Photodissociation dynamics of OCIO: O(3PJ) state and energy distributions

Article abstract of DOI:10.1063/1.474222

The photodissociation dynamics of OClO in the near ultraviolet (UV) was investigated in a state specific and energy selective manner. At a dissociation wavelength of 308 nm, OClO(X ²B₁) was excited to the OClO(A²(18,0,O)) state, from which it decays into ClO(X ²II_3/2,1/2) and O(³P_2,1,0). The nascent oxygen fragments were detected spin selectively by resonant enhanced multiphoton ionization and time of flight measurements (REMPI-TOF). Based on the measurements and the conservation of energy and linear momentum, the internal energy of the ClO partner fragment was obtained. On average, more than 60% of the available energy is transferred into internal excitation of the ClO radical. Nearly the whole internal energy is vibrational energy with vibrational levels populated up to the energetic limit. Besides, the internal energy depends on the oxygen spin-orbit state because the fraction of highly excited ClO fragments increases with increasing total angular momentum J. The bimodal behavior of the fragment energy distribution indicates two different dissociation pathways, in which one leads to ClO radicals excited up to u = 10 and the other one up to v = 15. Furthermore, the decay is anisotropic, which was proved by polarization experiments. This is a hint for a short decay time estimated to be in the order of a few hundred femtoseconds.

Full text of DOI:10.1063/1.474222

Photodissociation dynamics of OClO: O ( 3 P J ) state and energy distributions
Melanie Roth, Christof Maul, and Karl-Heinz Gericke
Citation: The Journal of Chemical Physics 107, 10582 (1997); doi: 10.1063/1.474222
View online: http://dx.doi.org/10.1063/1.474222
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/107/24?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation dynamics of 3-bromo-1,1,1-trifluoro-2-propanol and 2-(bromomethyl) hexafluoro-2-propanol at
234 nm: Resonance-enhanced multiphoton ionization detection of Br (2 P j )
J. Chem. Phys. 134, 194313 (2011); 10.1063/1.3591373
Femtosecond dynamics of Cu ( H 2 O ) 2
J. Chem. Phys. 122, 054310 (2005); 10.1063/1.1836759
Br ( 2 P j ) and Cl ( 2 P j ) atom formation dynamics of allyl bromide and chloride at 234 nm
J. Chem. Phys. 114, 10368 (2001); 10.1063/1.1374581
Femtosecond photodissociation dynamics of Cr(CO) 6
AIP Conf. Proc. 454, 89 (1998); 10.1063/1.57175
The photodissociation dynamics of OClO between 306 and 370 nm: Fragment translational energy release and
recoil anisotropy
J. Chem. Phys. 106, 6538 (1997); 10.1063/1.473652
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

3
Photodissociation dynamics of OClO: O„ P … state and energy distributions
J
Melanie Roth, Christof Maul, and Karl-Heinz Gericke
Institut f u¨ r Physikalische und Theoretische Chemie, Technische Universit a¨ t Braunschweig,
Hans-Sommer-Strasse 10, 38106 Braunschweig, Germany
͑
Received 18 August 1997; accepted 22 September 1997͒
The photodissociation dynamics of OClO in the near ultraviolet ͑UV͒ was investigated in a state
˜
2
specific and energy selective manner. At a dissociation wavelength of 308 nm, OClO(X B ) was
1
˜
2
2
excited to the OClO(A A (18,0,0)) state, from which it decays into ClO(X ⌸3/2,1/2) and
O( P2,1,0). The nascent oxygen fragments were detected spin selectively by resonant enhanced
2
3
multiphoton ionization and time of flight measurements ͑REMPI-TOF͒. Based on the measurements
and the conservation of energy and linear momentum, the internal energy of the ClO partner
fragment was obtained. On average, more than 60% of the available energy is transferred into
internal excitation of the ClO radical. Nearly the whole internal energy is vibrational energy with
vibrational levels populated up to the energetic limit. Besides, the internal energy depends on the
oxygen spin-orbit state because the fraction of highly excited ClO fragments increases with
increasing total angular momentum J. The bimodal behavior of the fragment energy distribution
indicates two different dissociation pathways, in which one leads to ClO radicals excited up to v
ϭ10 and the other one up to vϭ15. Furthermore, the decay is anisotropic, which was proved by
polarization experiments. This is a hint for a short decay time estimated to be in the order of a few
hundred femtoseconds. © 1997 American Institute of Physics. ͓S0021-9606͑97͒02148-X͔
INTRODUCTION
who observed infrared spectra of the fundamental (v
ϭ1←vϭ0) and first overtone (vϭ2←vϭ0) bands of the
Photodissociative reactions play an important role in at-
mospheric chemistry. As Cl and ClO radicals are crucially
involved in stratospheric ozone depletion, there is great in-
terest in the photolysis of OClO. The near ultraviolet and
visible photodissociation of the OClO molecule leads to two
product channels, one producing O and ClO radicals and the
2
2
ClO (X ⌸ , ⌸ ) radical.
3
/2
1/2
15
Baumg a¨ rtel and Gericke showed that the two-photon
LIF technique is a powerful method in detecting nascent and
completely state-resolved ClO because the two-photon exci-
tation eliminates the tedious creation of VUV photons, while
the fluorescence observation in the VUV by means of a solar
blind photomultiplier dramatically improves the signal to
noise ratio. Davis and Lee16 studied the dissociation dynam-
ics of OClO at wavelengths between 350 and 475 nm. At
wavelengths shorter than 400 nm they observed a release of
the excess energy essentially in relative translation and not in
vibration, whereas at wavelengths longer than 400 nm,
nearly all thermodynamically accessible levels are populated.
They supposed that most of the internal ClO energy is stored
in vibration and less in rotation.
other one generating Cl and O products:
2
3
OClOϩh␯→O͑ P ͒ϩClO
͑1͒
͑2͒
J
2
→
Cl͑ P ͒ϩO .
J
2
In gas phase photodissociations, the first channel is preferred
with a quantum yield of about 95%. Incoherent results have
1,2
been obtained on the quantum yield of the second channel.
3
Davis and Lee published a yield of 3.9%Ϯ0.8% at 404 nm
and less than 0.2% at wavelengths smaller than 370 nm,
agreeing well with results lately obtained in our laboratory,4
where a different detection scheme was used. Therefore,
channel 2 clearly is a secondary process in the studied wave-
length regime. The main channel, 1, has been investigated by
determining the vibrational and rotational population of the
nascent ClO fragment. The state specific detection of ClO is
based on early spectroscopic work by various groups analyz-
Delmdahl et al.17 successfully employed the two-photon
LIF excitation when investigating the state-to-state OClO
˜
2
photodissociation from the A A ͑18,0,0͒ and ͑11,0,0͒ lev-
2
els of the symmetric stretch motion where the symmetric
stretch motion is transferred into the dissociative asymmetric
˜
2
stretch motion on the A A surface leading to the O and
2
ClO products. For the dissociation originating from the
͑18,0,0͒ state, Delmdahl et al.17 found extremely vibra-
5
6
ing spectra obtained in the microwave, infrared ͑IR͒, and
the vacuum ultraviolet ͑VUV͒ region, characterizing the
electronically excited C ⌺ state and the X ⌸ ground state.
tionally excited ClO. The conclusion of this work was that
7
2
the energy above the dissociation barrier of the A state is
2
2
2
i
essentially transferred into ClO vibration. This observation is
8,9
18
In addition, there are laser induced fluorescence ͑LIF͒,
confirmed by measurements of Huber et al., who observed
1
0,11
12
REMPI,
and femtosecond real-time probing experi-
the ClO energy distribution when OClO is excited into the
13
ments. Based on some of those measurements, Coxon cal-
culated the molecular constants of ClO in its ground and
excited state for vϭ0 and vϭ1. The spectroscopic constants
of the ground level were completed by Burkholder et al.,14
␯ ϭ9 to ␯ ϭ18 levels. They used photofragment transla-
1
1
tional energy spectroscopy and detected the ClO fragment
with a conventional quadrupole mass spectrometer.
Though the OClO photodissociation has been investi-
10582
J. Chem. Phys. 107 (24), 22 December 1997
0021-9606/97/107(24)/10582/10/$10.00
© 1997 American Institute of Physics
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

Roth, Maul, and Gericke: Photodissociation dynamics of OClO
10583
1
9,20
gated with the REMPI technique
and the O atom is easily
column filled with NaClO and Raschig rings. While the gas
2
accessed by REMPI,2
1,22
the experimenters concentrated on
flowed through the column, OClO and NaCl were produced.
The gas was transferred into the reaction chamber without
further purification. The cell was evacuated by two oil diffu-
sion pumps ͑Leybold, 500 l/s͒ and a turbomolecular pump
͑Leybold Heraeus 360 CSV, 360 l/s͒ to a base pressure of
measuring ClO and Cl fragments, but no oxygen fragment
analysis has been performed. In this work, for the first time
REMPI spectra of the oxygen fragments are presented,
formed by dissociation of OClO (A A (18,0,0)) upon ex-
citation at 308 nm. We state selectively probed all spin-orbit
˜
2
2
Ϫ4
10 Pa. The OClO gas was expanded into the reaction
3
states of oxygen O( P ) with ͑2ϩ1͒-REMPI at wavelengths
chamber by a pulsed nozzle ͑General Valve͒ which was op-
erated at a stagnation pressure of 200–600 mbar and a pulse
duration of typically 250 ␮s. At a repetition rate of 10 Hz,
J
around 225 nm. Analyzing state specific velocity distribu-
tions by time of flight measurements, kinetic and internal
energies of the ClO partner fragment are obtained, taking
into account the conservation of energy and linear momen-
tum. Thus, the energetics and the ͑joint partner͒ product state
distributions of the photodissociation process can completely
be characterized, especially with respect to the dissociation
pathways leading to spectroscopically dark, highly vibra-
tionally excited ClO fragments. Furthermore, new dissocia-
tion pathways were found with a rise in excitation energy but
their dissociation geometry, decay time, and spatial distribu-
tion is still unknown. In addition, we present high resolution
REMPI spectra of the spectroscopically accessible ClO
which were investigated by Vaida et al.23 at lower resolu-
tion. In order to obtain some information on the dissociation
time, polarization measurements were performed, and the in-
fluence of the polarization vector of the probe laser on the
shape of the TOF profile, reflecting the spatial fragment dis-
tribution, was analyzed.
Ϫ3
the background pressure was 10 Pa with the nozzle in op-
eration. From linewidth measurements with CO, the parent
molecule temperature could be estimated to be about 10 K.
Thus, the internal energy of OClO is negligible.
The molecular beam propagates perpendicularly to the
axis of our time of flight spectrometer, which is described in
detail elsewhere.28 At the end of the spectrometer, micro-
channel plates in chevron configuration are used to enhance
the ion signal. The TOF spectrometer can be operated in
three different ways.
First, in the drift mode the ions drift through the spec-
trometer with a total length sϭ0.57m only because of their
initial recoil velocity. The signal is monitored with a multihit
time-to-digital converter ͑FAST 7885͒ with a time resolution
of 5 ns per channel, stored in a data buffer ͑FAST MCD/PC͒
and transferred to a personal computer. The drift mode di-
rectly yields the velocity distribution F (v) from the mea-
v
D
T
The generation of highly vibrationally excited ClO in the
photodissociation of OClO is of great importance in atmo-
spheric chemistry because of possible subsequent ClO reac-
tions with constituents of the atmosphere.
Choo and Leu suggested the following reactions of vi-
brationally excited ClO with Cl and O3:
sured TOF profile F (t) via the relation
D
T
2
F ͑v͒ϭF ͑t͒•t .
͑6͒
v
Second, in the acceleration mode, the ions are acceler-
ated along a region of 0.19 m (ϭs/3) by a homogeneous
electric field of variable strength. In the present measure-
ments the electric field was fixed at 2600 V/m. The accelera-
tion region is followed by a field free drift region of twice
the length of the acceleration region. The shape of the tem-
porally broadened ion signal was monitored by a digital tran-
sient wave form recorder ͑LeCroy 9450͒. The acceleration
2
4
ClO͑vϾ2͒ϩCl→Cl ϩO,
͑3͒
͑4͒
2
ClO͑vӷ0͒ϩO →ClOϩO ϩO.
Similarly, the UV dissociation of O at 226 nm (O3
3
2
3
2
5
ϩh␯(226 nm͒→O ϩO) is known to produce highly vi-
2
3
Ϫ
g
mode yields the distribution F (v ) of the velocity compo-
brationally excited O (X ⌺ ,vу26) molecules, which
v
x
x
2
could be a so far neglected source in the atmospheric ozone
production via the reaction,
nent v along the spectrometer axis from the TOF profile
x
A
F (⌬t) via the linear relationship
T
3
O ͑vу26͒ϩO →O ϩO͑ P͒.
͑5͒
ϭc •⌬t,
͑7͒
2
2
3
vx
Ј
Furthermore, a vibrationally excited oxygen molecule can be
2
0
where cЈϭ8s/3t ϭ8eU /m accounts for the acceleration
A
dissociated by near UV radiation, whereas O in its ground
2
voltage U , the spectrometer length s, particle mass m, and
A
level requires VUV26 radiation. As a result of the higher flux
of near UV light in the atmosphere, this oxygen dissociation
pathway becomes important for atmospheric chemistry.
Due to the formation of extremely vibrationally excited
charge e. ⌬tϭtϪt is the deviation from the center of the
acceleration profile.
Equation ͑7͒ holds in all practically relevant cases,
where the acceleration energy is large compared to the initial
kinetic energy of the observed fragment:
0
ClO radicals, reactions with the atmospheric molecules O or
N become possible, and in this case the reaction rate may
well exceed the relaxation rate.
2
2
1
2
mv2.
e•UAӷEkinϭ
͑8͒
F (v) is easily obtained from F (v ) if the spatial fragment
v
v
x
x
EXPERIMENT
distribution is known.
The OClO molecule was generated according to the
Third, in the Doppler mode the electric field is so large
that one cannot measure any difference in arrival time. In-
stead the total intensity of the ion signal is integrated by a
27
method of Derby and Hutchinson. A bulb containing a gas
mixture of about 10% Cl and 90% N was coupled to a
2
2
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

10584
Roth, Maul, and Gericke: Photodissociation dynamics of OClO
boxcar integrator and averager ͑Stanford Research System
SR 250͒ and digitized by an A/D converter while the laser is
tuned over the Doppler line. In the Doppler mode the distri-
bution F (v ) is obtained, similarly as in the preceding
v
y
y
paragraph, from the Doppler broadened line shape D(⌬␯)
via the relation,
c⌬␯
vyϭ
,
͑9͒
␯0
where c is the speed of light, v is the velocity component
y
along the laser beam, ␯ is the center of the Doppler line, and
0
⌬
␯ϭ␯Ϫ␯ is the frequency deviation from the line center. If
0
the dissociation process is isotropic, F (v ) and F (v ) are
v
x
v
y
x
y
equal.
The high temporal resolution of TOF profiles obtained in
the drift mode allows a very accurate determination of the
underlying velocity distribution. The acceleration profiles
compensate for the decreasing spectrometer transit efficiency
for low kinetic energies, although at the expense of temporal
resolution. Thus, operating the spectrometer in both modes
combines the respective advantages and yields highly accu-
rate, complete velocity distributions.29
FIG. 1. Term scheme of the oxygen atom which contains the ͑2ϩ1͒-REMPI
transitions of every spin-orbit state. The energy levels are not scaled.
An excimer pumped dye laser system ͑Lambda Physik
LPX 605i ͓XeCl 308 nm͔, LPD 3000͒ delivered the photoly-
sis and probe pulses. A small part of the pump light was
coupled out by a beam splitter and was used to photolyze the
molecule. An optical delay line formed the jitter-free timing
between photolysis and probe pulse. The laser pulses were
electronically delayed with respect to the molecular beam
rival times. The signals were normalized by measuring the
probe laser intensity.
RESULTS
͑
Stanford Research System DG 535͒.
In the photodissociation of OClO the oxygen fragment is
3
3
The oxygen fragments O( P2,1.0) in their different spin-
formed in its P ground state, where its angular momentum
J
orbit states were probed by ͑2ϩ1͒-REMPI in the wavelength
region of 225 nm. In order to obtain this wavelength we used
Coumarin 47 ͑440–484 nm, 5–10 mJ/pulse͒ and a ␤-barium
borate ͑BBO͒ crystal for second harmonic generation ͑50–
quantum number has the values Jϭ0,1,2. The energy levels
of these states are separated by 158 and 227 cm as it is
Ϫ1
indicated by the term scheme in Fig. 1. Using transitions
3
3
3
p P ←2p P with ⌬Jϭ0, we detected the different
J
J
3
200 ␮J/pulse͒. The light pulses counterpropagated in the re-
spin-orbit states by ͑2ϩ1͒-REMPI. The O( P ) state was
probed at 225.572, O( P ) at 225.974, and O( P ) at
2
3
3
action chamber and both intersected the spectrometer axis
perpendicularly. The molecular beam intersected the probe
pulse in an angle of 54°. In order to determine anisotropic
effects in the photodissociation of OClO we polarized the
photolysis laser parallel and perpendicular to the spectrom-
eter axis sជ . The polarization of the analyzing dye laser light
1
0
2
26.149 nm.
Conservation of linear momentum and energy allows a
calculation of maximum oxygen recoil velocity. The avail-
able energy Eav is the difference between photolysis energy
Ϫ1 30
h␯ and dissociation energy ED (19 311 cm ), added to
was always maintained parallel to the spectrometer axis.
The photolysis and probe pulse were focused in the re-
action chamber. The photolysis light was focused by a 150
mm lens in the Doppler and the acceleration mode, whereas
for the drift mode a 100 mm lens was used. The probe pulse
was focused by a 80 mm lens for the acceleration and drift
profiles, for the Doppler profiles by a 135 mm lens.
the internal energy of the parent molecule, E ͑OClO): E
int
av
ϭh␯ (308 nm)ϪE ϩE ͑OClO). In the molecular beam
D
int
the parent molecule is jet cooled, and thus its internal energy
Ϫ1
E ͑OClO) is negligible and a value of E ϭ13 157 cm is
int
av
calculated. The available energy is distributed into the ki-
netic energy Ekin and internal energy Eint of the fragments.
The kinetic energy of the ClO fragment is that of the oxygen
fragment weighted by the fragment mass ratio, which is a
consequence of the conservation of linear momentum. The
The other partner fragment ClO was detected in the
Doppler mode only by a one-color experiment with RDC
360 ͑342–373 nm, 7–10 mJ/pulse͒ as laser dye. In that case
internal energy E ͑O) of the oxygen atom is given by the
int
OClO photolysis and ClO detection occur at the same wave-
length. Since the flight times of 35ClO and ClO differ by
their masses, isotope specific spectra were observed simulta-
neously by adjusting two boxcar gates at the respective ar-
spin-orbit state energies ͑see Fig. 1͒,
37
mO
E ϭE ͑O)ϩ
ͩ
1ϩ
ͪ
Ekin͑O)ϩE ͑ClO).
͑10͒
av
int
int
mClO
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

Roth, Maul, and Gericke: Photodissociation dynamics of OClO
10585
FIG. 2. Drift profiles of oxygen in different spin-orbit states (O 3P ,J
J
ϭ0,1,2) containing three peaks. The first peak corresponds to photolysis of
FIG. 3. TOF profiles obtained by acceleration of the generated oxygen
OClO at 225 nm, the second one to secondary photolysis of ClO at 308 nm
and the main peak to photolysis at 308 nm.
fragments with an electric field ͑U ϭ500 V, unpolarized photolysis͒. For
b
3
O( P ) the highest contribution of slow O fragments is observed.
2
guished from the oxygen fragments formed in primary pho-
tolysis because of their low intensity.
The oxygen velocity is at its maximum when the ClO radical
has no internal energy. Thus, the maximum velocities of
At a first glance there is no remarkable difference be-
tween the profiles, so that the velocity distribution seems to
be independent of the spin-orbit state. However, these drift
profiles are not sensitive to slow recoil velocities and reliable
conclusions can only be drawn when the acceleration profiles
are taken into account. The results of these measurements are
shown in Fig. 3. These profiles were measured using unpo-
larized probe light, so that possible anisotropic effects of the
fragmentation process on the profiles are strongly reduced. In
3
3
3
O( P ), O( P ), and O( P ) are 3870, 3847 and 3836 m/s,
2
1
0
respectively.
In Fig. 2 the drift profiles of the oxygen in its spin-orbit
3
states (O( P ),Jϭ0,1,2) are shown. Of the three peaks, the
J
main peak is caused by photolysis of OClO at 308 nm. If the
ascending slope of this peak at short flight times is extrapo-
lated to the time axis, the maximum O recoil velocity can be
determined. The obtained value of 3900Ϯ100 m/s is in good
agreement with the calculated values according to Eq. ͑10͒.
The first peak in Fig. 2, corresponding to a value of 5200
m/s, is oxygen formed in the photolysis of OClO by a photon
of 225 nm ͑probe pulse͒, and the second peak represents the
fast oxygen radicals generated in the secondary photolysis at
3
3
3
contrast to the profiles of O( P ) and O( P ), the O( P )
1
0
2
profile shows an increase of intensity in the middle of the
peak, which indicates the preferred formation of slow atoms
generated in the Jϭ2 state.
In order to test any anisotropic behavior of the dissocia-
tion process we polarized the photolysis laser parallel and
perpendicular to the spectrometer axis ៝s . For an isotropic
dissociation pattern the profiles should have the same ap-
pearance. However, the obtained profiles presented in Fig. 4
and Fig. 5 differ significantly, which is an indication of an
anisotropic fragmentation. Moreover, this proves that the dis-
308 nm, where a second photon dissociates the ClO fragment
into Cl and O. In this case the total oxygen velocity is given
by the sum and difference of the ClO parent molecule veloc-
ity as a consequence of the primary photolysis and the oxy-
gen velocity of the secondary photolysis. The slow oxygen
radicals obtained via secondary photolysis cannot be distin-
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

10586
Roth, Maul, and Gericke: Photodissociation dynamics of OClO
FIG. 4. Acceleration profiles measured with the photolysis light being po-
larized perpendicular to the spectrometer axis, where the oxygen fragments
are preferentially generated perpendicular to the spectrometer axis.
FIG. 5. Acceleration profiles measured with the photolysis light polarized
parallel to the spectrometer axis, where the oxygen fragments are preferen-
tially generated in or against the direction of the detector.
sociation has to be fast because otherwise the anisotropic
effects would be strongly reduced due to parent molecular
rotation. The transition dipole moment of OClO
31
if no correlation between ␤ and F (v) exists. P (cos ␪) is
v
2
the second Legendre polynomial of the polar angle ␪ with
respect to the transition dipole moment ␮. The second term
in brackets has to vanish to obtain a distribution independent
˜
2
˜
2
(
A A ←X B ) lies in the molecular plane, parallel to the
2
1
line connecting the two oxygen atoms. Under the assumption
that the dissociation is very fast, the valence angle of the
ground state is maintained in the excitation. If the O frag-
ment recoils from the center of mass of the ClO partner, the
velocity vector forms an angle of 17.6° with respect to the
dipole moment ͑see Fig. 6͒. Consequently, the O fragments
are mainly generated perpendicular to the spectrometer axis
ជs if the photolysis light is polarized perpendicular to sជ , and
the signal intensity at the line center will be increased ͑Fig.
4͒. On the other hand, if the photolysis pulse is polarized
parallel to sជ , the O atoms are mainly formed in or against the
direction of sជ ͑Fig. 5͒. Consequently, there is a dip in the
center of the acceleration profile.
The angle dependent velocity distribution F (␪,v) can
v
be described by F (v) and a term containing the anisotropy
v
parameter ␤:
FIG. 6. Recoil angle of the O fragment calculated under the assumption that
the angle of the ground level remains constant during fragmentation and the
O fragment recoils from the center of mass of the ClO fragment.
F ͑v͒
v
F ͑␪,v͒ϰ
͓1ϩ␤P ͑cos ␪͔͒,
͑11͒
v
2
v
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

Roth, Maul, and Gericke: Photodissociation dynamics of OClO
10587
3
FIG. 8. O( P ) velocity distributions received by the combination of accel-
FIG. 7. Calculated acceleration profiles based on the profiles obtained by
polarizing the photolysis light parallel and perpendicular to the spectrometer
axis.
J
eration and drift profiles.
In order to test whether the high vibrational excitation of
the ClO radical is directly measurable in detecting ClO itself,
we observed REMPI one-color spectra of ClO in the wave-
length region of 338–356 nm, which are presented in Fig. 9.
The spectra are normalized to the laser intensity and are iso-
of the spatial distribution of the fragments. Since the anisot-
ropy parameter exhibits a value different from zero, the Leg-
endre polynomial has to fulfill the condition
1
2
P ͑cos ␪ ͒ϭ ͑3 cos ␪ Ϫ1͒ϭ0,
͑12͒
35
2
m
2
m
tope specific because of the difference in mass of ClO and
where ␪ ϭ54.7° is the so-called magic angle. In order to
m
obtain the acceleration profiles of the magic angle, profiles of
both polarizations were normalized separately to equal areas
and summed up with suitable factors according to a proce-
dure first described by Huber and co-workers.32 The result is
shown in Fig. 7.
Based on the acceleration and drift profiles, the velocity
distribution of every state is calculated and the results are
presented in Fig. 8. All obtained distributions show bimodal
3
behavior, but the velocity distribution of the O( P ) contains
2
3
3
more slow ions than that of the O( P ) or O( P ).
0
1
The obtained velocity distributions can be used to calcu-
late Doppler profiles. These calculated Doppler profiles are
in good agreement with the measured ones, indicating the
conformance between the drift, acceleration, and Doppler
profiles. Integration of Doppler profiles obtained under
equal conditions yields
a
population relation of
3
3
3
O( P ):O( P ):O( P )ϭ0.53: 0.32: 0.15, which is compa-
FIG. 9. One-color spectra of 35ClO and 37ClO with the natural isotope
2
1
0
relation 35Cl: Clϭ3:1 reflected in the intensity of the lines.
37
rable to a statistical distribution ͑0.56: 0.33: 0.11͒.
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

10588
Roth, Maul, and Gericke: Photodissociation dynamics of OClO
3
7
ClO. The signal intensity of both isotopes reflects the natu-
ral isotope relation of 35Cl: Clϭ3:1. The vibrational bands
are assigned in Fig. 9. The ⌸ spin system is energetically
higher than the ⌸1 spin system and is less populated. The
37
2
3/2
2
/2
observed ClO is vibrationally excited up to vϭ3, while the
intensity of the vibrational bands decreases with increasing
v. Comparatively few rotational lines are observed, so the
rotational excitation is low. This behavior is comparable to
the results of Vaida et al. at 339–342 nm, but the rotational
lines are monitored with much better resolution.2 Unfortu-
nately, unfavorable Franck–Condon factors do not allow di-
rect monitoring of the ClO fragments for vϾ5, as is shown
in earlier LIF measurements.15 Therefore, the analysis for the
highly vibrationally excited states of ClO has to rely on the
O atom observations.
3
DISCUSSION
3
The observed velocity distributions of the O( P ),
2
3
3
O͑ P ), and O( P ) products are bimodal in character,
1
0
which is also reflected in the energy distributions. The slow
oxygen fragment corresponds to a ClO fragment with ex-
tremely high internal energy. The contribution of slow ions
3
increases with increasing J from 20% for O( P ) to 32% for
0
3
3
O( P ), so that together with the O( P ) state the highest
2
2
portion of highly excited ClO radicals is found. Using Eq.
10͒ and the kinetic energy of the O fragments, the internal
͑
3
energy of the ClO fragment is calculated. The computed in-
ternal energy distributions are presented in Fig. 10. We
found ClO fragments with extremely high internal energy of
FIG. 10. Internal energy distribution of ClO for different O( PJ) partner
fragments. Remarkable is the bimodal behavior and the increasing fraction
of highly internally excited ClO products with increasing J of the partner
atom.
Ϫ1
about 8000 and 12 500 cm , respectively, in agreement
with Delmdahl et al..17 The decay of a molecule is influ-
enced by the potential surfaces of its electronic levels. Figure
formed into asymmetric core movement. As soon as enough
energy is transferred into the asymmetric bond length distor-
tion to overcome the barrier, the OClO can fragment into a
vibrationally excited ClO and an O radical.
11 shows cuts through ab initio potential surfaces calculated
by Peterson and Werner with respect to symmetric and
asymmetric bond length distortion and variation of the va-
3
3
lence angle. If the bond lengths are varied asymmetrically,
the molecule takes on C symmetry, otherwise C2v symme-
s
try. Both symmetries are described with different notations.
In Fig. 11 the C notation is used, whereas in the text we will
s
use C2v notation. The alternative notation is added in brack-
˜
2
2
ets to facilitate the comparison. The A A and B surfaces
2
2
are bonding with respect to both the symmetric OClO dis-
tance rClO and the valence angle ␸ because of their potential
˜
2
2
minimum. If the molecule fragments via the A A ͓ AЉ͔ and
2
2
2
A ͓ AЈ͔ surfaces with respect to asymmetric bond length
1
distortion RClO , it has to overcome an energy barrier. The
˜
2
potential surface of the A A state exhibits a barrier E with
2
b
respect to the RClO coordinate. If the excitation energy is
sufficiently high, the dissociation could follow a dissociation
channel, where a vibrational predissociation along the v co-
3
ordinate takes place. Although the molecule is initially
˜
2
FIG. 11. Cuts through ab initio potential energy surfaces calculated by
Peterson and Werner ͑Ref. 33͒ for variation of the symmetric stretch coor-
dinate RClO , the bending coordinate ␸, and the asymmetric stretch coordi-
bound in the A A level along the symmetric stretching
2
mode v , a small coupling between v and v is enough to
1
1
3
change into the asymmetric system while the energy is trans-
nate rClO .
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

Roth, Maul, and Gericke: Photodissociation dynamics of OClO
10589
3
TABLE I. State selective energy partitioning for the different O( P ) states.
J
Fraction of the
Spin-orbit
state
total oxygen
signal
Eint͑O)/Eav
Eint͑O͓͒cm
Eint͑ClO)/Eav
Eint͑ClO͓͒cm
Ekin͑O)/Eav
Ekin͑O͓͒cm
Ekin͑ClO)/Eav
Ekin͑ClO͓͒cm ]
Ϫ1
Ϫ1
Ϫ1
Ϫ1
]
]
]
3
O( P )
0.53
0.32
0.15
0%
cm
1%
58 cm
2%
27 cm
0.6%
70%
9122 cm
66%
8696 cm
63%
8337 cm
67.7%
23%
3071 cm
25%
3275 cm
27%
3496 cm
24.2%
7%
964 cm
8%
1028 cm
8%
1097 cm
7.5%
2
Ϫ1
Ϫ1
Ϫ1
Ϫ1
0
3
O( P )
1
Ϫ1
Ϫ1
Ϫ1
Ϫ1
1
2
3
O( P )
0
Ϫ1
Ϫ1
Ϫ1
Ϫ1
average
2
The minimum of the A1 surface with respect to the
valence angle is obtained for a value of 180° whereas the
m ϩm ͑1Ϫcos ␸ ͒
Cl
O
0
1
8
͗
E ͘ ϭ
rot ClO
␻2
,
͑14͒
mClO
˜
2
minimum of the A A level corresponds to an angle of
2
107°. A change of the valence angle induces a torque which
where ␻ is the energy of the bending vibrational mode v ,
2 2
should be reflected in high rotational energy. Since only low
and ␸ ϭ117.7° is the bending angle of the OClO ground
0
level.36 Based on Eq. ͑14͒ we calculate a value of 65 cm
.
Ϫ1
rotational energy was observed in the experiment, a decay on
2
the A surface is improbable. An intramolecular transition
In the impulsive model, a triatomic molecule ABC de-
cays because of an instantaneous development of a repulsive
force between the departing atom A, and the atom B it was
initially bound to. This force is directed along the breaking
bond. Therefore, a torque on the diatom BC around its center
of mass is induced. Consequently BC is rotationally excited
and the rotational angular momentum is directed perpendicu-
lar to the molecular plane. The third atom C is not involved
in the dissociation process. According to the impulsive
model, the limitations of which are shown by Schinke,37 the
rotational energy can be calculated by:
1
˜
2
˜
2
from the A A surface to the ground level X B and a
2
1
subsequent decay should lead to a statistically distributed
ClO fragment energy which is also in contrast to the
1
7
experiments, and therefore the mechanism is equally un-
likely.
Characteristic for our measurements are the extremely
high internal energy of ClO and the anisotropic distribution
of the products. The anisotropic behavior indicates a fast
decay, i.e., the lifetime of excited OClO is less than a rota-
tional period of the molecule. Otherwise the originally
aligned molecules would lose their alignment by rotation.
2
2
O
sin ␸ m
The average time of rotation T is given by
0
r
E ϭE
rot
.
͑15͒
av
2
2
2
͑
m ϩm ͒ Ϫm cos ␸
O
O Cl 0
Trϭ2␲ͱ͚ ͑␪ /3͒RT,
͑13͒
We obtain a rotational energy of 1037 cmϪ1. Both calcula-
i
i
tions indicate that the rotational energy is small compared to
Ϫ1
where ␪ are the moments of inertia, R is the gas constant,
the total internal ClO energy of about 8900 cm , agreeing
i
34
and T is the temperature. With a temperature of 10 K, a
rotational period of about 15 ps is calculated, which is an
upper limit. Using the linewidths of the absorption spectrum
at 296 K measured by Wahner et al., the lower limit of the
lifetime of the OClO͑18,0,0͒ state could be estimated to
with the results of Huber et al., which are based on a rota-
Ϫ1 18
tional energy of 650 cm
.
The energy difference (h␯-E ) between excitation en-
b
ergy h␯ and barrier E is mainly transformed into internal
b
energy of the ClO fragment. The remaining kinetic energy of
3
5
about 0.2 ps.
the fragments is equal to (E ϪE ). We observed a mean
b
D
Ϫ1
The bimodal behavior of the internal energy distribution
refers to different dissociation channels. Because of the fast
decay it is probable that the OClO molecule directly disso-
internal ClO energy of 8900 cm , which corresponds to a
vibrational excitation around vϭ10. That way an energy bar-
Ϫ1
rier of 23 500 cm results from dissociation energy and ki-
˜
2
ciates on the excited A A potential surface if the excitation
netic energy. The conclusion that the surplus energy (h␯
2
energy overcomes the barrier. This should be reflected in
highly vibrationally excited ClO radicals, which comes from
the initial vibrational excitation of OClO. In that case the
excited OClO molecule nearly decays under a perpendicular
ϪE ) is released as ClO vibration is supported by recent
experimental results. Huber et al. observed an increase of
internal ClO energy with increasing available energy,
b
whereas the average translational energy Ekin ͑ClOϩO͒ re-
˜
2
Ϫ1 18
geometry. The equilibrium angle of the A A state is about
mains constant, ranging from 4200 to 4600 cm
.
Delm-
2
1
18°. Therefore, the oxygen radical could hardly transfer vi-
dahl et al. examined the OClO dissociation of the
˜
2
brational energy into ClO while it is repelled.
(A A (11,0,0)) state where the OClO molecule was excited
2
17
One can estimate the rotational energy of the ClO frag-
ment by different model calculations. If the parent molecule
OClO does not rotate, the mean rotational energy can be
at 351 nm. From the energy balance they expected ClO
radicals vibrationally excited to vϭ3, which was manifested
by their measurements. In our experiment we observed the
percentages of the available energy which were transferred in
calculated from the zero point motion of the v vibration:
2
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

10590
Roth, Maul, and Gericke: Photodissociation dynamics of OClO
TABLE II. ClO fragments with sufficient energy to react with N2.
Fraction of the
total oxygen
Percentage of ClO radicals
following the second dissociation
Percentage of ClO radicals
with an internal energy higher
Ϫ1
Spin-orbit state
signal
pathway
than 8500 cm
3
O( P )
0.53
0.32
0.15
32%
24%
20%
62%
54%
49%
2
3
O( P )
1
3
O( P )
0
average
27.6%
57.5%
internal and kinetic energy shown in Table I. The main part
of about 68% is found in internal ClO excitation. The kinetic
energy is nearly constant and its value is in good agreement
with that of Huber et al..18
enthalpy is based on recommended values of the standard
enthalpy of formation for reactants in their respective ground
states.40 We computed the portion of ClO radicals having
enough energy to follow the reaction with N ; the results are
2
While the first maximum ͑at low internal energy͒ of the
ClO internal energy can be explained by this decay mecha-
also summarized in Table II. On average more than 50% of
the ClO fragments have sufficient energy to react with N2.
In order to determine the impact of extremely vibra-
tionally excited ClO on the atmosphere, the dissociation dy-
namics of other OClO vibrational states with respect to their
spin selective behavior should also be investigated. This is,
in principle, possible because it only requires another tunable
laser system so that the excitation energy is variable.
˜
2
nism on the A A surface, it seems very unlikely that the
2
second maximum ͑at high internal energy͒ can be explained
by a decay on the same surface.
The second dissociation mechanism might presumably
2
2
occur on the B ͓ AЈ͔ surface. At an excitation energy of
2
Ϫ1
˜
2
2
3
2 468 cm , the potential surfaces of the A A and A1
2
levels are close together and the crossing is smooth ͑see Fig.
˜
2
2
1
1͒. Thus, a fast A A ← A transition can be expected. A
2
1
CONCLUSION
possibly following vibronic, and, therefore, likewise fast
2
2
coupling with the B ͓ AЈ͔ surface leads to the dissociation.
2
The OClO dissociation was investigated by REMPI-TOF
On this surface the molecule decays at a bending angle of
about 180° that presumably causes high vibrational energy.
If it is assumed that the energy difference between the exci-
3
2
measurements of the O( P ) and ClO( ⌸ ) fragments. The
J
⍀
dissociation leads to highly vibrationally excited ClO and
3
correspondingly slow O( P ) fragments. The bimodal be-
J
2
2
tation energy and the minimum of the B ͓ AЈ͔ surface de-
2
havior of the velocity and energy distributions leads us to
conclude that the fragmentation follows two different disso-
ciation pathways with a broad velocity distribution. Channel-
ing the vibrational energy of the excited OClO into the
asymmetric stretch coordinate, the parent molecule decays
termines the internal energy of the ClO fragment, the internal
energy could receive nearly the whole available energy
Ϫ1
(
E ϭ13 157 cm ). Thus, the observed second peak in the
av
Ϫ1
energy distribution at about 12 500 cm is favorably ex-
plained by this decay mechanism.
˜
2
via a barrier on the A A potential energy surface, releasing
2
The amount of ClO fragments being generated via the
second dissociation pathway depends on the spin-orbit state
of the O partner atom and increases with increasing J. We
calculated the contribution of each spin-orbit state by inte-
the difference between photon energy and barrier height as
ClO vibration. Extremely vibrationally excited ClO is likely
to be produced following conversion of the OClO from the
2
2
initially excited A surface to the B surface, releasing the
2
2
Ϫ1
grating the internal energy from 11 000 cm to Eav and nor-
total available energy into ClO internal energy. The fragmen-
tation process is spin selective. The fraction of highly vibra-
tionally excited ClO increases with increasing J of the oxy-
gen partner fragment. As an anisotropic spatial distribution is
observable, the decay time is expected to be of the order of a
few hundred femtoseconds.
malizing it to the total internal ClO energy. The results are
3
shown in Table II. In the O( P ) state, 32%, in the other
2
states, 24% and 20% respectively, of the ClO fragments fol-
low the second dissociation pathway.
The high vibrational excitation of the ClO radical allows
reactions which are normally thermally prevented by the ac-
tivation energy, and endothermic reactions become possible.
A reaction with chlorine and ozone ͓see Eqs. ͑3͒ and ͑4͔͒ is
unlikely because relaxation measurements of ClO (vр6)
The extreme internal excitation of the ClO product opens
up new reaction schemes in atmospheric chemistry, such as
the reaction of ClO with N . Therefore the OClO dissocia-
2
tion needs to be considered as a possible chlorine atom
source in the atmosphere.
with N indicate that the vibrational relaxation rate is too
2
high to allow a significant amount of ClO (v) reactions.
However, the reaction with N is energetically allowed for
2
vϾ9, according to
ACKNOWLEDGMENTS
ClO͑vӷ0͒ϩN →N OϩCl, ⌬H ϭ8500 cmϪ1, ͑16͒
2
2
r
We acknowledge financial support of the Deutsche For-
schungsgemeinschaft. M.R. thanks the Fonds der Chemis-
chen Industrie for fellowship support.
with the Cl product itself catalyzing the ozone
3
8,39
decomposition.
The calculation of the standard reaction
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57

Roth, Maul, and Gericke: Photodissociation dynamics of OClO
10591
1
2
20
W. G. Lawrence, K. C. Clemitshaw, and V. A. Apakarian, J. Geophys.
Res. 95, 591 ͑1991͒.
E. Bishenden, P. D. Hammer, and C. J. Howard, J. Chem. Phys. 95, 2113
E. Bishenden and D. J. Donaldson, J. Chem. Phys. 101, 9565 ͑1994͒.
21
B. E. Forch and C. N. Merrow, J. Chem. Phys. 95, 3252 ͑1991͒.
T. Kinugawa, T. Sato, T. Arikawa, Y. Matsumi, and M. Kawasaki, J.
Chem. Phys. 93, 3289 ͑1990͒.
22
͑1991͒.
3
4
5
23
24
25
H. F. Davis and Y. T. Lee, J. Phys. Chem. 96, 5681 ͑1992͒.
R. F. Delmdahl ͑unpublished͒.
R. K. Kakar, E. A. Cohen, and M. Geller, J. Mol. Spectrosc. 70, 243
E. R u¨ hl, A. Jefferson, and V. Vaida, J. Phys. Chem. 94, 2990 ͑1990͒.
K. Y. Choo and M. T. Leu, J. Phys. Chem. 89, 4832 ͑1985͒.
R. L. Miller, A. G. Suits, P. L. Houston, R. Toumi, J. A. Mack, and A. M.
Wodtke, Science 263, 1831 ͑1994͒.
R. Toumi, P. L. Houston, and A. M. Wodtke, J. Chem. Phys. 104, 775
͑1996͒.
R. I. Derby and W. S. Hutchinson, Inorg. Synth. 4, 152 ͑1953͒.
T. Haas, C. Maul, and K.-H. Gericke, Chem. Phys. Lett. 202, 108 ͑1993͒.
C. Maul, T. Haas, K.-H. Gericke, and F. J. Comes, J. Chem. Phys. 102,
3238 ͑1995͒.
I. P. Fischer, Trans. Faraday Soc. 63, 684 ͑1967͒.
M. Mons and I. Dimicoli, Chem. Phys. Lett. 131, 298 ͑1986͒.
J. R. Dubs, U. Br u¨ hlmann, and J. R. Huber, J. Chem. Phys. 84, 3106
͑1978͒.
6
26
R. T. Menzies, J. S. Margolis, E. D. Hinkley, and R. A. Toth, Appl. Opt.
6, 523 ͑1977͒.
1
7
8
27
28
N. Basco and R. D. Morse, J. Mol. Spectrosc. 15, 35 ͑1973͒.
Y. Matsumi, S. M. Shamsuddin, and M. Kawasaki, J. Chem. Phys. 101,
29
8
262 ͑1994͒.
9
J. B. Nee and K. J. Hsu. J. Photochem. Photobiol. A: Chem. 55, 269
1991͒.
30
͑
1
1
1
0
1
2
31
32
M. T. Duignan and J. W. Hudgens, J. Chem. Phys. 82, 4426 ͑1985͒.
E. Bishenden and D. J. Donaldson, J. Chem. Phys. 101, 9565 ͑1994͒.
T. Baumert, J. L. Herek, and A. H. Zewail, J. Chem. Phys. 99, 4430
͑1986͒.
33
͑1993͒.
K. A. Peterson and H. J. Werner, J. Chem. Phys. 101, 9565 ͑1992͒.
K.-H. Gericke. Ph.D. thesis, Johann Wolfgang Goethe-Universit a¨ t Frank-
fur Main, 1981.
A. Wahner, G. S. Tyndall, and A. R. Ravishankara, J. Phys. Chem. 96,
8948 ͑1992͒.
R. Vasudev, R. N. Zare, and R. N. Dixon, J. Chem. Phys. 80, 4863 ͑1984͒.
R. Schinke, Comments At. Mol. Phys. 23, 15 ͑1989͒.
S. Solomon, Nature ͑London͒ 347, 347 ͑1990͒.
P. J. Crutzen and F. Arnold, Nature ͑London͒ 324, 651 ͑1986͒.
W. B. DeMore, S. P. Sander, D. M Golden, R. F. Hampson, M. J. Kurylo,
C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina. NASA
Panel for Data Evaluation.
1
1
3
4
34
J. A. Coxon, Can. J. Phys. 57, 1538 ͑1979͒.
J. B. Burkholder, P. D. Hammer, and C. J. Howard, J. Mol. Spectrosc.
35
1
24, 139 ͑1987͒.
1
1
1
5
6
7
S. Baumg a¨ rtel and K.-H. Gericke, Chem. Phys. Lett. 227, 461 ͑1994͒.
36
37
H. F. Davis and Y. T. Lee, J. Chem. Phys. 105, 8124 ͑1996͒.
R. F. Delmdahl, S. Baumg a¨ rtel, and K.-H. Gericke, J. Chem. Phys. 104,
38
2
883 ͑1996͒.
1
1
8
9
39
A. Furlan, H. A. Scheld, and J. R. Huber, J. Chem. Phys. 106, 6538
1997͒.
40
͑
V. Vaida, E. C. Richard, A. Jefferson, L. A. Cooper, R. Flesch, and E.
R u¨ hl, Ber. Bunsenges. Phys. Chem. 96, 391 ͑1992͒.
J. Chem. Phys., Vol. 107, No. 24, 22 December 1997
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
129.49.251.30 On: Fri, 05 Dec 2014 04:20:57