5848 J. Phys. Chem. B, Vol. 108, No. 19, 2004
Letters
IRAS spectra of the surface before and after exposure to CO.
After each CO adsorption experiment, the bimetallic Ru/Pt
surface was annealed at ca. 650 K for 2.5 h, including 6 min at
-
8
ca. 650 K in an O2 atmosphere (1 × 10 Torr) to remove
residual carbon. Although measurements were performed over
the range 0.24 < θRu < 0.73 and a range of CO exposures,
only results obtained for θRu ) 0.52 and saturation CO coverage
will be reported in this work. Infrared features attributed to CO
adsorption on sites generated by the presence of Ru were
identified by subtracting, from the recorded spectra, contribu-
tions due to CO on pristine Pt (100), for which the IRAS
spectrum at 130 K displayed an atop CO stretching band at
-
1
2
100 cm , in excellent agreement with data reported by other
26
groups for the same surface under very similar conditions.
Results and Discussion
Shown in Figure 1, curve a, is the IRAS spectrum of a Ru-
modified Pt(100) (θRu ) 0.52) prepared by the procedures
specified in the Experimental Section, which was first annealed
for 42 min at 650 K and then exposed to CO to achieve
saturation coverage at 130 K. The corresponding spectra for
the same Ru/Pt surface following identical annealing/CO ex-
posure experiments run consecutively are given in sequence in
curves b through d in this same figure. As clearly indicated,
Figure 1. IRAS spectra for a Ru-modified Pt(100) (θRu ) 0.52)
prepared by the procedures specified in the Experimental Section. The
Ru-modified Pt(100) first annealed for 20 min at 650 K and then
exposed to CO to achieve saturation coverage at 130 K (curve a). The
corresponding spectra of identical annealing/CO exposure experiments
run in sequence for the same Ru/Pt surface are given in order in curves
b-d in this same figure. Also shown in this plot is the spectrum of
CO on pristine Pt(100) recorded under otherwise identical conditions.
-1
these curves show a well-defined isosbestic point at 2096 cm ,
providing strong evidence for a quantitative conversion of one
type of site into the other. A statistical analysis of these data
nitrogen-cooled cryostat mounted on a UHV chamber (base
pressure from mid 10-11 to ca. 1 × 10
-10
Torr) equipped with
-
1
IRAS, Auger electron spectroscopy (AES), XPS, and low-energy
electron diffraction (LEED) described in detail elsewhere.
was performed in terms of a component at 2100 cm due to
CO on pristine Pt(100) extracted directly from the IRAS spectra
of CO adsorbed under the same conditions on the Ru-free Pt-
(100) surface and a second peak attributed to CO adsorption
on Ru-modified sites (see Figure 2). The dotted and dashed lines
in each of the panels in this figure represent best fit components
2
4
Bimetallic Ru/Pt surfaces were prepared by first cleaning the
+
Pt(100) crystal with cycles of Ar sputtering and annealing to
-
8
1
073 K in an O2 atmosphere (ca. 1 × 10 Torr). The Pt(100)
crystal was characterized by XPS and then dosed with trisru-
thenium dodecacarbonyl, Ru3(CO)12, using procedures devel-
oped in our laboratory.13 As specified in that article, a few
Ru3(CO)12 crystals are placed in a copper boat collimator
assembly mounted on a transfer arm and stored under a dry Ar
atmosphere in another independently pumped auxiliary chamber
while not in use. For the actual deposition, this latter chamber
was evacuated, the gate valve that connects to the main UHV
chamber was opened, and the tip of the boat-collimator was
placed ca. 1.5 cm from the surface of the clean Pt(100) crystal
cooled to ca. 130 K to allow Ru3(CO)12 vapors to condense on
the surface while monitoring the deposition process with time-
resolved IRAS. After deposition, the Ru3(CO)12 layer was
decomposed by irradiating the sample with X-rays, while
simultaneously heating the sample to ca. 650 K. After turning
the X-ray source off, the sample was annealed at ca. 650 K in
-1
attributed to CO adsorbed linearly on Pt(100) sites (2100 cm
)
-1
and on Ru-modified sites (2092 cm ), respectively. As shown
in Figure 3, the integrated area of these two peaks normalized
by the average total area under the composite feature is
consistent with a monotonic decrease in the Ru-modified sites
(or equivalently an increase in the Pt(100) sites) from 80% down
to 25%. In harmony with the presence of the isosbestic point,
and within the uncertainties of this analysis, the sum of the two
contributing peaks was constant for all four runs (see solid
squares in Figure 3). Also shown in this figure (see right
ordinate, open squares) is the amount of Ru determined from
the XPS data for each of the Ru/Pt(100) surfaces, which was
found to be virtually constant for the specimen following each
annealing/CO adsorption cycle. As shown by the elegant low-
energy ion-scattering (LEIS) work of Hayden and co-workers
-
8
27
28,29
an O2 atmosphere (ca. 1 × 10 Torr) for 5 min, and then XPS
spectra were recorded to check for carbon and oxygen impuri-
ties. Ru coverages, θRu, quoted in this work, were calculated
based on the homogeneous attenuation model25 and, therefore,
cannot be generally correlated with actual surface coverages
for both Pt(111) and Pt(110),
Ru does incorporate into
the Pt surface region for temperatures much lower 650 K,
leading to a net decrease in the Ru coverage and, thus, a
corresponding Pt surface enrichment. Both the IRAS and LEIS
data underscore the limitations of XPS for the determination of
true surface coverages for Ru-modified Pt surfaces. Further
support for this model is provided by the earlier data reported
by Lamouri et al., who monitored a surface Pt enrichment (Ru
depletion) by TPD for which the low-temperature desorption
peak attributed to Ru-modified sites decreased (and that cor-
responding to Pt sites increased) as the time of annealing of
the bimetallic Ru/Pt surface was increased.
(see below).
For a typical experiment, freshly prepared bimetallic Ru/Pt
surfaces were first cooled to 133 K, then flash-annealed to ca.
50 K, and subsequently allowed to cool back to 133 K before
exposure to CO. The latter was effected using a gas leak valve
6
(
Varian) allowing for control of the partial pressure of CO in
-
11
the low 10
Torr range. The partial pressure of CO during
the runs was monitored using an ionization gauge, while
recording real time IRAS spectra at ca. 20 s intervals. As is
customary, IRAS data were displayed in the form ∆R/R ) (Rsam
In summary, the IRAS results obtained in this study afford a
means of monitoring changes in the surface composition of Ru-
modified Pt surfaces induced by thermal annealing, which serves
to complement information derived from TPD. As shown by
-
Rref)/Rref, where Rref and Rsam in this case represented the
Yee, Nelson
