Accepted Manuscript

Liquid–liquid equilibria for the extraction of furfural from aqueous solution using
different solvents

Kun Xin, Yuhe Song, Fangfang Dai, Yingmin Yu, Qingsong Li

PII: S0378-3812(16)30316-8
DOI: 10.1016/j.fluid.2016.06.040
Reference: FLUID 11158

To appear in: Fluid Phase Equilibria

Received Date: 12 May 2016

Revised Date: 20 June 2016
Accepted Date: 20 June 2016

Please cite this article as: K. Xin, Y. Song, F. Dai, Y. Yu, Q. Li, Liquid–liquid equilibria for the extraction
of furfural from aqueous solution using different solvents, Fluid Phase Equilibria (2016), doi: 10.1016/
j.fluid.2016.06.040.

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
 ACCEPTED MANUSCRIPT

Submitted to “Fluid Phase Equilibria”

PT
Liquid–liquid equilibria for the extraction of furfural from

RI
aqueous solution using different solvents
Kun Xin a, Yuhe Song b, Fangfang Dai a, Yingmin Yua, Qingsong Li a*

SC
a
The State Key Lab of Heavy Oil Processing, College of Chemical Engineering,

U
China University of Petroleum – East China,
AN
Qingdao, Shandong 266580, China

b
M

College of Pharmacy and Chemistry & Chemical Engineering, Taizhou

University, Taizhou, Jiangsu 225300, China

D
TE
C EP
AC

* Corresponding author

E-mail: licup01@163.com

1
 ACCEPTED MANUSCRIPT

Abstract
Liquid-liquid equilibria data are reported for the furfural-water system with

p-xylene and also toluene as solvents at 298.15 and 323.15 K. Other solvents

including methyl iso-butyl carbinol (MIBC) at 298.15 K, benzyl alcohol and

PT
acetophenone at 323.15 K have been investigated as well. The distribution coefficient

RI
and selectivity were calculated according to the LLE data. Both the NRTL and the

UNIQUAC models were successfully applied to fit the data for these five ternary

SC
systems. All the root mean square deviation (RMSD) values of the two models were

U
less 0.01.
AN
Key words: Liquid-liquid equilibria; furfural; water; p-xylene; toluene; methyl
M

iso-butyl carbinol; benzyl alcohol; acetophenone

D
TE
C EP
AC

2
 ACCEPTED MANUSCRIPT

List of symbols

g, u interaction parameters

D the distribution coefficient

S the separation factor

PT
M number of tie lines

RI
OF objective function

RMSD root mean square deviation

SC
r volume parameter in UNIQUAC equation

U
q area parameter in UNIQUAC equation or randomly varying quantity
AN
described by a probability distribution

q arithmetic mean of n independent repeated observations q  of

M

randomly-varying quantity q
D

q kth independent repeated observation of randomly-varying quantity q

TE

sq experimental standard deviation of the mean q, equal to the positive square

root of s  q
EP

sq   experimental standard deviation, equal to the positive square root of s  q  

C

i)
s(X experimental standard deviation of input mean Xi , equal to the positive
AC


square root of s  X

u(xi) standard uncertainty of input estimate xi that estimates input quantity Xi

xi mole fraction composition of liquid phase or estimate input quantity Xi in

u(xi)

Xi i th input quantity on which the measurand depends

3
 ACCEPTED MANUSCRIPT

Xi estimate of the value of input quantity Xi

Greek letter

α non-randomness factor in NRTL equation

Superscripts

PT
calculated

RI
Subscript

i, j component identification: 1 pertains to water, 2 to furfural, and 3 to the organic

SC
solvent

U
α, β component identification: α for the aqueous phase, β for the organic phase
AN
k number of the tie line
M
D
TE
C EP
AC

4
 ACCEPTED MANUSCRIPT

Introduction

Furfural is a highly versatile and key derivative used in the manufacture of a wide

range of important chemicals where it is used as a solvent for extractive refining of

lubricating oils, as a raw material for pharmaceuticals and phenolic resins, as well as

PT
an intermediate in the manufacture of lubricants, nylon, adhesives, plastics and

RI
solvents [1–3].

Furfural can be produced by digestion of cellulose wastes with steam and sulfuric

SC
acid in the range of temperatures between 413.15 and 458.15 K, pressure between 3.5

U
and 10 atm, forming a dilute aqueous solution [4]. Solubility of furfural in water is
AN
limited to less than 8.3 wt % at 293.15 K. Separation of furfural from this solution can

be performed by azeotropic distillation, supercritical CO2 extraction, adsorption,

M

organic solvent extraction, and membrane separation. However, separation of furfural

D

from aqueous mixtures by azeotropic distillation is of large investment and high

TE

energy consumption [5]. The supercritical CO2 extraction technology is still in the

experimental stage due to its high equipment investment and operating cost [6].
EP

Besides, the adsorption separation and the membrane separation technology are
C

limited to a certain concentration of the solution and difficult to realize mass

AC

production [7–8]. In this way, liquid−liquid extraction of furfural appears as the most

convenient choice for furfural upgrading at the moment. To facilitate development of

furfural extraction units, many different solvents including chlorinated hydrocarbons,

various aliphatic alcohols and aliphatic ketones have been studied for their ternary

liquid−liquid equilibria with water and furfural [9–10].

5
 ACCEPTED MANUSCRIPT

As a result of the limited solubility of furfural in water, phase splitting occurs

easily at high content of furfural in aqueous solution. In this work, low concentration

of furfural in aqueous solution is studied which is practical and meaningful.

Experimental liquid–liquid equilibrium data were determined for five systems,

PT
including water + furfural + p-xylene and also toluene at 298.15 and 323.15 K , water

RI
+ furfural + methyl iso-butyl carbinol (MIBC) at 298.15 K, water + furfural + benzyl

alcohol and acetophenone at 323.15 K as well. Considering the similar density of

SC
benzyl alcohol and acetophenone with water, a higher temperature of 323.15 K was

U
used to speed up the molecular motion which greatly shortened the time to reach the
AN
equilibrium. To the best of our knowledge, all the LLE data determined in this paper

have not been reported up to now. Tie-lines were determined for the ternary systems
M

according to the LLE data. The distribution coefficient and separation factor were
D

calculated and were used as the standard to evaluate the separation efficiency. The
TE

experimental data were also correlated with Non-Random Two Liquids (NRTL) [11]

and Universal Quasi-Chemical (UNIQUAC) [12] activity coefficient models.

EP

2. Experimental
C

2.1 Materials
AC

Materials such as furfural, p-xylene, toluene, methyl iso-butyl carbinol, benzyl

alcohol and acetophenone were purchased from Sinopharm Chemical Reagent.

Double distilled water was employed throughout. All the chemicals were used without

further purifications. The major information of the used chemicals were displayed in

Table 1.

6
 ACCEPTED MANUSCRIPT

2.2 Apparatus and procedure

LLE data for the studied ternary system were obtained at desired temperature

under atmospheric pressure. The details about experimental equipment have been

presented in our previous work and the reliability of the experimental system has been

PT
evaluated [13]. The agitation and settling time were investigated, demonstrating that 3

RI
h was enough for furfural in water and solvents to reach liquid–liquid equilibrium.

Therefore, this mixture was agitated vigorously for at least 1 h and left for at least 4 h

SC
before GC measurements. Moreover, the air in the still was replaced by N2 in order to

U
prevent the oxidation of furfural. The evaporated compounds were completely
AN
condensed by the condenser to ensure the mass balance.

In each measurement, the samples were carefully withdrawn by syringes from the
M

organic phase and the aqueous phase, respectively. The internal standard method was
D

used to obtain quantitative results in the analysis, and isopropanol was chosen as the
TE

internal standard substance. The Agilent GC6820 gas chromatograph equipped with a

thermal conductivity detector and Porapak N column (3 mm × 3 m) was used to

EP

analyze the composition of both phases. Hydrogen was used as the carrier gas at the
C

rate of 60 mL/min. The temperatures of the injector and the detector were maintained
AC

at 523.15 K. After 1 min of holding at 373.15 K, the column temperature was raised to

376.15 K at a rate of 0.5 K/min, afterwards raised to 523.15 K and maintained at this

temperature for 1.5 min. Each sample was analyzed at least three times and the

average values were used as the sample composition. Besides, the standard

uncertainty of the liquid phase compositions was calculated according to GUM

7
 ACCEPTED MANUSCRIPT

standard [14]. A series of LLE data were obtained by changing the feed proportion at

desired temperatures.

2.3 Uncertainty calculation

There are two categories for calculating the uncertainty, named as “A” and “B”

PT
[14]. In this work, Type A was used to evaluate the uncertainty of each quantity q. The

RI
relevant calculation equations are listed as:

s  q   = ∑ q − q⁄n − 1 (1)

SC
s  q = s  q  ⁄n (2)

i)

U
u(xi)=s(X (3)
AN
where the experimental standard deviation (s(qk)) is calculated, describing the

variability of the observed values qk or more specifically, their dispersion about their
M

mean q. Thus, for an input quantity Xi, the Type A standard uncertainty u(xi) of its
D

i is u(xi)=s(X
estimate xi=X i).
TE

2.4 Separation efficiency calculation

The separation efficiency of solvents was evaluated by the distribution coefficient

EP

(D) and separation factor (S). They were calculated from the tie-line data, and defined
C

as the following equations:

AC


D =  (4)


 ⁄
S= (5)
 ⁄

Where  and  are the mole fractions of solute and diluent in the dilute phase,

respectively.  and  are the mole fractions of solute and diluent in the solvent

phase, respectively.

8
 ACCEPTED MANUSCRIPT

3. Thermodynamics modeling

The NRTL and UNIQUAC models were applied to correlate the experimental

mole fractions by using the Aspen Plus software (version 7.3).

The NRTL model is based on the local composition concept and there are three

PT
adjustable parameters g
, g
, and α
in it. The recommended α
values for different

RI
types of mixtures are commonly between 0.1 and 0.47.

A combinatorial term related to differences in size and shape between the

SC
components and a residual term accounting for energy differences between the

U
molecules contribute to the activity coefficient and the excess Gibbs energy in
AN
UNIQUAC model [12]. To fit the UNIQUAC interaction parameters, the structural

parameters (r and q) of pure components are needed. The r and q values are
M

recommended by DECHEMA [15] and listed in Table 2.

D

The interaction parameters between components are essential when the NRTL
TE

and the UNIQUAC models are applied. They can be determined by minimizing the

objective function (OF) of the composition. The objective function (OF) used is:
EP


OF = ∑+
' ∑& ∑%$%&' − 
 *
(%&' ) (6)
C

where M is the number of tie lines, x is the experimental mole fraction, and 
, is the
AC

calculated mole fraction. The subscripts i, j and k refer to the components, the phases

and the tie line, respectively. The experimental data were correlated at each

temperature respectively. The optimization results were judged by calculating the

corresponding percent-root-mean-square deviation (RMSD) values using the

following equation:

9
 ACCEPTED MANUSCRIPT

 /
* $234 5x
, 234 )
' ∑& ∑%

RMSD = 100 ∗ 1∑+ 8 (7)
7+

the corresponding M, x, x
,, i, j, k are the same with those in the OF equations.

4. Results and discussion

4.1 Experimental data

PT
The LLE data for the ternary systems (Water + Furfural + Solvents) at desired

RI
temperatures are shown in Table 3 and 4 with all concentrations expressed in mole

SC
fraction. The phase diagrams for the ternary mixtures are presented in Figs. 1–7. The

tie-lines and feed composition are also plotted in Figs. 1–7. As shown in these figures,

U
the points of the feed compositions agree the tie line with great accuracy, which is in
AN
accordance with the lever rule. The result indicates that the mass balance is satisfied
M

during the experimental operations, demonstrating the reliability of experimental data

[16].
D

The system studied in this paper is formed by homogeneous binary mixtures of

TE

solvents with furfural, partly miscible binary mixtures of furfural with water and a
EP

system with large miscibility gaps: solvent + water. The mutual solubility of solvents

especially p-xylene and toluene with water is very small. The ternary systems have
C
AC

the wide heterogeneous region between two branches of the bimodal, each of the

branches going from the water + solvent side of the concentration triangle to the

furfural + solvent side. Near the apex of pure water the homogeneous is narrow and

this area expands on furfural addition.

The values of D and S calculated from the experimental data are given in Table 3

and 4. As presented in these tables, it can be observed that generally the distribution
10
 ACCEPTED MANUSCRIPT

coefficients tend to increase when the concentration of furfural increases.

Simultaneously, the tie line slope factor reflects the same tendency. It is known that

the distribution coefficient is related with the solubility of solute which has an effect

on the tie line slope [17, 18]. As presented in Table 3 and 4, the separation factors in

PT
all cases are much more than one, revealing that the extraction process is possible.

RI
Meanwhile, generally the separation factor decrease as the concentration of furfural

increases. That’s because the two-phase region shrinks when the concentration of

SC
furfural increases, demonstrating that the separation capacity of the solvent is reduced.

U
Otherwise, the variation of separation factors occurred with the concentration of
AN
furfural increasing when MIBC or benzyl alcohol was used as solvent. That may be

caused by the relative high mutual solubility of alcohol and water. The experimental
M

separation factors versus furfural mole fraction in the solvent phase at desired
D

temperatures are demonstrated in Fig. 8. Moreover, at the same concentration of

TE

furfural, highest value of D is achieved when acetophenone is used as extractant,

That’s related with the appropriate polarity of acetophenone in comparison of furfural.

EP

But considering the low mutual solubility of aromatic hydrocarbon with water, for
C

water (1) + furfural (2) + p-xylene or toluene (3), the greater size of the two-phase
AC

region, greater separation factors and high furfural distribution coefficient indicate

that p-xylene and toluene present technical advantages to remove furfural from water.

In view of the poisonousness of toluene, p-xylene is an ideal extractant for separating

furfural which is first reported to the best of our knowledge.

4.2 Data Correlation

11
 ACCEPTED MANUSCRIPT

The data correlations for Water + Furfural +Solvents were carried out by using

both NRTL and UNIQUAC models. The results of data regression are listed in Table 5.

Generally, all the RMSD values were below 1.0 %, the calculated LLE data from the

NRTL and UNIQUAC model are also plotted in Figs. 1–7, demonstrating that the two

PT
models can correlate well with the experimental data.

RI
4. Conclusions

The liquid-liquid equilibria for Water + Furfural + Solvents ternary system were

SC
investigated at desired temperatures under atmospheric pressure. It was found that

U
furfural have a much higher solubility in selected solvents than that in water. All the
AN
separation factors were much larger than one, implying the feasibility of selected

solvents to extract the furfural from aqueous solution. Generally the distribution
M

coefficients tended to increase when the concentration of furfural increased.

D

Meanwhile, the opposite trend was usually observed for the separation factor. With
TE

low mutual solubility with water, great separation factors, high furfural distribution

coefficient and green environmental protection, p-xylene was an ideal extractant for
EP

separating furfural. Meanwhile, both the NRTL and the UNIQUAC models were
C

successfully applied to correlate experimental data and the corresponding optimum

AC

interaction parameters of models were obtained.

Acknowledgements

This study is supported by Qingdao postdoctoral fund (NO. T1504117) and the

Project of Natural Science Research of Higher Education Institutions of Jiangsu

Province (NO. 15KJD530001).

12
 ACCEPTED MANUSCRIPT

References

[1] K.J. Zeitsch, The chemistry and technology of furfural and its many by-products,

Elsevier, The Netherlands, 2000.

[2] B. Coto, R. van Grieken, J.L. Peña, J.J. Espada, Chem. Eng. Sci. 61 (2006) 8028–

PT
8039.

RI
[3] J. Lecomte, A. Finiels, C. Moreau, Industrial Crops and Products 9 (1999) 235–

241.

SC
[4] H.D. Mansilla, J. Baeza, S. Urzúa, Bioresource Technology 66 (1998) 189–193.

U
[5] J.R. Croker, R.G. Bowrey, IND. ENG. CHEM. RES. 23 (1984) 480–484.
AN
[6] Y.C. Kim, H.S. Lee, J. IND. ENG. CHEM. 7 (2001) 424–429.

[7] K. Jeřábek, L. Hankova, Z. Prokop, Reactive polymers 23 (1994) 107–112.

M

[8] U.K. Ghosh, N.C. Pradhan, B. Adhikari, Desalination 208 (2007) 146–158.
D

[9] B.F. Almeida, T.M. Waldrigui, T.C. Alves, L.H. Oliveira, M. Aznar, Fluid Phase
TE

Equilib. 334 (2012) 97–105.

[10] M. Männistö, J.P. Pokki, A. Creati, A. Voisin, A. Zaitseva, J. Chem. Eng. Data 61
EP

(2016) 903–911.
C

[11] H. Renon, J.M. Prausnitz, AIChE J. 14 (1968) 135–144.

AC

[12] D.S. Abrams, J.M. Prausnitz, AIChE J. 21 (1975) 116–128.

[13] F.F. Dai, K. Xin, Y.H. Song, M.D. Shi, Y.M. Yu, Q.S. Li, Fluid Phase Equilib.

409 (2016) 466–471.

[14] BIPM, IEC, IFCC, ILAC, ISO, IUPAC, IUPAP and OIML, Evaluation of

measurement data- Guide to the expression of uncertainty in measurement, JCGM.

13
 ACCEPTED MANUSCRIPT

100(2008) 2008.

[15] J. Gmehling, U. Onken, Vapor–Liquid Equilibrium Data Collection, DECHEMA,

Chemistry Data Series, Frankfurt, Germany, 1977.

[16] Y.X. Li, Q. Xu, S.L. Liu, H.M. Li, F.B. Zhang, G.L. Zhang, Q. Xia, Fluid Phase

PT
Equilib. 412 (2016) 199–204.

RI
[17] J.S. Santos, A.P.L. Craig, J.M.O. Santana, A.F. Santos, M.F. Heredia, M.L.

Corazza, E. Franceschi, C. Dariva, J. Chem. Eng. Data 60 (2015) 2050–2056.

SC
[18] A. Cháfer, J. de la Torre, E. Lladosa, J. B. Montón, Fluid Phase Equilib. 361

U
(2014) 23–29.
AN
M
D
TE
C EP
AC

14
 ACCEPTED MANUSCRIPT

Table 1 Materials description

Reported GC GC Purity
Component CAS Source
purity (mass%) (mass%)
Double-distilled
7732-18-5 Self-made 99.91
water
Furfural 98-01-1 Sinopharm 99.0 99.46
p-Xylene 106-42-3 Sinopharm 99.0 99.53

PT
Toluene 108-88-3 Sinopharm 99.5 99.91
Methyl iso-butyl
108-11-2 Sinopharm 99.0 99.49
carbinol

RI
Benzyl alcohol 100-51-6 Sinopharm 99.0 99.57
Acetophenone 98-86-2 Sinopharm 99.5 99.85

SC
Table 2 UNIQUAC structural parameters (r and q)
Component r q
Deionized water

U0.9200 1.3997
AN
Furfural 3.1681 2.4840
p-Xylene 4.6579 3.5360
Toluene 3.9228 2.9700
M

Benzyl alcohol 4.4832 3.4864

acetophenone 4.6941 3.6080
Methyl iso-butyl
4.8016 4.1240
D

carbinol
TE
C EP
AC

15
 ACCEPTED MANUSCRIPT

Table 3 Experimental LLE Data (mole Fraction) for Water (1) + Furfural (2) +Solvents (3) system
at 298.15 K under atmosphere pressure a
Organic phase Aqueous phase
Solvent D S
x1 x2 x3 x1 x2 x3
0.0056 0.0107 0.9837 0.9994 0.0005 0.0001 21.40 3819
0.0060 0.0216 0.9724 0.9990 0.0009 0.0001 24.00 3996

PT
0.0067 0.0410 0.9523 0.9982 0.0016 0.0001 25.63 3818
0.0071 0.0616 0.9313 0.9975 0.0024 0.0001 25.67 3606
p-Xylene 0.0072 0.0839 0.9089 0.9967 0.0031 0.0001 27.06 3747

RI
0.0080 0.1052 0.8867 0.9961 0.0038 0.0001 27.68 3447
0.0079 0.1284 0.8637 0.9954 0.0044 0.0001 29.18 3677

SC
0.0084 0.1641 0.8275 0.9946 0.0053 0.0001 30.96 3666

0.0068 0.0093 0.9839 0.9995 0.0003 0.0002 31.00 4557

0.0069 0.0200 0.9730 0.9990 0.0007 0.0003 28.57 4137

U
0.0077 0.0413 0.9510 0.9983 0.0014 0.0003 29.50 3825
AN
0.0084 0.0642 0.9274 0.9976 0.0021 0.0003 30.57 3631
Toluene
0.0088 0.0876 0.9035 0.9970 0.0026 0.0003 33.69 3817
0.0092 0.1112 0.8795 0.9965 0.0032 0.0003 34.75 3764
M

0.0097 0.1340 0.8563 0.9959 0.0038 0.0004 35.26 3620

0.0099 0.1744 0.8158 0.9950 0.0046 0.0003 37.91 3810
D

0.2676 0.0080 0.7244 0.9964 0.0005 0.0031 16.00 59.58

0.2752 0.0163 0.7086 0.9957 0.0010 0.0033 16.30 58.97
TE

0.2786 0.0343 0.6870 0.9945 0.0021 0.0034 16.33 58.30

0.2814 0.0528 0.6658 0.9935 0.0031 0.0034 17.03 60.13
MIBC
0.2878 0.0712 0.6410 0.9923 0.0040 0.0037 17.80 61.37
EP

0.2957 0.0900 0.6144 0.9910 0.0050 0.0040 18.00 60.32

0.3011 0.1084 0.5906 0.9902 0.0058 0.0039 18.69 61.46
0.3041 0.1367 0.5592 0.9884 0.0070 0.0046 19.53 63.47
C

a
Standard uncertainties u are u(T)= 0.1K, u(p)=0.3 kPa and u(x)= 0.0007.
AC

16
 ACCEPTED MANUSCRIPT
Table 4 Experimental LLE Data (mole Fraction) for Water (1) + Furfural (2) +Solvents (3) system
at 323.15 K under atmosphere pressure a
Organic phase Aqueous phase
Solvent D S
x1 x2 x3 x1 x2 x3
0.0079 0.0070 0.9852 0.9995 0.0003 0.0002 23.33 2952
0.0110 0.0270 0.9620 0.9987 0.0011 0.0002 24.55 2229
0.0132 0.0577 0.9291 0.9976 0.0022 0.0002 26.23 1982
0.0141 0.0887 0.8972 0.9963 0.0033 0.0004 26.88 1899

PT
p-Xylene
0.0151 0.1190 0.8659 0.9954 0.0042 0.0004 28.33 1868
0.0163 0.1510 0.8327 0.9943 0.0051 0.0006 29.61 1806
0.0168 0.1828 0.8004 0.9937 0.0059 0.0004 30.98 1833

RI
0.0175 0.2157 0.7668 0.9930 0.0066 0.0004 32.68 1854

SC
0.0115 0.0094 0.9791 0.9992 0.0003 0.0005 31.33 2722
0.0133 0.0196 0.9671 0.9987 0.0007 0.0006 28.00 2103
0.0156 0.0413 0.9431 0.9981 0.0014 0.0006 29.50 1887
0.0142 0.0638 0.9220 0.9974 0.0020 0.0006 31.90 2241

U
Toluene
0.0148 0.0869 0.8984 0.9967 0.0027 0.0006 32.19 2168
AN
0.0150 0.1118 0.8732 0.9960 0.0033 0.0007 33.88 2250
0.0174 0.1352 0.8474 0.9955 0.0040 0.0005 33.80 1934
0.0170 0.1718 0.8112 0.9947 0.0048 0.0006 35.79 2094
M

0.3909 0.0080 0.6011 0.9763 0.0005 0.0232 16.00 39.96

0.3925 0.0155 0.5920 0.9745 0.0010 0.0245 15.50 38.48
D

0.3911 0.0227 0.5862 0.9740 0.0014 0.0246 16.21 40.38

Benzyl 0.3935 0.0377 0.5688 0.9736 0.0023 0.0241 16.39 40.56
TE

alcohol 0.3972 0.0526 0.5503 0.9736 0.0032 0.0232 16.44 40.29

0.3989 0.0672 0.5339 0.9730 0.0041 0.0230 16.39 39.98
0.3971 0.0899 0.5130 0.9732 0.0054 0.0214 16.65 40.80
EP

0.4082 0.1103 0.4815 0.9737 0.0065 0.0198 16.97 40.48

0.1204 0.0115 0.8680 0.9961 0.0002 0.0037 57.50 475.7

C

0.1249 0.0236 0.8516 0.9957 0.0005 0.0039 47.20 376.3

0.1267 0.0482 0.8250 0.9954 0.0009 0.0037 53.56 420.8
AC

0.1337 0.0717 0.7946 0.9950 0.0014 0.0036 51.21 381.1

Acetophenone
0.1352 0.0947 0.7702 0.9946 0.0019 0.0035 49.84 366.7
0.1405 0.1164 0.7431 0.9942 0.0023 0.0035 50.61 358.1
0.1492 0.1388 0.7121 0.9937 0.0028 0.0035 49.57 330.2
0.1525 0.1717 0.6758 0.9929 0.0035 0.0037 49.06 319.4
a
Standard uncertainties u are u(T)= 0.1K, u(p)=0.3 kPa and u(x)= 0.0009.

17
 ACCEPTED MANUSCRIPT
Table 5 Binary energy parameters of NRTL and UNIQUAC models for the system Water (1) +
Furfural (2) + Solvents (3)
NRTL parameters UNIQUAC parameters
T/K i-j RMSD RMSD
gij-gjja gji-giia α uij-ujja uji-uiia

1-2 -3500.59 25202.64 0.20 -8147.10 41751.58

p-Xylene
1-3 501066.11 32640.07 0.16 0.05 -8423.30 -2838.93 0.03

PT
298.15
2-3 517806.87 12480.44 0.20 -3186.75 -5547.97

1-2 9022.52 2486.27 0.20 -1218.33 259043.46

RI
p-Xylene
1-3 381812.17 7301.50 0.25 0.03 1346.29 7079.90 0.03
323.15
2-3 518086.91 2759.91 0.20 -2199.81 4116.92

SC
1-2 11430.68 21187.10 0.30 -10859.71 259043.46
Toluene
1-3 1457433.7 -545411.55 0.20 0.16 8471.70 16481.24 0.03
298.15K
2-3 -11655.72 16675.72 0.30 -4849.82 10706.07

U
AN
1-2 321687.72 16042.78 0.30 -990.46 87293.79
Toluene
1-3 1456370.6 -545394.97 0.20 0.86 1784.88 6115.06 0.04
323.15K
2-3 -2708.40 -1957.82 0.30 -1713.72 2883.43
M

1-2 6919.27 1863.91 0.20 145.61 1961.41

MIBC
1-3 13044.95 -1899.06 0.20 0.22 -82.97 1341.22 0.20
D

298.15
2-3 -3442.48 8394.29 0.20 -433.29 2917.81
TE

Benzyl 1-2 11819.61 -1735.29 0.20 713.85 1131.28

alcohol 1-3 11756.08 -1432.88 0.20 0.14 -1004.83 3878.33 0.17
323.15 2-3 446.66 790.02 0.20 -1293.57 2258.91
EP

Acetopheno 1-2 8925.42 1371.51 0.20 1258.24 2357.88

ne 1-3 13487.40 1632.82 0.20 0.07 -57.85 3987.26 0.08
C

323.15 2-3 -5489.17 7999.04 0.20 13640.27 -1685.88

AC

a
the unit of parameters is J/mol.

18
 ACCEPTED MANUSCRIPT
Figure Caption:
Fig.1 Ternary phase diagram for Water + Furfural + p-Xylene system at 298.15 K, (a)

Integrated ternary phase diagram (b) Zoomed in view of the phase points in p-Xylene

phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the phase points in

aqueous phase ( ) feed composition; () experimental data; (○) NRTL model; (△)

PT
UNIQUAC model.

RI
Fig.2 Ternary phase diagram for Water + Furfural + p-Xylene system at 323.15 K, (a)

SC
Integrated ternary phase diagram (b) Zoomed in view of the phase points in p-Xylene

phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the phase points in

U
aqueous phase ( ) feed composition; () experimental data; (○) NRTL model; (△)
AN
UNIQUAC model.
M

Fig.3 Ternary phase diagram for Water + Furfural + Toluene system at 298.15 K, (a)

Integrated ternary phase diagram (b) Zoomed in view of the phase points in toluene
D

phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the phase points in
TE

aqueous phase ( ) feed composition; () experimental data; (○) NRTL model; (△)
EP

UNIQUAC model.

Fig.4 Ternary phase diagram for Water + Furfural + Toluene system at 323.15 K, (a)
C
AC

Integrated ternary phase diagram (b) Zoomed in view of the phase points in toluene

phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the phase points in

aqueous phase ( ) feed composition; () experimental data; (○) NRTL model; (△)

UNIQUAC model.

Fig.5 Ternary phase diagram for Water + Furfural + MIBC system at 298.15 K, (a)

Integrated ternary phase diagram (b) Zoomed in view of the phase points in MIBC
19
 ACCEPTED MANUSCRIPT

phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the phase points in

aqueous phase ( ) feed composition; () experimental data; (○) NRTL model; (△)

UNIQUAC model.

Fig.6 Ternary phase diagram for Water + Furfural + Benzyl alcohol system at 323.15

PT
K, (a) Integrated ternary phase diagram (b) Zoomed in view of the phase points in

RI
benzyl alcohol phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the

phase points in aqueous phase ( ) feed composition; () experimental data; (○)

SC
NRTL model; (△) UNIQUAC model.

U
Fig.7 Ternary phase diagram for Water + Furfural + Acetophenone system at 323.15 K,
AN
(a) Integrated ternary phase diagram (b) Zoomed in view of the phase points in

acetophenone phase (c) Zoomed in view of the feed spots (d) Zoomed in view of the
M

phase points in aqueous phase ( ) feed composition; () experimental data; (○)
D

NRTL model; (△) UNIQUAC model.

TE

Fig.8 Experimental separation factor versus the furfural mole fraction ( ) in the

solvent phase for the ternary system Water (1) + Furfural (2) + Solvents (3).
C EP
AC

20
 ACCEPTED MANUSCRIPT

Figure 1
0.00
1.00 0.8
0.2
(a) (b)
0.25
0.75
ne

ne
Fu

Fu
yle

yle
rfu

rfu
0.50 0.9
p-X

p-X
ra

ra
0.50

PT
l

l
0.1

0.75
0.25

RI
1.00 1.0
0.00 0.0
0.00 0.25 0.50 0.75 1.00 0.0 0.1 0.2

SC
Water Water
0.000

(c) (d)

U
0.005
ne

Fu

Fu
AN
ne
yle

rfu

rfu
0.14
yle
p-X

ra

ra
p-X

0.02
l

l
M

0.005

0.16
D

0.00 0.000
0.84 0.86 0.995 1.000
Water Water
TE
C EP
AC

21
 ACCEPTED MANUSCRIPT

Figure 2
0.00
1.00

(a) 0.25
(b) 0.8
0.2
0.75
ne

ne
Fu

Fu
yle

yle
rfu

rfu
0.50
p-X

p-X
ra

ra
0.50

PT
l

l
0.9
0.1
0.75
0.25

RI
1.00 1.0
0.00 0.0
0.0 0.1 Water 0.2

SC
0.00 0.25 0.50 0.75 1.00
Water
0.08 0.04 0.000

(c) (d)

U 0.005
ne

ne
Fu

Fu
AN
yle

yle
rfu

rfu
p-X

p-X
ral

ral
0.10 0.02

0.005
M
D

0.00 0.000
0.90 0.92 0.995 1.000
Water Water
TE
C EP
AC

22
 ACCEPTED MANUSCRIPT

Figure 3
0.00
1.00
0.8
(a) (b) 0.2
0.25
0.75
ne

ne
Fu

Fu
lue

lue
rfu

rfu
0.50

ra

ra
To

To

PT
0.50 0.9

l
0.1

0.75
0.25

RI
1.00 1.0
0.00 0.0

SC
0.00 0.25 0.50 0.75 1.00 0.0 0.1 0.2
Water Water
0.000
0.12
0.005

(c) (d)

U
AN
ne

ne
Fu

Fu
lue

lue
rf

rf
ura

ura
0.02
To

To
l

l
0.14
M

0.005
D

0.00 0.000
0.86 0.88 0.995 1.000
Water Water
TE

(c) (d
C EP
AC

23
 ACCEPTED MANUSCRIPT

Figure 4
0.00
1.00
0.8
0.2
(a) 0.25 (b)
0.75
ne

ne
Fu

Fu
lue

lue
rfu

rfu
0.50

ral

ral
To

To
0.50 0.9

PT
0.1

0.75
0.25

RI
1.00 1.0
0.00 0.0
0.00 0.25 0.50 0.75 1.00 0.0 0.1 Water 0.2

SC
Water 0.000

0.12
0.04
(c) (d)

U
ne

ne
Fu

Fu
AN
0.004
lue

lue
rfu

rfu
ral

ral
To

To

0.14
0.02 0.004
M
D

0.00 0.000
0.86 0.88 0.996 1.000
Water Water
TE

(c) (d)
C EP
AC

24
 ACCEPTED MANUSCRIPT

Figure 5
0.00
1.00

(a) 0.25 (b) 0.2

0.75 0.6

Fu

Fu
BC

BC
rfu

rfu
0.50
MI

MI
ral

ral
0.50

PT
0.1

0.75 0.7
0.25

RI
1.00
0.00 0.0
0.00 0.25 0.50 0.75 1.00 0.3 0.4
Water

SC
Water

0.04
0.010
(c) 0.10 (d)

U
0.005
Fu

Fu
BC

BC
rfu

rfu
AN
MI

MI
ral

ral
0.02 0.005
0.12

0.010
M

0.00 0.000
0.88 0.90 0.990 0.995
D

Water Water
TE
C EP
AC

25
 ACCEPTED MANUSCRIPT

Figure 6

0.00
1.00
0.15
(a) 0.25 (b) 0.50
0.75
l

l
ho

ho
lco

lco
0.10

Fu

Fu
la

la
rfu

rfu
0.50
y

y
ral

ral
0.55

PT
nz

nz
0.50
Be

Be
0.75 0.05
0.25

RI
0.60

1.00
0.00 0.00
0.00 0.25 0.50 0.75 1.00 0.40 0.45 0.50

SC
Water Water
0.020
0.04
(c) (d)

U
0.008
0.12
ol

l
ho
oh

lco
Fu

Fu
alc

AN
0.024
yl a
rfu

rfu
yl

ral

ral
nz

nz

0.02
Be

Be

0.004
M

0.14
0.028

0.00 0.000
D

0.86 0.88 0.972 0.976 0.980

Water Water
TE
C EP
AC

26
 ACCEPTED MANUSCRIPT

Figure 7
0.00
1.00

0.7 0.2
(a) 0.25 (b)
0.75
ne

ne
no

no
Fu

Fu
he

he
rfu

rfu
0.50

PT
p

p
ral

ral
eto

eto
0.50
Ac

Ac
0.8 0.1

0.75

RI
0.25

1.00

SC
0.00 0.0
0.00 0.25 0.50 0.75 1.00 0.2 0.3
Water Water

0.004

U
(c) 0.10 0.04 (d) 0.004
AN
e

e
on

on
Fu

Fu
n

n
he

he
rfu

rfu
p

p
ral

ral
eto

eto
Ac

Ac

0.002
M

0.12 0.02 0.006

D
TE

0.00 0.000
0.88 0.90 0.994 0.996
Water Water
C EP
AC

27
 ACCEPTED MANUSCRIPT

Figure 8

5000 p-Xylene-298.15K
Toluene-298.15K
p-Xylene-323.15K
Toluene-323.15K
MIBC-298.15K
4000 Benzyl alcohol-323.15K
Acetophenone-323.15K

3000

PT
S

2000

RI
400
350

SC
300
60

40
U
AN
0.00 0.05 0.10 0.15 0.20 0.25
x2β
M
D
TE
C EP
AC

28