Insights into the active sites of ordered mesoporous cobalt oxide catalysts for the total oxidation of o-xylene

Article abstract of DOI:10.1016/j.jcat.2017.05.016

Cobalt oxide is a typical transition metal oxide that exhibits high catalytic activity for the total oxidation of volatile organic compounds. In this study, a reduction process in a glycerol solution was adopted to generate mesoporous CoO (meso-CoO) or CoO_x (meso-CoO_x) from mesoporous Co₃O₄ (meso-Co₃O₄). The obtained samples were rich in Co²⁺ species and exhibited high catalytic activity for o-xylene oxidation. The meso-CoO_x sample with the largest surface Co²⁺ amount performed the best: The o-xylene conversion at 240?°C was 83%, and the reaction rate over meso-CoO_x was nine times higher than that over meso-Co₃O₄. It is found that the samples with more surface Co²⁺ species possessed better oxygen activation ability, and the Co²⁺ species were the active sites that favored the formation of highly active O₂^? and O₂^2? (especially O₂^?) species.

Full text of DOI:10.1016/j.jcat.2017.05.016

Journal of Catalysis 352 (2017) 282–292
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Insights into the active sites of ordered mesoporous cobalt oxide
catalysts for the total oxidation of o-xylene
⇑
⇑
⇑
Shaohua Xie, Yuxi Liu , Jiguang Deng , Jun Yang, Xingtian Zhao, Zhuo Han, Kunfeng Zhang, Hongxing Dai
Beijing Key Laboratory for Green Catalysis and Separation, Key Laboratory of Beijing on Regional Air Pollution Control, Key Laboratory of Advanced Functional Materials,
Education Ministry of China, and Laboratory of Catalysis Chemistry and Nanoscience, Department of Chemistry and Chemical Engineering, College of Environmental and
Energy Engineering, Beijing University of Technology, Beijing 100124, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Cobalt oxide is a typical transition metal oxide that exhibits high catalytic activity for the total oxidation
of volatile organic compounds. In this study, a reduction process in a glycerol solution was adopted to
generate mesoporous CoO (meso-CoO) or CoOx (meso-CoOx) from mesoporous Co3O4 (meso-Co3O4).
The obtained samples were rich in Co2+ species and exhibited high catalytic activity for o-xylene oxida-
tion. The meso-CoOx sample with the largest surface Co2+ amount performed the best: The o-xylene con-
version at 240 °C was 83%, and the reaction rate over meso-CoOx was nine times higher than that over
meso-Co3O4. It is found that the samples with more surface Co2+ species possessed better oxygen activa-
tion ability, and the Co2+ species were the active sites that favored the formation of highly active OÀ2 and
O22À (especially OÀ2 ) species.
Received 13 February 2017
Revised 24 May 2017
Accepted 25 May 2017
Keywords:
Mesoporous transition metal oxides
Cobalt monoxide
Divalent cobalt species
O-xylene oxidation
Ó 2017 Elsevier Inc. All rights reserved.
Oxygen activation ability
1. Introduction
or SBA-16) as a hard template, in which KIT-6 is mostly adopted
due to its three-dimensionally ordered mesoporous architecture
Most volatile organic compounds (VOCs) are harmful to the
environment and human health [1,2]. A number of technologies
are available to control VOC emission, such as absorption, adsorp-
tion, incineration, photocatalytic oxidation, and catalytic oxidation
[3–5]. Catalytic oxidation has been regarded as one of the most
promising pathways for the abatement of VOCs due to its high effi-
ciency and low cost [5,6], in which the key issue is the availability
of high-performance catalysts. Therefore, it is desirable to fabricate
highly active catalysts at low temperatures, thus reducing the high
energy consumption for the removal of low-concentration VOCs.
As a typical transition metal oxide, Co3O4 attracts much atten-
tion due to its excellent reducibility and large amount of oxygen
vacancies [7–10]. Since heterogeneous catalysis takes place on
the surface of a catalyst, increasing the surface area of a material
is beneficial for improving its physicochemical properties, and
hence enhancing the catalytic performance. A facile route to
obtaining a material with a high surface area is to have the mate-
rial fabricated in a mesoporous structure. Ordered mesoporous
Co3O4 (meso-Co3O4) networks are usually generated using the
nanocasting method with mesoporous silica (e.g., KIT-6, SBA-15,
[11,12]. Previously, our group used the vacuum- or ultrasound-
assisted silica-templating strategy to successfully obtain a series
of meso-Co3O4 materials [13–16] and observed that meso-Co3O4
catalysts with high surface areas performed excellently in the oxi-
dation of CO or typical VOCs.
Co3O4 possesses a spinel-type crystal structure, in which there
are Co3+ ions on the octahedral coordination sites and Co2+ ions
on the tetrahedral coordination sites. It has been reported that
Co3O4 is highly active for CO oxidation: Its surface Co3+ ions are
the active site favorable for CO adsorption, whereas its surface
Co2+ ions with a neighboring oxygen vacancy are easily reoxidized
by gas-phase O2 molecules to form adsorbed oxygen species
[17,18]. It is generally accepted that the oxidation of an organic
compound takes place via a redox cycle of catalyst on the surface
of transition metal oxides, and the organic molecules can react
with the chemisorbed oxygen species to yield the total oxidation
products [19,20]. Recently, Gu and co-workers fabricated ordered
mesoporous CoO by adopting H2 reduction of nanocast Co3O4,
and observed that the rock-salt structured CoO showed good cat-
alytic activity for CO oxidation, which was attributed to the oxida-
tion of surface Co2+ by oxygen to Co3+ [18]. Our previous works
have also demonstrated that defective Co3O4 (CoOx) showed high
catalytic activity for methane oxidation due to its good methane
adsorption ability and improved low-temperature reducibility
⇑
Corresponding authors. Fax: +86 10 6739 1983.
E-mail addresses: yxliu@bjut.edu.cn (Y. Liu), jgdeng@bjut.edu.cn (J. Deng),
hxdai@bjut.edu.cn (H. Dai).
http://dx.doi.org/10.1016/j.jcat.2017.05.016
0021-9517/Ó 2017 Elsevier Inc. All rights reserved.

S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
283
[21]. CoO exclusively containing Co2+ ions in octahedral coordina-
tion is generally believed to be inactive for CO oxidation. However,
the surface Co2+ species are easily oxidized to the Co3+ species by
gaseous oxygen [22,23], accompanied by the formation of chemi-
sorbed oxygen species. The surface-activated oxygen species are
needed to oxidize the VOCs.
Based on these considerations, we here propose a facile reduc-
tion process of meso-Co3O4 with a glycerol aqueous solution, in
which the CoO phase could be gradually formed from the surface
to the bulk. As a typical VOC, o-xylene was selected to evaluate
the catalytic activities of the as-prepared samples. All of the cata-
lysts were characterized by a number of techniques. Effects of
the reduction process on the physicochemical properties (such as
the texture, crystal phase, surface composition, redox ability, and
oxygen activation ability) of cobalt oxides were examined. Differ-
ently from CO oxidation, the active sites for VOCs total oxidation
over cobalt oxides have seldom been discussed. Hence, the active
sites of mesoporous cobalt oxides for o-xylene total oxidation are
studied.
ous solution (50 wt.%) was pumped into the reactor at a rate of
1.0 mL/h with a syringe pump (Longer Pump LSP-01-2A) and evap-
orated at the outlet of the capillary. The samples obtained after 3
and 12 h of reduction were denoted as meso-CoOx and meso-
CoO, respectively. Afterward, the reactor was kept at 320 °C for
1 h under a nitrogen flow of 50 mL/h to remove the adsorbed spe-
cies, followed by cooling to RT under the same atmosphere. The
brown color of the meso-CoO powders and the absence of CO2
products formation over meso-CoO during the calcination of
meso-CoO suggest that there was no significant carbon accumula-
tion during the reduction process (Fig. S1). Glycerol has a high
decomposition temperature (above 290 °C); however, accumula-
tion of carbon was hardly found on the meso-CoO surface at the
adopted temperature (320 °C). From the above results, we believe
that there was no significant residual glycerol on the sample sur-
face, and it did not influence the total oxidation of o-xylene.
All of the chemicals (A.R. in purity) were purchased from Beijing
Chemical Reagents Company and used without further
purification.
2.2. Catalyst characterization
2. Experimental
Physicochemical properties of the mesoporous Co3O4, CoOx, and
CoO samples were characterized. X-ray diffraction (XRD) patterns
of the samples were obtained on a Bruker D8 Advance diffractome-
2.1. Catalyst preparation
Mesoporous silica (KIT-6) was synthesized adopting the proce-
dures described in the literature [24]. In typical routes, a low HCl
concentration (0.5 M) in an aqueous solution using tetraethoxysi-
lane (TEOS, Acros 99%) as silicon source and a mixture of Pluronic
P123 and n-butanol (Aldrich, 99%) as structure-directing agent.
Briefly, 6.0 g of Pluronic P123 was dissolved in 217 mL of deionized
water and 9.83 mL of HCl (37%) under vigorous stirring. After com-
plete dissolution, 7.41 mL of n-butanol was added. The mixture
was stirred at 35 °C for 1 h, and then 13.8 mL of TEOS was slowly
added to the homogeneous clear solution. This mixture was further
stirred at 35 °C for 24 h, followed by aging at 100 °C for 24 h under
static conditions (this process is referred to as hydrothermal treat-
ment). The solid product was filtered, washed several times with
deionized water and alcohol, and dried at 100 °C for 24 h. The final
KIT-6 template was obtained by calcining the above powders at
550 °C for 4 h in air. Ordered mesoporous Co3O4 (meso-Co3O4)
was fabricated according to the strategy reported previously [15].
The typical fabrication procedures are as follows: After 1.0 g of
KIT-6 was suspended in 50 mL of toluene, ultrasonic irradiation
(a 100 kHz ultrasonic wave produced at output power 300 W)
was applied at room temperature (RT) for 0.5 h. After irradiation,
the solution was stirred at 65 °C for 0.5 h, then 2.0 g of Co(NO3)2-
Á6H2O was added under vigorous stirring, and finally the solution
was dried at 50 °C for 4 h, obtaining pink powders. The pink pow-
ders were put into a crucible and then calcined in a muffle furnace
at a ramp of 1 °C/min from RT to 600 °C and kept at this tempera-
ture for 6 h. The silica template was removed by etching twice with
a hot (80 °C) NaOH aqueous solution (2.0 mol/L). The meso-Co3O4
sample was obtained after centrifugation, washing three times
with deionized water and ethanol, and drying at 80 °C for 24 h.
The ordered mesoporous cobalt oxides were generated by the
reduction of meso-Co3O4 with glycerol [25]. However, this method,
first reported by Schüth and co-workers, had low productivity,
thus limiting its wide application. Here, we modify this strategy.
The ordered mesoporous cobalt oxides were prepared by adopting
ter with CuKa radiation and a nickel filter (k = 0.15406 nm). Scan-
ning electron microscopic (SEM) images of the samples were
recorded on a Gemini Zeiss Supra 55 apparatus (operated at
10 kV). BET (Brunauer–Emmett–Teller) surface areas of the sam-
ples were determined via N2 adsorption at À196 °C on
a
Micromeritics ASAP 2020 analyzer, with the samples being out-
gassed at 300 °C under vacuum for 2.5 h before measurement. X-
ray photoelectron spectroscopy (XPS, VG CLAM 4 MCD analyzer)
was used to determine the binding energies (BEs) of Co2p, O1s,
and C1s of surface species, using Mg Ka (hm = 1253.6 eV) as excita-
tion source. To remove the adsorbed water and carbonate species
on the surface, the samples were pretreated under N2 (flow
rate = 20 mL/min) at 300 °C for 1 h and then cooled to RT. The
pretreated samples were degassed in the preparation chamber
(10À5 Torr) for 0.5 h and then introduced into the analysis chamber
(3 Â 10À9 Torr) for XPS spectrum recording. The C1s signal at
284.6 eV was taken as a reference for BE calibration. Raman spec-
troscopy (Horiba HR Evolution) was applied to characterize the
nature of the catalysts and their adsorbed oxygen species.
Carbon monoxide temperature-programmed reduction (CO-TPR)
experiments were carried out on a chemical adsorption analyzer
(Autochem II 2920, Micromeritics) coupled with mass spectroscopy
(MS) (Hiden QGA). Before TPR measurement, ca. 0.03 g of the sample
(40–60 mesh) was loaded into
a quartz fixed-bed U-shaped
microreactor (i.d. = 4 mm) and pretreated in a N2 flow of 30 mL/min
at 300 °C for 1 h. After being cooled to RT under the same atmo-
sphere, the pretreated sample was exposed to a flow (30 mL/min)
of 10% CO–90% Ar (v/v) mixture and heated from RT to 900 °C at a
ramp of 10 °C/min. The alteration in the CO concentration of the
effluent was monitored online by the chemical adsorption analyzer.
The reduction peak was calibrated against the complete reduction
of a known standard, powdered CuO (Aldrich, 99.995%).
Oxygen temperature-programmed desorption (O2-TPD) was
carried out on the same apparatus as that used in the CO-TPR
experiments. Prior to O2-TPD experiment, 30 mg of the sample
was preheated under a N2 flow of 30 mL/min at 300 °C for 1 h. After
the sample cooled to RT, an O2 flow of 40 mL/min was employed
for 1 h, and then a He flow of 30 mL/min was used to purge the
residual O2 in the system for 1 h. The O2 desorption signals were
recorded as the sample was heated from RT to 900 °C at a ramp
of 10 °C/min and under the same flow. The desorbed amounts of
a
modified meso-Co3O4-reduction procedure with glycerol
(Scheme 1). The reduction was carried out in a tubular furnace,
in which 1.0 g of meso-Co3O4 was loaded onto a quartz boat. Before
the powders were heated at a ramp of 10 °C/min from RT to 320 °C,
a nitrogen flow of 50 mL/min was employed for 0.5 h to remove
the O2 in the system. After the N2 flow was cut off, a glycerol aque-

284
S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
Scheme 1. Illustration of the process of Co3O4 reduction.
oxygen were determined according to an oxygen pulse experiment
monitored by a thermal conductivity detector together with MS.
In situ diffuse reflectance Fourier transform infrared spectro-
scopic (DRIFT) experiments were carried on a Nicolet 6700 FT-IR
spectrometer with a liquid-nitrogen-cooled MCT detector. Before
the experiments, 30 mg of the sample was loaded into a high-
temperature IR cell with a KBr window and preheated under a
N2 flow of 30 mL/min at 300 °C for 1 h. Subsequently, the sample
was cooled to 240 °C and kept for 0.5 h, and then the background
spectrum was recorded. Finally, an air flow (20.0 mL/min) was
passed through the IR cell, and the in situ DRIFT spectra of the sam-
ples were recorded in an air flow with different times, accumulat-
calculated under the most unfavorable conditions, and the results
reveal that the external and internal diffusion (mass transfer
resistance) was negligible under the adopted reaction conditions.
In the present work, the concentration of o-xylene was low
(1000 ppm), and the kinetic calculation was obtained at low
o-xylene conversions (<11%). The reaction rates and TOFs were also
determined under o-xylene conversions below 41%. The low
o-xylene conversions and low o-xylene concentration in o-xylene
total oxidation would be efficient in avoiding heat transfer effects.
Furthermore, by varying the flow rates and catalyst mass (at the
same SV), the same o-xylene conversions were detected over the
as-prepared catalysts for o-xylene total oxidation at temperatures
below 230 °C. Therefore, we believe that there was no significant
heat or mass transfer limitation under the adopted reaction
conditions. Furthermore, the axial dispersion effect can be
neglected at a reactor/particle diameter ratio higher than 10 and
a catalyst bed length/particle diameter ratio higher than 50 [27].
In the present work, the above two ratios were 21 and 65,
respectively.
ing 32 scans and a spectrum resolution of 4 cmÀ1
.
2.3. Catalytic activity evaluation
Catalytic activities of the samples were measured in
a
continuous-flow fixed-bed quartz tubular microreactor (i.d.
= 6.0 mm). Before activity measurement, all of the samples were
pretreated undera N2 flow of 20 mL/min at 300 °C for 1 h. To min-
imize the effect of hot spots, 50 mg of the sample (40–60 mesh)
was diluted with 0.25 g of quartz sands (40–60 mesh). The reactant
mixture was composed of 1000 ppm o-xylene + 20.0 vol% O2 + N2
(balance). The 1000-ppm o-xylene was generated by passing a N2
flow through a pure o-xylene-containing bottle that was chilled
in an isothermal bath at 0 °C. The total flow was 16.6–166.0 mL/
min, giving a space velocity (SV) of 20,000–200,000 mL/(g h).
Reactants and products were analyzed online by gas chromatogra-
phy (GC-14C, Shimadzu) with a flame ionization detector (FID) and
a thermal conductivity detector (TCD), using a stabilwax-DA col-
umn (30 m in length) and a Carboxen 1000 column (3 m in length).
The balance of carbon throughout the catalytic system was
estimated to be 99.5 1.5%. The o-xylene conversion is defined as
(cinlet–coutlet)/cinlet  100%, where the cinlet and coutlet were the inlet
and outlet o-xylene concentrations in the feed stream, respectively.
The absence of a mass transfer limitation under reaction conditions
was verified both experimentally and theoretically. The estimation
of the Damköhler number under the most unfavorable conditions
reveals that the external diffusion mass transfer resistance was
negligible. Performing a Weisz–Prater analysis is a facile way to
evaluate the absence or presence of internal mass transfer limita-
tion. According to the Weisz–Prater criterion, the dimensionless
Weisz–Prater parameter (NW–P) value is less than 0.3, which can
be considered a sufficient condition for the absence of significant
pore diffusion limitation [26]. When o-xylene was oxidized at
230 °C over the meso-CoOx sample, the NW–P value was calculated
to be 0.18, less than 0.3. Effective diffusivity was calculated
according to the Maxwell–Gilliland formula and Nusselt number
(details of calculations are presented in Supplementary Material).
The Damköhler number and Weisz–Prater parameter were
3. Results and discussion
Fig. 1 shows the XRD patterns of the samples. The low-angle
XRD pattern of Co3O4 showed well-resolved diffraction peaks due
to the (211) and (332) crystal planes as well as the shoulder peak
Cubic Co3O4
Cubic CoO
(JCPDS PDF # 74-1657)
(JCPDS PDF # 48-1719)
meso-Co3O4
meso-CoOx
meso-CoO
1
3
5 10 20 30 40 50 60 70 80
2 Theta (Deg.)
Fig. 1. XRD patterns of the cobalt oxide samples.

S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
285
due to the (220) crystal plane, which were characteristic of a highly
ordered mesoporous structure with a cubic Ia3d symmetry. The
wide-angle XRD pattern of the Co3O4 materials displayed well-
resolved reflections that could be indexed to the cubic Co3O4 phase
(JCPDS PDF# 74-1657). Similar results were reported in one of our
previous works [15]. With the extension in reduction time, the
mesoporous structure was maintained, but the diffraction peak
due to the (332) crystal plane in the low-angle region disappeared
(indicating a decrease in structural ordering). It is observed that
the XRD pattern of meso-CoOx was quite similar to that of meso-
CoO, implying that a cubic CoO phase (JCPDS PDF # 48-1719)
was formed after reduction of Co3O4 for 3 h. However, it is hard
to determine the amount of the CoO phase formed in the meso-
CoOx sample due to the detection limits of the XRD technique.
Fig. 2A illustrates the CO-TPR profiles of the samples, and their
quantitative analysis results are also shown in the figure. There
were two main steps of meso-Co3O4 reduction, whereas only one
peak with a tail was recorded for the reduction of the meso-CoOx
and meso-CoO samples. The reduction band below 300 °C was
assigned to reduction of the surface Co3+ species, whereas the
one above 300 °C was ascribed to reduction of the bulk Co3+ and
Co2+ species [28,29]. A similar reduction profile was observed over
the meso-CoOx and meso-CoO samples. The absence of the reduc-
tion signal below 300 °C suggests the presence of only Co2+ species
in the meso-CoOx and meso-CoO samples. The broad reduction
band with a tail above 300 °C implies a sequential reduction by
CO of the bulk Co3+ and Co2+ species in meso-CoOx and meso-
CoO [29]. To estimate the amounts of different oxygen species,
O2-TPD profiles were recorded, as shown in Fig. 2B. For the
meso-Co3O4 sample, the weak peaks at 170 and 370 °C were due
to the surface-adsorbed oxygen species (OWS) bound to the Co2+
and Co3+ ions in different coordination environments, whereas
the main peak at 795 °C was assignable to the desorption of lattice
oxygen from Co3O4. Differently from the meso-Co3O4 sample, a
strong desorption at 430 °C of surface-adsorbed oxygen species
(OSS) bound to the Co3+ ions (which were generated from Co2+ ions
by the formation of adsorbed oxygen species in oxygen vacancies)
was detected in the meso-CoOx and meso-CoO samples, and a
small peak at 525 or 660 °C was due to the desorption of lattice
oxygen from CoO [8,29]. Moreover, a small peak similar to desorp-
tion of lattice oxygen from Co3O4 at 725 °C was also observed in
the meso-CoOx sample, suggesting the co-presence of Co3O4 and
CoO in this sample. No apparent desorption of lattice oxygen sim-
ilar to that from Co3O4 was detected in the meso-CoO sample,
implying the absence of the Co3O4 phase in this sample. In other
words, the meso-CoOx sample was a mixture of Co3O4 and CoO,
but the sample (meso-CoO) prepared after reduction for 12 h con-
tained a pure CoO phase. The oxygen desorption amounts of the
samples are summarized in Table 1. It is obvious that the reduction
treatment of meso-Co3O4 greatly enhanced the adsorbed oxygen
desorption amounts (148.0
173.1 mol/g from meso-CoO), which were over 20 times higher
than that (7.6 mol/g) from meso-Co3O4. Based on the surface area
lmol/g from meso-CoOx and
l
l
of the sample, oxygen desorption amounts from the samples were
calculated (Table 1). One can clearly observe that the oxygen des-
orption amount increased with the extension in reduction time
(from 0.077 to 1.55, and then to 3.60 l
mol/m2), suggesting that
the deeper reduction of the sample resulted in a greater adsorbed
oxygen species amount. It should be noted that the surface oxygen
species (OSS) bonded to the Co3+ species accounted for the higher
oxygen amount desorbed from the meso-CoOx and meso-CoO sam-
ples. That is, the surface Co2+ would be easily oxidized into Co3+ by
oxygen molecules, giving rise to formation of the surface oxygen
species (OSS) bonded to the Co3+ species. Such a high concentration
of adsorbed oxygen species on the reduced samples would be ben-
eficial for the redox-involving reactions [17,18].
The porous structures and morphologies of the meso-Co3O4,
meso-CoOx, and meso-CoO samples were further examined by
the TEM technique, and the results are presented in Fig. 3. It is
clearly seen from the TEM images that a well-ordered mesoporous
structure was generated in the meso-Co3O4, meso-CoOx, and meso-
CoO samples, in good agreement with the low-angle XRD results.
This result also demonstrates that the reduction treatment of
meso-Co3O4 could lead to production of a well-ordered meso-
porous CoO. TEM images (Fig. 3c and d) of the meso-CoOx sample
were recorded at different magnifications. There were a number
of irregular entities of meso-CoOx (Fig. 3c), and each entity was
composed of ordered mesopores (Fig. 3d). As shown in the high-
resolution TEM images of meso-Co3O4 and meso-CoO, the intrapla-
nar spacings (d values) were measured to be ca. 0.46 and 0.23 nm
(Fig. 3b and f), in good consistency with those of the (111) crystal
plane of the standard Co3O4 (JCPDS PDF# 74-1657) and CoO (JCPDS
PDF # 48-1719) samples, respectively. The nitrogen adsorption–
desorption isotherm of the meso-Co3O4, meso-CoOx, or meso-CoO
sample was a type IV isotherm with a clear H1 hysteresis loop at
a relative pressure (p/p0) of 0.4–0.9 (Fig. 4A), suggesting the pres-
ence of uniform mesopores. This result was in good agreement
with those of the low-angle XRD and TEM investigations. It is
(A)
(B)
430 oC
425 oC
CO comsumption
525 oC
430 oC
meso-CoO
meso-CoOx
12.3 mmol/g
meso-CoO
440 oC
660 oC
725 oC
11.3 mmol/g
meso-CoOx
795 oC
348 oC
370 oC
170 oC
270 oC
meso-Co3O4
14.9 mmol/g
meso-Co3O4
× 10
50 150 250 350 450 550 650 750 850
50 150 250 350 450 550 650 750 850 950
Temperature (oC)
Temperature (oC)
Fig. 2. (A) CO-TPR profiles and (B) O2-TPD profiles of the cobalt oxide samples.

286
S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
Table 1
Textural properties and oxygen desorption amounts of the samples.
Sample
Textural property
Surface area (m2/g)
Oxygen desorption amount
Adsorbed oxygen species
Pore diameter (nm)
Pore volume (cm3/g)
Lattice oxygen species
mol/g)/(
mol/m2)
(
l
mol/g)/(
l
mol/m2)
(l
l
meso-Co3O4
meso-CoOx
meso-CoO
98.1
95.6
48.1
3.9 and 10.2
3.9 and 7.6
3.9 and 8.5
0.21
0.17
0.090
7.6/0.077
148.0/1.55
173.1/3.60
16.0/0.16
3.8/0.040
30.4/0.63
(b)
(c)
(a)
Co3O4 (111)
0.23 nm
5 nm
200 nm
20 nm
(f)
(d)
(e)
0.46 nm
CoO (111)
10 nm
20 nm
5 nm
Fig. 3. TEM images of (a, b) meso-Co3O4, (c, d) meso-CoOx, and (e, f) meso-CoO.
140
(B)
(A)
120
100
80
60
40
20
0
8.5 nm
3.9 nm
meso-CoO
meso-Co3O4
7.6 nm
meso-CoOx
meso-CoOx
10.2 nm
meso-Co3O4
24 30
meso-CoO
0.8
0.0
0.2
0.4
0.6
1.0
0
6
12
18
Relative pressure (p/p0)
Pore diameter (nm)
Fig. 4. (A) Nitrogen adsorption–desorption isotherms and (B) pore-size distributions of the cobalt oxide samples.

S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
287
observed that the capillary condensation step of the CoO sample at
Cubic CoO
Cubic Co3O4
meso-CoO
slightly higher relative pressures, which might be related to the
larger pores, was most likely due to the partial sintering of the
mesoporous structure. As shown in Fig. 4B, the pore sizes of the
samples were 3.9 and 7.6–10.2 nm. After reduction of the meso-
Co3O4 sample, the peak at a pore size of 3.9 nm was weakened in
intensity, whereas the pore size distribution of the larger pores
in the meso-CoOx or meso-CoO sample was widened. These results
were also reflected from the changes in their adsorption–desorp-
tion isotherms. Table 1 summarizes the textural parameters of
the samples. The BET surface areas of the meso-Co3O4, meso-
CoOx, and meso-CoO samples were 98.1, 95.6, and 48.1 m2/g,
respectively. The drop in surface area of the meso-CoO sample
might be due to collapse of the smaller pores (3.9 nm).
360
330
300
270
240
210
180
150
120
90
As a typical VOC, o-xylene was selected for evaluating the cat-
alytic activities of the samples. It can be observed clearly from
Fig. 5A that o-xylene conversion increased with the rise in temper-
ature. The meso-CoOx and meso-CoO samples performed better
than the meso-Co3O4 sample. The meso-CoOx sample showed the
best activity at 240 °C (83% o-xylene conversion), which was
around six times that over the meso-Co3O4 sample (13% o-xylene
conversion at 240 °C). It is understandable that meso-CoOx per-
formed better than meso-CoO, since the two samples possessed
similar surface composition but the latter displayed a smaller sur-
face area. To examine the catalytic stability, we carried out on-
stream o-xylene oxidation at 240 °C over the meso-CoOx sample.
It is seen from Fig. 5B that no significant loss in activity was
observed over meso-CoOx after 30 h of on-stream reaction. Accord-
ing to the literature [30], CoO could be oxidized easily to Co3O4 in
air. To investigate the actual temperature for the phase transfor-
mation of CoO to Co3O4 in air, in situ XRD experiments were car-
ried out on the meso-CoO sample. As shown in Fig. 6, no
apparent change in CoO phase was observed with the rise in tem-
perature from RT to 270 °C, but a sudden phase transformation
occurred after the temperature rose to 300 °C. These results sug-
gest that the CoO phase in the meso-CoO sample could be oxidized
into the Co3O4 phase in air above 270 °C, indicating that the CoO
phase was stable in air below 270 °C. TEM images (Fig. S2) of the
used meso-CoOx sample were recorded, in which there were no
apparent changes in the pore structure after 30 h of on-stream o-
xylene oxidation at 240 °C. XPS of the used meso-CoOx sample
was also carried out (Fig. S3). Compared with the fresh sample,
only slight decreases in surface Co2+ and adsorbed oxygen species
60
25
10
20
30
40
50
60
70
80
2 Theta (Deg.)
Fig. 6. In situ XRD patterns of meso-CoO at different treatment temperatures in air.
concentrations were observed on the used meso-CoOx sample
(after 30 h of on-stream o-xylene oxidation at 240 °C). All of these
results suggest that the meso-CoOx and meso-CoO samples were
catalytically stable below 260 °C.
Considering that the surface Co2+ species could be oxidized
easily into the Co3+ species by the gas-phase oxygen molecules
(thus generating the active oxygen species) [18,30], it is important
to determine the surface element compositions and surface species
on the samples. Fig. 7 illustrates the Co2p3/2 and O1s XPS spectra of
the samples, and their surface element compositions are also put in
the figure. From Fig. 7A, one can see the Co2p3/2 signal of each sam-
ple at binding energy (BE) = 779.7 eV, indicative of surface Co3+
presence, whereas the Co2p3/2 signal at BE = 781.7 eV, together
with the shake-up satellite at BE = 786.3 eV, was due to the surface
Co2+ species [31]. The surface Co2+/Co3+ molar ratio (1.22) of meso-
CoOx was similar to that (1.20) of meso-CoO, but much higher than
that (0.56) of meso-Co3O4. As for the O1s XPS spectra (Fig. 7B), one
can decompose the asymmetrical O1s spectrum of each sample
into three components: the surface lattice oxygen (Olatt) species
at BE = 529.7 eV, the surface-adsorbed oxygen (Oads, OÀ2 , O22À, or
100
100
(A)
(B)
90
80
60
40
20
meso-CoOx
80
70
Reaction temperature: 240 oC
60
1000 ppm o-xylene in air
SV = 40,000 mL/(g h)
SV = 40,000 mL/(g h)
0
50
180
200
220
240
260
280
0
5
10
15
20
25
30
Reaction temperature (oC)
Time on stream (h)
Fig. 5. (A) O-xylene conversion as a function of temperature over the cobalt oxide samples and (B) 30 h of on-stream o-xylene oxidation over meso-CoOx.

288
S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
(A) O 1s
Co 2p
(B)
(Olatt
)
3/2
529.7 eV
(Co3+)
(Oads
)
531.4 eV
779.7 eV
781.7 eV
Co2+/Co3+ = 1.20
Satellite
(Co2+)
533.1 eV
meso-CoO
meso-CoOx
meso-CoO
Oads/Olatt = 0.85
Co2+/Co3+ = 1.22
meso-CoOx
meso-Co3O4
Oads/Olatt = 0.84
Co2+/Co3+ = 0.56
meso-Co3O4
Oads/Olatt = 0.47
774
777
780
783
786
789
792
526
528
530
532
534
536
Binding energy (eV)
Binding energy (eV)
Fig. 7. (A) O1s and (B) Co2p3/2 XPS spectra of the cobalt oxide samples.
OÀ) species at BE = 531.4 eV, and the surface-adsorbed molecular
water at BE = 533.1 eV [15,31]. It is generally believed that the
surface-adsorbed oxygen species are active for the oxidation of
hydrocarbons at low temperatures [21,32]. It was reported that
the catalyst with a higher activation ability for O2 molecules could
show better performance for the combustion of VOCs [32,33]. As
shown in Fig. 7B, the surface Oads/Olatt molar ratio (0.84) of
meso-CoOx was almost the same as that (0.85) of meso-CoO, but
much higher than that (0.47) of meso-Co3O4. The XPS results
demonstrate that the surface element compositions and surface
species were similar on the meso-CoOx and meso-CoO samples,
but the surface Co2+ and adsorbed oxygen species concentrations
were greatly enhanced after the cobalt oxide samples were
reduced for different times. It was understandable that the surface
Co3+ species over the reduced samples were generated via the oxi-
dation of surface Co2+ species by the gas-phase oxygen molecules.
That is, a larger surface Co2+ species amount would result in forma-
tion of a larger amount of the surface-adsorbed oxygen species.
The results are in good agreement with those of O2-TPD investiga-
tions. That is, a higher surface Co2+ species concentration was more
efficient for o-xylene oxidation, suggesting that the Co2+ species
were the active site for o-xylene oxidation.
It has been reported that oxidation of VOCs obeys a reaction
mechanism of the first order with respect to VOC concentration
and the zeroeth order with respect to O2 concentration under con-
ditions of excess oxygen [15,33–35]. For example, by varying the
VOC partial pressure between 0.004 and 0.018 atm and the partial
pressure of oxygen between 0.05 and 0.20 atm over the silver-
loaded zeolite (HY and HZSM-5) catalysts for the oxidation of butyl
acetate under excess oxygen conditions (O2/butyl acetate molar
ratio = 95.5–98.0), Wong and co-workers found that the reaction
order with respect to oxygen concentration was close to zero,
whereas the reaction order with respect to VOC concentration
was close to 1 [34]. Alifanti et al. also claimed that first order with
respect to toluene concentration and zeroeth order with respect to
oxygen concentration under excessive oxygen conditions (O2/toluene
molar ratio = 117.6) were obtained over the LaCoO3/Ce1–xZrxO2
catalysts for toluene oxidation [35]. Hence, it is reasonably
assumed that the oxidation of o-xylene at an o-xylene/O2 molar
ratio of 1/200 would follow a first-order reaction mechanism with
respect to o-xylene concentration (c): r = Àkc = (ÀA exp(ÀEa/RT))c,
where r, k, A, and Ea are the reaction rate (mol/s), rate constant
(sÀ1), pre-exponential factor (sÀ1), and apparent activation energy
(kJ/mol), respectively. The k values could be calculated from the
reaction rates and reactant conversions at different SVs (e.g.,
20,000, 40,000, and 80,000 mL/(g h)) and reaction temperatures.
To avoid mass transfer limitations, the temperatures used for
kinetic calculation were lower than 230 °C, at which the reactant
conversions over the samples were below 11%. Fig. S2 shows the
Arrhenius plots for o-xylene oxidation over the samples, and their
apparent activation energies are summarized in Table 2. A highly
linear relationship of ln k versus reverse reaction temperature
was observed from the Arrhenius plots for o-xylene oxidation over
the as-prepared samples, implying that o-xylene oxidation under
the present reaction conditions followed a first-order reaction
mechanism with respect to o-xylene concentration over these
samples. The apparent activation energies over meso-CoOx,
meso-CoO, and meso-Co3O4 were 105.2 5.4, 100.1 7.3, and
132.1 7.0 kJ/mol (at 95% confidence level), respectively. The Ea
values over meso-CoOx and meso-CoO were rather close, but much
lower than that over meso-Co3O4 for o-xylene combustion. Due to
similar surface elemental compositions on meso-CoOx and meso-
CoO, o-xylene oxidation would follow similar reaction pathways
at the same active sites of the two samples, suggesting that the
Co2+ species might be the active site for the addressed reaction.
Based on the above results, we could accurately calculate the turn-
over frequencies (TOF) according to the number of surface Co2+
sites in the samples. According to the surface atomic configuration
of the (111) plane (Fig. S4), there are 21 O2À ions, 9 Co3+ ions, and 5
Co2+ ions in meso-Co3O4, and 2 O2À ions and 9 Co2+ ions in
meso-CoO and meso-CoOx. The areas of the illustrated (111) plane
in meso-Co3O4, meso-CoO, and meso-CoOx are estimated to
be 3.3923 nm2 (i.e., (1.414 Â 0.808)2 Â 0.433 Â 6), 0.8481 nm2
(i.e., (1.414 Â 0.404)2 Â 0.433 Â 6), and 0.8481 nm2 (i.e., (1.414 Â
0.404)2 Â 0.433 Â 6), respectively. The surface areas of meso-
Co3O4, meso-CoO, and meso-CoOx were 98.12 Â 1018, 48.09 Â 1018
,
and 95.56 Â 1018 nm2/g, which were equal to 1.45 Â 1020 Co2+
ions per gram (0.24 Â 10À3 mol/g), 5.10 Â 1020 Co2+ ions per
gram (0.85 Â 10À3 mol/g), and 10.14 Â 1020 Co2+ ions per gram
(1.68 Â 10À3 mol/g), respectively. The TOF = xC0/n, where x is the

S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
289
Table 2
Reaction rates, TOFs, o-xylene conversions at 240 °C, and apparent activation energies for o-xylene oxidation of the samples.
Temperature (°C)
Reaction rate ( Â 10À9 mol/(g s))
TOF ( Â 10À6 (sÀ1))
meso-CoO
meso-CoOx
meso-Co3O4
meso-CoO
meso-CoOx
meso-Co3O4
200
210
220
230
8.01
14.20
30.60
71.98
29.24
50.30
86.26
187.64
83.3
2.20
3.70
7.74
15.61
9.4
16.7
36.0
84.7
–
17.4
30.0
51.3
111.7
–
9.2
15.4
32.2
64.3
–
Conversion at 240 °C (%)
Ea (kJ/mol)
A (molecule/(g s))
50.6
13.0
105.2 5.4
100.1 7.3
132.1 7.0
–
–
–
–
–
–
7.9 Â 109
9.0 Â 109
1.8 Â 1012
conversion at a certain temperature, C0 (mol/(g s)) is the initial
o-xylene molar amount per second and per gram (since the
catalytic activities of the samples were detected in a continuous
flow), and n (mol/g) is the molar amount of the surface Co2+ ions
per gram. The o-xylene oxidation rates can be calculated according
to the activity data and mass of the catalyst. For example,
according to the activity data obtained over meso-Co3O4 at
o-xylene concentration = 1000 ppm and SV = 40,000 mL/(g h), the
C0, x, and n were 0.454 Â 10À6 mol/(g s), 3.4% at 230 °C, and
0.24 Â 10À3 mol/g, respectively; hence the TOF obtained over
meso-Co3O4 at 230 °C was 64.3 Â 10À6 sÀ1. All of the TOFs and
reaction rates of the samples are summarized in Table 2. The
o-xylene oxidation rates at different temperatures over meso-CoOx
were two and nine times higher than those over meso-CoO
and meso-Co3O4, respectively. The TOF values estimated on
the basis of the surface Co2+ site were essentially similar for
the three samples, implying that the Co2+ species were the
active sites for o-xylene oxidation. The pre-exponential factor
(9.0 Â 109 molecules/(g s)) obtained over meso-CoOx was higher
than that (7.9 Â 109 molecules/(g s)) obtained over meso-CoO;
however, the Ea values obtained over both samples for o-xylene
oxidation were similar (Table 2). This strongly indicates that the
difference in reaction rate was solely due to the variation in
number of the active sites, and the significantly higher reaction
rate over meso-CoOx could be ascribed to the richness in Co2+
active sites on the surface of this sample [17]. All of the above
results indicate that the active sites of the cobalt oxide samples
for o-xylene oxidation were identical (i.e., the Co2+ species).
To further determine the nature of the active sites in meso-CoO
for o-xylene oxidation, we evaluated the catalytic performance of
the meso-CoO and meso-Co3O4 samples pretreated under different
atmospheres. Prior to evaluation, the samples were pretreated
under different atmospheres (nitrogen, oxygen, and 10 vol.% CO
+ 90 vol.% N2) at 300 °C for 1 h. According to the results of the
in situ XRD and CO-TPR tests, one can see that the CoO phase could
be oxidized into the Co3O4 phase after pretreatment in an air flow
at 300 °C for 1 h, and the Co3+ species in the meso-CoO and Co3O4
samples could be reduced to the Co2+ species in a 10 vol.% CO flow.
That is, the Co2+ species were rich in the samples pretreated in a 10
vol.% CO flow, whereas the amount of the Co2+ species decreased in
the oxidized meso-CoO sample. As shown in Fig. 8A, similar cat-
alytic activity was observed over the meso-Co3O4 sample pre-
treated in the nitrogen and oxygen flows, but a significant
enhancement in activity (a drop of 30 °C in T50%) was detected over
the sample pretreated in a CO flow. This result indicates that a lar-
ger amount of the Co2+ species was beneficial for o-xylene oxida-
tion. From Fig. 8B, one can realize that o-xylene oxidation was
favored to take place on the Co2+ sites, since the catalytic activity
decreased with the decrease in Co2+ concentration in the meso-
CoO sample. All of these results again support the conclusion that
the Co2+ species were the active site for o-xylene oxidation.
So far, many authors have focused on the relationship between
the formation of oxygen species and the catalytic activity of a sam-
ple for VOCs oxidation [6,21,36–40]. It has normally been accepted
that surface-adsorbed oxygen species play a key role in the oxida-
tion of VOCs at low temperatures [6,21,36–38], although some
researchers proposed that the lattice oxygen in a sample was more
active for VOC oxidation than the adsorbed species [39,40]. Consid-
ering the total removal of o-xylene below 300 °C over the cobalt
oxide samples, the active oxygen species might be generated at
lower temperatures. Based on the results of O2-TPD studies, there
were two main desorptions of oxygen species from meso-Co3O4
at low temperatures, which were related to the surface OWS and
OSS species, respectively. However, there was only one desorption
of oxygen species at a lower temperature from the meso-CoOx
and meso-CoO samples, which was due to the surface OSS species.
That is, the OSS species might be the active oxygen species, since
the samples with a larger amount of the OSS species performed
much better for o-xylene oxidation.
In order to identify the nature of the surface OSS species on the
typical samples, the laser Raman technique was used. As shown in
Fig. 9, a typical spectrum of Co3O4 (A1g: 686 cmÀ1, F2g: 523 cmÀ1
,
F2g: 480 cmÀ1, Eg: 205 cmÀ1, and F2g: 145 cmÀ1) was recorded on
the meso-Co3O4 sample [9,41,42]. After reduction of meso-Co3O4,
the Raman signals due to Co3O4 were weakened, and the Raman
bands were increased in intensity over the meso-CoO sample,
implying destruction of the spinel structure. It is obviously seen
that there were three new Raman bands at 1182, 1020, and
831 cmÀ1 of the meso-CoO sample, which could be assigned to
the superoxide species (OÀ2 ) and peroxide species (O22À) at the
one-electron defect site and the peroxide species (O22À) [43,44],
respectively. These results suggest that the oxygen species
adsorbed on meso-CoO were mainly superoxide and peroxide spe-
cies. In other words, surface Co2+ species were beneficial for oxy-
gen activation to form the superoxide and/or peroxide species.
To further understand the different effects of meso-Co3O4 and
meso-CoO on oxygen activation ability under reaction conditions,
an in situ DRIFT experiment was conducted. The in situ DRIFT pro-
files versus exposure time were recorded over the samples in an air
flow at 240 °C, as shown in Fig. 10. As the exposure time increased,
there was detection of adsorbed oxygen molecules at 1260 cmÀ1
and peroxide species at 860 cmÀ1 over the meso-Co3O4 sample
(Fig. 10A), whereas four adsorbed oxygen species were detected
over the meso-CoO sample (Fig. 10B). The absorption bands at
1260, 1160–1220, 1000, and 860 cmÀ1 were related to the
adsorbed oxygen molecules (O2), superoxide species (OÀ2 ), and per-
oxide species (O22À) at the one-electron defect site, and to peroxide
species [43–45]. It is obvious that differing from the meso-Co3O4
sample, meso-CoO exhibited a superability to activate oxygen
molecules to form superoxide species at the o-xylene oxidation
temperature (240 °C), although meso-Co3O4 could also activate
oxygen into peroxide species at high temperatures. There were
adsorbed oxygen molecules on meso-Co3O4, but the superoxide
species with a higher concentration existed on meso-CoO. These
results suggest that the surface Co2+ species would favor the acti-
vation of oxygen to form adsorbed oxygen species (especially the

290
S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
100
80
60
40
20
0
100
(B) meso-CoO
(A)
meso-Co3O4
80
Nitrogen flow
Oxygen flow
10 vol% CO flow
Nitrogen flow
Oxygen flow
10 vol% CO flow
60
40
20
0
SV = 40,000 mL/(g h)
SV = 40,000 mL/(g h)
160
180
200
220
240
260
280
160
180
200
220
240
260
280
Reaction temperature (oC)
Reaction temperature (oC)
Fig. 8. O-xylene conversion as a function of temperature over (A) meso-Co3O4 and (B) meso-CoO after pretreatment under different atmospheres at 300 °C for 1 h.
through the activation of oxygen molecules detected by the laser
Raman and in situ DRIFT techniques.
In the past several years, a number of studies have focused on
686
the effect of active oxygen species on oxidation reactions. Due to
the instability of the monoxide species (OÀ), the superoxide species
(OÀ2 ) and peroxide species (O22À) at the one-electron defect site
were active in the oxidation of CO and hydrocarbons (e.g., pro-
pane), and such active oxygen species are usually formed from
523
O2 molecules on the surface oxygen vacancies of transition metal
oxides [10,43–50]. For instance, after studying the reaction mech-
anism of low-temperature CO oxidation over a TiO2-supported Au
catalyst, Liu and co-workers pointed out that the O2 species
adsorbed onto the oxygen vacancies (Ovacancy) at the Ti oxide sur-
face were OÀ2 , and these superoxide species adjacent to the Au par-
ticles were responsible for the catalytic oxidation of CO over the
Au/Ti(OH)⁄4 catalyst at RT [47]. Schubert and co-workers investi-
gated the effect of support on oxygen supply in CO oxidation over
supported Au catalysts and found that the reducible transition
metal oxides (e.g., Fe2O3) exhibited supercatalytic activity for CO
oxidation due to their ability to provide reactive oxygen species.
These authors believed that the abundant Ovacancy in the vicinity
of the gold clusters on the reducible transition metal oxide sup-
ports could provide reactive oxygen species involved in the oxida-
tion of CO [50]. Based on the results obtained in the present studies
and reported in the literature, the adsorbed oxygen species on
meso-Co3O4 came from the gas-phase oxygen molecules activated
145
205
480
meso-Co3O4
1020
1182
831
meso-CoO
300
1300
1100
900
700
500
100
Raman shift (cm−1)
Fig. 9. Raman spectra of the meso-CoO and meso-Co3O4 samples.
superoxide species), coinciding with the results of laser Raman
investigations. Since the adsorbed oxygen species were easily
mutually transformed (OÀ2 M O22À) on the surface of a catalyst
[44], it is hard to elucidate whether the peroxide species were
transformed from superoxide species or were directly generated
meso-Co3O4 in air flow at 240 oC
meso-CoO
(A)
(B)
in air flow at 240 oC
40
40
30
30
25
20
25
20
16
12
16
12
8
8
6
4
2
1
6
4
2
1
4000
3600
1600
1200
800
4000
3600
1600
1200
800
Wavenumber (cm−1)
Wavenumber (cm−1)
Fig. 10. In situ DRIFT spectra of (A) meso-Co3O4 and (B) meso-CoO in a flow of air at 240 °C for different times.

S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
291
[7] Y.F. Wang, C.B. Zhang, F.D. Liu, H. He, Well-dispersed palladium supported on
ordered mesoporous Co3O4 for catalytic oxidation of o-xylene, Appl. Catal. B
142–143 (2013) 72–79.
on the surface Ovacancy of the cobalt oxide samples. The low surface
Ovacancy reasonably explains why there was a small amount of
adsorbed oxygen species desorption from the meso-Co3O4 sample.
The ultrahigh amount of adsorbed oxygen species desorption from
meso-CoO indicates that there was a large amount of Ovacancy in
meso-CoO. The Ovacancy near the Co2+ species could pick up elec-
trons from Co2+ to form OÀ2 via the pathway Co2+ + Ovacancy + O2 ?
OÀ2 + Co3+ (accompanied by the oxidation of Co2+ to Co3+), thus acti-
vating the gas-phase O2 molecules. Therefore, the surface Co2+ spe-
cies were the main center for oxygen activation, which well
explains the existence of surface Co3+ species (XPS results) and
the strong desorption signals of surface oxygen species bound to
the Co3+ ions (O2-TPD results) on the meso-CoOx and meso-CoO
samples.
[8] C.Y. Ma, Z. Mu, J.J. Li, Y.G. Jin, J. Cheng, G.Q. Lu, Z.P. Hao, S.Z. Qiao, Mesoporous
Co3O4 and Au/Co3O4 catalysts for low-temperature oxidation of trace ethylene,
J. Am. Chem. Soc. 132 (2010) 2608–2613.
[9] B.Y. Bai, J.H. Li, Positive effects of K+ ions on three-dimensional mesoporous Ag/
Co3O4 catalyst for HCHO oxidation, ACS Catal. 4 (2014) 2753–2762.
[10] J. Haber, W. Turvk, Kinetic studies as a method to differentiate between
oxygen species involved in the oxidation of propene, J. Catal. 190 (2000) 320–
326.
[11] A. Rumplecker, F. Kleitz, E.L. Salabas, F. Schüth, Hard templating pathways for
the synthesis of nanostructured porous Co3O4, Chem. Mater. 19 (2007) 485–
496.
[12] M.B. Zheng, J. Cao, S.T. Liao, J.S. Liu, H.Q. Chen, Y. Zhao, W.J. Dai, G.B. Ji, J.M. Cao,
J. Tao, Preparation of mesoporous Co3O4 nanoparticles via solid-liquid route
and effects of calcination temperature and textural parameters on their
electrochemical capacitive behaviors, J. Phys. Chem. C 113 (2009) 3887–3894.
[13] Y.S. Xia, H.X. Dai, H.Y. Jiang, L. Zhang, Three-dimensional ordered mesoporous
cobalt oxides: highly active catalysts for the oxidation of toluene and
methanol, Catal. Commun. 11 (2010) 1171–1175.
[14] J.G. Deng, L. Zhang, H.X. Dai, Y.S. Xia, H.Y. Jiang, H. Zhang, H. He, Ultrasound-
assisted nanocasting fabrication of ordered mesoporous MnO2 and Co3O4 with
high surface areas and polycrystalline walls, J. Phys. Chem. C 114 (2010) 2694–
2700.
[15] Y.X. Liu, H.X. Dai, J.G. Deng, S.H. Xie, H.G. Yang, W. Tan, W. Han, Y. Jiang, G.S.
Guo, Mesoporous Co3O4-supported gold nanocatalysts: highly active for the
oxidation of carbon monoxide, benzene, toluene, and o-xylene, J. Catal. 309
(2014) 408–418.
[16] Z.X. Wu, J.G. Deng, Y.X. Liu, S.H. Xie, Y. Jiang, X.T. Zhao, J. Yang, H. Arandiyan, G.
S. Guo, H.X. Dai, Three-dimensionally ordered mesoporous Co3O4-supported
Au–Pd alloy nanoparticles: high-performance catalysts for methane
combustion, J. Catal. 332 (2015) 13–24.
[17] X.W. Xie, Y. Li, Z.Q. Liu, M. Haruta, W.J. Shen, Low-temperature oxidation of CO
catalysed by Co3O4 nanorods, Nature 458 (2009) 746–749.
[18] D. Gu, C.J. Jia, C. Weidenthaler, H.J. Bongard, B. Spliethoff, W. Schmidt, F.
Schüth, Highly ordered mesoporous cobalt-containing oxides: structure,
catalytic properties, and active sites in oxidation of carbon monoxide, J. Am.
Chem. Soc. 137 (2015) 11407–11418.
4. Conclusions
meso-CoOx and meso-CoO samples were prepared via reduction
of meso-Co3O4 by glycerol at different times, and exhibited high
catalytic activities for o-xylene oxidation. The meso-CoOx sample
with a surface composition similar to the meso-CoO sample and
a surface area similar to the meso-Co3O4 sample performed best
for o-xylene oxidation, giving rise to an o-xylene conversion of
83% at 240 °C and a reaction rate nine times higher than that over
the meso-Co3O4 sample. The good catalytic performance of meso-
CoOx was associated with the Co2+ species concentration, which
was identified as the active site for o-xylene oxidation. The Co2+
species were highly active for the activation of gas-phase oxygen
molecules to active oxygen species (OÀ2 and O22À), especially the
superoxide species. Our results demonstrate that the surface Co2+
species were the main center for oxygen activation. This finding
may be useful for the design of novel and high-performance cata-
lysts for oxygen-involving reactions.
[19] S. Ordóñez, L. Bello, H. Sastre, R. Rosal, F.V. Dìez, Kinetics of the deep oxidation
of benzene, toluene, n-hexane and their binary mixtures over a platinum on c-
alumina catalyst, Appl. Catal. B 38 (2002) 139–149.
[20] C. He, P. Li, J. Cheng, Z.P. Hao, Z.P. Xu, A comprehensive study of deep catalytic
oxidation of benzene, toluene, ethyl acetate, and their mixtures over Pd/ZSM-5
Catalyst: mutual effects and kinetics, Water Air Soil Pollut. 209 (2010) 365–376.
[21] S.H. Xie, Y.X. Liu, J.G. Deng, X.T. Zhao, J. Yang, K.F. Zhang, Z. Han, H.X. Dai,
Acknowledgments
Three-dimensionally
nanoparticles: highly active catalysts for methane oxidation, J. Catal. 342
(2016) 17–26.
ordered
macroporous
CeO2-supported
Pd@Co
This work was supported by the NSF of China (21377008,
21677004, and 21477005), the National High Technology Research
[22] P. Broqvist, I. Panas, H.A. Persson, A DFT study on CO oxidation over Co3O4, J.
Catal. 210 (2002) 198–206.
[23] J. Jansson, Low-temperature CO oxidation over Co3O4/Al2O3, J. Catal. 194
(2000) 55–60.
and
Development
Program
of
China
(863
Program,
2015AA034603), the Beijing Nova Program (Z141109001814106),
the NSF of Beijing Municipal Commission of Education
(KM201410005008), and Scientific Research Base Construc-
tionÀScience and Technology Creation PlatformÀNational Materi-
als Research Base Construction.
[24] T.W. Kim, F. Kleitz, B. Paul, R. Ryoo, MCM-48-like large mesoporous silicas with
tailored pore structure: facile synthesis domain in
a ternary triblock
copolymer–butanol–water system, J. Am. Chem. Soc. 127 (2005) 7601–7610.
[25] H. Tüysüz, Y. Liu, C. Weidenthaler, F. Schüth, Pseudomorphic transformation of
highly ordered mesoporous Co3O4 to CoO via reduction with glycerol, J. Am.
Chem. Soc. 130 (2008) 14108–14110.
[26] J.R. Paredes, E. Díaz, F.V. Díez, S. Ordóñez, Combustion of methane in lean
mixtures over bulk transition-metal oxides: evaluation of the activity and self-
deactivation, Energy Fuels 23 (2009) 86–93.
[27] F. Kapteijn, J.A. Moulijn, A. Tarfaoui, Catalyst characterization and mimicking
pretreatment procedures by temperature-programmed techniques, Stud. Surf.
Sci. Catal. 123 (1999) 525–541.
[28] L. Lukashuk, K. Föttinger, E. Kolar, C. Rameshan, D. Teschner, M. Hävecker, A.
Knop-Gericke, N. Yigit, H. Li, E. McDermott, M. Stöger-Pollach, G. Rupprechter,
Operando XAS and NAP-XPS studies of preferential CO oxidation on Co3O4 and
CeO2-Co3O4 catalysts, J. Catal. 344 (2016) 1–15.
[29] Y.B. Yu, T. Takei, H. Ohashi, H. He, X.L. Zhang, M. Haruta, Pretreatments of
Co3O4 at moderate temperature for CO oxidation at À80 °C, J. Catal. 267
(2009) 121–128.
[30] X.L. Xu, J.Q. Li, DFT studies on H2O adsorption and its effect on CO oxidation
over spinel Co3O4 (110) surface, Surf. Sci. 605 (2011) 1962–1967.
[31] S.H. Xie, J.G. Deng, S.M. Zang, H.G. Yang, G.S. Guo, H. Arandiyan, H.X. Dai, Au–
Pd/3DOM Co3O4: highly active and stable nanocatalysts for toluene oxidation,
J. Catal. 322 (2015) 38–48.
[32] H.Y. Chen, A. Sayari, A. Adnot, F. Larachi, Composition–activity effects of Mn–
Ce–O composites on phenol catalytic wet oxidation, Appl. Catal. B 32 (2001)
195–204.
[33] Y. Wang, H. Arandiyan, J. Scott, M. Akia, H.X. Dai, J.G. Deng, K.F. Aguey-Zinsou,
R. Amal, High performance Au–Pd supported on 3D hybrid strontium-
substituted lanthanum manganite perovskite catalyst for methane
combustion, ACS Catal. 6 (2016) 6935–6947.
[34] C.T. Wong, A.Z. Abdullah, S. Bhatia, Catalytic oxidation of butyl acetate over
silver-loaded zeolites, J. Hazard. Mater. 157 (2008) 480–489.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2017.05.016.
References
[1] A.L. Bolden, C.F. Kwiatkowski, T. Colborn, New look at BTEX: are ambient levels
a problem?, Environ Sci. Technol. 49 (2015) 5261–5276.
[2] J.Y. Zheng, M. Shao, W.W. Che, L.J. Zhang, L.J. Zhong, Y.H. Zhang, D. Streets,
Speciated VOC emission inventory and spatial patterns of ozone formation
potential in the pearl river delta, China, Environ. Sci. Technol. 43 (2009) 8580–
8586.
[3] F.I. Khan, A.K. Ghoshal, Removal of volatile organic compounds from polluted
air, J. Loss Prev. Process Ind. 13 (2000) 527–545.
[4] H.L. Chen, H.M. Lee, S.H. Chen, M.B. Chang, S.J. Yu, S.N. Li, Removal of volatile
organic compounds by single-stage and two-stage plasma catalysis systems: a
review of the performance enhancement mechanisms, current status, and
suitable applications, Environ. Sci. Technol. 43 (2009) 2216–2227.
[5] S.C. Kim, W.G. Shim, Catalytic combustion of VOCs over a series of manganese
oxide catalysts, Appl. Catal. B 98 (2010) 180–185.
[6] H.X. Dai, Environmental catalysis: a solution for the removal of atmospheric
pollutants, Sci. Bull. 60 (2015) 1708–1710.

292
S. Xie et al. / Journal of Catalysis 352 (2017) 282–292
[35] M. Alifanti, M. Florea, S. Somacescu, V.I. Parvulescu, Supported perovskites for
total oxidation of toluene, Appl. Catal. B 60 (2005) 33–39.
[36] L. Wang, C.B. Zhang, H. He, F.D. Liu, C.X. Wang, Effect of doping metals on OMS-
[43] J. Guzman, S. Carrettin, A. Corma, Spectroscopic evidence for the supply of
reactive oxygen during CO oxidation catalyzed by gold supported on
nanocrystalline CeO2, J. Am. Chem. Soc. 127 (2005) 3286–3287.
2/
c
-Al2O3 catalysts for plasma-catalytic removal of o-xylene, J. Phys. Chem. C
[44] R.Q. Long, Y.P. Huang, H.L. Wan, Surface oxygen species over cerium oxide and
their reactivities with methane and ethane by means of in situ confocal
microprobe Raman spectroscopy, J. Raman Spectrosc. 28 (1997) 29–32.
[45] M. Che, A.J. Tench, Characterization and reactivity of molecular oxygen species
on oxide surfaces, Adv. Catal. 32 (1983) 1–148.
120 (2016) 6136–6144.
[37] Y.F. Wang, C.B. Zhang, Y.B. Yu, R.L. Yue, H. He, Ordered mesoporous and bulk
Co3O4 supported Pd catalysts for catalytic oxidation of o-xylene, Catal. Today
242 (2015) 294–299.
[38] S.H. Xie, H.X. Dai, J.G. Deng, Y.X. Liu, H.G. Yang, Y. Jiang, W. Tan, A.S. Ao, G.S.
Guo, Au/3DOM Co3O4: highly active nanocatalysts for the oxidation of carbon
monoxide and toluene, Nanoscale 5 (2013) 11207–11219.
[39] H.C. Genuino, S. Dharmarathna, E.C. Njagi, M.C. Mei, S.L. Suib, Gas-phase total
oxidation of benzene, toluene, ethylbenzene, and xylenes using shape-
selective manganese oxide and copper manganese oxide catalysts, J. Phys.
Chem. C 116 (2012) 12066–12078.
[40] Y.S. Wu, M. Liu, Z.C. Ma, S.T. Xing, Effect of alkali metal promoters on natural
manganese ore catalysts for the complete catalytic oxidation of o-xylene,
Catal. Today 175 (2011) 196–201.
[41] H. Shirai, Y. Morioka, I. Nakagawa, Infrared and raman spectra and lattice
vibrations of some oxide spinels, J. Phys. Soc. Japan 51 (1982) 592–597.
[42] V.G. Hadjiev, M.N. Iliev, I.V. Vergilov, The Raman spectra of Co3O4, J. Phys. C:
Solid State Phys. 21 (1988) L199–L201.
[46] A.B. López, K. Krishna, M. Makkee, J.A. Moulijn, Active oxygen from CeO2 and
its role in catalysed soot oxidation, Catal. Lett. 99 (2005) 203–206.
[47] H.C. Liu, A.I. Kozlov, A.P. Kozlova, T. Shido, K. Asakura, Y. Iwasawa, Active
oxygen species and mechanism for low-temperature CO oxidation reaction on
a
TiO2-supported Au catalyst prepared from Au(PPh3)(NO3) and as-
precipitated titanium hydroxide, J. Catal. 185 (1999) 252–264.
[48] Q. Fu, H. Saltsburg, M.F. Stephanopoulos, Active nonmetallic Au and Pt species
on ceria-based water-gas shift catalysts, Science 301 (2003) 935–938.
´
[49] B.G. Swierkosz, Nano-Au/oxide support catalysts in oxidation reactions:
provenance of active oxygen species, Catal. Today 112 (2006) 3–7.
[50] M.M. Schubert, S. Hackenberg, A.C. Veen, M. Muhler, V. Plzak, R.J. Behm, CO
oxidation over supported gold catalysts—‘‘Inert” and ‘‘Active” support
materials and their role for the oxygen supply during reaction, J. Catal. 197
(2001) 113–122.