The photolysis of NO2 at 193 nm

Article abstract of DOI:10.1016/S0009-2614(01)00149-X

Infrared kinetic spectroscopy has been used to investigate the photo-dissociation of NO₂ at 193 nm. The O(¹D)/(O(¹D)+O(³P)) branching ratio was measured as 0.55±0.03 by adding a large excess of H₂ to the system, and comparing the amount of NO₂ removed during the photolysis to the amount of OH formed by the subsequent fast reaction between O(¹D) and H₂. The rate constant for the reaction between O(¹D) and NO₂ was measured as (1.5±0.3)×10^-10cm^-3s^-1, and the 193 nm absorption cross-section for NO₂ was estimated as (2.9±1.2)×10^-19cm².

Full text of DOI:10.1016/S0009-2614(01)00149-X

30 March 2001
Chemical Physics Letters 337 +2001) 72±78
www.elsevier.nl/locate/cplett
The photolysis of NO2 at 193 nm
F. Sun, G.P. Glass, R.F. Curl *
Chemistry Department and Rice Quantum Institute, Rice University, Houston, TX 77005, USA
Received 14 December 2000
Abstract
Infrared kinetic spectroscopy has been used to investigate the photo-dissociation of NO2 at 193 nm. The O+1D)/
+O+1D)+O+3P)) branching ratio was measured as 0:55 Æ 0:03 by adding a large excess of H2 to the system, and com-
paring the amount of NO2 removed during the photolysis to the amount of OH formed by the subsequent fast reaction
between O+1D) and H2. The rate constant for the reaction between O+1D) and NO2 was measured as ꢀ1:5 Æ 0:3† Â
10À10 cmÀ3
s
À1, and the 193 nm absorption cross-section for NO2 was estimated as ꢀ2:9 Æ 1:2†  10À19 cm2. Ó 2001
Elsevier Science B.V. All rights reserved.
1. Introduction
yield for O+3P) production has been determined to
be almost unity [6]. At 244 nm, a second photo-
1
Chemical reactions of free radicals with NO2,
many of which are of importance in atmospheric
and combustion chemistry, are commonly studied
by ArF laser photolysis of mixtures containing
NO2 and a suitable radical precursor [1±5]. In or-
der to model such systems, it is necessary to
properly account for the reactions of the photo-
lysis products of NO2. Unfortunately, it is not
possible to do this because the O+1D) to O+3P)
branching ratio for NO2 photolysis at 193 nm is
unknown.
dissociation channel yielding NOꢀX2P† ‡ Oꢀ D†
opens. At wavelengths between 252 and 214 nm,
the quantum yield for production of O+1D) has
been investigated by Uselman and Lee [7]. Their
results suggest that the O+1D) to O+3P) branching
ratio increases from zero at 252 nm to approxi-
mately 0.5 at 243 nm. The branching ratio then
decreases somewhat before rising again to reach a
value of 0:5 Æ 0:1 at 214 nm. Over this entire
wavelength range, photo-dissociation appears to
2
~
occur from the B B2 state of the NO2 molecule [8].
At signi®cantly higher energies, photo-dissoci-
ation has been studied by Taherian et al. [9], who
used an F2 laser to excite NO2 at 157 nm. This
wavelength lies in the middle of a progression of
absorption bands, extending from 165 to 135 nm,
Photo-dissociation of NO2 has been thoroughly
investigated at a number of wavelengths other
than 193 nm. At wavelengths between 398 and 244
nm the only energetically allowed photo-dissocia-
tion products are NOꢀX2P† ‡ Oꢀ P†, and over the
3
greater part of this wavelength range, the quantum
that has been assigned to a 3pru Rydberg state
2
~
‡
+E Ru ) of NO2 [10]. The principal photo-dissoci-
ation products identi®ed by Taherian et al. were
ground state NO and O [9]. The quantum yield of
O+1S), which is accessible at wavelengths shorter
* Corresponding author. Fax: +1-713-348-5155.
E-mail address: rfcurl@rice.edu +R.F. Curl).
0009-2614/01/$ - see front matter Ó 2001 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 9 - 2 6 1 4 + 0 1 ) 0 0 1 4 9 - X

F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
73
than 170 nm, was measured as less than 1%, and
the quantum yield of O+1D) was determined to be
no greater than 10%.
In the present work, we report a measurement
of the O+1D) to O+3P) branching ratio at 193 nm, a
determination of the room temperature rate con-
stant for the subsequent reaction of O+1D) with
NO2, and an estimate of the 193 nm absorption
cross-section for NO2.
the appropriate Doppler width in order to estimate
the degree of line broadening caused by either
pressure broadening or unresolved hyper®ne
structure.
Reaction mixtures typically contained 30 mTorr
of NO2, 14±800 mTorr of H2, and 9 Torr of
He. Approximately 0.6% of the NO2 in the
excimer beam was dissociated by each laser pulse.
All ¯ow rates were measured using calibrated
MKS ¯ow meters. Partial pressures for individual
components were calculated from the relative
¯ow rate and the measured total pressure. The
equilibrium between N2O4 and NO2 was ac-
counted for by using the data of Verhoek and
Daniels [13].
2. Experimental
In these experiments, mixtures containing NO2,
H2 and excess He bu€er gas were photolysed at
193 nm, by using an ArF excimer laser. Laser-
based infrared absorption spectroscopy was used
to follow the time evolution of various reaction
products. The overall apparatus has been de-
scribed previously [11]. For these studies, a newly
developed infrared laser based on di€erence fre-
quency generation +DFG) in periodically poled
LiNbO3 was used [5]. Using this laser source, NO2
and OH were observed at 2915.854 cmÀ1 +the
m1 ‡ m3164;13 154;12F1 transition) and 3407.989
cmÀ1 +the P+4,5)1+transition), respectively [12].
The peak infrared absorption cross-sections for
Doppler broadened NO2 and OH at these
wavelengths have been measured [12] as 6:13 Â
10À19 cm2 and 4:37 Â 10À18 cm2, respectively. The
sensitivity of the apparatus to OH corresponds
to a S=N ˆ 1 for an OH concentration of
3. Methodology
The O+1D) to O+3P) branching ratio was deter-
mined by adding excess H2 to the photolysis cell
and comparing the amount of NO2 removed dur-
ing photolysis to the amount of OH formed by the
subsequent fast reaction between O+1D) and H2. In
the presence of H2, the following reactions occur.
First the photolysis
3
1
NO2 ‡ hmꢀ193 nm† ! NO ‡ aOꢀ P† ‡ bOꢀ D†
ꢀ1†
which is followed by the destruction of O+3P) by
NO2
3
1 Â 1011 cmÀ3
.
Oꢀ P† ‡ NO2 ! NO ‡ O2
ꢀ2†
and the destruction of O+1D) by NO2 through re-
action and quenching
Two kinds of data were obtained in this study:
time scans in which the probe frequency was tuned
to a ®xed frequency at or near the absorption
maximum, and frequency scans in which the probe
frequency scanned over the line and data was ac-
quired at several time intervals after the excimer
¯ash. The time scans provided a better S=N ratio
for a given collection period, and allowed data to
be extrapolated back to time t ˆ 0. The frequency
scans were least-squares ®tted to a Gaussian line-
shape, which removed any time variation in the
baseline and any uncertainty about whether the
probe was tuned to the peak of the absorption line.
The frequency scans also provided a measurement
of the linewidth, which could be compared with
1
Oꢀ D† ‡ NO2 ! NO ‡ O2
ꢀ3a†
ꢀ3b†
1
3
Oꢀ D† ‡ NO2 ! Oꢀ P† ‡ NO2
and the reaction of O+1D) with H2
1
Oꢀ D† ‡ H2 ! OH ‡ H
and the reaction of the H produced in +4) with
ꢀ4†
NO2
H ‡ NO2 ! NO ‡ OH
ꢀ5†
In the above mechanism, a ‡ b ˆ 1, and any O+3P)
produced in +3b) is removed by +2).

74
F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
When a very large excess of H2 is added to the
system, reaction +3a) and +3b) becomes negligible
as compared to reaction +4), and we have by
combining +1), +2), +4), and +5) remembering that
a ‡ b ˆ 1
2NO2 ‡ hm ‡ bH2 ! 2NO ‡ aO2 ‡ 2bOH
ꢀ6†
Thus the ratio of the OH formed to the NO2
consumed equals b, and the branching ratio for
production of O+1D) can be determined directly by
measuring the relative amounts of OH formed and
NO2 removed.
For smaller relative concentrations of H2, re-
actions +3a) and +3b) compete with reaction +4) for
O+1D). Then, keeping the concentration of NO2
and the laser ¯uence constant, the overall mecha-
nism can be analyzed to give the following Stern±
Volmer equation:
ÀD‰NO2Š=D‰OHŠ ˆ bÀ1ꢀ1 ‡ k3‰NO2Š=k4‰H2Š† ꢀ7†
from which k3=k4 can be estimated, if [OH] is de-
termined as a function of [H2]. The easiest way to
derive Eq. +7) is to recognize that for each photon
absorbed, two molecules of NO2 are always de-
stroyed and the number of molecules of OH pro-
duced will be equal to
Fig. 1. The time behavior of the OH and NO2 infrared signals
observed when using a mixture containing 31 mTorr of NO2,
548 mTorr of H2 and 9 Torr of He.
2k4‰H2Š
b
k4‰H2Š ‡ k3‰NO2Š
containing 31 mTorr of NO2, 548 mTorr of H2 and
9 Torr of He. As can be seen, the NO2 signal de-
creases rapidly to a relatively constant value. The
OH signal increases somewhat more slowly, before
reaching a maximum value from which it very
slowly decays. The slower increase in the OH ab-
sorbance undoubtedly arises because a signi®cant
fraction of the OH produced by reaction +4) is
vibrationally excited [14], and collisional relax-
ation of these excited molecules, predominantly by
NO2 [15], takes 70±80 ls. The very slow decay,
which exhibits an exponential decay constant of
approximately 2 ms, is consistent with OH loss
occurring primarily by the third order reaction [16]
In the experiments reported here, the infrared
absorbance of OH was measured rather than its
concentrations. Therefore, it proved to be conve-
nient to convert the OH and the NO2 concentra-
tion in Eq. +7) to an absorbance by using the
equations, AOH ˆ rOH‰OHŠL and ÀDANO ˆ rNO
2
2
ꢀÀD‰NO2Š†L. Then, after replacing ‰NO2Š=‰H2Š by
PNO =PH , Eq. +8) is obtained:
2
2
À1
A
ˆ fÀDANO ꢀrOH=rNO †gꢀ1 ‡ k3PNO =k4PH †
OH
2
2
2
2
ꢀ8†
In this equation, A represents the base e absor-
bance, r the peak infrared absorption cross-sec-
tion, and Px the partial pressure of component x.
OH ‡ NO2 ‡ He ! HNO3 ‡ He
ꢀ9†
4. Results
In order to validate the reaction mechanism
outlined in +1)±+5), the following two experiments
were performed. First, NO2 was monitored in both
the presence and the absence of H2. According to
Fig. 1 shows the time behavior of OH and NO2
following the 193 nm ¯ash photolysis of a mixture

F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
75
the proposed mechanism, D+NO2) should not be
a€ected by the addition of H2, since the amount of
NO2 removed in reaction +3a) and +3b) is the same
as that removed by the sequence +4) + +5). Experi-
mentally, the amount of NO2 removed in the ab-
sence of H2 was found to be 0:96 Æ 0:05 of that
removed in the presence of H2. A series of exper-
iments was also performed in which H2 was re-
placed by CH4. In these experiments, the following
reaction [17,18] sequence replaces +4) + +5).
1
Oꢀ D† ‡ CH4 ! OH ‡ CH3
ꢀ10†
ꢀ11†
CH3 ‡ NO2 ! CH3O ‡ NO
CH3O is fairly unreactive +reacting only with NO2
by a slow third order process [19]). The amount of
OH produced in the presence of CH4 is predicted
to be only one half of that produced in the pres-
ence of H2. In practice, the ratio was measured as
0:51 Æ 0:03.
Fig. 2. A Stern±Volmer plot of the OH absorbance versus the
inverse of the H2 partial pressure.
Having successfully validated the mechanism,
k3=k4 was then measured by photolyzing a number
of mixtures that contained 31 mTorr of NO2, 9
Torr of He, and various partial pressures of H2,
ranging from 14 to 548 mTorr. Table 1 lists the
absorbances of OH that were determined from
these experiments. For each experiment, the ab-
sorbance was estimated by extrapolating absor-
bance-time scans of the type shown in Fig. 1 back
to time t ˆ 0. This procedure typically increased
the absorbance by about 5% from that recorded at
70±80 ls.
À1
A
ˆ ꢀ62:7 Æ 0:5† ‡ ꢀ2:73 Æ 0:07†PHÀ1
ꢀ12†
OH
2
represents a weighted least squares ®t to the data.
Since the form of Eq. +12) is identical to that of
Eq. +8), it is clear that k3PNO =k4 can be estimated
by dividing the slope of the line in Fig. 2 by its
2
intercept. Therefore, k3PNO =k4 ˆ ꢀ2:73 Æ 0:07†=
2
ꢀ62:7 Æ 0:5† ˆ 0:044 Æ 0:002 Torr. When PNO is
2
set equal to its value of 0.031 Torr, a value of
1:4 Æ 0:1 is obtained for k3=k4.
The O+1D)/+O+1D) +O+3P)) branching ratio was
measured by comparing the peak OH absorption
obtained from a frequency scan recorded 105 ls
after the photolysis pulse to the decrease in NO2
absorption measured from an absorption-time
scan of the type shown in Fig. 1. Frequency scans
rather than time scans were used to ensure that the
true maximum peak absorption of the transient
OH species was actually sampled, since an analysis
of several OH time traces indicated that it was not
always possible to pre-set the probe laser to the
exact center of the OH absorption line. This di-
culty was not experienced when monitoring stable
species such as NO2, because it was possible to set
the laser to the frequency of maximum absorption
of such species before beginning the time scan.
The data shown in Table 1 are displayed in
Fig. 2 as a Stern±Volmer plot of AÀOH1 vs PHÀ1. In this
®gure, the straight line, which can be represented
by the equation
2
Table 1
The absorbance of OH as a function of the partial pressure of
H2
PH
+mTorr)
1=PH
1000 Â A
1/A
2
2
+TorrÀ1
)
14.1
25.9
45.9
73.1
143
71.0
38.6
21.8
13.7
7.0
4.31
6.05
7.48
10.0
12.2
13.8
14.9
232
165
134
100
82.0
282
548
3.6
1.8
72.7
67.0

76
F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
A number of measurements of the type de-
by NO2 in reaction mixtures containing contained
31 mTorr of NO2 and 548 mTorr of H2. This
percentage was determined by using the value of
k3=k4 determined above. After both corrections
were made, the branching ratio for production of
O+1D) was estimated as 0:55 Æ 0:03.
scribed above, and shown in Fig. 3, were made
using a reaction mixture that contained 31 mTorr
of NO2, 548 mTorr of H2 and 9 Torr of He, and
from these, the average relative absorbance,
A
OH=DANO , AOH=DANO , was determined to be
2
2
3:54 Æ 0:18. From this value, the ratio of D[OH]
to D[NO2] was estimated as 0:50 Æ 0:03 by using
Eq. +13), and substituting known values [12] for
In order to measure the 193 nm absorption
cross-section for NO2, the amount of OH pro-
duced by photolysis of a mixture containing
31 mTorr of NO2, and 548 mTorr of H2 was com-
pared to that produced by photolysis of a mixture
containing 13.5 mTorr of NH3, and 31 mTorr of
NO2. According to the mechanism outlined in re-
actions +1)±+5), in the ®rst mixture, 2b molecules of
OH are formed for each photon absorbed by NO2.
rNO and rOH
.
2
D‰OHŠ=D‰NO2Š ˆ ꢀAOH=DANO † Á ꢀrNO =rOH
†
2
2
ꢀ13†
In order to determine b, the branching ratio for
production of O+1D), two small corrections had to
be made to value of D[OH]/D[NO2] estimated
above. The ®rst took account of the small increase
in [OH] when the OH absorption signal was ex-
trapolated from time t ˆ 105 ls to t ˆ 0. The
second correction was made in order to account
for the small percentage of O+1D) that is removed
Therefore, [OH]1 is equal to 2brNO [NO2]. In the
second mixture, essentially all of the OH is pro-
2
duced by photolysis of NH3, because rNH is very
3
much greater than rNO [5]. In this mixture, the
following reactions occur.
2
NH3 ‡ hm ! NH2 ‡ H
H ‡ NO2 ! OH ‡ NO
ꢀ14†
ꢀ5†
and, one molecule of OH is formed for each
photon absorbed by NH3 +i.e., ‰OHŠ ˆ
2
rNH ‰NH3Š). Therefore,
3
r
‰NO2Š
‰NH3Š
r
r
P
NO2
NO2 NO2
‰OHŠ1=‰OHŠ ˆ 2b
ˆ 2b
2
r
NH3
P
NH3 NH3
ꢀ15†
In the present study [OH]1/[OH]2 was measured
as 0.079. When this value, the previously deter-
mined value of b, and the appropriate partial
pressures for NO2 and NH3 were substituted into
expression +15), rNO =rNH was estimated as
2
3
3:1 Â 10À2
.
5. Discussion
A recent review [17] of the reaction between
O+1D) and H2 identi®ed the major +>95%) reac-
tion channel [20] as +4), and recommended the
use of a rate constant for this channel of
1:1ꢀ2†  10À10 cm3 sÀ1. If this number is combined
with the value of k3=k4 measured in the present
study, a value of 1:5ꢀ3†  10À10 cm3 sÀ1 is ob-
Fig. 3. Frequency scans of the OH line at 3407.989 cmÀ1 at
various times after the photolysis ¯ash of a mixture containing
31 mTorr of NO2, 548 mTorr of H2 and 9 Torr. The ®ve scans
were made 10±20, 30±40, 50±60, 100±110, and 200±210 ls after
the ¯ash. The baseline of each successive scan was increased by
10 mV in order to improve visibility. The solid line though each
scan is a Gaussian ®t to the data.

F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
77
tained for k3. This value is about one half of that
measured by Heidner and Hussain [21], and later
recommended by Cvetanovic [22]. However, reso-
nance absorption measurements of the type used
in [21] do not directly measure a rate constant, but
instead determine the product ck, where c is a non-
trivial factor which relates the absorbance to the
concentration of O+1D). In 1982, Wine and Ravi-
shankara [23] noted that O+1D) rate measurements
made by using resonance absorption spectroscopy,
are frequently a factor of 2 higher than those ob-
tained by using a variety of other techniques. The
value for k3 obtained here is consistent with this
pattern.
bance from t ˆ 105 ls to t ˆ 0, that due to possi-
ble error in the value of b used in Eq. +15), and
uncertainty in our knowledge of rNH . This later
3
uncertainty arises primarily because the NH3
spectrum exhibits rotational structure over the
ArF laser line. Although this structure is di€use,
the absorption coecient has been reported to
vary by as much as 20% across the entire excimer
wavelength [26]. In view of the above uncertain-
ties, it would be unrealisticto specify bounds on
the measured value of rNO closer than Æ40%.
2
Even so, the measurement does suggest that the
cross-section lies near the lower end of the range
spanned by the previously reported measurements.
The branching ratio for O+1D) production was
estimated as 0:55 Æ 0:03. Here, the indicated error
represents the propagated standard deviation ob-
tained from a number of di€erent measurements of
6. Conclusions
The O+1D) branching ratio for ArF excimer
photolysis of NO2 at 193 nm has been measured
as 0:55 Æ 0:03. In addition, the rate constant for
the reaction between O+1D) and NO2 has been
AOH and DANO . A possible source of systematic
2
error arises from the neglect of pressure broaden-
ing. In the present study, the linewidths +1=e half-
widths) of OH and NO2 were measured by ®tting a
Gaussian lineshape to the appropriate frequency
scans as 185 Æ 6 and 97 Æ 2 MHz, respectively. At
296 K, the Doppler linewidths for OH and NO2
were calculated as 183.4 and 95.4 MHz, respec-
tively. Therefore, it is clear that pressure broad-
ening did not signi®cantly a€ect our intensity
measurements.
measured as ꢀ1:5 Æ 0:3†  10À10 cmÀ3
s
À1, and the
193 nm absorption cross-section for NO2 has been
estimated as ꢀ2:9 Æ 1:2†  10À19 cm2.
Acknowledgements
The O+1D) branching ratio measured at 193 nm
is 14% higher than that determined by Uselman
and Lee [7] at somewhat lower energies. In both
studies, photo-dissociation probably occurred
This work was supported by grants from the
Department of Energy and the Robert A. Welch
Foundation.
2
~
from the NO2+B B2) state. The somewhat higher
branching ratio measured in the present study may
result from the increased number of NO + O+1D)
product states available at higher energies.
References
[1] A.R. Ravishankara, F.L. Eisele, P.H. Wine, J. Chem. Phys.
73 +1980) 3743.
Previous measurements of the 193 nm absorp-
tion cross-section for NO2 have been thoroughly
reviewed by Schneider et al. [24]. It is generally
agreed that the cross-section lies between 2.7 and
5:4 Â 10À19 cm2. In this study, rNO =rNH was es-
[2] M. Wollenhaupt, J.C. Crowley, J. Phys. Chem. A 104
+2000) 6429.
[3] R.W. Quandt, J.F. Hershberger, J. Phys. Chem. 100 +1996)
9407.
[4] J. Park, M.C. Lin, J. Phys. Chem. 101 +1997) 2643.
[5] F. Sun, J. DeSain, G. Scott, P.Y. Hung, R.I. Thompson,
G.P. Glass, R.F. Curl, J. Phys. Chem. A submitted.
[6] R. Atkinson, D.L. Baulch, D.R.A. Cox, R.F. Hampson
Jr., J.A. Kerr, M.J. Rossi, J. Troe, J. Phys. Chem. Ref.
Data. 26 +1997) 1329.
2
3
timated as 3:1 Â 10À2. By using this value, together
with the well-known cross-section [25,26] for NH3,
rNO can be estimated as 2:9 Â 10À19 cm2. How-
2
ever, uncertainties in this estimate arise from sev-
eral sources. Among these are: the uncertainty
arising from the extrapolation of the OH absor-
[7] W.M. Uselman, E.K.C. Lee, Chem. Phys. Lett. 30 +1975)
212.

78
F. Sun et al. / Chemical Physics Letters 337 *2001) 72±78
[8] R.K. Ritchie, A.D. Walsh, P.A. Worsop, Proc. Roy. Soc.
A 266 +1962) 257.
[17] R. Atkinson, D.L. Baulch, D.R.A. Cox, R.F. Hampson
Jr., J.A. Kerr, M.J. Rossi, J. Troe, J. Phys. Chem. Ref.
Data. 26 +1997) 521.
[9] M.R. Taherian, P.C. Cosby, T.G. Slanger, J. Phys. Chem.
91 +1987) 2304.
[18] F. Yamada, I.R. Slagle, D. Gutman, Chem. Phys. Lett. 83
+1981) 409.
[10] R.K. Ritchie, A.D. Walsh, Proc. Roy. Soc. A 267 +1962)
395.
[19] R. Atkinson, J. Phys. Chem. Ref. Data. 26 +1997) 215.
[20] P.H. Wine, A.R. Ravishankara, Chem. Phys. 69 +1982)
365.
[11] J.D. DeSain, P.Y. Hung, R.I. Thompson, G.P. Glass,
G. Scuseria, R.F. Curl, J. Phys. Chem. A 104 +2000) 3356.
[12] L.S. Rothman, C.P. Rinsland, A. Goldman et al., J. Quant.
Spectros. Radiat. Transfer. 60 +1998) 665.
[13] F.H. Verhoek, F. Daniels, J. Am. Chem. Soc. 53 +1931)
1250.
[21] R.F. Heidner, D. Husain, Int. J. Chem. Kinetics 5 +1973)
819.
[22] R.J. Cvetanovic, Canad. J. Chem. 52 +1974) 1452.
[23] P.H. Wine, A.R. Ravishankara, Chem. Phys. Lett. 77
+1981) 103.
[14] J.E. Butler, R.G. Macdonald, D.J. Donaldson, J.J. Sloan,
Chem. Phys. Lett. 95 +1983) 183.
[24] W. Schneider, G. Moorgat, G.S. Tyndall, J.P. Burrows,
J. Photochem. Photobiol. A 40 +1987) 195.
[25] M. Suto, L.C. Lee, J. Chem. Phys. 78 +1983) 4515.
[26] D.F. Davidson, A.Y. Chang, K. Kohse-Honghaus, R.K.
Hansen, J. Quant. Spectros. Radiat. Transfer. 42 +1989)
267.
[15] D.H. Ja€er, I.W.M. Smith, Faraday Discuss Chem. Soc. 67
+1979) 212.
[16] R. Forster, M. Frost, D. Fulle, H.F. Hammann, H.F.
Hippler, A. Schlepegrell, J. Troe, J. Phys. Chem. 103 +1995)
2949.