The enhancement of the catalytic performance of CrOx/Al2O3 catalysts for ethylbenzene dehydrogenation through tailored coke deposition

Article abstract of DOI:10.1039/c5cy01157d

In previous work we have shown that ethylbenzene dehydrogenation over CrO_x/Al₂O₃ catalysts proceeds sequentially via cracking and then dehydrogenation reactions. The present work reports how tailored coke deposition on the catalyst surface can suppress undesired reactions such as cracking to benzene and coke during ethylbenzene dehydrogenation. Additionally, this approach also provides insights into the precursor molecules involved in the formation of carbonaceous deposits, hence providing further understanding of coke formation. Pre-coked catalysts were prepared by adsorbing the products of the ethylbenzene reaction (i.e., benzene, toluene, styrene, ethylene) as single components, in a flowing system at 600 °C over the fresh catalyst. The resulting pre-coked catalysts were then evaluated in the ethylbenzene dehydrogenation reaction and their performance compared with that of the catalyst without exposure to pre-treatment. Characterisation of pre-coked catalysts by elemental analysis, temperature-programmed oxidation (TPO), temperature-programmed desorption (TPD), Raman spectroscopy and diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) indicated that ethylene is the main coke precursor during ethylbenzene dehydrogenation and that ethylene-derived coke is associated with a reduction in selectivity to styrene as compared to the fresh catalyst. Coke deposited after pre-coking with aromatic molecules, and in particular with benzene, was beneficial for dehydrogenation activity, as shown by the increase in styrene selectivity relative to the fresh catalyst. This enhancement of dehydrogenation activity was correlated with deactivation of acid sites and the reduction of chromium from Cr(vi) to Cr(iii) (active species for dehydrogenation) as a result of the pre-coking procedure.

Full text of DOI:10.1039/c5cy01157d

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
The enhancement of the catalytic performance
of CrOx/Al2O3 catalysts for ethylbenzene
dehydrogenation through tailored coke
deposition†
Cite this: DOI: 10.1039/c5cy01157d
a
a
Sara Gomez Sanz, Liam McMillan, James McGregor,‡ a J. Axel Zeitler,a Nabil Al-Yassir,b
*
Sulaiman Al-Khattafb and Lynn F. Gladdena
In previous work we have shown that ethylbenzene dehydrogenation over CrOx/Al2O3 catalysts
proceeds sequentially via cracking and then dehydrogenation reactions. The present work reports how
tailored coke deposition on the catalyst surface can suppress undesired reactions such as cracking to
benzene and coke during ethylbenzene dehydrogenation. Additionally, this approach also provides
insights into the precursor molecules involved in the formation of carbonaceous deposits, hence
providing further understanding of coke formation. Pre-coked catalysts were prepared by adsorbing the
products of the ethylbenzene reaction (i.e., benzene, toluene, styrene, ethylene) as single components,
in a flowing system at 600 °C over the fresh catalyst. The resulting pre-coked catalysts were then eval-
uated in the ethylbenzene dehydrogenation reaction and their performance compared with that of the
catalyst without exposure to pre-treatment. Characterisation of pre-coked catalysts by elemental analy-
sis, temperature-programmed oxidation (TPO), temperature-programmed desorption (TPD), Raman
spectroscopy and diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) indicated that
ethylene is the main coke precursor during ethylbenzene dehydrogenation and that ethylene-derived
coke is associated with a reduction in selectivity to styrene as compared to the fresh catalyst. Coke
deposited after pre-coking with aromatic molecules, and in particular with benzene, was beneficial for
dehydrogenation activity, as shown by the increase in styrene selectivity relative to the fresh catalyst.
This enhancement of dehydrogenation activity was correlated with deactivation of acid sites and the
reduction of chromium from CrIJVI) to CrIJIII) (active species for dehydrogenation) as a result of the pre-
coking procedure.
Received 22nd July 2015,
Accepted 2nd September 2015
DOI: 10.1039/c5cy01157d
www.rsc.org/catalysis
desired product.5–8 Additionally, coke deposits have long been
postulated to promote catalytic reactions including hydroge-
1. Introduction
A large number of studies on the influence of carbonaceous
deposits on the catalytic activity and selectivity of zeolites1,2
and metal-oxide catalysts3,4 are reported in the literature.
Most of these investigations refer to the detrimental effect of
coke deposition on the catalytic performance by blocking
pores or by poisoning the active sites. However, increasingly
there is discussion around the idea that certain coke struc-
tures can play a beneficial role, and improve selectivity to the
nation through facilitating hydrogen transfer, and methanol-
to-hydrocarbon conversion through the hydrocarbon pool
mechanism.9,10 Industrially, there is considerable interest
that exists in developing methods to enhance the selectivity
of products by surface modification of zeolite catalysts. Such
methods include pre-treatments that typically imply the inac-
tivation of external active sites.11 The most effective ways to
deactivate non-selective acid sites in zeolites include chemical
vapour/liquid deposition of organosilicon compounds,12,13
pre-coking treatment,14,15 selective deactivation of acid sites
with basic molecules16 or impregnation with metals such as
cerium.17
a University of Cambridge, Department of Chemical Engineering and
Biotechnology, Cambridge CB2 3RA, UK. E-mail: james.mcgregor@sheffield.ac.uk
b King Fahd University of Petroleum & Minerals, Dhahran 31261, Saudi Arabia
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5cy01157d
Pre-coking or tailored carbon deposition refers to the con-
trolled laydown of carbonaceous deposits to enhance the
selectivity of cracking catalysts and zeolites in hydrocarbon
processing. This approach has been used in the past18 and it
‡ Present address: University of Sheffield, Department of Chemical and Biologi-
cal Engineering, Sheffield S1 3JD, UK.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
enables: i) identification of the molecular species responsible 2. Experimental
for the generation of the precursors of coke; and ii) enhance-
ment of catalytic performance through poisoning of acid
2.1. Catalyst synthesis
sites, pore blocking or deposition of active coke.19 Reactions
involved in hydrocarbon processing such as alkylation, hydro-
genation and isomerisation can be notably improved through
tailored pre-coking. For instance, Mandal et al. reported that
a level of 0.2–0.3% of coke on a regenerated FCC catalyst
maximised the fraction of medium distillates and limited
over-cracking.20 Elsewhere, Lee et al.21 showed that p-xylene
did not react over fresh FCC catalyst at 200 °C but alkylation
occurred if the catalyst had been pre-coked with isopropanol
at 500 °C. A further example where pre-coking has been suc-
cessfully applied is the selective toluene disproportionation
process (MSTDP™), where the catalyst is pre-coked with aro-
matic feedstocks at elevated temperature during the initial
stages of the treatment.22,23 In the work of Haag and co-
workers,24,25 pre-coking of HZSM-5 with an aromatic feed-
stock led to a coke content exceeding ~2 wt% which was suf-
ficient to increase the selectivity to p-xylenes from toluene
disproportionation. Despite the advantages offered by tai-
lored carbon deposition, there are also examples of ineffec-
tive pre-coking in the literature. Laforge et al., studied the
effect of pre-coking through pre-adsorption of n-heptane over
MCM-22 for m-xylene isomerization.26 It was concluded that
this pre-treatment was not able to selectively deactivate the
external acid sites of the zeolite. In addition, pre-coking can
cause pore narrowing, hence introducing diffusion limita-
tions which can diminish catalytic performance.27
The present work focuses on the surface modification of
CrOx/Al2O3 by pre-coking the catalyst through adsorbing, in
separate experiments, the products of the ethylbenzene dehy-
drogenation reaction: benzene, toluene, styrene and ethylene.
Characterisation of the catalyst surface after pre-coking, but
before exposure to the dehydrogenation reaction, and then
again post-reaction, enables us to identify the dominant pre-
cursor molecules in coke formation during ethylbenzene
dehydrogenation and hence how pre-coking might be
exploited to suppress undesired reactions such as cracking to
benzene and coke during the dehydrogenation reaction.
Characterisation of pre-coked catalysts by elemental analysis,
temperature-programmed oxidation (TPO), temperature-
programmed desorption (TPD), terahertz-time-domain
spectroscopy (THz-TDS), Raman spectroscopy and diffuse
reflectance infrared Fourier transform spectroscopy (DRIFTS)
provides information on the coke structures which improve
dehydrogenation activity and the precursor molecules which
are responsible for carbon deposition. X-ray photoelectron
spectroscopy (XPS) was used to determine the oxidation sates
of chromium.
An alumina-supported chromia catalyst was prepared by wet
impregnation using CrIJNO3)3·9H2O (Sigma-Aldrich, 99%); the
alumina support used was γ-Al2O3 (Condea Chemie GmbH,
Hamburg, Germany, BET surface area = 216 m2 g−1). Approxi-
mately 15.98 g of CrIJNO3)3·9H2O and 20 g of γ-Al2O3 were
added to 300 ml of water. The catalyst precursor was then
mixed for 3 h at room temperature to ensure a homogeneous
distribution of chromia on the support. The solution was
then vacuum filtered and the filtrate was washed with water
several times. The as-prepared solid was dried in air at 120 °C
overnight and calcined for 24 h at 600 °C, also in air. The
colour of the catalyst changed from green to yellow after cal-
cination. The chromium loading was 0.8 wt% which is signif-
icantly below monolayer coverage, ~12 wt%.28 It has been
found that below monolayer coverage the predominant sur-
face chromium species are expected to be polymeric and
monomeric CrIJVI).29 Characterisation of the fresh catalyst by
Raman spectroscopy showed bands in the 882–890 cm−1
range corresponding to monochromate and polychromate
species.30
2.2. Pre-coking and catalytic activity measurements
Carbonaceous deposits derived from pre-coking with differ-
ent precursor molecules were formed by pre-adsorbing sepa-
rately the products of the reaction (benzene, toluene, styrene
and ethylene) at 600 °C over CrOx/Al2O3. Pre-coking of CrOx/
Al2O3 was carried out using a fixed-bed reactor (inner diame-
ter: 14 mm; total length: 40.2 cm) connected to an on-line GC
(Agilent 6890N Series, column Agilent HP-5) equipped with
an FID detector and an Agilent HP-5 column. 1 g of CrOx/
Al2O3 (particle size, D50: 100 μm; height of bed: 12 mm) was
loaded into the reactor and was pretreated under 40 ml min−1
of helium at 600 °C for 30 min. Adsorbates (benzene,
toluene and styrene) were introduced in a saturated He flow
via a bubbler heated at 70 °C; the accuracy of the outlet com-
position was 5%. The flow was adjusted to feed the same
number of carbon molecules of precursor into the reactor for
a period of 360 min (3.2 × 1023 carbon molecules) and to
ensure a total volumetric flow rate of 45 ml min−1 of helium and
ethylbenzene. In the case of ethylene pre-coking, 16.5 ml min−1
of ethylene (99.99%, Air Liquide) were mixed with 28.5 ml min−1
of helium and preheated at 70 °C, thereby maintaining the
overall amount of carbon molecules fed to the reactor for 360
min (3.2 × 1023 carbon molecules).
Ethylbenzene dehydrogenation over CrOx/Al2O3 and pre-
coked catalysts was performed in the same reactor set-up as
used for pre-coking. Approximately 1 g of CrOx/Al2O3 catalyst
was pretreated under helium (1 barg, 40 ml min−1) at 600 °C
for 30 min. After this period, 40 ml min−1 of helium were
flowed through a saturator filled with ethylbenzene at 70 °C.
The reactor was fed with 45 ml min−1 of a mixture of helium
and ethylbenzene maintaining a helium/ethylbenzene ratio of
7.75 at atmospheric pressure. Gas chromatography
In summary, the aims of this work are to identify: i) if the
pre-coking approach can be used to determine the main pre-
cursor molecule of coke in ethylbenzene dehydrogenation;
and ii) whether or not a pre-coking treatment can suppress
undesired reactions, such as cracking to benzene and coke,
leading to an enhanced selectivity to dehydrogenation.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
measurements were taken every 20 min during a total reac-
tion time of 360 min. Conversion and selectivity to benzene,
toluene, styrene and coke were determined according to eqn
(1) and (2) and averaged over the steady-state regime. It
should be noted that conversion is calculated considering
conversion to all products including coke. Selectivity to
coke was calculated by applying a carbon mass balance as
shown in eqn (3). After 6 h the catalyst was cooled down
under 40 ml min−1 helium and removed for subsequent
characterisation.
Kaiser Optical Systems Inc. supplied by Clairet Scientific).
Samples were placed on a cover glass at room temperature.
Exposure time was 2 min and the laser power was set at
50 mW in order to minimise fluorescence/coke combustion
effects. Spectra of both the fresh catalyst and empty sample
holder were recorded for subsequent background subtrac-
tions. The spectra were fitted by using four components (G,
D1, D3 and D4) and analysed in terms of positions, relative
intensity and full width at half maximum amplitude
(FWHM).
DRIFTS spectra were obtained without sample dilution in
a Thermo Nicolet NEXUS 670 spectrometer equipped with a
diffuse reflectance attachment and MCT detector. The mea-
surements were performed at room temperature by adding
128 scans at a resolution of 4 cm−1. The spectrum of fresh
CrOx/Al2O3 was recorded as a background and subtracted
from the spectra of pre-coked catalysts, such that the
observed infrared absorptions could be attributed to the coke
deposits present on the catalyst surface.
(1)
(2)
XPS analyses were carried out at the Sheffield Surface
Analysis Centre, (SSAC), Sheffield. XPS experiments of coked
CrOx/Al2O3 after ethylbenzene dehydrogenation were
conducted to analyse the oxidation states of chromium.
These experiments were performed on an Axis Ultra DLD
(Kratos) with a monochromated Al-Kα X-ray source operated
in electrostatic mode. The vacuum during analysis was ca.
1 × 10−8 mbar. Spectral analysis was carried out using
CasaXPSsoftware (Casa, http://www.casaxps.com, UK); survey
and narrow scan C1s and O1s spectra were obtained.
where i = benzene, toluene, styrene and coke
ncoke = n(Cethylbenzene,in) − n(Cethylbenzene,out) − n(Cproducts,out)(3)
2.3. Characterisation techniques
The catalysts were characterised after pre-coking but before
exposure to the ethylbenzene dehydrogenation reaction, and
then again post-reaction, using a variety of techniques.
Elemental analysis (Microanalytical Department, Depart-
ment of Chemistry, University of Cambridge) was employed
in order to determine the percentages of carbon, hydrogen
and nitrogen present. The determination of C/H ratio was
conducted via combustion of the coked samples in pure oxy-
gen at 950 °C. The resulting carbon dioxide, water and nitro-
gen were measured using thermal conductivity detectors. The
instrument utilised was an Exeter Analytical CE-440 elemen-
tal analyser.
3. Results
3.1. Catalytic activity measurements
As discussed in a previous study,33 the reaction profile of eth-
ylbenzene dehydrogenation over fresh CrOx/Al2O3 at 600 °C
shows two well defined periods of activity (Fig. 1): (i) an
induction period before 3 h time-on-stream characterised by
cracking activity; ii) styrene breakthrough after 3 h time-on-
stream where dehydrogenation activity prevails. Cracking
TPO was conducted on the coked catalysts in a Catlab
Microreactor Module connected to an online quadrupole
mass spectrometer. A gas mixture of composition 5% O2/He
at a total flow rate of 40 ml min−1 was used. The heating rate
was 10 °C min−1 from 50 °C to 875 °C. TPD was performed in
the same equipment by heating the catalyst from 50 °C to
900 °C at a rate of 10 °C min−1 under a helium flow rate of
40 ml min−1
.
THz-TDS measurements were performed at 293 K on a
spectrometer described previously.31,32 Samples were pre-
pared by mixing 50 mg of coked catalyst with 150 mg of poly-
ethylene (Sigma-Aldrich) which is transparent to THz
radiation.
Raman spectroscopy experiments were conducted using a
Raman spectrometer employing a 784.41 nm laser excitation
(3 mW) at 5 cm−1 spectral resolution (Raman Rxn1 Systems,
Fig. 1 Conversion of ethylbenzene (△) and selectivity to toluene (■),
benzene (■) and styrene (■) after ethylbenzene dehydrogenation over
CrOx/Al2O3 at 600 °C.33
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 2 Classification of coke structures according to their C/H mass
ratio and % Ccoke. The square and round brackets show closed and open
intervals respectively
products such as benzene and toluene are mainly formed
during the first 3 h of reaction time with maximum selectivity
values of 23% and 5%. Selectivity to styrene reaches a maxi-
mum after 165 min (25%) prior to achieving a steady-state
value of 15% with an ethylbenzene conversion of 97%. Selec-
tivity to coke after 6 h time-on-stream is 81%.
C/H mass
Coke structure
ratio range
% Ccoke range
Linear alkanes (CnH2n+2
Linear alkenes with
)
(3, 6)
6
(75, 85.7)
85.7
1 double bond (CnH2n
Linear alkenes with
n double bonds
)
(6, 4n)
(85.7, 100)
3.2. Study of the nature of coke formed by tailored carbon
deposition
Cycloalkanes ((CH2)n)
6
85.7
Cycloalkenes (CnH2n−2
Benzene (C6H6)
Alkylaromatics
Linear polyaromatics with
6 membered-rings
Non-linear polyaromatics
with 6 membered-rings
Polyaromatics with n
membered-rings
(mixture of 4, 5, 6…n
membered-rings)
)
(6, 8)
12
(6, 12)
(15, 24)
(85.7, 88.9)
92.3
(85.7, 92.3)
(93.8, 96)
After pre-coking, but prior to exposure to the ethylbenzene
dehydrogenation reaction, the catalysts pre-coked with differ-
ent molecules were characterised using the methods
described in section 2. Alongside these data, catalytic activity
measurements were performed and are reported in section
3.3. These data are then used: i) to determine which reactant/
product is the main precursor molecule of coke in ethylben-
zene dehydrogenation and ii) to assess whether a pre-coking
method can be used to suppress undesired side reactions
such as cracking to benzene or coke.
(16.8, 24)
(12, 24)
(94.4, 96)
(92.3, 96)
The results of the characterisation of the pre-coked cata-
lysts before exposure to the ethylbenzene dehydrogenation
reaction are now presented.
on average, in the range corresponding to polyaromatic coke
compounds with n membered-rings. The C/H ratio and %
Ccoke for the catalyst pre-coked with toluene are 9.9 and
90.8%, values that on average, correspond to alkylaromatic
coke compounds. Since elemental analysis provides a quanti-
fication of average coke structures, complementary data from
other characterisation techniques (e.g., TPO, Raman, DRIFTS,
etc.) are required in order to gain a greater understanding of
the coke structures present.
Temperature-programmed oxidation (TPO). TPO measure-
ments are shown in Fig. 2. The oxidation temperature allows
discrimination between coke of different types;34 in particu-
lar, hydrogen-rich coke deposits are oxidised at temperatures
lower than 327 °C, while highly condensed or polyaromatic
coke is oxidised at higher temperatures of ~427 °C. The area
Elemental analysis. As reported in Table 1, CrOx/Al2O3 pre-
coked with aromatic molecules showed the highest carbon
contents (11.1, 10.7 and 8.9 wt%, for benzene, styrene and
toluene, respectively). In contrast, the coke content deposited
on the ethylene pre-coked catalyst is much lower, at 2.7 wt%.
This value approaches the sensitivity limit of the equipment,
therefore no accurate conclusion can be drawn as to the type
of coke structure for the catalyst pre-coked with ethylene. By
comparing the data in Table 1 with the listing of classifica-
tion of coke structures according to their C/H mass ratio and
% Ccoke, as tabulated in Table 2, some inferences as to the
nature of the deposited coke can be made. However, it is
important to note that the heterogeneous nature of carbona-
ceous structures makes their identification a complex pro-
cess. The values of C/H ratio and % Ccoke yielded by elemen-
tal analysis are average values corresponding to an average
coke structure, which may differ significantly from the actual
coke structure. Nevertheless, the values of C/H ratio and %
Ccoke for the catalyst pre-coked with benzene and styrene fall,
Table 1 Elemental microanalysis showing the wt% carbon and hydrogen
and C/H mass ratio after pre-coking with benzene, toluene, styrene and
ethylene over CrOx/Al2O3 catalysts at 600 °C for 6 h time-on-stream and
fresh CrOx/Al2O3 after ethylbenzene dehydrogenation at 600 °C for 6 h
time-on-stream
CrOx/Al2O3 pre-coked
with
% C
(wt%)
% H
(wt%)
C/H mass
ratio % Ccoke
Benzene
Toluene
Styrene
Ethylene
CrOx/Al2O3 after EB
dehydrogenation
11.1 0.1 0.8 0.1 13.9 0.9 93.3 0.4
8.9 0.1 0.9 0.1 9.9 0.2 90.8 0.2
10.7 0.1 0.7 0.1 15.2 0.9 93.9 0.3
2.7 0.1 0.6 0.1 4.5 0.2 81.8 0.7
6.3 0.2 0.6 0.1 10.5 0.9 91.3 0.6
Fig. 2 TPO of fresh CrOx/Al2O3 after ethylbenzene dehydrogenation
and after pre-coking with benzene, toluene, styrene and ethylene at
600 °C.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
of the TPO bands is related to the amount of carbon-
containing material present. By comparing the oxidation
temperature of each pre-coked sample with that of the fresh
catalyst coked at 600 °C after ethylbenzene dehydrogenation
(non-pre-coked catalyst), an indication as to the nature of
the precursor molecule of coke (benzene, toluene, styrene or
ethylene) can be elucidated. As seen from Fig. 2, the oxida-
tion temperature of the coked CrOx/Al2O3 catalyst after ethyl-
benzene dehydrogenation is ~455 °C. The temperature at
which the maximum intensity of the TPO peak falls for each
of the precursor molecules is 495, 491, 483 and 463 °C for
benzene, styrene, toluene, and ethylene, respectively. The
oxidation temperatures of the catalysts pre-coked with ben-
zene, toluene and styrene are significantly higher than the
oxidation temperature over the non-pre-coked catalyst,
which is indicative of the higher polyaromaticity of the coke
laid down over the pre-coked catalysts. In contrast, the oxi-
dation temperature of coke deposited from ethylene pre-
coking is very close to that of the fresh catalyst after ethyl-
benzene dehydrogenation. Additionally, we have previously
observed that oxidation temperature is a function of coke
precursor and does not vary significantly with time-on-
stream.33 This suggests that the chemical nature of coke laid
down over the non-pre-coked catalyst is similar to that
observed over the catalyst pre-coked with ethylene.
species for dehydrogenation over the pre-coked CrOx/Al2O3.
The species desorbed were analysed by mass spectrometry
and the mass-to-charge ratios monitored were m/z = 91 (tolu-
ene), m/z = 78 (benzene), m/z = 104 (styrene), m/z = 106 (ethyl-
benzene) and m/z = 28 (ethylene).
The TPD of the catalyst after pre-coking with toluene
(Fig. 3a) shows a peak at 165 °C which is associated with tol-
uene physisorption. The signal for m/z = 78, i.e. for coke
which has undergone decomposition into benzene, shows a
main band at 600 °C with a shoulder at 750 °C. For the cata-
lyst pre-coked with benzene (Fig. 3b) the TPD profile only
shows one band at 800 °C for m/z = 78. The TPD of CrOx/
Al2O3 after pre-coking with styrene (Fig. 3c) shows peaks at
210 °C (both for m/z = 104 and m/z = 78) due to styrene
physisorption and also bands at 600 and 750 °C for m/z =
78. Therefore, there are three adsorption sites (at 600, 750
and 800 °C) characterised by different adsorption energies
where coke is formed. Since the low energy adsorption site
(600 °C) is only present for the alkyaromatic adsorbates (tol-
uene and styrene) it is proposed that it corresponds to a
dealkylation site. In contrast, the 750 and 800 °C bands
appear for the catalysts pre-coked with styrene, benzene and
toluene, therefore these sites may correspond to cracking
centres with increasing adsorption strength. Finally, the TPD
profile of the catalyst after pre-coking with ethylene (Fig. 3d)
shows very poor signal-to-noise due to the low amount of
coke present. Small peaks observed between 170 and 300 °C
are attributed to physisorbed ethylene and the main band at
Temperature-programmed desorption (TPD). This tech-
nique allows the identification of adsorption sites on the cat-
alyst and provides information on the initial adsorbed
Fig. 3 TPD (helium flow rate = 40 ml min−1, 5 °C min−1) of pre-coked CrOx/Al2O3 with a) toluene, b) benzene, c) styrene and d) ethylene.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
450 °C is ascribed to CO released due to thermal decomposi-
tion of coke.
sp2–sp3 bonds or C–C and CC stretching vibrations of
polyene-like structures.37
Raman spectroscopy. The position of the Raman bands
and the D1/G and D1/D3 ratios of the pre-coked catalysts
after reaction are shown in Table S1.† The method of analysis
follows that reported previously.33 Deconvolution of D1, D3
and D4 bands was performed by considering Gaussian-
The Raman spectra of the fresh CrOx/Al2O3 catalyst after
ethylbenzene dehydrogenation (with no pre-coking) (dotted
lines) and the pre-coked catalysts (solid lines) are shown in
Fig. 4. The shape of the spectra as well as the position and
width of the D1 and G bands allows comparison between the
nature of coke derived from precursor molecules and that
from ethylbenzene dehydrogenation over the fresh catalyst.
The position of the G band for all the pre-coked catalysts is
very similar to that of the fresh catalyst after reaction. How-
ever, significant differences are observed in the position of
the D1 band, especially for the catalysts pre-coked with aro-
matic molecules (Fig. 4a–c). For instance, the D1 band for
the catalyst pre-coked with benzene appears at 1303 cm−1
whereas it is located at 1363 cm−1 for the fresh catalyst after
reaction (Fig. 4b). This is indicative of the more polyaromatic
nature of the coke deposited after pre-coking with benzene.
Conversely, both the D1 and G bands appear practically at
the same Raman shift for the CrOx/Al2O3 pre-coked with eth-
ylene and the fresh catalyst after reaction (Fig. 4d). In addi-
tion, their similar band widths and spectral shapes reflect
that the structure of coke laid down after reaction over the
fresh CrOx/Al2O3 catalyst is similar to that formed after pre-
coking with ethylene, in agreement with TPO data.
shaped peaks whereas the
G band was fitted with a
Lorentzian function. Both D1/G and D1/D3 ratios were
determined in order to characterise the degree of graphitic
order or structural organisation of the coke. At D1/G values
greater than 1.1 this parameter is not reliable for the deter-
mination of the degree of order of the carbonaceous
deposits present.35 Therefore D1/D3 is a better indicator of
the structural order of coke in this work. Fitting was
performed by setting the positions of these bands according
to the values reported in the literature for carbonaceous
materials36,37 and establishing a boundary of 10 cm−1.
Values of FWHM were allowed to vary freely. The analysis
suggests that coke formed over the catalyst pre-coked with
benzene is the most graphitic as it shows the highest D1/D3
ratio (4.0) whereas coke laid down after ethylene pre-coking
shows the lowest graphiticity (1.9). For all the pre-coked cat-
alysts, the G band is shifted to values higher than 1580 cm−1
which is indicative of
a low degree of condensation.
The D4 band is more intense for the catalysts pre-coked
with toluene and styrene; this feature is associated with
Diffuse-reflectance infrared Fourier transform spectro-
scopy (DRIFTS). Fig. 5 shows the DRIFTS difference spectra
Fig. 4 Raman spectra of fresh CrOx/Al2O3 after ethylbenzene dehydrogenation (---) and after pre-coking (–), with a) toluene, b) benzene, c) sty-
rene and d) ethylene.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Fig. 5 DRIFTS difference spectra of CrOx/Al2O3 after pre-coking with a) toluene, b) benzene, c) styrene and d) ethylene at 600 °C, obtained by
subtracting the spectrum of fresh CrOx/Al2O3 from those of the pre-coked catalysts, such that the resulting absorptions are attributed to the coke
deposits present on the catalyst surface.
of the pre-coked catalysts obtained by subtracting the spec-
trum of the fresh CrOx/Al2O3 catalyst from those of the pre-
coked catalysts. The difference spectra of the catalysts pre-
coked with aromatic molecules are very similar. The positive
absorptions at 1470–1430 cm−1 and 1380–1365 cm−1 are due
to methyl C–H asymmetric and symmetric bending, respec-
tively.38 The bands at 1580–1560 cm−1 are attributed to
stretching of aromatic rings. The bands at 958–953 cm−1 can
be ascribed to aromatic C–H in-plane bending.39 The absorp-
tions between 850 and 670 cm−1 are assigned to C–H out-of-
plane bending in an aromatic ring. The location of these
bands provides information about the nature of the substitu-
tion and allows discrimination between mono-, di- or poly-
substitutions.39 A band located around 816 cm−1 is observed
as a shoulder of the 958 cm−1 band for the catalyst pre-coked
with toluene suggesting para-substitution of the aromatic
rings. This band has a lower intensity for the catalysts pre-
coked with benzene and styrene indicating the presence of
less substituted polyaromatic coke compounds. The catalyst
pre-coked with ethylene shows an intense positive band at
1580 cm−1 revealing the presence of aromatic rings. There is
also strong absorption at 1470, 1430 and 1370 cm−1 associ-
ated with methyl C–H asymmetric and symmetric bending.
The single absorption located at 833 cm−1 reflects para-
substitution of aromatic rings. The small band at 1730 cm−1
is ascribed to carbonyl compounds.38
which is due to the contribution of CrO groups associated
with CrIJVI)–O species (~1000 cm−1) and Al–O stretching vibra-
tions (1200–700 cm−1);40 ii) the band at 1650 cm−1 associated
with uncharged molecular adsorbed oxygen on CrIJVI)–O
sites;41,42 iii) the weak-moderate absorption centered around
2100 cm−1 assigned to chromates or CrIJVI)–O sites;43 iv) the
broad band located at 3520 cm−1 indicative of hydroxyl
groups on the alumina support;44 and v) the band at 1650 cm−1
corresponding to the bending vibration of H2O.45 It is
clearly seen that pre-coking causes significant loss of these
five functionalities indicating that this treatment blocks the
acid sites (as shown by the removal of hydroxyl groups) and
reduces the catalyst to CrIJIII), as mirrored by the loss of Cr-
IJVI)–O sites. However, coke derived from aromatic molecules
(benzene, toluene and styrene) is more effective than coke
derived from ethylene at removing these functionalities as
shown by the greater negative intensity of OH, CrIJVI)–O and
Al–O (Fig. 5a–c).
The DRIFTS spectrum of the fresh catalyst after ethylben-
zene dehydrogenation at 600 °C is shown in Fig. 6. The posi-
tive bands at 1580 and 1500 cm−1 are associated with aro-
matic ring stretching. The absorptions at 1470, 1420 and the
shoulder at 1370 cm−1 correspond to methyl C–H asymmetric
and symmetric bending. The small band centred at 1710 cm−1
shows the presence of carbonyl compounds.39 By com-
paring the coke region (1700–1300 cm−1) of the spectrum
shown in Fig. 6 with those of the pre-coked catalysts (Fig. 5),
there are clear similarities between the coke deposited on
Five negative bands are also present for all the pre-coked
catalysts: i) a broad band centered at around 1100 cm−1
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
pre-coked with ethylene (Fig. 5d) and that of fresh CrOx/Al2O3
after ethylbenzene dehydrogenation (Fig. 6) are consistent
with a more discrete, molecular structure similar to that of
p-xylene (Fig. 7b). While the coke structures are likely to be
heterogeneous in nature, this analysis indicates likely differ-
ences in the degree of condensation of the coke formed.
3.3. Reaction of pre-coked catalysts and characterisation of
post-reaction materials
In the following section the catalytic performance of the pre-
coked catalysts is presented, along with the characterisation
of the pre-coked catalysts following exposure to the ethylben-
zene dehydrogenation reaction.
3.3.1. Ethylbenzene dehydrogenation over pre-coked cata-
lysts. Ethylbenzene dehydrogenation was performed over the
fresh CrOx/Al2O3 (Fig. 1) and the pre-coked catalysts (Fig. 8).
As shown in Fig. 1, the selectivities to benzene and toluene
are high over the cracking period (0–3 h) for the fresh cata-
lyst. Conversely, for all the pre-coked catalysts (Fig. 8) there is
no formation of benzene and toluene as evidenced by the fact
that no induction time is observed, which suggests that the
pre-coking treatment blocks the acid sites of the alumina
support. It is noted that for the catalyst pre-coked with tolu-
ene (Fig. 8a) the selectivity to styrene increases monotonically
to 42% at 165 min time-on-stream and then decreases,
levelling-off at ~24% at 245 min, in a similar manner to the
fresh catalyst after ethylbenzene dehydrogenation (Fig. 1). In
a previous study,33 it was observed that methane was pro-
duced as a co-product of ethylbenzene hydrogenolysis into
toluene and from CO methanation during ethylbenzene dehy-
drogenation. For the catalyst pre-coked with toluene, forma-
tion of methane was also observed during ethylbenzene dehy-
drogenation. Furthermore, water was present in the gas
phase due to dehydration of surface hydroxyl groups and fur-
ther chromium reduction by ethylbenzene.33 The change in
selectivity to styrene at 165 min (Fig. 8a) may be due to the
competition between, on the one hand, ethylbenzene steam
reforming and CO methanation and, on the other hand, eth-
ylbenzene dehydrogenation.
Fig. 6 Diffuse reflectance infrared Fourier transform spectroscopy of
fresh CrOx/Al2O3 after ethylbenzene dehydrogenation at 600 °C.
CrOx/Al2O3 during ethylbenzene dehydrogenation and on the
catalyst pre-coked with ethylene. The similarity between the
DRIFTS spectra of the pre-coked catalyst with ethylene and
the fresh catalyst after ethylbenzene dehydrogenation (Fig. 5d
and 6) is due to the presence of similar carbonaceous struc-
tures. To confirm that the similarity between the bands
shown in Fig. 5d and 6 can be unambiguously assigned to
the similar nature of the carbon deposited as opposed to the
amount of coke deposited, a sample of the catalyst with a
much higher level of pre-coking with the ethylene precursor
was analysed. The features obtained from DRIFTS analysis
was identical to those seen in Fig. 5d and 6.
In an attempt to assign chemical structures to the coke
deposited after pre-coking, the DRIFTS spectra shown in
Fig. 5 were compared with spectra of typical model coke com-
pounds including alkylbenzenes and polyaromatic com-
pounds (NIST Chemistry Web Book; http://webbook.nist.gov).
Fig. 7 shows the infrared spectra of 1-methyl naphthalene
(Fig. 7a) and p-xylene (Fig. 7b) as two examples. It can be
observed that the catalysts pre-coked with aromatic mole-
cules (Fig. 5a–c) exhibit a structure similar to 1-methyl naph-
thalene (Fig. 7a) as shown by the weak absorptions at 1570,
1450 and 1370 cm−1. In contrast, the spectra of the catalyst
The selectivity to styrene for the catalysts pre-coked with
aromatic molecules was compared to that of the fresh
Fig. 7 FTIR spectra of a) 1-methyl naphthalene and b) p-xylene (NIST Chemistry Web Book; http://webbook.nist.gov).
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Fig. 8 Selectivity to toluene (■), benzene (■) and styrene (◆) for ethylbenzene dehydrogenation (600 °C, 6 h time-on-stream) over CrOx/Al2O3
pre-coked with a) ethylene, b) toluene, c) benzene and d) styrene. The inset plots show the conversion of ethylbenzene (△) versus time-on-stream.
catalyst. The higher selectivity obtained over the catalysts pre-
coked with aromatic molecules was correlated with the deac-
tivation of CrO sites associated with CrIJVI)–O species
caused by the pre-coking treatment. These results were also
confirmed by DRIFTS studies of the pre-coked catalysts
(Fig. 5). Table 3 shows the values of conversion and selectivity
for the fresh and pre-coked catalysts after ethylbenzene dehy-
drogenation. The selectivities to styrene and coke for the
fresh catalyst were 15.1% and 80.1%, respectively. The com-
parison of these values with those of the pre-coked catalysts
identifies benzene as the best coking agent in terms of
improving catalytic performance, with the highest selectivity
to styrene (31.0%) and the lowest selectivity to coke (66.7%).
Similar values of catalytic activity are observed for the catalyst
pre-coked with styrene (29.1% selectivity to styrene and
69.1% selectivity to coke). In contrast, pre-coking with
Table 3 Comparison of catalytic performance of fresh and pre-coked CrOx/Al2O3 for ethylbenzene dehydrogenation at 600 °C. The average error in
conversion and selectivity values is ~5% of the stated value
Conversion and selectivity of ethylbenzene dehydrogenation over
Fresh
CrOx/Al2O3
Ethylene pre-coked
CrOx/Al2O3
Toluene pre-coked
CrOx/Al2O3
Benzene pre-coked
CrOx/Al2O3
Styrene pre-coked
CrOx/Al2O3
% conversion
97
98
90
87
94
% selectivity
Toluene
Benzene
Styrene
Methane
Ethylene
Coke
1.3
2.9
15.1
0.3
0.3
0.5
0.4
11.4
0.2
0.2
0.7
0.8
24.4
0.3
0.3
0.9
1.0
31.0
0.2
0.2
0.8
0.4
29.1
0.3
0.3
80.1
87.3
73.5
66.7
69.1
Time to reach steady state (min)
Induction time (min)
185
145
205
25
245
0
145
0
165
0
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 5 CrIJIII)/CrIJVI) ratios for fresh CrOx/Al2O3, after ethylbenzene dehy-
drogenation for 1, 3 and 6 h time-on-stream and after pre-coking with
benzene
ethylene gives the lowest selectivity to styrene (11.4%) and
the highest selectivity to coke (87.3%). Further investigation
was conducted to confirm the process responsible for the
improvement in styrene selectivity for the catalyst pre-coked
with benzene, i.e. whether it results from selective deactiva-
tion of acid sites or from a change in chromium oxidation
state. The catalyst acidity and the distribution of chromium
sites were studied by NH3-TPD and XPS, respectively. The
methods of analysis follow those described previously.33
Table 4 shows the number and distribution of acid sites, as
determined by NH3-TPD, for fresh CrOx/Al2O3, CrOx/Al2O3
after ethylbenzene dehydrogenation at different times-on-
stream and CrOx/Al2O3 pre-coked with benzene. Acid sites
were classified according to their desorption temperature
into low energy acid sites (<250 °C), medium energy acid
Sample
CrIJIII)/CrIJVI)
Fresh CrOx/Al2O3
0.3 0.02
1.1 0.07
2.2 0.13
2.0 0.13
1.2 0.06
CrOx/Al2O3 – 600 °C – 1 h time-on-stream
CrOx/Al2O3 – 600 °C – 3 h time-on-stream
CrOx/Al2O3 – 600 °C – 6 h time-on-stream
CrOx/Al2O3 pre-coked with benzene
value of 1.2 which indicates that chromium is partly reduced
to CrIJIII). Given that dehydrogenation activity was observed
for the catalyst pre-coked with benzene from the beginning
of the reaction, the lower value of the CrIJIII)/CrIJVI) ratio of the
pre-coked catalyst with benzene (1.2) as compared to the
fresh catalyst after 3 h of ethylbenzene dehydrogenation (2.2)
suggests that the extent of chromium reduction is not
directly correlated with: i) the observed selectivity to styrene;
or ii) the length of the induction time before dehydrogena-
tion occurs. Therefore, while the presence of CrIJIII) is
required for dehydrogenation, the selective deactivation of
acid sites, which inhibits side-reactions, plays a critical role
in obtaining high selectivities to styrene.
In order to identify the type of coke laid down during pre-
coking which is beneficial for dehydrogenation activity, the
selectivity to styrene over the pre-coked catalysts was corre-
lated with the chemical nature of coke deposited after pre-
coking. Fig. 9 shows the selectivity to styrene after ethylben-
zene dehydrogenation over the pre-coked catalysts versus four
characteristics of the deposited coke; namely: the C/H ratio,
oxidation temperature, THz absorption coefficient at 1 THz
and percentage of reduction of the OH band. Fig. 9a–c show
that higher selectivities to styrene are obtained when highly
ordered coke deposits are laid down during pre-coking. Fur-
thermore, selectivity to dehydrogenation increases with
decreasing concentration of hydroxyl groups (Fig. 9d). These
data suggest that highly-ordered carbonaceous deposits are
more effective at deactivating hydroxyl groups and hence
increase dehydrogenation selectivity.
sites (250–400 °C) and high energy acid sites (>400 °C).46
A
higher percentage of low and high energy acid sites are poi-
soned after benzene pre-coking as compared to medium
strength acid sites. The low strength acid sites decreased
from 202 to 86 μmol NH3 per g after pre-coking, while the
high energy acid sites decreased from 101 to 42 μmol NH3
per g. In contrast, the medium strength acid sites were poi-
soned to a lesser extent, decreasing from 191 to 115 μmol
NH3 per g after pre-coking. Fig. S1† shows the XPS spectra of
the fresh CrOx/Al2O3 and the catalyst pre-coked with benzene.
The oxidation states of chromia were analysed by
deconvoluting the peaks present between 570 and 600 eV.
The CrIJIII)/CrIJVI) ratios for the fresh and pre-coked catalyst
with benzene are shown in Table 5. For the fresh catalyst, the
CrIJIII)/CrIJVI) ratio is low (0.3) suggesting that after calcination
the main species in the catalyst is CrIJVI) as has been previ-
ously reported by other authors.47,48 However, after pre-
coking with benzene, the CrIJIII)/CrIJVI) ratio increases to a
Table 4 Number of acid sites and acidity distribution of CrOx/Al2O3. a)
Fresh, b) after 3 h of ethylbenzene dehydrogenation (cracking period), c)
after 6 h of ethylbenzene dehydrogenation and d) after pre-coking with
benzene
Acidity
μmol NH3
Sample
distribution
per g
3.3.2 Characterisation of pre-coked catalysts after reac-
Fresh CrOx/Al2O3
Low strength
Medium strength 191
High strength
Total
202
7
6
4
tion.
Elemental analysis. The data shown in Tables 1 and
7 allow comparison between the carbon-containing com-
pounds on the pre-coked catalyst before (Table 1) and after
(Table 6) the dehydrogenation reaction. As shown in Table 6
the carbon percentage after ethylbenzene dehydrogenation
over benzene pre-coked CrOx/Al2O3 (13.5%) only increases
slightly relative to the catalysts pre-coked with benzene
(11.1%). This demonstrates that more than 97% of coke
arises from the pre-coking treatment, suggesting that ben-
zene adsorbs strongly over the acid sites or causes pore
blockage, hence limiting coking during ethylbenzene dehy-
drogenation. Also, the nature of coke does not change sub-
stantially given the very similar C/H values of the pre-coked
catalyst with benzene (13.9) and the catalyst after reaction
(15.0). In contrast, for the catalysts pre-coked with toluene
101
494 17
CrOx/Al2O3 after cracking period (3 h) Low strength
167
6
7
4
Medium strength 207
High strength
Total
70
444 17
CrOx/Al2O3 after dehydrogenation
period (6 h)
Low strength
Medium strength 121
High strength
Total
103
5
2
1
8
40
264
CrOx/Al2O3 pre-coked with benzene
Low strength
Medium strength 115
High strength
Total
86
3
4
1
8
42
243
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Fig. 9 Selectivity to styrene over pre-coked CrOx/Al2O3 versus a) C/H ratio, b) oxidation temperature, and c) THz absorption coefficient at 1 THz
and d) % reduction of OH band (3520 cm−1) from DRIFTS.
Table 6 Elemental analysis of pre-coked and reacted CrOx/Al2O3 after ethylbenzene dehydrogenation at 600 °C
Ethylbenzene dehydrogenation over
% C (wt%)
% H (wt%)
C/H mass ratio
% Ccoke
Fresh CrOx/Al2O3 – 6 h, 600 °C
6.3 0.2
13.5 0.1
15.4 0.1
21.4 0.1
14.9 0.2
0.6 0.1
0.9 0.1
0.9 0.1
0.9 0.1
0.5 0.1
10.5 0.9
15.0 0.8
17.1 0.2
23.8 1.2
29.8 2.0
91.3 0.7
93.8 0.3
94.5 0.1
95.9 0.3
96.8 0.2
CrOx/Al2O3 pre-coked with benzene – 6 h, 600 °C
CrOx/Al2O3 pre-coked with toluene – 6 h, 600 °C
CrOx/Al2O3 pre-coked with styrene – 6 h, 600 °C
CrOx/Al2O3 pre-coked with ethylene – 6 h, 600 °C
Table 7 Position of the Raman bands, D1/G and D1/D3 ratios of pre-coked catalysts after ethylbenzene dehydrogenation at 600 °C for 6 h time-on-
stream
Ethylbenzene dehydrogenation over CrOx/Al2O3 pre-coked with
Benzene
Toluene
Styrene
Ethylene
D4
D1
D3
G
Position (cm−1
FWHM (cm−1
Position (cm−1
FWHM (cm−1
Position (cm−1
FWHM (cm−1
Position (cm−1
FWHM (cm−1
)
)
)
)
1208
58
1223
97
1223
98
1205
76
)
1309
148
1486
167
1598
56
1328
159
1475
172
1595
72
1329
153
1466
167
1592
76
1318
158
1501
170
1593
66
)
)
)
Intensity ratios
D1/G
D1/D3
3.0 0.1
3.7 0.2
2.2 0.1
2.3 0.1
2.0 0.1
2.8 0.1
1.9 0.1
4.9 0.2
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
and styrene, the percentages of carbon after ethylbenzene
dehydrogenation approximately doubled as compared to the
pre-coked catalyst (from 8.9 to 15.4 for toluene; and 10.7 to
21.4, for styrene), indicating that these two precursor mole-
cules have a lower ability to reduce coke deposition. The per-
centage of carbon deposited over the catalyst pre-coked with
ethylene after reaction is substantially higher (14.9%) as com-
pared to the pre-coked catalyst (2.7%). These data show that
pre-coking with benzene is most effective at reducing coke
formation during ethylbenzene dehydrogenation, whilst eth-
ylene is the least effective pre-coking agent in this regard.
Temperature-programmed oxidation (TPO). The TPO pro-
files measured over the pre-coked catalysts after ethylbenzene
dehydrogenation (Fig. 10) were compared to those of the pre-
coked catalysts (Fig. 2). The peak intensities observed for the
pre-coked CrOx/Al2O3 are lower than those of the reacted cata-
lysts, consistent with the increase in the carbon percentage
after reaction, as seen in the elemental analysis data (Table 6).
In addition, there is a significant shift in the oxidation tem-
perature between the styrene pre-coked CrOx/Al2O3 (483 °C)
and the reacted catalyst after pre-coking with styrene
(532 °C), reflecting a change in coke structure from more dis-
ordered to more ordered carbon deposits. A smaller change in
the oxidation temperature is observed between the pre-coked
and reacted catalysts for the toluene and benzene precursors:
from 483 to 498 °C for the pre-coked catalyst with toluene and
after reaction, respectively; and from 495 to 505 °C in the case
of benzene. In the case of the catalyst pre-coked with ethylene
there is an increase in the oxidation temperature from 450 °C
after pre-coking to 550 °C after reaction.
Fig. 11 Fig. 7. THz spectra of CrOx/Al2O3 after (a) ethylbenzene
dehydrogenation at 600 °C, after pre-coking at 600 °C with (b) tolu-
ene (c) ethylene, (d) styrene, (e) benzene; and ethylbenzene dehydro-
genation over pre-coked CrOx/Al2O3 with f) benzene, g) toluene, (h)
styrene and (i) ethylene.
disordered in nature. Furthermore, THz absorption is gener-
ally lower for the pre-coked catalysts (Fig. 11b–d and h) than
for the reacted CrOx/Al2O3 (Fig. 11e–g and i) which demon-
strates the gradual development of the carbonaceous network
after reaction. This increase in the absorption coefficient is
much less significant for CrOx/Al2O3 pre-coked with benzene
(Fig. 11d) and the same catalyst after reaction (Fig. 11e), indi-
cating that after ethylbenzene dehydrogenation over the cata-
lyst pre-coked with benzene the degree of order remains prac-
tically unaltered, in agreement with TPO data.
Raman spectroscopy. The maximum value of the D1/D3
ratio (Table 7) is obtained for the pre-coked catalyst with eth-
ylene after ethylbenzene dehydrogenation (4.9) which indi-
cates that the coke present on this sample has the highest
degree of graphiticity. The catalysts pre-coked with styrene
and toluene after reaction exhibit the lowest values of D1/D3
ratios (2.8 and 2.3, respectively) which reflects the presence
of coke with lower degree of structural organisation. These
results confirm the trends indicated by THz-TDS (Fig. 11)
and TPO (Fig. 10).
THz-time domain spectroscopy (THz-TDS). THz-TDS pro-
vides a qualitative assessment of the degree of graphitic order
of coke formed over the pre-coked and reacted catalysts. The
THz spectra of the pre-coked and reacted catalysts are shown
in Fig. 11. In general, the values of the THz absorption coeffi-
cient are relatively low, indicating that the coke structures are
4. Discussion
4.1 Use of pre-coking to identify the main precursor molecule
of carbon deposits
In this investigation, tailored carbon deposition was
employed to elucidate the precursor molecule associated with
coke deposition during ethylbenzene dehydrogenation. The
identification of the precursor molecule was performed
through the comparison between the nature of coke depos-
ited over the fresh catalyst during ethylbenzene dehydrogena-
tion and over the pre-coked catalysts. The different character-
isation techniques provided insights into this. For instance,
the TPO profile of the catalyst pre-coked with ethylene
showed a similar oxidation temperature (463 °C) to the fresh
Fig. 10 Temperature-programmed oxidation of pre-coked catalysts
after ethylbenzene dehydrogenation at 600 °C (40 ml min−1, 5% O2/
He, 5 °C min−1).
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
catalyst after ethylbenzene dehydrogenation (455 °C) (Fig. 2).
In contrast, the catalysts pre-coked with aromatic molecules
exhibited higher oxidation temperatures than the fresh CrOx/
Al2O3 after ethylbenzene dehydrogenation, consistent with
their higher polyaromaticity (Fig. 2). The clear resemblance
between the Raman spectra of the catalyst pre-coked with
ethylene and the fresh catalyst after reaction further confirms
the similarity between coke formed during ethylbenzene
dehydrogenation and ethylene-derived coke (Fig. 4d). The
analysis of the nature of coke by DRIFTS spectroscopy yields
the same conclusion. Similar bands were observed in the
1700–1300 cm−1 region for the catalyst pre-coked with ethyl-
ene and the fresh catalyst after reaction (Fig. 5d and 6). Coke
analysis by DRIFTS also showed that ethylene-derived coke is
formed by molecular and discrete carbonaceous structures
similar, e.g., to p-xylene (Fig. 5d and 7b) suggesting that a sig-
nificant quantity of the coke deposits comprise para-
substituted aromatics. The catalysts pre-coked with aromatic
molecules exhibited structures similar to naphthenic com-
pounds such as 1-methyl naphthalene (Fig. 5a–c and 7b), sig-
nificantly different from those on the fresh catalyst after eth-
ylbenzene dehydrogenation (Fig. 6). The data presented
provide strong evidence that ethylene is the main coke pre-
cursor during ethylbenzene dehydrogenation and that coke
derived from ethylene is associated with a reduction in selec-
tivity to styrene during ethylbenzene dehydrogenation com-
pared to the performance of the fresh catalyst.
Based on the characterisation of the ethylene-derived coke
structures it is also possible to propose a mechanism for coke
formation on the catalyst surface. The formation of alkylated
benzene species such as p-xylene could proceed through a
three-step mechanism as previously described:49 i) ethylene
trimerization into hexene; ii) Diels–Alder reaction between
2,4-hexadiene and ethylene to yield 3,6-dimethylcyclohexene;
and iii) dehydrogenation of 3,6-dimethylcyclohexene to form
p-xylene. Alternatively, chromium oxide–silica–alumina cata-
lysts are known to catalyse the oligomerization of ethylene,50
which can be followed by the transfer dehydrogenation/
Diels–Alder reactions to produce p-xylene.51,52 These mecha-
nisms provide potential routes from the ethylene precursor
to the coke structures identified on both the ethylene pre-
coked catalyst and the non-pre-coked catalyst after reaction.
In contrast, for the catalysts pre-coked with aromatic mole-
cules the coke deposits have been observed to differ from
those formed on the fresh catalyst after ethylbenzene dehy-
drogenation and instead show similar structures to naph-
thenic compounds. These species may form from either the
direct condensation of benzene rings of a hydrogen–C2H2
addition pathway.53
reactions catalysed by acid sites: i.e. cracking to benzene and
coke, hydrogenolysis of ethylbenzene into toluene and
dealkylation of ethylbenzene and toluene. As seen from the
reactor data, pre-coking with benzene was the most effective
in enhancing selectivity to benzene, but this pre-coking treat-
ment was only slightly better than when pre-coking with sty-
rene. The data shown in Fig. 9, show that when pre-coking
increases selectivity to styrene, the effect of pre-coking is to
produce coke with relatively higher C/H ratio, and which is
more structurally ordered. Furthermore, in these cases a
greater reduction in hydroxyl density on the catalyst surface is
observed. Whilst there is little difference between the perfor-
mance of styrene and benzene as pre-coking agents, the data
shown in Fig. 9b and c might suggest that benzene produces
a more ordered, thermally stable coke structure than styrene
and that this has a greater influence on selectivity than the
absolute reduction in the number of hydroxyl groups present.
5. Conclusions
This investigation has demonstrated that pre-coking is a use-
ful strategy to identify the main precursor molecule of coke
in ethylbenzene dehydrogenation over CrOx/Al2O3 catalysts.
The characterisation of pre-coked catalysts by elemental anal-
ysis, TPO, Raman spectroscopy and DRIFTS showed that eth-
ylene is the main coke precursor in ethylbenzene dehydroge-
nation over CrOx/Al2O3 catalysts and that the nature of coke
is para-substituted aromatic species. Furthermore, the pre-
coking treatment can be used to inhibit undesired reactions
(cracking, hydrogenolysis, dealkylation) and enhance selectiv-
ity to styrene. Pre-coking with benzene was the most effective
pre-treatment, with styrene demonstrating similar efficacy.
This is a result of the removal of hydroxyl groups due to the
presence of significant coke deposits, and the associated
change in chromium oxidation state to CrIJIII), the active spe-
cies for dehydrogenation. Pre-coking with aromatic molecules
led to the deposition of naphthenic coke structures which are
beneficial for dehydrogenation since they result in the deacti-
vation of acid sites and Cr–VIIJO) sites. Conversely, coke
formed from ethylene was found to be detrimental for dehy-
drogenation activity as shown by the decrease in styrene
selectivity as compared to the fresh catalyst.
Acknowledgements
The authors express their appreciation to the support from
the Ministry of Higher Education, Saudi Arabia, in establish-
ment of the Center of Research Excellence in Petroleum
Refining & Petrochemicals at King Fahd University of Petro-
leum & Minerals (KFUPM).
4.2 The effect of pre-coking on suppressing undesired
reactions
References
The catalytic evaluation of the pre-coked catalysts for the
dehydrogenation of ethylbenzene (Fig. 8) showed that pre-
coking with aromatic molecules led to the inhibition of side-
1 M. Guisnet and P. Magnoux, Appl. Catal., 1989, 54, 1–27.
2 E. G. Derouane, Catalysis by Acids and Bases, Elsevier,
Amsterdam, 1985, vol. 20.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
3 P. Marécot, A. Akhachane and J. Barbier, Catal. Lett.,
1996, 36, 37–39.
30 M. Cherian, M. S. Rao, A. M. Hirt, I. E. Wachs and G. Deo,
J. Catal., 2002, 211, 482–495.
4 C. Hettige, K. R. R. Mahanama and D. P. Dissanayake,
Chemosphere, 2001, 43, 1070–1083.
5 W. W. Kaeding, C. Chu, L. B. Young, B. Weinstein and S. A.
Buttery, J. Catal., 1981, 67, 159–174.
6 J. Houzvicka, R. Pestman and V. Ponec, Catal. Lett.,
1995, 30, 289–296.
7 D. Chen, H. P. Rebo, K. Moljord and A. Holmen, Stud. Surf.
Sci. Catal., 1997, 111, 159–166.
8 J. McGregor, Z. Huang, E. P. J. Parrott, J. A. Zeitler, K. L.
Nguyen, J. M. Rawson, A. Carley, T. W. Hansen, J.-P.
Tessonnier, D. S. Su, D. Teschner, E. M. Vass, A. Knop-Gericke,
R. Schlögl and L. F. Gladden, J. Catal., 2010, 269, 329–339.
9 S. J. Thomson and G. Webb, J. Chem. Soc., Chem. Commun.,
1976, 526.
10 I. M. Dahl and S. Kolboe, Catal. Lett., 1993, 20, 329–336.
11 F. Bauer, W.-H. Chen, H. Ernst, S.-J. Huang, A. Freyer and
S.-B. Liu, Microporous Mesoporous Mater., 2004, 72, 81–89.
12 M. Niwa, H. Ito, S. Kato, T. Hattori and Y. Murakami,
J. Chem. Soc., Chem. Commun., 1982, 819–820.
13 H. Begum, N. Katada and M. Niwa, Microporous Mesoporous
Mater., 2001, 46, 13–21.
31 P. C. Upadhya, K. L. Nguyen, Y. C. Shen, J. Obradovic, K.
Fukushige, R. Griffiths, L. F. Gladden, A. G. Davies and E. H.
Linfield, Spectrosc. Lett., 2006, 39, 215–224.
32 E. P. J. Parrott, J. A. Zeitler, J. McGregor, S.-P. Oei, H. E.
Unalan, S.-C. Tan, W. I. Milne, J.-P. Tessonnier, R. Schlögl
and L. F. Gladden, J. Phys. Chem. C, 2009, 113, 10554–10559.
33 S. Gomez, L. McMillan, J. McGregor, J. A. Zeitler, N. Al-
Yassir, S. Al-Khattaf and L. F. Gladden, Catal. Sci. Technol.,
2015, 5, 3782–3797.
34 O. Bayraktar and E. L. Kugler, Appl. Catal., A, 2002, 233,
197–213.
35 F. Tuinstra and J. L. Koenig, J. Chem. Phys., 1970, 53, 1126–1130.
36 O. Beyssac, B. Goffe, J.-P. Petitet, E. Froigneux, M. Moreau
and J.-N. Rouzaud, Spectrochim. Acta, Part A, 2003, 59A,
2267–2276.
37 A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner and R.
Pöschl, Carbon, 2005, 43, 1731–1742.
38 W. Shen, Z. Li and Y. Liu, Recent Pat. Chem. Eng., 2008, 1,
27–40.
39 J. Coates, in Encyclopedia of Analytical Chemistry, ed. R. A.
Meyers, John Wiley & Sons Ltd., Chichester, 2000, pp. 10815–
10837.
40 P. A. Schroeder, in Teaching Clay Science, ed. S. Guggenheim
and A. Rule, Clay Minerals Society Workshop Series, Aurora,
2002, vol. 11, pp. 181–202.
14 W. O. Haag, D. H. Olson and F. Gorra, US Pat., no. 4 097
543, 1982, assigned to Mobil Oil Corporation.
15 F. Gorra, L. L. Breckenridge, W. M. Guy and R. A. Sailor, Oil
Gas J., 1992, 90, 60–67.
16 G. Yaluris, R. B. Larson, J. M. Kobe, M. R. González, K. B.
Fogash and J. A. Dumesic, J. Catal., 1996, 158, 336–342.
17 J.-H. Kim, Y. Sugi, T. Matsuzaki, T. Hanaoka, Y. Kubota, X.
Tu, M. Matsumoto, S. Nakata, A. Kato, G. Seo and C. Pak,
Appl. Catal., A, 1995, 131, 15–32.
18 R. Fiedorow, R. Frański, A. Krawczyk and S. Beszterda,
J. Phys. Chem. Solids, 2004, 65, 627–632.
19 J. McGregor and L. F. Gladden, Appl. Catal., A, 2008, 345,
51–57.
41 A. A. Davydov, in Infrared Spectroscopy of Adsorbed Species on
the Surface of Transition Metal Oxides, ed. C. H. Rochester,
John Wiley & Sons, Inc., 1984, pp. 6–62.
42 R. B. Fahim, M. I. Zaki, N. E. Fouad, M. Abdel-Khali and N.
Sheppard, Appl. Catal., A, 2004, 265, 229–235.
43 T. J. Bandosz, in Surface Chemistry of carbon materials, ed. P.
Serp and J. L. Figueiredo, John Wiley & Sons, Inc., New
Jersey, NJ, USA, 2009, pp. 45–92.
44 S. M. K. Airaksinen, A. O. I. Krause, J. Sainio, J. Lahtinen,
K.-J. Chao, M. O. Guerrero-Pérez and M. A. Bañares, Phys.
Chem. Chem. Phys., 2003, 5, 4371–4377.
20 S. Mandal, A. K. Das and S. Ghosh, Ind. Eng. Chem. Res.,
1993, 32, 1018–1023.
21 C. K. Lee, L. F. Gladden and P. J. Barrie, Appl. Catal., A,
2004, 274, 269–274.
45 K. Ramasesha, L. De Marco, A. Mandal and A. Tokmakoff,
Nat. Chem., 2013, 5, 935–940.
22 C. Venkat, in 19th Dewitt Petrochemical Review, 1994, pp.
A1–A11.
23 C. D. Low, R. J. Lawson, P. J. Kuchar and G. L. Gray, US Pat.,
no. 5 321 184, 1994, assigned to UOP.
46 M. Muhler, J. Schütze, M. Wesemann, T. Rayment, A. Dent,
R. Schlögl and G. Ertl, J. Catal., 1990, 126, 339–360.
47 O. F. Gorriz, V. Cortes Corberan and J. L. G. Fierro, Ind. Eng.
Chem. Res., 1992, 31, 2670–2674.
24 W. O. Haag and D. H. Olson, US Pat., no. 4 117 026, 1978, to
Mobil Oil Corporation.
25 W. O. Haag, D. H. Olson and P. G. Rodewald, US Pat., no. 4
508 836, 1985, to Mobil Oil Corporation.
26 S. Laforge, D. Martin and M. Guisnet, Microporous
Mesoporous Mater., 2004, 67, 235–244.
27 F. Bauer, W. H. Chen, E. Bilz, A. Freyer, V. Sauerland and
S. B. Liu, J. Catal., 2007, 251, 258–270.
48 F. Cavani, M. Koutyrev, F. Trifirò, A. Bartolini, D. Ghisletti,
R. Iezzi, A. Santucci and G. Del Piero, J. Catal., 1996, 158,
236–250.
49 T. W. Lyons, D. Guironnet, M. Findlater and M. Brookhart,
J. Am. Chem. Soc., 2012, 134, 15708–15711.
50 A. Chakrabarti and I. E. Wachs, Catal. Lett., 2014, 145,
985–994.
51 H. S. Bloch, US Pat., 2920114, 1960.
28 B. M. Weckhuysen and R. A. Schoonheydt, Catal. Today,
1999, 51, 223–232.
52 W. K. Reagen, T. M. Pettijohn, J. W. Freeman and E. A.
Benham, US Pat., 5786431, 1998.
29 S. De Rossi, M. P. Casaletto, G. Ferraris, A. Cimino and G.
Minelli, Appl. Catal., A, 1998, 167, 257–270.
53 M. Frenklach, W. C. Gardiner, S. E. Stein, D. W. Clary and T.
Yuan, Combust. Sci. Technol., 1986, 50, 79–115.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015