Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920

Contents lists available at SciVerse ScienceDirect

Journal of Industrial and Engineering Chemistry

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

Comparison of CFD results and experimental data in a fixed bed Fischer–Tropsch

synthesis reactor
Ali Reza Miroliaei a, Farhad Shahraki a,*, Hossein Atashi a, Ramin Karimzadeh b
a
Department of Chemical Engineering, University of Sistan and Baluchestan, P.O. Box 98164-161, Zahedan, Iran
b
Department of Chemical Engineering, University of Tarbiat Modares, P.O. Box 14155-4838, Tehran, Iran

A R T I C L E I N F O A B S T R A C T

Article history: Hydrogenation of carbon monoxide in a fixed bed Fischer–Tropsch reactor has been simulated using
Received 7 January 2012 computational fluid dynamics. The geometry of fixed bed reactor was modeled using the discrete packing
Accepted 6 May 2012 method. The effects of reaction temperature, feed flow rate and H2/CO ratio on CO conversion and product
Available online 12 May 2012
selectivity were investigated. The mass fraction of products was also predicted by using the CFD model at
different H2/CO ratios. It was observed that the CO conversion increased with increasing temperature,
Keywords: while decreased with increasing mass flow rate. The selectivity and mass fraction of products also
Computational fluid dynamics
increased with increasing temperature and H2/CO ratio.
Discrete packing method
ß 2012 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights
Fischer–Tropsch synthesis
Fixed bed reactor reserved.
Synthesis gas

1. Introduction model to investigate parameter effects on commercial reactors. Bub

et al. [11] developed a two-dimensional, pseudo-homogeneous, plug
The Fischer–Tropsch synthesis (FTS) is a process that the flow model to predict product distribution. Liu et al. [12] studied the
synthesis gas (syngas), a mixture of carbon-monoxide and steady state and dynamic behavior of the fixed bed reactor for FTS
hydrogen, obtained from natural gas, coal, and biomass is process, applying a two-dimensional heterogeneous model. Pina
converted to a mixture of hydrocarbons. Fuels produced from et al. [13] studied the effect of heat flux on the performance of an
the FTS are of high quality due to a very low aromaticity and industrial methane steam reformer using a one-dimensional model.
absence of sulfur. The FTS process consists of either low They found that the profile of heat flux along the axial direction of the
temperature Fischer–Tropsch process, operated at 200–250 8C, reactor greatly affected the methane conversion and the tube wall
or high temperature Fischer–Tropsch process, operated at 300– temperature. Wang et al. [14] proposed a one-dimensional
350 8C, depending on the product required. Three reactor types, heterogeneous reactor model to simulate the performance of fixed
including multi-tubular fixed bed reactor, slurry bubble column bed reactors for hydrocarbon production. Jess and Kern [2] used a
reactor and fluidized bed reactors are used for the FTS process. The two-dimensional pseudo-homogeneous model for comparison of
FTS is usually performed on cobalt or iron based catalysts. Many the performance of cobalt and iron based catalysts in multi-tubular
inorganic oxides such as Al2O3, SiO2, TiO2, and ZrO2 are used as fixed bed reactor. They observed that the effective reaction rate with
supports for improving the structural stability of catalysts [1–4]. cobalt is higher than with iron. Guettel and Turek [15] compared
The process variables such as reaction temperature, gas velocity, different reactor types. They used a one-dimensional model in their
catalyst density, total pressure, and hydrogen to carbon monoxide simulations. Rahimpour et al. [16] optimized the operating
(H2/CO) ratio affect the product distribution [5,6]. conditions to maximize the gasoline production yield. They proposed
Many researchers have been studying on catalyst development a reactor consisting two fixed bed for the exothermic Fischer–
[7–9], and reactor modeling [10,11], but few studies have been made Tropsch reactions and the endothermic reaction of cyclohexane
on optimization of specific products. For example, Atwood and dehydrogenation to benzene. Zakeri et al. [17] studied the process
Benett [10] proposed a one-dimensional, heterogeneous plug flow variables such as type and concentration of binder and compression
pressure in pelletizing process on the selectivity of ethylene and
propylene. They found that the pelletizing by binder improved the
ethylene and propylene selectivities.
* Corresponding author. Tel.: +98 541 2447039; fax: +98 541 2447186.
E-mail addresses: [email protected] (A.R. Miroliaei),
CFD codes have also been used to model homogeneous and
[email protected] (F. Shahraki). heterogeneous reactions by several authors [18–24]. For example,

1226-086X/$ – see front matter ß 2012 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jiec.2012.05.003
 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920 1913

(SEM) and Brunauer–Emmett–Teller (BET) were also explained.

Nomenclature
The kinetic measurements were carried out in a fixed bed reactor
with an inner diameter of 12 mm. The schematic of the
C molar concentration [kg mol/m3] experimental set-up is shown in Fig. 1. All gas lines to the reactor
Dij mass diffusion coefficient [m2/s] bed were made from stainless steel tubing. Three mass flow
Ei activation energy [J/kg mol] controllers (Brooks, Model 5850E) equipped with a four-channel
gi gravity acceleration in direction i [m/s2] control panel (Brooks 0154) were used to adjust automatically the
hi fluid enthalpy of component i [kJ/kg] flow rate of the inlet gases comprising CO, H2, and N2 with purity of
Jij diffusion flux of component i [kg/m2 s] 99.99%. The mixed gases passed into the reactor tube, which was
kB Boltzmann constant [J/molecule K] placed inside a tubular furnace (Atbin, Model ATU 150-15) capable
of producing temperature up to 1300 8C and controlled by a digital
kf thermal conductivity of fluid [W/m K]
programmable controller. The reactor tube was constructed from
Mw molecular weight [kg/kg mol]
stainless steel tubing with the catalyst bed situated in the middle
m rate exponent [–] of the reactor. The reaction temperature was controlled by a
n rate exponent [–] thermocouple inserted into catalyst bed and visually monitored by
P static pressure [Pa] a computer. The catalyst was pre-reduced in situ at atmospheric
ri Arrhenius reaction rate of component i [kg mol/ pressure in a flowing H2–N2 stream (N2/H2 = 1, flow rate of each
gcat s] gas = 30 ml/min) at 400 8C for 16 h before syngas exposure.
R universal gas constant [J/kg mol K] Reactant and product steams were analyzed on-line using a gas
Ri reaction rate of component i [kg/gcat s] chromatograph (Varian, Model 3400 series) equipped with a 10-
Sh volumetric heat source [J/m3 s] port sampling valve (Supelco Company, USA, Visi Model), a sample
loop and a thermal conductivity detector. The contents of sample
t time [s]
loop were injected automatically into a packed column (Hayesep
T temperature [K]
DB, Alltech Company, USA). Helium was employed as a carrier gas
u superficial velocity [m/s]
for optimum sensitivity (flow rate = 30 ml/min). The calibration
x coordinate [m] was carried out using various calibration mixtures and pure
Yi local mass fraction of component i [–] compounds obtained from American Matheson Gas Company
(USA). GC controlling and collection of all chromatograms was
Greek letters done via an IF-2000 single channel data interface (TG Co., Tehran,
dij Kronecker delta function [–] Iran) at windows1 environment.
e Lennard–Jones energy parameter [J/molecule] Evidently, due to the complexity of FTS reaction system, a
r density [kg/m3] proper reaction model, from which the selectivity and heat transfer
s Lennard–Jones characteristic length [m] information can be determined in a quantitative fashion, has not
yet been presented. The FTS can be regarded as a surface
t deviatoric stress tensor [Pa]
polymerization reaction since monomer units are formed from
VD diffusion collision integral [–]
the reagents, hydrogen and carbon monoxide, in situ on the surface
of the catalyst, which is usually based on iron or cobalt. The
Dixon et al. [21] studied diffusion and reaction for the endothermic kinetics of the FTS on cobalt catalysts have been investigated due to
methane steam reforming in a fixed bed reactor. They observed high selectivity to lower molecular weight olefins [26–28].
that the strong temperature gradient happened in the particles In this study, the production of liquid hydrocarbons was
placed near the tube wall. Shi et al. [22] used n-hexadecane as feed neglected because of their small amounts compared to those of
for the production of hydrogen. They found that the thermal hydrocarbons. Water-gas shift reaction was also eliminated in the
conductivity of solid catalyst layer affected the temperature profile kinetic study which is a reasonable assumption for the cobalt
in the reactor, but its effect on product hydrogen concentration catalyst. The equations of chemical reactions are applied according
was negligible. Arzamendi et al. [23] used two CFD models to to:
simulate the integration of the steam reforming and combustion of
methane in a catalytic microchannel reactor. They considered the CO þ 3H2 ! CH4 þ H2 O (1)
effect of the gas velocity, catalyst load, steam to carbon ratio, and
flow arrangement on the microchannel performance. Esteban 2CO þ 4H2 ! C2 H4 þ 2H2 O (2)
Duran et al. [24] developed a CFD model for predicting the
performance of annular reactors with surface reaction. They 3CO þ 6H2 ! C3 H6 þ 3H2 O (3)
studied several hydrodynamic models to predict the kinetic
behavior of the reactor for laminar and turbulent flow regimes.
4CO þ 8H2 ! C4 H8 þ 4H2 O (4)
The present study has been focused on the CFD analysis of the
Fischer–Tropsch synthesis in a fixed bed reactor for production of
ethylene and propylene using the discrete packing method. The 5CO þ 10H2 ! C5 H10 þ 5H2 O (5)
effects of process variables such as reaction temperature, feed flow
The intrinsic reaction rates were described by using the general
rate, and H2/CO ratio on CO conversion and selectivity of products
equations and a function of the concentrations of carbon monoxide
have been investigated under laminar flow conditions.
and hydrogen as follows:
2. Experimental set up and kinetic study  
Ei m n
r i ¼ ki exp CCO CH2 (6)
RT
Experimental studies of the FTS for production of ethylene and
propylene on the optimal catalyst of cobalt–manganese were According to experimental measurements [25], the values of
reported in our previous work [25]. The characteristics of catalyst kinetic parameters obtained of linear regression analysis are given
such as X-ray diffraction (XRD), Scanning Electron Microscopy in Table 1.
1914 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920

Fig. 1. Schematic representation of the reactor used: (1) gas cylinders, (2) pressure regulators, (3) needle valves, (4) mass flow controllers (MFC), (5) monometers, (6) non
returns valves, (7) ball valves, (8) tubular furnace, (9) reactor, (10) catalyst bed, (11) trap, (12) condenser, (13) silica gel column, and (14) gas chromatograph (GC).

3. CFD modeling In this equation Sm contains the mass added through phase
changes or user defined sources. In general, and in the simulations
Modeling and simulation are the essential tools in the analysis described here, the source term was equal to zero.
and scale-up of reactors. The model must give the spatial
distribution of reactants, catalysts, inerts, and products in detail 3.1.2. Momentum equation
at all times. For better investigation of fluid flow and transport Navier–Stokes equations have been used for the description of
phenomena, a 3D model is necessary. It is important to realize, that momentum equations. The equation for conservation of momen-
as 3D models are more complex, they require more computational tum in the i-direction is defined by:
resource [29].
@ðrui Þ @ðrui u j Þ @P @t i j
þ ¼ þ þ rg i (8)
3.1. Governing equations @t @x j @xi @x j
where rgi is the gravitational body force, tij is the stress tensor. The
There are several commercial CFD codes such as finite element,
stress tensor for a Newtonian fluid is defined by:
finite difference and finite volume for process simulation. In this
study, we used finite volume technique for solving the equations.   
@ui @u j 2 @ul
The governing equations consist of: ti j ¼ m þ  m d (9)
@x j @xi 3 @xl i j

3.1.1. Continuity equation where dij is the Kronecker delta function.

The general form of continuity equation can be expressed as
follows: 3.1.3. Energy equation
The energy equation is solved in the form of a transport
@r @ðrui Þ
þ ¼ Sm (7) equation for static temperature. The temperature equation is
@t @xi obtained from the enthalpy equation, by taking the temperature as
Table 1 a dependent variable. The enthalpy equation is defined as:
Values of kinetic parameters. 0 1
Product ki Ei (J/kg mol) m n @ðrhÞ @ðrui hÞ @ @ @T X @p
þ ¼ kf  hi Ji j A þ ui þ Sh (10)
CH4 3.30551  106 27.38  106 0.30135 1.0951
@t @xi @xi @xi j
@xi
C2H4 2.65611  104 46.31  106 0.42481 0.52695
C3H6 3.21193  104 45.34  106 0.51780 0.54717 In this equation Sh includes heat of chemical reaction, any
C4H8 1.47879  105 33.88  106 0.98977 0.38180 interphase exchange of heat, and any other user-defined volumet-
C5H10 2.53525  106 30.08  106 0.39915 0.80525
ric heat sources.
 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920 1915

3.1.4. Species transport equation reactor outlet, outlet flow boundary condition was used. In this
The local mass fraction of each species is predicted by using the boundary condition, the flow field is assumed to be fully
solution of a convection–diffusion equation. The species transport developed. No-slip boundary condition was applied at the walls.
equation takes the following general form: The pressure–velocity coupling was performed by the SIMPLE
scheme [32]. To achieve higher accuracy results, Second-order
@ @ðrui Y i Þ @Ji j upwind schemes were used for the advection terms in the
ðrY i Þ þ ¼ þ Ri (11)
@t @xi @xi momentum, energy and species equations [33].
where Jij is the diffusion flux of species i in the j direction and Ri is
the net rate of production of species i by chemical reaction. The 4. Results and discussion
diffusion flux of species i was calculated based on Maxwell–Stefan
Equations. The kinetic theory was applied for description of mass In this section, the CFD results are compared with the
diffusion coefficient. The mass diffusion coefficient using kinetic experimental data. The effects of reaction temperature, mass flow
theory is calculated from a modification of the Chapman–Enskog rate, and H2/CO ratio on carbon monoxide conversion and
formula: selectivity of products are explained in a temperature rang of
463.15–523.15 K, mass flow rate of 1.9  106–6.5  106 kg/s,
1=2 and H2/CO ratio of 1–3. For study of the effect of carbon monoxide
½T 3 ðð1=Mw;i Þ þ ð1=M w; j ÞÞ
Di j ¼ 0:00188 (12) and hydrogen concentrations on the mass fraction of products, the
P abs s 2i j VD
carbon monoxide mole fraction was varied from 0.13 to 0.4 and the
where Pabs is the absolute pressure, s is the Lennard–Jones’ hydrogen mole fraction was varied from 0.2 to 0.6.
characteristic length, and VD is the diffusion collision integral,
which is a measure of the interaction of the molecules in the 4.1. Effect of reaction temperature on CO conversion
system. VD is a function of the quantity TD , where
The effects of reaction temperature on carbon monoxide
T conversion have been shown in Fig. 2. It can be seen from Fig. 2
TD ¼ (13)
ðe=kB Þi j that with increasing reaction temperature, the conversion of
carbon monoxide gradually increases. According to kinetic theory
where kB is the Boltzmann constant, and (e/kB)ij is the Lennard–
of gases, the kinetic energy of molecules increases with increasing
Jones’ energy parameter. The Lennard–Jones parameters for each
temperature, consequently the value of CO consumption increases.
species have been mentioned in Table 2 [30,31].
On the other hand, the increasing temperature promotes CO
dissociation, and provides more surface C atoms for forming
3.2. Geometry and boundary conditions
hydrocarbons. The percent deviations for CO conversion between
the experimental data and CFD simulations are 3.57%, 5.04%, 4.11%,
The 3D geometry comprising 44 spherical particles (four layers
7.05%, 6.81% for temperatures of 463.15 K, 473.15 K, 483.15 K,
with a 608 rotation around the reactor axis within each layer) with
493.15 K, and 523.15 K, respectively.
0.0038 m diameter based on discrete packing method was used in
CFD simulations. In this approach, we can achieve the actual flow
4.2. Effect of mass flow rate on CO conversion
field between the particles into the bed.
The dimensions of this geometry are the same to the
Fig. 3 shows the effects of feed flow rate on carbon monoxide
dimensions of experimental set-up [25]. The geometry and
conversion at different mass flow rates. As shown in Fig. 3, CO
unstructured tetrahedral mesh were generated by using commer-
conversion decreases with increasing mass flow rate because the
cial software, GAMBIT 2.0.4. The grid independence study was
residence time of reactants decreases on the catalyst surface into
performed. Finally the total number of controlled volumes was
the bed. Also, due to increase of flow channeling with increasing
674,806. In the present work, the continuity, momentum, energy,
mass flow rate, the conversion of carbon monoxide decreases.
and species equations were solved by using commercial CFD
package, ANSYS FLUENT 12.0.16, under laminar flow conditions
4.3. Effects of mass flow rate and reaction temperature on product
due to low Reynolds numbers. Physical properties of gaseous
selectivity
species were defined as temperature-dependant. The incompress-
ible ideal gas law was applied for dependency of density on the
The effects of mass flow rate, shown in Fig. 4, on the product
temperature. For mixture material, mass weighted-mixing law for
selectivity for two temperatures, i.e., 463.15 K and 483.15 K have
thermal conductivity and viscosity, mixing-law for specific heat
capacity were selected. It was assumed that the reactions occur
only on the surface of particles.
The boundary conditions were used at steady-state flow and
equal to the boundary conditions for physical experiments. The
condition of mass flow rate was used at the reactor inlet. At the

Table 2
Values of Lennard–Jones parameters.

Component si  1010 (m) (e/kB) (K)

CO 4 100
H2 2.92 38
CH4 3.758 148.6
C2H4 4 100
C3H6 4 100
C4H8 4 100
C5H10 4 100
Fig. 2. Carbon monoxide conversion as a function of reaction temperature, reaction
H2O 2.605 572.4
conditions: T = 463.15–523.15 K, H2/CO = 2, mass flow rate = 1.9  106 kg/s.
1916 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920

catalysts into the bed are significantly increased. The trend in Fig. 4
shows that the selectivity of products depends highly on
temperature and mass flow rate.

4.4. Effect of H2/CO ratio on product mass fraction

Hydrogen to carbon monoxide (H2/CO) ratio has a very

important effect on the mass fraction of species. Fig. 5 shows the
effect of H2 mole fraction on the mass fraction of products at the
CO mole fraction of 0.2. This obviously shows an important role
of H2 mole fraction on the mass fraction of products. It can be
seen that the mass fraction of products increases with increasing
H2 mole fraction. As H2 mole fraction increased, the H2 coverage
increased and CO coverage decreased. Consequently, the H2
Fig. 3. Carbon monoxide conversion as a function of mass flow rate, reaction
adsorption rate and FTS activity increased. As shown in Fig. 6,
conditions: T = 463.15 K, H2/CO = 2, mass flow rate = 1.9  106–6.5  106 kg/s.
the mass fraction of products exhibits negative trends with
increasing CO mole fraction at the H2 mole fraction of 0.4. The
been illustrated at different mass flow rates. It clearly indicates FTS activity achieved maximum at H2/CO ratio of 3. When H2/
that the selectivity of products increases as the temperature CO = 1, the CO adsorption rate is higher than hydrogen
increases. It is also observed that the selectivity of products adsorption rate, therefore, the scarce of H2 cause low
decreases with increasing mass flow rate. As mentioned before, the activity for FTS. The mass fraction of products also increases
residence time of reactants decreases with increasing flow rate. On with increasing reaction temperature because the reactions
the other hand, at higher mass flow rate, vacant sites on surface of progress.

Fig. 4. Selectivity of products as a function of mass flow rate at different temperatures, reaction conditions: T = 463.15 K and 483.15 K, H2/CO = 2, mass flow rate = 1.9  106–
6.5  106 kg/s.
 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920 1917

Fig. 5. Mass fraction of products as a function of H2 mole fraction, reaction conditions: T = 463.15–523.15 K, CO mole fraction = 0.2, mass flow rate = 1.9  106 kg/s.

4.5. Concentration contours of products concentration of products varies at the surface of particles. As
shown in Fig. 7, the molar concentration of products increases
Fig. 7 shows the predicted concentration contours of products along the length of the bed. From these results, it can be
at the surface of particles inside the reaction zone. Reactants flow concluded that the inhomogeneity in the flow around a catalytic
in the z direction. The contours of products at other conditions particle has a significant influence on the concentration
are similar to those shown in Fig. 7. It is seen that the distribution in the reaction zone.
1918 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920

Fig. 6. Mass fraction of products as a function of CO mole fraction, reaction conditions: T = 463.15–523.15 K, H2 mole fraction = 0.4, mass flow rate = 1.9  106 kg/s.
 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920 1919

Fig. 7. Contours of molar concentration of: (a) CH4, (b) C2H4, (c) C3H6, reaction conditions: T = 483.15 K, H2/CO = 2, mass flow rate = 1.9  106 kg/s. legend shows molar
concentration (kmol/m3).

5. Conclusions concluded that the value of product is not only controlled by the
reaction temperature but also by the feed ratio. In general, we can
A 3D computational fluid dynamics model on the basis of predict the distribution of products at different operating
discrete packing method was used to predict the conversion of conditions by using the CFD method.
carbon monoxide and distribution of products in the Fischer–
Tropsch synthesis process. The simulations were performed in the Appendix A
fixed bed reactor. The reactions were assumed to be surface
reaction type in the simulations. The effects of reaction tempera- The CO conversion (%) was calculated from the following formula:
ture, feed flow rate and H2/CO ratio on the value of products were
investigated under laminar flow conditions. It was found that the
reaction conditions strongly influence the product distribution. It Mole of COin  Mole of COout
CO conversion ð%Þ ¼  100
was observed that the conversion of carbon monoxide increased Mole of COin
with increasing temperature, while decreased with increasing
mass flow rate. The selectivity and mass fraction of products also The selectivity (%) towards the individual components on carbon-
increased with increasing temperature and H2/CO ratio. It can be basis was calculated according to:
1920 A.R. Miroliaei et al. / Journal of Industrial and Engineering Chemistry 18 (2012) 1912–1920

[16] M.R. Rahimpour, M.H. Khademi, A.M. Bahmanpour, Chemical Engineering Science
Mole of i product 65 (2010) 6206.
Selectivity of i product ð%Þ ¼
Mole of COin  Mole of COout [17] M. Zakeri, A. Samimi, M. Khorram, H. Atashi, A. Mirzaei, Powder Technology 200
(2010) 164.
 100 [18] J.A.M. Kuipers, W.P.M. van Swaaij, Advances in Chemical Engineering 24 (1998)
227.
[19] Y. Jiang, M.R. Khadilkar, M.H. Al-Dahhan, M.P. Dudukovic, AIChE Journal 48 (2002)
References 716.
[20] D.I. Gerogiorgis, B.E. Ydstie, Chemical Engineering Research and Design 86 (2005)
[1] M. Ahmadi Marvast, M. Sohrabi, S. Zarrinpashne, Gh. Baghmisheh, Chemical 603.
Engineering and Technology 28 (2005) 78. [21] A.G. Dixon, M.E. Taskin, E.H. Stitt, M. Nijemeisland, Chemical Engineering Science
[2] A. Jess, C. Kern, Chemical Engineering and Technology 32 (2009) 1164. 62 (2007) 4963.
[3] S.H. Kang, J.W. Bae, K.J. Woo, P.S. Sai Prasad, K.W. Jun, Fuel Processing Technology [22] L. Shi, D.J. Bayless, M.E. Prudich, International Journal of Hydrogen Energy 34
91 (2010) 399. (2009) 7666.
[4] G. Fan, B. Zou, S. Cheng, L. Zheng, Journal of Industrial and Engineering Chemistry [23] G. Arzamendi, P.M. Dieguez, M. Montes, J.A. Odriozola, E.F.S. Aguiar, L.M. Gandia,
16 (2010) 220. Chemical Engineering Journal 154 (2009) 168.
[5] Y.Y. Ji, H.W. Xiang, J.L. Yang, Y.Y. Xu, Y.W. Li, B. Zhong, Applied Catalysis A: General [24] J. Esteban Duran, M. Mohseni, F. Taghipour, Chemical Engineering Science 65
214 (2001) 77. (2010) 1201.
[6] S.H. Kang, K.J. Woo, K.W. Jun, Y. Kang, Korean Journal of Chemical Engineering 26 [25] H. Atashi, F. Siami, A.A. Mirzaei, M. Sarkari, Journal of Industrial and Engineering
(2009) 1533. Chemistry 16 (2010) 952.
[7] A.P. Raje, B.H. Davis, Catalysis Today 36 (1997) 335. [26] R. Zennaro, M. Tagliabue, C.H. Bartholomew, Catalysis Today 58 (2000) 309.
[8] E. van Steen, H. Schulz, Applied Catalysis 186 (1999) 309. [27] H. Atashi, M. Mansouri, S.H. Hosseini, M. Khorram, A.A. Mirzaei, M. Karimi, G.
[9] J.W. Bae, S.J. Park, S.H. Kang, Y.J. Lee, K.W. Jun, Y.W. Rhee, Journal of Industrial and Mansouri, Korean Journal of Chemical Engineering 29 (2012) 304.
Engineering Chemistry 15 (2009) 798. [28] I.C. Yates, C.N. Satterfield, Energy and Fuels 5 (1991) 158.
[10] H.E. Atwood, C.O. Bennett, Industrial and Engineering Chemistry Process Design [29] A.G. Dixon, M. Nijemeisland, E.H. Stitt, Industrial and Engineering Chemistry
and Development 18 (1979) 163. Research 44 (2001) 6342.
[11] G. Bub, M. Baerns, B. Bussemeier, C. Frohning, Chemical Engineering Science 33 [30] A.G. Dixon, M. Nijemeisland, E.H. Stitt, Advances in Chemical Engineering 31
(1980) 348. (2006) 307.
[12] Q.S. Liu, Z.X. Zhang, J.L. Zhou, Journal of Natural Gas Chemistry 8 (1999) 137. [31] ANSYS, Inc., ANSYS FLUENT 12.0.16, Theory Guide: Species Transport and Finite-
[13] J. Pina, N.S. Schbib, V. Bucala, D.O. Borio, Industrial and Engineering Chemistry rate Chemistry, ANSYS, Inc., 2009.
Research 40 (2001) 5215. [32] S.V. Patankar, Numerical Heat Transfer and Fluid Flow, Hemisphere, Washington,
[14] Y.N. Wang, Y.Y. Xu, Y.W. Li, Y.L. Zhao, B.J. Zhang, Chemical Engineering Science 58 DC, 1980.
(2003) 867. [33] H.K. Versteeg, W. Malalasekera, An Introduction to Computational Fluid Dynam-
[15] R. Guettel, T. Turek, Chemical Engineering Science 64 (2009) 955. ics, The Finite Volume Method, Longman, Essex, England, 1995.