State-to-state dissociation of OClO(qq 2A2ν1,0,0) → ClO(X 2ΠΩ,v,J) + O(3P)

Article abstract of DOI:10.1063/1.471649

The photolysis of OClO at 351nm and 308nm excites the electronic OClO(qq ²A₂←qq ²B₁) transition with simultaneous powerful excitation of only symmetric stretching modes of the qq ²A₂ state.

Full text of DOI:10.1063/1.471649

Statetostate dissociation of OClO(Ã2 A 2ν1,0,0)→ClO(X 2ΠΩ,v,J)+O(3 P)
Ralph Felix Delmdahl, Stephan Baumgärtel, and KarlHeinz Gericke
Citation: The Journal of Chemical Physics 104, 2883 (1996); doi: 10.1063/1.471649
View online: http://dx.doi.org/10.1063/1.471649
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/104/8?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Photodissociation dynamics of OClO
J. Chem. Phys. 105, 8142 (1996); 10.1063/1.472700
A study of the S(3 P 2,1,0;1 D 2) production in the 193 nm photodissociation of CH3S(X̃)
J. Chem. Phys. 97, 6283 (1992); 10.1063/1.463690
The multiplet state distribution of O(3 P J ) produced in the 193 nm photodissociation of SO2
J. Chem. Phys. 93, 868 (1990); 10.1063/1.459462
Optogalvanic spectroscopy of twophoton dissociation of Tl2 at 532 nm producing large population inversion
between Tl(72 S 1/2) and (62 P 3/2) states
J. Chem. Phys. 82, 2217 (1985); 10.1063/1.448314
Quenching of CN(A 2Π, v=0,1) by O2, H2, N2, NO, and CO2
J. Chem. Phys. 82, 2511 (1985); 10.1063/1.448299
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

2
2
3
˜
State-to-state dissociation of OClO(A A2␯1,0,0)˜ClO(X ⌸⍀ ,v,J)؉O( P)
¨
Ralph Felix Delmdahl, Stephan Baumgartel, and Karl-Heinz Gericke
Institut fur Physikalische und Theoretische Chemie, Johann Wolfgang Goethe-Universitat Frankfurt/Main,
Marie-Curie-Strasse 11, D-60439 Frankfurt/Main, Germany
¨
¨
͑Received 12 June 1995; accepted 6 November 1995͒
Applying the two-photon laser-induced fluorescence technique for nascent state resolved
2
2
˜
ClO(X ⌸ , ,J) detection, the photofragmentation dynamics of OClO in the (A A 11,0,0) and
v
⍀
2
2
˜
the (A A218,0,0) state is investigated at fixed photolysis wavelengths of 351 nm and 308 nm. In
both experiments the product fragments are formed in their electronic ground states, namely
ClO͑2⌸⍀͒ and O(3P). A complete product state analysis proves the ClO radicals originating from
2
˜
the OClO(A A 11,0,0) dissociation at 351 nm to be formed in ϭ0–4 vibrational states with a
v
2
spin–orbit ratio of P(2⌸3/2):P(2⌸1/2)ϭ3.8Ϯ0.5. The ClO fragment shows moderate rotational
excitation. The obtained ClO product state distributions and the relatively high translational energy
2
˜
of the fragments can be explained by predissociation of the (A A2␯1,0,0)-excited parent molecule
in the course of which the initial symmetric stretch motion ͑␯ ͒ of OClO is transferred into the
1
dissociative asymmetric stretching mode ͑␯ ͒. ClO line profile measurements indicate a dissociation
3
2
˜
time of less than 0.5 ps. Resulting from the OClO(A A218,0,0) dissociation at 308 nm ClO is
generated in very high vibrational states. The rotational excitation is comparable to that of the 351
nm photolysis study. © 1996 American Institute of Physics. ͓S0021-9606͑96͒00307-5͔
I. INTRODUCTION
subject of investigation. In their one-color resonance-
enhanced multiphoton ionization ͑REMPI͒ spectra between
By now, it is a well-known fact that the polar strato-
spheric ozone destruction is caused on a large scale by the
photochemistry of chlorine oxides.1–4
335 and 370 nm, Vaida et al. observed ClO from dissociation
2
of (A A2)-excited OClO in ϭ3–6 vibrational states.6,8
˜
v
However, ionic fragmentation of OClOϩ cannot be entirely
¨
In recent years, numerous attempts have been made to
understand the gas phase dynamics of OClO photodissocia-
tion from the first absorption band.5–11 This near-UV absorp-
tion spectrum extending from about 260 to 480 nm results
ruled out in this experiment. Recent results of Baumgartel
and Gericke,13 who applied the two-photon laser-induced
fluorescence (2h␯) LIF technique for ClO detection, indi-
cate that in the one-color near-UV photodissociation of
OClO the fragments are formed mainly in lower vibrational
states.
2
2
˜
˜
from the (A A2␯1 ,␯2 ,␯
X B10,0,0) electronic transi-
3
tion and shows considerable vibrational structure including
progressions of all three vibrational modes. Two decomposi-
Contrary to an early unsuccessful attempt in detecting
the ClO radical by the conventional one-photon LIF method
2
˜
tion pathways are observable for the (A A2)-excited parent
molecule leading to formation of ClOϩO ͑major channel͒
and ClϩO2 ͑minor channel͒, respectively. As examined by
Davis and Lee,9 the yield of the ClϩO2 product channel
reaches a maximum of 3.9Ϯ0.8% at a photolysis wavelength
of 404 nm and decreases at shorter wavelengths ͑e.g., Ͻ0.2%
in the range of 350–370 nm͒. Furthermore, a strong sensi-
tivity to the initial vibrational state of the electronically ex-
cited OClO molecule is seen in the branching ratio for the
two product channels.9,16 Davis and Lee9 found the forma-
tion of ClO tenfold greater from OClO with asymmetric
probing the predissociative (A ⌺) state of ClO,14 the (2h␯)
2
2
LIF technique uses the high-lying radiative (C ⌺) Rydberg
2
state by two-photon excitation of nascent ClO(X ⌸ , ,J)
v
⍀
fragments. With a more intricate experimental setup the
(C ⌺) state is also attainable in (1h␯) LIF detection of ClO
2
fragments in the VUV region using four-wave mixing as re-
ported recently by Matsumi et al.15 Although (2h␯) LIF rep-
resents an excellent way to achieve both vibrational and ro-
tational ClO spectral resolution, the one-color experiment in
the photodissociation of OClO still suffers from a particular
disadvantage.
stretching ͑␯ ͒ excitation than from neighboring vibrational
3
states with pure symmetric stretching ͑␯ ͒ excitation or sym-
As a consequence of scanning the photolysis wavelength
over a range of around 30 nm with regard to the absorption
spectrum of OClO, approximately ten different vibrational
(A A2␯1 ,␯2 ,␯3) states are passed. Under these conditions,
the initial state from which the photodissociation occurs is in
no way determined. However, it is an indelible fact that a
deep insight into the fragmentation dynamics is only avail-
able if a state-to-state experiment, requiring both the knowl-
edge of the initial and the final state, is carried out.
1
metric stretching and bending ͑␯ ͒ excitation. Bishenden and
2
Donaldson16 also reported that asymmetric stretching excita-
tion enhances the ClO production. A previously assumed
photoisomerization mechanism in the gas phase leading to
asymmetric ClOO and subsequently to formation of
ClϩO2,5,6 as obtained almost quantitatively from matrix-
isolated OClO molecules,12 has been shown to be extremely
unlikely under gaseous conditions and can therefore be
excluded.9 In our work, the ClOϩO product channel pre-
dominating in the near-UV OClO photofragmentation is the
2
˜
In order to meet this necessity, the OClO photodissocia-
tion reported in this paper is performed using constant wave-
J. Chem. Phys. 104 (8), 22 February 1996
0021-9606/96/104(8)/2883/8/$10.00
© 1996 American Institute of Physics
2883
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
2884
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2
˜
lengths of either 351 nm or 308 nm, exciting (X B10,0,0)
2
˜
ground state OClO either into the (A A211,0,0) or into the
2
˜
(A A218,0,0) state. Thus, the photo-induced decay of the
parent molecule is initiated at these particular photolysis
wavelengths by sole contribution of the symmetric stretching
mode ͑␯ ͒. Bishenden and Donaldson16 also detected ClO
1
2
˜
originating from distinct (A A2␯1 ,␯2 ,␯3) states of the
OClO parent molecule using a two-color REMPI experi-
ment. At a photolysis wavelength of 360 nm corresponding
2
˜
to (A A210,0,0)-excited OClO, they obtained ͑2ϩ1͒
REMPI spectra which indicate vibrational excitation of the
ClO fragments. However, the assignment of the observed
transitions proved to be a rather challenging task because
three close-lying electronic states, namely the D, E, and F
state of ClO, were involved in the spectral features. There-
fore, the exact determination of neither the vibrational nor
the rotational product state distribution was possible.
FIG. 1. ͑a͒ Two-photon laser-induced fluorescence spectrum of ClO origi-
nating from the OClO photolysis in a one-color experiment. ͑b͒ Two-color
(2h␯) LIF spectrum of ClO. The photolysis wavelength is fixed at 351 nm,
corresponding to an initial decay originating definitely from the excited
In the two-color (2h␯) LIF experiments described in
this work, the elementary dissociation process of
(A A2␯1,0,0)-excited OClO is investigated by the complete
product state analysis and by comparing the obtained results
with ab initio calculations of the upper potential surfaces.17
2
˜
(A 2A2,11,0,0) state of the parent molecule.
˜
tion beam entered the chamber focused in the same way by a
quartz lens ͑fϭ500 mm͒. At this point, the total enclosure of
the detection beam by the photolysis beam volume inside the
reaction cell was of great importance for providing actual
two-color experimental conditions. The dye laser bandwidth
of 0.4 cmϪ1 was optionally reducible to 0.08 cmϪ1 by addi-
II. EXPERIMENTAL DETAILS
According to the method of Derby and Hutchinson18
OClO was generated by pouring a 10% Cl2/N2 mixture
through a column containing NaClO2 and glass beads. The
actual yield of OClO was at least 70% as indicated by ab-
sorption measurements. Since the remainders, being Cl2 and
N2, show no influence on further events, the yellow product
was transferred to the photo-process without further steps of
purification.
A XeF excimer laser ͑Lambda Physik EMG 201 MSC͒
provided a photolysis wavelength of 351 nm. The OClO pho-
tolysis at 308 nm was carried out with a XeCl excimer laser
͑Lambda Physik EMG 101 MSC͒. The output energy of the
photolysis lasers was about 150 mJ of which 10% reached
the interior of the reaction chamber. A beam diameter of 0.02
cm2 and the corresponding OClO absorption coefficients
which are reported in Ref. 19 guaranteed the photolysis pro-
ceeding at saturation. When the probe laser was blocked,
ClO fragments were not observable in the focus of only the
photolysis laser. Furthermore, a variation of the photolysis
laser energy had no influence on the intensity of the observed
LIF signals. This fact and the relatively low photolysis en-
ergy of around 15 mJ in the reaction cell gives evidence
against our LIF signals resulting from multiphoton pro-
cesses. The probe energy of the tunable detection system
consisting of a dye laser ͑Lambda Physik FL 3002͒ pumped
by a XeCl excimer laser ͑Lambda Physik LPX 100͒ was
5–10 mJ, depending on the respective wavelength. The spec-
´
tional application of an intracavity etalon ͑Lambda Physik͒,
as necessary for rotational linewidth and Doppler profile
measurements.
The (2h␯) LIF spectrum ͓Fig. 1͑b͔͒ resulting from the
351 nm photolysis was recorded at a total cell pressure of 45
Pa and 100 ns delay between photolysis and detection.
Hence, collisions and secondary reactions affecting the na-
scent photolysis products were prevented and nascent prod-
uct state distributions were observable.
A solar-blind photomultiplier tube ͑EMR 542G-08-18͒
perpendicular to the laserbeams registered the fluorescence
light in the region of around 170 nm originating from
2
(C ⌺)-excited ClO radicals. The multiplier output signal
was integrated by a boxcar integrator and averager ͑Stanford
Research Systems SR 250͒ and after A/D conversion re-
corded by a personal computer ͑Siemens PC AT 386͒, the
latter also controlling the stepping motor for both grating and
´
etalon. Simultaneous recording of fluorescence as well as
laser intensity allowed to obtain normalized spectra. A home
made trigger device regulated the order of events. The jitter
between the photolysis and detection laser pulse was about
10 ns, primarily determined by the excimer laser.
At the photolysis wavelength of 308 nm a pressure of
200 Pa with 200 ns delay was additionally used. Under these
experimental conditions, the rotational state distribution is
influenced by collision processes whereas the fragment vi-
bration is unaffected. This was necessary because the spec-
2
2
tral extent of the observable ClO (C ⌺ X ⌸⍀) rovi-
bronic transitions required the usage of p-terphenyl in diox-
ane and RDC 360-neu ͑Radiant Dyes Laser & Acc. GmbH͒
in dioxane as active laser media. A quartz lens ͑fϭ500 mm͒
was employed to focus the photolysis beam into the reaction
chamber. Counterpropagating and time delayed, the detec-
tral features of the observable lower (0
)-transitions de-
v
creased as a consequence of the stronger vibronic excitation
of the fragments. Since the generation of vibrationally ex-
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2885
FIG. 2. Vibrational product state distribution of the ClO (X 2⌸⍀) radicals
generated in the 351 nm photodissociation of OClO.
FIG. 3. Boltzmann plot of the ͑0-0, ⍀ϭ3/2͒-transition in Fig. 1͑b͒.
cited ClO up to ϭ17 is energetically possible in the 308 nm
v
photolysis, the vibrational distribution of ClO including
higher vibronic levels was further investigated in an addi-
tional two-color ͑2ϩ1͒-REMPI/TOF ͑time-of-flight͒
experiment.20 OClO was prepared as described above and
carried to photolysis via a pulsed molecular beam nozzle.
The O(3P) partner fragments generated at a photolysis
wavelength of 308 nm were detected by resonant two-photon
nificant difference exists between this spectrum and that ob-
tained from the one-color experiment in Fig. 1͑a͒. In the
one-color OClO photodissociation, the ClO fragments are
formed mainly in lower vibrational states whereas the frag-
mentation process of OClO in the (A A211,0,0) state leads
to ClO radicals which are vibrationally hot.
The quantitative evaluation of the vibrational features in
Fig. 1 is based on the following consideration. The depen-
dence of the intensity of an isolated transition
2
˜
excitation of the 2p P2 3p PJ(Jϭ0,1,2) transition21 and
subsequent ionization by an additional photon at a detection
wavelength of 226 nm. The signal of the ions drifting to the
detector was then recorded. The pressure in the reaction
chamber was around 5ϫ10Ϫ3 Pa allowing the measurement
of the nascent translational energy distribution. A detailed
description of the experimental apparatus is given in Ref. 22.
3
3
( Ј,JЈ) ( ,J) on both the transition probability and the
v
v
initial population of the respective rovibronic level ( ,J) is
v
given by
v
v
Ј
Ј
IvJ J ϰBJ Jv•P ,J͒.
͑1͒
Ј
Ј
͑v
The desired vibrational populations P( ) are obtained by
integrating over the entire rotational envelope of the respec-
tive vibrational band
v
III. RESULTS
In Fig. 1͑a͒ the one-color (2h␯) LIF spectrum is shown.
The very efficient laser dye RDC 360-neu ͑Radiant Dyes
Laser & Acc. GmbH͒ was employed, resulting in higher in-
tensities for the ͑0 2͒- and the ͑0 3͒-band in comparison
to the recently reported one-color experiment.13 The normal-
ized (2h␯) LIF spectrum of nascent ClO resulting from the
two-color experimental setup at a fixed OClO photolysis
wavelength of 351 nm is depicted in Fig. 1͑b͒. Solely
v
v
Ј
Ј
IJ J ␯͒d␯ϰ•FC
͑v v
͑2͒
͑
͑v
͒•P ͒,
Ј
͵
␯͑v v͒
Ј
at which FC( Ј ) describes the corresponding Franck–
v v
Condon factors of the ClO (C X) vibrational transitions
which are taken from Ref. 23.
Figure 2 shows that the vibrational state distribution of
nascent ClO is inverted, possessing a maximum population
(0
)-transitions are observed. The assignment of the ap-
v
pearing spectral features is possible by comparison with the
simulated rovibrational transitions. Quite obviously, a sig-
at ϭ2. The vibrational product state distribution, the em-
v
ployed Franck–Condon factors as well as the ratio of the
populations of the spin–orbit systems are listed in Table I.
The population ratios of the spin–orbit systems are ob-
tained by separate integration over the corresponding lines.
Concerning the spin–orbit subbands of the ͑0 0͒-, ͑0 2͒-,
and the ͑0 3͒-vibronic transition; the population ratios
P(2⌸3/2):P(2⌸1/2) are readily obtainable from the spectral
features. The ͑0 1, ⍀ϭ3/2͒-subband is partially covered by
the ͑0 2͒-transition and the ͑0 4, ⍀ϭ3/2͒-subband is too
weak. Since the obtained values of the populations of the
spin–orbit ratios are identical within the limit of error, simi-
lar conditions are assumed in the case of the ͑0 1͒- and the
͑0 4͒-transition. Thus, in the OClO photolysis at 351 nm
TABLE I. Franck–Condon factors, populations, and spin–orbit ratios of the
observed vibrational transitions in the photolysis of OClO at 351 nm.
FCa
P( )exp/%
v
P( )bPrior/%
P(⌸3/2):P(⌸1/2)
vЈ
v
v
0
0
0
0
0
0
0.3107
0.3412
0.2074
0.0924
0.0336
6
5
52
27
10
20
17
15
13
10
4.32
͑3.81͒
3.76
3.35
͑3.81͒
1
2
3
4
aReference 23.
bRigid rotator harmonic oscillator ͑RRHO͒ approximation.
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2886
The respective transitions are simulated using the ClO mo-
lecular constants which are measured by Coxon24 assuming
thermal rotational state distributions of different tempera-
tures. The transitions calculated in this way are compared
with the observed experimental features. In each case, the
best correspondance is acquired at about 1000Ϯ300 K.
A further experiment serves as an indication for the ad-
equacy of the applied (2h␯) LIF method. Selecting a delay
of 30 ␮s between photolysis and ClO detection and increas-
ing the total cell pressure with argon up to 2 kPa should lead
to sole observation of completely relaxed product molecules;
in particular, this means no vibronic excitation as well as a
rotational state distribution corresponding to that at room
temperature. Both exclusive appearance of the ͑0 0͒-
transition and a very good agreement of obtained and ex-
pected rotational state distribution are found.
Figure 4͑a͒ presents an enlarged part of the ͑0 2, ⍀ϭ3/
2͒-transition of the spectrum in Fig. 1. The single R(48)
rotational line is depicted in Fig. 4͑b͒. Each high resolution
scan is recorded with the etalon equipped detection system at
a reduced laser bandwidth of 0.08 cmϪ1. The width as well
as the shape of the Doppler profile of the R(48) line are
determined from a least square fit. The applied fit procedure
which is described in Ref. 25 yields an experimental line-
width ͑FWHM͒ of 0.166Ϯ0.01 cmϪ1. This is in excellent
agreement with the theoretically calculated linewidth of
0.163 cmϪ1 which is obtained by considering the available
energy and the rotational energy of JClOϭ48.
FIG. 4. ͑a͒ High resolution scan of the ͑0-2, ⍀ϭ3/2͒-transition obtained
from the two-color experiment at 351 nm. The application of an etalon
reduced the laser bandwidth to 0.08 cmϪ1. ͑b͒ Enlarged view of the R(48)
rotational line, which is marked in ͑a͒. The analysis of the Doppler profile
yielded an upper limit of 0.5 ps for the lifetime of the excited OClO mol-
ecule.
the ClO spin–orbit system ⍀ϭ3/2 is preferred with a popu-
lation ratio P(2⌸3/2):P(2⌸1/2) of 3.8Ϯ0.5. In addition, the
2
(2h␯) LIF spectrum of the single ClO ͑C ⌺,
2
ϭ0 X ⌸ , ϭ0͒ transition recorded under completely
vЈ
v
⍀
rotationally and vibrationally relaxed conditions ͑30 ␮s delay
and 2 kPa͒ yields a ratio of 4.00. This result is in good
From the line shape, the anisotropy parameter ␤ is
deduced,25 giving evidence of the correlation between the
molecular vectors ␮parent and vfragment . An experimental value
of ␤expϭ1.14 is obtained. Provided that the initial alignment
agreement
with
the
theoretically
expected
P(2⌸3/2):P(2⌸1/2) value of 4.5 calculated for a Boltzmann
distribution at 300 K. The spectroscopic properties of the
ClO radical applicable for two-photon excitation are de-
scribed in the previous work.13
between ␮
and vfragment is fixed, a theoretical value
parent
␤
␤
theoϭ1.73 is calculated. The experimental parameter
expϭ1.14 resulting from the Doppler profile indicates that
In obtaining the rotational state distributions simulated
vibrational bands are adapted to the measured ones using a
least square fit procedure. The theoretical intensity at a fre-
quency ␯ is given by
the initial ␮, v-correlation is influenced by the motion of the
excited parent molecule during the fragmentation process. As
described in Ref. 26, the deviation of the anisotropy param-
eter ␤ from the theoretically expected value can be used to
estimate the lifetime ␶ of the electronically excited parent
molecule. A lifetime of ␶ϭ0.5 ps is calculated for the
Itheo ␯͒ϰ
P q͒•S •L ␯q qϪ␯,⌬␯͒,
͑3͒
͑
͑
͑
͚
q q
Ј
Ј
q ,q
Ј
2
˜
in which P(q) is the desired quantum state population,
(A A211,0,0)-excited OClO molecule. However, it is im-
portant to mention that geometrical rearrangements of the
dissociation complex as well as coupling with the electronic
(2A1) and (2B2) states also strongly reduce the ␤theo param-
eter. As a consequence, the calculated lifetime presents an
upper limit. Nevertheless, the obtained value is in agreement
with recent OClO femtosecond investigations, in which the
Sq q the transition probability, and L(␯
Gaussian line shape function centered at ␯
Ϫ ␯,⌬␯) the
and possessing
q q
Ј
Ј
q q
Ј
a linewidth ͑FWHM͒ of ⌬␯. Summation is carried out over
all transitions (qЈ q) contributing to the signal at the re-
spective frequency. The population P(q) of each quantum
state q is obtainable in this way.
For the spin–orbit subband ⍀ϭ3/2 of the ͑0 0͒-
transition, the nascent population of rotational states is found
to correspond to a Boltzmann distribution with a temperature
2
27
˜
lifetime of (A A2)-excited OClO was measured to 0.5 ps.
The photodissociation of OClO at a constant wavelength
of 308 nm corresponds to a higher symmetric stretch ͑␯ ͒
1
parameter of 480Ϯ30 K. The Boltzmann plot of nascent
excitation of the parent molecule. In this case, essentially
␯1ϭ18 and to a smaller extent ␯1ϭ17 is excited.19 The na-
scent rotational state distributions which are observed in the
308 nm (2h␯) LIF spectrum at 45 Pa and 100 ns delay
appear similar to those obtained from the 351 nm photolysis.
2
ClO͑X ⌸3/2
, ϭ0͒ originating from the 351 nm photolysis
v
is shown in Fig. 3. Unfortunately, precise molecular con-
stants of the higher vibronic levels of the electronic
2
ClO(X ⌸) ground state do not exist so far. Therefore, popu-
lations of rotational states belonging to Ͼ0 are not avail-
The vibrational product state distribution P( ) of the
v
v
able by a least square fit procedure but estimated as follows.
ClO fragments for the five lowest vibrational levels is pre-
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2887
involve excitation from the 1a2 , 8a1 , and 5b2 orbitals into
2
˜
the singly occupied 3b1 level. The transition from the X B1
2
state to the B2 state is dipole forbidden in C2v geometry.
2
2
2
2
˜
˜
˜
Both the A1 X B1 and the A A2 X B1 transition are
dipole allowed but the former is of perpendicular polariza-
tion and should be weak. In fact, spectroscopic experiments
2
observe only excitation to the A A2 state.6,8,10,27 The
˜
near-UV absorption spectrum of chlorine dioxide28 shows
sharp features originating from a strong progression of the
symmetric stretching mode ͑␯ ͒ and combinations of the
1
symmetric stretching mode with the bending mode ͑␯ ͒ and
2
the asymmetric stretching mode ͑␯ ͒.
3
A look at the OClO UV absorption spectrum indicates
that the applied photolysis wavelengths of both 351 nm and
308 nm exclusively affect the symmetric stretching mode
FIG. 5. Vibrational product state distribution of the first five vibronic levels
of ClO as a result of the 308 nm OClO photolysis.
2
˜
͑␯ ͒ of the OClO(A A2) parent molecule. At 351 nm only
1
␯1ϭ11 and at 308 nm primarily ␯1ϭ18 participates in the
excitation process. Resulting from this particular initial sym-
metric stretch excitation, the ClO radicals are formed in
sented in Fig. 5. Except for the population of the ϭ0 vi-
bronic level, the population of the vibrational states increases
v
ϭ0–4 vibrational states for the 351 nm photolysis and
v
continually up to ϭ4. Higher vibronic levels are not ob-
v
significantly vibrationally hotter for the 308 nm photolysis.
Both photolysis wavelengths generate the ClO fragments
with moderate rotational excitation.
The available energy Eav which has to be released in the
products of a photodissociation process is defined by
servable because the increase of the population numbers is
too weak to compensate the strongly decreasing Franck–
Condon factors. However, the line intensities are signifi-
cantly lower than those observed in the corresponding one-
color spectrum, indicating the formation of rather strong
vibronically excited ClO radicals.
Eavϭh␯ϩEint͑OClO)ϪD0 ,
͑4͒
In order to determine the ClO population in higher vi-
in which h␯ represents the energy of the photons used for the
photolysis, Eint͑OClO͒ describes the initial internal energy of
the parent molecule before excitation ͑EintϷErotϭ3/2RT͒
and D0 is the dissociation energy of the OClO molecule.
With a D0 value of 231Ϯ8 kJ/mol ͑Ref. 29͒ the energy avail-
able for the products at a photolysis wavelength of 351 nm
amounts to 114Ϯ8 kJ/mol. The complete product state analy-
sis, i.e., the determination of the internal energy of the ClO
brational levels than ϭ4 a two-color REMPI/TOF study is
v
performed at which the time-of-flight of the recoiled O(3P)
atoms is measured. The internal energy distribution of ClO
obtained from the observed time-of-flight distribution of the
O(3P) partner fragments is represented in Fig. 6. Although
the vibrational structure is not resolved, it is obvious from
the spectrum that ClO with an internal energy of up to at
least 8000 cmϪ1 is generated in the 308 nm photolysis. This
result corresponds to ClO fragments which are significantly
vibrationally excited and agrees with the tendency observed
in the 308 nm (2h␯) LIF experiment. The kinetic energy of
the O atoms corresponding to higher internal energies of ClO
is too low and hence the ions are not detectable in the drift
experiment. Detailed investigations of the OClO photolysis
using the REMPI/TOF method are in progress and will be
published elsewhere.20
IV. DISCUSSION
The ground state of OClO has the configuration in C2v
geometry17
2
2
2
2
X B1 2A
:
5b ͒ 8a ͒ 1a ͒ 3b ͒1,
͑ ͑ ͑ ͑
2 1 2 1
˜
͓
͔
Љ
at which the term in square brackets [2AЉ] corresponds to Cs
geometry of the molecule. The first three excited electronic
states,
FIG. 6. Internal energy distribution of ClO obtained from the time-of-flight
distribution of the O(3P) partner fragments. The recorded data points be-
come uncertain beyond 8000 cmϪ1 due to very low recoil velocities of the
O(3P) atoms corresponding to high internal excitation of the ClO frag-
ments.
2
2
2
1
A A2 2A
:
5b ͒ 8a ͒ 1a ͒ 3b ͒2,
͑ ͑ ͑ ͑
2 1 2 1
˜
͓
͔
Љ
2A1 2A
:
5b ͒ 8a ͒ 1a ͒ 3b ͒2,
͑ ͑ ͑
2 1 2 1
2
1
2
͓
͔
͔
͑
͑
Ј
Ј
2B2 2A
:
5b ͒ 8a ͒ 1a ͒ 3b ͒2,
͑ ͑ ͑
2 1 2 1
1
2
2
͓
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
2888
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
TABLE II. Distribution of the available energy over the fragments obtained in the state-to-state dissociation
2
OClO ClO (X
⌸
, ,J)ϩO(3P) at a constant photolysis wavelength of 351 nm ͑Eavϭ9540 cmϪ1͒.
v
⍀
͗Evib͘/cmϪ1
͗Erot͘/cmϪ1
͗Etrans͘/cmϪ1
fvib/%
frot/%
ftrans/%
Experiment
Prior
1920
2767
675
2767
6945
4006
20
29
7
29
73
42
3
2
2
˜
fragments, is possible for the 351 nm OClO photolysis. The
distribution of the available energy over the degrees of free-
dom of the fragments originating from (A A211,0,0) ex-
cited OClO is summarized in Table II.
RClO leads to O( P)ϩClO͑ ⌸͒. However, both the A A2
͓ AЉ in Cs geometry͔ and the A1 [2AЈ] surface exhibit a
barrier and only the unbounded B2 [ AЈ] surface is repul-
sive with respect to RClO ͑Fig. 7͒. For this reason, at low
excitation energies OClO can only decay on the B2 [ AЈ]
surface and Peterson and Werner17 proposed the following
three-step fragmentation mechanism promoted by the asym-
metric stretch: Internal conversion from the initially excited
2
2
2
2
2
˜
2
2
Provided that the internal energy of ClO is known, the
measured linewidth ͑FWHM͒ of the rotational line depicted
in Fig. 4͑b͒ allows the determination of the dissociation en-
ergy of OClO. The Doppler width of the R(48) transition is
measured at 0.166Ϯ0.01 cmϪ1 correlating with a ClO veloc-
ity of 880Ϯ50 m/s and an O velocity of 2830Ϯ170 m/s,
respectively. Thus, the entire translational energy of both
fragments amounts to 84Ϯ10 kJ/mol leading to an Eav value
of 121Ϯ10 kJ/mol and a D0 value of 224Ϯ10 kJ/mol. The
excellent agreement between D0 determined from the R(48)
linewidth and the D0 value ͑231Ϯ8 kJ/mol͒ in Ref. 29 addi-
tionally supports the application of the (2h␯) LIF technique
in the state-resolved detection of nascent ClO radicals.
Cuts through ab initio potential energy surfaces of the
OClO molecule have been calculated by Werner et al.17 and
are shown in Fig. 7 with respect to the symmetric, the asym-
metric bond length distortion, and the variation of the va-
2
2
2
2
˜
A A2 [ AЉ] state to the A1 [ AЈ] state through spin–orbit
coupling, followed by vibronic coupling of the A1 [2AЈ]
2
state with the reactive B2 [2AЈ] state via the asymmetric
2
2
2
stretch, and finally decay of OClO on the B2 [ AЈ] surface
into the OϩClO fragments. Since the asymmetric stretch is
the dissociative mode, excitation of combination bands of
symmetric and antisymmetric stretch should promote the dis-
sociation process. This behavior was experimentally
observed.9,16
The A A2 surface which is directly reached in the
present experiment exhibits a barrier of about 0.4 eV with
respect to asymmetric bond length distortion. Since the mini-
mum of the A A2 state is at 2.7 eV ͑right part of Fig. 7͒,
energies above 3.1 eV may be sufficient to overcome that
barrier. Indeed, the absorption spectrum of jet cooled OClO
shows increasing homogeneous linewidth broadening with
2
˜
2
˜
2
2
2
˜
lence angle. The A A2 , A1 , and B2 surfaces are bonding
with respect to both the symmetric O–Cl–O distance rClO
and the valence angle. Asymmetric bond length distortion
FIG. 7. Potential energy surface of several electronic states of the OClO molecule calculated by Peterson and Werner ͑Ref. 17͒. Depicted are one-dimensional
cuts with respect to symmetric distortion of both Cl–O bonds ͑left part, valence angle fixed at 106.4°͒, variation of the valence angle ͑middle part, rClOϭ1.56
Å͒, and asymmetric bond distortion of one Cl–O bond ͑right part, valence angle fixed at 117.4 and rClOϭ1.56 Å͒.
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2889
2
˜
˜
2
2
˜
˜
increasing excitation of the ␯ mode of the A A2 state
1
˜
X B1 0,0,0͒ϩh␯ A A2 1ӷ0,0,0͒
͑
͑v
2
2
͓␯1ϭ6, 7, 8 corresponds to (A A2␯1,0,0) (X B10,0,0)
excitation energies of 3.09 eV, 3.20 eV, 3.29 eV͔.36 At the
excitation wavelengths of our study ͑351 nmϵ3.53 eV, 308
nmϵ4.03 eV͒, the strongly vibrating OClO will be above the
barrier and the observed ClO products could be generated via
2
˜
X B1 ӷ0, ӷ0, 3ӷ0͒
͑v1 v2
v
2
3
ClO ⌸, ,J͒ϩO P͒.
͑
v
͑
In this case, the intermediate is a vibrationally hot OClO
molecule in the ground state. Donald and Innes30 concluded
from line width measurements that this decay channel is un-
likely at low energies. However, at higher excitation energies
the number of states increases considerably and the decay
rate via this channel should increase.27 A RRKM calculation
of Baumert et al.27 shows that the decay time decreases from
130 fs to 90 fs when the dissociation wavelength is varied
from 350 nm to 310 nm.33 This decay mechanism implies a
broad energy distribution within the X B1 ground state and
one might expect a statistical energy distribution for the ClO
product. A simple prior distribution is calculated using the
rigid rotator harmonic oscillator approximation ͑RRHO͒.
The results ͑Tables I, II; 351 nm photolysis͒ indicate that the
experimental distribution strongly deviates from this statisti-
cal distribution. The photodissociation of OClO at 308 nm
leads to formation of vibrationally extremely hot ClO radi-
cals ͑Fig. 6͒. The release of the available energy is very
specific and strongly deviates from a statistical distribution.
The ClO is generated essentially in very high vibrational
states. The rotational excitation is rather low and comparable
to that of the photolysis at 351 nm.
2
˜
vibrational predissociation of the A A2 state along the ␯
3
coordinate. Although the system is bound along the initially
excited symmetric stretching mode ␯1, the slightest coupling
between ␯ and ␯ is sufficient to achieve an asymmetric
1
3
system and energy will be transferred into the asymmetric
nuclear motion. As soon as sufficient energy is transferred
into the asymmetric bond length distortion in order to over-
come the barrier, the OClO molecule will decay into a vibra-
tionally excited ClO fragment and an O atom. If the energy
released in descending the exit barrier is essentially trans-
ferred into translation and rotation ͑but not vibration͒ then
one would expect highly vibrating ClO products where the
vibrational energy is essentially given by the energy differ-
2
˜
2
˜
ence between the excitation energy to reach the A A2 state
and the total height of the barrier. Due to the angular geom-
etry of OClO it is very likely that a simple repulsion of the O
atom will essentially induce translation and rotation but only
to a minor part vibration. At the photolysis wavelength of
351 nm the maximum of the ClO vibrational state distribu-
tion is found to be at ϭ3 corresponding to an energy of
v
ϳ0.4 eV. This energy is the difference between the photon
energy, 3.53 eV, and the total energy required to overcome
the barrier, 3.1 eV. The energy barrier is supposed to be
overcome the faster the higher the ␯1 mode is excited and the
vibrational excitation of ClO may be attributed to the initial
high vibronic excitation of the (A A2␯1,0,0) parent mol-
ecule. As a consequence, the photodissociation of OClO at
308 nm should generate vibrationally extremely hot ClO
radicals which, indeed, is observed in our experiment ͑Fig.
6͒.
The formation of ClO in very high vibrational states di-
rectly explains the results of Glownia et al.34 who used a
femtosecond pump-and-probe experiment to observe ClO
formation after OClO excitation at 308 nm. The build up
time of ClO was in the order of a nanosecond. This cannot be
explained by a long living intermediate but by collisional
relaxation of the vibrationally excited ClO.6
The observed extremely high vibrational excitation of
ClO at short photodissociation wavelengths may have an im-
pact on our understanding of the atmosphere because new
reaction channels are open when ClO is formed in high vi-
brational states. For example, the reaction of ClOϩN2 is en-
dothermic by 8500 cmϪ1 ͑Ref. 35͒ whereas the internal en-
ergy of the ClO fragment generated in the photodissociation
of OClO at 308 nm is even higher.
2
˜
A decay along ␯3 is also indicated by the observed mod-
erate rotational excitation of the ClO radicals. The rotational
state distribution of the fragments is probably initiated by the
decrease of the bond angle from around 117.6° to 107° oc-
2
2
˜
˜
curring in the electronic transition between X B1 and A A2
͑Ref. 32͒ and by the repulsion on the upper potential of
OClO leading to the OϩClO fragments. Vibronic coupling
involving the bending mode should lead to a much higher
rotational excitation of the ClO fragments than is observed in
the experiment. The minimum energy of the 2A1 surface with
respect to the valence angle is obtained for a linear O–Cl–O.
The torque which is induced on this surface by changing the
In summary, the photolysis at 351 nm and 308 nm ex-
2
2
˜
˜
cites the electronic OClO (A A2 X B1) transition with
simultaneous powerful excitation of only symmetric stretch-
2
˜
ing modes of the A A2 state. Slight disturbances provide the
fast vibrational predissociation of the parent molecule on the
2
A A2 (2AЉ) surface into the fragments, ClO(2⌸, ,J)
˜
v
ϩO(3P), of which the ClO is generated with high vibra-
tional but only moderate rotational energy.
2
˜
valence angle from 107° ͑minimum on the A A2 surface͒ to
180° should result in highly rotating ClO products. Accord-
ing to the calculations of Peterson and Werner, the energy
ACKNOWLEDGMENTS
2
2
˜
difference between A A2(107°) and A1(180°) amounts to
ϳ0.36 eV. However, the observed average ClO rotational
energy is 0.08 eV ͑Table II͒.
This work was supported by the Deutsche Forschungs-
gemeinschaft. S. B. thanks the Konrad-Adenauer-Stiftung for
fellowship support. Special thanks to T. Haas, C. Maul, and
M. Roth for the recording of the very helpful REMPI/TOF
spectra. We thank Professor Dr. F. J. Comes for material
support and helpful discussions.
In principle, another fundamental pathway involving
2
2
˜
˜
coupling of the (A A2) state with the (X B1) ground state
is possible,27
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25

¨
Delmdahl, Baumgartel, and Gericke: Dissociation of OClO
2890
1 S. Solomon, Nature 347, 347 ͑1990͒.
20 C. Maul, M. Roth, and K.-H. Gericke ͑unpublished͒.
21 T. Kinugawa, T. Sato, T. Arikawa, Y. Matsumi, and M. Kawasaki, J.
Chem. Phys. 93, 3289 ͑1990͒.
2 R. S. Stolarski and R. J. Cicerone, Can. J. Chem. 52, 1582 ͑1974͒.
3 J. G. Anderson, W. H. Brune, and M. H. Proffitt, J. Geophys. Res. 94,
11 465 ͑1989͒.
22 C. Maul, T. Haas, and K.-H. Gericke, J. Chem. Phys. 192, 3238 ͑1995͒.
23 J. B. Nee and K. J. Hsu, J. Photochem. Photobiol. A: Chem. 55, 269
͑1991͒.
4 G. L. Manney, L. Froidevaux, and J. W. Waters, Nature 370, 429 ͑1994͒.
5
¨
V. Vaida, S. Solomon, E. C. Richard, E. Ruhl, and A. Jefferson, Nature
342, 405 ͑1989͒.
24 J. A. Coxon, Can. J. Phys. 57, 1538 ͑1979͒.
6
E. Ruhl, A. Jefferson, and V. Vaida, J. Phys. Chem. 94, 2990 ͑1990͒.
25 K.-H. Gericke, T. Haas, M. Lock, R. Theinl, and F. J. Comes, J. Phys.
Chem. 95, 6104 ͑1991͒.
¨
7 E. Bishenden, J. Haddock, and D. J. Donaldson, J. Phys. Chem. 95, 2113
͑1991͒.
26 S.-C. Yang and R. Bersohn, J. Chem. Phys. 61, 460, 4400 ͑1974͒.
27 T. Baumert, J. L. Herek, and A. H. Zewail, J. Chem. Phys. 99, 4430
͑1993͒.
8 V. Vaida, E. C. Richard, A. Jefferson, L. A. Cooper, R. Flesch, and E.
¨
Ruhl, Ber. Bunsenges. Phys. Chem. 96, 391 ͑1992͒.
9 H. F. Davis and Y. T. Lee, J. Phys. Chem. 96, 5681 ͑1992͒.
28 D. S. Bethune, A. J. Schell-Sorokin, J. R. Lankard, M. M. T. Loy, and P.
P. Sorokin, in Advances in Laser Spectroscopy, edited by Garetz and Lom-
bardi ͑Heyden, Philadelphia, 1983͒, Vol. 2, p. 1.
10
¨
R. Flesch, B. Wassermann, B. Rothmund, and E. Ruhl, J. Phys. Chem. 98,
6263 ͑1994͒.
11 A. J. Colussi and S. P. Sander, Chem. Phys. Lett. 187, 85 ͑1991͒.
29 I. P. Fischer, Trans. Faraday Soc. 63, 684 ͑1967͒.
30 D. A. McDonald and K. K. Innes, Chem. Phys. Lett. 59, 562 ͑1987͒.
31 S. Michielsen, A. J. Merer, S. A. Rice, F. A. Novak, K. A. Freed, and Y. J.
Hamada, J. Chem. Phys. 74, 3089 ͑1981͒.
12 A. Arkell and I. Schwager, J. Am. Chem. Soc. 89, 5999 ͑1967͒.
13
¨
S. Baumgartel and K.-H. Gericke, Chem. Phys. Lett. 227, 461 ͑1994͒.
14 M. A. A. Clyne, I. S. McDermid, and A. H. Curran, J. Photochem. 5, 201
͑1976͒.
32 J. L. Gole, J. Phys. Chem. 84, 1333 ͑1980͒.
15 Y. Matsumi, S. M. Shamsuddin, and M. Kawasaki, J. Chem. Phys. 101,
8262 ͑1994͒.
33 The calculated decay times just serve as an estimation of the order of
magnitude.
16 E. Bishenden and D. J. Donaldson, J. Chem. Phys. 101, 9565 ͑1994͒.
17 K. A. Peterson and H. J. Werner, J. Chem. Phys. 96, 8948 ͑1992͒.
18 R. I. Derby and W. S. Hutchinson, Inorg. Synth. 4, 152 ͑1953͒.
19 A. Wahner, G. S. Tyndall, and A. R. Ravishankara, J. Phys. Chem. 91,
2734 ͑1987͒.
34 J. H. Glownia, J. Misewich, and P. P. Sorokin, in Supercontinuum Lasers,
edited by R. R. Alfano ͑Springer, Berlin, 1990͒, p. 370.
35 D. L. Baulch, R. A. Cox, R. F. Hampson, Jr., J. A. Kerr, J. Troe, and R. T.
Watson, J. Phys. Chem. Ref. Data 13, 1259 ͑1984͒.
36 E. C. Richard and V. Vaida, J. Chem. Phys. 94, 163 ͑1990͒.
J. Chem. Phys., Vol. 104, No. 8, 22 February 1996
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:
128.114.34.22 On: Mon, 24 Nov 2014 21:37:25