pubs.acs.

org/IECR Article

Novel Two-Stage Process for Manufacturing Butadiene from Ethanol

Alexandre C. Dimian, Nicoleta Ilona Bezedea, and Costin Sorin Bildea*
Cite This: Ind. Eng. Chem. Res. 2021, 60, 8475−8492 Read Online

ACCESS Metrics & More Article Recommendations *

sı Supporting Information

ABSTRACT: Replacing petrochemicals with products obtained from

renewable raw materials is a top challenge today. The paper deals with the
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

design and simulation of a sustainable process for manufacturing butadiene

from bioethanol with a capacity of 91,000 t/year. The two-stage process
employs a selective CuO/Cr2O3 catalyst for ethanol dehydrogenation and
recent yield-performant Ta2O5 catalyst for butadiene synthesis. The paper
Downloaded via TSINGHUA UNIV on August 11, 2021 at 12:59:05 (UTC).

develops original flowsheets suitable for industrial implementation based on

comprehensive kinetic models that provide a realistic description of the actual
chemistry. The development of a complex separation system makes use of a
systematic task-oriented methodology. Efficient separation methods are
applied according to the chemical nature and physical properties of species for removing numerous impurities generated in reactions.
The overall butadiene mass yield and carbon yield are 0.495 kg butadiene/kg ethanol (vs the maximum value 0.587) and 0.837,
respectively. The catalyst productivity is 0.253 kg butadiene/kg cat·h. Substantial energy savings are achieved by heat recovery
around the chemical reactors and by applying heat pumping for ethanol recovery and recycling. The sustainability analysis reveals
low material intensity (0.227 kg waste/kg butadiene), moderate energy intensity (0.562 kW h/kg electrical energy and 15.8 MJ/kg
thermal utilities), and low impact on the environment. The economic analysis pinpoints the possibility of achieving butadiene prices
competitive with petrochemical products if the process is integrated in a biorefinery.

1. INTRODUCTION In the perspective of reducing the dependence on fossil

Butadiene is a key chemical intermediate produced today
exclusively by oil refining and petrochemical processes. The
global butadiene production of 16 million tons in 2018 is
projected to increase to 20 million tons in 2020.1 Key
applications are styrene−butadiene rubbers for producing tires,
acrylonitrile−butadiene−styrene polymers, and many other
products.2,3 Because the above products will continue to be
largely used, it is expected that during the next decades the
demand for butadiene will grow even in a future “decarbo-
nated” economy.
The butadiene price reports show a large variability,
generally between 1000 and 2000 $/t but with peaks up to
4000 $/t, strongly depending on the oil market, as shown in
the Supporting Information.4 Currently, the main source of
butadiene is the oil refining, around 95% globally, isolated from
the C4 fraction obtained by the steam cracking of hydrocarbon
(HC) feedstock, together with other olefins.3 The main hurdle
is getting high-purity butadiene by employing complicated and
expensive separation schemes such as extractive distillation. In
the next years, the decline of the refining business should
create tensions with respect to butadiene availability, while the
needs of butadiene-derived products will remain high.
Therefore, a price range of 1400−1600 $/t may be assumed
for testing the profitability of a future process based on
bioethanol, which will be abbreviated here simply as ethanol-
to-butadiene (ETB).
resources, sustainable technologies based on renewable raw
materials (RMs) are required. This transition period raises
challenging aspects for the conceptual process design.5 Several
chemistries for butadiene manufacturing from (bio)ethanol
and (bio)butanol are presently being investigated. Notably,
bioethanol climbed in the last years to top chemicals with an
annual global production rate of 86.6 million tons in 2019.7
The top producers are the USA, Brazil, and European Union,
with, respectively, 47.1, 25.4, and 4.4 million tons/year. The
perspective of producing a substantial part of butadiene from
bioethanol sources seems highly realistic, at least in the three
regions mentioned above.
The butadiene manufacturing from ethanol has a long story,
starting at the beginning of the 20th century by the discoveries
of Russian chemists Ostromislensky8 (two-stage process) and
Lebedev9 (one-stage process). Industrial processes have been
developed in the period 1920−1930 in the USA and former
USSR, respectively. The boom of oil refining and petrochem-
istry after the World War II made these processes
Received: March 10, 2021
Revised: May 22, 2021
Accepted: May 24, 2021
Published: June 7, 2021

© 2021 American Chemical Society https://doi.org/10.1021/acs.iecr.1c00958

8475 Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 1. Block diagram of the process design problem.
uncompetitive. Today, the interest in manufacturing butadiene
from renewable sources, primarily bioethanol, comes back.
In the first stage of the Ostromislensky process, ethanol
dehydrogenates to acetaldehyde
C2H5OH ↔ C2H4O + H 2
(1)
Typical catalysts are based on CuO. The reaction takes place
at 250 to 280 °C with few, if any, secondary products. In the
second stage, acetaldehyde reacts with ethanol to produce
butadiene by eliminating water after a complicated chemistry
C2H4O + C2H5OH = C4 H6 + 2H 2O (2)
Typical catalysts are silica-supported Ta, Zr, and Nb oxides.
In the one-stage process, the two reactions occur
simultaneously. Old catalytic systems were based on MgO/
ZnO, but more recent formulations are M/MOx/SiO2, where
M and MOx are a metal and its oxide, suitable for
dehydrogenation and for butadiene synthesis, respectively.
The key issue in developing ETB technologies is finding
performant industrial catalysts in terms of yield, productivity,
and robustness. The research of catalysts was the object of
numerous papers in the last years, only interesting reviews
being cited here.4,10−13 In contrast, the analysis and process
design of an ETB technology taking advantage from the recent
progress in catalysis is currently missing. This is the goal of the
present research.
The few papers published so far addressing the process
design tackle more sustainability issues from a life-cycle
analysis perspective than the design process itself. A
comparison with previous works will be discussed in a final
section.
At the first sight, the one-stage process should have the
advantage of a single catalyst and an apparently simpler
technology. However, it can be argued that a two-stage process
employing different catalysts could bring the advantage of a
better individual performance. Our study will show indeed that
using two catalyst results in superior overall butadiene carbon
yield, over 83% and in a profitable product price, the
dehydrogenation step being limited to around 10% from the
capital and operation costs.
2. GOALS AND ACHIEVEMENTS
The goal of this research is the development of an ecoefficient
process for manufacturing butadiene from ethanol by a two-
step reaction technology. Performant catalysts ensure good
productivity and yield of acetaldehyde and butadiene. The
8476
pubs.acs.org/IECR Article
second stage, butadiene synthesis, is particularly complex as
chemistry. A comprehensive recent kinetic model is applied14
but completed with missing information concerning the
chemical equilibria and adapted to industrial conditions by
using catalyst efficiency. The paper develops original
flowsheets of industrial value based on kinetic simulation of
chemical reactors and on efficient separation schemes,
including the treatment of waste and impurities and integrating
energy-saving aspects. The performance features are captured
by economic assessment and sustainability analysis. The plant
production target is 11,400 kg/h butadiene with purity over
99.7%, which corresponds to 91,000 t/year at 8000 h annual
operating time.
The results of this study are process flowsheets supported by
comprehensive material and heat balances, sizing of units, as
well as capital investment and operation costs. The economic
and sustainability analyses demonstrate that the proposed two-
stage ETB process is feasible and profitable in the frame of
assumed prices for RMs and utilities. The integration in a
biorefinery frame enables the use of cheaper utilities and
sharing the infrastructure, thus making sure competitive
product price even by higher bioethanol prices.
3. METHODOLOGY
The strategy of process design follows the process synthesis
methodology presented in the authors’ book (Chapters 7 and
9).6 The emphasis is set on the reaction−separation−recycle
structure. This level deals also with the problem of heat
integration (HI) around the chemical reactor. The synthesis of
separations is solved by adopting a task-oriented methodology
aided by computer simulation including suitable thermody-
namic modeling. The process integration stage solves the
challenge of energy saving. The conceptual analysis and
flowsheet development presented in this research has been
supported by the comprehensive simulation software Aspen
Plus 11.0.
Figure 1 presents the process block diagram. Two reactors
for ethanol dehydrogenation and butadiene synthesis are
employed, each one with its separation system. Separating the
mixture obtained by the butadiene synthesis is by far more
difficult. By using the principle of the first separation step
(FSS), the flowsheet development is split in two subsystems,
the vapor separation system (VSS) and liquid separation
system (LSS).
The VSS tackles the separation of raw butadiene, as well as
of light byproducts and impurities. This is followed by
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 2. Ethanol dehydrogenation to acetaldehyde with hydrogen separation (flowsheet without HI).
butadiene purification, a step that could involve significant
costs, depending on impurities formed by the synthesis stage.
The synthesis of the LSS flowsheet is rather complex. The
butadiene synthesis generates a large amount of water that
entrains unconverted reactants and numerous impurities. The
goal of LSS is to recover and recycle ethanol (EtOH) and
acetaldehyde (AcH), as well as to remove intermediate-boiling
and heavier impurities, even in small amounts, preventing their
accumulation in recycles. Figure 1 shows that exit points
should be conveniently located in both the VSS and LSS. Some
impurities leave the plant in both subsystems. The VSS and
LSS are thus interconnected. The generation of splits in
process synthesis is assisted by heuristics, while the selection of
the suitable separation methods is based on the characteristic
properties of components. Computer simulation is used for the
assessment of separations.
4. FIRST STAGE: ETHANOL DEHYDROGENATION
Ethanol dehydrogenation is a reversible endothermal reaction
with the enthalpy of reaction ΔHR (298 K) = 12.1 kcal/mol. It
takes place in the presence of a CuO catalyst promoted with
different elements.15,16 The reaction conditions are usually
temperature of 250 to 350 °C and pressure of 2 to 3 bar. The
acetaldehyde yield is generally high, over 95%, with impurities
such as acetic acid, ethyl ether, ethyl acetate, and gaseous
species.
An alternative approach for getting acetaldehyde is the
oxidative dehydrogenation of ethanol. The reaction becomes
exothermal by in situ hydrogen combustion, and as a result,
there is no need of expensive high-temperature thermal agents.
The equilibrium shifts to reaction completion. However, there
are also disadvantages, the most important one being a lower
yield due to the formation of secondary oxidation products.
In this work, we prefer the direct dehydrogenation method,
selective and efficient, as proved in numerous applications.
When using CuO-based catalysts, the selectivity achieves
practically 100%.15,16 The hydrogen byproduct is valuable,
namely, as a future ecological fuel.
4.1. Dehydrogenation Reactor: Design and Sizing.
Figure 2 illustrates the conceptual flowsheet. The ethanol feed
consists of 15,000 kg/h aqueous solution with azeotropic
composition (95% wt ethanol).
8477
pubs.acs.org/IECR Article
The chemical reactor R-100 is built as a vertical heat
exchanger with the catalyst in tubes and thermal fluid in a shell.
The commercial CuO/Cr2O3 catalyst consists of cylindrical
pellets with both height and diameter of 3.5 mm, a solid
density of 4.3 g/cm3, and a porosity of 50%. The kinetic
parameters of a Langmuir−Hinshelwood−Hougen−Watson
model16 are given in the Supporting Information. The
experimental data were obtained in the range 225−285 °C
in an isothermal reactor fed with ethanol 95% wt. The kinetic
parameters consider the catalyst efficiency as well. In the above
conditions, the byproduct formation is negligible. The reaction
is controlled by the chemical equilibrium, being favored by low
pressure and high temperature.
A suitable thermal fluid is Dowtherm-A. This consists of a
mixture (C6H5)2O/(C6H5)2 that may operate as vapor
between 257 and 400 °C and as liquid between 15 and 400
°C, depending on the working pressure.
The reactor simulation (vapor phase at 1.7 bar and 250 °C)
shows that a conversion over 80% is achieved in a single
reactor with 3670 tubes of 38/34 mm outside diameter/inside
diameter and 6 m length, accommodated in a shell of 2400 mm
diameter. The catalyst amount is 44,130 kg. On the shell side,
39,600 kg/h Dowtherm A (saturated vapor at 5 bar and 344.4
°C) flows in counter current.
The tube-side heat-transfer coefficient (HTC) has a value of
245 W/m2 K (Supporting Information).17 On the shell side,
the partial HTC of condensing Dowtherm is 1000 W/m2 K, as
indicated in the technical documentation.18 By considering all
resistances, it results on overall HTC of 160 W/m2 C.
Figure 3 presents profiles of conversion and temperature in
the reaction tubes, as well as the temperature profile of thermal
fluid in the shell. Full lines correspond to the catalyst of 100%
activity. Note that the conversion rises vigorously on the first
half of the reactor, approaching the equilibrium conversion of
84%. In the entry zone, the temperature in tubes falls from 250
to 229 °C, but it recovers quickly to a constant value around
280 °C. The temperature profile reflects an initial fast reaction
rate that may lead up to 60% conversion in the entry reactor
zone. Here, the thermal fluid supplies most of the energy
needed by the reaction by vapor condensation and afterward
by liquid in convection. The reduction of temperature
difference between the shell and tubes is a measure of the
decay in the reaction rate.
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

4.2. Separation System. The task of the separation

Figure 3. Ethanol dehydrogenation to acetaldehyde: profiles of
ethanol conversion (left axis), reaction mixture, and thermal fluid
temperature (right axis) vs the dimensionless reactor length. Solid
lines are profiles for the fresh catalyst (100% activity). Dashed lines
are profiles for the deactivated catalyst (50% activity).
Hence, the reactor design ensures both preheating and
reaction functions. In addition, an exit temperature of 30 °C
above the inlet allows good temperature driving force for heat
integration purposes.
The conversion profile from Figure 3 (solid lines) seems to
indicate that the catalyst amount is oversized. However, this is
necessary to cope with uncertainty in design and operation.
Consider, for example, a much lower catalyst activity, say
reduced by 50% due to deactivation. As shown in Figure 3
(dashed lines), in this case, the full tube length is used for
approaching the equilibrium conversion. On the shell side, the
condensation zone is wider, while the liquid-phase temperature
drops slower. However, the temperature profile on the catalyst
side is similar to the previous case, which demonstrates that
the reactor operation capability is preserved.
The reactor productivity is 10,401/44,130 = 0.258 kg
acetaldehyde/kg cat·h.
system is recovering hydrogen as the byproduct and
acetaldehyde to be used as the reactant for the next step.
The strategy adopted here uses absorption with ethanol as the
solvent. Figure 2 displays the conceptual flowsheet that
highlights the main operations. Energy-saving aspects will be
treated later. The ethanol stream of 15,000 kg/h, preheated in
the unit H-101 to 250 °C, enters the reactor R-100. The
reactor effluent is cooled to 33 °C in H-102 with cooling water.
The first hydrogen separation takes place in the flash FL-101,
enhanced by injecting 12,000 kg/h cold ethanol solution of
91% wt, recycled from the butadiene synthesis stage. Since the
hydrogen stream still contains a substantial amount of
acetaldehyde, this is submitted to a second separation stage
in FL-102 after compression at 10 bar, cooling at 21 °C with
chilled water and injection of recycled ethanol solution. The
flow rate of ethanol used as the solvent was chosen to ensure
the flow rate of ethanol necessary in the second stage, which is
in large excess to the acetaldehyde. The ethanol recycled was
distributed between the absorption stages FL-101 and FL-102
so as to maximize the acetaldehyde recovery. Small amounts of
acetaldehyde and ethanol are recovered at hydrogen separation
by cooling at −30 °C with propane refrigeration. Although
small duty (−0.37 MW), this operation allows recovery of
significant amounts of ethanol (139 kg/h) and acetaldehyde
(579 kg/h). The hydrogen product stream still contains 10%
acetaldehyde, which can be further removed by using standard
methods. The outlet stream (AcH-S1) from this section
contains ethanol and acetaldehyde in the mass ratio of around
2:1. The recoveries of aldehyde and hydrogen are 99.5 and
98.6%, respectively. The energy analyzer tool in Aspen Plus
11.0 indicates that 5.87 MW or 49.9% of energy may be saved.
The large duties of the exchangers H-1 and H-2 indicate that
the heat recovery problem is of interest primarily around the
chemical reactor. The heat-integrated flowsheet (Section 7)
will develop this issue as a heat exchange network. This will
show much lower inlet/outlet temperature differences.

Table 1. Chemical Reactions and Kinetic Parameters for Butadiene Synthesis Using the Ta2O5 Catalysta
no reactions k Ea Keq k′ φ η
R1a 2AcH ↔ CRA + H2O 0.3 40 0.0684 4.387 1.41 0.5
R2a 2AcH ↔ MVK + H2O 0.064 40 8.4888 0.008 0.652 0.8
R1b CRA + EtOH ↔ 2-buten-1-ol + AcH 36 45 9.7260 3.70 28.5 0.035
R2b MVK + EtOH ↔ 3-buten-2-ol + AcH 36 45 0.0195 1842 28.5 0.035
R1c 2-buten-1-ol → BD + H2O 0.0094 55 <0.2 1.0
R2c 3-buten-2-ol → BD + H2O 0.0094 55 <0.2 1.0
R1d 2-buten-1-ol → butanal 0.00014 50 <0.2 1.0
R2d 3-buten-2-ol → MEK 0.00014 50 <0.2 1.0
R1e butanal + EtOH ↔ 1-butanol + AcH 36 45 3.440 0.096 28.5 0.035
R2e MEK + EtOH ↔ 2-butanol + AcH 36 45 0.200 180 28.5 0.035
R1f 1-butanol → 1-butene + H2O 0.0094 55 <0.2 1.0
R2f 2-butanol → 2c-butene + H2O 0.0094 55 <0.2 1.0
R3 2AcH ↔ Et-Ac 0.00036 41 0.767 0.0005 <0.2 1.0
R4 2AcH ↔ acetone + Formal 0.0045 50 0.007 0.6312 <0.2 1.0
R5 acetone + EtOH → C3 + AcH + H2O 0.65 50 none 1.0
R6 acetone + EtOH → C5H8 + H2O 0.54 40 none 1.0
R7 EtOH → C2 + H2O 8.0 × 10−6 157 none 1.0
R8 2EtOH ↔ Et2O + H2O 4.4 × 10−6 103 3.95 1.1 × 10−6 none 1.0
R9 2C5H8 → C10H16 (pinene) 10 0 10 none 1.0

a
k and k′ in (m3)2/kg/kmol/s (second-order reactions) or in m3/kg/s (first-order reactions); Ea in kJ/mol.

8478 https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 4. Flowsheet of the butadiene reaction section and FSS.
5. SECOND STAGE: BUTADIENE SYNTHESIS
5.1. Chemical Kinetics. Table 1 presents the detailed
kinetic model.14 This offers a comprehensive description of the
species involved in an industrial ETB process. The
stoichiometric equations are presented in Figure S1 (Support-
ing Information). The butadiene synthesis follows the
condensation reaction of acetaldehyde (AcH) along two
routes, leading to (a) croton aldehyde (CRA) and (b) methyl
vinyl ketone (MVK). These intermediates react further with
ethanol (EtOH) and give the unsaturated alcohols 2-buten-1-
ol and 3-buten-2-ol by hydrogen transfer to AcH (Meerwein−
Ponndorf−Verley reaction). Finally, 1,3-butadiene (BD) is
obtained by dehydration of butenols. The butenols are at the
origin of key impurities, such as butanal, 1-butanol, methyl-
ethyl ketone (MEK), and 2-butanol. Dehydration of 1- and 2-
butanol leads to butene isomers, such as 1-butene and 2-cis-
butene. The thermal degradation of acetaldehyde results in the
formation of ethyl-acetate (EtAc), acetone, formaldehyde
(Formal), and propylene (C3). A significant amount of
ethanol can dehydrate to ethylene (C2) and ethyl ether
(Et2O).
There also heavy species appearing by the combination of
byproducts and impurities. These are described in patents as
“black oil”.19 In the kinetic model, this is described by
pentadiene (C5H8) formation by the reaction of acetone and
ethanol, followed by dimerization to C10H16, a HC assimilated
for the simulation needs with α-pinene (normal bp 185 °C).
The kinetic models are of the power-law type. The
equilibrium constants (not reported in the original paper14)
were evaluated as follows: the RGIBBS model from Aspen Plus
was used to calculate the equilibrium composition at several
temperatures, from which the equilibrium constants were
calculated. From these values, the temperature-dependent
equilibrium constants (ln Keq = A + B/T) were determined by
linear regression and used to calculate the pre-exponential
factor and activation energy of the reverse reactions. The
standard enthalpy and free energy of formation necessary to
compute the chemical equilibrium were estimated by the Gani
method.
In this project, we consider that the catalyst consists of
spherical particles of 3 mm diameter. Since the original kinetic
parameters14 were obtained using a powder catalyst, these were
corrected by calculating the Thiele modulus φ and the
8479
pubs.acs.org/IECR Article
effectiveness factor η to account for the effect of the internal
diffusion. In two cases (ethylene and ethyl ether formation),
the original pre-exponential factors14 were adjusted to
compensate inconsistencies in product distribution. The
kinetic model implemented in Aspen Plus 11.0 has 27
reactions described by power-law rate equations.
The first four lines of Table 1 show the thermodynamic
equilibrium constants Keq at 340 °C for the key reactions of the
routes a and b, which involve the formation of CRA and
methyl viny ketone, respectively. Thus, the formation of CRA
(1a) is controlled by the chemical equilibrium, while the
reaction to 1-butenol (2a) is practically irreversible, the
equilibrium constants being 0.068 and 8.49, respectively. An
opposite situation takes place in the second path, where the
equilibrium constants of the reactions 1b and 2b are 9.73 and
0.019.
5.2. Reaction Section. Figure 4 depicts the flowsheet of
the reaction section. The butadiene synthesis takes place in the
reactor R-200. This receives an ethanol/acetaldehyde mixture
from dehydrogenation, already in the mass ratio 2:1, as well as
acetaldehyde and ethanol recycle streams. In addition, an
ethanol makeup stream is used to achieve the reactor-inlet
mass ratio ethanol/acetaldehyde of 3:1, found to be optimal.14
The pressure at the reactor inlet is 3 bar. The unit R-200 is a
multitubular heat-exchange reactor hosting the catalyst in
tubes. The temperature in the reactor varies in the range from
340 to 350 °C, maintained by using Dowtherm A.
5.3. Kinetic Reactor Design. Figure 5 presents the results
of simulating a plug flow reactor in view of determining the
necessary amount of the catalyst. The feed is a mixture of
41,500 kg/h (30,000 ethanol, 10,000 acetaldehyde and 1500
water) at 3 bar and 330 °C. The mass ratio of ethanol to
acetaldehyde is set to 3.14 The left vertical axis represents the
conversion of reactants, acetaldehyde and ethanol, as well as
the butadiene mass yield, while the right vertical axis (enlarged
scale) shows the mass yield of by products and impurities, such
as olefin (ethylene plus propylene), butenes, oxygenated
components, CRA plus MVK, and butenols.
Note that the component yield is reported referring to a
conventional mixture of 10,000 kg/h acetaldehyde and 10,000
kg/h ethanol. The conversion increases rapidly with the
amount of the catalyst, around 80% acetaldehyde conversion
being achieved for a catalyst load of 20,000 kg. At 40,000 kg of
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 5. Profiles of conversion and yield for 1,3 butadiene synthesis.
Reactant conversion and butadiene yield on the left axis. Yield of
other components on the right axis.
the catalyst, a state close to equilibrium is reached. At 100,000
kg of the catalyst, the acetaldehyde conversion and the
butadiene mass yield reach 94.2 and 55.5%, respectively. On
the contrary, the yield of light olefins (mostly ethylene) never
reaches equilibrium but increases linearly from zero up to 10%.
The formation of intermediates (CRA, MVK, and butenols)
passes through a maximum, which demonstrates the parallel
consecutive character of the complex reaction network. The
yields of oxygenated impurities and butenes remain practically
constant at maximum 1.7 and 0.6%, respectively.
We were also able to evaluate the contribution of the two
reaction routes in BD formation by substituting the
dehydration of 3-buten-2-ol to the 1,2-butadiene isomer in
the second path. The simulation shows that the CRA path
contributes around 60%, while the MVK path contributes the
rest of 40%, in agreement with the experimental values.14
On the basis of these results, we select a reactor hosting
45,000 kg of the catalyst. The reactor simulation indicates
conversion of acetaldehyde and ethanol 87.7 and 35.1%,
respectively, yield butadiene 51.1% and C2/C3 olefins 5.25%.
These results will be consolidated as rigorous material balance
including the recycle streams presented in Section 6.
5.4. Thermal Reactor Design. The butadiene synthesis
reaction is moderate endothermal. The reactor is a multi-
tubular heat exchanger with the catalyst in tubes and thermal
fluid in shell. The input mixture, obtained from the plant
simulation including recycles, has a flow rate of 50,000 kg/h,
and a composition of acetaldehyde 22.8, ethanol 68.5, water
8.7 all in % wt, entering at 330 °C and 3 bar. The reaction
temperature should be kept in the optimal range from 330 to
370 °C. An amount of 45,000 kg of the catalyst is hosted in
5300 tubes of 50/46 mm OD/ID and 7 m length, which can
be assembled as two shells of 2500 mm. The overall heat
transfer of 160 W/m2/K is controlled by the flow in tubes. The
Dowtherm flow is 50,400 kg/h saturated vapor at 8.5 bar and
382 °C.
Figure 6 presents the profiles of acetaldehyde conversion
and temperatures in tubes and the shell. The temperature
quickly reaches a peak of 367 °C at 20% tube length, while the
acetaldehyde conversion rises to 70%. Then, the temperature
falls gradually to the exit value of 340 °C, while the conversion
attains a steady state at 89%. On the shell side, one notes a
relatively short zone of condensing vapor, followed by liquid in
convection. Note that the temperature profile ensures a
8480
pubs.acs.org/IECR Article
Figure 6. Butadiene synthesis reactor: profiles of acetaldehyde
conversion (left axis) and temperatures in tubes and in the shell
(right axis) vs dimensionless length.
minimum 10 °C temperature difference between the outlet
and inlet streams, required to achieve energy saving by
implementing a feed-effluent heat exchanger (FEHE) (see
Section 7).
Although no information about catalyst deactivation by coke
deposition is given in the references regarding the kinetics,14,16
fees for purchase and/or regeneration are allocated in the
economic analysis. Both catalysts are metal oxides that should
allow simple regeneration by in situ coke burning with hot
air.35 At a higher deactivation rate, a circulating fluid bed
(CFB) reactor could be considered. In view of this approach,
an attrition-resistant Ta2O5 catalyst was made, and spherical
particles of 75 to 150 μm were tested in a cold device.36 The
design of innovative CFB reactors may follow the methods
outlined in a recent authors’ paper dealing with glycerol
dehydration to acroleine.37
5.5. Separation Section. The reactor outlet is a mixture
characterized by the presence of various chemical species, as
listed in Table 1. Three categories can be distinguished based
on volatility, as indicated by the normal boiling point (in °C).
(1) Volatiles: ethene (−103.7), propene (−47.7), form-
aldehyde (−19.3), 1-butene (−6.2), butadiene (−4.4), 2
cis-butene (3.7), and acetaldehyde (21.0).
(2) Intermediates: diethyl ether (34.4), pentadiene (44.1),
acetone (56.1), butanal (74.8), ethyl acetate (77.0),
ethanol (78.3), MEK (79.6), MVK (80.8).
(3) Heavies: 2-butenol (90.5), 3-butenol (97.0), 2-butanol
(99.7), water (100.0), CRA (102.5), 1-butanol (118.7),
and α-pinene (185.0) for black oil.
The non-random two-liquid (NRTL) method with Henry
components was selected in Aspen Plus to calculate the
physical properties (thermal, transport, and phase equilibria).
All components are in the Aspen database, together with the
parameters for calculating the pure-component physical
properties. The binary interaction coefficients were taken
from Aspen 11.0 and NIST databases or estimated by
UNIFAC, when not available.
5.5.1. First Separation Step. As shown in Figure 4, after
reaction and cooling at 15 °C, the mixture enters the FSS. The
task divides the separations into two subsystems, the VSS and
LSS. The first separates butadiene, light olefins, and volatile
impurities, while the second tackles the recovery of
acetaldehyde and ethanol reactants to be recycled, as well as
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 7. Flowsheet of the VSS.
the removal of the intermediate and heavier impurities. The
VSS and LSS are interconnected.
The characteristic property for performing the FSS is the
relative volatility. Suitable separation methods are distillation
and absorption, which are combined in the unit C-201. This is
a reboiled stripping column with 12 theoretical stages, with the
feed entering at the fifth stage, while 5000 kg/h cold water is
added on the top stage. The pressure is 2.1 bar and molar
reflux ratio 1.1. The separation is sharp. The vapor phase
quantitatively recovers butadiene and over 99% of acetalde-
hyde, olefins, butene, formaldehyde, and ethyl ether, as well as
the majority of acetone, ethyl acetate, 2-butanol, MEK, and
MVK. The liquid phase captures most of ethanol and water,
with heavier impurities as CRA, butenols, and 1-butanol.
Butanal is split equally between the two streams. One may
consider MEK as key light and CRA as key heavy. Note that
the black oil is soluble in an ethanol-rich solution but not in
water.19 The removal of black oil is compulsory before entering
the separation system. A suitable location is after the column
C-201 (HS unit in Figure 4). Standard liquid−liquid extraction
with a suitable solvent or adsorption on clays may be applied.
This subject is not handled here, but information about how to
deal with it may be found in patents.19
Acetaldehyde and light oxygenated components will be
recovered from the butadiene stream in the VSS and sent for
treatment to the LSS. On the other hand, the butadiene
entrained in the LSS will be recovered and sent to VSS. The
development of the VSS and LSS is presented in more details
in the next sections.
The numerous impurities generated in this process are
characterized by large differences in volatility and chemical
nature. All impurities must be removed, and suitable exit points
must be found, otherwise these accumulate in recycles creating
difficulties in controlling the plant operation.20 In addition, the
process produces a significant amount of wastewater that must
be treated in view of becoming a resource. This goal must be
attained when handling the other effluents as well.
8481
pubs.acs.org/IECR Article
The synthesis of distillation sequences should fulfill both the
desired recovery of species, related to the column design
(number of stages, reflux/boil-up), and the constraints of
temperatures in condenser and reboiler which depend on the
operating pressure. These temperatures are limited by the
available utilities, for cooling (water, air, chilled water, brine)
and for heating (low-pressure steam). These parameters were
found, for each separation unit, by rigorous simulation in
Aspen Plus.
5.5.2. Vapor Separation System. Figure 7 illustrates the
flowsheet. The vapor stream leaving C-201 in top is sent to
compression from 2.1 to 5 bar in CP-301 and, after cooling,
submitted to a new split in the unit C-301. The task is the
recovery of acetaldehyde and removal of light oxygenated
components from the butadiene stream. The method is
reboiled stripping assisted by absorption in cold water. The
unit C-301 has 15 stages fed with mixture on stage 12 and cold
water on top stage. Over 94% from butadiene goes in the top
stream with olefins and butenes. The acetaldehyde and other
oxygenated components are absorbed completely in water,
except di-ethyl ether for which the recovery is only 86%. Thus,
the butadiene is impurified with light HCs but only with traces
of oxygenated components, the majority being sent to the LSS
for complete removal (stream S-305, feed to column C-401).
Further, the butadiene stream is conditioned for olefin
separations, firstly by compression in CP-302 from 5 to 26 bar,
such to ensures suitable temperatures in condenser and
reboiler, then cleaned and dried in C-302. Note that heating
from 11.1 to 40.2 °C is required to avoid partial condensation
during compression. The unit C-302 may consist of several
adsorbent beds, each one with specific task. The most used
adsorbent for oxygenated species is the activated carbon.21
Silica gel is used for gas drying.
The olefin recovery by cryogenic distillation takes place in
the column C-303, equipped with 50 theoretical stages and
feed location on stage 15. The reflux ratio is 3. The ethylene
and propylene recoveries in the top product are 100 and 84%,
respectively. The butadiene is separated quantitatively in the
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 8. Flowsheet of the LSS.
bottom stream, together with butenes and the rest of
propylene. The bottom product is a “raw butadiene” with
99.15% purity, not sufficient for fulfilling the quality
specification (99.7%). The purification of butadiene is treated
in Section 6.
5.5.3. Liquid Separation System. The LSS depicted in
Figure 8 should ensure the recovery and the purification of the
recycled reactants (ethanol and acetaldehyde), the separation
of the large amount of water formed in the synthesis reaction,
as well as the removal of impurities in exit streams. The
strategy consists in building separate recycle paths for
acetaldehyde and ethanol/water streams. In addition, the
problem of impurities is better managed.
The first unit C-401 is a distillation column that receives the
bottom streams from C-201 and C-301. These contain large
amounts of water (resulting from reaction and absorption)
with ethanol, acetaldehyde and impurities. The task of this key
unit is separating the light components (acetaldehyde and
impurities) from an ethanol/water mixture, that allows further
recycling the large ethanol excess. The column is provided with
a side-stream that creates a separation path for acetaldehyde
and intermediate impurities. C-401 has 30 stages with feed on
9, and side-stream on 5. The vapor distillate recovers most of
the butadiene with some amounts of acetaldehyde and acetone.
After compression in CP-401 these components are removed
by absorption in cold water in the column C-402, while the
butadiene (stream S-406) is sent to VSS. The butadiene
recovery in the LSS is economically important since it
concerns approximately 5% of the final butadiene produced.
The side stream from C-401 contains most of the acetaldehyde
and diethyl ether. This impurity may form in significant
amounts and is difficult to separate. The bottom stream of C-
8482
pubs.acs.org/IECR Article
401 collects water and ethanol with heavier oxygenated species
such as alcohols, aldehydes, and ketones.
The recovery of the acetaldehyde resumes with the removal
of lights in the column C-403. As this column is operated at
lower pressure, cooling is required (H-402) to avoid flashing of
the liquid fed to C-403. The stream “lights” contains volatile
organic components (VOCs) such as formaldehyde, with some
acetaldehyde and butadiene. Note that C-403 is provided with
a small side stream in view of complete recovery of
acetaldehyde and diethyl ether. In this way, the bottom stream
WW1 is almost free of impurities and can be recycled as
process water (PW) for absorption.
The removal of diethyl ether from acetaldehyde takes place
in the column C-404 (30 stages), which receives the side
streams from the columns C-401 and C-403 (stage 20). It was
found that the diethyl ether may be absorbed in a paraffinic
HC fraction, in this case, dodecane. The spent solvent is
regenerated and recycled (flowsheet and material balance are
given in the Supporting Information). The top vapor
containing over 82% acetaldehyde and around 15% butadiene
may be conditioned in the unit C-405 by adsorption over
activated carbon and sent to the reactor.
The bottom stream of C-401 recovers the ethanol to be
recycled, as well as most of the water and oxygenated
impurities. This stream is sent to the column C-406 that has
the task to separate ethanol from a water solution of around
42% wt. This unit has 30 stages and a reflux ratio of 1.35. The
top vapor stream quantitatively recovers ethanol as 90.9% wt
with 8% wt water and 1.1% wt oxygenated species. The
oxygenated species are 85% wt carbonyl species (form-
aldehyde, acetaldehyde, butyraldehyde, diethyl ether, acetone,
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Figure 9. Flowsheet of the butadiene purification.
MVK, MEK, and ethyl acetate) and 15% wt hydroxylic C4
species.
Removing the oxygenated impurities is a challenge. Since
these show differences in size and polarity, physical methods
may be envisaged as adsorption on activated carbon or
hyperfiltration on membranes. However, chemical methods
can be more efficient when treating specific functional groups.6
A promising method for removing impurities containing
carbonyl groups, such as aldehydes and ketones, is the
treatment with hydrazine or derivatives. These can be converted
into the corresponding alkanes by the Wolff−Kishner reaction.
For illustration, we consider acetone. The reaction with
hydrazine under basic conditions gives an unstable inter-
mediate, hydrazone, and then propane by releasing nitrogen
(CH3)2 −CO + N2H4
→ (CH3)2 −CN−NH 2 + H 2O
→ CH3−CH 2−CH3 + N2 (3)
The method applicability may be illustrated by the selective
removal of carbonyl compounds from sweet wines using solid-
phase extraction with phenyl-sulfonyl-hydrazine.22 The solid
support may be a modified polymer, such as polystyrene, or
divinylbenzene resin. Thus, the unit C-407 consists of a
tandem of two columns, filled with the supporting solid, such
as spheres or packing, one working in purification, the other in
regeneration. The resulting ethanol stream may be sent directly
to the reactor, the HCs being eliminated in recycles, as
indicated. Note that the same treatment, this time for the
removal of carbonyl components from the butadiene stream,
may be applied by adding a suitable packing in the unit C-302.
The simulation of the whole flowsheet presented in Figures
4, 7, and 8 was performed by closing the recycles. A control
structure based on “design specification” was built in Aspen
Plus. The goal is setting up the mass ratio of ethanol to
acetaldehyde at the reactor inlet at the value needed for
8483
pubs.acs.org/IECR Article
optimal operation. The design specification structure is a key
plant-wide control structure of the material balance.23 The
results that follow are valid for an amount of 45,000 kg of the
catalyst with the ethanol-to-acetaldehyde ratio of 3 and
polytropic reactor operation around 340 °C. The acetaldehyde
fed to reactor consists of 11,827 kg/h acetaldehyde (10,179
kg/h from the ethanol dehydrogenation section and 1648 kg/h
recycled). The ethanol flow rate is 34,481 kg/h (30,000 kg/h
from the dehydrogenation, 2512 kg/h net recycled from
distillation, and 1919 kg/h as makeup). Since the flow rate of
butadiene at the reactor outlet is 11,510 kg/h from which 109
kg/h comes by the acetaldehyde recycle, one gets a net
butadiene production of 11,401 kg/h. The per-pass con-
versions are 85.5% for acetaldehyde and 34% for ethanol. For a
catalyst load of 45,000 kg, one gets a reactor productivity of
0.253 kg butadiene/kg catalyst·h. Full mass balance is given in
the Section 7 and in the Supporting Information.
6. BUTADIENE PURIFICATION
Typical purity specifications of commercial butadiene are:3 1,3
butadiene 99.6%, propyne 5 ppm, acetylenes 20 ppm, 1,2
butadiene 20 ppm, dimers max 250 ppm, and carbonyls max
25 ppm. The butenes are admitted as reported in the limit of
minimum 99.5% purity.2 It should be emphasized that the
acetylenes and 1,2 butadiene are harmful impurities by the
steam cracking of HC feedstock, whose removal requires a
considerable effort in equipment and energy.2,3 When
examining the chemistry of the ETB process (Table 1),
these impurities should not appear, which is a significant
advantage versus the petrochemical processes. Other impur-
ities produced in the ETB process are allowed in larger
concentration (e.g., 0.3% wt 2-cis-butene). However, even in
this case, the removal by simple distillation of butenes, namely,
of 2-cis-butene, is particularly difficult. The relative volatility to
butadiene of 1-butene is 1.1 and of 2-cis-butene 0.78. Using
extractive distillation with suitable solvents, such as N-methyl-
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
2-pyrrolidone and dimethylformamide (DMF), makes the
separation feasible but still demanding in energy.2,3
The challenge in simulation is finding a suitable thermody-
namic model. This subject was addressed in few papers.24−26
Here, we use the NRTL model with binary interaction
parameters from the Aspen v. 11 (DMF with 2-cis-butene and
butadiene), NIST v. 11 (DMF with 1-butene and propene,
butadiene with 1-butene, 2-cis-butene and propene, ethene/
propene, and 1-butene/propene), the others being estimated
by UNIFAC. The simulation shows that the relative volatilities
of 1-butene and 2-cis-butene to butadiene increase at 1.6 and
1.2, respectively.
The raw butadiene stream (raw-BD) obtained so far has a
purity of 99.15% wt, with 0.55% wt 2-cis-butene, 0.03% wt 1-
butene, and 0.27% wt ethene/propene. This may be acceptable
for some applications. However, the specification for 2-cis-
butene for typical purity is not satisfied. In the following, we
investigate the butadiene purification by employing DMF as
the solvent. Figure 9 presents the flowsheet. The extractive
distillation column C-501 operates at 6 bar and has 90
theoretical stages. The simulation shows that the separation
needs a large solvent ratio of 5.8. C2/C3 olefins and C4
impurities leave at the top, the split fractions being 98.3% for
1-butene but only 49.5% for 2-cis-butene. Purified butadiene is
obtained by solvent striping in C-502, while solvent traces are
removed by deep-cooling in the flash unit FL-501.
7. MATERIAL BALANCE
The results discussed in this section were obtained from the
stream report supplied by Aspen Plus 11.0. The relative error
in the closing material and energy balances was 0.0001. Table 2
Table 2. Material Balance Stage 1, Ethanol
Dehydrogenation
inputs outputs
components feed-1 solvent AcH-S1 H2 mass yielda
flow kg/h 15,000 22,000 36,477 525
ethanol 14,250 20,240 23,787
acetaldehyde 0 0 10,179 57 0.718
water 750 1760 2510 0.000
hydrogen 1 468 0.033
I/O reaction 14,250 14,252
I/O balance 37,000 37,003
a summing the cells of PW in inputs and in outputs minus the
kg product/kg ethanol in feed-1.
presents the material balance for stage 1, ethanol dehydrogen-
ation. The input is 15,000 kg/h solution ethanol 95% wt
containing 14,250 kg/h ethanol. The output stream AcH-S1
sent to stage 2 contains 10,179 kg/h acetaldehyde and traces of
hydrogen. The output stream H2 contains 468 kg/h hydrogen
and 57 kg/h acetaldehyde. Although the selectivity is 100%,
the mass yield is only 0.718 kg-acetaldehyde/kg-ethanol since
the reaction is equilibrium limited. No recycle is necessary
because the ethanol/acetaldehyde mixture is sent directly to
the second stage. Ethanol solution of 91% wt recycled from the
second stage is used as the solvent in separations. The result is
achieving a mixture with an ethanol-to-acetaldehyde mass ratio
of 2.3, close to value 3 required in the second stage.
Table 3 presents the material balance of the butadiene
synthesis section. Stream feed-2 results by combining the AcH-
S1 stream with a makeup stream MKP-S1 (see Figure 4)
8484
pubs.acs.org/IECR Article
containing 6540 kg/h ethanol 95% wt (6213 kg/h ethanol). In
total, 30,000 kg/h ethanol is obtained, so as to bring the
reactant mass ratio at the reactor inlet close to 3. Since the
recycle streams of acetaldehyde and ethanol are variable, an
ethanol makeup stream MKP-S2 is used to finely adjust the
ratio of reactants at the set value.
The material balance of reactants from Table 3 states that
the amounts of ethanol and aldehyde entering the process
must be equal to the amounts of butadiene, byproducts,
impurities leaving the process, and water formed by the
reaction and losses in purge streams.
Referring to the input of reactants, this is 3000 + 10,179 +
2570 = 42,749 kg/h. Subtracting 20,262 kg/h ethanol recycled
to stage 1 gives 22,487 kg/h as net input of reactants. On the
output side, the streams are grouped as VSS, LSS, and recycle
to S1. The VSS includes butadiene (raw BD), VOCs, and HCs.
The LSS consists mainly of water streams, waste oxygenated
impurities, and losses which refers to waste oxygenated
impurities, black oil, and unrecoverable species.
The flow rates of components can be found in the stream
report in the Supporting Information. Butadiene has a flow rate
of 11,401 kg/h that corresponds to a production of 91,215 t/
year. The HC stream includes ethylene (843 kg/h), propylene
(194 kg/h), butenes (65 kg/h), and C5+ (261 kg/h). The
oxygenated impurities are divided into carbonyl and hydroxyl
components, the first being by far more substantial, 824 kg/h,
while the second only 47 kg/h. Among the carbonyl
components, we note formaldehyde (254 kg/h), ethyl ether
(291 kg/h), butyraldehyde (132 kg/h), ketones (134 kg/h),
and ethyl acetate (13 kg/h). The C4 hydroxyl components,
such as butanols and butenols, are in tiny amounts.
Regarding the water, one may distinguish between water
generated by the reaction (WR) and water introduced in the
process (PW). The first results from the material balance
around the chemical reactor, and the second refers to the water
used for absorption and entering as ethanol solution. The PW
leaving in the LSS and VSS are PWL and PWV, respectively.
Accordingly, on the outputs side, the calculation gives VSS +
LSS + PW − (PWR + PWV) = 12,873 + 29,848 + 8548 −
(28,673 + 110) = 22,485 kg/h.
The I/O overall material balance results by summing the
headings of inputs and outputs in Table 3 at 64,721 and 64,720
kg/h, respectively. Finally, the balance of PW is obtained by
reaction water. Note that the above I/O material balances are
closed in the limit of a flow-sheeting error of 0.0001.
The yield values for stage 2 in Table 3 are obtained by
dividing the amounts formed by the reaction by the total
amount of input reactants, that is, 22,847 kg/h. Thus, the yield
of butadiene in the second stage is 50.7% kg/kg reactant.
The overall material balance shown in Table 4 results by
combining Tables 2 and 3, considering that the acetaldehyde
produced in the first stage is consumed in the second one. The
input of ethanol in stage 1 is 14,250 kg/h, to which one adds
6213 + 2570 = 8783 kg/h in the second stage, in total, 23,032
kg/h. The outputs are butadiene (11,401 kg/h), hydrogen
(468 kg/h), water (8548 kg/h), HC (1362 kg/h), oxygenated
impurities (870 kg/h), and losses (361 kg/h), in total, 23,010
kg/h. Note that closing the overall material balance was
performed manually since the simulation of the entire
flowsheet was impractical. The input/output difference is
only of 0.09%. This is reasonable from process design point of
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Table 3. Material Balance Stage 2, Butadiene Synthesis

inputs outputs
components feed-2 MKP-S2 water abs. VSS LSS recycle to S1 mass yielda
flow kg/h 43,016 2705 19,000 12,873 29,848 22,000
ethanol 30,000 2570 20,262
acetaldehyde 10,179
butadiene 11,401 0.507
water reaction 8548 0.380
water process 2,837 135 19,000 110 28,673 1,738
HC 1362 0.061
ox impurities 870 0.039
losses 304 0.014
I/O reactants 22,487 22,485
I/O overall balance 64,721 64,720
I/O water 21,972 21,973
a
kg product/kg RM fed to stage 2.

Table 4. Overall Process Material Balance Table 5. Price of Utilities

flow kg/h input output mass yieldb utility characteristics price ($/GJ)
ethanol 23,032 tower cooling water 23 to 37 °C 1.0
acetaldehyde chilled water 5 to 15 °C 3.4
butadiene 11,401 49.5 LP/MP steam 140 to 170 °C/3.5 to 8 bar 15.0
H2 468 2.0 Dowtherm 250 to 380 °C/250−290 kJ/kg 17.0
HCsa 1362 5.9 refrigeration 1 brine or glycol water −20 °C 10.0
oxygenated impurities 870 3.8 refrigeration 2 propane −40 °C 13.0
water 8548 37.1
losses 361 1.6
(oil) 5.87 MW has a potential saving of 47.9%. Figure 10
total 23,010 100.0
a
presents the heat-integrated flowsheet for ethanol dehydrogen-
Ethylene 843 kg/h. bkg product/kg ethanol. ation. The FEHEs highlight the energy saving measures. Thus,
the compressor CP-101 outlet enthalpy is used for evaporating
view and does not affect the ethanol-specific consumption, 16% of the ethanol feed in FE-202 (0.44 MW), the rest of
which is a key element in the economic analysis. ethanol being vaporized in H-106 (3.2 MW) by employing LP
In the end, the overall yield is 0.495 kg butadiene/kg steam. The remixed vapor is heated up in FE-203 (1.15 MW)
ethanol. Note that in an ideal process, the maximum yield is in counter current with the reactor effluent and then in H-101
0.587 kg butadiene/kg ethanol. Accordingly, the specific (0.09 MW). The remaining enthalpy of the reactor effluent is
ethanol consumption is 2.02 kg ethanol/kg butadiene, a used in FE-201 to bring the fresh feed to the boiling point
value that will be used in the economic analysis. (0.95 MW). Besides 3.4 MW cooling water, this stage uses
The carbon yield can be calculated as the atomic carbon 0.37 MW refrigeration (−35 °C). The total utility
passed into the product with respect to the initial amount in consumption is 7.89 MW, which is a saving of 38.5% with
the reactant. In this case, one gets the value of 84.3%, showing respect to the nonintegrated case. The hot utility duty (LPS)
that most of the carbon present in the RM is preserved and falls by 42.5% to 3.4 MW. In this way, most of the energy is
only a limited part is lost. saved.
Ethylene and propylene may be considered recoverable as Regarding the butadiene synthesis, the examination of the
olefin byproducts. The conversion of 843 kg/h ethylene that flowsheets in Figures 4, 7, and 8 reveals that the most
would give 1385 kg/h ethanol could increase butadiene interesting for energy saving are the chemical reactor zone and
production by around 6%. the ethanol recovery by distillation.
In the first case, the feed conditioning (preheating,
8. PROCESS INTEGRATION evaporation, and superheating) needs 22 MW duty that may
This section deals with the integration of process units in view be supplied by the hot reactor effluent. Figure 11 shows the
of saving energy. Pinch analysis is applied to the flowsheets flowsheet for heat recovery. Compared with Figure 2, three
developed so far.6 The solutions are constrained by economic FEHEs units are inserted for heat recovery. The first FE-101
and technological features. Table 5 presents the price of (duty 2.7 MW) is used for preheating the feed up to 85.5 °C,
utilities used in this work, where $ signifies USD. In addition, the bubble point. The FE-101 outlet stream is split and
the electric energy price is taken as 0.1 $/kW h. The prices are evaporated over the range 85.5 to 100.3 °C in two units, FE-
characteristic for Europe6 and USA27 but on the more 102 (duty 7.2 MW, process−process) and H-201 (duty 5.9
expensive side, as predicted for the next years. MW, LPS). The last unit FE-103 ensures superheating the
Concerning the ethanol dehydrogenation stage, the reactor reactor inlet stream at 330 °C (duty 6.4 MW). The cumulated
needs 4.56 MW fulfilled by using Dowtherm vapor at 5 bar and duty of the FEHE units is 16.3 MW, which represents 73.4% of
344.4 °C. The pinch analysis applied in Figure 3, excluding the the total 22.2 MW used for preheating. The cold utility usage
chemical reactor, reports an overall energy requirement of diminishes accordingly. The reactor itself needs only 3.0 MW
12.82 MW and saving potential of 43.85%, of which hot utility of heat supplied by Dowtherm. Note that the reactor design
8485 https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Figure 10. Heat-integrated flowsheet for ethanol dehydrogenation to acetaldehyde.

Figure 11. Heat-integrated reaction section for butadiene synthesis.

Figure 12. Heat-pumping energy saving for the ethanol distillation column.

ensures an outlet temperature of minimum 10 °C hotter than The distillation of the water/ethanol solution involves a
the inlet to enable heat recovery, as described above. large amount of hot utility, over 19 MW LP steam (bottoms at
8486 https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Table 6. Utilities’ Requirements and Costs6,27
unit stage 1
electricity kW 1373
chilled water MW 4.14
cooling water MW 0 4.87
steam MW 3.29
Dowtherm MW 4.66
refrigeration MW 0.38
thermal energy MW
hourly cost $/h
annual cost M$/year
1.5 bar and 102.3 °C), while an equivalent energy is lost in the
cooling water (top at 1.1 and 88.2 °C). This situation incites
the application of heat pumping (HP). Figure 12 presents the
simulation flowsheet in Aspen 11.0 to be compared with
column C-406 in Figure 8. The mixture comprises 53,370 kg/h
aqueous solution of ethanol of 44.3% wt, at 70 °C. The column
has 30 stages, with feed on stage 25. A classical setup at top
supplies (88.1 °C) 92% wt ethanol and (111.3 °C) 99.8% wt
water at bottom at a molar reflux ratio of 1.82. The reboiler
and condenser duties are 18.3 and 19.2 MW, respectively.
In the heat-pump-assisted column (Figure 12), the top
vapor, close to the azeotropic composition, enters a
compressor that increases the stream pressure from 1.1 to 5
bar, while the temperature increases at 173 °C. This is the hot
fluid providing 16.5 MW to the heat-integrated reboiler. After
condensation, the saturated liquid is depressurized to 1.1 bar,
which results in partial vaporization and temperature decrease.
An auxiliary cooler (Cond-Aux) finishes the condensation. The
rest of the necessary heat is provided by an auxiliary reboiler
(Reb-Aux, 2.1 MW), which can also be used for control
purposes. The mechanical energy required for compression is
2.9 MW. Accordingly, an excellent coefficient of performance
of 10 is achieved. The energy saving is 88.7%.
A shortcut calculation gives an idea of profitability. The
purchase compressor cost may be estimated by the relation CP
= 580,000 + 2000 × [P, kW]0.6 = 2.97 M$,6 where the
compressing power P is calculated with 0.72 isentropic
efficiency and 0.92 mechanical efficiency (default in Aspen
Plus). Considering an installation factor of 1.3 gives the
investment cost CI = 3.56 M$. The benefit of 4.81 M$/year is
the difference between the 7.13 M$/year cost of saved steam
(0.0165 GW × 8000 h/year × 15 $/GJ × 3600 s/h) and 2.32
M$/year the cost for compression (2900 kW × 0.1 $/kW h ×
8000 h/year). Neglecting the cost of the auxiliary reboiler, the
payback time (PBT) of investment is 3.56/4.81 = 0.62 years or
7.5 months. Thus, implementing the heat pump is fully
profitable.
The inspection of the remaining units (columns C-201, C-
301, C-303, C-401, C-402, C-403, and C-404) shows that only
LP steam is necessary for driving the reboilers. The cumulative
duty is 13.1 MW. This is a substantial amount. However, in a
larger frame of a biorefinery, this can be advantageously
supplied by a combined heat and power (CHP) system that
makes use of biomass waste for generating electricity, while LP
steam is extracted from back-pressure turbines. Details about
the balance of utilities are given in the Supporting Information.
9. SUSTAINABILITY ANALYSIS
Here, we examine the major sustainability metrics that may
characterize the environmental performance of a process
8487
pubs.acs.org/IECR Article
stage 2 purification process cost [$/h]
4106 89.6 5568 556.8
7.27 0 11.41 139.6
8.23 13.07 47.1
27.51 11.89 42.38 2288.3
3.1 0 7.68 470.0
0.37 2.05 2.80 130.9
77.34
3632.8
29.06
design:6 material intensity, energy intensity, water consump-
tion, global warming potential, toxicity, pollutants, and risks.
9.1. Material Intensity. The measure primarily regards the
RM consumption and the production of wastes. In Section 6,
the overall mass balance gave a mass yield of 0.495 kg
butadiene/kg ethanol. Accordingly, the specific consumption of
ethanol is 2.02 kg/kg butadiene (stoichiometric value 1.704).
A significant amount of water forms (0.371 kg/kg); this is not
waste, but a resource. Hydrogen is a saleable byproduct with a
yield of 0.02 kg/kg. If all other byproducts are considered
waste, then the material intensity would be 0.227 kg waste/kg
butadiene. An improvement can be obtained by the conversion
of some waste, namely, ethylene, in useful products.
9.2. Energy Intensity. Table 6 presents the utility needs
according to the three process stages examined so far. These
are collected and summarized for every operational unit by the
Aspen economic analyzer (AEA) (details in the Supporting
Information). Steam is the most important utility, mainly used
by stage 2 and butadiene purification. This is followed by the
reactor heating by Dowtherm and the energy for compression.
In total, the electric power is 5.57 MW, including the HP. The
total thermal duty is 77.3 MW, from which 50.1 MW is for hot
utilities (steam and Dowtherm).
A substantial part of hot energy may be covered by waste
combustion; 2593 kg/h waste (heat of combustion 35 MJ/kg,
efficiency 0.75) may produce 18.9 MW, reducing the hot utility
duty from 50.1 to 31.2 MW by 38%. Reported to metric ton of
butadiene, the energy intensity is 561.7 kW h/t electrical energy
and 15.8 GJ/t thermal utilities. The main consumer is ethanol
recovery by distillation. Improving the catalyst is desirable in
the direction of reducing the ethanol/acetaldehyde ratio at the
reactor inlet.
9.3. Global Warming Potential. The process itself does
not release greenhouse gas emissions by chemical reactions.
CO2 appears by the combustion of waste, estimated as 4760 kg
CO2/h (Supporting Information). However, the largest part of
CO2 originates from utilities. This is estimated by the AEA tool
at 13,436 kg/h. The specific CO2 emission amount is 1.60 kg/
kg.
9.4. Water. The ETB process supplies water as the
byproduct, namely, 8548 kg/h (see Table 4). To this, one
should add 1212 kg/h entering with ethanol. On the other
side, 19,000 kg/h recycled water is used as the absorption
agent. This leaves as bottoms from the distillation column for
ethanol recycling of high purity (over 99.99%) with only traces
of oxygenated impurities. As a consequence, there is no need
for costly chemical treatment. However, biological treatment
could be necessary to meet the environmental regulations.
9.5. Safety, Health, and Environment Risks. Butadiene
is considered hazardous by the Occupational Safety and Health
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Administration (OSHA): flammable, explosive mixture with
air, acute toxicity (4), germ cell mutagenicity (1B), and
carcinogenicity (1). There are also serious hazards for
acetaldehyde. Handling ethanol-concentrated solutions also
increases risk concerns. Safety, health, and environment
protection measures may demand more capital investment,
which are estimated as 5% from the inside battery limit (ISBL)
costs.
Referring to the present situation, the butadiene manufactur-
ing is dominated by refinery-based processes, mainly as a
byproduct by the steam cracking of HC fractions. From
sustainability viewpoint, the best processes are based on butane
dehydrogenation (Houdry) and butene oxy-dehydrogenation
(OXO-D process), however, marginal as market share. The
specific consumptions are 1.68 and 1.115 kg/kg corresponding
to material intensity of 0.64 and 0.07, respectively.28 It may be
observed that the proposed process generates much less waste
than butane dehydrogenation, and it has in general much
better material intensity, which controls the economics,
compared with the petrochemical counterparts.
10. COMPARISON WITH PREVIOUS WORKS
Some works claim that the one-stage process is superior to the
two-step process. Thus, using data issued from some patents
and a private report, it was argued that specific ethanol
consumption is very disadvantageous for the two-stage process
with respect to the one-stage process, 3.75 versus 2.06 t/t,
although the energy requirements were similar.29 From our
viewpoint, the assessment is questionable in the absence of
process design details, including material and energy balances.
A more recent paper published an analysis of a one-stage
process employing a proprietary highly active and selective Au/
MgO−SiO2 catalyst.30 The paper presents a simplified process
diagram that was submitted to simulation in Aspen Plus.
Optimal selectivity is reached at 350 °C and low ethanol
conversion (0.25), so there is presumably a substantial recycle
of ethanol. The specific consumption of 1.818 kg ethanol/kg
butadiene seems to be excellent, with a low ethylene formation
of 0.014 kg/kg. The hydrogen coproduct is 0.078 kg/kg
butadiene, around 4 times more than in our work. The energy
(thermal) requirements are only 6.18 MJ/kg (32% in reaction
and 68% in separations), which seems to be largely
underestimated. As before, in the absence of detailed process
design data, a quantitative evaluation cannot be made.
A recent research presents the assessment of the one-stage
process employing a Hf−Zn-based catalyst.31 The paper
focuses on economic and life cycle assessments for
implementation in the EU, Brazil, and the USA, compared
with the naphtha-cracking route. Experimental data obtained in
laboratory with pure ethanol32 are combined with similar
experiments with ethanol/water solution 92.5% wt and used
further as the RM in design. The chemical reactor assessment
is tackled by using a yield description. The stoichiometric
model considers, beside the key components, the formation of
byproducts, impurities, and waste. Unconverted ethanol is
recycled to the reactor as azeotrope solution. The acetaldehyde
is treated in two alternatives, either as the product or recycled
form. In the absence of kinetic equations, the species
selectivities are specified by regression equations as functions
of species concentrations. The Supporting Information
presents a flowsheet simulated in Aspen Plus. The flowsheet
shows a complex structure of separations, but no details are
given about the conceptual approach and modeling, simu-
8488
pubs.acs.org/IECR Article
lation, sizing, and performance of units. The mass balance with
reactants’ recycling gives an ethanol-specific consumption of
2.62 kg ethanol/kg butadiene and yields of C2/C3 olefin 0.246
kg/kg, butenes 0.062 kg/kg, H2 0.042 kg/kg, waste for fuel
0.235 kg/kg, the rest being water formed by the reaction. It
may be noted that the material consumption is 30% higher
than here, which would significantly affect the process
economics. The paper also deals with energy saving. Natural
gas is used for ethanol conditioning and reactor heating. The
utility consumption is 10.8 GJ/h and 14.7 GJ/t for hot and
cold utilities, as well as 500 kW h electricity per ton butadiene.
In our case, the energy consumption is somewhat higher due
to a large ethanol recycle and due to energy spent for
impurities’ removal. However, this is largely compensated by
better material intensity.
11. ECONOMIC ANALYSIS
The goal of this section is evaluating capital and operation
costs of the proposed technology and assessing the profit-
ability. The work is assisted by the AEA 11.0. The Supporting
Information presents the data centralized at the levels of stage
1, stage 2, purification, and whole process. The assessment
procedure is presented elsewhere.6
Table 7 shows the results obtained for the ISBL capital and
utility costs. First, the purchase equipment cost (PCE) is
Table 7. Sustainability Measures per Metric Ton of
Butadiene
measure unit value
specific consumption ethanol tonne 2.020
material intensity (waste) tonne 0.227
carbon-atoms yield % 84.3
electric energy kW h 561.7
hot utility (steam, Dowtherm) GJ 15.8
process water m3 1.67
cooling water m3 18.3
chilled water m3 29.3
refrigeration tonne 18.8
estimated for each equipment item. Then, installation cost
(IC) and direct cost (DC) are obtained by following the
internal AEA database. In the first place, the calculations are
done for stages 1, 2, and purifications for the nonheat-
integrated flowsheets. Then, the problem of saving energy by
HI around the chemical reactors is solved, which captures most
of the hot energy needs. In addition, for stage 2, the energy
saving by HP is accounted as well.
The items involved in the described flowsheets and designed
by computer-aided methods are designated as the base process
equipment (BPE). Typically, integrated separation units, such
as distillers, heat exchangers, compressors, and standard
reaction vessels are the key contributors to the capital
investment and the costs of utilities. It is interesting to note
that the final obtained ratio of DC/PCE of 3.02 is not far from
the factorial method used in academic books, which is 3.6 for
essentially fluid-type processes.
The data from Table 8 highlight the relation of energy saving
to the capital investment. Thus, for the non-HI stage, the PCE
is 2.11 M$, IC 3.56 M$, and DC 7.11 M$, while utility cost is
4.20 M$/year. A direct capital investment of 8.84 − 7.11 =
1.73 M$ has as a result cost-saving in utilities of 4.20 − 3.10 =
1.1 M$/year, or 26.1%. Similarly, in stage 2, an investment of
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research pubs.acs.org/IECR Article

Table 8. Estimation of ISBL and Utility Costs for the ETB Two-Stage Process (in M$)
stage purchased equipment costs installation costs direct costs utility costs energy cost saving % payback time (year)
stage 1 non-HI 2.11 3.56 7.11 4.20
stage 1 HR 2.74 4.56 8.84 3.10 26.1 1.59
stage 2 non-HI 6.767 15.1 29.1 34.26
stage 2 HR 7.45 16.54 33.54 23.99 30.0 0.43
stage 2 HR + HP 11.41 20.67 41.45 19.73 42.4 0.850
purifications 2.19 3.87 7.67 6.22
BPE cost 16.34 29.10 57.96 29.06
ISBL 19.61 34.92 69.56 33.4

12.35 M$ for HI and HP save 14.52 M$/year energy, or
42.4%. The short PBTs for recovering the investment fees
emphasize the efficiency of these measures. Putting together
the investment of the two stages gives a total capital of 57.96
M$. In addition, we should consider supplementary equipment
not included in simulation, such as storage and various vessels,
Dowtherm system, impurities, and VOC treatment, as well as
in situ catalyst regeneration and wastewater treatment. These
fees may be included in a factor of 20% over the BPE cost that
leads to an ISBL capital of 69.6 M$. Similarly, by including a
factor of 15% over the utility cost for BPE (29.06 M$), one
gets 33.4 M$ for the total cost of utilities.
Next, the estimation of the total capital investment (TCI)
may be done following established methods.6 We consider the
off-site costs (outside battery limit) OSBL = 0.45 × ISBL,
engineering, construction, and contingency costs as 0.25 ×
(ISBL + OSBL), which results in a fixed capital cost FC = ISBL
× (1 + 0.45) × (1 + 0.25) = 126.2 M$.
TCI includes fixed capital, as well as working and starting
capital contributions. The working capital includes fees for
purchasing ethanol, catalysts, adsorbent (activated carbon,
silica), DMF, HC solvent and hydrazine reactant (see also the
Supporting Information). The ethanol inventory considers 2
months of operation, or 16 M$. For each catalyst, an inventory
of 100 t covers two reactor fillings. Considering prices of 6000
$/t for CuO/Cr2O3 and 10,000 $/t for Ta2O5 gives catalyst
expenses of 1.6 M$. Similarly, for active carbon and DMF fees
one gets for each 0.1 M$. Including 0.2 M$ fees for other
chemicals gives a total cost estimation for catalysts and
chemicals of 2 M$. The starting costs and liabilities provisions
are taken to be 7.2 M$. The resulting working and startup
expenses are 25.2 M$ or 20% from FC. The final TCI is then
151.4 M$ (see also the Supporting Information).
The annual operation costs consist of fixed operation cost
(FOC) and variable operation cost. The FOC refers to the
labor, maintenance, overhead charges, insurance, and taxes.
The variable operation cost refers to the cost of RM,
consumable materials (CMs), as catalysts, solvents, adsorbent,
solvents, chemicals for water treatment, and so forth, as well as
the cost of utilities.
The FOC is estimated as 10.2 M$, of which the DC is 6.0 M
$ and indirect cost is 4.2 M$.
The RM cost primarily regards the purchasing cost of
ethanol. Considering the specific ethanol consumption of 2.02
kg/kg leads to an input ethanol rate of 179,897 t/year. The
estimated CM fees are 2.5 M$/year (approx. 3% from RM),
similar to those affected in the working capital (2 M$), to
which one may add fees for catalysts and adsorbent
regeneration (0.5 M$/year). Details are given in the
Supporting Information.
We also consider a credit for commercializing hydrogen.
The cost of green hydrogen today is evaluated between 2.50
and 6.80 $/kg.33 The production rate of 468 kg/h corresponds
to 3744 t/year or 4.1% from butadiene. Assuming a price of
4000 $/t gives a sale credit of 15 M$/year. At a mean
butadiene price of 1500 $/t, the hydrogen could bring 11%
supplementary revenue.
The above elements allow us to analyze the profitability of
the designed process. As a profitability measure, we consider
the pretax return on investment (PT-ROI). First, the gross
profit (GP) is obtained as the difference between revenues and
operation costs. Note that the GP appears on a company’s
income statement and can be calculated by subtracting the cost
of goods sold from the revenue (sales).34 Then, PT-ROI is
obtained as the GP divided by the TCI
PT‐ROI = GP/TCI × 100
= (revenues − operation costs)/TCI × 100
(4)
Thus, the PT-ROI depends only on the process design
efficiency, avoiding the inclusion of debatable accounting
elements, such as depreciation policy, taxation, the time value
of money. The pretax rate of return is the measure most cited
for investments in the financial world, while the return on the
invested capital ratio gives a sense of how a company makes
use of the capital to generate profits.34
Figure 13 shows a plot of the butadiene price versus ethanol
price in several scenarios. The region of interest is delimited by
the probable market prices. Currently, cheaper ethanol may be
found in the USA and Brazil, that is, 500−600 $/t, while in
Figure 13. Assessment of the profitability by butadiene manufacture
from ethanol.

8489 https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Europe, the best prices are 600−700 $/t.2 At present, the
butadiene price is correlated with the oil price since this comes
from oil at 95%. In the period from 2008 to 2018, lower
butadiene prices (around 1000 $/t) and higher prices (around
2000 $/t) were recorded for oil barrels between 40 and 60 $/t,
with spot prices up to 4000 $/t.4 An average value of 1520 $/t
over the same period is considered in the recent analysis of a
one-stage process.32
The minimum selling price (dot line) corresponds to zero
GP. A PT-ROI value of 15% should ensure good profitability.
Figure 13 also shows how the competitive performance of the
process can be improved. In the base case (BC), when the
production costs are the highest, the minimum butadiene
selling price is 1275 $/for ethanol price of 500 $/t, but it rises
over 1682 $/t at an ethanol price of 700 $/t.
When referring to PT-ROI 15%, in the BC, a higher ethanol
price, say 700 $/t, leads to a butadiene price of 1931 $/t.
Therefore, a reduction of the production costs is necessary.
This goal can be achieved by dropping the utility price and/or
the capital investment, as well as reducing the specific ethanol
consumption by employing more selective catalysts. The
results are presented in Figure 13. As an example, when
reducing the utility cost by 30%, TCI by 20%, and specific
ethanol consumption by 10%, the butadiene price drops by 7.2,
3.5, and 4%, respectively. Utilities may be reduced by tighter
HI, lower ethanol recycle, or cheaper utilities, for example, by
placing the plant in a biorefinery site and employing a CHP
system. Moreover, the capital cost may be reduced by sharing
storage and effluent treatment. These actions would bring an
advantage of 10.5%. Reducing the ethanol consumption is
more challenging since the actual value is not far from the
stoichiometric limit 1.704 kg ethanol/kg butadiene. However,
a decisive advantage may be taken from a lower ethanol price.
Indeed, the proposed process does not need high-purity
ethanol but only solutions of 90 to 95%. This is illustrated by
shifting the ethanol cost from 500 to 400 $/t, which allows
dropping the butadiene price by 13%.
We may also determine an after-tax ROI (AT-ROI). The
gain may be associated with the annual cash flow resulting
from the operation. The following equation applies
AT‐ROI = [GP × (1 − t R ) + D × t R ] /TCI × 100 (5)
We consider a linear depreciation of TCI over 10 years and a
tax rate of 35%. For the prices of ethanol and butadiene
obtained above, 500 and 1523 $/t, respectively, the ROI is
13.25% (Supporting Information). An AT-ROI value over 10%
is particularly good, as historically, the average ROI for the
S&P 500 companies has been around 10% per year.34
12. DISCUSSION OF RESULTS
1. In the first process stage, ethanol dehydrogenation,
acetaldehyde is obtained with high selectivity (100%) in
a multitubular reactor operating at 280 °C and 3 bar by
employing the CuO/Cr2O3 catalyst. The heat for the
endothermic reaction is supplied by a Dowtherm vapor
system. A catalyst productivity of 0.258 kg acetaldehyde/
kg catalyst·h is achieved. Hydrogen is obtained as a
byproduct with 99.8% wt purity. With an overall mass
yield of only 2%, hydrogen may bring a credit of around
6% in the RM costs. A sequence of flashes separates the
acetaldehyde by absorption in ethanol aqueous solution
of 91% wt recycled from the second stage (Figure 2). In
8490
pubs.acs.org/IECR Article
this way, an ethanol/acetaldehyde mixture of mass ratio
2:1 is prepared in stage 1 for stage 2. By building a heat
recovery scheme only around the chemical reactor
(Figure 10), an overall energy saving of 38.5% is
obtained (hot utility saving of 42.5%), which is 87.7%
from the maximum saving determined by pinch analysis.
2. The butadiene synthesis from ethanol and acetaldehyde
makes use of the performant Ta2O5 catalyst. The
reaction conditions are 330 °C and 3 bar with a mass
ratio of ethanol to acetaldehyde of 3:1. The catalytic
reactor design considers a comprehensive kinetic model
(27 parallel consecutive reactions and 24 species) that
delivers a realistic description of the outlet reaction
mixture. The simulation shows that 60% of butadiene
originates through aldol condensation via CRA and 40%
via MVK. The first route leads to 1-butene that may be
removed by simple distillation, while the second route
gives 2-cis-butene that needs extractive distillation. The
complex reaction network reveals the formation of many
byproducts and impurities as saturated and unsaturated
HCs and oxygenated species (alcohols, aldehydes,
ketones, ethers, and esters), all mixed with large amounts
of water. These are characterized by a wide volatility
range and occurrence of many azeotropes. Finally, the
material balance with closed recycles reveals mass yields
(in kg product/kg ethanol) of 50.7% butadiene, 38.5%
water, 4.7% C2/C3 olefins, 4.2% oxygenated species,
and 1.2% waste. The catalyst productivity is 0.253 kg
butadiene/kg cat·h.
3. Referring to the second stage, the work develops an
original and complex flowsheet for the separation system
by employing a systematic task-oriented methodology.7
The flowsheet is validated by computer simulation
(Figures 4, 7−9). The FSS that makes use of reboiled
striping column divides the process synthesis problem
into solving two subsystems, vapor and liquid separa-
tions (VSS and LSS). The VSS tackles the separation of
raw butadiene and HCs. Absorption in water is used for
capturing the entrained acetaldehyde together with other
oxygenated impurities to be treated in the LSS. After
vapor compression, cleaning, and drying, ethylene and
propylene are recovered as byproducts by cryogenic
distillation. The raw butadiene already has a good purity
of 99.1% with butene impurities (1-butene and 2-cis-
butene). The absence of vinyl acetylenes tremendously
simplifies the purification process. A final purity over
99.7% is obtained by extractive distillation with DMF.
4. The LSS handles the recovery and recycling of ethanol
and acetaldehyde to the synthesis reactor. A critical task
is the removal of numerous impurities in suitable exit
streams to prevent accumulation in recycles. The
solution provides different paths for acetaldehyde and
ethanol recycling, the first carrying the light volatile
impurities, while the second carries the higher boilers.
The removal of diethyl ether, an intermediate boiler, is
difficult and may cause losses in acetaldehyde, as
reported in earlier patents. We found that the ethyl
ether can be removed effectively by extractive distillation
with paraffinic solvent (dodecane or similar) in an
absorption/striping device. The heavier oxygenated
impurities, such as ketones and alcohols, are concen-
trated in the ethanol top distillate, from which it can be
removed by chemical conversion. The solution proposed
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
is the treatment with hydrazine that transforms the
carbonyl species to benign HCs. The remaining small
amounts of impurities can be removed by adsorption
over activated carbon.
5. The butadiene synthesis is an energy-intensive step,
namely, for supporting the chemical reaction, recycling
the reactants, and removing byproducts and impurities.
Thus, 73.4% of the total 22.2 MW used for feed
preheating can be saved by FEHE units. Ethanol
distillation is an energy-intensive operation, but the
application of heat-pumping allows reducing the LP
steam consumption by 88.7% while recovering the
investment in only 7.5 months. The water resulting from
processing contains only traces of organic impurities and
does not need costly treatment.
6. Butadiene purification by extractive distillation is costly.
However, because the vinyl acetylenes are absent, this
step is much less capital and energy-intensive than in a
petrochemical process where butadiene is produced by
steam cracking.
7. The sustainability measures demonstrate that the
proposed process is ecoefficient compared to petro-
chemical alternatives. The butadiene carbon yield of
83.7% is high, while the material intensity reflecting the
waste generation is moderate (0.277 kg/kg). The
specific consumption of 2.02 kg ethanol/kg butadiene
is rather low and contributes the most to good
profitability. The electrical energy consumption of
581.7 kW h/t and 15.3 GJ/t thermal energy are
moderate but significant in the product price. CO2
emissions are limited to the utilities’ production and to
the waste combustion.
8. The economic analysis reveals an ISBL cost of 69.6 M$
and TCI of 151.4 M$ corresponding to 763 and 1660
$/t, respectively, for a plant capacity of 91,000 t/year. In
the operation costs, the largest share belongs to RM
(71%), followed by utilities (21%) and fixed costs (8%).
9. Framing the economic analysis in a probable interval of
ethanol and butadiene prices shows a linear dependence.
In the base case, an ethanol price of 500 $/t (typical for
US and Brazil markets) may ensure a butadiene price of
around 1500 $/t with a pretax ROI of 15%. Significant
profitability improvement can be obtained by reducing
the cost of utilities, capital investment, and ethanol-
specific consumption. Placing the process in the frame of
a biorefinery offers cheaper utilities by implementing a
CHP cycle, as well as lower off-site capital costs. Thus,
the reduction by 30% in utilities’ cost and by 20% in
TCI cost results in dropping of the butadiene price by
10%. In this way, the process may become profitable for
higher ethanol prices, as 600 to 700 $/t, typical for
European markets.
10. The optimization of all design parameters is outside the
scope of this work. The flowsheets developed by
applying the methodology described in Section 2 may
be considered to be located in the optimal region,
although not obtained by a mathematical procedure. In
the first place, we optimized the operation conditions
and catalyst amount for achieving the best butadiene
yield in the reactor−separator−recycle structure. Pro-
cessing the reaction mixture was greatly simplified by
dividing the separations in the interconnected VSS and
LSS. The selection of separation methods was guided by
8491
pubs.acs.org/IECR Article
identifying suitable characteristic properties, for example,
chemical conversion of carbonyl components versus
physical separation. We optimized the performance of
each distillation/absorption column in terms of stages
and internals. The energy usage was minimized by
applying process integration methods. The economic
analysis showed that the efficient use of the RM (high
butadiene yield and limited impurities) is the most
important factor in maximizing the process profitability.
13. CONCLUSIONS
This work demonstrates by rigurous process design that
producing butadiene from ethanol by a two-stage process can
be competitive with the current oil-based technologies, as well
as with one-stage processes. In fact, the use of distinct catalysts
for ethanol dehydrogenation to acetaldehyde and for butadiene
synthesis from ethanol and acetaldehyde aims to achieve the
highest yield at each reaction stage. Thus, ethanol dehydrogen-
ation with the CuO/Cr2O3 catalyst is practically 100%
selective, while a carbon yield of 83.7% is reached in the
butadiene synthesis step by using an efficient Ta2O5/SiO2
catalyst. In addition, the metal oxide catalysts are cheaper and
easier to regenerate than metal/metal oxide catalysts.
The results of the design are comprehensive flowsheets,
material and heat balances, and key unit sizing, based on
detailed kinetic modeling of reactors, systematic development
of separation systems, advanced methods for energy saving,
and state-of-the art computer simulation. The economic
analysis shows that the contribution of the dehydrogenation
stage in the overall capital and operation costs is limited to
10%, which is compensated by higher yield achieved by the
butadiene synthesis stage. In addition, the hydrogen may bring
a credit of 6% in the feedstock cost. Compared to one-step
processes, the overall process is characterized by lower specific
ethanol consumption, 2.02 kg ethanol/kg butadiene and
reduced amounts of byproducts and impurities. By applying
efficient saving measures including heat pumping, the energy
intensity is substantially reduced to 16 MJ/kg thermal energy
and 0.6 kW h/kg electricity. The process profitability can be
boosted by implementation in a biorefinery frame that secures
lower prices for feedstock and utilities.
■
ASSOCIATED CONTENT
* Supporting Information
sı
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.iecr.1c00958.
Kinetic models; design of the ethanol dehydrogenation
reactor; material balances; sizing of distillation columns
capital and operating costs; butadiene prices; and
economic analysis (PDF)
■
AUTHOR INFORMATION
Corresponding Author
Costin Sorin Bildea − University Politehnica of Bucharest,
011061 Bucharest, Romania; orcid.org/0000-0001-
7707-1366; Email: sorin.bildea@upb.ro
Authors
Alexandre C. Dimian − University Politehnica of Bucharest,
011061 Bucharest, Romania
Nicoleta Ilona Bezedea − University Politehnica of Bucharest,
011061 Bucharest, Romania
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492
Industrial & Engineering Chemistry Research
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.iecr.1c00958
Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Statista. Production capacity of butadiene worldwide in 2018
and 2023. Available at https://www.statista.com (accessed Mar 02,
2021).
(2) American Chemistry Council (ACC). Butadiene product
stewardship guidance manual. 2019, available at https://www.
americanchemistry.com (last accessed Aug 26, 2020).
(3) Grub, J.; Loesser, E. Butadiene. Ullmans’ Encyclopedia of
Industrial Chemistry; Wiley-VCH, 2011.
(4) Pomalaza, G.; Arango Ponton, P.; Capron, M.; Dumeignil, F.
Ethanol-to-butadiene: the reaction and its catalysts. Catal. Sci. Technol.
2020, 10, 4860−4911.
(5) Dimian, A. C. Renewable raw materials, chance and challenge for
computer-aided process design engineering. Comput.-Aided Chem.
Eng. 2007, 24, 309−318.
(6) Dimian, A. C.; Bildea, C. S.; Kiss, A. A. Integrated Design and
Simulation of Chemical Processes; Elsevier, 2014.
(7) Alternative Fuels Data Center. Ethanol manufacturing; US
Department of Energy. Available at https://afdc.energy.gov (last
accessed Aug 14, 2020).
(8) Ostromislensky, J. Production of butadiene. J. Russ. Phys.-Chem.
Soc. 1915, 47, 1472−1506.
(9) Lebedev, S. V. Preparation of bivinyl directly from alcohol. I. J.
Gen. Chem. 1933, 3, 698−717.
(10) Angelici, C.; Weckhuysen, B. M.; Bruijnincx, P. C. A.
Chemocatalytic conversion of ethanol into butadiene and other
bulk chemicals. ChemSusChem 2013, 6, 1595−1614.
(11) Makshina, E. V.; Dusselier, M.; Janssens, W.; Degrève, J.;
Jacobs, P. A.; Sels, B. F. Review of old chemistry and new catalytic
advances in the synthesis of butadiene. Chem. Soc. Rev. 2014, 43,
7917−7953.
(12) Pomalaza, G.; Capron, M.; Ordomsky, V.; Dumeignil, F.
Recent breakthroughs in the conversion of ethanol to butadiene.
Catalysts 2016, 6, 203.
(13) Qi, Y.; Liu, Z.; Liu, S.; Cui, L.; Dai, Q.; He, J.; Dong, W.; Bai,
C. Synthesis of 1,3 butadiene and its substituted monomer for
synthetic rubbers. Catalysts 2019, 9, 97.
(14) Dussol, D.; Cadran, N.; Laloue, N.; Renaudot, L.; Schweitzer,
J.-M. New insights of butadiene production from ethanol: Elucidation
of concurrent reaction pathways and kinetic study. Chem. Eng. J. 2020,
391, 123586.
(15) Franckaerts, J.; Froment, G. F. Kinetic study of the
dehydrogenation of ethanol. Chem. Eng. Sci. 1964, 19, 807−818.
(16) Peloso, A.; Moresi, M.; Mustachi, C.; Soracco, B. Kinetics of
the dehydrogenation of ethanol to acetaldehyde on unsupported
catalyst. Can. J. Chem. Eng. 1979, 57, 159−164.
(17) Li, C.-H.; Finlayson, B. A. Heat transfer in packed bedsa re-
evaluation. Chem. Eng. Sci. 1977, 32, 1055−1066.
(18) Dow Chemicals. Dowtherm RP, Heat Transfer Fluid, product
technical data. Available at https://www.dow.com (last accessed Jan
20, 2021).
(19) Dastillung, R.; Fischer, B.; Jacquin, M.; Huyghe, R. Method for
the production of butadiene and hydrogen from ethanol in two low-
water and low-energy consumption reaction steps. U.S. Patent
10,358,396 B2, 2019.
(20) Dimian, A. C.; Groenendijk, A. J.; Iedema, P. D. Recycle
interaction effects in a complex plant. Ind. Eng. Chem. Res. 2001, 40,
5784−5794.
(21) Yang, R. T. Adsorbents: Fundamentals and Applications; Wiley-
Interscience, 2003.
8492
pubs.acs.org/IECR Article
(22) Blasi, M.; Barbe, J.-C.; Maillard, B.; Deleuze, H. New
methodology for removing carbonyl compounds from sweet wines.
J. Agric. Food Chem. 2007, 55, 10382−10387.
(23) Dimian, A. C.; Bildea, C. S. Component inventory control. In
Integrating Design and Control; Seferlis, P., Giorgiadis, M., Eds.;
Computer-Aided Chemical Engineering; Elsevier, 2004; Vol. 17, pp
401−430.
(24) Mathias, P. M.; Richard Elliott, J., Jr.; Klamt, A. Butadiene
Purification Using Polar Solvents. Analysis of Solution Nonideality
Using Data and Estimation Methods. Ind. Eng. Chem. Res. 2008, 47,
4996−5004.
(25) Tripathi, N.; Xu, Q.; Palanki, S. Modeling and Simulation of the
1,3-Butadiene Extraction Process at Turndown Capacity. Chem. Eng.
Technol. 2019, 42, 2649−2657.
(26) Kim, Y.; Kim, S.; Lee, B. Simulation of 1,3-butadiene extractive
distillation process using N-methyl-2-pyrrolidone solvent. Korean J.
Chem. Eng. 2012, 29, 1493−1499.
(27) Turton, R.; Bailie, R. C.; Whiting, W.; Shaeiwitz, J. A.;
Bhattacharyya, D. Analysis, Synthesis and Design of Chemical Processes;
Pearson, 2013.
(28) Chauvel, A.; Lefebvre, G.; Leprince, P. Petrochemical Processes;
Editions Technip, 1989.
(29) Cespi, D.; Passarini, F.; Vassura, I.; Cavani, F. Butadiene from
biomass, a life cycle perspective to address sustainability in the
chemical industry. Green Chem. 2016, 18, 1625−1638.
(30) Shylesh, S.; Gokhale, A. A.; Scown, C. D.; Kim, D.; Ho, C. R.;
Bell, A. T. From Sugars to Wheels: The Conversion of Ethanol to 1,3-
Butadiene over Metal-Promoted Magnesia-Silicate Catalysts. Chem-
SusChem 2016, 9, 1462−1472.
(31) De Baerdemaeker, T.; Feyen, M.; Müller, U.; Yilmaz, B.; Xiao,
F.-S.; Zhang, W.; Yokoi, T.; Bao, X.; Gies, H.; De Vos, D. E.
Bimetallic Zn and Hf on silica catalysts for the conversion of ethanol
to 1,3-butadiene. ACS Catal. 2015, 5, 3393−3397.
(32) Cabrera Camacho, C. E.; Alonso-Fariñas, B.; Villanueva
Perales, A. L.; Vidal-Barrero, F.; Ollero, P. Techno-economic and
life-cycle assessment of one-step production of 1,3-butadiene from
bioethanol using reaction data under industrial operating conditions.
ACS Sustainable Chem. Eng. 2020, 8, 10201−10211.
(33) Hydrogen Council. Path to hydrogen competitiveness. A cost
perspective. Available at https://hydrogencouncil.com (last accessed
Feb 19, 2021).
(34) Investopedia. Pretax of Return. https://www.investopedia.com
(last accessed Feb 20, 2021).
(35) Cheong, J. L.; Shao, Y.; Tan, S. J. R.; Li, X.; Zhang, Y.; Lee, S. S.
Highly active and selective Zr/MCF catalyst for production of 1,3-
butadiene from ethanol in a dual fixed bed reactor system. ACS
Sustainable Chem. Eng. 2016, 4, 4887−4894.
(36) Kim, T.-W.; Kim, J.-W.; Kim, S.-Y.; Chae, H.-J.; Kim, J.-R.;
Jeong, S.-Y.; Kim, C.-U. Butadiene production from bioethanol and
acetaldehyde over tantalum oxide-supported spherical silica catalysts
for circulating fluidized bed. Chem. Eng. J. 2015, 278, 217−223.
(37) Dimian, A. C.; Bildea, C. S. Sustainable process design for
manufacturing acrylic acid from glycerol. Chem. Eng. Res. Des. 2021,
166, 121−134.
https://doi.org/10.1021/acs.iecr.1c00958
Ind. Eng. Chem. Res. 2021, 60, 8475−8492