Short-wavelength photolysis of jet-cooled OClO(2A2 v1>20) →ClO(X2ΠΩ,v,J) + O(3PJ)

Article abstract of DOI:10.1063/1.1367393

Velocity map imaging was used to study the energy partitioning and anisotropy for the ClO (X²Π)+O (³P₂) product channel. Very high anisotropy parameters and extreme channeling of available energy were found. This observation indicates a fundamental change of the fragmentation dynamics of OClO near UV absorption band.

Full text of DOI:10.1063/1.1367393

Short-wavelength photolysis of jet-cooled OClO ( 2 A 2 ν 1 >20)→ ClO (X 2 Π Ω ,v,J)+ O
3 P J )
(
Ralph F. Delmdahl, David H. Parker, and André T. J. B. Eppink
Citation: The Journal of Chemical Physics 114, 8339 (2001); doi: 10.1063/1.1367393
View online: http://dx.doi.org/10.1063/1.1367393
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/114/19?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Communication: Direct measurements of nascent O(3P0,1,2) fine-structure distributions and branching ratios of
correlated spin-orbit resolved product channels CO(ã 3Π; v) + O(3P0,1,2) and CO( X
VUV photodissociation of CO2
̃ Σ + 1 ; v) + O(3P0,1,2) in
J. Chem. Phys. 140, 231101 (2014); 10.1063/1.4883515
The UV photodissociation dynamics of ClO radical using velocity map ion imaging
J. Chem. Phys. 123, 174303 (2005); 10.1063/1.2083487
Photolysis of oxalyl chloride ( ClCO ) 2 at 193 nm: Emission of CO (v6,J60) detected with time-resolved Fourier-
transform spectroscopy
J. Chem. Phys. 120, 6957 (2004); 10.1063/1.1669390
Vibrationally mediated photolysis dynamics of H 2 O in the v OH =3 manifold: Far off resonance
photodissociation cross sections and OH product state distributions
J. Chem. Phys. 110, 8564 (1999); 10.1063/1.478764
Photofragment excitation spectrum for O(1D) from the photodissociation of jet-cooled ozone in the wavelength
range 305–329 nm
J. Chem. Phys. 106, 6390 (1997); 10.1063/1.473629
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:
131.193.242.21 On: Thu, 27 Nov 2014 05:17:33

JOURNAL OF CHEMICAL PHYSICS
VOLUME 114, NUMBER 19
15 MAY 2001
2
Short-wavelength photolysis of jet-cooled OClO„ A ␯ Ì20…
2
1
2
3
\
ClO„X ⌸ ,v,J…¿O„ P …
⍀
J
a)
Ralph F. Delmdahl and David H. Parker
Department of Molecular and Laser Physics, University of Nijmegen, Toernooiveld 1,
NL-6525 ED Nijmegen, The Netherlands
Andr e´ T. J. B. Eppink
Faculty of Physics, University of Bielefeld, Universit a¨ tsstraße 25, D-33615 Bielefeld, Germany
͑
Received 7 December 2000; accepted 5 March 2001͒
Highly inverted vibrational level populations are found for ClO fragments resulting from the UV
2
2
3
photodissociation of OClO( A ␯ Ͼ20) into ClO(X ⌸ ,v,J) and O( P ) fragments. These
2
1
⍀
J
3
distributions depend significantly on the spin–orbit J state of the oxygen O( P ) partner atom. In
J
contrast, the ClO rotational excitation is modest. Distinct rotational structure is visible in the O
3
2
(
P ) photofragment yield spectrum recorded from the highly excited OClO ( A ␯ ϭ21) vibronic
J
2
1
band, which is indicative for hitherto unforeseen long dissociation lifetimes of very highly excited
OClO. The data point towards an unexpected nearly-linear and highly asymmetric dissociation
2
geometry. Carrying out near-threshold fragmentation experiments of OClO ( A 0,0,0)
2
2
3
→
ClO (X ⌸3 vϭ0, J)ϩO( P
) the dissociation energy D of OClO has been accurately
/2
2,1,0 0
determined to 247.3Ϯ0.5 kJ/mol. © 2001 American Institute of Physics.
͓
DOI: 10.1063/1.1367393͔
to affect the ozone layer.27 However, due to the negligible
fraction of the Cl channel found in gas-phase photodissocia-
tion of OClO, the atmospheric role of this trace has been
assumed constituent so far to be restricted to a nighttime
I. INTRODUCTION
The stratospheric trace gas chlorine dioxide ͑OClO͒ is a
strong absorber in the near ultraviolet ͑UV͒ spectral region
between 260 and 480 nm with the Franck–Condon maxi-
28
reservoir for inorganic chlorine.
Ϫ17
2
Ϫ1
1
mum at ϳ350 nm(␴ϳ1.2ϫ10
cm molecule ). It has
It is interesting to note at this point that several recent
experimental studies investigating OClO fragmentation have
confirmed that increasing amounts of internally excited ClO
been the subject of numerous laboratory and theoretical stud-
ies triggered by its potential role in stratospheric ozone
2
–5
loss.
Several sensitive fragment detection techniques
radicals are produced at photolysis wavelengths in the range
ranging from resonantly enhanced multiphoton ionization
of 330–308 nm.6
,10,16
In a recent state-to-state study of OClO
6
–8
9–12
͑
REMPI͒, laser induced fluorescence ͑LIF͒,
photofrag-
real-time femto-
and velocity map imaging have been
1
3–17
using velocity map imaging it has been shown that a virtu-
ally complete conversion of the surplus energy into high
fragment vibration takes place. ClO fragments are formed
almost exclusively in the three highest energetically possible
vibrational levels v ͑ClO͒ϭ19, 20, and 21 in the dissociation
ment translational energy spectroscopy,
1
8,19
20
second studies,
applied so far in addressing the near UV photodissociation of
gaseous OClO, which proceeds via the two fragmentation
channels:
2
0
process at a photolysis wavelength of 273.5 nm. At higher
photolysis energies the ClOϩO channel is known to be the
only one existent, and the internal energies found in the ClO
fragments are sufficient to thermodynamically open several
atmospherically important reactions. Implications for atmo-
2
3
CIO͑X ⌸͒ϩO͑ P ͒,
͑1͒
͑2͒
J
2
1
3
Ϫ 3
ϩ
Cl͑ P͒ϩO ͑ ⌬, ⌺ , ⌺ ͒.
2
By now, it has been established that channel ͑1͒ is dominant,
2
˜
exceeding 95% in the entire range of the OClO (A A
2
spheric chemistry arising from reactive ClO (vӷ0) frag-
2
11,14,21
ments have thus been suggested.6
,16,29,30
Evidence for the
�
X B ) absorption band.
Hence, a noticeable influ-
1
ence on the stratospheric ozone layer arising from channel
formation of nitrous oxide (N O) by the reaction of vibra-
2
͑2͒ is ruled out at present.
tionally excited ClO generated in the 308 nm photolysis of
OClO with ambient nitrogen has been reported.3 These find-
ings raise the question of the extent of ClO (vӷ0) radicals
generated from channel ͑1͒. Moreover, these vibrationally
hot radicals should exhibit a significant red shift in
1
In contrast to the gas-phase studies, considerably higher
matrix host dependent quantum yields of chlorine atom for-
mation have been measured following photoexcitation of
OClO in various solutions2
2,23
and low temperature
2
4,25
matrices,
where even a Cl yield of unity has been
32
absorptivity towards the near-UV region and this might
2
6
reported. These results suggest a heterogeneous photo-
chemistry of OClO on atmospheric aerosols which is likely
have additional implications on the atmosphere. An exhaus-
tive assessment of the respective significance of these pro-
cesses, however, requires extensive atmospheric modeling
taking into account also the fact that a further decrease of
a͒Electronic mail: parker@sci.kun.nl
0021-9606/2001/114(19)/8339/8/$18.00
8339
© 2001 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:
31.193.242.21 On: Thu, 27 Nov 2014 05:17:33
1

8340
J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Delmdahl, Parker, and Eppink
total stratospheric ozone will lead to a significant enhance-
ment of the relative solar flux, particularly in the relevant
ultraviolet region.3 From a more fundamental point of view,
examining this rather complex example of triatomic photof-
ragmentation dynamical behavior is particularly challenging,
since multiple excited electronic states are potentially in-
volved during the decay of OClO which is predissociative in
overlap of the two foci, thus, unambiguously proving the
3
absence of any alternative sources of O ( P ) fragment for-
J
3
3
mation. The different spin–orbit states of the O ( P ) pho-
J
toproducts generated from OClO photolysis were probed by
3
(2ϩ1)-REMPI using the well known (3p P
J
3
� 2p P , ⌬Jϭ0) transitions. Since the laser bandwidth
J
3
was always below the Doppler width of the ejected O ( P )
J
3
4–36
nature.
products, scanning the detection laser wavelength repeti-
tively across the entire Doppler profile of the respective O
(3PJ) transitions provided an equal sensitivity for each of
the fragment velocities. Ions were accelerated by an array of
three electrodes, acting as an electrostatic immersion lens, to
the same point of a position sensitive detector consisting of a
dual microchannel plate ͑MCP͒ phosphor screen ͑P20͒ detec-
tor assembly with a charged coupled device ͑CCD͒ camera
͑LaVision, 1024ϫ1280 pixels͒. In order to provide mass se-
lectivity, the front MCP was gated. Typically, 5 000–10 000
laser shots were averaged to collect the raw 2D ͑two-
dimensional͒ ion data. Since the 2D projection was perpen-
dicular to the electric dipole vectors of the pump and the
probe pulse, the full 3D fragment distribution inherent in the
raw 2D ion images could be reconstructed utilizing a Hankel
In the present work, velocity map imaging is applied to
study the energy partitioning and anisotropy for the ClO
2
3
(
(
X ⌸)ϩO ( P ) product channel. OClO is excited to
2
2
A ␯ Ͼ20) at photodissociation wavelengths between
2
1
2
73.5 and 293 nm covering ␯ ϭ21–25 symmetric stretch
1
levels in a cold molecular beam. Correspondingly, O
3
(
PJϭ2,1,0) photofragments resulting from OClO photolysis
at 280 nm are spin–orbit state selectively detected in order to
monitor its correlation with energy partitioning and recoil
anisotropy.
II. EXPERIMENT
OClO is synthesized following the method described by
Derby and Hutchinson.37 In brief, a mixture of neat Cl di-
2
41
inversion algorithm. The 3D image then represents a slice
luted in helium at 750 Torr is flowed through a glass column
through the middle of the reconstructed 3D distribution.
packed with sodium chlorite (NaClO ) and glass beads at
2
room temperature. The greenish reaction product is led via
teflon tubing directly to the molecular valve. Since the re-
III. RESULTS
maining Cl will not have any influence on the measure-
ments, no efforts of further purification of the reaction prod-
uct were necessary.
At the photolysis energies used in this work, the photof-
ragmentation of OClO from its first absorption band gener-
ates both the ClO and the O fragment in their electronic
ground states, X ⌸ and P, respectively. The experiments
described here focus on energy partitioning and recoil anisot-
ropy resulting from the short wavelength photolysis of jet-
2
2
3
The Nijmegen velocity map imaging apparatus as well
as the technique of velocity map imaging a recently devel-
oped high resolution variant of the ion imaging method,38
has been described in detail in an earlier work by Eppink and
cooled OClO. Wavelengths below ␭ϭ300 nm, covering
3
9,40
2
Parker.
Briefly, a pulsed supersonic molecular beam of
OClO ( A ) symmetric stretch vibronic levels from ͑21,0,0͒
2
3
ϳ5% OClO seeded in helium at atmospheric pressure was
directed along the axis of a time of flight spectrometer and
skimmed to 1 mm diam before encountering two counter
propagating pulsed dye laser beams at right angles. An oil
diffusion pump for evacuating the main chamber and further
evacuation of the time-of-flight region by a turbomolecular
pump provided a background pressure as low as 2
to ͑25,0,0͒ were used, and the O ( P ) dissociation products
J
were state-selectively monitored via (2ϩ1) REMPI detec-
tion.
3
A. O „ P… images: General features
3
Figure 1 depicts O ( P ) fragment 2D images together
2
with the corresponding slices through the reconstructed 3D
distributions as obtained from OClO photodissociation at se-
lected wavelengths between 273.5 and 293 nm. Both the
dissociation and the probe laser were vertically polarized
with their electric vectors parallel to the image plane, which
is necessary for reconstructing the 3D velocity distribution
Ϫ7
ϫ10 Torr during the measurement. A Nd:YAG-pumped
dye laser system ͑Spectra-Physics DCR-2A; PDL-2͒ oper-
ated with either Sulforhodamine B or Rhodamin 6G dye fol-
lowed by frequency doubling delivered photolysis wave-
lengths between 273 and 293 nm. A second Nd:YAG-
pumped dye laser system ͑Spectra-Physics GCR-11; PDL-2͒
operated with Coumarin 47 dye and frequency doubling de-
livered the detection laser light at wavelengths around 226
from the original 2D ion data. Ion intensities in these repre-
3
sentations increase from black to white. The O ( P ) ion
2
images shown in Fig. 1 were recorded under experimental
conditions such that even the highest energetically possible
O fragment velocities were still detectable on the active area
of the position sensitive detector. This ensures that the total
product quantum state distribution formed is recorded. In
each of the images, which contain the full fragment angular
recoil and kinetic energy information, several well-resolved
sharp rings peaking at small pixel radii close to the center of
the detector are discernible. The high ion intensities observed
3
nm for detection of O ( P2,1,0) fragments. Both laser systems
were operated at 10 Hz and quartz lenses with fϭ200 mm
and fϭ500 mm were used to focus the detection and the
photolysis laser, respectively. The time delay between the
detection laser ͑pulse duration ϳ5 ns, output energy after
doubling ϳ0.3 mJ͒ and the photolysis laser ͑pulse duration
ϳ5 ns, output energy after doubling ϳ2 mJ͒ was 20 ns.
Experiments were carried out under clean conditions in the
3
sense that no O ( P ) atomic products were detectable when
at small pixel radii indicate that the major fraction of O
J
3
the photolysis laser beam was blocked or in case of a non-
( P ) atoms is released with very low velocities in the dis-
2
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:
131.193.242.21 On: Thu, 27 Nov 2014 05:17:33

J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Photolysis of jet-cooled OClO
8341
FIG. 1. Raw 2D data and slices through the 3D recoil distributions of O
3
FIG. 2. Raw 2D data ͑upper trace͒ and slices through the 3D recoil distri-
butions ͑lower trace͒ of O ( P) atoms in spin–orbit states Jϭ2, 1, and 0
resulting from the photolysis of jet-cooled OClO at a photolysis wavelength
of 280 nm. The direction of the electric vector of the photolysis and probe
laser is vertical in the images.
(
P ) atoms resulting from the photolysis of jet-cooled OClO at the given
2
3
photolysis wavelengths. The electric vector of the photolysis and probe light
is always in the vertical direction, providing the necessary cylindrical sym-
metry. Darker regions correspond to higher signal levels.
sociation process in the entire investigated spectral region.
Furthermore, as is visible in Fig. 1 at a first glance, the an-
ropy, albeit, the respective features still reflect the expected
3
2
gular distributions of O ( P ) ions reveal a significant an-
ϳcos distribution.
2
3
isotropy of the recoiling ClO and O products with the pref-
erential recoil being along the polarization axis of the laser
While O ( P2,1,0) spin–orbit states considerably affected
the energy partitioning in the fragmentation process, the ap-
2
ϩ
beams. The measured ϳcos -type angular distributions re-
pearance of O ion images were not influenced upon altering
2
2
3
flect the parallel nature of the dipole allowed ( A � B )
the polarization geometry of our laser beams. Hence O ( P)
2
1
initial absorption step of OClO. Anisotropy parameters ␤
alignment, as has been reported recently for the dissociation
4
2
ϭ1.3, 1.4, 1.3, and 1.3, respectively, are determined for pho-
of NO at 212.8 nm, is not detectable in OClO photofrag-
2
3
tofragments paired with O ( P ) generated from OClO dis-
mentation at short wavelengths.
2
sociation at 273.5, 275.5, 277.5, and 279.5 nm, respectively,
by fitting the experimental data using the well-known equa-
B. Bond energy and kinetic energy distributions
tion I(␪)ϭ1ϩ␤P (cos ␪), where P2 is the second order
2
Legendre polynomial.
Due to the conservation of linear momentum and energy
3
A reduction in the recoil anisotropy of the products is
observed in OClO dissociation at longer wavelengths,
namely, 292.5 and 293 nm. The distribution at ␭
ϭ292.5 nm yields an overall value of ␤ϭ0.9. In case of the
in the jet-cooled OClO fragmentation, the measured O ( P )
J
kinetic energy distribution of the photoproducts is directly
correlated to the distribution of internal energy of the ClO
partners, which can be written as E (ClO)ϭE
int
h␯
293 nm photolysis a ␤ parameter of 0.8 for the two slower
ϪD (OClO)ϪE (O)ϪEkin(ClOϩO). Since experiments
0 int
fractions is obtained, whereas the fast fragment recoil is sur-
were carried out in a pulsed supersonic nozzle expansion, the
initial internal energy of the parent molecules prior to disso-
prisingly isotropic, corresponding to a ␤ value of merely 0.4.
3
The raw O ( P) velocity map images measured from the
ciation is negligible. The dissociation energy D of OClO
0
photolysis of OClO at a photolysis wavelength of 280 nm
and detecting O ( P) atoms in their Jϭ2, 1 and 0 spin–orbit
has been rather accurately determined in an independent
3
2
near-threshold OClO( A 0,0,0) dissociation experiment
2
2
states are represented in Fig. 2 together with the correspond-
ing 3D cuts. Most salient, the results given in Fig. 2 reveal a
remarkable change in the product velocity distribution ac-
where vibrationally and rotationally cold ClO(X ⌸3/2) frag-
ments are being produced.29 Consequently, the relation
D (OClO)ϭE ϪE ϪE (O) with E ϵE is appli-
0
h␯
av
int
av
kin,max
3
companied with the formation of spin–orbit excited O ( P)
cable and the bond dissociation energy can be determined
from the speed of the fastest fragments. By selectively im-
fragments in their sublevels Jϭ1 and Jϭ0. Compared to the
3
3
image taken for the O ( P ) fragment, many resolved rings
aging the O ( P) fragments in their spin–orbit sublevels J
2
of strong ion intensity at higher radii are clearly visible when
ϭ2, 1, and 0 a value for D (OClO) of 247.3Ϯ0.5 kJ/mol
0
3
3
O ( P ) and O ( P ) atoms are monitored. Consequently, O
was reliably calculated from the correspondingly fastest O
1
0
3
3
(
P ) fragments which are formed spin–orbit excited are
( P ) fragments observed. The obtained value agrees very
1
J
1
4,16
created with significantly higher average kinetic energies in
well with recently reported results,
however, is deter-
3
the fragmentation process than their O ( P ) ground state
mined with much higher accuracy in our near-threshold ex-
periment. It should be mentioned here that near-threshold
dissociation using velocity map imaging has very recently
been demonstrated to be an appropriate means in obtaining
extremely precise dissociation energies for molecular species
2
companions. Note at this point that the images represented in
Fig. 2 are close-ups for better clarity and are comparable in
size to the images at neighboring photolysis energies shown
3
in Fig. 1. Apparently, spin–orbit excitation of the O ( P)
4
4
products is also associated with a reduction in recoil anisot-
such as O ͑Ref. 43͒ and IBr, respectively.
3
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:
131.193.242.21 On: Thu, 27 Nov 2014 05:17:33

8342
J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Delmdahl, Parker, and Eppink
FIG. 3. Total kinetic energy distributions obtained when probing the O
3
(
P ) photofragments from OClO dissociation at the given photodissocia-
FIG. 4. Multiple Gaussian representation of the trimodal ClO vibrational
2
2
3
tion wavelengths. Vibrational level energies of ClO (X ⌸ ,v, jϭ0) are
distribution determined from the corresponding O ( P ) kinetic energy dis-
3
/2
2
indicated by combs.
tribution measured at a dissociation wavelength of 293 nm shown in Fig. 3.
The vibrational levels of ClO are assigned.
Both from the ͑0,0,0͒ and the ͑24,0,0͒ vibronic band of
3
OClO the branching between the different O ( P ) states
tion for OClO dissociation at a wavelength of 292.5 nm ͑see
J
was found to be very close to a ϳ(2Jϩ1) statistical distri-
Fig. 3͒ is shown together with
a
comb of the
3
2
bution. Similar values for the O ( P ) branching ratios have
ClO(X ⌸ ,v,Jϭ0) vibrational levels. Illustrating the tri-
J
3/2
been determined in a recent (2ϩ1) REMPI study at ␭
ϭ308 nm addressing the fragmentation of OClO ͑18,0,0͒.6
In Fig. 3 the product total kinetic energy distributions for
modal composition of this distribution, its overall appearance
is to a rather good approximation reproducible by simply
fitting a set of three different Gaussians having widths
3
2
the dissociation of OClO→O( P )ϩClO (X ⌸) as calcu-
͑FWHM͒ of 3.5 kJ/mol, 33 kJ/mol, and 56 kJ/mol, respec-
2
2
lated from the reconstructed ion images ͑Fig. 1͒ at the given
photolysis wavelengths are depicted. The comb given above
each distribution indicates the location of the vibrational lev-
tively. This fit implies that the initially ( A 21,0,0)-excited
2
parent molecule decays along three different fragmentation
channels, generating three types of ClO fragments. The
2
els of ClO(X ⌸ ,v,Jϭ0) photoproducts released in con-
slowest fraction is rotationally cold (T ϳ200 K) but ex-
3
/2
rot
3
junction with the O ( P ) partner fragments at the different
tremely hot in terms of vibrational excitation, resulting in the
vibrationally and spin–orbit state resolved structure close to
thermodynamic threshold ͓in this case v(ClO)ϭ13–17͔.
Underlying single rotational levels cannot be unambiguously
resolved, since the corresponding rotational constant of ClO,
2
photolysis energies. From the respective ClO vibrational
combs it is obvious that in the whole wavelength range the
ClO molecules are formed with vibrational excitation. As the
most striking result, in each case the overall population of
the ClO vibrational states shows no maximum before the
respective thermodynamic threshold. The vibrational spacing
observable in the kinetic energy distributions are in excellent
agreement with the known vibrational constants of the
Ϫ1
being on the order of 0.5 cm , is too small. The vibrational
distributions of the other two fractions of photofragments
peak at vϭ5 and vϭ10 cover ClO vibrational levels which
are rather likely broadened due to the formation of much
higher rotational levels. A rotational temperature of ϳ1000
K and a spin–orbit ratio ⌸1/2 :⌸3/2ϭ0.25 is assumed for
these far less resolvable ClO fractions. This reasoning has
strong support from previously reported rotationally resolved
two-photon laser induced fluorescence spectra taken at dis-
2
45
ClO(X ⌸ ,v) radical. As obvious, the relatively high
3
/2
3
9
resolution inherent in velocity map imaging allows many
of the vibrational levels corresponding to the low fragment
kinetic energy region to be resolved. At very low kinetic
energies, where the resolution of our velocity map imaging
detector is highest, part of the ClO fine structure, namely
sociation wavelengths of 351 and 308 nm ͑Ref. 10͒ and also
2
⍀
ϭ1/2 spin–orbit sublevels and rotational structure, be-
from (2ϩ1) REMPI spectra of ClO(X ⌸ v,J) fragments
⍀
comes apparent. Only a minor contribution of spin–orbit ex-
cited ClO(X ⌸1/2) is formed in this region. Furthermore, a
obtained from photolysis of OClO in one- and two-color
pump–probe experiments. These will be the subject of a fol-
lowing study.46 As Fig. 3 shows, at wavelengths below ␭
ϭ280 nm the formation of the latter two fractions of photo-
fragments is strongly supressed yielding dramatically in-
2
decrease in dissociation wavelength, i.e., an increase in pho-
tolysis energy, leads to a substantial shift of the total distri-
bution to higher average vibrational energy. In case of the
OClO ͑25,0,0͒ photolysis at ␭ϭ273.5 nm, essentially only
the three highest energetically possible ClO vibrational lev-
els are noticeably populated, i.e., virtually all of the available
energy is transferred to product vibration.
verted ClO vibrational state distributions formed in conjunc-
3
tion with O ( P ) atomic partners which are essentially
2
monomodal.
Figure 5 depicts the kinetic energy distributions obtained
at a photolysis wavelength of 280 nm when each of the three
In Fig. 4, the ClO internal energy distribution which has
3
3
3
been determined from the O ( P ) kinetic energy distribu-
spin–orbit excited atomic fragments O ( P ), O ( P ), and
2
1 1
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:
131.193.242.21 On: Thu, 27 Nov 2014 05:17:33

J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Photolysis of jet-cooled OClO
8343
3
FIG. 6. Normalized
O
( P ) photofragment yield spectra of the
2
2
2
2
(
A 21,0,0), ( A 12,0,0), and ( A 0,0,0) rovibronic transitions of jet-
2
2
2
cooled OClO. The smaller peak to the red of the ͑12, 0, 0͒ band is due to
3
7
O ClO.
spectrum. Figure 7 shows the experimentally obtained
2
(
A 0,0,0), photofragment yield spectrum ͑upper trace͒ to-
2
2
2
FIG. 5. Total kinetic energy distributions obtained as the result of probing O
gether with ( A 0,0,0)� ( B 0,0,0) rovibronic transitions
2
1
3
(
P) fragments in all three J levels generated from OClO photolysis at 280
35
͑
lower trace͒ simulated for the O ClO isotope and a Boltz-
2
nm. Vibrational level energies of ClO(X ⌸ ,v,Jϭ0) are indicated by
3
/2
mann rotational state distribution according to T ϭ40 K.
combs. Note that the vertical scaling is arbitrary and total J level populations
rot
are statistical ͑i.e., the integrated curves for jϭ2:1:0 are 5:3:1͒.
The good agreement is obvious and indicates the degree of
rotational cooling of OClO in helium upon supersonic nozzle
expansion in our apparatus. OClO molecular constants were
taken from the work of M u¨ ller et al.47
3
O ( P ) were state-selectively probed. The internal energies
0
3
of the different O ( P ) photofragments are ϳ1.9 kJ/mol for
J
3
3
O ( P ) and ϳ2.7 kJ/mol for O ( P ). All three channels
IV. DISCUSSION
1
0
correspond to a generation of internally excited ClO frag-
ments in vibrational states (vϾ9) which are well resolvable,
however, a much larger portion of faster photoproducts is
Based on the extensive information obtained previously
for OClO dissociation at longer wavelengths, the highlights
of which are summarized in this section, one would expect at
shorter dissociation wavelengths higher ClO rotational and
vibrational excitation, and an increase in the product angular
distribution anisotropy towards a maximum value of ␤
3
3
observed for spin–orbit excited O ( P ) and O ( P ) atoms.
1
0
2
3
While in the ClO(X ⌸v)ϩO ( P ) channel ClO vibrational
2
states vϭ18 and 19 are most prominent, the kinetic energy
3
distributions obtained by monitoring the O ( P ) and O
1
3
(
P ) partners are centered around the ClO vϭ14 vibra-
0
tional level. Moreover, the strong mutual resemblance is ob-
vious in the latter two distributions which is prevalent also in
3
terms of the measured recoil anisotropy. While the O ( P )
2
result yields a rather high ␤ value of 1.4, decreased anisot-
ropy parameters ␤ϭ1.0 and ␤ϭ1.1 were measured for O
3
3
(
P ) and O ( P ) fragments, respectively.
0 1
C. Excitation spectra
By recording the integrated O ion signal as a function of
the dissociation wavelength while keeping the detection laser
3
wavelength fixed and selectively probing O ( P ) atoms the
2
2
photofragment yield spectra of the OClO( A 21,0,0),
A 12,0,0), and ( A 0,0,0) vibronic bands depicted in Fig.
were measured. Interestingly, both the O ( P ) yield spec-
2
2
2
2
(
2 2
3
6
trum of the ͑0,0,0͒ band at the red end and also the ͑21,0,0͒
band far to the blue spectral region exhibit distinct rotational
structure. Not surprisingly, rotational contours have been
vanished for the ͑12,0,0͒ band located in the vicinity of the
Franck–Condon maximum of the OClO near-UV absorption
2
FIG. 7. Measured ( A 0,0,0) band ͑above͒ and a simulation of the
2
35
2
2
O ClO ( A 0,0,0� B 0,0,0) rovibronic transitions ͑below͒. The best
2
1
agreement is obtained for Trotϭ40 K. This indicates also the degree of ro-
tational cooling of the parent in the supersonic beam.
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:
31.193.242.21 On: Thu, 27 Nov 2014 05:17:33
1

8344
J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Delmdahl, Parker, and Eppink
2
3
TABLE I. OCIO( A ␯ Ͼ20)→ClOϩO( P ): Energy and angular distri-
2
2
1
2
crease in rotational linewidths for OClO( A ␯ ϭ5–10) vi-
2
1
butions.
bronic bands, i.e., at excitation energies covering the barrier
to decay along the asymmetric stretching coordinate.3
␭dissa
Eh␯b
Eav1b,c
Evib
b
Erot
b
fintd
fkind
͗
͘
͗
͘
␤
5
As calculations by Peterson and Werner show, upon
273.5 437.7
275.5 434.5
277.5 431.4
279.5 428.3
292.5 409.2
293.0 408.6
190.4
187.2
184.1
181.0
161.9
161.3
175
158
150
149
97
2
2
3
3
5
5
93%
7%
1.3
2
2
the optical transition from the B ground state to the A
excited state the equilibrium bond angle of OClO narrows
from ϳ117° to ϳ107°. Likewise, the repulsive B excited
state is very strongly bent, exhibiting an equilibrium angle of
only ϳ90°. The A excited state, on the contrary, which is
nearly degenerate in energy with the initially prepared A2
state exhibits a double minimum structure with respect to the
bending coordinate. Note that only a shallow barrier of about
1
2
85% 15% 1.4
83% 17% 1.3
84% 16% 1.3
63% 37% 0.9 ͑0.9͒
65% 35% 0.4 ͑0.8͒
2
2
100
2
1
2
a
Photolysis wavelength in nm.
Energies are given in kJ mol
D0 ͑O–ClO͒ of 247.3 kJ mol is used ͑see text͒.
Fraction of available energy.
b
c
Ϫ1
.
Ϫ1
d
2
2
0 kJ/mol has been calculated to separate OClO( A ) from a
1
bond angle of ϳ120° en route to an almost linear configura-
tion ͑170°͒.35
ϭ0.94. Furthermore, one would also expect a distribution of
2
2
The transition dipole moment for the ( A � B ) ab-
3
2
1
O ( P ) J states that is statistical and independent of the ClO
J
sorption step lies in the molecular plane along the line con-
necting the two centers of the O atoms. Accordingly, under
the condition of an instantaneous fragmentation of the parent
molecule, and assuming that the O fragment recoils along the
direction of the disrupting O–Cl bond, the maximum value
of the anisotropy parameter ␤max ͑determined by the angle ␹
between transition dipole moment and recoil velocity͒ can be
deduced from the equilibrium geometry of each electronic
state distribution. An overview of the results obtained in the
short wavelength gas-phase investigation of OClO photof-
ragmentation reported in this paper is given in Tables I and
II. Very few of these expectations are met, requiring a quite
different view of the dissociation of OClO at higher photon
energies.
A. Dissociation geometry and dynamics of
photoexcited OClO
state using the relation ␤maxϭ2P (cos ␹). Correspondingly,
2
2
2
the equilibrium decay geometry of 117°( B ), 107°( A ),
1
2
Theoretical calculations addressing the ClOϩO dissocia-
2
3
4,35
and 90°( B2) result in an upper value of ␤max of
tion channel
have shown that at the very high photon
2
2
2
1
.15( B ), 0.94( A ), and 0.5( B ), respectively. As is no-
1 2 2
energies used in this work three excited potential energy sur-
ticed from Tables I and II, for dissociation wavelengths at
␭ϭ280 nm and lower, the measured ␤ parameters either ex-
ceed these limiting values or come very close thus, requiring
near-zero dissociation lifetimes. This inevitably points to-
wards a significant fraction of OClO decaying from the op-
tically prepared symmetric stretch vibronic level of the A2
state via the less bent ͑ϳ120°͒ or quasilinear A1 state which
leads to higher ␤max values of ϳ1.25 and ϳ2.0, respectively.
faces ͑PES͒ are open for a decay of the initially
2
(
A ␯ ,0,0)-excited OClO parent along the asymmetric co-
2 1
ordinate. Following the strong, dipole-allowed parallel tran-
2
2
sition from the B electronic ground state to the initial A
1
2
excited state, the parent molecule can reach the nearly de-
2
2
generate A excited surface and subsequently the lower ly-
1
2
2
2
ing B excited state. The A state is accessible from the
2
1
2
2
initial A state by spin–orbit coupling, while the B state is
2
2
2
Although to date the optically weak perpendicular transition
reached from the A state by asymmetric stretch (␯ ) cou-
1
3
2
to the A state has not been spectroscopically detected, the
1
pling. A barrier of ϳ40 kJ/mol to ClOϩO formation along
first direct experimental evidence for its presence ͑and also
the asymmetric stretch coordinate of OClO has been calcu-
2
2
2
for its significant contribution in depopulating the initially
lated both for the A and the A state, whereas the B
2
1
2
2
48
excited A state͒ has been recently given by Reid et al. in
2
state has been found to be purely repulsive. Accordingly,
OClO when dissociated at photon energies below the abso-
lute barrier height ͑ϳ300 kJ/mol͒ should be restricted to the
above-mentioned nonadiabatic coupling mechanism includ-
a resonance Raman depolarization study.
2
B. Dissociation via the A state channel
2
2
2
1
ing all three excited surfaces, namely A , A , and B .
2
1
1
The emerging picture of the OClO parent molecule is
The barrier has been experimentally characterized in previ-
ous high-resolution absorption measurements of jet-cooled
2
focused on a predominant coupling to the A final state after
1
2
2
2
2
the initial ( A � B ) absorption step and evolution along
OClO( A ␯ ,␯ ,␯ )� ( B 0,0,0) transitions reported by
2
1
2
1
2
3
1
5
1
its bending coordinate towards a near-linear recoil geometry
in the exit channel. This picture, however, is not sufficient to
account for the extreme energy partitioning observed. First
of all, the optical signature of the dissociating state is that of
a near-linear state, i.e., a state yielding ␤ values Ͼ1.2. In
addition, for wavelengths below 280 nm, merely 17% ͑or
less͒ of the available energy is found in the translation of the
fragments ͑see Table I͒.
Richard and Vaida. These authors found a substantial in-
3
TABLE II. O ( P ) state, energy, and angular distributions from OClO at
J
␭
ϭ280 nm.
a
a
J
fJ
Eav1a
͗
Evib
͘
͗
Erot
͘
fint
fkin
␤
0
1
2
12%
28%
60%
177.5
178.3
180.2
127
128
147
3
3
2
73%
73%
83%
27%
27%
17%
1.0
1.1
1.4
These unprecedented channeling of vibrational energy
cannot be inferred in a straightforward manner just from the
increase of Cl–O bond length ͑ϳ0.16 Å͒ ͑Ref. 35͒ arising
aEnergies are given in kJ molϪ1
.
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:
31.193.242.21 On: Thu, 27 Nov 2014 05:17:33
1

J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Photolysis of jet-cooled OClO
8345
2
2
from the initial ( A � B ) electronic excitation. Further-
correspond to a very long dissociation lifetime in the order of
17 ps which is completely unexpected in case of the highly
excited 21,0,0 level. Should OClO decay from a near-linear
geometry, ␤max will be close to 2. Since ␤max is reduced by
a factor of 4 as a result of a long lifetime and a rotational
period of OClO in the order of ␻ϳ1.6 ps,14 the measured ␤
parameter of 0.4 ͑Table I͒ gives further evidence for a near-
2
1
more, the various impulsive models describing the expected
distributions of the surplus energy over the fragment degrees
of freedom in triatomic decay49,50 are far from sufficient to
account for the observed high percentages of vibrational en-
ergy in the ClO photofragments. As a consequence, impul-
sive fragment recoil is not capable of generating the ob-
served kinetic energy distributions and must at least be
accompanied by an additional mechanism to become a
source of the almost entirely inverted ClO vibrational state
populations which are seen at short photolysis wavelengths
below 280 nm.
2
linear decay along the A1 surface.
Both the trimodality of the vibrational distributions ob-
3
served in the O ( P2) kinetic energy curves, and the substan-
tially altered kinetic energy distributions resulting from prob-
3
3
ing the O ( P1) and O ( P0) fragments indicate that the
2
In view of these considerations, very recent ab initio
predissociative decay of OClO( A2 ␯1Ͼ20,0,0) proceeds via
more than one pathway. Similar O ( P) spin–orbit level-
2
3
theoretical calculations addressing the OClO A potential
2
energy surface reported by Xie and Guo36 appear remarkably
helpful with respect to a rational explanation of the observed
energy partitioning. Their results show that significant in-
dependent kinetic energy release has been observed recently
by Gericke and co-workers6 in the OClO decay at ␭
ϭ308 nm and also in the short-wavelength photolysis of O3
investigated by Houston and co-workers.52 No detailed mod-
els are currently available in describing these exit channel
separable amplitudes of antisymmetric stretching motion
prevail due to anharmonicity of the A state potential. Al-
though the OClO parent is in the first step excited to a purely
symmetric stretch vibronic level of the A state and the A
2
2
correlation effects and further theoretical work is demanded.
2
2
2
Given the strong fragment recoil anisotropy, the A surface
1
2
2
is nonetheless likely to act as the main exit channel at the
employed wavelengths thereby imprinting its near-linear
equilibrium angle on the fragmenting triatomic parent.
potential exhibits only a single C2v minimum, the wave
function has significant non-C2v symmetry. This is calcu-
2
lated to be the case for ( A ␯ Ͼ4,0,0) vibronic levels of
2
1
OClO. The coupling between symmetric and antisymmetric
stretching coordinate is already clearly visible as a bifurca-
tion in the wave function calculated for the relatively low
V. CONCLUSION
2
36
With the blue wing OClO dissociation study reported in
this article, virtually the entire spectral region of the
lying A 5,0,0 vibronic level of OClO. As a consequence,
2
2
the initial OClO( A ␯ Ͼ20,0,0) parent geometry is far off
2
1
2
2
UV( A � B ) absorption band of the OClO molecule has
2
1
from C2v with largely unequal Cl–O bond distances. Ac-
cordingly, the transition dipole and recoil vector become
closer to parallel, resulting in a higher beta value. The initial
asymmetric stretching motion of vibronically excited
been covered to date by the various state-to-state investiga-
2
tions of the nascent photoproducts ClO(X ⌸ ,v,J) and O
⍀
3
(
P ). Although available theory ͑see for example Refs. 34,
J
2
35, and 36, and references therein͒ is suitable to explain
much of the observations, there is a clear need for more
refined theory which can account for spin–orbit correlation
effects.
Summarizing, the observed very high anisotropy param-
eters and extreme channeling of available energy show a
fundamental change of the fragmentation dynamics of OClO
OClO( A ␯ Ͼ20,0,0) exceeds the bond dissociation energy
2
1
with the O fragment asymptotically released at the outer
turning point. This rationale agrees with the observed paucity
2
of fragment kinetic and rotational energy. OClO( A ␯
2
1
Ͼ20,0,0) with an expected severely distorted symmetry is
initially prepared, undergoes fast coupling to the near-
2
degenerate A surface, and subsequently decays under a
1
17
prevalent in the hitherto disregarded blue end of the OClO
near-linear geometry.
near UV absorption band. Furthermore, the amount of vibra-
tionally hot ClO radicals generated from OClO at short pho-
tolysis wavelengths should be scrutinized in light of possible
atmospheric mechanisms due to both their increased reactiv-
ity and absorptivity.
As obvious from Figs. 3 and 4, the kinetic energy distri-
butions obtained from OClO photodissociation at somewhat
longer photolysis wavelengths, namely, at 292.5 and 293 nm,
are the sum of three different contributions. As to the slowest
fraction, the decrease in the corresponding ␤ parameters with
increasing dissociation wavelength ͑see Table I͒ is presum-
ably due to a longer dissociation lifetime rather than the
ACKNOWLEDGMENTS
2
2
result of fast dissociation via the more bent A2 and B2
R.F.D. and A.T.J.B.S acknowledge financial support by
the European Union TMR program IMAGINE, Contract No.
ERB 4061 PL 97-0264. Technical assistance by Cor Sikkens
and Chris Timmer is gratefully acknowledged. We thank I.
Szydtowska, Dr. W. L. Meerts, and Dr. K. Remmers for
providing us with their polyatomic spectra fitting program.
states, respectively. This reasoning is supported by the recur-
rence of rotational resolution at high energies visible in the
photofragment yield spectrum of the 21,0,0 vibronic level
͑
Fig. 6͒ which is in the first instance quite puzzling with
5
1
regard to earlier work reported by Richard and Vaida. In
their high-resolution absorption study the rotational state
2
vanished for the OClO( A 8,0,0) band which is in agree-
2
1
A. Wahner, G. S. Tyndall, and A. R. Ravishankara, J. Phys. Chem. 91,
ment with the unresolved 12,0,0 band structure depicted in
2
734 ͑1987͒.
Fig. 6 ͑middle͒. The rotational linewidths of the 0,0,0 and of
2
M. B. McElroy, R. J. Salawitch, S. C. Wofsy, and J. A. Logan, Nature
Ϫ1
the 21,0,0 band shown in Fig. 6 are about 0.3 cm and
͑London͒ 321, 759 ͑1986͒.
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:
31.193.242.21 On: Thu, 27 Nov 2014 05:17:33
1

8346
J. Chem. Phys., Vol. 114, No. 19, 15 May 2001
Delmdahl, Parker, and Eppink
3
27
R. Brandtjen, T. Kl u¨ pfel, D. Perner, and B. M. Knudsen, Geophys. Res.
Lett. 21, 1363 ͑1994͒.
J. Sessler, M. P. Chipperfield, J. A. Pyle, and R. Toumi, Geophys. Res.
Lett. 22, 687 ͑1995͒.
S. Solomon, Nature ͑London͒ 347, 347 ͑1990͒.
M. Roth, C. Maul, and K.-H. Gericke, J. Chem. Phys. 107, 10582 ͑1997͒.
E. Bishenden and D. J. Donaldson, J. Chem. Phys. 99, 3129 ͑1993͒.
R. Flesch, M. C. Sch u¨ rmann, J. Plenge, M. Hunnekuhl, H. Meiss, M.
Bischof, and E. R u¨ hl, Phys. Chem. Chem. Phys. 1, 5423 ͑1999͒.
S. Baumg a¨ rtel and K.-H. Gericke, Chem. Phys. Lett. 227, 461 ͑1994͒.
R. F. Delmdahl, S. Baumg a¨ rtel, and K.-H. Gericke, J. Chem. Phys. 104,
V. Vaida and J. D. Simon, Science 268, 1443 ͑1995͒.
2
2
8
9
A. Wahner and C. Schiller, J. Geophys. Res. 97, 8047 ͑1992͒.
R. F. Delmdahl and K.-H. Gericke, Phys. Chem. Earth C 25, 229 ͑2000͒.
S. Baumgartel, R. F. Delmdahl, K.-H. Gericke, and A. Tribukait, Eur.
Phys. J. D 4, 199 ͑1998͒.
4
30
¨
5
6
7
8
31
R. F. Delmdahl and K.-H. Gericke, Chem. Phys. Lett. 281, 407 ͑1997͒.
M. Mandelmann and R. W. Nicholls, J. Quant. Spectrosc. Radiat. Transf.
3
3
3
2
3
4
17, 483 ͑1977͒.
9
0
V. E. Fioletov, J. B. Kerr, and D. I. Wardle, Geophys. Res. Lett. 24, 2997
͑1997͒.
J. L. Gole, J. Phys. Chem. 84, 1333 ͑1980͒.
1
1
1
1
2
883 ͑1996͒.
1
2
3
35
R. F. Delmdahl, S. Ullrich, and K.-H. Gericke, J. Phys. Chem. A 102,
680 ͑1998͒.
Y. Matsumi, S. M. Shamsuddin, and M. Kawasaki, J. Chem. Phys. 103,
490 ͑1995͒.
C. J. Kreher, R. T. Carter, and J. R. Huber, Chem. Phys. Lett. 286, 389
1998͒.
K. A. Peterson and H.-J. Werner, J. Chem. Phys. 96, 8948 ͑1992͒.
D. Xie and H. Guo, Chem. Phys. Lett. 307, 109 ͑1999͒.
R. I. Derby and W. S. Hutchinson, Inorg. Synth. 4, 152 ͑1953͒.
D. W. Chandler and P. L. Houston, J. Chem. Phys. 87, 1445 ͑1987͒.
A. T. J. B. Eppink and D. H. Parker, Rev. Sci. Instrum. 68, 3477 ͑1997͒.
D. H. Parker and A. T. J. B. Eppink, J. Chem. Phys. 107, 2357 ͑1997͒.
A. J. R. Heck and D. W. Chandler, Annu. Rev. Phys. Chem. 46, 335
͑1995͒.
M. Ahmed, D. S. Peterka, A. S. Bracker, O. S. Vasyutinskii, and A. G.
Suits, J. Chem. Phys. 110, 4115 ͑1999͒.
N. Taniguchi, K. Takahashi, Y. Matsumi, S. M. Dylewski, J. D. Geiser,
and P. L. Houston, J. Chem. Phys. 111, 6350 ͑1999͒.
E. Wrede, S. Laubach, S. Sculenburg, A. J. Orr-Ewing, and M. N. R.
Ashfold, Chem. Phys. Lett. 326, 22 ͑2000͒.
36
7
37
38
4
39
40
41
͑
1
1
1
4
5
6
H. F. Davis and Y. T. Lee, J. Chem. Phys. 105, 8142 ͑1996͒.
H. F. Davis and Y. T. Lee, J. Phys. Chem. 96, 5681 ͑1992͒.
A. Furlan, H. A. Scheld, and J. R. Huber, J. Chem. Phys. 106, 6538
4
4
2
3
͑1997͒.
1
1
7
8
K. Fenner, A. Furlan, and J. R. Huber, J. Phys. Chem. A 101, 5736 ͑1997͒.
P. Ludowise, M. Blackwell, and Y. Chen, Chem. Phys. Lett. 273, 211
͑1997͒.
1
2
9
0
44
45
T. Baumert, J. L. Herek, and A. H. Zewail, J. Chem. Phys. 99, 4430
1993͒.
R. F. Delmdahl, B. L. G. Bakker, and D. H. Parker, J. Chem. Phys. 112,
298 ͑2000͒. ͑The reported wavelength of 281 nm was actually 273.5 nm
͑
J. A. Coxon, Can. J. Phys. 57, 1538 ͑1979͒.
R. F. Delmdahl and D. H. Parker ͑in preparation͒.
46
5
due to an unrecognized dye laser miscalibration.͒
W. G. Lawrence, K. C. Clemitshaw, and V. A. Apkarian, J. Geophys. Res.
47
H. S. P. M u¨ ller, G. O. So”rensen, M. Birk, and R. R. Friedl, J. Mol.
2
2
2
1
2
3
Spectrosc. 186, 177 ͑1997͒.
9
5, 18 591 ͑1990͒.
48
P. J. Reid, A. P. Esposito, C. E. Foster, and R. A. Beckman, J. Chem.
Phys. 107, 8262 ͑1997͒.
G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3626 ͑1972͒.
R. Schinke, Photodissociation Dynamics ͑Cambridge University Press,
Cambridge, 1993͒.
E. C. Richard and V. Vaida, J. Chem. Phys. 94, 163 ͑1991͒.
J. D. Geiser, S. M. Dylewski, J. A. Mueller, R. J. Wilson, R. Toumi, and
P. L. Houston, J. Chem. Phys. 112, 1279 ͑2000͒.
S. C. Hayes, M. P. Philpott, S. G. Mayer, and P. J. Reid, J. Phys. Chem.
A 103, 5534 ͑1999͒.
R. C. Dunn, E. C. Richard, V. Vaida, and J. D. Simon, J. Phys. Chem. 95,
4
5
9
0
6
060 ͑1991͒.
2
2
2
4
5
6
E. J. Lanzendorf and A. C. Kummel, Geophys. Res. Lett. 23, 1251 ͑1996͒.
L.-H. Lai, C.-P. Liu, and Y.-P. Lee, J. Chem. Phys. 109, 988 ͑1998͒.
C. J. Pursell, J. Conyers, P. Alapat, and R. Parveen, J. Phys. Chem. 99,
51
52
1
0433 ͑1995͒.
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:
31.193.242.21 On: Thu, 27 Nov 2014 05:17:33
1