Preferential oxidation of carbon monoxide catalyzed by platinum nanoparticles in mesoporous silica

Article abstract of DOI:10.1021/ja0703123

Preferential oxidation (PROX) of CO is an important practical process to purify H₂ for use in polymer electrolyte fuel cells. Although many supported noble metal catalysts have been reported so far, their catalytic performances remain insuffici

Full text of DOI:10.1021/ja0703123

Published on Web 07/31/2007
Preferential Oxidation of Carbon Monoxide Catalyzed by
Platinum Nanoparticles in Mesoporous Silica
Atsushi Fukuoka,*,† Jun-ichi Kimura,‡ Tadashi Oshio,‡ Yuzuru Sakamoto,‡ and
Masaru Ichikawa†
Contribution from Catalysis Research Center, Hokkaido UniVersity, N-21 W-10,
Sapporo 001-0021, Japan, and DiVision of Chemistry, Graduate School of Science,
Hokkaido UniVersity, N-10 W-8, Sapporo 060-0810, Japan
Received January 15, 2007; E-mail: fukuoka@cat.hokudai.ac.jp
Abstract: Preferential oxidation (PROX) of CO is an important practical process to purify H2 for use in
polymer electrolyte fuel cells. Although many supported noble metal catalysts have been reported so far,
their catalytic performances remain insufficient for operation at low temperature. We found that Pt
nanoparticles in mesoporous silica give unprecedented activity, selectivity, and durability in the PROX
reaction below 353 K. We also studied the promotional effect of mesoporous silica in the Pt-catalyzed
PROX reaction by infrared spectroscopy using the isotopic tracer technique. Gas-phase O2 is not directly
used for CO oxidation, but the oxygen of mesoporous silica is incorporated into CO2. These results suggest
that CO oxidation is promoted by the attack of the surface OH groups to CO on Pt without forming water.
nanoclusters in mesoporous silicas.15-18 In this work, we
targeted the preferential oxidation of CO (PROX: CO + 1/2O2
f CO2) in excess H2 as a catalytic reaction, because the PROX
is important for the purification of H2 for polymer electrolyte
fuel cells (PEFCs). H2 produced from gasoline or natural gas
contains a small amount of CO,19 but CO is a strong poison to
Pt electrodes in the PEFCs. Currently, CO is decreased to ca.
10 ppm by the PROX reaction at 423 K, and then H2 is supplied
to the PEFCs at 353 K. If the PROX is operated below 353 K,
the number of cooling processes can be reduced in the practical
production of H2. Supported noble metals such as Pt,20-24 Ru,25
Au,26,27 and bimetallic Pt-Fe28,29 have been reported as the
1. Introduction
Since the discovery of FSM-161,2 and MCM-413 in the early
1990s, mesoporous silicas1-5 have been extensively explored
for practical applications in catalysis,6,7 sorption,8,9 electronics,
and so on.10 The mesoporous silicas are characterized by ordered
pores (2-10 nm) with high surface area (ca. 500-1000 m2 g-1),
which are attractive as catalysts and supports. In the reported
works on catalysis, however, the characteristic of high surface
area was simply used to give higher dispersion of active sites
than that over the conventional silica. The unique promotional
effect of the mesoporous silica was described in the litera-
ture,6,7,11-14 but the effect remains unclear at the molecular level.
To design heterogeneous catalysts capable of high activity
and selectivity, we have studied the template synthesis of metal
(15) Fukuoka, A.; Higashimoto, N.; Sakamoto, Y.; Inagaki, S.; Fukushima, Y.;
Ichikawa, M. Microporous Mesoporous Mater. 2001, 48, 171-179.
(16) Fukuoka, A.; Sakamoto, Y.; Guan, S.; Inagaki, S.; Sugimoto, N.; Fukushima,
Y.; Hirahara, K.; Iijima, S.; Ichikawa, M. J. Am. Chem. Soc. 2001, 123,
3373-3374.
(17) Sakamoto, Y.; Fukuoka, A.; Higuchi, T.; Shimomura, N.; Inagaki, S.;
Ichikawa, M. J. Phys. Chem. B 2004, 108, 853-858.
(18) Fukuoka, A.; Higuchi, T.; Ohtake, T.; Oshio, T.; Kimura, J.; Sakamoto,
Y.; Shimomura, N.; Inagaki, S.; Ichikawa, M. Chem. Mater. 2006, 18, 337-
343.
(19) Rosso, I.; Galleitti, C.; Saracco, G.; Garrone, E.; Specchia, V. Appl. Catal.,
B 2004, 48, 195-203.
(20) Oh, S. H.; Sinkevitch, R. M. J. Catal. 1993, 142, 254-262.
(21) Kahlich, M. J.; Gasterger, H. A.; Behm, R. J. J. Catal. 1997, 171, 93-
105.
† Catalysis Research Center.
‡ Division of Chemistry.
(1) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn.
1990, 63, 988-992.
(2) Inagaki, S.; Fukushima, Y.; Kuroda, K. J. Chem. Soc., Chem. Commun.
1993, 680-682.
(3) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S.
Nature 1992, 359, 710-712.
(4) Zhao, D.; Feng, J.; Huo, Q.; Melosh, N.; Fredrickson, G. H.; Chmelka, B.
F.; Stucky, G. D. Science 1998, 279, 548-552.
(5) Inagaki, S.; Guan, S.; Fukushima, Y.; Ohsuna, T.; Terasaki, O. J. Am. Chem.
Soc. 1999, 121, 9611-9614.
(22) Minemura, Y.; Ito, S.; Miyao, T.; Naito, S.; Tomishige, K.; Kunimori, K.
Chem. Commun. 2005, 1429-1431.
(6) Corma, A. Chem. ReV. 1997, 97, 2372-2419.
(7) Thomas, J. M.; Raja, R. Stud. Surf. Sci. Catal. 2004, 148, 163-211.
(8) Itoh, T.; Yano, K.; Inada, Y.; Fukushima, Y. J. Am. Chem. Soc. 2002,
124, 13437-13441.
(23) Pedrero, C.; Waku, T.; Iglesia, E. J. Catal. 2005, 233, 242-255.
(24) Pozdnyakova, O.; Teschner, D.; Wootsch, A.; Kro¨hnert, J.; Steinhauer, B.;
Sauer, H.; Toth, L.; Jentoft, F. C.; Knop-Gericke, A.; Paa´l, Z.; Schlo¨gl, R.
J. Catal. 2006, 237, 1-16.
(9) Mal, N. K.; Fujikawa, M.; Tanaka, Y. Nature 2003, 421, 350-353.
(10) Schu¨th, F.; Schmidt, W. AdV. Mater. 2002, 14, 629-638.
(11) Yamamoto, T.; Tanaka, T.; Funabiki, T.; Yoshida, S. J. Phys. Chem. B
1998, 102, 5830-5939.
(25) Echigo, M.; Tabata, T. Appl. Catal., A 2003, 251, 157-166.
(26) Chilukuri, S.; Joseph, T.; Malwadkar, S.; Damle, C.; Halligudi, S. B.; Rao,
B. S.; Sastry, M.; Ratnasamy, P. Stud. Surf. Sci. Catal. 2003, 146, 573-
576.
(12) Inaki, Y.; Yoshida, H.; Kimura, K.; Inagaki, S.; Fukushima, Y.; Hattori,
T. Phys. Chem. Chem. Phys. 2000, 2, 5293-5297.
(13) Iwamoto, M.; Tanaka, Y.; Sawamura, N.; Namba, S. J. Am. Chem. Soc.
2003, 125, 13032-13033.
(27) Landon, P.; Ferguson, J.; Solsona, B. E.; Garcia, T.; Carley, A. F.; Herzing,
A. A.; Kiely, C. J.; Golunski, S. E.; Hutchings, G. J. Chem. Commun. 2005,
3385-3387.
(28) Watanabe, M.; Uchida, H.; Ohkubo, K.; Igarashi, H. Appl. Catal., B 2003,
46, 595-600.
(14) Junges, U.; Jacobs, W.; Voigt-Martin, I.; Krutzsch, B.; Schu¨th, F. J. Chem.
Soc., Chem. Commun. 1995, 2283-2284.
9
10120
J. AM. CHEM. SOC. 2007, 129, 10120-10125
10.1021/ja0703123 CCC: $37.00 © 2007 American Chemical Society

PROX by Pt Nanoparticles in Mesoporous Silica
A R T I C L E S
Table 1. Structural Parameters of Supports and Catalysts
PROX catalysts, but their catalytic activity, selectivity, and
durability need further improvement.
surface area
pore diameter
(BJH, nm)b
pore volume
1
a
1
materials
(BET, m2 g-
)
(mL g-
)
CO/Pt
H/Pt
We explored the catalysis of Pt nanoparticles and nanowires
synthesized in FSM-16 in the PROX reaction. Mesoporous
silicas FSM-16 and MCM-41 are synthesized from different
precursors, but they have the same structure (two-dimensional
hexagonal structure (p6mm symmetry) of one-dimensional
channels).2,3 We also used mesoporous organosilica HMM-1
with bridging -CH2CH2- groups5 and conventional silica to
know the effect of the organic group and the mesoporous
structure on the catalytic performances. The Pt particles in FSM-
16 exhibit catalytic behavior markedly different from that of
Pt/SiO2. Extremely high CO conversion and selectivity are
obtained over Pt/FSM-16 at low temperature. The promotional
effect of the mesoporous silica was further studied by infrared
(IR) spectroscopy using isotopes of CO, O2, and H2, and we
propose a reaction mechanism with the involvement of support-
oxygen in the CO oxidation reaction.
FSM-16
HMM-1
SiO2
Al2O3
Pt(p)/FSM-16
Pt(w)/FSM-16
Pt(w,ex)/FSM-16
Pt(p)/HMM-1
Pt(w)/HMM-1
Pt/SiO2
1018
689
287
82
944
878
979
661
648
2.7
3.0
0.84
0.57
1.33
0.13
0.77
0.68
0.78
0.58
0.57
2.7
2.7
2.7
3.0
3.0
0.24 0.27
0.09 0.08
0.31
0.15 0.15
0.13 0.11
0.17
Pt/Al2O3
a Brunauer-Emmet-Teller surface area. b Barrett-Joyner-Halenda
method.
PROX Reaction. Catalytic PROX reactions were conducted in a
plug flow reactor (inner diameter 8 mm) made of Pyrex (Figure S1).
Mass flows of CO (99.9%), H2 (99.999%), O2 (99.999%), N2 (99.999%,
internal standard), and CO2 (99.95%) were controlled by mass flow
controllers (STEC). Typically, Pt(p)/FSM-16 (0.1 g, Pt 5 wt %) was
diluted with glass beads (diameter 1 mm, 1.5 g) and charged in the
reactor. The catalyst was reduced in H2 flow at 473 K for 1 h, treated
in N2 flow at 473 K for 30 min, and cooled to room temperature. Next,
a gas mixture (CO 1%, O2 0.5-1%, N2 5%, H2 balance, flow rate 20
mL min-1, space velocity (SV) 12 000 mL g-1 h-1, 0.1 MPa) was fed
to the reactor. Reactions with CO2 (20%) and water vapor (2%) were
also performed by adjusting the flow rate of H2, but SV was always
kept constant under various conditions. After reaching to the steady
state in ca. 1 h, the outlet gas was analyzed by on-line gas chroma-
tography (Shimadzu GC 8A, thermal conductivity detector, a molecular
sieve 13X column (4 m) for separation of H2, O2, N2, and CO and an
active carbon column (2 m) for CO2, temperature 323-453 K (16 K
min-1)). The CO conversion (XCO) and the CO selectivity (SCO) are
2. Experimental Section
Preparation of Supported Pt Catalysts. FSM-162 and HMM-15
were synthesized according to the literature. SiO2 (Fuji Silysia, Cariact
Q-10) and γ-Al2O3 (Nishio, A-11) were used as received. Supported
Pt catalysts were prepared as follows. Typically, FSM-16 (200 mg in
30 mL of water) was impregnated with H2PtCl6‚6H2O (Wako, 28 mg
in 30 mL of water). The mixture was stirred for 24 h, evaporated to
dryness, and dried under vacuum for 24 h. The resulting H2PtCl6/FSM-
16 was calcined in O2 flow at 473 K for 2 h, then reduced in H2 flow
at 473 K for 2 h to give Pt nanoparticles in FSM-16 (abbreviated Pt-
(p)/FSM-16). UV-irradiation (250-600 nm) to H2PtCl6/FSM-16 with
water and methanol vapors for 48 h yielded Pt nanowires in FSM-16
(Pt(w)/FSM-16).15 Pt particles and/or wires were also prepared in
HMM-1. Pt/SiO2 and Pt/Al2O3 were prepared by the similar impregna-
tion and hydrogen-reduction. The Pt loading was 5 wt % for all of the
catalysts.
defined as follows: XCO ) (NCO,in - NCO,out)/NCO,in × 100 (%), SCO
)
(1/2)(NCO,in - NCO,out)/(NO ,in - NO ,out) × 100 (%), where N is mol of
2
2
CO or O2.
Pt nanowires were extracted from FSM-16 with HF and again
deposited on the external surface of FSM-16.17,18 This material was
also used as a catalyst: Pt(w,ex)/FSM-16.
3. Results and Discussion
3.1. Synthesis and Characterization of Pt Particles and
Wires in Mesoporous Silica. Pt nanoparticles and nanowires
were selectively synthesized by the typical impregnation and
subsequent hydrogen- or photoreduction method.15-18 Nitrogen
adsorption isotherms for FSM-16 and Pt(p)/FSM-16 are shown
in Figure S2, and structural parameters of the supports and
catalysts are summarized in Table 1. TEM, SEM, and STEM
data for the catalysts are shown in Figures 1 and S3. From these
results, the formation of mesopores is clearly observed for FSM-
16 and HMM-1, and similar pore diameters are obtained from
the TEM and SEM images and the pore size distribution in the
nitrogen adsorption.
For Pt(p)/FSM-16, the formation of 2.5 nm Pt particles is
clearly observed in the TEM and STEM images (Figure 1a,d),
and the SEM image (Figure 1c) shows no formation of big Pt
particles on the external surface of FSM-16. These results
indicate that the Pt particles are located in the internal meso-
pores. On the other hand, Pt wires (2.5 nm × 100-200 nm)
were formed in FSM-16 by the photoreduction method (Figure
1b). We also synthesized Pt particles or wires in HMM-1 by
the similar hydrogen- or photoreduction, and Pt particles on SiO2
and Al2O3 by the hydrogen-reduction (TEM, Figure S3). In
contrast to the mesoporous silica supports, a broad distribution
of the particle size was observed for Pt/SiO2 (3-7 nm) and
Pt/Al2O3 (3-10 nm) in the TEM images.
Characterization. N2 adsorption was carried out with a Quantach-
rome Autosorb-6, and uptakes of CO (323 K) and H2 (298 K) were
measured with a Quantachrome Chembet-3000. Powder X-ray diffrac-
tion (XRD) patterns were recorded on a Rigaku Miniflex using Cu KR
radiation (λ ) 0.15418 nm) at 30 kV and 15 mV. Transmission electron
microscopy (TEM) was performed with a Hitachi H-800 and a JEOL
JEM-2000ES, and field-emission scanning electron microscopy (FE-
SEM) with a Hitachi S-5500. X-ray photoelectron spectroscopy (XPS)
was measured with a JEOL JPS-9010MC. Temperature-programmed
desorption (TPR) of CO was analyzed using a Bell Japan TPD-1-AT.
IR spectra were measured at 293 K with a JASCO FTIR-420
spectrometer (resolution 2-4 cm-1, integration 100 times). A self-
supported wafer was prepared from catalyst powder (50 mg). The wafer
sample was charged in an IR cell with CaF2 windows, and dried under
vacuum at 473 K for 30 min. Next, H2 (100 Torr) was added to the
cell, and the sample was reduced at 473 K for 1 h. After evacuation at
473 K for 30 min, the sample was cooled to 293 K, and IR spectra
were acquired after exposure of the sample to CO or 13CO for 10 min.
Next, other gases (H2, D2, O2, or 18O2) were added to the cell, and IR
spectra were similarly measured after 10 min. For each treatment, the
IR spectrum was obtained by subtracting the spectrum of gaseous CO
charged in the IR cell.
(29) Tanaka, K.; Moro-oka, Y.; Ishigure, K.; Yajima, T.; Okabe, Y.; Kato, Y.;
Hamano, H.; Sekiya, S.; Tanaka, H.; Matsumoto, Y.; Koinuma, H.; He,
H.; Zhang, C.; Feng, Q. Catal. Lett. 2004, 92, 115-121.
9
J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007 10121

A R T I C L E S
Fukuoka et al.
Figure 2. XRD patterns of FSM-16, Pt(p)/FSM-16, and Pt(w)/FSM-16 at
high and low 2θ angles.
smaller than the experimental results. Therefore, our results
suggest that the mesopores are not completely plugged by the
Pt particles. It is thus reasonable to assume that the Pt particles
and wires at the internal mesopores work as catalysts in the
gas-phase reactions containing CO and H2.
In the XRD patterns for Pt/FSM-16, no significant change
was observed at a low 2θ angle (Figure 2a), indicating that the
ordered mesoporous structure remains unchanged before and
after the incorporation of Pt particles and wires. Peaks at a high
2θ angle in Figure 2b were assigned to typical fcc Pt by
comparing with the JCPDS card (No. 04-0802). The XRD
patterns for other catalysts are summarized in Figure S4.
3.2. Catalytic Performances of the Supported Pt Catalysts
in the PROX Reactions. First, PROX reactions were carried
out under the conditions of twice excess O2 (O2/CO ) 1) over
the stoichiometry (CO + 1/2O2 f CO2) according to the condi-
tions of precedent papers.19-27 Figure 3 shows the CO conver-
sion over the supported Pt catalysts as a function of reaction
temperature. Pt(p)/FSM-16 and Pt(w)/FSM-16 provided high
CO conversions, while SiO2-, Al2O3-, and HMM-1-supported
catalysts needed 423 K for removal of a majority of CO. In
particular, Pt(p)/FSM-16 gave ca. 100% conversion of CO even
at 313 K, thus showing that this is one of the most active
catalysts for PROX. No decrease in the CO conversion was
observed over Pt(p)/FSM-16 under the practical condition
containing CO2 (20%) and water vapor (2%) at 333 K (1 in
Figure 3a).
Figure 1. TEM images of (a) Pt(p)/FSM-16 and (b) Pt(w)/FSM-16. (c)
High-resolution SEM image of Pt(p)/FSM-16 and (d) HAADF (high angle
annular dark field) STEM image for (c).
It is a task for the PROX catalysts to work only for CO
oxidation in the presence of excess H2, because the H2 oxidation
(H2 + /2O2 f H2O) is a competitive reaction. The stoichio-
As shown in Table 1, the dispersion of Pt is 24-27% for
Pt(p)/FSM-16 and 8-9% for Pt(w)/FSM-16 by the uptake of
CO and H2. The TEM images in Figure 1 and the XRD data in
Figure 2 (vide infra) indicate that the 2.5 nm fcc Pt particles
are located in the mesopores of FSM-16. The ratio of surface
to bulk is 45% for the 2.5 nm Pt particle with fcc lattice structure
(truncated-cuboctahedron model).30 By comparing this value
(45%) and the dispersion data (24-27%), it is suggested that
50-60% of the Pt surface is exposed to the gas phase for Pt-
(p)/FSM-16. Similarly, ca. 25% of the Pt surface is exposed to
the gas phase even for Pt(w)/FSM-16. If the Pt particle or wire
plugged the mesopores, the dispersion would be markedly
1
metric condition (O2/CO ) 1/2) is enough for the CO oxidation
alone, but excess O2 is needed for high conversion of CO due
to the consumption of O2 in the H2 oxidation. Hence, the
selectivity for CO oxidation (CO selectivity) is an important
factor to evaluate the performances of PROX catalysts, where
the CO selectivity is defined as the ratio of the consumed CO
to the consumed O2 (see Experimental Section). In our reactions
flowing the stoichiometric O2/CO (1/2) gas mixture, Pt(p)/FSM-
16 showed CO conversions over 95% at 298-423 K, while
Pt/SiO2 gave low conversions (Figure 3b). This indicates that
the CO selectivity is over 95% for Pt(p)/FSM-16. The CO
selectivity was also studied under excess O2 (O2/CO ) 1) by
keeping the O2 conversion at 20-40% at various temperature.
(30) Bergeret, G.; Gallezot, P. In Handbook of Heterogeneous Catalysis; Ertl,
G., Kno¨zinger, H., Weitkamp, J., Eds.; VCH: Weinheim, 1997; Vol. 1, p
439.
9
10122 J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007

PROX by Pt Nanoparticles in Mesoporous Silica
A R T I C L E S
Figure 4. (a) PROX and normal CO oxidation over Pt(p)/FSM-16.
Conditions of PROX: CO 1%, O2 1%, N2 5%, H2 (PROX) or He (normal
CO oxidation) balance, SV 12 000 mL g-1 h-1, 0.1 MPa. (b) H2 oxidation
and water-gas shift reaction over Pt(p)/FSM-16. Conditions of H2 oxida-
tion: O2 1%, N2 5%, H2 balance, SV 12 000 mL g-1 h-1, 0.1 MPa.
Conditions of WGSR: CO 1%, H2O 2%, N2 5%, He balance, SV 12 000
mL g-1 h-1, 0.1 MPa. (c) Addition of water to normal CO oxidation over
Pt(p)/FSM-16 at the reaction time of 0.5-2.0 h. Conditions: CO 1%, O2
1%, H2O 2%, N2 5%, He balance, SV 12 000 mL g-1 h-1, 0.1 MPa, 333
K.
Figure 3. PROX reactions over supported Pt catalysts. (a) Under O2/CO
) 1. (b) Under stoichiometric O2/CO ) 1/2. (c) Durability test of Pt(p)/
FSM-16 at 353 K under O2/CO ) 1. Conditions: CO 1%, O2 0.5-1%, N2
5%, H2 balance, SV 12 000 mL g-1 h-1, 0.1 MPa.
As depicted in Figure S5, high selectivity was obtained over
the Pt catalysts on FSM-16 and HMM-1, but Pt/SiO2 or Al2O3
showed 50-60% selectivity.
In the durability test of Pt(p)/FSM-16 at 353 K (Figure 3c),
the CO conversion decreased slightly from 100% to 90% in 96
h, but the catalyst was regenerated by the hydrogen-reduction
at 473 K for 1 h. No sintering of Pt particles or wires was
observed in the recovered catalysts (Figure S6). Because of the
high activity, selectivity, and durability of Pt(p)/FSM-16, we
can operate the PROX reaction at 313-353 K, which would
be beneficial to the production of hydrogen for fuel cells by re-
ducing the number of cooling processes.19 The catalytic perfor-
mances of Pt(p)/MCM-41 were the same as those of Pt(p)/FSM-
16, thus making no difference between FSM-16 and MCM-41
as a support. It is reported that Pt-Fe/mordenite showed high
conversion and selectivity below 353 K,28 but this catalyst was
not durable as described in ref 27, giving lower conversion and
selectivity than Pt(p)/FSM-16 in our test reactions.
on FSM-16.17,18 However, the resulting Pt(w,ex)/FSM-16 gave
lower activity and selectivity than Pt(w)/FSM-16 (Figure S7),
which indicates that the occlusion of Pt in the internal pores is
important for high performances.
Under the normal CO oxidation conditions using He flow
instead of H2, the CO conversion declined to 20% over Pt/FSM-
16 catalysts (Figure 4a), showing that H2 has a promotional
effect on the CO oxidation. One explanation for the H2 effect
is the involvement of the water-gas shift reaction (WGSR: CO
+ H2O f CO2 + H2). However, Pt(p)/FSM-16 was active for
the H2 oxidation but inactive for WGSR (Figure 4b). Further-
more, the addition of H2O to the normal CO oxidation over
Pt(p)/FSM-16 had no positive effect on the CO conversion
(Figure 4c). These results indicate that H2O is not involved in
the PROX mechanism by Pt(p)/FSM-16.
To elucidate the effect of the internal and external surfaces
of FSM-16, Pt wires were extracted from Pt(w)/FSM-16 by
dissolving silica with HF and the wires were again deposited
3.3. Mechanistic Study of the PROX Reaction by IR. We
have studied the PROX mechanism by IR spectroscopy. Figure
9
J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007 10123

A R T I C L E S
Fukuoka et al.
Figure 5. IR spectra of CO adsorption on supported Pt catalysts. (a) CO
(30 Torr) at 293 K. (b) Gaseous 12CO2 or 13CO2 formed in the adsorption
of 12CO (30 Torr) and/or 13CO (30 Torr) on Pt(p)/FSM-16.
5a shows the IR spectra of the supported Pt catalysts after
addition of CO (30 Torr, 1 Torr ) 133 Pa) at 293 K. All three
samples, active Pt(p)/FSM-16 and inactive Pt(p)/HMM-1 and
Pt/SiO2, showed a linear CO frequency on Pt at 2088 cm-1
,
indicating no difference in the electronic states at the Pt surfaces.
The XPS results (Figure S8) and the TPD of CO (Figure S9)
also indicate the similar electronic states of the three catalysts.
In the enlarged IR spectrum for Pt/(p)/FSM-16 at 2400-2200
cm-1 (Figure 5b), small gaseous CO2 peaks were observed at
ca. 2350 cm-1. In the adsorption of 13CO or a mixture of 12CO
and 13CO, gaseous 13CO2 peak was seen at 2280 cm-1. It should
be noted that no oxygen source was present except for the
support FSM-16 in the adsorption of 12CO or 13CO. One possible
explanation is that the oxygen of silica in FSM-16 is incorpo-
rated into CO2 in the CO adsorption. In contrast to the active
Pt(p)/FSM-16, no CO2 peaks were observed in the CO adsorp-
tion over the inactive Pt(p)/HMM-1 or Pt/SiO2 catalyst.
Figure 6. IR spectra of gaseous CO2 (a) and Pt-CO (b) in sequential
adsorption of CO (10 Torr), 18O2 (10 Torr), and D2 (60 Torr) on Pt(p)/
FSM-16 at 293 K. (c) IR spectra in sequential adsorption of CO (10 Torr)
and D2 (10 Torr) on various supported Pt catalysts at 293 K.
Figure 6b shows the change of Pt-CO peak in the sequential
adsorption of CO, 18O2, and D2 on Pt(p)/FSM-16. The Pt-D
peak was not directly observed due to the overlap with strong
Pt-CO peak. Even in the adsorption of H2 or D2 alone, the
Pt-H or Pt-D peaks were not observed in IR. However, we
have observed the shift of Pt-CO peak that suggests the site-
selective adsorption of H2 on Pt and O2 on FSM-16. After
admission of CO to Pt(p)/FSM-16, the Pt-CO peak appeared
at 2088 cm-1 and the peak was not shifted by the addition of
18O2 to the gas phase (Figure 6b). In contrast, by further addition
of D2 to the gas phase, the Pt-CO peak was shifted to 2084
cm-1. This low-frequency shift is attributable to the enhanced
back-donation from Pt to CO by forming a Pt-D bond in the
dissociative adsorption of D2 to Pt. If 18O2 were adsorbed on Pt
under the same conditions, the Pt-CO peak would be shifted
to higher frequency by the decreased back-donation from Pt to
CO. Accordingly, we conclude that oxygen is not adsorbed on
Pt but on the support FSM-16.
To confirm the incorporation of the oxygen of FSM-16 into
CO2, we carried out sequential addition of CO, 18O2, and D2 on
Pt(p)/FSM-16. After admission of CO (10 Torr) to the IR cell
at 293 K, gaseous C16O2 was formed as shown in Figure 6a.
By further addition of 18O2 (10 Torr) to the cell, the peak
intensity of C16O2 slightly increased, but no peak of C16O18O
was detected at 2330 cm-1 as an 18O-labeled CO2. After
introduction of D2 (60 Torr) to make the PROX conditions (CO
+
18O2 + D2), the C16O2 peak again increased but the peaks of
18O-labeled CO2 were not observed. If labeled water (D218O)
generated from D2 and 18O2 attacked CO, C16O18O would be a
main product of WGSR. Therefore, neither gas-phase oxygen
(18O2) nor water (D218O) reacts with C16O over Pt(p)/FSM-16,
and the surface silanol groups are the source of oxygen in the
CO oxidation.
We also tested the possibility of fast 16O-18O exchange
between gaseous C18O2 and siliceous FSM-16 (Si16O2), but no
exchange was observed in the admission of C18O2 to Pt(p)/FSM-
9
10124 J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007

PROX by Pt Nanoparticles in Mesoporous Silica
A R T I C L E S
If O2, H2, and CO were competitively adsorbed on Pt, the
formation of H2O would not be avoided to reduce the CO
selectivity. Hence, the key to the high CO selectivity over Pt-
(p)/FSM-16 is the site-selective adsorption of H2 on Pt and O2
on FSM-16 without the formation of H2O. Tanaka et al. reported
that a small amount of H2O works as a cocatalyst for PROX
over FeOx/Pt/TiO2 catalyst.32 In contrast to this catalyst, the
addition of H2O showed no improvement of CO oxidation over
Pt(p)/FSM-16 (Figure 4c). If H2O were formed, the CO
selectivity would be lower in the catalytic reactions, and
C16O18O would be formed from C16O and D218O in the
sequential adsorption of CO, 18O2, and D2. Therefore, it seems
unlikely that water is involved in our PROX mechanism.
4. Conclusion
Figure 7. Proposed mechanism for PROX over Pt(p)/FSM-16.
We have shown that mesoporous silica-supported Pt nano-
particles are highly active and selective catalysts in the PROX
of CO. The catalytic performances of supported Pt catalysts are
promising with an impact at the commercial level. We have
also demonstrated the incorporation of support-oxygen into the
product CO2 by the IR experiments using the isotopic tracer
technique. From the experimental results, we conclude that the
PROX is promoted by the attack of OH groups at the internal
surface of mesoporous silica toward CO on Pt. The high activity
of the surface OH groups would be applicable to other oxidation
reactions. Work along this line is now in progress.
16 (Figure S10). It is thus reasonable to conclude that the oxygen
of FSM-16 is incorporated into CO2 even at 293 K.
As shown in Figure 5a, Pt(p)/FSM-16 gave more OH groups
at 3800-3200 cm-1 than did Pt/SiO2 and Pt(p)/HMM-1. When
D2 was admitted to the IR cell after CO adsorption at 293 K, a
broad band of OD was detected over Pt(p)/FSM-16 at ca. 2600
cm-1 (Figure 6c), while the OD band was not seen over Pt/
SiO2, Pt(p)/HMM-1, or FSM-16 itself. Presumably D2 is
dissociatively adsorbed on the Pt surface even in the presence
of adsorbed CO, and the resulting D species spills over onto
the FSM-16 surface to be exchanged with the surface OH. We
found that Pt(p)/FSM-16 are active for both the H-D exchange
and the PROX reaction.
By considering the catalytic performances and the IR results,
we propose that the surface OH on FSM-16 is highly reactive
toward CO adsorbed on Pt nanoparticles at the interface of Pt
and FSM-16. As depicted in Figure 7, the OH groups attack
CO on Pt to form CO2 and H2, and a vacant site is formed on
FSM-16. O2 is adsorbed on FSM-16 as shown in Figure 6, and
it is suggested that the oxygen species migrates in FSM-16 to
form the OH groups with spill-over hydrogen from Pt. In this
mechanism, CO reacts with surface OH instead of H2O. Basset
et al. also reported that silanol groups (Si-18OH) of conventional
silica react with bis(allyl)Rh complex immobilized on the silica
to form Rh(C18O) complex.31 This also indicates the attack of
silanol groups on the surface species on metal at the interface
of metal and support. In our experiments, the mesoporous silica
has a greater effect on the PROX reaction than the conventional
silica (Figure 3a), thus showing that the OH groups at the
internal surface of mesoporous silica are more active than those
of the conventional silica.
Acknowledgment. We thank Mr. Y. Nodasaka for TEM,
Hitachi High-Technologies for FE-SEM, Dr. N. Iwasa for TPD,
and Dr. S. Inagaki, Prof. K. Tanaka, Dr. K. Yamakata, and Dr.
P. L. Dhepe for discussions on the reaction mechanism. This
work was supported by a Grant-in-Aid for Scientific Research
from the Ministry of Education, Culture, Sports, Science and
Technology, Japan (18560737, 18065001), and NOASTEC.
Supporting Information Available: Schematic representation
of the plug flow reactor (Figure S1); nitrogen adsorption
isotherms for FSM-16 and Pt(p)/FSM-16 (Figure S2); TEM
(Figure S3) and XRD (Figure S4) of Pt/HMM-1, Pt/Al2O3, and
Pt/SiO2; CO selectivity at 20-40% O2 conversion over the
supported Pt catalysts (Figure S5); TEM of the recovered Pt(p)
or Pt(w)/FSM-16 after PROX (Figure S6); PROX by Pt(w,ex)/
FSM-1 (Figure S7); XPS of Pt(p)/FSM-16 (Figure S8); TPD
of CO for the supported Pt catalysts (Figure S9); and IR of
C18O2 adsorption on Pt(p)/FSM-16 (Figure S10). This material
is available free of charge via the Internet at http://pubs.acs.org.
JA0703123
(31) Santini, C. C.; Scott, S. L.; Basset, J.-M. J. Mol. Catal. A 1996, 107, 263-
271.
(32) Tanaka, K.; Shou, M.; He, H.; Shi, X. Catal. Lett. 2006, 110, 185-190.
9
J. AM. CHEM. SOC. VOL. 129, NO. 33, 2007 10125