Removal pathways of surface nitrogen in a steady-state NO + CO reaction on Pd(110) and Rh(110): Angular and velocity distribution studies

Article abstract of DOI:10.1021/jp0379603

Knowledge of the relation of N₂ and N₂O formation is requisite for improving environmental catalysts. The angular and velocity distributions of desorbing products N₂ and CO₂ were investigated in a steady-state NO + CO reaction on Pd(110) and Rh(110) by cross-correlation time-of-flight methods. On Pd(110), N₂ desorption was split into two inclined components collimating at ± 40° in the plane along the [001] direction. The inclined N₂ formation originated from the N₂O intermediate. At low temperatures, the pathway through the N₂O intermediate prevailed, and, above 720 K, the associative nitrogen desorption started to dominate. N₂ desorption on Rh(110) was sharply collimated along the surface normal in a wide temperature region, indicating that N(a) was mostly removed through the associative process. On both surfaces, the translational temperature of desorbing N₂ was very high, reaching about 2500-3500 K. On the other hand, CO₂ desorption always collimated along the surface normal on both surfaces with the translational temperature at 1600-2000 K.

Full text of DOI:10.1021/jp0379603

14232
J. Phys. Chem. B 2004, 108, 14232-14243
Removal Pathways of Surface Nitrogen in a Steady-State NO + CO Reaction on Pd(110)
†
and Rh(110): Angular and Velocity Distribution Studies
‡
§
|
,§
Izabela I. Rze z´ nicka, Yunsheng Ma, Gengyu Cao, and Tatsuo Matsushima*
Graduate School of EnVironmental Earth Science, Hokkaido UniVersity, Sapporo 060-0810 Japan,
Catalysis Research Center, Hokkaido UniVersity, Sapporo 001-0021 Japan, and
Department of Chemical Physics, UniVersity of Science and Technology of China, Hefei 230026, China
ReceiVed: December 21, 2003; In Final Form: March 24, 2004
The angular and velocity distributions of desorbing products N
CO reaction on Pd(110) and Rh(110) by cross-correlation time-of-flight techniques. The CO
sharply collimated along the surface normal on both surfaces. On the other hand, N desorption on Pd(110)
sharply collimated along about 40° off the surface normal in the plane along the [001] direction below around
50 K, yielding a translational temperature of about 3600 K. At higher temperatures, the normally directed
desorption was relatively enhanced. On Rh(110), desorbing N sharply collimated along the surface normal,
yielding a translational temperature of about 2500 K. The inclined desorption was assigned to the decomposition
of the intermediate, N O(a) f N
2
and CO
2
were studied in a steady-state NO
+
2
desorption
2
6
2
2
2
(g) + O(a), and the normally directed component was proposed to be due
2
to the associative desorption of adsorbed nitrogen atoms, 2N(a) f N (g). The branching of these pathways
was analyzed on Pd(110).
I. Introduction
the reaction rate and are always related to the product desorption
5,10
step whenever any step becomes rate-determining. Thus, the
inclined N2 desorption is useful for examining the reaction
pathway of NO decomposition. In fact, from the inclined
desorption and similar velocity distributions of desorbing N2
Nitrous oxide, N2O, is formed when the emissions from diesel
or petrol engines (containing NO) are passed through catalytic
converters with active metals such as rhodium and palladium.
1
N2O is harmful and yields a remarkable greenhouse effect.
Knowledge of the relation of N2 and N2O formation is requisite
for improving such environmental catalysts. This paper reports
the first separation of three removal pathways of surface
nitrogen, i.e., the associative process of 2N(a) f N2(g), the
intermediate decomposition of N(a) + NO(a) f N2O(a) f
N2(g) + O(a), and its desorption without decomposition.
In the pathway through the N2O intermediate on Rh(110),
desorbing N2 is expected to collimate at 30-70° off the surface
normal because such large collimation angles were observed
in the N2O decomposition.2 However, neither inclined N2
desorption nor N2O production became noticeable up to about
1
1-14
in both NO and N2O decompositions on Pd(110),
Ohno et
al. concluded that the N2 emission in NO decomposition
proceeds through the intermediate N2O(a). However, the
surveyed steady-state conditions were limited to around 2 ×
-
7
1
0
Torr because of the use of an apparatus designed for angle-
12,15
resolved temperature-programmed desorption (AR-TPD).
the present work, angle-resolved steady-state desorption
AR-SSD) measurements were successfully performed for
steady-state NO decomposition up to 1 × 10 Torr using a
gas doser with a fine orifice and cross-correlation time-of-flight
technique.
In
(
-
4
-6
-
4
The angular and velocity distributions of desorbing products
in the course of catalyzed reactions have been mostly analyzed
1
× 10 Torr of NO. N2 desorption always collimated along
the surface normal, suggesting that the associative process is
predominant. On the other hand, on Pd(110), the inclined N2
16
7
with modulated molecular beams (MMB). For NO decompo-
sition on well-defined surfaces in an ultrahigh vacuum, the
phase-sensitive detection technique with beam modulations was
requisite so that the N2 signal could be monitored in the high
background due to NO decomposition on the reaction chamber
desorption and a significant amount of N2O formation were
observed in a wide pressure range, indicating that the pathway
through the N2O intermediate was operative except for high
temperatures.
The overall reaction rate at the steady-state NO + CO reaction
is mostly controlled by the dissociation of adsorbed NO. No
information of the removal of surface nitrogen can be obtained
from kinetic measurements at the steady-state conditions. The
angular and velocity distributions of desorbing products can
provide information on the reaction sites and also product
1
7
wall. Steady-state conditions, however, cannot be established
for the reaction because of the periodical change of reactants
on the surface. Similar ambiguities were also induced in AR-
TPD work in the presence of reactant gases, in which simple
TPD phenomena were not differentiated from kinetic behavior
8
1
8-20
at the steady-state conditions.
Thus, neither angular nor
9
velocity distributions have been reported under steady-state NO
desorption processes. These distributions do not directly involve
1
2,15
+
CO reactions except for our earlier work.
Isotope-tracer experiments have frequently been reported for
†
Part of the special issue “Gerhard Ertl Festschrift”.
*
To whom correspondence should be addressed. Fax: +81-11-706-
mechanism studies of NO reduction. Belton et al. could not
confirm in their isotope-TPD work that N2O produced on
9
120. E-mail: tatmatsu@cat.hokudai.ac.jp.
‡
Graduate School of Environmental Earth Science, Hokkaido University.
Catalysis Research Center, Hokkaido University.
University of Science and Technology of China.
2
1
§
Rh(111) yielded N2. Recently, Zaera and Gopinath reported
|
14
15
that the replacement of surface N(a) by N(a) upon shifting
1
0.1021/jp0379603 CCC: $27.50 © 2004 American Chemical Society
Published on Web 06/08/2004

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14233
Figure 1. Apparatus for angle-resolved product desorption combined with cross-correlation time-of-flight techniques. The insert shows the structure
around a sample crystal and a gas doser. REV-LEED (reverse-view low-energy electron diffraction optics); QMS (quadrupole mass spectrometer);
XPS (X-ray photoelectron spectroscopy optics); MCS (multichannel scalar); trigger (photocell and light-emitting diode); PC (personal computer).
3
-1
The chopper house is pumped at about 7 m s
.
from 14NO(g) to NO(g) in a steady-state CO + NO reaction
15
15
14 15
22
yielded N2 and N N in large amounts. This result indicates
that such isotope methods are not effective for differentiating
the above N(a) removal processes.
II. Experiment
Experiments were performed in an ultrahigh-vacuum system
composed of a reaction chamber, a chopper house, and an
5
,23
analyzer, which were separately pumped (Figure 1).
reaction chamber is a conventional ultrahigh-vacuum chamber
The
3
-1
with a high pumping rate (about 1.5 m s ). It has reverse-
view low-energy electron diffraction (LEED) and X-ray pho-
toelectron spectroscopy (XPS) optics, a quadrupole mass
+
spectrometer (QMS) for the angle-integrated (AI) form, an Ar
gun, a sample manipulator, and a gas-handling system. The
chopper house has a narrow slit facing the reaction chamber
2
4
and contains a cross-correlation random chopper blade. The
analyzer is equipped with another QMS operating in a pulse-
counting mode for AR and time-of-flight (TOF) analyses. The
arrival times at the ionizer of the QMS are registered in a
multichannel scalar synchronized with the chopper rotation. The
distance from the ionizer to the chopper blade is 377 mm. A
time resolution of 20 µs is obtained with a chopper frequency
of 98.023 Hz. A very large pumping rate is established in the
chopper house by cooling a copper plate below 40 K, yielding
3
-1
about 7.0 m s of the pumping rate. Therefore, only molecules
passing the first slit, the chopper slits, and the second drift tube
without scattering can enter the ionizer of the analyzer. The
analyzer is pumped by another turbo-molecular pump to keep
-
10
the base pressure below 1 × 10 Torr. In this apparatus, AR-
Figure 2. Steady-state reaction rate on Pd(110) versus surface
-
3
SSD measurements are possible up to 1 × 10 Torr.
13
15
temperature: (a) AI- and (b) AR-desorption signals of CO
2
,
2
N O
A gas mixture of 15NO and CO or NO and 13CO was
12
14
15
13
15
15
and
N
2
. The AR signals of CO
2 2 2
, N , and N O at θ ) 0° and also
1
5
12
15
prepared in a glass bottle. The NO/ CO mixture was used in
the experiment where the nitrogen signal was observed and the
of N at θ ) 41° off the surface normal toward the [001] direction
2
-
5
15
12
are shown. The total pressure is 7.5 × 10 Torr ( NO: CO ) 1:1 or
14
13
1
4
13
NO: CO ) 1:1). The lines in the figure are “guides to the eye”.
NO/ CO mixture was used when the CO2 signal was
monitored. One of the mixtures was introduced through a doser
with a small orifice (diameter, 0.1 mm) about 2 cm from the
sample crystal (the insert in Figure 1). Thus, the flux intensity
of incident reactants decreases when the desorption angle shifts
from the normal direction, where the desorption angle (polar
angle, θ) is defined as the angle from the surface normal along
the [001] direction. This effect was corrected as described in
eters was separately estimated by introducing N2O. It was only
about 25% of the N2O signal.
III. Results
(A) Pd(110). (A.1) General Features. The steady-state
formation rates of N2, CO2 and N2O were monitored in both
AI and AR forms as a function of the surface temperature (TS),
as shown in Figure 2. The effused flow of an equimolar mixture
of NO and CO from the small orifice was kept constant, yielding
15
13
section III(A.2). The product signals due to N2, CO2, and
1
5
N2O were monitored in both AI and AR forms. Hereafter,
these species are simply described as N2, CO2, and N2O in the
text. The fragmentation of N2O into N2 in both mass spectrom-
-
5
the pressure in front of the crystal at around 7.5 × 10 Torr.

14234 J. Phys. Chem. B, Vol. 108, No. 38, 2004
Rze z´ nicka et al.
The AI signal (Figure 2a) was determined by the QMS in the
reaction chamber as the difference in the signal between the
desired surface temperature and room temperature. The mass
sensitivity was corrected to each species. All the rates were
negligible below 500 K, increased rapidly around 550 K to a
maximum with increasing TS, and decreased again at TS values
above 600 K. The AI N2 signal was rather noisy because of the
high background signal, which reached about 80% of the
apparent total signal. Around the maximum rate, the ratio of
CO2:N2:N2O was 5:2.1:1. The N2O formation was significant
below 670 K.
The AR signal was obtained by the QMS in the analyzer as
the difference between the signal at the desired angle and the
signal when the crystal was away from the line-of-sight position.
In Figure 2b, the AR signals at θ ) 0° were shown for desorbing
CO2 and N2O, and the AR signal at θ ) 0 and 41° is displayed
for N2 because these were their collimation angles (the
maximum flux position), as described in section III(A.3).
Similarly to the AI signals, the AR signals were negligible below
5
00 K; thereupon, those abruptly reached their maximum at
about 600 K and slowly decreased toward zero at 800 K. No
hysteresis was found in their rates with decreasing TS values.
The background N2 signal level when the crystal was away was
about 10% of the total signal at θ ) 41° around 600 K.
Noticeable differences were found in the signal ratio between
the AI and AR forms. These are explained by considering the
differences in the sharpness of the angular distributions between
25
each species below 600 K. Above it, the ratio is also affected
by changes in the branching of the surface nitrogen removal
pathways with increasing TS. These will be explained after the
analysis of angular and velocity distributions.
Figure 3. (a) Apparent angular distributions of desorbing 15
2
N O and
(
b) its velocity distributions at different desorption angles in the steady-
15
12
Only the (1 × 1) pattern was observed in LEED measure-
state NO + CO reaction on Pd(110). The total pressure is 7.5 ×
-5 15 12
1
0
Torr ( NO: CO ) 1:1). The signal was not corrected by
ments under a steady-state CO + NO reaction at the total
-7
considering the changes in the reactant density and the surface area
falling into the acceptance angle of the analyzer (see the text).
pressure of 1 × 10 Torr of the equimolar mixture of NO and
CO and in the range of TS ) 400-800 K. This is very different
from the results in the CO + O2 reaction on Pd(110), where
LEED patterns changed from c(2×4)-O to (1×2)-CO and (1×1)
with increasing CO pressure.26
of cos(θ). The apparent signal should be invariant when the
distribution is in a cosine form, and the incident flux is constant
because the surface area falling into the acceptance angle
increases as the reciprocal of the cosine of the desorption angle.
The velocity of this product showed the Maxwellian distribution
at the surface temperature at any desorption angle (Figure 3b).
The cosine distribution is characteristic of the desorption after
(A.2) Angle-ResolVed Measurements. A gas mixture of NO
and CO was introduced through a gas doser with a small orifice
at a constant flow rate (the inset of Figure 1). The use of this
orifice was effective in reducing N2 formation on the reaction
chamber wall. Its formation with a NO dosage in the back-fill
type is usually high and obscures the observation of N2
formation on a small sample surface. The AR signal was slightly
improved under this construction. For the steady-state CO +
O2 reaction, however, this construction was not necessarily
requisite.2 Under this construction, the flux of incident reactants
toward the unit surface area decreases proportionally to the
cosine of the desorption angle when the angle is shifted from
the normal direction even if the incident reactant density in the
effused beam is homogeneous around the crystal sample.
24
thermalization of molecules to the surface temperature. Result-
ant desorbing molecules show a Maxwellian velocity distribution
at the surface temperature all over the desorption angle. Here,
we concluded that the density of the incident reactants around
the sample crystal in the effused beam was homogeneous and
then its incident flux to the unit area decreased in a form of
cos(θ) when the angle was shifted from the normal direction.
This effect must be considered in the angle-dependence analysis.
However, the shift due to this correction was not significant.
3
(A.3) Angular Distributions. The AR signals of desorbing CO2
We examined this point in the following way. Here, two
factors must be examined toward the surface composition: one
is the decrement of the incident flux, and the other is the incident
angle dependence of the adsorption rate. The latter factor is
negligible because no incident angle dependence is expected
for both CO and NO, which have the sticking probability close
and N2 were sensitive to the desorption angle θ. CO2 desorption
sharply collimated along the surface normal, whereas the latter
peaked around 40° off the surface normal toward the [001]
direction. Assuming the power series of the cosine of the
5
desorption angle to each desorption component, the CO2 signal
13
was approximated in a form of cos (θ) at TS ) 550 K and 7.5
27
to unity on Pd(110) and Rh(110).
-5
×
10 Torr with the (1:1) mixture (Figure 4a). The sharp
8
To examine the former factor, the angular distribution of N2O
was measured at a fixed flow of the mixed gas. The results are
shown in Figure 3a. The apparent AR signal of N2O, without
correction due to the increase of the surface area falling into
the acceptance angle of the analyzer, closely followed a form
distribution became slightly broader as cos (θ) at 650 K. On
the other hand, the angular distribution of desorbing N2 was in
2
8
a two-directional form approximated as cos (θ + 41) +
2
8
cos (θ - 41) (Figure 5a). The signal at the normal direction
was negligibly small at 550 K. A sharp two-directional form

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14235
Figure 4. Angular distributions (shown in polar coordinates) of
desorbing CO in the plane along the [001] direction at different T
2 S
Figure 5. Angular distributions of desorbing 15
2
N in the plane along
1
3
S
the [001] direction at different T values. The signal on the ordinate
values. The signal on the ordinate was normalized to that in the normal
was normalized to the maximum at the collimation angle of θ ) 40°
direction at 600 K. The reaction of 14NO + CO on Pd(110) was at
13
15
12
and at 600 K. The reaction of NO + CO on Pd(110) was at the
-
5
14
13
the steady state with the total pressure of 7.5 × 10 Torr ( NO: CO
-5
15
12
steady state with the total pressure of 7.5 × 10 Torr ( NO: CO )
)
1:1). The apparent distribution sharpness is indicated in 〈 〉. Typical
deconvolutions are drawn by broken curves. The solid lines show the
summation of components.
1
:1). Typical deconvolutions are drawn by broken curves. The solid
lines show the summation of components.
translational temperature was around 1600 K, about 3 times
higher than the surface temperature at the normal direction and
decreased quickly with an increasing desorption angle, being
collimated around (40° is consistent with the previous results
-
7
12,15
at 2 × 10 Torr.
The distributions at TS ) 600 and 650 K
28,29
are shown in Figure 5b,c. With increasing TS, the inclined
desorption signal decreased at TS values higher than 600 K, but
it was still significant at 700 K. The collimation angle of the
inclined desorption slightly decreased at higher TS. The signal
intensity at the normal direction increased with increasing TS,
indicating the presence of the normally directed component. The
consistent with the normally directed distribution.
The values
slightly increased with TS. On the other hand, desorbing N2O
showed a Maxwellian distribution at the surface temperature.
(B) Rh(110). (B.1) Temperature Dependence. The maximum
N2 or CO2 formation rate on Rh(110) was about 1 order less
than that on Pd(110). However, the Rh (110) surface showed
noticeable activity already around 450 K, where no activity was
found on Pd(110). Thus, it became difficult to determine the
AI N2 signal on Rh(110) because of the relatively enhanced
background. On this surface, a pronounced hysteresis was
observed in the reaction rate with surface temperature changes.
Figure 8 shows the AR signals of products N2 and CO2 at their
collimation angle θ ) 0° under steady-state conditions at the
28
total signal at 650 K was approximated by 0.85{cos (θ + 40)
2
8
5
+
cos (θ - 40)} + 0.16 cos (θ) + 0.16 cos(θ) as referred to
the velocity analysis in the next section.
A.4) Velocity Distributions. Both desorbing CO2 and N2
(
showed hyperthermal energy. The translational temperature of
N2 was very high, more than twice that of CO2. Typical velocity
curves of N2 are shown in Figure 6. The apparent translational
temperature calculated from the average kinetic energy (〈E〉)
as T〈E〉 ) 〈E〉/2k, is shown in 〈 〉 in the figure, where k is the
Boltzmann constant. The value was maximized around the
collimation angle to the value of 3570 K at TS ) 550 K (Figure
-5
total pressure of 7.5 × 10 Torr. No detectable N2O formation
was found under these conditions. For both CO2 and N2
products, the formation rate was negligible below 400 K and
then suddenly increased around 450 K, reaching the first
maximum at 580-590 K. After this maximum, the rate once
decreased and started to increase again around 700 K, reaching
the second maximum at 740 K. Above it, the rate slowly
decreased. When the surface temperature decreased, the rate
was reproduced at the second maximum around 740 K. It started
to decrease around 600 K and did not reach the first maximum
around 580 K. The reaction was highly retarded below 600 K
when the surface was cooled. A similar hysteresis was also
6
a) and a similar value at TS ) 650 K (Figure 6b). It decreased
quickly with an increasing shift from the collimation angle. The
value at the normal direction decreased to 1100 K at TS ) 550
K. It was highly underestimated because of the relatively high
background signal. The angle dependence is consistent with the
inclined desorption.
The velocity distributions of CO2 at TS ) 550 and 650 K are
shown at different desorption angles in Figure 7. The average

14236 J. Phys. Chem. B, Vol. 108, No. 38, 2004
Rze z´ nicka et al.
Figure 6. Velocity distributions of desorbing 15N
at various desorption angles in a steady-state 15NO + 12CO reaction on Pd(110) at the total
2
-5
15
12
pressure of 7.5 × 10 Torr ( NO: CO ) 1:1) at (a) T
S
) 550 K and (b) T ) 650 K. The angle θ was scanned in the plane along the [001]
S
direction. The average kinetic energy is given in 〈 〉 in temperature units. Typical deconvolutions into two fast components and a slow component
are given by broken curves. The translational temperature for each component is also given. Open circles represent observed signals, and closed
ones indicate the signals after subtraction of the thermalized component.
2
7
found in the CO2 formation rate (Figure 8b). A sharp increase
in the CO2 signal was observed around 470 K at the jump
position of the N2 signal before the first N2 maximum at 590
K. This CO2 spike was reproducible, and the intensity was kept
constant; however, it was difficult to see a similar spike in the
N2 signal because of the high background. The spike in the N2
signal was more pronounced when more NO was in the mixture
gas. The CO2 formation was also reduced below 600 K when
the surface was cooled.
(1:1) mixture, from cos (θ) at 470 K to cos (θ) at 750 K (Figure
10a). On the other hand, the sharpness of the CO2 distribution
was insensitive to the TS value approximated as cos (θ).
3
The velocity distributions of desorbing N2 showed the average
translational temperature around 2600-2700 K, which was
surprisingly insensitive to the desorption angle, as shown in
Figure 11a. It increased with increasing surface temperature with
a slope of about 2 (Figure 12a). The average CO2 translational
temperature reached around 1310 K at the surface normal
direction at TS ) 470 and 1920 K at 750 K (Figure 12b). The
value increased with increasing surface temperature with a slope
similar to that of N2. It decreased quickly with increasing
desorption angle, consistent with the normally directed desorp-
tion (Figure 13).
In LEED measurements at the total pressure of 1 × 10-
7
Torr with the (1:1) mixture gas, a (2 × 1) pattern appeared at
3
60 K, where no reaction proceeded (Figure 9). It was due to
3
0
NO adsorption, mostly to reconstruction by N adsorption. It
was transformed into a (3 × 1) pattern due to reconstruction
by N(a) at 380-400 K. At 400 K, a mixed pattern of (2 × 1)
and (1 × 3) appeared. Above 600 K, it transformed into a (1 ×
IV. Discussion
2
) pattern. In the direction of surface cooling, the (1 × 2) pattern
was stable up to 620 K, and then a mixed pattern of (1 × 3)
and c(4 × 2) was observed. These are similar to those observed
by oxygen adsorption.31 The retardation seems to be due to
surface oxide formation.
(A) Desorption Components. (A.1) Pd(110). The velocity
distributions of CO2 and N2 products showed different features.
The distribution of desorbing N2 is too broad as a single
desorption component. To examine this point, the thermalized
component to the surface temperature was first subtracted from
the observed distribution. The presence of this thermalized one
is clear from the data at θ ) 0° as shown in Figure 6a,b. The
resultant distributions yielded the translational temperature of
(B.2) Angular and Velocity Distributions. Both N2 and CO2
products desorbed sharply along the surface normal over the
conditions studied. The N2 desorption became sharper with
-
5
increasing TS at the total pressure of 7.5 × 10 Torr with the

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14237
The velocity distribution after the subtraction of the thermal-
ized component was still broad, yielding the speed ratio (SR),
2
2
1/2
1/2
defined as (〈V 〉/〈V〉 -1) /(32/9π -1) , of 1.09 at 550 K and
.34 at 650 K, even at the collimation position. Here, V is the
1
2
velocity of the molecule, 〈V〉 is the mean velocity, and 〈V 〉 is
the mean square velocity. This SR value indicates the width of
the distribution curve and yields unity when the distribution is
2
4
in a Maxwellian form. The SR yields in most cases values
smaller than unity for the desorption component with a
hyperthermal energy at the collimation angle; i.e., lower-velocity
molecules are less populated than those at the Maxwellian
distribution at the translational temperature. However, the
velocity distribution of the fast components yields the SR value
larger than unity, although the translational temperature is very
high. This strongly suggests that the velocity distribution after
the subtraction of the thermalized component still involves two
or more desorption components.
(A.2) DeconVolution of Velocity Distribution. Generally, it
is difficult to uniquely deconvolute one velocity distribution
curve into two or more distribution curves because two
parameters are required to construct each distribution curve on
the basis of the modified Maxwellian form. This distribution
is in the form of
Figure 7. Velocity distributions of desorbing 13CO
at various
2
1
4
13
desorption angles in a steady-state NO + CO reaction on Pd(110)
-
5
14
13
at the total pressure of 7.5 × 10 Torr ( NO: CO ) 1:1) and at (a)
2
4
T
S
) 550 K and (b) T ) 650 K. The angle θ was scanned in the plane
S
along the [001] direction. The average kinetic energy is given in 〈 〉 in
temperature units. Typical deconvolutions into a fast component and a
slow component are given by broken curves. The translational
temperature of the fast component is also given.
3
2
2
f(V) ) V exp{-(V - V ) /R }
0
where f(V) is the modified Maxwellian distribution function, Vo
is the stream velocity, and R is the width parameter. The stream
velocity determines the peak position of the distribution. This
distribution function is widely used to describe observed velocity
distributions. Of course, one parameter is enough to describe a
Maxwellian form (when Vo ) 0), but no suitable deconvolution
is successful when only two Maxwellian forms with different
temperatures are used. One of the best ways is to deconvolute
into two modified Maxwellians form; in this case, only three
parameters are separately determined by fitting the data points
at high- and low-velocity sides. There are several ways to omit
the remaining parameter. Here, we simply assume a common
width parameter. In this case, the energy of the faster component
would be somewhat underestimated, and that of the other should
be somewhat overestimated because the faster component is
likely to yield a smaller R value. The resultant deconvolutions
are shown by broken curves in Figure 6. The fastest component
obtained in this way yielded a translational temperature of 6250
K (1.02 eV) maximizing at θ ) 40-50° (Figure 13) and the
SR of 0.66. The second component showed around 2500 K (0.43
eV) and the SR ) 0.9. There are significant differences in the
velocity curves of desorbing N2 between the present work and
1
2
the previous one. In the latter, no signals were noticed above
-
1
2
.5 km s . On the other hand, in the present work, the signal
-
1
was still significant in the range of 2.5-4 km s , yielding a
new faster component. The observation of these high-velocity
signals might be obscured in the high background in the previous
work because of the small signal intensity at low reactant
pressures.
The translational temperature of fast CO2 was simply
estimated by subtracting the thermalized component. The energy
of the resultant fast component was maximized to 1840 K at
the normal direction and decreased quickly with increasing
desorption angle (Figure 13). This is only one-third of that of
the fastest component of N2.
1
5
Figure 8. Steady-state AR signals at the surface normal of (a)
N
2
1
3
-5
and (b) CO
2
as a function of T
S
13
at the total pressure of 7.5 × 10
1
5
12
14
Torr ( NO: CO ) 1:1 or NO: CO ) 1:1) on Rh(110). The signals
observed in the direction of the increasing surface temperature are
designated by closed symbols, and those in the downward direction,
by open symbols. The lines in the figure are “guides to the eye”.
(A.3) Rh(110). The velocity distributions of N2 and CO2
the fast components. The translational temperature of this
products on Rh(110) were close to each other. Both translational
temperatures were maximized at the normal direction. The fast
component of CO2 after the subtraction of the thermalized
remaining component reached about 3300-3500 K at θ ) 40-
5
0° (Figure 13).

14238 J. Phys. Chem. B, Vol. 108, No. 38, 2004
Rze z´ nicka et al.
Figure 9. LEED patterns from Rh(110) in a steady-state NO + CO reaction. Diffraction patterns were recorded at the accelerating voltage of 85
-
7
eV at the total pressure of 1 × 10 Torr (NO:CO )1:1). T
S
/K ) (a) 360; (2×1)-N, (b) 380; (3×1)-N, (c) 580; (2×1)-N/(1×3)-O, (d) 740;
(
1×2)-O, (e) 840; (1×2)-O. In the decreasing temperature, (f) 600, c(4×2)+(1×2); and (g) 400, c(4×2)+(1×3).
2 2
N in the plane along the [001] direction
Figure 10. Angular distributions (shown in polar coordinates) on Rh(110) of desorbing (a) 15 and (b) 13CO
at different T
S
values. The signal on the ordinate was normalized to that in the normal direction at 550 K. The reaction of NO + CO was at the
-
5 15 12 14 13
steady state with the total pressure of 7.5 × 10 Torr ( NO: CO ) 1:1 or NO: CO ) 1:1). The apparent distribution is indicated in 〈 〉. Typical
deconvolutions are drawn by broken curves. The solid lines show the summation of components.
component reached 1940 K at the normal direction and
decreased quickly with increasing desorption angle as shown
by the broken curve in Figure 11b. This is consistent with the
normally directed desorption.

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14239
Figure 13. Translational temperature of 15N
opened marks) on Pd(110) versus the desorption angle in the plane
along the [001] direction. The reaction was in the steady state at the
(full marks) and 13CO
2
2
(
-
5
15
12
14
13
total pressure of 7.5 × 10 Torr ( NO: CO ) 1:1 or NO: CO )
1
:1). The average value was calculated over all molecules before
deconvolution of the velocity curve. The value for the fast component
was estimated after deconvolution. The vertical bars indicate typical
experimental errors. The lines in the figure are “guides to the eye”.
Figure 11. Velocity distributions at various desorption angles on Rh-
1
5
13
(
110) of (a)
the total pressure of 7.5 × 10 Torr ( NO: CO ) 1:1 or NO: CO
1:1) and at T ) 550 K. The angle θ was scanned in the plane along
N and (b) CO in a steady-state NO + CO reaction at
2
2
-
5 15 12 14 13
(
B) Reaction Pathways. The following steps were proposed
)
S
8
to proceed NO decomposition on Pd and Rh surfaces.
the [001] direction. The average kinetic energy is given in 〈 〉 in
temperature units. Typical deconvolutions are given by broken curves.
The translational temperature for each component is also given.
NO(g) f NO(a)
(i)
(ii)
NO(a) f N(a) + O(a)
2
N(a) f N (g)
(iii)
2
N(a) + NO(a) f N O(a)
(iv)
(v)
2
N O(a) f N O(g)
2
2
N O(a) f N (g) + O(a)
(vi)
2
2
To remove the surface oxygen, some reducing reagent, CO
in the present case, is necessary for the catalytic cycle to
proceed.
CO(g) f CO(a)
(vii)
CO(a) + O(a) f CO (g)
(viii)
2
The oxygen removal step (viii) has been well-analyzed under
steady-state conditions and also transient conditions. At tem-
peratures above 400 K for steady-state reactions, O(a) on Pd-
Figure 12. Velocity distributions at various temperatures on Rh(110)
1
5
13
15
13
of (a)
total pressure of 7.5 × 10 Torr ( NO: CO ) 1:1 or NO: CO )
:1). The angle was fixed at θ ) 0°. The average kinetic energy is
N and (b) CO in a steady-state NO + CO reaction at the
2
2
(110) is quickly removed by CO. Desorbing CO2 in CO
-
5 15 12 14 13
oxidation on Pd(110) collimated sharply along the surface
1
normal when O(a) . CO(a), but it showed a cosine distribution
given in 〈 〉 in temperature units. Typical deconvolutions are given by
broken curves. The translational temperature for each component is
also given.
2
8,29
when O(a) , CO(a).
Similar phenomena were observed on
Rh(110) except for bidirectional CO2 desorption at the limited
3
2
conditions. Desorbing CO2 in the NO + CO reaction showed
angular and velocity distributions very similar to those in CO
oxidation in the active region on both surfaces.
On the other hand, the fast component of desorbing N2 after
subtraction of the thermalized component reached 2520 K at
the normal direction (Figure 11a). Surprisingly, this value did
not decrease with increasing desorption angle. It increased from
Steps (i), (ii), and (iii) have been well-characterized on Pd-
(110) and Rh(110). Step (ii) is in most cases rate-determining
8
and subject to surface oxygen. It proceeds above 420 K on
3
3
2
520 K at the normal direction to 2700 K at θ ) 30°, although
Rh(110) and 450 K on Pd(110) at the steady-state reaction.
Desorbing N2 from step (iii) sharply collimated along the surface
the flux decreased in this way. This suggests the presence of
inclined desorption showing high velocity off the normal
direction. In fact, the angular distribution became sharper at
7
normal. Both Pd(110) and Rh(110) surfaces are converted into
a missing-row form stabilized by oxygen when the surface is
3
1,34
750 K, suggesting the disappearance of the inclined component.
exposed to oxygen above 500 K.

1
4240 J. Phys. Chem. B, Vol. 108, No. 38, 2004
Rze z´ nicka et al.
(
C) Intermediate N2O. The knowledge of steps (iv), (v), and
K. However, the collimation of desorbing N2 on the resultant
surface shifted only about 30° off the normal in the plane along
the [001] direction. The desorption mechanism of this 30°
component is not clear at present. The adsorption structure of
declining N2O, which interacts with the oxygen-modified site
through the oxygen end, might be induced.
(
vi) in relation to N2O(a) is limited on noble metals. Step (iv),
N(a) + NO(a) f N2O(a), took place around 100 K on Pt(335),
around 300-350 K on Rh(111), and above 400 K on Pd(110),
where N(a) was formed from electron bombardments to NO(a)
or by exciting N2(g).7
,21,35
N2O decomposition, step (v), is sensi-
13
tive to the kind of metal and its surface structure. On Pd((110),
Rh(110), Ru(001), W(110), Cu(110), and stepped Ni(557),
the dissociation proceeds at or below 100 K. Incident N2O was
In fact, such inclined N2O was predicted on positively charged
2,3
36
37
38
53
Ag clusters. This kind of positive metal atoms may be induced
5
4
in rhodium oxides. Such an adsorption structure of N O has
2
39,40
decomposed on Ni(110), Ni(100), and Rh(110) below 200 K
never been proposed on clean noble metals.
41
but not on Rh(111). N2O(a) was desorbed at 90-120 K on
Here, we may expect the contribution of this inclined
desorption of N2 in the steady-state NO + CO reaction because
of the presence of surface oxygen.
Pt(111),4 Ir(111), Ni(111), and Ag(111) without dissocia-
tion. On Pt(335) ) [(S)4(111)×(001)], step (iii), 2N(a) f N2(g),
takes place around 400 K, showing high reactivity of surface
nitrogen on platinum.35 On Rh(110), Rh(100), and Rh(111), N2O
formation was observed in a steady-state NO + CO reaction at
2
43
44
45
However, we could not find N2O formation on Rh(110) in
-
7
-4
the pressure range of 10 -10 Torr of NO and TS ) 300-
00 K. The possibility of N2O formation on Rh(110) cannot be
8
-
2
46
NO pressures above 10 Torr.
completely ruled out because of the rapid decomposition on
this surface and the observed sharpening of the angular
distribution of desorbing N2 at higher temperatures, which
suggests the presence of inclined desorption. Furthermore, this
possibility is also supported by the slow decay of the fast
component with increasing desorption angle, as seen in Figure
N2O(a) was proposed to adsorb via its terminal N atom on
metal surfaces in a standing form or in an inclined one.
Prior to decomposition, N2O must lie to release oxygen on the
surface. Ohno et al. proposed that N2O is oriented along the
001] direction before decomposition and then the nascent N2
receives repulsive forces along the ruptured N2-O axis. This
aligned N2O in the [001] direction was first proposed on Pd-
110)(1×1) as the precursor of inclined N2 emission. The
thermal decomposition of adsorbed N2O is completed below
60 K on both Rh(110) and Pd(110); however, the presence of
short-lived N2O is not necessarily ruled out in the steady-state
NO + CO reaction, even above 500 K, because adsorbed N2O
can be stabilized by surface oxygen.3 A recent density
functional theory (DFT) calculation with a generalized gradient
approximation by Kokalj shows that the parallel form of N2O
on Pd(110) is in a bending configuration, bridging atomic
troughs extending in the [1 1h 0] direction.4 It is as stable as a
standing or tilting form with bonding via the terminal nitrogen
atom. Recent near-edge X-ray absorption fine-structure spec-
troscopy (NEXAFS) measurements of N2O on Pd(110) around
0 K showed remarkable anisotropy in the X-ray polarization
dependence of two π resonance (N 1s f 3π*) NEXAFS
peaks. This suggests that the major N2O is lying along the
001] direction and a minor species is standing, which is
consistent with the DFT results.
4
2-45,47
[
1
1
1
1a. Thus, the angular distribution at TS ) 470 K was
7
deconvoluted into three components, a cos (θ) form, a cos(θ)
form, and the inclined cos (θ ( 30) form, as shown by the
(
12
broken curves. Here, the deconvolution was performed by
assuming the maximum contribution of the inclined component
on Rh(110).
1
N(a) on Rh(110) is not highly reactive with NO(a); rather, it
is reactive with N(a), forming N2(g). This suggests that the
reactivity of N(a) is enhanced at high pressures because of the
significant N2O formation at high pressures. This enhancement
is not necessarily induced by increased NO(a) or N(a) popula-
tion. It may be enhanced by subsurface oxygen (or in oxide
forms) or nitrogen.
,48
9,50
(E) Branching of Surface-Nitrogen Removal. On Pd(110),
the three removal pathways of surface nitrogen are operative:
6
(
(
I) through the associative process of N(a), i.e., 2NO(a) f 2N-
a) + 2O(a) f N2(g) + 2O(a); (II) via the intermediate
5
1
decomposition of N2O(a), i.e., 2NO(a) f O(a) + N(a) + NO-
a) f N2O(a) + O(a) f N2(g) + 2O(a); and (III) through the
[
(
desorption of N2O(a), i.e., 2NO(a) f O(a) + N(a) + NO(a) f
N2O(a) + O(a) f N2O(g) + O(a). Here, the removal of O(a)
as CO2 is omitted for simplicity. According to these reaction
pathways, the product ratio must be CO2:N2:N2O ) (2x + 2y
The angular distribution of desorbing N2 in the thermal
decomposition of N2O on Pd(110) was sharply collimated at
4
0-50° off the surface normal in the plane along the [001]
13
direction. The product N2 showed high excess translational
+
z):(x + y):z, where x and y represent the progress of N2
energy.1
1,14
Both angular and velocity distributions are very
formation in the former two processes, I and II, and z indicates
that of N2O formation in the latter, III. The above-observed value
on Pd(110) at 600 K is 5:2:1, consistent with this stoichiometry,
yielding the ratio of 2:1 for (x + y):z. This ratio and the quantity
of [CO2]/{2[N2] + [N2O]} are shown in Figure 14a as a function
of the surface temperature. [CO ], [N ], and [N O] designate
similar to those observed in the steady-state NO + CO reaction
on Pd(110) in the present work. We conclude that the inclined
N2 desorption comes from the intermediate N2O decomposition
in the steady-state NO + CO reaction.
(
D) Inclined N2 Desorption. Adsorbed N2O is partly
2
2
2
decomposed on Rh(110) even around 60 K, as confirmed by
recent NEXAFS measurements.52 The angular distribution of
desorbing product N2 in N2O decomposition on this surface was
more complex than that on Pd(110), as recently reported by
2 2 2
the amount of produced CO , N , and N O, respectively,
estimated from either AI or AR signals. The latter quantity
should always be unity when no other reactions take place. This
was confirmed for the observed values from the AI signals, as
shown by open squares in Figure 14a. The values are scattered
around unity because of the large error involved in the AI-N2
signal. The (x + y):z ratio represents the branching of N :N O
2
,3,6
AR-TPD.
When the clean surface was exposed to N2O, the
product N2 desorption in the subsequent TPD procedure
collimated along about 70° off the surface normal toward the
2
2
[001] direction. This desorption was proposed to be due to the
formation. The observed ratio steeply increased with increasing
temperature above around 600 K; at higher surface temperatures,
the N2O formation was relatively more suppressed than the total
N2 formation. However, this branching parameter does not
present how much N2 is formed through the N2O intermediate
decomposition of surface-parallel N2O oriented along the [001]
direction. The N2O decomposition was largely retarded on
oxygen-covered Rh(110). The surface reactivity toward the N2O
decomposition was recovered after being heated above 1100

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14241
TABLE 1: Anisotropic Desorption Parameters on Pd(110)
direction
CO
2
N
2
(inclined)
N
2
(normally)
N
2
(cosine)
1
5
28
5
1
[
[
001]
1 1h 0]
cos (θ)
cos (θ)
cos (θ)
cos (θ)
3
14
a
5
1
cos (θ)
cos (θ)
cos (θ)
cos (θ)
refs
this work,55 this work,7
this work
this work
a
This sharpness represents the angle dependence in the declining
plane passing the collimation position and along the [1 1h 0] direction.
including the thermalized component, and show the degree to
which the above two pathways, I and II, share the total N2 yield.
The branching in these two pathways can be estimated from
angular distribution analysis. The observed angular distribution
in the plane along the [001] direction was approximated as
2
8
28
5
0
.95{cos (θ + 41) + cos (θ - 41)} + 0.03 cos (θ) + 0.07
cos(θ) at TS ) 550 K. The prefactors indicate the relative
intensity. The contribution of the cosine component was
estimated from the deconvolution of the velocity distributions.
Below 600 K, N2 is mostly desorbed through the inclined way.
2
8
At 650 K, the distribution became 0.85{cos (θ + 40) +
28
5
cos (θ - 40)} + 0.16 cos (θ) + 0.16 cos(θ); i.e., the normally
directed components were enhanced. This tendency became
clearer at higher temperatures, indicating a sharing shift in the
reaction pathways. At higher temperatures, the surface nitrogen
removal was highly contributed from the associative process
of N(a).
The total amount of CO2 produced was estimated from the
AR-signal intensity at the collimation angle, the sharpness of
the angular distribution around the collimation axis, and the mass
sensitivity correction due to velocity. The total amount of
n
m
desorbing species in the form of cos (θ) cos (θ) is given by
-
1/2
-1/2
Q ) I (n + 1) (m + 1)
0
where Q is the total amount and n and m are the power value
of the power series of cos(θ) in the plane along the definite
crystal azimuth and that perpendicular to it, respectively.25 By
considering the velocity differences among N2 and N2O and
the anisotropy of the angular distribution around the collimation
5
,15,55
axis for N2,
the formation of N2 in the inclined way, the
normally directed form, and the cosine form can be separately
estimated. The amounts of CO2 and N2O were also estimated
from the AR signals in a similar way. The used desorption
parameters of CO2 and N2 are summarized in Table 1.
-
5
The results at the 7.5 × 10 Torr of the (1:1) mixture are
shown in Figure 14b. The N2 formation in each pathway was
separately estimated. The results were again examined by the
stoichiometry in Figure 14a. The contribution from the normally
directed component in a cos (θ) form is always minor because
of the formation in a small amount. The anisotropic distribution
Figure 14. Stoichiometry and branching of the steady-state NO +
-
5
CO reaction on Pd(110) at the total pressure of 7.5 × 10 Torr with
1
5
12
14
13
5
a
NO: CO ) 1:1 or NO: CO ) 1:1 mixture. (a) Ratio of the amount
1
3
15
15
of CO
2
to that of (2 N
2
+
2
N O) calculated from (0) AI signals
1
5
and (9) AR signals. The amount of
2
N from the AR signals was
of desorbing CO2 has been well-studied and is shown in Table
estimated by assuming the value of 14 for the power of the power
series of the desorption angle in the inclined plane passing the
collimation position and along the [1 1h 0] direction (see text). The
9
1
. The distribution of desorbing N2 was frequently assumed to
be symmetric around the collimation axis at about 40° because
of the difficulty in getting reliable data of angle-resolved
measurements in the inclined plane passing the collimation
position. To get the stoichiometry, the power in the power series
of the cosine of the desorption angle along the [1 1h 0] direction
was adjusted to the value of 14 ( 5 at 600 K. It agreed well
1
5
15
branching ratio by ( N
2 2
):( N O) is also shown. (b) Integrated amounts
1
5
of nitrogen evolved in different ways: (9)
[) normally directed desorption; (b) cosine component; and (0) N O
2
in a cosine distribution. (c) Branching of
2 2
N formation: (0) through the intermediate N O (inclined desorp-
2
N inclined desorption;
15
(
1
5
2
N formation. Fractions of
15
15
tion); (4) through the associative process of N(a) (the normally directed
desorption); and (9) the thermalized cosine desorption. The former
process prevails at low temperatures, whereas, at high temperatures,
the cosine component becomes predominant. Vertical bars indicate
experimental errors. The lines in the figure are “guides to the eye”.
2
0
with the report by Ikai and Tanaka.
Below 600 K, N2 is mostly formed from the N2O decomposi-
tion, and N2O formation is also significant. Above this tem-
perature, both the N2 formation from N2O(a) and the N2O
desorption decrease rapidly, whereas the N2 signal at the normal
direction over that shared by the former process becomes
noticeable. The N2 desorption in the cosine form increases more
pathway; i.e., the values of x and y are not separately determined.
In the following analysis, we separately discuss these quantities,

14242 J. Phys. Chem. B, Vol. 108, No. 38, 2004
Rze z´ nicka et al.
rapidly. The resultant branching of N2 formation is shown in
Figure 14c. The results above 750 K involved large uncertainty.
In the figure, the inclined desorption represents the N2 emission
from the reaction pathway II through the decomposition of
receives more energy into the vibrational modes. This branching
change may compensate the rapid decrease of the fast compo-
nent at large desorption angles.
Such highly excited molecules originate from a large energy
dissipation in the N2 desorption processes. For the process of
N2O(a) f N2(g) + O(a), the energy which the product N2 can
carry out is presented by ∆ET, which is given as ∆ET ) EN2(g)
+ EO(a) - EN2O(a,TS), where EN2(g), EO(a), and EN2O(a;TS) are the
potential energies of N2(g), O(a), and the transition state of N2O-
5
intermediate N2O. The normally directed component in a cos -
(θ) form is ascribed to the reaction pathway I via the N(a)
associative process. The remaining cosine component becomes
major at high temperatures. This cosine component showed
similar kinetic behavior to that of pathway I with respect to the
surface temperature and the PCO/PNO ratio, suggesting that it
was also due to the associative process. There must be a
mechanism whereby the associative process is likely to yield a
thermalized component. In general, the N2 formation through
the associative process of N(a) is enhanced at higher temper-
atures.
-
1
(a) dissociation, respectively. By assuming 400-500 kJ mol
6
1
as the bond energy of O-metal, the available energy was
-
1
estimated to be 240-340 kJ mol because the dissociation of
3
N2O(g) f N2(g) + O(g)( P) is endothermic by about 160 kJ
-
1
mol and the heat of adsorption of N2O on both Pd(110) and
Rh(110) is close to the activation energy of N2O(a) dissociation.5
In this process, the emitted energy mostly comes from the
metal-O bond formation. In the associative processes of N(a),
the released energy may become higher because of the high
On Rh(110), the N2 formation mostly proceeds through the
associative process in the NO pressure range studied, although
the contribution from the inclined N2 desorption pathways is
highly possible at lower temperatures. AR-SSD measurements
-
1
bond energy of N-N (945 k J mol ) and the low binding of
5
8-60
-
4
-2
N-noble metals.
are desirable in the pressure range of 10 -10 Torr and at
lower temperatures.
V. Summary
(F) Desorption Dynamics. Two or three fast-desorption
components were already found in desorbing product N2
The angular and velocity distributions of desorbing products
N and CO were successfully analyzed in the steady-state NO
+ CO reaction on Pd(110) and Rh(110) in a wide range of NO
pressures by means of angle-resolved product desorption with
cross-correlation time-of-flight techniques and LEED observa-
tions. The following results were obtained.
56
molecules in the NO + CO reaction on Pt(100), in the N2O
2
2
1
1,14
decomposition on Pd(110),
on Pd(110).
and in the NO decomposition
1
1,12
These N2 molecules have very high velocity;
furthermore, their velocity distribution curve is commonly wider
than that of the Maxwellian form at a definite temperature. These
fast components were proposed to be due to the different
vibrational states because of the energy difference close to the
vibrational excitation of N2, 0.28 eV (1610 K or 26.9 kJ); i.e.,
the faster component is in the lower vibrational energy state.
The velocity distribution of N2 in the present work is wide
enough to involve several vibrational excited levels. In fact,
desorbing N2 in the associative process of N(a) on Cu(111),
Ag(111), and Ru(001) showed very wide translational energy
(1) On Pd(110), N desorption was split into two inclined
2
components collimating at (40° in the plane along the [001]
direction. On the basis of the close similarity of angular and
velocity distributions in N O decomposition on Pd(110), the
2
inclined N formation is proposed to originate from the N O
2
2
intermediate. At low temperatures, the pathway through the N O
2
intermediate prevails, and, above 720 K, the associative nitrogen
desorption starts to dominate.
57-59
distributions and high populations of vibrational excited states.
(2) N desorption on Rh(110) was sharply collimated along
2
Higher time-resolution TOF measurements in the state-selective
the surface normal in a wide temperature region, indicating that
form are required.60
N(a) is mostly removed through the associative process.
(3) On both surfaces, the translational temperature of des-
orbing N2 was very high, reaching about 2500-3500 K. On
the other hand, the CO2 desorption always collimated along the
surface normal on both surfaces with the translational temper-
ature in the range of 1600-2000 K.
The composition with respect to different velocity components
of desorbing N2 was, surprisingly, fairly insensitive toward the
desorption angle on both Rh(110) and Pd(110). Here are several
possibilities for this occurrence. On Rh(110), the presence of
inclined desorption in the reaction N2O(a) f N2(g) + O(a) is
highly possible.3 The observed T〈E〉 would be insensitive to
the desorption angle if the inclined component of N2 contributed
to T〈E〉 at large desorption angles because the translational energy
should be maximized at the collimation angle. In fact, the
angular distribution becomes sharp at higher temperatures
,6
Acknowledgment. I.I.R. is indebted to the Ministry of
Education, Sports, and Culture of Japan for her scholarship
(2001-2004). This work was partly supported by a 1996 COE
special equipment program of the said Ministry. It was also
supported in part by Grant-in-Aid No. 13640493 for General
Scientific Research from the Japan Society for the Promotion
of Science. The authors thank Ms. Atsuko Hiratsuka for drawing
the figures.
(Figure 10). The suppression of the inclined component is
reasonably expected at higher temperatures. On the other hand,
the angular distribution representing a single process usually
becomes broader at higher temperatures, as seen in the CO2
velocity distributions on both Pd(110) and Rh(110), because of
the enhanced thermal motion.5,9
References and Notes
For the inclined desorption on Pd(110), the following
possibilities are expected: (1) significant anisotropy in velocity
distributions of each desorption component is present, i.e., a
(1) Rahkamaa, K.; Salmi, T. Chem. Eng. Sci. 1999, 54, 4343.
(2) Horino, H.; Rze z´ nicka, I.; Kokalj, A.; Kobal, I.; Hiratsuka, A.;
Ohno, Y.; Matsushima, T. J. Vac. Sci. Technol., A 2002, 20, 1592.
5
(3) Imamura, K.; Horino, H.; Rze z´ nicka, I.; Kobal, I.; Kokalj, A.; Ohno,
rapid decrease of the fastest component would be observed in
the plane along the [0 1h 1] direction rotated 90° from the present
azimuth; (2) a new energy transfer mechanism is operative, i.e.,
the energy transfer from the transition state into the translational
mode may be preferred at larger shifts from the collimation
angle, whereas desorbing N2 around the collimation angle
Y.; Nieuwenhuys, B. E.; Hiratsuka, A.; Matsushima, T. Surf. Sci., 2004, in
press.
(
(
(
4) Matsushima, T. Catal. SurV. Jpn. 2002, 5, 71.
5) Matsushima, T. Surf. Sci. Rep. 2003, 52, 1.
6) Liu, S.; Horino, H.; Kokalj, A.; Rze z´ nicka, I.; Imamura, K.; Ma,
Y.-S.; Kobal, I.; Ohno, Y.; Hiratsuka, A.; Matsushima, T. J. Phys. Chem.
B 2004, 108, 3828.

NO + CO Reaction on Pd(110) and Rh(110)
J. Phys. Chem. B, Vol. 108, No. 38, 2004 14243
(
(
(
7) Ikai, M.; Tanaka, K. I. J. Phys. Chem. B 1999, 103, 8277.
8) Garin, F. Appl. Catal., A 2001, 222, 183.
9) Matsushima, T. Heterog. Chem. ReV. 1995, 2, 51.
(34) Dhanak, V. R.; Comelli, G.; Paolucci, G.; Prince, K. C.; Rosei, R.
Surf. Sci. 1992, 260, L24.
(35) Wang, H.; Tobin, R. G.; DiMaggio, C. L.; Fisher, G. B.; Lambert,
D. K. J. Chem. Phys. 1997, 107, 9569.
(36) Umbach, E.; Menzel, D. Chem. Phys. Lett. 1981, 84, 491.
(37) Ceballos, G.; Wende, H.; Baberschke, K.; Arvanitis, D. Surf. Sci.
2001, 482-485, 15.
(38) Kodama, C.; Orita, H.; Nozoye, H. Appl. Surf. Sci. 1997, 121/122,
579.
(39) Sau, R.; Hudson, J. B. J. Vac. Sci. Technol. 1981, 18, 607.
(40) Hoffman, D. A.; Hudson, J. B. Surf. Sci. 1987, 180, 77.
(41) Li, Y.; Bowker, M. Surf. Sci. 1996, 348, 67.
(42) Avery, N. R. Surf. Sci. 1983, 131, 501.
(43) Cornish, J. C. L.; Avery, N. R. Surf. Sci. 1990, 235, 209.
(44) V a¨ terlein, P.; Krause, T.; B a¨ ssler, M.; Fink, R.; Umbach, E.;
Taborski, J.; W u¨ stenhagen, V.; Wurth, W. Phys. ReV. Lett. 1996, 76, 4749.
(45) Schwaner, L.; Mahmood, W.; White, J. M. Surf. Sci. 1996, 351,
228.
(
10) Kobal, I.; Rze z´ nicka, I.; Matsushima, T. Recent Research DeVelop-
ments in Physical Chemistry; Transworld Research Network: Trivandrum,
India, 2002; Vol. 6, p 391.
11) Ohno, Y.; Kimura, K.; Bi, M.; Matsushima, T. J. Chem. Phys. 1999,
10, 8221.
12) Kobal, I.; Kimura, K.; Ohno, Y.; Matsushima, T. Surf. Sci. 2000,
45, 472.
13) Horino, H.; Liu, S.; Hiratsuka, A.; Ohno, Y.; Matsushima, T. Chem.
Phys. Lett. 2001, 341, 419.
14) Ohno, Y.; Kobal, I.; Horino, H.; Rze z´ nicka, I.; Matsushima, T. Appl.
Surf. Sci. 2001, 169/170, 273.
15) Kobal, I.; Kimura, K.; Ohno, Y.; Horino, H.; Rzeznicka, I.;
Matsushima, T. Stud. Surf. Sci. Catal. 2000, 130, 1337.
(
1
4
(
(
(
(
(
(
16) Arumainayagam, C. R.; Madix, R. J. Prog. Surf. Sci. 1991, 38, 1.
17) Colonell, J. I.; Gibson, K. D.; Sibener, S. J. J. Chem. Phys. 1996,
1
(46) Herman, J. S.; Peden, C. H. F.; Schmieg, S. J.; Belton, D. N. Catal.
Lett. 1999, 62, 131.
(47) Haq, S.; Hodgson, A. Surf. Sci. 2000, 463, 1.
(48) Huang, H. H.; Seet, C. S.; Zou, Z.; Xu, G. Q. Surf. Sci. 1996, 356,
181.
Chem. Phys. 1997, 106, 311.
(
21) Belton, D. N.; DiMaggio, C. L.; Schmieg, S. J.; Simon Ng, K. Y.
(49) Kokalj, A.; Kobal, I.; Horino, H.; Ohno, Y.; Matsushima, T. Surf.
Sci. 2002, 506, 196.
J. Catal. 1995, 157, 559.
22) Zaera, F.; Gopinath, C. S. Chem. Phys. Lett. 2000, 332, 209.
(50) Kokalj, A.; Kobal, I.; Matsushima, T. J. Phys. Chem. B 2003, 107,
2741.
T. J. Chem. Phys. 2003, 119, 9829.
(51) Horino, H.; Rze z´ nicka, I.; Matsushima, T.; Takahashi, K.; Naka-
mura, E. UVSOR ActiVity Report 2002; Okazaki, Japan, 2003; p 209.
(52) Horino, H.; Rze z´ nicka, I.; Matsushima, T.; Takahashi, K.; Naka-
mura, E. UVSOR ActiVity Report 2002: Okazaki, Japan, 2003; p 211.
(53) Zhanpeisov, N. U.; Martra, G.; Ju, W. S.; Matsuoka, M.; Coluccia,
S.; Anpo, M. J. Mol. Catal. A: Chem. 2003, 201, 237.
(54) Yakovlev, A. L.; Zhidomirov, G. M.; Van Santen, R. A. Catal.
Lett. 2001, 75, 45.
(55) Rze z´ nicka, I.; Ohno, Y.; Horino, H.; Kobal, I.; Kimura, K.;
Matsushima, T., Proc. 9th Int. Heterog. Catal. Varna 2000, 97.
(56) Ohno, Y.; Sarawut, P.; Horino, H.; Kobal, I.; Hiratsuka, A.;
Matsushima, T. Chem. Phys. Lett. 2003, 373, 161.
(57) Murphy, M. J.; Skelly, J. F.; Hodgson, A. Chem. Phys. Lett. 1997,
279, 112.
(
24) Comsa, G.; David, R. Surf. Sci. Rep. 1985, 5, 145.
(25) Cao, G.; Moula, Md. G.; Ohno, Y.; Matsushima, T. J. Phys. Chem.
B 1999, 103, 3235.
26) Rze z´ nicka, I.; Matsushima, T. Chem. Phys. Lett. 2003, 377,
79.
27) Ehrlich, G. In Chemistry and Physics of Solid Surfaces; Vancelow,
(
2
(
R., How, R., Eds.; Springer-Verlag: Berlin, Heidelberg, Germany, 1988;
Vol. VII, p 1.
(
28) Moula, Md. G.; Wako, S.; Cao, G.; Kimura, K.; Ohno, Y.; Kobal,
I.; Matsushima, T. Phys. Chem. Chem. Phys. 1999, 1, 3677.
29) Wako, S.; Moula, Md. G.; Cao, G.; Ohno, Y.; Kobal, I.; Mat-
sushima, T. Langmuir 2000, 16, 2689.
30) Lizzit, S.; Comelli, G.; Hofmann, Ph.; Paolucci, G.; Kiskinova, M.;
Rosei, R. Surf. Sci. 1992, 276, 144.
31) Dhanak, V. R.; Comelli, G.; Cautero, G.; Paulucci, G.; Prince, K.
C.; Kiskinova, M.; Rosei, R. Chem. Phys. Lett. 1992, 188, 237.
(
(
(58) Carter, R. N.; Murphy, M. J.; Hodgson, A. Surf. Sci. 1997, 387,
102.
(59) Murphy, M. J.; Skelly, J. F.; Hodgson, A.; Hammer, B. J. Chem.
Phys. 1999, 110, 6954.
(
(
(
32) Rze z´ nicka, I.; Matsushima, T. J. Phys. Chem. B 2003, 107, 8479.
33) Bowker, M.; Guo, Q.; Li, Y.; Joyner, R. W. J. Chem. Soc., Faraday
(60) Hodgson, H. Prog. Surf. Sci. 2000, 63, 1.
Trans. 1995, 91, 3663.
(61) Brown, W. A.; Kose, R.; King, D. A. Chem. ReV. 1998, 98, 797.