Role of phenoxy radicals in PCDD/F formation

Article abstract of DOI:10.1002/kin.10083

In this work, the role of phenoxy radicals in polychlorinated dibenzo-p-dioxins and polychlorinated dibenzofurans (PCDD/F) formation was investigated by studying the slow oxidation of 2-chlorophenol (2-CP) and 2-chloroanisole (2-CA) at a gas-phase concentration of 4 ppm (～2.1 × 10⁴ μg/m³) over a temperature range of 400-800°C. Residence times were maintained at 2.0 ± 0.10 s. PCDD/F reaction products were dibenzofuran, dibenzo-p-dioxin, 4-chlorodibenzofuran, 1-chlorodibenzo-p-dioxin, 4,6-dichlorodibenzofuran, and 1,6-dichlorodibenzo-p-dioxin (1,6-DCDD). Major products observed in these experiments were 2,6-dichlorophenol, 3-phenyl-2-propenal, 1-indanone, 1,3-isobenzofurandione, and 3-phenyl-2-propenoyl chloride. The 2-CP and 2-CA experiments, along with the variable concentration 2-CA experiments, showed that the concentration of radicals present in the oxidation system has a significant effect on the PCDD/F product distribution and ultimately the PCDD/PCDF ratio. Also, the observation of dichlorinated phenoxy phenol and dichlorinated dihydroxybiphenyl, the proposed intermediate species in the radical-radical mechanism, suggests that radical-radical mechanism dominates gas-phase PCDD/F formation. This information will be helpful in constructing a detailed kinetic mechanism of PCDD/F formation/destruction in combustor postcombustion zone.

Full text of DOI:10.1002/kin.10083

Role of Phenoxy Radicals
in PCDD/F Formation
SUKH SIDHU, PHIL EDWARDS
Environmental Science and Engineering, 300 College Park, University of Dayton, Dayton, Ohio 45469-0132
Received 11 February 2002; accepted 6 June 2002
DOI 10.1002/kin.10083
ABSTRACT: In this work, the role of phenoxy radicals in polychlorinated dibenzo-p-dioxins
and polychlorinated dibenzofurans (PCDD/F) formation was investigated by studying the slow
oxidation of 2-chlorophenol (2-CP) and 2-chloroanisole (2-CA) at a gas-phase concentra-
tion of 4 ppm ( 2.1 104 µg/m3) over a temperature range of 400–800 C. Residence times
were maintained at 2.0 0.10 s. PCDD/F reaction products were dibenzofuran, dibenzo-p-
dioxin, 4-chlorodibenzofuran, 1-chlorodibenzo-p-dioxin, 4,6-dichlorodibenzofuran, and 1,6-
dichlorodibenzo-p-dioxin (1,6-DCDD). Major products observed in these experiments were
2,6-dichlorophenol, 3-phenyl-2-propenal, 1-indanone, 1,3-isobenzofurandione, and 3-phenyl-
2-propenoyl chloride. The 2-CP and 2-CA experiments, along with the variable concentration
2-CA experiments, showed that the concentration of radicals present in the oxidation system
has a significant effect on the PCDD/F product distribution and ultimately the PCDD/PCDF ratio.
Also, the observation of dichlorinated phenoxy phenol and dichlorinated dihydroxybiphenyl,
the proposed intermediate species in the radical–radical mechanism, suggests that radical–
radical mechanism dominates gas-phase PCDD/F formation. This information will be helpful
in constructing a detailed kinetic mechanism of PCDD/F formation/destruction in combustor
C
postcombustion zone. 2002 Wiley Periodicals, Inc. Int J Chem Kinet 34: 531–541, 2002
formed plays an important role in the formation of
soot and polyaromatic hydrocarbons like naphthalene
[6,7]. A second path for consumption of phenoxy radi-
cals is dimerization, which leads to the formation of
dibenzofurans (DF), dibenzo-p-dioxins (DD), dihy-
droxybiphenyls, and phenoxyphenols, with dibenzo-
furan being the major product under postcombustion
conditions [8]. Dimerization of chlorinated phenoxy
radicals can thus lead to the formation of polychlori-
nated dibenzo-p-dioxins and polychlorinated dibenzo-
furans (PCDD/F), a class of 210 species that are known
to cause adverse health effects in humans and animals
[9,10]. Similarly, chlorinated cyclopentadienyl radicals
formedfromthedecompositionofchlorinatedphenoxy
radicals can form polychlorinated napthalenes, which
are also known to cause adverse health effects in hu-
mans and animals [11].
INTRODUCTION
The phenoxy radical is considered a key intermediate
species in the oxidation of aromatic species that are
an important fraction of automotive and aviation fu-
els [1–3]. The formation and destruction of phenoxy
radicals are important steps that control the overall
conversion of the aromatic fuel fraction into lower
molecular weight aliphatic fragments. Reactions of
phenoxy radicals are key reactions for the formation
of many pollutants, namely soot, dioxins, and dioxin-
like compounds. A phenoxy radical can thermally de-
compose and eliminate CO and form a cyclopentadi-
enyl radical [4,5]. The cyclopentadienyl radical thus
Correspondence to: Sukh Sidhu; e-mail: sidhu@udri.udayton.
edu.
The structure of phenoxy radical is a hybrid of
one oxygen-centered (1) and two carbon-centered (2
and 3) radicals. These resonance structures provide
Contract grant sponsor: USEPA Northeast Hazardous Substance
Research Center and NSF.
c
2002 Wiley Periodicals, Inc.

532
SIDHU AND EDWARDS
16 kcal/mol of resonance stabilization energy to the
phenoxy radical [12].
elsewhere [17]. Briefly, the reactor system consists of
four integrated units: (1) A control console for precise
adjustment of temperature, pressure, residence time,
and the respective gas flows; (2) A thermal reactor com-
partment containing a high temperature Lindberg fur-
nace housed within a gas chromatograph (HP 5890) to
allow precise sample introduction into and out of the
quartz reactor; (3) A gas chromatographic oven (HP
5890) containing a capillary column (J & W Scientific
DB-5, 30 m, 0.25 mm ID, 0.25 ␮m film thickness)
for separation of organic reaction products; and (4) A
mass spectrometer (HP 5970B MSD) for product iden-
tification and quantification. For this study, the reactor
system was fitted with a 1.0 cm ID, 27 cm3 quartz gas-
phase reactor. The residence time distribution in the
reactor simulates plug-flow behavior. Previous studies
in which the reactor diameter was varied from 0.1 to
1.2 cm have shown that surface reactions of chlorinated
hydrocarbons are insignificant for a reactor diameter
1.0 cm [18]. Also, the high temperature region was
strictly confined to the wide bore section of the reactor
to prohibit end wall effects. Therefore, the system is
completely homogeneous. The residence time of the
reaction gas in the reactor is calculated by dividing
the reactor volume by the volumetric flow rate of the
reaction gas, corrected for temperature and pressure.
Residence times were maintained at 2.0 0.10 s over
a temperature range of 400–800 C. At each reaction
temperature, the flow rate was adjusted to maintain the
2 s residence time. Each experiment was run for 60
min. Reactions were conducted in an atmosphere of
4% oxygen in helium.
O
O
O
.
H
.
.
(3)
H
(1)
(2)
Because of this resonance stabilization energy, phe-
noxy radicals are usually present at higher concen-
trations than nonstabilized radicals like phenyl. The
relative high concentration and stability enables phe-
noxy radicals to participate in many reactions that are
key steps in pollutant formation in combustion sys-
tems [13]. Understanding these phenoxy reactions will
benefit researchers who are conducting theoretical or
experimental investigations of fuel ignition, soot, and
pollutant formation.
It is an impossible task to study reactions of all
phenoxy radicals with and without varying degrees of
chlorination and other substitutions. An acceptable ap-
proach is to select a particular type of phenoxy radical
based on its importance in pollutant formation and then
study its reactions in detail. In this paper, we present the
reactions of monochlorinated phenoxy radicals under
postcombustion zone conditions. The chlorinated phe-
noxy radical in this study was generated from slow ox-
idation of 2-chlorophenol (2-CP) and 2-chloroanisole
(2-CA). Both 2-CP and 2-CA were chosen as phenoxy
precursors because of their presence in waste combus-
tion effluents and also because only one of their ortho
sites is chlorinated [14]. Theoretical and experimen-
tal studies of nonchlorinated phenoxy radicals have
shown that phenoxy radicals exhibit a high probability
for ortho–ortho C C coupling, which then leads to the
formation of DF and DD [15,16]. Since understanding
the role of phenoxy radicals in the formation of chlo-
rinated DF and DD is the main objective of this study,
we wanted to study a phenoxy radical possessing one
chlorinated and one nonchlorinated ortho site.
The phenoxy precursors, namely 2-CP and 2-CA,
were purchased from Aldrich Chemical and used with-
out any further treatment. To achieve low precursor
concentration and stay within detection limits, the gas-
phase precursor concentration in the reaction gas was
maintained at 4 1 ppm (2.1 104 ␮g/m3 at RTP).
A gas-phase sample of a precursor in helium was pre-
pared in a 0.52 l sample bulb for each run. During the
runs, the sample was injected into the reactor at a con-
stant rate by using a syringe pump (Sage Instruments
Model 341A). For each run, a syringe pump flow rate
setting was selected and the required concentration of
precursor in the sample bulb was calculated to maintain
a precursor concentration in the reaction gas of 4 ppm.
The equivalence ratios for 2-CP and 2-CA were
4
EXPERIMENTAL
6.76 10 and 8.26 10 4, respectively (fuel lean
conditions).
All experiments were performed using a high-
temperature flow reactor coupled with an in-line gas
chromatograph mass spectrometer (GC–MS) system
designed to simulate the reaction conditions in a com-
bustor postflame zone. The setup is described in detail
The reaction products and unreacted precursor were
trapped with a composite adsorptive trap consisting of
0.12 g of TENAX-TAR and 0.24 g of CARBOTRAP R.
After each run, the trap was desorbed at 300 C for
20 min, and the reaction products were trapped at

ROLE OF PHENOXY RADICALS IN PCDD/F FORMATION
533
30 C on a J&W Scientific DB-5 column (30 m, 0.25
mm ID, 0.25 ␮m film thickness). Products were an-
alyzed using GC–MS with a temperature program-
ming of 30 C (1 min hold) to 300 C (10 min hold)
at a rate of 15 C/min. Selected ion monitoring (SIM)
was used to identify and quantitate reaction products.
For PCDD/F analysis, the molecular weight (M) and
molecular weight +2 (M + 2) ions, along with one
other prominent ion, were monitored. For other prod-
ucts, the molecular weight ion and two other prominent
ions in the compound’s spectrum were monitored. In
SIM mode, we can detect up to 0.1 ng (on column)
of any PCDD/F species, which means that we should
observe any PCDD/F product that has a yield better
than 0.00003 mol%. Quantitation was performed using
external standard calibration curves for the respective
product. A few minor reaction products were quanti-
tated using standard calibration for chemically similar
compounds.
at 2.0 0.10 s. Major products observed in these
experiments were 2,6-dichlorophenol (2,6-DCP), 3-
phenyl-2-propenal, 1-indanone, 1,3-isobenzofurandi-
one, and 3-phenyl-2-propenoyl chloride. PCDD/F
products observed were DF, DD, 4-chlorodibenzofuran
(4-CDF), 1-chlorodibenzo-p-dioxin (1-CDD), 4,6-
dichlorodibenzofuran (4,6-DCDF), and 1,6-dichloro-
dibenzo-p-dioxin (1,6-DCDD).
The product profiles of non-PCDD/F products from
2-CP oxidation experiments are shown in Fig. 1. The
gas-phase 2-CP is fairly stable, with slight decomposi-
tion beginning at 500 C but rapid decomposition only
after 650 C. The major non-PCDD/F product is 3-
phenyl-2-propenal, with a mole percent yield of 0.32%
at 650 C. Other products formed have mole percent
yields between 0.0001 and 0.1%. Maximum yields
of non-PCDD/F products occurred between 650 and
700 C. The PCDD/F product profiles from 2-CP oxida-
tion are shown in Fig. 2. The major PCDD/F product is
1-CDD, with a maximum mole percent yield of 0.022%
at 650 C. Other PCDD/F products are produced in
yields of 0.0001–0.003%, with the maximum yield of
PCDD/F products occurring between 600 and 650 C.
All data points in these figures are an average of repli-
cate experiments ( 10%). The observation of 1-CDD
as a major PCDD/F product from 2-CP oxidation is not
consistent with the results of 2-CP oxidation and pyrol-
ysis studies conducted by the Louw, Mulholland, and
RESULT AND DISCUSSION
In this work, the role of phenoxy radicals in PCDD/F
formation was investigated by studying the slow ox-
idation of 2-CP and 2-CA at a gas-phase concentra-
tion of 4 ppm ( 2.1 104 ␮g/m3) over a temperature
range of 400–800 C. Residence times were maintained
1
0.1
0.01
0.001
0.0001
0.00001
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1050
Temperature (°C)
Figure 1 Thermal oxidation profiles for 2-CP and non-PCDD/F products; product yields in mole percent. [2-CP]0 = 2.1
104 ␮g/m3 (RTP), tr = 2 s in 4% O2/He.

534
SIDHU AND EDWARDS
1
0.1
0.01
DF
DD
4-CDF
1-CDD
4,6-DCDF
1,6-DCDD
2,6'-DC2PP
2-CP/100
0.001
0.0001
0.00001
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1050
Temperature (°C )
Figure 2 Thermal oxidation profiles for 2-CP and PCDD/F products; product yields in mole percent. [2-CP]0 = 2.1
104 ␮g/m3 (RTP), tr = 2 s in 4% O2/He.
Hagenmaier groups, in which they found 4,6-DCDF
to be the major PCDD/F product [19–23]. However,
all the previous 2-CP studies were conducted at much
higher 2-CP inlet concentrations (>1000 vs. 4 ppm)
and this difference in inlet concentration could be re-
sponsible for different PCDD/F product distribution.
This is discussed in detail later.
4,6-dichlorodibenzofuran, with a maximum mole per-
cent yield of 0.001% at 700 C (the major product from
2-CP was 1-CDD at 0.022% at 650 C). Other PCDD/F
products are produced in yields of 0.00002–0.0005%,
with the maximum yield of PCDD/F products occur-
ring from 500–700 C. The observation of 4,6-DCDF
as a major PCDD/F product from 2-CA oxidation is
consistent with the results of Louw, Mulholland, and
Hagenmaier [19–23].
This difference in product distribution between 2-
CA and 2-CP experiments can be attributed to weaker
O CH3 bond strength of 2-CA ( 64 kcal/mol) com-
pared to O H bond strength in 2-CP ( 88.7 kcal/mol)
[24,25]. Because of this weaker O CH3 bond strength,
2-CA generates phenoxy radicals and methyl radicals
at lower temperatures than from 2-CP. In 2-CP oxi-
dation experiments, the destruction of 2-CP occurred
initially via thermal decomposition,
The product profiles of non-PCDD/F products for
2-chloranisole oxidation are shown in Fig. 3. The gas-
phase 2-CA is less stable than 2-CP, with decompo-
sition beginning at 500 C (similar to 2-CP), but with
very rapid decomposition thereafter. 2-CP is the ma-
jor non-PCDD/F products observed in the 2-CA ex-
periments. Other non-PCDD/F observed in the 2-CA
experiments are similar to 2-CP experiments but with
lower mole percent yields. Yields of non-PCDD/F
products from 2-CA experiments range from 0.002 to
9% at 600 C. Maximum yields of non-PCDD/F prod-
ucts from 2-CA and 2-CP oxidation occurs around
650 C, but product formation begins at lower tem-
peratures (500 C) for 2-CA than 2-CP. The PCDD/F
product profiles from 2-CA oxidation are shown in
Fig. 4. Similar PCDD/F products are observed as in the
2-CP experiments, but yields from 2-CA oxidation
are lower. The major PCDD/F product from 2-CA is
C6H5OH → C 6 H 5 O + H
but the recombination rate of phenoxy radical and H is
1
very high (2.5 1014 cm3 mol
s
1) and this is the
reasonthatrapiddestructionof2-CPisobservedonlyat
higher temperatures (>650 C) [26]. On the other hand,

ROLE OF PHENOXY RADICALS IN PCDD/F FORMATION
535
1
DCP
0.1
3-Ph-2-propenal
1-Indanone
1,3-Isobenzo-furandione
3-Ph-2-propenoyl chloride
0.01
2-CP/10
2-CA/100
0.001
0.0001
0.00001
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1050
Temperature (°C)
Figure 3 Thermal oxidation profiles for 2-CA and non-PCDD/F products; product yields in mole percent. [2-CA]0 = 2.1
104 ␮g/m3 (RTP), tr = 2 s in 4% O2/He.
1
0.1
DF
DD
4-CDF
0.01
1-CDD
4,6-DCDF
1,6-DCDD
2-CA/100
0.001
0.0001
0.00001
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1050
Temperature (°C )
Figure 4 Thermal oxidation profiles for 2-CA and PCDD/F products; product yields in mole percent. [2-CP]0 = 2.1
104 ␮g/m3 (RTP), tr = 2 s in 4% O2/He.

536
SIDHU AND EDWARDS
the 2-CA thermal decomposition profile shows a rapid
decomposition of 2-CA because the initial decompo-
sition products of 2-CA, phenoxy and methyl radical
recombine but mostly form cresol and not anisole. This
is why product formation from 2-CA occurs over a
broader temperature range than from 2-CP. The methyl
radicals formed from the initial dissociation of 2-CA
create a radical pool that is not present in chlorophenol,
in which many reactions take place such as
at 600 C. From Fig. 5 it can also be observed that
dioxin formation is reduced more than furan formation
such that the major PCDD/F product from 2-CA is 4,6-
DCDF, where it was 1-CDD in the 2-CP experiments,
and the furans in general become dominant PCDD/F
products in 2-CA oxidation.
PCDD/F products observed in both 2-CP and 2-CA
experiments are probably formed via radical–
molecule or radical–radical reactions between the
2-chlorophenoxy radicals and the 2-CP/2-CA molecu-
les. Thermochemical and kinetic analysis has previ-
ously concluded that the molecule–molecule reaction
is infeasible in 2-CP/2-CA oxidation under the reaction
conditions of this study [27]. In the radical–molecule
reaction (see Fig. 6), the initial step is the displacement
of the molecule’s chlorine by the phenoxy radical. The
resulting compound can then either abstract HCl to di-
rectly form DD or can lose a hydrogen atom by OH
abstraction to form a phenoxy ether intermediate (I1 in
Fig. 6). The intermediate then undergoes Smiles rear-
rangement to form either DD through intraring elimi-
nation of the second chlorine atom or 1-CDD through
intraring elimination of the hydrogen atom. Smiles re-
arrangement is a well-documented occurrence in these
typesofreactions[8,21–23,28]. Alternatively, indiben-
zofuran formation the initial step is instead the dis-
placement ofOH bythe phenoxyradical. The resulting
CH3 + O2 → CH2O + OH ;
CH3 + O2 → CH3O + O
O + CH3
→ CH2O + H ;
CH3O + M → CH2O + H + M, etc.
H radical thus formed can also combine with phenoxy
radical to form 2-CP, which is the major product ob-
served in 2-CA experiments. A weaker initial dissoci-
ation bond combined with a lack of recombination to
form parent compounds and reactive methyl radicals
create a very reactive system, in which it is more diffi-
cult to form large molecular growth molecules such as
PCDD/F. This can be clearly observed in Fig. 5, which
compares product yields (mole percent yields based
on precursor input) from 2-CP and 2-CA oxidation
Figure 5 Yield comparison for 2-CP and 2-CA. T = 600 C, [2-CP, 2-CA]0 2.1 104 ␮g/m3, tr = 2 s in 4% O2/He.

ROLE OF PHENOXY RADICALS IN PCDD/F FORMATION
537
OH
O
Cl
Cl
+
-Cl
-OH
Cl
O
O
OH
Cl
Cl
H
-HCl
-H
+OH
-HCl
+H
-Cl
Cl
O
O
I1
I2
O
Cl
H
-Cl
-Cl
-H
Smiles
-H
Smiles
Cl
Cl
O
O
O
O
O
O
DF
4-CDF
1-CDD
DD
(Major)
(Minor)
Figure 6 Radical–molecule pathway to PCDD/F formation from oxidation of 2-CP.
compound can then either abstract HCl to directly form
4-CDF or lose hydrogen with the help of OH to form
a phenyl ether intermediate (I2 in Fig. 6). The inter-
mediate then undergoes Smiles rearrangement to form
either DF through intraring elimination of the second
chlorine atom or 4-CDF through intraring elimination
of hydrogen. However, the PCDD/F product distribu-
tion obtained from radical–molecule mechanism is not
consistent with the results of this study, as it gives DD
and 4-CDF as main PCDD/F products, whereas the ex-
perimental results show 1-CDD and 4,6-DCDF as main
PCDD/F products. Another problem with the radical–
molecule mechanism is that it requires chlorine and
hydroxyl displacement as first steps, and these steps
are not energetically favored under the conditions of
the present study [8].
Exclusion of the radical–molecule mechanism
leaves us with a radical–radical mechanism to ex-
plain PCDD/F formation in the present study. Phenoxy
or chlorophenoxy radicals exist as oxygen-centered
radicals (“eno” form) or as carbon-centered radicals
(“keto” form). Theoretical and experimental studies
have shown that keto form is the more stable form of the
phenoxy radical [29,30]. In the radical–radical reaction
to PCDD, keto and eno react to form an ether adduct,
which is then followed by hydrogen tautomerism.
The resulting compound, dichlorophenoxyphenol (I3
in Fig. 7), can then either abstract HCl to directly
form 1-CDD or lose a hydrogen atom by OH abstrac-
tion to form a chlorophenyl–chlorophenoxy ether in-
termediate. This intermediate then undergoes Smiles
rearrangement to form either 1-CDD through intrar-
ing elimination of the second chlorine atom or form a
DCDD (1,6- or 1,9-) through intraring elimination of
hydrogen. The rotation of the five-membered Smiles
rearrangement ring at the apex results in two possi-
ble DCDD isomers. However, we observed only one
DCDD isomer (1,6-DCDD); this perhaps could be at-
tributed to higher stability of 1,6-DCDD than 1,9-
DCDD [31]. The intermediate I3 (2,60-dichloro-2-
phenoxyphenol; listed as 2,60-DC2PP in Fig. 2) has
been consistently observed in the 2-CP experiments
but not in 2-CA experiments. 2,60-DC2PP was iden-
tified based on a good match of product spectrum
with NIST mass spectral library. This identification
was later verified and quantified using a synthesized
standard.
In the radical–radical reaction to PCDF (see Fig. 8),
chlorophenoxy radicals in the keto form combine at
the ortho-carbon site to form dichlorinated biphe-
noxy intermediate (with two ketone groups). This
intermediate then undergoes hydrogen tautomerism
to form dichloro-2,20-dihydroxybiphenyl. Dichloro-
2,20-dihydroxybiphenyl can lose a hydrogen atom via

538
SIDHU AND EDWARDS
O
O
Cl
Cl
H
eno-
keto-
O
O
Cl
Cl
H
+
O
H
O
Cl
Cl
OH
O
O
I3
Cl
-OR-
Cl
HO
Cl
Cl
H
+O
O
-H
O
Cl
Cl
Smiles
-HCl
-H
-Cl
Cl
Cl
O
O
O
O
Cl
1-CDD
(Major)
1,6-DCDD
(Minor)
Figure 7 Radical–radical pathway to PCDD formation from oxidation of 2-CP.
abstraction and then eliminate a hydroxyl radical to
form 4,6-dichlorodibenzofuran or eliminate water di-
rectly to form 4,6-dichlorodibenzofuran. Similarly, a
chlorophenoxy radical can react with a nonchlorinated
phenoxy radical to form 4-CDF, and two nonchlori-
nated phenoxy radicals can form DF. Dichloro-2,20-
dihydroxybiphenyl was observed in both the 2-CP
and 2-CA experiments with yields very close to
detection limit (0.00003 mol%). We were able to
identify dichloro-2,20-dihydroxybiphenyl but were not
able to quantify it within our acceptable experi-
mental scatter limits ( 10%). This is the reason
that dichloro-2,20-dihydroxybiphenyl is not shown in
Figs. 2 and 4.
From the earlier discussion, it is evident that the
radical–radical mechanism is consistent with the ob-
served intermediates and PCDD/F product distribu-
tion. Radical–radical PCDD/F formation mechanism
can also explain why 1-CDD was the major PCDD/F
product in our 2-CP experiments, whereas 4,6-DCDF
was the major product in the 2-CA experiments and 2-
CP oxidation studies conducted by Louw, Mulholland,
and Hagenmaier groups [19–23]. As shown in Fig. 7,
dioxin formation occurs via combination of eno and
keto forms of phenoxy radicals and thus relies upon
the eno–keto radical equilibrium. As the radical pool
increases because of large initial concentrations of phe-
noxy precursor (33 and 684 ppm 2-CP experiments)

ROLE OF PHENOXY RADICALS IN PCDD/F FORMATION
539
O
To further test our hypothesis that radical concentra-
tion can impact the PCDD/PCDF ratio and to better un-
derstand the role of radical concentration on PCDD/F
formation, experiments were conducted for 2-CP ox-
idation at a temperature of 600 C and variable inlet
2-CP precursor concentrations of 4, 33, and 684 ppm
(2.1 104, 1.73 105, and 3.57 106 ng/m3 at RTP).
As postulated, similar reaction products were observed
at all three concentrations, but the yield distribution of
PCDD/F products did change. The dominant PCDD/F
product in the 4 ppm 2-CP oxidation experiments was
found to be 1-CDD, while the dominant PCDD/F prod-
uct in the 33 and 684 ppm experiments was found to
be 4,6-DCDF. This reaffirms the previous conclusion
that when the concentration is increased, the increased
number of radicals available hinders the formation of
PCDD more than the formation of PCDF, which is not
dependent on chemistry involving the less stable eno
form of phenoxy radicals. However, the reduction of
dioxin products with an increase in 2-CP concentra-
tion is much less noticeable than when a new species
such as hexane or anisole is introduced. PCDD/F prod-
uct distributions for 4, 33, and 684 ppm experiments are
shown in Fig. 9. Figure 9 also shows that no 1,6-DCDD
was observed in 33 ppm experiments. Mechanistically
it is not possible for 1,6-DCDD to not form in 33 ppm
2-CP experiments. The most plausible explanation is
that 1,6-DCDD was formed, but was just below the
detection limit.
O
Cl
Cl
+
O
O
Cl
Cl
OH
HO
Cl
Cl
-H O
2
Cl
Cl
O
4,6-DCDF
Figure 8 Radical–radicalpathwaytoPCDFformationfrom
oxidation of 2-CP.
The 2-CP and 2-CA experiments, along with the
variable concentration 2-CA experiments, showed that
the concentration of radicals present in the oxida-
tion system has a significant effect on the PCDD/F
product distribution and ultimately the PCDD/PCDF
ratio. Also, the observation of dichlorinated phe-
noxy phenol and dichlorinated dihydroxybiphenyl, the
proposed intermediate species in the radical–radical
mechanism, suggests that radical–radical mechanism
dominates gas-phase PCDD/F formation in the com-
bustor postcombustion zone. The only scenario in
which radical–molecule PCDD/F formation mecha-
nism might be dominant would be the combustion of
phenol whose ortho and para sites are chlorinated (e.g.,
2,4,6-trichlorophenol or pentachlorophenol) such that
the radical–radical combination at ortho sites would
require chlorine tautomerism, a step that is thermody-
namically not feasible under the postcombustion zone
conditions. However, in most practical combustion
situations, other hydrocarbon species will be present
along with 2,4,6-trichlorophenol or pentachlorophenol
and the chlorine at the ortho or para sites of these
phenols could be easily displaced, opening reaction
routes for radical–radical mechanism. Thermochemi-
calandkineticmodelingofthisdataanddatafromother
or because of weak initial dissociation bond of phe-
noxy precursor (2-CA), the reactions of phenoxy radi-
cal with other nonphenoxy radicals also increases. This
depletes the total phenoxy radical concentration and
further lowers the chances of phenoxy–phenoxy rad-
ical recombination to form PCDD/Fs. The concentra-
tion of the eno form of the phenoxy radical decreases
more than that of the keto form, as keto is the more
stable form of the phenoxy radical [29,30]. Since furan
formation involves only the more stable keto form of
the radical, the decrease in furan yields is less than the
observed decrease in dioxins. This higher decrease in
dioxin yields when compared to furan yields is clearly
shown Fig. 5, which compares product yields from
2-CP and 2-CA experiments at 600 C. This is con-
sistent with results from our previous study in which
oxidation of 2,4,6-trichlorophenol (2,4,6-TCP) with
and without hexane was investigated. The introduc-
tion of the extra radicals produced from the decom-
position of hexane caused a decrease in the yields
of PCDD products, while PCDF yields did not show
the same effect, and were even observed to increase
[32].

540
SIDHU AND EDWARDS
Figure 9 Yield comparison for PCDD/F products at various initial 2-CP concentrations. T = 600 C, tr = 2 s in 4% O2/He.
chlorophenol studies would be useful in clarifying as-
pects of the various pathways to PCDD/F formation
from precursors.
9. Schwetz, B. A.; Norris, J. M.; Sparschu, G. L.; Rowe, V.
K.; Gehring, P. J.; Emerson, J. L.; Gebrig, C. G. Environ
Health Perspect 1973, 5, 87.
10. Rozman, K.; Roth, W. L.; Stahl, B. U.; Greim, H.; Doull,
J. Toxicology 1993, 77, 39.
11. Williams, D. T.; Kennedy, B.; LeBel, G. L. Chemosphere
1993, 27, 795.
12. Mallard, W. G. (Ed.); NIST Chemistry Web Book; NIST
Standard Reference, Database 69, November 1998.
13. Bozzelli, J. W.; Desai, V.; Ritter, E. R.; Dean, A. M. ACS
Div Fuel Chem 1991, 36(4), 1494.
14. Dellinger, B.; Taylor, P. H.; Tirey, D. A. Minimiza-
tion and Control of Hazardous Combustion By-products;
RiskReductionLaboratory, Cincinnati, Ohio, 1991;EPA
600-52-039.
15. Tuppurainen, K.; Halonen, I.; Tarhanen, J.; Ruuskanen,
J. J Mol Struct (Theochem) 1994, 312, 139.
16. Armstrong, D. R.; Cameron, C.; Nonhebel, D. C.;
Perkins, P. G. J Chem Soc, Perkin Trans 2 1983, 563.
17. Rubey, W. A.; Grant, R. A. Rev Sci Instrum 1988, 59,
265.
The authors thank Ms. Joy Klosterman for her help in prepar-
ing this manuscript.
BIBLIOGRAPHY
1. Bittker, D. A. Combust Sci Technol 1991, 79, 49.
2. Lindstedt, R. P.; Skevis, G. Combust Flame 1994, 99,
551.
3. Emdee, J. L.; Brezinsky, K.; Glassman, I. J Phys Chem
1992, 96, 2151.
4. Lin, C.-Y.; Lin, M. C. Int J Chem Kinet 1985, 17, 1025.
5. Frank, P.; Herzler, J.; Just, T.; Wahl, C. In Twenty-Fifth
Symposium (International) on Combustion; The Com-
bustion Institute: Pittsburgh, PA, 1994; p. 833.
6. McEnally, C. S.; Pfefferle, L. D. Combust Sci Technol
1998, 131, 323.
7. Melius, C. F.; Colvin, M. E.; Marinov, N. M.; Pitz, W.
J.; Senkan, S. M. In Twenty-Sixth Symposium (Interna-
tional) on Combustion; The Combustion Institute: Pitts-
burgh, PA, 1996; p. 685.
18. Taylor, P. H.; Tirey, D. A.; Rubey, W. A.; Dellinger, B.
Combust Sci Technol 1994, 101, 75.
19. Yang, Y.; Mulholland, J. A.; Akki, U. In Twenty-Seventh
Symposium (International) on Combustion; The Com-
bustion Institute: Pittsburgh, PA, 1998; p. 1761.
20. Akki, U.; Mulholland, J. A. Organohalogen Compd
1997, 31, 475.
8. Wiater, I.; Born, J. G. P.; Louw, R. Eur J Org Chem 2000,
921.
21. Born, J. G. P.; Louw, R.; Mulder, P. Chemosphere 1989,
19, 401.

ROLE OF PHENOXY RADICALS IN PCDD/F FORMATION
541
22. Born, J. G. P. PhD. dissertation, Leiden University,
Netherlands, 1992.
27. Sidhu, S.; Maqsud, L.; Dellinger, B.; Mascolo, G. Com-
bust Flame 1995, 100, 11.
23. Weber, R.; Hagenmaier, H. Chemosphere 1999, 38,
529.
28. Gray, A. P.; Cepa, S. P.; Cantrell, J. S. Tetrahedron Lett
1975, 33, 2875.
24. Pecullan, M.; Brezinsky, K.; Glassman, I. J Phys Chem
A 1997, 101, 3305.
29. Zhong, X.; Bozzelli, J. W. Chem Phys Processes Com-
bust 1993, 125.
25. Himo, F.; Eriksson, L. A.; Blomborg, M. R. A.;
Siegbahn, P. E. M. Int J Quant Chem 2000, 76,
714.
30. Mackie, J. C. J Phys Chem 1989, 93, 664.
31. Mulholland, J. A.; Akki, U.; Yang, Y.; Ryu, J.-Y. Chemo-
sphere 2001, 42, 719.
26. He, Y. Z.; Mallard, W. G.; Tsang, W. J Phys Chem 1988,
92, 2196.
32. Sidhu, S. S.; Dellinger, H. B. Organohalogen Compd
1997, 31, 469.