Kinetic study of the mechanism of the low-temperature pyrolysis of vinyl bromide

Article abstract of DOI:10.1021/jp0031911

The pyrolysis of vinyl bromide has been examined in the temperature range 637-733 K and at pressures from 6 to 86 kPa. The yields of the major hydrocarbon products, C₂H₂, C₂H₄, and 1,3-C₄H₆, are second order in vinyl bromide over the entire range of temperatures investigated. At the higher temperatures, initiation by molecular elimination of HBr dominates, while at lower temperatures a free radical initiation channel becomes increasingly important. Our data for the overall process leading to HBr fit the relation ln(k) = (30.7 ± 4.8) - ((26.6 ± 3.3) x 10³)/T, with the rate constant in the units L mol^-1 s^-1, indicating an activation energy of 220 kJ mol^-1 ± 12% for the HBr elimination. A simple Arrhenius extrapolation is close to previous results at temperatures from 800 to over 2000 K. The combination of our data and the earlier measurements of the HBr elimination is reasonably represented by ln(k) = 37 - (3 × 10⁴)/T. Our data suggest that the free radical pathway is disproportionation rather than unimolecular cleavage of the C-Br bond, a situation analogous to that in the low-temperature thermal decomposition of ethylene. Kinetic analysis indicates that the activation energy of this new free radical initiation channel is approximately 150 kJ mol^-1, much less than the C-Br bond energy.

Full text of DOI:10.1021/jp0031911

1830
J. Phys. Chem. A 2001, 105, 1830-1837
Kinetic Study of the Mechanism of the Low-Temperature Pyrolysis of Vinyl Bromide
Patricia Ann Laws, Bradley D. Hayley, Lori M. Anthony, and John M. Roscoe*
Department of Chemistry, Acadia UniVersity, WolfVille, NoVa Scotia, Canada B0P 1X0
ReceiVed: September 5, 2000; In Final Form: December 15, 2000
The pyrolysis of vinyl bromide has been examined in the temperature range 637-733 K and at pressures
from 6 to 86 kPa. The yields of the major hydrocarbon products, C2H2, C2H4, and 1,3-C4H6, are second order
in vinyl bromide over the entire range of temperatures investigated. At the higher temperatures, initiation by
molecular elimination of HBr dominates, while at lower temperatures a free radical initiation channel becomes
increasingly important. Our data for the overall process leading to HBr fit the relation ln(k) ) (30.7 ( 4.8)
- ((26.6 ( 3.3) × 103)/T, with the rate constant in the units L mol-1 s-1, indicating an activation energy of
220 kJ mol-1 ( 12% for the HBr elimination. A simple Arrhenius extrapolation is close to previous results
at temperatures from 800 to over 2000 K. The combination of our data and the earlier measurements of the
HBr elimination is reasonably represented by ln(k) ) 37 - (3 × 104)/T. Our data suggest that the free radical
pathway is disproportionation rather than unimolecular cleavage of the C-Br bond, a situation analogous to
that in the low-temperature thermal decomposition of ethylene. Kinetic analysis indicates that the activation
energy of this new free radical initiation channel is approximately 150 kJ mol-1, much less than the C-Br
bond energy.
Introduction
Shilov and Sabirova1 studied the decomposition of C2H3Br
from approximately 900 to 1100 K at pressures of 1-7 kPa,
using the toluene jet technique. They found that the reaction
was kinetically second order and did not depend on the surface/
volume ratio. Their results were independent of whether toluene
or benzene was used in the jet. They observed ethylene as well
as acetylene in the decomposition of C2H3I but only acetylene
in the decomposition of C2H3Br and took this to indicate that
only the molecular decomposition of C2H3Br to C2H2 and HBr
was significant. They reported an activation energy of 274 kJ
mol-1 and a preexponential factor of 1.0 × 1017 L mol-1 s-1
for the molecular decomposition.
The vinyl radical figures prominently as a reactive intermedi-
ate in a wide range of chemical processes, including combustion
and petroleum refining operations such as cracking. It has also
been implicated in the formation of soot in flames and in the
deposition of carbon in industrial pyrolysis systems. Its chem-
istry is of fundamental importance too since it is the simplest
unsaturated free radical. Despite its importance, very few good
sources of the vinyl radical are available for kinetic or
mechanistic work. Those sources that have been examined
frequently are accompanied by parallel complicating reactions
that reduce the yield of vinyl radicals and introduce species that
interfere with the vinyl radical chemistry one would like to
study. Direct measurement of this radical by nondestructive
methods such as optical spectroscopy remains difficult due to
relatively low detection sensitivity and interference due to
absorption of light by the precursor or by other species that are
formed at the same time. The high chemical reactivity of the
vinyl radical also imposes a limit on the concentrations that
can be obtained from many sources.
In view of their importance in combustion and industrial
pyrolysis, it is desirable to identify a source of vinyl radicals
that could be used for kinetic and mechanistic experiments at
temperatures of a few hundred degrees Celsius. Preliminary
experiments in our laboratory to evaluate the range of temper-
atures over which vinyl bromide could be used as a photo-
chemical source of vinyl radicals indicated that a free radical
thermal decomposition process became important at tempera-
tures above approximately 350 °C. A survey of the literature
indicated that the only thermal decomposition process that has
been identified experimentally is the molecular elimination of
HBr.1-4 The experiments that have been reported were all done
at comparatively high temperatures.
Cadman and Engelbrecht2 studied the decomposition of C2H3-
Br in a reflected shock wave at temperatures from 1630 to 2100
K. They assumed that the decomposition was unimolecular and
found no evidence for any channel other than the one giving
acetylene and HBr. They reported an activation energy of 262
kJ mol-1 but, since this was a brief note, they did not provide
enough experimental information to permit calculation of a
bimolecular preexponential factor from their reported unimo-
lecular value of 9.5 × 1012 s-1
.
Lippiatt and Wells3 studied the reaction in a shock tube
coupled to a time-of-flight mass spectrometer at temperatures
from 1329 to 2128 K at pressures close to 20 kPa. They found
only C2H2 and HBr as products of the reaction and reported
that the rates of formation of these products mirrored the rate
of disappearance of C2H3Br. The activation energy was reported
to be 160 kJ mol-1 with a bimolecular preexponential factor of
1.0 × 1012 L mol-1 s-1
.
The most recent investigation of the decomposition of C2H3-
Br also used a reflected shock wave.4 The reaction was studied
from 1300 to 2000 K at pressures of a few hundred kilopascals.
They considered the possibility of a channel giving C2H3 + Br
but found no evidence for it. They concluded that the decom-
position of C2H3Br occurred entirely by the molecular channel
giving C2H2 and HBr and reported an activation energy of 174
* To whom correspondence should be addressed. Telephone: (902) 585-
1353. Fax: (902) 585-1114. E-mail john.roscoe@acadiau.ca.
10.1021/jp0031911 CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/02/2001

Low-Temperature Pyrolysis of Vinyl Bromide
J. Phys. Chem. A, Vol. 105, No. 10, 2001 1831
kJ mol-1 with a preexponential factor of 9.5 × 1010 L mol-1
s-1 for this reaction.
While the thermal decomposition of C2H3Br has been studied
from approximately 900 to over 2000 K, no evidence has been
reported for a decomposition channel other than the molecular
one leading to C2H2 and HBr. This contrasts with our observa-
tion of substantial yields of C2H4 and 1,3-C4H6 in the decom-
position at temperatures as low as 630 K, suggesting the
occurrence of a free radical decomposition channel that produces
C2H3 radicals. There is also some disagreement concerning the
activation energy for the molecular process. We have therefore
undertaken an experimental study of the lower temperature
decomposition of vinyl bromide below 800 K to measure the
relative importance of the molecular and free radical channels
and to clarify the kinetics and mechanism of the free radical
process. We report here a kinetic analysis of our data which
separates the molecular and free radical channels and provides
an empirical kinetic characterization of each.
Figure 1. Effect of varying the surface/volume ratio. Open symbols
are for the unpacked reaction vessel, S/V ) 1.0 cm-1, at a temperature
of 711 ( 2 K and a pressure of 16.1 ( 0.4 kPa of C2H3Br. The larger
closed symbols are for the packed reaction vessel, S/V ) 11 cm-1, at
a temperature of 704 ( 2 K and a pressure of 16.2 ( 0.6 kPa of C2H3-
Br. The smaller closed symbols are for the packed reaction vessel,
corrected to refer to the temperature of 711 K used in the unpacked
reaction vessel, using the experimental activation energies for each of
the reaction products. O, b, C2H2: 0, 9, C2H4: 4, 2, 1,3-C4H6.
Experimental Section
The pyrolysis experiments were made in Pyrex reaction
vessels that were heated electrically and were well-insulated.
The surface/volume ratios of these reactors ranged from 1.0 to
11 cm-1. The temperature was monitored continuously with
calibrated iron-constantan thermocouples that were inserted into
thermocouple wells extending into the center of the pyrolysis
vessel. The temperature remained constant during an experiment
to within the measurement precision which was better than
(0.5%. Pressures were measured with piezoelectric pressure
transducers whose calibration was checked daily. Kinetic
measurements made under identical conditions in different
reaction vessels were indistinguishable.
Experiments were all made in batch mode. A reaction mixture
of known composition was introduced to the reaction vessel at
a measured pressure. It was then allowed to react for a measured
length of time that was at least 10 times the settling time for
the pressure. At the end of the desired reaction time, a measured
pressure of the reaction mixture was transferred to a gas sample
loop and injected onto the chromatographic column. This
procedure prevented loss of products such as ethylene that do
not trap quantitatively with liquid nitrogen. When reactions
required the use of mixtures of vinyl bromide with other gases
such as argon, the gases were added to a reservoir at known
pressures and were allowed to mix for a sufficiently long time
that experiments using both longer and shorter mixing times
gave identical results.
the reaction products were determined initially by comparison
of their chromatographic retention times with those of authentic
samples, and these identifications were confirmed by obtaining
the mass spectra of the individual product peaks as they eluted
from the chromatographic column. The resolution of the
products was also checked by observing the mass spectra
obtained during successive scans of the mass spectrum as a
given peak eluted from the column.
Vinyl bromide was obtained from the Aldrich Chemical Co.
and had a stated purity of 98%. It contained methyl ethyl
hydroquinone as an inhibitor. It was purified before use by
freeze-pump-thaw cycles followed by bulb to bulb distillation
retaining only the middle part. Gas chromatographic analysis
of the purified material on the SP1700 column gave only a single
peak, while analysis on the Hayesep T column gave a small
additional peak, representing about 0.2% of the C2H3Br peak,
with a retention time slightly less than that of the C2H3Br peak.
This additional peak did not interfere with those of the reaction
products being measured and its area, relative to that of the
C2H3Br peak, did not change when the vinyl bromide was
pyrolized. It was concluded that the substance responsible for
the small additional peak when the Hayesep T column was used
for analysis did not affect either the chemical analysis or the
chemical behavior of the reaction under study.
The reaction mixtures were analyzed by gas chromatography
using either a Perkin-Elmer Sigma 3 or a Hewlett-Packard
HP5880 gas chromatograph. Both instruments used a flame
ionization detector, and the HP5880 was also interfaced to a
Ametek Dycor MA200M quadrupole mass spectrometer. Most
Results and Discussion
1
of the analyses used a 30 foot, /8 in. diameter column packed
Ethylene, acetylene, 1,3-butadiene, and vinyl acetylene were
the hydrocarbon reaction products obtained in the experiments.
Their yields were measured as a function of both the pressure
of vinyl bromide and reaction time at each temperature. The
extent of decomposition of vinyl bromide was at most 5% and
in most cases was much less than this. While HBr is presumably
formed in the reaction, we were not able to detect it or measure
its concentration with the available instrumentation. Representa-
tive plots of product yields as a function of reaction time are
shown in Figures 1 and 2. The standard deviations of the points
on such plots were typically of the order of 10-25%, depending
on the temperature, specific reaction product, and extent of
with SP1700, but a few analyses used columns packed with
Carbosieve G, Porapak Q, or Hayesep T to verify acceptable
resolution of the components to be analyzed in the reaction
mixtures. The chromatographs were calibrated daily with gas
mixtures containing known amounts of the desired analytes
diluted in helium. Typical chromatographic conditions with the
SP1700 column were 75 psi helium carrier, temperature program
50 °C for 20 min, ramp to 70 °C at 10 °C/min, and hold at 70
°C for 30 min. Typical retention times under these conditions
were as follows: ethylene, 6.5 min; acetylene, 8.7 min; 1,3-
butadiene, 20.7 min; vinyl bromide, 46 min. The identities of

1832 J. Phys. Chem. A, Vol. 105, No. 10, 2001
Laws et al.
on this chromatographic column from the very much larger peak
due to unreacted vinyl bromide at the small conversions of vinyl
bromide used in our experiments. Similar GC/MS experiments
with purified vinyl bromide gave no indication of the presence
of C2H4Br2, and the gas chromatogram gave no indication of
the unresolved peak in the tail of the vinyl bromide peak.
While the incomplete resolution of the chromatographic peak
tentatively assigned to C2H4Br2 prevented its quantitation and
limited the detail that could be obtained in its mass spectrum,
a few analyses were made of mixtures of vinyl bromide with
1,2-dibromoethane to assess the likelihood that these results
could reasonably be attributed to an isomer of C2H4Br2. Gas
chromatograms of mixtures of vinyl bromide and 1,2-dibromo-
ethane also showed an incompletely resolved peak in the tail
of the peak due to vinyl bromide. This unresolved peak had
the same retention time as the unresolved peak, tentatively
attributed to dibromoethane, in the reaction products. Although
the poor resolution of this peak from that due to vinyl bromide
prevented its use for analytical purposes, the combination of
the gas chromatographic and mass spectrometric measurements
is consistent with the formation of 1,2-dibromoethane as a
product of the thermal decomposition of vinyl bromide under
our conditions. These limited experiments do not rule out the
possibility that 1,1-dibromoethane is formed as well, but the
limited resolution of the peak attributed to C2H4Br2 from the
tail of the peak for vinyl bromide suggested that a more complete
series of experiments was unlikely to provide additional
significant information. Vinyl acetylene was a significant
reaction product only at the higher temperatures. It was identified
by its mass spectrum, but we were unable to obtain a sample
of the authentic material to use for calibration of the gas
chromatograph.
Figure 1 shows yields of products as a function of reaction
time obtained at approximately the same temperature and
pressure in reaction vessels of widely different surface/volume
ratios. The open plotting symbols, representing the results for
the reaction vessel with the small S/V ratio, lie consistently above
the filled plotting symbols, representing the results obtained in
the reaction vessel with the larger S/V ratio. However, the results
represented by the open plotting symbols were also obtained at
a temperature 7 ( 4 °C higher than that at which the data for
the filled plotting symbols were obtained. When the measured
variation of yields with temperature is considered, the small
increase in yield in the reaction vessel having the small S/V
ratio is entirely accounted for by the slightly higher reaction
temperature used with that reaction vessel. This is indicated by
the smaller filled plotting symbols in Figure 1. These represent
product yields that were corrected for the small variation in
temperature using the activation energy for each product
obtained from the data presented in Figures 6-8. The results
obtained with reaction vessels of different S/V ratios suggest
that surface effects are negligible in our experiments.
Figure 2. (a) Representative plots of the dependence of product yields
on reaction time at a total pressure of 16 kPa and (a, top) 636, (b,
middle) 711, and (c, bottom) 734 K: b, C2H2; 9, C2H4; 2, 1,3-C4H6.
Figure 2 shows representative plots of the variation of product
yields with time in the upper, lower, and middle parts of the
temperature range examined. The yield of acetylene is clearly
the largest over the entire temperature range. However, its yield
becomes proportionately smaller as the temperature decreases.
In contrast, the ratio of yields of the more minor products,
ethylene and 1,3-butadiene, shows much less variation with
temperature. This suggests that the kinetically limiting processes
leading to acetylene are different from those leading to ethylene
and butadiene and that the kinetically limiting processes leading
to the latter two products are similar. The plots of which those
in Figure 2 are representative were all linear over the range of
reaction. When replicate experiments were made, their standard
deviations were also of this magnitude.
Chromatograms of the mixture of reaction products usually
showed a small, unresolved peak with a retention time ap-
proximately 4 min longer than that of vinyl bromide. Sequential
scans with the mass spectrometer as this unresolved peak eluted
from the SP1700 column gave a series of peaks at masses of
186, 188, and 190 with the intensity ratio to be expected of a
dibrominated compound. The masses and intensity distribution
of these peaks in the mass spectrum indicated that C2H4Br2 was
probably present in the reaction products but was not resolved

Low-Temperature Pyrolysis of Vinyl Bromide
J. Phys. Chem. A, Vol. 105, No. 10, 2001 1833
Figure 3. Representative plots to verify second-order dependence on
[C2H3Br]. The temperature for these experiments was 732 K: b,
C2H2: 9, C2H4: 2, 1,3-C4H6.
Figure 4. Effect of replacing C2H3Br with Ar as a bath gas. The filled
symbols are for experiments with 16 kPa of C2H3Br and varying
pressures of added Ar. The open symbols are for experiments with
pure C2H3Br. The temperature for these experiments was 732 K. O,
b, C2H2; 0, 9, C2H4; 4, 2, 1,3-C4H6.
temperatures, pressures, and reaction times investigated, indicat-
ing that secondary reactions involving the measured reaction
products were either negligible or had a constant stoichiometric
effect on the measured yields. We conclude from this behavior
that, at the small conversions of C2H3Br in our experiments,
the slopes of such plots give reasonable estimates of the initial
rates of formation of the reaction products.
only thermal decomposition channel that has been reported
experimentally is
C2H3Br f C2H2 + HBr
(1)
Reaction rates were measured as a function of initial
concentration of vinyl bromide over the full range of temper-
atures studied. The concentration of vinyl bromide was varied
by roughly an order of magnitude in these experiments. Plots
of reaction rate against the square of the vinyl bromide
concentration were accurately linear for each of the measured
reaction products. The formation of each of these products is
therefore second order in vinyl bromide, and empirical second-
order rate constants can be calculated by dividing the initial
rate by the square of the vinyl bromide concentration. A
representative set of such plots is shown in Figure 3. The
precision of the individual data points in such plots was of the
order of 10-25%, depending on reaction conditions. The
precision of the slopes of plots such as those in Figure 3 was
also of this magnitude. Experiments were also made in which
the pressure of vinyl bromide was kept constant while the total
pressure was increased by adding argon. If the second-order
behavior is due to reactions requiring energy transfer from the
bath gas, the empirical second-order rate constants will increase
as the argon pressure is increased. However, if the second-order
kinetic behavior is due to a chemical interaction between two
vinyl bromide molecules rather than energy transfer, the second-
order rate constants should be independent of the pressure of
added argon. The data in Figure 4 indicate that the rate constants
based on the yields of C2H4 and 1,3-C4H6 are independent of
the pressure of added argon while the rate constants based on
the yield of C2H2 depend strongly on the pressure of added
argon. These results indicate that some reactions that have
kinetic control over the yield of C2H2 depend on energy transfer
while the reactions leading to C2H4 and 1,3-C4H6 result from a
chemical interaction between two molecules of C2H3Br. The
second-order behavior of the rate of production of C2H2 is
consistent with the results of three of the literature reports1,3,4
in which the effect of pressure on the measured rate constant
for formation of C2H2 was examined, and the reaction was found
to be second order under their conditions.
The observation of large yields of C2H2 in our work is consistent
with this being an important decomposition process.
The reaction
C2H3Br f C2HBr + H2
(2)
has been reported in matrix isolation work.5,6 A careful search
for H2 as a product in our experiments, using a gas chromato-
graph equipped with a 5A molecular sieve column and a thermal
conductivity detector, was negative. Our detection limits
indicated that we should have been able to detect H2 at a
concentration of at least 0.8 µmol/L. We conclude on this basis
that channel 2 does not occur significantly under our conditions.
This also indicates that the channel
C2H3Br f C2H2 + H + Br
(3)
does not take place appreciably since the hydrogen atom
produced in this reaction would almost certainly lead to
formation of H2, either as a result of hydrogen abstraction from
vinyl bromide or by recombination. These results are consistent
with the conclusion of a theoretical analysis of decomposition
of vinyl bromide on the ground-state potential energy surface7
in which HBr elimination was found to have a potential barrier
approximately 100 kJ mol-1 lower than H2 elimination and even
farther below reaction 3.
The observation of significant quantities of ethylene, 1,3-
butadiene, and vinyl acetylene as reaction products leads to the
conclusion that a free radical initiation process that forms vinyl
radicals must be occurring. The initiation reaction
C2H3Br f C2H3 + Br
(4)
We must first evaluate the possible initiation steps in the
decomposition of vinyl bromide. It was noted earlier that the
would provide the necessary free radicals for a chain process
leading to the observed products. This would then be followed

1834 J. Phys. Chem. A, Vol. 105, No. 10, 2001
Laws et al.
5. The temperature dependence of the relative rates of formation
of C2H4 and 1,3-C4H6 seems to be qualitatively consistent with
reactions 9 and 10 as the main source of 1,3-C4H6.
We next consider the alternative of reaction 11 as the
dominant source of 1,3-C4H6. In that case, the ratio of rates of
formation of C2H4 and 1,3-C4H6 would be given by
ν
k5[C2H3Br]
k11[C2H3]
C2H4
)
ν
1,3-C4H6
In this case, the activation energy obtained from the relative
rate measurements would reflect the difference Ea(5) - Ea(11)
- Ea(C2H3). Reaction 11 is a simple recombination reaction
and is expected to have no activation energy. The activation
energy for production of C2H3, Ea(C2H3), is expected to be
comparable to that for the initiation reaction producing C2H3
and should be substantially larger than that for reaction 5. If
reaction 11 were the main source of 1,3-C4H6, the ratio of rates
for C2H4 and 1,3-C4H6 should either be nearly independent of
temperature or should decrease with increasing temperature. We
conclude that the temperature dependence of the relative rates
of formation of C2H4 and 1,3-C4H6 is more consistent with the
principal formation of 1,3-C4H6 through reactions 9 and 10 than
with its formation in reaction 11.
Figure 5. Temperature dependence of the ratio of the rate of formation
of C2H4 to that of 1,3-C4H6.
by the reactions below, leading to C2H4 and additional C2H2
C2H3 + C2H3Br f C2H4 + C2H2Br
Br + C2H3Br f HBr + C2H2Br
C2H2Br f C2H2 + Br
(5)
(6)
(7)
The basic reaction scheme above suggests that decomposition
of the C2H2Br radical provides a source of C2H2 in addition to
that arising from reaction 1. There are two sources of this radical,
reactions 5 and 6. Assuming that decomposition of C2H2Br in
reaction 7 is rapid and that reaction 5 is the only significant
source of C2H4, the yield of C2H4 provides an estimate of the
additional C2H2 produced via (5) and (7). There is no compa-
rable way of estimating the extra C2H2 produced via (6) and
(7) since no unique products result. However, examination of
the literature on reactions of olefins with halogen atoms leads
one to the conclusion that the addition process, reaction 8, is
likely to dominate the abstraction process, reaction 6, by at least
an order of magnitude (see, for example, refs 9-11). With this
assumption, one may estimate the yield of C2H2 from reaction
1 by subtracting the yield of C2H4 from the measured total yield
of C2H2. The plots of net C2H2, calculated in this way, against
reaction time were linear, and their slopes were divided by the
square of the C2H3Br concentration to obtain estimates of the
second-order rate constant for reaction 1.
Figure 6 presents our kinetic data for reaction 1 in Arrhenius
format and provides an inset for comparison with the earlier
studies in which second-order rate constants were reported. Our
results yield a preexponential factor of 2.5 × 1013 L mol-1 s-1
and an activation energy of 220 kJ mol-1 (13%. This degree
of uncertainty in the slope of Figure 6 gives a large extrapolation
uncertainty in determining the preexponential factor and is a
result of the limited temperature range accessible in our
experiments. This makes our estimate above of the preexpo-
nential factor quite approximate. Our experimental values may
as well as the following reactions giving 1,3-C4H6
Br + C2H3Br f C2H3Br2
(8)
(9)
C2H3 + C2H3Br f C4H6Br
C4H6Br f 1,3-C4H6 + Br
2 C2H3 f 1,3-C4H6
(10)
(11)
The relative importance of reactions 10 and 11 can be inferred
from the temperature dependence of the ratio of rates of
formation of C2H4 and 1,3-C4H6. The relevant data are shown
in Figure 5. The data are more scattered than those for the
Arrhenius plots of Figures 6-8 because the latter depend on
the measurement of the yield of only one reaction product while
the data of Figure 5 require the yields of two reaction products
of comparable magnitude. However, it is clear that the ratio of
rates, νC2H4/ν1,3-C4H6, increases with increasing temperature. The
slope of the line in Figure 5 leads to an energy of 46 kJ mol-1
( 31%, which represents the difference in activation energy of
the reaction paths leading to C2H4 and 1,3-C4H6.
We first examine the formation of 1,3-C4H6 via reactions 9
and 10. If it is assumed that reaction 5 determines the proportion
of C2H3 that produces C2H4 while reaction 9 determines the
proportion of the channel leading to 1,3-C4H6, the ratio of rates
νC2H4/ν1,3-C4H6 should be approximately equal to the ratio k5/
k9. Since reaction 5 is a hydrogen abstraction while reaction 9
is a free radical addition, reaction 5 should have the larger
activation energy and the ratio of rates should increase with
increasing temperature as observed. The slope of the plot in
Figure 5 would then represent the difference in activation energy
of reactions 5 and 9. While kinetic data for reactions of C2H3
are scarce, the difference in activation energy for the comparable
abstraction and addition reactions of C2H3 with C3H6 has been
estimated8 to be approximately 26 kJ mol-1, which is similar
to the value estimated from the slope of Figure 5. As an
alternative comparison, the difference in activation energies for
the addition and abstraction reactions of C2H5 with C2H4 is9 40
kJ mol-1, which is also similar to the value derived from Figure
be compared with those of Shilov and Sabirova1 (1.0 × 1017
L
mol-1 s-1 and 274 kJ mol-1), Lippiatt and Wells3 (1.0 × 1012
L mol-1 s-1 and 163 kJ mol-1), and of Saito et al.4 (9.5 × 1013
L mol-1 s-1 and 170 kJ mol-1) obtained at much higher
temperatures. As shown in the inset of Figure 6, our data and
those of the three earlier studies cluster about a straight line
which gives a preexponential factor of 8.2 × 1015 L mol-1 s-1
and an activation energy of 250 kJ mol-1. While it is clear that
these data could also be accommodated by a line with gentle
upward curvature, the quality of the data does not justify more
than a linear representation. The agreement between our results

Low-Temperature Pyrolysis of Vinyl Bromide
J. Phys. Chem. A, Vol. 105, No. 10, 2001 1835
although experiments with vinyl fluoride13 observed infrared
emission from vibrationally excited acetylene. They found that
this emission was consistent with the approximately 170 kJ
mol-1 of the exothermicity of reaction 13, which they took as
evidence that the initial decomposition gave vinylidene that
rapidly isomerized to acetylene. Recent IRMPD experiments
on vinyl bromide14 found only acetylene, HBr, and a noncon-
densable gas, which was assumed to be hydrogen, under their
normal irradiation conditions. When the laser was tightly
focused, diacetylene and a black soot deposit were found but
these products were absent in experiments with a long focal
length lens. Both the thermal activation experiments1-4 and the
IRMPD work12-14 access higher energies than our experiments,
and one might expect them to produce larger yields of vinylidene
than would be obtained in our work. The absence of stable
reaction products that could be attributed to vinylidene in the
higher energy experiments suggest that vinylidene isomerizes
so rapidly that there was insufficient time to produce stable
products in bimolecular reactions.
Vinylidene is thought to be formed in its singlet state and
would be expected to undergo rapid addition to compounds such
as vinyl bromide. The adduct might then decompose giving HBr
and C4H4, vinyl acetylene, which was observed in our experi-
ments but for which quantitative measurement was not possible.
It would not, however, be expected that vinylidene would
produce either ethylene or 1,3-butadiene in the quantities that
were measured. It should be noted that the chromatographic
peaks for vinyl acetylene were no larger than those for 1,3-
butadiene, the product formed in the smallest amount, and were
usually much smaller. The main effect of such reactions would
be to make the extent of the HBr elimination reaction estimated
from the acetylene yield too small. This, in turn, would make
our estimates of the rate constants for that reaction too small.
The observation that our kinetic data for the HBr elimination
extrapolate well to those obtained at the much higher temper-
atures of the shock tube and toluene jet experiments, as indicated
in Figure 6, suggests that our assumption that vinylidene
isomerizes to acetylene so rapidly that it does not participate
appreciably in other reactions in the system is not seriously in
error. The formation of vinyl acetylene in our system can equally
well be accounted for by the reaction of C2H3 with acetylene,
as discussed later.
Figure 6. Temperature dependence of the rate constants estimated for
reaction 1. The inset compares our results with the kinetic data reported
from earlier work done at much higher temperatures. The dashed line
provides a composite representation of our work and that of the three
earlier studies.
and the earlier work is good, particularly considering the lengthy
extrapolation involved, the limited temperature range accessed
by our experiments, the extent of disagreement between the two
shock tube studies, the variety of experimental methods used,
and our indirect method of estimating the rate constant for
reaction 1.
While the principal objective of this work is the separation
and experimental characterization of the free radical component
of the thermal decomposition of vinyl bromide at comparatively
low temperatures, the nature of the HBr elimination path requires
some comment. Although this has traditionally been regarded
as a four-center elimination, there is now a significant body of
work indicating that the three-center elimination of HBr is a
lower energy reaction path. This includes both experimental
work12-14 using infrared multiphoton dissociation (IRMPD) of
vinyl halides and theoretical calculations on an analytic potential
energy surface for vinyl bromide.7,15 While IRMPD accesses
higher energies than conventional pyrolysis experiments such
as ours, it presumably still occurs on the ground-state potential
surface, and many of the conclusions reached from these studies
should still be relevant to thermal activation experiments.
While a four-center decomposition process would be com-
pletely molecular, producing HBr and acetylene, a three-center
decomposition would yield HBr and vinylidene. The isomer-
ization of vinylidene to acetylene has a low energy barrier
estimated to be in the range7 of 19.7-2.3 kJ mol-1, indicating
that it is likely to be rapid relative to the three-center process.
The overall decomposition process leading to HBr and acetylene
would then comprise the reactions
We now turn our attention to the free radical component of
the reaction. If reaction 4 is kinetically limiting for production
of C2H3, the subsequent reactions consuming C2H3 should yield
products whose kinetic behavior contains information about
reaction 4. We first consider C2H4. This is produced in the
reaction sequence
C2H3Br f C2H3 + Br
(4)
(5)
C2H3 + C2H3Br f C2H4 + C2H2Br
C2H3Br f CH2C: + HBr
CH2C: f CHCH
(12)
(13)
The formation of C2H4 was found to be second order in C2H3-
Br, so the initial rate of formation of C2H4 divided by the square
of the concentration of C2H3Br will give an effective rate
constant which, when plotted in Arrhenius form, will yield
information about the kinetically limiting step. Such a plot is
found in Figure 7. While the intercept of such a plot has no
simple interpretation since it does not necessarily describe a
single elementary reaction, the activation energy should reflect
that of the kinetically limiting step. The activation energy of
reaction 4, as written, should be comparable to the C-Br bond
in which our experiments indicate that reaction 12 would be in
its second-order regime. The possible intervention in our
experiments of reactions initiated by vinylidene needs to be
examined. The key question is whether the isomerization
reaction is sufficiently rapid that reaction 13 dominates over
alternative loss routes for vinylidene.
The shock tube2-4 and toluene jet1 experiments did not report
any products other than HBr and acetylene. The IRMPD
experiments on vinyl halides also were unable to detect products
other than acetylene that could have resulted from vinylidene,
energy, which has been estimated16 to be about 320 kJ mol-1
,
.
while that of reaction 5 should be roughly9 20-30 kJ mol-1

1836 J. Phys. Chem. A, Vol. 105, No. 10, 2001
Laws et al.
Figure 8. Temperature dependence of the empirical second-order rate
constants based on the rate of formation of 1,3-C4H6.
Figure 7. Temperature dependence of the empirical second order rate
constants based on the rate of formation of C2H4.
Reaction 4 would therefore be kinetically limiting for C2H4,
and the activation energy for production of C2H4 would be
roughly 300 kJ mol-1 if reaction 4 were a unimolecular C-Br
bond scission. The slope of the plot in Figure 7 gives an
activation energy of 150 kJ mol-1 ( 14%. This, together with
the measured reaction order of 2, suggests that the source of
C2H3 is not the unimolecular dissociation of C2H3Br as presented
in reaction 4.
initiation is dominated by direct C-H bond cleavage with its
much larger activation energy and larger preexponential factor.
This view is consistent with recent detailed kinetic modeling
of ethylene pyrolysis in the lower temperature regime.21,22
The unexpectedly small activation energy for initiation of the
free radical component of the pyrolysis of vinyl bromide is
consistent with a similar explanation. Reaction 4 is stoichio-
metrically equivalent to the following sequence of reactions.
A similar conclusion can be reached by examining the yields
of 1,3-C4H6. The production of 1,3-C4H6 was found to be second
order in C2H3Br, so an empirical second-order rate constant can
be calculated by dividing the measured initial rate of formation
of 1,3-C4H6 by the square of the concentration of C2H3Br. Figure
8 shows the temperature dependence, in Arrhenius format, of
these empirical rate constants based on the formation of 1,3-
C4H6. The slope of the line in this plot gives an activation energy
of 136 kJ mol-1 ( 15%, a value that is comparable with that
obtained from the rates of formation of C2H4 and again much
smaller than the C-Br bond energy in C2H3Br. The minimum
set of chemical reactions required to produce 1,3-C4H6 in this
system is similar to that for production of C2H4, namely,
2C2H3Br f C2H3 + C2H3Br2
C2H3Br2 f C2H3Br + Br
(14)
(15)
As with reaction 10, the activation energy for reaction 15 is
anticipated9 to be in the range of 20-30 kJ mol-1 and reaction
14 would be responsible for the larger value of about 140-150
kJ mol-1 which we measure for the production of 1,3-C4H6 and
C2H4. Reaction 14 would then be the rate-limiting step and
would result in the second-order kinetics observed for production
of C2H4 and 1,3-C4H6. Since reaction 14 requires a chemical
interaction between two C2H3Br molecules, its rate constant
should not depend appreciably on energy transfer from the bath
gas. This is consistent with our observation that the rate
constants calculated from the rates of formation of C2H4 and
1,3-C4H6 do not depend on whether the bath gas is argon or
C2H3Br. The tentative identification of C2H4Br2 by its mass
spectrum in the effluent from the gas chromatograph, referred
to at the beginning of this section, would also be consistent
with the occurrence of reactions 8 and 14.
C2H3Br f C2H3 + Br
C2H3 + C2H3Br f C4H6Br
C4H6Br f 1,3-C4H6 + Br
(4)
(9)
(10)
The activation energy of reaction 9 is expected9 to be of the
order of 10-20 kJ mol-1, while that for reaction 10 is expected9
to be of the order of 20-30 kJ mol-1. Again, the activation
energy for the production of 1,3-C4H6 should be dominated by
that of reaction 4 and would be roughly 300 kJ mol-1 if C2H3
radicals were to result from unimolecular dissociation of C2H3-
Br. Clearly a lower energy pathway must be responsible for
production of the vinyl radical in this system.
Br + C2H3Br f C2H3Br2
(8)
2C2H3Br f C2H3 + C2H3Br2
(14)
The C2H3Br2 radical formed in these reactions is produced either
directly or indirectly from products of this free radical decom-
position process. It would be expected to abstract a hydrogen
atom from C2H3Br or some other hydrogen source in the reaction
mixture producing C2H4Br2. A similar hydrogen abstraction
reaction involving the vinyl radical, reaction 5, is proposed as
the main source of ethylene.
A similar situation exists in the pyrolysis of ethylene for
which the initiation step has been found to be second order with
a measured activation energy of approximately 240 kJ mol-1 17,18
,
much less than the C-H bond energy in C2H4. It has been
suggested19,20 that the pyrolysis of ethylene occurs by a different
mechanism at the high temperatures of shock tube experiments
than at the lower temperatures of conventional pyrolysis
experiments and that, below roughly 1000 K, initiation is
strongly dominated by disproportionation to ethyl and vinyl
radicals. Under shock tube conditions, on the other hand,
C2H3 + C2H3Br f C2H4 + C2H2Br
(5)
The presence of vinyl acetylene in the reaction products,
particularly at the higher temperatures, is consistent with the

Low-Temperature Pyrolysis of Vinyl Bromide
reaction sequence
J. Phys. Chem. A, Vol. 105, No. 10, 2001 1837
from disporportionation of vinyl bromide with an activation
energy of approximately 150 kJ mol-1, much less than the C-Br
energy in vinyl bromide, and reminiscent of the situation that
has been shown to occur in the pyrolysis of ethylene at
temperatures below 1000 K.
C2H3 + C2H2 f C4H5
(16)
(17)
C4H5 f C4H4 + H
Acknowledgment. The authors are indebted to the Imperial
Oil Charitable Foundation for major funding of this project.
Additional financial support was obtained from the Natural
Sciences and Engineering Research Council of Canada and from
Acadia University.
This would be most important at the higher temperatures where
the yield of C2H2 is proportionately the largest and the
temperature would be large enough to compensate for the
expected comparatively large activation energy of reaction 17.
Vinyl acetylene is thus regarded as a secondary product resulting
from a reaction involving C2H2, the stable primary reaction
product formed in the largest concentration.
References and Notes
(1) Shilov, A. E.; Sabirova, R. D. Kinet. Catal. 1964, 5, 32.
(2) Cadman, P.; Engelbrecht, W. J. Chem. Commun. 1970, 453.
(3) Lippiatt, J. H.; Wells, D. E. Dyn. Mass Spectrom. 1976, 4, 273.
(4) Saito, K.; Yokubo, T.; Fuse, T.; Tahara, H.; Kondo, O.; Higashihara,
T.; Murakami I. Bull. Chem. Soc. Jpn. 1979, 52, 3507.
(5) Abrash, S. A.; Carr, C. M.; McMahon, M. T.; Zehner, R. W. J.
Phys. Chem. 1994, 98, 11909.
(6) Agrawal, P. M.; Sorescu, D. C.; Raff, L. M.; Abrash, S. A. J. Phys.
Chem. 1995, 99, 14959.
(7) Abrash, S. A.; Zehner, R. W.; Mains, G. J.; Raff, L. M. J. Phys.
Chem. 1995, 99, 2959.
(8) Tsang, W. J. Phys. Chem. Ref. Data 1991, 20, 221.
(9) Westley, F.; Herron, J. T.; Frizzell, D.; Hampson, R. F.; Mallard,
W. J. NIST Standard Reference Database 17-2Q98; National Institute of
Standards and Technology: Washington, DC, 1998.
(10) Barnes, I.; Bastian, V.; Becker, K. H.; Overath, R.; Tong, Z. Int. J.
Chem. Kinet. 1989, 21, 499.
(11) Yarwood, G.; Peng, N.; Niki, H. Int. J. Chem. Kinet. 1992, 24,
369.
(12) Reiser, C.; Lussier, F. M.; Jensen, C. C.; Steinfeld, J. I. J. Am.
Chem. Soc. 1979, 101, 350.
(13) Caballero, J. F.; Wittig, C. J. Chem. Phys. 1985, 82, 1332.
(14) Samoudi, B.; D´ıaz, L.; Oujja, M.; Santos, M. J. Photochem.
Photobiol. A: Chem. 1999, 125, 1.
(15) Kay, R. D.; Raff, L. M. J. Phys. Chem. A 1997, 101, 1007.
(16) Wodtke, A. M.; Hintsa, E. J.; Somorjai, J.; Lee, Y. T. Isr. J. Chem.
1989, 29, 383.
(17) Ayranci, G.; Back, M. H. Int. J. Chem. Kinet. 1981, 13, 897.
(18) Ayranci, G.; Back, M. H. Int. J. Chem. Kinet. 1983, 15, 83.
(19) Benson, S. W.; Haugen, G. R. J. Phys. Chem. 1967, 71, 1735.
(20) Boyd, M. L.; Wu, T.-M.; Back, M. H. Can. J. Chem. 1968, 46,
2415.
(21) Roscoe, J. M.; Jayaweera, I. S.; MacKenzie, A. L.; Pacey, P. D.
Int. J. Chem. Kinet. 1996, 28, 181.
Summary
We have extended experimental kinetic measurements of the
thermal decomposition of vinyl bromide to much lower tem-
peratures than have previously been examined. As the temper-
ature decreases, the extent of decomposition via the molecular
elimination of HBr also decreases, making it possible to observe
a new, second-order, free radical channel leading to bromine
atoms and vinyl radicals. We are unaware of any previous
characterization of this free radical channel for the decomposi-
tion of vinyl bromide. We have separated the kinetic behavior
of the molecular elimination of HBr from this free radical
pathway and report estimates of the rate constants for the
molecular channel that extrapolate well to the earlier, high-
temperature results obtained by quite different experimental
methods. Our own experiments over a limited range of low
temperatures lead to an activation energy of 220 kJ mol-1 for
the molecular channel, while a linear Arrhenius representation
of both our data and the higher temperature results leads to an
activation energy of 250 kJ mol-1. The molecular channel was
found to be second order, and its rate constant depended on
whether the bath gas was vinyl bromide or argon. We therefore
conclude that the molecular channel depends strongly on energy
transfer from the bath gas. On the other hand, the rate constants
for the formation of products characteristic of the free radical
path are second order and independent of the nature of the bath
gas. We propose that the generation of vinyl radicals results
(22) Roscoe, J. M.; Bossard, A. R.; Back, M. H. Can. J. Chem., 2000,
78, 16.