Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews

journal homepage: www.elsevier.com/locate/rser

A systematic review on CO2 capture with ionic liquids: Current status and T
future prospects
⁎
Mahsa Aghaie, Nima Rezaei, Sohrab Zendehboudi
Faculty of Engineering and Applied Science, Memorial University, St. John's, NL, Canada

A R T I C LE I N FO A B S T R A C T

Keywords: Global warming due to the emission of greenhouse gases, especially carbon dioxide (CO2), has a significant effect
CO2 capture on the climate change and has become a widespread concern in the recent years. Carbon capture, utilization, and
Climate change sequestration (CCUS) strategy appears to be effective in decreasing the carbon dioxide level in the atmosphere.
Absorption Despite a great progress in this field, there are still major limitations in commercialized the CO2 capture methods
Ionic liquids
that rely on absorption phenomena. High capital costs of for the CO2 capture, low absorption and desorption
Physical and thermodynamic properties
rates (which require large facilities), solvent losses due to evaporation, and the use of corrosive solvents are
among main obstructions. Recently, CO2 capture with ionic liquids (ILs) has appreciably attracted researchers’
attention. The distinct properties of ILs such as negligible vapor pressure and their affinity to capture the CO2
molecules make them a feasible alternative for currently available solvents including, different amines.
This paper covers a brief review of previous engineering and research works on various CO2 capture tech-
niques, the description of CO2 capture process using ILs, mechanisms of the CO2 capture with ILs at molecular
level, CO2 and ILs properties, characterization of the CO2/IL systems, impacts of operating and fluids conditions
on CO2 absorption capacity by ILs, and CO2 solubility and selectivity in ILs. Moreover, the technical and eco-
nomic aspects of the CO2 capture with ILs, screening criteria for ILs/CO2 systems, and important results obtained
from previous studies will form the last parts of this manuscript. This review offers a proper/systematic guideline
that assists researchers and engineers to comprehensively understand and to effectively design the CO2/ILs
processes, focusing on the thermodynamic and mass transfer aspects.

1. Introduction removing the carbon dioxide from flue-gases produced by power plants
can effectively decrease the carbon level in the Earth's atmosphere. The
It is believed that global warming is mainly resulted from the production, processing, and utilization of fossil fuel sources such as oils,
greenhouse gases (water vapor, carbon dioxide, methane, and nitrous coals, and natural gases undoubtedly will continue. Thus, it seems ne-
oxide) which are trapping the heat in the Earth's atmosphere. Carbon cessary to use them as clean as feasible through employing effective
dioxide (CO2), which remains longer than other gases in the atmo- strategies for carbon and other pollutants reduction. Carbon capture,
sphere, is more responsible for global warming and climate changes utilization, and storage (CCUS) is an emerging method which helps to
[1]. Carbon dioxide is naturally produced and consumed in the carbon better manage CO2 concentration through various ways such as cap-
cycle. As humans are currently consuming the long-buried carbons in turing carbon, sequestration, or/and turning it to useful chemicals
the forms of fossil fuels and coal in significant amounts, excessive [4,5].
carbon is released in the atmosphere. On the other hand, the oceans (as Carbon capture processes, in general, are divided into three main
a natural absorbent of CO2) are not able to absorb all emitted carbon categories including pre-combustion, post-combustion, and oxy-com-
dioxide from the atmosphere [2]. bustion. In the post-combustion and oxy-combustion methods, the fuel
Intergovernmental Panel on Climate Change (IPCC) report shows is burned with air and oxygen, respectively. The heat released in these
that about 79% of the total CO2 emissions come from fossil fuels and methods is used to generate electricity. The main difference between
minerals that are being used for power generation [3]. Coal power the post-combustion and oxy-combustion processes is the composition
plants (operated by natural gas and oil) with 60% of the total CO2 of the flue-gas. In the latter technique, the flue-gas contains highly
emissions have a great contribution to the global warming [3,4]. Hence, concentrated CO2, which seems suitable for underground storage.

⁎
Corresponding author.
E-mail addresses: [email protected], [email protected] (S. Zendehboudi).

https://doi.org/10.1016/j.rser.2018.07.004
Received 30 December 2017; Received in revised form 18 March 2018; Accepted 1 July 2018
Available online 29 August 2018
1364-0321/ © 2018 Elsevier Ltd. All rights reserved.
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Nomenclatures CF3-CFH2CH3SO3− methylsulfate anion

[CH3CH2NH3][NO3] ethylammonium nitrate
Acronyms CO2 carbon dioxide
CO32- carbonate
AAD average absolute deviation [C2Mpyrr][Br] 1-ethyl-1-methylpyrrolidinium bromide
AIL aprotic ionic liquid [CnCnMIM]+ 1-alkyl-3-alkylimidazolium cation
ANN artificial neural network (n = carbon chain length)
CILs conventional ionic liquids [CnMIM]+ 1-alkyl-3-methylimidazolium cation
CC capital cost (n = carbon chain length)
CCS carbon capture and storage [C2mim][NTf2] 1-ethyl-3-methylimidazolium bis(trifluorome
CCUS carbon capture, utilization, and storage thylsulfonyl)imide
DE differential evolution [emim][EtSO4] 1-ethyl-3-methylimidazolium ethyl sulfate
DESs deep eutectic solvents [emim][TfO] 1-ethyl-3-methylimidazolium trifluorome
DETL differential evolution with Tabu List thanesulfonate
EOS equation of state HCO3- bicarbonate
GA genetic algorithm Hg mercury
GHG greenhouse gas hmg hexamethylguanidinium
HBDs hydrogen bond donors [hmim][BF4] 1-hexyl-3-methylimidazolium tetrafluoro
IGCC integrated gasification combined cycle borat
IL ionic liquid [hmim] [Ntf2] 1-hexyl-3-methylimidazolium trifluorome
IPCC Intergovernmental Panel on Climate Change thanesulfonate
IUPAC International Union of Pure and Applied [hmim][PF6] 1-hexyl-3-methylimidazolium hexafluoro
Chemistry phosphate
LLE liquid-liquid equilibria HSO4− hydrogen sulfate anion
LM Levenberg-Marquart MDEA methyl diethanolamine
MD molecular dynamics MEA monoethanolamine
MOFs metal-organic frameworks [N-bupy][BF4] N-butylpyridinium tetrafluoroborate
NRTL non-random two liquid equation of state N(CN)2− dicyanamide anion
OC operating cost N4111 tetrabutylammonium
OME operating and maintenance expenses NaCl sodium chloride
PC-SAFT perturbed chain statistical association fluid Nn,n,n,n+ tetraalkylammonium cation (n = carbon
theory chain length)
PILs protic ILs NO2 nitrogen dioxide
PR Peng-Robinson equation of state NPf2− bis(pentafluoroethylsulfonyl)imide anion
PT Patel-Teja O2 oxygen
PRSV Peng-Robinson-Stryjek-Vera equation of state OAc− acetate anion
PSO particle swarm optimization omim 1-octyl-3-methyl imidazolium
RK Redlich-Kwong equation of state [omim][BF4] 1-methyl-3-octylimidazolium
SA simulated annealing tetrafluoroborate
SAFT statistical association fluid theory [omim][Cl] 1-octyl-3-methylimidazolium chloride
SRK Soave-Redlich-Kwong equation of state PDMS polydimethylsiloxane
TGA thermogravimetric analysis [PF6]- hexafluorophosphate
TSIL task-specific ionic liquid pmg pentamethylguanidinium
vdW van der Waals PP polypropylene
VL van Laar equation of state ppg pentamethylpropylguanidinium
VLE vapor-liquid equilibria PPO polyphenyleneoxide
WS Wong-Sandler mixing rule [P14666][2-CNpyr] trihexyl(tetradecyl)phosphonium
2-cyanopyrrolide
Chemicals [P14666][3-CF3pyra] trihexyl(tetradecyl)phosphonium
3-(trifluoromethyl) pyrazolide
AMP 2-amino-2-methl-1-propanol [P14666][Br] trihexyl(tetradecyl)phosphonium bromide
BF4− tetra fluoroborate anion [P14666][Cl] trihexyl(tetradecyl)phosphonium chloride
[bmim][BF4] 1-butyl-3-methylimidazolium tetrafluoroborate [P14666][Gly] trihexyl(tetradecyl)phosphonium glycinate
[bmim] [Br] 1-butyl-3-methylimidazolium bromide [P14666][Ile] trihexyl(tetradecyl)phosphonium isoleucinate
[bmim] [dca] 1-butyl-3-methylimidazolium dicyanamid Pn,n,n,n+ tetraalkylphosphonium cation (n = carbon
[bmim][NO3] 1-butyl-3- methylimidazolium nitrate chain length)
[bmim][NTf2] 1-butyl-3-methylimidazolium bis(trifluorome [Pyn,n]+ 1-alkyl-1-alkylpyrrolidinium cation (n = carbon
thylsulfonyl)imide chain length)
[bmim][CF3SO3] 1-butyl-3-methylimidazolium trifluorome [Rmim][Tf2N] 1-alkyl-3-methylimidazolium (alkyl =
thanesulfonate methyl, butyl, octyl, and dodecyl)
[bmim][PF6] 1-butyl-3-methylimidazolium hexafluorophosphate bis(trifluoromethylsulfonyl)imides
[bmim][TfO] 1-butyl-3-methylimidazolium trifluorome SCN− thiocyanate anion
thanesulfonate SO2 sulphur dioxide
Br− bromide anion tba tetrabutylammonium

503
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

[TBMA][Cl] tributylmethylammonium chloride and (A.10)

tbp tetrabutylphosphonium Iji binary interaction coefficient between molecules
tea tetraethylammonium j and i in Eq. (A.8) and Table A.1
tma tetramethylammonium k transmission coefficient in Eq. (24)
tmg tetramethylguanidinium k1 coefficient in energy parameter model in Table 8
ttp trihexyl(tetradecyl)phosphonium kij binary interaction coefficient in Eqs. (A.1),
(A.3), (A.4), (A.7), (A.14) and in Table A.1
Variables/Parameters Kij’ binary interaction coefficient in Eq. (A.15)
M molecular weight in Eqs. (19), (24), and (26)
A Antoine constant in Eq. (6); viscosity model Mref molecular weight of sample in Eq. (27)
coefficient in Eq. (10) MS molecular weight of sample in Eq. (27)
AJ, AL, AM critical temperature model constants in Eqs. (11), m mass in Eqs. (24) and (26)
(14) and (18), respectively mref mass of reference in Eq. (27)
a constant in Eq. (11); constant in vapor pressure mS mass of sample in Eq. (27)
model in Eq. (25); energy parameter in Eqs. (33), N number of atoms in molecule in Eq. (15)
(A.1), (A.7), (A.8), and Table A.1 ni number of occurrences of group i in Eqs.
a1, a2, and a3 constants in dentistry-temperature correlation in (11)–(20)
Eq. (8) npts number of data points in Eq. (34)
ac a constant in energy parameter in Eq. (30) P pressure
ai, and aj a coefficient in reduced effective cross section Pb bubble point pressure
model in Eq. (3); energy parameter for compo- Pc critical pressure
nent i and j in Eqs. (A.1), (A.3), (A.6), (A.7), PJi, PLi, PMi contributions of component i to critical pressure
(A.8), (A.12), (A.14) and Table A.1 in Eqs. (12), (15), and (19), respectively
aij attractive forces between molecules i and j in Psat saturation pressure
Eqs. (A.3), (A.12), (A.14), and Table A.1 R universal gas constant
∼
a Helmholtz free energy in Eqs. (32) and (33) S orifice area in Eq. (24)
∼
a assoc association contribution of Helmholtz free en- S’ surface area to liquid in Eq. (24)
ergy in Eq. (32) T temperature
∼
a disp dispersion contribution of Helmholtz free energy T* reduced temperature in Eqs. (2)–(5)
in Eq. (32) Tb boiling point temperature
∼
a hc hard chain molecules contribution of Helmholtz Tc critical temperature
free energy in Eq. (32) TJi, TLi, TMi contributions of component i to critical
B coefficient of vapor pressure model in Eqs. (6) temperature in Eqs. (12), (14), and (18),
and (23); parameter in density model in Eqs. (7) respectively
and (9) Tr reduced temperature
b coefficient in vapor pressure model in Eq. (25); t time in Eqs. (26) and (27)
volume parameter in Eqs. (28), (31), (A.5), u parameter in PRSV model in Table 6
(A.6), (A.7), (A.9) and in Tables 6, 7, and A.1 V volume
b1 and b2 volume parameter for component i and j in Eqs. Vc critical volume
(A.2), (A.4), (A.5), (A.9), (A.13), (A.15) and VJi, VLi, VMi contributions of component i to critical volume
Table A.1 in Eqs. (13), (16), and (20), respectively
bi and bj volume parameter for component i and j in Eqs. v parameter in PRSV model in Table 6
(A.2), (A.4), (A.5), (A.9), (A.13), (A.15) and V2∞ partial molar volume of solute at infinite dila-
Table A.1 tation in Eqs. (35) and (A.10)
bij binary interaction coefficient of molecules i and j w parameter in PRSV model in Table 6
in Eqs. (A.2), (A.4), (A.6), (A.13), (A.15) and xi and xj mole fraction of component i and j, respectively
Table A.1
C coefficient of vapor pressure model in Eq. (6); Greek Letters
coefficient in density model in Eq. (7); para-
meter in Eq. (A.6), Eq. (A.7), and Table A.1 α vaporization coefficient in Eqs. (24) and (26);
CJ, CL, CM critical pressure model constants in Eqs. (12), energy parameter in Eq. (29) and Table 7
(15), and (19), respectively Δ change or difference
Cp specific heat capacity at constant pressure in Eq. ϕ fugacity coefficient
(25) σμ *(T *) reduced effective cross section in Eqs. (2) and (3)
c volume parameter in Tables 6, 7 δ parameter in five parameter EOSs in Eq. (28)
DJ parameter in critical pressure model in Eq. (15) and Table 7
dij parameter in viscosity-temperature model in Eq. ε particle energy in Eq. (4); activation energy in
(5) Eq. (10); a parameter in five parameter EOS in
EJ, EL, EM critical volume model constants in Eqs. (13), Eq. (28) and Table 7
(16) and (20), respectively η apparent viscosity in Eq. (10); a parameter in
f2 fugacity of component 2 (solute) in Eqs. (35), five parameter EOS in Eq. (28) and Table 7
(36), (A.10), and (A.11) κ conductivity in Table 4
GE Gibbs excess free energy in Eq. (A.6), Eq. (A.7), ρ density
and Table A.1 τ elapsed time in Eq. (24)
H2 enthalpy of component 2 (solute) in Eqs. (35) ω acentric factor

504
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Subscripts and Superscripts E excess

exp experiment
0 reference state, and dilute hc hard chain
* reduced LV liquid-vapor
∞ infinite dilution m melting
assoc association r reduced
b boiling ref reference
c critical res residual
calc calculation s sample
disp dispersion sat saturation

However, in the post-combustion method, further stages/process units

are required to capture the carbon from the flue-gas [6].
The pre-combustion method is associated with the integrated gasifica-
tion combined cycle (IGCC). In the pre-combustion, the fuel is gasified to
produce syngas [7]. The syngas, which contains CO and H2, is utilized as a
feed for the water-gas-shift reactors. In the presence of steam, CO is con-
verted to CO2 and H2. CO2 is then captured from the outlet stream and H2 is
combusted to generate the energy. Post-combustion is easier to be im-
plemented as a retrofit option in the existing power plants, compared to
other two approaches [6]. In other words, pre-combustion and oxy-com-
bustion methods can only be applied in new power plants [8].
As illustrated in Fig. 1, the post-combustion carbon capture tech-
nique is divided into several main subcategories such as absorption,
microbial (algal), adsorption, membrane, and cryogenics. Among these
techniques, absorption is more popular and is widely employed in the
commercial post-combustion CO2 capture plants. In this method, the
flue-gas in the tall columns or towers (known as absorbers) is in contact
with the solvent, which captures CO2 from the flue-gas. The rate of CO2
absorption with solvent is a crucial parameter in the chemical absorp- Fig. 2. Schematic figure of absorption/desorption columns in a CO2 capture
tion processes. High absorption rate not only can reduce the capital process (modified after [12]).
costs of CO2 capturing, but is also considered as an asset for an ab-
sorption process to be operated at the industrial scale [9,10]. Fig. 2 also Table 1
illustrates a simple schematic of CO2 absorption and desorption col- General properties of ionic liquids [19].
umns in a typical carbon capture plant. Property General characters
Various solvents in the chemical absorption of CO2 capture pro-
cesses have been introduced [13]. Amines due to their high reactivity Salt ions large cations and anions
Freezing temperature < 100 °C
with CO2 molecules, high thermal stability, and high absorption capa-
Liquidous temperature > 200 °C
city are widely used as the absorbent for carbon capture operations Thermal stability high
[14]. Among all amine groups, monoethanolamine (MEA), tertiary Viscosity < 100 cP, workable
amines such as methyl diethanolamine (MDEA), and sterically hindered Dielectric constant < 30
amines such as 2-amino-2-methl-1-propanol (AMP) are the common Polarity moderate
Specific conductivity < 10 mS/cm, good
absorbents for CO2 capture systems [15]. Application of amines faces
Vapor pressure negligible
several drawbacks; including, high construction cost, high equipment Solvency strong
corrosion rate, and amine degradation by SO2, NO2, and O2 in the flue- Catalytical character excellent (for organic reactions)
gases which induces a high absorbent make-up rate and high energy

Fig. 1. Diagram of main carbon separation/capture methods in the post-combustion category (modified after [11]). [MOFs: metal-organic frameworks; PDMS:
polydimethylsiloxane; PPO: polyphenyleneoxide; PP: polypropylene].

505
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

consumption during high-temperature absorbent regeneration pro- results and experimental data at pressures 0–100 MPa. Shiftlet et al.
cesses where large equipment are required [16]. These disadvantages [31,32] utilized the most popular EOSs including Redlich–Kwong (RK),
are linked with the inherent properties of amines, which have high Soave-Redlich-Kwong (SRK), van der Waals (vdW), and Peng-Robinson
vapor pressures, corrosive nature, and high energy inputs for re- (PR) with proper mixing rules for multi-component mixtures. They
generation [17]. Several research studies have been conducted to re- concluded that the experimental solubility data of CO2, SO2, CF3-CFH2,
place amines with superior solvents, which have lower volatility, better and NH3 in ILs are well correlated with the cubic (modified) EOSs.
thermal stability, lower corrosive characteristic, lower degradation Andrew et al. [33] examined Soft-SAFT EOS ability to estimate the
rate, and lower regeneration cost [12–15]. Recently, ionic liquids (ILs) solubility of CO2 in [Cn-mim][NTf2] using one binary parameter, re-
have been proposed as a potential replacement solvent for amines in the sulting in a good match with the experimental data. Chen et al. [6]
carbon capture processes [18]. The general properties of a majority of checked the competence of PC-SAFT with and without associating
ILs are listed in Table 1 [19]. parameters in determining the solubility of CO2 in imidazolium based
Strong ion-ion interactions for ILs (compared to other common in- ILs where the temperature-dependent binary interaction parameters
termolecular forces such as London forces and ion-dipole interactions in were taken into account. The results revealed that PC-SAFT is able to
organic solvents) lead to negligible evaporation at the ambient condi- acceptably forecast the solubility so that the modeling results were very
tions [20]. Other appealing properties of ILs are their high thermal close to the real data. It was also found that PC-SAFT with the asso-
stability, large electrochemical window, and their ability to dissolve ciating parameters describes VLE more precise, compared to that
compounds with various polarities [21]. Furthermore, physical and without the associating parameters.
chemical properties of ILs can be adjusted by altering the cation and The solubility of gases in ILs can be increased by extending the in-
anion parts to make them suitable for particular applications [22]. terface area of gas/ILs systems and by increasing the operating pres-
From the environmental point of view, ILs are inflammable, non-vola- sure. Maginn et al. [18] examined the solubility of nine different gases
tile, and recyclable which make them environmentally friendly, com- into [bmim][PF6] IL and found out that water and CO2 have strong
pared to other chemicals such as amines [23,24]. interactions. CO2 in ILs exhibits higher solubility, compared to solubi-
In a research work, Blanchard et al. [25] reported that CO2 gas can lity of ethylene, ethane, methane, carbon monoxide, hydrogen, and
be considerably dissolved into the certain ILs. Since then, there is a nitrogen in ILs. Argon and oxygen have also very low solubility in ILs.
growing interest in exploring and understanding the solubility of var- They also reported Henry's law constants of gases in ILs. The gas so-
ious gases in ILs [13,22,23]. lubility can be explained based on the molecules interactions in terms of
A drawback to use ILs for CO2 capture is their high viscosity due to polarizability and dipole and quadrupole moments between the gas and
their complex synthesis and purification processes. ILs have higher solvent molecules. It should be noted that the mixing enthalpy and
viscosity, compared to conventional solvents [26]. This character entropy can also indicate the strength of interactions between CO2 and
makes them less suitable for industrial CO2 capture since the rate, at ILs. Yokozeki et al. [34] successfully correlated the experimental solu-
which the CO2 is absorbed by ILs, is much slower. Solvents that have a bility of CO2 in 18 ILs to fluids properties and operating conditions
lower viscosity tend to be more favorable for this separation purpose as through using Redlich-Kwong equation of state (RK-EOS). In addition,
the pumping costs are lower and the mass transfer rates are normally they presented the data in terms of excess functions (excess Gibbs free
higher [27]. According to Krupiczka et al. [28], the viscosity of ILs can energy, GE; excess enthalpy, HE; and excess entropy, SE). Depending on
be modified by using a proper combination of cations and anions. The the magnitudes of these excess functions, the CO2 absorption might be a
cation has generally a lower impact on ILs properties than the anion. chemical or physical process.
With increasing the alkyl chain of cations, the viscosity of ILs increases. The thermodynamic behaviors of CO2/ILs systems such as physical
The effect of anion on viscosity in imidazolium based ILs is reported as and chemical properties, phase equilibrium of CO2/ILs mixtures, and
[bmim][NTf2] < [bmim][CF3SO3] < [bmim][BF4] < [bmim][PF6]. solubility and selectivity of CO2 in ILs should be properly investigated.
The viscosity of ILs also changes by adding a co-solvent. The effect of A suitable equation of state (EOS) and mixing rules are required to
co-solvent in decreasing viscosity depends on the solvent polarity and accurately model the thermodynamic behaviors of CO2 and IL systems.
the extent of dissociation of ILs into ions. Polar solvents such as water A part of the common EOSs for research investigations in this area are
lower the viscosity more than nonpolar solvents; including, benzene PR EOS, RK EOS [35], SRK EOS [36], Predictive Peng-Robinson EOS,
and toluene. For instance, the role of water in reducing the viscosity of Peng-Robinson-Stryjek-Vera EOS, Sanchez-Lacombe EOS [37], PC-SAFT
hydrophobic ILs is prominent. EOS, extended PC-SAFT (PC-PSAFT) EOS, and truncated PC-PSAFT EOS
Conventional ILs capture the carbon through a physical absorption [38]. The solubility of CO2 in most of ILs is not determined yet, since a
mechanism. Although adjusting cations/anions in the ILs can enhance majority of the previous studies have assessed the solubility of CO2 in
the solubility of CO2 in conventional ILs, compared to the current common imidazolium based ILs [5,14,30]. Thus, modified versions of
commercial carbon capture technologies such as amine-based solvents, various EOSs need to be utilized to obtain the solubility of CO2 in
their CO2 capture capacity is still low [29]. A comprehensive under- phosphonium, pyridinium, ammonium, and guanidinium based ILs.
standing of molecular structure/interactions and accurate determina- In this review paper, CO2 capture using ILs is briefly described in
tion of thermodynamic and physical characters such as solubility, terms of mechanisms, advantages, and disadvantages. The past research
vapor-liquid equilibria (VLE), liquid-liquid equilibria (LLE), selectivity, works in this research field are discussed and their main findings are
density, and viscosity are necessary to employ appropriate ILs in dif- reported. Comprehensive information on the properties and thermo-
ferent applications/cases. To use ILs in gas separation processes, it is dynamic behaviors of CO2, ILs, and CO2 / ILs systems is also provided.
required to understand the phase behavior of ILs/gas systems. A well- The last sections summarize the practical, economic, and theoretical
developed methodology to obtain phase equilibrium conditions is aspects of CO2 capture with ILs.
thermodynamic modeling using appropriate equations of state (EOSs).
A large number of research works have focused on the solubility of 2. CO2 emissions: sources and statistical data
gases in common imidazolium-based ILs with [BF4]- and [PF6]- anions
[26–28]. Karakastani et al. [30] predicted the solubility of CO2 in some Global greenhouse gas (GHG) emissions are mainly due to human
ILs through employing perturbed chain-statistical associating fluid activities, including fossil fuel burning, deforestation, and industrial
theory (PC-SAFT) EOS where they considered the quadrupole interac- processes. Among different sources of GHG emissions, the energy sector
tions between CO2 molecules, dipolar interactions between ionic liquid has the largest contribution [39,40]. The other shares come from
(IL) molecules, and the Lewis acid-base type of association between the agriculture, industrial processes, and solid biomass burning for gen-
IL and the CO2. They found a great agreement between the modeling eration of heat and energy. Due to economic growth and population

506
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Fig. 5. Regional distribution of 2014 global CO2 emissions from fossil fuels
combustion and industrial processes (modified after [45]).

depicted in Fig. 3(a) and Fig. 3(b), respectively [40]. Electricity and
heat generation, and transportation together are two major sectors,
producing about two-third of the global CO2 emissions from fuel
combustion. A significant extent of the GHG emissions in heat and
electricity generation is from coal, while the road transportation is re-
sponsible for most of the CO2 emissions in transportation sector.
In the atmosphere, the average volume concentration of CO2 is
about 385 ppm (or, 582 ppm by mass); the total mass of CO2 in the
atmosphere is therefore about 3.0 × 1015 kg [36,37]. The concentra-
tion of CO2 varies by season, geographical region, and human-induced
activities that produce carbon [41]. Fig. 4 demonstrates the CO2
emissions by sectors in the period of 1971–2014 [40,42]. According to
Fig. 4, the global carbon emissions from fossil fuels have been increased
by about 90% since 1971. The carbon emissions from all sectors have
exhibited a considerable increase from 1971 to 2014. In 2016, the
annual rate of CO2 emissions increased from 3.2 ppm at the beginning
Fig. 3. Schematic plot of: (a) shares of global anthropogenic GHG emissions and
(b) world CO2 emissions from fuel combustion by sector in 2014 (modified after
of 2016 to 3.6 ppm towards the end [43].
[40]). Total carbon dioxide emitted by region in 2014 is illustrated in
Fig. 5. China has the highest contribution to the CO2 emissions. It is
clear that increasing the atmospheric carbon dioxide (CO2) concentra-
increase, the energy demand is increasing. Global total primary energy
tion will lead to an increase in CO2 concentration in the ocean surface,
supply has increased by 150% from 1971 to 2014 [39]. CO2 has the
because of the gas exchange between the air and oceans. Hence, the
largest share of the anthropogenic GHG emissions. The growing world
dissolved inorganic carbon (e.g., CO2, HCO3-, and CO32-) will be in-
energy demand is mainly responsible for increasing the CO2 emissions.
creased and subsequently the pH of the ocean surface will decrease,
In 2014, global CO2 emissions reached 32.4 Gt CO2, of which 82%
resulting in the ocean acidification [44].
comes from energy sector, mainly due to burning of fossil fuels. Among
the fossil fuels, 34% is from oil, 46% from coal, 19% from gas fuels, and
the remaining 1% has been resulted from nuclear, hydro, geothermal, 3. CO2 properties
solar, wind, biofuels, and waste [39]. Global GHG emissions from
various sectors and world CO2 emissions from fuel combustion are Carbon dioxide (CO2) is an odorless and colorless gas at the atmo-
spheric condition. This chemical component is relatively nontoxic,

Fig. 4. Global CO2 emissions by sector in the period 1971–2014 (modified after [40,42]).

507
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

noncombustible, and soluble in water. Its molecular weight is 44.009 g/ and 31.1 °C [51,52].
mol. Normal melting point of CO2 is −56.5 °C and its solubility in water To reasonably forecast the thermodynamic behavior of any com-
is 2.9 g/L at 25 °C [46]. CO2 is also miscible with hydrocarbons and ponent or mixture through using equations of state (EOSs), the critical
organic liquids at particular thermodynamic conditions [57,58]. The properties and acentric factor of the component/mixture are required.
density of gas CO2 is 1.976 g/l at 25 °C. Its vapor pressure and viscosity The critical properties and acentric factor of CO2 are listed in Table 3
are 5720 kPa (56.5 atm) at 20 °C and 0.01503 cP at the atmospheric [52].
condition, respectively [47]. The surface tension of CO2 is 0.0162 N/m
at the melting point [58]. 4. Ionic liquids (ILs) characteristics

3.1. Viscosity Based on the molecular structure, ILs are classified into proton-do-
nating (protic ILs (PILs)) and nonproton-donating (aprotic ILs (AILs))
The viscosity of CO2 in other conditions can be calculated using Eq. solvents. A general structure of cations and anions in ILs molecular
(1), where it expresses μ (ρ , T ) as a summation of three terms. The first configuration is shown in Fig. 7.
term μ0 (T ) is the dilute gas viscosity at a given temperature, Δμ (ρ , T ) is In Table 4, the characteristics of ILs are listed and compared to the
the excess viscosity which signifies the effect of density (or pressure), atomic, molecular, and molten salts. As clear from Table 4, ILs due to
and Δμc (ρ , T ) is the critical enhancement due to fluctuations near the their various structures and ions have different solvent properties,
critical point [48]. Eq. (1) represents the above description in the compared to the atomic salts; however, they are similar to the polar
mathematical form: protic liquids and molten salts in terms of physical characteristics.
μ (ρ , T ) = μ0 (T ) + Δμ (ρ , T ) + Δμc (ρ , T ) (1)
4.1. Density
Fenghour et al. [49] proposed the following correlation to calculate
the viscosity of carbon dioxide in dilute regions: ILs are denser than water, except pyrrolidinium dicyano-diamide
and guanidinium with a density ranging from 0.90 g/cm3 to 0.97 g/
1.00697T1/2
μ0 (T ) = cm3. As shown in Fig. 8, the density of ILs declines while the number of
σμ *(T *) (2) carbons in the alkyl chain increases. Furthermore, the density of 1-
where μ0 (T ) is the zero-density viscosity in μPa. s , T is the temperature methylimidazolium ILs decreases linearly with increasing temperature.
in Kelvin, and σμ *(T *) denotes the reduced effective cross section, which An increase in the viscosity value of ILs is noticed with increasing the
is represented by the following empirical equation: number of carbons in the alkyl group, which exhibits a different trend,
4
compared to the viscosity behavior for conventional organic solvents.
ln σμ* (T *) = ∑ ai (ln T *)i High viscosity of ILs is attributed to the van der Waals interactions and
i=0 (3) hydrogen bonding. Generally, the molecular interactions increase with
alkyl lengthening or/and fluorination.
The reduced temperature (T*) can be calculated using the following
Luciana Tome et al. [54] showed that experimental density data of
expression:
ILs are in good agreement with the Tait equation [55] for density of
kT liquids which is given below.
T* =
ε (4)
ρ (T , P = 0.1MPa)
ρ=
where ε is the particle energy, k stands for the Boltzmann constant, and (B + P )
ε ⎡1 − C ln ⎤
T represents the absolute temperature ( k = 251.196 K ). ai is a coefficient ⎣ (B + 0.1) ⎦ (7)
in Eq. (3) which is listed in Table 2.
ρ (T , P = 0.1MPa) = a1 + a2 T + a3 T 2 (8)
The excess viscosity describes the viscosity behavior change with
the density outside the critical region as given by the following ex- b2
pression: B = b1 +
T (9)
n m
dij where a1, a2, and a3 can be fitted to the experimental density data for
Δμ (ρ , T ) = ∑ bi (T ) ρi where bi = ∑
i=1 j=1
T *(j − 1) (5) different ILs versus temperature. C, b1, and b2 are also obtained if the
density is plotted against temperature and pressure, using Eqs. (7)–(9).
where T* stands for the reduced temperature as defined by Eq. (4) and
dij denotes a coefficient, which is obtained through curve fitting of the 4.2. Viscosity
excess viscosity experimental data versus the reduced temperature.
The empirical equation for the apparent dynamic viscosity (η in cP)
3.2. Vapor pressure of ILs to describe its temperature dependency is as follows [56]:

The vapor pressure (Psat) of CO2 can be calculated using the Antoine η = A e ε / RT (10)
equation as follows [50]: where A represents a constant for a given liquid, ε is the activation
B energy for viscous flow of liquid, T is the temperature in Kelvin, and R
log P sat (mmHg) =A−
C + T ( oC) (6)
Table 2
where A, B, and C are the constants (For CO2: A = 9.8206, Coefficients used in Eq. (3) to obtain viscosity
B = 1347.790, and C =273.00) [50]. of CO2 at low density [49].
i ai
3.3. Thermodynamic behavior and critical properties
0 0.235156
Fig. 6 presents the pressure-temperature diagram of CO2. As de- 1 − 0.491266
2 5.211155 × 10-2
picted in Fig. 6, the triple point represents the temperature and pressure
3 5.347906 × 10-2
(− 56.7 °C and 5.11 atm) that three phases can coexist at the thermo- 4 -1.537102 × 10-2
dynamic equilibrium. The critical point of CO2 corresponds to 73.8 atm

508
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

researchers to determine the critical properties of ILs. For instance,

Valderrama and Robles [58] employed an extended group contribution
method based on the concepts developed by Lydersen [59] and Joback
and Reid [60] to obtain the critical properties, normal boiling points,
and acentric factors of ILs. In this method, only structure of the ILs and
their molecular weights are required [61]. It is accurate enough to be
used in the generalized correlations and methods based on equations of
state. The absolute average deviation (AAD) of calculated and experi-
mental liquid densities is 5–6% [61]. The critical properties and acen-
tric factors for some common ILs are tabulated in Table 5.
The group contribution method for estimation of critical properties
was developed by Lydersen [59]. In this approach, the critical proper-
ties of a compound are calculated based on the contribution of certain
Fig. 6. CO2 phase diagram (reprinted with permission from [51]). groups of atoms and the number of frequency of group of atoms in a
molecule, through using the following expressions:
Table 3 Tb
Critical properties of CO2 [52]. Tc =
AL + ∑ ni ΔTLi − (∑ ni ΔTLi )2 (11)
Critical parameter Designation/symbol Value
M
Pc =
Temperature Tc (K) 304.2 [CL + ∑ ni ΔPLi]2 (12)
Pressure Pc (bar) 73.8
Volume Vc (cm3 mol−1) 94.0 Vc = EL + ∑ ni ΔVLi (13)
Acentric factor ω 0.2236
in which, ni stands for the number of times that a group appears in the
molecule, Tb represents the normal boiling temperature, ΔTLi denotes
the contribution of group i to the critical temperature, ΔPLi symbolizes
denotes the universal gas constant in J/(mol K). Temperature and the contribution of group i to the critical pressure, ΔVLi is the con-
chemical additives strongly affect the viscosity of ILs. For instance, tribution of group i to the critical volume, M is the molecular weight,
increasing temperature or/and adding organic solvents lead to a re- and AL, CL, and EL are the constants equal to 0.567, 0.34, and 40, re-
duction in the viscosity. spectively [59]. Eqs. (14)–(16) were also proposed by Joback and Reid
The surface tension data of ILs is rare in the literature. It has been [60] to determine the critical properties:
reported that ILs have a higher surface tension than conventional sol- Tb
Tc =
vents (e.g., hexane with a surface tension of 1.8 Pa.cm); but, not as AJ + ∑ ni ΔTJi − (∑ ni ΔTJi )2 (14)
much as water (7.3 Pa.cm) [57]. For instance, the surface tensions of
most imidazolium based ILs at 298.15 K are in the range of 2.5 (Pa.cm) 1
Pc =
to 3.5 (Pa.cm). [CJ + DJ N − ∑ ni ΔPJi]2 (15)

Vc = EJ + ∑ ni ΔVJi (16)
4.3. Critical properties
where N is the number of atoms in the molecule and AJ, BJ, CJ, DJ, and
The critical properties of ILs are not readily available and should be EL represent the constants which are equal to 0.584, 0.965, 0.113,
estimated due to the decomposition phenomenon at temperatures near 0.0032, and 17.5, respectively [60].
the boiling point. There are certain techniques/methods introduced by The combination of Lydersen's method, Joback Reid method, and

Fig. 7. Structures of cations and anions of most common ionic liquids [53].

509
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 4
A comparison between the physical properties of different solvents [53].
Parametera Solvent type

Atomic Molecular Molten salt Ionic liquid

Hg H2O NaCl PIL AIL

[CH3CH2NH3][NO3] [C4mim][PF6]

Tm (°C) -38.8 0 801 12 10

Tb (°C) 356.7 100 1413 255 409
ρ (g/cm3) 13.53 0.997 1.539 1.21 1.366
μ (Pa.s) 1.526×10−3 8.95×10−4 12.5×10−4 35.9×10−4 36.9×10−4
Psat (Pa) 2.67 3173 12700 0.49 < 10-2
D×10−6 (cm2/s) 85 22.99 Na+: 80.1 [CH3CH2NH3+]: 0.158 [NO3−]: 0.151 [C4mim+]: 1.5
Cl−: 63.5 [PF6−]: 1.8
γLV (mN/m) 486.5 72.8 111.7 47.3 43.8
κ (S/cm) 105 5.5×10−4 0.256 2.69×10−2 1.4×10−3
Cp (J/mol.K) 27.98 75.3 66.9 206 406

a
Tm: melting temperature; Tb: boiling temperature; ρ: density; μ: viscosity; Psat: vapor pressure; D: diffusion coefficient; γLV: interfacial tension of liquid-vapor; κ:
conductivity; and Cp: specific heat capacity at constant pressure.

Vc = EM + ∑ ni ΔVMi (20)

The constant parameters in Eqs. (18)–(20) are AM = 0.5703,

BM = 1.0121, CM = 0.2573, and EM = 6.75 [58].
Valderrama et al. [58] extended the “modified Lydersen-Joback-
Reid method” to ILs. Following this strategy, the acentric factor is
calculated using Eq. (21):

(Tb − 43)(Tc − 43) P (T − 43) P P

ω= log ⎛ c ⎞ − c
⎜ ⎟ log ⎛ c ⎞ + log ⎛ c ⎞
⎜ ⎟ ⎜ ⎟

(Tc − Tb)(0.7Tc − 43) ⎝ Pb ⎠ (Tc − Tb) ⎝ Pb ⎠ ⎝ Pb ⎠

−1 (21)

Valderrama et al. [64] used the group contribution method with 44

defined groups and prepared the most complete database for the critical
properties of ILs.

Fig. 8. Density variation of ILs with number of carbons [21]. 4.4. Vapor pressure

the technique proposed by Alvarez and Valderrama [63] was resulted in The vapor pressure of chemical components is required in many
a proper strategy, called “modified Lydersen-Joback-Reid” method, as applications such as separation, solvent selection, and design and op-
presented by the following equations: eration of process equipment. According to the literature, ILs have a
negligible vapor pressure. There are a few models to predict the vapor
Tb = 198.2 + ∑ ni ΔTbMi (17) pressure of ILs.
The vapor pressure of ILs might be estimated through the equili-
Tb brium condition concept, implying the same fugacity of each compo-
Tc =
AM + ∑ ni ΔTMi − (∑ ni ΔTMi )2 (18) nent in both vapor and liquid phases [65].
The vapor pressure of ILs can also be calculated using the Antoine
M
Pc = equation (logPSat = A – B/(T + C)) with water as a reference (C = 43).
[CM + ∑ ni ΔPMi]2 (19) Valderrama and Sanga [62] suggested Eqs. (22) and (23) to determine

Table 5
The critical properties of common imidazolium based ionic liquids [58,62].
IL IUPAC Namea Tc (K) Pc (bar) Vc (cm3/mol) ω

[bmim][PF6] 1-n-butyl-3-methylimidazolium hexafluorophosphate 708.9 17.3 779.5 0.7553

[hmim][PF6] 1-hexyl-3-methylimidazolium hexafluorophosphate 754.3 15.5 893.7 0.8352
[bmim][BF4] 1-butyl-3-methylimidazolium tetrafluoroborate 623.3 20.4 672.0 0.8489
[hmim][BF4] 1-hexyl-3-methylimidazolium tetrafluoroborat 679.1 17.9 786.2 0.9258
[N-bupy][BF4] N-butylpyridinium tetrafluoroborate 597.6 20.3 648.1 0.8307
[omim][Cl] 1-octyl-3-methylimidazolium chloride 860.1 20.3 814.2 0.6190
[emim][EtSO4] 1-ethyl-3-methylimidazolium ethyl sulfate 1061.1 40.4 676.8 0.3368
[bmim][TfO] 1-butyl-3-methylimidazolium trifluoromethanesulfonate 1016.3 29.4 767.6 0.3677
[emim][TfO] 1-ethyl-3-methylimidazolium trifluoromethanesulfonate 985.2 35.8 653.4 0.2891
[bmim] [Ntf2] 1-butyl-3-methylimidazolium trifluoromethanesulfonate 1265.0 27.6 1007.1 0.2656
[hmim] [Ntf2] 1-hexyl-3-methylimidazolium trifluoromethanesulfonate 1287.3 23.9 1121.3 0.3539
[bmim] [Br] 1-butyl-3-methylimidazolium bromide 834.9 29.8 583.3 0.4891
[bmim] [dca] 1-butyl-3-methylimidazolium dicyanamid 1035.8 24.4 712.0 0.8419

a
IUPAC is an acronym for International Union of Pure and Applied Chemistry.

510
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

the two constants of A and B for twenty ILs, respectively: sample, respectively. They concluded that the vapor pressures obtained
through evaporation data are reliable where the TGA method is em-
A = log [Pc (Tc − 43)(Tc − Tb)] (22)
ployed [71].
(T −43) ⎤ In general, physical and thermodynamic properties of ILs including heat
B = log ⎡Pc (Tc−43) b capacity, viscosity, density, solubility, excess enthalpies, and vapor-liquid
⎢
⎣ (Tc−Tb) ⎥
⎦ (23)
equilibrium data of the water/CO2/ILs mixtures are necessary to evaluate
The values for vapor pressure are accurate based on the generalized ILs for CO2 capture and to design the absorption processes [72].
correlations and EOSs [66].
Knudsen strategy introduces another method to determine the vapor
pressure of ILs. Knudson equation for calculation of ILs vapor pressure 5. Characterization of CO2/IL systems
is as follows [67]:
5.1. Experimental methods
kS ⎞ Δm 2πRT
P sat = ⎛1 +
⎝ S′α ⎠ kSτ M (24) It is vital to investigate the thermodynamic properties of the sub-
where Δm stands for the mass loss during an elapsed time τ, S refers to stances which are utilized in the CO2 capture processes so that the
the orifice area, k is the transmission coefficient, M signifies the mo- appropriate equipment are chosen to attain the ultimate separation
lecular weight of a vapor, S′ denotes the surface area of a liquid which is goal. There are several experimental methods developed by researchers
assumed to be equal to the cross-section area of a cell, and α represents and research centers to determine the thermodynamic properties of
the vaporization coefficient. various ILs, CO2, and their mixtures.
Vladimir et al. [68] measured the vapor pressure using the tran- High pressure phase equilibria experiments are traditionally classi-
spiration method and compared them with the values obtained from the fied as analytical systems when the composition is identified. They are
Knudsen technique. They also introduced a relationship to calculate the referred to the synthetic systems if only the overall composition is
saturation (vapor) pressure (Psat) at different temperatures. The fol- known. Sub-classifications are based on the mobility of the sample. In
lowing equation is utilized to obtain the vapor pressure [78]: the dynamic systems, one or more phases are recirculated outside the
equilibrium cell. However, in the static methods, the sample is agitated
b T in a closed cell. The volume of the cell can be constant or variable. A
R ln P sat = a + + Δlg Cp ln ⎛ ⎞
⎜ ⎟

T ⎝ T0 ⎠ (25) summary of common techniques for obtaining thermodynamic phase

equilibria parameters is depicted in Fig. 9.
In Eq. (25), T0 is the reference temperature (298.15 K) and Δlg Cp
The common methods to determine the CO2 solubility in an IL
represents the difference between the heat capacities of gas and liquid
without sampling are the gravimetric analysis, pressure drop method,
phases [68].
and gas chromatography. Gravimetric analysis is an analytical method
Thermogravimetric analysis (TGA) is also a promising method for
which describes the quantitative determination of gas solubility
rapid determination of vapor pressure. Eq. (26) proposed by Langmuir
through measuring the overall weight variation of a sample over the
can be used to calculate the vapor pressure based on the evaporation
absorption process [17]. The schematic of the basic components of a
rate [69]:
gravimetric system is demonstrated in Fig. 10(a). The pressure drop
dm M method is a synthetic technique. In this methodology, the system vo-
=α P sat
dt 3πRT (26) lume and temperature are held constant and the pressure difference is
recorded during gas absorption [74]. A schematic of a pressure drop
where dm/dt is the evaporation rate, α is the evaporation coefficient, M equipment is shown in Fig. 10(b). Since ILs have a negligible vapor
is the molecular weight of vapor, and T and Psat denote the temperature pressure, the changes in pressure are assumed to be due to the gas
and vapor pressure, respectively. α depends on experimental condi- absorption. Thus, the gas absorbed by the IL can be determined. Solu-
tions. Phang et al. [70] suggested a reference substance similar to the bility data from a gas chromatography equipment can be achieved
experimental substance in terms of boiling and melting points and through employing a systematic procedure. The solvent (IL) is saturated
molecular structure so that α can be assumed constant. Thus, the vapor with the solute (CO2) and then coated on a column. A non-absorbing
pressure of a substance S (as a sample) can be expressed by the fol- carrier gas is directed into the column to extract CO2 from the IL. The
lowing equation: carrier gas is then analyzed in the gas chromatograph. This procedure
determines the amount of absorbed CO2 [74]. The pressure drop and
PSsat = ⎛
⎜
Mref ⎞
⎟
0.5 ( )
dm
dt S
synthetic methods are simpler than the gravimetric analysis and their
⎝ MS ⎠ ( )
dm
dt ref (27)
results have higher sensitivity; but, their precision is lower [74].
According to Blanchard, Gu, and Brennecke [75], a pycnometer
Mref and Ms represent the molecular weight of the reference and with a known volume can be used to determine the density of each pure

Fig. 9. Experimental methods to obtain phase equilibria data (modified after [73]).

511
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Fig. 10. : Schematics of: (a) a simple gravimetric system, including basic components, and (b) diagram for pressure drop method [P and T correspond to the pressure
and temperature sensors, correspondingly] (modified after [17]).

Table 6 Table 7
Summary of cubic EOSs widely used for VLE calculations [78]. Parameters of EOSs [87].
EOS Name EOS or P = P(V,T) Reference EOS General EOS parameters

vdW van der Waals (1837) P=

RT
−
a [78,79] δ ε θ η
V−b V2
PR Peng-Robinson (1976) P=
RT
−
a [79] van der Waals (1890) 0 0 a b
V−b V (V + b) + b (V − b)
PT Patel-Teja (1982) RT a (T ) [80] Redlich and Kwong (1949) 0 0 a/(Tr0.5) b
P= − Soave (1972) b 0 aα(T) b
V−b V (V + b) + c (V − b)
RK Redlich-Kwong (1942) RT a / T 0.5 [81] Peng and Robinson (1976) 2b -b2 aα(T) b
P= −
V−b V (V + b) Patel and Teja (1982) b+c -bc aα(T) b
SRK Soave-Redlich-Kwong P=
RT
−
a [82] Stryjek and Vera (1986) 2b -b2 aα(T) b
V+c−b (V + c )(V + b + c )
(1972)
PRSV Peng–Robinson-Stryjek-Vera P=
RT
−
a (T ) [83]
(1986) V−b (V + ub)(v + wb)
results obtained from EOSs are well matched with experimental data.
Ren et al. [84] used the PR EOS with the van der Waals 2-parameter
mixing rule to correlate VLE data for the CO2-IL systems. A great
IL. The pycnometer is immersed in a constant-temperature bath and the
agreement was attained between the experimental data and modeling
amount of IL in the pycnometer is gravimetrically obtained where a
outputs at all isotherms and pressures. However, Freitas et al. [117]
thermal equilibrium is maintained [76]. The density of ILs can also be
claimed that the combination of the van der Waals 2-parameter mixing
measured using a density measuring module such as DMA 4500
rule and the PR EOS does not lead to acceptable VLE results for some
(manufacturer: Anton Paar GmbH). Another method for determination
multi-component systems. Using a more complicated mixing rule such
of IL density is with the aid of the proven oscillating U-tube principle
as the Wong-Sandler (WS) mixing rule was proposed for a number of
that leads to a high accuracy [77].
multi-component mixtures. Based on Maia et al. [85], the PR EOS
combined with the van der Waals and Wong-Sandler mixing rules is
5.2. Equations of state accurate at low pressures to model the VLE of binary systems; however,
it does not offer reliable results at high pressures.
It is important to study the vapor-liquid equilibrium (VLE) for the Another EOS, which is often addressed in the literature, is the
CO2 capture processes to design and operate the separation stages [52]. Soave-Redlich-Kwong (SRK) EOS. According to Maia et al. [86], the PR
One of the most common approaches to correlate and predict phase EOS and SRK EOS are able to model the VLE data for most ternary
equilibrium is using equations of state (EOSs). According to Li [78], an systems with a total standard deviation of less than 0.03 and 0.04, re-
EOS with a simple structure and reasonable accuracy is preferable, spectively.
suggesting a cubic EOS is well suited to model phase equilibrium of The general form of five-parameter EOSs is given below [87]:
multi-component systems. EOSs have been used to obtain the physical
RT θ (V − η)
and thermodynamic characteristics of pure ILs and mixtures containing P= +
V−b (V − b) (V 2 + δV + ε ) (28)
water, gases, and amines [61]. There are various EOSs; however, the
most common EOSs are the cubic equations derived from van der Waals The parameters ε, θ , δ , and b might be constant or zero for a single
EOS. Among them, it has been proven that the Peng-Robinson (PR) EOS component. They may depend on the composition of the mixture. These
combines the simplicity and accuracy required for thermodynamic parameters are listed in Table 7.
equilibrium calculations [52]. The well-known cubic EOSs are listed in In all EOSs listed in Tables 6 and 7, parameter b is a positive con-
Table 6. Li [78] reported that there are four cubic EOSs which are stant and equal to η. Parameters a and b are dependent on the critical
widely used in oil and gas industries, including: Peng-Robinson; Patel- properties of the component as expressed by the following equations:
Teja, Redlich-Kwong; and Soave-Redlich-Kwong.
The Peng-Robinson (PR) EOS can be used with different mixing a = ac α (Tr , ω) (29)
rules to model the VLE of mixtures so that it can accurately predict
C1 RTc2
important parameters such as gas and liquid volumes and vapor pres- ac =
sure [78]. In many chemical and petroleum engineering cases, the Pc (30)

512
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 8
Temperature dependent parameter (α (Tr , ω) ) for different EOSs [87].
Equation of state α(Tr, ω)

van der Waals (1890) 1

Redlich and Kwong (1949) Tr−1/2
Soave (1972) [1 + (0.48 + 1.574ω − 0.176ω2)(1 − Tr1/2)]
2

Peng and Robinson (1976) [1 + (0.37646 + 1.5422ω − 0.2699ω2)(1 − Tr1/2)]

2

Patel and Teja (1982) 2 2

(1 + [1 + (0.452413 + 1.38092ω − 0.295937ω2)(1 − Tr1/2)] (1 − Tr1/2))
Stryjek and Vera (1986) [1 + (0.37889 + 1.48971ω − 0.0.171318ω2 + 0.019655ω3)
2
(1 − Tr1/2)] + k1 (1 − Tr )(0.7 − Tr )

C2 R Tc In the literature, different models for obtaining VLE data with the
b=
Pc (31) aid of experimental data have been proposed. Generally, the differential
evolution (DE) and differential evolution with Tabu List (DETL) show a
where the parameter α is a function of temperature and acentric factor.
better performance than simulated annealing (SA), genetic algorithm
C1 and C2 are the constants, depending on the type of EOS. C1 is
(GA), and particle swarm optimization (PSO) [95]. The DE approach is
0.0778 for Peng-Robinson (PR) [79] and 0.0833 for Soave-Redlich-
based on the difference between two individuals and adding it to an-
Kwong (SRK) [88]. C2 is 0.4572 for PR and 0.4218 for SRK. α (Tr , ω) is a
other individual and generating a new individual.
temperature dependent parameter expressed in terms of the reduced
The pressure-temperature (P-T) diagram is depicted in Fig. 11 for
temperature (Tr ), as tabulated in Table 8.
the CO2/[omim][BF4] system, based on the experimental data. The
Recently, some molecular-based models have been introduced
measured P-T data (isopleths) indicate the values at which the last tiny
which are more suitable for both complex and simple molecule struc-
CO2 bubble is completely dissolved in [omim][BF4] at the equilibrium
tures. For instance, Chapman et al. [89] proposed the statistical asso-
conditions [96]. A homogeneous liquid phase appears at higher pres-
ciating fluid theory (SAFT), which uses a reference fluid that in-
sures whereas at lower pressures, gas and liquid phases co-exist: a CO2-
corporates the chain length and molecular association. The reference
rich phase and a [omim][BF4]-rich phase [96]. Gutkowski et al. [97]
fluid utilized in SAFT is simpler than the hard sphere reference fluid
found no critical points for any of the measured isopleths. It was con-
used in most of EOSs. This EOS provides useful thermodynamic basis
cluded that the critical condition will not happen at temperatures below
through employing the chemical potential or fugacity concept to take
363 K and pressures up to 100 MPa.
into account the effects of molecular structure and interactions on the
Fig. 12 shows a pressure-composition diagram for CO2 in ten
phase behavior and properties of fluids [6]. In late 1990, different
common imidazolium based ILs where the pressure is plotted against
versions of SAFT such as SAFT-VR (for potentials of variable attractive
the CO2 mole fraction at 298 K [98].
range),[90] Soft-SAFT (built from a reference Lennard-Jones fluid)
Another approach to attain appropriate and reliable properties of
[91,92], and PC-SAFT (perturbed-chain SAFT) [93] have been devel-
ILs and to experience high gas solubility of amines is using mixtures of
oped for improving the accuracy of the reference fluid term. According
ILs and amines. For example, Camper et al. [99] used two mixtures of
to the literature, the perturbed-chain statistical associating fluid theory
alkanolamines, namely MEA and DEA with [Rmim][Tf2N] ILs. Fig. 13
(PC-SAFT) is a suitable model to simulate thermodynamic behaviors of
demonstrates CO2 uptake by two mixtures of ILs-amine. They observed
pure ILs and binary systems of CO2 and ILs.
that the absorption performance of ILs-amine mixture is higher than
The PC-SAFT equation is written in terms of Helmholtz free energy
pure amines and the desorption energy is lower. Furthermore, they do
as the following:
not exhibit the drawbacks of functionalized ILs.
a͠ res = a͠ hc + a͠ disp + a͠ assoc (32) Optimization algorithms are an important element for properly
in which, a͠ res
is the residual Helmholtz free energy of the system. The controlling processes and developing mathematical models with op-
superscripts hc, disp, and assoc refer to a reference hard-chain con- timal variables [100]. Algorithms simplify calculations and can be used
tribution, a dispersion contribution, and an associating contribution, to help researchers determine the VLE data. This is vital as pressure-
respectively. The following relationship also exists, according to PC- temperature -composition diagrams can provide important
SAFT:
a
a͠ =
RT (33)
The binary interaction parameters are normally used to match the
modeling results with the experimental data. In some cases, if the van
der Waals 2-parameter mixing rules are incorporated in EOSs, modeling
of VLE through EOS is not feasible. Based on the previous studies, the
Wong-Sandler mixing rules are a proper choice while studying CO2-IL
systems [94]. Further information on the mixing rules is found in
Appendix.

5.3. Pressure-temperature-composition diagrams

To further understand the thermodynamic behaviors of CO2/IL

systems, it seems necessary to plot pressure-temperature -composition
diagrams through experimental and modeling methods so that re-
searchers/engineers can specify the thermodynamic state of mixtures at Fig. 11. P-T diagram for the CO2/[omim][BF4] system at different isopleths for
various process and thermodynamic conditions without conducting x CO2 :• 0.101, ○ 0.200, ■ 0.299, □ 0.412, ▲ 0.505, △ 0.602, ◆ 0.702, ◊ 0.752
timely and difficult computational tasks. [96].

513
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 9
Average absolute deviations of experimental and calculated data for equili-
brium (saturation) pressure [6,106].
Models Ionic liquids AAD (%)

303.15 K 313.15 K Average

PR/WS/vL [C2-mim][Tf2N] 4.50 4.10 4.30

[C4-mim][Tf2N] 5.41 3.91 4.66
[C6-mim][Tf2N] 5.80 4.81 5.31
[C8-mim][Tf2N] 4.81 4.98 4.89
PR/vdW [C2-mim][Tf2N] 12.10 13.50 12.75
[C4-mim][Tf2N] 11.71 11.00 11.35
[C6-mim][Tf2N] 9.70 9.61 9.65
[C8-mim][Tf2N] 10.61 9.90 10.25
PC-SAFT [C2-mim][Tf2N] 4.42 3.90 4.16
[C4-mim][Tf2N] 4.93 6.92 5.93
Fig. 12. Effect of the anion and pressure on the solubility of CO2 in [bmim] [C6-mim][Tf2N] 8.23 6.71 7.47
cation based ILs at a temperature of 298 K [98]. [C8-mim][Tf2N] 8.06 9.17 8.61

thermodynamic data through a simple, straightforward, and under-

standable manner.
Particle swarm optimization (PSO) is an algorithm based on a social-
psychological metaphor; a population of individuals (referred to par-
ticles) adapts by returning stochastically towards previously successful
regions [101]. According to Lazzús [101], the PSO algorithm is ap-
propriate to obtain the optimal values of important parameters in-
volved in the VLE of binary systems containing supercritical fluids and
ILs. Providing further information, the proposed PSO algorithm evalu-
ates the binary interaction parameters with high accuracy. Lazzùs [101]
compared the results obtained from the PSO and Levenberg-Marquart
(LM) algorithms. It was found that the PSO algorithm estimates VLE
information of the system more accurate than LM. The flow diagram of
the PSO algorithm to determine VLE data is shown in Fig. 14.
Lazzus [102] also introduced a hybrid model including artificial
neural network (ANN) and particle swarm optimization (PSO) for pre-
diction of the thermodynamic properties.
Fig. 13. CO2 uptake in equimolar solutions of [C2mim][Tf2N]-MEA and
[Rmim][Tf2N]-DEA (modified after [99]).
5.4. Experimental and modeling data for thermodynamic properties

There are advantages and disadvantages with both experimental

and modeling approaches to forecast the thermodynamic properties and
behaviors of CO2-IL systems. Experiments provide useful thermo-
dynamic information. However, conducting experiments is costly and
time consuming. The costs associated with the experiments also limit
the amount of ILs to be used in the tests. On the other hand, the
modeling data are associated with some inherent errors due to as-
sumptions, convergence, and solution method drawbacks [103]. Tables
9–11 represent the experimental CO2 solubility data for imidazolium,
phosphonium, pyridinium, pyrrolidinium, and functionalized ILs. It is
important to compare experimental and modeling data to ensure that
the appropriate model is chosen for the process and the model is ac-
curate enough to predict the system's behaviors. Table 9 shows the
absolute average deviations, in percentage, between the experimental
and modeling data of the PC-SAFT EOS and PR EOS where Wong-
Sandler/van Laar (WS/vL) mixing rules and van der Waals (vdW)
mixing rules are employed. These results show that the PR/WS/vL
model predicts the high-pressure solubility of CO2 in [Cn-mim][Tf2N]
more accurate than PR/vdW model under pressures up to the super-
critical region of CO2 [94]. PC-SAFT EOS was used to compare the
experimental solubility data of CO2 in the imidazolium based ILs. It was
found that PC-SAFT EOS is able to estimate the solubility in a good
agreement with experimental data [6]. PC-SAFT parameters can be
determined using a correlation in terms of density and vapor pressure
data of ILs. Hence, the parameters can be obtained without carrying out
experiments. The binary interaction of ILs (as discussed before) is
Fig. 14. : PSO algorithm diagram for vapor-liquid equilibrium modeling [101]. temperature dependent and increases linearly with temperature. The

514
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 10
CO2 solubility experimental data for imidazolium-based ionic liquids.

Ionic liquid T (K) P (bar) Max solubilitya H (bar) Reference

(mol CO2/mol IL)

[hmim][FEP] 283.15, 298.6, 323.2 0.297-18.1 0.517 18.5 [107]

[bmpyrr][FEP] 0.498 20.3
[ETT][FEP] 0.497 20.8
[bmim][PF6] 293-393 1-97 0.555 12.0 [108]
[bmim][PF6] 283-348 0.096-19.99 0.379 34.0 [109]
[bmim][BF4] 0.364 38.0
[bmim][PF6] 313.15, 323.15, 333.15 0.97-92 0.729 N/A [75]
[C8mim][PF6] 0.755
[C8mim][BF4] 0.513
[bmim][NO3] 0.708
[emim][EtSO4] 0.579
[N-bupy][BF4] 0.423
[C2-mim][Tf2N] 301-344.4 10-43 0.761 N/A [110]
[C4-mim][Tf2N] 0.803
[C6-mim][Tf2N] 0.833
[C8-mim][Tf2N] 0.845
[bmim] [DCA] 298.15, 313.15, 333.15 1-120 0.550 N/A [98]
[bmim] [NO3] 0.475
[bmim] [BF4] 0.520
[bmim] [PF6] 0.640
[bmim] [TfO] 0.641
[bmim] [Tf2N] 0.681
[bmim] [methide] 0.761
[bmim][NO3] 0.442

a
Maximum solubility in the range of tested operating conditions.

binary interaction can be obtained by curve-fitting of the experimental volume and energy, and can potentially predict a more precise and
vapor- liquid equilibrium data, and a correlation can be determined reliable phase behavior when the association between the molecules
using the group contribution method. As shown in Table 9, PC-SAFT becomes important. The SAFT parameters have the physical sig-
EOS estimates the solubility of CO2 in ILs with a higher accuracy, nificance, featuring diameter, number of segments, and segment-seg-
compared to cubic EOSs. ment interactions [105]. Therefore, the SAFT EOSs consider the effect
In recent years, advanced EOS have been developed in the frame- of molecular structure and interactions on the bulk properties and
work of statistical mechanics and have been proven promising due to phase behavior unlike the cubic EOSs that assume molecules as a hard
their capability to accurately simulate complicated systems [104,105]. sphere [6].
Arguably, the most widely used EOSs in this context are the Statistical Based on the experimental and modeling investigations available in
Associating Fluid Theory (SAFT) and the Perturbed Chain-SAFT (PC- the open sources, the proper cubic EOSs to reasonably simulate the
SAFT) [104]. thermodynamic behaviors of CO2-IL systems are also reported in this
The SAFT and PC-SAFT EOSs were proposed to overcome the study (see Table 6) where modifications on the mixing rules and in-
drawbacks of cubic EOSs, especially in estimating the density and teraction terms of the cubic EOSs have been made to attain a good
compressibility factor of liquid phases. The cubic EOSs do not consider agreement between the real data and modeling outputs [78].
the effect of association and cross association of molecules on the phase As Eq. (34) shows, AAD in Table 9 stands for the average absolute
equilibria. This will be an important feature in the IL-CO2 systems. The deviation which represents the difference between the experimental
PC-SAFT EOS incorporates the influence of molecular association on the and calculated target parameter(s).

Table 11
CO2 solubility for phosphonium, pyridinium, and pyrrolidinium ionic liquids.
Ionic liquid T(K) P(bar) Max Solubilitya (mol CO2/mol IL) H (bar) References

[THTDP][NTf2] 293.2–363.2 1.06–375 0.879 19.8 [111]

[THTDP][Cl] 0.800 18.9
[N-bupy][BF4] 313.15,323.15, 333.15 0.97–91.9 0.581 [75]
[P(14)666][Cl] 303.15 1 36 [112]
[P(14)666][DCA] 0.033 29
[P(14)666][Tf2N] 37
[P(2)444][DEP] 59
[P(14)444][DBS] 29
[N(4)113][Tf2N] 65
[N(6)113][Tf2N] 53
[N(10)113][Tf2N] 43
[N(4)111][Tf2N] 61
[N(6)111][Tf2N] N/A
[N(10)111][Tf2N] 44
[N(1)888][Tf2N] 0.034 29
[N(6)222][Tf2N] 60
[hmpy][Tf2N] 283 1–10 0.200 25.8 [113]

a
Maximum solubility in the range of tested operating conditions.

515
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

npts
100 Pisat ,exp − Pisat , cal the selectivity and solubility of CO2 in ILs with respect to various gases
AAD (%) = ∑ such as N2, CH4, H2, and H2S should be evaluated. In this section, so-
npts i=1
Pisat ,exp (34)
lubility of CO2 in ILs and its selectivity in the presence of different gases
in which, npts represents the number of data points; Psat,exp represents will be discussed.
the experimental saturation pressure; and Psat,cal denotes the calculated
saturation pressure. 6.1. CO2 solubility in ILs
Table 12 shows the CO2 solubility for functionalized ionic liquids at
two different temperatures and pressures. Providing further clarifica- It has been found that anions have a higher impact, compared to
tion on Table 12, functionality of NH2 -Cation means that the group of cations, on the solubility of CO2 in ILs [121]. The CO2 solubility in ten
NH2 is attached to a chemical group in the IL to form the cation part of distinct imidazolium-based ILs was evaluated by Sudhir et al. [98].
the functionalized IL. NH2 -Anion refers to the combination of NH2 They observed that the highest and lowest CO2 solubility can be ob-
group and a chemical branch in the IL structure to hold a negative tained in the ILs with fluoroalkyl groups in the anion part (e.g., [me-
charge, leading to formation of the anion part of the functionalized IL. thide] and [Tf2N]−) and nonfluorinated inorganic anions (e.g.,
[NO3]− and [DCA]−), respectively. Their experiments also showed
5.5. Effect of ions on thermodynamic behaviors that the solubility strongly depends on the type of anions. However, in
the smaller extent, the CO2 solubility can be affected by the selection of
ILs are composed solely of anions and cations. It implies that there alkyl groups on the cation part. In general, when the length of alkyl
are many different combinations of ions that can be used to tailor ILs for chain increases from butyl to octyl, the CO2 solubility increases [98].
various purposes. According to the literature, there are certain anions Molecular dynamic simulations show that the imidazolium-based ILs
and cations that offer a higher performance for CO2 capture processes. have better performance in terms of CO2 capture than phosphonium-
Brennecke and Maginn [116] found from their research study that the based ILs, where the same cation and anion are in the structure of ILs.
types of cations and anions significantly affect the chemical, physical, For instance, [Tf2N]− is more efficient, compared to [PF6]− [122].
and thermodynamic properties of ILs. For instance, the anion part of ILs Zhang et al. [107] used COSMO-RS and conducted a series of experi-
has the greatest effect on the solubility. According to Freitas et al. ments to investigate the CO2 capture by ILs. They determined the
[117], the solubility is strongly related to the energy of vaporization Henry's law constants of CO2 with different combinations of the cations
and molar volume. As most of ILs have a high energy of vaporization, and anions. Their results implied that [FEP] anion has the greatest CO2
ILs with larger anions have higher molar volumes, leading to a greater solubility. Their experiments on three different cations with [FEP]
CO2 dissolution. It has been confirmed that the alkyl chain length of the anion revealed that CO2 is more soluble in [hmim][FEP] (imidazolium
cation has a major impact on the solubility of CO2 in ILs. The alkyl based IL) than other ILs. Solubility and other properties including
group increases the dispersion forces of the cation for better interaction Henry's law constant, enthalpies, and entropies of nine gases dissolved
with CO2, implying that growing the length of the alkyl chain leads to in common imidazolium based IL ([bmim][PF6]) were studied where
an increase in the CO2 solubility [118]. the pressure was up to 13 bar and three distinct temperatures (10, 25,
The type of ions can also influence the melting point of the ILs. It and 50 °C) were examined. The experimental results showed that the
was found that the melting point shows an ascending trend with in- highest and lowest solubilities in [bmim][PF6] are obtained for carbon
creasing the chain length. When an IL contains a halide as an anion, the dioxide and oxygen, respectively. It was also concluded that [bmim]
melting point is higher, compared to the cases having other anions [PF6] (compared to common organic solvents) is not a very good sol-
[119]. vent for reactions containing H2, O2, and CO (e.g., hydrogenation re-
The miscibility and hydrophobicity of ILs with water are also af- actions, oxidation reactions, and hydroformylation reactions) for in-
fected by the type of ions. For example, 1-alkyl-3-methylimidazolium dustrial purposes at extremely high pressures [18].
hexafluorophosphate ILs are immiscible with water, while 1-alkyl-3- Solubility of CO2 in ILs is an important factor to screen proper
methylimidazolium nitrate ILs are completely miscible with water. solvents and to design an efficient absorption system in CO2 capture
Similarly, the hydrophobicity of an IL increases with increasing the plants [123]. The solubility of gases in liquid solvents using Krichevsky-
chain length of the alkyl chain [120]. Kasarnovsky [124] equation is obtained as follows:
There is a correlation between the cation and the density where
f2 (T , P ) V ∞ (P − P sat )
Uygur [87] concluded that the density of the IL lowers with increasing ln = ln H2 (P sat ) + 2
the alkyl chain length for the imidazolium cation case. x2 RT (35)

where f2(T, P) represents the fugacity of gas solute in the gas phase at
6. CO2 solubility and selectivity temperature T and pressure P; x2 is the mole fraction of the gas dis-
solved in the liquid solvent; Psat denotes the saturation or vapor pres-
Flue-gas in practice contains various gases and impurities. Hence, sure of the liquid solvents; V2∞ refers to the partial molar volume of the

Table 12
CO2 solubility for functionalized ionic liquids.
Ionic liquid Functionality P(bar) T(K) Max solubility (mol CO2/mol IL) H (bar) References

[P66614][Ile] NH2 -Anion 0–1.5 295.15 0.5 26 [114]

[P66614][Sar] NH2 -Anion 0.523 1000
[P66614][Gly] NH2 -Anion 0.574 5.5
[P66614][Pro] NH2 -Anion 0.56 1000
[P66614][Met] NH2 -Anion 0.568 18
APMim[NTf2] NH2 -Cation 0–10 303 0.26 [115]
APMim[DCA] NH2 -Cation 0.29
APMim[BF4] NH2 -Cation 0.33
AEMPyrr[BF4] NH2 -Cation 0.27
MeImNet2[BF4] NR3- -Cation 0.25
Bmim[Tau] NH2 -Anion 0.44
Bmim[Gly] NH2 -Anion 0.4

516
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

gas at the infinite dilution of the liquid solvents; H2 stands for the It seems necessary to optimize CO2 selectivity in ILs in terms of gas
Henry's law constant of the gas in the liquid solvents; R signifies the composition and operating conditions. The selectivity generally in-
universal gas constant; and T is the temperature of the system. The creases the number of stages required for CO2 capture processes and
Krichevsky-Kasarnovsky equation is often used to obtain the Henry's consequently the operating and capital costs. For instance, in the case of
law constant from solubility data [124]. Thus, this equation is only CO2/H2S /CH4 systems at high temperatures, an additional step is re-
limited to predicating the solubility of a gas in a liquid at low solubility quired to separate H2S and CH4 from CO2 [127].
conditions and should be avoided for the cases of CO2 solubility in an IL
with high magnitudes of solubility. It should be noted that the Henry's 7. CO2 capture using ILs: performance and modifications
law is valid only if the concentration of a component in the liquid phase
is minor. As discussed in the introduction, carbon capture processes, in gen-
The fugacity (f) of a pure gas can be determined using the following eral, are divided into three main categories, including pre-combustion,
equation: post-combustion, and oxy-combustion. Among these methods, the post-
f (T , P ) = ϕ (T , P ) P (36) combustion is appropriate for retrofit plants and the industrialized
absorption methods mainly employ the post-combustion technology.
in which, ϕ refers to the fugacity coefficient at pressure P and tem- The current solutions for absorption method are amine-based solvents,
perature T. This thermodynamic property can be obtained using various since they have a high CO2 capacity. The practical applications of ILs
EOSs such as SRK and PR [36]. for CO2 capture depend on their effectiveness, compared to the amines.
Maia et al. [86] claimed that all cubic-modified EOSs lead to almost It is believed that ILs can be tailored to have a comparable CO2 capacity
the same accuracy while determining the solubility. In many cases, the without having the disadvantages associated with amine-based sol-
Peng-Robinson (PR) and modified PR EOSs are appropriate EOSs to vents.
model the gas solubility. The PR EOS combined with the Wong- Due to the volatility of amine-based solvents, the processes, which
Sandler/UNIQUAC mixing rules can be used to estimate the Henry's law use this type of solvents, consume a high amount of energy during the
constant, which provides the gas solubility in a liquid solvent when the solvent regeneration stage, which leads to an extra cost to the entire
gas concentration in the liquid phase is small. This strategy helps to CO2 capture process. This cost can be significantly lowered or even
obtain the limiting slope defined by Henry's law as the solubility ap- eliminated through using ILs as the non-volatile solvents. The amine-
proaches zero [125]. According to Carvalho et al. [111], this metho- based solvents are not environmentally friendly due to their volatility
dology results in uncertainty that considerably affects the accuracy and nature. They are unstable in nature, resulting in producing dangerous
reliability of the model outputs. Other models such as Predictive Peng- toxic materials such as nitrosamines, nitramines, and amides. It should
Robinson, Peng-Robinson-Stryjek-Vera, and Sanchez-Lacombe EOSs be noted that nitrosamines are of the most concern as they are carci-
have been utilized to determine the gas solubility in ILs, exhibiting an nogenic and noxious to humans even at low concentrations [17]. ILs
error percentage less than 9.0% [37]. due to their negligibly low vapor pressure are considered non-volatile.
It means that they do not vaporize and therefore do not contaminate the
6.2. CO2 selectivity atmosphere. Therefore, solvent losses are negligible. Amines also par-
ticipate in various parts of absorption processes and produce waste
Due to the presence of gases and impurities in the flue-gas, the se- which can eventually corrode the equipment. However, ILs are not
lectivity of CO2 in ILs appears to be vital in carbon capture technique usually corrosive due to their high thermal and chemical stability. ILs
with ILs [126]. To systematically study the solubility of CO2 in ILs, CO2 are normally degraded at temperatures higher than 300 °C and only
selectivity in ILs needs to be investigated [127]. According to Zhang cause corrosion to equipment if there are impurities in the ILs [17].
et al. [128], conventional ILs have a greater CO2 selectivity, compared Conventional ILs capture carbon through physical absorption me-
to other gases (e.g., CO, H2, and O2) due to the large quadrupole mo- chanisms. Although adjusting cations/anions in the ILs can enhance the
ment of CO2. Since there is a mixture of different gases in the flue-gas, solubility of CO2 in conventional ILs, their CO2 capture capacity is still
this high CO2 selectivity is interesting, implying that ILs absorb CO2 low in comparison with the current commercial carbon capture tech-
easier than other gases. In general, the solubility of CO2 in ILs is higher nologies such as amine-based solvents [29]. Indeed, there are strong
than N2 and H2. However, the solubility of H2 and N2 exhibits an in- chemical interactions between amine and CO2 (−ΔH = 80–130 kJ/
crease with increasing temperature [32,129]. In CO2/H2S systems, CO2 mol) in the amine-based solvents which make amines favorable for CO2
selectivity is highly dependent on CO2/H2S feed ratio and CO2 solubi- molecules. However, in the case of conventional ILs/ CO2 systems, there
lity is remarkably decreased in the presence of H2S [35]. Ramdin et al. is a comparatively weak forces (−ΔH = ~ 15 kJ/mol) between ILs and
[130] showed that the selectivity of CO2/CH4 in different phosphonium CO2 molecules [26]. Among alkanolamines as the most common sol-
and imidazolium-based ILs significantly decreases with increasing vents for the absorption purpose, mono ethanol amine (MEA) is more
temperature. efficient than other amines. The order of absorption rate for amines is as

Table 13
Comparison of properties of ILs with common solvents used for CO2 capture [128,132,133].
Variable Chemicals for CO2 Capture

Conventional ILs Task-Specific ILs MEA (30 %wt) DEPG PC

Absorption type Physical Chemical Chemical Physical Physical

Viscosity (cP) 20-2000 50-2000 18.98 5.8 3.0
Vapor pressure (bar) at 25oC 1.33×10-9 1.33×10-9 8.5×10-4 9.73×10-7 1.13×10-4
ΔHabs (kJ/mol CO2) at 1bar and 40ºC ~10- 20 ~40-50 ~85 ~14.3 ~15.9
CO2 solubility (mol/mol) at 1bar and 20-40 ºC > 2.51 1.6 50-85 3.63 3.41
Selectivity CO2/CH4 8-35 n/a 15 26
H2S/CO2 2-10 1-2 8.8 3.3
CO2/H2 50-150 n/a 77 126
CO2/N2 30-100 n/a 50 117
Cost 200 $/5 gr 400 $/5gr 370 $/1L 80 $/1L 117 $/1L

517
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

following: MEA > DEA > AMP > DIPA > MDEA[131]. Popular sol-
vents for the physical absorption processes are DEPG (Selexol process),
methanol (Rectisol process), NMP or N-methyl-2-pyrrolidone (Purisol
process), Morphysorb process, and Propylene Carbonate (Flour process)
[132]. General comparison between CO2 capture processes through
conventional ILs and the most common used commercial solvents are
provided in Table 13 [128].
As provided in Table 13, pure conventional ILs are not a promising
option for the CO2 capture processes. To overcome this drawback, task-
specific or functionalized ILs with the ability of chemisorption rather
than physisorption have been developed. In this type of ILs, suitable
moieties such as amines are introduced in the conventional ILs [134]. In
this property-modification process, CO2 -philic functional groups are
attached to ILs which can significantly increase the CO2 capture capa- Fig. 15. CO2 absorption mechanism of super basic ILs [144].
city of ILs [135]. Task-specific ILs in the broad classification are divided
into three groups including; cation-functionalized ILs, anion functio-
nalized ILs, and dual amino-functionalized ILs. The experimental results
show that the reaction mechanisms between the amine functionalized
IL and CO2 are similar to the mechanisms between the organic amine
and CO2. Carbon capture capacity of task-specific ILs is also comparable
to alkanolamines [4].
Complicated synthesis and purification steps, high viscosity which Fig. 16. General reaction leading to production of amino acid IL [143].
leads to slow CO2 diffusion mass transfer, and high production cost are
the main challenges to use functionalized ILs for CO2 capture [136]. For
CO2 capture, the energy consumption is fairly high which can be con-
instance, Camper et al. [99] concluded that the synthesis of amine-
siderably decreased by utilizing an IL-amine system [99]. Based on the
functionalized imidazolium-based ILs requires several steps, implying
literature, when the content of ILs is 30 wt%, the energy consumption
that the process is not economical compared to the absorption processes
lowers by 27%, in contrast to the conventional MEA processes. How-
that utilize conventional solvents such as MEA. Therefore, in order to
ever, if the concentration of ILs is lower than 10 wt%, the presence of
improve CO2 capture capacity and to avoid these drawbacks, mixing ILs
ILs in aqueous 30 wt% MEA might decrease the required energy during
with other appropriate compounds such as water and amines are pro-
the regeneration process without affecting the absorption efficiency/
posed. In this approach, solubility increases due to the viscosity re-
performance [141]. This increase in the energy efficiency would be
duction of the IL in the presence of water [128]. This mixture preserves
beneficial as it would cause a decrease in the total costs of CO2 capture
the favorable properties of ILs for CO2 capture and overcomes the in-
processes. Table 14 summarizes various characters of three types of IL
herent drawbacks of ILs including high viscosity and cost [137]. A
mixtures discussed in this section.
mixture of alkanolamines such as MEA and an IL can also increase
Recently, a new generation of the task specific ILs such as super base
energy efficiency, absorption rate, and CO2 absorption capacity [6].
and amino acid ILs with an equimolar CO2 absorption capacity has been
There are some research studies in the literature, focusing
introduced. In the super basic task-specific ILs, the CO2 molecule from C
on this challenging topic [99,138–140]. Common IL-amine mixtures
atom is bonded to the nitrogen atoms in the anionic part of ILs, as
are MEA + 1-butyl-3-methylimidazolium tetrafluoroborate ([bmim]
shown in Fig. 15. Amino acids contain a carboxylic acid and an amino
[BF4]) + H2O and MEA + 1-(2-hydroxyethyl)-3-methylimidazolium
group which can be utilized as either an anion or a cation. Amino acid
([hemim][DCA]) + H2O. Xu et al. [141] conducted several experiments
ILs have a promising carbon capture characteristic due to the presence
and found that the viscosity of two ILs, known as 1-(2-hydroxyethyl)-3-
of two functional groups, including carboxylic and amino groups.
methyl-imidazolium dicyanamide ([C2OHmim][DCA]) and 1-butyl-3-
Furthermore, these ILs are more biodegradable, biocompatible, and
methylimidazolium ([Bmim][DCA]), which are 50 mPa.s and 7 mPa.s
easily producible in the large scales with low costs [142]. The general
at 313.15 K respectively, lowers to 3.54 mPa.s through mixing with
reaction of amino acid IL production is illustrated in Fig. 16 [143].
amine (MEA). According to another study carried out by Yang et al.
Recently, deep eutectic solvents (DESs) have been introduced as a
[139], the viscosity of an IL-amine solution including [bmim][BF4] (as
new class of ILs. They have most of the promising properties of ILs and
an IL) is 3.54 mPa.s at 323 K, which is much lower than the viscosity of
they are economically comparable to the amines for CO2 capture. DESs
the conventional [bmim][BF4] (e.g., 31.08 mPa.s). The low viscosity of
are synthesized by ammonium and phosphonium salts with different
the IL-amine systems conveys the message that a higher mass transfer
hydrogen bond donors (HBDs) such as urea. Some DESs offer a higher
occurs where they are in contact with CO2 such that more CO2 is ab-
CO2 solubility and a lower viscosity, compared to ILs. Moreover,
sorbed at a faster rate, which is favorable for the CO2 capture opera-
functionalized DESs promote a higher solubility for CO2, compared to
tions.
the conventional DESs [143,145,146]. It is important to note that CO2
It should be noted that when the conventional amines are used for

Table 14
Summary of advantages and disadvantages of three ionic liquid (IL) mixture systems [17].
Type Advantages Disadvantages

Conventional ILs (CILs) • High CO selectivity • High viscosity

• Lower
2
energy consumptions than conventional solvents
Task-specific ILs • functional
Higher CO storage capacity, compared to CILs (due to • Higher viscosities, compared to CILs
• Complicated
2
groups) synthesis processes
IL-alkanolamine-water mixture • Lower viscosities, compared to CILs • Not environmentally friendly as pure ILs
• Higher CO dissolution, compared to CILs
2 • Amine loss and energy consumption are lower for this type of ILs than amines
but higher than CILs

518
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 15
Prices for imidazolium, phosphonium, pyridinium, and guanidium-based ILs and tailored ILs [148].
IL-based Cation Anion Price (CAD$) Quantity (g)

Imidazolium 1-Allyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide 160 5

bromide 151 5
chloride 113 5
dicyanamide 150 5
iodide 225 5
chloride 90 5
1-Benzyl-3-methylimidazolium hexafluorophosphate 105 5
tetrafluoroborate 125 5
chloride 225 5
1-Butyl-2,3-dimethylimidazolium hexafluorophosphate 83 5
tetrafluoroborate 99.4 (> 99%), 33(> 97%) 5
4-(3-Butyl-1-imidazolio) butanesulfonate 303 5
acetate 331 5
bis(trifluoromethylsulfonyl)imide 142 5
bromide 71 5
chloride 73 5
dicyanamide 305 5
1-Butyl-3-methylimidazolium hexafluoroantimonate 105 5
hexafluorophosphate 90(> 97%) 5
hydrogen sulfate 149 5
iodide 68 5
tetrafluoroborate 45 5
1,3-Dimethylimidazolium dimethyl phosphate 46 5
methyl sulfate 188 5
tetrafluoroborate 43 5
bis(pentafluoroethylsulfonyl)imide 339 5
bis(trifluoromethylsulfonyl)imide 221 5
1-Ethyl-2,3-dimethylimidazolium chloride 40 5
bromide 40 5
dicyanamide 313 5
tetrafluoroborate 275 5
trifluoromethanesulfonate 103 5
bis(trifluormethylsulfonyl)imide 185 5
1-Hexyl-3-methylimidazolium chloride 194 5
hexafluorophosphate 87 5
1-Methylimidazolium chloride 135 5
Phosphonium Tetrabutylphosphonium Methanesulfonate 84 5
Tetrafluoroborate 338 5
p-toluenesulfonate 122 5
bis(trifluoromethylsulfonyl)amide 56 5
bis(2,4,4-trimethylpentyl)phosphinate 27 5
Trihexyltetradecylphosphonium Bromide 90 5
Chloride 25 5
Decanoate 90 5
Dicyanamide 33 5
Triphenylphosphonio propane-1-sulfonate 258 5
Triphenylphosphonio propane-1-sulfonic 963 50
Pyridinium 1-Butyl-3-methylpyridinium bis(trifluormethylsulfonyl)imide 175 1
Hexafluorophosphate 156 5
1-Butyl-4-methylpyridinium Iodide 59 5
Tetrafluoroborate 110 5
1-Butylpyridinium Bromide 97 5
1-(3-Cyanopropyl)pyridinium Chloride 158 5
1-Ethylpyridinium Tetrafluoroborate 90 5
3-Methyl-1-propylpyridinium bis(trifluormethylsulfonyl)imide 183 1
Guanidinium Guanidinium iodide 82.50 5
Guanidinium chloride 97 5
Guanidine thiocyanate 120 100
Guanidine thiocyanate (solution) 107.50 50 ml
Tailored 1-(3-Cyanopropyl)-3-methylimidazolium bis(trifluoromethylsulfonyl)amide 439 5
1-(3-Cyanopropyl)-3-methylimidazolium chloride 390 5

solubility in amine-based DESs such as ChCl-DEA is much higher than (Now MilliporeSigma), DuPont, Fluka, Scionix, Chemada, and Degussa
conventional DESs and 30% (by weight) aqueous amines [145,146]. that provide ILs for various purposes; however, most of ILs are pro-
duced at the laboratory scales and their prices are in the range of $1/g-
8. Economic analysis of CO2 capture using ionic liquids $10/g. The cost should be reduced by a factor of approximately 100 or
greater to be comparable with that of the conventional solvents.
Currently, ILs are not as commercially available to the same extent Ramdin et al. [26] reported that the lab-scale cost of ILs is around
as the amine-based solvents that are conventionally used in CO2 cap- $1000/kg, which is 100–1000 times more expensive, compared to the
ture. They are usually used at the lab scale which is very far from the conventional solvents. This price is expected to drop by $40/kg to
industrial scale in terms of efficiency, operation, and cost. According to $100/kg according to BASF company when ILs are produced at larger
Arshad [147], there are several vendors such as Merck, Sigma Aldrich scales; however, it is still higher than the price of conventional solvents

519
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

by a factor of 10–20 due to the complexity and cost of the synthesis and IL will work at ideal conditions, and selectivity of CO2 in the presence of
purification stages. The prices of imidazolium based, phosphonium various gases by the IL. There are also other factors such as the toxicity
based, pyridinium based, and tailored ILs are listed in Table 15. As and biodegradability of the ILs if it is planned to use ILs in large
shown in Table 15, the conventional ILs are cheaper than the functio- quantities.
nalized ones. In general, the tailored ILs are at least five times more
expensive than most conventional ILs. It is worth mentioning that all
9.1. Economic prospect
prices are based on the official website of Sigma Aldrich (Now Milli-
poreSigma) company, accessed June 2017 [148].
As discussed before, ILs are currently expensive solvents. Their
Protic ILs with their simple acid-base chemistry are used in in-
production processes are fairly difficult, which is a major obstacle while
dustrial processes and have a low cost (< $1.23/kg) which is com-
utilizing ILs for CO2 capture at large scales. Many ILs are only available
parable to that of the organic solvents such as toluene and acetone
at the lab scales. Hence, it is not easy to estimate the total purchase cost
[149]. Also, the cost for synthesizing ILs with hydrogen sulfate (as the
if they are used in industrialized operations. The supplier Sigma-Aldrich
anion group) through acid-base neutralization is comparable with that
is one of the ILs suppliers that provides the typical costs of a majority of
of the organic solvents such as acetone and ethyl acetate ($1.34/kg to
ILs. It is clear that as the amount of the requested ILs increases, the
$1.4 /kg). For example, the prices of triethylammonium hydrogen
price per gram drops significantly.
sulfate and 1-methylimidazolium hydrogen sulfate are $1.24/kg, and
According to the prices listed for ILs by Sigma Aldrich (Now
$2.96/kg to $5.88/kg, respectively [150].
MilliporeSigma), the five best ILs in terms of economic aspect are: tri-
The total expenses of power plants include the costs for CO2 capture
butylmethylammonium chloride ([TBMA][Cl]); 1-butyl-3-methylimi-
from flue-gas (by amines) in the capital cost (CC) and operating costs
dazolium tetrafluoroborate ([bmim][BF4]); trihexyl(tetradecyl)phos-
(OC) [139]. Using ILs, some parts of the capital costs are less than those
phonium bromide ([P14666][Br]); trihexyl(tetradecyl)phosphonium
with amines. For example, in desorption process, ILs will be recovered
chloride ([P14666][Cl]); and 1-ethyl-1-methylpyrrolidinium bromide
completely due to low vapor pressure of the ILs. However, amines are
([C2Mpyrr][Br]). The prices of ILs at different scales are listed in
volatile and some quantity of amines will be lost during the desorption
Table 16.
process, leading to an increase in material and energy demands. Re-
searchers found that using mixtures of ILs and amines, the energy
consumption of the desorption stage lowers by more than 37%, com- 9.2. Solubility and selectivity prospects
pared to that with the amine solution [139].
Even though ILs are currently more expensive than the amines, their The CO2 capture capacity of an IL is an important factor while
price is expected to be considerably decreased in the future. There are a choosing an IL for a CO2 capture process. Researchers have studied
large number of ILs available that can be synthesized through different many different ILs to determine what are the best in terms of their CO2
ways. They have multiple-form markets such as their utilization as a solubility for the post-combustion CO2 capture processes. It has been
solvent and catalyst with applications to reaction systems, separation found that the anion of ILs has the greatest impact on the CO2 capture
systems, bio-refinery systems, energy storage, and GHG reducing capacity of ILs. According to the literature, the more fluorinated the
media. These are great incentives for commercialization of functional anion is, the higher the CO2 solubility is. It was also concluded that
ILs with low costs. Some of the benefits of ILs, compared to the amines, task-specific ionic liquids (TSILs) have a much greater CO2 solubility
in CO2 capture are as follows [17]: than conventional ILs, since they are usually modified to absorb more
CO2.
• Less energy is required for regeneration of diluted ILs to remove the Oliferenko et al. [153] provided a spreadsheet which lists many ILs
and their corresponding Henry's law constants. Based on the data/in-
absorbed CO2, than amines due to their physical absorption me-
chanisms. formation, the best ILs in terms of solubility are: 1-pentyl-3-methyli-
• There is a low chance of being involved in a reaction with impurities midazolium tris(nonafluorobutyl)trifluorophosphate ([p5mim][bFAP]);
trihexyl(tetradecyl)phosphonium chloride ([P14666][Cl]); 1-hexyl-3-
and low probability of corrosion occurrence in the equipment due to
the high thermal and chemical stability (> 300 °C) of ILs. methylimidazolium tris(heptafluoropropyl)trifluorophosphate ([hmim]
• There is no loss in the gas stream during regeneration due to the low [pFAP]); trihexyl(tetradecyl)phosphonium bis(trifluoromethylsulfonyl)
imide ([P14666][Tf2N]); and methyltrioctylammonium bis(tri-
vapor pressure of ILs.
• There are high potentials to produce a task-specific IL with manip- fluoromethylsulfonyl)imide ([N1888][Tf2N]). The Henry's law constants
for these ILs are presented in Table 17. It should be noted that the
ulating anions and cations due to tunable nature of the ILs.
Henry's law constant represents the amount of CO2 which is physically
The separation cost consists of two parts including direct costs and absorbed by an IL. In fact, it does not account for chemical absorption
indirect costs. The direct costs correspond to the costs of absorption and which can occur in TSILs.
regeneration columns, lean/rich solvent heat exchangers, pumps, lean- The promising ILs in terms of selectivity of CO2 and H2S over other
solvent coolers, CO2 compression unit, and other major equipment. The gases (e.g., methane) include BF4, NO3, and CH3SO4 anions and cations
indirect costs include the general facilities, emergency, and owner's such as Tetrabutylammonium (N4111), Pentamethylguanidinium
costs [151]. In addition, the operating and maintenance expenses
(OME) should be considered. The OME consists of operating labor, Table 16
maintenance costs, administrative and support labor costs, IL makeup, Prices for five potential ILs for CO2 capture [148].
chemicals, solid waste treatment, power, and CO2 transport and storage Ionic liquid Weight (g) Price (CAD$) Price per weight (CAD$/g)
[151].
[TBMA][Cl] 100 129 1.29
500 595 1.19
9. Screening of ionic liquids for CO2 capture [bmim][BF4] 100 180.5 1.81
1000 108.5 1.09
18
Due to the tunability of ILs, room temperature ILs have 10 pos- [P14666] [Br] 50 107.5 3.25
sible structures [152]. This makes selection of the best ILs for CO2 [P14666] [Cl] 5 35.5 7.10
50 112.5 2.25
capture difficult. However, there are particular factors that should be
[C2Mpyrr] [Br] 5 36.2 7.24
considered to choose the most suitable ILs for CO2 capture processes. 50 136 2.72
These features include cost, ability of the IL to absorb CO2, how well the

520
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table 17 absorb a high amount of CO2. However, this is not appropriate for post-
Henry's law constant for potential ILs in terms of solubility [153]. combustion processes, because increasing the pressure will increase the
Ionic Liquid T (K) Henry's law constant (bar) operating costs. Therefore, it is necessary to employ ILs that can attain
an acceptable solubility of > 40% at the moderate temperature range
[p5mim][bFAP] 313 26.92 of 293–313 K and the moderate pressure range of 5–10 bar. Several
303 22.44a
researchers have studied important absorption and characterization
[hmim][pFAP] 313 28.84
303 24.01a
aspects of the amine-functionalized ILs, since they exhibit higher CO2
[[P14666][Cl] 303 23.99 absorption capacities due to chemical absorption, compared to con-
[P14666][Tf2N] 303 26.30 ventional ILs. A part of functionalized ILs used for CO2 capture are
[N1888][Tf2N] 303 38.18 trihexyl(tetradecyl)phosphonium glycinate ([P14666][Gly]), iso-
a leucinate ([P14666][Ile]), 2-cyanopyrrolide ([P14666][2-CNpyr]), 3-(tri-
Linearly interpolated for comparison.
fluoromethyl) ([P14666][3-CF3pyra]), prolinate ([P14666][Pro]), and
methionate ([P14666][Met]) [26,114,135]. These studies reveal that
non-conventional ILs are able to absorb CO2 in a 1:1 stoichiometry at
295 K under low pressure conditions.

10. Main technical and practical challenges

Although thermodynamic models are very useful to characterize

CO2 and IL systems, there are still major challenges in this area.
Adequate knowledge is required to select appropriate EOSs and mixing
rules for complex systems such as H2O/CO2/IL. Inappropriate selection
results in significant errors while predicting the thermodynamic char-
acteristics [158,159].
The critical properties of CO2 are available; however, finding the
critical properties of the ILs is a challenging task, since they are de-
composed at the normal boiling point. Hence, efficient methodologies
need to be developed to accurately obtain the critical properties of ILs.
Fig. 17. Calculated CO2 solubility in various ILs having an anion of Tf2N [155]. Thermodynamic models are often complex to be solved in phase
equilibria for determination of the corresponding parameters. There are
(pmg), and Tetramethylguanidinium (tmg) [154]. computational and software tools that may assist researchers. However,
Maiti [155] concluded that ILs containing pentamethylpropylgua- most of these approaches are expensive, difficult, and time-consuming.
nidinium [ppg] cation and [BF4], [FEP], and [PF6] anions have a This poses another major challenge while employing thermodynamic
higher CO2 solubility. The solubility data are illustrated in Fig. 17 and models.
Fig. 18. Thus, the functionalized-guanidinium cations show a higher High viscosity of ILs, which leads to a decline in solubility of CO2 in
solubility than the imidazolium-based ILs [156,157]. Among anions, ILs, is one of the main challenges while utilizing ILs. There are some
BF4 and PF6 are the best where the CO2 solubility is the main selection solutions for this challenge, which they need to be examined and in-
criterion. vestigated properly. One of the solutions is using mixture of water and
amines with ILs, which the optimum composition and process condi-
tions should be determined for these mixtures.
9.3. Operational condition prospects Another disadvantage of using ILs for CO2 capture is the cost. ILs are
more expensive, compared to amines. Although, the price of ILs in the
Since high pressure and low temperature operations facilitate CO2 large scale (less than $40/kg) can be much lower than currently
absorption, the high pressure and low temperature conditions favor to available lab scale price (~ $1000/kg), the cost is still 10–20 higher

Fig. 18. CO2 solubility in different ILs having a cation of bmim [155].

521
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

than the cost for conventional solvents [157]. Researchers need to find • Current CO 2 capture techniques generally suffer from high capital
simple and economic synthesis solutions for CO2 capture where the and operating costs, solvent degradation, high solvent losses, cor-
targeted solvent is an IL. To use ILs for absorption method in large rosive nature of solvents, toxicity of solvents, need of large ab-
scales, an appropriate system design and optimal operating conditions sorption equipment, and high energy consumption in solvent re-
should be selected. generation.
Recently, there have been new developments in CO2 capture and • Among the EOSs, the results obtained by using PC-SAFT EOS
utilization techniques such as introduction of efficient solid/liquid/gas showed great agreement with the experimental data at atmospheric
systems with high CO2 capture efficiency, utilization of a single liquid and high pressures. Also, PR EOS results in good predictions where
absorber to simultaneously capture CO2 and SO2, hydrogen production the Wong-Sandler/van Laar mixing rules are utilized.
through employing chemical looping technique for CO2 capture, elec- • Although the functionalized or task-specific ILs are more suitable for
trofuels production from CO2, and use of CO2 for algae culturing the absorption of CO2, they are 2–4 times more expensive than the
[160–164]. However, more systematic engineering and research ac- conventional ILs.
tivities need to be conducted to assess the viability of the new strategies • The best ILs in terms of technical and economic prospectives are
in terms of practical implication and economic and environmental screened. Based on this investigation, it is concluded that guanidi-
prospects. nium cations and fluorine anions including [BF4], [FEP], and [PF6]
are the best cation and anion, respectively as they exhibit a high CO2
11. Conclusions solubility.
• To commercialize CO2 capture technology with ILs, it is vital to find
CO2 capture processes involve several challenges in terms of tech- cheap but suitable IL/additive mixtures with a relatively high so-
nical, economical, and environmental prospects. Novel options for re- lubility and selectivity where important mass transfer and thermo-
moval of CO2 from flue-gases, including absorption with ionic liquids dynamic prospects of CO2/H2O/IL/additive systems (diffusivity,
(ILs) have been proposed and investigated from thermodynamic, mass transfer rate, vapor pressure, and equilibrium conditions) are
transport phenomena, and economic points of view. The CO2 capture well explored through experimental and modeling investigations.
process using ILs is briefly explained in this review manuscript where • ILs, as green solvents, offer a promising CO2 capture medium where
the absorption mechanisms and the main practical and economic ad- an efficient absorption rate can be attained if a proper ion or che-
vantages and advantages are highlighted. The physical and chemical mical is added to IL. In addition to engineering design aspects of
characteristics and the critical properties of CO2 and ILs (and their carbon capture processes, their high viscosity, cost, availability,
mixtures) obtained from both experimental and modeling approaches compatibility, and purity are the main challenges to develop ap-
are reported. The influences of important parameters such as type of propriate absorption systems to capture CO2 using ILs at an in-
anions and cations, ions length (size), IL volatility, and operating con- dustrial scale.
ditions on CO2 solubility and selectivity are also discussed. Proper
equations of state (EOSs) to study the vapor-liquid equilibrium of CO2/
ILs systems are reviewed in this paper. In addition, a comparison is Acknowledgements
made between the prices of the most common ILs that are considered as
potential solvents in CO2 capture processes. The cost of ILs plays an The authors would like to acknowledge the financial assistance of
important role in their commercial utilization to absorb CO2. The main the Natural Sciences and Engineering Research Council of Canada
conclusions are drawn as follows: (NSERC), InnovateNL, and Memorial University (MUN) in Canada.

Appendix. Mixing rules for equations of state

To use an EOS for mixtures, EOS constants should be obtained using mixing rules. Mixing rules include the binary interaction terms which
represent the attractive and repulsive forces [158]. In general, the mixing rules are divided into two main categories including random and non-
random mixing rules. The non-random mixing rules are used when the system contains asymmetric compounds such as water and methanol. The
non-random mixing rules are represented by two terms; namely, conventional random mixing term and asymmetric term due to polarity.
The random mixing rules offer acceptable results in many practical cases. It is suggested that the mixing rules used in any EOS should attain the
same form as that of the virial equation at conditions at which both equations are valid. The interactions between molecules in the multicomponent
systems (the coefficients of virial equations) at low pressures are described as follows [159]:

a= ∑ ∑ xi xj (ai . aj)0.5 (1−kij)

i j (A.1)

b= ∑ ∑ xi xj bij = ∑ ∑ xi xj (bi + bj)/2 = ∑ xi bi

i j i j i (A.2)

aij = (ai aj )1/2 (1 − kij ) (A.3)

bij = [(bi + bj )/2](1 − kij ) (A.4)

b= ∑ ∑ xi xj (bi + bj)/2
i j (A.5)

(
RT ∑i ∑j x i x j b −
a
RT ij)
b=
⎡
RT − ⎢∑i x i
⎣
( )+
ai
aj
GE ⎤
C ⎥
⎦ (A.6)

522
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

Table A.1
Different mixing rules for EOSs [159].
Mixing rule Equation

van der Waals a = ∑i ∑j xi xj (ai . aj )0.5 (1 − kij )

b = ∑i ∑j xi xj (bi + bj )/2
Wong-Sandler a GE ⎤
a = b ⎡∑i xi ⎛ i ⎞ +
⎜ ⎟

⎢
⎣ ⎝ aj ⎠ C ⎥
⎦
a
RT ∑i ∑j xi xj ⎛b − ⎞
⎝ RT ⎠ij
b=
⎡ ⎛ a ⎞ GE ⎤
RT − ⎢∑i xi ⎜ i ⎟ + ⎥
a C
⎣ ⎝ j⎠ ⎦
Huron-Vidal a GE ⎤
a = b ⎡∑i xi ⎛ i ⎞ +
⎜ ⎟

⎢
⎣ ⎝ aj ⎠ C ⎥
⎦
Mathias-Klotz-Prausnitz 1/3
3
a = ∑i = 1 ∑ j = 1 xi xj ai . aj (1 − kij ) + ∑i = 1 xi ⎛⎜∑ j = 1 xj ( ai aj I ji ) ⎞⎟
N N N N

⎝ ⎠
b = ∑i xi bi

Once the parameters a and b are determined, they can be used in EOSs.

⎡ a GE ⎤
a = b ⎢∑ x i ⎜⎛ i ⎟⎞ +
a
⎝ j⎠ C ⎥ (A.7)
⎣ i ⎦

N N N N 1/3 3
⎛ ⎞
a= ∑ ∑ xi xj ai . aj (1 − kij ) + ∑ xi ⎜∑ xj ( ai aj I ji )
⎟
i=1 j=1 i=1 ⎝ j=1 ⎠ (A.8)

b= ∑ xi bi
i (A.9)

f2 (T , P ) V∞
ln = ln H2 + 2 P
x2 RT (A.10)

f2 (T , P ) = ϕ (T , P ) P (A.11)
where bij represents the coefficients due to interaction between molecules i and j.
The attractive (aij) and repulsive (bij) forces between molecules i and j in the EOS can be expressed by the simple geometric and arithmetic
average, respectively, through the following equations [159]:
aij = (ai aj )1/2 (A.12)

bij = (bi + bj )/2 (A.13)

Considering the attractive term between two non-similar molecules, the following binary interaction parameter (kij) should be added, as well
[159]:
aij = (ai aj )1/2 (1 − kij ) (A.14)
For the repulsive term, if there is a high concentration of CO2 in the system, another binary interaction parameter for the repulsion term (K ij′ )
should be used as follows:
bij = [(bi + bj )/2](1 − K ′ij ) (A.15)
Common mixing rules that can be applied for ILs-CO2 systems are given in Table A.1.

References [8] Sreedhar I, et al. Process and engineering trends in membrane based carbon cap-
ture. Renew Sustain Energy Rev 2017;68:659–84.
[9] (a) C.A.T. Force, Advanced post-combustion CO2 capture, 2009.
[1] Figueroa JD, et al. Advances in CO 2 capture technology—the US department of (b) Freund H, Bauer J, Zeiser T, Emig G. Detailed simulation of transport processes
energy's carbon sequestration program. Int J Greenh Gas Control 2008;2(1):9–20. in fixed-beds. Ind Eng Chem Res 2009;44(16):6423–34
[2] Bolin B, Eriksson E. Changes in the carbon dioxide content of the atmosphere and [10] Sreedhar I, et al. Carbon capture by absorption–path covered and ahead. Renew
sea due to fossil fuel combustion. atmosphere sea Motion 1959;1:30–142. Sustain Energy Rev 2017;76:1080–107.
[3] Coninck Hd. et al., IPCC special report on carbon dioxide capture and storage. [11] Mirzaei S, Shamiri A, Aroua MK. A review of different solvents, mass transfer, and
Intergovernmental Panel on Climate Change; 2005. hydrodynamics for postcombustion CO2 capture. Rev Chem Eng
[4] Yu KMK, et al. Recent advances in CO2 capture and utilization. ChemSusChem 2015;31(6):521–61.
2008;1(11):893–9. [12] Global C, Institute (GCCSI)(2012) The Global Status of CCS: 2012. Canberra,
[5] Rahman FA, et al. Pollution to solution: capture and sequestration of carbon di- Australia.
oxide (CO2) and its utilization as a renewable energy source for a sustainable fu- [13] Rochelle GT. Amine scrubbing for CO2 capture. Science 2009;325(5948):1652–4.
ture. Renew Sustain Energy Rev 2017;71:112–26. [14] Riemer PW, Ormerod WG. International perspectives and the results of carbon
[6] Chen Y, Mutelet F, Jaubert J-Nl. Modeling the solubility of carbon dioxide in dioxide capture disposal and utilisation studies. Energy Convers Manag
imidazolium-based ionic liquids with the PC-SAFT equation of state. J Phys Chem 1995;36(6):813–8.
B 2012;116(49):14375–88. [15] Kothandaraman A. Carbon dioxide capture by chemical absorption: a solvent
[7] Bailera M, et al. Power to Gas projects review: lab, pilot and demo plants for comparison study. Massachusetts Institute of Technology; 2010.
storing renewable energy and CO2. Renew Sustain Energy Rev 2017;69:292–312. [16] Yu C-H, Huang C-H, Tan C-S. A review of CO2 capture by absorption and

523
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

adsorption. Aerosol Air Qual Res 2012;12(5):745–69. [57] Huddleston JG, et al. Characterization and comparison of hydrophilic and hy-
[17] Torralba-Calleja E, Skinner J, Gutiérrez-Tauste D. CO2 capture in ionic liquids: a drophobic room temperature ionic liquids incorporating the imidazolium cation.
review of solubilities and experimental methods. J Chem 2013;2013. Green Chem 2001;3(4):156–64.
[18] Anthony JL, Maginn EJ, Brennecke JF. Solubilities and thermodynamic properties [58] Valderrama J, Robles P. Critical properties, normal boiling temperatures, and
of gases in the ionic liquid 1-n-butyl-3-methylimidazolium hexafluorophosphate. J acentric factors of fifty ionic liquids. Ind Eng Chem Res 2007;46(4):1338–44.
Phys Chem B 2002;106(29):7315–20. [59] Lydersen AL, Station MEE. Estimation of critical properties of organic compounds
[19] Johnson KE. What's an ionic liquid? Interface-Electrochem Soc 2007;16(1):38–41. by the method of group contributions. University of Wisconsin; 1955.
[20] Henni A, et al. Quantum-mechanical and thermodynamic study of amines and [60] Joback KG, Reid RC. Estimation of pure-component properties from group-con-
ionic liquids for CO2 capture. Faculty of Graduate Studies and Research, University tributions. Chem Eng Commun 1987;57(1–6):233–43.
of Regina; 2013. [61] Maia FM, et al. Modeling of ionic liquid systems: phase equilibria and physical
[21] Zhang S, et al. Physical properties of ionic liquids: database and evaluation. J Phys properties. INTECH Open Access Publisher; 2013.
Chem Ref data 2006;35(4):1475–517. [62] Valderrama JO, Sanga WW, Lazzús JA. Critical properties, normal boiling tem-
[22] Wasserscheid P, Welton T. Ionic liquids in synthesis. 1. Wiley Online Library; perature, and acentric factor of another 200 ionic liquids. Ind Eng Chem Res
2008. 2008;47(4):1318–30.
[23] Ghandi K. A review of ionic liquids, their limits and applications. Green Sustain [63] Alvarez V, Valderrama J. A modified Lydersen-Joback-Reid method to estimate the
Chem 2014;2014. critical properties of biomolecules. Alimentaria 2004;254:55–66.
[24] Welton T. Room-temperature ionic liquids. Solvents for synthesis and catalysis. [64] Valderrama JO, Rojas RE. Critical properties of ionic liquids. revisited. Ind Eng
Chem Rev 1999;99(8):2071–84. Chem Res 2009;48(14):6890–900.
[25] Blanchard LA, et al. Green processing using ionic liquids and CO2. Nature [65] Valderrama JO, Forero LA. An analytical expression for the vapor pressure of ionic
1999;399(6731):28–9. liquids based on an equation of state. Fluid Phase Equilibria 2012;317:77–83.
[26] Ramdin M, de Loos TW, Vlugt TJ. State-of-the-art of CO2 capture with ionic li- [66] Valderrama JO, Sanga WW. Critical properties and vapor pressure of twenty
quids. Ind Eng Chem Res 2012;51(24):8149–77. imidazolium based ionic liquids used in extraction bioprocesses. Frankfurt:
[27] Kirchner B. Ionic liquids. 290. Springer; 2009. Dechema Veranstaltungen; 2008.
[28] Krupiczka R, Rotkegel A, Ziobrowski Z. Comparative study of CO 2 absorption in [67] Paulechka Y, et al. Vapor pressure and thermal stability of ionic liquid 1-butyl-3-
packed column using imidazolium based ionic liquids and MEA solution. Sep Purif methylimidazolium Bis (trifluoromethylsulfonyl) amide. Thermochim Acta
Technol 2015;149:228–36. 2005;439(1):158–60.
[29] Shannon MS, Bara JE. Properties of alkylimidazoles as solvents for CO2 capture [68] Emel'yanenko VN, Verevkin SP, Heintz A. The gaseous enthalpy of formation of the
and comparisons to imidazolium-based ionic liquids. Ind Eng Chem Res ionic liquid 1-butyl-3-methylimidazolium dicyanamide from combustion calori-
2011;50(14):8665–77. metry, vapor pressure measurements, and ab initio calculations. J Am Chem Soc
[30] Kroon MC, et al. Modeling of the carbon dioxide solubility in imidazolium-based 2007;129(13):3930–7.
ionic liquids with the tPC-PSAFT equation of state. J Phys Chem B [69] Langmuir I. The vapor pressure of metallic tungsten. Phys Rev 1913;2(5):329.
2006;110(18):9262–9. [70] Phang P, Dollimore D, Evans SJ. A comparative method for developing vapor
[31] Yokozeki A, Shiflett MB. Gas solubilities in ionic liquids using a generic van der pressure curves based on evaporation data obtained from a simultaneous TG–DTA
Waals equation of state. J Supercrit Fluids 2010;55(2):846–51. unit. Thermochim Acta 2002;392:119–25.
[32] Shiflett MB, Yokozeki A. Separation of carbon dioxide and sulfur dioxide using [71] Aschenbrenner O, et al. Measurement of vapour pressures of ionic liquids and
room-temperature ionic liquid [bmim][MeSO4]. Energy Fuels 2009;24(2):1001–8. other low vapour pressure solvents. Green Chem 2009;11(8):1217–21.
[33] Andreu JS, Vega LF. Modeling the solubility behavior of CO2, H2, and Xe in [C n- [72] Ficke LE, Novak RR, Brennecke JF. Thermodynamic and thermophysical proper-
mim][Tf2N] ionic liquids. J Phys Chem B 2008;112(48):15398–406. ties of ionic liquid+ water systems. J Chem Eng Data 2010;55(11):4946–50.
[34] Yokozeki A, et al. Physical and chemical absorptions of carbon dioxide in room- [73] Dohrn R, Fonseca JM, Peper S. Experimental methods for phase equilibria at high
temperature ionic liquids. J Phys Chem B 2008;112(51):16654–63. pressures. Annu Rev Chem Biomol Eng 2012;3:343–67.
[35] Shiflett MB, Yokozeki A. Separation of CO2 and H2S using room-temperature ionic [74] Torralba-Calleja E, Skinner J, Gutiérrez-Tauste D. CO2 Capture in ionic liquids: a
liquid [bmim][PF6]. Fluid Phase Equilibria 2010;294(1):105–13. review of solubilities and experimental methods. J Chem 2013:1–16.
[36] Zhou L, Fan J, Shang X. CO2 Capture and separation properties in the ionic liquid [75] Blanchard LA, Gu Z, Brennecke JF. High-pressure phase behavior of ionic liquid/
1-n-Butyl-3-methylimidazolium nonafluorobutylsulfonate. Materials CO2 systems. J Phys Chem B 2001;105(12):2437–44.
2014;7:3867–80. [76] Blanchard LA, Gu Z, Brennecke JF. High-Pressure Phase Behavior of Ionic Liquid/
[37] Valderrama JO, Urbina F, Faúndez CA. Gas-liquid equilibrium modeling of mix- CO2 Systems. J Phys Chem 2001;105:2437–44.
tures containing supercritical carbon dioxide and an ionic liquid. J Supercrit Fluids [77] Density Measuring Module: DMA 4500 ME, in Anton Paar.; 2016.
2012;64:32–8. [78] Li H. Thermodynamic properties of CO2 mixtures and their applications in ad-
[38] Karakatsani EK, et al. tPC-PSAFT modeling of gas solubility in imidazolium-based vanced power cycles with CO2 capture processes.; 2008.
ionic liquids. J Phys Chem C 2007;111(43):15487–92. [79] Peng D-Y, Robinson DB. A new two-constant equation of state. Ind Eng Chem
[39] Statistics I. CO2 emissions from fuel combustion-highlights. IEA, Pari; 2011. Fundam 1976;15(1):59–64.
〈http://www.iea.org/co2highlights/co2highlights.pdf〉. CitedJuly. [80] Patel NC, Teja AS. A new cubic equation of state for fluids and fluid mixtures.
[40] Statistics IEAI. CO2 emissions from fuel combustion-highlights.: Paris; 2016. Chem Eng Sci 1982;37(3):463–73.
[41] EIA U. International energy statistics [available at]. Washington, DC: US Energy [81] Redlich O, Kwong JN. On the thermodynamics of solutions. V. An equation of
and Information Administration; 2011 [available at]. 〈http://www.eia.gov/ state. fugacities of gaseous solutions. Chem Rev 1949;44(1):233–44.
countries/data.cfm〉. [82] Soave G. Equilibrium constants from a modified Redlich-Kwong equation of state.
[42] Statistics O. Greenhouse gas emissions; 2017. Chem Eng Sci 1972;27(6):1197–203.
[43] 2016 on Track for Record Rate of Atmospheric CO2 Increase. 2016; Available [83] Stryjek R, Vera J. PRSV2: a cubic equation of state for accurate vapor—liquid
from: 〈https://robertscribbler.com/2016/11/01/〉. equilibria calculations. Can J Chem Eng 1986;64(5):820–6.
[44] Kurihara H, Shirayama Y. Effects of increased atmospheric CO2 on sea urchin early [84] Ren W, Sensenich B, Scurto AM. High-pressure phase equilibria of {carbon dioxide
development. Mar Ecol Progress Ser 2004;274:161–9. (CO 2)+ n-alkyl-imidazolium bis (trifluoromethylsulfonyl) amide} ionic liquids. J
[45] Matin A, et al. Canada's greenhouse gas inventory 1990–2002. Canada: Chem Thermodyn 2010;42(3):305–11.
Greenhouse Gas Division, Environment Canada Ottawa; 2004. [85] Maia FM, et al. Modeling of ionic liquid systems: phase equilibria and physical
[46] Yalkowski S, He Y, Jain P. Handbook of aqueous solubility data second edition. properties, in Ionic Liquids-New Aspects for the Future. InTech; 2013.
New York: CRC Press; 2010. [86] Maia FM, et al. Equation of state modelling of systems with ionic liquids: literature
[47] Haynes WM. CRC handbook of chemistry and physics. CRC press; 2014. review and application with the Cubic Plus Association (CPA) model. Fluid Phase
[48] Sovova H, Prochazka J. Calculations of compressed carbon dioxide viscosities. Ind Equilibria 2012;332:128–43.
Eng Chem Res 1993;32(12):3162–9. [87] Uygur TJ. Experimental and thermodynamic studies of ionic liquids for carbon
[49] Fenghour A, Wakeham WA, Vesovic V. The viscosity of carbon dioxide. J Phys dioxide separation. Faculty of Graduate Studies and Research, University of
Chem Ref Data 1998;27(1):31–44. Regina; 2013.
[50] Austgen DM, et al. Model of vapor-liquid equilibria for aqueous acid gas-alkano- [88] Soave G. Improvement of the van der Waals equation of state. Chem Eng Sci
lamine systems using the electrolyte-NRTL equation. Ind Eng Chem Res 1984;39(2):357–69.
1989;28(7):1060–73. [89] Chapman WG, Jackson G, Gubbins KE. Phase equilibria of associating fluids: chain
[51] Wilcox J. Carbon capture. Springer Science & Business Media; 2012. molecules with multiple bonding sites. Mol Phys 1988;65(5):1057–79.
[52] Palma Chilla L, Lazzús J, Pérez Ponce A. Particle swarm modeling of vapor-liquid [90] Gil-Villegas A, et al. Statistical associating fluid theory for chain molecules with
equilibrium data of binary systems containing CO 2+ imidazolium ionic liquids attractive potentials of variable range. J Chem Phys 1997;106(10):4168–86.
based on bis [(trifluoromethyl) sulfonyl] imide anion. J Eng Thermophys [91] Pàmies JC, Vega LF. Vapor− liquid equilibria and critical behavior of heavy n-
2011;20(4):487–500. alkanes using transferable parameters from the soft-SAFT equation of state. Ind
[53] Hayes R, Warr GG, Atkin R. Structure and nanostructure in ionic liquids. Chem Rev Eng Chem Res 2001;40(11):2532–43.
2015;115(13):6357–426. [92] Felipe J. Thermodynamic behaviour of homonuclear and heteronuclear Lennard-
[54] Tomé LI, et al. Measurements and correlation of high-pressure densities of imi- Jones chains with association sites from simulation and theory. Mol Phys
dazolium-based ionic liquids. J Chem Eng Data 2008;53(8):1914–21. 1997;92(1):135–50.
[55] Dymond J, Malhotra R. The Tait equation: 100 years on. Int J Thermophys [93] Gross J, Sadowski G. Perturbed-chain SAFT: an equation of state based on a per-
1988;9(6):941–51. turbation theory for chain molecules. Ind Eng Chem Res 2001;40(4):1244–60.
[56] Bonhote P, et al. Hydrophobic, highly conductive ambient-temperature molten [94] Yazdizadeh M, Rahmani F, Forghani AA. Thermodynamic modeling of CO2 solu-
salts. Inorg Chem 1996;35(5):1168–78. bility in ionic liquid ([Cn-mim][Tf2N]; n = 2, 4, 6, 8) using Wong-Sandler mixing

524
M. Aghaie et al. Renewable and Sustainable Energy Reviews 96 (2018) 502–525

rule, Peng-Robinson equation of state (EOS) and differential evolution (DE) cryogenic separation and distillation theory. Energies 2014;7(5):3484–502.
method. Korean J Chem Eng 2011;1:246–51. [127] Ramdin M. CO2 Capture with ionic liquids: experiments and molecular simula-
[95] Zhang H, et al. Novel bare-bones particle swarm optimization and its performance tions. TU Delft: Delft University of Technology; 2015.
for modeling vapor–liquid equilibrium data. Fluid Phase Equilibria [128] Zhang X, et al. Carbon capture with ionic liquids: overview and progress. Energy
2011;301(1):33–45. Environ Sci 2012;5(5):6668–81.
[96] Gutkowski KI, Shariati A, Peters CJ. High-pressure phase behavior of the binary [129] Kumełan J, Tuma D, Maurer G. Simultaneous solubility of carbon dioxide and
ionic liquid system 1-octyl-3-methylimidazolium tetrafluoroborate + carbon di- hydrogen in the ionic liquid [hmim][Tf 2 N]: experimental results and correlation.
oxide. J Supercrit Fluids 2006;39:187–91. Fluid Phase Equilibria 2011;311:9–16.
[97] Gutkowski K, Shariati A, Peters C. High-pressure phase behavior of the binary ionic [130] Ramdin M, et al. Solubility of CO2 and CH4 in ionic liquids: ideal CO2/CH4 se-
liquid system 1-octyl-3-methylimidazolium tetrafluoroborate+ carbon dioxide. J lectivity. Ind Eng Chem Res 2014;53(40):15427–35.
Supercrit Fluids 2006;39(2):187–91. [131] Rosli A, et al. Advances in Liquid Absorbents for CO^ sub 2^ Capture: a Review. J
[98] Aki SN, et al. High-pressure phase behavior of carbon dioxide with imidazolium- Phys Sci 2017;28:121.
based ionic liquids. J Phys Chem B 2004;108(52):20355–65. [132] Burr B, Lyddon L. A comparison of physical solvents for acid gas removal. In:
[99] Camper D, et al. Room-temperature ionic liquid− amine solutions: tunable sol- Proceedings of the 87th annual gas processors association convention, Grapevine,
vents for efficient and reversible capture of CO2. Ind Eng Chem Res TX, March.; 2008.
2008;47(21):8496–8. [133] Xie Y, et al. Energy consumption analysis for CO2 separation using imidazolium-
[100] Lazzús JA. Thermodynamic modeling based on particle swarm optimization to based ionic liquids. Appl Energy 2014;136:325–35.
predict phase equilibrium of binary systems containing ionic liquids. J Mol Liq [134] Zhang Y, et al. Dual amino-functionalised phosphonium ionic liquids for CO2
2013;186:44–51. capture. Chem Eur J 2009;15(12):3003–11.
[101] Lazzús JA. Optimization of high-pressure vapor-liquid equilibrium modelling of [135] Gurkan BE, et al. Equimolar CO2 absorption by anion-functionalized ionic liquids.
binary mixtures (supercritical fluid + ionic liquid) by particle swarm algorithm. J Am Chem Soc 2010;132(7):2116–7.
Commun Math Comput Chem 2015;73:663–88. [136] Bates ED, et al. CO2 capture by a task-specific ionic liquid. J Am Chem Soc
[102] Lazzús JA. Neural network-particle swarm modeling to predict thermal properties. 2002;124(6):926–7.
Math Comput Model 2013;57(9):2408–18. [137] Zhang X, et al. Gas–liquid mass-transfer properties in CO2 absorption system with
[103] Klamt A, Eckert F, Arlt W. COSMO-RS: an alternative to simulation for calculating ionic liquids. AIChE J 2014;60(8):2929–39.
thermodynamic properties of liquid mixtures. Annu Rev Chem Biomol Eng [138] Bara JE, et al. Room-temperature ionic liquids and composite materials: platform
2010;1:101–22. technologies for CO2 capture. Acc Chem Res 2009;43(1):152–9.
[104] Diamantonis NI, et al. Evaluation of cubic, SAFT, and PC-SAFT equations of state [139] Yang J, et al. CO2 capture using amine solution mixed with ionic liquid. Ind Eng
for the vapor–liquid equilibrium modeling of CO2 mixtures with other gases. Ind Chem Res 2014;53(7):2790–9.
Eng Chem Res 2013;52(10):3933–42. [140] Zhao Y, et al. Density, viscosity, and performances of carbon dioxide capture in 16
[105] Ting PD, et al. Application of the PC-SAFT equation of state to asphaltene phase absorbents of amine+ ionic liquid+ H2O, ionic liquid+ H2O, and amine+ H2O
behavior, in Asphaltenes, heavy oils, and petroleomics. Springer; 2007. p. 301–27. systems. J Chem Eng Data 2010;55(9):3513–9.
[106] Yazdizadeh M, Rahmani F, Forghani AA. Thermodynamic modeling of CO2 solu- [141] Xu F, et al. Solubility of CO 2 in aqueous mixtures of monoethanolamine and
bility in ionic liquid ([C n-mim][Tf2N]; n = 2, 4, 6, 8) with using Wong-Sandler dicyanamide-based ionic liquids. Fluid Phase Equilibria 2014;365:80–7.
mixing rule, Peng-Rabinson equation of state (EOS) and differential evolution (DE) [142] Shahrom M, Wilfred C. Synthesis and thermal properties of amino acids ionic li-
method. Korean J Chem Eng 2011;28(1):246–51. quids (AAILS). J Appl Sci 2014;14(10):1067.
[107] Zhang X, Liu Z, Wang W. Screening of ionic liquids to capture CO2 by COSMO-RS [143] Fukumoto K, Yoshizawa M, Ohno H. Room temperature ionic liquids from 20
and experiments. AIChE J 2008;54(10):2717–28. natural amino acids. J Am Chem Soc 2005;127(8):2398–9.
[108] Perez-Salado Kamps A, et al. Solubility of CO2 in the ionic liquid [bmim][PF6]. J [144] Wang C, et al. Carbon dioxide capture by superbase-derived protic ionic liquids.
Chem Eng Data 2003;48(3):746–9. Angew Chem 2010;122(34):6114–7.
[109] Shiflett MB, Yokozeki A. Solubilities and diffusivities of carbon dioxide in ionic [145] Ma C, et al. Development of low-cost deep eutectic solvents for CO2 capture.
liquids:[bmim][PF6] and [bmim][BF4]. Ind Eng Chem Res 2005;44(12):4453–64. Energy Procedia 2017;142:3320–5.
[110] Shin E-K, Lee B-C, Lim JS. High-pressure solubilities of carbon dioxide in ionic [146] Adeyemi I, Abu-Zahra MR, Alnashef I. Novel green solvents for CO2 capture.
liquids: 1-alkyl-3-methylimidazolium bis (trifluoromethylsulfonyl) imide. J Energy Procedia 2017;114:2552–60.
Supercrit Fluids 2008;45(3):282–92. [147] Arshad MW. CO2 capture using Ionic Liquids. Lyngby, Denmark: Technical
[111] Carvalho PJ, et al. High carbon dioxide solubilities in trihexyltetradecylpho- University of Denmark, DTU, DK-2800 Kgs; 2009.
sphonium-based ionic liquids. J Supercrit Fluids 2010;52(3):258–65. [148] Sigma-Aldrich. 2017. Available from: 〈https://www.sigmaaldrich.com/〉.
[112] Kilaru PK, Scovazzo P. Correlations of low-pressure carbon dioxide and hydro- [149] Hallett JP. Tuning ionic liquids for low cost: applications in lignocellulose de-
carbon solubilities in imidazolium-, phosphonium-, and ammonium-based room- construction, In: Proceedings of the International Symposium on Tunable Solvents
temperature ionic liquids. Part 2. Using activation energy of viscosity. Ind Eng For Green Processing. InPROMPT 20162/3 June 2016 Berlin; 2016.
Chem Res 2008;47(3):910–9. [150] Chen L, et al. Inexpensive ionic liquids: [hso 4]−-based solvent production at bulk
[113] Anderson JL, Dixon JK, Brennecke JF. Solubility of CO2, CH4, C2H6, C2H4, O2, scale. Green Chem 2014;16(6):3098–106.
and N2 in 1-Hexyl-3-methylpyridinium Bis (trifluoromethylsulfonyl) imide: com- [151] Zhai H, Rubin ES. Systems analysis of ionic liquids for post-combustion CO 2
parison to other ionic liquids. Acc Chem Res 2007;40(11):1208–16. capture at coal-fired power plants. Energy Procedia 2014;63:1321–8.
[114] Goodrich BF, et al. Experimental measurements of amine-functionalized anion- [152] Wappel D, et al. Ionic liquids for post-combustion CO2 absorption. Int J Greenh
tethered ionic liquids with carbon dioxide. Ind Eng Chem Res 2010;50(1):111–8. Gas Control 2010;4(3):486–94.
[115] Galan-Sanchez LM. Functionalised ionic liquids, absorptionsolvents for CO2 and [153] Oliferenko AA, et al. Prediction of gas solubilities in ionic liquids. Phys Chem
olefin separation Ph.D. Thesis; 2008. Chem Phys 2011;13(38):17262–72.
[116] Brennecke JF, Maginn EJ. Ionic liquids: innovative fluids for chemical processing. [154] Mortazavi-Manesh S, Satyro MA, Marriott RA. Screening ionic liquids as candi-
AIChE J 2001;47(11):2384–9. dates for separation of acid gases: solubility of hydrogen sulfide, methane, and
[117] Freitas A, et al. Modeling vapor liquid equilibrium of ionic liquids+ gas binary ethane. AIChE J 2013;59(8):2993–3005.
systems at high pressure with cubic equations of state. Braz J Chem Eng [155] Maiti A. Theoretical screening of ionic liquid solvents for carbon capture.
2013;30(1):63–73. ChemSusChem 2009;2(7):628–31.
[118] Ren W, Sensenich B, Scurto AM. High-pressure phase equilibria of {carbon dioxide [156] Sairi NA, et al. Low pressure solubilities of CO2 in guanidinium
(CO2) + n-alkyl-imidazolium bis(trifluoromethylsulfonyl)amide} ionic liquids. J trifluoromethanesulfonate–MDEA systems. Fluid Phase Equilibria
Chem Thermodyn 2010;42:305–11. 2015;385:79–91.
[119] Brennecke JF, Maginn EJ. Ionic liquids: innovative fluids for chemical processing. [157] Ramdin M, de Loos TW, Vlugt TJH. State-of-the-Art of CO2 Capture with Ionic
Am Inst Chem Eng 2001;47(11):2384–9. Liquids. Ind Eng Chem Res 2012;51:8149–77.
[120] Shariati A, Gutkowski K, Peters CJ. Comparison of the phase behavior of some [158] Maia FM, et al. Modeling of ionic liquid systems: phase equilibria and physical
selected binary systems with ionic liquids. Am Inst Chem Eng properties. Ionic Liquids-New Aspects for the Future. InTech; 2013. p. 31–60.
2005;51(5):1532–40. [159] Danesh A. PVT and phase behaviour of petroleum reservoir fluids. 47. Elsevier;
[121] Cadena C, et al. Why is CO2 so soluble in imidazolium-based ionic liquids? J Am 1998.
Chem Soc 2004;126(16):5300–8. [160] Qi G, Wang Sh. Modeling solid-liquid equilibrium of the NH3-CO2-SO2-K2SO4-
[122] YUE Z-g. et al., Molecular simulation on microstructure of ionic liquids in capture H2O system and its application to combined capture of CO2 and SO2 using aqu-
of CO2. 过程工程学报. 11(4); 2011. eous ammonia. Energy Procedia 2017;114:834–9.
[123] Palma Chilla LO, Lazzús JA, Pérez Ponce AA. Particle swarm modeling of vapor- [161] Pearson P, Moscatelli J, Garg B, Cousins A, Feron PHM. An update on the devel-
liquid equilibrium data of binary systems containing CO2 + imidazolium ionic opment of the CSIRO's CS-Cap combined CO2 and SO2 capture process. Energy
liquids based on Bis[(trifluoromethyl)sulfonyl]imide anion. J Eng Thermophys Procedia 2017;114:1721–8.
2011;20(4):487–500. [162] Luo M, Yi Y, Wang Sh, Wang Z, Wang Q. Review of hydrogen production using
[124] Carroll JJ, Mather AE. The system carbon dioxide-water and the Krichevsky- chemical-looping technology. Renew Sustain Energy Rev 2018;81(2):3186–214.
Kasarnovsky equation. J Solut Chem 1992;21(7):607–21. [163] Pires JCM. COP21: The algae opportunity? Renew Sustain Energy Rev
[125] Carvalho PJ, et al. High carbon dioxide solubilities in trihexyltetradecylpho- 2017;79:867–77.
sphonium-based ionic liquids. J Supercrit Fluids 2010;52:258–65. [164] Brynolf S, Taljegard M, Grahn M, Hansson J. Electrofuels for the transport sector: a
[126] Xu G, et al. An improved CO2 separation and purification system based on review of production costs. Renew Sustain Energy Rev 2018;81(2):1887–905.

525