WO3-TiO2 vs. TiO2 photocatalysts: Effect of the W precursor and amount on the photocatalytic activity of mixed oxides

Article abstract of DOI:10.1016/j.cattod.2013.01.008

Aiming at producing TiO₂-based photocatalytic materials with reduced charge carriers recombination, WO₃-TiO₂ mixed oxides were synthesized by a sol-gel method employing either an inorganic salt, Na₂WO₄, or an organic alkoxide, W(OC₂H ₅)₆, as tungsten precursor, with different W/Ti ratios. The so-obtained materials were characterized by XRPD, BET, UV-vis reflectance, XPS and EDX analyses and their photoactivity was tested under UV-visible irradiation in both the mineralization of formic acid in aqueous suspension and the gas phase oxidation of acetaldehyde. Both photoactivity results and photocurrent measurements point to a superior performance of photocatalysts obtained from the organic precursor with an optimal tungsten content (3%). The formation of an intimately mixed oxide, as revealed by XRPD analysis, results in photoactivity higher than that of pure TiO₂, and also of benchmark P25 TiO₂, consequent to a better charge separation due to the migration of photoproduced holes from WO₃ domains to TiO₂ and of photopromoted electrons in the opposite direction. The persistence of pure anatase phase in W-containing photocatalysts also after calcination at 700 C and their higher surface area with respect to pure TiO₂ also contribute in increasing the photocatalytic activity of the WO ₃-TiO₂ mixed oxides.

Full text of DOI:10.1016/j.cattod.2013.01.008

Catalysis Today 209 (2013) 28–34
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
WO3–TiO2 vs. TiO2 photocatalysts: effect of the W precursor and
amount on the photocatalytic activity of mixed oxides
Francesca Ribonia, Luca G. Bettinib,c, Detlef W. Bahnemannd, Elena Sellia,c,∗
a Dipartimento di Chimica, Università degli Studi di Milano, Via Golgi 19, I-20133 Milano, Italy
b Dipartimento di Fisica, Università degli Studi di Milano, Via Celoria 16, I-20133 Milano, Italy
c CIMaINa, Università degli Studi di Milano, Via Celoria 16, I-20133 Milano, Italy
d Institut für Technische Chemie, Leibniz Universität Hannover, Callinstrasse 3, D-30167 Hannover, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 30 June 2012
Received in revised form 7 January 2013
Accepted 8 January 2013
Available online 13 April 2013
Aiming at producing TiO2-based photocatalytic materials with reduced charge carriers recombination,
WO3–TiO2 mixed oxides were synthesized by a sol–gel method employing either an inorganic salt,
Na2WO4, or an organic alkoxide, W(OC2H5)6, as tungsten precursor, with different W/Ti ratios. The so-
obtained materials were characterized by XRPD, BET, UV–vis reflectance, XPS and EDX analyses and their
photoactivity was tested under UV–visible irradiation in both the mineralization of formic acid in aque-
ous suspension and the gas phase oxidation of acetaldehyde. Both photoactivity results and photocurrent
measurements point to a superior performance of photocatalysts obtained from the organic precursor
with an optimal tungsten content (3%). The formation of an intimately mixed oxide, as revealed by XRPD
analysis, results in photoactivity higher than that of pure TiO2, and also of benchmark P25 TiO2, conse-
quent to a better charge separation due to the migration of photoproduced holes from WO3 domains to
TiO2 and of photopromoted electrons in the opposite direction. The persistence of pure anatase phase in
W-containing photocatalysts also after calcination at 700 ◦C and their higher surface area with respect
to pure TiO2 also contribute in increasing the photocatalytic activity of the WO3–TiO2 mixed oxides.
© 2013 Elsevier B.V. All rights reserved.
Keywords:
WO3/TiO2 photocatalyst
Formic acid mineralization
Acetaldehyde photocatalytic oxidation
Photocurrent measurements
1. Introduction
rate of electron transfer to reducible species (e.g. oxygen in oxida-
tion reactions, protons in hydrogen production from water) and a
Photocatalysis has been widely studied in the last decades as
one of the most effective techniques for the abatement of organic
and inorganic pollutants of water and air, as well as a mean to
convert solar into chemical energy. Photocatalytic reactions occur-
ring on the surface of semiconductor materials notoriously proceed
through the photoproduction of electron/hole (e-/h+) pairs after
absorption of photons with energy equal to or greater than the
semiconductor band gap (Eg). This is followed by space–charge sep-
aration by the traps available on the semiconductor surface and
by charge transfer processes involving the separated electrons and
holes and species adsorbed on the semiconductor surface [1,2].
Titanium dioxide is by far the most used semiconductor,
although it has two main defects. Firstly, it is a relatively high
energy band gap material (Eg ≈ 3.2 eV) that can be excited under
UV irradiation (ꢀ < 387 nm). This practically rules out the use of sun-
light to efficiently induce photocatalytic reactions. Secondly, a low
high recombination rate of the photoproduced e−/h+ pairs limit the
rate of photocatalytic reactions.
In fact, a photocatalyst can be envisaged as a microreactor per-
forming both reduction and oxidation half-reactions at expressly
designed surface sites [3]. Since both processes usually occur
simultaneously on sites often separated by a distance of only few
angstroms, designing efficient photocatalysts with enhanced sepa-
ration between the photoproduced charge carriers would increase
the efficiency of photocatalytic processes.
Among the strategies adopted to minimize these drawbacks in
TiO2-based photocatalysts, coupling TiO2 with another metal oxide
appears one of the most promising routes. For instance, by choosing
a metal oxide having the conduction band (CB) lower in energy than
the CB of TiO2 and the valence band (VB) almost at the same value or
even lower in energy than the VB of TiO2, photopromoted electrons
are transferred from the CB of TiO2 to the CB of the coupled metal
oxide, while photogenerated holes might be trapped within the
TiO2 particles.
Recent examples of effective mixed oxides photocatalysts with
increased photocatalytic activity in either liquid or gas phase reac-
tions include sol–gel synthesized Fe2O3/TiO2 and Y2O3/Fe2O3/TiO2
[4], ball-milling and sol–gel synthesized ZnO/TiO2 photocatalysts,
∗
Corresponding author at: Dipartimento di Chimica, Università degli Studi di
Milano, Via Golgi 19, I-20133 Milano, Italy. Tel.: +39 02 503 14237;
fax: +39 02 503 14300.
E-mail address: elena.selli@unimi.it (E. Selli).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.01.008

F. Riboni et al. / Catalysis Today 209 (2013) 28–34
29
the activity of which has been measured in both reduction and
oxidation processes [5,6], and the mesoporous PdO/TiO2 nanocom-
posites synthesized by a simple sol–gel process in the presence of
a copolymer as templating agent, which were employed in CH3OH
photooxidation [7].
In the present study W/Ti mixed oxides, obtained starting from
different precursors and synthesis routes, have been studied in
photocatalytic oxidative degradation reactions [8–10]. As WO3 is
about 15 times more acidic than TiO2, the Lewis surface acidity
of WO3–TiO2 increases with increasing the WO3 fraction up to a
plateau level of the WO3 amount in the mixed oxides [11]. Indeed,
the point of zero charge of TiO2 and WO3 are at pH 6.2 and 0.3–0.5,
respectively [12]. Moreover, WO3 can easily act as an electron
acceptor, with a consequently increased efficiency of photopro-
duced charge carriers separation [13], since W(VI) can be easily
reduced to W(V) [8].
The sol–gel WO3–TiO2 powders prepared starting from two dif-
ferent W(VI) precursors also with different W/Ti molar ratios have
been fully characterized by different surface and bulk techniques
and tested as photocatalysts in a liquid phase and in a gas phase
oxidation reaction. Formic acid (FA), chosen because it undergoes
direct photomineralization without forming any stable intermedi-
ate species, was used as a model water pollutant. The photocatalytic
abatement of acetaldehyde (AC), one of the principal odor-causing
gases in indoor air, particularly in cigarette smoke [14], which can
be envisaged as a model chemical pollutant of air, was employed
as a gas phase photocatalytic test reaction. Photoelectrochemical
measurements were also performed on photoanodes coated with
the synthesized photocatalysts, to determine the incident photon
to current conversion efficiency (IPCE) and compare it with their
photoactivity in the two test reactions.
adopted for the preparation and calcination of the pure TiO2 T700
photocatalyst.
Most of the employed chemicals were purchased from
Sigma–Aldrich and used as received. All solutions were prepared
employing ultra-pure water (18.2 MO cm), supplied by a Millipore
Direct-Q 3 water purification system.
2.2. Photoanodes preparation
The photoanodes to be employed in IPCE measurements were
prepared by depositing the synthesized photocatalysts on ITO con-
ductive glass (1.5 cm × 1.5 cm), which was preliminarily cleaned
in three consecutive steps (i.e. with distilled water, acetone and
ethanol), under ultrasound for 10 min, in a Falc, model LBS1, appa-
ratus and finally dried under N2 flow. The ITO specimens were
manually dip-coated 30 times in well dispersed 2-propanol suspen-
sions containing 1.5 g L−1 of photocatalyst, which were preliminary
sonicated for 30 min. The ITO specimens were dried under air flow
after each coating step, in order to obtain homogeneously deposited
films, and finally calcined at 450 ◦C for 4 h in air, to ensure the
adhesion of the powders on the ITO specimens. The uniformity of
all photocatalyst films deposited on ITO was checked by means of
depth-profiling measurements. Their mean thickness was 2 ␮m.
2.3. Photocatalysts characterization
XRPD measurements were performed using a Philips PW3020
powder diffractometer, operating at 40 kV and 40 mA, employing
˚
Cu K␣ radiation (ꢀ = 1.54056 A) as X-ray source. The diffractograms
were recorded by continuous scanning between 20◦ and 80◦ 2ꢁ
angles, with a 0.05◦ step. Quantitative phase analysis was made by
the Rietveld refinement method [16], using the “Quanto” software;
the average crystallite size was obtained by applying the Scherrer
equation.
2. Experimental
X-ray photoelectron spectroscopy experiments were performed
in a Leybold LHS 10/12 spectrometer equipped with a hemispher-
ical electron energy analyzer using Al K␣ radiation (ꢀ = 1486.3 eV).
Photoelectron spectra were recorded in constant analyzer energy
mode. The binding energies were referred to the C 1s peak at
284.6 eV. Energy dispersive X-ray (EDX) spectroscopy experiments,
performed using a Zeiss Sigma Scanning Electron Microscope
2.1. Photocatalysts synthesis
A pure TiO2 photocatalyst, labeled as T700, was synthesized by
a sol–gel method, starting from titanium(IV) isopropoxide (TTIP
Aldrich, purity 97%) as titanium precursor, following a procedure
similar to that already described [15]. An anhydrous ethanol solu-
tion (100 mL, purity > 99.8%) containing 10 mL of dissolved TTIP
was heated at 30 ◦C under vigorous stirring. Then 34 mL of water
were added dropwise in order to attain a Ti/H2O = 1/58 molar ratio.
After stirring and refluxing for one hour, the suspension was con-
centrated under reduced pressure at 35 ◦C. The so-obtained white
slurry was kept in oven at 70 ◦C overnight to eliminate organic com-
pounds and then calcined at 700 ◦C under a 100 mL min−1 air flow
for 4 h.
Two different WO3 precursors were used to prepare WO3–TiO2
photocatalysts: i) an inorganic salt, sodium tungstate dihydrate
(Na2WO4 × 2H2O, Carlo Erba, purity > 99.0%) to synthesize the sam-
ple labeled as TWa, and ii) an organic precursor, tungsten(VI)
hexa-ethoxide (W(OC2H5)6, Alfa-Aesar, 99.8% (metal basis), 5 wt.%
in ethanol), to prepare the TWbx% photocatalyst series, with x
referring to the W/Ti percent molar ratio. The TWa sample was pre-
pared according to the procedure followed to prepare pure TiO2
T700, with the only difference that 0.324 g of sodium tungstate
were dissolved in water (34 mL) prior to its addition to the TTIP
ethanol solution. The nominal W/Ti molar ratio of TWa was
thus 3%.
equipped with a Bruker Quantax 400 EDS detector (30 mm2
X
Flash silicon drift detector), provided information on the surface
composition of the photocatalysts, in terms of atomic percent
amounts.
The BET surface area was determined by N2 adsorption in a
Micromeritics Tristar II 3020 V1.03 apparatus. The optical prop-
erties of the powders were determined by a UV–visible reflectance
spectrophotometer (Jasco, V-650), equipped with an integrating
sphere (ISV-722), employing BaSO4 as reference not absorbing
material.
2.4. Photocurrent measurements
The photoelectrochemical characterization of both pure TiO2
and WO3–TiO2 mixed oxide photoanodes was performed in a
two electrodes newly home-built photoelectrochemical cell, with
a platinum coil as counter electrode, similar to those already
described in the literature [17]. The cell was filled with a KOH 1.0 M
solution. The TiO2 and WO3–TiO2 working electrode films were illu-
minated from the back side of the ITO support by means of a 150 W
Xe lamp (Lot-Oriel Arc Lamp), the emission of which was selected
through a Lot-Oriel Omni-␭ 300 monochromator. The incident pho-
tons to current conversion efficiency (IPCE) spectra were collected
in the 320–440 nm range by means of a digital Agilent 34410A mul-
timeter, without any applied external bias. The digital multimeter
The TWbx% photocatalysts series was obtained by adding the
required amount of tungsten(VI) hexa-ethoxide to the TTIP ethanol
solution, in order to produce WO3–TiO2 mixed oxides with nominal
W/Ti molar ratios equal to 1, 3 and 5%. The so-obtained solu-
tions were then treated following step by step the same procedure

30
F. Riboni et al. / Catalysis Today 209 (2013) 28–34
and the monochromator were simultaneously operated by a
LabView software, in order to acquire the steady short circuit pho-
tocurrent at each wavelength. Measurements were performed with
a 5 nm step. The lamp irradiance on the cell at each wavelength was
measured with a Newport (818-UV) calibrated photodetector.
The TiO2 and WO3–TiO2 photoanodes were prepared by
following the procedure described in Section 2.2. Photoelectro-
chemical measurements were performed also on a photoanode,
prepared by the same procedure, employing P25 from Degussa
(Evonik).
2.5. Photocatalytic activity tests
FA degradation runs were performed under atmospheric condi-
tions in a magnetically stirred 100 mL cylindrical quartz reactor,
using the equipment and following the procedure described
in detail elsewhere [15,18]. The photocatalyst amount within
the aqueous suspension and the FA initial concentration were
fixed at 0.1 g L−1 and 1.0 × 10−3 mol L−1, respectively. The full
emission intensity on the photoreactor of the employed irradi-
ation source (an Osram Powerstar HCI-T 150 W/NDL lamp) was
2.53 × 10−7 Einstein s−1 cm−2. This lamp mainly emits visible light
above 400 nm, but also has a small emission in the 350–400 nm
range, as evidenced in the emission spectrum reported in Ref. [18].
The residual amount of formate anions contained in samples with-
drawn from the photoreactor at different times was determined by
ion chromatography with conductivity detection, after centrifuga-
tion. The pH of the suspensions, measured by means of an Amel
model 2335 pHmeter, did not vary significantly during FA degra-
dation, always being in the 3.2–3.6 range. All photocatalytic runs
lasted 5 h and were repeated at least twice, to check their repro-
ducibility.
The photocatalytic degradation of AC in the gas phase was inves-
tigated in a flow system irradiated with a 450 W Xenon lamp
(Philips Cleo Compact), with the typical emission at ꢀ > 320 nm.
The irradiation intensity of the lamp was fixed at 1.0 mW/cm2
and checked prior to every photocatalytic run. The photocatalysts
were used in film form, prepared by pressing the powders within
a Plexiglass reactor with an optical window (6 cm2 surface area
available for irradiation), followed by pre-irradiation for 72 h with
a UVA lamp (Philips, Cleo Performance UV Type 3, 100 W, intensity
1 mW/cm2), to clean their surface from any adsorbed pollutant.
A mixture of dry air, wet air and AC (50 mL min−1, correspond-
ing to 1 ppm) was used to purge the reactor containing the film
before starting the photocatalytic tests, until a steady AC concen-
tration was attained in the gas phase outgoing from the reactor (ca.
20 min), and continued during the runs. Then a 2 h-long irradiation
began. AC abatement was monitored by sampling the gas phase
at the exit of the photoreactor and analyzing it in an on-line gas
chromatograph (Synthec Spectras GC 955) [19].
Fig. 1. XRPD patterns of (a) pure TiO2 undoped T700, (b) Na2WO4, (c) TWa, (d)
TWb1%, (e) TWb3%, (f) TWb5%. A and R indicate the anatase and rutile phases,
respectively. Na2WO4 reflections are indicated by *. The b pattern was collected
after annealing sodium tungstate at 700 ◦C, for better comparison.
occurring around 600 ◦C in pure TiO2. An increase of the TiO2
anatase phase thermal stability induced by the addition of tung-
sten, hindering both the crystal growth and the anatase to rutile
phase transition, has recently been reported [20,21]. Furthermore,
no reflections corresponding to the formation of new phases or pure
WO3 were seen in the XRPD patterns. Tungsten oxide may either
be amorphous or in concentration below the detection limit of the
XRPD technique.
W6+ can in principle be easily introduced in the titania lattice,
substituting Ti4+, the ionic radius of W6+ being 0.060 nm, that of Ti4+
0.0605 nm [22]. Indeed, as suggested by Li et al. [23], the fact that
no features in the XRPD pattern of mixed oxides is attributable to
WO3 may imply that tungsten ions could be incorporated into the
titania lattice substituting titanium ions to form W–O–Ti bonds or
could be located at interstitial sites. However, focusing on the cell
parameters reported in Table 1, which were calculated from the
XRPD data employing the Rietveld refinement, a nonmonotonous
dependence on the W percent amount can be envisaged, probably
consequent to some anisotropic distortion of the anatase lattice
induced by the different interactions among W cations and/or W
cations and O vacancies [24]. The very minor effects on the results
of XRPD analysis observed upon W addition point to a very limited
W insertion in the TiO2 lattice and to the prevailing formation of
WO3 domains.
On the contrary, concerning the TWa sample (Fig. 1(c)), in
addition to the anatase reflections mentioned above for the TWb
samples, typical peaks belonging to the Na2WO4 precursor were
also recorded (at 2ꢁ = 27.6◦, 32.5◦, [25]) as evidenced by the b and c
diffraction patterns in Fig. 1. The phase composition analysis high-
lighted the presence of ca. 2 wt.% of Na2WO4 (Table 1).
Finally, the crystallite size dimensions of pure TiO2 and
WO3–TiO2 samples calculated from XRPD data employing the
Scherrer equation (Table 1) indicate a small decrease in anatase
particle size with increasing the W amount [26].
3. Results and discussion
3.1. Crystal structure and physical properties
In line with this observation, BET analysis showed an increase
in the specific surface area (SSA) of the TiO2-based photocatalysts
with increasing the W percent amount in the TWb series (Table 1).
Also in TWa, containing 3% of W, the presence of small amounts
of W inhibited the sintering process induced by calcination at high
temperature. This leads to the conclusion that even a modest W
content is beneficial in order to obtain samples with SSA values
higher than that of pure TiO2 (Table 1), after annealing at 700 ◦C.
The XPS survey spectra of the investigated samples exhibited
the expected signals at binding energies corresponding to titanium,
oxygen and tungsten photoemissions, together with a small carbon
The XRPD patterns reported in Fig. 1 clearly evidence the role
of W in modifying the phase composition of the synthesized
photocatalysts. Pure TiO2 T700 mainly consisted of rutile phase,
as revealed by the (1 1 0) reflection at 2ꢁ = 27.4◦ in Fig. 1(a), and
also contained ca. 10 wt.% of anatase phase (signal at 2ꢁ = 25.5◦).
However, as shown by the d–f diffraction patterns in Fig. 1, all W-
containing titania photocatalysts of the TWb series were composed
of pure anatase (2ꢁ = 25.5◦, 37.9◦), suggesting that the presence of
small amounts of W, in the 1–5 mol% range, completely inhibited
the transformation of the anatase into the rutile phase, typically

F. Riboni et al. / Catalysis Today 209 (2013) 28–34
31
Table 1
Phase composition (A = anatase, R = rutile), anatase particles dimensions, dA, anatase lattice cell parameters and cell volume, all obtained from XRPD analysis, assuming the
absence of amorphous material; surface specific area, SSA, from BET analysis of the synthesized TiO2 and WO3–TiO2 photocatalysts.
Sample
Phase composition (wt.%)
dA (nm)
a (Å)
c (Å)
Cell vol. (Å3)
SSA (m2g−1
)
T700
TWa
TWb1%
TWb3%
TWb5%
10% A + 90% R
98% A + 2% Na2WO4
100% A
100% A
100% A
39
29
24
20
15
3.7827
3.7846
3.7852
3.7844
3.7890
9.5102
9.5063
9.5155
9.5027
9.5041
136.08
136.16
136.34
136.09
136.44
6.7
19.2
14.8
28.7
52.9
Fig. 2. XPS spectra in the (A) Ti 2p, (B) O 1s and (C) W 4f binding energy ranges of (a) T700, (b) TWa, (c) TWb1%, (d) TWb3% and (e) TWb5%.
contribution arising from contaminants and the Na 1s signal
(6.2 at.%) only in the spectrum of TWa, arising from the tungsten
precursor, Na2WO4. The W/Ti atomic ratios, 0.13, 0.08, 0.14 and
0.18 for TWa, TWb1%, TWb3% and TWb5%, respectively, though
much higher than the nominal molar amounts, confirmed that
TWa and TWb3% contained similar tungsten amounts, both higher
than that contained in TWb1% and lower than that contained in
TWb5%.
The XPS spectra in the Ti 2p1/2 and Ti 2p3/2 binding energy region
can be compared in Fig. 2A. The Ti 2p doublet signals showed very
similar profiles, with the two components picking at 465.1 eV and
459.5 eV. This points to the formation of titanium dioxide with the
correct TiO2 stoichiometry [27], also in the case of the mixed oxides.
The XPS signal of pure TiO2 in the O 1s binding energy region
(Fig. 2B, trace a) was composed of a main band peaking at 530.6 eV,
typical of oxygen bound to Ti, and of a shoulder at higher binding
energies, mainly originated from oxygen in surface hydroxyl groups
[27]. On the contrary, in the XPS spectra of the W–Ti mixed oxides,
the main O 1s band also peaked at 530.6 eV (see Fig. 2B(b–e)),
while the contribution arising from hydroxyl groups was strongly
reduced, as a consequence of the tungsten-induced reduction of
surface hydration.
The EDX spectrum of TWb3% shown in Fig. 3 displays the signals
belonging to W, together with those due to Ti and O, and that of C
arising from contaminants. This confirms that tungsten is present
on the surface regions of the oxide, although a minor incorpora-
tion into the TiO2 lattice cannot be excluded [29], and suggests the
formation of WO3 microdomains that create a bi-component com-
posite structure in WO3–TiO2 rather than producing TiO2 doping.
The atomic percent amounts of the species detected by EDX in
TWb3% (see Fig. 3) indicate a W/Ti atomic ratio equal to 2.7%, which
confirms the W nominal amount assessed in the synthesis proce-
dure. Furthermore, the O/Ti atomic ratio is below 2, suggesting the
possible presence of surface oxygen vacancies in our mixed oxide
sample calcined at 700 ◦C.
The absorption spectra of the investigated photocatalyst mate-
rials are all depicted in Fig. 4. Pure TiO2 T700, mainly composed of
Finally, Fig. 2C shows the XPS signals collected in the binding
energy range between 42 eV and 32 eV. A couple of bands located
at 37.7 eV and 36.0 eV can be recognized in the spectra of W–Ti
samples (see Fig. 2C(b–e)), corresponding to the W 4f5/2 and W
4f7/2 signals, respectively [28], while the XPS spectrum of pure TiO2
T700 (not shown) is perfectly flat in this region. This is compati-
ble with the formation of small clusters of tungsten oxide in the
mixed oxides. A shift of about 0.4 eV toward higher binding ener-
gies was detected in the spectrum of TWa, Fig. 2C(b), probably due
2-
to the presence of the WO4 species revealed by XRPD analysis
(Fig. 1(c)).
Fig. 3. EDX spectrum of the TWb3% sample.

32
F. Riboni et al. / Catalysis Today 209 (2013) 28–34
Fig. 5. IPCE spectra of synthesized TiO2 and WO3–TiO2 samples measured without
any external bias.
Fig. 4. UV–vis absorption spectra of pure TiO2 and WO3–TiO2 samples. Inset:
(Ahꢂ)1/2 vs. hꢂ plots.
calculated from photoelectrochemical measurements according to
Eq. (2):
ꢀ
ꢁ
1240 × Isc
(Plight × ꢀ)
rutile (Table 1), exhibited an absorption threshold red-shifted with
respect to those of the other samples, consisting of anatase phase.
On the other hand, all WO3–TiO2 mixed oxides showed almost the
same absorption properties, only TWa exhibiting a somewhat blue-
shifted absorption edge, mirroring the behavior of pure anatase
TiO2, in line with a lower or null TiO2–WO3 mixing in this case.
As both TiO2 and WO3 are indirect band gap semiconductors, the
band gap energies of the here investigated photocatalytic materials
were estimated from Tauc plots:
IPCE (%) =
× 100
(2)
where ꢀ is the wavelength of the incident light, Plight is the
illumination power density at the specific wavelength and Isc
is the measured short circuit photocurrent density at the same
wavelength [17]. These measurements were carried out with the
electrochemical cell filled with a KOH 1.0 M solution. Both TiO2
and WO3 can be safely used as photoanodes under such conditions,
without any photocorrosion risk, since both semiconductors have
a VB maximum lower than the (O2/H2O) potential [34].
(Ahv)1/2 = Bi(hv − Eg)
(1)
Two types of IPCE response were observed. In fact, all WO3–TiO2
mixed oxide photoanodes presented an IPCE onset (IPCE > 1%) at
about 380 nm, consistent with the absorption edge of the TiO2
anatase phase, mirroring the IPCE onset of the photoanode pre-
pared by deposition of commercial P25 TiO2. This feature leads
to the conclusion that enhanced charge carriers separation results
from the presence of WO3, the absorption of light and the conse-
quent photoproduction of charge carriers being related to the TiO2
features only. Furthermore, at wavelengths shorter than 380 nm,
TWb1%, TWb3% and TWb5% presented much higher IPCE values
compared to TWa, highlighting the influence of the employed tung-
sten precursor on the photoactivity of the material, the relatively
lower IPCE of TWa most probably to be related to the presence of
Na2WO4 domains rather than WO3 domains.
Besides this effect, the height of IPCE curves strongly depended
on the amount of WO3 mixed in the TiO2 matrix, with a maxi-
mum for a W/Ti nominal molar ratio of 3% (TWb3%), in line with
previous reports [20,35]. The maximum IPCE value (16%) attained
with TWb3% was higher than that measured with commercial P25
powder under identical conditions (Fig. 5).
A standing for absorbance, Eg being the band gap energy, Bi the
absorption constant for indirect transitions and hꢂ the photon
energy in electronvolt. Eg can be calculated by extrapolating the
linear portion of the plots shown in the inset of Fig. 4 down to the
abscissa axis (to zero absorbance) [30]. The resulting Eg values are
listed in Table 2. The value of 2.80 eV obtained for T700, mainly
consisting of rutile phase, appears surprisingly low, the usually
reported band gap value of rutile being 3.0 eV [31]. In W-containing
samples, as a consequence of the stabilization of the anatase phase,
the band gap energy values increased up to about 3 eV, however
being always lower than that reported for the pure anatase phase
(Eg = 3.2 eV [31]). The presence of surface oxygen vacancies in our
pure TiO2 and mixed oxide photocatalysts calcined at 700 ◦C is com-
patible with band gap values smaller than those of the pure rutile
and anatase polymorphs, as recently demonstrated in the literature
[32,33].
3.2. Photoelectrochemical characterization
On the other hand, pure TiO2 T700 exhibited a photocurrent
onset (IPCE ≥ 1%) at longer wavelength, i.e. at about 410 nm, clearly
due to the presence of the rutile phase, in full agreement with its
absorption spectrum (Fig. 4). Its maximum IPCE value, however,
was much lower than that attained with TWb3%.
The incident photons to current conversion efficiency (IPCE)
expresses the number of collected electrons per number of inci-
dent photons at a given irradiation wavelength. Fig. 5 reports the
IPCE vs. wavelength curves of the TiO2 and WO3–TiO2 samples,
Table 2
3.3. Photocatalytic activity tests
Band gap energy (Eg) of the TiO2 and WO3–TiO2 photocatalysts, calculated applying
Eq. (1) to the absorption spectra, and results of the photoactivity tests, in terms of
zero order rate constants of formic acid (FA) degradation, kFA, with R2 values, and of
percent amount of acetaldehyde (AC) degraded after 2 h.
FA photocatalytic degradation occurred according to a zero
order rate law, i.e. FA concentration decreased at constant rate dur-
ing irradiation (see Fig. 6), as in previous studies [15,18,36]. Thus,
the photocatalytic activity of both pure TiO2 and WO3–TiO2 mixed
oxides can be compared in terms of the zero order rate constants
kFA collected in Table 2. Apart from the case of TWb1%, the photo-
catalytic activity of the TiO2-based photocatalysts appeared to be
enhanced by W addition. The highest photoactivity was achieved
Sample
Eg (eV)
kFA × 108 (M s−1
)
R2
AC degradation (%)
T700
TWa
TWb1%
TWb3%
TWb5%
2.80
3.02
2.95
2.92
2.93
1.87 0.03
3.56 0.15
0.58 0.02
3.98 0.17
2.15 0.06
0.9984
0.9910
0.9954
0.9870
0.9958
18
14
−
31
−

F. Riboni et al. / Catalysis Today 209 (2013) 28–34
33
3.4. Origin of the W-induced increase in TiO2 photoactivity
The better performance of the Ti–W oxides can be partly
ascribed to the structure-preserving effect of tungsten, ensuring a
full anatase composition of the material up to calcination at 700 ◦C
and a very smooth modification of the anatase cell parameters
(Table 1). Highly crystalline materials with high intrinsic pho-
toactivity would thus be obtained, due to effects similar to those
recently observed upon addition of low amounts of p-block ele-
ments to TiO2 [15,38]. Furthermore, the presence of tungsten also
led to a decrease in anatase particles dimensions, paralleled by an
increase in the specific surface area of the photocatalysts (Table 1),
which notoriously plays a remarkable positive role in enhancing
the rate of all heterogeneous catalytic reactions.
Furthermore, W addition creates a bi-component composite
structure in TiO2. We attribute the observed higher photoactivity
of WO3–TiO2 mixed oxides mainly to the vectorial electron transfer
from the photoexcited TiO2 particles to the WO3 regions [39]. Since
the band gap energy of TiO2 and WO3 are 3.2 eV and 2.8 eV, respec-
tively [40], with a lower-lying VB and CB of WO3 with respect to the
VB and CB of TiO2, the usually accepted photoexcitation mechanism
[41,42] implies that photopromoted electrons are transferred from
the CB of TiO2 to the CB of coupled WO3, while photogenerated
holes migrate in the opposite direction, i.e. from the lower-lying
VB of WO3 to the VB of TiO2, with a consequent increase of charge
separation efficiency [41]. This photocatalytic activity increasing
mechanism is predominant in mixed oxides samples containing
WO3 domains, obtained from organic W precursors.
Fig. 6. Photocatalytic degradation of FA in the presence of different photocatalysts,
confirming the zero order rate law.
with the TWa and TWb3% photocatalysts, both characterized by a
3% nominal percent molar W/Ti ratio, in line with previous reports
[20,35].
The photocatalytic activity trend observed with the TWb photo-
catalysts series can be discussed in comparison with the IPCE curves
obtained with the same samples (Fig. 5). An increase in tungsten
content from 1 up to 3% produced an enhancement of the photon-
to-current conversion, whereas a higher tungsten content (5%) led
to a photoactivity decrease, most probably due to the increased
extent of defects in the oxide structure, acting as recombination
centers of photogenerated charge carriers. Both lower IPCE val-
ues and a lower photocatalytic reaction rate resulted under such
conditions.
The highest photocatalytic performance attained in the case of
3% molar tungsten amount could not be related to more extended
light absorption, as the calculated Eg values in the TWb series
(Table 2) are almost independent of the W amount. On the other
hand, according to the results of calculations based on XRPD anal-
ysis, the anatase cell in TWb3% was the least distorted in the TWb
series (Table 1), i.e. the a lattice cell parameter and the cell vol-
ume had the smallest values, suggesting a low degree of structural
defects with respect to pure anatase [37]. Also the TWa sample had
a cell structure very similar to that of pure anatase (Table 1).
To better understand the role of W in enhancing the photocat-
alytic activity of TiO2, the best performing photocatalysts were also
tested in gas phase AC photodegradation and the results were com-
pared to those obtained with pure TiO2 T700, in terms of percent
amount of substrate degraded after 2 h (Table 2). AC photodegrada-
tion is known to proceed via a radical chain mechanism on the TiO2
surface, leading to the formation of CO2, preceded by the generation
of relatively large amounts of acetic acid [14].
4. Conclusions
The sol–gel synthesis of mixed WO3–TiO2 photocatalysts start-
ing from organic precursors of the two metals led to materials
with an intimate contact between the two oxides, exhibiting higher
photoactivity paralleled by increased IPCE values, with respect to
pure TiO2 prepared by the same route. The decrease of particles
size observed with increasing the W amount, the parallel benefi-
cial increase of surface area and the persistence of highly active
crystalline anatase phase, even after calcination at 700 ◦C, con-
tribute to the high photocatalytic performance of the materials. A
better photoproduced charge separation occurs, if separate WO3
microdomains form in the TiO2 structure, as is the case of materi-
als obtained from organic precursors. By contrast, a less performing
photocatalyst was obtained when sodium tungstate was employed
as W precursor in the sol–gel synthesis, due to its persistence in
XRD-detectable separate domains within the TiO2 structure. The
optimal tungsten content, corresponding to a 3% W/Ti molar ratio,
was attained with samples characterized also by minimal distortion
of the lattice cell parameters with respect to pure anatase TiO2.
Acknowledgements
The best photocatalytic activity in the gas-phase reaction was
once again achieved with TWb3%, surprisingly even higher than
that of P25 TiO2, ensuring only 26% AC degradation after 2 h. Of
course, the higher specific surface area of W-containing TiO2 sam-
ples with respect to T700 (Table 2) played a fundamental role in
this case, substrate adsorption on the photocatalyst surface being a
crucial step in any gas phase photocatalytic reaction. The different
results obtained with TWa and TWb3% may once again be related
to the presence of tungstate domains in TWa, detected by XRPD
analysis.
This work received financial support from the Cariplo Foun-
dation through the project Development of second generation
photocatalysts for energy and environment and from the 2009PASLSN
MIUR-PRIN project. The collaboration of Dr. Cristina Lenardi in XPS
and EDX analyses and of Dr. Jonathan Bloh in AC degradation mea-
surements is gratefully acknowledged.
References
Also the fact that pure TiO2 T700, though characterized by lower
SSA (Table 1), was more active than TWa in AC photodegradation
(Table 2), should be related to the presence of tungstate domains
on the TWa surface, which might result in suppressed AC adsorp-
tion on the photocatalyst surface due to the repulsion between two
electron-rich species (i.e. CH3CHO and WO42-).
[1] D.W. Bahnemann, M. Hilgendorff, R. Memming, Journal of Physical Chemistry
B 101 (1997) 4265–4275.
[2] A. Fujishima, X. Zhang, D.A. Tryk, Surface Science Reports 63 (2008) 515–582.
[3] S. Linic, P. Cristopher, D.B. Ingram, Nature Materials 10 (2011) 911–921.
[4] A.A. Ismail, Applied Catalysis B: Environmental 58 (2005) 115–121.
[5] C. Shifu, Z. Wei, L. Wei, Z. Sujuan, Applied Surface Science 255 (2008)
2478–2484.

34
F. Riboni et al. / Catalysis Today 209 (2013) 28–34
[6] T. Ivanova, A. Harizanova, T. Koutzarova, B. Vertruyen, Journal of Non-
Crystalline Solids 357 (2011) 2840–2845.
[26] B. Tryba, M. Piszcz, A.W. Morawski, International Journal of Photoenergy 2009
(2009) 1–7.
[7] A.A. Ismail, Applied Catalysis B: Environmental 117–118 (2012) 67–72.
[8] V. Keller, P. Bernhardt, F. Garin, Journal of Catalysis 215 (2003) 129–138.
[9] C. Yumin, Rare Metals 25 (2006) 649–653.
[10] C.F. Lin, C.H. Wu, Z.N. Onn, Journal of Hazardous Materials 154 (2008)
1033–1039.
[11] Y.T. Kwon, K.Y. Song, W.I. Lee, G.J. Choi, Y.R. Doz, Journal of Catalysis 191 (2000)
192–199.
[12] G.A. Parks, Chemical Reviews 65 (1965) 177–198.
[13] B. Gao, Y. Ma, Y. Cao, W. Yang, J. Yao, Journal of Physical Chemistry B 110 (2006)
14391–14397.7.
[27] T. Caruso, C. Lenardi, R.G. Agostino, M. Amati, G. Bongiorno, T. Mazza, A. Poli-
ˇ
cicchio, V. Formoso, E. Maccallini, E. Colavita, G. Chiarello, P. Finetti, F. Sutara,
T. Skála, P. Piseri, K.C. Prince, P. Milani, Journal of Chemical Physics 128 (2008)
0947041–947048.
[28] H. Ishihara, G.K. Kannarpady, K.R. Khedir, J. Woo, S. Trigwellb, A.S. Biris, Physical
Chemistry Chemical Physics 13 (2011) 19553–19560.
[29] J. Choi, H. Park, M.R. Hoffmann, Journal of Physical Chemistry C 114 (2010)
783–792.
[30] D. Dasgupta, F. Demichelis, C.F. Pirri, A. Tagliaferro, Physical Review B 43 (1991)
2131–2135.
[14] Y. Ohko, D.A. Tryk, K. Hashimoto, A. Fujishima, Journal of Physical Chemistry B
102 (1998) 2699–2704.
[31] N. Serpone, D. Lawless, R. Khairutdinov, Journal of Physical Chemistry 99 (1995)
16646–16654.
[15] M.V. Dozzi, S. Livraghi, E. Giamello, E. Selli, Photochemical & Photobiological
Sciences 10 (2011) 343–349.
[16] H.M. Rietveld, Journal of Applied Crystallography 2 (1969) 65–71.
[17] O. Enea, J. Moser, M. Grätzel, Journal of Electroanalytical Chemistry 259 (1989)
59–65.
[32] X. Chen, L. Liu, P.Y. Yu, S.S. Mao, Science 331 (2011) 746–750.
[33] X. Jiang, Y. Zhang, J. Jiang, Y. Rong, Y. Wang, Y. Wu, C. Pan, Journal of Physical
Chemistry C 116 (2012) 22619–22624.
[34] S. Chen, L.W. Wang, Chemistry of Materials 24 (2012) 3659–3666.
[35] X.Z. Li, F.B. Li, C.L. Yang, W.K. Ge, Journal of Photochemistry and Photobiology
A 141 (2001) 209–217.
[18] M.V. Dozzi, L. Prati, P. Canton, E. Selli, Physical Chemistry Chemical Physics 11
(2009) 7171–7180.
[36] M. Mrowetz, E. Selli, Journal of Photochemistry and Photobiology A 180 (2006)
[19] A.A. Ismail, D.W. Bahnemann, J. Rathousky, V. Yarovyid, M. Wark, Journal of
Materials Chemistry 21 (2011) 7802–7810.
15–22.
ˇ
[37] V. Stengl, J. Velická, M. Marˇíková, T.M. Grygar, Applied Materials & Interfaces 3
[20] K.K. Akurati, A. Vital, J.P. Dellemann, K. Michalowa, T. Graule, D. Ferri, A. Baiker,
Applied Catalysis B: Environmental 79 (2008) 53–62.
[21] N. Couselo, F.S. García Einschlag, R.J. Candal, M. Jobbágy, Journal of Physical
Chemistry C 112 (2008) 1094–1100.
[22] R.D. Shannon, Acta Crystallographica Section A 32 (1976) 751–767.
[23] J. Li, J. Xu, W.L. Dai, H. Li, K. Fanppl, Applied Catalysis B: Environmental 82 (2008)
233–243.
(2011) 4014–4023.
[38] M.V. Dozzi, B. Ohtani, E. Selli, Physical Chemistry Chemical Physics 13 (2011)
18217–18227.
[39] N.R. de Tacconi, C.R. Chenthamarakshan, K. Rajeshwar, T. Pauporté, D. Lincot,
Electrochemistry Communications 5 (2003) 220–224.
[40] A. Rampaul, I.P. Parkin, S.A. O’Neill, J. De Souza, A. Mills, N. Elliott, Polyhedron
22 (2003) 35–44.
[24] H. Kamisaka, T. Suenaga, H. Nakamura, K. Yamashita, Journal of Physical Chem-
istry C 114 (2010) 12777–12783.
[41] C. Shifu, C. Lei, G. Shen, C. Gengyu, Powder Technology 160 (2005)
198–202.
[25] H. Liu, D. Yang, R. Gao, L. Chen, S. Zhang, X. Wang, Catalysis Communications 9
(2008) 1302–1306.
[42] J.A. Rengifo-Herrera, M.N. Blanco, L.R. Pizzio, Applied Catalysis B: Environmen-
tal 110 (2011) 126–132.