Applied Catalysis A: General 366 (2009) 193–200

Contents lists available at ScienceDirect

Applied Catalysis A: General

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

Dehydrogenation of ethylbenzene in the presence of CO2 using a catalyst

synthesized by polymeric precursor method
Tiago Pinheiro Braga, Antônio Narcı́sio Pinheiro, Camila Vieira Teixeira, Antoninho Valentini *
Departamento de Quı´mica Analı´tica e Fı´sico-Quı´mica, Universidade Federal do Ceará, Campus do Pici, CEP 60455-970, Fortaleza, CE, Brazil

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

Article history: Catalysts of iron oxide dispersed on Al or Si oxides were prepared via a polymeric precursor derived from
Received 29 April 2009 the Pechini method and tested in the dehydrogenation of ethylbenzene in the presence of CO2, in order to
Received in revised form 3 July 2009 contribute with the studies of this reaction. The catalysts were characterized by thermogravimetric
Accepted 5 July 2009
analysis (TG), temperature-programmed reduction (TPR), X-ray diffraction (DRX) and temperature-
Available online 10 July 2009
programmed desorption of CO2 (TPD-CO2). Analysis of the spent catalysts by TG and Fourier transformed
infrared spectroscopy (FT-IR) pointed to the contribution of CO2 to the coke deposition. The catalytic
Keywords:
results suggest that the high initial ethylbenzene conversion is due to the contribution of basic sites, and
Carbon dioxide
Dehydrogenation
the CO2 adsorption in the basic site (lattice oxygen) may compete with the oxidative dehydrogenation of
Ethylbenzene ethylbenzene. Although CO2 provides the appropriate conditions to lower the consumption of the basic
Styrene site, it is not able to promote the Fe2+ oxidation or to regenerate the basic site (lattice oxygen) in the iron
oxide dispersed on Al or Si oxide catalysts.
ß 2009 Elsevier B.V. All rights reserved.

1. Introduction Several catalysts have been tested in ethylbenzene dehydro-

The worldwide capacity for the production of the styrene
monomer is more than 20 Mt/year [1]. The high production of
styrene is justified by the many different polymeric materials that
are synthesized from this monomer [2], making it an important
industrial chemical product.
The catalytic process for the dehydrogenation of ethylbenzene
in the presence of steam is the main commercial route for styrene
production [3]. This method requires a great deal of energy due to
the excess of steam used, which is estimated to consume 10% of
production cost. Many research groups have been devoted to
synthesizing a catalyst with better stability than those actually
used [4–6], which may contribute to a decrease in process costs.
Additionally, in order to introduce an alternative route to the steam
process, researchers have dedicated their time to study the use of
carbon dioxide (CO2) as gas co-feeder in the ethylbenzene
dehydrogenation reaction [7–9]. The results of these studies
pointed to the possibility of saving energy. While the energy
required for the styrene production by a steam process is estimated
to be 1.5  109 cal/t-styrene, the process using CO2 is estimated to
be 1.9  108 cal/t-styrene [10]. In addition, this process also
presents the possibility of using a major greenhouse gas.
genation in the presence of CO2 [7,11–15]. Catalysts based on iron
oxide, supported on aluminum oxide, present good activity in the
dehydrogenation reaction [6,9,16]. Furthermore, the catalysts of
iron oxide supported on silicon oxides are also active and selective
in the above reaction [17].
On the other hand, further research is necessary to improve
catalytic stability and consequently decrease the operational costs.
One strategy to improve the catalytic performance is to add a
dopant [18] to increase CO2 adsorption by the catalyst, thus
achieving higher catalytic stability.
The method used to synthesize the catalysts directly affects its
activity and stability. The method of the polymeric precursor has
shown good results in the catalytic process [19,20]. Nevertheless,
no results have been published in the ethylbenzene dehydrogena-
tion reaction with catalysts synthesized by this route. Therefore,
with the aim of exploring this route of synthesis, we present the
results observed in the ethylbenzene dehydrogenation reaction for
the catalysts synthesized by the polymeric precursor method.
2. Experimental
2.1. Sample preparation

The catalysts were synthesized by the polymeric precursor

* Corresponding author. Tel.: +55 85 3366 9951; fax: +55 85 3366 9982. method [19,20]. This method is based on the chelation of cations
E-mail address: valent@ufc.br (A. Valentini). (metals) by citric acid in a water solution. Aluminum nitrate

0926-860X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2009.07.004
194 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200
nonahydrate {Al(NO3)39H2O}, iron nitrate nonahydrate
{Fe(NO3)39H2O}, tetraethylorthosilicate (TEOS), citric acid mono-
hydrate (CA) {C6H8O7H2O}, and ethylene glycol (EG) {C2H6O2}
were used as starting chemicals. A CA/metal ratio of 1.5:1 (mol)
was used for all the samples. The metal amount is the sum of Fe, Al
and/or Si. The mass ratio of CA/EG was kept at 2:3.
For the synthesis process of the FeAl-14 sample, where 14 is the
Al/Fe molar ratio, 0.070 and 0.005 mol of Al and Fe salts,
respectively, were dissolved in distilled water at room tempera-
ture. CA (0.1125 mol) was dissolved in ethanol at 50 8C. Later the
aqueous solution was added to the CA–ethanol solution and stirred
for 15 min at 50 8C. Subsequently, the EG was added, and the
mixture was stirred for 3 h at 100 8C until it became a viscous resin.
The resin was treated at 200 8C for 1 h under air. The resulting
precursor composite was ground and treated at 500 8C under air
flow for 1 h.
Similar conditions were used to prepare the other samples,
FeAl-4, FeSi-3 and FeSi-12.
2.2. Catalyst characterization
The pyrolysis step was followed by thermogravimetric analysis
(TG), using a 10 8C/min heating rate, under an air flow of 50 ml/min
and 10 mg of sample. The crystal structure of the metal oxides was
characterized by X-ray diffraction analysis (XRD), with CuKa
irradiation source (l = 1.540 Å, 30 kV and 20 mA).
Specific surface area (BET) and pore volume of the catalysts
were determined by N2 adsorption/desorption isotherms, at liquid
nitrogen temperature. The catalyst sample (40–50 mg) was
evacuated at 200 8C for 2 h prior to the experiments.
Temperature-programmed reduction (TPR) analyses of the
catalysts were carried out in the range of 100–950 8C in a quartz
reactor, by passing a 5% H2/N2 flow (30 ml/min), at a heating
rate of 10 8C/min. The H2 consumption was monitored by a
thermal conductivity detector (TCD) connected to the reactor
outlet.
The CO2 temperature-programmed desorption was carried
out in the range of 40–400 8C under a helium flow (10 8C/min,
30 ml/min). The catalysts (200 mg) were preheated under
helium flow (30 ml/min) at 550 8C for 1 h, after which the
temperature was decreased to 80 8C, and a flow of pure CO2
(30 ml/min) was subsequently introduced into the reactor for
0.5 h. The desorbed CO2 was detected by an online thermal
conductivity detector (TCD) after passing by a trap (20 8C) to
remove any trace water.
Fig. 1. Thermal analysis (TG) profile and the respective DTG of the precursors of the catalysts.
2.3. Catalytic activity
Catalytic tests were carried out in a fixed bed stainless steel
microreactor, using 150 mg of sample. The experiments were
carried out at 550 8C at atmospheric pressure. The catalyst was
preheated under N2 flow (30 ml/min) up to the reaction
temperature, and then the reaction mixture (CO2 and ethylben-
zene) was introduced at a flow rate of 30 ml/min. The gas phase
composition of the feed was ethylbenzene (1.09 mmol/h), using a
CO2 to ethylbenzene molar ratio of 30. The catalytic ethylbenzene
conversion was analyzed by gas chromatography, using an
instrument equipped with a FID and a nonpolar capillary column.
3. Results and discussion
3.1. Thermal analysis
The thermal analysis result, presented in Fig. 1, indicates three
main events. The weight elimination in the range of 40–200 8C,
with a maximum of around 100 8C, is related to the volatile
compounds (water) adsorbed in the surface of the catalyst sample.
In the second temperature range (between 200 and 500 8C), the
decomposition of the residual organic material is favored [19,20]
and shows two peaks in the DTG profile. The event observed near
290 8C is probably due to the incomplete organic precursor
decomposition during the sample preparation (heat treated at
200 8C). Such temperature of heat treating may result in a sample
which presents high amounts of oxygen in the composition of the
residual material.
The peak in the DTG profile observed near 400 8C is probably
related to the residual organic substance decomposition, but with
low amount of oxygen in the composition [21].
It is interesting to point out, in Fig. 1, the percentage of residual
weight loss for the samples. The increase in iron ratio in the
catalyst composition promotes a lowering in the residual weight
elimination. A similar behavior is observed due to the change of the
support precursor (Al or Si). The iron likely works as a catalyst in
the precursor elimination during the heat treatment at 200 8C.
Such a statement suggests that the difference of the metal ratio
(Fe), as well as the nature of the component used as support (Al or
Si), can affect the final morphology of the samples after thermal
treatment (surface area, diameter and volume of pores), as
indicated in a previous work [21]. The use of organic compounds
(CA and EG) in the preparation of the catalysts is responsible for the
morphological behavior displayed by the material. The removal of
 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200
Fig. 2. X-ray diffraction profile of catalysts after calcination process under air flow.
volatile materials, eliminated during the pyrolysis step, results in
cavities and simultaneous rearrangement of the solid.
3.2. X-ray diffraction
After the calcination process was carried out at 500 8C under air
flow, as suggested by the TG analysis, the samples were submitted
to the X-ray diffraction analysis (DRX). The results are presented in
Fig. 2.
The FeAl-14 catalyst presented the profile of an amorphous
solid, whereas the FeAl-4 sample showed broader peaks, a fact that
can be assigned to a small crystal size or a poorly crystalline
material. The 2u degree value of the broader peaks and the relative
intensities suggests Al2O3 phase formation for the FeAl-4 sample.
On the other hand, the FeSi-12 and FeSi-3 catalysts showed the
pattern of hematite (a-Fe2O3, JCPDS 86-2368). Additionally, the
FeSi-3 sample presented the maghemite phase (g-Fe2O3, JCPDS 39-
1346).
Fig. 3. N2 adsorption/desorption isotherms (a) and pore diameter distribution (b).
195
Table 1
Relative weight percentage of phases, determined through Rietveld refinement
method.
Samples Crystalline phase (wt %)
Al2O3 a-Fe2O3 g-Fe2O3
FeAl-14 – – –
FeAl-4 100 – –
FeSi-12 100 –
FeSi-3 11.5 88.5
This apparent aluminum influence in the hematite crystal-
lization can be assigned to the ionic radii. The radii of Al3+ (0.53 Å)
species are closer to the radii of Fe3+ (0.67 Å) species, if compared
to Si4+ (0.39 Å). The similar ionic radii enable the insertion of Al3+
into the lattice in place of Fe3+, which makes the hematite
crystallization difficult. Nevertheless, if the hematite phase is
formed in the FeAl-14 and FeAl-4 samples, its crystal size is very
small [22].
In spite of the low crystalline organization presented by some
catalysts, the Rietveld method [23] was applied for the refinement
of the diffraction profile. By this method, it was possible to quantify
the different phase formed, and the results are displayed in Table 1.
The results observed by the DRX analysis point to a high
dispersion of the iron oxide in the FeAl-14, FeAl-4 and FeSi-12
samples; additionally, the maghemite phase in the FeSi-3 sample
points to a promising catalytic performance in the ethylbenzene
dehydrogenation reaction. This indication comes from the fact that
these materials may present a high concentration of atomic surface
defects, which are also responsible for catalytic activity [24,25].
3.3. Specific surface area measurements
By applying nitrogen adsorption/desorption isotherms, which
are presented in Fig. 3, the pore diameter and volume as well as the
surface area of the synthesized materials were determined. All the
samples showed isotherm profiles characteristic of mesoporous
materials (type IV according to IUPAC classification), which are
interesting for the catalysts processes.
196 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200
Table 2
Textural properties of the catalysts determined by N2 adsorption isotherms.
Samples S (m2/g) Vp (cm3/g)
FeAl-14 145 0.23
FeAl-4 166 0.18
FeSi-12 323 0.80
FeSi-3 217 0.39
The profiles of pore diameter distribution presented in Fig. 3b
(according to the BJH method), exhibit a broad pore size
distribution, with the majority of the porous diameter located in
the range of 30–150 Å.
An increase in the amount of iron in the sample composition
promoted changes in the hysteresis behavior, which can be
attributed to the closing of the pores by iron oxide. Table 2 lists the
values of the surface area, total pore volume and the average pore
diameter of the samples, which points to the textural effect,
promoted by an increase in the amount of iron.
Such textural effects in the porosity of the synthesized solids
may be related to the information evidenced in the TG analysis
(Fig. 1). Samples with higher iron amount were seen to
demonstrate a lower ratio of residual organic substances. After
the calcination process, according to the isotherm curves, this can
result in differences in the characteristics of the pores [21], which
are perceptible in the values of the surface area, volume and
diameter.
In spite of the decrease in the textural property values due to an
increase in the amount of iron, the samples possess a reasonable
surface area, which can be positive for the catalytic process. A
larger surface area can result in a higher capacity of carbon dioxide
adsorption. The improvement in the CO2 adsorption capacity may
be interesting for the reaction proposed, since CO2 is one of the
reactants.
3.4. Temperature-programmed reduction
The temperature-programmed reduction (TPR) experiments for
the FeSi-12, FeSi-3 and FeAl-4 samples produced a well-defined H2
consumption peak with a maximum near 420 8C (Fig. 4).
Conversely, the TPR profile for the FeAl-14 catalyst showed a
broad peak of H2 consumption, with a maximum at 475 8C. The
peaks (a) in the range of 420–475 8C are assigned to the reduction
of the Fe3+ species due to magnetite (Fe3O4) formation [3] and/or
FeO [26]. Otherwise, the formation of a small amount of metallic
iron in the same temperature range (300–500 8C) should be
considered [26].
Fig. 4. Temperature-programmed reduction curves of the catalysts.
Additionally, the samples FeSi-12 and FeSi-3 show two H2
consumption peaks (b and g) in the temperature range of 520–
Dp (Å) 800 8C, both of them associated with the Fe2+ reduction to produce
metallic iron [3,18,26]. Nevertheless, the amount of H2 consump-
65
49 tion in the peaks b and g is lower than the H2 consumption in peak
83 a, suggesting incomplete Fe2+ reduction or that in the peak a takes
75 place mainly in the formation of FeO instead of Fe3O4 phase.
A possible explanation for the presence of two H2 consumption
peaks at 575 and 695 8C for the FeSi-3 and at 605 and 710 8C for the
FeSi-12 sample is their textural properties. Metal oxide particles
placed inside of porous materials are more difficult to reduce as
compared to metal oxides positioned at external surfaces [27].
Thus, the b peak may be due to Fe2+ reduction outside the porous
channel, while the peak g may be due to the reduction of Fe2+
inside the channel.
The FeAl-14 and FeAl-4 catalysts do not show a well-defined H2
consumption peak at temperatures higher than 520 8C. On the
other hand, the TPR base line level at higher temperatures points to
a continuous iron oxide reduction [26].
Assuming that the Fe3+ species is the active site for the
ethylbenzene dehydrogenation reaction [3,28], the TPR profiles
suggest that the aluminum addition can delay the deactivation
process during the catalytic test.
3.5. Catalytic tests
The catalytic performances of the samples are shown in Fig. 5a
(ethylbenzene conversion) and Fig. 5b (styrene selectivity). The
ethylbenzene conversion observed for the FeSi-12 and FeAl-14
samples, both with low percentages of iron, suggests that high
surface area alone is not a determinative factor in greater catalytic
conversion. In spite of the higher surface area presented by the
FeSi-12 catalyst (Table 2) relative to the FeAl-14 sample, the
ethylbenzene conversion values were very similar (Fig. 5a) for both
samples. Moreover, for samples with superior ratio of iron (FeSi-3,
FeAl-4), the ethylbenzene conversion was higher for samples with
lower surface area (Fig. 5a).
The higher ethylbenzene conversion observed for the FeAl-4
sample, relative to the FeSi-3 catalyst, is compatible with the
DRX results (Fig. 2). Since the DRX profile of the FeSi-3 pointed
to a superior diameter of iron oxide particles, the FeSi-3 sample
is expected to have a lower iron oxide surface area. In other
words, different catalysts with the same iron content do not
necessarily have the same ratio of accessible active sites for the
reaction.
Additionally, as suggested by the TPR profile, the FeSi-3 catalyst
undergoes an inherent phase change (Fe3+ reduction) under
reaction conditions (550 8C), more easily than FeAl-4 sample. This
fact can explain the superior ratio of deactivation showed by the
FeSi-3 sample. The Fe3+ reduction can promote Fe3O4 phase
formation, which exhibits only a minor catalytic activity [1,2], or
the formation of FeO, which is catalytically inactive [29,30] in the
dehydrogenation of ethylbenzene.
Thus, the decrease in the ethylbenzene conversion with time on
stream, which is distinct for the different catalysts, can be
indicative of the reaction pathway followed by the dehydrogena-
tion of ethylbenzene with CO2.
Publications of several groups [25,30,31] agree that the
elementary step of dehydrogenation of the ethyl group to an
ethylene group takes place at the oxygen sites. This process
produces hydroxyl groups and reduced iron species in the catalyst
surface, but the method of reoxidation of the iron and desorption of
the hydrogen is not clear.
The ethylbenzene dehydrogenation reaction (in the presence of
steam) was recently proposed to occur through one of two
reactions: one is the faster oxidative dehydrogenation (1), as
 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200 197

Fig. 5. Catalytic performance as a function of time for the catalysts: (a) ethylbenzene conversion, (b) styrene selectivity. Reaction conditions: T = 550 8C, W/F = 137 (gcat h/
mol), CO2/EB = 30 (mol/mol).

mentioned above, and the second is simple dehydrogenation (2) submitted to the catalytic test in the absence of CO2 (Fig. 6). The
[6]. observed profile points to the CO2 competitive adsorption for basic
sites, since the stability of the ethylbenzene conversion is reached
C8 H10 þ O2
lattice ! C8 H8 þ H2 O þ 2e

(1) within 50 min in the absence CO2, while in its presence, the initial
conversion is superior and the apparent stabilization takes place
C8 H10 ! C8 H8 þ H2 (2) after 100 min and at identical ethylbenzene conversion of the
3+ process without CO2.
Reaction (2) can also promote Fe reduction, as verified by
This result corroborates previous data, which shows that
analyzing the catalysts after the reaction process. These results
the CO2 adsorption in the basic site competes with the
suggest that the incorporation of a doping compound, or the
oxidative dehydrogenation reaction (Reaction (1)), and the
formation of a phase that makes the Fe3+ reduction process
CO2 adsorption contributes to a higher catalytic conversion;
difficult, contributes positively for the catalytic stabilization
however, the CO2 is not able to regenerate the basic site of the
[3,16,18].
catalyst surface.
Considering the Reactions (1) and (2) in the dehydrogenation of
In order to verify the correlation between the CO2 adsorption
ethylbenzene with CO2, the CO2 can take part in the process in
ability and the ethylbenzene conversion of the catalysts, experi-
different ways: one is through the reverse water–gas shift reaction
ments of temperature-programmed desorption of CO2 (TPD-CO2)
(RWGS), which produces H2O and CO (3); the second is CO2
were carried out (Fig. 7). The data in Fig. 7 are in arbitrary units, and
reduction to form a radical anion adsorbed in metallic oxide (4),
the intensity of the deflection was not quantified with regard to the
which can regenerate the basic site (O2lattice ). number of moles of CO2. Nevertheless, the experimental conditions
CO2 þ H2 ! CO þ H2 O (3) were rigorously equivalent for all samples; therefore, it was
possible to perform the comparison between the areas of the TPD
CO2 þ e ! CO (4) profiles.
2

Nevertheless, thermodynamically, Reaction (3) is more favor-

able than Reaction (4), since the radical anion formation requires a
highly negative potential [32,33]. Therefore the regeneration of the
basic site by the CO2 is unlikely.
On the other hand, the conversion decrease observed in the
Fig. 5 can be related to a competitive process between the CO2
adsorption and Reaction (1) for the basic site occupation. Initially,
all the basic sites are free, thus permitting the CO2 to be adsorbed in
such sites [34]; if the CO2 is adsorbed, its reaction with hydrogen
from ethylbenzene can be favored. Concomitantly, the oxidative
dehydrogenation reaction occurs, which promotes consumption of
the basic sites and the Fe3+ reduction; consequently, ethylbenzene
conversion decreases. In addition, due to the basic site consump-
tion, the CO2 adsorption ability by the catalyst is unfavorable; thus,
the reaction between hydrogen and CO2 would decrease.
Therefore, the basic sites can contribute significantly for the
greater initial ethylbenzene conversion observed, and the CO2
presence at this reaction stage provides the conditions to reduce
the ratio of basic site consumption. Fig. 6. Ethylbenzene conversion for the FeSi-3 catalyst, under distinct atmospheres.
In order to achieve more information about the role of CO2 in Reaction in CO2: T = 550 8C, W/F = 137 (gcat h/mol), CO2/EB = 30 (mol/mol). Reaction
the ethylbenzene dehydrogenation process, the FeSi-3 sample was in N2: T = 550 8C, W/F = 137 (gcat h/mol).
198 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200
Fig. 7. Temperature-programmed desorption of CO2 (TPD-CO2) for the catalysts. The
inserted numerical values indicate the relative area between the samples.
Just as we observed in the catalytic performance (Fig. 5), Fig. 7
points to a support and the influence of the amount of iron on the
CO2 adsorption ability.
In spite of the higher surface area and the iron percentage
presented by the FeSi-3 sample, its CO2 adsorption capacity was
lower than FeAl-14 catalyst. This result can be indicative of the
influence of the support in the properties of the catalyst [16];
however, the contribution for the adsorption of CO2 by the
aluminum oxide in the samples FeAl-14 and FeAl-4 should be
considered [31,34].
Similar behavior was observed in the styrene selectivity for the
different catalysts, as shown in Fig. 5b and emphasized on Table 3,
demonstrating that a support change (Si or Al oxide) does not
significantly affect the properties of the active site. On the other
hand, the difference observed in the initial catalytic behavior and
the adsorption of CO2 profile point to the support influence. Thus, if
one takes into account only the reaction time lower than 2 h, a
higher difference in the catalytic selectivity is observed, suggesting
the support influence in the catalytic properties, as pointed out by
the ethylbenzene conversion (Fig. 5) and the TPD-CO2 (Fig. 7).
Recently [35], the catalytic performance of materials containing
aluminum and iron oxide in a similar reaction conditions for the
ethylbenzene dehydrogenation was reported. The results of
catalytic activity (Table 3) and stability, reported in the present
paper, show a similar behavior for the samples with a comparable
composition [35]. On the other hand, the styrene selectivity
observed for the samples synthesized by the polymeric precursor
method and reported here is higher.
In order to obtain information concerning the reasons of the
decrease in the ethylbenzene conversion, mainly for the samples
Fig. 8. Thermogravimetric analysis profile of the samples after 5 h of catalytic test.
Table 3
Catalytic selectivity and activity in the ethylbenzene dehydrogenation reactiona.
Samples Selectivityb (%) Activityc (mmol/h gcat)
Benzene Toluene Styrene
FeAl-14 4 3 92 0.575 (0.818)
FeAl-4 4 1 94 1.783 (1.834)
FeSi-12 4 - 98 0.546
FeSi-3 4 - 88 0.997
a
Reaction conditions: T = 550 8C, W/F = 137 (gcat h/mol), CO2/EB = 30 (mol/mol).
b
Average values only for reaction period higher than 180 min.
c
The values in the parenthesis are from reference [35].
FeAl-4 and FeSi-3, the TG and DRX analyses were carried out for all
the catalysts, after the catalytic test.
The thermal analysis results (Fig. 8) point to a lower coke
deposition in the catalysts containing SiO2, even for the FeSi-3
sample, which presented a high ethylbenzene conversion and a
superior iron amount. These results suggest that coke deposition is
a function of the iron amount and also of the support used in the
sample.
The TG profile for all the catalysts showed coke elimination
(burning) in the temperature range of 300–420 8C, although the
samples FeSi-3 and FeAl-14 pointed to a graphitic carbon presence
with an inflection point at 540 8C [27]. This temperature range for
coke burning is characteristic of carbon deposits with a low
molecular weight [34,36,37], but also of carbonaceous material
containing oxygen in its structure [16,34].
The negative effects of coke on catalyst conversion are well
understood, but the chemical nature is not sufficiently known.
Thus, with the aim of determining the specific functional groups
present in the carbonaceous deposits, the spent catalyst was
dissolved in concentrated HF, and the material (coke) extracted
was analyzed by Fourier transformed infrared (FT-IR). The
spectroscopy results reveal the presence of C–H aliphatic
(2918 cm1 with low intensity) and aromatic (750 cm1) groups;
additionally, a band at 1580 cm1 is observed, which is assigned to
C55C vibrations in polyaromatic coke. A weak absorption band at
1400 cm1, which may correspond to the C–H in-plane bending
mode of vinyl group, is also identified [38].
On the other hand, in the range of 1500–1000 cm1, the FT-IR
spectra reveal other absorption bands (1438, 1257, 1160 and
1110 cm1). The band at 1438 cm1 can be assigned to the C–H
bending mode of a vinyl group bonded to oxygen [39]; however,
this band may also be assigned to the C5 5O group of carboxylate-
and carbonate-type species [23,40]. The presence of C5 5O group in
the carbonaceous material is reinforced by the band with low
intensity at 1705 cm1. The bands at 1257, 1160 and 1110 cm1
are assigned to the OH, COC and CO groups [41]. The presence of
OH group in the carbonaceous samples is suggested by a wide band
 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200
at 3408 cm1, but the possible presence of water in the
carbonaceous sample cannot be ruled out.
The results from TG and FT-IR analysis point to the possibility
that the coke is generated from the ethylbenzene and CO2.
The carbon deposition from CO2 may be due to the consecutive
reactions of the CO2 with H2, which produce H2O and C (Eqs. (3)
and (5)). A good indication that such reactions take place in a
significant ratio was the low fraction of CO and H2 experimentally
observed, and the high ratio of H2O detected in the reaction
products.
CO þ H2 ! C þ H2 O (5)
In addition to the coke deposits, the catalytic deactivation in the
ethylbenzene dehydrogenation reaction has been attributed to the
reduction of the Fe3+ species [3,10], though the reduction of the
Fe3+ takes place as an effect of the basic site consumption (Eq. (1))
[25,30,31]. Nonetheless, knowledge of the iron reduction is needed
to find information about catalyst stability. Thus, with the aim of
confirming the occurrence of the reduction process (Fe3+ to Fe2+),
DRX analysis of the spent catalysts was carried out.
The profiles presented in Fig. 9 point to Fe3+ reduction, and for
an increase in the crystallite diameter during the reaction process
for the FeAl-4, FeSi-3 and FeSi-12 samples. On the other hand, the
sample FeAl-14 did not show diffraction peaks correlated with
phases which contain iron. Nonetheless, the reduction process
(Fe3+ to Fe2+) may have taken place in the FeAl-14 sample.
The diffraction patterns of the FeAl-4 sample point to the
presence of maghemite phase (g-Fe2O3 JCPDS 39-1346), and
hercynite phase (AlFe2O4 JCPDS 82-0582). In spite of the Fe3+
reduction observed for the sample FeAl-4, the main phase formed
was the hercynite. The formation of the AlFe2O4 phase contributes
to make the growth of the crystals of the Fe3O4 phase difficult.
As was demonstrated previously [42,43], the AlFe2O4 ferrite
phase presents a superior catalytic stability in the ethylbenzene
dehydrogenation reaction as compared with hematite (Fe2O3)
phase. Therefore, the superior catalytic conversion observed for the
sample FeAl-4 (Fig. 5) may be correlated with the formation of the
phase AlFe2O4.
The higher catalytic stability of the AlFe2O4 can be related to a
possible larger energy required for the abstraction of the basic site
(Reaction (1)); it can facilitate CO2 adsorption in such sites,
providing superior ethylbenzene conversion. On the other hand,
with the progress of the reaction process, the coke deposition can
promote an obstruction of the active sites, which is reflected in
conversion decrease.
Fig. 9. X-ray diffraction patterns of the samples after the catalytic process.
199
The diffraction pattern of the FeSi-12 and FeSi-3 samples points
to the magnetite phase formation (Fe3O4 JCPDS 01-1111);
however, the maghemite phase is still detected (g-Fe2O3 JCPDS
39-1346), as was observed for the FeAl-4 sample.
Catalysts based on iron oxide normally show the magnetite
phase formation during the deactivation process in the dehy-
drogenation of ethylbenzene; however, the maghemite phase
presence after the catalytic process was not reported before. Thus,
the presence of the maghemite phase, after the reaction process, is
probably due to the formation of a magnetite shell in the FeSi-3 and
FeSi-12 samples, or a hercynite shell phase in the sample FeAl-4,
during the reaction process.
4. Conclusion
Further studies are necessary for the elucidation of the
mechanism of ethylbenzene dehydrogenation in the presence of
CO2. However, the results presented here point to the necessity of
introducing a dopant for the oxidation of Fe2+ by the CO2. In
addition, it is necessary to improve the CO2 adsorption in order to
increase the catalytic stability, with the purpose of avoid
carbonaceous deposition.
The textural and structural action of aluminum is pointed out
by the high surface area and pore volume of the catalyst, and the in
situ formation of the ferrite AlFe2O4, which presents higher
catalytic stability.
Finally, the results of catalytic ethylbenzene conversion and
styrene selectivity presented by the samples synthesized by the
polymeric precursor method are promising.
Acknowledgments
The authors acknowledge the ‘‘Universidade Federal do Ceará’’
(UFC), CNPq/CT-PETRO, Dr. J.M. Sasaki (Laboratório de Raios X) for
the DRX measurements. T.P. Braga, C.V. Teixeira and A.N. Pinheiro
express gratitude for the scholarship from CNPq.
References
[1] O. Shekhah, W. Ranke, R. Schlogl, J. Catal. 225 (2004) 56.
[2] G.R. Meima, P.G. Menon, Appl. Catal. A 212 (2001) 239.
[3] M.S. Ramos, M.S. Santos, L.P. Gomes, A. Albornoz, M.C. Rangel, Appl. Catal. A 341
(2008) 12.
[4] T. Hirano, Appl. Catal. 40 (1988) 247.
[5] N. Mimura, I. Takahara, M. Saito, T. Hattori, K. Ohkuma, M. Ando, Catal. Today 45
(1998) 61.
[6] Y. Sekine, R. Watanabe, M. Matsukata, E. Kikuchi, Catal. Lett. 125 (2008) 215.
[7] S. Sato, M. Ohhara, T. Sodesawa, F. Nozaki, Appl. Catal. 37 (1988) 207.
[8] M. Sugino, H. Shimada, T. Tsuruda, H. Miura, N. Ikenaga, T. Suzuki, Appl. Catal. A
121 (1995) 125.
[9] M. Saito, H. Kimura, N. Mimura, H. Miura, J. Wu, K. Murata, Appl. Catal. A 239
(2003) 71.
[10] N. Mimura, M. Saito, Catal. Lett. 58 (1999) 59.
[11] J.S. Chang, S.E. Park, M.S. Park, Chem. Lett. 26 (1997) 1123.
[12] T. Badstube, H. Papp, P. Kustrowski, R. Dziembaj, Catal. Lett. 55 (1998) 169.
[13] T. Badstube, H. Papp, R. Dziembaj, P. Kustrowski, Appl. Catal. A 204 (2000) 153.
[14] Y. Sakurai, T. Suzaki, N.O. Ikenaga, T. Suzuki, Appl. Catal. A 192 (2000) 281.
[15] N. Mimura, M. Saito, Catal. Today 55 (2000) 173.
[16] F. Cavani, F. Trifirò, Appl. Catal. A 133 (1995) 219.
[17] A.C. Oliveira, J.L.G. Fierro, A. Valentini, P.S.S. Nobre, M.C. Rangel, Catal. Today 85
(2003) 49.
[18] R. Dziembaj, P. Kustrowski, L. Chmielarz, Appl. Catal. A 255 (2003) 35.
[19] A. Valentini, N.L.V. Carreño, L.F.D. Probst, A. Barison, A.G. Ferreira, E.R. Leite, E.
Longo, Appl. Catal. A 310 (2006) 174.
[20] A. Valentini, N.L.V. Carreño, L.F.D. Prost, P.N. Lisboa, W.H. Schreiner, E.R. Leite, E.
Longo, Appl. Catal. A 255 (2003) 211.
[21] A. Valentini, N.L.V. Carreño, L.F.D. Probst, E.R. Leite, E. Longo, Micropor. Mesopor.
Mater. 68 (2004) 151.
[22] M. Jarvinen, J. Appl. Crystallogr. 26 (1993) 525.
[23] H.M. Rietveld, J. Appl. Crystallogr. 2 (1967) 65.
[24] C. Kuhrs, Y. Arita, W. Weiss, W. Ranke, R. Schlögl, Top. Catal. 14 (2001) 111.
[25] W. Weiss, D. Zscherpel, R. Schlogl, Catal. Lett. 52 (1998) 215.
[26] J.C. Ryu, D.H. Lee, K.S. Kang, C.S. Park, J.W. Kim, Y.H. Kim, J. Ind. Eng. Chem. 14
(2008) 252.
200 T.P. Braga et al. / Applied Catalysis A: General 366 (2009) 193–200

[27] B. Pawelec, L. Daza, J.L.G. Fierro, J.A. Anderson, Appl. Catal. A 145 (1996) 307. [34] H. Hattori, Chem. Rev. 95 (1995) 537.
[28] B.D. Herzog, H.F. Raso, Ind. Eng. Chem. Prod. Res. Dev. 23 (1984) 187. [35] T.P. Braga, E. Longhinotti, A.N. Pinheiro, A. Valentini, Appl. Catal. A 362 (2009) 139.
[29] Y. Joseph, C. Kuhrs, W. Ranke, M. Ritter, W. Weiss, Chem. Phys. Lett. 314 (1999) [36] A. Effendi, K. Hellgardt, Z.-G. Zhang, T. Yoshida, Catal. Commun. 4 (2003) 203.
195. [37] M. Guisnet, P. Magnoux, Appl. Catal. A 212 (2001) 83.
[30] Y. Joseph, M. Wuhn, A. Niklewski, W. Ranke, W. Weiss, C. Woll, R. Schlogla, Phys. [38] M. Baghalha, O. Ebrahimpour, Appl. Catal. A 326 (2007) 143.
Chem. Chem. Phys. 2 (2000) 5314. [39] S.-H. Wang, P.R. Griffiths, Fuel 64 (1985) 229.
[31] A. Sun, Z. Qin, S. Chen, J. Wang, J. Mol. Catal. A 210 (2004) 189. [40] W. Wei, J.A. Moulijn, G. Mul, J. Catal. 262 (2009) 1.
[32] M. Hammouche, D. Lexa, M. Momenteau, J.M. Saveant, J. Am. Chem. Soc. 113 [41] T. Jiang, K. Xu, Carbon 33 (1995) 1663.
(1991) 8455. [42] E.H. Lee, Catal. Rev. Sci. Eng. 8 (1973) 285.
[33] C. Amatore, J.M. Saveant, J. Am. Chem. Soc. 103 (1981) 5021. [43] A. Miyakoshi, A. Ueno, M. Ichikawa, Appl. Catal. A 216 (2001) 137.