Gas phase photocatalytic removal of toluene effluents on sulfated titania

Article abstract of DOI:10.1016/j.jcat.2005.08.017

Photocatalytic removal of toluene in the gas phase was carried out over UV-illuminated sulfated titania materials in a cylinder-like continuous reactor. A series of SO42 - TiO₂ samples was obtained from the addition of H₂SO₄ on an amorphous titanium hydroxide gel synthesized according to a classical sol-gel procedure. The wide variety of materials led to varying photocatalytic behaviors depending strongly on the experimental synthesis parameters, having a determinant influence on the surface specific area, the crystallinity of the material, the crystallographic nature of TiO ₂, and the sulfate surface content. Optimization of the experimental parameters, such as the molarity of the sulfation solution, varying in the range 0.25-5 M leading to surface sulfate coverage of 2.5-14 wt%, and the calcination temperature ranging from 400 to 800 °C, promoted enhanced photocatalytic performance toward toluene removal as compared with commercially available P25 TiO₂ and sulfate-free sol-gel TiO₂. The most efficient photocatalyst was obtained for a near-monolayer sulfate coverage corresponding to the presence of both TiO₂ and well-dispersed SO42- with optimized contact between SO42- and TiO₂ domains. Furthermore, a positive role of sulfates is attributed both to an electron transfer from titanium to sulfates, leading to a positive charge trap effect, and to better desorption of electron-rich sp2-bound carbon aromatic poisons, thus limiting deactivation.

Full text of DOI:10.1016/j.jcat.2005.08.017

Journal of Catalysis 235 (2005) 318–326
www.elsevier.com/locate/jcat
Gas phase photocatalytic removal of toluene effluents on sulfated titania
∗
Elodie Barraud, Florence Bosc, David Edwards, Nicolas Keller, Valérie Keller
Laboratoire des Matériaux, Surfaces et Procédés pour la Catalyse, UMR 7515 CNRS and ELCASS (European Laboratory for Catalysis and Surface Science),
Louis Pasteur University, 25 rue Becquerel, BP 08, 67087 Strasbourg cedex 2, France
Received 18 April 2005; revised 18 August 2005; accepted 19 August 2005
Available online 27 September 2005
Abstract
Photocatalytic removal of toluene in the gas phase was carried out over UV-illuminated sulfated titania materials in a cylinder-like continuous
2−
reactor. A series of SO
–TiO samples was obtained from the addition of H SO on an amorphous titanium hydroxide gel synthesized accord-
4
2
2
4
ing to a classical sol–gel procedure. The wide variety of materials led to varying photocatalytic behaviors depending strongly on the experimental
synthesis parameters, having a determinant influence on the surface specific area, the crystallinity of the material, the crystallographic nature of
TiO , and the sulfate surface content. Optimization of the experimental parameters, such as the molarity of the sulfation solution, varying in the
2
◦
range 0.25–5 M leading to surface sulfate coverage of 2.5–14 wt%, and the calcination temperature ranging from 400 to 800 C, promoted en-
hanced photocatalytic performance toward toluene removal as compared with commercially available P25 TiO and sulfate-free sol–gel TiO . The
2
2
most efficient photocatalyst was obtained for a near-monolayer sulfate coverage corresponding to the presence of both TiO and well-dispersed
2
2−
2−
SO
with optimized contact between SO
and TiO domains. Furthermore, a positive role of sulfates is attributed both to an electron transfer
from titanium to sulfates, leading to a positive charge trap effect, and to better desorption of electron-rich sp -bound carbon aromatic poisons,
4
4
2
2
thus limiting deactivation.

2005 Elsevier Inc. All rights reserved.
Keywords: Titania; Sulfates; Photocatalysis; Toluene; Electron transfer; Photogenerated charge recombination
1
. Introduction
today, having high photocatalytic efficiency, stability toward
photocorrosion and chemicals, insolubility in water, low tox-
icity, and low cost. Its band gap energy of 3.2 eV leads to
photoexcitation requiring wavelengths less than ca. 385 nm,
corresponding to a near-UV illumination. Absorption of UV
light excites electrons from the valence to the conduction band,
creating electron–hole pairs with a high oxidative potential of
3 V associated with the photogenerated holes, which initiate re-
dox reactions with adsorbed surface species [6].
Nonetheless, TiO2 does not have sufficiently high photocat-
alytic activity to allow its use as such for industrial purposes. Its
behavior is limited mainly by the recombination of photogen-
erated charges [7] and by adsorption–desorption problems of
both reactants and reaction products [8]. Developing new, high-
efficiency photocatalytic materials based on the TiO2 semicon-
ductor is thus of immense importance for sustainable develop-
ment.
The emission of volatile organic compounds by industrial
processes in the United States is regulated by the Clean Air
Act of 1990, which has created a strong incentive for research
in this area in both the industrial and academic communi-
ties involved in innovative sustainable environmental research.
Among the advanced oxidation processes developed to meet the
ever-stricter antipollution legislation required by the pressure
for environmental protection, photocatalytic oxidation shows
much promise for the purification of contaminated wastewater
containing organic pollutants and the removal of air contami-
nants [1–5]. Photocatalysis is advantageous because the energy
required is supplied by the direct absorption of light at room
temperature, which requires the use of semiconductor materials
with adequate band gaps as photocatalysts. Titanium dioxide
(TiO2) is the most attractive and efficient semiconductor used
Heterogeneous photocatalysis has been used for the toluene
degradation in either the liquid [9–11] or vapor [12–19] phase.
The photocatalytic systems generally lead to moderate conver-
*
Corresponding author. Fax: +33 3 90 24 27 61.
E-mail address: vkeller@chimie.u-strasbg.fr (V. Keller).
0021-9517/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.08.017

E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
319
sions of gaseous toluene with apparent deactivation on stream.
Ollis et al. [17] reported that co-feeding of toluene with more
reactive chlorinated olefins strongly increased the decomposi-
tion rate, taking inspiration from the precursor work of Bernan
et al. [20], which added trichloroethylene to significantly in-
crease the removal of nonaromatic organics. Among the alter-
native catalytic systems designed to improve the removal effi-
ciency over that of pure TiO2, a recent approach developed by
Ollis et al. to avoid the hazardous use of chlorinated olefins in-
volves pretreating the reference TiO2 catalyst with hydrochloric
acid to introduce chloride at the TiO2 surface, leading to highly
reactive and oxidizing chlorine radicals. Those radicals act in
A more systematic control of the sulfation parameters is re-
ported and demonstrates the positive role played by the sulfate
species.
2. Experimental
2
.1. Catalyst preparation
The sulfated titania catalysts, SO42 –TiO2, were prepared
−
according to the two-step method used by Arata for sulfated
oxide catalysts [21,22]. An amorphous Ti(OH)4 hydroxide was
prepared by a sol–gel procedure in which the Ti[OCH(CH3)2]4
titanium isopropoxide precursor (TIP, Fluka, purum) was com-
bined with water and ethanol at a basic pH (maintained at
around 9 by dropwise addition of ammonia). After drying
·
the same way as OH radicals and thus react with adsorbed aro-
matics [19]. However, the positive influence of this treatment
reported by Ollis et al. is effective only for low aromatic con-
◦
3
overnight at room temperature and further at 110 C, the amor-
centrations (ca. 14 ppm, corresponding to 50 mg/m in flowing
phous Ti(OH)4 hydroxide was sulfated under vigorous stir-
ring by an aqueous solution of sulfuric acid (VWR-Prolabo,
air) and during the early stage of the photocatalytic reaction,
with the depletion of chloride radicals contributing to catalyst
deactivation [19].
>
96.9%), using 1.5 mL of sulfation solution for 1 g of hydrox-
ide solid. The molarity of the aqueous sulfation solution was set
Sparked by the growing worldwide interest in solid acid
catalysis, mainly sulfated oxide materials and especially sul-
fated zirconia [21,22], sulfated titania materials recently have
been used as superacid solid catalysts for several organic re-
actions [23,24]. However, in the photocatalysis field, several
authors have emphasized that the role of the sulfate species
in photocatalytic efficiency is not clear and remains controver-
sial [25]. Indeed, Fu et al. [26] attributed the improved perfor-
mance of sulfated titania for CH3Br degradation to a greater
surface area and a larger fraction of the UV photoactive anatase
phase. In parallel to that finding, Muggli et al. [27] recently
reported that the use of sulfated TiO2 was nonbeneficial for
the gas phase oxidation of toluene at room temperature. These
authors suggested that the strong acid sites of sulfated titania
play an important role by increasing the adsorption strengths
and therefore the surface coverage by the organic pollutants,
but also that the sulfates are detrimental to photocatalytic ef-
ficiency at room temperature and have only a positive role at
at 0.25, 0.5, 1, and 5 M. After evaporation of excess solution at
◦
1
00 C, the materials were further calcined in air at tempera-
◦
tures ranging from 400 to 800 C for 5 h.
The performance of the sulfated titania photocatalysts as
a function of the synthesis parameters is compared to that of
those obtained on pure TiO2 synthesized by the corresponding
◦
sulfate-free sol–gel procedure with calcination at 500 C for 5 h
and the TiO2 P25 reference from Degussa.
2
.2. Characterization techniques
Structural characterization was done by powder X-ray dif-
fraction (XRD) measurements, carried out with a Siemens dif-
fractometer (model D-5000), using a Cu-Kα radiation source
in a θ/2θ mode, with a long duration scan (10 s) and a small
◦
step scan (0.02 2θ). The mean TiO2 particle size (i.e., the av-
erage size of the coherent diffracting domains) was determined
using Scherrer’s equation with the normal assumption of spher-
ical crystallites
◦
temperatures above 100 C. In addition, they showed that the
photocatalytic behavior of sulfated titania was strongly depen-
dent on the nature of the molecules to transform. Recently,
Colon et al. [25] ruled out the role of any surface acidity due
to the presence of sulfate groups. These authors also claimed
that the sulfate species have no active role during the photo-
catalytic process, and that sulfation can only be considered a
positive photocatalyst pretreatment that affects the TiO2 struc-
ture by delaying its crystallization and consequently leads to
stabilization of the anatase phase at higher temperatures with
a rather high surface area. They reported that the photocat-
alytic efficiency improved with sulfation pretreatment with an
Dc = Kλ/bhkl cosθ,
where Dc is the average particle size (nm), K is the constant of
diffraction taken to be unity, λ is the X-ray wavelength (nm),
and bhkl is the peak width at half-maximum for the hkl peak,
corrected from instrument broadening and broadening due to
the Kα2 radiation. The content of the rutile and anatase phases
was determined from the peak intensities of the most intense
diffraction peaks of both the rutile and anatase phases, that is,
◦
◦
the (110) peak at 27.6 and the (101) peak at 25.4 respectively,
according to JCPDS files 21-1276 and 21-1275.
◦
increasing temperature of calcination up to 700 C, at which
The surface area measurements were performed on a Coul-
point the sulfate species were removed from the material. They
claimed that calcining at higher temperature eliminated much
of the recombination process occurring in the crystal defects of
the TiO2 structure.
The present article reports on the use of sulfate-promoted
titania for improving the efficiency of titania in room tempera-
ture photocatalytic degradation of gas phase toluene on stream.
ter SA-3100 porosimeter using N as an adsorbent. Before
each measurement, the sample was outgassed at 300 C for
2
◦
3 h. Surface areas were calculated from the nitrogen adsorption
isotherms using the BET method (SBET). The micropore sur-
face areas were calculated using the t-plot method [28]. A de-
tailed study of the method, concerning the correctness of the
different parameters used, has published elsewhere [29].

320
E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
Thermogravimetric analysis (TGA) was performed using a
around 380 nm, located inside the inner tube of the reactor.
The reaction products were analyzed on-line every 2 min by
a thermal conductivity detector on a micro-gas chromatograph
(HP microCG M200H), allowing detection and quantification
SETARAM thermoanalyzer. Each sample was placed in a plat-
◦
inum crucible and heated from room temperature to 900 C at
◦
a rate of 5 C/min, using a 20% (v/v) O2/N2 mixture or 100%
Ar at a flow rate of 100 mL/min.
of toluene, water, CO , and organic byproducts on OV1 and
2
X-Ray photoelectron spectroscopy (XPS) surface character-
ization was performed using a ThermoVG Scientific apparatus
equipped with Al-Kα (1487) and Mg-Kα (1254 eV) sources
Poraplot Q columns. The photocatalytic experiments were per-
formed for 150-min durations.
(
pass energy of 20 eV). All of the spectra were decomposed
3
3
3
. Results and discussion
assuming several contributions, each having a Doniach–Sunjic
shape [30] and a Shirley background subtraction [31]. Ar
bombardment was performed for 3 min using a source work-
ing at 2 kV and 18 mA of ionic current. The sulfur-to-titanium
+
.1. Sulfated titania characterizations
.1.1. Structural characterizations and B.E.T determinations
Table 1 summarizes the influence of the calcination temper-
(S/Ti) surface atomic ratios were calculated for the sulfated ti-
tania photocatalysts using the sensitivity factors, as determined
by Scofield [32]. The subtraction of the energy shift due to
electrostatic charging was determined using the contamination
carbon C 1s band at 284.6 eV as a reference.
ature of the sulfated amorphous Ti(OH)4 hydroxide gel on the
specific surface area, the anatase-to-rutile molar ratio, and the
mean particle size, using an aqueous sulfation solution of 0.5 M
molarity. The sulfated titania obtained was 0.5 M SO4 –TiO2.
Characteristics of commercially available TiO2 (P25; Degussa)
are also given for comparison. A maximum BET specific sur-
2
−
2
.3. Experimental device and procedure
◦
face area was obtained for a calcination temperature of 500 C
The photocatalytic reaction was carried out in a 300-mm
2
(i.e., 116 m /g), associated with an anatase mean particle size
long cylindrical concentric tubular Pyrex reactor composed of
two coaxial tubes 4 mm apart, between which the reactant mix-
ture passes through. Detailed descriptions of the photocatalytic
reactor and device can be found elsewhere [33,34]. Photocat-
alytic material (440 mg) was evenly coated on the internal side
of the 35-mm diameter external tube by evaporating a catalyst-
containing aqueous slurry to dryness. The catalyst-coated re-
of 14 nm. Higher calcination temperatures led to a drastic de-
crease in surface area, down to 8 m /g, with a strong increase
in particle size, corresponding to material sintering. A low
calcination temperature (e.g., 400 C) resulted in a lower sur-
2
◦
2
face area of 49 m /g with a mean particle size of 15 nm,
which could be attributed to the low crystallinity of the sulfated
material, which was proved to have a different pore network
structure, due to a delay in crystallization compared with non-
sulfated materials [22]. Indeed, sulfation of TiO2 shifted the
anatase → rutile phase transition toward higher temperatures,
◦
actor was then dried at 110 C for 1 h in air. Toluene (Carlo
Erba, >99.5%) and water were fed at ambient temperature and
atmospheric pressure by bubbling air through two saturators,
then mixed with additional air (using Tylan MFC 260 mass-
flow controllers) to obtain the required toluene-water-air ratios
with a constant total air flow of 200 cm /min. The toluene con-
tent was set at 110 ppm, corresponding to 400 mg of toluene
per m of flowing air. The relative humidity was set at 30%,
with 100% of relative humidity defined as the saturated vapor
pressure of water at 25 C, corresponding to about 24 Torr, that
is, about 3% relative to the total atmospheric pressure. Before
the photocatalytic reaction, the catalyst was first exposed to the
polluted air stream with no illumination until dark-adsorption
equilibrium was reached. Afterward, the UV illumination was
switched on. Illumination was provided by a commercially
available 8-W black light tube with a spectral peak centered
◦
as rutile phase appeared after calcination at 600 C on simi-
3
◦
lar sulfate-free sol–gel TiO2 [35], compared with 800 C in our
case for SO42 –TiO2, in agreement with the results of Colon et
−
3
al. [25].
◦
Keeping the calcination temperature at 500 C, decreasing
◦
the molarity of the H2SO4 sulfation solution to 0.25 M resulted
in a 71 m /g surface area of sulfated titania, whereas increasing
the molarity to 1 and 5 M produced sulfated titania surface ar-
eas of 60 and 45 m /g, respectively. No significant changes in
the structural characterizations was reported; only the anatase
phase was observed with an unchanged anatase mean particle
size (i.e., between 12 and 15 nm).
2
2
Table 1
Influence of the calcination temperature of the amorphous Ti(OH) hydroxide gel after the sulfation procedure starting with a sulfation solution of 0.5 M, on the
4
BET specific surface area, the anatase-to-rutile molar ratio and the mean particle size. Comparison with commercial P25 Degussa and the sol–gel TiO obtained by
2
the corresponding sulfate-free sol–gel procedure
◦
Calcination temperature ( C)
Commercial
P25 Degussa
Sol–gel
TiO
2
400
500
600
700
800
2
BET specific surface area (m /g)
[
49 [0]
116 [0]
15 [0]
18 [0]
8 [0]
46 [0]
60 [0]
2
microporous content in m /g
derived from the t-plot method]
Anatase fraction (%)
Mean particle size (nm)
100
15
100
14
100
24
100
38
40
46
80
32
100
13

E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
321
2−
Fig. 1. Thermal gravimetry analysis of SO
–TiO samples under air, as a
4
2
Fig. 2. Total and surface sulfate loadings as a function of molarity of H SO
2
4
function of the molarity of the H SO sulfation solution.
2
4
2−
◦
–TiO samples calcined in air at 500 C for 5 h,
2
sulfation solution for SO
4
obtained by AAS/TGA and XPS, respectively.
3
.1.2. Thermogravimetric analysis
Fig. 1 shows TGA performed in air on SO4 –TiO2 mate-
2−
surface sulfate content first increased rapidly for low H2SO4
concentrations and more slightly for concentrations >0.5 M,
whereas the total sulfate loading increased quasilinearly up to
rials synthesized using aqueous sulfation solutions of different
molarities (0.25, 0.5, 1, and 5 M). The sulfated materials were
stable in air at least up to 550 C, before the weight loss oc-
◦
2
1 wt% for the 5 M H2SO4 solution. The significant slowing
curred, attributed exclusively to the sulfate removal. The be-
ginning of the weight loss was slightly shifted toward higher
temperatures for the lowest concentration of sulfation solution.
The weight loss allowed the determination of the total amount
of the slope of the curve representing the evolution of sur-
face sulfate species could be correlated with the deposition of a
monolayer of sulfate species on the titania surface. Indeed, tak-
ing into account the saturation coverage of sulfate species of
of SO42 species, which were equal to 2.5, 3.5, 5, and 21 wt%
for H2SO4 solutions with molarity 0.25, 0.5, 1, and 5 M, respec-
tively. Fig. 1 thus leads to the conclusion that a progressive de-
−
1
8
2
2
.4 × 10 atom/m , the sulfate equivalent monolayer cover-
2−
age corresponds to 4.4 wt% SO4 . The change in gradient of
the surface sulfate content curve corresponds to a total sulfate
content of 3.5 wt%, that is, a near-monolayer sulfate coverage.
Such a double-gradient XPS pattern has already been reported
for different systems like zirconia-supported tungsten oxide cat-
alysts [36,37].
pletion of SO42 started from calcination temperatures varying
−
◦
from 550 to 600 C, depending on the molarity of the sulfation
solution. For this reason, the standard calcination temperature
◦
in the present study was fixed at 500 C, so as to be ensure the
use and investigation of a material that is not partially desul-
2−
The Ti 2p XPS spectra reported in Fig. 3a show that the un-
sulfated TiO2 synthesized by the analogue sulfate-free sol–gel
TiO2 procedure, taken as a reference, exhibited a signal that can
be fitted with only two components, the largest one at 458.4 eV
for the Ti 2p3/2 and another one at 464.1 eV for the Ti 2p1/2,
corresponding to the Ti 2p spin-orbit components of Ti(IV) sur-
face species. Comparison of the unsulfated sol–gel TiO2 sample
fated. Note that the depletion SO4 species was achieved at
◦
8
00 C in all cases. The sulfate contents determined by TGA
were in close agreement with those deduced from the atomic
absorption spectroscopy (AAS) measurements performed in the
Service Central d’Analyse of the CNRS in Vernaison (France),
both of which are shown in Fig. 2.
◦
and the sulfated samples calcined at 500 C for 5 h, varying
3
.1.3. X-Ray photoelectron spectroscopy studies
in the sulfate loadings, showed an increasing upward shift of
Ti 2p (IV) binding energies of 0.2, 0.4, and 0.5 eV for sul-
fation solution molarities of 0.25, 0.5, and 1 M, respectively,
as summarized in Fig. 4. This shift to higher binding energy
could be explained by the increase in effective positive charge
around Ti(IV) surface species, suggesting the direct coordina-
tion of titanium atoms to strongly electron-withdrawing sulfate
centers. It thus can be proposed that sulfation of TiO2 led to the
electron transfer from TiO2 to sulfate anions. Further increas-
ing the surface sulfate species, by increasing the molarity of
the sulfation solution up to 5 M, totally annealed this upward
shift, which decreased to 0 eV. The volcano curve obtained in-
dicated that the direct coordination, and thus the effect of the
Fig. 2 shows the total and surface sulfate content as a
function of the molarity of the sulfation solution for samples
◦
calcined at 500 C for 5 h. Sulfur and titanium atoms being
2−
present at the surface as SO4
and TiO2 species, respec-
tively, the surface sulfate contents were deduced from sulfur-
to-titanium (S/Ti) atomic surface ratios, using the appropri-
ate molar weight-correction factors. It could be observed that
the surface sulfate loading exceeded the total sulfate loading
in all cases except for the 5 M sulfated material, indicating
a model in which most of the sulfur should be located at the
surface for low sulfation solution concentration. The incorpo-
ration of sulfates into the bulk titania framework should be
more important for the highest-molarity sulfation solutions. The

322
E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
2
−
Fig. 3. XPS spectra: (a) the Ti 2p region of TiO synthesized by the corresponding sulfate-free sol–gel procedure and sulfated-SO
–TiO samples calcined in
2
2
4
◦
2−
air at 500 C for 5 h with different molarities of the H SO solution; (b) the S 2p region of sulfated SO
4
◦
–TiO samples calcined in air at 500 C for 5 h as a
–TiO samples calcined in air at 500 C for 5 h as a
2
4
2
2−
◦
function of the molarity of the H SO sulfation solution; (c) the O 1s region of unsulfated and sulfated-SO
2
4
4
2
function of the surface sulfate loading. The surface and total amount of sulfate (wt%) are added into brackets and square brackets, respectively.
electron-withdrawing sulfate centers, was attenuated when in-
creasing the surface sulfate levels to totally disappear on the
highest-loaded sample.
The S 2p region of the XPS spectra (Fig. 3b; summary
in Fig. 4) exhibited a broad signal with a maximum centered
around 168.8 eV for the 0.25, 0.5, and 1 M SO42 –TiO2 cata-
lysts, whereas the maximum for the binding energy was shifted
to higher binding energies for the 5 M sulfated titania material
−
(
around 169.3 eV). This shift clearly showed a modification in
the sulfur environment at the surface of the high-molarity solu-
tion 5 M sulfated titania, probably resulting from the further
evolution of polysulfates to a deeper bulk sulfation. In con-
trast, low molarity of the sulfation solution did not influence
the electronic state of sulfur atoms, meaning that the sulfur en-
vironment was similar in the resulting low-molarity sulfated
materials. Similar results have been reported by Jung et al. [38]
Fig. 4. S 2p binding energy together with the positive/negative binding energy
shifts for XPS Ti 2p and O 1s contributions, as a function of the sulfation
solution molarity, for the different resulting sulfated-titania materials calcined
◦
at 500 C for 5 h.

E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
323
and Ecormier et al. [39] for sulfated titania and sulfated zirconia
materials, respectively, for which the binding energy of sulfur
was not affected by low sulfate content, up to 6.6 and 12 wt%,
respectively.
Fig. 3c shows the O 1s spectra obtained on the different
TiO2-based materials. The nonsulfated TiO2 displayed a dom-
inant peak at about 530.2 eV, corresponding to oxygen bonded
to titanium atoms [40]. Sulfation of TiO2 led to the appearance
of a peak shoulder around 532 eV, assigned to oxygen bonded
to the central atom of sulfur [41].
In addition, a low-molarity solution ranging from 0.25 to
1
M resulted in a slight shift toward lower energies of the
titanium-bonded oxygen contribution, from 530.2 eV down to
29.8 eV (Fig. 4). This lower energy shift was weaker for
5
higher-molarity sulfation solutions. The increase in sulfation re-
sulted in the gradual attenuation of this main O 1s peak and a
concomitant increase in the higher binding energy shoulder as-
signed to surface sulfates. For low-molarity sulfation solutions,
the evolution with increasing molarity of the higher energy shift
of the Ti 2p XPS signal, with the sulfate S 2p contribution kept
constant, led to the hypothesis that the electron accumulation
phenomena resulting from the electron transfer from TiO2 to
sulfate occurred exclusively on the oxygen atoms of the sul-
fate groups. Furthermore, the electron deficiency of the titanium
atoms resulting from the electron transfer to sulfate led to shift
the electronic density of the titanium–oxygen bond toward the
electronegative oxygen atoms, and also could explain the slight
shift toward lower energy of the titanium-bonded oxygen con-
tribution at 530.2 eV. Nevertheless, increasing the molarity of
the sulfation solution above 1 M decreased this binding en-
ergy shift, probably due to the appearance of new polysulfate
species with different oxygen electronic states and with no elec-
tron transfer from titania to sulfate, in agreement with the Ti 2p
XPS signal of the 5 M sample.
Fig. 5. Photocatalytic removal efficiency in terms of outlet toluene concentra-
tion as a function of time on stream over the commercial Degussa P25 reference
catalyst (2), sol–gel TiO (ꢀ), and 0.5 M-sulfated titania catalysts calcinated
2
◦
◦
◦
◦
2−
at 400 C ("), 500 C (!), 600 C (a), 800 C (Q), and SO
nated at 500 C using 0.25 M (P), 1 M (e) and 5 M (1) solutions.
–TiO calci-
4
2
◦
on stream under UV illumination. The increased molarity of the
sulfation solution up to 5 M (13.7 surface and 21 total sulfate
wt%) led to a drastic decrease in the photocatalytic efficiency
down to a near-zero level. In contrast, decreasing this value
to 0.25 M (corresponding to a catalyst with surface and to-
tal sulfate contents of 7.7 and 2.5 wt%, respectively) yielded
complete removal of toluene in 26 min. The use of calcination
◦
temperatures other than 500 C led to photocatalysts displaying
poor efficiency for the toluene removal, with decreasing perfor-
◦
mances for temperatures increasing from 400 to 800 C.
It should be noted that, for example, the titania-based ma-
terials synthesized using 0.5 and 1 M sulfation solutions dis-
played total toluene removal for about 42 min, whereas they
2
had surface areas of 116 and 60 m /g, respectively. In contrast,
3
.2. Photocatalytic performances of sulfated titania materials
photocatalysts with similar specific surface areas demonstrated
2−
totally different photocatalytic behaviors (e.g., the 1 M SO4
TiO and the sol–gel TiO photocatalysts). Taking into account
–
Fig. 5 shows the photocatalytic performances for toluene re-
2
2
moval in terms of outlet toluene concentration as a function of
time on stream, with an inlet toluene concentration of 110 ppm
and a relative humidity ratio of 30%, diluted in 200 ml/min
these results, the specific surface area of the materials should be
ruled out from advanced hypotheses and cannot be considered
a major parameter for explaining the photocatalytic efficiency
for the toluene removal, with another explanation given later in
this paper.
flowing air. The on-stream performance of SO42 –TiO2 pho-
tocatalysts as a function of the calcination temperature and the
sulfation solution molarity was also compared with that of the
widespread commercially available TiO2 P25 (Degussa) and
the corresponding sulfate-free sol–gel TiO2. The reference P25
photocatalyst displayed only 90% of total toluene removal and
rapidly deactivated on stream. The unsulfated TiO2 sol–gel ma-
terial exhibited slightly a better activity, reaching total removal
of the pollutant for only a few minutes, followed by rapid de-
activation. The sulfated titania catalysts demonstrated the best
efficiency for the 0.5 and 1 M sulfation solution molarity and a
−
It is worth noting that only CO and H O were detected
2
2
as gaseous reaction products whatever photocatalyst was used.
Fig. 6 shows the evolution of the selectivity toward CO to-
2
gether with the toluene concentration in the outlet stream over
2
−
the most efficient material, the 0.5 M SO4 –TiO sample cal-
2
◦
cined at 500 C. An appreciable deficit in carbon occurred,
suggesting the deposition of carbonaceous species on the cata-
lyst surface. Indeed, no carbon deposition on the surface should
correspond to 100% selectivity. No carbon deposit occurred,
whereas the overshoot in selectivity (>100%) at the beginning
◦
calcination temperature of 500 C, with total toluene removal
2−
within 42 min before deactivation. The 0.5 M SO4 –TiO2
of the test meant that CO was formed in excess during the first
2
(
11.1 surface and 3.5 total sulfate wt%) photocatalytic material
period of the catalytic test. The same experiment performed
without preadsorption of the toluene before illumination (i.e.,
by directly switching on the UV lamp when the reaction mix-
2−
exhibited a less pronounced deactivation than the 1 M SO4
TiO2 (11.5 surface and 5 total sulfate wt%) material with time
–

324
E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
Fig. 6. Outlet stream toluene concentration in ppm and selectivity toward CO
2
2−
◦
–TiO catalyst calcined at 500 C.
2
on the 0.5 M SO
4
2
−
Fig. 8. TGA (Thermal Gravimetry Analysis) of the 0.5 M SO
–TiO sample
4
2
◦
calcined at 500 C, under air (bold) and argon (fine), after reaction of toluene.
ture passes through the reactor) yielded the disappearance of
this initial CO2 excess. This observation led to the conclusion
that the initial CO2 excess was due to the oxidation of part of
the preadsorbed toluene on the fresh catalyst sample. After re-
recirculation or multiple-pass reactor. This could probably ex-
plain the absence of any reaction intermediates in the gas phase
even though they were observed on the catalyst surface after
testing.
2−
action, FTIR spectra of the 0.5 M SO4 –TiO2 photocatalyst
◦
calcined at 500 C, exhibited the presence of adsorbed reaction
intermediates such as benzaldehyde and benzoic acid, as shown
From Fig. 8, showing the TGA analysis of the 0.5 M SO42
−
–
2
in Fig. 7 and reported by Marci et al. [40]. Such sp -bound
TiO2 deactivated photocatalyst, it could be deduced that re-
moval of these adsorbed reaction intermediates was possible
in either air or argon, whereas the phenomenon was strongly
dependent on the nature of the carrier gas. Indeed, raising
the temperature in argon or in air from room temperature to
carbon aromatic molecules were poisons for the photocata-
lyst and thus were responsible for the catalyst deactivation by
blocking the active sites. This observation led to the conclusion
that the photocatalytic degradation of toluene would be mainly
due to the oxidation of the methyl group (i.e., the exo-cyclic
carbon–carbon bond breaking), as confirmed by complemen-
tary experiments carried out with benzene (not shown here).
The improvement of the intermediate desorption on sulfated
titania could lead to multiple adsorption–desorption steps for
the reactant and the reaction intermediates, through the 300-
mm photoreactor. The reactor can thus be considered to act as a
◦
900 C showed two or three different weight loss domains, re-
◦
spectively. In both cases, the weight loss in the 100–300 C
range was attributed to the desorption of adsorbed CO2 and
toluene, as observed by complementary thermodesorption ex-
periments (not shown). Indeed, a fraction of the produced CO2
has been reported to remain adsorbed on the photocatalyst dur-
ing volatile organic compound oxidation [33,34]. It is worth
2−
◦
Fig. 7. Fourier Transform Infra Red spectra of the 0.5 M SO
–TiO sample calcined at 500 C after reaction with toluene.
4
2

E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
325
noting that whatever the flowing gas, both thermal treatments
yielded the same total weight loss (about 9%), meaning that the
same initial species were removed from the deactivated cata-
lyst, despite totally different behavior patterns for the weight
loss and heat flow evolution. Under air, the weight loss at 300–
◦
5
00 C could be assigned to the exothermic combustion of the
strongly adsorbed oxygenated aromatic intermediates, before
the nonexothermic removal of sulfate groups occurred at higher
temperatures corresponding to the third weight loss domain, in
agreement with the XPS and TGA results on fresh sulfated-
titania materials. Using argon as carrier gas resulted in a unique
◦
weight loss domain at temperatures >300 C, attributed to the
simultaneous desorption of the adsorbed reaction intermediates
and removal of sulfates. The removal of sulfates at a lower
temperature in argon than in air could be explained by the weak-
ening or destabilization of the oxygen-to-sulfur bond due to the
desorption of the electron-rich oxygenated aromatics. Conse-
quently, the stability in air of the sulfates at the TiO2 surface up
2
−
Fig. 9. Influence of the calcination temperature of the 0.5 M SO
–TiO cat-
4
2
alyst on the S/Ti atomic surface ratio (1), the BET surface area ("), the mean
particle size derived from the XRD peak broadening using the Scherrer calcu-
lation and the anatase molar fraction.
◦
to 580 C allowed regeneration of the photocatalyst to be per-
formed by combusting the strongly adsorbed intermediates into
CO2 for 2 h at 450 C without destroying the sulfate-containing
Based on XPS analyses, the low-molarity sulfation proce-
◦
dure led to an electron transfer from TiO to the oxygen atoms
2
surface. One must underline that the regeneration of the 0.5 M
of sulfate ions, causing delocalization and a new electron distri-
2−
SO4 –TiO2 catalyst under air led to the restitution of the ini-
tial photocatalytic efficiency toward toluene removal.
bution on the surface of TiO , as described by Jung et al. [38].
2
Those authors attributed this electron transfer to sulfate ions,
◦
The heat release around 700–800 C at a constant sample
leading to an electronic deficiency on Ti atoms and to the for-
mation of Lewis acid sites. We assumed that this irreversible
charge transfer acted as a charge trap [42], increasing the spa-
tial charge separation between the photogenerated electrons and
holes and thus limiting their recombination, which is one of
the major limiting factors in photocatalysis. The sulfated tita-
nia catalysts with the highest shift in Ti2p binding energies
showed the best photocatalytic efficiencies. Taking into ac-
count the saturation coverage of sulfate species, the optimum
efficiency for toluene photocatalytic removal corresponds to a
weight could be assigned to the anatase-to-rutile transforma-
tion. Because the sulfate species, which are known to delay the
anatase-to-rutile transformation [25,26], are released under air
at a higher temperature than under argon, the heat release was
observed at a higher temperature under air.
3
.3. The role of the sulfate species
Sulfated titania materials led to completely different photo-
2−
near-monolayer surface coverage of the TiO2 surface by SO4
catalytic behaviors for the highly concentrated toluene effluent
removal depending on various synthesis parameters, such as
the molarity of the H2SO4 sulfation solution and the calcina-
tion temperature of the amorphous sulfated Ti(OH)4 hydroxide.
This could explain the strong discrepancy reported in the liter-
ature concerning the role of sulfates in photocatalysis and thus
required a systematic control of the synthesis procedure.
2−
species (i.e., the 0.5 M and the 1 M SO4 –TiO2 samples with
0
.8 and 1.2 equivalent of monolayer, respectively), keeping in
mind that the latter is certainly starting to built up polysulfate
species.
It has been shown that an important limitation is catalyst de-
2
activation due to strongly adsorbed sp -bound carbon aromatic
poisons derived from benzenes, such as benzaldehyde and ben-
zoic acid. The electron-rich nature of the sulfate species could
+
2−
Performing Ar bombardment on the 0.5 M SO4 –TiO2
titania (not shown) led to the removal of the sulfated top lay-
ers and resulted in a drastic decrease in the S/Ti atomic surface
ratio down to zero, proving the exclusive surface location of
2
thus favor the desorption of electron-rich sp -bound poisons
and limit the deactivation of the photocatalyst on stream.
To summarize, a large variety of sulfated titania materials
can be obtained with totally different photocatalytic behaviors
depending strongly on the experimental synthesis parameters,
with a main influence on the surface specific area, the crys-
tallinity of the material, the crystallographic nature of TiO2
(anatase or rutile), and the sulfate surface content.
2−
SO4 species. This assumption is confirmed in Fig. 2, which
shows the characteristic pattern of sulfated titania with strong
surface anchorage at low concentration and for which a deeper
bulk sulfation required a high molarity sulfation solution, as
reported by Ecormier et al. [39] for SO42 –ZrO2 materials.
Low molarity led exclusively to surface sulfation, whereas in-
creasing the molarity solution yielded progressive formation of
polysulfates and finally to bulk sulfation, as supposed for the 1
and 5 M sulfated samples, respectively, in agreement with Jung
et al. [38]. In this latter case, no drastic decrease in the S/Ti
−
Fig. 9 summarizes the influence of the calcination temper-
ature on some physicochemical properties, including the S/Ti
atomic surface ratio, BET specific surface area, anatase con-
tent, and mean particle size. For a 0.5 M sulfation solution,
a maximum S/Ti atomic surface ratio was observed for ma-
+
◦
atomic surface ratio was observed after Ar bombardment.
terials calcined at 400–500 C, with the material calcined at

326
E. Barraud et al. / Journal of Catalysis 235 (2005) 318–326
◦
4
00 C demonstrating a lower specific surface area than the ma-
UMR 7504 CNRS) and C. Petit (LMPSC) for helping with
the TGA characterizations and discussions, and P. Bernhardt
(LMSPC) for helping with the XPS characterizations.
◦
terial calcined at 500 C. Higher calcination temperatures led
to a decrease in the S/Ti surface ratio, down to zero at 800 C,
◦
due to progressive desulfation of the catalyst beginning around
◦
6
00 C, as confirmed by the TGA characterization (Fig. 1). The
References
optimum calcination temperature can result from the highest
specific surface area and good crystallinity, obtained with an
adequate S/Ti atomic surface ratio (i.e., without desulfation),
thus leading to higher adsorption capacity and photoefficiency.
On the optimized catalysts, the main roles of TiO2 are in
band gap excitation in the particle bulk and in reactant ad-
sorption and transformation at the particle surface, whereas the
surface sulfate groups are supposed to play a double role of a
charge trap and poison desorption promoter. The explanation
for this that has been put forward is that an optimum surface
sulfate coverage is necessary to provide the most efficient pho-
tocatalytic material and has been found close to the equivalent
monolayer coverage. This model implies that the presence of
[
[
[
[
1] M. Schiavello (Ed.), Photocatalysis and Environment. Trends and Appli-
cations, Kluwer-Academic Publishers, Dordrecht, 1988.
2] E. Pelizzetti, N. Serpone (Eds.), Photocatalysis. Fundamentals and Appli-
cations, Wiley, New York, 1989.
3] D.F. Ollis, E. Pelizzetti, N. Serpone, Environ. Sci. Technol. 25 (1991)
1522.
4] D.F. Ollis, H. Al-Ekabi (Eds.), Photocatalytic Purification and Treatment
of Water and Air, Elsevier, Amsterdam, 1993.
[
[
[
5] J.M. Herrmann, Catal. Today 53 (1999) 115.
6] W.H. Strehlow, E.L. Cook, J. Phys. Chem. 2 (1) (1979) 163.
7] B.J. Liu, T. Torimoto, H. Yoneyama, J. Photochem. Photobiol. A 113
(1998) 93.
[8] J.A. Navio, G. Colon, J.M. Herrmann, J. Photochem. Photobiol. A 108
(1997) 179.
[9] M. Fujihira, Y. Satoh, T. Osa, Nature 293 (1981) 206.
2−
both TiO2 and well-dispersed SO4 with optimized contact
[
10] A. Navio, M. Garcia Gomez, M.A. Pradera Adrian, J. Fuentes Mota,
2
−
between SO4
and TiO2 domains is required. Excessively
J. Mol. Catal. 104 (1996) 329.
high SO42 surface coverage decreases the number of TiO2
surface sites available for the photocatalytic reaction, whereas
a low sulfate content reduces the connection (i.e., the intimate
contact) between sulfate ions and TiO2.
−
[11] L. Cao, Z. Gao, S.L. Suib, T.N. Obee, S.O. Hay, J.D. Freihaut, J. Catal. 196
2000) 253.
12] T. Ibusuki, K. Takeuchi, Atm. Environ. 20 (9) (1986) 1711.
13] T.N. Obee, R.T. Brown, Environ. Sci. Technol. 29 (1995) 1223.
14] V. Augugliaro, S. Coluccia, V. Loddo, G. Martra, L. Palmisano, M. Schi-
avello, Appl. Catal. B 20 (1999) 15.
(
[
[
[
4
. Conclusion
[15] V. Augugliaro, V. Loddo, G. Marci, L. Palmisano, C. Sbriziolo, M. Schi-
avello, M.L. Turco Liveri, Stud. Surf. Sci. Catal. 130 (2000) 1973.
[
[
[
[
16] H. Einaga, S. Futamura, T. Ibusuki, Appl. Catal. B 38 (2002) 215.
17] O. d’Hennezel, D.F. Ollis, J. Catal. 167 (1997) 118.
18] O. d’Hennezel. PhD thesis, North Carolina State University, 1998.
19] M. Lewandowski, D.F. Ollis, J. Catal. 217 (2003) 38.
It has been shown that systematic control of the sulfation
procedure parameters allowed optimization of the synthesis of
2−
SO4 –TiO2 catalysts and led to enhanced photocatalytic prop-
erties toward concentrated toluene effluent removal by both in-
creasing the photocatalytic total efficiency and attenuating the
[20] E. Berman, J. Dong, in: W.W. Eckenfelder, A.R. Bowers, J.A. Roth (Eds.),
rd Int. Symp. Chem. Oxidation, Technologies for the Nineties, vol. 3,
3
deactivation with time on stream. The most efficient SO42
−
–
Technomic, Lancaster, PA, 1993, p. 183.
[
[
[
21] K. Arata, Appl. Catal. 146 (1996) 3.
22] K. Arata, Adv. Catal. 37 (1990) 165.
23] L.K. Noda, R.M. De Almeida, L.F.D. Probst, N.S. Gonçalves, J. Mol.
Catal. A 225 (2005) 39.
TiO2 photocatalyst was obtained for a near-monolayer sulfate
coverage. The positive role of sulfates has been attributed to
i) an electron transfer from titanium to sulfates, leading to
(
[
[
[
[
[
24] A. Corma, A. Martinez, C. Martinez, Appl. Catal. A 144 (1996) 249.
25] G. Colon, M.C. Hidalgo, J.A. Navio, Appl. Catal. B 45 (2003) 39.
26] X. Fu, Z. Ding, W. Su, Chin. J. Catal. 20 (1999) 321.
27] D.S. Muggli, L. Ding, Appl. Catal. B 32 (2001) 181.
28] E.P. Barrett, L.G. Joyner, P.H. Halenda, J. Am. Chem. Soc. 73 (1951) 373.
a positive charge trap effect and thus hindering photogener-
ated electron-hole recombination, and (ii) a better desorption of
electron-rich sp -bound carbon aromatic poisons on electron-
rich sulfates, thus limiting deactivation. The better stability in
air toward temperature of the sulfates compared with the aro-
matic deposit allowed regeneration of the photocatalyst. This
halide-free sulfate-promoted TiO2 photocatalyst is thus a re-
generable and highly efficient alternative for improving the on-
stream gas phase toluene removal, for which commercial TiO2
and other alternatives lack efficiency.
2
[29] R.S.H. Mikhail, S. Brunauer, E.E. Bodor, J. Colloid Interface Sci. 26
1968) 45.
(
[
[
[
[
30] S. Doniach, M. Sunjic, J. Phys. C: Solid State Phys. 3 (2) (1970) 285.
31] D.A. Shirley, Phys. Rev. B 5 (1972) 4709.
32] J.H. Scofield, J. Electron. Spectrosc. Relat. Phenom. 8 (1976) 129.
33] V. Keller, F. Garin, Catal. Commun. 4 (2003) 377.
[34] V. Keller, P. Bernhardt, F. Garin, J. Catal. 215 (2003) 129.
[35] N. Keller, V. Keller, E. Barraud, F. Garin, M.J. Ledoux, J. Mater. Chem.
1
4 (12) (2004) 1887.
Acknowledgments
[
[
[
36] F. Di Gregorio, V. Keller, J. Catal. 225 (2004) 45.
37] Y. Xie, Y. Tang, Adv. Catal. 37 (1990) 1.
38] S.M. Jung, P. Grange, Catal. Lett. 76 (2001) 1.
[39] M.A. Ecormier, K. Wilson, A.F. Lee, J. Catal. 215 (2003) 57.
[40] G. Marci, M. Addamo, V. Augugliaro, S. Coluccia, E. Garcia-Lopez,
V. Loddo, G. Marta, L. Palmisano, M. Schiavello, J. Photochem. Photo-
biol. A 160 (2003) 105.
41] S.M. Jung, P. Grange, Catal. Today 59 (2000) 305.
42] R. Gomez, T. Lopez, E. Ortiz-Islas, J. Navarrete, F. Tzompanztzi, X. Bok-
himi, J. Mol. Catal. A 193 (2003) 217.
This work was carried out in an “Action Concertée Incita-
tive” framework under the auspices of the French Educational
Ministry, which E. Barraud thanks for providing a postdoctoral
fellowship grant. The authors are grateful to the Chemistry De-
partment of CNRS for providing a postdoctoral fellowship grant
to F. Bosc. The authors also thank M. Bacry (LMSPC) for per-
forming the surface area measurements, D. Burger (IPCMS,
[
[