Journal Pre-proofs

Research Paper

Reactive-extractive distillation processes design for aqueous ternary azeotrope

separation

Yanlei Zhu, Hao Chen, Ning Li, Yong Liu, Rui Wang

PII: S1359-4311(25)01295-5
DOI: https://doi.org/10.1016/j.applthermaleng.2025.126703
Reference: ATE 126703

To appear in: Applied Thermal Engineering

Received Date: 2 December 2024

Revised Date: 17 April 2025
Accepted Date: 1 May 2025

Please cite this article as: Y. Zhu, H. Chen, N. Li, Y. Liu, R. Wang, Reactive-extractive distillation processes
design for aqueous ternary azeotrope separation, Applied Thermal Engineering (2025), doi: https://doi.org/
10.1016/j.applthermaleng.2025.126703

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2025 Published by Elsevier Ltd.

 Reactive-extractive Distillation Processes Design for Aqueous

Ternary Azeotrope Separation

Yanlei Zhu1, Hao Chen2, Ning Li1, Yong Liu1, Rui Wang1*

1 School of Chemistry & Chemical Engineering, Tianjin University of Technology,

Tianjin 300384, P. R. China

2 China Tianchen Engineering Corporation, Tianjin 300400, P. R. China

*Corresponding author: wrui@tju.edu.cn

ABSTRACT

In the production of ethyl tert-butyl ether using ethanol and tert-butyl alcohol,
aqueous ternary azeotrope is generated, and efficient separation methods are crucial for
resource recycling and environmental protection. This study proposes an innovative
reactive-extractive distillation process, in which ethylene oxide hydration reaction is
introduced to consume water and produce ethylene glycol as an extractant to enhance
the relative volatility between ethanol and tert-butyl alcohol, thereby facilitating
efficient separation of the azeotropic system. The accuracy of the UNIQUAC
thermodynamic model was verified and subsequently utilized to evaluate the feasibility
of the reactive-extractive distillation coupled process through ternary phase diagrams
with residual curves. A genetic algorithm was employed to optimize the process by
minimizing the total annual cost, and the overall performance was evaluated through
economic, environmental, and exergy analyses. Two thermally integrated processes
with preheat and intermediate heat exchange were designed based on the advanced
exergy analysis. Results show that the intermediate heat exchange intensified reactive-
extractive distillation process achieves significant improvements over the traditional
three-column extractive distillation, including reductions of 76.74% in total annual
costs, 72.43% in CO₂ emissions, and 29.23% in exergy destruction, thereby offering
practical solutions for process intensification and industrial application.

Keywords: Reactive-extractive distillation; advanced exergy analysis; genetic

algorithm optimization; tert-butanol/ethanol/water

1. Introduction

In the synthesis of ethyl tert-butyl ether (ETBE) via tert-butanol (TBA) and
ethanol (EtOH)[1,2], aqueous ternary azeotropic mixture containing TBA and EtOH is
generated[3]. Since EtOH and TBA are key reactants, recovering and recycling these
components is essential for improving the ETBE process's atom economy. However,
conventional distillation is ineffective due to the azeotropes. And the separation of these
components via conventional distillation is challenging[4].

1
 Special distillation techniques such as pressure-swing distillation (PSD)[5],
extractive distillation (ED)[6], and reactive extractive distillation (RED)[7], which are
extensively studied for separating complex azeotropic systems. PSD, where no new
substances are introduced, is effective for systems where the azeotropic point changes
significantly with pressure and is widely used for separating binary or ternary
azeotropic systems[8,9]. ED is applicable in various fields, including chemical,
pharmaceutical and biological[10–12]. RED has emerged as an innovative technique for
aqueous ternary azeotropic mixtures separation, in which the hydration reaction of
ethylene oxide (EO) is leveraged to consume water and simplify the separation
complexity[13], and ethylene glycol (EG) is synthesized as an extractant to increase the
relative volatility of the residual components[14,15]. It has demonstrated promising
results in the separation of acetonitrile/EtOH/H2O[16], benzene/isopropanol/water[17]
and benzene/n-propanol/water[18] et al.

The configuration of RED evolves from the triple column reactive extractive
distillation (TCRED) to the double column RED (DCRED)[19] and the latest highly
intensified dividing wall reactive extractive distillation (DW-RED)[19,20]. Wang et al.[16]
studied the utilization of DCRED for recovery of ethyl acetate (EA) and EtOH from
water. Yang et al.[21] designed the DW-RED to separate a ternary azeotropic mixture of
TBA/EtOH/H2O. Liu et al.[19] studied the separation of the ternary mixture consisting
of ethyl acetate/EtOH/H2O, findings indicated that the intensified DW-RED resulted in
a reduction of the TAC of about 8.16% with approximately 3% higher energy
consumption than DCRED, which revealed that it does not offer energy conservation
benefits over the DCRED method, contradicting common expectations. It drives us to
explore effective analytics to identify high-energy consumption locations and
accurately determine improvement possibilities to propose exergy-efficient
configurations.

Energy-based methods are generally inadequate for addressing the real

mechanisms behind environmental impacts, thermodynamic inefficiencies, and costs[22].
Exergy analysis, as a powerful thermodynamic tool based on the second law of
thermodynamics, pinpoints the areas, extent, origins of thermodynamic inefficiencies
within a system and elucidates the interplay among various system components[22–25].
The exergy analysis has been validated in key scenarios such as active and passive solar
distillation systems[26], desalination processes[27] solar tracking parabolic dish
collectors[28]. Advanced exergy analysis (AEA) uncovers the true potential of
improvement by excluding unavoidable exergy destruction caused by economic and
technical limitations, which has been applied to various chemical processes to guide
the development of more sustainable and energy-efficient methods[29–32]. AEA has been
successfully applied to the TCED process for acetonitrile/EtOH/H₂O separation,
demonstrating its ability to identify avoidable exergy destruction and improve process
efficiency, its potential for analyzing and optimizing more complex processes such as
RED remains largely unexplored[33].

Therefore, this study aims to apply AEA to a RED process design and heat
integration for the recovery of TBA and EtOH from wastewater, identifying
2
opportunities to enhance its exergy performance. Specifically, the feasibility of RED
was verified using ternary residue curve maps. Based on a literature-reported kinetic
model[34] and a verified thermodynamic model[3], DCRED are simulated, optimized and
evaluated from economic, environmental, and exergy perspectives. The genetic
algorithm (GA) was utilized determine the most effective operational parameters by
minimizing the TAC[35]. AEA was performed to identify the energy-intensive areas
within the DCRED process, and two heat-integrated configurations were substantially
proposed to significantly improve process performance.

2. Methodology

2.1 Process Simulation

The RED processes were modeled and simulated using the RadFrac module within
the Aspen Plus® V12 platform. Company downloads site gives in supplementary
material.

Thermodynamics. Selecting the appropriate thermodynamic model and accurately

determining the binary interaction parameters are crucial for simulation calculations, as
the accuracy directly impacts the calculation results. In this study, the UNIQUAC
model is applied to describe the vapor-liquid phase equilibrium behavior and the binary
interaction parameters are given in Table 1. The UNIFAC method was employed to
estimate the missing binary interaction parameters, which are essential for accurately
modeling vapor-liquid equilibriums. This approach ensures that the predicted data
aligns closely with the experimental values obtained from the NIST database within
Aspen Plus®, demonstrating a high degree of consistency, as illustrated in Figure 1.

Table 1. Regressed binary interaction parameters for the UNIQUAC model[36].

Comp i Comp j Aij Aji Bij Bji

EtOH TBA 0.0000 0.0000 -62.3572 61.6401

EtOH H2O 2.0046 -2.4936 -728.9705 756.9477

EtOH EG -8.2308 2.6876 2632.9255 -959.5647

TBA H2O 3.6099 -3.8725 -1345.2574 1209.3901

H2O EO 0.0000 0.0000 133.7554 -881.8782

H2O EG -0.6018 0.6018 120.7787 -18.6714

EO EG 0 0 54.5850 -60.2742

3
 Comp i Comp j Aij Aji Bij Bji

EO TBA 0 0 -249.6648 157.6299

EO EtOH 0 0 -296.8506 203.5622

EG TBA 0 0 -156.2496 69.4207

(a) (b)

(c) (d)

Figure 1. Comparison of experimental and predicted data on binary vapor-liquid

equilibrium at 1 atm.
Reaction Kinetics. The hydration reaction of EO can be carried out under catalytic
and non-catalytic conditions. Non-catalytic reactions are hard to control precisely the
operating conditions, resulting in low reaction selectivity and the generation of
undesired by-products diethylene glycol and triethylene glycol. Research shows that a
carbon nanotube-reinforced composite catalyst improves this reaction selectivity
effectively[34]; the reaction equation and kinetic parameters used in this article are given
in Eq. 1-2. The reaction rate units and [Ci] basis adopt kmol·m-3·s-1 and mole fraction,
respectively.

C2 H 4O( EO) + H 2O ® C2 H 6O2 ( EG ) (1)

r  mol/m3 × h  = 2.80 ´1013 exp  -71.7 / RT  xEO xWater (2)

where r is the reaction rate; x EO and xWater are the mole fraction of EO and H2O,
4
respectively; the gas constant R is equal to 8.314 kJ·(kmol·K)-1; T is the reaction
temperature, K.

2.2 Optimization

The genetic algorithm (GA)[37,38] was implemented for process optimization by

integrating MATLAB® and Aspen Plus®. The genetic algorithm optimization is based
on MATLAB®. The Aspen Plus® V12 user interface operates as an Active X
automation server, allowing MATLAB® to interact with Aspen Plus through the MAP
toolkit[39]. MATLAB® calls the GA to perform the optimization, thus indirectly
optimizing the process simulation. Some of the genetic algorithm code and company
downloads site are available at supplementary material.

Figure 2 illustrates the optimization framework of the configurations. The GA

manipulated discrete variables and continuous variables[40]. After each set of
parameters was simulated in Aspen Plus®, results were passed to MATLAB® for
objective function and constraint evaluation[41]. The GA then updated the decision
variables (x) and iterated until the stopping criteria were met. In this study, the
optimization goal is to minimize TAC. The purities of the distillates and bottom streams,
defined as constraints in the GA optimization framework (Eq. 3), are maintained at their
specified targets. The detailed calculation of TAC can be found in the supporting
material.

Figure 2. The optimization framework of the configurations.

5
 xTBA ³ 99.50 mol %
xEtOH ³ 99.50 mol % (3)
xEG ³ 99.94 mol %

Tables S2 to S4 delineate the constraints that specify the upper and lower bounds
for both discrete and continuous decision variables within the optimization framework.
The discrete decision variables encompass the number of stages in each column (NT1
and NT2), the stage at which side feed enters the SRC (NS1 and NS2), and the positions
where feeds are introduced to the columns (NS01, NFF, NEO, NW1, and NF2). The
continuous decision variables comprise the side reflux rate in the SRC (R1) and the
rates of distillate production (D1).

2.3 Process Evaluation Index

2.3.1 Economic Evaluation

The economic index known as TAC, as introduced by Douglas, encompasses both

the total capital cost (TCC) and the annual operating cost (AOC), as depicted in Eq. 4.
This index serves as a comprehensive metric for assessing the economic viability of
various processes.

TAC = TCC/payback period+AOC (4)

The TCC includes heat exchanger (condenser and reboiler) costs and the column
costs. The AOC provides steam and cooling water costs[42]. Furthermore, it is presumed
that the payback period will be three years. The comprehensive formulas for the
computation of the TAC are outlined in Table S1 of the supplementary material.

2.3.2 Environmental Analysis

To improve the manufacturing process in terms of environmental impact, CO2

emissions were closely tied to energy consumption rates, and this method can be
employed to assess the environmental performance of process design in this study[43].
Specific explanations and descriptions of the specific meanings of the variables and
symbols used in the formulas (Ep. 5 and 6) are provided in the supporting material
section, which should be consulted.

 CO 2 emissions = Q fuel / NHV + (C % / 100)a (5)

Q fuel = Q proc / l proc ´ ( h proc - 419) ´ (TFTB - T0 ) / (TFTB - Tstack ) (6)

2.3.3 Exergy Analysis

Actual processes exhibit irreversibility due to chemical reactions, heat transfer

across finite temperature gradients, mixing substances with different compositions or
6
phases, unrestricted expansion, and friction. Therefore, using exergy analysis to
identify ways to conserve energy and reduce irreversible destruction is essential. Figure
3 shows that exergy analysis includes conventional and advanced methods[44].

Figure 3. Structure flowchart of the exergy-based methods.

Conventional Exergy Analysis. Conventional exergy analysis (CEA) is

categorized into physical, chemical, kinetic, and potential energy. Kinetic and potential
energy can be ignored due to their small effects. The exergy analysis equation is as
follows (given in Ep. 7):
· · ·
E = E ph + E ch (7)

· · ·
E , E p h , and Ech represent exergy, physical exergy, and chemical exergy,
respectively. Eq. 8 [45] and Eq. 9 [46] calculate physical and chemical exergy,
respectively.

·
E ph = m ëé h - h0  + T0  s - s0  ûù (8)

·
Ech = å  xi ex 0,i + RT0 xi lnxi  (9)
m

Where h is the enthalpy of the stream and s represents the entropy. h0 , s0 represent

the reference enthalpy and reference entropy of streams, respectively. e x 0 ,i is the

standard chemical exergy of each pure substance (given in Table S5 (Supplementary

Material)) and xi represents the mole fraction of a mixture component.

·
The exergy equation for the Kth component is Eq. 10[47]: where ED,K is exergy

7
 · ·
destruction of Kth component, EF,K , EP,K are the input exergy, and output exergy of
the Kth component, respectively.

· · ·
E F ,K = E P,K + E D,K (10)

Advanced Exergy Analysis. AEA primarily involves the breakdown of exergy

destruction for each component and the entire system into two distinct categories: the
·
first category involves distinguishing between endogenous exergy destruction ( E DEN, K )
·
and exogenous exergy destruction ( E DEX, K ), thereby highlighting the interdependencies
among system components[48]; the second category separates exergy destruction into
avoidable and unavoidable components, thereby identifying the true potential for
enhancing the system's efficiency. the second category separates exergy destruction
· ·
into avoidable exergy destruction ( E DAV, K )and unavoidable exergy destruction ( E DUN, K ),
thereby identifying the true potential for enhancing the system's efficiency[49]. The
exergy destruction rate equation for the Kth component is as follow Eq. 11[50] and Eq.
12[51]:

· · ·
E D , K = E DEX, K + E DEN, K (11)

· · ·
E D , K = E DUN, K + E DAV, K (12)

To identify the exergy destruction efficiency across various components, it is

essential to examine models representing ideal, unavoidable, and real operating
conditions, as detailed in Table S6 (Supplementary Material).

The combination of avoidable/unavoidable and endogenous/exogenous exergy

destruction gives rise to four novel notions: unavoidable endogenous exergy destruction
g g
( E DUN, K, EN ), unavoidable exogenous exergy destruction ( E DUN, K, EX ), avoidable
g
endogenous exergy destruction ( E DAV, K, EN ), and avoidable exogenous exergy
g
destruction ( E DAV, K, EX ) (given in Figure 5)[52]. By introducing the value of the
unavoidable endogenous condition in Eq. 13, the values of the other three are calculated
through Eq. 14-16.
· · æ · · ö
E DUN, K, EN = E DUN, K × ç E DEN, K / E D , K ÷ (13)
è ø

· · ·
E DUN, K, EX = E DUN, K - E DUN, K, EN (14)
8
 · · ·
E DAV, K, EN = E DEN, K - E DUN, K, EN (15)

· · ·
E DAV, K, EX = E DEX, K - E DUN, K, EX (16)

The thermodynamic efficiency of a system is assessed using the exergy efficiency

h (Eq. 17), the exergy destruction ratio y (Eq. 18), and the reduction in the total
efficiency associated with thermodynamic inefficiency y * (Eq. 19)[53] where the
higher the exergy efficiency and the lower the exergy destruction ratio, the more fully
the energy is utilized[54].
· ·
h = E P,K / E F ,K (17)

· ·
y = E D , K / E F ,TOT (18)

· ·
y* = E D , K / E D ,TOT (19)

2.4 Process Configurations

Several processes were designed for separation of EtOH/TBA/H2O azeotrope,

depicted in Figure 4. The triple column extractive distillation (TCED, Figure 4a)
comprises the three extractive distillation columns (EDC1, EDC2, and EDC3)[33]. The
fresh feed and solvent are introduced into the EDC1. The bottom mixture from EDC1
is then fed to EDC2, where further separation occurs. The resulting EG/H2O mixture is
separated in EDC3, and EG is recovered from the bottom of EDC3 and recycled to
EDC1 and EDC2. Figure 4b is the DCRED, which includes ERDC and SRC. EO, H2O,
and solvent are fed into different sections of the ERDC. Within the ERDC, water reacts
with EO to form EG, which then acts in situ as a solvent to separate the mixture (A and
B). High-purity product A is recovered as the ERDC distillate, while the remaining
bottom stream, containing primarily EG and product B, is sent to the SRC for the
extraction of high-purity product B as its distillate. EG is obtained from the bottom of
SRC and partially recycled back to the ERDC, with excessive amount of solvent purged
out of the system. Figure 4(c-d) are variations of the DCRED process with integrated
heat exchange networks (DCREDHI1 and DCREDHI2) to improve energy efficiency.

9
 (a) (b)
Purge
Solvent Solvent Makeup

A B A B
Solvent
ERDC SRC
EDC1 EDC2 EDC3 Feed
Feed
Extractive Extractive Reactive-
section section Reactant extractive
section

(c) (d)
Purge Purge

A B A B
Solvent Solvent

Feed ERDC SRC Feed ERDC SRC

Reactive-
extractive
Reactant Reactant section
Reactive-
extractive
section

Figure 4. Process flow diagram for (a) TCED, (b) DCRED, (c) DCREDHI1, and
(d) DCREDHI2.

To ensure consistency, the initial feed flow rates and composition of

configurations proposed in this paper were kept identical and. are the same as that of
Shi et al[2]. The feed is set at a flow rate of 100 Kmol/h, consisting of 35 mol% EtOH,
35 mol% TBA, and 30 mol% H2O, at a temperature of 320 K and a feed pressure of 1
atm[55]. The target purity after separation and purification must not fall below 99.5
mol%. Condenser employs cooling water within the temperature range of 293.15 K to
298.15 K as its cooling agent, while the reboiler utilizes low-pressure steam as its heat
source. Atmospheric pressure operation is prioritized to minimize the heat load on the
reboiler. Additionally, the tray pressure drop is set to 0.0068 atm[36].

3. Result and Discussion

3.1 Feasibility Analysis and Baseline Process

The ternary system of TBA/EtOH/H2O forms two azeotropes at atmospheric

pressure, as given in Figure 5a: TBA/H2O and EtOH/H2O. The EtOH/H2O azeotropic
point is the unstable point, while the TBA/H2O azeotropic point is the saddle. The
boiling points of TBA/EtOH are close (given in Table 2). The resulting EG can then be
used as an extractant to effectively increase the relative volatility of TBA and EtOH, as
shown in Figure 5b.

10
 Figure 5. Ternary diagram of (a) TBA/EtOH/H2O and (b) TBA/EtOH/EG.

Table 2. Azeotropes in TBA/EtOH/H2O system.

Mole composition
Temperature
Classification
(K)
TBA EtOH H2O

351.31 Unstable node 0.000 0.900 0.100

351.46 Saddle 0.000 1.000 0.000

353.45 Saddle 0.629 0.000 0.371

355.62 Stable node 1.000 0.000 0.000

373.15 Stable node 0.000 0.000 1.000

Baseline Process. The triple column extractive distillation (shown in Figure 6)

provides a basis for comparison with the strategy proposed in this study, illustrating the
effectiveness of the intensified and heat-integrated configurations in reducing energy
consumption and costs[36].

11
 260 kmol/h 310.019 kmol/h 0.049 kmol/h
0.9999 EG 0.9999 EG Cooler 1 EG
Con1 Con2 T = 320K Con3
QC = -4132.46 W
2 2 2
4 D1 351.50 K 5 D2 355.56 K D3 373.15 K
35.17 kmol/h 35.01 kmol/h 29.85 kmol/h
EDC1 0.995 EtOH EDC2 0.995 TBA EDC3 0.9995 Water
RR1 = 0.27 RR2 = 1.30 RR3 = 2.00
21 D1 = 0.95 m 16 D2 = 0.99 m 7 D3 = 0.87 m
FF 320K QC1 = -486.72 kW QC2 = -453.69 kW QC3 = -1016.66 kW
1 atm QR1 = 1952.80 kW QR2 = 2834.86kW QR3 = 1471.01 kW
100 kmol/h
0.3 Water
0.35 TBA
0.35 EtOH 73 W1 413.62K 26 19
324.83 kmol/h W2 457.48 K W3 474.52 K
0.0924 Water 599.82 kmol/h 569.97 kmol/h
Reb1 0.8003 EG Reb2 0.0498 Water Reb3 0.0001 Water
0.1073 TBA 0.9502 EG 0.9999 EG

Figure 6. Flowsheet of the optimal TCED configuration.

3.2 Double Column Reactive Extractive Distillation

Figure 7 shows the DCRED with the comprehensive stream data and
specifications for the distillation columns. Figure 8 shows the variation of the total
annual cost (TAC) for the DCRED process with each generation in the genetic
algorithm, with optimization ceasing at the 100th generation. Figure 9 depicts the
profiles of temperature and liquid/vapor compositions, demonstrating that purity
requirements are satisfied and the configuration is stable and efficient. As a result, the
process yields 99.5 mol% purity EtOH at the top of the ERDC column (D1). Moreover,
the SRC column achieves purities of 99.5 mol% for TBA and 99.94 mol% for EG.

140.00 kmol/h Cooler T=320K

0.9994 EG QC = -901.71 kW
Con1 D1 356.50 K Con2 D2 340.40 K
1.2150 atm 0.2200 atm
35.00 kmol/h 35.01 kmol/h
2 0.995 EtOH 2 0.995 TBA
3
FF 320K 1atm ERDC SRC
100 kmol/h 13
0.35 TBA RR1 = 2.68 RR2 = 0.43
0.35EtOH 16 D1 = 1.08 D2 = 0. 77
0.3 Water QC1 = -1391.85 kW 6 QC2 = -547.63 kW
QR1 = 1660.81 kW QR2 = 666.36 kW
EO 320K
30.00 kmol/h 33

53 W1 414.15 K 10
1.5754 atm W2 453.01 K
205.00 kmol/h 0.2880 atm
Reb1 0.1704 TBA Reb2 169.99 kmol/h
0.8287 EG 0.9994 EG

Figure 7. Flowsheet of the optimal DCRED configuration.

12
 Figure 8. The optimization result of the DCRED configuration.

Figure 9. The liquid, vapor composition profiles and temperature profiles of the
optimal DCRED configuration.
13
3.2.1 Conventional Exergy Analysis of DCRED

The CEA was conducted based on the DCRED simulation findings, and the
outcomes are displayed in Table 3 and Figure 10a. The total exergy destruction of the
DCRED configuration is 2088.93 kW. The peak exergy destruction, amounting to
500.47 kW, takes place within the reboiler unit. In distillation column operations, a
pronounced thermal gradient is observed between the base stream of the column and
the cooling agent, predominantly accounting for the considerable exergy destruction
within the reboiler[56]. As for the other elements, the substantial thermal driving force,
stemming from temperature disparities, and the mass transfer driving force, arising
from chemical potential disparities within the vapor-liquid equilibrium phase, are
identified as the principal factors leading to exergy destruction within the column[56].
Figure 10a shows that for the DCRED configuration, the priority for component
optimization is ERDC, followed by Reb1, SRC, Con1, Con2, Cooler, and lastly Reb2.

Table 3. CEA results of the DCRED configuration.

·
· E P,K ·
Component E F , K (kW) E D , K (kW) h ( k ) (% ) y ( k ) (% ) y ( k ) * (% )
(kW)

ERDC 268706.71 268215.89 490.82 99.82 0.05 23.50

Con1 67588.77 67301.55 287.22 99.58 0.03 13.75

Reb1 201455.56 200955.08 500.47 99.75 0.05 23.96

SRC 207139.19 206785.36 353.83 99.83 0.04 16.94

Con2 110951.50 110760.97 190.53 99.83 0.02 9.12

Reb2 96534.38 96463.05 71.33 99.93 0.01 3.41

Cooler 47038.25 46843.53 194.72 99.59 0.02 9.32

TOT 999414.36 97325.43 2088.93 99.79 0.21 100.00

14
 Figure 10. Distribution map of exergy destruction within components for the (a)
DCRED, (b) DCREDHI1, and (c) DCREDHI2.

3.2.2 Advanced Exergy Analysis of DCRED.

Table 4 gives the AEA results for all components of DCRED. Figure 11a
g g
illustrates the E DEX, K and E DEN, K of the components within the DCRED configuration.
g
E DEX, K constitutes 29.50% of the overall exergy destruction. The top three components
g
account E DEX, K are SRC (227.13 kW), Con2 (130.53 kW), and Reb1 (126.05 kW),
respectively, which are the primary components that need paying attention to for heat
integration. 70.50% of the total exergy destruction is endogenous, suggesting that
enhancing the interplay between various pieces of equipment can increase the overall
g g
process's thermodynamic irreversibility. Figure 11b shows E DAV, K and E DUN, K of
g
components within the DCRED configuration. E DAV, K accounts for 30.52% of the total
g
exergy destruction. SRC (241.96 kW) is the largest E DAV, K component, indicating has
g
great potential for modification. E DUN, K that accounts for large components are ERDC
(458.69 kW), Reb1 (396.36 kW), and Con1 (207.73 kW), implying to boost the
efficiency of other process components to reduce destruction and pointing to significant
potential for improvement.

15
 Table 4. AEA results of the DCRED configuration.

g g g g g g g g g
Compone ED,K E DEN, K E DEX, K E DUN, K E DAV, K E DUN, K, EN E DUN, K, EX E DAV, K, EN E DAV, K, EX
nt
(kW) (kW) (kW) (kW) (kW) (kW) (kW) (kW) (kW)

490.8 490.6 458.6

ERDC 0.14 32.13 458.55 0.13 32.13 0.01
2 8 9

287.2 207.1 207.7

Con1 80.07 79.49 149.82 57.91 57.33 22.16
2 5 3

500.4 374.4 126.0 396.3 104.0

Reb1 296.55 99.84 77.87 26.22
7 2 5 9 9

353.8 126.7 227.1 111.8 241.9

SRC 40.06 71.82 86.64 155.32
3 0 3 7 6

190.5 130.5 130.5

Con2 60.00 59.97 18.89 41.09 41.11 89.44
3 3 5

Reb2 71.33 25.28 46.05 28.28 43.06 10.02 18.26 15.26 27.80

194.7 188.5 188.5

Cooler 6.19 6.19 182.53 6.00 6.00 0.20
2 3 3

2088. 1472. 616.1 1451. 637.4 1156.4

TOT 295.03 316.34 321.14
93 76 7 45 8 2

Figure 11. The exergy destruction within components of DCRED configuration.

16
 g g g g
Figure 12 presents the E DUN, K, EN , E DUN, K, EX , E DAV, K, EN and E DAV, K, EX of the
components within the DCRED configuration. Avoidable endogenous exergy
destruction raises concerns because it shows each component's potential for
g
autonomous improvement. The SRC has the largest E DAV, K, EN of 89.37 kW, followed by
Reb1 (77.87 kW) and Con1 (57.33 kW). Despite the significant exergy destruction
associated with the Red1 cooler, the portion of this destruction that can be mitigated
through intrinsic improvements is minimal. Consequently, enhancing the operational
parameters of this equipment is unlikely to yield substantial energy conservation
g
benefits. The top two components account for E DAV, K, EX are SRC (155.32 kW) and Con2
(89.44 kW). Therefore, it is imperative to concentrate efforts on enhancing the
performance of other components to mitigate their exergy destruction. SRC has the
g g g g
largest E DAV, K, EN and E DAV, K, EX . With E DAV, K, EX is higher than E DAV, K, EN for SRC in mind,
can achieve better energy efficiency results. Figure 12 shows the priority order of
DCRED configuration component optimization is SRC, followed by Con2, Reb1, Con1,
Reb2, ERDC, and lastly, Cooler. This order is different from the CEA results of the
DCRED configuration.

Figure 12. Results of splitting the exergy destruction within components of DCRED
configuration.

3.3 DCRED with Heat Integration 1

17
 The largest source of exergy destruction of DCRED process is the significant
temperature difference between the overhead vapor and the cooling agent. Optimizing
condenser and HEATX cooler operation is key to reducing this destruction in the
DCRED process. Higher exergy destruction is also seen within the column itself,
especially in the rectifying (saturated vapor feed) and stripping (saturated liquid feed)
sections. To reduce energy consumption and exergy losses, this research proposes a
modified preheat process (DCREDHI1) based on AEA, which alters the thermal
conditions of the feed entering the distillation column (Figure 13). The thermal stream
within the HEATX cooler is utilized to preheat the incoming feed stream W1, which is
initially at 414.51K, resulting in an elevated temperature for stream F2, reaching
424.81K, which better utilizes the sensible heat of the EG stream.

The reduction of TAC with the number of generations for the DCREDHI1 process
is given in Figure 14, and the optimization is terminated at 100 generations. The optimal
DCREDHI1 process for separating the ternary azeotropic mixture TBA/EtOH/H2O is
demonstrated in Figure 13. Figure 15 illustrates the profiles of liquid and vapor
compositions along with temperature data for the optimal separation process, which
indicates that purity requirements have been met and the configuration is stable and
efficient. The heat duty of the DCREDHI1 configuration is 5060.18 kW, reducing 58.86%
compared with the TCED configuration.

Cooler T=320K
QC = -635.33 kW Con1 Con2 D2 338.18 K
0.2200 atm
320K 35.01 kmol/h
140.00 kmol/h 2 2 0.995 TBA
0.9994 EG D1 356.50 K
3
1.2150 atm
FF 320K 1atm ERDC 35.00 kmol/h SRC
100 kmol/h 13
0.995 EtOH
0.35 TBA
0.35EtOH 16 F2
0.3 Water 424.81K RR2 = 0.59
6 D2 = 0.78 m
EO 320K QC2 = -642.44 kW
30.00 kmol/h 33 HEATX QR2 = 484.71 kW
QR = 246.15 kW
RR1 = 2.68
D1 = 1.08 m
QC1 = -1391.29 kW 53 10
QR1 = 1660.26 kW W1 414.15 K
1.5754 atm W2 450.32 K
204.98 kmol/h 0.2880 atm
Reb1 0.1704 TBA Reb2 169.99 kmol/h
0.8288 EG 0.9994 EG

Figure 13. Flowsheet of the optimal DCREDHI1 configuration.

18
 Figure 14. The optimization result of the DCREDHI1 configuration.

Figure 15. The liquid, vapor composition profiles and temperature profiles of the
19
 optimal DCREDHI1 configuration.

3.3.1 Conventional Exergy Analysis of DCREDHI1

The CEA of DCREDHI1 is executed, and the findings are presented in Table 5
and Figure 10b. The total exergy destruction of the DCREDHI1 configuration is
2020.93 kW. Compared with DCRED, DCREDHI1 configuration decreased by 3.26%
in exergy destruction, which demonstrates that preheating the feed stream can reduce
the exergy destruction in configuration can be reduced. Figure 10b shows that for the
DCREDHI1 configuration, the optimization priority is ERDC, followed by Reb1, SRC,
Con1, Con2, Cooler, Reb2, and lastly, HEATX.

Table 5. CEA results of the DCREDHI1 configuration.

· ·
E F ,K E P,K ·
h ( k ) (% ) y ( k ) (% ) y ( k ) * (% )
Component E D, K (kW)
(kW) (kW)

ERDC 256512.78 256018.86 493.92 99.81 0.04 24.44

Con1 62913.07 62645.84 267.23 99.58 0.02 13.22

Reb1 193878.20 193414.73 463.46 99.76 0.04 22.93

SRC 216629.31 216243.95 385.35 99.82 0.03 19.07

Con2 121014.40 120815.78 198.62 99.84 0.02 9.83

Reb2 95919.82 95846.20 73.62 99.92 0.01 3.64

Cooler 46980.12 46847.01 133.11 99.72 0.01 6.59

HEATX 129713.88 129708.27 5.61 100.00 0.00 0.28

TOT 1123561.58 1121540.65 2020.93 99.82 0.18 100.00

3.3.2 Advanced Exergy Analysis of DCREDHI1.

Table 6 presents the performance data of each component, which is derived from
the AEA associated with DCREDHI1. DCREDHI1 is engineered to reduce the exergy

20
 g g
destruction with the SRC column. Figure 16a illustrates the E DEX, K and E DEN, K of the
g
components within the DCREDHI1 configuration. E DEX, K accounts for 18.64% of the
overall exergy destruction. The SRC has been reduced to 20.72 kW, which is 90.88%
g
lower than the DCRED configuration. The top two components account for E DEX, K are
Con2 (134.59 kW) and Reb1 (88.44 kW). 81.36% of the total exergy destruction is
g g
endogenous. Figure 16b shows E DAV, K and E DUN, K of components within the DCREDHI1
g
configuration. E DAV, K accounts for 28.98% of the total exergy destruction. SRC (244.05
g g
kW) is also the largest E DAV, K component. E DUN, K that accounts for large components are
ERDC (458.43 kW), Reb1 (396.40 kW), and Con1 (207.81 kW).

Table 6. AEA results of the DCREDHI1 configuration.

g g g
g g g
Compone
g EN g UN AV
g E DUN, K, EX E DAV, K, EN E DAV, K, EX
E D,K E D,K E EX
D,K E D,K E D,K E UN , EN
D ,K
nt (kW) (kW)
(kW) (kW) (kW) (kW)
(kW) (kW) (kW)

493.9 493.8 458.4

ERDC 0.09 35.49 458.34 0.09 35.48 0.01
2 2 3

267.2 207.0 207.8

Con1 60.17 59.43 161.02 46.79 46.05 13.38
3 7 1

463.4 375.0 396.4

Reb1 88.44 67.07 320.76 75.64 54.27 12.80
6 3 0

385.3 364.6 141.3 244.0

SRC 20.72 133.71 7.60 230.93 13.12
5 4 0 5

198.6 134.5 131.8

Con2 64.03 66.74 21.51 45.22 42.51 89.37
2 9 8

Reb2 73.62 10.54 63.08 32.14 41.49 4.60 27.53 5.94 35.54

133.1 123.6 126.8

Cooler 9.51 6.28 117.77 9.06 5.83 0.45
1 1 3

21
 g g g
g g g
Compone
g EN g UN AV
g E DUN, K, EX E DAV, K, EN E DAV, K, EX
E D,K E D,K E EX
D,K E D,K E D,K E UN , EN
D ,K
nt (kW) (kW)
(kW) (kW) (kW) (kW)
(kW) (kW) (kW)

HEATX 5.61 5.59 0.02 5.61 0.00 5.59 0.02 0.00 0.00

2020. 1644. 376.6 1435. 585.6 1223.3

TOT 211.94 421.01 164.67
93 32 1 25 8 1

Figure 16. The exergy destruction within components of DCREDHI1 configuration.

g g g g
Figure 17 shows the E DUN, K, EN , E DUN, K, EX , E DAV, K, EN , and E DAV, K, EX of the components

g
within the DCREDHI1 configuration. The SRC has the largest E DAV, K, EN of 230.93 kW,
followed by Reb1 (54.27 kW), and Con2 (42.51 kW). The top two components account
g
for E DAV, K, EX are Con2 (89.37 kW) and Reb2 (35.54 kW). The result presents that the
total exergy destruction decreased with the heat exchanger applied in processes. Figure
17 shows the priority order of DCREDHI1 configuration component optimization is
SRC, followed by Con2, Reb1, Con1, Reb2, ERDC, Cooler, and lastly HEATX. This
order is different from the CEA results of the DCREDHI1 configuration.

22
 Figure 17. Results of splitting the exergy destruction within components of
DCREDHI1 configuration.

3.4 DCRED with Heat Integration 2

g
Based on the above theoretical analysis, the SRC (244.05 kW) is the largest E DAV, K
component. It is also observed that EG has a much higher boiling point than EtOH and
TBA, leading to steep temperature distributions in the SRC (given in Figure 18) due to
the wide boiling point range of the mixture. These steep temperature distributions
enable efficient intermediate heating[57]. The DCRED with heat integration 2
(DCREDHI2) configuration is designed, which is intermediate heat exchange (given in
Figure 19). The W2 stream is split into two parts: one for heating the ERDC reboiler
and the other for the SRC side reboiler. The two streams are then mixed and cooled by
the cooler. Figure 18 shows that the temperature profile of DCREDHI2 exhibits a more
homogeneous and stable distribution compared to DCREDHI1, indicating that
DCREDHI2 provides better thermodynamic efficiency.

23
 Figure 18. SRC temperature profiles in the configurations DCREDHI1 and
DCREDHI2.

Cooler T=320K
QC = -546.13 kW Con1 D1 356.50 K Con2 D2 338.19 K
1.2150 atm 0.4800 atm
320K 35.02 kmol/h 35.01 kmol/h
140.00 kmol/h 2 0.995 EtOH 2 0.995 TBA
0.9994 EG 3
FF 320K 1atm ERDC RR1 = 2.68 SRC RR2 = 0.37
100 kmol/h 13 D1 = 1.08 m D2 = 0.63 m
0.35 TBA QC1 = -1389.02 kW QC2 = -558.04 kW
0.35EtOH 16 QR1 = 1659.43 kW QR2 = 533.52 kW
0.3 Water 6
EO 320K
30.00 kmol/h 33
HEATX
QR = 111.44 kW W2-1

53 W1 414.15 K 10
1.5754 atm W2 450.32 K
204.99 kmol/h 0.5480 atm
Reb1 0.1704 TBA Reb2 169.98 kmol/h
0.8288 EG 0.9994 EG

W2-2

Figure 19. Flowsheet of the optimal DCREDHI2 configuration.

The reduction of TAC with the number of generations for the DCREDHI2 process
is given in Figure 20, and the optimization is terminated at 100 generations. The optimal
DCREDHI2 process for separating the TBA/EtOH/H2O is demonstrated in Figure 20
with comprehensive stream details and column specifications as evidenced. Figure 21
shows the liquid, vapor composition, and temperature profiles of the optimal separation
process, which indicates that purity requirements have been met and the configuration
is stable and efficient.

24
 Figure 20. The optimization result of the DCREDHI2 configuration.

Figure 21. The liquid, vapor composition profiles and temperature profiles of the
25
 optimal DCREDHI2 configuration.

3.4.1 Conventional Exergy Analysis of DCREDHI2

The CEA of DCREDHI2 is carried out, and the results are shown in Table 7 and
Figure 10c. The total exergy destruction of the DCREDHI2 configuration is 1377.95
kW. Compared with DCRED, the DCREDHI2 configuration showed a significant
exergy destruction reduction of 34.04%. The DCREDHI2 configuration has the lowest
exergy destruction compared with other configurations.

Table 7. CEA results of the DCREDHI2 configuration.

· ·
E F ,K E P,K ·
Component E D , K (kW) h ( k ) (% ) y ( k ) (%) y ( k ) * (%)
(kW) (kW)

ERDC 220028.51 219408.78 619.73 99.72 0.06 44.97

Con1 48802.07 48595.78 206.29 99.58 0.02 14.97

Reb1 245257.87 245009.14 248.73 99.90 0.02 18.05

SRC 163766.71 163657.10 109.61 99.93 0.01 7.95

Con2 40195.76 40129.54 66.22 99.84 0.01 4.81

Reb2 69598.92 69575.24 23.68 99.97 0.00 1.72

Cooler 46918.58 46834.18 84.39 99.82 0.01 6.12

HEATX 207722.31 207703.01 19.30 99.99 0.00 1.40

TOT 1042290.72 1040912.77 1377.95 99.87 0.13 100.00

3.4.2 Advanced Exergy Analysis of DCREDHI2.

The performance data of each component based on the AEA of DCREDHI2 are
g g
presented in Table 8. Figure 22a illustrates the E DEX, K and E DEN, K of the components within
g g
the DCREDHI2 configuration. E DEX, K and E DEN, K accounts for 8.07% and 91.93% of the
26
 g
overall exergy destruction. The top two components account for E DEX, K are SRC (41.72
g
kW) and Reb1 (32.49 kW). The reason of the E DEX, K growth of SRC can be attributed to
the impact of integrating a side reboiler, which modifies the temperature profile within
the column and subsequently elevates the vapor flow rate in the section situated above
g
the side reboiler. The E DEX, K of Con2 (134.59 kW) and Reb1 (88.44 kW) are reduced by
g g
93.90% and 78.36%. Figure 22b shows E DAV, K and E DUN, K of components within the
g g
DCREDHI2 configuration. E DAV, K and E DUN, K accounts for 4.42% and 95.58% of the total
g g g g
exergy destruction. Figure 23 shows the E DUN, K, EN , E DUN, K, EX , E DAV, K, EN and E DAV, K, EX of

g
components within the DCREDHI2 configuration. The Reb1 has the largest E DAV, K, EN of

g
26.50 kW. The SRC has the largest E DAV, K, EX of 5.29 kW. The destruction of components
of the DCREDHI2 configuration is mostly unavoidable and is the best process at the
present stage.

Table 8. AEA results of the DCREDHI2 configuration.

g
Compone
g g
EN g g
UN E DAV, K g g g g
E D,K E D,K E EX
D,K E D,K E DUN, K, EN E DUN, K, EX E DAV, K, EN E DAV, K, EX
nt (kW) (kW) (kW
(kW) (kW) (kW) (kW) (kW) (kW)
)

ERDC 619.73 618.87 0.86 618.44 1.29 617.58 0.86 1.29 0.00

Con1 206.29 206.29 0.00 206.29 0.00 206.29 0.00 0.00 0.00

30.4
Reb1 248.73 216.24 32.49 218.25 189.74 28.51 26.50 3.98
8

13.9
SRC 109.61 67.89 41.72 95.71 59.28 36.43 8.61 5.29
0

Con2 66.22 58.01 8.21 58.49 7.74 51.23 7.25 6.78 0.96

Reb2 23.68 10.03 13.65 21.30 2.37 9.02 12.28 1.01 1.37

27
 g
Compone
g g
EN g g
UN E DAV, K g g g g
E D,K E D,K E EX
D,K E D,K E DUN, K, EN E DUN, K, EX E DAV, K, EN E DAV, K, EX
nt (kW) (kW) (kW
(kW) (kW) (kW) (kW) (kW) (kW)
)

Cooler 84.39 74.93 9.47 80.10 4.29 71.12 8.99 3.81 0.48

HEATX 19.30 14.47 4.83 18.48 0.82 13.86 4.63 0.61 0.20

1377.9 1266.7 111.2 1317.0 60.8 1218.1

TOT 98.94 48.60 12.29
5 3 3 7 8 3

Figure 22. The exergy destruction within the component of DCREDHI2

configuration.

28
 Figure 23. Results of splitting the exergy destruction within components of
DCREDHI2 configuration.

4. Process Evaluations

Economic Evaluation. Figure 24a presented the economic evaluation results of

DCRED, DCREDHI1, and DCREDHI2 configurations, the TACs are 2.75×10^6 $/y,
1.49×10^6 $/y, and 1.77×10^6 $/y, reduction of 63.86%, 80.42%, and 76.74%
compared to the TCED configuration, respectively. The DCRED process differs from
the TCED process by the integrating reactive and extractive distillation within a single
column, significantly reducing equipment investment costs. Simultaneously, the
removal of water from the azeotropic mixture in the ERDC facilitates EG production,
substantially lowering the energy consumption of the separation process. Compared to
the DCRED process, the DCREDHI1 process achieves a greater reduction in TAC by
modifying the thermal condition and fluid dynamics within the column. This
necessitates adjusting the column diameter; the DCREDHI1 configuration features a
reduced column diameter to ensure uniform fluid distribution and effective gas-liquid
contact. The DCREDHI2 configuration, however, yields a less favorable cost reduction
than DCREDHI1, primarily due to the added side-stream intermediate heat exchanger
for the SRC, which requires a larger heat transfer area than the heat exchanger used in
DCREDHI1.

29
 Figure 24. Comparison of the (a) TACs, (b) CO2 emission, (c) total exergy
destruction, and (d) the heat duty for the TCED and modified configurations.

Environmental Evaluation. The environmental analysis results (conducted with

CO2 emission as an indicator) are given in Figure 24b. The CO2 emissions for TCED,
DCRED, DCREDHI1, and DCREDHI2 are 0.4523 kg/s, 0.1416 kg/s, 0.1395 kg/s, and
0.1247 kg/s, respectively. Compared with the TCED configuration, DCRED,
DCREDHI1, and DCREDHI2 show decreases of 68.69%, 69.16%, and 72.43% in CO2
emissions. The DCREDHI2 configuration offers the most favorable environmental
benefits.

Exergy Evaluation. The total exergy destruction of TCED, DCRED, DCREDHI1,

and DCREDHI2 configurations are 1945.98 kW, 2088.93 kW, 2020.93 kW, and
1377.95 kW, respectively (given in Figure 24c). The DCRED and DCREDHI1
configurations have increased to varying degrees, primarily due to the hot and cold flow
streams of the reboiler and the condenser between the temperature difference is larger.
The DCREDHI2 configuration has decreased by 29.23 %, which got the least amount
of exergy to lose.

As shown in Figure 24d, by comparing and analyzing the heat load of the TCED
and the modified configurations, it can be clearly observed that the modified
configurations present a significant advantage in the isentropic efficiency index. The
experimental data show a reduction of 57.97%, 58.86%, and 60.98% for DCRED,
DCREDHI1 and DCREDHI2, respectively, compared to TCED. This data provides a

30
reliable theoretical basis for the subsequent optimization of the energy consumption of
industrial-scale thermal systems.

5. Conclusions

In the ethyl tert-butyl ether synthesis process, substantial amounts of industrial

wastewater containing tert-butanol and ethanol are generated, and efficient recovery
methods are crucial. Therefore, this work investigated the design and heat integration
of an advanced exergy analysis-based reactive-extractive distillation processes for the
recovery of tert-butanol and ethanol from wastewater. The accuracy UNIQUAC model
was validated and employed to assess the feasibility of the reactive extractive
distillation via ternary residue curve maps. Advanced exergy analysis was employed to
pinpoint high-energy-consuming components within the reactive extractive distillation
configuration, such as the distillation column, condenser, and reboiler, and two heat
intensified processes with preheat and intermediate heat exchange were designed to
improve energy efficiency. A genetic algorithm was applied to optimize flowsheet
parameters for minimizing total annual costs, followed by a comprehensive evaluation
from economic, environmental, and exergy perspectives. Results indicated that the
reactive extractive distillation with intermediate heat exchange has optimal
performance, reduced heat duty by 60.98%, total annual cost by 76.74%, CO2 emissions
by 72.43%, and exergy destruction by 29.23%, compared to the traditional triple
column process. This work promotes the sustainable industrial recovery of ethanol and
tert-butanol from wastewater, and provides new method for the separation of ternary
aqueous azeotropic mixtures.

CRediT authorship contribution statement

Yanlei Zhu: Conceptualization, Methodology, Software, Investigation, Writing.

Validation. Hao Chen: Conceptualization, Methodology, review & editing, Supervision.
Ning Li: Investigation. Yong Liu: Validation, Supervision. Rui Wang: Resources,
Methodology, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or
personal relationships that could have appeared to influence the work reported in this
paper.

Acknowledgments

This research was financially supported by the Natural Science Foundation of

Tianjin (22JCQNJC00720), Tianjin University of Technology Graduate Research
Innovation Practice Project Fund (NO. YJ2359).

Acronyms

AEA = Advanced exergy analysis

AOC = Annual operating cost [$/y]

CEA = Conventional exergy analysis

31
COM = Component object model

Con = Condenser

DCRED = Double column reactive extractive distillation

DCREDHI1 = DCRED with heat integration 1

DCREDHI2 = DCRED with heat integration 2

DWC = Dividing wall column

DW-RED = Double-column reactive extractive distillation

DW-DCRED = Dividing wall-double column reactive-extractive distillation

ED = Extractive distillation

EDC=Extractive distillation column

ERC = Solvent recovery column

ERDC = Extractive and reactive distillation column

GA=Genetic algorithm

NIST = National Institute of Standards and Technology

PSD = Pressure-swing distillation

Reb = Reboiler

RED = Reactive extractive distillation

RD = Reactive distillation

RR = Reflux ratio

SRC = Solvent recovery column

TCED = Triple column extractive distillation

TCRED = Triple column reactive extractive distillation

TAC=Total annual cost [$/y]

TBA = Tert-butyl alcohol

TCC = Total capital cost [$/y]

UNIQUAC = Universal Quasi-Chemical Activity Coefficient

UNIFAC = UNIQUAC Functional-group Activity Coefficients

32
Symbols
·
E DEN, K = Endogenous exergy destruction

·
E DEX, K = Exogenous exergy destruction

·
E DAV, K = Avoidable exergy destruction

·
E DUN, K = Unavoidable exergy destruction

g
E DUN, K, EN = Unavoidable endogenous exergy destruction

g
E DUN, K, EX = Unavoidable exogenous exergy destruction

g
E DAV, K, EN = Avoidable endogenous exergy destruction

g
E DAV, K, EX = Avoidable exogenous exergy destruction

h = Exergy efficiency

y = Exergy destruction ratio

y * = Thermodynamic inefficiency

Reference

[1] Bartling J, Esperschütz J, Wilke B M, et al. ETBE (ethyl tert butyl ether) and
TAME (tert amyl methyl ether) affect microbial community structure and function
in soils[J]. Journal of Hazardous Materials, 2011, 187(1-3): 488-494,
https://doi.org/10.1016/j.jhazmat.2011.01.058.

[2] Domingues L, Pinheiro C I C, Oliveira N M C, et al. Model development and

validation of ethyl tert -butyl ether production reactors using industrial plant
data[J]. Industrial & Engineering Chemistry Research, 2012, 51(46): 15018-
15031, https://dx.doi.org/10.1021/ie302030j.

[3] Shi X, Zhu X, Zhao X, et al. Performance evaluation of different extractive

distillation processes for separating ethanol/tert-butanol/water mixture[J]. Process

33
 Safety and Environmental Protection, 2020, 137: 246-260,
https://doi.org/10.1016/j.psep.2020.02.015.

[4] Wang C, Zhuang Y, Qin Y, et al. Optimization and eco-efficiency analysis of

extractive distillation processes with different solvents for separating the ternary
mixture embedding two azeotropes[J]. Separation and Purification Technology,
2021, 269: 11876, https://doi.org/10.1016/j.seppur.2021.118763.

[5] Cheng H, Wang K, Wang Y, et al. Mechanism analysis and multiobjective

optimization of efficient and energy-saving separation of green fuel additives via
extractive pressure swing distillation with an ionic liquid mixed entrainer[J].
Industrial & Engineering Chemistry Research, 2024, 63(12): 5261-5275,
https://pubs.acs.org/doi/10.1021/acs.iecr.4c00009.

[6] Wang C, Zhuang Y, Dong Y, et al. Conceptual design and simulation-based opti
mization of indirect extractive distillation processes considering preconcentration
for separating ternary azeotropic systems[J]. Industrial & Engineering Chemistry
Research, 2023, 62(44): 18493-18510, https://pubs.acs.org/doi/10.1021/acs.iecr.3
c01748.

[7] Neyestani F, Eslamloueyan R. A novel reactive-extractive distillation process for

separation of water/methanol/tetrahydrofuran mixtures[J]. Scientific Reports,
2024, 14(1): 1931, https://doi.org/10.1038/s41598-024-52427-3.

[8] Yang A, Shen W, Wei S, et al. Design and control of pressure‐swing distillation
for separating ternary systems with three binary minimum azeotropes[J]. AIChE
Journal, 2019, 65(4): 1281-1293, https://doi.org/10.1002/aic.16526.

[9] Luo H, Liang K, Li W, et al. Comparison of pressure-swing distillation and

extractive distillation methods for isopropyl alcohol/diisopropyl ether
separation[J]. Industrial & Engineering Chemistry Research, 2014, 53(39):
15167-15182, https://pubs.acs.org/doi/10.1021/ie502735g.

[10] Su Y, Yang A, Jin S, et al. Investigation on ternary system

tetrahydrofuran/ethanol/water with three azeotropes separation via the
combination of reactive and extractive distillation[J]. Journal of Cleaner
Production, 2020, 273(5760): 123145,
https://doi.org/10.1016/j.jclepro.2020.123145.

[11] Kaymak D B. Design and control of an alternative bioethanol purification proces

s via reactive distillation from fermentation broth[J]. Industrial & Engineering C
hemistry Research, 2019, 58(4): 1675-1685, http://dx.doi.org/10.1021/acs.iecr.8b
04832.

[12] Huang K, Nakaiwa M, Wang S, et al. Reactive distillation design with

34
 considerations of heats of reaction[J]. AIChE Journal, 2006, 52(7): 2518-2534,
https://doi.org/10.1002/aic.10885.

[13] Sánchez-Ramírez E, Sun S, Sim J Y, et al. A more appropriate way to optimize

the hybrid reactive-extractive distillation system[J]. Separation and Purification
Technology, 2024, 344: 127184, https://doi.org/10.1016/j.seppur.2024.127184.

[14] Kiss A A, Jobson M, Gao X. Reactive distillation: stepping up to the next level of
process intensification[J]. Industrial & Engineering Chemistry Research, 2019,
58(15): 5909-5918, http://dx.doi.org/10.1021/acs.iecr.8b05450.

[15] Melles S, Grievink J, Schrans S M. Optimisation of the conceptual design of

reactive distillation columns[J]. Chemical Engineering Science, 2000, 55(11):
2089-2097, https://doi.org/10.1016/S0009-2509(99)00497-2.

[16] Wang C, Zhuang Y, Liu L, et al. Design and comparison of energy-saving double
column and triple column reactive-extractive hybrid distillation processes for
ternary multi-azeotrope dehydration[J]. Separation and Purification Technology,
2021, 259: 118211, https://doi.org/10.1016/j.seppur.2020.118211.

[17] Yan J, Liu J, Ren J, et al. Design and multi-objective optimization of hybrid
reactive-extractive distillation process for separating wastewater containing
benzene and isopropanol[J]. Separation and Purification Technology, 2022, 290:
120915, https://doi.org/10.1016/j.seppur.2022.120915.

[18] Zhang Z, Xu Y, Cheng X, et al. Design and optimization of heat pump assisted
reactive extractive distillation for separating benzene/n-propanol/water ternary
azeotropes[J]. Separation and Purification Technology, 2025, 360: 130994,
https://doi.org/10.1016/j.seppur.2024.130994.

[19] Liu J, Yan J, Liu W, et al. Design and multi-objective optimization of reactive-e
xtractive dividing wall column with organic Rankine cycles considering safety[J]
. Separation and Purification Technology, 2022, 287: 120512, https://doi.org/10.1
016/j.seppur.2022.120512.

[20] Kong Z Y, Sunarso J, Yang A. Recent progress on hybrid reactive-extractive

distillation for azeotropic separation: A short review[J]. Frontiers in Chemical
Engineering, 2022, 4: 986411, https://doi.org/10.3389/fceng.2022.986411.

[21] Yang A, Su Y, Sun S, et al. Towards sustainable separation of the ternary

azeotropic mixture based on the intensified reactive-extractive distillation
configurations and multi-objective particle swarm optimization[J]. Journal of
Cleaner Production, 2022, 332: 130116,
https://doi.org/10.1016/j.jclepro.2021.130116.

35
[22] Morosuk T, Tsatsaronis G, Schult M. Conventional and Advanced Exergetic
Analyses: Theory and Application[J]. Arabian Journal for Science and
Engineering, 2013, 38(2): 395-404, https://doi.org/10.1007/s13369-012-0441-9.

[23] Terzi R. Application of exergy analysis to energy systems[M]//Taner T.

Application of Exergy. InTech, 2018, http://dx.doi.org/10.5772/intechopen.74433.

[24] Elwardany M. Enhancing steam boiler efficiency through comprehensive energy

and exergy analysis: A review[J]. Process Safety and Environmental Protection,
2024, 184: 1222-1250, https://doi.org/10.1016/j.psep.2024.01.102.

[25] Salehi N, Mahmoudi M, Bazargan A, et al. Exergy and life cycle-based

analysis[M]//Hussain C M. Handbook of Environmental Materials Management.
Cham: Springer International Publishing, 2018: 1-22,
https://doi.org/10.1007/978-3-319-58538-3_84-2.

[26] Tiwari G N, Dimri V, Chel A. Parametric study of an active and passive solar
distillation system: Energy and exergy analysis[J]. Desalination, 2009, 242(1-3):
1-18, https://doi.org/10.1016/j.desal.2008.03.027.

[27] Gude V G. Exergy evaluation of desalination processes[J]. ChemEngineering,

2018, 2(2): 28, https://doi.org/10.3390/chemengineering2020028.

[28] Alic E. Experimental and numerical performance assessment of a hybrid parabolic

dish collector with photovoltaic for a distiller[J]. Energy Sources, Part A:
Recovery, Utilization, and Environmental Effects, 2023, 45(2): 3994-4016,
https://doi.org/10.1080/15567036.2023.2200737.

[29] Açıkkalp E, Hepbasli A, Yucer C T, et al. Advanced exergoenvironmental assess

ment of a building from the primary energy transformation to the environment[J].
Energy and Buildings, 2015, 89: 1-8, https://doi.org/10.1016/j.enbuild.2014.12.0
20.

[30] Anvari S, Khoshbakhti Saray R, Bahlouli K. Conventional and advanced exergetic

and exergoeconomic analyses applied to a tri-generation cycle for heat, cold and
power production[J]. Energy, 2015, 91: 925-939,
https://doi.org/10.1016/j.energy.2015.08.108.

[31] Wang Z, Xiong W, Ting D S K, et al. Conventional and advanced exergy analyses
of an underwater compressed air energy storage system[J]. Applied Energy, 2016,
180: 810-822, https://doi.org/10.1016/j.apenergy.2016.08.014.

[32] Li L, Liu Z, Deng C, et al. Conventional and advanced exergy analyses of a

vehicular proton exchange membrane fuel cell power system[J]. Energy, 2021,
222: 119939, https://doi.org/10.1016/j.energy.2021.119939.
36
[33] Yuan B, Yang Z, Yang A, et al. Target localization optimization of a
superstructure triple-column extractive distillation with four-parallel evaporator
organic Rankine cycles system based on advanced exergy analysis[J]. Separation
and Purification Technology, 2021, 272: 118894,
https://doi.org/10.1016/j.seppur.2021.118894.

[34] He Wenjun, Fei Taikang, Wang Jiahua, et al. Heterogeneous catalytic hydration
kinetics of ethylene oxide and thermal-stability of homogeneous temperature
reactor[J]. Chemical Reaction Engineering and Technology, 2017, 33(4): 326-334,
http://dx.doi.org/10.11730/j.issn.1001-7631.2017.04.0326.09.

[35] Babaie O, Nasr Esfahany M. Optimum process configuration for ETBE

production based on TAC minimization[J]. Separation and Purification
Technology, 2021, 256: 117744, https://doi.org/10.1016/j.seppur.2020.117744.

[36] Shi X, Zhu X, Zhao X, et al. Performance evaluation of different extractive distil
lation processes for separating ethanol/tert-butanol/water mixture[J]. Process Saf
ety and Environmental Protection, 2020, 137: 246-260, https://doi.org/10.1016/j.
psep.2020.02.015.

[37] Kong J, Wu Y, Xia T, et al. Optimization and control of a stabilization column for
a propylene-producing fluid catalytic cracking process[J]. Ind. Eng. Chem. Res.,
2024, 63(49): 21553-21564, https://doi.org/10.1021/acs.iecr.4c03009.

[38] Yang A, Su Y, Shen W, et al. Multi-objective optimization of organic Rankine

cycle system for the waste heat recovery in the heat pump assisted reactive
dividing wall column[J]. Energy Conversion and Management, 2019, 199: 112041,
https://doi.org/10.1016/j.enconman.2019.112041.

[39] Su Y, Jin S, Zhang X, et al. Stakeholder-oriented multi-objective process optimi

zation based on an improved genetic algorithm[J]. Computers & Chemical Engin
eering, 2020, 132: 106618, https://doi.org/10.1016/j.compchemeng.2019.106618.

[40] Yang A, Su Y, Shi T, et al. Energy-efficient recovery of tetrahydrofuran and ethyl

acetate by triple-column extractive distillation: entrainer design and process
optimization[J]. Frontiers of Chemical Science and Engineering, 2022, 16(2):
303-315, https://doi.org/10.1007/s11705-021-2044-z.

[41] Feng Z, Li Q, Rangaiah G P, et al. Economic and exergy analysis of three-product

dividing-wall column and double liquid-only side-stream distillation processes[J].
Chemical Engineering Research and Design, 2023, 200: 729-740,
https://doi.org/10.1016/j.cherd.2023.11.022.

[42] Wu Y, Meng D, Yao D, et al. Mechanism analysis, economic optimization, and

environmental assessment of hybrid extractive distillation–pervaporation
37
 processes for dehydration of n -propanol[J]. ACS Sustainable Chemistry &
Engineering, 2020, 8(11): 4561-4571,
https://dx.doi.org/10.1021/acssuschemeng.0c00263.

[43] Gadalla M A, Olujic Z, Jansens P J, et al. Reducing CO2 emissions and energy
consumption of heat-integrated distillation systems[J]. Environmental Science &
Technology, 2005, 39(17): 6860-6870, https://dx.doi.org/10.1021/es049795q.

[44] Olaleye A K, Wang M, Kelsall G. Steady state simulation and exergy analysis of
supercritical coal-fired power plant with CO2 capture[J]. Fuel, 2015, 151: 57-72,
https://doi.org/10.1016/j.fuel.2015.01.013.

[45] Mehrpooya M, Gharagheizi F, Vatani A. Thermoeconomic analysis of a large in

dustrial propane refrigeration cycle used in NGL recovery plant[J]. International
Journal of Energy Research, 2009, 33(11): 960-977, https://doi.org/10.1002/er.15
23.

[46] Kim J, Park J, Qi M, et al. Process Integration of an autothermal reforming

hydrogen production system with cryogenic air separation and carbon dioxide
capture using liquefied natural gas cold energy[J]. Industrial & Engineering
Chemistry Research, 2021, 60(19): 7257-7274,
https://doi.org/10.1021/acs.iecr.0c06265.

[47] Mehrpooya M, Ansarinasab H. Exergoeconomic evaluation of single mixed

refrigerant natural gas liquefaction processes[J]. Energy Conversion and
Management, 2015, 99: 400-413,
https://doi.org/10.1016/j.enconman.2015.04.038.

[48] Kelly S, Tsatsaronis G, Morosuk T. Advanced exergetic analysis: approaches for

splitting the exergy destruction into endogenous and exogenous parts[J]. Energy,
2009, 34(3): 384-391, https://doi.org/10.1016/j.energy.2008.12.007.

[49] Chen J, Havtun H, Palm B. Conventional and advanced exergy analysis of an

ejector refrigeration system[J]. Applied Energy, 2015, 144: 139-151,
https://doi.org/10.1016/j.apenergy.2015.01.139.

[50] Petrakopoulou F, Tsatsaronis G, Morosuk T, et al. Conventional and advanced

exergetic analyses applied to a combined cycle power plant[J]. Energy, 2012,
41(1): 146-152, https://doi.org/10.1016/j.energy.2011.05.028.

[51] Gungor A, Erbay Z, Hepbasli A, et al. Splitting the exergy destruction into
avoidable and unavoidable parts of a gas engine heat pump (GEHP) for food
drying processes based on experimental values[J]. Energy Conversion and
Management, 2013, 73: 309-316,
https://doi.org/10.1016/j.enconman.2013.04.033.
38
[52] Fallah M, Mahmoudi S M S, Yari M. A comparative advanced exergy analysis f
or a solid oxide fuel cell using the engineering and modified hybrid methods[J].
Energy Conversion and Management, 2018, 168: 576-587, https://doi.org/10.101
6/j.enconman.2018.04.114.

[53] Mehrpooya M, Sharifzadeh M M M, Ansarinasab H. Investigation of a novel

integrated process configuration for natural gas liquefaction and nitrogen removal
by advanced exergoeconomic analysis[J]. Applied Thermal Engineering, 2018,
128: 1249-1262, https://doi.org/10.1016/j.applthermaleng.2017.09.088.

[54] Sun H, Yan F, Yu H, et al. Analysis of exergy loss of gasoline surrogate

combustion process based on detailed chemical kinetics[J]. Applied Energy, 2015,
152: 11-19, https://doi.org/10.1016/j.apenergy.2015.04.087.

[55] Zhang Y R, Wu T W, Chien I L. Intensified hybrid reactive-extractive distillation

process for the separation of water-containing ternary mixtures[J]. Separation an
d Purification Technology, 2021, 279: 119712, https://doi.org/10.1016/j.seppur.2
021.119712.

[56] Ashrafizadeh S A, Amidpour M, Abolmashadi M. Exergy analysis of distillation

column using concept of driving forces[J]. journal of chemical engineering of
japan, 2013, 46(7): 434-443, http://dx.doi.org/10.1252/jcej.11we038.

[57] Yu J, Wang S J, Huang K, et al. Improving the Performance of Extractive

Dividing-Wall Columns with Intermediate Heating[J]. Ind. Eng. Chem. Res.,
2015, 54(10): 2709-2723, http://dx.doi.org/10.1021/ie503148t.

Highlights

 Reactive extractive distillation for separating water-containing ternary azeotropic

mixture.

 Feasibility for TBA/EtOH/water system validated by residue curve maps.

 Advanced exergy analysis performed to identify exergy destruction components.

 Genetic algorithm employed to optimize the processes by minimizing total annual

cost.

 Intermediate heat exchange-integrated reactive extractive distillation found to be

the best strategy.

39
Declaration of interests

☒ The authors declare that they have no known competing financial interests or personal
relationships that could have appeared to influence the work reported in this paper.

☐ The authors declare the following financial interests/personal relationships which may be
considered as potential competing interests:

40