Pretreatments of Co3O4 at moderate temperature for CO oxidation at -80 °C

Article abstract of DOI:10.1016/j.jcat.2009.08.003

Heterogeneous catalysts that can work at ambient temperature are useful for in-door air quality control, pure gas production for fuel cells and semiconductors, gas sensing, and so forth. Deposition of gold nanoparticles on base metal oxide is known to pro

Full text of DOI:10.1016/j.jcat.2009.08.003

Journal of Catalysis 267 (2009) 121–128
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Pretreatments of Co3O4 at moderate temperature for CO oxidation at ꢀ80 °C
Yunbo Yu a,1, Takashi Takei a, Hironori Ohashi a,b, Hong He c, Xiuli Zhang c, Masatake Haruta a,
*
a Department of Applied Chemistry, Graduate School of Urban Environmental Sciences, Tokyo Metropolitan University, 1-1 Minami-osawa, Hachioji, Tokyo 192-0397, Japan
b Japan Science and Technology Agency (JST), CREST, 4-1-8 Hon-cho, Kawaguchi, Saitama 332-0012, Japan
c Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, 18 Shuangqing Road, Haidian District, Beijing 100085, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 April 2009
Revised 31 July 2009
Accepted 3 August 2009
Available online 8 September 2009
Heterogeneous catalysts that can work at ambient temperature are useful for in-door air quality control,
pure gas production for fuel cells and semiconductors, gas sensing, and so forth. Deposition of gold nano-
particles on base metal oxide is known to provide highly active catalysts for CO oxidation at temperatures
below room temperature. Here we report that some select base metal oxides such as Co3O4, MnO2, and
NiO are intrinsically active catalysts for CO oxidation below 50 °C when pretreated at moderate temper-
ature between 150 and 250 °C in a stream of non-reducing dry gases, for example, dry air, CO in air, or N2.
These metal oxides are p-type semiconductors and form surface excess oxygen adsorbed at oxygen
vacancies which are created by the above pretreatments.
Keywords:
CO oxidation
Moisture
Pretreatment
Ó 2009 Elsevier Inc. All rights reserved.
Surface oxygen vacancy
Tricobalt tetraoxide
1. Introduction
Pretreatment in oxidative or reductive atmosphere at high tem-
peratures (>500 °C) has often been used to change the structure of
Heterogeneous catalysts that can work at ambient temperature
are useful for in-door air quality control [1], purification of H2 for
fuel cells [2], pure N2 and O2 gas production from air in the semi-
conductor industry [3], gas sensing [4] and so forth. Until now, only
methanol and its decomposed derivatives such as HCHO, HCOOH,
H2, and CO can be oxidized at room temperature over noble metal
catalysts and base metal oxide catalysts [5]. Especially, CO oxida-
tion has been most extensively and intensively studied because it
is a structure sensitive reaction and the simplest reaction for
mechanistic study.
Some base metal oxides such as Co3O4, NiO, and a mixture of
MnO2 and CuO (called Hopcalite) are known to be active at room
temperature for CO oxidation [6], but not in the presence of mois-
ture. We have already reported that Co3O4 and NiO show catalytic
activity for CO oxidation below ꢀ40 °C when moisture content is
below 1.0 ppm [7]. Recently, Co3O4 nanorods calcined at 450 °C
have been found to exhibit surprisingly high activity for CO oxida-
tion at ꢀ77 °C even in a moist atmosphere [8]. This presents the
second condition that a select base metal oxide is intrinsically ac-
tive as a catalyst below room temperature when specifically active
crystalline planes are preferentially exposed at the surface.
the catalyst particles [9,10] in order to improve the catalytic activ-
ity. Pretreatment in air at moderate temperatures such as 200 °C
[11,12] has also been performed to yield clean surfaces; catalyst
reconstruction during this process does not take place substan-
tially. Sadykov and co-workers [13,14] have shown that pretreat-
ment in He at 350 °C resulted in the surface reconstruction of
Co3O4 to enhance the formation of weakly bound oxygen species
and therefore the catalytic activity. These findings have motivated
us to consider that pretreatment in neutral (neither oxidizing nor
reducing) atmospheres may be more effective for enhancing cata-
lytic activity than those in oxidizing and reducing atmospheres;
however, to date no work has been yet focused on this issue.
In this study, CO oxidation has been chosen to study the influ-
ence of moderate temperature pretreatments (6300 °C) on the cat-
alytic activity of pure base metal oxides. Particular attention was
paid to the effect of the kinds of gases, temperature, and moisture
level for the pretreatments on the catalytic activity of Co3O4. Sur-
prisingly, pretreatments in dry air and even in inert or reducing-
oxidizing dry atmospheres (N2 or CO/air) in the temperature range
of 150–250 °C dramatically enhance the catalytic activity of Co3O4,
MnO2, and NiO, when they were calcined at relatively low temper-
atures below 400 °C. CO oxidation can take place even at a temper-
ature as low as ꢀ80 °C in the case of Co3O4. This enhancement can
be ascribed to the formation of surface oxygen vacancies, which
may provide a novel guideline for the creation of base metal oxide
catalysts with high activity at low temperatures.
* Corresponding author. Fax: +81 42 677 2821.
E-mail address: haruta-masatake@center.tmu.ac.jp (M. Haruta).
Present address: Research Center for Eco-Environmental Sciences, Chinese
1
Academy of Sciences, 18 Shuangqing Road, Haidian District, Beijing 100085, China.
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.08.003

122
Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
The conversion of CO was measured at the steady state; usually
2. Experimental
2.1. Catalyst preparation
30 min after the catalyst bed temperature was settled down. The
temperature was usually lowered from room temperature by using
methanol/liquid nitrogen mixture contained in a vacuum bottle. To
maintain constant temperature (change within 1 °C), small amount
of liquid nitrogen was carefully added into the bottle for 10 min
each time, and two vacuum bottles containing the mixture were
used alternately. When CO conversion at room temperature was
below 100%, the catalyst was heated by a resistive heater until
100% CO conversion is reached. The concentrations of CO and
CO2 in the inlet and outlet streams were measured by using an
automatic sampling gas chromatograph (GC8A, Shimadzu
Corporation).
Tricobalt tetraoxide (Co3O4) was prepared by calcination at
300 °C of the precipitates obtained by adding an aqueous solution
of Co(NO3)2 to an aqueous solution of Na2CO3 (1.2 times of the stoi-
chiometric amount) [11]. Typically, cobalt nitrate hexahydrate
[Co(NO3)2ꢁ6H2O, 98.0%, Wako Pure Chemical Industries Ltd.],
17.361 g was dissolved in 500 ml of distilled water. The aqueous
solution of Co(NO3)2 was heated to 70 °C and poured into 500 ml
of aqueous Na2CO3 solution heated at 70 °C, followed by further
aging at 70 °C for 1 h. During preparation, the solution was vigor-
ously stirred. The precipitate was thoroughly washed with distilled
water and filtered. The solid obtained was dried at 100 °C over-
night in air and calcined in air at 300 °C for 4 h to transform into
Co3O4. Other metal oxides such as Mn2O3, Fe2O3, and NiO were pre-
pared in a similar manner to those for Co3O4 and by calcination in
air at 800, 400, and 400 °C for 4 h, respectively. Active MnO2 was
prepared by the oxidation of Mn(NO3)2 in aqueous solution with
KMnO4 followed by calcination in air at 300 °C [15].
2.4. Kinetic measurements
Kinetic measurements were carried out in different tempera-
ture ranges depending on the pretreatment conditions: in the
range of ꢀ23 to 23 °C after pretreatments at room temperature,
in the range of ꢀ41 to ꢀ21 °C after pretreatment at 150 °C in wet
air, and in the range of ꢀ81 to ꢀ49 °C (after operated at ꢀ90 to
ꢀ49 °C for 5 h for obtaining steady state) after pretreatment at
150 °C in dry air, CO in air, or N2. The stream containing 0.2–
1.0 vol% CO in air (excess oxygen) and the stream containing 1.0–
20 vol% O2 and 1.0 vol% CO in N2 (excess oxygen) were fed for
the measurements of rate dependency on the partial pressure of
CO and of O2, respectively. The gas hourly space velocity was chan-
ged in the range of 1.2 ꢂ 104–6.1 ꢂ 104 hꢀ1 ml/g cat in order to
keep CO conversion below 15% for realizing a differential reactor
assumption.
Titanium dioxide (P25, 50 m2/g in BET surface area), Fe3O4
(4.8 m2/g), CuO nanopowder (33 nm in diameter, 29 m2/g), and
CeO2 (166.4 m2/g) were purchased from Nippon Aerosil, Wako
Pure Chemical Industries, Sigma-aldrich, and Daiichi Kigenso Kaga-
ku Kogyo Co. Ltd., respectively.
2.2. Characterization of metal oxides
BET (Brunauer, Emmett, and Teller) surface areas were deter-
mined by nitrogen adsorption at ꢀ196 °C by using a surface area
and porosity analyzer (Micromeritics Tristar, Shimadzu Corpora-
tion, JP). The low concentration of moisture in the dry gas was
measured by a cryogenic optical dew point meter (Hycosmo II
S1200, Osaka Sanso Kogyo Ltd, JP) after flowing the gas for 2 h at
a rate of 100 ml/min, while the high concentration of moisture in
wet air was measured by a dew point analyzer (DPO-3D, Osaka
Sanso Kogyo Ltd, JP) under the same conditions.
2.5. Oxygen temperature-programmed desorption (O2-TPD) and CO
temperature-programmed reduction (CO-TPR)
O2-TPD and CO-TPR experiments were preformed at Automated
Catalyst Characterization System (Autochem 2920, Micromeritics,
USA) equipped with a mass spectrometer (QIC 20, Hiden, UK) by
using 0.20 g catalyst powder. After pretreatments at different con-
ditions, the sample was cooled down to ꢀ70 °C, and then exposed
to O2 at the same temperature. After 60 min, the feed gas was
switched to He (or 5.0 vol% CO in He for CO-TPR experiments)
for purging the system for 60 min and then the temperature was
raised at a ramp of 10 °C/min from ꢀ70 to 600 °C in a stream of
He at 30 ml/min (or 5.0 vol% CO in He for CO-TPR experiments)
and the mass signal of O2 (m/z = 32) (or mass signal of CO (m/
z = 28), O2 (m/z = 32), and CO2 (m/z = 44) for CO-TPR) was recorded
simultaneously. The purity of O2, N2, and He in cylinders was
above 99.99%. To estimate the amount of different oxygen species,
the O2-TPD profiles of the samples after different pretreatments
were deconvoluted and then the integrated peak area was
obtained.
Transmission electron microscopy (TEM) observation was car-
ried out by using a JEOL/JEM-2000FX. X-ray photoelectron spectra
(XPS) were measured on a Shimadzu/Kratos ESCA 3400 using Mg
Ka X-ray source (100 W) after pretreatment in different atmo-
spheres such as in air, N2, and CO/air for 40 min within the temper-
ature range of 100–250 °C, which is similar to the conditions of
catalytic activity tests for CO oxidation. Binding energies are refer-
enced to the C(1s) binding energy of adventitious carbon contam-
ination taken to be 284.6 eV. X-ray diffraction (XRD) analyses were
performed with a RINT-TTR III diffractometer (Rigaku Corporation,
JP) using Cu Ka radiation at 50 kV and 300 mA after different pre-
treatments which are similar to XPS measurements.
2.3. CO oxidation after different pretreatments
3. Results
The catalytic tests for CO oxidation were carried out in a fixed-
bed quartz reactor (i.d. = 5 mm), containing 0.15 g of catalyst sam-
ples. A standard reaction gas containing 1.0 vol% CO, 99 vol% air,
and 1.8 ppm H2O was fed directly from the cylinder without puri-
fication. In all the cases pretreatments were performed at desired
temperatures for 40 min in a stream of dry air (3.4 ppm H2O),
wet air (13,840–15,780 ppm H2O), 1.0 vol% CO in air, N2(1.6 ppm
H2O), 1.0 vol% CO in N2 (2.6 ppm H2O), or 5.0 vol% H2 in N2
(2.8 ppm H2O) at a flow rate of 50 min/ml. After pretreatments un-
der different conditions, the catalyst sample was cooled down to
room temperature and then the feed gas was switched to reaction
gas.
3.1. The catalytic activities of metal oxides for CO oxidation after
different pretreatments
The morphology and crystalline nature were investigated by
using TEM and XRD. The mean diameter of Co3O4 calculated from
the peak width by using the Scherrer’s equation was 11 nm. A spe-
cific surface area of 129 m2/g was obtained by BET analysis. After
pretreatments in different atmospheres at 150 °C for 40 min, CO
oxidation over Co3O4 was performed in a reaction stream of air
containing 1.0 vol% CO and 1.8 ppm moisture. The reaction curves
are shown in Fig. 1. Dry air pretreatment gives surprisingly high

Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
123
Table 2
100
80
60
40
20
0
Conversion of CO in CO oxidation at ꢀ80 °C over Co3O4 after pretreatments at
different temperatures.
Pretreatment atmosphere
Pretreatment temperature (°C)
25
50
100
150
200
250
N2 (%)
0.4
0.5
0.3
0.3
0.4
0.6
0.2
0.9
77.0
78.8
0.6
100
100
100
1.3
89.9
98.6
100
1.2
80.1
98.0
81.3
2.3
CO/air (%)
Dry air (%)
Wet air (%)
1.0
Note: Before CO oxidation measurements, the catalysts were pretreated in different
atmospheres at different temperatures for 40 min. Reaction conditions: 0.15 g
catalyst sample, 1.0 vol% CO in air, 50 ml/min.
The influence of pretreatment temperature was also studied on
CO oxidation over Co3O4 at ꢀ80 °C (Table 2). Pretreatments in dry
air, CO/air, or N2 at 150, 200, or 250 °C give surprisingly high cata-
lytic activity (CO conversion above 80%) at ꢀ80 °C, whereas pre-
treatments at room temperature and at 50 °C under all
atmospheres result in low conversion of CO, below 1.0%. At
100 °C, pretreatments are effective only in CO/air and N2, namely,
under reducing–oxidizing or inert gas atmospheres. These results
as a whole imply that the formation of oxygen vacancies can be
correlated to the genesis of low temperature catalytic activity. In
contrast, after pretreatments by wet air at room temperature to
250 °C, CO conversions at ꢀ80 °C are always below 3%, demon-
strating that the concentration of moisture in the pretreatment
gas has a great influence on the catalytic activity of Co3O4.
After pretreatment in CO/air at 150 °C, CO oxidation was per-
formed at ꢀ73 °C (Fig. 2). The CO conversion remains to be 100%
during the initial 3 h time-on-stream (TOS) and then decreased
gradually to 50% after 8 h. After dry air pretreatment at 150 °C,
100% CO conversion at 0 °C was maintained for 37 h. Interestingly,
pretreatment in N2 and in CO/air at 150 °C provides much longer
durability than air pretreatment, maintaining 100% CO conversion
at 0 °C until 70 and 98 h. Pretreatment in CO/air shows the longest
durability among the three kinds of pretreatment atmospheres,
therefore providing a simple and in situ regeneration method for
the practical use of Co3O4 catalysts for CO oxidation.
-120
-100
-80
-60
-40
-20
0
20
40
60
Temperature (oC)
Fig. 1. The influence of pretreatment atmosphere on CO oxidation over Co3O4.
Reaction conditions: 0.15 g catalyst sample, 1.0 vol% CO in air, 50 ml/min. In all the
cases pretreatment was performed at 150 °C for 40 min in dry air (M), wet air (N),
1.0 vol% CO/air (e), N2 (s), 1.0 vol% CO/N2 (h), or 5.0 vol% H2/N2 (O).
catalytic activity for CO oxidation. Even at a temperature as low as
ꢀ80 °C, 100% CO conversion was obtained. Pretreatments in the
reactant gas, 1.0 vol% CO/air and in N2 also lead to 100% CO conver-
sion at ꢀ80 °C. In contrast, after the pretreatment in 1.0 vol% CO/
N2, complete oxidation of CO occurred only above room tempera-
ture. A similar result was obtained by the pretreatment in
5.0 vol% H2/N2, indicating that the reductive pretreatments depress
the catalytic activity of Co3O4.
Pretreatment in wet air supplied by an air compressor, contain-
ing 13,840–15,780 ppm H2O, leads to 100% CO conversion at 22 °C,
which is 102 °C higher than in the case of dry air pretreatment. It
has been reported that the concentration of moisture in the feed
gas has a great influence on the catalytic activity of Co3O4 for CO
oxidation [11,16]. Our present results show that the concentration
of moisture in the pretreatment gases also greatly affects the rate
of CO oxidation. The catalytic activity after pretreatment in air,
CO/air, or N2 is much higher than that previously obtained under
similar conditions, where 100% CO conversion was achieved only
when the catalyst temperature was raised to about 60 °C [11]. It
was reported that calcination temperature has strong influence
on the catalytic activity of Co3O4 for low temperature CO oxidation
[17]. We have also studied the influence of calcination temperature
and found, as shown in Table 1, that the calcination at 300 °C gave
much higher catalytic activity for CO oxidation than the calcination
at 400 °C.
Pretreatment temperature also changes the durability of Co3O4
for CO oxidation. After pretreatment in N2 at 200 °C, 100% CO con-
version was maintained at 0 °C for 91 h, which is longer than after
the pretreatment in the same atmosphere at 150 °C. Wang et al.
[12] reported that Co3O4 calcined at 300 °C was able to maintain
100% conversion in CO oxidation longer than 8 h at 25 °C and about
4 h at ꢀ78 °C, but under dry conditions. So far, the best perfor-
mance for low temperature CO oxidation in the presence of mois-
ture (3 ppm) has been achieved by Co3O4 nanorod catalysts,
prepared by morphology-controlled synthesis using ethylene gly-
Table 1
The catalytic performance of Co3O4 and Au/Co3O4 previously reported for low temperature CO oxidation.
Catalyst
Calcination temperature
(°C)
Pretreatment
conditions
SV
Moisture in the
feed (ppm)
T50 (°C)
T100 (°C)
Ref.
(hꢀ1 ml/g cat)
a
Co3O4
400
Air, 200 °C
20,000
0.85
3
ꢀ54
[11]
50
60
6000
90
120
a
Au/Co3O4
CoOx/Al2O3 honeycomb
Co3O4
Co3O4 nanorod
Co3O4
400
827
300
450
300
300
400
Air, 200 °C
2% O2, 227 °C
Air, 200 °C
20% O2/He, 450 °C
Air, 200 °C
Wet air, 200 °C
Air, 200 °C
20,000
0.85
ꢀ76, 45 min
ꢀ43b
ꢀ78, 4 h
ꢀ77, 6 h
ꢀ80
a
a
ꢀ63b
[18]
[12]
[8]
This work
a
5000
Driedc
3
1.8
1.8
1.8
a
15,000
20,000
20,000
20,000
ꢀ95
ꢀ13
ꢀ58
28
0
a
b
c
No information.
Measured for transient reaction with a temperature ramp of 20 °C/min from ꢀ123 to 227 °C.
Pretreatment and reaction gases were dried by using molecular sieve and silica gel.

124
Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
alytic activity. Over TiO2 (P25) CO conversion was below 2% even at
100
80
60
40
20
0
300 °C. As a result, the data for TiO2 are not shown in Fig. 3.
3.2. Kinetic study of CO oxidation over Co3O4 after different
pretreatments
To understand the relationship between pretreatment and the
origin of high catalytic activity of Co3O4 for low temperature CO
oxidation, XRD, XPS, O2-TPD, CO-TPR, and kinetic study were per-
formed after different pretreatments. XRD patterns show that pre-
treatment in air, CO/air, N2, CO/N2, or H2/N2 within a temperature
range of 150–250 °C hardly changed the bulk structure of Co3O4.
XPS results showed that the binding energy of cobalt and oxygen
was hardly changed by pretreatment in air, N2, or CO/air within a
temperature range of 100–250 °C. Fig. 4 shows Arrhenius plots
for the rate of CO oxidation over Co3O4 with different pretreat-
ments. Pretreatments at 150 °C in dry air, CO/air, and N2 show al-
most the same reaction rates and apparent activation energy, 23–
24 kJ/mol, suggesting a similar reaction pathway for CO oxidation.
Xie et al. reported that Co3O4 nanorods showed almost the same
activation energy of 22 kJ/mol for CO oxidation [8]. Pretreatments
at room temperature and at 150 °C in wet air gave much higher
activation energy, 50–54 kJ/mol, suggesting that reaction pathway
and/or rate-determining step might be changed. Pretreatments at
room temperature in dry air, CO/air, and N2 gave intermediate acti-
vation energies between 34 and 37 kJ/mol.
0
2
4
6
8
40 60 80 100 120 140 160
Time on stream (h)
Fig. 2. Durability of Co3O4 for CO oxidation under different conditions: at ꢀ73 °C
after pretreatment in 1.0 vol% CO/air at 150 °C (e), at 0 °C after pretreatment in dry
air at 150 °C (M), at 0 °C after pretreatment in N2 at 150 °C (s), at 0 °C after
pretreatment in 1.0 vol% CO/air at 150 °C (ꢀ), and at 0 °C after pretreatment in N2 at
200 °C (d).
col as an organic solvent and by calcination at 450 °C. Complete
conversion of CO was maintained for 6 h at ꢀ77 °C and for more
than 55 h at 25 °C in the presence of 3 ppm moisture [8].
The dependence of CO oxidation rate on the concentrations of
CO and O2 after different pretreatments is shown in Table 3. As
for dry air pretreatment, the reaction orders of CO and O2 were
0.30 and 0.31, respectively. This result suggests a possibility that
large amounts of CO and O2 were adsorbed on the surfaces of
Co3O4, leading to weak dependencies of reaction rate on CO and
O2 concentrations and that the reaction rate is substantially con-
trolled by the surface reaction. Pretreatment in CO/air and N2
shows a little lower reaction order for O2 (both 0.16) while a little
higher order for CO (0.47 and 0.48, respectively) than dry air pre-
treatment. In contrast to the pretreatment in dry air, pretreatment
in wet air obviously increases the reaction order of CO to 1.0 while
hardly changes the reaction order of O2, 0.30, indicating that the
pre-adsorption of moisture strongly suppresses the adsorption of
CO and that CO and O2 adsorb on different sites. Takita et al. [19]
In this study, significantly large enhancement in the catalytic
activity and durability for low temperature CO oxidation (such as
at around room temperature) was achieved over usual Co3O4 pow-
der in the presence of moisture simply by optimizing pretreatment
conditions. The influence of pretreatment conditions on CO oxida-
tion was further studied over different kinds of base metal oxides
and is shown in Fig. 3. In the case of Co3O4, MnO2, and NiO, pre-
treatment in N2 or CO/air always leads to higher catalytic activity
for CO oxidation than pretreatment in wet air or reductive atmo-
spheres. In the case of Fe3O4, Fe2O3, CuO, and CeO2, pretreatment
in N2 or in CO/air provided only marginal improvement in the cat-
-100
-50
0
Co3O4
-13.5
-14.0
-14.5
-15.0
-15.5
-16.0
-16.5
-17.0
-17.5
-18.0
-18.5
-19.0
MnO2
NiO (200oC)
CuO
50
100
150
200
250
300
350
Fe2O3 (300oC)
Mn2O3 (250oC)
CeO2
Fe3O4
3.25
3.50
3.75
4.00
4.25
4.50
4.75
5.00
5.25
5.50
1000/T (/K)
300
200
100
0
-100
Fig. 4. Arrhenius plots for the rate of CO oxidation over Co3O4 after pretreatment in
dry air at 150 °C (M), wet air at 150 °C (N), dry air at room temperature (h), wet air
at room temperature (j), 1.0 vol% CO/air at 150 °C (e), 1.0 vol% CO/air at room
temperature (ꢀ), N2 at 150 °C (s), or N2 at room temperature (d). The feed gas
consisted of 1.0 vol% CO and air as a balance and the gas hourly space velocity was
changed in the range of 1.2 ꢂ 104–6.1 ꢂ 104 hꢀ1 ml/g cat in order to obtain CO
conversion below 15% for a differential reactor assumption.
T1/2 after wet air or reductive treatment (oC)
Fig. 3. Temperature for 50% CO conversion (T1/2) after pretreatment in N2 (solid
mark) or 1.0 vol% CO/air (open mark) versus after pretreatment in wet air (or
reductive atmosphere). The pretreatments were performed at 150 °C unless
otherwise noted in parentheses.

Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
125
Table 3
Kinetic parameters for CO oxidation over Co3O4 after different pretreatments.
Pretreatment atmosphere at 150 °C
Temperature range (°C)
Activation energy (kJ/mol)
Reaction ordera
CO
Temperature (°C)
O2
Temperature (°C)
b
N2
ꢀ51 to ꢀ81
ꢀ49 to ꢀ81
ꢀ51 to ꢀ80
ꢀ21 to ꢀ41
23
24
24
50
0.48
0.47
0.30
1.0
ꢀ60
ꢀ61
ꢀ71
ꢀ26
0.16
0.16
0.31
0.30
ꢀ49
ꢀ61
ꢀ66
28
CO/airb
Dry airb
Wet airc
a
The stream containing 0.2–1.0 vol% CO in air (excess oxygen) and the stream containing 1.0–20 vol% O2 and CO (1.0 vol%) in N2 were fed for the measurements of rate
dependency on the partial pressure of CO and of O2, respectively. The gas hourly space velocity was changed in the range of 1.2 ꢂ 104–6.1 ꢂ 104 hꢀ1 ml/g cat in order to keep
CO conversion below 15% for realizing a differential reactor assumption.
b
Before rate measurements, CO oxidation was performed within the temperature range of ꢀ90 to ꢀ49 °C for 5 h to obtain steady state.
c
Before rate measurements, CO oxidation was performed within the temperature range of ꢀ41 to 28 °C for 30 min to obtain steady state.
also showed by TPD experiments that the pre-adsorption of water
did not change the total amount of oxygen desorbed from the sur-
face of Co3O4 even though it resulted in the rearrangement of dif-
ferent oxygen species.
at 150 °C, in addition to peaks around 177 and 380 °C, a small
amount of vO2 (1.1%) was observed which desorbed at around
ꢀ30 °C. After N2 pretreatment at 50 °C, vO2 was hardly observed,
whereas increasing the pretreatment temperature to 150 °C in N2
results in the formation of large amount of vO2 (6.5%, desorbed
from ꢀ60 °C and reached maximum at ꢀ10 °C). Further increase
in the pretreatment temperature to 250 °C gradually raised the
desorption temperature of vO2 as well as the amount of vO2. Taking
into account that pretreatment at 150 °C in N2 led to the highest
catalytic activity for CO oxidation at ꢀ80 °C when pretreatments
were carried out in a temperature range of room temperature to
250 °C, it is likely that weakly bound molecular oxygen species
(vO2) play a crucial role in CO oxidation at low temperatures and
their amount and bonding strength can be tuned by the pretreat-
ments. Namely, the lower the desorption temperature of vO2 is,
the more active Co3O4 is for low temperature CO oxidation. Mean-
while, pretreatment in N2 at 150 °C forms more vO2 than dry air
pretreatment at the same temperature, which can explain why
pretreatment in N2 gives a little lower reaction order of O2 than
pretreatment in dry air.
3.3. Low temperature O2-TPD and CO-TPR studies over Co3O4 after
different pretreatments
To further investigate the relationship between adsorbed oxy-
gen and pretreatment conditions, O2-TPD experiments were car-
ried out starting at ꢀ70 °C. In Fig. 5, four kinds of oxygen species
are observed at the temperature ranges of ꢀ60 to 40 °C, 100–
200 °C, 250–400 °C, and at above 400 °C, which could be assigned
to molecular oxygen species adsorbed on oxygen vacancy (vO2),
surface oxygen ion bound with Co2+ and Co3+ (OWS), surface oxygen
ion bound with three Co3+ cations (OSS), and bulk oxygen (OB),
respectively. The ratios of different oxygen species after different
pretreatments were further estimated by the area of corresponding
peak and are summarized in Table 4. After pretreatment in dry air
So far many papers have discussed the relationship between the
formation of oxygen species and the catalytic activity of Co3O4 for
CO oxidation [13,17,20–24]. It has generally been accepted that ad-
sorbed surface oxygen species play a key role in CO oxidation
[13,17,20,21,24,25], while few have proposed that lattice oxygen
has higher catalytic activity for CO oxidation than adsorbed species
[22,23]. In view of the fact that 100% oxidation of CO over Co3O4
can occur at temperatures far below room temperature, the active
oxygen species may desorb at lower temperatures. However, the
desorption of oxygen on Co3O4 was investigated with a conven-
tional TPD only above room temperature [13,19,24]. In this study,
our TPD instrumentation has enabled us to observe that the
desorption of oxygen species weakly bound with Co3O4 can occur
at ꢀ60 °C after O2 preadsorption at ꢀ70 °C. Further CO-TPR exper-
iments show that this weakly bound oxygen could react with CO to
form CO2 at ꢀ40 °C (Fig. 6). It should be noted that, after O2 pread-
sorption at ꢀ70 °C, the sample was purged with 5% CO/He at the
same temperature for 60 min. Even in this purging process (the in-
serted graph in Fig. 6), CO2 was formed at the initial stage, reaching
a maximum amount soon after, and then the amount decreased
gradually, which indicates sufficiently high reactivity of weakly
bound oxygen with CO at ꢀ70 °C.
x6
OSS
380 oC
OWS
vO2
17 oC
157 oC
Air 150 o
C
0
OB
-60 -30
N2 250 o
30
C
2 oC
172 oC
177 oC
N2 200 o
C
N2 150 o
N2 50 o
C
-10 oC
C
Air 150 o
C
-100
0
100
200
300
400
500
600
Temperature (oC)
Fig. 5. O2-TPD profiles of Co3O4 after different pretreatments. The procedures were
as follows: (1) pretreatment of 0.2 g Co3O4 under different conditions for 40 min;
(2) cooling down to ꢀ70 °C and O2 adsorption for 1 h; (3) purging with He for 1 h;
and (4) heating at a ramp of 10 °C/min from ꢀ70 to 600 °C in He and recording mass
signal of O2 (m/z = 32) simultaneously.
Table 4
The ratios of different oxygen species estimated by the area of corresponding peak.
Air at 150 °C (%)
N2 at 50 °C (%)
N2 at 150 °C (%)
N2 at 200 °C (%)
N2 at 250 °C (%)
vO2
OWS
OSS
OB
1.1
18.0
42.3
38.6
0
6.1
41.2
52.7
6.5
4.3
34.4
54.8
11.8
7.1
44.8
36.3
23.4
14.9
31.3
30.4

126
Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
XRD patterns show that Co3O4 is the only phase after calcina-
tion of the precipitate Co(OH)(CO3) at 300 °C in air. Tricobalt tetra-
-70oC
CO2
oxide is a spinel type oxide which consists of cobalt (II) in
tetrahedral sites (Co2+) and cobalt (III) in octahedral sites (Co3+
)
[20,37–39]. By successively substituting the Co2+ and Co3+ ions in
Co3O4 with less active Zn2+ and Al3+ ions, respectively, it has been
proved that the surface Co3+ of cobalt spinel oxide is essential for
the genesis of high activity for CO oxidation while Co2+ is not useful
for this reaction [37]. Intensive studies on the low temperature CO
oxidation over Co3O4 have been performed by Jansson et al.
[21,25,38], and it has been proposed that CO adsorption and reac-
tion at low temperature take place on the surface octahedral sites
(Co3+). Preferred adsorption of CO is also found to occur at surface-
exposed Co3+ sites while other sites such as sublayer Co3+ ions are
sterically hindered and inaccessible for CO adsorption based on a
DFT study [39].
0:13:29 0:27:38 0:44:58
Purging Time (min)
CO2
CO
400
O2
-100
0
100
200
300
500
600
Temperature (oC)
As shown in Fig. 7 in the left-hand side structure presented by
Broqvist et al. [39], there are two kinds of oxygen ions on Co3O4
(110) surface: one is bonded to one Co2+ and one Co3+ ion, while
another is bound to three Co3+ ions as the nearest neighbors. Thus,
the two surface oxygen ions on the Co3O4 (110) plane experience
different crystal fields, which cause their electropositivities to dif-
fer, and thereby also their reactivities. In Fig. 7, these surface oxy-
gen ions are hereafter denoted as OWS and OSS, representing the
weaker and stronger local crystal field sites, respectively. As a re-
sult, two different surface oxygen vacancies can be formed by
the oxygen abstraction from either OWS or OSS site, but the OWS
vacancies formation may be more feasible and lead to significantly
longer lifetime than the OSS vacancies [39]. The desorption peaks at
157–177 °C in TPD study (Fig. 5) may correspond to the surface
oxygen abstraction from OWS site, which also gives an evidence
that pretreatment at above 150 °C is highly effective for CO oxida-
tion by enhancing the formation of surface oxygen vacancies. The
oxygen desorption starting at 250–300 °C can be assigned to the
surface oxygen abstraction from OSS. The shoulder observed at
around 420 °C can be ascribed to the bulk oxygen (OB) since this
sample was calcined at 300 °C.
Based on the above results in combination with the hypotheses
of CO oxidation over metal oxides at low temperatures [33,34], a
possible scheme is proposed for the role of pretreatment and for
the pathways of CO oxidation over Co3O4. As shown in Fig. 7, pre-
treatment in inert and reducing–oxidizing atmosphere (N2 and CO/
air) at moderate temperatures enhances the formation of surface
OWS vacancies by oxygen removal from the OWS sites and then
leads to the formation of weakly bound oxygen species such as
bridged Co3þ—Oꢀ2 —Co2þ, which exhibits high reactivity toward ad-
sorbed ‘‘end on” CO species on the surface-exposed Co3+ sites to
form CO2 even at ꢀ80 °C. During this process, the dissociation of
adsorbed Oꢀ2 occurs and then atomic oxygen anion leaves the sur-
face vacancies (Co3þ—Oꢀ—Co2þ). Considering that the reactivity of
adsorbed Oꢀ is a little higher than that of Oꢀ2 [34], the CO oxidation
can easily undergo to produce CO2, and then the surface OWS
vacancies are available again for O2 adsorption. Also, it is indicated
that the formation of adsorbed atomic oxygen anion may be the
rate-determining step during the low temperature CO oxidation
over Co3O4 after pretreatment in dry air, N2, and CO/air at above
150 °C. It is reasonable that pretreatment in dry air is less efficient
for the formation of surface OWS vacancies than pretreatment in N2,
or CO/air. Water molecules are more strongly adsorbed on Co3+ ac-
tive sites than CO and thereafter retard the adsorption of CO [40],
causing a sharp increase in the reaction order of CO.
Fig. 6. CO-TPR profiles of Co3O4 after pretreatment in N2 at 150 °C. The procedures
were as follows: (1) pretreatment of 0.2 g Co3O4 in N2 at 150 °C for 40 min; (2)
cooling down to ꢀ70 °C and O2 adsorption for 1 h; (3) purging with 5.0 vol% CO in
He for 1 h; and (4) heating at a ramp of 10 °C/min from ꢀ70 to 600 °C in 5.0 vol% CO
in He and recording mass signal of CO (m/z = 28), O2 (m/z = 32), and CO2 (m/z = 44),
simultaneously.
4. Discussion
It is well known that supported gold catalysts exhibit high cat-
alytic activity for CO oxidation at temperatures far below room
temperature. The adsorption and activation of molecular oxygen
are thought to be crucial to gold catalysts for low temperature
CO oxidation [26–30]. The formation of adsorbed reactive oxygen
species such as superoxide ion ðOꢀ2 Þ can be correlated to the pres-
ence of surface oxygen vacancies on metal oxide support [29] or on
the metal–support interfaces [26–29]. By using temporal analysis
of product (TAP) techniques, Widmann et al. [31] found that the
catalytic activity of Au/CeO2 for CO oxidation was significantly im-
proved by the removal of about 7% of the surface oxygen through
pre-reduction with a few CO pulses, whereas over-reduction
caused lower catalytic activity at the initial stage. Meanwhile, dif-
ferent kinds of superoxide species were also observed on the sur-
faces of CoO–MgO dilute solid solutions between ꢀ196 and 25 °C
by IR and EPR [32], and it was hypothesized that the superoxide
species played an important role in CO oxidation over metal oxides
at low temperatures [33,34]. These previous studies indicate that
pretreatments in inert and reducing–oxidizing atmosphere such
as N2 and CO/air may lead to the creation of surface oxygen vacan-
cies, where molecular oxygen is activated at low temperatures,
resulting in the improved catalytic activity and/or durability.
Sadykov and co-workers [13,14] also studied the influence of
pretreatment atmosphere on the formation of adsorbed oxygen
species over Co3O4 calcined at 400 °C. In comparison with O2 pre-
treatment at 350 °C, pretreatment in He at the same temperature
enhanced the formation of weakly bound oxygen species over
Co3O4. In TPD starting at 25 °C, the desorption of O2 occurred at
27–97 °C, which was much higher than the temperature for the
desorption of molecular oxygen species adsorbed on Co3O4 cal-
cined at 300 °C in our TPD experiments. It is well known that the
metal oxides prepared in air inevitably bear surface hydroxyl
groups, which has been clearly identified by IR measurement
[35]. As reported by Ciuparu et al. [36], a part of hydroxyls is
known to be easily removed by pretreatment in the flow of H2 at
temperature below 700 °C, thus creating surface oxygen vacancies.
The relationship between the removal of surface hydroxyls and the
formation of oxygen vacancies over Co3O4 during different pre-
treatment will be further investigated by IR spectroscopy in the
near future.
As shown in Fig. 3, the catalytic activity of base metal oxides for
CO oxidation decreased in the following sequence: Co3O4 > M-
nO2 > NiO > CuO > Fe2O3 > Mn2O3 > Fe3O4 > CeO2 > TiO2,
which
approximately coincides with an increasing sequence of metal–
oxygen bond energy of base metal oxides as reported by Boreskov

Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
127
OWS Vacancy
OSS
OSS
OSS
OWS
CO
150-250oC
-80oC
Air, N2 or CO/air
-80oC
-80oC
CO2
-
O2 + ( )
(O2 )
(COCo3+)
CO2
rate-determining
CO + Co3+
OSS
OSS
(O2 ) + (COCo3+)
-
CO2 + (O-) + Co3+
CO2 + ( ) + Co3+
CO
(O-) + (COCo3+)
-80oC
Co3+ Co2+
OWS
OSS
O
H
C
Fig. 7. A possible scheme for the role of pretreatment and CO oxidation pathways over Co3O4. OSS: surface oxygen ions of Co3O4 bonded to three Co3+ ions as the nearest
neighbors, representing the stronger local crystal field sites; OWS: surface oxygen ions of Co3O4 bonded to one Co2+ and one Co3+ ion, representing the weaker local crystal
field sites. The oxygen removal from OWS sites to form surface OWS vacancies may occur preferentially to that from OSS sites.
[41]: Co3O4 < CuO < NiO ꢃ MnO2 < Fe2O3 < TiO2. By taking into ac-
count the lower metal–oxygen bond energy of Co3O4, MnO2 and
NiO, it is reasonably understood that pretreatment in N2 or CO/
air at moderate temperature is efficient for oxygen abstraction
from the surface of these metal oxides to form surface oxygen
vacancies, finally enhancing the adsorption of O2 for CO oxidation.
Cupric oxide, CuO, has metal–oxygen bond energy of 76–80 kJ/mol,
which is similar to that of MnO2 and NiO (80–84 kJ/mol) [41,42],
however, pretreatment in N2 or in CO/air offered only a marginal
improvement in the catalytic activity. Sadykov et al. [14] proposed
that pretreatment and activity measurement conditions can
broadly vary the surface layer stoichiometry of Co3O4, NiO, and
MnO2, while CuO possesses a very narrow range of the surface
non-stoichiometry, indicating that the surface non-stoichiometric
variability may also contribute to the effect of pretreatment.
measurements. They also thank Prof. Y. Shimakawa and Dr. Y. Teng
of Kyoto University for their advices concerning metal oxides prep-
aration. Y. Yu also thanks Japan Society for the Promotion of Sci-
ence (JSPS) for the financial support to his stay at TMU for 2
years as a postdoctoral fellow. This work was partly supported
by Grant-in-Aid for Specially Promoted Research (19001005) of
the Ministry of Education, Culture, Science and Sports, Japan. It
was also partly supported by the National Natural Science Founda-
tion for Creative Research Groups of China (50621804).
References
[1] A.V. Nero Jr., Scientific Am. 258 (1998) 24.
[2] M. Jacoby, Chem. Eng. News 2003, January 20, p. 32.
[3] R. Jain, US Patent 5 110 569, 1992.
[4] N. Funasaki, A. Henmi, S. Ito, Y. Asano, S. Yamashita, T. Kobayashi, M. Haruta,
Sensor. Actuat. B 14 (1993) 536.
[5] M. Haruta, A. Ueda, S. Tsubota, R.M. Torres Sanchez, Catal. Today 29 (1996)
443.
5. Conclusions
[6] F.S. Stone, Adv. Catal. 13 (1962) 1.
[7] M. Haruta, M. Yoshizaki, D.A.H. Cunningham, T. Iwasaki, Ultra Clean Technol. 8
(1996) 117 (in Japanese).
Atmospheres, temperature, and the moisture content in the pre-
treatments have a great influence on the catalytic activity of Co3O4,
MnO2, and NiO which are calcined at relatively low temperatures of
300 or 400 °C. They are p-type semiconductors having smaller me-
tal–oxygen bond energy than other base metal oxides and are char-
acterized by relatively large surface non-stoichiometric variability.
These two properties can be correlated to the formation and stabil-
ity of surface oxygen vacancies and lead to the high catalytic activ-
ity of Co3O4, MnO2, and NiO for CO oxidation at a temperature as
low as ꢀ80 °C. The other metal oxides are intrinsically much less
active for CO oxidation at temperatures below 120 °C, and there-
fore pretreatment effect is small. This finding may provide a new
guideline for the creation of low-temperature catalysts without
using noble metals by optimizing pretreatment conditions.
[8] X. Xie, Y. Li, Z.-Q. Liu, M. Haruta, W. Shen, Nature 458 (2009) 746.
[9] Z. Qu, M. Cheng, W. Huang, X. Bao, J. Catal. 299 (2005) 446.
[10] S.H. Kim, S.W. Nam, T.H. Lim, H.I. Lee, Appl. Catal. B 81 (2008) 97.
[11] D.A.H. Cunningham, T. Kobayashi, N. Kamijo, M. Haruta, Catal. Lett. 25 (1994)
257.
[12] Y.-Z. Wang, Y.-X. Zhao, C.-G. Gao, D.-S. Liu, Catal. Lett. 116 (2007) 136.
[13] V.A. Razdobarov, V.A. Sadykov, S.A. Veniaminov, N.N. Bulgakov, O.N.
Kovalenko, Yu.D. Pankratiev, V.V. Popovskii, G.N. Kryukova, S.F. Tikhov,
React. Kinet. Catal. Lett. 37 (1988) 109.
[14] V.A. Sadykov, S.F. Tikhov, S.V. Tsybulya, G.N. Kryukova, S.A. Veniaminov, V.N.
Kolomiichuk, N.N. Bulgakov, E.A. Paukshtis, V.P. Ivanov, S.V. Koshcheev, V.I.
Zaikovskii, L.A. Isupova, L.B. Burgina, J. Mol. Catal. A 158 (2000) 361.
[15] S.B. Kanungo, J. Catal. 58 (1970) 419.
[16] F. Grillo, M.M. Natile, A. Glisenti, Appl. Catal. B 48 (2004) 267.
[17] M.J. Pollard, B.A. Weinstock, T.E. Bitterwolf, P.R. Griffiths, A.P. Newbery, J.B.
Paine III, J. Catal. 254 (2008) 218.
[18] P. Thormählen, M. Skoglundh, E. Fridell, B. Andersson, J. Catal. 188 (1999) 300.
[19] Y. Takita, T. Tashiro, Y. Saito, F. Hori, J. Catal. 97 (1986) 25.
[20] D. Perti, R.L. Kabel, G.J. Mccarthy, AIChE J. 31 (1985) 1435.
[21] J. Jansson, M. Skoglundh, E. Fridell, P. Thormählen, Top. Catal. 16/17 (2001)
385.
Acknowledgments
The authors would like to acknowledge Mr. Y. Man of Beijing
Research Institute of Chemical Industry for O2-TPD and CO-TPR
[22] H.-K. Lin, C.-B. Wang, H.-C. Chiu, S.-H. Chien, Catal. Lett. 86 (2003) 63.

128
Y. Yu et al. / Journal of Catalysis 267 (2009) 121–128
[23] C.-B. Wang, C.-W. Tang, H.-C. Tsai, M.-C. Kuo, S.-H. Chien, Catal. Lett. 107
(2006) 31.
[24] Y.-Z. Wang, Y.-X. Zhao, C.-G. Gao, D.-S. Liu, Catal. Lett. 125 (2008) 134.
[25] J. Jansson, J. Catal. 194 (2000) 55.
[34] J.V.D. Berg, A.J.V. Dillen, J.V.D. Meijden, J.W. Geus, in: J.P. Bonnelle, B. Delmon,
E. Derouane (Eds.), Surface Properties and Catalysis by Non-Metals, D. Reidel
Publishing Company, 1983, p. 493.
[35] M.I. Zaki, H. Knözinger, Mater. Chem. Phys. 17 (1987) 201.
[36] D. Ciuparu, Y. Chen, S. Lim, Y. Yang, G.L. Haller, L. Pfefferle, J. Phys. Chem. B 108
(2004) 15565.
[37] K. Omata, T. Takada, S. Kasahara, M. Yamada, Appl. Catal. A 146 (1996) 255.
[38] J. Jansson, A.E.C. Palmqvist, E. Fridell, M. Skoglundh, L. Österlund, P.
Thormählen, V. Langer, J. Catal. 211 (2002) 387.
[39] P. Broqvist, I. Panas, H. Persson, J. Catal. 210 (2002) 198.
[40] A.J. Goodsel, J. Catal. 30 (1973) 175.
[26] M. Haruta, CATTECH 6 (2002) 102.
[27] G.C. Bond, D.T. Thompson, Gold Bull. 33 (2000) 41.
[28] J. Guzman, S. Carrettin, J.C. Fierro-Gonzalez, Y. Hao, B.C. Gates, A. Corma,
Angew. Chem., Int. Ed. 44 (2005) 4778.
[29] J. Guzman, S. Carrettin, A. Corma, J. Am. Chem. Soc. 127 (2005) 3286.
[30] J.A. van Bokhoven, C. Louis, J.T. Miller, M. Tromp, O.V. Safonova, P. Glatzel,
Angew. Chem., Int. Ed. 45 (2006) 4651.
[31] D. Widmann, R. Leppelt, R.J. Behm, J. Catal. 251 (2007) 437.
[32] E. Giamello, Z. Soijka, M. Che, A. Zecchina, J. Phys. Chem. 90 (1986) 6084.
[33] G.I. Golodets, J.R.H. Ross, Language (Eds.), Heterogeneous Catalytic Reaction
Involving Molecular Oxygen, Studies in Surface Science and Catalysis, vol. 15,
Elsevier, Amsterdam–Oxford–New York, 1983, p. 280.
[41] G.K. Boreskov, Kinet. Catal. 8 (1967) 878.
[42] A. Bielanski, J. Haber, Catal. Rev. Sci. Eng. 19 (1979) 1.