Low-temperature selective oxidation of CO in H2-rich gases over Ag/SiO2 catalysts

Article abstract of DOI:10.1016/j.molcata.2005.05.033

Factors influencing basic catalytic activity of Ag/SiO₂ catalysts for CO selective oxidation in H₂-rich gases, such as support, silver loading, oxygen concentration and reaction temperature, etc. were investigated. Among the supports used, SiO₂ and the pure silica zeolite are better supports for silver catalyst. Ag/SiO₂ catalyst has a better stabilization for CO selective oxidation at low temperatures. Carbon monoxide in the presence of large excess of hydrogen can be stoichiometrically oxidized by oxygen over Ag/SiO₂ catalysts at low temperatures, and the catalytic activities increase with the increase of the silver loading. Low reaction temperature favors this process of selective oxidation of CO in H ₂-rich gas, and the selectivity of O₂ to oxidize CO can reach 60-80%. The maximum conversion of CO under such selectivity levels for oxygen to oxidize CO also increases with the silver loading. XRD results show that the main species on all catalysts is crystallite silver, and the size of the silver particles increases with the silver loading. It is found that the CO conversion increases with increasing oxygen contents properly, however, CO is not consumed completely even when the overall concentration of O₂ is increased to 3%. Pretreatment conditions have been found to strongly influence the activity and selectivity of the catalysts. The catalysts deactivate after pretreatment with hydrogen at 500°C, whereas a better activity is obtained after pretreatment with oxygen or He at the same temperature.

Full text of DOI:10.1016/j.molcata.2005.05.033

Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
Low-temperature selective oxidation of CO in H2-rich
gases over Ag/SiO2 catalysts
Zhenping Qu1, Mojie Cheng, Chuan Shi, Xinhe Bao∗
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences,
457 Zhongshan Road, P.O. Box 110, Dalian 116023, China
Received 23 July 2004; received in revised form 21 April 2005; accepted 27 May 2005
Available online 5 July 2005
Abstract
Factors influencing basic catalytic activity of Ag/SiO2 catalysts for CO selective oxidation in H2-rich gases, such as support, silver loading,
oxygen concentration and reaction temperature, etc. were investigated. Among the supports used, SiO2 and the pure silica zeolite are better
supports for silver catalyst. Ag/SiO2 catalyst has a better stabilization for CO selective oxidation at low temperatures. Carbon monoxide in
the presence of large excess of hydrogen can be stoichiometrically oxidized by oxygen over Ag/SiO2 catalysts at low temperatures, and the
catalytic activities increase with the increase of the silver loading. Low reaction temperature favors this process of selective oxidation of CO
in H2-rich gas, and the selectivity of O2 to oxidize CO can reach 60–80%. The maximum conversion of CO under such selectivity levels for
oxygen to oxidize CO also increases with the silver loading. XRD results show that the main species on all catalysts is crystallite silver, and
the size of the silver particles increases with the silver loading. It is found that the CO conversion increases with increasing oxygen contents
properly, however, CO is not consumed completely even when the overall concentration of O2 is increased to 3%. Pretreatment conditions
have been found to strongly influence the activity and selectivity of the catalysts. The catalysts deactivate after pretreatment with hydrogen at
500 ◦C, whereas a better activity is obtained after pretreatment with oxygen or He at the same temperature.
© 2005 Published by Elsevier B.V.
Keywords: Ag/SiO2; CO selective oxidation; H2-rich gas
1. Introduction
tial to develop a method that can lower effectively the CO
concentration in the reforming gases.
Polymer electrolyte membrane fuel cells (PEFCs), operat-
ing at a relatively low-temperature (e.g. 60–80 ◦C) with pure
hydrogen or reforming gases as fuels, have attracted much
attention as power sources for electric vehicles [1,2]. How-
ever, at low-temperatures, the Pt anodes of the PEFC are
extremely sensitive to and easily poisoned by traces of CO
in the fuels [3]. When employing reforming gases, the CO
concentration is generally ca. 1%, which is much higher than
the 10 ppm level desired for the PEFC. Therefore, it is essen-
Among the various methods available for CO removing,
catalytic selective oxidation of CO seems to be a promis-
ing one and has been investigated extensively. Most cata-
lysts employed previously are based on noble metals [4–9].
Pt/Al2O3 catalysts were reported to have a high activity and
selectivity for the selective oxidation of CO; however, it
requires more than 2% of oxygen to oxidize 1% of CO in
H2-rich gases, which might result in a wasteful consump-
tion of H2 and a risk of explosion [4,5]. Oh and Sinkevitch
reported that Ru/Al2O3 and Rh/Al2O3 catalysts exhibited
high activities for the selective oxidation of CO in H2-rich
gases (900 ppm CO, 800 ppm O2 and 0.85 vol% H2) [6,7].
However, the reaction temperature (200–250 ◦C) over Pt, Rh
catalysts is relatively high compared with that of PEFCs.
∗
Corresponding author. Tel.: +86 411 84686637; fax: +86 411 84694447.
E-mail address: xhbao@dicp.ac.cn (X. Bao).
Present address: Institute of Physical Chemistry, University of Stuttgart,
1
70569 Stuttgart, Germany.
1381-1169/$ – see front matter © 2005 Published by Elsevier B.V.
doi:10.1016/j.molcata.2005.05.033

Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
23
Bethke has reported that selectivity of ca. 50% at 100 ◦C was
obtained over Au/␥-Al2O3 catalysts containing Au particles
of about 5–10 nm in size [8]. Basic factors influencing the
activity and selectivity of catalysts for CO selective oxida-
tion in H2 have also been widely investigated, which is very
important for the application of the catalyst in the PEFCs
and the research of the properties of the catalyst. The effect
of the supports has also been investigated on the Au and Pt
catalysts [9–11]. Igarashi and co-workers have shown that at
high temperatures noble metal catalysts supported on mor-
denite and other zeolites possessed good selectivity for CO
oxidation in H2 [7,10]. The better activity and selectivity can
be obtained on the Au catalyst supported on the reducible
oxide [11]. Reaction temperature and oxygen concentration
in the feed gas were found to play an important role in the
activity. Manasilp and Gulari have found that the best reac-
tion activity of Pt/Al2O3 was obtained at 170 ◦C, and CO can
be completely oxidized when the oxygen concentration was
increased to 1.35%, however, the selectivity increases with
the decrease of the oxygen concentration [12].
monochromatic detector. Copper K␣ radiation was used with
power setting at 40 kV and 100 mA. Typical operation param-
eters were: a scan rate of 5◦ min−1. The size of the Ag
particles was calculated from the half-width of the main peak
at 2θ = 38◦ according to the Scherrer’s method [22].
Adsorption studies of CO and O2 were conducted in a
Chemisorption 2800 instrument. Typically, the catalyst was
treated in a flow of He at 500 ◦C for 2 h and degassed for
1 h at the same temperature. The uptakes of CO and O2 were
then measured at 35 ◦C.
TEM experiments were performed with the JEM-2000EX
microscope, using an accelerating voltage of 100 kV. For the
TEM analysis, a sample was dispersed in absolute alcohol
by an ultrasonic bath and deposited on a Formvar coated 200
mesh Cu grid.
2.2. Activity tests
The reactions of selective oxidation of CO were carried
out in a conventional flow reactor. The reactants were fed with
a volume ratio of H2/CO/O2 = 98.5/1/0.5, which were con-
trolled by an independent thermal mass flow controller. While
in the case of experiment for the effect of oxygen concentra-
tion on the activity and selectivity of catalysts, it consisted of
1 vol% CO, 77.08 vol% H2, O2 with varying O2:CO molar
ratio (λ = 1–4, 6) and He balance. The parameter λ is defined
as the amount of oxygen present divided by the amount of
oxygen needed to completely oxidize all of the CO in the
feed gas (λ = 2[O2]/[CO]), where λ = 1 is sufficient to pro-
vide for the complete oxidation of CO to CO2 in the absence
of an oxygen consuming side reaction. The total flow rate
was 50 ml/min. Typically, 0.25 g of catalyst was used and
pretreated under He, or 30% O2 in He, or highly pure hydro-
gen, at 773 K for 2 h. The temperature of the catalyst bed was
monitored with a thermocouple that was put on the center of
the reaction bed.
Silver has been investigated and utilized as an oxida-
tion catalyst [13,14], and supported silver catalysts have
been proved to possess a stable activity for CO oxidation at
low temperatures [15–17], e.g. complete CO oxidation over
¨
¨
Ag/Mn/perovskite can be attained below 373 K. Guldur and
Balikci studied the Ag/Co and Ag/Mn composite oxide cata-
lysts used in removing CO in H2-rich gases between 110 and
200 ◦C. However, no obvious activity for CO selective oxi-
dation was observed at the temperatures below 100 ◦C [18].
Recently, we have shown that silver catalyst is a promising
catalyst for the removal of CO in the fuels for PEFCs at low-
temperatures [19–21]. We mainly focused on the activations
of oxygen species and silver catalysts, and the effect of the
formation of subsurface oxygen species on the CO oxida-
tion. The formation of subsurface oxygen species increases
the activity of the silver catalyst, and more reactant species
can adsorb on the surface of silver catalyst. The presence
of subsurface oxygen species supplies another way for CO
oxidation. However, the basic factors influencing the activity
of silver catalysts for CO selective oxidation are not care-
fully examined. In this paper, we concretely investigate the
influence of reaction conditions, e.g. support, oxygen con-
centration, reaction temperature and silver loading, on the
activity and selectivity.
On-line gas chromatograph with a TCD detector was
employed to measure the reactor inlet and outlet effluent gas
streams. Two columns were used: a molecular sieve 5A col-
umn (3 m × 4 mm) was used to separate hydrogen, oxygen
and carbon monoxide. A Porapak Q column (3 m × 4 mm)
was used to analyze carbon dioxide and H2O. Both columns
were operated at a constant temperature of 30 ◦C.
Two parallel reactions occur in the present catalytic sys-
tem:
1
2. Experimental
CO + O2 → CO2
(1)
(2)
2
1
2
2.1. Catalyst preparation and characterization
H2 + O2 → H2O
Silver catalysts supported on different supports were pre-
pared by an incipient wetness impregnation method. The sup-
port was contacted with a certain amount of AgNO3 precursor
solution under stirring. The as prepared AgNO3/support sam-
ple was then aged at RT and dried at 80 ◦C for 12 h.
In this paper, the CO conversion was calculated from the
change of the CO concentration:
[CO]in − [CO]out
CO conversion =
× 100%,
[CO]in
X-ray powder diffraction (XRD) experiments were car-
ried out on a Rigaku D/max-␥b X-ray diffractometer with
where [CO]in is the inlet CO concentration and [CO]out is the
outlet CO concentration.

24
Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
The O2 conversion is based on oxygen consumption:
ported on the CeO2, Al2O3 and H-ZSM-5, indicating the
silver species are finely dispersed on the support. The peaks
for metallic silver particles appear on the SiO2, MgO, C and
TiO2. According to the Scherrer’s method, the size of silver
particles is about 8 nm for Ag/SiO2 catalyst. The half-width
of the diffraction peak at 2θ = 38◦ becomes sharp when sup-
ported on the MgO, C and TiO2, and the larger silver particles
are formed. Different particle size is obtained when silver
is supported on different supports. It is reasonable to think
this phenomenon is related with the different properties of
supports. Lower BET surface area of MgO (38.22 m2/g), C
(5.49 m2/g) and TiO2 (11.76 m2/g) will result in the poor dis-
persion of silver particles. However, a fine dispersion can
still be obtained, though the BET surface area of CeO2
(74.54 m2/g), Al2O3 (240.23 m2/g) is lower than that of SiO2
(495 m2/g). A strong interaction of silver species with CeO2
and Al2O3 is easily formed [23,24], meanwhile the properties
of silver particles are changed.
[O2]in − [O2]out
O2 conversion =
× 100%,
[O2]in
where [O2]in is the inlet O2 concentration and [O2]out is the
outlet O2 concentration.
Selectivity towards CO2 was defined as the ratio of oxygen
used to oxidize CO to CO2 over the total amount of oxygen
consumption.
(CO)
ꢀO2
S =
(CO)
(H2O)
ꢀO2
+ ꢀO2
The amount of oxygen not used in oxidizing CO to CO2
was used to oxidize H2, which would reduce the amount of
H2 used to feed into fuel cell. Then the selectivity should be
large enough not to reduce the fuel efficiency. Because of the
large error in the quantification of H2O, the selectivity of CO
oxidation was calculated from the oxygen mass balance:
Powder XRD patterns of the Ag/SiO2 catalysts with var-
ious Ag loadings and that of a reference SiO2 sample are
shown in Fig. 2. All samples were pretreated under He atmo-
sphere at 500 ◦C. No diffraction peaks of silver or silver
compounds are detected on the 2% Ag/SiO2 sample, indi-
cating that silver species are well dispersed. Metallic silver,
the main peak of which is at 2θ = 38◦ and corresponding
to the (1 1 1) plane of silver, appears in the high loading
silver/silica samples. The crystallite sizes, determined by
Scherrer’s method from the XRD peak broadening, are 5.7,
8, 14, 16.5 and 18 nm, respectively. The particle size of sil-
ver catalyst increases with the silver loading. When the silver
loading of the catalyst is higher than 16%, the average size of
the Ag particles does not change obviously (not shown here).
A small peak appears at 2θ = 33◦ for the Ag/SiO2 catalysts
with Ag loadings higher than 8%.
0.5 × [CO2]
0.5 × CCO × [CO]in
S =
× 100% =
[O2]in − [O2]out
CCO
CO × [O2]in
2
=
λ × CO
2
where CCO is CO conversion and CO is O2 conversion.
A definition comparison of λ and S demonstrates that λ
and S are interrelated.
2
3. Results and discussion
3.1. Catalyst characterization
Fig. 1 shows the XRD patterns of silver catalysts (weight
loading of 6%) supported on different supports. All catalysts
were pretreated with He at 500 ◦C for 2 h. No diffraction
peaks of silver species are observed when the silver is sup-
Table 1 shows the amounts of CO and O2 adsorbed per
gram of silver on silica-supported catalysts with different sil-
ver loadings. As can be seen, the reversible adsorption for
Fig. 2. XRD patterns of Ag/SiO2 catalysts with different loadings pretreated
with He at 500 ◦C for 2 h.
Fig. 1. XRD patterns of silver catalysts supported on different supports.

Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
25
Table 1
Amounts of CO and O2 sorbed per gram of silver on the silica-supported
silver catalysts with different loading
Extrapolation
(cc/G STP)
2%
4%
6%
8%
12%
CO
Total
Reversible
Irreversible
1.4386
1.3801
0.0366
1.7282
1.5765
0.1460
2.0275
1.8469
0.1845
2.8596
2.6272
0.2568
3.1243
2.6802
0.4417
O2
Total
Reversible
Irreversible
0.9434
0.9103
0.0366
1.0106
1.0033
0.0080
1.1152
1.1047
0.0113
1.3532
1.3428
0.0104
1.5005
1.4898
0.0107
Differencea
0.4952
0.7176
0.9123
1.5064
1.6238
a
The difference between the total amounts of CO adsorbed and that of O2
adsorbed.
oxygen is the dominant adsorption at this temperature. More-
over, both the total amounts of adsorption of CO and O2 at
35 ◦C and the difference between the adsorbed amounts of
CO and O2 increases with the silver loading.
To understand the difference in characteristics between
the Ag/SiO2 catalysts treated with hydrogen and that with
oxygen, the catalysts are characterized by XRD and TEM.
The size distribution of silver particles after different treat-
ment is shown in Table 2. Only a range of the mean size of
the silver particles is given in Table 2, due to the limit of the
particle numbers. Fig. 3 shows the XRD patterns of the 16%
Ag/SiO2 pretreated under oxygen and hydrogen atmospheres
at 500 ◦C, respectively. It can be clearly seen that the atmo-
sphere employed in the pretreatment has a great influence on
the size and structure of the Ag particles. The XRD pattern
of the 16% Ag/SiO2 treated with oxygen at 500 ◦C for 2 h
shows diffraction peaks corresponding to metal silver parti-
cles with average size of about 18 nm, as well as ultrafine
silver oxide species (2θ = 33◦) (Fig. 3A(a)). After the sample
is purged with hydrogen at the same temperature for 1 h, the
diffraction peak of the silver oxide species disappears, and
the average size of the silver particles decreases to about 8 nm
(Fig. 3A(b)). If the 16% Ag/SiO2 sample is firstly pretreated
with highly pure hydrogen at 500 ◦C for 2 h, only metal sil-
ver particles with average size of about 12 nm are observed
by XRD (Fig. 3B(c)). The average size of the metallic silver
particles (about 17 nm) obviously increases after the sample
(Fig. 3B(c)) is purged by oxygen at the same temperature for
Fig. 3. XRD patterns of 16% Ag/SiO2 treated with oxygen and hydrogen
atmospheres at 7500 ◦C. (A) Firstly treated with oxygen for 2 h (a), followed
by hydrogen atmosphere at the same temperature for 1 h (b). (B) Firstly
treated with hydrogen for 2 h (c), followed by oxygen atmosphere at the
same temperature for 1 h (d).
1 h (Fig. 3B(d)). However, in this case, no diffraction peak of
silver oxide can be detected. Yang et al. have reported on the
thermal stabilities of uniform silver clusters prepared on oxi-
dized silicon and aluminum surface in oxidizing and reducing
atmospheres [25]. The thermal stability of the silver clus-
ter was significantly lower under the oxidizing conditions,
and heating above 623–673 K under the oxidizing conditions
induced a migration of the silver cluster. Aoyama et al. have
also shown that oxidation induced a size growth of the par-
ticles. On the contrary, metallic silver particles were finely
dispersed by the reduction process [26]. For the Ag/SiO2
catalysts, it can also be obtained according to the character-
ization results that pretreatment at 500 ◦C in oxygen favors
the migration and aggregation of metal silver particles on the
support, while following pretreatment at 500 ◦C in hydrogen
favors the dispersion of the silver particles. TEM results also
confirm this conclusion. As shown in Fig. 4, the mean silver
particle size of the Ag/SiO2 catalyst treated with oxygen is
larger than that treated with hydrogen. The detailed expla-
nation about the aggregation of silver particle after oxygen
treatment at high temperatures can be found in Ref. [20].
Table 2
Characteristicsofsilver/silvercompoundsontheAg/SiO2 catalystpretreated
with different gases at 500 ◦C analyzed with XRD and TEM
Oxygen
Followed
by H2
Hydrogen
Followed
by O2
TEM
Particle size (nm)
11–18
6–10
XRD
State
Metal,
oxide
18
Metal
8
Metal
12
Metal
17
Particle size (nm)

26
Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
Fig. 5. Variations of CO conversion as a function of different SiO2/Al2O3
ratio of zeolite H-ZSM-5 at 50 ◦C.
conversion, the corresponding O2 selectivity and the reaction
temperature are presented. No CO conversion is observed
on the silver catalyst supported on the zeolite H-ZSM-5
(SiO2/Al2O3 = 24) and activated carbon during the detection
temperatures, moreover the oxygen is completely conversed
only at the temperature higher than 140 and 220 ◦C. The
catalytic activity of silver catalyst supported on different
supports follows: SiO2 > Al2O3 > CeO2 > TiO2 > MgO ꢀ H-
ZSM-5 (SiO2/Al2O3 = 24) and activated carbon. It has been
known from XRD patterns that a fine dispersion of silver
particles on the Al2O3, CeO2 and H-ZSM-5 is obtained and
larger silver particle is observed on the MgO and C supports.
So it is thought that the main reason for the higher catalytic
activity for CO selective oxidation over silver catalyst is
not due to the smaller or larger size of silver particles, and
should be related with the properties of support and the
interaction between support and silver particles.
The catalytic activities of silver catalyst supported on the
H-ZSM-5 with different SiO2/Al2O3 ratio are also tested, as
shown in Fig. 5. The best catalytic activity is obtained on the
silver catalyst supported on the pure-silica-zeolite H-ZSM-5,
and the activity decreases with the decreasing of SiO2/Al2O3
ratio. The Al2O3 in the zeolite easily interacts with the silver
species supported, and changes the properties of silver par-
ticles dispersed on the support. From the above results, we
can also get that the reaction for CO selective oxidation over
silver catalyst does not need the acidity of the support. Then
the MCM-41 and SBA-15 zeolites are also chosen to support
silver particles to examine the activity of silver catalyst and
the effect of support. The CO conversion of about 30% at
Fig. 4. TEM micrographs of Ag/SiO2 catalysts treated with oxygen (a), and
hydrogen (b) at 500 ◦C for 2 h.
3.2. Catalytic properties
3.2.1. Effects of supports on the activity and selectivity
Table 3 shows the catalytic activity of silver catalyst
supported on the different supports. The maximum CO
Table 3
Catalytic activities of silver catalysts supported on different supports
Catalysts
␥-Al2O3
CeO2
H-ZSM-5 (SiO2/Al2O3)=24)
C
MgO
SiO2
TiO2
He 500 ◦C/2 h
CCO (%)
S (%)
13
15.85
65
7.94
9.07
95
–
–
–
–
–
–
4.48
4.48
150
23.35
29.75
60
5.11
39.36
130
TCO (◦C)

Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
27
45–50 ◦C can be obtained, which is very similar with that of
silver catalyst supported on SiO2. Anyway, the silica and the
pure silica zeolites are better supports for silver catalyst. The
effects of support on the properties of silver species and the
catalytic activity need to be further investigated in the future.
In the following research, SiO2 is chosen as the support of
silver catalyst.
3.2.2. Effects of silver loading
The catalytic behaviors of the Ag/SiO2 catalysts with var-
ious Ag loadings for the selective oxidation of CO in H2
atmosphere are shown in Fig. 6. The same pretreatment is
Fig. 7. Maximum CO conversion and the corresponding selectivity over
Ag/SiO2 catalysts as a function of silver loading.
Fig. 8. The temperatures for 100% conversion of oxygen and the maximum
conversion of CO as a function of silver loading.
Fig. 6. Variations of O2 conversion (a), CO conversion (b), and selectivity
S (c) over Ag/SiO2 catalysts with different loadings as a function of the
reaction temperature. CO 1.0%, O2 0.5%, H2 balance. Flow rate: 50 ml/min,
pretreatment with He at 500 ◦C for 2 h.
Fig. 9. The reaction rate of silver catalysts with different loadings for CO
selective oxidation at 30 ◦C.

28
Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
done before the activity tests. No activity was found on the
SiO2 support for the selective oxidation of CO. Over all
catalysts, the O2 conversion increases monotonously with
the reaction temperature, and oxygen can be totally con-
sumed above a certain temperature. The oxygen conversion
at the same temperature also increases with the silver load-
ing. It is interesting to note that the CO conversion curves are
“volcano-shaped”onallofthecatalysts. Thetemperaturecor-
responding to maximum CO conversion decreases with the
increase of the silver loading. Furthermore, the maximum CO
conversion again increases with the silver loading, which is
consistent with the result of Table 1, namely the adsorption
amounts of CO and O2 increase with the silver loading. It is
found that there is little change for the oxygen and CO con-
Fig. 10. Conversion of CO and O2, and the selectivity towards CO2 vs. reaction temperature on 16% Ag/SiO2 catalysts pretreated with He at 500 ◦C for
different λ.

Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
29
versions over the catalysts with silver loadings higher than
16% (not shown here). On the other hand, the selectivity
for CO oxidation of the catalysts decreases with the reaction
temperature, which indicates that the consumption of oxygen
at high temperatures is mostly contributed by the oxidation
of hydrogen. Therefore, for CO selective oxidation over the
Ag/SiO2 catalysts, oxygen preferentially reacts with carbon
monoxide at low temperatures, but with hydrogen at high
temperatures. In other words, Ag/SiO2 catalysts seem to be
a kind of promising catalysts for the selective oxidation of
carbon monoxide at low temperatures.
The effects of silver loading on the activity of Ag/SiO2 cat-
alysts for selective CO oxidation can be seen more clearly in
the Figs. 7 and 8. Fig. 7 shows the maximum CO conversion
and corresponding selectivity over Ag/SiO2 catalysts with
different silver loadings. Both the maximum CO conversion
and the corresponding selectivity increase with the Ag load-
ing. The temperatures corresponding to the oxygen complete
conversion and the maximum CO conversion decrease with
the increase of the Ag loading, as shown in Fig. 8. The reac-
tion rates of Ag/SiO2 with different silver loading at 30 ◦C
are shown in Fig. 9. The reaction rate of the catalyst for CO
selective oxidation increases with silver loading at low tem-
peratures, however, the increase is not linearity. That is to say
that the silver species loaded on supports does not completely
exert the role of catalytic sites for CO selective oxidation. It
has been found that more bulk oxygen species form, and the
initial temperature of reaction for CO with surface oxygen
species decreases obviously for silver catalysts with higher
loading after treated with He at 500 ◦C [27]. When the silver
loading is increased, the ability of supplying active oxygen
speciesofsilvercatalystsincreasesobviously. Webelievethat
the activity increase of the silver catalyst should be related
with the structure (oxygen species) of silver particles.
Fig. 11. Dependencies of the CO conversion (a), and the selectivity (b) as a
function of O2 concentration at 45 ◦C. CO 1%, H2 77.08% and He balance.
sion increases from 43.72 to 61.88%. At high O2:CO ratio
(λ = 3) the CO conversion is 64.96%. However, no obvious
increase of CO conversion is observed for increasing oxy-
gen concentration further. Only 0.5% oxygen is necessary
for stoichiometric oxidation of 1% CO, but CO still remains
in the outlet gas even in case of the 3% oxygen on silver cat-
alysts. The CO is not converted completely even in excess of
oxygen. The selectivity is plotted against the oxygen concen-
tration in Fig. 11b. The selectivity is increased by reducing
the oxygen concentration at 45 ◦C, and it decreases sharply
when the oxygen concentration is increased to 1.5%. The
oxygen concentration of 1% in the feed gas is suitable to oxi-
dize CO in a large excess of hydrogen on Ag/SiO2 catalysts,
taking into consideration both the high CO conversion, the
high selectivity and the high safety for the oxidation of CO.
3.2.3. O2 effect
To avoid an accidental explosion, the feed gas consisted
of 1% CO, 77.08% H2, oxygen and He balance. Fig. 10
shows the CO conversion, oxygen conversion and selectiv-
ity for different λ as a function of temperature. In all cases,
there exists a maximum value for CO conversion, and oxygen
can be totally consumed above a certain temperature (50 ◦C).
The selectivity towards CO2 is not changed obviously at low
temperatures. However, the slopes of the curves for oxygen
conversion, CO conversion and the selectivity become larger
above 45 ◦C with increasing λ. That is to say that above cer-
tain temperature hydrogen can react with oxygen quickly, at
the same time the CO conversion and selectivity decrease
sharply with increasing oxygen concentration. Oxygen pref-
erentially reacts with hydrogen at high temperatures; more-
over it becomes stronger and more obvious with high oxygen
concentration. This means that most of the oxygen added is
used for the oxidation of hydrogen under such a condition.
Changes in the activities for the oxidation of CO on the
oxygen content in the feed gas at 45 ◦C are shown in Fig. 11.
When increasing the oxygen content to 1%, the CO conver-
3.2.4. Activity after pretreatment with H2, O2 and He at
500 ◦C
Fig. 12 shows the catalytic behaviors of the 16% Ag/SiO2
pretreated in different atmospheres (O2, H2, He) at 500 ◦C. It
can be clearly seen that the pretreatment conditions strongly

30
Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
It has been reported that larger silver particles were present
on Ag/ZrO2 with high loadings, but were not present at
lower silver contents, and the catalysts with high loadings
were more active for the complete oxidation of methane
[29]. Therefore, the small number of surface defects and the
low reactivity of the predominant (1 1 1) face could explain
the low activities of the catalysts, and larger silver particles
appear to be polycrystalline with a large number of defects
and grain boundaries, which increased the reaction activity
[29]. In this paper, the size of silver particles and activity
of catalysts also increase with the silver loading, moreover
XRD results indicate that the metal silver particle sizes for the
O2 pretreated and He pretreated catalysts are similar and are
much larger than that of the H2 pretreated catalyst (Fig. 3).
From the results above, it is believed that the polycrystalline
formation due to the aggregation of silver particles is in favor
of the reaction for CO selective oxidation and the increase
of the adsorption ability of the catalyst. On the other hand,
it has been reported that the formation of subsurface oxy-
gen species on the silver surface would result in an activity
increase of some reactions [30,31]. Moreover, in our research
we found the formation of the subsurface oxygen species
after pretreatment with oxygen or He at certain temperatures
and its higher activity [20], and also observed the increas-
ing of the amount of subsurface oxygen species with silver
loading [27]. Reports of strongly adsorbed subsurface oxy-
gen species associated with silver catalysts are prevalent in
the literature, and evidence that large silver particles favors
the formation of subsurface oxygen and oxygen can remain
occluded in large silver particles after prolonged evacua-
tion at 500 ◦C [32]. Therefore, small silver particles have
to aggregate into large one prior to the formation of the spe-
cial silver-oxygen compound. Higher oxygen pretreatment
temperature is required on the silver catalyst with low load-
ing, which possesses smaller silver particle size, in order to
obtain a higher reaction activity [20]. Interestingly, the deac-
tivated catalyst can be re-activated after re-pretreatment with
oxygen. The deactivation of silver catalyst after pretreatment
with H2 at high temperatures is mainly due to the structure
Fig. 12. O2 conversion (a), CO conversion (b), and selectivity S (c) over 16%
Ag/SiO2 catalysts treated with different gases as a function of the reaction
temperature.
influenced the activity and selectivity of the catalysts towards
the selective oxidation of CO. The O2 conversion can reach
100% for all catalysts treated under the three different gases.
However, the O2 pretreated and the He pretreated catalysts
exhibit higher activity and selectivity at low temperatures
than those of the H2 pretreated catalyst. Meima et al. have
reported that a severe deactivation of an alumina supported
silver catalyst for CO oxidation occurred after reduction at
400 ◦C and subsequent cooling in nitrogen [28]. Accord-
ingly, similar deactivation phenomenon of the Ag/SiO2 cat-
alyst treated with hydrogen at high temperatures can also be
observed.
Fig. 13. Stability of the 8% Ag/SiO2 catalyst for CO selective oxidation at
50 ◦C.

Z. Qu et al. / Journal of Molecular Catalysis A: Chemical 239 (2005) 22–31
31
change of silver particles compared with that after oxygen
References
pretreatment, and the absence of subsurface oxygen species.
The stability of the 8% Ag/SiO2 catalyst for the selective
oxidation of CO at a 50 ◦C is shown in Fig. 13. During this
study, all of the catalysts show good stabilities. Only an initial
activity decreasing within the first hour of the stream time was
observed. Then a steady state is acquired. These observations
of ours are consistent with some previous observations, in
which the deactivation processes were investigated during
the kinetic studies on CO selective oxidation on Au/␣-Fe2O3
and CO oxidation on Au/TiO2, and the authors concluded
that the steady states could only be attained after a period of
time [33,34].
[1] R.A. Lemons, J. Power Sources 29 (1990) 251.
[2] A.J. Appleby, F.R. Foulkes, Fuel Cell Handbook, van Nostrand Rein-
hold, New York, 1989.
[3] S. Gottesfeld, J. Pafford, J. Electrochem. Soc. 135 (1988)
2651.
[4] M.J. Kahlich, H.A. Gasteiger, R.J. Behm, J. Catal. 171 (1997)
93.
[5] M.M. Schubert, H.A. Gasteiger, R.J. Behm, J. Catal. 172 (1997)
256.
[6] S.H. Oh, R.M. Sinkevitch, J. Catal. 142 (1993) 254.
[7] H. Igarashi, H. Uchida, M. Watanabe, Chem. Lett. 29 (2000)
1262.
[8] G.K. Bethke, H.H. Kung, Appl. Catal. A 194 (2000) 43.
[9] H. Igarashi, H. Uchida, M. Suzuki, Y. Sasaki, M. Watanabe, Appl.
Catal. A 159 (1997) 159.
[10] M. Watanabe, H. Uchida, H. Igarashi, M. Suzuki, Chem. Lett. 24
(1995) 21.
4. Conclusions
[11] M.M. Schubert, V. Plzak, J. Garche, R.J. Behm, Catal. Lett. 76
(2001) 143.
[12] A. Manasilp, E.E. Gulari, Appl. Catal. B 37 (2002) 17.
[13] K. Tamaru, Appl. Catal. A 151 (1997) 167.
[14] J.G. Serafin, A.C. Liu, S.R. Seyedmonir, J. Mol. Catal. A 131 (1998)
157.
[15] G.G. Xia, Y.G. Yin, W.S. Willis, J.Y. Wang, S.L. Suib, J. Catal. 185
(1999) 91.
[16] S. Imamura, H. Yamada, K. Utani, Appl. Catal. A 192 (2000)
226.
[17] K. Song, S. Kang, S.D. Kim, Catal. Lett. 49 (1997) 65.
[18] C. Gu¨ldu¨r, F. Balikci, Int. J. Hydrogen Energy 27 (2002) 219.
[19] Z.P. Qu, M.J. Cheng, X.L. Dong, X.H. Bao, Catal. Today 93 (2004)
247.
[20] Z.P. Qu, M.J. Cheng, W.X. Huang, X.H. Bao, J. Catal. 229 (2005)
446.
[21] Z.P. Qu, S.T. Zhou, W.C. Wu, C. Li, X.H. Bao, Catal. Lett. 101
(2005) 21.
[22] P. Scherrer, Goettingen Nachr. 2 (1918) 98.
[23] S. Imamura, H. Yamada, K. Utani, Appl. Catal. A 192 (2000) 221.
[24] M.C. Kung, K.A. Bethke, J. Yan, J.H. Lee, H.H. Kung, Appl. Surf.
Sci. 121 (1997) 261.
[25] M.X. Yang, P.W. Jacobs, C. Yoon, L. Muray, E. Anderson, D.
Attwood, G.A. Somorfai, Catal. Lett. 45 (1997) 5.
[26] N. Aoyama, K. Yoshida, A. Abe, T. Miyadera, Catal. Lett. 43 (1997)
249.
[27] Z.P. Qu, M.J. Cheng, C. Shi, X.H. Bao, Chin. J. Catal. 23 (2002)
460–464.
[28] G.R. Meima, L.K. Knijff, A.J. Van Dillen, J.W. Geus, J. Chem. Soc.,
Faraday Trans. I 85 (1989) 293.
[29] Lj. Kundakovic, M. Flytzani-Stephanopoulos, Appl. Catal. A 183
(1999) 35.
The effects of the supports, silver loading, oxygen con-
centration and pretreatment conditions on the activity and
selectivity for CO selective oxidation are investigated in this
paper. SiO2 and the pure silica zeolites are found to be better
supportsforsilvercatalyst. TheactivityandselectivityforCO
oxidation as well as the CO maximum conversion increase
with the silver loading. XRD patterns show that the intensity
of the metallic silver peak and the size of the silver particles
increase with the silver loading. The CO conversion increases
with increasing oxygen concentration properly. However, the
CO is not consumed completely even in case of 3% oxygen
in the feed gas. The selectivity at low temperatures is not
changed obviously, but the selectivity above 45 ◦C decreases
with oxygen concentration sharply. Above a certain tem-
perature hydrogen reacts with oxygen easily and sharply. A
pretreatment process exerts a strong influence on the activi-
ties of these catalysts towards CO selective oxidation. Under
different pretreatment conditions, a change in the particle size
and structure ofsilver occurs, which could influence the activ-
ity for CO oxidation. We also find that oxidation induces the
aggregation of silver particles at high temperatures, whereas
metallic silver can be caused to be finely dispersed by the
reduction process. High activity and selectivity for CO oxi-
dation are obtained over an Ag/SiO2 catalyst pretreated under
oxygen atmosphere. The reversibility of changes in activity
and selectivity for CO oxidation indicates that the activity
and selectivity of silver catalysts might be related with the
structure of silver particles.
[30] C. Backx, J. Moolhuysen, P. Geenen, R.A. Van Santen, J. Catal. 72
(1981) 364.
[31] X. Bao, J. Barth, G. Lempfuhl, R. Schuster, Y. Uchida, R. Schlo¨gl,
G. Ertl, Surf. Sci. 284 (1993) 14.
[32] Y.L. Sandler, W.M. Hickam, in: W.M.H. Sachtler, G.C.A. Schuit,
P. Zwietering (Eds.), Proceeding of Third International Congress on
Catalysis, Amsterdam, vol. 1, 1964, p. 227.
[33] M.J. Kahlich, H.A. Gasteiger, R.J. Behm, J. Catal. 182 (1999)
430.
Acknowledgements
We are grateful to the Chinese Ministry of Science and
Technology and the Natural Science Foundation of China for
support of this work.
[34] M.A. Bollinger, M.A. Vannice, Appl. Catal. B 8 (1996) 417.