Effects of reduction temperature and metal-support interactions on the catalytic activity of Pt/γ-Al2O2 and Pt/TiO 2 for the oxidation of CO in the presence and absence of H 2

Article abstract of DOI:10.1021/jp054888v

TiO₂- and γ-Al₂O₃-supported Pt catalysts were characterized by HRTEM, XPS, EXAFS, and in situ FTIR spectroscopy after activation at various conditions, and their catalytic properties were examined for the oxidation of CO i

Full text of DOI:10.1021/jp054888v

23430
J. Phys. Chem. B 2005, 109, 23430-23443
Effects of Reduction Temperature and Metal-Support Interactions on the Catalytic
Activity of Pt/γ-Al2O3 and Pt/TiO2 for the Oxidation of CO in the Presence and Absence
of H2
†
†
‡
†
Oleg S. Alexeev, Soo Yin Chin, Mark H. Engelhard, Lorna Ortiz-Soto, and
Michael D. Amiridis*,†
Department of Chemical Engineering, UniVersity of South Carolina, Columbia, South Carolina 29208, and
W. R. Wiley EnVironmental Molecular Sciences Laboratory, Pacific Northwest National Laboratory,
P.O. Box 999, MS K8-93, Richland, Washington 99352
ReceiVed: August 29, 2005; In Final Form: September 30, 2005
TiO
spectroscopy after activation at various conditions, and their catalytic properties were examined for the oxidation
of CO in the absence and presence of H (PROX). When γ-Al was used as the support, the catalytic,
electronic, and structural properties of the Pt particles formed were not affected substantially by the pretreatment
conditions. In contrast, the surface properties and catalytic activity of Pt/TiO were strongly influenced by
2
- and γ-Al
2 3
O -supported Pt catalysts were characterized by HRTEM, XPS, EXAFS, and in situ FTIR
2
2 3
O
2
the pretreatment conditions. In this case, an increase in the reduction temperature led to higher electron density
on Pt, altering its chemisorptive properties, weakening the Pt-CO bonds, and increasing its activity for the
oxidation of CO. The in situ FTIR data suggest that both the terminal and bridging CO species adsorbed on
fully reduced Pt are active for this reaction. The high activity of Pt/TiO
attributed to the ability of TiO to provide or stabilize highly reactive oxygen species at the metal-support
interface. However, such species appear to be more reactive toward H than CO. Consequently, Pt/TiO shows
substantially lower selectivities toward CO oxidation under PROX conditions than Pt/γ-Al
2
for the oxidation of CO can also be
2
2
2
O .
2 3
Introduction
particles. However, the question of how big metal particles
should be for their catalytic properties to become independent
of the nature of the support remains open. Furthermore, limited
attention has been paid to the influence of metal-support
interactions on reactions taking place in oxidative environments.
The goal of the work summarized in this paper was to
determine how the nature of the metal-support interactions
could affect the dispersion, electronic, chemisorptive, and
catalytic properties of Pt in reactions in which O2 participates
as a reactant. More specifically, our interest is focused on the
oxidation of CO in the presence and absence of H2 over TiO2-
and γ-Al2O3-supported Pt catalysts. These reactions have
attracted substantial attention in recent years due to their
Supported Pt catalysts are employed in several large-scale
industrial applications including among others oxidation,
hydrogenation, and hydrocarbon rearrangement reactions. The
properties of Pt in these catalysts are determined not only by
the metal surface structure and dispersion, but also by the
metal-support interaction. For example, more than two decades
ago it was demonstrated that platinum-group metals supported
on TiO2 exhibit a strong dependence of their CO and H2
chemisorptive properties on the reduction protocol used to treat
these catalysts. This phenomenon, described by the term Strong
Metal-Support Interaction (SMSI), has been the subject of
extensive investigation. The results reported in the literature
indicate that the SMSI effect also influences the catalytic
performance of supported metals in hydrogenation and hydro-
1
2
3
potential application in hydrogen purification schemes for fuel
9,10
cell use, as well as indoor/cabin air cleanup.
The TiO2- and
γ-Al2O3-supported platinum catalysts were characterized in
detail by HRTEM, XPS, EXAFS, and FTIR spectroscopy.
Results from these efforts were combined with catalytic data
to determine how various factors influence the activity of the
supported Pt particles in the oxidation of CO in air, as well as
in H2-rich mixtures.
3
genolysis reactions. Several explanations suggested for the
interpretation of SMSI were based on electronic and structural
effects involving both the active metal and fragments of the
3
,4
support. Results of more recent work focusing on cluster-
derived catalysts suggest that reactants, as well as fragments of
the support, may be viewed as ligands capable of affecting the
structure and, therefore, the catalytic properties of supported
Experimental Methods
5
-7
metals.
This effect is maximized for small supported metal
Reagents and Materials. The γ-Al2O3 support with a BET
clusters incorporating only a few metal atoms, which have
properties distinctly different from those of bulk metals. As
the size of these clusters grows, the fraction of the metal atoms
located at the metal-support interface progressively decreases,
hence the effect of the support becomes negligible for very large
2
8
surface area of 100 m /g was prepared by forming a paste of
aluminum oxide C (Degussa) in dionized water, followed by
drying at 120 °C and calcination at 500 °C in air for 24 h. The
2
TiO2 powder (Kemira) with a BET surface area of 72 m /g was
also calcined in air at 500 °C for 12 h prior to use. H2, N2, and
He (all UHP grade; National Welders) were purified prior to
their use by passage through oxygen/moisture traps (Model
OT3-2, Agilent) capable of removing traces of O2 and water to
*
Address correspondence to this author.
University of South Carolina.
Pacific Northwest National Laboratory.
†
‡
1
0.1021/jp054888v CCC: $30.25 © 2005 American Chemical Society
Published on Web 11/22/2005

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
5 and 25 ppb, respectively. O2 and CO (National Welders)
were purified prior to their use by passage through a moisture
trap (Model GMT-2GCHP, Agilent) capable of removing water
to 25 ppb. The H2PtCl6‚6H2O precursor (99.95% purity, Alfa
Aesar) was used as supplied.
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23431
1
the catalyst bed. Samples in a powder form (0.077 g) were
diluted 90 times by weight with quartz particles (60-80 mesh)
to keep the catalyst bed isothermal. The total volumetric flow
rate of the reactant mixture was held at 154 mL/min (1atm, 25
°C) yielding a corresponding Gas Hourly Space Velocity
(
GHSV) of 120 000 mL/g‚h. Reacting gases were mixed in a
Catalyst Preparation. Pt/TiO2 and Pt/γ-Al2O3 were prepared
by incipient wetness impregnation of the corresponding supports
with an aqueous solution of H2PtCl6‚6H2O, followed by drying
in air at 110 °C for 24 h. The amounts of the precursor were
chosen to yield samples containing 1 wt % of Pt. Prior to further
characterization, each sample was treated with H2 as described
below.
gas distribution system, while the flow of each gas was
controlled by a mass flow controller (Model 201-DKASVBAA,
Porter) to create an accurate and reproducible feed containing
0
.5% of CO, 0.5% of O2, 45% of H2, and 54% of N2. Before
mixing, the CO/N2 mixture was heated to 350 °C in a quartz
trap to eliminate any carbonyls that may have been formed in
the storage cylinder. The feed and the reaction products were
analyzed with on-line single beam NDIR CO (Ultramat 23,
Siemens) and O2 (Model 201, AMI) analyzers capable of
detecting CO and O2 in the 0-250 and 0-1000 ppm ranges,
respectively. The outputs from both analyzers and the temper-
ature controller were linked to a user interface with Labview
software. The reaction selectivity toward the formation of CO2
was calculated as the amount of O2 consumed in the CO
oxidation reaction (calculated from the CO balance) over the
total amount of O2 consumed.
FTIR Spectroscopy. FTIR spectra were collected with a
Nicolet Nexus 470 spectrometer equipped with a MCT-B
detector cooled by liquid nitrogen. Powder samples were pressed
into self-supported wafers and mounted in the IR cell, which
was connected to a gas distribution manifold. The design of
11
the IR cell has been described elsewhere. Samples were
pretreated in H2 flow at the 200-400 °C temperature range for
2
h prior to each measurement. After the reduction treatment
was completed, the cell was flushed with He for 15 min and
cooled to room temperature in flowing He. Spectra were
-
1
The catalytic oxidation of CO in the absence of H was carried
recorded at a spectral resolution of 2 cm accumulating 64
scans per spectrum.
2
out in equipment similar to that described above, at atmospheric
pressure, a GHSV of 120 000 mL/g‚h, and temperatures between
3
X-ray Photoelectron Spectroscopy (XPS). XPS spectra were
collected with an X-ray High-Resolution Photoelectron Spec-
trometer (Quantum 2000, Physical Electronics), using a focused
monochromatic Al KR beam with an energy of 1486.7 eV, that
0 and 300 °C. The reaction feed contained 1% of CO balanced
with air. Both the reaction feed and products were analyzed
with an on-line single-beam NDIR (Ultramat 23, Siemens)
analyzer capable of detecting CO in the 0-500 ppm and 0-5%
ranges and CO2 in the 0-5% range.
Prior to any catalytic measurements, the samples were treated
with H2 or O2, while the temperature was ramped at 5 °C/min
and held at the desired temperature for 2 h. In the absence of a
catalyst, there was no measurable conversion of CO for either
reaction in the temperature range of interest (i.e., up to 300 °C).
EXAFS Spectroscopy. EXAFS data were collected at X-ray
beamline X-18B of the National Synchrotron Light Source
(NSLS), Brookhaven National laboratory, Upton, NY. The
storage ring energy was 2.8 GeV and the ring current 110-
270 mA.
2
was rastered over 0.28 mm of the sample area. The spectrom-
eter was equipped with a Spherical Capacitor Energy Analyzer
incorporating a 16-element multichannel detector. Prior to these
measurements, each sample was treated with H2 at 200-400
°
C or O2 at 300 °C in a glass reactor connected to a gas
distribution system. After the treatment was completed, the
reactor was sealed and transferred into a glovebox, where the
sample was loaded on the standard sample holder that was
transferred to the analysis chamber without air exposure. The
-
8
residual gas pressure in the analysis chamber was 5 × 10
Torr. The Ti (2p) line (458.7 eV) was used to calibrate the
binding energies that were measured with an accuracy of (0.2
eV. Low-energy electrons and argon ions were used for
specimen neutralization.
The Pt/γ-Al O samples were loaded in a wafer form into an
2
3
1
3
in situ EXAFS cell connected to a gas distribution system,
allowing for the in situ treatment of samples in various
atmospheres. After the desired treatment was completed, the
EXAFS cell was evacuated to 10 Torr at room temperature,
cooled to liquid nitrogen temperature, and aligned in the X-ray
beam. The EXAFS data were recorded in the transmission mode
with an appropriate amount of sample to give an absorbance of
approximately 2.5 at the Pt L3 (11563.7 eV) edge.
Transmission Electron Microscopy (TEM). TEM images
were obtained with a 200 keV high-resolution transmission
electron microscope (JEOL 2010F) having a specific point-to-
point resolution of 0.194 nm. All images were digitally recorded
with a multiscan charge-coupled device camera (Model 794,
Gatan) with 1024 × 1024 pixel resolution and further analyzed
with a digital micrograph (Gatan). During a typical sample
preparation for TEM imaging, approximately 15 mg of sample
was finely dispersed and embedded in a L. R. White resin. Then,
the sample was sectioned by using a RMC MT 6000 microtome
to slices with a thickness of approximately 50 nm and deposited
onto a Cu grid. Average particle sizes were determined by
measuring at least 200 particles from 20 different micrographs.
The statistical analysis used to determine the surface-mean
-
5
Data for Pt/TiO samples were recorded in the fluorescence
2
mode with a 13th element Ge detector. Samples in powder form
were reduced in H at the desired temperature, purged with He
at the same temperature, cooled to room temperature in He flow,
2
and mounted on the sample holder while exposed to air. Data
were collected at room temperature with a Si(111) double-crystal
monochromator that was detuned by 20% to minimize the
effects of higher harmonics in the X-ray beam.
The EXAFS data were analyzed with experimentally deter-
mined reference files obtained from EXAFS data characterizing
materials of known structure, as stated elsewhere. The EXAFS
parameters were extracted from the raw data with the aid of
the XDAP software. The methods used to extract the EXAFS
function from raw data are essentially the same as those reported
1
2
particle sizes was similar to that reported elsewhere. The
accuracy of determining the mean size was estimated to be on
the order of (15%.
14
Catalytic Measurements. Steady-state catalytic activity
measurements for the preferential oxidation of CO under excess
H2 (PROX) were performed in a quartz single-pass fixed-bed
microreactor at atmospheric pressure. The temperature inside
the reactor was monitored by a thermocouple extending into
1
5
1
6
elsewhere. The data used for each sample were the average

23432 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
TABLE 1: Average Size of Pt Particles in Pt/γ-Al
2 3
O and
Pt/TiO Following Reduction at Different Temperatures
2
Average Pt diameter (Å)
reduction
sample
temp (°C)
HRTEM
EXAFS
1
1
.0% Pt/γ-Al
O
2 3
200
9
11
11
25
26
3
4
00
00
.0% Pt/TiO
2
200
00
25
14
4
of five scans. The data at the Pt L3 edge were analyzed with a
maximum of 16 free parameters over the ranges of 3.50 < k <
-
1
1
5.50 Å (k is the wave vector) and 0.0 < r < 5.0 Å (r is the
distance from the absorbing atom; in this case Pt). The
statistically justified number of free parameters, n, was found
1
7,18
to be 39, as estimated on the basis of the Nyquist theorem:
n ) (2∆k∆r/π) + 1, where ∆k and ∆r, respectively, are the k
and r ranges used in the data fitting. The parameters character-
izing both the high-Z (Pt-Pt) and low-Z (Pt-Osupport) contribu-
tions were determined by multiple-shell fitting in R-space with
1
3
application of k and k weighting in the Fourier transforma-
1
6
tions. The fit was optimized by use of a difference file
technique with phase- and amplitude-corrected Fourier trans-
forms of the data.
Figure 1. Particle size distributions as estimated by TEM for 1 wt %
Pt/TiO samples reduced with H at 200 and 400 °C.
2 2
1
9,20
Results
Platinum Dispersion. The average Pt particle sizes in the
Pt/γ-Al2O3 and Pt/TiO2 systems after various pretreatments were
estimated from EXAFS data based on known relationships
between the average size of metal particles and the metal-metal
first-shell coordination number.2 The results obtained (Table
1,22
1
) indicate that when γ-Al2O3 was used as the support, small
platinum particles with an average diameter of approximately
Å were formed after reduction at 200 °C. An increase of the
9
reduction temperature up to 400 °C had no substantial effect
on the Pt particle size and dispersion (i.e., the fraction of metal
atoms exposed at the surface) with the average Pt particle size
remaining at approximately 11 Å. Assuming a spherical shape
for these metal particles, the Pt dispersion in these samples was
found to be almost 100%.
Following the hydrogen treatment at 200 °C Pt particles with
an average diameter of approximately 25 Å were formed in the
case of Pt/TiO2 as indicated by both the HRTEM and EXAFS
data (Table 1). The HRTEM data (Table 1, Figure 1) further
indicate that an increase of the reduction temperature to 400
Figure 2. XPS spectra of the Pt(4f) core level region for 1 wt % Pt/
TiO samples treated with O or H at various temperatures.
2
2
2
°
C was accompanied by a decrease of the average Pt size from
5 Å to approximately 14 Å. However, the EXAFS data suggest
2
TABLE 2: Summary of XPS Data Obtained for Pt/TiO2
that the size of Pt particles remained unchanged (Table 1). This
discrepancy between the results obtained from the two different
techniques can be rationalized in terms of SMSI behavior, as
discussed in detail later.
treatment gas/
temp (°C)
Pt 4f5/2 (eV)
Pt 4f7/2 (eV)
Pt/Ti
O
2
H
2
H
2
/300
/200
/400
75.4
75.0
74.6
72.1
71.7
71.3
0.012
0.011
0.009
XPS Measurements. The binding energies of the Pt (4f) core
level region and the Pt/Ti atomic ratios for Pt/TiO2 samples
treated with O2 or H2 at various temperatures are reported in
Table 2. The corresponding XPS spectra of the Pt (4f) region
are shown in Figure 2. When the Pt/TiO2 sample was treated
with O2 at 300 °C, the XPS spectrum was characterized by two
peaks corresponding to the Pt (4f5/2) and Pt (4f7/2) photoemission
lines centered at 75.4 and 72.1 eV, respectively. Following
exposure of the same sample to H2 at 200 °C both peaks shifted
to lower binding energies by approximately 0.4 eV. Subsequent
treatment with H2 at 400 °C resulted to an additional shift of
the Pt (4f) lines to even lower binding energies of 74.6 and
surface Pt (i.e., Pt/Ti atomic ratio) between samples treated in
O2 at 300 °C and in H2 at 200 °C. However, the Pt/Ti atomic
ratio decreased beyond the margin of error as the reduction
temperature was increased to 400 °C (Table 2).
Infrared Spectra of CO Adsorbed on Pt/γ-Al2O3. When
Pt/γ-Al2O3 was reduced in H2 at 200 °C and exposed to 75
Torr of CO at room temperature, three bands were observed in
the νCO region at 2110 (vw), 2062 (s), and approximately 1820
-
1
(w) cm (Figure 3A). These bands can be assigned to linear
CO on Pt and Pt sites, and bridging CO on Pt0 sites,
2
+
0
2
3
7
2
1.3 eV for the Pt (4f5/2) and Pt (4f7/2) peaks, respectively (Figure
). No differences were observed in the relative amounts of the
respectively. The intensity of the peak associated with CO
2
+
coordinated on Pt sites was small and its contribution to the

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23433
Figure 4. FTIR spectra collected during the desorption of CO from
(
2 3 2
A) 1 wt % Pt/γ-Al O reduced at 400 °C and (B) 1 wt % Pt/TiO
Figure 3. FTIR spectra of CO adsorbed at room temperature on (A)
reduced at 200 °C, following treatment in He at various temperatures.
1
wt % Pt/γ-Al
2 3 2 2
O and (B) 1 wt % Pt/TiO reduced in H at different
temperatures.
TABLE 3: Thermal Stability of Adsorbed CO Species
reduction
temp (°C)
temp of complete
CO desorption (°C)
total absorption was less than 1%, consistent with the notion
sample
that almost all Pt was effectively reduced at this temperature to
the metallic state. Furthermore, the band at 2110 cm-1 was not
1.0% Pt/γ-Al
2
O
3
200
400
400
400
60
25
25
3
4
00
00
detectable in the infrared spectra obtained with samples reduced
at higher temperatures, indicating reduction of Pt to an even
higher extent under these conditions. The only difference
observed in this case was a decrease in the intensity of the two
1
.0% Pt/TiO
2
200
300
400
0
bands associated with Pt for the sample reduced at 400 °C,
adsorbed on fully reduced Pt sites were observed at 2076 (s)
and 1830 (w) cm , respectively (Figure 3B). An increase of
indicating that some sintering of the Pt particles has taken place
during this pretreatment.
-1
the reduction temperature to 200 °C had no substantial effect
on either the intensity or frequency of these bands. However,
when the reduction temperature was further increased to 300
FTIR data further show that the CO species are relatively
weakly adsorbed on Pt/γ-Al2O3 and can be removed by heating
of the sample in He flow. The desorption of CO did not depend
on the reduction temperature and typically proceeded as shown
in Figure 4A. The bridging CO band at 1820 cm- disappeared
first at approximately 285 °C, while the linear CO band at 2062
-
1
°C, the band at 2076 cm declined dramatically in intensity
and appeared as a shoulder to a new band observed at 2060
cm (Figure 3B, spectrum 3). The presence of bridging CO
1
-1
species remained evident in the spectrum, although the intensity
of the corresponding band was very low. The intensities of all
these bands further declined as the reduction temperature was
increased to 500 °C, although the characteristic frequencies of
these bands remained unchanged (Figure 3B, spectra 3-5).
The FTIR data further indicate that the various CO species
adsorbed on Pt/TiO2 can be removed by thermal treatment in
He (Figure 4B). However, the pattern observed during the
-
1
-1
cm declined in intensity, broadened, shifted to 2030 cm ,
and finally disappeared at approximately 400 °C (Figure 4A).
These results are also summarized in Table 3.
Infrared Spectra of CO Adsorbed on Pt/TiO2. Similar to
the results obtained with Pt/γ-Al2O3, when Pt/TiO2 was reduced
in H2 at 100 °C and exposed to 75 Torr of CO at room
temperature, bands of terminal and bridging CO species

23434 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
desorption of CO in this case appeared to be different from that
observed for the Pt/γ-Al2O3 samples. For example, the adsorp-
tion of CO on the Pt/TiO2 sample reduced at 200 °C was much
weaker and complete desorption of CO was achieved at much
lower temperatures. The bridging species were no longer
detectable as the temperature was increased to 38 °C, while the
terminal CO species completely disappeared at temperatures as
low as 60 °C (Figure 4B). Moreover, as the intensity of the
-
1
linear CO band at 2076 cm decreased with desorption
temperature, no apparent shift of the band position was observed
(Figure 4B). Similar behavior was also observed with the
samples reduced at higher temperatures, although complete
desorption of CO was observed in these cases even at room
temperature (Table 3).
Kinetic and in Situ FTIR Monitoring of the Oxidation of
CO by Air. The CO conversions observed at various temper-
atures over 1 wt % Pt/γ-Al2O3 samples pretreated under different
conditions are shown in Figures 5A and 5B. Almost identical
results were obtained for the samples reduced in H2 at 300 or
4
2
00 °C with 100% conversion being reached in each case at
40 °C. Reduction at lower temperature (200 °C) did not
influence the light-off curve in the 50-200 °C temperature
range. However, after attaining approximately 60% CO conver-
sion at 216 °C, the reaction rate was decreased in the 216-235
°
C temperature range, and then again increased with 100% CO
conversion eventually reached at 260 °C. When an oxidation
pretreatment at 300 °C was used instead of reduction, the light-
off curve characteristics were almost identical with those
observed with the sample reduced at 200 °C (Figure 5B).
However, a subsequent treatment in H2 at 400 °C resulted in a
sample indistinguishable from the one originally pretreated in
H2 at 400 °C (Figures 5A and 5B).
Similar light-off curves obtained with 1 wt % Pt/TiO2 samples
pretreated under different conditions are shown in Figure 6. The
light-off curves in this case are shifted toward lower tempera-
tures, indicating the higher activity of the Pt/TiO2 samples for
the oxidation of CO. Furthermore, the performance of these
samples was substantially affected by the pretreatment protocol.
More specifically, while complete CO conversion was reached
at 210 °C for the sample pretreated in O2 at 300 °C, over the
samples reduced in H2 at 200 °C or 300 °C complete CO
conversion was observed at 200 and 178 °C, respectively.
Further increase of the reduction temperature to 400 °C did not
result in any further shift of the light-off temperature, although
this sample exhibited noticeably higher CO conversions in the
Figure 5. CO conversions versus temperature observed during the
2 3
oxidation of CO in air over 1 wt % Pt/γ-Al O samples pretreated (A)
in H at (2) 200, (b) 300, and (9) 400 °C and (B) in (2) H at 200
2
2
°C, (b) O
2
at 300 °C, and (9) O
2
at 300 °C followed by H
2
at 400 °C.
cm-1 declined in intensity and finally disappeared at ap-
proximately 200 °C, with the former being removed first. In
contrast, the low-intensity band, initially observed in the
5
0-150 °C temperature range than samples reduced at lower
-1
spectrum at 2121 cm , increased substantially in intensity as
temperatures (Figure 6).
the temperature was raised to 170 °C, and remained present at
temperatures as high as 300 °C (Figure 8B).
Figure 7 provides a direct comparison of the CO oxidation
over Pt/TiO2 and Pt/γ-Al2O3 samples pretreated in H2 at 400
Similarly, when Pt/TiO2 was exposed to the reaction mixture
at room temperature, three carbonyl bands were observed at
°
C. The results demonstrate the higher activity of the Pt/TiO2
sample. The pure TiO2 and γ-Al2O3 supports also exhibited
limited activity at temperatures above 300 °C.
-1
2
125 (wv), 2080 (s), and 1835 (w) cm (Figure 8A, spectrum
1). These bands remained unchanged as the reaction temperature
In situ FTIR experiments were conducted in parallel to the
kinetic measurements under similar experimental conditions with
the infrared cell serving as a flow reactor. Spectra of Pt/TiO2
and Pt/γ-Al2O3 samples exposed to the reaction mixture (1%
CO/air) at different temperatures are shown in Figure 8, parts
A and B, respectively. Exposure of Pt/γ-Al2O3 to the reaction
mixture at 25 °C led to the appearance of three bands at 2121
was increased to 105 °C. A substantial change was observed
-
1
beyond this point, with the band at 1835 cm disappearing
-
1
from the spectrum and the band at 2080 cm significantly
declining in intensity, and eventually disappearing as the reaction
temperature reached 140 °C (Figure 8A, spectrum 6). The band
-1
at 2125 cm also declined in intensity with increasing reaction
temperature, but remained visible in the spectrum even at 160
°C (Figure 8A, spectrum 7).
-
1
(wv), 2067 (s), and 1832 (w) cm (Figure 8B), which are
similar to those observed during the adsorption of CO in the
absence of O2 (Figure 4A). The spectrum remained essentially
unchanged at temperatures up to 115 °C. As the reaction
temperature was increased further, the bands at 1832 and 2067
Curves characterizing the removal of CO species from the
surface of Pt/TiO2 as a function of the reaction temperature are
compared in Figure 9. These data were extracted from in situ
FTIR spectra and show that when the sample was reduced at

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23435
Figure 6. Light-off curves characterizing the oxidation of CO in air
over 1 wt % Pt/TiO
[) 200, (2) 300, and (9) 400 °C. (GHSV) 120 000 mL/g‚h; 1% CO
balanced with air).
2 2 2
samples pretreated in (b) O at 300 °C, or H at
(
Figure 8. In situ FTIR spectra collected at various reaction temper-
atures in 1% CO in air over (A) a 1 wt % Pt/TiO
2
sample prereduced
at 200 °C and (B) a 1 wt % Pt/γ-Al sample prereduced at 300 °C.
2
O
3
remained at 100%. No differences were observed with samples
reduced at temperatures between 200 and 400 °C.
Figure 7. Light-off curves demonstrating the effect of the support on
the oxidation of CO in air over (9) 1% Pt/TiO , (2) 1% Pt/γ-Al
pretreated with H at 400 °C, ([) pure TiO treated in H at 300 °C,
and (1) pure γ-Al treated in H at 300 °C.
Similar light-off curves obtained over bare TiO2 and a 1 wt
Pt/TiO2 sample are shown in Figure 11. Over the TiO2
2
2 3
O
%
2
2
2
support pretreated in H2 at 300 °C the CO conversion became
measurable only at temperatures above 200 °C (Figure 11). The
consumption of O2 was slightly higher than that of CO at every
reaction temperature examined, reflecting the relatively low
selectivity for CO oxidation, which was in the range 16-32%.
The 1 wt % Pt/TiO2 sample pretreated in H2 at 200 °C
exhibited much higher activity than that of the bare TiO2 support
with a maximum CO conversion of approximately 62%
observed at 252 °C. Substantial consumption of O2 was observed
with this sample at reaction temperatures as low as 70 °C, and
complete O2 conversion was reached at 140 °C (Figure 11).
Some significant changes were observed in the light-off
characteristics of both CO and O2 when the 1 wt % Pt/TiO2
sample was pretreated in H2 at 300 °C. Approximately 10% of
the O2 in this case was already consumed at 30 °C, while
complete O2 conversion was observed at 115 °C. In contrast,
the CO light-off curve exhibited two maxima, one of ap-
proximately 30% at 70 °C, and one (similar to what was
observed with the sample pretreated at 200 °C) of approximately
O
2 3
2
300 °C or higher temperatures, almost all CO species were
removed from the Pt surface at 80 °C. However, when the
sample was reduced at 200 °C or, alternatively, oxidized at 300
°
C, temperatures of about 140 and 225 °C, respectively, were
required. Similar information extracted from FTIR spectra of
the Pt/γ-Al2O3 (not shown) indicate no substantial differences
between samples reduced at various temperatures.
Selective Oxidation of CO in H2-Rich Mixtures. The
conversions of CO and O2 at different temperatures during the
selective oxidation of CO in the presence of excess H2 over the
1
wt % Pt/γ-Al2O3 sample reduced in H2 at 300 °C are shown
in Figure 10. In both cases the conversions increased in parallel,
to almost 100% as the temperature was increased to 210 °C.
During this process the selectivity remained in the neighborhood
of 50%. Further increase of the reaction temperature to 233 °C
was accompanied by a decrease in the CO conversion to 85%
and the CO selectivity to 42%, while the O2 conversion

23436 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
Figure 9. Curves characterizing the removal of CO from the Pt surface
during CO oxidation by air over 1 wt % Pt/TiO treated with (b) O
at 300 °C, (9) H at 200 °C, and (2) H at 300 °C. The surface
Figure 11. Dependence of (9, 2, 3) CO and (b, [, O) O
on the reaction temperature during the selective oxidation of CO in a
-rich stream over TiO (3, O) treated with H at 300 °C and over
1% Pt/TiO reduced with H at (b, 2) 200 °C and (9, [) 300 °C.
2
conversion
2
2
2
2
H
2
2
2
concentration of CO species was determined from in situ IR data
2
2
recorded at various reaction temperatures and GHSV of approximately
7
5 000 mL/g‚h.
H2, O2, or CO), HRTEM, and EXAFS.21,24 The data of Table 1
indicate that when an aqueous solution of H2PtCl6‚6H2O was
used as the precursor for the preparation of Pt/γ-Al2O3, the
subsequent drying and reduction with H2 at 200 °C led to the
formation of small Pt particles with an average diameter of 9
Å. Similarly sized particles were obtained when an air exclusion
technique was used to prepare Pt/γ-Al2O3 from a [PtCl2(PhCN)2]
6
precursor. When the reduction temperature was increased to
3
00 °C, the average diameter of the metal particles was also
increased to approximately 11 Å, indicating that limited
agglomeration of Pt took place after such a treatment. Further
increase of the reduction temperature to 400 °C appeared to
have no additional effect on the Pt dispersion (Table 1),
indicating the relatively high thermal stability of Pt on the
γ-Al2O3 surface under these conditions. These results are in good
agreement with earlier reports illustrating the high resistance
of γ-Al2O3-supported Pt to sintering even after oxidation-
2
1
reduction treatments at 200-400 °C.
Larger Pt particles were obtained with the Pt/TiO2 system.
The HRTEM data indicate that the average Pt particle size in
this case was approximately 26 Å after reduction at 200 °C
(Table 1). The EXAFS data (Table 4) indicate that this sample
is characterized by a first-shell Pt-Pt coordination number of
Figure 10. Dependence of (2) CO and (b) O
reaction temperature during the selective oxidation of CO in a H
stream over 1% Pt/γ-Al reduced with H at 300 °C.
2
conversion on the
2
-rich
O
2 3
2
9
.1 at a distance of 2.75 Å, which in turn yields an average Pt
60% at 250 °C. Almost identical behavior was also observed
22
particle size of approximately 25 Å, in good agreement with
the HRTEM results. Substantial Pt-O contributions were also
observed in this sample at 2.16 and 2.25 Å signifying interac-
tions of Pt with the oxygen atoms of the support, as well as
with oxygen atoms adsorbed on Pt due to its exposure to air
during the data acquisition process in the fluorescence mode
for this particular sample, although at this point we cannot
distinguish between chemisorbed oxygen and oxygen associated
with the support. Nevertheless, the presence of substantial Pt-
Pt contributions in this sample is consistent with the presence
of relatively large Pt particles the structure of which is not
affected substantially by interactions with O2. In contrast, in
the case of small metal clusters or particles, it is expected that
the interaction with O2 will lead to gradual oxidative fragmenta-
tion of the platinum clusters, ultimately leading to a complete
loss of Pt-Pt bonds and to the formation of platinum oxide
with the sample pretreated in H2 at 400 °C.
Structural Characterization by EXAFS. The structural
parameters characterizing the Pt species formed in the Pt/γ-
Al2O3 and Pt/TiO2 samples after various treatments are sum-
marized in Tables 4 and 5. The error bounds in these parameters
represent precisions determined from statistical analysis of the
EXAFS data with the XDAP software,15 and not accuracies.
The estimated accuracies in the determination of the EXAFS
parameters of the main contributions are as follows: coordina-
tion number (N), (20%; distance (R), (1%; Debye-Waller
2
factor (∆σ ), (30%; inner potential correction (∆E0), (10%.
Discussion
Platinum Dispersion. The dispersion of Pt in supported
catalysts can be reliably determined from a variety of different
techniques including the chemisorption of probe molecules (e.g.,
2
5
species. It has been reported, for example, that exposure to

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23437
Samples after Various Treatmentsa
2
TABLE 4: Summary of the EXAFS Data Characterizing the Pt/TiO
treatment gas/
temp (°C)
3
2
2
shell
Pt-Pt
N
R (Å)
10 ∆σ (Å )
0
∆E (eV)
H
2
/200
/400
9.1 ( 0.4
2.75 ( 0.01
5.4 ( 0.4
-0.1 ( 0.3
Pt-Osupport
Pt-O1
3.1 ( 0.9
5.3 ( 0.3
9.3 ( 0.5
2.16 ( 0.02
2.25 ( 0.02
2.76 ( 0.01
5.9 ( 0.8
10.0 ( 1.1
4.7 ( 0.4
-13.2 ( 0.5
9.3 ( 0.8
-0.3 ( 0.3
Pt-O2
Pt-Pt
H
2
Pt-Osupport
Pt-O1
3.4 ( 0.1
5.9 ( 0.3
2.18 ( 0.02
2.28 ( 0.02
6.8 ( 0.1
10.0 ( 1.2
-17.2 ( 0.4
3.8 ( 0.6
Pt-O2
a
2
0
Notation: N, coordination number; R, distance between absorber and backscatterer atoms; ∆σ , Debye-Waller factor; ∆E , inner potential
correction.
2 3
TABLE 5: Summary of the EXAFS Data Characterizing the Pt/γ-Al O
Samples after Various Treatmentsa
3
2
2
treatment gas/temp (°C)
shell
N
R (Å)
10 ∆σ (Å )
0
∆E (eV)
H
H
H
2
2
2
/400
/300
/200
Pt-Pt first
Pt-Pt second
Pt-Osupport
6.1 ( 0.1
2.2 ( 0.1
2.72 ( 0.01
3.91 ( 0.01
5.8 ( 0.1
9.0 ( 0.3
0.7 ( 0.2
-8.0 ( 0.3
Pt-O
Pt-O
s
0.7 ( 0.1
0.8 ( 0.1
6.0 ( 0.1
2.1 ( 0.1
2.12 ( 0.01
2.74 ( 0.01
2.72 ( 0.01
3.88 ( 0.02
10.0 ( 1.0
1.0 ( 0.6
5.5 ( 0.1
10.0 ( 1.0
-1.6 ( 0.2
-2.2 ( 0.5
-1.3 ( 0.1
-2.6 ( 0.5
l
Pt-Pt first
Pt-Pt second
Pt-Osupport
Pt-O
Pt-O
s
0.6 ( 0.1
1.1 ( 0.1
4.5 ( 0.1
2.2 ( 0.1
2.10 ( 0.01
2.70 ( 0.01
2.72 ( 0.01
3.91 ( 0.01
10.0 ( 1.3
3.8 ( 0.5
4.0 ( 0.1
9.0 ( 0.3
0.7 ( 0.3
2.1 ( 0.4
-3.9 ( 0.1
-8.0 ( 0.3
l
Pt-Pt first
Pt-Pt second
Pt-Osupport
Pt-O
Pt-O
s
0.9 ( 0.1
1.1 ( 0.1
4.1 ( 0.1
2.13 ( 0.01
2.65 ( 0.01
2.79 ( 0.01
10.0 ( 1.1
0.2 ( 0.3
7.7 ( 0.2
-2.0 ( 0.3
4.5 ( 0.4
-4.6 ( 0.2
l
H
2
/200 then CO + air/160
/200 then CO + air/280
Pt-Pt
Pt-Osupport
Pt-O
Pt-O
s
2.0 ( 0.1
1.6 ( 0.2
3.1 ( 0.1
2.25 ( 0.02
2.64 ( 0.01
2.78 ( 0.01
-0.6 ( 0.1
10.0 ( 1.0
0.8 ( 0.1
-13.4 ( 0.4
-4.7 ( 0.4
-1.0 ( 0.2
l
H
2
Pt-Pt
Pt-Osupport
Pt-O
Pt-O
s
2.8 ( 0.1
1.4 ( 0.1
2.19 ( 0.01
2.68 ( 0.01
1.0 ( 0.1
2.8 ( 0.7
-8.5 ( 0.2
-6.41 ( 0.4
l
a
Notation as in Table 4; the subscripts s and l refer to short and long, respectively.
O2 at room temperature was sufficient to break all the Pt-Pt
bonds in γ-Al2O3-supported platinum clusters with an average
diameter of about 10 Å.26 This disintegration process was
followed by EXAFS spectroscopy, where the formation of
isolated platinum oxide clusters was confirmed.6
consistent with the presence of metal particles having an average
diameter of aproximately 26 Å. One possible explanation for
this apparent discrepancy between HRTEM and EXAFS data
2
in the case of Pt/TiO sample is the unique nature of the
interactions between Pt and the TiO at elevated reduction
2
When the reduction temperature of Pt/TiO2 was increased to
00 °C, the average diameter of the Pt particles estimated from
temperatures, as suggested earlier for catalysts showing SMSI-
31
4
type behavior. The reduction of pure TiO2 usually takes place
2
7
HRTEM measurements was substantially different from that
calculated based on the EXAFS data (Table 1). The HRTEM
data indicate that the size of the Pt particles becomes smaller
following reduction at 400 °C (Table 1, Figure 1). A redispersion
of Pt has been previously observed in oxygen-rich atmospheres
at elevated temperatures, and has been attributed to the oxidation
of Pt particles to volatile species and the subsequent transport
at temperatures higher than 600 °C. However, this process is
accelerated in the presence of Pt. Temperature-programmed
reduction (TPR) data, for example, indicate that the reduction
of TiO2 in this case begins at 300 °C, with a maximum observed
2
7
at approximately 460 °C. It has been further suggested that
H2 atoms chemisorbed on Pt can interact with oxygen atoms
from the support located at the metal-support interface to form
3
+
31
of such species over relatively large distances on the support
oxygen vacancies and partially reduced Ti
sites. The
surface.2
7-30
It also has been shown that the presence of chlorine
3+
subsequent strong interaction between exposed Ti cations and
during the oxidation can significantly enhance the Pt redispersion
Pt atoms located at the metal-support interface may then lead
IV
on γ-Al2O3 or TiO2 due to the formation of [Pt OxCly] species
to the migration of partially reduced TiO support fragments
x
2
8-30
from PtO2.
However, despite the presence of chloride ions
onto the metal, resulting in a decrease of the exposed metal
3
in our case due to the use of the H2PtCl6‚6H2O precursor, it is
very unlikely that these species played an important role in the
oxygen-free environment used. Moreover, in such a case a
similar behavior would be expected from the Pt/γ-Al2O3 sample,
which was not observed.
surface. If such decoration was to take place, it would result
in underestimations of the metal particle sizes by HRTEM,
consistent with our observations. At the same time, the EXAFS
results should remain unchanged, since EXAFS probes the bulk
of the Pt particles present. Such a decoration model is also
consistent with the XPS and FTIR data discussed below.
Nevertheless, additional detailed worksbeyond the scope of the
current papersis needed to further probe the reasons for the
In contrast to the HRTEM data, the EXAFS results (Table
4
) indicate that after reduction at 400 °C, the first-shell Pt-Pt
coordination number and distance remain virtually unchanged,

2
3438 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
observed discrepancy between HRTEM and EXAFS measure-
ments and to reach definitive conclusions.
of Pt, consistent with our XPS data. Similar behavior was also
observed in Rh/TiO2, Rh/Ta2O5, Ni/Nb2O5, and Ni/TiO2 samples.
In these cases EXAFS results provide strong evidence for the
structural reorganization of the support in the vicinity of the
metal particles, enabling a direct bonding between metal atoms
and partially reduced cations from the support, as evidenced
Electronic Structure of Platinum Particles. The photo-
emission doublet observed at 75.4 and 72.1 eV in the Pt (4f)
region of the O2-treated Pt/TiO2 sample (Figure 2, Table 2) can
2
+
be assigned to Pt species present on the TiO2 surface in the
form of PtO. TPR data reported previously for a Pt/TiO2 sample
prepared from a similar precursor indicate the complete reduc-
tion of Pt at temperatures as low as 100 °C.27 Consequently,
the Pt 4f5/2 and Pt 4f7/2 peaks observed in the spectrum of the
sample reduced with H2 at 200 °C at 75.0 and 71.7 eV,
respectively (Table 2, Figure 2), can be assigned to the fully
reduced Pt species.
n+
n+
n+
by the detection of Rh-Ti , Rh-Ta , Ni-Nb , and Ni-
n+
36-38
Ti contributions in the EXAFS spectra.
CO Adsorption on Pt. Additional information regarding the
state of Pt in Pt/TiO2 and Pt/γ-Al2O3 can be obtained from the
FTIR spectra recorded during the adsorption and desorption of
3
9
CO as a probe molecule. No differences were observed
between Pt/γ-Al2O3 samples reduced in the 200-400 °C
-
1
temperature range and bands observed at 2062 and 1829 cm
Further reduction at 400 °C results in a shift of the Pt (4f)
lines to lower binding energies by approximately 0.4 eV. In
addition to changes in the oxidation state and/or electron density,
changes in metal dispersion may also result in shifts in the XPS
spectra, since it has been shown that the Pt (4f) doublet is located
can be assigned to terminal and bridging CO species adsorbed
on metallic Pt, respectively (Figure 3A). The red shift observed
with these samples in the terminal CO band during desorption
(Figure 4A) is consistent with a decrease in the dipole-dipole
32
coupling effect between adsorbed CO molecules due to a
decrease in the CO surface coverage. In general, these data
indicate that the Pt-support interactions in the Pt/γ-Al2O3
sample were relatively weak and did not alter the surface
properties of Pt to a noticeable extent.
at higher binding energies for smaller Pt particles. In contrast,
a shift toward lower binding energies is observed when Pt
particles become electron rich. It is possible that particle size
and charge transfer effects may compensate each other to a
certain degree. However, since Pt particle size remained virtually
unchanged in the Pt/TiO2 sample after reduction at 200-400
In contrast, substantial differences were observed between
Pt/TiO2 samples reduced at different temperatures. More specif-
ically, the intensities of both the linear and bridged CO bands
were decreased by 40% and 70%, respectively, as the reduction
temperature was increased from 300 to 500 °C, indicating a
significant change in the chemisorptive properties of Pt (Figure
3B). Furthermore, no apparent shift was observed during
desorption (Figure 4B), indicating the absence of any significant
dipole-dipole coupling between CO molecules in these cases.
Thus, the infrared data provide strong evidence for a decrease
of the total amount of CO adsorbed on Pt/TiO2 with increasing
reduction temperature, which is consistent with earlier literature
°
C, as evidenced by the EXAFS data discussed in the previous
section, the shift toward lower energies observed in the XPS
spectra can be attributed to an increase in the electron density
on Pt following reduction at elevated temperatures, consistent
with an increased interaction between the Pt atoms and reduced
fragments of the support. TiO2 reduction in the presence of Pt
33
has been confirmed previously by ESR measurements. In fact,
it was shown that after reduction at 300 °C the concentration
3
+
of Ti cations is 30 times higher in the presence of Pt than on
the pure support. More recent XPS data show that formation of
3+
Ti cations in the presence of Pt occurs at temperatures as low
32
3,4
as 170 °C at the surface layers of TiO2. Therefore, an electron
reports.
3+
transfer can be expected in this case from Ti to Pt, if the Pt
The reduction temperature of 300 °C appears to represent a
critical point at which the change of Pt properties in the Pt/
TiO2 samples takes place. At this point a substantial decrease
3
+
atoms and Ti cations were to interact with each other. Such
a transfer would enhance the electron density on Pt and result
in a shift of the Pt (4f) doublet toward lower energies, consistent
with our observations (Table 2).
-1
is observed in the intensity of the band at 2076 cm and a
-1
new band appears at 2060 cm (Figure 3B). The appearance
It is worthy of note that the Pt/Ti atomic ratio, reflecting the
relative amount of the surface Pt atoms, decreased as the
reduction temperature was increased to 400 °C (Table 2). Similar
results have been reported already for TiO2- and CeO2-supported
Pt catalysts exhibiting the SMSI effects and were interpreted
in terms of a screening effect for the emitted photoelectrons
of this new band representing linear CO species at lower
wavenumbers indicates the creation of new electron-rich Pt sites
in the sample. Such an increase of the electron density of Pt
results in an increase of the back-donation of the metal electrons
into 2π* antibonding orbitals of the CO molecule and a
weakening of the CdO bond. The presence of electron-rich Pt
sites in Pt/TiO2 following reduction at temperatures above 300
°C suggests the existence of electronic interactions between Pt
and the support in agreement with the XPS data discussed
earlier. Still, however, the electronic properties of a small
fraction of Pt sites remained unchanged, as indicated by the
presence of a shoulder observed at the original location of the
terminal CO band (i.e., 2076 cm ) (Figure 3B). These
arguments are further supported by the results obtained during
the desorption of CO. According to the Blyholder model39
describing the interaction of CO with transition metals, an
increase in the electron density of the metal does not only result
3
2,34
from Pt atoms.
Such a screening could also be observed
upon decreasing the amount of exposed Pt atoms due to
sintering. More likely, however, the observed decrease of the
Pt/Ti ratio as a function of the reduction temperature can be
attributed to a decoration of Pt by TiOx fragments of the support
in agreement with earlier suggestions.3 It is interesting that
our HRTEM and XPS data suggest that such a decoration of Pt
can occur at temperatures as low as 400 °C, while typically
this process is believed to take place at temperatures higher than
,4
-
1
5
00 °C.35 Nevertheless, recent XPS, soft X-ray photoelectron
spectroscopy, and low-energy ion scattering data confirm the
possibility of Pt decoration by CeOx and TiOx suboxides in
in a weakening of the CdO bond, but also a strengthening of
3
2,34
40,41
model catalysts at temperatures as low as 170 °C.
Such
the Pt-CO bond.
However, this model is disputed by more
decoration of Pt by a partially reduced TiOx species would
substantially increase the metal-support interface, and therefore,
the number of possible contact points between Pt atoms and
recent theoretical calculations, which indicate that it is possible
that changes of the electron environment of transition metals
may result in an increase of back-donation, but not necessarily
3+
42
Ti cations, resulting in the modification of electronic properties
a strengthening of the M-CO bond. Consistent with this

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23439
prediction, binding energies of CO on various bimetallics were
found to be smaller than those determined for the corresponding
monometallic surfaces.43-46
for the oxidation of CO was the one reduced in H2 at 400 °C
(Figure 6). This sample was characterized by a reduced Pt/Ti
ratio and a negative shift (by 0.4 eV) of the Pt (4f) lines (Table
2), consistent with the assumption that Pt particles in this case
are partially covered by fragments of partially reduced support
(TiOx) (although it is not clear at this point whether such
decoration of Pt by TiOx takes place under CO oxidation
conditions). Electronic interactions between Pt and these frag-
ments are expected to result in an increase of the electron density
on Pt and, therefore, a weakening of the CO adsorption (as
already shown by the FTIR results discussed previously). Since
the reaction between adsorbed CO and oxygen atoms is believed
to be the rate-determining step during oxidation of CO, it is
expected that the presence of weakly bonded CO and/or O2
species on the Pt surface would favor higher reaction rates, in
agreement with our observations.
Our infrared data further show that when γ-Al2O3 was used
as the support, the reduction temperature had no effect on the
thermal stability of the adsorbed CO species, which can be
completely removed from the Pt surface at approximately 400
°
C. This result is in good agreement with similar data reported
21
previously for Pt/γ-Al2O3. In contrast, the stability of the
adsorbed CO species was found to be substantially different
when TiO2 was used as the support. For example, CO adsorbed
on Pt/TiO2 samples that were previously reduced in H2 at 200
°
C can be completely removed from the surface at ap-
proximately 60 °C (Table 3). The strength of CO adsorption
decreases even further on Pt/TiO2 samples reduced in H2 in the
3
00-500 °C range. Room temperature treatment in He was
enough in these cases to completely remove the adsorbed CO,
Reactivity of Various CO Species. Our in situ FTIR results
provide valuable information regarding the reactivity of the
various CO species adsorbed on the Pt surface. When freshly
reduced Pt/TiO2 and Pt/γ-Al2O3 samples were exposed to the
consistent with a very week Pt-CO bonding in these samples.
In summary, the infrared results demonstrate that supported
platinum particles interact much stronger with TiO2 than with
γ-Al2O3. These metal-support interactions depend on the
reduction temperature and led to an increase in the electron
density of Pt following reduction at 300 °C or above, altering
its chemisorptive properties and weakening the Pt-CO bonding.
1
% CO/He flow, the intensity of the bands corresponding to
the terminal and bridging CO species adsorbed on Pt was the
same even when the adsorption was performed at temperatures
as high as 300 °C, consistent with the high heat of CO
Effect of Support on the Oxidation of CO. It has been
previously suggested that the oxidation of CO proceeds through
a Langmuir-Hinshelwood (LH) type mechanism,47 involving
the reaction between CO molecules and oxygen atoms adsorbed
on a metal surface. Alternatively, the reaction has also been
proposed to proceed through an Eley-Rideal (ER) type mech-
57
adsorption on Pt. When the CO adsorption was conducted in
the presence of air (Figures 8A and 8B), the CO bands for both
terminal and bridging species were observed at slightly higher
-1
wavenumbers (a shift of approximately 5 cm ), signifying the
51
presence of O2 on the Pt surface, consistent with earlier reports.
Furthermore, the appearance of the weak band at approximately
48
anism involving chemisorbed oxygen atoms and gas-phase CO.
A growing database of reported experimental data in more recent
years has provided additional insight and may suggest that both
these kinetic models represent simplifications of a much more
complex mechanism. For example, neither of these two mech-
anisms can account for the higher activity of catalysts prepared
on easily reduced metal oxides as opposed to catalysts prepared
on γ-Al2O3 or SiO2. Moreover, it has been shown that
preadsorbed CO (or O2) species are not displaced by subsequent
exposure to O2 (or CO) and that the heat of adsorption of CO
does not change significantly in the presence of O2, suggesting
that two different types of adsorption sites on the Pt surface
-1
2
121 cm in the spectra of both Pt/TiO2 and Pt/γ-Al2O3
samples indicates that at room temperature a small fraction of
0
2+
23
Pt sites was converted to Pt in the CO/Air mixture. It has
been previously shown that Pt clusters with an average diameter
of 10 Å can be completely disintegrated after exposure to O2 at
6
room temperature. However, the heat of adsorption of O2 on
Pt, which typically ranges between 214 and 335 kJ/mol
5
8
51
depending on coverage, can be as low as 30 kJ/mol in the
presence of CO. Such weakly adsorbed oxygen species appear
to have no significant effect on the structure of supported Pt
2
+
particles, and only a small fraction of Pt species is formed
(as indicated by the 2121 cm band) regardless of particle size
or the support used.
-1
4
9-52
are involved in the accommodation of CO and O2.
Such
data imply that there is no competition for the adsorption sites
between CO and O2, contradicting earlier suggestions that some
adsorbed CO species must be desorbed from the metal surface
to free adsorption sites for O2.
The in situ FTIR measurements conducted under CO oxida-
tion conditions over both Pt/TiO2 and Pt/γ-Al2O3 (Figure 8A)
indicate that the reaction between CO and oxygen was negligible
at room temperature as evidenced by the presence of substantial
In the case of Pt/TiO2 it is likely that TiO2 can participate in
the catalytic process by providing highly active oxygen species.
It is known that when oxygen vacancies are formed at the
metal-support interface of Pt/TiO2, oxygen from the gas phase
could be dissociatively adsorbed on such deffects with one of
the oxygen atoms filling the vacancy and the other being
-
1
amounts of terminal (band at 2080 cm ) and bridging (band
-1
at 1835 cm ) CO species on the Pt surface. As the temperature
was increased, the intensity of both bands was decreased,
indicating relatively low surface coverages at elevated temper-
atures due to the reaction between CO and oxygen, implying
that both terminal and bridging CO species adsorbed on Pt are
active in the oxidation of CO. However, the bridging CO species
were less thermally stable on the Pt surface (Figure 4B) and
the first to disappear from the spectra under reaction conditions,
suggesting that the weakly adsorbed species may be more
reactive for the oxidation of CO. At the same time, the intensity
coordinated to the five-coordinate Ti4 site as an adatom.
+
33,53
Other surface species, such as molecularly adsorbed oxygen in
-
the form of O2 , can also be stable at relatively high temper-
5
3,54
atures.
Moreover, it has been shown that the presence of Pt
-
-
55
can stabilize O and O3 species photoformed on TiO2.
Nevertheless, regardless of the nature of the oxygen species
formed, one may assume that the presence of such species
conveniently located at the metal support interface contributes
to the higher activity of Pt/TiO2 for the oxidation of CO.
In contrast to earlier observations,56 our results show that
among the different Pt/TiO2 samples examined, the most active
-
1
of the band at 2125 cm remained almost unaffected by the
reaction temperature, implying that the CO species adsorbed
2
+
on Pt sites are not involved in the reaction. The catalytic and
infrared data show similar trends (Figures 6 and 9) indicating
that reduced samples were more active in the oxidation of CO

2
3440 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
as compared to the oxidized ones. Furthermore, the activity of
reduced samples was increased with increasing reduction
temperature.
(Table 5). In this case, the sample was characterized by a NPt-Pt
of 3.1 and by a NPt-Os of 2.8 consistent with the suggestion
that Pt particles were disintegrated at these conditions to a larger
extent. As a result, oxide-like Pt species were formed, as
evidenced by the appearance of the characteristic band for the
Similar behavior for the terminal and bridging CO species
was observed by in situ FTIR on the Pt/γ-Al2O3 sample (Figure
2
+
-1
8
B). Furthermore, no substantial differences were observed
Pt -CO species at 2121 cm in the FTIR spectrum (Figure
B). However, the EXAFS data indicate that at high reaction
between Pt/γ-Al2O3 samples reduced at various temperatures.
Finally, no substantial differences were observed in the catalytic
data (Figure 5A) obtained with these samples, suggesting that
in the case of Pt/γ-Al2O3 pretreatment at different conditions
did not induce any metal-support interactions capable of
affecting either the chemisorptive or catalytic properties of Pt.
Structural Changes of Pt/γ-Al2O3 under CO Oxidation
Conditions. The in situ FTIR results (Figure 8B) show that
under reaction conditions the disappearance of the terminal and
bridging CO species was accompanied by the appearance of a
8
temperatures and high CO conversions, the surface of the
catalyst resembles extremely small Pt clusters incorporating no
more than 4-6 Pt atoms (Table 5). Such Pt clusters can be
61
positively charged and therefore be less active for the oxidation
of CO as opposed to completely reduced Pt species. This
conclusion is in agreement with our catalytic data demonstrating
that the reduced Pt/TiO2 samples were more active than the
oxidized ones.
-
1
It is interesting that the low- and the high-temperature features
of the light-off curves characterizing the Pt/γ-Al2O3 after the
reduction at 200 °C can be completely reproduced upon
decreasing the reaction temperature, indicating the reversible
nature of the suggested structural changes. We associate such
reversibility with the ability of CO to reduce partially or
completely oxidized Pt species formed during the course of the
catalytic reaction. Such reduction presumably takes place at low
reaction temperatures, while at high reaction temperatures (and
therefore high CO conversions) the lack of CO in the gas phase
once again leads to the partial oxidation of Pt. Similarly, when
initially oxidized Pt/γ-Al2O3 was exposed to the CO/air mixture
at room temperature, initial contact with the CO led to the partial
reduction of Pt. This essentially explains why the light-off curves
characterizing the oxidized Pt/γ-Al2O3 were almost identical
with those of the sample reduced at 200 °C (Figure 5B) and
suggests that these two samples have a similar structure under
reaction conditions.
strong band at 2112 cm indicating the oxidation of Pt. This
process was more noticeable with increasing reaction temper-
ature for Pt/γ-Al2O3 than for Pt/TiO2. This difference can be
attributed to differences in the Pt dispersion (Table 1) since the
smaller Pt particles observed on Pt/γ-Al2O3 are more susceptible
to oxidation. Moreover, previous literature data reported for Pt/
γ-Al2O3 show that at high reaction temperatures and, therefore,
high CO conversions, oxygen species dominate the coverage
5
9
on the Pt surface. At these conditions, the majority of the
oxygen species are strongly bonded to Pt due to high heats of
oxygen adsorption in the absence of CO species. One would
expect that such strong O2 adsorption will mainly have an affect
on the structure of small Pt particles as opposed to big ones, in
agreement with our observations.
The infrared and EXAFS data reported here lead to a better
understanding of the light-off curves characterizing the CO
oxidation over the Pt/γ-Al2O3 samples that were reduced at 200
°
C or oxidized at 300 °C (Figures 5A and 5B). The inflection
points observed in the light-off curves of these samples in the
16-235 °C temperature range could be explained by structural
changes of Pt before and after the ignition temperature similar
Furthermore, the EXAFS data show that under reaction
conditions the Pt-Pt bond distance in Pt/γ-Al2O3 was about
2.78 Å, which is 0.06 Å larger than that observed in the freshly
reduced samples (Table 5). It is not quite clear at the moment
why the Pt-Pt bond distance was increased under reaction
conditions. If these changes were due to the oxidation of Pt by
O2, the Pt-Pt contributions would be expected at a shorter
2
59
to what has been suggested elsewhere. Reduction of Pt/γ-Al2O3
at 200 °C leads to the formation of highly dispersed supported
Pt particles having an average diameter of approximately 9 Å
with nearly all the Pt atoms exposed (Table 1). Such small
particles were characterized by a first-shell Pt-Pt coordination
number of 4.5 at a bond distance of 2.72 Å and by a second-
shell Pt-Pt coordination number of 2.2 at a bond distance of
6
distance. Changes in the Pt-Pt bond distance were observed
previously by EXAFS when Pt-containing samples were ex-
3
.91 Å (Table 5). The Pt-O contributions were observed in
the spectrum at short (Pt-Os) and long (Pt-Ol) distances of
.13 and 2.65 Å, respectively, signifying interactions of Pt with
posed to H , but were not detected upon exposure to various
2
6
types of hydrocarbons. On the other hand, the experimental
2
structural parameters determined from EXAFS data indicate that
the metal-metal bond distances in fully carbonylated Ru3(CO)12
or Rh6(CO)16 clusters are typically larger than those observed
60
the support, which are typical for supported metals. Exposure
of the sample to a CO/air mixture at 160 °C had some impact
on the structure of the surface Pt species. The second-shell Pt-
Pt contributions were no longer present in the spectrum,
indicating some disintegration of the Pt particles under these
conditions. The observed increase of the coordination number
for the Pt-Os contributions to 2.0 (Table 5) may be further
regarded as evidence that interactions with O2 have caused some
structural changes of the metal particles. However, the degree
of such changes was relatively small as the first-shell Pt-Pt
coordination number remained nearly unchanged (Table 5).
Thus, our EXAFS data allowed us to infer that at relatively
low reaction temperatures, and therefore low CO conversions
with plenty of CO present in a gas phase, O2 as a reactant
has a small impact on the structure of highly dispersed Pt
particles.
60
in corresponding decarbonylated species. This allows us to
suggest that under our experimental conditions the observed
increase of the RPt-Pt could be explained by a possible formation
of Pt carbonyl clusters with structures that cannot be precisely
identified from our EXAFS data. Infrared data reported by
6
2
Ichikawa et al. provide evidence for the formation of
2
-
[
Pt3(CO)6] from H2PtCl6 after treatment with CO and H2O at
50 °C, although such structural data as the Pt-Pt bond distances
were not reported for this cluster. The process of the formation
of Pt carbonyls could be similar to that of the formation of
γ-Al2O3-supported Ir4 or Ir6 carbonyl clusters upon reductive
6
1
carbonylation of mononuclear Ir species.
Finally, the fact that an unusual shape of the light-off curves
was observed only for Pt/γ-Al2O3 having dispersed Pt particles
is consistent with the inference that only exposed metal surfaces
Stronger changes in the value of NPt-Pt were observed when
the same sample was exposed to the CO/air flow at 280 °C

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23441
are vulnerable to the structural changes induced by reactants as
that the rate of hydrogen oxidation is decreased at these
temperatures, the observed conversion of CO is likely related
to the onset of side reactions such as methanation and/or the
water-gas shift reactions.
was shown by EXAFS for Pt/γ-Al2O3 during the hydrogenation
6
of alkenes.
Preferential Oxidation of CO in Hydrogen-Rich Streams
(
PROX). The catalytic data of Figure 10 indicate that Pt/γ-
Al2O3 can effectively and selectively oxidize CO in the presence
CO + 3H
CO + H O T H + CO
2
2
f CH
4
+ H
2
O
(1)
(2)
63
of hydrogen in agreement with previous literature reports. The
results show that the CO and O2 conversions proceed in parallel
with increasing reaction temperature up to 210 °C. At that point
complete conversion of both CO and O2 was reached simulta-
neously, yielding a selectivity of approximately 50%. At low
reaction temperatures, the surface of Pt is nearly covered to
saturation by CO due to the high heat of CO adsorption as
compared to H2. Therefore, the lack of sufficient H2 surface
concentration is believed to be limiting the H2 oxidation activity
under these conditions. However, at higher reaction tempera-
tures, the removal of CO from the Pt surface due to the reaction
with O2 or partial desorption creates vacant active sites capable
of accommodating H2, and therefore, accelerating the H2
oxidation reaction. Thus, the observed decrease in CO conver-
sion with increasing temperature (Figure 10) leaves a very
narrow window for successful operation of this catalyst in the
selective oxidation of CO and requires a precise control of the
reaction temperature. A simple comparison of the results
obtained over the same Pt/γ-Al2O3 sample for the oxidation of
CO in the absence (Figure 8A) and presence of H2 (Figure 10)
indicates that in the presence of H2 the light-off curve is shifted
to lower temperatures by approximately 30 °C, consistent with
previous literature reports.65 A subsequent increase of the
reaction temperature led to a decrease in the CO conversion
2
2
The results of independent kinetic measurements conducted with
a 0.5% CO-45% H mixture over the same catalysts indicate
2
no substantial activity for the methanation reaction (1) at
temperatures below 300 °C. This result is consistent with
previous reports indicating that the activity of Pt/γ-Al O for
6
4
2
3
CH formation under PROX conditions is negligible at tem-
4
70
peratures below 225 °C. Above this temperature, the formation
of methane increases slowly to a maximum of 0.025% at
7
0
approximately 350 °C.
It is thus more likely that the water-gas shift reaction (2) is
a major contributor to the observed conversions of CO in the
170-300 °C temperature range. The H O, which is required
2
for the reaction to proceed, is produced in situ as the product
of H oxidation. The equilibrium constant reported for this
reaction suggests that at aproximately 200 °C the forward
2
reaction should be dominant in the absence of large amounts
7
1
of both H and CO in the reaction stream. Indeed substantial
2
2
activity for the water-gas shift reaction has been reported
72
previously for Pt/TiO in the 200-400 °C range, closely
2
resembling our catalytic data (Figure 11). The water-gas shift
reaction appears to be structure insensitive, as the turnover
frequencies determined at differential conditions were essentially
(Figure 10), consistent with the loss of O2 due to the competing
7
2
oxidation of H2, leading to reduced selectivities at higher
constant over a wide range of Pt dispersions. However, it was
shown that these turnover frequencies depend strongly on the
type of TiO2 used, suggesting that the support plays an important
temperatures.6
6-68
When Pt/TiO2 was reduced at 200 °C, no activity for the
selective oxidation of CO was observed at low temperatures
and the light-off curve obtained resembled that observed for
the Pt/γ-Al2O3 sample, with a delay of approximately 30 °C
and a maximum CO conversion of only 62%. In contrast, when
the sample was reduced at 300 °C or higher, a low-temperature
activity was evident in the light-off curve, characterized by a
maximum of 30% CO conversion at 70 °C (Figure 11). These
results are in agreement with recent reports, illustrating the
7
2
role in this reaction. In general, the water-gas shift reaction
7
3
has been described by a redox mechanism, involving the
oxidation of CO adsorbed on Pt by oxygen from the support
followed by the restoration of the support surface with oxygen
7
4
atoms from H2O. Alternatively, an associative mechanism has
also been suggested involving the interaction of CO adsorbed
on the metal surface with hydroxyls on the support leading to
the formation of a formate-like intermediate, the subsequent
decomposition of which results in the formation of CO2 and
ability of Pt/TiO2 to selectively oxidize CO in hydrogen-rich
7
5
streams over a broad range of temperatures.69 The low-
H . In both cases, the support plays an important role and the
function of Pt is most likely restricted to the supply of CO
2
temperature activity observed is directly related to the reduction
temperature of the sample and 300 °C is the lowest reduction
temperature at which Pt/TiO2 shows such activity.
molecules to the metal support interface, emphasizing the
bifunctional nature of the reaction. Such a scheme is consistent
with our catalytic data demonstrating the lack of an effect by
the reduction temperature on the CO conversion due to the
water-gas shift reaction in the 170-300 °C temperature range
over Pt/TiO2 (Figure 11).
The striking result here is that in the case of Pt/TiO2 the light-
off curves for CO and oxygen do not resemble each other, as
was observed for Pt/γ-Al2O3 (Figure 10). Complete O2 con-
sumption, for example, is observed at 150 °C for the sample
reduced at 200 and 120 °C for the samples reduced at 300 and
Conclusions
4
00 °C (Figure 11), indicating that the oxidation of H2 to form
H2O proceeds with much higher rates over Pt/TiO2 than over
Pt/γ-Al2O3. Furthermore, the oxidation of H2 is affected by
the strength of the Pt-support interactions, since the light-off
curves characterizing the O2 consumption are shifted to lower
temperatures with increasing reduction temperature of Pt/TiO2
Platinum particles of various sizes were formed on γ-Al2O3
and TiO2 from a H2PtCl6‚6H2O precursor following reduction
in H2 at different temperatures in the 200-400 °C range. These
particles were characterized by FTIR, HRTEM, XPS, and
EXAFS and their catalytic properties were examined for the
oxidation of CO in the absence and presence of H2.
(Figure 11).
Finally, we should point out that substantial conversion of
When γ-Al2O3 was used as the support, Pt particles as small
as 9-11 Å on average were formed. In situ FTIR data show
that both the terminal and bridging CO species adsorbed on Pt
can be effectively oxidized. EXAFS data further show that
during the course of the CO oxidation reaction these Pt particles
CO with a maximum of about 62% was observed over the
various Pt/TiO2 samples in the 170-300 °C temperature range
with the light-off curves in this region being insensitive to the
reduction temperature (Figure 11). With no reason to believe

23442 J. Phys. Chem. B, Vol. 109, No. 49, 2005
Alexeev et al.
can undergo substantial structural changes induced by the strong
adsorption of the reactants. Exposure of Pt/γ-Al2O3 to a CO/
air mixture at 280 °C, for example, led to partial fragmentation
of the metal particles. As a result, extremely small platinum
clusters incorporating no more than 4-6 metal atoms strongly
interacting with the support were formed. Being exposed to O2
under these reaction conditions, such small Pt clusters can be
positively charged, and therefore, be less active for CO oxidation
than fully reduced Pt. The transformation of metal clusters to
these oxide-like species appeared to be reversible and mainly
responsible for the inflection points observed in the light-off
curves in the 215-235 °C temperature range. In contrast,
exposure of Pt/γ-Al2O3 to the same CO/air mixture at low
reaction temperatures did not result in any substantial changes
in the morphology of the platinum particles, since the Pt-Pt
first-shell coordination number remained almost unchanged.
However, the Pt-Pt bond distance was increased due to the
possible formation of Pt carbonyl species.
(11) Captain, D. K.; Roberts, K. L.; Amiridis, M. D. Catal. Today 1998,
4
2, 93.
(12) Zaikovskii V. I.; Ryndin, Yu. A.; Kovalchuk, V. I.; Plyasova, L.
M.; Kuznetsov, B. N.; Yermakov, Yu. I. Kinet. Katal. 1981, 22, 443.
(13) Jentoft, R. E.; Deutsch, S. E.; Gates, B. C. ReV. Sci. Instrum. 1996,
7, 2111.
6
(
(
14) Alexeev, O.; Panjabi, G.; Gates, B. C. J. Catal. 1998, 173, 196.
15) Vaarkamp, M.; Linders, J. C.; Koningsberger, D. C. Phys. B 1995,
2
08-209, 159.
(16) Koningsberger, D. C. In Synchrotron Techniques in Interfacial
Electrochemistry; Melendres, C. A., Tadjeddine, A., Eds.; Kluwer: Dor-
drecht, The Netherlands, 1994; p 181.
(
17) Stern, E. A. Phys. ReV. B 1993, 48, 9825.
(18) Brigham, E. O. The Fast Fourier Transform; Prentice Hall:
Englewood Cliffs, NJ, 1974.
19) Kirlin, P. S.; van Zon, F. B. M.; Koningsberger, D. C.; Gates, B.
C. J. Phys. Chem. 1990, 94, 8439.
20) van Zon, J. B. A. D.; Koningsberger, D. C.; van’t Blik, H. F. J.;
(
(
Sayers, D. E. J. Chem. Phys. 1985, 82, 5742.
(21) Alexeev, O.; Kim, D.-W.; Graham, G. W.; Shelef, M.; Gates, B.
C. J. Catal. 1999, 185, 170.
(22) Kip, B. J.; Duivenvoorden, F. B. M.; Koningsberger, D. C.; Prins,
R. J. Catal. 1987, 105, 26.
Platinum particles with an average diameter of approximately
6 Å were formed on the TiO2 support. In this case, the strength
(23) Little, L. H. Infrared Spectra of Adsorbed Species; Academic
Press: London, UK, 1975.
2
(24) Baker, R. T.; Bernal, S.; Calvino, J. J.; Perez-Omil, J. A.; Lopez-
of the Pt-support interactions depends mainly on the reduction
temperature and can lead to an increased electron density on
Pt, altering its chemisorptive properties and leading to a
weakening of the Pt-CO bonding. These factors appear to
promote the oxidation of CO. In addition, the higher activity
of TiO2-supported samples can also be attributed to the ability
of TiO2 to provide or stabilize highly reactive oxygen species
presumably located at the metal-support interface. However,
such oxygen species appear to be much more reactive toward
H2 than CO. As a result, the Pt/TiO2 samples examined show
poor selectivities for the CO oxidation under PROX conditions.
Cartes, C. Nanotechnol. Catal. 2004, 2, 403.
(25) Weber, R. S.; Boudart, M.; Gallezot, P. In Growth and Properties
of Metal Clusters; Bourdon J., Ed.; Elsevier: Amsterdam, The Netherlands,
1
980; p 415.
(26) Fukushima, T.; Katzer, J. R.; Sayers, D. E.; Cook. J. In Proceedings
of the 7th International Congress on Catalysis; Seiyama, T., Tanabe, K.,
Eds.; Elsevier: Amsterdam, The Netherlands, 1981; p 79.
(
27) Lee, T. J.; Kim, Y. G. J. Catal. 1984, 90, 279.
(28) Lietz, G.; Lieske, H.; Spindler, H.; Hanke, W.; V o¨ lter, J. J. Catal.
1
983, 81, 17.
29) Lieske, H.; Lietz, G.; Spindler, H.; V o¨ lter, J. J. Catal. 1983, 81, 8.
(30) D’Aniello, M. J.; Monroe, D. R.; Carr, C. J.; Krueger, M. H. J.
Catal. 1988, 109, 407.
31) Baker, R. T. K.; Prestridge, E. B.; Garten, R. L. J. Catal. 1979,
(
(
5
9, 293.
Acknowledgment. This work was supported by the U.S.
Department of Energy, Office of Basic Energy Sciences (DE-
FG02-00ER14980). Use of the National Synchrotron Light
Source, Brookhaven National Laboratory, was supported by the
U.S. Department of Energy, Office of Science, Office of Basic
Energy Sciences, under contract No. DE-AC02-98CH10886.
The authors further acknowledge the staff of beamline X-18B.
The EXAFS data were analyzed with the XDAP software
developed by Vaarkamp et al.15 A portion of this work was
also performed at the W.R. Wiley Environmental Molecular
Sciences Laboratory, a national scientific user facility sponsored
by the U.S. Department of Energy’s Office of Biological and
Environmental Research and located at Pacific Northwest
National Laboratory. Finally, the technical assistance of Dr.
Chongmin Wang and Ms. Sarah Holton in carrying out some
of the experiments is gratefully acknowledged.
(32) Pesty, F.; Steinr u¨ ck, H.-P.; Madey, T. E. Surf. Sci. 1995, 339, 83.
(33) Huizinga, T.; Prins, R. J. Phys. Chem. 1981, 85, 2156.
(
34) Mullins, D. R.; Zhang, K. Z. Surf. Sci. 2002, 513, 163.
(35) Bitter, J. H.; Cauqui, M. A.; Gatica, J. M.; Bernal, S.; Ramaker,
D. E.; Koningsberger, D. C. Stud. Surf. Sci. Catal. 2000, 130, 3183.
(36) Koningsberger, D. C.; Martens, J. H. A.; Prins, R.; Short, D. R.;
Sayers, D. E. J. Phys. Chem. 1986, 90, 3047.
(
37) Martens, J. H. A.; Prins, R.; Koningsberger, D. C. Catal. Lett. 1989,
, 211.
38) Sankar, G.; Vasudevan, S.; Rao, C. N. R. J. Phys. Chem. 1988,
92, 1878.
2
(
(
(
(
(
39) Blyholder, G. J. Phys. Chem. 1964, 68, 2772.
40) Linsebigler, A.; Lu, G.; Yates, J. T. Surf. Sci. 1993, 294, 284.
41) Alexeev, O.; Shelef, M.; Gates, B. C. J. Catal. 1996, 164. 1.
42) Delbecq, F.; Moraweck, B.; V e´ rit e´ , L. Surf. Sci. 1998, 396, 156.
(43) Xu, C.; Koel, B. E. Surf. Sci. 1994, 310, 198.
(
44) Xu, C.; Koel, B. E. Surf. Sci. 1995, 327, 38.
(45) Atli, A.; Abon, M.; Beccat, P.; Bertolini, J. C.; Tardy, B. Surf. Sci.
1
994, 302, 121.
46) Noordermeer, A.; Kok, G. A.; Nieuwenhuys, B. E. Surf. Sci. 1986,
72, 349.
(
1
(
47) Engel, T.; Ertl, G. AdV. Catal. 1979, 28, 2.
References and Notes
(48) Atkins, P. W. Physical Chemistr; Oxford University Press: Oxford,
UK, 1998, Chapter 28.
(1) Rotman, D. Chem. Week 1992, 8, 150.
(
(
(
(
(
49) Bourane, A.; Bianchi, D. J. Catal. 2002, 209, 126.
50) Bourane, A.; Bianchi, D. J. Catal. 2002, 209, 114.
51) Bourane, A.; Bianchi, D. J. Catal. 2001, 202, 34.
52) Bourane, A.; Dulaurent, O.; Bianchi, D. J. Catal. 2000, 196, 115.
53) Henderson, M. A.; Epling, W. S.; Perkins, C. L.; Peden, C. H. F.;
(
2) Tauster, S. J.; Fung, S. C.; Garten, R. L. J. Am. Chem. Soc. 1978,
1
00, 170.
3) Strong Metal Support Interactions; Baker, R. T. K., Tauster, S. J.,
Dumesic, J. A., Eds; American Chemical Society: Washington, D.C., 1986;
(
2
38 pp.
Diebold, U. J. Phys. Chem. B 1999, 103, 5328.
(
(
4) Alekseev, O. S.; Ryndin, Yu. A. Usp. Khim. 1992, 61, 765.
5) Argo, A. M.; Odzak, J. F.; Gates, B. C. Nature (London) 2002,
(
54) Uner, D.; Tapan, N. A.; O¨ zen, I˙ .; U¨ ner, M. Appl. Catal., A 2003,
2
51, 225.
4
2
2
15, 623.
(55) Einaga, H.; Ogata, A.; Futamura, S.; Ibusuki, T. Chem. Phys. Lett.
(6) Alexeev, O. S.; Li, F.; Amiridis, M. D.; Gates, B. C. J. Phys. Chem.
2
001, 338, 303.
005, 109, 2338.
(
56) Lane, G. S.; Wolf, E. E. J. Catal. 1987, 105, 386.
(
7) Vayssilov, G. N.; Gates, B. C.; R o¨ sch, N. Angew. Chem., Int. Ed.
003, 42, 1391.
8) Xu, Z.; Xiao, F.-S.; Purnell, S. K.; Alexeev, O.; Kawi, S.; Deutsch,
S. E.; Gates, B. C. Nature (London) 1994, 372, 346.
9) Santra, A. K.; Goodman, D. W. Electrochim. Acta 2002, 47, 3595.
10) Song, C. Catal. Today 2002, 77, 17.
(57) Chafik, T.; Dulaurent, O.; Gass, J. L.; Bianchi, D. J. Catal. 1998,
1
79, 503.
(
(58) Wartnaby, C. E.; Stuck, A.; Yeo, Y. Y.; King, D. A. J. Chem. Phys.
1
995, 102, 1855.
(59) Bourane, A.; Bianchi, D. J. Catal. 2004, 222, 499.
(60) Alexeev, O.; Gates, B. C. Top. Catal. 2000, 10, 273.
(
(

Study of TiO2- and γ-Al2O3-Supported Pt Catalysts
J. Phys. Chem. B, Vol. 109, No. 49, 2005 23443
61) Gates, B. C. J. Mol. Catal. A 2000, 163, 55.
(68) Kim, D. H.; Lim, M. S. Appl. Catal., A 2002, 224, 27.
(69) Epling, W. S.; Cheekatamarla, P. K.; Lane, A. M. Chem. Eng. 2003,
Ichikawa, M. Inorg. Chim. Acta 1999, 294, 281.
93, 61.
(
63) Korotkikh, O.; Farrauto, R. Catal. Today 2000, 62, 249.
64) Poelsema, B.; Mechtersheimer, G.; Comsa, G. Surf. Sci. 1981, 111,
(70) Han, Y.-F.; Kahlich, M. J.; Kinne, M.; Behm, R. J. Appl. Catal., B
2004, 50, 209.
(
5
9
19.
(71) Choi, Y.; Stenger, H. G. J. Power Sources 2004, 129, 246.
(72) Panagiotopoulou, P.; Kondarides, D. J. Catal. 2004, 225, 327.
(73) Barbier, J., Jr.; Duprez, D. Appl. Catal., B 1993, 3, 61.
(74) Wang, X.; Gorte, R. J.; Wagner, J. P. J. Catal. 2002, 212, 225.
(75) Shido, T.; Iwasawa, Y. J. Catal. 1993, 141, 71.
(
65) Oh, S. H.; Sinkevitch, R. M. J. Catal. 1993, 142, 254.
(
3.
(
66) Kahlich, M. J.; Gasteiger, H. A.; Behm, R. J. J. Catal. 1997, 171,
67) Manasilp, A.; Gulari, E. Appl. Catal., B 2002, 37, 17.