Spectroscopic characterization-catalytic activity correlation of molybdena based catalysts

Article abstract of DOI:10.1016/j.molcata.2012.11.013

Titania-supported molybdena catalyst was prepared by calcination at 773 K of ammonium heptamolybdate-impregnated TiO₂ pellets with the equivalent of 5 monolayers of MoO₃. Then, potassium-modified versions of the catalyst were prepared at 0.5-5 wt%-K. The modified and unmodified catalysts were characterized, before and after H₂-reduction at 673 K, by means of ISS, XPS, UPS and FT-IR spectroscopy techniques. Results obtained enabled revealing the presence of density of states (DOS) at the Fermi level and, hence, metallic properties related to the deformed rutile structure of MoO₂. On the other hand, Br?nsted (MoOH) acidic function was found to form on the surface of the MoO₂ structure, thereby producing a metal-acid (bifunctional) MoO_2-x(OH)_y catalyst. The presence of Br?nsted acid MoOH groups was evidenced by in situ IR spectra of adsorbed pyridine molecules. On the other hand, the K-modification was found to enhance the molybdena dispersity and reducibility, but to suppress the surface acidity by transforming the bifunctional MoO_2-x(OH) _y species into the monofunctional MoO_2-x(OK)_y species. Subsequent catalytic measurements employing hydrotreating reactions of 1-hexene and n-hexane could confirm the exposure of bifunctional or monofunctional sites on the catalysts tested. The combination of research methods adopted in the present investigation has shown competency in unravelling the nature of active sites on heterogeneous catalysts.

Full text of DOI:10.1016/j.molcata.2012.11.013

Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Spectroscopic characterization–catalytic activity correlation of molybdena
based catalysts
H. Al-Kandarib, A.M. Mohameda, S. Al-Kandaria, F. Al-Kharafia, G.A. Mekhemerc, M.I. Zakic, A. Katriba,∗
a Chemistry Department, Kuwait University, P.O. Box 5969 Safat 13060, Kuwait
b College of Health Sciences, PAAET, P.O. Box 1428, Faiha 72853, Kuwait
c Chemistry Department, Faculty of Science, Minia University, El-Minia 61519, Egypt
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 9 May 2012
Received in revised form
13 November 2012
Accepted 18 November 2012
Available online 26 November 2012
Titania-supported molybdena catalyst was prepared by calcination at 773 K of ammonium
heptamolybdate-impregnated TiO2 pellets with the equivalent of
5 monolayers of MoO3. Then,
potassium-modified versions of the catalyst were prepared at 0.5–5 wt%-K. The modified and unmo-
dified catalysts were characterized, before and after H2-reduction at 673 K, by means of ISS, XPS, UPS
and FT-IR spectroscopy techniques. Results obtained enabled revealing the presence of density of states
(DOS) at the Fermi level and, hence, metallic properties related to the deformed rutile structure of
MoO2. On the other hand, Brønsted (Mo OH) acidic function was found to form on the surface of the
MoO2 structure, thereby producing a metal-acid (bifunctional) MoO2−x(OH)y catalyst. The presence of
Brønsted acid Mo OH groups was evidenced by in situ IR spectra of adsorbed pyridine molecules. On
the other hand, the K-modification was found to enhance the molybdena dispersity and reducibility, but
to suppress the surface acidity by transforming the bifunctional MoO2−x(OH)y species into the mono-
functional MoO2−x(OK)y species. Subsequent catalytic measurements employing hydrotreating reactions
of 1-hexene and n-hexane could confirm the exposure of bifunctional or monofunctional sites on the
catalysts tested. The combination of research methods adopted in the present investigation has shown
competency in unravelling the nature of active sites on heterogeneous catalysts.
Keywords:
Catalytic properties of MoO2−x(OH)y
MoO2−x(OK)y
XPS
UPS
ISS
FT-IR
MoO3
TiO2
Bifunctional catalyst
1-Hexene
n-Hexane
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
above Mo atoms placed along the c-axis of the deformed rutile
structure of this phase [4,5]. Formation of Brønsted acid Mo OH
Identification of the active sites in heterogeneous catalytic
processes constitutes a major challenge. The complexity of the sit-
uation results from the fact that catalytic active sites are formed, in
situ, under specific activation condition(s). Moreover, the chemical
structure of the catalyst is composed, quite often, of several ele-
ments having variable oxidation states. Also, the catalytic reaction
takes place on the outermost surface layer, which is not necessar-
ily similar to the bulk structure. In general, catalysts are classified
as metallic, acidic or bifunctional (metal-acid). In this respect, we
have developed bifunctional catalysts based on Mo and W oxides
[1–3]. The catalytic active sites are formed throughout the hydro-
gen reduction of MO3 (M = Mo or W) at temperatures up to 673 K in
the case of MoO3. Initial activity is triggered following the reduction
of MoO3 to MoO2 phase, which has metallic properties. The metallic
character of MoO2, measured as density of state (DOS) structure at
the Fermi-level, is rationalized in terms of delocalized ␲-electrons
sites on the surface is effected following dissociative chemisorption
of hydrogen molecules during the reduction process. Consequently,
the reduction of MoO3/TiO2 leads to the formation of metallic-and-
acidic (bifunctional) structure along the Mo Mo atoms, which is
depicted as being MoO2−x(OH)y/TiO2. TiO2 is used as a support
material for the catalysts thus generated in order to improve their
surface accessibility and mechanical strength. The most adequate
XPS–UPS, ISS, FT-IR, HRTEM and XRD spectroscopic techniques are
implemented to identify the occurrence of the metallic and acidic
functions on this catalyst system.
On the other hand, well-established catalytic isomerization
reactions undertaking the conversion of light linear C5–C7 hydro-
carbon compounds into more valuable branched isomers are
facilitated via the bifunctional mechanism [6]. These reactions take
place in a concerted mechanism requiring the availability of both
metallic and acidic functions. In contrast, hydrogenation of alkenes
to alkane compounds requires the availability of metallic function
only. Therefore, the present investigation was designed to apply
above-mentioned spectroscopic techniques at similar experimen-
tal conditions, at which the catalyst conduct in the hydrogenation
∗
Corresponding author. Tel.: +965 24985582; fax: +965 24816482.
E-mail address: ali.katrib@ku.edu.kw (A. Katrib).
1381-1169/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2012.11.013

2
H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
of 1-hexene and the isomerization of n-hexane is activated. The
prime objective of the investigation is to correlate the catalyst
surface structure, activity and selectivity, in hopes of a better under-
standing of the nature and type of the active site(s) in view of
its optimization. Moreover, conclusions drawn are consolidated by
spectroscopic observation of changes conceded by the sought cor-
relations as a result of potassium-modification of the catalyst [7].
The exposure of different acid sites on TiO2-supported molybde-
num oxides has been reported by several authors and research
works [8,9]. Miyata et al. [8] did probe the presence of different
Lewis and Brønsted acid sites on TiO2-supported molybdena via
IR spectra of adsorbed ammonia or pyridine. It is postulated that
relatively strong Mon+ Lewis acid sites are converted to Brønsted
type acid sites following the exposure to water vapor atmosphere.
On the other hand, pure titania with or without water pretreatment
did not show any Brønsted acidity by the adsorption of ammonia or
pyridine [8]. It is noteworthy that the required bifunctional (metal-
acidic) properties for the hydroisomerization of alkane molecules,
as performed by MoO2−x(OH)y/TiO2, takes place only after the
formation of the metallic function in MoO2, following hydrogen
treatment of MoO3/TiO2 at higher temperatures than 573 K. Such
catalytic activity has not been observed using helium gas for exam-
ple. In other words, the metallic function seems, in this case, to
be the initial state for the dissociation of hydrogen molecules into
active hydrogen atoms, which are then bonded to surface oxygen
atoms of MoO2 to generate Mo OH Brønsted acid sites. Just to
mention that water pretreatment is not required in this case. Nev-
ertheless, the formation of different Lewis and Brønsted acid sites
on MoO3/TiO2, before and after hydrogen (and helium) treatments
at different temperatures has been further probed in the present
investigation, employing a combination of catalytic test reactions of
1-hexene and n-hexane, and surface characterization by XPS–UPS,
ISS and in situ FTIR techniques. To the best of our knowledge, this
combination of research techniques has not been adopted in such
a single literature report.
source was monochromaticof AlK␣ operatingat the powerof 300 W
(15 kV, 20 mA). Ultraviolet Photoelectron Spectroscopy (UPS) He(I)
resonance 584 A radiation of 21.217 eV was employed for the
˚
Valence Band (VB) energy region measurements. Vacuum in the
analysis chamber was better than 7 × 10−9 mbar during all mea-
surements. The in situ reduction was carried out in a high-pressure
gas cell housed in the preparation chamber. Binding energies
were based on the carbon contamination C1s at 284.8 eV within
an experimental error of 0.2 eV. Ion scattering spectroscopy
(ISS) measurements were performed on the same test sample
using He+ with a kinetic energy of 1 KeV. Ex and in situ Fourier-
transform Infrared (IR) spectra were recorded at 4000–400 cm−1
by averaging 200 scans at the resolution of 4 cm−1, using a Gen-
esis II Thermo Mattson FT-IR spectrophotomer (USA) powered
with a WinFirst Lite software (Mattson Corp.) for data acquisi-
tion and handling. The ex situ spectra were taken of lightly loaded
(<1 wt %) KBr-supported discs of the test materials. Whereas,
the in situ spectra were taken of thin (30 mg/cm2), but intact,
self-supporting wafers of the test samples mounted inside a spe-
cially designed IR-cell equipped with CaF2 windows, before and
after adsorption of pyridine (Py) molecules. Difference spectra
of irreversibly adsorbed Py species were obtained by absorption
subtraction of background spectra, facilitated by the installed soft-
ware.
2.3. Catalyst tests
Catalytic hydrotreating reactions of 1-hexene and n-hexane
were studied on the test catalysts by introducing the reactants as
pluses of 5 ␮L and time on stream modes over a fixed-bed quartz
reactor under atmospheric hydrogen pressure. A continuous H2
flow of 40 cm3/min was allowed through 500 mg of the catalyst,
which contains 65 mg of Mo. The reaction mixture was separated
and analyzed with an on-line gas chromatograph Chemito, India
1000 equipped with a Petrocol-DH column and a flame ionization
detector.
2. Experimental
2.1. Catalyst preparation
3. Results and discussion
The supported MoO3/TiO2 catalyst system was prepared by cal-
cination at 773 K for 12 h of wet impregnated titania pellets with
the equivalent of 5 monolayers of MoO3 of ammonium hepta-
molybdate ((NH4)6Mo7O24·4H2O, 99.9% pure product of STREM).
The impregnation was carried out according to Pines et al. [10]
and the titania support was Degussa P25 (75% anatase) with pore
volume of 0.5 cm3/g and BET surface area of 50 5 m2/g. It is
worth mentioning that, the loaded ammonium heptamolybdate
was quantitatively decomposed into MoO3 species during the cal-
cination process. For simplicity, the calcination product is denoted
MoTi. The set of potassium-modified MoTi catalysts was prepared
by post-impregnation of the dried heptamolybdate-impregnated
TiO2 pellets with different concentrations of KNO3 prior to the
eventual calcination process at 773 K. The potassium concentra-
tions used were calculated to amount to 0.5, 1, 2.5 and 5% of the
molybdena content of the catalyst. The K-modified catalysts thus
obtained are denoted x-KMoTi, where x stands for the %-K. In situ
reduction of portions of the pure and K-modified catalysts thus
obtained was carried out by heating at 673 K for 12 h of the cat-
alyst in a flow of 40 cm3 H2/min. The hydrogen gas was a 99.9%
pure product of KOAC (Kuwait).
3.1. Catalyst characterization
3.1.1. ISS spectra
The impregnation of the equivalent of 5 monolayers of MoO3
on titania pellets seems to have been inhomogeneous as could be
inferred from the display of O, Mo and Ti peaks in the ISS spectrum
obtained for the calcined MoTi sample (Fig. 1a). This implies that
MoO3 deposition is, most probably, in form of aggregates and not
uniformly spread on the support surface [11]. Extensive research
work has been reported in the literature concerning the disper-
sion of Mo oxide on different types of titania [9,12,13] by oxygen
chemisorption. Chary et al. [9] did observe that in the case of loading
below 6 wt%-MoO3, molybdenum oxide particles disperse better on
TiO2 (anatase) than on TiO2 (rutile). Moreover, the dispersed Mo
oxide is present in the form of an isolated tetrahedral coordination
environment on TiO2 (rutile), whereas the dispersion is in the form
of polymeric species on TiO2 (anatase) [12]. The different states of
dispersion are attributed to the different dehydration orders of OH
groups of the molybdena and surface OH groups of the rutile and
anatase.
It is important to note at this point, that BET-surface area
measured for the equivalent of five monolayers of MoO3/TiO2
was 34 m2/g before reduction with hydrogen, which increased
up to 48 m2/g after reduction at 673 K for 12 h. This reduction by
hydrogen corresponds to the activation procedure of the catalyst.
The surface area of the catalyst is thus still less than that (50 m2/g)
2.2. Catalyst characterization
X-ray photoelectron spectroscopy (XPS) was conducted using
a Thermo Scientific ESCALAB-250Xi spectrometer. The radiation

H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
3
(c)
O
K
Ti
Mo
(d)
(c)
(b)
(a)
250
400
550
700
850
1000
(b)
(a)
Kinetic Energy (eV)
Fig. 1. ISS spectra obtained for the calcined MoTi (a), 1%-KMoTi (b), an 5%-KMoTi
(c) catalyst samples.
of the pure TiO2 support. These results may explain the reason why
titanium atoms are part of the outermost surface area as measured
by ISS. The exact reason(s) of the absence of complete coverage
of the TiO2 by the equivalent five monolayer of MoO3 could be
attributed to the formation of a “para-crystalline MoO3” phase that
resembles bulk MoO3 and takes the form of needle-like crystallites
growing away from the support surface as postulated by Bond
et al. [14]. Apparently, the interaction between molybdenum oxide
and titania surface is less than in the case of in situ evaporated
Mo atoms over TiO2. In the latter case, one monolayer equivalent
of Mo(0) was sufficient to completely cover the titania surface as
observed by Hercules et al. using ISS technique [11].
454
458
462
466
470
Binding Energy (eV)
Fig. 2. XPS Ti (2p) spectra obtained for TiO2 before reduction (a), after reduction in:
MoTi (b), 1%-KMoTi (c) and 5%-KMoTi (d).
for MoO2 and 227.7 and 230,85 eV for Mo(0) [14]. On the other
hand, Ti(2p3/2,1/2) binding energy values are observed at 458.5 and
The display of an extra peak assignable to potassium in the spec-
tra given rise by 1%- and 5%-KMoTi indicates that K atoms are
also part of the outermost surface composition of the test sam-
ples (Fig. 1b and c). Subsequent in situ exposure of the test samples
to hydrogen at 673 K for 12 h did not cause any change to the ISS
spectra shown in Fig. 1. These results may reveal that the outermost
surface of the test catalysts is composed possibly of one or more dif-
ferent chemical entities containing Mo, O, Ti and K. Accordingly,
identification of the possible chemical species of the outermost
surface before and after reduction using XPS–UPS and FT-IR tech-
niques is essential in order to elucidate the sought catalyst surface
structure-activity relationship(s).
464.2 eV for TiO2 [15].
XPS Ti(2p) spectra obtained for pure TiO2 support, before and
after H2-reduction at 673 K for 12 h, are quite similar in monitoring
two well-defined spectral lines at 458.5 and 464.2 eV (Fig. 2a). The
two peaks are thus due, respectively, to Ti(2p3/2) and Ti(2p1/2) spin-
orbit components of TiO2 [15]. Compatibly, in situ IR spectra taken
of the TiO2, before and after reduction, are almost identical (Fig. S1,
in the linked Supporting Information File). These results are consis-
tent with literature data [16] showing that H2-reduction of TiO2 to
lower valance state(s), such as Ti2O3, does commence at tempera-
tures no less than 1273 K. This may be correlated with our previous
observation [17] demonstrating that TiO2 has no detectable cat-
alytic activity in hydrotreating reactions of alkanes. On the other
hand, The Ti(2p) peak broadening and binding energies remained
invariant in XPS spectra obtained for MoTi and x-KMoTi before and
after reduction (Fig. 2).
XPS of the Mo(3d) energy region spectrum obtained for calcined
MoTi (Fig. 3a) monitor two well-defined spectral lines at 232.65
and 235.85 eV characteristics of Mo(3d5/2,3/2) spin-orbit compo-
nents of MoO3 [14]. In situ sample exposure to hydrogen at 673 K
for 12 h results in partial reduction of MoO3 to MoO2 [3] as could be
observed from the peaks resolved at 232.5 and 229.6 eV (Fig. 3b).
However, the reduction process seems to be incomplete, because
two peaks are additionally resolved at 233.9 and 231.7 eV, which
3.1.2. XPS–UPS spectra
XPS is the adequate surface analytical technique to gain infor-
mation about the nature of chemical species established on the
test surface. Monitoring changes conceded by the oxidation state
of the ISS-identified Mo, O, Ti and K atoms (Fig. 1), before and
after hydrogen-reduction is essential to define the surface chemical
composition(s). Moreover, the presence of metallic function(s) is
manifested by density of states (DOS) structure at the Fermi-level
in the XPS–UPS valence band (VB) energy region. Reference XPS
binding energies for the Mo (3d5/2
, 3/2) are: 232.65 and 235.85 eV
for MoO3, 231.7 and 234.9 eV for Mo2O5, 229.1 and 232.3 eV

4
H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
(d)
(c)
(b)
(a)
0
2
4
6
8
10
12
Binding Energy (eV)
Fig. 4. XPS VB spectra obtained for calcined MoTi (a), reduced MoTi (b), 1%-KMoTi
(c) and 5%-KMoTi (d).
(c)
Fig. 3. XPS Mo(3d) spectra obtained for calcined MoTi (a), reduced MoTi (b), 1%-
KMoTi (c) and 5%-KMoTi (d).
(b)
(a)
are assignable to the occurrence of Mo2O5-like species. It is most
likely that, the Mo2O5 is formed as an intermediate surface species
in the reduction course of MoO3 to MoO2 [18,19]. The formation
of MoO2 as a result of this reduction process is also observed in
the XRD of the reduced sample. It is interesting to note that for
the same initial amount of MoO3 on TiO2, the extent of reduc-
tion to MoO2 is increased with the amount of potassium additive
(0.5–5%; Fig. 3c and d). Apparently, substantial reduction of MoO3
to MoO2 is shown to take place following the modification with
5%-K (Fig. 3d). XPS K(2p) spectra obtained for the x-KMoTi catalyst
samples (x = 0.5–5%) (not shown), before and after reduction, were
similar in monitoring two peaks at 292.9 and 295.7 eV due, respec-
tively, to K2p3/2 and K2p1/2 spin-orbit components. These peaks
account for K O bond formation [20] in surface species such as
-1 -0.5
0
0.5
1
1.5
2
2.5
Binding Energy (eV)
Fig. 5. High-resolution UPS spectra obtained over the VB energy regime for calcined
MoTi (a), and reduced MoTi (b), and 2.5%-KMoTi (c).
K
O Mo. The possible formation of isolated molecular species in

H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
5
Fig. 6. In situ IR ␯CCN difference spectra obtained for irreversibly adsorbed Py on calcined samples of the support (Ti), and the catalysts MoTi and x-KMoTi (x = 1–5%-K) at RT
and 200 ◦C [HPy = hydrogen-bonded Py; LPy = Py bound to Lewis acid sites; BPy = Py bound to Bronsted acid sites].
this case could result in better dispersion of the modified KMoTi
structure. Consequently, a better reduction of Mo(VI) on the x-
KMoTi samples (Fig. 3c and d) is expected as compared to the case
on the unmodified MoTi sample (Fig. 3b).
3.1.3. FT-IR spectra
Several ex and in situ IR spectra were taken of the various test
catalyst samples before and after irreversible adsorption of pyri-
dine (Py) molecules. The obtained ex situ spectra are compared in
Fig. S2. Based on relevant spectral data reported by Stencel [21],
Massoth [22], and Gadsden [23], the set of ex situ spectra displayed
in Fig. S2 may disclose that the molybdena load of the calcined
MoTi is partly aggregated into MoO3 particles, whereas the rest
is dispersed in the form of 2-dimensional structures constituting
molybdate monomers and polymers. The modification with potas-
sium additives is shown to increase the fraction of the monomeric
species. Compatibly, Martin et al. [7] have found potassium addi-
tives to interact with MoO3/TiO2 at 773 K leading to the formation
of monomeric species of K2MoO4 and polymeric K2Mo4O13. The
ex situ spectra (Fig. S2) show, moreover, that MoO3 is more easily
reduced when dispersed on the TiO2 support than when it is self-
supported bulk phase under the conditions applied, and that the
potassium modification seems to further improve its reducibility.
In situ IR spectra were taken of the calcined and reduced pure
TiO2 support, MoTi, KMoTi and unsupported MoO3 before and
XPS of the VB energy region obtained for the calcined, K-
modified and unmodified catalyst samples do not show any DOS
structure at the Fermi level (Fig. 4a). However, a pronounced
DOS structure, characteristic of the metallic properties of MoO2,
is observed in the spectra obtained following the samples in situ
exposure to hydrogen at 673 K for 12 h (Fig. 4b–d). These spectra
reveal, moreover, that the intensity of the DOS peak increases with
the potassium proportion (Fig. 4b–d). This correlation is most prob-
ably related to the concomitant increase of the MoO2 proportion
(vide supra).
High-resolution UPS measurements facilitated resolving two
maxima at 0.4 and 1.4 eV for the observed DOS peak (Fig. 5).
These energy values are assignable, respectively, to delo-
calized ␲-electrons (metallic function) above the Mo Mo
atoms and ␴-bonds between Mo Mo in the MoO2 structure
[4,5].

6
H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
Conversion
nC6
100
80
60
40
20
0
100
Conversion
nC6
80
iC6
60
40
20
0
iC4-iC5
iC6
iC4-iC5
C1-C5
700
C1-C5
0
1
2
3
4
5
450
500
550
600
650
Concentration of K (%)
Temperature (K)
Fig. 8. Product distributions of 1-hexene catalytic reactions at 623 K on reduced
x-KMoTi as a function of the amount of the potassium additive (=x).
Fig. 7. Products distribution of 1-hexene on MoTi at different reaction temperatures.
the complete conversion of the reactant encompasses 84.2% hydro-
genation to n-hexane and 15.8% to saturated branched products,
leading essentially to the formation of 2- and 3-methyl pentanes
(MPs) to the thermodynamically controlled ratio (i.e., 2-MP/3-
MP = 1.5). Fig. 7 shows, moreover, that at 623 K the conversion of
1-hexene gone to 59.9% MPs, 21.9% n-hexane, 12.9% hydrocracking
C1–C6, and 5.3% iC4-iC5 products. At higher reaction tempera-
tures, relatively higher amounts of the hydrocracking products
are favored. This may be ascribed to thermodynamic and kinetic
considerations. It is worth noting that, at near ambient reaction
temperatures only the hydrogenation of 1-hexene to n-hexane
takes place.
The above results may be explained as follows. The hydrogena-
tion of 1-hexene to n-hexane, which is catalyzed by the metallic
function of MoO2−x(OH)y at relatively low reaction temperatures,
seems to adopt the easiest surface reaction pathway. On the other
hand, the 1-hexene branching to 2- and 3-MP is suggested to
be undertaken via the following concerted multi-step process: (i)
hydrogenation of 1-hexene to n-hexane, (ii) dehydrogenation of the
formed n-hexane to 1-hexene [both step-(i) and -(ii)] are catalyzed
by the metallic function of MoO2−x(OH)y, (iii) isomerzation of 1-
hexene to i-hexene via carbenium ion mechanism facilitated by the
Brønsted Mo OH acidic function of MoO2−x(OH)y, and (iv) hydro-
genation of the i-hexene isomers to i-C6 saturated isomers, which is
catalyzed by the metallic function of the bifunctional active surface
species. The catalytic steps (ii)–(iv) are known as the bifunctional
mechanism of the isomerization reactions of alkane compounds
[6].
Different catalytic behaviors were exhibited towards 1-hexene
following the addition of increasing amounts of K to MoTi (i.e.,
the formation of x-KMoTi catalysts) and the subjection to the
same hydrogen reduction process at 673 K for 12 h. The relative
effect of the K proportion (0.5–5%-K) is visualized by the 1-hexene
conversion product distributions at 623 K as displayed in Fig. 8. Irre-
spective of the amount of the K-additive, a complete conversion of
1-hexene molecules is observed. However, the product distribu-
tion is varied with the K-proportion. For instance, 31.0% selectivity
to MPs formation and 62.0% hydrogenation to n-hexane are shown
to occur on the 0.5%-KMoTi catalyst. This is to be compared with
the occurrence of 2% selectivity to MPs formation and 93.4% hydro-
genation to n-hexane on the 5%-KMoTi catalyst (Fig. 8).
after Py adsorption. The spectra obtained are given in Figs. S3–S6.
The outcome of the information communicated by these spec-
tra could justify drawing the following conclusions: (i) dispersion
of molybdena species on the titania support develops the sur-
face Lewis acidity by exposing Mo6+ sites, and generates strong
Brønsted acidity via exposed Mo6+ OH groups, (ii) potassium addi-
tives weaken both the Lewis and Brønsted acid sites observed on the
MoTi catalyst, which may be attributed to formation of potassium
molybdate interaction species whereby Mo OH groups are turned
into Mo OK groups, and (iii) H2-reduction of MoTi decreases pro-
portions of the strong Lewis and Brønsted acid sites, whereas it
eliminates the acid sites on KMoTi. The IR results in support of
these conclusions are reviewed in Fig. 6. The spectrum obtained for
Py/MoTi at RT monitors two absorptions at 1636 and 1536 cm−1
for the formation of pyridinium ions [24], thus revealing the pres-
ence on MoTi of strong Brønsted acidic sites (viz., Mo OH). On
the other hand, potassium-alkalization of MoTi (i.e., formation of
KMoTi) is shown to weaken these two absorptions as a function of
the amount of the potassium additive (1–5%-K). The observed eas-
ier thermal desorption of pyridine bound to Brønsted acid sites on
KMoTi than MoTi at 473 K (Fig. 6), and the increase of this easiness
with the amount of the potassium additive, may lend a strong sup-
port to the XPS results suggesting formation of Mo
O K bonds at
the expense of the acidic Mo H groups. It is worth noting that,
O
the reduction of 5%-KMoTi rendered the material surface incapable
of irreversibly adsorbs pyridine molecules at RT. This latter result
indicates beyond doubt that the generated surface is void of strong
Lewis and Brønsted acid sites.
3.2. Catalytic measurements
Formation of the bifuctional (metal-acid) MoO2−x(OH)y sur-
face species on the reduced MoTi catalyst and the modification(s)
brought about by K additives (x-KMoTi catalyst) were assessed by
testing the catalytic activity and selectivity of the respective cat-
alysts in hydrotreating reactions of 1-hexene and n-hexane. The
results obtained are rationalized below in terms of the availability
of acidic, metallic, or dual metal-acid catalytic functions on the test
catalysts.
3.2.1. 1-hexene reactions
The above results may be explained as follows. At low reac-
tion temperatures, all of the 1-hexene molecules are catalytically
hydrogenated to n-hexane on the surface metallic function. The
same situation applies to the subsequent dehydrogenation process
(step-(ii)) observed to occur on the unmodified MoO2−x(OH)y (i.e.,
the reduced MoTi catalyst). As a matter of fact, isomerization of the
Hydrogen reduction of MoO3/TiO2 at 673 K for 12 h gener-
ates the metal-acid bifunctional MoO2−x(OH)y surface species.
Introduction of 1-hexene reactant on this surface result on 100%
conversion of 1-hexene at all reaction temperatures probed
(473–675 K, Fig. 7). Fig. 7 reveals that at the temperature of 473 K

H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
7
n-hexane is observed to encompass 97.5% isomerization selec-
tivity to the formation of MPs and di-methyl butanes (Fig. 9).
The conversion of 72.7% n-hexane with isomerization selectivity
amounting to 84.4% is observed to occur at 623 K (Fig. 9). These rel-
atively high conversion and isomerization selectivity of n-hexane
are explicable by the availability on the reduced MoTi catalyst of the
bifunctional (metal-acid) MoO2−x(OH)y species. On the contrary,
the conversion is decreased to 20.2% and the isomerization selec-
tivity to 78.9% when the reaction was conducted on 0.5%-KMoTi at
the same temperature of 623 K (Fig. 10). On
0.5%-KMoTi the MoO2−x(OH)y species are partly converted to
MoO2−x(OK)y. Thus, the observed catalytic inactivity of the 5%-
KMoTi towards n-hexane at 623 K (Fig. 10) may signify that
almost all of the strong Brønsted Mo OH groups present in the
MoO2−x(OH)y surface species are transformed into MoO2−x(OK)y. In
other words, the absence of exposed strongly acidic Mo OH groups
renders the KMoTi catalyst inactive in isomerization reactions of
n-hexane.
100
80
60
40
20
0
Conversion
Cracking
iC6
550
600
650
700
Temperature (K)
Fig. 9. Product distributions of n-hexane catalytic reactions on reduced MoTi as a
function of the reaction temperature.
4. Conclusion
produced n-hexane via the carbenium ion mechanism necessitates
the co-existence of Brønsted Mo OH acidic function. Consequently,
the observed decrease in the production of MPs as a function of
the amount of K-additive (Fig. 8), must be related to a consequent
decrease in the population of Brønsted Mo OH acidic sites. Hence,
both MoO2−x(OH)y and MoO2−x(OK)y species should co-exist on
the 0.5%-KMoTi catalyst, on which 31.0% MPs are produced (Fig. 8).
Whereas, the observed complete suppression of the formation of
MPs on the 5%-KMoTi should be correlated with an almost complete
absence of the Brøønsted Mo OH group of MoO2−x(OH)y in favor of
the formation of MoO2−x(OK)y species. These catalytic results are
in complete agreement with the above presented and discussed IR
results.
Combination of XPS–UPS, ISS and FT-IR spectroscopic studies
in parallel with catalytic measurements, carried out following the
application of the same in situ pretreatment, enable identifying
the catalytic active sites exposed on partially reduced MoO3 bulk
or dispersed particles on TiO2. A bifunctional (metal-acid) surface
species, represented here as MoO2−x(OH)y structure, is formed fol-
lowing the H2-reduction at 673 K for 12 h of the equivalent of 5
monolayers of MoO3/TiO2. The presence of the metallic function
is monitored by XPS–UPS spectra showing a DOS structure at the
Fermi level. The availability of the acid function (Brønsted Mo OH
groups) is IR-probed by the observance of two ␯CCN absorptions
at 1636 and 1536 cm−1 diagnostic for the formation of pyridinum
ions (PyH+) on Brønsted acid sites. K+ ion additives transform
the bifunctional MoO2−x(OH)y/TiO2 catalyst into a monofunctional
(metallic) catalyst by converting the associated Brønsted Mo OH
groups into the inactive Mo OK groups, as probed by the sup-
pression of the IR diagnostic absorptions of adsorbed PyH+ ions.
Furthermore, the potassium-modification improves the dispersion
and reducibility of MoO3/TiO2, as evidenced by XPS measurements.
The generation of the metal-acid bifunctional surface species
(MoO2−x(OH)y) and its K-assisted conversion into the metal mono-
functional metallic MoO2−x(OK)y species is probed by catalytic
activity and selectivity measurements towards hydrotreating reac-
tions of 1-hexene and n-hexane. Only hydrogenation reactions
of 1-hexene (metallic function catalyzed) are observed on the
potassium-modified MoO2−x(OK)y/TiO2, i.e. the, x-KMoTi catalysts,
whereas hydrogenation and isomerization reactions producing 2-
and 3-methyl pentanes are observed for 1-hexene on the K-free
bifunctional MoO2−x(OH)y/TiO2 (MoTi) catalyst. Using n-hexane as
a reactant, only isomerization reactions to methyl pentanes take
place on the MoO2−x(OH)y/TiO2 catalyst, whereas no catalytic activ-
ity is shown by x-KMoTi catalyst towards n-hexane molecules.
3.2.2. n-Hexane reactions
According to the above discussed catalytic results, it is expected
that isomerization reactions of n-hexane take place only on the
reduced MoTi (i.e., MoO2−x(OH)y/TiO2 catalyst). On the other
hand, K-modified MoTi catalysts should not show catalytic activ-
ity towards n-hexane isomerization, especially on the 5%-KMoTi
catalyst, on which almost all of the strongly acidic Mo OH groups
are converted into Mo OK groups. Indeed, the onset of the cat-
alytic activity of MoTi is observed at 573 K, at which temperature
the conversion of 39.1%
100
80
60
40
20
0
iC6
Cracking
Acknowledgements
The support by Kuwait University through research grant #
SC03/09 and GFS project # GS02/08 is gratefully acknowledged.
Conversion
3
0
1
2
4
5
Supplementary data
Concentration of K (%)
Supplementary material related to this article found, in the
online version, at http://dx.doi.org/10.1016/j.molcata.2012.11.013
Fig. 10. Product distributions of n-hexane catalytic reactions at 623 K on x-KMoTi
as a function of the amount of the K-additive (x) at 623 K.

8
H. Al-Kandari et al. / Journal of Molecular Catalysis A: Chemical 368–369 (2013) 1–8
References
[13] D.S. Kim, Y. Kurusu, I.E. Wachs, F.D. Hardcastle, K. Segawa, J. Catal. 120 (1989)
325–336.
[14] G.C. Bond, S. Flamerz, L. van Wijk, Catal. Today 1 (1987) 229–243.
[15] G.C. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, in: G.E. Muilenberg (Ed.),
Handbook of XPS, Perkin-Elmer, USA, 1979.
[16] C. Hauf, R. Kniep, G. Pfaff, J. Mater. Sci. 34 (1999) 1287–1292.
[17] H. Al-Kandari, S. Al-Kandari, F. Al-Kharafi, A. Katrib, Energy Fuels 23 (2009)
5737–5742.
[18] R. Tokarz-Sobieraj, K. Hermann, M. Witko, A. Blume, G. Metsl, R. Schlögl, Surf.
Sci. 489 (2001) 107–125.
[19] M.I. Zaki, H. Knözinger, Proceding of the 21st International Conference on
Chemical Reaction Engineering, vol. 1, Wiley-Eastern, New Delhi, 1987, pp.
19–30.
[20] A. Jordan, M.I. Zaki, C. Kappenstein, C. Geron, Phys. Chem. Chem. Phys. 5 (2003)
1708–1715.
[21] J.M. Stencel, Raman Spectroscopy for Catalysts, Van Nostrand Reinhold Catal-
ysis Series, New York, 1990, pp. 53–79.
[22] F.E. Massoth, Adv. Catal. 27 (1978) 265–309.
[23] J.A. Gadsden, Infrared Spectra of Minerals and Related Inorganic Compounds,
Butterworths, London, 1975.
[1] A. Katrib, P. Leflaive, L. Hilaire, G. Maire, Catal. Lett. 38 (1996) 95–99.
[2] H. Belatel, H. Al-Kandari, F. Al-Kharafi, A. Katrib, F. Garin, Appl. Catal. A 275
(2004) 141–147.
[3] H. Al-Kandari, F. Al-Kharafi, A. Katrib, Catal. Commun. 9 (2008) 847–852.
[4] A. Guilino, S. Parker, F.H. Jones, R.G. Egdell, J. Chem. Soc., Faraday Trans. 12
(1996) 2137–2141.
[5] F.H. Jones, R.G. Egdell, A. Brown, F.R. Wondre, Surf. Sci. 374 (1997) 80–94.
[6] F. Gault, Adv. Catal. 30 (1980) 1–95.
[7] C. Martin, I. Martin, V. Rives, B. Grzybowska, I. Gressel, Spectrochim. Acta 52
(1996) 733–740.
[8] H. Miyata, T. Mukati, T. Ono, Y. Kubokawa, J. Chem. Soc., Faraday Trans. 1 (1988)
4137–4143.
[9] K.V.R. Chary, T. Bhaskar, K.K. Seela, K.S. Lakshmi, K.R. Reddy, Appl. Catal. A 208
(2001) 291–305.
[10] H. Pines, R.C. Olberg, N.V. Ipatieff, J. Am. Chem. Soc. 70 (1948) 533–537.
[11] S. Rondon, M. Houalla, D.M. Hercules, Surf. Interface Anal. 26 (1998) 329–
334.
[12] H. Zhu, M. Shen, Y. Wu, X. Li, J. Hong, B. Liu, X. Wu, L. Dong, Y. Chen, J. Phys.
Chem. B 109 (2005) 11720–11726.
[24] M.I. Zaki, H. Knözinger, Spectrochim. Acta A 13 (1987) 1455–1459.