Thermal stability of larger carbonyl compounds: 2-methylbutyraldehyde

Article abstract of DOI:10.1002/kin.20828

We have decomposed 2-methylbutyraldehye in single pulse shock tube experiments, at temperatures between 1075 K to 1250 K and pressures of 1.5 and 6 atm. Five reaction channels have been identified from the unimolecular products that have been detected. The products involve C-C bond cleavage together with a retro-ene reaction. Particularly remarkable is the detection of small amount of butane; the first time that an alkane has been detected in these high-temperature experiments. The formation of butane represents evidence of the roaming radical reaction or more generally the high-temperature radical disproportionation reaction. Ethylene is present in large yields, and this is traced to a 1,2 H-transfer reaction involving the highly exothermic transfer of a carbonyl hydrogen to an alkyl site. The results are consistent with earlier single pulse shock tube experiments on the 2-ketones. In particular, they confirm the earlier observation that substituting a methyl and hydrogen atom with a doubly bonded oxygen atom have very little effect on the rate constant of the adjoining C-C bond. The yields of ethylene are consistent with a very low value of the carbonyl resonance energy.

Full text of DOI:10.1002/kin.20828

Thermal Stability of Larger
Carbonyl Compounds:
2-Methylbutyraldehyde
CLAUDETTE M. ROSADO-REYES, WING TSANG
National Institute of Standards and Technology, Gaithersburg, MD 20899
Received 31 May 2013; revised 4 September 2013; accepted 6 September 2013
DOI 10.1002/kin.20828
Published online 10 January 2014 in Wiley Online Library (wileyonlinelibrary.com).
ABSTRACT: We have decomposed 2-methylbutyraldehye in single pulse shock tube experi-
ments, at temperatures between 1075 K to 1250 K and pressures of 1.5 and 6 atm. Five reaction
channels have been identified from the unimolecular products that have been detected. The
–
products involve C C bond cleavage together with a retro-ene reaction. Particularly remarkable
is the detection of small amount of butane; the first time that an alkane has been detected
in these high-temperature experiments. The formation of butane represents evidence of the
roaming radical reaction or more generally the high-temperature radical disproportionation
reaction. Ethylene is present in large yields, and this is traced to a 1,2 H-transfer reaction
involving the highly exothermic transfer of a carbonyl hydrogen to an alkyl site. The results
are consistent with earlier single pulse shock tube experiments on the 2-ketones. In particu-
lar, they confirm the earlier observation that substituting a methyl and hydrogen atom with a
–
doubly bonded oxygen atom have very little effect on the rate constant of the adjoining C C
bond. The yields of ethylene are consistent with a very low value of the carbonyl resonance
ꢀC
energy. 2014 Wiley Periodicals, Inc. Int J Chem Kinet 46: 285–293, 2014
–
the C C bond being broken and the A factors show
INTRODUCTION
lower values as the temperature increases [2]. A sum-
mary of relevant bond dissociation energies can be
found in a review published over a decade ago [3] and
it shows the differences with values that were gener-
ally used for many years [4]. While there has been
little change in recent years, theoretical work has now
furnished additional support [5] for smaller radicals
at least. Thus for unsaturated hydrocarbons, there are
processes involving six center transition states, usually
termed the retro-ene reaction. In the case of alcohols,
This paper is a continuation of earlier studies on the
quantitative details of the decomposition of larger or-
ganic compounds in the gas phase and in the absence of
radical-induced processes. The aim is to determine the
mechanism and rate expressions for these processes.
Hitherto, we have studied alkanes, alkenes, alkynes,
branched aromatics, alcohols, ketones, amines, and
alkyl halides [1]. These decompose by various molec-
ular and bond cleavage processes. The rate expressions
for the latter are consistent with the bond energies of
–
there are two channels, dehydration and C C bond
cleavage. We have recently established the rate rules
for these processes [6,7]. In the case of the ketones, [8]
–
we have found that C C bond cleavage processes are
Correspondence to: Claudette M. Rosado-Reyes; e-mail:
the predominant decomposition mechanism with only
a minor retro-ene decomposition channel. One would
claudette.rosado-reyes@nist.gov.
C
ꢀ
2014 Wiley Periodicals, Inc.

286
ROSADO-REYES AND TSANG
expect that the aldehydes will be a simple extension of
these observations.
reaction releases sec-butyl radical into the system.
Beta bond scission releases propene and methyl radi-
cals. A minor reaction channel releases H atoms and
2-butenes. In addition, the breaking of appropriate
More recent studies on the so-called “roaming” re-
actions suggest that there may well be far richer chem-
istry [9,10]. This is due to the discovery of the reac-
tion of methyl and formyl radicals, where the main
reaction channel turns out to produce methane and car-
bon monoxide. Theoretical explanation of these re-
sults have been provided by Bowman and Shepler [9].
Sivaramakrishnan et al. [10] provided evidence for
small contributions for this type of reaction for the
decomposition of isobutane and neopentane. Although
not acknowledged, the general issue is related to the
discovery of disproportionation products during the
combination of alkyl radicals at room temperature with
much discussion regarding the nature of the transition
state [11].
The presence of molecular products in thermal de-
compositions that are initiated by bond breaking is di-
rectly related to disproportionation reactions. Thus the
failure to detect smaller alkanes from the decomposi-
tion of the larger alkanes from single pulse shock tube
experiments would seem to rule out this possibility.
In single pulse shock tube studies, methane cannot be
used as a marker of the reaction since methyl radicals
will always be converted to methane by reaction with a
reaction scavenger. On the other hand, any larger alkyl
radicals with an ethyl terminal will always decompose
into an olefin due to the high temperatures, character-
istic of single pulse shock tube experiments. There is
very little question that smaller alkanes are more stable
than the associated larger compounds. Hence any de-
tection of such small alkanes except for methane and
ethane is indicative of their direct formation and to
some extent may be considered as evidence for their
formation during their initial decomposition process.
The presence of molecular channels in the decom-
position of hydrocarbons is of great importance in high-
temperature processes because it will reduce the con-
centration of radicals that are formed. These radicals
dictate the course of the chain branching. Combustion
is of course the best example of this.
–
C C bonds will yield radical products, except for
methyl, that will readily decompose into appropriate
olefins.
Another interesting issue involves the carbonyl res-
onance energy. The experimental studies on ketones
have indicated that the carbonyl resonance energy is
much smaller than the allylic resonance energy [4].
However, ab initio calculations indicate an equal or
larger value than for allyl radicals [13]. Finally, we
note that the carbonyl-H bond energy is much lower
than that for hydrocarbons, with an actual magnitude
on the order of 42 kJ/mol [3]. This discrepancy raises
the question of possible contribution from 1-2 and 1-3
hydrogen transfer processes leading to the formation of
a carbonyl-type radical that will subsequently release
CO directly. It is quite clear that a number of inter-
esting new elements are introduced by the addition of
a formyl group to a hydrocarbon framework, and this
study is carried out with the hope of obtaining insights
into the nature of these processes.
Table I summarizes some past work on related sys-
tems, such as ketones, alkanes, and alkenes. Note the
similarities and differences in the rate constants and ex-
pressions that involve the addition of a carbonyl system
into the basic alkane framework. Such comparisons
lead to very suggestive rate rules that are expected to
be manifested in the course of the experiments. The
two significant structural elements are the introduction
of the oxygen atom and the double bond. It is inter-
esting that the bond energy of a carbon-vinyl bond
–
is enormously larger than the C carbonyl bond be-
ing broken in these studies. In comparable olefins, the
48 kJ/mol resonance energy is so much larger than an
–
ordinary C C bond. Thus the overwhelming process
is the breaking of the carbon allylic bond. The overall
effect of substituting the doubly bonded oxygen atom
for a methyl group and hydrogen atom is to lower the
rate constant by 50%. This is a small change and is
even more unexpected than in the case of the alcohols.
The experiments have been carried out in a sin-
gle pulse shock tube under conditions that prevent
radical-induced decomposition and isolate the initial
unimolecular pyrolitic reaction for detailed study. This
isolation is achieved by carrying out the shock tube
studies under high dilution and in the presence of
a chemical inhibitor that removes all active radicals
from the reaction mixture and also allows the use of
an internal standard for estimating the temperature
behind the reflected shock wave. The chemical in-
hibitor utilized in this study is 1,3,5-trimethylbenzene
The subject of this study is 2-methylbutyraldehyde.
From the previous study of ketones [8] predictions can
–
be made on the various C C bond cleavage reactions
that can occur. A particularly interesting reaction is in-
–
volving the breaking of the C C bond next to the car-
bonyl group that will lead to the formation of a formyl
and secondary butyl radical. If roaming occurs, an im-
portant product that will be formed is normal butane.
This is of course analogous to the formation of methane
and carbon monoxide from the combination of formyl
and methyl radicals. We have previously studied the
decomposition of 3,4-dimethylhexane [12]. The main
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

THERMAL STABILITY OF LARGER CARBONYL COMPOUNDS
287
Table I Past Work on Rate Expressions Rate Constants (1150 K) for Reactions Analogous to the Reactions Studies in
This Study Dealing with 2-Methylbutyraldehyde
Reaction
Rate Expression (s−1
)
Rate Constant (s−1) (1150 K)
3-Methylpentanon-2 → CH3CO + s-C4H9 [8]
5-Methylhexanone-2 → CH3COCH2 + i-C4H9 [8]
5-Methylhexanone-2 → CH3COCH3 + i-C4H8 [8]
Hexanone-2 → CH3COCH3 + C3H6 [8]
3-Methylpentane → CH3CH2 + s-C4H9 [12]
2,3-Dimethylpentane → (CH3)2CH + s-C4H9 [12]
2-Methylbutane → CH3 + s-C4H9 [12]
1-Hexene → allyl + n-propyl [14]
1016.4 exp(−38,300 K/T)
1016.6 exp(−40,600 K/T)
1012.56 exp(−31,600 K/T)
1013.28 exp(−32,400 K/T)
1016.7 exp(−40,000 K/T)
1016.1 exp(−38,100 K/T)
1016.4 exp(−42,100 K/T)
1015.9 exp(−35,600 K/T)
1012.64 exp(−28,900 K/T)
35.9
18.6
4.2
1.1
39.1/2
53.1
3.15/2
288
1-Hexene → 2 propene [14]
48.8
The rate constants are divided by 2 to take into account the reactions where there are two pathways.
(135TMB). Radical attack on 135TMB leads to the
formation of the benzyl-type reaction that is unreac-
tive under the present environment.
where the rate constant is defined as kt = −ln (1−([1,3-
butadiene]/[cyclohexene]i)) = A exp(−Ea/RT) and t is
the residence time of approximately 500 μs. Note that
this is the same residence time that is used in defin-
ing the rate constant for the decomposition process
involving the target molecule 2-methylbutyraldehyde.
Since the target and standard molecules are in the same
sample, they must also experience the same tempera-
ture. There is thus large-scale canceling of the usual
kinetic uncertainties. Hence the data precision is very
high and the scatter is only from the gas chromato-
graphic measurements. The addition of the radical
inhibitor removes mechanistic artifacts from affect-
ing the rate expressions that are obtained. In addi-
tion, since both target and standard compounds must
have suffered the same temperature and time resolu-
tion, a drastic reduction of the experimental scatter is
achieved.
Postshock samples are immediately extracted from
the shock tube through the sampling port into the vac-
uum valve and loop sampling system, consisting of
two valves and loop samplers, where the undiluted
gas sample is compressed to 1 atm of pressure, sepa-
rated into two loop volumes of 1 mL each, and then
injected into two gas chromatographic columns for
species separation. The gas chromatography system
(Agilent Technologies, Santa Clara, CA, USA; model
6890N) was equipped with twin flame ionization de-
tectors (FID) and an Agilent Technologies 5973 inert
mass selective (MS) detector for mass spectral anal-
yses. C4 and larger compounds are separated in a
J&W Scientific 30 m × 0.53 mm i.d. DB-1 fused sil-
ica (100% dimethylpolysiloxane) column. The eluting
fraction from DB-1 is split for simultaneous FID and
MS analyses. Lighter compounds are separated effi-
ciently in a Restek 30 m × 0.53 mm i.d. RT-Alumina
(aluminum oxide porous layer) capillary column. Peak
EXPERIMENTAL∗
The basic procedures used in carrying out the com-
parative rate single pulse shock tube experiments have
been discussed in detail in earlier studies [1,15,16].
The reaction mixtures used in these studies are 500
μL/L 2-methylbutyraldehyde (target species; ≥95%;
Sigma-Aldrich, St. Louis, MO, USA), 200 μL/L of cy-
clohexene (internal standard; 99+%; Sigma-Aldrich),
and 10,000 μL/L 135TMB (chemical inhibitor; 99%;
Sigma-Aldrich) at 1.5 and 6 atm of argon (99.999%;
Paxair, Danbury, CT, USA) pressure. Some experi-
ments were also carried out without cyclohexene to
verify the product distribution. The cyclohexene de-
compose via the reverse Diels–Alder reaction is char-
acterized by the rate expression [16]
k(cyclohexene → ethylene + 1, 3 butadiene)
= 1.4 × 1015 exp( − 33,500 K/T) s−1
We can then extract the reaction temperature from
the relation
ꢀ
ꢀ
ꢃꢃ
1
−R
−1
[1, 3 − butadiene]
ꢁ
ꢂ
=
ln
ln 1 −
T
Ea
At
cyclohexene
i
∗Disclaimer: Certain commercial materials and equipment are
identified in this paper to specify adequately the experimental proce-
dure. In no case does such identification imply recommendation of
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.
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

288
ROSADO-REYES AND TSANG
rapid decomposition of traces of a labile unknown im-
purity with concentrations of less than 1 or 2%. Note
that by studying reactions under high inhibition con-
ditions, this process cannot contribute to the decom-
position mechanism. The large yields of butenes pre-
vent us from determining the contributions of these
compounds from the sec-butyl radicals and were not
considered for product analysis. We have previously
found that they contribute about 10% to the decompo-
sition of the secondary butyl radicals where the main
process is beta bond scission to form propene. How-
ever, even ignoring the butenes formed in this manner
did not remove the small upward shift of the ethylene
and downward shift of the propene yields (see Fig. 2
below). We have accordingly only analyzed quantita-
tively the results from 1136 K upward, purely arbitrary.
Typical results at various extents of decomposi-
tion are presented in Fig. 1 in terms of the total
product yield and placed on a more quantitative ba-
sis for specific products in Fig. 2. At high extents of
decompositions on a purely molar basis, more prod-
ucts are formed compared with the quantities of 2-
methylbutyraldehyde that has disappeared. Obviously,
there has been some double counting; the breakup of
2-methylbutyraldehyde leads to the presence of more
than one detectable fragment. Thus, for example, the
detection of propionaldehyde suggests the simultane-
ous formation of ethylene via a retro-ene reaction.
As seen in Fig. 3, there are of course multiple chan-
nels that can lead to the formation of more than one sta-
ble fragment. One of these stable fragments is ethylene;
it is the product in the highest concentration. Consider
the situation when an ethyl radical is formed upon the
cleavage of the appropriate C-ethyl bond. The associ-
ated product to this cleavage is the propionaldehyde-2
radical. Beta bond scission will yield acrolein, and this
was indeed detected. Thus acrolein and ethylene can
form from the same 2-methylbutyraldehyde. However,
the acrolein to ethylene ratio is not equal to unity.
The retro-ene reaction will provide equal amounts of
ethylene and propionaldehyde. Here again, the excess
ethylene cannot be accounted for.
Figure 1 Molar product fractions formed from the thermal
decomposition of 2-methylbutyraldehyde.
areas from FID are determined using Agilent Tech-
nologies ChemStation software and converted to mo-
lar quantities based on experimentally determined re-
sponse factors. Retention times and molar FID re-
sponses of all starting substrates and products were
determined from standard samples. The n-alkanes and
n-alkenes were calibrated with commercial samples
(GRACE, Inc.) containing 100 μL/L of each compo-
nent. For other components, calibrations were made
from measured vapor pressures of degassed compo-
nents to prepare in-house mixtures. Experimental re-
sponse factors were determined for ethylene, propene,
n-butane, 1-butene, 2-butene, 1,3-butadiene, and cyclo-
hexene. The estimated standard uncertainty (1σ) in the
analysis of these components is no more than 3%. The
FID sensitivities for acrolein, propionaldehyde, and 2-
methylbutyraldehyde were obtained from previously
measured FID relative molar response factors [17].
RESULTS
The detected products from 2-methylbutyraldehyde
decomposition are ethylene, propene, acrolein, propi-
onaldehyde, n-butane, 1-butene, and 2-butene. Ethy-
lene is the major product, and its yield is particularly
prominent as the temperature is increased. Carbon
monoxide and formaldehyde are stable products that
cannot be detected with our analytical system. Large
quantities of methane and ethane are present. These
can be formed from methyl radicals in the system that
abstract H atoms from the scavenger and their recom-
bination and will not be used in the present analysis.
Considerable amounts of the butene isomers (1-butene
and 2-butene) are also present. They appear at a very
low extent of decomposition. We attribute this to the
There remains the possibility of a 1,2-hydrogen shift
involving the hydrogen atom adjacent to the carbonyl
group. There has been no experimental observation of
this process. Theory has shown that for alkyl radicals,
1,2-hydrogen shifts will require an activation energy
of nearly 170 kJ/mol [18]. This is substantially higher
than the usual competitive beta bond scission reaction
–
–
involving C C and C H bond scission [3]. However,
for alkyl radicals this is essentially a thermoneutral
reaction. In contrast, the binding energy of the car-
–
bonyl C H bond is extraordinarily small, of the order
of 370 kJ/mol. This is much smaller than that of the
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

THERMAL STABILITY OF LARGER CARBONYL COMPOUNDS
289
Figure 2 Fractional yields of product from 2-methylbutyraldehyde decomposition.
Figure 3 Mechanism for the decomposition of 2-methylbutyraldehyde.
–
secondary C H bond of the order of 410 kJ/mol. If this
Once the carbonyl radical is formed, the subsequent
decomposition reaction can be considered to be instan-
taneous. Two ethylene molecules are formed for every
ethyl-propionaldehyde-2 bond that is broken, and this
will in fact bring about a satisfactory mass balance.
Note that the data in Fig. 1 become insensitive to these
effects at a low extent of reaction. In the case of a very
difference is directly transferable to the activation ener-
gies of H atom transfer, then it will bring the subsequent
radical destruction process into the same range of val-
–
–
ues as beta C H and C C cleavages. Note that the sub-
sequent decomposition of the carbonyl compound has
an extremely small barrier of less than 55 kJ/mol [3].
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

290
ROSADO-REYES AND TSANG
small extent of reaction, the mass balance considera-
tions discussed here will be inapplicable.
–
Aside from C C bond cleavage yielding an ethyl
radical, the ejection of a methyl grouping will leave
a butyraldehyde-2 radical. Beta bond cleavage of a
–
C C bond will lead to acrolein. The latter is a more
–
likely process than the ejection of a C H bond from the
propionaldehyde-2 radical since there is an activation
barrier difference of at least 25 kJ/mol. This process
will only lead to acrolein formation and therefore will
not help in reducing the excess products that have been
detected. Thus the ejection of ethyl and methyl can
both lead to acrolein formation and the issue is then
the partitioning of the acrolein between the sources.
However, there are not sufficient data to decide on the
proper data partitioning and analysis. For the present
purposes, it will be assumed that ethyl ejection will
lead to another ethylene being formed as a consequence
of 1,2 hydrogen shift and methyl ejection is the sole
source of acrolein. This assumption leads to an ethyl
and methyl rate constants for ejection of a factor of two
difference and appears to be somewhat small based on
hydrocarbon data. It exactly balances the measured
mole fractions formed and destroyed. The partitioning
of the methyl and ethyl ejection channels will lead to
the possibility of some untabulated channels.
The data in Fig. 2 show the fractional yield for the
particular product channel of interest. The results at
lower temperatures depart from linearity, showing a
smaller ethylene yield and larger yields of the other
products. A simple explanation is that at the lower
temperatures the two ethylenes are no longer produced
following the cleavage of the ethyl-propionaldehyde-2
bond. The general mechanism is shown in Fig. 3.
The most interesting product is undoubtedly n-
butane. This is the first time that we have detected a
larger alkane in all the studies carried out in our single
pulse shock tube experiments. n-Butane is indicative
of a direct molecular reaction leading to the forma-
tion of carbon monoxide. Nevertheless, it is present
in much smaller quantities than the olefins. Because
of the isomeric butenes that are formed at the lower
temperature we have assumed that for every propene
molecule that is detected, approximately a further 10%
Figure 4 Rate constants as a function of temperature for the
processes of interest. Ethyl refers to the process that yields
ethylene (reaction (2)). Methyl deals with reactions that yield
acrolein (reaction (3)). The retro-ene reaction involves direct
formation of propionaldehyde (reaction (1)). The propene is
derived from the sec-butyl radical (reaction (4)), and butane
is the roaming reaction that accompanies the bond breaking
(reaction (5)). Circles are reactions at 3 atm, triangles are
for reactions at 6 atm, and squares are for processes at 1.5
atm. Rate measurements are arbitrarily truncated at 1180 K
due to reasons discussed in the text. Rate measurements are
truncated at 1135 K due to reasons discussed in the text.
where
k(R1) = [propionaldehyde]retroene reaction
k(R2) = [ethylene − propionaldehyde]/2
k(R3) = [acrolein]
–
of the sec butyl radical is formed from the hydrogen-
ejection channel, a ratio that we have found earlier
when a compound such as 3,4-dimethylhexane is de-
composed in single pulse shock tube experiments [12].
Figure 4 contains Arrhenius plots of rate constants
as a function of temperature. We follow the standard
procedure for treating multichannel unimolecular de-
composition processes. Specifically, we define
k(R4) = [s − butyl + formyl] = 1.1 × [propene]
where we have multiplied the propene yield by the
factor of 1.1 to take into account the butane isomers
formed during the course of s-butyl radical decompo-
sition as determined from earlier studies:
k(total) = k(R1) + k(R2) + k(R3) + k(R4) + k(R5)
k(R5) = [n − butane]
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

THERMAL STABILITY OF LARGER CARBONYL COMPOUNDS
291
Table II Summary Data on the Various Processes Involved in the Decomposition of 2-Methylbutyraldehyde in Single
Pulse Shock Tube Experiments
Rate Constant (s−1
(1150 K)
)
Reaction
Rate Expression (s−1
)
0.14
R1: 2-me-butyraldehyde → ethylene + propionaldehyde
R2: 2-me-butyraldehyde → ethyl + n-propionaldehyde-2
R3: 2-me-butyraldehyde → methyl + n-butyraldehyde-2
R4; 2-me-butyraldehyde → s-butyl + formyl
1013.01
1016.35
1015.85
1015.32
1014.83
exp(−31,540 200 K/T)
exp(−38,830 310 K/T)
exp(−38,500 300 K/T)
12.3
49.9
20.5
34.5
2.39
0.04
2
1
1.5
0.06
0.11
0.15
0.1
exp(−36,540 300 K/T)
0.08
R5: 2-me-butyraldehyde → n-butane + CO
exp(−38,400 230 K/T)
This furnishes a basis for comparison with the data summarized in Table I.
k(R5) is the roaming reaction and should be in-
cluded with the rate constant for s-butyl radical for-
mation as a total rate constant for the breaking of the
secondary butyl–formyl bond.
extremely small directly due to the internal standard
technique. Overall, they lead to absolute uncertain-
ties of approximately a factor of 2 in the A factor and
6 kJ/mol in the activation energy. However, practically
all uncertainties in rate constants and expressions arise
from mechanistic ambiguities. The present discussion
is fully indicative of these problems.
From the experimental data, we obtain
k(total) = −ln(1 − [totalproducts/
initial 2 − methylbutyraldehyde])/t
DISCUSSION
where t is the residence time of 500 μs and the total
products include ethylene, propionaldehyde, acrolein,
propene, and n-butane.
A large number of products are formed during 2-
methylbutyraldehyde decomposition in a single pulse
shock tube. The analysis of the results is helped by the
detection of smaller aldehydes. Except for the direct
formation of butane, the rate expressions and constants
obtained here can be directly compared with values in
Table I. Specifically, there does not appear to be any
drastic change in the rate expressions or constants be-
tween the aldehydes and the ketones.
There is somewhat larger spread of A factors than
that in the data included in Table I. For the bond cleav-
age reactions, this is undoubtedly due to the formation
of the butenes at a low extent of conversions that we
have ignored in the treatment of the bond cleavage re-
actions. More support for this assumption is furnished
by the analogous heat of formation for the formyl radi-
cal described below. It is of course well known that
rate constants are more accurately determined than
their temperature dependence (slope). More valid di-
rect comparisons can be made in terms of the rate con-
stants at a common temperature. These are the actual
experimental values that are directly measured. This
is equivalent to comparing equilibrium constants as
opposed to the thermodynamic quantities that may be
inferred. However, even on this finer basis, the general
trends are very similar. Unfortunately, except for the
The individual rate constants can then be obtained
from the relation
k(process)j/k(total) = concentrationj/
total product concentration,
where the subscript j refers to the individual processes
described above.
The Arrhenius plots of the five individual processes
are included in Fig. 4. The reader is reminded that the
acrolein yield is assigned to the formation of methyl
radical and is probably an overestimate because methyl
will also be released during the formation of propene.
Most of the above discussion center about the assump-
tions that must be made with regard to the ethylene
that is formed. They thus impinge on the reactions in-
volved in the ejection of methyl and ethyl radicals.
They do not necessarily impinge on the processes that
deal with the breaking of the sec–butyl–formyl bond or
the retro-ene reaction that leads to the direct formation
of propionaldehyde and ethylene. However, because of
the apparent drop in formation of ethylene and the cor-
responding increase in the relative mole fractions of the
other products, the rate expressions may be effected.
–
The rate expressions that we have obtained are
given in Table II. We have used the same format as in
Table I to facilitate comparisons. The uncertainties in-
cluded are the result of the least square analysis and are
C C bond cleavages of the formyl-s-butyl and acetyl-
butyl bonds, exact comparisons cannot be made. The
rate constant for formyl and acetyl substitution leads to
values that are virtually unchanged. Apparently, methyl
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

292
ROSADO-REYES AND TSANG
substitution has minimal effects on rate constants of
carbonyl compounds. This is somewhat equivalent in
going from an ethyl to isopropyl situation in the alka-
nes where the rate constants change only minimally.
Comparison of the carbonyl compounds to the alka-
nes indicates the same minimal effects. In the case of
ketones, there is a decrease of a factor of 1.5 in rate
constants which is not observed for the aldehyde. It is
indeed somewhat surprising that such drastic structural
change such as going from a methylene to a carbonyl
structure should lead to such small changes. There ap-
pears to be a very simple rate rule in going from alka-
nes to carbonyl compounds for the bonds adjacent to
the carbonyl compound. The retroene reactions show
larger effect with the rate constants for carbonyl com-
pounds being considerably smaller. Intuitively, this is
consistent with smaller resonance energy.
For the decomposition of the sec–butyl–acetyl
(Table I) and sec–butyl–formyl (Table II) bonds, there
appears to be an order of magnitude difference in the A
factor and a 13 kJ/mol difference in the activation en-
ergies. However, the rate constants differ by less than
a factor of 1.5. The question naturally arises whether
this is due to a compensation effect arising from the
problems described earlier for measuring the rate con-
stant as a function of temperature. For this purpose,
we can use our results to determine the heat of forma-
tion of the formyl radical and use this as a check of
the derived rate expression. Assuming that the reverse
combination reaction between formyl and the s-butyl
radical has no activation energy, then
so many different systems should be very close to
each other. Indeed, the formyl group probably has
characteristics of a diatomic species that would be
expected to have smaller A factors. Apparently, the
magnitude for such effects is not large. Indeed, in the
alkenes it is possible to recognize the effect of vinyl
substitution and consequently allylic resonance by
roughly a factor of three smaller A factors. There is no
such evidence in the case of the carbonyl compound.
The presence of normal butane provides evidence
for the presence of a roaming channel in the decom-
position of a larger polyatomic molecule. It is in such
a small amount that for practical purposes, it will not
make much of a contribution to a high-temperature
system. Note that we do not report on the other possi-
ble roaming product, the abstraction of hydrogen atom
from the s–butyl radical by the formyl radical. We do
note that at the highest temperatures the butene iso-
mers are slightly in excess (factor of 2 ) than expected,
roughly 10% of the propene yield in other s-butyl sys-
tems. This cannot be a major channel. In the present
analysis, the s-butyl-formyl reaction channel has been
treated at the sole bond-breaking process. Presumably,
the “roaming” reactions should be included. It is, how-
ever, such a small component and thus will not affect
the conclusions drawn above.
Finally, we note that although there does not appear
to be large contributions for a carbonyl resonance en-
ergy, our results do not rule out a small contribution
of the order of 10 kJ/mol. This will probably be man-
ifested in a smaller activation energy for the relevant
–
C C bond.
ꢀH = ꢀE + RT = 305 + 9.6
= 314.6 kJ/mol at 1150 K
BIBLIOGRAPHY
From the heat of formation of 2-
methylbutyraldehyde of −257.3 kJ/mol [19] at
1150 K, a heat of formation of the secondary
butyl radical of 39.5 kJ/mol [5] at 1150 K, and
the bond dissociation energy (BDE) (s-butyl-H) =
414.2 kJ/mol [20], we find the heat of formation at
1150 K for the formyl radical to be 17.8 kJ/mol.
Correcting to room temperature, this leads to the value
of 23 kJ/mol. The well-established heat of formation
of formyl radical is 42 kJ/mol. Our results deviates
by 19 kJ/mol, and if this arises from an error in the
slope of the determination of the activation energy
the results would require raising the A factor to 1.6 ×
10 [15]. This adjustment will place our formyl radical
heat of formation well within the range of values
for the ketones and alkanes tabulated in Table I and
numerous other studies on alkane decomposition.
There is of course no reason that the A factors for
1. Tsang, W. Annu Rev Phys Chem 1990, 41, 559.
2. Tsang, W. Int J Chem Kinet 1978, 10, 821.
3. Tsang, W. In Energetics of Organic Free Radicals; Mar-
tinho Simoes, J. A., Greenberg, A., Liebman, J. F., Eds.;
Blackie Academic and Professional: London, 1996;
pp. 22.
4. McMillen, D. F.; Golden, D. M. Annu Rev Phys Chem
1982, 33, 493.
5. Johnson, R. D., III. Available at http://cccbdb.nist.gov/.
NIST Standard Reference Database Number 101: ac-
cessed January 2013.
6. Rosado-Reyes, C. M.; Tsang, W.; Alecu, I. M.; Mer-
chant, S. S.; Green, W. H. J Phys Chem A 2013, 117,
6724.
7. Rosado-Reyes, C. M.; Tsang, W. J Phys Chem A 2013,
117, 10170.
8. Tsang, W. Int J Chem Kinet 1984, 16, 1543.
9. Bowman, J. M.; Shepler, B. C. Ann Rev Phys Chem
2011, 62, 531.
International Journal of Chemical Kinetics DOI 10.1002/kin.20828

THERMAL STABILITY OF LARGER CARBONYL COMPOUNDS
293
10. Sivaramakrishnan, R.; Michael, J. V.; Harding, L. B.;
Klippenstein, S. J. J Phys Chem A 2012, 116, 5981.
11. Gibian, M. J.; Corley, R. C. Chem Rev 1973, 73, 441.
12. Tsang, W. J Phys Chem 1972, 76, 143.
13. El-Nahas, A. M.; Bozzelli, J. W.; Simmie, J. M.;
Navarro, M. V.; Black, G.; Curran, H. J. J Phys Chem A
2006, 110, 13618.
14. Tsang, W. Int J Chem Kinet 1978, 10, 1119.
15. Rosado-Reyes, C. M.; Tsang, W. J Phys Chem A 2012,
116, 9599.
17. Schofield, K. Prog Energy Combust Sci 2008, 34, 330.
18. Davis, A. C.; Francisco, J. S.J Phys Chem AJ Phys Chem
A 2011, 115, 2966.
19. Kroenlein, K.; Muzny, C. D.; Kazakov, A. F.; Diky, V.;
Chirico, R. D.; Magee, J. W.; Abdulagatov, I.; Frenkel,
M. NIST/TRC Web Thermo Tables (WTT), NIST Stan-
dard Reference Subscription Database 3—Professional
Edition, Version 2-2012-1-Pro, 2012.
20. Stull, D. R.; Prophet, H.; Office of Standard Reference
Data; National Bureau of Standards: Washington, DC,
1971.
16. Tsang, W. Int J Chem Kinet 1973, 5, 651.
International Journal of Chemical Kinetics DOI 10.1002/kin.20828