Received: 20 July 2020 Revised: 24 November 2020 Accepted: 26 November 2020

DOI: 10.1002/apj.2609

RESEARCH ARTICLE

Extractive distillation of benzene, toluene, and xylenes from

pyrolysis gasoline using methylsulfonylethane as a
cosolvent

Siliang Gao | Wencheng Tang | Ming Zhao | Siyuan Qie | Weiwei Pang |
Longsheng Tian

Research Institute of Petroleum and

Processing, Sinopec, Beijing, China Abstract
Highly efficient separation of benzene, toluene, and xylenes (BTXs) from
Correspondence
pyrolysis gasoline is very important in petrochemical industries. Though the
Gao Siliang, Research Institute of
Petroleum and Processing, Sinopec, extractive distillation (ED) process is simpler and consumes less energy
Beijing 100083, China. compared with liquid–liquid extraction process, it is difficult for a single
Email: [email protected]
solvent, for example, sulfolane, to achieve both high purity and high yield of
Funding information BTXs. In this work, methylsulfonylethane (MSE) was chosen as a cosolvent to
Sinopec Corp., Grant/Award Number: improve the selectivity of sulfolane after solvent screening, and factors that
417016-3
may affect the selectivity of the composite solvent were fully investigated, such
as the content of cosolvent, solvent to feed ratio, the composition of the feed,
and temperature. Furthermore, 240 h of continuous extractive distillation and
solvent recovery experiment was carried out using sulfolane (85 wt%)–MES
(15 wt%) mixture as solvent. The purity of mixed aromatics obtained was
99.83%, and the yield was as high as 99.7%.

KEYWORDS
aromatics, composite solvent, extractive distillation, methylsulfonylethane, pyrolysis gasoline

1 | INTRODUCTION consumption of the process increases dramatically when

the content of aromatics in the feed is higher than 65%.
BTXs (benzene, toluene, and xylenes) are important One of the reasons is that a certain amount of the
chemicals for petrochemical industries, and more than raffinate product needs to be remixed with the feed in
40 million of metric tonnes of benzene and xylenes and order to maintain two liquid phases in the extractor.
approximately the half for toluene are produced per year Besides, the extractive stripper needs to evaporate a
globally.1 The C6–C8 cut of pyrolysis gasoline with BTXs large amount of “aromatics reflux,” which returns back
content generally above 80% is one of the main sources. to the lower part of the extractor to ensure the high
The conventional process for the separation of these purity of BTXs, and the mass ratio of the refluxed
aromatics from pyrolysis gasoline is liquid–liquid aromatics to the feed is about 0.6–1.0 according to the
extraction, and the sulfolane process is the most widely industrial experience.
practiced currently.2,3 In this process, the separation of To reduce energy consumption, many efforts have
BTXs is made in an extractor followed by an extractive been made by researchers on the new design of separa-
stripper and a recovery column. However, the energy tion process.4 Extractive distillation (ED) process is

Asia-Pac J Chem Eng. 2021;16:e2609. wileyonlinelibrary.com/journal/apj © 2021 Curtin University and John Wiley & Sons, Ltd. 1 of 12
https://doi.org/10.1002/apj.2609
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
2 of 12 GAO ET AL.

believed to be more suitable if the content of aromatics to select a suitable cosolvent to improve separation
in the hydrocarbon feed is between 65% and 90%5 for its efficiency of the ED process for BTXs recovery.
low energy consumption. Pablo Navarro et al.6 proposed This study focused on the ED process aided by
the use of the ED process with ionic liquids (ILs) for the sulfolane and cosolvent to obtain aromatics with high
separation of aromatics from pyrolysis gasoline to over- purity and high yield, while avoiding energy loss caused
come liquid–liquid extraction limitations, and quantum by back-mixing. Seventeen organic compounds were
chemical-based simulation methodology (COSMO-based/ screened as cosolvents, and methylsulfonylethane (MSE)
Aspen Plus) was employed to simulate the ED process was found to achieve the highest improvement in
rigorously. Their results showed that extractive distilla- selectivity with sulfolane. The selectivity of the novel
tion stands as a more effective process configuration to composite solvent was investigated under different condi-
separate aromatics from aliphatics, even requiring less tions. Continuous experiment of extractive distillation
operating costs and investment. Ismael Díaz et al.7 and solvent recovery was carried out to evaluate the
made a conceptual design of ED process for the actual separation performance of the composite solvent,
aromatic–aliphatic separation from naphtha using using C6–C8 cut of pyrolysis gasoline as feedstock.
the 1-ethyl-3-methylimidazolium dycianamide [EMIM] Finally, the solvent to feed ratio on each stage in the ED
[DCA] IL as entrainer. They concluded that for the high column was discussed in detail to explore more informa-
aromatic content naphtha, good separation performance tion of the separation process.
was obtained with a significant reduction of the
energy required per ton of naphtha fed to the system in
comparison with the separation by liquid–liquid 2 | MATERIALS AND METHODS
extraction. Though a wide number of works has
been done to characterize the phase equilibria for 2.1 | Materials
(aliphatic + aromatic + ILs) systems as well as the IL
application in the ED process,8–14 to the best of Sulfolane (mass fraction purity: 99.0%) was supplied by
our knowledge, continuous experiment of aromatics Liaoning Guanghua Chem. Co., Ltd. Its decomposition
separation and solvent recovery using ILs as solvent has temperature is 220
C.31 The sources and mass fraction
not been performed yet; thus, the actual purity and yield purity of 17 organic compounds used in this work as
of the aromatic products remain uncertain. cosolvents were listed in Table 1. All the cosolvents have
On the other hand, organic compounds used as sol- good thermal stability with low toxicity, and their physi-
vents or cosolvents for aromatics separation have always cal properties were listed in Table S1. Three organic com-
attracted attention,15–24 such as sulfolane, N-methyl pounds (ethylene carbonate, 1-cyclohexyl-2-pyrrolidone,
pyrrolidone (NMP), N-formyl morpholine, ethylene and propylene carbonate) recommended in literature21,24
glycol, or propylene carbonate. However, most of them as solvents were purchased from Tokyo Chemical
were applied in the liquid–liquid extraction process.25,26 Industry, and their mass fraction purity were all higher
Their use in the ED process has not drawn attention until than 99%. Benzene (AR) and ethylcyclohexane (AR) were
recently. The sulfolane extractive distillation (SED) purchased from Sinopharm Chemical Reagent Co., Ltd.
process was studied by Tian's group 27 and mainly (Beijing, China). The C6–C8 cut of pyrolysis gasoline was
applied for the recovery of benzene or benzene and tolu- obtained from a Sinopec refinery, and its composition
ene.28,29 Gaile et al.30 also explored the application of was listed in Table 2.
sulfolane–NMP mixtures to extractive rectification of the
benzene fraction from reformed naphtha. The results
showed that after benzene separation, the distillate con- 2.2 | Screening of cosolvents
tains less than 1 vol% benzene and can be used as a com-
ponent of automobile gasolines. Apparently, sulfolane As described in Section 1, among all hydrocarbons from
has resulted the most extensively used due to its proper C6 to C8, it is the most difficult for benzene (B) and
capacity of aromatics and polarity, good thermal stability, ethylcyclohexane (ECH) to be separated, so they were
and nontoxicity. But when the fraction of the feed chosen as the model components of aromatics and nonar-
becomes wider, sulfolane or sulfolane–NMP composite omatics, respectively. Considering that in the ED column,
solvent is not selective enough to keep benzene (boiling if there was still a small amount of nonaromatics at the
point: 80.1
C) at the bottom of the ED column, while bottom, they would be sent to the recovery column with
evaporating C8 cycloalkanes, such as ethylcyclohexane the rich solvent and lower the purity of the final aromatic
(boiling point: 131.8
C), to the top. Thus, BTXs and non- products; therefore, the selectivity of solvent at the
aromatics cannot be totally separated. So it is necessary bottom of ED column is especially important. Because
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GAO ET AL. 3 of 12

TABLE 1 Compound description, mass fraction purity, and 2.3 | Measurement of relative volatility
supplier αECH/B
Mass
fraction In the ED process, the selectivity of solvent can be
Chemical purity Supplier characterized by the relative volatility of nonaromatics
2-Pyrrolidinone >99.0% Alfa Aesar Co. to aromatics. In this work, the headspace gas chroma-
tography (GC) was used to determine the relative vola-
N-methylpyrrolidone AR Sinopharm Chemical
(NMP) Reagent Co.
tility. The main instruments were as follows: Agilent
7694 headspace sampler; HP6890 GC (The chromato-
Dimethyl sulfone >99.0% Tokyo Chemical
graphic column is a capillary column with the length of
Industry
60 m and the inner diameter of 0.32 mm, and the sta-
MSE >98.0% Tokyo Chemical
tionary liquid is the modified PEG20000), FID hydrogen
Industry
flame detector. The measurement procedure was as
Dipropyl sulfone >99.0% Tokyo Chemical
follows: appropriate amount of hydrocarbon feed and
Industry
composite solvent were taken respectively and mixed
Isopropyl methyl >97.0% Tokyo Chemical
according to a certain mass ratio. A mixed solution
sulfone Industry
(10 ml) was placed in the balancer of the headspace
Diethylene glycol >98.0% Sinopharm Chemical
GC. The balancer had a volume of 20 ml, and its lid
Reagent Co.
would be sealed after the prepared solution was added.
Triethylene glycol AR Sinopharm Chemical By shaking and letting stand for 1 h at a certain tem-
Reagent Co.
perature, vapor–liquid equilibrium was achieved. Then
Dipropylene glycol CP Sinopharm Chemical the sample of the vapor phase in the balancer was
Reagent Co.
taken by a special needle supplied with the instrument,
Ethylene glycol >99.0% Tokyo Chemical and its composition was analyzed by GC. The detector
monobutyl ether Industry
temperature of GC was 280
C, column temperature was
Ethylene glycol >98.0% Tokyo Chemical 80
C, and it held for 20 min. The carrier gas was high-
monophenyl ether Industry purity nitrogen with a flow rate of 25 ml/min; the com-
Diethylene glycol >99.0% Aladdin Biochemical bustion gas was high-purity hydrogen with a flow rate
monomethyl ether Technology Co., Ltd. of 30 ml/min; and the assistant combustion gas was
Diethylene glycol >98.0% Alfa Aesar Co. purified air with a flow rate of 350 ml/min.
monoethyl ether Because the amount of hydrocarbons that enter
Diethylene glycol CP Sinopharm Chemical the vapor phase due to volatilization was little, its
monobutyl ether Reagent Co. influence on the composition of the liquid phase can
Triethylene glycol >98.0% Tokyo Chemical be ignored. The relative volatility αECH/B of the two
monomethyl ether Industry hydrocarbon components was calculated according to
Tetraethylene glycol >98.0% Alfa Aesar Co. Equation 1.
monomethyl ether
Tetraethylene glycol >98.0% Alfa Aesar Co.
αNA=A = ðyNA =yA Þ=ðx NA =x A Þ = ðANA =AA Þ=ðx NA =x A Þ ð1Þ
dimethyl ether

In the formula, the subscripts NA and A stand for

nonaromatics and aromatics, respectively; x represents
the aromatics were enriched in that zone, benzene–ECH the mass fraction of the component in the liquid phase;
solution with a mass ratio of 9:1 was prepared as the y represents the mass fraction of the component in the
simulated hydrocarbon feed. vapor phase; A represents the chromatographic peak area
The solution of sulfolane and a certain proportion of the vapor sample in the balancer. The value of αNA/A
of cosolvent were prepared as the composite solvent. can characterize the solvents selectively. The larger its
Then it was added into the simulated hydrocarbon value is, the easier it is for nonaromatics to be evaporated
feed at a mass ratio of 5:1. The mixture was to the top of ED column whereas the aromatics still
equilibrated at 100
C for 1 h before relative volatility remain in the solvent at the bottom; conversely, the
measurement. The solvent that could achieve the closer the value comes to 1, the more difficult it is for the
highest relative volatility was considered to be the two hydrocarbons to be separated by extractive
most suitable. distillation.
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
4 of 12 GAO ET AL.

TABLE 2 Composition of pyrolysis gasoline C6–C8 cut

Weight percentage/%

C number N-alkane Isoparaffin Alkene Cycloalkane Aromatics Total

5 0.03 0 0 1.43 0 1.46
6 1.59 2.20 0.04 6.18 46.71 56.72
7 0.41 0.44 0.01 1.54 21.68 24.08
8 0.11 0.82 0.01 0.69 15.64 17.27
9 0.15 0.23 0 0.04 0 0.42
10 0 0.05 0 0 0 0.05
Total 2.29 3.74 0.06 9.88 84.03 100.00

TABLE 3 Parameters of experimental columns

Item ED column SR column

Diameter/mm 30 40
Packing type Triangle packing Triangle packing
Packing size 1.5 mm × 1.5 mm 1.5 mm × 1.5 mm
Number of 40 40
theoretical plates

was injected into the bottom of recovery column at a

mass flow rate of 50 g/h. Both columns were operated
under normal pressure.
FIGURE 1 Schematic process flow diagram of the continuous The equipment ran stably for 240 h, and the samples
experiment of mixed aromatics and raffinate were taken every 2 h.
Their compositions were analyzed by HP6890 GC. The
solvent to feed ratio (Sc) was defined as the mass flow
2.4 | Continuous experiment of rate ratio of the solvent to the gasoline feedstock and can
extractive distillation and solvent recovery be adjusted by changing the flow rate of the gasoline
feedstock. After each adjustment of Sc, the equipment
The schematic process flow diagram of the continuous ran stably for at least 48 h, and 10 samples obtained
experiment was shown in Figure 1. The hydrogenated within 10 consecutive hours of stable operation were cho-
pyrolysis gasoline C6–C8 fractions were sent to the middle sen to represent the separation result, half of which were
part of the ED column at the mass flow rate of 150– raffinate and the other half were mixed aromatics. It was
200 g/h after preheating to 60
C, and the sulfolane/MSE worth mentioning that Sc was not exactly the same as S,
composite solvent with MSE mass fraction of 15% was which will be discussed in detail later. The parameters of
sent to the upper part of the column at a fixed flow rate the columns used in the experiment were listed in
(1200 g/h). Two mass flowmeters (Krohne, OPTIMASS Table 3.
3000) were used to measure the mass flow rate of the At the end of the experiment, the sample of the lean
above two streams respectively to facilitate the control. solvent was taken and analyzed by GC to examine
After extractive distillation, the nonaromatics (raffinate) whether the solvent was thermodynamically stable and
were obtained at the top of the ED column, and the rich whether there were aromatics with higher boiling point
solvent containing high-purity aromatics was obtained at (such as O-xylene) left in the solvent. The GC
the bottom. Then the rich solvent was sent to the solvent analysis conditions were basically the same as those for
recovery column (SR column), where the mixed aro- analyzing hydrocarbon composition, except that the
matics and solvent were separated. The solvent was sent column temperature rose to 200
C.
back to the ED column, and the mixed aromatics were The yield of the mixed aromatics and each aromatic
collected as products. A small stream of stripping water hydrocarbon was calculated based on Equation 2.
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GAO ET AL. 5 of 12

xb xf − xr TABLE 4 Changes in αECH/B before and after adding solvents
yield of aromatics = × 100% ð2Þ
x f ðx b − x r Þ Mass fraction of
Solvent cosolvent/% αECH/B
(none) — 0.140
Sulfolane 0 1.537
x: mass fraction of aromatics.
Subscripts. Sulfolane/2-pyrrolidinone 10 1.583
f: gasoline feedstock; b: aromatics product; r: raffinate Sulfolane/NMP 10 1.509
product. Sulfolane/dimethyl sulfone 10 1.533
Sulfolane/MSE 10 1.693
Sulfolane/dipropyl sulfone 10 1.540
3 | R E S U L T A N D DI S C U S S I O N
Sulfolane/isopropyl methyl 10 1.627
sulfone
3.1 | Cosolvent screening
Sulfolane/diethylene glycol 10 1.474

The screening of cosolvents was conducted according to Sulfolane/triethylene glycol 10 1.458

the following aspects. (1) Proper polarity, which is benefi- Sulfolane/dipropylene glycol 10 1.475
cial to aromatics separation; (2) a significant difference in Sulfolane/ethylene glycol 10 1.272
boiling point from aromatics (>200
C), which is helpful monobutyl ether
to solvent recovery; (3) good thermal stability and Sulfolane/ethylene glycol 10 1.402
chemical stability to ensure that aromatics are not monophenyl ether
contaminated by degradable substances; (4) nontoxic, Sulfolane/diethylene glycol 10 1.345
noncorrosive. Therefore, several different solvent types monomethyl ether
containing different functional groups were screened in Sulfolane/diethylene glycol 10 1.290
this study, for example, sulfones (solvents containing monoethyl ether
sulfur atoms), pyrrolidones (solvents containing nitrogen Sulfolane/diethylene glycol 10 1.268
atoms), glycols, and ethers. Solvents containing halogen monobutyl ether
atoms were not considered due to their toxicity. Alcohols, Sulfolane/triethylene glycol 10 1.438
aldehydes, phenols, and esters were not taken into monomethyl ether
account because they were either easily oxidized or easily Sulfolane/tetraethylene glycol 10 1.410
hydrolyzed. Acid and amine are not suitable to be used methyl ether
as cosolvents because they are corrosive. Therefore, Sulfolane/tetraethylene glycol 10 1.476
17 organic compounds as representatives of cosolvents dimethyl ether
were screened in this study.
Ethylene 35 1.405
Table 4 listed the values of αECH/B before and after carbonate/1-cyclohexyl-
adding different solvent. It can be seen that αECH/B was 2-pyrrolidone
only 0.140 without any solvent, indicating that benzene Propylene carbonate/diethylene 15 1.378
will be distillated to the top of the column rather than glycol
ECH. After adding sulfolane, αECH/B jumped to 1.537,
showing that sulfolane was a favorable main solvent.
Compared with pure sulfolane, the addition of
2-pyrrolidinone, MSE, and isopropyl methyl sulfone sulfolane/MSE composite solvent is more selective
could further increase αECH/B. Among them, the improve- than sulfolane and can be used for BTXs recovery. It
ment to αECH/B by MSE was the most significant, and the should be pointed out that we did not find clear
value of αECH/B further increased to 1.693. The change of advantages of sulfolane/NMP composite solvent in the
αECH/B was small after adding dimethyl sulfone and ED process according to αECH/B, though it had been
dipropyl sulfone; other cosolvents may even cause a recommended in the literature.30,33 In addition, ethylene
decrease in αECH/B. Values of αECH/B after adding differ- carbonate/1-cyclohexyl-2-pyrrolidone and propylene
ent amount of cosolvent (except MSE) were shown in carbonate/diethylene glycol composite solvents rec-
Table S2. It can be seen that sulfolane/MSE is the best ommended in the liquid–liquid extraction process did not
composite solvent within the scope of the experiment. show high selectivity in the ED process either. Therefore,
Kang reported that ED process is industrially applicable sulfolane/MSE composite solvent will be mainly studied
when α is greater than 1.2.32 The results showed that in the following research.
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
6 of 12 GAO ET AL.

It is known that the relative volatility can be industrial application, the sulfolane/MSE composite sol-
increased by either inducing positive deviations from vent with an MSE mass fraction of 15% was used as the
Raoult's law for the interaction of the solvent with the solvent in the following experiments.
low-boiler or by inducing negative deviations from The use of MSE in combination with sulfolane con-
Raoult's law for the interaction with the high-boiler. tributes to the creation of moderate polar medium.
Also, a combination of both is possible. According to Yang et al.35 studied the dipolarity/polarizability and
Everson's research,34 the dielectric constant and the the hydrogen-bond basicity of the composite solvent to
dipole moment, which related to the polarity of the determine the microscopic solvent properties. They
molecules, have the greatest influence on the relative found that the increase of basicity and the increase of
volatility. Because benzene is more polar than ECH and dipolarity/polarizability had inverse effects on the sepa-
is more easily polarized by solvents, the interaction ration of the solute. In our study, using sulfolane/MSE
between benzene and solvent may be the reason of the composite solvent with an MSE mass fraction of 95% as
change in relative volatility. Under the induction of solvent, αECH/B was relatively small due to the relatively
solvent, the higher the polarization degree of benzene, weak hydrogen-bond basicity; with pure sulfolane as
the greater the value of αECH/B. The polarity of the com- solvent, αECH/B was also relatively small due to its rela-
posite solvent can be slightly changed due to different tively strong dipolarity/polarizability. As a result of the
cosolvent, which may also cause changes in αECH/B. combined effect of both basicity and dipolarity/polariz-
ability, αECH/B can achieve a larger value than pure
cases. That is why there was an optimal value for the
3.2 | Selectivity of sulfolane/MSE cosolvent content.
composite solvent

3.2.1 | Effect of MSE content on αECH/B 3.2.2 | Effect of S on αECH/B

Figure 2 showed the effect of composite solvent composi- In the ED column, the mass flow of the solvent on each
tion on αECH/B, and the mass ratio of solvent to feedstock stage could be considered unchanged due to its non-
(solvent ratio, S) was 5.0, whereas the vapor–liquid equi- volatility. However, with the separation of the aromatics
librium temperature varied from 80
C to 120
C. It can be and nonaromatics, the mass flow of the hydrocarbons on
seen that after adding 15(wt)% of MSE in the sulfolane, each stage varied. Thus, the mass ratio of the solvent to
αECH/B could increase by about 10%–15% compared with hydrocarbons at different location of the column was not
sulfolane, which was very beneficial for the separation of a constant. Figure 3 showed the effect of S on αECH/B,
aromatics and nonaromatics. With the increase of MSE and the solution of benzene–ECH at the mass ratio of 9:1
content, the value of αECH/B tended to be stable and then was still used as simulated feed. The range of vapor–
dropped dramatically, and the maximum value at 100
C liquid equilibrium temperature was 80
C–120
C. It can
was 1.809. Considering the cost of adding cosolvent in be seen that when S is 3.2 at 100
C, αECH/B was only

F I G U R E 2 αECH/B using sulfolane/MSE composite solvent FIGURE 3 αECH/B using sulfolane/MSE composite solvent as a
with different MSE content function of S
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GAO ET AL. 7 of 12

1.323, which was a little difficult to achieve full separa-

tion of ECH and benzene. With the increase of S, the
value of αECH/B went up significantly. After S reached 7.3,
αECH/B could be as high as 2.161 at 100
C, and the diffi-
culty of separation may be greatly reduced. If S rose con-
tinuously, the selectivity of the composite solvent
increased slowly and then tended to stabilize, but the
energy consumption would continue to increase accord-
ingly. Therefore, S should be kept within a reasonable
range in industrial application.

3.2.3 | Effect of feed composition on

αECH/B
FIGURE 4 αECH/B of hydrocarbon feed with different
Obviously, the concentration of aromatics was higher at composition
the lower part of the ED column compared with other
locations, and the concentration of nonaromatics
reached the maximum at the top. Thus, there was a benzene mass fraction of 30%, temperature hardly
certain distribution of the hydrocarbon composition in showed effect on αECH/B. It was observed that the feed
the column. The final purity and yield of the aromatic containing 20%(w) benzene cannot be totally dissolved by
products depended on whether the sulfolane/MSE the solvent until the temperature was higher than 100
C,
composite solvent could keep a high selectivity for whereas the feed containing 40%(w) benzene and the sol-
hydrocarbon feed with different compositions. It is vent will become homogeneous when the temperature
worth mentioning that the composition of feed not only was higher than 80
C. Thus, the effect of temperature
affected αECH/B but also had significant influence on seemed to relate with whether the liquid phase is homo-
the phase state of the whole system. More specifically, geneous or heterogeneous.
increasing the content of benzene will lead the solvent– Navarro et al.14 used ILs in the extractive distilla-
hydrocarbon system to vapor–liquid equilibria (VLE) tion process for n-heptane/toluene separation, and they
and further from vapor–liquid–liquid equilibria during also measured relative volatility in both VLE and
the measurement of αECH/B, in any case for the same vapor–liquid–liquid equilibria cases. According to their
solvent ratio. When the mass fraction of benzene in the research, there were two driving forces regarding the
feed was as high as 90%, the whole system was always use of a solvent: capacity and selectivity; capacity
homogeneous under all conditions within the experi- worked preferably for low solubility cases (heteroge-
mental range. On the contrary, if the content of ECH neous), and selectivity was the essential property of a
increased, two liquid phases were likely to form due to mass agent for high solubility (homogeneous) ones.
the poor solubility of ECH in the solvent. Increasing temperature was beneficial to improve the
Figure 4 showed the changes of αECH/B after adding capacity of the solvent, so it showed positive effect on
the composite solvent to benzene–ECH solutions with αECH/B when the liquid phase was heterogeneous;
different composition at a solvent ratio (S) of 15.0. It can whereas after all the hydrocarbon feed were dissolved,
be seen that as the mass fraction of benzene decreased increasing temperature will result in a decrease in
gradually, αECH/B significantly went down. The result selectivity of the solvent.36 Therefore, the value of
suggested that under the same condition, the composite αECH/B dropped. In addition, the feed with a benzene
solvent has a high selectivity for the feed enriched with mass fraction of 20%–40% were in a critically mutually
aromatics and a low selectivity when the content of soluble state at a solvent ratio of 15.0 and a tempera-
aromatics went down. ture range of 80
C–100
C, and the result showed that
Moreover, it is interesting to find that if the mass the impact of temperature was negligible.
fraction of benzene was less than 30%, increasing temper- Finally, it can also be seen that if the mass fraction
ature led to a higher αECH/B, although the influence of of ECH was 90% in the feed, two liquid phases still
temperature was less and less obvious as benzene content formed even S was as high as 15.0 at the temperature
went up. When the mass fraction of benzene exceeded of 120
C, indicating that the coexistence of vapor–
30%, temperature had a negative effect on αECH/B for any liquid–liquid was inevitable in the separation of C6–C8
feed with a certain composition. For the feed with a fractions.
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
8 of 12 GAO ET AL.

3.2.4 | Effect of temperature on αECH/B αECH/B cannot be obtained at a higher temperature due
to the upper measurement limitation of our experimental
To further confirm the different trends of αECH/B when instrument, the trend of αECH/B with temperature can be
the liquid phase is homogeneous and heterogeneous extrapolated in a wider range. If other conditions remain
phases, the value of ECH/benzene relative volatility with unchanged, αECH/B will be lower when the temperature
temperature at two different solvent ratios was measured, gets higher because the liquid phase is homogenous at
and the results were shown in Figure 5. The hydrocarbon the bottom of ED column.
feed was benzene–ECH solution with a mass ratio of 7:3.
It can be seen that when the solvent ratio was high and
the hydrocarbon feed can be completely dissolved in the 3.3 | Continuous extractive distillation
solvent at room temperature, the decrease in equilibrium and solvent recovery using sulfolane/MSE
temperature could contribute to the increase in αECH/B. composite solvent
The αECH/B could rise to 3.315
C at 40
C. The trend
obtained in the experiment was consistent with the litera- Because the aromatics distillation unit was not included
ture.37 The reason might be the change in polarity of the in our experimental equipment, only the mixed aromatics
composite solvent with temperature. The forces among and the raffinate can be obtained in the end. Two indexes
polar molecules mainly include the orientation force and were used to evaluate the separation performance. The
the hydrogen bonds. In this system, the orientation force first one was the purity of mixed aromatics, which was
is dominant, and it increases with the polarity of mole- defined as the mass fraction of BTXs in mixed aromatic
cules. When the temperature decreases, the absolute products (the same below). The second one was the mass
value of interaction energy among molecules becomes fraction of aromatics (benzene, mainly) in the raffinate,
larger,38 resulting in the increase of the polarity of the which had a direct effect on the yield of mixed aromatics.
solvent. Thus, the selectivity turns out to be stronger. On Increasing the evaporation of the rich solvent at the bot-
the other hand, when the solvent ratio was as low as 0.5 tom of ED column could enable more nonaromatics to be
and the liquid phase was always heterogeneous, increas- distillated to the top and contribute to an increase in the
ing temperature helped to improve αECH/B, though the purity of mixed aromatics. However, it might also cause
improvement was not large. The trends were consistent more loss of benzene whose boiling point is the lowest
with the experiment results of Section 3.2.3. The cross- among all aromatics and reduce the yield. Moreover, the
relationship of temperature and solvent ratio was also raffinate cannot be qualified as gasoline blending compo-
reported and fully discussed in Navarro's research.14 nent due to the excessive benzene content. On the con-
For the application in industry, the solvent ratio trary, reducing the heat duty of the reboiler may result in
cannot be too low for BTXs separation. Besides, it is an increase in the yield of aromatics, but more nonaro-
difficult to conduct extractive distillation at low tempera- matics may be left in the aromatics; thus, the purity of
tures such as 40
C. Instead, the actual temperature at ED BTXs may decrease. Therefore, the two aspects should be
column bottom is higher than 120
C generally. Though both taken into consideration during the operation, and
effort needed to be made to maintain a balance of aro-
matics purity and their yield.
The effect of Sc on the separation was investigated,
and the main operation conditions were listed in Table 5.
It can be seen that the temperatures both at the top (Ttop)
and bottom (Tbottom) of ED column changed most signifi-
cantly with Sc. With the increase of Sc, Tbottom rose gradu-
ally. The reason may be that the concentration of the
solvent at the bottom of the ED column was elevated due
to less gasoline feedstock entering into the column, and
the boiling point of the solvent was higher than 200
C.
Ttop basically remained the same when Sc was 6.0 and
7.0, but as Sc increased further, Ttop also went
up. Because the temperature changes of distillation col-
umn reflect the changes in composition directly, the
results indicated an obvious influence of Sc on the raffi-
F I G U R E 5 αECH/B using sulfolane/MSE composite solvent as a nate. Higher Sc may result in a larger αECH/B, and when
function of temperature the selectivity was improved to a certain extent, the
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GAO ET AL. 9 of 12

TABLE 5 Main operating parameters of the continuous automobile gasolines, and the best result obtained in our
experiment experiment was very close to but still a little higher than
that value, we recommended adding auxiliary equipment
Value
at the top of ED column, such as a membrane separation
Items Sc = 6.0 SC = 7.0 SC = 8.0 unit39 or a small extraction column, to recover the extra
ED column benzene.
Ptop (absolute)/kPa 101 101 101 The composition and yield of the mixed aromatic
Reflux ratio 0.5 0.5 0.5 products and the raffinate obtained at the highest Sc were


listed in Table 8. It can be seen that the purity of mixed
Ttop/ C 80–84 80–85 84–87


aromatic products was 99.83%, and the major impurity
Tbottom/ C 122–124 123–125 126–128
was cycloalkanes. Although the contents of C6, C7, and
Solvent recovery column C8 cycloalkanes in the feedstock dropped in order, their
Ptop (absolute)/kPa 101 101 101 contents increased successively in the mixed aromatic
Reflux ratio 1.0 1.0 1.0 products due to the difficulties in separation. As the con-


Ttop/ C 73–77 73–77 73–77 tent of C9 cycloalkanes was extremely low in feedstock,
their content in mixed aromatic products were still low
Tbottom/
C 174–176 174–176 174–176
even if they can hardly be separated.
It should be pointed out that the concentration of
other aromatics except benzene was too low to be
amount of light aromatics (e.g., benzene) evaporated to measured in the raffinate, meaning that the yield of
the top may be dramatically decreased. Therefore, the toluene and xylenes could easily exceed 99.9% in the
temperature rose up. extractive distillation process using sulfolane/MSE com-
The purity of mixed aromatic products obtained and posite solvent, whereas the yield of benzene was lower.
the mass fraction of benzene in the raffinate at three dif- That is a little different from the yield of aromatics in the
ferent Sc were shown in Tables 6 and 7, respectively. To liquid–liquid extraction process, in which the yield of
facilitate comparison, the purity of mixed aromatics was benzene is always the highest. Besides, the mass fraction
maintained at the same level by adjusting the heat duty of sulfolane and MSE in the lean solvent was higher than
of the ED column reboiler. It can be seen from Table 6 99.7% after the continuous experiment, and no decompo-
that the purity of the mixed aromatic products ranging sition products were detected, showing the efficient
from 99.8% to 99.9% was quite stable for at least 10 h in solvent recovery process and good thermal stability of the
every condition. However, the mass fraction of benzene composite solvent.
in raffinate was significantly reduced with the increase of
Sc, as shown in Table 7. When the Sc was 6.0, the average
mass fraction of benzene in the raffinate was as high as 3.4 | Discussion on the solvent ratio in
8.10% with large fluctuation. The reason may be that two the ED column
liquid phases appeared in the upper part of the ED col-
umn at the lower solvent ratio due to the poor solubility According to the experiment result in Section 3.2, when
of the composite solvent for the nonaromatics, which the solvent ratio (S) was 7.2, αECH/B was already up to
could lead to unstable operation of distillation. The aver- 1.982
C at 120
C for hydrocarbon feed that enriched with
age mass fraction of benzene in raffinate was reduced to aromatics, and it would continue to increase as S became
3.31% when the Sc was 7.0, and that value dropped to even larger. So if the mass ratio of solvent to hydrocar-
1.50% after Sc further increased to 8.0. Because it is gener- bons at the bottom of ED column could have been higher
ally accepted that raffinate, which contains less than than 7.0, it would be easy to separate benzene and ECH,
1 vol% benzene, can be used as a component of not to mention the separation of other aromatics and

T A B L E 6 Purity of mixed aromatic

Purity of mixed aromatic products/%
products under different conditions
Sc Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
6.0 99.823 99.856 99.838 99.809 99.806
7.0 99.837 99.801 99.827 99.838 99.807
8.0 99.849 99.823 99.814 99.827 99.823
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
10 of 12 GAO ET AL.

TABLE 7 Loss of benzene in the

Mass fraction of benzene in the raffinate/%
raffinate
Sc Sample 1 Sample 2 Sample 3 Sample 4 Sample 5
6.0 7.484 5.765 11.801 10.015 5.423
7.0 3.242 2.285 3.574 3.812 3.640
8.0 1.961 0.944 1.252 1.861 1.481

TABLE 8 Purity and yield of mixed aromatics (Sc = 8.0) detailed information in the ED column, PRO/II soft-
ware was used to simulate the separation process, and
Item Experimental value
the nonrandom-two-liquid model and binary interaction
Composition of mixed aromatics parameters regressed from the unpublished VLE experi-
Aromatics/% 99.83 mental data were applied for calculation. Because MSE
C6 cycloalkanes/(μg/g) 97 did not exist in the original databank of the software, it
Dimethylcyclohexane/(μg/g) 268 was simulated as the pseudo-component. The feedstock
ECH/(μg/g) 905 and the operating conditions were the same as those of
the continuous experiment, and the relative deviation
C9 cycloalkanes/(μg/g) 430
in mass flow rate and aromatics content between the
Composition of raffinate
simulated and the actually separated raffinate was less
Benzene/% 1.5 than 2%, indicating the simulation was basically consis-
Toluene/% 0 tent with the experiment.
C8 aromatics/% 0 In the simulation model, the theoretical trays were
Yield of aromatics/% 99.7 numbered 1–40 from the top of the column to the
Yield of benzene/% 99.5
bottom. The gasoline feedstock entered the ED column
on the 20th stage, and the composite solvent entered on
Yield of toluene/% 100
the second stage. The result showed that the S cannot be
Yield of C8 aromatics/% 100 maintained at 8.0 from top to bottom of the ED column.
As shown in Figure 6, S was about 7.0–10.0 between the
feeding position of pyrolysis gasoline and the seventh
nonaromatics. However, the result of continuous experi- stage, and it dropped to only about 3.9 below the 20th
ment showed that the concentration of C6–C8 nonaro- stage. It can be also found from Figure 6 that the mass
matics was still 1270 μg/g in mixed aromatic product fraction of aromatics in hydrocarbons in the liquid phase
even if the solvent ratio (Sc) was 8.0. The confusing was higher than 90% below the seventh stage, and the
results can be explained from the difference in definition temperature in that zone was from 100
C to 120
C
of the two “solvent ratios.” S is the concept of solvent
ratio in theory, and it determined whether aromatics can
be totally separated, although Sc is more convenient to
measure in industrial operation. Though it has been
widely accepted by refineries, it is not the mass ratio of
solvent to hydrocarbons in the ED column, especially on
each stage.
In fact, because the solvent can be considered non-
volatile and remained in the liquid phase on each tray
of the ED column, its mass flow rate can be regarded
as a constant. With the evaporation of nonaromatics
into the vapor phase, the amount of hydrocarbons in
the liquid phase will be less and less from the bottom
to the top of the ED column. Thus, S may change
significantly with the separation of BTXs and nonaro-
matics at different positions in the column. This is
caused by the flow of substances in the column, not F I G U R E 6 S and mass fraction of aromatics on each stage of
related to the solvent or cosolvent. To obtain the the ED column (Sc = 8.0)
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
GAO ET AL. 11 of 12

according to both of the simulation and experiment ORCID

results. Considering the trend of αECH/B obtained in Siliang Gao https://orcid.org/0000-0001-7143-4867
Section 3.2, it can be inferred that real αECH/B could only
reach about 1.4–1.5 at the lower part of ED column. RE FER EN CES
Therefore, the selectivity of composite solvent reduced 1. Global Chemicals Outlook, Towards Sound Management of
significantly, and the mixed aromatics obtained still had Chemicals, United Nations Environment Programme; 2013.
a certain number of nonaromatics that cannot be 2. Gary J, Handwerk G, Kaiser M. Petroleum Refining Technology
completely removed. and Economics. fifth ed. Boca Raton, FL: CRC Press; 2007:1024.
Besides, it is also interesting to find out that S was 3. Meyers RA. Handbook of Petroleum Refining Processes. third ed.
New York: McGraw-Hill; 2004:753.
higher than 10.0 above the seventh stage. The reason
4. Wang Q, Chen JY, Pan M, et al. A new sulfolane aromatic
might be the low amount of hydrocarbons in the liquid
extractive distillation process and optimization for better
phase in this zone. Because the mass fraction of energy utilization. Chem Eng Process. 2018;128:80-95.
aromatics in the pyrolysis gasoline was 84.03%, about 5. Weissermel K, Arpe HJ. Aromatics-Production and Conversion.
16% of the feedstock (nonaromatics) would be distillated Industrial Organic Chemistry (fourth ed.). Wiley-VCH:
as the raffinate product. Considering the reflux ratio was Weinheim; 2003:58.
0.5, only 8% of the feedstock existed in the liquid phase. 6. Navarroa P, García ID, Larriba M, et al. Dearomatization of
Thus, S should be very high on the premise that the pyrolysis gasoline by extractive distillation with 1-ethyl-
3-methylimidazolium tricyanomethanide. Fuel Process Technol.
solvent mass flow on each stage was basically the same.
2019;195:106156-106163.
Though the concentration of the nonaromatics in the 7. Díaz I, Palomar J, Rodríguez M, Riva J, Ferro V, González EJ.
upper part of the ED column was high, which may result Ionic liquids as entrainers for the separation of aromatic–
in a drop in αECH/B according to Section 3.2.3, the actual aliphatic hydrocarbon mixtures by extractive distillation. Chem
αECH/B should not be too low that the aromatics cannot Eng Res des. 2016;115:382-393.
be separated due to the increase of S. 8. Larriba M, Navarro P, García J, Rodríguez F. Liquid-liquid
extraction of toluene from n-heptane by {[emim][TCM]
+[emim][DCA]} binary ionic liquid mixtures. Fluid Phase
Equilib. 2014;364:48-54.
4 | C ON C L U S I ON 9. Larriba M, Navarro P, García J, Rodríguez F. Separation of tol-
uene from n-heptane, 2, 3-dimethylpentane, and cyclohexane
Sulfolane/MSE composite solvent was applied as using binary mixtures of [4empy][Tf2N] and [emim][DCA]
entrainer for BTXs recovery, and the mixture of sulfolane ionic liquids as extraction solvents. Sep Purif Technol. 2013;120:
(85 wt%)–MSE (15 wt%) showed excellent separation per- 392-401.
formance. Factors such as solvent ratio (S), composition 10. Yao C, Hou Y, Ren S, Wu W, Liu H. Selective extraction of aro-
of the feed, and temperature all have impacts on the matics from aliphatics using dicationic ionic liquid-solvent
composite extractants. J Mol Liq. 2019;291:111267-111273.
selectivity. The extractive distillation process of BTXs
11. Navarro P, Larriba M, García J, Rodríguez F. Thermal stability,
from pyrolysis gasoline C6–C8 cut using sulfolane–MES
specific heats, and surface tensions of ([emim][DCA]+[4empy]
composite solvent was investigated, and the results of [Tf2N]) ionic liquid mixtures. J Chem Thermodyn. 2014;76:
continuous experiment showed that Sc had influence on 152-160.
the operation condition as well as the BTXs purity and 12. Navarro P, Larriba M, García J, Rodríguez F. Thermal stability
yield. Mixed aromatics with purity of 99.83%, and a yield and specific heats of {[emim] [DCA] + [emim][TCM]} mixed
of 99.7% can be obtained when Sc was 8.0. Though the ionic liquids. Thermochim Acta. 2014;588:22-27.
mass fraction of benzene in the raffinate was still a little 13. Ayuso M, Barcala AC, Larriba M, et al. Enhanced separation of
benzene and cyclohexane by homogeneous extractive distilla-
higher than 1 vol% under the optimized operating condi-
tion using ionic liquids as entrainers. Sep Purif Technol. 2020;
tions, it is believed that the ED process can be further 240:116583-116592.
improved by adding auxiliary equipment to meet the 14. Navarro P, Larriba M, Mellado ND, et al. Experimental screen-
requirements of the raffinate product. Through detailed ing towards developing ionic liquid-based extractive distillation
simulation of the ED column, it was found that though in the dearomatization of refinery streams. Sep Purif Technol.
Sc was as high as 8.0, S was only about 1.4–1.5 at the 2018;201:268-275.
lower part of the ED column, which can explain why a 15. Yorulmaz Y, Karpuzcu F. Sulpholane versus diethylene
small amount of nonaromatics remained in the mixed glycol in recovery of aromatics. Chem Eng Res des. 1985;63:
184-190.
aromatic product.
16. Antosik M, Malanowski S. Liquid-liquid equilibria in hydrocar-
bon+3-methoxypropionitrile systems, correlation and predic-
ACK NO WLE DGE MEN TS tion. Fluid Phase Equilib. 1993;85:161-170.
The authors of this work want to thank Sinopec Corp. for 17. Graham HL. Extraction of benzene with di-and tri-ethylene
financial support of this project (417016-3). glycols. J Chem Eng Data. 1962;7(2):214-217.
 19322143, 2021, 3, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/apj.2609 by South African Medical Research, Wiley Online Library on [18/07/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
12 of 12 GAO ET AL.

18. AI-Zayied TA, Al-Sahhaf TA, Fahim MA. Measurement of 31. Cheng NL. Solvent Handbook. fifth ed. Beijing: Chemical
phase equilibrium in multicomponent systems of aromatics Industry Press; 2015:914.
with N-methyl pyrrolidone and predictions with UNIFAC. 32. Kang YM, Jeon Y, Lee G, et al. Quantitative structure relative
Fluid Phase Equilib. 1990;61(1-2):131-144. volatility relationship model for extractive distillation of eth-
19. Naithani J, Khanna MK, Nanoti SM, Rawat BS. Quaternary ylbenzene/p-xylene mixtures. Ind Eng Chem Res. 2014;53(27):
liquid-liquid equilibrium studies on hydrocarbon-solvent sys- 11159-11166.
tems. J Chem Eng Data. 1992;37(1):104-106. 33. Wang Q, Zhang BJ, He CC, Chen QL. Optimal design of a new
20. Wang W, Gou Z, Zhu S. Liquid-liquid equilibria for aromatics aromatic extractive distillation process aided by a co-solvent
extraction with tetraethylene glycol. J Chem Eng Data. 1998; mixture. Energy Procedia. 2017;105:4927-4934.
43(1):81-83. 34. Everson RC, van der Merwe BJ. The effects of selected solvents
21. Radwan GM, Al-Muhtaseb SA, Dowaidar AM, Fahim MA. on the relative volatility of a binary system consisting of
Extraction of aromatics from petroleum naphtha reformate by 1-octene and 2-hexanone. Fluid Phase Equilib. 1998;143(1-2):
a 1-cyclohexyl-2-pyrrolidone/ethylene carbonate mixed solvent. 173-184.
Ind EngChem Res. 1997;36(2):414-418. 35. Yang QW, Xing HB, Su BG, et al. Improved separation effi-
22. Saha M, Rawat BS, Khanna MK, Nautiyal BR. Liquid-liquid ciency using ionic liquid-cosolvent mixtures as the extractant
equilibrium studies on toluene heptane solvent. J Chem Eng in liquid-liquid extraction: a multiple adjustment and synergis-
Data. 1998;43(3):422-426. tic effect. Chem Eng J. 2012;181-182:334-342.
23. Ali SH, Lababidi HMS, Merchant SQ, Fahim MA. Extraction of 36. Larriba M, Navarro P, García J, Rodríguez F. Liquid-liquid
aromatics from naphtha reformate using propylene carbonate. extraction of BTEX from reformer gasoline using binary mix-
Fluid Phase Equilib. 2003;214(1):25-38. tures of [4empy][Tf2N] and [emim][DCA] ionic liquids. Energy
24. Lababidi HMS, Ali SH, Fahim MA. Optimization of aromatics Fuel. 2014;28(10):6666-6676.
extraction of naphtha reformate by propylene 37. Zhang ZG, Xu SM, Li XG, Zhang WJ. Influence of temperature
carbonate/diethylene glycol mixed solvent. Ind Eng Chem Res. on solvent selectivity for extractive distillation. J Chem Ind Eng
2006;45(14):5086-5097. (China). 2004;55(10):1740-1742.
25. Gaile AA, Erzhenkov AS, Koldobskaya LL. Extraction of tolu- 38. Reichardt C. Solvent Effects in Organic Chemistry. Verlag
ene and xylenes from the toluene xylene fraction of the Chemie Weinheim: Weinheim; 1979:256.
reforming catalyzate with mixtures of triethylene glycol with 39. Wang NX, Wu T, Wang L, et al. Hyperbranched polymer com-
sulfolane. Russ J Appl Chem. 2007;80:608-611. posite membrane using water as solvent for separating
26. Mahmoudi J, Lotfollahi MN. (liquid+liquid) equilibria of aromatic/aliphatic hydrocarbon mixtures. Sep Purif Technol.
(sulfolane+benzene+n-hexane), (N-formylmorpholine+ben- 2017;179:225-235.
zene+n-hexane), and (sulfolane+N-formylmorpholine+ben-
zene+n-hexane) at temperatures ranging from (298.15 to
318.15)K: experimental results and correlation. J Chem SU PP O R TI N G I N F O RMA TI O N
Thermodyn. 2010;42(4):466-471. Additional supporting information may be found online
27. Tian LS, Tang WC, Wu SJ. Study on extractive distillation for in the Supporting Information section at the end of this
recovering pure benzene and solvent oil. Chin Pet Process Pe article.
Technol. 2001;7:5-8.
28. Wang JY, Tian LS, Tang WC, Gui SX. Commercial application
of a combined process of extractive distillation and liquid- How to cite this article: Gao S, Tang W, Zhao M,
liquid extraction using sulfolane. Chin Pet Process Pe Technol. Qie S, Pang W, Tian L. Extractive distillation of
2002;33:19-22. benzene, toluene, and xylenes from pyrolysis
29. Tian LS, He SB, Tang WC, Lin XF. Development and applica-
gasoline using methylsulfonylethane as a
tion of an extractive distillation process for recovering benzene
from reformate. Chin Pet Process Pe Technol. 2003;34:1-5.
cosolvent. Asia-Pac J Chem Eng. 2021;16:e2609.
30. Gaile AA, Zalishchevskii GD, Erzhenkov AS, Koldobskaya LL. https://doi.org/10.1002/apj.2609
Benzene separation from the benzene fraction of reformer
naphtha by extractive rectification with N-methylpyrrolidone-
sulfolane mixtures. Russ J Appl Chem. 2008;81:1296-1302.