Surface-nitrogen removal in a steady-state NO + H2 reaction on Pd(110)

Article abstract of DOI:10.1021/jp0461450

Surface-nitrogen removal steps were analyzed in the course of a catalyzed NO + H₂ reaction on Pd(110) by angle-resolved mass spectroscopy combined with cross-correlation time-of-flight techniques. Four removal steps, i.e., (i) the associative process of nitrogen atoms, 2N(a) → N ₂(g), (ii) the decomposition of the intermediate, NO(a) + N(a) → N₂O(a) → N₂(g) + O(a), (iii) its desorption, N ₂O(a) → N₂O(g), and (iv) the desorption as ammonia, N(a) + 3H(a) → NH₃(g), are operative in a comparable order. Above 600 K, process (i) is predominant, whereas the others largely contribute below 600 K. Process (iv) becomes significant at H₂ pressures above a critical value, about half the NO pressure. Hydrogen was a stronger reagent than CO toward NO reduction and relatively enhanced the N(a) associative process.

Full text of DOI:10.1021/jp0461450

1256
J. Phys. Chem. B 2005, 109, 1256-1261
Surface-Nitrogen Removal in a Steady-State NO + H2 Reaction on Pd(110)
Yunsheng Ma and Tatsuo Matsushima*
Catalysis Research Center, Hokkaido UniVersity, Sapporo 001-0021, Japan
ReceiVed: August 26, 2004; In Final Form: October 16, 2004
Surface-nitrogen removal steps were analyzed in the course of a catalyzed NO + H2 reaction on Pd(110) by
angle-resolved mass spectroscopy combined with cross-correlation time-of-flight techniques. Four removal
steps, i.e., (i) the associative process of nitrogen atoms, 2N(a) f N2(g), (ii) the decomposition of the
intermediate, NO(a) + N(a) f N2O(a) f N2(g) + O(a), (iii) its desorption, N2O(a) f N2O(g), and (iv) the
desorption as ammonia, N(a) + 3H(a) f NH3(g), are operative in a comparable order. Above 600 K, process
(i) is predominant, whereas the others largely contribute below 600 K. Process (iv) becomes significant at H2
pressures above a critical value, about half the NO pressure. Hydrogen was a stronger reagent than CO toward
NO reduction and relatively enhanced the N(a) associative process.
I. Introduction
structure was later confirmed by density functional theory
(DFT) calculations and near-edge X-ray absorption fine structure
(NEXAFS) and scanning tunneling microscopy (STM) work.13-15
This peculiar N2 desorption is useful to analyze the pathway
of the removal of surface nitrogen because the associative
desorption of N(a) emits N2 sharply along the surface normal.8,16
Furthermore, the angular distribution is always related to the
product desorption step whenever any step becomes rate-deter-
mining.17,18 The angular and velocity distributions of desorbing
products in the NO decomposition have been analyzed with
several relaxation methods, such as modulated molecular
beams19 and AR-TPR.8-11,20-22 Steady-state conditions, how-
ever, could not be established for the reaction, and then simple
phenomena were not differentiated from kinetic behavior under
steady-state conditions.8,22 In the present work, AR product
desorption measurements were successfully performed for a
steady-state NO + H2 reaction.
The NO reduction by CO, H2, and hydrocarbon on rhodium
and palladium surfaces has received much attention because of
its importance in controlling automobile exhaust gas.1 In this
catalytic reduction, N2O is concomitantly produced as one of
the undesired byproducts. N2O itself is harmful and has a
remarkable greenhouse effect. In the NO + H2 reaction, another
undesired product, NH3, is formed. Therefore, it is necessary
to improve the selectivity to N2 in these catalytic processes;
however, knowledge of the reaction mechanism is still limited,
especially regarding the steps for the removal of surface
nitrogen. These are difficult to analyze because of the presence
of several rapid pathways after the slow NO dissociation. This
paper contains the first confirmation that the N2O intermediate
decomposition is the final step of the main pathway in a steady-
state NO + H2 reaction on Pd(110).
Many publications have dealt with the NO + H2 reaction on
Pt, Rh, and Pt-Rh alloy surfaces.2-8 However, most of them
have focused on the nonlinear behavior of surface reactions, in
which the removal of surface nitrogen was tacitly assumed to
be rapid enough not to play a rate-limiting role. Ikai et al. found
that, in the course of heating in their angle-resolved (AR)
temperature-programmed reaction (TPR) of NO and H2 or CO
on Pd(110), the N2 peak at 490 K involved desorption collimated
at 38° off normal toward the [001] direction and desorbing N2
in the other peak at around 600 K collimated at the surface
normal, whereas the off-normal peak at 490 K was absent in
the subsequent cooling.8 Furthermore, they confirmed that the
reaction between 14N(a) and 15NO(a) emitted the product 14N15N
in an inclined way, while the associative nitrogen desorption
collimated along the surface normal. The authors argued that
the inclined N2 desorption originated in the desorption-mediated
reaction without the formation of the intermediate N2O. On the
other hand, our previous study revealed that desorbing N2 from
N2O dissociation, NO decomposition, and a steady-state NO +
CO reaction on Pd(110) commonly showed identical angular
and velocity distributions.9-12 Thus, N2O(a) was proposed to
be oriented along the [001] direction before dissociation. This
II. Experiments
The apparatus has three separately pumped chambers.17,23 The
reaction chamber is equipped with reverse-view low-energy
electron diffraction (LEED) and X-ray photoelectron spectros-
copy (XPS) optics, an ion gun, and a quadrupole mass spec-
trometer (QMS) for angle-integrated (AI) measurements. The
chopper house, which has a large pumping rate of about 7 m3‚s-1
for a high angle resolution,24 has a narrow slit facing the reaction
chamber and a cross-correlation random chopper blade. Another
QMS was used in the analyzer connected through a narrow tube
for AR product desorption and time-of-flight analyses. The
distance from the ionizer to the chopper blade was 377 mm,
and the time resolution was selected at 20 µs.
15NO was introduced through a doser with a small orifice
(diameter 0.1 mm) about 2 cm from a sample crystal while D2
or CO was backfilled. The product 15N2, 15ND3, D2O, and 15N2O
signals were monitored in both AI and AR forms. Hereafter,
these are simply described as N2, ND3, D2O, and N2O in the
text. The desorption angle (polar angle, θ) was scanned in the
plane along the [001] direction.17 The N2 signals in both QMSs
were corrected by the contribution due to the fragmentation of
N2O. The pressures of reactant gases were also corrected by
their mass spectrometer sensitivities.
* To whom correspondence should be addressed. Fax: +81-11-706-
9120. E-mail: tatmatsu@cat.hokudai.ac.jp.
10.1021/jp0461450 CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/29/2004

Nitrogen Removal on Pd(110)
J. Phys. Chem. B, Vol. 109, No. 3, 2005 1257
Figure 1. TS dependence of (a) AI and (b) AR signals of products at the collimation positions in a steady-state 15NO + D2 reaction at PNO ) 5
× 10-6 Torr with the pressure ratio 15NO/D2 ) 2. Angular distributions of desorbing 15N2 at (c) 530 K and (d) 640 K. Typical deconvolutions
shown by broken curves are based on the velocity distribution analysis. The ordinate was normalized to the maximum at 530 K.
III. Results
noticed that the AR N2O signal at θ ) 0° followed a TS
dependence similar to that of N2 at θ ) 40°.
A. General Features. The AI signal was determined by the
QMS in the reaction chamber as the difference in the signal
between the desired surface temperature (TS) and room tem-
perature. The AI signals are shown for desorbing N2, N2O, ND3,
and D2O at a fixed NO pressure of 5 × 10-6 Torr and the
Only the (1 × 1) pattern was observed in LEED measure-
ments under a steady-state NO + D2 reaction at the NO pressure
of 5 × 10-8 Torr and PNO/PD ) 2 in the range of TS ) 450-
2
800 K. This is very different from the results of the CO + O2
reaction showing c(2 × 4) or (1 × n) structures.25
pressure ratio PNO/PD ) 2 (Figure 1a). All the rates are
2
B. Angular Distribution. Both N2O and ND3 desorption
showed the cosine distribution characteristic of the desorption
after thermalization to TS. It was confirmed by the velocity
distribution, which was fully described by a Maxwellian
distribution at the surface temperature; i.e., N2O and ND3 were
trapped on the surface before desorption. On the other hand,
the angular distribution of desorbing N2 changed significantly
with increasing temperature. The results are shown in Figure
1c,d. At TS ) 530 K, N2 desorption mostly collimated around
40°, while at 640 K (Figure 1d), the intensity of the N2 signal
at the normal direction increased. The distribution was decon-
voluted into three components on the basis of the following
velocity distribution analysis.
negligible below 490 K, increase rapidly at around 510 K to a
maximum with increasing TS, and decrease again above 650
K. No hysteresis was found in their rates with decreasing TS.
With the pressure ratio PNO/PD ) 2, ND3 formation was
negligible in the temperature range studied. In fact, ND3
formation was observed only at higher PD as described in the
following section.
2
2
On the other hand, the AR signal was obtained by the QMS
in the analyzer as the difference between the signal at the desired
angle and that when the crystal was away from the line-of-
sight position. The AR N2 signals at θ ) 0° and 40° are
displayed versus TS because these are their collimation angles
(Figure 1b). There are noticeable differences in their temperature
dependence. The AR N2 signal at 40° becomes noticeable at
500 K, increases quickly to the maximum at 530 K, and then
begins to decrease above 550 K. On the other hand, the N2 signal
at θ ) 0° increases slowly above 500 K and reaches the
maximum at around 640 K. This indicates remarkable changes
in the angular distributions with increasing TS. Both N2 signals
at different desorption angles were highly reduced at higher
temperatures but still noticeable even at 800 K. It should be
Typical velocity distribution curves at different desorption
angles at TS ) 640 K are shown in Figure 2. The apparent
translational temperature calculated from the average kinetic
energy ( E ) as T E ) E /2k, is shown in angular brackets in
the figure, where k is the Boltzmann constant. The value was
maximized at around the collimation angle to the value of 2200
K at TS ) 640 K. It decreased quickly with an increasing shift
from the collimation angle. This angle dependence was con-
sistent with the inclined desorption. The velocity distribution

1258 J. Phys. Chem. B, Vol. 109, No. 3, 2005
Ma and Matsushima
Figure 3. PD dependence of each component formation at a fixed
15NO pressure2of 5 × 10-6 Torr: (a) AR signals of different products
at TS ) 550 K and (b) component flux after integration, (c) component
flux after integration at TS ) 640 K. The normally directed 15N2
involved the cosine component. The vertical broken line indicates the
kinetic transition point.
Figure 2. Velocity distributions of desorbing 15N2 at different angles.
The average kinetic energy is indicated within angular brackets in the
temperature units. Typical deconvolutions (see the text) are drawn by
broken curves. The solid line indicates their summation. The condition
is at TS ) 640 K in Figure 1.
was approximated as 0.51 cos θ + 0.26 cos5 θ + 0.3{cos28(θ
+ 40) + cos28(θ - 40)}. The inclined N2 desorption was due
to N2O decomposition, and the normally directed N2 component
comes from the associative desorption of N(a). These compo-
nents showed different behavior toward the variation of the
surface temperature and the pressures of D2 and NO, which
also suggests that they come from different nitrogen removal
steps. The cosine component, however, cannot be assigned to
a definite process from the angular distribution.
involved components faster than those expected by the Maxwell
distribution at the surface temperature of 640 K. The latter
component was expected to follow the cosine distribution, and
it was first subtracted from the velocity distribution curves. The
resultant velocity curve provided the flux of the fast component,
which peaked at θ ) 0° and 40°. The fast component with
translational temperature about 3 times higher than the surface
temperature was observed at the normal direction. On the other
hand, the fast component at θ ) 40° reached the maximum
with a translational temperature of about 3200 K.
C. Hydrogen Effect. The product formation kinetics changed
sharply at a critical D2 pressure. The AR signals of different
products in a steady-state NO + D2 reaction at 550 K are
displayed as a function of the deuterium pressure in Figure 3a,
where the NO pressure was fixed at 5 × 10-6 Torr. With
Assuming a power series of the cosine of the desorption angle
to each component, the N2 distribution was in a two-directional
form approximated as cos28(θ + 40) + cos28(θ - 40) and the
N2 signal at the normal direction was very small below 600 K.
The total signal at 530 K, for example, can be described as
0.95{cos28(θ + 40) + cos28(θ - 40)} + 0.03 cos5(θ) + 0.07
cos(θ). At higher TS ) 640 K, the normally directed component
was enhanced. The signal intensity at the normal direction
increased with increasing TS, indicating the presence of the
normally directed fast component. The total signal at 640 K
increasing PD , the AR N2 signal at 40° increases, is maximized
2
at PD ) 1 × 10-6 Torr, and slightly decreases above this level.
2
Both the N2 and N2O signals at θ ) 0° also showed similar
dependence. These AR signals were integrated around their
collimation angles, yielding the total formation of each com-
ponent as shown in Figure 3b.26 Here, the normally directed
N2 desorption in this figure was shown as the summation

Nitrogen Removal on Pd(110)
J. Phys. Chem. B, Vol. 109, No. 3, 2005 1259
the other hand, the normally directed N2 desorption can be
enhanced by hydrogen more than by CO.
IV. Discussion
A. Desorption Components. Four different desorption
pathways of N-containing products are operative in a steady-
state NO + D2 reaction on Pd(110). For the N2 formation, two
different pathways work and show different spatial distribu-
tions: one is the associative process, which emits N2 along the
surface normal, and the other is due to decomposition of the
N2O(a) intermediate and emits N2 along 40° off the surface
normal in the plane along the [001] direction. In their AR-TPR
work of NO + H2, Ikai et al. argued that the inclined desorption
collimated at 38° was due to the desorption-mediated reaction
without passing the intermediate N2O(a).8 In our experiments,
however, the N2 desorption with different collimation angles
was assigned to different N2 formation channels. Ammonia was
Figure 4. PD2 dependence of the flux ratio of the cos θ component to
that of the cos5 θ component at θ ) 0°. Each flux is designated by
5
I
cos θ and Icos θ. These were derived from velocity distribution analysis
at TS ) 640 K.
the main N-containing product at higher PD , as reported in the
2
work of Ikai et al., where the H2 pressure was about 2 orders
including the cosine N2 component because of the same kinetic
behavior as described below. Below the critical pressure, the
inclined N2 desorption and N2O formation increase with a slope
of about 1.5 and slightly decrease in a similar way above it.
Furthermore, these components showed a similar TS dependence.
higher than that of NO.8
By state-resolved desorption measurements, Hodgson reported
that the N2 formed in the NO + H2 reaction on Pd(110) carried
considerable vibrational excitation with no excess translational
and rotational energy.28 They proposed that N2 that was formed
by the recombination process desorbed via a molecular chemi-
sorption state with an extended N2 bond. It is difficult to
compare their results with ours because of the lack of detailed
information regarding the conditions of the experiment. Ac-
cording to our experiments, the apparent translational energy
of the products depends on the surface temperature, the NO/H2
pressure ratio, and the desorption angle. For example, the
thermalized component would contribute mainly to the N2
desorption at higher TS and higher H2/NO pressure ratio for
non-angle-resolved measurements.
On the other hand, the N2 that desorbed along the normal
direction increases more rapidly with increasing PD below the
2
critical point, remains invariant above it, and decreases rapidly
with a further increase. This characteristic became clear at 640
K, where the slope of about 2 was much larger than that for the
inclined N2 desorption and N2O formation. The ammonia
formation became significant only above the critical pressure
at 550 K and started at slightly higher D2 pressure at 640 K.
It should be noticed that the cosine component also increased
with increasing D2 pressure. The AR N2 signal at the normal
direction consisted of the normally directed associative desorp-
tion and the thermalized component. Each contribution was
separated by velocity distribution analysis, as described in the
former subsection. The flux ratio of the cos θ component at θ
Recently, Goodman and co-workers reported infrared reflec-
tion (IR) spectroscopy work in a steady-state NO + CO reaction
on Pd(111), where a possible intermediate of -NCO(a) was
proposed according to the absorption signal at around 2255
) 0° to that of the cos5 θ component is shown versus PD in
cm-1 29,30
.
However, this species was not observed below 2 ×
2
10-2 Torr, although the reaction rapidly proceeded even far
below this pressure, and no pathways emitting N-containing
products were proposed. Unfortunately, the N2O(a) species
giving the absorption in a similar frequency range did not
yield a high cross-section for IR absorption compared with
-NCO(a). It should be noted that the inclined N2 emission along
about 40° off normal is commonly observed in NO + CO and
NO + H2 reactions on Pd(110). Below about 600 K and at PNO
Figure 4. It was fairly constant. This constancy was also
observed versus the surface temperature.
D. NO + CO Reaction. For a comparison, the AR signals
of N-containing products in a steady-state NO + CO reaction
are displayed against TS in Figure 5. All the signals show a
tendency similar to that of the signals in the NO + D2 reaction.
The AR N2 signal at the collimation angle of θ ) 41° increases
steeply at around 520 K, is maximized at 550 K, and rapidly
decreases at higher temperatures. The N2O formation follows a
similar TS dependence. On the other hand, the AR N2 signal at
θ ) 0° increases slowly and is maximized at 610 K. The relative
N2 signal at θ ) 41° to that at 0° was sensitive to the PNO/PCO
ratio. The normally directed N2 desorption was enhanced at
higher CO pressures. In fact, the critical CO pressure was very
close to the NO pressure.27 This situation became clearer in the
angular distribution above 600 K. Below 550 K, the N2
desorption was merely described by the inclined ways, whereas
the normally directed component was enhanced at 600 K when
> PCO or PH , the N2 emission is common; on the other hand,
at PNO , PH , surface nitrogen is highly converted into NH3.
2
2
B. Kinetics and Branching. There is a kinetic transition point
in the NO + D2 reaction that is quite similar to that in a NO +
CO reaction.27 The overall reaction is mostly controlled by NO
dissociation. At hydrogen pressures below the critical value,
the surface is fairly covered by NO(a) and O(a). In this region,
O(a) has a retarding effect on the NO dissociation. This
dissociation or O(a) removal is rate-determining. With increasing
hydrogen pressure, the amount of O(a) decreases and the
coverage of surface nitrogen increases steeply, which is sug-
gested by the kinetic characteristics described in sections III.C
and III.D. However, no hydrogen accumulates on the surface
because of the fast H2O formation and the small heats of
adsorption for H2 and H2O. This yields increased active (vacant)
areas for NO dissociation, and most of N(a) is removed by the
reaction with NO(a) via the N2O(a) intermediate. N2O(a) is
P
NO/PCO ) 0.5. This is consistent with the results in the NO +
D2 reaction. The normally directed desorption was enhanced at
higher temperatures and higher hydrogen pressures.
The angular distribution at TS ) 640 K can be approximated
as 0.23 cos θ + 0.13 cos5 θ + 0.8{cos28(θ + 41) + cos28(θ -
41)}. The maximum absolute intensity of the inclined N2
desorption was almost the same for both reducing reagents. On

1260 J. Phys. Chem. B, Vol. 109, No. 3, 2005
Ma and Matsushima
Figure 5. TS dependence of AR signals of products at their collimation positions in a steady-state 15NO + CO reaction at PNO ) 5 × 10-6 Torr
with 15NO/CO ) (a) 0.5 and (b) 2. Angular distributions of desorbing 15N2 at 550 and 640 K at these 15NO/CO ratios are also shown. Typical
deconvolutions are drawn by broken curves. The ordinate was normalized to the maximum at 550 K.
easily decomposed on vacant parts on Pd(110), emitting N2 in
an inclined way.11 In addition, the reaction rate of the associative
N(a) desorption, process (i), is enhanced more quickly than those
of processes (ii) and (iii) since the associative desorption
involves two nitrogen atoms, the other processes are coupled
with one nitrogen atom and NO(a), and the amount of NO(a)
does not increase at a fixed NO pressure. This was actually
observed in Figure 3. At the critical point, the formation of O(a)
from the NO dissociation is balanced by its removal by H(a)
(or CO(a) when it is used). Above the critical point, on the other
hand, the surface is deficient in O(a) because of the high
hydrogen pressure, and H(a) can then accumulate, yielding NH3.
The N(a) amount decreases slowly with increasing D2, keeping
a fairly constant N2 and N2O formation. This behavior is quite
similar to the kinetic switching observed in the CO oxidation.17
It happens when the product formation step is rapid compared
with the supply of reactants on the surface.
a large amount of energy is released in process (ii) due to the
formation of the strong O-metal bond.17,31 On the other hand,
the released energy in process (i) is less because of the strong
N-metal bond.32
Both reactions of NO + CO and NO + H2 have identical
intermediates and common surface-nitrogen removal pathways,
showing similar kinetic behavior. A remarkable difference is
in the associative process of N(a), which was more favored in
the NO + H2 reaction, while the activities for N2 inclined
desorption and N2O formation were almost the same. This means
that more N(a) can be formed on the surface in a NO + H2
reaction. H(a) can remove O(a) as quickly as CO(a), but H(a)
does not accumulate at higher temperatures and shows no
retarding effect on the NO dissociation, whereas CO(a) occupies
the surface. Here, the adsorption heat of hydrogen is less than
that of CO. A similar observation was reported on Pt(100) by
using synchrotron XPS.3
At higher hydrogen pressures, however, the formation of
undesired product ammonia becomes remarkable and the surface
may involve NHX (X ) 1, 2, or 3) species.
The two N2 formation pathways show different kinetic
behavior. The normally directed desorption is favored at higher
surface temperatures because the associative process has higher
activation energy. In fact, both the formation and decomposition
of N2O have smaller activation energies.17,33 The kinetic
behavior of the thermalized component (in the cos θ form) is
always similar to that of the normally directed N2 desorption
(in the cos5 θ form) from process (i), suggesting the formation
of both components from a common process; i.e., the branching
may take place after the reaction barrier is passed. The lack of
the thermalized N2 component in a steady-state N2O + CO
reaction would support this conclusion. The cosine component
showing the desorption after thermalization comes from the
associative process. In other words, the product N2 from process
(ii) is not trapped on the surface, and the N2 from process (i) is
largely trapped. From its angular distribution, about 80% of the
C. Velocity Components. The velocity curve after subtraction
of the thermalized component is still wide at around 40° as
compared with a Maxwellian distribution. The speed ratio
(SR) defined as ( V2 / V 2 - 1)1/2/(32/9π - 1)1/2 still had larger
values than unity, where V is the velocity of the molecule, V
is the mean velocity, and V2 is the mean square velocity. The
SR value is unity for a Maxwellian distribution and becomes
smaller for distributions of molecules with hyperthermal en-
ergy.34 Thus, the distribution after subtraction of the slow
component was further deconvoluted into two components
by assuming the modified Maxwellian distribution, f(V) )
V3 exp{-(V - V0)2/R2}, where V0 is the stream velocity and R
is the width parameter.
N2 from this process was estimated to be trapped above PD
)
Generally, it is difficult to uniquely deconvolute one velocity
distribution curve into two distribution curves because two
2
1 × 10-6 Torr in Figure 3c.26 This is not unreasonable because

Nitrogen Removal on Pd(110)
J. Phys. Chem. B, Vol. 109, No. 3, 2005 1261
(2) De Wolf, C. A.; Nieuwenhuys, B. E. Catal. Today 2001, 70, 287.
(3) Rienks, E. D. L.; Bakker, J. W.; Baraldi, A.; Carabinerio, S. A. C.;
Lizzit, S.; Weststrate, C. J.; Nieuwenhuys, B. E. J. Chem. Phys. 2003, 119,
6245.
(4) Lombardo, S. J.; Fink, T.; Imbihl, R. J. Chem. Phys. 1993, 98,
5526.
(5) Zemlyanov, D. Y.; Smirnov, M. Y.; Gorodetskii, V. V.; Block, J.
H. Surf. Sci. 1995, 329, 61.
(6) Hallock, A. J.; Matthews, C. M.; Balzer, F.; Zare, R. N. J. Phys.
Chem. B 2001, 105, 8725.
(7) Colonell, J. I.; Gibson, K. D.; Sibener, S. J. J. Chem. Phys. 1996,
104, 6822.
(8) Ikai, M.; Tanaka, K. I. J. Phys. Chem. B 1999, 103, 8277.
(9) Ohno, Y.; Kimura, K.; Bi, M.; Matsushima, T. J. Chem. Phys. 1999,
110, 8221.
(10) Ohno, Y.; Kobal, I.; Horino, H.; Rzeznicka, I.; Matsushima, T. Appl.
Surf. Sci. 2001, 169/170, 273.
(11) Horino, H.; Liu, S.; Hiratsuka, A.; Ohno, Y.; Matsushima, T. Chem.
Phys. Lett. 2001, 341, 419.
(12) Kobal, I.; Kimura, K.; Ohno, Y.; Matsushima, T. Surf. Sci. 2000,
445, 472.
(13) Kokalj, A.; Kobal, I.; Matsushima, T. J. Phys. Chem. B 2003, 107,
2741.
(14) Horino, H.; Rzeznicka, I.; Matsushima, T.; Takahashi, K.; Naka-
mura, E. UVSOR ActiVity Report (2002); Institute for Molecular Science:
Okazaki, Japan, 2003; p 209.
(15) Matsushima, T.; Rzeznicka, I.; Watanabe, K.; Inokuchi, Y.;
Ohshimo, K.; Nishi, N. Nano-technology Support Report 2004; Institute
for Molecular Science: Okazaki, Japan, 2004; p 37.
(16) Matsushima, T. Surf. Sci. 1988, 197, L287.
(17) Matsushima, T. Surf. Sci. Rep. 2003, 52, 1.
(18) Kobal, I.; Rzeznicka, I.; Matsushima, T. Recent Research DeVelop-
ments in Physical Chemistry; Transworld Research Network: Trivandrum,
India, 2002; Vol. 6, p 391.
(19) Colonell, J. I.; Gibson, K. D.; Sibener, S. J. J. Chem. Phys. 1996,
104, 6822.
(20) Ikai, M.; Tanaka, K.-I. Surf. Sci. 1996, 357/358, 781.
(21) Ikai, M.; Janssen, N. M. H.; Nieuwenhuys, B. E.; Tanaka, K. J.
Chem. Phys. 1997, 106, 311.
parameters are required for each distribution on the basis of
the modified form. In this case, only three parameters were
separately determined by fitting the data points at high- and
low-velocity sides. To omit the fourth parameter, we simply
assumed a common width parameter.27 The resultant deconvo-
lutions are shown by broken curves in Figure 2b-d. At θ )
41° where the fast components merely come from the inclined
desorption, i.e., almost no contribution from the normally
directed component, the faster component shows a translational
temperature of 5460-6150 K and the slower one a translational
temperature of 2200-2310 K. This result is quite similar to
that for the NO + CO reaction on Pd(110).27 The value for the
faster component was estimated to be a similar temperature but
with an uncertainty of (800 K and that for the slower
component with an uncertainty of (300 K when four parameters
were separately adjusted. These fast components were proposed
to be due to different vibrational states because of the energy
difference close to the vibrational excitation of N2, 1600 K.12
This desorption characteristic of desorbing N2 in the inclined
way is common in NO + CO and NO + H2 reactions. It is not
affected by replacement of CO with H2 as a reducing reagent,
supporting the conclusion that the off-normal desorption is due
to N2O(a), which involves neither CO nor hydrogen.
V. Summary
The analysis of both angular and velocity distributions of
desorbing products N2, NH3, and N2O has discriminated among
four surface nitrogen-removal processes involved in a steady-
state NO + H2 reaction on Pd(110). The pathway through the
intermediate N2O(a) prevails below 600 K and at lower
hydrogen pressures, whereas, at high temperatures, the associa-
tive desorption of nitrogen adatoms dominates the removal. At
high hydrogen pressures, the NH3 formation becomes major.
As compared with CO, hydrogen shows not only a higher
reactivity in NO reduction but also a higher selectivity to the
associative N2 desorption.
(22) Ikai, M.; Tanaka, K. I. J. Chem. Phys. 1999, 110, 7031.
(23) Rzeznicka, I.; Moula, Md. G.; Morales, L.; Ohno, Y.; Matsushima,
T. J. Chem. Phys. 2003, 119, 9829.
(24) Kobayashi, M.; Tuzi, Y. J. Vac. Sci. Technol. 1979, 16, 685.
(25) Rzeznicka, I.; Matsushima, T. Chem. Phys. Lett. 2003, 377, 279.
(26) Cao, G.; Moula, Md. G.; Ohno, Y.; Matsushima, T. J. Phys. Chem.
B 1999, 103, 3235.
(27) Ma, Y.-S.; Rzeznicka, I. I.; Matsushima, T. Chem. Phys. Lett. 2004,
388, 201.
(28) Hodgson, A. Prog. Surf. Sci. 2000, 63, 1
(29) Ozensoy, E.; Hess, C.; Goodman, D. W. J. Am. Chem. Soc. 2002,
124, 8524.
(30) Ozensoy, E.; Goodman, D. W. Phys. Chem. Chem. Phys. 2004, 6,
3765.
(31) Liu, S.; Horino, H.; Kokalj, A.; Rzeznicka, I.; Imamura, K.; Ma,
Y.-S.; Kobal, I.; Ohno, Y.; Hiratsuka, A.; Matsushima, T. J. Phys. Chem.
B 2004, 108, 3828.
Acknowledgment. We thank Professor I. Kobal (J. Stefan
Institute) for his stimulating discussion. This work was partly
supported by a 1996 COE special equipment program of the
Ministry of Education, Sports, and Culture of Japan. 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.
(32) Shustorovich, E.; Sellers, E. Surf. Sci. Rep. 1998, 31, 1.
(33) Wang, H.; Tobin, R. G.; DiMaggio, C. L.; Fisher, G. B.; Lambert,
D. K. J. Chem. Phys. 1997, 107, 9569.
References and Notes
(1) Paˆrvulescu, V. I.; Grange, P.; Delmon, B. Catal. Today 1998, 46,
233.
(34) Comsa, G.; David, R. Surf. Sci. Rep. 1985, 5, 145.