Comparison of catalytic properties of supported metal oxides for benzene oxidation using ozone

Article abstract of DOI:10.1039/c5cy00315f

Catalytic oxidation of gaseous benzene using ozone over SiO₂-supported metal oxides (oxides of Mn, Fe, Co, Ni, and Cu) was performed at 343 K. The supported metal oxides were prepared from metal nitrates via impregnation followed by calcination at 873 K. X-ray diffraction studies revealed that aggregated metal oxides were formed on the supports. The quantities of the exposed metal oxides on SiO₂ were determined using temperature-programmed desorption of formic acid species adsorbed on the catalysts. The turnover frequencies (TOFs) for ozone decomposition and catalytic oxidation of benzene with ozone were also compared for the supported metal oxides. The TOFs for ozone decomposition with the supported metal oxides decreased in the order: Cu > Co ~ Ni > Fe > Mn. Conversely, the SiO₂-supported Mn oxide (Mn/SiO₂) catalyst exhibited the highest activity for catalytic oxidation of benzene with ozone and the highest efficiency for ozone utilization. FTIR studies revealed that benzene ring cleavage continuously proceeded on the Mn/SiO₂ catalyst, and oxygen-containing species, which were readily oxidized to CO₂ and CO in the presence of ozone, accumulated on the catalyst surface. The Fe-, Co-, Ni-, and Cu-oxides suffered from catalyst deactivation due to the build-up of by-product compounds, which was attributed to their low activity for oxidation of these substances. These results suggest that the most important steps in benzene oxidation with ozone on metal oxides are the formation and oxidation of by-product compounds. The higher activity of the Mn oxide catalyst for benzene oxidation is ascribed to the facile oxidation of these by-product compounds.

Full text of DOI:10.1039/c5cy00315f

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Comparison of catalytic properties of supported
metal oxides for benzene oxidation using ozone†
Cite this: Catal. Sci. Technol., 2015,
5, 3147
*
Hisahiro Einaga, Nanako Maeda and Yusuke Nagai
Catalytic oxidation of gaseous benzene using ozone over SiO2-supported metal oxides (oxides of Mn, Fe,
Co, Ni, and Cu) was performed at 343 K. The supported metal oxides were prepared from metal nitrates
via impregnation followed by calcination at 873 K. X-ray diffraction studies revealed that aggregated metal
oxides were formed on the supports. The quantities of the exposed metal oxides on SiO2 were determined
using temperature-programmed desorption of formic acid species adsorbed on the catalysts. The turnover
frequencies (TOFs) for ozone decomposition and catalytic oxidation of benzene with ozone were also
compared for the supported metal oxides. The TOFs for ozone decomposition with the supported metal
oxides decreased in the order: Cu > Co ~ Ni > Fe > Mn. Conversely, the SiO2-supported Mn oxide (Mn/
SiO2) catalyst exhibited the highest activity for catalytic oxidation of benzene with ozone and the highest
efficiency for ozone utilization. FTIR studies revealed that benzene ring cleavage continuously proceeded
on the Mn/SiO2 catalyst, and oxygen-containing species, which were readily oxidized to CO2 and CO in
the presence of ozone, accumulated on the catalyst surface. The Fe-, Co-, Ni-, and Cu-oxides suffered
from catalyst deactivation due to the build-up of by-product compounds, which was attributed to their
low activity for oxidation of these substances. These results suggest that the most important steps in ben-
zene oxidation with ozone on metal oxides are the formation and oxidation of by-product compounds.
The higher activity of the Mn oxide catalyst for benzene oxidation is ascribed to the facile oxidation of
these by-product compounds.
Received 4th March 2015,
Accepted 26th March 2015
DOI: 10.1039/c5cy00315f
www.rsc.org/catalysis
constituent metals.3 It has been well understood that in the
catalytic oxidation process on metal oxides, O2 is adsorbed
1. Introduction
Supported metal oxides are widely used in industrial pro-
cesses as heterogeneous catalysts for various types of reac-
tions, including redox,1 acid–base,2 and photocatalytic trans-
formation.3 As a result, investigation and control of their
catalytic performance are still important topics in the field of
catalysis. The textural and catalytic properties of supported
metal oxides generally depend on the acidity, basicity, and
redox properties of the metal oxides, the oxidation states of
the metals, the sizes and structures of the metal oxides, and
the interactions between the metal oxides and the supporting
materials. The relationship between textural and catalytic
properties has been clarified.
on the metal oxides. As the catalyst temperature increases,
the adsorbed O2 dissociates to generate atomic oxygen spe-
cies and the lattice oxygen atoms in the metal oxides are acti-
vated.4 All of these oxygen species participate in the oxidation
reaction. Therefore, a high temperature is required for the
generation of active oxygen species on metal oxide surfaces.
Supported metal oxide catalysts are also effective for oxida-
tion processes using ozone (catalytic ozonation process).5–13 In
this process, ozone is decomposed at low temperatures on the
metal oxide surface to afford active oxygen species that are
capable of oxidizing organic compounds.14–18 In fact, some
catalytic ozonation processes even take place at room temper-
ature and are effective for complete oxidation of volatile organic
compounds at low concentration levels under mild conditions.
For these catalytic ozonation processes, the oxides of metals in
the first transition series, such as Mn, have been mainly used
as the active species.19–22
Supported oxides of metals in the first transition series
(Mn, Fe, Co, Ni, and Cu) have been used for catalytic oxida-
tion processes due to the good redox properties of the
Department of Energy and Material Sciences, Faculty of Engineering Sciences,
Kyushu University, Kasuga, Fukuoka 816-8580, Japan.
E-mail: einaga.hisahiro.399@m.kyushu-u.ac.jp; Fax: +81 92 583 8853;
Tel: +81 92 583 7525
We have investigated the relationship between the struc-
tures and catalytic properties of supported metal oxides. Our
studies have revealed that manganese oxides supported on
SiO2 and zeolites are effective for oxidation of benzene and
cyclohexane.23–27 The effects of the structures of manganese
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5cy00315f
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3147

View Article Online
Paper
Catalysis Science & Technology
oxides on their catalytic properties have also been studied.
Both highly dispersed manganese oxides and aggregated
manganese oxides exhibit catalytic ozonation reactivity. The
former exhibited much higher activities than the latter when
the reaction rate was compared at the same Mn loading
levels because the number of metal oxides located on the cat-
alyst surface was much higher in the highly dispersed sys-
tems. Conversely, the selectivity for CO2 formation and the
efficiency for ozone utilization were only slightly affected by
the structure of the manganese oxides. This finding indicates
that one of the decisive factors controlling the catalytic prop-
erties of supported metal oxides is the dispersion of the
metal oxides, which determines their size and location.
Accordingly, to compare the inherent activity of metal oxides
for catalytic ozonation, the dispersion of the metal oxides
should be considered and the activity per active site should
be compared.
(Fuji Silysia, Q-6) with metal nitrates (Wako Pure Chem.
Ind.). The supported catalysts were dried at 373 K for 24 h
and calcined at 873 K for 5 h in air. The amount of metal
loaded on the SiO2 support was determined from the amount
of metal nitrates and SiO2 used in catalyst preparation. The
total quantity of metal on the SiO2 support was 3.64 mmol
per 1.0 g of SiO2, which corresponds to a metal loading of
20 wt% for Mn/SiO2.
2.2. Catalyst characterization
The surface area of each sample was determined from the N2
adsorption isotherm obtained at 77 K using the BET method
with a Belsorp mini (BEL Japan Co., Ltd.). X-Ray diffraction
patterns were obtained using a JEOL RINT2200 diffractome-
ter with CuKα radiation. The crystalline sizes (D) of the metal
oxide phases were determined using the Scherrer equation
(eqn (1)):
The number of active sites on a supported catalyst is gen-
erally estimated from the amount of probe molecules
chemisorbed on the active sites. Unlike supported metal cata-
lysts for which CO chemisorption techniques can be utilized,
suitable probe molecules are not available for many
supported metal oxides, and thus estimation of their active
sites is relatively difficult. For estimating the quantity of a
metal oxide on a support, methanol and formic acid have
been used as probe molecules as the mechanism for forma-
tion of adsorbed species has been well investigated with these
compounds; once adsorbed on oxide surfaces, these com-
K
 cos
D 
,
(1)
where λ, β, θ, and K are the wavelength of the incident X-ray
beam, the peak width, the angle between the beam and the
crystal surface, and a constant, respectively. In this study, a
value of 0.94 was used for the constant K.
pounds form strongly bound methoxides and formates.28–30
A
2.3. Catalytic test reactions
strong correlation has been observed between the surface area
of metal oxides and the amount of formate species formed
on their surfaces, which enables the determination of the
relative surface area. As a result, it is now possible to esti-
mate the number of exposed metal oxides on a catalyst using
formic acid as a probe molecule.
Catalytic oxidation of benzene was performed in a fixed-bed
flow reactor. Details of the reaction system were reported
previously.8–12 The reaction gases for the catalytic reactions
consisted of benzene in N2 (>99.9995%, total hydrocarbon
<1 ppm) and O2 (>99.9995%, total hydrocarbon <1 ppm) in
cylinders. Ozone was synthesized from pure O2 using an
ozone generator. Prior to benzene oxidation with ozone, the
catalyst was heated at 573 K in O2 flow (50 mL min−1) for 1 h,
and then thermostated at the reaction temperature (343 K)
using a water bath or a heating apparatus. The concentra-
tions of benzene and ozone were 150 and 2250 ppm, respec-
tively. The content of O2, which was supplied at a gas flow
rate of 500 ml min−1, in the samples was 10% and the cata-
lyst weighed 0.10 g. Gas samples were analyzed using a FTIR
spectrometer (Perkin-Elmer Spectrum One) equipped with a
2.4 m optical length gas cell (volume of 100 mL). In this sys-
tem, the homogeneous gaseous reaction of benzene with
ozone can be neglected.
In the present study, the catalytic oxidation of benzene
using ozone over SiO2-supported metal oxides (Mn, Fe, Co,
Ni, and Cu) was investigated, and their catalytic activities
were compared on the basis of their turnover frequencies
(TOFs), which is tentatively defined as the number of mole-
cules converted divided by the number of molecules on the
exposed metal oxide surface. For estimation of the number of
exposed metal oxides on the catalysts, chemisorption of
formic acid was carried out as formic acid gets strongly
adsorbed on metal oxides but not on SiO2. The benzene oxi-
dation processes on the supported metal oxides were also
investigated by FTIR spectroscopy and temperature-
programmed oxidation (TPO) studies. The obtained results
revealed the factors controlling the catalytic properties of the
metal oxide catalysts for benzene oxidation with ozone.
For the determination of product compounds on the cata-
lyst surface after reaction, extraction with methanol was
performed and the products were analyzed using GC-MS
(Shimadzu QP-5000).
TPO analyses were performed on the SiO2-supported metal
oxide catalysts used for benzene oxidation with ozone. After
catalytic ozonation of benzene was performed at 343 K, the
catalyst used was then heated to 773 K at a rate of 10 K
min−1 in a 10% O2–N2 gas mixture flowing at a rate of 500 ml
2. Experimental
2.1. Catalyst preparation
SiO2-supported metal oxides (Mn/SiO2, Fe/SiO2, Co/SiO2, Ni/
SiO2, and Cu/SiO2) were prepared via impregnation of SiO2
3148 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
min−1. The products in the effluent gases were analyzed
using the FTIR spectrometer equipped with the gas cell.
Temperature-programmed desorption (TPD) analyses for
formic acid were also conducted on the SiO2-supported metal
oxide catalysts. Prior to TPD measurement, each catalyst was
heated in O2 flow at 573 K and then in feed gas saturated
with formic acid vapor for 30 min at 343 K, which was suffi-
cient for full coverage of the metal oxide surface with formic
acid.
were used as precursors for the preparation of supported
metal oxides, aggregated manganese oxides were formed, and
their structures changed as the calcination temperature
increased.24,25 Previously, we also reported that for SiO2-
supported manganese oxide catalysts, the highest catalytic
performance was obtained when the catalyst samples were
calcined at 873 K. In the present study, therefore, catalyst cal-
cination was performed at 873 K for all of the SiO2-supported
metal oxide catalysts.
Fig. 1 shows the XRD patterns of the SiO2-supported metal
oxides prepared using an impregnation method with metal
nitrate precursors. For the Mn/SiO2 catalysts, an Mn2O3
phase was predominantly formed, as reported previously.26
Crystalline phases were also observed for the other SiO2-
supported metal oxides; Fe2O3, CuO, Co3O4, and NiO species
were observed for Fe/SiO2, Cu/SiO2, Co/SiO2, and Ni/SiO2,
respectively, along with broadened peaks of the SiO2 support.
Thus, aggregated metal oxides were formed on the SiO2 sup-
ports in all of the catalysts when metal nitrates were used as
the starting materials. No obvious peaks for metal silicate
phases were observed for the supported metal oxides. The
crystalline sizes of the metal oxides were evaluated using
X-ray line-broadening analysis and are listed in Table 1.
According to the Scherrer equation, the crystalline sizes were
in the range of 12–23 nm. The BET surface areas of all of the
supported metal oxide catalysts were nearly the same and
were slightly lower than that of the SiO2 support precalcined
at 873 K because of the reduced fraction of SiO2 in the
supported catalysts.
2.4. FTIR studies
FTIR studies of the adsorbed species formed on the
supported metal oxides during benzene oxidation with ozone
and post-treatment were conducted using an in situ FTIR cell
with KBr windows. The cell was connected to the reaction
system for catalytic ozonation of benzene. Catalyst samples
were pressed into self-supporting disks (ϕ = 20 mm) and
mounted in the FTIR cell. Prior to analysis, each sample was
heated at 573 K in O2 for 1 h. After the gases (benzene 150
ppm, O2 10%–N2 balance) were fed into the in situ FTIR cell
and gas–solid adsorption–desorption equilibrium was
achieved, ozone and benzene-containing gases were passed
through the cell. FTIR spectra were recorded using a Perkin-
Elmer Spectrum One spectrometer equipped with a TGS
detector. The spectra were collected at room temperature
with a resolution of 4 cm−1
.
2.5. Determination of turnover frequency
The dispersion of the metal oxides on the SiO2 supports
was estimated using TPD measurement of formic acid pre-
adsorbed on the SiO2-supported metal oxides. Fig. 2 shows
Turnover frequencies (TOFs) are generally defined as the
number of molecules converted per number of active sites
per second. In the present study, the number of active sites
on each catalyst was conveniently approximated to be the
number of exposed metal sites, which are considered to be
the active species. To determine the number of exposed
metal sites on the SiO2-supported catalysts, formic acid was
used as the probe molecule for chemisorption as it adsorbs
onto metal oxides to form strongly adsorbed formates while
forming few strongly bound species on SiO2. The number of
exposed metal oxide sites was determined from the number
of chemisorbed formates on the metal oxide surfaces, which
was estimated from the TPD curves of formic acid pre-
adsorbed on the supported catalysts.
Because the TOF depends on the reaction conditions and
the catalyst weight, the TOF values were obtained when ben-
zene conversion was linear with respect to the W/F value (cat-
alyst weight per gas flow rate).
3. Results and discussion
3.1. Structural analysis and estimation of the numbers of
active sites on the SiO2-supported metal oxides
The structure of supported manganese oxides and the disper-
sion of manganese oxide species have been reported to
depend strongly on preparation conditions, supporting mate-
rials, and metal precursors.21–23 When manganese nitrates
Fig. 1 XRD patterns of the SiO2-supported metal oxides.
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3149
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
Catalysis Science & Technology
Table 1 Textual properties of SiO2-supported metal oxides
Amount of gas phase products/× 10 μmol
−1a
gcatalyst
Surface
Crystalline
phase
Crystalline
size/nm
Catalyst
CO2
CO
HCOOH
Total
Dispersion/%
area/m2 g−1
Mn/SiO2
Co/SiO2
Cu/SiO2
Ni/SiO2
Fe/SiO2
SiO2
15.4
21.0
21.7
18.6
16.0
n. d.
3.2
2.9
3.3
4.8
7.9
5.7
4.2
22.8
34.4
36.6
27.1
28.4
8.3
4.5
6.4
6.8
6.2
6.4
—
388
396
364
419
391
462
Mn2O3
Co3O4
CuO
NiO
Fe2O3
—
22.9
11.9
22.8
19.9
17.2
—
10.5
11.6
3.7
4.4
2.6
aCatalysts have been used for formic acid adsorption for 30 min. The amount was determined from the amount of gas phase products TPD
profiles.
the TPD profile of the Mn/SiO2 catalyst (Fig. 2(a)). For com-
parison, the profile of the SiO2 support calcined at 873 K is
also shown. The desorption profiles of formic acid were dif-
ferent for the two samples. For the SiO2 support, formic acid
was observed in the temperature range of 373–573 K, and CO
was observed in the range of 403–773 K, whereas CO2 was
not detected in the TPD profile. In the case of the Mn/SiO2
catalyst, the profile for formic acid desorption was similar to
that for the SiO2 support, and CO formation was observed in
the same temperature range as was detected for the SiO2
support. In addition, a profile for CO2 formation was ob-
served, indicating that strongly bound species were formed
on the metal oxides and were oxidized to CO2 upon heating.
Formic acid has been reported to form weakly and
strongly bound formate species once it is adsorbed on the
surfaces of metal oxides.27 On oxides of metals in the first
transition series, formic acid is known to dissociatively
adsorb and form three types of strongly bound adsorbed spe-
cies: bidentate, monodentate, and surface-bridged species.
Chemisorption studies have revealed that the quantity of
adsorbed formic acid normalized by surface area is on aver-
age ~5–6 μmol m−2 and is independent of the type of metal
oxide. In the present study, as the temperature in the N2 flow
was increased during TPD, these adsorbed species were
desorbed and/or decomposed to CO2 and CO. Conversely,
formic acid on the SiO2 support was evolved in the low tem-
perature range because some of the adsorbed species were
only weakly bound, and CO was observed at higher tempera-
tures because of decomposition of the remaining adsorbed
species that were strongly bound to the SiO2 surface. In addi-
tion, in the TPD profile of the Mn/SiO2 catalyst, as the cata-
lyst temperature was increased further, strongly bound for-
mates on the manganese oxides were mainly decomposed to
CO2 and H2O because of the reaction of lattice oxygen atoms
with surface formates. Peaks for CO2 formation at 547 and
726 K were observed, with the former being much larger than
the latter. These results suggest that two types of strongly
adsorbed species were formed on the Mn oxide sites.
The quantities of evolved formic acid and generated CO2
and CO for each of the supported metal oxides were esti-
mated from the corresponding TPD profiles and are listed in
Table 1. Based on the considerations described above, the
quantity of formic acid was considered to represent the num-
ber of sites on the Mn oxides able to form weak bonds with
formic acid, while the quantity of CO2 generated in the tem-
perature range of 403–773 K was assumed to correspond to
the number of sites able to bond strongly to formic acid. It
can be seen from Table 1 that the number of adsorption sites
for weak binding of formic acid on the Mn/SiO2 catalyst was
comparable to or slightly larger than that on the SiO2 sup-
port, suggesting that the adsorption sites capable of forming
weak bonds with formic acid were mostly located on the SiO2
Fig. 2 TPD profiles for formic acid pre-adsorbed on (a) Mn/SiO2 and
(b) SiO2. Catalyst weight = 0.10 g, [O2] = 10% in N2, gas flow rate =
500 mL min−1
.
3150 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
support. From the quantities of CO2 and CO formed in the
TPD profiles, the quantities of adsorption sites capable of
forming strongly bound formate species were estimated to be
−1
57 and 186 μmol gcatalyst for the SiO2 and Mn/SiO2 catalysts,
respectively. Thus, the adsorption sites for strong binding of
formate species on the Mn/SiO2 catalyst were mainly located
on the manganese oxide surfaces.
For the Co/SiO2, Ni/SiO2, Cu/SiO2, and Fe/SiO2 catalysts,
similar profiles were observed; the evolution of weakly bound
formic acid was followed by oxidative decomposition of
strongly adsorbed species and formation of CO2 (ESI†). Thus,
both the sites capable of weakly binding with formic acid
and two types of adsorption sites with the ability to strongly
bind with formic acid were also present in these catalysts.
However, note that the peak temperatures for CO2 formation
were slightly different for each catalyst, reflecting the oxidiz-
ing ability of the metal oxides and/or the interactions
between the adsorbed species and the metal oxide surfaces.
The quantities of formic acid, CO, and CO2 evolved in the
TPD profiles of these catalysts are also listed in Table 1. For
all of the catalysts, the quantities of CO formed were similar
to that for the SiO2 support, while the quantities of CO2 gen-
erated over the supported metal oxides were much greater
than that for SiO2, indicating that strongly bound formates
were also located on the Co, Ni, Cu, and Fe oxides. With
respect to the evolution of formic acid, the quantities for the
Fe/SiO2 and Ni/SiO2 catalysts were nearly the same as that for
the SiO2 support, whereas the quantities for the Co/SiO2 and
Cu/SiO2 catalysts were much greater. These results suggest
that sites capable of weakly binding with formic acid were
formed on the surfaces of the Co and Cu oxides.
The numbers of exposed metal atoms on the catalyst sur-
faces were estimated from the quantities of strongly bound
surface formates on the supported metal oxides and are com-
pared in Table 1. The metal oxide dispersion, defined as the
number of exposed metal atoms divided by the total number
of metal atoms in each metal oxide, was similar for all of the
catalysts and was in the range of 4.5–6.8%. This result indi-
cates that the metal oxides on the SiO2 support were highly
aggregated, which was consistent with the results of XRD
analyses, as described above. Assuming that the metal oxide
particles were composed of crystallites, the particle sizes of
the metal oxides were found to be less than 40 nm, which
was in good agreement with the values estimated from X-ray
line-broadening analyses (12–23 nm).
Fig.
3
Time course for ozone decomposition using the SiO2-
supported metal oxides. Conditions: [O3] = 2250 ppm, [O2] = 10% in
N2, catalyst weight = 0.020 g, gas flow rate = 500 mL min−1 (GHSV =
1500 L g−1 h−1), temperature = 343 K. (○) Mn, (△) Fe, (■) Co, (□) Ni, and
(●) Cu.
range of 0.14–0.37 s−1 for the supported metal oxides and
decreased in the order of Cu > Co ~ Ni > Fe > Mn. It has
been reported that metal oxides with p-type conductivity
exhibit high activity for ozone decomposition, and negatively
charged oxygen species are formed during the reaction.
Therefore, the high activity of the transition metal oxides
Mn2O3, Fe2O3, Co3O4, and NiO for ozone decomposition can
be explained in terms of their high p-type conductivities.
Note that Cu oxide exhibited the highest activity for ozone
decomposition even though unsupported Cu oxides have
n-type conductivity and exhibit very low activity for ozone
decomposition. It is possible that interactions between the
Cu oxide and the SiO2 support affect the electronic state of
the metal.
Fig. 4 shows the time course for benzene oxidation with
ozone over the SiO2-supported metal oxides at 343 K. For
each catalyst, the benzene oxidation profiles differed widely.
The Mn/SiO2 catalyst exhibited steady-state activity for ben-
zene oxidation after an initial gradual decrease in benzene
conversion, whereas the other supported catalysts suffered
from severe deactivation. The catalytic activities for benzene
oxidation are summarized in Table 2. The rate of benzene
oxidation was determined for each reaction when the
conversion correlated linearly with the W/F value. The rates
of benzene oxidation decreased in the order of Mn > Co >
Ni > Cu > Fe and were much different from those of
ozone decomposition in the absence of organic compounds.
Thus, the catalytically active sites with high activity for
ozone decomposition were not effective for catalytic ozona-
tion of benzene. Conversely, while the Mn oxide catalyst
exhibited the lowest activity for ozone decomposition, it
showed the highest activity for catalytic ozonation of ben-
zene. The SiO2 support without metal loading exhibited
3.2. Catalytic benzene oxidation with ozone
Fig. 3 shows the time course for ozone decomposition using
the SiO2-supported metal oxides in the absence of an organic
substrate at 343 K. All of the supported metal oxide catalysts
exhibited steady-state activity for ozone decomposition, and
the SiO2 support precalcined at 873 K without any metal
oxides exhibited a much lower activity, indicating that ozone
decomposition mainly proceeded on the metal oxide sur-
faces. The TOF values for ozone decomposition were in the
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3151

View Article Online
Paper
Catalysis Science & Technology
participate in the oxidation of organic compounds on the
catalysts:14,15
O3 + * → O2 + O*
O* + O3 → O2 + O2*
O2* → O2 + *
(2)
(3)
(4)
(5)
O* + C6H6 → decomposition.
(*refers to catalytically active sites)
During catalytic decomposition of ozone in the absence of
an organic substrate, reaction steps (2)–(4) proceeded without
significant changes in their reaction rates. In catalytic ozona-
tion of benzene, the reaction of active oxygen species with
benzene adsorbed on the catalyst surface also proceeded
(step (5)). In this process, benzene was oxidized to form by-
product compounds, including oxygen-containing species, on
the catalysts and the build-up of these by-products on the cat-
alytically active sites resulted in catalyst deactivation (see
below).
Fig. 4 Time course for benzene oxidation with ozone over the SiO2-
supported metal oxides. [Benzene] = 150 ppm, [O3] = 2250 ppm, [O2] =
10%, temperature = 343 K. (○) Mn, (△) Fe, (■) Co, (□) Ni, and (●) Cu.
Given the reaction mechanism described by eqn (2)–(5),
the efficiency for ozone utilization during catalytic oxidation
of benzene was then considered. When one oxygen atom of
ozone reacts with an organic compound, the reaction stoichi-
ometry is expressed by the following equation (eqn (6)):
almost no activity for benzene oxidation, indicating that
metal deposition was crucial for obtaining catalytic activity. It
should be noted that the rates of ozone decomposition rap-
idly decreased in the initial period and the rates were much
lower than those in the absence of benzene (ESI†).
C6H6 + (15 − x)O3 → (6 − x)CO2 + xCO + (15 − x)O2 + 3H2O.(6)
In catalytic ozonation of benzene, CO2, CO, and formic
acid were formed as the gas phase products for all of the cat-
alysts (Table 2). However, the product distribution for ben-
zene oxidation was slightly different for each catalyst. The
mole fraction of CO2 was the highest and the formation of
formic acid was much suppressed for the Mn/SiO2 catalyst.
The formation ratios of CO2 and CO that were formed using
the Co- and Ni/SiO2 catalysts were similar to that using
the Mn/SiO2 catalyst; the mole fraction of formic acid was
much greater for the Co and Ni/SiO2 catalysts. In addition,
the Fe/SiO2 catalyst generated the highest mole fractions of
CO and formic acid among the catalysts.
For the Mn/SiO2 and Co/SiO2 catalysts, the ratio of the
quantities of decomposed ozone and reacted benzene (ozone/
benzene decomposition ratio) was 10
1, and the value
remained fairly constant during the reaction. Given that the
CO2-to-CO formation ratio was 2, the theoretical value for the
ozone/benzene decomposition ratio is approximately 13
assuming that one of the oxygen atoms in each ozone mole-
cule reacted during benzene oxidation. The fact that the
experimental value for the ozone/benzene decomposition
ratio was less than the theoretical value indicates that ozone
was efficiently consumed during benzene oxidation.
It has been reported that ozone is decomposed on Mn
oxide catalysts to form oxygen species (eqn (2)–(5)) that
It is likely that O2 was involved in the autoxidation pro-
cesses in which the radical intermediates formed in benzene
Table 2 Activities for benzene oxidation with ozone over SiO2-supported metal oxide catalystsa
Ozone decompositiona
Rate/× 10−1 mol g−1 h−1
Benzene oxidationb
Rate/× 10−4 mol g−1 h−1
Product selectivityd /%
O3/C6H6
ratioc
Catalyst
TOFb /× 10−1 s−1
TOF/× 10−3 s−1
CO2
CO
HCOOH
Mn/SiO2
Co/SiO2
Cu/SiO2
Ni/SiO2
Fe/SiO2
SiO2
0.77
1.95
2.67
2.02
1.66
0.12
1.41
2.80
3.67
2.99
2.34
—
19.6
4.9
1.6
1.6
0.7
—
4.21
0.75
0.23
0.25
0.10
—
10
10
23
35
44
—
66
57
48
59
36
—
32
29
31
28
34
—
2
14
21
13
30
—
a
Conditions: [O3] = 2250 ppm, [O2] = 10%, catalyst weight = 0.020 g, gas flow rate = 500 mL min−1 (GHSV = 150 L g−1 h−1). b Conditions: ijC6H6] =
150 ppm, [O3] = 2250 ppm, [O2] = 10%, catalyst weight = 0.10 g, gas flow rate = 500 mL min−1 (GHSV = 30 L g−1 h−1). Data were taken at 160 min.
c
d
The ratio was determined from the amount of ozone consumed divided by that of benzene consumed. Product selectivity was defined as the
mole fraction of gas phase products.
3152 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
oxidation are oxidized by O2, as in the case of alumina-
supported manganese oxides.20 Conversely, for the Cu/SiO2,
Fe/SiO2, and Ni/SiO2 catalysts, benzene oxidation activity
greatly decreased with time-on-stream and the ozone/benzene
decomposition ratio increased from 15 to 23–44, indicating
that self-decomposition of ozone to O2 also proceeded along
with the decrease in benzene oxidation activity.
of the catalysts after they were used for benzene oxidation
with ozone at 343 K. As seen in Fig. 5, weakly bound formic
acid and CO were generated in the temperature range of 343–
543 K on the Mn/SiO2 catalyst, while more strongly adsorbed
oxygen-containing species, such as adsorbed formates and
carboxylates, were oxidized to CO2 and CO in the higher tem-
perature range (~773 K). Benzene desorption was not
observed in the TPO profiles.
3.3. TPO of the by-product compounds on the catalyst surfaces
A similar behavior was observed for the other metal oxide
catalysts; formic acid and CO were observed to form in the
low temperature range (390–610 K), whereas CO2 was formed
The quantities of by-product compounds formed on the SiO2-
supported metal oxide catalysts were determined using TPO
Fig. 5 TPO profiles of the SiO2-supported catalysts used after benzene oxidation with ozone. Catalyst weight = 0.10 g, gas flow rate = 500 mL
min−1, 10% O2–N2 balance.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3153

View Article Online
Paper
Catalysis Science & Technology
in the higher temperature range. It should be noted that the
peak temperature for CO2 formation was slightly different for
each catalyst, reflecting the difference in the oxidizing abili-
ties of the metal oxides.
catalysts during benzene oxidation with ozone. Fig. 6 shows
the FTIR spectra for oxidation of benzene with ozone at room
temperature over the Mn/SiO2 catalyst (294 K). Peaks assign-
able to aliphatic CO (1754 and 1772 cm−1) and C–O (1275–
1480 cm−1) stretching frequencies were observed immedi-
ately, and the intensities of these bands gradually increased
with time-on-stream (Fig. 6(a)). This result indicates that ben-
zene ring cleavage proceeded without an induction period via
reaction of the benzene ring with the oxygen species formed
during ozone decomposition. In a separate experiment, GC-
MS analysis of the catalyst surface after benzene oxidation
with ozone clarified that carboxylic acids and acid anhydrides
were formed on the catalyst surfaces during benzene oxida-
tion with ozone.
When the catalyst sample was heated to 343 K in a feed
stream containing benzene and ozone, the intensities of the
peaks for the oxygen-containing species (CO and C–O)
decreased (Fig. 6(b)), indicating that the quantities of these
oxygen-containing species decreased due to further oxidation
of these species upon heating the catalyst. Next, the Mn/SiO2
catalyst was exposed to ozone at 343 K in the absence of ben-
zene. In this case, while the intensities of the bands for the
The quantities of the by-product compounds formed on
the SiO2-supported metal oxide catalysts were determined
using the TPO profile data in Fig. 5 and are listed in Table 3.
The carbon density, defined as the quantity of organic carbon
per number of active sites, for each of the catalysts is also
listed in Table 3. The amount of by-products adsorbed on the
catalysts decreased in the order of Co > Mn > Cu > Ni > Fe
and was not consistent with the catalytic activity for benzene
oxidation of ozone. On the Mn and Co catalysts, which
exhibited steady-state activity for benzene oxidation, signifi-
cant quantities of by-product compounds were deposited. On
the Cu/SiO2, Ni/SiO2 and Fe/SiO2 catalysts, which exhibited
much lower activities and suffered from significant deactiva-
tion during benzene oxidation, the amount of carbonaceous
materials was rather much lower than that on the Mn/SiO2
catalyst. These findings indicate that the catalytically active
sites on the Cu/SiO2, Ni/SiO2, and Fe/SiO2 catalysts were
much more easily poisoned by the formation of by-product
compounds than those on the Mn/SiO2 and Co/SiO2 catalysts.
The ratio of ozone/benzene decomposition, which is an
indicator of the efficiency for ozone utilization, remained
nearly unchanged during benzene oxidation over the Mn/SiO2
and Co/SiO2 catalysts, as described above. This finding
implies that the formation of by-product compounds also did
not affect the ozone/benzene decomposition ratio, and that
ozone was efficiently consumed during benzene oxidation
after the catalytically active sites were covered by the by-
product compounds. Conversely, the build-up of by-product
compounds on the Cu/SiO2, Ni/SiO2, and Fe/SiO2 catalysts
appeared to prevent the reaction of benzene with the oxygen
species formed during benzene oxidation (step (5)) and pro-
mote self-decomposition of ozone to molecular oxygen (O3 → O2)
on the catalyst surface.
3.4. Oxidation behavior of the by-product compounds on the
catalyst surfaces
FTIR spectroscopic studies provided information of the
chemical species formed on the SiO2-supported metal oxide
Table 3 Amount of gas phase products in TPO profilesa
Amountb /× 10 μmol gcatalyst
−1
Carbon
Catalyst
CO2
CO
HCOOH
Total
densityc
Mn/SiO2
Co/SiO2
Cu/SiO2
Ni/SiO2
Fe/SiO2
94.0
120.7
47.7
43.1
34.4
14.1
16.4
6.5
6.9
4.8
9.6
18.0
6.2
7.7
4.3
117.6
155.1
60.4
57.7
43.5
9.1
8.6
3.1
3.3
2.4
Fig. 6 FTIR spectra of the Mn/SiO2 catalyst in benzene oxidation with
ozone. Conditions: [benzene] = 150 ppm, [O3] = 2250 ppm, [O2] =
10%, gas flow rate = 500 mL min−1, 10% O2–N2 balance. (a) Spectral
changes at room temperature (interval of 2.5 min). (b) Spectra after
reaction at 343 K.
a
Catalysts have been used for benzene oxidation with ozone for 3 h.
The amount was determined from TPO profiles. The amount of
b
c
carbon normalized by that of active sites.
3154 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
oxygen-containing species decreased with time, the bands
still remained, even after prolonged feeding of ozone.
During benzene oxidation with ozone using the Ni/SiO2,
Fe/SiO2, and Co/SiO2 catalysts, peaks due to aliphatic CO
(1754 and 1772 cm−1) and C–O (1275–1480 cm−1) stretching
also appeared, indicating that benzene ring cleavage
proceeded on these catalysts as well (Fig. 7–9). The band
intensities also decreased after heating these catalysts at 343
K and upon their exposure to an ozone feed without benzene.
Conversely, during benzene oxidation with ozone over the
Cu/SiO2 catalyst, in addition to the band at 1736 cm−1 attrib-
uted to aliphatic CO groups, peaks in the range of 1550–
1700 cm−1 that were assigned to aromatic CO groups or
other types of oxygen-containing species were observed in the
IR spectrum (Fig. 10). Thus, the oxidation pathways for ben-
zene for the Ni/SiO2, Fe/SiO2, and Co/SiO2 catalysts were simi-
lar to those for the Mn/SiO2 catalyst, while those for the
Cu/SiO2 catalyst were different.
The changes in the intensities of the IR peaks in the range
of 1736–1750 cm−1 (assigned to aliphatic CO groups) indi-
cated the rates of benzene ring cleavage for the different cata-
lysts. Therefore, the band intensities assigned to CO
groups in the IR spectra of the five types of SiO2-supported
Fig. 8 FTIR spectra of the Co/SiO2 catalyst in benzene oxidation with
ozone. (a) Spectral changes at room temperature (interval of 2.5 min).
(b) Spectra after reaction at 343 K. Conditions were the same as in
Fig. 6.
metal oxide catalysts during benzene oxidation with ozone at
room temperature (294 K) are compared in Fig. 11. The rela-
tive intensities were determined using the Si–O stretching
bands of SiO2 at 1870 cm−1 (overtone). The intensity of the
CO stretching peak monotonically increased with time for
the Mn/SiO2 catalyst, while for the Ni/SiO2, Fe/SiO2, and
Cu/SiO2 catalysts, the band intensities initially increased, but
were nearly saturated after 10 min and were much lower than
the intensity of the peak for the Mn/SiO2 catalyst. These
results indicate that benzene ring cleavage proceeded only in
the initial period but then suddenly ceased on the Ni/SiO2,
Cu/SiO2, and Fe/SiO2 catalysts. The Fe oxide catalyst showed
the lowest activity for benzene ring cleavage among all of the
catalysts. Conversely, the band intensity of the CO
stretching peak in the IR spectrum of the Co/SiO2 catalyst
was higher than that of the peak in the IR spectrum of the
Mn/SiO2 catalyst during the initial period, suggesting that
the Co/SiO2 catalyst initially exhibited a higher activity for
benzene ring cleavage than the Mn/SiO2 catalyst. However,
the increase in the band intensity was much smaller after
10 min, indicating that the rate of benzene ring cleavage
decreased as the reaction proceeded and by-product com-
pounds were accumulated on the Co/SiO2 catalyst.
Fig. 7 FTIR spectra of the Fe/SiO2 catalyst in benzene oxidation with
ozone. (a) Spectral changes at room temperature (interval of 2.5 min).
(b) Spectra after reaction at 343 K. Conditions were the same as in
Fig. 6.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3155

View Article Online
Paper
Catalysis Science & Technology
Fig. 9 FTIR spectra of the Ni/SiO2 catalyst in benzene oxidation with
ozone. (a) Spectral changes at room temperature (interval of 2.5 min).
(b) Spectra after reaction at 343 K. Conditions were the same as in
Fig. 6.
Fig. 10 FTIR spectra of the Cu/SiO2 catalyst in benzene oxidation with
ozone. (a) Spectral changes at room temperature (interval of 2.5 min).
(b) Spectra after reaction at 343 K. Conditions were the same as in
Fig. 6.
The FTIR studies also provided information about the
effect of catalyst heating on the behavior of the by-product
compounds. As described above, benzene oxidation activity
was found to increase with increasing catalyst temperature,
and the increase in the catalytic activity was explained in
terms of the oxidation behavior of the by-product com-
pounds.9,23 When the catalyst temperature was raised from
294 K to 343 K during benzene oxidation with ozone, the
intensity of the IR band assigned to CO groups greatly
decreased (to 0.4 times the original) for the Mn/SiO2 catalyst,
indicating that oxidation of the oxygen-containing by-product
compounds was promoted. This result indicates that catalyst
heating during the reaction was significant for the Mn/SiO2
catalyst. In addition, feeding of ozone in the absence of ben-
zene at 343 K further promoted the oxidation processes.
However, some oxygen-containing by-products remained after
prolonged exposure to a feed of ozone.
heating is not as significant as it is for the Mn/SiO2 catalyst.
Subsequent feeding of ozone at 343 K also decreased the
intensities of the CO stretching peaks for the other
For the other metal oxide catalysts, oxidation of the by-
product compounds was also promoted by catalyst heating at
343 K. However, the decrease in the intensity of the CO
stretching peak for each of them was much less than that for
the Mn/SiO2 catalyst. This finding indicates that for these
metal oxide catalysts, the promotional effect due to catalyst
Fig. 11 Time course for changes in FTIR band intensity of the
supported metal oxides.
3156 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Fig. 12 Time course for benzene oxidation over Mn/SiO2 and Co/SiO2 catalysts. With benzene feed: [benzene] = 150 ppm, [O3] = 2250 ppm,
[O2] = 10%; without benzene feed: [O3] = 2250 ppm, [O2] = 10%. Catalyst weight = 0.10 g, gas flow rate 500 mL min−1 (GHSV = 300 L g−1 h−1),
temperature = 343 K.
catalysts, but to a lesser degree than what was observed for
the Mn/SiO2 catalyst.
The results described above indicate the important factors
affecting the activity of metal oxide catalysts for benzene oxi-
dation. Benzene oxidation proceeded on Mn, Fe, Co, Ni, and
Cu oxides in the presence of ozone, with benzene ring cleav-
age occurring rapidly, which was likely due to the high activ-
ity of the oxygen species formed during ozone decomposition
on the metal oxides. However, the oxidation behavior of the
oxygen-containing by-product compounds on the catalysts
differed. For the Mn/SiO2 catalyst, the oxygen-containing by-
products were readily oxidized to CO2 and CO in the presence
of ozone and benzene ring cleavage proceeded continuously,
even though the manganese oxide sites were covered by the
by-products. Thus, the Mn oxides exhibited high durability
for benzene oxidation, and the high catalytic activity of the
Mn/SiO2 catalyst was ascribed to facile oxidation of the
oxygen-containing by-product compounds, as shown in
Fig. 12. Although the Co/SiO2 catalyst exhibited higher perfor-
mance than the Mn/SiO2 catalyst for benzene ring cleavage in
the presence of ozone, the rate of the by-product oxidation
reaction was lower than that for the Mn/SiO2 catalyst. With
the Fe, Ni, and Cu oxides, benzene oxidation immediately
ceased after formation of the by-product compounds on the
catalysts. Thus, the high performance of the SiO2-supported
metal oxide catalysts depended not only on their activities for
benzene ring cleavage but also on their activities for oxida-
tion of the by-product compounds on their surfaces in the
presence of ozone.
As described above, the Co/SiO2 catalyst exhibited a higher
activity for benzene ring cleavage, but a lower steady-state
activity than the Mn/SiO2 catalyst. This finding urged us to
compare the oxidation behavior of the by-product com-
pounds on the Mn/SiO2 and Co/SiO2 catalysts in the presence
of an ozone feed. Fig. 12 shows the time course for benzene
oxidation with ozone over the Mn/SiO2 and Co/SiO2 catalysts.
After benzene oxidation was performed at 343 K, the reaction
was repeated with only an ozone feed. For the Mn/SiO2 cata-
lyst, CO2 was formed via oxidative decomposition of the by-
product compounds on the catalyst. The quantity of gener-
ated CO2 monotonically decreased with time-on-stream.
Thus, the oxygen-containing species on the Mn/SiO2 catalyst
were readily oxidized to CO2 with ozone. Conversely, for the
Co/SiO2 catalyst, the CO2 formation behavior was much dif-
ferent; the quantity of generated CO2 increased during the
initial period (3–10 min) and then monotonically decreased
with time. This behavior implies that the oxygen-containing
species were first transformed into other types of species and
then oxidized to CO2. Thus, oxidation of the by-product com-
pounds to CO2 on the Mn/SiO2 catalyst was much faster than
that on the Co/SiO2 catalyst. Note that the rate of CO2 forma-
tion via oxidation of the by-product compounds was also
much lower for the Cu/SiO2, Fe/SiO2, and Ni/SiO2 catalysts
than that for the Mn/SiO2 catalyst.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 3147–3158 | 3157

View Article Online
Paper
Catalysis Science & Technology
Conclusions
References
The catalytic properties of SiO2-supported metal oxides (Mn,
Fe, Co, Ni, and Cu) for benzene oxidation with ozone were
compared. The supported metal oxides were prepared using
an impregnation method with metal nitrate precursors.
Aggregated metal oxide phases were formed on the SiO2 sup-
ports after calcination at 873 K. The dispersion of the metal
oxides on the SiO2 supports was estimated by using TPD of
adsorbed formate species on the supported metal oxides,
which confirmed the formation of aggregated metal oxides
and revealed that the dispersion was not strongly affected by
the type of the metal oxide.
1 I. E. Wachs and T. Kim, Oxidation Reactions over Supported
Metal Oxide Catalysts: Molecular/electronic Structure-
activity/selectivity Relationships, in Metal oxide Catalysis, ed.
S. D. Jackson and J. S. J. Hargreaves, Wiley-VCH, Weinheim,
2009, vol. 2.
2 K. Arata, Green Chem., 2009, 11, 1719.
3 T. Kamegawa, J. Morishima, M. Matsuoka, J. M. Thomas and
M. Anpo, J. Phys. Chem. C, 2007, 111, 1076.
4 A. Bielanski and J. Haber, Catal. Rev.: Sci. Eng., 1979, 19, 1.
5 Y. Xi, C. Reed, Y.-K. Lee and S. T. Oyama, J. Phys. Chem. B,
2005, 109, 17587.
The SiO2-supported metal oxides showed steady-state
activity for ozone decomposition at 343 K, and similar TOF
values for ozone decomposition were obtained. However, the
TOFs for benzene oxidation with ozone varied significantly
with the metal oxide. The supported manganese oxide cata-
lyst exhibited steady-state activity for benzene oxidation and
a much higher activity than the other metal oxide catalysts.
Although the Co/SiO2 catalyst also exhibited steady-state
activity for benzene oxidation and ozone was efficiently con-
sumed during benzene oxidation on the Mn/SiO2 and Co/
SiO2 catalysts, the Co/SiO2 catalyst exhibited a lower activity
than the Mn/SiO2 catalyst. The Ni, Cu, and Fe oxides on SiO2
suffered from severe deactivation and self-decomposition of
ozone to molecular O2 on the catalyst surfaces.
6 C. Reed, Y. Xi and S. T. Oyama, J. Catal., 2005, 235, 378.
7 C. Reed, Y.-K. Lee and S. T. Oyama, J. Phys. Chem. B, 2006,
110, 4207.
8 S. Dimitrova, G. Ivanov and D. Mehandjiev, Appl. Catal., A,
2004, 266, 81.
9 E. Rezaei and J. Soltan, Chem. Eng. J., 2012, 198–199, 482.
10 J. H. Park, J. M. Kim, M. Jin, J. K. Jeon, S. S. Kim, S. H. Park,
S. C. Kim and Y. K. Park, Nanoscale Res. Lett., 2012, 7, 14.
11 J. H. Park, J. Jurng, G. N. Bae, S. H. Park, J. K. Jeon, S. C.
Kim, J. M. Kim and Y. K. Park, J. Nanosci. Nanotechnol.,
2012, 12, 5942.
12 C. R. Lee, J. Jurng, G. N. Bae, J. K. Jeon, S. C. Kim, J. M. Kim,
M. Jin and Y. K. Park, J. Nanosci. Nanotechnol., 2011, 11,
7303.
In situ FTIR spectroscopic studies and TPO of the by-
product compounds formed on the catalysts during benzene
oxidation with ozone revealed that, for the Mn/SiO2 catalyst,
benzene ring cleavage proceeded promptly to produce
oxygen-containing by-product compounds that covered the
catalytically active sites, but these by-products were readily
oxidized to CO2 and CO in the presence of ozone, as com-
pared with other metal oxide catalysts. For the Co/SiO2 cata-
lyst, the rate of benzene ring cleavage was higher than that
for Mn/SiO2 catalyst in the initial period. However, the oxida-
tion rate of the by-product compounds was much lower than
that with the Mn/SiO2 catalyst. For the Ni/SiO2 and Cu/SiO2
catalysts, the rate of benzene ring cleavage was similar to that
for the Mn/SiO2 catalyst, but both were rapidly and signifi-
cantly deactivated by the formation of lower amounts of by-
product compounds. The Fe/SiO2 catalyst not only exhibited
a much lower activity for benzene ring cleavage than the
Mn/SiO2 catalyst but also almost completely suppressed the
oxidation process due to a small amount of by-product com-
pounds. Thus, the formation of oxygen-containing by-product
compounds significantly affected the catalytic activities of the
SiO2-supported metal oxides for benzene oxidation with ozone.
13 M. Li, K. N. Hui, K. S. Hui, S. K. Lee, Y. R. Cho, H. Lee, W.
Zhou, S. Cho, C. Y. H. Chao and Y. Li, Appl. Catal., B, 2011,
107, 245.
14 W. Li, G. V. Gibbs and S. T. Oyama, J. Am. Chem. Soc., 1998,
120, 9041.
15 W. Li and S. T. Oyama, J. Am. Chem. Soc., 1998, 120, 9047.
16 R. Radhakrishnan, S. T. Oyama, J. G. Chen and K. Asakura,
J. Phys. Chem. B, 2001, 105, 4245.
17 S. Imamura, M. Ikebata, T. Ito and T. Ogita, Ind. Eng. Chem.
Res., 1991, 30, 217.
18 A. Naydenov, R. Stoyanova and D. Mehandjiev, J. Mol. Catal.
A: Chem., 1995, 98, 9.
19 Y. Liu, X.-S. Li, C. Shi, J.-L. Liu, A.-M. Zhu and B. W.-L. Jang,
Catal. Sci. Technol., 2014, 4, 2589.
20 H. Einaga and S. Futamura, J. Catal., 2004, 227, 304.
21 H. Einaga and S. Futamura, Appl. Catal., B, 2005, 60, 49.
22 H. Einaga, M. Harada and A. Ogata, Catal. Lett., 2004, 129,
422.
23 H. Einaga and A. Ogata, J. Hazard. Mater., 2009, 164, 1236.
24 H. Einaga and S. Futamura, Catal. Commun., 2007, 8, 557.
25 H. Einaga, Y. Teraoka and A. Ogata, Catal. Today, 2011,
164, 571.
26 H. Einaga, Y. Teraoka and A. Ogata, J. Catal., 2013, 305, 227.
27 H. Einaga, S. Yamamoto, N. Maeda and Y. Teraoka, Catal.
Today, 2015, 242, 287.
Acknowledgements
This work was partially supported by the Steel Foundation for
Environmental Protection Technology. This work was also
supported by JSPS KAKENHI (grant numbers 22560765 and
25550071).
28 I. E. Wachs and K. Routray, ACS Catal., 2012, 2, 1235.
29 D. E. Fein and I. E. Wachs, J. Catal., 2002, 210, 241.
30 I. E. Wachs, J.-M. Jehng and W. Ueda, J. Phys. Chem. B, 2005,
109, 2275.
3158 | Catal. Sci. Technol., 2015, 5, 3147–3158
This journal is © The Royal Society of Chemistry 2015