Parabenzoquinone pyrolysis and oxidation in a flow reactor

Article abstract of DOI:10.1002/(SICI)1097-4601(1998)30:9<683::AID-KIN9>3.0.CO;2-O

An experimental and theoretical study of the pyrolysis and oxidation of parabenzoquinone has been performed. The experiments were conducted in an isothermal quartz flow reactor at atmospheric pressure in the temperature range 600-1500 K. The main variables considered are temperature, oxygen concentration, and presence of CO. A detailed reaction mechanism for the pyrolysis and oxidation chemistry of parabenzoquinone is proposed, which provides a good description of the experimental results. Both the experimental work and the kinetic mechanism proposed for the pyrolysis and oxidation of parabenzoquinone represent the first systematic study carried out for this important aromatic compound. Our pyrolysis results confirm that the primary dissociation channel for p-benzoquinone leads to CO and a C₅H₄O isomer, presumably cyclopentadienone. However, significant formation of CO₂ during the pyrolysis may indicate the existence of a secondary dissociation channel leading to CO₂ and a C₅H₄ isomer. Under oxidizing conditions, consumption of p-benzoquinone occurs mainly by dissociation at lower temperatures. As the temperature increases interaction of OC₆H₄O with the radical pool becomes more significant, occurring primarily through hydrogen abstraction reactions followed by ring opening reactions of the OC₆H₃O radical.

Full text of DOI:10.1002/(SICI)1097-4601(1998)30:9<683::AID-KIN9>3.0.CO;2-O

Parabenzoquinone
Pyrolysis and Oxidation
in a Flow Reactor
MARIA U. ALZUETA,1 MIRIAM OLIVA,2 PETER GLARBORG3
1,2Department of Chemical and Environmental Engineering, University of Zaragoza, Centro Polite´cnico Superior. Mar´ıa de
Luna, 3, 50015 Zaragoza, Spain
3Department of Chemical Engineering, Technical University of Denmark, 2800 Lyngby, Denmark
Received 29 September 1997; accepted 7 April 1998
ABSTRACT: An experimental and theoretical study of the pyrolysis and oxidation of paraben-
zoquinone has been performed. The experiments were conducted in an isothermal quartz flow
reactor at atmospheric pressure in the temperature range 600–1500 K. The main variables
considered are temperature, oxygen concentration, and presence of CO. A detailed reaction
mechanism for the pyrolysis and oxidation chemistry of parabenzoquinone is proposed, which
provides a good description of the experimental results. Both the experimental work and the
kinetic mechanism proposed for the pyrolysis and oxidation of parabenzoquinone represent
the first systematic study carried out for this important aromatic compound.
Our pyrolysis results confirm that the primary dissociation channel for p-benzoquinone
leads to CO and a C5H4O isomer, presumably cyclopentadienone. However, significant for-
mation of CO2 during the pyrolysis may indicate the existence of a secondary dissociation
channel leading to CO2 and a C5H4 isomer. Under oxidizing conditions, consumption of p-
benzoquinone occurs mainly by dissociation at lower temperatures. As the temperature in-
creases interaction of OC6H4O with the radical pool becomes more significant, occurring pri-
marily through hydrogen abstraction reactions followed by ring opening reactions of the
OC6H3O radical. ᭧ 1998 John Wiley & Sons, Inc. Int J Chem Kinet: 30: 683–697, 1998
INTRODUCTION
about some of the intermediates formed during the ox-
idation process. A number of recent studies have
pointed to the importance of benzoquinone, OC6H4O,
which appears to be an important product in the re-
action between phenyl and oxygen [12,13],
Understanding of the detailed oxidation chemistry of
aromatic compounds is still a major challenge. Despite
significant efforts during the last two decades [1–9]
large uncertainties remain in the reaction mechanism
for even the simplest aromatic species, such as ben-
zene. While the initial steps in benzene oxidation are
now better understood [4–6, 8–11], little is known
C6H5 ϩ O2 EF OC6H4O ϩ H,
and may also be formed by reaction of the phenoxy
radical with oxygen atoms [14],
Correspondence to: M. U. Alzueta
Contract grant Sponsor: Department of Chemical and Environ-
mental Engineering of the University of Zaragoza and Department
of Chemical Engineering of the Technical University of Denmark
(CHEC). The CHEC research program is cofunded by the Danish
Technical Research Council, Elsam, Elkraft, and the Danish Min-
istry of Energy
C6H5O ϩ O EF OC6H4O ϩ H.
In addition to being an important intermediate in ben-
zene oxidation, benzoquinone has been identified as a
potential pollutant from combustion facilities such as
᭧ 1998 John Wiley & Sons, Inc.
CCC 0538-8066/98/090683-15

684
ALZUETA, OLIVA, AND GLARBORG
been performed under well-defined reaction condi-
tions, making it difficult to derive reaction specific in-
formation.
It appears that the two isomers of benzoquinone
have different reaction characteristics, even though
little mechanistic information is available. Flash ther-
molysis of the two isomers at temperatures around
1100 K [18,19,26,27] indicate that the pyrolysis of p-
benzoquinone and o-benzoquinone may proceed along
fundamentally different pathways, perhaps because
different C5H4O isomers are formed in the initial py-
rolysis step [18,19]. Parabenzoquinone is more ther-
mally stable [12] than orthobenzoquinone, which has
been reported to dissociate readily [26,28]. This is in
agreement with the general observation that o-qui-
nones are decarbonylated more readily than p-qui-
nones [23].
The objective of the present work is to perform an
experimental and theoretical study of parabenzoqui-
none pyrolysis and oxidation. Experiments are con-
ducted in a flow reactor in the temperature range from
600 to 1500 K at atmospheric pressure. Based on the
present data as well as results from the literature, a
detailed reaction mechanism for the chemistry of par-
abenzoquinone is proposed, that provides a good de-
scription of the experimental results under both py-
rolysis and oxidation conditions.
Figure 1 Diagram of the quartz-flow reactor. 1: Main flow.
2: Gas injectors. 3: Mixing area. 4: Reaction zone. 5: Gas
outlet. 6: Cooling air inlet.
EXPERIMENTAL
The experimental installation basically consists of a
gas-feeding system, a gas-reaction system, and a gas-
analysis system. Pure gases from gas cylinders are
dosed with mass-flow controllers and fed into the re-
actor. The addition of p-benzoquinone together with
water vapor was done by passing a nitrogen stream
through a saturated solution of p-benzoquinone at a
temperature of 303 K.
The reaction system contains a quartz-flow reactor
which is located inside an electrically heated oven that
allows us to reach temperatures up to 1500 K. The
reactor, which has been constructed following the de-
sign of Kristensen et al. [29], is shown schematically
in Figure 1. It has a reaction zone of 8.7 mm inside
diameter and 200 mm in length. Reaction conditions
are carefully controlled. The reactants are led sepa-
rately to the reactor and rapidly and efficiently mixed
at the entrance of the reaction zone. The temperature
is kept uniform along the reaction zone within 7 K.
The gas entering into the reaction zone consists of
a mixture of p-benzoquinone, water vapor, and nitro-
gen, with oxygen added in the oxidation experiments.
The reactants are very diluted in order to keep iso-
sewage sludge incinerators [15]. Despite the potential
importance of benzoquinone, very little is currently
known about the chemistry of this component. Two
isomers of benzoquinone may be of significance, par-
abenzoquinone (p-OC6H4O, in this article just termed
OC6H4O), and orthobenzoquinone (o-OC6H4O). It is
not clear at present which form of benzoquinone is
favored in the C6H5 ϩ O2 and C6H5O ϩ O reactions,
but it is likely that both isomers play a role [16,17].
A limited amount of experimental data on the
chemistry of benzoquinone can be found in the liter-
ature. The only reaction specific work reported is a
shock-tube pyrolysis study of parabenzoquinone by
Frank et al. [12]. Further studies on p-benzoquinone
include flash-thermolysis experiments [18–20] and
thermal-hydrogenation experiments [21–25]. For o-
benzoquinone, to our knowledge only thermolysis data
have been reported [26,27]. While the reported studies
on benzoquinone have provided important information
on the pyrolysis chemistry, no data are available on
the oxidation behavior. Furthermore, except for the
shock-tube study, none of these experiments have

PARABENZOQUINONE PYROLYSIS AND OXIDATION
685
thermal conditions along the reaction zone. At the out-
let of the reactor, the reaction is quenched using cool-
ing air.
860 K in the presence of hydrogen by Sakai and co-
workers [21–24], point to a possible different inter-
pretation of the shock-tube results of Frank et al. [12].
We believe that a secondary pyrolysis channel may be
active,
The analysis of the products after conditioning is
performed by means of continuous analyzers for CO
and CO2, with an estimated uncertainty of the mea-
surements of 5%, but not less than 5 ppm. Further-
more, the products are analyzed using a FTIR (Fourier
Transform Infra-Red) analyzer. The reactor tempera-
ture is measured with a K thermocouple, and the pres-
sure with an absolute pressure transducer. Tempera-
ture control and data acquisition are performed in a PC
system.
OC6H4O EF C5H4 ϩ CO2,
(R2)
with a rate comparable to that of (R1) at high temper-
atures. This secondary channel was not detected by
Frank and co-workers, partly because they did not an-
alyze for the products of (R2) and partly because their
assumption of a high-thermal stability of C5H4O may
have been in error. As discussed in detail below,
we estimate rate constants for (R1) of 3.7 · 1011
1
exp(Ϫ29700/T) sϪ
and (R2) of 3.5 · 1012
REACTION MECHANISM
1
exp(Ϫ33700/T) sϪ , based on pyrolysis results carried
out in the present study as well as our interpretation
of the results of Sakai et al. and Frank et al. However,
the occurrence of reaction (R2), which would involve
the formation of a 1,4 bridge of O2, is presently only
a working hypothesis, and it needs to be confirmed
independently.
The reaction mechanism for benzoquinone pyrolysis
and oxidation developed in the present work consists
of an oxidation subset for C3 9C6 hydrocarbons,
listed in Appendix A, together with a C2 hydrocarbon
subset, which includes C1 hydrocarbon chemistry as
well as oxidation of CO/H2. The C2 subset was
adopted without changes from Glarborg et al. [30].
The C3 9C4 reaction subset was largely taken from
work of Miller and co-workers [31,32]. Thermody-
namic data were taken from different sources
[30,33,34], and Appendix B lists thermodynamic
properties (⌬Hf298, S298, and Cp) for selected species.
Reactions of benzoquinone and the subsequent C6 and
C5 intermediates are discussed in detail below.
As stated in the introduction, very little is currently
known about the detailed chemistry of benzoquinone.
Only pyrolysis and hydrogenation studies have been
reported in literature, and results from these studies
are not conclusive.
The thermal stability of parabenzoquinone was in-
vestigated in a shock-tube study by Frank et al. [12].
The temporal absorption of H, O, and CO was moni-
tored during the unimolecular decomposition of low
initial concentrations of p-benzoquinone. Frank et al.
[12] detected only CO in measurable quantities, with
[CO]end/[OC6H4O]0 ϭ 1.0 Ϯ 0.1 in the temperature
range 1550–1900 K. Based on this information and
assuming a high-thermal stability of C5H4O, they de-
duced that the reaction
To our knowledge there are no data available in the
literature on reactions of benzoquinone with radicals.
The reaction of OC6H4O with H atoms presumably has
two product channels,
OC6H4O ϩ H EF C5H5O ϩ CO
OC6H4O ϩ H !: OC6H3O ϩ H2
(R3)
(R4)
For reaction (R3) we have assumed that the H addition
to an unsubstituted ring position was rate limiting and
we have chosen a rate of four times the value for a
position in phenol [35,36]. The H atom may also add
to the oxygen in OC6H4O, which following H atom
transfer and subsequent dissociation would result in
the same products as (R3). For the H abstraction chan-
nel the rate constant was estimated by analogy with
the C6H6 ϩ H reaction. Due to the lack of thermody-
namic data for the OC6H3O radical, we have assumed
the reactions of this species to be irreversible.
The reaction of benzoquinone with O atoms was
also estimated by analogy with the corresponding ben-
zene reaction. The C6H6 ϩ O reaction has two impor-
tant product channels, one forming a C6H6O adduct,
which yields a phenoxy radical plus an H atom [37],
and an H abstraction channel leading to phenyl and
OH. By analogy we obtain for the OC6H4O ϩ O re-
action,
OC6H4O EF C5H4O ϩ CO
(R1)
was the only important dissociation channel. From the
pseudo-first-order increase of CO, they found a rate
1
constant for (R1) of 7.4 · 1011 exp(Ϫ29700/T) sϪ .
OC6H4O ϩ O !: C6H3O3 ϩ H
(R5)
(R6)
Results of the present work, supported by thermal
decarbonylation studies of parabenzoquinone at
OC6H4O ϩ O !: OC6H3O ϩ OH.

686
ALZUETA, OLIVA, AND GLARBORG
1
The reaction of parabenzoquinone with OH can be as-
sumed to proceed as an addition reaction at low tem-
perature, similar to the addition to the Ͼ C"C Ͻ
double bond in 1,4-napthoquinone [38,39]. However,
under the conditions of the present work, the adduct
formed presumably dissociates rapidly back to reac-
tants. At higher temperatures we assume that the H
abstraction reaction,
with a rate of 1.0·1015exp(Ϫ39300/T) sϪ . Reactions
of C5H4O with H and O radicals are expected to lead
to ring opening,
C5H4O ϩ H !: CH2CHCCH2 ϩ CO (R13)
C5H4O ϩ O !: CH2CHCCH ϩ CO2 (R14)
We assume that C5H4 formed in reaction (R2) is the
noncyclic 1,4 pentadiyne isomer. Reactions of this
component with the O/H radical pool presumably pro-
ceed as H abstraction reactions producing the C5H3
radical, which subsequently may react with O2 feeding
into the C2 hydrocarbon subset.
OC6H4O ϩ OH !: OC6H3O ϩ H2O, (R7)
is the only active channel. The rate for this reaction
can be estimated at 1050–1100 K from the present
work (see below), yielding
a
value of
1.5 · 1011 cm3/mole-s at 1075 K. We have extrapo-
lated this value, assuming a T2 dependence. It is note-
worthy that the reaction is significantly slower than
the C6H6 ϩ OH reaction.
The OC6H3O radical is expected to react very fast
with H and O radicals, leading to ring opening
RESULTS AND DISCUSSION
The pyrolysis and oxidation of p-benzoquinone has
been studied in the temperature range 600–1500 K at
atmospheric pressure in a flow reactor. The experi-
ments were conducted keeping a constant mass flow
rate and varying the reaction temperature. This means
that the residence time in the reactor is a function of
temperature, being about 200 ms at 1000 K. The
model calculations were performed using Senkin [40],
a plug-flow code that runs in conjunction with the
Chemkin library [41].
OC6H3O ϩ H !: 2C2H2 ϩ 2CO
(R8)
OC6H3O ϩ O !: C2H2 ϩ HCCO ϩ 2CO (R9)
The C6H3O3 radical formed in (R5) with two
Ͼ C"O Ͻ double bonds and a Ͼ C—O Ͻ single
bond, presumably has a low-thermal stability; we ex-
pect it to dissociate rapidly in the ring opening step,
Pyrolysis of p-Benzoquinone
C6H3O3 !: C2H2 ϩ HCCO ϩ 2CO (R10)
The key issues in the pyrolysis of parabenzoquinone
are the branching fraction between the two potential
dissociation channels,
The fate of C5H4O formed in the primary dissociation
step of parabenzoquinone is not certain. There are in-
dications that cyclopentadienone dimerizes fast [28],
but results are not conclusive. A flash-thermolysis
study at 1580 K by de Champlain and de Mayo [20]
identified cyclopentadienone, isolated as the dimer, as
the main product. However, at lower temperatures,
down to 1420 K, the yield of the dimer progressively
decreased. In a similar study performed at 1120 K,
Hageman and Wiersum [18,19] failed to detect the cy-
clopentadienone dimer and concluded that under these
conditions a different isomer of (C5H4O), most likely
a cyclopropanone, was the precursor of the major final
product, vinylacetylene.
OC6H4O EF C5H4O ϩ CO
OC6H4O EF C5H4 ϩ CO2,
(R1)
(R2)
and the fate of C5H4O formed in reaction (R1). Frank
et al. [12] proposed, based on high-temperature shock-
tube results for p-benzoquinone dissociation, that (R1)
is the dominating dissociation channel, and that cyclo-
pentadienone is thermally very stable. Hydrogenation
studies by Sakai and co-workers [21–24] at 860 K
confirm that CO is the dominating pyrolysis product,
but also significant amounts of CO2 were observed.
Sakai et al. [22,24] suggest that the CO2 is formed by
an intramolecular mechanism, i.e. (R2), since CO2 was
not observed in the pyrolysis of reactants such as phe-
nol, 1-napthol, or fluorenone with only one oxygen
molecule in their structure and since in the case of 1,4-
naphthoquinone pyrolysis, the rate of CO2 formation
was practically independent of the H2 dilution ratio.
Under the conditions of the present experiments,
we believe that the dimerization of C5H4O is too slow
to compete with thermal dissociation or reaction with
radicals. Our results support the suggestion of Emdee
et al. [4], that C5H4O dissociates to C2H2 and CO,
C5H4O !: C2H2 ϩ C2H2 ϩ CO, (R12)

PARABENZOQUINONE PYROLYSIS AND OXIDATION
687
(a)
(b)
Figure 2 (a) Comparison of experimental data and model predictions for CO and CO2 formation in the pyrolysis of p-
benzoquinone. Symbols denote experimental data and lines denote model predictions. Inlet concentrations: 52 ppm OC6H4O;
50 ppm O2; 2.4% H2O; and balance N2. Pressure is 1.03 atm. Residence time (s) ϭ 210/T(K). Solid lines: calculations made
with the mechanism of the present work (App. A). Short dashed lines: calculations made omitting reaction 2, with the reaction
rate recommended by Frank et al for (R1), and with a reaction rate of 1.0 · 1012exp(Ϫ39300/T) for reaction (R12). Long dashed
lines: calculations made with a different product channel for reaction (R12), i.e., C5H4O K CH2CHCCH ϩ CO. (b) Inlet
concentrations: 108 ppm OC6H4O; 50 ppm O2; 5.1% H2O; and balance N2. Pressure is 1.03 atm. Residence time (s) ϭ
202/T(K). The rest as in Figure 2(a).
Figure 2(a) and (b) show results from the present
study for CO and CO2 in the pyrolysis of p-benzoqui-
none. Symbols denote experimental results and lines
model calculations. The decomposition of p-benzo-
quinone, evidenced by the onset of CO formation,
starts slowly at a temperature of approximately
950 K, and becomes faster for temperatures higher
than 1100 K. Formation of CO2 is observed at tem-
peratures above 1050 K, and at 1400 K the amount of
CO2 is comparable to that of CO.
Unfortunately, the interpretation of the results of
Figure 2(a) and (b) is complicated by the presence of
other sources of CO2 than reaction (R2). Due to im-
purities in the N2 gas cylinders and to the oxygen dis-
solved in the aqueous solution of p-benzoquinone
through which the nitrogen stream is saturated, a little
amount of oxygen can be present in the experiments.
We have estimated the O2 impurity to be 50 Ϯ
10 ppm. Due to the presence of O2, an O/H radical
pool builds up, and CO and other intermediates formed
during the pyrolysis may lead to production of CO2.
The presence of H2O may further enhance OH radical
formation at high temperatures.
The presence of oxygen in trace concentrations is
difficult to avoid experimentally. However, the oxy-
gen impurity is not directly associated with the p-ben-
zoquinone feed, and, therefore, the impact of the O2
impurity will decrease as the concentration of OC6H4O
increases for similar experimental conditions. Unfor-
tunately, it was not possible to assess the significance
of the O2 impurity in this way, because experimentally
the p-benzoquinone level could not be raised suffi-
ciently. Figure 2(a) and (b) show results with 52 and
108 ppm of OC6H4O, respectively. At the higher p-
benzoquinone level (Fig. 2(b)), the CO2/OC6H4O ratio
at high temperatures is diminished, suggesting that
some of the CO2 formation can be attributed to the
presence of impurities.

688
ALZUETA, OLIVA, AND GLARBORG
largely as release of CO/CO2 from benzoquinone and
subsequent ring intermediates; little further oxidation
of carbon from the original ring structure takes place.
The recommendation of Frank et al. [12] with only
(R1) active as dissociation channel for parabenzoqui-
none and high-thermal stability for cyclopentadienone
leads to a [COx]out/[OC6H4O]in ratio of 1.0, unless sec-
ondary reactions of C5H4O produces CO/CO2. Even
with very fast rates, reactions of C5H4O with the O/H
radical pool leads to insufficient formation of CO/CO2,
due to the low radical concentrations. Another poten-
tial consumption channel of C5H4O is dimerization.
The dimerization reaction has been suggested to be an
important reaction pathway for cyclopentadienone un-
der pyrolysis and hydrogenolysis conditions [20,28].
In order to explain the observed COx formation by this
pathway, assuming fast-thermal dissociation of the di-
mer to produce two CO molecules, a dimerization rate
of about 3.0 · 1010 cm3/mole-s is required at 1300 K.
From analogy with other Diels–Alder reactions [42]
we estimate an upper limit of 1.0 · 1012exp(Ϫ7550/T)
cm3/mole-s. This rate constant results in a value of
3.0·109 cm3/mole-s at 1300 K, i.e., too slow to ac-
count for the observed results. For this reason, a high-
thermal stability of C5H4O appears to be in disagree-
ment with our experimental results. However, it
should be noted that a very fast dimerization rate for
C5H4O, 8.5·1011 cm3/mole-s at 850 K, has been sug-
gested in literature [28] and further work is needed in
order to resolve this issue.
Figure 3 CO and CO2 concentrations in the pyrolysis of
p-benzoquinone in presence of excess NH3. Symbols denote
experimental data and lines denote model predictions. Inlet
concentrations: 68 ppm OC6H4O; 3850 ppm NH3; 50 ppm
O2; 3.8% H2O; and balance N2. Pressure is 1.03 atm. Resi-
dence time (s) ϭ 240/T(K). Solid lines: calculations made
with a simplified three reaction model, including OC6H4O
and C5H4O dissociation (R1, R2, R12). Short dashed lines:
calculations made with the mechanism of the present work,
plus a subset for NH3 oxidation, and a rate constant for
C5H3 ϩ O2 (R20) of 1.0 · 1010.
In order to minimize the effect of side reactions
leading to CO2, a p-benzoquinone pyrolysis experi-
ment was carried out in the presence of a large excess
of NH3. The amines act as a radical sink, reducing the
impact of side reactions. The results are shown in Fig-
ure 3. As the temperature increases from 1300 to
1480 K, CO2 increases from 17 ppm to about
30 ppm, i.e., a significant fraction of the benzoquinone
input. Still, comparison with the results of Figure 2(a)
and (b) shows that the presence of NH3 suppresses
CO2 formation considerably at higher temperatures. At
1300 K, the [CO2]out/[OC6H4O]in ratio decreases from
55–70% to 25%, and the effect is even more pro-
nounced at higher temperatures. Furthermore, [CO ϩ
CO2]out/[OC6H4O]in drops from 2.2–2.7 to 1.8 at
1300 K and from 3.4–3.8 to 1.8 at 1400 K.
Reinterpretation of the shock-tube results of Frank
et al. [12] in terms of two dissociation channels for
parabenzoquinone, (R1) and (R2), with comparable
rates, followed by dissociation of C5H4O, suggests that
the measured rate applied to the following global re-
action,
1
1
OC6H4O !: C2H2 ϩ CO ϩ C5H4 ϩ CO2.
2
2
The rate constant for (R1) at high temperatures is, thus,
roughly half of that determined by Frank et al. Adopt-
ing the activation energy suggested by the shock-tube
results, we obtain a rate constant for reaction (R1) of
1
3.7·1011 exp(Ϫ29700/T) sϪ , which is also the value
We have investigated different possible routes to
CO and CO2 formation under the reaction conditions
of Figure 3. It is noteworthy that the sum CO ϩ CO2
remains constant within a few ppm in the whole tem-
perature range of 1300–1480 K in Figure 3. Further-
more, within the experimental uncertainty the sum cor-
responds to twice the inlet concentration of OC6H4O.
This suggests that formation of COx (x ϭ 1,2) arises
listed in Appendix A. Based on the results shown in
Figure 2(a) and (b) and matching the concentration of
CO with the mechanism of Appendix A at 1100 K,
where no or very little CO2 is formed, we obtain a rate
1
for reaction (R1) of 1.9 sϪ for the results of Figure
2(a) and of 0.9 sϪ1 for the results of Figure 2(b). From
these data, an optimum value of 1.4 sϪ1 for the reaction
rate at 1100 K is obtained. Within the uncertainty of

PARABENZOQUINONE PYROLYSIS AND OXIDATION
689
the determination, this value agrees with the k1 ex-
pression adopted in this work.
Routes to CO2 other than direct dissociation of par-
abenzoquinone (R2) were evaluated. Due to the low
radical concentrations, little oxidation of CO to CO2
is predicted. Assuming that only CO is formed from
parabenzoquinone dissociation and that cyclopenta-
dienone further leads to release of CO/CO2, either di-
rectly through dissociation and/or reaction with the
radical pool or through a fast dimerization, the level
of CO2 predicted at 1300 K for the conditions of Fig-
ure 2(a) is below 1 ppm, i.e., much lower than the
observed value of 17 ppm. Direct CO2 formation from
dissociation of the dimer appears to be unlikely, bear-
ing in mind the structure of the dimer [28].
We also performed a flash-pyrolysis experiment in
a laboratory atmospheric fluidized bed reactor at a re-
action temperature of 1085 Ϯ 10 K. A flash-pyrolysis
experiment of solid p-benzoquinone is suitable for de-
termining the CO and CO2 distribution when these
products are formed, since the solid p-benzoquinone
fed into the reactor reaches instantaneously the desired
reaction temperature. A detailed description of the ex-
perimental setup used can be found elsewhere [43]. P-
benzoquinone particles with a diameter smaller than
0.5 mm were transported into the reactor at a feeding
rate of 1–2 g/min, using nitrogen as fluidizing agent
with a flow rate of approximately 1000 ml/min. Both
CO and CO2 were detected and a CO/CO2 ratio be-
tween 6.1 and 7.5 was found in this experiment.
From this discussion, we suggest that a secondary
dissociation channel for parabenzoquinone, leading to
CO2, is active, i.e.,
Figure 4 Branching fraction k2/(k1 ϩ k2) for dissociation
of OC6H4O as a function of temperature. The values asso-
ciated with the work of Sakai et al. [21–24] and Frank et al.
[12] are based on our interpretation of their data, as ex-
plained in the text.
1.0·1010 cm3/mole-s to avoid prediction of excessive
oxidation to CO and CO2 at high temperatures. Pre-
sumably the C5H3 radical reacts mainly with the amine
pool under the conditions of Figure 3. The FTIR spec-
tra obtained indicate the presence of CN bonds formed
from such hydrocarbon/amine interactions.
In the modeling only the CO2 formation was con-
sidered. For each temperature the value of k2 was ad-
justed to obtain the best fit between the predicted value
of CO2 and the measurement. The results in Figure 4
show the branching fraction k2/(k1 ϩ k2) as function
of temperature. The values obtained with the 3-step
model can be considered upper limits for the branch-
ing fraction, since pathways to CO2 competing with
(R2) were neglected. The results obtained with the full
mechanism are also uncertain, partly due to the un-
certainties in the parabenzoquinone pyrolysis subset,
partly because the amine pyrolysis chemistry is not yet
well established [45]. However, the results obtained,
with a branching fraction increasing from 0.13 at
1085 K, and from 0.25–0.30 at 1300 K to 0.40–0.45
at 1500 K, are in reasonable agreement with the data
derived from the hydrogenation study of Sakai et al.
and the shock-tube study by Frank et al., as discussed
below.
OC6H4O EF C5H4 ϩ CO2.
(R2)
In order to estimate the rate constant for reaction (R2),
we have analyzed the results of Figure 3, the results
of the flash-pyrolysis experiment, as well as those of
the hydrogenation study by Sakai et al. [21–24] at
860 K and the shock-tube study by Frank et al. [12]
at 1550–1900 K.
The results of Figure 3 were analyzed partly in
terms of the detailed chemistry, partly in terms of a
simple three reaction model including only benzoqui-
none dissociation (reactions (R1) and (R2)) and cyclo-
pentadienone dissociation (R12). In both models, the
rate constant for (R1) was the estimate of the present
work, i.e., half the value recommended by Frank et al.
[12], while k12 was taken from Emdee et al. [4]. The
detailed modeling involved two changes compared
to the Appendix A mechanism: a NH3 reaction
subset was included [44], and the rate constant
for the C5H3 ϩ O2 reaction was lowered to
The shock-tube study by Frank et al. [12] showed
that in the 1550–1900 K range, only CO was detect-
able as pyrolysis product (among CO, H, and O), and
[CO]end/[OC6H4O]0 ϭ 1.0 Ϯ 0.1. If the assumption
that C5H4O dissociates rapidly compared to OC6H4O

690
ALZUETA, OLIVA, AND GLARBORG
at these temperatures is correct, these results indicate
that the branching fraction ␣ is 0.5 Ϯ 0.05 under the
conditions of the shock-tube experiments.
The results of Sakai et al. [21–24] must be inter-
preted with care, partly because the purity of the p-
benzoquinone was not better than 80–90% and partly
because surface reactions may have had some effect.
Furthermore, a considerable formation of high molec-
ular weight substances was detected, and these may
conceivably also have been a source for CO2. Obser-
vations for anthraquinone and phenanthrenequinone
[23] indicate that silica-glass surface reactions act to
increase the overall conversion rate and may also cat-
alyze the formation of the high molecular compounds.
However, the effect of wall reactions on p-benzoqui-
none pyrolysis or hydrogenolysis has not been inves-
tigated.
Despite these uncertainties, the rate of formation of
CO and CO2 in the hydrogenation experiments pro-
vides some information of the branching fraction of
the parabenzoquinone dissociation. Assuming that re-
actions other than the two dissociation steps (R1) and
(R2) can be neglected as sources of CO and CO2 in
the early stages of reaction, we derive a value of
␣ ϭ 0.10 Ϯ 0.05 at 863 K.
Figure 5 Reaction path diagram for p-benzoquinone py-
rolysis.
1
of 7.4·1011exp(Ϫ29790/T) sϪ for reaction (R1),
omitting reaction (R2), and assuming a high-thermal
stability for cyclopentadienone, corresponding to a
1
dissociation rate of 1.0·1012exp(Ϫ39300/T) sϪ . To
avoid significant CO production from dissociation of
C5H4O (below 10% of the total CO formation) under
the shock-tube conditions, the rate of (R12) must be
three orders of magnitude lower than estimated by
Emdee et al. The predictions conducted with these as-
sumptions imply very little formation of CO2 and a
considerably lower CO formation at temperatures
higher than 1150 K than is observed experimentally.
We also tested another product channel for C5H4O
dissociation, i.e.,
Figure 4 summarizes the results on the branching
fraction for parabenzoquinone dissociation. From the
data we have estimated a rate constant for reaction
C5H4O EF CH2CHCCH ϩ CO (R12a)
1
(R2) of 3.5·1012exp(Ϫ33700/T) sϪ .
The solid lines of Figure 2(a) and (b) denote model
calculations with the mechanism proposed in this work
(Appendix A). The agreement between model calcu-
lations and experimental results is fairly good, even
though at high temperatures CO is slightly overpre-
dicted and CO2 is somewhat underpredicted. Despite
these differences, the model succeeds in predicting the
initiation temperature of p-benzoquinone thermal de-
composition as well as the concentrations up to tem-
peratures of 1250 K.
Our interpretation of the shock-tube parabenzoqui-
none dissociation data requires that cyclopentadienone
dissociates fairly rapidly, releasing CO. The rate and
product channel for C5H4O dissociation proposed by
Emdee et al. [4], i.e., that the dissociation proceeds as
C5H4O !: C2H2 ϩ C2H2 ϩ CO, (R12)
1
with a rate of 1.0·1015exp(Ϫ39300/T) sϪ , is in agree-
ment with this interpretation and is supported by the
present results, as evidenced by the solid line model
predictions in Figure 2(a) and (b).
The short-dashed lines in Figure 2(a) and (b) denote
model predictions carried out according to the rec-
ommendation of Frank et al., i.e., with a rate constant
Figure 6 Comparison of experimental data and model pre-
dictions for CO and CO2 formation in the oxidation of p-
benzoquinone. Symbols denote experimental data and lines
denote model predictions. Inlet concentrations: 81 ppm
OC6H4O; 0.1% O2; 1.8% H2O; and balance N2. Pressure is
1.03 atm. Residence time (s) ϭ 240/T(K).

PARABENZOQUINONE PYROLYSIS AND OXIDATION
691
the estimated upper limit for the dimerization reaction
(1.0·1012exp(Ϫ7550/T) cm3/mole-s), and assuming
the dimer to dissociate thermally to C6H5C2H3 ϩ
1
2CO with a rate of 1.0·1012exp(Ϫ30000/T) sϪ . The
results roughly coincided with the results obtained
with the mechanism listed in Appendix A. Even with
the high dimerization rate and assuming irreversible
reactions, the dimerization process has little impact
under the present experimental conditions.
Figure 5 shows a reaction pathway diagram of p-
benzoquinone pyrolysis for a temperature of 1050 K,
based on the present mechanism. At low temperature,
p-benzoquinone is mainly consumed through the re-
action OC6H4O K C5H4O ϩ CO (R1), but as the tem-
perature increases, also the secondary dissociation
channel, OC6H4O K C5H4 ϩ CO2 (R2), becomes im-
portant. Cyclopentadienone is mainly consumed by
thermal decomposition, leading to acetylene and car-
bon monoxide. The C5H4 formed through reaction
(R2) is thermally stable. In the presence of radicals,
caused by the oxygen impurity, C5H4 may by reaction
with the O/H radical pool be converted to the C5H3
radical, which presumably feeds into the C2 hydrocar-
bon pool.
Figure 7 Comparison of experimental data and model pre-
dictions for CO and CO2 formation in the oxidation of p-
benzoquinone. Symbols denote experimental data and lines
denote model predictions. Inlet concentrations: 64 ppm
OC6H4O; 4.1% O2; 1.8% H2O; and balance N2. Pressure is
1.03 atm. Residence time (s) ϭ 197/T(K).
The results with (R12a) replacing (R12) are shown in
Figure 2(a) and (b) as long-dashed lines. At tempera-
tures higher than 1250 K, discrepancies for CO and
CO2 are important, indicating that the product channel
suggested by Emdee and co-workers is the most likely.
However, uncertainties in the CH2CHCCH oxidation
subset may partly be responsible for the observed dif-
ference, and more work is needed to verify the rate
and product channel for cyclopentadienone dissocia-
tion.
Oxidation of p-Benzoquinone
Figures 6 and 7 show results for CO and CO2 obtained
in p-benzoquinone oxidation experiments with two
different oxygen concentrations. The onset of reaction
as well as the oxidation of p-benzoquinone at lower
temperatures are roughly independent of the O2 con-
centration. The formation of CO is initiated at a tem-
perature of approximately 1050 K and reaches its
maximum concentration at around 1150 K. Above
1250 K all the CO is converted to CO2. Even though
the oxygen level has little effect on the CO profiles, it
The importance of the dimerization channel was
also investigated. Calculations were performed using
Figure 8 Reaction path diagram for p-benzoquinone oxidation.

692
ALZUETA, OLIVA, AND GLARBORG
is seen that an increase in oxygen concentration results
in a shift of the onset of CO2 formation towards lower
temperatures.
The model predictions, carried out with the mech-
anism in Appendix A, are in good agreement with the
experiments. Through analysis of the calculations it is
possible to identify the rate limiting processes in the
benzoquinone oxidation. Figure 8 shows a reaction
pathway diagram for the benzoquinone oxidation.
Independent of the oxygen level, OC6H4O is largely
consumed through thermal dissociation at lower tem-
peratures. At the lowest oxygen concentration, thermal
decomposition dominates at temperatures up to
1150 K. As the temperature is increased, the radical
pool is replenished and reactions between OC6H4O
and the O/H radical pool eventually become signifi-
cant. At higher temperatures, coinciding with the onset
of CO oxidation to CO2, these reactions become re-
sponsible for a large fraction of the conversion of p-
benzoquinone. At high oxygen concentrations this
shift occurs at lower temperatures, since the presence
of O2 promotes the radical formation.
The interaction of p-benzoquinone with the radical
pool proceeds mainly trough the following reactions:
Figure 9 Comparison of experimental data and model pre-
dictions for CO and CO2 formation in the oxidation of p-
benzoquinone in presence of CO. Symbols denote experi-
mental data and lines denote model predictions. Inlet con-
centrations: 40 ppm OC6H4O; 140 ppm CO; 3.8% O2; 1.6%
H2O; and balance N2. Pressure is 1.03 atm. Residence time
(s) ϭ 220/T(K). Solid lines denote calculations made with
the mechanism of the present work. Short dashed lines: cal-
culations made with a rate for reaction (R7) ten times faster
than the constant used in the present work. Long dashed
lines: calculations made with a rate for reaction (R7) ten
times slower than the constant used in the present work.
OC6H4O ϩ OH EF OC6H3O ϩ H2O (R7)
OC6H4O ϩ O !: C6H3O3 ϩ H
(R5)
(R6)
OC6H4O ϩ O !: OC6H3O ϩ OH.
Subsequently, we assume rapid ring opening of the
OC6H3O and C6H3O3 radicals by reaction with the rad-
ical pool (R8, R9) or by thermal dissociation (R10).
These reactions feed into the C2 hydrocarbon pool,
which is quickly oxidized under the conditions stud-
ied. For the O2 concentration of 4%, this reaction se-
quence becomes important at temperatures of around
1050 K, while for the low oxygen concentration it be-
comes significant above 1100 K.
In order to obtain an estimate of the reaction of
OC6H4O with OH, the effects of CO on the oxidation
of p-benzoquinone were studied. The results are
shown in Figure 9. The presence of CO enhances the
consumption of OC6H4O significantly, due to the
chain branching sequence
100 K, compared to the absence of CO. P-benzoqui-
none is completely oxidized at 1100 K under the
present reaction conditions. Calculations reveal that
the major pathways for OC6H4O oxidation in the pres-
ence of CO are similar to those when CO is absent,
even though they occur at lower temperatures.
The competition between parabenzoquinone and
carbon monoxide for OH affects the chain branching
in the system and thereby the oxidation rate. If the
OC6H4O ϩ OH reaction (R7) is fast, the CO induced
chain branching becomes less significant, and the on-
set of CO oxidation is shifted to higher temperatures.
On the other hand, if reaction (R7) is very slow the
CO oxidation will be largely unaffected by the pres-
ence of OC6H4O. From the experimental results it is,
thus, possible to obtain an estimate of k7. From the
results shown in Figure 9, we have derived a value for
CO ϩ OH EF CO2 ϩ H
H ϩ O2 EF O ϩ OH,
which contributes to replenish the radical pool. As
consequence of that, the oxidation regime of p-ben-
zoquinone is shifted to lower temperatures, about

PARABENZOQUINONE PYROLYSIS AND OXIDATION
693
Appendix A Mechanism Subset Describing the Oxidation of C3 9C6 Hydrocarbons. Rate Constants are Expressed as k ϭ
AT␤exp(ϪEA/RT). Units are cal, cm3, mole, and sec

694
ALZUETA, OLIVA, AND GLARBORG
Appendix A (Continued)

PARABENZOQUINONE PYROLYSIS AND OXIDATION
695
Appendix B Thermodynamic Properties for Selected Species. Units are cal, mole, and Kelvin
k7 of 1.5·1011 cm3/mole-s at 1075 K. Unfortunately,
the observed CO and CO2 profiles are only sensitive
to this competition in a fairly narrow temperature
range between 1000 and 1100 K. This fact, together
with uncertainties in the reaction mechanism used,
limits the accuracy of this determination. The model
predictions shown as solid lines in Figure 9 are carried
out with an extrapolated value for k7 of
1.0·106 ·T2exp(Ϫ2000/T) cm3/mole-s. The sensitivity
of the results to the rate of this reaction is also seen in
Figure 9, where calculations with a reaction rate 10
times faster (short-dashed lines) and slower (long-
dashed lines) are represented.
benzoquinone has been proposed, which provides a
good description of the experimental results obtained
in this work.
A key issue in the pyrolysis and oxidation of par-
abenzoquinone is the branching fraction between the
dissociation channels OC6H4O K C5H4O ϩ CO, and
OC6H4O K C5H4 ϩ CO2, and the fate of C5H4O. Our
results indicate that an intramolecular channel to CO2
is active, increasing in importance with temperature.
However, the data interpretation depends on assump-
tions of thermal stability and dissociation products for
C5H4O as well as the potential importance of dimeri-
zation of this species.
Under oxidizing conditions and at high tempera-
tures, the interaction between parabenzoquinone and
the radical pool becomes significant. Under the present
conditions, the most important step is the H abstraction
by reaction with OH, leading to the OC6H3O radical
with subsequent ring opening.
CONCLUSION
A combined experimental study of the pyrolysis and
oxidation of parabenzoquinone has been performed in
the temperature range 600–1500 K. A detailed reac-
tion mechanism for oxidation and pyrolysis of para-
The experimental and theoretical work presented
here constitute the first systematic study of paraben-

696
ALZUETA, OLIVA, AND GLARBORG
Symposium (Int.) on Combustion, The Combustion In-
stitute, Pittsburgh, Pennsylvania, 1994, p. 841.
15. B. Dellinger, S. L. Mazer, and R. A. Dobbs, J. Air Man-
age. Assoc. 41, 838 (1991).
16. A. M. Mebel and M. C. Lin, J. Am. Chem. Soc. 116,
9577 (1994).
zoquinone chemistry. Parabenzoquinone may be im-
portant intermediate species in the oxidation of aro-
matic compounds such as benzene, and the results
obtained extend the knowledge of parabenzoquinone
behavior under pyrolysis and oxidation conditions.
17. M. C. Lin and A. M. Mebel, J. Phys. Org. Chem. 8, 407
(1995).
18. H. J. Hageman and U. E. Wiersum, Tetrahedron Lett.
45, 4329 (1971).
19. H. J. Hageman and U.e. Wiersum, Angew. Chem. 7, 314
(1972).
20. P. de Champlain and P. de Mayo, Can. J. Chem. 50,
270 (1972).
21. T. Sakai and M. Hattori, Chem. Lett. 1103 (1975).
22. T. Sakai and M. Hattori, Chem. Lett. 1153 (1976).
23. T. Sakai, M. Hattori, and N. Yamane, J. Japan Petrol.
Inst. 23, 44 (1980).
The work is carried out as part of the combustion research
programs at the Department of Chemical and Environmental
Engineering of the University of Zaragoza and at the De-
partment of Chemical Engineering of the Technical Univer-
sity of Denmark (CHEC). The CHEC research program is
cofunded by the Danish Technical Research Council, Elsam,
Elkraft, and the Danish Ministry of Energy. The authors
wish to thank Dr. J. Ruiz for his help with the flash-pyrolysis
experiment, and Dr. P. Frank for helpful discussions con-
cerning p-benzoquinone pyrolysis.
24. T. Sakai, M. Hattori, and N. Yamane, J. Japan Petrol.
Inst. 23, 334 (1980).
25. W. D. Gra¨ber and K. J. Hu¨ttinger; Erdo¨l Kohle Erdgas
Petrochem. 33, 416 (1980).
BIBLIOGRAPHY
26. D. C. de Jongh, R. Y. van Fossen, and C. F. Bourgeois;
Tetrahedron Lett. 3, 271 (1967).
27. D. C. de Jongh and D. A. Brent, J. Org. Chem. 35, 4204
(1970).
1. J. D. Bittner and J. B. Howard, Eighteenth Symposium
(Int.) on Combustion, The Combustion Institute, Pitts-
burgh, Pennsylvania, 1981, p. 1105.
2. C. Venkat, K. Brezinsky, and I. Glassman, Nineteenth
Symposium (Int.) on Combustion, The Combustion In-
stitute, Pittsburgh, Pennsylvania, 1982, p. 143.
3. K. Brezinsky, Prog. Energy Combust. Sci., 12, 1 (1986).
4. J. L. Emdee, K. Brezinsky, and I. Glassman, J. Phys.
Chem. 96, 2151 (1992).
28. G.-J. Schraa, I. W. C. E. Arends, and P. Mulder, Chem.
Soc. Perkin Trans. 2, 189 (1994).
29. P. G. Kristensen, P. Glarborg, and K. Dam-Johansen,
Combust. Flame 107, 211 (1996).
30. P. Glarborg, M. U. Alzueta, K. Dam-Johansen, and J.
A. Miller, Combust. Flame; 115, 1 (1998).
31. J. A. Miller and C. F. Melius, Combust. Flame 91, 21
(1992).
5. R. P. Lindstedt and G. Skevis, Combust. Flame 99, 551
(1994).
6. H.-Y. Zhang and J. T. McKinnon, Combust. Sci. Techn.
107, 261 (1995).
32. J.-F. Pauwels, J. V. Volponi, and J. A. Miller, Combust.
Sci. Techn. 110, 249 (1995).
7. R. P. Lindstedt and L. Q. Maurice, Combust. Sci. Techn.
120, 119 (1996).
8. Y. Tan and P. Frank, Twenty-Sixth Symposium (Int.) on
Combustion, The Combustion Institute, Pittsburgh,
Pennsylvania, 1996, p. 677.
9. S. G. Davis, H. Wang, K. Brezinsky, and C. K. Law,
Twenty-Sixth Symposium (Int.) on Combustion, The
Combustion Institute, Pittsburgh, Pennsylvania, 1996,
p. 1025.
10. D. L. Baulch, C. J. Cobos, R. A. Cox, Th. Just, J. A.
Kerr, M. J. Pilling, J. Troe, R. W. Walker, and J. War-
natz, J. Phys. Chem. Ref. Data 21, 411 (1992).
11. D. L. Baulch, C. J. Cobos, R. A. Cox, P. Frank, G.
Hayman, Th. Just, J. A. Kerr, T. Murrells, M. J. Pilling,
J. Troe, R. W. Walker, and J. Warnatz, J. Phys. Chem.
Ref. Data 23, 847 (1994).
33. R. J. Kee, F. M. Rupley, and J. A. Miller, The Chemkin
Thermodynamic Data Base, Sandia Report SAND87-
8215, Sandia National Laboratories, Livermore, Cali-
fornia, 1991.
34. A. Burcat and B. McBride, 1995 Ideal Gas Thermody-
namic Data for Combustion and Air-Pollution Use;
Aerospace Engineering Report TAE 732, Technian, Is-
rael. Institute of Technology, 1995 and updates.
35. J. A. Manion and R. Louw, J. Phys. Chem. 94, 4127
(1990).
36. I. W. C. E. Arends, R. Louw, and P. Mulder, J. Phys.
Chem. 97, 7914 (1993).
37. P. N. Bajaj and A. Fontijn, Combust. Flame 105, 239
(1996).
38. R. Atkinson, Int. J. Chem. Kinet. 19, 799 (1987).
39. R. Atkinson, S. M. Aschmann, J. Arey, and B. Zielin-
ska, Atmospheric Environment 23, 2679 (1989).
40. A. E. Lutz, R. J. Kee, and J. A. Miller, SENKIN: A
Fortran Program for Predicting Homogenous Gas
Phase Chemical Kinetics with Sensitivity Analysis, San-
dia Report SAND87-8248, Sandia National Laborato-
ries, Livermore, California, 1987.
12. P. Frank, J. Herzler, Th. Just, and C. Wahl, Twenty-Fifth
Symposium (Int.) on Combustion The Combustion In-
stitute, Pittsburgh, Pennsylvania, 1994, p. 833.
13. S. S. Kumaran and J. V. Michael, Proceedings of the
21st Int. Symp. on Shock Waves, Great Keppel Island,
Australia, 1997.
14. R. Buth, K. Hoyermann, and J. Seeba, Twenty-Fifth
41. R. J. Lutz, F. M. Rupley, and J. A. Miller, CHEMKIN-

PARABENZOQUINONE PYROLYSIS AND OXIDATION
697
II: A Fortran Chemical Kinetics Package for the Anal-
ysis of Gas-Phase Chemical Kinetics, Sandia Report
SAND89-8009, Sandia National Laboratories, Liver-
more, California, 1989.
44. J. A. Miller and P. Glarborg, in “Gas-Phase Chemical
Reaction Systems; Experiments and Modeling 100
Years after Max Bodenstein”, (Eds. J. Wolfrum, H.-R.
Volpp, R. Rannacher, J. Warnatz), Springer Series in
Chemical Physics 61, 318 (1996).
45. A. M. Dean and J. W. Bozzelli, “Combustion Chemistry
of Nitrogen”, Combustion Chemistry II, (ed. W. C. Gar-
diner, Jr.), Springer Verlag, in press (1997).
42. J. Sauer and R. Sustmann, Angew. Chem. 92, 773
(1980).
43. J. Arauzo, D. Radlein, J. Piskorz, and D. S. Scott, En-
ergy Fuels 8, 1192 (1994).