Absolute Rate Constant and Product Branching Fractions for the Reaction between F and C2H4 at T = 202-298 K

Article abstract of DOI:10.1021/jp9901747

The discharge-flow kinetic technique coupled to mass-spectrometric detection has been used to determine the variable-temperature dependence of the rate constant and product branching fractions for the reaction between F(²P) and C₂H₄ at P = 1 Torr nominal pressure (He). The reaction was studied at T = 202 and 236 K by monitoring the decay of C₂H₄ in the presence of a large excess of F(²P). The overall rate coefficients were determined to be k₁(202 K) = (1.7 ± 0.4) x 10^-10 cm³ molecule^-1 s^-1 and k₁(236 K) = (2.1 ± 0.5) x 10^-10 cm³ molecule^-1 s^-1 with the quoted uncertainty representing total errors. Further, the branching fractions for the two observed reaction channels F + C₂H₄ → C₂H₃ + HF (1a) and F + C₂H₄ → C₂H₃F + H (1b) were determined by quantitatively measuring the yield of C₂H₃F under conditions of excess C₂H₄. The stabilized adduct, C₂H₄F, was not detected at T = 202 K. The derived branching fractions were Γ_1a(202 K) = 0.25 ± 0.09, Γ_1b (202 K) = 0.75 ± 0.16, and Γ_1a(236 K) = 0.27 ± 0.13, and Γ_1b (236 K) = 0.73 ± 0.20, where the quoted uncertainty represents total errors. By inclusion of k₁(298 K) = (3.0 ± 0.8) x 10^-10 cm³ molecule^-1 s^-1, a revised value that used data from our previous study and Γ_1a(298 K) = 0.35 ± 0.04 and Γ_1b (298 K) = 0.65 ± 0.04 from a laser photolysis/photoionization mass spectrometry study, we obtain the Arrhenius expressions k_1a(T) = (7.5 ± 4.0) x 10^-10 exp[(-1.2 ± 0.3)/(RT)] and k_1b(T) = (5.2 ± 1.0) x 10^-10 exp[(-0.6 ± 0.1)/(RT)] in units of cm³ molecule^-1 s^-1 for k and in units of kcal mol^-1 for activation energy. The quoted uncertainty represents total errors at 1σ precision errors plus 15% systematic errors. RRKM calculations have shown that the critical energy for H addition to C₂H₃F is less than 6 kcal mol^-1 larger than that for the addition of F to C₂H₄ and that the competitive decomposition of chemically activated C₂H₄F radicals favor C-H bond rupture by a factor greater than 1000 over that for C-F bond rupture.

Full text of DOI:10.1021/jp9901747

4470
J. Phys. Chem. A 1999, 103, 4470-4479
Absolute Rate Constant and Product Branching Fractions for the Reaction between F and
C2H4 at T ) 202-298 K
Fred L. Nesbitt†
Department of Natural Sciences, Coppin State College, Baltimore, Maryland 21216
R. Peyton Thorn, Jr.*,‡ and Walter A. Payne, Jr.§
Laboratory for Extraterrestrial Physics, NASA/Goddard Space Flight Center, Greenbelt, Maryland 20771
D. C. Tardy*,|
Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
ReceiVed: January 12, 1999; In Final Form: April 15, 1999
The discharge-flow kinetic technique coupled to mass-spectrometric detection has been used to determine
the variable-temperature dependence of the rate constant and product branching fractions for the reaction
between F(2P) and C2H4 at P ) 1 Torr nominal pressure (He). The reaction was studied at T ) 202 and 236
K by monitoring the decay of C2H4 in the presence of a large excess of F(2P). The overall rate coefficients
were determined to be k1(202 K) ) (1.7 ( 0.4) × 10-10 cm3 molecule-1 s-1 and k1(236 K) ) (2.1 ( 0.5) ×
10-10 cm3 molecule-1 s-1 with the quoted uncertainty representing total errors. Further, the branching fractions
for the two observed reaction channels F + C2H4 f C2H3 + HF (1a) and F + C2H4 f C2H3F + H (1b) were
determined by quantitatively measuring the yield of C2H3F under conditions of excess C2H4. The stabilized
adduct, C2H4F, was not detected at T ) 202 K. The derived branching fractions were Γ1a(202 K) ) 0.25 (
0.09, Γ1b (202 K) ) 0.75 ( 0.16, and Γ1a(236 K) ) 0.27 ( 0.13, and Γ1b (236 K) ) 0.73 ( 0.20, where the
quoted uncertainty represents total errors. By inclusion of k1(298 K) ) (3.0 ( 0.8) × 10-10 cm3 molecule-1
s-1, a revised value that used data from our previous study and Γ1a(298 K) ) 0.35 ( 0.04 and Γ1b (298 K)
) 0.65 ( 0.04 from a laser photolysis/photoionization mass spectrometry study, we obtain the Arrhenius
expressions k1a(T) ) (7.5 ( 4.0) × 10-10 exp[(-1.2 ( 0.3)/(RT)] and k1b(T) ) (5.2 ( 1.0) × 10-10 exp[(-
0.6 ( 0.1)/(RT)] in units of cm3 molecule-1 s-1 for k and in units of kcal mol-1 for activation energy. The
quoted uncertainty represents total errors at 1σ precision errors plus 15% systematic errors. RRKM calculations
have shown that the critical energy for H addition to C2H3F is less than 6 kcal mol-1 larger than that for the
addition of F to C2H4 and that the competitive decomposition of chemically activated C2H4F radicals favor
C-H bond rupture by a factor greater than 1000 over that for C-F bond rupture.
Introduction
F + C2H6 reaction, k(298 K) ) 2.2 × 10-10 cm3 molecule-1
s-1.7 Although the prompt radical generation is very desirable
to consume the initial F present and thus prevent secondary
chemistry, the high reactivity of F also leads to a low selectivity,
i.e., multiple pathways. In contrast to the alkanes, the reaction
of fluorine atoms with the an alkene, e.g., C2H4, can occur by
two processes: (1) direct H atom abstraction to form HF and
the corresponding free radical, (2) addition of fluorine atoms
to the double bond to yield an energetic adduct, C2H4F*.8 The
C2H4F* adduct can then decompose through the loss of a H
atom to form vinyl fluoride, C2H3F, or it can be collisionally
stabilized to form an adduct-like product C2H4F.9,10 This
mechanism results in three possible exothermic product path-
ways:
The kinetics of small C2 radicals such as vinyl (C2H3) have
implicit importance in the atmospheric chemistry of the outer
planets,1 in high-temperature chemistry of combustion proc-
esses,2 and in ultralow-temperature chemistry of dense inter-
stellar clouds.3 These C2 radical species are generated either
by thermal or vacuum ultraviolet dissociation from a stable
precursor molecule or by chemical reaction. For example, in
Titan’s atmosphere, C2H3 is produced by the termolecular
association reaction of H with C2H2.1,4 Once produced in such
systems, these radicals serve to interconvert hydrocarbon
species.1
The reaction of fluorine atoms with hydrocarbons is an
important laboratory source of hydrocarbon radicals. For the
reaction of fluorine atoms with alkanes only one pathway is
available: a H atom abstraction with the production of HF and
an alkyl radical.5 The high reactivity of F atoms leads to rate
coefficients that typically are at the collision rate,6 e.g., for the
F(2P) + C2H4 f C2H3 + HF
f [C2H4-F]* f C2H3F + H
(1a)
(1b)
* To whom correspondence should be sent.
† Email: fnesb29280@aol.com.
f [C2H4-F]* + M f C2H4F + M* (1c)
‡ Email: ysrpt@lepvax.gsfc.nasa.gov.
§ Email: u1wap@lepvax.gsfc.nasa.gov.
| Email: dwight-tardy@uiowa.edu.
Thus, the F + C2H4 reaction system illustrates the three
categories of a bimolecular reaction: a metathesis (abstraction),
10.1021/jp9901747 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/22/1999

Reaction between F and C2H4
J. Phys. Chem. A, Vol. 103, No. 23, 1999 4471
a displacement (addition followed by decomposition), and an
association (addition followed by stabilization) reaction.11
Previous reaction dynamic studies have shown that the F +
C2H4 reaction proceeds by parallel abstraction and addition
mechanisms.8 Specifically, a crossed molecular beam study of
F atoms with C2H4 and C2D4 by Parson and Lee8 determined
that the addition adduct was a long-lived complex, which after
several rotational periods eventually released a H atom and vinyl
fluoride. This study also determined that the abstraction channel
produced a highly vibrationally excited DF (populating ν′ ) 2,
the highest thermochemically accessible level) and a vinyl
radical with very little internal energy. The reaction of F atoms
with C2H4 has also been extensively studied by infrared
chemiluminescence techniques.12-15 A chemiluminescence study12
and a high collision energy (2.3-12.1 kcal mol-1) crossed
molecular beam study16 have shown that the energy in the vinyl
fluoride product from reaction 1b was not completely random-
ized because of the exit channel barrier of 5-6 kcal mol-1. A
using a pulsed IR laser photolysis-photoionization mass
spectrometry technique. In this study, the F atoms were
generated by IR multiple photon decomposition of C6F5Cl, and
the absolute branching fraction was determined from measure-
ments of both the C2H4 depletion and the C2H3F product
formation. In an earlier study, Moehlmann and McDonald13
measured the integrated HF infrared chemiluminescence and
after deconvolution of the data determined an addition-
decomposition/abstraction cross-section ratio of 3, i.e., Γ1b
)
0.75. The accuracy of this value is not known because of a need
to assume populations of the ν ) 0 state of HF and the difficulty
in determining absolute Einstein coefficients for C2H3F. In the
first product branching study, which measured the yields of the
18F-containing product compounds by radio gas chromatography,
Williams and Rowland9 reported a total addition channel
branching fraction, Γadd ) Γ1b + Γ1c, of 0.65.
However, there has not been any kinetic or product branching
fraction studies at low temperatures. The objective of this study
is to make direct measurements of the absolute rate constant
and product branching fractions of reaction 1 as a function of
temperature. Measurements were made using the discharge flow
mass spectrometric technique at 1 Torr total pressure. These
studies confirm that reaction 1 is a convenient and quantitative
laboratory source of the C2H3 radical over the temperature range
T ) 202-298 K.
more recent, lower collision energy (0.8-2.5 kcal mol-1
)
crossed molecular beam study17 has established an upper limit
of 0.8 kcal mol-1 as the potential energy barrier to F atom
addition to C2H4.
Despite the intense interest from the reaction dynamics
viewpoint in the F + C2H4 reaction system, there have been
few bulk gas-phase kinetic and product branching fraction
measurements. To date, there are two relative rate measure-
ments18,19 and only one absolute rate measurement20 at T ) 298
K for reaction 1 reported in the literature. Milstein et al.18
reported a relative rate of 0.82 ( 0.02 in 4000 Torr of SF6 for
addition reaction via reaction 1c relative to
Experimental Section
Discharge Flow Reactor. All experiments were performed
in a Pyrex flow tube 60 cm long and 2.8 cm in diameter, the
inner surface of the flow tube being lined with Teflon FEP.
The flow tube was coupled via a two-stage stainless steel
collision-free sampling system to a recently installed computer-
controlled quadrupole mass spectrometer (Merlin mass spec-
trometer, ABB Extrel Corp.) that was operated at low electron
energies (typically less than 20 eV). Ions were detected by an
off-axis conversion dynode/channeltron multiplier (Detector
Technology Corp.). The flow tube has a Pyrex movable injector
for the introduction of the C2H4 reactant, which could be
changed from a distance between 2 and 40 cm from the sampling
pinhole. Helium carrier gas was flowed at 945 sccm into the
reaction flow tube through ports at the rear of the flow tube.
All gas flows were measured and controlled by mass flow
controllers (MKS Instruments). At a typical total pressure of 1
Torr the linear flow velocity was between 2360 and 3000 cm
s-1. This system has been described in detail previously.22
Atomic F Production and Titration. Fluorine atoms were
generated by passing molecular fluorine (ca. 5% diluted in
helium) or CF4 (ca. 10% diluted in helium) through a sidearm
at the upstream end of the flow tube that contained a microwave
discharge (∼50 W, 2450 MHz, Opthos Instruments). The
F(2P) + C2H2 f C2H2F
(2)
and Smith et al.19 reported a relative rate of 0.52 ( 0.08 for
only the abstraction reaction 1a, k1a, relative to
F(2P) + CH4 f HF + CH3
(3)
in ∼1 Torr Ar carrier gas. The previous absolute rate coefficient
measurement,20 performed in this laboratory, yielded the result
k1(total) ) (2.7 ( 0.5) × 10-10 cm3 molecule-1 s-1 using the
same technique and similar conditions described below. The
relative rate measurement of Milstein et al.18 is consistent with
our previous k1(total) value when combined with a value of Γ1c
) k1c/k1(total) = Γ1b ) k1b/k1(total) ) 0.65 from ref 21 and a
recent measurement of k2.20 As discussed later, the total addition
channel rate coefficient, kadd ) k1b + k1c, is believed to be
pressure-independent, and therefore, the only effect of pressure
is the partitioning between the stabilization and decomposition
pathways of the adduct. Thus, the relative measurement of k1c
by Milstein et al.18 at high pressure can approximate a
measurement of k1b at low pressure. The other relative rate
measurement by Smith et al.19 yielded only fair agreement with
our previous k1a value, which was calculated from a value of
Γ1a ) k1a/k1(total) ) 0.35 from ref 21 and our previous k1(total)
value.20
3
discharge region consisted of a /8 in. ceramic tube coupled to
a glass discharge arm. When CF4 was used, a recombination
volume was placed downstream from the microwave discharge
to allow CFx to recombine. The volume was 10 cm in length,
7 cm in diameter Pyrex glass, giving a residence time of ca. 60
ms.
There have been three previous product branching studies
carried out at T ) 295-298 K over a wide range of pressures
and carrier gases: in 100-4000 Torr of SF6 and 240-1580
Torr of CF4,9 in 0.7 Torr of He,21 and in 2 × 10-4 Torr of
C2H4.13 These three studies are in good agreement with one
another and have shown that the addition processes in this
reaction, channels 1b and 1c, occur about twice as frequently
as does abstraction. The most recent and accurate measurement
of Γ1b ) 0.65 ((0.06) was obtained by Slagle and Gutman21
The concentration of fluorine atoms in the kinetic studies was
determined by measuring the Cl2 consumption in the fast
titration reaction
F + Cl2 f FCl + Cl
(4)
k4(298 K) ) 1.6 × 10-10 cm3 molecule-1 s-1 (ref 23)
With Cl2 in excess, the F atom concentration was determined

4472 J. Phys. Chem. A, Vol. 103, No. 23, 1999
Nesbitt et al.
+
by measuring the decrease in the Cl2 signal (m/z ) 70) at an
and O2 (99.999%, Scientific Gas Products, UHP) were used
without further purification. Cl2 (VLSI grade, Air Products),
C2H4 (99%, Air Products), CF4 (99.9%, Matheson), and C2H3F
(98%, PCR Inc.) were degassed using repeated freeze-pump-
thaw cycles at liquid nitrogen temperature.
electron energy of ∼14 eV when the discharge was initiated.
The dilute Cl2/He mixture was admitted via the movable injector.
The position of the injector was chosen to ensure that reaction
4 went to completion and that the position was close to the
middle of the decay range for C2H4 under reaction conditions.
The absolute F concentration is given by [F] ) [Cl2]Disc.Off
-
Results
[Cl2]Disc.On ≡ (∆Cl2 signal)[Cl2]Disc.Off. As discussed previously
for N atom studies,24 a number of precautions were taken in
order to avoid systematic errors in this type of measurement.
Typically, 80-96% of the F2 was dissociated and initial F atom
concentrations were (1.0-5.0) × 1012 molecule cm-3 for the
kinetic studies. All F atom titrations for the kinetic studies were
conducted in the presence of the same oxygen concentration as
used in the decay experiments as discussed below.
Kinetic Studies. The rate measurements were performed
under pseudo-first-order conditions with [F]0 > [C2H4]0 and [F]0/
[C2H4]0 values ranging from 8.4 to 13.6. The decay of C2H4 is
given by the expression
ln[C2H4]t ) -kobs(d/V) + ln[C2H4]0
(5)
where kobs is the measured pseudo-first-order decay constant, d
is the distance from the tip of the movable injector to the
sampling pinhole, and V is the linear velocity. Linear least-
squares analysis of plots of ln(C2H4 signal) at m/z ) 28 vs
contact time yielded the observed pseudo-first-order rate
constant, kobs. Corrections (0.5-3%) were made to kobs to
account for axial diffusion to give kcorr according to the method
of Lewis et al.26 The diffusion coefficient for C2H4 in He was
estimated to be D ) 288 cm2 s-1 and D ) 216 cm2 s-1 at T )
236 K and at T ) 202 K, respectively. Corrections for radial
diffusion were not necessary, since they were always smaller
than axial diffusion.
For the product branching studies, the concentration of
fluorine atoms was determined by measuring the decrease in
the F2+ signal (m/z ) 38) at an electron energy of ∼20 eV when
the discharge was initiated. This method was preferred because
the product yield measurements were performed with the injector
position between d ) 3 and 5 cm from the sampling pinhole as
discussed later in Results. If Cl2 had been used as a titrant at
this injector position, there would not be sufficient time for
reaction 4 to go to completion. Separate experiments with the
injector at 30 cm showed good agreement between the ∆Cl2
and the ∆F2 methods. The absolute F concentration is given by
[F] ) 2([F2]Disc.Off - [F2]Disc.On) ≡ 2(∆F2 signal)[F2]Disc.Off
.
As described previously,20 nonlinearity in the C2H4 decay
curves can be due to regeneration of ethylene via the rapid vinyl
self-reaction
Typically, 80-96% of the F2 was dissociated and initial F
radical concentrations were (1.6-2.6) × 1012 molecule cm-3
for the product branching studies.
In a few kinetic decay and product branching experiments at
T ) 202 K, CF4 was used as the F atom precursor. Typically,
20-30% of the CF4 was dissociated and the initial F atom
concentrations were 1.8 ×1012 and 2.7 × 1012 molecule cm-3
for the kinetic studies and 4.25 × 1012 molecule cm-3 for the
two product branching measurements. For the kinetic studies,
the initial F concentration was determined by Cl2 consumption
as described above. However, for the product branching studies
a ∆CF4+ signal decrease method (analogous to the ∆F2+ signal
decrease method described above) could not be used, since there
C2H3 + C2H3 f C2H2 + C2H4
(6)
k6(298 K) ) 1.41 × 10-10 cm3 molecule-1 s-1 (ref 27)
Molecular oxygen, [O2]0 ) (3.6-3.9) × 1014 molecule cm-3
,
was added to scavenge C2H3.
C2H3 + O2 f HCO + H2CO
(7)
k7(298 K) ) 1.0 × 10-11 cm3 molecule-1 s-1 (ref 28)
+
is not a CF4 parent ion.25 Instead, initial concentrations of F
atoms were determined by measurements of the ClF (m/z )
54) generated in rapid reaction 4.23 As in the Cl2 consumption
method, a dilute Cl2/He mixture was admitted via the movable
injector that was positioned to ensure that reaction 4 went to
completion, typically at d ) 30 cm. Although this injector
position is 25 cm further upstream from where the product yield
measurements were performed, previous experiments have
shown that the F atom concentration profile in the flow tube is
constant. This technique for measuring [F] by ClF formation,
which avoids calibration with an external ClF reagent, has been
thoroughly discussed by Appelman and Clyne.23 This technique
consists of two sequential measurements of the ClF signal. First,
the Cl2 reagent flow was set so that an excess of Cl2 was present
in the flow tube, [Cl2]0/[F]0 g 2, therefore converting all F atoms
to ClF. Then under the same mass spectrometer conditions and
the same [Cl2]0, the [F]0 was greatly increased, thereby
consuming all the Cl2 and the ClF signal measured. This second
step determined the ClF signal calibration, since [ClF] product
) [Cl2]0. Thus, the ClF signal from F + Cl2 could be used to
measure absolute F atom concentration in the 1011-1013
molecule cm-3 range used in this product branching study.
In the presence of O2 the observed first-order decays were
strictly linear as required by eq 5 (see Figure 1). Possible
contributions from the reaction
H + C2H4 + M f C2H5 + M
(8)
k8(1 Torr of He) ) 6.0 × 10-14 cm3 molecule-1 s-1 (ref 29)
to the depletion of C2H4 are negligible (<1%) under the
conditions of the experiment.
To investigate the possibility of additional F atom loss
processes, product observation experiments under the same
pseudo-first-order conditions and at T ) 202 and 236 K were
performed by monitoring the fluoroethylene species C2H3F (m/z
) 46), C2H2F2 (m/z ) 64), and C2HF3 (m/z ) 82). The ionizer
energy used was IE ≈ 14.0 eV, which is above the ionization
energy of vinyl fluoride (IE ) 10.36 eV), all three C2H2F2
isomers, vinylidene fluoride (IE ) 10.29 eV), (Z)-1,2-difluo-
roethylene (IE ) 10.23 eV), (E)-1,2-difluoroethylene (IE )
10.21 eV), and trifluoroethylene (IE ) 10.14 eV).25 The results,
shown in Figure 2, demonstrate that the consumption of C2H3F,
which is the dominant product of reaction 1, occurs simulta-
neously and on approximately the same time scale as the C2H4
Materials. Helium (99.9995%, Air Products) was drawn
through a trap held at 77 K. F2 (4.92% in helium, Air Products)

Reaction between F and C2H4
J. Phys. Chem. A, Vol. 103, No. 23, 1999 4473
TABLE 1: Summary of Rate Data for the F(2P) + C2H4
Reaction at T ) 236 K and T ) 202 Ka
temp atomic F
K
[F]mean
[C2H4]0
kcorr
precursor 1012 molecule cm-3 1011 molecule cm-3 s-1
236
236
236
236
236
236
202
202
202
202
202
202
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
CF4
CF4
1.11
1.18
1.81
2.57
2.67
3.29
1.98
2.71
3.58
4.60
1.61
2.44
1.16
1.57
1.94
2.54
2.76
2.64
1.56
2.44
2.96
4.91
2.80
3.00
190
268
288
491
509
681
325
498
642
757
299
363
a Excess O2 added to scavenge C2H3 radical and prevent regeneration
of C2H4; see text.
the C2HF3 net signal, implying that it is a major product of the
F + C2H2F2 reaction
Figure 1. Plots of ln(C2H4 net signal) vs reaction time at T ) 202 K
and P ) 1 Torr. Concentrations are in units of 1011 molecule cm-3
:
F + C2H2F2 f C2HF3 + H
f other products
(10a)
(10b)
[F]mean ) (a) 19.6, (b) 26.8, (c) 45.4; [C2H4]0 ) (a) 1.56, (b) 2.44, (c)
4.91; [O2]0 ) (a) 393, (b) 392, (c) 394. Solid lines are obtained from
linear least-squares analyses and give the following pseudo-first-order
C2H4 decay rates in units of s-1: (a) 321, (b) 489, (c) 737. For clarity,
traces a and c are shifted on the vertical axis; the actual net signal
counts for trace a are twice that shown and for trace c are half that
shown.
However, reaction 10 becomes significant only near the comple-
tion of reaction 1, and therefore, consumption of F by reaction
10 was neglected in determining [F]mean
.
To allow for the small depletion of F caused by reaction with
C2H4 and with C2H3F as discussed above, measured concentra-
tions of F were corrected according to
[F]mean ) [F]0 - 0.5[C2H4]0 - 0.5[C2H3F]
(11)
The value of [C2H3F] resulting from reaction 1b can be
approximated from the product branching fraction, Γ1b, and the
initial [C2H4]0:
[C2H3F] ) Γ1b[C2H4]0
(12)
For the T ) 298 K data, Γ1b ) 0.65 (ref 21) and thus
[F]mean ) [F]0 - 0.825[C2H4]0
(13)
Since our results from the product branching studies showed
that Γ1b increased slightly with temperature, [F]mean values were
slightly lower at T ) 202 than at T ) 236 and 298 K for the
same [C2H4]0 and [F]0. The range for the correction was 6-13%.
This analysis to determine [F]mean via eqs 11 and 12 was
applied to data from our previous study at T ) 298 K.20 The
[F]mean values are lower than the initial [F]0 by 16-31%. This
is a larger correction than was used for the low-temperature
data because lower [F]0/[C2H4]0 ratios were used in the previous
study.
Figure 2. Plot of observed products at T ) 236 K and P ) 1 Torr
with [F]0 > [C2H4]0. Concentrations are in units of molecule cm-3
[F]0 ) 3.45 × 1012; [C2H4]0 ) 3.40 × 1011; [O2]0 ) 3.66 × 1014.
:
decay. Furthermore, the increasing C2H2F2 net signal correlates
with the decreasing C2H3F net signal, implying that it is a major
product of the F + C2H3F reaction at low pressures
The bimolecular rate constant, k1, is calculated from
kcorr ) k1[F]mean + kw
(14)
F + C2H3F f C2H2F2 + H
f other products
(9a)
(9b)
where kw is a first-order rate constant that accounts for the loss
of C2H4 on the walls of the flow tube or other sources. Table 1
summarizes the rate data and experimental condition for T )
236 and 202 K. Figures 1 and 2 show typical first-order decays
of C2H4 in excess [F]. Figure 3 shows the variation in the
pseudo-first-order rate constant kcorr with [F]mean for reaction 1
at T ) 202-298 K, respectively. A linear least-squares analysis
of the data in Table 1 according to eq 14 gives a bimolecular
rate constant of k1(236 K) ) (2.1 ( 0.5) × 10-10 cm3
Although k9 has not been measured here and has not been
reported in the literature, its value should be between that of
the Cl + C2H3F reaction in the high-pressure limit, 1.85 × 10-10
cm3 molecule-1 s-1 at T ) 298 K (ref 30), and that of reaction
1, k1(298 K) ) 2.7 × 10-10 cm3 molecule-1 s-1 20
.
Similarly,
the decay of the C2H2F2 net signal correlates with the rise of

4474 J. Phys. Chem. A, Vol. 103, No. 23, 1999
Nesbitt et al.
Product Branching Studies. The only products of reaction
1 observed at both T ) 202 and T ) 236 K were C2H3F and
C2H3. These products are consistent with previous stud-
ies.8,9,16-18,20,21 To determine the product branching fractions,
measurements were made by monitoring the C2H3F net signal
at m/z ) 46 directly as a function of distance (reaction time).
The experiments were conducted with [C2H4]0 > [F]0 to avoid
potential loss of C2H3F via the secondary reaction 9. [C2H4]0/
[F]0 values ranged from 14 to 41. As in the kinetic experiments,
molecular oxygen, [O2]0 ) (3.6-5.5) × 1014 molecule cm-3
,
was added to scavenge C2H3. With [C2H4]0 ) (5.3-6.9) × 1013
molecule cm-3, the C2H3F signal profile leveled off between t
) 1.1 and 1.8 ms, indicating that the F + C2H4 reaction had
gone to completion. At longer reaction times, t > 2 ms, the net
C2H3F signal increased above this constant level. This growth
in signal was especially noticeable when the percent dissociation
of F2 in the microwave discharge was <90%, suggesting that
atomic hydrogen, formed along with C2H3F in reaction 1b, reacts
with undissociated F2
Figure 3. Summary plot of the corrected pseudo-first-order rate
constant kcorr vs [F]mean at P ) 1 Torr. The data points at T ) 202, 236,
and 298 K are indicated by circles, open squares, and open triangles,
respectively. Open circle data points used F2 as atomic F precursor.
Solid circle data points used CF4 as atomic F precursor. The lines are
obtained from a linear least-squares analysis. At T ) 202 K the slope
of the dotted dashed line yields k1 ) (1.66 ( 0.41) × 10-10 cm3
molecule-1 s-1 and the intercept yields kw ) +13 ( 51 s-1. At T )
236 K the slope yields k1 ) (2.07 ( 0.53) × 10-10 cm3 molecule-1 s-1
and the intercept yields kw ) -32 ( 54 s-1. At T ) 298 K the slope
yields k1 ) (3.03 ( 0.78) × 10-10 cm3 molecule-1 s-1 and the intercept
yields kw ) -114 ( 106 s-1. Quoted uncertainties are 1σ plus 15%.
H + F2 f HF + F
(15)
k15(T) ) 1.46 × 10-12 exp(-1210/T)
cm3 molecule-1 s-1 (ref 31)
This reaction, followed by reaction 1, leads to additional C2H3F
formation. Consequently, only experiments that used CF4 as a
F atom source or had a high F2 dissociation (and thus low
residual [F2]) yielded accurate product branching fractions.
The final product C2H3F signal levels were taken as the
average of the signals in the plateau region. The magnitudes of
the product signals were calibrated using a range of appropriate
known concentrations of a reference C2H3F/He mixture under
similar flow conditions. The product branching fraction, Γ1b,
was determined from the equation
TABLE 2: Summary of Revised Rate Data for the F(2P) +
C2H4 Reaction at T ) 298 Ka
b
atomic F
precursor
[F]mean
[C2H4]0
kcorr
1012 molecule cm-3
1011 molecule cm-3
s-1
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
F2
2.35
0.84
1.93
2.94
3.79
3.61
4.22
2.22
3.11
3.23
3.45
2.79
2.21
1.03
1.35
2.40
9.31
4.33
7.23
9.20
14.3
12.9
17.8
5.79
16.9
12.5
18.1
10.6
6.55
3.60
4.38
5.72
559
188
498
461
963
862
1384
552
915
808
1068
794
513
250
318
599
Γ1b ) [C2H3F]/[F]0
For the abstraction channel
Γ1a ) 1 - Γ1b
(16)
(17)
and since C2H4F was not detected as a stable product at P ) 1
Torr, then Γ1c ) 0. The product branching fraction results are
summarized in Table 3 and give the following branching
fractions: Γ1b(236 K) ) 0.73 ( 0.20 and Γ1b(202 K) ) 0.75 (
0.16. Using eq 17 to determine the abstraction channel branching
fraction gives Γ1a(236 K) ) 0.27 ( 0.13 and Γ1a(202 K) )
0.25 ( 0.09. The quoted uncertainties are statistical at 1σ and
include an additional 15% for estimated systematic errors.
At an ionization energy of ∼14 eV and P ) 1 Torr (He), a
net signal at m/z ) 47 was detected at T ) 202 K. This signal
was previously reported in our studies20 at T ) 298 K but was
unidentified. Two experiments were performed to determine
whether the net signal observed at m/z ) 47 was due to the
C2H4F addition-stabilization product of reaction 1c. First,
simultaneous measurements of both m/z ) 46 and 47 signals at
this low temperature and under the vinyl fluoride yield
experimental condition described above showed that the tem-
poral profiles of the two signals matched over the range 2.0 <
t < 15.9 ms. The averaged ratio of m/z ) 47 to m/z ) 46 net
signal was 0.0233 ( 0.0010. This value is in close agreement
with the natural abundance of the 13C isotope in the C2H4 reagent
(2.2%) and, hence, in the C2H3F vinyl fluoride reaction product.
Second, further verification of the absence of C2H4F was
a Original data from ref 20. Only [F]mean values have been revised.
b Previous [F]mean ) [F]0 - 0.5[C2H4]0; revised [F]mean ) [F]0
-
0.825[C2H4]0 (see eq 13).
molecule-1 s-1 and k1(202 K) ) (1.7 ( 0.4) × 10-10 cm3
molecule-1 s-1. Our previous room-temperature measurement
for this reaction20 was k1(298 K) ) (2.7 ( 0.5) × 10-10 cm3
molecule-1 s-1. The reanalyzed data from this study is presented
in Table 2 and Figure 3. A linear least-squares analysis of the
revised data in Table 2 according to eq 14 gives a bimolecular
rate constant of k1(298 K) ) (3.0 ( 0.8) × 10-10 cm3
molecule-1 s-1. The intercepts kw ) +13 ( 51, -32 ( 54,
and -114 ( 106 s-1 at T ) 202, 236, and 298 K, respectively,
are statistically insignificant, thus showing that there are no
additional C2H4 loss processes in this system. Quoted uncertain-
ties are statistical at the 1σ level plus systematic errors estimated
to be about 15%.

Reaction between F and C2H4
J. Phys. Chem. A, Vol. 103, No. 23, 1999 4475
stronger than that in ethane by only 10.9 kcal mol-1 33
Also,
the activation barrier for F atom addition, Ea,add ) 0.57 kcal
mol-1 is only slightly less than the upper limit of 0.8 kcal mol-1
established by Robinson et al.17 in a low-energy crossed
molecular beam study.
Alternatively, the temperature dependence of this reaction can
be parametrized by a power dependence expression, k(T) ) bTm.
Applying least-squares analysis to the rate constants given in
Table 4 gives
TABLE 3: Summary of the Experimentally Determined
Product Branching Fractions Γ1a and Γ1b at T ) 202 K and
T ) 236 K for the Reaction F + C2H4 f C2H3 + HF (1a)
and F + C2H4 f C2H3F + H (1b)
.
a
[F]0
[C2H4]0
temp 1012 molecule 1013 molecule
cm-3
cm-3
Γ1a
Γ1b
b
c
K
236
236
236
236
236
236
1.11
1.11
1.78
1.78
2.58
2.58
3.54
3.54
7.04
7.04
7.05
7.05
0.39
0.37
0.20
0.26
0.20
0.18
0.61
0.63
0.80
0.74
0.80
0.82
k1a(T) ) (1.16 × 10-16)T 2.41 cm3 molecule-1 s-1
k1b(T) ) (2.55 × 10-13)T 1.17 cm3 molecule-1 s-1
d
d
0.27 ( 0.13
0.73 ( 0.20
202
202
202
202
202
202
202
202
202
202
1.62
1.62
1.63
1.63
2.37
2.37
2.50
2.50
4.25e
4.25e
6.60
6.60
6.57
6.48
6.58
6.49
6.59
6.59
5.94
5.94
0.24
0.26
0.29
0.24
0.21
0.17
0.24
0.23
0.76
0.75
0.71
0.76
0.79
0.83
0.76
0.77
This power dependence expression for k1b may be preferred over
the Arrhenius expression because of the absence of an energy
barrier at the entrance of the potential energy surface for the F
+ C2H4 system.17
RRKM Model. The observed branching ratio, k1c/k1b, can
be used to provide information on the thermokinetics for the
addition-decomposition reactions. The addition of F to C2H4
and its subsequent decomposition (D) to H + C2H3F or
stabilization (S) to C2H4F can be modeled as a simple chemical
activation system34 in which the adduct is formed with an
internal energy distribution of populated states.
0.36
0.23
0.64
0.77
d
d
0.25 ( 0.09
0.75 ( 0.16
a Unless noted, F2 used as F atom precursor. [O2]0 ) 4.5 × 1014
c
molecule cm-3
.
b Γ1a ) 1 - Γ1b. Γ1b ) [C2H3F]/[F]0. d Quoted errors
are 1σ statistical plus 15%. Nominal pressure ) 1 Torr (He). e CF4
used as F atom precursor. [O2]0 ) 5.5 × 1014 molecule cm-3
.
F + C2H4 f C2H4F(E),
kf
(1b′)
(1b′′)
obtained by the addition of a large excess of O2 as a scavenger.
C2H4F(E) f H + C2H3F, D, k(E)
C2H4F(E) f F + C2H4, D′, k′(E)
C2H4F + O2 f products (18)
(-1b′)
A signal decrease at m/z ) 47 was not observed at [O2]0 ) 1.5
× 1015 molecule cm-3 and at t ) 2 ms. Thus, the m/z ) 47
signal seen here and in our previous study is not due the
presence of the pressure-stabilized free radical adduct C2H4F
but due to the isotopic vinyl fluoride product 13C2H3F.
C2H4F(E) + M f C2H4F(E′) + M, S, k2(E′,E) (1c′)
where kf is the rate coefficient for the formation reaction, k(E)
and k′(E) are the microscopic unimolecular rate coefficients for
decompositions, and k2(E′,E) is the rate coefficient for inter-
molecular energy from internal energy E to E′. A potential
energy profile with defining energies is exhibited in Figure 5.
The critical energies are designated as either E0 or E0(X), while
reaction energy changes are designated by ∆E00(reactant;product).
The excess energy E+ is the internal energy of the transition
state and is the difference between the internal energy of reactant
and the critical energy for reaction, i.e. E+ ) E - E0; the
Discussion
Arrhenius Expression. The results from this study at low
temperatures are combined with the revised k1(298 K) values
using data from our previous study20 and the previous product
branching studies.21 The rate constants of the two separate
channels are obtained from the total rate constant and the
corresponding branching fraction, i.e., k1a(T) ) k1(T)Γ1a(T) and
k1b(T) ) k1(T)Γ1b(T) as shown in Table 4. The rate coefficients
for reactions 1a and 1b are plotted in Figure 4 as a function of
reciprocal temperature, T-1. As can be seen from Figure 4, both
the abstraction and the addition channels have rate coefficients
that increase with temperature. The lines in Figure 4 are obtained
from a linear least-squares analysis of the ln k1a vs T-1 and ln k1b
vs T-1 data; these analyses yield the following Arrhenius
expressions:
minimum excess energy, E+
is the difference in critical
min
energies: E+min ) E0′ - E0. The k(E)’s can be computed with
the RRKM model,35
P(ꢀ+)
F(E)
+
+
r Q ∑
k(E) )
h Q
where r+ is the reaction path degeneracy, Q+/Q is the adiabatic
partition function ratio for rotations, ∑P(ꢀ+) is the sum of all
active internal energy eigenstates of the transition complex with
total energy E+, and F(E) is the density of states for the
energized reactant with energy E. If the addition product is
formed by thermalized reactants, then the distribution of internal
energy states for C2H4F is given by36
k1a(T) ) (7.5 ( 4.0) × 10-10 exp[(-1.18 ( 0.35)/(RT)]
cm3 molecule-1 s-1
k1b(T) ) (5.2 ( 1.0) × 10-10 exp[(-0.57 ( 0.10)/(RT)]
cm3 molecule-1 s-1
k′(E) B(E)
f(E) )
for E ) E0′ to ∞
where the errors are quoted at the 1σ plus 15% level and units
of kcal mol-1 are used for the activation energy.
k′(E) B(E)
∑
If the Arrhenius activation energy parameter is used as a
barrier energy height, then these results show that the activation
barrier for H atom abstraction, Ea,absr ) 1.18 kcal mol-1, is
almost twice as large as that for the very fast F + C2H6 reaction,
where B(E) is the Boltzmann distribution
F(E) exp(-E/(RT))
B(E) )
for E ) 0 to ∞
Ea,absr ) 0.7 kcal mol-1 32
.
Yet the C-H bond in ethylene is
∑
F(E) exp(-E/(RT))

4476 J. Phys. Chem. A, Vol. 103, No. 23, 1999
Nesbitt et al.
TABLE 4: Summary of Values for k1(T) and Product Branching Fractions for the F(2P) + C2H4 Reaction
k1 /10-10 cm3
k1ad/10-11 cm3
k1bd/10-10 cm3
a
molecule-1 s-1
Γ1a
Γ1b
molecule-1 s-1
molecule-1 s-1
b
c
temp K
298
236
202
3.0 ( 0.8e
2.1 ( 0.5g
1.7 ( 0.4g
0.35 ( 0.04f
0.27 ( 0.13g
0.25 ( 0.09g
0.65 ( 0.06f
0.73 ( 0.20g
0.75 ( 0.16g
10.6 ( 2.8
5.6 ( 1.6
4.2 ( 1.1
1.97 ( 0.54
1.51 ( 0.57
1.25 ( 0.41
b
a Quoted uncertainties are statistical at one standard deviation plus 15% for systematic errors. Γ1a is the branching fraction for the abstraction
product channel forming C2H3 + HF. c Γ1b is the branching fraction for the addition product channel forming C2H3F + H. d The combined uncertainties
in k1a and k1b are calculated as the product of the absolute k values and the relative combined uncertainty values. The relative combined uncertainty
values are obtained as the square root of the sum of the squares of the individual relative uncertainties. e Reanalysis of data from ref 20 as discussed
in text and reported in Table 2. f Reference 21; quoted uncertainties are at 10%. g This study. Quoted uncertainties are statistical at one standard
deviation plus 15% for systematic errors.
d[N(E)]/dt ) f(E) + k(E,E′)[M][N(E′)] -
∑
k(E′,E)[M][N(E)] - k(E)[N(E)] - k′(E)[N(E)]
∑
for all E. For steady-state conditions, all d[N(E)]/dt ) 0, the
resulting coupled algebraic equations can be solved for [N(E)]ss.38
The steady-state populations can then be used to calculate the
branching ratios, i.e., S/D and D′/D.
S ) k(E′,E)[N(E)]
for all E g E0 and E′ < E0
for all E g E0
∑
ss
D ) k(E)[N(E)]
∑
ss
D′ ) k′(E)[N(E)]
for all E g E0′
∑
ss
S + D + D′ ) 1
Thus, the amounts of decomposition and stabilization can be
calculated from the k(E)’s and the Boltzmann distribution at
ambient temperature and pressure.
Figure 4. Arrhenius plot for the F(2P) + C2H4 reaction. Separate plots
are shown for the H abstraction channel (open circles), k1a, and for the
addition-decomposition channel (open squares), k1b. Solid lines are
obtained from linear least-squares analyses of the ln k1a vs T-1 and the
ln k1b vs T-1 data and yield the Arrhenius expressions given in the text.
Error bars indicate (1σ plus 15% for both k1a and k1b.
To calculate the k(E)’s, vibrational frequencies and moments
of inertia for the transition states and C2H4F radical must be
known. There are no direct experiments; however, ab initio
calculations for the radical and transition states have been
reported.39 In the present kinetic calculations the previously
reported vibrational frequencies and geometries were used; the
energetics were optimized with respect to the present experi-
mental observations.
The details of k(E′,E) are not known a priori; often the
observed pressure dependence of S/D or D′/D is used to
parametrize k(E′,E). Typically, three parameters are used to
parametrize k(E′,E):40 the Lennard-Jones collision frequency (σ
and ꢀ), the shape of the energy-transfer distribution (exponential,
Gaussian, or step ladder), and the average energy removed per
collision, ∆Ed . An exponential model is often chosen for weak
colliders such as the rare gases, and the Lennard-Jones collision
parameters can be estimated. There are no reported values for
∆Ed in the C2H4F system; however, an estimate can be made
by comparing similar systems, i.e. radicals with similar excita-
tion and critical energies. The chemically activated ethyl41 and
butyl42 radicals with ∼40 kcal mol-1 of internal energy are
formed by the addition of H to the appropriate olefin. The critical
energies for decomposition are 40 and 33 kcal mol-1 for the
ethyl and butyl radicals, respectively. The reported results
indicate that ∆Ed for helium is independent of temperature
between 78 and 300 K and is ∼400 cm-1. The Michael
group43-45 has also extensively studied and performed RRKM
calculations on chemically activated ethyl radicals. However,
their strong collider calculations used a pressure-independent
collisional efficiency factor for helium, i.e. simple pressure
displacement. Thus, their calculations did not include a colli-
sional deactivation cascade; this is important for low-pressure
systems.
Figure 5. Potential energy profile for C2H4F system depicting energies
for reactants, intermediates, transition states, and products.
In these experiments the deactivator (M), helium, is known to
be a weak collider;37 i.e., helium does not remove sufficient
energy from C2H4F to completely quench reactions 1b′′ and
-1b′. The strong collision assumption requires that stabilization
is the result of a single collision. Thus, for a weak collider,
stabilization results from sequential collisions; at low collision
rates, i.e., low pressures, the unimolecular processes are
enhanced.
The populations can be calculated by solving the master
equation

Reaction between F and C2H4
J. Phys. Chem. A, Vol. 103, No. 23, 1999 4477
Figure 6. Plots of log(S/D) vs log(pressure) at 202 K for strong collider
(dashed line) for exponential models with ∆Ed ) 327 (dashed-dotted
line), 414 (solid line), and 498 (dash-dot-dotted line) cm-1 and for
the 414 exponential model at 298 K (dotted line). The critical energies
are E0 ) 38.2 kcal mol-1 and E0′ ) 46.0 kcal mol-1 with E+min ) 7.8
Figure 9. Plots of D′/D vs log(pressure) as a function of temperature
for a strong collider: T ) 202 K (solid line), 236 K (long dashed line),
and 298 K (short dashed line). Similar plots for an exponential model
with ∆Ed ) 414 cm-1 as a function of temperature are also shown:
T ) 202 K (dashed-dotted line), 236 K (dash-dot-dotted line), and
298 K (dotted line). The critical energies are E0 ) 38.2 kcal mol-1 and
kcal mol-1
.
E0′ ) 46.0 kcal mol-1 with E+ ) 7.8 kcal mol-1
.
min
collision efficiency is <0.01. Also to be noted is that for all
pressures a decrease in ambient temperature produces an
increase in S/D; thus, in this work where an upper limit for S
of 0.005 is reported (i.e., S/D ≈ 5 × 10-3, since D ≈ 1) the
lowest temperature (202 K) is used for the comparison. A change
in S/D with decreasing temperature is due to two effects: the
collision number and the average energy of reacting radicals.
A decrease in temperature produces a net increase in the
collision frequency, which results in an increase of S/D, while
a decrease in temperature also reduces the average energy of
radicals so that fewer “weak” collisions are required for
stabilization, i.e., S/D increases with decreasing temperature for
a constant ∆Ed . Thus, both factors predict that S/D will
increase with decreasing temperature.
Figure 7. Plots of log(S/D) vs log(pressure) at T ) 202 K for an
exponential model with ∆Ed ) 414 cm-1: E0′ ) 46.0 kcal mol-1
with E0 ) 40.2 (dashed-dot-dotted line), 38.2 (solid line), and 36.2
(dash-dotted line) kcal mol-1 and E0 ) 38.2 kcal mol-1 with E0′ )
The dependence of S/D vs pressure at 202 K on E0 and E0′
are shown in Figure 7. An increase in E0′ or a decrease in E0
decreases S/D. There are a number of combinations of E0 and
E0′ that are consistent with the upper limit for the experimental
observation; in fact, only an upper limit of E0 for a given E0′ or
a lower limit of E0′ for a given E0 can be estimated. In Figure
44.0 (dotted line) and E0′ ) 48.0 (dashed line) kcal mol-1
.
8 a plot of E+ vs E0′ for three values of S/D (0.002, 0.005,
min
0.008) is shown; for the present experimental conditions S/D
≈ S, since D ≈ 1. For the indicated range of S/D these plots
illustrate that S/D is more sensitive to E+min than it is to E0′. To
maintain a constant S/D, a change in E+ of 1 kcal mol-1 is
min
equivalent to a change of 4.5 kcal mol-1 in E0′. Thus, the present
experiments can provide an estimate for E+
.
min
By use of reported values for enthalpies of formation at 0 K
for F, C2H4, H, and C2H3F, ∆E0 ) 13.0 kcal mol-1 and the
0
best estimate for E0′ ) 46 kcal mol-1 gives E+ > 7 kcal
min
mol-1. From the potential energy profile, it can be seen that
Figure 8. Plots of E+min vs E0′ for three values of S/D: S/D ) 0.008
(solid line), 0.005 (dotted line), and 0.002 (dashed line).
0
E+min ) E0(F) - E0(H) + ∆E0 (C2H4;C2H3F)
The results of the steady-state calculations are summarized
in Figures 6-9. The effect of the dependence of S/D on pressure,
temperature, and ∆Ed is shown in Figure 6. For all collision
models (strong collider and exponential models with ∆Ed of
327, 414, and 498 cm-1) S/D decreases with decreasing pressure;
the strong collider exhibits near-linearity over the whole pressure
range, while S/D for the weak colliders exhibit a large deviation
from the strong collider as the pressure decreases. At high
pressures (>100 Torr) a scale factor (collisional efficiency) of
∼0.2 can be used in converting the weak collider pressure to
an effective strong collider pressure. However, below 100 Torr,
the collision efficiency is pressure-dependent; at 1 Torr the
so that
E0(H) - E0(F) )
∆E0 (C2H4;C2H3F) - E+min < 6 kcal mol-1
0
or with E0(F) ) 1 kcal mol-1, then
E0(H) )
0
E0(F) + ∆E0 (C2H4;C2H3F) - E+min < 7 kcal mol-1

4478 J. Phys. Chem. A, Vol. 103, No. 23, 1999
Nesbitt et al.
Thus, the critical energy for H addition to C2H3F is less than 6
kcal mol-1 larger than that for the addition of F to C2H4, i.e.,
the electronegative fluorine atom reduces the π electron density
and thus increases the critical energy for addition by less than
Perhaps a better, but not perfect, analogue of the F atom is
the cyano radical, CN, often referred to as a pseudo-halogen51
because of its high electron affinity (3.86 eV, ref 52), which is
comparable to that of F (3.40 eV, ref 53). There have been
numerous kinetic studies on the CN + C2H4 reaction,54-62 but
there has been only one product branching fraction study.63 As
in the F + C2H4 reaction, the total rate of this reaction has been
found to be independent of total pressure.56,59,60,62 However, a
comparison of the temperature dependency of these two reaction
rates shows major differences. In contrast to the CN + C2H4
reaction, which has a reported Ea of -0.34 kcal mol-1 (ref 62),
we have observed a positive temperature dependence in the F
+ C2H4 reaction overall as well as separately in the abstraction
channel and in the addition channel. There are significant
differences in the mechanisms between the F + C2H4 and the
CN + C2H4 reactions in the partitioning between the abstraction
and the addition-decomposition channels. While the total rate
coefficient at T ) 298 K for both reactions are almost the same,
k1(298 K) ) 3.0 × 10-10 cm3 molecule-1 s-1 and k20(298 K)
) 2.5 × 10-10 cm3 molecule-1 s-1 (average of seven studies;
refs 56-62), the product branching fractions for the addition-
decomposition channels (Γadd.-decomp) are quite different. For
the F reaction, Γadd.-decomp is 0.65 but only 0.20 for the CN
reaction.63 The F atom reaction occurs via two parallel
processes: direct H atom abstraction (reaction 1a) and addition
to the C-C double bond (reactions 1b and 1c).8,16,19 For the
CN reaction, all products are suggested to arise from a single
activated complex,60,63 i.e.,
6 kcal mol-1
.
These energetics also predict the pressure dependence of D′/D
as a function of temperature as shown in Figure 9. For a given
pressure, D′/D increases with increasing temperature. This is
understood, since the average energy of reacting molecules
increases with increasing temperature and the change in the slope
of k(E) with energy increases more for reaction -1b′ than it
does for 1b′′. The pressure dependence of D′/D is more complex;
it involves both the increase of the average energy of reacting
molecules with increasing pressure, which favors D′ and the
increase in steady-state population below E0′, which favors D.
At 1 Torr of helium these effects cancel one another and it is
also observed that D′/D is nearly independent of the energy-
transfer model and D′/D increases from 7 × 10-4 at 202 K to
1.5 × 10-3 at 298 K. Thus, reaction -1b′ is unimportant. Details
of this reaction could be obtained by studying the addition of F
to cis-C2H2D2 or trans-C2H2D2.
Comparison of the Reactions of F, Cl, and CN with C2H4.
For any new kinetic results for an elementary reaction, it is
valuable to compare the new results to different but related
reactions. As discussed in our previous paper,20 the reaction of
F with C2H4 is not analogous to the reactions of the other
halogen atoms Cl and Br with ethylene. For these reactions,
until recently, only the pressure-stabilized adducts C2H4Cl46-48
or C2H4Br48,49 were reported as products. Thermochemical
calculations based on a heat of formation of the vinyl radical
of 70.6 kcal mol-1 (ref 47) show that the addition-decomposi-
tion channel and the abstraction channel are endothermic for
these reactions. Recently, however, low-pressure experiments
(P ) 0.2-20 Torr) by Kaiser and Wallington47 and by Pilgrim
and Taatjes50 have verified the presence of an abstraction
channel in the Cl + C2H4 reaction
CN + C2H4 f [C2H4-CN]* f C2H3 + HCN
f [C2H4-CN]* f C2H3CN + H
(20a)
(20b)
Evidence for this mechanism is twofold. First, all four temper-
ature-dependent rate studies59-62 observed a slightly negative
temperature dependence ultimately down to T ≈ 50 K61 and a
pressure independence up to P ) 500 Torr of Ar.62 Second,
even though the C-H bond strength is larger in C2H2 than in
C2H4, the overall rate coefficients for both CN + C2H4 and CN
+ C2H2 are identical over the temperature range T ) 100-704
K.59-62
Cl + C2H4 f C2H3 + HCl
(19a)
and have measured a rate coefficient at T ) 297-383 K.
k
19a(T) ) 6.0 × 10-11 exp(-3270/T)
Conclusion
cm3 molecule-1 s-1 (ref 47)
The primary results of this study are threefold. First, the data
from our discharge flow mass spectrometry experiments at low
temperatures show that the F + C2H4 rate constant increases
with temperature. If an Arrhenius expression is used to fit the
data, activation barriers of 1.2 and 0.6 kcal mol-1 are observed
for the H atom abstraction and the F atom addition channels,
respectively.
The second conclusion is that the branching fraction for the
addition-decomposition channel increases only marginally with
a decrease in temperature. Also, the addition-stabilization
product, C2H4F, was not observed at low temperatures and at
P ) 1 Torr. Thus, the F + C2H4 reaction can be used as a
convenient laboratory source of C2H3 radicals at low temper-
atures if one recognizes complications from the production via
reaction 1b of hydrogen atoms that can react rapidly with C2H3.
Finally, we have shown via RRKM calculations that the
critical energy for H addition to C2H3F is less than 6 kcal mol-1
larger than that for the addition of F to C2H4 and that the
competitive decomposition of chemically activated C2H4F
radicals favors C-H bond rupture by a factor greater than 1000
over that for C-F bond rupture.
H atom abstraction is the minor channel for both Cl + C2H4
and F + C2H4. For Cl + C2H4, the abstraction channel has a
branching fraction of only 0.0035 at P ) 1 Torr, whereas it is
0.35 for F + C2H4, a 100-fold difference.21
Since both the abstraction and the addition-decomposition
channel are endothermic in the Cl + C2H4 reaction, the
addition-stabilization channel is dominant at P > 3 mTorr
Cl + C2H4 (+M) f C2H4Cl (+M)
(19b)
Kaiser and Wallington47 have reported a limiting high-pressure
rate coefficient for Cl + C2H4 of k∞,19 ) 5.7 × 10-10 cm3
molecule-1 s-1, which is almost a factor of 2 faster than the
total rate constant for F + C2H4, k1 ) 3.0 × 10-10 cm3
molecule-1 s-1 at T ) 298 K. The only effect of pressure in
the F + C2H4 system will be on the partitioning between
addition-decomposition (reaction 1b) and addition-stabilization
(reaction 1c). This results from the observation that k(addition)/
9
k(total) for F + C2H4 is 0.65 in 4000 Torr of SF6 and in 0.7
Torr of He19 and is ∼0.75 in 2 × 10-4 Torr of C2H4.13

Reaction between F and C2H4
J. Phys. Chem. A, Vol. 103, No. 23, 1999 4479
Warnatz, J. J. Phys. Chem. Ref. Data 1995, 24, 1609.
(30) Iyer, R. S.; Chen, C.-Y.; Rowland, F. S. J. Phys. Chem. 1985, 89,
2042.
(31) Baulch, D. L.; Duxbury, J.; Grant, S. J.; Montague, D. C. J. Phys.
Chem Ref. Data, Suppl. 1981 10, 1.
(32) Maricq, M. M.; Szente, J. J. J. Phys. Chem 1994, 98, 2078.
(33) NIST Webbook, http://webbook.nist.gov.
Acknowledgment. R.P.T. and F.L.N. acknowledge the
support of NASA Cooperative Agreement NCC5-68 to the
Catholic University of America and the NASA Planetary
Atmospheres Program. All authors thank Dr. Louis J. Stief for
the use of laboratory facilities and for helpful discussions.
(34) Rabinovitch, B. S.; Flowers, M. C. Q. ReV. 1964, 18, 122.
(35) (a) Marcus, R. A.; Rice, O. K. J. Phys. Colloid Chem. 1951, 55,
894. (b) Marcus, R. A. J. Chem. Phys. 1952, 20, 359.
(36) Rabinovitch, B. S.; Diesen, R. W. J. Chem. Phys. 1959, 30, 735.
(37) Tardy, D. C.; Rabinovitch, B. S. Chem. ReV. 1977, 77, 369.
(38) Tardy, D. C.; Rabinovitch, B. S. J. Chem. Phys. 1968, 48, 5194.
(39) (a) Schlegel, H. B. J. Phys. Chem. 1982, 86, 4878. (b) Schlegel,
H. B.; Bhalla, K. C.; Hase, W. L. J. Phys. Chem. 1982, 86, 4883.
(40) Oref, I.; Tardy, D. C. Chem. ReV. 1990, 90, 1407.
(41) Current, J. H.; Rabinovitch, B. S. J. Chem. Phys. 1964, 40, 2742.
(42) Kohlmaier, G.; Rabinovitch, B. S. J. Chem. Phys. 1963, 38, 1692,
1709.
(43) Michael, J. V.; Suess, G. N. J. Chem. Phys. 1973, 58, 2807.
(44) Cowfer, J. A.; Michael, J. V. J. Chem. Phys. 1975, 62, 3504.
(45) Lee, J. H.; Michael, J. V.; Payne, W. A.; Stief, L. J. J. Chem. Phys.
1978, 68, 1817.
(46) Maricq, M. M.; Szente, J. J.; Kaiser, E. W. J. Phys. Chem. 1993,
97, 7970.
(47) Kaiser, E. W.; Wallington, T. J. J. Phys. Chem. 1996, 100, 4111.
(48) Yarwood, G.; Peng, N.; Niki, H. Int. J. Chem. Kinet. 1992, 24,
369.
(49) Barnes, I.; Bastian, K. H.; Overath, R.; Tong, Z. Int. J. Chem. Kinet.
1989, 21, 499.
(50) Pilgrim, J. S.; Taatjes, C. A. J. Phys. Chem. A 1997, 101, 4172.
(51) Bersohn, R. In Molecular Energy Transfer; Levine, R. D., Jortner,
J., Eds.; Wiley: New York, 1976; p 154.
(52) Bradforth, S. E.; Kim, E. H.; Arnold, D. W.; Neumark, D. M. J.
Chem. Phys. 1993, 98, 800.
References and Notes
(1) Yung, Y. L.; Allen, M.; Pinto, J. P. Astrophys. J., Suppl. Ser. 1984,
55, 465. Lara, L. M.; Lellouch, E.; Lopez-Moreno, J. J.; Rodrigo, R. J.
Geophys. Res. 1996, 101, 23261. Romani, P. N.; Bishop, J.; Be´zard, B.;
Atreya, S. Icarus 1993, 106, 442.
(2) Tsang, W.; Hampson, R. F. J. Phys. Chem. Ref. Data 1986, 15,
1087.
(3) Herbst, E.; Leung, C. M. Mon. Not. R. Astron. Soc. 1986, 222,
689. Herbst, E.; Leung, C. M. Astrophys. J., Suppl. Ser. 1989, 69, 271.
(4) Payne, W. A.; Stief, L. J. J. Chem. Phys. 1976, 64, 1150.
(5) Jones, W. E.; Skolnik, E. G. Chem ReV. 1976, 76, 562.
(6) Fettis, G. C.; Knox, J. H. Prog. React. Kinet. 1964, 2, 2.
(7) Maricq, M. M.; Szente, J. J. J. Phys. Chem. 1994, 98, 2078.
(8) Parson, J. M.; Lee, Y. T. J. Chem. Phys. 1972, 56, 4658.
(9) Williams, R. L.; Rowland, F. S. J. Phys. Chem. 1972, 76, 3509.
(10) Flores, A. L.; deB. Darwent, B. J. Phys. Chem. 1969, 73, 2203.
(11) Benson, S. W. Thermochemical Kinetics, 2nd ed.; J. Wiley and
Sons: New York, 1976.
(12) Moehlmann, J. G.; Gleaves, J. T.; Hudgens, J. W.; McDonald, J.
D. J. Chem. Phys. 1974, 60, 4790.
(13) Moehlmann, J. G.; McDonald, J. D. J. Chem. Phys. 1975, 62, 3061.
(14) Bogan, D. J.; Setser, D. W. J. Chem. Phys. 1976, 64, 586.
(15) Donaldson, D. J.; Watson, D. G.; Sloan, J. J. Chem. Phys. 1982,
68, 95.
(16) Farrar, J. M.; Lee, Y. T. J. Chem. Phys. 1976, 65, 1414.
(17) Robinson, G. N.; Contininette, R. E.; Lee, Y. T. J. Chem. Phys.
1990, 92, 275.
(18) Milstein, R.; Williams, R. L.; Rowland, F. S. J. Phys. Chem. 1974,
78, 857.
(53) Blondel, C.; Cacciani, P.; Delsort, C.; Trainham, R. Phys. ReV. A
1989, 40, 3698.
(19) Smith, D. J.; Setser, D. W.; Kim, K. C.; Bogan, D. J. J. Phys. Chem.
1977, 81, 898.
(20) Nesbitt, F. L.; Monks, P. S.; Scanlon, M.; Stief, L. J. J. Phys. Chem.
1994, 98, 4307.
(54) Bullock, G. E.; Cooper, R. J. Chem. Soc., Faraday Trans. 1 1971,
67, 3258.
(21) Slagle, I. R.; Gutman, D. J. Phys. Chem. 1983, 87, 1818.
(22) Brunning, J.; Stief, L J. J. Chem. Phys. 1986, 84, 4371.
(23) Appelman, E. H.; Clyne, M. A. A. J. Chem. Soc., Faraday Trans.
1 1975, 71, 2072.
(24) Marston, G.; Nesbitt, F. L.; Nava, D. F.; Payne, W. A.; Stief, L. J.
J. Phys. Chem. 1989, 93, 5769.
(25) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin,
R. D.; Mallard, W. G. J. Phys. Chem. Ref. Data 1988, 17 (Suppl. 1).
(26) Lewis, R. S.; Sander, S. P.; Wagner, W.; Watson, R. T. J. Phys.
Chem. 1980, 84, 2009.
(27) Thorn, R. P.; Payne, W. A.; Stief, L. J.; Tardy, D. C. J. Phys. Chem
1996, 100, 13594.
(55) Schacke, H.; Wagner, H. G.; Wolfrum, J. Ber. Bunsen-Ges. Phys.
Chem. 1977, 81, 670.
(56) Lichtin, D. A.; Lin, M. C. Chem. Phys. 1985, 96, 473.
(57) Sayad, N.; Li, X.; Caballer, J. F.; Jackson, W. M. J. Photochem.
Photobiol., A 1988, 45, 177.
(58) North, S. W.; Fei, R.; Sears, T. J.; Hall, G. Int. J. Chem. Kinet.
1996, 29, 127.
(59) Lichtin, D. A.; Lin, M. C. Chem. Phys. 1986, 104, 325.
(60) Herbert, L.; Smith, I. W. M.; Spencer-Smith, R. D. Int. J. Chem.
Kinet. 1992, 24, 791.
(61) Sims, I. R.; Queffelec, J.-L.; Traver, S.; Rowe, B. R.; Herbert, L.
B.; Kartha¨user, J.; Smith, I. W. M. Chem. Phys. Lett. 1993, 211, 461.
(62) Yang, D. L.; Yu, T.; Wang, N. S.; Lin, M. C. Chem. Phys. 1992,
160, 317.
(28) Slagle, I. R.; Park, J.-Y.; Heaven, M. C.; Gutman, D. J. Am. Chem.
Soc. 1984, 160, 4356. Krueger, H.; Weitz, E. J. Chem. Phys. 1988, 88,
1608.
(29) Baulch, D. L.; Cobos, C. J.; Cox, R. A.; Frank, P.; Hayman, G.;
Just, Th.; Kerr, J. A,; Murrells, T.; Pilling, M. J.; Troe, J.; Walker, R. W.;
(63) Monks, P. S.; Romani, P. N.; Nesbitt, F. L.; Scanlon, M.; Stief, L.
J. J. Geophys. Res. 1993, 98, 17115.