3 46.069 085 85 391.844 391 2 1,267.473 745 1,903.020 452 1,256.163 441 242.462 559 3 6 605.117 779 9 5,087.008 598 16,217.305 15 23,885.063 01 15,336.690 52 2,828.288 422

! (17)

The term a~t in Eqn (6) is calculated from a~t ¼

5 X 2 X

 Dmn

m¼2 n¼1

u kB T

m  n z3 s

(18)

in which Dmn’s are universal constants listed in Table 2, s is the close-packed reduced density (¼ 21/2 p/6), and u/kT is evaluated in the spirit of the van der Waals one fluid theory   PP u xa xb kBabT vab a b u ¼ PP (19) kT xa xb vab a

b

in which  1=3 #3 ðva Þ1=3 þ vb ¼ 2 "

vab

(20)

The molar volume of segment a (va) is related to the segment diameter as follows: va ¼

p 3 s NAv 6s a

(21)

CO2 Storage in Deep Saline Aquifers 309 Table 2: Universal coefficients Dmn in Eqn (18)38 n\m

2

3

4

1 2

2.420747 9.955897

4.151326 1.520369

2.501130 0

5 0.462574 0

2.1.1.2 Chain Term a~chain 44,45

The chain term is calcsulated by h X    i SW a~chain ¼  Xi ðmi  1Þ ln gSW g s  ln ab i 0;i sab

(22)

i

and   X   Bab;i lngSW sab ¼ ln gSW i ab sab

(23)

ba

in which gSW ab ðsab Þ is the square-well radial distribution function calculated at contact, and SW SW g0 is g evaluated at zero density. The pair radial distribution function for a mixture of square-well segments is determined as follows:       hs gSW (24) ab sab ¼ gab sab þ buab g1;ab sab in which ghs ab ðsab Þ is the pair radial distribution function for a mixture of hard spheres given by    3sa sb sa sb 2 ðz2 Þ2 1 z2 hs þ þ2 (25) gab sab ¼ 1  z3 sa þ sb ð1  z3 Þ2 sa þ sb ð1  z3 Þ3 and g1,ab(sab) is the perturbation term,   1 g1;ab sab ¼ 4uab

"

vadisp va1;ab lab 1;ab  P   P ps3 N p vr vl 3 X m ab i i 3 6 sab NAv r Xi mi 6 ab Av 1

i

# (26)

i

In Eqn (23), Bab,i is the bond fraction of type ab in molecule of component i, and it is equal to one because each component studied in this work is a homosegmented molecule. 2.1.1.3 Association Term a~assoc 46,47

The association term is calculated by Adidharma and Radosz47 # " Ai  X X X nðG Þ i Xi ln X Ai  a~assoc ¼ þ 2 2 i A ˛G i

i

(27)

310 Chapter 10 in which n(Gi) is the number of association sites on molecule i and XAi is the mole fraction of molecules i not bonded at site Ai given by X Ai ¼ 1 þ rn

P

1  P  B Ai Bj  Xj X jD

j

(28)

Bj ˛ Gj

in which DAi Bj is the association strength between site Ai at molecule i and site Bj at molecule j given by    DAi Bj ¼ F Ai Bj ebaðrÞuij ghs sij s3ij kAi Bj

(29)

in which  

F Ai Bj ¼ exp ε Ai Bj kB T  1

(30)

aðrÞ ¼ 1 þ 0:1044r  2:8469ðr Þ2 þ 2:3787ðr Þ3

(31)

and

in which r* is the reduced density related to the segment density rs by r ¼

6 XX xa xb s3ab 20 p a b

(32)

The Lorentz and Berthelot combining rules are used for the size and energy parameters, that is,  pffiffiffi 13 2 oo v (33) si ¼ NAv 2 1=3 33   3 kAi Bi 1=3 þ s3 kAj Bj s j 6 i 7 (34) s3ij kAi Bj ¼ 4 5 2 ε Ai Bj ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi εAi Bi εAj Bj

(35)

2.1.1.4 Ionic Term a~ion

The restricted primitive model is used in SAFT2 to account for long-range Coulombic interactions48 a~ion ¼ 

3X 2 þ 6X þ 2  2ð1 þ 2XÞ3=2 12prm NAv d 3

(36)

CO2 Storage in Deep Saline Aquifers 311 in which d is the effective (hydrated) diameter defined by X x0i di d¼

(37)

i

in which x0i is the mole fraction of ion i on a solvent-free basis and the summation is over all ions. In Eqn (36), X is the dimensionless quantity defined by X ¼ kd

(38)

in which k is the Debye inverse screening length given by k2 ¼

4p X 4pe2 X rn; j q2j ¼ r z2 εw kT j εw kB T j n; j j

(39)

in which εw is the dielectric constant of water, rn, j is the number density of ion j, qj is the charge of ion j (¼ zje), zj is the valence of the ion j, e is the charge of an electron, and the summation is over all ions in the mixture. The density at a certain temperature and pressure can be iterated based on Eqn (2), and other properties can be derived from the residual Helmholtz energy. To calculate phase equilibrium, the fugacity coefficient of component i is calculated by !  res  res X v~ a v~ a  xj þ Z  1  ln Z lnb 4 i ¼ a~res þ vxi T;r;xj6¼i vxj j

(40)

T;r;xk6¼j

in which Z is the compressibility factor and calculated with  res  v~ a Z ¼1þr vr T;x

(41)

3. Model Results and Discussion In modeling, each component is modeled as one kind of segment with parameters of segment number m, segment volume voo, segment energy u/K, and the reduced range of the potential well l. For molecules with association interactions, two additional parameters are used, that is, the well depth of the association siteesite potential ε, and the parameter related to the volume available for bonding k. For ions, one additional parameter exists, effective diameter d.

312 Chapter 10 Table 3: Parameters of CO2, H2S, and H2O

CO249 H2S32 H2O38

M

n00, cc/mol

u0/k, K

l

ε/k, K

1.3513 1.2882 1.0

11.137 11.7602 9.8307

219.992 254.318 311.959

1.4220 1.5240 1.5369

217.7834 135.62 1481.41

k 0.18817 0.004382 0.04682

3.1 Pure Components In CO2-storage-related systems, the components of H2S, H2O, and CO2 have been investigated,38,49,50 in which H2S was modeled as a molecule with four association sites, two sites of type S and two sites of type H.50 H2O was also modeled as a molecule with four association sites, two sites of type O and two sites of type H.38 CO2 was modeled as a molecule with three association sites, two sites of type O and one site of type C.49 Sites of the same type do not associate with each other. The parameters of each pure component were obtained from the fitting of the experimental data of saturation pressure and saturation liquid densities. The fitted parameters of H2S, H2O, and CO2 are summarized in Table 3. To further illustrate the model performance, the model results of vapor pressure and liquid density of CO2 and H2S were compared with the experimental data and shown in Figures 1 and 2, respectively, which implies that the model can reliably represent both vapor pressure and density.

Figure 1 (a) Vapor pressure (P) and (b) saturated liquid density (r) of CO2. n, experimental data51; d, calculated results from the SAFT2 model.

CO2 Storage in Deep Saline Aquifers 313

Figure 2 (a) Vapor pressure (P) and (b) saturated liquid density (r) of H2S. A, Experimental data by Lemmon et al.52; d, calculated results from the SAFT2 model.

3.2 Aqueous Electrolyte Solutions In the field of geological carbon storage, it is common that the brine is simplified as, or refers to, aqueous NaCl solution because NaCl is usually the main component. This is feasible when the gas solubility and aqueous solution density are the main research focus. However, when the dissolved CO2 reacts with rock, it is possible that an existent rock is dissolved and new minerals precipitate. In this case, it is also necessary to know the properties of other ions, such as their activities (¼ the product of concentration and activity coefficient). This is because the pressure is high, which makes the activity considerably different from concentration. Thus, to cover the whole interests in geochemistry, it is necessary to study the properties of “true” brine. Generally, the components in brines include Naþ, Kþ, Ca2þ, Mg2þ, Cl, and SO2 4 . Based þ on SAFT2, the properties of aqueous solutions with electrolytes including Li , Naþ, Kþ, 2 2 Ca2þ, Mg2þ, Cl, Br, HCO 3 , SO4 , and CO3 were studied at temperatures from 298.15 to 473.15 K and at pressures up to 1000 bars with 0e6 mol/kgH2O electrolyte in ionic strength. The work was described in detail in Ref. 41 and only a brief description is summarized in the following text. In modeling, each ion was modeled as charged, but nonassociating spherical segments. To account for the effect of pressure on the properties of electrolyte solutions, a shortrange interaction between cation and anion was allowed, but the cationecation and

314 Chapter 10 anioneanion short-range interactions were neglected, that is, kab ¼ 1 in Eqn (10). The short-range interaction between ion and water was fully considered, that is, kab ¼ 0 in Eqn (10). kab for cationeanion, short-range interaction is expected to be in the range of zero to unity. For practical purposes, this cationeanion short-range interaction was set to be 0.5, that is, kab ¼ 0.5 in Eqn (10), to reduce the number of parameters. Moreover, the parameters of n, u, l, and d for each ion were temperature-dependent with the following equations: h i v ¼ v25 1 þ b1 ðT  298:15Þ þ b2 ðT  298:15Þ2 (42) h i (43) u ¼ u25 1 þ b3 ðT  298:15Þ þ b4 ðT  298:15Þ2 h i l ¼ l25 1 þ b5 ðT  298:15Þ þ b6 ðT  298:15Þ2 (44) h i d ¼ d25 1 þ b7 ðT  298:15Þ þ b8 ðT  298:15Þ2 (45) in which b1 to b8 are the ion-specific coefficients. In parameter fitting, first, one set of parameters for the group of ions (Liþ, Naþ, Kþ, Ca2þ, Mg2þ, Cl, Br, I, NO3  ; HCO3  ; SO4 2 ; CO3 2 ) at 298.15 K was fitted to a group of experimental data and liquid density (see reference in Ref. 41). Second, the coefficients for the temperature-dependent parameters were fitted to the experimental activity coefficients and liquid densities at temperatures up to 473.15 K and low pressures. The parameters for Naþ, Kþ, Ca2þ, Mg2þ, Cl, and SO2 4 are listed in Tables 4 and 5, in which l25,iw is the arithmetic mean of the reduced well range for watereion interactions (l25,iw ¼ 0.5(l25,w þ l25,i), in which w refers to water and i refers to cation or anion). Using SAFT2 with fitted parameters, the density and osmotic coefficients of aqueous electrolyte solutions can be predicted. Figure 3(a) illustrates the density of NaCleH2O predicted with the model up to 1000 bars and the comparison of the well-known Pitzer model53 represented with symbols of temperatures at 298.15, 373.15, and 473.15 K. The comparison shows that the model prediction agrees well with the model results of Pitzer.53 As all the main components (ions) in brine have been covered in the study, the properties  Table 4: Parameters for Kþ, Ca2þ, Mg2þ, Cl, and SO2 4 at 298.15 K (25 C)

I þ

Na Kþ Ca2þ Mg2þ Cl SO2 4

v25, cc

(u25/k), K

l25,i-w

˚ d25, A

1.6961 3.9721 4.8193 6.8018 7.6771 4.3173

3321.67 390.008 5148.90 3645.86 1228.19 2219.72

1.8860 1.8126 1.9707 2.0071 1.0924 1.0185

4.6806 3.3779 5.3193 4.8892 5.6428 2.8347

CO2 Storage in Deep Saline Aquifers 315 Table 5: Coefficients in Eqns (42e45) for Kþ, Ca2þ, Mg2þ, Cl, and SO2 4

þ

Na Kþ Ca2þ Mg2þ Cl SO2 4

103 b1

105 b2

103 b3

105 b4

104 b5

106 b6

104 b7

106 b8

K1

K2

K1

K2

K1

K2

K1

K2

3.8224 7.0994 2.9048 1.0726 1.5282 2.8561

0.0022 1.2548 0.0497 0.4615 0.1448 e

4.8026 1.5538 2.8213 60.890 4.7972 5.5736

2.5607 0.4560 0.7746 6.9648 2.1132 3.6565

7.7334 28.076 8.9577 9.7640 0.9565 e

3.0528 7.5747 0.9250 13.587 9.7282 e

0.6083 14.922 25.313 50.801 0.0827 25.377

5.0683 4.0720 6.8974 20.117 5.0773 3.5871

of brine can be predicted. The density of brine at different temperatures was predicted with the typical composition taken from literature54 and compared with available experimental data from 298.15 to 423.15 K. The model prediction of brine agrees well with experimental data55 as shown in Figure 3(b), in which the chlorinity is the mass of chlorine in 1 kg solution. The detailed evaluation of the model preformation was described in our previous work.41 As mentioned above, the properties of brine are often assumed those for aqueous NaCl solution. To study the error caused by this assumption, the properties for both brine with

Figure 3 Density for (a) aqueous NaCl solution at 298.15 K (circles), 373.15 K (diamonds), and 473.15 K (squares); and (b) brine/seawater at 298.15 K (circles), 373.15 K (diamonds), and 423.15 K (squares). Symbols: (a) model results of Pitzer, (b) experimental data55; Curves: model prediction.

316 Chapter 10

Figure 4 Properties for aqueous NaCl solutions (dashed line) and brine/seawater (solid line) at 298.15 K. Symbols: experimental data56,57; Curves: model prediction.

typical compositions and aqueous NaCl solutions were predicted and then compared. The results are depicted in Figure 4. As shown in Figure 4(a), the activity of water in aqueous NaCl solution is similar to that of brine at a different salinity (¼ total mass of salts in 1 kg solution), which proves that the simplification of brine to aqueous NaCl solution is reasonable when phase equilibrium is the research focus. The density of brine is slightly higher than that for aqueous NaCl solution at high chlorinity as shown in Figure 4(b), and the largest error is up to 3.5% under the conditions studied, which is still acceptable. Therefore, the comparison confirms that it is reasonable to assume brine to be aqueous NaCl solutions. The activity coefficient is used to adjust concentration when the nonideal behavior can-not be neglected. When the activity coefficient is almost a constant in the concentration range of interest, the adjustment can be merged with other parameters that are independent of concentration. In general, compared to NaCl, the concentration of Ca (CaCl2) or Mg (MgCl2) in brine is low, but it is still unclear whether the activity coefficient is a constant or not. Using the developed model, the mean activity coefficients of both CaCl2 and MgCl2 in brine were predicted and illustrated in Figure 5(a). For comparison, the mean activity coefficients of NaCl in brine with the typical composition taken from literature54 are also illustrated in Figure 5(a). When the ratio of Ca2þ/Mg2þ to Naþ is fixed, with increasing salinity, the mean activity coefficients of both CaCl2 and MgCl2 decrease significantly and then increase. For further discussion, the mean activity coefficients of

CO2 Storage in Deep Saline Aquifers 317

Figure 5 Mean activity coefficients for (a) NaCl (circles), CaCl2 (triangle), MgCl2 (diamond) in brine with typical compositions taken from literature54 at 298.15 K and (b) CaCl2 (circles) and MgCl2 (triangle) in H2O at 400 bars and 323.15 K, symbols: experimental data: circles,58; triangles,59; curves: model prediction.

CaCl2 and MgCl2 in water were predicted at 400 bars and 323.15 K, respectively, and the prediction was compared with experimental data58,59 in Figure 5(b). Again, the activity coefficients vary from 1 to 0.45 when the concentration changes from 0 to 0.5 mol/kgH2O. All of these observations imply that it is necessary to study the properties of other main components besides NaCl in water.

3.3 CO2eBrine (H2OeNaCl) As NaCl is the main component of brine, the phase equilibrium and properties of CO2eH2OeNaCl play an important role in CO2 geological storage. Models based on the equation of state have been proposed, but the predictive capacity is quite limited for most of them. One of the main reasons is the difficulty in the description of systems with charged species under high pressures. For the cubic equation of state, the reference state is the ideal gas state, which does not make the prediction at high pressures promising. In addition, the involvement of NaCl, a charged species, increases the challenge of the research work. The Pitzer equation53 is a successful method for the system with charged species, and the extension to high pressures has been carried out. The intrinsic characterization of the GE model, however, limits the description of other thermodynamic properties like density.

318 Chapter 10 SAFT2 has been proposed to represent the properties of aqueous NaCl solutions up to high pressures. Meanwhile, it is worth mentioning that the hard sphere is used as a reference state in SAFT-based models, which ensures reliable prediction at high pressures because the higher the pressure, the closer the system goes to hard sphere. SAFT2 has been extended to CO2(2)eH2O(3)eNaCl(4) systems, in which one type of cross association was assigned, between the site of type O in CO2 and the site of type H in H2O. The cross association is temperature dependent in the following equations: εOH 23 ¼ C1 þ C2 T 1 þ C3 T 2 þ C4 T 3 k 2 3 kOH 23 ¼ C3 þ C4 T þ C5 T þ C6 T

(46) (47)

in which Cs are constants, k is the Boltzmann constant, and T is the temperature in Kelvin. A temperature-dependent binary interaction parameter k23 was used to adjust the crossdispersive energy between CO2(2) and H2O(3). k23 ¼ C9 þ C10 T þ C11 T 2 þ C12 T 3

(48)

A temperature-dependent binary interaction parameter, k24, the same for both CO2eNaþ and CO2eCl pairs, was used to adjust the short-range interactions between CO2(2) and ions(4). u24þ ¼ ðu2 u4þ Þ1=2 ð1  k24 Þ

(49a)

1=2

u24 ¼ ðu2 u4 Þ ð1  k24 Þ k24 ¼ C13 þ ðT  273:15Þ$C14

(49b) (50)

The coefficients to calculate temperature-dependent cross parameters were obtained from the fitting of experimental data33 and are listed in Table 6. The modeling results for CO2eH2OeNaCl have been discussed earlier.33,34 To further illustrate the model performance, Figure 6 shows the two examples. As shown in Table 6: Cs in Eqns (46e58) Parameter

Value

Parameter

C1 C2 (T) C3 (T2) C4 (T3) C5 C6 (T1) C7 (T2) C8 (T3) C9

9.1029  10 9.0662  106 3.4164  109 4.3221  1011 1.5329  100 1.2150  102 3.1406  105 2.5884  108 9.2661  101 3

1

C10 (T ) C11 (T2) C12 (T3) C13 C14 (T1) C15 C16 (T1) C17 C18 (T1)

Value

Parameter 3

9.3469  10 2.7383  105 2.3398  108 8.0059  102 1.4236  103 1.5154  102 2.7904  100 1.8143  103 9.6423  105

C19 C20 (T1) C21 C22 C23 C24 (K1) C25 C26 (K1) e

Value 1.9109  101 8.1684  104 6.5349  102 1.783  103 3.72163  102 5.5909  107 4.2800  101 1.6560  103 e

CO2 Storage in Deep Saline Aquifers 319

Figure 6 (a) Mole fractions of H2O in CO2-rich phase ðyH2 O Þ for CO2eH2O at 308.15 K and pressures (P) up to 600 bars. Experimental data: diamonds60; triangles61;d, calculated. (b) Mole fractions of CO2 in H2O-rich phase ðxCO2 Þ for CO2eH2OeNaCl at 313.15 K, different pressures (P), and salt concentration (molality, m). Experimental data: squares, m ¼ 062; leftward triangles, m ¼ 0.5263; rightward triangles, m ¼ 0.5364; circles: m ¼ 2.563; downward triangles: m ¼ 463; diamonds, m ¼ 465; upward triangles m ¼ 6.65d, calculated (m ¼ 0.53, 2.5, 4.0, and 5.999). ┄, calculated (m ¼ 0).

Figure 6(a), the model can represent the minimum H2O concentration in the CO2-rich phase in CO2eH2O. For the CO2eH2OeNaCl system, the CO2 solubility decreases with increasing salt concentration (molality, m, mol/kgH2O), known as the salt-out effect, and the model captures not only the pressure effect on the CO2 solubility but also the saltingout effect as shown in Figure 6(b). It has been shown earlier that the activity of water in brine and the density of brine are almost the same values as those for aqueous NaCl solution, which means that it is reasonable to simplify brine as aqueous NaCl solutions when the thermodynamic properties and phase equilibrium are the research focus. To further verify this statement, CO2 solubility and aqueous solution density with dissolved CO2 calculated for CO2eH2OeNaCl were compared with the experimental data of CO2 solubility in brine and the solution density of brine with dissolved CO2 measured by Yang and Gu.66 The comparison is shown in Figure 7 with good agreement. This observation proves again that it is reasonable to simplify brine as aqueous NaCl solutions when the thermodynamic properties and phase equilibrium are the research focus.

320 Chapter 10

Figure 7 CO2 solubility in brine and the aqueous solution density with dissolved CO2. Symbols: (a) experimental CO2 solubility in brine and (b) experimental brine density with dissolved CO2.66 Curves: model prediction for CO2eH2OeNaCl (mNaCl ¼ 0.07414 mol/kgH2O).

Using the SAFT2 model, the phase equilibria of CO2eH2O at 298.15, 323.15, 348.15, and 373.15 K and at pressures up to 400 bars were predicted. The corresponding equilibrium compositions in H2O-rich (aqueous) and CO2-rich (nonaqueous) phases are illustrated in Figure 8. The solubility of H2O in CO2 decreases and then increases with increasing pressure, and it increases with increasing temperature. The equilibrium H2O content in CO2 is a limiting factor for CO2 transportation. To avoid the risk of corrosion and hydrate formation, it is better to choose the pressure range in which the H2O content increases with increasing pressure. The CO2 solubility in H2O increases rapidly and then slowly with increasing pressure at low temperatures. With increasing temperature, the CO2 solubility in water decreases in the low-pressure range, and it decreases first and then increases in the high-pressure range. This is because at low temperatures, the phase equilibrium is changed from a vaporeliquid equilibrium to liquideliquid equilibrium with increasing pressures. With increasing temperature, this phenomenon weakens and disappears. The CO2 solubility in aqueous NaCl solutions and the solution density with dissolved CO2 were predicted at temperatures from 298.15 to 373.15 K and at pressures up to 400 bars with 2 mol/kgH2O of NaCl. As shown in Figure 9, similar to the CO2 solubility in water, the CO2 solubility in aqueous NaCl solutions increases rapidly and then slowly with increasing pressure at low temperatures. With increasing temperature,

CO2 Storage in Deep Saline Aquifers 321

Figure 8 Phase equilibrium for CO2eH2O.

Figure 9 CO2 solubility in aqueous NaCl solutions and the aqueous solution density with dissolved CO2.

the CO2 solubility in aqueous NaCl solutions decreases at low pressures, and it decreases first and then increases at high pressures. Compared to CO2 solubility in water, the CO2 solubility in aqueous NaCl solutions decreases with an increasing concentration of NaCl.

322 Chapter 10 1068

1000

ρ, kg/m3

1066

(b)

xCO2’ = 1 800

xCO2’ = 0.75

1064

xCO2’ = 0.5

1062 xCO2’ = 0.25 1060

0.01 0.02 xCO2+H2S

600 H2S%: 0, 20, 40, 60, 80 323.15 K

400 200

xCO2’ = 0 1058 0

ρ, kg/m3

(a)

0.03

0

0

100

200 P, bar

300

400

Figure 10 Density of (a) aqueous solution with dissolved gas at 135 bars, 348.15 K and mNaCl ¼2 mol/kgH2O and (b) mixture of H2S þ CO2 at 323.15 K.

The density of solution at equilibrium is depicted in Figure 9(b). The density at equilibrium increases with increasing pressure and decreasing temperature. At a fixed temperature, with increasing pressure, the CO2 solubility increases. From the results shown in Figure 9(b), it is not clear whether the increase in dissolved CO2 will cause an increase or decrease in the solution density. To study specifically the effect of dissolved CO2 on the solution density, a special case was chosen at 348.15 K, 135 bars and with 2 mol/kgH2O NaCl. The model results are shown in Figure 10(a). It is clear that the dissolved CO2 will increase solution density. This is very important for the CO2dissolution process. With CO2 dissolution, the aqueous solution density becomes denser, which can induce natural convection and hence enhance CO2 dissolution (solubility trapping) in the reservoir.

3.4 H2S(1)eCO2(2)e Brine (H2O(3)eNaCl(4)) H2S is one of the most common components in natural gas and products derived from oil processing and production. For example, the natural gas industry in the Alberta Basin, Canada, produces a significant amount of sour gases (H2S þ CO2). Because surface desulfurization and surface storage is not economical and carries significant liability, acid gas (mixture of CO2 and H2S, also referred to as “sour gas”) disposal is carried out through injections into depleted oil and gas reservoirs and deep saline formations. In addition, CO2 streams from power stations and other CO2 intensive industries are generally mixtures, and the separation of CO2 from gas mixtures is the main expense of the CO2-capture process. In fact, it is estimated that three-quarters of the total cost is spent on CO2 separation from gas streams. H2S is one of the main impurities of flue-gas streams

CO2 Storage in Deep Saline Aquifers 323 from thermal power plants. The injection of H2SeCO2 mixtures can lead to the potential for great capital and energy savings in capture, but it may also seriously increase the costs and risks of CO2 transportation and storage. For example, the inclusion of H2S will increase the risk of pipeline corrosion, and the dissolution of H2S into the brine in the reservoirs seriously alters the geochemistry. To understand the effects of H2S on CO2 transportation and geological storage, the SAFT-based model has been developed for H2SeH2O, H2SeH2OeNaCl, and H2SeCO2 systems.32,34,35 The model results have been verified with available experimental data to ensure their accuracy. In modeling H2S(1)eH2O(3),32 cross association between the site H in H2S and the site O in H2O was allowed, and two temperature-dependent parameters were used to describe this cross association. Meanwhile, a temperature-dependent binaryinteraction parameter was used to adjust the cross-dispersive energy for this binary system with the equations: εHO 13 ¼ C15 þ C16 T k kHO 13 ¼ C17 þ C18 T k13 ¼ C19 þ C20 T

(51) (52) (53)

Cross parameters were fitted to mole fractions both in H2S-rich/vapor phases and H2O-rich phases measured experimentally. The fitted results are listed in Table 6. The model is found to represent well the phase equilibria of the H2SeH2O system from 273 to 630 K and at pressures up to 200 bars. For H2S(1)eCO2(2), we assumed only one type of cross association (i.e., between the site of type H in H2S and the site of type O in CO2). Because the self-association in pure H2S and CO2 is of different origin (i.e., only the former is due to hydrogen bonding), we prefer not to use the common mixing rules for the association parameters. The cross-association parameters were directly fitted to the experimental data and found to be constants: εHO 12 ¼ C21 k kHO 12 ¼ C22

(54) (55)

in which εHO 12 is the well depth of the siteesite interaction potential energy between a site of type H in component 1 and a site of type O in component 2, and kHO 12 is a measure of the volume available for bonding between a site of type H in component 1 and a site of type O in component 2. A temperature-dependent binary-interaction parameter k12 is used to adjust the crossdispersive energy for this binary system: k12 ¼ C23 þ C24 T

(56)

324 Chapter 10 One additional binary interaction constant, the same k14, was used for both H2SeNaþ and H2SeCl pairs to adjust the short-range interactions between segments H2S and Naþ/Cl, u14þ ¼ ðu1 u4þ Þ1=2 ð1  k14 Þ u14 ¼ ðu1 u4 Þ þ

1=2

ð1  k14 Þ

(57a) (57b)



in which subscripts 4þ and 4 denote the Na and Cl , respectively. The temperature-dependent cross parameter was allowed with the following equation: k14 ¼ c25 þ c26 =T

(58)

The model has been extended to CO2eH2SeH2OeNaCl, the effect of H2S on both CO2 transportation and geological storage was discussed in recent publications,34,35 and the main observations are summarized here. To study the effect of H2S on CO2 transportation, using the developed model, the phase state and density for CO2eH2S streams were predicted. The model prediction shows that with inclusion of H2S, phase split may occur, and the density increases and then decreases with increasing pressure. Figure 10(b) shows the model prediction at 323.15 K in which the dashed lines represent two-phase flow and the solid lines represent density in the single-phase region. In CO2 transportation, two-phase flow in the pipeline could cause cavitation and pressure peaks and would damage the pipeline. Two-phase flow is also problematic in the operation of pumps, as well as compressors and injection wells. Therefore, the inclusion of H2S increases the difficulty for CO2 transportation. The density of the (supercritical)- gas phase reflects the amount of fluid that can be stored at a fixed volume in reservoirs, as well as the transportation capacity. As shown in Figure 10(b), due to the inclusion of H2S the mass density increases at low pressures, which means that the volume of the mixture decreases, leading to an increase in the amount of the gases that can be stored at a fixed volume in the reservoirs. Thus, the inclusion of H2S in this low-pressure range is favorable for CO2 geological storage in terms of gas volume. The inclusion of H2S will also affect the gas solubility as well as the aqueous solution density with dissolved gas. Figure 10(a) shows the density of aqueous solution density with dissolved H2S þ CO2 at 348.15 K, 135 bars and with 2 mol/kgH2O NaCl. With inclusion of H2S, the aqueous solution density with dissolved gas decreases and may even be less than that without dissolved gas. This is very important for the gas-dissolution process. If the dissolution of gas into brine causes an increase in brine density, this density difference may also be large enough to trigger Rayleigh instability, which can strongly enhance dissolution processes due to mixing. However, if the dissolution of gas into the brine decreases the brine mass density, the brine can be subject to buoyancy-driven

CO2 Storage in Deep Saline Aquifers 325 0.05

0.03

xCO2 or xH2S or xCO2+H2S

xCO2 + xH2S

0.04

0.03

0.02 H2S%: 0, 20, 40, 60, 80 323.15 K

0.02

0.01

0.01

0

0

100

200 P, bar

300

400

0

0

0.2

0.4 0.6 xH2S/(xH2S+xCO2)

0.8

1

Figure 11 Gas solubility (a) at 323.15 K, mNaCl ¼ 0; (b) at 348.15 K, 135 bars, mNaCl ¼ 2 mol/kgH2O.

migration (and potential escape from formation) associated with separate gas phases. Thus, the inclusion of H2S is not good for the dissolution process. As shown in Figure 11, H2S is more soluble in H2O or brine than is CO2, and CO2 can also be dissolved in liquid H2S. With the inclusion of H2S, the solubility of the gas mixture of (H2S þ CO2) is higher than that for pure CO2. With increasing inclusion of H2S, the H2S solubility increases, whereas the CO2 solubility decreases.

3.5 Other Related Work Accurate evaluation of the capacity of a saline aquifer for CO2 storage and the fate of the injected fluids in sedimentary basins requires analysis of the thermodynamic properties (density, phase equilibrium, interfacial tension, etc.) and transport properties (viscosity, diffusion coefficients, thermal conductivities, etc.). For any multiphase flow system, viscosity plays an important role. Viscosity characterizes the fluids’ resistance with respect to deformation under shear stress. The lower a fluid’s viscosity, the lower is its resistance to flow and displacement by another fluid. Therefore, accurate prediction of viscosity is extremely important. Interfacial tension is one of the important properties linked to CO2-storage capacity and release risk. Physical trapping by capillary forces is one of the key physical processes, which control the capacity of a saline aquifer for CO2 storage. This process is controlled by interfacial tension, wettability, and pore size. In addition, to ensure the stability and long-term viability of geologic carbon storage, injected CO2 must be kept in place by an overlying caprock of very low permeability. Capillary forces in the caprock act to prevent

326 Chapter 10 upward migration and escape of the stored supercritical fluid, with interfacial tension between the aqueous brine phase and the CO2 phase being the primary control. Beyond caprock sealing capacity, the flow of CO2 through reservoir rock and residual CO2 trapping following plume imbibition are also strongly impacted by capillary effects that directly relate to interfacial tension. Meanwhile, deep saline formation is a porous rock formation, and the properties of porous rock will affect the CO2-storage process. For example, the porosity will affect the CO2storage capacity. The mineral surface will absorb fluids and then the confined space will lead to an inhomogeneous profile of fluid inside the pores. The fluid behavior and the corresponding properties in confined pores are different from the bulk phase and depend strongly on the pore size, surface of rock (substrates), as well as shape. 3.5.1 SAFT-Density Functional Theory (DFT) Model Representing Fluid Properties in Confined Pores Based on SAFT, several groups have developed DFT models, but with the focus on the interfacial properties for which the external field is weak. For example, Gloor et al.67e69 developed a DFT model on the basis of SAFT-variable range (VR) with a local density approximation, and the model accurately predicts the interfacial tension. Based on perturbed-chain polar statistical associating fluid theory (PCP-SAFT), Gross70 proposed a DFT reproducing the surface tension of nonpolar and polar substances. The free energy functional for dispersion attraction is expressed as a first-order perturbation term. As the functional is not consistent with bulk PC-SAFT, a local density approximation was used to compensate for the difference. For confined fluids in which the external field plays an important role, a hybrid model PC-SAFT-DFT has been developed 71,72 in which the fluid behavior in nanopores and bulk phase can be represented consistently and simultaneously. In the developed model, the modified fundamental measure theory was used for the hard-sphere contribution, and the dispersion free-energy functional was represented with weighted-density approximation by averaging the density in the range of interaction. In addition, the chain-free-energy functional from interfacial SAFT was used to account for the chain connectivity, and the ionic term was based on local density approximation. The model performance for a real substance, especially for the small molecules, was further investigated by studying the gas adsorption in activated carbons and zeolite silicates.73 The substance of gas was modeled as chain molecules with the parameters taken from the bulk, whereas the parameters of solid surface were obtained from the fitting of gas-adsorption isotherms measured experimentally. The results show that the model can reliably reproduce the confined behaviors of physically existing substances

CO2 Storage in Deep Saline Aquifers 327 in nanopores. Further investigation of the effect of pore-size distribution on the model performance reveals that consideration of pore-size distribution can improve the model performance but costs much more computing time. 3.5.2 SAFT-Based Model Representing Surface Tension and Viscosity The SAFT-based model has been developed to represent surface tensions of fluid and mixtures. Two methods exist: the coupling of density gradient theory (DGT) and the density function theory (DFT). It is more common to use DGT to study the surface tension, and PCP-SAFT,74 soft-SAFT,75 PC-SAFT76, and SAFT-VR77 have been studied. Particularly, the surface tensions of CO2eH2O, and CO2 þ hydrocarbon have been studied by several works.74,76e78 Meanwhile, based on SAFT, DFT has been developed to study the surface tension as described in the foregoing section. Only one publication, however, was on CO2eH2Oen-alkanes.79 The free-volume theory and the friction theory can be combined with SAFT to represent the viscosity of fluids and mixtures, which has been described in the review article.80 Since 2008, the study of viscosity for CO-storage-related systems has not made much progress, and more research needs to be done in this area.

4. Summary and Perspectives Deep saline aquifers hold the largest capacity for CO2 storage. However, injection of CO2 into geologic formations will result in a number of processes such as multiphase flow, solute transport, and chemical reactions among CO2ebrineerocks. Both multiphase flow and chemical reactions require us to understand the thermodynamics of mixed fluids, which is the focus of this review chapter. For the thermodynamics and kinetics of mineraleCO2eH2O interactions, please refer to Lu et al.81 and Liu et al.82 Although many experimental data and EOS are available for the major components of the CO2eNaCleH2O fluids relevant to CCS, both experimental data and theoretical models are lacking for systems containing impurities such as H2O, H2S, SOx, NOx, CO, amine, CH4, H2, O2, Mono-Ethylene Glycol (MEG), or Tri-Ethylene Glycol (TEG). The dissolution of a trace amount of some impurities can drastically alter the brine chemistry. The challenge for CCS is to evaluate these impurities because of the urgent need for both technical operation and regulatory review of impure CO2 streams for injection.

Acknowledgment Xiaoyan Ji acknowledges the support of the Swedish Research Council. Chen Zhu acknowledges the support of the U.S. Department of Energy grant DE-FE0004381, the Norwegian Center of Excellence Subsurface CO2 storagedCritical Elements and Superior Strategy (SUCCESS), and the State Key

328 Chapter 10 Laboratory of Ore Deposits at the Institute of Geochemistry, Chinese Academy of Sciences. Although the work was partly sponsored by an agency of the United States Government, the views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

Symbols and Nomenclature a~res Bab,i ci cij d Dmn e gSW ab k kab m mi NAv n(Gi) P qj T T* R u u/k ua uab voo va X xi (Xi) x0i xa X Aai zj Z b ε εw bi 4 kAB k l lab rm rn r*

dimensionless residual Helmholtz free energy bond fraction of type ab in molecule of component i constant for calculating kab, u/k, or cross-association parameters ε and k universal constants listed in Table 1 effective (hydrated) diameter universal constants listed in Table 2 charge of an electron square-well radial distribution function Boltzmann constant binary interaction parameter segment number number of segments of component i Avogadro number number of association sites on molecule i pressure in bars charge of ion j absolute temperature in Kelvin dimensionless temperature (¼ kT/u) gas constant well depth of square-well potential segment energy segment energy of segment a well depth of square-well potential for the aeb interaction segment volume molar volume of segment a dimensionless quantity (¼ kd) mole fraction of component i mole fraction of ion i on a solvent-free basis segment fraction mole fraction of molecule of component i not bonded at side A of segment a valence of the ion j compressibility factor 1/kT well depth of the association siteesite potential dielectric constant of water fugacity coefficient of component i parameter related to the volume available for bonding between sites A and B Debye inverse screening length segment reduced range of the potential well reduced range of the potential well for the aeb interaction molar density number density reduced density

CO2 Storage in Deep Saline Aquifers 329 sa sab DAai Bbj DAai Bbj

diameter of segment a distance between centers of segment a and b at contact close-packed reduced density (¼ 21/2p/6) association strength between site A a at molecule of component i and site B b at molecule of component j

References 1. 2. 3. 4. 5. 6.

7.

8. 9.

10. 11. 12. 13.

14.

15.

16. 17. 18.

IPCC fourth assessment report: climate change. 2007. Directive 2009/31/EC on the geological storage of carbon dioxide. 2009. IEA. Energy technology perspectives. 2010. Bachu S, Gunter WD. Acid gas injection in the Alberta basin, Canada: a CO2-storage experience. Geological storage of carbon dioxide, vol. 233. London: Geological Society, Special Publications; 2004. Bachu S, Watson TL. Review of failures for wells used for CO2 and acid gas injection in Alberta, Canada. Energy Procedia 2009;1(1):3531e7. Bachu S, Carroll J. In-situ phase and thermodynamic properties of resident brine and acid gases (CO2 and H2S) injected in geological formations in western Canada: demonstration of CO2 geological storage, carbon dioxide capture for storage in deep geologic formationsdresults from the CO2 capture Project. In: Benson SM, editor. Geologic storage of carbon dioxide with monitoring and verification, vol. 2. London: Elsevier; 2005. p. 867e76. IPCC. IPCC special report on carbon dioxide capture and storage. In: Metz B, Davidson O, de Coninck HC, Loos M, Meyer LA, editors. Prepared by working group III of the intergovernmental panel on climate change. Cambridge University Press; 2005. Ji YH, Ji XY, Feng X, Liu C, Lu LH, Lu XH. Progress in the study on the phase equilibria of the CO2-H2O and CO2-H2O-NaCl systems. Chin J Chem Eng 2007;15(3):439e48. Song YC, Zhan YC, Zhang Y, Liu SY, Jian WW, Liu Y, et al. Measurements of CO2-H2O-NaCl solution densities over a wide range of temperatures, pressures, and NaCl concentrations. J Chem Eng Data 2013;58(12):3342e50. Hou SX, Maitland GC, Trusler JPM. Phase equilibria of (CO2 þ H2O þ NaCl) and (CO2 þ H2O þ KCl): measurements and modeling. J Supercrit Fluids 2013;78:78e88. Akinfiev NN, Diamond LW. Thermodynamic model of aqueous CO2-H2O-NaCl solutions from 22 to 100 degrees C and from 0.1 to 100 MPa. Fluid Phase Equilibria 2010;295(1):104e24. Duan ZH, Sun R, Liu R, Zhu C. Accurate thermodynamic model for the calculation of H2S solubility in pure water and brines. Energy & Fuels 2007;21(4):2056e65. Liu YH, Hou MQ, Ning H, Yang DZ, Yang GY, Han BX. Phase equilibria of CO2 þ N2 þ H2O and N2 þ CO2 þ H2O þ NaCl þ KCl þ CaCl2 systems at different temperatures and pressures. J Chem Eng Data 2012;57(7):1928e32. Duan ZH, Li DD. Coupled phase and aqueous species equilibrium of the H2O-CO2-NaCl-CaCO3 system from 0 to 250 degrees C, 1 to 1000 bar with NaCl concentrations up to saturation of halite. Geochim Cosmochim Acta 2008;72(20):5128e45. Dubessy J, Tarantola A, Sterpenich J. Modelling of liquid-vapour equilibria in the H2O-CO2-NaCl and H2O-H2S-NaCl systems to 270 degrees C. Oil Gas Sci Technol-Revue De L Institut Francais Du Petrole 2005;60(2):339e55. Spycher N, Pruess K. A phase-partitioning model for CO2-brine mixtures at elevated temperatures and pressures: application to CO2-enhanced geothermal systems. Transp Porous Media 2010;82(1):173e96. Mao SD, Zhang DH, Li YQ, Liu NQ. An improved model for calculating CO2 solubility in aqueous NaCl solutions and the application to CO2-H2O-NaCl fluid inclusions. Chem Geol 2013;347:43e58. Dubacq B, Bickle MJ, Evans KA. An activity model for phase equilibria in the H2O-CO2-NaCl system. Geochim Cosmochim Acta 2013;110:229e52.

330 Chapter 10 19. Spycher N, Pruess K, Ennis-King J. CO2-H2O mixtures in the geological sequestration of CO2. I. Assessment and calculation of mutual solubilities from 12 to 100 degrees C and up to 600 bar. Geochim Cosmochim Acta 2003;67(16):3015e31. 20. Zirrahi M, Azin R, Hassanzadeha H, Moshfeghian M. Mutual solubility of CH4, CO2, H2S, and their mixtures in brine under subsurface disposal conditions. Fluid Phase Equilibria 2012;324:80e93. 21. Li J, Duan ZH. A thermodynamic model for the prediction of phase equilibria and speciation in the H2O-CO2-NaCl-CaCO3-CaSO4 system from 0 to 250 degrees C, 1 to 1000 bar with NaCl concentrations up to halite saturation. Geochim Cosmochim Acta 2011;75(15):4351e76. 22. Perfetti E, Thiery R, Dubessy J. Equation of state taking into account dipolar interactions and association by hydrogen bonding: II - modelling liquid-vapour equilibria in the H2O-H2S, H2O-CH4 and H2O-CO2 systems. Chem Geol 2008;251(1e4):50e7. 23. Perfetti E, Thiery R, Dubessy J. Equation of state taking into account dipolar interactions and association by hydrogen bonding. I: application to pure water and hydrogen sulfide. Chem Geol 2008;251(1e4):58e66. 24. Li ZD, Firoozabadi A. Cubic-plus-association equation of state for water-containing mixtures: is “cross association” necessary? Aiche J 2009;55(7):1803e13. 25. Pappa GD, Perakisa C, Tsimpanogiannisb IN, Voutsas EC. Thermodynamic modeling of the vapor-liquid equilibrium of the CO2/H2O mixture. Fluid Phase Equilibria 2009;284(1):56e63. 26. Vega LF, Jackson G. 20 years of the SAFT equation of state-recent advances and challenges symposium held in Bellaterra, Barcelona, 19e21 September 2010 Foreword. Fluid Phase Equilibria 2011;306(1):1e3. 27. Sun R, Dubessy J. Prediction of vapor-liquid equilibrium and PVTx properties of geological fluid system with SAFT-LJ EOS including multi-polar contribution. Part I: application to H2O-CO2 system. Geochim Cosmochim Acta 2010;74(7):1982e98. 28. Sun R, Dubessy J. Prediction of vapor-liquid equilibrium and PVTx properties of geological fluid system with SAFT-LJ EOS including multi-polar contribution. Part II: application to H2O-NaCl and CO2-H2O-NaCl System. Geochim Cosmochim Acta 2012;88:130e45. 29. dos Ramos MC, McCabe C. Modeling the phase behavior, excess enthalpies and Henry’s constants of the H2O þ H2S binary mixture using the SAFT-VR plus D approach. Fluid Phase Equilibria 2010;290(1e2): 137e47. 30. Tang XH, Gross J. Modeling the phase equilibria of hydrogen sulfide and carbon dioxide in mixture with hydrocarbons and water using the PCP-SAFT equation of state. Fluid Phase Equilibria 2010;293(1):11e21. 31. Tan SP, Yao Y, Piri M. Modeling the solubility of SO2 þ CO2 mixtures in brine at elevated pressures and temperatures. Ind Eng Chem Res 2013;52(31):10864e72. 32. Ji XY, Zhu C. Modelling of phase equilibria in the H2S-H2O system with statistical associating fluid theory. Energy & Fuels 2010;24:6208e13. 33. Ji XY, Tan SP, Adidharma H, Radosz M. SAFT1-RPM approximation extended to phase equilibria and densities of CO2-H2O and CO2-H2O-NaCl systems. Ind Eng Chem Res 2005;44(22):8419e27. 34. Ji XY, Zhu C. A SAFT equation of state for the quaternary H2S-CO2-H2O-NaCl system. Geochimica Cosmochimica Acta 2012;91:40e59. 35. Ji XY, Zhu C. Predicting possible effects of H2S impurity on CO2 transportation and geological storage. Environ Sci Technol 2013;47(1):55e62. 36. Ji X, Tan SP, Adidharma H, Radosz M. Statistical associating fluid theory coupled with restricted primitive model to represent aqueous strong electrolytes: multiple-salt solutions. Ind Eng Chem Res 2005;44(19):7584e90. 37. Ji XY, Adidharma H. Ion-based SAFT2 to represent aqueous single- and multiple-salt solutions at 298.15 K. Ind Eng Chem Res 2006;45(22):7719e28. 38. Tan SP, Adidharma H, Radosz M. Statistical associating fluid theory coupled with restrictive primitive model extended to bivalent ions. SAFT2: 1. Single salt plus water solutions. J Phys Chem B 2006;110(33):16694e9. 39. Ji X, Tan SP, Adidharma H, Radosz M. Statistical associating fluid theory coupled with restrictive primitive model extended to bivalent ions. SAFT2: II. Brine/Seawater properties predicted. J Phys Chem B 2006;110:16700e6.

CO2 Storage in Deep Saline Aquifers 331 40. Ji X, Adidharma H. Ion-based SAFT2 to represent aqueous single-salt solutions at temperatures and pressures up to 473.15 K and 1000 bar. Ind Eng Chem Res 2006;46(13):4667e77. 41. Ji XY, Adidharma H. Ion-based statistical associating fluid theory (SAFT2) to represent aqueous single-salt solutions at temperatures and pressures up to 473.15 K and 1000 bar. Ind Eng Chem Res 2007;46(13):4667e77. 42. Ji XY, Adidharma H. Ion-based SAFT2 to represent aqueous multiple-salt solutions at ambient and elevated temperatures and pressures. Chem Eng Sci 2008;63(1):131e40. 43. Le Thi C, Tamouza S, Passarello JP, Pascal Tobaly P, Hemptinne GC. Modeling phase equilibrium of H2þn-alkane and CO2þn-alkane binary mixtures using a group contribution statistical association fluid theory equation of state (GC-SAFT-EOS) with a k(ij) group contribution method. Ind Eng Chem Res 2006;45(20):6803e10. 44. Adidharma H, Radosz M. Prototype of an engineering equation of state for heterosegmented polymers. Ind Eng Chem Res 1998;37(11):4453e62. 45. Adidharma H, Radosz M. Square-well SAFT equation of state for homopolymeric and heteropolymeric fluids. Fluid Phase Equilibria 1999;158e160:165e74. 46. Adidharma H, Radosz M. A study of square-well statistical associating fluid theory approximations. Fluid Phase Equilibria 1999;161(1):1e20. 47. Adidharma H, Radosz M. SAFT1 assoc fluids: alkanols. J Phys Chem B 2001;105:9822e7. 48. Tan SP, Adidharma H, Radosz M. Statistical associating fluid theory coupled with restricted primitive model to represent aqueous strong electrolytes. Ind Eng Chem Res 2005;44(12):4442e52. 49. Ji XY, Adidharma H. Thermodynamic modeling of CO2 solubility in ionic liquid with heterosegmented statistical associating fluid theory. Fluid Phase Equilibria 2010;293(2):141e50. 50. Ji X, Zhu C. Modeling of phase equilibria in the H2SH2O system with the statistical associating fluid theory. Energy & Fuels 2010;24(11):6208e13. 51. Lemmon EW, McLinden MO, Friend DG. Thermophysical properties of fluid systems in NIST chemistry webBook, in NIST standard reference database 2003. Gaithersburg MD 20899: National Institute of Standards and Technology; March 2003. http://webbook.nist.gov. 52. Lemmon EW, McLinden MO, Friend DG. Thermophysical properties of fluid systems. In: Linstrom PJ, Mallard WG, editors. NIST chemistry webbook, NIST standard reference database number 69. Gaithersburg MD 20899: National Institute of Standards and Technology http://webbook.nist.gov. [cited 21.09.09]. 53. Pitzer KS, Peiper JC, Busey RH. Thermodynamic properties of aqueous sodium chloride solutions. J Phys Chem Ref Data 1984;13:1e102. 54. Platford RF. The activity coefficient of sodium chloride in seawater. J Mar Res 1965;23(2):55e62. 55. Fabuss BM, Korosi A, Huq AKMS. Densities of binary and ternary aqueous solutions of NaCl, Na2SO4, and MgSO4, of sea waters, and sea water concentrates. J Chem Eng Data 1966;11:325e31. 56. Bromley LA, Singh D, Ray P, Sridhar S, Read SM. Thermodynamic properties of sea salt solutions. AIChE J 1974;20(2):326e35. 57. Millero FJ, L.W.H. The thermodynamics of seawater at one atmosphere. Am J Sci 1976;276(9):1035e77. 58. Holmes HF, Busey RH, Simonson JM, Mesmer RE. CaCl2(aq) at elevated temperatures. Enthalpies of dilution, isopiestic molalities, and thermodynamic properties. J Chem Thermodyn 1994;26:271e98. 59. Holmes HF, Mesmer RE. Aqueous solutions of the alkaline-earth metal chlorides at elevated temperatures. Isopiestic molalities and thermodynamic properties. J Chem Thermodyn 1996;28:1325e58. 60. Valtz A, Chapoya A, Coqueleta C, Paricaudb P, Richon D. Vapour-liquid equilibria in the carbon dioxide-water system, measurement and modelling from 278.2 to 318.2 K. Fluid Phase Equilibria 2004;226:333e44. 61. King MB, Mubarak A, Kim JD, Bott TR. The mutual solubilities of water with supercritical and liquid carbon dioxide. J Supercrit Fluids 1992;5(4):296e302. 62. Wiebe R, Gaddy VL. The solubility of carbon dioxide in water at various temperatures from 12 to 40 and at pressures to 500 atmospheres. Critical phenomena. J Am Chem Soc 1940;62:815e7. 63. Kiepe J, Horstmann S, Fischer K, Gmehling J. Experimental determination and prediction of Gas solubility data for CO2 þ H2O mixtures containing NaCl or KCl at temperatures between 313 and 393 K and pressures up to 10 MPa. Ind Eng Chem Res 2002;41(17):4393e8.

332 Chapter 10 64. Bando S, Takemura F, Nishio M, Hihara E, Akai M. Solubility of CO2 in aqueous solutions of NaCl at (30 to 60)  C and (10 to 20) MPa. J Chem Eng Data 2003;48(3):576e9. ¨ cal C, Maurer G. Solubility of carbon dioxide in aqueous solutions of sodium 65. Rumpf B, Nicolaisen H, O chloride: experimental results and correlation. J Solut Chem 1994;23(3):431e48. 66. Yang CD, Gu YG. Accelerated mass transfer of CO2 in reservoir brine due to density-driven natural convection at high pressures and elevated temperatures. Ind Eng Chem Res 2006;45(8):2430e6. 67. Llovell F, Galindo A, Blas FJ, Jackson G. Classical density functional theory for the prediction of the surface tension and interfacial properties of fluids mixtures of chain molecules based on the statistical associating fluid theory for potentials of variable range. J Chem Phys 2010;133(2). 68. Gloor GJ, Jackson G, Blas FJ, Del Rı´o EM, De Miguel E. An accurate density functional theory for the vapor-liquid interface of associating chain molecules based on the statistical associating fluid theory for potentials of variable range. J Chem Phys 2004;121(24):12740e59. 69. Gloor GJ, Jackson G, Blas FJ, Del Rı´o EM, De Miguel E. Prediction of the vapor-liquid interfacial tension of nonassociating and associating fluids with the SAFT-VR density functional theory. J Phys Chem C 2007;111(43):15513e22. 70. Gross J. A density functional theory for vapor-liquid interfaces using the PCP-SAFT equation of state. J Chem Phys 2009;131(20). 71. Shen GL, Ji XY, Lu XH. A hybrid perturbed-chain SAFT density functional theory for representing fluid behavior in nanopores. J Chem Phys 2013;138(22). 72. Shen GL, Ji XY, Oberg S, Lu XH. A hybrid perturbed-chain SAFT density functional theory for representing fluid behavior in nanopores: mixtures. J Chem Phys 2013;139(19). 73. Shen GL, Ji XY, Lu XH. Modeling of molecular gas adsorption isotherms on porous materials with hybrid PC-SAFT-DFT. Fluid Phase Equilibria 2014;382:116e26. 74. Nino-Amezquita G, van Putten D, Enders S. Phase equilibrium and interfacial properties of water þ CO2 mixtures. Fluid Phase Equilibria 2012;332:40e7. 75. Llovell F, Vilasecaa O, Junga N, Vegaa LF. Waterþ1-alkanol systems: modeling the phase, interface and viscosity properties. Fluid Phase Equilibria 2013;360:367e78. 76. Li XS, Liu JM, Fu D. Investigation of interfacial tensions for carbon dioxide aqueous solutions by perturbed-chain statistical associating fluid theory combined with density-gradient theory. Ind Eng Chem Res 2008;47(22):8911e7. 77. Lafitte T, Mendiboure B, Pineiro M, Bessieres D, Miqueu C. Interfacial properties of Water/CO2: a comprehensive description through a gradient theory-SAFT-VR Mie approach. J Phys Chem B 2010;114(34):11110e6. 78. Fu D, Wei YZ. Investigation of vapor-liquid surface tension for carbon dioxide and hydrocarbon mixtures by perturbed-chain statistical associating fluid theory combined with density-gradient theory. Ind Eng Chem Res 2008;47(13):4490e5. 79. Amezquita OGN, Enders S, Jaeger PT, Eggers R. Interfacial properties of mixtures containing supercritical gases. J Supercrit Fluids 2010;55(2):724e34. 80. Tan SP, Adidharma H, Radosz M. Recent advances and applications of statistical associating fluid theory. Ind Eng Chem Res 2008;47(21):8063e82. 81. Lu P, Fu Q, Seyfroed Jr WE, Hedges SW, Soong Y, Jones K, et al. Coupled alkali feldspar dissolution and secondary mineral precipitation in batch systems: 2. New experiments with supercritical CO2 and implications for carbon sequestration. Appl Geochem 2013;30:75e90. 82. Liu FY, Liu P, Zhu C, Xiao YT. Coupled reactive transport modeling of CO2 sequestration in the Mt Simon Sandstone formation, Midwest U.S.A. Int J Greenh Gas Control 2011;52:297e307.