Enhanced catalytic activity of TAPO-5 in the oxidation of cyclohexene with hydrogen peroxide under anhydrous conditions

Article abstract of DOI:10.1016/j.cattod.2013.05.001

This work studies the influence of the amount of water accompanying and stabilizing the conventional oxidant agent H₂O₂ on the activity of different TAPO-5 materials in cyclohexene oxidation, in comparison to that of the par excellen

Full text of DOI:10.1016/j.cattod.2013.05.001

Catalysis Today 213 (2013) 211–218
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Enhanced catalytic activity of TAPO-5 in the oxidation of cyclohexene
with hydrogen peroxide under anhydrous conditions
Almudena Alfayate, Carlos Márquez-Álvarez, Marisol Grande-Casas,
∗
Beatriz Bernardo-Maestro, 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:
This work studies the influence of the amount of water accompanying and stabilizing the conventional
oxidant agent H2O2 on the activity of different TAPO-5 materials in cyclohexene oxidation, in comparison
to that of the par excellence large-pore Al-free Ti-beta zeolite. It was proved that the catalytic activity of
a TAPO-5 markedly and systematically increases as the amount of the oxidant H2O2 (30 wt.% in aqueous
solution) decreases, suggesting the strong inhibitor role of water in the catalytic activity of TAPOs. Fol-
lowing this noteworthy result, we have developed a method for removing water from the mixture solvent
Received 17 December 2012
Received in revised form 10 May 2013
Accepted 10 May 2013
Available online 7 June 2013
Keywords:
Hydrophilic/hydrophobic conditions
TAPO-5
Water as inhibitor
Soxhlet
Allylic oxidation
Ti(III)APO-5
(acetonitrile)/H2O2 30 wt.% (aqueous solution) based on a Soxhlet extraction. Water content was quanti-
fied by FTIR, which certified the reproducible reaching of a nearly anhydrous system (5 mg H2O/g CH3CN).
Under these conditions, TAPO-5 catalysts were able to convert cyclohexene in an extension higher that
half of that given by Ti-beta. The product selectivity given by TAPOs and Ti-beta is completely different,
as the former favors the cyclohexene reaction through radicalary mechanism and Ti-beta almost exclu-
sively gives epoxidation products. The catalytic features of the conventional Ti(IV)APO-5 is improved by
the new so-called Ti(III)APO-5 catalyst, synthesized from gels containing Ti(III).
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
framework in contrast to the hydrophobic nature of all-SiO -based
2
materials [13,14]; or (c) the undesired (for redox catalytic purposes)
4
+
The incorporation of Ti ions into zeolite frameworks [1] allowed
acidity [10,12] associated to the most accepted Ti ion incorpo-
microporous materials to extend their applications as heteroge-
neous catalysts to redox processes under mild conditions [2]. Along
the time, potential and/or real industrial applications of Ti-zeolites
ration in P5+ sites of an AlPO4 framework [15], in contrast to the
4+
charge balance arising from the Ti ions incorporation occupying
isoelectronic Si4+ sites. The lower activity of more hydrophilic and
acidic Al-containing Ti-beta against its Al-free homologue zeolite
[10,12] does not conclusively resolve anything in this sense, as it
could be due to the reasons (a), (b) and/or (c). Similarly, Ti-MCM-
41 was not able to reach the catalytic conversions of Ti-zeolites
when the latter have no diffusional problems [5,16]; it is also dif-
ferent from a Ti-zeolite in Ti environment (but not in the siliceous
nature of their first coordination sphere) and in its hydrophilicity
due to its abundant silanols. In any case, all these results sug-
gest that the Ti environment is probably not the only factor that
accounts for the catalytic features of the material in which it is
incorporated. Therefore, it is likely that the low activity of TAPOs
is susceptible of being improved by, for instance, modifying the
hydrophilicity of the catalytic reaction media. In this sense, out-
standing differences in activity have been found by simply changing
different reaction conditions, such as reactants electrophilicity [10]
or solvent nature [17]. Even removal of water from reaction liq-
uid medium, which is one of the subjects of this work, has been
reported to improve the catalytic activity of Ti-MCM-41 catalysts
[18].
[
3,4] and Ti-containing silica-based mesoporous materials [5] in
the field of Fine Chemistry have continued increasing. These cat-
alysts are able to efficiently catalyze a large number of oxidation
reactions, providing the added value of the shape-selectivity [2–4].
That success has not been shared by their homologue AlPO -based
4
microporous materials, in spite of Ti incorporation into the AlPO4
framework is out of any doubt [6–8]. Interestingly, its homologue
Fe or Mn-doped AlPO -5 materials have been shown as active
4
in oxidation reactions under similar conditions to those used for
Ti-zeolite catalysts [9]. This uneven behavior between the dif-
ferent Ti-containing microporous materials as a function of the
framework composition can be attributed to diverse causes: (a)
the different Ti environment [10,11], Ti(OSi)4 being particularly
favorable; (b) the higher hydrophilic character [11,12] of AlPO4
∗ Corresponding author. Tel.: +34 915854795; fax: +34 915854760.
E-mail addresses: manuel.sanchez@icp.csic.es, manuel.sanchez.sanchez@urjc.es
M. Sánchez-Sánchez).
(
0
920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.05.001

2
12
A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
The number of publications about TAPOs as catalysts in oxi-
one hour and the final gel was introduced in Teflon-lined stainless-
steel autoclaves and subsequently hydrothermally treated at 448 K
under autogeneous pressure for 18 h. The resultant purple solid,
whose color indicated that at least part of Ti present in the solid
maintain its oxidation state 3+ after crystallization process, was
recovered by filtration and was washed with deionized water. Once
dried, the solid became yellowish white, due to the complete oxi-
dation of Ti to its oxidation state 4+.
Conventional catalysts, Ti(IV)APO-5 and Ti-beta, were prepared
strictly following the procedures described elsewhere [27,28]
with Ti content either having showed high catalytic activity [27]
or specifically designed to have a Ti content similar to that
of Ti(III)APO-5, respectively. In order to clarify the differences
between the samples Ti(III)APO-5 and Ti(IV)APO-5, a detailed
description of the synthesis procedure for preparing Ti(IV)APO-5
is also described next. Ti(IV)APO-5 was prepared from a gel with a
dation reactions is rather scarce [13,19–21]. It is surprising that,
chronologically speaking, the first publication of them is not about
a proper TAPO catalyst but about Ti-containing SAPO-5 catalyst
(
the so-called TAPSO-5 [13]) attempting to imitate the siliceous
Ti environment of Ti-zeolites. The TAPSO approach in principle
means the reduction at some extent of every above-mentioned pos-
sible sources of the low activity of TAPOs, as it would make more
siliceous the Ti environment, increase the hydrophobic character
of its framework, and decrease the acidity directly associated to
the Ti incorporation in P sites. Unfortunately, that study was not
able to shed light on the ultimate cause responsible of the low cat-
alytic activity of TAPOs as TAPSO activity was not compared to that
of any of the two pertinent catalysts: its homologue Si-free TAPO-
5
and a Ti-zeolite of a similar pore openings (large pore defined
by 12-membered rings). The same is applicable for the study of
catalytic activity of Ti-substituted SAPO-5 and Si-VPI-5 tested in
phenol hydroxylation [20].
molar composition of 1.0 Al:1.0 P:0.02 Ti (IV):0.5 TPAOH:20 H O,
2
using titanium tetraisopropoxide as Ti source. First, pseuboehmite
(Catapal B 73.56 wt.% Al O ) was added over an aqueous solution
Recently, the combination of different advanced characteriza-
tion techniques applied to the study of Ti environments in TAPOs
has renewed the interest for these materials [7,8,22]. It has been
2
3
of phosphoric acid (85%, Aldrich) in deionized water. Then, tita-
nium isopropoxide (Aldrich) and subsequent TPAOH as SDA were
added dropwise. After 1 h of vigorous stirring, the homogeneous gel
was transferred into a Teflon-lined stainless-steel autoclave, which
4
+
suggested that, unlike Ti incorporation in Ti-zeolites but just like
4
+
4+
Si incorporation in an AlPO4 framework, Ti ions have differ-
ent incorporation mechanisms including that by pairs (known as
mechanism SIII in SAPO materials [23]), which implies the exist-
◦
were hydrothermally treated at 175 C for 192 h.
The most extended nomenclature for denoting the here-
called Ti(IV)APO-5 is TAPO-5 (sometimes, TiAPO-5). However,
in this article, we will use the term TAPO-5 to encompass
both Ti(III)APO-5 and Ti(IV)APO-5. The final composition of the
inorganic frameworks as determined by ICP analyses were:
Ti0.026Al0.497P0.477O2 for Ti(III)APO-5, Ti0.007Al0.521P0.473O2 for
Ti(IV)APO-5 and Ti0.017Si0.983O2 for Ti-beta.
ence of Ti
O Ti bonds into the framework of the conventional
TAPO materials.
We have been developing a research project aimed to con-
tribute to the knowledge behind the much lower catalytic activity
of the conventional TAPOs against Ti-zeolites, and accordingly to
design strategies to make TAPOs competitive oxidation catalysts.
Our approaches are based on the rational design of both new Ti
environments in the catalyst and of the reaction conditions. The
generation of new Ti environments in AlPOs as well as their cat-
alytic implications is deeply discussed elsewhere [24,25], and a
typical catalyst resultant of applying that strategy is also used in the
current article (denoted as Ti(III)APO-5), which has been prepared
from gels containing Ti(III) ions instead of the conventional Ti(IV)
ones. This work focuses on the effect of reaction conditions, partic-
ularly water content of the oxidant agent H O and reaction media,
Before catalytic tests SDA molecules were eliminated from all
the samples by calcination at 823 K for 5 h under an air flow of
−
1
100 mL min . Previously the samples had been heated at a rate of
−
1
−1
3 K min under a N2 flow of 100 mL min up to 823 K and main-
tained at this temperature for 1 h under these same conditions.
The nature of crystalline phases was determined by pow-
der X-ray diffraction (PXRD) patterns, collected by a PANalytical
X’Pert Pro diffractometer using Cu K␣ radiation. UV–visible dif-
fuse reflectance spectra were registered on a Cary 5000 Varian
spectrophotometer equipped with an integrating sphere and the
obtained data were converted to corresponding spectra expressed
according to Kubelka–Munk function.
2
2
on the catalytic behavior of both conventional TAPO-5 and the so-
called Ti(III)APO-5. This study presents some singularities respect
to the previous publication studying the effect of water removal in
the activity of Ti-MCM-41 in similar reactions but using as-received
‘
anhydrous’ TBHP [18]: the magnitude of water to remove is much
2.2. Water extraction from acetonitrile–H O 30 wt.% mixture.
2
2
higher in our case, H O decomposes in direct contact to the Molec-
Quantification of H O and H O
2
2
2
2
2
1
ular Sieve, and the characterization technique used in ref. 18,
H
NMR, has intrinsic limitations for quantifying H O and H O from
Some catalytic experiments were carried out over an
2
2
2
a mixture solution [26].
acetonitrile–H O2 30 wt.% mixture, after water extraction in a
2
Soxhlet system. Molecular Sieve 3A was used as adsorbent selective
to water, being able to discriminate between that small molecule
and the more volatile CH CN and less volatile H O ones. To
2
. Experimental
3
2
2
perform the experiment, 30 g of Molecular Sieve, previously dehy-
drated at 453 K, were wrapped within a filter paper bag and placed
2.1. Catalysts preparation and characterization
in the Soxhlet thimble whereas a mixture solution of CH CN and
3
Ti(III)APO-5 sample was synthesized [24,25] by hydrothermal
treatment of a gel with the following molar composition: 0.96
H O 30 wt.% aqueous solution in a proportion of 30 mmol of H O
2
2
2
2
per mol of CH CN was put into the still pot. (Due to the risks of
3
Al:1.0 P:0.04 Ti(III):0.8 MCHA:25 H O. Both the preparation of the
heating concentrated hydrogen peroxide, especially in the presence
of organic compounds, it is not recommendable to carry out this
2
gel and the subsequent autoclave sealing were carried out under
N2 atmosphere. In a typical gel preparation, TiCl3 aqueous solu-
tion (∼10 wt.%) in HCl (20–30 wt.%) (supplied by Aldrich with the
exact composition) was added over a solution of phosphoric acid
experiment in absence of acetonitrile or with low acetonitrile/H O2
2
ratio). The amount of Molecular Sieve 3A was chosen supposing a
theoretical water adsorption capacity of 15 wt.% (real adsorption
of ∼20 wt.%). The mixture was heated up to 349 K under mag-
netic stirring. Small aliquots were taken out from the still pot after
every Soxhlet extraction cycle and they were analyzed by FTIR in
(
85%, Aldrich) in deionized water turning it into deep purple color.
Next, Al(OH) ·xH O (Sigma) and straightaway the organic structure
3
2
directing agent (SDA) N-methyldicyclohexylamine (MCHA) were
added under vigorous stirring. The resulting mixture was stirred for
order to quantify H O2 and the remaining amount of water until a
2

A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
213
minimum constant value was reached. H O2 concentration in the
2
final solution was also determined by iodometric titration. Follow-
ing this procedure, it was possible to avoid the H O decomposition,
2
2
which takes place at higher concentrations, thanks to the dilu-
tion in acetonitrile. Substantial H O2 decomposition was detected
2
when Molecular Sieve was present into the mixture even at room
temperatures. This fact as well as the potential interference of the
Molecular Sieve in the catalytic activity, forced us to remove water
in a step previous to the reaction tests.
Attenuated total reflection-Fourier transform infrared (ATR-
FTIR) spectra of liquid samples were recorded using a Thermo
Nicolet Nexus 670 FTIR spectrometer equipped with a SensIR
Technologies DurasamplIR horizontal ATR accessory and a liquid
nitrogen-cooled MCT detector. The internal reflection element (IRE)
used was a diamond-faced ZnSe prism, mounted in a stainless steel
◦
trough plate, the incidence angle was 45 and total number of
reflections was 9. The trough above the ATR crystal was ca. 4 mm in
diameter and had an approximate capacity of 50 ␮L. For every anal-
ysis, ca. 10 ␮L of sample solution was deposited on the crystal, and
a cover was used to avoid solvent evaporation during acquisition of
spectrum. Each spectrum was obtained as the average of 128 scans
with 4 cm resolution in the 4000–650 cm wavenumber range.
Infrared spectra were recorded at room temperature and normal-
ized against a background spectrum acquired using the bare IRE
crystal.
Fig. 1. PXRD patterns of the as-synthesized samples Ti(III)APO-5, Ti(IV)APO-5 and
Ti-beta.
−1
−1
3. Results and discussion
3.1. Materials characterization
2.3. Catalytic reaction conditions
Fig. 1 shows the powder X-ray diffraction (PXRD) patterns of
the as-synthesized materials that, once calcined, were tested as
catalysts in the oxidation of cyclohexene. The patterns are the typ-
ical ones of the BEA and AFI-structured materials. No signal of any
impurity was detected in any of the patterns. The slight differences
in relative intensities and/or peak broadening between the patterns
of both TAPO-5 must be attributed to small differences in crystal
morphology and crystal size, respectively. Calcination process did
not lead to the appearance of any extra phase (PXRD patterns not
shown).
Cyclohexene oxidation reactions were performed in batch mode
using a round bottom flask immersed in a temperature controlled
bath and equipped with a thermometer, a magnetic stirrer and a
reflux condenser provided with a drying tube on the top. All these
items were completely dried before the reaction. Circulating water
through the condenser is maintained at 278 K in order to avoid any
(
selective) evaporation of the chemicals from the reaction mixture,
which has been proved to be meaningful, particularly in the case of
cyclohexene, if simply running tap water is used. So cool circulation
causes an important water condensation in the condenser coming
from ambient humidity, what is basically prevented with the dry-
ing tube. The catalysts were previously activated overnight in the
same reaction flask at 433 K under flowing N2 in order to remove
any contained water. Under these conditions, recovery of reaction
mixture from a blank reaction (in absence of catalyst) was over 98%
after 30 h of reaction.
The aim of using Ti(III) or Ti(IV) sources was to favor the incor-
poration of Ti ions in Al(III) or P(V) positions of a theoretical AlPO4
framework, respectively [24,25]. The purple color of as-crystallized
Ti(III)APO-5 would suggest a higher probability of isomorphic sub-
stitution of Al(III) by Ti ions. The posterior sample color turning to
white during drying or calcination processes, implies the whole oxi-
dation of Ti to Ti(IV) but it is foreseeable that Ti ions do not migrate
and therefore they should maintain their framework ion neigh-
bors. Fig. 2 shows the 200–400 nm region of the diffuse-reflectance
UV–visible spectra of the calcined Ti(IV)APO-5, Ti(III)APO-5 and Ti-
beta. The maximum of the O → Ti charge transference transition
is different for each sample. In good agreement with the litera-
ture [20,29], that transition takes place at lower wavelength in
the case of Ti-beta (maximum at 204 nm) in comparison with the
TAPO-5 materials. That maximum is also slightly different in the
spectra of the conventional Ti(IV)APO-5 (maximum at 233 nm) and
Ti(III)APO-5 (225 nm), suggesting that effectively some modifica-
tion in the Ti environment has been afforded with our new method
[24,25]. Indeed, for the first time an O → Ti charge transference
band is centered below 230 nm for a calcined TAPO-5 material,
whose band usually appears at higher wavelength than that of
the as-prepared one. The sense of the maxima shift is the right
one for catalytic purposes, assuming that at least part of the high
activity of Ti-beta is due to its Ti environment characterized by hav-
ing O → Ti transitions at low wavelength. In any case, the spectra
of both TAPO-5 samples follow a relatively similar pattern from
which that of Ti-beta is away. Apart from the differences in max-
ima, the bands are quite large and asymmetric, having a shoulder
at higher wavelength (at ca. 280 nm) which is more pronounced in
In a typical experiment, 3.0 g of cyclohexene, CH CN and H O
3
2
2
(
or, instead of both, the resultant mixture of CH CN and H O
3 2
2
from Soxhlet extraction if the reaction was carried out under water
controlled conditions) in molar ratio CH CN:H O :cyclohexene of
3
2
2
2
0:0.55:1 and toluene as the internal standard (and, if necessary,
extra acetonitrile as solvent to complete the CH CN/cyclohexene
3
ratio of 20 in the case of previously Soxhlet-treated mixtures)
were added over the activated catalysts (cyclohexene/catalyst
weight ratio of 10). The mixture was heated under stirring up
to 343 K and when the reaction temperature was reached, an
aliquot was taken as time zero. Successively small aliquots were
taken at different reaction times and catalyst particles sepa-
rated by filtration. Samples were analyzed by gas chromatography
using a Varian CP-3380 gas chromatograph equipped with a 15 m
length, 0.25 mm i.d. and 1 ␮m thickness dimethylpolysiloxane film
Factor Four (Varian VF-1 ms) capillary column and a flame ion-
ization detector (FID). Reactants and products were identified
by mass spectrometry, infrared and 1H and C NMR spectro-
scopies (Supplementary Information). Cyclohexene conversion was
calculated but from the corresponding yields of the resultant prod-
ucts.
13

2
14
A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
2
1
1
0
5
0
5
0
Oxidant/cyclohexene molar ratio - % H O in oxidant
2
04 nm
2
2
25 nm
0
.55 H O - anhydrous
2
33 nm
2 2
Ti(IV)APO-5
1.1 TBHP - anhydrous
0
.55 H O - 70 %
2
2
Ti-beta
1
.1 H O - 70 %
2
2
2
.2 H O - 70 %
Ti(III)APO-5
2 2
0
5
10
15
20
25
30
t / h
Fig. 3. Kinetics of cyclohexene conversion in presence of the Ti(III)APO-5 catalyst
under different conditions of oxidant/cyclohexene ratio, water content and oxidant
nature. These conditions are explicitly shown in the figure for each experiment.
2
00
250
300
350
400
Wavelength / nm
Fig. 2. UV region of the diffuse-reflectance UV–visible spectra of the calcined sam-
ples Ti(III)APO-5, Ti(IV)APO-5 and Ti-beta. (For interpretation of the references to
spectra in figure legend, the reader is referred to the web version of the article.)
(mainly its hydration degree) on the catalytic activity for a given
catalyst. Ti(III)APO-5 produces very low and kinetically slow con-
version of cyclohexene when H O2 30 wt.% is the oxidant agent
2
and starting from a H O /cyclohexene ratio of 1.1 (circles in Fig. 3).
2
2
the spectrum of the Ti-richer Ti(III)APO-5 sample. The more recent
studies attributes that shoulder to the presence of more than one
Ti environment into the AlPO4 framework [7,8,23].
A deeper characterization of the materials is out of the scope
of this article. The reader is remitted to the literature for detailed
characterization of the samples Ti-beta [11,28] and Ti(IV)APO-5
There is a very long induction period and, once it is surpassed, the
conversion of cyclohexene continues being poor, scarcely reaching
4% after 30 h. Despite we expected an improvement by decreasing
water from the system; it is still paradoxical that the conversion
increases by reducing the amount of oxidant agent H O2 (and
2
consequently also water) (square symbols) and that it diminishes
when the oxidant increases (triangles). Obviously, it means that
the inhibitor role of water is much more powerful than the kinetic
advantages arisen from increasing the amount of one of the reac-
tants. Another indication of that strong inhibitor role is given by
the experiment in which commercial anhydrous TBHP is used as
oxidant agent (stars). In spite of its much higher kinetic diameter
compared to H2O2, which one could expect to be a key factor in
a 1D pore system microporous catalyst, it is able to oxidize cyclo-
hexene more effectively than H2O2 of a concentration of 30 wt.%
in water, at any of the above-mentioned H2O2/cyclohexene ratios.
Besides, unlike the experiments with H2O2 30 wt.%, the conversion
of cyclohexene with TBHP lacks induction period, undoubtedly due
to the absence of any inhibitor species. The even better conversion
obtained with H2O2 as oxidant agent under anhydrous conditions
(see below) (diamonds in Fig. 3) does ratify the inhibitor role of
water for Ti(III)APO-5 catalyst.
[
19,27], which herein have been prepared by exactly following the
published procedures. Regarding the sample Ti(III)APO-5, a deep
characterization of its Ti sites is presented and discussed elsewhere
[
24,25].
3.2. Influence of water in catalytic activity of Ti-beta and TAPO-5
catalysts
It has been already mentioned the outstanding influence of the
solvent nature in the catalytic behavior of a given Ti-zeolite. Rather
than the electrophilic or the protic/aprotic nature of the solvent, its
hydrophobic character seemed to be particularly influential [10].
Accordingly, it is expected to afford higher effect by modifying the
amount of water present in the commercial samples of the oxidant
agent used in this reaction, H O . Moreover, such effect should be
2
2
more pronounced in hydrophilic AlPO -based materials, in which
4
H O will be preferably adsorbed against H O , than in hydrophobic
2
2
2
Al-free Ti-containing zeolites, where H O will not become a seri-
3.3. Water removal from the reaction mixture
2
ous adsorbent competitor of H O . In this context, it is required
2
2
to optimize the reaction conditions in order to exploit the real
potential of the intrinsic activity of Ti sites beyond the limitations
attributable to the framework nature. Otherwise, the comparison
Although the afforded activity of this TAPO-5 catalyst with com-
mercial anhydrous TBHP oxidant could be considered as acceptable
in some aspects, any potential application of TAPOs against Ti-
between the catalytic activity of Ti-containing AlPO -5 and Ti-beta
zeolites must necessarily imply the use of H O2 as oxidant agent
4
2
catalysts would be based on two aspects: (i) the influence of water
could be presumably of different magnitude for the two tested cat-
alyst frameworks; (ii) the selectivity of both kind of catalysts would
be hardly comparable if the conversions become as different as
expected. Consequently, as a previous step of our work, we will
optimize the activity of a TAPO-5 by regulating the hydrophobicity
of the reaction media.
because of its much more viable price, in addition to its already-
commented lower diffusion problems within the AlPO framework.
4
Moreover, TBHP can be supplied with an undetermined amount of
water [18]. In this context, we have made an effort for developing
the method described in experimental part in order to remove H O
2
from the reaction system as a previous step to the proper catalytic
reaction.
Fig. 3 shows the evolution with time (up to 30 h) of the
conversion of cyclohexene in presence of different amounts
and concentrations of H O in aqueous solution using the
Ti(III)APO-5 catalyst. An experiment in which anhydrous tert-
butylhydroperoxide (TBHP) was used as oxidant agent is also
included in Fig. 3. At first glance, the most remarkable feature of
that figure is indeed the huge influence of the oxidant conditions
The solutions initially formed by as-received commercial ace-
tonitrile and commercial hydrogen peroxide 30 wt.% in aqueous
solution were treated in a Soxhlet containing Molecular Sieve 3A
and were analyzed by FTIR in order to evaluate the extent of
water removal along the water extraction process. Aliquots of the
solutions were withdrawn from the still pot after every Soxhlet
extraction cycle and their FTIR spectra were recorded using the ATR
2
2

A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
215
0
.1
e
d
c
b
a
3500
3000
2500
2000
1500
1000
-1
Wavenumber / cm
Fig. 5. Calibration curves for water and hydrogen peroxide elaborated from relative
integrated absorbances of their representative infrared bands (A1632 for water and
A869 for hydrogen peroxide) respect to that of acetonitrile (A919) determined from
ATR-FTIR spectra of standard solutions of water or hydrogen peroxide in acetonitrile.
Fig. 4. ATR-FTIR spectra of acetonitrile (a) and acetonitrile solutions containing
0 mg H2O per g CH3CN (b) and 71 mg of 30 wt.% H2O2 aqueous solution per g CH3CN
c). The spectra of the last sample after 1 and 7 water extraction cycles are shown
2
(
in (d) and (e), respectively.
From the linear regression of these data, a slope of 44.7 mg H O/g
2
cell. Fig. 4 shows the ATR-FTIR spectrum of the initial mixture (spec-
CH CN per unit of the ratio A1632/A919 was obtained. In the case
3
trum c) in which the H O2 concentration should be of ca. 30 mmol
2
of hydrogen peroxide, the mass ratio vs. the ratio of integrated
absorbances A869/A919 shows some deviation from linearity in the
H O per mol CH CN. For reference purposes, the spectra of the
2
2
3
commercial acetonitrile (Fig. 4a) and a solution containing 20 mg
of water per gram of acetonitrile (Fig. 4b) are also shown. Water
can be unambiguously identified in these spectra by the two broad,
0
–40 mg H O /g CH CN range (Fig. 5). Therefore, data were fitted
2 2 3
to a second degree polynomial equation.
Using the above calibration curves, the evolution of water
and hydrogen peroxide concentration during the water extraction
treatment of solutions prepared from hydrogen peroxide aqueous
solution and acetonitrile has been determined from the series of
ATR-FTIR spectra recorded along the treatment. In Fig. 6, the calcu-
lated water/acetonitrile and hydrogen peroxide/acetonitrile ratios
have been plotted as a function of the number of Soxhlet extraction
cycles for one of the samples prepared. For comparison purposes,
the hydrogen peroxide to acetonitrile ratio determined by titra-
tion has also been plotted for the initial solution and after seven
extraction cycles. Hydrogen peroxide concentrations determined
by the ATR-FTIR method are in close agreement with those deter-
mined by titration. Fig. 6 shows that, after the first extraction cycle,
−
1
partly overlapping bands at 3621 and 3543 cm , corresponding to
−
1
O
H stretching modes, and the HOH bending band at 1632 cm
.
The presence of these bands in the spectrum 4a shows that the
commercial acetonitrile contains a certain amount of water, which
was taken into account to determine the total content of water
of the standard solutions used to carry out the quantitative anal-
ysis based on ATR-FTIR spectra (see below). Hydrogen peroxide
−
1
is characterized by a broad, strong band at 3405 cm associated
−
1
to O H stretching, the broad HOOH bending band at 1335 cm
,
which appears as a shoulder on the strong acetonitrile CH defor-
3
−1
mation band, and the weak band at 869 cm that corresponds to
the O O stretching mode.
When the solution prepared by mixing the hydrogen peroxide
aqueous solution and acetonitrile undergoes the water extraction
treatment, a progressive decrease of intensity of the water bands
occurs (spectra d and e in Fig. 4) evidencing the successful removal
of water. On the other hand, the spectra indicate that hydrogen
peroxide remains stable during the treatment as the intensity of its
bands is maintained.
In order to carry out quantitative determinations of water and
hydrogen peroxide in acetonitrile solutions, ATR-FTIR spectra of
standard solutions containing known amounts of hydrogen perox-
ide and/or water in acetonitrile were recorded. The corresponding
calibrations were elaborated from the integrated absorbances of
selected bands representative of acetonitrile, water and hydrogen
peroxide. Narrow and non-overlapping bands were chosen to facil-
itate an accurate determination of integrated absorbances. Thus,
−
1
the hydrogen peroxide ␯O O band (869 cm ), the water ıHOH
band (1632 cm−1) and the acetonitrile ␯C C band (919 cm ) were
selected as representative of each compound. In Fig. 5, the water
to acetonitrile mass ratio has been plotted vs. the ratio of inte-
grated absorbances of water and acetonitrile bands (A1632/A919) for
a set of standard solutions of water in acetonitrile. As shown in
Fig. 5, there exists a linear relationship between the ratio of inte-
−1
Fig. 6. Evolution of water/acetonitrile mass ratio (open squares) and hydrogen per-
oxide/acetonitrile mass ratio (open circles) determined by ATR-FTIR analysis of a
solution of hydrogen peroxide aqueous solution in acetonitrile along the Soxhlet
extraction cycles. The hydrogen peroxide/acetonitrile mass ratios determined by
titration are shown as full circles.
grated absorbances of water and acetonitrile bands (A1632/A919
)
and the water to acetonitrile mass ratio up to 50 mg H O/g CH CN.
2
3

2
16
A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
3
2
2
1
1
0
5
0
5
0
5
0
catalyst, then attributable to the inhibitor role of water. Since
water is basically absent under these reaction conditions, the still
remaining induction period could be (at least partially) related to
diffusional problems proper of the 1D-pore dimensionality of the
AFI materials against a much more permeable 3D-pore beta struc-
ture. Moreover, Fig. 7 shows a substantially higher conversion for
Ti(III)APO-5 than Ti(IV)APO-5. Although the different Ti content of
the TAPO-5 samples could explain such difference in conversion,
a careful examination of the evolution of conversion with time in
Fig. 7 evidences that again the induction period is the main cause
of that difference. These differences in the induction period can-
not be attributed to diffusional problems, since both materials are
isostructural and moreover, the size of the polycrystalline aggre-
gates of Ti(IV)APO-5 (of ca. 20 ␮m) are smaller than those forming
Ti(III)APO-5 (ca. 33 ␮m) (SEM pictures not shown). It means that
Ti centers convert cyclohexene much quicker in Ti(III)APO-5 than
in Ti(IV)APO-5, suggesting the different nature of the catalytically
active centers. A further discussion about the differences in both Ti
environment and catalytic activity of the samples Ti(IV)APO-5 and
Ti(III)APO-5 is given elsewhere [24,25].
Ti-beta
Ti(III)APO-5
Ti(IV)APO-5
0
5
10
15
t / h
20
25
30
Fig. 7. Kinetics of cyclohexene conversion in presence of the Ti(III)APO-5 (squares),
Ti(IV)APO-5 (circles) and Ti-beta (triangles) catalysts under anhydrous conditions.
Fig. 8 shows the cyclohexene conversion and the yields of
the products generated by different reaction mechanisms for
Ti(III)APO-5 (left), Ti(IV)APO-5 (middle) and Ti-beta (right). These
mechanisms and their corresponding detected products are sum-
marized in Scheme 1. It is well known that the oxidation of
the hydrogen peroxide/acetonitrile ratio shows an increase of 27%.
This increase can be explained considering that the solution that
remains in the still pot after an extraction cycle contains around
cyclohexene with H O2 can follow either (1) a homolytic mech-
2
3
0% less acetonitrile than the initial solution because part of the
anism governed by free-radical species, leading to the allylic
oxidation products, or (2) a heterolytic mechanism in which an
solvent is retained in the initially-empty Soxhlet chamber and con-
denser at the end of every extraction cycle. The ATR-FTIR analysis of
the sample taken after the first extraction cycle shows a 65% lower
water/acetonitrile ratio compared to the starting solution. Bearing
in mind the above mentioned decrease in the amount of acetoni-
trile, this result implies that the first extraction cycle removes
ca. 75% of the water present in the initial solution. Subsequent
extraction cycles render progressively lower water removal yields
and the water/acetonitrile mass ratio reaches a nearly constant
value, close to 5 mg H O/g CH CN (i.e., around 90% water removal)
oxygen transfer from H O2 to cyclohexene promotes the forma-
2
tion of the corresponding epoxide. For all three tested catalysts,
the epoxidation of cyclohexene (with or without subsequent epox-
ide aperture) is the unique mechanism occurred at the very first
steps of the reaction (diamonds in Fig. 8). In other words, the mag-
nitude of epoxidation process makes the difference between the
conversion of Ti-beta and TAPO-5 catalysts during the first hours
of reaction.
On the other hand, cyclohexene conversion through allylic
mechanism (Scheme 1), in which cyclohexene reacts through
a carbon atom adjacent to the double bond, begins at shorter
times in the case of TAPO-5 catalysts. That time is particularly
short for Ti(III)APO-5 (2 h), and in the next studied time (4 h) it
becomes the most abundant reaction product. This product is firstly
detected after 6 h of reaction in presence of Ti(IV)APO-5 and after
2
3
after 6 extraction cycles. That ‘irremovable’ amount of water is
lower than that found in the as-received acetonitrile (Fig. 4a
and e), so these reached experimental conditions are more anhy-
drous than those using a full-anhydrous oxidant in acetonitrile.
The hydrogen peroxide to acetonitrile mass ratio shows a small
decrease as the number of cycles increases, which indicates that
this water extraction method produces little hydrogen peroxide
decomposition.
The reaction medium used in the experiment under anhydrous
conditions with Ti(III)APO-5 as catalyst (diamonds in Fig. 3) was
that resulting after seven extraction cycles. The relatively high
activity of Ti(III)APO-5 catalyst in such reaction medium, exceed-
ing that of the same catalyst when double molar ratio of anhydrous
TBHP was used as oxidant agent (black stars), justifies the effort
8
h in presence of Ti-beta. In good agreement with Scheme 1,
the occurrence of the homolytic mechanism was detected by
the presence of 2-cyclohexenyl hydroperoxide, also known as
3
-hydroperoxy-1-cyclohexene, (stars in Fig. 8). The identification
of that non-commercial product was based on mass spectrometry
and IR and 1H and 13C NMR spectroscopies (Supporting Informa-
tion). The other two products generated by homolytic mechanism
were identified as 2-cylohexen-1-one and 2-cyclohexen-1-ol. Their
formation as products of the cyclohexene oxidation has been
proposed to take place through a subsequent transformation
of 2-cyclohexenyl hydroperoxide [30], what explains their later
appearance in our system. That transformation is given at minor
extension, and consequently a high selectivity to peroxide is
reached. Such high selectivity toward this peroxide is uncommon
in the oxidation of cyclohexene. Apart from the academic inter-
est of this singular catalytic behavior, the high selectivity to an
isolable non-commercial product that is a potential intermediate
of the industrially-demanded products such as 2-cyclohexen-1-one
[31] or even cyclohexanone [32], gives a substantial importance to
these results. The subsequent selective reaction of the intermediate
peroxide to any of these products could be carried out either by a
new process or in the same pot by using a co-catalyst or simply a
bifunctional catalyst.
made for making anhydrous the acetonitrile–H O2 system.
2
3.4. Cyclohexene oxidation with Ti(III)APO-5, Ti(IV)APO-5 and
Ti-beta catalysts under anhydrous conditions
Once optimized the reaction conditions for having a close to
anhydrous reaction media, we carried out the reaction tests with
our three catalysts under these conditions. The cyclohexene con-
versions, determined by the sum of the detected products, are
plotted against the reaction time for the three catalysts in Fig. 7. In
spite of the reactions conditions are adequate for favoring TAPO-5
catalysts, the catalytic activity continues being significantly higher
for Ti-beta. The differences in activity did mainly take place at
the first stages of the reaction, despite Fig. 3 showed an outstand-
ing improvement in limiting the induction period for Ti(III)APO-5

A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
217
20
15
10
5
0
30
Ti(III)APO-5
Ti(IV)APO-5
Ti-beta
25
20
15
10
5
0
1
0
5
0
Epoxidation
Allylic oxidation
Peroxide
0
10
20
30
0
10
20
30
0
10
20
30
t / h
t / h
t / h
Fig. 8. Kinetics of global cyclohexene conversion (full symbols) for Ti(III)APO-5 (left-hand side, squares), Ti(IV)APO-5 (middle, circles) and Ti-beta (right-hand side, triangles)
catalysts, and cyclohexene conversion divided in different reaction mechanisms (open symbols): epoxidation (diamonds), and allylic oxidation (hexagons). To remark the
high selectivity to cyclohexenyl hydroperoxide amongst the products obtained through allyllic mechanism, it is also plotted (as empty starts).
Despite the formation of the 2-cyclohexenyl hydroperoxide had
been often catalyzed by homogeneous compounds, it does not nec-
essarily mean that Ti leaching from our heterogeneous catalysis is
taken place. Indeed, we discard that possibility because no traces
of Ti were detected at all in the reaction mixtures by X-ray fluo-
rescence technique, which however was able to correctly detected
yields of that product. In addition, the selectivity to the global
amount of products formed by heterolytic mechanism (that is, the
sum of hydroperoxide and products through allylic oxidation) is
higher for Ti(III)APO-5 (84% after 30 h) than for Ti(IV)APO-5 (81%).
In short, our Ti(III)APO-5 catalyst accentuates all the catalytic fea-
tures of Ti(IV)APO-5, which could make this family of catalysts an
alternative to Ti-zeolites for obtaining certain products, particularly
those formed through C-H bond activation at allylic positions.
0.1 ppm of a solution deliberately prepared of that concentration
from an ICP standard solution.
Although compared to Ti-beta, both tested TAPO-5 materials
follow a similar catalytic behavior pattern, there are some dif-
ferences between them, which could be crucial to evaluate the
success of the strategy for preparing a TAPO-5 material (Ti(III)APO-
4
. Conclusions
The catalytic activity of a conventional TAPO-5, here denoted
5
) from gels containing Ti(III) instead of Ti(IV) sources. Thus,
as Ti(IV)APO-5, and a so-called Ti(III)APO-5 catalysts in the oxida-
tion of cyclohexene with commercial hydrogen peroxide 30 wt.% in
aqueous solution has been investigated. By systematic variation of
H O 30 wt.%/cyclohexene ratio, it was found that cyclohexene con-
version decreases with the increase of the content of H O2 30 wt.%
in the mixture. Consequently, a strong inhibitor role of water in
this reaction was deduced and explained in terms of the higher
we have already commented the higher cyclohexene conversion
reached by Ti(III)APO-5. Furthermore, compared to the conven-
tional Ti(IV)APO-5 catalyst, it is able to generate higher amounts
of products from any of the considered mechanisms. Even the
above-remarked high selectivity of Ti(IV)APO-5 for 2-cyclohexenyl
hydroperoxide (up to 76%) is scarcely superior to that given by
Ti(III)APO-5 (70%), and even the latter generates higher absolute
2
2
2
hydrophilic character of the AlPO -based microporous framework.
4
In order to explore the real catalytic potential of TAPOs under
favorable conditions, we developed both an experimental proce-
dure and an analytical FTIR-based methodology for reproducible
OH
O
OOH
water removal and water-H O2 quantification, respectively, from
2
the system acetonitrile-H O2 30 wt.%. Under these conditions, the
2
catalytic activity of TAPOs increases, becoming approximately half
of that given by an Al-free Ti-beta zeolite. Besides, the selectiv-
ity is very much different. Whereas Ti-beta basically epoxides
cyclohexene, in presence of TAPOs the through-radical mecha-
nism dominates the cyclohexene oxidation, with a high selectivity
to the intermediate 2-cyclohexenyl hydroperoxide. Ti(III)APO-5
resulted more active and more selective to allylic transformation
of cyclohexene than the conventional Ti(IV)APO-5, so the former
is postulated as a good candidate for further researching of this
catalytic alternative to the cyclohexene epoxidation by Ti-zeolites.
(
1)
2)
H O
2
2
catal.
(
OH
O
O
H O
2
OH
OH
Acknowledgments
The authors are grateful to Dr. Enrique Sastre and Prof. Félix
Sánchez for their valuable help in the identification of the products
by mass spectrometry and GC, as well as Dr. José Miguel Cam-
pos and Dr. Marta Iglesias for some GC-MS measurements. They
Scheme 1. Proposed reactions for the generation of the identified products in the
catalytic conversion of cyclohexene with H2O2 under anhydrous conditions in pres-
ence of Ti(III)APO-5, Ti(IV)APO-5 or Ti-beta catalysts.

2
18
A. Alfayate et al. / Catalysis Today 213 (2013) 211–218
also acknowledge Spanish Government for financial support (MAT-
009-13569 and MAT-2012-31127). A. Alfayate acknowledges CSIC
for a PhD JAE-predoc fellowship. B. Bernardo-Maestro acknowl-
edges Spanish Society of Catalysis (SECAT) for an ‘Introduction to
Research’ grant.
[13] A. Tuel, Zeolites 15 (1995) 228–235.
[
[
[
[
[
14] M. Sanchez-Sanchez, G. Sankar, L. Gomez-Hortiguela, Microporous and Meso-
porous Materials 114 (2008) 485–494.
15] M.H. Zahedi-Niaki, S.M. Javaid Zaidi, S. Kaliaguine, Microporous and Meso-
porous Materials 32 (1999) 251–255.
16] A. Corma, M.T. Navarro, J. Perez-Pariente, Journal of the Chemical Society,
Chemical Communications (1994) 147–148.
17] A. Corma, P. Esteve, A. Martinez, Journal of Catalysis 161 (1996)
11–19.
2
Appendix A. Supplementary data
18] A. Corma, M. Domine, J.A. Gaona, J.L. Jorda, M.T. Navarro, F. Rey, J. Perez-
Pariente, J. Tsuji, B. McCulloch, L.T. Nemeth, Chemical Communications (1998)
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.cattod.
2211–2212.
[19] M.H. Zahedi-Niaki, M.P. Kapoor, S. Kaliaguine, Journal of Catalysis 177 (1998)
231–239.
2
013.05.001.
[
20] B.Y. Hsu, S. Cheng, J.M. Chen, Journal of Molecular Catalysis A: Chemical 149
1999) 7–23.
(
References
[21] S.O. Lee, R. Raja, K.D.M. Harris, J.M. Thomas, B.F.G. Johnson, G. Sankar, Ange-
wandte Chemie International Edition 42 (2003) 1520–1523.
[22] S. Maurelli, M. Chiesa, E. Giamello, R.M. Leithall, Chemical Communications
(2012) 8700–8702.
[23] L. Sierra de Saldarriaga, C. Saldarriaga, M.E. Davis, Journal of the American
Chemical Society 109 (1987) 2686–2691.
[24] A. Alfayate, M. Sanchez-Sanchez, J. Perez-Pariente, ES Patent
PCT/ES2012/070436 (2011).
[25] A. Alfayate, M. Sanchez-Sanchez, J. Perez-Pariente, (manuscript in preparation).
[26] N.A. Stephenson, A.T. Bell, Analytical and Bioanalytical Chemistry 381 (2005)
1289–1293.
[27] A.M. Prakash, L. Kevan, M.H. Zahedi-Niaki, S. Kaliaguine, Journal of Physical
Chemistry B 103 (1999) 831–837.
[28] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, A. Martínez, C. Prieto, S. Valencia,
Chemical Communications (1996) 2367–2368.
[29] Z. Chang, R. Koodali, R.M. Krishna, L. Kevan, Journal of Physical Chemistry B 104
(2000) 5579–5585.
[30] J.M. Fraile, J.I. Garcia, J.A. Mayoral, E. Vispe, Journal of Catalysis 204 (2001)
146–156.
[
[
[
1] M. Taramasso, G. Perego, B. Notari, U.S. Patent 4,410,501 (1983).
2] B. Notari, Advances in Catalysis 41 (1996) 253–334.
3] C. Perego, A. Carati, P. Ingallina, M.A. Mantegazza, G. Bellussi, Applied Catalysis
A 221 (2001) 63–72.
[
[
[
[
4] A. Corma, Journal of Catalysis 216 (2003) 298–312.
5] C. Perego, R. Millini, Chemical Society Reviews 42 (2013) 3956–3976.
6] M. Hartmann, L. Kevan, Chemical Reviews 99 (1999) 635–663.
7] S. Maurelli, M. Vishnuvarthan, M. Chiesa, G. Berlier, S. Van Doorslaer, Journal
of the American Chemical Society 133 (2011) 7340–7343.
8] S. Maurelli, M. Vishnuvarthan, M. Chiesa, G. Berlier, Physical Chemistry Chem-
ical Physics 14 (2012) 987–995.
[
[
9] R. Raja, S.O. Lee, M. Sanchez-Sanchez, G. Sankar, K.D.M. Harris, B.F.G. Johnson,
J.M. Thomas, Topics in Catalysis 20 (2002) 85–88.
[
[
10] A. Corma, M.A. Camblor, P. Esteve, A. Martinez, J. Perez-Pariente, Journal of
Catalysis 145 (1994) 151–158.
11] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martinez, J.A.
Perdigon-Melon, S. Valencia, Journal of Physical Chemistry B 102 (1998)
7
5–88.
12] J.C. Torres, D. Cardoso, R. Pereira, Microporous and Mesoporous Materials 136
2010) 97–105.
[31] H.E.B. Lempers, J.D. Chen, R.A. Shedon, Studies in Surface Science and Catalysis
94 (1995) 705–712.
[32] J.D. Chen, J. Dakka, R.A. Sheldon, Applied Catalysis A 108 (1994) L1–L6.
[
(