Article
pubs.acs.org/jced
Extraction of Phenolic Pollutants (Phenol and p‑Chlorophenol)
from Industrial Wastewater
Khalid Farhod Chasib*
Chemical Engineering Department, University of Technology, Baghdad, Iraq
ABSTRACT: The efficiency of five new solvents as a selective
solvent in the extraction of phenol and p-chlorophenol from
wastewater was investigated. The phenols samples were
collected from real petroleum refinery wastewater and from
an experimentaly prepared aqueous phenol solution. In this
work, data have been estimated for 10 systems containing,
phenol + water or p-chlorophenol + water as a common component liquid and + five solvents [ethylene glycol, diethylene
glycol, poly(ethylene glycol) (200), dimethylsulfoxide and
tetramethylene sulfolone (sulfolane)]. The consistency and accuracy of the tie-line data were evaluated using three correlation
relations namely, Bachman, Hand, and Othmer, and Tobias correlation. The Plait Point for each ternary system was estimated.
Among the five solvents used to extract the phenol or p-chlorophenol from wastewater, diethylene glycol (DEG) has the highest
selectivity and distribution coefficient and the greatest differences between its boiling point and density and those of phenol or
p-chlorophenol. It can therefore, be regarded as an excellent solvent for extracting phenol or p-chlorophenol from wastewater.
The liquid−liquid equilibrium data have been predicated using the nonrandom-two-liquid (NRTL) model and universal-quasichemical (UNIQUAC) model. The binary interaction parameters have been calculated using the Maximum Likelihood Principle
technique. The experimental data fitted by the NRTL model are more accurate than the UNIQUAC model.
1. INTRODUCTION
If the pollutants are to be recycled, liquid−liquid extraction can
be used because it is cost-efficient for the extraction of a wide
range of phenol concentrations.
For phenol removal from wastewater, liquid−liquid equilibria
data of ternary water−phenols−solvent system are important in
the modeling and design of the extraction process.11
Solvent extraction is the most economic nondestructive
process and has been applied with good results for recovering
phenol from industrial effluents when the phenol content in the
effluent is above 50 mg/L.12
Various wastewater industries have phenols, such as
petrochemicals (3.9 mg/L to 1230 mg/L), coking operations
(29 mg/L to 3950 mg/L), wood products, paint, pharmaceutical,
pulp and paper industries (0.2 mg/L to 1700 mg/L), plastics,
coal processing (10 mg/L to 6900 mg/L), and refineries (5 mg/L
to 600 mg/L).13 Several organic solvents, such as diisopropyl
ether (DIPE), methyl isobutyl ketone (MIBK), ethylbenzene,
cumene, di-isopropyl ether, isopropyl acetate, iso-pentyl-acetate,
n-hexane, toluene, methyl-iso-butyl ketone, benzene, n-octylpyrrolidone, ethyl acetate, cyclohexane, n-butyl acetate, n-hexyl
acetate, n-pentyl-acetate, cyclo-hexyl acetate, acetate esters, TBP,
and butyl acetate are in common use for recovering the phenolic
compounds in the wastewater by solvent extraction technology.14−16
It is necessary to study new solvents and new experiments
related to the removal of phenolic compounds because the
industrial wastewater from petroleum refining and coking plants
Industrial wastewaters including those from petroleum refining and coking plants contain appreciable amounts of phenols
(especially phenol or chlorophenol isomers) which have been
identified as hazardous compounds for many aquatic organisms
by environmental protection agencies.1
Phenols, particularly phenol and chlorophenols can be
considered as a serious pollutant of water and soil. They come
from the chlorination of water or from industrial and agricultural
sources.2,3 Water that has a concentration usually less than
0.02 mg·L−1 of phenols is considered unpolluted water.4 The
level of phenols in dinking water is considered by WHO’s
guidelines for drinking water quality as 0.001 mg·L−1.5
Phenol and chlorophenols are the starting material for many
chemical industries. For example, additives for rubber chemicals,
emulsifiers, dyes, detergents, adhesives, flavors and impregnating resins are heavily used.6 Five million tons per year is the
phenol worldwide production.7 It is used for the production of
caprolactcam and epoxy resins.8 The widespread use of
chlorophenols as the chlorination of municipal and industrial
wastewater and drinking water, degradation products of chlorinated herbicides, wood preservative, the chlorination of lignin or
the use of slimicicles in paper or pulp mill plants were the major
sources of environmental contamination.9
Various methods have recently been applied for phenolic
compounds removal like biological-based processes, membrane
extraction, distillation, adsorption, ozonation, electrochemical
methods, fenton, pervaporation, and liquid−liquid extraction.10
© 2013 American Chemical Society
Received: October 16, 2012
Accepted: May 13, 2013
Published: May 29, 2013
1549
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Table 1. Physical Properties of the Pure Chemicals at 298.15 K Compared with Literature Values
density ρ/g·cm−3
chemicals
P
ClP
W
EG
DEG
DMSO
TMS
a
obs.
1.1320
1.2953
0.99713
1.1112
1.1163
1.0959
1.2631
viscosity η/cp
refractive index nD
lit.
1.1320
1.3061
0.99707
1.1100
1.1164b
1.0958
1.2623c
obs.
lit.
a
1.5411
1.5581
1.3328
1.4317
1.4434
1.4780
1.4813
1.5403
1.5579
1.3329
1.4306
1.4461
1.4773
1.4820
obs.
lit.
b.p./°C
4.110
6.153
1.010
13.542
29.799
2.098
10.211
4.076
6.150a
0.89
13.550
30.000
2.121
10.286c
181.8
220
100
197.3
244.8
189.0
217.3
At 318.15 K. bAt 393.15 K. cAt 303.15 K.
is composed of petrochemicals containing an appreciable
amount of phenol or chlorophenols which have been identified
as hazardous compounds for many aquatic organisms by
Environment Protection Agencies. To our knowledge, the new
experimental information available in the literature for liquid−
liquid equilibria of ternary mixtures containing the pair phenol−
water is limited. Therefore the purpose of the present investigation is to generate the data for the water−phenol pair with new
different solvents to aid the correlation of liquid−liquid
equilibria, including phase diagrams, distribution coefficients of
phenol, tie-lines data, and selectivity of the solvents for the
aqueous phenol system.
The aim of this study is to extract phenols, mainly phenol or
p-chlorophenol, from aqueous and real petroleum refinery wastewater using different solvents. We present new experimental
liquid−liquid equilibrium (LLE) data for different ternary
systems including phenol or p-chlorophenol + water + solvent.
Five different solvents were used, ethylene glycol (EG),
diethylene glycol (DEG), poly(ethylene glycol) 200 (PEG),
dimethylsulfoxide (DMSO), and tetramethylsulfon (sulfolane)
(TMS). The phase equilibrium diagrams, distribution coefficient
of phenols, tie-line data, and selectivity and solvency of solvent
for each system were generated. To our knowledge there is no
literature data on the systems studied.
The UNIQUAC and the NRTL models of the liquid-phase
activity coefficients for the multicomponent mixtures of nonpolar and polar liquids were used to correlate the experimental data.
Table 2. Characterization of Used Refinery Wastewater
characteristic
value
pH
COD (mg/L)
phenols (mg/L)
TSS (g/L)
TDS (g/Ll)
8.9
4612
94
0.13
17
Figure 1. Cloud point titrator.
2. EXPERIMENTAL SECTION
2.1. Chemicals. Phenol and p-chlorophenol with purity of
99 % obtained from Aldrich Company were used without further
purification. Bidistilled water with conductivity (≤ 1 μs) was used
in all experiments. Ethylene glycol (stated purity 99.3 %), diethylene
glycol (stated purity 99 %), dimethyl sulfoxide (stated purity
99.5 %), sulfolane (stated purity >99.5 %) were obtained from
Fluka Company and were stored under 4 Å molecular sieves and
filtered before use. Poly(ethylene glycol) was obtained from
Aldrich Company in liquid form and the average molecular
weight reported by the manufacturer was 200 and used without
further treatment. The physical properties of the chemicals used
in this study are listed in Table 1 and compared with values in the
literature.17−19
Samples of the wastewater were collected from a local petroleum
refinery (Midland Refineries CompanyAl-Daura Refinery, Iraq)
and the samples were preserved at room temperature in dark color
plastic containers. The refinery wastewater samples characteristics
are given in Table 2.
Two different initial phenol concentrations from Real refinery
wastewater samples were tested namely 94 and 43 (± 0.5) mg/L.
A series of different initial concentrations of phenol solution was
prepared from phenol stock solution for the aqueous wastewater
ranging from 100 mg/L to 300 mg/L.
2.2. Density Measurements. The densities of the pure
chemicals were measured with an Anton Paar Digital Density
Meter (model DMA 602) at 298.15 K, except for sulfolane which
was measured at 303.15 K (uncertainty ± 1·10−4 g·cm−3). The
details of using this instrument were described elsewhere.20
2.3. Refractive Index Measurements. Refractive indices,
nD of the pure component liquid and mixtures were measured at
298.15 K using Abbe refractometer (Tafesa, Germany) by the
reflection method using sodium line (λ = 589.3 nm) with a
precision of the reading of ± 0.0002. It was calibrated before
measuring the refractive indices of sample using standard pure
liquids (uncertainty ± 5·10−4).
In all refractive index measurements, the temperature was
kept constant within ± 0.01 K using a Schott-Gerate CT 1150
thermostat water bath, and a Hewlett-Packard model 201 A
quartz thermometer.
2.4. Viscosity Measurements. The viscosities of the
pure liquids were determined using a suspended Ubbelhode
viscometer in a bath controlled to ± 0.01 K at specific
1550
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 2. Determination of the tie lines.
Figure 3. Plots of refractive indices versus concentration of solvents for
phenol + water + sulfolane system at 308.15 K.
Figure 4. Equilibrium compositions in a phenol + water binary system
with an upper critical solution temperature.
temperature (uncertainty ± 3·10−4 cp). The times were
determined electronically using an electronic timer (SchottGerete model AVS300) with a precision ± 0.01 s.
2.5. Apparatus and Procedure. 2.5.1. Determination of
the Binodal Curves. The binodal curves of the ternary system
were determined by the apparatus of titration as described by
Haddad and Admister.21 The cloud point is the temperature at
which haziness is first observed at the bottom of the test jar. The
apparatus shown diagrammatically in Figure 1 was kept within
± 0.1 °C of the experiment temperature by circulating water
through a constant temperature bath. A microburet calibrated to
0.01 mL was used to ensure that the liquid was perfectly titrated.
During the addition, the solution in the jacket jar was mixed with
a Teflon-coated magnetic rod.
The binodal curve of the ternary system was determined by the
method of Othmer et al.22 The phenol was placed in the sample
bottle inside the constant temperature water jacket and was
titrated using a microburet while the solution was being stirred by
a magnetic stirring bar. As soon as the cloud point was reached,
the liquid was reclarified by a slow titrating solvent. Water was
then added to make the liquid cloudy once again. The procedure
was repeated many times, and the successive cloud points were
connected up to produce a binodal curve.
2.5.2. Determination of the Tie Lines. To determine lines of
conjugate layers, a ternary two phase system was analyzed by the
so-called cross-section method by Radecki et al.23 The method
consists of plotting the refractive index of a mixture against
concentration of the components at a constant ratio of the remaining two components. As known, sections A, B, and C (as
shown in Figure 2), developed from a point of triangle
representing a solvent concentration in a mixture are loci of
1551
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 5. Binodal curve and tie lines for ternary system (P + W + EG) at 308.15 K.
Figure 6. Binodal curve and tie lines for ternary system (P + W + DEG) at 308.15 K.
points lie on three sections (A, B, and C) but they belong to the
same tie lines. Composition of the conjugate layers is read from
the intersections of the tie lines with a binodal curve illustrating
the equilibrium between one- and two-phase systems.
To obtain the average composition of each sample, as well as to
investigate the repeatability, all the measurements were repeated
three times in the same condition. The composition analysis
deviation in the measurement was less than 2 %.
points characterized by a constant ratio of the remaining two
components. For each section, a series of mixtures is prepared
which occurs in a two-phase region and differs in the solvent
content [70:30 (A), 50:50 (B) and 40:60 (C)]. The solution was
stirred for 3 h at constant temperature and left overnight to reach
the ternary liquid−liquid equilibrium. As the concentration of
solvent was increased, samples were taken from each of the two
phases and the refractive indices of each equilibrium phase are
measured and the relationship between the refractive index and
solvent concentration for a given phase is obtained as shown in
Figure 3.
Since the composition of equilibrium phases lying on one tie
line is constant, arbitrary values of the refractive index (nD)
shown in Figure 3 may be chosen in the plot of nD vs (Csolvent)
which indicates the curves of the corresponding points
determining the solvent content in the phases considered. The
3. RESULTS AND DISCUSSION
3.1. Binary Systems. The experimental solubility data (mg/L)
of the phenol + water binary mixtures is plotted in Figure 4. The
critical solution temperature for this binary system is 74.0 °C.
From Figure 4 it can be seen that the two-phase region of
phenol with water ranges between 23 % and 91 %. The temperature
1552
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 7. Binodal curve and tie lines for ternary system (P + W + PEG) at 308.15 K.
Figure 8. Binodal curve and tie lines for ternary system (P + W + DMSO) at 308.15 K.
35 °C was chosen for determining liquid−liquid equilibria for all
systems.
3.2. Ternary Systems. 3.2.1. Mutual Solubility. The liquid−
liquid equilibria data (mutual solubility data) of the following 10
ternary systems have been studied at 308.15 K.
1. ethylene glycol (EG) + water (W) + phenol (P)
2. diethylene glycol (DEG) + water (W) + phenol (P)
3. poly(ethylene glycol) (PEG) + water (W) + phenol (P)
4. dimethylsulfoxide (DMSO) + water (W) + phenol (P)
5. sulfolane (TMS) + water (W) + phenol (P)
6. ethylene glycol (EG) + water (W) + p-chlorophenol (ClP)
7. diethylene glycol (DEG) + water (W) + p-chlorophenol (ClP)
8. poly(ethylene glycol) (PEG) + water (W) + p-chlorophenol
(ClP)
9. dimethylsulfoxide (DMSO) + water (W) + p-chlorophenol
(ClP)
10. sulfolane (TMS) + water (W) + p-chlorophenol (ClP)
Mutual solubility curves and mixture compositions for these
ternary systems have been plotted on a triangular diagram as
shown in Figures 5 to 14.
The minimum concentration (in % mass) for the solubility of
the phenol over the whole composition range in the mixture
(water + phenol + solvent) was found to be 9.970, 48.765,
37.262, 23.432, and 46.955 for EG, DEG, PEG, DMSO,
and TMS, respectively, and that for the mixture (water +
p-chlorophenol + solvent) it was found to be 23.417, 60.401,
34.019, 35.973, and 57.527 for EG, DEG, PEG, DMSO, and
TMS respectively. This reflects the magnitude of the area of the
two-phase region. This region increases for the systems (water +
phenol + solvent) in the order DEG > TMS > PEG > DMSO >
EG, while for the systems (water + p-chlorophenl + solvent) in
the order DEG > TMS > DMSO > PEG > EG. In the two sets of
data, the order of immiscibility is similar except for interchange of
DMSO and PEG. It seems to be that PEG is more interactable
1553
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 9. Binodal curve and tie lines for ternary system (P + W + TMS) at 308.15 K.
Figure 10. Binodal curve and tie lines for ternary system (ClP + W + EG) at 308.15 K.
with phenol than with p-chlorophenol. However, the shape of
binodal curve seems to be strongly dependent on the type of
solvent used. These results are in contrast with these solvents in
their ternary systems containing aromatic and aliphatic hydrocarbon mixtures.24−27
3.2.2. Effect of Temperature. The binodal curve results for the
ternary system (phenol + H2O + PEG) at 308.15, 318.15, and
328.15 K are presented in Figure 15 as a representative
example of the 10 systems. It can be seen from this figure that
the heterogeneous region at temperature 308.15 K is greater
than that of the other temperature. This is the reason why
we chose the temperature 308.15 K for the investigation
of the liquid−liquid equilibrium data for the systems under
consideration.
3.2.3. Tie-Line Data. The liquid−liquid equilibrium data (tieline data) indicating the composition of the two phases (solvent-rich
phase and water-rich phase) obtained experimentally for the 10
ternary systems at 308.15 K were plotted on equilateral triangles
following the method of Francies28 and presented in Figure 5 to
14. These data for the systems fit well on the binodal curves
indicating the accuracy of the experimental tie-line data. The
slopes of the tie lines of the systems (phenol + water + solvent)
are inclined toward the solvent in the case of three systems only,
namely (EG + W +P), (PEG + W + P), and (DMSO +W + P);
thus, indicating the separation of phenol from water by extraction
with these solvents can be achieved in fewer stages.
3.2.3.1. Tie-Line Correlations. The tie-line data of the 10
ternary mixtures were correlated by Bachman,29 Hand,30 and
Othmer and Tobias31 as follows:
The first empirical equation describing the distribution of
components in the three components two-phase liquid systems
was due to Bachman.
1554
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 11. Binodal curve and tie lines for ternary system (ClP + W + DEG) at 308.15 K.
Figure 12. Binodal curve and tie lines for ternary system (ClP + W + PEG) at 308.15 K.
⎛ xS,S ⎞
⎟⎟
xS,S = a1 + b1⎜⎜
⎝ x W,W ⎠
⎛ 1 − xP,S ⎞
⎛ 1 − x W,W ⎞
⎟⎟ = a3 + b3 log⎜⎜
⎟⎟
log⎜⎜
⎝ x P,S ⎠
⎝ x W,W ⎠
(1)
where xS,S and xW,W are the concentration (mass fraction
percent) of components solvent (S) in the solvent-rich phase and
water (W) in the water-rich phase, respectively; a1 and b1 are
constants.
Hand correlated the concentration of the solute in two
conjugate solutions by the following equation:
⎞
⎛ x P,S ⎞
⎛x
⎟⎟ = a 2 + b2 ·log⎜⎜ P,W ⎟⎟
log⎜⎜
⎝ xS,S ⎠
⎝ x W,W ⎠
(3)
where a3 and b3 are constants.
The parameters aj and bj (j = 1−10) of eqs 1 to 3 were obtained
by using the method of the Maximum Likelihood Principle
technique32 starting from the experimental tie-line data. The
parameters and the correlation coefficients, Rj, are given in
Table 3. Since the data show little scattering from a straight line,
they are judged acceptable on an empirical basis, indicating internal
consistency of the experimental data.
3.2.4. Distribution Coefficient and Selectivity. Selectivity is
the ability of solvent to dissolve one material in preference
to another. This is of prime importance, since the greater
the selectivity the more easily a desired separation can be
made.
(2)
where xP,S and xP,W are the mass fractions of phenol and water in
the solvent-rich phase and the water-rich phase, respectively; a2
and b2 are constants.
Othmer and Tobias proposed the following correlation:
1555
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 13. Binodal curve and tie lines for ternary system (ClP + W + DMSO) at 308.15 K.
Figure 14. Binodal curve and tie lines for ternary system (ClP + W + TMS) at 308.15 K.
(kW) of the phenol and water, respectively, and the selectivity β of
the solvent. The distribution coefficient of phenol and water is
given by the formula:
The effectiveness of the solvent for extraction can be expressed
in terms of the distribution coefficient of the solute and the
selectivity of the solvent.
Capacity or solvency means the ability of the solvent to dissolve
reasonable amounts of the material to be separated while still maintaining a good selectivity. Solvent capacity has an even greater
influence on the extraction process. If the capacity is too low, an
excessive amount of solvent is required for a given separation so the
capacity determines the rate of circulating solvent and the size of most
of the plant equipment. In addition, the operating costs are affected by
the heat needed for heating the solvent from extraction to distillation
temperature. Most solvents with high capacity have low selectivity and
vice versa. So, if the solvent has high capacity and selectivity, it is
considered the more ideal solvent (an excellent solvent).33
The effectiveness of the solvent for the extraction of phenol
can be expressed in terms of the distribution coefficient (kP) and
kP =
x P,S
phenol mass % in solvent layer
=
phenol mass % in water layer
x P,W
(4)
kW =
x W,S
water mass % in solvent layer
=
water mass % in water layer
x W,W
(5)
The selectivity β which is a measure of the ability of the solvent
for the removal of phenol from the phenol−water solution is
expressed in terms of the ratio of phenol mass percent in solventfree solvent layer to phenol mass percent in solvent-free water
layer as follows:
1556
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 15. The effect of temperature on liquid−liquid equilibria of P + W + PEG ternary system.
Table 3. Results of the Maximum Likelihood Principle Analysis for the Systems W + Solvents + P or ClP at 308.15 K
correlation
Bachman
solvent
a1
b1
Hand
R1
a2
b2
Othmer and Tobais
R2
a3
b3
R3
0.992
0.975
0.996
0.990
0.989
0.116
0.243
0.160
0.240
0.321
0.507
0.788
0.606
0.651
0.721
0.989
0.969
0.970
0.965
0.984
0.979
0.976
0.991
0.996
0.988
0.464
0.103
0.157
0.245
0.107
1.190
0.521
0.768
0.915
0.526
0.980
0.933
0.992
0.986
0.979
P + W + Solvent
EG
DEG
PEG
DMSO
TMS
−0.285
−0.435
−0.416
−0.376
−0.400
0.399
0.466
0.458
0.440
0.450
0.987
0.999
0.997
0.998
0.998
EG
DEG
PEG
DMSO
TMS
−0.469
−0.413
−0.415
−0.427
−0.438
0.480
0.461
0.486
0.454
0.473
0.999
0.966
0.999
0.999
0.991
1.540
1.020
0.555
0.476
1.971
1.866
2.507
2.213
1.942
1.543
ClP + W + Solvent
0.650
0.684
0.343
0.416
0.662
0.686
0.422
0.531
0.451
0.503
increases, the value of (ln KP − ln KW) therefore, decreases as the
concentration of phenol in one of the phases increases.
Selectivity decreases with the increase of concentration of the
component to be extracted.
The distribution coefficients for phenol or p-chlorophenol
between water and solvent indicate that the preference of
extraction of phenol lies in the following order: DEG > PEG >
TMS > DMSO > EG, while for p-chlorophenol it lies in the
following order: DEG > TMS > PEG > DMSO > EG. The
selectivity for these two systems follows the same trend TMS >
DEG > PEG > DMSO > EG which can be obtained from the
selectivity diagram illustrated in Figures 18 and 19 which reveal
that TMS and DEG are more selective for phenols in the
presence of water.
3.2.5. Plait Point. The plait point is the point on the binodal
curve at which the tie lines connect two compositions of the
equilibrium phases become extremely short. At this point the two
phases in equilibrium become identical in composition.
Interpolation of tie lines will lead to estimate the position of
the plait point. The plait point composition on the binodal curve
was determined from the formula:34
x P,S
phenol mass % in solvent free solvent layer
β=
=
phenol mass % in solvent free water layer
x P,S + x W,S
x P,W
x P,W + x W,W
(6)
The obtained results of the distribution coefficients and the
selectivity are listed in Table 4. The distribution coefficient
of phenol and the selectivity of the solvents, as calculated from
the experimental tie-line data using eqs 4 to 6 are plotted in
Figures 16, 17, 18, and 19 respectively.
The selectivity of a solvent as mentioned above is its ability to
separate the components of a given mixture. The selectivity
depends mainly on the distribution coefficient of phenol (KP)
and water (KW) and thus could be defined as the difference between the logarithms of KP and KW. Separation by extraction
becomes easier as this value increases. The selectivity of the
solvent toward components phenol and water depends on the
two factors affecting the distribution coefficient values: the concentration and the temperature, (where temperature is constant
in this study, this value is affected only by concentration). As the
concentration of the phenol in a heterogeneous ternary system
1557
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Table 4. The Distribution Coefficients and Selectivity for the Systems Solvent + W + P or ClP at 308.15 K
KP
KW
EG + W + P
0.670
0.244
0.215
0.218
DEG + W + P
0.684
0.318
0.265
0.254
0.248
0.245
0.242
PEG + W + P
0.831
0.521
0.461
0.414
0.362
0.344
0.308
DMSO + W + P
0.902
0.400
0.344
0.338
0.317
TMS + W + P
0.736
0.455
0.413
0.361
0.329
0.322
0.291
EG + W + ClP
0.926
0.410
22.369
13.346
12.994
12.011
5.441
5.588
13.691
22.103
18.994
16.899
10.674
3.977
3.455
4.286
5.483
7.213
7.569
9.722
1.587
3.579
5.154
5.953
6.112
6.460
3.458
4.851
7.388
8.810
10.147
12.074
6.242
10.080
x P,W
x W,W
=
x P,S
x W,S
=
xP
xW
β
KP
24.314
13.928
13.479
12.489
12.861
15.364
16.241
13.164
14.132
6.209
6.020
15.089
23.837
20.584
18.214
11.861
11.769
12.080
35.709
27.514
26.716
25.798
17.780
3.970
2.902
3.636
4.741
6.315
6.651
8.915
7.987
8.914
11.716
14.930
21.351
20.300
21.858
1.551
3.209
4.643
5.247
5.588
18.172
19.958
38.633
38.835
35.583
28.664
23.320
6.699
2.840
3.747
5.917
7.492
8.544
10.883
10.582
6.182
4.336
44.144
52.757
56.582
29.145
6.403
10.344
KW
EG + W + ClP
0.246
0.272
0.231
0.229
0.223
DEG + W + ClP
0.357
0.127
0.135
0.132
0.130
0.139
0.143
PEG + W + ClP
0.790
0.260
0.248
0.221
0.212
0.206
0.215
DMSO + W + ClP
0.802
0.360
0.283
0.248
0.226
0.217
0.208
TMS + W + ClP
0.535
0.196
0.178
0.150
0.167
0.172
0.174
β
15.288
16.328
16.936
13.696
14.656
18.200
13.645
41.678
33.065
33.546
31.063
21.762
8.149
9.857
12.778
16.249
23.229
21.885
22.570
20.005
21.416
41.713
43.261
39.976
31.203
25.050
13.753
6.542
4.922
49.352
56.607
60.842
32.008
the method of Treybal with the correlation of Hand requires
experimental binodal curve data in addition to the tie-line data.
This is why the estimation of the compositions of plait points for
the ternary liquid−liquid equilibria (LLE) systems by the two
methods are different.
In general, the plait points resulting from correlation of the tie
lines with the thermodynamic models are not located on the
experimentally determined binodal curves.
3.3. Selection of Solvent. The result of an extraction is
affected in the first place by the extent of the two-phase area. For
the separation of a given mixture, the size of this area is dependent on temperature and the nature of the solvent (in this study
the temperature is constant).
In this study the slopes of tie lines are steeper in the DEG +
water system than the corresponding solvents (see Figures 5
to 14), the plait points are located in the region of mixtures
containing more DEG in the case of DEG + water + phenol or
p-chlorophenol system. So the separation of P or ClP form W + P
or ClP by extraction with DEG can be achieved in fewer stages
than the separation of P or ClP by other solvents. On the other
hand, the solubility gaps (Figures 5 to 14), of DEG + W + P or
(7)
The simplest methods for the interpolation and correlation of tie
lines are those based on the construction of the conjugate line,
which can be drawn in many ways.35 These methods are to be
preferred since they are quick and reliable, and the position of the
plait point is readily found.
Values of the plait points of the ternary systems under study
are listed in Table 5.
Treybal et al.,36 proposed a method for estimating plait points
using the coordinate system of Hand. The estimation of the plait
points for the ternary systems using Hand’s correlation is also
listed in Table 5 for comparison.
The experimental method (construction method) is a
graphical method which uses a conjugate line to interpolate tie
lines and it is helpful for interpolation and limited extrapolation
when equilibrium data are scarce. With this method, plait points
of several ternary systems are estimated, and it turns out that
the results are favorably comparable with the results obtained by
the method proposed by Treybal. However, it needs to emphasize that the experimental method uses only tie-line data while
1558
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 16. Distribution of phenol between water and solvent layer at
308.15 K.
Figure 18. Selectivity curves for solvents + water + phenol at 308.15 K.
Figure 19. Selectivity curves for solvents + water + p-chlorophenol at
308.15 K.
Figure 17. Distribution of p-chlorophenol between water and solvent
layer at 308.15 K.
efficient solvent for the extraction of P or ClP from their mixtures
with water.
DEG has the highest distribution coefficient and selectivity
when it is compared with the other solvents. Although, TMS has
the highest distribution coefficient of the other solvents rather
than DEG and highest selectivity, it is unsuitable as solvent, since
the differences between the boiling point and density of the TMS
and P or ClP are too small and also because of its toxicity. The
boiling point and density of DEG differ greatly from those of
phenols (see Table 1), and so DEG is a more suitable solvent for
extraction of phenol and p-chlorophenol from their mixtures
with water.
Phenols consist of two parts, an alkyl/aryl group and a
hydroxyl group (difunctional compounds; the hydroxyl group
and the aromatic ring). The properties of phenols are chiefly due
ClP (heterogeneous region) is larger than those of the heterogeneous region of the other solvents.
For an extraction study, the selection of a solvent depends on
the solvent power measured by the solute distribution coefficient
and also on its selectivity. In the case of recovery of solute from
diluent (water), a solvent with the largest possible distribution coefficient and highest selectivity toward phenol or
p-chlorophenol is preferred. The distribution coefficient of the
systems (solvents + W + P or ClP) increases in the order DEG >
PEG > TMS > DMSO > EG or DEG > TMS > DMSO > PEG >
EG, respectively, while the selectivity of both systems, respectively,
is in the following sequence TMS > DEG > PEG > DMSO > EG.
However, comparisons of the solubility gaps (Figures 5
to 14), distribution coefficient (Figures 16 and 17) and selectivity
(Figures 18 and 19) show that the DEG solvent is the most
1559
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
and electron releasing groups decrease it. In substituted phenols,
the presence of electron withdrawing groups such as a chloro
group, enhances the acidic strength of the phenol. This effect is
more pronounced when such a group is present at the ortho and
para positions. It is due to the effective delocalization of the
negative charge in the phenoxide ion.
Compounds that can form intermolecular hydrogen bonds
have higher melting points and boiling points than compounds
that cannot, such as p-chlorophenol (44 °C, 220 °C). It forms an
intermolecular hydrogen bond between the chloro group and the
phenol function which reduces its ability to form intermolecular
bonds. While a halogen atom or an electron-withdrawing group
increases the acidity (pKa(PhOH) = 9.95), the effect greatly
varies with the position. The ortho isomers are usually less acidic
than the para isomers because an intermolecular hydrogen bond
makes it more difficult to remove the phenolic hydrogen. The
para and the meta isomers dissolve well in polar solvents and
poorly in nonpolar solvents. This is why PEG is more interactable with phenol than with p-chlorophenol, because of the
electron withdrawing inductive effect of chlorine.
The selection of a solvent for an extraction study depends on
the solvent power measured by the solute distribution coefficient
and also on its selectivity. The solvent must be miscible with the
separated components. The basic concept of miscibility is “like
dissolves like” consisting of hydrogen bonding, polar, and
dispersion interaction forces. In considering whether A component will dissolve in a liquid, three possible interactions should
be considered. In two-component systems namely A and B, there
are three interaction among these two components which are
A−A, B−B, and A−B. If A−B interaction is strong or comparable
to A−A or B−B associations then two liquid components are
likely to be miscible and mixed with each other.
The first interaction force between two molecules, which is
common for all component pairs, is the instantaneous dipole−
induced dipole interaction or dispersion force. When two components are attached together, the partial positive charge of one
dipole will attract the partial negative in the neighboring
molecule or vice versa.
The second interaction force is the permanent dipole and
permanent dipole interactions (hydrogen bonding). If two atoms
constituting a bond have significantly different electronegativities, the bond will be permanently polar and produce a permanent polar molecule. Generally, this type of interaction force
occurs with a hydrogen compound which is attached with oxygen,
nitrogen, and the halide group. The last one is the interaction force
Table 5. Composition of the Plait Point for the Ternary
Systems Solvents + Water + Phenol or P-Chlorophenol
at 308.15 K
experimental method
solvent
xS
xW
xP(or ClP)
Treybal method
xS
xW
xP(or ClP)
0.698
0.712
0.743
0.699
0.067
0.203
0.206
0.159
0.798
0.718
0.814
0.792
0.789
0.648
0.068
0.201
0.147
0.115
0.135
0.231
P + W + Solvent
EG
DEG
PEG
DMSO
0.189
0.080
0.070
0.171
0.784
0.718
0.796
0.740
EG
DEG
PEG
DMSO
TMS
TMS
0.126
0.051
0.028
0.069
0.038
0.126
0.830
0.779
0.899
0.845
0.819
0.657
0.028
0.235
0.202
0.085
0.135
0.051
0.089
0.142
ClP + W + Solvent
0.045
0.134
0.170
0.081
0.073
0.039
0.086
0.093
0.150
0.076
0.216
0.121
to the hydroxyl group. The nature of the alkyl and aryl groups
simply modify these properties. The solubility of phenols in
water is due to their ability to form intermolecular hydrogen
bonds with water molecules. The solubility decreases with an
increase in size of alkyl/aryl (hydrophobic) groups. Properties
like cohesion (intermolecular force between like molecules) and
adhesion (intermolecular force between unlike molecules) are
also a result of weak intermolecular forces. The physical properties of phenols are strongly influenced by the hydroxyl group,
which permits phenols to form hydrogen bonds with other
phenol molecules and with water. Thus, phenols have higher
melting points and boiling points and are more soluble in water.
The presence of the −OH group in phenols activates the
aromatic ring toward electrophilic substitution and directs the
incoming group to the ortho and para positions due to a
resonance effect. Because of the higher electronegativity of sp
hybridized carbon of phenol to which −OH is attached, electron
density decreases on oxygen. This increases the polarity of the
O−H bond and results in an increase in ionization of phenols.
Thus, phenols are polar compounds. The −OH group of phenols
makes it possible for them to participate in hydrogen bonding.
This contributes to the higher boiling points and greater watersolubility of phenolic compounds
Phenols are acidic in nature. In fact, phenols are Brönsted
acids; that is, they can donate a proton to a stronger base. Electron withdrawing groups in phenol increase its acidic strength
Table 6. UNIQUAC Parameter Uij (J·mol−1) and RMSD Values for the Systems Solvents + W + P or ClP at 308.15 Ka.
solvent
U11
U22
U33
U12
U13
U23
RMSD
5.942
6.642
4.021
3.156
6.524
3.215
1.342
2.308
4.081
4.376
5.442
2.711
2.327
6.272
2.094
3.656
3.946
3.716
4.411
2.627
5.123
1.865
1.311
5.231
9.286
6.314
6.608
6.731
5.061
2.905
P + W + Solvent
a
EG
DEG
PEG
DMSO
3.968
4.393
1.219
5.550
1.497
7.984
9.066
1.427
EG
DEG
PEG
DMSO
TMS
TMS
3.563
6.137
1.783
2.764
3.455
1.504
3.527
7.527
1.021
7.662
9.806
1.151
6.865
1.863
3.658
3.966
2.935
1.864
3.588
1.4129
ClP + W + Solvent
2.953
2.746
2.779
4.581
1.0211
4.301
1.877
1.667
1.258
1.657
4.258
8.834
Note: all U values are multiplied by 103.
1560
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Table 7. NRTL Parameter gij (J mol−1), αij and RMSD Values for the Systems Solvents + W + P or ClP at 308.15 Ka
g11
solvent
g22
g33
g12
g13
g23
α12
α13
α23
RMSD
1.110
1.099
1.134
9.211
0.456
0.460
0.460
0.360
0.459
0.425
0.255
0.460
0.431
0.431
0.388
0.429
0.112
0.067
0.073
0.099
1.346
1.140
1.123
1.180
1.159
1.194
0.444
0.446
0.460
0.460
0.457
0.455
0.186
0.403
0.437
0.432
0.409
0.417
0.430
0.437
0.414
0.449
0.425
0.395
0.687
0.093
0.113
0.512
0.169
0.029
P + W + Solvent
a
EG
DEG
PEG
DMSO
2.740
2.430
3.535
4.816
1.729
2.610
0.548
3.775
7.570
1.902
3.288
2.521
6.314
6.299
2.373
4.997
EG
DEG
PEG
DMSO
TMS
TMS
6.558
1.631
2.508
3.368
2.328
3.743
1.016
8.464
3.461
2.660
1.789
1.611
1.404
1.012
2.398
2.568
1.092
3.384
5.537
6.677
8.751
7.098
6.932
2.845
6.969
8.596
7.246
9.104
ClP + W + Solvent
8.045
8.957
6.089
7.966
9.305
9.257
Note: all g values are multiplied by 103.
Table 8. Calculated Tie Lines Using the UNIQUAC Model for
Solvents + W + P Systems at 308.15 K
solvent-rich layer
solvent
EG
DEG
PEG
DMSO
TMS
Table 9. Calculated Tie Lines Using the UNIQUAC Model for
Solvents + W + ClP systems at 308.15 K
water-rich layer
solvent-rich layer
xS
xW
xP
xS
xW
xP
0.090
0.211
0.119
0.055
0.263
0.465
0.428
0.361
0.289
0.207
0.090
0.021
0.269
0.237
0.187
0.092
0.066
0.030
0.099
0.134
0.083
0.020
0.009
0.163
0.331
0.258
0.180
0.136
0.089
0.018
0.824
0.400
0.360
0.356
0.674
0.335
0.378
0.389
0.396
0.398
0.405
0.923
0.457
0.489
0.536
0.531
0.511
0.540
0.763
0.441
0.468
0.488
0.472
0.824
0.355
0.337
0.371
0.415
0.374
0.330
0.085
0.387
0.520
0.588
0.062
0.199
0.192
0.248
0.313
0.393
0.504
0.540
0.273
0.273
0.276
0.376
0.421
0.29
0.164
0.424
0.447
0.490
0.518
0.011
0.312
0.404
0.448
0.448
0.536
0.651
0.055
0.059
0.043
0.032
0.077
0.231
0.149
0.132
0.097
0.073
0.038
0.015
0.193
0.181
0.166
0.153
0.194
0.123
1.040
1.470
1.250
1.680
1.077
0.112
0.314
0.287
0.247
0.234
0.211
0.098
0.728
0.446
0.471
0.505
0.624
0.359
0.386
0.402
0.431
0.463
0.493
0.772
0.328
0.320
0.321
0.351
0.365
0.376
0.655
0.353
0.368
0.377
0.413
0.635
0.260
0.304
0.350
0.354
0.425
0.599
0.216
0.494
0.485
0.461
0.298
0.409
0.464
0.465
0.470
0.463
0.467
0.211
0.477
0.497
0.512
0.495
0.484
0.500
0.240
0.498
0.506
0.454
0.508
0.252
0.425
0.408
0.401
0.411
0.362
0.301
solvent
EG
DEG
PEG
DMSO
TMS
produced by a permanent dipole and induced dipole interactions.
This type of interaction force can be found between a hydrocarbon and highly polar components such as HCl, phenols,
alcohol component, and carboxylic acid component.
The solvent prefers to pair with the component that uses the
lowest cohesive energy when compared with the other two pairs.
Therefore, it can be concluded that the lower is the mixing
energy, the easier it is for the two components to attach together.
To find the suitable solvent for a given system, one of the
interaction forces between the components in the closed boiling
water-rich layer
xS
xW
xP
xS
xW
xP
0.123
0.206
0.291
0.151
0.098
0.066
0.042
0.423
0.643
0.614
0.511
0.423
0.316
0.153
0.027
0.314
0.376
0.296
0.198
0.097
0.039
0.148
0.528
0.468
0.327
0.228
0.132
0.038
0.290
0.607
0.364
0.558
0.324
0.196
0.082
0.871
0.691
0.701
0.689
0.674
0.661
0.657
0.479
0.151
0.205
0.228
0.250
0.259
0.274
0.903
0.528
0.492
0.466
0.463
0.427
0.371
0.838
0.397
0.354
0.348
0.335
0.316
0.303
0.632
0.169
0.201
0.214
0.254
0.270
0.278
0.005
0.101
0.006
0.158
0.227
0.271
0.299
0.097
0.201
0.180
0.260
0.325
0.424
0.571
0.069
0.157
0.131
0.237
0.338
0.478
0.588
0.012
0.074
0.177
0.324
0.436
0.550
0.658
0.076
0.223
0.434
0.226
0.421
0.533
0.638
0.052
0.0127
0.093
0.071
0.048
0.021
0.009
0.097
0.228
0.153
0.104
0.074
0.060
0.450
0.014
0.135
0.089
0.071
0.057
0.041
0.009
0.100
0.122
0.094
0.071
0.059
0.044
0.030
0.104
0.288
0.169
0.156
0.088
0.054
0.0.38
0.903
0.379
0.252
0.304
0.280
0.293
0.298
0.408
0.307
0.346
0.377
0.391
0.427
0.450
0.809
0.290
0.321
0.338
0.364
0.414
0.529
0.761
0.506
0.517
0.540
0.554
0.592
0.626
0.506
0.317
0.383
0.369
0.468
0.528
0.560
0.044
0.493
0.653
0.624
0.671
0.684
0.692
0.493
0.464
0.499
0.517
0.534
0.11
0.504
0.176
0.574
0.588
0.589
0.578
0.543
0.461
0.138
0.370
0.388
0.388
0.386
0.363
0.343
0.389
0.394
0.447
0.473
0.442
0.416
0.400
point system and solvent has to be estimated. Thus, the external
factor contains two effects; that is, the electrostatic stabilization
of the ionic form and the donor−acceptor interaction of solvent
molecules with the free electron pair of the phenolate oxygen
atom.
1561
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Table 10. Calculated Tie Lines Using the NRTL Model for
Solvents + W + P Systems at 308.15 K
solvent-rich layer
solvent
EG
DEG
PEG
DMSO
TMS
Table 11. Calculated Tie Lines Using the NRTL Model for
Solvents + W + ClP Systems at 308.15 K
water-rich layer
solvent-rich layer
xS
xW
xP
xS
xW
xP
0.127
0.140
0.095
0.059
0.243
0.330
0.282
0.248
0.212
0.171
0.120
0.016
0.163
0.139
0.111
0.073
0.063
0.038
0.089
0.119
0.067
0.019
0.009
0.153
0.273
0.209
0.147
0.125
0.077
0.016
0.613
0.232
0.217
0.232
0.563
0.200
0.201
0.205
0.211
0.220
0.232
0.764
0.276
0.271
0.274
0.259
0.249
0.250
0.658
0.258
0.255
0.264
0.254
0.613
0.207
0.207
0.222
0.227
0.236
0.243
0.259
0.627
0.687
0.708
0.192
0.469
0.516
0.546
0.576
0.608
0.647
0.219
0.560
0.589
0.614
0.666
0.627
0.711
0.252
0.629
0.677
0.715
0.736
0.234
0.518
0.583
0.629
0.646
0.686
0.740
0.037
0.089
0.053
0.030
0.103
0.325
0.226
0.191
0.133
0.088
0.029
0.021
0.320
0.309
0.279
0.192
0.157
0.097
0.116
0.177
0.156
0.177
0.079
0.120
0.381
0.354
0.302
0.254
0.244
0.110
0.946
0.771
0.779
0.776
0.878
0.600
0.728
0.764
0.810
0.837
0.860
0.933
0.543
0.578
0.628
0.719
0.750
0.812
0.732
0.604
0.676
0.696
0.769
0.855
0.444
0.494
0.585
0.646
0.673
0.814
0.016
0.139
0.166
0.192
0.018
0.073
0.045
0.043
0.056
0.073
0.110
0.045
0.136
0.112
0.091
0.088
0.091
0.089
0.151
0.218
0.167
0.126
0.151
0.024
0.173
0.150
0.112
0.098
0.081
0.074
solvent
EG
DEG
PEG
DMSO
TMS
The polarity difference between the solvent and phenols
should not be too high for effective extraction. A low polarity
difference between the solvent and phenols results in attractive
forces between the different molecules, and as a result the
phenols are preferentially pulled toward the solvent. Indeed the
hydrogen bonds system formation and the polarity difference
between the solvents and phenols support the above arguments.
It is clear from what has been considered that the extent in
which mixtures deviate from ideality governs the distribution of a
solute (phenol) between two solvents (solvent + water).
However, it is then possible to predict the nature of extraction
of phenols from the mixture on the basis of their hydrogen
bonding potentialities. Because, in the case of our ternary systems
using five solvents the main molecular interactions are the
hydrogen bond forces. Thus, the ability of extraction depends
mainly on the solvent used. Therefore, hydrogen bond molecular
interactions are the unique force predominant in the extraction
process. It is worthwhile to mention that the liquid−liquid
equilibrium in the presence of water is determined by intermolecular forces, predominantly hydrogen bonds. The addition
of solvents to a mixture of water + phenols enhances the formation of the hydrogen-bonded system.
3.4. Data Correlation. Thermodynamic models, namely
the universal-quasi-chemical (UNIQUAC) equation proposed
by Anderson and Prausnitz37 and the nonrandom-two-liquid
(NRTL) equation proposed by Renon and Prausnitz38 activity
water-rich layer
xS
xW
xP
xS
xW
xP
0.097
0.170
0.251
0.157
0.122
0.070
0.030
0.436
0.395
0.347
0.310
0.312
0.258
0.201
0.032
0.223
0.200
0.174
0.159
0.130
0.036
0.203
0.244
0.219
0.207
0.205
0.152
0.076
0.329
0.402
0.304
0.317
0.204
0.151
0.113
0.853
0.343
0.231
0.266
0.248
0.270
0.281
0.302
0.076
0.097
0.107
0.113
0.129
0.143
0.743
0.202
0.199
0.185
0.185
0.185
0.203
0.701
0.302
0.265
0.254
0.243
0.249
0.258
0.448
0.101
0.116
0.108
0.145
0.168
0.181
0.048
0.485
0.516
0.576
0.628
0.659
0.687
0.261
0.527
0.554
0.582
0.574
0.612
0.654
0.223
0.574
0.600
0.639
0.654
0.683
0.760
0.095
0.452
0.515
0.539
0.551
0.598
0.665
0.221
0.496
0.579
0.574
0.650
0.680
0.705
0.065
0.154
0.108
0.068
0.038
0.020
0.013
0.141
0.372
0.270
0.172
0.101
0.074
0.034
0.011
0.190
0.168
0.121
0.071
0.031
0.010
0.073
0.265
0.201
0.112
0.065
0.038
0.015
0.212
0.434
0.203
0.270
0.140
0.070
0.028
0.922
0.763
0.765
0.789
0.759
0.719
0.695
0.828
0.619
0.728
0.806
0.865
0.857
0.860
0.983
0.759
0.793
0.580
0.909
0.955
0.966
0.915
0.664
0.690
0.740
0.764
0.752
0.735
0.876
0.531
0.661
0.723
0.819
0.848
0.860
0.011
0.082
0.126
0.142
0.201
0.260
0.291
0.030
0.007
0.001
0.021
0.032
0.068
0.105
0.004
0.051
0.038
0.028
0.018
0.013
0.023
0.010
0.070
0.108
0.146
0.169
0.209
0.249
0.002
0.033
0.135
0.006
0.039
0.080
0.111
coefficient models, were used to correlate the experimental data
for the 10 systems studied.
The difference between the experimental and calculated mass
fractions was minimized by the use of the objective function F:
n
F=
i−1
xexp
jL (i)
3
2
∑ min ∑ ∑ (xjLexp(i) − xjLcal(i))2
(8)
j=1 L=1
xcal
jL (i)
is the experimental mass fraction,
is the calculated
mass fraction, and n is the number of the experimental tie lines.
This objective function is minimized by the values of the
parameters sought, using both the UNIQUAC model and NRTL
model. Using the method of the Maximum Likelihood Principle
technique,32 the values of the six parameters for UNIQUAC
model: U11, U22, U33, U12, U13, U23 (J·mol−1) were calculated. The
values of the nine parameters for the NRTL model are g11, g22, g33,
g12, g13, g23, a12, a13, a23 (J·mol−1). As for the ternary liquid−liquid
system, they were calculated by using the procedure proposed by
Prausnitz et. al.39 The parameters calculated in this way are listed
in Tables 6 and 7.
The root-mean-square deviation (RMSD) is also included in
the Tables and they are defined as
1562
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
Figure 20. Comparison of experimental tie line data with the calculated values using UNIQUAC (■) and NRTL (▲) model for TMS + W + P at 308.15 K.
1/2 ⎞
⎛ n 3 2 ⎛ exp
cal
(xjL (i) − xjL
(i))2 ⎞ ⎟
⎜
⎟
RMSD = 100⎜∑ ∑ ∑ ⎜⎜
⎟ ⎟⎟
⎜ i=1 j=1 L=1 ⎝
6n
⎠ ⎠
⎝
The binary interaction parameters have been calculated using
the method of Maximum Likelihood Principle technique.
Of the five solvents used to extract the phenol or
p-chlorophenol from wastewater, DEG has the highest selectivity
and distribution coefficient and the greatest differences between
its boiling point and density and those of phenol or
p-chlorophenol. It is therefore an excellent solvent for extracting
phenol or p-chlorophenol from wastewater.
It was found that an increase in solvent concentration decreased the concentration of phenol in the wastewater. This is
expected. At a fixed phenol initial concentration increasing the
solvent concentration provided more solvent capacity for phenol
and hence the removal is enhanced. This effect was stronger for
aqueous wastewater than for refinery wastewater for the same
amount of solvent. This is due to the existence of other
compounds in the real refinery wastewater that compete with
phenol for the amount of solvent, and hence lead to a less amount
of phenol removed from the wastewater.
(9)
The agreement between the experimental and the calculated
values is measured by the (RMSD).
The calculated tie lines data obtained by UNIQUAC and
NRTL are listed in Tables 8 to 11 and one of them is presented in
Figure 20 as a representative example of the other studied ternary
systems. The calculation based on both UNIQUAC model and
NRTL model gave a good representation of the tie lines data for
the systems studied here. However, the NRTL model fitted to
the experimental data is more accurate than the UNIQUAC
model, as can be seen from the results shown in Tables 6 and 7.
The NRTL correlation gives better RMSD than those from
UNIQUAC. It indicates that the liquid−liquid equilibria data are
better correlated with NRTL than UNIQUAC models
■
4. CONCLUSIONS
A survey of the literature indicates that no liquid−liquid
equilibrium data are available for the use of different solvents
EG, DEG, PEG, DMSO, and TMS for extraction of phenol or
p-chlorophenol from wastewater.
Phenol samples are collected from a real petroleum refinery
and experimentally prepared aqueous solution wastewater. The
effectiveness of selective solvent for a given phenol samples was
evaluated.
The consistency of experimental tie-lines data for the ternary
systems (solvents + W + P or ClP) was confirmed for five
solvents using Bachman, Hand, and Othmer-Tobias correlations.
The good fit confirms the reliability of the data.
The calculation based on both the UNIQUAC model and
NRTL model gave a good representation of the tie-line data for
the systems (solvents + W + P or ClP). However, the calculated
values based on the NRTL model are found to be better than
those based on the UNIQUAC model. A relative deviation between experimental and calculated distribution coefficient values
of phenols is less than 2 %.
AUTHOR INFORMATION
Corresponding Author
*Tel.: +964 790 2 89 80 18. E-mail: khalid_farhod@
uotechnology.edu.iq.
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
The author acknowledges Mr. K. M. Ahmed’s efforts for his
support and help during the revision of this manuscript.
■
REFERENCES
(1) Chapman, D. R. Separation Process in Practice; Richard Publisher:
New York, 1961.
(2) Mathews, A. P.; Su, C. A. Prediction of competitive adsorption
kinetics for two priority pollutants. Environ Progress 1983, 2, 257.
(3) Zielke, R.; Pinnavaia, T. Modified Clays for the Adsorption of
Environmental Toxicants: Binding of Chlorophenols to Pillared,
Delaminated, and Hydroxy-Interlayered Smectites. Clays Clay Miner.
1988, 36, 403−408.
1563
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564
Journal of Chemical & Engineering Data
Article
(4) Chapman, D. R. Water Quality Assessment; Champan and Hall:
London, 1992.
(5) WHO Guidelines for drinking-water quality; Vol. 1, Recommendation, World Health Organization: Geneva, 1993.
(6) Elvers, B.; Hawkins, S.; Schulz, G. Ullmanns Encyclopedia of
Industrial Chemistry, 5th ed.; V.C.H.: Weinheim, Germany, 1991; Vol.
A19.
(7) Rippen, G. Umweltchemikalien, Ecobase Media Explorer; Verlage
Cheimi: Weinheium, Germany, 1998.
(8) Weissermel, K.; Arpe, H. J. Industrielle Organische Chemic:
Bedeutende Vor und Zwischenprodukfe, 5th ed.; V.C.H.: Weinheim,
Germany, 1998.
(9) Exon, J. H. Review of Chlorinated Phenols. Vet. Hum. Toxicol.
1984, 26, 508.
(10) Balasubramanian, A.; Venkatesan, S. Removal of phenolic
compounds from aqueous solutions by emulsion liquid membrane
containing ionic liquid [BMIM]+[PF6]− in tributyl phosphate.
Desalination 2012, 289, 27−34.
(11) Yang, C.; Qian, Y.; Jiang, Y.; Zhang, L. Liquid−liquid equilibria for
the quaternary system methyl isobutyl ketone−water−phenol−hydroquinone. Fluid Phase Equilib. 2007, 258, 73−77.
(12) Reis, M. T. A.; Freitas, O. M. F.; Agarwal, S.; Ferreira, L. M.;
Ismael, M. R. C.; Machado, R.; Carvalho, J. M. R. Removal of phenols
from aqueous solutions by emulsion liquid membranes. J. Hazard.
Mater. 2011, 192, 986−994.
(13) Moussavi, G.; Barikbin, B.; Mahmoudi, M. The removal of high
concentrations of phenol from saline wastewater using aerobic granular
SBR. Chem. Eng. J. 2010, 158, 498−504.
(14) Palma, M. S. A.; Paiva, J. L.; Zilli, M.; Converti, A. Batch phenol
removal from methyl isobutyl ketone by liquid−liquid extraction with
chemical reaction. Chem. Eng. Process. 2007, 46, 764−768.
(15) Busca, G.; Berardinelli, S.; Resini, C.; Arrighi, L. Technologies for
the removal of phenol from fluid streams: A short review of recent
developments. J. Hazard. Mater. 2008, 160, 265−288.
(16) El-Naas, M. H.; Al-Zuhair, S.; Alhaija, M. A. Removal of phenol
from petroleum refinery wastewater through adsorption on date-pit
activated carbon. Chem. Eng. J. 2010, 162, 997−1005.
(17) Timmermans, J.; Physicochemical Constants of Pure Organic
Compounds; Elsevier: Amsterdam, Vol. II, 1965.
(18) Riddick, J. A.; Buuger, W. B.; Sakano, T. K. Organic Solvents;
Wiley: New York, 4th ed., 1986.
(19) Assael, M. J.; Trusler, J. P. M.; Tsoalkis, T. F. Thermophysical
Properties of Fluid, An Introduction to their Predication; Imperial College
Press: London, 1998.
(20) Al - Dujaili, A. H.; Awwad, A. M.; Yassen, A. A. Refractive Indices,
Densities, and Excess Properties of 4-(2-Hydroxyethyl) Morpholine +
Water at (298.15, 308.15, and 318.15) K. J. Chem. Eng. Data 2000, 45,
647−649.
(21) Haddad, P. O.; Edmister, W. C. Phase Equilibriums in Acetic
Acid−Diethyl Ketone−Water System. J. Chem. Eng. Data 1972, 17, 275.
(22) Othmer, D. F.; Tobias, P. E. Liquid−Liquid Extraction Data
Partial Pressures of Ternary Liquid Systems and the Prediction of Tie
Lines. Ind. Eng. Chem. 1942, 34, 696−700.
(23) Radecki, A.; Kaczmark, B.; Grzybowski, J. Liquid−Liquid Phase
Equilibrium for Ternary Systems Hexamethyldisiloxane−Acetic Acid
(Propionic Acid)−Water. J. Chem. Eng. Data 1975, 20, 163−165.
(24) Chen, J.; Li, Z.; Duan, L. Liquid−Liquid Equilibria of Ternary and
Quaternary Systems Including Cyclohexane, 1-Heptene, Benzene,
Toluene, and Sulfolane at 298.15 K. J. Chem. Eng. Data 2000, 45,
689−692.
(25) Colombo, A.; Battilana, P.; Ragaini, V.; Bianchi, C. L.; Carvoli, G.
Liquid−Liquid Equilibria of the Ternary Systems Water + Acetic Acid +
Ethyl Acetate and Water + Acetic Acid + Isophorone (3,5,5-Trimethyl2-cyclohexen-1-one). J. Chem. Eng. Data 1999, 44, 35−39.
(26) Aspi, K. K.; Surana, N. M.; Ethirajulu, K.; Vennila, V. Liquid−
Liquid Equilibria for the Cyclohexane + Benzene + Dimethylformamide
+ Ethylene Glycol System. J. Chem. Eng. Data 1998, 43, 925−927.
(27) Wang, W.; Gou, Z.; Zhu, S. Liquid−Liquid Equilibria for
Aromatics Extraction Systems with Tetraethylene Glycol. J. Chem. Eng.
Data 1998, 43, 81−83.
(28) Francis, A. W. Liquid-Liquid Equilibrium; Interscience Publisher:
New York, 1987.
(29) Bachman, I. Tie Lines in Ternary Liquid Systems. Ind. Eng. Chem.
Anal. Ed. 1940, 12, 38.
(30) Hand, D. B. The Distribution of a Consolute Liquid between Two
Immiscible liquids. J. Phys. Chem. 1930, 34, 1961−2000.
(31) Othmer, D. F.; Tobias, P. E. Tie-Line Correlation. Ind. Eng. Chem.
1942, 34, 693−700.
(32) Anderson, T. F.; Abrams, D. S.; Grens, E. A. Evaluation of
Parameters for Nonlinear Thermodynamic Models. AIChE J. 1978, 24,
20.
(33) Alders, L. Liquid−Liquid Extraction, 2nd ed.; Elsevier Publishing
Co.: Amsterdam, The Netherlands, 1959.
(34) International Critical Tables 3; McGraw Hill: New York, 398,
1928.
(35) Marcus, Y. The Properties of Solvents; Wiley, New York, 1998.
(36) Treybal, R. E. Liquid Extraction, 2nd ed.; McGraw Hill: New York,
1963.
(37) Anderson, T. F.; Prausnitz, J. M. Application of the UNIQUAC
Equation to Calculation of Multicomponent Phase Equilibria. 1. Vapor−
Liquid Equilibria. Ind. Eng. Chem. Proc. Des. Dev. 1978, 17, 552−561.
(38) Renon, H.; Prausnitz, J. M. Local Compositions in Thermodynamic Excess Functions for Liquid Mixtures. AIChE J. 1968, 14, 135−
144.
(39) Prausnitz, J. M.; Eckert, C. A.; Orye, R. V.; O’Connell, J. P.
Computer Calculations for Multicomponent Vapor−Liquid and Liquid−
Liquid Equilibria; Prentice-Hall: Englewood Cliffs, NJ, 1980.
1564
dx.doi.org/10.1021/je4001284 | J. Chem. Eng. Data 2013, 58, 1549−1564