Catalytic activity of Pt-deposited Rh(110) bimetallic surface for NO + H2 reaction

Article abstract of DOI:10.1021/jp962829z

The Pt/Rh(110) bimetallic surface was prepared by electrochemical deposition of Pt ions on at Rh(110) surface, The Pt/Rh(110) surface was stable at 760 K in vacuum, but Rh atoms were extracted onto the Pt layer a 760 K by heating in O₂. The LEED pattern sequentially changed from p(1 × 1) with high background to a p(1 × 3) and a c(2 × 2) by heating in O₂. When the c(2 × 2) surface was heated at 760 K in H₂, the LEED pattern changed to p(1 × n). Extraction of the Rh atom occurred also during the NO + H₂ reaction which was monitored by cyclic voltammograms. The p(1 × n) Pt/Rh(110) surface is highly active for the NO + H₂ reaction, which suggests that the extraction of Rh atoms onto Pt layer is responsible for the activation of the surface for the NO + H₂ reaction.

Full text of DOI:10.1021/jp962829z

1186
J. Phys. Chem. B 1997, 101, 1186-1189
Catalytic Activity of Pt-Deposited Rh(110) Bimetallic Surface for NO + H2 Reaction
Akira Sasahara, Hiroyuki Tamura, and Ken-ichi Tanaka*
The Institute for Solid State Physics, The UniVersity of Tokyo, 7-22-1 Roppongi, Minato-ku, Tokyo 106, Japan
ReceiVed: September 13, 1996; In Final Form: December 4, 1996X
The Pt/Rh(110) bimetallic surface was prepared by electrochemical deposition of Pt ions on a Rh(110) surface.
The Pt/Rh(110) surface was stable at 760 K in vacuum, but Rh atoms were extracted onto the Pt layer at 760
K by heating in O2. The LEED pattern sequentially changed from p(1×1) with high background to a p(1×3)
and a c(2×2) by heating in O2. When the c(2×2) surface was heated at 760 K in H2, the LEED pattern
changed to p(1×n). Extraction of the Rh atom occurred also during the NO + H2 reaction, which was
monitored by cyclic voltammograms. The p(1×n) Pt/Rh(110) surface is highly active for the NO + H2
reaction, which suggests that the extraction of Rh atoms onto Pt layer is responsible for the activation of the
surface for the NO + H2 reaction.
1. Introduction
combined p(3×1) + p4g p(2×2) structure. When this p(3×1)
+ p4g p(2×2) surface is exposed to H2 at room temperature,
the p(3×1) + p4g p(2×2) pattern is readily changed to the
(1×1) structure. When this p(1×1) surface is exposed to O2,
the p(3×1) pattern is recovered at room temperature. From
these results, it was deduced that the p(3×1) + p4g p(2×2)
structure is attributable to an oxygen-induced reconstruction of
Rh containing a (100)-like Pt-enriched layer. We studied the
catalytic activity of Pt0.25Rh0.75(100), Rh/Pt(100),7,8 and Pt/Rh-
(100)8,9 surfaces for the NO + H2 reaction as a model catalyst
for the NO reduction. It is worthy of note that the real catalysis
Pt-Rh catalyst is known by the name of the three-way
catalyst, which was developed for environmental protection from
NOx, hydrocarbon, and CO exhaust from automotive vehicles.1
A real Pt-Rh catalyst is prepared by impregnation of Pt and
Rh ions on Al2O3 or CeO2-Al2O3.2 It is an interesting feature
that a small amount of Rh is indispensable for getting prominent
catalytic activity of the Pt-based catalyst for NOx reduction or
CO oxidation. For example, it was shown that a supported Pt/
Rh catalyst has a higher activity than a supported Pt and a
supported Rh catalyst for the oxidation of CO in the presence
of NO.3 Dispersion of Rh or Pt atoms is also important for the
practical catalyst to get high catalytic activity. D’Aniello et
al.4 demonstrated that the catalytic activity was recovered by
redispersion of Pt or Rh atoms for an aged Pt-Rh by treating
with Cl2. So far, many efforts have been applied to get the
Pt-Rh catalyst with higher activity and longer life. However,
the role of Rh in the three-way catalyst is still not clear. To
throw light on this subject, we precisely studied the structure
and catalysis of the Pt/Rh(100) alloy surface and Rh-deposited
Pt(100) and Pt-deposited Rh(100) bimetallic surfaces. It was
shown that a Pt0.45Rh0.55 alloy tip annealed at 973 K is composed
of a Pt-enriched topmost layer and a Pt-depleted (Rh enriched)
second layer.5 On the other hand, it was proved that when a
Pt0.25Rh0.75(100) surface is annealed at higher than 950 K, the
Pt fraction of the alloy surface is lowered with increasing
annealing temperature and is almost equal to the Pt fraction of
the bulk composition at 1300 K. From these results, we deduced
that the alloy surface will be in equilibrium at higher than 950
K and that the equilibrium Pt fraction decreases with increasing
in the exhaust gas and the reaction of NO with H2 on Pt0.25
-
Rh0.75(100), Rh/Pt(100), and Pt/Rh(100) surfaces work under
entirely different circumstances; that is, the real catalyst works
in the presence of hydrocarbons, NOx, CO, SOx, and particulates
at rather high pressures.
Although the working conditions for the model catalyst and
the real catalyst are largely different, the phenomena observed
on the single-crystal surfaces may reflect the essential role of
Rh. In our previous studies, it was shown that Rh/Pt(100), Pt/
Rh(100), and Pt-Rh(100) surfaces are activated by reconstruc-
tion; that is, a common p(3×1) surface is established on these
three surfaces by O2 or NO. Therefore, it is interesting whether
other crystallographic planes are also activated by the recon-
struction. From this point of view, the structure and the catalytic
activity for NO + H2 reaction were studied on the Pt/Rh(110)
bimetallic surface that was prepared by electrochemical deposi-
tion of Pt ions on a p(1×1) clean Rh(110) surface. In our series
of experiments, it will be shown that the voltammogram is a
promising method for diagnosing the reconstructed surface
because it is sensitive to the composition of the topmost layer
of the working electrode. In this paper, the catalysis and surface
structure of the Pt/Rh(110) surface were studied by combining
the LEED and XPS analyses and the electrochemical method.10
annealing temperature. However, when a Pt-enriched Pt0.25
-
Rh0.75(100) surface is exposed to O2 or NO, a clear p(3×1)
LEED pattern appears with increasing Rh fraction on the surface
even at 450 K.6 It was also shown that a similar p(3×1)
structure is accomplished on a Rh-deposited Pt(100) surface8
when the surface is heated in O2 or NO at 400 K. From these
results, it was deduced that the p(3×1) pattern reflects the
reconstructive growth of a common surface structure such as
Rh-O/Pt-layer by the reaction of Rh atoms with oxygen.
Furthermore, it was proved that when a Pt-deposited Rh(100)
surface was heated in O2 at 600 K, the Rh atom was extracted
by reacting with oxygen and the LEED pattern changed to a
2. Experimental Section
A clean Rh(110) surface was obtained by repeated Ar+
sputtering and annealing in ultrahigh vacuum (UHV) at 1300
K. After that, the Rh(110) disk was transferred into a small
volume high-pressure cubic reactor attached to the main UHV
chamber, and the cubic reactor was filled with 1 atm of highly
pure Ar gas. After that, an electrochemical glass cell containing
2-
a solution of 0.05 M H2SO4 and 3.0 × 10-5 M PtCl6 was
* To whom the correspondence should be addressed.
X Abstract published in AdVance ACS Abstracts, January 1, 1997.
lifted up into the cubic reactor so that the solution surface
S1089-5647(96)02829-5 CCC: $14.00 © 1997 American Chemical Society

Catalytic Activity
J. Phys. Chem. B, Vol. 101, No. 7, 1997 1187
Figure 1. (a) Cyclic voltammogram of p(1×1) Rh(110) surface in 5 × 10-2 M H2SO4. (b) Cyclic voltammogram of Pt-deposited Rh(110) surface
in 5 × 10-2 M H2SO4.
meniscus makes contact with one side of Rh disk as described
in previous paper.7-9 Then the electrochemical deposition of
Pt ions was carried out by applying cyclic voltammetry. After
the deposition of a certain amount of Pt on the Rh(110) surface,
the Rh disk was washed in the cubic reactor with pure water to
avoid the effects of Cl- ions,11 and then the cyclic voltammo-
gram of the Pt-deposited Rh(110) surface in a 0.05 M H2SO4
solution was recorded. After the electrochemical measurement
in the H2SO4 solution, the Rh disk was pulled back to the UHV
chamber for characterization by LEED and XPS. No detectable
contamination was observed on the Pt-deposited Rh(110)
surface, provided that we used ultrapure water and highly pure
Ar gas.
Catalytic reduction of NO with H2 was carried out in the
Figure 2. N2/NO ratio in TPR on Rh(110) and Rh/Pt(110) surfaces.
prepared in the UHV system showed a clear p(1×1) LEED
pattern, but the spots were a little diffuse.
cubic cell by raising the temperature of the Pt/Rh(110) crystal
in a flow of a mixture of 5.8 × 10-9 Torr of NO and 1.6 ×
It is worthy of note that the hydrogen adsorption/desorption
10-8 Torr of H2. The temperature-programmed reaction (TPR)
of NO + H2 was monitored by measuring N2 with a quadrupole
mass analyzer, where the crystal temperature was raised in a
rate of 1 K/s.
peaks have almost equal potentials at the peak maximum on
both Rh(110) and Rh(100) surfaces.13 It is quite a contrast to
that on Pt(110) and Pt(100) surfaces, where the voltammogram
on Pt(110) is entirely different from that on Pt(100) surface.14,15
This fact indicates that the redox reaction of hydrogen, H+
+
e- a H, takes place at the same potential on (1×1) Rh(100)
and (1×1) Rh(110) surfaces but not on the Pt surfaces.
When a Rh(110) surface was heated in a flow of 5.8 × 10-9
Torr of NO and 1.6 × 10-8 Torr of H2, the N2 pressure started
to increase at ca. 360 K and made a peak at about 440 K and
was followed by a continuous increase with a maximum at about
550 K as shown in Figure 2 with open circles. The N2 peak
appearing at 440 K is given by a desorption-mediated reaction
of adsorbed NO, which is similar to that observed on the Rh-
(110) surface.16 The catalytic reaction of NO with H2 on Rh-
(110) was brought about at 550 K, which is lower than the
corresponding temperature of 600 K on Rh(100) surface.13 This
result indicates that the NO + H2 reaction on Rh(110) and Rh-
(100) surfaces is also as structure sensitive as that on Pt(100)
and Pt(110) surfaces,17 but it is not as remarkable compared to
that on Pt(110) and Pt(100) surfaces.
3. Results and Discussion
A cyclic voltammogram of a clean Rh(110) surface in a 0.05
M H2SO4 solution is shown in Figure 1a. Sharp peaks at -0.24
and -0.21 V(SCE) are attributable to the hydrogen adsorption
(H+ + e- f H(a)) and desorption (H(a) f H+ + e-) processes.
The charge for the hydrogen desorption peak at -0.21 V(SCE)
in Figure 1a was about 161 mC cm-2, which is almost equal to
the calculated value of 157 mC cm-2 for a flat Rh(110) surface
by assuming one hydrogen atom for each Rh atom on the (1×1)
Rh(100) surface. The obtained voltammogram (Figure 1a) is
very similar in shape to that reported by Go´mez et al.,12 but
peaks at -0.17 V(SCE) of their voltammogram are more sharp
compared to the corresponding peaks at -0.21 V in our
voltammogram, which may be reflected by the quality of the
Rh(110) surface. In our experiment, the Rh(110) surface

1188 J. Phys. Chem. B, Vol. 101, No. 7, 1997
Sasahara et al.
The as-deposited surface of ca. 1.5 monolayers of Pt ions on
Rh(110) gave a XPS peak ratio of Rh 3d5/2/Pt 4d3/2 ) 1.1, but
this surface showed no LEED pattern. No appreciable con-
tamination was detected on this surface by XPS. This reveals
that the deposition of Pt on Rh(110) surface is not pseudomor-
phic.
It was shown that the XPS peak for the Pt of a Pt-deposited
Rh(100) surface changes little by heating to 1050 K in UHV8,9
but is drastically changed by heating in O2, and a p(3×1) LEED
pattern is established at 600 K with increasing Rh atoms by
reacting with oxygen.8,9 Recently, our STM image shows that
the p(3×1) surface is reflected by an ordered arrangement of
Pt and Rh-O rows on the surface.18 Therefore, hereafter the
reconstructed surface is described by a (Pt-Rh-O-Pt). If we
compare the catalytic activity of Rh(100), Pt/Rh(100), and (Pt-
Rh-O-Pt)/Rh(100) for the reaction of NO + H2, it is clear
that the (Pt-Rh-O-Pt)/Rh(100) is extremely active for the NO
+ H2 reaction.13 From these results, it was deduced that the
reconstruction is responsible for the prominent catalytic activity
of the Pt-Rh catalyst for the NO + H2 reaction.
Figure 3. Change of the Rh 3d5/2/Pt 4d3/2 ratio of the XPS spectrum
and LEED pattern of Pt/Rh(110) surfaces.
As mentioned above, the Pt layer on the Pt/Rh(100) surface
is stable at ca. 1000 K in vacuum,8 but the XPS ratio of Rh
3d5/2/Pt 4d3/2 for the Pt/Rh(110) surface was increased by
annealing at 1000 K for 30 min. This fact indicates that the Pt
atoms on the Rh(110) surface are less stable than those on the
Rh(100) surface so that the Pt atoms are diffused into the Rh
crystal. In other words, thermal stability of the Pt layer on the
Rh surface depends on the crystallographic planes. A Pt/Rh-
(110) surface annealed at 1000 K, which gave a p(1×1) LEED
pattern, did not show a very high activity for the reduction of
NO with H2, as shown in Figure 2 with solid circles. It is
interesting that a characteristic N2 peak observed on the Rh-
(110) surface also appeared at ca. 440 K, as shown in Figure 2
with open circles, which may suggest that the Pt/Rh(110)
annealed at 1000 K is almost covered with Rh atoms instead of
Pt.
structure is established. To compare the catalytic activity of
the Pt/Rh(110) surface, the reaction of NO with H2 was
performed on a p(1×n) Pt/Rh(110) surface having a ratio of
Rh 3d5/2/Pt 4d3/2 ) 2.7, where the p(1×n) surface was previously
prepared by heating the c(2×2) surface in 1 × 10-7 Torr of H2
at 760 K. As mentioned above, the Rh/Pt ratio of the p(1×n)
Pt/Rh(110) surface was changed little by treating with H2 at
760 K, and it is the same on the c(2×2) surface obtained by
heating in O2. However, as shown in Figure 2 with triangles,
the p(1×n) Pt/Rh(110) surface showed a remarkable catalytic
activity for the reaction and N2 evolution occurred at ca. 400
K, which is almost the same temperature for the NO + H2
reaction on the p(3×1) (Rh-O-Pt)/Rh(100) surface.12 This
fact may reveal that similar active sites are formed on both (Rh-
O-Pt)/Rh(100) and c(2×2) or p(1×n) Pt/Rh(110) surfaces.
When an as-deposited Pt/Rh(110) surface was heated in 1 ×
10-7 Torr of H2 at 760 K for 10 min, the LEED pattern changed
to a p(1×n) but the Rh 3d5/2/Pt 4d3/2 ratio of the XPS peaks
changed little. This p(1×n) LEED pattern was stable at 760 K
in UHV and no O 1s XPS peak was detected. When the p(1×n)
surface was annealed at 760 K in 1 × 10-7 Torr of O2 for 60
min, the LEED pattern sequentially changed from the p(1×n)
to a p(1×3) pattern and finally to a c(2×2) pattern as shown in
Figure 3. In accordance with the LEED pattern change, the
Rh 3d5/2/Pt 4d3/2 ratio and the O 1s peak of the XPS spectrum
increased concomitantly as shown in Figures 3 and 4. Taking
account of the fact that the Pt and Rh atoms cannot diffuse very
rapidly at temperatures lower than 900 K,19 the rapid increase
of the Rh 3d5/2/Pt 4d3/2 ratio at 760 K is caused by the reaction
of Rh with O2. The p(1×n) and the p(1×3) patterns were stable
even at 760 K in H2, but the c(2×2) surface changed to a (1×n)
structure in 1 × 10-7 Torr of H2 at 760 K. It should be pointed
out that the c(2×2) is not influenced at room temperature by
exposure to H2. Contrary to the Pt/Rh(110) surface, the Pt/
Rh(100) surface gives the p(3×1) + p4g p(2×2) LEED pattern
by heating in O2 at 600 K and this surface readily changes to
the p(1×1) structure at room temperature by exposure to H2.
By reaction of Rh atoms with O, Rh atoms were segregated
on the Pt0.25Rh0.75(100) alloy surface by exposure to O2 or NO,20
and it is the same on the Pt/Rh(110) surface. The extraction of
Rh atoms on the Pt/Rh(110) surface by reacting with O2 can be
diagnosed by the cyclic voltammogram for the Pt/Rh(110)
2-
surface. To avoid the influence of Cl- ions from PtCl6 on
the voltammogram,11 the Pt/Rh(110) disk was washed with
superpure water and then the cyclic voltammogram was obtained
in 0.05 M H2SO4 solution. It is known that the peak at -0.21
V(SCE) attributable to the Rh(110) surface (Figure 1a) is
suppressed as the deposition of Pt atoms increases and a shoulder
appears at around -0.15 V(SCE), as shown in Figure 1b. After
the washing of this surface with superpure water, the Pt/Rh-
(110) disk was transferred to the main UHV chamber and was
characterized by LEED and XPS. The amount of Pt deposited
on the Rh(110) surface was estimated to be 0.9 monolayer by
the XPS, where no clear LEED pattern was observed.
Although this Pt/Rh(110) surface was heated at a rate of 1
K/s to 760 K in UHV, no clear LEED pattern was observed.
The Rh 3d5/2/Pt 4d3/2 ratio is not very sensitive to the
compositional change of the topmost layer,21 but the cyclic
voltammogram may be sensitive to the topmost layer. In fact,
the cyclic voltammogram for the surface annealed at 760 K in
UHV gave a new peak at ca. -0.15 V(SCE) as shown in Figure
5a, although the ratio of Rh/Pt changed little. By comparison
of the cyclic voltammogram of the annealed surface to that of
Rh(110) surface, the new peak at -0.15 V(SCE) is responsible
for the characteristic peak of the Pt layer. Therefore, the Pt
As mentioned above, Pt layer formed on the Rh surface has
different stability depending on the crystallographic planes of
Rh crystal in vacuum. However, the extraction of Rh by O2
was commonly observed on both the Pt/Rh(110) and Pt/Rh-
(100) surfaces. The Pt/Rh(100) surface gave high catalytic
activity for the NO reduction when Rh atoms are extracted by
reacting with O2; that is, the p(3×1) (Pt-Rh-O-Pt)/Rh(100)

Catalytic Activity
J. Phys. Chem. B, Vol. 101, No. 7, 1997 1189
The Pt/Rh(110) surface was subjected to the TPR of NO +
H2 by raising the temperature at a rate of 1 K/s. The Rh 3d5/2
/
Pt 4d3/2 ratio changed little during the TPR, but a new peak
appeared in the voltammogram at -0.22 V(SCE) as shown in
Figure 5b, which indicates the segregation of Rh atoms during
TPR; that is, the cyclic voltammetry sensitively proves the
segregation of Rh atoms on the topmost layer during the NO +
H2. From these results, we can conclude that a chemical
reconstruction of the Pt/Rh(110) surface may be responsible for
the activation of the Pt/Rh(110) catalyst, and (Pt-Rh-O-Pt)
constitution might be formed on these surfaces by reacting with
O2 or NO, which is the active surface for the reaction of NO +
H2. Considering the chemical activation forming a specific
constitution observed on Pt0.25Rh0.75(100), Rh/Pt(100), Pt/Rh-
(100), and Pt/Rh(110) surfaces, a similar activation could be
expected on the Pt-Rh/washcoat catalyst.
References and Notes
(1) King, D. A.; Woodruff, D. P. The Chemical Physics of Solid
Surfaces and Heterogeneous Catalysis; Elsevier: Amsterdam, 1982; Vol.
4.
(2) Keiski, R. L.; Ha¨rko¨nen, M.; Lahti, A.; Maunula, T.; Savima¨ki,
A.; Slotte, T. Catalysis and AutomotiVe Pollution Control III; Studies in
Surface Science and Catalysis, Vol. 96; Elsevier Science B.V.: Amsterdam,
1995.
Figure 4. Change of the O 1s peak of the XPS spectrum of Pt/Rh-
(110) surfaces: (i) Rh/Pt ) 1.3; (ii) Rh/Pt ) 1.8; (iii) Rh/Pt ) 2.5.
(3) Oh, S. H.; Carpenter E. J. Catal. 1986, 98, 178.
(4) D’Aniello, M. Jr.; Monroe, D. R.; Carr, C. J.; Krueger, M. H. J.
Catal. 1988, 109, 407
(5) Tsong, T. T.; Ren, D. M.; Ahmd, M. Phys. ReV. B 1988, 38, 7428.
(6) Yamada, T.; Hirano, H.; Tanaka, K. Surf. Sci. 1990, 226, 1.
(7) Taniguchi, M.; Kuzembaev, E. K.; Tanaka, K. Surf. Sci. Lett. 1993,
290, L711.
(8) Tamura, H.; Sasahara, A.; Tanaka, K. Langmuir 1994, 10, 4530.
(9) Tamura, H.; Sasahara, A.; Tanaka, K. Surf. Sci. Lett. 1994, 303,
L379.
(10) Stickney, J. L.; Rosasco, S. D.; Song, D.; Soriaga, M. P.; Hubbard,
A. T. Surf. Sci. 1983, 130, 326.
(11) Markovic, N.; Hanson, M.; McDougall, G.; Yeager, E. J. Elec-
troanal. Chem. 1986, 214, 555 .
(12) Go´mez, R.; Rodes, A.; Perez, J. M.; Feliu, J. M.; Aldaz, A. Surf.
Sci. 1995, 327, 202.
(13) Tamura, H.; Tanaka, K. J. Electroanal. Chem. 1995, 381, 95.
(14) Aramata, A.; Furuya, N. J. Electroanal. Chem. 1990, 295, 317.
(15) Yamamoto, K.; Kolb, D. M.; Kotz, R.; Lehmpful, G. J. Electroanal.
Chem. 1979, 96, 233.
(16) Baird, R. J.; Ku, R. C.; Wynblatt, P. Surf. Sci. 1980, 97, 346.
(17) Sasahara, A.; Tamura, H.; Tanaka, K. Catal. Lett. 1994, 28, 161.
Figure 5. (a) Cyclic voltammogram of Pt/Rh(110) surface (θPt ) 0.9)
heated to 760 K in 5 × 10-2 M H2SO4, where the Pt/Rh(110) was
annealed at 760 K in vacuum. (b) Cyclic voltammogram of Pt/Rh-
(110) surface (θPt ) 0.9) in 5 × 10-2 M H2SO4 after the TPR experiment
in a flow of NO and H2 to 760 K.
(18) Matsumoto, Y.; Okawa, Y.; Fujita, T.; Tanaka, K. Surf. Sci. 1996,
355, 109.
(19) Nieuwenhuys, B. E.; Siera, J.; Wolf, R. M. ISIJ Int. 1989, 29, 550.
(20) Hirano, H.; Yamada, T.; Tanaka, K. Surf. Sci. Lett. 1989, 222, L804.
atoms might arrange themselves in an array over the surface
by annealing.
(21) Tanuma, S.; Powell, C. J.; Penn, D. R. Surf. Interface Anal. 1993,
20, 77.