Chemical Engineering and Processing 50 (2011) 113123

Contents lists available at ScienceDirect

Chemical Engineering and Processing:
Process Intensication
j our nal homepage: www. el sevi er . com/ l ocat e/ cep
DME synthesis and cyclohexane dehydrogenation reaction in an optimized
thermally coupled reactor
M.H. Khademi
a
, M. Farsi
a
, M.R. Rahimpour
a,b,
, A. Jahanmiri
a,b
a
Department of Chemical Engineering, School of Chemical and Petroleum Engineering, Shiraz University, Shiraz 71345, Iran
b
Gas Center of Excellence, Shiraz University, Shiraz 71345, Iran
a r t i c l e i n f o
Article history:
Received 15 June 2010
Received in revised form 11 October 2010
Accepted 5 December 2010
Available online 13 December 2010
Keywords:
Recuperative Coupling
Co-current heat-exchanger reactors
Cyclohexane dehydrogenation
DME synthesis
Optimization
Differential evolution
a b s t r a c t
This paper presents a study on optimization of DME synthesis and cyclohexane dehydrogenation in a
thermally coupled reactor. A steady-state heterogeneous model has been performed in order to evaluate
the optimal operating conditions and enhancement of DME and benzene production. In this work, the
catalytic methanol dehydration to DME is coupled with the catalytic dehydrogenation of cyclohexane
to benzene in a heat exchanger reactor formed of two xed beds separated by a wall, where heat is
transferred across the surface of tube. The optimization results are compared with corresponding pre-
dictions for a conventional (industrial) methanol dehydration adiabatic reactor operated at the same feed
conditions. The differential evolution (DE), an exceptionally simple evolution strategy, is applied to opti-
mize thermally recuperative coupled reactor considering DME and benzene mole fractions as the main
objectives. The simulation results have been shown that there are optimum values of initial molar ow
rate and inlet temperature of exothermic and endothermic sides to maximize the objective function. The
results suggest that optimal coupling of these reactions could be feasible and benecial and improves
the thermal efciency of process. Experimental proof-of-concept is needed to establish the validity and
safe operation of the novel reactor.
2010 Elsevier B.V. All rights reserved.
1. Introduction
Dimethyl ether (DME) is a colorless gas at the ambient condi-
tion and easily liqueed under low pressure. Recently, DME is an
attractive topic in academic and industrial research due to global
environment pollution and energy supply problem. DME does not
produce any particulate matter and toxic gases such as NO
x
at
burning, when used as a fuel. It can be produced from a variety
of feed-stock such as natural gas, crude oil, residual oil, coal, waste
products and bio-mass [1]. It is useful for a variety of application
such as LPG substitute, transportation fuel, propellant; chemical
feedstock and fuel cell [2].
DME production in the conventional process involves a
methanol dehydration reaction as an indirect method. Currently,
DME synthesize from syngas as a direct method is developed to
produce it at low cost. DME production through direct synthesis
from syngas using a dual catalyst system permits both methanol
synthesis and dehydration in a single process, with no methanol
purication. At present, DME is commercially produced by dehy-
drationof methanol inthe adiabatic packedbedreactor using acidic
porous catalysts [3].

Corresponding author. Tel.: +98 711 2303071; fax: +98 711 6287294.
E-mail address: rahimpor@shirazu.ac.ir (M.R. Rahimpour).
There are several articles in the literature that discuss modeling
of DMEcatalytic packedbedreactors. Eslamlueyan[4] modeledand
simulated DME synthesis in adiabatic xed bed reactor and shown
that difference between one-dimensional and two-dimensional
modeling is negligible. Liu et al. [5] modeled and designed a three-
phase bubble column reactor for direct synthesis of DME from
syngas with considering the inuence of inert carrier backmix-
ing on transfer and the inuence of catalyst grain sedimentation
on reaction. Simulation of uidized-bed reactor for DME synthe-
sis from syngas has been performed by Lu et al. [6] using plug
owmodel, which shows the great advantage of uidized-bed over
xed-bed or slurry reactor. Moradi et al. [7] studied DME synthesis
fromsynthesis gas inslurryreactor experimentallyanddetermined
the optimum operating conditions of DME synthesis. Omata et al.
[8] studied DME production fromsyngas in a temperature gradient
reactor for overcoming both the equilibriumlimit of the reaction at
high temperature and low activity of the catalyst at low tempera-
ture. Then, they optimized the reactor for higher CO conversion by
combined genetic algorithm and neural network.
Multifunctional reactors integrate, in one vessel, one or more
transport processes and a reaction system and are widely used in
industries as process intensication tools [9,10]. These multifunc-
tional reactors make the process more efcient and compact and
result in large savings in the operational and capital costs [11].
A multifunctional reactor can be used, for example, for coupling
0255-2701/$ see front matter 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.cep.2010.12.004
114 M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123
Nomenclature
a
v
Specic surface area of catalyst pellet (m
2
m
3
)
A
c
Cross section area of each tube (m
2
)
A
i
Inside area of inner tube (m
2
)
A
o
Outside area of inner tube (m
2
)
c Total concentration (mol m
3
)
C
i
Molar concentration of component i (mol m
3
)
C
p
Specic heat of thegas at constant pressure(J mol
1
)
d
p
Particle diameter (m)
D
i
Tube inside diameter (m)
D
ij
Binary diffusion coefcient of component i in j (m
2
s
1
)
D
im
Diffusion coefcient of component i in the mixture
(m
2
s
1
)
D
o
Tube outside diameter (m)
D
sh
Shell inside diameter (m)
F Total molar ow rate (mol s
1
)
G Mass velocity (kg m
2
s
1
)
h
f
Gas-solid heat transfer coefcient (W m
2
K
1
)
h
i
Heat transfer coefcient between uid phase and
reactor wall in exothermic side (W m
2
K
1
)
h
o
Heat transfer coefcient between uid phase and
reactor wall in endothermic side (W m
2
K
1
)
^H
f,i
Enthalpy of formation of component i (J mol
1
)
k Rate constant of dehydrogenation reaction (mol
m
3
Pa
1
s
1
)
k
1
Rate constant for the rate of methanol dehydration
reaction (mol
2
kg
1
s
1
m
3
)
k
g
Mass transfer coefcient for component i (m s
1
)
K Conductivity of uid phase (W m
1
K
1
)
K
B
Adsorption equilibriumconstant for benzene (Pa
1
)
K
eq
Reaction equilibrium constant for methanol dehy-
dration reaction (mol m
3
)
K
i
Adsorption equilibriumconstant for component i in
methanol dehydration reaction (m
3
mol
1
)
K
p
Equilibrium constant for dehydrogenation reaction
(Pa
3
)
K
w
Thermal conductivity of reactor wall (W m
1
K
1
)
L Reactor length (m)
M
i
Molecular weight of component i (g mol
1
)
N Number of components (N=3 for both dehydration
and dehydrogenation reaction)
P Total pressure (for exothermic side: bar; for
endothermic side: Pa)
P
i
Partial pressure of component i (Pa)
r
1
Rate of reaction for DME synthesis (mol kg
1
s
1
)
r
2
Rate of reactionfor dehydrogenationof cyclohexane
(mol m
3
s
1
)
r
i
Reaction rate of component i (for exothermic reac-
tion: mol kg
1
s
1
; for endothermic reaction: mol
m
3
s
1
)
R Universal gas constant (J mol
1
K
1
)
R
p
Particle radius (m)
Re Reynolds number
Sc
i
Schmidt number of component i
T Temperature (K)
u Supercial velocity of uid phase (m s
1
)
u
g
Linear velocity of uid phase (m s
1
)
U Overall heat transfer coefcient between exother-
mic and endothermic sides (W m
2
K
1
)
v
ci
Critical volume of component i (cm
3
mol
1
)
y
i
Mole fraction of component i (mol mol
1
)
z Axial reactor coordinate (m)
Greek letters
Viscosity of uid phase (kg m
1
s
1
)
Density of uid phase (kg m
3
)

b
Density of catalytic bed (kg m
3
)
z Tortuosity of catalyst
Superscripts
g In bulk gas phase
s At surface catalyst
Subscripts
0 Inlet conditions
B Benzene
C Cyclohexane
i Chemical species
j Reactor side (1: exothermic side, 2: endothermic
side)
k Reaction number index
exothermic and endothermic reactions. In this type of reactor, an
exothermic reaction is used as the heat producing source to drive
the endothermic reaction(s). In the last years promising concepts
for the recuperative coupling of exothermic and endothermic reac-
tions have been published [1222]. A distributed mathematical
model for thermally coupled membrane reactor that is composed
of three sides is developed for methanol and benzene synthesis by
Khademi et al. [23]. Methanol synthesis takes place in the exother-
mic side and supplies the necessary heat for the endothermic
dehydrogenation of cyclohexane reaction. Selective permeation of
hydrogen through the Pd/Ag membrane is achieved by co-current
ow of sweep gas through the permeation side. Also, recently
Farsi et al. [24] investigated the effect of various parameters on
the performance of a novel recuperative reactor. In this reac-
tor, catalytic DME synthesis reaction is coupled with the catalytic
dehydrogenation of cyclohexane to benzene. From these previ-
ous studies, coupling of endothermic and exothermic reactions
may enable both the concentration and temperature proles along
the reactor to be manipulated, shifting the conversion of ther-
modynamically limited reactions to higher values, and efciently
using the heat liberated by an exothermic reaction side to pro-
vide the endothermic heat requirements of the other reaction
side [25].
In the last few years, a Differential Evolution (DE) algorithm for
dealing with optimization problems has been proposed. DE algo-
rithmis a stochastic optimization method minimizing an objective
function that can model the problems objectives while incorporat-
ingconstraints. The algorithmmainlyhas three advantages; nding
the true global minimumregardless of the initial parameter values,
fast convergence, and using a few control parameters. Being sim-
ple, fast, easy to use, very easily adaptable for integrand discrete
optimization, quite effective in nonlinear constraint optimization
including penalty functions and useful for optimizing multi-modal
search spaces are the other important features of DE algorithm.
Recently, Khademi et al. [26] optimized the methanol synthesis
reaction and cyclohexane dehydrogenation in a thermally coupled
reactor using differential evolution(DE) method. Also, a novel ther-
mally coupled membrane reactor that is composed of three sides
is optimized for methanol, benzene and hydrogen productions by
Khademi et al. [27] using DE algorithm.
In our previous work [24], a distributed mathematical model
for thermally coupled reactor -containing DME synthesis reactions
and cyclohexane dehydrogenation- is developed. The effect of var-
ious key operating variables on the performance of the reactor is
M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123 115
Fig. 1. A schematic diagram of (a) a traditional adiabatic DME reactor and (b) a thermally coupled reactor conguration.
numerically investigated. As DME and benzene are the important
products, thecapacityof theplant is usuallyveryhigh, as aresult the
investment cost is also very high, and even a small enhancement in
the plant operation can generate major income. While the optimal
operating conditions can increase the production and reduce the
operating cost so, the purpose of this study is optimization of oper-
ating conditions for thermally coupled reactor using DE method,
as a strong optimization method. The inlet temperature and initial
molar ow rate of exothermic and endothermic sides have been
considered as decision variables to reach maximum mole fraction
of DME and benzene.
Rigorous mathematical models are excellent tools for the explo-
rationof the basic characteristics of suchnovel congurations. Such
anexplorationcanachieve considerable savings inmoney and time
during the expensive stage of pilot plant development. The con-
tinuous development of the model in conjunction with the pilot
plant optimal utilization can also achieve considerable benets on
the road towards the successful commercializationof suchefcient
novel congurations.
The paper is organized as follows: process description and
reactions scheme and kinetics are shown in Section 2 and 3. Math-
ematical model and concepts about DE are explained in Sections 4
and 5, respectively, followed by optimization of thermally coupled
reactor in Section 6. Numerical solution is presented in Section 7
with results and discussion in Section 8 and conclusions are drawn
in Section 9.
2. Process description
2.1. Industrial adiabatic DME reactor
Fig. 1(a) shows the schematic diagram of a conventional adia-
batic methanol dehydration reactor. In the conventional adiabatic
reactor, the catalyst is packed in the reactor. The methanol dehy-
dration reaction is carried out over a commercial -Al
2
O
3
catalyst.
In a heat exchanger, the heat of product stream is used to preheat
the feed stream.
2.2. Thermally coupled reactor
Fig. 1(b) shows a schematic diagram of thermally coupled reac-
tor conguration. It consist a shell compartment surrounding tube
compartments. Catalytic dehydrogenation of cyclohexane to ben-
zene is assumed to take place in the shell, whereas methanol
dehydration occurs inside the tube, with xed bed of different
catalysts on both sides. Heat is transferred continuously from
the exothermic reaction to the endothermic reaction. The clear
advantages of this integrated catalytic reactor include: achieving a
multiple reactants multiple products conguration and possibility
of achievinghigher degree of insituenergyintegrationbetweenthe
coupled endothermic dehydrogenation reaction and the exother-
mic DME synthesis reactions.
3. Reaction scheme and kinetics
3.1. Methanol dehydration (DME synthesis)
The reaction of DME synthesis is mainly dehydration of
methanol that is exothermic and equilibrium reaction. Many
researches are focused on DME synthesis reaction [2830]. In the
current work, the rate expressions have been selected from Ber ci c
and Levec [31].
2CH
3
OH CH
3
OCH
3
+H
2
O zH = 23.4kJ]mol (1)
The following reaction rate equation for methanol dehydration,
r
1
, is used [31]:
r
1
=
k
1
K
2
CH
3
OH

C
CH
3
OH

C
C
2
H
6
O
C
H
2
O

]K
eq

1 +2

K
CH
3
OH
C
CH
3
OH

5
+K
H
2
O
C
H
2
O

4
(2)
116 M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123
Table 1
The reaction rate constant, the adsorption equilibrium constant, and the reaction
equilibrium constant for methanol dehydration reaction.
k =Aexp( E/RT) A E (J mol
1
)
k
1
3.710
1
mol
2
kg
1
s
1
m
3
-105000
K
CH
3
OH
7.910
4
m
3
mol
1
70500
K
H
2
O
0.8410
1
m
3
mol
1
41100
Table 2
The reaction rate constant, the adsorption equilibrium constant, and the reaction
equilibrium constant for dehydrogenation of cyclohexane reaction.
k =Aexp(B/T) A B (K)
k 0.221mol m
3
Pa
1
s
1
-4270
KB 2.0310
1
Pa
1
6270
KP 4.8910
35
Pa
3
3190
where the reaction equilibrium constant, K
eq
is as follows:
ln

K
eq

= 0.86 log 1 +
3138
1
+1.33 10
3
1
1.23 10
5
1
2
+3.5 10
10
T (3)
k
1
, K
CH
3
OH
and K
H
2
O
are, respectively, the reaction rate constant
and the adsorption equilibrium constant for methanol and water
vapor that are tabulated in Table 1. Commercially, -Al
2
O
3
catalyst
is used in methanol dehydration reaction.
3.2. Cyclohexane dehydrogenation
Hydrogen is an optimumlarge scale fuel for the future, although
there remain some problems in transport and long-term storage.
One has to develop the use of alternative fuels that are easily trans-
formed into hydrogen and that can be stored in liquid form, and,
thus, more safely and economically. One of these fuels is cyclohex-
ane.
The reaction scheme for the dehydrogenation of cyclohexane to
benzene is as follows.
C
6
H
12
C
6
H
6
+3H
2
zH
298
= +206.2kJ/mol (4)
The following reaction rate equation of cyclohexane, r
2
, is used
[32]:
r
2
=
k(K
P
P
C
]P
3
H
2
P
8
)
1 +(K
8
K
P
P
C
]P
3
H
2
)
(5)
where k, K
B
, and K
p
are, respectively, the reaction rate constant,
the adsorption equilibrium constant for benzene, and the reac-
tion equilibrium constant that are tabulated in Table 2. p
i
is the
partial pressure of component i in Pa. The reaction temperature
is in the range of 423523K and the total pressure in the reac-
tor is maintained at 101.3 kPa. The catalyst for this cyclohexane
dehydrogenation reaction is Pt/Al
2
O
3
[33].
4. Mathematical model
Fig. 2 shows a schematic diagram for the co-current mode of
a heat-exchanger reactor conguration. A one-dimensional het-
erogeneous model, which is a conventional model for a catalytic
reactor withheat andmass transfer resistances, has beendeveloped
for this reactor inorder todeterminetheconcentrationandtemper-
ature distributions inside the reactor. In this model the following
assumptions are used:
The gas mixture is an ideal gas in both catalytic reactor sections.
Both sections of the reactor are operated at steady state condi-
tions.
Table 3
Strategy and parameters used for DE.
Strategy DE/best/1/bin
Population size (NP) 100
Scaling factor (F) 0.8
Crossover constant (CR) 1.0
Radial variations in both beds are negligible (one-dimensional
model).
With due attention to high gas velocity, axial diffusion of mass
and heat are negligible in both sections.
Bed porosity in axial and radial directions is constant.
Plug ow is employed in both endothermic and exothermic
sides.
The chemical reactions are assumed to take place only in the
catalyst particles and homogenous reactions are neglected.
Heat loss to surrounding is neglected.
To obtain the mole and energy balance equations, a differential
element along the axial direction inside the reactor was consid-
ered. The balances typicallyaccount for convection, transport tothe
solid-phase and reaction. The mass and energy balances, pressure
drop equation and boundary conditions for solid and uid phases
are summarized in Appendix A. To complete the simulation, aux-
iliary correlations should be added to the model. These auxiliary
correlations are presented in Appendix B.
5. Differential evolution
Differential evolution (DE) method, a recent optimization tech-
nique, is anexceptionally simple andeasy to use evolutionstrategy,
whichis signicantly fast androbust innumerical optimizationand
is more likely to nd a functions true global optimum [34]. Differ-
ent strategies can be adopted in DE algorithm depending upon the
type of problem for which DE is applied. The strategies can vary
based on the vector to be perturbed, number of difference vectors
considered for perturbation, and nally the type of crossover used.
Choosing populationsize (NP), scaling factor (F), andcrossover con-
stant (CR) depends on the specic problem applied, and are often
difcult. But some general guidelines are available. Normally, NP
should be about 510 times the number of parameters in a vector.
As for F, it lies in the range 0.41.0. Initially F =0.5 can be tried then
F and/or NP is increased if the population converges prematurely. A
good rst choice for CR is 0.1, but in general CR should be as large as
possible [35]. More details of basic version of DE (pseudo code), its
strategies, and choosing of operating parameters are reported by
Babu and Angira [34,36], and Babu and Munawar [35]. In this opti-
mization, the strategy and parameters used for DE are presented in
Table 3. Already, DE has been successfully applied for solving sev-
eral complex problems and is now being identied as a potential
source for accurate and faster optimization.
6. Optimization of recuperative reactor
In this study, maximization of the objective namely, summa-
tion of DME and benzene mole fractions in the reactor outlet is
considered. The objective is as follows:
j = y
DM
+y
C
6
H
6
(6)
Four decisionvariables namely, inlet temperature of exothermic
side T
01
, inlet temperature of endothermic side T
02
, initial molar
ow rate of exothermic side F
01
, and initial molar ow rate of
endothermic side F
02
are considered for optimization. For tubular
and exothermic reactors, temperature is an important parameter
that changes during the reactor length and it has a direct effect on
thermodynamic equilibrium and catalyst activity. The main rea-
M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123 117







C6H12
C6H6
H2
Ar
C6H12
Ar
CH3OH
DME
H2O
CH3OH
Heat transfer
Endothermic side
Endothermic side
Exothermic side
Fig. 2. A schematic diagram of the co-current mode for a recuperative reactor conguration.
son to develop an inlet optimal temperature in exothermic and
endothermic sides is energy saving and reducing in size of pre-
heater. While the heat transferred from the exothermic side to
endothermic side is dependent on the ratio of the exothermic-to-
endothermic side ow rates, these two parameters are selected as
decision variables. The ranges of decision variables are:
495 - 1
01
- 550K (7)
423 - 1
02
- 523K (8)
0.03 - F
01
- 0.15mol]s (9)
0.03 - F
02
- 0.15mol]s (10)
To ensure that the temperature of methanol gas at the reactor
inlet, T
01
is not too lowfor the DME synthesis reaction to occur, the
lower boundoninlet temperature of exothermic side is set at 495K.
At high temperatures catalyst starts deactivation [37], hence an
upper bound of 550Kis chosen for inlet temperature of exothermic
side.
While the activity of the Pt/Al
2
O
3
catalyst in the temperature
range of 423523K was evaluated in a conventional packed bed
and no signicant deactivation was found after at least 3days of
use [33], the bounds for the inlet temperature of endothermic side,
T
02
is chosen. The lower and upper bounds for the initial molar ow
rate of exothermic and endothermic sides have been selected with
no prior intention.
Three constraints are also considered for optimization:
1
2
- 1
1
(11)
495 - 1
1
- 550K (12)
423 - 1
2
- 523K (13)
The constraints on temperature (Eq. (11)) are based on the min-
imum temperature required in order to make a driving force for
heat transfer from the solid wall.
The optimization problem considered above is then reformu-
lated so as to include the constraints. Penalty function method
is employed for handling constraints. The constraints in Eqs.
(11)(13) are incorporated into the objective function (Eq. (6))
using penalty functions. This methodinvolves penalizing the objec-
tive function in proportion to the extent of constraint violation
(i.e., the penalty function takes a nite value when a constraint
is violated and a value of zero when constraint is satised). In the
present case, we used 10
7
as our penalty parameter. But this value
depends on order of magnitude of the variables in the problem; it
may change from problem to problem.
The objective function considered for minimization, nally, is
thus:
Minimize
j = j +10
7
5

i=1
C
2
i
(14)
where
C
1
= max{0, (1
2
1
1
)} (15)
C
2
= max{0, (495 1
1
)} (16)
C
3
= max{0, (1
1
550)} (17)
C
4
= max{0, (423 1
2
)} (18)
C
5
= max{0, (1
2
523)} (19)
The resulting optimization problems are solved using the pro-
posed algorithm, DE.
7. Numerical solution
Theformulatedmodel composedof 8ordinarydifferential equa-
tions and the associated boundary conditions lends itself to be
an initial value problem. The algebraic equations in the model
incorporate the initial conditions, the reaction rates, the ideal gas
assumption, as well as aforementioned correlations for the heat
and mass transfer coefcients and the physical properties of u-
ids. These equations alongwiththe discretizedordinarydifferential
equations using backward nite difference forma set of non-linear
algebraic equations. The reactor length is then divided into 100
separate sections and the Gauss-Newton method in MATLAB pro-
gramming environment is used to solve the non-linear algebraic
equations in each section.
8. Results and discussions
8.1. Model validation
The model of methanol dehydration side is validated against
conventional adiabatic DME synthesis reactor under the design
specications and input data listed inTable 4. It is observed that the
model performedsatisfactorily well under special case of industrial
conditions and the observed plant data were in good agreement
with simulation data. The comparison between steady state simu-
lation results and plant data for the conventional adiabatic reactor
is shown in Table 5. The plant data used for evaluation is taken
from Zagros Petrochemical Complex. As it is seen, the maximum
absolute error is about 1.95% which is acceptable.
In this section, various steady-state behaviors observed in the
co-current optimized thermally coupled reactor, are analyzed and
the predictedmole fractionandtemperature proles are presented.
118 M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123
Table 4
Operating conditions for methanol dehydration process (exothermic side).
Parameter Value
Gas phase
Feed composition (mole fraction)
CH
3
OH 0.936
DME 0.055
H
2
O 0.009
Inlet pressure (bar) 18.18
Catalyst particle
Density (kg m
3
) 2010
Particle diameter (m) 0.317510
2
Specic surface area (m
2
m
3
) 673
Ratio of void fraction to tortuosity of catalyst particle 0.066
Length of reactor (m) 8.08
Bed void fraction 0.5
Density of catalyst bed (kg m
3
) 1005
Tube inner diameter (m) 3.810
2
Tube outer diameter (m) 4.3210
2
Wall thermal conductivity (W m
1
K
1
) 48
Theperformanceof thethermallycoupledreactor is analyzed, using
different operating variables, for methanol and cyclohexane con-
version as follows:
Mcthuncl ccnvcrsicn =
F
CH
3
OH,in
F
CH
3
OH,cut
F
CH
3
OH,in
(20)
Cyclchcxunc ccnvcrsicn =
F
C
6
H
12
,in
F
C
6
H
12
,cut
F
C
6
H
12
,in
(21)
An obvious measure for the performance of the reactor concept
is how much heat has to be supplied through the exothermic reac-
tion to maintain the endothermic reaction. The relative heat supply
is dened by the fuel ratio 4:
4 =
available heat of exothermic reaction
maximumrequiredheat of endothermic reaction
(22)
As efciency of the reactor we dene.
=
heat actuallyconsumedfor endothermic reaction
heat actuallyreleasedfor exothermic reaction
(23)
Optimal conditions imply 41
+
and 1

.
8.2. Base case
In order to establish a reference point, optimization is carried
out for a base case and the operating conditions used for both
sides of the reactor are given in Table 4 and Table 6. Operating con-
ditions for the methanol dehydration side are similar to those used
by Farsi et al. [38]. The inlet composition of the methanol dehydra-
tion reaction is typical of industrial methanol dehydration process.
On the endothermic side, the inlet mole fraction of cyclohexane
that is diluted with argon is the same as that presented by Jeong
et al. [39]. Thus, the base case aims to investigate the situation
when the cyclohexane dehydrogenation is used to consumption
the generated heat frommethanol dehydrationand to cool downit,
resultinginahigher temperatureat rst parts of exothermic sidefor
higher kinetics constants and then reducing temperature gradually
Table 5
Comparison between steady state simulation results and plant data for the conven-
tional adiabatic methanol dehydration reactor.
Simulation
result
Plant
data
Absolute
error
DME molar ow rate (kmol/hr) 2457 2506 1.95%
MeOH molar ow rate (kmol/hr) 940.6 937.7 0.31%
Temperature (

C) 652.2 644 1.27%

Table 6
Operating conditions for dehydrogenation of cyclohexane to benzene (endothermic
side).
parameter Value
Gas phase
Feed composition
a
(mole fraction)
C
6
H
12
0.1
C
6
H
6
0.0
H
2
0.0
Ar 0.9
Inlet pressure
a
(Pa) 1.01310
5
Particle diameter
b
(m) 3.5510
3
Bed void fraction 0.39
Specic surface area (m
2
m
3
) 825
Shell inner diameter (m) 8.810
2
a
Obtained from Jeong et al. [39]
b
Obtained from Koukou et al. [40]
at the end parts of reactor for increasing thermodynamics equilib-
rium which is not similar to the temperature prole along a tube
lled with catalyst within a methanol dehydration adiabatic reac-
tor. This allows comparison of the methanol dehydration process
in the optimized thermally coupled reactor (OTCR) with adiabatic
reactor (AR) for similar thermal behavior. The optimization and
simulation results of the reactor in the endothermic side are not
compared with any reference case.
8.3. Simulation and optimization
With due attention to subjects of Section 5, the optimization
approach is to nd optimal temperature proles along the exother-
mic and endothermic sides to maximize DME and benzene mole
fractions through the optimization of initial molar ow rate and
inlet temperature of exothermic and endothermic sides. Differen-
tial evolution method is applied to determine the optimal reactor
operating conditions for DME and benzene production process in a
thermally coupled reactor. Figs. 3 and 4 show proles of objective
functioninterms of inlet temperature of exothermic andendother-
mic sides in the thermally coupled reactor, respectively. As shown
in these gures, the objective function is maximized in one point.
This means there is anoptimal inlet temperature of exothermic and
endothermic sides and their values are 524.7 and 470.3K, respec-
tively. Also Figs. 5 and 6 shows variation of objective function in
term of initial molar ow rate of exothermic and endothermic
495 503 511 519 527 535 543 550
0.2
0.5
0.8
1.1
1.4
1.7
2
Exothermic side temperature, K
O
b
j
e
c
t
i
v
e

F
u
n
c
t
i
o
n

Optimal
Fig. 3. Optimal inlet temperature of exothermic side for OTCR.
M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123 119
423 443 463 483 503 523
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
Endothermic side temperature, K
O
b
j
e
c
t
i
v
e

F
u
n
c
t
i
o
n
Optimal
Fig. 4. Optimal inlet temperature of endothermic side for OTCR.
0.02 0.04 0.06 0.08 0.1 0.12 0.14
1.5
1.55
1.6
1.65
1.7
1.75
1.8
1.85
Exothermic side flow rate, mol s
-1
O
b
j
e
c
t
i
v
e

F
u
n
c
t
i
o
n
Optimal
Fig. 5. Optimal initial molar ow rate of exothermic side for OTCR.
stream in thermally coupled reactor. As it can be seen in these g-
ures, the optimumvalue of initial molar owrate of exothermic and
endothermic streams are 0.0337 and 0.0459mol s
1
. The results of
the optimization (using differential evolution method and MATLAB
programming) are summarized in Table 7.
The simulation of thermally coupled reactor is carried out using
optimization results in Table 7 and the results of this simulation
are shown in several gures. Fig. 7(a)(c) shows the comparison of
mole fraction of components in exothermic side of optimized ther-
mally coupled reactor (OTCR) with adiabatic reactor (AR). Fig. 7(a)
Table 7
The optimized parameters for thermally coupled reactor.
Parameter Value
Inlet temperature of exothermic side (T
01
), K 524.7
Inlet temperature of endothermic side (T
02
), K 470.3
Initial molar ow rate of exothermic side (F
01
), mol s
1
0.0337
Initial molar ow rate of endothermic side (F
02
), mol s
1
0.0459
Objective function 1.845
0.02 0.04 0.06 0.08 0.1 0.12 0.14
0.2
0.5
0.8
1.1
1.4
1.7
2
Endothermic side flwo rate, mol s
-1
O
b
j
e
c
t
i
v
e

F
u
n
c
t
i
o
n
Optimal
Fig. 6. Optimal initial molar ow rate of endothermic side for OTCR.
illustrates the mole fraction prole of DME along the reactor, at
steady-state for exothermic side of OTCR and CMR. Fig. 7(b) and (c)
presents similar results for methanol andH
2
Ocomponents, respec-
tively while Fig. 7(d) is simultaneous plots of mole fraction for
cyclohexane, benzene and hydrogen in the endothermic side along
the reactor axis.
As shown, it is observed that there is not a considerable dif-
ference between the behavior of variables in the exothermic
side of OTCR and AR that is, the proles of mole fraction of
components along the reactor have the same patterns in the
both reactor under steady-state conditions. As it can be seen
in Fig. 7(a), the comparison of DME mole fraction in exother-
mic side of OTCR with AR shows that the DME mole fraction
in output of OTCR is increased by 2.43%. The mole fractions of
hydrogen and benzene are increased and cyclohexane is decreased.
As the reaction scheme for cyclohexane dehydrogenation indi-
cates, the increase of hydrogen mole fraction is higher than
benzene.
Fig. 8 shows axial temperature proles in the exothermic
and endothermic sides of OTCR and AR. In the adiabatic reac-
tor, the temperature increases up to equilibrium temperature
(647.6K). After a certain position along the reactor (dimension-
less length=0.1 in Fig. 8), the temperature maintain constant. In
addition, the highest temperature is observed at the exothermic
side, since this is where heat is generated. Part of this heat is
used to drive the endothermic reaction and the rest to heat the
reaction mixtures in both sides. The temperature of the dehydro-
genation side is always lower than that of the exothermic side
in order to make a driving force for heat transfer from the solid
wall. Along the exothermic side, temperature increases rapidly
and a hot spot develops as demonstrated in Fig. 8 and then
decreases to 424K. Also, the temperature pattern in endother-
mic side is the same as temperature pattern in exothermic
side; at the entrance of dehydrogenation side, the temperature
increases smoothly and a hot spot forms and then the temperature
decreases. Note that in the endothermic side, the outlet temper-
ature nearly is the same as the outlet temperature in exothermic
side (424K).
Fig. 9(a) shows the variation of rate of reaction for both sides
of OTCR. Near the reactor entrance, both the methanol dehydra-
tion and cyclohexane dehydrogenation reactions are fast. After
a certain position along the reactor (dimensionless length=0.15
120 M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123
Fig. 7. Comparison of (a) DME, (b) methanol, (c) H
2
O mole fraction along the reactor axis between exothermic side of OTCR and AR and (d) mole fraction of cyclohexane,
benzene and hydrogen in the endothermic side.
in Fig. 9(a)), the rate of methanol dehydration reaction is
zero.
Fig. 9(b) illustrates the variation of the generated and consumed
heat ux in the exothermic and the endothermic reaction, respec-
tively, and transferred heat from the solid wall along the reactor
for OTCR. At the reactor entrance (dimensionless length=00.1
in Fig. 9(b)), methanol dehydration reaction proceeds faster than
dehydrogenation and as a result more heat is produced by the
exothermic reaction than consumed by the endothermic one. The
excess heat raises the temperature of the system in the rst
half of the reactor as illustrated by the temperature prole in
Fig. 8.
In this region, the generated heat ux is higher than the con-
sumed one. The system heats up and a peak in the generated
heat ux is observed. Afterward, the generated heat ux decreases
rapidly, mainly due to fuel depletion. The opposite situation
occurs when the consumed heat ux is higher than the gener-
ated one. If the consumed heat ux is higher than the generated
one, the system starts to cool down resulting to low tempera-
ture, which in turn decreases both reaction rates. Thus, after a
certain position along the reactor (dimensionless length=0.1 in
Fig. 9(b)) the generated heat ux becomes lower than the con-
sumed one, which coincides with a hot spot development (see
Fig. 8).
Along the reactor length, the heat values consumed by the
endothermic side and transferred from the solid wall are close
to each other. This demonstrates the efcient thermal commu-
nication between the both sides, and which is due to high solid
wall thermal conductivity and the relatively small shell diame-
ter. At the reactor entrance, the transferred heat from the solid
wall is higher than the consumed heat by the endothermic side.
Almost, a maximum in the reaction heat uxes consumed and
transferred from the solid wall are located at the same axial posi-
tion, namely0.1. After this positionalong the reactor, the consumed
heat by the dehydrogenation side becomes larger than the trans-
0 0.2 0.4 0.6 0.8 1
400
450
500
550
600
650
700
T
e
m
p
e
r
a
t
u
r
e
,

K
Dimensionless length


Exothermic side of OTCR
Endothermic side of OTCR
AR
Fig. 8. Variation of temperature for exothermic and endothermic sides of OTCR, and
AR along the reactor.
M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123 121
Fig. 9. Variation of (a) rate of reaction for both sides and (b) generated and consumed heat ux and also transferred heat from solid wall along the reactor for the base case.
ferred heat from the solid wall and the system starts to cool down
(see Fig. 8).
In Fig. 10 cyclohexane conversion in the endothermic side of
OTCR, and methanol conversion in the exothermic side of OTCR
and AR along the reactor are shown. Cyclohexane reaches 100%
conversion, and methanol 84.55% and 82.27% conversion in the
exothermic side of OTCR and AR, respectively. The comparison of
methanol conversion in OTCR with AR shows that the methanol
conversion in output of OTCR is increased by 2.78%. The values
of fuel ratio and efciency of OTCR are 0.95 and 1.08, respec-
tively which are close to the optimal conditions. The heat actually
consumed for endothermic reaction is higher than heat actually
released for exothermic reaction. Thus, the reactor efciency is
higher than unit. On the other hand, since the transferred heat
from the solid wall isnt enough for endothermic reaction thus the
some of required heat for endothermic reaction take from its sen-
sible heat and temperature of endothermic side decreases. Overall,
the optimized operating parameters for this case lead to efcient
coupling of the two reactions.
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
M
e
t
h
a
n
o
l

C
o
n
v
e
r
s
i
o
n
Dimensionless length
0 0.2 0.4 0.6 0.8 1
0
0.2
0.4
0.6
0.8
1
C
y
c
l
o
h
e
x
a
n
e

C
o
n
v
e
r
s
i
o
n


OTCR
AR
Fig. 10. Cyclohexane and methanol conversion in OTCR and AR along the reactor
length.
122 M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123
Table 8
Mass and energy balances, pressure drop equation and boundary conditions for solid and uid phases in both reactor sides.
Mass and energy balances for solid phase in both sides uvc
j
k
gi,j
(y
g
i,j
y
s
i,j
) +q
j
r
i,j

b
= 0 (24)
uvh
j
(1
g
j
1
s
j
) +
b
N

i1
q
j
r
i,j
(zH
j,i
) = 0 (25)
Mass and energy balances for uid phase in both sides
F
j
/c
y
g
i,j
z
+uvc
j
k
gi,j
(y
s
i,j
y
g
i,j
) = 0 (26)

F
j
/c
C
g
pj
1
g
j
z
+uvh
j
(1
s
j
1
g
j
)
D
i
/c
U(1
g
2
1
g
1
) = 0 (27)
Pressure drop (Ergun momentum balance)
dP
dz
= 150
(1)
2
ug

3
d
2
p
+1.75
(1)u
2
g

3
dp
(28)
Boundary conditions z = 0, y
g
i,j
= y
g
i0,j
, 1
g
= 1
g
0
, P
g
= P
g
0
(29)
Table 9
Physical properties, mass and heat transfer correlations.
Parameter Equation Reference
Component heat capacity Cp =a +bT +cT
2
+dT
3
Mixture heat capacity Based on local compositions
Viscosity of reaction mixtures Based on local compositions
Mixture thermal conductivity Lindsay and Bromley [41]
Mass transfer coefcient between gas and solid phases k
gi
= 1.17Rc
0.42
Sc
0.67
i
ug 10
3
Cussler [42]
Rc =
2Rpug

Sc
i
=

D
im
10
4
D
im
=
1y
i

i=j
y
i
D
ij
[43]
D
ij
=
10
7
1
3]2

1]M
i
+1]M
j
P(v
3]2
ci
+v
2]3
cj
)
2
Reid et al. [44]
Overall heat transfer coefcient
1
U
=
1
h
i
+
/
i
ln(Dc]D
i
)
2LKw
+
/
i
/c
1
hc
Heat transfer coefcient between gas phase and reactor wall
h
Cp

Cp
K

2]3
=
0.458

udp

0.407
[45]
9. Conclusion
Optimization of DME synthesis reaction coupled with dehy-
drogenation of cyclohexane to benzene by means of indirect heat
transfer in a catalytic heat exchanger reactor was studied by a one-
dimensional model. The reactor consists of two separated sides for
exothermic and endothermic reactions. A base case was generated
considering similar operating conditions to industrial methanol
dehydration reactor. Differential evolution (DE) method, an excep-
tionally simple evolution strategy, is applied to determine the
optimal reactor operating conditions. Maximization of DME and
benzene mole fractions in the reactor outlet is considered. It is
shownthat suitable amount of initial molar owrate andinlet tem-
perature of both sides can provide the necessary heat to heat up
the mixtures and to drive the endothermic process at the same
time. The short distance between the heat sink and transferred
heat increases the efciency of heat transfer. This new cong-
uration represents some important improvement in comparison
to conventional adiabatic methanol dehydration reactor as fol-
lows: reduces the size of the reactors; lower outlet temperature
of product stream leads to shift the thermodynamics equilibrium;
hydrogen and benzene are also produced as an additional valu-
able product; and auto-thermality is achieved within the reactor.
The optimization method has enhanced the DME mole fraction by
2.43% and methanol conversion by 2.78%. The results indicate that
DMEsynthesis reactionandcyclohexane dehydrogenationina heat
exchanger reactor is feasible and benecial.
Appendix A. Mass and energy balances
The mass and energy balances, pressure drop equation and
boundary conditions for solid and uid phases are summarized in
Table8. Inequations (24) and(25), qis effectiveness factor (theratio
of the reaction rate observed to the real rate of reaction), which is
obtained from a dusty gas model calculations. The detail of such a
dusty gas model is given by Farsi et al. [24]. In equation (27), the
positive sign is used for the exothermic side and the negative sign
for the endothermic side.
Appendix B. Auxiliary correlations
In the heterogeneous model, because of transfer phenomena,
the correlations of estimation of heat and mass transfer between
two phases and estimation of physical properties of chemical
species and overall heat transfer coefcient between shell and tube
sides shouldbe considered. The correlations usedfor physical prop-
erties and mass and heat transfer coefcients are summarized in
Table 9. The heat transfer coefcient between gas phase and reac-
tor wall is applicable for the heat transfer coefcient between bulk
gas phase and catalyst phase (h
f
).
References
[1] C. Arcoumanis, C. Bae, R. Crookes, E. Kinoshita, The potential of dimethyl ether
(DME) as an alternative fuel for compression-ignition engines: A review, Fuel
87 (2008) 10141030.
[2] K.L. Ng, D. Chadwick, B.A. Toseland, Kinetics and modeling of Dimethyl ether
synthesis from synthesis Gas, Chem. Eng. Sci. 54 (1999) 35873592.
[3] W.Z. Lu, L.H. Teng, W.D. Xiao, Simulation and experiment study of Dimethyl
ether synthesis fromsyngas ina uidized-bedreactor, Chem. Eng. Sci. 59(2004)
54555464.
[4] S.M. Nasehi, R. Eslamlueyan, A. Jahanmiri, Simulation of DME reactor from
methanol. Proceeding of the 11
nd
chemical engineering conference, Iran, Kish
Island, 2006.
[5] D. Liu, X. Hua, D. Fang, Mathematical Simulation Design of Three-Phase Bubble
Column Reactor for Direct Synthesis of Dimethyl Ether fromSyngas, J. Nat. Gas
Chem. 16 (2) (2007) 193199.
[6] W.Z. Lu, L.H. Teng, W.D. Xiao, Theoretical analysis of uidized bed reactor for
dimethyl ether synthesis fromSyngas. Int. J. Chem. Reactor Eng. 1(2003) Article
S2.
M.H. Khademi et al. / Chemical Engineering and Processing 50 (2011) 113123 123
[7] G.R. Moradi, R. Ghanei, F. Yaripour, Determination of the Optimum Operating
Conditions for Direct Synthesis of Dimethyl Ether from Syngas. Int. J. Chem.
Reactor Eng. 5 (2007) Article A14.
[8] K. Omata, T. Ozak, T. Umegaki, Y. Watanabe, N. Nukui, M. Yamada, Optimization
of the temperature prole of a temperature gradient reactor for DME synthesis
using a simple genetic algorithm assisted by a neural network: High-quality
transportation fuels, Energy Fuels 17 (4) (2003) 836841.
[9] D.W. Agar, Multifunctional reactors: old preconceptions and new dimensions,
Chem. Eng. Sci. 54 (10) (1999) 12991305.
[10] M. Zanr, A. Gavriilidis, Catalytic combustion assisted methane steam reform-
ing in a catalytic plate reactor, Chem. Eng. Sci. 58 (17) (2003) 39473960.
[11] F.M. Dautzenberg, M. Mukherjee, Process intensication using multifunctional
reactors, Chem. Eng. Sci. 56 (10) (2001) 251267.
[12] J.B. Hunter, G. McGuire, Method and apparatus for catalytic heat-exchange. US
Patent 4 (1980) 214 867.
[13] N. Itoh, T.H. Wu, An adiabatic type of palladiummembrane reactor for coupling
endothermic and exothermic reactions, J. Membr. Sci. 124 (1997) 213222.
[14] J. Frauhammer, G. Eigenberger, L. Hippel, D. Arntz, A new reactor concept for
endothermic high-temperature reactions, Chem. Eng. Sci. 54 (1516) (1999)
36613670.
[15] G. Kolios, J. Frauhammer, G. Eigenberger, Asimpliedprocedure for the optimal
design of auto-thermal reactors for endothermic high-temperature reactions,
Chem. Eng. Sci. 56 (2) (2001) 351357.
[16] G. Kolios, J. Frauhammer, G. Eigenberger, Efcient reactor concepts for cou-
pling of endothermic and exothermic reactions, Chem. Eng. Sci. 57 (9) (2002)
15051510.
[17] S.S.E.H. Elnashaie, T. Moustafa, T. Alsoudani, S.S. Elshishini, Modeling and basic
characteristics of novel integrated dehydrogenationhydrogenation mem-
brane catalytic reactors, Comput. Chem. Eng. 24 (27) (2000) 12931300.
[18] T.M. Moustafa, S.S.E.H. Elnashaie, Simulation production of styrene and cyclo-
hexane in an integrated membrane reactor, J. Membr. Sci. 178 (2000) 171184.
[19] C. Fukuhara, A. Igarashi, Performance simulation of a wall type reactor in which
exothermic and endothermic reactions proceed simultaneously, comparing
with that of a xed-bed reactor, Chem. Eng. Sci. 60 (2005) 68246834.
[20] R.C. Ramaswamy, P.A. Ramachandran, M.P. Dudukovic, Recuperative coupling
of exothermic and endothermic reactions, Chem. Eng. Sci. 61 (2006) 459472.
[21] R.C. Ramaswamy, P.A. Ramachandran, M.P. Dudukovic, Coupling exothermic
and endothermic reactions in adiabatic reactors, Chem. Eng. Sci. 63 (6) (2008)
16541667.
[22] N.S. Abo-Ghander, J.R. Grace, S.S.E.H. Elnashaie, C.J. Lim, Modeling of a novel
membrane reactor to integrate dehydrogenation of ethylbenzene to styrene
with hydrogenation of nitrobenzene to aniline, Chem. Eng. Sci. 63 (2008)
18171826.
[23] M.H. Khademi, A. Jahanmiri, M.R. Rahimpour, A novel conguration for
hydrogen production from coupling of methanol and benzene synthesis in a
hydrogen-permselective membrane reactor, Int. J. Hydrogen Energy 34 (2009)
50915107.
[24] M. Farsi, M.H. Khademi, A. Jahanmiri, M.R. Rahimpour, A Novel Recuperative
Conguration for Coupling of Methanol Dehydration to DME with Cyclohexane
Dehydrogenation to Benzene, Ind. Eng. Chem. Res. 49 (2010) 46334643.
[25] B. Amon, H. Redlingshofer, E. Klemm, E. Dieterich, G. Emig, Kinetic investigation
of deactivation by coking of a noble metal catalyst in the catalytic hydrogena-
tion of nitrobenzene using a catalytic wall reactor, Chem. Eng. Process 38 (4)
(1999) 395404.
[26] M.H. Khademi, P. Setoodeh, A. Jahanmiri, M.R. Rahimpour, Optimization of
methanol synthesis and cyclohexane dehydrogenation in a thermally coupled
reactor using differential evolution (DE) method, Int. J. Hydrogen Energy 34
(2009) 69306944.
[27] M.H. Khademi, M.R. Rahimpour, A. Jahanmiri, Differential evolution (DE) strat-
egy for optimization of hydrogen production, cyclohexane dehydrogenation
and methanol synthesis in a hydrogen-permselective membrane thermally
coupled reactor, Int. J. Hydrogen Energy 35 (2010) 19361950.
[28] M. Xu, D.W. Goodman, A. Bhattacharyya, Catalytic dehydration of methanol to
dimethyl ether (DME) over Pd/Cab-O-Sil catalysts, Appl. Catal. A: Gen. 149 (2)
(1997) 303309.
[29] K.W. Jun, H.S. Lee, H.S. Roh, A.E. Park, Catalyst dehydration of methanol to
dimethyl ether (DME) over solid-acid catalysts, Bull. Korean Chem. Soc. 23 (6)
(2006) 803806.
[30] J. Bondiera, C. Naccache, Kinetics of Methanol Dehydration in Dealuminated
H-Mordenite: Model with Acid and Basic Active Centers, Appl. Catal. 69 (1991)
139148.
[31] G. Ber ci c, J. Levec, Intrinsic and global reaction rate of methanol dehydration
over Al
2
O
3
pellets, Ind. Eng. Chem. Res. 31 (1992) 10351040.
[32] N. Itoh, A membrane reactor using palladium, AIChE J. 33 (1987) 15761578.
[33] B.H. Jeong, K.I. Sotowa, K. Kusakabe, Catalytic dehydrogenation of cyclohexane
in an FAU-type zeolite membrane reactor, J. Membr Sci. 224 (2003) 151158.
[34] B.V. Babu, R. Angira, Modied differential evolution (MDE) for optimization of
non-linear chemical processes, Comput. Chem. Eng. 30 (2006) 9891002.
[35] B.V. Babu, S.A. Munawar, Differential evolution strategies for optimal design of
shell-and-tube heat exchangers, Chem. Eng. Sci. 62 (2007) 37203739.
[36] R. Angira, B.V. Babu, Optimization of process synthesis and design problems: A
modieddifferential evolutionapproach, Chem. Eng. Sci. 61(2006) 47074721.
[37] F. Yaripour, F. Baghaei, I. Schmidt, J. Perregaard, Catalytic dehydration of
methanol to dimethyl ether (DME) over solid-acid catalysts, Catal. Commun. 6
(2005) 147152.
[38] M. Farsi, A. Jahanmiri, R. Eslamlueyan, Modeling and Optimization of MeOH to
DME inIsothermal Fixed-bed Reactor. Int. J. Chem. Reactor Eng. 8 (2010) Article
A79.
[39] B.H. Jeong, K.I. Sotowa, K. Kusakabe, Modelingof anFAU-typezeolitemembrane
reactor for the catalytic dehydrogenation of cyclohexane, Chem. Eng. J. 103
(2004) 6975.
[40] M.K. Koukou, G. Chaloulon, N. Papayannakos, N.C. Markatos, Mathematical
modeling of the performance of non-isothermal membrane reactors, Int. J. Heat
Mass Transfer 40 (10) (1997) 24072417.
[41] A.L. Lindsay, L.A. Bromley, Thermal conductivity of gas mixture, Ind. Eng. Chem.
42 (1950) 15081510.
[42] E.L. Cussler, Diffusion Mass Transfer in Fluid Systems, University Press, Cam-
bridge, 1984.
[43] C.R. Wilke, Estimation of liquid diffusion coefcients, Chem. Eng. Progress 45
(1949) 218224.
[44] R.C. Reid, T.K. Sherwood, J. Prausnitz, The Properties of Gases and Liquids, third
ed, McGraw-Hill, New York, 1977.
[45] J.M. Smith, Chemical Engineering Kinetics, McGraw-Hill, New York, 1980.