Formation of HCNO and HCN in the 193 nm photolysis of H2CCO in the presence of NO

Article abstract of DOI:10.1016/S0009-2614(02)00387-1

NO-reburning is one of the main strategies for removing NO_x from the combustion of fossil fuels. The products derived from the ArF laser photolysis of H₂CCO in the presence of a large excess of NO at 298 K and 3.5 mbar in He bath gas were studied in a tubular flow reactor using time-resolved MS. HCNO and HCN were the major reaction products observed in the presence of NO. The observed yields, HCNO = 0.18 and HCN = 0.04 were attributed to the contributions of the reactions CH₂ + NO → HCNO + H, CH₂ + NO → HCN + OH, and HCCO + NO → HCNO + CO (HCCO + NO → HCN + CO₂), which confirmed results of earlier FTIR end-product analyses. The efficient production of HCNO by the reactions CH₂ + NO → products and HCCO + NO → products should be taken into account in improved NO_x-reburn models.

Full text of DOI:10.1016/S0009-2614(02)00387-1

15 April 2002
Chemical Physics Letters 356 (2002) 181–187
www.elsevier.com/locate/cplett
Formation of HCNO and HCN in the 193 nm photolysis
of H2CCO in the presence of NO
*
€
G. Eshchenko, T. Kocher, C. Kerst, F. Temps
Institut f€ur Physikalische Chemie, Christian-Albrechts-Universit€at zu Kiel, Olshausenstr. 40, D-24098 Kiel, Germany
Received 7 February 2002; in final form 28 February 2002
Abstract
The products obtained from the ArF laser photolysis of H2CCO in the presence of a large excess of NO at
T ¼ 298 Kand p ¼ 3:5 mbar in He bath gas were investigated in a tubular flow reactor by time-resolved mass
spectrometry. The main reaction products observed in the presence of NO were HCNO and HCN. The measured yields,
CHCNO ¼ 0:78 Æ 0:18 and CHCN ¼ 0:19 Æ 0:04, respectively, were attributed to the contributions of the reactions
CH2 þ NO ! HCNO þ H
CH2 þ NO ! HCN þ OH
and
ð1aÞ
ð1bÞ
HCCO þ NO ! HCNO þ CO
ðHCCO þ NO ! HCN þ CO2
ð2aÞ
ð2bÞ
The time-resolved measurements are compared to results of preceding end product analyses by Fourier transform IR
spectroscopy. Ó 2002 Published by Elsevier Science B.V.
1. Introduction
NH2 and NO is reduced via the reaction
NH2 þ NOþ ! N2 þ H2O [8].
The two reactions
NO-reburning is one of the main strategies for
the removal of NOx from the combustion of fossil
fuels (see, e.g., [1–7]). The process depends on the
reactions of small hydrocarbon radicals, especially
CH2 [4], HCCO [5,7], and CH3 [4–6] which recycle
NO into the combustion mechanism. As a result,
consecutive reactions may occur which lead to
CH2 þ NO ! products
and
ð1Þ
HCCO þ NO ! products
ð2Þ
which play key roles in the NOx-reburn process
have been investigated in the last years by a
number of research groups [9–31]. Their main
product channels have, however, remained un-
known for a rather long time [14,17,29].
* Corresponding author. Fax: +49-431-880-1704.
E-mail address: temps@phc.uni-kiel.de (F. Temps).
0009-2614/02/$ - see front matter Ó 2002 Published by Elsevier Science B.V.
PII: S0009-2614(02)00387-1

182
G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
In recent publications from this laboratory, we
were generated by low energy electron impact
ionization. The ions were extracted at right angles,
mass filtered in the quadrupole analyzer, and
monitored using a channeltron detector, connected
via a preamplifier/discriminator (Stanford Re-
search SR445) to a transient recorder (Spectrum
PAD280A) PC card for single ion counting and
averaging. The ion signals from 1000 to 10 000
laser shots were accumulated to record the kinetic
concentration–time profiles at a selected m/z value.
The complete experiment, including gas flows,
pressure, timing pulses, and the data acquisition
reported on the detection of HCNO and HCN in
end product studies of reactions (1) and (2) by
Fourier transform infrared (FTIR) spectroscopy
[14,17,29]. The products were determined in the
photolysis of mixtures of H2CCO=NO=Ar and
NO2=NO=C2H2=Ar, respectively, using light from a
high pressureHg lamp at k ¼ 313 nm inastationary
cell. While the molecules could be identified from
their FTIR spectra without ambiguities, the exper-
imental method did not permit direct time-resolved
measurements and contributions of slow consecu-
tive reactions could not be entirely ruled out. This
Letter reports on a reinvestigation of reactions (1)
and (2) using time-resolved mass spectrometry with
ArF laser photolysis of H2CCO in the presence of a
large excess of NO in a tubular slow flow reactor to
substantiate the previous conclusions.
was controlled by LABVIEW software.
The He carrier gas was of the highest com-
mercially available purity (99.9999%, Messer
Griesheim). NO (99.5%, Messer Griesheim) was
purified by repeated fractional distillation until no
impurities of NO2 or N2O were detectable in FTIR
spectra. H2CCO was prepared by pyrolysis of
acetic acid anhydride [14]. Reference samples of
HCNO and HCN for determination of their mass
spectrometric detection sensitivities were synthe-
sized by pyrolysis of 3-phenyl-4-oxoiminoisoxazol-
5-(4H)-one and by heating mixtures of NaCN with
stearic acid, respectively [14]. The NO, H2CCO,
HCNO, and HCN were premixed with He and
taken from glass storage bulbs. Mixtures of
HCNO were used immediately to avoid decom-
position; all other mixtures were allowed to mix
thoroughly before use. The gases were regulated
with calibrated mass flow controllers (Aera), the
pressure in the reactor was measured with a ca-
pacitance pressure transducer (MKS Baratron).
2. Experimental
The experimental investigations were carried
out in a 75 cm long, 1.7 cm i.d. quartz reactor
connected to a molecular beam sampling quadru-
pole mass spectrometer (Extrel C50 and Bruker
MM1) [32]. Radicals were produced by pulsed
excimer laser photodissociation of H2CCO in He
inert carrier gas at k ¼ 193 nm in the absence and
presence of a large excess of NO. The laser beam
was steered along the reactor axis using dielectri-
cally coated mirrors through a series of apertures
to ensure a homogeneous radial intensity profile.
The fused silica windows at both ends were purged
with He inert gas to avoid deposition of carbon on
the windows. The laser pulse energies were mea-
sured with a calibrated power meter (coherent)
before and behind the reactor. Corrected for the
light reflected at the windows, both values agreed
within 65%. The laser was operated at a repetition
rate of 3 Hz and slow flow conditions were used to
replace the gases between the laser shots.
3. Results
3.1. Reaction system
The photolysis of H2CCO at k ¼ 193 nm can
proceed via the following channels:
H2CCO þ hm ð193 nmÞ
A molecular beam into the high vacuum
chamber (p 6 10À6 mbar) of the mass spectrome-
ter was formed by the expansion of a small
amount of the gases through a 0.7 mm i.d. conical
pinhole in the reactor wall, a differential pumping
stage, and a 0.5 mm i.d. skimmer. Molecular ions
3
! CH2 þ CO U3a ¼ 0:63
ð3aÞ
ð3bÞ
ð3cÞ
ð3dÞ
1
! CH2 þ CO U3b ¼ 0:19
! HCCO þ H U3c ¼ 0:11
! CCO þ H2 U3d ¼ 0:07

G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
183
The photodissociation yields U3a–d were taken
from work by Glass et al. [33]. A small yield of
HCCO was also found by Rim and Hershberger
[30]. Accordingly, CH2+CO are the main products
(82%). The CH2 is supposed to be formed mainly
The reactions of 3CH2 (1t)and 1CH2 (1s) have been
written separately here, but they are expected to
give the same products [17] so that the two spin
states do not need to be distinguished in the fol-
lowing. The rate constants have been determined
in an earlier work (k1t ¼ 2 Â 1013 cm3 molÀ1 sÀ1
[11,12], k1s ¼ 1:5 Â 1014 cm3 molÀ1 sÀ1 [18–20],
3
3
e
in its X B1 ground electronic state ð CH2Þ, only a
1
~
side channel [33] should lead to CH2 in its a A1
1
excited state ð CH2Þ. The CH2 (a) is rapidly de-
k2 ꢀ 1:8 Â 1013 cm3 molÀ1
s
À1 [25–27]). The fate of
~
activated by the He bath gas [18–20]. A possible
1
the H atoms and OH radicals produced in reaction
(1) has been discussed previously [14,17]. Their
main reaction pathways are
~
formation of CH2 in the b B1 state is included in
e
the CH2 (X) and (a) yields because of its short
~
radiative lifetime [33]. The resulting radical con-
centrations were determined using the H2CCO
absorption cross-section ðr193 nm ¼ ð1:0 Æ 0:1Þ Â
10À18 cm2Þ, which was measured with a UV/VIS
desktop spectrometer (Shimadzu UV-2101PC).
The calculated concentrations were checked in
other experiments [32] by comparing the HCl from
the reaction Cl þ SiH4 in the 193 nm photolysis of
mixtures of CCl4 or ðCOClÞ2 with SiH4 by (i) di-
rect mass spectrometric calibration and (ii) calcu-
H þ H2CCO ! CH3 þ CO
ð6Þ
OH þ H2CCO ! CH2OH þ CO
! CH3 þ CO2
No interferences are expected from these reactions
regarding the yields of HCNO and HCN, except
perhaps (see discussion) from the reaction
ð7aÞ
ð7bÞ
H þ HCNO ! HCN þ OH
ð8Þ
The CCO radicals from (3d) and their products
with NO cannot play a significant role in the
reaction system due to their small concentra-
tions.
lation from the known CCl4 [34] or ðCOClÞ [35]
2
absorption cross-sections.
In the absence of NO, the CH2, HCCO, and
CCO radicals from the photolysis (3a),(3b),(3c),
(3d) can react only with the H2CCO, which is
present in large excess, or, at higher concentra-
tions, undergo different mutual combination re-
actions. In particular, the reactions of CH2 (either
3.2. HCNO and HCN yields
Experimental measurements were performed at
T ¼ 296 Kand p ¼ 3:5 mbar. The concentrations
of H2CCO were varied in the range 7:7 Â 10À11 mol
1
3CH2 or CH2) with H2CCO give C2H4 [20]
CH2 þ H2CCO ! C2H4 þ CO
and the reactions
ð4Þ
cmÀ3 6 ½H2CCOꢁ 6 6:6 Â 10À10 mol cmÀ3, with ½NOꢁ ¼
0
4:22 Â 1010 mol cmÀ3. With photolysis pulse ener-
gies of typically 37 mJ (average values before and
after the reactor, Æ5 mJ variations between differ-
ent experimental runs), radical concentrations were
3
3CH2 þ CH2 ! C2H2 þ H2
ð5aÞ
ð5bÞ
! C2H2 þ 2H
generated in the range 7:5 Â 10À13 mol cmÀ3
6
lead to C2H2 [20]. The H atoms and HCCO and
CCO radicals are mostly consumed by the
H2CCO or in different fast radical–radical reac-
tions.
In the presence of a large excess of NO, the
system is dominated by the fast reactions
½CH2ꢁ 6 6:4 Â 10À12 mol cmÀ3 and 1:0 Â 10À13 mol
0
cmÀ3 6 ½HCCOꢁ0 6 8:4 Â 10À13 mol cmÀ3, respec-
tively. Under these conditions, the CH2 and HCCO
decays due to reactions (1) and (2) were practically
complete in reaction times of Dt 6 1 ms.
Without NO, strong mass peaks of reaction
products were detected only at m=z ¼ 28ðC2H4Þ;
26ðC2H2Þ, and 15 (CH3; see above). An additional
weak signal at m=z ¼ 27 was attributed to C2H3,
which can be produced in the system in secondary
steps.
3CH2 þ NO ! products
1CH2 þ NO ! products
HCCO þ NO ! products
ð1tÞ
ð1sÞ
ð2Þ

184
G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
Upon addition of NO, the above products dis-
appeared, and strong mass peaks appeared at
m=z ¼ 43 and 27. In agreement with the FTIR
observations [14,17,29], these product signals were
attributed to HCNO and HCN from the reaction
channels
CH2 þ NO ! HCNO þ H
! HCN þ OH
ð1aÞ
ð1bÞ
HCCO þ NO ! HCNO þ CO
! HCN þ CO2
ð2aÞ
ð2bÞ
Typical concentration–time profiles which show
the fast product formation after the photolysis
pulse at t ¼ 0 are depicted in Fig. 1. The observed
rise times are determined by the ensuing reaction
kinetics and by the time constant of the experi-
mental setup. Both reaction products appear to be
stable on the time scale of the experiments (up to
Fig. 2. Measured concentrations of HCNO and HCN plotted
as function of the radical concentrations ð½CH2ꢁ þ ½HCCOꢁÞ
0
from the photolysis of the H2CCO.
Dt ¼ 46 ms). Although the observed time profiles
show significant scatter, the product yields could
be determined with good precision (6Æ 3%) due to
the large number of points in a profile. The final
product yields were determined by taking the av-
erage values from the measured time profiles be-
tween 5 and 45 ms reaction time. As a caveat, it is
noted that the m/z ¼ 43 signal (HCNO) was ob-
served on a relatively high background signal,
which was attributed to ketene precursor mole-
cules with one D or one 13C atom at the same m/z
value.
In Fig. 2, the measured HCNO and HCN
concentrations have been plotted vs. the concen-
trations of ð½CH2ꢁ þ ½HCCOꢁÞ0. As seen, the data
points can be nicely described by straight lines
through the origin. The slopes give values for the
product yields with respect to ð½CH2ꢁ þ ½HCCOꢁÞ
0
of
CHCNO ¼ 0:78 Æ 0:18 and CHCN ¼ 0:19 Æ 0:04:
These yields arise from the simultaneous contri-
butions of reactions (1) and (2) in the system.
4. Discussion
The highly exothermic reactions of CH2 and
HCCO with NO can, in principle, lead to a large
number of products. OH radicals and H atoms
Fig. 1. Measured concentration–time profiles of HCNO (open
circles) and HCN (filled circles) at two H2CCO concentrations.

G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
185
were first detected in laser magnetic resonance
(LMR) and electron spin resonance (ESR) mea-
surements [11]. Atakan et al. [13] also reported on
the appearance of OH and HCN and estimated a
3CH2 þ NO rate constant value from the OH
profile. An observed mass signal at m=z ¼ 44 was
attributed to long-lived H2CNO complexes [13].
Bauerle et al. [15] investigated the formation of H
atoms and OH radicals in shock tube experiments.
Su et al. [16] reported on an FTIR emission study
and suggested a number of other product chan-
nels, but questions have been noted regarding the
assignments [17]. Eventually, Grußdorf et al. [14]
and Fikri et al. [17] determined quantitative
product yields of HCNO (1a) and HCN (1b) in the
photolysis of mixtures of H2CCO/NO/Ar by FTIR
spectroscopy. With small corrections for second-
ary reactions, they arrived at branching ratios of
k1a=k1 ¼ 0:89 Æ 0:06 and k1b=k1 ¼ 0:11 Æ 0:06.
Unimolecular rate calculations [17] using input
from quantum chemical potential energy calcula-
tions [21–24] gave channel branching ratios in
good agreement with the measurements.
The product distribution for HCCO þ NO have
been analyzed by mass spectrometry [27] and by
FTIR [29] and diode laser spectroscopy [30].
Boullart et al. [27] first proposed HNCO þ CO as
the main channel, but their mass spectra can be
reassigned to HCNO þ CO [31]. With this as-
sumption, the reported measurements [27,29,30]
for the HCNO þ CO (2a) and HCN þ CO2 (2b)
channels are in agreement. According to the FTIR
work, the branching ratios are k2a=k2 ¼ 0:64 Æ
0:12 and k2b=k2 ¼ 0:28 Æ 0:10 [29].
(Æ10% including estimated systematic error lim-
its), the determination of the absolute photolysis
pulse energies (Æ10%), the measurements of the
product time profiles (Æ10%, including statistical
errors and estimated systematic uncertainties), and
the absolute mass spectrometric calibrations of the
HCNO (Æ10%) and HCN (Æ5%). Uncertainties
resulting from errors of the gas pressures and flow
1
rates (Æ5%) or the effect of the reaction of CH2
with the H2CCO (6Æ 5%) are comparatively
small. Errors of the CH2 and HCCO quantum
yields in the H2CCO photolysis taken from Glass
et al. [33] are less important only as long as the
total yield of CH2 þ HCCO does not change, since
the CH2 þ NO and HCCO þ NO reactions both
give HCNO and HCN. It is only an assumption
here that U3a þ U3b þ U3c ¼ 0:93 Æ 0:05. With
these premises, the root mean square error limits
of the HCNO and HCN product yields CHCNO and
CHCN are of the order of Æ20%, as quoted above.
Adopting the 81% CH2 yield from the
k ¼ 193 nm H2CCO photolysis of Glass et al. [33],
the present results apply essentially directly to re-
action (1). More precisely, however, the yields in
the present work arise from the simultaneous
contributions of reactions (1) and (2) in the system.
A better comparison is made by combining the
quoted branching ratios for the k ¼ 193nm
H2CCO photolysis with the FTIR branching ra-
tios, which gives
ðU3a þ U3bÞ Â k1a=k1 þ U3c  k2a=k2
CHCNO
¼
U3a þ U3b þ U3c
¼ 0:86 Æ 0:08;
The time-resolved observations of HCNO and
HCN in the present experimental study show that
both molecules are direct reaction products. Thus,
the conclusions from the FTIR determinations
[17,29] are confirmed. The time resolution was fast
enough to rule out slow secondary reactions (for
example heterogeneous reactions), although it was
too slow for direct kinetics information to be de-
termined.
A few comments are useful regarding the error
limits of the experiments. Measurements of abso-
lute product yields are generally difficult. The main
sources of errors in the experiments are due to the
uncertainty of the H2CCO absorption coefficient
ðU3a þ U3bÞ Â k1b=k1 þ U3c  k2b=k2
U3a þ U3b þ U3c
CHCN
¼
¼ 0:13 Æ 0:08:
It is seen that the yields for HCNO from the pre-
sent work and from the FTIR analyses are in good
agreement within their error limits. The present
yield for HCN is somewhat larger than that ex-
pected from the FTIR data, but still in agreement
within the combined uncertainties. Only little
room seems to be left for other product channels.
The remaining small difference of the CHCN
values may be explained by several effects. First,

186
G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
5. Conclusions
regarding the present work, it is possible that a
small amount of the HCNO from (1a) and (2a)
may have been converted to HCN via the reac-
tion H+HCNO (8). Reaction (8) is assumed in
the current NOx-reburn mechanisms [7]. No
measurements are available for its rate constant
to the best of our knowledge, so that its quan-
titative effect cannot be taken into account very
easily. Assuming a rate constant in the range of
The products of the reactions CH2 þ NO (1)
and HCCO+NO (2) were investigated at T ¼
298 Kand p ¼ 3:5 mbar by ArF excimer laser
photolysis of H2CCO in the presence of a large
excess of NO. The reaction products HCNO and
HCN were detected by time-resolved mass spect-
rometry. The observed yields of CHCNO ¼ 0:78 Æ
0:18 and CHCN ¼ 0:19 Æ 0:04 with respect to
ð½CH2ꢁ þ ½HCCOꢁÞ0 confirm results of earlier
FTIR end product analyses [14,17,29]. The effi-
cient production of HCNO by reactions (1) and (2)
has to be taken into account in improved NOx-
reburn models.
k8 ꢀ 5 Â 1012 cm3 molÀ1
s
À1, some conversion of
HCNO to HCN (620%) would indeed be ex-
pected for the experimental conditions used. A
careful inspection of the HCN time profiles at
initial high radical concentrations indicates a
slow increase after the initial fast rise immedi-
ately after the laser pulse, which may be due to
reaction (8). That this rise is not reflected by a
corresponding decrease of the HCNO time pro-
files is due to the poor signal-to-noise ratio (see
above) for this product.
Second, it is noted that the channel branching
ratios were determined in the FTIR work from the
initial HCNO and HCN formation rates at short
photolysis times, because the yields at longer times
were affected by a slow decomposition of the un-
stable HCNO. It is possible that the HCNO de-
composition was slightly ‘‘overcorrected’’ by this
analysis [17].
Acknowledgements
The financial support of this work by the
Deutsche Forschungsgemeinschaft and the Fonds
der Chemischen Industrie is gratefully acknowl-
edged.
References
[1] J.O.L. Wendt, C.V. Sternling, M.A. Matovich, Proc.
Combust. Inst. 14 (1972) 897.
A comment is in order eventually regarding the
k ¼ 193 nm H2CCO photolysis, which was em-
ployed in this work, compared to a photolysis at
k ¼ 308 nm or 351 nm, which would have pro-
[2] S.L. Chen, J.M. McCarthy, W.D. Clark, M.P. Heap, W.R.
Seeker, D.W. Pershing, Proc. Combust. Inst. 21 (1986)
1159.
[3] T. Kolb, P. Jansohn, W. Leuckel, Proc. Combust. Inst. 22
(1988) 1193.
1
duced practically only CH2 (at k ¼ 308 nm) or
3
[4] T. Etzkorn, S. Muris, J. Wolfrum, C. Dembny, H.
Bockhorn, P.F. Nelson, A. Attia-Shahin, J. Warnatz,
Proc. Combust. Inst. 24 (1992) 925.
only CH2 þ CO (at k ¼ 351 nm). However, the
much lower H2CCO absorption coefficients [34] at
those wavelengths require much larger precursor
concentrations to detect any reaction products
[5] D. Stapf, W. Leuckel, Proc. Combust. Inst. 26 (1996) 2083.
[6] J.A. Miller, C.T. Bowman, Progr. Energy Sci. Combust. 15
(1989) 287.
1
with NO. Under those conditions, the CH2 reac-
[7] P. Dagout, F. Lecomte, S. Chevalier, M. Cathonnet,
Combust. Sci. Technol. 139 (1998) 329.
tion system (at k ¼ 308 nm) is governed by the
1CH2 þ H2CCO reaction and the main products
are simply C2H4 þ CO. Tentative measurement
performed at k ¼ 351 nm did show a formation of
HCN from reaction (1b), but observations of
HCNO were impossible due to the large back-
ground mass signal at m=z ¼ 43 at the high
½H2CCOꢁ. Quantitative HCNO and HCN yield
measurements could only be performed by
H2CCO photolysis at k ¼ 193 nm.
[8] M. Gehring, K. Hoyermann, H. Schacke, J. Wolfrum,
Proc. Combust. Inst. 14 (1972) 99.
[9] A.H. Laufer, A.M. Bass, J. Phys. Chem. 78 (1974) 1344.
[10] C. Vinckier, W.J. Debruyn, J. Phys. Chem. 83 (1979) 2057.
[11] V. Seidler, F. Temps, H.Gg. Wagner, M. Wolf, J. Phys.
Chem. 93 (1989) 1070.
[12] D.C. Darwin, A. Young, H.S. Johnston, C.B. Moore, J.
Phys. Chem. 93 (1989) 1074.
[13] B. Atakan, D. Kocis, J. Wolfrum, P. Nelson, Proc.
Combust. Inst. 24 (1992) 691.

G. Eshchenko et al. / Chemical Physics Letters 356 (2002) 181–187
187
[14] J. Grußdorf, F. Temps, H.Gg. Wagner, Ber. Bunsenges.
Phys. Chem. 101 (1997) 134.
[25] K.G. Unfried, G.P. Glass, R.F. Curl, Chem. Phys. Lett.
177 (1991) 33.
[15] S. Bauerle, M. Klatt, H.Gg. Wagner, Ber. Bunsenges. Phys.
Chem. 99 (1995) 97.
[26] F. Temps, H.Gg. Wagner, M. Wolf, Z. Phys. Chem. 176
(1992) 27.
[16] H. Su, F. Kong, B. Chen, M.-B. Huang, Y. Liu, J. Chem.
Phys. 113 (2000) 1885.
[27] W. Boullart, M.T. Nguyen, J. Peeters, J. Phys. Chem. 98
(1994) 8036.
[17] M. Fikri, St. Meyer, J. Roggenbuck, F. Temps, Faraday
Disc. 110 (2001) 223.
[28] M.T. Nguyen, W. Boullart, J. Peeters, J. Phys. Chem. 98
(1994) 8030.
[18] M.N.R. Ashfold, M.A. Fullstone, G. Hancock, G.W.
Ketley, Chem. Phys. 55 (1981) 245.
[29] U. Eickhoff, F. Temps, Phys. Chem. Chem. Phys. 1 (1999)
243.
[19] A.O. Langford, H. Petek, C.B. Moore, J. Chem. Phys. 78
(1983) 6650.
[30] K.T. Rim, J.F. Herschberger, J. Phys. Chem. A 104 (2000)
293.
€
[20] H.-H. Carstensen, Dissertation, Max-Planck-Institut fur
€
[31] L. Vereecken, R. Sumathy, S.A. Carl, J. Peeters, Chem.
Phys. Lett. 344 (2001) 400.
€
Stromungsforschung, Report 4/95, Gottingen, 1995.
[21] J. Roggenbuck, F. Temps, Chem. Phys. Lett. 285 (1998)
€
[32] C. Kerst, T. Kocher, and F. Temps, to be published.
422.
[22] W.A. Shapley, G.B. Bacskay, Theor. Chim. Acta 100
[33] G.P. Glass, S.S. Kumaran, J.V. Michael, J. Phys. Chem. A
104 (2000) 8360.
(1998) 212.
[23] W.A. Shapley, G.B. Bacskay, J. Phys. Chem. A 103 (1999)
[34] H. Okabe, Photochemistry of Small Molecules, Wiley,
New York, 1978.
4505.
[24] W.A. Shapley, G.B. Bacskay, J. Phys. Chem. A 103 (1999)
4515.
[35] A.V. Baklanov, L.N. Krasnoperov, J. Phys. Chem. A 105
(2001) 4917.