Experimental and kinetic modeling study of the oxidation of benzene

Article abstract of DOI:10.1002/1097-4601(2000)32:8<498::AID-KIN8>3.0.CO;2-H

The oxidation of benzene under flow-reactor conditions has been studied experimentally and in terms of a detailed chemical kinetic model. The experiments were performed under plug-flow conditions, at excess air ratios ranging from close to stoichiometric to very lean. The temperature range was 900-1450 K and the residence time of the order of 150 ms. The radical pool was perturbed by means of varying the concentration of water vapor and by adding NO. Furthermore, a few experiments were conducted on pyrolysis and oxidation of phenol. Benzene oxidation is initiated at approximately 1000 K; the temperature for complete oxidation depends on stoichiometry, ranging from 1100 K (very lean conditions) to 1300 K (close to stoichiometric). The water vapor level and the presence of NO have only a minor impact on the temperature regime for oxidation. The proposed chemical kinetic model was validated by comparison with the present experimental data as well as flow reactor and mixed reactor data from literature. The model provides a reasonably good description of the overall oxidation behavior of benzene over the range of conditions investigated. However, before details of the oxidation behavior can be predicted satisfactorily, a number of kinetic issues need to be resolved. These include product channels and rates for the reactions of phenyl and cyclopentadienyl with molecular oxygen as well as reaction chemistry for the oxygenated cyclic compounds formed as intermediates in the oxidation process.

Full text of DOI:10.1002/1097-4601(2000)32:8<498::AID-KIN8>3.0.CO;2-H

Experimental and Kinetic
Modeling Study of the
Oxidation of Benzene
MARIA U. ALZUETA,* PETER GLARBORG, KIM DAM-JOHANSEN
Department of Chemical Engineering, Technical University of Denmark, 2800 Lyngby, Denmark
Received 16 November 1999; accepted 20 March 2000
ABSTRACT: The oxidation of benzene under flow-reactor conditions has been studied exper-
imentally and in terms of a detailed chemical kinetic model. The experiments were performed
under plug-flow conditions, at excess air ratios ranging from close to stoichiometric to very
lean. The temperature range was 900–1450 K and the residence time of the order of 150 ms.
The radical pool was perturbed by means of varying the concentration of water vapor and by
adding NO. Furthermore, a few experiments were conducted on pyrolysis and oxidation of
phenol. Benzene oxidation is initiated at ϳ1000 K; the temperature for complete oxidation
depends on stoichiometry, ranging from 1100 K (very lean conditions) to 1300 K (close to
stoichiometric). The water vapor level and the presence of NO have only a minor impact on
the temperature regime for oxidation. The proposed chemical kinetic model was validated by
comparison with the present experimental data as well as flow reactor and mixed reactor data
from literature. The model provides a reasonably good description of the overall oxidation
behavior of benzene over the range of conditions investigated. However, before details of the
oxidation behavior can be predicted satisfactorily, a number of kinetic issues need to be
resolved. These include product channels and rates for the reactions of phenyl and cyclopen-
tadienyl with molecular oxygen as well as reaction chemistry for the oxygenated cyclic com-
pounds formed as intermediates in the oxidation process. ᭧ 2000 John Wiley & Sons, Inc. Int J
Chem Kinet 32: 498–522, 2000
INTRODUCTION
The oxidation behavior ofaromatic compounds is
atypical compared to other hydrocarbons, due to the
characteristics ofthe aromatic ring and its rupture.
During the last two decades, significant experimental
and theoretical efforts [1–19] have been aimed at im-
proving our understanding ofthe oxidation behavior
ofthese species. In his early review on aromatics ox-
idation, Brezinsky [4] identified a number ofimportant
unresolved issues. At present, more than a decade
later, many ofthese issues remain unresolved. This can
mainly be attributed to the complexity ofthe aromatics
chemistry but also the limited availability ofwell-
characterized experimental data has limited progress
in our understanding. Large uncertainties still remain
in the reaction mechanisms for even the simplest ar-
omatic species, benzene and toluene.
The oxidation chemistry ofaromatic compounds is of
significant theoretical and practical interest. Aromatic
compounds are known to be harmful to the environ-
ment and the emission ofthese species fr om a number
ofcombustion systems is a significant concern. Fur-
thermore, aromatic species are important precursors to
dioxins and to soot formation. Aromatic compounds
are formed to some extent in most combustion pro-
cesses; in addition, they are added in considerable
quantities to unleaded gasolines in order to increase
the octane number and prevent knock in engines.
Correspondence to: P. Glarborg (pgl@kt.dtu.dk)
*
Present Address: Department ofChemical and Environmental
Engineering, Centro Politecnico Superior, University ofZaragoza,
0015 Zaragoza, Spain
2000 John Wiley & Sons, Inc.
Previous experimental studies dealing with benzene
oxidation include flow reactor studies [2,4,6,7]; well-
5
᭧

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
499
stirred reactor experiments [19]; shock tube CO data
3] and ignition delay data [5]; and low-pressure flame
composition profiles [1,15] as well as flame speed
measurements [16].
securing a uniform temperature profile within 10 K
over the reactor. The reactor temperature was mea-
sured close to the reactor with a Type-K thermocouple.
The gaseous components were fed to the reactor in
four separate streams. The main flow usually con-
tained nitrogen and water vapor, and the rest ofreac-
tants were supplied through side injector tubes. In the
experiments for analyzing phenol oxidation, phenol
was dissolved in water and a nitrogen stream was sat-
urated in the phenol solution at a given temperature.
The inlet concentration ofphenol was determined by
complete oxidation at very lean conditions and high
[
A number ofmodeling studies ofbenzene oxidation
have been reported. In an early effort, Bittker [8] pro-
posed a simplified mechanism to model the ignition
delay time measurements ofBurcat et al. [5] and the
flow reactor data ofLovell et al. [6]. This mechanism
was in part empirical, since the rate constants for a
number ofthe most important reactions were adjusted
to fit the experimental data. Based on the review [4]
and flow-reactor experiments [6], Brezinsky and co-
workers proposed the first detailed reaction mecha-
nism for benzene and toluene oxidation [9]. This
mechanism has been the starting point ofmost sub-
sequent modeling studies, including those ofLindstedt
and Skevis [10], Zhang and McKinnon [12], Tan and
Frank [14], and Davis et al. [16]. In these studies, ben-
zene oxidation mechanisms were tested against data
from flames, mainly the low-pressure flame species
profiles ofBittner and Howard [1]. In general, these
comparisons confirm that important reaction paths are
missing in the mechanism.
temperature (i.e., 50% O and 1500 K). In most ex-
2
periments, very diluted gas mixtures were used in or-
der to maintain isothermal conditions along the reac-
tion zone. The reactor tube used in the present
experiments had a diameter of0.89 cm and a length
of19 cm. The product gas was quenched at the outlet
ofthe reactor by heat exchange with air.
The concentrations ofCO, CO , O , and NO were
2
2
measured continuously by spectrophotometric and
paramagnetic analyzers, with an accuracy of Ϯ3% but
not less than 5 ppm. Furthermore, FTIR (Fourier-
transformed infrared) spectra were taken, from which
the concentrations ofbenzene have been determined.
The uncertainty ofthese measurements has been esti-
mated to be Ϯ10% but not less than 10 ppm. The
reactor pressure was measured with an absolute pres-
sure transducer. Temperature control and data acqui-
sition were automated, using a personal computer.
Progress in the understanding ofthe benzene oxi-
dation chemistry has been facilitated by recent studies
ofimportant reaction subsets involving C and C spe-
6
5
cies, [17,18,20–23] as well as the improved reliability
ofthe C –C hydrocarbon reaction subset [24–29].
1
4
The availability ofthermodynamic properties is also
enhanced [17,18,23,28,30,31], although significant
uncertainties remain.
The objective ofthe present work is to per fo rm an
experimental and theoretical study ofbenzene oxida-
tion under lean conditions and temperatures charac-
teristic ofpostflame conditions. Perturbation experi-
ments on benzene oxidation, supplemented with
experiments on pyrolysis and oxidation ofphenol, are
carried out in an atmospheric pressure flow reactor in
the temperature range 900–1400 K. Experimental data
from the literature on flow reactor [6] and well-stirred
reactor [19] experiments are also considered. A de-
tailed reaction mechanism for the oxidation of benzene
is proposed, and some important issues in our under-
standing ofthe oxidation benzene are discussed.
REACTION MECHANISM
The reaction mechanism for benzene oxidation devel-
oped in the present work consists ofan oxidation sub-
set for C –C hydrocarbons, together with a small
1
6
subset for reactions involving NO. The C –C6
3
hydrocarbon subset is listed in Appendix A; the full
reaction mechanism can be obtained directly from the
authors in Chemkin format.
The model is an extension ofthe reaction mecha-
nism developed for parabenzoquinone pyrolysis and
oxidation by Alzueta et al. [22], drawing on the ben-
zene oxidation scheme by Emdee et al. [9]. The C5
hydrocarbon subset is mainly adopted from the recent
work ofZhong and Bozzelli [17,18,35], while the C –
3
EXPERIMENTAL
C hydrocarbon chemistry is based on the work of
4
Miller and co-workers [24,27] and Marinov et al. [28].
The C –C /NO subset has been adopted without
The flow reactor and the experimental procedure are
described elsewhere [32,33,34], and only a briefde-
scription is given here. A quartz flow reactor designed
for obtaining plug flow in the laminar flow regime [32]
was placed in a three-zone electrically heated oven,
1
2
changes from Glarborg et al. [29].
The reverse rate constants were obtained from the
forward rate constants and thermodynamic equilib-
rium constants. Thermodynamic data were taken

5
00
ALZUETA, GLARBORG, AND DAM-JOHANSEN
mostly from the Sandia Thermodynamic Database
30], Burcat and McBride [31], and Zhong and Boz-
zelli [17,18,35]. Appendix B lists thermodynamic
The reaction ofC H with OH radicals proceeds
6 6
[
mainly through hydrogen abstraction (3), even though
the product channel forming phenol (Ϫ22) is almost
as fast. The rate constants for these steps are accurate
within a factor of two [26].
properties (⌬Hf298, Sf298, and C ) for selected species.
p
The availability and accuracy ofthe aromatics’ ther-
mochemistry have improved considerably in recent
years, but for a number of important intermediates,
large uncertainties remain. The uncertainty in the heat
of fo rmation ofC H is now significantly reduced due
Reactions ofthe phenyl radical at elevated temper-
atures are not well characterized. Under lean condi-
tions, C H is mainly consumed by reaction with O .
6
5
2
This reaction, a key step in aromatics oxidation and
PAH growth [41,42], was until recently believed to be
comparatively slow and to produce exclusively phen-
oxy radicals at higher temperatures [43]:
6
5
to the work of Heckmann et al. [36], but large differ-
ences in reported data are still observed for species
such as C H , C H O, and OC H O.
5
5
5
4
6
4
The elementary reactions in the oxidation scheme
that are most important for the present calculations are
discussed in detail next. The numbering ofreactions
corresponds to the listing in Appendix A. The discus-
sion will focus mainly on the C /C reaction subset,
C H ϩ O EF C H O ϩ O
(13)
6
5
2
6
5
However, recent shock tubes studies by Frank et al.
[20] and Kumaran and Michael [21] prove the C H ϩ
6
5
6
5
even though the chemistry ofthe smaller hydrocarbons
also has some significance.
O reaction to be comparatively fast and indicate an
important secondary channel producing H atoms:
2
Under the conditions ofthe present study, the main
consumption steps for benzene are the reactions with
H, O, and OH radicals:
C H ϩ O !: (product) ϩ H
(14a)
6
5
2
Both Frank et al. [20] and Kumaran and Michael [21]
C H ϩ H EF C H ϩ H
(Ϫ12)
(1)
derived a branching ratio, k /k14a, ofabout 3 under the
6
6
6
5
2
13
conditions oftheir investigation (1000–1500 K), but
no direct identification ofthe products ofreaction
(14a) could be made. Frank et al. proposed that the
reaction formed parabenzoquinone, that is,
C H ϩ O EF C H O ϩ H
6
6
6
5
C H ϩ O EF C H ϩ OH
(2)
6
6
6
5
C H ϩ OH EF C H ϩ H O
(3)
6
6
6
5
2
C H ϩ OH EF C H OH ϩ H
(Ϫ22)
C H ϩ O EF pOC H O ϩ H
(14b)
6
6
6
5
6
5
2
6
4
The C H ϩ H reaction (12) is well characterizedboth
This product channel was preferred to formation of the
6
5
2
experimentally and theoretically [37], and the uncer-
isomer orthobenzoquinone (oOC H O),
6
4
tainties for the reverse step, C H ϩ H, are mainly
6
6
related to the thermochemistry ofthe phenyl radical.
The product channels ofthe reaction ofC H with O
C H ϩ O EF oOC H O ϩ H
(14)
6
5
2
6
4
6
6
atoms are still uncertain. The main reaction channel
involves addition ofan O atom to fo rm a C H O ad-
since parabenzoquinone is thermally more stable. The
theoretical study on C H O isomers by Mebel and Lin
6
6
6
5
2
duct, which can follow either a rearrangement to pro-
duce phenol or suffer a ring hydrogen abstraction re-
sulting in phenoxy. No quantitative measurements of
the products have been reported. Earlier work [4,38]
proposed the addition plus rearrangement channel
leading to phenol as the preferred mechanism. Re-
cently, Bajaj and Fontijn [39] concluded from a flow
reactor study at 405 K that phenoxy is the main pri-
mary product ofthis step, although the of rmation of
phenol cannot be excluded. Following Bajaj and Fon-
tijn, we assume the phenoxy channel (1) to be domi-
nating. The direct hydrogen abstraction by O radicals,
reaction (2), is significantly slower than the addition
reaction (1). We have adopted the rate constant esti-
mated by Leidreiter and Wagner [40] from a shock
tube study in the temperature range 1200–1450 K.
[44] is not conclusive regarding the products ofthe
C H ϩ O reaction, but indicates that several mech-
6
5
2
anisms may be involved.
In previous modeling studies [6,8,9,10,13,16], re-
action (13) was the only product channel for the
C H ϩ O reaction. We have included the secondary
6
5
2
H-forming channel, basing our choice of products on
the fact that Chai and Pfefferle [19] detected only trace
amounts ofbenzoquinone in their recent jet-stirred re-
actor study oflean benzene oxidation at 1000–1200
K. If of rmed as a major product in the C H ϩ O2
6
5
reaction, parabenzoquinone would be expected to
build up in considerable concentrations; pOC H O is
6
4
comparatively stable at these temperatures [22]. For
this reason, we presently consider formation of or-
thobenzoquinone (14) to be the most likely secondary

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
501
product channel ofthe phenyl ϩ O reaction. How-
ever, further work is clearly required in order to de-
C H OH EF C H ϩ CO
(20)
(15)
2
6
5
5
6
C H O EF C H ϩ CO
6
5
5
5
termine the products ofthis important reaction.
At lower temperatures, below 1000 K, addition of
O to phenyl followed by stabilization of C H O ,
Shock tube experiments on phenol pyrolysis experi-
ments [48] indicate that (20) is the major dissociation
pathway, with a secondary channel forming phenoxy
and H atoms (Ϫ16). However, the importance ofre-
action (20) has been questioned [15] , based on results
from modeling low-pressure benzene flames. This is-
sue is discussed further later.
2
6
5
2
C H ϩ O EF C H O
2
6
5
2
6
5
is the dominant product channel [44,45,46]. However,
in the shock tube studies at 1000–1500 K [20,21] no
evidence ofthis reaction channel was fo und. A number
ofC H O isomers may potentially be important at
For thermal decomposition ofC H O (15), we have
6
5
6
5
2
adopted the measurement ofFrank et al. [20]. Recent
theoretical studies [49,50] indicate that the activation
energy for this reaction should be considerably higher
than the value ofapproximately 44 kcal/mol in fe rred
experimentally [20,43]; this issue needs to be ad-
dressed.
lower temperatures [44], but very little is known about
thermochemistry and reaction pathways for these com-
ponents. Based on the shock tube results, we have
omitted formation and reactions of phenyl peroxy in
the present mechanism, but further work is required to
establish the importance ofthis species.
Phenoxy may also lead to benzoquinone by reac-
tion with O atoms,
Other consumption steps for C H include thermal
6
5
dissociation, interaction with the radical pool, and re-
actions with aromatic and linear hydrocarbons. How-
ever, the impact ofthese reactions in the present cal-
culations (stoichiometric to lean conditions) is small;
phenyl is largely consumed by reaction with molecular
oxygen.
C H O ϩ O EF pOC H O ϩ H
(17)
(18)
6
5
6
4
C H O ϩ O EF oOC H O ϩ H
6
5
6
4
According to the work ofButh et al. [51], the
Oxygenated aromatics such as phenol and phenoxy
constitute important intermediates in benzene oxida-
tion, and pathways through these components may
lead to aromatic ring opening. A significant amount of
recycling between C H OH and C H O may occur.
C H O ϩ O reaction is very fast, producing benzo-
6
5
quinone and H atoms [51]. The branching fraction be-
tween the two potential product channels (17, 18) is
uncertain; as a first approximation we have assumed
that the two benzoquinone isomers are formed in equal
quantities. The formation of the two benzoquinone
isomers is supported by the ab initio study ofLin and
Mebel [52]; it is noteworthy, however, that Lin and
Mebel predict a much lower reaction rate than that
determined experimentally by Buth et al.
6
5
6
5
Phenol is converted to C H O by reaction with the
6
5
radical pool, mainly OH,
C H OH ϩ OH EF C H O ϩ H O
(24)
6
5
6
5
2
while phenoxy is recycled back to phenol mainly by
the reactions
The benzoquinone isomer, which has been detected
as intermediate in benzene oxidation [19], is presum-
ably parabenzoquinone; orthobenzoquinone is not
thermally stable, as we will discuss. In a previous flow
reactor study [22], we analyzed the chemistry ofthis
compound under pyrolysis and oxidation conditions.
In the temperature range ofthe present study, disso-
ciation is a key step in parabenzoquinone consump-
tion, both under pyrolysis and oxidation conditions.
The main dissociation pathway is [20,22],
C H O ϩ H EF C H OH
(16)
6
5
6
5
C H O ϩ HO EF C H OH ϩ O
2
(Ϫ25)
6
5
2
6
5
Overall this reaction sequence serves as a radical sink.
The most important chain-terminating step is the
C H O ϩ H reaction (16), which proceeds close to
6
5
collision frequency; the rate constant recommended by
Baulch et al. [26] has been used. The rate constant for
reaction (24) is also fairly well established [47], while
for reaction (Ϫ25) an estimated rate for the reverse
reaction was used, in the absence ofmeasured data.
The recycle reactions compete with a number of
steps that limit the size ofthe phenol/phenoxy pool.
At higher temperatures, phenol and phenoxy may dis-
sociate through CO elimination,
pOC H O EF C H O ϩ CO
(31)
6
4
5
4
but also a secondary channel may be active [22],
pOC H O EF C H ϩ CO
(32)
6
4
5
4
2
At high temperatures and lean conditions, reactions of

5
02
ALZUETA, GLARBORG, AND DAM-JOHANSEN
pOC H O with the radical pool become significant.
The main steps are proposed to be [22],
Emdee et al. [9]. Due to the comparatively weak bond-
ing ofthe hydrogen atoms, Emdee et al. assumed that
the consumption ofC H occurred largely through hy-
6
4
5
6
pOC H O ϩ H EF C H O ϩ CO
(33)
(34)
(35)
(36)
drogen abstraction reactions, mainly
6
4
5
5
pOC H O ϩ H EF OC H O ϩ H
2
6
4
6
3
C H ϩ H EF C H ϩ H
2
(49)
(53)
(55)
(57)
(58)
5
6
5
5
pOC H O ϩ O !: C H O ϩ H
6
4
6
3
3
C H ϩ O EF C H ϩ OH
5
6
5
5
pOC H O ϩ O EF OC H O ϩ OH
6
4
6
3
C H ϩ OH EF C H ϩ H O
5
6
5
5
2
pOC H O ϩ OH EF OC H O ϩ H O (37)
6
4
6
3
2
C H ϩ HO EF C H ϩ H O
5
6
2
5
5
2
2
Subsequently, C H O and OC H O are assumed to
C
5
H
6
ϩ O
2
EF C
5
H
5
ϩ HO
2
6
3
3
6
3
react by dissociation or reaction with the radical pool
38–42), leading to ring opening [22]. Details ofthe
reaction subset for parabenzoquinone are still uncer-
tain, but the overall pyrolysis and oxidation behavior
is fairly well described [22].
The rate coefficients were estimated by analogies to
(
CH O reactions [9]. The recent theoretical work of
2
Zhong and Bozzelli [17] indicates, however, that a
number ofaddition reactions compete with the me-
tathesis steps. Zhong and Bozzelli predict from QRRK
theory that a significant fraction of the cyclopentadi-
ene may react to form oxygenated compounds and
propose a number ofoxidation pathways not consid-
ered in previous modeling studies. These reactions,
which in general have not been characterized experi-
mentally, limit formation of the cyclopentadienyl rad-
ical and open up a number ofalternative ring-opening
oxidation pathways.
Orthobenzoquinone, if fo rmed in reactions (14) or
(18), would be expected to react differently from para-
benzoquinone. Orthobenzoquinone dissociates readily
even at comparatively low temperatures. De Jongh et
al. [53] reported thermolysis of oOC H O to be very
6
4
fast at 1073 K, with the major product being cyclo-
pentadienone (detected as the dimer),
oOC H O EF C H O ϩ CO
(30)
6
4
5
4
For most ofthe abstraction reactions (53, 55, 57,
5
8), rate constants still rely on estimates (Appendix
The rapid dissociation of oOC H O and the formation
6
4
A). Only for the C H ϩ H reaction (49) are experi-
ofC H O as the major product have been confirmed
5
6
5
4
mentally determined rate coefficients now available
56]. For the addition reactions discussed next, exper-
by subsequent studies [54]. The lower thermal stability
[
of oOC H O compared to pOC H O is in agreement
6
4
6
4
imental data are also very limited; we largely rely on
the QRRK estimates ofZhong and Bozzelli [17,18].
with the general observation that o-quinones are de-
carbonylated more readily than p-quinones [55]. In the
absence ofdirect measurements of r the rate constant
of(30), we employ an estimated value of r this exo-
thermic reaction. Our estimated activation energy cor-
responds roughly to the C9C bond strength between
the two carbonyl elements in oOC H O, which is
The C H ϩ H abstraction reaction (49) competes
5
6
with addition ofthe hydrogen atom to the double bond
in C H , resulting mainly in stabilization ofC H [17],
either the cyclic isomer,
C H ϩ H EF C H
7
5
6
5
7
6
4
(50)
5
6
5
about 40.6 kcal/mol [54]. With this rate constant,
oOC H O dissociates readily under the conditions in-
6
4
or, after ring opening, a linear isomer,
vestigated in the present study, and no other reactions
for this component were included in the mechanism.
Cyclopentadiene (C H ) and the derived radical,
˙
C H ϩ H EF CH CHCHCHCH
2
(51)
5
6
2
5
6
cyclopentadienyl (C H ), are major intermediates
5
5
Addition ofan oxygen atom to C H leads mainly to
5
6
formed in the high-temperature oxidation of benzene,
and their fate is critical for the oxidation process. Re-
cent studies ofthe pyrolysis [23,56,57,58] and oxi-
dation [56] ofcyclopentadiene have improved our
knowledge ofthe chemistry ofthese components, but
significant uncertainties remain, both in their reaction
mechanism and in the thermochemistry ofthese im-
portant intermediates.
formation of cyclic C H O [17]:
5
5
C H ϩ O EF C H O ϩ H
(52)
5
6
5
5
In the present work, no distinction is made between
the cyclic C H O isomers formed in this reaction. Ad-
5
5
dition ofOH to cyclopentadiene leads to a linear al-
cohol [17]:
The cyclopentadiene reaction subset used in pre-
vious models has generally relied on the estimates of
˙
C H ϩ OH EF CH CHCHCHCHOH (54)
5
6
2

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
503
while addition ofHO to C H mainly results in a cyc-
reaction pathway, is endothermic and significantly
slower [18].
2
5
6
lic C H radical and O :
5
7
2
As discussed by Zhong and Bozzelli [18], the re-
action ofC H with HO may be a key reaction in the
C H ϩ HO EF C H ϩ O
(56)
5
6
2
5
7
2
5
5
2
lean oxidation ofbenzene. Both ofthese radicals are
comparatively unreactive and may build up to rela-
tively large concentrations. The addition ofhydro-
peroxy radical to C H forms a complex, which either
dissociates, liberating OH radicals or a water mole-
cule, or recycles back to reactants. The dominating
product channel appears to be
The cyclopentadienyl radical (C H ), which is an
5
5
important intermediate in the transition from cyclic to
linear chain kinetics, is a resonance-stabilized radical.
It has been assumed to be consumed mainly through
radical/radical reactions with the O/H radical pool [4].
However, reaction with molecular oxygen [18] or
isomerization to its linear form [56,58] have also been
suggested to be important.
5
5
C H ϩ HO EF C H O ϩ OH
(71)
5
5
2
5
5
The reactions ofC H with H and O radicals have
5
5
been extensively studied in recent years [18,21,56,58,
9]. The C H ϩ H reaction,
This reaction is comparatively fast and will act as a
pseudochain branching step in the oxidation process.
The reaction ofthe cyclopentadienyl radical with
molecular oxygen has been suggested to be a major
consumption route for C H [2,56], but little is known
5
5
5
C H ϩ H EF C H
6
(66)
5
5
5
5
5
has been measured in both forward and reverse direc-
tions, and the rate constant is fairly well established.
Reported data [21,59] agree within a factor of 2 at a
about it. Roy and Frank [56] suggest that the reaction
mainly produces C H O:
5
4
1
4
temperature of1100 K, ranging r fo m 1.0·10
2
to
C H ϩ O EF C H O ϩ OH
1
4
3
5
5
2
5
4
.0·10 cm /mols.
Similarly, to C H ϩ H, the reaction ofC H with
5
5
5
5
and derive a set ofrate coe fif cients with a very high
preexponential factor from shock tube measurements.
QRRK calculations [18] indicate that the addition of
O to C H forms an energized peroxy radical, which
O radicals proceeds close to collision frequency.
Shock tube studies [20,21] indicate that the major
product channel involves formation of H atoms:
2
5
5
predominantly dissociates back to reactants. However,
a minor fraction of the adduct formed may undergo a
ring-opening reaction [18]:
C H ϩ O EF C H O ϩ H
(68)
5
5
5
4
Zhong and Bozelli [18] confirm reaction (68) as the
main product channel but identify also a second re-
action pathway, involving ring opening:
C H ϩ O EF CH CHCHCO ϩ HCO (72)
5
5
2
2
˙
C H ϩ O EF CH CHCHCH ϩ CO (67)
Despite a low reaction rate, about two orders ofmag-
nitude lower than the value derived by Roy and Frank,
this step may consume a significant fraction of the
cyclopentadienyl under lean conditions and thus acts
as an important path for the conversion of cyclopen-
tadienyl radical into linear species.
5
5
2
The association ofhydroxyl radicals to the cyclopen-
tadienyl radical,
C H ϩ OH EF (C H OH)
(69)
5
5
5
5
The fate of the cyclic compounds formed from
C H and C H (i.e., C H OH, C H O, C H OH, and
is proposed to result in an energized cyclopentadienol
radical [18]. This adduct can stabilize or isomerize to
cyclopentadiene-1-ol and cyclopentadiene-2-ol. Ac-
cording to the QRRK calculations [18], these three
exothermic product channels have comparable reac-
tion rates. In the mechanism, we do not distinguish
between these isomers. The channel leading to the cy-
clopentadienol and H,
5
5
5
6
5
5
5
5
5
4
C H O) is uncertain. There are no indications either in
5
4
the present experimental work or from results reported
in literature that any ofthese components build up to
significant concentrations during benzene oxidation,
not even in the 1000–1200 K range. Presumably,
C H O and C H OH dissociate fairly rapidly, while
5
5
5
4
C H OH and C H O are thermally more stable. In par-
5
5
5
4
ticular, reactions ofcyclopentadienone are important,
since this species appears to be an important link be-
tween the aromatic and nonaromatic hydrocarbons
during the benzene oxidation process. Emdee et al.
C H ϩ OH EF C H OH ϩ H
(70)
5
5
5
4
which in previous models was assumed to be the major

5
04
ALZUETA, GLARBORG, AND DAM-JOHANSEN
[
9] assumed C H O to open and lose the CO group
C H O !: C H ϩ CO
(86b)
5
4
5
4
4
4
according to
Since a large fraction of the oxidation route for ben-
zene appears to proceed through cyclopentadienone
under the conditions ofthe present study, such a global
consumption step would have significant implications
for the predicted oxidation behavior. We have chosen
not to include such an overall reaction; however,
clearly the development ofa reliable aromatics oxi-
dation mechanism requires a better characterization of
the reactions ofC H O.
C H O !: C H ϩ C H ϩ CO
(86)
5
4
2
2
2
2
In a more recent work, Wang and Brezinsky [23] pro-
posed, based on molecular orbital and RRKM theory,
that the thermal dissociation ofcyclopentadienone
leads mainly to cyclobutadiene:
C H O !: cC H ϩ CO
(86a)
5
4
4
4
5
4
The ring rupture ofthe cyclic compounds results in
with a rate constant in fairly good agreement with the
early estimate [9]. In the present work, we have
adopted the rate constant estimated by Wang and Bre-
zinsky but assume cyclobutadiene to dissociate readily
to acetylene. However, dissociation ofC H O has a
a number oflinear C 4 and C hydrocarbons, some of
5
which are oxygenated. Very few reactions of these
species have been characterized experimentally, and
this subset ofthe mechanism relies largely on QRRK
estimates [35] or estimates based on analogies with
C –C hydrocarbon reactions. Even though few of
5
4
large activation energy due to the endothermicity of
the reaction and is fairly slow under the conditions of
interest in the present study. Ifreaction (86) is used as
the main consumption step for cyclopentadienone, as
proposed by Emdee and co-workers, C H O builds up
1
3
these reactions show up in sensitivity analyses for ben-
zene oxidation under the present conditions, the un-
certainties in the C –C subset limit the reliability of
3
5
the model predictions and will have to be dealt with
in further work. Another potential limitation of the
mechanism is that formation of polyaromatic com-
5
4
to large concentrations in our modeling predictions.
This is in contrast to the present experimental data set
as well as the data of Chai and Pfefferle [19]. We have
added a subset for reactions of cyclopentadienone with
the radical pool (reactions 87–91), either hydrogen
abstraction (analogy with benzene steps) or ring open-
ing. However, even with very fast rates, these reac-
tions do not prevent C H O from accumulating under
pounds and hydrocarbons higher than C was not con-
6
sidered. Even under slightly lean conditions, such spe-
cies have been detected in minor concentrations [19],
and they may affect the benzene oxidation chemistry.
5
4
certain conditions in our calculations.
RESULTS AND DISCUSSION
Results from thermal decomposition of C H O pre-
5
4
cursors, oOC H O [53] and dihydro-benzodioxin [54],
The oxidation ofbenzene has been studied in the tem-
perature range 900–1400 K at atmospheric pressure
in a flow reactor. The experiments were conducted
keeping a constant mass flow rate and varying the re-
action temperature. This means that the residence time
in the reactor is a function of temperature, being about
150 ms at 1200 K. The benzene experiments were per-
formed as perturbation studies, varying the oxygen
concentration between 800 ppm and 50% (correspond-
ing to excess-air ratios ranging from about stoichio-
metric to 600); the concentration ofwater vapor (0.5
and 5%); and the concentration ofNO (0 and 100
ppm). In addition to the benzene experiments, a few
experiments were performed to study the pyrolysis and
oxidation ofphenol. Table I shows the experimental
matrix.
6
4
show that cyclopentadienone is rapidly consumed,
even at low temperatures (850–1100 K) and low rad-
ical concentrations. The fast consumption of C H O
5
4
under these conditions was attributed to a fast dimer-
ization reaction; while no or small amounts ofC H O
5
4
were detected, the dimer or decomposition/rearrange-
ment products from the dimer were identified [53,54].
Based on their results, Schraa et al. [54] derived a very
11
3
fast dimerization rate for C H O, 4.0·10 cm /mol-s
5
4
at 850 K, a value much higher than for other Diels-
Alder reactions. Whether or not this very high value
can be confirmed, the results indicate a high reactivity
ofcyclopentadienone.
In flames, cyclopentadienone is consumed rapidly
by thermal dissociation (86), but in the temperature
range ofthe flow reaction (i.e., 1000–1400 K) and jet-
stirred reactor studies, this step is too slow to account
for the consumption of C H O. Due to the lack ofde-
To complement the experimental data obtained in
the present study, additional flow reactor data from
Lovell et al. [6] and well-stirred reactor data from Chai
and Pfefferle [19] were analyzed. In these studies, a
smaller range ofreaction conditions was covered, but
contrary to the present study, concentrations ofimpor-
5
4
tailed kinetic information, Tan and Frank [14] de-
scribed C H O consumption in terms ofa global re-
5
4
action step:

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
505
Table I Experimental Conditions*
C H6
(ppm)
C H OH
(ppm)
O2
(ppm)
H O
(%)
NO
(ppm)
Residence time
(s)
6
6
5
2
Set
␭
1
2
3
4
5
6
7
8
9
50
28
trace
250000
870
830
1290
2.00
4.70
0.53
0.51
4.87
4.65
0.50
0.47
4.70
4.70
1.23
0
0
0
104
0
101
0
109
0
139
0
165[K]/T
115[K]/T
178[K]T
180[K]/T
185[K]/T
177[K]/T
181[K]/T
173[K]/T
187[K]/T
186[K]/T
184[K]/T
1275
1.04
1.04
1.60
1.60
49.6
49.4
48.3
47.4
600
111
106
107
102
105
101
108
109
110
1226
39100
37400
39100
38800
491000
1
1
0
1
*
The experiments are conducted at constant mass flow, and thereby the residence time is dependent on the reaction temperature, as listed.
The pressure in the experiments is 1.05 bar.
tant intermediate species formed during benzene oxi-
dation were measured. The significant range ofcon-
ditions covered by the experimental data allows an
assessment ofthe model capabilities, but a thorough
validation ofthe reaction mechanism would require a
more comprehensive characterization ofstable and in-
termediate species.
The model calculations were performed using Sen-
kin [60] for plug-flow calculations and PSR [61] for
well-stirred reactor modeling; both ofthese codes run
in conjunction with the Chemkin library [62].
Phenol Experiments
Phenol is an important intermediate during the com-
bustion ofaromatics [48,63], but it was not possible
to measure the concentration ofthis component in our
benzene oxidation experiments. As we will discuss,
the present reaction mechanism describes satisfacto-
rily the phenol profile in the benzene oxidation flow-
reactor experiments ofLovell et al. [6] but grossly
overpredicts the phenol concentration under the
stirred-reactor conditions ofthe experiments ofChai
and Pfefferle [19]. For this reason, we performed sep-
arate experiments to study the pyrolysis and the oxi-
dation ofphenol in the temperature range 900–1450
K. Thereby, the phenol chemistry subset ofthe model
could be evaluated independently ofthe benzene re-
actions.
Figure 1 Comparison between experimental data and
model predictions for CO and CO during pyrolysis ofphe-
2
Figure 1 compares experimental (symbols) and pre-
nol as function of temperature. Solid lines denote model pre-
dictions with the Appendix A mechanism; dashed lines
denote calculations without the reaction C H OH L
dicted (lines) concentrations ofCO and CO from py-
2
rolysis ofphenol. In addition to phenol and nitrogen,
the inlet gas contained about 2% ofwater vapor (Table
I, Set 1). Oxygen impurities may be present in the re-
6
5
C H ϩ CO (20). The inlet conditions are listed as Set 1 in
5
6
Table I.

5
06
ALZUETA, GLARBORG, AND DAM-JOHANSEN
actant mixture in trace amounts (10–50 ppm) due to
oxygen dissolved in the aqueous solution ofphenol.
However, from kinetic modeling calculations we con-
clude that the oxygen impurities are much lower than
tween the dissociation channels ofphenol, considering
the uncertainty in secondary reactions.
We have chosen to retain reaction (20) in the re-
action mechanism, since the shock tube measurements
ofHorn et al. [48] are considered to be more sensitive
to the phenol dissociation products than the low-pres-
sure benzene flames [15].
Figure 2 shows results for oxidation of phenol un-
der very lean conditions. The formation of CO is ini-
tiated at temperatures above 950 K and reaches its
maximum concentration at about 1100 K. At temper-
atures above 1200 K, all the phenol is oxidized to CO2.
The model predictions are in good agreement with the
experimental results. Under the conditions studied,
phenol is mainly consumed by hydrogen abstraction
reactions:
5
0 ppm.
Formation ofCO occurs at temperatures higher
than 1100 K and increases monotonically with tem-
perature. No maximum in CO is attained for the con-
ditions studied. At the highest temperatures, CO is
2
also formed. Also, a number of intermediate species
are formed; FTIR spectra indicate the presence of sig-
nificant amounts ofCH 4 and C H , as well as other
2
2
unidentified species.
Model predictions performed with the Appendix A
mechanism are shown as solid lines on Figure 1. The
onset and profile ofboth CO and CO 2 are well pre-
dicted by the model. The calculations indicate that the
formation of CO is due to the presence ofwater and
C H OH ϩ OH EF C H O ϩ H O
(24)
(23)
(26)
2
6
5
6
5
2
the subsequent H/O radical pool generated.
There is some controversy regarding the primary
product channel for phenol dissociation. The chain
initiation channel to phenoxy and H,
C H OH ϩ O EF C H O ϩ OH
6
5
6
5
C H OH ϩ HO EF C H O ϩ H O
2
6
5
2
6
5
2
C H OH EF C H O ϩ H
(Ϫ16)
6
5
6
5
is well established, through measurements in the re-
verse direction. A product channel forming cyclopen-
tadiene and carbon monoxide,
C H OH EF C H ϩ CO
(20)
6
5
5
6
was proposed by Cypr e` s and Bettens [64] from iso-
tope-marked experiments around 1000 K. This prod-
uct channel was later confirmed by Horn et al. [48]
based on H and CO concentration profiles in phenol
pyrolysis shock tube experiments. However, Shan-
dross et al. [15] found channel (20) to be inconsistent
with data from low-pressure benzene flames.
A first-order sensitivity analysis for CO in phenol
pyrolysis under the conditions ofFigure 1 is shown in
Table II. The analysis predicts a high sensitivity ofthe
predicted CO concentration towards the rate constant
ofreaction (20). However, this is an arti af ct ofthe
sensitivity analysis; as pointed out recently by Brezin-
sky et al. [65], flow-reactor results for CO (or C H )
5
6
cannot distinguish between the phenol dissociation
channels, since both channels ultimately lead to the
same set ofstable species. Reaction (20) corresponds
to the reaction sequence C H OH L C H O ϩ H
6
5
6
5
(
Ϫ16), C H O L C H ϩ CO (15), and C H ϩ H L
6 5 5 5 5 5
C H (66). This is confirmed by the results ofFigure
5
6
Figure 2 Comparison between experimental data and
1
, where the dashed lines denote modeling predictions
model predictions for CO and CO during oxidation ofphe-
2
excluding reaction (20). The difference between the
two sets ofcalculations is too small to distinguish be-
nol as function of temperature. The inlet conditions are listed
as Set 2 in Table I.

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
507
Table II Linear Sensitivity Coefficients for CO at Initiation Conditions for Selected Sets of Conditions*
Set 1
Set 2
Set 3
Set 5
Set 7
Set 9
Set 4 Set 10
Reaction
1150 K 1025 K 1050 K 1000 K 975 K 950 K 1000 K 900 K
1
2
3
.
.
.
C H ϩ O L C H O ϩ H
Ϫ0.08 Ϫ0.21
Ϫ0.21 Ϫ1.58 Ϫ0.16
0.03
0.64
6
6
6
5
C H ϩ O L C H ϩ OH
0.07
1.19
0.02
1.33
6
6
6
5
C H ϩ OH L C H ϩ H O
0.25
0.05
0.03
0.51
0.02
0.11
0.47
0.02
6
6
6
5
2
1
1
1
1
1
1
1
2
2
2
2
2
4
5
6
6
6
7
7
2. C H ϩ H L C H ϩ H
3. C H ϩ O L C H O ϩ O
6 5 2 6 6
0.27
0.47
0.11 Ϫ0.01
6
6
2
6
5
4. C H ϩ O L oOC H O ϩ H
Ϫ0.02 Ϫ0.07
0.19 0.37
0.17 Ϫ0.28
0.01 Ϫ0.05
Ϫ0.01
Ϫ0.27 Ϫ0.47 Ϫ0.07
1.37 1.04 0.28
Ϫ0.24 Ϫ0.19 Ϫ0.23 Ϫ0.51
Ϫ0.21 Ϫ0.23 Ϫ0.04 Ϫ0.20
0.02
0.74
6
5
2
6
4
5. C H O L C H ϩ CO
0.24
0.06
6
5
6
5
6. C H O ϩ H L C H OH
0.02
0.92
6
5
6
5
7. C H O ϩ O L pC H O ϩ H
6
5
6
4
8. C H O ϩ O L oC H O ϩ H
0.08 Ϫ0.12 Ϫ0.02
0.04
6
5
6
4
0. C H OH L C H ϩ CO
6
5
5
6
2. C H OH ϩ H L C H ϩ OH
Ϫ0.06 Ϫ0.15
0.20 Ϫ0.02 Ϫ0.07
Ϫ0.15 Ϫ0.37 Ϫ0.14
Ϫ0.04 Ϫ0.05 Ϫ0.06 Ϫ0.14
Ϫ0.13 Ϫ0.21
0.12
6
5
6
6
4. C H OH ϩ OH L C H O ϩ H O
6
5
6
5
2
5. C H OH ϩ O L C H O ϩ HO
2
0.78
0.20
Ϫ0.01
Ϫ0.03
6
5
2
6
5
6. C H OH ϩ HO L C H O ϩ H O
6
5
2
6
5
2
2
9. C H ϩ H L C H ϩ H
Ϫ0.07 Ϫ0.02
0.02 Ϫ0.01 Ϫ0.01
Ϫ0.56 Ϫ0.28
Ϫ0.01
5
6
6
5
2
8. C H ϩ O L C H ϩ HO
2
Ϫ0.01 Ϫ0.61
Ϫ0.01
Ϫ0.02
5
6
2
5
5
6. C H ϩ H L C H
6
Ϫ0.31 Ϫ0.01
Ϫ0.05 Ϫ0.03
Ϫ0.11 Ϫ0.01
5
5
5
8. C H ϩ O L C H O ϩ H
Ϫ0.02 Ϫ0.04
Ϫ0.05 Ϫ0.10
5
5
5
4
9. C H ϩ OH L C H OH
Ϫ0.01
5
5
5
5
1. C H ϩ HO L C H O ϩ OH
0.05
Ϫ0.04
0.10
0.52
0.08
0.36
0.22
0.15
5
5
2
5
5
2. C H ϩ O L CH CHCHCO ϩ HCO
0.01 Ϫ0.05
0.47
Ϫ0.02
0.03
0.36
0.15
0.09
5
5
2
2
1
1
43. CH CHCHCO ϩ H L CH CHC.H ϩ CO
46. CH CHCHCO ϩ O L CH HCO ϩ HCCO
Ϫ0.05 Ϫ0.02
2
2
2
0.03
0.84
0.04
0.14
0.04
0.03
0.41
0.15
0.36
0.13
0.09
4.39
0.04
2.15
0.08
0.62
2
2
O ϩ OH L H ϩ O2
0.10
10.59
OH ϩ H L H O ϩ H
2
2
2
OH L O ϩ H O
0.93
0.33
0.22
2
H ϩ O ϩ M L HO ϩ M
0.05 Ϫ10.26 Ϫ4.27
0.09 0.08
0.07 Ϫ0.07
2
2
H ϩ HO L OH ϩ OH
2
O ϩ HO L O ϩ OH
Ϫ0.01
Ϫ0.19 Ϫ0.01 Ϫ0.02
Ϫ0.06 Ϫ0.05
Ϫ0.20 Ϫ0.32
2
2
OH ϩ HO L H O ϩ O
2
Ϫ0.01
2
2
HO ϩ HO L H O ϩ O
2
0.26
0.26
0.17
0.95
0.28
2
2
2
2
H O ϩ M L OH ϩ OH ϩ M
0.08
0.06 Ϫ0.02 Ϫ0.02
Ϫ0.06 0.02 0.02
0.15
2
2
HCO ϩ M L H ϩ CO ϩ M
CH ϩ HO L CH ϩ H O
0.23 Ϫ0.03
Ϫ0.95 Ϫ0.23
4
2
3
2
2
NO ϩ O ϩ M L NO ϩ M
Ϫ0.03 Ϫ0.11
Ϫ0.07
2
NO ϩ OH ϩ M L HONO ϩ M
NO ϩ HO L NO ϩ OH
0.10
2
2
*
The sensitivity coefficients are given as A ␦Y /Y ␦A , where A is the preexponential constant for reaction i and Y is the mass fraction
i j j i i j
ofthe jth species. Therefore, the sensitivity coefficients listed can be interpreted as the relative change in predicted concentration for the
species j caused by increasing the rate constant for reactoin i by a factor of 2.
and, at higher temperatures, also by thermal dissocia-
tion:
marily by reaction with the radical pool, while
oOC H O decomposes thermally. Cyclopentadienyl is
6
4
consumed by reaction with the O/H radical pool or O2.
The calculations are most sensitive to the reactions
with HO and O (Table II), forming oxygenated cyc-
C H OH EF C H ϩ CO
(20)
6
5
5
6
2
2
Subsequently, C H O and C H may be converted to
lic and linear species:
6
5
5
6
C H by CO elimination (15) or H abstraction reac-
5
5
tions, respectively. The phenoxy radical may also react
with O atoms, forming benzoquinone (17, 18). Under
the conditions ofFigure 2, pOC H O is consumed pri-
C H ϩ HO EF C H O ϩ OH
(71)
5
5
2
5
5
C H ϩ O EF CH CHCHCO ϩ HCO (72)
6
4
5
5
2
2

5
08
ALZUETA, GLARBORG, AND DAM-JOHANSEN
A minor consumption pathway occurs through isom-
erization followed by ring opening:
˙
C H EF CHCCHCHCH
2
(65)
5
5
˙
CHCCHCHCH EF H CCCH ϩ C H
2
(112)
2
2
2
The cyclic oxygenated compounds are mostly con-
verted by H abstraction/H elimination reactions
(
C H OH, C H OH, and C H O) or by CO elimination
5 5 5 4 5 5
(
oOC H O) to C H O, which is the most stable of
6
4
5
4
these species. Cyclopentadienone, as well as paraben-
zoquinone, builds up at lower temperatures; but at
higher temperatures, it leads to ring opening, mainly
by reaction with the radical pool.
The good agreement between experimental data
and modeling predictions for pyrolysis and oxidation
ofphenol indicate that the overall fe atures ofthe phe-
nol chemistry are adequately described by the present
reaction mechanism. However, a thorough validation
ofthe phenol oxidation subset ofthe mechanism
would require comparison ofmodeling predictions to
a larger and more detailed experimental database. The
phenol chemistry is quite complex and the uncertain-
ties related to the reactions by cyclopentadienyl and
oxygenated intermediates are significant.
Benzene Oxidation
The oxidation ofbenzene under flow-reactor condi-
tions was studied over a large range ofstoichiometries
in the temperature range 900–1400 K. In addition to
varying the stoichiometry, the radical pool composi-
tion (at a fixed stoichiometry) was perturbed by chang-
ing the relative proportion between oxygen and water
vapor or by adding NO. The perturbation experiments
emphasize particular subsets ofthe oxidation chem-
istry and may provide important data for model de-
velopment and validation.
Figure 3 Comparison between experimental data and
model predictions for C H , CO, and CO during the oxi-
6
6
2
dation of benzene for different stoichiometries as a function
oftemperature. The inlet conditions are listed in Table I
(␭ ϭ 1.04, Set 3; ␭ ϭ 1.6, Set 5; ␭ ϭ 49.6, Set 7; ␭ ϭ 600,
Set 11).
Figure 3 compares experimental results and model
calculations for C H , CO, and CO as a function of
6
6
2
temperature obtained for stoichiometries ranging from
CO and CO among the two low excess-air sets, a
2
stoichiometric (␭ ϭ 1.04) to very lean conditions
significant shift in C H profiles is observed.
6
6
(
␭ ϭ 600). Independent ofthe stoichiometry, benzene
The model predicts well the experimental trends,
both the temperature regime for oxidation and the con-
centration profiles. However, for the lean conditions
and at lower temperatures, a carbon balance indicates
that the model underpredicts the conversion of
intermediate C and C compounds. In particular,
oxidation is initiated at about 1000 K. As the temper-
ature increases, the rate ofconsumption ofbenzene
depends strongly on the oxygen concentration. At high
O levels, benzene is almost completely consumed at
1
2
100 K; but at stoichiometric conditions, this temper-
4
5
ature is shifted almost 200 K towards higher temper-
atures. A similar shift in temperature is observed for
pOC H O and C H O appear to be overpredicted;
6 4 5 4
these compounds build up in concentration in the cal-
culations and cause underprediction ofCO and CO 2.
Furthermore, the C H oxidation rate for the ␭ ϭ 1.6
the profiles ofCO and CO . The two high excess-air-
2
ratio experiments provide very similar results. How-
ever, although no significant differences are seen for
6
6
set is somewhat underpredicted.

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
509
Figure 4 shows a reaction path diagram for the ox-
C H ϩ O EF C H O ϩ H
(1)
6
6
6
5
idation ofbenzene, based on rate-o f- production anal-
ysis ofthe calculations. The analysis indicates that the
major pathways for benzene oxidation (shown as solid
lines in the diagram) are similar for the variety of con-
ditions studied in this work. In addition to these re-
action paths, a number ofminor pathways are active
close to stoichiometric conditions; these are shown in
Figure 4 as dotted lines.
Phenyl reacts solely with O , forming mainly phen-
oxy:
2
C H ϩ O EF C H O ϩ O
(13)
6
5
2
6
5
but with a significant secondary channel, presumably
forming orthobenzoquinone:
Benzene oxidation is initiated by reaction with the
radical pool, forming phenyl or feeding into the phe-
nol/phenoxy pool:
C H ϩ O EF oOC H O ϩ H
(14)
6
5
2
6
4
Most ofthe phenol is converted to phenoxy through
reactions with the radical pool:
C H ϩ OH EF C H ϩ H O
(3)
6
6
6
5
2
C H ϩ OH EF C H OH ϩ H
(Ϫ22)
C H OH ϩ OH EF C H O ϩ H O
(24)
6
6
6
5
6
5
6
5
2
Figure 4 Reaction path diagram for C H oxidation, based on calculations with the Appendix A mechanism for the conditions
6
6
ofFigure 3. Solid lines denote the major oxidation pathways; dotted lines denote minor reaction pathways important only
under the close-to-stoichiometric conditions.

5
10
ALZUETA, GLARBORG, AND DAM-JOHANSEN
C H OH ϩ O EF C H O ϩ OH
(23)
(26)
tion reactions with H and OH radicals become impor-
tant:
6
5
6
5
C H OH ϩ HO EF C H O ϩ H O
2
6
5
2
6
5
2
C H ϩ H EF C H
6
(66)
(69)
5
5
5
Phenoxy radicals formed in significant quantities from
phenyl and phenol thereby become an important link
between the aromatic and nonaromatic hydrocarbons
during the benzene oxidation process. They are largely
converted to cyclopentadienyl radicals and benzoqui-
none:
C H ϩ OH EF C H OH
5
5
5
5
The reaction sequence, involving reaction (66) fol-
lowed by recycling ofcyclopentadiene to cyclopen-
tadienyl by reaction with the radical pool (49, 53, 55),
constitutes an important chain-terminating process.
The fate of C H OH and the derived radicals, C H O
C H O EF C H ϩ CO
(15)
6
5
5
5
5
5
5
5
and C H OH, is not well established. Following Zhong
5
4
C H O ϩ O EF OC H O ϩ H (17, 18)
6
5
6
4
and Bozzelli [35], we assume C H OH to be consumed
5
5
by abstraction reactions:
Due to the uncertainties ofthe reactions C H ϩ O
2
6
5
(
14) and C H O ϩ O (17, 18), the amount ofbenzo-
6 5
C H OH ϩ H, O, OH
quinone formed and the ratio between the two isomers
are quite uncertain. However, our calculations indicate
that the model predictions for the overall oxidation
behavior ofbenzene (i.e., the concentrations ofC H ,
CO, and CO ) are not very sensitive to these uncer-
tainties in the mechanism. pOC H O is consumed pri-
5
5
EF C H O/C H OHϩH , OH, H O
5
5
5
4
2
2
(
75–80)
6
6
The C H O and C H OH radicals are largely converted
5
5
5
4
2
to C H O by thermal decomposition or reaction with
5
4
6
4
O2:
marily by reaction with the radical pool, presumably
leading to ring opening; oOC H O decomposes by CO
6
4
C H O EF C H O ϩ H
(73)
(92)
(94)
5
5
5
4
elimination, forming C H O. Despite these differences
5
4
in oxidation pathways, model predictions varying the
products (parabenzoquinone or orthobenzoquinone) or
the rate (Appendix A value or zero) ofreaction (14)
provide similar results for C H , CO, and CO . This is
C H OH EF C H O ϩ H
5
4
5
4
C H OH ϩ O EF C H O ϩ HO
2
5
4
2
5
4
6
6
2
with a smaller fraction of C H O eliminating CO and
5
5
consistent with the fact that none of the benzoquinone
consumption reactions show up in the sensitivity anal-
ysis (Table II).
leading to ring opening:
˙
C H O EF CH CHCHCH ϩ CO
(74)
5
5
2
The cyclopentadienyl radical formed in reaction
(
15) is a key intermediate in the oxidation process, and
According to the mechanism, cyclopentadienone is
consumed mainly by reaction with the radical pool,
modeling predictions are very sensitive to the reac-
tions involving this radical (Table II). The chemistry
ofcyclopentadienyl is very complicated and has been
shown to involve numerous reaction pathways and
species [18]. It is worth noticing that the introduction
into the mechanism ofthe C H subset from Zhong
C H O ϩ OH EF C H O ϩ H O
(91)
5
4
5
3
2
˙
C H O ϩ H EF C CHCHCH ϩ CO (87)
5
4
2
5
5
C H O ϩ O EF C CHCCH ϩ CO
2
(89)
5
4
2
and Bozzelli [18] significantly improved model pre-
dictions for the present conditions.
while thermal dissociation (86) is less important under
these conditions. In the modeling, cyclopentadienone
is the most stable ofthese oxygenated cyclic com-
pounds. However, as discussed earlier, there are ex-
perimental indications that significant consumption
steps for this component are missing from the mech-
anism.
Under the conditions ofthis work, C H is con-
5
5
sumed mostly by reaction with HO , forming C H O:
2
5
5
C H ϩ HO EF C H O ϩ OH
(71)
5
5
2
5
5
or by ring-opening reactions,
C H ϩ O EF CH CHCHCH ϩ CO (67)
˙
The first-order sensitivity coefficients for CO listed
in Table II correspond to the initiation conditions of
selected benzene oxidation experiments. The results
indicate that the oxidation ofbenzene is mainly sen-
sitive to the C –C reaction subset. Apart from the
5
5
2
C H ϩ O EF CH CHCHCO ϩ HCO (72)
5
5
2
2
˙
C H EF CHCCHCHCH
(65)
5
5
2
6
5
Under close to stoichiometric conditions, recombina-
initiation reactions ofbenzene, the calculations are

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
511
sensitive to reactions involving C H , C H O, and
the radical pool mainly by its influence on the OH/O
ratio. Under lean conditions, the presence ofwater va-
por acts to convert one O radical into two hydroxyl
radicals through the fast shuffle reaction:
6
5
6
5
C H OH and to the fate of the cyclopentadienyl radi-
6
5
cal. However, the sensitivity analysis should be inter-
preted very cautiously due to the large uncertainties in
the reaction mechanism. As long as even the major
oxidation route for benzene oxidation is in question, a
sensitivity analysis may provide misleading results.
In order to further investigate the oxidation char-
acteristics ofbenzene, we perturbed the radical pool
O ϩ H O EF OH ϩ OH
2
This step shifts the OH/O ratio to higher values. The
effect is difficult to quantify, because it depends on
reactant composition, temperature, and oxidation rate;
but for the lean conditions, the OH/O ratio is generally
in the experiments by varying the amount ofH O and
2
by adding NO. Figure 5 compares experimental and
calculated results for water vapor levels of 0.5 and
increased by a factor of 2–5 by increasing the H O
2
4
.8%, respectively, for very lean conditions.
The water vapor level affects the composition of
level from 0.5 to 5%. In high concentrations, water
may also affect the composition of the radical pool by
acting as an efficient collision partner in the reaction:
H ϩ O ϩ M EF HO ϩ M
2
2
converting the highly reactive H atom into the com-
paratively unreactive HO radical.
2
The results ofFigure 5 show that even under very
lean conditions, where the effect would be expected
to be most pronounced, the impact ofthe water vapor
level is only minor and within the experimental un-
certainty. A small shift in the temperature for onset of
oxidation ofbenzene is observed, but the CO and CO 2
profiles are largely unchanged. The model correctly
predicts the small shift in the benzene concentration
profile but overestimates the effect of the increased
OH/O ratio at the higher H O level on the oxidation
2
ofCO to CO 2.
The influence ofNO addition on oxidation ofben-
zene is shown in Figure 6, which compares results
obtained in the absence and in the presence of100 ppm
NO as a function of temperature for stoichiometric and
very lean conditions. The presence ofNO changes the
composition ofthe radical pool through the reaction
NO ϩ HO EF NO ϩ OH
2
2
This step converts the unreactive HO radical into OH.
2
Thereby, the presence ofNO may significantly pro-
mote fuel oxidation under lean conditions in the 900–
1
200 K range [66,67]. Subsequently, NO is partially
2
recycled to NO by reaction with O or H atoms [66,67].
Under stoichiometric or fuel-rich conditions the con-
centration ofthe HO 2 radical is smaller and the pro-
moting effect of NO by this mechanism diminishes.
The presence ofNO may also promote oxidation by
direct interaction with the fuel oxidation chemistry
Figure 5 Comparison between experimental data and
model predictions for C H , CO, and CO profiles during the
6
6
2
[68,69]. Such an interaction is not believed to be sig-
oxidation of benzene for different stoichiometries and water
vapor levels as function of temperature. The inlet conditions
nificant under the present conditions, even though
some reaction between benzene-derived radicals and
NO is likely to occur.
are listed in Table I (␭ ϭ 49.6 and 0.5% H O: Set 7; ␭ ϭ
2
4
8.3 and 4.8% H O; Set 9).
2

5
12
ALZUETA, GLARBORG, AND DAM-JOHANSEN
Figure 6 shows that under very lean conditions the
presence ofNO shi tf s the oxidation regime ofC 6H6
model predicts correctly the shift in the temperature
for onset of oxidation caused by the NO under lean
conditions as well as the slight inhibition observed for
␭ ϭ 1.6. Compared to other fuels [66–69], the pro-
moting effect of NO under lean conditions is fairly
small for benzene. This indicates that other pseudo-
and the formation of CO and CO towards lower tem-
2
peratures (ϳ60–80 K). Between 900 and 1000 K, a
significant fraction, 10–30%, of the inlet NO is emit-
ted as NO . At excess-air ratios close to stoichiometric,
2
no sensitization ofthe C H oxidation by NO is seen
chain-branching steps involving HO compete with the
6
6
2
and very little NO is predicted. Under these condi-
NO ϩ HO reaction. Apart from the reaction with NO,
2
2
tions, presence ofNO causes a slight inhibition ofthe
oxidation, due to NO-catalyzed recombination ofrad-
icals [67].
HO is primarily removed through the following re-
actions:
2
The modeling predictions are in reasonably good
agreement with the experimental observations. The
C H O ϩ HO EF C H OH ϩ O
2
(Ϫ25)
(26)
6
5
2
6
5
C H OH ϩ HO EF C H O ϩ H O
2
6
5
2
6
5
2
C H ϩ HO EF C H O ϩ OH
(71)
5
5
2
5
5
C H ϩ HO EF C H ϩ O
2
(Ϫ58)
5
5
2
5
6
Ofthese reactions, only reaction (71), which produces
C H O and OH radicals, is chain propagating; the
5
5
other reactions are chain terminating. However, they
all contribute to remove HO and thereby minimize
2
the influence ofthe NO ϩ HO reaction.
2
In summary, the results ofFigures 5 and 6 indicate
that the oxidation rate ofbenzene under the conditions
investigated is mainly controlled by the stoichiometry
and the temperature. The radical pool composition,
specifically the O/OH ratio and the HO level, is com-
2
paratively less important.
The Flow Reactor Data of Lovell et al. The experi-
mental results ofLovell et al. [6] on lean benzene ox-
idation were used to further evaluate the benzene ox-
idation scheme. These data were obtained in a
turbulent flow reactor under near-adiabatic conditions.
The concentrations ofC H , C H OH, C H , C /C hy-
6
6
6
5
5
6
4
2
drocarbons, and CO were measured as function of ax-
ial distance in the reactor. Due to some uncertainty in
the velocity profile for the experiments [9], the reac-
tion time is only accurate within 15–20%.
The initial mixing ofreactants in the reactor is
known to affect the chemical induction time. For this
reason, modeling results are shifted in time to match
the measured benzene level at the point ofthe highest
concentration gradient. This approach has been shown
to be valid, since the induction chemistry has only
a negligible effect on the postinduction species pro-
files [9].
Figure 7 compares experimental results [6] and
model predictions for lean benzene oxidation at 1100
K. The overall agreement is reasonable, considering
the uncertainties in the mechanism just discussed. The
predicted benzene consumption rate is slightly lower
than observed; this discrepancy affects also the pre-
Figure 6 Comparison between experimental data and
model predictions for C H , CO, and CO during the oxi-
6
6
2
dation of benzene for different stoichiometries and NO lev-
els as function of temperature. The inlet conditions are listed
in Table I (␭ ϭ 1.04 and 0 ppm NO: Set 3; ␭ ϭ 48.3 and
0
4
ppm NO: Set 9; ␭ ϭ 1.6 and 100 ppm NO: Set 4; ␭ ϭ
7.4 and 100 ppm NO: Set 10).

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
513
detailed characterization ofthe product composition
was feasible.
Figure 8 compares experimental results reported by
Chai and Pfefferle [19] for lean benzene oxidation
with model predictions. Benzene oxidation is initiated
at ϳ1000 K, and benzene is largely depleted at 1100
K. The major intermediates are C H and CO, which
2
2
reach peak values of1000 and 10,000 ppm, respec-
tively. Apart from benzene, phenol is the most abun-
dant aromatic compound (around 40 ppm). From the
figure, it is evident that the model overpredicts the
benzene oxidation rate, shifting the temperature for
complete consumption ofC H almost 50 K to lower
6
6
temperatures. However, the modeling results must be
interpreted with some caution due to the occurrence of
multiple solutions. The solution shown is the one with
the greatest extent ofreaction.
The most remarkable discrepancies between exper-
imental data and modeling predictions concern phenol
and benzoquinone (shown in Fig. 8) as well as cyclo-
pentadienone and acetylene (not shown). The present
mechanism grossly overpredicts the concentration of
the cyclic intermediates C H OH, OC H O, and
6
5
6
4
C H O (note the logarithmic scale used for the bottom
part ofFig. 8). Correspondingly, the levels ofC 2H2
5
4
Figure 7 Comparison between experimental data [6] and
model predictions for oxidation of benzene in an adiabatic
flow reactor. Inlet composition is C H ϭ 1571 ppm, O ϭ
6
6
2
1
5,900 ppm; balance N . Inlet temperature is 1098 K, at-
2
mospheric pressure.
dictions for C H OH, C H , and CO. However, the
6
5
5
6
peak levels ofC H OH and C H are fairly well pre-
6
5
5
6
dicted. At longer reaction times, CO is significantly
underpredicted; this can be attributed mostly to
buildup ofcyclic intermediates such as pOC H O and
6
4
C H O in the calculations.
5
4
The Stirred Reactor Data of Chai and Pfef-
ferle. Even at comparable temperatures, experimental
conditions in a jet-stirred reactor are quite different
from those of a flow reactor. For this reason, we
wished to supplement the flow-reactor data from the
present study and those ofLovell et al. with the well-
mixed-reactor results reported by Chai and Pfefferle
[19]. Chai and Pfefferle performed their experiments
in a micro-jet reactor, which approaches a well-stirred
reactor under the conditions investigated. Their ex-
periments, which covered the temperature range 900–
Figure 8 Comparison between experimental data [19] and
model predictions for oxidation of benzene in a jet-stirred
reactor. Inlet composition is C H ϭ 5100 ppm, O ϭ 2.0%
(vol); balance Ar. Pressure is 0.46 atm; nominal residence
time is 50 ms.
1
300 K, were performed at nominal residence times
of ϳ50 ms and pressures of350 torr. The products of
reaction were measured by a residual gas analyzer and
time-of-flight mass spectrometers. Thereby, a very
6
6
2

5
14
ALZUETA, GLARBORG, AND DAM-JOHANSEN
Appendix A The C –C Reaction Subset in the Benzene Oxidation Mechanism
3
6
A
␤
Ea
Source
Cyclic C species
6
1
2
3
4
5
6
7
8
9
.
.
.
.
.
.
.
.
.
C H ϩ O L C H O ϩ H
2.2E13
2.0E13
1.6E08
6.0E13
1.4E15
2.0E14
2.0E07
1.0E14
5.0E13
1.0E07
3.0E13
5.7E04
2.6E13
3.0E13
7.4E11
2.5E14
8.5E13
8.5E13
1.0E13
1.0E12
1.2E14
2.2E13
1.3E13
3.0E06
1.0E13
1.0E12
6.0E12
6.0E12
4.9E12
1.0E12
3.7E11
3.5E12
2.5E13
2.0E12
1.5E13
1.4E13
1.0E06
1.0E14
1.0E14
1.0E14
1.0E12
1.0E14
0.00
0.00
1.42
0.00
0.00
0.00
2.00
0.00
0.00
2.00
0.00
2.43
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.00
0.00
0.00
0.00
0.00
0.00
4530
[26, 39]
[40]
[25]
est
[21], est
[21, 70]
[27] est
[20]
[27] est
[27] est
est
[37]
[20]
[20], see text
[20]
[25]
[51], est
[51], est
[51]
[48]
[25]
6
6
6
5
C H ϩ O L C H O ϩ OH
14,700
1450
63,400
83,860
0
6
6
6
5
C H ϩ OH L C H ϩ H O
6
6
6
5
2
C H ϩ O L C H ϩ HO
2
6
6
2
6
5
C H L C H ϩ H
6
5
6
4
C H ϩ H L C H
6
6
5
6
C H ϩ H L C H ϩ H
2
1000
0
6
5
6
4
C H ϩ O L C H ϩ CO
6
5
5
5
C H ϩ OH L C H O ϩ H
0
6
5
6
6
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
0.
1.
2.
3.
4.
5.
6.
7.
8.
9.
0.
1.
2.
3.
4.
5.
6.
7.
8.
9.
0.
1.
2.
3.
4.
5.
6.
7.
8.
9.
0.
1.
2.
C H ϩ OH L C H ϩ H O
1000
0
6
5
6
4
2
C H ϩ HO L C H O ϩ OH
6
5
2
6
5
C H ϩ H L C H ϩ H
6280
6100
9000
43,900
0
6
5
2
6
6
C H ϩ O L C H O ϩ O
6
5
2
6
5
C H ϩ O L oOC H O ϩ H
6
5
2
6
4
C H O L C H ϩ CO
6
5
5
5
C H O ϩ H L C H OH
6
5
6
5
C H O ϩ O L pOC H O ϩ H
0
0
0
6
5
6
4
C H O ϩ O L oOC H O ϩ H
6
5
6
4
C H O ϩ O L C H ϩ CO
2
6
5
5
5
C H OH L C H O ϩ CO
60,800
12,400
7930
2900
Ϫ1310
38,000
1000
0
6
5
5
6
C H OH ϩ H L C H O ϩ H
2
6
5
6
5
C H OH ϩ H L C H ϩ OH
[25]
[26]
[47]
est
6
5
6
6
C H OH ϩ O L C H O ϩ OH
6
5
6
5
C H OH ϩ OH L C H O ϩ H O
6
5
6
5
2
C H OH ϩ O L C H O ϩ HO
2
6
5
2
6
5
C H OH ϩ HO L C H O ϩ H O
2
est
6
5
2
6
5
2
C H OH ϩ C H L C H O ϩ C H
4
[9] est
[9] est
[71]
est
[22]
[22]
[22] est
[22] estb
6
5
2
3
6
5
2
C H OH ϩ CH CHCB CH L C H O ϩ CH CHCHCH
0
6
5
2
2
6
5
2
2
C H OH ϩ C H L C H O ϩ C H
6
4400
40,000
59,000
67,000
4700
8100
4530
14,700
4000
0
6
5
6
5
6
5
6
oOC H O L C H O ϩ CO
6
4
5
4
pOC H O L C H O ϩ CO
6
4
5
4
pOC H O L C H ϩ CO
6
4
5
4
2
a
pOC H O ϩ H L C H O ϩ CO
6
4
5
5
pOC H O ϩ H : OC H O ϩ H
2
6
4
6
3
b
pOC H O ϩ O : C H O ϩ H
[22] est
6
4
6
3
3
pOC H O ϩ O L OC H O ϩ OH
[22] estb
[22]
6
4
6
3
pOC H O ϩ OH L OC H O ϩ H O
6
4
6
3
2
OC H O ϩ H : pOC H O ϩ OH
est
6
3
6
4
OC H O ϩ H : 2C H ϩ 2CO
0
0
[22] est
[22] est
[22] est
est
6
3
2
2
OC H O ϩ O : C H ϩ HCCO ϩ 2CO
6
3
2
2
C H O : C H ϩ HCCO ϩ 2CO
50,000
0
6
3
3
2
2
C H O ϩ H : C H ϩ CH CO ϩ 2CO
6
3
3
2
2
2
Cyclic C species
5
a
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
C H7 L CH CHCB HCHCH
3.2E15
3.6E12
1.0E13
2.4E13
1.3E11
8.0E24
7.2E13
2.4E73
1.1E14
8.9E12
5.6E12
0.00
0.00
0.00
0.00
0.00
39,500
0
[72]
estc
est
[35] est
estc
[35] qrrk
[56]
[18] qrrk
[18] qrrk
[18] qrrk
[18] qrrk
5
2
2
C H ϩ H L C H ϩ H
2
5
7
5
6
C H ϩ O L C H ϩ OH
0
0
0
5
7
5
6
c
C H ϩ OH L C H ϩ H O
5
7
5
6
2
C H ϩ O L C H ϩ HO
2
5
7
2
5
6
C H ϩ O L CHOCH CH CB HCHCO
Ϫ3.80
0.00
Ϫ17.85
Ϫ0.16
Ϫ0.15
Ϫ.06
20,000
3500
31,500
3100
590
200
5
7
2
2
2
C H ϩ H L C H ϩ H
5
6
5
5
2
C H ϩ H L C H
5
6
5
7
C H ϩ H L CH CHCB HCHCH
5
6
2
2
C H ϩ O L C H O ϩ H
5
6
5
5
C H ϩ O L C H O ϩ H
5
6
5
5
(Continued)

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
515
Appendix A (Continued)
A
␤
Ea
Source
Duplicate reactiona
53.
54.
55.
56.
57.
58.
59.
C H ϩ O L C H ϩ OH
4.8E04
1.1E13
3.1E06
1.3E15
1.1E04
4.0E13
1.8E-1
2.71
Ϫ0.07
2.00
Ϫ1.07
2.60
1100
870
0
9530
12,900
37,150
0
[18] estc
[18] qrrk
[18] estd
[18] qrrk
[18] estc
5
6
5
5
C H ϩ OH L CH CHCB HCHCHOH
5
6
2
C H ϩ OH L C H ϩ H O
5
6
5
5
2
C H ϩ HO L C H ϩ O
5
6
2
5
7
2
2
2
C H ϩ HO L C H ϩ O
5
6
2
5
5
c
C H ϩ O L C H ϩ HO
0.00
4.00
[18] est
[18] estc
5
6
2
5
5
C H ϩ CH L C H ϩ CH
4
5
6
3
5
5
6
6
0.
1.
C H ϩ C H L C H ϩ CH
3
2.1E67
1.2E-1
Ϫ16.08
42,460
0
[18] estc
[18] estc
5
6
2
3
6
6
C H ϩ C H L C H ϩ C H
4.00
5
6
2
3
5
5
2
4
6
2.
C H ϩ C H L C H ϩ C H
1.0E-1
4.00
0
[18] estc
5
6
6
5
6
6
5
5
6
6
6
6
6
6
6
3.
4.
5.
6.
7.
8.
9.
C H ϩ CH CHCB CH L C H ϩ CH CHCHCH
6.0E12
3.2E11
1.0E12
1.5E14
3.2E13
5.8E13
6.5E14
1.1E43
1.1E59
0.00
0.00
1.00
0
8000
77,000
0
440
20
Ϫ2730
18,730
33,450
[9] est
[9] est
[56]
[21, 59]
[18] qrrk
[18] qrrk
[18] qrrk
[18] qrrk
[18] qrrk
5
6
2
2
5
5
2
2
C H ϩ C H O L C H ϩ C H OH
5
6
6
5
5
5
6
5
C H L CHCCHCHCB H
5
5
2
C H ϩ H L C H
6
0.00
5
5
5
C H ϩ O L CH CHCHCB H ϩ CO
Ϫ0.17
Ϫ0.02
Ϫ0.85
Ϫ8.76
Ϫ13.08
5
5
2
C H ϩ O L C H O ϩ H
5
5
5
4
C H ϩ OH L C H OH
5
5
5
5
C H ϩ OH L C H OH
5
5
5
5
C H ϩ OH L C H OH
5
5
5
5
Duplicate reactiona
C H ϩ OH L C H OH ϩ H
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
3.5E57
6.3E29
1.2E19
2.9E32
1.1E79
3.2E12
4.0E13
4.7E11
1.0E13
5.5E12
1.0E13
3.6E03
1.0E13
1.0E06
1.0E06
1.0E06
5.7E32
6.2E41
Ϫ12.18
Ϫ4.69
Ϫ2.48
Ϫ6.50
Ϫ19.62
0.00
0.00
0.00
0.00
0.00
0.00
2.55
0.00
2.50
2.50
2.00
Ϫ6.76
Ϫ7.87
48,350
11,650
10,970
21,220
66,250
0
[18] qrrk
[18] qrrk
[18] qrrk
[35] qrrk
[35] qrrk
5
5
5
4
C H ϩ HO L C H O ϩ OH
5
5
2
5
5
C H ϩ O L CH CHCHCO ϩ HCO
5
5
2
2
C H O L C H O ϩ H
5
5
5
4
C H O L CH CHCHCB H ϩ CO
5
5
2
c
C H OH ϩ H L C H OH ϩ H
2
[35] est
5
5
5
4
C H OH ϩ H L C H O ϩ H
2
6094
0
[35] estd
[35] estd
[35] estd
5
5
5
5
C H OH ϩ O L C H OH ϩ OH
5
5
5
4
C H OH ϩ O L C H O ϩ OH
4683
1731
1697
10,531
15,800
5000
3000
0
5
5
5
5
d
C H OH ϩ OH L C H OH ϩ H O
[35] est
5
5
5
4
2
C H OH ϩ OH L C H O ϩ H O
[35] estd
[35] est
[35] est
[22] est
[22] est
[22] est
[23]e
5
5
5
5
2
C H OH ϩ HO L C H OH ϩ H O
2
5
5
2
5
4
2
C H OH ϩ HO L C H O ϩ H O
5
5
2
5
5
2
2
2
2
C H ϩ H L C H ϩ H
5
4
5
3
C H ϩ O L C H ϩ H
5
4
5
3
C H ϩ OH L C H ϩ H O
5
4
5
3
2
C H O : C H ϩ C H ϩ CO
68,500
98,700
5
4
2
2
2
2
C H O : C H ϩ C H ϩ CO
5
4
2
2
2
2
Duplicate reaction
C H O ϩ H L CH CHCHCB H ϩ CO
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
2.1E61
2.0E12
1.0E13
1.4E13
1.1E08
2.1E13
1.0E14
3.0E13
1.0E14
9.7E58
1.0E14
1.0E12
Ϫ13.27
0.00
40,810
8100
2000
14,700
1450
54,000
0
5000
0
38,180
0
[35] qrrk
est
5
4
2
b
C H O ϩ H L C H O ϩ H
2
5
4
5
3
C H O ϩ O L CH CHCCH ϩ CO
2
0.00
0.00
1.42
0.00
0.00
0.00
0.00
[22] est
5
4
2
b
C H O ϩ O L C H O ϩ OH
est
5
4
5
3
b
C H O ϩ OH L C H O ϩ H O
est
5
4
5
3
2
f
C H OH ϩ H L C H O ϩ H
[9], adj
5
4
5
4
C H OH ϩ H L C H OH
est
est
est
[35] qrrk
est
5
4
5
5
C H OH ϩ O L C H O ϩ HO
2
5
4
2
5
4
C H O ϩ H L C H O
5
3
5
4
C H O ϩ O L CB HCHCHCO ϩ CO
Ϫ13.47
0.00
5
3
2
2
C H ϩ H L C H
4
5
3
5
C H ϩ O L C H ϩ HCCO ϩ CO
0.00
0
[22] est
5
3
2
2
2
(Continued)

5
16
ALZUETA, GLARBORG, AND DAM-JOHANSEN
Appendix A (Continued)
A
␤
Ea
Source
Linear C species
5
9
9.
CH CHCHCHCH ϩ H L CH CHCHCH ϩ CH
3
5.2E71
7.0E06
7.0E06
5.4E11
1.2E36
2.3E20
2.9E26
2.0E14
1.5E13
8.2E10
2.5E34
6.3E05
1.2E36
1.0E12
Ϫ16.38
2.00
51,000
5000
0
[35] qrrk
2
3
2
2
c
1
1
1
1
1
1
1
1
1
1
1
1
1
00.
01.
02.
03.
04.
05.
06.
07.
08.
09.
10.
11.
12.
CH CHCHCHCH ϩ H L CH CHCHCHCB H ϩ H
[28] est
[28] estc
[35]
2
3
2
2
2
CH CHCHCHCH ϩ OH L CH CHCHCHCB H ϩ H O
2.00
2
3
2
2
2
CH CHCb HCHCH L CH CHCHCHCB H
Ϫ0.70
Ϫ7.25
Ϫ1.60
Ϫ2.18
0.00
0.00
0.18
Ϫ6.12
1.65
Ϫ7.25
1.00
60
2
2
2
2
CH CHCB HCHCH ϩ O L CH CHCHO ϩ CH HCO
33,600
3020
36,770
0
[35] qrrk
[35] qrrk
[35] qrrk
[28] estc
[35] estc
[35] qrrk
[35] qrrk
[35] est
[35] est
[56]
2
2
2
2
2
CH CHCHCHCb H ϩ H L CH CHCHCHCH
2
2
2
3
CH CHCHCHCB H ϩ H L CH CHCHCb H ϩ CH
2
2
2
3
CH CHCHCHCB H ϩ O L CH CHCHO ϩ C H
2
2
2
2
3
CH CHCHCHCb H ϩ OH L CH CHCHCHCH OH
0
2
2
2
2
CH CHCHCHCB H ϩ O L CH CHCB HCHCHO ϩ CH O
9140
16,250
17,480
33,600
31,000
2
2
2
2
2
CH CHCHCHCH OH ϩ H L CH CHCHCH ϩ CH OH
2
2
2
2
2
CHOCH CH CB HCHO ϩ O L CHOCH CH CHO ϩ HCOO
2
2
2
2
2
CH CHCB HCHCHOH ϩ O L CHOCHCHOH ϩ CH HCO
2
2
2
CHCCHCHCB H L H CCCH ϩ C H
2
2
2
2
C hydrocarbon oxidation subset
4
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
CH CHCHCH ϩ H L CH CHCHCB H ϩ H
3.0E07
3.0E07
6.0E08
1.0E12
2.0E07
2.0E07
3.0E13
3.0E13
2.3E10
3.5E09
1.0E14
1.0E14
2.00
2.00
1.45
0.00
2.00
2.00
0.00
0.00
1.05
1.18
0.00
0.00
13,000
6000
Ϫ858
0
5000
2000
57,800
45,850
3279
Ϫ447
37,000
30,000
[27]
[27]
[28]
[28]
[27]
[27]
est
2
2
2
2
CH CHCHCH ϩ H L CH CHCb CH ϩ H
2
2
2
2
2
CH CHCHCH ϩ O L CH CHCB H ϩ HCO
2
2
2
2
CH CHCHCH ϩ CH HCO ϩ C H
3
2
2
2
2
CH CHCHCH ϩ OH L CH CHCHCB H ϩ H O
2
2
2
2
CH CHCHCH ϩ OH L CH CHCB CH ϩ H O
2
2
2
2
2
CH CHCHCH ϩ O L CH CHCHCB H ϩ HO
2
2
2
2
2
CH CHCHCH ϩ O L CH CHCB CH ϩ HO
est
2
2
2
2
2
2
g
CHOCH CH CHO ϩ H L CHOCH CH CB O ϩ H
[35] est
[35] estg
[27]
2
2
2
2
2
CHOCH CH CHO ϩ OH L CHOCH CH Cb O ϩ H O
2
2
2
2
2
CH CHCHCB H(ϩM) ϭ CH CHCCH ϩ H(ϩM)
2
2
Low pressure limit:
Enhanced third-body efficiencies:
H O ϭ 5
2
124.
125.
126.
127.
128.
129.
CH CHCHCb H ϩ H L CH CHCB CH ϩ H
1.0E14
3.0E07
2.0E07
1.0E12
1.0E07
1.0E14
2.0E15
0.00
2.00
2.00
0.00
2.00
0.00
0.00
0
1000
1000
[27]
[27]
[27]
[28]
[28]
[27]
2
2
2
CH CHCHCb H ϩ H L CH CHCCH ϩ H
2
2
2
CH CHCHCb H ϩ OH L CH CHCCH ϩ H O
2
2
2
CH CHCHCb H ϩ O L CB HCHCHO ϩ CH O
0
2
2
2
CH CHCHCB H ϩ O : CH CHCCH ϩ HO
10,000
50,000
42,000
2
2
2
2
CH CHCB CH (ϩM) ϭ CH CHCCH ϩ H(ϩM)
2
2
2
Low pressure limit:
Enhanced third-body efficiencies:
H O ϭ 5
2
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
CH CHCb CH ϩ H L H CCCB H ϩ CH
1.0E14
3.0E13
6.1E05
1.2E36
1.6E45
2.0E07
3.0E07
1.4E07
8.1E06
7.5E06
1.0E07
3.0E13
3.0E13
6.6E13
1.5E07
0.00
0.00
0.92
Ϫ7.25
Ϫ9.92
2.00
2.00
2.00
1.88
2.00
2.00
0.00
0.00
Ϫ0.02
2.00
0
0
[27]
[27]
est
est k98
[35] qrrk
[27]
2
2
2
3
CH CHCb CH ϩ OH L CH CHCCH ϩ H O
2
2
2
2
h
CH CHCb HCHO : CH CHCB H ϩ CO
Ϫ1120
33,600
20,670
15,000
5000
1900
180
5000
2000
2
2
2
CH CHCB HCHO ϩ O : CH CHCHO ϩ HCOO
2
2
2
CHOCH CH CB O ϩ O L C H ϩ O CCHOOB
2
2
2
2
4
2
CH CHCCH ϩ H L HCB CHCCH ϩ H
2
2
CH CHCCH ϩ H L H CCB CCH ϩ H
[27]
2
2
2
i
CH CHCCH ϩ O L HCCO ϩ C H
3
est
2
2
i
CH CHCCH ϩ O L H CCCB H ϩ HCO
est
2
2
CH CHCCH ϩ OH L HCB CHCCH ϩ H O
[27]
[27]
est
est
[35] qrrk
est 0.5k109
2
2
CH CHCCH ϩ OH L H CCB CCH ϩ H O
2
2
2
CH CHCCH ϩ O L HCB CHCCH ϩ HO
60,750
42,200
2740
2
2
2
CH CHCCH ϩ O L H CCB CCH ϩ HO
2
2
2
2
CH CHCHCO ϩ H L CH CHCB H ϩ CO
2
2
2
CH CHCHCO ϩ H L CH CB CHCO ϩ H
6000
2
2
2
(Continued)

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
517
Appendix A (Continued)
A
␤
Ea
1300
Ϫ860
5000
2000
0
Source
esti
est 0.5k110
est 0.5k112
1
1
1
1
1
1
45.
46.
47.
48.
49.
50.
CH CHCHCO ϩ H L CH CHCB H ϩ CO
5.9E06
3.0E08
1.0E07
1.0E07
3.0E12
1.0E14
1.0E14
2.00
1.45
2.00
2.00
0.00
0.00
0.00
2
2
2
CH CHCHCO ϩ O L CH HCO ϩ HCCO
2
2
CH CHCHCO ϩ OH L CB HCHCHCO ϩ H O
2
2
CH CHCHCO ϩ OH L CH CB CHCO ϩ H O
est 0.5k
est
2
2
2
113
i
CH CHCHCO ϩ OH L CH CHCB H ϩ CO
2
2
2
2
HCB CHCCH(ϩM) L C H ϩ H(ϩM)
36,000
30,000
[27]
4
2
Low pressure limit:
Enhanced third-body efficiencies:
H O ϭ 5
2
151.
152.
153.
154.
155.
HCb CHCCH ϩ H L H CCB CCH ϩ H
1.0E14
1.0E13
3.0E12
2.8E03
1.0E14
2.0E15
0.00
0.00
0.00
2.90
0.00
0.00
0
0
0
[27]
est
est, glob
[27]
[27]
2
HCb CHCCH ϩ OH L C H ϩ H O
4
2
2
HCb CHCCH ϩ O : C H ϩ CO ϩ HCO
2
2
2
HCB CHCCH ϩ C H L C H
1400
55,000
48,000
2
2
6
5
H CCB CCH(ϩM) L C H ϩ H(ϩM)
2
4
2
Low pressure limit:
Enhanced third-body efficiencies:
H O ϭ 5
2
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
179.
180.
181.
H CCb CCH ϩ H L C H ϩ H
5.0E13
2.0E13
2.0E13
3.0E13
1.0E12
2.0E13
1.0E14
3.0E13
6.1E05
1.0E14
3.0E07
2.0E07
1.0E07
2.0E07
1.2E12
6.7E12
1.0E07
5.0E13
1.0E07
3.0E07
1.0E07
5.0E13
1.0E13
1.0E13
1.0E12
1.0E13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.92
0.00
2.00
2.00
2.00
2.00
0.00
0.00
2.00
0.00
2.00
2.00
2.00
0.00
0.00
0.00
0.00
0.00
0
0
0
0
0
0
0
0
[27]
[27]
[27]
[27]
[27]
[27]
est k125
2
4
2
2
H CCb CCH ϩ O L CH CO ϩ C H
2
2
2
H CCb CCH ϩ O L H C O ϩ H
2
2
4
H CCb CCH ϩ OH L C H ϩ H O
2
4
2
2
H CCb CCH ϩ O L CH CO ϩ HCCO
2
2
2
H CCb CCH ϩ CH L C H ϩ C H
2
2
3
4
2
CH Cb CHCO ϩ H : H CCCB H ϩ HCO
2
2
CH CB CHCO ϩ OH : CHCCHCO ϩ H O
est k
2
2
126
j
Cb HCHCHCO L H CCCB H ϩ CO
Ϫ1120
0
est
2
Cb HCHCHCO ϩ H L CH CB CHCO ϩ H
est k119
est k120
est k121
est k122
[27]
[27]
[27]
[27]
[27]
2
Cb HCHCHCO ϩ H L CHCCHCO ϩ H
1000
1000
10,000
2000
0
Ϫ410
1000
3000
2000
5000
2000
0
2
Cb HCHCHCO ϩ OH L CHCCHCO ϩ H O
2
CB HCHCHCO ϩ O L CHCCHCO ϩ HO
2
2
C H ϩ H L C H ϩ H
2
4
2
4
C H ϩ O L C H ϩ CO
4
2
3
2
C H ϩ OH L H C O ϩ H
4
2
2
4
C H ϩ OH L C H ϩ H O
4
2
4
2
H C O ϩ H L C H ϩ HCCO
2
4
2
2
H C O ϩ OH L CH CO ϩ HCCO
[27]
2
4
2
CHCCHCO ϩ H : CHCCB CO ϩ H2
est k131
est k135
est
est
est
CHCCHCO ϩ OH : CHCCB CO ϩ H O
2
CHCCb CO ϩ H : C H ϩ CO
3
2
CHCCb CO ϩ O : C H ϩ CO ϩ CO
0
0
0
0
2
CHCCb CO ϩ OH : C H ϩ CO ϩ CO
2
2
CHCCB CO ϩ O : HCCO ϩ CO ϩ CO
est
[27]
2
C H ϩ O L CO ϩ CO ϩ C H
4
2
2
C hydrocarbon oxidation subset
3
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
C H ϩ H L C H ϩ CH
3
7.2E12
1.7E05
4.1E05
8.0E05
1.6E07
5.2E11
1.2E11
6.0E10
3.1E06
1.1E06
0.00
2.50
2.50
2.50
1.76
0.70
0.70
0.70
2.00
2.00
1300
2490
9800
12,300
Ϫ1216
5884
8960
7630
Ϫ298
1451
[73]
[73]
[73]
[73]
[73]
[73]
[73]
[73]
[73]
[73]
3
6
2
4
C H ϩ H L CH CHCB H ϩ H
3
6
2
2
2
C H ϩ H L CH CB CH ϩ H
3
6
2
3
2
C H ϩ H L CB HCHCH ϩ H
3
6
3
2
C H ϩ O L C H ϩ HCO
3
6
2
5
C H ϩ O L CH CHCB H ϩ OH
3
6
2
2
C H ϩ O L CB HCHCH ϩ OH
3
6
3
C H ϩ O L CH CB CH ϩ OH
3
6
2
3
C H ϩ OH L CH CHCB H ϩ H O
3
6
2
2
2
C H ϩ OH L CH CB CH ϩ H O
6
6
2
3
2
(Continued)

5
18
ALZUETA, GLARBORG, AND DAM-JOHANSEN
Appendix A (Continued)
A
␤
Ea
2778
13,909
47,600
47,600
47,600
0
Source
[73]
192.
193.
194.
195.
196.
197.
198.
199.
200.
201.
202.
203.
204.
205.
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
217.
C H ϩ OH L CB HCHCH ϩ H O
2.1E06
9.6E03
2.0E13
2.0E13
2.0E13
5.0E13
1.9E26
1.8E14
1.0E13
1.0E13
1.8E13
5.0E15
1.1E10
2.8E25
1.0E14
2.0E13
1.0E13
1.1E23
1.0E14
4.0E13
1.0E14
2.0E13
1.1E22
2.2E14
1.0E13
1.2E11
5.6E33
2.00
2.60
0.00
0.00
0.00
0.00
Ϫ3.60
0.00
0.00
0.00
Ϫ0.41
Ϫ1.40
0.34
Ϫ4.80
0.00
0.00
0.00
Ϫ3.29
0.00
0.00
3
6
3
2
C H ϩ HO L CH CHCB H ϩ H O
3
[73]
[73]
[73]
[73]
[28]
[28]
[28]
[73]
[73]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[27]
[28]
6
2
2
2
2
2
C H ϩ O L CH CHCb H ϩ HO
3
6
2
2
2
2
C H ϩ O L CB HCHCH ϩ HO
3
6
2
3
2
C H ϩ O L CH CB CH ϩ HO
3
6
2
2
3
2
CH CHCb H ϩ H L H CCCH ϩ H
2
2
2
2
2
CH CHCb H ϩ H L C H
5470
0
2
2
3
6
CH CHCb H ϩ O L CH CHCHO ϩ H
2
2
2
CH CHCb H ϩ OH L H CCCH ϩ H O
0
0
2
2
2
2
2
CH CHCb H ϩ HO L CH CHCHO ϩ H ϩ OH
2
2
2
2
CH CHCb H ϩ O L CH CHCHO ϩ OH
22,860
22,430
12,840
15,470
0
0
0
3900
0
2
2
2
2
CH CHCb H ϩ O L H CCCH ϩ HO
2
2
2
2
2
2
CH CHCb H ϩ O L CH HCO ϩ CH O
2
2
2
2
2
CH CHCB H ϩ O L C H ϩ CH O ϩ OH
2
2
2
2
2
2
Cb HCHCH ϩ H L CH CHCB H ϩ H
3
2
2
Cb HCHCH ϩ H L H CCCH ϩ H
3
3
2
Cb HCHCH ϩ OH L H CCCH ϩ H O
3
3
2
CB HCHCH ϩ O L CH HCO ϩ HCO
3
2
3
CH Cb CH ϩ H L CH CHCB H ϩ H
2
3
2
2
CH Cb CH ϩ H L H CCCH ϩ H
0
0
0
2
3
3
2
CH Cb CH ϩ O L CH CO ϩ CH
0.00
0.00
Ϫ3.29
0.00
0.00
2
3
2
3
CH Cb CH ϩ OH L H CCCH ϩ H O
2
3
3
2
CH CB CH ϩ O L CH CO ϩ CH O
3900
68,100
5000
3000
4450
2
3
2
3
2
H CCCH L CH CCH
2
2
3
H CCCH ϩ H L CH CCH ϩ H
2
2
3
H CCCH ϩ H(ϩM) L CH CHCB H (ϩM)
0.69
Ϫ5.00
2
2
2
2
Low pressure limit:
Troe parameters: 0.7086 134 1784 5740
2
18.
H CCCH ϩ H(ϩM) L CH CB CH (ϩM)
8.5E12
1.1E34
0.00
Ϫ5.00
2000
4450
[28]
2
2
2
3
Low pressure limit:
Troe parameters: 0.7086 134 1784 5740
219.
220.
221.
222.
223.
224.
H CCCH ϩ O L C H ϩ CO
1.3E07
3.0E07
2.0E07
3.0E07
1.0E14
6.5E12
8.5E39
1.88
2.00
2.00
2.00
0.00
180
5000
1000
5000
4000
2000
6580
[28]
[27]
[27]
[27]
[27]
[28]
2
2
2
4
H CCCH ϩ H L H CCCB H ϩ H
2
2
2
2
H CCCH ϩ OH L H CCCB H ϩ H O
2
2
2
2
CH CCH ϩ H L H CCCB H ϩ H
3
2
2
CH CCH ϩ H L CH ϩ C H
2
3
3
2
CH CCH ϩ H(ϩM) L CH CB CH (ϩM)
0.00
Ϫ7.27
3
2
3
Low pressure limit:
Troe parameters: 0.7086 134 1784 5740
225.
226.
227.
228.
229.
230.
231.
232.
233.
CH CCH ϩ OH L H CCCB H ϩ H O
2.0E07
1.5E13
2.0E13
4.0E13
7.2E12
5.0E07
1.0E13
3.0E13
1.0E17
3.5E55
2.00
0.00
0.00
0.00
0.00
1.76
0.00
0.00
Ϫ0.82
Ϫ4.88
1000
2100
3500
4200
1970
76
[27]
[28]
[28]
[28]
[28]
[28]
[28]
[28]
[27]
3
2
2
CH CCH ϩ O L C H ϩ CO
3
2
4
CH CHCHO ϩ H L C H ϩ HCO
2
2
4
CH CHCHO ϩ H L CH CHCb O ϩ H
2
2
2
CH CHCHO ϩ O L CH CHCB O ϩ OH
2
2
CH CHCHO ϩ O L CH CO ϩ HCO ϩ H
2
2
CH CHCHO ϩ OH L CH CHCB O ϩ H O
0
2
2
2
CH CHCHO ϩ O L CH CHCB O ϩ HO
36,000
315
2225
2
2
2
2
H CCCB H ϩ H(ϩM) L H CCCH (ϩM)
2
2
2
Low pressure limit:
Troe parameters: 0.7086 134 1784 5740
Enhanced third-body efficiencies:
H O ϭ 8.6
2
2
34.
H CCCB H ϩ H(ϩM) L CH CCH(ϩM)
1.0E17
3.5E55
Ϫ0.82
Ϫ4.88
315
2225
[27]
2
3
Low pressure limit:
Troe parameters: 0.7086 134 1784 5740
(Continued)

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
519
Appendix A (Continued)
A
␤
Ea
Source
Enhanced third-body efficiencies:
H O ϭ 8.6
2
235.
236.
237.
238.
239.
240.
241.
242.
243.
244.
245.
246.
247.
248.
249.
H CCCB H ϩ H L C H ϩ H
5.0E13
1.4E14
2.0E13
3.0E10
1.0E13
1.0E14
1.0E14
5.4E20
1.0E14
5.0E13
2.0E12
2.0E13
3.0E13
1.0E13
3.0E13
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Ϫ2.72
0.00
0.00
0.00
0.00
0.00
0.00
0.00
1000
[27]
[27]
[27]
[27]
[27]
[28]
[28]
[35] qrrk
[27]
[27]
[27]
[27]
[27]
[27]
[35] est
2
3
2
2
H CCCb H ϩ O L CH O ϩ C H
2
0
0
2868
2
2
H CCCb H ϩ OH L C H ϩ H O
2
3
2
2
H CCCb H ϩ O L CH CO ϩ HCO
2
2
2
H CCCB H ϩ H CCCB H L C H ϩ H
0
2
2
6
5
CH CHCb O L C H ϩ CO
34,000
2
2
3
CH CHCb O ϩ O L C H ϩ CO
0
7000
0
2
2
3
2
CH CHCB O ϩ O L CH HCO ϩ CO
2
2
2
2
C H ϩ O L C H ϩ CO
3
2
2
2
C H ϩ OH L C H ϩ HCO
0
3
2
2
2
C H ϩ O L HCCO ϩ CO ϩ H
1000
0
0
0
4000
3
2
2
C H ϩ CH L CH CHCCH ϩ H
3
2
3
2
C H ϩ HCCO L HCB CHCCH ϩ CO
3
2
C H ϩ H CCCB H L C H ϩ H
3
2
2
6
4
O CCHOOB L HCOO ϩ CO
2
2
a
b
c
d
e
f
In the mechanism, no distinction is made between isomers ofthe cyclic compounds C H O, C H OH, and C H .
5
5
5
5
5
7
Estimated by analogy to C H reactions.
6
6
Estimated by analogy to C H hydrocarbon reactions (x ϭ 5–8).
3
x
Estimated by analogy to CH OH or C H OH reactions.
3
2
5
Lumped reaction, assuming rapid dissociation ofcyclobutadiene.
To bring the rate constant in accordance with the revised heat of formation for C H O, the activation energy was raised by 6 kcal/mol.
5
4
g
h
i
Estimated by analogy to CH O reactions.
2
Based on estimate for CH CHCHCB HO : CH CHCB H ϩ CO [35].
2
2
2
Estimated by analogy to C hydrocarbon reactions (C H , C H , or CH CO).
2
2
4
2
2
2
j
Based on estimate for CH CHCH CB O L H CCCB H ϩ CO [35].
2
2
2
For “duplicate reactions” the resulting rate constant is the sum ofthe rate constants listed.
(not shown) and CO are significantly underpredicted.
CONCLUSIONS
These deviations may partly be attributed to the dif-
ference in temperature for observed and predicted on-
set ofbenzene oxidation. Furthermore, the uncertainty
in the overall oxidation pathways for benzene at these
temperatures and the probability ofmissing consump-
tion steps in the mechanism will affect the predictions
for benzoquinone and cyclopentadienone. However,
the large discrepancy observed for phenol is difficult
to understand. The jet-stirred-reactor data appear to be
incompatible with the flow-reactor data ofFigure 7,
which are fairly well predicted by the mechanism.
Phenol is formed and consumed by the same reactions
under well-mixed conditions as under flow-reactor
conditions (Fig. 4), even though the relative impor-
tance of reactions is slightly different due to differ-
ences in the concentration and composition ofthe rad-
ical pool. With a few exceptions, these reactions are
fairly well characterized and we were not able to ob-
tain agreement between the measured and calculated
phenol profile by adjusting any ofthese steps within
their uncertainty limits. Further work is required to
resolve this issue.
The oxidation ofbenzene has been studied in an iso-
thermal flow reactor at atmospheric pressure and tem-
peratures in the range 900–1400 K. The excess-air
ratio was varied from close to stoichiometric to very
lean conditions. In addition, the radical pool compo-
sition was perturbed by varying the levels ofwater
vapor (0.5 to 5%) and by adding NO (0 and 100 ppm).
The results were analyzed in terms ofa detailed chem-
ical kinetic model, which has been developed in this
work.
Results indicate that the oxidation rate ofbenzene
is mainly controlled by temperature and stoichiometry.
The onset ofoxidation occurs at ϳ1000 K; complete
oxidation occurs between 1100 and 1300 K, depend-
ing on the stoichiometry. Under lean conditions, pres-
ence ofNO uf rther lowers the initiation temperature
by 60–80 K. The effect of the OH/O ratio is less im-
portant.
A chemical kinetic model for benzene oxidation
has been developed. It was evaluated by comparison
with experimental data from the present work as well
as flow-reactor and mixed-reactor data from literature.

5
20
ALZUETA, GLARBORG, AND DAM-JOHANSEN
Appendix B Thermodynamic Properties for Selected Species
Species
⌬Hf298
S298
Cp300
Cp400
Cp500
Cp600
Cp800
Cp1000
Cp1500
Source
C H6
19.81
80.73
121.76
10.35
Ϫ25.01
Ϫ29.37
Ϫ25.00
27.03
39.88
31.99
62.00
42.94
Ϫ9.02
13.20
15.90
131.79
68.95
166.75
18.19
49.31
49.11
97.09
7.18
64.35
19.92
21.01
18.61
24.79
25.43
26.04
27.09
27.06
24.33
31.31
32.11
32.20
33.25
32.43
29.19
37.08
38.09
37.67
38.38
37.05
33.19
42.01
43.28
42.27
45.87
43.90
38.90
49.25
51.12
48.87
51.05
47.77
42.81
53.28
55.68
53.42
58.31
53.26
48.24
58.98
62.24
59.48
[30]
[30,26]
[30]
[30]
[30]
[31]
[74]
[35]
[31]
[30]
[23,30]
[18]
[18]
[23,31]
[23,31]
[31]
[35]
[31]
[31]
[17]
[17]
[31]
[17]
[35]
[35]
[30]
[30]
[30]
[30]
[30]
[30]
[35]
[35]
[18]
[35]
[18]
[35]
[30]
[30]
[30]
[30]
[30]
[30]
[30]
[30]
[30]
[18]
[35]
[18]
[18]
[35]
6
C H5
69.82
69.53
74.87
76.93
79.62
6
C H4
6
C H O
6
5
C H OH
6
5
pOC H O
6
4
a
oOC H O
6
4
OC H O
77.15
73.65
64.45
59.43
72.73
75.33
66.49
74.08
66.81
69.20
67.32
76.48
75.93
77.20
73.38
85.00
87.20
102.84
70.44
75.31
73.06
67.33
69.06
72.94
87.61
90.11
76.87
81.79
71.96
73.81
66.43
61.51
64.73
69.24
68.74
57.94
58.89
61.48
64.81
67.39
76.87
69.24
68.07
88.68
24.86
19.23
16.67
34.75
20.61
22.09
19.70
22.98
17.60
18.69
17.04
24.23
21.07
22.52
21.94
25.02
27.68
30.36
18.58
19.49
19.38
17.32
18.02
20.24
26.38
25.55
19.43
20.18
21.66
21.99
17.27
15.46
16.07
15.45
15.54
14.25
14.51
15.84
14.93
16.98
20.40
17.48
15.91
21.91
31.62
26.40
23.12
36.13
27.19
29.03
25.79
29.34
22.64
24.13
21.74
30.73
27.67
29.00
27.24
32.27
34.31
35.62
23.16
23.12
23.47
20.61
20.82
22.43
30.97
29.54
24.28
24.69
26.31
26.19
19.62
19.27
19.55
18.38
18.56
16.97
17.06
17.74
16.10
20.89
24.07
20.61
19.28
23.72
37.21
32.57
28.70
37.46
32.59
34.71
30.69
34.50
27.00
28.51
25.59
36.36
33.14
34.25
31.43
38.09
39.88
40.36
27.42
26.47
27.21
23.62
23.29
24.44
35.03
33.07
28.43
28.55
29.92
29.09
21.79
22.73
22.72
21.21
21.44
19.46
19.40
19.47
16.91
24.09
27.22
23.12
22.00
25.54
41.37
37.64
33.39
38.75
36.99
39.34
34.62
38.52
30.53
32.04
28.60
41.10
37.68
38.56
34.80
42.76
44.55
44.59
31.26
29.48
30.54
26.30
25.41
26.23
38.61
36.19
31.96
31.85
32.71
31.08
23.73
25.80
25.53
23.89
24.12
21.71
21.54
21.01
17.55
26.70
29.89
25.11
24.19
27.29
47.25
44.82
40.31
41.18
43.45
46.16
40.25
43.83
35.25
37.09
32.48
47.91
44.57
45.20
39.89
49.62
51.79
51.63
37.44
34.32
35.79
30.53
28.56
29.10
44.46
41.31
37.47
37.05
36.53
33.65
26.81
30.78
29.99
28.50
28.65
25.45
25.16
23.43
18.72
30.57
34.09
27.94
27.37
30.33
51.64
49.77
45.10
43.39
47.72
50.71
43.90
47.57
38.60
40.34
35.15
51.19
49.43
50.19
43.18
54.33
57.03
56.94
41.29
37.37
38.89
33.20
30.46
30.93
48.88
45.17
41.38
40.80
38.95
35.66
28.73
34.52
32.89
31.60
31.68
28.20
27.90
25.00
19.74
33.22
37.11
29.76
29.47
32.54
6
3
C H7
57.31
51.82
47.92
53.79
57.40
49.00
52.86
43.20
44.70
38.73
60.25
56.81
57.42
48.29
61.52
65.21
64.64
46.79
41.84
43.16
37.25
33.46
33.68
55.45
50.88
46.93
46.33
43.25
38.78
31.51
40.14
37.43
36.09
36.13
32.06
31.79
27.55
21.22
37.44
41.57
32.54
32.73
34.45
5
C H6
5
C H5
5
C H O
5
5
C H OH
5
5
C H O
5
4
C H OH
5
4
C H4
5
C H O
5
3
C H3
5
CH CHCHCHCH3
2
CH CHCB HCHCH
2
2
CH CHCHCHCB H
2
2
CHCCHCHCB H
2
CH CHCB HCHCHOH
2
CH CHCHCHCH OH
Ϫ16.25
Ϫ27.18
28.29
74.14
86.09
2
2
CHOCH CH CB HCHO
2
2
CH CHCHCH
2
2
CH CHCB CH
2
2
CH CHCHCB H
2
CH CHCCH
69.14
2
HCCHCCH
129.88
111.32
Ϫ68.60
Ϫ31.70
15.45
18.85
1.82
60.92
54.59
4.89
38.64
61.09
64.75
47.63
45.77
83.04
H CCCCH
2
CHOCH CH CHO
2
2
CHOCH CH CB O
2
2
CH CHCB HCHO
2
CH CHCH CB O
2
2
CH CHCHCO
2
CB HCHCHCO
H C O
2
4
CH CHCH3
2
CH CHCB H
2
2
CH CB CH
2
3
CB HCHCH
3
H CCCH
2
2
H CCCH
3
H CCCH
2
C H2
129.60
Ϫ20.32
Ϫ62.45
38.78
10.58
Ϫ61.69
3
CH CHCHO
2
CHOCHCHOH
CB HCHCHO
CH CHCB O
2
O CCHOOB
2
a
As an approximation, entropy and heat of formation for orthobenzoquinone were assumed similar to parabenzoquinone.

EXPERIMENTAL AND KINETIC MODELING STUDY OF THE OXIDATION OF BENZENE
521
The model provides a fairly good description of the
overall oxidation behavior ofbenzene over the range
ofconditions investigated. Thereby, the mechanism
can be used with some confidence to simulate benzene
oxidation under postflame conditions. However, a
number ofkinetic issues need to be addressed be fo re
concentrations ofcyclic and linear intermediates can
be predicted satisfactorily. These include rate and
product channels for the reactions of phenyl and cy-
clopentadienyl with molecule oxygen as well as re-
12. Zhang, H.-Y.; McKinnon, J. T. Combust Sci Techn
995, 107, 261.
3. Lindstedt, R. P.; Maurice, L. Q. Combust Sci Techn
996, 120, 119.
4. Tan, Y.; Frank, P. Twenty-Sixth Symposium (Interna-
tional) on Combustion; The Combustion Institute: Pitts-
burgh, PA, 1996; p 677.
1
1
1
1
1
1
5. Shandross, R. A. S.; Lovell, J. P.; Howard, J. B.
Twenty-Sixth Symposium (International) on Combus-
tion; The Combustion Institute: Pittsburgh, PA, 1996; p
711.
6. Davis, S. G.; Wang, H.; Brezinsky, K.; Law, C. K.
Twenty-Sixth Symposium (International) on Combus-
tion; The Combustion Institute: Pittsburgh, PA, 1996; p
action chemistry for the oxygenated cyclic C and C5
6
compounds.
1
025.
7. Zhong, X.; Bozzelli, J. W. Int J Chem Kinet 1997, 29,
93.
8. Zhong, X.; Bozzelli, J. W. J Phys Chem A 1998, 102,
537.
9. Chai, Y.; Pfefferle, L. D. Fuel 1998, 77, 313.
0. Frank, P.; Herzler, J.; Just, Th.; Wahl, C. Twenty-Fifth
Symposium (International) on Combustion; The Com-
bustion Institute: Pittsburgh, PA, 1994; p 833.
1
1
8
The authors would like to thank Professor J. W. Bozzelli for
sharing results on aromatics chemistry prior to publication.
The work was carried out as part ofthe combustion research
program ofthe Department ofChemical Engineering ofthe
Technical University ofDenmark (CHEC). The CHEC re-
search program is cofunded by the Danish Technical Re-
search Council, Elsam, Elkraft, and the Danish Ministry of
Energy. M. U. Alzueta acknowledges CAI and CONSID for
the “Ayuda Europa” awarded for carrying out part of this
research. In addition, the work was funded by the Human
Capital and Mobility program ofthe European Union (Con-
tract No. CHRX-CT94-0436).
3
1
2
21. Kumaran, S. S.; Michael, J. V. In Phenyl Radical Ther-
molysis and Rate Constants for Phenyl ϩ O , Proceed-
2
ings ofthe 21st International Symposium on Shock
Waves, Great Keppel Island, Australia, 1997.
2. Alzueta, M. U.; Oliva, M.; Glarborg, P. Int J Chem Ki-
net 1998, 30, 683.
2
2
2
2
3. Wang, H.; Brezinsky, K. J. Phys Chem A 1998, 102,
1
530.
4. Miller, J. A.; Melius, C. F. Combust Flame 1992, 91,
1.
BIBLIOGRAPHY
2
5. Baulch, D. C.; Cobos, C. J.; Cox, R. A.; Esser, C.;
Frank, P.; Just, T.; Kerr, J. A.; Pilling, M. J.; Troe, J.;
Walker, R. W.; Warnatz, J. J. Phys Chem RefData
1
2
3
. Bittner, J. D.; Howard, J. B. Eighteenth Symposium (In-
ternational) on Combustion; The Combustion Institute:
Pittsburgh, PA, 1981; p 1105.
. Venkat, C.; Brezinsky, K.; Glassman, I. Nineteenth
Symposium (International) on Combustion; The Com-
bustion Institute: Pittsburgh, PA, 1982; p 143.
1992, 21, 411.
2
6. Baulch, D. C.; Cobos, C. J.; Cox, R. A.; Frank, P.; Hay-
man, C.; Just, T.; Kerr, J. A.; Pilling, M. J.; Troe, J.;
Walker, R. W.; Warnatz, J. Combust Flame 1994, 98,
. Hsu, H.; Lin, C. Y.; Lin, M. C. Twentieth Symposium
59.
(International) on Combustion; The Combustion Insti-
2
2
7. Pauwels, J.-F.; Volponi, J. V.; Miller, J. A. Combust Sci
Techn 1995, 110, 249.
8. Marinov, N. M.; Pitz, W. J.; Westbrook, C. K.; Castaldi,
M. J.; Senkan, S. M. Combust Sci Techn 1996, 116,
tute: Pittsburgh, PA, 1984; p 623.
4
5
. Brezinsky, K. Prog Energy Combust Sci 1986, 12, 1.
. Burcat, A.; Snyder, C.; McBride, B. J. Ignition delay
times ofbenzene and toluene with oxygen in argon
mixtures; NASA TM-87312, 1986.
211.
2
3
9. Glarborg, P.; Alzueta, M. U.; Dam-Johansen K.; Miller,
J. A. Combust Flame 1998, 115, 1.
0. Kee, R. J.; Rupley, F. M.; Miller, J. A. The Chemkin
Thermodynamic Data Base, Sandia Report SAND87-
6
7
. Lovell, A. B.; Brezinsky, K.; Glassman, I. Twenty-Sec-
ond Symposium (International) on Combustion; The
Combustion Institute: Pittsburgh, PA, 1988; p 1063.
. Lyon, R. K. Twenty-Third Symposium (International)
on Combustion; The Combustion Institute: Pittsburgh,
PA, 1990; p 903.
8
1
215; Sandia National Laboratories: Livermore, CA,
991.
3
1. Burcat, A.; McBride, B. 1995 Ideal Gas Thermody-
namic Data for Combustion and Air-Pollution Use,
Aerospace Engineering Report TAE 732; Institute of
Technology: Technian, Israel, 1995 and updates.
32. Kristensen, P. G.; Glarborg, P.; Dam-Johansen, K.
Combust Flame 1996, 107, 211.
8
9
. Bittker, D. A. Combust Sci Techn 1991, 79, 49.
. Emdee, J. L.; Brezinsky, K.; Glassman, I. J. Phys Chem
1
992, 96, 2151.
0. Lindstedt, R. P.; Skevis, G. Combust Flame 1994, 99,
51.
1. Leung K. M.; Lindstedt, R. P. Combust Flame 1995,
02, 129.
1
1
5
33. Alzueta, M. U.; Glarborg, P.; Dam-Johansen, K. Com-
bust Flame 1997, 109, 25.
1

5
22
ALZUETA, GLARBORG, AND DAM-JOHANSEN
3
3
4. Alzueta, M. U.; Røjel, H.; Kristensen, P.; Glarborg, P.;
Dam-Johansen, K. Energy Fuels 1997, 11, 716.
5. Bozzelli, J. Private communication ofpreliminary re-
sults. See also Zhong, X. Ph.D. Thesis, Rutgers Uni-
versity: Newark, NJ, May 1998.
6. Heckmann, E.; Hippler, H.; Troe, J. Twenty-Sixth Sym-
posium (International) on Combustion; The Combus-
tion Institute: Pittsburgh, PA, 1996; p 543.
7. Mebel, A. M.; Lin, M. C.; Yu, T.; Morokuma, K. J. Phys
Chem A 1997, 101, 3189.
8. Nicovich, J. M.; Gump, C. A.; Ravishankara, A. R. J
Phys Chem 1982, 86, 1684.
56. Roy, K.; Frank, P. Twenty-First International Sympo-
sium on Shock Waves, Great Keppel Island, Australia,
1997, paper 1560.
57. Horn, C.; Roy, K.; Frank, P.; Slutsky, V.; Just, Th.
Twenty-Seventh Symposium (International) on Com-
bustion; The Combustion Institute: Pittsburgh, PA,
1998; p 329.
3
58. Burcat, A.; Dvinyaninov, M. Int J Chem Kinet 1997,
29, 505.
3
3
3
4
4
4
59. Bruinsma, O-S. L.; Tromp, P. J. J.; de Sauvage Nolting,
H. J. J.; Moulijn, J. A. Fuel 1988, 67, 334.
60. Lutz, A. E.; Kee, R. J.; Miller, J. A. SENKIN: A Fortran
Program for Predicting Homogeneous Gas Phase Chem-
ical Kinetics with Sensitivity Analysis, Sandia Report
SAND87-8248; Sandia National Laboratories: Liver-
more, CA, 1987.
61. Glarborg, P.; Kee, R. J.; Grcar, J. F.; Miller; J. A. PSR:
A Fortran Program for Modeling Well-Stirred Reactors,
Sandia Report SAND86-8209; Sandia National Labo-
ratories: Livermore, CA, 1986.
62. Kee, R. J.; Rupley, F. M.; Miller J. A. CHEMKIN-II:
A Fortran Chemical Kinetics Package for the Analysis
ofGas-Phase Chemical Kinetics, Sandia Report
SAND89-8009; Sandia National Laboratories: Liver-
more, CA, 1989.
9. Bajaj, P. N.; Fontijn, A. Combust Flame 1996, 105,
2
39.
0. Leidreiter, H. I.; Wagner, H. Gg. Z Phys Chem (Neue
Folge) 1989, 165, 1.
1. Frenklach, M.; Clary, D. W.; Yuan, T.; Gardiner, W. C.
Stein, S. E. Combust Sci Techn 1986, 50, 79.
2. Castaldi, M. C.; Marinov, N. M.; Melius, C. F.; Huang,
J.; Senkan, S. M.; Pitz, W.; Westbrook, C. K. Twenty-
Fifth Symposium (International) on Combustion; The
Combustion Institute: Pittsburgh, PA, 1994; p 693.
3. Lin, C. Y.; Lin, M. C. J Phys Chem 1986, 90, 425.
4. Mebel, A. M.; Lin, M. C. J Am Chem Soc 1994, 116,
4
4
9
577.
4
4
5. Yu, T.; Lin, M. C. J Am Chem Soc 1994, 116, 9571.
6. Tranter, R. S.; Grotheer, H. H.; Just, Th. Fourteenth In-
ternational Gas Kinetics Symposium, Leeds 1996, paper
F7.
7. Knispel, R.; Koch, R.; Siese, M.; Zetzsch, C. Ber Bun-
senges Phys Chem 1990, 94, 1375.
8. Horn, C.; Roy, K.; Frank, P.; Just, Th. Twenty-Seventh
Symposium (International) on Combustion; The Com-
bustion Institute: Pittsburgh, PA, 1998; p 321.
9. Olivella, S.; Sol e` , A.; Garcia-Raso, A. J Phys Chem
63. He, Y. Z.; Mallard, W. G.; Tsang, W. J Phys Chem
1988, 92, 2196.
64. Cypr e` s, R.; Bettens, B. Tetrathedron 1974, 30, 1253.
65. Brezinsky, K.; Pecullan, M.; Glassman, I. J Phys Chem
A 1998, 102, 8614.
66. Hjuler, K.; Glarborg, P.; Dam-Johansen, K. Ind Engn
Chem Res 1995, 34, 1882.
67. Glarborg, P.; Kristensen, P. G.; Kubel, D.; Hansen, J.;
Dam-Johansen, K. Combust Sci Techn 1995, 110–111,
46.
68. Alzueta, M. U.; Muro, J.; Bilbao, R.; Glarborg, P. Israel
J Chem 1999, 39, 73.
69. Bendtsen, A. B.; Glarborg, P.; Dam-Johansen, K. Com-
bust Sci Technol 2000, 151, 31–72.
70. Ackerman, L.; Hippler, H.; Reihs, C.; Troe, J. J Phys
Chem 1990, 94, 5247.
4
4
4
5
5
1
995, 99, 10549.
0. Liu, R.; Morokuma, K.; Mebel, A. M.; Lin, M. C. J Phys
Chem 1996, 100, 9314.
1. Buth, R.; Hoyermann, K.; Seeba, J. Twenty-Fifth Sym-
posium (International) on Combustion; The Combus-
tion Institute: Pittsburgh, PA, 1994; p 841.
5
5
5
5
2. Lin, M. C.; Mebel, A. M. J Phys Org Chem 1995, 8,
71. Fahr, A.; Stein, S. E. J Phys Chem 1988, 92, 4951.
72. Arends, I. W. C. E.; Louw, R.; Mulder, P. J Phys Chem
1993, 97, 7914.
73. Tsang, W. J Phys Chem RefData 1991, 20, 221.
74. Lias, S. G.; Liebman, J. F.; Levin, R. D.; Kafafi, S. A.
NIST Standard Reference Database 25: Structures and
Properties, version 2.02, 1994.
4
07.
3. de Jongh, D. C.; van Fossen, R. Y., Bourgeois, C. F.
Tetrahedron Lett 1967, 3, 271.
4. Schraa, G.-J.; Arends, I. W. C. E.; Mulder, P. J Chem
Soc Perkin Trans 1994, 2, 189.
5. Sakai, T.; Hattori, M.; Yamane, N. J. Japan Petrol Int
1
980, 23, 44.