Photocatalytic properties of sodium decatungstate supported on sol-gel silica in the oxidation of glycerol

Article abstract of DOI:10.1016/j.cattod.2011.11.033

Photoexcitation of Na₄W₁₀O₃₂ dissolved in water produces a powerful oxidizing reagent that is able to form hydroxyl radicals from water. OH radicals formation has been demonstrated by EPR spin-trapping technique using DMPO as spin trap. The photooxidation of glycerol occurs through its reaction with OH radicals and is characterized by low selectivity and partial degradation of the substrate to CO₂. Entrapment of Na₄W₁₀O₃₂ inside a silica matrix by a sol-gel procedure gives a new heterogeneous photocatalyst characterized by the presence of micropores (7 and 13 ) and mesopores (30 ). Interaction between polyoxoanion and protonated silanol groups are strong and prevent any kind of leaching. The solid support affects the selectivity of the O₂- assisted photooxidation process of glycerol. In particular, silica surface enhances alcohol adsorption improving its local concentration, so favouring its reaction with photogenerated OH radicals that leads to glyceraldehyde and dihydroxyacetone. At the same time, carbon dioxide is not formed, thus indicating that it is possible to control the high oxidizing power of sodium decatungstate through its heterogenization on a suitable support.

Full text of DOI:10.1016/j.cattod.2011.11.033

Catalysis Today 206 (2013) 46–52
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Photocatalytic properties of sodium decatungstate supported on sol–gel silica in
the oxidation of glycerol
Alessandra Molinaria,∗, Andrea Maldottia,∗, Amra Bratovcica, Giuliana Magnaccab
a Dipartimento di Chimica, Università di Ferrara, Via Luigi Borsari 46, 44123 Ferrara, Italy
b Dipartimento di Chimica IFM; NIS, Center of Excellence, Università di Torino, Via P. Giuria 7, 10125 Torino, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 5 July 2011
Received in revised form 18 October 2011
Accepted 29 November 2011
Available online 2 January 2012
Photoexcitation of Na4W10O32 dissolved in water produces a powerful oxidizing reagent that is able
to form hydroxyl radicals from water. OH• radicals formation has been demonstrated by EPR spin-
trapping technique using DMPO as spin trap. The photooxidation of glycerol occurs through its reaction
with OH• radicals and is characterized by low selectivity and partial degradation of the substrate
to CO2. Entrapment of Na4W10O32 inside a silica matrix by a sol–gel procedure gives a new het-
˚
erogeneous photocatalyst characterized by the presence of micropores (7 and 13 A) and mesopores
Keywords:
˚
(30 A). Interaction between polyoxoanion and protonated silanol groups are strong and prevent any
Heterogeneous photocatalysis
Sodium decatungstate
Glycerol oxidation
EPR spin-trapping
kind of leaching. The solid support affects the selectivity of the O2-assisted photooxidation process
of glycerol. In particular, silica surface enhances alcohol adsorption improving its local concentration,
so favouring its reaction with photogenerated OH• radicals that leads to glyceraldehyde and dihy-
droxyacetone. At the same time, carbon dioxide is not formed, thus indicating that it is possible to
control the high oxidizing power of sodium decatungstate through its heterogenization on a suitable
support.
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
recyclable and that they can be employed in reaction media where
they are insoluble. Moreover, it has been well established that
Photocatalysis with polyoxotungstates for oxidation reactions
continues to be a topic of intense research. It moves toward a “sus-
tainable chemistry” opening to new oxidative routes that require
O2 as available and cheap oxidant, mild temperature and pressure
heterogenization represents a suitable means to tailor efficiency
and selectivity of photocatalytic processes through the control of
the microscopic environment surrounding the photoactive poly-
oxoanion. More specifically, high specific surface area, typical of
micro and mesoporous materials, and the appropriate ratio of sur-
face hydrophobic/hydrophilic character are benefits for improving
efficiency and selectivity of the process, the latter favouring the
enrichment of substrates in the proximity of the photoactive poly-
oxotungstate.
We reported recently about the entrapment of (nBu4N)4W10O32
in a silica matrix by a sol–gel procedure and demonstrated that
the solid support has a strong effect on the oxygen-assisted pho-
tooxidation of 1-pentanol and 3-pentanol in organic solvent [22].
In the present work, we follow an analogous procedure to entrap
Na4W10O32 in a silica matrix. Detailed results on the character-
ization of this material are presented. Our attention is focussed
on the photoreactivity of Na4W10O32 and on the possible effects
of its heterogenization. The photocatalytic activity of the material
is assessed in the photooxidation of glycerol in aqueous medium.
This choice is ascribed to the fact that the extensive functionaliza-
tion of this alcohol with similarly reactive hydroxy groups renders
its selective oxidation particularly difficult. As a consequence, any
improvement in selectivity is a noticeable issue. Moreover, glycerol
conditions and use of light as a renewable source of energy [1–3].
4−
The decatungstate anion W10O32
has been intensively studied
as very attractive photocatalyst for the transformation of organic
4−
compounds. A peculiar property of W10O32
is that it presents
an absorption spectrum in the near UV region (ꢀmax = 323 nm)
that partially overlaps the UV solar emission spectrum, open-
ing the possibility to carry out benign solar-photoassisted
applications [2–9].
Considerable attention has been recently devoted to the het-
4−
erogenization of W10O32 following several procedures [10–22]:
impregnation of porous inorganic networks, incapsulation in silica
matrices by sol–gel methods, immobilization by ionic exchange,
and occlusion in polymeric membranes. Main features of these
photoactive materials are that they are easily handled and
∗
Corresponding authors. Tel.: +39 0532455147; fax: +39 0532240709.
E-mail addresses: alessandra.molinari@unife.it (A. Molinari),
andrea.maldotti@unife.it, mla@unife.it (A. Maldotti).
0920-5861/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2011.11.033

A. Molinari et al. / Catalysis Today 206 (2013) 46–52
47
is the primary by-product in biodiesel production and it is rapidly
2.3. Photocatalytic experiments
becoming a waste product with a related disposal cost. Thus, the
development of new processes able to take advantage of this largely
available and cheap molecule is a urgent need. Research is following
different strategies to exploit glycerol: (i) its selective conver-
sion into oxygenated derivatives that are higher value products
[23,24], (ii) its use, in deaerated conditions, as sacrificial molecule
in reforming [25–27] and photo-reforming processes [28–30] for
H2 production.
Photocatalytic experiments were carried out inside a closed
Pyrex tube of 15 mL capacity at 298 1 K. The desired amount of
photocatalyst (4 × 10−4 M in the case of Na4W10O32 solution or
8 g/L in the case of Na4W10O32/SiO2 suspension) was placed in 3 mL
of an aqueous solution containing glycerol (1 × 10−2 M), and mag-
netically stirred for 20 min. The pyrex photoreactor was filled with
O2 and then joined through an inlet tube to a balloon filled with O2.
Up to now, only TiO2 has been used for the photocatalytic oxi-
dation of glycerol in aerated conditions [31,32]. A common issue is
the need to control the high oxidation power of this photoexcited
semiconductor, which may induce over-oxidation of the primary
products up to carbon dioxide. Good selectivity to dihydroxyace-
tone and glyceraldehyde was obtained with surface fluorinated
TiO2 [31], while the use of rutile or anatase-rutile polymorphic
phases did not prevent degradation of glycerol to CO2 [32]. In this
Photochemical excitation (120 min) was performed by an exter-
nal Helios Q400 Italquartz medium-pressure Hg lamp, selecting
wavelengths higher than 290 nm with a cut off filter. The pho-
ton flux, measured with a MACAMUV203X ultraviolet radiometer,
was 15 mW cm−2. At the end of the photocatalytic experiment,
the sample was analyzed by HPLC system equipped with a pho-
todiode array detector. The column used was an Alltech IOA-1000
Organic Acids, 300 × 7.8 cm by GRACE. Double distilled water was
used as eluent, filtered before use by a vacuum filtration system
equipped with Whatman Nylon Membrane Filters 0.2 ␮m. The
detected products were glyceraldehyde, dihydroxyacetone, glyc-
eric acid. Quantitative analyses were carried out by calibration
curves with commercial products. Each photocatalytic experiment
was repeated three times in order to evaluate the errors, which
never exceeded 5%. Control experiments were run irradiating
SiO2 suspended in the solution containing glycerol (1 × 10−2 M)
or keeping (120 min) the photocatalyst dispersed in the solution
in the dark. When Na4W10O32/SiO2 was employed as a photocat-
alyst, the possible release of polyoxoanion in the solution phase
was evaluated by UV–vis analysis of the solution at the end of
irradiation.
In order to establish the amount of CO2 eventually formed, sam-
ples were irradiated maintaining the reactor firmly closed. At the
end of irradiation 1.5 ml of a NaOH solution (0.1 M) were put inside
the reactor with a syringe and mixed with the irradiated solution.
Then, 3.5 mL of the aqueous solution were taken and kept into a
vial. After the addition of 1 mL of saturated citric acid solution, car-
bon dioxide released was measured by pH meter BasiC 20 CRISON
equipped with the gas sensing probe (Crison 9666). Quantitative
analysis was made through a calibration curve built from standard
solutions of NaHCO3 treated as described above. Results obtained
from the irradiated samples were compared with those coming
from analogous samples kept in the dark for the same period. The
yield of CO2 was referred to the number of carbon atoms present
in the alcohol molecule.
respect, heterogenized Na4W10
alternative to titania systems.
O32 has been evaluated to be a valid
2. Experimental
2.1. Photocatalyst preparation and characterization
Reagents and solvents were purchased in the highest puri-
ties available from Sigma and used without further purification.
Sodium decatungstate (Na4W10O32
) was prepared following
a literature procedure [33]. The heterogeneous photocatalyst
(Na4W10O32/SiO2) was prepared by hydrolysis of tetraethy-
lorthosilicate (TEOS) in the presence of an acid aqueous solution
of Na4W10O32, following a procedure recently published in detail
for the entrapment of (nBu4N)4W10O32 in silica [22]. On the basis
of the solid mass obtained at the end of the preparation and the
initial amount of Na4W10O32 employed (0.5 g), we could estimate
that the prepared material Na4W10O32/SiO2 contained 30% (w/w)
of decatungstate. The photocatalyst was washed several times with
aliquots of water. UV–vis spectra of these aliquots showed that the
release of Na4W10O32 into the solution was negligible. A sample
not including decatungstate (SiO2) was also prepared, following
the procedure described above but without adding Na4W10O32
.
Diffuse reflectance UV–vis spectra were recorded with a Jasco
V-570 using an integrating sphere and BaSO4 as reference. The plot-
ted spectra were obtained by the Kubelka–Munk transformation
(F(R) = 1 − R2/2R) versus the wavelength. N2 adsorption–desorption
experiments were carried out at 77 K by means of ASAP2020 instru-
ment (Micromeritics). Before each measurement, the sample was
outgassed overnight at 150 ◦C with a rotative pump (residual pres-
sure about 10−2 mbar).
3. Results and discussion
3.1. Textural characterization
Fig. 1a shows the UV–vis spectrum of heterogeneous
Na4W10O32/SiO2 (curve A). Spectra of homogeneous Na4W10O32
and of silica, obtained with the sol–gel procedure, are also reported
for a comparison (curves B and C, respectively). It is seen that
2.2. EPR spin-trapping experiments
EPR spin-trapping experiments were carried out with
Bruker 220 SE spectrometer equipped with a TE 201 resonator,
at microwave frequency of 9.4 GHz. The samples for epr
a
Na4W10O32/SiO2 presents an intense absorption at wavelengths
4−
a
lower than 400 nm that is typical of W10O32
and that is not
spin-trapping experiments were H2O/CH3CN (85/15) solutions
containing 5,5-dimethylpyrroline N-oxide (DMPO, 3 × 10−2 M) as
spin trap, Na4W10O32 (2.4 × 10−3 M) and, when requested, glyc-
erol (1 × 10−2 M). CH3CN was necessary in order to dissolve the
spin trap. In heterogeneous experiments, the solid Na4W10O32/SiO2
was suspended in the solution containing DMPO and, when neces-
sary, glycerol as described above. Then, samples were put into a flat
quartz cell and directly irradiated into the EPR cavity at ꢀ > 290 nm
with a Hg medium pressure mercury lamp. No EPR signals were
obtained in the dark and during irradiation of the solution in the
absence of photocatalyst.
present in the silica matrix. Fig. 1b reports the visible spectrum
of Na4W10O32/SiO2 irradiated in the presence of glycerol under
anaerobic conditions (curve A). It is observed an important
increase of absorption above 600 nm, in the region where the blue
5−
W10O32
typically absorbs (see for comparison curve C relative
to homogeneous reduced decatungstate). Exposure to air of the
irradiated powder leads to the reoxidation of decatungstate in its
oxidized state that not absorbs in the visible region (curve B). These
results are in line with the known primary photochemical process
of decatungstate anion [4] and represent a further evidence that
the polyoxoanion is unchanged in nature after its heterogenization

48
A. Molinari et al. / Catalysis Today 206 (2013) 46–52
Fig. 2 shows the incremental pore volume as a function of the
a
pore width obtained via the DFT method that allows us to rec-
2,0
1,5
1,0
0,5
0,0
˚
ognize the fraction of pores with widths between 10 and 20 A.
Na4W10O32/SiO2 is characterized by micropores with width of
˚
A
about 7 A, in agreement with the previous literature data on the
entrapment of polyoxometalates in silica matrices via sol–gel tech-
niques [18]. In addition to these micropores, Na4W10O32/SiO2 also
˚
˚
contains micropores around 13 A and mesopores around 30 A.
Table 1 and Fig. 2 show that the morphological features of
Na4W10O32/SiO2 differ markedly from those of its parent mate-
˚
rial SiO2. In fact, SiO2 does not present 7 A micropores, but 75%
B
of its porosity is characterized by micropores with width of
C
˚
11–12 A. Moreover, the degree of mesoporosity is lower than in
Na4W10O32/SiO2. These results indicate that the incorporation of
sodium decatungstate during the synthesis of the silica-based pho-
tocatalyst induces an important modification of the porous texture
of the siliceous material, suggesting that sodium decatungstate can
act as a templating agent.
300
400
500
600
wavelength (nm)
It has been reported in previous works that the formation
of microporous decatungstate is due to the chemical interaction
between sodium decatungstate and the silica network [18]. Since
during the preparation the hydrolysis of TEOS in the presence of
Na4W10O32 is carried out at pH 2, the silanol group ( Si OH)
b
A
0,4
0,2
0,0
is protonated in the acidic medium to form Si OH2+. This can,
4−
in turn, act as a counterion for W10O32
(
yielding the couple
Si OH2+) (Na3W10O32−) by acid–base reaction. This strong inter-
action could explain why the salt is no longer released although
Na4W10O32/SiO2 is dispersed in aqueous media. It is noteworthy
to evidence that the substitution of Na+ ions with organic and less
polar groups such as tetrabutylammonium ions [22] decreases the
C
B
4−
strength of the interaction between W10O32 and SiO2 matrix, as
observed in the effect induced by CH3CN washing of the material
(loss of 20% of the starting decatungstate amount).
400
500
600
700
800
3.2. EPR spin-trapping investigation
wavelength (nm)
EPR spin-trapping investigation is a powerful technique for
detecting the formation of short-lived radicals [35] and has been
fruitfully employed in photochemical studies on polyoxometalates
in order to better understand photochemical primary processes
[36–41]. This technique is based on the ability of some molecules
such as nitrones to trap radicals to give nitroxides stable enough
to be successfully detected and studied. The nature of the trapped
radical can be often identified by the parameters obtainable from
the EPR spectrum.
Evidences of radical formation as a consequence of photoexci-
tation of Na4W10O32 in homogeneous solution have been obtained
irradiating (ꢀ > 290 nm) H2O/CH3CN (85/15, v/v) solutions contain-
ing Na4W10O32 (2.4 × 10−3 M), glycerol (1 × 10−2 M) and DMPO
(3 × 10−2 M) as a spin trap, directly inside the EPR cavity. Illu-
mination causes the prompt formation of a quartet (1:2:2:1,
aN = aH = 14.5 gauss), which is shown in Fig. 3. Signal pattern and
coupling constant values are in agreement with the trapping of OH•
radicals by DMPO to form the paramagnetic adduct [DMPO–OH]•
according to step 2 of Scheme 1 [35]. No signals relative to the trap-
ping of radicals coming from glycerol are detected even after some
minutes irradiation. The same quartet is also obtained during irra-
diation of Na4W10O32 dissolved in an aqueous solution containing
the spin trap but in the absence of glycerol, suggesting that the
Fig. 1. (a) UV–vis spectra of Na4W10O32/SiO2 (curve A), of homogeneous Na4W10O32
(curve B) and of SiO2 (curve C). (b) Visible spectra of reduced Na4W10O32/SiO2 (curve
A), of oxidized Na4W10O32/SiO2 (curve B) and of reduced homogeneous Na4W10O32
(curve C).
Gas-volumetric adsorption of N2 at 77 K gives useful information
about the morphological features of Na4W10O32/SiO2 in compari-
son with SiO2. Table 1 reports that the introduction of Na4W10O32
into the silica matrix causes a decrease in surface specific area value.
At the same time it is observed a 35% increase in the total pore
volume.
Na4W10O32/SiO2 material presents both mesoporosity and
microporosity (Table 1). To evaluate both and, in particular, to
reveal the porosity eventually present at the boundary between
micro and mesoporosity regions, we employ here the same method
recently adopted by us for the investigation of (n-Bu4N)4W10O32
encapsulated inside a silica matrix [22] that consists in the isotherm
deconvolutions in accordance to density functional theory (DFT)
[34]. The pore volumes reported in Table 1 and Fig. 2 are classi-
fied on the basis of the pore size distribution determined via the
DFT method and quantified considering the DFT cumulative pore
volume curves.
Table 1
Morphological features of Na4W10O32/SiO2.
SSA/m2g−1
Vtot/cm3g−1
Vmicro (<18 A width, cm3 g−1
)
Vmeso (>18 A width, cm3 g−1
)
˚
˚
Sample
SiO2
Na4W10O32/SiO2
772
521
0.176
0.240
0.131
0.162
0.045
0.078

A. Molinari et al. / Catalysis Today 206 (2013) 46–52
49
0.020
0.015
0.010
0.005
0.000
10
20
30
40
50
60
70
80
90
100
Pore Width (Angstroms)
˚
Fig. 2. DFT pore analysis for Na4W10O32/SiO2 (cross symbols) and for SiO2 (solid line). The vertical broken line represents the threshold value considered (18 A) between
micro and mesopores.
alcohol does not take part directly to the primary photoprocess.
Control experiments show that no signal is observed neither in the
dark nor during irradiation but in the absence of decatungstate.
The formation of OH• radicals may be ascribed to
the reaction between photoexcited polyoxoanion and
water according to Scheme
obtain OH• radicals by direct reaction between photoex-
cited pentakis(isopropylammonium)hydrogenhexatungstate
1 (step 1). The possibility to
[NH3Pri]5[W6O20(OH)] and water was pointed out by Yamase
in the early eighties [42]. More recently, Papaconstantinou’s
group and other researchers proposed that polyoxometalates and
TiO2 exhibit overall similar photocatalytic behaviour in aqueous
solution [1,5,43,44]. This similarity was attributed to the common
formation of OH• radicals. In those works, the formation of OH•
radicals was suggested by the detection of hydroxylation products
in photocatalytic experiments with aromatic hydrocarbons, and
on the basis of the excited state potentials [43].
3425
3450
3475
3500
magnetic field (gauss)
The photogenerated OH• radicals can initiate glycerol oxida-
tion by abstraction of a hydrogen atom from one of the available
C␣ atoms, producing a hydroxy alkyl radical (step 3). Then, these
radicals may be further oxidized to the corresponding carbonylic
compounds (step 4). Reaction between reduced decatungstate and
O2 regenerates the photocatalyst in its initial state (step 5). In our
Fig. 3. EPR spectrum of [DMPO–OH]• obtained upon irradiation (ꢀ > 290 nm) of
Na4W10O32 dissolved in an aqueous solution containing glycerol.
4−
W10O324-* + H2O
photocatalytic system, direct reaction of photoexcited W10O32
.
[DMPO-OH]
with the alcoholic substrate seems not to be the main pathway for
glycerol oxidation since no radicals coming from the poly-alcohol
are trapped by DMPO.
(1)
(2)
DMPO
.
W10O325- + OH + H+
O2 + H+
Fig. 4 shows that addition of glycerol causes a decrease of sev-
eral times of the signal intensity of [DMPO–OH]•. This indicates
that glycerol is a very good competitor for the reaction with OH•
radical with respect to the spin trap (steps 2 and 3). It is also seen
that the signal intensity of the paramagnetic [DMPO–OH]• adduct
decreases rapidly after the first 120 s irradiation. Likely, this adduct
can react with other radical species formed during irradiation (see
for example step 5) or with the decatungstate itself, as reported in
the past by some researchers [45].
h
ν
(5)
(3) Glycerol
.
4-
W10O32
HO2
OH
OH
+ H2O
OH
OH
OH
HO
HO
HO
+
.
.
(4)
(4)
It has been previously reported that the [DMPO–OH]• species
may originate from degradation of the adduct [DMPO–OOH]•
between DMPO and O2−. [46,47]. This paramagnetic species
presents a typical 12-lines EPR spectrum [48]. On the other hand,
this source of [DMPO–OH]• is expected to be negligible in our
photocatalytic system on the basis of literature data indicating
that a maximum level of approximately 3% [DMPO–OH]• should
originate from [DMPO–OOH]• [46,49]. In line with this statement,
Fig. 4 shows that the [DMPO–OH]• spectrum is the only observed
O
OH
O
OH
(DHA)
(GAD)
Scheme 1. Proposed reaction mechanism for glycerol oxidation by photoexcited
Na4W10O32 in aqueous medium.

50
A. Molinari et al. / Catalysis Today 206 (2013) 46–52
4,5E+04
3,0E+04
1,5E+04
0,0E+00
its local concentration in proximity of the decatungstate, so
favouring its reaction with the photogenerated OH• (Scheme 1, step
3) in competition with trapping by DMPO (step 2).
Adsorption of alcohols and diols on various siliceous matri-
ces is an efficient phenomenon that has been already pointed
out [12,14,50]. For this reason, glycerol is expected to adsorb on
Na4W10O32/SiO2 too. A qualitative confirmation of this behaviour
is obtained by the experiment described in the following: a
huge amount of Na4W10O32/SiO2 (70 g/L) was suspended in a
small volume (0.7 mL) of the aqueous solution containing glycerol
(1 × 10−2 M) and stirred in the dark for 30 min. From HPLC analy-
sis of the solution, performed before and after the contact with the
photocatalyst, we observed a decrease of about 35% of the initial
amount of glycerol, indicating that it can adsorb and concentrate
on the surface.
0
100
200
300
400
500
irradiation time (s)
Fig. 4. Fixed-field EPR signal intensity of the [DMPO–OH]• adduct in time upon irra-
diation (ꢀ > 290 nm) of a H2O/CH3CN solution (85/15, v/v) containing Na4W10O32
(2.4 × 10−3 M) and DMPO (3 × 10−2 M) (squares) and of an analogous solution con-
taining Na4W10O32 (2.4 × 10−3 M) and DMPO (3 × 10−2 M) and glycerol (1 × 10−2 M)
(circles).
3.3. Photocatalytic activity
Typical photocatalytic experiments were carried out irradiating
(ꢀ > 290 nm) Na4W10O32 (4 × 10−4 M) or Na4W10O32/SiO2 (8 g/L) in
aqueous solutions containing glycerol (1 × 10−2 M) at room tem-
perature and under 760 Torr of O2. Owing to the presence of three
oxidizable hydroxyl groups, a broad range of glycerol derivatives
can be formed. Among them, we decided to follow by HPLC the
formation of the carbonylic compounds, glyceraldehyde (GAD) and
dihydroxyacetone (DHA), and of glyceric acid (GA), the less oxi-
dized among acids that is obtained from further oxidation of the
aldehyde. Moreover, CO2 has been evaluated as final degradation
product.
Irradiation of Na4W10O32 in homogeneous solution leads to the
formation of GAD and DHA as primary oxidized products (Table 2).
They represent about 75% of all the monitored compounds. Accord-
ing to EPR spin-trapping evidences, the formation of GAD and DHA
should occur through steps 3 and 4 of Scheme 1. The subsequent
easy oxidation of GAD may lead to the corresponding carboxylic
acid. In fact, GA is about 10% of the converted glycerol.
This photocatalytic system shows quite strong oxidizing abil-
ity that it is difficult to control for synthetic purposes. In fact, the
total yield to the primary photoproducts GAD, DHA and GA is only
the 47% of the converted glycerol, which, likely undergoes over-
oxidation to yield a mixture of undetected C3 and C2 oxidized
products, such as mesoxalic, tartronic, oxalic and glycolic acids. This
statement is confirmed by the data of Table 2, which show that the
photocatalytic experiment leads also to the formation of significant
amounts of carbon dioxide.
These results are in line with EPR spin-trapping findings that
demonstrate the formation of OH• radicals, which are very strong
and unselective oxidants able to degrade both the initial substrate
and the reaction intermediates. This photocatalytic behaviour of
Na4W10O32 is also in agreement to what claimed by Papacon-
stantinou’s group, who attributed the photocatalytic reactivity of
polyoxotungstate anions toward phenols and 2-propanol in water
to the formation of OH• radicals [43].
already after few seconds irradiation and that the presence of a
scavenger of OH• radicals such as glycerol reduces significantly its
intensity.
Analogous EPR spin-trapping experiments with heterogeneous
Na4W10O32/SiO2 were carried out suspending the photocatalyst
in the solution containing DMPO and glycerol, when necessary.
Results obtained are qualitatively similar to those in homoge-
neous phase: the unique signal observed is the quartet relative to
the formation of [DMPO–OH]•, and in the presence of glycerol a
severe decrease of intensity is observed again (Fig. 5). This indi-
cates that the mechanism proposed in homogeneous phase is still
working when decatungstate is entrapped inside the silica matrix.
Intensities of signal obtained upon irradiation of Na4W10O32/SiO2
are significantly lower than those observed in homogeneous con-
ditions. This result can be tentatively ascribed to the fact that
adsorption phenomena of glycerol on silica are expected to enhance
7500
5000
2500
0
0
100
200
300
400
500
irradiation time (s)
Fig. 5. Fixed-field EPR signal intensity of the [DMPO–OH]• adduct in time upon irra-
diation (ꢀ > 290 nm) of Na4W10O32/SiO2 dispersed in a H2O/CH3CN solution (85/15,
v/v) containing DMPO (3 × 10−2 M) (squares) and in an analogous solution contain-
ing also glycerol (1 × 10−2 M) (circles).
Table 2
Photocatalytic oxidation of glycerol by Na4W10O32 and Na4W10O32/SiO2.a
Photocatalytic system
Converted glycerol (␮mol)
Detected products ␮mol (selectivity %)b
c
GAD
DHA
GA
CO2
Na4W10O32
Na4W10O32/SiO2
19.2
9.6
6.3(32.8)
5.7(59.4)
0.54 (2.8)
0.54 (5.6)
1.8 (9.4)
0.51 (5.3)
0.45 (2.3)
< 0.03 (0.3)
a
In a typical experiment Na4W10O32 (4 × 10−4 M) or Na4W10O32/SiO2 (8 g/L) were put in aqueous solutions (3 mL) containing glycerol (1 × 10−2 M) and irradiated (120 min)
at ꢀ > 290 nm at 298 1 K and 760 Torr of O2. Reported values are the mean of three repeated experiments.
b
Selectivity % is expressed as ␮mol of product divided by ␮mol of converted glycerol. The remaining to 100% is represented by other oxidized derivatives that have not
been identified.
c
In order to compare directly the values of carbon dioxide with those of GAD, DHA and GA, the ␮mol of CO2 have been divided by three.

A. Molinari et al. / Catalysis Today 206 (2013) 46–52
51
Photocatalytic properties of Na4W10O32 in the oxidation of
glycerol change significantly after its entrapment inside the sil-
ica matrix. Typical experiments were carried out irradiating the
photocatalyst (8 g/L) suspended in an aqueous solution contain-
ing the alcohol. The amount of 8 g/L was chosen on the basis of
measurements with an ultraviolet radiometer that indicate that
photocatalyst absorbs 90% of the impinging radiation at 313 nm,
which is the emission line of the employed light source closest
to the absorption maximum of decatungstate at 323 nm. Control
experiments give evidences that (i) irradiation of a dispersion of
SiO2 does not lead to the formation of detectable amounts of
oxidation products, allowing us to exclude any kind of photoac-
tivation of the silica matrix; (ii) no oxidation of glycerol occurs
after the contact between the heterogenized decatungstate and the
alcohol in the absence of light. Moreover, UV–vis spectra of solu-
tions recovered after the irradiation of Na4W10O32/SiO2 show that
decatungstate is not released from the support. Further irradiation
of these solutions does not accumulate extra oxidation products,
thus indicating that the photocatalytic process is truly heteroge-
neous in nature.
indicating that incorporation of decatungstate induces an impor-
tant modification of the porous texture of the siliceous material.
Interactions between polyoxoanion and protonated silanol groups
are strong and the photocatalyst is never released from the solid
matrix.
The photochemical characterization of Na4W10O32/SiO2 in the
O2-assisted oxidation of glycerol shows that the solid support
has a strong effect on the selectivity. In particular, silica surface
favours adsorption of glycerol, enhancing its local concentration.
This facilitates its reaction with OH• radicals photogenerated by
decatungstate placed on the surface or inside pores. As a conse-
quence, a significative improvement in selectivity toward primary
oxidation compounds is obtained with negligible amounts of car-
bon dioxide. The results reported in this study provide new insights
into the role of textural features in tuning the photocatalytic prop-
erties of Na4W10O32. Comprehension of the morphology and of
surface properties can represent the starting point for developing
selective photocatalytic systems, an issue of particular relevance
when the oxidizable substrates present large functionalization with
groups of similar reactivity.
Irradiation of Na4W10O32/SiO2 leads mainly to the accumula-
tion of GAD and DHA. In fact, they represent more than 90% of
the monitored products. Further oxidation of GAD–GA occurs only
in low degree, since the amount of acid accumulated during the
photocatalytic experiment is four times lesser than that formed in
the homogeneous sample. Moreover, the sum of GAD, DHA and GA
is about the 70% of the converted glycerol and only a negligible
amount of carbon dioxide is detected. All these findings indicate
that heterogenization has an important effect in decreasing the
oxidizing ability of sodium decatungstate.
References
[1] A. Hiskia, A. Mylonas, E. Papaconstantinou, Chem. Soc. Rev. 30 (2001) 62–69.
[2] A. Maldotti, A. Molinari, R. Amadelli, Chem. Rev. 102 (2002) 3811–3836.
[3] M. Fagnoni, D. Dondi, D. Ravelli, A. Albini, Chem. Rev. 107 (2007) 2725–2756.
[4] C. Tanielian, Coord. Chem. Rev. 178–180 (1998) 1165–1181.
[5] P. Kormali, A. Troupis, T. Triantis, A. Hiskia, E. Papaconstantinou, Catal. Today
124 (2007) 149–155.
[6] A. Maldotti, A. Molinari, Top. Curr. Chem. 303 (2011) 185.
[7] C. Tanielian, I.N. Likakis, R. Seghrouchni, F. Cougnon, M. Orfanopoulos, J. Mol.
Catal. A: Chem. 262 (2007) 170–175.
[8] I.N. Likakis, C. Tanielian, R. Seghrouchni, M. Orfanopoulos, J. Mol. Catal. A: Chem.
262 (2007) 176–184.
[9] M.D. Tzirakis, I.N. Lykakis, M. Orfanopoulos, Chem. Soc. Rev. 38 (2009)
2609–2621.
[10] A. Maldotti, A. Molinari, G. Varani, M. Lenarda, L. Storaro, F. Bigi, R. Maggi, A.
Mazzacani, G. Sartori, J. Catal. 209 (2002) 210–216.
[11] A. Maldotti, R. Amadelli, I. Vitali, L. Borgatti, A. Molinari, J. Mol. Catal. A: Chem.
204–205 (2003) 703–711.
[12] A. Molinari, R. Amadelli, A. Mazzacani, G. Sartori, A. Maldotti, Langmuir 18
(2002) 5400–5405.
[13] A. Molinari, G. Varani, E. Polo, S. Vaccari, A. Maldotti, J. Mol. Catal. A: Chem. 262
(2007) 156–163.
[14] A. Maldotti, A. Molinari, F. Bigi, J. Catal. 253 (2008) 312–317.
[15] A. Maldotti, A. Molinari, in: D.L. Marmaduke (Ed.), Progress in Heterogeneous
Catalysis, Nova Science Publishers, New York, 2008, pp. 1–15.
[16] M. Bonchio, M. Carraro, G. Scorrano, A. Bagno, Adv. Synth. Catal. 345 (2003)
1119–1126.
[17] M. Bonchio, M. Carraro, M. Gardan, G. Scorrano, E. Orioli, E. Fontananova, Top.
Catal. 40 (2006) 133–140.
[18] Y. Guo, C. Hu, X. Wang, Y. Wang, Y. Zou, H. Ding, S. Feng, Chem. Mater. 13 (2001)
4058–4064.
[19] Y. Guo, C. Hu, J. Mol. Catal. A: Chem. 262 (2007) 136–148.
[20] L. Ni, J. Ni, Y. Lv, P. Yang, Y. Cao, Chem. Commun. (2009) 2171–2173.
[21] S. Farhadi, M. Zaidi, Appl. Catal. A: Chem. 354 (2009) 119–126.
[22] A. Molinari, A. Bratovcic, G. Magnacca, A. Maldotti, Dalton Trans. 39 (2010)
7826–7833.
The EPR spin-trapping investigation described above indicates
that heterogenization does not affect the ability of Na4W10O32 to
oxidize water to hydroxyl radicals (Scheme 1, step 1). Therefore, it
should be not surprising that differences in selectivity can critically
depend on textural features that allow the tuning of photocatalytic
properties of Na4W10O32 through the control of surface interac-
tions with substrates and intermediates [40,43,44]. In particular,
EPR spin-trapping experiments give the indication that the reac-
tion between OH• radical and glycerol occurs efficiently also in the
heterogeneous system (Fig. 5). In agreement with this result, it is
found that photocatalytic efficiency undergoes a decrease of only
25% after heterogenization. Likely, as mentioned above, adsorption
phenomena of glycerol on silica enhance its local concentration
in proximity of decatungstate. This substrate accumulation on the
surface can favour, on the one hand, the reaction between glyc-
erol and the photogenerated OH• radical and, on the other hand,
can prevent the subsequent oxidation of the formed carbonylic
compounds.
[23] M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi, C. Della Pina, Angew. Chem. Int.
Ed. 46 (2007) 4434–4440.
4. Conclusions
[24] C. Zhou, J.N. Beltramini, Y. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008) 527–549.
[25] D.A. Simonetti, E.L. Kunkes, J.A. Dumesic, J. Catal. 247 (2007) 298–306.
[26] E.L. Kunkes, R.R. Soares, D.A. Simonetti, J.A. Dumesic, Appl. Catal. B: Environ. 90
(2009) 693–698.
[27] T. Montini, R. Singh, P. Das, B. Lorenzut, N. Bertero, P. Riello, A. Benedetti, G.
Giambastiani, C. Bianchini, S. Zinoviev, S. Miertus, P. Fornasiero, Chem. Sus.
Chem. 3 (2010) 619–628.
[28] V.M. Daskalaki, D.I. Kondarides, Catal. Today 144 (2009) 75–80.
[29] V. Gombac, L. Sordelli, T. Montini, J.J. Delgado, A. Adamski, G. Adami, M.
Cargnello, S. Bernal, P. Fornasiero, J. Phys. Chem. A 114 (2010) 3916–3925.
[30] T. Montini, V. Gombac, L. Sordelli, J.J. Delgado, X. Chen, G. Adami, P. Fornasiero,
Chem. Catal. Chem. 3 (2011) 574–577.
[31] V. Maurino, A. Bedini, M. Minella, F. Rubertelli, E. Pelizzetti, C. Minero, J. Adv.
Oxid. Technol. 11 (2008) 184–192.
[32] V. Augugliaro, H.A. Hamed El Nazer, V. Loddo, A. Mele, G. Palmisano, L.
Palmisano, S. Yurdakal, Catal. Today 151 (2010) 21–28.
[33] F. Bigi, A. Corradini, C. Quarantelli, G. Sartori, J. Catal. 250 (2007) 222–230.
[34] D.L. Ou, P.M. Chevlier, I.A. Mackinnon, K. Eguchi, R. Boisvert, K. Su, J. Sol–gel Sci.
Technol. 26 (2003) 407–412.
Photoexcitation of Na4W10O32 dissolved in water produces a
powerful oxidizing reagent. The oxidizing ability is manifested
through formation of hydroxyl radicals arising from the reaction
of excited Na4W10O32 and water. OH• radical formation has been
demonstrated by EPR spin-trapping technique using DMPO as spin
trap. Photoexcited homogeneous Na4W10O32 is an efficient catalyst
for the oxidation of glycerol but, since its photocatalytic activity
depends on OH• radicals, it shows low selectivity and leads to over
oxidation and complete degradation of the substrate.
Entrapment of Na4W10O32 inside a silica matrix by a sol–gel
procedure gives a heterogeneous photocatalyst characterized by
˚
˚
the presence of micropores of about 7 A and 13 A and mesopores
˚
of about 30 A. The morphological features of Na4W10O32/SiO2 dif-
fer from those of SiO2 prepared following the same procedure,

52
A. Molinari et al. / Catalysis Today 206 (2013) 46–52
[35] E.G. Janzen, Acc. Chem. Res. 4 (1971) 31–40.
[36] A. Maldotti, R. Amadelli, V. Carassiti, A. Molinari, Inorg. Chim. Acta 256 (1997)
309–312.
[37] A. Molinari, R. Amadelli, V. Carassiti, A. Maldotti, Eur. J. Inorg. Chem. (2000)
91–96.
[43] A. Mylonas, A. Hiskia, E. Androulaki, D. Dimotikali, E. Papaconstantinou, Phys.
Chem. Chem. Phys. 1 (1999) 437–440.
[44] S. Kim, H. Park, W. Choi, J. Phys. Chem. B 108 (2004) 6402–6411.
[45] J. Zakrzewski, C. Giannotti, J. Photochem. Photobiol. A: Chem. 57 (1991)
453–455.
[38] M.L. Ganadu, L. Andreotti, I. Vitali, A. Maldotti, A. Molinari, G.M. Mura, Pho-
tochem. Photobiol. Sci. 1 (2002) 951–954.
[46] E. Finkelstein, G.M. Rose, E.J. Rauckman, Mol. Pharmacol. 21 (1982) 262–
265.
[39] D. Dondi, M. Fagnoni, A. Molinari, A. Maldotti, A. Albini, Chem. Eur. J. 10 (2004)
142–148.
[47] A. Lipovsky, Z. Tzitrinovich, H. Friedmann, G. Applerot, A. Gedanken, R. Lubart,
J. Phys. Chem. C 113 (2009) 15997–16001.
[40] A. Molinari, M. Montoncello, H. Rezala, A. Maldotti, Photochem. Photobiol. Sci.
8 (2009) 613–619.
[48] G.R. Buettner, L.W. Oberley, S.W.H.C. Leuthauser, Photochem. Photobiol. 28
(1978) 693–695.
[41] D. Dondi, D. Ravelli, M. Fagnoni, M. Mella, A. Molinari, A. Maldotti, A. Albini,
Chem. Eur. J. 15 (2009) 7949–7957.
[42] T. Yamase, Inorg. Chim. Acta 76 (1983) L25–L27.
[49] G.R. Buettner, Free Radic. Res. Commun. 19 (1993) S79–S87.
[50] A. Molinari, A. Maldotti, A. Bratovcic, G. Magnacca, Catal. Today 161 (2011)
64–69.