Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Separation On Aromatics and Non-Aromatics by Extractive Distillation With NMP

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

http://www.paper.edu.

cn
Research Paper

Journal of Chemical Engineering of Japan, Vol. 36, No. 1, pp. 2024, 2003

Separation on Aromatics and Non-aromatics by Extractive


Distillation with NMP
BIAOHUA CHEN, ZHIGANG LEI AND JIANWEI LI
The Key Laboratory of Science and Technology of Controllable
Chemical Reactions, Ministry of Education,
Beijing University of Chemical Technology, Beijing 100029, China

Keywords: Extractive Distillation, NMP, Aromatics, Non-Aromatics, Separation


Vapor-liquid-equilibrium (VLE) at normal pressure for the system formed by n-heptane, benzene and
N-methyl-pyrrolidone (NMP) are investigated. The results show that NMP is a good extractive solvent
that reverses the volatility of the binary system of n-heptane + benzene. A UNIFAC model is selected and
proved that the group interaction parameters are reliable. A modified extractive distillation (MED) process, in which a little of water is added to lower the boiling point of reboilers in aromatics recovery column, is proposed on the basis of the original extractive distillation (OED) process. The process simulation shows that the proposed process raises the yield ratio of benzene from 94.2% to 98.2%. Since the
modification is not complicated, the MED process have a lasting value in industry.

Introduction
In industry the methods for separating aromatics
and non-aromatics are liquid-liquid extraction and extractive distillation (Rawat and Gulati, 1981;
Wardencki and Tameesh, 1981; Berg, 1983). Liquidliquid extraction as an old method is almost replaced
by extractive distillation in most cases. For instance,
in the past the method of separating aromatics from
cracking oil and reformed naphthas was usually liquid-liquid extraction. But today extractive distillation
is widely used for this separation. The extractive distillation process is simple and can deal with large
amount of feedstock. But it requires the addition of a
third component, the solvent, that modifies the relative
volatilities of the different components in the mixture.
So it is evident that selection of a suitable solvent is
very important to ensure an effective and economical
design of extractive distillation.
In this work it is assumed that n-heptane and benzene are taken on as the representation of aromatics
and non-aromatics respectively because aromatics and
non-aromatics are mainly composed of them and the
separation mechanism is consistent. However, it is
found from some researchers (Liu et al., 1995; Zheng
et al., 1997; Blanco et al., 2000) that the solvents N,Ndimethylformamide (DMF), N-formylmorpholine
(NFM) are suitable for the purpose of separating aromatics and paraffin. But we know that DMF is toxic to
some degrees and does harm to human health, which
Received on March 18, 2002. Correspondence concerning this
article should be addressed to B. Chen (E-mail address:
chenbh@mal.buct.edu.cn).

20

doesnt conform to the requirement of current ecological chemical engineering. It is a pity for NMF that its
price is very expensive in chemical markets. Therefore,
it is very reluctant for many plants to employ these
two kinds of solvents. In this work a new solvent, Nmethyl-2-pyrrolidone (NMP) is selected as the separating agent. NMP has many outstanding advantages,
i.e. no toxicity, no harm and facility of ecological treatment. Thus, we investigate the separation of n-heptane
and benzene by extractive distillation with NMP.
1.

Experiment Section

The chemicals used in this work were purchased


from the Beijing Chemical Regents Shop, Beijing,
PRC, and all of the analytical purity in concentration
is above 99.5%.
A typical recycling vapor-liquid equilibrium cell
reported by Lei and Zhou (2000), having a volume of
100 ml and stirred by magnetism force, was utilized to
measure the vapor-liquid-equilibrium (VLE) of the nheptane(1)/benzene(2) system and n-heptane(1)/
benzene(2)/NMP(3) system at normal pressure. NHeptane/benzene and NMP were blended with the feed/
solvent volume ratio (i.e. solvent ratio) 1:2 and 1:3
respectively. In the cell the mixture with about 50 ml
was heated to boil at normal pressure. One hour later,
the desired phase equilibrium was achieved. At that
time, small samples in the vapor and liquid phases were
removed by a transfer pipet with 1 ml in order to make
the equilibrium not to be destroyed. Each sample was
analyzed by local 2305 GC, using a flame ionization
detector (FID). Air is supplied as the compressed gas
and hydrogen as the carrier gas. The capillary column
Copyright 2003 The Society of Chemical Engineers, Japan

http://www.paper.edu.cn
Table 1

No.

Fig. 1

The experimental flow sheet of extractive distillation process with two columns: 1, extractive distillation column; 2, solvent recovery column

was at 160C and the detector at 160C. The data were


dealt with by a SC1100 workstation. In terms of the
peak area of the samples the component compositions
were deduced.
On the basis of the VLE experiment, the experimental flow sheet of extractive distillation process with
two columns, extractive distillation column and solvent
recovery column, has been established in the laboratory and illustrated in Fig. 1 where components A and
B to be separated are respectively obtained from the
top of two columns, and the solvent S is recovered in
the solvent recovery column and recycled. The extractive distillation column was composed of three sections, rectifying section 25 mm (diameter) 400 mm
(height), stripping section 25 mm (diameter) 400 mm
(height) and scrubbing section 25 mm (diameter) 200
mm (height). The solvent recovery column was composed of two sections, rectifying section 25 mm (diameter) 400 mm (height) and stripping section 25
mm (diameter) 400 mm (height). The two columns
were all packed by a type of ring-shape packing with
the size of 3 mm (width) 3 mm (height). Since the
extractive distillation column plays an important role
in the separation process, its theoretical plates are determined by use of the system of n-heptane and
methylcyclohexane at infinite reflux, having 11 theoretical plates.
2.

Experiment Result

The experimental data at normal pressure (101.33


kPa) for different solvent ratios are listed in Table 1.
The xy curves are illustrated in Fig. 2 on a solvent
free basis. It can be seen from Fig. 2 that NMP can
apparently change the relative volatility of n-heptane
and benzene with the result that the equilibrium curves
are far away from the diagonal when the feed/solvent
ratios are 1:2 or 1:3. Therefore, NMP is a good separating agent. Since the fluidity of electron cloud is
different for n-heptane and benzene, the interaction
force between benzene and NMP is larger than that of

The experimental data at normal pressure for


different solvent ratios
T/K

x1

y1

353.25
355.50
357.70
359.75
361.70
363.50
365.30
366.95
368.55
370.05
370.85
371.55

0.0
0.125
0.243
0.355
0.461
0.562
0.658
0.750
0.837
0.920
0.961
1.0

0.0
0.104
0.188
0.263
0.338
0.415
0.506
0.606
0.720
0.850
0.921
1.0

The feed/solvent volume ratio 1:2


1
433.80
0.0
2
435.40
0.080
3
436.85
0.155
4
439.65
0.303
5
442.00
0.434
6
443.45
0.518
7
444.75
0.593
8
446.90
0.724
9
449.25
0.879
10
451.00
1.0

0.0
0.158
0.268
0.510
0.622
0.707
0.724
0.921
0.975
1.0

The feed/solvent volume ratio 1:3


1
444.40
0.0
2
445.10
0.042
3
445.70
0.078
4
447.05
0.161
5
448.85
0.281
6
451.20
0.441
7
453.50
0.613
8
455.80
0.796
9
457.35
0.926
10
458.15
1.0

0.0
0.091
0.197
0.159
0.532
0.755
0.880
0.971
0.991
1.0

No solvent
1
2
3
4
5
6
7
8
9
10
11
12

n-heptane and NMP. As a result, n-heptane is brought


out a light component and benzene a heavy component, although the boiling point of n-heptane is higher
than that of benzene. That is, n-heptane would be obtained as the overhead product in the extractive distillation column, being benzene and the solvent, NMP,
the bottom product.
The results of extractive distillation column at
different operation conditions are listed in Table 2,
where the feeding concentration is n-heptane 44.5 wt%
and benzene 55.5 wt%, and the reflux ratio is just 0.5
to ensure the high solvent concentration along the column. Table 2 manifests that NMP can effectively separate n-heptane and benzene, and results in a high purity of benzene with a concentration of above 99.5 wt%
at least. Therefore it is reasonable to select NMP as a
solvent.
21

http://www.paper.edu.cn
Table 2

Solvent ratio

1.7
2
3

Experiment results of extractive distillation column

Weight fraction of product

Extractive distillation column

n-Heptane

Benzene

Tto p [K]

Tb o t [K]

0.986
0.992
0.998

0.960
0.958
0.992

370.7
371.2
371.7

378.2
387.7
421.2

Fig. 3

The modified extractive distillation process for


separating benzene and n-heptane by extractive distillation: 1, extractive distillation column; 2, aromatics recovery column; 3, solvent recovery column

In general, for separating a two-component system with close boiling points, the extractive distillation process with two columns, i.e. extractive distillation column and solvent recovery column, is used.
However, the solvent NMP has a disadvantage, that is,
a high boiling point of 477.2 K. It indicates that a water
stream with a higher temperature is needed to heat NMP
to reach its boiling point at normal pressure. That means
much energy has to be consumed. If the solvent
recovery column is operated at below normal pressure,
the degree of vacuum pressure must not be too high
due to the restriction for temperature of condensing
water at the top of extractive distillation column. In
the end, NMP at the bottom of the solvent recovery
column often exists containing a little of benzene in
order to decrease reboiler temperature. Accordingly,
the yield ratio of benzene is generally low. In addition,
because benzene is easy to volatilize, some benzene
may be drawn into the atmosphere by a vacuum pump
in the case of operation at vacuum pressure, which does
not conform to the requirement of ecological chemical

engineering and should be avoided. So extractive


distillation process with two columns must be
improved.
It is known that when a little of water is added
into the solvent NMP, the boiling point of the mixture
will decrease greatly. By virtue of this character, one
column (i.e. aromatics recovery column) is added on
the basis of the original extractive distillation (OED)
process. A modified extractive distillation (MED) process with three columns is diagrammed in Fig. 3.
Some water is gone into the lower section of aromatics recovery column. This column is operated at
normal pressure with only 413.2 K on the bottom, and
formerly the temperature will be higher to become
477.2 K without water. The mixture of NMP and water
on the bottom enters into the solvent recovery column
which is at vacuum pressure. In the solvent recovery
column the pure NMP is obtained on the bottom and
recycled into the extractive distillation column.
To simulate the process shown in Fig. 3, the accurate VLE model should be selected. The UNIFAC
model (Reid et al., 1987; Lei et al., 2002) is generally
used for the extractive distillation process and adopted
in this work. For the studied system of n-heptane(1)/
benzene(2)/NMP(3), the UNIFAC group volume parameters Rk, area parameters Qk and interaction parameters a mn are listed in Tables 3 and 4 respectively
(Jorgensen et al., 1979; Gmehling, 1982; Macedo et

22

JOURNAL OF CHEMICAL ENGINEERING OF JAPAN

Fig. 2

3.

VLE curves of n-heptane/benzene/NMP at different solvent ratios: , feed/solvent volume ratio 1:3;
, feed/solvent volume ratio 1:2; , no solvent

Separation Process Design

http://www.paper.edu.cn

Table 3

The volume and area parameters Rk, Q k


of the UNIFAC group for n-heptane(1)/
benzene(2)/NMP(3) system

NMP
CH2
CH3
ACH
H2 O

Rk

Qk

3.9810
0.6744
0.9011
0.5313
0.9200

3.200
0.540
0.848
0.400
1.400

Table 5

No.

1
2
3
4
5
6
7
8

Table 4
m

D2 x D2
Fx F

VOL. 36 NO. 1 2003

n
NMP

NMP
H2 O
CH2
ACH

H2 O

0.00
452.20
220.30
30.04

832.20
0.00
1318.00
903.80

CH2

ACH

13.89
300.00
0.00
11.12

23.88
362.30
61.13
0.00

Comparison of the yield ratio of benzene between the OED and MED
processes
D2 [kgh 1 ]

[%]

x D2 [wt%]

OED

MED

OED

MED

OED

MED

56
55
54
53
52
51
50
49

56
55
54
53
52
51
50
49

0.9316
0.9471
0.9631
0.9796
0.9963
0.9997
0.9999
0.9999

0.9749
0.9921
0.9999
0.9999
0.9999
0.9999
0.9999
0.9999

94.9
94.7
94.6
94.4
94.2
92.7
90.9
89.1

99.3
99.2
98.2
96.4
94.5
92.7
90.9
89.1

al., 1983; Tiegs et al., 1987).


In order to ensure accurate simulation, we use the
Herrington method to verify the consistency of thermodynamics by MS Excel 97 software that is convenient for this purpose. By this method it is proved that
this model is consistent. Moreover, the accuracy of
UNIFAC group interaction parameters for predicting
VLE is farther investigated. By comparison with the
VLE data, it is shown that the calculated values using
the UNIFAC model and the experimental values given
in Table 1 are in good agreement. So the UNIFAC
model is reliable and can be used for process
simulation.
The OED process with two columns, i.e. extractive distillation column and solvent recovery column,
and the MED process with three columns, i.e. extractive distillation column with 30 theoretical plates, aromatics recovery column with 15 theoretical plates and
solvent recovery column with 10 theoretical plates, are
respectively simulated to compare the yield ratio of
benzene under the same operation conditions. In the
process simulation, the feeding mixture are composed
of 45.0 wt% n-heptane and 55.0 wt% benzene with a
flow rate 100 kgh 1, and the feed/solvent volume ratio
is 1:3. The comparison of yield ratio is listed in Table 5 where is defined as

The UNIFAC group interaction parameters for the


system n-heptane(1)/benzene(2)/NMP(3)

(1)

Here x D2 and xF represent the weight fraction of benzene at the top of aromatics recovery column and the
feeding mixture respectively.
When the MED process just reaches the benzene
concentration of above 99.5%, it is only 96.3% in the
OED process. At the same time the yield ratio of the
MED process is 98.2% while the OED process only
94.6%. Therefore, the MED process is more advantageous than the OED process not only in purity but also
in the yield ratio of benzene. However, the MED
process contributes to more energy and investment
costs than the OED process because one column
(aromatics recovery column) is added in the MED
process.
Conclusion
In this work the solvent NMP is selected as the
separating agent for separating aromatics and non-aromatics by the extractive distillation process. An
important reason to choose NMP is that this solvent
can meet the requirement of current ecological
chemical engineering. The experiments of VLE and
extractive distillation process have been done, and it
is shown that NMP is effective to separate aromatics
and non-aromatics. The UNIFAC model is selected, and
verified to be reliable and can be used for process
simulation.
The OED process with two columns has many disadvantages such as low aromatics yield ratio and eas23

http://www.paper.edu.cn

ing to pollute the environment and so on. A new MED


process is put forward and compared with the OED
process. The result is that both the purity and the yield
ratio of benzene in the MED process are apparently
greater than those in the OED process under the same
conditions, but the energy and investment costs of the
MED process are more expensive than those of the
OED process. Because the modification of the MED
process is not so complicated and only one column is
added, it is convenient to be carried out in industry,
which had a lasting value in industry.
Nomenclature
amn
= interaction parameter of the UNIFAC group m and
group n
[K]
D
= flow rate of the product at the top of the column
[kgh1]
F
= flow rate of the feed
[kgh1]
Qk
= area parameter of the UNIFAC group
[]
= volume parameter of the UNIFAC group
[]
Rk
T
= temperature
[K]
x
= molar fraction in the liquid phase with solvent free
[]
y
= molar fraction in the vapor phase with solvent free
[]

benzene yield ratio

<Subscript>
k
=
m
=
n
=

the UNIFAC group


the UNIFAC group
the UNIFAC group

1
2
3

n-heptane
benzene
NMP

24

=
=
=

[]

Literature Cited
Berg, L.; Separation of Benzene and Toluene from Close Boiling
Nonaromatics by Extractive Distillation, AIChE J., 29, 961
966 (1983)
Blanco, B., M. T. Sanz, S. Beltran, J. L. Cabezas and J. Coca;
Vapor-Liquid Equilibria for the Ternary System Benzene
+ n-Heptane + N,N-Dimethylformamide at 101.33 kPa, Fluid
Phase Equilibria, 175, 117124 (2000)
Gmehling, J; Vapor-Liquid Equilibria by UNIFAC Group Contribution. Revision and Extension 2, Ind. Eng. Process Des. Dev.,
21, 118127 (1982)
Jorgensen, S. S., B. Kolbe, J. Gmehling and P. Rasmussen; VaporLiquid Equilibria by UNIFAC Group Contribution. Revision
and Extension, Ind. Eng. Process Des. Dev., 18, 714722
(1979)
Lei, Z. G. and R. Q. Zhou; Influence of Solvent with Salts on the
Vapor-Liquid Equilibrium of Alcohol-Water Systems, Fine
Chemicals, 17, 307309 (2000)
Lei, Z. G., R. Q. Zhou and Z. T. Duan; Separating 1-Butene and
1,3-Butadiene with DMF and DMF with Salt by Extractive Distillation, J. Chem. Eng. Japan, 35, 211216 (2002)
Liu, B., Y. Zheng and W. Zhao; Relative Volatility of Benzene and
Thiophene in Aqueous DMF Solvent, Journal of Chemical
Engineering of Chinese Universities, 9, 259262 (1995)
Macedo, E. A., U. Weidlich, J. Gmehling and P. Rasmussen; VaporLiquid Equilibria by UNIFAC Group Contribution. Revision
and Extension 3, Ind. Eng. Process Des. Dev., 22, 676678
(1983)
Rawat, B. S. and I. B. Gulati; Studies on the Extraction of Aromatics with Sulpholane and its Combination with Thiodiglycol,
J. Chem. Tech. Biotechnol., 31, 2532 (1981)
Reid, R. C., J. M. Prausnitz and T. K. Sherwood; The Properties of
Gases and Liquids, pp. 5256, McGraw-Hill, New York, USA
(1987)
Tiegs, D., J. Gmehling and P. Rasmussen; Vapor-Liquid Equilibrium by UNIFAC Group Contribution 4, Ind. Eng. Chem. Res.,
26, 159161 (1987)
Wardencki, W. and A. H. H. Tameesh; Evaluation of Extraction
Properties of NMP and DOH Mixtures by Gas Chromatography, J. Chem. Tech. Biotechnol., 31, 8692 (1981)
Zheng, Y., R. Cai and W. Zhao; Separation of Alkane from
Hydrorefining Raffinate Using NFM with Extractive Distillation, Fuel and Chemical Process, 28, 9092 (1997)

You might also like