Journal of Flow Chemistry

https://doi.org/10.1007/s41981-018-00029-2

FULL PAPER

A new reactor concept for the combined production of ammonia

and methyl ethyl ketone
Roozbeh Ghani 1 & Davood Iranshahi 1

Received: 18 November 2018 / Accepted: 28 December 2018

# Akadémiai Kiadó 2019

Abstract
In this study, simultaneous production of ammonia and methyl ethyl ketone (MEK) in a multi tubular thermally coupled reactor is
presented. Based on this new configuration, the released heat from the ammonia synthesis reaction as an extremely exothermic
reaction in the inner tube is employed to supply the required heat for the endothermic 2-butanol dehydrogenation reaction in the
outer tube. MEK and hydrogen are produced by the dehydrogenation reaction of 2-butanol and the produced hydrogen is used to
supply 30.72% of the required hydrogen for the ammonia synthesis. Furthermore, in spite of the conventional ammonia synthesis
and 2-butanol dehydrogenation plants, interstage coolers and furnace are not required for the proposed configuration. Therefore,
operational costs, energy consumption and furnace emissions like CO, CO2 and NOx are significantly decreased. It should be also
stressed that the multi-objective optimization is employed to enhance the performance of the reactor with the aid of maximizing
the summation of the production rate and yield for each side of the reactor. Besides, the effect of the main process parameters
variation on the reactor performance has been studied.

Keywords Energy consumption minimization . Ammonia synthesis process . Methyl ethyl ketone (MEK) production . 2-butanol
dehydrogenation . Thermally coupled reactor . Multi-objective optimization

Introduction pressures using iron-promoted or recently ruthenium-

based catalysts [5–7]. The reaction is reversible and re-
Ammonia synthesis leases plenty of heat. Therefore, the existence of a re-
actor cooling system or a cooling approach to remove
Ammonia as one of the most valuable chemical mate- the released heat by the reaction is essential [8]. In
rials is widely used in the chemical and agricultural addition, auto thermal reactors which employ the hot
industries. It plays a big role in manufacturing fertil- reactor effluent to heat up the cold reactor feed to the
izers, explosive materials, and other organic nitrogen required reaction temperature, can be considered as an
compounds. It is also used to produce inorganic prod- effective method to utilize the released heat of the exo-
ucts such as nitric acid and sodium nitrate [1–4]. thermic reaction. The reactor feed can be heated up in a
Ammonia is mainly produced directly from nitrogen separate heat exchanger or by integration of the regen-
and hydrogen by a well-known industrial process called erative or recuperative heat exchange into the reactor.
Haber-Bosch. It is known as the main industrial method The integration of the heat exchange into the reactor
for ammonia production. In this process, hydrogen and has some advantageous over the auto thermal reactors
nitrogen are subjected to high temperatures and high with a separate heat exchanger [9]. Several studies have
worked on the ammonia synthesis process. Babu and
Angira simulated an auto-thermal ammonia synthesis re-
actor by employing the Quasi-Newton method to obtain
* Davood Iranshahi the optimum value of the reactor length at the various
[email protected]
top temperatures [10]. Also, Carvalho et al. proposed a
1
Department of Chemical Engineering, Amirkabir University of
new algorithm to find a solution for the optimization-
Technology (Tehran Polytechnic), No. 424, Hafez Avenue, constrained differential equations in an autothermal am-
Tehran 15914, Iran monia synthesis reactor [2]. In 2017, Muhammad Aziz
 J Flow Chem

et al. presented a high energy efficiency system which coupling process, the endothermic reaction’s required heat is
integrates nitrogen production, ammonia synthesis, and supplied by the released heat from the exothermic reaction.
power generation processes to convert hydrogen to am- Thermally coupled reactors enable us to employ highly exo-
monia by employing process integration technique as thermic reactions to be coupled with appropriate endothermic
well as utilizing the produced heat of the ammonia synthesis reactions resulting in reasonable temperature profiles and con-
reaction [7]. Recently, Humphreys et al. introduced a versions while autothermal reactors can be used only for
BaZr0.1Ce0.7Y0.2O3 − δ (BZCY) proton conducting support to weakly to moderately exothermic reactions [9]. In addition,
promote the catalytic activity of Ni towards ammonia the thermal coupling can increase the reaction rate as well as
synthesis from hydrogen and nitrogen by employing the reactor equilibrium conversion when both the reactions in
no gas purification [6]. the endothermic and exothermic sides of the reactor are re-
versible. Simultaneous production of different products as
MEK production well as preventing the formation of unwanted products in
the process are another advantages of this type of reactors
MEK is the most significant commercially manufactured ke- compared to the autothermal reactors [23–27].
tone. The main usage of MEK is as an important industrial There are several studies on the simulation and optimiza-
solvent which can dissolve plastics, resins, paints and adhe- tion of the thermally coupled reactors by employing various
sives. MEK and hydrogen are produced by the dehydrogena- reactions and optimization methods. Patel et al. suggested a
tion reaction of 2-butanol. It is an endothermic catalytic reac- thermally coupled membrane reactor in which the steam
tion which requires a certain amount of heat to occur and uses reforming of methane is considered as the endothermic reac-
copper, zinc, or brass as the catalyst [11–14]. Gas dehydroge- tion which is coupled with the methane combustion [28].
nation and liquid dehydrogenation are the two different ways Afterward, Ramaswamy et al. studied the behavior of the adi-
for the dehydrogenation of 2-butanol, but due to lower invest- abatic packed bed coupled reactors in both the steady-state
ment requirements and simpler process, gas dehydrogenation and dynamic conditions [29]. Recently, in 2017, Iranshahi
is the more common method [15]. Several studies have been et al. presented a moving bed thermally coupled reactor based
carried out in attempts to improve the dehydrogenation pro- on the coupling of naphtha reforming with the exothermic
cess of 2-butanol. Keuler et al. investigated the dehydrogena- reaction of toluene hydrodealkylation [24]. Also, Saeedi
tion reaction of 2-butanol in a Pd–Ag membrane reactor and et al. proposed a thermally coupled reactor in which the cata-
compared the simulation results to a plug flow reactor at a lytic naphtha reforming is coupled with the oxidation of sulfur
specific temperature range. The membrane reactor showed dioxide as the exothermic reaction. Furthermore, they applied
better performance from the 2-butanol conversion view point multi-objective optimization to maximize the production rates
[16]. Ravi et al. presented a kinetic study of the 2-butanol of hydrogen and aromatics and obtain the minimum value of
dehydrogenation reaction on a zinc oxide catalyst in a light-ends [30].
continuous stirred tank reactor. The experiments were This work aims to integrate the ammonia synthesis and the
done in the temperature range from 400 °C to 490 °C 2-butanol dehydrogenation processes in a novel thermally
[17]. Also, Keuler et al. investigated the dehydrogena- coupled reactor concept in which the required heat for
tion reaction of 2-butanol in which Cu-based catalysts the dehydrogenation reaction is supplied by the ammo-
had been used. They also studied that how the 2- nia synthesis reaction. As a result, this configuration
butanol dehydrogenation reaction is affected by param- leads to a significant decrease in operational costs and
eters like temperature, the feed flow rate and particle energy consumption due to the elimination of inter-stage
size [18]. Afterward, Wang et al. employed the impreg- coolers and furnace which were essential for the ammo-
nation method to prepare Cu-ZnO-Cr2O3/SiO2 catalysts nia synthesis and the 2-butanol dehydrogenation con-
for the dehydrogenation of 2-butanol [19]. Besides, in ventional plants, respectively. Besides, a part of the re-
2015, Geravand et al. prepared copper-silica (Cu/SiO2) quired hydrogen for the ammonia synthesis process is
nanosized catalysts and evaluated the catalytic perfor- supplied by the produced hydrogen in the dehydrogenation
mance of these catalysts in the 2-butanol dehydrogenation side.
reaction [20].

Thermally coupled reactors Process description

The thermal coupling which can be applied in recuperative, Ammonia synthesis

regenerative, and direct methods is an efficient way to reduce
energy consumption and operational costs of the process [21, Figure 1 shows an industrial ammonia synthesis plant. The
22]. It also decreases the size of the reactor [23]. In thermal reactor, as the main part of the plant, consists of the three
J Flow Chem

Fig. 1 A schematic diagram of the conventional ammonia synthesis plant

catalytic beds and the two interstage coolers to cool down the ammonia synthesis and 2-butanol dehydrogenation plants are
output gases from the first and second beds. The output gas eliminated, respectively. Furthermore, the produced hydrogen
from the reactor enters a heat exchanger to be cooled down. It by the dehydrogenation process is used as the ammonia syn-
will be also liquefied in the condenser and the ammonia will be thesis feed. The operating conditions and characteristics of the
separated as the product. In addition, a recycle stream consists reactor in both the exothermic [8, 34, 35] and endothermic
of the unreacted nitrogen and hydrogen is considered to be [33, 36] sides are presented in Table 3. As can be seen, the
warmed up in the heat exchanger and then enters the reactor operating conditions of the exothermic side are same as the
[31, 32]. The operating conditions and characteristics of the conventional ammonia synthesis reactor which already
conventional ammonia synthesis plant based on the Khorasan described in Table 1.
petrochemical plant data are described in Table 1 [8].
Table 1 The operating conditions and characteristics of the
MEK production conventional ammonia synthesis plant

Parameter Value
MEK and hydrogen are obtained as the main products of the
conventional 2-butanol dehydrogenation plant which is illus- Total feed flow rate, kg hr.−1 189,932.4
trated in Fig. 2. As seen, the reaction takes place in a multi- Total feed molar flow rate, kmol hr.−1 21,853.3
tubular reactor while its required heat is supplied by a furnace Reactor inlet temperature, K 533
[11]. The operating conditions and characteristics of the con- First bed inlet temperature (after heating), K 640
ventional reactor are presented in Table 2 [33]. Inlet pressure, atm 125.8
Feed composition (mole fraction)
Thermally coupled reactor H2 0.6672
N2 0.2297
Figures 3 and 4 show the schematic diagrams of the suggested NH3 0.0222
process and the thermally coupled reactor, respectively. In the First bed length, m 0.93
inner tube, the exothermic reaction of ammonia synthesis pro- Second bed length, m 0.93
vides the required heat for the outer tube’s endothermic reac- Third bed length, m 0.93
tion of 2-butanol dehydrogenation. As a result, the interstage Reactor diameter, m 2.8
coolers and furnace which were employed in the conventional
 J Flow Chem

Fig. 2 A schematic diagram of the conventional 2-butanol dehydrogenation plant

Kinetic model log10 K a ¼ −2:691122log10 T −5:519265  10−5 T

2001:6
Ammonia synthesis þ 1:848863  10−7 T 2 þ þ 2:6899 ð4Þ
T
The ammonia synthesis reaction can be expressed as follows [1]:
0:5N 2 þ 1:5H 2 ↔N H 3 ΔH 298 ¼ −46:22 KJ =mol ð1Þ

Table 2 The operating conditions and characteristics of the

In order to calculate the reaction rate, the modified Temkin conventional 2-butanol dehydrogenation plant
equation has been employed by the following rate expression
[34]: Parameter Value
2 !α !1−α 3
a3H 2 a2NH 3 Feed flow rate, gr hr.−1 200 to 600
RN H 3 ¼ k 2 4K 2a aN 2 − 5 ð2Þ Inlet temperature, K 644
a2N H 3 a3H 2
Inlet pressure, atm 1
where α is a constant which takes a value between 0.5 and Feed composition (mole fraction)
0.75 [34]. 2-butanol 1
k2 and Ka, as the reverse reaction rate and equilibrium H2 0
constants, are determined as follows, respectively [34]: MEK 0
  Reactor length, m 1.016
− 40765
Reactor sleeves diameter, m 0.013 and 0.025
k 2 ¼ 1:7698  1015 e ð3Þ
Rg T
J Flow Chem

Fig. 3 A schematic diagram of the proposed new configuration

Fig. 4 A conceptual diagram of the proposed thermally coupled reactor

 J Flow Chem

Table 3 The operating conditions and characteristics of the thermally Table 4 Rate constants for the 2-butanol dehydrogenation reaction
coupled reactor
Rate constants Ref
Parameter Value log10 C ¼ − T 10735
(7) [33]
ðo RÞþ7:776
Exothermic side (Ammonia synthesis): log10 K AK ¼ 486
T ðo RÞ−0:1986
(8) [33]
−3
Catalyst density, kg m 4900 Equilibrium constants
Diameter of catalyst particle, mm 8 log10 K ¼ − T2790
ðK Þ þ 1:510  log10 T ðK Þ þ 1:865 (9) [33]
Bed porosity 0.46 log10 K A ¼ − T o RÞþ5:327
6165
(10) [33]
Feed composition(mole fraction) ð
H2 0.6672
N2 0.2297
NH3 0.0222 Mathematical model
Inlet temperature, K 640
Inlet pressure, atm 125.8 Mathematical modeling plays an essential role in analyzing
−1
Total feed flow rate, kg hr. 189,932.4 the reactor behavior and improves its performance. In this
Total feed molar flow rate, kmol hr.−1 21,853.3 work, modeling the multi-tubular thermally coupled packed
Endothermic side (2-butanol dehydrogenation): bed reactor is done according to a one-dimensional homoge-
Catalyst density, kg m−3 8500 neous model. In addition, according to a differential element
Diameter of catalyst particle, mm 3.17 which is shown in Fig. 4, mass and energy balances for the
Porosity of the bed 0.39 exothermic and endothermic sides of the reactor are derived
Feed composition (mole fraction) separately. Due to axially move (in the z-direction) of the feed
2-butanol 1 in the catalytic bed assumption, the obtained equations are of
H2 0 the ordinary type (ODE). The 4th order Runge-Kutta method
MEK 0 is employed to solve these equations. The following assump-
Inlet temperature, K 644 tions are also applied in modeling:
Inlet pressure, atm 1
Total feed flow rate, kg hr.−1
85,255.1 & Steady state condition.
Total feed molar flow rate, kmol hr.−1 1150.2 & Variable physical properties during the process
Inner tube (exothermic side) diameter, m 0.576 & Ergun’s equation is employed to obtain the pressure drops
Outer tube (endothermic side) diameter, m 0.781 & Homogenous catalyst bed
Reactor length, m 2.79 & Heat loss from the reactor is negligible
& The effectiveness factor for the endothermic side is
assumed to be one

MEK production
Table 5 Mass and energy balances as well as the physical properties of
MEK and hydrogen are produced through the following the endothermic side
2-butanol dehydrogenation reaction [36]: Mass balance Ref
∂F j
C4 H 10 O↔C 4 H 8 O þ H 2 ΔH 298 ¼ 54 KJ =mol ð5Þ ∂Z ¼ ν j RA AEndo
c j ¼ 1; 2; 3; :::; m (11) [37]

Energy balance
hh  Exo Endo i i
The reaction rate can be expressed by the following ðΔH Endo RA Þþ ð ri Þ
−T
ðM Endo AEndo Þ
2U T
c
∂T Endo
equation [33]: ∂Z ¼ mEndo C Endo
(12) [37]
P mix

Components heat capacity

   2  2
PK i PH i
C P Ai − C oP j ¼ C 1 þ C 2 sinhC3C=T=T Endo þ C 4 coshC 5C=T=T Endo
Endo Endo
(13) [38]
K ð 3 Þ ð 5 Þ
RA ¼   ð6Þ
PA Viscosity of reaction mixture
PK i 1 þ K A PAi þ K AK i
PK i C
C 1 T Endo 2
μ¼ 2 (14) [38]
1þC 3 =T Endo þC 4 =T Endo

where, PAi , PK i and PH i refer to the partial pressures of Pressure drop

2-butanol, MEK and hydrogen, respectively. Other ki- ∂P
¼ 150μ ð1−εÞ Endo
u
2
þ 1:75ρð1−εÞ Endo 2
∂Z φ2 d 2 ε3 ϕ d p ε3 u s
(15) [39]
netic parameters are presented in Table 4 [33]. s p
J Flow Chem

According to the aforementioned assumptions, the model- Khorasan petrochemical plant data [8] in Table 7. As can be
ing equations, as well as the physical properties of the endo- seen, the values of the parameters approve the accuracy of the
thermic and exothermic sides, are presented in Tables 5 and 6, presented model.
respectively.
MEK production

Model validation The experimental data, as well as the simulation results of the
2-butanol conversion at different S/F (Surface-Feed) ratios,
Ammonia synthesis are demonstrated in Table 7. These data are obtained based
on the pure 2-butanol feed, under the atmospheric pressure
To determine the validity and correctness of the presented and at the temperature of 700 °F [33]. As it can be seen, the
model for the ammonia synthesis reactor, the simulation re- maximum relative error between the simulation results and the
sults including the molar fractions, temperature, and the nitro- experimental data is only 0.96%, which shows the correctness
gen conversion at the reactor’s outlet are compared to the and accuracy of the model.

Table 6 Mass and energy balances as well as the physical properties of the exothermic side

Mass balance Ref

∂ Fi
∂Z ¼ ν i RN H 3 ηAc i ¼ 1; 2; 3; :::; n
Exo
(16) [37]
Energy balance
hh  i i
2U ðT Exo −T Endo Þ
ð−ΔH Exo RN H 3 Þ− ri ðM Exo Ac Exo Þ
∂T Exo
∂Z ¼ mExo C Exo
(17) [37]
P mix

Overall heat transfer coefficient

ðDo =Di Þ
1
U ¼ h1i þ Ai ln2πLK w
þ AAoi  h1o (18) [40]

Heat transfer coefficient between gas phase and reactor wall

  
K ρug d p 0:365
h ¼ 4:21 dpf μ (19) [41]

Components activity
ai = fi = yi × Fii × PExo (20) [34]
Fugacity coefficients of components
      P 
P2 þ 300eð−0:011901T −5:941Þ e−300 −1
Exo0:125 Exo0:5
FiΗ 2 ¼ exp e −8:8402T þ0:541
P−e −0:1263T −15:980 Exo
(21) [34]
−3 −3 −6 2 −6 2
FiN 2 ¼ 0:93431737 þ 0:3101804  10 T Exo þ 0:295896  10 P−0:2707279  10 T Exo þ 0:477507  10 P (22) [34]

FiN H 3 ¼ 0:1438996 þ 0:2028538  10−2 T Exo −0:448762  10−3 P−0:1142945  10−5 T Exo þ 0:2761216  10−6 P2
2
(23) [34]
Specific heats for pure components
 
CpH 2 ¼ 4:184 6:952−0:04576  10−2 T Exo þ 0:09563  10−5 T Exo −0:2079  10−9 T Exo
2 3
(24) [42]
 
CpN 2 ¼ 4:184 6:903−0:03753  10−2 T Exo þ 0:1930  10−5 T Exo −0:6861  10−9 T Exo
2 3
(25) [42]
0 1
 10−2 T Exo þ 0:23663  10−5 T Exo −1:5981 −9 Exo3
2
"
6:5846−0:61251  
# 10 T
B 96:1778−0:067571P þ −0:2225 þ 1:6847  10−4 P T Exo C
CpN H 3 ¼ 4:184@ A (26) [42]
þ  −4 −7
 Exo2
þ 1:289  10 −1:009510 P T
Heat of reaction
ΔH Exo ¼ 9723:24½−23840:57
 þ ðP−300:0Þð1:08 þ ðP−300:0Þð0:01305
   (27) [43]
þðP−300:0Þ 0:83502  10ð−5Þ þ ðP−300:0Þ  0:65934  10ð−7Þ ÞÞ þ 4:5 1391:0−T Exo 
Effectiveness factor
 
η ¼ −17:5390 þ 0:0769T Exo þ 6:9005X − 1:08279  10−4 T Exo
2

  (28) [34]
−26:4247X 2 þ 4:92765  10−8 T Exo þ 38:9373X 3
3

Pressure drop
2
∂P ð1−εÞ Exo ð1−εÞ Exo
¼ 150μ u þ 1:75ρ
2
∂Z φ2 d 2 ε3
s p
ϕ d p ε3 us
(29) [39]
 J Flow Chem

Table 7 The modeling results and plant data of the conventional plants

Plant data Modeling results Relative errors (%)

Ammonia synthesis:
Mole fraction
H2 0.5684 0.5627 1.00
N2 0.1976 0.1952 1.21
NH3 0.1435 0.1458 1.60
Temperature, K 723 717.9 0.71
Nitrogen conversion 0.237 0.241 1.65
Experimental data Modeling results Relative errors (%)
2-butanol dehydrogenation:
S/F ratios
1 0.022538 0.022540 0.009
2 0.045076 0.045081 0.011
3 0.0684601 0.0684689 0.013
4 0.0918441 0.0918588 0.016
5 0.114838 0.114869 0.027
6 0.138546 0.138611 0.047
7 0.161540 0.162011 0.292
8 0.184468 0.185412 0.512
9 0.207462 0.208989 0.736
10 0.230846 0.233066 0.962

Results and discussion Figure 6 demonstrates the conversion of nitrogen for the
exothermic side and the conventional ammonia synthesis re-
The simulation results of the proposed thermally coupled actor. Based on the conventional reactor, the conversion has
reactor in the exothermic (ammonia synthesis) and endo- been slightly decreased from 0.241 to 0.231 for the exother-
thermic (2-butanol dehydrogenation) sides are compared to mic side caused by heat transfer from this side to the endo-
the conventional ammonia synthesis and single 2-butanol thermic side which results in the lower temperature profile
dehydrogenation reactors, respectively. The single 2- (Fig. 5), the lower reaction rate, and finally the lower
butanol dehydrogenation reactor has the same operat- conversion.
ing conditions and reactor specifications as the endo- The temperature profiles of the endothermic side and the
thermic side of the thermally coupled reactor described single 2-butanol dehydrogenation reactor are presented in
in Table 3. Fig. 7. At the beginning of the endothermic side, the reaction
The temperature profiles of the industrial ammonia heat requirement is more than the transferred heat from the
synthesis reactor and the exothermic side of the ther- exothermic side, thus the temperature decreases until it gets to
mally coupled reactor are presented in Fig. 5. At first, a minimum value at 620 K. Afterward, due to the decrease in
temperature increases rapidly in the conventional reactor the 2-butanol dehydrogenation reaction rate, and gradually
due to the high initial concentrations of reactants and increase in the temperature difference between two sides,
the high exothermic reaction rate. Afterward, the first which acts as a heat transfer driving force, the temperature
bed’s outlet enters a cooler and the temperature drops rises. Finally, the temperature remains constant at 2.20 m
from 763 k to 670 k. On the other hand, in the coupled length from the entrance of the endothermic side to the end
reactor, the temperature increases at the beginning of the of the reactor when the required heat and the transferred heat
exothermic side since the generated heat by this side is are equal to each other.
greater than the transferred heat to the endothermic side. The conversion of 2-butanol in the single dehydrogenation
Finally, when the produced heat becomes equal to the reactor is compared to the endothermic side in Fig. 8. The
transferred heat, the temperature remains constant. It conversion has increased from 0.891 for the single reactor to
should be also noted that as a result of heat transfer 0.957 for the coupled one. The higher conversion is caused by
between two sides, the exothermic side’s temperature heat transfer from the exothermic side to the endothermic side
profile is lower than the conventional reactor. of the thermally coupled reactor which results in the higher
J Flow Chem

Fig. 5 The temperature profiles

for the conventional ammonia
synthesis reactor and the
exothermic side

temperature profile (Fig. 7), the higher reaction rate, and fi- respectively. As seen in Fig. 9a, as the inlet temperature of
nally, the higher 2-butanol conversion compared to the single the exothermic side increases, the nitrogen conversion in-
reactor. creases due to the increased reaction rate at higher tempera-
It should be also pointed out that the proposed con- tures. Also, it is apparent that by increasing the inlet temper-
figuration requires 30.72% less hydrogen as the ammo- ature of the endothermic side, the conversion of nitrogen will
nia synthesis feed due to the hydrogen transfer from the increase due to the lower temperature difference and heat
outlet of the dehydrogenation side to the inlet of the transfer between the reactor sides which provides the higher
exothermic side. mean temperature and conversion in the exothermic side.
In this work, the effects of the main process parameters According to Fig. 9b, by increasing the inlet temperature of
variation including the inlet temperatures and the flow veloc- the endothermic side, the conversion of 2-butanol increases.
ities on the reactor performance are determined. Figure 9 dem- On the other hand, the higher 2-butanol conversion is
onstrates the influence of variation in the inlet temperatures on achieved by increasing the inlet temperature of the exothermic
the nitrogen and 2-butanol conversions. The inlet tempera- side due to increasing the temperature difference and heat
tures of the exothermic and endothermic sides are considered transfer which leads to the higher mean temperature in the
to change in the range of 620–660 K and 624–664 K, endothermic side.

Fig. 6 Nitrogen conversion along

the conventional ammonia
synthesis and the coupled reactors
 J Flow Chem

Fig. 7 The temperature profiles

for the single 2-butanol
dehydrogenation reactor and the
endothermic side

Figure 10a, b show the effect of the flow velocities varia- Fig. 10b, the lower flow velocity in the endothermic side
tion on the conversions. The flow velocities in the exothermic results in the higher reactor residence time and increases the
and endothermic sides (uExo, uEndo) are 7.21 m/s and 11.93 m/s, conversion of 2-butanol. Also, increasing the exothermic side
respectively. Therefore, the exothermic and endothermic flow velocity leads to the higher 2-butanol conversion, which
sides flow velocities are considered to change in the range of is due to an increase in the heat transfer coefficient. As a result,
5–9 m/s and 10–14 m/s, respectively. As expected, by decreas- the 2-butanol conversion is increased by increasing the heat
ing the exothermic side flow velocity, the nitrogen conversion transfer from the exothermic side.
will increase due to increasing the reactor residence time. This Optimization techniques are beneficial methods to
fact can be seen in Fig. 10a. Also, the lower endother- find the optimum operating conditions of the reactor
mic side flow velocity leads to the lower heat transfer and improve its performance. Depending on the purpose, dif-
coefficient in the endothermic side which results in the higher ferent optimization methods can be used [44]. Multi-objective
mean temperature and conversion in the exothermic side. In optimization which uses genetic algorithm (GA) can find a

Fig. 8 2-butanol conversion

along the single 2-butanol
dehydrogenation and the coupled
reactors
J Flow Chem

Fig. 9 The effect of the inlet

temperatures variation on the (a)
nitrogen and (b) 2-butanol
conversions

group of acceptable solutions for a variety of complicated function while the summation of the MEK production and out-
optimization problems in which objective functions are con- let yield is calculated by the second objective function in the
flicting with each other [30]. Finding a specific solution for other side. Both the objective functions are considered to be
multiple-objective problems by applying single objective maximized. Moreover, to normalize the production rates and
method leads to unacceptable results. Multi-objective optimi- give the same order of magnitude to them, the Bweighted sum
zation performs simultaneous optimization of each objective method^ has been employed by giving proper weights to the
by two general methods. The first method is based on combi- objective function parameters [45]. Therefore, production rates
nation of objective functions to a single function which can be of ammonia and MEK are divided by 2400 and 1100, respec-
determined by Butility theory^ or Bweighted sum^ techniques. tively. The aforementioned functions are as below:
The second general method determines a Pareto optimal set 
Prod N H 3
which includes a group of nondominated solutions. Obj1 ¼ − Y iNH 3 þ ð30Þ
Identifying solutions in the Pareto optimal set is the final goal 2400
of the multi-objective optimization algorithm [45]. 
Prod MEK
In this study, higher performance for the thermally coupled Obj2 ¼ − Y iMEK þ ð31Þ
1100
reactor is achieved with the aid of the multi-objective optimi-
zation. In the exothermic side, the summation of the ammonia where Prod and Yi refer to the production rate and
production and outlet yield is determined by the first objective yield, respectively.
 J Flow Chem

Fig. 10 The effect of the flow

velocities variation on the (a)
nitrogen and (b) 2-butanol
conversions

Furthermore, six decision variables including the en- reactor with a high degree of channelling, a meaning value of
dothermic and exothermic sides inlet temperatures LOD is considered in the current work. The ratio between
(T Endo
0 , T Exo
0 ), the total feed molar flow rates (F Endo
0 , F Exo
0 ), length and diameter may vary, but ratios between 3 and 10
length to inner tube diameter ratio (LOD) and the outer tube are most common [46].
diameter (DO) are selected. The variation ranges of the deci- In addition, in order to prevent high pressure drop during
sion variables are presented in Table 8. Since the reactor vol- maximizing the production rates by increasing the feed molar
ume is considered to be constant through the optimization, the flow rates, pressure drops in both sides of the optimized reac-
length changes with respect to the inner tube diameter. Also, it tor are considered as the constraints. A major disadvantage of
is apparent that both of the inner and outer tubes have the same the packed bed reactors is high pressure drop in the reactor.
length. Therefore, the optimized values of the LOD and the High pressure drop will result in both the high pumping-
outer tube diameter (DO) give the optimum length of the re- energy costs and lowering partial pressure of gaseous reactant,
actor as one of the design parameters. Also, since the more affecting the reactor performance [47]. In the present study,
LOD value results in a taller and narrower reactor with higher the pressure drop constraints are applied for both sides of the
pressure drop, and the fewer LOD value leads to the shorter reactor to avoid increasing the recompressing costs of the
J Flow Chem

Table 8 Variation ranges of the decision variables as well as the The optimization result is demonstrated as a Pareto
optimized values
front in Fig. 11. As seen, the curve consists of several
Range A B points with different production rate and yield values.
Each of the points can be accepted as a solution of
Parameter the optimization depending on the operating conditions
T Endo
0 ,K 624–664 664 664 and requirements. According to Fig. 11, point A at the
T Exo
0 , K 620–660 660 660 beginning of the curve, shows the maximum summation
F Endo
0 , kmol hr.−1 1075–1200 1176.2 1190.6 of the ammonia production and yield based on the first
−1
F Exo
0 , kmol hr. 21,750–21,950 21,925.2 21,864.1 objective function, while point B which is located at the
LOD 3–6 3.78 4.12 end of the curve, shows the maximum summation of the
Do, m 0.65–0.90 0.90 0.90 MEK production and yield according to the second ob-
Constraints jective function. The optimized values of the decision
ΔPExo, atm ≤ 2.5 2.39 2.31 variables as well as the constraints at these points are
ΔPEndo, atm ≤ 0.5 0.42 0.47 presented in Table 8.
YiNH 3 ≥ 0.072 0.114 0.102

Conclusion
reactor outputs. Besides, the lower pressure drop is generally
favorable to reach an acceptable simulation of packed bed A multi-tubular thermally coupled packed bed reactor in
reactors. These constraints are defined based on the operating which the endothermic reaction of 2-butanol dehydroge-
conventional reactors in the petrochemical plants to keep nation performs in the outer tube and the ammonia syn-
the pressure drops lower than 2.5 atm and 0.5 atm in thesis as the exothermic reaction, playing the heat sup-
the exothermic and endothermic sides, respectively. plier role, takes place in the inner tube, was simulated
Moreover, another constraint is applied to keep the pro- and studied. The required furnace and cooler which are
duction yield of ammonia in the optimized reactor employed in the conventional dehydrogenation and am-
higher than its value in the conventional reactor. The monia synthesis plants were eliminated, respectively. As
aforementioned constraints are as follows: the other advantage of the proposed configuration, the
produced hydrogen in the dehydrogenation side was
employed to provide a part of the required hydrogen
ΔPExo ≤ 2:5 atm ð32Þ
for the ammonia synthesis process. According to the
simulation results, compared to the single 2-butanol de-
ΔPEndo ≤ 0:5 atm ð33Þ hydrogenation reactor which operates at the same oper-
ating conditions and reactor specifications as the endo-
thermic side of the thermally coupled reactor, 2-butanol
Y iNH 3 ≥0:072 ð34Þ conversion and MEK production were increased. On the

Fig. 11 Pareto front for the multi-

objective optimization
 J Flow Chem

other hand, in comparison to the industrial ammonia 3. Peng P, Cheng Y, Hatzenbeller R, Addy M, Zhou N, Schiappacasse
C et al (2017) Ru-based multifunctional mesoporous catalyst for
synthesis plant, the nitrogen conversion has been slight-
low-pressure and non-thermal plasma synthesis of ammonia. Int J
ly decreased. Finally, the higher production rates and Hydrog Energy 42(30):19056–19066
yields for both sides of the proposed reactor were 4. Shimoda N, Kimura Y, Kobayashi Y, Kubota J, Satokawa S (2017)
achieved by employing the multi-objective optimization. Ammonia synthesis over yttrium-doped barium zirconate and
cerate-based perovskite-type oxide supported ruthenium catalysts.
Int J Hydrog Energy 42(50):29745–29755
5. Nikačević N, Jovanović M, Petkovska M (2011) Enhanced ammo-
Nomenclature ai, Activity of component i; AExo c , Cross-section nia synthesis in multifunctional reactor with in situ adsorption.
area of inner tube, m2; AEndo c , Cross-section area of outer tube, Chem Eng Res Des 89(4):398–404
m2; Ai, Inside area of inner tube, m2; Ao, Outside area of inner
6. Humphreys J, Lan R, Du D, Xu W, Tao S (2018) Promotion effect
tube, m2; C PN 2 , C PH 2 , C PNH 3 , Specific heat capacity of nitro-
of proton-conducting oxide BaZr0.1Ce0.7Y0.2O3−δ on the catalytic
gen, hydrogen and ammonia, kJ kmol−1 K−1; C 0P j , Heat capacity
activity of Ni towards ammonia synthesis from hydrogen and ni-
of component j in endothermic side, J kmol−1 K−1; CP, Specific heat of the
trogen. Int J Hydrog Energy 43(37):17726–17736
gas at constant pressure, J mol−1; C Endo Pmix , Specific heat of gas 7. Aziz M, Putranto A, Biddinika MK, Wijayanta AT (2017) Energy-
mixture in endothermic side, kJ kmol−1 K−1; C Exo Pmix , Specific heat saving combination of N2 production, NH3 synthesis, and power
of gas mixture in exothermic side, kJ kmol−1 K−1; C, Rate constant; generation. Int J Hydrog Energy 42(44):27174–27183
dp, Particle diameter, m; Do, Tube outside diameter, m; Di, Tube 8. Azarhoosh MJ, Farivar F, Ale EH (2014) Simulation and optimiza-
inside diameter, m; fi, Fugacity of component i, atm; Fi, Molar flow rate of tion of a horizontal ammonia synthesis reactor using genetic algo-
component i in exothermic side, kmol hr.−1; Fj, Molar flow rate of com- rithm. RSC Adv 4(26):13419–13429
ponent j in endothermic side, kmol hr.−1; Fii, Fugacity coefficient of 9. Kolios G, Frauhammer J, Eigenberger G (2000) Autothermal fixed-
component i; h, Heat transfer coefficient between fluid phase and reactor bed reactor concepts. Chem Eng Sci 55(24):5945–5967
wall, W m−2 K−1; ΔHExo, Heat of reaction in exothermic side, kJ kmol−1; 10. Babu BV, Angira R (2005) Optimal design of an auto-thermal am-
ΔHEndo, Heat of reaction in endothermic side, kJ kmol−1; Ka, Equilibrium monia synthesis reactor. Comput Chem Eng 29(5):1041–1045
constant of reaction; KA, Adsorption equilibrium constant for 2-butanol; 11. Neier W, Strehlke G (2002) 2-Butanone. Ullmann's encyclopedia of
K, Chemical equilibrium constant; KAK, Rate constant; Kw, Thermal con- industrial chemistry. 6 ed: Wiley
ductivity of reactor wall, W m−1 K−1; Kf, Fluid thermal conductivity,
12. Bai G, Wang Y, Li F, Zhao Z, Chen G, Li N et al (2013) Influence of
W m−1 K−1; L, Reactor length, m; m, Number of components in
acid–base properties of the support on copper-based catalysts for
endothermic side; mExo, Mass flow rate of the feed in exothermic side,
catalytic dehydrogenation of 2-butanol. Catal Lett 143(1):101–107
kg hr.−1; mEndo, Mass flow rate of the feed in endothermic side,
13. Fang D, Ren W, Liu Z, Xu X, Xu L, Lü H et al (2009) Synthesis and
kg hr.−1; MExo, Mean molecular weight in the flow in exothermic side,
applications of mesoporous cu-Zn-Al2O3 catalyst for dehydrogena-
kg kmol−1; MEndo, Mean molecular weight in the flow in endothermic
tion of 2-butanol. J Nat Gas Chem 18(2):179–182
side, kg kmol−1; n, Number of components in exothermic side; P, Total
14. Yang J, Zeng T, Cai D, Li L, Tang W, Hong R et al (2016)
pressure, atm; PAi , Partial pressure of 2-butanol at interface, atm; PK i ,
Supported ionic liquids as green catalyst for 2-butanol synthesis
Partial pressure of MEK at interface, atm; PH i , Partial pressure of from transesterification of sec-butyl acetate. Asia Pac J Chem Eng
hydrogen at interface, atm; RA, 2-butanol dehydrogenation reac- 11(6):901–909
tion rate, lbmol ft.−2 h−1; RNH 3 , Ammonia synthesis reaction 15. Liu Z, Huo W, Ma H, Qiao K (2006) Development and commercial
rate, kmol m−3 h−1; Rg, Gas constant, J mol−1 K−1; ri, Inside tube application of methyl-ethyl-ketone production technology. Chin J
radius, m; TExo, Temperature of exothermic reaction side, K; TEndo, Chem Eng 14(5):676–684
Temperature of endothermic reaction side, K; uExo, Axial velocity in 16. Keuler JN, Lorenzen L (2002) The dehydrogenation of 2-butanol in
exothermic side, m s−1; uEndo, Axial velocity in endothermic side, a Pd–ag membrane reactor. J Membr Sci 202(1–2):17–26
m s−1; ug, Velocity of gas phase, m s−1; U, Overall heat transfer coeffi- 17. Ravi S, Raghunathan T (1988) Dehydrogenation of butan-2-ol on
cient, W m−2 K−1; X, Conversion of nitrogen; yi, Mole fraction of com- zinc oxide catalyst: a continuous stirred tank reactor study. Ind Eng
ponent i; Z, Axial reactor coordinate, m Chem Res 27(11):2050–2055
Superscript Exo, Endo, Exothermic and endothermic reaction sides 18. Keuler J, Lorenzen L, Miachon S (2001) The dehydrogenation of 2-
Subscript i, Numerator for component in exothermic side; j, Numerator butanol over copper-based catalysts: optimising catalyst composi-
for component in endothermic side tion and determining kinetic parameters. Appl Catal A Gen 218(1–
Greek letters μ, Fluid viscosity, pa sec; ρ, Density, kg m−3; φs, 2):171–180
Sphericity; η, Effectiveness factor; νi, Stoichiometric coefficient of com- 19. Wang Z, Ma H, Zhu W, Wang G (2002) Characterization of
ponent i in exothermic side; νj, Stoichiometric coefficient of component j cu-ZnO-Cr2O3/SiO2 catalysts and application to dehydroge-
in endothermic side nation of 2-butanol to 2-butanone. React Kinet Catal Lett
76(2):271–279
Publisher’s note Springer Nature remains neutral with regard to jurisdic-
20. Geravand E, Shariatinia Z, Yaripour F, Sahebdelfar S (2015)
tional claims in published maps and institutional affiliations.
Copper-based nanocatalysts for 2-butanol dehydrogenation:
screening and optimization of preparation parameters by re-
sponse surface methodology. Korean J Chem Eng 32(12):
2418–2428
References 21. Chen J, Yan L, Song W, Xu D (2017) Methane steam reforming
thermally coupled with catalytic combustion in catalytic
1. Ammonia AM (2006) Ullmann's encyclopedia of industrial chem- microreactors for hydrogen production. Int J Hydrog Energy
istry. Wiley-VCH, Weinheim 42(1):664–680
2. Carvalho EP, Borges C, Andrade D, Yuan JY, Ravagnani MA 22. Nimkar SC, Mewada RK, Rosen MA (2017) Exergy and
(2014) Modeling and optimization of an ammonia reactor using a exergoeconomic analyses of thermally coupled reactors for meth-
penalty-like method. Appl Math Comput 237:330–339 anol synthesis. Int J Hydrog Energy 42(47):28113–28127
J Flow Chem

23. Karimi M, Rahimpour MR, Rafiei R, Shariati A, Iranshahi D 33. Perona JJ, Thodos G (1957) Reaction kinetic studies: catalytic de-
(2016) Improving thermal efficiency and increasing produc- hydrogenation of sec-butyl alcohol to methyl ethyl ketone. AICHE
tion rate in the double moving beds thermally coupled reac- J 3(2):230–235
tors by using differential evolution (DE) technique. Appl 34. Dyson D, Simon J (1968) Kinetic expression with diffusion correc-
Therm Eng 94:543–558 tion for ammonia synthesis on industrial catalyst. Ind Eng Chem
24. Iranshahi D, Saeedi R, Azizi K, Nategh M (2017) A novel integrat- Fundam 7(4):605–610
ed thermally coupled moving bed reactors for naphtha reforming 35. Rase HF (1977) Chemical reactor design for process plants: case
process with hydrodealkylation of toluene. Appl Therm Eng 112: studies and design data. Wiley, New York
1040–1056 36. Olsson H (2013) Advanced process integration aspects of tubular
25. Izurieta EM, Borio DO, Pedernera MN, López E (2017) Parallel reactors: Department of Energy and Environment. Chalmers
plates reactor simulation: ethanol steam reforming thermally University of Technology
coupled with ethanol combustion. Int J Hydrog Energy 42(30): 37. Fogler HS (2006) Elements of chemical reaction engineering. 4 ed:
18794–18804 Prentice Hall PTR
26. Chen J, Yan L, Song W, Xu D (2018) Comparisons between meth- 38. Green DW, Perry RH (2007) Perry's chemical Engineers' hand-
ane and methanol steam reforming in thermally integrated book. Eighth Edition, McGraw-Hill Education
microchannel reactors for hydrogen production: a computational 39. Ergun S, Orning AA (1949) Fluid flow through randomly packed
fluid dynamics study. Int J Hydrog Energy 43(31):14710–14728 columns and fluidized beds. Ind Eng Chem 41(6):1179–1184
27. Wiranarongkorn K, Authayanun S, Assabumrungrat S, 40. Holman JP (2009) Heat transfer. 10 ed: McGraw-Hill Education
Arpornwichanop A (2016) Analysis of thermally coupling steam 41. Wen D, Ding Y (2006) Heat transfer of gas flow through a packed
and tri-reforming processes for the production of hydrogen from bed. Chem Eng Sci 61(11):3532–3542
bio-oil. Int J Hydrog Energy 41(41):18370–18379 42. Shah M (1967) Control simulation in ammonia production. Ind Eng
28. Patel KS, Sunol AK (2007) Modeling and simulation of methane Chem 59(1):72–83
steam reforming in a thermally coupled membrane reactor. Int J 43. Gaines LD (1977) Optimal temperatures for ammonia synthesis
Hydrog Energy 32(13):2344–2358 converters. Ind Eng Chem Process Des Dev 16(3):381–389
29. Ramaswamy R, Ramachandran P, Duduković M (2008) Coupling 44. Amirabadi S, Kabiri S, Vakili R, Iranshahi D, Rahimpour MR
exothermic and endothermic reactions in adiabatic reactors. Chem (2013) Differential evolution strategy for optimization of hydrogen
Eng Sci 63(6):1654–1667 production via coupling of methylcyclohexane dehydrogenation
30. Saeedi R, Iranshahi D (2017) Multi-objective optimization of ther- reaction and methanol synthesis process in a thermally coupled
mally coupled reactor of CCR naphtha reforming in presence of double membrane reactor. Ind Eng Chem Res 52(4):1508–1522
SO2 oxidation to boost the gasoline octane number and hydrogen. 45. Konak A, Coit DW, Smith AE (2006) Multi-objective opti-
Fuel 206:580–592 mization using genetic algorithms: a tutorial. Reliab Eng
31. Elnashaie SS, Abashar ME, Al-Ubaid AS (1988) Simulation and Syst Saf 91(9):992–1007
optimization of an industrial ammonia reactor. Ind Eng Chem Res 46. Deckwer W-D, Field RW (1992) Bubble column reactors, vol 200.
27(11):2015–2022 Wiley, New York
32. Ludwig EE (2011) Applied process design for chemical and petro- 47. Cybulski A, Moulijn JA (2005) Structured catalysts and
chemical plants. 4 ed: Elsevier Science reactors2nd edn. CRC press, Boca Raton