Fluid Phase Equilibria 459 (2018) 44e50

Contents lists available at ScienceDirect

Fluid Phase Equilibria

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / fl u i d

Measurement and correlation of isobaric vapoureliquid equilibrium

for systems of o-cresol, m-cresol and 2, 6-dimethylphenol at 20.0 kPa
Jun Wang, Jie Ou, Xueni Sun, Chunxiang Huang, Jiarui Sun, Hui Shao, Yixin Leng*
Changzhou University School of Petrochemical Engineering, Changzhou City, Jiangsu Province 213143, China

a r t i c l e i n f o a b s t r a c t

Article history: Isobaric vapoureliquid equilibrium (VLE) data for binary systems containing o-cresol (OC) þ m-cresol
Received 23 October 2017 (MC), OC þ 2, 6-dimethylphenol (DMP), and MC þ DMP at 20.0 kPa were measured by using a modified
Received in revised form Othmer still. Its reliability was confirmed by measuring the saturated vapour pressure data for the
6 December 2017
components of MC and DMP. Herington area test was used to confirm the thermodynamic consistency
Accepted 7 December 2017
for these binary systems, and the results showed good thermodynamic consistency. The binary VLE
Available online 8 December 2017
experimental data were correlated and predicted by Wilson, NRTL, and UNIFAC models. The results
correlated by Wilson and NRTL models agree well with the experimental values. The results predicted by
Keywords:
O-cresol
UNIFAC are in accordance with experimental data with rational deviations for systems OC þ MC and
M-cresol OC þ DMP but show relatively large deviations between the predicted values and experimental data for
2, 6-dimethylphenol MC þ DMP system. By defining the group ¼COH in DMP and MC as two different groups, the experi-
Vapour-liquid equilibrium mental data were correlated, and new groups with binary interaction parameters were obtained. The
Group results calculated by UNIFAC with new defining groups are in accordance with the experimental data.
© 2017 Published by Elsevier B.V.

1. Introduction
O-cresol (OC), m-cresol (MC), and 2,6-dimethylphenol (DMP)
are important organic synthesis materials. OC is mainly used as
synthetic resin and as a raw material for the synthesis of pesticides,
plastic antioxidants, and polymerization inhibitors et al. [1] MC can
be used to manufacture antioxidants, pesticides, and vitamin E et al.
[2] DMP is used as a monomer to synthesize engineering plastic
PPO, the raw material for the synthesis of pesticide intermediates 2,
6-diphenylamine and antiarrhythmic drugs [3,4].
OC, MC, and DMP are mainly derived from chemical synthesis
and byproduct of coal tar fractionation. Separation or purification of
these phenols is important due to their wide application. In general,
vacuum distillation can be used to separate OC and MC/DMP due to
their relatively high boiling point difference(about 10.0 K), namely,
464.15 K for OC, 475.35 K for MC, and 474.15 K for DMP [5]. Sepa-
rating MC and DMP using ordinary distillation is difficult because
the normal boiling point of DMP is close to that of MC. However,
according to the comparison of their saturated vapour pressure
data, we found that the difference of boiling point between MC and
* Corresponding author.
E-mail address: lengyixin61@163.com (Y. Leng).
DMP gradually increases with pressure reduction (see details in the
next section). In this case, vacuum distillation could be used to
separate MC and DMP. Vapoureliquid equilibrium (VLE) data are
important on the design and optimization of distillation process.
According to literature and Dortmund databank, Fox et al. [6]
measured the Tex experimental data (bubble point) for the sys-
tem of OC þ MC at 101.0 kPa in 1917. Ye et al. [7] measured the
Texey experimental data of OC þ DMP at 10.0 and 25.0 kPa in 1998.
To our knowledge, the VLE data for DMP þ MC are limited and may
not meet the requirements of designing the separation process.
In this work, the reliability of the modified Othmer still was
confirmed by measuring the saturated pressure data for the com-
ponents of MC and DMP. Isobaric VLE data for the binary systems
OC þ MC, OC þ DMP, and DMP þ MC at 20.0 kPa were measured by
using the equilibrium apparatus. Herington area test was used to
confirm the thermodynamic consistency of these binary systems.
Wilson [8] and NRTL [9] models were used to correlate the VLE
data, and the interaction parameters involved in models were ob-
tained. UNIFAC [10] model was used to predict the VLE of these
three binary systems. The groups of ¼COH were distinguished as
two different groups in the system of DMP þ MC, and the group
binary interaction parameters were obtained by correlating the
binary VLE data.

https://doi.org/10.1016/j.fluid.2017.12.011
0378-3812/© 2017 Published by Elsevier B.V.
 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50 45

Table 1
Mass fraction purity of the chemical reagent used.

Material Source CAS RN Mass fraction purity Purification method Analysis method

O-Cresol Sinopharm Chemical Regent Co., Ltd. 95-48-7 0.990 none GC

M-Cresol Sinopharm Chemical Regent Co., Ltd. 108-39-4 0.990 none GC
2, 6-Dimethylphenol Aladdin Reagent Co., Ltd 576-26-1 0.990 none GC

2. Experimental
2.1. Materials
The chemicals used in this work were OC, MC, and DMP. The OC
(99.0% purity of mass fraction) and MC (99.0% purity of mass
fraction) were supplied by Sinopharm Chemical Regent Co., Ltd.,
and DMP (99.0% purity of mass fraction) was provided by Aladdin
Reagent Co., Ltd. These chemicals were stated pure from factory.
Detailed information on these chemicals is listed in Table 1. All
reagents were used without further purification in this work.
2.2. VLE measurements
A modified Othmer still (c.f., Fig. 1.) was used to measure the VLE
of systems consisting of OC þ MC, OC þ DMP, and MC þ DMP. In this
still, both liquid and condensed vapour phases were recirculated
successively to sufficiently mix the phases, thus ensuring that
equilibrium state could be achieved as early as possible. System
pressure and temperature were measured by using a U-shaped
differential manometer (filled with mercury) with uncertainty of
±0.0665 kPa and a high precision mercury thermometer with un-
certainty of ±0.05 K. In each VLE measurement, the mixed solution
was placed into the boiling chamber at first, and the system pres-
sure was fixed and held constant by using a vacuum pump (2XZ-4,
Shanghai Instrument, China). Heat was provided, and a suitable
power of heater was defined to achieve reasonable condensation
reflux rate (i.e., about 2e3 drops per second). When constant
temperature and pressure were maintained for more than 50 min,
the equilibrium state was assumed, and the samples can be taken
from the sampling ports at this time.
Vapour pressure data for the components of DMP and MC was
measured to determine the reliability of equilibrium apparatus
(Table 2). For the convenience of evaluation, the experimental data
of this work were compared with previously published experi-
mental results, and the compared results are shown in Fig. 2. In this
figure, the experimental data of vapour pressure are in good
agreement with those in literature [11e15]. Therefore, verification
showed that the equilibrium was reliable.
2.3. Sample analysis
The gas chromatograph (Agilent 7890B) equipped with a Flame
Ionization Detector and a HP-ULTRA 2
(25 m  0.320 mm  0.52 mm) chromatographic column was used
to analyze the vapour and liquid samples. The temperatures of

Fig. 1. Experimental device for VLE measurement. 1. heating rod; 2. liquid phase sampling port; 3. equilibrium chamber; 4. mercury thermometer; 5. glycerol; 6. vapour phase
sampling port; 7. condensing coil; 8. valve; 9. buffer bottle (erlenmeyer flask); 10. U-shaped differential manometer; 11. vacuum pump.
46 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50

Table 2
Saturated vapour pressure of DMP and MC measured by this work.a

DMP MC

T/K p/kPa T/K p/kPa

390.06 6.70 402.24 9.29

396.14 8.87 408.55 11.93
401.43 10.89 413.39 14.57
405.64 12.74 417.47 16.68
409.97 14.98 420.54 18.79
413.33 16.80 423.55 20.91
416.65 18.91 426.04 23.02
419.43 21.02 428.56 25.13
422.01 22.83 e e
436.55 33.02
438.90 35.64
a
Standard uncertainties u of temperature T, pressure p are u(T) ¼ 0.05 K and
u(p) ¼ 0.0665 kPa.

vaporization and the chromatographic column were adjusted to

Fig. 3. Experimental and correlated/predicted VLE data for the binary system of DMP
533.15 K and 383.15 K, respectively. The temperature of detection (1) þ MC (2) at 20.0 kPa.
was set as 533.15 K. To ensure the reliability of the analysis, each
sample was analyzed three times and the average value was ob-
tained unless the measuring difference among the three samples of pure component i at equilibrium temperature T, respectively; gi
was less than 0.5% in mole fraction. The calculated area fraction represents the activity coefficient of component i; ViL is the molar
data were converted to mole fraction. volume of component i in pure liquid phase; R is the ideal gas
constant. Since the system pressure is 20.0 kPa, the fugacity co-
3. Results and discussion efficients can be ignored in the vapour phase, i.e., 4i and 4si are equal
to 1.0. Moreover, this term exp[VLi (p-psi )/RT] is approximately to be
3.1. VLE data of binary systems 1.0 because of the low total pressure. As a result, Eq. (1) can be
simplified as,
The VLE data were obtained at the pressure of 20.0 kPa for three
binary systems (i.e., OC þ MC, OC þ DMP, and DMP þ MC). The yi p ¼ gi xi psi (2)
measured VLE data of these systems are shown and listed in
Figs. 3e5 and Tables 3e5. The activity coefficient gi of component i The vapour pressure for pure component i (psi )
at the equilib-
at equilibrium state can be expressed as follows: rium temperature can be calculated by Antoine equation (i.e., Eq.
"  (3)),
#
ViL p  psi
4i yi p ¼ gi xi psi 4si exp (1)
RT    B
ln psi kPa ¼ A þ (3)
T=K þ C
where yi, xi are the vapour and liquid phase molar fractions,
respectively, of pure component i; 4i and 4si are the fugacity co- in which the equation parameters can be obtained by fitting the
efficients of component i in mixture and pure vapour phases, experimental data available in Refs. [11e17]. This process was
respectively; p and psi are the system pressure and vapour pressure performed in Origin 8.0 using nonlinear least square to correlate

Fig. 2. Compared MC and DMP vapour pressure data measured by this work with Fig. 4. Experimental and correlated/predicted VLE data for the binary system of OC
those of literatures. (1) þ MC (2) at 20.0 kPa.
 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50 47

Table 4
Experimental VLE data for binary system of OC (1) þ MC (2) at 20.0 kPa.a

No T/K x1 x2 y1 y2 g1 g2
1 411.26 1.0000 0.0000 1.0000 0.0000 0.9966 e
2 413.35 0.7573 0.2427 0.8193 0.1807 1.0009 1.0275
3 414.25 0.6670 0.3330 0.7435 0.2565 0.9991 1.0278
4 415.26 0.5589 0.4411 0.6502 0.3498 1.0064 1.0192
5 416.35 0.4567 0.5433 0.5523 0.4477 1.0072 1.0172
6 417.25 0.3768 0.6232 0.4706 0.5294 1.0083 1.0145
7 417.95 0.3141 0.6859 0.4038 0.5962 1.0132 1.0119
8 418.86 0.2381 0.7619 0.3208 0.6792 1.0293 1.0040
9 419.45 0.1898 0.8102 0.2601 0.7399 1.0260 1.0068
10 419.96 0.1554 0.8446 0.2167 0.7833 1.0261 1.0038
11 422.28 0.0000 1.0000 0.0000 1.0000 e 0.9963
a
Standard uncertainties u of temperature T, composition x1, y1 and pressure p are
u(T) ¼ 0.05 K, u(x1) ¼ u(y1) ¼ 0.001 and u(p) ¼ 0.0665 kPa.

Herington analysis [18]. The equations for Herington analysis can

be expressed as:
Fig. 5. Experimental and correlated/predicted VLE data for the binary system of OC
(1) þ DMP (2) at 20.0 kPa.  
Z 1 
 g 
 ln 1
dx
 g2 
 0
the experimental data. The fitted parameters and results are listed D ¼ 100 Z 1   (4)
in Table 6 and Fig. 6. The results fitted by this work are in good  g1 
ln dx
agreement with the experimental data with maximum relative  g2
0
error of 0.057% for OC, 0.094% for MC, and 0.076% for DMP. Thus, the
Antoine constants of OC, MC, and DMP could be used to correlate  
Tmax  Tmin 
the VLE data for these binary systems. From Fig. 6, we can see that J ¼ 150 
 (5)
Tmin
the bubble point difference of MC and DMP is small at pressure
~101.0 kPa. However, the temperature difference increases with the where Tmax and Tmin are the highest and lowest temperature in the
decreasing system pressure, which indicates that relative volatility experiment, respectively. According to the Wisniak's analysis [19],
may be improved when pressure is reduced. For the system of the isobaric VLE data passed the consistency test when the checked
OC þ DMP, the bubble point difference increases with the result jDeJj was less than 10. We used Eqs. (2)e(5) to calculate D
increasing the system pressure, which suggests that relative vola- and J, and the results are listed in Table 7. In Table 7, jDeJj values are
tility may be decreased by reducing the pressure in this system. 3.29 for DMP (1) þ MC (2) system, 1.84 for OC (1) þ DMP (2) system,
We used Eqs. (2) and (3) to calculate activity coefficients, and the and 2.73 for OC (1) þ MC (2) system. These results suggested that
results are given in Tables 3e5. Table 3 shows that the activity co- the experimental data for these binary systems at 20.0 kPa passed
efficients in DMP (1) þ MC (2) system are slightly larger than 1.0, the thermodynamic consistency.
which indicates that this system shows a positive deviation from
Raoult's law. Tables 4 and 5 show that the activity coefficients in OC
(1) þ DMP (2) and OC (1) þ MC (2) mixtures are both approximately
equal to 1.0, which indicates that these two systems are considered 3.3. Correlation and prediction for binary VLE data
ideal.
The isobaric experimental VLE values for three binary systems
DMP þ MC, OC þ MC, and OC þ DMP at 20.0 kPa were correlated by
3.2. Thermodynamic consistency test the Wilson and NRTL models using the maximum likelihood
method. The objective function for binary system is expressed as
Thermodynamic consistency test was conducted using follows [20]:

Table 3 Table 5
Experimental VLE data for binary system of DMP (1) þ MC (2) at 20.0 kPa.a Experimental VLE data for binary system of OC (1) þ DMP (2) at 20.0 kPa.a

No T/K x1 x2 y1 y2 g1 g2 No T/K x1 x2 y1 y2 g1 g2
1 418.11 1.0000 0.0000 1.0000 0.0000 0.9994 e 1 411.26 1.0000 0.0000 1.0000 0.0000 0.9966 e
2 418.26 0.9291 0.0709 0.9343 0.0657 0.9998 1.0667 2 412.07 0.8251 0.1749 0.8538 0.1462 1.0019 1.0312
3 418.36 0.8893 0.1107 0.8973 0.1027 0.9998 1.0640 3 412.56 0.7481 0.2519 0.7889 0.2111 1.0033 1.0160
4 418.80 0.7237 0.2763 0.7418 0.2582 1.0005 1.0548 4 413.33 0.6219 0.3781 0.6775 0.3225 1.0086 1.0063
5 419.15 0.6089 0.3911 0.6331 0.3669 1.0029 1.0455 5 414.00 0.5219 0.4781 0.5842 0.4158 1.0121 1.0022
6 419.56 0.4575 0.5425 0.4914 0.5086 1.0217 1.0295 6 414.60 0.4341 0.5659 0.4985 0.5015 1.0166 0.9999
7 420.05 0.3424 0.6576 0.3751 0.6249 1.0250 1.0252 7 415.18 0.3559 0.6441 0.4173 0.5827 1.0172 1.0003
8 420.45 0.2501 0.7499 0.2848 0.7152 1.0511 1.0143 8 415.93 0.2648 0.7352 0.3195 0.6805 1.0197 0.9971
9 421.00 0.1521 0.8479 0.1826 0.8174 1.0879 1.0052 9 416.49 0.1853 0.8147 0.2286 0.7714 1.0225 1.0005
10 421.55 0.0754 0.9246 0.0946 0.9054 1.1158 1.0012 10 417.12 0.1129 0.8871 0.1426 0.8574 1.0243 0.9993
11 422.16 0.0000 1.0000 0.0000 1.0000 e 1.0005 11 418.11 0.0000 1.0000 0.0000 1.0000 e 0.9994
a a
Standard uncertainties u of temperature T, composition x1, y1 and pressure p are Standard uncertainties u of temperature T, composition x1, y1 and pressure p are
u(T) ¼ 0.05 K, u(x1) ¼ u(y1) ¼ 0.001 and u(p) ¼ 0.0665 kPa. u(T) ¼ 0.05 K, u(x1) ¼ u(y1) ¼ 0.001 and u(p) ¼ 0.0665 kPa.
48 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50

Table 6
Fitted Antoine constants for OC, MC, and DMP.

Components A B C RMSDa damax Tmin ~ Tmax(K)

OC 14.11812 3449.13646 101.05827 0.132 0.057 380.98 ~ 464.15

MC 14.49411 3723.17365 98.37434 0.187 0.094 382.58 ~ 475.46
DMP 13.89112 3471.04716 99.51255 0.280 0.076 390.06 ~ 476.68
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a PN cal: exp: 2
RMSDðTÞ ¼ i¼1 ðT i T Þ =N .
! i
jTical Tiexp j
dðTÞmax ¼ max T exp
 100%.
i

" These phenomena occur because the isomers cannot be distin-

N 
X   
Ti;c  Ti;e 2 pi;c  pi;e 2
OF ¼ þ þ
i¼1
sT sp
 2 #
yi;c  yi;e
þ (6)
sy
The optimization process was performed in the ASPEN Plus 7.3
[21], and the optimized interaction parameters involved in these
two models were obtained and are listed in Table 8. The two activity
coefficient models can be used to calculate the bubble point tem-
perature and vapour phase composition by using the previously
obtained binary interaction energy parameters. The calculated
Texey diagrams of these binary systems at 20.0 kPa are shown in
Figs. 3e5, and the root mean square deviation (RMSD), the average
absolute deviations (AAD) of the equilibrium temperature, and
vapour phase mole fraction of the three systems were obtained and
are listed in Table 8. Figs. 3e5 and the compared findings of AAD
and RMSD showed that the results correlated by the Wilson and
NRTL models agree well with the experimental values with minor
deviations. This result indicated that both models could satisfac-
torily describe the VLE behavior of these systems at 20.0 kPa. As
shown in Figs. 3e5, no azeotropic behavior was found in these
binary systems.
Meanwhile, we used the UNIFAC model to predict the temper-
ature and vapour phase composition of bubble point at 20.0 kPa.
The groups and group interaction parameters involved in the
UNIFAC model are listed in Table 9 [21]. The predicted results are
shown and listed in Figs. 3e5 and Table 8. The results predicted by
UNIFAC are in accordance with the experimental data with rational
deviations for OC (1) þ MC (2) and OC (1) þ DMP (2) systems.
However, for DMP (1) þ MC (2) system, relatively large deviations
were found between the predicted values and experimental data.
 2
xi;c  xi;e guished due to the shortage of the UNIFAC model [22e24].
sx Figs. 3e5 show that the activity coefficients of these systems pre-
dicted by the UNIFAC model are all close to/equal to 1.0, which
indicates that these systems are all processed into ideal systems by
the UNIFAC model. This treatment may be rational for the systems
of OC þ MC and OC þ DMP. However, for the system of DMP þ MC,
this treatment may generate large deviations (c.f. Table 8) because
the hydroxyl in the DMP is largely different from that in the MC due
to the effects of the two methyl groups in DMP (steric hindrance
and electron-donating effects). This difference would affect the
interaction of group ¼COH with other groups. Therefore, we treated
the group ¼COH in DMP and MC as two different groups (i.e.,
COHDMP and COHMC) as listed in Table 10. For convenience, the
interactions between COHDMP and COHMC were only considered in
this work because the groups COHDMP and COHMC could form a
hydrogen bond, which could significantly affect the thermody-
namic behavior of this system. The group interaction parameters of
COHMC and COHDMP were correlated by means of MATLAB 2009,
and the fitted parameters are listed in Table 10. The correlated re-
sults of VLE data for the system of DMP þ MC are shown in Fig. 3
and Table 8, which suggest that the results calculated by the UNI-
FAC model with new defining group are in accordance with the
experimental data with rational deviations, and the calculated ac-
curacy of the modified UNIFAC model is apparently improved
compared with that of the original UNIFAC model. This information
can be useful in screening rational solvents when separating the
mixture of DMP and MC through extractive distillation or adductive
crystallization.
The mean relative volatility (am), one of the key points for
distillation, of these binary systems at different pressure values has
been calculated using the following equations:
y1i ð1  x1i Þ
ai ¼ (7)
ð1  y1i Þx1i

Z1
am ¼ ai dx (8)
0

where x1i, y1i were calculated by the Wilson model with the help of
binary parameters listed in Table 8. The calculated results of am are
shown in Fig. 7. For the systems DMP þ MC and OC þ MC, am in-
creases with decreasing pressure, especially when the total pres-
sure is less than ~22.5 kPa. This result can be explained by the
higher high saturated vapour pressure ratio at low temperatures.

Table 7
Thermodynamic consistency test for experimental data.

System p/kPa DTmax/K Tmin/K D/% J/% rD-Jr

DMP þ MC 20.0 4.05 418.11 4.74 1.45 3.29

OC þ DMP 20.0 6.85 411.26 0.66 2.50 1.84
Fig. 6. Compared OC, MC and DMP saturated vapour pressure data correlated by this
OC þ MC 20.0 11.02 411.26 6.75 4.02 2.73
work with those of experiment.
 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50 49

Table 8
The quaternion interaction parameters, the root mean square deviation (RMSD) and the average absolute deviations (AAD) for the equilibrium temperature (T) and the vapour
phase composition (y1) of the NRTL, Wilson and UNIFAC activity coefficient models.e

Model p Quaternion interaction parameters RMSD AAD

DMP (1) þ MC(2) kPa bij/K bji/K aij y1a T/Kb d(y1)c d(T)d
Wilson 20.0 307.5700 190.7245 e 0.0012 0.021 0.0009 0.016
NRTL 20.0 267.7275 387.0912 0.3 0.0013 0.020 0.0009 0.015
UNIFAC 20.0 e e e 0.0067 0.457 0.0055 0.378
UNIFAC(modified) 20.0 e e e 0.0026 0.071 0.0019 0.061
OC(1) þ MC(2)
Wilson 20.0 237.8134 169.1448 e 0.0015 0.053 0.0013 0.037
NRTL 20.0 226.8805 294.6405 0.3 0.0015 0.053 0.0014 0.037
UNIFAC 20.0 e e e 0.0033 0.271 0.0026 0.255
OC(1) þ DMP(2)
Wlison 20.0 164.1672 223.8732 e 0.0008 0.035 0.0007 0.025
NRTL 20.0 284.1971 223.3437 0.3 0.0007 0.035 0.0007 0.025
UNIFAC 20.0 e e e 0.0032 0.159 0.0027 0.146
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
a PN cal:  yexp: Þ2 =N .
RMSDðyÞ ¼ i¼1 ðyi i
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
b PN cal: exp: 2
RMSDðTÞ ¼ i¼1 ðT i  T i Þ =N .

c PN  cal 
exp 
dðyÞ ¼ 1 ðyi  yi Þ=N.
PN  cal 

d
dðTÞ ¼ 1 ðTi  Tiexp Þ=N.
e
Energy interaction parameter for Wilson is lnLij ¼ bij/T, for NRTL is Dgij/RT ¼ bij/T.

Table 9
Original groups and binary interaction parameters for the system of DMP þ MC.a

Group i/j ¼CH ¼CCH3 ¼COH

¼CH 0 167.0 1329

¼CCH3 146.8 0 884.9
¼COH 25.34 244.2 0
a
Group interaction parameter for UNIFAC is lntij ¼ bij/T, and the unit for the
parameter bij is K.

Table 10
Modified groups and binary interaction parameters for the system of DMP þ MC.b

Group i/j ¼CH ¼CCH3 COHMC COHDMP

¼CH 0 167.0 1329 1329

¼CCH3 146.8 0 884.9 884.9
COHMC 25.34 244.2 0 100.0a
COHDMP 25.34 244.2 12.3a 0
a
Values fitted by this work.
b
Group interaction parameter for UNIFAC is lntij ¼ bij/T, and the unit for the
parameter bij is K. Fig. 7. The effect of pressure on mean relative volatility of systems OC þ MC,
OC þ DMP and DMP þ MC.

However, for the system OC þ DMP, Fig. 7 shows that the am slightly
decreases with decreasing pressure. This phenomenon occurred were predicted by using the UNIFAC model, and the predicted re-
because the saturated vapour pressure difference decreases with sults are in accordance with the experimental data with rational
decreasing temperatures, thus inducing low pure component deviations except for system DMP þ MC. However, the UNIFAC
vapour pressure ratio at low temperatures (Fig. 6). model obtained more accurate results with the modified groups
than with the original groups. The main reason for this improve-
4. Conclusions ment is the groups ¼COH in molecules of DMP and MC were
differentiated.
In this work, the isobaric vapour-liquid equilibrium data for
binary systems DMP þ MC, OC þ MC, and OC þ DMP at 20.0 kPa
Acknowledgments
were measured by using a modified Othmer still. The binary VLE
values passed the consistency test of thermodynamics. The values
The authors are grateful to the Nature Science Foundation of
of activity coefficient indicated that the systems of OC þ MC and
Jiangsu Province for financial support under Project BK 20150262.
OC þ DMP could be considered as ideal, and system DMP þ MC was
a positive deviation system at 20.0 kPa. Moreover, the experimental
binary VLE data were correlated by using the Wilson and NRTL References
models. The interaction parameters of these models were obtained.
[1] Ullmann's Encyclopedia of Industrial Chemistry A8, 1987, pp. 25e29.
The results correlated by the Wilson and NRTL models are in good [2] K.Y. Wang, Y.P. Xu, Z.M. Liu, US 9212116 B2, 2015.
agreement with the experimental values. These binary VLE data [3] A.S. Hay, Prog. Polym. Sci. 24 (1999) 45e80.
50 J. Wang et al. / Fluid Phase Equilibria 459 (2018) 44e50
_
[4] W. Zukowski, G. Berkowicz, J. Baron, S. Kandefer, D. Jamanek, S. Szarlik,
Z. Wielgosz, M. Zielecka, Chem. Cent. J. 8 (2014) 51.
[5] L.S. Lee, C.W. Lin, C.H. Kao, Ind. Eng. Chem. Res. 39 (2000) 2068e2075.
[6] J.J. Fox, M.F. Barker, J. Soc. Chem. Ind. Lond. 36 (1917) 842.
[7] R.Q. Ye, M.S. Xia, D.H. Yu, H.L. Liu, Y. Hu, Chin. J. Chem. Eng. 2 (1998) 183e188.
[8] G.M. Wilson, J. Am. Chem. Soc. 86 (1964) 127e130.
[9] H. Renon, J.M. Prausnitz, AIChE J. 14 (1968) 135e144.
[10] A. Fredenslund, R.L. Jones, J.M. Prausnitz, AIChE J. 21 (1975) 1086e1099.
[11] J.L. Cabezas, S. Beltran, J. Coca, J. Chem. Eng. Data 35 (1990) 389.
[12] K.B. Goldblum, R.W. Martin, R.B. Young, Ind. Eng. Chem. 39 (1947) 1474.
[13] E. Cepeda, C. Gonzalez, J.M. Resa, J. Chem. Eng. Data 34 (1989) 270.
[14] G.W.A.Z. Kahlbaum, Phys. Chem. 26 (1898) 577e658.
[15] R.J.L. Andon, D.P. Biddiscombe, J.D. Cox, R. Handley, D. Harrop,
E.F.G. Herington, J.F. Martin, J. Chem. Soc. (1960) 5246e5254.
[16] S.M. Ashraf, T.E.V. Prasad, M. Ramakrishna, D.H.L. Prasad, Fluid Phase Equil.
129 (1997) 165e174.
[17] T.E.V. Prasad, G. Harish, N. Krupavaram, S. Prasad, A. Jaiswal, D.H.L. Prasad,
Fluid Phase Equil. 238 (2005) 45e51.
[18] E.F.G. Herington, J. Chem. Eng. Data 37 (1951) 457e470.
rez-Correa, A. Mejía, H. Segura, J. Chem. Eng. Data 58 (2013)
[19] J. Wisniak, J.R. Pe
3563e3566.
[20] L.Z. Zhang, D.M. Xu, J. Gao, L.W. Zhao, Z.S. Zhang, C.L. Li, J. Chem. Thermodyn.
102 (2016) 155e163.
[21] Version 7.3, Aspen Plus Software, Aspen Technology Inc, Burlington, MA,
2001.
[22] A. Fredenslund, Fluid Phase Equil. 52 (1989) 135e149.
[23] R.S. Wu, L.L. Lee, Fluid Phase Equil. 61 (1990) 17e51.
[24] C. Marcolli, T. Peter, Atmos. Chem. Phys. 5 (2005) 1545e1555.