chen2016.pdf

chemical engineering research and design 1 1 1 ( 2 0 1 6 ) 184–195
Contents lists available at ScienceDirect
Chemical Engineering Research and Design
journal homepage: www.elsevier.com/locate/cherd
Comparison of heteroazeotropic and extractive
distillation for the dehydration of propylene glycol
methyl ether
Yi-Chun Chen, Bor-Yih Yu, Chung-Chih Hsu, I-Lung Chien∗
Department of Chemical Engineering, National Taiwan University, Taipei 106, Taiwan
a r t i c l e i n f o
Article history:
Received 3 February 2016
Received in revised form 2 May 2016
Accepted 4 May 2016
Available online 11 May 2016
Keywords:
Azeotropes
Extractive distillation
Heterogeneous azeotropic
distillation
Propylene glycol methyl ether
Process design
a b s t r a c t
Separation section often is the most energy intensive section of an industrial chemical pro-
cess, even so when there is an azeotropic mixture needed to be separated. In order to further
save energy in a particular azeotropic separation system, various heat integration methods
were proposed. However, because of more complex nature of these heat integration meth-
ods, the operation and control of such process may be hindered. Since there are various ways
which can be used to separate a particular azeotropic mixture, this paper uses an industrial
relevant propylene glycol methyl ether dehydration system as an example to demonstrate
that selecting the most effective separation method is crucially important in attempting
to save energy of the separation system. The drawback of using heterogeneous azeotropic
distillation for this separation task can easily be revealed with process understanding. A
more effective separation method via extractive distillation is proposed to significantly save
steam cost by 39.7% and reduce the total annual cost by 32.7%. Simple feed-effluent heat
exchanger can be added in the proposed extractive distillation system to further save energy
without much complication of the process design flowsheet.
© 2016 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.
1. Introduction
Separation section often is the most energy intensive section
of an industrial chemical process, even so when there is an
azeotropic mixture needed to be separated. Simple regular
distillation cannot be used to achieve complete separation. A
recent book by Luyben and Chien (2010) summarizes the feasi-
ble ways used in industry to achieve such separation. Although
for a particular azeotropic mixture, there may be more than
one way to achieve the separation task. However, significant
energy saving and reduction of the total annual cost of the
separation process can easily be obtained if the most suitable
separation method can be selected. In this paper, an industrial
relevant propylene glycol methyl ether dehydration system
will be used as an example to demonstrate the above claim.
∗
Corresponding author. Tel.: +886 2 3366 3063; fax: +886 2 3366 3040.
E-mail address: ilungchien@ntu.edu.tw (I.-L. Chien).
Propylene glycol methyl ether (PM) is widely used as a
solvent and emulsifier in industrial and consumer products
(Chiavone-Filho et al., 1993; Doan et al., 2009; Timofeeva et al.,
2014). It can also be used to react with acetic acid via ester-
ification reaction to produce propylene glycol methyl ether
acetate, which is one of the most commonly used ester with a
high industrial demand in surface coatings (Hsieh et al., 2006;
Tochigi et al., 2007; Oh et al., 2015). In the above-mentioned
usages, dehydration of PM is an important task in the sep-
aration process. PM and water formed a minimum-boiling
azeotrope. This azeotropic mixture is customarily separated in
industry via heterogeneous azeotropic distillation by adding a
light entrainer.
In this paper, the drawback of using heterogeneous
azeotropic distillation for this separation system can be
http://dx.doi.org/10.1016/j.cherd.2016.05.003
0263-8762/© 2016 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

chemical engineering research and design 1 1 1 ( 2 0 1 6 ) 184–195 185
revealed by simply drawing the material balance lines with
the help of simulation tool giving the residual curve maps,
liquid–liquid envelope and tie lines. Another design flowsheet
via extractive distillation will be proposed for this separation
task. By adding a suitable heavy entrainer into the separation
system without introducing additional azeotrope, the relative
volatility of the original mixture can be greatly enhanced, thus,
a two-column system can be designed to separate PM and
water. This ternary vapor–liquid equilibrium diagram belongs
to a so-called 1.0-1a in Serafimov’s classification (Serafimov,
1970). There are large numbers of paper in open literatures
proposed to study this 1.0-1a extractive distillation system
to separate azeotropic mixtures (e.g. Laroche et al., 1991;
Arifin and Chien, 2008; Kossack et al., 2008; Luyben, 2008;
Gutiérrez-Guerrz et al., 2009; Rodríguez-Donis et al., 2009;
Hsu et al., 2010; Shen et al., 2013; Petlyuk et al., 2015). Some
papers also described of adjusting operating pressure in the
extractive-distillation system to improve separation efficiency
(e.g. Luyben and Chien, 2010; Luo et al., 2014; You et al., 2015,
2016). However, to the best of our knowledge; none of the paper
in open literature has investigated into this separation sys-
tem. Other light or intermediate boiling entrainers may also
be added (Laroche et al., 1991), however, a heavy entrainer is
most common and will be proposed for this separation task.
2. Separation via heterogeneous azeotropic
distillation
An industrial relevant separation problem is to obtain propyl-
ene glycol methyl ether (PM) from a diluted aqueous solution.
The composition of the PM in the fresh feed stream is at
7.8 mol% with the rest to be water. The feed flow rate is
assumed to be 1000 kmol/h. The desired product specification
of PM is set to be 99.9 mol%.
For the following simulation studies conduct in this paper,
The NRTL thermodynamic model is selected to describe
the non-ideal vapor–liquid and vapor–liquid–liquid behav-
ior in the studied system and the HOC model (Hayden and
O’Connell, 1975) is used to account for non-ideality in vapor
phase. The Aspen built-in NRTL parameters for the PM–water
pair were not satisfactory to describe the VLE experimental
data as published by Tochigi et al. (2007). Therefore, parameter
regression was conducted to obtain the suitable NRTL param-
eters for this pair. It is shown in Fig. 1 that the T–xy plots at
three operating pressures (93.3 kPa, 53.3 kPa and 26.7 kPa) are
all predicted quite well using the obtained parameters. For
other component pairs which do not have any experimental
data, UNIFAC group contribution method was used to estimate
the NRTL parameters. All NRTL parameters used in this paper
are listed in Table 1.
In a particular chemical plant, isopropyl acetate (iPrAc) was
used as a light entrainer for this separation system via hetero-
geneous azeotropic distillation. As shown in Fig. 2, the design
flowsheet includes a preconcentrator column and a hetero-
geneous azeotropic distillation column. The purpose of the
preconcentrator column is to remove portions of water in
the fresh feed through column bottoms. The limiting distil-
late composition of this preconcentrator column will approach
azeotropic composition of PM–water. In practice, the actual
distillate composition is an important design variable to bal-
ance the costs associated with the preconcentrator column
and the heterogeneous azeotropic column.
Fig. 1 – T–xy plots of PM–water system at various operating pressures.

186 chemical engineering research and design 1 1 1 ( 2 0 1 6 ) 184–195
Table 1 – NRTL model parameters of the studied systems.
Comp. i PM PM iPrAc PM Water
Comp. j Water iPuAc Water Sulfolane Sulfolane
Source From plant From plant From plant UNIFAC estimated ASPEN VLE-IG
aij −1.530 0 0 0 0
aji 7.626 0 0 0 0
bij (K) 431.045 −60.437 186.759 −126.099 432.733
bji (K) −1998.746 338.092 1483.132 −270.389 333.416
cij 0.3 0.3 0.2 0.3 0.3
Comp. i PM Water PM Water
Comp. j NMP NMP DMSO DMSO
Source UNIFAC estimated ASPEN VLE-IG UNIFAC estimated ASPEN VLE-IG
aij 1.844761 7.4081 0 −1.2449
aji −0.50235 −2.2438 0 1.7524
bij (K) −949.405 −2279.96 −150.833 586.801
bji (K) 26.03436 533.8474 −211.593 −1130.22
cij 0.3 0.3 0.3 0.3
Aspen Plus NRTL: ln i =

j
xijiGji

k
xkGki
+

j
xiGji

k
xkGki

ij −

m
xmmiGmj

k
xkGkj
where : Gij = exp(−˛ijij)
ij = aij +
bij
T
˛ij = cij, ii = 0, Gii = 1
Fig. 2 – RCM, LLE, material balance lines and conceptual design ﬂowsheet via heterogeneous azeotropic distillation.

Fig. 3 – Design flowsheet via heterogeneous azeotropic distillation for the studied separation system.
This distillate stream is fed into a heterogeneous
azeotropic column. By adding isopropyl acetate (iPrAc) into
the system, an additional azeotrope is formed that has the
minimum azeotropic temperature of 76.46 ◦C in the system.
This azeotrope is also inside the VLLE envelope. With the aid
of iPrAc into the column through an organic reflux stream, the
bottom composition can approach pure PM composition and
the column top vapor can approach the lowest temperature
of the ternary system, which is the azeotropic composition of
iPrAc-water.
This azeotrope is heterogeneous. Thus by condensing and
cooling the top vapor to 40 ◦C and feed the condensate to
a decanter, natural liquid–liquid separation into an organic
phase and an aqueous phase will occur. The organic phase,
which is rich in iPrAc, is recycled back to the heterogeneous
azeotropic column. The composition of the aqueous phase is
pure enough and drawn out of the system as aqueous product.
A small entrainer makeup stream is needed to feed into the
system at the decanter to balance the small iPrAc loss from the
outlet streams. The material balance lines for the two columns
(shown in Fig. 2) clearly explain the conceptual design of this
separation flowsheet.
With process understanding of the separation principle, it
can be revealed by the lever rule that the top vapor flow rate
of the heterogeneous azeotropic column will be much greater
than that of its bottom flow rate. This large top vapor flow rate
can be estimated by the large ratio of (b/a) in Fig. 2. In order
to generate such a large top vapor flow rate, the reboiler duty
must be quite high resulting in potentially in-efficient separa-
tion by using this separation method. The design flowsheet of
this plant in Fig. 3 showed that the reboiler duty of second col-
umn is at 11,870 kW. This heterogeneous azeotropic column
will be used as a base case for comparison with an improved
design proposed later in this paper. From Fig. 3, it is shown
that the feed composition to the heterogeneous azeotropic
column was determined to be 19.17 mol% PM to balance the
costs associated with the two columns. The feed flow rate is
at 403.8 kmol/h. This feed condition will be set to be exactly
the same in the following investigations for fair comparison.
It is possible to further save the energy requirement of
preconcentrator column by heat integration methods, e.g. via
multi-effect distillation. However, this is beyond the scope of
this paper and will not be investigated further.
3. Separation via extractive distillation
system
Even without conducting any rigorous simulation, it is demon-
strated that the potential drawback of using heterogeneous
azeotorpic distillation for the separation of this PM–water
mixture can be revealed with ternary diagram of the stud-
ied system including the information of residual curve maps,
liquid–liquid envelope with tie lines, and material balance
lines. Another separation method for azeotropic mixture will
be investigated in this section by adding a heavy entrainer into
the system via extractive distillation.
3.1. Entrainer selection
The most important decision in this design is to choose a most
effective entrainer in order to greatly enhance the relative
volatility of water (lighter component) vs. PM (heavier compo-
nent). Chapter 10 in Luyben and Chien (2010) used isovolatility
and equivolatility curves to compare candidate entrainers. An
even easier plot to generate using simulation tool is shown in
Fig. 4. In this plot, the starting composition is set as the feed
K(Water)/K(PM)
Entrainer/Azeotrope Feed Ratio
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Relative
Volatility
0
2
4
6
8
10
12
14
DMSO
NMP
Sulfolane
Fig. 4 – Enhancement of relative volatility for the candidate
heavy entrainers at 1 atm.

composition of heterogeneous azeotropic column in the last
section. By gradually adding entrainer into the system, we can
perform a bubble-point calculation in a flasher to obtain the
enhancement of relative volatility at various feed ratio. The
potential candidates are screened from some common heavy
entrainers in Gmehling and Möllmann (1998). Fig. 4 demon-
strates that sulfolane (SUL) or N-methyl-2-pyrrolidone (NMP)
are much more effective entrainers as compared to dimethyl
sulfoxide (DMSO) for this separation system.
This figure can also be used to determine the distillate
products in the two-column system. By observing the rela-
tive volatility of water over PM to be greater than one in Fig. 4,
one would know that water should be the distillate product of
extractive distillation column and PM should be the distillate
product of recovery column. If PM were the distillate product
of extractive distillation column, the relative volatility of water
over PM should become less than one by adding the candidate
entrainer.
Other factors that affect the entrainer performance are the
y–x or T–xy plots. The y–x or T–xy plots for the water-entrainer
pair can be used to determine the ease of separation in the
rectifying section of extractive distillation column. The y–x or
T–xy plots for the PM-entrainer pair can be used to determine
if the separation in the entrainer recovery column is easy or
not. Of course, thermally stable, nontoxic, low price, and other
favorable physical properties should also be considered in the
selection of proper entrainer.
The T–xy plots of the two potential entrainers (SUL and
NMP) versus water or PM are shown in Fig. 5. It is observed
that the separations of water-entrainer or PM-entrainer are
all fairly easy. One problem by using sulfolane as entrainer
is that the temperature in the reboiler of entrainer recovery
column is very high if this column operated at atmospheric
pressure. If the assumption of the high-pressure steam is at
41 barg (254 ◦C), this column has to be operated at vacuum
pressure. In Fig. 5, it can be seen that by setting the top pres-
sure of this column at 0.06 atm, cooling water can still be used
in the condenser and high-pressure steam can be used in the
reboiler. As for the NMP case, there is no problem to operate
this column at atmospheric pressure.
3.2. Extractive distillation system using sulfolane as
entrainer (P2 at vacuum)
Since sulfolane is the entrainer which can enhance the relative
volatility the most, we will investigate its performance first.
In an extractive distillation system, there are two columns.
The first column is an extractive distillation column with
both fresh feed and entrainer feeding into this column. The
feed location of the entrainer is normally higher because the
entrainer is heavy, while that of the fresh feed location is lower.
In the column section between two feed locations (called
extractive section), the concentration of entrainer is high so
that water will go up the column because of the enhancement
of its relative volatility while that of PM will go down with the
entrainer. In the second entrainer recovery column, the PM
and entrainer are separated to obtain top PM product and bot-
tom entrainer. This bottom stream can be recycled back to the
extractive distillation column. A cooler is normally placed in
this recycle loop so that the entrainer feed temperature is 10 K
below the top temperature of extractive distillation column
as suggested by Doherty and Malone (2001). A small entrainer
makeup stream is also necessary to balance the small loss of
entrainer through two outlet streams.
Table 2 – Basis of economics and equipment sizing.
Column diameter (D): Aspen tray sizing
Column length (L): NT trays with 2-ft spacing plus 20% extra length
Column and other vessel (D and L are in meters)
Capital cost = 17,640(D)1.066
(L)0.802
Condensers (area in m2
)
Heat-transfer coefficient = 0.852 kW/K m2
Differential temperature = reflux-drum temperature − 315 K
Capital cost = 7296(area)0.65
Reboilers (area in m2
)
Differential temperature = steam temperature − base
temperature
Heat exchangers (area in m2
)
Differential temperature = LMTD of (inlet and outlet temperature
differences)
Water coolers (area in m2
)
Differential temperature = LMTD of (inlet or outlet
temperature − 315 K)
Energy cost
HP steam = $9.88/GJ (41 barg, 254 ◦
C)
MP steam = $8.22/GJ (10 barg, 184 ◦
C)
LP steam = $7.78/GJ (5 barg, 160 ◦
C)
Cooling water = $0.354/GJ
Electricity = $16.9/GJ
TAC = (capital cost/payback period) + energy cost
Payback period = 3 year
There are quite a few design variables that need to be
determined in this flowsheet. The most important one is the
entrainer feed flow rate (or the ratio between entrainer and
fresh feed). When this entrainer feed flow rate is larger, the
enhancement of the relative volatility is greater which can
cut the costs associated with the extractive distillation col-
umn. However, larger entrainer feed flow rate will increase the
loading of second column, hence to increase the costs associ-
ated with the entrainer recovery column. The proposed design
flowsheet will be established next to find a low total annual
cost of this two-column system.
The TAC of the overall system includes the annualized cap-
ital costs and the operating costs. The annualized capital costs
include: costs of two columns, two reboilers, two condensers
and a cooler. The payback period was assumed to be 3 years
in the calculation. The operating costs include: steam for two
reboilers, cooling water for two condensers and a cooler, and
also a very small entrainer makeup cost. Note that to be more
accurate in the TAC calculation, the costs associated with the
vacuum system to generate operating pressure at a vacuum
pressure should also be included. However, by following the
procedure outlined on page 589–590 of Seider, et al.’s book
(Seider et al., 2009), the annualized capital cost of the liquid-
ring vacuum pump and also the annual electricity cost of this
vacuum system can be neglected in comparison with other
main terms in the TAC calculation. The bases for calculating all
terms in TAC were mainly adopted from Chapter 5 of Luyben’s
book (Luyben, 2011) and are summarized in Table 2.
Sequential iterative procedure was used to obtain the pro-
posed design flowsheet with the procedure outlined in Fig. 6.
At each entrainer flow rate, the other design variables for

Fig. 5 – T–xy plots of entrainer-PM and entrainer-water.
the extractive distillation column are determined ﬁrst. These
design variables include: total stages of extractive distilla-
tion column (NT1), fresh feed location (NF1), and entrainer
feed location (NFE). After these three design variables are
determined at each entrainer ﬂow rate, the design vari-
ables associated with the entrainer recovery column can be
determined next. Since the boiling temperature of sulfolane
at 1 atm is quite high (at 560 K), even high-pressure steam in
Table 2 cannot be use in the reboiler of this column. Thus, it
is preferable to operate this column at vacuum pressure. This
top pressure is an important design variable for this column.
If it is too low, the diameter of the column will increase which

Fig. 6 – Flowchart of iterative sequential optimization
procedure.
raises the capital cost of this column. However, if this top pres-
sure is set to be too high, the temperature difference between
the high-pressure steam and the reboiler temperature will be
small which will increase capital cost of the reboiler. The two
other design variables are the total stages and the feed location
of the entrainer recovery column.
As for top pressure of the extractive distillation column, it
was conveniently set to be at atmospheric pressure for easier
column operation. However, if a plot similar to Fig. 4 reveals
that the enhancement of relative volatility is much greater
when the operating pressure is reduced, a more suitable vac-
uum pressure should be an extra design variable need to be
investigated to minimize the total annual cost. Note that the
lowest pressure in the investigation should be the one which
can still result in boiling-point temperature of water to be
90 100 110 120 130 140 150
2300
2320
2340
2360
2380
2400
Overall TAC
FE flowrate (kmol/h)
Overall
TAC
(1000$)
Fig. 7 – Summarized TAC plots at various entrainer feed
rates (sulfolane as entrainer).
greater than 320 K so that cheap cooling water can be used
in the condenser. In this study, lowering the pressure did not
see significant enhancement of relative volatility of water over
PM by adding sulfolane into the system.
By collecting the optimized results at each entrainer flow
rate, a summarized plot of overall TAC which includes the
TAC for the two columns and also the cost associated with
the cooler and the makeup cost can be seen in Fig. 7. The
optimized entrainer flow rate is at 110 kmol/h, this gives the
entrainer-to-feed ratio of only at 0.27. At this entrainer flow
rate, the detail plots associated with the extractive distilla-
tion column can be seen in Fig. 8. It is shown that total stages
should be 26 with the two feed locations at: NF1 = 17 and
NFE = 3. Also at this entrainer flow rate the detail plots associ-
ated with the entrainer recovery column can be seen in Fig. 9.
Note that the optimized top pressure of this column should be
at 0.06 atm, with total stages to be 9 and feed location at 4th
stage. The final optimized design flowsheet is shown in Fig. 10.
All itemized TAC terms are given in Table 3. The economic
comparison of this design flowsheet to the heterogeneous
azeotropic column in Fig. 3 will be discussed in Section 3.4.
The liquid composition profiles of the two columns are
shown in Fig. 11. It is observed that high concentration of sul-
folane in the extractive section is necessary to prevent PM to
go up the column. The separation of water and sulfolane in
the rectifying section of extractive distillation column is rather
easy, thus, only two stages are required. The separation of PM
and sulfolane is also easy with only nine stages needed for the
second column.
3.3. Extractive distillation system using NMP as
entrainer (P2 = 1 atm)
From the plot of relative volatility for various entrainers
in Fig. 4, NMP can also enhance the relative volatility of
water versus PM to some extent. The advantage of NMP over
sulfolane is that NMP’s boiling temperature at 1 atm is consid-
erably lower (at 478 K). This means that the entrainer recovery
column of this system can be operated at atmospheric pres-
sure and still able to use high-pressure steam as heat source
of reboiler. There is no need to operate the second column
at vacuum pressure, thus avoiding the tendency of increasing
column diameter at lower operating pressure.

2 3 4 5 6
1880
1900
1920
1940
NFE
TAC1
(1000$)
FE=110kmol/h, NT1=24
NF1=15
NF1=16
NF1=17
2 3 4 5 6
1880
1900
1920
1940
NFE
TAC1
(1000$)
NF1=15
NF1=16
NF1=17
NF1=18
2 3 4 5 6
1880
1900
1920
1940
NFE
TAC1
(1000$)
NF1=16
NF1=17
NF1=18
2 3 4 5 6
1880
1900
1920
1940
NFE
TAC1
(1000$)
NF1=16
NF1=17
NF1=18
NF1=19
23 24 25 26 27 28
1885
1886
1887
1888
1889
1890
FE=110kmol/h
NT1
TAC1
(1000$)
Fig. 8 – Detailed plots of the extractive distillation column at FE = 110 kmol/h (sulfolane as entrainer).
Table 3 – Itemized TAC terms of various design cases.
Heterogeneous
azeotropic
distillation column
(sulfolane as solvent)
P2 = 0.06 atm
(NMP as solvent)
P = 1 atm
C1 C2 C1 C2
Annualized column
capital cost
1000$/year 101.35 123.12 43.04 226.14 64.44
Annualized reboiler
capital cost
1000$/year 111.46 170.26 20.63 339.24 41.04
Annualized condenser
capital cost
1000$/year 100.91 44.68 73.39 54.61 19.04
Steam cost 1000$/year 2912.28 1487.06 268.19 1880.11 556.26
Condenser cooling
water cost
1000$/year 129.13 58.73 14.75 73.69 19.58
Reboiler duty MW 11.87 5.74 0.86 7.25 1.79
Annualized decanter
capital cost
1000$/year 94.98 – –
Annualized cooler
capital cost
1000$/year – 9.86 13.94
Cooler cooling water
cost
1000$/year – 8.32 15.03
Makeup cost 1000$/year 0.02 0.01 0.09
Total reboiler duty MW 11.87 6.60 9.04
Total steam cost 1000$/year 2912.28 (0%) 1755.24 (−39.7%) 2436.37
Annualized total capital
cost
1000$/year 408.71 (0%) 484.98 (+18.6%) 758.45
TAC 1000$/year 3450.14 (0%) 2322.02 (−32.7%) 3303.12

Fig. 9 – Detailed plots of the entrainer recovery column at FE = 110 kmol/h (sulfolane as entrainer).
The same iterative sequential optimization procedure as
outlined in Fig. 6 was used for this system. The resulting opti-
mized design flowsheet by using NMP as entrainer can be seen
in Fig. 12. Notice that the optimized entrainer flow rate has to
be increased from 110 kmol/h to 170 kmol/h. Even with this
higher entrainer flow rate, the total stages and the stages of
the extractive section are much greater as compared to the
result in Fig. 10. This is because of not enough enhancement
of relative volatility resulting in requiring larger stages to meet
stringent purity specifications. All itemized TAC terms of this
system are also given in Table 3. It is quite obvious that this
system cannot compete with the design flowsheet in Fig. 10.
3.4. Comparison of separation using extractive
distillation (with SUL as entrainer) vs. heterogeneous
azeotropic distillation (with iPrAc as entrainer)
In this section, the alternative design flowsheets in Figs. 3 and
10 will be compared. As explained in Section 2, the draw-
back of separation via heterogeneous azeotropic distillation
Fig. 10 – Proposed design flowsheet of this studied system via extractive distillation (sulfolane as entrainer).

0 5 10 15 20 25
0.0
0.2
0.4
0.6
0.8
1.0
0 2 4 6 8 10
0.0
0.2
0.4
0.6
0.8
1.0
C2 Composition Profile
C1 Composition Profile
Liquid
Mole
Fraction
(-)
C1 Stages
PM
Water
Sul
Liquid
Mole
Fraction
(-)
C2 Stages
PM
Sul
Fig. 11 – Liquid composition profiles of the two columns in
Fig. 10.
for this separation system can be revealed by simply drawing
the material balance lines with the help of simulation tool
giving the residual curve maps, liquid–liquid envelope and
tie lines. The results from rigorous simulations confirmed the
shortcomings of heterogeneous azeotropic distillation.
Although in Fig. 3, only one column was shown to obtain
the main product of PM at 99.9 mol%. However, the purity of
aqueous outlet product (99.36 mol%) was lower than that of
the water product in Figs. 10 and 11 (99.9 mol%). Thus, another
column should be needed to further purify this aqueous out-
let product for fair comparison. Even at this lower water
purity in the heterogeneous azeotropic distillation column of
Fig. 3, the reboiler duty (11,870.4 kW) is much greater than the
total reboiler duty in Fig. 10 (5736.53 + 860.74 = 6597.3 kW). By
adjusting the price difference of different steam grades, the
steam cost of the extractive distillation system using sulfolane
as entrainer is still much lower than that of the heterogeneous
azeotropic distillation system. The details of the itemized TAC
of all the alternative separation systems for this application
can be found in Table 3.
From Table 3, it is shown that steam cost is the dominant
one in all the itemized TAC terms. The proposed design of
extractive distillation system in Fig. 10 can reduce the steam
cost by 39.7% as compared to the heterogeneous azeotorpic
distillation column in Fig. 3. Although the annualized capital
cost of the proposed design is a little higher than the heteroge-
neous azeotropic distillation column, the TAC of the propose
design is still significantly lower (32.7%) than that of the base
case design in Fig. 3. In this analysis, the payback period of
the capital cost was assumed to be 3 years. If a longer pay-
back period is assumed, the reduction of the TAC can even be
greater.
3.5. Proposed design using sulfolane as entrainer with
simple further heat integration
It is illustrated that by just selecting a more appropriate
separation method, the steam cost and the TAC of this PM
dehydration process can significantly be reduced without
introducing any complication in the design configurations via
process intensification technology. In this section, a simple
heat integration method is applied to the proposed extractive
distillation system in Fig. 10.
Since the bottom temperature of entrainer recovery col-
umn is high, a feed-effluent heat exchanger can be added to
transform the energy in this hot stream to fresh feed. In this
way, the reboiler duty of extractive distillation column can be
reduced. This added heat exchanger is designed so that the
minimum approach temperature is at 10 K. Since the temper-
ature of this entrainer recycle stream is not cooled down to
90 ◦C as originally designed, another small cooler can fulfill
this task. The design flowsheet by just adding a feed-effluent
heat exchanger can be seen in Fig. 13.
Applying this simply heat integration concept, the total
steam cost can be reduced from $1.755 × 106 in Fig. 10 to
$1.598 × 106 using the flowsheet in Fig. 13. With the calcula-
tion formula of installed capital cost in Table 2, the additional
capital investment of this extra heat exchanger is only at
Fig. 12 – Design flowsheet via extractive distillation using NMP as entrainer.

Fig. 13 – Proposed design with simple heat integration (sulfolane as entrainer).
$4.89 × 104. This means that the payback period of this addi-
tional capital investment is only 0.31 year. Also, this simple
heat integration configuration will not lose any important
control degree-of-freedom in the system. With the broad
investigation of the control strategy of extractive distillation
system in open literature, there would not be any foreseeable
problem in maintaining the two product purities in the face of
feed flow rate and feed composition disturbances.
Another energy-saving design is to thermally-coupled the
two-column system (in Fig. 10). In order to combine the two
reboilers, a vapor sidedraw from the second column needs to
be fed to the bottom of extractive distillation column in order
to provide vapor traffic in this column. However in our pro-
posed system as in Fig. 10, an expensive compressor needs
to be installed so that vapor sidedraw can then be fed from a
vacuum pressure to a higher pressure. Since it is quite obvious
that the alternative design by thermally-coupled this two-
column system will not be economical, we did not investigate
this energy-saving option any further.
4. Conclusions
In this paper, an industrial relevant separation process of
dehydrating propylene glycol methyl ether was investigated.
The common way for this separation system is to utilize a
light entrainer (in this case, isopropyl acetate) via heteroge-
neous azeotropic distillation. Even without conducting any
rigorous simulation, it is demonstrated in this paper that the
potential drawback of using this separation method can be
revealed with ternary diagram of the studied system including
the information of residual curve maps, liquid–liquid envelope
with tie lines, and material balance lines.
By using a different separation method via extractive distil-
lation, it is demonstrated that significant reduction in steam
cost of 39.7% and TAC cut by 32.7% can be realized by the pro-
posed design in Fig. 10. Sulfolane was found to be the effective
heavy entrainer to greatly enhance the relative volatility of
water versus PM. A simple further heat integration method
can be applied to the proposed extractive distillation system to
further cut the steam cost by 8.9% without introducing much
complication into the design flowsheet.
This investigation is appropriate as in the conceptual
design stage of a particular separation process. To make sure
about the ternary vapor–liquid equilibrium behavior of this
separation system for the enhancement of relative volatil-
ity, further thermodynamic experimental evaluation of this
PM–water–sulfolane ternary mixture should be conducted so
that detail design of this separation process can be followed.
Acknowledgement
The research funding from the Ministry of Science and Tech-
nology of R.O.C. under grant no. MOST 104-2221-E-002-187 is
greatly appreciated.
References
Arifin, S., Chien, I.L., 2008. Design and control of an isopropy
alcohol dehydration process via extractive distillation using
dimethyl sulfoxide as an entrainer. Ind. Eng. Chem. Res. 47,
790–803.
Chiavone-Filho, O., Proust, P., Rasmussen, P., 1993. Vapor–liquid
equilibria for glycol ether + water systems. J. Chem. Eng. Data
38, 128–131.
Doan, H.D., Weli, A., Wu, J., 2009. A combined photocatalytic and
electrochemical treatment of wasterwater containing
propylene glycol methyl ether and metal ions. Chem. Eng. J.
151, 51–58.
Doherty, M.F., Malone, M.F., 2001. Conceptual Design of
Distillations. McGraw-Hill, New York, pp. 183–256.
Gmehling, J., Möllmann, C., 1998. Synthesis of distillation
processes using thermodynamic models and the Dortmund
data bank. Ind. Eng. Chem. Res. 37, 3112–3123.
Gutiérrez-Guerrz, R., Segovia-Hernández, J.G., Hernández, S.,
2009. Reducing energy consumption and CO2 emissions in
extractive distillation. Chem. Eng. Res. Des. 87, 145–152.
Hayden, J.G., O’Connell, J.P., 1975. A generalized method for
predicting second virial coefficients. Ind. Eng. Chem. Process
Des. Dev. 14, 209–216.
Hsieh, C.T., Lee, M.J., Lin, H.M., 2006. Multiphase equilibria for
mixtures containing acetic acid, water, propylene glycol
monomethyl ether, and proylene glycol methyl ether acetate.
Ind. Eng. Chem. Res. 45, 2123–2130.
Hsu, K.Y., Hsiao, Y.C., Chien, I.L., 2010. Design and control of
dimethyl carbonate-methanol separation via extractive

distillation in the dimethyl carbonate reactive-distillation
process. Ind. Eng. Chem. Res. 49, 735–749.
Kossack, S., Kramer, K., Gani, R., Marquardt, W., 2008. A
systematic synthesis framework for extractive distillation
processes. Chem. Eng. Res. Des. 86, 781–792.
Laroche, L., Bekiaris, N., Andersen, H.W., Morari, M., 1991.
Homogeneous azeotropic distillation. Comparing entrainers.
Can. J. Chem. Eng. 69, 1302–1319.
Luo, H., Liang, K., Li, W., Xia, M., Xu, C., 2014. Comparison of
pressure-swing distillation and extractive distillation method
for isopropyl alcohol/diisopropyl ether separation. Ind. Eng.
Chem. Res. 53, 15167–15182.
Luyben, W.L., 2008. Comparison of extractive distillation and
pressure-swing distillation for acetone-methanol separation.
Ind. Eng. Chem. Res. 47, 2696–2707.
Luyben, W.L., Chien, I.L., 2010. Design and Control of Distillation
Systems for Separating Azeotropes. Wiley, Hoboken, NJ.
Luyben, W.L., 2011. Principles and Case Studies of Simultaneous
Design. Wiley, Hoboken, NJ.
Oh, J., Agrawal, G., Sreedhar, B., Donaldson, M.E., Schultz, A.K.,
Frank, T.C., Bommarius, A.S., Kawajiri, Y., 2015. Conversion
improvement for catalytic synthesis of propylene glycol
methyl ether acetate by reactive chromatography:
experiments and parameter estimation. Chem. Eng. J. 259,
397–409.
Petlyuk, F., Danilov, R., Burger, J., 2015. A novel method for the
search and identification of feasible splits of extractive
distillations in ternary mixtures. Chem. Eng. Res. Des. 99,
132–148.
Rodríguez-Donis, I., Gerbaud, V., Joulia, X., 2009. Thermodynamic
insights on the feasibility of homogeneous batch extractive
distillation. 1. Azeotropic mixtures with heavy entrainer. Ind.
Eng. Chem. Res. 48, 3544–3559.
Seider, W.D., Seader, J.D., Lewin, D.R., Widagdo, S., 2009. Product
and Process Design Principles, 3rd ed. Wiley, Hoboken, NJ.
Serafimov, L.A., 1970. The azeotropic rule and the classification of
multicomponent mixtures. VII. Diagrams for ternary
mixtures. Russ. J. Phys. Chem. 44, 567–571.
Shen, W., Benyounes, H., Gerbaud, V., 2013. Extension of
thermodynamic insights on batch extractive distillation to
continuous operation. 1. Azeotropic mixtures with a heavy
entrainer. Ind. Eng. Chem. Res. 52, 4606–4622.
Timofeeva, M.N., Panchenko, V.N., Matrosova, M.M., Andreev,
A.S., Tsybulya, S.V., Gil, A., Vicente, M.A., 2014. Factors
affecting the catalytic performance of Zr, Al-pillared clays in
the synthesis of propylene glycol methyl ether. Ind. Eng.
Chem. Res. 53, 13565–13574.
Tochigi, K., Takahara, H., Shiga, Y., Kawase, Y., 2007. Isobaric
vapor–liquid equilibria for water + propylene glycol
monomethyl ether (PGMA), water + proplyleneglycol
monomethyl ether acetate (PGMEA), and PGME + PGMEA and
reduced pressures. Fluid Phase Equilib. 260, 65–69.
You, X.Q., Rodríguez-Donis, I., Gerbaud, V., 2015. Improved design
and efficiency of the extractive distillation process for
acetone-methanol with water. Ind. Eng. Chem. Res. 54,
491–501.
You, X.Q., Rodríguez-Donis, I., Gerbaud, V., 2016. Reducing
process cost and CO2 emissions for extractive distillation by
double-effect heat integration and mechanical heat pump.
Appl. Energy 166, 128–140.

chen2016.pdf

Recommended

Recommended

More Related Content

Similar to chen2016.pdf

Similar to chen2016.pdf (20)

Recently uploaded

Recently uploaded (20)

chen2016.pdf