Sensitivity of catalysis to surface structure: The example of CO oxidation on Rh
under realistic conditions
J. Gustafson,
1,
* R. Westerström,
2
A. Mikkelsen,
2
X. Torrelles,
3
O. Balmes,
4
N. Bovet,
5
J. N. Andersen,
2
C. J. Baddeley,
1
and E. Lundgren
2
1
EaStCHEM School of Chemistry, University of St Andrews, St Andrews, Fife KY16 9ST, United Kingdom
2
Department of Synchrotron Radiation Research, Lund University, Box 118, SE-221 00 Lund, Sweden
3
Instituto de Ciencia de Materiales de Barcelona (CSIC), 08193 Bellaterra, Barcelona, Spain
4
ESRF, 6 rue Jules Horowitz, F-38043 Grenoble Cedex, France
5
MAX-lab, Lund University, Box 118, SE-221 00 Lund, Sweden
Received 17 March 2008; published 22 July 2008
Using a combination of surface x-ray diffraction and mass spectrometry at realistic pressures, the CO
oxidation reactivity of Rh111 and Rh100 model catalysts has been studied in conjunction with the surface
structure. The measurements show that the presence of a specific thin surface oxide is crucial for the high
activity of the Rh based CO oxidation. As this oxide is readily formed on all Rh facets, we conclude that the
specific Rh crystal planes exposed during catalysis will not directly influence the reactivity. This is fortified by
the very close similarity between the Rh111 and the Rh100 results.
DOI: 10.1103/PhysRevB.78.045423 PACS numbers: 82.65.+r, 61.05.cp, 68.47.Gh, 82.80.Ms
I. INTRODUCTION
Transition-metal based catalysts, consisting of dispersed
active metal nanoparticles on an insulating oxide support,
form the basis of much of modern chemistry. These nanopar-
ticles, or nanocrystals, expose various facets, the most abun-
dant of which, in the case of an fcc metal, are the 111 and
100 surfaces due to their low surface energies, according to
the Wulff construction.
1
The catalysis related properties of
different facets have therefore been extensively explored un-
der ultrahigh vacuum UHV conditions, using model single-
crystal surfaces, demonstrating that for many reactions, the
surface orientations present on the nanoparticle have a strong
impact on the catalytic activity.
2–6
At more realistic pres-
sures, the effects of the surface orientation on the catalytic
properties is in principle unexplored, both experimentally
and theoretically.
A very recent in situ reflectance absorption infrared spec-
troscopy RAIRS study of CO oxidation on Pt-group metals
under relevant conditions shows the existence of a “hyperac-
tive” oxygen-covered phase.
7
This could agree well with ear-
lier studies showing that the CO oxidation reaction over Pt
and Pd model catalysts is more efficient when a thin oxide is
present on the substrate surface,
8–10
a situation which may
correspond to that of a real catalyst at work. Although such
thin oxides have been shown to exist on a number of
transition-metal surfaces,
11
their role in catalysis at realistic
conditions is under debate.
12
In the case of Rh, one of the active components in auto-
motive catalytic converters,
13,14
no such studies have been
performed previously. It has, however, been shown that in
high oxygen partial pressures, a similar thin trilayer O-Rh-O
surface oxide is formed on all Rh surface orientations inves-
tigated so far.
15–19
These studies include low-index surfaces
as well as the 553 and 223 high-index surfaces, where the
stepped surface structure completely vanishes during the sur-
face oxide formation.
18,19
A very recent density-functional-
theory study by Mittendorfer et al.
20
showed that this also
applies to Rh nanoparticles.
In the present paper, we have used surface x-ray diffrac-
tion SXRD and mass spectrometry to investigate the rela-
tion between the presence of oxide structures and changes in
the CO oxidation activity over Rh111 and Rh100 surfaces
in situ at catalytically relevant pressures. Although there are
some minor differences, the general behavior is practically
identical for the two surface orientations. Starting with the
surface oxide in pure O
2
, introducing CO reduces the oxide
and leaves the surface in a metallic phase. At this point, mass
spectrometry reveals a low CO
2
production. As the reaction
proceeds, the O
2
/ CO ratio in the chamber rises and at one
point the surface oxide is reformed. Concurrently, a large
increase in the CO oxidation rate can be observed.
The results show that for CO oxidation, the Rh surface is
much more active in the surface oxide phase than in the
metallic phase. Since a similar surface oxide is formed on all
Rh surfaces, including nanoparticles, our results strongly
suggest that the reactivity of a Rh based CO oxidation cata-
lyst is not sensitive to the specific crystal planes exposed but
is governed by the surface oxide formation. This conclusion
is fortified by a very close similarity between the Rh111
and the Rh100 results.
II. EXPERIMENT
The measurements were performed in the high-pressure
chamber
21
at the surface diffraction beamline ID3 Ref. 22
at the European Synchrotron Radiation Facility ESRF in
Grenoble, France. The wavelength of the incident x rays was
set to 0.724 Å. The sample was aligned according to the
bulk Bragg reflections of the Rh substrates. The coordinates
H , K , L in reciprocal space refer to a basis b
1
, b
2
, b
3
with
b
1
and b
2
spanning the surface lattice of the Rh substrate, as
shown in Figs. 1b and 1c for 111 and 100, respec-
tively, and with b
3
perpendicular to the surface plane.
The CO oxidation measurements were performed in a so-
called batch reaction chamber, in which the system is first
stabilized in the presence of pure O
2
. CO is then introduced
PHYSICAL REVIEW B 78, 045423 2008
1098-0121/2008/784/0454236 ©2008 The American Physical Society 045423-1