Fast-flow study of the C+NO and C+O2 reactions

Article abstract of DOI:10.1016/S0009-2614(99)00586-2

The C+NO and C+O₂ reactions were studied, at room temperature, in a low-pressure fast-flow reactor. C atoms were obtained from the reaction of CBr₄ or CCl₄ with potassium atoms and probed by their VUV resonance fluorescence. For C+NO, the overall rate constant was determined as (5.4±0.8)×10^-11 cm³ molecule^-1 s^-1 and for C+O₂ as (2.5±0.4)×10^-11 cm³ molecule s^-1. For the C+NO reaction, yielding N(²D, ⁴S)+CO and O(³P)+CN, the atomic products were also probed by their VUV resonance fluorescence. The [N(²D)+N(⁴S)]/[O(³P)] branching ratio was estimated as 1.5±0.3.

Full text of DOI:10.1016/S0009-2614(99)00586-2

16 July 1999
Ž
.
Chemical Physics Letters 308 1999 7–12
www.elsevier.nlrlocatercplett
Fast-flow study of the CqNO and CqO2 reactions
)
A. Bergeat, T. Calvo, G. Dorthe, J.C. Loison
Laboratoire de Physico-Chimie Moleculaire, UMR 5804, CNRS-UniÕersite Bordeaux I, 351 Cours de la Liberation, F-33405 Talence
´
´
´
Cedex, France
Received 15 April 1999; in final form 20 May 1999
Abstract
The CqNO and CqO2 reactions were studied, at room temperature, in a low-pressure fast-flow reactor. C atoms were
obtained from the reaction of CBr4 or CCl4 with potassium atoms and probed by their VUV resonance fluorescence. For
y
11
Ž
.
CqNO, the overall rate constant was determined as 5.4"0.8 =10
cm3 moleculey1 sy1 and for CqO2 as
4
Ž
.
Ž2
.
Ž3
.
y
11
2.5"0.4 =10
cm3 molecule sy1. For the CqNO reaction, yielding N D, S qCO and O P qCN, the atomic
w Ž2
.
Ž4 .x w Ž3 .x
products were also probed by their VUV resonance fluorescence. The
N
D qN S r O P branching ratio was
estimated as 1.5"0.3. q 1999 Elsevier Science B.V. All rights reserved.
1
™ N 2 P qCO Sq
,
,
Dr H2o98 sy1.05 eV ,
1. Introduction
Ž
.
Ž
.
.
1c
Ž
.
.
Atomic carbon reactions are involved in combus-
tion as well as in the interstellar medium for those
exhibiting no barrier in the entrance channel. How-
ever, for many of them, there are still uncertainties
about the overall rate constants and the product
branching ratios, mainly because the generation of
ground-state carbon atoms clean of other reactive
species is not easy. This is the case for CqNO and
CqO2 whose channel exothermicities are hereafter
given with respect to ground-state products:
2
™ O 3P qCN Sq
Dr H2o98 sy1.24 eV ,
Ž
.
Ž
1d
Ž
1
2 Ž3
.
Ž
.
.
Ž
.
C
3P qO Sy ™ O 3P qCO Sq
Dr H2o98 sy5.99 eV ,
,
Ž
.
2a
Ž
.
.
1
™ O 1D qCO Sq
,
Dr H2o98 sy4.02 eV .
Ž
.
Ž
2b
Ž
For CqNO, the reported values of the overall
rate constant at 300 K, expressed in units of 10y11
w x
2 , 2.7"0.2 3 . A first study in a shock tube, over
the temperature range 2720–3810 K, found no tem-
perature dependence of the rate constant at a value of
1
C
3P qNO 2 P ™ N 4S qCO Sq
,
Ž
.
Ž
.
Ž
.
Ž
.
cm3 moleculey1 sy1, are: 4.8"0.8 1 , 1.6"0.2
Dr H2o98 sy4.62 eV ,
1a
Ž
.
.
w x
w x
1
™ N 2 D qCO Sq
,
Dr H2o98 sy2.24 eV ,
Ž
.
Ž
.
1b
Ž
w x
3.3"1.2 4 . A further shock tube study, over the
range 1550–4050 K, again found no temperature
)
Corresponding author. Fax: q33
jc@lpcm.u-bordeaux.fr
5
5684 6645; e-mail:
dependence of the rate constant at a value of 7.7"3.5
w x
5 . In the latter study, the product branching ratio
0009-2614r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.
Ž
.
PII: S0009-2614 99 00586-2

(
)
8
A. Bergeat et al.rChemical Physics Letters 308 1999 7–12
could be determined in the temperature range 2430–
4040 K, as 60% for the COqN product channel and
thus 40% for CNqO. For CqO2 , the reported rate
constants at 300 K, expressed in units of 10y11 cm3
The CqNO and CqO2 kinetics were studied, at
room temperature, in a fast-flow reactor. C atoms
were obtained by the successive abstractions of Br or
Cl atoms from CBr4 or CCl4 by K atoms and
detected by their VUV fluorescence. In the case of
CqNO, an estimation of the branching ratio be-
tween NqCO and CNqO was made from the
moleculey1 sy1, are the following: 2.6"0.3 1 ,
w x
w x
w x
w x
4.7"0.3 2 , 1.6"0.2 3 , 4.68 6 . The latter value
has been obtained by the CRESU technique from
experiments in the temperature range 27–295 K for
which the rate coefficient fitted the expression 4.68
Ž4
.
Ž2
.
Ž3
.
detection of N S , N D and O P by their VUV
resonance fluorescence.
=10y11 Tr298 K y0.40. No value at temperatures
exceeding room temperature is available. In a crossed
beam experiment in our laboratory, Costes and Naulin
found no potential energy barrier for the CqO2
Ž
.
2. Experimental
w x
reaction 7 .
2.1. Fast-flow reactor
w x
A theoretical study by Halvick et al. 8 reported
no potential energy barrier for the collinear approach
of C towards the N atom of NO for the CqNO ™
The fast-flow reactor has been described in detail
w
x
elsewhere 12 and only a brief description is thus
given. It consisted of a hollowed-out stainless-steel
block, in which a 36 mm inner-diameter Teflon tube
was inserted, with four perpendicular optical ports
for chemiluminescence and laser-induced fluores-
cence detection. The flow velocity in the reactor was
w x
CNqO channel. Later, Persson et al. 9 calculated
the potential energy profiles for either approach to-
wards the N and O atom. In both cases, they found
no potential energy barrier. However, they pointed
out that although the absence of a barrier to the
formation of CON by the approach of C towards O
would allow the formation of COqN in a direct
process, this channel is entropically disfavoured with
respect to the initial formation of CNO, then iso-
26.5 m sy1, with He )99.995% as carrier gas, at a
total pressure of 2.0 Torr. C atoms were obtained by
the successive abstractions of halogen atoms from
CCl4 or CBr4 by atomic potassium vapour in a
microfurnace ending into a nozzle. A glass tube
introduced the halogenated compound between the
furnace and the nozzle exit. The whole device mix-
ing C atoms escaping from a nozzle with NO or O2
slid along the Teflon inner wall of the reactor. The
distance between the window detection and the noz-
zle exit could vary over the range 0–100 mm with a
0.5 mm precision.
Ž
.
Ž2
.
merising to NCO P which further dissociates into
1
Ž2
.
Ž
.
q
N D qCO X S . In a further study of this group
w
x
Ž2
.
Ž1
.
q
10 , the rate coefficients for the N D qCO S
Ž3
.
Ž2
.
q
and O P qCN S
productions were calculated
over the temperature range 200–4500 K. The
branching ratio for COqN was found not to depart
significantly from 0.6 over the whole temperature
range, in good agreement with Dean et al.’s experi-
w x
mental results 5 . The sum of their rate coefficient
Atoms were detected by their resonance fluores-
cence. Atom excitation was achieved with the mi-
crowave discharge lamp previously used to probe the
expressions for the two product channels leads to an
overall rate constant of 7.8=10y11 cm3 moleculey1
sy1 at 4500 K, again in agreement with Dean et al.’s
values. Unfortunately, at 300 K, the calculated over-
all rate constant amounts to 12.36=10y11 cm3
moleculey1 sy1, far in excess of any experimental
value. Since the reaction is barrierless, capture mod-
w
x
atomic products of the CHqNO reaction 12 . The
flowing gas mixture previously used to get intense
emission lines of N, H and O atoms also gave C
emission lines owing to CO and CO2 impurities. The
general procedure for atomic detection has also been
w
x
w x
els were used by Beghin et al. 11 to calculate the
CqNO rate coefficient over the temperature range
50–500 K. According to these calculations, the rate
coefficient should increase from 3.43=10y11 cm3
moleculey1 sy1, at 300 K, to 13.6=10y11 cm3
moleculey1 sy1, at 50 K.
detailed previously 12 . We would only mention that
the conditions of the presently reported experiments
ensure the linear dependence of the atomic fluores-
cence versus both the lamp emission intensity and
Ž3
3
.
the C, N and O atom concentrations. For C P , the
3
transitions at 156.11 nm 2p D0 l 2p2 P and at
3
Ž
.

(
)
A. Bergeat et al.rChemical Physics Letters 308 1999 7–12
9
3
3
2
0
Ž
.
Ž4
.
.
165.72 nm 3s P l 2p P were used. For N S
reactions will be detailed in a further article. The
variations of the CBr2 , CBr and C signals with the
distance between the end of the glass tube injecting
the halomethane in the K vapour and the nozzle
aperture are given in Fig. 1. In this figure, signals
have been normalised to their maximum. The de-
crease of C atoms was due to the wall removal. The
C production efficiency was very similar for CBr4
and CCl4, but, in the case of CBr4, the limiting step
should be either the second or the third Br abstrac-
tion, while in the case of CCl4 it should be the last
Cl abstraction. The chemical lifetime of CBr in the
oven was found to be smaller than that of CCl for
similar conditions, so that the production of C2 by
CqCBr is smaller than that by CqCCl. Actually,
the reactions C2 qNO ™ CqNCO, C2OqN, and
C2 qO2 ™ CqCO2 , C2OqO could interfere in
the atom probing either for the atomic carbon decay
as well as for the N and O product branching of
CqNO. The relative importance of C2 thus was
estimated. For that purpose, C2 was produced from
C2Cl4 at a concentration equivalent to that of C
atoms from CCl4 and then compared to C2 as a
product of subsequent reactions of C atoms. This
comparison was made through the CN chemilumi-
nescence from the C2 qNO ™ COqCN reaction
4
4
Ž
.
2
Ž2
that at 120.00 nm 3s P l 2p S0 and for N D
3
2
that at 149.33 nm 3s P l 2p D0 were em-
3
Ž
.
3
Ž3
.
Ž
0
ployed. O P was monitored at 130.35 nm 3s S
3
l 2p4 P .
.
2.2. Source of C atoms
C atoms were produced in the reactant injector
nozzle from the CBr4 q4K ™ Cq4KBr or CCl4 q
4K ™ Cq4KCl overall reactions resulting from the
successive elementary steps:
CBr4 qK ™ CBr3 qKBr ,
Dr H2o98 sy1.50"0.30 eV ,
CBr3 qK ™ CBr2 qKBr ,
CBr2 qK ™ CBrqKBr ,
3a
3b
Ž
Ž
.
.
Dr H2o98 3b qD H o 3c sy1.15"0.3 eV ,
Ž
.
298 Ž
.
r
3c
3d
4a
4b
Ž
Ž
Ž
Ž
.
.
.
.
CBrqK ™ CqKBr ,
Dr H2o98 sy1.15"0.3 eV ,
CCl4 qK ™ CCl3 qKCl ,
Dr H2o98 sy1.32"0.10 eV ,
CCl3 qK ™ CCl2 qKCl ,
Dr H2o98 sy1.0"0.10 eV ,
CCl2 qK ™ CClqKCl ,
Dr H2o98 sy1.21"0.10 eV ,
w
x
13,14 . In the conditions used for the C pseudo-
first-order decay determination, C2 was found at
concentrations three orders of magnitude lower than
4c
4d
Ž
Ž
.
.
CClqK ™ CqKCl ,
Dr H2o98 sy0.30"0.10 eV .
Whatever the microfurnace temperature, an ex-
w x Žw x w x.
cess of potassium atoms K )20 CCl4 , CBr4
,
was used. CBr2 , CBr or CCl radicals and C atoms
were probed while varying the different parameters
such as the oven temperature, the halomethane flow
and the corresponding carrier gas flows. CBr2 radi-
cal evolution was followed by the BrU2 chemilumi-
nescence at 292 nm given by the CBr2 qO ™ CO
qBrU2 reaction. The evolution of CBr was followed
2
2
q
via the CN B S ™ X Sq chemiluminescence at
Ž
.
Ž
.
387 nm given by CBrqN ™ CN B, A, X qBr. The
CCl radical was followed by the same chemilumi-
nescent reaction or by LIF using the A 2D § X 2 P
transition near 278 nm. C atoms were detected by
their VUV fluorescence. All these chemiluminescent
Fig. 1. Evolutions of CBr2 , CBr and C with the distance between
Ž
the glass tube injecting CBr4 in K vapour and the nozzle see Fig.
x.
w
1 in Ref. 12 .

(
)
10
A. Bergeat et al.rChemical Physics Letters 308 1999 7–12
Ž
.
out further purification. CBr4 99% and CCl4
Ž
.
99.5% were also used without further purification.
3. Overall rate constants of the CHNO and CH
O2 reactions
The pseudo-first-order decays of atomic carbon
fluorescence were monitored at different concentra-
tions of NO and O2. Owing to the time needed to
achieve the mixing of reactants, the decay became
exponential at 3 cm from the nozzle through which
Ž
.
escaped C atoms Fig. 2 . The pseudo-first-order rate
constants were corrected for radial and axial diffu-
sion using Keyser’s formula:
Fig. 2. Examples of the C fluorescence decay with NO added in
excess.
a2
D
Õ2
that of C so that its influence on C decay measure-
ments could be neglected. In the conditions used for
the determination of the CqNO product branching,
the halomethane concentration was increased so that
the C2 concentration was only two orders of magni-
tude lower than that of C. It was still of no influence
in the estimation of the CqNO product branching
from the N and O atom probing, not only because
the C2 concentration remained much lower than that
of C but also because the C2 qNO reaction was
found to produce mainly COqCN. Another diffi-
culty was the overlapping of C fluorescence lines by
kcor skobs 1q
q
kobs ,
ž 48D
/
where kobs is the observed first-order rate constant, a
the radius of the reactor, Õ the average flow velocity,
D the diffusion coefficient of C atoms, equal to
860rP cm2 sy1, where P is the total pressure in
w x
Torr 3 .
Ž
For CqNO, an overall rate constant of 5.4"
cm3 moleculey1 sy1 was found Fig.
y11
.
Ž
0.8 =10
.
y11
3 . For CqO2 , it was 2.5"0.4 =10
cm3
Ž
.
moleculey1 sy1 Fig. 4 .
Ž
.
With the reactor used in the present study, Dorthe
1
1
q
Ž
.
the CO A Pu ™ X S
chemiluminescence. For
w x
et al. 3 had previously found for CqNO a value of
1
Ž
.
the CO A Pu state, lying at 8.02 eV above
y11
Ž
.
2.7"0.2 =10
cm3 moleculey1 sy1 and for
1
q
Ž
.
CO X S , only the reactions CBrqO ™ COq
Br, exoergic by 8.2 eV, and C2 qO2 ™ COqCO,
exoergic by 10.8 eV, could contribute to the ob-
served chemiluminescence, in addition to the CqO
q M ™ CO q M reaction. To minimise these
chemiluminescent reactions, we used low CBr4 or
CCl4 concentrations with the maximum distance be-
tween the end of the glass tube and the nozzle
aperture. Under these conditions, the C partial pres-
sure was below 0.005 mTorr with a negligible amount
CqO2 a value of 1.6"0.2 =10
cm3 mole-
y11
Ž
.
culey1 sy1. Atomic carbon decay was followed then
Ž
.
of CBr CCl and C2 radicals. In the case of the
CqO2 reaction, we could not completely eliminate
the CO chemiluminescence from C2 qO2 which
remained one order of magnitude smaller than the C
fluorescence. Each scan was thus recorded with and
without the lamp to subtract the CO chemilumines-
Ž
.
cence from the C fluorescence. NO 99.90% and O2
Fig. 3. Plot of the pseudo-first-order rate constant of the CqNO
reaction against NO concentration.
Ž
.
99.90% were used directly from the cylinders with-

(
)
A. Bergeat et al.rChemical Physics Letters 308 1999 7–12
11
3
Ž
.
via the CS a P chemiluminescence from CqOCS
3
Ž
.
™ CS a P qCO, observed when OCS was added
to NO. The discrepancy between the two sets of
3
Ž
.
experiments is ascribed to the fact that CS a P
being long-lived, has a chemiluminescence decay
which is slower than that of atomic carbon. Our new
determinations are consistent with the values re-
y11
w x Ž
.
ported by Husain and Young 1 : 4.8"0.8 =10
cm3 moleculey1 sy1 for CqNO and 2.6"0.3 =
Ž
.
10y11 cm3 moleculey1 sy1 for CqO2.
4. Product channels of CHNO
Ž3
.
Ž4
.
Ž2
.
Ž3
.
Fig. 5. Relative evolutions of C P , N S , N D and O P .
Ž4 Ž3
.
For the CqNO reaction, the N S , N D , O P
.
Ž2
.
atomic products were also probed. However, the
Ž2
.
tained from C2Cl4. In the latter case, we did not
observe N atom production from C2 qNO ™ C2O
qN and the C production from C2 qNO ™ NCO
qC was at least 50 times smaller than that of the C
directly obtained from CBr4, thus suggesting that
CNqCO is the main product channel of the C2 q
NO reaction. It can be concluded that C2 did not
actually interfere with our atomic product measure-
ments of the CqNO reaction. The CCl and CBr
radicals could also react with NO with a rate con-
situation was rather complex because N D could
Ž4
.
be converted by inelastic collisions into N S and
removed by reactive collisions with NO to give
Ž4
.
OqN2 , as occurred for N S for which the rate
w
x
constant at 300 K 15 has been determined as
3.65=10y11 cm3 moleculey1 sy1. Furthermore, the
Ž2
.
N D fluorescence line was partially overlapped by
the CO chemiluminescence from CqOqM while
our lamprdetection system gave smaller signals for
O and N atoms than for C atoms. To enhance the
signal from N and O atoms, higher concentrations of
CCl4 or CBr4 were used. As a consequence, the C2
concentration increased and CCl or CBr radicals
were not completely removed. Using the CN chemi-
luminescence given by C2 qNO, we checked that
the concentration of C2 was two orders of magnitude
lower than that of C, by comparison with C2 ob-
1
stant similar to that of C atoms and give O atoms
but not N atoms. As their concentrations are at least
10 times lower than the C concentration, they could
only slightly interfere with the O atoms from Cq
NO. Typical atom kinetics are given in Fig. 5, the
intensities are normalised on the initial C atomic
concentration. If N is removed by NO to yield
N2 qO, the removal of O by OqNOqM ™ NO2
qM is negligible. The observed O decay is thus due
to the wall removal. From the evolution of the
stationary concentrations of N and O atoms, we
Ž2
estimate their nascent branching ratio to N D q
.
Ž4
.
Ž3
.
N S s60"20% and O P s40"20%, in good
w x
agreement with the results of Dean et al. 5 in the
2570–3790 K range and with theoretical calculations
10 . There is no doubt that N D , and O P are
w
x
Ž2
.
Ž3
.
w x
primary products, as predicted by Persson et al. 9 .
Ž4
.
The direct production of N S remains questionable.
1
y11
cmy3 moleculey1 sy1 and
Ž
.
k CClqNO,298K s3.7=10
y11
cmy3 moleculey1 sy1, see
Ž
.
Fig. 4. Plot of the pseudo-first-order rate constant of the CqO2
reaction against O2 concentration.
k CBrqNO,298K s2.16=10
Refs. 16,17 , respectively.
w
x

(
)
12
A. Bergeat et al.rChemical Physics Letters 308 1999 7–12
Ž2
The nascent electronic branching between N D and
N S is difficult to estimate since we could not
.
w x
2
K.H. Becker, K.J. Brockmann, P. Wiesen, J. Chem. Soc.,
Ž4
.
Ž
.
Faraday Trans. 2 84 1988 455.
w x
3
G. Dorthe, Ph. Caubet, Th. Vias, B. Barrere, J. Marchais, J.
`
evaluate the effective contribution of the relaxation
Ž
.
Phys. Chem. 95 1991 5109.
Ž2
.
Ž4
.
from N D to N S to the observed relative station-
ary concentrations of N atoms. It appears, however,
that the nascent electronic branching must be largely
w x
4
D. Lindackers, M. Burmeister, P. Roth, 23rd Int. Symp. on
Combustion, The Combustion Institute, Pittsburgh, PA, 1990,
p. 251.
Ž2
.
w x
5
A.J. Dean, R.K. Hanson, C.T. Bowman, J. Phys. Chem. 95
in favour of N D as suggested by theoreticians
who discarded the pathway leading to N S 9 .
Ž
.
1991 3180.
Ž4 . w x
w x
6
D. Chastaing, P.L. James, I.R. Sims, I.W.M. Smith, Faraday
Discuss. No. 109, Chemistry and Physics of Molecules and
Grains in Space, Nottingham, Apr. 15–17, 1998, p. 47.
w x
7
M. Costes, C. Naulin, C.R. Acad. Sci., Paris, Ser. IIc 1
´
5. Conclusions
Ž
.
1998 771.
w x
8
P. Halvick, J.C. Rayez, E.M. Evleth, J. Chem. Phys. 81
The new CqNO and CqO2 rate constant val-
ues obtained in our fast-flow reactor from the direct
detection of C atoms are found closer to Husain and
Young’s values than those previously reported from
experiments where C atoms were not directly probed.
The product branching ratio of CqNO is found to
be ;60"20% for COqN and thus ;40"20%
for CNqO. For the CqNO ™ COqN channel,
Ž
.
1984 728.
w x
9
B.J. Persson, B.O. Roos, M. Simonson, Chem. Phys. Lett.
Ž
.
234 1995 382.
w
w
w
w
x
10 M. Simonson, N. Markovic, S. Nordholm, B.J. Persson,
Ž
.
Chem. Phys. 200 1995 141.
x
Ž
.
11 A. Beghin, T. Stoecklin, J.C. Rayez, Chem. Phys. 195 1995
259.
x
12 A. Bergeat, T. Calvo, J.C. Loison, G. Dorthe, J. Phys. Chem.
Ž
.
A 102 1998 8124.
Ž2
.
Ž4
.
the electronic branching between N D and N S
x
Ž
.
13 H. Reisler, M. Mangir, C. Wittig, J. Chem. Phys. 71 1979
Ž2
.
2109.
appears largely in favour of N D .
w
w
x
Ž
.
14 H. Krause, J. Chem. Phys. 70 1979 3871.
x
Ž
.
15 P.O. Wenberg, J.G. Anderson, J. Geophys. Res. 99 1994
18839.
References
w
x
16 J.J. Tiee, F.B. Wampler, W.W. Rice Jr., Chem. Phys. Lett.
Ž
.
73 1980 519.
w x
1
w
x
D. Husain, A.N. Young, J. Chem. Soc., Faraday Trans. 2 71
17 R.S. McDaniel, R. Dickson, F.C. James, O.P. Strausz, T.N.
Ž
.
Ž
.
1975 525.
Bell, Chem. Phys. Lett. 43 1976 130.