Pyrolyses of aromatic azines: Pyrazine, pyrimidine, and pyridine

Article abstract of DOI:10.1021/jp970211z

The thermal decompositions of pyrazine, pyrimidine, and pyridine in shock waves have been investigated using the complementary techniques of laser-schlieren (LS) densitometry and time-of-flight (TOF) mass spectrometry (1600-2300 K, 150-350 Torr). A free radical chain reaction with initiation by ring C-H fission in the pyrolyses of all three azines is proposed. The measured C-H fission rates are compared and analyzed by RRKM theory. Barriers of 103 ± 2 kcal/mol for pyrazine, 98 ± 2 for pyrimidine, and 105 ± 2 for pyridine have been determined, supporting values lower than the barrier for C-H fission in benzene, 112 kcal/mol. The lower barriers for the azines are explained by the additional contributions of resonance structures of azyl radicals due to neighboring N-C interactions, which serve to further stabilize the azyl radicals. Detailed chain mechanisms are constructed to model the LS profiles and the TOF concentration profiles of the major products, hydrogen cyanide, acetylene, cyanoacetylene, and diacetylene. Of particular interest are the TOF observations and the mechanistic explanation of temperature dependent maxima seen in the formation of cyanoacetylene in the presence or absence of excess H₂.

Full text of DOI:10.1021/jp970211z

J. Phys. Chem. A 1997, 101, 7061-7073
7061
Pyrolyses of Aromatic Azines: Pyrazine, Pyrimidine, and Pyridine
J. H. Kiefer* and Q. Zhang†
Department of Chemical Engineering, UniVersity of Illinois at Chicago, Chicago, Illinois 60607
R. D. Kern,* J. Yao, and B. Jursic
Department of Chemistry, UniVersity of New Orleans, New Orleans, Louisiana 70148
ReceiVed: January 14, 1997; In Final Form: April 22, 1997X
The thermal decompositions of pyrazine, pyrimidine, and pyridine in shock waves have been investigated
using the complementary techniques of laser-schlieren (LS) densitometry and time-of-flight (TOF) mass
spectrometry (1600-2300 K, 150-350 Torr). A free radical chain reaction with initiation by ring C-H
fission in the pyrolyses of all three azines is proposed. The measured C-H fission rates are compared and
analyzed by RRKM theory. Barriers of 103 ( 2 kcal/mol for pyrazine, 98 ( 2 for pyrimidine, and 105 (
2 for pyridine have been determined, supporting values lower than the barrier for C-H fission in benzene,
112 kcal/mol. The lower barriers for the azines are explained by the additional contributions of resonance
structures of azyl radicals due to neighboring N-C interactions, which serve to further stabilize the azyl
radicals. Detailed chain mechanisms are constructed to model the LS profiles and the TOF concentration
profiles of the major products, hydrogen cyanide, acetylene, cyanoacetylene, and diacetylene. Of particular
interest are the TOF observations and the mechanistic explanation of temperature dependent maxima seen in
the formation of cyanoacetylene in the presence or absence of excess H2.
Introduction
The major products of the pyrazine and pyrimidine pyrolyses,
HCN, C2H2, and CHCCN, were first observed by Doughty et
An increasing interest is evident in the investigation of the
pyrolyses of N-heterocycles containing one or more nitrogen
atoms in the aromatic ring such as pyrrole, pyridine, pyrazine,
and pyrimidine. This is in large part from their relevance in
the formation of various NOx species during the combustion of
heavy oils and of coal. Basic research in this field was first
conducted on pyridine with its relative structural simplicity; early
studies1-7 reported different decomposition rates depending on
reaction conditions. These investigations clearly demonstrated
the radical chain nature of the decomposition; some studies in
fact proposed that the pyrolysis is initiated primarily via C-H
fission,4-7 most likely from the carbon atom adjacent to the
nitrogen atom.7
al.10,11 in their single-pulse shock-tube studies covering the range
1200-1550 K and a total pressure of 13-14 atm. Both
decompositions were considered to be free radical mechanisms
initiated by low-barrier C-H fission (∼95 kcal/mol), which was
again attributed to the effect of the adjacent N-C interaction.
A theoretical study by Jones et al.9 supports this proposal;
however, this predicted somewhat higher barriers of ∼104 kcal/
mol for the initial dissociations of both molecules.
Doughty et al.10,11 reported a greater rate of pyrazine
dissociation relative to pyrimidine; this faster rate was ascribed
to a rapid abstraction of H atoms from the diazine parents by
the CN radical. To account for the difference in the rates,
Doughty et al.10,11 adopted an “A” factor for the initiation rate
in pyrazine which is higher by a factor of ∼3 than that for
pyrimidine in their model.
An experimental determination of the lower barrier depends
critically upon an accurate measurement of initiation rate. In
the single-pulse studies10,11 of the diazines, initiation rate
constants were derived by matching calculated and inferred
product yields for various reaction temperatures at the very late
times of 600-1000 µs. Here the product yield and distribution
will be largely governed by secondary chain processes and thus
have but a remote connection to the initiation rate. Moreover,
The o-pyridyl radical is believed to be the most stable of the
three possible pyridyls and the only major pyridyl radical whose
fission leads to the primary decomposition products, HCN, C2H2,
and CHCCN.7 The results of Kern et al. and Mackie et al.6,7
also revealed a very low barrier of 93-98 kcal/mol for the initial
C-H fission, much lower than the 112 kcal/mol for benzene.
The reduced strength of the C-H bond was explained by
Kikuchi et al.,8 Mackie et al.,7 and Jones et al.9 as arising from
the interaction of the nitrogen lone pair with the neighboring C
atom, implying an important chemical effect of the ring N atom
and its subsequent effect on the pyrolysis of azines.
* Corresponding authors.
† Present address: Department of Chemistry, University of New Orleans.
X Abstract published in AdVance ACS Abstracts, August 15, 1997.
S1089-5639(97)00211-9 CCC: $14.00 © 1997 American Chemical Society

7062 J. Phys. Chem. A, Vol. 101, No. 38, 1997
Kiefer et al.
TABLE 1: Sumary of Experimental Conditions
TOF. The chemicals used for the TOF experiments were
obtained from several sources: pyrazine, pyrimidine, pyridine
(Aldrich, >99%), H2 (Matheson, >99%), C2H2 (Matheson,
99.6%), Ar (Matheson, 99.998%), and Ne (Matheson, 99.999%).
Solid pyrazine was sublimed into a gas-handling system and
purified by bulb-to-bulb distillation. The middle fraction was
retained. A similar procedure was applied to purifying pyri-
midine and pyridine vapors. Before the shock-tube experiments,
the mixtures were analyzed by the TOF mass spectrometer at
room temperature. No impurities were recorded within the static
detection limit of the TOF, ca. 300 ppm.
mixture compositions
temperature (K)
pressure (Torr)
Laser-Schlierena
1689-2362
1680-2279
1679-2301
1666-2134
1745-2212
1722-2272
4% pyrazine
4% pyrimidine
4% pyridine
131-198
313-342
136-214
266-349
107-140
309-377
Time-of-Flightb
1460-2267
1491-2216
1731-2037
1641-2158
1603-2217
1706-2054
1695-2176
2% pyrazine
2% pyrimidine
2% pyridine
2% pyrazine-5% H2
2% pyrimidine-5% H2
2% pyridine-5% H2
2% pyridine-5% D2
189-376
196-365
250-322
239-367
232-382
255-334
254-372
The mass spectrometric calibration factor, also known as the
sensitivity factor, for one of the major products observed
throughout the TOF experiments, C2H2, was measured using a
2% C2H2-3% Ar-95% Ne mixture under nonreacting shock
conditions. The sensitivity factors for HCN, cyanoacetylene,
and diacetylene were determined by nitrogen and carbon atom
balances, respectively, during shock experiments.
a Mixtures balanced by Kr. b Diluent consists of 3% Ar balanced
by Ne.
The recorded mass spectra were transferred from a Tektronix
digitizer (RTD720A) to a PC and automatically searched for
the mass peaks of interest. The peak heights of preselected
masses are stored in a multicolumn data file. The converted
concentration profiles are stored in a separate file for the purpose
of comparison with modeling results.
kinetic study of such a chain reaction could also be distorted,
especially at low temperatures, due to initiation by impurities
which more easily release radicals.
The above problems can be most easily avoided by observing
the reactions at higher temperatures, where impurities are
unimportant and the chain contribution is comparatively minor.
To explore the characteristics of the unimolecular initiation and
the dynamic formation of its products, a recent study of pyrazine
pyrolysis by Kiefer et al.12a was performed at higher tempera-
tures of 1600-2300 K and lower total pressures of 150-350
Torr, utilizing the time-resolved techniques of laser-schlieren
(LS) densitometry and time-of-flight mass spectrometry (TOF).
This study confirmed the low barriers for both diazine fissions
relative to benzene dissociation, but proposed higher values than
those reported by Doughty et al.10,11 Nonetheless, a nearly
doubled barrier reduction was reported for pyrimidine dissocia-
tion. With the aid of time-resolved TOF product analysis, an
improved pyrazine chain mechanism was constructed featuring
a different role of the important chain carrier CN in the
formation of one of the key products, cyanoacetylene, CHCCN.
We present here comprehensive mechanisms to describe the
pyrolyses for all three azines including in detail the RRKM
parameters employed to determine their respective C-H fission
barriers. The LS experimental data for pyrazine and pyrimidine12a
have been corrected herein for a systematic error (an incorrect
calibration factor). The results for all three shock-tube tech-
niques, LS, TOF, and single-pulse, are compared to model
calculations using the mechanisms proposed in the present work.
Results
Some typical LS semilog gradient profiles for azine dissocia-
tion are shown in Figure 1, where the measured density gradient,
dF/dx, is plotted against time. All the azine dissociations can
be resolved to ∼2300 K; below ∼1600 K, the gradients are too
small for detection. The most important feature evident in these
experiments is that the density gradients all go through a clearly
discernible shallow maximum after an initial rise, descending
at late time. In our lower to moderate temperature range, this
process is more easily resolved than at higher temperatures,
where the positions of the maxima are shifted toward t ) 0
and the initial rise of the density gradients becomes difficult to
see. These maxima, often well resolved within the first 5 µs,
are an important indication of chain decomposition which have
also been observed in benzene dissociation.15 As the chain
reaction proceeds, the increasing rate of endothermic reaction
causes continuously increasing positive density gradients at the
early stage of the reactions, whereas at the late times the slow-
down from the temperature drop produces negative gradients
and thus generates the maxima. This evidence for a chain
reaction is consistent with the conclusions of Doughty et al.7,10,11
based on their product distribution.
Experimental Section
TOF concentration profiles of the major product species from
two pyrimidine experiments are displayed in Figure 2. The
major product species seen here are HCN, C2H2, and CHCCN,
the same as those identified in the single-pulse studies of
Doughty et al.10,11 At high temperatures, C4H2 is also detectable
in the present study. Mass peaks attributable to cyanogen and
acrylonitrile were observed within a concentration range, (0.1-
1) × 10-9 mol/cm3, which is within the TOF detection limit.
Their detection values are rough upper limit to their amounts
formed, although the scatter of the points is too high to construct
quantitative reaction profiles.
The experimental apparatus and the procedures for both the
LS and TOF techniques have been described previously.12a,13
The conditions for the azine experiments are summarized in
Table 1; some additional details are as follows.
LS. Pyridine, pyrazine, and pyrimidine were obtained from
Aldrich (purity > 99%) and were vacuum degassed; only the
middle fractions were retained. Krypton was Spectra-Gases
excimer grade. Mixtures were prepared manometrically in a
50 L glass bulb fitted with a Teflon-coated magnetic stirrer,
and the uncertainty of mixture composition was less than 1%
of the C5H5N and C4H4N2 fractions. The molar refractivities
of pyrazine and pyrimidine were taken as a constant, 22.73 and
23.44, respectively.14 Although these must vary during decom-
position, the mixture refractivity values are generally >87%
from that of Kr. A total of 56 runs of 2% and 4% pyrazine in
Kr and 48 runs of 4% pyrimidine in Kr were analyzed. The
pyridine data were published previously.6
A 2% C5D5N-3% Ar-Ne mixture was analyzed to distinguish
the minor products of pyridine pyrolysis. Several species have
the same molecular mass in nondeuterated pyridine, for example,
C4H3 and cyanoacetylene (m/e 51), or C2N2 and C4H4 (m/e 52).
The TOF results revealed that C4H3 and C4H4 concentrations
are negligibly low. Thus, the mass peak m/e 51 observed in

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7063
Figure 1. Laser-schlieren semilog density gradient profiles for 4% pyridine/diazine-Kr experiments. Experimental gradients are indicated by the
symbol (×); calculated gradients are shown by the solid lines.
the nondeuterated pyridine experiments is primarily from
cyanoacetylene. The mass peaks at m/e 52 and m/e 53 are
attributed mostly to the cracking pattern of the parent pyridine.
The concentrations of HCN, C2H2, C4H2, and CHCCN at low
temperatures all increase monotonically with reaction time and
eventually level out, whereas, at high temperatures, CHCCN
exhibits a sharp early maximum before a final drop to a constant
value.
The effect of hydrogen addition, reducing dramatically the
yield of CHCCN for all three azines, is shown in Figure 3.
However, this reduction effect is not significant for the other
species produced in the reaction.
Further ring opening and cleavage of the azyl radicals is
expected to generate various radicals with acyclic structures
containing six to eight carbon and nitrogen atoms with the
radical site on the N atom. A group additivity estimation of
the heats of formation of these species requires a value for the
imine N-H bond dissociation energy D0. As in the work on
pyrazine pyrolysis,12a we shall adopt a value of 89 kcal/mol,
that calculated using density functional theory.12b This value
is now found to be fully consistent with the latest ab initio
calculation by Melius.16
Another ambiguity arises from the uncertainty of the ther-
modynamic contribution of the (imine) Ni-Cd group, which is
also not available for the group additivity method.16 In this
study its value is assumed to be the same as that of Ni-CB,
recognizing that the same value has been assigned for both the
Cd-CdH and Cd-CBH groups.17
The heat of formation of CHCCN was reported as 96 kcal/
mol in the pyrazine work,12a based on a matching of the late-
time high-temperature CHCCH TOF profiles to modeling
calculations. Here the distribution of products is nearly
independent of the kinetics of the involved reactions and largely
determined by the thermochemistry of the species. The
sensitivity of ∆fH°298(CHCCN) to the late concentration profile
is illustrated in Figure 4. It can be seen that the late-time
concentration value would be a factor of ∼2 too high if a value
of 93 kcal/mol was used. Although the ∆fH°298(CHCCN)
Thermochemistry
The successful kinetic simulation of the complex free-radical
chain revealed by the LS results for all three azines relies on
the completeness and accuracy of the thermochemical informa-
tion of the involved species. As discussed in the Introduction,
the proposed stabilization effect due to the interaction between
a N atom lone electron pair and a neighboring C atom unpaired
electron in the azyl radicals implies that the usual bond-energy
methods cannot accurately estimate the energies of these
radicals. The C-H bond energies of these azines are regarded
here as primary unknowns whose values are to be determined
by kinetic modeling and the application of unimolecular rate
theory.

7064 J. Phys. Chem. A, Vol. 101, No. 38, 1997
Kiefer et al.
Figure 2. TOF product profiles from two experiments in 2% pyrimidine-3%Ar-Ne at 1635 K, 228 Torr (4) and 2042 K, 327 Torr (2). The solid
lines show simulations using the mechanism of Table 4.
derived here is indeed higher by ∼12 kcal/mol than an earlier
measurement of 84 reported by Harland,18 it should be noted
that the usual group additivity suggests a value of 91 kcal/mol,17
which is also consistent with the estimates of 90.8 and 91.3 by
Dorofeeva et al.19 and Dibeler et al.,20 respectively. The
reliability of the additivity value essentially depends on the
accuracy of the group value Ct-CN, 64 kcal/mol, suggested
by Benson17 from the measured heat of formation of 128.1 kcal/
mol for NC-CC-CN (2Ct-CN groups) given by Chu and Stull
et al.21 All of these indicate that ∆fH°298(CHCCN) could well
be higher than 90 kcal/mol and therefore partially supports our
conclusion.
for the initial dissociations of the azines are taken directly from
the zero-time gradients obtained by extrapolation of the LS
profiles using the mechanisms in Tables 3, 4, and 5 (discussed
below) to indicate the correct path through the first unresolved
microsecond,6,15,22 which lies within the induction time for this
chain. This procedure of model-assisted extrapolation is
particularly important in this work in order to define the zero-
time gradients, because in the first few microseconds, the
reaction is dominated by chain acceleration.
With a well-described initial rise and location of the
maximum, as in Figure 1, the model-assisted extrapolation
provides reliable values for (dF/dx)t)0. To convert these values
to rates of initial dissociation, the assumptions of vibrationally
relaxed and chemically frozen postincident shock conditions are
employed.
Some important thermochemical data are listed in Table 2;
other properties were estimated from standard statistical ther-
modynamic formulas.
First-order rate constants for pyrazine and pyrimidine have
been reported.12a A set of new values derived as above, and
corrected for a systematic error, are displayed in Figure 5. The
systematic error caused slightly lower density gradients by a
factor of ∼1.5. The present Arrhenius expressions and their
Kinetics
(1) Initiation Rates. The observed overall shape of the LS
azine profiles is indicative of a chain reaction which we assume
to be initiated by C-H fission of the parent. The rate constants

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7065
Figure 3. Effect of H2 addition on the CHCCN profiles for pyrazine, pyrimidine, and pyridine. Solid lines show simulations using the mechanisms
of Tables 3, 4, and 5, respectively.
However, the correction applied herein results in a much better
fit of the LS pyrazine and pyrimidine profiles without a change
in the mechanism. In fact, the problems with the late-time
calculations, noted in ref 12, are no longer present.
313-342 Torr
k (s-1) ) 1056.93T-12.17e-116.84(kcal/mol)/RT (rms 2.57%)
The expressions for pyrimidines are
136-214 Torr
k (s-1) ) 1040.46T-8.10e-90.07(kcal/mol)/RT (rms 4.91%)
266-349 Torr
k (s-1) ) 108.84T0.5e-59.32(kcal/mol)/RT (rms 5.70)
Figure 5 clearly exhibits not only a pressure dependence, i.e.,
falloff behavior for both diazines, but also an evident difference
in their initiation rates; pyrimidine dissociation appears faster
Figure 4. Effect of various values of ∆fH°298 for cyanoacetylene on
than pyrazine under similar conditions, and the temperature
dependence of the pyrazine rate is stronger than that of
pyrimidine. This implies that the initial pyrazine dissociation
must have a larger activation energy.
the modeling calculations for a 2% pyrazine experiment (2) at 1998
K and 308 Torr.
respective rms values for pyrazine dissociation are as follows:
A model-assisted extrapolation of the pyridine LS density
gradients to t ) 0 was also performed on the LS data for
pyridine pyrolysis published by Kern et al.,6 using the improved
131-198 Torr
k (s-1) ) 1065.97T-14.72e-124.08(kcal/mol)/RT (rms 6.06%)

7066 J. Phys. Chem. A, Vol. 101, No. 38, 1997
Kiefer et al.
TABLE 2: Important Thermochemical Data in Azine
Pyrolyses
pressures are:
107-140 Torr
∆fH°298
S298 (cal/
(kcal/mol) (mol K))
species
ref
35, 36
35, 37
6
k (s-1) ) 10149.62T-37.19e-211.83(kcal/mol)/RT (rms 12%)
C4H4N2 (pyrazine)
C4H4N2 (pyrimidine)
C5H5N (pyridine)
C4H3N2 (pyrazyl)
2-C4H3N2 (2-pyrimidyl)
4-C4H3N2 (4-pyrimidyl)
5-C4H3N2 (5-pyrimidyl)
o-C5H4N (o-pyridyl)
m,p-C5H4N (m,p-pyridyl)
C2N2
47.0
46.0
33.5
98.0
92.0
97.0
105.9
86.3
93.4
76.0
32.3
54.5
111.0
114.0
96.0
43.7
90.6
102.6
115.3
65.8
67.2
67.5
69.7
68.5
69.8
69.8
69.4
69.2
57.7
48.2
48.0
59.7
65.5
60.1
65.3
65.9
66.2
65.8
309-377 Torr
k (s-1) ) 10101.06T-23.98e-165.23(kcal/mol)/RT (rms 9%)
pwa and 36
pw and 37
pw and 37
pw and 37
pw
These initial rates are found to be generally lower by ∼30%
than the previous LS results.6 However, the LS pyridine data
are still modeled satisfactorily by the solid lines as shown in
Figure 1.
The results of RRKM analyses of the measured azine
initiation rates are also depicted in Figures 5 and 6. The
parameters of these calculations, chosen to fit the measured rate
constants, are detailed in Table 6. All the models have the usual
restricted-rotor Gorin transition state appropriate for bond
pw
16, 46
46
46
6
pw and 34
pw
38
pw and 39
pw and 38
pw and 17, 39
HCN
C2H2
C4H2
C4H3
CHtC-CN
CH2dCH-CN
NC-CHdN
CHdCH-CN
NC-NdCH
a pw ) present work.
fission. These calculations assume the same ∆E
) 400
down
cm-1 (average -∆E ≈ 53 cm-1, for 1500-2300 K) and η
all
) 0.008 for all three azines, values typical for large hydrocar-
bons.23 This indicates the similarity of the reaction and the
energy transfer behavior of these large molecules; only the
barriers were varied. This value of ∆E down is also supported
mechanism shown in Table 5. The newly derived initial rate
constants for pyridine decomposition are found to be ∼30%
lower than the previous ones and are not affected by the
aforementioned systematic error which applies to only the LS
pyrazine and pyrimidine data. The modified initiation rates
are displayed in Figure 6. The least-squares fits for both
by Miller and Barker24 in their recent calculation of -∆E
all
(∼55 cm-1) for pyrazine. Fits to the measured rates in both
high- and low-pressure ranges are excellent, with a barrier of
103 ( 2 kcal/mol for C-H fission in pyrazine, 98 ( 2 kcal/
TABLE 3: Reaction Mechanism for Pyrazine Pyrolysis
no.
reaction
A (cgs)
Initiation
n
Ea (kcal/mol)
---see text---
ref
(1)
C4H4N2 + M ) C4H3N2 + H + M
Propagation
1.00E12a
(2)
(3)
C4H3N2 ) C2H2 + NCCHN
0.00
-7.73
-6.45
0.00
30.0
50.3
44.1
0.0
cf. text
RRKM
RRKM
cf. text
RRKM
RRKM
est.
NCCHN + M ) CN + HCN + M
NCCHN + M ) H + C2N2 + M
CN + C4H4N2 ) 2HCN + CHCHCN
CHCHCN + M ) C2H2 + CN + M
CHCHCN + M ) CHCCN + H + M
H + C4H4N2 ) C4H3N2 + H2
CN + C4H4N2 ) HCN + C4H3N2
CHCHCN + C4H4N2 ) CH2CHCN + C4H3N2
CN + HCN ) H + C2N2
3.55E42
5.62E37
2.00E14
2.57E59
2.57E46
3.16E12
7.94E5
1.00E4
7.94E7
5.49E2
7.52E22
7.52E22
9.68E23
1.2E14
6.61E5
7.94E5
4.07E5
3.16E13
(4)
(5)
(6)
-11.91
-8.55
0.00
84.2
64.0
0.0
(7)
(8)
(9)
2.30
0.0
41
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
2.40
1.0
est.
42, est.
43
1.60
0.0
CN + H2 ) HCN + H
3.20
-0.2
CN + C2H2 ) CHCCN + H
-3.00
-3.00
-3.00
0.00
0.0
41, est.
41, est.
41, est.
44
45, est.
41, est.
44, est.
est.
C2H + HCN ) H + CHCCN
C2H + HCN ) CN + C2H2
0.0
0.0
C2H + C2H2 ) C4H2 + H
0.0
CH2CHCN + H ) CHCHCN + H2
CH2CHCN + CN ) CHCHCN + HCN
H2 + C2H ) C2H2 + H
2.50
6.0
2.30
0.0
2.40
0.2
C2H + C2N2 ) CHCCN + CN
0.00
0.0
Termination
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
H + NCCHN ) H2 + C2N2
3.97E13
3.13E13
1.00E14
3.97E13
3.15E13
1.00E12
1.00E14
3.15E13
3.15E13
1.58E13
1.99E17
7.24E16
4.57E15
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.0
0.0
est.
H + NCCHN ) 2HCN
est.
CN + NCCHN ) HCN + C2N2
H + CHCHCN ) H2 + CHCCN
H + CHCHCN ) HCN + C2H2
H + CHCHCN ) CH2CHCN
CN + CHCHCN ) HCN + CHCCN
CN + CHCHCN ) C2N2 + C2H2
2CHCHCN ) CHCCN + CH2CHCN
NCCHN + CHCHCN ) C2N2 + CH2CHCN
C2H + CN + M ) CHCCN + M
C2N2 + M ) 2CN + M
0.0
est.
0.0
est.
0.0
est.
0.0
est.
0.0
est.
0.0
est.
0.0
est.
0.0
est.
0.0
est.
100.3
105.2
extrap. 45
extrap. 45
HCN + M ) H + CN + M
Molecular Channels
8.89E11
(34)
(35)
CH2CHCN ) C2H2 + HCN
CH2CHCN ) CHCCN + H2
0.00
0.00
68.0
77.0
extrap. 45
extrap. 45
1.26E13
a Read as 1.00 × 1012.

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7067
TABLE 4: Reaction Mechanism for Pyrimidine Pyrolysis
no.
reaction
A (cgs)
Initiation
n
Ea (kcal/mol)
---see text---
ref
(36)
C4H4N2 + M ) 2-C4H3N2 + H + M
Propagation
2.82E12a
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
(48)
(49)
(50)
(51)
2-C4H3N2 ) C2H2 + NCCHN
0.00
0.00
0.00
-3.66
0.00
0.00
0.00
0.00
0.00
2.30
2.30
2.30
2.40
2.40
2.40
40.0
30.0
45.0
25.6
0.0
0.0
0.0
0.0
8.0
0.0
0.0
cf. text
cf. text
cf. text
RRKM
cf. text
cf. text
est.
est.
est.
41
41
4-C4H3N2 ) HCN + CHCHCN
2.51E11
1.00E13
3.63E27
1.00E14
2.00E14
1.58E12
3.16E12
1.58E12
2.00E5
3.98E5
2.00E5
2.51E3
5.01E3
2.51E3
5-C4H3N2 ) H + HCN + CHCCN
NCNCH + M ) CN + HCN + M
CN + C4H4N2 ) HCN ) C2H2 + NCCHN
CN + C4H4N2 ) 2HCN + CHCHCN
H + C4H4N2 + H2 + 2-C4H3N2
H + C4H4N2 ) H2 + 4-C4H3N2
H + C4H4N2 ) H2 + 5-C4H3N2
CN + C4H4N2 ) HCN + 2-C4H3N2
CN + C4H4N2 ) HCN + 4-C5H3N2
CN + C4H4N2 ) HCN + 5-C4H3N2
CHCHCN + C4H4N2 ) CH2CHCN + 2-C4H3N2
CHCHCN + C4H4N2 ) CH2CHCN ) 4-C4H3N2
CHCHCN + C4H4N2 ) CH2CHCN + 5-C4H3N2
1.0
0.0
3.0
8.0
41
est.
est.
est.
Termination
(52)
(53)
(54)
(55)
H + NCNCH ) H2 + C2N2
3.97E13
3.15E13
1.00E14
2.00E13
0.00
0.00
0.00
0.00
0.0
0.0
0.0
0.0
est.
est.
est.
est.
H + NCNCH ) 2HCN
CN + NCNCH ) HCN + C2N2
NCNCH + CHCHCN ) C2N2 + CH2CHCN
a Read as 2.82 × 1012.
TABLE 5: Reaction Mechanism for Pyridine Pyrolysis
no.
reaction
A (cgs)
n
Ea (kcal/mol)
ref
Initiation
(56)
C5H5N + M ) o-C5H4N + H + M
---see text---
Propagation
1.00E13a
(57)
(58)
(59)
(60)
(61)
(62)
(63)
(64)
(65)
(66)
(67)
(68)
(69)
(70)
o-C5H4N ) C2H2 + CHCHCN
0.00
0.00
0.00
-5.40
0.00
0.00
0.00
0.00
2.30
2.30
2.30
2.40
2.40
2.40
50.0
40.0
40.0
37.0
0.0
0.7
8.0
8.0
0.0
0.0
0.0
cf. text
cf. text
cf. text
RRKM
cf. text
est.
est.
est.
est.
est.
est.
est.
est.
est.
m-C5H4N ) C4H3 + HCN
3.16E12
3.16E12
7.67E33
1.00E14
6.31E12
6.31E12
3.15E12
3.98E5
3.98E5
2.00E5
5.01E3
5.01E3
2.51E3
p-C5H4N ) C4H3 + HCN
C4H3 + M ) C4H2 + H + M
CN + C5H5N ) HCN + C2H2 + CHCHCN
H + C5H5N ) o-C5H4N + H2
H + C5H5N ) m-C5H4N + H2
H + C5H5N ) p-C5H4N + H2
CN + C5H5N ) HCN + o-C5H4N
CN + C5H5N ) HCN + m-C5H4N
CN + C5H5N ) HCN + p-C5H4N
CHCHCN + C5H5N ) CH2CHCN + o-C5H4N
CHCHCN + C5H5N ) CH2CHCN + m-C5H4N
CHCHCN + C5H5N ) CH2CHCN + p-C5H4N
0.0
5.0
4.5
Termination
(71)
(72)
(73)
(74)
(75)
H + C4H3 ) H2 + C4H2
3.97E13
1.99E13
1.00E14
3.15E13
1.99E13
0.00
0.00
0.00
0.00
0.00
0.0
0.0
0.0
0.0
0.0
est.
est.
est.
est.
est.
H + C4H3 ) 2C2H2
CN + C4H3 ) HCN + C4H2
CN + C4H3 ) CHCCN + C2H2
CHCHCN + C4H3 ) C4H2 + CH2CHCN
a Read as 1.00 × 1013.
mol for pyrimidine, and 105 ( 2 kcal/mol for pyridine. The
larger rate and lower Ea of pyrimidine are clearly evident in
Figure 5.
origins and rate constants are discussed in the following sections.
Other reaction rates are based primarily on literature values or
estimates.
(2) Reaction Mechanism. The complete reaction mecha-
nisms used to simulate the LS and TOF data are listed in Table
3 for pyrazine, Table 4 for pyrimidine, and Table 5 for pyridine.
Some of the reactions in Table 3, (3)-(7) and (11)-(35), are
common to pyrimidine; (6), (7), (11)-(20), (23)-(29), and
(31)-(35) are common to pyridine.
In the mechanisms of Tables 3-5, H and CN radicals appear
as the two primary chain carriers. The complexity of azine
pyrolyses undoubtedly incurs numerous additional species and
reactions. However, to adequately explain the kinetic behavior
of the major products HCN, C2H2, CHCCN, and C4H2, it is
only necessary to consider a few important reactions whose
Azyl Radical Dissociations. In contrast to the simple,
dominant pyrazyl dissociation path indicated by the bold arrow
steps in Scheme 1,12a there are three main pathways for
pyrimidyl dissociation, all of which are illustrated in Scheme 2
(a-c). Standard heats of reactions (kcal/mol) for the major and
minor pathways are in parentheses. In Scheme 2, only (a), (c),
and the bold-arrow channels in (b) are included. It should be
noted that the heats of reaction of these first azyl ring-opening
steps in Schemes 1 and 2 are all < 36 kcal/mol, whereas that
of phenyl dissociation, shown in Scheme 3, is almost 62 kcal/
mol. This strongly implies much lower barriers for the azyl
dissociations.

7068 J. Phys. Chem. A, Vol. 101, No. 38, 1997
Kiefer et al.
Figure 6. Arrhenius plots of first-order rate constants for C-H fission
of pyridine: (0) 309-377 Torr and (b) 107-140 Torr. The solid and
dashed lines illustrate the results of RRKM calculations with ∆E
down
) 400 cm-1, the long dashed line for 350 Torr, and the short dashed
for 150 Torr. The extrapolated values for k∞ are also shown above the
rate constant data.
Figure 5. Arrhenius plot of first-order rate constants for the C-H
fission of pyrazine, reaction 1 in Table 3, and pyrimidine, reaction 36
in Table 4: (b) pyrazine, 313-342 Torr; (9) pyrazine, 133-198 Torr;
(O) pyrimidine, 266-349 Torr; (0) pyrimidine, 136-214 Torr. The
solid and dashed lines illustrate the results of RRKM calculations with
∆E down ) 400 cm-1 for both pyrimidine and pyrazine. The solid lines
represent extrapolated values for k∞, rate constants at 305 Torr (upper
solid) and 165 Torr (lower solid) for pyrimidine; the dashed lines
represent extrapolated values for k∞, rate constants at 329 Torr (upper
dashed) and 156 Torr (lower dashed) for pyrazine.
migration from middle carbon to form NCCCH2. However,
direct H-C fission from the middle carbon is much more
entropy-favorable and therefore faster than H-migration, which
in fact results in the formation of one of the important products,
NCCCH, cyanoacetylene.
Pyridyl dissociations are similarly considered in Scheme 4.
The modeling results demonstrate that it is in fact the accelerat-
ing chain and endothermicity released by the fast dissociations
in these reactions and those of the subsequent linear four-
membered radicals (see below) that are largely responsible for
the early rise of the LS density gradients. This rise is less
extreme in azine experiments than that seen in benzene
pyrolysis,15 not because the reactions are slower, but because
these heats of reaction are much lower than that for the phenyl
dissociation in Scheme 3.
The overall rate chosen previously12a for pyrazyl dissociation,
1012e-30(kcal/mol)/RT (s-1), is in the falloff regime, despite the very
large rate. This takes into account the fact that such a large
radical can be directly formed pre-excited through thermoneutral
or exothermic abstraction, reactions 8-10 of Table 3, 43-51
of Table 4, or 62-70 of Table 5, without the usual need for
collisional excitation.25 To reflect the same ease in the
pyrimidyl dissociations, very fast rates are also adopted for the
reactions in Scheme 2, namely, reactions 37-39 in Table 4.
The ring openings of the various azyl radicals formed in this
pyrolysis either by dissociation or abstraction are extremely
facile, with both low barriers and high entropies of reaction.
An estimation of the HPL rate constant for such openings will
exceed 1010 s-1 at present temperatures. Collisional activation
cannot support such a large rate, and the reaction would
normally fall off sharply to something near the LPL, with a
much smaller rate. At the lower pressures used herein, this rate
Unimolecular Dissociations of Key Small Radicals. One of
the important features in azine pyrolyses is the generation of
chain radicals containing a total of four C and N atoms, i.e.,
NCCHN, CHCHCN, and NCNCH, as shown in Schemes 1, 2,
and 4. The dissociations of these radicals, reactions 3-6 in
Table 3 and 40 in Table 4, largely control the overall rate of
chain propagation. Unfortunately, no theoretical study of the
rate of these reactions has so far been reported. The rates of
these reactions are accordingly estimated using a vibration model
RRKM calculation chosen because these unimolecular dissocia-
tions are not simple bond fission processes; they must have
barriers to their reverse reactions and local transition states.
These barriers were estimated by Jursic as described previously.12a
Important parameters are given in Tables 7-9. The calculations,
performed with a consistent value of ∆E down ) 250 cm-1 (as
employed for many other molecules26), demonstrate that the rate
constants of these reactions are all near the low-pressure limit,
but are still very fast. Together with the azyl dissociations, these
decompositions assist the rise of the LS density gradients after
the first microsecond by contributing additional endothermicity
and promoting chain propagation.
could drop to around 107 s-1
.
However, radicals formed by near-thermoneutral abstraction
retain the thermal energy of the large parent molecule; most
are thus formed with more than enough internal energy to open.
The rates estimated herein take this situation into account; the
assumed rates are larger than is reasonable for activation-
restricted dissociation, but not so large as the HPL rates. This
seems a reasonable compromise. This idea is relatively new;
its origin and a more detailed description is available.25
The species NCCHCH in eq b of Scheme 2 is one of the
important radicals generated by the bond fission of its precursor
radical. It is possible that this radical could undergo an H atom

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7069
TABLE 6: Gorin Model RRKM Parameters for Azine Dissociations: Pyridine/Pyrazine/Pyrimidine + M f o-Pyridyl/Pyrazyl/
2-Pyrimidyl + H + M
pyridine
pyrazine
pyrimidine
a
frequencies (cm-1
)
(374, 405, 563, 605, 700, 749, 886, (350, 418, 602, 704, 756, 785,
(344, 399, 623, 678, 721, 811,
960, 980, 992, 1065, 1071,
1139, 1159, 1225, 1370, 1398,
1466, 1568, 3038, 3052, 3074),
1004, 1564, 308637
981, 981, 992, 1030, 1068,1085,
1146, 1218, 1377, 1440, 1483,
1571, 1583, 3036, 3055, 3055,
3055), 3035, 1218, 8856
960, 983, 1016, 1018, 1063,
1130, 1149, 1346, 1411, 1483,
1525, 1580, 3012, 3069, 3055),
927, 1233, 304036
transition state moments of inertia 2.84, 1.39, 1.456
(10-38 gm cm2)
2.73, 1.31, 1.4236
2.72, 1.38, 1.3440
reaction path degeneracy
2
4
1
η
0.008
400
104.9 ( 2
0.008
400
103.0 ( 2
0.008
400
97.8 ( 2
∆E down (cm-1
)
barrier, Eo (kcal/mol)
fit of RRKM results (rms < 0.05%)
pyridine
pyrazine:
pyrimidine:
log k∞ (s-1) ) 16.41 - 103.6 (kcal/mol)/2.3RT
log k∞ (s-1) ) 16.72 - 103.6 (kcal/mol)/2.3RT
log k∞ (s-1) ) 16.24 - 98.8 (kcal/mol)/2.3RT
a Frequencies in parentheses are also used for transition states.
SCHEME 1
radical site around the ring via the following three resonance
structures.
SCHEME 2
The resonance energy in cyclohexadienyl is estimated to be ∼23
kcal/mol according to group additivity.17
It is reasonable to propose that similar additions must also
occur in the azines. Specifically, various radicals “X” may add
to either a N or C atom in the ring. However, because the N-X
bond is much weaker than a C-X attachment, the X addition
to C atom should be more important. Scheme 6 illustrates the
additions of CN, H, NCCHN, and CHCHCN to the C atom of
parent pyrazine.
These adduct radicals are further stabilized through two
resonance structures involving the N atom lone pair, in addition
to the structures analogous to those of cyclohexadienyl illustrated
above.
SCHEME 3
SCHEME 4
These structures will also provide resonance energy in addition
to the 23 kcal/mol derived above for cyclohexadienyl. With
an additional 5 kcal/mol of resonance energy assumed for this
in pyrazine, the heats of formation of the resulting adducts can
be then roughly estimated by the group additivity method as
98.1 kcal/mol for CN, 67.1 for H, 129.6 for NCCHN, and 116.9
for the CHCHCN adduct. The resulting heats of reaction for
each addition are also shown in Scheme 6. Due to the low
concentrations of NCCHN and CHCHCN and the large negative
entropy accompanying the addition of the large and unstable
NCCHN and CHCHCN radicals, addition of these radicals to
the pyrazine parent will hardly play an important role. There-
fore, these addition reactions are neglected in the mechanisms;
only the CN and H additions are recognized here. Furthermore,
it is seen that CN addition is very exothermic and will ultimately
lead to the formation of the important CHCHCN radical (see
below). The overall result of CN addition to pyrazine, indicated
by bold arrows in Scheme 7, has been included as a single step
(5) in Table 3.
Additions of CN to Azine. In the earlier paper on pyrazine,12a
a new path to CHCCN was proposed, initiated by CN addition
to the parent molecule. Here this scheme is considered in more
detail and extended to the other azines.
It has been well established that H can add to the benzene
ring to form the resonance-stabilized cyclohexadienyl radical
(Scheme 5).27-30 The∆fH°298 of cyclohexadienyl has been
determined as ∼47 kcal/mol by Nicovich et al.27 and Tsang.31
This adduct radical is in fact stabilized by delocalization of the
Based on the consideration that further decomposition of the
highly activated CN addition product, cyclohexadienylnitrile
radical, could have a rate close to the high-pressure limit, the
overall rate was assumed to be limited by CN addition; the

7070 J. Phys. Chem. A, Vol. 101, No. 38, 1997
Kiefer et al.
TABLE 7: Vibration Model RRKM Parameters for CHCHCN Dissociations: CHCHCN + M f CN + C2H2 + M (6);
CHCHCN + M f H + CHCCN + M (7)
molecular frequencies (cm-1
molecular moments of inertia (10-38 gm cm2)
)
242, 362, 570, 683, 869, 954, 972, 1282, 1615, 2239, 3042, 312538
1.52, 1.67, 0.1539
(6)
(7)
transition state frequencies (cm-1
)
100, 150, 200, 500, 500, 683, 800,
242, 362, 500, 500, 570, 683, 800,
1800, 2239, 3100, 3125
3.10, 3.40, 0.15
2.04
65.0
1
1000, 1800, 2239, 3125
1.52, 1.67, 0.15
1.00
50.0
1
transition state moments of inertia (10-38 gm cm2)
I+/I
barrier, Eo (kcal/mol)
reaction path degeneracy
∆E down (cm-1
)
250
250
fits of RRKM results of 200 Torr (rms < 0.05%):
CHCHCN + M f CN + C2H2 + M
log k (cm3/(mol s)) ) 59.41-11.91 log T - 84.19 (kcal/mol)/2.3RT
CHCHCN + M f H + CHCCN + M
log k (cm3/(mol s)) ) 46.41-8.55 log T - 64.04 (kcal/mol)/2.3RT
TABLE 8: Vibration Model RRKM Parameters for NCCHN Dissociations: NCCHN + M f CN + HCN + M (3); NCCHN +
M f H + C2N2 + M (4)
molecular frequencies (cm-1
molecular moments of inertia (10-38 gm cm2)
)
3000, 2243, 1609, 1386, 945, 908, 614, 333.5, 23239
1.52, 1.67, 0.1539
(3)
(4)
transition state frequencies (cm-1
)
3150, 2243, 1800, 100, 945, 700, 50,50
3.10, 3.40, 0.15
2.04
46.0
1
2243, 1800, 800, 945, 850, 614, 350, 250
1.52, 1.67, 0.15
1.00
41.0
1
transition state moments of inertia (10-38 gm cm2)
I+/I
barrier, Eo (kcal/mol)
reaction path degeneracy
∆E down (cm-1
)
250
250
fits of RRKM results of 200 Torr (rms < 0.05%):
NCCHN + M f HCN + CN + M
log k (cm3/(mol s)) ) 42.55-7.73 log T - 50.26 (kcal/mol/2.3RT
NCCHN + M f H + C2N2 + M
log k (cm3/(mol s)) ) 37.75-6.45 log T - 44.12 (kcal/mol)/2.3RT
TABLE 9: Vibration Model RRKM Parameters for
NCNCH Dissociation: NCNCH + M f CN + HCN + M
SCHEME 7
molecular frequencies (cm-1
)
241, 358, 636, 754, 939, 1459.5,
1616.5, 2211, 314439
transition state frequencies (cm-1
)
200, 241, 300, 754, 1459.5, 1800,
2211, 3144
molecular moments of inertia
(10-38 gm cm2)
1.52, 1.67, 0.1539
transition state moments of inertia 3.10, 3.40, 0.15
(10-38 gm cm2)
reaction path degeneracy
I+/I
1
2.04
32.0
250
barrier, Eo (kcal/mol)
SCHEME 8
∆E down (cm-1
)
fit of RRKM results of 200 Torr (rms < 0.05%):
log k (cm3/(mol s)) ) 27.56-3.660 log T - 25.602 (kcal/
mol)2.3RT
SCHEME 5
SCHEME 6
magnitude of the rate was selected to be typical of the rate of
CN addition to olefins, ∼1014 cm3/mol-s.32
opening channels again favor the bold-arrow routes because of
their low energy. The final overall consequence of these
channels is either the formation of C2H2, HCN, and NCCHN,
or 2 HCN and CHCHCN, indicated as reactions 41 and 42,
respectively in Table 4. It is again assumed that CN addition
Addition of CN to pyrimidine is similarly considered in
Scheme 8. Here a group additivity analysis reveals that the
three isomers have similar heats of formation, 103.1 kcal/mol
for 2-, 95.8 for 4-, and 93.1 for the 5-adduct. However, ring-

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7071
SCHEME 9
is the rate-controlling step with a rate similar to that in pyrazine.
CN addition routes to pyridine are presented in Scheme 9.
Clearly in Scheme 9 only the bold-arrow channels are important,
and these lead to the products HCN, C2H2, and CHCHCN. The
dominant pathways are represented by a single reaction, (61),
in Table 5.
The important result in Schemes 7-9 is the formation of the
CHCHCN radical, a principal precursor to the observed product
CHCCN.
Discussion
(1) Initial C-H Fission. Mackie et al.7 and Doughty et
al.10,11 consistently inferred a lower barrier for the initial C-H
bond fissions from their pyridine and diazine experiments
compared with the well-known C-H bond energy of 112 kcal/
mol in benzene. This has been explained by Jones et al.9 as
arising from the stabilization effect of the interaction of the
adjacent C atom radical site with the neighboring N lone electron
pair. Their computational study of the geometries of azyl
radicals at the SCF/3-21G level of theory shows a noticeable
shortening of the N-C bond in these radicals, which is
indicative of such an interaction.
It is instructive to examine the azyl radicals as hybrids of
the canonical structures illustrated in Figure 7. In addition to
the benzene analogue structures (I, II, IV, V, VIII, IX, XI, XII,
XIII, XIV, XVI, XVII, XVIII, and XIX), the interaction of the
N lone pair with the neighboring unpaired C electron provides
additional unique resonance structures, i.e., III of pyrazyl, VI
and VII of 2-pyrimidyl, X of 4-pyrimidyl, and XV of o-pyridyl,
which serve to further delocalize the unpaired electron. There-
fore, the stabilization effect could also be reasonably explained,
at least qualitatively, by additional contributions from these
resonance structures. It seems possible to anticipate, by simply
counting these additional structures, that the barrier to forming
2-pyrimidyl should be less than the barriers required to form
pyrazyl, 4-pyrimidyl, and o-pyridyl. It also follows that the
barriers involved in forming 5-pyrimidyl, m-pyridyl, and p-
pyridyl are similar to the barrier for phenyl radical formation
(112 kcal/mol), since additional resonance structures are not
available for these radicals. The proposed C-H fission barriers
for each channel inferred and derived therein are compared in
Table 10.
In Table 10, the lower barrier for pyrimidine f 4-pyrimidyl
is assumed to be the same as that for pyrazine dissociation. The
barriers for pyrimidine/pyridine f 5-pyrimidyl/o-, p-pyridyl are
the same as for benzene f phenyl. This is consistent with the
discussion of additional resonance contributions of the azyls in
Figure 7. The derived barriers for forming o-pyridyl, pyrazyl,
and 2-pyrimidyl are indeed lower than the 112 kcal/mol of
benzene, but not as low as the values reported by Mackie et
al.7 and Doughty et al.10,11 Furthermore, the difference of the
barrier reduction between pyrazine f pyrazyl and pyrimidine
f 2-pyrimidyl revealed by RRKM analysis (103 vs 98 kcal/
mol) clearly indicates an increased stabilization effect in
Figure 7. Resonance structures for various azyl radicals.
2-pyrimidyl, where C-H fission from ortho-C produces a radical
site adjacent to two N atoms, resulting in additional delocal-
ization. This was not seen in either the ab initio calculations
or the single-pulse modeling (cf. Table 10). Note that the barrier
of pyridine f o-pyridyl, 105 kcal/mol, came from an indepen-
dent RRKM calculation on the refined rates of pyridine
dissociation. This value, nonetheless, is quite consistent with
the 103 kcal/mol for pyrazine. This can now be understood by
realizing that both C-H fissions actually come from the same
environment, the ortho-C, indicating a consistently reduced
barrier during the formation of these radicals. All of these
findings support the above discussion of proposed azyl reso-
nance structures.
(2) Chain Mechanism. With the initiation rates defined by
LS experiments and the time-resolved behavior of the stable

7072 J. Phys. Chem. A, Vol. 101, No. 38, 1997
TABLE 10: Azine C-H Fission Barriers
Kiefer et al.
3. This has been explained through the formation of its
precursor CHCHCN radical by CN addition to the azines.12a
This explanation was inspired in part by the reasoning that its
formation must require the direct involvement of a parent azine
in order to produce an early maximum, since only the reactant
has a high concentration at early reaction time. The formation
of CHCCN is then accomplished by subsequent loss of H from
CHCHCN. The overall CN reaction with azines is thought to
be limited by the rate of initial CN addition to the ring, which
is then followed by fast dissociation of the chemically activated
CN adduct. Many studies have indeed shown that CN reaction
with olefins is dominated by exothermic addition, rather than
abstraction.32
C-H fission
ab initio9 single-pulse10,11 this work
N
N
N
N
•
•
104.0a
109.7
107.8
98
105
112
112
+ H
+ H
+ H
N
N
•
N
N
•
•
+ H
+ H
+ H
+ H
103.8
104.3
103.5
110.7
96.45
95.08
103
98
N
N
N
N
Not only is the above mechanism of CHCHCN formation
theoretically acceptable but it also describes the CHCCN TOF
profiles beautifully for all the azines. As already reported in
pyrazine pyrolysis,12a a considerable decline in CHCCN con-
centration was observed in the presence of excess H2; this
behavior is also observed in the pyridine and pyrimidine
experiments exhibited in Figure 3. This supports the earlier
explanation for pyrazine;12a the excess H2 serves to convert CN
to H atom via reaction 12, thereby reducing the contribution of
the CN addition to the formation of CHCCN. The consistency
of the observed drop in the initial peak in CHCCN due to H2
addition in all the azine experiments verifies the important role
of CN addition. This reaction, forming CHCCN and, at low
temperatures, CH2CHCN, is the unique contribution of CN as
chain carrier.
N
N
N
N
N
N
N
•
102.08
103
112
N
N
N
•
a Values in kcal/mol.
TABLE 11: Comparison of Product Yields Using the
Mechanisms in Tables 3-5 with Single-Pulse Experimentsa
ratios
species
conversion (%)
HCN/C2H2
HCN/CHCCN
pyridine
pyrazine
pyrimidine
6 (11)b
36 (70)
47 (22)
0.5 (0.5)
2.4 (2.3)
2.7 (2.9)
1.6 (1.0)
10.4 (8.1)
9.8 (5.0)
Additional supporting evidence for our mechanism comes
from the observation by Doughty et al. of an increased yield of
acrylonitrile, CH2CHCN, with excess H2 (5%) in their single-
pulse pyrazine experiments.10,11 Using the mechanism herein,
modeling calculations reveal that CH2CHCN concentration is
increased by a factor of ∼4 at 1450 K, the upper temperature
end of the single-pulse experiments. This increase occurs
because the dominant CH2CHCN removal steps, reactions 17
and 18 of Table 3, are reversed by the presence of excess H2.
a Reaction conditions: 0.7%pyridine-Ar, 1450 K and 9 atm, 6 ×
10-4 s; 0.3% diazine-Ar, 1450 K and 14 atm, 8.50 × 10-4 s for
pyrazine and 9.25 × 10-4 s for pyrimidine. b Single-pulse results7,10,11
shown in parentheses.
products by the TOF data, the mechanisms constructed for the
azine decompositions reproduce these product profiles satis-
factorily over the broad temperature range 1600-2300 K. This
demonstrates that the mechanisms offered herein describe the
branching ratios and overall formation rates of the products very
closely. As further demonstration, Table 11 compares the
single-pulse data of Doughty et al.10,11 with our extrapolated
simulation at 1450 K for the three azines. It is apparent that
our model predicts lower conversions of pyridine and pyrazine
than the observations, but a higher conversion of pyrimidine,
both by ∼50%. Compared with pyrazine conversion, the higher
conversion of pyrimidine is anticipated since the previous
RRKM analysis clearly shows that the rate constant of pyrimi-
dine dissociation, when extrapolated to 1450 K, is slightly higher
than that of pyrazine. A slightly faster dissociation of pyrimi-
dine is also consistent with the results of Lifshitz in his
unpublished single-pulse shock-tube study of diazine pyroly-
ses.33 However, Doughty et al., had to adopt a rate constant of
pyrazine almost 3 times higher than that of pyrimidine to model
their high conversion of pyrazine.10,11 As discussed in the
Introduction, the sensitivity of the single-pulse shock-tube
technique to impurities in chain reactions might be responsible
for their observed higher pyrazine conversion.
C2N2 is another trace product in all the azine experiments,
but was not quantified in the single-pulse work of Doughty et
al. This is surprising given the presence of CN radicals and
the inclusion of dissociation or abstraction channels in our
mechanism which inevitably produce C2N2. Our simulation
shows that the dissociation of NCCHN, reaction 4 of Table 3,
is the most important source of C2N2 due to its low dissociation
barrier of 41 kcal/mol. This channel produces a significant
amount of C2N2, especially at the lower temperatures where
there are relatively large concentrations of NCCHN radical.
However, despite the importance of this C2N2-producing chan-
nel, the mechanisms herein are still able to hold C2N2
concentrations to acceptable levels for all the azines over the
entire temperature range, typically ∼2 × 10-9 mol/cm3, by
involving the reverse of reaction 11, CN + HCN f H + C2N2,
to destroy the excess C2N2. This behavior is a consequence of
the large concentrations of H atoms generated in the present
pyrolysis modeling.
Finally, it is worth summarizing the role of CN radical in
azine pyrolyses. CN is an important chain carrier in azine
decomposition, whereas in benzene pyrolysis the only significant
chain carrier is the H atom. In the Doughty et al. study,10,11
the importance of CN as the other chain carrier in addition to
H was attributed to its efficient abstraction of H atoms from
diazine molecules, which was stressed to be the main reason
for the observed fast overall disappearance of diazines. The
present mechanisms, however, invoke the exothermic channels
of CN addition to the parent azines. An analysis of the rate of
Despite the inconsistent prediction of azine conversion, our
modeling gives a satisfactory agreement of the ratios of the main
products with those reported by Doughty et al., especially for
pyridine, as shown in Table 11. This again supports the product
formation channels in our mechanisms.
As shown in Figure 2, a sharp early maximum of CHCCN is
consistently observed at high temperatures in pyrimidine py-
rolysis and also appears for the other azines, as seen in Figure

Pyrolyses of Aromatic Azines
J. Phys. Chem. A, Vol. 101, No. 38, 1997 7073
(9) Jones, J.; Bacskay, G. B.; Mackie, J. C.; Doughty, A. J. Chem.
Soc., Faraday Trans. 1995, 91, 1587.
(10) Doughty, A.; Mackie, J. C. J. Chem. Soc., Faraday Trans. 1994,
90, 541.
(11) Doughty, A.; Mackie, J. C.; Palmer, J. M. 25th International
Symposium on Combustion; The Combustion Institute: Pittsburgh, 1994;
p 893.
the reactions involved reveals that the real reason for the faster
consumption of the azines, besides the reduced C-H fission
barriers, is the ease of dissociation of all the large radicals,
including the azyls and the adducts formed by CN addition to
the azines.
(12) (a) Kiefer, J. H.; Zhang, Q.; Kern, R. D.; Chen, H.; Yao, J.; Jursic,
B. S. 26th International Symposium on Combustion; The Combustion
Institute: Pittsburgh, 1996; Vol. I, p 651. (b) Jursic, B. S. J. Phys. Chem.
A 1997, 101, 2345.
(13) Kiefer, J. H.; Manson, A. C. ReV. Sci. Instrum. 1981, 52, 1392.
(14) Bru¨hl, J. W. Z. Phys. Chem. 1897, 22, 481.
Summary and Conclusions
This study has investigated the pyrolyses of pyrazine,
pyrimidine, and pyridine using the combined LS and TOF
techniques in the shock tube. The experiments cover the
temperature range 1600-2300 K for pressures of 150-350 Torr.
The LS gradient data clearly substantiate the chain nature of
the azine pryolyses which are initiated mainly by ring C-H
fission with a well-resolved induction period. The LS data
further allow an unambiguous comparison of the initial C-H
fission rates for azines. RRKM fits to these rate constants, with
(15) Kiefer, J. H.; Mizerka, L. J.; Patel, M. R.; Wei, H. C. J. Phys.
Chem. 1985, 89, 2013.
(16) Melius, C. F. Calculation of Heat of Formation and Free Energy,
http://herzberg-ca.sandia.gov/∼melius/index.html, June, 1996.
(17) Benson, S. W. Thermochemical Kinetics; John Wiley: New York,
1976.
(18) Harland, P. W. Int. J. Mass Spectrom. Ion Processes 1986, 70,
231.
a constant ∆E
of 400 cm-1 and η (Gorin restriction
down
parameter) of 0.008, both usual for large molecules, yield fission
barriers of 103 ( 2 kcal/mol for pyrazine, 98 ( 2 for pyrimidine,
and 105 ( 2 for pyridine. Comparing with the 112 kcal/mol
for the benzene dissociation barrier, the derived barriers agree
with the assessment by Doughty et al.10,11 that the nitrogen lone
pair reduces the adjacent C-H bond energy in these hetero-
cycles. The magnitudes of the azine barriers are attributed to
the contribution of additional resonance structures of azyl
radicals resulting from the interaction between the nitrogen lone
pair(s) and the neighboring C-H bond.
Furthermore, substantially improved chain mechanisms,
including pathways to the major products HCN, C2H2, CHCCN,
and C4H2, have been proposed to successfully model the LS
and TOF data of all the azines. The modeling is assisted by
high-level calculations for the imine N-H bond energy, which
is then used to estimate the properties of several key radicals
whose dissociation rates and branching ratios largely control
the formation of chain carriers. In particular, the mechanisms
confirm the importance of CN in azine pyrolyses; its primary
role is its highly exothermic addition to a parent azine followed
by a series of very facile dissociations which finally generate
the CHCHCN radical, a precursor to CHCCN and CH2CHCN.
This scheme is consistent with the TOF observation of a severe
reduction of CHCCN concentration in mixtures containing initial
amounts of H2 for all the azine experiments at high temperatures.
The fast overall decomposition of azines is explained in part
by the reduced C-H fission barriers, but mainly by the ease of
dissociation of all large secondary radicals formed here by
dissociation and abstraction reactions.
(19) Dorofeeva, O. V.; Gurvich, L. V. Thermochim. Acta 1991, 178,
273.
(20) Dibeler, V. H.; Reese, R. M.; Franklin, J. L. J. Am. Chem. Soc.
1961, 83, 1813.
(21) (a) Chu, J. Y.; Nguyen, T. T.; King, K. D. J. Phys. Chem. 1982,
86, 443. (b) Stull, D. R.; Westrum, E. F., Jr.; Sinke, G. C. The Chemical
Thermodynamics of Organic Conpounds; John Wiley: New York, 1969.
(22) Besseris, G.; J.; Kiefer, J. H.; Zhang, Q.; Walker, J. A.; Tsang, W.
Int. J. Chem. Kinet. 1995, 27, 691.
(23) Tsang, W. Combust. Flame 1989, 78, 71.
(24) Miller, L. A.; Barker, J. R. J. Chem. Phys. 1996, 105, 1383.
(25) Tsang, W.; Bedanov, V.; Zachariah, M. R. J. Phys. Chem. 1996,
100, 4011.
(26) (a) Tsang, W.; Hampson, R. F. J. Phys. Chem. Ref. Data 1986, 15,
No. 3. (b) Tsang, W. 1987, 16, No. 3; 17, No. 2; 19, No. 1; 20, No. 2.
(27) Nicovich, J. M.; Ravishankara, A. R. J. Phys. Chem. 1984, 88,
2534.
(28) Duran, R. P.; Amorebieti, V. T.; Colussi, A. J. J. Phys. Chem. 1988,
92, 636.
(29) Westmoreland, P. R.; Dein, A. M.; Howard, J. B.; Longwell, J. P.
J. Phys. Chem. 1989, 93, 8171.
(30) Ackermann, L.; Hippler, H.; Pagsberg, P.; Reihs, C.; Troe, J. J.
Phys. Chem. 1990, 94, 5247.
(31) Tsang, W. J. Phys. Chem. 1986, 90, 1152.
(32) (a) Westley, F.; Frizzell, D. H.; Herron, J. T.; Hampson, R.F.;
Mallard, W. J. NIST Chemical Kinetics Database; NIST Standard Reference
Data: Gaithersburg, MD, 1993. (b) Sims, I. R.; Queffelec, J.-L.; Travers,
D.; Rowe, B. R.; Herbert, L. B.; Karthauser, J.; Smith, I. W. M. Chem.
Phys. Lett. 1993, 211, 461. (c) Herbert, L.; Smith, I. W. M.; Spencer-Smith,
R. D. Int. J.Chem. Kinet. 1992, 24, 791. (d) Lichtin, D. A.; Lin, M. C.
Chem. Phys. 1985, 96, 473.
(33) Lifshitz, A. Private communication, 1995.
(34) Ha, T. K.; Gey, E. J. Mol. Struct. (THEOCHEM) 1994, 306, 197.
(35) Pedley, J. B.; Naylor, R. D.; Kirby, S. P. Thermochemical Data of
Organic Compounds, 2nd ed.; Chapman and Hall: London, 1986.
(36) Hewett, K. B.; Shen, M.; Brummel, C. L.; Philips, L. A. J. Chem.
Phys. 1994, 100, 4077.
Acknowledgment. Research was supported by the U.S.
Department of Energy under DE-FGO2-85ER13384 (J.H.K.)
and DE-FGO5-85ER13400 (R.D.K.).
(37) Ponger, G.; Fogarasi, G. Spectrochim. Acta 1992, 48A, 243.
(38) Halverson, F.; Stamm, R. F.; Whalen, J. J. J. Chem. Phys. 1948,
16, 808.
References and Notes
(39) Evans, R. A.; Lorencak, P.; Ha, T.-K.; Wentrup, C. J. Am. Chem.
Soc. 1991, 113, 7261.
(40) Fernholt, L.; Rømming, C. Acta Chem. Acad. A 1978, 32, 271.
(41) Herbert, L.; Smith, I. W. M.; Spencer-Smith, R. D. Int. J. Chem.
Kinet. 1992, 24, 791.
(42) Zabarnick, S.; Lin, M. C. J. Chem. Phys. 1989, 134, 185.
(43) Wooldridge, S. T.; Hanson, R. K.; Bowman, C. T. Int. J. Chem.
Kinet. 1995, 27, 1075.
(44) Kiefer, J. H.; Sidhu, S. S.; Kern, R. D.; Xie, K.; Chen, H.; Harding,
L. B. Combust. Sci. Technol. 1992, 82, 101.
(45) Lifshitz, A.; Bidani, M.; Suslensky, A.; Tamburu, C. J. Phys. Chem.
1989, 93, 1369.
(1) Axworthy, A. E.; Dayan, V. H.; Martin, G. B. Fuel 1987, 57, 29.
(2) Houser, T. J.; McCarville, M. E.; Biftu, T. Int. J. Chem. Kinet.
1980, 12, 555.
(3) Houser, T. J.; Hull, M.; Alway, R. M.; Biftu, T. Int. J. Chem. Kinet.
1980, 12, 569.
(4) Leidreiter, H. I.; Wagner, H. Gg. Z. Phys. Chem. Neue Folge 1987,
153, 99.
(5) Rawlins, W. T.; Tanzawa, T.; Schertzer, S. P.; Kreck, R. H. Final
Report to DOE, 1983, Contract No. DEAC228OPC30392.
(6) Kern, R. D.; Yong, J. N.; Kiefer, J. H.; Shah, J. N. 16th Int. Symp.
Shock Tubes WaVes 1987, 43.
(7) Mackie, J. C.; Colket, M. B.; Nelson, P. F. J. Phys. Chem. 1990,
94, 4099.
(46) Chase, M. W., Jr.; Davies, C. A.; Downey, J. R., Jr.; Frurip, D. J.;
McDonald, R. A.; Syverud, A. N. J. Phys. Chem. Ref. Data 1985, 14, Suppl.
1.
(8) Kikuchi, O.; Hondo, Y.; Morihashi, K.; Nakayama, M. Bull. Chem.
Soc. Jpn. 1988, 61, 291.