Kinetics and mechanism of NO reduction with CO on Ir surfaces

Article abstract of DOI:10.1016/j.jcat.2007.10.011

The adsorption and thermal reactivity of NO and CO and the kinetics of the NO reduction with CO on Ir surfaces were studied using X-ray photoelectron spectroscopy, polarization modulation infrared reflection-absorption spectroscopy, and temperature programmed desorption. The NO adsorption and dissociation activity was strongly dependent on the Ir surface structure. The NO dissociation activity of the Ir planes decreased in the order (100) > (211) ? (111). In contrast, the type of the CO adsorption site was independent of the Ir surface structure. The activity of Ir(111) for N₂ and CO₂ production from the NO + CO reaction was low compared with the activities of Ir(100) and Ir(211). The kinetic data for an Ir/SiO₂ powder catalyst were similar to data obtained for Ir(211). The order of the turnover frequencies for N₂ and CO₂ formation for the Ir planes was in good agreement with the order for NO dissociation activity, and this agreement indicates that the catalytic activity for NO reduction was dependent on NO dissociation. A kinetic study of the elementary steps indicated that the rate-limiting step for NO reduction with CO was the NO dissociation step.

Full text of DOI:10.1016/j.jcat.2007.10.011

Journal of Catalysis 253 (2008) 139–147
www.elsevier.com/locate/jcat
Kinetics and mechanism of NO reduction with CO on Ir surfaces
T. Fujitani ∗, I. Nakamura, A. Takahashi, M. Haneda, H. Hamada
Research Institute for Innovation in Sustainable Chemistry, National Institute of Advanced Industrial Science and Technology (AIST), 16-1 Onogawa, Tsukuba,
Ibaraki 305-8569, Japan
Received 4 September 2007; revised 10 October 2007; accepted 12 October 2007
Available online 19 November 2007
Abstract
The adsorption and thermal reactivity of NO and CO and the kinetics of the NO reduction with CO on Ir surfaces were studied using X-ray
photoelectron spectroscopy, polarization modulation infrared reflection–absorption spectroscopy, and temperature programmed desorption. The
NO adsorption and dissociation activity was strongly dependent on the Ir surface structure. The NO dissociation activity of the Ir planes decreased
in the order (100) > (211) ꢀ (111). In contrast, the type of the CO adsorption site was independent of the Ir surface structure. The activity of
Ir(111) for N and CO production from the NO + CO reaction was low compared with the activities of Ir(100) and Ir(211). The kinetic data for
2
2
an Ir/SiO powder catalyst were similar to data obtained for Ir(211). The order of the turnover frequencies for N and CO formation for the Ir
2
2
2
planes was in good agreement with the order for NO dissociation activity, and this agreement indicates that the catalytic activity for NO reduction
was dependent on NO dissociation. A kinetic study of the elementary steps indicated that the rate-limiting step for NO reduction with CO was the
NO dissociation step.
© 2007 Elsevier Inc. All rights reserved.
Keywords: Ir single crystal; Ir/SiO ; X-ray photoelectron spectroscopy; Polarization modulation infrared reflection–absorption spectroscopy
2
1. Introduction
Shimokawabe and Umeda [7] examined the reduction of NO
by CO in the presence of excess O2 and found that of seven
metals (Ag, Cu, Fe, Ir, Pd, Pt, and Rh) supported on metal ox-
ides, Ir/WO3, Ir/ZnO, and Rh/Al2O3 show the most pronounced
catalytic activities for the reduction of NO by CO. Takahashi et
al. [8] also reported that the catalytic activity for the selective
reduction of NO with CO over Ir/SiO2 is greatly increased by
the addition of a W species.
The reaction of NO and CO has been studied over various
transition- and noble-metal catalysts because of its importance
for three-way catalytic converters [9–13]. However, clarifying
the reaction mechanism, kinetics, and the role of the active site
using only industrially applied heterogeneous catalysts is diffi-
cult. Single crystals and model catalysts provide well-defined
surface geometries and allow detailed atomic-level characteri-
zation of the post-reaction surface and adsorbate-converted sur-
face to be performed. These types of studies have clarified the
dependence of the catalytic behavior on the crystal structure,
in addition to providing information regarding the identity and
role of the active site, reaction intermediates, promoters, and
inhibiters [14,15]. Hopstaken et al. [16] and Root et al. [17] re-
ported that the dissociation and desorption of NO from Rh(111)
Selective catalytic reduction is a promising method for re-
moving NO, a pollutant emitted from diesel engines, lean-burn
engines, and combustors. Numerous studies have investigated
the selective catalytic reduction of NO using reductants such
as NH3 and hydrocarbons [1–3]. H2 and CO have not been re-
garded as effective reductants in the presence of O2, because
they are more easily oxidized by O2 than by NO. Recently,
Ir-based catalysts were found to be effective for the selective
reduction of NO with CO in the presence of O2. Ogura et al. [4]
found that NO is selectively reduced with CO on Ir/silicalite in
an oxidizing atmosphere. Wang et al. [5] investigated the re-
duction of NO with CO over Pd/ZSM-5, Pt/ZSM-5, Ir/ZSM-5,
and Rh/ZSM-5 catalysts in the presence of excess oxygen; of
these catalysts, Ir/ZSM-5 exhibits the highest activity. Ir/SiO2
has also been reported to be an excellent catalyst for the re-
duction of NO by CO in the presence of O2 and SO2 [6].
*
Corresponding author. Fax: +81 298 61 8172.
E-mail address: t-fujitani@aist.go.jp (T. Fujitani).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.10.011

140
T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
are not markedly influenced by the presence of CO, whereas
repulsion between NO and CO destabilizes CO. Lintz et al.
[18,19] studied the elementary step for the NO + CO reaction
using Pt, Pd, and Rh. In addition, the NO + CO reaction on
model surfaces under high pressure has been studied, and these
studies provide information on real reaction kinetics and in situ
adsorption behavior. Goodman et al. [20,21] carried out a high-
pressure kinetic study of the CO + NO reaction on Pd single
crystals, model planar catalysts, and conventional high-surface-
area Pd/Al2O3 catalysts. The study showed that the reaction rate
is enhanced over the Pd(111) surface and over larger particles in
the model supported catalysts relative to the rates over Pd(100),
Pd(110), and smaller supported particles; these results indicate
that an inactive atomic N species plays a role in determining the
reaction rate. Schmieg et al. [22] studied the NO + CO reaction
at high-pressure over Rh(111) and Rh(110) to investigate the
effect of the Rh surface structure and selectivity; they showed
the NO + CO reaction is sensitive to the Rh structure primarily
with regard to the selectivity of the reaction for N2O production.
However, the reaction mechanism, kinetics, and adsorption be-
havior for the NO + CO reaction on Ir-based catalyst surfaces
are not fully understood [23,24].
equipment (<8×10−11 Torr); and a reaction chamber in which
the NO and CO adsorption experiments and reactions un-
der high-pressure conditions (<1000 Torr) were performed.
The pressure of the reaction chamber could be varied from
8 × 10−10 to 800 Torr. An IR spectrometer and the polar-
ization modulation (PM) optics were situated in the reaction
chamber (GWC Technologies, Model SSD-50; Hinds Instru-
ments, Model PEM-50 Controller). The IR beam reflected from
the surface was collected by a liquid-nitrogen-cooled mercury
cadmium telluride IR detector. The photoelastic modulator con-
troller was set to 2500 cm−1. The PM-IRA spectra for the
NO + CO reaction were recorded at a resolution of 4 cm−1
with 500 scans over a total measurement time of 5 min. The
IRA spectra for the NO and CO adsorption experiments were
recorded at a resolution of 4 cm−1 with 100 scans over a total
measurement time of 30 s.
2.2. Kinetic measurements and testing of surface reaction
properties
The NO + CO reaction to form CO2 and N2 was per-
formed in a batch mode. The standard reaction gas pressures of
NO (99.99%, TAIYO NIPPON SANSO Corporation) and CO
(99.999% pure, TAIYO NIPPON SANSO Corporation) were 5
and 5 Torr, respectively. The NO was purified by five freeze–
pump–thaw cycles using liquid nitrogen. The CO was in an
Al cylinder and was trapped with a liquid nitrogen bath to ex-
clude any metal carbonyls. The samples were controlled to the
reaction temperature and the gate valve to reaction chamber
and turbomolecular pump was closed, and then reactants were
dosed in the reaction chamber. The reaction temperature was
573–723 K, and the reaction time was 60–90 min. After the
reaction, He gas was fed through the reaction chamber at a to-
tal pressure of 760 Torr. The reaction gases diluted with He
were analyzed with a micro gas chromatograph (µ-GC, Agilent,
3000A), which was directly connected to the reaction chamber.
The µ-GC was equipped with a Molecular Sieve 5A column
(for the analysis of N2, NO, and CO) and a Porapak Q col-
umn (for the analysis of CO2 and N2O). Turnover frequencies
(TOFs) were obtained by normalizing the gas-phase production
per second to the total number of surface sites. The kinetic data
for reaction on the Ir/SiO2 catalyst were acquired in the flow
mode with CO and NO partial pressures of 5 and 5 Torr, re-
spectively, in a He carrier gas.
In this study, we investigated the adsorption and thermal re-
activity of NO and CO over Ir single crystals to clarify the
influence of the Ir surface structure. Furthermore, we examined
the kinetics and mechanism of the NO + CO reaction over Ir
single crystals at high-pressure and compared the results with
results for a conventional Ir/SiO2 powder catalyst.
2. Experimental
2.1. Apparatus
X-ray photoelectron spectroscopy (XPS) experiments were
performed in an ultrahigh-vacuum (UHV) apparatus composed
of two chambers (VG, ESCALAB 220-i): a surface-analysis
chamber (<1 × 10−10 Torr) and a preparation chamber (<1 ×
10−8 Torr). The analysis chamber was equipped with an ion
gun for Ar+ sputtering, a photoelectron analyzer for XPS, and
dosers for adjusting the NO and CO exposures. XPS spectra
were measured with MgKα radiation. Coverage was estimated
by XPS on the basis of the oxygen-buildup curve at 300 K; the
oxygen saturation coverage on the Ir surfaces was 0.5 (a cov-
erage value (Θ) of 1 corresponds to the number of surface Ir
atoms) [25]. The nitrogen, oxygen, and carbon coverages were
determined from the N 1s/Ir 4f7/2, O 1s/Ir 4f7/2, and C 1s/Ir 4f7/2
peak area ratios, using the O 1s/Ir 4f7/2 peak area ratio obtained
from the oxygen saturation coverage and the sensitivity factors
for N 1s, O 1s, and C 1s [26].
The infrared reflection–absorption spectroscopy (IRAS,
Mattson, RS/2) and temperature-programmed desorption (TPD)
experiments were carried out in a UHV apparatus composed
of four chambers: a load–lock chamber for changing samples
(<5 × 10−10 Torr); a preparation chamber equipped with an
ion gun for Ar+ sputtering; an analysis chamber for the Auger
electron spectroscopy (AES, OMICRON, SPECTALEED), and
quadrupole mass spectroscopy (SPECTRA, VAC-CHECK)
Adsorption experiments were carried out with 15NO (99.9%
pure, TAIYO NIPPON SANSO Corporation) and CO at 5 ×
10−9 to 5 × 10−8 Torr (0–30 L; 1 L = 10−6 Torrs) and a sam-
ple temperature of 223 K. 15NO gas was used to distinguish
between N2 and CO desorption peaks and between N2O and
CO2 desorption peaks during TPD. The TPD experiments were
performed with a 1.0 K/s heating rate.
2.3. Sample preparation
We used single-crystal discs of Ir(111), Ir(100), and Ir(211)
(8-mm diameter, 1-mm thickness, 99.999% purity) polished
on one side only. Ir(211) consists of three-atom-wide (111)

T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
141
terraces separated by one-atom-high (100) steps. The crystal
orientations were accurate to within 1◦, and the surface rough-
nesses were <0.03 µm. The Ir surfaces were cleaned by cycles
of Ar+ sputtering and annealing in oxygen at 1000 K and then
by annealing at 1250 K in a vacuum. The surface cleanliness of
the samples was verified by XPS or AES.
Silica-supported Ir catalyst was prepared by impregnating
SiO2 (Fuji Silysia Chemicals, Cariact G-10, 300 m2/g) with an
aqueous solution of H2IrCl6·6H2O (Soekawa Chemicals). The
impregnated catalyst precursors were dried at 383 K and cal-
cined at 873 K for 8 h in air. The loading of noble metals was
fixed at 0.5 wt%. The number of exposed Ir atoms for Ir/SiO2
was estimated by the amount of CO adsorption. The Ir disper-
sion was calculated to be 0.33% and the total number of active
sites in the catalyst was estimated to be 2 × 1020 sites/g-Ir.
15 L. However, our previous XPS results indicated that the NO
adsorption on hollow site was present even at low NO cover-
age [23]. Thus, it is possible that the NO simultaneously ad-
sorbed on atop and hollow sites. From the O 1s peaks for XPS
measurements, the saturation coverages of atop NO (532.8 eV)
and hollow NO (530.4 eV) were estimated to be 0.15 and 0.32,
respectively.
We observed two peaks, at 1570–1620 and 1760–1790 cm−1
during exposure of Ir(100) to 0.2–15 L of NO. The peak at low
frequency is in good agreement with peaks in the frequency re-
gion reported for NO adsorption on bridge sites of Rh(100) [27]
and Pt(100) [28]. We therefore assigned the peak at 1570–
1620 cm−1 to NO adsorbed on the bridge site of Ir(100). The
peak at high frequency was observed at about 30 cm−1 lower
than the peak observed for atop NO on Ir(111). Nyberg and
Uvdal [29] reported that the frequency of atop NO on Pd(100)
appears at 1667 cm−1, which is about 30 cm−1 lower than
for atop NO on Pd(111). We thus assigned the peak at 1760–
1790 cm−1 to NO adsorbed on the atop site on Ir(100). The
peak intensities of both atop NO and bridge NO increased with
increasing NO exposure and saturated at about 3 L. We also in-
vestigated the NO adsorption process on Ir(100) using XPS and
observed two O 1s peaks, at 531.2 and 532.8 eV, which were in
good agreement with reported values for the binding energies
of bridge NO and atop NO over Ni(100) [30] and Pt(111) [31],
respectively. Saturation coverage was obtained at 3 L, where
the saturation coverages of atop NO and bridge NO were 0.15
and 0.30, respectively.
,
3. Results and discussion
3.1. NO reactivity on the Ir surfaces
We examined the NO adsorption behavior on the Ir(111),
Ir(100), and Ir(211) surfaces using IRAS. Fig. 1 shows IRA
spectra of the Ir surfaces exposed to NO at 223 K. We observed
two peaks, at 1400–1444 and 1799–1820 cm−1, during expo-
sure of Ir(111) to 0.2–15 L of NO. The peak at 1799–1820 cm−1
was assigned to NO adsorbed on the atop site, and the peak
at 1400–1444 cm−1 was due to NO adsorbed on the hollow
site [23]. The intensity of the peak for atop NO increased with
increasing NO exposure, and then saturated at 2 L. After the
adsorption of atop NO was saturated, NO began to adsorb on
the hollow site, and NO saturation coverage was obtained at
On the Ir(211) surface, we observed two peaks, at 1547–
1604 and 1768–1809 cm−1, during exposure to 0.2–15 L of NO
at 223 K. The peak at 1768–1809 cm−1 was assigned to NO
adsorption on the atop site of the (111) terrace of Ir(211) [24].
The peak at 1547–1604 cm−1 was due to NO adsorbed on the
bridge site of the (100) step over Ir(211) [24]. The coverage
of both atop NO and bridge NO saturated at about 3 L. The
saturation coverage of atop NO and bridge NO were 0.19 and
0.21, respectively, which were estimated from O 1s peaks at
531.2 and 532.8 eV.
We next examined the thermal reactivity of NO adsorbed
on Ir(111), Ir(100), and Ir(211) (Fig. 2). In the TPD spectra
for 15NO and 15N2 after saturation of the Ir(111) surface with
15NO at 223 K, we did not observe any desorption peaks for
15N2O and O2 up to 1300 K, which indicated that atomic oxy-
gen (Oa) dissociated from NO diffused into the Ir subsurface.
Two desorption peaks for 15NO were observed, at 393 and
455 K, which were assigned to desorption of 15NO adsorbed
on the hollow and atop sites, respectively [23]. Simultaneously
with 15NO desorption, 15N2 desorption peaks were observed at
471 and 574 K. We assigned these two 15N2 peaks to a dis-
proportionation reaction between atop NO and atomic nitrogen
(Na) dissociated from hollow NO and to the recombination of
Na, respectively [23]. A desorption peak for 15N2 was observed
at 390 K on Ir(100), whereas no desorption peak for 15NO ap-
peared. The adsorbed NO on the atop and bridge sites on Ir(100)
completely dissociated to Na at 350 K, as indicated by the XPS
measurement. Therefore we assigned the desorption peak for
15N2 at 390 K to the recombination of Na from both adsorbed
15
Fig. 1. IRA spectra of Ir(111), Ir(211), and Ir(100) after exposure to NO at
223 K.

142
T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
15
Fig. 3. IRA spectra of Ir(111), Ir(211), and Ir(100) after exposure to CO at 15 L
and 223 K.
Fig. 2. TPD spectra for NO adsorbed on Ir(111), Ir(211), and Ir(100).
NO species on Ir(100). Furthermore, the desorption tempera-
ture for recombination of Na on Ir(100) was 200 K lower than
that on Ir(111), which indicates that the Ir(100) surface showed
a high efficiency for N2 formation from NO. In TPD spectra on
Ir(211), the 15N2 desorption peak was observed at 590 K, and
a small 15NO desorption peak was seen at 440 K, which indi-
cates that almost all the 15NO adsorbed on Ir(211) decomposed
to 15N2. We thus found that both atop NO and bridge NO dis-
sociated on Ir(211). About twice as much N2 formed on Ir(211)
as on Ir(111), which indicates that the NO dissociation activity
was higher over Ir(211) than over Ir(111).
The NO dissociation activity of the Ir planes decreased in the
order (100) > (211) > (111). The thermal dissociation of NO is
known not to occur on the low index planes such as Pd(111) and
Pd(100) [32]. Stepped surfaces such as Pd(311) and Pd(112) are
active for the thermal dissociation of NO [32,33]. In contrast,
the (100) plane was the most active Ir surface structure for NO
dissociation. However, the details of the dependence of the NO
dissociation activity on surface structure were not clear.
Ir(111) surface, which indicates that the CO was adsorbed on
atop sites on Ir(100) and Ir(211).
We thus believe that CO adsorption site on the Ir surface
was not dependent on the Ir surface structures. This result was
very different from other transition metal surfaces such as Pd
[34,35], Pt [36,37], and Rh [38,39]. It is very hard to discuss
the difference in adsorption properties of Ir surface from only
experimental results. We have been investigating the unique
specificity of CO adsorption on the Ir surface using density
functional theory (DFT) calculation.
On Ir(211), we could not determine whether the adsorption
plane was the (111) terrace or the (100) step from the frequency
of atop CO. The intensities of the atop-CO peak were dependent
on the Ir surface structure. However, from the XPS results, we
estimated that the saturated atop-CO coverage on all the Ir sur-
faces was 0.33, which suggests that the difference of the peak
intensities for atop CO on the Ir surfaces was not due to ad-
sorbed CO coverage. We also measured the initial adsorption
rate for atop CO on the Ir surfaces from the build-up curve for
CO adsorption using XPS (Fig. 4). From the slope of the line at
zero coverage in Fig. 4, we estimated initial adsorption rate to
be 7.3×10−4–7.6×10−4 (molecules/(site s)) for all the Ir sur-
faces. Thus, we clearly show that the CO adsorption properties
such as adsorption site, adsorption rate, and saturation coverage
were not dependent on the Ir surface structure, which were dif-
ferent from the adsorption properties of other transition metal
surfaces [34–39]. At present, the reason for the difference in
adsorption properties of Ir surface is not clear.
3.2. CO reactivity on the Ir surfaces
We also investigated the adsorption properties of CO on
Ir(111), Ir(100), and Ir(211). Fig. 3 shows the IRA spectra for
the Ir surfaces at CO exposure of 1, 3, and 15 L at 223 K. The
adsorbed CO peaks shifted to higher frequency with increasing
CO exposure for all the Ir surfaces. The intensities and frequen-
cies of adsorbed CO peaks on the Ir surfaces became constant
above 10 L. For saturation CO coverage, a peak of CO adsorbed
on atop sites was observed at 2073 cm−1 in the IRA spec-
tra for the Ir(111) surface [23]. Peaks were observed at 2085
and 2083 cm−1 in the IRA spectra for Ir(100) and Ir(211), and
these peaks were close to the peak for atop CO observed on the
We examined the thermal reaction properties of atop CO on
the Ir surfaces. TPD spectra were obtained for the Ir surfaces
after various CO exposures at 223 K (Fig. 5). CO desorption
peaks at 550 and 640 K were observed for the Ir(111) and

T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
143
cates that the CO desorption peak at 670 K is due to atop CO
on the (100) step over Ir(211). Above 2 L, a new desorption
peak appeared at 560 K, and the area of the peak increased with
increasing CO exposure up to 10 L. The temperature for this
new desorption peak was close to the desorption temperature
of CO adsorbed on Ir(111). We therefore identified the peak at
560 K as atop CO on the (111) terrace. We estimated the cover-
age for atop CO on the (111) terrace and (100) step over Ir(211)
from the TPD peak area for the saturated CO-covered Ir(211)
surface. The CO desorption ratios for the (111) terrace and the
(100) step were about 2 and 1, respectively, and these values are
in good agreement with the atomic ratios for the (111) terrace
and (100) step over Ir(211). At low CO coverage, CO adsorbed
only on the atop site of the (100) step. CO adsorption on the
(100) step saturated at 2 L, and then CO began to adsorb at the
(111) terrace above 2 L. We thus clarified the CO adsorption
behavior by TPD experiments.
Fig. 4. Buildup of CO on Ir(111) ("), Ir(211) (Q), and Ir(100) (F) at 223 K.
3.3. Reaction kinetics of NO reduction with CO on the Ir
surfaces
The kinetics of the NO reduction with CO over the Ir single
crystal and the powder Ir/SiO2 catalyst were investigated under
real reaction conditions. N2 and CO2 were formed by the re-
action of NO + CO. No N2O formation was observed on any
of the Ir surfaces. The production ratios for N2 and CO2 were
about 1 and 2, respectively, which indicates that the following
reaction proceeded on the Ir metal surfaces:
NO + CO → (1/2)N2 + CO2.
In addition to N2 and CO2, N2O is produced from the
NO + CO reaction (PNO = PCO = 8 Torr) over Rh(110) and
Rh(111) [22]. Furthermore, high N2O formation from the NO
+ CO reaction (PNO = PCO = 1 Torr) is observed on Pd single-
crystal surfaces [20]. In particular, the N2O selectivity on the Pd
surface increases with increasing reaction temperatures. Thus,
in comparison, the Ir metal surface has excellent selectivity for
N2 formation from the NO + CO reaction.
Fig. 6 shows Arrhenius plots of the TOFs for N2 and CO2
formation from the NO + CO reaction on the Ir surfaces as
well as on an Ir/SiO2 powder catalyst. As can be seen from the
figure, the Ir(111) plane showed low activity for N2 and CO2
production compared with the Ir(100) and Ir(211) planes. The
TOFs for N2 and CO2 formation for the Ir planes decreased in
the order (100) ꢀ (211) ꢀ (111); for example, the TOF for
N2 formation at 600 K on Ir(100) was about 3 and 50 times
the TOFs on Ir(211) and Ir(111), respectively. Table 1 sum-
marizes the apparent activation energies and pre-exponential
factors for N2 and CO2 formation over the Ir surfaces. The
apparent activation energies for N2 and CO2 formation were es-
timated to be 107–115 and 110–120 kJ/mol, respectively; and
these values were almost the same, regardless of the changes
of the Ir surface structure. The kinetics of the NO + CO reac-
tion have been studied using Pd [20] and Rh [22] single-crystal
surfaces. In these studies, the apparent activation energies for
N2 and CO2 formation were found to be almost independent
Fig. 5. TPD spectra for CO adsorbed on Ir(111), Ir(211), and Ir(100).
Ir(100) surfaces, respectively, exposed to 15 L of CO. Although
the CO adsorption site was similar for both surfaces, the ther-
mal stability of the CO adsorbed on Ir(100) was higher than
that of the CO adsorbed on Ir(111). Orita calculated the adsorp-
tion energy of atop CO on Ir(111) and Ir(100) using DFT [40].
The CO adsorption energy on Ir(100) was 0.2–0.5 eV higher
than that on Ir(111), which was in agreement with the ther-
mal stability of the CO adsorbed on the Ir surfaces. A single
CO desorption peak at 670 K was observed for the Ir(211) sur-
face exposed to 1.0 L of CO; this temperature agrees with the
desorption temperature of CO adsorbed on Ir(100), which indi-

144
T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
Table 1
mation
Apparent activation energies and pre-exponential factors for N and CO for-
2
2
−1
)
Catalyst E (kJ/mol)
υ (s
formation CO formation
a
N
formation CO formation
N
2
2
2
2
a
9.3 0.6
7.5 1.7
9.9 1.5
9.1 1.4
10.2 0.6
8.1 1.5
Ir/SiO
117.2 4.1
107.0 10.6 109.6 9.6
115.3 9.5
111.7 8.7
123.1 4.1
10
10
10
10
10
10
10
10
2
b
Ir(111)
Ir(100)
Ir(211)
a
b
b
10.5 1.3
10.3 1.3
117.8 8.0
120.3 8.6
Flow reaction at P = P = 5 Torr in a He carrier gas.
NO
CO
b
Batch reaction at P = P = 5 Torr.
NO
CO
reaction proceeds on the Ir metallic surface over an Ir/SiO2
powder catalyst. The TOFs for N2 and CO2 formation over
Ir/SiO2 were almost same for those for Ir(211), which indi-
cates that the Ir surface structure of Ir/SiO2 was close to that
of Ir(211). That is, the Ir(211) surface can be regarded as an
appropriate model catalyst for the NO + CO reaction over an
Ir/SiO2 catalyst because the apparent activation energies and
the TOFs for N2 and CO2 formation over Ir(211) were in good
agreement with those over the Ir/SiO2 catalyst. It is generally
considered that most of the N2 is produced via NO dissociation
and subsequent combination of Na. If the N2 is produced from
this mechanism on the Ir surfaces, the reaction temperature for
N2 formation is dependent on the Ir surface structure because
the N2 desorption temperature on Ir(100) is 200 K lower than
those on Ir(211) and Ir(111). However, N2 formation activities
were observed at the same reaction temperature region for all
the Ir surfaces, suggesting that the N2 was not formed by the
recombination of Na during NO + CO reaction.
To clarify the details of the reaction mechanism, we mea-
sured surface adsorbates on Ir(211) under real reaction condi-
tions using in situ PM-IRAS. We observed isocyanate (NCO)
species at 2268 cm−1 [21,41–45] in addition to adsorbed CO
species at reaction temperatures above 400 K (Fig. 7). The
peak intensity of NCO was decreased at 600 K, which clearly
indicates that NCO formation depended on the reaction tem-
perature. Solymosi et al. reported the IR band of NCO species
on metal and metal oxide surface, and they found that the pre-
adsorbed oxygen atoms significantly stabilize the NCO species
on Rh(111) [44,45]. In Fig. 8, the peak intensity for NCO
species and the TOFs for N2 formation over Ir(211) are plot-
ted as a function of reaction temperature. The peak intensity for
NCO species increased with increasing the reaction tempera-
ture and reached a maximum value at 600 K. Above 600 K, the
peak intensity for NCO species greatly decreased with reaction
temperature. In contrast, N2 formation activity appeared above
600 K and greatly increased with increasing reaction tempera-
ture. The NCO species is an intermediate for NO reduction with
hydrocarbons, and this intermediate reacts with NO to form N2
and CO2 on a Cu-ZSM-5 surface [41]. Surface isocyanates are
intermediates for the NO + CO reaction, which react with NO
to form gas-phase N2O over a Rh surface [43]. On the other
hand, Goodman et al. [21] reported that the isocyanate species
plays the role of a spectator rather than an intermediate in CO
+ NO reaction over the Pd crystal because of the stability of
the isocyanate species. However, the isocyanate species formed
Fig. 6. Arrhenius plots of an Ir/SiO powder catalyst (2), Ir(111) ("), Ir(211)
2
(Q), and Ir(100) (F) for N and CO formation from the NO + CO reaction.
2
2
The powder catalyst data were obtained in the flow mode (P
in a He carrier gas), and the single-crystal data were acquired in the batch mode
= P = 5 Torr
NO
CO
(P = P = 5 Torr).
NO
CO
of the surface structure, whereas the pre-exponential factor was
strongly dependent on the surface structure. In our study, the
pre-exponential factor was dependent on the Ir surface struc-
ture. Furthermore, the order of the pre-exponential factors for
the various surface structures was in good agreement with the
order of the planes with respect to NO dissociation activity.
Goodman et al. [20] reported that removal of an inactive atomic
nitrogen species may be important in determining the activity
for N2 and CO2 formation over Pd surfaces. However, we found
that the catalytic activity for NO reduction over Ir surfaces was
dependent on the NO dissociation activity.
The activation energies for N2 and CO2 formation over an
Ir/SiO2 catalyst were estimated to be 117.2 4.1 and 123.1
4.1 kJ/mol, respectively, and these values were similar to those
of the Ir single crystal. This result indicates that the NO + CO

T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
145
Fig. 8. Temperature dependence of NCO peak intensity (2) and TOF for N
2
formation (") over Ir(211) during exposure to a NO/CO (= 1) gas mixture at
10 Torr.
Table 2
a
Apparent activation energies and pre-exponential factors for NO dissociation
−1
2.5 0.3
4.2 0.6
3.5 0.1
Catalyst
E (kJ/mol)
υ (s
)
a
Fig. 7. In situ PM-IRA spectra of Ir(211) during exposure to a NO/CO (= 1)
Ir(111)
Ir(100)
Ir(211)
a
36.6 1.9
36.0 3.8
35.4 0.7
10
10
10
gas mixture at 10 Torr.
on the Ir surfaces is not so much stable under the reaction tem-
perature region from the result of Fig. 8. Furthermore, Jentys et
al. also reported [46] that the surface concentration of the iso-
cyanate species was strongly correlated to the activity of the
Pt/MCM-41 catalyst, whose relationship is in good agreement
with the result of Fig. 8. They concluded that the isocyanate
species was a reaction intermediate during the NO reduction.
We thus believe that the NCO species is a reaction intermedi-
ate in our system and that the NO + CO reaction consists of the
following elementary steps:
The kinetic data were obtained at a constant temperature between 255 and
310 K in a vacuum. The initial NO coverage on the Ir surfaces was 0.25.
Table 3
a
Apparent activation energies and pre-exponential factors for NCO formation
−1
Catalyst
E (kJ/mol)
a
υ (s
)
3.6 0.4
3.4 0.5
3.1 0.2
Ir(111)
Ir(100)
Ir(211)
a
43.8 2.5
42.6 1.6
41.2 2.9
10
10
10
The initial N coverage was 0.25. The Ir surfaces were exposed to 5 Torr of
a
CO at a constant temperature between 325 and 350 K.
NOa → Na + Oa,
(1)
(2)
(3)
(4)
Na + COa → NCOa,
NCOa + NOa → N2g + CO2g,
Oa + COa → CO2g.
We thus considered the adsorbed NO species, both bridge NO
and atop NO, were effectively dissociated on the Ir(100) sur-
face. This order was in good agreement with that for the TOFs
for NO reduction catalytic activity, which suggests that NO dis-
sociation activity determined the total rate for the NO + CO
reaction.
Next, we measured the formation rate of NCO using IRAS.
First, we prepared atomic nitrogen by the dissociation of NO at
325–450 K. The initial Na coverage was 0.25. Then the Ir sur-
faces were exposed to 5 Torr of CO at a constant temperature
between 325 and 350 K. We measured the intensity of NCO
species using IRAS, and we obtained the initial formation rates
of NCO species from the slope of a plot of INCO versus t at
zero coverage. The kinetic data are shown in Table 3. The ac-
tivation energy and pre-exponential factor were determined to
We measured kinetic data for steps (1)–(4). First, we mea-
sured the NO dissociation kinetic data on the Ir surfaces. The
dissociation kinetics of NO were measured at a constant tem-
perature between 255 and 310 K in a vacuum. The initial NO
coverage on the Ir surfaces was 0.25. Atomic nitrogen and oxy-
gen were the only products formed by the dissociation of NO,
and the formation ratio of Oa and Na was almost 1. We mea-
sured the Na coverage using XPS in isothermal decomposition
experiments, and the rate constant was obtained from the slope
of a plot of log ΘNa versus t. Table 2 summarizes the kinetic
data on the Ir surfaces. The activation energy was estimated to
be 36 kJ/mol for all the Ir surfaces, and the pre-exponential
factor and the reaction rates were strongly dependent on the
surface structures, which indicated that the NO dissociation re-
action was structure sensitive. The NO dissociation activity of
the Ir planes decreased in the order (100) > (211) ꢀ (111).
be 41–44 kJ/mol and 103.1–103.6 −1, respectively. Values for
s
all the surfaces were in good agreement with one another. Thus,
we believe that the formation of NCO intermediates was inde-
pendent of the Ir surface structures.

146
T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
Table 4
in good agreement with the order for NO dissociation activ-
ity, which indicates that the NO-reduction catalytic activity
was dependent on the NO dissociation.
Apparent activation energies and pre-exponential factors for NCO consump-
a
tion
−1
8.5 0.4
7.9 0.5
8.0 0.2
Catalyst
E (kJ/mol)
υ (s
)
a
(4) The TOFs for N2 and CO2 formation from the NO +
CO reaction over Ir/SiO2 were almost same for those over
Ir(211), which indicates that the Ir surface structure of
Ir/SiO2 was similar to the surface structure of Ir(211).
Thus, we believe that the active site for the NO + CO reac-
tion over Ir/SiO2 was a metallic Ir surface.
Ir(111)
Ir(100)
Ir(211)
75.4 2.3
69.8 2.8
70.6 1.0
10
10
10
a
−7
The Ir surfaces were exposed to 1 × 10 Torr of NO at a constant temper-
ature between 300 and 325 K.
(5) An NCO species was an intermediate for NO reduction
with CO, and the NCO species reacted with NO to form N2
and CO2. From the kinetic study of the elementary steps,
we concluded that the rate-limiting step for NO reduction
with CO was the NO dissociation step. That is, the NO dis-
sociation activity determined the total reaction rate.
Finally, we measured the rate of NCO consumption. We con-
firmed that the NCO species did not decompose at temperatures
below 333 K. However, we did observe a decrease in NCO
coverage with exposure to NO on the Ir surface at tempera-
tures between 300 and 325 K, which indicates that elementary
step 3 proceeded on the Ir surfaces. We thus measured the in-
tensity of NCO species with NO exposure at a constant tem-
perature and determined the rate constant from the slope of a
plot of logINCO versus t. The NO exposure pressure was 1 ×
10−7 Torr. The apparent activation energy and pre-exponential
Acknowledgment
This work was supported by the Japan Society for the Pro-
motion of Science (JSPS-KAKENHI 17350079).
factor for Ir(100) were determined to be 69.8 2.8 kJ/mol
0.5
and 107.9
s
−1, respectively, and these values are in good
References
agreement with those obtained for Ir(211) (70.6 1.0 kJ/mol
and 108.0 0.2 s−1) (Table 4). The apparent activation energy for
Ir(111) (75.4 2.3 kJ/mol) was slightly higher than the values
for (100) and (211). In contrast, the NCO consumption rate on
Ir(111) was slower than the rates on (100) and (211). Because
the NO dissociation reaction was included in this elementary
step, the reaction rate for NCO consumption was dependent on
the Ir surface structure. We thus have elucidated the detailed ki-
netics and mechanism of the NO + CO reaction on Ir-based
catalysts using surface-science techniques.
[1] V.I. Pârvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233.
[2] H. Hamada, Catal. Today 22 (1994) 21.
[3] R. Burch, J.P. Breen, F.C. Meunier, Appl. Catal. B 39 (2002) 283.
[4] M. Ogura, A. Kawamura, M. Matsukata, E. Kikuchi, Chem. Lett. (2000)
146.
[5] A. Wang, L. Ma, Y. Cong, T. Zhang, D. Liang, Appl. Catal. B 40 (2003)
319.
[6] M. Haneda, Pusparatu, Y. Kintaichi, I. Nakamura, M. Sasaki, T. Fujitani,
H. Hamada, J. Catal. 229 (2005) 197.
[7] M. Shimokawabe, N. Umeda, Chem. Lett. 33 (2004) 534.
[8] A. Takahashi, I. Nakamura, M. Haneda, T. Fujitani, H. Hamada, Catal.
Lett. 112 (2006) 133.
[9] H.S. Oh, C.C. Eickel, J. Catal. 128 (1991) 526.
[10] H.S. Oh, B.G. Fisher, E.J. Carpenter, W.D. Goodman, J. Catal. 100 (1986)
360.
4. Conclusions
(1) NO adsorbed on the atop and hollow sites of Ir(111). On
Ir(100), NO adsorbed on atop and bridge sites. NO ad-
sorbed on the atop site of the (111) terrace and the bridge
site of the (100) step over Ir(211). Only hollow NO de-
sorbed to N2 at 471 K and 574 K on Ir(111). Both atop NO
and bridge NO almost completely desorbed to N2 at 400
and 600 K on Ir(100) and Ir(211), respectively. The NO
dissociation activity of the Ir planes decreased in the order
(100) > (211) ꢀ (111). The NO adsorption and dissoci-
ation activities were strongly dependent on the Ir surface
structure.
[11] K.B. Cho, J. Catal. 148 (1994) 697.
[12] F. Schuth, E. Wicke, J. Phys. Chem. 144 (1985) 239.
[13] Y.K. Ng, N.D. Belton, J.S. Schmieg, B.G. Fisher, J. Catal. 146 (1994) 394.
[14] W.D. Goodman, Surf. Sci. 299/300 (1994) 837.
[15] W.D. Goodman, Chem. Rev. 95 (1995) 523.
[16] M.J.P. Hopstaken, W.J.H. van Gennip, J.W. Niemantsverdriet, Surf.
Sci. 433–435 (1999) 69.
[17] T.W. Root, L.D. Schmidt, Surf. Sci. 150 (1985) 173.
[18] L.M. Carballo, T. Hahn, H.-G. Lintz, Appl. Surf. Sci. 40 (1989) 53.
[19] T. Hahn, H.-G. Lintz, Appl. Surf. Sci. 40 (1989) 59.
[20] D.R. Rainer, S.M. Vesecky, M. Koranne, W.S. Oh, D.W. Goodman,
J. Catal. 167 (1997) 234.
[21] C. Hess, E. Ozensoy, D.W. Goodman, J. Phys. Chem. B 107 (2003) 2759.
[22] C.H.F. Peden, D.N. Belton, S.J. Schmieg, J. Catal. 155 (1995) 204.
[23] T. Fujitani, I. Nakamura, Y. Kobayashi, A. Takahashi, M. Haneda, H. Ha-
mada, J. Phys. Chem. B 109 (2005) 17603.
[24] I. Nakamura, K. Suzuki, A. Takahashi, M. Haneda, H. Hamada, T. Fuji-
tani, J. Vac. Sci. Technol. A 25 (2007) 1143.
[25] Ts. Marinova, L.K. Kostov, Surf. Sci. 185 (1987) 203.
[26] J.E. Muilenberg (Ed.), Handbook of X-Ray Photoelectron Spectroscopy,
Perkin-Elmer, Eden Prairie, MN, 1979.
[27] J.S. Villarrubia, W. Ho, J. Chem. Phys. 87 (1987) 750.
[28] D.Y. Zemlyanov, M.Y. Smirnov, V.V. Gorodetskii, J.H. Block, Surf. Sci.
329 (1995) 61.
(2) CO adsorbed only on the atop site over the Ir surfaces at
223 K, indicating that CO adsorption was independent of
the Ir surface structures. On Ir(211), CO adsorbed initially
on the (100) step and then began to adsorb on the (111)
terrace after the step sites were saturated. The thermal sta-
bility of the adsorbed CO on Ir(100) was higher than that
on Ir(111).
(3) Compared to Ir(100) and Ir(211), Ir(111) showed low activ-
ity for N2 and CO2 production from the NO + CO reaction.
The TOFs for N2 and CO2 formation for the Ir planes de-
creased in the order (100) ꢀ (211) ꢀ (111). This order was
[29] C. Nyberg, P. Uvdal, Surf. Sci. 204 (1988) 517.

T. Fujitani et al. / Journal of Catalysis 253 (2008) 139–147
147
[30] A. Sandell, A. Nilsson, N. Mårtensson, Surf. Sci. 251/252 (1991) 971.
[31] J.F. Zhu, M. Kinne, T. Fuhrmann, R. Denecke, H.-P. Steinrück, Surf.
Sci. 529 (2003) 384.
[32] I. Nakamura, T. Fujitani, H. Hamada, Surf. Sci. 514 (2002) 409.
[33] D. Ramsier, Q. Gao, H.N. Waltenburg, J.T. Yate Jr., J. Chem. Phys. 100
(1994) 6837.
[34] R.J. Behm, K. Christman, G. Ertl, J. Chem. Phys. 73 (1980) 2984.
[35] S. Surnev, M. Sock, M.G. Ramsey, F.P. Netzer, M. Wiklund, M. Borg, J.N.
Andersen, Surf. Sci. 470 (2000) 171.
[36] H. Steinger, S. Lehwald, H. Ibach, Surf. Sci. 123 (1982) 264.
[37] R. Martin, P. Gardner, A.M. Bradshaw, Surf. Sci. 342 (1995) 69.
[38] A. Beutler, E. Lundgren, R. Nyholm, J.N. Andersen, B.J. Setlik, D. Hes-
kett, Surf. Sci. 396 (1998) 117.
[39] F. Strisland, A. Ramstad, T. Ramsvik, A. Borg, Surf. Sci. 415 (1998)
L1020.
[40] H. Orita, in preparation.
[41] H. Oka, T. Okada, K. Hori, J. Mol. Catal. 109 (1996) 51.
[42] H. Celio, K. Mudalige, P. Mills, M. Trenary, Surf. Sci. 394 (1997) L168.
[43] D.I. Kondarides, T. Chafik, X.E. Verykios, J. Catal. 193 (2000) 303.
[44] J. Kiss, F. Solymosi, J. Catal. 179 (1998) 277.
[45] F. Solymosi, T. Bánsági, J. Catal. 202 (2001) 205.
[46] W. Schießer, H. Vinek, A. Jentys, Appl. Catal. B 33 (2001) 263.