Decatungstate catalyst supported on silica and γ-alumina: Efficient photocatalytic oxidation of benzyl alcohols

Article abstract of DOI:10.1016/j.jcat.2007.09.023

Four supported catalysts with the same tungsten loading were prepared by depositing decatungstate species W₁₀O^4-₃₂, through wet impregnation, on the surface of γ-alumina and silica at different pH values. The prepared samples were characterized using BET measurements as well as XRD, UV-vis DR, and XP spectroscopies. Higher dispersion of W(VI) oxo-species was obtained in the silica-supported catalysts compared with the corresponding alumina-supported ones. Within the same support, the dispersion was higher when the impregnation pH is lower than the point of zero charge (pzc) of the support. The decatungstate anions were present mainly on the silica surface without any modification, whereas these underwent a partial depolymerization on their deposition on the γ-alumina surface. The extent of depolymerization was less in the sample prepared at pH above pzc. These findings were explained in terms of the mode of deposition of the W(VI) species from the solution onto the support surface. The photocatalytic activity of the aforementioned catalysts, concerning the photooxidation of 1-phenylethanol, depends on the fraction of the W₁₀O^4-₃₂ supported species rather than on the W(VI) dispersion. Thus, extremely high conversions have been obtained over the silica-based catalysts and also over the γ-alumina-based catalyst prepared at relatively high pH. These catalysts also are very effective in the photooxidation of a series of secondary and primary benzyl alcohols, in which benzyl ketones and benzoic acids were formed as the only or major products, respectively. The easy separation of the solid catalyst from the reaction mixture, the high activity, selectivity, and stability as well as the retained activity in subsequent catalytic cycles, make these supported catalysts suitable for a small-scale synthesis. Based on product analysis and kinetic data on the heterogeneous oxidation of benzyl alcohols, we suggest that a hydrogen abstraction transfer (HAT) mechanism predominates with respect to an electron transfer (ET) one in these reactions.

Full text of DOI:10.1016/j.jcat.2007.09.023

Journal of Catalysis 252 (2007) 178–189
www.elsevier.com/locate/jcat
Decatungstate catalyst supported on silica and γ -alumina:
Efficient photocatalytic oxidation of benzyl alcohols
Manolis D. Tzirakis a, Ioannis N. Lykakis a, George D. Panagiotou b, Kyriakos Bourikas b,
Alexis Lycourghiotis b, Christos Kordulis b,c,∗, Michael Orfanopoulos a,∗
a
Department of Chemistry, University of Crete, Iraklion, Voutes 71003, Crete, Greece
b
Department of Chemistry, University of Patras, GR 26500, Patras, Greece
c
Institute of Chemical Engineering and High Temperature Chemical Processes (FORTH/ICE-HT), P.O. Box 1414, GR 265 00, Patras, Greece
Received 17 May 2007; revised 27 September 2007; accepted 29 September 2007
Abstract
Four supported catalysts with the same tungsten loading were prepared by depositing decatungstate species W
4−
O
, through wet impregnation,
10
32
on the surface of γ -alumina and silica at different pH values. The prepared samples were characterized using BET measurements as well as XRD,
UV–vis DR, and XP spectroscopies. Higher dispersion of W(VI) oxo-species was obtained in the silica-supported catalysts compared with the
corresponding alumina-supported ones. Within the same support, the dispersion was higher when the impregnation pH is lower than the point
of zero charge (pzc) of the support. The decatungstate anions were present mainly on the silica surface without any modification, whereas these
underwent a partial depolymerization on their deposition on the γ -alumina surface. The extent of depolymerization was less in the sample prepared
at pH above pzc. These findings were explained in terms of the mode of deposition of the W(VI) species from the solution onto the support
surface. The photocatalytic activity of the aforementioned catalysts, concerning the photooxidation of 1-phenylethanol, depends on the fraction of
4−
the W
O
supported species rather than on the W(VI) dispersion. Thus, extremely high conversions have been obtained over the silica-based
10
32
catalysts and also over the γ -alumina-based catalyst prepared at relatively high pH. These catalysts also are very effective in the photooxidation
of a series of secondary and primary benzyl alcohols, in which benzyl ketones and benzoic acids were formed as the only or major products,
respectively. The easy separation of the solid catalyst from the reaction mixture, the high activity, selectivity, and stability as well as the retained
activity in subsequent catalytic cycles, make these supported catalysts suitable for a small-scale synthesis. Based on product analysis and kinetic
data on the heterogeneous oxidation of benzyl alcohols, we suggest that a hydrogen abstraction transfer (HAT) mechanism predominates with
respect to an electron transfer (ET) one in these reactions.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Decatungstate; Heterogeneous photooxidation; Benzyl alcohols; Catalysis; Silica; Alumina
1. Introduction
thesis of fine chemicals and pharmaceutical intermediates. The
objective is to develop industrial chemical processes environ-
mentally acceptable. Under these circumstances, the develop-
ment of heterogeneous catalytic processes is the most suitable
solution.
For catalytic oxidation, an attractive approach is the use of a
solid, recyclable catalyst and environmentally friendly reagent,
such as molecular oxygen as the oxidant, that is, without any
other additive [1–6]. Working at room temperature and at-
mospheric pressure is obviously desirable. Another important
aim is to achieve the highest possible yield and thus minimize
The main shortcoming of homogeneous catalytic processes
comes from the requirement of catalyst recovery or separation.
In many cases, catalyst separation from the reaction mixture is
technically and/or economically unfeasible. Moreover, there is
a growing need for environmentally benign catalysts in the syn-
*
Corresponding authors.
E-mail addresses: kordulis@chemistry.upatras.gr (C. Kordulis),
orfanop@chemistry.uoc.gr (M. Orfanopoulos).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.09.023

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
179
the additional costs of separation and waste removal that may
2. Experimental
be a critical factor in a small-scale synthesis [7–10].
In this challenging field, a number of groups are investi-
gating polyoxometalates (POMs) as catalysts with a significant
activity in the oxidation of organic compounds [11–18]. Poly-
oxometalates (POMs) form a class of inorganic compounds that
can be envisioned as soluble molecular semiconducting oxides.
Their ability to undergo photoinduced multielectron transfer
without changing their structure make them very attractive cat-
alysts in the oxidation of organic substrates in the presence of
molecular oxygen. Among POMs, decatungstate W10O432− ex-
hibits interesting photocatalytic properties, because its absorp-
tion spectrum partially overlaps the UV solar emission spec-
trum, opening the potential route for environmentally benign
solar-photo assisted applications. Heterogeneous catalytic oxi-
dation of organic substrates using W10O432− is currently an area
of intense interest. Nevertheless, only a few examples have been
reported to date concerning the synthesis of heterogenized de-
catungstate catalysts [19–24].
The oxidation of alcohols into the corresponding aldehy-
des, ketones, and acids is a crucial transformation in organic
chemistry with both academic and industrial relevance [25–27].
Following our previous work on the homogeneous photooxida-
tion of aryl alkanols by decatungstate [28–30], here we present
a detailed study on the heterogeneous photocatalyzed oxidation
of a series of secondary and primary benzyl alcohols. To this
purpose, decatungstate was deposited for the first time on sil-
ica and γ -alumina surface by wet impregnation at various pH
values.
More specifically, in this work we deal with (i) the depo-
sition, through wet impregnation, of the decatungstate species
on the surface of γ -alumina and silica at different pH values
(two catalysts based on silica and two based on γ -alumina were
prepared); (ii) the characterization of the prepared catalysts
using BET measurements, X-ray powder diffraction (XRD),
UV–vis diffuse reflectance (DRS), and X-ray photoelectron
(XPS) spectroscopies (in this context, the state of dispersion
and the structure of the active phase and its dependence on
the impregnation pH were investigated); (iii) the photooxida-
tion of the 1-phenylethanol, used as a model compound, over
the above catalysts (in this framework, the relationship be-
tween the physicochemical characteristics of the solid catalysts
and its photocalytic activity is investigated); (iv) the stability,
the possibility for recycling, the dependence of the photocat-
alytic activity on the catalyst amount, and the solvent effect
on the catalyst activity with respect to the most active cata-
lysts, one supported on silica and the other one supported on
γ -alumina; (v) the catalytic behavior of the aforementioned
more active catalysts regarding the heterogeneous photooxida-
tions of a series of secondary and primary benzyl alcohols (in
this context, various benzyl alcohols were synthesized, whereas
the interest is mainly concentrated on the selectivity of the
aforementioned more active catalysts); and (vi) the mechanism
of the heterogeneous catalyzed photooxidation of benzyl alco-
hols.
2.1. Preparation of supported decatungstate catalysts
Four catalysts were prepared by supporting tetrabutylam-
monium decatungstate on SiO2 (Alfa Aesar, 99.8% (metal ba-
sis), specific surface area: 225 m2 g−1, particle size: ∼325
mesh) and γ -Al2O3 (AKZO γ -alumina extrudates, HDS-000-
1.5 mm E, specific surface area: 265 m2 g−1, crushed to ob-
tain powder with particle size ∼100 mesh). The tetrabutylam-
monium decatungstate was synthesized and purified following
standard literature procedures [31]. The tetrabutylammonium
decatungstate prepared was identified by FTIR spectroscopy
[21]. The wet impregnation method was followed for the de-
position of the active component on the surface of the supports.
Two pH values of the impregnating solutions were selected for
each support, one higher and one lower than the correspond-
ing point of zero charge (pcz) of the supports (pzc = ∼7.5 for
γ -Al2O3 [32] and pzc = ∼3 for SiO2 [33]).
In preparation procedure, 0.5 g of [Bu4N]4W10O32 was dis-
solved in a spherical flask of a rotary evaporator that contained
60 mL CH3CN and 15 mL of bi-distilled water. The pH of this
solution is about 3, and it was regulated at the desired pH value
(1.9 and 5.3 in the case of SiO2; 5.2 and 9.0 in the case of
γ -Al2O3) by adding NH4OH or HNO3. Then 5 g of the support
(SiO2 or γ -Al2O3) was added, and the pH of the suspension
was regulated once again at the desired value. The suspension
was left under rotation for 3 h, after which the solvents were re-
moved by evaporation within 4 h, at 30 ◦C and 35 mbar. The
solids obtained were dried in air at 110 ◦C for 24 h. The pre-
pared samples contain 5 wt% W and are symbolized by the
formula WOx/SiO2-Y or WOx/Al2O3-Y, where Y indicates the
pH value of the impregnation solution.
2.2. Preparation of benzyl alcohols
The synthetic procedures of benzyl alcohols (1–15) used as
substrates in the present study, as well as 1H NMR, 13C NMR,
and MS data, are presented in the supplementary material. The
spectroscopic data of the corresponding oxidation products also
are presented.
2.3. Physicochemical characterization of the supported
decatungstate catalysts
Specific surface area (SSA) measurements were performed
in a laboratory constructed apparatus by the three-point dy-
namic BET method. Pure nitrogen and helium (Air Liquide)
were used as adsorption and carrier gas, respectively. A ther-
mal conductivity detector (VICI, Valco Instruments Co., Inc.)
was used for the detection of the adsorbed amount of nitrogen
at a given partial pressure. The water pore volumes (PVs) of
the samples were determined following the incipient wetness
impregnation method.
Diffuse reflectance spectra (DRS) of the samples studied
were recorded in the range of 200–800 nm at room temperature

180
M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
using a UV–vis spectrophotometer (Varian Cary 3) equipped
with an integration sphere. The corresponding supports were
used as references for the catalytic samples.
Table 1
The catalysts prepared, their specific surface areas (SSA) and pore volumes
(PV), the percentage of the tungsten atoms on their surfaces determined by
XPS (W ) and the W-phase surface areas per gram of catalyst (W
)
s.a.
SSA
FTIR spectrum of the tetrabutylammonium decatungstate
synthesized was recorded in a Perkin–Elmer 16 PC FTIR spec-
trophotometer using wafer of the sample in KBr. The spectrum
was recorded in the range 500–4500 cm−1, with a nominal res-
olution of 2 cm−1 and averaging 10 spectra.
a
b
Catalyst
SSA
PV
W
(%)
W
SSA
(m g )
s.a.
2
−1
3
−1
2
−1
(m g
)
(cm g
)
WO /SiO -1.9
157
130
180
120
1.39
1.30
0.52
0.46
6.5
7.3
4.9
5.3
10.2
9.5
8.8
6.4
x
2
WO /SiO -5.3
x
2
WO /Al O -5.2
x
2
3
X-ray diffraction (XRD) patterns were obtained in the range
15◦–80◦ (scanning rate 2◦ min−1) with a Philips PW 1480 au-
tomated diffractometer using CuKα (λ = 1.4518 Å) radiation
filtered through Ni. X-ray photoelectron spectra (XPS) were
recorded at room temperature in an UHV chamber (base pres-
sure 8 × 10−10 mbar) consisting of a fast specimen entry as-
sembly, a preparation, and an analysis chamber. The residual
pressure in the analysis chamber was below 10−8 mbar. The lat-
ter was equipped with a hemispherical electron energy analyzer
(SPECS, LH 10) and a twin-anode X-ray gun for XPS. The
unmonochromatized MgKα line at 1253.6 eV and a constant
pass energy mode for the analyzer were used in the experi-
ments. Pass energies of 36 and 97 eV resulted in full width
at half maximum (FWHM) of 0.9 and 1.6 eV, respectively,
for the Ag3d5/2 peak of a reference foil. The binding energies
were calculated with respect to the C1s peak (C–C, C–H) set
at 284.6 eV.
WO /Al O -9.0
x
2
3
a
2
−1
2
−1
−1
SSA of SiO = 225 m g , SSA of Al O = 265 m g
.
.
2
2 3
b
3
−1
3
PV of SiO = 1.6 cm g , PV of Al O = 0.65 cm g
2
2 3
3. Results and discussion
3.1. Catalyst characterization
3.1.1. Dispersion of the active phase
The BET, XPS, and XRD results allowed investigation of
the state of dispersion of the active phase. The specific surface
areas (SSAs) of the samples are summarized in Table 1. This
table shows that deposition of the decatungstate salt provoked
a drastic diminution of the SSAs of the supports. This likely
indicates that thin pores of the supports were plugged by the
deposited phase, as confirmed by a corresponding decrease in
the pore volumes of the samples, also summarized in Table 1.
However, the decrease in SSA and PV are relatively low in the
cases where the impregnation pH used is below the pzc of the
corresponding support. This indicates that better dispersion of
the supported phase was achieved when the support surfaces
were positively charged.
UV–vis spectra were recorded on a Hitachi U-2001 spec-
trophotometer with bandwidth 195–1100 nm for determining
the concentration of the W10O43−2 species in solution. This de-
termination was performed after evaluating the stability of the
supported catalysts (see Section 3.3).
The aforementioned influence of the impregnation pH on the
dispersion of the supported phase is confirmed by the XPS re-
sults presented in Table 1. In fact, based on the normalized areas
of the XPS peaks of the W4f, Al2p, Si2p and O1s core electrons
and the corresponding atomic sensitivity factors [34] we have
calculated the percentage of the tungsten atoms on the surfaces
of the catalysts prepared (Ws.a.). Assuming that the W, Al, Si,
and O surface atoms occupy similar areas on the surface and by
multiplying the Ws.a./100 values with the specific surface ar-
eas of the catalysts, the SSA of the active phase (WSSA) was
estimated for each sample (Table 1). It can be seen that this pa-
rameter is higher when the impregnation pH is lower than the
pzc of the support. The influence of the impregnation pH on the
dispersion of the supported phase implies that electrostatic in-
teractions are exerted on impregnation between the positively
charged surface hydroxyls of the supports and the W-oxo an-
ions. In effect, these interactions could be responsible for the
high dispersion of the W-phase [35]. We discuss this point fur-
ther later in the paper.
2.4. Photocatalytic reactions
The catalytic reactions were carried out in an open cap vial
equipped with a Teflon septum. The reaction mixture was ir-
radiated under molecular oxygen bubbling. A Variac Cermax
300-W xenon lamp (λ > 300 nm) was used. Most of the exper-
iments used HPLC-grade acetonitrile as the solvent; however,
acetone, dichloromethane, diethyl ether and ethyl acetate were
also used as solvents in several experiments. The reaction mix-
ture (4 mL) contained 0.05 M of substrate and an amount of
catalyst corresponding to 5.5 × 10−4 M [Bu4N]4W10O32. The
catalytic grains were kept in suspension by vigorous magnetic
stirring. The photocatalytic reactions were carried out at 5–
10 ◦C in an ice water bath.
During irradiation, the reaction mixture was monitored by
gas chromatography and 1H NMR spectroscopy. 1H NMR and
13C NMR spectra were recorded on a Bruker AMX 500 MHz
spectrometer, in CDCl3. GC analysis was conducted on a Shi-
madzu GC-17A equipped with a 60-m HP-5 capillary col-
umn. GC-MS analysis was performed on a Shimadzu GC MS-
QP5050A apparatus equipped with a 50-m HP-5 capillary col-
umn and a 5971A MS detector. Detailed data concerning the
product analysis of decatungstate-catalyzed photooxidations
are presented in supplementary material.
The XRD patterns (not presented here) of the alumina-
supported samples showed no peaks assigned to the decatung-
state crystallites. The only peaks observed in these patterns
were due to the γ -alumina used as support. On the other hand,
peaks assigned to the decatungstate crystallites were hardly de-
tected in the SiO2-supported catalysts, whereas XRD peaks due
to the support were not detected, because silica is an amorphous

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
181
Fig. 1. The diffuse reflectance spectra of the samples prepared (WO /SiO -Y and WO /Al O -Y) and the absorption spectrum of (nBu N) W
O
10 32
in the
x
x
2
2
3
4
4
4−
impregnation solution (W
O
).
10
32
material. The above suggest that relatively good dispersion of
the supported W-phase was achieved in all cases though, as al-
ready mentioned, its magnitude depended on the impregnation
pH and the support used.
Concerning the deposition of the W-oxo species from aque-
ous solutions on the γ -alumina surface during equilibration,
a mechanism reported in the literature [38] proposes that at
impregnation pHs > ∼4.5, the deposition occurs mainly via
the monomeric WO24− species, even though at acidic pH, the
polymeric W-oxo species are predominant in the impregnation
solution as it is indicated by equilibrium:
3.1.2. Structure of the active phase
Fig. 1 presents the DR spectra recorded for the samples,
along with the spectrum of the impregnation solution that con-
tained dissolved tetrabutylammonium decatungstate (W10O432−).
The fact that the samples did not absorb in the range 400–
800 nm indicates that W-oxo species with valence <6 were not
practically formed on the surface of both supports irrespective
of the impregnation pH. This is in excellent agreement with our
XPS results, which in all cases showed a binding energy equal
to 35.0 0.1 eV for the W4f7/2 photoelectrons corresponding
to W(VI) [36].
xWO24− + nH+ ꢀ HpWxOy−2x+n + [(n − p)/2]H2O.
(1)
Thus, according to the aforementioned deposition mechanism
at impregnation pH near 5, the selective adsorption of the
monomeric WO24− species on the positively charged surface hy-
droxyls is the main deposition mode. This is illustrated by the
equilibrium
AlOH2+ + WO24− ꢀ AlOH+2 ···WO42−
.
(2)
The above adsorption shifts the equilibrium (1) to the left side
promoting thus the depolymerization of decatungstate ions. At
this point, it should be stressed that the surface of oxide sup-
ports are covered by hydroxyl groups when they are in con-
tact with an electrolyte solution. These groups may be proto-
nated or deprotonated resulting into positively and negatively
charged hydroxyls, respectively. The relative concentration of
the positively (negatively) charged surface hydroxyls increases
(decreases) as pH decreases. Thus, at pH 5.2, which is con-
siderably lower than the pzc = 7.5 of alumina, the positively
charged hydroxyls are predominant, favoring the adsorption of
the monomeric WO24− species on the alumina surface via equi-
librium (2). This in effect is in agreement with the blue shift
observed in the DR spectrum of the dried WOx/Al2O3-5.2 sam-
ple, as well as with the relatively high dispersion of the W-phase
detected by XPS (Table 1).
3.1.2.1. Structure of the active phase supported on γ -alumina
The absorption spectrum of impregnation solution (W10O34−2
)
exhibited two main peaks at ∼272 and ∼321 nm. In contrast,
the DR spectra of the WOx/Al2O3-Y samples exhibited a single
absorbance at ∼235 nm. In a previous work [37], we showed
that the depolymerization of the W-oxo species is related to a
blue shift of their absorbance from ∼380–400 nm observed in
WO3 with octahedral symmetry to ∼205–225 nm observed in
Na2WO4 with tetrahedral symmetry. Thus, inspecting Fig. 1,
we can conclude that the decatungstate species underwent par-
tial depolymerization on their deposition on the alumina sur-
face. Taking into account the impregnation method followed,
a portion of the W-phase could be expected to have been de-
posited on the alumina surface during the 3 h equilibration of
the impregnation suspension at the selected pH, whereas the
remainder portion has been deposited by precipitation on the
drying step due to the evaporation of the solvents.
On the other hand, at impregnation pH 9.0, used for the
preparation of WOx/Al2O3-9.0 sample, the monomeric WO42−

182
M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
Scheme 1. Photooxidation of 1-phenylethanol (1) to acetophenone (1a).
species are predominant in the impregnation solution in accor-
dance to the equilibrium (1), while the population of the posi-
tively charged hydroxyls on the γ -alumina surface is negligible.
This is because the pzc of the γ -alumina used is ∼7.5. In view
of the above the development of attractive electrostatic interac-
tions of the monomeric WO24− species with the support surface
sites is not expectable. In fact, according to the aforementioned
deposition mechanism [38] at this high pH only a small fraction
of the supported W-phase could be deposited via surface reac-
tion of the monomeric WO24− ions with two adjacent neutral
surface hydroxyls as it is illustrated by the equilibrium:
Fig. 2. Time profiles of the 1-phenylethanol photooxidation in oxygen saturated
4−
CH CN in the presence of (nBu N) W
O
(W
O
) and over the hetero-
3
4
4
10 32
10
32
geneous catalysts (WO /SiO -Y and WO /Al O -Y).
x
x
2
2 3
Table 2
2AlOH + WO24− ꢀ AlO–(WO2)–OAl + 2OH−.
(3)
Heterogeneous photooxidation of 1-phenylethanol over different supported
(WO /SiO and WO /Al O ) decatungstate catalysts
x
x
2
2 3
The rest W-phase was deposited via precipitation on evapora-
tion of the solvents. The above mechanism explains both the
depolymerization of decatungstate species detected by DRS in
this sample, as well as the lower WSSA in comparison to that of
the WOx/Al2O3-5.2 sample observed by XPS (see Table 1). In
fact, it is accepted that precipitation results to larger supported
crystallites than adsorption [35].
a
Catalyst
Irradiation
time (min)
Conversion
(%)
Selectivity
(1a) (%)
4−
W
O
90
99
>99
10
32
WO /SiO -1.9
90
90
90
90
92
95
12
65
>99
>99
>99
>99
x
2
WO /SiO -5.3
x
2
WO /Al O -5.2
x
2
3
WO /Al O -9.0
x
2
3
a
1
Determined by H NMR spectroscopy and gas chromatography, with error
3.1.2.2. Structure of the active phase supported on silica The
DR spectra of the WOx/SiO2-Y samples are identical with the
spectrum (W10O342−) of the (nBu4N)4W10O32 in the impregna-
tion solution (see the absorption peaks at ∼272 and ∼321 nm;
Fig. 1). This shows that the decatungstate anions are present
mainly on the silica surface with no modification. However, the
slightly lower intensities of the absorption peaks observed in
the spectrum of WOx/SiO2-1.9 sample compared with those of
the WOx/SiO2-5.3 sample probably demonstrates that a small
fraction of the W-phase is deposited in the former sample via
adsorption of W-oxo anions on the positively charged hydroxyls
of the SiO2 followed by depolymerization. This is in accor-
dance with the relatively high WSSA value achieved for the
WOx/SiO2-1.9 sample (see Table 1). The above results prob-
ably show that when the deposition of (nBu4N)4W10O32 on
the silica surface takes place by impregnation in aqueous so-
lution at pH below pzc, a direct interaction is exerted between
the W-oxo anions and the protonated surface hydroxyls of the
support. However, other researchers [19,22,39] have shown that
the quaternary ammonium cations of the salt interact with the
support surface, and these cations may act as a bridge between
the surface and the decatungstate when nonaqueous impregna-
tion solutions are used.
4% and 1%, respectively. Naphthalene was used as internal standard.
under homogeneous conditions yields acetophenone (1a) as the
only product (Scheme 1) [28]. All of the heterogeneous cat-
alytic reactions were carried out under the same conditions,
using 0.05 M of 1 and 36.5 mg (ca. 5.5 × 10−4 M) of the sup-
ported catalyst. The kinetic curves obtained are shown in Fig. 2.
In these curves, the % conversion of 1 is plotted as a function
of reaction time. For comparison, the corresponding homoge-
neous reaction was also evaluated (dashed line in Fig. 2). When
the same reaction was run in the presence of pure SiO2 or Al2O3
no oxidation products were obtained. This observation indicates
that the catalyst’s support does not interfere to the photocat-
alyzed reaction. Control experiments in the absence of catalyst
showed no oxidation of the substrate in the presence of molec-
ular oxygen. In addition, no oxidation products were observed
when blank experiments were run in the dark.
In all cases, irradiation of 1 in acetonitrile gave the corre-
sponding ketone 1a as the only product. The percentage conver-
sions for this reaction were found to be equal to 92 and 95%, re-
spectively, over WOx/SiO2-1.9 and WOx/SiO2-5.3 samples at
90 min irradiation (Table 2). These results demonstrate that the
activity of the silica-supported catalysts is surprisingly as high
as that of the homogeneous precursor (Table 2). In contrast,
the Al2O3-supported catalyst exhibited lower activity than that
of the homogeneous or silica-supported catalysts. The percent-
age conversion over the sample WOx/Al2O3-5.2 did not exceed
12%; however, a percentage conversion equal to 65% regarding
3.2. Photocatalytic activity
The photocatalytic activity of the above four catalytic sys-
tems, has been assessed studying the photooxidation of 1-phen-
ylethanol (1), as a model reaction. This reaction is known that,

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
183
(a)
(b)
Fig. 3. Catalytic performance of recycled (a) WO /SiO -5.3 and (b) WO /Al O -9.0 catalysts in the heterogeneous photooxidation of 1.
x
x
2
2 3
the WOx/Al2O3-9.0 catalyst is quite important, demonstrating
the influence of the impregnation pH (Table 2).
previous conclusion that leaching did not result in significant
removal of the decatungstate catalyst from the support surface.
The above results suggest that quite strong chemical inter-
actions occurred between supported W-oxo species and support
surfaces on heating at 100 ◦C. Moreover, these results indicate
that the contribution of the homogeneous reaction to the het-
erogeneously catalyzed reaction was actually negligible. Thus,
we conclude that the catalytic activity of the systems studied
here can be attributed almost exclusively to the heterogenized
decatungstate sites.
To confirm the feasibility of catalyst recycling, 1 (0.1 M) was
irradiated for 90 min in the presence of 73 mg WOx/SiO2-5.3 or
WOx/Al2O3-9.0 (11 × 10−4 M (nBu4N)4W10O32) in oxygen-
saturated CH3CN, at 5–10 ◦C. The resulting solution was cen-
trifuged, the supernatant was removed, and the solid catalyst
was rinsed thoroughly 4 times with CH3CN. The catalyst was
then reused in a new catalytic cycle. The above procedure was
repeated twice for each catalyst (Fig. 3).
The activity behavior of the supported catalysts could be
explained taking into account their physicochemical character-
istics. More precisely, the lower activity of the WOx/Al2O3-Y
samples compared with the WOx/SiO2-Y samples could be
easily attributed to the depolymerization of the decatungstate
species taking place on deposition of the active phase on the
alumina surface. The relatively low activity of the catalysts
prepared by impregnation at pHs lower than the pzc of the cor-
responding supports is probably due to the stronger supported
phase–support interactions developed in these cases. Such in-
teractions stabilize the oxidation state of the supported phase.
This in turn decreases the ability of the supported phase to
change its oxidation state and thus to participate in the oxidation
catalytic cycle needed for the heterogeneous photooxidation of
1-phenylethanol.
Inspection of Fig. 3 shows that the catalysts could be reused
at least twice without apparent loss of catalytic activity. The
first catalytic cycle over the WOx/SiO2-5.3 and WOx/Al2O3-
9.0 samples resulted in conversion of 78 and 50%, respectively,
within 90 min of irradiation. In the second cycle, the conver-
sions were slightly decreased, to 70 and 42%, respectively.
However, in the third catalytic cycle, 17 and 27% conversions
were achieved within the same reaction time. Although the
activity over the sample WOx/SiO2-5.3 was notably dimin-
ished in the third catalytic cycle, the activity over the sample
WOx/Al2O3-9.0 was not diminished proportionately. These re-
sults are indicative of a greater stability of the active phase when
alumina was used as the catalyst support.
3.3. Stability and reusability of the catalysts
Among the various pH values and supports, the WOx/SiO2-
5.3 and WOx/Al2O3-9.0 were found to be the most efficient
catalysts. Thus, we investigated the stability and reusability
of these two catalysts. To test the catalysts’ stability, sepa-
rate solutions of WOx/Al2O3-9.0 and WOx/SiO2-5.3 catalysts
(5.5 × 10−4 M or 36.5 mg in 4 mL CH3CN) were irradiated for
90 min under continuous oxygen bubbling at 22 1 ◦C. Both
catalysts were kept in suspension by magnetic stirring. After
the irradiation experiments, the resulting solutions were cen-
trifuged, and the UV/vis spectra of the supernatant solutions
were compared with a standard solution of (nBu4N)4W10O32
(5.5 × 10−4 M), indicating that <3% of the decatungstate had
been lost from the silica or alumina.
3.4. Effect of catalyst amount on the photocatalytic activity
To further confirm our results on the leaching of the cata-
lysts, both catalysts were mixed with acetonitrile in separate
solutions and stirred for 5 h at 22 1 ◦C. Then the catalysts
were filtered and removed, while 0.05 M of 1-phenylethanol
was added into the filtrates. The resulting solutions were irradi-
ated for 2 h in the presence of molecular oxygen. Only a small
amount of acetophenone 1a (4–8%) was detected by gas chro-
matography. The above control experiments corroborate the
To investigate the dependence of the photocatalytic activity
on the amount of catalyst in the oxidation of 1, the reaction was
carried out using five different suspensions of varying catalyst
concentrations. These results are illustrated in Fig. 4. Generally
speaking, a higher concentration of the supported decatungstate
in the reaction mixture is expected to increase the percentage
conversion. As seen in Fig. 4, an increase in catalyst concentra-

184
M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
(a)
(b)
Fig. 4. Time profiles of the 1-phenylethanol photooxidation catalyzed by various masses of the (a) WO /SiO -5.3 and (b) WO /Al O -9.0 catalysts, in oxygen
x
x
2
2 3
◦
saturated CH CN at 5–10 C.
3
(a)
(b)
Fig. 5. Time profiles of the 1-phenylethanol photooxidation over 7.3 mg of the (a) WO /SiO -5.3 and (b) WO /Al O -9.0 catalysts in various solvents saturated
x
x
2
2 3
◦
with oxygen at 5–10 C.
tion led to higher conversions, especially in the initial stages of
the reaction.
31 to 56%, respectively. However, an even greater increase in
WOx/Al2O3-9.0 (from 7.3 to 36.5 mg) resulted in only a small
increase in the final conversion (from 56 to 65%). Therefore,
in the case of Al2O3-supported catalyst, the more pronounced
shadow effects could be attributed to the lower transparency of
the alumina compared with silica particles.
In particular, working with 3.65 mg of WOx/SiO2-5.3 cata-
lyst [concentration of active phase in the reaction mixture equal
to 5.5×10−4 M (nBu4N)4W10O32], almost 1 h was required for
35% completion of the reaction, compared with only 10 min
when using 36.5 mg [5.5 × 10−4 M (nBu4N)4W10O32]. On
the other hand, an additional increase in the concentration of
the active phase beyond 5.5 × 10−4 M (nBu4N)4W10O32 did
not affect the rate of the oxidation reaction to the same ex-
tent. Thus, when the concentration value of the active phase
was 5.5 × 10−4 M (nBu4N)4W10O32, the reaction conversion
reached almost maximum value (Fig. 4a). We could attribute
this observation to shadow effects. In fact, as the amount of the
solid catalyst increased in the reaction mixture, a continuously
higher fraction of the catalyst particles could not be illumi-
nated being in the shadow of the particles suspended in front
of them. Similar and even more pronounced shadow phenom-
ena should be operative in the case of WOx/Al2O3-9.0 catalyst,
as inferred by the catalytic results depicted in Fig. 4b. For exam-
ple, an increase in the amount of WOx/Al2O3-9.0 from 3.65 mg
to 7.3 mg resulted in an increase of the final conversion from
3.5. Solvent effect on the catalyst activity
Various solvents were evaluated for the catalytic photooxi-
dation of 1-phenylethanol 1. The kinetic results are illustrated
in Fig. 5. The solvent effect on catalysts activity is actu-
ally important. Acetonitrile and acetone were found to be the
most effective solvents for both catalysts (WOx/SiO2-5.3 and
WOx/Al2O3-9.0). In contrast, significantly lower conversions
were achieved using dichloromethane, diethyl ether, or ethyl
acetate.
3.6. Heterogeneous photooxidation of p-substituted phenyl
ethanols
We extended our study of the catalytic behavior of the sam-
ples WOx/SiO2-5.3 and WOx/Al2O3-9.0 to the p-substituted

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
185
suggests the formation of a radical intermediate through a hy-
drogen atom transfer (HAT) mechanism, which is better stabi-
lized by electron-withdrawing substituents (such as –CF3).
3.7. Heterogeneous photooxidation of p-alkyl-substituted
benzyl alcohols
Scheme 2. Heterogeneous photooxidation of p-substituted phenylethanols 1–3.
So far, we have reported the catalytic behavior of the WOx/
SiO2-5.3 and WOx/Al2O3-9.0 catalysts with respect to the pho-
tooxidation of p-substituted phenyl ethanols 1–3 bearing one
benzylic hydrogen atom. To further probe the applicability of
these heterogeneous catalytic systems in a small-scale synthe-
sis, a series of p-alkyl-substituted benzyl alcohols, 4–10, were
oxidized under the experimental conditions described above.
These substrates bear two distinguishable benzylic hydrogen
atoms, one on the alcohol carbon and one on the p-alkyl sub-
stituent, both of which potentially can be cleaved under the pho-
tooxidation conditions. Therefore, the selectivity of the catalyst
by means of preferential hydrogen abstraction can be tested.
The corresponding ketone would be expected to be formed ex-
clusively in hydrogen atom abstraction from the CH–OH car-
bon. On the other hand, hydrogen abstraction from the benzylic
carbon of the p-substituent, led to the formation of dioxy-
genated products, as shown in Scheme 3.
1-phenylethanols, 1-[4-(trifluoromethyl)phenyl]ethanol (2) and
1-(4-methoxyphenyl)ethanol (3). In this context, we studied the
heterogeneous photooxidation of 2 and 3 (see Scheme 2). The
synthetic procedure for these alcohols is described in supple-
mentary materials.
All of the catalytic oxidations were carried out under the
same experimental conditions, using 36.5 mg of the catalyst
[concentration of the active phase 5.5 × 10−4 M (nBu4N)4-
W10O32] and 0.05 M of the alcohol in 4 mL CH3CN. The
activity results are compiled in Table 3, which also reports the
oxidation of 1 for comparison. In all cases, the corresponding
aryl ketones, 2a and 3a, were formed as the only products.
The kinetic results indicate that heterogenized decatungstate
may be generally used for the catalytic photooxidation of ben-
zyl alcohols. Concerning the silica-based sample, it can be seen
that the effect of the p-substituent on the percentage conversion
was rather negligible. As for the γ -alumina-based sample, as
the electron-withdrawing ability of the p-substituent increased,
the photooxidation reaction rate was accelerated. This result
Scheme 4 presents the photooxidation reactions of the p-al-
kyl-substituted benzyl alcohols. The results, summarized in Ta-
ble 4, show that both catalysts tested were active in the photoox-
idation of benzyl alcohols 4–10. Moreover, the silica-supported
catalyst was more active than the alumina-supported one for all
of photooxidation reactions studied.
Table 3
Results of heterogeneous photooxidation of 1-phenylethanols 1–3 catalyzed by
WO /SiO -5.3 and WO /Al O -9.0 in oxygen saturated CH CN
x
x
2
2
3
3
As shown in Table 4, the photooxidation of a series of
para-substituted benzyl alcohols 4–10 showed a strong pref-
erence for oxidation of the carbon that bears the hydroxy
group over both WOx/SiO2-5.3 and WOx/Al2O3-9.0 catalysts,
resulting in the corresponding benzyl ketones 4a–10a. 1-(4-
Methylphenyl)ethanol 4 impressively illustrates this point. Sim-
ilarly, alkyl-substituted phenyl ethanols 5 and 6 gave again the
corresponding benzyl ketones 5a and 6a as the major products.
However, the selectivity depends on the extent of the reaction
time (results not shown). In example, for substrates 5 and 6, at
low irradiation time, high selectivities for benzyl ketones were
observed (entries 3 and 5, Table 4). The same trend also was
a
Substrate
Catalyst
Irradiation
time (min)
Conversion
(%)
1
1
2
2
2
3
3
3
WO /SiO -5.3
90
90
90
90
180
90
95
61
100
48
84
96
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
90
180
25
39
x
2
3
WO /Al O -9.0
x
2
3
a
1
Determined by H NMR spectroscopy and gas chromatography, with error
4% and 1%, respectively.
Scheme 3. Two benzylic hydrogen atoms, both potentially able to be cleaved, in the photooxidation of p-alkyl-substituted benzyl alcohols.
Scheme 4. Photooxidation reactions of the p-alkyl-substituted benzyl alcohols 4–10.

186
M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
Table 4
Product analysis of the heterogeneous photooxidation of aromatic alcohols 4–10 catalyzed by WO /SiO -5.3 and WO /Al O -9.0 in oxygen saturated CH CN
x
x
2
2
3
3
a
b,c
b
Entry
Substrate
Catalyst
Irradiation
time (min)
Selectivity (%)
Conversion
(%)
a
b
c
d
e
1
2
3
4
5
6
7
8
9
10
11
12
13
14
4
4
5
5
6
6
7
7
8
8
9
9
10
10
WO /SiO -5.3
90
120
30
90
30
90
60
60
60
60
60
60
60
98
97
80
70
77
41
93
73
93
44
85
63
92
88
100
100
82
68
80
84
76
78
98
99
84
85
76
70
–
–
1
12
–
11
–
–
2
1
–
–
–
5
11
10
3
4
2
–
–
–
–
–
–
–
–
–
2
6
–
–
2
4
–
–
x
2
WO /Al O -9.0
–
12
9
10
2
18
14
–
–
13
9
11
13
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
1
2
13
17
x
2
WO /Al O -9.0
–
–
–
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
120
x
2
3
a
−4
Arylalcohol (0.05 M) was irradiated (xenon lamp, 300 W, λ > 300 nm) in the presence of WO /SiO -5.3 and WO /Al O -9.0 (5.5 × 10
M), in 4 mL
x
x
2
2 3
◦
CH CN, at 5–10 C.
b
3
1
Determined by H NMR spectroscopy and gas chromatography after reduction of hydroperoxides with PPh , with error 4% and 1%, respectively.
3
c
See Scheme 4 for a, b, c, d, and e.
noted in substrates 7 and 8, where the aforementioned selectiv-
ity, as well as the % conversion, were even higher at only 60 min
of irradiation (entries 7 and 9, Table 4). In addition, the pres-
ence of a bulky phenyl group next to the benzylalcohol carbon
did not change this selectivity, as demonstrated by substrates
9 and 10. Similar selectivities also were observed in the photo-
catalytic oxidations of benzyl alcohols 4–10, over WOx/Al2O3-
9.0. These results demonstrate a stronger and more consistent
preference for hydrogen abstraction from the benzylalcohol
carbon compared with the para-alkyl substituent of substrates
4–10. To rationalize these results, we propose that the polar sur-
face of the solid supports, probably favors accumulation of the
polar alcohol moiety, resulting in the selective oxidation to the
corresponding ketones. These results clearly demonstrate the
important role of the present catalytic systems in synthetically
useful oxidative transformations.
Scheme 5. Photooxidation reactions of the primary benzyl alcohols 11–14.
Table 5
Product analysis of the heterogeneous photooxidation of benzyl alcohols 11–14
catalyzed by WO /SiO -5.3 or WO /Al O -9.0 in oxygen saturated CH CN
x
x
2
2
3
3
a
b,c
b
Substrate
Catalyst
Selectivity (%)
Conversion
(%)
a
b, c
11
11
12
12
13
13
14
14
WO /SiO -5.3
99
92
99
98
97
86
92
65
2
3
1
2
4
5
3
98
97
99
98
96
95
97
44
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
WO /SiO -5.3
x
2
WO /Al O -9.0
x
2
3
3.8. Heterogeneous photooxidation of primary benzyl alcohols
WO /SiO -5.3
x
2
WO /Al O -9.0
56
x
2
3
a
In the previous subsections, we reported on the catalytic
ability of the WOx/SiO2-5.3 and WOx/Al2O3-9.0 catalysts re-
garding the heterogeneous photooxidation of secondary ben-
zyl alcohols. In this subsection we extend our study to the
photooxidation of primary benzyl alcohols. This is necessary
to accomplish the evaluation of these catalysts with respect
to the photooxidation of benzyl alcohols. The catalytic tests
were performed under the reaction conditions described previ-
ously. Benzyl alcohol 11, 4-trifluoromethylbenzyl alcohol 12,
4-methylbenzyl alcohol 13, and 4-methoxybenzyl alcohol 14
(see Scheme 5) have been used as probe molecules.
Table 5 compiles the catalytic results. It may be seen that
in most cases, the photooxidation of the primary benzyl al-
cohols 11–14 resulted in the corresponding para-substituted
benzoic acids 11b–13b. Only in the photooxidation of 14 with
WOx/Al2O3-9.0 was the corresponding aldehyde 14a the major
product. It is remarkable that in all cases, the carboxylic acid
Benzyl alcohol (0.05 M) was irradiated (xenon lamp, 300 W, λ > 300 nm)
for 90 min in the presence of WO /SiO -5.3 or WO /Al O -9.0 (5.5 ×
x
x
2
2 3
−4
◦
10 M), in 4 mL CH CN, at 5–10 C.
b
3
1
Determined by H NMR spectroscopy.
c
See Scheme 5 for a, b, and c.
(11b–14b) versus the corresponding peroxo-compound (11c–
14c) were formed at ratios up to 15/1. This ratio was determined
by H NMR spectroscopy. These results are not in agreement
1
with recent work on the decatungstate/SiO2-catalyzed oxida-
tion of primary benzyl alcohols. In that study, benzyl aldehydes
were reported to be the only products with no further oxida-
tion to the carboxylic acids [40]. The observed disagreement
could be related to the different method followed for catalyst
preparation. In fact, the decatungstate/SiO2 catalyst used in the
latter case was prepared by entrapment of (nBu4N)4W10O32
into the SiO2 matrix developed following the sol–gel procedure

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
187
involving the hydrolysis of tetraethylortho-silicate (Si(OEt)4,
TEOS).
tinguishing an ET from a HAT mechanism in the side-chain
oxidation of arylalkanes [49] and arylalkanols [28,49–53]. It
is known that in the side-chain oxidation of aromatic alcohols,
which involves an ET mechanism, a Cα–Cβ bond cleavage
leads to the corresponding aldehydes through a radical cation
intermediate [50–53] (Scheme 7). On the other hand, a HAT
mechanism leads only to the corresponding ketones [28,49].
The products likely obtained on photooxidation of 15 are
presented in Scheme 8. The results obtained from these re-
actions are summarized in Table 6. For comparison purposes,
the table also reports product analysis of the corresponding ho-
3.9. Reaction mechanism
It is known that in the homogeneous decatungstate-catalyzed
photooxidation, the reaction between the reactive wO interme-
diate with organic substrates XH may occur either by (a) hy-
drogen abstraction transfer (HAT) and/or (b) electron transfer
(ET), according to the proposed mechanistic schemes [16,41–
47] shown in Scheme 6. Both mechanisms give rise to the same
1 electron-reduced species and to the corresponding substrate-
derived radical. Recent studies based on kinetic data have sug-
gested that a HAT mechanism involves the oxidation of aryl
alcohols [29,30].
To elucidate the reaction mechanism of the present heteroge-
neous catalysts, we performed the photooxidation reaction of 1
in the presence of 1,3,5-trimethoxybenzene (TMB), which is a
well-established electron-transfer (ET) quencher. If ET occurs,
TMB with E1/2(ox) versus SCE 1.12 V [48] (oxidation poten-
tial for the 1-arylalcohols is 1.6–1.7 V, [49]) should retard the
photooxidation by back-electron transfer to the 1 radical cation
according to
mogeneous photooxidation using (nBu4N)4W10O32 (W10O4−
)
[28] and 9,10-dicyanoanthracene (DCA, a well established3E2T
photosensitizer [48,54–58]) as catalysts. As shown in Table 6,
the heterogeneous photooxidation of 15 led to the correspond-
ing ketone a as the major product (80 and 82% relative yield),
along with small amounts of the products b and c. This signif-
icant amount of the products b and c indicates the contribution
1
•+ + TMB ꢀ 1 + TMB•+
.
(4)
The reaction quenching plots are shown in Fig. 6. In both cat-
alytic reactions, a significant decrease in the reaction rate can be
seen, suggesting that ET likely takes part in the reaction mech-
anism.
To further illuminate the reaction mechanism, we investi-
gated the WOx/Al2O3-9.0 and WOx/SiO2-5.3 catalyzed pho-
tooxidation of 1-(4-methoxyphenyl)-2,2-dimethyl-1-propanol
(15). This alcohol has been used as mechanistic tool for dis-
Fig. 6. Time profiles of the 1-phenylethanol photooxidation in oxygen saturated
◦
CH CN at 5–10 C over 7.3 mg of the WO /SiO -5.3 and WO /Al O -9.0
x
x
3
2
2 3
catalysts in absence and in the presence of TMB in the reaction mixture.
Table 6
Product analysis of the 1-(4-methoxyphenyl)-2,2-dimethyl-1-propanol 15 pho-
tooxidation catalyzed by WO /Al O -9.0, WO /SiO -5.3, (nBu N) W
O
10 32
x
x
2
3
2
4
4
and DCA in oxygen saturated CH CN
3
b,c
b
Catalyst
Time
(min)
Selectivity (%)
Conversion
(%)
a
a
b
c
Scheme 6. Reaction of wO intermediate with organic substrates XH.
4−
W
O
20
60
60
20
30
25
20
30
97
80
82
25
3
5
7
–
10
32
WO /SiO -5.3
15
11
10
x
2
WO /Al O -9.0
x
2
3
DCA
65
a
1-(4-Methoxyphenyl)-2,2-dimethyl-1-propanol (0.05 M) was irradiated
−4
(xenon lamp, 300 W, λ > 300 nm) in the presence of catalyst (5.5 × 10 M),
◦
in 4 mL CH CN, at 5–10 C.
3
b
1
Scheme 7. Oxidation of 1-(4-methoxyphenyl)-2,2-dimethyl-1-propanol (15)
through an electron transfer (ET) process.
Determined by H NMR spectroscopy with error 1%, respectively.
c
See Scheme 8 for a, b, and c.
Scheme 8. Photooxidation of 1-(4-methoxyphenyl)-2,2-dimethyl-1-propanol 15 catalyzed by WA-9.0, WS-5.3, (nBu N) W
O
10 32
and DCA in oxygen saturated
4
4
CH CN.
3

188
M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
of an ET mechanism to the overall mechanistic pathway. In par-
ticular, in the case of an ET mechanism (DCA), products b and
c would be formed as the major ones in 75% relative yield [58],
whereas in the case of a HAT mechanism, product a would
be expected to be the only one [28]. Thus, it is clear that the
heterogeneous photocatalyzed oxidation of benzyl alcohols by
decatungstate supported on silica and alumina proceeded via
HAT and ET mechanisms, with the former as the predominant
one.
hydrogen atoms, a highly selective oxidation to the cor-
responding benzyl ketones was observed. Concerning the
heterogeneous photooxidation of primary benzyl alcohols,
we found that in most of the cases, the corresponding ben-
zoic acids were formed selectively.
6. Based on product analysis and kinetic data, we suggest
that the heterogeneous decatungstate-catalyzed oxidation
of benzyl alcohols proceeded via a HAT and ET mecha-
nism, with the former as the predominant one.
7. The easy separation from the reaction mixture, along with
the high activity, selectivity, stability, and retained activity
in subsequent catalytic cycles, make these supported cata-
lysts suitable for small-scale synthesis.
4. Conclusion
Based on our findings in the present study, we can draw the
following conclusions:
Supplementary material
1. The state of dispersion of the decatungstate phase de-
posited on the silica and γ -alumina surface by impregna-
tion from aqueous solutions depended on the nature of the
support and also on the impregnation pH. Higher dispersion
was obtained in the silica-supported catalysts compared
with the corresponding alumina-supported ones. Within the
same support, the dispersion was higher when the impreg-
nation pH was lower than the pzc of the support.
2. Concerning the structure of the supported phase, we found
that the decatungstate anions were present mainly on the
silica surface with no modification. Only a very small frac-
tion of the W-oxo species, deposited through electrostatic
adsorption at pH below the pzc of silica, was depolymer-
ized. In contrast, the decatungstate species underwent par-
tial depolymerization on their deposition on the γ -alumina
surface. The extent of depolymerization was smaller in the
sample prepared at pH higher than pzc.
The online version of this article contains additional supple-
mentary material.
Please visit DOI: 10.1016/j.jcat.2007.09.023.
References
[1] L.J. Huerta, P. Amorós, D. Beltrán-Porter, V. Cortés Corberán, Catal. To-
day 117 (2006) 180.
[2] C. Mateos-Pedrero, C. Cellier, P. Ruiz, Catal. Today 117 (2006) 362.
[3] C. Cellier, B. Blangy, C. Mateos-Pedrero, P. Ruiz, Catal. Today 112 (2006)
112.
[4] V. Cortés Corberán, Catal. Today 99 (2005) 33.
[5] V.P. Vislovskiy, N.T. Shamilov, A.M. Sardarly, R.M. Talyshinskii, V.Yu.
Bychkov, P. Ruiz, V. Cortés Corberán, Z. Schay, Zs. Koppany, Appl. Catal.
A Gen. 250 (2003) 143.
[6] L. Moens, P. Ruiz, B. Delmon, M. Devillers, Appl. Catal. A Gen. 249
(2003) 365.
[7] M. Besson, P. Gallezot, Catal. Today 57 (2000) 127.
[8] B.M. Choudary, M.L. Kantam, P.L. Santhi, Catal. Today 57 (2000) 17.
[9] B.Z. Zhan, A. Thompson, Tetrahedron 60 (2004) 2917.
[10] T. Mallat, A. Baiker, Chem. Rev. 104 (2004) 3037.
[11] C.L. Hill (Ed.), Chem. Rev. 98 (1998) 1 (a special issue of Chemical Re-
views is devoted to polyoxometalates).
3. The differences in the state of dispersion and the structure
of the deposited phase may be explained in terms of the
mode of deposition of the W-oxo species from the solution
onto the support surface.
4. Using 1-phenylethanol as a model compound, it was con-
cluded that the photocatalytic activity depends on the frac-
tion of the W10O432−-supported species rather than on the
total area of the W phase. Thus, surprisingly, the activ-
ity of the silica-supported catalysts was as high as that of
the homogeneous precursor, whereas the Al2O3-supported
catalyst exhibited lower activity compared with the homo-
geneous or silica-supported catalysts. Moreover, within the
same support, the photocatalytic activity increased with
increasing impregnation pH. Extremely high conversions
were obtained over the silica-based catalysts and over the
γ -alumina-based catalyst prepared at relatively high pH.
We further used the most active catalyst supported on silica
and the most active one based on γ -alumina to investigate
critical aspects of photooxidation.
[12] M.T. Pope, in: Heteropoly and Isopoly Oxometallates, Springer-Verlag,
Berlin, 1993.
[13] C.L. Hill, C.M. Prosser-McCartha, in: Photosensitization and Photocatal-
ysis Using Inorganic and Organometallic Compounds, Kluwer Academic,
Dordrecht, 1993.
[14] R. Neumann, M. Dahan, Nature 388 (1997) 353.
[15] C.L. Hill, C.M. Prosser-McCartha, Coord. Chem. Rev. 143 (1995) 407.
[16] C. Tanielian, Coord. Chem. Rev. 178 (1998) 1165.
[17] A. Hiskia, A. Mylonas, E. Papaconstantinou, Chem. Soc. Rev. 30 (2001)
62.
[18] C.L. Hill (Ed.), J. Mol. Catal. 262 (2007) 1 (a special issue is devoted to
polyoxometalates).
[19] A. Molinari, R. Amadelli, L. Andreotti, A. Maldotti, J. Chem. Soc. Dalton
Trans. (1999) 1203.
[20] A. Molinari, R. Amadelli, V. Carassiti, A. Maldotti, Eur. J. Inorg. Chem. 1
(2000) 91.
[21] A. Molinari, R. Amadelli, A. Mazzacani, G. Sartori, A. Maldotti, Lang-
muir 18 (2002) 5400.
5. We proved that these two supported catalysts also were
very effective for the photooxidation of three series of ben-
zyl alcohols (secondary benzyl alcohols bearing one or
two distinguishable benzylic hydrogen atoms and primary
benzyl alcohols). Concerning the heterogeneous photooxi-
dation of secondary benzyl alcohols bearing two benzylic
[22] A. Maldotti, A. Molinari, G. Varani, M. Lenarda, L. Storaro, F. Bigi,
R. Maggi, A. Mazzacani, G. Sartori, J. Catal. 209 (2002) 210.
[23] Y. Guo, C. Hu, X. Wang, Y. Wang, E. Wang, Y. Zou, H. Ding, S. Feng,
Chem. Mater. 13 (2001) 4058.
[24] M. Bonchio, M. Carraro, G. Scorrano, E. Fontananova, E. Drioli, Adv.
Synth. Catal. 345 (2003) 1119.

M.D. Tzirakis et al. / Journal of Catalysis 252 (2007) 178–189
189
[25] M. Hudlicky, Oxidations in Organic Chemistry, Am. Chem. Soc., Wash-
ington, DC, 1990.
[41] I. Texier, J.F. Delouis, J.A. Delaire, C. Giannotti, P. Plaza, M.M. Martin,
Chem. Phys. Lett. 311 (1999) 139.
[26] S.V. Ley, J. Norman, W.P. Griffith, S.P. Marsden, Synthesis 7 (1994) 639.
[27] G. Tojo, M. Fernández, in: G. Tojo (Ed.), Oxidation of Alcohols to Alde-
hydes and Ketones: A Guide to Current Common Practice, in: Science &
Business Media, Springer, 2006.
[42] I. Texier, J.A. Delaire, C. Giannotti, Phys. Chem. Chem. Phys. 2 (2000)
1205.
[43] C. Tanielian, R. Mechin, R. Seghrouchni, C. Schweitzer, Photochem. Pho-
tobiol. 71 (2000) 12.
[28] I.N. Lykakis, C. Tanielian, M. Orfanopoulos, Org. Lett. 5 (2003) 2875.
[29] C. Tanielian, I.N. Lykakis, R. Seghrouchni, F. Cougnon, M. Orfanopoulos,
J. Mol. Catal. 262 (2007) 170.
[30] I.N. Lykakis, C. Tanielian, R. Seghrouchni, M. Orfanopoulos, J. Mol.
Catal. 262 (2007) 176.
[44] C. Tanielian, C. Schweitzer, R. Seghrouchni, M. Esch, R. Mechin, Pho-
tochem. Photobiol. Sci. 2 (2003) 297.
[45] C. Tanielian, K. Duffy, A. Jones, J. Phys. Chem. B 101 (1997) 4276.
[46] C. Tanielian, R. Seghrouchni, C. Schweitzer, Phys. Chem. A 107 (2003)
1102.
[31] A. Chemseddine, C. Sanchez, J. Livage, J.P. Launay, M. Fournier, Inorg.
Chem. 23 (1984) 2609.
[32] J. Vakros, K. Bourikas, S. Perlepes, C. Kordulis, A. Lycourghiotis, Lang-
muir 20 (2004) 10542.
[33] K. Bourikas, C. Kordulis, A. Lycourghiotis, Environ. Sci. Technol. 39
(2005) 4100.
[34] C.D. Wagner, L.E. Davis, M.V. Zeller, J.A. Taylor, R.H. Raymond, L.H.
Gale, Surf. Interface Anal. 3 (1981) 211.
[35] K. Bourikas, C. Kordulis, A. Lycourghiotis, Catal. Rev. 48 (2006) 363.
[36] P.J.C. Chappell, M.H. Kibel, B.G. Baker, J. Catal. 110 (1998) 139.
[37] L. Karakonstantis, H. Matralis, C. Kordulis, A. Lycourghiotis, J. Catal. 162
(1996) 306.
[38] L. Karakonstantis, K. Bourikas, A. Lycourghiotis, J. Catal. 162 (1996)
295.
[39] M. Fournier, R. Thouvenot, C. Rocchiccioli-Deltcheff, J. Chem. Soc. Fara-
day Trans. 87 (1991) 49.
[40] S. Farhadi, M. Afshari, J. Chem. Res. S 3 (2006) 188.
[47] D.C. Duncan, M.A. Fox, J. Phys. Chem. A 102 (1998) 102.
[48] J. Eriksen, C.S. Foote, J. Phys. Chem. 82 (1978) 2659.
[49] E. Baciocchi, M. Bietti, O. Lanzalunga, Acc. Chem. Res. 33 (2000) 243.
[50] A.L. Perrott, D.R. Arnold, Can. J. Chem. 70 (1992) 272.
[51] R. Popielarz, D.R. Arnold, J. Am. Chem. Soc. 112 (1990) 3068.
[52] E. Baciocchi, F. D’Acunzo, C. Galli, O. Lanzalunga, J. Chem. Soc. Perkin
Trans. 2 (1995) 133.
[53] E. Baciocchi, M. Mattioli, R. Ruzziconi, Tetrahedron Lett. 33 (1992)
1237.
[54] C.S. Foote, Tetrahedron 41 (1985) 2221.
[55] J. Eriksen, C.S. Foote, J. Am. Chem. Soc. 102 (1980) 6083.
[56] J.W. Boyd, P.W. Schmalzl, L.L. Miller, J. Am. Chem. Soc. 102 (1980)
3856.
[57] I. Leray, M. Ayadim, C. Ottermans, H.J.L. Jiwan, J.P. Soumillion, J. Pho-
tochem. Photobiol. A 132 (2000) 43.
[58] I.N. Lykakis, S. Lestakis, M. Orfanopoulos, Tetrahedron Lett. 44 (2003)
6247.