Thermal decomposition of quinoline and isoquinoline. The role of 1-indene imine radical

Article abstract of DOI:10.1021/jp9729831

The thermal reactions of quinoline and isoquinoline were studied behind reflected shock waves in a pressurized driver single pulse shock tube over the temperature range 1275-1700 K and densities of ～3 ×10^-5 mol/ cm³. The decomposition products found in the postshock mixtures of quinoline and isoquinoline and their production rates were identical for both isomers. They were C₂H₂, C₆H₅CN, HC≡CCN, C₆H₆, HCN, C₆H₅-C≡CH and C₄H₂. Trace quantities of C₆H₄, C₅H₅N and C₅H₄N-C≡CH were also found. The total disappearance rates of quinoline and isoquinoline are the same, and in terms of a first-order rate constant they are given by k_total = 10^13.0exp(-75.5 × 10³/RT) s^-1 where R is expressed in units of cal/(K mol). The same product distribution in the two isomers can be accounted for if the production of 1-indene imine radical as an intermediate is assumed. A kinetic scheme containing the reactions of both quinoline and isoquinoline with 72 species and 148 elementary reactions accounts for the observed product distribution. The reaction scheme is given, and the results of computer simulation and sensitivity analysis are shown.

Full text of DOI:10.1021/jp9729831

928
J. Phys. Chem. A 1998, 102, 928-946
Thermal Decomposition of Quinoline and Isoquinoline. The Role of 1-Indene Imine Radical
Alexander Laskin† and Assa Lifshitz*
Department of Physical Chemistry, The Hebrew UniVersity, Jerusalem 91904, Israel
ReceiVed: September 12, 1997; In Final Form: NoVember 13, 1997
The thermal reactions of quinoline and isoquinoline were studied behind reflected shock waves in a pressurized
driver single pulse shock tube over the temperature range 1275-1700 K and densities of ∼3 × 10-5 mol/
cm3. The decomposition products found in the postshock mixtures of quinoline and isoquinoline and their
production rates were identical for both isomers. They were C2H2, C6H5CN, HCtCCN, C6H6, HCN, C6H5-
CtCH, and C4H2. Trace quantities of C6H4, C5H5N and C5H4N-CtCH were also found. The total
disappearance rates of quinoline and isoquinoline are the same, and in terms of a first-order rate constant
they are given by ktotal ) 1013.0exp(-75.5 × 103/RT) s-1 where R is expressed in units of cal/(K mol). The
same product distribution in the two isomers can be accounted for if the production of 1-indene imine radical
as an intermediate is assumed. A kinetic scheme containing the reactions of both quinoline and isoquinoline
with 72 species and 148 elementary reactions accounts for the observed product distribution. The reaction
scheme is given, and the results of computer simulation and sensitivity analysis are shown.
I. Introduction
mechanism were given. The thermal decompositions of benzene
and pyridine, on the other hand, have been extensively studied,
discussed, and summarized.9-11 They are essential for interpret-
ing the results of the present investigation.
The present investigation on the decomposition of quinoline
and isoquinoline is a continuation of our investigation on the
decomposition of nitrogen containing aromatic compounds. It
is well-known1-3 that fuel nitrogen in coal is predominantly
pyrrole and pyridine type nitrogen, with a slightly higher weight
to compounds containing the pyrrole ring.3 The study of the
pyrolysis of benzopyrrole (indole) and benzopyridine (quinoline)
as the basic nitrogen containing components of the coal matrix
is an important element in the understanding coal combustion.
We have recently published a detailed investigation of the
thermal reactions of indole.4 These reactions were found to be
similar to those of pyrrole, and as far as the reactions of the
pyrrole ring are concerned, they could be predicted from the
mechanism of the pyrolysis of pyrrole.5,6 The thermal reactions
of indole are mostly isomerizations at low temperatures (low
conversions) and as the temperature increases fragmentations
take over. The assumption of the existence of an indole T
indolenine tautomerism was necessary to explain the production
of the isomerization products.
The initiation step in the thermal decomposition of benzene
involves an ejection of a hydrogen atom from one of the six
C-H bonds in the molecule, forming phenyl radical and a
hydrogen atom.9,11 The fragmentation of the phenyl radical is
preceded by a â-scission of the ring forming a linear chain
•
l-C6H5 radical. This radical decomposes via several reaction
channels yielding mainly C2H2 and C4H2 as final products.
The mechanism of pyridine decomposition is very similar to
that of benzene. It is somewhat more complicated, however,
owing to the decreased symmetry in pyridine as compared to
benzene. As a result of the H-atom ejection from the pyridine
ring, ortho, meta, and para pyridyl radicals can be formed. The
thermochemistry of the pyridyl radicals shows that the produc-
tion of ortho pyridyl radical is preferred, owing to its stabiliza-
tion by the neighboring nitrogen. Ab initio calculations12 give
a value of 6 kcal/mol for the stabilization energy.
In the present article, we report on a study of the pyrolysis
of quinoline and isoquinoline. Both molecules contain a
pyridine ring fused to benzene. The nitrogen atom occupies
an R-position in quinoline and a â-position in isoquinoline. Both
molecules have approximately the same thermal stability, their
standard heats of formation7 being 52.2 and 50.2 kcal/mol,
respectively.
We are aware of only one study on the thermal decomposition
of quinoline, by Bruinsma et al.,8 who reported on its overall
decomposition rate and its Arrhenius parameters. No data on
the distribution of reaction products or on the decomposition
† In partial fulfillment of the requirement of a Ph.D. Thesis to be
submitted to the Senate of the Hebrew University of Jerusalem by A. Laskin.
S1089-5639(97)02983-6 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/22/1998

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 929
The â-scissions of the various pyridyl radicals lead to five
different open ring structures with different locations of the
nitrogen in the open chain. The thermochemistry of the
â-scission of the ortho radical is quite different from those of
the meta and para radicals. Out of the five possible â-scis-
sion channels only one forms the strong -CtN bond. This is
the C-N bond scission in the ortho pyridyl radical (channel
b). This scission makes the open-chain radical much more
stable than all other radicals obtained from the meta and para
pyridyls.
Therefore, the formation of the ortho pyridyl radical in the
first step, as well as the â-scission of its C-N bond in the second
step, is the preferred pathway for pyridine decomposition. The
decomposition of the open chain cyano radical leads mainly to
the production of C2H2 and HCtCCN.
Since quinoline and isoquinoline are built from these two
rings, it is reasonable to assume that similar reaction pathways
operate in their thermal decompositions. In this report we
describe the thermal reactions of quinoline and isoquinoline,
compose a reaction scheme for the overall decomposition and
describe the results of computer simulations that support the
suggested mechanism. It will be shown that the product
distribution can be accounted for only if production of 1-indene
imine radical as an intermediate is assumed.
The thermal reactions of quinoline and isoquinoline are
initiated by an H-atom ejection from the pyridine ring. In both
molecules the H atom in the ortho position to the nitrogen is
the preferred ejection site owing to the resonance stabilization
of the remaining radical. It will be shown later that, following
the preferred â-scission of the C-N bond, the main decomposi-
tion products in quinoline are expected to be HCtCCN and
C6H6 and in isoquinoline C2H2 and C6H5CN. However, the
experimental results show that the concentrations of all the
decomposition products are absolutely the same irrespective of
whether the original reactant is quinoline or isoquinoline. To
describe this fact we assume the existence of 1-indene imine
radical intermediate which plays the central role in the thermal
decomposition of quinoline and isoquinoline.
Figure 1. Gas chromatograms of a postshock mixture of 0.3%
quinoline in argon heated to 1590 K, taken with FID and NPD.
Isoquinoline is absent from the products. The broken-line peak shows
its retention time on the column relative to quinoline. The numbers in
parentheses indicate multiplication factors.
tank was connected to the driven section at 45° angle toward
the driver near the diaphragm holder in order to prevent
reflection of transmitted shocks and in order to reduce the final
pressure in the tube. The driven section was separated from
the driver by Mylar polyester films of various thickness
depending upon the desired shock strength.
Prior to performing an experiment, the tube and the gas-
handling system were pumped down to ∼3 × 10-5 torr. The
reaction mixtures were introduced into the driven section
between the ball valve and the end plate, and pure argon was
introduced into the section between the diaphragm and the valve,
including the dump tank. After each experiment a gas sample
was transferred from the tube through a heated injection system
to a Hewlett-Packard model 5890A gas chromatograph operating
with flame ionization (FID) and nitrogen phosphor (NPD)
detectors.
The temperatures and densities of the gas behind the reflected
shocks were calculated from the measured incident shock
velocities using the three conservation equations and the ideal
gas equation of state. The molar enthalpies of quinoline and
isoquinoline were taken from the NIST-Structures and Properties
Code.7 Incident shock velocities were measured with two
miniature high-frequency pressure transducers (P. C. B. model
113A26) located 230 mm apart near the end plate of the driven
section. The signals generated by the shock wave passing over
the transducers were fed through an amplifier to a Nicolet model
3091 digital oscilloscope. Time intervals between the two
signals shown on the oscilloscope were obtained digitally with
an accuracy of 2 ms, corresponding to approximately 20 K. A
third transducer placed in the center of the end plate provided
II. Experimental Section
Apparatus. The thermal decompositions of quinoline and
isoquinoline were studied behind reflected shocks in a pressur-
ized driver, 52-mm i.d. single-pulse shock tube made of Double
Tough Pyrex tubing. The tube and its gas-handling system were
maintained at 150 ( 1 °C with a heating system with 15
independent computer-controlled heating elements.
The shock tube had a 4-m long driven section divided in the
middle by a 52-mm i.d. ball valve. The driver section had a
variable length up to a maximum of 2.7 m and could be varied
in small steps in order to obtain the best cooling conditions.
Cooling rates were approximately 5 × 105 K/s. A 36-L dump

930 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
TABLE 1: Experimental Conditions and Product Distribution in Postshock Mixtures of Quinolinea
T5 (K)
dwell time (ms)
C5 × 105/cm3
quinoline
C2H2
C6H5-CN
HCtCCN
C6H6
HCN
C6H5CtCH
C4H2
1290
1295
1310
1320
1325
1335
1335
1362
1365
1385
1385
1410
1410
1415
1445
1457
1465
1490
1490
1500
1500
1535
1550
1550
1565
1570
1607
1650
2.00
1.96
2.02
2.03
2.02
2.00
2.00
2.04
2.00
2.00
2.00
2.00
1.99
2.0
2.00
1.94
2.00
2.00
2.00
1.99
1.96
2.00
1.96
1.93
1.95
1.99
1.96
1.98
2.50
2.56
2.31
2.49
2.30
2.45
2.56
2.57
2.45
2.40
2.59
2.40
2.27
2.33
2.37
2.53
2.37
2.53
2.45
2.18
2.12
2.25
2.20
2.15
2.25
2.07
2.14
2.09
99.29
99.46
99.20
99.14
99.03
98.48
98.23
98.07
97.42
97.71
97.15
93.28
86.05
94.62
90.46
91.13
87.10
77.77
77.20
71.22
66.77
68.89
52.66
39.99
39.94
23.94
16.65
12.85
0.45
0.31
0.42
0.45
0.52
0.85
0.89
1.00
1.38
1.34
1.54
3.35
6.72
2.92
4.47
3.79
6.43
10.00
11.10
12.41
14.21
13.65
20.87
28.17
24.76
33.57
46.29
44.00
0.15
0.13
0.18
0.22
0.22
0.27
0.34
0.41
0.40
0.38
0.46
1.15
1.49
0.75
1.24
1.36
1.31
1.95
1.71
2.72
2.88
2.34
3.96
4.32
3.91
3.37
2.98
2.16
0.029
0.018
0.023
0.030
0.036
0.10
0.15
0.083
0.16
0.12
0.16
0.34
0.74
0.28
0.61
0.35
0.85
1.11
1.40
1.55
1.68
2.15
2.27
2.55
2.66
4.22
3.23
2.75
0.021
0.011
0.030
0.056
0.046
0.041
0.061
0.13
0.18
0.06
0.21
0.60
1.98
0.18
1.08
1.48
1.29
3.52
2.41
5.23
6.38
5.41
9.67
10.43
12.90
16.37
18.57
19.30
0.059
0.016
0.048
0.033
0.043
0.073
0.070
0.081
0.094
0.15
0.12
0.34
0.61
0.31
0.37
0.38
0.52
0.72
0.92
1.15
1.37
1.14
1.68
2.11
1.79
1.75
1.87
1.92
0.053
0.10
0.072
0.073
0.14
0.14
0.19
0.19
0.15
0.22
0.52
1.01
0.47
0.66
0.66
0.84
0.90
1.23
1.71
1.99
1.39
1.87
2.50
1.65
1.67
1.27
0.92
0.033
0.038
0.065
0.154
0.166
0.085
0.13
0.41
1.42
0.48
1.11
0.86
1.65
4.02
4.03
4.00
4.70
5.05
7.02
9.93
10.77
15.10
9.14
16.09
a In mole percent. H2, C6H4, C5H5N, and C5H4N-CtCH are not included in the table.
measurements of the reaction dwell time (approximately 2 ms)
with an accuracy of (5%.
prior to decomposition (eq 3) where A(pri)t is the peak area of
a product i in the shocked sample, S(pri) is its sensitivity relative
to that of the reactant, N(pri) is the number of its carbon atoms,
F5/F1 is the compression behind the reflected shock, and T1 is
the initial temperature, which was 423 K in the present series
of experiments.
The FID and NPD sensitivities of the products relative to
the reactant were determined from standard mixtures. They
were estimated for C4H2, HCtC-CN, and C5H4N-CtCH,
based on comparisons to the relative sensitivities of similar com-
pounds. GC peak areas were recorded with a Spectra Physics
model SP4200 computing integrator and transferred after each
analysis to a PC for data reduction and graphical presentation.
Materials and Analysis. Reaction mixtures containing 0.3%
quinoline (or isoquinoline) diluted in argon, were prepared in
12 L glass bulbs and stored at 150 ( 1 °C and 700 torr. Both
the bulbs and the line were pumped down to approximately 10-5
torr before the preparation of the mixtures. 2.1 torr (0.3% of
700 torr) of quinoline or isoquinoline in the gas phase
correspond approximately to 5% of their equilibrium vapor
pressure at 150 °C, so that wall condensation is negligible.
Quinoline, listed as 98.0% pure, and isoquinoline, 97.0% pure,
were obtained from Aldrich Chemical Co. Each showed only
one GC peak. The argon used was Matheson ultrahigh purity
grade, listed as 99.9995%, and the helium was Matheson pure
grade, listed as 99.999%. All materials were used without
further purification.
III. Results
In order to determine the product distribution in the thermal
decompositions of quinoline and isoquinoline, some 30 experi-
ments were carried out with quinoline as a reactant molecule
and 20 with isoquinoline. The temperature range covered in
these experiments was 1275-1700 K at overall densities behind
the reflected shock wave of ∼3 × 10-5 mol/cm3.
Gas chromatographic analyses of the postshock mixtures were
performed on two 1-m Porapak N columns with flame ionization
and nitrogen phosphor detectors. Identification of reaction
products was based on their GC retention times assisted by a
Hewlett-Packard model 5970 mass selective detector. Typical
FID and NPD chromatograms of a mixture containing 0.3%
quinoline in argon, shock-heated to 1520 K, are shown in
Figure 1.
Details of the experimental conditions and the product
distribution are given in Tables 1 and 2 for quinoline and
isoquinoline, respectively. The mole percents given in the tables
correspond to those of the products in the postshock mix-
tures irrespective of the number of their carbon atoms. Mo-
lecular hydrogen was not measured and was not included in
the tables.
The concentrations of the reaction products C5(pri) were
calculated from their GC peak areas using the relations:13
C5(pri)) A(pri)/S5(pri) × {C5(reactant)0/A(reactant)0} (1)
C5(reactant)0 ) p1{%(reactant)F5/F1}/100RT1
(2)
The decomposition products found in the postshock mixtures
were identical in the two series of experiments, both in
distribution and production rates. No quinoline T isoquinoline
isomerization occurred. Neither traces of isoquinoline in shock-
heated mixtures of quinoline nor traces of quinoline in shock-
heated mixtures of isoquinoline were found. In both series, GC
analyses revealed the presence of C2H2, C6H5-CN, HCtC-
CN, C6H6, HCN, C6H5-CtCH, and C4H2 as major decomposi-
A(reactant) ) A(reactant) + (1/9) × N(pr )A(pr ) /S(pr )
∑
0
t
i
i t
i
(3)
In these relations, C5(reactant)0 is the concentration of
quinoline or isoquinoline behind the reflected shock prior to
decomposition and A(reactant)0 is their calculated GC peak area

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 931
TABLE 2: Experimental Conditions and Product Distribution in Postshock Mixtures of Isoquinolinea
T5 (K) dwell time (ms) C5 × 10-5 mol/cm3 isoquinoline C2H2 C6H5-CN HCtCCN C6H6
HCN
C6H5-CtCH
C4H2
1275
1300
1335
1345
1360
1365
1365
1400
1415
1440
1460
1465
1480
1490
1505
1515
1550
1590
1600
1600
1630
1675
1700
2.00
2.04
2.03
2.00
2.04
2.04
2.01
2.00
2.00
2.00
2.00
1.99
2.00
1.99
1.99
1.98
1.99
2.00
1.98
1.99
1.98
1.97
1.98
2.48
2.62
2.73
2.65
2.38
2.58
2.50
2.51
2.33
2.34
2.30
2.28
2.26
2.25
2.10
2.37
1.97
2.28
2.53
2.57
2.23
2.14
2.00
99.56
99.09
98.06
97.90
98.06
97.05
96.96
91.01
89.01
89.95
89.74
78.28
79.70
76.62
66.25
61.12
56.02
30.80
26.21
22.34
22.53
17.62
12.47
0.23
0.48
0.075
0.12
0.25
0.30
0.27
0.35
0.36
1.21
1.22
1.02
1.02
1.93
1.66
2.29
2.87
3.27
2.92
4.01
4.74
4.87
3.56
2.20
2.06
0.012
0.040
0.087
0.11
0.095
0.16
0.17
0.55
0.72
0.65
0.70
1.63
1.68
1.57
2.27
2.68
2.99
4.98
4.78
4.77
5.40
4.93
5.02
0.033
0.060
0.12
0.12
0.10
0.15
0.16
0.42
0.55
0.49
0.50
0.94
0.91
1.28
1.60
1.52
1.66
2.12
2.72
2.60
2.08
1.66
1.39
0.021
0.055
0.11
0.12
0.13
0.25
0.32
0.82
0.94
1.30
1.31
3.52
4.12
3.44
7.43
7.26
11.73
20.21
20.51
22.09
23.22
25.66
22.01
0.065
0.14
0.29
0.32
0.32
0.39
0.45
1.11
1.51
1.16
1.08
2.08
1.72
2.28
2.72
3.20
2.27
2.85
3.40
2.53
2.32
1.17
1.10
1.01
1.04
0.94
1.52
1.42
3.58
5.21
4.51
4.53
9.31
8.04
9.85
13.10
16.38
17.16
25.59
27.25
30.60
29.74
34.58
37.82
0.057
0.067
0.079
0.11
0.15
0.41
0.84
0.93
1.11
2.31
2.15
2.67
3.75
4.57
5.24
9.44
10.39
10.19
11.15
12.23
18.12
a In mole percent. H2, C6H4, C5H5N, and C5H4N-CtCH are not included in the table.
Figure 3. Nitrogen-carbon mass balance among the decomposition
products. 0, quinoline; 9, isoquinoline.
Figure 2. Recovery of reaction products tested in several runs with
quinoline against the reactant loss, using 0.3% xenon in mixtures
containing 0.3% quinoline in argon as an internal standard.
all the decomposition products, each multiplied by the number
of its carbon atoms. The 45° line in the figure represents a
complete nitrogen-carbon mass balance. As can be seen,
within the limit of the experimental scatter, there is no significant
deviation from a complete mass balance.
tion products. Trace quantities of C6H4, C5H5N, and C5H4N-
CtCH were also found in the postshock mixtures, particularly
in high-temperature shocks.
The possible loss of products by adsorption on the shock tube
walls or in the gas chromatograph was examined in the
following manner. In several tests, 0.3% xenon was added to
the reaction mixture and ratios [quinoline]/[xenon] were mea-
sured by GC-MS (using 30 cm Porapak N columns) in the
unshocked and in several shocked samples. By comparing these
ratios in the shocked and the unshocked samples, we could
determine the relative loss of the reactant in each test. This
relative loss was compared to the corresponding values calcu-
lated by the sum (1/9)∑N(pri)A(pri)t/S(pri), which represents the
concentrations of all the products normalized by the number of
their carbon atoms (see Experimental Section, eq 3). These
comparisons are shown graphically in Figure 2. Except for two
tests, the ratios are practically the same, showing scatter in both
directions. It seems therefore that there is no significant loss
of products in the analyses.
Figure 4 shows an Arrhenius plot of the overall decomposition
of quinoline and isoquinoline calculated as first-order rate
constants from the relation
ktotal ) -ln([reactant]t/[reactant]0)/t
(4)
As can be seen the rates for quinoline and isoquinoline are
practically the same. The value obtained (for the two molecules)
is ktotal ) 1013.0 exp(-75.5 × 103/RT) s-1 where R is expressed
in units of cal/(K mol). The value obtained by Bruinsma et
al.8 for the total decomposition of quinoline in a low-temperature
study is also shown for comparison. There is a good agreement
between the values of the rate constants in the two studies on
the basis of an extrapolation of the low-temperature data to the
data of the present high-temperature study.
IV. Discussion
The mass balance of nitrogen vs carbon is shown in Figure
3. The concentrations of the nitrogen containing species are
plotted against one-ninth of the sum of the concentrations of
1. Reaction Mechanism. A. Formation of Quinolyl and
Isoquinolyl Radicals. Similar to other aromatic systems, it is

932 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
Similar representations were introduced for the isoquinolyl
radicals. The R-ortho and â-ortho isoquinolyl radicals are
considered to form by H-atom ejection or abstraction.
The meta isoquinolyl radical in the pyridine ring then forms
from the â-ortho isoquinolyl radical by fast equilibration.
Figure 4. First-order Arrhenius plot of the overall decomposition of
quinoline and isoquinoline. 0 and broken line, quinoline; 9 and solid
Quinolyl and isoquinolyl radicals result also from H-atom
abstractions. At relatively low temperatures, H atoms are
abstracted mainly by other hydrogen atoms. As the temperature
increases and the concentrations of additional radicals in the
system build up, hydrogen atoms are also abstracted by less
line, isoquinoline. The low-temperature rate constant (ktotal ) 1013.2
-
exp(-77.9 × 103/RT), s-1) obtained by Bruinsma et al.8 is shown for
comparison.
assumed here too that the initiation step in the thermal decom-
positions of quinoline and isoquinoline involves ejection of a
hydrogen atom from the reactant by C-H bond cleavage. There
are seven such bonds in each molecule, three in the pyridine
ring, and four in the benzene ring. Thus, seven different radicals
can in be formed as a result of H-atom ejection from each
molecule. Each one can then decompose via its own decom-
position pathway.
•
reactive radicals such as C6H5 , C6H4CN•, C5H4N•, and others.
H-atom abstractions by different radicals (R•) were introduced
into the reaction scheme according to our formalism by forming
two quinolyl and three isoquinolyl radicals. Forty nine such
abstractions appear in the reaction scheme.
To simplify the overall decomposition mechanism and re-
duce the number of elementary reactions in the final kinetic
scheme, we made two assumptions: 1. Since all four hydro-
gen atoms on the benzene ring are bound with the same bond
energy (∼112 kcal/mol), we have used only one radical in the
benzene ring instead of four.
27
2. Meta and para quinolyl radicals in the pyridine ring are
less stable than the ortho radical and should therefore be con-
sidered as separate entities. However, to reduce the number of
elementary reactions that produce these radicals, particularly
abstraction reactions, we assumed that the only radical formed
either by H-atom ejection or abstraction is the ortho quinolyl
radical, which then quickly equilibrates with the meta and para
radicals.
B. Preferred Decomposition Pathways. It has already been
noted in the Introduction that the preferred pathway for the
decomposition of the pyridine ring involves a H-atom ejection
from the ortho position followed by â-scission of the C-N bond
in the pyridyl radical.
In quinoline there is only one ortho site from which a H atom
can be ejected. The preferred â-scission is the rupture of the
C-N bond, similar to the process in pyridine. This leads to an
open-chain radical, for which the further decomposition pathway
of the lowest energy is a 1,4 H-atom migration in the first step
•
and the formation of C6H5 and HCtCCN in the second step.
Isoquinoline has two ortho positions with respect to the
benzene ring, R and â (see reactions 6 and 7). Both are
subjected to â-scission of the C-N bond. However, in
contradistinction to the â-scission in the R-ortho radical, which
opens without affecting the benzene ring, the â-scission of the
C-N bond in the â-ortho radical results in the formation of a
Note that reactions 4 and 5 are not necessarily elementary
reactions. They simply provide a means to form the meta and
the para radicals as if they were formed by H-atom ejection or
abstraction.

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 933
Figure 5. The reaction scheme showing the coupling between quinoline and isoquinoline in the “preferred pathways”. Numbers in parentheses
indicate reaction numbers as they appear in the reaction scheme.
Figure 6. The reaction scheme showing the coupling between para and meta quinolyl and their reaction products. Numbers in parentheses indicate
reaction numbers as they appear in the reaction scheme.
CdC bond conjugated to the benzene ring. It fixes the bonds
in the benzene ring and destroys its aromaticity.
There must be a fast coupling between quinoline and
isoquinoline reactions along their preferred pathways at early
stages, before ring fragmentation occurs. This coupling equili-
brates the open chain radicals in the two pathways so that they
lose their identity. The coupling suggested in Figure 5 is based
on the well-known15 phenomenon that open-chain radicals
having π-bond in the chain undergo very fast radical-induced
cyclization forming two possible cyclic structures:
This process requires ∼20 kcal/mol (AM1 estimation) more
than the â-scission of the C-N bond in the R-ortho isoquinolyl
radical. It appears that the preferred pathway for isoquinoline
decomposition involves the R-ortho radical rather than the b.
Its decomposition begins with a H-atom ejection from the
R-ortho position in the pyridine ring, followed by â-scission of
the C-N bond and decomposition of the open-chain radical to
produce C2H2 and C6H4CN•.
In the present case the open-chain ortho quinolyl and R-ortho
isoquinolyl radicals undergo cyclization to yield the same
intermediate, namely, 1-indene imine radical.
Out of the two possible cyclizations in each one of the of
the open-chain radicals, one is simply the back reaction of the
opening of the ortho quinolyl and R-ortho isoquinolyl radicals
and the other is the formation of indene imine radical. Once
the latter has been formed, it can undergo two â-scissions to
yield the two different open-chain radicals.
Since the nitrogen is outside the five-membered ring, the
â-scission location determines whether the product of the
decyclization is a member of the quinoline or isoquinoline
pathway.
C. Additional Radical Decomposition Pathways. Figure 6
shows the ring cleavage of meta and para quinolyl and the
consequent fragmentation of the open-chain radicals. The
overall decomposition pathways were constructed in such a way
that both the ring opening and the fragmentation would follow
the â-scission rule, as assumed for the ortho quinolyl radical.
The C6H5• and C6H4CN• radicals are stable enough9,14 to be
able to abstract or recombine with a hydrogen atom before
decomposing.
The direct outcome of this analysis is that the preferred
decomposition pathways of the two isomers yield different final
products, HCtCCN and C6H6 in quinoline and C2H2 and C6H5-
CN in isoquinoline.
There is clear experimental evidence, however, not only that
the relative product abundances in the two isomers are identical,
but also that the production rates of all the products are exactly
the same, regardless of whether the reactant is quinoline or
isoquinoline. This “discrepancy” could be accounted for if a
very fast quinoline T isoquinoline isomerization would take
place. However, such an isomerization was not evident. No
isomerization products were identified in postshock mixtures.

934 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
Figure 7. The reaction scheme showing the coupling between â-ortho
and meta isoquinolyl and their reaction products. Numbers in paren-
theses indicate reaction numbers as they appear in the reaction scheme.
Figure 9. Four reaction pathways in the decomposition of isoquinolyl
with the radical sites on the benzene ring. Numbers in parentheses
indicate reaction numbers as they appear in the reaction scheme.
which is highly endothermic (∼120 kcal/mol), and the second
leads to the destruction of the resonance of the benzene ring
owing to the formation of a CdC bond adjacent to the ring.
Thus, decomposition of the â-ortho isoquinolyl radical was not
included in the reaction scheme, its only reaction being its
equilibration with the meta radical.
Figures 8 and 9 show the decomposition channels of the four
different radical sites in quinoline and isoquinoline arising from
removal of a hydrogen atom from the benzene ring. As has
been mentioned before, since all C-H bonds in the benzene
ring are of the same energy (∼112 kcal/mol), we used a single
radical in each isomer without specifying the radical site.
Nonetheless, we examined the routes of all four possibilities in
each isomer. Formation of C4H2 can be accounted for only by
decomposition from radicals obtained by removal of a hydrogen
atom from the benzene ring, as shown later when we discuss
the computer modeling results.
2. Computer Modeling. A. Reaction Scheme. To account
for the distribution of reaction products in the decomposition
of both quinoline and isoquinoline, a common reaction scheme
containing 72 species and 148 elementary reactions was
composed (Table 3).
The Arrhenius parameters for the majority of the reactions
were estimated by comparison with similar reactions with known
rate parameters. Additional ones were taken from kinetic
schemes that describe the decomposition mechanisms of ben-
zene,9,16 benzonitrile,14 phenyl acetylene,17 and pyridine.10
Several parameters were taken from the NIST Chemical Kinetic
Data Base.18
Thermodynamic properties of most the species were taken
from literature sources.7,10,12,19,20 The heats of formation of
several species were estimated using the NIST Structures and
Properties program.5
B. Comparison of Model Calculations and Experimental
Results. Figures 10-13 show the experimental and calculated
mole percents of the four products formed in the “preferred”
decomposition pathways (Figure 5), acetylene, cyanoacetylene,
benzene, and benzonitrile. The filled squares (9) are the
experimental points obtained with isoquinoline as a reactant and
the open squares (0) with quinoline. The lines represent
calculated mole percents at 25 K intervals. The solid lines are
the model calculations for isoquinoline and the dashed lines
for quinoline. In all of the four figures, the upper part A shows
the calculations without the coupling of the quinoline and
isoquinoline pathways (Figure 5), namely, with reactions 14 and
15 removed from the scheme. The lower part of the figures B
shows the results of the calculations with the coupling.
Figure 8. Four reaction pathways in the decomposition of quinolyl
with the radical sites on the benzene ring. Numbers in parentheses
indicate reaction numbers as they appear in the reaction scheme.
As can be seen in Figure 6, the open-chain radicals obtained
by â-scissions in the meta and para quinolyl radicals can also
undergo radical-induced cyclization T decyclization reactions
similar to the mechanism shown in Figure 5 and thus again
couple the two pathways shown in Figure 6. There is, however,
a marked difference in the nature of the coupling between the
two pairs of pathways that expresses itself by the different
locations of the nitrogen atom in the intermediate species. In
the “preferred” pathway starting with the ortho radicals (Figure
5), the nitrogen atom in the intermediate specie is located outside
the ring. It can undergo two different â-scissions. The specific
bond cleavage and the consequent recyclization determines
whether the final product is quinolyl or isoquinolyl. The
coupling is thus between the quinoline and isoquinoline
pathways. In the pathways of the meta and para quinolyl
radicals, on the other hand (Figure 6), the nitrogen atom in the
intermediate is part of the ring. It can also undergo two different
â-scissions, but none of which changes the relative position of
the nitrogen atom with respect to the benzene ring. Therefore,
the specific bond cleavage in this intermediate and the conse-
quent cyclization determines whether the final product is meta
or para quinolyl and not whether it is quinolyl or isoquinolyl.
The crossing from quinoline to isoquinoline and vice versa can
take place in the ortho radicals only. Since it is believed that
there is a fast equilibration of meta and para quinolyl radicals,
the coupling of the two pathways shown in Figure 6 and the
intermediate is insignificant from kinetic viewpoint. It was
therefore not included in the reaction scheme.
Three isoquinolyl radicals are formed by removal of a
hydrogen atom from the pyridine ring, R-ortho, â-ortho, and
meta. Figure 5 shows the pathway for the R-ortho isoquinolyl
radical, Figure 7 shows the pathways for the â-ortho and the
meta radicals. The pathways of these two radicals are also
coupled by an intermediate. However, the two possible
â-scissions of the â-ortho isoquinolyl radical both lead to dead
ends. One leads to formation of benzyne and isocyanide radical,

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 935
TABLE 3: Reaction Scheme for the Decomposition of Quinoline and Isoquinolinea (Values Are Given at 1400 K)

936 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
TABLE 3: (Continued)

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 937
TABLE 3: (Continued)

938 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
TABLE 3: (Continued)

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 939
TABLE 3: (Continued)

940 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
TABLE 3: (Continued)

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 941
a ∆Hr° and ∆Sr° are expressed in units of kcal/mol and cal/(K mol), respectively. Rate constants are expressed as k ) A exp(-Ea/RT) in units of
cm3, s, mol, kcal. b The point in the middle of the ring means that the corresponding radical was introduced into the reaction scheme in “common”
form, namely, irrespective of its exact location on the ring. c Estimation based on H-atom ejection from benzene (ref 18). d Estimation based on
H-atom ejection from ortho position in pyridine (ref 10). e Estimation based on rupture of ortho pyridyl radical (refs 10 and 12). f Estimation based
on rupture of phenyl radical (refs 9,18, and 21) with respect to the corresponding reaction thermochemistry. g Estimation based on H-atom abstraction
from benzene (refs 9 and 16) and pyridine (ref 10) rings. h H-atom abstraction from aromatic molecules by phenyl were reported in the work of
i
Fahr and Stein;22 such abstractions by phenyl, pyridyl, and their derivatives are also introduced into the reaction scheme. Reaction is not elementary:
j Reaction is not elementary:
k Reaction is not elementary, see text (Discussion, part 1A.).
Figure 10. Comparison between experimental and calculated mole
Figure 11. Comparison between experimental and calculated mole
percent of benzonitrile. 0 and broken line, quinoline; 9 and solid line,
isoquinoline. The upper part (A) shows results of model calculations
without coupling between quinoline and isoquinoline (Figure 5). The
lower part (B) shows the results with coupling.
percent of acetylene. 0 and broken line, quinoline; 9 and solid line,
isoquinoline. The upper part (A) shows results of model calculations
without coupling between quinoline and isoquinoline (Figure 5). The
lower part (B) shows the results with coupling.
As can readily be seen, as expected, without the coupling of
two pathways shown in Figure 5, the calculated product
distributions are different for the two isomers. On the other
hand, when the two pathways are coupled the agreement with
experiment (in the sense that both quinoline and isoquinoline
show almost identical distribution of reaction products) is very
good. The effect of coupling is particularly pronounced in
benzonitrile, as the R-ortho radical pathway in isoquinoline is
the only route by which benzonitrile can be formed when
quinoline is the starting material. It is less pronounced in the
other three products, as they are also produced by other channels,
in which this coupling does not play any role.
Figures 14-16 show the experimental and calculated results
for diacetylene, hydrogen cyanide, and phenyl acetylene. These
products are formed via other reaction channels, and the
coupling of the ortho radical pathways (Figure 5) affects their
concentrations only very slightly. The symbols and the lines
in the figures are the same as those in the previous four figures.
The additional line at the bottom of Figure 14 shows the
calculated mole percent of diacetylene after eliminating all the
reactions of quinolyl and isoquinolyl with the radical sites on
the benzene ring. The diacetylene formed by reaction channels
•
involving acetylene, such as C2H2 + C2H2 f C4H3 + H
followed by C4H3 f C4H2 + H•, are not fast enough at the
•
temperatures of this study to account for the observed C4H2.
Figure 17 shows the overall decomposition of quinoline and
isoquinoline.
C. Open Questions. There are still some open questions
regarding the experimental results and the model calculations.
1. Why does the coupling between quinoline and isoquinoline
not promote a quinoline T isoquinoline isomerization? 2. Why
were indene imine and other intermediates not identified in the
postshock mixtures? 3. If the reactions of quinolyl and
isoquinolyl with radical sites on the benzene ring are responsible
for the production of diacetylene (Figures 8 and 9), why are
pyridine and pyridyl acetylene formed in these reactions absent
from the products?

942 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
Figure 14. Comparison between experimental and calculated mole
percent of diacetylene. 0 and broken line, quinoline; 9 and solid line,
isoquinoline. The lower broken line shows the model calculations where
all the reactions in the two schemes shown in Figures 8 and 9 are
removed from the overall reaction scheme. Production of diacetylene
by reactions of acetylene alone cannot account for its observed mole
percent.
that its mole percent is around 1 × 10-3, below the detection
limit. This observation is somewhat surprising. The coupling
mechanism shown in Figure 5 should have led to a quinoline
isoquinoline isomerization since the open-chain radical coming
from â-scission of the ortho quinolyl exchanges to the equivalent
radical in isoquinoline. Cyclization of the latter and recombina-
tion with a hydrogen atom (or abstraction of one) should lead
to isomerization.
This discrepancy can be clarified by examining the competi-
tion between the two parallel reactions of the open-chain
radicals, fragmentation on one hand, and cyclization followed
by H-atom recombination on the other hand.
Figure 12. Comparison between experimental and calculated mole
percent of cyano acetylene. 0 and broken line, quinoline; 9 and solid
line, isoquinoline. The upper part (A) shows results of model calcula-
tions without coupling between quinoline and isoquinoline (Figure 5).
The lower part (B) shows the results with coupling.
Since the cyclization-decyclization processes of the radicals
are very fast at the temperature range of the present investiga-
tion, the species involved in these processes are effectively
equilibrated. The rate of formation of isoquinoline in tests
where quinoline is the reactant is thus given by (see Figure 5)
d[isoquinoline]
dt
k–12
k12
•
N
k–6[H•] + (k52 + k63)[quinoline]
=
kisoquinoline
where
k–12
k12
•
N
=
N
•
The two terms in the curled parenthesis represent the rates
of recombination with and abstraction of hydrogen atoms. The
fragmentation rate of the open-chain radical to acetylene and
benzonitrile radical is
d[C2H2]
•
N
Figure 13. Comparison between experimental and calculated mole
percent of benzene. 0 and broken line, quinoline; 9 and solid line,
isoquinoline. The upper part (A) shows results of model calculations
without coupling between quinoline and isoquinoline (Figure 5). The
lower part (B) shows the results with coupling.
= k13
dt
The computer simulation shows that the ratio k13/kisoquinoline
is around 3 × 103 at 1300 K and 8 × 103 at 1550 K. This
result is due to the very low concentration of hydrogen atoms
and a very low rate of abstraction by large radical species.
Figure 18 shows the calculated mole percent of isoquinoline
in tests with quinoline as the reactant. The calculations show

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 943
Figure 18. Calculated mole percent of pyridine, isoquinoline and the
indene imine intermediate with quinoline as the decomposing com-
pound. The mole percents of all the three compounds are very small;
for two, they are below the detection limit. Pyridine was identified in
the postshock mixtures but in minute quantities that did not allow a
quantitative analysis.
Figure 15. Comparison between experimental and calculated mole
percent of hydrogen cyanide. 0 and broken line, quinoline; 9 and solid
line, isoquinoline. The modeling is done with coupling.
Figure 19. Calculated mole percents of ortho, meta, and para pyri-
dile acetylene. Pyridyl acetylene was identified in the postshock
mixtures but in minute quantities that did not allow a quantitative
analysis.
Figure 16. Comparison between experimental and calculated mole
percent of phenyl acetylene. 0 and broken line, quinoline; 9 and solid
line, isoquinoline. A slight difference in the mole percent of phenyl
acetylene in decomposition of quinoline and isoquinoline is seen in
both the experiment and the calculations. The modeling is done with
coupling.
kcal/mol) but also its equilibrium concentration is extremely
small. The expected mole percent of indene imine is very low
and practically unrecoverable, as can be seen in Figure 18.
The two pairs of products, diacetylene and pyridyl radical
and acetylene and pyridyl acetylene radical, are formed in
theprocess of opening the benzene ring in both quinoline and
isoquinoline. On the other hand, opening of the pyridine ring,
in the main channel, leads to the production of phenyl and
cyanoacetylene and to benzonitrile radical and acetylene.
Whereas phenyl and benzonitrile radicals show up as benzene
and benzonitrile in the line of the products, pyridine and pyridyl
acetylene were identified only in minute quantities. The reason
for this behavior is the very low stability of pyridyl and pyridyl
acetylene radicals as compared to phenyl and benzonitrile
radicals, which do not have a nitrogen atom in the ring. The
decomposition of pyridyl requires only 35 kcal/mol,12 whereas
the decomposition of phenyl requires some 65 kcal/mol.9
Pyridyl decomposes almost completely before it has the chance
to recombine or to abstract a hydrogen atom, whereas phenyl
and benzonitrile radicals are stable enough and have a long
enough lifetime to recombine or to abstract a hydrogen atom
from the reactant.
Figure 17. Comparison between experimental and calculated mole
percent of reactant loss. 0 and broken line, quinoline; 9 and solid line,
isoquinoline.
The considerations regarding our ability to identify the
intermediate indene imine in the postshock mixtures are similar
to those of indentifying isoquinoline. The computations show
that not only the abstraction of hydrogen atoms from quinoline
by indene imine radical is very slow (it is endothermic by 8-14
Figures 18-19 show calculated yields of pyridine, ortho,
meta, and para pyridyl acetylenes produced according to the
present scheme from quinoline as a reactant. As can be seen,
only pyridine and ortho pyridyl acetylene are expected to be
produced in detectable amounts. Compounds having m/z ) 79

TABLE 4: Sensitivity Factors Sij ) ∆logCi/∆logki at 1300/1550 K for Quinoline. k Is Changed by a Factor of 3

Thermal Decomposition of Quinoline
J. Phys. Chem. A, Vol. 102, No. 6, 1998 945
and 103 corresponding to pyridine and pyridyl acetylene were
indeed identified in the postshock mixtures, but only in trace
amounts that could not be analyzed quantitatively.
D. SensitiVity Analysis. Table 4 shows the sensitivity spec-
trum of the reaction scheme for the decomposition of quinoline
as the reactant, at 1300 and 1500 K. The sensitivity factor Si,j
is defined in the table as: ∆log Ci/∆log kj at 2 ms. It was
evaluated here by changing kj by a factor of 3. (Si,j ) 1 means
that a factor of 3 change in kj causes a factor of 3 change in
Ci.) Reactions for which Si,j is less that 0.1 for all the products
both at low and high temperatures are not included in the table.
The sensitivity spectrum for the decomposition of isoquinoline
is practically identical in its context to that of quinoline and it
is thus not shown.
Most of the features in the sensitivity spectrum are self-
evident. It is of interest, however, to examine some of the
details in the mechanism in light of the information given in
Table 4. Since the overall mechanism of the decomposition of
quinoline is essentially a free radical mechanism, ejection of a
hydrogen atom from the molecule, being the initiation step, is
a very important step to which the entire system is very sensitive.
However, since ejection of H atom from the ortho position in
the pyridine ring (reaction 3) is faster than ejection of H from
other sites in the pyridine and benzene rings, the contribution
of reaction 1, for example, to the concentration of H atoms in
the system is no more than 5-7% and does not affect the
product concentrations.
Although diacetylene (C4H2) is a direct product of the quinolyl
radical which is formed in reaction 1, its formation is more
sensitive to the value of k3 than to the value of k1. The reason
for this is that reaction 55, which also produces quinolyl with
the radical site on the benzene ring, depends on the total
concentration of H atoms which is determined mainly by the
rate of reaction 3. Indeed, as can be seen in Table 4, diacetylene
production is sensitive to reaction 55 with roughly the same
sensitivity factor as reaction 1.
Another interesting feature is the competition between the
decomposition channels of the two open-chain radicals at the
two ends of the indene imine radical, (i.e., reactions 11 and
13). Reaction 13, which is responsible for the formation of
acetylene and benzonitrile, strongly inhibits the products of the
second open-chain radical (reaction 11). Since cyclization T
decyclization of indene imine radical is fast, the three radicals
involved are practically in equilibrium with one another, and
increasing or decreasing the rate of these processes has no effect
on the product distribution.
V. Conclusion
The mechanism of the thermal decomposition of quinoline
and isoquinoline can be summarized as follows.
1. The decomposition is initiated by H-atom ejections from
sites in both the pyridine and the benzene rings, with ejection
from the ortho position relative to the nitrogen in the pyridine
ring preferred.
2. The distribution of reaction products in quinoline is
identical with that of isoquinoline even though the preferred
decomposition routes in the two isomers are different.
3. An intermediate indene imine radical couples, in fast
cyclization T decyclization processes, the decomposition routes
of the two isomers and is thus responsible for the identical
product distributions.
4. A combined reaction scheme for quinoline and isoquino-
line containing 72 species and 148 elementary reactions

946 J. Phys. Chem. A, Vol. 102, No. 6, 1998
Laskin and Lifshitz
(11) Kern, R. D.; Xie, K. Prog. Energy Combust. Sci. 1991, 17, 191.
(12) Jones, J.; Bacskay, G. B.; Mackie, J. M. Isr. J. Chem. 1996, 36,
239.
(13) Lifshitz, A.; Tamburu, C.; Frank, P.; Just, Th. J. Phys. Chem. 1993,
97, 4085.
(14) Lifshitz, A.; Cohen, Y.; Braun-Unkhoff, M.; Frank, P. Proceedings
of the 26th International Symposium on Combustion; The Combustion
Institute: Pittsburgh, PA, 1996; p 659.
successfully accounts for the product distribution as a function
of temperature.
Acknowledgment. This research was supported by a grant
from the Israel Coal Supply Co. The authors thank Professor J.
A. Berson of Yale University for very helpful discussions.
(15) Giese, B.; Kopping, B.; Go¨bel, T.; Dickhant, J.; Thoma, G.; Kulicke,
K. J.; Trach, F. Radical Cyclization Reactions. In Organic Reactions; John
Wiley and Sons: New York, 1996; Vol. 48, p 303.
(16) Kiefer, J. H.; Mizerka, L. J.; Patel, M. R.; Wei, H. C. J. Phys.
Chem. 1985, 89, 2013.
References and Notes
(1) Attar, A.; Hendrickson, G. G. In Coal Structure; Meyers, R. A.,
Ed.; Academic Press: New York, 1982; p 132.
(2) Given, P. H. In Coal Science; Gorbaty, M. L., Larsen, J. W.,
Wender, I., Eds.; Academic Press: New York, 1982; Vol. 3, p 65.
(3) Unsworth, J. F. In Coal Quality and Combustion Performance;
Unsworth, J. F.; Barrat, D. J., Roberts, P. T., Eds.; Elsevier Science
Publishers: Amsterdam, 1991; Chapter 4.2, p 206.
(4) Laskin, A.; Lifshitz, A. J. Phys. Chem. A. 1997, 101, 7787.
(5) Lifshitz, A.; Tamburu, C.; Suslensky, A. J. Phys. Chem. 1989, 93,
5802.
(6) Mackie, J. C.; Colket, M. B., III; Nelson, P. F.; Esler, M. Int. J.
Chem. Kinet. 1991, 23, 733.
(7) Stein, S. E.; Rukkers, J. M.; Brown, R. L. NIST-Standard Reference
Data Base 25; National Institute of Standards and Technology: Washington,
DC, 1993.
(8) Bruinsma, O. S. L.; Tromp, P. J. J.; de Sauvage Nolting, H. J. J.;
Moulijn, J. A. Fuel 1988, 67, 334.
(17) Hertzler, J.; Frank, P. Ber. Bunsen-Ges. Phys. Chem. 1992, 96, 1333.
(18) Westly, F.; Herron, J. T.; Cvetanovich, R. J.; Hampson, R. F.;
Mallard, W. G. NIST-Chemical Kinetics Standard Reference Data Base
17, Version 5.0, National Institute of Standards and Technology: Wash-
ington, DC, 1985.
(19) Burcat, A.; McBride, B. 1995 Ideal Gas Thermodynamics Data
for Combustion and Air-Pollution Use; TechnionsIsrael Institute of
Technology: Haifa, 1995. (TEA 732).
(20) Wang, H.; Frenklach, M. J. Phys. Chem. 1994, 98, 11465.
(21) Braun-Unkhoff, M.; Frank, P.; Just, Th. Proceedings of the
International 22th Symposium on Combustion; The Combustion Institute:
Pittsburgh, PA, 1988; p 1053.
(22) Fahr, H.; Stein, S. E. J. Phys. Chem. 1988, 92, 4951.
(23) Back, M. H. Can. J. Chem. 1971, 49, 2199.
(24) Weissman, M.; Benson, S. W. Int. J. Chem. Kinet. 1984, 16, 307.
(25) Benson, S. W. Int. J. Chem. Kinet. 1989, 21, 233.
(26) Laufer, A. H.; Bass, A. M. J. Phys. Chem. 1979, 83, 310.
(27) Number in parentheses indicates the number of the reaction in the
kinetic scheme.
(9) Laskin, A.; Lifshitz, A. Proceedings of the 26th International
Symposium on Combustion; The Combustion Institute: Pittsburgh, PA,
1996; p 669 and refs 1-11 cited therein.
(10) Mackie, J. C.; Colket, M. B., III; Nelson; P. F. J. Phys. Chem.
1990, 94, 4099 and refs 6-10 cited therein.