Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of Ti source and Ti content

Article abstract of DOI:10.1016/j.cattod.2013.09.034

Different TAPO-5 materials, prepared from Ti(III) chloride, have been tested as catalysts in the oxidation of cyclohexene with hydrogen peroxide under anhydrous conditions. Solid TiCl₃ was shown to render better synthesis reproducibility and higher catalytic activity of the resultant materials compared to TiCl₃ aqueous solution. The synthesis of these materials was carried out under N₂ atmosphere to preserve the initial oxidation state of Ti during the crystallization process. As a consequence, the so-called Ti(III)APO-5 materials have Ti environments different to those found in conventional TAPO-5 (Ti(IV)APO-5). Indeed, their catalytic activity in the oxidation of cyclohexene markedly overcomes that of the Ti(IV)APO-5 at any Ti content and after any reaction time. Turnover number (TON) of Ti(III)APO-5 samples exponentially increases as Ti content decreases, the Ti-poorer sample (Ti/(Ti + Al + P) molar ratio of 0.003) reaching TON values higher than 200 after 6 h of reaction at 343 K and higher than 550 after 24 h. The interest of Ti(III)APO-5 catalysts lies on their high selectivity to products formed through allylic oxidation of cyclohexene. The main product of this reaction, 2-cyclohexenyl hydroperoxide, was obtained with more than 80% selectivity over Ti(III)APO-5 catalysts. This behavior is in contrast with the well-known strong tendency of the more active Ti-zeolites to epoxidize the double bond.

Full text of DOI:10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2013) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of
cyclohexene: Effect of Ti source and Ti content
Almudena Alfayate, Carlos Márquez-Álvarez, Marisol Grande-Casas,
∗
Manuel Sánchez-Sánchez, Joaquín Pérez-Pariente
Instituto de Catálisis y Petroleoquímica, ICP-CSIC, C/Marie Curie 2, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 July 2013
Received in revised form
Different TAPO-5 materials, prepared from Ti(III) chloride, have been tested as catalysts in the oxidation
of cyclohexene with hydrogen peroxide under anhydrous conditions. Solid TiCl3 was shown to render
better synthesis reproducibility and higher catalytic activity of the resultant materials compared to TiCl3
aqueous solution. The synthesis of these materials was carried out under N2 atmosphere to preserve the
initial oxidation state of Ti during the crystallization process. As a consequence, the so-called Ti(III)APO-5
materials have Ti environments different to those found in conventional TAPO-5 (Ti(IV)APO-5). Indeed,
their catalytic activity in the oxidation of cyclohexene markedly overcomes that of the Ti(IV)APO-5 at
any Ti content and after any reaction time. Turnover number (TON) of Ti(III)APO-5 samples exponentially
increases as Ti content decreases, the Ti-poorer sample (Ti/(Ti + Al + P) molar ratio of 0.003) reaching
TON values higher than 200 after 6 h of reaction at 343 K and higher than 550 after 24 h. The interest
of Ti(III)APO-5 catalysts lies on their high selectivity to products formed through allylic oxidation of
cyclohexene. The main product of this reaction, 2-cyclohexenyl hydroperoxide, was obtained with more
than 80% selectivity over Ti(III)APO-5 catalysts. This behavior is in contrast with the well-known strong
tendency of the more active Ti-zeolites to epoxidize the double bond.
1
9 September 2013
Accepted 20 September 2013
Available online xxx
Keywords:
TAPO-5
Ti(III)APO-5
Cyclohexene oxidation
Allylic oxidation
Ti(III) source
2
-Cyclohexenyl hydroperoxide
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
catalytic performance. First, both networks are inherently differ-
ent in hydrophobic/hydrophilic terms, in such a way that the
Since the discovery in the early eighties of the ability of TS-1
more hydrophilic AlPO framework favors the selective adsorption
4
to catalyze the selective oxidation of organic compounds [1], the
application of Ti-zeolites as heterogeneous catalysts for the oxi-
dation of organic substrates under mild conditions has been the
subject of a large number of studies [2–5]. The interest in this kind
of zeolitic materials led to explore other potential Ti-containing
materials, such as silica-based mesoporous ones, which would
allow processing larger molecules [6–8]. However, in contrast to
the continuous development of highly active Ti-zeolites and Ti-
mesoporous materials, the attempts carried out with the related
of the most polar molecules. The influence of hydrophobic-
ity/hydrophilicity on the catalytic activity has been already made
clear in Ti-zeolites by modifying the polarity of their frameworks
or the reaction conditions [4,15–18]. A second difference between
TAPOs and Ti-zeolites is the Ti environment obtained by the isomor-
phous substitution of tetrahedral atoms in both materials. When
Ti(IV) ions are introduced in zeolitic frameworks, they can only
substitute isoelectronic Si(IV) atoms, necessarily leading to Ti(OSi)4
environments. However, Ti(IV) replaces more likely P(V) atoms in
an AlPO framework, generating Ti(OAl) environments. Such envi-
group of titanium-doped AlPO -based microporous zeotypes have
4
4
4
failed to render interesting catalysts. In spite of the large number
of the so-called TAPO materials with different structures that have
been prepared [9–11], up to date, their catalytic performance in
those oxidation processes [12–15] has been shown poor compared
to that of Ti-zeolites.
ronments are thought to inhibit the activity of Ti centers as deduced
from the lower catalytic activity of Ti-zeolites containing aluminum
compared with their Al-free analogues [4,19]. Recently, Chiesa and
co-workers [20–22] have reported some spectroscopic studies indi-
cating that Ti(IV) in an AlPO4 material can also be incorporated by
pairs, so that two Ti atoms replace a P(V) atom and its contiguous
Several aspects that make TAPOs different to Ti-zeolites have
to be considered in order to develop strategies to enhance their
Al(III) generating Ti
SAPO materials [23,24]. Moreover, though isolated titanium centers
in P(V) positions and/or Ti Ti pairs are found in the material,
O Ti units, just like Si(IV) can do in the case of
O
∗ Corresponding author. Tel.: +34 915854784; fax: +34 915854760.
E-mail address: jperez@icp.csic.es (J. Pérez-Pariente).
titanium atoms in aluminum sites have been proven to play the
most important role for redox catalytic activity [22].
0
920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.09.034
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
2
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
Recently, we have presented some innovative strategies in
TAPO-5 synthesis [25,26], which take advantage of the higher
versatility of AlPO4 frameworks compared to zeolite ones to incor-
porate heteroatom ions with different charges. Such versatility
would, in principle, allow to change the Ti environment within
the aluminophosphate framework by means of its incorporation
as Ti(III). This cation would occupy preferably Al(III) sites in order
to maintain the charge balance of the framework, and would give
Conventional Ti(IV)APO-5 catalysts were prepared strictly fol-
lowing the procedure described elsewhere [30], with the molar gel
composition 1.0Al:1.0P:xTi(IV):0.5TPAOH:20H O, using titanium
2
isopropoxide (Aldrich) as Ti(IV) source.
The samples prepared with Ti(III) sources will be named as
Ti(III)APO-5, while Ti(IV)APO-5 will denote the conventional sam-
ples prepared with the Ti(IV). The general term TAPO-5 will be
used to refer to both Ti(III)APO-5 and Ti(IV)APO-5 materials. The
nomenclature used for every particular solid will be as follows: the
oxidation state of titanium source, followed by a letter only in the
case of Ti(III)APO-5 samples to indicate the type of TiCl3 source
used (‘p’ for powder and ‘s’ for solution), next a number that will
indicate the titanium/phosphorous percent molar ratio in the gel
(×100 times) and finally the crystallization time will be indicated
by the number of hours (h) or days (d). As an example, Ti(III)p-1-4h
will then denote the sample prepared with TiCl3 powder, with a
titanium content in the gel x = 0.01 and after 4 h of crystallization.
Prior to catalytic tests, all TAPO-5 samples were calcined, in
order to eliminate the SDA molecules. The solid samples were
rise to Ti(OP)4 environments. In addition, the presence of Ti
O Ti
pairs would be prevented since replacing contiguous P(V) and Al(III)
by two Ti(III) ions would entail an excessive negative charge in
the framework, hardly compensated by a mono-protonable amine,
which is the most common structure directing agent in AlPOs.
We have reported elsewhere that this approach indeed gener-
ates TAPO-5 catalysts that are more active than the conventional
ones, being able to oxidize cyclohexene at a rate of the same
order as large-pore Ti-beta [27] under suitable conditions. This
work extends the study of TAPO-5 materials prepared from gels
containing Ti(III) ions, here called Ti(III)APO-5, as catalysts in the
cyclohexene oxidation with H O under anhydrous conditions. The
−
−
1
heated under a N2 flow of 100 mL min from room temperature
2
2
1
influence that Ti content in Ti(III)APO-5 materials has on the intrin-
sic activity of Ti centers has been also investigated, discussed, and
critically compared with the conventional TAPO-5 catalysts pre-
pared using Ti(IV) sources.
to 823 K (at a heating rate of 3 K min ) and kept at this tem-
perature for 1 h. Then they were maintained under an airflow of
−
1
100 mL min
at 823 K for 5 h. Complete removal of the organic
molecules was certified by thermogravimetric analysis.
2.2. Characterization techniques
2
. Experimental
Nature and purity of crystalline phases were studied by powder
2
.1. Catalysts preparation
X-ray diffraction (PXRD) using a PANalytical X’Pert Pro diffrac-
tometer (Cu K␣ radiation). Diffuse reflectance UV-visible (DRUV)
spectra were recorded with a Cary 5000 Varian spectrophotometer
equipped with an integrating sphere using the synthetic poly-
mer Spectralon as reference. The spectra were corrected applying
the Kubelka–Munk function. The chemical composition of the
solids was determined by inductively coupled plasma (ICP-OES)
spectrometry with an ICP Winlab Optima 3300 DV Perkin–Elmer
spectrometer. Scanning electron microscopy (SEM) micrographs
were taken in a FEI Nova NANOSEM microscope with a vCD detec-
tor. Nitrogen adsorption/desorption isotherms were measured at
77 K in a Micromeritics ASAP 2010 equipment. Calcined samples
were previously degassed at 623 K for 16 h. Surface areas were esti-
mated by applying the BET method, and micropore volume was
obtained by application of the t-plot method to the N2 adsorption
data.
Ti(III)APO-5 samples were prepared with different titanium
content by hydrothermal treatment using titanium trichloride
as Ti(III) source [25,26], in one of these two formulations:
(
i) as TiCl3 (∼10 wt%) in HCl (20–30 wt%) aqueous solu-
tion (supplied by Aldrich with the exact composition), and
(
ii) TiCl3 powder (Aldrich). The general gel composition was
(
1 − x)Al:1.0P:xTi(III):mMCHA:25H O:nHCl for the former TiCl
2
3
source, where x denotes the Ti/P molar ratio, n is the HCl/P molar
ratio resultant after its inevitable addition with TiCl in hydrochlo-
3
ric acid solution, MCHA refers to N-methyldicyclohexylamine and
m designates the MCHA/P molar ratio, which was varied to com-
pensate the changes of pH introduced by the HCl added, so that the
pH value of the gel was forced to be in the range 6.5–7.0. When
the TiCl3 powder was used, the gel composition was simplified
to (1 − x)Al:1.0P:xTi(III):0.8MCHA:25H O having pH values in the
2
same range. Both the preparation of the gel and the following auto-
claves sealing were carried out under inert atmosphere, in a glove
bag filled with nitrogen. The TiCl3 source was added over an aque-
ous solution of phosphoric acid (85 wt%, Sigma) in deionized water,
which turns into purple color. Next, Al(OH) ·xH O (Sigma–Aldrich)
2.3. Catalytic experiments
Catalysts were tested in the oxidation of cyclohexene with H O2
2
under anhydrous conditions as described elsewhere [27]. These
conditions were reached after removing almost all the water that
accompanies the commercial 30 wt% H O aqueous solution. Water
3
2
was added over the solution and the resultant suspension was vig-
orously stirred for ca. 10 min, followed by the dropwise addition of
N-methyldicyclohexylamine (MCHA). This amine was selected as
structure directing agent (SDA) because of its high specificity to AFI-
2
2
was removed by Soxhlet extraction of solutions prepared dissolving
the 30 wt% H O2 solution in acetonitrile. The reactions were car-
2
ried out in batch mode using a round bottom flask equipped with
a magnetic stirrer, a thermometer and a covered reflux condenser
provided with a drying tube on top. The later was used in order to
avoid the condensation of ambient humidity inside the condenser
caused by the low temperature of water circulating through it. The
coolant water temperature was controlled at 278 K to prevent any
evaporation of the chemicals during the reaction. Calcined catalysts
were activated overnight inside the reaction system at 433 K under
a flow of N2 to remove any water adsorbed into the catalyst chan-
nels. Under the reaction conditions used, mass balance was ca. 98%
after 30 h of reaction.
structured AlPO -based materials [28]. In particular, this specificity
4
has already been proven for AlPO4 materials doped with the same
heteroatom, Sn, in two different oxidation states [29]. The resul-
tant gel was stirred for 1 h before transferring it into Teflon-lined
stainless steel autoclaves for the hydrothermal treatment at 448 K
under autogeneous pressure. The obtained purple solids, whose
color indicated that at least part of the Ti present in the solid main-
tains its oxidation state 3+ after the crystallization process, were
recovered by filtration and washed with deionized water. After dry-
ing under ambient atmosphere, the solids became yellowish white,
suggesting that Ti(III) was oxidized to the most stable titanium
oxidation state, Ti(IV).
In a typical experiment, 3.0 g of cyclohexene, the correspond-
ing amount of H O2 in CH CN solution (previously submitted to
2
3
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
3
Table 1
Ti, Al and P content of the synthesis gels and corresponding values determined by
ICP-OES for as-prepared Ti(III)APO-5 and Ti(IV)APO-5 samples.
Ti(III)s-4-18h
Sample
Gel compositiona
Solid compositiona
Ti(III)p-4-18h
Ti(III)p-4-4h
Ti
Al
P
Ti
Al
P
Ti(III)p-0.5-4h
Ti(III)p-1-4h
Ti(III)p-2-4h
Ti(III)p-4-4h
Ti(III)p-4-18h
Ti(III)s-4-18h
Ti(IV)-1-8d
0.005
0.01
0.02
0.04
0.04
0.04
0.01
0.02
0.995
0.99
0.98
0.96
0.96
0.96
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.006
0.009
0.034
0.069
0.065
0.062
0.005
0.014
1.012
1.011
1.025
1.026
0.928
0.988
1.018
1.042
0.982
0.980
0.941
0.905
1.007
0.950
0.977
0.944
Ti(III)p-2-4h
Ti(III)p-1-4h
Ti(III)p-0.5-4h
Ti(IV)-2-8d
Ti(IV)-2-8d
a
Ti, Al and P contents, either in gel or in solid, are expressed considering that
the sum of all three 2, in such a way that, with the values given in the table, the
composition corresponds to the general formula TixAlyPzO4.
Soxhlet extraction of water) and additional acetonitrile, if neces-
sary, were mixed to reach a molar ratio CH CN:H O :cyclohexene
Ti(IV)-1-8d
3
2
2
of 20:0.55:1. Toluene was added as internal standard. This solu-
tion was poured into the flask containing the activated catalyst
10
20
30
40
2θ / º
(
cyclohexene/catalyst weight ratio of 10). The reaction mixture was
heated up to 343 K under continuous stirring. The reaction progress
was followed by taking different aliquots under stirring. The time
zero aliquot was taken when the setpoint temperature was reached
in the mixture.
Fig. 1. PXRD patterns of the as-prepared TAPO-5 samples synthesized with different
Ti(III) and Ti(IV) content.
materials, which were purple-colored after the crystallization pro-
cess [25,26], became white due to the oxidation of Ti(III) to
Ti(IV) during the drying process under ambient atmosphere or
with the subsequent calcination process. The most intense band
in the spectra (found at 205–232 nm) corresponds to charge-
transfer transition between oxygen and tetrahedrally coordinated
isolated titanium. Its maximum is red shifted when titanium load-
ing increases. In the spectra of samples with lower content of the
heteroatom, the maxima are found at a particularly low wavelength
Products were separated and analyzed by gas chromatography
in a Varian CP-3380 gas chromatograph equipped with a capillary
column (dimethylpolysiloxane) with length of 15 m, i.d. of 0.25 mm
and thickness of 1 ␮m, and a flame ionization detector (FID). Prod-
1
13
ucts were identified by mass spectrometry, H and C NMR and
infrared spectroscopies, as described elsewhere [27]. Conversions
were calculated from the yields of the resultant products. In order to
study the possible leaching of Ti from the heterogeneous catalysts
during the reaction, all the final reaction mixtures were filtered and
analyzed by total X-ray fluorescence. No traces of titanium were
found in these solutions for any of the catalysts tested (detection
limit of the technique was below 0.1 ppm). A blank experiment was
performed under the same reaction conditions resulting in a total
conversion lower than 5% after 30 h.
(
205 nm). That position is unlike in conventional TAPO-5, partic-
ularly if they are calcined, and it is typical of Al-free Ti-zeolites
18]. For a similar titanium content, the maxima of this band is
[
red-shifted in the spectra of the Ti(IV)APO-5 samples with respect
to these of the Ti(III)APO-5 samples: from 205 nm in Ti(III)p-0.5-
4
h to 215 nm in Ti(IV)-1-8d and from 215 nm in Ti(III)p-1-4h to
3
. Results and discussion
232nm
3.1. Catalysts characterization
280nm
Table 1 summarizes the composition of the gels and the as-
prepared catalysts. Titanium incorporation was less effective when
Ti(IV) source was used, the titanium content in the solid being lower
than in the starting gel. In contrast, the titanium content of cata-
lysts obtained with Ti(III) sources was in all cases equal to or higher
than that of the corresponding gel. For this reason, later discussions
about characterization and catalytic activity results will be done
comparing Ti(III)APO-5 and Ti(IV)APO-5 samples with similar tita-
nium content in the solid recovered, instead of those prepared with
the same Ti content in the synthesis gel. Hence Ti(III)p-0.5-4h will
be compared with Ti(IV)-1-8d and Ti(III)p-1-4h with Ti(IV)-2-8d.
Powder X-ray diffraction patterns of as-prepared samples are
shown in Fig. 1. All the materials show the typical pattern of pure
AFI phase and no signals of impurities appear in the diffractograms.
After calcination, all samples kept the pure AFI phase. It is notice-
able that at high content of titanium longer crystallization times
were necessary to achieve crystallinities similar to those obtained
at lower content of this heteroatom.
2
05nm
Ti(III)s-4-18h
Ti(III)p-4-18h
Ti(III)p-4-4h
Ti(III)p-2-4h
Ti(III)p-1-4h
Ti(III)p-0.5-4h
Ti(IV)-2-8d
Ti(IV)-1-8d
200
250
300
Wavelength / nm
350
400
Fig. 2 shows the UV region of the diffuse reflectance UV-visible
spectra of the calcined samples. It is worth recalling that titanium
is in its 4+ oxidation state in all these samples, for Ti(III)APO-5
Fig. 2. UV region of the DRUV-visible spectra of the calcined samples Ti(III)APO-5
and Ti(IV)APO-5 with different titanium content.
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
4
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
Fig. 3. Scanning electron micrographs of Ti(III)APO-5 and Ti(IV)APO-5 samples with different Ti content.
2
33 nm in Ti(IV)-2-8d. This systematic shift infers that titanium
TiO -containing species, has been recently assigned to Ti O Ti
2
atoms have a different local coordination environment when the
TAPO-5 is synthesized using a Ti(III) source [25,26] instead of the
Ti(IV) one. Additionally, the UV-visible spectra of all the samples
present a contribution centered at ca. 280 nm, which is more pro-
nounced in the spectra of samples with higher titanium content,
these prepared at longer crystallization times and the Ti(IV)APO-5
ones. It can be noticed that the crystallization time seems to have a
specific effect on the UV spectra. The two samples synthesized with
the largest amount of Ti(III) in the gel have practically the same Ti
content in the solids but the contribution at ca. 280 nm is much
more pronounced in the sample synthesized at 18 h. The nature
of this contribution, traditionally attributed to extra framework
bond within the framework [20,21]. These species are anyway
undesired in Ti-zeolites from the point of view of catalytic activity,
the isolated titanium centers being preferred [31].
SEM micrographs (Fig. 3) show that the materials having the
three lowest titanium contents, Ti(III)p-0.5-4h, Ti(III)p-1-4h and
Ti(III)p-2-4h, are obtained as polycrystalline spherical-shape aggre-
gates of elongated rod-like crystals. This morphology has been
already observed in other MeAPO-5 materials prepared with the
same SDA [32]. The particles of the samples with the highest tita-
nium content synthesized with both TiCl3 sources (powder and in
an aqueous solution of hydrochloric acid) are of hexagonal prism-
like shape and formed by non-isolable fused crystals. The range of
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
5
Table 2
The explanation to this apparent anomaly between catalytic behav-
ior and spectroscopic features should be based on the singularities
introduced by Ti(III) incorporation into the AlPO4 framework with
respect to Ti(IV) [26].
Specific surface areas and micropore volumes of Ti(III)APO-5 samples estimated
from N2 adsorption/desorption isotherms.
SBET (m2
g
−1
)
Micropore volume
Sample
3
−1
)
(
cm
g
Moreover, it must be noted that the only difference between
both samples is the nature of the used TiCl source. The TiCl aque-
Ti(III)p-0.5-4h
Ti(III)p-1-4h
Ti(III)p-2-4h
Ti(III)p-4-4h
Ti(III)p-4-18h
Ti(III)s-4-18h
Ti(IV)APO-5
336
317
245
207
302
285
305
0.115
0.111
0.083
0.065
0.112
0.091
0.116
3
3
ous solution has a high and variable concentration of HCl which
differs in the range of 20–30 wt% from one commercial batch to
another, making difficult to control certain synthesis conditions
−
such as Cl concentration or pH, which can affect the subsequent
crystallization process. Despite the relatively long crystallization
time used to obtain the sample Ti(III)s-4-18h, the lack of control
over the synthesis conditions could affect its crystallinity and its
textural properties. As shown in Table 2, its micropore volume is
below the typical value of conventional Ti(IV)APO-5 [10]. In other
words, the use of solid TiCl3 is recommended instead of TiCl3 sta-
bilized in HCl aqueous solution, as the former does not introduce
any extra species in the gel, does not need to alter the general gel
composition to compensate/control the starting pH and, maybe as
a consequence of the two previous reasons, generates materials in
a much more reproducible way.
the aggregates sizes is from 10 to 30 ␮m for these samples and are
comparable to those of conventional catalysts synthesized with the
Ti(IV) source, which also consist of aggregates of ca. 20 ␮m.
Textural properties of the calcined samples were studied by N2
adsorption/desorption at 77 K. BET surface areas and micropore
volumes estimated by the t-plot method are collected in Table 2.
The sample Ti(III)p-4-4h has specific surface area and micropore
volume lower than the values reported for conventional Ti(IV)APO-
5
materials [10]. The comparison with its homologue crystallized
On the other hand, the crystallization time also has an effect on
the subsequent catalytic behavior. There is a significant difference
in the conversion given by samples prepared with the same Ti(III)
source but crystallized for 4 h (Ti(III)p-4-4h) and 18 h (Ti(III)p-4-
during 18 h suggests that this is likely due to an uncompleted crys-
tallization as it was also reflected in the XRD patterns (Fig. 1).
3.2. Effect of the Ti(III) source on the catalytic activity
1
8h). Since this cannot be attributed to different titanium content
or to an important effect of crystal size, the lower catalytic activity
of the sample Ti(III)p-4-4h should be due to its lower crystallinity
TAPO-5 catalysts were tested in the oxidation of cyclohexene
with H O under anhydrous conditions as described in the exper-
2
2
(Table 2 and Fig. 1). Considering UV-visible spectra (Fig. 2), again
imental section. The catalytic tests were designed to adequate the
reaction conditions to the hydrophilic nature of the TAPO catalysts
27]. In our previous work, both conversion and selectivity of the
sample here denoted as Ti(III)s-4-18h were compared to those of
the conventional Ti(IV)APO-5 and a large-pore Ti-beta zeolite. This
current work explores the catalytic activity of different Ti(III)APO-
one would expect just the opposite result according to the intensity
of the band at ca. 280 nm, which is higher for the sample crys-
tallized for 18 h. Therefore, the low crystallinity and accordingly
low surface area and micropore volume of sample Ti(III)p-4-4h
[
(Table 2) seem to have a more important effect on the overall
activity than the higher content of the more active isolated Ti
species suggested by the UV-visible spectra. In any case, the fact
that samples of high crystallinity are more active than similar ones
with lower crystallinity indicates that, whatever is the role of the
Ti(III)APO-5 catalysts in the general mechanism of the reaction, the
catalytically-active part of the sample is the crystalline one.
5
catalysts prepared with different Ti content and different Ti(III)
sources. Ti(III)APO-5 catalysts synthesized under identical crys-
tallization conditions but with different Ti(III) sources, that is
Ti(III)s-4-18h and Ti(III)p-4-18h, showed very different activity
Fig. 4), the latter being more active. According to either traditional
33,34] or recent [20–22] interpretation given to the contribution
(
[
at 280 nm in the DRUV spectra (Fig. 2), and assuming that Ti centers
forming part of Ti Ti bonds possess lower catalytic activity [31],
the relative catalytic activity of those two samples was unexpected.
O
3.3. Effect of the Ti(III) content on the catalytic activity
The catalytic activity of the Ti(III)APO-5 samples with differ-
25
20
15
10
5
0
ent titanium content has been also studied in the oxidation of
cyclohexene with H O2 under anhydrous conditions. Fig. 5 shows
2
Ti(III)p-4-18h
Ti(III)s-4-18h
the cyclohexene conversion with the reaction time over differ-
ent TAPO-5 catalysts prepared from TiCl3 powder. The samples
with the two lowest titanium contents, Ti(III)p-0.5-4h and Ti(III)p-
1
-4h, together with the sample Ti(III)p-4-18h, gave the highest
conversions of the olefin. However, their titanium content is quite
different, so the intrinsic activity per Ti center, that is, the turnover
number (TON), is markedly different (Fig. 6A and B). After 6 h of
reaction, an inverse correlation, with almost exponential decay,
between the turnover numbers and the titanium content in the
Ti(III)APO-5 samples (full squares in Fig. 6A) is observed. This result
is in good agreement with the UV spectra of this series of samples
as regards the shift to higher wavelength of the maximum of the
charge transfer band and the increase of the contribution of the
Ti(III)p-4-4h
0
5
10
15
t / h
20
25
30
signal at ca. 280 nm, probably due to the presence of Ti
O Ti pairs
[
20–22]. Taking these aspects into account, it can be inferred that
lower titanium content produces higher proportion of isolated tita-
nium centers in the material, with a singular environment within
the framework that makes them very active in the cyclohexene
oxidation reaction.
Fig. 4. Cyclohexene conversion over TAPO-5 catalysts prepared with different Ti(III)
sources: Ti(III)s-4-18h (squares), prepared with aqueous hydrochloric acid solution
of TiCl3, and Ti(III)p-4-4h (diamonds) and Ti(III)p-4-18h (stars) prepared with TiCl3
powder.
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
6
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
30
25
20
15
10
5
0
the preference of TAPO-5 to favor the reaction through the radical
Ti(III)p-1-4h
pathway, mostly leading to the allylic oxidation products, instead
of those formed through the epoxidation of the double bond.
Table 3 summarizes the total conversion results, the yields corre-
sponding to both reaction mechanisms and the product selectivity
obtained for the Ti(III)APO-5 materials at the end of the reaction,
after 30 h. Selectivity to the reaction products are quite similar
for all catalysts, the most abundant product being 2-cyclohexenyl
hydroperoxide [27] in all cases. Other side-products such as 2-
cyclohexen-1-ol and 2-cyclohexen-1-one are also detected, surely
coming from the hydroperoxide [39,40]. Cyclohexanol and some-
times 1,2-cyclohexanediol, formed from the epoxide, are also
detected at limited or negligible extension. These results could
make Ti(III)APO-5 materials to become potential catalysts for cer-
tain applications that require the activation of the allylic position
without affecting the double bond.
Ti(III)p-4-18h
Ti(III)p-0.5-4h
Ti(III)p-2-4h
Ti(III)p-4-4h
0
5
10
15
20
t / h
25
30
35
3.4. Influence of oxidation state of Ti source on the catalytic
Fig. 5. Cyclohexene conversion over Ti(III)APO-5 catalysts prepared with TiCl3 pow-
der and different titanium content: Ti(III)p-05-4h (circles), Ti(III)p-1-4h (squares),
Ti(III)p-2-4h (triangles), Ti(III)p-4-4h (diamonds) and Ti(III)p-4-18h (stars).
activity of TAPO-5
This section focuses on a more extended comparison of the
catalytic behavior of Ti(III)APO-5 and Ti(IV)APO-5 materials with
different titanium content. The cyclohexene conversion levels
reached by samples Ti(III)p-0.5-4h and Ti(III)p-1-4h are compared
with those of Ti(IV)-1-8d and Ti(IV)-2-8d in Fig. 7. Regardless the
titanium content, Ti(III)APO-5 catalysts (full symbols) are always
more active than the conventional Ti(IV)APO-5 (open symbols), at
any tested reaction time.
On the other hand, the catalytic activity per titanium center is
systematically much higher for the materials Ti(III)APO-5 at any
time of reaction (Fig. 6). On the contrary, no significant differ-
ences in the selectivities to the different reaction products were
observed (Table 3). Thus, the most abundant product obtained with
Ti(III)APO-5 materials, 2-cyclohexenyl hydroperoxide, is also the
one with the highest selectivity for Ti(IV)APO-5 catalysts.
After 24 h of reaction, the pattern of TON evolution with the
titanium content (Fig. 6B) is similar to that found after 6 h of reac-
tion (Fig. 6A), although, obviously, absolute values of TON increased
from 6 h to 24 h. The comparison of catalytic activity per Ti center of
the Ti-richest Ti(III)APO-5 samples shows that TON values are sub-
stantially affected by the crystallization time, resulting the sample
prepared at shorter time (and not fully crystallized) with lower TON
values, as expected from the discussion of the previous section.
Cyclohexene oxidation can take place through two different
mechanisms [27,35–38]: (i) a homolytic one involving radical
species, and (ii) an heterolytic mechanism implying the oxidation
of the double bond to give the epoxide as either final product or
intermediate of different products formed by epoxide ring aper-
ture. It is well-known [19,31] that the latter mechanism is directed
by Ti-zeolite catalysts while, on the contrary, the studied TAPO-5
materials mainly induced the reaction through the radical mecha-
nism under anhydrous conditions [27]. All tested catalysts certify
Although Table 3 breaks down the conversion into the selectiv-
ity to different products for different TAPO-5 catalysts after 30 h of
reaction, a valid comparison of selectivity between different sam-
ples strictly requires referring it to the same conversion rather than
6
00
00
A
B
6
h of reaction
24h of reaction
200
150
100
5
400
300
200
100
0
5
0
0
0
.00
0.01
0.02
Ti / (Ti+Al+P)
0.03
0.04
0.00
0.01
0.02
0.03
0.04
Ti / (Ti+Al+P)
Fig. 6. Cyclohexene conversion per Ti center (TON) in Ti(III)APO-5 (full squares) and Ti(IV)APO-5 (open circles) samples with different titanium content after 6 h (A) and 24 h
(
B) of reaction.
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
7
30
25
20
15
10
5
0
25
Ti(III)p-1-4h
20
15
10
5
0
Radical
Radical
Ti(III)p-0.5-4h
Ti(IV)-1-8d
Ti(IV)-2-8d
Epoxidation
Epoxidation
0
5
10
15
20
25
30
35
0
5
10
15
t / h
20
25
30
t / h
Fig. 8. Conversion (thick lines) and sum of the yield of products (thin lines) formed
through radical (diamonds) and epoxidation (squares) mechanisms of the Ti-poorest
Ti(III)APO-5 (full symbols) and Ti(IV)APO-5 (empty symbols) samples.
Fig. 7. Cyclohexene conversion catalyzed by Ti(III)APO-5 samples: Ti(III)p-0.5-4h
(full circles) and Ti(III)p-1-4h (full squares), and Ti(IV)APO-5 materials: Ti(IV)-1-8d
open triangles) and Ti(IV)-2-8d (open hexagons).
(
the same reaction time. However, the here addressed reaction is
very unusual in this sense, as two different reaction mechanisms
coexist and they are taking place almost in consecutive sequence
rather than in parallel. It means that imposing the condition of
equal conversion level for selectivity comparison purposes could
not be the best solution in this particular case, if it implies to com-
pare very different reaction times. Consequently, we compare in
Fig. 8 the conversion and selectivity (to the sum of products formed
by one or the other mechanism) along the reaction time, for the
Ti(III)APO-5 and Ti(IV)APO-5 materials with the lowest and similar
Ti content, instead of comparing the selectivities at a given similar
conversion. The samples with the lowest Ti content were selected
due to the exceptionally high TON values given by Ti(III)p-0.5-4h
path products (83 and 74%, respectively, after 30 h, Table 3), what
makes Ti(III)APO-5 materials much more promising candidates as
TAPO-based heterogeneous catalysts alternative to Ti-zeolites in
the oxidation of cyclohexene with H O2 [27].
2
Given that the different catalytic activity shown by Ti(III)APO-5
and Ti(IV)APO-5 cannot be assigned to a drastic difference in crystal
size (Fig. 3), it could be attributed to the distinct mechanism of Ti
incorporation imposed by the use of Ti sources with different tita-
nium oxidation state. Some consequences of the different oxidation
state of the titanium incorporated in the AlPO4 framework were
already noticed in the UV spectra (Fig. 2). The spectra of calcined
samples Ti(III)p-0.5-4h and Ti(III)p-1-4h presented narrower bands
at lower wavelength and with a less intense contribution at 280 nm.
All these catalytic and spectroscopic features of both Ti(III)APO-5
and Ti(IV)APO-5 materials strongly evidence some differences in
Ti environment, suggesting the presence of titanium centers with
higher intrinsic activity in the material Ti(III)APO-5. Admitting that
titanium occupying aluminum sites are the centers with the most
relevant redox potential [20–22], the higher conversion of the sam-
ples Ti(III)APO-5 should imply a higher amount of titanium in that
environment within the AlPO4 framework. In principle, that goal
can be only achieved by the incorporation of the metal in a lower
oxidation state, as Ti(III), in the synthesis gel.
(Fig. 6), which is likely due to the higher proportion of Ti occupy-
ing Al sites in the AlPO4 framework in this sample. Apart from the
already commented higher conversion of the Ti(III)APO-5 sample,
both catalysts have similar behavior in terms of selectivity. Thus, at
long reaction times, in both cases the radical mechanism ends up
dominating over the epoxidation one, which is however compet-
ing with the former at short times. It means that the selectivity to
products formed via the radical mechanism (mainly 2-cyclohexenyl
hydroperoxide, Table 3) simply increases by prolonging the reac-
tion time. That strategy is based on the different kinetic trend of
both mechanisms, since the epoxidation products yield curve is a
kind of downward horizontal asymptote whereas that of the rad-
ical mechanism follows a linear tendency with time. Moreover, a
detailed analysis of the difference in selectivity between both sam-
ples shows that, for this low Ti content, Ti(III)APO-5 is not only more
active than Ti(IV)APO-5 but it is also more selective to the radical
In summary, the approach of incorporating Ti(III) instead of
Ti(IV) ions into the AlPO -5 framework was successfully achieved,
4
producing TAPO-5 catalysts of promising catalytic behavior. How-
ever, the exact nature of the Ti sites, as well as the relationship
between the Ti environment and the catalytic activity, are far from
Table 3
Conversion and product selectivity in the oxidation of cyclohexene with H2O2 under anhydrous conditions over Ti(III)APO-5 catalysts after 30 h of reaction.
Sample
Total conversiona (%)
Yieldb (%)
Product selectivityc (%)
Epoxidation
Radical
Epox
Chol
Diol
Enol
Enone
Chhp
Ti(III)p-0.5-4h
Ti(III)p-1-4h
Ti(III)p-2-4h
Ti(III)p-4-4h
Ti(III)p-4-18h
Ti(IV)-1-8d
24
27
22
15
24
16
12
4
4
3
4
2
4
2
20
23
19
11
22
12
10
14
12
12
14
8
2
2
2
7
2
5
1
1
–
–
3
–
–
–
3
3
5
7
3
3
1
5
8
11
12
6
75
75
70
57
81
69
76
21
18
2
4
Ti(IV)-2-8d
a
Conversion calculated by the total addition of the yields of the resultant products.
Sum of the yields of the products obtained through every mechanism.
Epox, cyclohexene oxide; Chol, cyclohexanol; Diol, 1,2-cyclohexanediol; Enol, 2-cyclohexen-1-ol; Enone, 2-cyclohexen-1-one; Chhp, 2-cyclohexenyl hydroperoxide.
b
c
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034

GModel
CATTOD-8596; No. of Pages8
ARTICLE IN PRESS
8
A. Alfayate et al. / Catalysis Today xxx (2013) xxx–xxx
being known. Nevertheless, some insights based on different char-
acterization techniques are given elsewhere [26].
[5] J.C. van der Waal, M.S. Rigutto, H. van Bekkum, Appl. Catal. A 167 (1998)
31–342.
3
[
[
6] V. Hulea, F. Fajula, J. Bousquet, J. Catal. 198 (2001) 179–186.
7] V. Hulea, E. Dumitriu, Appl. Catal. A 277 (2004) 99–106.
4
. Conclusions
[8] T. Blasco, A. Corma, M.T. Navarro, J. Pérez-Pariente, J. Catal. 156 (1995) 65–74.
9] B.M.T. Lok, B.K. Marcus, E. Flanigen, U.S. Patent 4,500,651 (1985).
[
[
10] M.H. Zahedi-Niaki, P.N. Joshi, S. Kaliaguine, Progress in zeolite and microporous
materials, in: S.K.I. Hakze Chon, U. Young Sun (Eds.), Studies in Surface Science
and Catalysis, vol. 105, Elsevier, Seoul, 1997, pp. 1013–1020.
Ti(III)APO-5 materials have been successfully prepared using
different Ti(III) sources, in a wide range of Ti content and after
different crystallization times, using the AFI-specific structure
directing agent N-methyldicyclohexylamine (MCHA) and prepar-
ing the starting gel under inert atmosphere. At least part of Ti
seems to be incorporated as Ti(III) into the AlPO4 framework.
Although this Ti(III) is oxidized to Ti(IV) as soon as the sample is
dried under ambient atmosphere, the aim of rerouting the conven-
tional Ti incorporation in P sites of an AlPO4 framework towards
its incorporation in Al sites has been somehow successful, produc-
ing TAPO-5 materials with unprecedented catalytic activity under
the tested conditions. Thus, Ti(III)APO-5 materials resulted signif-
icantly more active in the oxidation of cyclohexene with H O
[
[
11] F.J. Luna, S.E. Ukawa, M. Wallau, U. Schuchardt, J. Mol. Catal. A: Chem. 117
(
1997) 405–411.
12] M.H. Zahedi-Niaki, M.P. Kapoor, S. Kaliaguine, J. Catal. 177 (1998) 231–239.
[13] A. Tuel, Zeolites 15 (1995) 228–235.
[14] S.O. Lee, R. Raja, K.D.M. Harris, J.M. Thomas, B.F.G. Johnson, G. Sankar, Angew.
Chem. Int. Ed. 42 (2003) 1520–1523.
[
15] B.Y. Hsu, S. Cheng, J.M. Chen, J. Mol. Catal. A: Chem. 149 (1999) 7–23.
[16] A. Corma, P. Esteve, A. Martínez, J. Catal. 161 (1996) 11–19.
[17] J.C. van der Waal, H. van Bekkum, J. Mol. Catal. A: Chem. 124 (1997) 137–146.
[
[
[
[
[
[
18] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martínez, J.A. Perdigón-
Melón, S. Valencia, J. Phys. Chem. B 102 (1998) 75–88.
19] J.C. Torres, D. Cardoso, R. Pereira, Microporous Mesoporous Mater. 136 (2010)
97–105.
20] S. Maurelli, M. Vishnuvarthan, M. Chiesa, G. Berlier, S. Van Doorslaer, J. Am.
Chem. Soc. 133 (2011) 7340–7343.
2
2
under anhydrous conditions than conventional Ti(IV)APO-5 ones.
Moreover, very high activity per Ti center was achieved when
Ti(III)APO-5 catalysts with low Ti content were used. Furthermore,
these catalysts favor the oxidation of cyclohexene through the
allylic position of the double bond, unlike Ti-containing silica-based
nanoporous materials that favor the oxidation of cyclohexene
through double bond epoxidation. Thinking in possible industrial
applications, interestingly Ti(III)APO-5 catalysts are quite selec-
tive (even above 80%) to 2-cyclohexenyl hydroperoxide, which
can easily further react to give some oxidized cyclohexene deriva-
tives having intact the double bond. Practically the same selectivity
is also given by the less-active Ti(IV)APO-5 catalyst, which sug-
gests that only Ti atoms surrounded by P (presumably abundant in
Ti(III)APO-5 and scarce in Ti(IV)APO-5) are the catalytically-active
centers.
21] S. Maurelli, M. Vishnuvarthan, G. Berlier, M. Chiesa, Phys. Chem. Chem. Phys.
14 (2012) 987–995.
22] C. Novara, A. Alfayate, G. Berlier, S. Maurelli, M. Chiesa, Phys. Chem. Chem. Phys.
15 (2013) 11099–11105.
23] S. del Val, T. Blasco, E. Sastre, J. Pérez-Pariente, J. Chem. Soc., Chem. Commun.
(1995) 731–732.
[
[
24] A.M. Prakash, S. Unnikrishnan, K.V. Rao, Appl. Catal. A 110 (1994) 1–10.
25] A. Alfayate, M. Sánchez-Sánchez, J. Pérez-Pariente, ES Patent
PCT/ES2012/070436 (2011).
[26] A. Alfayate, M. Sánchez-Sánchez, J. Pérez-Pariente, Microporus Mesoporous
Mater., submitted.
[
27] A. Alfayate, C. Márquez-Álvarez, M. Grande-Casas, B. Bernardo-Maestro, M.
Sánchez-Sánchez, J. Pérez-Pariente, Catal. Today 213 (2013) 211–218.
[28] M. Sánchez-Sánchez, G. Sankar, A. Simperler, R.G. Bell, C.R.A. Catlow, J.M.
Thomas, Catal. Lett. 88 (2003) 163–167.
[
[
[
29] M. Sánchez-Sánchez, R. van Grieken, D.P. Serrano, J.A. Melero, J. Mater. Chem.
19 (2009) 6833–6841.
30] A.M. Prakash, L. Kevan, M.H. Zahedi-Niaki, S. Kaliaguine, J. Phys. Chem. B 103
(1999) 831–837.
31] B. Notari, Innovation in zeolite materials science, in: P.J. Grobet, W.J. Mortier,
E.F. Vansant, G. Schulz-Ekloff (Eds.), Studies in Surface Science and Catalysis,
vol. 37, Elsevier, Belgium, 1988, pp. 413–425.
Acknowledgments
[
[
32] A. Manjón-Sanz, M. Sánchez-Sánchez, P. Mu n˜ oz-Gómez, R. García, E. Sastre,
Microporous Mesoporous Mater. 131 (2010) 331–341.
33] E. Gianotti, A. Frache, S. Coluccia, J.M. Thomas, T. Maschmeyer, L. Marchese, J.
Mol. Catal. A: Chem. 204–205 (2003) 483–489.
The authors thank Beatriz Bernardo-Maestro and Rubén
Sepúlveda for their contributions in the preparation of the
materials. They also acknowledge the Spanish Government for
financial support (Projects MAT-2009-13569 and MAT-2012-
[34] S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina, F. Boscherini, F.
Buffa, F. Genoni, G. Leofanti, J. Phys. Chem. 98 (1994) 4125–4132.
[
35] R.R. Sever, R. Alcala, J.A. Dumesic, T.W. Root, Microporous Mesoporous Mater.
6 (2003) 53–67.
3
1127). A. Alfayate acknowledges CSIC for a PhD Jae-Predoc
6
fellowship.
[36] J. Vernimmen, M. Guidotti, J. Silvestre-Albero, E.O. Jardim, M. Mertens, O.I.
Lebedev, G. Van Tendeloo, R. Psaro, F. Rodriguez-Reinoso, V. Meynen, P. Cool,
Langmuir 27 (2011) 3618–3625.
References
[
[
37] E. Jorda, A. Tuel, R. Teissier, J. Kervennal, J. Catal. 175 (1998) 93–107.
38] J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, J. Catal. 204 (2001) 146–156.
[
[
[
[
1] M. Taramasso, G. Perego, B. Notari, U.S. Patent 4,410,501 (1983).
2] P. Ratnasamy, D. Srinivas, H. Knözinger, Adv. Catal. 48 (2004) 1–169.
3] B. Notari, Adv. Catal. 41 (1996) 253–334.
[39] S.M. Mahajani, M.M. Sharma, T. Sridhar, Chem. Eng. Sci. 54 (1999) 3967–3976.
[40] H.E.B. Lempers, J.D. Chen, R.A. Sheldon, Catalysis by microporous materials, in:
H.K. Beyer, H.G. Karge, I. Kiricsi, J.B. Nagy (Eds.), Studies in Surface Science and
Catalysis, vol. 94, Elsevier, Hungary, 1995, pp. 705–712.
4] A. Corma, M.A. Camblor, P. Esteve, A. Martínez, J. Pérez-Pariente, J. Catal. 145
(
1994) 151–158.
Please cite this article in press as: A. Alfayate, et al., Ti(III)APO-5 materials as selective catalysts for the allylic oxidation of cyclohexene: Effect of
Ti source and Ti content, Catal. Today (2013), http://dx.doi.org/10.1016/j.cattod.2013.09.034