Aromatic hydrocarbon growth from indene

Article abstract of DOI:10.1016/S0045-6535(00)00236-8

Aromatic hydrocarbon growth from indene (C₉H₈), which contains the five-membered ring cyclopentadienyl moiety, was investigated experimentally in a 4 s flow reactor over a temperature range 650-850°C. Major products observed were three C₁₈H₁₂ isomers (chrysene, benz[a]anthracene and benzo[c]phenanthrene), two C₁₇H₁₂ isomers (benzo[a]fluorene and benzo[b]fluorene), and two C₁₀H₈ isomers (naphthalene and benzofulvene). Reaction pathways to these products are proposed. Indenyl radical addition to indene produces a resonance-stabilized radical intermediate which further reacts by one of two routes. Rearrangement by intramolecular addition produces a bridged structure that leads to the formation of C₁₇H₁₂ and C₁₀H₈ products. Alternatively, β scission produces biindenyl, which leads to the formation of C₁₈H₁₂ products by a ring condensation mechanism analogous to that proposed for cyclopentadiene-to-naphthalene conversion. Temperature dependencies of both the partitioning between these two routes and the product isomer distributions are consistent with thermochemical modeling using semi-empirical molecular orbital methods. The results further illustrate the role of resonance-stabilized radical rearrangement in aromatic growth and condensation of systems with cyclopentadienyl moieties.

Full text of DOI:10.1016/S0045-6535(00)00236-8

Chemosphere 42 (2001) 625±633
q
Aromatic hydrocarbon growth from indene
Mingming Lu, James A. Mulholland *
Environmental Engineering, Georgia Institute of Technology, 200 Bobby Dodd Way, Atlanta, GA 30332-0512, USA
Received 1 August 1999; received in revised form 1 November 1999
Abstract
Aromatic hydrocarbon growth from indene (C
was investigated experimentally in a 4 s ¯ow reactor over a temperature range 650±850°C. Major products observed
were three C18 12 isomers (chrysene, benz[a]anthracene and benzo[c]phenanthrene), two C17 12 isomers (ben-
zo[a]¯uorene and benzo[b]¯uorene), and two C10 isomers (naphthalene and benzofulvene). Reaction pathways to
9 8
H ), which contains the ®ve-membered ring cyclopentadienyl moiety,
H
H
H
8
these products are proposed. Indenyl radical addition to indene produces a resonance-stabilized radical intermediate
which further reacts by one of two routes. Rearrangement by intramolecular addition produces a bridged structure that
8
leads to the formation of C17H12 and C10H products. Alternatively, b scission produces biindenyl, which leads to the
formation of C18H12 products by a ring condensation mechanism analogous to that proposed for cyclopentadiene-to-
naphthalene conversion. Temperature dependencies of both the partitioning between these two routes and the product
isomer distributions are consistent with thermochemical modeling using semi-empirical molecular orbital methods. The
results further illustrate the role of resonance-stabilized radical rearrangement in aromatic growth and condensation of
systems with cyclopentadienyl moieties. Ó 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Indene; PAH; Soot; Semi-empirical MO modeling; Toxic combustion byproducts
1
. Introduction
of PAH formation and growth in ¯ames focussed on
acetylene addition pathways (Frenklach et al., 1984).
More recently, the importance of resonance-stabilized,
Formation of polycyclic aromatic hydrocarbons
PAHs) and soot in ¯ames continues to be studied due to
(
5 5
or neutral, cyclopentadienyl radicals (C H ) in aromatic
the complexity of the reaction mechanisms. These
products of incomplete combustion are of interest be-
cause several PAHs are known to be mutagens and
carcinogens and exposure to ®ne soot particles, which
can carry PAHs deep into the lung, has been linked to
increased rates of morbidity and mortality. Early study
ring formation has been studied. Fuels containing cy-
clopentadienyl moieties have high sooting tendencies
(McEnally and Pfe€erle, 1998). Cyclopentadienyl radi-
cals are also produced in ¯ames by phenoxy radical ring
contraction (Manion and Louw, 1989). Unimolecular
decomposition of cyclopentadienyl radical at high tem-
peratures produces acetylene and resonance-stabilized
propargyl radical (Roy et al., 1998) which contribute to
aromatic growth. Cyclopentadienyl radicals decompose
slowly, however, due to their stability; therefore, they
can contribute to PAH growth directly by combining to
form naphthalene (Manion and Louw, 1989; Castaldi
et al., 1996; Marinov et al., 1996; Melius et al., 1996).
While there has been detailed study of the chemical
mechanism of naphthalene formation from cyclopent-
q
Presented at The Sixth International Congress on Toxic
Combustion Byproducts June 27±30, 1999, University of
Karlsruhe, Session 6: Soot.
*
Corresponding author. Tel.: +1-404-894-1695; fax: +1-404-
8
94-8266.
E-mail address: james.mulholland@ce.gatech.edu (J.A.
Mulholland).
0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 2 3 6 - 8

6
26
M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
adiene (Melius et al., 1996), there has been little study of
the reactivity of its benzo derivatives, such as indene.
Laskin and Lifshitz (1998) studied the thermal decom-
position of indene, but not carbon growth by indenyl
radical addition. McEnally and Pfe€erle (1998) studied
naphthalene formation from indene and methane, but
did not report on higher molecular weight products. In
this paper, we present experimental and computational
results on aromatic growth from indene. We extend the
cyclopentadiene-to-naphthalene mechanism of Melius
hydrofulvalene (1), which rearranges to form the more
stable 1,10-dihydrofulvalene (2). Loss of H-atom from
the 10-carbon site produces a tertiary, resonance-stabi-
lized radical (3). Isomerization by 1,3-H shift produces
(4), which undergoes three-membered ring closure to
form (5). Ring opening by b scission of a C±C bond
produces a six-membered ring (6). Reactions producing
(5) and (6) are repeated to produce (7) and (8), respec-
tively, converting a second ®ve-membered ring to six-
membered ring. In the ®nal step, b scission yields
naphthalene (9) and H-atom.
9 8
to the conversion of indene (C H ) to C18H12 isomers
that contain four six-membered rings. In addition, a
second reaction pathway is proposed for the formation
of C17
H
12 and C10
H
8
products from indene. Semi-em-
2.2. Comparison of heats of reaction calculated by Ab
initio and semi-empirical methods
pirical molecular orbital modeling results are presented
to explain the temperature dependence of partitioning
between the two pathways as well as the distribution of
Heats of formation of stable intermediates calculated
by Melius using the ab initio BAC-MP2 method are
compared with those calculated using semi-empirical
methods in Fig. 2. Three semi-empirical methods based
on the neglect of diatomic di€erential overlap approxi-
mation were evaluated: modi®ed neglect of diatomic
overlap (MNDO), Austin model 1 (AM1) and para-
metric method 3 (PM3). While the absolute values of the
heats of formation calculated using the ab initio BAC-
MP2 method and three semi-empirical methods di€er
signi®cantly, the relative energy levels, or heats of re-
action, are similar (Table 1). For example, fulvalene
formation is not favored thermodynamically by any of
the methods (Fig. 2). Biggest discrepancies between
semi-empirical and BAC-MP2 heats of reaction are for
the ®rst and last steps: the combination of two cyclo-
pentadienyl radicals to form 9,10-dihydrofulvalene (1)
and the formation of naphthalene (9) by C±H bond b
scission. Overall, PM3 and AM1 heats of reaction cor-
relate well with the BAC-MP2 values (0.95 and 0.92 R-
squared values, respectively); the MNDO R-squared
value is 0.83.
18
C H12 isomer products.
2
. Naphthalene formation from cyclopentadienyl radicals
.1. Reaction mechanism
2
Melius et al. (1996) studied cyclopentadienyl radical
reactions using the quantum chemical BAC-MP2 and
BAC-MP4 methods. After investigating reaction ener-
gies of alternative routes, they proposed a nine-step re-
action mechanism for the conversion of two
cyclopentadienyl radicals to naphthalene via rearrange-
ments involving three-membered ring closing and
opening of resonance-stabilized radicals (Fig. 1). Di-
merization of cyclopentadienyl radicals yields 9,10-di-
Modeling of the cyclopentadiene system demon-
strates that the formation of ®ve-membered ring dimers
and their subsequent rearrangement to ortho-fused six-
membered ring structures via three-membered ring in-
termediates is a plausible mechanism for aromatic
growth. Here, we apply this mechanism to indene, a
benzo derivative of cyclopentadiene. Experiments were
performed to test the hypothesis that aromatic growth
from indene would occur by biindenyl formation and
rearrangement to C18H12 products that contain four
ortho-fused six-membered rings. Due to the large num-
ber of atoms in this system, computational investigation
using ab initio methods was not feasible. Therefore,
semi-empirical molecular orbital methods were used to
estimate reaction path thermodynamic barriers. We
have also used semi-empirical methods to estimate
transition state barriers; these results are being reported
elsewhere.
Fig. 1. Reaction mechanism proposed by Melius et al. (1996)
for the conversion of two cyclopentadienyl radicals to naph-
thalene (9) via 9,10-dihydrofulvalene (1) and 1,10-di-
hydrofulvalene (2).

M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
627
Fig. 2. Thermodynamic energy diagram for the conversion of two cyclopentadienyl radicals to naphthalene (9). Structures are shown
in Fig. 1, with the exception of fulvalene (C10 ). Heats of formation of stable and radical species, but not transition states, were
calculated using the BAC-MP2 ab initio method of Melius et al. (1996) and the MNDO, AM1 and PM3 semi-empirical methods.
H
8
Table 1
Standard heats of reaction for cyclopentadiene-to-naphthalene
mechanism
Unreacted indene, aromatic products and soot were
collected on an inert ®lter (0.2 lm) followed by two ice-
cooled dichloromethane (DCM) impingers in series.
Filter, impinger and rinse samples were combined and
soot, de®ned as the DCM-insoluble fraction, was sepa-
rated by vacuum ®ltration and measured gravimetrical-
ly. Samples were then analyzed using a Varian 2000 GC/
MS with DB5-ms capillary column and a Hewlett±
Packard 1050 HPLC with UV-vis diode array detector.
Products were identi®ed and quanti®ed using chemical
standards, except for biindenyl and methyl benzo¯uo-
rene products which were identi®ed on the basis of
elution order and spectral data and quanti®ed using GC/
MS response factors for compounds of similar size and
structure.
DHr;298°, kcal/mol
a
BAC-MP2
MNDO AMI
PM3
2
C
5
H
5
® (1) )52.0
)22.9
)7.3
51.3
16.2
10.2
)20.8
5.5
39.7
)3.6
59.3
)38.8
)4.5
60.1
(
(
(
(
(
(
(
(
1) ® (2)
2) ® (3) + H
3) ® (4)
4) ® (5)
5) ® (6)
6) ® (7)
7) ® (8)
8) ® (9) + H
)2.2
62.6
17.2
5.3
17.7
12.8
15.3
12.1
)18.4
13.1
)33.9
18.9
)33.0
13.7
)24.6
12.7
)32.7
42.5
)41.7
35.4
)33.3
33.7
a
BAC-MP2 data from Melius et al. (1996).
A temperature range 650±850°C was studied, the
range over which 90% of the indene was reacted. Below
3
. Experimental methods
6
50°C, thermal decomposition of indene was not ob-
A laminar ¯ow, isothermal quartz tube reactor was
served. Above 850°C, ring fragmentation pathways be-
came signi®cant. Two to four experiments were
performed at each temperature. A total carbon balance
was calculated for each experiment, with results ranging
from 100% for the lower temperature experiments to
used to study indene pyrolysis chemistry. Indene vapor
was generated by loading 20 ll of indene (99% purity)
into a glass bulb heated to 140°C (40°C below the indene
boiling point), and then passing through helium carrier
gas at a rate of 390 standard ml/min until all of the
indene was vaporized and delivered to the reactor
through heated transport lines. The gas stream con-
taining 0.2 mol% (nominal) indene vapor had a 4 s
residence time in the reactor. An experiment took ap-
proximately 5 min to complete, with the entire product
stream immediately quenched at the outlet of the reactor
and sampled. Gas samples taken during an experiment
demonstrated that the indene vapor concentration de-
livered to the reactor was constant over 90% of the ex-
periment.
8
5% at 850°C.
4. Results and discussion
4.1. Pyrolysis product yields
Yields of unreacted indene, total aromatic products,
and soot are shown in Fig. 3 for the temperature range
650±850°C. Data are presented on a percent carbon feed

6
28
M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
basis, with soot assumed to be 100% carbon. Also shown
is total carbon recovery. Average values are indicated by
the points, with the range of values from three experi-
ments shown at 825°C. At 650°C, less than 3% of the
indene was reacted; at 850°C, about 90% of the indene
was reacted. Total aromatic product and soot yields
increased with temperature. Unrecovered carbon, which
did not exceed 15% for any experiment, was primarily
due to unrecovered soot that could not be rinsed from
the from the quartz tube wall at the point of quenching.
For temperatures above 850°C, low carbon recovery and
a broad range of aromatic products were observed, in-
dicative of indene ring fragmentation.
Yields of aromatic product groups, representing over
99% of the total aromatic product yield, are shown in
Fig. 4. As expected based on the cyclopentadiene-to-
naphthalene mechanism, three C18
found as major products, and both 1,1- and 1,2-biin-
denyls (C18 14) were observed. Two C10 isomers, two
H12 isomers were
H
H
8
C
C
17
18
H
12 isomers (benzo¯uorenes), and two additional
14 isomers that were identi®ed as methyl benzo¯u-
H
orenes were also observed. Impurities in the indene re-
actant are likely responsible for the 1% yields of C10
species at low temperatures. At 750°C and above, the
12 product yields were of similar mag-
nitude, suggesting that C10 and C17 12 products may
H
8
10 8
C H and C17H
H
8
H
Fig. 3. Total carbon recovery and yields of unreacted indene, total aromatic products and soot from indene pyrolysis. Average values
are shown, with a range of values given at 825°C.
Fig. 4. Aromatic product yields from indene pyrolysis. In parentheses, the numbers of isomers detected are given. Average yields are
shown, with a range of values indicated at one temperature.

M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
629
have been formed together from three indene molecules.
The temperature dependence of the biindenyl and the
methyl benzo¯uorene yields suggests that these com-
adiene-to-naphthalene mechanism shown in Fig. 1 is
extended to the indene system, with analogous num-
bering of reaction intermediates shown in Fig. 6. In the
indene mechanism, four routes are possible for the sec-
ond ring closure step producing three-membered ring
intermediates (7). Subsequent ring opening produces (8),
followed by H-atom loss to produce the three observed
pounds are intermediates to C18
products, respectively.
Isomer fractions for C10
products are shown in Fig. 5. In the top panel, the C10
H12 products and
17 8
C H12/C10H
8 12
H , C17H12 and C18H
H
8
product isomer distribution shows a decreasing benzo-
fulvene-to-naphthalene ratio with increasing tempera-
ture. McEnally and Pfe€erle (1998) have proposed a
route of indene conversion to naphthalene via benzo-
fulvene, which is consistent with our results. The single
carbon required for indene conversion to benzofulvene
C
18
H
12 isomers (9); benz[a]anthracene has a reaction
path degeneracy of 2, consistent with its lack of sym-
metry. The fact that only these three C18 12 isomers
H
were observed supports the proposed mechanism.
A second pathway of aromatic growth is proposed to
explain the formation of C10
8
H and C17H12 products.
is likely generated by the formation of two C17
H
12 iso-
Instead of indenyl radical±radical combination to form
1,1-biindenyl (1), indenyl radical addition to the non-
aromatic indene double bond can occur. Addition is
favored at the 2-carbon site of indene such that a reso-
nance-stabilized radical is produced. This radical can
undergo either intramolecular addition to form a
bridged radical intermediate (route A) or b scission of a
C±H bond to produce 1,2-biindenyl (route B). The latter
mers, benzo[a]¯uorene and benzo[b]¯uorene, from
methyl benzo¯uorene decomposition. Benzo[b]¯uorene
was produced in greater quantities than benzo[a]¯uorene
by a factor of 3 and more (middle panel, Fig. 5), con-
sistent with the thermodynamic properties of these iso-
mers.
Three of the ®ve ortho-fused C18H12 isomers were
produced: chrysene, benz[a]anthracene and ben-
zo[c]phenanthrene (bottom panel, Fig. 5). Two other
route proceeds to the C18H12 isomer products as before.
The bridged radical intermediate, on the other hand, can
rearrange to form methyl benzo[b]¯uorenes. Subsequent
C
18
H12 isomers, triphenylene and naphthacene, were not
found. The relative yields of the three isomers that were
formed are consistent with their relative stabilities, with
chrysene, the most stable isomer, produced in highest
yield and benzo[c]phenanthrene, the least stable isomer,
produced in lowest yield.
loss of CH
which can isomerize to form benzo[a]¯uorene. Addition
of CH to indene leads to the formation of benzoful-
2
results in the formation of benzo[b]¯uorene,
2
vene, which rearranges to form naphthalene.
Reaction path thermodynamic energies calculated
using the PM3 method are shown in Fig. 7. Reaction
enthalpies favor the formation of benzo¯uorene prod-
ucts via the bridged intermediate (A) over the formation
4
.2. Reaction pathways
Reaction pathways are proposed to account for the
observed aromatic products. The nine-step cyclopent-
of C18
for indenyl radical conversion to C18 12 products is very
H
12 products via biindenyl (B). Heats of reaction
H
8
Fig. 5. Isomer fractions of C10H , C17H12 and C18H12 products from indene pyrolysis. Average values are shown, with a range of
values indicated at one temperature.

6
30
M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
Fig. 6. Proposed reaction mechanism for the formation of C10
pounds that were measured in the experiment.
8
H , C17H12 and C18H12 products from indene. Boxes designate com-
similar to that of cyclopentadienyl radical conversion to
naphthalene, as expected. The reaction step that deter-
(a), which leads to chrysene, is favored by 9.7 kcal/mol
over route (c), which leads to benz[a]anthracene, and by
15 kcal/mol over route (b), which leads to benzo[c]phe-
nanthrene. Route (d) has the highest thermodynamic
mines which C18H12 product isomer is formed is the
rearrangement of intermediate (6) to produce (7). Route
Fig. 7. PM3 thermodynamic energy diagram for the formation of C18
refer to structures shown in Fig. 6.
H12 products and methyl benzo¯uorenes from indene. Labels

M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
631
energy barrier. E€ects of reaction temperature on the
partitioning between benzo¯uorene and C18 12 product
pathways and the C18 12 isomer distribution are ex-
amined in the following sections.
4.4. Temperature dependence of C18
distribution
H
12 isomer product
H
H
To evaluate the degree to which thermodynamic
barriers control the rate of formation of C18 12 product
isomers for the ane set of reactions proposed, ther-
modynamic properties of the C18 12 products and rad-
H
4
.3. Temperature dependence of partitioning between
H
benzo¯uorene and C18 12 product pathways
H
ical intermediates were calculated. Di€erences in isomer
heats of formation and symmetry numbers, whose ratios
represent statistical factors (Bishop and Laidler, 1964),
a€ect the thermodynamic distribution of isomers. In
Table 2, experimental measurements of the heats of
formation of three-ring and four-ring ortho-fused PAHs
are compared with calculated values using ab initio and
semi-empirical methods. While the BAC-MP2 ab initio
method is clearly superior in predicting absolute heat of
formation, the AM1 and PM3 semi-empirical methods
predict di€erences in the heats of formation of these
isomers that agree well with the experimental di€erences.
Experimental and calculated thermodynamic isomer
The ratio of the total C18H12 product yield to the
total benzo¯uorene product yield, shown in Fig. 8,
demonstrates that reaction path (A) is favored at low
temperatures and reaction path (B) is favored at high
temperatures. The results are consistent with kinetic
parameters for path (A) that are 30 kcal/mol lower in
enthalpy (favorable) and 30 cal/mol/K lower in entropy
(
unfavorable) than for path (B). Thermodynamically,
reaction (A) is found to be less endothermic than reac-
tion (B) by 48 kcal/mol (Fig. 7), as calculated by the
PM3 method, and path (A) is expected to have an un-
favorable entropy of reaction due to loss of internal
rotation. This entropy loss is estimated to be 8 cal/mol/K
using the method of Pitzer (1953) for free internal ro-
tation and modi®ed for hindered rotation (Mulholland
et al., 1993). Transition state calculations are needed to
assess actual kinetic barriers. The di€erence in kinetic
enthalpy barriers is expected to be smaller than the
di€erence in thermodynamic heats of reaction since the
activation energy of reaction (A) should have a larger
electronic component than that of reaction (B). The
di€erence in entropy barriers is expected to be greater
than that due only to the loss of internal rotation en-
tropy in reaction (A) since this reaction also involves a
strained bridged structure.
fractions for C18H12 products are compared in Fig. 9. At
high temperatures and high indene conversion, the ex-
perimental distribution appears to approach a thermo-
dynamic distribution of products. At low temperature
and low conversion, chrysene is favored by an amount
consistent with the thermodynamic barriers for the for-
mation of radical intermediates (7). The range in en-
thalpies of the intermediates is greater than that of the
products due largely to di€erences in aromaticity of
these intermediates, as shown in Fig. 6 by the number six
p-electron systems in each of the structures. At low
temperatures, when the relative rates of intermediate (7)
formation are controlling, and at high temperatures,
when the relative rates of C18H12 product (9) formation
Fig. 8. Ratio of total C18H12 product yield to total benzo¯uorene product yield. The curve represents a ®t to the data.

6
32
M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
Table 2
Symmetry numbers and standard heats of formation for C14
H10 and C18
H
12 compounds
Sym.
DHf;298°, kcal/mol
a
b
Exp
BAC-MP2
MNDO
AMI
PM3
Phenanthrene
Anthracene
Triphenylene
2
4
6
2
1
2
4
49.5
54.4
61.9
62.8
69.5
69.8
72.3
47.5
54.4
60.1
61.4
64.6
65.2
75.2
55.7
58.8
76.9
75.2
75.1
79.8
81
57.4
62.9
75.5
76.0
78.4
81.0
87.0
55.0
61.7
68.3
70.9
74.7
77.6
84.4
Chrysene
Benz[a]anthracene
Benzo[c]phenanthrene
Naphthacene
a
Experimental data from TRC Thermodynamic Table (1995) and Cox and Pilcher (1970).
BAC-MP2 data from Melius et al. (1996).
b
Fig. 9. Thermodynamic isomer distributions of C18H12 products (solid lines), calculated by BAC-MP2, AM1 and PM3 methods, and
radical intermediates (dashed lines ± PM3). Experimental data are shown by symbols: chrysene ± circles; benz[a]anthracene ± triangles;
benzo[c]phenanthrene ± squares. Radical intermediates are those labeled (7) in Fig. 6.
are controlling, the respective thermodynamic barriers
appear to predict the experimental results. Consistent
with this result, Melius found the activation energies,
calculated using the BAC-MP2 method, of the two
analogous cyclopentadiene-to-naphathalene reactions to
be the same as the thermodynamic barriers for these
reactions.
to six-membered ring structures. In addition, we pro-
pose a second pathway to explain the formation of
benzo¯uorene products which retain one cyclopentadi-
1
ene moiety and provide a C fragment that adds to
indene to produce naphthalene via benzofulvene. E€ects
of temperature on the overall relative rates of the two
mechanisms and on the distribution of C18H12 isomer
products are shown to be consistent with thermody-
namic properties estimated by semi-empirical quantum
chemical methods.
5
. Conclusion
Our results o€er further support the importance of
resonance-stabilized radicals in aromatic and soot
growth mechanisms. We have extended a mechanism of
naphthalene formation from cyclopentadiene to the
Acknowledgements
The support of the National Science Foundation
under grant no. CTS-9457028 is gratefully acknowl-
edged.
formation of three C18H12 isomers from indene, in-
volving the conversion of ®ve-membered ring moieties

M. Lu, J.A. Mulholland / Chemosphere 42 (2001) 625±633
633
References
aromatic hydrocarbon formation in premixed methane and
ethane ¯ames. Combust. Sci. Technol. 116±117, 211±287.
McEnally, C.S., Pfe€erle, L.D., 1998. An experimental study in
non-premixed ¯ames of hydrocarbon growth processes that
involve ®ve-membered carbon rings. Combust. Sci. Technol.
Bishop, D.M., Laidler, K.J., 1964. Symmetry numbers and
statistical factors in rate theory. J. Chem. Phys. 42, 1688±
1
691.
1
31, 323±344.
Castaldi, M.J., Marinov, N.M., Melius, C.F., Huang, J.,
Senkan, S.M., Pitz, W.M., Westbrook, C.K., 1996. Exper-
imental and modeling investigation of aromatic and poly-
cyclic aromatic hydrocarbon formation in a premixed
ethylene ¯ame. Proc. Combust. Inst. 26, 693±702.
Melius, C.F., Colvin, M.E., Marinov, N.M., Pitz, W.J., Senkan,
S.M., 1996. Reaction mechanisms in aromatic hydrocarbon
formation involving the C
5 5
H cyclopentadienyl moiety.
Proc. Combust. Inst. 26, 685±692.
Mulholland, J.A., Mukherjee, J., Wornat, M.J., Saro®m, A.F.,
Rutledge, G.C., 1993. Semiempirical molecular orbital
estimation of the relative stability of bianthryls produced
by anthracene pyrolysis. Combust. Flame 94, 233±243.
Pitzer, K.S., 1953. Quantum Chemistry, Prentice-Hall, New
York, pp. 443±444.
Cox, J.D., Pilcher, G., 1970. Thermochemistry of organic and
organometallic compounds. Academic Press, London, pp.
1
80±191.
Frenklach, M., Clary, D.W., Gardiner, W.C., Stein, S.E., 1984.
Detailed kinetic modeling of soot formation in shock-tube
pyrolysis of acetylene. Proc. Combust. Inst. 20, 887±901.
Laskin, A., Lifshitz, A., 1998. Thermal decomposition of
indene. Experimental results and kinetic modeling. Proc.
Combust. Inst. 27, 313±320.
Roy, K., Horn, C., Frank, P., Slutsky, V.G., Just, T., 1998.
High-temperature investigations on the pyrolysis of cyclo-
pentadiene. Proc. Combust. Inst. 27, 329±336.
TRC, 1995. TRC Thermodynamic tables ± hydrocarbons,
thermodynamic research center. The Texas University
System, College Station, TX, pp. 3800,3880,3970.
Manion, J.A., Louw, R., 1989. Rates, products, and mecha-
nisms in the gas-phase hydrogenolysis of phenol between
9
22 and 1175 K. J. Phys. Chem. 93, 3563±3574.
Marinov, N.M., Pitz, W.J., Westbrook, C.K., Castaldi, M.J.,
Senkan, S.M., 1996. Modeling of aromatic and polycyclic