1,2-Pentadiene decomposition

Article abstract of DOI:10.1002/kin.1072

1,2-Pentadiene was decomposed in single pulse shock tube experiments. There appear to be a large number of parallel decomposition and isomerization channels. It was shown that the resonance energy of the 1,3-butadiene-2-yl radical is smaller than that of allyl radical by an amount equal to the π bond conjugation energy of butadiene.

Full text of DOI:10.1002/kin.1072

1,2-Pentadiene
Decomposition
JUERGEN HERZLER, JEFFREY A. MANION, WING TSANG
National Institute of Standards and Technology, Gaithersburg, Maryland 20899
Received 16 May 2001; accepted 18 June 2001
ABSTRACT: 1,2-Pentadiene has been decomposed in single pulse shock tube experiments.
There appear to be a large number of parallel decomposition and isomerization channels. The
following rate expressions for isomerization in the temperature range 1100–1250 K and pres-
sures between 200 and 300 kPa have been obtained
1
k(1,2-pentadiene → 1-pentyne) = 1.0 1013 exp( 32300 900 K/T ) s
k(1,2-pentadiene → cis/trans-1,3-pentadiene = 2.2 1014 exp( 33900 900 K/T ) s
k(1,2-pentadiene → cyclopentene) = 1.8 1011 exp( 26080 900 K/T )s
1
1
Overall, we estimate the 2 uncertainty in the absolute rate constants to be about 30%. From
detailed balance the rate expression for 1,3 to 1,2-pentadiene isomerization has been found to
be
1
k(1,3-pentadiene → 1,2-pentadiene) = 1.2 1015 exp( 42070 900 K/T ) s
The significance of the 1,3 to 1,2 isomerization process as a bridge between the two-carbon and
three-carbon species in high temperature hydrocarbon decomposition systems is discussed.
The decomposition processes involve bond breaking and retroene reactions. The rate expres-
sion for the latter has been found to be
1
k(1,2-pentadiene → propyne + ethylene) = 6.6 1012 exp( 29240 900 K/T ) s
The rate constants for fission of the C3 C4 and C4 C5 bonds appear to be close to each other.
Our results indicate that the resonance energy of the 1,3-butadiene-2-yl radical is smaller than
C
that of allyl radical by an amount equal to the bond conjugation energy of butadiene. 2001
John Wiley & Sons, Inc. Int J Chem Kinet 33: 755–767, 2001
channels becoming important as the degree of unsatu-
ration is increased. For example, when one compares
1-pentene with n-pentane, it is found that the former
species undergoes much faster unimolecular decompo-
sition. Not only does allylic resonance energy weaken
the 3,4 C C bond, but 1-pentene can also decompose
via a retroene reaction [1]. In a similar fashion, uni-
molecular stability decreases rapidly in the series cy-
clohexane, cyclohexene [1], and 1,4-cyclohexadiene
[2], where the respective initial products are 1-hexene,
butadiene and ethylene, and benzene. The common
INTRODUCTION
This paper is concerned with the decomposition
and isomerization of 1,2-pentadiene. Interest in this
molecule arises from the possibility of new reaction
Correspondence to: Jeffrey A. Manion; e-mail: jeffrey.
manion@nist.gov.
c
2001 John Wiley & Sons, Inc. This article is a US Government
work and, as such, is in the public domain in the United States of
America.

756
HERZLER, MANION, AND TSANG
characteristic in these rate progressions is the increas-
ing importance of molecular channels involving the
direct elimination of stable molecules.
assessment of the role of each of the radicals in high
temperature situations.
The starting point for the prediction of the rate con-
stants for these processes under all conditions is the
determination of the unimolecular rate expressions at
the high-pressure limit. We chose to study the 1,2-
pentadiene instead of 1,2-butadiene directly because
the latter compound is no longer commercially avail-
able short of custom synthesis. As will be seen subse-
quently, this choice leads to some interesting and un-
expected new information. The use of 1,2-pentadiene
does, however, make the data analysis more complex
since competitive C C bond fission becomes a more
important process.
Despite their importance, there has been little work
on the isomerization reactions of larger polyunsatu-
rated compounds. The interconversion of allene and
propyne is a well-studied [7] reaction, however. It has
been suggested that the reaction proceeds by way of
cyclopropene and involves a carbene intermediate [8].
Hidaka et al. [9] have recently used the shock tube
to study the decomposition of 1,2-butadiene at tem-
peratures between 1100 and 1600 K and pressures of
120–230 kPa. They analyzed the results on the basis
of a 82-step reaction mechanism. Because of the com-
plexity of the reaction, the authors were forced to fit
their data on the basis of assumed activation energies
for the main two processes. On this basis they obtained
the following rate expressions:
In the case of homolytic bond breaking, knowledge
of the thermodynamics of the process together with
analogies regarding radical combination rate constants
leads to a certain degree of predictability [3]. By con-
trast, the situation is much less certain for molecular
reactions where there are simultaneous bond breaking
and bond forming processes. A particularly interesting
class of reactions is isomerization of highly unsatu-
rated molecules. Such species tend to be formed under
severe pyrolytic conditions, for example, in fuel-rich
combustion regions of flames. Reactions of such com-
pounds are of particular importance immediately prior
and subsequent to the formation of aromatics and soot.
Our interest in such processes is the main rationale for
the present study.
1,2-Pentadiene can be formed in high temperature
systems from the combination of methyl and 1,3-
butadienyl radicals or ethyl and propargyl radicals.
Both propargyl and methyl radicals are extremely sta-
ble in the unimolecular sense, and hence tend to adopt
relatively high concentrations and undergo a variety
of such combination processes. We are especially in-
terested in the rate expression for transformation of
the 1,2-diene structure to the 1,3-diene and the reverse.
The prototypical example of this reaction is the conver-
sion of 1,2-butadiene to 1,3-butadiene. Although the
present experiments directly concern the five carbon
compound, it is expected that isomerization rate for
the diene structure should be similar to that of anal-
ogous hydrocarbons. Hence, with the present results
and detailed balance it should be possible to estimate
a rate expression for 1,3-butadiene → 1,2-butadiene, a
process difficult to study directly but likely to be of con-
siderable importance in a variety of high temperature
combustion systems.
k(1,2-butadiene → 1,3-butadiene)
1
= 2.5 1013 exp( 31702 K/T ) s
k(1,2-butadiene → propargyl + methyl)
1
= 2.0 1015 exp( 37741 K/T ) s
The authors warn the reader that the results are
valid only under their specific conditions and should
not be extrapolated or imprudently applied to other
conditions.
Vinyl and propargyl radicals are important species
in many high temperature organic systems. Mecha-
nisms for their reaction to form benzene and higher
aromatics have been proposed [4,5]. Combination of
two vinyl radicals will lead to the formation of “hot”
1,3-butadiene that can isomerize to 1,2-butadiene and
thereafter decompose to propargyl and methyl radicals
[6]:
Unlike the allene to propyne conversion, which is
essentially thermoneutral, the 1,2 to 1,3 isomerization
reactions are highly exothermic. The reverse reaction
consequently has a large activation energy and is likely
to be important only at high temperatures, for example,
in combustion processes. Following up on their initial
suggestion [9], Hidaka et al. [10] appear to confirm the
importance of the isomerization mechanism in studies
on 1,3-butadiene decomposition. Here again the results
mustbeextractedonthebasisofacomplexmechanism.
A very important consequence of the work of Hidaka
et al. [10] is that it suggests that propargyl radicals
could be an important product arising from butadiene
→
→
2C2H3 → [1,3-butadiene]
[1,2-butadiene]
C3H3 + CH3
At high temperatures, despite its endothermicity, the
reverse process may also play a role. Information on
these interconversion processes will permit a proper

1,2-PENTADIENE DECOMPOSITION
757
decomposition. In their study of this reaction, Kiefer
et al. [11,12] originally attributed their surprisingly low
experimental activation energy to a low heat of forma-
tion of vinyl radical, where they assumed that the pro-
cess under study involved the direct formation of two
vinyl radicals. It is now clear that this cannot be the
case, however, since subsequent work has now firmly
established the “higher” heat of formation of vinyl rad-
ical [13]. There is therefore a need for an alternative
decomposition channel for 1,3-butadiene. Our initial
analysis of the present data led us to carry out more de-
tailed modeling of the 1,3-butadiene system. That work
is presented elsewhere [14] and shows that the decom-
position channel to propargyl and methyl radicals is not
sufficient to account for the results of Kiefer et al. There
is in fact the need to invoke a direct molecular channel.
Nevertheless it is clear that the vinyl radical production
channel does not contribute to 1,3-butadiene decompo-
sition and that the propargyl radical channel is the main
radical production pathway.
Although the original intention was to obtain in-
formation on the isomerization reactions, it will be
seen subsequently that bond breaking processes are
also important. There are two carbon–carbon single
bonds in 1,2-pentadiene which must be considered.
The thermodynamic parameters relevant to scission of
the propargyl–ethyl bond (CH2 C CH CH2CH3 →
C3H3 + C2H5), are available [1,15] and we find
energy is calculated to be between 295 and 314 kJ/mol.
This leads to the conclusion that the energies of the
two C C bonds are not very far apart, despite the
fact that in the first case a vinyl–carbon bond is being
broken.
EXPERIMENTAL
We have studied 1,2-pentadiene decomposition using
the comparative rate single pulse shock tube methodo-
logy. The single pulse shock tube [1] has unique capa-
bilities for studying the kinetic stability of intermediate
or large sized organic molecules. Heating is by a shock
wave, thus the walls are cold. The heating time is short,
about 500 ␮s for the present configuration. Hence re-
actants cannot migrate to or from the cold wall. Thus
all reactions must occur in the gas phase. Subsequent to
the heating pulse, samples can be removed for detailed
analysis. In the present case, gas chromatography with
flameionizationisused, thusmakingitpossibletowork
with very dilute mixtures. It also means that chemical
inhibitors, which will capture all reactive radicals, can
be added in vast excess. Thus only unimolecular reac-
tions are observed. This also means that it is possible
to simultaneously study two unimolecular process at
once. If the rate expression of one is well known, then
that of the other can be determined with great accuracy
since the known reaction can now serve as the internal
standard. Our focus on isomerization reactions means
that small amounts of several similar products need
to be simultaneously determined with quantitative pre-
cision. Chromatographic analysis is probably the only
technique by which this can be achieved and the present
type of studies the only means to obtain reliable rate
data.
The reaction mixtures that were studied contained
280 and 400 ppm of 1,2-pentadiene in the presence
of a 2% mixture of m-xylene (the inhibitor) together
with 100 ppm of cyclohexene serving as the internal
standard. The diluent gas was argon. Some experiments
were also carried out in the absence of cyclohexene.
This permitted an estimate to be made of the quantities
of ethylene and butadiene formed in the absence of the
standard. The reaction temperature ranged from 1100
to 1250 K and the pressure during reaction was between
200 and 300 kPa. The presence of the inhibitor serves
as a sink for reactive radicals through the metathesis
BDE = 1f H(C3H3) + 1f H(C2H5)
1f H(1,2-pentadiene)
= 347 [1] + 119 [1] 146 [15] = 320 kJ/mol
For fission of the butadienyl–methyl bond (CH2
C
CHCH2 CH3 → C4H5 + CH3) there is an element of
uncertainty in the bond energy that arises since buta-
dienyl radical can be expressed in the following two
resonance forms
˙
˙
H2C C CH CH2 ↔ H2C C CH CH2
To estimate the heat of formation of the radical one
can modify prototypical C H bond energies of the re-
spective parent hydrocarbons of each of the above reso-
nance forms, assuming in each case that the butadienyl
radical has the same resonance energy as allyl radical,
50 kJ/mol [1]. Using this procedure, we derive the heat
of formation of the left-most form to be 313 kJ/mol,
whereas that on the right is calculated as 294 kJ/mol.
Clearly, the above approach is not completely satis-
factory. Nonetheless, if the true heat of formation of
butadienyl radical is in this range, the C-methyl bond
Certain commercial materials and equipment are identified in
this paper in order to specify adequately the experimental proce-
dure. In no case does such identification imply recommendation or
endorsement by the National Institute of Standards and Technology,
nor does it imply that the material or equipment is necessarily the
best available for the purpose.

758
HERZLER, MANION, AND TSANG
reaction:
R + CH3C6H4CH3 → CH3C6H4CH2 + RH
RESULTS
Mechanisms and General Observations
Typical product yields at several temperatures can be
found in Table I. Also found but not tabulated are small
amounts of unsaturated compounds such as acety-
lene, propene, butene, butyne, and benzene. Except for
benzene their yields were smaller than those from the
minor isomerization channels. For benzene, the con-
centration levels are near the levels of the latter. These
minor products are treated only in a qualitative man-
ner. To a large extent this is because it is difficult to
exclusively associate their formation with a single re-
action channel. This is not to say that all the other com-
pounds can be so associated. These problems will be
discussed subsequently. The complex distribution of
products clearly demonstrates that many reactions are
occurring in parallel and in series. It appears that along
with isomerization, direct molecular decomposition as
well asbondbreakingprocessesall make contributions.
In fact, the decomposition processes are more impor-
tant than the isomerizations.
The general proposed mechanism is illustrated
in Fig. 1. Aside from the isomerization reactions,
propargyl–ethyl and butadienyl–methyl bonds are be-
ing broken. This is demonstrated by the presence of
methane and ethane, toluene, m-ethylmethylbenzene
and, as will be seen subsequently, the excess of ethy-
lene over that formed from what will be identified
as direct molecular channels. From the product spec-
trum it is clear that hydrogen atoms and methyl rad-
icals are released into the system. The former can be
definitively established from the presence of toluene
which, as noted before, can only be formed as a re-
sult of the displacement of the methyl radical from the
m-xylene inhibitor by hydrogen atoms. Methyl radi-
cals from this source, as well as those directly ejected
from bond fission in 1,2-pentadiene, lead to the forma-
tion of methane, ethane, and m-ethylmethylbenzene.
Ethane, of course, is the product of methyl radical re-
combination whereas the latter species is derived from
combination of methyl radical with m-methylbenzyl
radical, a primary species formed in the radical inhi-
bition pathway. The presence of significant quantities
of vinylacetylene is suggestive of a sequence of reac-
tions that begin with the breaking of a C-methyl bond
leading to the formation of the 2-butadienyl radical and
methyl radical.
This process quickly converts highly reactive radicals
.
(R ) such as H-atoms or methyl radicals into resonance
stabilized benzyl-type radicals. Under the present con-
ditions such stabilized radicals cannot abstract hydro-
gen atoms or add to the 1,2-pentadiene but instead
only recombine with other radicals. This is thus an-
other route for the inhibition of radical reactions. In
addition to reacting as above, hydrogen atoms can also
displace methyl groups attached to the aromatic ring:
H + CH3C6H4CH3 → C6H5CH3 + CH3
Near 1100 K this reaction is about 50% the rate of
abstraction of the benzylic hydrogens [16]. Because
this is the only realistic route to toluene in our system,
this product will serve as a convenient marker for the
presence of hydrogen atoms.
Analysis was by gas chromatography with 30 m
Q-plot and alumina columns in the programmed tem-
perature mode. Identification of the listed major prod-
ucts was made by comparison with authentic samples.
All products up to C9 aromatics were determined with
flame ionization detection. Cyclohexene decomposi-
tion was the internal standard. This is the substance
that is added to the reaction mixture as a means of de-
termining the reaction temperature. The rate expression
for cyclohexene decomposition is well established [1]
as
k(cyclohexene → butadiene + ethylene)
1
= 1.41 1015 exp( 33500/T ) s
If we now determine its extent of decomposition, the
reaction temperature can be readily determined from
the relation
1/T = (15.15 log k)/33500
with
k = t 1 ln [1/(1 (butadiene)f/(cyclohexene)i)]
where the subscripts refer to initial and final concen-
trations and t is the heating time of 500 ␮s.
The 1,2-pentadiene and cyclohexene were obtained
from Chemical Samples and were of stated purity of
99.9%. m-Xylene was from Aldrich and was of 99+%
purity. Gas chromatographic analysis did not reveal the
presence of any significant impurities. The argon was
Matheson High-purity Grade.
˙
CH2 C CHCH2 CH3 → CH2 C CH CH2 + CH3
The decomposition of the former leads to the formation
of vinylacetylene:
˙
CH2 C CH CH2 → HC C CH CH2 + H

1,2-PENTADIENE DECOMPOSITION
759
Table I Typical Product Yields in ppm
Mixtures and Conditions
400 ppm 1,2-pentadiene
100 ppm cyclohexene
2% meta-xylene at
400 ppm 1,2-pentadiene
100 ppm cyclohexene
2% meta-xylene at
283 ppm 1,2-pentadiene
2% meta-xylene at
Products
1213 K and 250 kPa
1231 K and 250 kPa
1171 K and 250 kPa
Methane
Ethane
Ethylene
Allene
34.9
2.2
53.2
3.1
25.6
4.1
16.6
172.4
N.A.a
4.8
8.7
9.4
N.A.a
N.A.a
N.A.a
66.1
13.4
143
N.A.a
N.A.a
51.1
N.A.a
16.3
21.5
5.2
286
N.A.a
2.3
6.2
5.6
79.5
16.5
N.A.a
6.1
Propyne
44.7
56.7
27.2
195
3.7
9.6
16.4
14.9
38.9
54.4
50.2
1,3-Butadiene
Vinyl Acetylene
1,2-Pentadiene
1-Pentyne
1,3-Pentadiene-cis
1,3-Pentadiene-trans
Cyclopentene
Cyclohexene
Toluene
Ethylmethylbenzene
a Not analyzed.
One direct source of ethylene in the present system
is the retroene reaction: 1,2-pentadiene → propyne
+ C2H4. However, comparison of the molar quanti-
ties of ethylene with the C3 species attributable to the
retroene reaction shows that there is a significant ex-
cess of ethylene. This indicates an additional source
of ethylene. It is almost certainly ethyl radicals that
rapidly decompose, C2H5 → C2H4 + H, thus also re-
leasing H atoms into the system.
All of the minor products can be explained in simi-
lar fashion. For example, the presence of acetylene can
be attributed to the decomposition of the 2-butadienyl
radical which, upon isomerization to 1-butadienyl rad-
ical, will then decompose to form acetylene and vinyl
radicals. The vinyl radical can then decompose to form
another acetylene and a hydrogen atom or abstract a
hydrogen to form ethylene. As our main interest is in
the initial steps in 1,2-pentadiene decomposition, we
have not attempted to develop a detailed model of such
secondary processes, and note only that the processes
are consistent with the current understanding of radical
hydrocarbon chemistry.
The major portion of the propyne that is found ap-
parently arises from the retroene decomposition: 1,2-
pentadiene → HC CCH3 + C2H4. Ofcourse, propyne
can also arise from propargyl radicals that are formed.
It is highly likely that the observed allene is derived
from propargyl radicals [17]. It cannot be derived from
propyne if we use the published kinetic results on
propyne to allene conversion [7].
The benzene that is found arises from the combina-
tion of propargyl radicals [4,5]. Finally, it is suspected
that the small amounts of other olefins are due either
to the hydrogen-atom induced decomposition of 1,2-
pentadiene or radical combination reactions involving
the vinyl radical.
Figure 1 Primary processes during the thermal decompo-
sition and isomerization of 1,2-pentadiene.

760
HERZLER, MANION, AND TSANG
Finally, it should be noted that although 1-pentyne was
observed, 2-pentyne was not detected as a product. In
the work of Hidaka et al. [9] on 1,2-butadiene decom-
position both 1-butyne and 2-butyne were found. A
rationale for this difference will be offered in the Dis-
cussion section.
If one follows the general line of reasoning given
here it will be possible to make some estimate of
the contributions from bond breaking on the basis of
the yields of methane, ethane, ethylene, toluene, and
m-ethylmethylbenzene. These will have limited accu-
racy because most of the numbers are derived from
differences. What is found is that the recovered prod-
uct yields from the breaking of the butadienyl–methyl
and propargy–methyl bonds are very close to each
other. It is suspected, however, that all the products
from the bond breaking reactions have not been recov-
ered. Specifically, some of the precursor to vinylacety-
lene, 1,3-butadien-2-yl, has been converted to unidenti-
fied products. Combination with methylbenzyl radical
(present in large amounts as part of the radical chain
inhibition chemistry) may be responsible.
Figure 2 Rate constants pertinent to the isomerization of
1,2-pentadiene to 1,3-pentadiene. Triangles (N) represent
data from yields of trans-1,3-pentadiene. Squares ( ) are
from yields of cis-1,3-pentadiene. The dotted line is the re-
sult from the sum of both 1,3-pentadiene isomers. The solid
linesrepresenttheleastsquaresfitstotheexperimentalpoints.
The dashed lines are results derived from the assumption that
the trans-isomer is the sole product and that the cis-isomer is
formed subsequently through the isomerization process. See
text.
Rate Expressions for Overall Reaction
the retroene decomposition of 1,2-pentadiene, which
leads directly to formation of ethylene and propyne,
we have slightly modified the raw data to account for
a secondary source of product. Ethylene formation is
a poor marker for this channel, as it is formed primar-
ily via bond fission. We therefore rely on production
of propyne to obtain rate constants for this pathway.
However an additional minor source of propyne is from
As previously derived [18] for a mathematically anal-
ogous system, rate constants for individual channels
can be calculated from our data using the following
relation:
ki = ( t) 1 ln [1/(1 ␣Xi)],
where Xi is the fractional conversion to a particular
P
species and = 1 +
X ).
=i
In the present case, the compounds that we con-
sider are the major direct stable products or those prod-
ucts that are directly formed from sequences of uni-
molecular reactions. Specifically, these compounds are
propyne, cis- and trans-1,3-pentadiene (Z- and E-1,3-
pentadiene), cyclopentene, 1-pentyne, and vinylacety-
lene. At high extents of decomposition they total only
about half of the 1,2-pentadiene that has disappeared.
The missing compounds are treated as a group for the
purposes of the calculations using the equation given
above. This does not introduce serious errors because
from an examination of the relation given above it can
be seen that at low conversions where the 50% unac-
counted compounds value used here may be in error,
the uncertainty of the unaccounted reaction channels
becomes very small.
Figure 3 Rate constants pertinent to the isomerization of
1,2-pentadiene to form cyclopentene ( ) and 1-pentyne ( ).
The solid lines represent the least squares fits to the experi-
mental points. The dashed line indicates the rate expression
derived for 1-pentyne formation after correction for its de-
composition. See text.
The results are given in the Arrhenius plots in
Figs. 2–4. These primary data have not been adjusted
for subsequent chemistry of the products. Regarding

1,2-PENTADIENE DECOMPOSITION
761
products produced in small amounts. Indeed the main
reactions involve bond breaking. Overall, we estimate
the 2 uncertainty in the absolute rate constants to be
about 30%.
Rate Expressions for Elementary Processes
In the subsequent discussion we convert several of the
above overall rate expressions to true elementary rate
expressions. In all cases we have considered the pos-
sible disappearance of the products used as a markers
for the particular reaction channels. This has been done
by modeling the subsequent reactions using rate con-
stants derived from appropriate literature data. At times
this has led to some changes in the rate expressions
and in other cases proved to be unnecessary. These
effects are all noted in the course of the subsequent
discussion.
Therateexpressionforconversionof1,2-pentadiene
to cis-1,3-pentadiene appears to have a disproportion-
ately large A-factor and activation energy, whereas the
parameters for the trans isomer are smaller than might
be expected. We suspect that this is due to the initial
product being the trans isomer and that interconversion
to the cis compound occurs from this source as op-
posed to formation directly from 1,2-pentadiene. This
cis–trans isomerization of 1,3-pentadiene was studied
by Marley and Jeffers [19] many years ago. Results of
calculations using their rate constants for cis–trans iso-
merization where it is assumed that the trans compound
is the first product and that the cis compound is formed
from it are given in Fig. 2 (dotted line). It can be seen
that the experimental distribution of cis and trans com-
pounds is reproduced. Note that the equilibrium con-
stant for this reaction is near 1 under our conditions [15]
so the product distribution is still kinetically controlled
even at the highest temperatures studied. Although we
cannot rule out that some minor fraction of the product
is initially the cis species, the data are consistent with
isomerization of 1,2-pentadiene to 1,3-pentadiene be-
ing stereospecific for the trans isomer. This leads to the
following corrected rate expression for the conversion
of 1,2-pentadiene to trans-1,3-pentadiene.
Figure 4 Rate constants pertinent to the retroene decom-
position of 1,2-pentadiene to form propyne and ethylene ( )
and vinylacetylene ( ).
propargyl radicals. To correct for this we make use of
previous studies [17], which suggest that reactions
of propargyl radicals should result in about equal yields
of allene and propyne. Hence to obtain the amount of
propyne formed from the retroene reaction we have
subtracted the allene yield from the total formation of
propyne. This adjustment is in the range of 10–15%. As
shown in Fig. 4, after this small correction, we obtain
k(1,2-pentadiene → propyne + ethylene)
1
= 6.6 1012 exp( 29240 900 K/T ) s
The Arrhenius rate expressions for the other processes
are as follows:
k(1,2-pentadiene → trans-1,3-pentadiene)
1
= 1.6 1013 exp( 31283 900 K/T ) s
k(1, 2-pentadiene → cis-1,3-pentadiene)
1
= 2.7 1016 exp( 41000 900 K/T ) s
k(1,2-pentadiene → 1-pentyne)
1
= 6.3 1011 exp( 29164 900 K/T ) s
k(1,2-pentadiene → cyclopentene)
k(1,2-pentadiene → trans-1,3-pentadiene)
1
= 1.8 1011 exp( 26080 900 K/T ) s
1
= 2.2 1014 exp( 33900 K/T ) s
k(1,2-pentadiene → vinylacetylene)
= 1.3 1018 exp( 44550 900/T ) s
1
Note that the above parameters comprise the trans-
ferable rate expression. The individual rate parameters
for the cis and trans isomers derived directly from our
data are valid only in our system and should not be
used elsewhere. Because the 1,3-pentadienes are much
more thermodynamically stable than 1,2-pentadiene,
The uncertainties in the activation energies are a factor
of 3 larger than those usually obtained from similar
single pulse shock tube experiments [1]. Part of the
problem is due to the large number of isomerization

762
HERZLER, MANION, AND TSANG
the reverse isomerization to 1,2-pentadiene is very slow
and it is not necessary to correct the data for that
process.
Relative Importance of C C Bond
Fission Channels
As indicated in Fig. 1, the two possible C C bond fis-
sions in 1,2-pentadiene lead to production of methyl
and ethyl radicals, respectively. To the extent that we
are able to count the number of these species formed
directly from 1,2-pentadiene, we can gauge the rel-
ative importance of the two channels. Table II sum-
marizes the results in terms of methyl and ethyl radi-
cals formed during the decomposition process. These
numbers are not very precise because it is necessary
to derive these values by taking into account multi-
ple sources and sinks for these radicals. For example,
methyl radicals are formed by fission of a C C bond
in the starting compound, but also by the reaction of
H atoms with the inhibitor. Although this latter chan-
nel is conveniently marked by toluene formation, the
correction adds to the uncertainty. Thus the quantity
of methyl radicals directly attributable to bond fission
in 1,2-pentadiene is derived from the sum of methane,
ethane, and m-ethylmethylbenzene yields and reduced
by the amount of toluene formed. It is also necessary
to subtract the small amount of background methane.
The alternate C C bond fission leads to ethyl radicals,
which rapidly decompose to ethylene. However, ethy-
lene is also formed in the retroene decomposition of
1,2-pentadiene. Further we cannot simply account for
this based on propyne formation, since propargyl rad-
icals also lead to some propyne (see previous mech-
anistic discussion). Following our earlier discussion,
propargyl radical decomposition gives approximately
equal amounts of allene and propyne. Thus for the ethyl
radical channel we use
The rate constant for the 1-pentyne formation
channel is affected by the subsequent decomposition
of 1-pentyne to form propargyl and ethyl radicals
as well allene and ethylene. The rate constants for
the decomposition of 1-pentyne have been taken as
equal to those found [1] for 1-hexyne decomposition,
and we estimate k(1-pentyne → propargyl + ethyl) =
1
7.9 1015 exp( 36300/T ) s and k(1-pentyne →
1
allene + ethylene) = 5.0 1012 exp( 28400/T )s
.
When these reactions are taken into account one ob-
tains the Arrhenius plot given in Fig. 3. The new rate
expression is
k(1,2-pentadiene → 1-pentyne)
1
= 1.0 1013 exp( 32300/T ) s
Although the equilibrium constants of formation of
1,2-pentadiene and 1-pentyne are very close to each
other, and hence the above reaction is reversible, the
low degree of conversion to 1-pentyne assures that the
reverse process may be ignored.
As noted above, in the case of the 1,3-pentadiene
product it proved unnecessary to correct the data for the
reverse isomerization. This is also the case for cyclo-
pentene, which is much more thermodynamically
stable than 1,2-pentadiene. Cyclopentene, in any
case, would preferentially undergo the well-studied
conversion to cyclopentadiene rather than reform
1,2-pentadiene. That process is also too slow to be
important, however, in agreement with the literature
data [20] and the absence of this species in the product
spectrum.
EthylC
EthyleneTotal
C
(Propyne Allene) Butadiene
We have suggested that vinylacetylene arises
through fission of the butadienyl–methyl bond of
1,2-pentadiene followed by ␤ fission of H from the
radical intermediate (Fig. 1). The rate expression for
this process is characterized by extraordinarily large
parameters, however, much larger than would be ex-
pected for a “normal” C C bond fission. What must be
happening is that, unlike simple alkyl radicals, the buta-
dienyl radical intermediate is stable enough that it does
not immediately eject a hydrogen atom. At low tem-
peratures in particular its unimolecular lifetime is long
enough that bimolecular reaction channels become im-
portant. The large activation energy is thus a reflection
of an ever increasing fraction of the radical interme-
diate undergoing decomposition to vinyl acetylene as
the temperature is increased. Clearly the directly deter-
mined rate expression is a composite value and should
not be used in other systems.
The butadiene correction is only necessary in the
cases where the retroene reaction of cyclohexene (→
C4H6 + C2H4) has been used as the standard. It is es-
timated that the uncertainties in the ratios of contribu-
tions from each channel may be as large as a factor
of 1.3. Nevertheless, the observation that the ratios of
Table II Methyl and Ethyl Radical Yields from
1,2-Pentadiene Decomposition
Percent Methyl
Radicals from
1,2-Pentadiene
Percent Ethyl
Radicals from
1,2-Pentadiene
Temperature (K)
1170
1174
1199
1204
28
32
31
32
22
23
27
34

1,2-PENTADIENE DECOMPOSITION
763
methyl to ethyl radicals are fairly close to each other
is suggestive that the two bonds must have similar
strengths.
If the first of these rate expressions can be taken
to be that for the conversion of 1,3-butadiene to
1,2-butadiene, then some conclusions can be drawn re-
garding the decomposition chemistry of 1,3-butadiene.
Specifically the rate of isomerization can be compared
with that for C C bond fission to give two vinyl radi-
cals. The rate expression for this latter process can be
derived from the reported recombination rate of vinyl
radicals [21] and the now established thermochemistry
of C2H3 [13]. We find
DISCUSSION
Rates of Isomerization
Comparison of the present result on the conversion of
1,2-pentadieneto1,3-pentadienewiththatestimatedby
Hidaka et al. [9] for the analogous butadiene compound
can be found in Fig. 5A. Agreement is very good ex-
cept for a small difference in the activation energy. This
strongly suggests that methyl substitution has a mini-
mal effect on the kinetics of the isomerization process.
On the basis of detailed balance, the present results on
the isomerization reactions of 1,2-pentadiene can be
used to calculate the rate expressions for the reverse
reactions. For this purpose we use the thermodynam-
ics of the stable compounds from Stull et al. [15] The
following rate expressions, valid for 1100–1250 K, are
derived:
k(1,3-butadiene → 2C2H3)
1
= 2.8 1017 exp( 56300 K/T ) s
.
The comparison is shown graphically in Fig. 5B.
Clearly the isomerization must be favored under all ap-
plicable conditions. Thus in the decomposition of 1,3-
butadiene, isomerizationto1,2-butadienemustbemore
important than direct formation of vinyl radical. Once
formed, 1,2-butadiene can decompose by C C bond
cleavage to form propargyl and methyl radicals. This
latter reaction is endothermic by 335 kJ/mol. Because
1,3-butadiene is more stable than the 1,2 compound
by 55 kJ/mol [15] the net conversion of 1,3-butadiene
to propargyl and methyl radicals will require a total
energy of 390 kJ/mol. This is very close to the acti-
k(trans-1,3-pentadiene → 1,2-pentadiene)
1
= 1.2 1015 exp( 42070 K/T ) s
k(1-pentyne → 1,2-pentadiene)
vation energy, E = 391–393 kJ/mol, determined by
∞
1
= 2.0 1013 exp( 32700 K/T ) s
Kiefer et al. [11,12] for 1,3-butadiene decomposition.
Although on the surface this appears compatible, there
remain some problems. To fit the data of Kiefer et al.
with the above mechanism it would require that the
k(cyclopentene → 1,2-pentadiene)
1
= 4.8 1014 exp( 40950 K/T ) s
Figure 5 Some rate constant comparisons. Graph A: (1) dashed line, the retroene decomposition: 1,2-pentadiene → propyne +
ethylene; (2) dash-dot-dash line, 1,2-pentadiene → 1-pentyne; (3) dotted line, 1,2-pentadiene → 1,3-pentadiene; (4) solid line,
1,2-butadiene → cyclopentene; (5) dash-dot-dot-dash line, 1,2-butadiene → 1,3-butadiene from Hidaka et al. [9]. Graph B: solid
line, 1,3-pentadiene → 1,2-pentadiene; dashed line, estimated rate constant for 1,3-pentadiene → C2H3 + 1-propenyl.

764
HERZLER, MANION, AND TSANG
isomerization 1,2-butadiene → 1,3-butadiene be much
slower than the decomposition process 1,2-butadiene
→ propargyl + methyl. Although the latter process
has a much larger A-factor, this is compensated for by
the 45 kJ/mol lower activation energy of the isomer-
ization. The end result is that the overall process for
formation of propargyl and methyl radicals cannot be
the main channel observed by Kiefer et al. As noted
earlier, more detailed calculations [14] have confirmed
this and suggest that the main channel in 1,3-butadiene
decomposition involves a molecular channel which di-
rectly forms acetylene and ethylene.
The present analysis is compatible with the mecha-
nism given by Hidaka and coworkers for 1,3-butadiene
decomposition since propargyl and methyl radicals are
the important radicals released into the system during
1,3-butadienedecomposition. Theseresultsshouldalso
permit the determination of the branching ratio for the
reactions
Mechanism of Isomerization
In the Introduction we mentioned the possibility
of a carbene mechanism for the allene to propyne
isomerization reaction. Historically, the postulated
mechanism for this reaction has evolved from direct
H transfer, to the participation of first diradical and then
carbene intermediates. Consequently, mechanistic
discussion in the past literature should be treated
cautiously and, indeed, the issues are still not fully
settled. The current thinking was largely establi-
shed in a series of papers in the mid to late 1980s
by Honjou et al. [22–24] in which they used ab initio
calculations to explore the C3H4 surface in detail.
They proposed that the dominant mechanism is that
shown in Fig. 6, involving carbene and cyclopropene
intermediates. Subsequent theoretical [25–27] and
experimental [28,29] work seem to broadly support
this view. A mechanism of the same general type
can obviously accommodate the interconversion of
1,2-pentadiene and 1-pentyne. A general scheme is
given Fig. 7. It is immediately apparent how complex
the C5H8 surface is in comparison to that of C3H4. Note
that the cyclopropenes and cyclobutenes are proposed
as intermediates only and are far too unstable at our
temperatures to survive as end products. The gas phase
decomposition of one of the possible intermediates,
1-ethylcyclopropene, has been studied at much lower
temperatures (472 K) by Hopf et al. [28]. They obtained
mass balances of (100 2)% and found 2-pentyne,
trans- and cis-1,3-pentadiene in near-equilibrium
proportions, and 1,3-dimethylcyclopropene as major
products. Traces of isoprene (2-methylbutadiene) and
1,2-pentadiene were also observed, but cyclopentene
was not a product. Hopf et al. discussed formation
of 2-pentyne and 1,3-dimethylcyclopropene, but
did not propose a mechanism for production of the
1,3-pentadienes.
C2H3 + C2H3 ⇔ C2H3-C2H3
⇔ C3H3-CH3 ⇔ C3H3 + CH3
We are carrying out such calculations as the present
time.
The rate expressions given above for formation of
1,2-pentadiene from 1-pentyne and cyclopentene can
be compared in a similar fashion with known rate
expressions for other decomposition or isomerization
processes. These are
k(1-pentyne → propargyl + C2H5)
= 7.9 1015 exp( 36300 K/T ) s 1 [1]
k(cyclopentene → cyclopentadiene + H2)
= 2.24 1013 exp( 30193 K/T ) s 1 [21]
A detailed discussion of the full mechanism would
be speculative and beyond the intended scope of this
work but some comments are in order. First, the ob-
servation that in our system formation of trans-1,3-
pentadiene is heavily favored over the cis isomer is
suggestive that the primary route to this product is
via 3-methylcyclobutene, which is known to isomerize
It can be seen that the rate constants for both of these
processesare larger thanthe onescalculatedfor isomer-
ization to 1,2-pentadiene, at least under higher temper-
ature conditions. Thus the two isomerization reactions
are not likely to be of importance in most applications.
We have assigned the propyne yields predominantly
to the retroene reaction. The activation energy for this
process is very much like those for the retroene reac-
tions of other alkenes and alkynes. A direct comparison
of the rate the retroene channel relative to some of the
isomerization pathways of 1,2-pentadiene can be found
in Fig. 5A.
Figure 6 Mechanism proposed by Honjou et al. [22–24]
→
for the allene
propyne isomerization reaction.

1,2-PENTADIENE DECOMPOSITION
765
Figure 7 Suggested mechanism for the isomerization reactions of 1,2-pentadiene. The main observed isomerization products
are indicated in bold. See text for discussion.
stereospecifically to trans-1,3-pentadiene [30]. Two
other possible routes are put forth in Fig. 7: both involve
1,2-hydrogen shifts, either from a carbene intermedi-
ate or from the ethyl group of 1-ethylcyclopropene.
These latter paths would be expected to generate a less
stereospecific product distribution, contrary to what
is observed. In this regard it is noteworthy that Hopf
et al. [28] observed a near-equilibrium proportion of
1,3-pentadienes in the low temperature decomposition
of 1-ethylcyclopropene. The gas phase isomerization
rate of 1,3-pentadienes has been measured [18,31] and
is far too slow to account for the equilibrium product
distribution found by Hopf et al. Although it is possible
that the long time scale (15 h) of the Hopf et al. stud-
ies could lead to surface mediated isomerization, the
1,2-hydrogen shift proposed in Fig. 7 would equally
well explain their results. Hopf et al. did not find any
cyclobutenes as products, but these compounds have
half-lives of only about 10 s under their conditions [30].
A second noteworthy observation is that we did not ob-
serve 2-pentyne as a product, while it would perhaps
be expected on the basis of the Hidaka et al. studies,
as well as the Hopf et al. work, if 1-ethylcyclopropene
were an important intermediate. A third observation is
that Hopf et al. report no cyclopentene while we find it
to be a relatively major pathway. An examination of the
mechanism of Fig. 7 presents a possible rationale for
the above observations. The differences with the low
temperature data of Hopf et al. are suggestive that in the
high temperature isomerization of 1,2-pentadiene the
1-ethylcyclopropene part of the C5H8 energy surface
is minimally sampled. This would require either that
the initial C3 → C2 H shift is heavily favored over
the C1 → C2 shift in 1,2-pentadiene (see Fig. 7) and
thus that the chemistry of 3-ethylcyclopropene domi-
nates, or that competing 1,3-, 1,4- and 1,5-insertion re-
actions dominate the chemistry of the initially formed
carbene. This latter suggestion is not fully satisfying as
the current thinking from low temperature experiments
[28,29] is that the insertion reactions of carbenes are
somewhat slower than 1,2-H transfer reactions, which
are in turn slower than simple closure to cyclopropenes.
The former suggestion, on the other hand, seems to
be supported by the work of Hopf et al. [28,29].
This postulate would not be at odds with the Hidaka
et al. work provided that in the C5 system the initially
formedcarbenepreferentiallyundergoesringclosureto
3-ethylcyclopropene, or, again, that insertion reactions
are dominant.
To summarize, our best explanation, at present, is
that the principal isomerization chemistry we observe
is channeled through the 3-ethylcyclopropene portion
of the energy surface. This suggests that secondary hy-
drogens preferentially undergo the initial H shift reac-
tion. Such patterns would be expected to be general in
nature and should help predict the reactivity of larger
1,2-dienes. A more detailed examination of the C5H8
energy surface would seem warranted and could per-
haps clarify the situation. A secondary issue that we
have not considered in detail is the possible role of

766
HERZLER, MANION, AND TSANG
1,2-alkyl transfers in the carbenes, a process that may
be important in some instances [29]. A better under-
standing of the C5H8 surface would again be helpful.
It is interesting that all the isomerization activation
energies are in the 240–280 kJ/mol range. This is very
close to the ␲-bond energy of olefins and is consistent
with the mechanisms involving fission or rearrange-
ment of this bond in some manner. This may well be
a general feature of the isomerizations of polyunsatu-
rated hydrocarbons.
Conclusion—Some Implications Regarding
Hydrocarbon Decomposition
The greater importance of the fragmentation reactions
in comparison to the isomerization processes is of con-
siderable importance in the context of general mech-
anisms for hydrocarbon decomposition. It means that
alkyl groups that are appended to sites of unsaturation
will be rapidly released into the reacting system. This
is brought about by the resonance stabilization of the
propargyl radical. This sets definite limitations on the
numbers and types of species that are important in
high temperature hydrocarbon environments even for
the largest and most complex of mixtures. A limit
can then be set on the number and types of species
that must be considered in complex models of such
systems.
The cyclopentene formation reaction is character-
ized by a low A-factor, 1.8 1011
s
1. Presumably
this is a reflection of a tighter transition state struc-
ture than that for conversion to trans-1,3-pentadiene
(A = 2.2 1014
s
1). However, the proposed transi-
tion states both involve carbene insertions and it is not
clear how these structures may be so different. We have
investigated the possibility that secondary decompo-
sition of cyclopentene to form cyclopentadiene [20]
could be perturbing our derived rate expression for cy-
clopentene formation. The published rate constants are
not large enough to make a significant contribution,
however, and it is not obvious what other reaction could
play a role.
BIBLIOGRAPHY
1. Tsang, W. In Shock Waves in Chemistry; Lifshitz, A.
(Ed.); Marcel Dekker: New York, 1981; pp. 59–129.
2. Benson, S. W.; O’Neal, H. E. NSRDS-NBS 21; U. S.
Government Printing Office, 20402, 1970.
3. Tsang, W.; Kiefer J. H. In The Chemical Dynamics and
Kinetics of Small Radicals; Liu, K.; Wagner, A. (Eds.);
1995; p. 58.
Bond Fission Reactions—Bond Energies
The relative equality in the yields of products from the
breaking of the two C C single bond strongly suggests
that these reactions will be characterized by nearly the
same rate expressions. Subject to the uncertainties that
have been mentioned earlier, the bond dissociation en-
ergies would therefore have to be similar. Note that pos-
sible differences in the A-factors are unlikely, since it
has been found [1] that bond breaking reactions leading
to the formation of a stiffened resonance stabilized rad-
ical have completely “typical” values. This then leads
to a butadienyl-CH3 bond dissociation energy of 318
6 kJ/mol or a heat of formation of butadienyl radical
(at 298 K) of 317 6 kJ/mol. This means that the sec-
ondary hydrogens in butadiene have bond strengths of
418 kJ/mol. We estimate that the bond strength of the
vinyl C H bond in 1-butene is 452 kJ/mol. The res-
onance energy of butadienyl radical is therefore only
34 kJ/mol. This is consistent with the 13–20 kJ/mol sta-
bilization energy of butadiene (due to the conjugated
␲ bonds) being lost in the radical. Classical molec-
ular orbital theory provides a ready explanation: the
sp-hybridized carbon in the radical contains two or-
thogonal p orbitals, only one of which can overlap with
the adjacent ␲ bond. Thus the radical must make a
“choice” of forms and the net observed stabilization is
the allylic resonance energy minus the ␲ bond conju-
gation energy.
4. Frenklach, M.; Clary, D. W.; Gardiner, W. C.; Stein, S. E.
In 20th Symp (Int) on Combustion, The Combustion
Institute, Pittsburg, PA, 1984; p. 887.
5. Stein, S. E.; Walker, J. A.; Suryan, M. M.; Fahr, A. In
23rd Symp (Int) on Combustion, The Combustion Insti-
tute, Pittsburg, PA, 1990; p. 85.
6. Kubitza, C. DasVinylradikal—Massenspektrometrische
Untersuchungen seiner Radikal-Radikal-Reaktionen
und seiner Additionsreaktionen mit ungesattigten
Kohlenwassestoffen in der Gasphase; PhD Dissertation;
Technischen Hochschule Darmastadt, 1995.
7. Lifshitz, A.; Frenklach, M.; Burcat, A. J Phys Chem
1975, 79, 1148.
8. Bailey, I. M.; Walsh, R. J Chem Soc Faraday Trans I
1978, 74, 1146.
9. Hidaka, Y.; Higashihara, T.; Nimomiya, N.; Oki, T.;
Kawano, H. Int J Chem Kinet 1995, 27, 331.
10. Hidaka, Y.; Higashihara, T.; Nimomiya, N.; Masaoka,
H.; Nakamura, T.; Kawano, H. Int J Chem Kinet 1996,
28, 137.
11. Kiefer, J. H.; Wei, H. C.; Kern, R. D.; Wu, C. H. Int J
Chem Kinet 1985, 17, 225.
12. Kiefer, J. H.; Mitchell, K. I.; Wei, H. C. Int J Chem Kinet
1988, 20, 787.
13. Kaiser, E. W.; Wallington, T.; J Phys Chem 1995, 99,
10549.
14. Tsang, W.; Mokrushin, V. In 28th Symp (Int’l) on Com-
bustion, The Combustion Institute, Pittsburg, PA, Part 2,
2000, 28, 1717.

1,2-PENTADIENE DECOMPOSITION
767
15. Stull, D. R.; Westrum, E. F. Jr.; Sinke, G. C. The Chem-
ical Thermo dynamics of Organic Compounds; Wiley:
New York, 1969.
16. Robaugh, D.; Tsang, W. J Phys Chem 1986, 90, 4159.
17. Tsang, W. Int J Chem Kinet 1970, 2, 23.
18. Tsang, W. J Chem Phys 1965, 43, 352.
19. Marley, W. M.; Jeffers, P. M. J Phys Chem 1975, 79,
2085.
20. Lewis, D. K., Bergmann, J.; Manjoney, R.; Paddock, R.;
Kaira, B. L. J Phys Chem 1984, 88, 4112.
21. Fahr, A.; Laufer, A.; Klein, R.; Braun, W. J Phys Chem
1991, 95, 3218.
22. Honjou, N.; Pacansky, J.; Yoshimine, M. J Am Chem
Soc 1984, 106, 5361.
24. Yoshimine, M.; Pacansky, J.; Honjou, N. J Am Chem
Soc 1985, 107, 5332.
25. Kakumoto, T.; Ushirogouchi, T.; Saito, K.; Imamura, A.
J Phys Chem 1987, 91, 183.
26. Melius, C. F.; Miller, J. A.; Evleth, E. M. 24th
Symp (Int’l) on Combustion, The Combustion Institute,
Pittsburg, PA, 1992; p. 621.
27. Davis, S. G.; Law, C. K.; Wang, H. J Phys Chem 1999,
103, 5889.
28. Hopf, H.; Plagens, A.; Walsh, R. Liebigs Ann 1996,
825.
29. Hopf, H.; Plagens, A.; Walsh, R. Angew Chem Int Ed
Engl 1996, 36, 381.
30. Frey, H. M. Trans Faraday Soc 1964, 60, 83.
31. Frey, H. M.; Lamont, A. M.; Walsh, R. J Chem Soc A
1971, 2642.
23. Honjou, N.; Pacansky, J.; Yoshimine, M. J Am Chem
Soc 1985, 107, 5332.