This article is licensed under CC-BY 4.

pubs.acs.org/jced Article

Liquid−Liquid Equilibrium Studies for Potential Solvents/Cosolvents

for the Separation of Aromatics and Alkanes
Blessing Mcebo Paile, Caleb Narasigadu, and Naadhira Seedat*

Cite This: https://doi.org/10.1021/acs.jced.4c00614 Read Online

ACCESS Metrics & More Article Recommendations

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: Liquid−liquid equilibria (LLE) data for ternary systems of n-heptane +

a solvent mixture + toluene were measured at 298.15 and 313.15 K and 101.3 kPa.
The solvents investigated are mixtures of N-methylpyrrolidone and glycerol, with
ratios of N-methylpyrrolidone to glycerol of 90:10, 70:30, and 50:50. LLE
Downloaded via 201.68.16.53 on May 12, 2025 at 13:35:45 (UTC).

measurements were carried out using the direct analytical method, and the
composition analysis was performed using gas chromatography. The effectiveness of
the solvent mixtures of N-methylpyrrolidone and glycerol in extracting toluene from
n-heptane was evaluated by determining the solvent selectivity and capacity and
evaluating the extent of the two-phase region. All systems investigated were found to
exhibit type I LLE behavior. The binodal curves were correlated to the Hlavatý β
function and log γ equations. The tie-lines were correlated with the NRTL and
UNIQUAC activity coefficient models. A comparison of the root-mean-square
deviation (rmsd) showed that the NRTL model provided a better correlation for all
systems.

1. INTRODUCTION is separated via extractive distillation, and azeotropic

Aromatic compounds are essential chemicals in the petro- distillation is used to separate naphtha with aromatic content
chemical industries, and this is due to their wide range of greater than 90 wt %.4,6 Naphtha with aromatic content lower
applications in the production of several petrochemical than 20 wt % exists but is technologically challenging to
products, which include plastics and fibers, detergents, separate.7 Extractive distillation and liquid−liquid extraction
fertilizers, explosives, pesticides, food and feed additives, require the inclusion of a solvent to alter the relative volatility
nutraceuticals, and pharmaceuticals and many other important of the mixture and create partitioning, respectively. The
substances.1−4 Aromatic compounds constitute about 30% of 8 commonly used organic solvents, which are usually polar, are
million known organic compounds; among these compounds sulfolane and glycols, while others include ethylene glycols
are benzene, toluene, ethylbenzene, and xylene isomers dimethylformamide, N-formyl morpholine, and N-methyl
(BTEX), which are often coproduced along other aromatic pyrrolidine.6 Desirable characteristics of ionic liquids, such as
and nonaromatic compounds during petroleum refining and low vapor pressure, wide liquid range, and relatively high
coal conversion.5 The major sources of BTEX are pyrolysis density when compared with aromatic and aliphatic com-
gasoline, reformer gasoline, and coke. The compositions of pounds, placed them as alternatives to organic solvents. Ionic
these sources vary for different feedstocks; typical reformer liquids that have been tested include imidazolium, pyridinium,
gasoline has a composition of about 25% toluene and 45% ammonium, and pyrrolidinium, which have also been
nonaromatic compounds.3,5 There is a need to separate BTEX employed as solvents in the extraction of aromatic com-
from other constituents; such separation often involves pounds.8 The major drawback of ionic liquids as solvents for
aromatic and nonaromatic compounds, and the compositions liquid−liquid extraction is the difficulty in finding ionic liquids
usually determine the method of separation. Due to relative with high selectivity, low toxicity, low viscosity, and high
volatilities near unity and azeotrope formations between capacity.9
aromatic and aliphatic compounds, conventional separation
processes like distillation are not practical, Rather, they can be
separated by liquid−liquid extraction, extractive distillation, or Received: October 27, 2024
azeotropic distillation, the composition of the aromatic Revised: April 15, 2025
compound in the mixtures often determines the type of Accepted: April 22, 2025
separation applied.6 Liquid−liquid ionic extraction of aromatic
compounds with ionic solvent has been applied for 20−65 wt
% aromatic content, 65−90 wt % of aromatic content naphtha
© XXXX The Authors. Published by
American Chemical Society https://doi.org/10.1021/acs.jced.4c00614
A J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

There have been several applications of organic solvents for Table 1. Chemical Purity Analysis at 101.3 kPa
the extraction of aromatic compounds. Alkhaldi et al.10
Refractive Index (nD)a
represented the Kuwait middle distillate with a multi-
component mixture with dodecane, hexadecane, 1,3,5- CAS Mass
Reg. Chemical fraction
trimethylbenzene (mesitylene), and butylbenzene as its Chemical No. supplier purityb Measuredc Lit
constituents, and N-methyl-2-pyrrolidone (NMP) was used
Glycerol 56-81-5 Sigma- >0.99 1.4736 1.473021
as an organic solvent for the liquid−liquid extraction process. Aldrich
Sulfolane was used as a solvent for the extraction of toluene in NMP 872- Sigma- >0.99 1.4689 1.469021
a mixture of toluene and heptane by Meindersma et al.11 50-4 Aldrich
DongChu et al.12 carried out the liquid−liquid equilibrium of Toluene 108- Glassworld >0.99 1.4963d 1.496722
the ternary mixture of heptane, xylene, and N-formylmorpho- 88-3
line, this is to determine the ability of N-formylmorpholine to n- 142- Sigma- >0.99 1.3853 1.385522
Heptane 82-5 Aldrich
extract xylene. Mohsen-Nia et al.13 used each of tetramethylene a
Refractive index at 298.15 K, u(T) = 0.1 K. bAs stated by supplier
sulfone, dimethyl sulfoxide (DMSO), and ethylene carbonate
standard uncertainties are. cu(nD) = 0.0001. dRefractive index at
solvents for the extraction of toluene or m-xylene from its 293.15 K, u(T) = 0.1 K.
mixtures with n-heptane or n-octane or cyclohexane.
Mesquita et al.14 evaluated 2-hydroxy ethylammonium
formate (2-HEAF) and glycerol as solvents for extracting by Narasigadu et al.24 The design of this cell allowed for
toluene from decane and toluene mixtures. temperature control by the circulation of water through the
The success of a solvent extraction process depends on the jacket of the cell. Temperature measurement with a precision
selection of a suitable solvent. Properties of such a solvent of 0.1 K20 was achieved by a HellermannTyton T235H
include a high selectivity for aromatics, high capacity, ability to multimeter sensor housed in the thermowell of the cell. The
form two phases at reasonable temperatures, and low risks to sensor was calibrated with the WIKA CTH6500 thermometer.
health, safety, and the environment (HSE).15 The Nexis GC-2030 gas chromatograph with flame ionizing
However, these solvents do not possess all of the properties detector (FID) from Shimadzu Corporation was used for the
of an ideal solvent. For instance, NMP has a very high capacity composition analysis. The calibration of the detector followed
and a relatively low selectivity.16 The combination of solvents the method detailed by Raal and Mü hlbauer,25 which
can overcome such cases.17 employed area ratios of single-phase samples of known
In a solvent screening methodology, Brijmohan et al.,18 composition across the composition range of 0 to 1. The
identified glycerol as a potential solvent in aromatics recovery. uncertainty in the mole fraction compositions, which may be
However, the use of glycerol is accompanied by a low- attributed to instrument stability, sample preparation, and
performance index due to low solvent capacities.19 To repeatability, was 0.006.20 The column used was the SH-Rxi-5
overcome this challenge and increase the degree of BTEX ms from Restek with the dimensions: 30 m, 0.25 mm ID, and
separation from its mixtures, exploring other solvents and the 0.25 μm column thickness. The agitation and equilibrium
effect of cosolvents via the liquid−liquid extraction technique waiting times for each measurement were 1 h each. A
is pertinent. In this study, the extraction strength of glycerol minimum of three 1 μL samples were analyzed for each phase.
and NMP was investigated in different ratios and temperatures, The relative standard deviation (RSD) values for the area
glycerol was used as a cosolvent with NMP. The system ratios of each phase were <2%. The computation of the RSD
under consideration is n-heptane + (NMP + glycerol) + followed the method outlined by Box et al.26 The experimental
toluene, and the LLE measurements were carried out at 298.15 setup is illustrated in Figure 1.
and 313.15 K and 101.3 kPa using different NMP to glycerol 2.3. Data Correlation. Empirical functions of varying
mass ratios of 90:10, 70:30, and 50:50. The LLE data reported complexity have been developed for the correlation of binodal
in this work serve to fill the knowledge gap pertaining to the curve data.27−31 This study employed the Hlavatý,28 β
equilibrium properties of an n-alkane + NMP + glycerol + function,30 and log γ29 equations, represented by eqs 1-3, for
aromatic system. the correlation of the binodal curve data. These models are
purely mathematical and have no thermodynamic basis. In
2. EXPERIMENTAL AND COMPUTATIONAL their application, these models have been found to display
METHODS good precision, passing through the solubility curve on both
2.1. Materials. Refractive index measurements ascertained phases and require few coefficients.32 Furthermore, the
the chemical purity of the chemicals used in this study. An appropriate choice of independent variables for the β function
Atago refractometer (model RX 7000i) with an overall and log γ equation avoids the problem of having highly
uncertainty of 0.0001 was used.20 Table 1 lists the respective intercorrelated variables.32,33 The solvent and cosolvent were
chemical supplier information and refractive index measure- treated as a pseudocomponent in the correlation of the binodal
ments. The chemicals were used without further purification. curve and tie-line data as has been done previously.34
2.2. Equipment. The LLE measurements were undertaken Hlavatý28 equation with coefficients Ai
by using the direct analytical technique. The experimental x3 = A1xA lnxA + A 2xB lnxB + A3xAxB (1)
procedure employed is that of Alders.23 Solvent mixtures were 30
prepared by mixing the solvent (NMP) with the cosolvent β function equation of Letcher et al. with coefficients Bi
(glycerol) to create mixtures with NMP to glycerol mass ratios
x3 = B1(1 xA)B2 xA B3 (2)
of 90:10 (NMP + 10% glycerol), 70:30 (NMP + 30%
29
glycerol), and 50:50 (NMP + 50% glycerol). The LLE log γ equation of Letcher et al., with coefficients Ci
measurements were carried out in a double-walled glass cell C2 C3
depicted in Figure 1c. The adoption of this cell was proposed x3 = C1( lnxA) xA (3)

B https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

Figure 1. Experimental setup: multimeter (a), magnetic stirrer (b), equilibrium cell (c), silicon pipes (d), water bath (e), balance (f), and
refractometer (g). [Image taken by one of the authors].

Table 2. Experimental LLE Data for N-Heptane (1) + Solvent[90% NMP (2) + 10% Glycerol (3)] + Toluene (4) System at
298.15 and 313.15 K and 101.3 kPaa
Solvent-Rich Phase n-Heptane-Rich Phase
x1 x2 x3 x4 x1 x2 x3 x4 ω K2
298.15 K
0.1369 0.7709 0.0922 0.0000 0.9027 0.0869 0.0104 0.0000 - -
0.1395 0.7409 0.0886 0.0310 0.8313 0.1106 0.0132 0.0449 4.1143 0.6904
0.1481 0.7052 0.0844 0.0623 0.7463 0.1463 0.0175 0.0899 3.4921 0.6930
0.1635 0.6442 0.0771 0.1152 0.6151 0.2070 0.0248 0.1531 2.8307 0.7524
0.1848 0.5948 0.0711 0.1493 0.5249 0.2504 0.0300 0.1947 2.1781 0.7668
0.2304 0.5160 0.0617 0.1919 0.4347 0.3039 0.0364 0.2250 1.6092 0.8529
313.15 K
0.1522 0.7572 0.0906 0.0000 0.8898 0.0984 0.0118 0.0000 - -
0.1540 0.7198 0.0861 0.0401 0.8068 0.1286 0.0154 0.0492 4.2700 0.8150
0.1475 0.7034 0.0841 0.0650 0.7133 0.1601 0.0191 0.1075 2.9241 0.6047
0.1605 0.6594 0.0789 0.1012 0.6118 0.2127 0.0254 0.1501 2.5700 0.6742
0.1863 0.6031 0.0721 0.1385 0.5413 0.2451 0.0293 0.1843 2.1835 0.7515
0.2298 0.5276 0.0631 0.1795 0.4620 0.2950 0.0353 0.2077 1.7375 0.8642
a
Standard uncertainties are u(T) = 0.1 K; u(x) = 0.006.

where where x3 represents the mole fraction of toluene; n represents

1
the number of data points; and 3 represents the number of
xA =
(x 2 + 2 x3 x 20) estimated coefficients.35 The tie-lines were correlated with the
0
(x 22 x 20) (4)
NTRL model of Renon and Prausnitz36 and the UNIQUAC
activity coefficient model of Anderson and Prausnitz37 using
1 ASPEN Plus software. The correlation procedure employed
xB =
( 0
x 22 x2 x
2 3 ) the Britt−Leucke algorithm,38 the Deming initialization
0
(x22 x 20) (5) method,39 and the minimization of a nonlinear least-squares
objective function.40 In the NRTL model regression, the value
In eqs 1−5, x2 refers to the mole fraction composition of the of the nonrandomness parameter (αij) was set to be equal for
solvent; x3 refers to the mole fraction composition of toluene; all binary pairs and fixed at either 0.20, 0.30, 0.40, or 0.48 as
and x022 and x02 represent the values of x2 on the binodal curve suggested by Walas.41 The best-fit correlation was indicated by
that cut the x3 = 0 axis. The best fit correlation was based on
the lowest value of the root-mean-square deviation (rmsd),
the lowest value of the standard deviation (σ) which is
represented by eq 6: computed by eq 7:

l xabc(calc))2 |
1/2
l
o |1/2
x3(exp)]k2 o o
o (xabc(exp) o
o
=o
n [x3(calc) o a b c
m
o }
o rmsd = m
o }
o
o k=1 (n 3) o o 6k o
n ~ (6) n ~ (7)

C https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

Table 3. Experimental LLE Data for N-Heptane (1) + Solvent[70%NMP(2) + 30% Glycerol(3)] + Toluene (4) System at
298.15 and 313.15 K and 101.3 kPaa
Solvent-Rich Phase n-Heptane-Rich Phase
x1 x2 x3 x4 x1 x2 x3 x4 ω K2
298.15 K
0.0919 0.6214 0.2867 0.0000 0.9481 0.0355 0.0164 0.0000 - -
0.1107 0.5866 0.2706 0.0321 0.8863 0.0458 0.0211 0.0468 5.4915 0.6859
0.1222 0.5514 0.2544 0.0720 0.7686 0.0864 0.0399 0.1051 4.3088 0.6851
0.0923 0.5383 0.2484 0.1210 0.6931 0.1005 0.0464 0.1600 5.6788 0.7563
0.1221 0.4955 0.2286 0.1538 0.5886 0.1308 0.0603 0.2203 3.3655 0.6981
0.1196 0.4634 0.2138 0.2032 0.4731 0.1727 0.0796 0.2746 2.9272 0.7400
0.1476 0.4040 0.1864 0.2620 0.3783 0.2131 0.0983 0.3103 2.1641 0.8443
313.15 K
0.1305 0.5950 0.2745 0.0000 0.9369 0.0432 0.0199 0.0000 - -
0.1412 0.5647 0.2605 0.0336 0.8242 0.0747 0.0344 0.0666 2.9452 0.5045
0.1378 0.5199 0.2399 0.1024 0.7040 0.1085 0.0500 0.1375 3.8047 0.7447
0.1403 0.4659 0.2150 0.1788 0.5633 0.1549 0.0714 0.2104 3.4120 0.8498
0.1647 0.4120 0.1901 0.2332 0.4570 0.1867 0.0861 0.2702 2.3948 0.8631
0.1866 0.3786 0.1746 0.2602 0.3844 0.2174 0.1003 0.2979 1.7993 0.8735
a
Standard uncertainties are u(T) = 0.1 K; u(x) = 0.006.

Table 4. Experimental LLE Data for N-Heptane (1) + Solvent[50%NMP (2) + 50% Glycerol (3)] + Toluene (4) System at
298.15 and 313.15 K and 101.3 kPaa
Solvent-Rich Phase n-Heptane-Rich Phase
x1 x2 x3 x4 x1 x2 x3 x4 ω K2
298.15 K
0.0767 0.4447 0.4786 0.0000 0.9750 0.0120 0.0130 0.0000 - -
0.0668 0.4346 0.4678 0.0308 0.9275 0.0159 0.0171 0.0395 10.8266 0.7797
0.1084 0.4048 0.4357 0.0511 0.8634 0.0240 0.0259 0.0867 4.6944 0.5894
0.1324 0.3605 0.3880 0.1191 0.7762 0.0292 0.0315 0.1631 4.2810 0.7302
0.1292 0.3019 0.3250 0.2439 0.6001 0.0651 0.0701 0.2647 4.2798 0.9214
0.1453 0.2623 0.2824 0.3100 0.4814 0.0870 0.0937 0.3379 3.0396 0.9174
0.1610 0.2364 0.2545 0.3481 0.4031 0.1097 0.1181 0.3691 2.3613 0.9431
313.15 K
0.1098 0.4287 0.4615 0.0000 0.9620 0.0183 0.0197 0.0000 - -
0.1303 0.3855 0.4150 0.0692 0.8235 0.0315 0.0340 0.1110 3.9400 0.6234
0.1487 0.3369 0.3626 0.1518 0.7087 0.0461 0.0497 0.1955 3.7001 0.7765
0.1585 0.2984 0.3213 0.2218 0.6066 0.0659 0.0709 0.2566 3.3081 0.8644
0.1766 0.2590 0.2789 0.2855 0.4759 0.0944 0.1016 0.3281 2.3449 0.8702
0.1977 0.2317 0.2495 0.3211 0.3991 0.1152 0.1240 0.3617 1.7921 0.8878
a
Standard uncertainties are u(T) = 0.1 K; u(x) = 0.006.

where k is the number of experimental points, x is the mole (x3)II

fraction, and the subscripts a, b, and c represent component, K2 =
(x3)I (8)
phase, and tie-line, respectively.
2.4. Selectivity and Capacity. The solvent capacity, (x3)II /(x3)I
represented by the distribution coefficient (K), is a measure of =
(x1)II /(x1)I (9)
the solvent power, it indicates the maximum volume of
aromatics that the solvent can contain.18 Therefore, its value where subscripts 1, 2, and 3 represent n-heptane, solvent
(NMP and glycerol as a pseudocomponent), and toluene,
should preferably be high so that a low ratio of solvent to feed respectively. Superscripts I and II represent n-heptane-rich and
may be used for an effective separation.42 The ratio of the solvent-rich phases, respectively.
distribution coefficient of the aromatics and nonaromatics A value of unity for the selectivity represents the plait point.
yields the selectivity (ω) of the solvent. In the present work, At this point, the two phases have identical compositions;
the selectivity measures the ability of the solvent to separate therefore, no separation can be achieved. This point was
determined using the graphical Coolidge method.43
toluene from n-heptane. Thus, a value greater than unity
indicates effective toluene recovery. Solvent capacity and 3. RESULTS AND DISCUSSION
selectivity are mathematically expressed by eqs 8 and 9, In this section, systems are referenced in terms of their solvent
respectively: mixing ratios. The experimental LLE data and respective
D https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

selectivities and capacities for the systems under investigation

are represented in Tables 2−4. Estimations of the plait point
compositions obtained from the graphical Coolidge method
are presented in Table 5. Compositions are presented in terms

Table 5. Estimated Plait Point Composition Using the

Graphical Coolidge Methodab
T (K) x1 x2 x3 x4
n-heptane (1) + NMP(2) + glycerol(3) + toluene (4)
298.1500 0.3194 0.3976 0.0475 0.2355
313.1500 0.3759 0.3582 0.0428 0.2231
n-heptane (1) + NMP(2) + glycerol(3) + toluene (4)
298.1500 0.2628 0.2814 0.1298 0.3260
313.1500 0.2379 0.3045 0.1405 0.3171 Figure 3. LLE for n-heptane (1) + solvent mixture [90% NMP(2) +
n-heptane (1) + NMP(2) + glycerol(3) + toluene (4) 10% glycerol(3)] + toluene (4) system at 313.15 K and 101.3 kPa. ●,
298.1500 0.2619 0.1711 0.1841 0.3829
Experimental data; ▲, plait point; ―, β function; •••x•••, NRTL
(α = 0.2).
313.1500 0.2710 0.1757 0.1891 0.3642
a
Standard uncertainty is u(T) = 0.1 K. bN-heptane (1) + solvent
[NMP(2) + glycerol(3)]+ toluene (4) system at 298.15 and 313.15 K
and 101.3 kPa.

of mole fractions where x1, x2, x3, and x4 represent the mole
fractions of n-heptane, NMP, glycerol (solvent mixtures), and
toluene, respectively.
Figures 2−7 graphically present the experimental and
modeled data of the best fit for the systems under

Figure 4. LLE for n-heptane (1) + solvent mixture [70% NMP(2) +

30% glycerol(3)] + toluene (4) system at 298.15 K and 101.3 kPa. ●,
Experimental data; ▲, plait point; ―, β function; •••x•••, NRTL
(α = 0.2).

Figure 2. LLE for n-heptane (1) + solvent mixture [90% NMP(2) +

10% glycerol(3)] + toluene (5) system at 298.15 K and 101.3 kPa. ●,
Experimental data; ▲, plait point; ―, β function; •••x•••, NRTL
(α = 0.2).

investigation, while their Hlavatý equation, β function, log γ

equation, NRTL, and UNIQUAC parameters are tabulated in
Tables 6−11. As the systems are quaternary, the activity
coefficient interaction parameters are presented for 4 Figure 5. LLE for n-heptane (1) + solvent mixture [70% NMP(2) +
components where 1, 2, 3, and 4 represent n-heptane, NMP, 30% glycerol(3)] + toluene (4) system at 313.15 K and 101.3 kPa. ●,
glycerol, and toluene, respectively. All systems produced one Experimental data; ▲, plait point; ―, β function; •••x•••, NRTL
pair of partially miscible liquids at 298.15 and 313.15 K and (α = 0.2).
101.3 kPa. Therefore, they are classified as type I LLE
systems.44 The β function equation provided a better
correlation for the binodal curves of all systems. The NRTL are represented by negative and positive values of the binary
model provided a better correlation for the tie-lines of all interaction parameters, respectively.45 As such, interaction
systems. This can be attributed to the nonrandomness parameters for the immiscible n-heptane and glycerol pair
parameter (αij). Proper selection of αij makes the NRTL move from negative to positive with the increase in glycerol
model applicable to a variety of mixtures, while other models content and the miscible n-heptane and toluene pair have
appear to be limited to specific types.36 In the NRTL model, negative interaction parameters, as shown in Tables 9− 11.
favorable and unfavorable interactions between species i and j Figures 2−7 show that the size of the two-phase region
E https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

Table 7. Binodal Curve Correlation Parameters for N-

Heptane (1) + (NMP + 30% Glycerol) (2) + Toluene (3)
System at 298.15 and 313.15 K and 101.3 kPa
Hlavatý Equation β Function Equation Log γ Equation
298.15 K
A1 0.4874 B1 1.6958 C1 1.5698
A2 0.0759 B2 1.0465 C2 1.0163
A3 1.9507 B3 1.3598 C3 1.7589
σ 0.0201 σ 0.0177 σ 0.0184
313.15 K
A1 0.6783 B1 1.5766 C1 1.4751
A2 0.1104 B2 0.9798 C2 0.9536
A3 2.1675 B3 1.4212 C3 1.8007
Figure 6. LLE for the n-heptane (1) + solvent mixture [50% NMP(2)
+ 50% glycerol(3)] + toluene (4) system at 298.15 K and 101.3 kPa. σ 0.0184 σ 0.0107 σ 0.0116
●, Experimental data; ▲, plait point; ―, β function; •••x•••,
NRTL (α = 0.2). Table 8. Binodal Curve Correlation Parameters for N-
Heptane (1) + (NMP + 50% Glycerol) (2) + Toluene (3)
System at 298.15 and 313.15 K and 101.3 kPa
Hlavatý Equation β Function Equation Log γ Equation
298.15 K
A1 0.3748 B1 2.1457 C1 1.8906
A2 0.2583 B2 1.1824 C2 1.1236
A3 2.0964 B3 1.2629 C3 1.6763
σ 0.0236 σ 0.0228 σ 0.0244
313.15 K
A1 0.3596 B1 1.8962 C1 1.7109
A2 0.1879 B2 1.1303 C2 1.0849
A3 2.2379 B3 1.2216 C3 1.6339
σ 0.0109 σ 0.0101 σ 0.0116
Figure 7. LLE for n-heptane (1) + solvent mixture [50% NMP(2) +
50% glycerol(3)] + toluene (3) system at 313.15 K and 101.3 kPa. ●, Table 9. NRTL and UNIQUAC Interaction Parameters for
Experimental data; ▲, plait point; ―, β function; •••x•••, NRTL the N-Heptane (1) + (NMP (2) + 10% Glycerol (3)) +
(α = 0.2). Toluene (4) System at 298.15 and 313.15 K and 101.3 kPa
Compo-
Table 6. Binodal Curve Correlation Parameters for the N- nent NRTL Parameters (α = 0.2) UNIQUAC Parameters
Heptane (1) + (NMP + 10% Glycerol) (2) + Toluene (3)
ij ji gij- gjj gji- gii rmsd uij- ujj uji- uii rmsd
System at 298.15 and 313.15 K and 101.3 kPa
298.15 K
Hlavatý Equation β Function Equation Log γ Equation 12 21 −8.588 1.544 0.0058 15.30 −1.8314 0.0423
298.15 K 13 31 −1.731 4.011 −0.2239 1.090
A1 0.3155 B1 0.9209 C1 0.8566 14 41 −0.3302 5.086 −1.506 −0.8734
A2 0.2377 B2 0.8891 C2 0.8603 23 32 50.77 827.1 −3.356 420.3
A3 1.3791 B3 1.1376 C3 1.4691 24 42 79.16 −31.64 68.96 68.96
σ 0.0103 σ 0.0038 σ 0.0046 34 43 1.208 12.05 0.2442 13.16
313.15 K 313.15 K
A1 0.3445 B1 0.7429 C1 0.6999 12 21 19.70 −2.372 0.0066 −222.5 3.058 0.0931
A2 0.0016 B2 0.7699 C2 0.7479 13 31 −0.7364 6.149 63.92 4.883
A3 1.4121 B3 1.0207 C3 1.3105 14 41 −0.6616 3.740 −0.4512 −0.1915
σ 0.0163 σ 0.0048 σ 0.0058 23 32 −4.636 75.96 −2.567 364.3
24 42 3920 63.92 601.9 39.06
34 43 1.361 4.023 2.240 2.405
increased with the increase in the glycerol content of the
solvent. Brijmohan et al.19 found the application of pure
glycerol in aromatics recovery demonstrated type II LLE In Figures 8−11, the selectivity and capacity of the systems
behavior44 which is attributed to glycerol’s high density.16 This investigated are compared with those of pure NMP46 and
translates to a large area over which the separation of toluene glycerol19 as well as the conventional solvents sulfolane47 and
from n-heptane can occur. Therefore, this property of glycerol NFM.42 Figures 8 and 9 show that the solvents under
was intensified when the glycerol content was increased in the investigation showed higher or comparable selectivities with an
solvent mixtures under investigation. The size of the two-phase increasing glycerol content. Furthermore, Figures 8 and 9 show
regions, listed in decreasing order of magnitude, is as follows: that the selectivity decreases with increasing toluene content in
NMP + 50% glycerol > NMP + 30% glycerol > NMP + 10% the feed as well as with increasing temperature. This is the
glycerol. A similar trend was also observed by Naidoo et al.34 result of the increased miscibility between the toluene and
F https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

Table 10. NRTL and UNIQUAC Interaction Parameters for

N-Heptane (1) + (NMP (2) + 30% Glycerol (3)) + Toluene
(4) System at 298.15 and 313.15 K and 101.3 kPa
Compo-
nent NRTL Parameters (α = 0.2) UNIQUAC Parameters
ij ji gij- gjj gji- gii rmsd uij- ujj uji- uii rmsd
298.15 K
12 21 −43.08 −1.205 0.0031 −36.02 −6.022 0.0623
13 31 0.8653 3.074 2.084 1.734
14 41 −1.172 4.900 −1.735 2.490
23 32 91.66 −37.46 −3.963 −40.65
24 42 −1.875 7412 −5.326 −27.18
34 43 1.107 1.479 0.9353 7.457
313.15 K
12 21 −41.39 −1.595 0.0033 −12.47 3.027 0.0967
13 31 1.031 3.053 1.123 4.229
14 41 −1.181 2.445 0.5545 0.2949 Figure 8. Selectivity comparison with respect to toluene content in
23 32 −2.091 −39.16 2.319 −19.22
the n-heptane-rich phase for the system n-heptane (1) + solvent (2) +
toluene (3) at 298.15 K and 101.3 kPa. Solvents: ■, pure glycerol; -,
24 42 −2.079 7411 30.16 −5.590
pure NMP; ○, NMP + 10% glycerol; ×, NMP + 30% glycerol; ●,
34 43 1.272 2.137 2.146 1.796 NMP + 50% glycerol; ▲, sulfolane; ⧫, NFM.

Table 11. NRTL and UNIQUAC Interaction Parameters for

N-Heptane (1) + (NMP (2) + 50% Glycerol (3)) + Toluene
(4) System at 298.15 and 313.15 K and 101.3 kPa
Compo-
nent NRTL Parameters (α = 0.2) UNIQUAC Parameters
ij ji gij- gjj gji- gii rmsd uij- ujj uji- uii rmsd
298.15 K
12 21 −15.23 −2.217 0.0044 −8.615 −1.376 0.0104
13 31 2.485 1.078 6.499 1.400
14 41 −1.150 3.607 −0.9841 6.790
23 32 −3.611 −18.24 −3.056 57.94
24 42 −2.430 −20.43 −4.665 −4.665
34 43 2.471 8.525 2.236 12.96
313.15 K
12 21 −17.28 −2.025 0.0030 −18.00 5.566 0.054
13 31 0.1991 3.108 1.349 2.920
14 41 −0.9879 2.002 −0.9821 0.6230 Figure 9. Selectivity comparison with respect to toluene content in
23 32 −3.064 −85.58 −0.8788 −332.7 the n-heptane-rich phase for the system n-heptane (1) + solvent (2) +
24 42 −3.035 −18.76 −7.848 −18.13 toluene (3) at 313.15 K and 101.3 kPa. Solvents: ■, pure glycerol; -,
34 43 2.235 4.435 2.170 3.900 pure NMP; ○, NMP + 10% glycerol; ×, NMP + 30% glycerol; ●,
NMP + 50% glycerol; ▲, sulfolane; ⧫, NFM.

NMP. A similar phenomenon is observed in the work of Furthermore, the good capacities demonstrated by the solvents
Brijmohan et al.45 The sharp decrease in selectivity observed under investigation can be attributed to the very high
with pure NMP is less pronounced in the NMP + 10% capacities associated with pure NMP.49 Saha et al.,50 observed
glycerol, NMP + 30% glycerol, and NMP + 50% glycerol a similar trend and reported higher capacities for toluene
systems, suggesting that the improved selectivity can be recovery using a solvent mixture of NMP and tetraethylene
attributed to glycerol in the solvent mixtures. A similar compared to the use of pure tetraethylene glycol. The changes
observation can be drawn from the work of Brijmohan et al.45 in the extent of the two-phase region at increased temperatures
This improvement in selectivity is in the order: NMP + 50% are evident in the capacity, as shown in Figures 10 and 11,
glycerol > NMP + 30% glycerol > NMP + 10% glycerol. suggesting that there was a slight decrease in this region. At
Although the selectivity has shown an improvement, conven- 313.15 K, the solubility of toluene in the solvent mixtures is
tional solvents still have superior selectivities. The solvents higher than the solubility of n-heptane in the solvent mixtures.
under investigation showed good capacities, especially NMP + Overall, a trade-off between a high selectivity and a low
50% glycerol. These findings were outside the scope of the capacity versus a low selectivity and a high capacity is seen
initial predictions, as the poor capacities associated with among all solvents at both temperatures.
glycerol were expected to carry over with an increase in As previously mentioned, the properties that govern solvent
glycerol content. According to Green and Southard,48 a selection are high selectivity for aromatics, high capacity, the
relatively large two-phase region is indicative of the magnitude ability to form two phases at reasonable temperatures, and low
of the solvent capacity. Hence, the NMP + 50% glycerol risks to health, safety, and the environment (HSE). NMP +
system, having the largest two-phase region, demonstrated this. 50% glycerol demonstrated the best selectivity with increasing
G https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

the expense of reducing other desired properties. Therefore,

among the solvents investigated, the most suitable solvent
would be a trade-off between these properties.

4. CONCLUSION
This investigation focused on identifying solvent mixtures of
NMP and glycerol as alternatives in aromatic recovery to
encourage the use of greener solvents. This was done by the
effective separation of toluene from n-heptane using solvent
mixtures of NMP and glycerol, with ratios of NMP to glycerol
at 90:10, 70:30, and 50:50. New LLE data were measured for
systems n-heptane + the solvent mixture + toluene at 298.15
and 313.15 K and 101.3 kPa using the direct analytical
technique. The LLE data were found to be better correlated
with the β function equation and the NRTL activity coefficient
model. All systems exhibited type I ternary LLE behavior. The
Figure 10. Capacity comparison with respect to toluene content in selectivity was greater than unity across all systems and
the n-heptane-rich phase for the system n-heptane (1) + solvent (2) + temperatures, indicating an effective separation. The systems
toluene (3) at 298.15 K and 101.3 kPa. Solvents: ■, pure glycerol; -, demonstrated improved selectivities and capacities with an
pure NMP; ○, NMP + 10% glycerol; ×, NMP + 30% glycerol; ●, increase in glycerol. Therefore, the selection of the ideal
NMP + 50% glycerol; ▲, sulfolane; ⧫, NFM. solvent would be a trade-off between solvent performance,
process design, impact on HSE, and capital and operational
costs.

■ AUTHOR INFORMATION
Corresponding Author
Naadhira Seedat − Department of Chemical Engineering,
Faculty of Engineering and the Built Environment, University
of Johannesburg, Doornfontein 2028, South Africa;
Department of Chemical Engineering, University of Pretoria,
Pretoria 0028, South Africa; orcid.org/0000-0001-8745-
1127; Email: [email protected]
Authors
Blessing Mcebo Paile − Department of Chemical Engineering,
Faculty of Engineering and the Built Environment, University
of Johannesburg, Doornfontein 2028, South Africa
Caleb Narasigadu − School of Chemical and Minerals
Figure 11. Capacity comparison with respect to toluene content in Engineering, Faculty of Engineering, Potchefstroom Campus,
the n-heptane-rich phase for the system n-heptane (1) + solvent (2) + North-West University, Potchefstroom 2351, South Africa;
toluene (3) at 313.15 K and 101.3 kPa. Solvents: ■, pure glycerol; -, orcid.org/0000-0001-9224-136X
pure NMP; ○, NMP + 10% glycerol; ×, NMP + 30% glycerol; ●,
NMP + 50% glycerol; ▲, sulfolane; ⧫, NFM.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jced.4c00614

aromatic content in the feed. This would mean that its Funding
application could lead to reduced capital costs due to the high B.P.: Conceptualization, formal analysis, investigation, writing
selectivity and the absence of additional equipment needed to − original draft. C. N.: Conceptualization, supervision,
recover glycerol’s negligible losses.18,19 NMP + 50% glycerol resources, methodology, writing − reviewing and editing.
also demonstrated the largest area for separation and the best N.S.: Supervision, resources, project administration, writing −
capacity at low temperatures. However, these capacities are not reviewing and editing.
comparable to those of conventional solvents. Therefore, the Notes
use of the solvents investigated could be accompanied by high The authors declare no competing financial interest.
solvent-to-feed ratios. NMP + 50% glycerol system contained
more glycerol, which encourages the use of greener solvents.
However, its application might not be practical on an industrial
■ ACKNOWLEDGMENTS
This work is based upon research supported by the National
scale due to the high viscosities associated with glycerol. Research Foundation of South Africa.
Glycerol has a viscosity of approximately 582.60 mPa·s, which
is large compared to conventional solvents.18 The high
viscosity of glycerol reduces the solubility and molecular
■ REFERENCES
(1) Wang, H.; Liu, W.; Shi, F.; Huang, L.; Lian, J.; Qu, L.; Cai, J.;
transfer of solutes in the solvent-rich phase, thereby requiring Xu, Z. Metabolic pathway engineering for high-level production of 5-
high energy consumption and pumping costs.51 This effect can hydroxytryptophan in Escherichia coli. Metab. Eng. 2018, 48, 279−287.
be mitigated by reducing the amount of glycerol and avoiding (2) Averesch, N. J. H.; Kromer, J. O. Metabolic Engineering of the
it as a major constituent of the solvent. However, this comes at Shikimate Pathway for Production of Aromatics and Derived

H https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

Compounds�Present and Future Strain Construction Strategies. (24) Narasigadu, C.; Naidoo, M.; Ramjugernath, D. Ternary
Front. Bioeng. Biotechnol. 2018, 6, 32. Liquid−Liquid Equilibrium Data for the Water + Acetonitrile +
(3) Nekhaev, A. I.; Maksimov, A. L. Production of Aromatic {Butan-1-ol or 2-Methylpropan-1-ol} Systems at (303.2, 323.2, 343.2)
Hydrocarbons from Biomass. Pet. Chem. 2021, 61 (1), 15−34. K and 1 atm. J. Chem. Eng. Data 2014, 59 (11), 3820−3824.
(4) Brijmohan, N.; Narasigadu, C. Ternary Liquid−Liquid (25) Mühlbauer, A. L.; Raal, J. D.; Phase equilibria: Measurement and
Equilibrium Data for the N-Formylmorpholine + Toluene + {n- computation; Taylor & Francis, 1998.
Nonane or n-Decane} Systems at (303.2, 323.2, and 343.2) K and (26) Box, G. E. P.; Hunter, J. S.; Hunter, W. G. Statistics for
101.3 kPa. J. Chem. Eng. Data 2020, 65 (2), 788−792. experimenters: Design, innovation, and discovery; Wiley-Interscience,
(5) Franck, H.-G.; Stadelhofer, J. W. Industrial Aromatic Chemistry: 2005.
raw Materials · Processes · Products; Springer Science & Business (27) Bulatov, S. N.; Yachmenev, L. T. Approximation of binodal
Media, 2012. curves of heterogeneous ternary systems. Teor. Osn. Khim. Technol.
(6) Meindersma, G. W.; de Haan, A. B. Conceptual process design 1971, 5, 644−650.
for aromatic/aliphatic separation with ionic liquids. Chem. Eng. Res. (28) Hlavatý, K. Correlation of the binodal curve in a ternary liquid
Des. 2008, 86, 745−752. mixture with one pair of immiscible liquids. Collect. Czech. Chem.
(7) Díaz, I.; Palomar, J.; Rodríguez, M.; De Riva, J.; Ferro, V.; Commun. 1972, 37 (12), 4005−4007.
González, E. J. Ionic liquids as entrainers for the separation of (29) Letcher, T. M.; Heyward, C.; Wootton, S. Phase separation in
aromatic−aliphatic hydrocarbon mixtures by extractive distillation. petrol-alcohol blends. S. Afr. J. Chem. 1986, 39 (1), 19−22.
Chem. Eng. Res. Des. 2016, 115, 382−393. (30) Letcher, T. M.; Siswana, P. M.; Van Der Watt, P.; Radloff, S.
(8) Seddon, K. R. Ionic liquids for clean technology. J. Chem. Phase equilibria for (an alkanol + p-xylene + water) at 298.2 K. J.
Technol. Biotechnol. 1997, 68 (4), 351−356. Chem. Thermodyn. 1989, 21 (10), 1053−1060.
(9) Canales, R. I.; Brennecke, J. F. Comparison of Ionic Liquids to (31) Procházka, J.; Heyberger, A. Correlation of ternary liquid-liquid
Conventional Organic Solvents for Extraction of Aromatics from equilibria in system isobutyl acetate-acetic acid-water. Chem. Eng. Sci.
Aliphatics. J. Chem. Eng. Data 2016, 61 (5), 1685−1699. 1996, 51 (6), 893−903.
(10) Alkhaldi, K. H. A. E.; Fandary, M. S.; Al-Jimaz, A. S.; Al- (32) Olán-Acosta, M. D. L. Á .; Castrejón-González, E. O.; Alvarado,
Tuwaim, M. S.; Fahim, M. A. Liquid−liquid equilibria of aromatics J. F. J.; Rico-Ramírez, V. Approximate design method for reactive
removal from middle distillate using NMP. Fluid Phase Equilib. 2009, liquid extractors based on thermodynamic equilibrium correlations.
286 (2), 190−195. Chem. Eng. Res. Des. 2016, 109, 443−454.
(11) Meindersma, G. W.; Podt, A. J. G.; de Haan, A. B. Ternary (33) Narasigadu, C.; Raal, J. D.; Naidoo, P.; Ramjugernath, D.
liquid−liquid equilibria for mixtures of toluene+n-heptane+an ionic Ternary liquid− liquid equilibria of acetonitrile and water with
liquid. Fluid Phase Equilib. 2006, 247 (1−2), 158−168. heptanoic acid and nonanol at 323.15 K and 1 atm. J. Chem. Eng. Data
(12) DongChu, C.; HongQi, Y.; Hao, W. (Liquid+ liquid) equilibria 2009, 54 (3), 735−738.
of {heptane+ xylene+ N-formylmorpholine} ternary system. J. Chem. (34) Naidoo, R. D.; Letcher, T. M.; Ramjugernath, D. Ternary
Thermodyn. 2007, 39 (12), 1571−1577. Liquid−Liquid Equilibra for Pseudoternary Mixtures Containing an
(13) Mohsen-Nia, M.; Modarress, H.; Doulabi, F.; Bagheri, H. n-Alkane + an Aromatic Hydrocarbon + {N-Methyl-2-pyrrolidinone +
Liquid + liquid equilibria for ternary mixtures of (solvent + aromatic a Solvent} at 298.2 K and 1 atm. J. Chem. Eng. Data 2001, 46 (6),
hydrocarbon + alkane). J. Chem. Thermodyn. 2005, 37 (10), 1111− 1375−1380.
1118. (35) Sen, A.; Srivastava, M. Regression Analysis: theory, Methods, and
(14) Mesquita, F. M. R.; Pinheiro, R. S.; Santiago-Aguiar, R. S.; de Applications; Springer: New York, 1990. .
Sant’ana, H. B. Measurement of phase equilibria data for the (36) Renon, H.; Prausnitz, J. M. Local compositions in
extraction of toluene from alkane using different solvents. Fluid Phase thermodynamic excess functions for liquid mixtures. AichE J. 1968,
Equilib. 2015, 404, 49−54. 14 (1), 135−144.
(15) Saha, M.; Rawat, B.; Khanna, M.; Nautiyal, B. Liquid− liquid (37) Anderson, T. F.; Prausnitz, J. M. Application of the UNIQUAC
equilibrium studies on toluene+ heptane+ solvent. J. Chem. Eng. Data equation to calculation of multicomponent phase equilibria. 1. Vapor-
1998, 43 (3), 422−426. liquid equilibria. Ind. Eng. Chem. Process Des. Dev. 1978, 17 (4), 552−
(16) Reddy, P.; Letcher, T. M. Chapter 5 - Phase Equilibrium 561.
Studies on Ionic Liquid Systems for Industrial Separation Processes of (38) Britt, H. I.; Luecke, R. H. The estimation of parameters in
Complex Organic Mixtures. In Thermodynamics, Solubility and nonlinear, implicit models. Technometrics 1973, 15 (2), 233−247.
Environmental Issues; Letcher, T. M. Ed.; Elsevier, 2007, pp. 85−111. (39) Deming, W. E. Statistical adjustment of data; John Wiley &
(17) Mueller, E.; Hochfield, G. Aromatics Extraction with Solvent Sons, 1943.
Combinations. In 7th World Petroleum Congress; OnePetro, 1967. (40) Novák, J. P.; Matouš, J.; Pick, J. Liquid-liquid Equilibria;
(18) Brijmohan, N.; Moodley, K.; Narasigadu, C. Identification and Elsevier, 1987.
Screening of Potential Organic Solvents for the Liquid−Liquid (41) Walas, S. M. Phase equilibria in chemical engineering;
Extraction of Aromatics. Org. Process Res. Dev. 2021, 25 (10), 2230− Butterworth, 1985.
2248. (42) DongChu, C.; HongQi, Y.; Hao, W. (Liquid+ liquid) equilibria
(19) Brijmohan, N.; Moodley, K.; Narasigadu, C. Ternary Liquid− of three ternary systems: (heptane+ benzene+ N-formylmorpholine),
Liquid Equilibrium Data for the n-Heptane+ Toluene+(Butane-1, 4- (heptane+ toluene+ N-formylmorpholine),(heptane+ xylene+ N-
diol or Glycerol) Systems at 298.2, 313.2, and 333.2 K and 0.1 MPa. J. formylmorpholine) from T = (298.15 to 353.15) K. J. Chem.
Chem. Eng. Data 2022, 67 (4), 975−983. Thermodyn. 2007, 39 (8), 1182−1188.
(20) Taylor, B. N.; Kuyatt, C. E. Guidelines for evaluating and (43) Coolidge, A. S. Phase equilibria for three component systems
expressing the uncertainty of NIST measurement results; US Department containing two liquid phases, the complete composition of at least
of Commerce, Technology Administration, National Institute of one of which is known. International Critical Tables Of Numerical
Standards and Technology, 1994. Data, Physics, Chemistry And Technology; McGraw-Hill: New York,
(21) O’Neil, M. J. The Merck index: an encyclopedia of chemicals, 1928; 398−417.
drugs, and biologicals; RSC Publishing, 2013. (44) Treybal, R. E. Liquid extraction; McGraw-Hill, 1963.
(22) Haynes, W. M. CRC handbook of chemistry and physics; CRC (45) Brijmohan, N.; Moodley, K.; Narasigadu, C. Use of Glycerol +
press, 2016. 2-Methylpentane-2,4-diol Liquid Mixtures in the Separation of
(23) Alders, L. Liquid-liquid extraction: theory and laboratory practice; Toluene from n-Heptane via Liquid−Liquid Extraction. J. Chem.
Elsevier Publishing Company, 1959. Eng. Data 2023, 68 (11), 2934−2945.

I https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Journal of Chemical & Engineering Data pubs.acs.org/jced Article

(46) Ferreira, P. O.; Ferreira, J. B.; Medina, A. G. Liquid-liquid

equilibria for the system N-methylpyrrolidone + toluene + n-heptane:
UNIFAC interaction parameters for N-methylpyrrolidone. Fluid Phase
Equilib 1984, 16 (3), 369−379.
(47) Tripathi, R. P.; Ram, A. R.; Rao, P. B. Liquid-liquid
equilibriums in ternary system toluene-n-heptane-sulfolane. J. Chem.
Eng. Data 1975, 20 (3), 261−264.
(48) Green, D. W.; Southard, M. Z. Perry’s Chemical Engineers’
Handbook; McGraw-Hill Education, 2019.
(49) Nagpal, J. M.; Rawat, B. S. Liquid-liquid equilibria for toluene-
heptane-N-methyl pyrrolidone and benzene-heptane solvents. J.
Chem. Technol. Biotechnol. 1981, 31 (1), 146−150.
(50) Saha, M.; Rawat, B. S.; Khanna, M. K.; Nautiyal, B. R. Liquid−
Liquid Equilibrium Studies on Toluene + Heptane + Solvent. J. Chem.
Eng. Data 1998, 43 (3), 422−426.
(51) Kislik, V. S. Chapter 13 - Advances in Development of Solvents
for Liquid−Liquid Extraction. In Solvent Extraction; Kislik, V. S. Ed.;
Elsevier, 2012. pp. 451−481.

J https://doi.org/10.1021/acs.jced.4c00614
J. Chem. Eng. Data XXXX, XXX, XXX−XXX