Reaction of methylidyne CH(X2π) radical with CH4 and H2S: Overall rate constant and absolute atomic hydrogen production

Article abstract of DOI:10.1016/S0301-0104(02)00443-3

The CH + CH₄ and H₂S reactions were studied, at room temperature, in a low-pressure fast-flow reactor. CH (X²π, v = 0) radicals were obtained from the reaction of CHBr₃ with potassium atoms. The overall rate constants were found at 330 K to be (0.76 ± 0.20) x 10-¹⁰ and (2.8 ± 0.8) x l0-¹⁰ cm³ molecule^-1 s^-1, respectively. The absolute atomic hydrogen productions were determined by resonance fluorescence in the vacuum ultraviolet: H production from the CH + CH₄ reaction is 100% and from the CH + H₂S reaction is 99_-4⁺¹%, the H production from the CH + H₂ reaction being the reference. Ab initio studies of the different stationary points relevant to the CH + CH₄ reaction have been performed at the CCSD(T)/cc-pVTZ level and comparison is made with experimental results. The experimental results for the CH + H₂S reaction is compared with those of a recent theoretical study [Chem. Phys. 242 (1999) 1].

Full text of DOI:10.1016/S0301-0104(02)00443-3

Chemical Physics 279 (2002) 87–99
www.elsevier.com/locate/chemphys
Reaction of methylidyne CHðX2PÞ radical with CH4 and H2S:
overall rate constant and absolute atomic
hydrogen production
Paul Fleurat-Lessard, Jean-Claude Rayez, Astrid Bergeat,
Jean-Christophe Loison*
Laboratoire de Physico-Chimie Moleꢀculaire, CNRS UMR 5803, Universiteꢀ Bordeaux I, F-33405 Talence Cedex, France
Received 12 December 2001; in final form 14 March 2002
Abstract
The CH þ CH4 and H2S reactions were studied, at room temperature, in a low-pressure fast-flow reactor.
CHðX2P; v ¼ 0Þ radicals were obtained from the reaction of CHBr3 with potassium atoms. The overall rate constants
were found at 330 K to be ð0:76 Æ 0:20Þ Â 10À10 and ð2:8 Æ 0:8Þ Â 10À10 cm3 moleculeÀ1
s
À1, respectively. The absolute
atomic hydrogen productions were determined by resonance fluorescence in the vacuum ultraviolet: H production from
the CH þ CH4 reaction is 100% and from the CH þ H2S reaction is 99þÀ41%, the H production from the CH þ H2 re-
action being the reference. Ab initio studies of the different stationary points relevant to the CH þ CH4 reaction have
been performed at the CCSD(T)/cc-pVTZ level and comparison is made with experimental results. The experimental
results for the CH þ H2S reaction is compared with those of a recent theoretical study [Chem. Phys. 242 (1999)
1]. Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction
the C atom, allowing addition on p-bonds and
insertion in r-bonds with no barrier. For example,
the CH þ CH4 reaction is important for the Titan
atmosphere chemistry [2], and CH reactions are
very important for the NOx production during
combustion process [3]. The CH þ H2S reaction is
a typical H2X-type reaction and is thus a proto-
type of the CH þ H2O reaction. This reaction is
also an important reaction for combustion of oil
and coals, and may also be concerning by sulfur
interstellar chemistry. Despite their importance,
there was, to our knowledge, no experimental
study of the product branching ratio for these two
reactions of CH with CH4 or H2S.
The CH radical is one of the most reactive
radical, and it is well known as an important in-
termediate in various reaction systems in fields
from planetary atmospheres to hydrocarbon
combustion [1]. This unusually high reactivity is
due to the presence of one singly occupied and one
vacant non-bonding molecular orbital localized on
* Corresponding author. Fax: +33-5-57-96-2521.
E-mail address: jc@lpcm.u-bordeaux.fr (J.-C. Loison).
0301-0104/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved.
PII: S0301-0104(02)00443-3

88
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
The possible reaction channels are ðDrH0 in
formed new kinetics experiments using a selective
kJ molÀ1Þ [4–10]:
source of CH radicals (from the reaction of CHBr3
with potassium atoms), in a low-pressure fast-flow
reactor at room temperature. The overall rate
constants were determined by following the decay
of the CH radicals laser induced fluorescence sig-
nal, the methane or the hydrogen sulfide being
introduced in excess, the diffusion corrections
having been validated in a previous study. Abso-
lute product branching ratios were estimated over
the channels yielding H atoms by comparison with
the CH þ H2 ! CH2 þ H reaction [14,15], H at-
oms being probed by resonance fluorescence in the
vacuum ultraviolet. New ab initio studies were
thus performed in order to assess more precisely
the role of the CH þ CH4 ! C2H3 þ H2 reaction
channel. The use of the CH þ CH4 reaction for
calibration of hydrogen atoms production yield in
CH reactions is discussed.
CHðX2PÞ þ CH4ðX1AgÞ
0
2
! C2H5ð A Þ
À 391:9
À 246:6
2
! Hð SÞ þ C2H4ðX1A1gÞ
! H2ðX1Rþg Þ þ C2H3ðX2A0Þ À 219:8
! CH2ðX3B2Þ þ CH3ðX2A020Þ þ 11:3
CHðX2PÞ þ H2SðX1A1Þ
! HCSðX2A0Þ þ H2ðX1RþÞ À 277:8
2
! H2CSðX1A1Þ þ Hð SÞ
À 255:2
À 152:3
00
! CH3ðX2A2Þ þ Sð PÞ
3
! CSHðX2A0Þ þ H2ðX1RþÞ À 113:0
! CH2ðX3B1Þ þ SHðX2PÞ
À 41:8
Many experiments have been carried out to deter-
mine the overall rate constant of the CH þ CH4
reaction with various techniques [11]. At 300 K,
there is a general agreement wiÀth1 a rate constant of
s
À1, and a signifi-
2. Theoritical and experimental methodology
about 9 Â 10À11 cm3 molecule
cant temperature dependence over the range 13–
770 K [12,13]. In contrast with the fact that the
overall rate constant has been accurately deter-
mined over a wide range of temperature, few is
available about the product branching ratios, al-
though they are needed to model the combustion or
interstellar chemistry processes. It has been com-
monly accepted that the CH þ CH4 reaction could
only lead to H þ C2H4 products at low pressure,
the stabilization of the C2H5 ethyl radical being
estimated negligible below 105 Torr [13], and the
slightly endothermic channel CH2 þ CH3 being
negligible below 2000 K [13]. However no experi-
mental studies have been performed to confirm it,
and the various ab initio calculations do not totally
rule out the possibility of the C2H3 þ H2 channel.
Sato et al. [10] have recently determined the
overall rate constant of the CH þ H2S reaction at
room temperature associated with ab initio and
RRKM studies on the branching ratios. The rate
constant was found to be ð3:2 Æ 0:8Þ Â 10À10 cm3
moleculeÀ1 sÀ1 and the RRKM calculations sug-
gest that the main product channel is H2CS þ H.
In order to have a better understanding of the
mechanism of the title reactions, we have per-
2.1. Ab initio calculations
Stationary points along the CHðX2PÞ þ CH4
ðX1AgÞ reaction path have first been found using
the coupled cluster method including single and
double excitations and a perturbational estimate of
triple excitations, CCSD(T) [15], with the correla-
tion consistent polarized valence double zeta (cc-
pVDZ) basis set of Dunning [17]. The harmonic
vibrational wavenumbers have been computed at
the same level of theory in order to characterize the
stationary points as minima or saddle points and to
obtain zero-point vibrational energy (ZPVE).
Structures and total energies were then refined at
the CCSD(T)/cc-pVTZ level [16,17], using the re-
stricted open shell formalism. ZPVE correction to
energies is used without any scaling factor. Unless
otherwise noted, the discussed energetics are those
obtained at the CCSD(T)/cc-pVTZ with CCSD(T)/
cc-pVDZ ZPVE corrections. For CCSD(T) calcu-
lations the core electrons have been frozen. All
CCSD(T)/cc-pVDZ calculations have been carried
out with the Gaussian 98 program [18], and all
CCSD(T)/cc-pVTZ calculations have been carried
out with the Molpro 2000.1 program [19–21].

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
89
much lower than the CHðX2PÞ reactivity toward
CH4, H2 and H2O (and then H2S) [26] and thus
will not contribute to H production. Additionally
LIF detection of CHðX2PÞ radical, using a
ND:YAG laser (Quantel YG 581C) pumped dye
laser and exciting the CHðA2D X2PÞ near 431
nm, allow us to determine that the CHðX2PÞ is
fully vibrationally and rotationally relaxed.
The CHBr3, diluted in He, is mixed inside the
injector with potassium vapor, coming from a
micro furnace [22]. The injector conditions are the
following: P ¼ 1:4 Torr, ½KŠ ¼ 1 mTorr, ½CHBr3Š ¼
0:05 mTorr, which give in the fast-flow reactor:
P ¼ 1:4 Torr, ½KŠ < 0:1 mTorr, ½CHBr3; CHBr2;
CHBrŠ ꢀ 0:001 mTorr and ½CHŠ ꢁ 0:003 mTorr.
As a large excess of potassium is introduced in the
injector compared to the CHBr3 concentration,
the precursors (CHBr3, CHBr2, CHBr) concen-
trations in the fast-flow reactor are very small. The
difference between CH concentration in the fast-
flow reactor and CHBr3 concentration introduced
is due to wall reactions in the micro furnace
leading to 30% efficiency conversion, and dilution
of the injector flow in the reactor. As the K þ H2S
reaction presents a barrier in the gas phase, the
only possible reaction of K atoms leads to the
formation of an adduct, which is absolutely neg-
ligible in our low-pressure experimental condi-
tions.
2.2. Experimental measurements
The experimental setup has been described in
detail previously [22–24], and only a summary is
given here. The setup consists of a fast-flow re-
actor, i.e., a 36-mm inner tube with four optical
ports for detection. The CH radicals are pro-
duced in an ‘‘injector’’ which slides in the reactor.
At the end of the injector, the CH radicals are
mixed with the methane or hydrogen sulfide flow.
Then, the distance between the end of the injector
and the observation windows is directly propor-
tional to the reaction time. The distance (d)
between the windows detection and the injector
nozzle aperture could be varied over the range
0–100 mm with 0.5-mm precision. The pressure
was measured by
a capacitance manometer
(Barocel 0–10 Torr) and the flow rates were
adjusted by thermal mass flow meters. Before
each experiment, the vacuum and the leak-
plus-out gassing rate were checked with a Pirani
gauge (respectively, <0.05 and <10 mTorr
minÀ1).
The CH radicals were produced in the reactant
injector nozzle from the CHBr3 þ 3K ! CHþ
3KBr overall reaction which can be separated into
the elementary steps ðDrH0 in kJ molÀ1 [25]):
CHBr3 þ K ! CHBr2 þ KBr DrH2098 ¼ À96 Æ 8
ð1Þ
CH4 and H2S were used directly from the cyl-
inders with a purity >99.995% for CH4 and
>97.5% for H2S (the main impurities were CS2 and
OCS below 2000 ppv). CHBr3 (99%) was used
without any further purification.
CHBr2 þ K ! CHBr þ KBr DrH2098 ¼ À75 Æ 8
ð2Þ
CHBr þ K ! CH þ KBr DrH2098 ¼ À37 Æ 16
ð3Þ
Hydrogen atoms were detected by their reso-
nance fluorescence using the 2p1 2P0 $ 1s1 2S
transition at 121.6 nm. Atom excitation was
achieved with the microwave discharge lamp pre-
viously used to probe the atomic products of the
CH + NO reaction [27]. A pure helium flow (purity
of 99.9999%) was used for the branching ratio
determination. The general procedure for atomic
detection has also been detailed previously [23,27].
We would only mention that the conditions of the
presently reported experiments ensure the linear
dependence of the atomic fluorescence versus both
the lamp emission intensity and the H atoms
concentration.
As all the K þ CHBrx ! KH þ CBrx ðx P 0Þ
reactions are endoergic, this source can only pro-
duce CH radicals. As the sum of the exothermici-
ties of the three abstractions is 208 kJ molÀ1, the
production of CHða4RÀÞ radicals, which is
69:9 kJ molÀ1 above the ground state, is possible.
However, as the last Br abstraction is only
37 kJ mÀolÀ1 above the ground state, production of
CHða4R Þ requires high concentration of meta-
stable electronic excited CHBr in the oven, which
is unlikely. Moreover, the CHða4RÀÞ reactivity is

90
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
3. Results
[4–9,30,36], while the energy diagram along the
reaction path is shown in Fig. 2.
3.1. Potential energy surface
3.1.1. Geometries
The CCSD(T)/cc-pVTZ optimized structures of
the different minima and transition structures lying
on the doublet surface are displayed in Fig. 1
together with the available experimental values
3.1.1.1. Reactants and products. Experimentally
determined geometries are available for each re-
actant and product. For the ethyl radical, only the
C–C bond distance is known from recent laser
ꢁ
Fig. 1. Optimized geometries. Distances in A and angles in degrees. Experimental values are indicated in parenthesis. For TS3, the
parameters are those obtained at the CCSD(T)/cc-pVDZ level.

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
91
spectroscopy experiments [30]. For all the geo-
metric parameters, agreement between the
CCSD(T)/cc-pVTZ optimized geometries (Fig. 1)
ꢁ
and experimental values are within 0.02 A for
bond lengths and 1.0° for bond angles.
Moreover, the agreement with earlier theoreti-
cal results at similar levels of calculation is very
good: on the average, the bond distances differ by
ꢁ
less than 0.005 A and the bond angles by less than
0.1° [28,29,31–35,37].
3.1.1.2. Transition states. The transition state TS1
is depicted in Fig. 1 and Table 1. It corresponds to
an hydrogen loss from the ethyl radical. It has Cs
symmetry with the mirror plane containing the two
carbon atoms and the departing hydrogen atom.
The structure of this transition state corresponds to
a late barrier: the geometry of the C2H4 fragment
nearly coincides with that of the free ethylene
Fig. 2. Schematic picture of the potential energy surface (PES)
correlating with methylidyne CHðX2PÞ and methane CH4
ðX1AgÞ ground states.
ꢁ
molecule, and the leaving H atom is about 2 A
away. This late character is in accord with the
Table 1
Main structural parameters for TS1
UMP2
6-31G(d,p)a
QCISD
6-311G(d,p)b
QCISD(T)
6-311G(d,p)c
CCSD(T)
VTZd
d(CC)
dðCHaÞ
1.335
1.865
1.079
1.079
107.2
121.5
121.5
Æ96:7
Æ88:3
1.353
1.976
1.087
1.087
106.7
121.1
121.4
Æ95:9
Æ88:7
1.357
1.967
1.088
1.088
106.7
1.351
1.950
1.083
1.083
106.6
121.1
121.3
Æ95:5
Æ88:5
dðCHbÞ
dðCHcÞ
aðCCHaÞ
aðCCHbÞ
aðCCHcÞ
HðHaCCHbÞ
HðHaCCHcÞ
e
–
–
–
–
Barrier height (relative to C2H5)
180.9f
167.9g
155.7h
156.2i
À1
ꢁ
Distances in A and angles in degrees. Barrier height in kJ mol
.
a From [33].
b From [28,31,32].
c From [34].
d This work. Absolute electronic energy: )78.934182 a.u.
e A – indicates that this value is not given in the original article.
f ZPVE: UMP2/6-31G(d,p).
g ZPVE: QCISD/6-311G(d,p).
h ZPVE: MP2/6-31G(d,p).
i ZPVE: CCSD(T)/cc-pVDZ.

92
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
reaction endothermicity ð144 kJ molÀ1 at our level
of calculations, see below). Agreement with previ-
ous theoretical studies is good [8,28,32,33].
TS2 is the transition state for the abstraction
reaction C2H4 þ H ! C2H3 þ H2. It possesses a
planar geometry and its main structural parame-
ters are shown in Table 2. This structure can be
described as a H2 molecule departing from a C2H3
radical; at our level of calculation, the distance
between the carbon atom and the first hydrogen
TS3 is the transition state for 1,1-elimination of
a H2 molecule from the ethyl radical: C2H5 !
C2H3 þ H2.
It is worth noting that Morokuma and co-
workers [28] have unsuccessfully searched this
transition state within the Cs symmetry. The
transition state structures they obtained corre-
spond to an excited electronic state and further
optimization of these structures without any
symmetry constraint leads to either TS1 or TS2.
This is in agreement with the fact that, as predicted
by a Woodward–Hoffman correlation diagram,
the ground state of the ethyl radical could not lead
to the ground state of C2H3 þ H2 under the Cs
symmetry constraint (see Appendix A).
ꢁ
atom in this transition state is 1.45 A, longer by
ꢁ
0.35 A than a typical C–H bond length. Mean-
while, the H–H bond in H2 is elongated only by
ꢁ
ꢁ
0.13 A, as compared with the free H2 (0.741 A)
[38]. This transition state has a late character; in
agreement with the endothermicity of this reaction
(26 kJ molÀ1, see below). Agreement with previ-
ous theoretical studies is good [8,28].
Therefore, we have undergone the search for a
direct H2-elimination without any symmetry con-
straint. However, as the TS1 and TS2 structures
Table 2
Main structural parameters for TS2
MP2
6-31G(d,p)a
QCISD
6-311G(d,p)b
CCSD(T)
VTZc
d(CC)
dðCH1Þ
dðH1H2Þ
a12
1.290
1.326
1.427
0.872
175.9
1.324
1.446
0.865
174.4
1.444
0.848
175.6
dðCHaÞ
1.084
1.082
1.080
121.8
121.5
131.5
115.7
1.091
1.088
1.086
1.087
1.084
1.082
121.3
121.6
131.0
115.7
dðCHbÞ
dðCHcÞ
d
aðCCHaÞ
aðCCHbÞ
aðCCHcÞ
aðCCH1Þ
–
–
130.8
115.6
Barrier height (relative to C2H4 þ H)
65.2e
73.7f
66.6g
Distances in A and angles in degrees. Barrier height in kJ molÀ1. (e) From [34].
a From [8].
ꢁ
b From [28].
c This work. Absolute electronic energy: )78.91084 a.u.
d A – indicates that this value is not given in the original article.
e ZPVE: MP2/6-31G(d,p).
f ZPVE: MP2/6-31G(d,p) scaled by 0.95.
g ZPVE: CCSD(T)/cc-pVDZ.

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
93
Table 3
Experimental enthalpies of formation at 0 K, in kJ molÀ1
are much lower in energy than the expected tran-
sition state, a direct unconstraint geometry opti-
mization starting from an approximate structure
will eventually lead to either TS1 or TS2. We have
thus started from a structure derived from that of
the 1,1-H2 elimination from ethane [42]. In order
to prevent the optimization to lead to the more
stable structures TS1 or TS2, we have first frozen
the departing H2 molecule, while optimizing only
the C2H3 fragment. Starting from the resulting
structure, we have then done a full optimization
which led to TS3. However, as this structure is
located much higher in energy than TS1 or TS2, its
geometry was optimized at the CCSD(T)/cc-pVDZ
level only. We have then performed a single point
calculation using this geometry at the CCSD(T)/
cc-pVTZ level. The main structural parameters are
shown in Fig. 1.
Experimental Df H0 at 0 K
ðkJ molÀ1
Þ
2
CHð PÞ
590.8a
CH4
C2H5ð A Þ
)67.0a
0
131:9 Æ 2:1b;c
2
C2H4
C2H3ð A Þ
61.1a
0
304:0 Æ 3:3b;d
386.0a
2
3
CH2ð B1Þ
00
2
CH3ð A2Þ
149.0a
216.0a
H
a From [4].
b From [5].
c From [9].
d From [6–8].
Table 4
Experimental and calculated enthalpies of formation at 0 K, in
kJ molÀ1
This transition state is better described as a H2
molecule departing from the C2H5 radical. As in
the ethyl radical, the remaining C2H3 fragment is
non-planar, which corresponds to a distorted
structure compared to the equilibrium geometry of
the vinyl radical. This transition state corresponds
to an asymmetric process, one C–H bond ðC1H5Þ
being much longer than the other one ðC1H4Þ:
Species
Experimental
(0 K)
This work
(0 K)
2
CHð PÞ þ CH4
0
0a
0
2
C2H5ð A Þ
TS1
À391:9 Æ 2:1b
)383.1
)226.7
)238.6
)172.1
)213.1
+8.3
)232.8c
C2H4 þ H
TS2
)246.6b
)176.3d
C2H3 þ H2
À219:8 Æ 3:3b
ꢁ
ꢁ
1.665 A compared to 1.434 A. As in the TS2
structure, these distances are much longer than a
typical C–H bond. The H–H bond in H2 is equal
TS3
CH2ð B1Þþ
3
11.3b
14.6
00
2
CH3ð A2Þ
ꢁ
a Absolute electronic energy: )78.84834 a.u.
to 0.849 A, which is similar to that found in the
ꢁ
transition state TS2: 0.865 A.
b From Table 6.
c From [41].
d From [28].
3.1.2. Relative stability and thermochemistry
The energy diagram along the reaction path is
shown in Fig. 2. The experimental enthalpies of
formation at 0 K, Df H00, from [4–9] are reported in
Table 3. Using these enthalpies, the relative sta-
bility of the main species can be estimated. In
Table 4, the experimental heats of reaction at 0 K
ðDrH00Þ are compared to those obtained at the
CCSD(T)/cc-pVTZ level. At this level, computed
values for each DrH00 are within 9 kJ molÀ1 of the
2
CHð PÞ molecule. However, the value of the
T1(CCSD) diagnostic, which gives an indication of
the validity of the correlation method based on a
single reference [39,40], has been computed to be
always less than recommended value of 0.02 [40],
even in the case of the open shell molecules.
3.1.3. Reaction mechanism
literature values, the RMS error being 5 kJ molÀ1
.
Fig. 2 displays a schematic picture of the po-
tential energy surface (PES) correlating with
methylidyne CHðX2PÞ and methane CH4ðX1A1Þ
ground states. On the entrance channel, the ethyl
radical C2H5 is formed without any barrier. This is
in agreement with the large rate constant at low
It is worth noting that the error on relative ener-
gies of C2H4 þ H and C2H3 þ H2 towards C2H5 is
less than 2 kJ molÀ1. The larger error on the rel-
ative energy of C2H5 towards CH þ CH4 could
come from the multireference character of the

94
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
temperature and shows the ability of CH radical to
insert into a C–H bond. Starting from the ethyl
radical, several channels are conceivable. First, the
been pointed out by Knyasev et al. [8] who un-
derlined that a better agreement with their experi-
mental data was found by lowering these values by
5 to 68:6 kJ molÀ1. It is worth noting that, as the
C2H4 þ H channel is open, the hydrogen atom will
go away and, although the transition state TS2 is
located 172:0 kJ molÀ1 below the entrance chan-
nel, this second step will be inaccessible.
3
C–C bond can dissociate to form the CH2ð B1Þ and
00
2
CH3ð A2Þ fragments. However, as seen in the pre-
vious part, this channel is endothermic. Second,
loss of an hydrogen atom could lead to CH2CH2 þ
H or CH3CH þ H. Between these two channels,
only the first one is exothermic, the second one
being endothermic by at least 500 kJ molÀ10 at the
Starting from the ethyl radical, the other
channel is the loss of a H2 molecule leading to
C2H3 þ H2. Two distinct paths would also be
possible; 1,1-H2 loss where the two departing hy-
drogen atoms are from the same carbon atom and
1,2-H2 loss where the hydrogen atoms come from
different carbon atoms. The 1,1-H2 loss can lead to
HCCH2 þ H2 or H3CC þ H2 isomers. The
1
CCSD(T)/cc-pVDZ level; the singlet ð A Þ and
00
3
triplet ð A Þ CH3CH isomers lie, respectively, 750
and 880 kJ molÀ1 higher in energy than the ethyl-
ene form. As the reaction leading to CH3CH þ H is
largely endothermic (at least 540 kJ molÀ1), it is not
considered further. A transition state (TS1) is lo-
cated connecting the ethyl radical to an ethylene
molecule and a hydrogen atom. The energy barrier
is 156:2 kJ molÀ1; the relative energy of TS1 being
À226:7 kJ molÀ1 with respect to the entrance
channel. This channel is much lower in energy than
the dissociation of the C–C bond, and should have
a larger contribution. The value of the energy
barrier is in very good agreement with the experi-
mental values deduced from a Transition State
modeling of the dissociation. Berman and Lin [41]
estimated a barrier height of 159 kJ molÀ1, while
Hase et al. [31] found a barrier of 158 kJ molÀ1. In
a second step, the hydrogen atom can abstract a
second hydrogen atom from C2H4 to form the vinyl
radical C2H3 and a hydrogen molecule H2. The
barrier height for this abstraction reaction is equal
to 67 kJ molÀ1 at the CCSD(T)/cc-pVTZ level at 0
K. This value is in good agreement with the ex-
perimental value of 65 kJ molÀ1. This experimen-
tal value corresponds to the value of 63 kJ molÀ1
obtained by Knyasev et al. [8] at 298 K and ex-
trapolated to 0 K using the thermal energy calcu-
lated at the CCSD(T)/cc-pVDZ level. (Conversely,
the barrier height at 298 K equals 64 kJ molÀ1 at
the CCSD(T)/cc-pVTZ level, in good agreement
with the original value of Knyasev et al.) Our en-
ergy barrier is about 8 kJ molÀ1 lower than the
previous calculations of Morokuma and co-work-
ers [28]: 74 kJ molÀ1 at the CCSD(T)/cc-pVTZ//
QCISD(T)/6-311G(d,p) and QCISD(T)/6-311G
(d,p) levels, using the MP2/6-31G(d,p) ZPVE
scaled by 0.95. This overestimation has already
0
00
2
4
CH3Cð A Þ and CH3Cð A Þ species are located 200
and 328 kJ molÀ1 higher in energy than the
CH2CH ground state, which means that their rel-
ative energies towards CH þ CH4 are, respectively,
0
À1
00
À12 kJ molÀ1 ð A Þ and 114 kJ mol ð A Þ. Al-
2
4
0
2
though the reaction leading to the CH3Cð A Þ
isomer is exothermic, the loss of a H2 molecule
corresponds to an important electronic reorgani-
zation and should proceed through a transition
state high in energy. Therefore, none of these iso-
mers is considered further.
As the H2 molecule leaves the C2H3 moiety,
there is a formation of a pC–C bond, as in the
elimination of H2 from ethane to form C2H4 þ H2.
Therefore, the activation energy of this process
should be comparable to the barrier of the
C2H6 ! C2H4 þ H2 reaction, which equals 445:9
kJ molÀ1 [42]. Indeed, we found that the transition
state TS3 corresponding the 1,1-H2 loss is located
391:4 kJ molÀ1 above the C2H5 radical, that is to
say 8:3 kJ molÀ1 above the entrance channel. As
its relative energy towards the ethyl radical is
much higher than that of TS1 and TS2: 156.4 and
211:0 kJ molÀ1, respectively, this channel will not
play any role in this reaction. Previous studies for
H2 loss from C2H6 and C2H4 have shown that the
channels corresponding to a 1,2-H2 loss are always
much higher in energy than those corresponding to
a 1,1-H2 elimination [42,43], therefore these tran-
sition states have not been investigated further.
This indicates that all channels for a direct H2 loss
are located high in energy. As a consequence the

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
95
only observable channel in our experiment is the
dissociation of the ethyl radical to form a molecule
of ethylene and a hydrogen atom.
3.2. Overall rate constant
The pseudo-first-order decays of CH radical
fluorescence signal were monitored at different
concentrations of CH4 or H2S introduced in large
excess. To get rid of the mixing effects, only the last
stages of the decay (after 3 cm from the injector
exit) have been taken to determine the pseudo-
first-order rate constants. The measured rate con-
stants were then corrected from radial and axial
diffusions from Keyser’s formula [44], as done
previously, with good results, for the CH + NO
[23] and CH þ O2 [45] reactions.
Fig. 4. Plot of the pseudo-first-order rate constant of the
CH þ H2S reaction versus the H2S concentration. The gradient
of the fitted line yields to the second-order rate constant,
k ¼ ð2:8 Æ 0:8Þ Â 10À10 cm3 moleculeÀ1 sÀ1
.
The results of our experiments are displayed in
Figs. 3 and 4, where axial and radial corrected
pseudo-first-order rate constants are plotted versus
the CH4 and H2S concentrations, respectively. The
main source of errors in our measurements is the
important radial and axial diffusions corrections.
Moreover, the high wall removal rate constant
associated with these diffusions leads to the limited
conditions of the plug-flow approximation and the
errors quoted take in account these uncertainties.
The second-order rate constants are thus ð7:6 Æ
2:0Þ Â 10À11, ð2:8 Æ 0:8Þ Â 10À10 cm3 moleculeÀ1
respectively. The present results are in fair agree-
ment with previous measurements [10,13,46,47],
even if they are slightly lower.
3.3. Hydrogen atomic production
Relative hydrogen production of the CH þ CH4
and CH þ H2S were determined by the ratio to the
H production from the CH þ H2 reaction by reso-
nance fluorescence in the vacuum ultraviolet. As the
H atoms branching ratio is known for the CH þ H2
reaction [14], even if this ratio is very dependent on
temperature and pressure, the determination of the
absolute branching ratios for CH þ CH4 and
CH þ H2S reactions can thus be worked out.
s
À1, for the CH þ CH4 and CH þ H2S reactions,
To measure the relative H atoms production,
their fluorescence signals are recorded successively
for the three reactions, the CH4, H2, and H2S
concentrations being adjusted in order to have
equivalent global first-order rate constant. This is
possible because our CH production is constant
within a period of more than 1 h. This operation
was repeated several times, alternately for different
CH4, H2S, or H2 concentrations, under different
pressures and different CHBr3 concentrations.
As the CH þ H2 reaction has a much lower glo-
bal rate constant than the CH þ CH4 and
CH þ H2S reactions, we have performed the com-
parison of CH þ CH4 and CH þ H2 at 350 K by
heating the He buffer gas flow (at 1.4 Torr and 350 K
Fig. 3. Plot of the pseudo-first-order rate constant of the
CH þ CH4 reaction versus the CH4 concentration. The gradient
of the fitted line yields to the second-order rate constant,
k ¼ ð7:6 Æ 2:0Þ Â 10À11 cm3 moleculeÀ1 sÀ1
.

96
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
kðCH þ H2Þ ¼ 2:0 Â 10À12 [14,15] and kðCH þ
C2H þ CH4 reaction give only C2H2 and CH3
[52]).
CH4Þ ¼ 7:5 Â 10À11 cm3 moleculeÀ1
s
À1). An ex-
ample of the traces of H atoms concentrations,
deduced from the fluorescence intensities, versus
the distance, i.e., the reaction time, is shown in Fig.
5. In our experimental conditions, we cannot ne-
glect the CH + CH reaction contributions, neither
as a direct H atoms source nor as C2H source
which could also perturb the H measurement by
the C2H þ H2, C2H þ CH4 or C2H þ H2S reac-
tions. The experiments are performed with first-
order global rate constant value, for the
CH þ CH4, CH þ H2 and CH þ H2S reactions,
around 3000 sÀ1 which corresponds to 1 mTorr of
CH4, 40 mTorr of H2 or 0.3 mTorr of H2S. Then,
considering a thermalized C2H, the first-order
global rate constant is around 1200 sÀ1 for the
C2H þ H2 [48,49] reaction and around 100 sÀ1 for
the C2H þ CH4 [50] and the C2H þ H2S reactions
(the C2H þ H2S ! HSþ C2H2 or H + HCCSH
As the H atom and C2H radical production rate
by the CH + CH reaction is related to the C2Hꢂ
chemiluminescence signal ðCH þ CH ! C2Hꢂ þ
HÞ [24], integration of the C2Hꢂ chemilumines-
cence signal is then proportional to the H and C2H
productions. In Fig. 5 is shown the H concentra-
tion evolutions from the CH þ H2, CH þ CH4 and
CH + CH reactions. The estimated H contribution
from the CH + CH reaction when CH4 or H2 is
added is also represented. This contribution is
calculated by integrating the C2Hꢂ chemilumines-
cence signal convoluted with a wall loss rate con-
stant of 200 sÀ1, which is the H atoms wall
reaction rate in our reactor. This contribution is
then scaled with the H atomic fluorescence signal
when only CH radicals are present in the reactor,
using the C2Hꢂ chemiluminescence integration
ratio with and without co-reactants. For the C2H
radicals produced by the CH þ CH reaction, we
consider that all the C2H radicals react with H2 to
give C2H2 þ H. For thermalized C2H this is done
in the first three centimeters. However as C2H is
produced in highly vibrationally excited states [24],
its reactivity toward H2 is enhanced [53]. In Fig. 6
is shown the corrected H atoms concentrations of
the CH þ H2 and CH þ CH4 reactions, correc-
tions due to the CH þ CH ! C2H þ H and
C2H þ H2 ! C2H2 þ H contributions, and also
reaction
could
be
estimated
close
to
10À11 cm3 moleculeÀ1
s
À1) [51]. The only impor-
tant secondary source of H atoms is thus the
C2H þ H2 ! H þ C2H2 reaction, the two other
reactions having too small rate constants com-
pared with the CH reactions (additionally the
Fig. 5. H atoms fluorescence signal from ðꢃÞ: CH + CH (when
no co-reactant is added); (dashed line): H production from
CH + CH, estimated from C2Hꢂ chemiluminescence signal,
during the CH þ CH4 or CH þ H2 reaction (see text), ðjÞ:
CH þ CH4, ðꢂÞ: CH þ H2 and C2Hꢂ chemiluminescence signal
(+): when no co-reactants is added, ðÂÞ: with CH4 or H2 added,
in function of the distance d in the reactor.
Fig. 6. Corrected H atom concentration from CH þ H2 ðꢂÞ and
CH þ CH4 ðjÞ reactions, and the ratio of these two curves
ðꢃÞ, in function of the distance, d, in the reactor.

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
97
the ratio of these two curves which is quite con-
stant from the second centimeter to the end of the
curve. We must notice that the continuous curve
linking the experimental points is the profile of the
H atoms production deduced by integrating the
experimental decay of CH radicals with additional
wall loss of 200 sÀ1 and then by scaling it to the
H fluorescence signal, with a converted factor 0.84
less important for CH þ H2 reaction than for the
CH þ CH4 reaction.
if we took the Bronwnsword et al. data [15]), we
deduce from the experimental data that the H
branching ratio of the CH þ CH4 reaction is
comprise between 96% and ‘‘106’’%. To estimate
more precisely the H atoms branching ratio, we
should use the theoretical information. First, due
to the large exothermicity and the low-exit barrier
of the C2H4 þ H channel, the stabilization of C2H5
adduct could be neglected, except at very high
pressure [13]. Secondly, the endothermic channel
CH2 þ CH3 is negligible below 2000 K [13]. And
finally, our ab initio calculations show that the
CH þ CH4 ! C2H3 þ H2 channel, corresponding
to the direct reaction C2H5 ! C2H3 þ H2, is not
accessible. Then, we conclude that, at low tem-
perature, the H atom is the exclusive product of
the CH þ CH4 reaction.
To check the validity of this correction, we have
measured the evolution of the ratio of [H] pro-
duction of the CH þ H2 and CH þ CH4 reactions,
in function of the CH initial concentration, at a
distance in the reactor of four centimeters. The
results, presented in Fig. 7, show a good conver-
gence of the corrected ratio toward 0.84, and un-
derline that the main correction is due to the
C2H þ H2 contribution, which could arise 50%.
The good convergence of the corrected and raw
ratio curves makes us quite confident about the
way to estimate the ratio of H production.
Finally we obtain a ratio of H production be-
tween the CH þ H2 reaction and the CH þ CH4
reaction equal to 0:84 Æ 0:06. As the H branching
ratio of the CH þ H2 reaction has been estimated
between 81% and 89% at 350 K and 1.4 Torr (81%
if we took the Fulle and Hippler data [14] and 89%
To confirm the use of the CH þ CH4 reaction as
a calibration of product yield, the CH þ H2S has
been studied by comparison with the CH þ CH4
reaction. Actually, a recent RRKM study coupled
to ab initio calculations [10] estimates a branching
ratio for H atoms production close to 100%. As the
CH þ H2S reaction is very similar experimental
case than CH þ CH4 reaction (the C2H þ H2S re-
action contribution is negligible), the correction
between the CH þ H2S and the CH þ CH4 reac-
tion is only due to the CH þ CH ! H þ C2H re-
action. Moreover, as we use equal first-order rate
constant, and as the H atoms production are sim-
ilar for the two reactions, the uncertainties are very
small in this case due to equivalent corrections.
With a very similar procedure than for the com-
parison between CH þ CH4 and CH þ H2 reac-
tions, a ratio of H production between the
CH þ H2S reaction and the CH þ CH4 reaction is
obtained equal to 0:99 Æ 0:04. This result is in ex-
cellent agreement with the recent RRKM calcula-
tions [10] predicting a branching ratio for the
CH2S þ H channel equal to 99.6%. It should be
noted that this good agreement for this open
system (five exit channels) is a good proof of the
validity of RRKM calculation when intermediate
species have enough long lifetimes (in this case
around 2 ps for CH3S and CH2SH) for the
assumption of a rapid intramolecular transfer
leading to energy randomization among internal
degrees.
Fig. 7. H atom ratio (CH þ H2=CH þ CH4 production) as a
function of the initial CH concentration [in arbitrary unit], with
no correction ðdÞ, corrected for C2H þ H2 ðjÞ reaction and
corrected for C2H þ H2 and CH + CH reactions ðꢃÞ.

98
P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
Fig. 8. Paths for the C2H5 ! C2H3 þ H2 reaction under the Cs symmetry.
4. Conclusions
the departing H2 molecule is perpendicular to that
plane. In the first channel, the HbHc fragment is
perpendicular to the symmetry plane whereas it
belongs to it in the second channel.
The first channel leads to a non-planar structure
corresponding to an excited state of the C2H3
molecule. The Woodward–Hoffman correlation
diagram corresponding to the second channel is
depicted in Fig. 9. This diagram clearly shows that
the ground state of C2H5 correlates once again
with an excited state of the C2H3 molecule.
Despite the fact that we have used the CH þ H2
reaction to determine absolute H branching ratios
of other reactions, its branching ratio between CH3
and CH2 þ H channels is very sensitive to several
parameters [14], and another reaction is necessary
for calibration of H production for CH reactions.
From our study, the CH þ CH4 reaction is, without
any doubt, a much better calibration reaction. First,
the possible complications arising from the
CH2 þ CH3, C2H5 stabilization and C2H3 þ H2
channels could be ruled out for low-pressure and
low-temperature measurements. Secondly, the
products of the reaction, C2H4 and H, do not react
with CH4 and then secondary reactions are very
limited if an excess of CH4 and a selective source of
CH radicals are used. Then we could consider that
this reaction proceeds to yield ethene and H atoms
exclusively. Moreover, the overall rate constant
(and thus detailed rate constants) dependence on
temperature is much less critical for the CH þ CH4
reaction than for the CH þ H2 reaction.
A good illustration is our determination of the
H branching ratio of the CH þ H2S reaction, which
have been measured equal to 99þÀ41%, in excellent
agreement with recent RRKM calculations [10].
Appendix A
When looking for the transition state leading to
C2H3 þ H2 from the C2H5 molecule under the Cs
symmetry constraint, two different paths are con-
ceivable as indicated in Fig. 8. In both cases, the
C1C2Ha fragment lies in the symmetry plane and
Fig. 9. Woodward–Hoffman correlation diagram for the
C2H5 ! C2H3 þ H2 reaction under the Cs symmetry.

P. Fleurat-Lessard et al. / Chemical Physics 279 (2002) 87–99
99
[23] A. Bergeat, T. Calvo, N. Daugey, J.-C. Loison, G. Dorthe,
J. Phys. Chem. A 102 (1998) 8124.
References
[24] A. Bergeat, T. Calvo, G. Dorthe, J.-C. Loison, J. Phys.
Chem. A 103 (1999) 6360.
[1] W.A. Sanders, M.C. Lin, in: Z.B. Alfassi (Ed.), Chemical
Kinetic of Small Organic Radicals, vol. III, CRC, Boca
Raton, FL, 1986.
[25] M. Born, S. Ingemann, N.M. Nibbering, J. Am. Chem.
Soc. 116 (1994) 7210;
[2] R.P. Wayne, Chemistry of Atmospheres, second ed.,
Clarendon Press, Oxford, 1991.
E. Tschuikow-Roux, S. Paddison, Int. J. Chem. Kinet. 19
(1987) 15.
[3] J.A. Miller, C.T. Bowman, Prog. Energy Combust. Sci. 15
(1989) 287.
[26] Zhou, K.D. Bayes, J. Phys. Chem. 97 (1993) 1896.
[27] A. Bergeat, T. Calvo, G. Dorthe, J.-C. Loison, Chem.
Phys. Lett. 308 (1999) 7.
[4] M.W. Chase Jr., NIST-JANAF Thermochemical Tables,
fourth ed., J. Phys. Chem. Ref. Data, Monograph 9, 1998,
1.
[28] A.M. Mebel, K. Morokuma, M.C. Lin, J. Chem. Phys. 103
(1995) 3440.
[5] W. Tsang, in: J.A. Martinho Simoes, A. Greenberg, J.F.
Liebman (Eds.), Heats of Formation of Organic Free
Radicals by Kinetic Methods in Energetics of Organic Free
Radicals, Blackie Academic and Professional, London,
1996, p. 22.
[29] J.C. Rienstra-Kiracofe, W.D. Allen, H.F. Schaefer III, J.
Phys. Chem. A 104 (2000) 9823.
[30] S. Davis, D. Uy, D.J. Nesbitt, J. Chem. Phys. 112 (2000)
1823.
[31] W.L. Hase, H.B. Schlegel, V. Balbyshev, M. Page, J. Phys.
Chem. 100 (1996) 5354.
[6] J.S. Pilgrim, C.A. Taatjes, J. Phys. Chem. 101 (1997) 8741.
[7] E.W. Kaiser, T.J. Wallington, J. Phys. Chem. 100 (1996)
4111.
[32] B.S. Jursic, J. Chem. Soc. Perkin Trans. 2 (1997) 637.
[33] Z.-X. Wang, M.-B. Huang, R.-Z. Liu, Can. J. Chem. 75
(1997) 996.
[8] V.D. Knyasev, A. Bencsura, S.I. Stoliarov, I.R. Slagle, J.
Phys. Chem. 100 (1996) 11346.
ꢂ
ꢀ
[34] J. Villa, A. Gonzalez-Lafont, J.M. Lluch, D.G. Truhlar, J.
Am. Chem. Soc. 120 (1998) 5559.
[9] P.W. Seakins, M.J. Pilling, J.T. Niiranen, D. Gutman,
L.N. Krasnoperov, J. Phys. Chem. 96 (1992) 9847.
[10] K. Sato, S. Wakabayashi, T. Matsubara, M. Sugiura, S.
Tsunashima, Y. Kurosaki, T. Takayanagi, Chem. Phys.
242 (1999) 1.
[35] P.J. Marshall, Phys. Chem. A 103 (1999) 4560.
[36] H. Kanamori, Y. Endo, E. Hirota, J. Chem. Phys. 92
(1990) 197.
[37] J.-H. Wang, H.-C. Chang, Y.-T. Chen, Chem. Phys. 206
(1996) 43.
[11] NIST DataBase (http://www.kinetics.nist.gov).
[12] A. Canosa, I.R. Sims, D. Travers, I.W.M. Smith, B.R.
Rowe, Astron. Astrophys. 323 (1997) 644.
[38] Handbook of Chemistry and Physics, CRC Press, Boca
Raton, Florida, 1996.
[13] M.A. Blitz, D.G. Johnson, M. Pesa, M.J. Pilling, S.H.
Robertson, P.W. Seakins, J. Chem. Soc. Faraday Trans. 93
(1997) 1473.
[39] T.J. Lee, P.R. Taylor, Int. J. Quantum Chem. Quantum
Chem. Symp. 23 (1989) 199.
[40] T.J. Lee, J.E. Rice, G.E. Scuseria, H.F. Schaefer, Theor.
Chim. Acta 75 (1989) 81.
[14] D. Fulle, H. Hippler, J. Chem. Phys. 106 (1997) 8691.
[15] R.A. Brownsword, A. Canosa, B. Rowe, I.R. Sims, W.M.
Smith, D.W.A. Stewart, A.C. Symonds, D. Travers, J.
Chem. Phys. 106 (1997) 7662.
[41] M.R. Berman, M.C. Lin, Chem. Phys. 82 (1983) 435.
[42] M.S. Gordon, T.N. Truong, J.A. Pople, Chem. Phys. Lett.
130 (1986) 245.
[16] K. Raghavachari, G.W. Trucks, J.A. Pople, M. Head-
Gordon, Chem. Phys. Lett. 157 (1989) 479.
[43] B.S. Jursic, J. Mol. Struc. Theochem 538 (2001) 19.
[44] L.F. Keyser, J. Phys. Chem. 88 (1984) 4750.
[45] A. Bergeat, T. Calvo, F. Caralp, J.-H. Fillon, G. Dorthe,
J.-C. Loison, Faraday Disc. 119 (2001) 67.
[46] D.L. Baulch, C.J. Cobos, R.A. Cox, C. Esser, P. Frank,
Th. Just, J.A. Kerr, M.J. Pilling, J. Troe, R.W. Walker, J.
Warnatz, J. Phys. Chem. Ref. Data 21 (1992) 411.
[47] H. Thiesemann, J. MacNamara, C.A. Taatjes, J. Phys.
Chem. A 101 (1997) 1881.
[17] T.H. Dunning Jr., J. Chem. Phys. 90 (1989) 1007.
[18] M.J. Frisch et al., Gaussian 98, Revision A.9, Gaussian
Inc., Pittsburgh, PA, 1998.
[19] MOLPRO is a package of ab initio programs written by
H.-J. Werner, P.J. Knowles, with contributions from R.D.
Amos, A. Bernhardsson, A. Berning, P. Celani, D.L.
Cooper, M.J.O. Deegan, A.J. Dobbyn, F. Eckert, C.
Hampel, G. Hetzer, T. Korona, R. Lindh, A.W. Lloyd, S.J.
McNicholas, F.R. Manby, W. Meyer, M.E. Mura, A.
[48] J. Peeters, H. Van Look, B. Ceursters, J. Phys. Chem. 100
(1996) 15124.
€
Nicklass, P. Palmieri, R. Pitzer, G. Rauhut, M. Schutz, H.
[49] B.J. Opansky, S.R. Leone, J. Phys. Chem. 100 (1996) 19904.
[50] B.J. Opansky, S.R. Leone, J. Phys. Chem. 100 (1996) 4888.
[51] By comparison with CH þ H2O reaction.
[52] B. Ceursters, H.M.T. Nguyen, J. Peeters, M.T. Nguyen,
Chem. Phys. Lett. 329 (2000) 412.
Stoll, A.J. Stone, R. Tarroni, T. Thorsteinsson.
[20] C. Hampel, K. Peterson, H.-J. Werner, Chem. Phys. Lett.
190 (1992) 1.
[21] P.J. Knowles, C. Hampel, H.-J. Werner, J. Chem. Phys. 99
(1993) 5219.
[53] S. Tasaki, S. Satyapal, R. Bersohn, Can. J. Chem. 72
(1994) 611.
[22] N. Daugey, A. Bergeat, A. Schuck, P. Caubet, G. Dorthe,
Chem. Phys. 87 (1997) 222.