Promoting mechanism of potassium in preferential CO oxidation on Pt/Al2O3

Article abstract of DOI:10.1016/j.jcat.2007.09.001

The addition of potassium enhanced the activity of preferential CO oxidation on Pt/Al₂O₃. The additive effect of potassium weakened the interaction between CO and Pt, and also changed the CO adsorption site. FTIR observations under the PROX conditions suggested the presence of adsorbed species derived from O₂ and H₂ such as OH species, which can be related to the promoting effect of H₂ present in the PROX reaction.

Full text of DOI:10.1016/j.jcat.2007.09.001

Journal of Catalysis 252 (2007) 39–48
www.elsevier.com/locate/jcat
Promoting mechanism of potassium in preferential CO oxidation
on Pt/Al O3
2
a
a
a
b
c
Masatoshi Kuriyama , Hisanori Tanaka , Shin-ichi Ito , Takeshi Kubota , Toshihiro Miyao ,
c
a,∗
a,∗
Shuichi Naito , Keiichi Tomishige , Kimio Kunimori
a
Institute of Materials Science, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8573, Japan
b
Department of Material Science, Shimane University, Matsue 690-8504, Japan
c
Department of Applied Chemistry, Faculty of Engineering, Kanagawa University, 3-27-1 Rokkakubashi, Kanagawa-ku, Yokohama, Kanagawa 221-8686, Japan
Received 5 May 2007; revised 5 July 2007; accepted 4 September 2007
Available online 29 September 2007
Abstract
The addition of potassium enhanced the activity of preferential CO oxidation on Pt/Al O . The additive effect of potassium weakened the
2
3
interaction between CO and Pt, and also changed the CO adsorption site. FTIR observations under the PROX conditions suggested the presence of
adsorbed species derived from O and H such as OH species, which can be related to the promoting effect of H present in the PROX reaction.
2
2
2
©
2007 Elsevier Inc. All rights reserved.
Keywords: Preferential CO oxidation; Platinum; Al O ; Potassium
2
3
1
. Introduction
Hydrogen is a promising energy carrier associated with fuel
ported [6–8,34,36–39]. Pt clusters supported on Cs-modified
SiO2 showed higher turnover rate and selectivity than clusters
of similar size dispersed on SiO and Al O ; the modifica-
2
2
3
cell technology [1], where one hydrogen production method is
the steam-reforming reaction of hydrocarbons and oxygenates.
In this case, the product of gases can contain CO as an impurity,
which can be considered a poison for the anode catalyst [2–4].
Preferential CO oxidation in H2-rich mixtures is an important
reaction in terms of the purification of fuel gas [5].
tion effect can be explained by the disruption of CO mono-
layer growth through promotion of the formation of unreactive
chemisorbed carbon via CO dissociations and disproportion-
ation [6]. In contrast, in the case of Pt/Al2O3 modified with
potassium, the adsorption site of CO on Pt was changed dras-
tically, and the interaction between CO and Pt became weaker
with the addition of potassium [51]. Our group has reported
that the optimized K–Pt/Al2O3 catalyst was very effective in
decreasing CO concentration to <10 ppm in fuel gas; in partic-
ular, the activity of CO oxidation was promoted drastically by
the presence of H2 [7,8]. In the present work, we investigated
the effect of potassium on catalytic properties in preferential
CO oxidation by means of transmission electron microscopy
(TEM), extended X-ray adsorption fine structure (EXAFS), X-
ray near-edge structure (XANES), and in situ Fourier transform
infrared spectroscopy (FTIR). In particular, we attempted to
elucidate the promoting mechanism of potassium on Pt/Al2O3
in preferential CO oxidation.
There have been many reports on supported noble metal
catalysts, such as Pt [6–33], Pd [15,34], Ru [15,35] and Rh
[
36–39], for preferential CO oxidation in H2-rich gas. In par-
ticular, the effect of various additives has been investigated
6,11,18,19,24,25,27]. In addition, Au [40–46] and Cu [47–50]
[
catalysts also exhibit high performance, which generally can be
explained by the weaker interaction with CO compared with
noble metal catalysts. Recently, the addition of alkali metal to
noble metal catalysts in preferential CO oxidation has been re-
*
Corresponding authors. Faxes: +81 29 853 4490; +81 29 853 5030.
E-mail addresses: tomi@tulip.sannet.ne.jp (K. Tomishige),
kunimori@ims.tsukuba.ac.jp (K. Kunimori).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.09.001

40
M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
2
. Experimental
2.3. Catalyst characterization
2
.3.1. Measurement of adsorption amount of CO and H2
The catalysts were characterized by the amounts of H2 and
2
.1. Catalyst preparation
CO adsorption. The amount of the irreversible adsorption of H2
and CO (H/Pt, CO/Pt) was measured at room temperature in a
vacuum system (sample weight, 0.15 g; dead volume, 65 cm )
by the volumetric method [38]. The samples are pretreated in
O2 at 773 K for 1 h, followed by H2 reduction at 773 K for 1 h
at atmospheric pressure. The equilibrium pressure of H2 and
CO was about 1.0 kPa.
The Al2O3 support (JRC-ALO-4 from Japan Reference Cat-
2
alyst [JRC]; BET surface area, 170 m /g) was used as support
materials. Before impregnation, Al2O3 was calcined for 3 h in
air at 873 K. Pt/Al2O3 catalyst was prepared by impregnat-
ing the Al2O3 with an aqueous solution of Pt(NO2)2(NH3)2
Soekawa Chemical Co., Ltd.). After impregnation, the sam-
ple was dried at 383 K for 12 h and then calcined in air at
73, 823, and 873 K for 3 h. The calcination temperature
3
(
7
2
.3.2. TEM observation
of the catalysts is denoted in an angle bracket, for example,
Pt/Al2O3 [773]. The loading amount of Pt was 2 wt%. The
Pt/Al2O3 catalysts modified with potassium also were prepared,
and the modification method was as follows. After the impreg-
nation with the aqueous solution of Pt(NO2)2(NH3)2, the sam-
ple was dried at 383 K for 12 h. After further impregnation
of the dried sample with an aqueous solution of KNO3 (Wako
Pure Chemical Industries, Ltd.), it was dried again at 383 K
for 12 h and calcined at 773 K for 3 h. The loading amount
of potassium is described in the molar ratio to Pt (K/Pt = 10).
The Pt/Al2O3 catalysts modified with potassium are denoted
as K–Pt/Al2O3 and the molar ratio of K/Pt is shown in paren-
theses like K–Pt/Al2O3 (10). The catalysts were reduced with
hydrogen at 773 K for 1 h in the reactor before the activity
test.
Transmission electron microscope (TEM) images were ob-
tained using a JEOL JEM-2010F microscope operated at
00 kV. First, the catalysts were reduced by H2 pretreatment at
73 K for 1 h in a fixed-bed reactor. After the reduction, sam-
2
7
ples were stored under vacuum until the measurements were
made. Samples were dispersed in 2-propanol using supersonic
waves, and they were put on Cu grids for TEM observation
under air atmosphere. Average particle size (d) is calculated
ꢀ
ꢀ
3
i
2
by d =
n d / nidi (n , number of pieces; d , particle
i
i
i
size) [52].
2.3.3. EXAFS and XANES
Pt L -edge EXAFS and XANES spectra was measured at
3
the BL-12C station of the Photon Factory at the High-Energy
Accelerator Research Organization (proposal 2005G041). The
storage ring was operated at 2.5 GeV. A Si(111) single crystal
was used to obtain a monochromatic X-ray beam. The mono-
chromator was detuned to 60% maximum intensity to avoid
higher harmonics in the X-ray beam. Two ion chambers filled
with Ar and 15% Ar diluted N2 for Pt L3-edge EXAFS were
used as detectors of I and I0, respectively. The samples for
the measurement were prepared by pressing 200 mg of cata-
lyst powder to disks. The thickness of the samples was chosen
as 0.6–0.7 mm (10 mm ϕ), to give an edge jump of 0.7. The
samples were pretreated at 773 K with H2 for 1 h. After the
pretreatment, we transferred the samples to the measurement
cell without exposing the sample disk to air, using a glove box
filled with nitrogen. EXAFS and XANES data were collected
in transmission mode at room temperature. For EXAFS analy-
sis, the oscillation was first extracted from the EXAFS data by a
spline smoothing method [53]. The oscillation was normalized
by the edge height at ca. 50 eV. The Fourier transformation of
2
.2. Activity test of preferential CO oxidation and related
reactions
Preferential CO oxidation in H2-rich gas was carried out in
a fixed-bed flow reaction system at atmospheric pressure using
3
1
(
00 mg of the catalyst at the total flow rate of 100 cm /min
−
1
STP) (GHSV = 30,000 h ). The feed stream contained 0.2%
CO, 0.2% O2, and 75% H2, and it was balanced with helium.
The effluent gas was analyzed using an online gas chromato-
graph (GC) system equipped with a thermal conductivity de-
tector. The activity was evaluated by CO and O2 conversions,
which can be calculated on the basis of CO and O2 concen-
trations in the reactant gas and the effluent gas. The selectivity
of CO oxidation is defined as the ratio of O2 consumption for
the CO oxidation to the total O2 consumption. The preferen-
tial CO oxidation in H2-rich stream is denoted as PROX here.
As a reference, we also carried out the activity test of CO ox-
idation in the absence of H2. Here the feed stream contained
3
the k -weighted EXAFS oscillation from k space to r space was
−
1
performed over the range of 30–160 nm to obtain a radial
distribution function. The inversely Fourier filtered data were
analyzed by a common curve-fitting method [54,55]. For the
curve-fitting analysis, the empirical phase shift and amplitude
functions for Pt–Pt and Pt–O bonds were extracted from the
data for Pt foil and Na Pt(OH) , respectively. In the analysis
0
.2% CO, 0.2% O2, and it was balanced with helium. This CO
oxidation in the absence of H2 is denoted as CO + O2. As an-
other reference, the activity of the water–gas shift reaction was
also tested; the feed contained 0.2% CO, 3% H2O, and balance
He. The steam is introduced by a gas-stripping method. In all
reactions, the activity was observed for 30 min at each condi-
tion. Because almost no deactivation was observed, the results
of the activity test correspond to those under the steady-state
condition.
2
6
of XANES spectra, the normalized spectra were obtained by
subtracting the pre-edge background from the raw data with
a modified Victoreen equation and normalizing them by the
edge height [56–59]. The EXAFS data were analyzed using
REX2000 version 2.3.3 (Rigaku Corp.).

M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
41
Fig. 1. Reaction temperature dependence of CO conversion (2, P) and selectivity of CO oxidation (F) in PROX and CO + O reactions. (a) K–Pt/Al O (10),
2
2 3
(
0
b) Pt/Al O [773], (c) Pt/Al O [823], (d) Pt/Al O [873] Reaction conditions: PROX: 0.2% CO, 0.2% O , 75% H , He balance (2, F); CO + O : 0.2% CO,
2
3
2
3
2
3
2
2
2
.2% O , He balance (P).
2
2
.3.4. FTIR measurements
FTIR spectra were recorded with a Nicolet Magna 550 spec-
the temperature range of 383–430 K. This behavior is useful,
because it avoids the need for careful temperature control. In
particular, according our previous report, K–Pt/Al O (10) gave
−
1
trometer equipped with a MCT detector (resolution, 4 cm ) in
a transmission mode, using an in situ IR quartz cell with CaF2
windows. All samples for the IR measurements were pressed
into self-supporting 20-mm-diameter, 300-mg wafers. The sam-
ple was transferred to the IR cell connected to the closed cir-
culation systems and then reduced with H2 at 773 K for 1 h at
2
3
CO concentrations <10 ppm as a result of the PROX reaction
in the temperature range of 375–410 K [7]. As shown in Fig. 1a,
CO conversion was much higher in the PROX reaction than in
the CO + O reaction, suggesting that the presence of H sig-
2
2
nificantly promotes CO oxidation on K–Pt/Al O (10). In con-
2
3
1
3 kPa. After pretreatment, the sample was cooled to 313 K, and
trast, as shown in Fig. 1b, CO conversion in the PROX reaction
on Pt/Al2O3 [773] was lower than that in the CO + O2 reaction
at the reaction temperatures above 420 K. This can be related to
the low selectivity of CO oxidation in PROX on Pt/Al2O3 [773].
The amounts of CO and H2 adsorption on K–Pt/Al2O3 (10) and
three Pt/Al2O3 catalysts are given in Table 1. The adsorption of
the Pt/Al2O3 catalysts decreased gradually with increasing cal-
cination temperature due to aggregation of Pt metal particles.
Note that the CO adsorption of K–Pt/Al2O3 (10) was compara-
ble to that of Pt/Al2O3 [823]. Based on the result of CO adsorp-
tion, it is possible to estimate the turnover frequency (TOF) of
CO oxidation, as also given in Table 1. Regarding the Pt/Al2O3
catalysts, the TOF in PROX increased slightly with decreas-
ing adsorption and metal dispersion. This tendency agrees with
previously reported findings [33]. It has been reported that flat
surfaces have a significantly higher CO oxidation activity per
surface site compared with low-coordination edge and corner
atoms in the PROX reaction [33]. In addition, it is characteristic
that K–Pt/Al2O3 (10) gave much higher TOF of CO oxidation
the gases (CO, CO+O2, CO+O2 +H2) were introduced to the
IR cell connected to the flow and evacuating system. Temper-
ature dependence of the FTIR spectra during the reactions was
measured for 15 min at each temperature. The FTIR spectra of
adsorbed species were obtained by subtracting the spectra be-
fore the reaction at the same temperature. The total gas flowing
−
1
rate in all of the reactions was adjusted to GHSV = 30,000 h
,
which was used for the activity tests in the fixed-bed reactor.
3
. Results and discussion
3
.1. Catalytic performance in PROX and CO+O2
Fig. 1 shows the effects of reaction temperature on the per-
formance of the PROX and CO + O2 reactions. K–Pt/Al2O3
10) showed much higher activity and selectivity than all of the
(
Pt/Al2O3 catalysts in the PROX reaction [7,8]. The CO conver-
sion in the PROX on K–Pt/Al2O3 (10) reached almost 100% in

42
M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
Table 1
Adsorption amount of H , CO and TOF of CO formation over K–Pt/Al O and Pt/Al O catalysts at 363 K
2
2
2
3
2 3
−
3
−1 a
s )
Catalyst
Calcination
temperature (K)
Adsorption amount
b
TOF (×10
b
c
d
e
H/Pt
CO/Pt
CO + O2
CO + O + H
CO + H O
1.3
0.03
–
2
2
2
f
K–Pt/Al O (10)
773
773
823
873
0.38
0.70
0.53
0.14
0.40
0.52
0.37
0.14
5.4
1.0
–
32
2
3
Pt/Al O [773]
1.9
2.2
2.6
2
3
Pt/Al O [823]
2
3
Pt/Al O [873]
–
–
2
3
a
b
c
d
e
f
TOF of CO oxidation is calculated on the basis of CO/Pt.
Irreversible adsorption of H and CO at room temperature.
2
0
0
0
.2% CO, 0.2% O , He balance.
2
.2% CO, 0.2% O , 75% H , He balance.
2
2
.2% CO, 3% H O, He balance.
2
The TOF was calculated by an extrapolating method in the Arrhenius plot at the temperature range of 323–353 K because K–Pt/Al O (10) gave 88% CO
2
3
conversion at 363 K.
in PROX than the Pt/Al2O3 catalysts. Furthermore, the TOF in
PROX was much higher than that in CO + O2 on K–Pt/Al2O3
(
10). One of the promoting mechanisms of H2 to CO oxidation
may be due to the water–gas shift reaction originating from H2
oxidation to H2O on the basis of previous reports [9,23–27].
Therefore, the activity of the water–gas shift reaction was also
measured; the TOFs obtained are given in Table 1. The TOFs of
the water–gas shift reaction were much lower than those in the
PROX and CO + O2 reactions, and the promotion of H2 to CO
oxidation cannot be explained by the water–gas shift reaction
in the present case.
3
.2. Catalyst characterization by TEM, EXAFS, and XANES
Fig. 2 shows TEM images of the Pt/Al2O3 [773] and K–Pt/
Al2O3 (10) catalysts. The average particle sizes of Pt/Al2O3
773] and K–Pt/Al2O3 (10) were determined to be 0.5 ± 0.1
and 2.0 ± 0.3 nm, respectively. The tendency for K–Pt/Al2O3
10) to have lower dispersion than Pt/Al2O3 [773] agrees well
[
(
with the results on adsorption (Table 1).
Fig. 3 shows the result of Pt L3-edge EXAFS analysis of
these catalysts, and Table 2 gives curve-fitting results. The
Pt–Pt and Pt–O bonds were required for the curve fitting of
Pt/Al2O3 [773] and [823], although the contribution of the Pt–O
bond was rather small. This suggests that highly dispersed Pt
metal particles interact with the surface of the Al2O3 support.
The coordination number of the Pt–Pt bond on Pt/Al2O3 in-
creased monotonously with increasing calcination temperature.
We conclude that aggregation of Pt metal particles occurs by
calcination at higher temperatures. The coordination number of
the Pt–Pt bond on K–Pt/Al2O3 (10) was located between that
of Pt/Al2O3 [773] and [823], and the tendency was similar to
that of the adsorption amount of CO. This may be related to
the neutralization of the acidic properties of Al2O3, which can
contribute to the enhanced metal dispersion.
Fig. 2. TEM images of (a) fresh Pt/Al O [773] and (b) K–Pt/Al O (10) after
H reduction.
2
2
3
2 3
Fig. 4 shows the results of Pt L3-egde XANES spectra, and
Table 3 gives the results of the XANES analysis. The white
line intensity of Pt L3-edge is known to be an informative in-
dication of the electronic state of Pt; the larger white line is
due to the greater electron vacancy in d-orbital [60]. As re-
ported previously, a relative electron deficiency of the Pt species
can be determined based on the white line intensity [56–59].
For the Pt/Al2O3 catalysts, the relative electron deficiency de-
creased gradually with increasing calcination temperature. This
may be associated with the decreased contribution of the Pt–O
bond in EXAFS results as well as the decreased metal disper-
sion. On the other hand, the relative electron deficiency of K–
Pt/Al2O3 (10) was positioned between that of Pt/Al2O3 [773]
and [823].

M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
43
3
Fig. 3. Results of Pt L -edge EXAFS analysis of K–Pt/Al O (10) and Pt/Al O catalysts after the H reduction. (I) k -weighted EXAFS oscillations. (II) Fourier
3
2
3
2
3
2
3
−1
transforms of k -weighted Pt L -edge EXAFS, FT range: 30–160 nm . (III) Fourier filtered EXAFS data (—) and calculated data (- - -). Fourier filtering range:
3
0
.16–0.31 nm: (a) K–Pt/Al O (10), (b) Pt/Al O [773], (c) Pt/Al O [823], (d) Pt/Al O [873].
2 3 2 3 2 3 2 3
Table 2
a
Curve fitting results of Pt L -edge EXAFS of K–Pt/Al O (10) and Pt/Al O catalysts after H reduction
3
2
3
2
3
2
b
−1
c
−1
d
e
f
Catalysts
Shells
C.N.
R (10 nm)
σ (10 nm)
ꢀE (eV)
R (%)
f
0
Pt/Al O [773]
Pt–Pt
Pt–O
5.9 ± 0.7
2.65 ± 0.007
0.094 ± 0.007
−4.7 ± 1.8
2
3
0.9 ± 0.2
2.03 ± 0.026
0.065 ± 0.044
−5.7 ± 7.1
0.94
Pt/Al O [823]
Pt–Pt
Pt–O
7.6 ± 0.6
0.6 ± 0.5
2.78 ± 0.002
2.12 ± 0.030
0.093 ± 0.002
0.092 ± 0.066
6.5 ± 0.5
2.0 ± 6.9
2
3
1.0
Pt/Al O [873]
Pt–Pt
Pt–Pt
9.5 ± 0.4
8.5 ± 0.3
2.78 ± 0.002
2.74 ± 0.003
0.082 ± 0.001
0.080 ± 0.012
1.6 ± 0.6
−0.2 ± 0.8
0.66
0.69
2
3
K–Pt/Al O (10)
2
3
a
b
c
d
e
f
Sample pretreatment: reduction (H , 773 K, 1 h).
Coordination number.
Bond distance.
2
Debye–Waller factor.
Difference in the origin of photoelectron energy between the reference and the sample.
Residual factor, Fourier filtering range: 0.16–0.38 nm.
Table 3
Results of Pt L -edge XANES analysis of K–Pt/Al O (10) and Pt/Al O cat-
3
2
3
2 3
alysts
a
b
c
ꢀA/A
Catalysts
ꢀA
A
K–Pt/Al O (10)
0.78
0.83
0.66
0.58
0.0
–
–
–
–
0.102
0.108
0.085
0.074
0
2
3
Pt/Al O [773]
2
3
Pt/Al O [823]
2
3
Pt/Al O [873]
2
3
Pt foil
7.70
a
Area difference between catalysts and Pt-foil (11540–11566 eV).
White line area of Pt foil.
Relative electron deficiency.
b
c
To facilitate discussion of the electronic state of the catalysts
considering the Pt metal particle size, Fig. 5 shows the rela-
tionship between the coordination number of the Pt–Pt bond
obtained from EXAFS and the relative electron deficiency from
XANES. In the Pt/Al2O3 catalysts, the relative electron defi-
ciency decreased gradually with increasing coordination num-
Fig. 4. Pt L -edge XANES spectra of K–Pt/Al O (10) and Pt/Al O catalysts:
3
2
3
2 3
(
a) K–Pt/Al O (10), (b) Pt/Al O [773], (c) Pt/Al O [823], (d) Pt/Al O
2 3
ber of the Pt–Pt bond. In contrast, K–Pt/Al O (10) is far differ-
ent than the relationship of the Pt/Al2O3 catalysts, which means
2
3
2
3
2
3
2
3
[873].

44
M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
Fig. 5. Relation between the relative electron deficiency from XANES and the
coordination number of the Pt–Pt bond from EXAFS of K–Pt/Al O (10) and
2
3
Pt/Al O catalysts.
2
3
that Pt metal particles in K–Pt/Al2O3 (10) have a greater elec-
tron deficiency for the large Pt–Pt coordination number, pos-
sibly due to the modification of potassium. The additive effect
of alkali metal to metal surfaces can be explained by electron
transfer from the alkali metal to the metal surface [61–63]. This
tendency is opposite to the results given earlier. The difference
is due to the chemical state of alkali metal; here potassium is
present as a cation, not metallic. The chemical state of the potas-
sium was investigated by TPR with H2. In the profile, the H2
consumption peak due to the reduction of potassium ions to
metal species was not observed, indicating that the potassium
was not metallic even after H2 reduction.
3
.3. Behavior of adsorbed CO during the reaction using in
situ FTIR
Fig. 6 shows the effect of evacuation temperature on IR spec-
tra of CO adsorption over K–Pt/Al2O3 (10) and Pt/Al2O3 [773].
On K–Pt/Al2O3 (10), three peaks were observed. The peaks at
−
1
2
059 and 1979 cm can be assigned to linear and bridge CO on
−
1
Pt, respectively [51]. Note that the peak observed at 1760 cm
was not observed on Pt/Al2O3; it has been suggested that this
peak can be assigned to a threefold-coordinated CO species on
the Pt atoms interacting with potassium species [51]. Linear,
bridge, and threefold CO are denoted as L-CO, B-CO, and T-
CO, respectively. The changes in the peak intensities are plotted
in Fig. 6III. Almost all of the CO adsorbed on K–Pt/Al2O3 (10)
was desorbed at 513 K; in contrast, a significant amount of CO
remained on Pt/Al2O3 [773] even at 633 K. This indicates that
the interaction of CO with K–Pt/Al2O3 (10) is much weaker
than that with Pt/Al2O3 [773]. This finding agrees with previ-
ously reported results [51] and also can be related to the higher
relative electron deficiency of K–Pt/Al2O3 (10).
Fig. 7 shows the temperature dependence of IR spectra of
K–Pt/Al2O3 (10) during the PROX and CO + O2 reactions and
the introduction of CO. The spectrum during the introduction
of CO at 313 K was almost the same as that shown in Fig. 6I,
but with the difference that CO can be adsorbed even at 453 K
Fig. 6. Effect of evacuation temperature on IR spectra of adsorbed CO on
(I) K–Pt/Al2O3 (10) and (II) Pt/Al2O3 [773] catalysts, and (III) temperature
dependence of total area of the CO adsorption peaks on the catalysts. The sam-
ples were exposed to 0.3 kPa CO at 313 K and evacuated.
in the presence of gas-phase CO (Fig. 7c). The peak intensi-
ties of L-CO, B-CO, and T-CO are plotted in Fig. 7f. The order
of the peak intensity in this temperature range was as follows:
T-CO > B-CO > L-CO. In contrast, in the CO + O2 reaction
(Fig. 7b), the shape of the spectra differed from that for the in-
troduction of CO. As shown in Fig. 7e, the order of the peak

M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
45
Fig. 7. Effect of temperature on IR spectra of K–Pt/Al O (10) catalyst during PROX (a), CO + O (b) reactions and CO (c) introduction, (d–f) area of each peak
2
3
2
as a function of temperature. Reaction conditions: PROX: 0.2% CO, 0.2% O , 75% H , He balance; CO + O : 0.2% CO, 0.2% O , He balance; CO: 0.2% CO,
2
2
2
2
He balance.
to higher wavenumber (about 20–60 cm 1) during the CO+O2
reaction compared with during the introduction of CO. Ac-
cording to the previous report, adsorbed oxygen atoms occupy
the threefold hollow sites on the Pt(111) surface [64]. In the
−
intensity at 313–373 K was B-CO > L-CO ≈ T-CO. The inten-
sity of T-CO decreased and that of B-CO increased significantly
compared with that seen during the introduction of CO. An-
other important point is that all these three peaks were shifted

46
M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
Fig. 8. Effect of temperature on IR spectra of Pt/Al O catalyst during PROX (a), CO + O (b) reactions and CO (c) introduction, (d) area of the linear CO peak
2
3
2
as a function of temperature. (1) L-CO(PROX), (P) L-CO(CO + O ), (E) L-CO(CO). Reaction conditions: PROX: 0.2% CO, 0.2% O , 75% H , He balance;
2
2
2
CO + O : 0.2% CO, 0.2% O , He balance; CO: CO 0.2%, He balance.
2
2
coadsorption of oxygen and CO atoms, oxygen atoms are ad-
sorbed on the threefold site and CO atoms are adsorbed mainly
on the on-top site over Pt(111), Pd(111), and Rh(111) [64,65].
The coadsorption significantly affects the site of CO adsorp-
tion, suggesting that the CO adsorption site can be shifted to
a site with lower coordination number. This can explain the
change in adsorption site during the CO + O2 reaction. In addi-
tion, the peak position of adsorbed CO was shifted to a higher
wavenumber by the coordination of oxygen atoms [65]. From
the comparison of the CO + O2 reaction and the introduction
of CO, the peak shift and the ratio change in the CO adsorption
sites suggest that coadsorption of CO and oxygen occurred dur-
Figs. 7a and 7b show the behavior during this PROX reac-
tion. At 313 K, the spectrum of PROX was similar to that of
CO + O2, with the order B-CO > L-CO ≈ T-CO. However, the
shift of B-CO and T-CO was smaller in PROX than in CO+O2.
Note that the total peak area (L-CO+B-CO+T-CO; Fig. 7d) of
PROX decreased slightly with increasing reaction temperature,
and at 313–373 K, the decrease in PROX was more significant
than that in CO + O2. As shown in Fig. 7c, CO can be adsorbed
on K–Pt/Al2O3 (10) when CO is present in the gas phase. This
result suggests the presence of another adsorbed species with
considerable coverage during the PROX reaction. Considering
that the peak shift of PROX is smaller than that of CO+O , the
2
ing the CO+O reaction. In addition, the increased absorbance
adsorbed species can have less ability to accept electrons from
2
−1
at around 1700 cm may be due to adsorbed carbonate species
originating from CO2 produced by CO oxidation. Furthermore,
in the case of CO + O2, the peaks disappeared suddenly above
Pt than oxygen atoms. It is expected that adsorbed species can
be formed in the simultaneous presence of H2 and O2, and that
candidates of the species are H2O2 and OOH. However, consid-
ering that Pt is usually unsuitable for the production of H2O2
3
93 K (Fig. 7b). This represents the increased oxygen cover-
age and decreased CO coverage at higher reaction temperatures,
which is expected based on the kinetics of CO oxidation with
O [66,67].
2
2 2
from H and O , these species are unlikely. Another plausible
candidate is hydroxide species (OH). The OH group may be
observed in FTIR spectra in higher wavenumber regions; how-

M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
47
ever, such observation is difficult, because the peaks due to H2O
a byproduct of the PROX reaction) overlap, and thus we do not
4. Conclusion
(
yet have direct evidence of the OH group on Pt. The species and
the reaction mechanism are discussed in more detail later.
Fig. 8 shows the temperature dependence of IR spectra of
Pt/Al O [773] during the PROX and CO + O reactions and
The addition of potassium to Pt/Al2O3 significantly en-
hanced the activity of preferential CO oxidation in an H2-rich
stream. In particular, the presence of H2 significantly promoted
CO oxidation. According to the catalyst characterization by
TEM and EXAFS, K–Pt/Al2O3 (10) had ca. 2-nm-diameter
metal particles. Pt L3-edge white line analysis revealed that
Pt metal particles on K–Pt/Al2O3 were electron-deficient com-
pared with those on the Pt/Al2O3 with similar metal parti-
cle sizes. FTIR observation of CO-TPD demonstrated much
weaker CO adsorption on K–Pt/Al2O3 than on Pt/Al2O3, possi-
bly related to the electron deficiency of K–Pt/Al2O3. The FTIR
spectra suggest that CO was adsorbed on the bridge and three-
fold hollow sites as well as on the on-top site on K–Pt/Al2O3.
This behavior differed from that on Pt/Al2O3. In situ FTIR ob-
servations under the PROX conditions on K–Pt/Al2O3 at the
low temperatures at which the PROX reaction proceeds indi-
cated less CO adsorption compared with that in the CO + O2
reaction and the introduction of CO. This suggests that the
adsorbed species originating from H2 and O2 (e.g., the OH
species) could be present on the surface and could promote CO
oxidation.
2
3
2
the introduction of CO. For Pt/Al2O3 [773], the spectra under
these three conditions seemed to be similar below 433 K. This
means that the coverage of adsorbed CO was close to the satu-
ration level, and the coverage of coadsorbed species was rather
low. The very small differences in the peak positions support
this interpretation. At temperatures above 443 K in the CO+O2
reaction, the peaks disappeared, as they did for K–Pt/Al2O3
(
10), although the temperature on Pt/Al2O3 [773] was much
higher than that on K–Pt/Al2O3 (10). This can be explained by
the stronger interaction of CO with Pt/Al2O3 [773], as shown
in Fig. 6.
3
.4. Promoting effect of potassium on preferential CO
oxidation over Pt/Al O
2
3
The addition of potassium weakened the interaction of Pt
and CO, possibly due to the greater electron deficiency of the
Pt species. This can contribute to high activity at lower reaction
temperatures. Another important finding is that the adsorption
References
[
1] R. Farrauto, S. Hwang, L. Shore, W. Ruettinger, J. Lampert, T. Giroux,
Y. Liu, O. Ilinich, Annu. Rev. Mater. Res. 33 (2003) 1.
2] B. Rohland, V. Plzak, J. Power Source 84 (1999) 183.
3] J. Divisek, H.-F. Oetjen, V. Peinecke, V.M. Schmidt, U. Stimming, Elec-
trochim. Acta 43 (1998) 3811.
site of CO was changed drastically. On K–Pt/Al O (10), bridge
2
3
and three-fold hollow CO species were clearly observed. The
reason for the adsorption site change is not clear at present,
but it may be related to the change in the electronic state of
Pt modified by potassium. Based on the results of in situ FTIR
[
[
[4] C. Song, Catal. Today 77 (2002) 17.
[
[
[
5] D.L. Trimm, Appl. Catal. A 296 (2005) 1.
observation, the coverage of CO on K–Pt/Al O (10) during
2
3
6] C. Pedrero, T. Waku, E. Iglesia, J. Catal. 233 (2005) 242.
7] Y. Minemura, S. Ito, T. Miyao, S. Naito, K. Tomishige, K. Kunimori,
Chem. Commun. (2005) 1429.
the PROX reaction was decreased, suggesting the presence of
another coadsorbed species than oxygen atom. This species can
promote CO oxidation. One possible candidate is the OH group,
which can be formed by adsorbed hydrogen and oxygen atoms.
It has been reported that the OH group promoted CO oxidation
on Pt(111) [63,68], as described below:
[8] Y. Minemura, M. Kuriyama, S. Ito, K. Tomishige, K. Kunimori, Catal.
Commun. 7 (2006) 623.
[9] A. Manasilp, E. Gulari, Appl. Catal. B 37 (2002) 17.
[10] H. Igarashi, H. Uchida, M. Suzuki, Y. Sasaki, M. Watanabe, Appl. Catal.
A 159 (1997) 159.
[
11] I.H. Son, A.M. Lane, Catal. Lett. 76 (2001) 151.
Had + Oad → OHad
and
(1)
[12] X. Liu, O. Korotkikh, R. Farrauto, Appl. Catal. A 226 (2002) 293.
13] A. Sirijaruphan, J.G. Goodwin Jr., R.W. Rice, J. Catal. 221 (2004) 288.
[14] M.J. Kahlich, H.A. Gasteiger, R.J. Behm, J. Catal. 171 (1997) 93.
[
[15] S.H. Oh, R.M. Sinkevitch, J. Catal. 142 (1993) 254.
[16] O. Korotkikh, R. Farrauto, Catal. Today 62 (2000) 249.
[17] A. Fukuoka, M. Ichikawa, Top. Catal. 40 (2006) 103.
[18] M. Kotobuki, A. Watanabe, H. Uchida, H. Yamashita, M. Watanabe,
J. Catal. 236 (2005) 262.
COad + OHad → CO2g + Had.
(2)
These reaction formulas correspond to the reaction route via
OH species as an autocatalytic mechanism [68]. In fact, the
presence of H2 enhanced the CO oxidation activity on Pt/Al2O3
at reaction temperatures below 420 K (Fig. 1b); however, the
effect was as significant as that on K–Pt/Al2O3 (10) (Fig. 1a).
This difference in the promoting effect of H2 may be due to
the difference in coverage of the OH species. Potassium plays
a very important role in the enhancement of OH coverage, for
[
19] M. Kotobuki, A. Watanabe, H. Uchida, H. Yamashita, M. Watanabe, Appl.
Catal. A 307 (2006) 275.
[
[
20] A. Sirijaruphan, J.G. Goodwin Jr., R.W. Rice, J. Catal. 224 (2004) 304.
21] M.M. Schubert, M.J. Kahlich, G. Feldmeyer, M. Huttner, S. Hackenberg,
H.A. Gasteiger, R.J. Behm, Phys. Chem. Chem. Phys. 3 (2001) 1123.
22] C. Kwak, T.J. Park, D.J. Suh, Appl. Catal. A 278 (2005) 186.
[
[23] E. Simsek, S. Ozkara, A.E. Aksoylu, Z.I. Onsan, Appl. Catal. A 316
(2006) 169.
[24] M. Shou, K. Tanaka, Catal. Lett. 111 (2006) 115.
−
+
example, by a Coulomb interaction between OH and K . In
addition, judging from the low activity of the water–gas shift
reaction, this active OH cannot be formed from H O in the gas
[25] S.H. Cho, J.S. Park, S.H. Choi, S.H. Kim, J. Power Sources 156 (2006)
260.
2
[26] J.L. Ayastuy, M.P. Gonzalez-Marcos, A. Gil-Rodriguez, J.R. Gonzalez-
Velasco, M.A. Gutierrez-Ortiz, Catal. Today 116 (2006) 391.
phase.

48
M. Kuriyama et al. / Journal of Catalysis 252 (2007) 39–48
[
[
27] J.L. Ayastuy, M.P. Gonzalez-Marcos, J.R. Gonzalez-Velasco, M.A. Gu-
tierrez-Ortiz, Appl. Catal. B 70 (2007) 532.
28] J.L. Ayastuy, A. Gil-Rodriguez, M.P. Gonzalez-Marcos, M.A. Gutierrez-
Ortiz, Int. J. Hydrogen Energy 31 (2006) 2231.
[46] C. Rossignol, S. Arrii, F. Morfin, L. Piccolo, V. Caps, J. Rousset, J. Ca-
tal. 239 (2006) 307.
[47] Y. Tanaka, T. Utaka, R. Kikuchi, K. Sasaki, K. Eguchi, Appl. Catal. A 238
(2003) 11.
[
[
[
29] G. Uysal, A.N. Akin, Z.I. Onsan, R. Yildirim, Catal. Lett. 108 (2006) 193.
30] I.H. Son, J. Power Sources 159 (2006) 1266.
31] S. Monyanon, S. Pongstabodee, A. Luengnaruemitchai, J. Power
Sources 163 (2006) 547.
32] E.Y. Ko, E.D. Park, K.W. Seo, H.C. Lee, D. Lee, S. Kim, Catal. Today 116
[48] J. Papavasiliou, G. Avgouropoulos, T. Ioannides, Appl. Catal. B 66 (2006)
167.
[49] Y. Liu, Q. Fu, M.F. Stephanopoulos, Catal. Today 93–95 (2004) 241.
[50] J.W. Park, J.H. Jeong, W.L. Yoon, C.S. Kim, D.K. Lee, Y. K Park, Y.W.
Rhee, Int. J. Hydrogen Energy 30 (2005) 209.
[
(
2006) 377.
[51] S. Derrouiche, P. Gravejat, B. Bassou, D. Bianchi, Appl. Surf. Sci. 253
(2007) 5894.
[
[
[
33] B. Atalik, D. Uner, J. Catal. 241 (2006) 268.
34] N. Iwasa, S. Arai, M. Arai, Catal. Commun. 7 (2006) 839.
35] Y.F. Han, M.J. Kahlich, M. Kinne, R.J. Behm, Phys. Chem. Chem. Phys. 4
[52] Y. Chen, K. Tomishige, K. Yokoyama, K. Fujimoto, Appl. Catal. A Gen.
165 (1997) 335.
(
2002) 389.
[53] J.W. Cook, D.E. Sayers, J. Appl. Phys. 52 (1981) 5024.
[54] K. Okumura, J. Amano, N. Yasunobu, M. Niwa, J. Phys. Chem. B 104
(2000) 1050.
[
[
[
[
[
[
[
[
[
[
36] H. Tanaka, S. Ito, S. Kameoka, K. Tomishige, K. Kunimori, Catal. Com-
mun. 4 (2003) 1.
37] H. Tanaka, S. Ito, S. Kameoka, K. Tomishige, K. Kunimori, Appl. Catal.
A 250 (2003) 255.
38] S. Ito, H. Tanaka, Y. Minemura, S. Kameoka, K. Tomishige, K. Kunimori,
Appl. Catal. A 273 (2004) 295.
39] S. Ito, T. Fujimori, K. Nagashima, K. Yuzaki, K. Kunimori, Catal. To-
day 57 (2000) 247.
40] M.M. Schubert, V. Plzak, J. Garche, R.J. Behm, Catal. Today 76 (2001)
[55] K. Okumura, S. Matsumoto, N. Nishiaki, M. Niwa, Appl. Catal. B 40
(2003) 151.
[56] T. Kubota, K. Asakura, N. Ichikuni, Y. Iwasawa, Chem. Phys. Lett. 256
(1996) 445.
[57] T. Kubota, K. Asakura, Y. Iwasawa, Catal. Lett. 46 (1997) 141.
[58] A.N. Mansour, J.W. Cook, D.E. Sayers Jr., J. Phys. Chem. 88 (1984) 2330.
[59] J.A. Horsely, J. Chem. Phys. 76 (1982) 1451.
[60] H. Yoshida, Y. Yazawa, T. Hattori, Catal. Today 87 (2003) 19.
[61] L.F. Liotta, G.A. Martin, G. Deganello, J. Catal. 164 (1996) 322.
[62] N. Pavlenko, P.P. Kostrobij, Y. Suchorski, R. Imbihl, Surf. Sci. 489 (2001)
29.
[63] I.N. Yakovkin, V.I. Chernyi, A.G. Naumovets, Surf. Sci. 442 (1999) 81.
[64] K.L. Kostov, P. Jakob, D. Menzel, Surf. Sci. 377 (1997) 802.
[65] J. Libuda, H.-J. Freund, Surf. Sci. Rep. 57 (2005) 157.
[66] K. Nakao, S. Ito, K. Tomishige, K. Kunimori, J. Phys. Chem. B 109 (2005)
24002.
1
43.
41] B. Schumacher, Y. Denkwitz, V. Plzak, M. Kinne, R.J. Behm, J. Catal. 224
2004) 449.
42] R.M. Torres Sanchez, A. Ueda, K. Tanaka, M. Haruta, J. Catal. 168 (1997)
25.
(
1
43] F. Arena, P. Famulari, G. Trunfio, G. Borura, F. Frusteri, L. Spadaro, Appl.
Catal. B 66 (2006) 81.
44] M. Shou, H. Takekawa, D.Y. Ju, T. Hagiwara, D.L. Lu, K. Tanaka, Catal.
Lett. 108 (2006) 119.
45] M. Azar, V. Caps, F. Morfin, J. Rousset, A. Piednoir, J. Bertolini, L. Pic-
colo, J. Catal. 239 (2006) 307.
[67] T. Engel, G. Ertl, Adv. Catal. 28 (1979) 1.
[68] J. Bergeld, B. Kasemo, D.V. Chakarov, Surf. Sci. 495 (2001) L815.