Measurement of the rate coefficient for collisional removal of O 2 (X3∑g-, ν=1) by O( 3P)

Article abstract of DOI:10.1063/1.2110227

We report a laboratory measurement of the rate coefficient for the collisional removal of O2 (X Σg-3, ν=1) by O (P3) atoms. In the experiments, 266-nm laser light photodissociates ozone in a mixture of molecular oxygen and ozone. The photolysis step produces vibrationally excited O2 (a Δg1) that is rapidly converted to O2 (X Σg-3, ν=1-3) in a near-resonant electronic energy-transfer process with ground-state O2. In parallel, a large amount of O (D1) atoms is generated that promptly relaxes to O (P3). Under the conditions of the experiments, only collisions with the photolytically produced O (P3) atoms control the lifetime of O2 (X Σg-3, ν=1), because its removal by molecular oxygen at room temperature is extremely slow. Tunable 193-nm laser light monitors the temporal evolution of the O2 (X Σg-3, ν=1) population by detection of laser-induced fluorescence near 360 nm. The removal rate coefficient for O2 (X Σg-3, ν=1) by O (P3) atoms is (3.2±1.0) × 10-12 cm3 s-1 (2δ) at a temperature of 315±15 K (2δ). This result is essential for the analysis and correct interpretation of the 6.3-μm H2 O (ν2) band emission in the Earth's mesosphere and indicates that the deactivation of O2 (X Σg-3, ν=1) by O (P3) atoms is significantly faster than the nominal values recently used in atmospheric models.

Full text of DOI:10.1063/1.2110227

Measurement of the rate coefficient for collisional removal of O 2 ( X Σ g − 3 , υ = 1 ) by
O ( P 3 )
Konstantinos S. Kalogerakis, Richard A. Copeland, and Tom G. Slanger
Citation: The Journal of Chemical Physics 123, 194303 (2005); doi: 10.1063/1.2110227
View online: http://dx.doi.org/10.1063/1.2110227
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/123/19?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Kinetic and dynamic studies of the Cl(2 P u) + H2O( X ̃ 1 A 1) → HCl( X ̃ 1Σ+) + OH( X ̃ 2Π) reaction on an ab
initio based full-dimensional global potential energy surface of the ground electronic state of ClH2O
J. Chem. Phys. 139, 074302 (2013); 10.1063/1.4817967
Nonadiabatic dynamics of O(1 D) + N2 ( X Σ g + 1 ) → O(3 P) + N2 ( X Σ g + 1 ) on three coupled potential
surfaces: Symmetry, Coriolis, spin-orbit, and Renner-Teller effects
J. Chem. Phys. 136, 054308 (2012); 10.1063/1.3682467
Theoretical determination of rate constants for vibrational relaxation and reaction of O H ( X Π 2 , v = 1 ) with O (
P 3 ) atoms
J. Chem. Phys. 129, 064306 (2008); 10.1063/1.2957901
Vibrational energy transfer in O 2 ( X Σ g − 3 , υ = 2 , 3 ) + O 2 collisions at 330 K
J. Chem. Phys. 123, 044309 (2005); 10.1063/1.1982788
Quantum reactive scattering calculations of cross sections and rate constants for the N ( 2 D)+ O 2 (X 3 Σ g − )→
O ( 3 P)+ NO (X 2 Π) reaction
J. Chem. Phys. 118, 3111 (2003); 10.1063/1.1530575
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:
69.166.47.134 On: Tue, 16 Dec 2014 09:19:12

THE JOURNAL OF CHEMICAL PHYSICS 123, 194303 ͑2005͒
Measurement of the rate coefficient for collisional removal
3
−
3
of O „X ⌺ ,␷=1… by O„ P…
2
g
a͒
Konstantinos S. Kalogerakis, Richard A. Copeland, and Tom G. Slanger
Aeronomy Group, Molecular Physics Laboratory, Stanford Research Institute SRI International,
Menlo Park, California 94025-3493
͑
Received 23 August 2005; accepted 14 September 2005; published online 11 November 2005͒
We report a laboratory measurement of the rate coefficient for the collisional removal of
3
−
3
O ͑X ⌺ ,␷=1͒ by O͑ P͒ atoms. In the experiments, 266-nm laser light photodissociates ozone in a
2
g
mixture of molecular oxygen and ozone. The photolysis step produces vibrationally excited
1
3
−
g
O ͑a ⌬ ͒ that is rapidly converted to O ͑X ⌺ ,␷=1–3͒ in a near-resonant electronic energy-transfer
2
g
2
1
process with ground-state O . In parallel, a large amount of O͑ D͒ atoms is generated that promptly
2
3
relaxes to O͑ P͒. Under the conditions of the experiments, only collisions with the photolytically
3
3
−
g
produced O͑ P͒ atoms control the lifetime of O ͑X ⌺ ,␷=1͒, because its removal by molecular
2
oxygen at room temperature is extremely slow. Tunable 193-nm laser light monitors the temporal
3
−
g
evolution of the O ͑X ⌺ ,␷=1͒ population by detection of laser-induced fluorescence near 360 nm.
2
3
−
g
3
The removal rate coefficient for O ͑X ⌺ ,␷=1͒ by O͑ P͒ atoms is ͑3.2±1.0͒
2
−
12
3 −1
ϫ10 cm s ͑2␴͒ at a temperature of 315±15 K ͑2␴͒. This result is essential for the analysis and
correct interpretation of the 6.3-␮m H O͑␯ ͒ band emission in the Earth’s mesosphere and indicates
2
2
3
−
g
3
that the deactivation of O ͑X ⌺ ,␷=1͒ by O͑ P͒ atoms is significantly faster than the nominal
2
values recently used in atmospheric models. © 2005 American Institute of Physics.
DOI: 10.1063/1.2110227͔
͓
I. INTRODUCTION
hand, there has been considerable effort on the part of theo-
reticians and also experimentalists to relate the processes of
vibrational relaxation, isotopic exchange, and atomic
Planetary upper atmospheres are active environments,
containing molecular species excited in different ways.
These species are generated both from the action of sunlight,
causing photodissociation, and from recombination pro-
cesses involving atoms and molecules. In the terrestrial at-
mosphere there is constant interplay between O and O spe-
cies as energy is exchanged, and these processes also involve
minor species, for example, CO , NO, and H O.
5
–7
recombination.
In the present study, we are interested in investigating
3
−
g
3
the V-T process, where O ͑X ⌺ ,␷=1͒ is relaxed by O͑ P͒.
2
This is a fundamental interaction in reaction dynamics, yet
there are no low-temperature measurements where this ques-
tion has been directly addressed. The reason that this process
has potential atmospheric importance is related to the fast
2
2
2
3
−
After formation, the electronically and vibrationally ex-
interaction between O ͑X ⌺ ,␷=1͒ and water. The H O͑␯ ͒
2 2 2
g
cited O molecules are collisionally relaxed, and the energy
mode can be excited in this manner, and the intensity of the
2
repositories are the lowest vibrational levels of the ground
resulting 6.3-␮m radiation can be interpreted as a measure of
state for O molecules in a triplet configuration and the
the local H O density. In order to determine how much
2
2
1
3
−
␷
=0 level of the a ⌬ state for singlet molecules. The latter
O ͑X ⌺ ,␷=1͒ is available for energy transfer to H O, we
g
2
2
g
loses all its energy by radiation above 90 km because elec-
tronic quenching is quite ineffective, but at lower altitudes it
is also converted to the ground state. There is thus a consid-
erable flux of molecules passing through the ground-state
need to know the relative relaxation rates for the competing
3
colliders O͑ P͒ and H O.
2
One of the tasks of the Sounding of the Atmosphere
using Broadband Emission Radiometry ͑SABER͒ instrument
in NASA’s Thermosphere, Ionosphere, Mesosphere, Energet-
ics and Dynamics ͑TIMED͒ mission is to measure atmo-
spheric water vapor by observing the H O͑␯ ͒ emission.
␷
=1 level, and the loss of this population by vibrational-to-
translational ͑V-T͒ energy transfer in a pure O /N atmo-
2
2
1
,2
sphere is very slow, which is why it is important to quan-
2
2
3
3,4
tify the losses by O͑ P͒ and water.
Knowledge of the water concentration profile is crucial to an
understanding of the chemistry and energy flow in the terres-
trial stratosphere and mesosphere. In turn, to relate the
SABER observations to actual water profiles, it is necessary
to understand the various non-local-thermodynamic-
equilibrium ͑non-LTE͒ processes that govern the production
3
−
g
Studies on O ͑X ⌺ ,␷=1͒ have historically been carried
2
out in shock tubes and thus at very high temperatures, but
atmospheric applications typically require kinetic informa-
tion at low temperatures, below 300 K. There have therefore
been few attempts to make measurements under conditions
directly applicable to atmospheric conditions. On the other
8
and loss of H O͑␯ ͒. Zhou et al. have demonstrated that a
2
2
significant fraction of the emission in the 6.3-␮m region ap-
pears to reflect lack of LTE. Making the assumption that the
a͒Electronic mail: ksk@sri.com
0
021-9606/2005/123͑19͒/194303/7/$22.50
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
123, 194303-1
© 2005 American Institute of Physics
6

194303-2
Kalogerakis, Copeland, and Slanger
J. Chem. Phys. 123, 194303 ͑2005͒
system is in LTE could lead to significant errors in deducing
water concentration profiles. Mlynczak et al. have high-
cused, 30-mJ/pulse, 266-nm laser beam, so that knowledge
of the initial ozone density provides the requisite O-atom
density. Reducing the photolysis laser beam energy by 20%
8
lighted the need for an improved laboratory measurement of
3
−
3
−
the collisional removal rate coefficient of O ͑X ⌺ ,␷=1͒ by
has no effect in the intensity of the O ͑X ⌺ ,␷=1͒ fluores-
2
g
2
g
oxygen atoms as essential to a reliable interpretation of the
SABER data. They determined that knowledge of this rate
cence signal, whereas the reduction of the pulse energy by
30% results in a reduction of fluorescence by just 3%. Vary-
ing the concentration of ozone allows for control of the
O-atom concentration.
3
coefficient to accuracy better than ±50% is needed for a re-
3
liable analysis of data from SABER.
For the vibrational relaxation ͑V-T͒ rates, there is essen-
tially no data at low temperature. Breen et al. obtained a
In the photolysis step, the fourth harmonic of a Nd:YAG
͑yttrium aluminum garnet͒ laser ͑Quanta Ray DCR-3͒ with
energy of approximately 30 mJ/pulse is focused into the cell
with a lens of 40 cm in focal length. The photodissociation of
−
12
3 −1
value of 1.4ϫ10 cm s in extrapolating to 300 K from
9
the 1000–3000 K data, while the calculated 300 K value is
6
1
four to five times larger.
ozone at 266 nm leads to the production of O in the a ⌬
2
g
This paper reports the first laboratory measurements of
state with a quantum yield of approximately 90% in levels
␷=0–3, with a minor channel involving the production of O2
in highly vibrationally excited levels of the ground state
3
−
the rate coefficient for the removal of O ͑X ⌺ ,␷=1͒ by O
2
g
atoms near room temperature. In the experiments, laser ra-
diation at 266 nm photodissociates ozone in a mixture of
3
− 10
X ⌺ . The corresponding atomic product partners for these
g
3
1
3
oxygen and ozone, so that only collisions with O͑ P͒ atoms
two channels are O͑ D͒ and O͑ P͒, respectively. Molecular
oxygen in the a ⌬ state is produced in ␷=0 with a fraction
3
−
1
control the lifetime of O ͑X ⌺ ,␷=1͒. A tunable 193-nm la-
2
g
g
3
−
ser monitors the temporal evolution of the O ͑X ⌺ ,␷=1͒
of approximately 55% and 25%, 13%, and 12% in ␷=1, 2,
2
g
11–13
population by detection of laser-induced fluorescence near
60 nm. The results obtained for the rate coefficient indicate
and 3, respectively.
Vibrationally excited O ͑a ⌬ ͒ rapidly relaxes to the
1
3
2
g
3
−
g
that the deactivation of O ͑X ⌺ ,␷=1͒ by O atoms is signifi-
corresponding vibrational states of the ground state in a near-
2
1
4
cantly faster than the nominal values used until recently in
atmospheric models and bear out the hypothesis that for the
resonant electronic energy-transfer process,
1
3
−
g
O ͑a ⌬ ,␷ = 1,2, or 3͒ + OЈ͑X ⌺ ,␷ = 0͒
2
g
2
O+O system, V-T vibrational energy relaxation, limiting
2
3
−
1
high-pressure recombination, and isotopic exchange exhibit
very similar rate coefficients.
→
3
O ͑X ⌺ ,␷ = 1,2, or 3͒ + OЈ͑a ⌬ ,␷ = 0͒. ͑1͒
2
g
2
g
−
g
Thus, O ͑X ⌺ ,␷=1–3͒ is produced promptly, permitting ki-
2
netic studies of these levels at ordinary temperatures, in con-
trast to the very high temperatures encountered in shock tube
experiments. We have recently reported measurements of the
II. EXPERIMENT
The experiments are performed in a stainless-steel flow
cell evacuated by a two-stage mechanical pump. Ozone is
produced by flowing high-purity oxygen through an ozonator
3
−
collisional removal of O ͑X ⌺ ,␷=2,3͒ by O near room
2
g
2
+
−
0.6
0.4
temperature. The determined rate coefficients were 2.0
ϫ10− cm s and ͑2.6±0.4͒ϫ10 cm s for ␷=2 and
13
3
−1
−13
3 −1
͑Welsbach model T-408͒ and is stored in a glass trap with
1
5
3
−
g
3
, respectively. In contrast, the removal of O ͑X ⌺ ,␷=1͒
2
silica gel, kept at low temperature by a dry ice/isopropyl
alcohol bath. Ozone is introduced into the cell by flowing a
by O is extremely slow, with a rate coefficient of
2
͑
2.7±0.4͒ϫ10− cm s , because the only available re-
18
3
−1 16
carrier gas, typically He or O , at slightly above atmospheric
2
laxation path is V-T energy transfer. Therefore, following the
photodissociation of ozone at 266 nm in mixtures containing
significant amount of molecular oxygen, all the nascent vi-
pressure through the trap. To minimize ozone decomposition
on surfaces, 1/4-inch-diam Teflon tubing is used and the
number of metal joints is kept to a minimum. A separate line
is used to introduce another gas into the cell, oxygen in most
cases. The flow of the high-purity gases employed in the
experiments is regulated and monitored by calibrated mass
flow meters ͑Tylan FM-380͒; typical flow rates are approxi-
mately 0.5 standard liters per minute. The gas lines merge
before the flow cell and gases are well mixed as they flow
through Teflon tubing of approximately 1 m in length.
The concentration of ozone in the cell is determined by
the relative absorption measurement of 253.7-nm light gen-
erated by a mercury lamp and is used to infer the O-atom
concentration. For this measurement, the absorption path be-
tween the mercury lamp and a filtered photomultiplier tube
goes through the center of the cell and intersects the photoly-
sis laser beam at an angle of 45°. In the experiments, no
difference is observed in the measured concentration of
ozone in the cell whether the photolysis laser is turned off or
is running continuously. Under the experimental conditions,
ozone in the cell is completely dissociated within the fo-
1
3
−
brationally excited O ͑a ⌬ ͒ will yield O ͑X ⌺ ,␷=1͒.
2
g
2
g
1
O ͑a ⌬ ,␷=0͒ is removed very slowly by O with a rate
2
g
2
−
18
3
−1 10
coefficient of ͑1.7±0.3͒ϫ10 cm s .
1
1
Fast energy transfer from O͑ D͒ to O relaxes O͑ D͒ to
2
3
1
+
g
O͑ P͒ and generates the ␷=0 and 1 levels of the O ͑b ⌺ ͒
2
state. This process has a rate coefficient of ͑4.0+0.8͒
ϫ10− cm s and has been reported to produce the ␷=0
11
3 −1
1
0,17
and 1 levels in approximately equal yields,
1
3
1
+
g
O͑ D͒ + O → O͑ P͒ + O ͑b ⌺ ,␷ = 0,1͒.
͑2͒
2
2
Recent unpublished investigations in our group have indi-
cated that the branching ratio of this process favors ␷=1. The
1
+
O ͑b ⌺ ,␷=0͒ level is removed extremely slowly by O
2
g
2
−
17
3
−1 10
with a rate coefficient of ͑3.9±2.0͒ϫ10 cm s . In con-
trast, experiments at Stanford Research Institute SRI Interna-
1
+
g
tional have shown that O ͑b ⌺ ,␷=1͒ is removed very effi-
2
3
−
ciently by O , and converted to O ͑X ⌺ ,␷=1͒ with a rate
2
2
g
−
11
3
−1 18
coefficient of ͑1.57±0.06͒ϫ10 cm s . This dramatic
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6

194303-3
Collisional removal of O by O
J. Chem. Phys. 123, 194303 ͑2005͒
2
difference of more than five orders in magnitude was attrib-
uted to a near-resonant electronic energy exchange relaxation
1
4
mechanism, with no change in vibrational quanta, i.e.,
1
+
g
3
−
O ͑b ⌺ ,␷ = 1͒ + OЈ͑X ⌺ ,␷ = 0͒
2
2
3
g
−
1
+
g
→
O ͑X ⌺ ,␷ = 1͒ + OЈ͑b ⌺ ,␷ = 0͒.
͑3͒
2
g
2
From the kinetics of the processes described in the pre-
vious paragraphs, it is evident that photolysis with 266-nm
radiation of a mixture containing molecular oxygen and
ozone ͑e.g., 500:1 ratio of partial pressures and a total pres-
sure of 100 Torr͒ will evolve after a few microseconds into a
3
mixture that for each ozone molecule contains 1 O͑ P͒ atom,
3
−
1
+
1
0
.4 O ͑X ⌺ ,␷=1͒ molecules, 0.9 O ͑b ⌺ ,␷=0͒ molecules,
g g
2 2
1
.9 O ͑a ⌬ ,␷=0͒ molecules, and a small fraction of 0.1 O
2
2
g
molecules in highly vibrationally excited states. Such a mix-
ture is appropriate for kinetic studies of the energy transfer
3
−
g
3
between O ͑X ⌺ ,␷=1͒ and O͑ P͒ atoms. The presence of
2
other excited species is not expected to interfere with the
measurements, as will be detailed in Sec. III.
3
−
g
For the monitoring of O ͑X ⌺ ,␷=1͒, the output radia-
2
tion of a tunable ArF excimer laser ͑Lambda Physik EMG
50 TMSC͒ is used. This laser has been modified in-house to
allow for triple-pass amplification of the tunable output of
1
FIG. 1. Excitation spectra of O /O mixture ͑60 Torr O /200 mTorr O ͒
1
9,20
3
2
2
3
the oscillator ͑192.5–194.2 nm͒ in the amplifier cavity.
near 193 nm, following O photodissociation at 266 nm. The features asso-
3
When optimized, it can produce radiation with a maximum
energy of approximately 100 mJ/pulse and a linewidth of
ciated with O ground-state levels ␷=0, 1, and 2 are marked. Data for
2
3
−
O ͑X ⌺ ,␷=1͒ removal by
O
atoms were taken for excitation of
2
g
3
−
3
−
3
−
−
1
O2͑X ⌺ ,␷=1͒ via the P͑19͒ and R͑21͒ triplets of the ͑7, 1͒ B ⌺ –X ⌺
g u
3 − 3 −
g
1
cm . Under the conditions of the experiments, the laser
was operated with a typical maximum energy output of
0 mJ/pulse. Given the absorption cross section of molecu-
band at 193.05 nm. Fluorescence from the ͑7, 19͒ B ⌺ –X ⌺ band was
u
g
detected by a filtered photomultiplier tube ͑PMT͒ at 360±10 nm.
3
lar oxygen at 193 nm, it is necessary to propagate the laser in
a beam path isolated and continuously purged with dry nitro-
gen gas. The radiation output of the tunable ArF laser excites
time of the fluorescence signal, which is contained within the
approximately 20-ns duration of each laser pulse. A delay
generator ͑SRS DG-535͒ controls the time delay between the
excitation and ionization laser pulses. The boxcar averager
and the delay generator are both interfaced to a personal
computer. For the investigations of the collisional removal
rates, the wavelengths of the two lasers are kept fixed, and
the time delay between the two lasers is varied while the
fluorescence is recorded. For each time point of such a delay
scan, the computer records the average fluorescence signal
resulting from 7–13 pairs of shots from the two lasers and
can sum consecutive delay scans. For the present experi-
ments, a set of 10–20 individual scans are summed to gen-
erate a data set ready for analysis. Delay scans are recorded
with the tunable ArF laser tuned on and off resonance and, as
detailed in the following section, a subtraction is performed
to remove the unwanted background fluorescence signal.
3
−
O ͑X ⌺ ,␷=1͒ via the P͑19͒ and R͑21͒ triplets of the ͑7, 1͒
2
g
3
−
3
−
B ⌺ –X ⌺ band, and the fluorescence from the ͑7, 19͒
u
g
3
−
u
3
−
g
B ⌺ –X ⌺ band is detected through a 360-nm interference
filter by a photomultiplier tube. The choice of the
7, 19͒ band ensures the most favorable combination of
͑
2
1
Franck-Condon factors and emission wavelength signifi-
cantly different from the 266-nm photodissociation laser.
Following propagation through the purged path using
193-nm mirrors, the beam is restricted to an approximate
diameter of 1–2 mm using apertures. The two counterpropa-
gating laser beams overlap in the flow cell and the diameter
of the probe beam is two to three times smaller than that of
the photolysis laser. The probe beam alignment is adjusted so
that it is centered with respect to the photolysis beam. Mea-
surements using different relative laser beam sizes and focus-
ing show no indication of diffusion effects. Typical values of
the energy of the probe beam just before it enters the cell are
in the range of 1–2 mJ/pulse. The time evolution of the
III. RESULTS AND DISCUSSION
3
−
g
population of O ͑X ⌺ ,␷=1͒ is monitored by varying the
2
delay between the photodissociation and the probe-laser
pulses.
The laser-induced fluorescence is collected with high-
quality quartz optics and detected by a photomultiplier tube
Figure 1 presents excitation spectra at different pump-
probe delay times with broadband fluorescence detection us-
ing a 360-nm interference filter. The total pressure is 60 Torr
in a mixture of O with 200 mTorr O . The features associ-
2
3
͑
RCA 1P28B͒. The signal is preamplified ͑SRS SR445͒ and
ated with the ␷=0, 1, and 2 levels of ground-state molecular
oxygen are indicated in the figure. As the delay time between
the pump and probe lasers is increased, the ␷=2 features
processed by a gated integrator/boxcar averager ͑SRS
SR250͒. The gate width is significantly larger than the life-
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6

194303-4
Kalogerakis, Copeland, and Slanger
J. Chem. Phys. 123, 194303 ͑2005͒
3
−
FIG. 3. O ͑X ⌺ ,␷=1͒ number density as a function of delay time for two
2
g
FIG. 2. LIF signal as a function of the delay time in a mixture with 94 Torr
O2 and 150 mTorr O3 for the probe-laser on resonance with the
different
O /O
mixtures
͑89 Torr O /115 mTorr O
and
3
2
2
3
94 Torr O /150 mTorr O ͒. The data for the latter mixture is identical to
that shown in Fig. 2. Each point is an average of 56 laser shots. The number
density scale is calibrated using the signal from thermal O ͑X ⌺ ,␷=1͒
2
3
3
−
O ͑X ⌺ ,␷=1͒ feature at 193.05 nm ͑open circles͒ and off resonance at
2
g
1
93.00 nm ͑open triangles͒. The bottom curve shows the difference between
the on- and off-resonance scans ͑closed circles͒. Each point is an average of
6 laser shots.
3
−
g
2
before firing the pump laser. Open arrows indicate the predicted signal in-
tensity after ozone photodissociation. Closed arrows indicate the range of
5
3
−
g
estimated number density for thermal O ͑X ⌺ ,␷=1͒ after relaxation. The
2
magnitude of the estimated temperature jump is presented in Table I. The
solid lines correspond to single-exponential fits that yield the decay
constants.
decrease rapidly in intensity because of collisions with O2,
leaving population in ␷=1 that is slowly removed by the O
atoms present in the mixture.
Figure 2 presents the recorded laser-induced fluores-
cence ͑LIF͒ signal as a function of the delay time in a mix-
ture with 94 Torr O and 150 mTorr O for ͑a͒ the probe laser
of energy available for rapid heating is approximately
25–30% of the total available energy of the photodissociating
photon, or approximately 1 eV for each event; the rest of the
energy is converted into electronic energy that is not relaxed
in the time frame of the measurement. Energy release caus-
ing a 10% temperature increase above room temperature
would result in a twofold increase in the fraction of thermal
2
3
3
−
on resonance with the O ͑X ⌺ ,␷=1͒ feature at
2
g
193.05 nm and ͑b͒ off resonance at 193.00 nm. The bottom
panel shows the difference between the on- and off-
resonance scans and represents the temporal evolution of the
3
−
g
3
−
O ͑X ⌺ ,␷=1͒ population. Because the duration of the scan
O ͑X ⌺ ,␷=1͒. Under the range of experimental conditions
2
2
g
is long ͑ϳ250 ␮s͒, it is difficult to obtain a straight baseline
in individual scans despite careful optimization of the elec-
trical connections to eliminate ground loops and electrical
pickup. The same behavior persists when the probe laser is
detuned off resonance or even when the laser beams are
blocked. We attribute the observed baseline oscillation to
electrical pickup related to the firing of the Nd:YAG laser
used for the generation of the 266-nm pulse. Nevertheless,
the baseline variation is reproducible between the on- and
off-resonance scans and allows the extraction of the fluores-
employed, we estimate the local temperature to be
315±15 K͑2␴͒. The open arrows of Fig. 3 indicate the ex-
pected initial signal intensity given the ozone concentration
used. The solid arrows show the expected signal intensity for
3
−
the final thermal O ͑X ⌺ ,␷=1͒ population following ozone
2
g
photodissociation, and correspond to the estimated range of
final temperature for the gas mixture used.
The solid line curves of Fig. 3 correspond to single-
exponential fits that yield the decay constants ͑i.e., the in-
verse lifetime of the exponential decay͒. Detailed informa-
tion on the measurements performed is listed in Table I. In
the data analysis, the O-atom concentration was adjusted to
account for the resultant rapid density change following
ozone photodissociation. For the calculation of the average
value of the rate coefficient, each measurement listed in
Table I was weighted proportionally to the inverse of the
square of the standard deviation resulting from the data fit.
The confidence limits were estimated by combining the error
limits from the exponential fits to the data with the estimated
uncertainties for the calculated temperature jump and the
measurement of the ozone concentration ͑2␴ uncertainty
within 5% and 6%, respectively, for each experiment͒. The
3
−
cence signal from O ͑X ⌺ ,␷=1͒.
2
g
3
−
g
Figure 3 presents the O ͑X ⌺ ,␷=1͒ number density as
a function of the delay time for two different ozone concen-
trations in the mixture. The signal from thermal O ͑X ⌺ ,
=1͒ present before firing the photodissociation laser is used
2
3
−
g
2
␷
to calibrate the number density scale. From the figure, we
can estimate that our detection technique has a sensitivity of
3–4 mTorr.
The offset in the baseline of Fig. 3 between signal inten-
sity before ozone photodissociation and signal intensity at
long-time delays is attributed to the temperature jump occur-
ring in the gas mixture following ozone photodissociation.
Given the energetics of the processes involved, the amount
3
−
g
removal rate coefficient for O ͑X ⌺ ,␷=1͒ by relaxation O
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6

194303-5
Collisional removal of O by O
J. Chem. Phys. 123, 194303 ͑2005͒
2
3
−
g
TABLE I. Experimental results for O ͑X ⌺ ,␷=1͒+O removal rate coefficient.
2
O pressure ͑mTorr͒ O pressure ͑Torr͒ Temperature estimate ͑K͒
Decay constant
Rate coefficient
͑10 cm s ͒
3
2
−
1
−12
3 −1
±6%͑2␴͒
±1 Torr
±5%͑2␴͒
͑␮s
͒
a
a
1
10
90
89
99
104
90
94
89
94
67
66
94
98
94
309
309
309
308
311
310
312
312
319
319
313
312
313
0.0115±0.003͑2␴͒ 3.34±0.87͑2␴͒
115
0.0138±0.0028
0.0110±0.0042
0.0107±0.0038
0.0131±0.0032
0.0143±0.0038
0.0130±0.0018
0.0124±0.0030
0.0102±0.0028
0.0124±0.005
0.0172±0.0024
0.0144±0.0036
0.0167±0.0032
3.84±0.78
2.86±1.09
2.71±0.96
3.32±0.81
3.56±0.94
3.00±0.41
2.84±0.68
2.37±0.65
2.86±1.16
3.71±0.52
3.06±0.76
3.54±0.68
1
1
1
1
1
1
1
1
1
1
1
23
26
27
29
40
41
42
43
50
52
53
a
Error estimates derived only from uncertainty in exponential fit.
3
−
atoms at 315±15 K͑2␴͒ is found to have a value of
and allows us to anticipate an increase of the O ͑X ⌺ ,
2
g
−
12
3 −1
͑
3.2±1.0͒ϫ10 cm s at the 95% confidence level ͑2␴͒.
␷=1͒+O rate coefficient of approximately 30% as the tem-
perature is decreased from room temperature to 230 K.
The nominal value used in atmospheric modeling of
TIMED/SABER results for the removal rate coefficient of
Experiments using total O pressure as low as 20 Torr
2
produced similar results within uncertainty. These experi-
ments involve temperatures approximately 10% higher than
the range presented here. Given that no significant change in
the measured rate coefficient is expected for such a tempera-
ture change, we interpret these results as supporting the as-
3
−
g
−12
3
−1 4
O2͑X ⌺ ,␷=1͒ by O atoms at 230 K is 1ϫ10 cm s .
This value is slightly smaller than that obtained from experi-
ments in high-temperature shock experiments, where 1000–
3
−
g
9
sumption that the measured lifetime of O ͑X ⌺ ,␷=1͒ is not
3000 K data were extrapolated to 300 K. The rate coeffi-
cient determined here is three times larger than the nominal
value used in the recent models. At issue is the amount of
O2͑X ⌺g ,␷=1͒ that is available for energy transfer to water
and subsequent IR radiation, and the increased loss rate for
O2͑X ⌺g ,␷=1͒ with O atoms decreases this amount by a
significant factor. In more recent modeling calculations using
the new rate coefficient presented here, the SABER team
2
influenced by diffusion out of the probe region in the flow
cell. In addition, the rate coefficient measured in a 1:3 mix-
ture of O and N at a total pressure of 48 Torr was not
3
−
2
2
affected by the presence of N , indicating that the relaxation
2
3
−
g
3
−
of O ͑X ⌺ ,␷=1͒ in collisions with N is at least as ineffi-
2
2
cient as in the case of O . This observation is in agreement
2
with high-temperature experiments available in the
2
2
found that ͑1͒ the H O͑010͒ vibrational temperature in the
literature.
Anderson et al. have pointed out that the isotopic ex-
2
5
mesosphere is closer to LTE, i.e., generally lower, ͑2͒ the
effect is large in daytime ͑when atomic oxygen is more abun-
dant͒, and ͑3͒ the effects are larger for a colder mesopause
and, hence, larger in the daytime polar summer, where the
*
*
change rate coefficient for the O +O →O+O O process
2
should be equal to that of the limiting high-pressure rate
coefficient for O+O recombination. The latter should also
2
change in vibrational temperature for H O͑010͒ can be as
much as 10 K near 80 km. Such a difference produces a
significant change in the retrieved H O from SABER, up to
be equivalent to the rate coefficient for vibrational relaxation
2
2
6
3
−
g
3
of O ͑X ⌺ ,␷=1͒ by O͑ P͒, assuming that all of these pro-
2
cesses have the same O transition state. From their study, the
2
2
2
6
5
0% for optically thin conditions. The overall conclusion is
isotopic exchange rate at 300
ϫ10 cm s , in remarkably good agreement with the de-
K
is ͑2.9±0.5͒
−
12
3 −1
3
−
termined limiting high-pressure recombination rate coeffi-
TABLE II. Experimental results for O2͑X ⌺g ,␷=1͒ relaxation, isotopic ex-
change, and the recombination high-pressure limit ͑300 K͒.
−
12
3
−1 7,23
cient value of ͑2.8±1.0͒ϫ10 cm s . Furthermore, the
results of calculations of isotope exchange rates and high-
pressure limit recombination rates are also in general agree-
Rate coefficient
−12
3 −1
System
͑10 cm s
͒
Reference
5
ment with the experimental values. Table II lists experimen-
High-pressure limit
1.7
Hippler and Troe ͑1971͒
Cobos and Troe ͑1985͒
tal values determined for the three processes, and there
is general agreement that in each case a number near
2
.8±1.0͑1␴͒
−
12
3 −1
3
ϫ10 cm s is a good representation of the 300 K rate
coefficient.
The agreement indicates that formation of the collision
Isotope exchange
2.9±0.5͑1␴͒
Anderson et al. ͑1985͒
2.66±0.37͑2␴͒ Wiegell et al. ͑1997͒
complex is the rate-determining step for the vibrational re-
laxation process. This conclusion is also consistent with the
available experiments and theoretical calculations of the tem-
3.4±0.6͑2␴͒
Fleurat-Lessard et al. ͑2003͒
Vibrational relaxation
3.2±1.0͑2␴͒
This work
2
4,25
perature dependence of the isotope exchange reaction
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6

194303-6
Kalogerakis, Copeland, and Slanger
J. Chem. Phys. 123, 194303 ͑2005͒
TABLE III. Comparison of ␷=1 relaxation by
O
atoms for
IV. CONCLUSIONS
The removal rate coefficient of O ͑X ⌺ ,␷=1͒ by O
3
−
1
1
+
O ͑X ⌺ ,a ⌬ ,b ⌺ ͒ at 300 K.
2
g
g
g
3
−
g
2
Rate coefficient
atoms has been measured for the first time near room tem-
perature. It has approximately the same value as the rate
coefficients for O+O2 isotopic exchange and the high-
pressure limit for three-body O+O2 recombination. The
value is larger by a factor of three than the nominal values
−
12
3 −1
State
͑10 cm s
͒
Reference
3
−
X ⌺
3.2±1.0͑2␴͒
ഛ1.5
This work
g
1
a
a ⌬
g
1
+
b
b ⌺
4.5͑+0.8/−1.3͒͑2␴͒
g
3
used in past atmospheric modeling studies, where relative
a
Reference 37.
Reference 28.
3
−
b
rate coefficients for the removal of O ͑X ⌺ ,␷=1͒ by O at-
2
g
oms and water are important in assessing H O density re-
2
trieval from IR observations of H O emission. Thus, revision
2
that new faster rate coefficient brings the vibrational tem-
peratures of both O and H O closer to LTE.
of the input to the model is necessary. In addition, because a
collision complex is most likely to be involved in the energy-
transfer process, the rate coefficient at the appropriate strato-
spheric temperature ͑ϳ230 K͒ may diverge even further
from the room-temperature value. This expectation is sup-
ported by all available information on the temperature de-
2
2
Because we have previously determined values for
3
1
1
+
O͑ P͒ relaxation of the ␷=1 levels of the O a ⌬ and b ⌺
2
g
g
2
7,28
states,
we can now compare the rate coefficient for the
three lowest electronic states of O states. In each case, we
2
5
,24,25,36
pendence of the isotopic exchange.
Experiments to
assume the dominant process is one of V-T transfer; even for
1
+
determine the temperature dependence of the rate coefficient
are currently underway, in which a new experimental ap-
proach is used that exploits the equilibrium between the
the b ⌺ state, there is not enough energy available for ex-
g
3
1
citation of the O͑ P͒ atom to O͑ D͒. Table III shows the
results of the determinations, and we see that the deactiva-
3
−
1
27
1
+
O ͑X ⌺ ,␷=1͒ and O ͑a ⌬ ,␷=1͒ levels.
2 2
g
g
tion of O ͑b ⌺ ,␷=1͒ is slightly faster than that of
2
g
3
−
1
O ͑X ⌺ ,␷=1͒, with O ͑a ⌬ ,␷=1͒ deactivation being
2
g
2
g
slower. For the ground state the presumption is that the pro-
ACKNOWLEDGMENTS
1
+
cess is V-T. For O ͑b ⌺ ,␷=1͒ there are obviously more
2
g
The NASA Ionospheric, Thermospheric and Mesos-
pheric Physics, Supporting Research and Technology Pro-
gram supported this work under Grant Nos. NAG5-9262,
and NAG5-13002. One of the authors ͑K.S.K.͒ acknowl-
edges partial support for his time from NSF Grant No. ATM-
options available to the system, yet the rate coefficient agrees
3
−
with that of O ͑X ⌺ ,␷=1͒ within the two standard devia-
2
g
tion uncertainty level. Of particular interest is the fact that
1
+
−1
3
−
O ͑b ⌺ ,␷=1͒ is resonant within 4 cm , with O ͑X ⌺ ,
␷
2
g
2
g
2
9
ϭ10͒, yet the rate coefficient does not reflect the view that
0
209229. We thank Philip C. Cosby for help with spectral
this adiabatic channel is important. The same is true when
assignments and David L. Huestis, Dušan A. Pejaković, Mar-
tin M. Mlynczak, Maya García-Comas, and Manuel Lopez-
Puertas for helpful comments and discussions.
1
+
the O ͑b ⌺ ,␷=1͒ level is removed by O , which is a very
2
g
2
3
0
3
−
fast process, yet apparently does not involve O ͑X ⌺ ,
2
g
␷
=10͒ excitation, because the deactivation of O in that
2
31
1
ground-state level is not anomalously fast, as would be the
case via microscopic reversibility if the O ͑b ⌺ ,
J. A. Kunc, J. Phys. B 24, 3741 ͑1991͒.
A. B. Callear, in Photochemistry and Reaction Kinetics, edited by P. G.
2
1
+
g
2
Ashmore, F. S. Dainton, and T. M. Sugden ͑Cambridge University Press,
Cambridge, 1967͒.
M. G. Mlynczak, D. K. Zhou, M. Lopez-Puertas, G. Zaragoza, and I. J.
3
−
g
␷
=1͒↔O ͑X ⌺ ,␷=10͒ transfer were significant. Because
2
1
3
3
the O ͑a ⌬ ,␷=1͒+O͑ P͒ rate coefficient is an upper limit,
2
g
M. Russell, Geophys. Res. Lett. 26, 63 ͑1999͒.
M. Lopez-Puertas, G. Zaragoza, B. J. Kerridge, and F. W. Taylor,
we do not conjecture on the mechanism, but because the O3
4
transition states for the three O electronic states are presum-
2
J. Geophys. Res. 100, 9131 ͑1995͒.
S. M. Anderson, F. S. Klein, and F. Kaufman, J. Chem. Phys. 83, 1648
5
ably different, differences in the loss rate coefficients are
hardly surprising.
͑
1985͒.
6
7
8
M. Quack and J. Troe, Ber. Bunsenges. Phys. Chem. 81, 329 ͑1977͒.
C. J. Cobos and J. Troe, J. Chem. Phys. 83, 1010 ͑1985͒.
D. K. Zhou, M. G. Mlynczak, M. Lopez-Puertas, and G. Zaragoza, Geo-
phys. Res. Lett. 26, 67 ͑1999͒.
As already mentioned, the much larger rate coefficient
3
−
g
for the relaxation of O ͑X ⌺ ,␷=1͒ by O atoms compared to
2
V-T relaxation by O is attributed to the formation of a tran-
2
9
J. E. Breen, R. N. Quy, and G. P. Glass, J. Chem. Phys. 59, 556 ͑1973͒.
S. P. Sander, R. R. Friedl, D. M. Golden et al., JPL Publication Report
No. 02-25, 2003; ͑http://jpldataeval.jpl.nasa.gov/index.html͒.
J. J. Valentini, D. P. Gerrity, D. L. Phillips, J.-C. Nieh, and K. D. Tabor,
J. Chem. Phys. 86, 6745 ͑1987͒.
R. K. Sparks, L. R. Carlson, K. Shobatake, M. L. Kowalczyk, and Y. T.
Lee, J. Chem. Phys. 72, 1401 ͑1980͒.
sient bound collision complex involving low-lying excited
states of ozone. We can make interesting comparisons for the
interaction of other simple diatomic molecules with atomic
oxygen. For example, the value of the rate coefficient for the
10
11
12
13
14
15
1
+
vibrational relaxation of N ͑X ⌺ ,␷=1͒ by O atoms at room
2
g
−
15
3
−1 32
temperature is ͑3.2±1.3͒ϫ10 cm s , whereas that for
S. M. Dylewski, J. D. Geiser, and P. L. Houston, J. Chem. Phys. 115,
7460 ͑2001͒.
1
+
−14
3
−1 33
CO͑X ⌺ ,␷=1͒+O is ͑2.8±0.9͒ϫ10 cm s . In con-
T. G. Slanger and R. A. Copeland, Chem. Rev. ͑Washington, D.C.͒ 103,
trast, extremely fast relaxation is observed in the case of the
4
731 ͑2003͒.
2
“
molecular free radical” NO͑X ⌸,␷=1͒, which is vibra-
K. S. Kalogerakis, T. G. Slanger, and R. A. Copeland, J. Chem. Phys.
123, 044309 ͑2005͒.
tionally relaxed in collisions with O atoms more than three
1
6
D. L. Huestis ͑unpublished͒.
1
+
orders of magnitude faster than CO͑X ⌺ ,␷=1͒ and one or-
17
18
L. C. Lee and T. G. Slanger, J. Chem. Phys. 69, 4053 ͑1978͒.
D. A. Pejaković, E. R. Wouters, K. E. Phillips, T. G. Slanger, R. A.
3
4,35
der of magnitude faster than O2.
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6

1
94303-7
Collisional removal of O by O
J. Chem. Phys. 123, 194303 ͑2005͒
Chem. Phys. Lett. 405, 372 ͑2005͒.
T. G. Slanger and P. C. Cosby, J. Phys. Chem. 92, 267 ͑1988͒.
H. I. Bloemink, R. A. Copeland, and T. G. Slanger, J. Chem. Phys. 109,
4237 ͑1998͒.
M. Klatt, I. W. M. Smith, A. C. Symonds, R. P. Tuckett, and G. N. Ward,
J. Chem. Soc., Faraday Trans. 92, 193 ͑1996͒.
R. J. McNeal, M. E. Whitson, Jr., and G. R. Cook, J. Geophys. Res. 79,
1527 ͑1974͒.
M. E. Lewittes, C. C. Davis, and R. A. McFarlane, J. Chem. Phys. 69,
1952 ͑1978͒.
R. P. Fernando and I. W. M. Smith, Chem. Phys. Lett. 66, 218 ͑1979͒.
E. S. Hwang, K. J. Castle, and J. A. Dodd, J. Geophys. Res. 108, 1109
͑2003͒.
P. Fleurat-Lessard, S. Y. Grebenshchikov, R. Siebert, R. Schinke, and
N. Halberstadt, J. Chem. Phys. 118, 610 ͑2003͒.
D. A. Pejaković, Z. Campbell, K. S. Kalogerakis, R. A. Copeland, and T.
G. Slanger ͑unpublished͒.
2
Copeland, and K. S. Kalogerakis, J. Geophys. Res. 110, A03308 ͑2005͒.
G. W. Faris and M. J. Dyer, J. Opt. Soc. Am. B 10, 2273 ͑1993͒.
G. W. Faris and M. J. Dyer, Opt. Lett. 18, 382 ͑1993͒.
1
9
0
1
2
3
4
29
30
2
2
2
2
2
P. H. Krupenie, J. Phys. Chem. Ref. Data 1, 423 ͑1972͒.
R. C. Millikan and D. R. White, J. Chem. Phys. 39, 3209 ͑1963͒.
H. v. Hippler and J. Troe, Ber. Bunsenges. Phys. Chem. 75, 27 ͑1971͒.
M. R. Wiegell, N. Wessel Larsen, T. Pedersen, and H. Esgaard, Int. J.
Chem. Kinet. 29, 745 ͑1997͒.
31
32
33
34
2
2
5
6
P. Fleurat-Lessard, S. Y. Grebenschikov, and R. Schinke, J. Chem. Phys.
119, 4700 ͑2003͒.
M. García-Comas, M. Lopez-Puertas, C. J. Mertens, M. G. Mlynczak, J.
M. Russell, III, E. E. Remsberg, L. L. Gordley, K. S. Kalogerakis, R. A.
Copeland, and T. G. Slanger, EOS Trans. Am. Geophys. Union 83,
SA72A-0512 ͑2002͒.
3
3
3
5
6
7
2
2
7
8
D. A. Pejaković, Z. Campbell, K. S. Kalogerakis, R. A. Copeland, and T.
G. Slanger, EOS Trans. Am. Geophys. Union 85, SA41A-1032 ͑2004͒.
D. A. Pejaković, R. A. Copeland, T. G. Slanger, and K. S. Kalogerakis,
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:
9.166.47.134 On: Tue, 16 Dec 2014 09:19:12
6