Comparison of Heterogeneous Azeotropic Distillation and Extractive

Computers and Chemical Engineering 100 (2017) 27–37

Contents lists available at ScienceDirect

Computers and Chemical Engineering

journal homepage: www.elsevier.com/locate/compchemeng

Comparison of heterogeneous azeotropic distillation and extractive

distillation methods for ternary azeotrope ethanol/toluene/water separation

∗

Lei Zhao, Xinyu Lyu, Wencheng Wang, Jun Shan, Tao Qiu

Jiangsu Key Laboratory of Advanced Catalytic Materials and Technology, School of Petrochemical Engineering, Changzhou University, No.1, Gehu Road,

Wujin District, Changzhou, Jiangsu 213164, China

a r t i c l e i n f o a b s t r a c t

Article history: Ethanol and toluene are important organic solvents in chemical industry. Ethanol, toluene and water

Received 31 October 2016

present a ternary minimum-boiling azeotrope and three binary azeotropes. In this article, two methods

Received in revised form 1 February 2017

are studied to separate the ternary azeotrope: heterogeneous azeotropic distillation using the toluene,

Accepted 4 February 2017

which is in the system itself, as entrainer and extractive distillation with glycerol as the only one solvent.

Available online 6 February 2017

In the heterogeneous azeotropic distillation, the pressures of the three columns are changed respectively

to achieve heat integration. In the extractive distillation, only one solvent and two columns are used

Keywords:

to separate the ternary mixture. Experiment shows that the partially heat-integrated heterogeneous

Ternary azeotrope

azeotropic distillation reduces the energy cost and total annual cost by 27.7% and 13.4% respectively

Heterogeneous azeotropic distillation

compared with the conventional heterogeneous azeotropic distillation. Unexpectedly, the extractive dis-

Extractive distillation

Ethanol/toluene/water tillation can save 18.8% and 39.3% in the energy cost and total annual cost respectively compared with

the partially heat-integrated heterogeneous azeotropic distillation.

1. Introduction rities. This special HAD method is applied to separate azeotrope

or close-boiling mixtures using the entrainer, which is in the sys-

Ethanol and toluene are widely used in chemical, pharmaceuti- tem itself, to generate a lower-boiling azeotrope with the one or

cal, dye and other ﬁelds as good solvents. Ethanol can deliquesce two components in the mixture. The minimum-boiling azeotrope

easily and absorb moisture from the air quickly because of the pres- produces two immiscible liquid phases in the decanter. Aque-

ence of hydrogen bond. Since ethanol, toluene and water present ous phase and organic phase both ﬂow back to the column with

three binary azeotropes and one ternary azeotrope, ordinary dis- optimum reﬂux ratios and the rest are separated and fed into

tillation methods can not separate the mixture effectively. Some different columns to obtain pure products. HAD for azeotropic mix-

other kinds of distillation methods should be studied to separate ture is expected to cross the distillation boundary (Doherty and

the ethanol/toluene/water mixture. Malone, 2001; Widagdo and Seider, 1996). Kiss used HAD method

Many methods can be used to separate azeotropic mixture, for in bioethanol dehydration with n-pentane as entrainer (Kiss et al.,

example, pressure-swing distillation (Yu et al., 2012), azeotropic 2012). Luyben also used benzene as the entrainer in the separation

distillation (Abu-Eishah and Luyben, 1985), extractive distillation of ethanol/water azeotrope by HAD (Luyben, 2006a, 2006b).

(Meirelles et al., 1992), adsorption (Garg and Ausikaitis, 1983), per- ED is generally applied to separate binary azeotrope and other

vaporation using membrane (Fleming, 1992) and some other new close-boiling mixtures with the relative volatility of light compo-

separation techniques. In this article, heterogeneous azeotropic dis- nents and heavy components below 1.1 (Seader and Henley, 1998;

tillation (HAD) and extractive distillation (ED) methods are studied Errico et al., 2013a, 2013b; Kossack et al., 2008; Lei et al., 2003;

to separate the ethanol/toluene/water mixture. Liang et al., 2014). The solvent added in the mixture improves

It is innovative to use toluene, which is in the system itself, as the relative volatility of light components and heavy components

entrainer for the HAD, which will not bring in any other impu- because the solvent interacts differently with the light components

and heavy components in the mixture. When the relative volatility

is improved, the light components are acquired in the top of the

∗ extractive column (EC) while the heavy components and solvent

Corresponding author.

are acquired from the bottom of the EC for subsequent separation.

E-mail address: [email protected] (T. Qiu).

http://dx.doi.org/10.1016/j.compchemeng.2017.02.007

28 L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37

Fig. 1. Process ﬂow sheet of HAD.

Table 1

Azeotropic Temperature and Composition of the System at 1 atm.

Temp (K) Type No. Comp. Toluene (mass%) Ethanol (mass%) Water (mass%)

347 Heterogeneous 3 47.49 42.75 9.77

357.68 Heterogeneous 2 80.09 0 19.91

351.3 Homogeneous 2 0 95.62 4.38

350.01 Homogeneous 2 31.92 68.08 0

Solvent and heavy components are separated by a solvent recov- optimization programs are used to get optimal conﬁgurations of

ery column (SRC) easily. Then, the solvent obtained in the bottom the two distillation methods.

of SRC is recycled to the EC (Kossack et al., 2008). Luyben used

dimethyl sulfoxide as solvent to separate maximum-boiling ace-

2. Simulation

tone/chloroform mixture by ED (Luyben, 2008). Figueroa used ED

method with ionic liquids as solvent to produce anhydrous ethanol

2.1. Heterogeneous azeotropic distillation

and saved energy a lot (Figueroa et al., 2012). Lei used ED method

with solid inorganic salt and ionic liquid as solvents to separate 2.1.1. Process design and ternary diagrams

In the separation of the ternary mixture with

ethanol/water mixture (Lei et al., 2014). Long retroﬁtted conven-

ethanol/toluene/water, the toluene can also serve as an entrainer

tional ED to a thermally coupled distillation scheme and reduced

and form a low-boiling ternary azeotrope with ethanol and water.

the energy cost signiﬁcantly (Duc Long and Lee, 2013).

So, it does not need to add any other entrainer, which will not

Separation of the ethanol/toluene/water mixture has visible

bring in any other impurities. The ethanol/toluene/water mixture

economic benefit, social benefit and environmental benefit. There

presents three binary azeotropes and one ternary azeotrope. One

are a lot of papers discussing the separation of binary azeotrope

ternary azeotrope (ethanol/toluene/water) is satisfactory because

via HAD or ED. However, the articles on the separation of ternary

it is heterogeneous and its azeotropic temperature (347 K) is

azeotrope using HAD and ED are quite limited. So, it is neces-

the lowest in the system. Another three binary azeotropes are

sary to study some workable distillation methods to separate the

also formed with azeotropic temperatures of 357.68, 351.3 and

ethanol/toluene/water ternary azeotrope, with the emphasis on

350.01 K, respectively. Table 1 shows the azeotropic temperature

economic optimization.

and composition of ethanol/toluene/water mixture at 1 atm. When

In order to separate the ethanol/toluene/water ternary

the pressure is 2.18 atm or 0.33 atm, the ternary azeotrope is still

azeotrope, two special distillation methods are studied in this arti-

heterogeneous and its azeotropic temperature is still the lowest

cle: HAD and ED methods. According to the total annual cost (TAC),

in the system. The NRTL physical property model is used in the

simulation for the separation of the ternary azeotrope.

L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37 29

Table 2

Goal Functions and Necessary Parameters for Economic Evaluation.

Parameters Data

Condensers Heat transfer

coefﬁcient = 0.852 kW/K m

2 0.65

Capital cost = 7296 × (A, m )

Heat transfer area: A = (K×t)

Reboilers Heat transfer

coefﬁcient = 0.568 kW/K m

2 0.65

Capital cost = 7296 × (A, m )

Heat transfer area: A = (K×t)

Column vessel Capital

1.066 0.802

cost = 17640 × (D, m) × (L, m)

Length: H = 1.2 × 0.61 × (NT − 2)

Energy cost LP stream (433 K): $7.78 per GJ

MP stream (457 K): $8.22 per GJ

HP stream (527 K): $9.88 per GJ

Chilled-water (253.15 K): $7.89 per GJ

totalcapitalcost

TAC paybackperiod + annual energy cost

Payback period 3 years

The process ﬂow sheet of the HAD using toluene as entrainer is

shown in Fig. 1 with three columns and one decanter. The pressures

of column 2 (C2) and column 3 (C3) are both reduced to 0.33 atm to

ensure that the temperature of the condensers is about 323.75 K.

So, cooling water can still be used in the condensers. The pressure

of column 1 (C1) is raised to 2.18 atm to ensure that the tempera-

ture in the top stage of C1 is about 20 K more than the temperature

in the bottom of C2 and C3. So, heat integration can be achieved in

the HAD in the next section. The C1 has a decanter and the feed-

stock is fed into this column (Stage 1 is the top tray). High-purity

ethanol product is obtained from the bottom of C1 while the min-

imum boiling ethanol/toluene/water heterogeneous azeotrope is

obtained from the top of C1. Then, the overhead vapor stream of

C1 is cooled, and two liquid phases are separated in the decanter.

Some of both the aqueous and organic phases are reﬂuxed to the

top tray. The rest of the aqueous phase is fed into C2 and the rest

of the organic phase is fed into C3. High-purity water and toluene

products are obtained in the bottom of C2 and C3 respectively while

the top products of these two columns are combined and back to

C1.

The residue curve map (RCM) of ethanol/toluene/water mixture

is shown in Fig. 2, which is divided into three distillation areas by

the distillation boundaries. The critical regions are enlarged in the

ternary diagrams. Fig. 2a is with a pressure of 2.18 atm, the oper-

ating pressure in C1. Because the temperature in the decanter is

313.15 K, Fig. 2b is with a pressure of 0.23 atm, where the temper-

ature of the ternary azeotrope is 313.59 K (Luyben, 2006a, 2006b).

Fig. 2c is with a pressure of 0.33 atm, the operating pressure in C2

and C3. Fig. 2 also shows the azeotropic temperature and compo-

sition of ethanol/toluene/water mixture at 2.18 atm and 0.33 atm.

Distillation boundary represents the limit of the separation in a

column. The composition points of overhead product and bot-

tom product in one column must lie in the same distillation area

(Luyben, 2006a, 2006b). The ethanol product is near the top corner

in area 1. The water product is near the lower left corner in area 2.

The toluene product is near the lower right corner in area 3. This

means that three columns at least must be used for the desired

separation.

In Fig. 2a, the composition points in C1 are analyzed without

Fig. 2. Residue curve maps for ethanol/toluene/water mixture at (a) 2.18 atm, (b) the decanter. The system has four inputs: F1, FR, aqueous reﬂux

0.23 atm, (c) 0.33 atm. (AR) and organic reﬂux (OR), and it has two outputs: B1 and V. M

(F1 + FR + AR + OR), B1 and V composition points lie on the straight

line and in the same area of the RCM, so M can be separated in one

column and the ethanol product can be obtained in the bottom of

C1. Because Fig. 2a is with a pressure of 2.18 atm, where the temper-

ature of the ternary azeotrope is 368.41 K, not the 313.15 K in the

30 L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37

Fix P1=2.18 atm P2=0.33 atm P3=0.33 at m

Give NT3

Give NF3 No No

Give NT2 No

Give NF2

Is TAC minimal with Give NT1 NT3, NF3?

Give N , N

FR F1 Yes

Vary RRa, RRo, RR2, RR3 to meet

Is TAC minimal with design specifications of column C1, C2 No N ? and C3 T3 No

Yes No Is TAC minimal with NT1, NT2, NF2, NT3, NF3?

Is TAC minimal?

Yes Is TAC minimal with NT2, YES Yes NF2, NT3, NF3?

Get the optimal NT1, NT2, NT3, NF1,

N , N , N and TAC

Yes FR F2 F3 Is TAC minimal with N , T2 Over

NT3, NF3?

Fig. 3. Sequential iterative optimization procedure for HAD.

decanter, AR and OR composition points are not on the liquid–liquid ethanol/toluene/water, with the ﬂow rate 1000 kg/h. The three

equilibrium envelope. product speciﬁcations are set as follows: the ethanol product has a

In Fig. 2b, AR + F2 and OR + F3 composition points will lie on a purity of 99.9 mass%, the toluene product has a purity of 99.9 mass%

straight line and pass through V composition point. However, the and the water product has a purity of 99.5 mass% by varying reﬂux

three composition points are in different areas of the RCM. The ratio of aqueous phase (RRa) and reﬂux ratio of organic phase (RRo)

decanter permits AR + F2, OR + F3 and V composition points to cross in C1, and reﬂux ratios in C2 and C3 (RR2, RR3) at the same time. The

the distillation boundary. In addition, AR + F2 and OR + F3 composi- assumed tray pressure drop of each column is 0.0068 atm. The dis-

tion points are on the liquid–liquid equilibrium envelope. Because tillation process is simulated by Aspen Plus. Luyben’s book explains

Fig. 2b is with a pressure of 0.23 atm, V composition point is not in how to use the software expertly (Luyben, 2013).

area 1. In this article, the reﬂux drum is regarded as the ﬁrst stage

Fig. 2c shows the composition points in C2 and C3. F2 is separated and the reboiler is regarded as the last stage (Luyben, 2009; Yang

into B2 and D2. F3 is separated into B3 and D3. F2, B2 and D2 com- et al., 2013). TAC is minimized in design of the distillation to get

position points lie on the straight line and in the same area of the a better process. TAC is minimized by changing the number of

RCM, so F2 can be separated in one column and water product can stages and the feed stage, etc. to get the optimum design. Con-

be obtained in the bottom of C2. F3, B3 and D3 composition points ventionally, TAC consists of process variable costs and ﬁxed capital

lie on a straight line and in the same area of the RCM, so F3 can be investment. Because the price of the decanter, reﬂux drum, pumps,

separated in one column and toluene product can be obtained in valves and pipes is usually far less than the price of the towers

C3. This means that the three products can be obtained using three and heat exchangers, these equipments are usually not considered.

columns and one decanter. The most important equipments include column vessels, reboilers

and condensers. Table 2 shows the goal functions for the economic

evaluation (Luyben, 2006a, 2006b).

2.1.2. Process study and economic analysis

In C1, the variables contain the number of stages (NT1), fresh feed

The following data is used for the distillation process

stage (NF1) and recycled stream feed stage (NFR). In C2, the variables

simulation: the feedstock consists of 61.52/32.97/5.51 mass%

L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37 31

(a)

3.770 (b) 3.81 3.76 5 3.80

3.79 3.76 0

3.78 3.755 $/y) $/y) 5 5 3.77 3.750

3.76

TAC(10 TAC(10 3.74 5

3.75 3.74 0 3.74

3.735 3.73

68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 10 11 12 13 14 15 16 17 18 19 20 21 22 N (the col umn C ) N (the column C ) T1 1 T2 2

3.750

3.745

3.740 $/y) 5 0 3.735 AC(1 T 3.730

3.725

3.720

234 5 678 9 10 11 N (the column C )

T3 3

Fig. 4. TAC with stages of (a) C1, (b) C2 and (c) C3.

contain the number of stages (NT2) and feed stage (NF2). In C3, the and C3 with energy. Two auxiliary condensers are added in C1 to

variables contain the number of stages (NT3), and feed stage (NF3). replenish necessary condensation.

These seven variables are optimized in order to minimize TAC of The optimal conﬁgurations for the partially heat-integrated

the HAD. An economic calculation sequence is set up to optimize HAD are found in the conventional HAD. The energy cost of the

the economics when the system has so many variables. Fig. 3 shows partially heat-integrated HAD is $1.54 × 10 /y, and the TAC is

the sequential iterative optimization procedure expressly. $3.23 × 10 /y. Optimal ﬂow sheet for partially heat-integrated HAD

Fig. 4a shows that the optimal NT1 is 75 with optimum NFR with equipment data, steam data, reﬂux ratio, heat duties, internal

and NF1. Fig. 4b shows that the optimal NT2 is 15 with optimum diameter (ID) and operating pressures (P), etc. is shown in Fig. 6.

NF2. Fig. 4c shows that the optimal NT3 is 5 with optimum NF3. Comparing the two optimal HAD methods, the research indi-

The optimal economic parameters are obtained: the optimal design cates that partially heat-integrated HAD is more economical than

variables are NT1 = 75, NT2 = 15 and NT3 = 5. The energy cost of the conventional HAD. Experiment shows that the partially heat-

5 5

HAD is $2.13 × 10 /y, and the TAC is $3.73 × 10 /y. Optimal ﬂow integrated HAD offers 27.7% and 13.4% reduction in the energy cost

sheet for HAD with equipment data, steam data, reﬂux ratio, heat and TAC respectively.

duties, internal diameter (ID) and operating pressures (P), etc. is

shown in Fig. 5. 2.2. Extractive distillation

2.1.3. Partially heat-integrated heterogeneous azeotropic 2.2.1. Solvent selection

distillation More time should be spent selecting the potential solvents

Many scientists have made great efforts to increase the efﬁ- because the solvent plays a vital role in ED.

ciency of energy utilization for the distillation methods in the past The solvent should be less volatile than components in the

decades. It is generally known that the columns of the pressure- mixture. The solvent should increase the relative volatility of the

swing distillation are set at different pressures. So, the energy components in the mixture and not generate an azeotrope with

utilization efﬁciency of the pressure-swing distillation can be any component in the mixture. Also, the solvent should be readily

increased by heat integration. Different operating pressures can available and cheap.

also be applied to HAD for heat integration. Originally, two solvents should be used to increase the relative

The ﬂow sheet for partially heat-integrated HAD is similar to volatility of ethanol/toluene and ethanol/water respectively. How-

the conventional HAD mentioned above. In order to achieve heat ever, experiment shows that only one solvent is needed to increase

integration, the overhead vapor from C1 provides reboilers of C2 the relative volatility of ethanol/toluene and ethanol/water respec-

32 L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37

Fig. 5. Optimal ﬂow sheet for HAD.

Fig. 6. Optimal ﬂow sheet for partially heat-integrated HAD.

L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37 33

(a) 1.0 (b) 1.0

0.9 0.9

0.8 0.8

0.7 0.7

0.6 0.6 l) ol) 0.5 0.5 hano han 0.4 0.4 y(Et y(Et

0.3 0.3

0.2 0.2

ethanol-toluene-glycerol ethanol- water-glycerol 0.1 ethanol-toluene 0.1 ethanol -water

0.0 0.0

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

x(Ethanol) x(Ethanol)

Fig. 7. x–y diagram for (a) ethanol + toluene + glycerol, (b) ethanol + water + glycerol system.

Fig. 8. Process ﬂow sheet of ED.

Fig. 9. Effect of S and RR1 in the EC (NT1 = 45, NF1 = 26, NFE = 5) on (a) ethanol purity in D1 and (b) impurity of ethanol in B1.

34 L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37

Fig. 10. Sequential iterative optimization procedure for ED.

Table 3

Results of the Solvent Selection for ED by Aspen Plus.

solvent Tb/K ␣1,2 (ethanol/toluene) ␣1,3 (ethanol/water)

glycerol 564.05 5.19 5.82

ethylene glycol 470.45 0.296 4.9

n-propylbenzene 432.35 2.166 0.196

n-butylbenzene 455.25 2.128 0.321

tively. Four solvents are discussed in this article for the separation 2.2.2. Process design and economic analysis

of ethanol/toluene/water mixture: glycerol (Gil et al., 2012), ethy- More attention should be paid to the process design and eco-

lene glycol, n-propylbenzene and n-butylbenzene. Table 3 shows nomic analysis after glycerol is selected as the solvent. Only two

␣ ␣

their boiling point (Tb) and the relative volatility ( 1,2, 1,3 the columns will be used in the ED thanks to the immiscibility of

relative volatility of ethanol/toluene and ethanol/water with sol- toluene with water. The energy cost and ﬁxed capital investment

vent S). Glycerol is chosen as the solvent in the simulation because will be reduced signiﬁcantly because one column is cut down. This

glycerol does not bring in any other azeotropes in the mixture is infrequent in the separation of ternary azeotrope. The ED process

and can increase the relative volatility of ethanol/toluene (5.19) with two columns and one decanter is shown in Fig. 8 clearly. The

and ethanol/water (5.82) simultaneously. Fig. 7 shows that glycerol fresh stream is fed at the middle stage of the EC while the solvent is

modiﬁes the vapor-liquid equilibrium curve obviously and elimi- fed at the top stage of the EC. In the EC, glycerol increases the volatil-

nates ethanol/water and ethanol/toluene azeotrope. The effect of ity of ethanol/water and ethanol/toluene respectively and makes

glycerol allows obtaining high purity ethanol. So, only one sol- the separation easier. Glycerol, toluene and water ﬂow down and

vent will be used in the ED. This characteristic will bring in fewer outﬂow in the bottom of the EC while high purity ethanol prod-

impurities and greatly reduce the energy cost. uct is obtained from the top of the EC. The SRC has a decanter and

removes water and toluene from glycerol. This separation is very

L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37 35

(a) 2.05 (b) 2.000

2.04 1.995

2.03 1.990

2.02 1.985 ) 2.01 1.980 $/y $/y) 5 5 10 10

2.00 ( 1.975 C( TAC TA 1.99 1.970

1.98 1.965

1.97 1.960

1.96 1.955

1.2 1.3 1.4 1.5 1.6 1.7 1.8 37 38 39 40 41 42 43 44 45 46 47 N (the extractive column) solvent ratio(S/F) T1

2.04

2.03

2.02

2.01 $/y)

5 2.00

1.99 TAC(10 1.98

1.97

1.96

1.95

234 567 8 N (the solvent recov ery column)

Fig. 11. TAC with (a) S/F ratio (mass), stages of (b) the EC (c) SRC.

easy because water and toluene is much more volatile than glyc- ethanol in B1 if S could be higher. In order to acquire the 99.5 mass%

erol. The overhead vapor stream of toluene/water is cooled, and two water product in SRC, the mass ﬂow rate of ethanol in B1 must be

liquid phases are separated in the decanter because of the immisci- less than 0.316 kg/h. So, the S must exceed 1280 kg/h.

bility of toluene with water. Some of both the aqueous and organic Fig. 9a shows that an optimal RR1 makes the maximum ethanol

phases are reﬂuxed to the top tray. The rest of the aqueous phase purity with a given S. The higher the S is, the higher the ethanol

and organic phase are water and toluene products. Then, the sol- purity could be obtained. Fig. 9a shows that the purity of ethanol

vent is cooled and recycled to the EC. Finally, little pure solvent could be more than 99.9% with an optimum RR1 when S is

make-up is added to maintain the solvent ﬂow rate/fresh feed ﬂow 1280 kg/h.

rate (S/F) ratio constant. In the EC, the bottom stream is held at 0.316 kg/h ethanol and

Three product speciﬁcations are set as follows: the ethanol prod- the distillate stream is held at 99.9 mass% ethanol by varying reﬂux

uct has a purity of 99.9 mass%, the toluene product has a purity of ratio (RR1) and distillate ﬂow rate (D1). In the SRC, the bottoms

99.9 mass% and the water product has a purity of 99.5 mass%. The stream is held at 99.9 mass% glycerol by varying reﬂux ratio of

pressure of the EC is operated at 1 atm. The pressure of SRC is oper- aqueous phase (RRa) and reﬂux ratio of organic phase (RRO). The

ated at 0.05 atm to avoid the thermal degradation of glycerol due to distillate stream is held at 0.01 mass% glycerol by vary distillate

the high temperature (Gil et al., 2012). The assumed tray pressure ﬂow rate (D2).

drop of each column is 0.0068 atm. A cooler is required to cool the recycled solvent. Knight and

There are two design degrees for the EC when the number of Doherty suggested that the temperature of solvent stream should

stages (NT1), fresh feed stage (NF1), solvent feed stage (NFE) and be 5–15 K less than the temperature in the top of the EC (Knight

operating pressure (P1) are given: solvent ﬂow rate (S) and reﬂux and Doherty, 1989). So, the temperature of the solvent stream is

ratio (RR1). Much time is spent exploring the effect of S and RR1 set at 336.15 K.

on compositions of overhead stream and bottom stream before In the EC, the variables contain the number of stages (NT1), fresh

the strict simulation. The effect of S and RR1 on the compositions feed stage (NF1), and solvent feed stage (NFE). In the SRC, the vari-

of overhead stream and bottom stream in the EC with NT1 = 45, ables contain the number of stages (NT2) and feed stage (NF2). These

NF1 = 26, NFE = 5 is shown in Fig. 9. The effect of S and RR1 on the ﬁve variables are optimized to make TAC minimal for the ED. An

impurity of ethanol in B1 is shown in Fig. 9b. There will be less economic calculation sequence is set up to optimize the economics

36 L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37

Fig. 12. Optimal ﬂow sheet for ED.

Table 4

Optimal Results for Partially Heat-Integrated HAD and ED.

Parameter Partially Heat-Integrated HAD ED

C1 C2 C3 EC SRC

NT 75 15 5 41 3

NF 5 9 2 25 2

NFR/NFE 6 N/A N/A 3 N/A

RR 1.72 (RRa)1.72 (RRo) 0.37 1.00 0.27 0.053 (RRa)0.053 (RRo)

Reboiler heat (kW) 735.78 245.78 36.05 346.85 108.31

Total capital cost (10 $) 5.07 2.14

Total energy cost (10 $/y) 1.54 1.25

TAC (10 $/y) 3.23 1.96

when the system has so many variables. Fig. 10 shows the sequen- compared with the partially heat-integrated HAD. This is mainly

tial iterative optimization procedure expressly. QR1 is the reboiler because only two columns and one solvent are used to separate

heat for the EC. QR2 is the reboiler heat for the SRC. the ternary mixture in the ED. Also, evaporation capacity in ED is

Fig. 11a shows that the optimal S/F ratio is 1.5. Fig. 11b shows less than evaporation capacity in HAD signiﬁcantly. It means that

that the optimal NT1 is 41 with optimum NFE and NF1. Fig. 11c the ﬁxed capital investment and energy cost will be reduced a lot

shows that the optimal NT2 is 3 with optimum NF2. The optimal respectively.

economic parameters are obtained: the optimal design variables

are NT1 = 41 and NT2 = 3. The energy cost of the ED is $1.25 × 10 /y,

and the TAC is $1.96 × 10 /y. Optimal ﬂow sheet for ED with equip-

4. Conclusion

ment data, steam data, reﬂux ratio, heat duties, internal diameter

(ID) and operating pressures (P), etc. is shown in Fig. 12. It is impor-

Two methods for the separation of ethanol/toluene/water

tant to observe that the temperature of the condenser in the SRC is

azeotrope are investigated: partially heat-integrated heteroge-

281.41 K. More expensive chilled-water should be used to condense

neous azeotropic distillation and extractive distillation. The two

the overhead vapor stream.

distillation methods are optimized by a proposed optimization

method with TAC as the objective function.

3. Process comparison In the partially heat-integrated heterogeneous azeotropic dis-

tillation, the optimal number of stages is 75 for C1, 15 for C2 and

In this chapter, we will compare partially heat-integrated HAD 5 for C3. The pressures of C1, C2 and C3 are 2.18 atm, 0.33 atm and

with ED according to the economics. The optimal processes of the 0.33 atm respectively.

two kinds of special distillation are acquired by the same standard Glycerol is chosen as the suitable solvent for the extractive dis-

and evaluation procedures. Table 4 shows the optimal result of par- tillation. The optimal number of stages is 41 for the EC and 3 for the

tially heat-integrated HAD and ED. It is worth mentioning that the SRC. The pressures of EC and SRC are 1 atm and 0.05 atm respec-

ED has reduced energy cost and TAC by 18.8% and 39.3% respectively tively.

L. Zhao et al. / Computers and Chemical Engineering 100 (2017) 27–37 37

The partially heat-integrated heterogeneous azeotropic dis- Gil, I.D., Gómez, J.M., Rodríguez, G., 2012. Control of an extractive distillation

process to dehydrate ethanol using glycerol as entrainer. Comput. Chem. Eng.

tillation has reduced the energy cost and TAC by 27.7% and

39, 129–142.

13.4% respectively compared with the conventional heterogeneous

Kiss, A.A., David, J., Suszwalak, P., 2012. Enhanced bioethanol dehydration by

azeotropic distillation. The extractive distillation has reduced extractive and azeotropic distillation in dividing-wall columns. Sep. Purif.

Technol. 86, 70–78.

the energy cost and TAC by 18.8% and 39.3% respectively com-

Knight, J.R., Doherty, M.F., 1989. Optimal design and synthesis of homogeneous

pared with the partially heat-integrated heterogeneous azeotropic

azeotropic distillation sequences. Ind. Eng. Chem. Res. 28, 564–572.

distillation. It is concluded that the extractive distillation is Kossack, S., Kraemer, K., Gani, R., Marquardt, W., 2008. A systematic synthesis

framework for extractive distillation processes. Chem. Eng. Res. Des. 86,

more attractive than the partially heat-integrated heterogeneous

781–792.

azeotropic distillation in terms of economics.

Lei, Z., Li, C., Chen, B., 2003. Extractive distillation: a review. Sep. Purif. Rev. 32,

121–213.

Acknowledgments Lei, Z., Xi, X., Dai, C., Zhu, J., Chen, B., 2014. Extractive distillation with the mixture

of ionic liquid and solid inorganic salt as entrainers. AIChE J. 60, 2994–3004.

Liang, K., Li, W., Luo, H., Xia, M., Xu, C., 2014. Energy-efﬁcient extractive distillation

We are thankful for assistance from the staff at Jiangsu Key Labo- process by combining preconcentration column and entrainer recovery

ratory of Advanced Catalytic Materials and Technology (Changzhou column. Ind. Eng. Chem. Res. 53, 7121–7131.

Luyben, W.L., 2006a. Distillation Economic Optimization. John Wiley & Sons, Inc,

University). This research did not receive any speciﬁc grant from

Chapter 4.

funding agencies in the public, commercial, or not-for-proﬁt sec-

Luyben, W.L., 2006b. Control of a multiunit heterogeneous azeotropic distillation

tors. process. AIChE J. 52, 623–637.

Luyben, W.L., 2008. Control of the maximum-boiling acetone/chloroform

azeotropic distillation system. Ind. Eng. Chem. Res. 47, 6140–6149.

References

Luyben, W.L., 2009. Design and control of the monoisopropylamine process. Ind.

Eng. Chem. Res. 48, 10551–10563.

Abu-Eishah, S.I., Luyben, W.L., 1985. Design and control of a two-column Luyben, W.L., 2013. Distillation Design and Control Using Aspen Simulation. John

azeotropic distillation system. Ind. Eng. Chem. Process Des. Dev. 24, 132–140. Wiley & Sons.

Doherty, M.F., Malone, M.F., 2001. Conceptual Design of Distillation Systems. Meirelles, A., Weiss, S., Herfurth, H., 1992. Ethanol dehydration by extractive

McGraw-Hill Science/Engineering/Math. distillation. J. Chem. Technol. Biotechnol. 53, 181–188.

Duc Long, N., Lee, M., 2013. Optimal retroﬁt design of extractive distillation to Seader, J.D., Henley, E.J., 1998. Separation Process Principles. John Wiley & Sons,

energy efﬁcient thermally coupled distillation scheme. AIChE J. 59, 1175–1182. New York, pp. 604.

Errico, M., Rong, B.-G., Tola, G., Spano, M., 2013a. Optimal synthesis of distillation Widagdo, S., Seider, W.D., 1996. Journal review. azeotropic distillation. AIChE J. 42,

systems for bioethanol separation. Part 1: extractive distillation with simple 96–130.

columns. Ind. Eng. Chem. Res. 52, 1612–1619. Yang, S., Wang, Y., Bai, G., Zhu, Y., 2013. Design and control of an extractive

Errico, M., Rong, B.-G., Tola, G., Spano, M., 2013b. Optimal synthesis of distillation distillation system for benzene/acetonitrile separation using dimethyl

systems for bioethanol separation. Part 2. Extractive distillation with complex sulfoxide as an entrainer. Ind. Eng. Chem. Res. 52, 13102–13112.

columns. Ind. Eng. Chem. Res. 52, 1620–1626. Yu, B., Wang, Q., Xu, C., 2012. Design and control of distillation system for

Figueroa, J., Lunelli, B.H., Maciel Filho, R., Maciel, M.W., 2012. Improvements on methylal/methanol separation. Part 2: pressure swing distillation with full

anhydrous ethanol production by extractive distillation using ionic liquid as heat integration. Ind. Eng. Chem. Res. 51, 1293–1310.

solvent. Procedia Eng. 42, 1016–1026.

Fleming, H.L., 1992. Consider membrane pervaporation. Chem. Eng. Prog. 88,

46–52.

Garg, D., Ausikaitis, J., 1983. Molecular-sieve dehydration cycle for high

water-content streams. Chem. Eng. Prog. 79, 60–65.