Chemical Engineering Journal 171 (2011) 1387–1398

Contents lists available at ScienceDirect

Chemical Engineering Journal

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

Kinetic modeling of ethylbenzene dehydrogenation over hydrotalcite catalysts

Luqman Atanda, Nabil Al-Yassir, Sulaiman Al-Khattaf ∗
KAUST Center in Development, King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia

a r t i c l e i n f o a b s t r a c t

Article history: Kinetics of ethylbenzene dehydrogenation to styrene was investigated over a series of quaternary mixed
Received 30 January 2011 oxides of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) catalysts prepared by calcination of hydrotalcite-
Received in revised form 13 April 2011 like compounds and compared with commercial catalyst. The study was carried out in the absence
Accepted 12 May 2011
of steam using a riser simulator at 400, 450, 500 and 550 ◦ C for reaction times of 5, 10, 15 and 20 s.
Mg3 Fe0.25 Mn0.25 Al0.5 afforded the highest ethylbenzene conversion of 19.7% at 550 ◦ C. Kinetic parame-
Keywords:
ters for the dehydrogenation process were determined using the catalyst deactivation function based
Hydrotalcite
on reactant conversion model. The apparent activation energies for styrene production were found to
Dehydrogenation
Ethylbenzene
decrease as follows: E1-Ni > E1-Co > E1-Mn .
Kinetic modeling © 2011 Elsevier B.V. All rights reserved.

1. Introduction
The increasing application of styrene monomer in the produc-
tion of polystyrene, ABS resins, synthetic rubbers, etc. has made
the demand for styrene increase tremendously with current pro-
duction capacity exceeding 25 × 106 tonnes/year [1]. Styrene is
manufactured commercially through two process routes – cat-
alytic dehydrogenation of ethylbenzene and coproduction with
propylene oxide from the reaction of propene with oxidized ethyl-
benzene to ethylbenzenedroperoxide, with the former accounting
for 90% of world production capacity. Catalytic dehydrogenation
of ethylbenzene (EB) is carried out over supported and unsup-
ported Fe-based catalyst in the presence of superheated steam
at temperature of 600–700 ◦ C. The process is carried out at high
steam-to-hydrocarbon ratio to increase activity, selectivity and sta-
bility of the catalyst [2]. However, the use of excess steam consumes
high energy which has been estimated to be 1.5 × 109 cal/styrene
tonne [3]. In addition to that, it was found that most examined cata-
lysts suffered from low stability, fast deactivation and subsequently
low catalytic activities [4–7].
Therefore, there is need to develop new catalyst systems that
allows for uniform dispersion of the active phase as well as high
stability. Hydrotalcite-like compounds (HTlcs) as precursors for
mixed oxide catalysts offer unique basic properties that make
them very attractive for catalytic applications. Fresh hydrotalcites
(HTs) have high water content which often makes them inactive
Abbreviations: EB, ethylbenzene; cat, catalyst; DEB, diethylbenzene; HT, hydro-
talcite; HTlc, hydrotalcite-like compound.
∗ Corresponding author. Tel.: +966 3 860 1429; fax: +966 3 860 4234.
E-mail address: skhattaf@kfupm.edu.sa (S. Al-Khattaf).
for certain base-catalyzed reactions [8,9], thus they are activated
by thermal decomposition. Therefore, the resultant oxides after
calcinations offer improved catalytic performance due to the for-
mation of small crystal size stable to thermal treatments. The
oxides also possess high surface area, basic properties and further
form small and thermally stable metal crystallites by reduction
[10]. Moreover, they are potentially recoverable and recyclable
[11]. Mixed oxides of thermally decomposed HT precursors are
potential catalysts with growing applications in various reactions
as reviewed by Cavani et al. [10]. HTs of several reducible biva-
lent (Ni, Cu, Co) and trivalent (Fe, Cr) cations in their structure
together with the classical ones (Mg, Zn, Al) serves as precur-
sors for the preparation of different mixed oxides active for
oxidation and hydrogenation/dehydrogenation reactions [12]. Sup-
ported iron catalyst prepared from Mg–Fe–Al HTlcs as precursor
has been successfully applied in the EB dehydrogenation to styrene
[6]. Upon calcinations, mixed oxides with a high surface area as
well as mesoporous character were formed. XRD analysis indicated
mainly the presence of periclase Mg(Fe, Al)O phase. Reduction pre-
treatment afforded the catalyst with somewhat higher activity as
well as higher stability. EB conversion of 60% and a styrene selec-
tivity of 95% were stably observed over Mg3.0 Fe0.5 Al0.5 at 550 ◦ C. A
recent study on the activity of various FeOx –MeOy /Mg(Al)O cat-
alysts derived from HTs showed FeOx –CoOy /Mg(Al)O have high
activity due to the formation of Fe–Co bimetallic species [13]. Addi-
tion of Co further facilitated the reduction–oxidation of Fe3+ /Fe2+ ,
and the active Fe3+ species was stabilized. The role of MgO as a
base support has been reported effective to initiate dehydrogena-
tion by H+ abstraction on the basic site, and accelerated by the
reduction-oxidation of Fe3+ active species [14].
Carrà and Forni [15] performed kinetic studies in the tem-
perature range of 770–900 K over the industrial catalyst, Shell

1385-8947/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cej.2011.05.053
1388 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398
Nomenclature
Ai pre-exponential factor for the ith reaction (m3 /kg of
catalyst s)
Ci concentration of specie i in the riser simulator
(mol/m3 )
CL confidence limit
Ei apparent activation energy of the ith reaction
(kJ/mol)
Kc equilibrium constant
ki apparent rate constant for the ith reaction (m3 /kg of
catalyst s)
k0i pre-exponential factor for the ith reaction after re-
parameterization (m3 /kg of catalyst s)
MWi molecular weight of specie i
r2 correlation coefficient
ri rate of reaction for species i
R universal gas constant (kJ/kmol K)
t reaction time (s)
T reaction temperature (K)
To average temperature of the experiment
V volume of the riser (45 cm3 )
Wc mass of the catalyst (0.81 g)
Whc total mass of the hydrocarbon injected the riser
(0.162 g)
yi mass fraction of ith component
Greek letters
ϕ apparent deactivation function
 catalyst deactivation constant (RC model)
105. The intrinsic rate of styrene formation was developed, based
upon Langmuir–Hinshelwood kinetics with activation energy of
45 kcal/mol. Coulter et al. [16] studied the kinetics using unpro-
moted and K-promoted polycrystalline catalysts. The unpromoted
catalyst yielded apparent activation energy of 39 kcal/kmol while
for K-promoted catalyst, the activation energy was 36 kcal/mol.
The authors concluded that the active sites of unpromoted and
promoted catalysts are identical which was later confirmed by
Shekhah et al. [17] As found in the work of Addiego et al. [18], the
increase of potassium loading initially decreases the apparent acti-
vation energy to 21 kcal/mol while further addition of potassium
led to an increase of the apparent activation energy to 34 kcal/mol.
Shekhah et al. [17] suggested that the increase of potassium loading
leads to the decrease in initial conversion rate due to the cov-
erage of active sites by excess potassium. Miyakoshi et al. [19]
studied the effect of some transition metals on Fe-K catalyst pre-
pared through sol–gel method; Mn was reported to enhance the
catalytic activity and suppressed coke formation. The activation
energies determined from the Arrhenius plot are 22.4 kcal/mol and
21.9 kcal/mol for 20% Mn-substituted Fe-K oxide and Fe-K oxide,
respectively. The authors concluded that the increased catalytic
activity resulted from the increased number of active sites as it
was observed that the difference in the activation energies was
not appreciable. Lee and Froment [20] developed a set of intrinsic
rate equations for a commercial potassium-promoted iron cata-
lyst based on Hougen–Watson model. Apparent activation energy
of 41.9 kcal/mol was obtained. The model was applied in the sim-
ulation of the dehydrogenation in industrial multibed adiabatic
reactors with either axial or radial flow, or accounting also for
thermal radical-type reactions, internal diffusion limitations, coke
formation, and gasification.
The use of fluidized bed has been suggested to improve EB
conversion [21]. The application of selective membranes further
increases styrene selectivity and consequently reduces the yield
of toluene due to the removal of hydrogen through the mem-
branes. It was shown that increase in styrene production over that
of the industrial fixed bed unit is achievable, if design and operating
parameters are properly chosen. In view of this, the present study
is aimed at investigating the kinetics of EB dehydrogenation over
quaternary mixed metal oxide catalysts of Mg3 Fe0.25 Me0.25 Al0.5
(Me = Co, Mn and Ni) obtained from HT precursors in a CREC flu-
idized riser simulator operated at temperatures below that of the
industrial process, as previous studies [6] have shown that opti-
mum condition was achieved at 550 ◦ C. Injection of the reactant and
withdrawal of products were for short period of time. It was envis-
aged that for short reaction time, the potential of side reactions
taking place will be limited. Kinetic modeling of the dehydrogena-
tion process was done using the catalyst activity decay function
based on reactant conversion.
2. Experimental
2.1. Materials
All the precursors used in the synthesis of the HT quater-
nary mixed oxides Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) were
obtained from Sigma–Aldrich and used as supplied. The chemical
feedstock (ethylbenzene, 99%) was also supplied by Sigma–Aldrich.
2.2. Catalyst preparation
The HT precursors of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and
Ni) catalysts were prepared by co-precipitation of metal nitrates,
following the method by Miyata and Okada [22] after minor modi-
fications. An aqueous solution containing the nitrates of Mg2+ , Fe3+ ,
Me2+ (Co2+ , Mn2+ or Ni2+ ) and Al3+ (200 ml) was added slowly with
vigorous stirring into an aqueous solution of sodium carbonate
(400 ml). The pH of the solution was adjusted at 10.0 by dropping
a 1 M aqueous solution of sodium hydroxide, leading to a precipi-
tation of heavy slurry. After the solution was aged at 60 ◦ C for 24 h,
the solution was filtrated and the precipitates were washed with
de-ionized water (1000 ml), dried in air at 100 ◦ C for 4 h, and cal-
cined at 550 ◦ C for 12 h in a muffle furnace in a static air atmosphere.
The concentration of Na2+ in the catalysts after the calcination was
confirmed to be less than 10 ppm by atomic absorption.
2.3. Catalyst characterization
All catalyst characterizations were performed for calcined cata-
lysts, which were calcined at 550 ◦ C for 12 h.
The chemical compositions of synthesized HT quaternary mixed
oxides of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) were deter-
mined by atomic absorption spectroscopy, using the Perkin-Elmer
equipment (Model AAnalyst 100).
Surface areas were calculated by the Brunauer–Emmett–Teller
(BET) method. N2 adsorption–desorption isotherms at −196 ◦ C
were measured using a conventional volumetric apparatus (Bel
Japan, BELSORP Mini). Adsorption measurements were done after
samples (ca. 0.1 g) were heated at 400 ◦ C for 10 h under nitrogen
flow.
XRD was recorded on a Mac Science MX18XHF-SRA pow-
der diffractometer with monochromatized Cu K␣ radiation
( = 0.154 nm) at 40 kV and 30 mA. The diffraction pattern was iden-
tified through comparison with those included in the JCPDS (Joint
Committee of Powder Diffraction Standards) database and litera-
tures.
H2 -TPR of the calcined catalysts was carried out by flow-
ing 5 vol.% H2 /N2 (30 cm3 min−1 ) in the temperature range of
100–900 ◦ C. The sample temperature increased with a rate of
 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398 1389

Table 1
Specific surface area, metal composition, and XPS analytical data of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni).

Catalyst SBET (m2 gcat −1 ) Molar ratioa Binding energyb (eV) Surface molar ratiob,c

Mg/Fe Mg/Al Mg/Me Mg 2p3/2 Fe 2p3/2 Me 2p3/2 Mg/Fe Mg/Me

Mg3 Fe0.25 Co0.25 Al0.5 187 11.5 5.5 11.5 48.9 710.6 779.6 6.5 5.0
Mg3 Fe0.25 Mn0.25 Al0.5 158 11.1 5.1 11.6 49.0 710.6 641.5 7.1 8.3d
Mg3 Fe0.25 Ni0.25 Al0.5 197 10.2 5.8 11.3 49.0 710.6 853.8 8.0 2.4
a
By atomic absorbance analyses.
b
By XPS analyses.
c
After ASF correction.
d
The data scattered between 6.25 and 12.5.

10 ◦ C min−1 and the amount of H2 consumed was monitored by a
thermal conductivity detector of a gas chromatograph (Shimadzu,
GC-8AIT).
XPS measurements were performed on a Perkin Elmer 1600E
spectrometer using Mg K␣ radiation as excitation source. In charge-
up correction, the calibration of binding energy (BE) of the spectra
was referenced to the C 1s electron bond energy corresponding to
graphitic carbon at 284.5 eV. In addition, relative atomic sensitivity
factors (ASF) were corrected to determine practically more accurate
chemical compositions on the surface.
2.4. Catalytic experiments
2.4.1. Experimental set-up (riser simulator)
All of the experimental runs were carried out in the riser simu-
lator reactor. This reactor is novel bench-scale equipment with an
internal recycle unit invented by de Lasa [23]. The detailed descrip-
tion of the simulator riser has been previously described [24].
2.4.2. Testing procedures and product identifications
In a typical experiment, the reactor was charged with 0.8 g of
calcined catalyst previously crushed and sieved to a particle size
of ∼ 60 ␮m diameter. The catalyst/reactant ratio was 5 (weight
of catalyst = 0.81 g, weight of reactant injected = 0.162 g). The reac-
tion temperature was varied in the range of 400–550 ◦ C, and the
reaction time was varied from 5 to 20 s. For all experimental runs
at pre-determined temperature, pressure changes are monitored
through pressure transducers installed in both the reactor and the
vacuum box chambers. The feedstock was injected into the reac-
tor at ∼14.7 psi (atmospheric pressure). The feedstock vaporizes
almost instantaneously, causing an abrupt increase in pressure to
∼26 psi. After vaporization of the reactant, the pressure further
increased to ∼35 psi due to increased number of molecules, which
is as a result of dehydrogenation and cracking reaction. A sudden
drop in reactor pressure was noticed after the stipulated time of
reaction. The quantitative analysis of the reaction products were
carried out online using a Agilent GC with FID (Agilent Chromato-
graph Model 6890N), equipped with an HP-INNOWax capillary
column (polyethylene glycol (PEG)) (length 60 m × internal diam-
eter 0.32 mm × film thickness 0.50 ␮m). During the course of the
investigation, a number of runs were repeated to check for repro-
ducibility in the experiment results, which were found to be
excellent. Typical errors were in the range of ±2%.
The ethylbenzene (EB) conversion (mol% or wt%) is expressed as
follows:
moles of converted ethylbenzene
XEB = × 100
moles of ethylbenzene fed
The selectivity (to product i) is expressed as follows:
number of carbon atoms of product i
Si = × 100
number of carbon atoms of converted products
The yield (to product i) is expressed as follows:
Si
Yi = × 100
XEB
2.4.3. Coke analysis
The amount of coke deposited on the spent catalysts was deter-
mined by a common combustion method using a carbon analyzer
multi EA 2000 (Analytikjena). The multi EA 2000 with CS module
is a specially developed system to permit simultaneous or separate
determination of the total carbon and total sulfur in samples of
solids, pastes, and liquids by means of high temperature oxidation

A B

c c
Intensity (a.u.)
Intensity (a.u.)

b b

a a

0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100

2 theta (°) 2 theta (°)

Fig. 1. XRD patterns of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) after (A) drying and (B) calcination at 550 ◦ C (: hydrotalcites; : unknown and : periclase).
1390 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398

Toluene + CH 4

-1
50 µmol g cat +H2 k3
k1
EB Styrene + H 2
e
k-1
k2
H2 consumed

d
Benzene + C 2H 4

c Scheme 1. Catalytic reaction scheme of ethylbenzene (EB) dehydrogenation.

b the ability of Mg–Al and Mg–Fe to form the hydrotalcite structure,

a bulk manganese oxide particles. This was suggested by XRD and
0 100 200 300 400 500 600 700
Temperature / (°C)
Fig. 2. H2 -TPR of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn, and Ni) catalysts. (a)
Mg3 Fe1 ; (b) Mg3 Fe0.5 Al0.5 ; (c) Mg3 Fe0.25 Mn0.25 Al0.5 ; (d) Mg3 Fe0.25 Co0.25 Al0.5 ; and (e)
Mg3 Fe0.25 Ni0.25 Al0.5 .
in a current of oxygen supplied to the unit directly. A small amount
of the spent catalyst (0.09–0.1 g) is used for the analysis. The coke
laid out on the sample during reaction experiments is burned com-
pletely, converting the carbonaceous deposit into carbon dioxide.
The amount of coke formed is determined by measuring the num-
ber of moles of CO2 released. It is worth noting that hydrogen is also
produced and converted into water. However, this method does not
detect hydrogen, and thus it is difficult to determine the C/H and
consequently comment on the nature of coke.
3. Results and discussions
3.1. Physico-chemical properties
The textural properties along with metal composition of
Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) catalysts after the cal-
cination at 550 ◦ C are shown in Table 1. It can be clearly seen
from Table 1 that all Mg3 Fe0.5 Me0.5 Al0.5 catalysts exhibit large SBET
between 150 and 200 m2 gcat −1 . This can be mainly attributed to
resulting in large surface area after the calcination. The relatively
low SBET (158 m2 g−1 ) of Mg3 Fe0.25 Mn0.25 Al0.5 compared to other
catalysts is probably due to plugging of the pores with segregated
H2 -TPR, which are discussed in the following paragraphs. It is also
800 900 1000
worth mentioning that the metal compositions are in complete
agreement with the nominal values. This is due to the formation of
hydrotalcite structure as the precursors during the coprecipitation,
which accommodates many metal cations in the structure.
XRD patterns of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) cat-
alysts after drying are shown in Fig. 1A. It can be noted that all
Mg3 Fe0.25 Me0.25 Al0.5 showed typical hydrotalcite reflections, indi-
cating that the second metal components were incorporated into
the hydrotalcite structure except Mg3 Fe0.25 Mn0.25 Al0.5 . Mn2+ pos-
sesses a larger ionic radii of 0.80 Å than the other Me2+ [25] and
requires the highest pH for its complete precipitation as hydrox-
ides among the Me cations used [10]. After the calcination at 550 ◦ C
(Fig. 1B), all samples showed the periclase reflections together with
weak spinel reflections. This indicates that iron was incorporated
in the periclase and the spinel.
H2 -TPR data of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Mn, Ni, and Co) are
shown in Fig. 2. Mg3 Fe1 showed a reduction peak at 418 ◦ C and
a broad peak above 650 ◦ C (Fig. 2a). The first peak is attributed to
the reduction of spinel (MgFe3+ 2 O4 ) to periclase (Mg1−x Fe2+ x O) and
the latter is to the reduction of Mg1−x Fe2+ x O to Fe0 , as reported
by Shen et al. [26] Mg3 Fe0.25 Ni0.25 Al0.5 showed a weak reduction
peak at 353 ◦ C and broad reduction peak above 800 ◦ C. The latter
is attributed to the reduction of Ni2+ in the Mg(Ni)O periclase [27].
Therefore, the one observed at 353 ◦ C can be attributed the reduc-
tion of Fe3+ species. Mg3 Fe0.25 Mn0.25 Al0.5 exhibited rather strong
two reduction peaks at 257 ◦ C and 428 ◦ C (Fig. 2c). This catalyst

Table 2
Product distribution (wt%) at various reaction conditions for the dehydrogenation of ethylbenzene over Mg3 Fe0.25 Co0.25 Al0.5 mixed oxides catalyst.

Temp. (◦ C) Reaction time (s) EB conv. (%) Styrene Benzene Toluene Gases DEB

400 5 3.18 2.58 0.44 0.09 – 0.12

10 4.66 3.66 0.74 0.08 – 0.18
15 5.19 3.85 0.89 0.11 0.06 0.22
20 5.51 3.79 1.10 0.12 0.06 0.38

450 5 5.48 4.14 0.85 0.16 0.09 0.13

10 6.47 4.82 1.14 0.23 0.12 0.16
15 6.61 4.48 1.49 0.19 0.13 0.25
20 7.12 4.44 1.58 0.18 0.14 0.41

500 5 6.99 4.89 1.45 0.23 0.19 0.15

10 8.55 5.67 1.85 0.24 0.24 0.21
15 10.71 6.71 2.69 0.33 0.44 0.24
20 11.7 7.35 2.99 0.37 0.39 0.37

550 5 8.65 5.83 1.95 0.24 0.33 0.22

10 11.38 7.11 2.83 0.36 0.48 0.17
15 12.83 7.84 3.39 0.42 0.56 0.25
20 15.46 8.84 4.56 0.60 0.88 0.34
 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398 1391

Table 3
Product distribution (wt%) at various reaction conditions for the dehydrogenation of ethylbenzene over Mg3 Fe0.25 Mn0.25 Al0.5 mixed oxides catalyst.

Temp. (◦ C) Reaction time (s) EB conv. (%) Styrene Benzene Toluene Gases DEB

400 5 4.06 3.76 0.36 – – –

10 5.82 5.19 0.49 – – 0.04
15 6.25 5.65 0.55 – – 0.05
20 6.64 5.93 0.65 – – 0.06

450 5 5.64 4.97 0.55 0.05 – –

10 9.00 7.58 1.11 0.11 0.08 0.07
15 8.98 7.44 1.12 0.09 0.08 0.13
20 10.19 8.25 1.28 0.12 0.07 0.19

500 5 8.70 6.96 1.40 0.12 0.14 –

10 10.65 8.33 1.77 0.18 0.15 0.07
15 12.93 9.69 2.39 0.21 0.20 0.15
20 17.25 12.01 3.03 0.39 0.25 0.24

550 5 11.46 8.13 2.59 0.25 0.30 –

10 15.02 10.27 3.58 0.39 0.41 0.13
15 18.63 11.50 4.66 0.58 0.54 0.16
20 19.70 12.28 4.95 0.70 0.54 0.19

exhibited unknown reflections together with those of hydrotalcite
after drying (Fig. 1B), suggesting that manganese oxides were sep-
arated from periclase after calcination. It was reported that a mixed
oxide of Mg0.2 Mn1.8 Al1 composition derived from hydrotalcite and
calcined at 500 ◦ C showed two reduction peaks at 333 ◦ C and 460 ◦ C
[28]. This also suggests that manganese oxides were separated from
periclase structure in Mg3 Fe0.25 Mn0.25 Al0.5 and showed the reduc-
tion peaks at 257 ◦ C and 428 ◦ C.
The XPS analytical results of Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn
and Ni) after calcination at 550 ◦ C are summarized in Table 1. The
surface molar ratios of Mg/Fe and Mg/Me calculated from the XPS
results exhibited lower values than the bulk molar ratios obtained
by AA analyses (Table 1). This indicates that both Fe and second
component metal (Me) species are localized in the surface layer of
the catalyst particles. On all Mg3 Fe0.25 Me0.25 Al0.5 catalysts except
Mg3 Fe0.25 Ni0.25 Al0.5 , Fe species showed similar surface enriched
localization, i.e., almost 1.4 times higher concentration in the sur-
face layer than the bulk phase. Surface enrichment of Fe was not
so significant for Mg3 Fe0.25 Ni0.25 Al0.5 . Contrarily, some of the sec-
ond component metals exhibited far enriched distribution in the
surface layer, i.e., the highest concentration in the surface layer
was obtained by Ni (5 times), followed by Co (2.2 times) and Mn
(1.3 times). The scattered experimental values for Mn clearly sug-
gest that Mn dispersion on the catalyst was not homogeneous.
Detailed synthetic approaches and physico-chemical properties of
these hydrotalcite based materials are given in Balasamy et al. [13].
3.2. Catalytic activity
The results from the dehydrogenation of EB over
Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) catalysts at differ-
ent reaction temperatures and reaction times are presented in
Tables 2–4. The experimental results showed styrene to be the
main reaction product, although small amounts of benzene were
obtained as by product (Fig. 3). Traces of toluene, gaseous hydro-
carbons, diethylbenzene (DEB) were also observed; however the
yields of these products were neglected in subsequent analysis. A
possible reaction scheme is shown in Scheme 1. In all the catalysts
investigated, EB conversion increased with both reaction temper-
ature and time and maximum conversion of approximately 15.5,
19.7, and 18.0% was attained at 20 s of time on stream and temper-
ature of 550 ◦ C for Mg3 Fe0.25 Co0.25 Al0.5 , Mg3 Fe0.25 Mn0.25 Al0.5 , and
Mg3 Fe0.25 Ni0.25 Al0.5 catalysts respectively as shown in Fig. 4(A–C).
The mixed oxides catalyst consisting of Mn gave the highest activ-
ity regardless of its low surface area. This may be accounted for by
the isolation of manganese oxides from the periclase structure, as
suggested by XRD, XPS and H2 -TPR. Therefore, the inhomogeneous
dispersion of Mn on the catalyst (i.e. presence of segregated (bulk)
Mn oxide) is assumed to increase the active sites of the catalyst.
One may suggest that the segregated Mn oxide further contributes
to the basicity of MgO which has been reported to initiate H+
abstraction [14]. The investigation of iron oxide catalysts contain-
ing various transition metal oxides by Hirano [29] also showed

Table 4
Product distribution (wt%) at various reaction conditions for the dehydrogenation of ethylbenzene over Mg3 Fe0.25 Ni0.25 Al0.5 mixed oxides catalyst.

Temp. (◦ C) Reaction time (s) EB conv. (%) Styrene Benzene Toluene Gases DEB

400 5 2.14 1.70 0.35 0.05 – 0.06

10 3.23 2.39 0.58 0.07 – 0.21
15 3.86 2.64 0.76 0.08 – 0.27
20 4.21 2.78 0.89 0.12 0.05 0.32

450 5 4.08 3.17 0.65 0.07 0.08 0.07

10 4.82 3.36 0.88 0.13 0.07 0.23
15 6.25 4.24 1.23 0.14 0.12 0.32
20 7.01 4.59 1.52 0.17 0.17 0.52

500 5 5.75 4.45 0.99 0.12 0.15 0.06

10 8.14 5.58 1.68 0.19 0.25 0.24
15 9.74 6.26 2.23 0.25 0.31 0.40
20 10.87 7.05 2.46 0.47 0.35 0.36

550 5 10.04 6.86 2.12 0.35 0.39 0.13

10 13.27 8.74 3.03 0.44 0.55 0.20
15 16.17 10.18 3.82 0.67 0.67 0.31
20 18.04 11.16 4.31 0.86 0.76 0.37
1392 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398
Fig. 3. Reactions occurring during EB dehydrogenation over Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni).
that increase in basicity due to reduction in electronegativity of the
promoter gave rise to the dehydrogenation activity of the catalyst.
As a control, a commercial catalyst was also tested. It showed a low
activity as shown in Fig. 5 which can be attributed to the absence
of steam.
3.2.1. Styrene/benzene selectivity
The product selectivity during the dehydrogenation of EB over
Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) and commercial cata-
lysts were compared in Fig. 6 at constant conversion level of ∼6%.
The results show that styrene is obtained as the most predominant
product on all the HT catalysts with the best selectivity observed
in Mg3 Fe0.25 Mn0.25 Al0.5 . The selectivity towards styrene exhibits
a trend opposite to EB conversion for all the HT catalysts inves-
tigated. It decreased when the reaction temperature was raised
due to cracking reaction. Fig. 7 shows selectivity as a function of
the EB conversion for all the reaction temperatures for Mn substi-
tuted catalyst, showing a maximum styrene selectivity of 92.7% at
reaction temperature of 400 ◦ C which corresponds to 4.9% EB con-
version. However, benzene selectivity increased as the temperature
was raised. This confirmed that cracking of EB to benzene becomes
pronounced at high temperature as shown in Fig. 7. A maximum
benzene selectivity of 25.1% was attained at 550 ◦ C corresponding
to 19.7% EB conversion. On the contrary, the commercial catalyst
has low selectivity towards styrene which was due to significant
cracking and such an occurrence has been suggested by Lee [30]
to be as a result of increase in acidity of the catalyst due to loss of
potassium from the catalyst surface.
It is also worth mentioning that Mn based catalyst exhibited
much lower selectivity to diethylbenzene (DEB, three isomers) than
the other two catalysts (Co and Ni). Therefore, it can be inferred
from these results that Mn does not favor the disproportionation
reaction. For instance, the selectivity to DEB (temperature = 400 ◦ C
and residence time = 20 s) over Mn, Co, and Ni based oxides were
1.0 wt%, 6.9 wt% and 7.6 wt%, respectively (Tables 2–4).
3.2.2. Styrene/benzene yield
The yield of styrene was found to increase with both
temperature and time, in a similar manner with EB con-
version. Maximum styrene yield of approximately 8.8, 12.3,
and 11.2% for Mg3 Fe0.25 Co0.25 Al0.5 , Mg3 Fe0.25 Mn0.25 Al0.5 and
Mg3 Fe0.25 Ni0.25 Al0.5 catalysts, respectively were attained at
reaction temperature of 550 ◦ C and 20 s time on stream.
Mg3 Fe0.25 Mn0.25 Al0.5 gave the maximum styrene yield, and its vari-
ation with temperature is shown in Fig. 8A. Similarly, benzene yield
was equally increasing with both temperature and time with a max-
imum of approximately 4.6, 4.9, and 4.3% for Mg3 Fe0.25 Co0.25 Al0.5 ,
Mg3 Fe0.25 Mn0.25 Al0.5 , and Mg3 Fe0.25 Ni0.25 Al0.5 catalysts, respec-
tively at 550 ◦ C for reaction time of 20 s. Fig. 8B shows the variation
of benzene yield with temperature for Mg3 Fe0.25 Mn0.25 Al0.5 . It can
thus be concluded that for longer reaction time, EB undergoes both
dehydrogenation and dealkylation reactions to give styrene and
benzene, respectively. It can also be seen from Tables 2–4 that both
 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398 1393

18 25
400 C 400 C
450 C A B
450 C
16 500 C 500 C
550 C
Ethylbenzene Conversion (%)

20 550 C
14 400 C Pred

Ethylbenzene Conversion (%)

400 C Pred
450 C Pred 450 C Pred
12 500 C pred 500 C Pred
550 C Pred 550 C Pred
15
10

8
10
6

4
5
2

0 0
0 5 10 15 20 25 0 5 10 15 20 25
Reaction Time (s) Reaction Time (s)

20
400 C C
18 450 C
500 C
550 C
Ethylbenzene Conversion (%)

16 400 C Pred
450 C Pred
14 500 C Pred
550 C Pred
12

10

0
0 5 10 15 20 25
Reaction Time (s)
Fig. 4. Conversion of ethylbenzene with respect to time at various temperatures over different hydrotalcites mixed oxides: (A) Mg3 Fe0.25 Co0.25 Al0.5 , (B) Mg3 Fe0.25 Mn0.25 Al0.5 ,
and (C) Mg3 Fe0.25 Ni0.25 Al0.5 .

the yield of DEB (disproportionation product) and toluene (cracking approach is more sound than time-on-stream approach which has
product) increased with temperature. been widely used in literature to account for catalyst deactivation.
Full description of the development of this mathematical model
3.2.3. Coke content measurement is shown in Ref. [31]. In developing the kinetic model, only the
Table 5 shows the amount of coke deposited for the catalysts most important reactions taking place were considered. Toluene
investigated at 550 ◦ C. The data revealed low coke yield, and the formation was not considered owing to its very low concentra-
ratio of coke weight percent to percent conversion is small. Simi- tion in the reaction product. It can thus be assumed that due to
larly, coking on the commercial catalyst was low. This implies that the short reaction time, the styrene produced did not undergo suc-
dehydrogenation of ethylbenzene over the catalysts is not accom- cessive cracking to toluene. Hence, reaction in Scheme 1 has been
panied by substantial coke deposition. It can thus be inferred that adapted, neglecting toluene formation, in determining the kinetic
loss of activity of this commercial catalyst in the absence of steam parameters.
was not due to coking but as a result of solid state transformation, The following set of species balances and catalytic reactions can
though the exact deactivation mechanism was not studied. be written:

3.3. Kinetic modeling Rate of EB disappearance, rEB

3.3.1. Model formulation V dCEB

− = (k1 CEB − k−1 CST CH + k2 CEB )ϕ (1)
Generally, the accepted catalytic reaction scheme for EB dehy- Wc dt
drogenation is shown in Scheme 1. The experimental results were
modeled using catalyst deactivation function model based on Rate of styrene appearance, rST
reactant conversion developed by Al-Khattaf and de Lasa [31].
According to this model, the deactivation is based on the conversion V dCST
= (k1 CEB − k−1 CST CH )ϕ (2)
of ethylbenzene which is directly related to coke formation. This Wc dt
1394 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398

Table 5
Coke formation and properties for ethylbenzene dehydrogenation reaction at different reaction conditions.

Catalysts Temp. (◦ C) Reaction time (s) Conversion (%) Coke wt (% C) Coke wt/conversion

Co 550 10 11.38 0.503 0.044

20 15.46 0.488 0.032

Mn 550 10 15.02 0.618 0.041

20 19.70 0.694 0.035

Ni 550 10 13.27 0.582 0.044

20 18.04 0.452 0.025

Comm. 550 10 7.87 0.382 0.049

20 11.15 0.446 0.040

20 100
400 C
18 comm ercial catalyst 90 450 C
Co
Mn 500 C
16 80 styrene
Ni 550 C

Selectivity (%)
14 70
EB conversion (%)

12 60

10 50

8 40

6 30

4 20

2 10 benzene

0 0
0 5 10 15 20 25 0 5 10 15 20 25
Reaction Time (sec) Ethy lbenzene Conversion (%)
◦
Fig. 5. Conversion of ethylbenzene with respect to time at 500 C. Comparison Fig. 7. Selectivity with respect to conversion and temperature for
between commercial and HT catalysts. Mg3 Fe0.25 Mn0.25 Al0.5 .

Rate of formation of benzene, rB

V dCB
= k2 CEB ϕ (3)
Wc dt

where Ci is molar concentration of each of the species in the riser

simulator, V is the volume of the riser (45 cm3 ), Wc is the mass of the
catalyst (0.81 g cat), t is time (s), k is the rate constant (cm3 /(g cat s)),
and ϕ, the catalyst decay function, which accounts for the loss of
catalytic activity as a result of deactivation due to coking. For the
reactant conversion model, the catalyst deactivation function is
given as:

ϕ = exp(−(1 − yEB )) (4)

where  is catalyst deactivation constant.

By definition, the molar concentration Ci of each species can
be written in terms of its mass fraction yi , which are the measur-
able variables from our chromatographic analysis, by the following
relation:
yi Whc
Ci = (5)
MWi V

where Whc is the weight of feedstock injected into the reactor,

Fig. 6. Product selectivity of ethylbenzene dehydrogenation over the different cat- MWi is the molecular weights of the individual species and V is
alysts at ∼6% ethylbenzene conversion.
the volume of the riser simulator.
 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398 1395

14 6
400 C A 400 C B
450 C
12 450 C
500 C 5 500 C
550 C
550 C
10 400 C Pred
400 C Pred
450 C Pred 4
Styrene yield (%)

450 C Pred

Benzene yield (%)

500 C Pred
8 500 C Pred
550 C Pred
550 C Pred
3
6

2
4

1
2

0 0
0 5 10 15 20 25 0 5 10 15 20 25

Ethylbenzene Conversion (%) Ethylbenzene Conversion (%)

Fig. 8. Styrene (A) and benzene (B) yield vs. ethylbenzene conversion at various temperatures Mg3 Fe0.25 Mn0.25 Al0.5 .

To ensure thermodynamic consistency at equilibrium, the dyB Wc

= G4 k2 yEB exp(−(1 − yEB )) (10)
temperature dependent equilibrium constant for the reversible dt V
reaction was found to be:
G1 , G2 , G3 , and G4 are lumped constants given as:
k1
KC = (6) MWEB Whc
k−1 G1 =
MWST MWH V
which has been calculated from literature as [32]:
   MWST
122700 − 126.3T − 2.194 × 10−3 T 2 G2 =
KC = exp − × 105 Pa MWEB
RT
(7)
Whc
G3 =
MWH V
Substituting Eqs. (4)–(6) into Eqs. (1)–(3) and simplifying gives
the following:
MWB
dyEB yST yH
   W G4 =
c MWEB
= − k1 yEB − G1 + k2 yEB exp(−(1 − yEB ))
dt KC V
The activation energy, Ei , for the dehydrogenation reaction is
(8)
related to temperature dependent rate constants ki according to
the Arrhenius equation given below:
dyST yST yH
 W  E 
c i
= k1 G2 yEB − G3 exp(−(1 − yEB )) (9) ki = Ai exp − (11)
dt KC V RT

Table 6
Estimated kinetic parameters for hydrotalcite mixed oxides catalyst based on reactant conversion model.

Parameters Values

k1 k2 

Mg3 Fe0.25 Co0.25 Al0.5

Ei-Co (kJ/mol) 76.37 93.07 30.86
95% CL 7.67 9.99 3.52
k0i a [×103 m3 /(kg of catalyst s)] 1.38 0.64
95% CL 0.30 0.15
Mg3 Fe0.25 Mn0.25 Al0.5
Ei-Mn (kJ/mol) 65.2 111.8 17.18
95% CL 6.41 13.87 2.10
k0i a [×103 m3 /(kg of catalyst s)] 1.24 0.32
95% CL 0.20 0.07
Mg3 Fe0.25 Ni0.25 Al0.5
Ei-Ni (kJ/mol) 87.46 94.74 16.51
95% CL 6.24 9.63 1.9
k0i a [×103 m3 /(kg of catalyst s)] 0.52 0.23
95% CL 0.06 0.03
a
Pre-exponential factor as obtained from Eq. (12); unit for second order (m6 /(kg of catalyst) s).
1396 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398

Table 7
Correlation matrix for ethylbenzene dehydrogenation over hydrotalcite mixed oxides catalyst.

k1 E1 k2 E2 

Mg3 Fe0.25 Co0.25 Al0.5

k1-Co 1.0000 0.8342 0.9290 0.6689 0.9788
E1-Co 0.8342 1.0000 0.8147 0.6937 0.8645
k2-Co 0.9290 0.8147 1.0000 0.4810 0.9251
E2-Co 0.6689 0.6937 0.4810 1.0000 0.6714
 0.9788 0.8645 0.9251 0.6714 1.0000
Mg3 Fe0.25 Mn0.25 Al0.5
k1-Mn 1.0000 0.7903 0.6898 0.4180 0.9696
E1-Mn 0.7903 1.0000 0.6074 0.4035 0.8374
k2-Mn 0.6898 0.6074 1.0000 −0.2421 0.6962
E2-Mn 0.4180 0.4035 −0.2421 1.0000 0.4154
 0.9696 0.8374 0.6962 0.4154 1.0000
Mg3 Fe0.25 Ni0.25 Al0.5
k1-Ni 1.0000 0.6215 0.6349 0.5320 0.9260
E1-Ni 0.6215 1.0000 0.6500 0.3865 0.8041
k2-Ni 0.6349 0.6500 1.0000 −0.0659 0.7173
E2-Ni 0.5320 0.3865 −0.0659 1.0000 0.5097
 0.9260 0.8041 0.7173 0.5097 1.0000

where Ai is called the pre-exponential factor. Agarwal and Brisk
have suggested re-parameterization of equation above during
kinetic modeling to help reduce parameter interaction. Therefore,
ki constants were re-parameterized by centering the values around
k0i which is the value of the rate constant for reaction i at the
average temperature T0 of the investigated temperatures [33]:
 E 1 1
 can be assumed that the extensive bed mixing and heat distribu-
i
ki = k0i exp − − (12)
R T T0
Since the experimental runs were done at 400, 450, 500 and
550 ◦ C, To was calculated to be 475 ◦ C.
The above reaction model equations were derived based on the
following assumptions:
1. Catalyst deactivation was assumed to be a function of reactant
conversion. And, a single deactivation function was defined for
all the reactions.
2. The model assumes only catalytic reactions and neglects thermal
conversion.
3. The reactor operates under isothermal conditions, justified by
the negligible temperature change observed during the reaction.
4. Toluene yield is negligible.
5. The model also neglects the disproportionation reaction.
promoted Fe2 O3 . Dehydrogenation of EB over Mg3 Fe0.25 Co0.25 Al0.5
to styrene was conducted using a continuous gas-flow fixed bed
reactor at atmospheric pressure [13]. The reaction was carried
out for 3 h of time-on-stream at 550 ◦ C with the catalyst after
been pre-treated with a He gas. Activation energy of 29.1 kcal/mol
was estimated for Mg3 Fe0.25 Co0.25 Al0.5 using the Arrhenius plot. It
tion within the fluidized bed enhanced reactant–catalyst contact
and uniform product quality, hence reduction of activation energy
towards styrene formation. The active specie is still identified to be
the metastable Fe3+ .
The main difference between the three catalysts presented is the
nature of the divalent cation involved in the matrix composition.
Mn2+ has been reported to possess a higher ionic radius than the
other Me2+ which may lead to incomplete incorporation of Mn2+
into the HT structure [13]. From the estimated activation ener-
gies, the addition of Mn2+ best improved the catalyst activity even
though both Co2+ and Ni2+ are well incorporated into the HT struc-
ture. It can thus be suggested that the segregation of Mn2+ serves as
20
Ethylbenzene Conversion (%)
18
Styrene yield (%)

16 Benzene yield (%)

3.3.2. Determination of model parameters
The kinetic parameters k0i , Ei and  for the reactions were 14
determined by fitting experimental results into the rate Eqs.
(8)–(10) using non-linear regression analysis with the aid of MAT- 12
LAB package. The values of the model parameters along with
Model

their corresponding 95% confidence limits (CLs) are shown in 10

Table 6 while the resulting cross-correlation matrices are also
given in Table 7 for Mg3 Fe0.25 Co0.25 Al0.5 , Mg3 Fe0.25 Mn0.25 Al0.5 and 8
Mg3 Fe0.25 Ni0.25 Al0.5 catalysts respectively. Table 7 shows the corre-
lations between k1 and E1 , E1 and  and k1 and  for all the estimated 6
parameters of dehydrogenation reaction to produce styrene. Sim-
ilarly, correlations between k2 and E2 , E2 and  and k2 and  are 4
shown for parameters of cracking reaction. It can be observed that
the cross-correlation matrices presented in this study revealed 2
low and moderate level of parameter interaction with only a few
exceptions. From the results of the kinetic parameters, the esti- 0
mated apparent energies of 18.3, 15.6, and 20.9 kcal/mol for Co, 0 5 10 15 20
Mn, Ni substituted catalysts, respectively are found to be lower
Experimental
compared to those reported in most literatures, however, the result
obtained for Ni is essentially the same as that reported by Addiego Fig. 9. Overall comparison between the experimental results and model predictions
et al. [18], an apparent activation energy of 21 kcal/mol for 30% K- of all HTs catalysts for Scheme 1.
 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398
additional active sites which increased the basicity of the catalyst,
responsible for initiating dehydrogenation by H+ abstraction [14].
The activation energies estimated for dealkylation reactions
which are 22.3, 26.7, and 22.7 kcal/mol for Mg3 Fe0.25 Co0.25 Al0.5 ,
Mg3 Fe0.25 Mn0.25 Al0.5 , and Mg3 Fe0.25 Ni0.25 Al0.5 catalysts respec-
tively, are found to be higher than that for dehydrogenation
reaction. Similar trend has been reported by Hirano [4,34] both for
unpromoted and promoted iron oxide catalysts. This indicates that
high activation energy must be overcome for the cracking reaction
to proceed, which was found to be highest in Mg3 Fe0.25 Mn0.25 Al0.5 .
This depicts that Mn2+ gave the best selectivity towards styrene
formation which is consistent with the experimental data.
The estimated kinetic parameters for the fitted parameters sub-
stituted into the developed model for the reaction scheme and
equations were validated through numerical solution using fourth-
order-Runge–Kutta routine. The numerical results were compared
with experimental data as shown in Fig. 4(A–C). The figure shows
the model predictions have a very good match with the experi-
mental values. Further comparisons between model predictions
and experimental data are presented in Fig. 8(A and B). This
demonstrates that the proposed kinetic model fits well into our
experimental observations. The reconciliation plot (Fig. 9) shows
the overall agreement of experimental data and model predictions.
The two agrees very well as suggested by the value of regression
coefficient, r2 (0.99).
4. Conclusion
The kinetics of ethylbenzene has been carried out over
Mg3 Fe0.25 Me0.25 Al0.5 (Me = Co, Mn and Ni) quaternary mixed oxide
catalysts prepared from hydrotalcites as precursors. The following
conclusions were drawn from the study:
(1) The EB conversion increases significantly with the increase of
reaction temperature and time, but the selectivity to styrene
decreased.
(2) The amount of benzene and toluene in the reaction product
increased as the temperature was raised, showing that cracking
reaction is favored at higher temperature. However, the yield
of toluene was not appreciable and hence was neglected.
(3) Incorporation of Mn2+ was the most effective giving superior
activity and selectivity towards styrene. A maximum EB con-
version of 19.7% at 550 ◦ C, and styrene selectivity of 92.7% at
400 ◦ C were attained.
(4) Kinetic parameters for the ethylbenzene dehydrogenation
showed that apparent energies for styrene formation were
found to decrease as follows: E1-Ni > E1-Co > E1-Mn . Also, activa-
tion energies for the cracking reaction were found to decrease
as follows: E2-Mn > E2-Ni > E2-Co .
(5) Activation energy for cracking was found to be higher than
dehydrogenation to styrene for all the HT catalyst, i.e.,
E2-Me > E1-Me (Me = Co, Mn and Ni). This trend agrees with lit-
erature.
Acknowledgments
This publication was based on work supported in part by Award
No. K-C1-019-12 made by King Abdullah University of Science
and Technology (KAUST). In addition, the authors express their
appreciation for the financial support from the Centre for Research
Excellence in Petroleum Refining and Petrochemicals at King Fahd
University of Petroleum & Minerals. The authors also acknowl-
edge the contribution of Prof. Emeritus K. Takehira (Hiroshima
University, Higashi-Hiroshima, Japan). Mr. Mariano Gica is also
acknowledged for his help during the experimental work.
1397
References
[1] S.J. Liao, T. Chen, C.X. Miao, W.M. Yang, Z.K. Xie, Q.L. Chen, Effect of TiO2 on the
structure and catalytic behavior of iron-potassium oxide catalyst for dehydro-
genation of ethylbenzene to styrene, Catal. Commun. 9 (2008) 1817–1821.
[2] F. Cavani, F. Trifiò, Alternative processes for the production of styrene, Appl.
Catal. A 133 (1995) 219–239.
[3] N. Mimura, I. Takahara, M. Saito, T. Hattori, K. Ohkuma, M. Ando, Dehydrogena-
tion of ethylbenzene over iron oxide-based catalyst in the presence of carbon
dioxide, Catal. Today 45 (1998) 61–64.
[4] T. Hirano, Roles of potassium in potassium-promoted iron-oxide catalyst for
dehydrogenation of ethylbenzene, Appl. Catal. 26 (1986) 65–79.
[5] K. Shibata, T. Kiyoura, Effect of potassium promoter on iron oxide catalysts for
dehydrogenation of ethylbenzene to styrene, Bull. Chem. Soc. Jpn. 42 (1969)
871–874.
[6] Y. Ohishi, T. Kawabata, T. Shishido, K. Takaki, Q. Zhang, Y. Wang, K. Nomura, K.
Takehira, Mg–Fe–Al mixed oxides with mesoporous properties prepared from
hydrotalcite as precursors: catalytic behavior in ethylbenzene dehydrogena-
tion, Appl. Catal. A 288 (2005) 220–231.
[7] B.D. Herzog, H.F. Rase, In situ catalyst reactivation: used ethylbenzene dehy-
drogenation catalyst with agglomerated potassium promoter, Ind. Eng. Chem.
Prod. Res. Dev. 23 (1984) 187–196.
[8] J.C.A.A. Roelofs, D.J. Lensveld, A.J. van Dillen, K.P. de Jong, On the structure of
activated hydrotalcites as solid base catalysts for liquid-phase aldol condensa-
tion, J. Catal. 203 (2001) 184–191.
[9] P. Kuśtrowski, D. Sułkowska, L. Chmielarz, A. Rafalska-Lasocha, B. Dudek,
R. Dziembaj, Influence of thermal treatment conditions on the activity of
hydrotalcite-derived Mg–Al oxides in the aldol condensation of acetone, Micro-
por. Mesopor. Mater. 78 (2005) 11–22.
[10] F. Cavani, F. Trifiro, A. Vaccari, Hydrotalcite-type anionic clays: preparation,
properties and applications, Catal. Today 11 (1991) 173–301.
[11] Y. Liu, E. Lotero, J.G. Goodwin, X. Mo, Transesterification of poultry fat with
methanol using Mg–Al hydrotalcite derived catalysts, Appl. Catal. A 331 (2007)
138–148.
[12] D.P. Debecker, E.M. Gaigneaux, G. Busca, Exploring, tuning, and exploiting the
basicity of hydrotalcites for applications in heterogeneous catalysis, Chem. Eur.
J. 15 (2009) 3920–3935.
[13] R.J. Balasamy, A. Khurshid, A.A.S. Al-Ali, L.A. Atanda, K. Sagata, H. Yahiro, K.
Nomura, T. Sano, K. Takehira, S.S. Al-Khattaf, Ethylbenzene dehydrogenation
over binary FeOx –MeOy /Mg(Al)O catalysts derived from hydrotalcites, Appl.
Catal. A 390 (2010) 225–234.
[14] R.J. Balasamy, A. Khurshid, A.A.S. Al-Ali, L.A. Atanda, K. Sagata, H. Yahiro, K.
Nomura, T. Sano, K. Takehira, S.S. Al-Khattaf, Ethylbenzene dehydrogenation
over FeOx /(Mg,Zn)(Al)O catalysts derived from hydrotalcites: role of MgO as
the catalyst support, Appl. Catal. A 398 (2011) 113–122.
[15] S. Carrà, L. Forni, Kinetics of catalytic dehydrogenation of ethylbenzene to
styrene, Ind. Eng. Chem. Process Des. Dev. 4 (1965) 281–285.
[16] K. Coulter, D.W. Goodman, R.G. Moore, Kinetics of the dehydrogenation of
ethylbenzene to styrene over unpromoted and K promoted model iron-oxide
catalysts, Catal. Lett. 31 (1995) 1–8.
[17] O. Shekhah, W. Ranke, R. Schlögl, Styrene synthesis: in situ characterization and
reactivity studies of unpromoted and potassium-promoted iron oxide model
catalysts, J. Catal. 225 (2004) 56–68.
[18] W.P. Addiego, C.A. Estrada, D.W. Goodman, M.P. Rosynek, R.G. Windham, An
infrared study of the dehydrogenation of ethylbenzene to styrene over iron-
based catalysts, J. Catal. 146 (1994) 407–414.
[19] A. Miyakoshi, A. Ueno, M. Ichikawa, Mn-substituted Fe–K mixed oxide catalysts
for dehydrogenation of ethylbenzene towards styrene, Appl. Catal. A 216 (2001)
137–146.
[20] W.J. Lee, G.F. Froment, Ethylbenzene dehydrogenation to styrene: kinetic mod-
eling and reactor simulation, Ind. Eng. Chem. Res. 47 (2008) 9183–9194.
[21] B.K. Abdalla, S.S.E.H. Elnashaie, Fluidized bed reactors without and with selec-
tive membranes for the catalytic dehydrogenation of ethylbenzene to styrene,
J. Membr. Sci. 101 (1995) 31–42.
[22] S. Miyata, A. Okada, Synthesis of hydrotalcite-like compounds and their physic-
ochemical properties—the systems Mg2+ –Al3+ –SO4 2− and Mg2+ –Al3+ –CrO4 2− ,
Clays Clay Miner. 25 (1977) 14–18.
[23] H.I. de Lasa, Riser simulator for catalytic cracking studies. U.S. Patent 5,102,628
(1992).
[24] D.W. Kraemer, H.I. de Lasa, Catalytic cracking of hydrocarbons in a riser simu-
lator, Ind. Eng. Chem. Res. 27 (1988) 2002–2008.
[25] R.D. Shannon, Revised effective ionic radii and systematic studies of interatomic
distances in halides and chalcogenides, Acta Crystallogr. A 32 (1976) 751–767.
[26] X. Ge, M. Li, J. Shen, The reduction of Mg–Fe–O and Mg–Fe–Al–O complex
oxides studied by temperature-programmed reduction combined with in situ
Mössbauer spectroscopy, J. Solid State Chem. 161 (2001) 38–44.
[27] D. Li, I. Atake, T. Shishido, Y. Oumi, T. Sano, K. Takehira, Self-regenerative activity
of Ni/Mg(Al)O catalysts with trace Ru during daily start-up and shut-down
operation of CH4 steam reforming, J. Catal. 250 (2007) 299–312.
[28] A. Chen, H. Xu, Y. Yue, W. Hua, W. Shen, Z. Gao, Hydrogenation of methyl
benzoate to benzaldehyde over manganese oxide catalysts prepared from
Mg/Mn/Al hydrotalcite-like compounds, Appl. Catal. A 274 (2004) 101–109.
[29] T. Hirano, Dehydrogenation of ethylbenzene on potassium-promoted iron-
oxide catalysts containing various transitional-metal oxides, Bull. Chem. Soc.
Jpn. 59 (1986) 1653–1655.
1398 L. Atanda et al. / Chemical Engineering Journal 171 (2011) 1387–1398

[30] E.H. Lee, Iron oxide catalysts for dehydrogenation of ethylbenzene in the pres-
ence of steam, Catal. Rev. Eng. Sci. 8 (1973) 285–305.
[31] S. Al-Khattaf, H. de Lasa, Catalytic cracking of cumene in a riser simulator: a
catalyst activity decay model, Ind. Eng. Chem. Res. 40 (2001) 5398–5404.
[32] R. Dittmeyer, V. Höllein, P. Quicker, G. Emig, G. Hausinger, F. Schmidt,
Factors controlling the performance of catalytic dehydrogenation of ethylben-
zene in palladium composite membrane reactors, Chem. Eng. Sci. 54 (1999)
1431–1439.
[33] A.K. Agarwal, M.L. Brisk, Sequential experimental design for precise parameter
estimation. 1. Use of reparameterization, Ind. Eng. Chem. Process Des. Dev. 24
(1985) 203–207.
[34] T. Hirano, Dehydrogenation of ethylbenzene over potassium-promoted iron
oxide containing cerium and molybdenum oxides, Appl. Catal. 28 (1986)
119–132.