The effect of translational energy on collision-induced dissociation of highly excited NO2 on MgO(100)

Article abstract of DOI:10.1016/S0009-2614(97)01277-3

Collision-induced dissociation of highly excited NO₂ (i.e., mixed ²B₂/²A₁ molecular eigenstates just below D₀) impinging on MgO(100) surfaces has been studied as a function of NO₂ internal excitation at an incident translational energy of 4400 cm^-1 by using state-selective NO detection. NO internal energy distributions as well as the average energy transferred per activating collision have been obtained. The results, in particular the NO [²Π_1/2]/[²Π_3/2] population ratios, indicate the presence of exit-channel interactions with the surface that are dependent on collision energy.

Full text of DOI:10.1016/S0009-2614(97)01277-3

16 January 1998
Ž
.
Chemical Physics Letters 282 1998 313–317
The effect of translational energy on collision-induced
ž
/
dissociation of highly excited NO2 on MgO 100
1
D.W. Arnold , M. Korolik, C. Wittig, H. Reisler
Department of Chemistry, UniÕersity of Southern California, Los Angeles, CA 90089-0482, USA
Received 8 September 1997
Abstract
2
Collision-induced dissociation of highly excited NO2 i.e., mixed B2r 2A1 molecular eigenstates just below D0
Ž
.
Ž
.
impinging on MgO 100 surfaces has been studied as a function of NO2 internal excitation at an incident translational
energy of 4400 cmy by using state-selective NO detection. NO internal energy distributions as well as the average energy
1
transferred per activating collision have been obtained. The results, in particular the NO 2P1
P3r2 population ratios,
2
w
x w
r
x
r2
indicate the presence of exit-channel interactions with the surface that are dependent on collision energy. q 1998 Elsevier
Science B.V.
1. Introduction
can be varied independently, the former by changing
the supersonic expansion conditions and the latter by
varying the excitation laser wavelength. Such experi-
ments shed light on the survival probability of highly
excited molecules near surfaces, and they are rele-
vant to reverse processes, e.g., can products of sur-
face reactions leave the surface in excited states?
Experimental studies of collision-induced dissoci-
Ž
.
ation CID at well-defined hyperthermal energies
and with state-selective product detection have pro-
vided insight into energy transfer mechanisms and
w
x
their dependence on environment 1–7 . A recent
refinement — the optical preparation of molecules
with selected amounts of internal excitation at speci-
fied hyperthermal energies — has enabled studies of
CID to be carried out at internal energies near the
dissociation threshold D0 , in which case only a
small amount of energy needs to be converted from
translation to vibration T™V to effect CID 1–4 .
In these studies, both the incident translational en-
ergy, Einc , and the internal energy of the molecules
Ž
The CID of internally excited NO2 hereafter
denoted NO2) has been studied previously in the gas
.
w
x
w x
phase 2–4 and with crystalline surfaces 1 . In
those experiments, hyperthermal NO2 in a molecular
2
beam was excited to mixed B2r 2A1 molecular
Ž
.
eigenstates near D0 by using laser excitation, and the
NO product was detected state-selectively. The gas-
phase results have led to the suggestion that CID
involves two largely independent steps: collisional
activation followed by unimolecular decomposition.
During the decomposition step, the gaseous ‘col-
Ž
.
w
x
1
Ž
.
lider’ i.e., the species colliding with NO2 is at a
Present address: Sandia National Laboratory, Mail Stop 9671,
Ž
.
large enough distance on average to minimize sec-
Livermore, CA 94551-0969, USA.
0009-2614r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.
Ž
.
PII S0009-2614 97 01277-3

(
)
314
D.W. Arnold et al.rChemical Physics Letters 282 1998 313–317
ondary interactions with the excited NO2 or its
fragments.
the surface and the resulting CID products to reach
the detection region. The relatively long fluorescence
Studies of NO2) colliding with MgO 100 sur-
faces at Einc s1940 cmy1 have shown subtle differ-
ences with gas-phase CID, especially in the internal
lifetime of NO2) ;50 ms 8 ensures that a signifi-
cant fraction of the molecules remains in the excited
state upon arrival at the surface. NO fragments are
detected state-selectively 3–5 mm from the surface
Ž
.
Ž
. w x
w x
excitations of the NO fragment 1 . In this Letter, we
report additional studies of the CID of NO2) on
2
q
Ž
.
by 1q1 226 nmq280 nm REMPI via the A S
2
Ž
.
w
x
MgO 100 aimed at understanding the molecule–
surface CID mechanism and how it differs from the
gas-phase CID mechanism. We argue that in the
molecule–surface case the collisionally activated
NO2 andror its fragments experience significant
§ X P transition as described before 5–7 . The
probe beams are produced by a dye laser whose
Ž
.
doubled output ;280 nm is Raman shifted to
;226 nm. After ionization, NOq is detected with a
bare channeltron. The probe beams propagate paral-
lel to the surface, and angular distributions are ob-
tained by measuring the NO signal along an arc
around the collision center.
Ž
.
Ž
interactions with the MgO 100 surface which de-
.
pend on Einc , and that this is the main source of the
observed differences.
Several competing processes and background sig-
nals must be accounted for in the data analysis, the
most important arising from the background NO in
2. Experimental details
w x
the NO2 beam 3 . Expansion-cooling ensures that
Ž2
.
only the lowest NO P1r2 rotational levels have
significant populations; nonetheless, collisions excite
background NO from low J into higher rotational
The experiments were carried out in a UHV
apparatus that has been described in detail elsewhere
w
x
w x
1,5–7 . Pulsed supersonic expansion of an NO2rO2
states 9 , necessitating corrections to the monitored
Ž
w
x w
x
.
mixture with NO2 r O2 ;1 seeded in He pro-
duces a molecular beam of cold NO2 whose transla-
tional energy is controlled by changing the seed ratio
of the gas mixture. In the present work Einc s4400
cmy1 was employed, and the incident angle was
Qinc s158. The rotational temperature of NO2 in the
NO signal. Background signals of different origins
are singled out by operating the pump laser in onroff
sequences, and they are subtracted from the total
signal for each data acquisition cycle, thereby mini-
mizing errors due to long-term drifts. Data process-
ing also includes normalization by the pump and
probe pulse energies.
Ž
beam estimated from the rotational temperature of
the background NO P1r2 which is always present
in NO2 samples is F10 K. The gas pulse 150–200
ms is skimmed and collimated before entering the
UHV chamber ; 10
occur with an ex situ cleaved MgO 100 surface. The
Ž2
.
.
Ž
.
y10
3. Results
Ž
.
Torr , where collisions
Ž
.
crystal is treated by heating in O2 as described
Key to the success of these experiments is that
CID occurs only with NO2) ; ground-state NO2 can-
not acquire sufficient energy via the collision to
dissociate. Thus, CID occurs only at excitation wave-
lengths that coincide with NO2 absorption features,
as illustrated in Fig. 1. Fig. 1a displays the laser-in-
w x
previously 1 , and its quality is checked by He
diffraction and Auger spectroscopy. A surface tem-
perature of 300 K is maintained in these experi-
ments.
Jet-cooled NO2 is excited into the mixed 2 B2r 2A1
manifold of eigenstates at ;400 nm, by using the
Ž
.
duced fluorescence LIF spectrum of NO2 near D0;
the spectrum is congested, but exhibits distinct, nar-
row bands. In Fig. 1b, the corresponding CID yield
Ž
output of an excimer laser pumped dye laser 0.4
cmy1 linewidth, 3–4 mJ . The laser beam, colli-
.
Ž2
spectrum, obtained by monitoring NO P1r2 , Js
mated to a 2 mm width along the molecular beam
axis with a cylindrical lens, excites the molecules
;20 mm from the MgO surface. A 14–20 ms
pump–probe delay allows the NO2) to collide with
.
5.5 and corrected to account for spurious signals as
w
x
described before 1,2 , is shown. The upper scale
D0 yhn indicates the T™V NO2 energy re-
Ž
.
Ž
.

(
)
D.W. Arnold et al.rChemical Physics Letters 282 1998 313–317
315
reveals that even the finer features in the LIF spectra
are reproduced in the CID yield spectrum. Also, as
w
x
in previous cases 3,4 , the CID yield signal de-
creases rapidly as the required T™V energy in-
creases.
A measure of the average energy transferred per
²
:
activating collision, D E , can be obtained from the
semilog plot of the point-by-point ratio of the CID
yield spectrum and the LIF spectrum. The higher
Ž
.
T™V NO2 efficiency expected for the higher Einc
of the present experiment, combined with higher
²
:
detection sensitivity, enable D E to be obtained for
2
the first time in gas-surface CID of NO2). The
Ž
.
Ž
.
linear dependence on average of ln ICIDrILIF on
excitation energy implies that the fraction of NO2)
molecules that undergoes CID decreases exponen-
tially with increasing energy deficiency according to:
²
:
.
ICID AILIF exp y D yhn r D E
1
Ž .
Ž
.
0
Such behavior is common in gas-phase energy trans-
fer — the phenomenological exponential gap law
Ž .
Ž2
NO P1r2
Fig. 1. Jet-cooled NO2 LIF excitation spectrum.
b
;
y1
w
x
10 . A linear least-squares fit in the range 17 cm
)
.
Õs0; Js5.5 CID yield as a function of NO2 excitation energy
y1
Ž
.
² :
F D0 yhn F240 cm
gives D E s115"35
hn at Einc s4400 cmy1. c Semilog plot of the point-by-point
Ž
.
Ž .
cmy1
.
As with Einc s1940 cmy1, the CID yield
3
ratio of the CID and LIF spectra. The top axis indicates energy
below D0.
signal peaks near the specular scattering angle.
NO rotational and spin–orbit distributions are ob-
tained by setting the excitation laser frequency at a
prominent absorption peak and scanning the probe
laser. Fig. 2 displays the rotational distributions ob-
tained by using the Q11 qP21 and Q22 qR12 lines
originating from the 2 P1r2 and 2 P3r2 states, respec-
quired to effect dissociation. The correspondence
between the peak positions in the two spectra is
evident. Closer inspection, especially in the region of
small D0 yhn, where the CID signal is largest,
tively, and obtained at D0 yhns42 cmy1 hns
Ž
.
398.57 nm . Similar distributions are obtained for
other rotational lines within each spin–orbit state,
and also at excitation energies 137 and 157 cmy1
below D0 . Since the distributions appear
Boltzmann-like, they are characterized by parameters
T
obtained from the slopes of semilog plots such
rot
as those shown in Fig. 2; the kT values are 190"
rot
2
40 cmy1 and 290"60 cmy1 for the P1r2 and
2
The fraction of the excited molecules whose fluorescence is
detected does not vary significantly throughout the narrow excita-
tion region employed here.
2
2
Ž
.
Fig. 2. Boltzmann plots of NO Õs0 P1r2 and P3r2 rota-
tional populations measured at a fixed NO2 excitation energy of
D0 yhn s41 cmy1. Open and filled circles display the popula-
tions of the Q11 qP21 and Q22 qR12 lines, respectively.
3
We exclude the region D0 yhn s0–17 cmy1 from the fit,
as it is dominated by photodissociation of rotationally excited
w
x
NO2 in the beam 3,4 .

(
)
316
D.W. Arnold et al.rChemical Physics Letters 282 1998 313–317
2 P3r2 states, respectively. These values are compa-
spin–orbit population ratios are different than in both
gas-phase CID and collisionless unimolecular de-
composition, and depend on Einc. These differences
suggest that product state distributions are influenced
by exit channel interactions between the decompos-
ing NO2 andror its fragments with the surface.
From the many studies of photoinitiated uni-
molecular decomposition of gaseous NO2 , a robust
rable to, or slightly lower than, those obtained in
scattering from MgO 100 at 1940 cmy1 1 and
Ž
.
w x
w
x
with gas-phase colliders 3,4 .
By summing the populations of all rotational lev-
els in each NO spin–orbit state, the ratio
2
2
w
x w
x
P1r2 r P3r2 ;0.4"0.05 is obtained. This
value lies between the unity ratio obtained in
molecule–surface CID at Einc s1940 cmy1 1 and
signature has emerged: the P3r2 r P1r2 ratio of
2
2
w x
w
x w
x
the ratios obtained in the gas-phase experiments
the NO product is always smaller than the statistical
w
x
w x
3,4 . It is also slightly higher than the ratio of
approximately 0.3 obtained in the unimolecular de-
value of unity 11,12 . This propensity is also present
Ž
in gas phase CID with small collision partners e.g.,
composition of NO2 at excess energy ;2000 cmy1
Ar 3 , and is in qualitative agreement with the
. w x
w
x
11 .
picture of collisional activation followed by uni-
molecular decomposition. However, when CID devi-
ates from the separable, two-step process described
above, significant interactions between the collision
partner and the CID products can affect energy
partitioning, resulting, for example in more ‘equi-
4. Discussion
w
x
The present results, taken together with the gas-
librated’ NO spin–orbit populations 3,4 .
w
x
phase CID studies 2–4 and the molecule–surface
w x
Exit-channel interactions are more complex in the
molecule–surface CID of NO2) than in gas-phase
CID. Molecule–surface binding energies of the par-
CID results obtained at Einc s1940 cmy1 1 , enable
Ž
.
us to obtain a more complete albeit still qualitative
Ž
.
picture of the reaction mechanism. It is therefore
useful to summarize the previous studies.
ent and its fragments are significant, and energy
transfer to the surface may be substantial. In addi-
tion, molecule–surface forces are likely to be stronger
and longer-range for highly excited molecules than
for unexcited molecules. Thus, one cannot assume a
priori that decomposition following collision with
MgO proceeds in a way that parallels gas phase CID.
Measurements at different collision energies reveal
the influence of molecule–surface exit channel inter-
actions on the state distributions of the NO product.
It is probable that some fraction of the NO will
interact strongly with the surface, since its binding
Gas-phase CID can be described as a two-step
Ž .
process: 1 collisional activation creates a distribu-
Ž .
tion of excited NO2 levels above D0 and 2 subse-
quent unimolecular decomposition yields NOqO.
The collider plays mostly a spectator role, and a
statistical model, based on the assumption that the
fractional population of NO2 above D0 decreases
exponentially with increasing energy, reproduces the
w x
experimental observations rather well 3 . Indications
of minor exit-channel interactions with the collider
are revealed mainly through the dependence of the
NO spin–orbit ratio on the nature of the collider.
The average energy transferred per collision depends
on the stiffness of the collision partner, with Ar
w
x
energy is ;0.2 eV 14 . For example, suppose some
of the post-collision NO2 molecules rebound from
˚
the surface with speeds ;6 Arps, and recall that
the NO2 collision-free unimolecular decomposition
rate several hundred cmy1 above D0 is ;1012 sy1
²
:
exhibiting the largest D E values, while poly-
w
x
13 . In this case, some decomposition occurs before
the parent and fragments are free from the surface.
Moreover, studies of expansion-cooled NO scattering
Ž
.
The molecule–surface results reveal some intrigu-
from M gO 100
have shown that the
2
2
Ž .
w
x w
x
ing differences: i the amount of energy transferred
P3r2 r P1r2 ratio for the scattered NO is near
Ž .
w
x
per collision is smaller than in the gas phase; ii
rotational excitation does not increase and may even
the statistical value of unity 15 . This is similar to
the value obtained in NO2) CID on MgO 100 at
Ž
Ž
.
Einc s1940 cmy1, but in contrast to the lower value
.
Ž .
decrease with increasing Einc; and iii the NO

(
)
D.W. Arnold et al.rChemical Physics Letters 282 1998 313–317
317
Ž
.
0.3 obtained in photoinitiated NO2 unimolecular
before they can free themselves from the environs of
the surface.
w
x
decomposition 11,12 .
The observed product state distributions reported
here for the higher Einc value of 4400 cmy1 is
consistent with this picture. The similarity of the
Acknowledgements
Ž
.
spin–orbit ratio 0.4 to the unimolecular decomposi-
This research was supported by the US Air Force
Office of Scientific Research. The authors thank K.
Mikhaylichenko for obtaining the NO2 LIF spectrum
and Andrei Sanov for stimulating discussions.
Ž
.
tion value 0.3 suggests that at this higher collision
energy the dissociating NO2 molecules escape from
the surface prior to decomposition. Clearly the extent
of secondary interactions with the surface depends
on Einc. It should be kept in mind that the present
experiments detect only those NO2) molecules whose
References
Ž
.
internal energy exceeds D0 following T™V NO2
w x
1
H. Ferkel, J.T. Singleton, H. Reisler, C. Wittig, Chem. Phys.
and subsequent interactions with the surface. Given
the smallness of the signals, this fraction is expected
to be modest. Note that the peaking of the NO signal
near the specular angle means that the dissociating
molecules retain some memory of the momentum of
the incident NO2) ; i.e., trapping–desorption, if it
occurs at all, is a minor channel.
Ž
.
Lett. 221 1994 447.
w x
2
C.R. Bieler, A. Sanov, M. Hunter, H. Reisler, J. Phys. Chem.
Ž
.
98 1994 1058.
w x
3
Ž
.
A. Sanov, C.R. Bieler, H. Reisler, J. Phys. Chem. 99 1995
7339.
w x
4
C.R. Bieler, A. Sanov, C. Capellos, H. Reisler, J. Phys.
Ž
.
Chem. 100 1996 3882.
w x
5
E. Kolodney, D. Baugh, P.S. Powers, H. Reisler, C. Wittig,
The value of D E obtained at Einc s4400 cmy1
²
:
Ž
.
J. Chem. Phys. 90 1989 3883.
w x
6
is lower than the values obtained with gas phase
E. Kolodney, P.S. Powers, L. Hodgson, H. Reisler, C. Wit-
Ž
.
tig, J. Chem. Phys. 94 1991 2330.
colliders at about half this energy. For example,
D E values range from 135 cm for NH3 to 307
w x
7
y1
P.S. Powers, E. Kolodney, L. Hodgson, G. Ziegler, H.
²
:
Ž
.
Reisler, C. Wittig, J. Phys. Chem. 95 1991 8387.
cmy1 for Ar 3,4 , pointing again to the significant
role played by energy transfer to the surface in the
molecule–surface collisions. In contrast to gas-phase
CID, where the NO internal excitation increases with
Einc , not only are the spin–orbit populations colder
at higher Einc , but the rotational excitation of the NO
product does not increase, and perhaps even de-
creases slightly, at higher Einc. This again reflects
the subtle influence of interactions with the surface
and their dependence on Einc. Thus, all of the evi-
dence collected to date is consistent with a dissocia-
tion mechanism in which NO2) gains energy in the
collision and some of these molecules decompose
w
x
w x
8
K.O. Patten Jr., J.D. Burley, H.S. Johnston, J. Phys. Chem.
Ž
.
94 1990 7960.
w x
9
C.R. Bieler, A. Sanov, H. Reisler, Chem. Phys. Lett. 235
Ž
.
1995 175.
w
w
x
Ž
.
10 J. Troe, J. Chem. Phys. 66 1977 4745.
x
11 M. Hunter, S.A. Reid, D.C. Robie, H. Reisler, J. Chem.
Ž
.
Phys. 99 1993 1093.
w
w
w
w
x
12 S.A. Reid, D.C. Robie, H. Reisler, J. Chem. Phys. 100
Ž
.
1994 4256.
x
13 S.I. Ionov, G.A. Brucker, C. Jaques, Y. Chen, C. Wittig, J.
Ž
.
Chem. Phys. 99 1993 3420.
x
14 H. Ferkel, L. Hodgson, J. Singleton, P. Blass, H. Reisler, C.
Ž
.
Wittig, J. Chem. Phys. 100 1994 9228.
x
15 E. Kolodney, D. Baugh, P.S. Powers, H. Reisler, C. Wittig,
Ž
.
Chem. Phys. Lett. 145 1988 177.