Dynamics of NO produced by photodissociation of t-butyl nitrite on Ag(111)

Article abstract of DOI:10.1016/S0009-2614(99)00418-2

The dynamics of NO ejected by the 248 nm photodissociation of t-butyl nitrite, (CH₃)₃CONO, adsorbed on Ag(111) has been examined. A trimodal NO velocity distribution is observed. The velocity distribution varies with the detection angle and coverage and is consistent with a model in which direct excitation of t-butyl nitrite is followed by prompt dissociation to form nitric oxide which desorbs along one of three distinguishable trajectories involving: (1) no intervening capture or collisions, (2) one or a few intervening collisions without complete thermalization, and (3) equilibration with the substrate temperature.

Full text of DOI:10.1016/S0009-2614(99)00418-2

4 June 1999
Ž
.
Chemical Physics Letters 306 1999 103–110
Dynamics of NO produced by photodissociation of t-butyl nitrite
ž
/
on Ag 111
1
)
Jeffrey E. Fieberg , J.M. White
Center for Materials Chemistry, Department of Chemistry and Biochemistry, UniÕersity of Texas, Austin, TX 78712, USA
Received 5 November 1998; in final form 23 March 1999
Abstract
Ž
.
Ž
.
The dynamics of NO ejected by the 248 nm photodissociation of t-butyl nitrite, CH3 3CONO, adsorbed on Ag 111 has
been examined. A trimodal NO velocity distribution is observed. The velocity distribution varies with the detection angle
and coverage and is consistent with a model in which direct excitation of t-butyl nitrite is followed by prompt dissociation to
Ž .
form nitric oxide which desorbs along one of three distinguishable trajectories involving: 1 no intervening capture or
Ž .
Ž .
collisions, 2 one or a few intervening collisions without complete thermalization, and 3 equilibration with the substrate
temperature. q 1999 Elsevier Science B.V. All rights reserved.
1. Introduction
metal creating hot holes and electrons which interact
Ž
with the adsorbate and lead to intra-adsorbate e.g.,
.
carbon–halogen bond breaking. While photon ab-
Over the past two decades, non-thermal reactions
of adsorbates on metals have been widely studied.
Both photons and electrons have been utilized to
initiate these processes and there is also evidence for
non-thermal events initiated thermally 1–3 . On
metals, such chemical reactions, e.g., photodissocia-
tion of alkyl halides, often involve important contri-
butions from substrate-mediated processes in which,
for example, the incident photon is absorbed in the
sorption by the adsorbate may contribute, seldom is
there a situation where this so-called direct photodis-
sociation dominates on metals. An alkyl nitrite ad-
Ž
.
w
x
sorbed on Ag 111 is such a case 4–6 .
w
x
Comparisons with gas-phase photo properties are
of interest, especially when examining the influence
of the surface environment on the dynamics of prod-
uct ejection. The gas-phase UV absorption spectrum
of alkyl nitrites involves two excited states, typically
w
x
Ž
denoted S1 and S2 7–9 . Absorption into S1 l)300
.
Ž
nm exhibits vibrational structure while S2 l-300
.
nm is unstructured. The chemistry ensuing after
excitation is simple and the same for both S1 and S2
– the internal N–O bond cleaves to form alkoxy and
NO. As expected, the dynamics for dissociation via
S1 and S2 differ, reflecting differences in the excited
)
Corresponding author. Fax: q1 512 471 9495; e-mail:
jmwhite@mail.utexas.edu
1 Current address: Department of Chemistry, Hillsdale College,
Hillsdale, MI 49242, USA.
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00418-2

(
)
104
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
Ž
state potential energy surfaces structured and un-
2. Experiment
The experimental apparatus dynamics ultra-high
.
structured absorption spectra and the higher photon
energies required to reach S2. Significantly, nitrite
dissociation via S2 occurs in ;10 fs 7,9 .
The photochemistry of alkyl nitrites adsorbed on
Ag 111 4–6,10 and MgF2 11,12 has been re-
Ž
w
x
Ž
.
.
vacuum UHV chamber and excimer laser have
w x
been described previously 6 . Briefly, the experi-
Ž
. w
x
w
x
ments were done in a two-level UHV chamber which
operates at a base pressure of -2=10y10 Torr. An
ion gun, reverse-view low-energy electron diffraction
Ž
.
ported. On Ag 111 , the substrate of interest here,
NO is ejected and the accompanying alkoxy frag-
ment is retained. For both methyl nitrite, CH3ONO,
and t-butyl nitrite, CH3 3CONO, the NO velocity
distributions are made up of strong thermal and
hyperthermal contributions. The wavelength and
coverage variations of the NO velocity distributions
are consistent with direct excitation as the dominant
excitation mechanism.
For CH3ONO on Ag 111 , a trimodal NO veloc-
ity distribution, dependent on the polar detection
angle, is required to describe the 248 nm photolysis
of CH3ONO at all coverages from 0.15 to 84 ML
Ž
.
LEED apparatus, and Auger electron spectrometer
Ž
.
comprise the upper level. Photolysis is done on the
lower level that has windows for passing pulsed
excimer laser light and a rotatable and differentially
pumped housing that contains a quadrupole mass
Ž
.
Ž
.
spectrometer QMS for time-of-flight TOF experi-
Ž
.
ments. The QMS axis is perpendicular to the Ag 111
Ž
.
surface normal.
Standard substrate mounting and cleaning proce-
Ž
.
dures were followed. The Ag 111 substrate was
cooled to as low as 89 K by contact with a liquid
nitrogen reservoir. Heating was accomplished resis-
tively and was limited to 800 K in order to avoid Ag
evaporation. The crystal was cleaned by sputtering
with 2.5 keV argon ions and annealing to 700 K for
w
x
6,13 . All three components are attributed to direct
photodissociation; the fastest component is assigned
to NO ejected without intervening collisions, the
intermediate velocity component to NO released af-
ter one or a few collisions with surrounding species,
and the slowest component to thermally accommo-
dated NO.
The dynamics of NO ejected near the surface
normal during irradiation of t-butyl nitrite, CH3
CONO, adsorbed on Ag 111 has been reported in an
elegant study by Jenniskens et al. 5,14 . For wave-
lengths between 266 and 532 nm, direct excitation
dominates for all coverages – sub-monolayers to
thick multilayers 50 ML, where ML denotes mono-
layer – and bimodal NO velocity distributions ade-
Ž
.
17 min. A 1=1 LEED pattern was observed over
Ž
.
the entire crystal. The t-butyl nitrite Aldrich was
freeze–pump–thawed several times before introduc-
tion into the UHV chamber. Dosing was performed
through a leak valve connected to a 3 mm ID tube
Ž
.
-
3
Y
Ž
.
that was directed at and displaced ;1 from the
w
x
Ž
.
Ag 111 crystal face.
In our experiments, a saturated monolayer was
formed by dosing with the substrate held at 146 K,
which is high enough to prevent multilayer forma-
Ž
.
Ž
.
tion. Temperature-programmed desorption TPD
showed monolayer desorption at 162 K, in excellent
quately describe the results. Separately, reflection
Ž
. w
x
absorption infrared data taken before and after irradi-
agreement with other work 160 K 5,14 . A cover-
age of 27 ML was produced when the substrate was
held at 100 K and the chamber pressure was raised
to 8=10y9 Torr for 5 min. The multilayer TPD
feature peaked at 141 K, consistent with earlier work
Ž
.
ation of CH3 3CONO multilayers shows the conver-
Ž
.
Ž
.
sion of trans- CH3 3CONO to cis- CH3 3CONO.
The latter evidences the importance of caging in
Ž
.
CH3 3CONO, an effect which inhibits separation
Ž
.
w
x
and promotes reassembly of NO and CH3 3CO.
5,14 .
Ž
.
In this Letter, we extend the work of Jenniskens
Ultraviolet 248 nm, 15 ns pulses from an ex-
cimer laser were incident 458 from the Ag 111
surface normal. The beam was apertured to provide
w x
Ž
.
et al. 5 and report angular distributions of NO for
248 nm photolysis of monolayer and multilayer
Ž
.
Ž
.
Ž
CH3 3CONO on Ag 111 . As for CH3ONO, a tri-
uniform irradiation 0.1–1.0 mJrpulse at the sur-
2
.
Ž
.
modal distribution is required, particularly off-nor-
mal, to account for the appearance of a very fast NO
component.
face over the central 80% ;0.9 cm of the
Ag 111 surface. Ejected neutral NO molecules were
ionized, extracted perpendicular to the flight path,
Ž
.

(
)
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
105
Ž
.
counted versus time 1.28 ms intervals , and aver-
aged over many laser pulses. Subsequently, the ion
flight time in the QMS was subtracted, and the
resulting density domain data were fit to n modified
Maxwell–Boltzmann distributions:
3. Results
Ž
.
For monolayer CH3 3CONO, the density do-
main NO TOF spectra, taken at 08 and 408 with
respect to the Ag 111 surface normal, are shown in
Ž
.
2
n
the left-hand panel of Fig. 1. Similar spectra for 27
d
I t sty4
Ž .
a exp yb
yÕo,i
1
Ž .
Ý
Ž
.
i
i
ML thick multilayer of CH3 3CONO are shown in
the right-hand panel of Fig. 1. While each of these
four curves have three characteristics in common –
an onset time near 50 ms, one or two local maxima,
and a long decay – there are significant differences.
Consider the 1 ML case. First, the onset of the NO
signal is much steeper at 408 than at 08. Second,
ž t
/
i
s1
Ž .
where I t is the total signal from n distributions, t
is the flight time, ai is the amplitude of the ith
component, bi smr2kBTi where Ti is the transla-
tional energy of the ith component , d is the dis-
tance from the sample to the QMS ionizer, and Õo,i
is the offset velocity for the ith component. The
fitting procedure involved least-squares fits with cer-
tain constraints, e.g., excellent fits to the leading
edges of TOF spectra.
Ž
.
Ž
there are two local maxima arrows A and B in Fig.
.
1 in the spectrum taken at 408, whereas at 08, there
Ž
.
is a weak, but reproducible shoulder arrow C near
the onset that is followed by a strong maximum
Ž
.
The angle-resolved TOF spectra were recorded by
arrow D . Third, beyond 200 ms, there is a rela-
Ž
rotating the QMS in the plane defined by the inci-
tively flat region to 400 ms, beyond which the NO
TOF signals decay smoothly to 1350 ms, only a
portion of which is shown in Fig. 1. Finally, the total
integrated signal at 08 exceeds that at 408 by a factor
.
dent and reflected laser light around the sample. The
angular resolution of the TOF mass spectrometer is
"2.08.
Ž
.
Ž
.
Ž
.
Fig. 1. TOF distributions raw data for NO photoejected from CH3 3CONO on Ag 111 . The left-hand panel displays spectra for an initial
coverage of 1 ML. The detector is 408 off-normal for the upper spectrum and is normal to the surface for the lower spectrum. Arrows A and
Ž
.
Ž
.
C correspond to the fast1 channel f1 and arrows B and D correspond to the fast2 channel f2 . The right-hand panel displays spectra for
an initial coverage of 27 ML. The upper spectrum is taken 408 off the surface normal and the lower spectrum is taken with the detector
normal to the surface. Arrow E labels f1 and arrows F and G label f2.

(
)
106
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
Ž
.
Ž
.
Fig. 2. TOF distributions for NO photoejected from CH3 3CONO on Ag 111 . The raw data from Fig. 1 is reproduced with trimodal fits to
the data. The left panel again displays spectra for 1 ML coverage and the right panel displays spectra for 27 ML coverage. The narrow lines
Ž .
Ž .
are the fits to Eq. 1 as described in the text and the bold line sums the three narrow lines. The fast1 channel f1 is emphasized by the
hatched areas.
of 2.3. Turning to the 27 ML case, the general
characteristics remain the same with the following
exceptions – a sharp onset is evident for both the 08
times 0200 ms, particularly to account for the
broad shoulder that is readily evident in the 1 ML
spectra and, upon more detailed inspection is also
present in the 27 ML spectra.
Ž
.
and 408 spectra and the first maximum arrow E is
much more evident at 408 than for the 408 monolayer
case. The maxima at F and G correspond to B and D
in the monolayer spectra.
The characteristic energies for the three channels
Ž
.
generally agree between the four spectra Table 1 .
Due to considerable overlap between the first two
channels for the 27 ML, 08 spectrum, the uncertainty
Ž
Adequate fitting of these four spectra and others
Ž .
.
not shown requires the use of n03 in Eq. 1 .
Ž
.
Ž .
Restricted to ns3, denoted fast1 f1 , fast2 f2
and slow s channels, the resulting fits are shown in
Fig. 2. The individual components are shown as thin
lines and their sum as the bold line. The fit parame-
Ž .
Table 1
NO dynamics: relative populations and energies
Coverage Angle
Component Pop. Etrans Etrans r2k
Ž
.
Ž .
%
Ž
eV
.
Ž .
K
degrees
Ž
.
ters populations, characteristic energies are listed in
Table 1 and illustrate several points. A very fast
contribution, f1, is required to fit the 50 ms interval
27 ML
1.0 ML
0
Fast1, f1
Fast2, f2
Slow, s
Fast1, f1
Fast2, f2
Slow, s
Fast1, f1
Fast2, f2
Slow, s
Fast1, f1
Fast2, f2
Slow, s
20
60
20
31
47
22
2
70
28
17
57
26
0.84 4900
0.30 1800
0.02
130
Ž
.
following the onset hatched areas in Fig. 2 . While
there is considerable uncertainty in the hatched area
for the 1 ML, 08 case, including it is necessary to fit
40
0
0.92 5400
0.24 1400
0.02
120
Ž
.
the shoulder at arrow C Fig. 1 . A second contribu-
tion, f2 , is necessary to fit the B and D regions of
the 1 ML case and the corresponding regions in the
0.92 5400
0.24 1400
0.02
120
40
0.92 5400
0.28 1600
Ž
.
NO spectra taken for multilayers F and G . In
addition to these two contributions, a third, slower
contribution, s, is needed to capture the intensity for
0.02
140

(
)
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
107
in the fitting parameters for these two peaks is higher
and does not agree as well with the other three
spectra. The f1 channel energies for the other three
NO velocity distributions exhibit resolvable fast and
slow components. Since these channels are present
Ž
.
for photolysis of CH3 3CONO that is either in
Ž
Ž
.
spectra agree extremely well all are 0.92"0.08
contact with or displaced far from Ag 111 , we
concur that direct photolysis dominates and that dis-
sociation via substrate-mediated processes is mini-
mal. Independent of coverage, the characteristic tem-
peratures of the slow channels, 125"10 K at 248
.
Ž
eV , the f2 energies are nearly the same 0.27"0.03
eV , and the s channel energies are all 0.020"0.005
.
eV. As discussed below, it is also significant to note
that the characteristic energy of f2 is 0.28"0.03 of
f1.
w
x
nm and 110"15 K at 266 nm 14 , are in quantita-
tive agreement and are interpreted in terms of NO
formed by direct photolysis and thermalized to the
substrate temperature by the Ag or, in the case of
Table 1 also indicates the percentage contribu-
tions of each of the three components normalized to
the total intensity of each TOF curve. The slow
channel shows a coverage dependence; it contributes
21% for the 27 ML case and 27% for the 1 ML case.
While there is minimal change in the angular depen-
dence of the relative population of the slow channel,
the absolute population of the slow channel de-
creases monotonically with increasing detector angle
Ž
.
multilayers, by the CH3 3CONO film. A slow chan-
nel is present in all the alkyl nitrite systems studied
w
x
to date 4–6,10–14 and dominates the TOF spectra
w
x
for 351 nm photolysis of CH3ONO 6,13 . There-
fore, the slow channel is not entirely attributable to
the pumping characteristics of the vacuum systems.
It is informative to compare previous work at 266
nm with our work at 248 nm. The characteristic
energy of the hyperthermal channel is larger at 248
nm since direct excitation into S2 at 248 nm has 0.33
eV more available energy. Since the 266 nm data
was fit with one hyperthermal channel, we combined
2Ž .
away from the surface normal, similar to the cos u
dependence of the slow channel observed for photol-
Ž
. w x
ysis of CH3ONO on Ag 111 6 . Removing the
slow channel leaves 79% and 73% of the total,
respectively, to be divided between the two fast
channels. There is a strong angular dependence in
the associated percentages. For the 1 ML case, the f1
contribution to the total increases by a factor of 8 in
Ž
.
not shown the two hyperthermal channels of Fig. 2
into a single distribution for better comparison. When
described in this way, the characteristic hyperthermal
energy at 248 nm increases with coverage. For 266
nm photolysis, the characteristic hyperthermal en-
ergy varies from 0.17 to 0.24 eV in passing from 1
Ž
.
Ž
.
passing from 08 2% to 408 17% . For the 27 ML
data, the increase is a factor of 1.6. Finally, inspec-
tion of the 27 ML curve fits in Fig. 2 indicates that
the integrated intensity of f1 at 408 is approximately
70% of that at 08.
w
x
to 10 ML 14 . When Fig. 1 is fit with a single
hyperthermal channel, i.e., the sum of f1 and f2 in
Fig. 2, the variation is from 0.25 to 0.35 eV in
passing from 1 to 27 ML. While at this level of
analysis, the results at 266 and 248 nm are compati-
ble, our data at 248 nm indicate that the coverage
variation of the hyperthermal channel energy re-
4. Discussion
Fitting the NO TOF distributions as a function of
angle requires at least three modified Maxwell–
Boltzmann components, i.e., the distributions are
w
x
w
x
trimodal. In earlier work 5,10,14 , limited to detec-
tion near the surface normal and with 15 ms time
resolution compared to 1.28 ms in our work, bimodal
ported earlier 14 , is actually the result of changing
populations in two hyperthermal distributions, f1 and
f2. At 408, where no previous data has been reported,
our data at both 1 and 27 ML require a minimum of
two hyperthermal channels. When the 08 data is also
analyzed in this way, the coverage variations of the
characteristic energies are small while the relative
intensities assigned to f1 and f2 , on the other hand,
Ž
.
.
velocity distributions hyperthermal and thermal
provided satisfactory fits. It is the off-normal 408
dependence of the NO TOF distributions reported
here that clearly require multiple hyperthermal com-
ponents.
Ž
Ž
.
At 08, where comparisons can be made, our re-
vary strongly with coverage Table 1 .
sults using 248 nm photons compare favorably with
For both 08 and 408 detection angles, our data are
w
x
previous work using 266 nm radiation 14 . For both,
describable in terms of direct excitation to the S2

(
)
108
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
Ž
.
excited state of CH3 3CONO which dissociates in
;10 fs along a trajectory coincident with the axis of
the internal NO bond. The laboratory direction of the
velocity vector of the nascent NO will, thus, depend
monolayers, Ag atoms are among the nearest neigh-
Ž
.
bors and can align CH3 3CONO and can also alter
the trajectory dynamics of electronically excited
Ž
.
CH3 3CONO as it undergoes dissociation. More-
Ž
.
Ž
. w x
on the orientation of the CH3 3CONO from which it
is formed. The two hyperthermal channels are then
attributable as follows: f1 is ascribed to NO trajecto-
ries involving no collisions, while f2 is ascribed to
NO trajectories involving scattering, primarily off
CH3, groups before exiting into vacuum. In the 248
nm gas-phase monomer photolysis, the center-of-
mass translational energy carried by the fragment
over, the large dipole moment 3.4 D 19 , suggests
that interadsorbate interactions, i.e., dipole–dipole
forces, will be significant in determining the orienta-
Ž
.
-
3
tion and islanding behavior of monolayer CH3
Ž
.
CONO on Ag 111 .
The orientation of monolayer t-butyl nitrite on
Ž
.
Ag 111 has been examined using reflection–absorp-
Ž
. w
x
tion infrared spectroscopy RAIRS 5,14 . Strong
trans N5O and N–O stretches were observed which
requires an adsorption geometry in which neither of
Ž
.
w x
pair, NO and CH3 3CO, is 1.26 eV 15 . Assuming
a pseudo-diatomic model and linear momentum con-
Ž
.
Ž .
servation, 71% of this 0.89 eV would be carried by
NO. This compares favorably to the characteristic
energy of f1 which ranges from 0.84 to 1.0 eV, and
supports our assignment of f1 to NO ejected without
collision.
these two bonds lies parallel to the Ag 111 surface.
Ž
.
The proposed geometry Fig. 3 aligns the adsorbate
Ž
. w x
such that, as for CH3ONO on Ag 111 20 , the Cs
plane defined by the C–O–N5O lies nearly perpen-
dicular to the surface. In this alignment, excitation to
S2 will lead to NO motion in the direction indicated
by the arrows, i.e., off the surface normal. Some of
these dissociative trajectories may escape without
collision and contribute to the f1 intensity, while
others will undergo strong energy and momentum
transfer to methyl groups of neighboring t-butyl
nitrite. Some of these NO species will scatter to-
wards the surface normal and contribute to f2; others
may scatter towards the substrate and contribute to
f2 or, if thermalized, to s. The relative intensity of f2
increases toward the surface normal, consistent with
NO that has undergone non-thermalizing collisions
before desorbing. For multilayers, we hypothesize
The ratio of the characteristic energy of f2 to the
characteristic energy of f1 is 0.28 Table 1 . For
Ž
.
w
x
photolysis of adsorbed CH3ONO 6,13 , a similar
relation obtains and, while the energies of f1 and f2
both decrease with increasing excitation wavelength,
the ratio of the characteristic energies remains con-
stant. In a head-on pseudo-binary collision of NO
Ž
.
Ž
.
30 amu with CH3 15 amu , the final NO energy
would be 0.33 of the initial energy. While this
description is clearly oversimplified, it does indicate
that a model attributing f2 to NO trajectories domi-
nated by one to a few non-thermalizing collisions
with CH3 is reasonable.
Ž .
Considering multilayers, f1 is attributable to di-
the alignment of the CH3 3CONO molecules at and
Ž
.
Ž
.
rect photodissociation of CH3 3CONO located at
the vacuum–adsorbate interface and oriented so that
the NO group can escape without a collision, while
f2 involves NO formed at or just below the surface
and directed such that non-thermalizing collisions
occur before exiting. The NO distribution, f3, is
attributed to NO formed at or below the vacuum–ad-
sorbate interface that fully thermalizes through colli-
sions prior to ejection. A similar model has been
used to describe the rotational energy distributions of
near the vacuum– CH3 3CONO interface is more
random for one monolayer. Thus, for multilayers,
Ž
.
-
3
NO ejected from photolyzed clusters of CH3
w
x
CONO and CH3ONO 16–18 .
Ž
.
Fig. 3. Schematic of proposed orientation of CH3 3CONO on
The multilayer model can also be used to charac-
Ž
.
Ag 111 . It is hypothesized that t-butyl nitrite bonds to the
substrate via the lone-pair electrons of the oxygen atoms as
indicated by the dashed lines.
Ž
.
terize ejection of NO from monolayer CH3 3CONO.
However, differences are anticipated since, for

(
)
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
109
collisionless ejection of NO contributes strongly at
both 08 and 408.
For monolayer coverage, the relative strength of
f1 at 408 compared to 08 is consistent with the
the fast channel and is attributed to trajectories domi-
nated by one to a few collisions between NO and
neighboring CH3 groups. The slow channel, s, is
attributed to NO thermalized by multiple collisions
before desorption.
Ž
.
orientation proposed in Fig. 3, i.e., CH3 3CONO is
aligned with its internal O–N bond far away from
the surface normal and located with respect to other
Ž
.
CH3 3CONO such that it can escape without under-
going collisions. The weakness of the f1 channel at
Acknowledgements
Ž
.
08 indicates that few monolayer CH3 3CONO
molecules are adsorbed with the internal O–N bond
normal to the surface.
In this context, it is important to keep in mind the
low thermal desorption temperature and the small
shifts of vibrational frequencies from gas-phase val-
ues. Both provide a reminder that alignment is due to
relatively weak polarization forces between the ad-
sorbate and substrate. In turn, a broad distribution of
angles between the Cs plane and the surface is
Support by the National Science Foundation,
Grant No. CHE9319640, and by the Robert A. Welch
Foundation is gratefully acknowledged.
References
w x
1
Ž
.
W. Ho, Surface photochemistry, in: H.-L. Dai, W. Ho Eds. ,
Laser Spectroscopy and Photochemistry on Metal Surfaces,
Part II, World Scientific, Singapore, 1995.
Ž
.
expected for adsorbed CH3 3CONO and, as a result,
some NO will be ejected along a direction more
w x
2
J.C. Polanyi, Y. Zeiri, Dynamics of adsorbate photochem-
Ž
.
nearly parallel to or even toward the Ag 111 surface
Ž
.
istry, in: H.-L. Dai, W. Ho Eds. , Laser Spectroscopy and
Photochemistry on Metal Surfaces, Part II, World Scientific,
Singapore, 1995.
Ž
.
-
3
where the influence of Ag and neighboring CH3
CONO will lead to a residence time sufficiently long
to thermalize NO prior to desorption.
w x
3
X.L. Zhou, J.M. White, Photodissociation and photoreaction
of molecules attached to metal surfaces, in: H.-L. Dai, W. Ho
Ž
.
Eds. , Laser Spectroscopy and Photochemistry on Metal
Surfaces, Part II, World Scientific, Singapore, 1995.
w x
4
Ž
.
L.A. Pressley, E.D. Pylant, J.M. White, Surf. Sci. 367 1996
1.
5. Summary
w x
5
H.G. Jenniskens, L. Philippe, W. van Essenberg, M. Kadod-
Ž
.
wala, A.W. Kleyn, J. Chem. Phys. 108 1998 1688.
w x
6
J.E. Fieberg, G.J. Szulczewski, J.M. White, Chem. Phys.
The TOF distributions of NO ejected by the 248
Ž
.
Lett. 290 1998 268.
Ž
.
-
3
w x
nm photodissociation of t-butyl nitrite, CH3
7
J.M. Mestdagh, M. Berdah, I. Dimicoli, M. Mons, P. Mey-
nadier, P. d’Oliveria, F. Piuzzi, J.P. Visticot, C. Jouvet, C.
Lardeux-Dedonder, S. Martrenchard-Barra, B. Soep, D. Sol-
Ž
.
CONO, adsorbed on Ag 111 depend on angle of
detection and coverage. The velocity distributions
are adequately fit using three modified Maxwell–
Boltzmann distributions. These trimodal distributions
are consistent with a model in which direct excita-
tion of t-butyl nitrite into its S2 electronic state is
followed by prompt dissociation to form nitric oxide
directed along the axis of the breaking bond. This
NO desorbs with or without local collisions with
neighboring adsorbates andror the substrate. The
fastest NO, f1, has a characteristic energy close to
that found for photodissociation of gas-phase
Ž
.
gadi, J. Chem. Phys. 103 1995 1013.
w x
8
Ž
.
B.A. Keller, P. Felder, J.R. Huber, J. Phys. Chem. 91 1987
1114.
w x
9
R. Schinke, S. Hennig, A. Untch, M. Nonella, J.R. Huber, J.
Ž
.
Chem. Phys. 91 1989 2016.
w
w
w
w
w
x
10 H.G. Jenniskens, W. van Essenberg, M. Kadodwala, A.W.
Ž
.
Kleyn, Chem. Phys. Lett. 268 1997 7.
x
11 C.J.S.M. Simpson, P.T. Griffiths, H.L. Wallaart, M. Towrie,
Ž
.
Chem. Phys. Lett. 234 1995 203.
x
12 C.J.S.M. Simpson, P.T. Griffiths, H.L. Wallaart, M. Towrie,
Ž
.
Chem. Phys. Lett. 263 1996 19.
x
13 J.E. Fieberg, P.M. Blass, J.M. White, manuscript in prepara-
tion.
Ž
.
CH3 3CONO and is attributed to NO ejected with-
out intervening collisions. The intermediate NO, f2 ,
has an energy that, on the average, is ;0.28 that of
x
14 H.G. Jenniskens, Surface Photochemistry of Physisorbed
Molecules, Ph.D. Dissertation, FOM-Institute for Atomic and
Molecular Physics, 1997.

(
)
110
J.E. Fieberg, J.M. WhiterChemical Physics Letters 306 1999 103–110
w
w
w
x
w
x
Ž
.
.
15 C.S. Effenhauser, P. Felder, J.R. Huber, J. Phys. Chem. 94
18 E. Kades, M. Rosslein, J.R. Huber, J. Phys. Chem. 98 1994
Ž
.
1990 296.
13556.
x
w
x Ž .
19 R. Cotting, J.R. Huber, J. Chem. Phys. 104 1996 6208.
20 J.E. Fieberg, J.M. White, J. Vac. Sci. Technol. A 15 1997
1674.
16 E. Kades, M. Rosslein, U. Bruhlmann, J.R. Huber, J. Phys.
Ž
.
w
x
Ž
Chem. 97 1993 989.
x
Ž
.
17 K. Bergmann, J.R. Huber, J. Phys. Chem. A 101 1997 259.