Korean J. Chem. Eng.

, 19(6), 996-1000 (2002)

Simulation of the Aromatic Recovery Process by Extractive Distillation

Min Su Ko, Sangyoup Na, Jungho Cho* and Hwayong Kim
School of Chemical Engineering & Institute of Chemical Processes, Seoul National University, Seoul 151-742, Korea
*Department of Chemical Engineering, Dongyang University, Youngju, Kyoungbuk 750-711, Korea
(Received 2 May 2002 accepted 23 July 2002)
Abstract The NFM (N-formylmorpholine) extractive distillation process is used for the purpose of recovering
aromatics, such as benzene, toluene, and xylene. The thermodynamic equation used in the simulation is the NRTL
equation, and NRTL interaction binary parameters were obtained by experimental data. The missing thermodynamic
properties are predicted by group contribution method. Simulation results of the extractive distillation column were
in good agreement with the operation data.
Key words: N-formylmorpholine (NFM), Extractive Distillation, Process Simulation, NRTL, Liquid-Liquid Equilibria

INTRODUCTION
There has been an ever-increasing demand for a higher purity of
aromatics as a feedstock for chemical synthesis. Many solvents
such as sulfolane [Lee and Kim, 1995, 1998], N-methylpyrrolidone
(NMP) [Muller, 1973], glycol [Symoniak et al., 1959; Taher and
Emina, 1996], and N-formylmorpholine (NFM) [Cinelli at al., 1972;
Park and Gmehling, 1989; Mohamed et al., 1995; Manoranjan et
al., 1998 ] are used to extract aromatics such as benzene, toluene
and xylene from hydrocarbon mixtures [Muller and Hoehfeld, 1963].
The NFM extractive distillation process separates aromatics from
the reformate [Morrison and Brown, 1996; Taher and Emina, 1996].
It minimizes aromatic content in gasoline and refines the extracted
aromatics, which are used as raw materials of petrochemical processes. The operating cost and the initial equipment cost for the overall plant can be reduced by the decision of its optimum value. Commercial process simulators such as PRO II (SimSci), HYSYS (Hyprotech) and ASPEN-PLUS (Aspentech) [Oonk, 1981] have been
used to determine the optimum solvent rate. However, LLE data
are not available for these mixtures.
The purpose of this work was to obtain LLE data for each binary
system at the operating temperature range of the NFM extractive
distillation process, and simulate the NFM extractive distillation
process. LLE data for the binary systems were measured at the temperature range from 298.15 K to 413.15 K, and correlated with the
NRTL [Renon and Prausnitz, 1968] model.

Fig. 1. Schematic flowsheet for simulation of NFM extractive distillation process.

products. A solvent, named NFM, is used to extract aromatics. The

NFM process is appropriate for separating aromatics by the extractive distillation process because NFM has selective affinity with
aromatics.
Injection of solvents that have good affinity with aromatics and
less affinity with non-aromatics reduces the relative volatility of
aromatics. This causes the collection of aromatics and solvents at
the bottom of extractive distillation column. At the first column,
extractive distillation occurs. Aromatics gather at the top and NFM
at the bottom in subsequent stripper column, which is connected
with the bottom of the first column. Then NFM is recycled to the
extractive distillation column.
2. Simulation Input Data
Simulation of the extractive distillation column is performed by
using the rigorous distillation model RADFRAC from the ASPEN
PLUS simulator. NFM is not built in simulator library and then the
user defined component wizard is used. The simulator required physical properties data and property parameters for non-databank components, such as normal boiling point, molecular weight, standard
liquid density, critical temperature, critical pressure, acentric factor,
Rackett parameter, vapor pressure, and heat capacity. We supplied
the missing thermodynamic properties with both methods of the
following: property estimation method to estimate the required properties data and data regression method to regress experimental data
to obtain the parameters. The critical properties (Tc and Pc) are es-

SIMULATION
1. Flowsheet for the NFM Extractive Distillation Process
Fig. 1 shows the flowsheet for the NFM process. The extractive
distillation process separates aromatics from reformate produced
by reformer. It minimizes aromatic content in gasoline and refines
extracted aromatics to be used as raw materials of petrochemical

To whom correspondence should be addressed.

E-mail: [email protected]

This paper is dedicated to Dr. Youn Yong Lee on the occasion of his
retirement from Korea Institute of Science and Technology.

996

Simulation of the Aromatic Recovery Process by Extractive Distillation

997

Table 1. Thermodynamic properties used in the simulation

Thermodynamic properties

Method using in the simulation

Critical properties
Tc
Pc

Pure vapor pressure of NFM

Binary parameters for LLE and VLE

Joback method (molecular structure)

761.1 K
49.5 bar
0.454
Experimental data [Park and Gmehling, 1989]
Databanks, estimation, and experimental data
Eight binary systems: pentane, hexane, heptane, octane,
cyclopentane, cyclohexane, benzene, and toluene with NFM

1989; Twu et al., 1994]. We used the analytical expressions developed by Ambrose and Walton (1989) [Poling et al., 2001] and the
estimated critical properties for Tc and Pc. The results are shown
in Table 1.
The thermodynamic equation used in the NFM process simulation is the NRTL equation, and the NRTL interaction binary parameters are obtained by experimental data, estimation, and databanks.

timated with modified Joback method and experimental normal

boiling point. The acentric factor is found from Eq. (1)
0
1
2 2
ln (Pc 1.01325) = f ( Tbr ) + f (T br ) + f (T br )
( )

( )

( )

(1)

where Pc is critical pressure in bars, Tbr is Tb/Tc in absolute temperature, K. Although a number of analytical expressions have been
suggested for f(0), f(1) and f(2) [Brandani, 1993; Schreiber and Pitzer,

Table 2. Binary parameters for all components in the NFM extractive distillation process

NC5
NC6
NC7
NC8
CP5
CH6
MCP
MCH
ECH
DMCH
BE
TOL
EBE
NFM

NC5

NC6

NC7

NC8

CP5

CH6

MCP

MCH

ECH

DMCH

BE

TOL

EBE

A1
B2
C3
C
C
A
A
C
C
A
A
C
D4

A
A
C
A
A
A
C
C
A
A
C
D

A
C
A
C
A
C
C
A
A
A
D

C
A
C
C
A
C
A
C
A
D

A
C
C
C
C
A
C
C
D

A
A
C
C
A
A
A
D

C
C
C
A
A
C
C

C
C
A
A
C
C

C
C
C
A
C

C
C
C
C

C
A
D

C
D

NFM

The VLE-IG and VLE-LIT are databanks for vapor-liquid applications. A : The databank VLE_IG was developed by AspenTech using binary
vapor liquid equilibrium data from the Dortmund databank. B2: The databank VLE_LIT contains binary parameters obtained from the literature.
The both of VLE-IG and VLE-LIT are used with thermodynamic models, such as Wilson, NRTL, and UNIQUAC models and with ideal-gas
model for vapor phase model. C3: R-PCES is obtained by UNIFAC model. D4: USER is obtained by regressing of experimental data.
2
NC5: Pentane, NC6: Hexane, NC7: Heptane, NC8: Octane, CP5: Cyclopentane, CH6: Cyclohexane, MCP: Methylcyclopentane, MCH:
Methylcyclohexane, DMCH: Dimethylcyclohexane, BZ: Benzene, TOL: Toluene, EBE: Ethylbenzene, NFM: N-Formylmorpholine.
3
Total number of binary systems is 91 (=14C2).
Table 3. User defined NRTL binary parameters for systems containing N-formylmorpholin
Component i
Component j

NC5
NFM

NC6
NFM

aij
aji
bij
bji
cij
dij
eij
eji

142.273778
219.392189
6157.33676
9897.63006
0.3
0.0001
20.8054802
32.27468113

4.54662813
35.2816182
1173.06664
1381.77553
0.3
0.0009
0.68991512
4.91089597

NC7
NFM

NC8
NFM

CP5
NFM

CH6
NFM

BE
NFM

TOL
NFM

9.76074132 95.8720696 52.2508155

25.9668
2.37871337
178.9185
16.019515 13.2889643 102.830391 174.840172 1.55606667 12.5085029
1359.53594 6056.27081 891.612371 6567.53477 433.899709
5225
1917.75121 1739.02684 3869.66766 7658.69462 960.692624
-2695
0.3
0.300084047 0.030018311
0.3
0.3
0.3
0.000443
0.000241
0.000536
0.000827
0.0001
0.0001
1.87500939 13.8563414 8.02588056 26.8772204 0.00011 5.00536868
2.24960645 1.90640856 15.5077994 25.853971 0.000108 1.83778013
Korean J. Chem. Eng.(Vol. 19, No. 6)

998

M. S. Ko et al.

Fig. 3. Temperature and pressure profile in extractive distillation

column.

The binary parameters have the temperature dependence function,

for NRTL model:

Fig. 2. Configuration of extractive distillation column.

The results of the existence of binary parameters for all components

in the NFM extractive distillation process are shown Table 2. NRTL
binary parameters of component i and j were calculated by means
of the NRTL equation whose parameters are reported in Table 3.

21G 21
12G 12
gE =
------- + --------------------x 1x2 -------------------RT
x 1 + x2G 21 x2 + x1G 12

(2)

G ij = exp ( ij ij )

(3)

g ij g ji
ij = -------------RT

(4)

ij = aij + bij T + eij lnT

(5)

ij = cij + dij (T 273.15 K )

(6)

Fig. 2 shows the column configuration for simulation of extractive distillation, and Table 4 shows simulation input data. For the
extractive distillation column of this process, the number of practical stages is 77 (number of stages for packed section and tray section, 34 and 45, respectively). The extractive distillation column is
composed of the packed section with structured packing, Gempack
2A of Glitsch Co., Ltd. and the tray section with sieve tray. The
height equivalent of a theoretical plate (HETP) for the packed sec-

Table 4. Feed conditions and column specifications of Radfrac extractive distillation column
Column specification
Configuration
Number of stages
Condenser
Reboiler
Valid phases
Convergence
Distillation rate [kg/hr]
Reboiler duty [kcal/hr]
Streams
5213-5
vap-feed
liq-feed
5207
5209-1
5206
Column type
2 to 31
32 to 77
November, 2002

77
Partial-vapor
Thermosyphon
Vapor-liquid-liquid
Strongly non-ideal liquid
1031
1769600
6
48
35
2
77
1
Packed-column
Tray-column

Column specification
Pressure [kg/cm2]
Total pressure drop
Section pressure drop
2 to 6
7 to 31
32 to 77
Rebolier
Outlet temperature [K]
Reboiler pressure [kg/cm2]
3-Phase
Stages
Key components
Side duties
Stage
Duty [kcal/hr]
Convergence
Algorithm
Inside loop

1.4
0.01
0.14
0.35
162
1.9
8 to 18
Feed components
76
1785000
Nonideal
Newton

Simulation of the Aromatic Recovery Process by Extractive Distillation

Fig. 4. Vapor-composition profile in extractive distillation of methylcyclohexane, benzene, with N-formylmorpholine as solvent.

999

Fig. 5. Liquid-composition profile in extractive distillation of methylcyclohexane, benzene, with N-formylmorpholine as solvent.

tion of extractive distillation is found from the Strigle method [Strigle, 1994]. The valid phases from introduction stage of lean solvent, 6, to underneath stage of liquid feed, 36, are vapor-liquid-liquid
phases. Condenser pressure is 1.4 kg/cm2, and column pressure drop
is 0.5 kg/cm2. In extractive distillation, stage-to-stage [MESH] calculations start at the bottom since the feed species, such as methylcyclohexane and benzene are the keys for second liquid phase and
the N-formylmorpholine (NFM) is a heavy nonkey.
RESULTS AND DISCUSSION
As shown in Fig. 3, the temperature changes most rapidly at the
very top, at the bottom of the column, and in the vicinity of the lean
solvent point and the feed point for the extractive distillation. The
temperature profiles inside the column, the temperatures at the reboiler and condenser, are 425.98 and 355.05 K, respectively. The

Fig. 6. Relative volatility profile in extractive distillation of methylcyclohexane, benzene, with N-formylmorpholine as solvent.

Table 5. Mass balance at the extractive distillation column

Stream number
Pentane
Hexane
Heptane
Octane
Cyclopentane
Cyclohexane
Methylcyclopentane
Methylcyclohexane
Ethylcyclohexane
Dimethylcyclohexane
Benzene
Toluene
Ethylbenzene
N-formylmorpholine
Total flow (kg/hr)
Temperature (K)

Operation data

Simulation calculation

5207

5206

5209-1

5208-1

5207

5206

5209-1

5208-1

002.5
020.6
011.0
000.0
341.8
378.6
04.2
056.5
000.0
000.7
042.9
000.7
000.0
000.0

003.0
024.7
013.3
000.1
410.1
454.3
005.0
067.8
000.0
000.0
051.5
000.9
000.0
000.0

00000.0
00000.0
00000.0
00007.2
00000.0
00000.0
00000.0
00001.5
00000.5
00002.0
13342
2392.9
00005.0
75626

000.5
004.1
002.2
000.0
068.4
075.7
000.8
011.3
000.0
000.0
008.6
000.1
000.0
000.0

002.5
020.5
011.0
000.1
342.0
378.5
004.0
048.6
000.0
000.0
045.0
007.0
000.0
000.0

0003.0
0024.6
0013.2
0000.1
0410.4
0454.2
0004.8
0058.3
0000.0
0000.0
0054.0
0008.4
0000.0
0000.0

00000.0
00000.0
00000.0
00007.3
00000.0
00000.0
00000.0
00003.1
00000.5
00002.0
13341.6
02391.6
00005.0
75626.0

000.5
004.1
002.2
000.0
068.4
075.7
000.8
009.7
000.0
000.0
009.0
001.4
000.0
000.0

859.5
0345.15

1031.00
0355.15

91377.1
00424.15

171.7
345.15

859.2
346.29

1031.0
0355.05

91377.1
00425.98

171.8
346.29

Korean J. Chem. Eng.(Vol. 19, No. 6)

1000

M. S. Ko et al.

pressure increases gradually toward bottom of the column.

Considering Figs. 4 and 5, above the feed the benzene in both
liquid and vapor dies out rapidly. Because of low relative volatility
with response to all the other components present, this component
does not enter the up flowing vapor on the stages above the feed
to any large extent. However, the benzene in both liquid and vapor
increases on the stages above lean solvent point due to vapor-liquid
phase region. While above the feed the methylcyclohexane in both
liquid and vapor increases rapidly till feed stage of lean solvent. In
1 to 5 stages occurs vapor-2nd liquid phase, 6 to 36 vapor-1st liquid2nd liquid phase and 37 to 77 stages vapor-1st liquid phase. The 1st
liquid contains NFM and methylcyclohexane and aromatics to be
highly soluble in NFM. The 2nd liquid contains paraffin and naphthene components.
Fig. 6. shows the relative volatility on the basis of the benzene.
The relative volatility is less for the benzene than for the other components, because of the difference of the mutual solubility among
aromatics, non-aromatics and NFM.
The results are presented in Table 5 for the mass balance, the total
flow rate, and temperature of each stream. Simulation results of the
extractive distillation column of the NFM process were in good
agreement with the operation data.
CONCLUSIONS
The NFM process was simulated with ASPEN PLUS 10.1 a sequential modular simulation software package. The NRTL models
were useful in simulating the NFM extractive distillation. The optimum solvent to feed ratio by mass of this process is around 3.5. The
better optimum operation condition will require a good amount of
experimental data, which are LLE data.
ACKNOWLEDGMENT
This work was supported by the BK21 project of Ministry of
Education and the National Research Laboratory (NRL) Program
of Korea Institute of Science & Technology Evaluation and Planning.
NOMENCLATURE
gE
: excess Gibbs energy [J/mol]
T
: absolute temperature [K]
Tb : normal boiling point [K]
R
: universal gas constant [J/mol-K]
x
: mole fraction [-]
gij
: parameter for interaction between components [J/mol]
Tc : critical temperature [K]
Pc : critical pressure [bar]
f(0), f(1), f(2): functions in Pitzers correlation for Pvp
Greek Letters
ij : nonrandomness parameter [-]

: acentric factor
REFERENCES
November, 2002

Alberto, C., Marcello, M., Giacomo, C., Bruno, M., Franco, M. and
Marco, S., Liquid-Liquid Equilibria of Hydrocarbons with Nformylmorpholine, J. Chem. Eng. Data, 44, 480 (1999).
Brandani, C., A Generalized Vapor Pressure Correlation for Nonpolar
Substances Based on Three-parameter Corresponding States, Ind.
Eng. Chem. Res., 32, 756 (1993),
Cinelli, E., Noe, S. and Paret, G., Extract Aromatics with FM, Hydrocarbon Process, 51, 141 (1972).
Cord, K., Petlef, T. and Gmehling, J., Measurement of Using GasLiquid Chromatography 3. Results for the Stationary Phases 10Nonadecanone, N-formylmorpholine, 1-Pentanol, m-Xylene, and
Toluene, J. Chem. Eng. Data, 34, 240 (1989).
Lee, S. and Kim, H., Liquid-liquid Equilibria for the Ternary Systems
Sulfolane+Octane+Benzene, Sulfolane+Octane+Toluene, and Sulfolane+Octane+p-Xylene, J. Chem. Eng. Data, 40, 499 (1995).
Lee, S. and Kim, H., Liquid-liquid Equilibria for the Ternary Systems
Sulfolane+Octane+Benzene, Sulfolane+Octane+Toluene, and Sulfolane+Octane+p-Xylene at Elevated Temperatures, J. Chem. Eng.
Data, 43, 358 (1998).
Manoranjan, S., Rawt, B. S., Khanna, M. K. and Nautiyal, B. R., Liquid-Lqiuid Equilibrium Studies on Toluene+Heptane+Solvent, J.
Chem. Eng. Data, 43, 422 (1998).
Mohamed, A. Q., Taher, A. A. and Mohamed, A. F., Liquid-Liquid
Equilibria in Some Binary and Ternary Mixtures with N-Formylmorpholine, J. Chem. Eng. Data, 40, 88 (1995).
Morrison, S. G. and Brown, W. E., Reforming Process For Producing
High Purity Benzene, Zeolites, 17, 403 (1996).
Muller, E. and Hoehfeld, G., Screening of Solvents for BTX Aromatics Extraction, Proceeding of the Senventh World Petroleum Congress, 4, 13 (1963).
Muller, E., Use of N-Methylpyrrolidone for Aromatics Extraction,
Chem. Ind., 518 (1973).
Oonk, H. A. J., Phase Theory: The thermodynamics of Heterogeneous
Equilibria, Elsevier Scientic Publishing Company, Amsterdam
(1981).
Park, S. J. and Gmehling, J., Isobaric Vapor-Liquid Equilibrium Data
for the Binary Systems 1,3,5-Trimethylbenzene/N-formylmorpholine and m-Xylene/N-formylmorpholine, J. Chem. Eng. Data, 34,
399 (1989).
Poling, B. E., Prausnitz, J. M. and OConnell, J. P., The Properties of
Gases and Liquids, 5th ed., McGraw-Hill, New York, NY (2001).
Renon, H. and Prausnitz, J. M., Local Compositions in Thermodynamic Excess Functions for Liquid Mixtures, AIChE J., 14, 135
(1968).
Schreiber, D. R. and Pitzer, K. S., Equation of State in the Acentric
Factor System, Fluid Phase Equilibria, 46, 113 (1989).
Strigle, R. F., Packed Tower Design and Applications-Random and
Structured Packing, 2nd ed., Gulf Pubulishing Company (1994).
Symoniak, M. F., Ganju, Y. N. and Vidueira, J. A., Plant Data for Tetra
Process, Hydrocarbon Process, 60, 139 (1981).
Taher, A. A. S. and Emina, K., Measurement and Prediction of Phase
Equilibria in the Extraction of Aromatics from Naphtha Reformate
by Tetraethyleneglcol, Fluid Phase Equilibria, 118, 271 (1996).
Twu, C. H., Coon, J. E. and Cunningham, J. R., A Generalized Vapor
Pressure Equation for Heavy Hydrocarbons, Fluid Phase Equilibria, 96, 19 (1994).