Download as pdf or txt
Download as pdf or txt
You are on page 1of 13

Correlation

Cite This: Ind. Eng. Chem. Res. 2019, 58, 4341−4353 pubs.acs.org/IECR

Modeling the Water Solubility in Imidazolium-Based Ionic Liquids


Using the Peng−Robinson Equation of State
Jeremías Martínez,†,‡ María A. Zúñiga-Hinojosa,§ and Ricardo Macías-Salinas*,§

Facultad de Química, Universidad Autónoma del Estado de México, Paseo Colón y Paseo Tollocan S/N, Toluca, Estado de México
C.P. 50120, México

Centro Conjunto de Investigación en Química Sustentable UAEM-UNAM, Carretera Toluca-Atlacomulco, km 14.5, Toluca,
Estado de México C.P. 50200, México
§
Instituto Politécnico Nacional, Departamento de Ingeniería Química, ESIQIE, Ciudad de México, Mexico City C.P. 07738, México
Downloaded via NATL TAIWAN UNIV SCIENCE & TECHLGY on August 2, 2023 at 03:56:53 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: Ionic liquids (ILs), which are also known as liquid salts or ionic
fluids, are organic salts with a low fusion point. They behave as a liquid at low
temperature or ambient temperature. Accordingly, they are prominent solvents
to be used in the green chemical processes, because of their attractive
physicochemical properties such as low vapor pressure, high thermal stability,
an excellent solvation behavior, and high gas solubility. Recently, various
experimental researches have reported the water solubility in different
imidazolium-based ILs at different temperature, pressure, and composition
conditions. In this study, we present the modeling of the vapor−liquid
equilibrium of the H2O−IL system using the Peng−Robinson cubic equation
of state (EoS), coupled with the Wong-Sandler mixing rules; eight binary
systems were studied for this purpose. In addition, improved temperature-
dependent parameters were introduced into the EoS as those proposed by
Stryjek and Vera for water [Can. J. Chem. Eng. 1986, 64, 323] and by Yokozeki
for the IL [Int. J. Thermophys. 2001, 22, 1057]. The studied ILs were [CxMIM][Cl] (x = 2, 4, 6), [C4MIM][PF6],
[C2MIM][BF4], [C4MIM][BF4], [OHC2MIM][BF4], and [OHC2MIM][Cl]. The obtained results showed a satisfactory
agreement between the experimental and the calculated solubility data using the present modeling approach under different
conditions of temperature, pressure, and composition.

■ INTRODUCTION
Ionic liquids (ILs) have recently attracted the attention within
This peculiar characteristic is due to a strong electrostatic
charge and a high tendency to create hydrogen bonds.11
ILs are organic salts formed by large and asymmetric organic
the research and industrial community, since they are
cations, e.g., imidazolium cation, bonded to alkyl groups, and
considered as new potential solvents for replacing the
organic and inorganic anions. Therefore, synthesizing a
traditional organic solvents. They present an extremely low significant amount of ILs for fitting some physicochemical
vapor pressure, which prevents their evaporation at room properties is possible.14,15 The physical and chemical proper-
conditions; accordingly, ILs are called “green solvents.”1,2 ties of ILs present many advantages over traditional solvents.
Other properties include their high thermal stability and their Accordingly, it is necessary to have a vast knowledge of ILs
high capacity for solubilizing several gases; in addition, they before overtaking in a process at the industrial scale. However,
remain in a liquid state over a wide range of temperatures. In to create an extensive phase equilibrium database of H2O−IL
this context, the vapor−liquid equilibrium (VLE) of the H2O− systems, it is necessary to perform an extensive experimental
IL system has potential applications in different separation investigation; consequently, this is time- and money-consum-
process, e.g., the extractive distillation (breaking azeo- ing. In this context, modeling is of vital importance, allowing
tropes),2−6 and absorption.7 ILs could be also used in one to calculate and predict physicochemical properties, and
refrigeration and air conditioning industries, in which the phase equilibrium behavior. Despite the complexity in
refrigerant−absorbent pairs such as H2O−LiBr and H2O− predicting the water solubility in the ionic liquids, several
NH38,9 are commonly used. Traditional refrigerant−absorbent thermodynamic models have been used, such as activity
pairs have some disadvantages such as crystal formation, coefficient models, equations of state (EoS), and unimolecular
toxicity, and flammability.10 Several studies have shown that
ILs can be used in the gas industry to prevent hydrate Received: October 18, 2018
formation, since they modify the dissociation equilibrium curve Revised: January 10, 2019
of gas hydrates. That is, ILs show a dual function: behaving Accepted: February 15, 2019
simultaneously as thermodynamic and kinetic inhibitors.7,11−13 Published: February 15, 2019

© 2019 American Chemical Society 4341 DOI: 10.1021/acs.iecr.8b05153


Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

quantum chemical calculations. Calvar et al.3 and Carvalho et ylimidazolium tetrafluoroborate [OHC2MIM][BF4]; and 1-(2-
al.16 used the Nonrandom Two-Liquid (NRTL) model to hydroxyethyl)-3-methylimidazolium chloride [OHC2MIM]-
correlate the experimental data of the H2O−[CxMIM][Cl] (x [Cl].
= 2, 4, 6) systems, obtaining absolute average deviation (AAD)
values of <1%. Whereas, Yokozeki and Shiflett17,18 used the
Redlich−Kwong EoS19 with the empirical modification in the
■ DESCRIPTION OF THE MODEL
The vapor−liquid phase behavior between water and IL was
temperature-dependent parameter of the pure component, and determined by the PR EoS,31 expressed as
the modified van der Waals-Berthelot mixing rule.20 Vetere21
calculated the critical properties of ILs by using a RT a(T )
P= −
compressibility factor of 0.256. The solubilities of the v−b v(v + b) + b(v − b) (1)
[C2MIM][BF4] and [C4MIM][BF4] ILs in water were
where a and b are the constants for pure components given by
predicted with AAD values of <0.02%.
The Statistical Associating Fluid Theory (SAFT) proposed RTc
a = 0.45723552821 α(T )
by Chapman et al.22 and Huan and Radosz23 in their variations Pc (2)
have also been used for several research groups to model the
H2O−IL solution. Passos et al.24 applied the PC-SAFT and
approach for calculating water solubility in diverse ILs using RTc
the cation [C4MIM+] as a base. In this case, ILs were b = 0.077960739039
Pc (3)
considered as associating species with two association sites.
The model parameters were calculated by simultaneous 35
The Soave correlation is defined as
adjustment to pure ILs densities, water activity coefficients at
T
298.15 K, and VLE experimental data at 0.1 MPa. Their results α(T ) = [1 + κ(1 − Tr 0.5)] , Tr ≡
showed a satisfactory agreement between the experimental and Tc (4)
calculated data obtaining AAD values between 1.01 and 7.85. where κ is only a function of the acentric factor ω. To model
Vega et al.25,26 modeled the water−IL system using the soft- the water, we use the Stryjek and Vera32 expression, in which
SAFT equation, where the IL was considered as a homonuclear the κ parameter was modified:
chain, i.e., cation and anion as one molecule with a specific
association number of sites; independently, the alkyl chain κ = κ0 + κ1(1 + Tr 0.5)(0.7 − Tr) (5)
length in imidazolium cation and the corresponding anion
were modeled. With regard to the equation parameters of the where
model for the system [CxMIM][Cl] (x = 2, 4, 6), the segment κ0 = 0.378893 + 1.4897153ω − 0.17131848ω2 + 0.0196554ω3
diameter parameter was set at an average value, whereas a (6)
variation in the chain length and the dispersive energy
κ1 has a specific value for pure components, i.e., when the
parameters were observed as the alkyl chain length of the
critical properties and the acentric factor of the components
cation increased. In this particular case, the AAD value
were known, the experimental vapor pressures were fitted to
obtained by Vega et al. was 0.1%.25,26 Freire et al.27 modeled
obtain the κ1 value (see Table 1). For water, the κ1 and ω
the H2O−[C4MIM][BF4] and H2O−[C4MIM][PF6] systems values are −0.06635 (ref 32) and 0.3438, (ref 36), respectively.
using the Conductor-like Screening for Real Solvents The following modification in the temperature-dependent a
(COSMO-RS) model; nevertheless, they only showed parameter of the pure component in the PR EoS17,18 was used
qualitative results. This model, based on the statistical for the eight ILs:20
i Tc T yzz
∑ βkjjjjj
thermodynamics, calculates the chemical potential of any

z
Tc zz{
species in any mixture and does not have a concentration ≤3 k

kT
dependency in the excess Gibbs free-energy function α (T ) = −
(UNIFAC).28 In contrast, Zhou et al.29 and Khan et al.30 k=0 (7)
obtained results with a better precision using COSMO-RS. The critical properties were taken from Valderrama et al. The 38
In the present work, we calculated the water solubility in authors calculated such properties using a group contribution
imidazolium-based ILs using the Peng−Robinson equation of method (see Table 1); the βk values presented in this work
state (PR EoS)31 at different temperature, pressure, and were fitted to PTx experimental solubility data of each binary
composition conditions. Regarding the proposed modeling system. We observed that β0 values approach to one; while, β2
approach for the H2O−IL systems, we incorporated in the PR and β3 values did not significantly affect the vapor−liquid
EoS the modification in the temperature-dependent parameter equilibrium calculations. So that, in the temperature dependent
a (cohesive energy) of the pure IL, as suggested by Yokozeki,20 correlation β1 was the only adjustable parameter, which will be
whereas, we included the α expression of Stryjek and Vera for further discussed. Therefore, we applied the equation proposed
water.32 Furthermore, we applied the Wong-Sandler (WS) by Yokozeki in solubility calculations.20
ij T T yz
modern mixing rules33 in combination with the UNIQUAC
α(T ) = 1 + β1jjj c − zzz
jT Tc z{
model34 for estimating the excess free energy. In this modeling
k
effort, the studied ILs were 1-ethyl-3-methylimidazolium
(8)
chloride, [C2MIM][Cl]; 1-butyl-3-methylimidazolium chlor-
ide, [C4MIM][Cl]; 1-hexyl-3-methylimidazolium chloride, On the other hand, Wong and Sandler33 proposed a set of
[C6MIM][Cl]; 1-butyl-3-methylimidazolium hexafluorophos- mixing rules to better represent the nonideal behavior of
phate, [C4MIM][PF6]; 1-ethyl-3-methylimidazolium tetra- mixtures such as the water−ILs system. These mixing rules
fluoroborate, [C2MIM][BF4]; 1-butyl-3-methylimidazolium combine an excess free-energy model with a cubic EoS to
tetrafluoroborate [C4MIM][BF4]; 1-(2-hydroxyethyl)-3-meth- produce the desired EoS behavior at both low and high
4342 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Table 1. Properties of Pure Components


component Tc [K] Pc [bar] r q q′
H2O 647.286a 220.8975a 0.920b 1.400b 1.000b
[C2MIM]Cl 748.6 34.17 12.089 10.488 7.489
[C4MIM]Cl 789.0 27.85 13.438 11.568 8.256
[C6MIM]Cl 829.2 23.50 14.787 12.648 9.031
[C4MIM][PF6] 719.4 17.28 17.405 11.006 7.858
[C2MIM][BF4] 596.2 23.59 13.808 8.550 6.105
[C4MIM][BF4] 643.2 20.38 15.157 9.630 6.876
[OHC2MIM][BF4] 691.7 24.67 14.582 9.442 6.742
[OHC2MIM]Cl 832.1 36.21 12.862 11.380 8.126
a
Poling et al.36 bPrausnitz et al.37

densities without being density-dependent. These rules allow where AEEoS represents the excess Helmholtz free energy
extrapolation over a wide range of temperature and pressure. derived from an EoS, whereas, the AEγ and GEγ expressions refer
WS mixing rules are based on two significant observations. The to excess free energies derived from an activity coefficient
first states that the van der Waals one-fluid mixing rules (eqs 9 model.
and 10) are sufficient conditions to ensure the proper In the WS mixing rules context, for the infinite pressure limit
composition dependence of the second virial coefficient (eq of an EoS, the molecules in the liquid solution are assumed to
11) and that they are not necessary conditions, since they be so tightly packed that there is no free volume, i.e., lim v =
impose constraints on the a and b parameters to satisfy eq 12. P →∞
b.
a= ∑ ∑ xixjaij Finally, from the above set of equations, Wong and
Sandler33 obtained the following expressions for am and bm:
i D yz
am = RT ·Q jjj
i j (9)

zz
∑ ∑ xixjbij k1 − D {
b= (16)
i j (10)
Q
B(x , T ) = ∑ ∑ xixjBij(T ) bm =
1−D (17)
i j (11)

i a y
∑ xixjjjjb − zzz
where

k RT {ij
a am
bm − m = ∑ Q = bm −
RT i j (12) RT (18)

i ai yz
∑ xijjjjj zz + G
z
z
To apply the combining rule (eq 12), an interaction parameter, E

k bi RT { C RT
kij, is introduced as follows: D=

ij a yz
(19)

jjb − zz =
ai aj
bi − ( ) + (bj − RT ) (1 − k )
k RT {ij
RT
ij
where C is an EoS-specific constant that is dependent on the
2 (13) used EoS; the C value for PR EoS is [ln( 2 − 1)]/ 2 .
To calculate the excess Gibbs free energy for water−IL
Note that the composition dependence of the second virial systems, the UNIQUAC equation34 was used:
coefficient, that is, the limit condition to low density, is
satisfied. Moreover, eq 12 does not provide any relationship to GE GE,combinatorial GE,residual
calculate the a and b parameters separately, but only for the = +
RT RT RT (20)
sum (b − a/RT); thus, an additional equation is needed.
The second observation indicates that the excess of the where the combinatorial term describes the dominant entropic
Helmholtz free energy of a mixture is less pressure dependent contribution (eq 21); which consists of two compositional
than the excess Gibbs free energy, i.e., variables: area fraction θ, and segment fraction ϕ, i.e.,
combinatorial part is determined by the composition, and
GE(T , P = low, xi) = AE(T , P = low, xi) molecular shape and size only. Moreover, z, is the coordination

ÅÄÅ É
= AE(T , P = ∞ , xi) number (equal to 10).
ij θ yzÑÑÑÑ
(14)
ji ϕ zy ji ϕ zy z ÅÅ ij θ yz
= x1 lnjjjj 1 zzzz + x 2 lnjjjj 2 zzzz + ÅÅÅÅx1q1 lnjjjj 1 zzzz + x 2q2 lnjjjj 2 zzzzÑÑÑÑ
k x 2 { 2 ÅÅÅÇ j ϕ1 z j ϕ2 zÑÑ
The first term comes from the fact that GE = AE + PvE, and the GE,combinatorial
PvE expression is negligible at low pressures; whereas, the RT k x1 { k { k {ÑÖ
second one is a result of the essential pressure independence of (21)
AE term at high densities. Accordingly, the second equation for
where the segment fraction, ϕ, and area fractions θ and θ’ are
a and b parameters comes from the condition
given by
E
AEoS (T , P = ∞ , x) = A γE(T , P = ∞ , x) xiqi
θi = i , j = 1, 2
= A γE(T , P = low, x) = GγE(T , P = low, x) (15) xiqi + xjqj (22)

4343 DOI: 10.1021/acs.iecr.8b05153


Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Table 2. Experimental H2O Solubilities in Different Ionic Liquids (ILs)


x-H2O
ionic liquids T [K] P [bar] maximum source
[C2MIM]Cl 355.58−409.3, 364.02−421.93, 373.66−435.03 0.5, 0.7, 1.0 0.998 Carvalho et al.16
[C4MIM]Cl 354.92−405.02, 363.49−414.97, 372.66−429.69 0.5, 0.7, 1.0 0.988 Carvalho et al.16
[C6MIM]Cl 354.64−383.26, 363.46−393.61, 373.13−405.3, 0.5, 0.7, 1.0, 1.01 0.998 Carvalho et al.,16
373.3−415.51 Calvar et al.41
[C4MIM][PF6] 283.1, 298.15, 323.15 0.002−0.010, 0.002−0.020, 0.005−0.050 0.151 Anthony et al.42
[C2MIM][BF4] 323.15, 328.15, 333.15 0.057−0.097, 0.073−0.124, 0.094−0.159 0.8 Han et al.43
373.15 0.025−0.756 0.778 Jork et al.5
[C4MIM][BF4] 308.17−443.65, 298.16−433.68, 285.13−433.92, 0.008−0.520, 0.013−1.228, 0.009−2.001, 0.582 Guan et al.44
284.95−433.28 0.011−3.192
[OHC2MIM][BF4] 392.6−464.8, 403.0−427.4, 361.6−402.2, 0.122−0.879, 0.483−1.008, 0.267−1.0, 0.8 Kim et al.45
325.8−383.8 0.092−0.988
[OHC2MIM]Cl 318.57−398.32, 314.82−391.60, 300.73−380.1, 0.033−0.805, 0.041−1.023, 0.028−1.011, 0.988 Nie et al.46
301.40−375.60, 301.87−373.56 0.036−1.008, 0.039−1.008

xiqi′ purposes of the H2O−IL systems. We present in, Table 2, the


θi′ = i , j = 1, 2 corresponding temperature and pressure ranges, the maximum
xiqi′ + xjq′j (23) H2O solubility measured, and the source of the experimental
xiri solubility data for all ILs considered in this study. As shown in
ϕi = i , j = 1, 2 Table 2, the temperature range is moderate (283.15−443.15
xiri + xjrj (24) K), whereas the pressure range is low (0.008−3.192 bar). The
Parameters r, q, and q’ are pure-component molecular maximum H2O solubility is reported for the H2O−IL−Cl−
structural constants, which are dependent on the molecular system at a water molar fraction of 0.998.
size and external surface areas. The UNIQUAC structural The simplex optimization procedure of Nelder and Mead47
parameters of the ionic species were obtained from Lei et al.,39 was used in the computations for searching the minimum of
by correlating the activity coefficients of solutes at infinite the following objective functions, according to the reported
dilution in ILs at different temperatures; whereas the r and q experimental data:
values for water were taken from Prausnitz et al.37 With regard N N
to the q′ values of both ILs and water, these were estimated by min f = ∑ [P jexp − P jcalc]2 + ∑ [1 − yHcalcO,j ]2
following expression: 0.7 × q.37 j=1 j=1
2
(27)
On the other hand, the residual term mainly describes the
intermolecular forces that are responsible for the enthalpy of or
mixing and is given by N N

G E,residual min f = ∑ [T jexp − T jcalc]2 + ∑ [1 − yHcalcO,j ]2


= −q1x1 ln(θ1 + θ2τ2,1) − q2x 2 ln(θ2 + θ1τ1,2) j=1 j=1
2
(28)
RT
(25)
where N is the number of experimental points, Pexp and Pcalc
where τ1,2 and τ2,1 are adjusted parameters, which are expressed are experimental and calculated pressures, Texp and Tcalc stand
as a function of characteristic energies ΔUi,j: for experimental and calculated temperatures, and ycalc
H2O is the
ΔUi , j equilibrium composition of H2O in the vapor phase. Note that,
ln τi , j = − i , j = 1, 2 in this modeling approach, only water is considered in the
RT (26)
vapor phase. In a first attempt to obtain the value of the β1,
In most of the cases, the above equation gives the primary ΔU 1,2 , ΔU2,1 , and k 12 parameters, we correlated the
effect of temperature on τi,j. This work aims to predict the experimental solubilities at 298 K, as suggested by Macias- ́
water solubility in ILs using the PR EoS coupled with WS 48
Salinas. However, we found that the experimental solubility
mixing rules. This modeling approach is restricted to β1 data of the H2O−[C4MIM][PF6] system was only reported at
parameter first; subsequently, to both UNIQUAC parameters 298 K in the literature. Therefore, we decided to use the
(ΔU1,2 and ΔU2,1) and finally, to the interaction parameter for solubilities at the lowest temperature, pressure, or H2O molar
the second virial coefficient k12 = k21. The above-mentioned fraction, according to the available experimental data. We
parameters were fitted to experimental solubility data. All found that the β1 value must be our priority, because of its
vapor−liquid equilibrium calculations were performed using reasonable representation of the cohesive energy (α) for ILs.
the phi-phi method. The required expressions for the fugacity Once optimized, β1 was set constant, and the ΔU1,2, ΔU2,1, and
coefficients of each species in the mixture, based on Wong− k12 parameters were fitted using the same initial values in each
Sandler approach, are given elsewhere.33,40 isotherm, isobar, or isopleth. Note that the ΔU1,2 and ΔU2,1

■ RESULTS AND DISCUSSION


The ILs considered in this work are based on the imidazolium
values showed similar values in each isobar of H2O−
[CxMIM][Cl] (x = 2, 4, 6) systems; therefore, we set either
ΔU1,2 or ΔU2,1 to optimize again the ΔU1,2 or ΔU2,1, and k12
cation bonded to the [Cl−], [PF−6 ], and [BF−4 ] anions. In this parameters. When we obtained the suitable β1, ΔU1,2, and
context, the anion plays a more significant role than the alkyl ΔU2,1 values, we set them constant, and finally, we fitted the
chain length; in addition, the shape and size of the anion k12 parameter. For purposes of modeling, we chose the
determine its miscibility extent in a solvent.5 For modeling appropriate values of the adjusted parameters based on a k12
4344 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

value of <1 and the lowest percent absolute average deviation Figure 1 shows the experimental and calculated solubilities
(AAD%) for each water solubility in IL calculation. Table 3 with the present modeling approach for H2O−[CxMIM][Cl]
(x = 2, 4, 6) systems at 0.5, 0.7, and 1 bar. In Figure 1a, a
Table 3. Adjusted Parameters with the PR-WS Model for reasonable agreement is observed between the experimental
H2O−[CxMIM][Cl] (x = 2, 4, 6) Systems and calculated temperatures using the 3-methylimidazolium
cation and the chloride anion, irrespective of the alkyl chain
ionic liquid, IL [C2MIM]Cl [C4MIM]Cl [C6MIM]Cl total
length of the cation and pressure. A characteristic, per se, of the
Na 93 81 123 297 WS approach, consists of reproducing the desired equation of
β1 10.27 0.78 10.07 state behavior at both low and high densities without being
ΔU12 [cal/mol] −1512.82 −842.77 −612.67 density-dependent and allows extrapolation over a wide range
ΔU21 [cal/mol] 1209.13 −422.43 115.16 of temperature and pressure, as shown in Figure 1. Carvalho et
k1,2 0.663 0.876 0.810 al.16 showed in the vapor−liquid equilibrium for H2O−
AAD, T% 0.14 0.22 0.18 [CxMIM][Cl] (x = 2, 4, 6) systems that, as the alkyl chain
a
N = number of experimental points. length increases in the cation, the boiling points decrease,
because of weak solvation of the cations. This phenomenon
can be explained through the UNIQUAC model. As
shows the correlating results of the model parameters β1, mentioned above, the residual term (eq 25) describes the
ΔU1,2, ΔU2,1, and k12, and the AAD% between experimental intermolecular forces that are responsible for the enthalpy of
and calculated temperatures of H2O−[CxMIM][Cl] (x = 2, 4, mixing; whereas, the combinatorial term describes the entropic
6) systems. As can be seen, the β1 value for H2O− contribution, which is determined by the composition,
[C4MIM][Cl] is smaller than the β1 values for H2O− molecular shape, and size. Although the UNIQUAC model
[C2MIM][Cl] and H2O−[C6MIM][Cl] systems. We observed was not developed to describe the behavior of mixtures
an increase in ΔU1,2 values, as the alkyl chain length increases containing electrolyte species, it can be applied in systems of
in the 1-alkyl-3-methylimidazolium cation; whereas ΔU2,1 water with IL solutions, since, in ILs, the ion charge is usually
values do not show any tendency. Table 3 also shows k12 dispersed, and the long-range electrostatic forces are weak,
values, which are >0.5. In addition, the predictive capability of compared with the short-range intermolecular forces, so that
the modeling approach to represent the experimental data of they can be neglected. Figure 2 presents the analysis of the
the water solubility in ILs with the Cl− anion is rather good, enthalpic and entropic contributions for H2O−[CxMIM][Cl]
because of the maximum AAD value of 0.22%. (x = 2, 4, 6) systems using the UNIQUAC model at 1 bar. As

Figure 1. Solubility of the systems with the PR-WS model: (a) H2O−[C2MIM][Cl], (b) H2O−[C4MIM][Cl] at 0.5, 0.7, and 1.0 bar, and (c)
H2O−[C6MIM][Cl] at 0.5, 0.7, 1.0, and 1.01 bar. Symbols represent experimental data, and solid lines represent calculated data.

4345 DOI: 10.1021/acs.iecr.8b05153


Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

is 0.75 for H2O−[C4MIM][Cl] system by using molecular


dynamics.
Regarding H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4]
systems, when we obtained the optimized β1 value, we only
observed in the H2O−[C4MIM][PF6] system similar values of
the ΔU2,1 parameter. In a first attempt, we set β1 and ΔU2,1
values and optimized ΔU1,2 and k12 parameters in each
isotherm to know the parameter variation, as a function of
temperature. However, we found that a great deal of scatter
existed among the adjusted ΔU1,2 parameters. This fact
notwithstanding, we correlated the ΔU1,2 and k12 parameters
to a straight-line equation, as a function of temperature. To
reduce the number of fitted parameters, after extensive testing
of solubility calculations, we set the k12 parameter constant at
the highest value (0.645). Figure 3 shows the behavior of the

Figure 2. Analysis of the entropic and enthalpic contributions for


H2O−[CxMIM][Cl] (x = 2, 4, 6) systems using the UNIQUAC
model at 1 bar.

can be seen in this figure, the water solubility in IL is


dependent on the enthalpic contribution noticeably. Fur-
thermore, all systems show an exothermic behavior,
irrespective of alkyl chain length of the imidazolium cation.
An examination of this figure also shows that, as the alkyl chain
length of cation increases, the enthalpic contribution increases,
while the entropic contribution decreases to a lesser extent. A
possible explanation is that the molecular interactions between
H2O and IL decrease as the size of IL increases, because of an
increase in the alkyl chain length cation, regardless of the Cl−
anion. Freire et al.27,49 showed that the decrease of the water
solubility in IL is due to the increase of hydrophobic character
of the cation, which is caused by the alkyl chain length. Figure 3. Fitted ΔU2,1 parameter, as a function of temperature for
Therefore, such behavior can be related to the decrease in the H2O−[C4MIM][PF6] system.
polarity of the IL, no matter the type of anion. According to
Zhang et al.,50 a smaller alkyl chain length of the cation results
in a stronger interaction between water and IL. It can be seen
in Figure 2 that, as the molar fraction of the H2O increases, the ΔU1,2 parameter, as a function of the temperature from the
enthalpic contributions decrease up to a minimum value, and H2O−[C4MIM][PF6] system at 283.15, 298.15, and 323.15 K.
subsequently, they approach zero, irrespective of alkyl chain This figure shows that the ΔU1,2 parameters become less
length of the cation. A maximum interaction exists between positive, and even negative as the temperature increases.
H2O and the ILs: [C 2 MIM][Cl], [C4MIM][Cl], and Therefore, we correlated them using the following straight-line
[C6MIM][Cl] at 0.69, 0.75, and 0.79 molar fraction of the equation:
H2O, respectively. This behavior suggests that, at the molar ΔU1,2 = c·T + d (29)
fractions above-mentioned, the interaction between H2O and
IL is 1:1. In addition, as the amount of water increases, the where c and d are the constants of the correlation, and T is the
H2O−H2O interactions are dominant, i.e., the solvation temperature. The adjusted parameters of the modeling
behavior is predominant. Khan et al.30 predicted the activity approach for the H2O−[C2MIM][BF4] system presented a
coefficient of water in ILs and the excess enthalpies using the similar behavior to the H2O−[C4MIM][PF6] system. The
COSMO-RS approach. This study consisted of setting the main difference was that the ΔU1,2 values were comparable in
[C4MIM+] cation; whereas the anion was varied, with one of each isotherm and the ΔU2,1 values did not show any
them being the Cl− anion. The results of this study showed tendency. Thus, we followed the above-mentioned approach
that the addition of H2O to binary solution reduces the for the fitting of the parameter. First, we fixed the value of the
electrostatic interaction between the [C4MIM+] cation and the ΔU1,2 parameter, and then the k12 parameter at the highest
Cl− anion, and both species are attracted to H2O. Moreover, value (0.553). Finally, we adjusted the ΔU2,1 parameter as a
the hydrogen of water presents a greater preference to bond function of temperature. The corresponding results are shown
with the lone pairs of electrons of the Cl− anion. Accordingly, in Figure 4, where it can be seen that all values are positive and
as the amount of water increases, the hydrogens of the they become less positive as the temperature increases. This
molecules of water cluster around Cl− anions up to a saturation figure also showed that the values decrease in a scattered
limit. The saturation limit obtained in this work is 0.75, similar manner as temperature increases. This fact notwithstanding,
to those reported in the literature. Niazi et al.51 showed that we correlated them according to the following straight-line
the maximum saturation limit of the molar fraction of the H2O expression, as a function of temperature:
4346 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Figure 4. Fitted ΔU2,1 parameter, as a function of temperature for


Figure 5. Solubility of the H2O−[C4MIM][PF6] system at 283.1,
H2O−[C2MIM][BF4] system.
298.15, and 323.15 K with the PR-WS model. Symbols represent
experimental data, and solid lines represent the calculated data.
ΔU2,1 = e·T + g (30)
where e and g are the constants of the correlation, and T is the
temperature. Table 4 presents the β1, ΔU1,2, ΔU2,1, and k12

Table 4. Fitted Model Parameters for H2O−[C4MIM][PF6]


and H2O−[C2MIM][BF4] Systems
ionic liquid coefficient [C4MIM][PF6] [C2MIM][BF4] total
Na 33 27 60
β1 4.26 13.79
ΔU12 [cal/mol c −5.10
d 1571.38 −793.89
ΔU21 [cal/mol] e −5.43
g 1439.82 3618.87
k12 0.645 0.554
AAD, P% 2.50 3.90
a
N = number of experimental points.

Figure 6. Solubility of the H2O−[C2MIM][BF4] system at 323.15,


values for H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4] 328.15, and 333.15 K with the PR-WS model. Symbols represent
systems. As seen in Table 4, the predictive capability of the experimental data, and solid lines represent the calculated data.
present approach for calculating the H2O solubility in ILs is
quite good.
Figure 5 shows the H2O solubility in [C4MIM][PF6] at positive, because of the stronger molecular interactions
283.1, 298.15, and 323.15 K; whereas, Figure 6 presents the between H2O and [C4MIM][PF6]. These molecular inter-
H2O solubility in [C2MIM][BF4] at 323.15, 328.15, and actions increase, in a linear manner, as the molar fraction of the
333.15 K. These figures show that the pressure increases as the H2O increases. The behavior of the enthalpic contributions is
molar fraction of the H2O increases. Both figures present a entirely different for the H2O−[C2MIM][BF4] system as
satisfactory agreement between experimental and calculated shown in Figure 9. This figure presents that all enthalpic
solubilities, even at the dilution region, as demonstrated in contributions are negative, and they decrease as the temper-
Figure 5. ature increases at a 0.4 molar fraction of the H2O; that is, the
Figures 7 and 8 depict the analysis of the entropic and molecular interactions between H2O and [C2MIM][BF4]
enthalpic contributions with the UNIQUAC model for H2O− increase as the temperature increases. It is important to note
[C4MIM][PF6] and H2O−[C2MIM][BF4] systems, respec- that, as the molar fraction of the H2O increases at 0.5, the
tively. As shown in both figures, the behavior of the entropic values of enthalpic contributions are similar; meaning that, the
contributions is the same at any temperature, because neither molecular interactions are equal, regardless of the temperature.
the size or shape of the ILs are different. However, the At molar fractions of >0.5, the enthalpic contributions begin to
behavior of the enthalpic contributions exhibits a variation as increase with temperature. This variation indicates that the
the molar fraction of the H2O is increased, as revealed in molecular interactions between H2O and [C2MIM][BF4] are
Figures 7 and 8. As can be seen in Figure 7, all enthalpic less favorable at higher temperatures. The latter behavior also
contributions are positive; this means that the behavior of shows that the effect of the composition is stronger than that of
these systems is endothermic. Also, it is shown that as temperature, since the slope of the straight line in the range
temperature increases, the enthalpic contributions are less from 0.4 to 0.7 molar fraction of the H2O does not change,
4347 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Figure 9. Analysis of the enthalpic contribution for H2O−[C2MIM]-


Figure 7. Analysis of the entropic and enthalpic contributions for [BF4] system at 323.15, 328.15, and 333.15 K, using the UNIQUAC
H2O−[C4MIM][PF6] system at 283.15, 298.15, and 323.15, K using model.
the UNIQUAC model. Dotted lines represent data from the entropic
contribution, and solid lines represent data from the enthalpic
contribution.

Figure 10. Analysis of the entropic and enthalpic contributions for the
H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4] systems at 323.15 K
and the H2O−[C2MIM][Cl] and H2O−[C4MIM][Cl] systems at 1
Figure 8. Analysis of the entropic and enthalpic contributions for bar, using the UNIQUAC model. Dotted lines represent data from
H2O−[C2MIM][BF4] system at 323.15, 328.15, and 333.15 K using the entropic contribution, and solid lines represent data from the
the UNIQUAC model. Dotted lines represent data from the entropic enthalpic contribution.
contribution, and solid lines represent data from the enthalpic
contribution.
enthalpic contributions are negative for the H2O−[C2MIM]-
[BF4] system, leading to more favorable interactions between
regardless of temperature. In Figure 9, it is also shown that H2O and [C2MIM][BF4] molecules. It is important to
there is little effect on the enthalpic contributions at H2O highlight that the binary solutions do not present a comparable
molar fractions of >0.7. According to Cammarata et al.,52 who range of molar fractions of the H2O, cation or anion, although
studied the H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4] both anions are fluorinated. However, when the H2O−
systems with the ATR-IR spectroscopy, the increase of the [C2MIM][BF4], and H2O−[C2MIM][Cl] systems are com-
H2O solubility in imidazolium-based ILs is dependent on the pared (Figure 10), it can be observed that the entropic
formation of the hydrogen bond between H2O and the anion contribution for the H2O−[C2MIM][BF4] system is greater
of IL. than for the H2O−[C2MIM][Cl] system. Figure 10 also shows
Figure 10 shows the entropic and enthalpic contributions for that the entropic contribution for the H2O−[C2MIM][BF4]
H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4] systems at system is greater than for the H2O−[C4MIM][Cl] system;
323.15 K and H2O−[C2MIM][Cl] and H2O−[C4MIM][Cl] even though the [C4MIM+] cation has a larger alkyl chain
systems at 1 bar. As can be seen, the entropic contributions are length. The enthalpic contributions are negative for the three
positive for H2O−[C4MIM][PF6] and H2O−[C2MIM][BF4] systems, what indicates an exothermic behavior and favorable
systems. Regarding the enthalpic contributions, they are molecular interactions. It can also be observed in Figure 10
positive for the H2O−[C4MIM][PF6] system, which indicates that the H2O solubility degree in the H2O−[C2MIM][Cl] is
its hydrophobic character in the H2O solubility; whereas the more significant than in the H2O−[C2MIM][BF4] system.
4348 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Figure 11. (a) Fitted ΔU2,1 parameter for the H2O−[C4MIM][BF4] system at H2O molar fractions of 0.112, 0.280, 0.398, and 0.582, and (b) fitted
ΔU2,1 parameter at H2O molar fractions of 0.112, 0.280, and 0.398 for the H2O−[C4MIM][BF4] system.

Figure 12. (a) Fitted ΔU2,1 parameter for the H2O−[OHC2MIM][BF4] system at H2O molar fractions of 0.2, 0.4, 0.6, and 0.8 , and (b) fitted
ΔU2,1 parameter for H2O−[OHC2MIM][BF4] system at H2O molar fractions of 0.2, 0.4, and 0.6.

Figure 13. (a) Fitted ΔU1,2 parameter for the H2O−[OHC2MIM][Cl] system at H2O molar fractions of 0.693, 0.795, 0.9, 0.955, and 0.988, and
(b) fitted ΔU1,2 parameter at H2O molar fractions of 0.693, 0.795, 0.9, and 0.955 for the H2O−[OHC2MIM][Cl].

According to Khan et al.,30 the increase in the size of the Freire et al.49 noted that the size or shape of anion has a more
halogenated anions causes a lesser interaction with H2O. considerable effect than the alkyl chain length of the
Furthermore, they showed that the fluorinated anions decrease imidazolium cation in the solubility of the above-mentioned
the interaction force between H2O and the IL. As evidenced in systems.
Figure 10, the UNIQUAC model was able to capture the size To obtain the fitted parameters of the modeling approach at
and shape effects of the anion (via entropic contribution), as constant composition for the H2O−[C4MIM][BF4], H2O−
well as the force in the molecular interactions between the [OHC2MIM][BF4], and H2O−[OHC2MIM][Cl] systems, we
H2O and IL, regardless of the type of anion. Jork et al.5 and followed the above-mentioned methodology. First, we
4349 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

Table 5. Fitted Model Parameters for the H2O−[C4MIM][BF4], H2O−[OHC2MIM][BF4], and H2O−[OHC2MIM][Cl]
Systems
ionic liquid coefficient [C4MIM][BF4] [OHC2MIM][BF4] [OHC2MIM][Cl] total
Na 45 35 48 128
β1 11.95 13.71 2.75
ΔU12 [cal/mol] h 2171.10
l 108.48 −757.65 −1778.30
ΔU21 [cal/mol] n 1952.74 265.16
o 432.67 −327.66 −48.05
k12 −0.517 0.748 0.591
AAD, P% 11.65 8.64 1.93
a
N = number of experimental points.

Figure 14. Solubility of the systems with the PR-WS model: (a) H2O−C4MIM][BF4] system at H2O molar fractions of 0.112, 0.280, 0.398, and
0.582 , (b) H2O−[OHC2MIM][BF4] system at H2O molar fractions of 0.2, 0.4, 0.6, and 0.8, and (c) H2O−[OHC2MIM][Cl] system at H2O
molar fractions of 0.693, 0.795, 0.9, 0.955, and 0.988. Symbols represent experimental data, and solid lines represent calculated data.

optimized the β1, ΔU1,2, ΔU2,1, and k12 parameters, and then and they increase as the molar fraction of the H2O increases.
we chose the β1 value at the lowest AAD, P%. Subsequently, Figure 11a shows that, at a molar fraction of 0.582, there is not
we set the β1 value constant, and ΔU1,2, ΔU2,1, and k12 a linear trend of the ΔU2,1 parameters. Figure 12 presents the
parameters were then again optimized. At this point, we ΔU2,1 values as a function of the molar fraction of the H2O for
observed that ΔU1,2 showed similar values for the H2O− the H2O−[OHC2MIM][BF4] system. Figure 12a shows that
[C4MIM][BF4] and H2O−[OHC2MIM][BF4] systems, so that all values are negative, and they became less negative as the
it was set constant for both systems. After that, we adjusted molar fraction of the H2O increases, except at 0.8. Figures 11a
ΔU2,1 and k12, and we observed a great deal of scatter among and 12a show that the H2O−[C4MIM][BF4] and H2O−
them; regardless of it, we correlated ΔU2,1 values to a straight- [OHC2MIM][BF4] systems show a deviation of the linear
line equation as a function of the molar fraction of the H2O, equation at the highest molar fraction of the H2O. This
and then we fitted the k12 parameter. Finally, we chose the behavior could be attributed to the solvation effect between
highest k12 value and set it constant to optimize the ΔU2,1 the H2O and the [BF−4 ] anion, which has been evidenced as
parameter again. Figure 11 shows the fitted ΔU2,1 values, as a the amount of water is increased.30 Since this modeling
function of the molar fraction of the H2O for the H2O− approach only takes into account the short-range interactions,
[C4MIM][BF4] system. As can be seen, all values are positive, we decide to exclude the obtained ΔU2,1 values at the highest
4350 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

molar fraction of H2O. A similar procedure for the fitting of


parameters was followed for the H2O−[OHC2MIM][Cl]
system. The main difference is that we set the ΔU2,1 parameter
constant and adjusted the ΔU1,2 parameter, as a function of the
molar fraction of the H2O. Figure 13a shows the fitted ΔU1,2
values as a function of the molar fraction of the H2O. It is
observed that the ΔU1,2 values increase linearly from negative
to positive, when the molar fraction of the H2O is increased,
except at 0.988. We suggested that the presence of the
hydroxyl group in the cation and the high concentration of
H2O, regardless of the size of the anion, produce such
behavior. However, the proposed model was not able to
describe this trend, because it includes only the short-range
interactions. Therefore, we fitted the ΔU1,2 values to a linear
equation without taking into account the last point, as shown
in Figure 13b. The following straight-line equation was used to
correlate ΔU1,2 and ΔU2,1 parameters as a function of the Figure 15. Analysis of the entropic and enthalpic contributions using
composition: the UNIQUAC model for the H2O−[C4MIM][BF4], H2O−
[OHC2MIM][BF4], and [OHC2MIM][Cl] systems at H2O molar
ΔU1,2 = h·x + l (31) fractions of 0.582, 0.6, and 0.693, respectively. Symbols represent data
from the entropic contribution, and solid lines represent data from the
ΔU2,1 = n·x + o (32) enthalpic contributions.

where h, l, n, and o are the correlation constants and x is the


molar fraction of the H2O. The correlating results of the β1, [C4MIM][BF4] system; this behavior shows that the molecular
ΔU1,2, ΔU2,1, and k12 parameters and the AAD, P% are listed in interactions are unfavorable between H2O and [C4MIM]-
Table 5. This table shows that the AAD, P% values are high for [BF4]. An examination of this figure also shows a poor
the H2O−[C4MIM][BF4] and H2O−[OHC2MIM][BF4] variation in the values of the enthalpic contributions for all
systems: 11.65% and 8.64%, even though the ΔU1,2 parameter systems, especially in the H2O−[OHC2MIM][Cl] system.
was correlated as a function of the molar fraction of H2O. Thus, the behavior of the entropic and enthalpic contributions
Figure 14 shows the experimental and calculated solubilities are not well represented by the UNIQUAC model using
experimental solubility data at constant composition.


with the modeling approach for the H2O−[C4MIM][BF4]
system at H2O molar fractions of 0.112, 0.280, 0.398, and
0.582, the H2O−[OHC2MIM][BF4] system at H2O molar CONCLUSIONS
fractions of 0.2, 0.4, 0.6, and 0.8, and the H2O−[OHC2MIM]- A formal modeling approach was devised here to calculate and
[Cl] at H2O molar fractions of 0.693, 0.795, 0.900, 0.955, and predict the H2O solubility in imidazolium-based ILs. It made
0.988. Figures 14a and 14b show that there exist a poor use of the Peng−Robinson cubic equation of state (EoS) and
agreement between experimental and calculated solubilities at the Wong−Sandler mixing rules under different conditions of
the highest isopleths. A possible explanation of these temperature, pressure, and composition. The following
discrepancies is that the solvation interactions are not taken conclusions can be drawn from this work:
into account in the modeling approach. These solvation forces (1) The agreement between experimental and calculated
are present as the concentration of water increase, especially in solubility data was remarkably good when the anion of
systems with bulky anions, e.g., [PF−6 ] and [BF−4 ]. On the other the ILs are small, such as halides. However, the model
hand, Figure 14c shows good agreement between the requires the use of temperature-dependent ΔU values
experimental and calculated solubilities, probably because the for improving H2O−solubility estimations when the ILs
size of the halide ion, such as Cl−, is smaller than the bulky present bulky anions or hydroxyl groups in cations.
anions. (2) Results of the comparison showed a poor agreement
Figure 15 depicts entropic and enthalpic contributions for between the experimental data and the calculated values
the H2O−[C4MIM][BF4], H2O−[OHC2MIM][BF4], and when the IL exhibits (i) a bulky anion, and (ii) a
[OHC2MIM][Cl] systems at H2O molar fractions of 0.582, hydroxyl group in the cation.
0.6, and 0.693, respectively. Although it is not valid to compare (3) The UNIQUAC model was able to predict the behavior
such systems, because they present different compositions, we of the molecular interactions (enthalpic contribution) at
only exhibited them to understand the behavior of their constant temperature and pressure, as well as the effect
contributions as a function of the molar composition of the of the alkyl chain length in cation and the size of the
H2O. As can be seen in this figure, the entropic contributions anion (entropic contribution).


are represented by a point, because the molar fractions are
constant. The entropic contributions for the binary systems AUTHOR INFORMATION
with the anion [BF−4 ] are positive, whereas the solution with
Corresponding Author
the Cl− anion is negative. On the other hand, the enthalpic
*E-mail: rms@ipn.mx.
contributions are described by solid lines. These contributions
are negative for the H2O−[OHC2MIM][BF4] and H2O− ORCID
[OHC2MIM][Cl] systems, which indicate that the molecular Ricardo Macías-Salinas: 0000-0002-0372-8190
interactions between H2O and the IL are favorable. However, Notes
the enthalpic contributions are positive for the H2O− The authors declare no competing financial interest.
4351 DOI: 10.1021/acs.iecr.8b05153
Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

■ ACKNOWLEDGMENTS
J.M. and M.A.Z.-H. gratefully acknowledge the National
exp = experimental value
residual = residual property
Council for Science and Technology of Mexico (CONACyT)
for providing financial support for this work. R.M.S. is also
grateful for the financial support provided by the Instituto
■ REFERENCES
(1) Marsh, K. N.; Boxall, J. A.; Lichtenthaler, R. Room Temperature
Politécnico National during the realization of this work.


Ionic Liquids and Their Mixtures: A Review. Fluid Phase Equilib.
2004, 219, 93.
LIST OF SYMBOLS (2) Quijada-Maldonado, E.; Meindersma, G. W.; de Haan, A. B.
a = attractive parameter in the cubic equation of state Ionic Liquid Effects on Mass Transfer Efficiency in Extractive
A = Helmholtz free energy Distillation of Water-Ethanol Mixtures. Comput. Chem. Eng. 2014, 71,
210.
b = co-volume parameter in the cubic equation of state (3) Calvar, N.; González, B.; Gómez, E.; Domínguez, A. Vapor-
B = as defined in eq 11 Liquid Equilibria for the Ternary System Ethanol+Water+1-Butyl-3-
c = as defined in eq 29 methylimidazolium Chloride and the Corresponding Binary Systems
C = constant specific to the cubic equation of state at 101.3 kPa. J. Chem. Eng. Data 2006, 51, 2178.
d = as defined in eq 29 (4) Seiler, M.; Jork, C.; Kavarnou, A.; Arlt, A.; Hirsch, A. Separation
D = as defined in eq 19 of Azeotropic Mixtures Using Hyperbranched Polymers or Ionic
e = as defined in eq 30 Liquids. AIChE J. 2004, 50, 2439.
f = objective function (5) Jork, C.; Seiler, M.; Beste, Y.-A.; Arlt, W. Influence of Ionic
g = as defined in eq 30 Liquids on the Phase Behavior of Aqueous Azeotropic Systems. J.
G = Gibbs free energy Chem. Eng. Data 2004, 49, 852.
h = as defined in eq 31 (6) Lei, Z.; Dai, C.; Zhu, J.; Chen, B. Extractive Distillation with
Ionic Liquids: A Review. AIChE J. 2014, 60, 3312.
k = interaction parameter in the cubic equation of state (7) Krannich, M.; Heym, F.; Jess, A. Characterization of Six
l = as defined in eq 31 Hygroscopic Ionic Liquid with Regard to Their Suitability for Gas
n = as defined in eq 32 Dehydration: Density, Viscosity, Thermal and Oxidative Stability,
N = number of data points Vapor Pressure, Diffusion Coefficient, and Activity Coefficient of
o = as defined in eq 32 Water. J. Chem. Eng. Data 2016, 61 (3), 1162.
P = pressure (8) Seiler, M.; Kühn, A.; Ziegler, F.; Wang, X. Sustainable Cooling
q = surface area parameter in UNIQUAC Strategies Using New Chemical System Solutions. Ind. Eng. Chem. Res.
q′ = prime surface area parameter in UNIQUAC 2013, 52, 16519.
Q = as defined in eq 18 (9) Oliveira, M. B.; Crespo, E. A.; Llovell, F.; Vega, L. F.; Coutinho,
r = volume parameter in UNIQUAC J. A. P. Modeling the Vapor-Liquid Equilibria and Water Activity
R = universal gas constant Coefficients of Alternative Refrigerant-Absorbent Ionic Liquid Water
Pairs for Absorption Systems. Fluid Phase Equilib. 2016, 426, 100.
T = temperature (10) Sun, J.; Fu, L.; Zhang, S. A Review of Working Fluids of
v = molar volume Absorption Cycles. Renewable Sustainable Energy Rev. 2012, 16, 1899.
x = number of alkyl groups (11) Xiao, C.; Adidharma, H. Dual Function Inhibitors for Methane
x = liquid mole fraction Hydrate. Chem. Eng. Sci. 2009, 64, 1522.
y = gas mole fraction (12) Xiao, C.; Wibisono, N.; Adidharma, H. Dialkylimidazolium
z = coordination number Halide Ionic Liquids as Dual Function Inhibitors for Methane
Subscripts Hydrate. Chem. Eng. Sci. 2010, 65, 3080.
(13) Richard, A. R.; Adidharma, H. The Performance of Ionic
c = critical property
Liquids and Their Mixtures in Inhibiting Methane Hydrate
EoS = equation of state Formation. Chem. Eng. Sci. 2013, 87, 270.
i, j = i−j pair interaction (14) Weingärtner, H. Understanding Ionic Liquids at the Molecular
m = mixture property Level: Facts, Problems, and Controversies. Angew. Chem., Int. Ed.
r = reduced property 2008, 47, 654.
1 = pertains to H2O (15) Chiappe, C.; Pieraccini, D. Ionic Liquids: Solvent Properties
2 = pertains to ionic liquid and Organic Reactivity. J. Phys. Org. Chem. 2005, 18, 275.
Greek Letters (16) Carvalho, P. J.; Khan, I.; Morais, A.; Granjo, J. F. O.; Oliveira,
N. M. C.; Santos, L. M. N. B. F.; Coutinho, J. A. P. A New
α = temperature-dependent correlation of the parameter α
Microebulliometer for the Measurement of the Vapor-Liquid
β0, β1, β2, β3 = fitted parameters to PTx experimental Equilibrium of Ionic Liquid Systems. Fluid Phase Equilib. 2013, 354,
solubility data 156.
ΔU = energy interaction parameter in UNIQUAC (17) Yokozeki, A.; Shiflett, M. B. Water Solubility in Ionic Liquids
γ = activity coefficient and Application to Absorption Cycles. Ind. Eng. Chem. Res. 2010, 49,
ϕ = segment fraction 9496.
κ0, κ1 = parameters suggested by Stryjek and Vera32 (18) Shiflett, M. B.; Yokozeki, A. Solubilities and Diffusivities of
θ = area fraction Carbon Dioxide in Ionic Liquids: [bmim][PF6] and [bmim][BF4].
θ′ = prime area fraction Ind. Eng. Chem. Res. 2005, 44, 4453.
τ = dimensionless energy parameter in UNIQUAC (19) Redlich, O.; Kwong, J. N. S. On the Thermodynamics of
ω = acentric factor Solutions. V. An Equation of State. Fugacities of Gaseous Solutions.
Chem. Rev. 1949, 44, 233.
Superscripts (20) Yokozeki, A. Solubility of Refrigerants in Various Lubricants.
calc = calculated value Int. J. Thermophys. 2001, 22, 1057.
combinatorial = combinatorial contribution (21) Vetere, A. Again the Rackett Equation. Chem. Eng. J. 1992, 49,
E = excess property 27.

4352 DOI: 10.1021/acs.iecr.8b05153


Ind. Eng. Chem. Res. 2019, 58, 4341−4353
Industrial & Engineering Chemistry Research Correlation

(22) Chapman, W. G.; Gubbins, K. E.; Jackson, G.; Radosz, M. New (42) Anthony, J. L.; Anderson, J. L.; Maginn, E. J.; Brennecke, J. F.
Reference Equation of State for Associating Liquids. Ind. Eng. Chem. Anion Effects on Gas Solubility in Ionic Liquids. J. Phys. Chem. B
Res. 1990, 29, 1709. 2005, 109, 6366.
(23) Huang, S. H.; Radosz, M. Equation of State for Small, Large, (43) Han, J.; Lei, Z.; Dai, C.; Li, J. Vapor Pressure Measurements for
Polydisperse, and Associating Molecules. Ind. Eng. Chem. Res. 1990, Binary Mixtures Containing Ionic Liquid and Predictions by the
29, 2284. Conductor-like Screening Model for Real Solvents. J. Chem. Eng. Data
(24) Passos, H.; Khan, I.; Mutelet, F.; Oliveira, M. B.; Carvalho, P. 2016, 61, 1117.
J.; Santos, L. M. N. B. F.; Held, C.; Sadowski, G.; Freire, M. G.; (44) Guan, T. T.; Sun, L.; Huangfu, L. X.; Guo, K. H. Experiment on
Coutinho, J. A. P. Vapor-Liquid Equilibria of Water+Alkylimidazo- Vapor-Liquid Phase Equilibrium of [BMIM]BF4 + H2O system.
lium-Based Ionic Liquids: Measurements and Perturbed-Chain Diwen Wuli Xuebao 2011, 33, 194.
Statistical Associating Fluid Theory Modeling. Ind. Eng. Chem. Res. (45) Kim, K.-S.; Park, S.-Y.; Choi, S.; Lee, H. Vapor Pressures of the
1-butyl-3-methylimidazolium Bromide+Water, 1-Butyl-3-methylimi-
2014, 53, 3737.
dazolium Tetrafluoroborate+Water, and 1-(2-Hydroxyethyl)-3-meth-
(25) MacDowell, N.; Llovell, F.; Sun, N.; Hallett, J. P.; George, A.;
ylimidazolium Tetrafluoroborate+Water Systems. J. Chem. Eng. Data
Hunt, P. A.; Welton, T.; Simmons, B. A.; Vega, L. F. New 2004, 49, 1550.
Experimental Density Data and Soft-SAFT Models of Alkylimidazo- (46) Nie, N.; Zheng, D.; Dong, L.; Li, Y. Thermodynamic Properties
lium ([CnC1im]+) Chloride (Cl−), Methylsulfate ([MeSO4]−), and of the Water + 1-(2-Hydroxyethyl)-3-methylimidazolium Chloride
Dimethylphosphate ([Me2PO4]−) Based Ionic Liquids. J. Phys. Chem. System. J. Chem. Eng. Data 2012, 57, 3598.
B 2014, 118, 6206. (47) Nelder, J. A.; Mead, R. A Simplex Method for Function
(26) Vega, L. F.; Llovell, F. Review and New Insights into the Minimization. Comput. J. 1965, 7, 308.
Application of Molecular-Based Equations of State to Water and (48) Macías-Salinas, R.; Chávez-Velasco, J. A.; Aquino-Olivos, M. A.;
Aqueous Solutions. Fluid Phase Equilib. 2016, 416, 150. Mendoza de la Cruz, J. L.; Sánchez-Ochoa, J. C. Accurate Modeling of
(27) Freire, M. G.; Ventura, S. P. M.; Santos, L. M. N. B. F.; CO2 Solubility on Ionic Liquids Using a Cubic EoS. Ind. Eng. Chem.
Marrucho, I. M.; Coutinho, J. A. P. Evaluation of COSMO-RS for the Res. 2013, 52, 7593.
Prediction of LLE and VLE of Water and Ionic Liquids Binary (49) Freire, M. G.; Santos, L. M. N. B. F.; Fernandes, A. M.;
Systems. Fluid Phase Equilib. 2008, 268, 74. Coutinho, J. A. P.; Marrucho, I. M. An Overview of the Mutual
(28) Banerjee, T.; Singh, M. K.; Khanna, A. Prediction of Binary Solubilities of Water-Imidazolium-Based Ionic Liquids Systems. Fluid
VLE for Imidazolium Based Ionic Liquids Systems Using COSMO- Phase Equilib. 2007, 261, 449.
RS. Ind. Eng. Chem. Res. 2006, 45, 3207. (50) Zhang, L.; Han, J.; Wang, R.; Qiu, X.; Ji, J. Isobaric Vapor-
(29) Zhou, T.; Chen, L.; Ye, Y.; Chen, L.; Qi, Z.; Freund, H.; Liquid Equilibria for Three Ternary Systems: Water+2-Propanol+1-
Sundmacher, K. An Overview of Mutual Solubility of Ionic Liquids Ethyl-3-methylimidazolium Tetrafluoroborate, Water+1-Propanol+1-
and Water Predicted by COSMO-RS. Ind. Eng. Chem. Res. 2012, 51, Ethyl-3-methylimidazolium Tetrafluoroborate, and Water+1-Propanol
6256. +1-Butyl-3-methylimidazolium Tetrafluoroborate. J. Chem. Eng. Data
(30) Khan, I.; Kurnia, K. A.; Mutelet, F.; Pinho, S. P.; Coutinho, J. A. 2007, 52, 1401.
P. Probing the Interactions between Ionic Liquids and Water: (51) Niazi, A. A.; Rabideau, B. D.; Ismail, A. E. Effects of Water
Experimental and Quantum Chemical Approach. J. Phys. Chem. B Concentration on the Structural and Diffusion Properties of
Imidazolium-Based Ionic Liquids-Water Mixtures. J. Phys. Chem. B
2014, 118, 1848.
2013, 117, 1378.
(31) Peng, D. Y.; Robinson, D. B. A New Two-Constant Equation of
(52) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T.
State. Ind. Eng. Chem. Fundam. 1976, 15, 59.
Molecular States of Water in Room Temperature Ionic Liquids. Phys.
(32) Stryjek, R.; Vera, J. H. PRSV: An Improved Peng-Robinson
Chem. Chem. Phys. 2001, 3, 5192.
Equation of State for Pure Compounds and Mixtures. Can. J. Chem.
Eng. 1986, 64, 323.
(33) Wong, D. S. H.; Sandler, S. I. A. Theoretically Correct Mixing
Rule for Cubic Equations of State. AIChE J. 1992, 38, 671.
(34) Abrams, D. S.; Prausnitz, J. M. Statistical Thermodynamics of
Liquid Mixtures: A New Expression for the Excess Gibbs Energy of
Partly or Completely Miscible Systems. AIChE J. 1975, 21, 116.
(35) Soave, G. Equilibrium Constants from a Modified Redlich-
Kwong Equation of State. Chem. Eng. Sci. 1972, 27 (6), 1197.
(36) Poling, B. E.; Prausnitz, J. M.; O’Connell, J. P. The Properties of
Gases and Liquids; McGraw−Hill: New York, 2001.
(37) Prausnitz, J.; Anderson, T.; Grens, E.; Eckert, C.; Hsieh, R.;
O’Connell, J. Computer Calculations for Multicomponent Vapor−Liquid
and Liquid−Liquid Equilibria; Prentice-Hall International Series in the
Physical and Chemical Engineering Sciences; Prentice-Hall: Engle-
wood Cliffs, NJ, 1980.
(38) Valderrama, J. O.; Forero, L. A.; Rojas, R. E. Extension of a
Group Contribution Method To Estimate the Critical Properties of
Ionic Liquids of High Molecular Mass. Ind. Eng. Chem. Res. 2015, 54,
3480.
(39) Lei, Z.; Zhang, J.; Li, Q.; Chen, B. UNIFAC Model for Ionic
Liquids. Ind. Eng. Chem. Res. 2009, 48, 2697.
(40) Orbey, H.; Sandler, S. I. Modeling Vapor−Liquid Equilibria:
Cubic Equations of State and Their Mixing Rules. Ed.; Cambridge
University Press: New York, 1998.
(41) Calvar, N.; González, B.; Gómez, E.; Domínguez, A. Study of
the Behaviour of the Azeotropic Mixture Ethanol-Water with
Imidazolium-Based Ionic Liquids. Fluid Phase Equilib. 2007, 259, 51.

4353 DOI: 10.1021/acs.iecr.8b05153


Ind. Eng. Chem. Res. 2019, 58, 4341−4353

You might also like