Ambient oxidation of benzene to phenol by photocatalysis on Au/Ti0.98V0.02O2: Role of holes

Article abstract of DOI:10.1021/cs500724z

A potential photocatalyst with 2 atom % vanadium incorporated into the lattice of disordered mesoporous titania, Ti_0.98V_0.02O₂, (TV2) was synthesized. Au was deposited on TV2 (Au/TV2) through a photodeposition method. Stru

Full text of DOI:10.1021/cs500724z

Research Article
pubs.acs.org/acscatalysis
Ambient Oxidation of Benzene to Phenol by Photocatalysis on
Au/Ti0.98V0.02O2: Role of Holes
Perumal Devaraji,† Naveen K. Sathu,† and Chinnakonda S. Gopinath*,†,‡,§
†Catalysis Division, National Chemical Laboratory, Dr. Homi Bhabha Road, Pune 411 008, India
‡Network of Institutes for Solar Energy (NISE), NCL Campus, Dr. Homi Bhabha Road, Pune 411 008, India
§Centre of Excellence on Surface Science, National Chemical Laboratory, Dr. Homi Bhabha Road, Pune 411 008, India
S
* Supporting Information
ABSTRACT: A potential photocatalyst with 2 atom % vanadium incorporated into the
lattice of disordered mesoporous titania, Ti0.98V0.02O2, (TV2) was synthesized. Au was
deposited on TV2 (Au/TV2) through a photodeposition method. Structural, microscopy,
and spectroscopy techniques support the incorporation of vanadium into the TiO2 lattice,
and Au was deposited on the surfaces of TV2. Photocatalytic oxidation of benzene was
conducted at ambient temperature under UV and/or visible light to demonstrate the
catalytic activity of the Au/TV2 catalyst. The TV2 lattice exhibits a quantum jump in
benzene to phenol oxidation compared to that of TiO2, highlighting the importance of V
for oxidation. Introduction of Au onto TV2 further increases the benzene to phenol
oxidation and phenol yield by a factor of 2 under UV light compared to those of bare TV2.
No significant phenol production was observed in visible light with or without gold,
indicating the role of gold is indirect toward charge separation and electron storage. Nano
gold clusters on TV2 selectively store photoexcited electrons and in turn maximize holes
utilization on TiO2. The high photocatalytic activity of Au/TV2 is mainly attributed to the presence of Schottky junctions,
disordered mesoporosity, and short diffusion lengths for charge carriers.
KEYWORDS: titania, photocatalysis, phenol, vanadium, Schottky junction, nano gold
visible light with layered titanate containing immobilized gold
nanoparticles in the interlayer space. However, this result is
possible only when the initial amount of benzene is a few
hundred parts per million (x) but in the presence of large
excess of a seed amount of phenol (≥10x). When there is no
seed phenol added, there is no phenol production observed.4a
Ide et al. also reported benzene oxidation over Au/TiO2 with a
13% yield and a 89% phenol selectivity after 24 h under a CO2
1. INTRODUCTION
In green chemistry, photocatalysis plays an important role in
solving some of the environmental issues. Of most of the
semiconductor photocatalysts, TiO2 plays a major role because
of its extraordinary chemical stability by minimizing photo-
corrosion and its high oxidizing ability through holes produced
upon light absorption.1 The oxidation of benzene to phenol by
photocatalysis is one of the important reactions attempted by
many researchers, but without a significant breakthrough.2−9
Direct oxidation of benzene through the cumene process10 is a
currently used method for the production of phenol that
involves three steps; in this process, benzene and propylene are
compressed to a pressure of 30 bar at 250 °C, in the presence
of an acid catalyst, such as phosphoric acid. The disadvantage of
this synthesis process is the formation of cumene hydro-
peroxide, which is an explosive, and acetone as a side product;
this necessitates the discovery of a suitable new green method
for phenol synthesis. Benzene to phenol oxidation on layered
double hydroxides in one step has been reviewed recently.11
As early as 1981, Fujihira et al. demonstrated photocatalytic
oxidation of toluene with TiO2 in the presence of H2O2 and UV
light from 500 W Xe lamp.2 Palmisano et al. reported selective
photocatalytic oxidation of 4-methoxy benzyl alcohol to the
corresponding aldehyde in an aqueous suspension of TiO2
catalyst under UV light.3 Ide et al.4 showed a 62% yield with a
96% phenol selectivity for photocatalytic benzene oxidation in
atmosphere (230 kPa) under 1 sun irradiation conditions.4b
A
high phenol yield (∼60%) with highly dispersed Au on TiO2 in
visible light was demonstrated by Zheng et al.;5 a high level of
Au dispersion was possible due to prereduction of titania to
blue titania (contains Ti3+), either by UV photoreduction5a or
by Zn2+ doping in titania,5b and helps to reduce gold ions in a
uniform manner with Schottky junctions. Benzene oxidation
was conducted in a manner similar to that described in ref 4a.
Fe-doped g-carbon nitride coated on SBA-15 exhibits 11.9%
benzene conversion with 20.7% phenol selectivity at 60 °C in
visible light.7 Selective photocatalytic oxidation of benzene to
phenol (38% benzene conversion and 97% phenol selectivity)
was demonstrated on an FeCl3/mpg-C3N4 hybrid with λ ≥ 420
nm at 60 °C; the same catalyst shows only 4% benzene
Received: January 22, 2014
Revised: July 3, 2014
Published: July 15, 2014
© 2014 American Chemical Society
2844
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
conversion.8 The survey described above highlights the
challenges involved in designing a new photocatalyst that
gives a high benzene conversion and a high phenol selectivity at
ambient temperature and pressure in the field of photocatalysis.
Monomeric or isolated V is known to be a good partial
oxidation or oxidative dehydrogenation catalyst.12 Because of
the high partial oxidation nature associated with isolated V, it is
worth exploring vanadium doped into the lattice of TiO2 (V/
TiO2) under the photocatalytic conditions. V/TiO2 is useful
not only for benzene oxidation but also for many other
oxidation reactions.13−18 Earlier, Tanarungsun et al. showed the
photocatalytic activity of vanadium-impregnated TiO2 for
benzene oxidation under a UV light source with a 0.19%
yield and a 90% selectivity.6 It is important to keep the V in an
isolated or monomeric state for the desired partial and selective
oxidation. It is also essential to prepare the V/TiO2 system with
mesoporosity for better activity. Although the ordered
mesoporous material shows a large surface area, the reactant
and product molecules have to diffuse a long distance to and
from the active sites that are present in the mesopores. This is
due to the availability of more active sites on the surface of
porous channels. Hence, the presence of a short diffusion
length associated with disordered mesopores is likely to be
helpful in minimizing the diffusion problem.19−21 This is
expected to improve the catalytic activity and selectivity of the
desired product.
In this study, we have synthesized 2 atom % V doped in
anatase TiO2 (TV2, Ti0.98V0.02O2) by the solution combustion
method (SCM)12,22 and analyzed its photocatalytic activity for
the oxidation of benzene to phenol in a biphasic system under
UV−visible light under ambient conditions. The photo-
deposition method was used to deposit Au on TV2 (Au/
TV2), producing 18% benzene conversion with phenol
selectivity and yields of 88.1 and 15.9 mol %, respectively,
after UV irradiation with Au/TV2 for 18 h. This study is a part
of various efforts from our laboratories to synthesize rationally
designed photocatalytic materials.23
Two samples were prepared by changing the molar
concentration of HAuCl4. The sample with a concentration
of 0.0005 M was named 1Au/TV2, and the sample with a molar
concentration of 0.0015 M was named 3Au/TV2. ICP analysis
shows the Au content to be 0.95 and 2.92 wt %, and they are
denoted as 1Au and 3Au, respectively, as shown above.
2.2. Characterization. To explore the structural properties
of as-synthesized materials, different characterization techni-
ques have been used, such as high-resolution transmission
electron microscopy (HRTEM), scanning electron microscopy
(SEM), powder X-ray diffraction (PXRD), N2 adsorption−
desorption isotherms, photoluminescence (PL), energy dis-
persive X-ray analysis (EDX), and X-ray photoelectron
spectroscopy (XPS). PXRD data were collected from a PAN
analytical X’pert Pro dual goniometer X-ray diffractometer,
using Cu Kα (1.5418 Å) radiation with a Ni filter, and the data
were collected using a flat holder in Bragg−Brentano geometry
(0.2°). The SEM system equipped with an EDX attachment
(FEI, model Quanta 200 3D) was used for morphological and
chemical composition. The surface area and pore size
distribution of the materials were analyzed on a Quadrasorb
SI instrument; the measurement was conducted after the
materials had been degassed at 200 °C. HRTEM of the
materials was conducted on a FEI TECNAI 3010 electron
microscope operating at 300 kV (Cs = 0.6 mm; 1.6 Å
resolution). Raman spectra of the samples were analyzed with a
Horiba JY LabRAM HR 800 Raman spectrometer coupled with
a microscope in reflectance mode with a 633 nm excitation
laser source and a spectral resolution of 0.3 cm−1. XPS
measurements have been taken with a custom-built ambient-
pressure XPS system from Prevac and equipped with a VG
Scienta SAX 100 emission controller monochromator using an
Al Kα anode (1486.6 eV) in transmission lens mode.24 The
photoelectrons are energy analyzed using VG Scienta’s R3000
differentially pumped analyzer. PL measurements were
performed using a Horiba Jobin Yuon Fluorolog 3 spectropho-
tometer equipped with a 450 W xenon lamp with a tunable
excitation wavelength. All PL measurements were taken with
325 nm excitation pulses. Diffuse reflectance UV−vis measure-
ments were taken on a Shimadzu spectrophotometer (model
UV-2550) with spectral-grade BaSO4 as reference material.
2.3. Photocatalytic Activity and Photocurrent Meas-
urements. Photocatalytic oxidation reactions were conducted
in a quartz RB flask for different periods of time with a 400 W
mercury lamp (λ = 200−400 nm) using a flow-type double-
jacketed quartz reactor.23 In a typical experiment, 30 mg of the
catalyst was dispersed in a quartz RB flask, containing an initial
reactant mixture of 2 mL of CH3CN, 1 mL of benzene, and 2
mL of 25% H2O2. Unless mentioned otherwise, dissolved
oxygen in the solution described above was not removed. Few
irradiation experiments were conducted with a WG.305 filter to
block the radiation between 200 and 309 nm. The temperature
of the double-jacketed reactor and quartz RB was maintained at
25 °C by allowing water circulation. Before any irradiation, the
reaction mixture was stirred for 30 min in the dark and
irradiation was conducted for different lengths of time between
6 and 24 h. After irradiation, organic and aqueous layers were
separated with a separating funnel. Oxidized products of
benzene were analyzed with an Agilent 6890 gas chromato-
graph (GC) equipped with a flame ionization detector (FID)
and an HP-5.5% phenyl methyl siloxane capillary column. The
presence of phenol was confirmed by analyzing both organic
and aqueous layers. It is to be noted that the presence of any
2. EXPERIMENTAL SECTION
2.1. Preparation of the Catalysts. All the chemicals
employed in this study were of analytical grade and used as
such. Titanium isopropoxide as a Ti precursor, ammonium
metavanadate as a V precursor, and urea as a fuel were used.
The procedure followed for the synthesis of TV2 by SCM was
previously reported by our group.12 The synthesis of V-doped
TiO2 was conducted with a 1:1 molar ratio of urea to metal ions
(Ti and V). The required amount of reactants was taken in a
250 mL beaker, and the aqueous solution was stirred for 1 h,
followed by introduction of the solution described above into a
muffle furnace preheated at 400 °C. Evaporation of water starts
at the beginning followed by smoldering-type combustion that
results in a powder material. Similarly, TiO2 was synthesized
without adding any V precursor.
Gold deposition was conducted by the photodeposition
method using HAuCl4 as the gold precursor. In a 250 mL
quartz round-bottom flask, 30 mL of required HAuCl4 was
taken in 120 mL of MeOH with 500 mg of catalyst. The
solution described above was bubbled with Ar at a rate of 50
mL/min for 1 h to remove any dissolved oxygen and generate
an inert atmosphere. Irradiation with a 400 W UV lamp was
conducted for 1 h (λ = 200−400 nm). After photodeposition,
the powder sample was separated by centrifugation, and finally,
the Au-deposited powder sample was dried at 40 °C for 12 h.
2845
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
other liquid products in the chromatogram was included in the
quantitative estimation.
materials, such as SBA-15 and MCM-48, all TV2 and titania
materials exhibit only one peak around 0.8° without any extra
peaks.25 This highlights the presence of disordered meso-
porosity12,19−21 for all TV2 materials.
Chronoamperometry measurments were recorded at 0 V
using a Gamry potentiostat (model 3000 potentiostat/
galvanostat/ZRA) in a conventional three-electrode test cell
with Ag/AgCl as the reference electrode and a platinum foil as
the counter electrodes with an Oriel Instruments 300 W xenon
arc lamp equipped with an A.M 1.5 filter.23 The immediate
response for shutter oscillation highlights the photofunctional
behavior of the material. As long as the material is irradiated,
the current remains constant, demonstrating production of the
current is exclusively due to the photoresponse.
3.1.1. Surface Area Analysis. N2 adsorption−desorption
studies have been conducted to investigate the textural
properties of mesoporous TiO2, TV2, 1Au/TV2, and 3Au/
TV2 materials. Figure 2a shows the results of N2 adsorption−
desorption isotherm studies conducted at 77 K and analyzed by
the BET method for surface area, and Figure 2b the Barret−
Joyner−Halenda (BJH) pore size distribution of TiO2 and all
TV2 materials. The surface area and porosity have been
calculated for all of the materials, and the results are listed in
Table 1. All TV2 materials show a type IV adsorption−
desorption isotherm with an H2 hysteresis loop that clearly
indicates the presence of mesopores.21,23 An average pore size
distribution in the range of 4.5 2 nm and an overall pore
volume of 0.26 0.02 cm3/g were observed. TV2 exhibits a
surface area (166 m2 g−1) larger than that of TiO2 (117 m2
g−1). This indicates that when gold was introduced onto the
surface of TV2, the surface area decreased marginally compared
to that of TV2. Because of the high level of dispersion of gold
nanoparticles on TV2, the surface area of 3Au/TV2 (151 m2
g−1) is similar to that of 1Au/TV2 (156 m2 g−1). The pore size
distribution profile indicates the presence of mesopores in all of
the catalysts. A significant shift in pore size toward a lower value
is evident for 3Au/TV2 (Figure 2b), highlighting the
introduction of gold nanoclusters into pores, too.
3. RESULTS AND DISCUSSION
3.1. Structural and Textural Properties Determined by
X-ray Diffraction Studies. PXRD studies have been
conducted to explore the structural aspects of prepared
catalysts. Figure 1 shows the PXRD pattern of all catalysts
3.1.2. HRTEM and SEM. The morphology and textural
properties of TV2 and Au/TV2 materials were analyzed by
HRTEM and SEM, and representative images are shown in
Figure 3. Generally, TV2 and 1Au/TV2 (and TiO2, not shown)
exhibit spherical particles and disordered mesoporous struc-
tures (Figure 3a−c). TV2 and Au/TV2 materials show textural
characteristics similar to those of TiO2.21 Disordered
mesoporosity arises from the intergrowth of fundamental
particles, and the same leads to aggregates with significant extra
framework void space.20,26,27 The nanocrystalline nature and
anatase phase of the materials were confirmed by the selected-
area electron diffraction (SAED) pattern. HRTEM shows a
particle size of ∼14 nm with a d spacing of 0.35 nm, which
corresponds to the (101) plane of the TiO2. These results are
supported well by XRD as well as N2 adsorption isotherm
results. The disordered mesoporous structure has additional
advantages like low diffusional barriers, because the depth of
mesopores is minimal, a few nanometers, unlike values of
several hundred nanometers in conventional ordered meso-
porous materials, like SBA-15.25 These types of mesopores are
known as pseudo three-dimensional (p3D) mesopores.21 This
disordered p3D mesoporous framework provides an easy route
for the diffusion of reactants due to lower diffusion barriers as
well as short diffusion lengths due to the small meso channel
depth. Figure 3b demonstrates the particles are electrically
interconnected together, which helps for the fast mobility of
charge carriers to the surface of the catalyst where the reaction
could occur.
Figure 1. XRD pattern of different materials, namely, TiO2, TV2,
1Au/TV2, and 3Au/TV2. The inset shows the low-angle XRD pattern,
indicating its disordered mesoporous nature.
prepared by SCM. The PXRD pattern of all materials was
found to match that of the anatase phase (JCPDS Card No. 21-
1272). No peaks corresponding to oxides of vanadium are
observed, which indicates that TiO2 and TV2 retain the anatase
phase with no V2O5 phase. No V2O5 formation in TV2
indicates the potential of the SCM method to introduce V into
the TiO2 lattice. The broad nature of the peaks confirms the
presence of nanocrystalline particles. No peaks were observed
for metallic Au in 1Au/TV2 or 3Au/TV2, which indicates a
high level of dispersion of Au on TV2. The respective average
crystallite size, calculated from the peak width with the Scherrer
equation, is observed to be between 8 and 12 nm. The inset of
Figure 1 shows the low-angle XRD pattern. A single diffraction
peak observed around 0.8° demonstrates the mesoporous
nature of the Ti1−xVxO2 materials. Unlike ordered mesoporous
The presence of gold is clearly visible by HRTEM (Figure
3b,c). The measured d spacing is 0.24 nm, which corresponds
to the (111) plane of metallic Au. The size of the gold particle
is approximately 7−8 nm; however, smaller particles were also
observed. The binding energy for typical metallic gold was
observed in XPS (vide infra) underscores the nature of gold
particles to be metallic on Au/TV2. Very importantly, the Au−
2846
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
Figure 2. (a) N2 adsorption−desorption isotherm and (b) pore size distribution pattern of TiO2, TV2, and Au/TV2 materials.
Table 1. Physicochemical Properties of TiO2, TV2, and Au/
TV2 Catalysts
BET surface area
(m2/g)
pore size
(nm)
pore volume
(cm3/g)
catalyst
TiO2
117
166
156
151
6.2
4.6
6.2
4.4
0.24
0.27
0.28
0.25
TV2
1Au/TV2
3Au/TV2
TiO2 metal−semiconductor (Schottky) junction was observed
easily with Au/TV2; (111) facets of nano gold are in direct
contact with (101) facets of titania shown in panels d and e of
Figure 3. Indeed, this is a criticial feature of Au/TV2 catalysts
that is essential for the separation of the charge carriers and
hence enhances the diffusion of charge carriers toward the
surface of the catalysts from the bulk. Both V and Au could act
as electron-trapping centers, and hence, the extent of hole
utilization increases significantly. This also leads to a decrease
in the level of charge recombination. In fact, this observation is
in good agreement with PL results, which show an increase in
the level of electron transfer and hence an increase in the
number of emission features compared to that of TV2.
3.2. Absorption and Emission Properties Determined
by UV−Visible Absorption Studies. UV−visible absorption
was measured to study the optical properties and examine the
deposition and doping effect of Au and V, respectively, on
TiO2. Figure 4 depicts the UV−vis absorption spectra, and the
inset shows the color associated with TiO2, TV2, and Au/TV2
materials. TiO2 shows absorption up to 380 nm in the UV
region with a 3.2 eV band gap, which is typical for titania. V
doping increases the level of visible light absorption up to 560
nm in TV2 catalysts. The visible light absorption in TV2 is
attributed to low-energy charge transfer bands originating from
charge transfer between O 2p and V 3d. This band position is
influenced by the coordination and oxidation state of V. The
V5+ state leads to absorption up to 550 nm, whereas V4+ leads
to a d−d transition band between 550 and 800 nm;12c,28 the
latter is absent in our study, suggesting the oxidation state to be
V5+ in TV2. This is also in agreement with XPS studies. The
Figure 3. TEM images of (a) TV2 and (b and c) Au/TV2, HRTEM
images of (d) 1Au/TV2 and (e) 3Au/TV2, and (f) SEM image of
1Au/TV2. Abundantly available Schottky junctions between (111)
facets of nano Au cluster and (101) facets of titania particles are shown
in panels d and e. The SAED pattern given in panel d shows (101) is
the major facet of TiO2 exposed on the surfaces.
2847
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
wavelength 325 nm. TiO2, TV2, and Au/TV2 composites show
emission features around 415, 441, and 470 nm, respectively.
The emission peak due to the band gap transition appears at
380 nm for all materials (not shown). A high-intensity emission
feature at 441 nm originates from the charge transfer transition
from Ti3+ to the oxygen anion in a [TiO6]8− complex present
on the material along (101) planes.29 The emission band at 415
nm is attributed to the free exciton emission of TiO2. After the
sample had been doped with V, a significant red shift in the first
emission band to 422 nm was observed, while no shift was
observed with the latter two emission features.
This observation underscores the introduction of new
acceptor states below the CB, due to the reduction of V5+,
which is supported by UV−vis absorption and XPS studies.
Further, the onset of emission of the first band also shifts from
titania to TV2 by the same extent. All three emission features
show the lowest intensity for TV2 and the highest for undoped
TiO2, indicating a significant electron trapping by doped V in
TV2. This process makes the holes available for oxidation on
the titania surface and leads to charge separation. Nonetheless,
upon deposition of Au on TV2, the intensity of all emission
features was observed to be higher than that of TV2;25b 1Au/
TV2 exhibits an emission intensity higher than that of 3Au/
TV2. It is to be underscored here that the anatase (101) facet is
the most abundant plane (Figure 3) and corrugated with
alternating rows of five-coordinated Ti (Ti3+) and bridging
oxygen at the edges of corrugation.30 This fact suggests the
reduction of Ti due to V doping. The five-coordinated surface
Ti3+ makes (101) an ideal facet for interaction with deposited
nano Au clusters due to a high level of surface unsaturation. In
fact, the interaction described above also reiterates the presence
of Schottky junctions in Au/TV2 nanocomposites, which helps
in electron−hole separation at the interface of Au and TiO2.5,24
The presence of Schottky junctions was directly supported by
HRTEM measurements, which supports the conclusion
described above. Indeed, this is an important observation that
indicates the surface electronic interaction between nano Au
clusters and TV2. Further, as nano metal clusters are a good
electron sink, they separate and store electrons and make the
holes available predominantly for oxidation. In essence,
electron trapping by V5+ on TV2 and electron storage by
gold clusters on Au/TV2 through Schottky junctions work
simultaneously and lead to an overall increase in hole
availability and/or level of utilization for the oxidation reaction.
When additional Au is loaded (3Au/TV2), the intensity of all
emission features decreases. Nonetheless, Au/TV2 exhibits
emission features with an intensity higher than that of TV2,
which underscores the transfer of an electron from nano Au to
the CB of titania. The decrease in the intensity of 3Au/TV2
emission features is due to an increase in the size of the gold
cluster (Figure 3e) and hence a weakened interaction with the
surface. On the basis of PL spectra, it can be inferred that
materials prepared by the SCM method show a substantial
decrease in their level of charge recombination due to selective
electron trapping by V and nano gold clusters, which enhances
the utilization of holes for oxidation.
Figure 4. UV−vis spectra of TiO2, TV2, 1Au/TV2, and 3Au/TV2
materials.
absorption intensity in the visible region, between 500 and 600
nm, is significantly enhanced after the deposition of Au on
TV2, and this is attributed to the surface plasmon resonance
(SPR) absorption by nano gold clusters.25 Strong interaction
between Au and TV2 leads to the enhancement of visible light
absorption, and Schottky junction helps for the separation of
charge carriers. However, nano gold does not absorb significant
UV light to demonstrate SPR, because of the poor absorption
coefficient.
3.2.1. Photoluminescence Studies. Figure 5 shows the PL
emission spectra of TV2 materials measured with an excitation
3.2.2. Raman Spectroscopy. Figure 6 shows the Raman
spectra of all catalysts compared with that of V2O5 as a
reference. All the catalysts show Raman active fundamental
modes12,31 for the anatase phase of titania. It is observed that
the intensity of all the Raman active modes decreases, and
broadening occurs after V doping. This decrease in the peak
intensity for TV2 and Au/TV2 reveals symmetry breaking of
Figure 5. Photoluminescence (PL) spectra of TiO2, TV2, 1Au/TV2,
and 3Au/TV2 materials.
2848
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
observed for V2O5, which indicates the absence of any V2O5
phase with the TV2 or Au/TV2 catalyst.32
3.3. X-ray Photoelectron Spectroscopy. XPS studies
were conducted to analyze the electronic states of TV2
catalysts. Figure 7 shows XPS spectra of Ti 2p, V 2p3/2, and Au
4f core levels of TV2 and Au/TV2 materials. The Ti 2p3/2 core
level of TiO2, TV2, and Au/TV2 appears at a binding energy
(BE) of 459.1 0.1 eV, and it confirms the presence of Ti4+ in
all of the materials.33 V doping in the titania lattice shows a V
2p3/2 peak at a BE of 517.3 eV for TV2 as well as Au/TV2,
indicating the oxidation state of V is predominantly +5 (Figure
7a, inset). It should be noted that the surface atom percent of V
increases from 0.4% on TV2 to 0.7% on 3Au/TV2; however,
the bulk atom percent of V remains at 1.9%, measured from
ICP analysis. The O 1s core level was observed at a BE of ∼530
0.2 eV for all materials (data not shown). Au 4f (Figure 7b)
core level spectra of 1Au/TV2 show at BE of Au 4f7/2 at 83.8
eV; it corresponds well to the BE of metallic gold at 84.0 eV.34
These results suggests that gold and V are present in the
metallic state (Au0) and V5+, respectively.25 The electronic
interaction described in the preceding sections does not induce
an observable shift in the BE in XPS.
3.4. Photocatalytic Benzene Oxidation. Panels a and b
of Figure 8 show the photocatalytic activity measured with a
TV2 catalyst at different time intervals for benzene to phenol
oxidation under UV irradiation. Figure 8c shows the phenol
yield observed with TV2 and Au/TV2 catalysts after irradiation
for 18 h. TV2 shows a 3% benzene conversion with a 100%
selectivity after irradiation for 6 h. The level of benzene
conversion increased linearly from 3 to 13% with an increasing
irradiation time from 6 to 24 h, respectively; however, a
significant decrease in phenol selectivity [from 100 to 85%
(Table 2)] was also observed. The decrease in phenol
selectivity is mainly due to further oxidation of phenol to
other side products, such as hydroquinone. Also, a qualitative
Figure 6. Raman spectra of the V2O5, TiO2, TV2, 1Au/TV2, and 3Au/
TV2 samples.
TiO2 due to incorporation of “V” into the crystallographic
lattice position of TiO2.12a,31 Hence, Raman results support the
lattice doping of V in TiO2. Also, peak broadening confirms the
presence of a nanocrystalline particle, and these results are
consistent with XRD results. In the case of TV2 and Au/TV2, a
small shift in the peak at 145 cm−1 is observed, and it confirms
doping of “V” into the TiO2 lattice. No characteristic peaks are
Figure 7. XPS spectra recorded for (a) Ti 2p and V 2p in the inset and (b) Au 4f core levels of TiO2, TV2, and Au/TV2 catalysts.
2849
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
Figure 8. UV irradiation time dependence of photocatalytic benzene oxidation results for (a) benzene conversion and phenol selectivity with TV2
and (b) phenol yield with TV2. (c) Phenol yield observed with TV2, 1Au/TV2, and 3Au/TV2 after UV irradiation for 18 h.
attributed to two reasons, namely, the formation of large gold
clusters on TV2 due to a large amount of gold deposition
(Figure 3e). Further, a large amount of gold also hinders the
absorption of light, and hence, the extent of light utilization
becomes relatively low. This reiterates the necessity of an
optimal amount of gold.
Table 2. Photocatalytic Benzene Oxidation Measured under
Different Conditions
phenol
catalyst and irradiation
t
benzene
conversion (%)
yield
(%)
selectivity
(%)
a
conditions
(h)
TV2 with UV
6
18
24
18
18
24
24
24
18
18
18
3.0
9.0
3.0
8.2
100
91
Many control experiments were also conducted, such as
benzene oxidation without a catalyst. In the dark or without
irradiation, there is no reaction whatsoever. Irradiation for 18 h
without a catalyst induces photochemical oxidation of benzene,
and the results are surprisingly comparable to those for 3Au/
TV2. However, a large amount of side products is observed in
the case without a catalyst, indicating the possibility of an
exclusive free radical mechanism due to photochemical
conversion and homogeneous cleavage of hydrogen peroxide
decomposition. The spectral output of a UV lamp source shows
the UV C radiation between 240 and 300 nm (see Figure S1 of
the Supporting Information). When the incident photons are
filtered for λ < 309 nm, very little photochemical conversion
was observed; however, no significant change in conversion or
selectivity was observed with TV2 and Au/TV2 catalysts. This
indicates the photochemical conversion was high with high-
energy UVC radiation, even without a catalyst. In the presence
of visible light on 1Au/TV2, the activity is also poor, with a 2%
phenol yield. Indeed, this highlights that gold does not have any
role as a catalyst. Although Au nanoclusters absorb visible light
due to SPR, only electrons can be transferred to the CB of
titania, leaving the holes in gold clusters. This aspect is verified
by photoelectrochemical measurements. Chronoamperometry
measurements were taken on TV2 and 1Au/TV2 under AM1.5
irradiation (1 sun condition), and the results are shown in
Figure 9. Compared to TV2, 1Au/TV2 generates an at least 5-
fold higher current under irradiation conditions, demonstrating
the effective absorption of visible light by gold in Au/TV2.
When the light was cut off by closing the shutter, no current
was produced, demonstrating the photofunctional behavior of
Au/TV2 in visible light. A small amount of current produced by
TV2 is likely due to absorption of a small amount of UV by
titania. Because TV2 is the active part of the catalyst and no
holes are available on the same under visible light conditions,
TV2 with UV
TV2 with UV
12.7
18.0
5.0
10.8
15.9
4.5
85
1Au/TV2 with UV
3Au/TV2 with UV
TiO2 with UV
88
90
0.3
0.3
100
100
100
61
TV2 with visible
1Au/TV2 without visible
no catalyst with UV
0.4
0.4
2.0
2.0
5.5
3.3
b
1Au/TV2 with UV
3.0
2.9
97
no catalyst with UV
(λ > 300 nm)
1.2
0.6
50
a
Unless mentioned otherwise, the reaction was conducted with H2O2,
b
H2O, and CH3CN. Dissolved oxygen was removed by argon bubbling
for 30 min.
analysis of gaseous products shows hydrogen and CO2,
indicating the re-formation of organic compounds via
secondary reactions. This diverts some amount of charge
carrier to the secondary reactions described above and hence
affects the selectivity and yield of the primary product (phenol).
However, we believe it occurs mainly to maintain the charge
neutrality of the system, which prevents the accumulation of
electrons in gold clusters and TV2. Table 2 displays a set of
benzene oxidation results with different catalysts using UV and
visible light radiation. 1Au/TV2 (TV2) shows 18% (9%)
benzene conversion and 16% (8.2%) phenol yield after UV
irradiation for 18 h. A simple doubling of the catalytic activity
after gold deposition highlights the importance of Au in
selectively separating and storing electrons from excitons. The
catalyst with high concentration of Au (3Au/TV2) gives a low
benzene conversion (5%) and a low phenol yield (4.5%); these
results are reproduced with catalysts prepared in the same and
different batches. The low activity observed with 3Au/TV2 is
2850
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
which is a well-known pathway.35 Total oxidation of the
aqueous solution of benzene to CO2 through phenol with
molecular oxygen was addressed by Bui et al.,9 and this study
demonstrates the necessity of molecular oxygen and the oxygen
from water for oxidation. Nonetheless, without a peroxide
source, there is no reaction observed with dissolved oxygen
alone. Our experiments demonstrate that the co-presence of
molecular oxygen with H2O2 is necessary for high benzene
conversion as well as phenol yield. This indicates that both act
synergistically to increase the activity. Indeed, the role of
dissolved oxygen makes this reaction a total green chemistry
reaction.
3.5. Proposed Mechanism. On the basis of the photo-
catalysis experiments conducted with different catalysts and
conditions, material characterization results through different
spectroscopy and structural techniques, chronoamperometry
results under visible light, and available experimental results
such as oxidation potential of benzene, from literature, a
possible mechanism of benzene oxidation is proposed in Figure
10. In the proposed mechanism, charge carriers are generated
upon absorption of light by TiO2. The generated holes in the
VB of titania react with benzene molecule and generate
benzene radical ions; electrons are consumed by peroxide (not
shown) as well as by oxygen molecules to form hydroxyl
radicals via superoxide formation (Figure 10).35 The superoxide
formed above interacts with protons or water on the surface of
catalyst to produce hydroxyl radicals. It should be noted that no
benzene oxidation occurs without peroxide as well as without
UV irradiation, and both are required to be present together; in
addition, molecular oxygen is required for high photocatalytic
activity. The inert atmosphere created by bubbling Ar decreases
the benzene conversion to a great extent, underscoring the role
of molecular oxygen to produce hydroxyl radicals. Even though
hydrogen peroxide was present in the experiment described
above, the low benzene conversion demands the the in situ
generation of hydroxyl radical for a larger amount of benzene
conversion. The benzene radical ions react with hydroxyl
Figure 9. Chronoamperometry measurements taken at zero voltage
for TV2 and Au/TV2 with simulated sunlight with an AM1.5 filter.
the activity is poor. The highest activity observed in UV light
underscores the necessity of energetic holes on the titania
surface, which can be met only by UV radiation. The low
activity in visible light with Au/TV2 confirms the importance of
the role of holes on TV2 surfaces. TiO2 without V doping also
shows minor activity, underscoring the role of V for charge
separation.
Another important aspect of the reaction is the presence of
dissolved oxygen in the reactant solution. Argon was bubbled
through the reactant solution to remove dissolved oxygen, and
photocatalytic conversion was measured with 1Au/TV2 (Table
2). Surprisingly, the benzene conversion and phenol yield
decreased, while high selectivity was maintained. This result
underscores the role of oxygen9 in producing hydroxyl radicals,
Figure 10. Possible mechanism of the photocatalytic oxidation of benzene to phenol, along with secondary reactions, proposed on the basis of
experimental observations for Au/TV2. Primary reactions (benzene and O2 conversion) occur on TV2 surfaces and are indicated by large green
arrows; secondary reactions (CO2 and H2 formation) could occur on titania and gold surfaces. The energy gap shown for metallic gold indicates the
surface plasmon resonance states. Schottky junction at the interface of Au and titania facilitates charge carrier separation upon light absorption.
Because of the energy overlap between gold SPR states and the CB of TV2, electron transfer can occur between them, but predominantly from gold
to TV2 (indicated by the larger arrow).
2851
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
radical and form phenol, likely through the deprotonation of an
unstable intermediate. Photocatalytic re-formation of acetoni-
trile, benzene, and/or phenol (mentioned as a carbon source in
Figure 10) is the possible secondary reaction via the
consumption of some holes on titania as well as gold, as
there some amounts of CO2 and H2 were also observed in the
gas phase products after irradiation.
Although nano gold absorbs visible light through SPR, Au/TV2
is not active for benzene oxidation under visible light because of
the unavailability of holes on titania surfaces. Au and V act as
electron sinks/traps, help in charge separation, and hence
enhance the diffusion of holes on the titania surface and its
utilization for oxidation. Nano gold does not seem to be
catalytically active for benzene oxidation; rather, it helps to
achieve a high phenol yield by charge carrier separation and
increases the availability of holes on titania for oxidation.
Although nano gold can absorb visible light, much lower
photocatalytic activity was observed on Au/TV2 in visible light.
Enhancement of the emission features of Au/TV2 in PL
reiterates the predominant transfer of the electron from gold to
titania, leaving the holes in the VB of nano gold. However,
because of the presence of Schottky junctions, the transfer of
the electron from the CB of TV2 to Au can also trap the
electrons for any reduction reactions on Au surfaces. It should
be noted that the oxidation potential of holes in gold clusters is
significantly lower than that of titania, suggesting the necessity
of phenol formation in a highly facile manner exclusively on
titania. A recent feature article reports the oxidation potential of
benzene in an acetonitrile solvent to be 2.48 0.03 eV.36 This
underscores the necessity of wide band gap semiconductors
with a valence band maximum of ≥2.5 eV, which can be met
only by titania in the system presented here. Above the benzene
oxidation potential value is the possibility of oxidation
exclusively on semiconductors, rather than metallic gold,
which exhibits a significantly lower VB maximal energy (≤2
eV). Although visible light absorption does not help benzene
oxidation for the reasons given above, it helps in reducing
molecular oxygen to H2O2 as it occurs at a reduction potential
of −0.695 eV. Parallel production of hydroxyl radicals on gold
clusters and titania enhances benzene oxidation, and hence, a
high level of conversion and a high phenol yield were observed
on Au/TV2. Electrons in the CB of TiO2 and plasmon excited
state of gold could undergo three processes under the reaction
conditions. (a) It may simply undergo recombination. (b) It
could be consumed by H2O2 and/or by molecular oxygen to
produce H2O2. (c) It could be trapped by V. Doubling of the
phenol yield from 8% (TV2) to 16% (Au/TV2) underscores
the prevalence of steps b and c, which aids the reaction.
Nonetheless, we suggest time-resolved electron paramagnetic
resonance studies of the reaction to improve our understanding
and to precipitate further development of the catalyst.
ASSOCIATED CONTENT
* Supporting Information
Spectral output from the UV source (Figure S1). This material
is available free of charge via the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*Phone: +0091-20-2590 2043. Fax: +0091-20-2590 2633. E-
mail: cs.gopinath@ncl.res.in.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
P.D. thanks UGC, New Delhi, for a research fellowship. We
acknowledge the financial support from CSIR, New Delhi, to
the TAPSUN program under Project NWP0056.
REFERENCES
■
(1) Fujishima, A.; Rao, T. N.; Tryk, D. A. J. Photochem. Photobiol., C
2000, 1, 1−21.
(2) Fujihira, M.; Satoh, Y.; Osa, T. Nature 1981, 293, 206−208.
(3) Palmisano, G.; Palmisano, L. J. Adv. Synth. Catal. 2007, 349, 964−
970.
(4) (a) Ide, Y.; Matsuoka, M.; Ogawa, M. J. Am. Chem. Soc. 2010,
132, 16762−16764. (b) Ide, Y.; Nakamura, N.; Hattori, H.; Ogino, R.;
Ogawa, M.; Sadakane, M.; Sano, T. Chem. Commun. 2011, 47, 11531−
11533.
(5) (a) Zheng, Z.; Huang, B.; Qin, X.; Zhang, X.; Dai, Y.; Whangbo,
M.-H. J. Mater. Chem. 2011, 21, 9079−9087. (b) Zheng, Z.; Huang, B.;
Meng, X.; Wang, J.; Wang, S.; Lou, Z.; Wang, Z.; Qin, X.; Zhang, X.;
Dai, Y. Chem. Commun. 2013, 49, 868−870.
(6) Tanarungsun, G.; Kiatkittipong, W.; Assabumrungrat, S.; Yamada,
H.; Tagawa, T.; Praserthdam, P. J. Chem. Eng. Jpn. 2007, 40, 415−421.
(7) Chen, X.; Zhang, J.; Fu, X.; Antonietti, M.; Wang, X. J. Am. Chem.
Soc. 2009, 131, 11658−11659.
(8) Zhang, P.; Gong, Y.; Li, H.; Chen, Z.; Wang, Y. RSC Adv. 2013, 3,
5121−5126.
(9) Bui, T. D.; Kimura, A.; Ikeda, S.; Matsumura, M. J. Am. Chem. Soc.
2010, 132, 8453−8458.
(10) (a) Herron, N.; Tolman, C. A. J. Am. Chem. Soc. 1987, 109,
2837−2839. (b) Mimoun, H.; Saussine, L.; Daire, E.; Postel, M.;
Fischer, J.; Weiss, R. J. Am. Chem. Soc. 1983, 105, 3101−3110.
(11) Antonyraj, C. A.; Srinivasan, K. Catal. Surv. Asia 2013, 17, 47−
70.
4. CONCLUSION
We synthesized the UV active Ti0.98V0.02O2 (TV2) semi-
conductor by a simple solution combustion synthesis protocol
and Au/Ti0.98V0.02O2 composites by photodeposition of gold on
TV2. HRTEM and photoluminescence studies reveal the
presence of Schottky junctions and an electron transfer
mechanism between Au and titania, respectively. Charge
carrier, especially hole, mobility is enhanced due to Schottky
junctions on the disordered mesoporous structure along with
electrically interconnected and electronically integrated nano-
crystallites of Au and Ti0.98V0.02O2. Moreover, the introduction
of gold reduces the level of recombination of excited electron
and hole and increases the separation and migration of
electrons and holes to the surface of the catalyst for further
reaction. The photocatalytic activity of Au/TV2 was studied by
oxidation of benzene to phenol under UV irradiation. The
introduction of vanadium and gold enhances the conversion of
benzene and selectivity of phenol; 18% benzene conversion and
88.1 and 15.9% phenol selectivity and yield, respectively, in the
presence of UV light, were obtained at ambient temperatures.
(12) (a) Sivaranjani, K.; Verma, A.; Gopinath, C. S. Green Chem.
2012, 14, 461−471. (b) Negi, S. S.; Sivaranjani, K.; Singh, A. P.;
Gopinath, C. S. Appl. Catal., A 2013, 452, 132−138. (c) Shiju, N. R.;
Anilkumar, M.; Mirajkar, S. P.; Gopinath, C. S.; Satyanarayana, C. V.;
Rao, B. S. J. Catal. 2005, 230, 484−492.
(13) Cavani, F.; Cortelli, C.; Frattini, A.; Panzacchi, B.; Ravaglia, V.;
Trifiro, F.; Fumagalli, C.; Leanza, R.; Mazzoni, G. Catal. Today 2006,
118, 298−306.
(14) Reddy, B. M.; Rao, K. N.; Reddy, G. K.; Bharali, P. J. Mol. Catal.
A: Chem. 2006, 253, 44−51.
(15) Nielsen, U. G.; Topsoe, N. Y.; Brorson, M.; Skibsted, J.;
Jakobsen, H. J. J. Am. Chem. Soc. 2004, 126, 4926−4933.
2852
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853

ACS Catalysis
Research Article
(16) Oganowski, W.; Hanuza, J.; Drulis, H.; Mista, W.; Macalik, L.
Appl. Catal., A 1996, 136, 143−159.
(17) Kumar, C. P.; Reddy, K. R.; Rao, V. V.; Chary, K. V. R. Green
Chem. 2002, 4, 513−516.
(18) Ye, X.; Yue, Y.; Miao, C.; Xie, Z.; Hua, W.; Gao, Z. Green Chem.
2005, 7, 524−528.
(19) Kim, S. S.; Pauly, T. R.; Pinnavaia, T. Chem. Commun. 2000,
835−836.
(20) Zhang, Z.; Pinnavia, T. J. Am. Chem. Soc. 2002, 124, 12294−
12301.
(21) (a) Sivaranjani, K.; Agarkar, S.; Ogale, S. B.; Gopinath, C. S. J.
Phys. Chem. C 2012, 116, 2581−2587. (b) Gnanakumar, E. S.; John, J.;
Raja, T.; Gopinath, C. S. J. Nanosci. Nanotechnol. 2013, 13, 2682−
2688. (c) Mathew, T.; Sivaranjani, K.; Gnanakumar, E. S.; Yamada, Y.;
Kobayashi, T.; Gopinath, C. S. J. Mater. Chem. 2012, 22, 13484−
13493.
(22) (a) Mapa, M.; Gopinath, C. S. Chem. Mater. 2009, 21, 351−359.
(b) Mapa, M.; Thushara, K. S.; Saha, B.; Chakraborty, P.; Janet, C. M.;
Viswanath, R. P.; Nair, R. P.; Murty, K. V. G. K.; Gopinath, C. S. Chem.
Mater. 2009, 21, 2973−2979. (c) Mapa, M.; Sivaranjani, K.; Bhange,
D. S.; Saha, B.; Chakraborty, P.; Kasi Viswanath, A.; Gopinath, C. S.
Chem. Mater. 2010, 22, 565−578.
(23) (a) Sivaranjani, K.; RajaAmbal, S.; Das, T.; Roy, K.;
Bhattacharyya, S.; Gopinath, C. S. ChemCatChem 2014, 6, 522−530.
(b) Sivaranjani, K.; Gopinath, C. S. J. Mater. Chem. 2011, 21, 2639−
2647. (c) Naik, B.; Parida, K. M.; Gopinath, C. S. J. Phys. Chem. C
2010, 114, 19473−19482. (d) Satish, M.; Viswanathan, B.; Viswanath,
R. P.; Gopinath, C. S. Chem. Mater. 2005, 17, 6349−6353.
(e) Kulkarni, D. G.; Murugan, A. V.; Viswanath, A. K.; Gopinath, C.
S. J. Nanosci. Nanotechnol. 2009, 9, 371−377. (f) Rajaambal, S.; Mapa,
M.; Gopinath, C. S. Dalton Trans. 2014, DOI: 10.1039/c4dt01268b.
(24) (a) Roy, K.; Vinod, C. P.; Gopinath, C. S. J. Phys. Chem. C 2013,
117, 4717−4726. (b) Roy, K.; Gopinath, C. S. Anal. Chem. 2014, 86,
3683−3687.
(25) (a) Maity, N.; Rajamohanan, P. R.; Ganapathy, S.; Gopinath, C.
S.; Bhaduri, S.; Lahiri, G. K. J. Phys. Chem. C 2008, 112, 9428−9433.
(b) Maity, P.; Gopinath, C. S.; Bhaduri, S.; Lahiri, G. K. Green Chem.
2009, 11, 554−561.
(26) Bulushev, D. A.; Kiwi Minsker, L.; Rainone, F.; Renken, A. J.
Catal. 2002, 205, 115−122.
(27) Christodoulakis, A.; Machli, M.; Lemonidou, A. A.; Boghosian,
S. J. Catal. 2004, 222, 293−306.
(28) Pantazidis, A.; Burrows, A.; Kiely, C. J.; Mirodatos, C. J. Catal.
1998, 177, 325−330.
(29) Yu, J. C.; Yu, J.; Ho, W.; Jiang, Z.; Zhang, L. Chem. Mater. 2002,
14, 3808−3816.
(30) Fujishima, A.; Zhang, X.; Tryk, D. A. Surf. Sci. Rep. 2008, 63,
515−582.
(31) (a) Bhattacharya, K.; Varma, S.; Tripathi, A. K.; Bharadwaj, S.
R.; Tyagi, A. K. J. Phys. Chem. C 2008, 112, 19102−19112.
(b) Zeferino, R. S.; Flores, M. B.; Pal, U. J. Appl. Phys. 2011, 109,
014308.
(32) (a) Das, N.; Eckert, H.; Hu, H.; Wachs, I. E.; Walzer, J. F.;
Feher, F. J. J. Phys. Chem. 1993, 97, 8240−8243. (b) Rodella, C. B.;
Franco, R. W. A.; Magon, C. J.; Donoso, J. P.; Nunes, L. A. O.; Saeki,
M. J.; Aegerter, A.; Sargentelli, V.; Florentino, A. O. J. Sol-Gel Sci.
Technol. 2003, 25, 83−88.
(33) (a) Satish, M.; Viswanath, R. P.; Gopinath, C. S. J. Nanosci.
Nanotechnol. 2009, 9, 423−432. (b) Gopinath, C. S. J. Phys. Chem. B
2006, 110, 7079−7080.
(34) (a) Pandikumar, A.; Sivaranjani, K.; Gopinath, C. S.; Ramaraj, R.
RSC Adv. 2013, 3, 13390−13398. (b) Sunil Sekhar, A. C.; Sivaranjani,
K.; Gopinath, C. S.; Vinod, C. P. Catal. Today 2012, 198, 92−97.
(35) (a) Zhang, Y.; Zhang, N.; Tang, Z.-R.; Xu, Y.-J. Chem. Sci. 2012,
3, 2812−2822. (b) Sun, S.; Wang, W.; Zhang, L.; Zhou, L.; Yin, W.;
Shang, M. Environ. Sci. Technol. 2009, 43, 2005−2010.
(36) Merkel, P. B.; Luo, P.; Dinnocenzo, J. P.; Farid, S. J. Org. Chem.
2009, 74, 5163−5173 and references cited therein.
2853
dx.doi.org/10.1021/cs500724z | ACS Catal. 2014, 4, 2844−2853