In Situ Formed Soot Deposit as a Carbon Source for Polychlorinated Dibenzo-p-dioxins and Dibenzofurans

Article abstract of DOI:10.1021/es034564p

The aim of this study was to investigate the role of in situ formed soot deposits generated during a combustion process for the formation of polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDDs/Fs). In situ formed soot deposits were generated in an entrained flow reactor by using a sooting methane (CH₄) flame (sooting phase), with or without chlorine doped into the flame, and fly ash added into the gas phase. The presence of fly ash in the soot deposit was found to be critical, as a catalyst for formation and/ or a chlorinating agent. The presence of chlorinated aromatic structures in the soot matrix was not enough to promote de novo formation of PCDDs/Fs without the presence of fly ash. PCDFs were formed via direct release of the molecule backbone structure from the soot. PCDDs were formed via a similar mechanism as well as an equally important formation pathway of condensation reactions of C₆ compounds. The formation rate of the soot/ash deposit was still at half its original activity 34 h after the deposits were formed, suggesting a persistent de novo formation occurring for a long time after the sooting incidences (memory effect).

Full text of DOI:10.1021/es034564p

Environ. Sci. Technol. 2004, 38, 2097-2101
lysts, such as perylene, coronene, anthracene, and chloro-
In Situ Formed Soot Deposit as a
Carbon Source for Polychlorinated
Dibenzo-p-dioxins and Dibenzofurans
E V A L E N A W I K S T R O¨ M , † , ‡ S H A W N R YA N , ‡
A B D E R R A H M A N E T O U A T I , § A N D
B R I A N K . G U L L E T T * , ‡
anthracene have shown significant formation of PCDDs/ Fs
(mainly PCDFs) (3, 6-8). PCDD/ F byproducts can be formed
via carbon oxidation reactions, either from direct release
(volatilization) of the C12 backbone ring structure or via
volatilization, condensation, and subsequent reaction of
smaller carbon fragments. Past efforts to understand the
origin of carbon in de novo PCDD/ F formation using a
mixture of 12C and 13C amorphous carbon added to an artificial
fly ash showed no scrambled 12C and 13C PCDF within the
benzenic structure (9), confirming that the three-ring carbon
“backbone” of the PCDF molecule is derived from large (gC12)
carbon structures (e.g., PAHs) present in the matrix rather
than from a composite product of smaller carbon structures.
National Risk Management Research Laboratory,
Air Pollution Prevention and Control Division, Air Pollution
Technology Branch, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711, and
ARCADIS Geraghty & Miller, Inc., P.O. Box 13109,
Research Triangle Park, North Carolina 27709
The PCDDs, however, resulted in both unmixed (all 12C12 or
13
all
C
structures) and mixed (12C6 plus 13C6) structures,
12
indicating an additional formation (9) pathway other than
direct release of the PCDD backbone from the fly ash’s carbon
matrix. These results, however, are in apparent contradiction
with others’ results (10), although these other authors
attributed this discrepancy to their inability to sufficiently
mix the 12C and 13C reactants. Further experiments (9)
suggested that the mixed 12C6/ 13C6 PCDD structures were
the result of an additional formation pathway via a surface
condensation of an aromatic ring structure, such as poly-
chlorinated phenol (PCPh), followed by reaction with a fly
ash-bound aromatic structure.
A link between soot formation and PCDD/ F levels has
been demonstrated through both laboratory experiments (11,
12) and field measurements (13). This may provide a partial
explanation for unpredictably high and temporally persistent
PCDD/ F yields in waste incinerators, sometimes referred to
as a “memory effect” (13-15). The newly formed soot/ fly
ash matrix present on the walls of the incinerator, due to a
poor combustion incident, could then act as an active carbon
source for formation of PCDDs/ Fs, even a long time after the
incident has passed.
The aim of this study was to investigate the role of in situ
formed soot deposits generated during a combustion
process for the formation of polychlorinated dibenzo-p-
dioxins and dibenzofurans (PCDDs/Fs). In situ formed soot
deposits were generated in an entrained flow reactor
by using a sooting methane (CH ) flame (sooting phase),
4
with or without chlorine doped into the flame, and fly ash
added into the gas phase. The presence of fly ash in
the soot deposit was found to be critical, as a catalyst
for formation and/or a chlorinating agent. The presence of
chlorinated aromatic structures in the soot matrix was
not enough to promote de novo formation of PCDDs/Fs
without the presence of fly ash. PCDFs were formed via
direct release of the molecule backbone structure from the
soot. PCDDs were formed via a similar mechanism as
well as an equally important formation pathway of
condensation reactions of C compounds. The formation
6
rate of the soot/ash deposit was still at half its original activity
34 h after the deposits were formed, suggesting a
persistent de novo formation occurring for a long time
after the sooting incidences (memory effect).
Significant questions remain regarding the importance
of de novo reactions for the total PCDD/ F emission in an
incinerator. The focus of this paper is understanding the
mechanistic aspects of de novo synthesis with in situ formed
soot and fly ash mixtures, such as what carbon structures
(C6, gC12, or <C6) are involved in such formation and the
role of gas-phase chlorine. The importance of de novo
formation with respect to time for PCDD/ F emissions is also
addressed.
Introduction
Reaction studies with field-collected or artificial fly ash in
microscale quartz reactors (1-5) have demonstrated forma-
tion de novo of polychlorinated dibenzo-p-dioxins and di-
benzofurans (PCDDs/ Fs) from carbon, chlorine, and cata-
lysts. Experiments under more realistic combustion condi-
tions of time and temperature, in which fly ash and ash-
bound soot are formed in situ, are necessary to clarify the
role of carbon sources and structures contributing to PCDD/ F
emissions.
Fly ash/ soot formed during combustion contains a com-
plex, macromolecular carbon matrix that can interact with
the gas phase by adsorbing and desorbing volatile and semi-
volatile compounds, such as polyaromatic hydrocarbons
(PAHs), during the continuous carbon oxidation reactions.
Indeed, studies on individual PAHs, in the presence of cata-
Materials and Methods
The total net output of PCDDs/ Fs emitted (i.e., a function
of formation, destruction, and desorption reactions occurring
simultaneously) via de novo synthesis from solid carbon
derived from both combustion-generated in situ formed soot
and municipal waste combustor fly ash was studied in an
entrained flow reactor (EFR). The EFR consists of a horizontal
reactor (HR) equipped with a diffusion-type burner and a
vertical reactor (VR). A constant temperature of 1000 °C was
employed in the HR, while a quenched profile, 650-240 °C,
was used in the VR to simulate postcombustion conditions.
A detailed description of the EFR can be found elsewhere (5).
Prior to all de novo experiments, in situ formed soot
deposits on the walls of the VR were generated by combustion
of methane (CH4) at an equivalence ratio (Φ) of 0.86 in the
presence or absence of fly ash or gas-phase chorine (Cl2)
(termed the “sooting phase”). Thus, a soot deposit with or
without preexisting organic chlorine (C-Cl) bonds and with
or without catalytic activity (fly ash) was generated in the
* Corresponding author phone: (919) 541-1534, fax: (919) 541-
0554, e-mail: gullett.brian@epa.gov.
† Joint program with the Oak Ridge Institute for Science and
Education postdoctoral program, Oak Ridge, TN 37831.
‡ U.S. Environmental Protection Agency.
§
ARCADIS Geraghty & Miller, Inc.
9
10.1021/es034564p CCC: $27.50
Published on Web 02/28/2004
2004 Am erican Chem ical Society
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 9 7

TABLE 1. Operating Conditions Used in the Sooting Phase and the de Novo Experimenta
parameter
setting
Sooting Phase
CH4 flow
1.0 L/m in
O2 flow
2.3 L/m in, burner O2; 1.0 L/m in, secondary O2
N2 flow
6.7 L/m in
Cl2 flow (10% in N2)
total flow
burner equivalence ratio, Φ
fly ash feeding
0 or 25 m L/m in (equal to 0 or 200 ppm Cl2 added)
11 L/m in
0.86b
0 or 1 g/h
com bustion tim e
tem perature in HR/VR
gas-phase residence tim e, HR/VR
60 m in
1000 °C/550-240 °C, quenched
0.7 s/1.0 s
De Novo Experiment
O2 flow
0 or 3.3 L/m in
N2 flow
7.7 or 11 L/m in
Cl2 flow (10% in N2)
total flow
25 m L/m in (equal to 200 ppm Cl2 added)
11 L/m in
sam pling tim e
120 m in
tem perature in HR/VR
gas-phase residence tim e, HR/VR
1000 °C/550-240 °C
0.7 s/1.0 s
a
b
All units are presented at STP. Due to the chem ical properties of CH4 and the burner constrution, a significant am ount of soot is form ed at
a Φ of 0.86.
EFR depending on the conditions employed in the sooting
phase. Combustion of CH4 can result in an unstable diffusion
flame and varying amounts of soot formation despite a rather
low value of Φ (0.86) (16).
results, the SIM ion windows were modified to include all
possible PCDD/ F ions, irrespective of isotopic distribution,
for the experiments conducted with 37Cl2 and 13CH4.
Isotopically labeled 13CH4 (99% purity, Cambridge Isotope
Laboratories, CAS No. 6532-48-5) and 37Cl2 gas (95% purity,
Cambridge Isotope Laboratories, CAS No. 13981-73-2) were
employed in selected experiments to study the importance
of carbon and chlorine forms, respectively, for de novo
formation of PCDDs/ Fs. No yields are reported for experi-
ments conducted with labeled 13CH4 or 37Cl2; the levels are
semiquantitative since only external standards could be
employed and, thus, only valid for relative comparisons.
Results are presented as relative homologue profiles, cal-
culated as percentages of the homologue molar yields to the
total PCDD and PCDF molar formation.
Table 1 lists the operating conditions used in the sooting
phase and the experimental settings used for the de novo
experiments. In all experiments except where noted, a
toluene-extracted (EX) fly ash was fed into the VR simulta-
neously with the formation of the soot deposit, to provide
a catalytic environment. The amount of chlorinated organics
present in the EX ash was virtually negligible (less than 50
pmol/ g of PCDDs/ Fs). The toluene extraction reduced the
total amount of carbon to 1.0 wt % as compared to the 1.2
wt % present in the as-received ash (a detailed description
of the EX ash can be found in Wikstro¨ m et al. (5)). Each de
novo experiment was conducted for 120 min (unless other-
wise indicated) in a N2 or air environment with or without
200 ppm Cl2 added into the gas environment (Table 1). It
should be noted that low parts per million levels of oxygen
could be present in the N2 environment due to unavoidable
small leaks in the experimental setup.
Results and Discussion
Table 2 lists the amount of PCDDs and PCDFs sampled,
expressed as picomoles per minute (pmol/ min), formed in
each experiment conducted, where the denominator rep-
resents the sampling time. The molar-based, average degree
of chlorination (#ClPCDD and #ClPCDF) is also listed to indicate
the effect of changes in the experimental conditions on the
homologue profile.
Origin of Chlorine for de Novo Form ation of PCDDs/Fs.
De novo formation of PCDDs/ Fs in an air/ 200 ppm Cl2
environment on a soot/ ash matrix, formed with or without
the presence of gas-phase Cl2 during the sooting phase, results
in about the same PCDD/ F yields (experiment 1b vs ex-
periment 5). Soot formation in the presence of Cl2 results in
a prechlorinated soot matrix (existing C-Cl bonds prior to
the de novo experiment), while the soot/ ash deposit gener-
ated without Cl2 likely has significantly lower C-Cl amounts
on the soot. This is supported by the reported percent level
uptake of gas-phase chlorine by carbon (18), while carbon
chlorination by chloride in fly ash alone results in only parts
per million levels of organic chlorine (without the addition
of gas-phase chlorine) (19). The similar PCDD/ F yields irre-
spective of initial C-Cl amounts in the soot/ ash deposits
suggest that fast reactions between the solid matrix and the
gas-phase chlorine (1 s gas-phase residence time in VR) occur
during the de novo experiment. The importance of fast gas-
phase chlorine reactions for de novo formation of PCDDs/
Fs is thoroughly discussed in a related paper (20).
The initial chlorine content of the soot matrix (prechlo-
rinated vs unchlorinated) was observed to affect the PCDD/ F
homologue distribution (but not the overall yields). Signifi-
cantly higher yields of octachlorinated dibenzofuran (OCDF)
and octachlorinated dibenzo-p-dioxin (OCDD) congeners
were formed in experiments conducted with the prechlo-
rinated soot compared to the unchlorinated soot, despite
the fact that both experiments were conducted with sup-
plementary Cl2 (200 ppm) added into the air environment
(Figure 1).
The samples from each de novo experiment consisted
solely of the collected gases since no soot generation or fly
ash feeding occurred during the actual experiments. In
addition, the sampled values represent only emissions, and
may not reflect the additional formation that remains within
the deposits. The cleanup and analyses of the mono-octa-
PCDD/ F homologues were conducted according to a slightly
modified version of U.S. EPA Method 0023a (17). A high-
resolution gas chromatograph (Hewlett-Packard 589)/ low-
resolution mass spectrometer (Agilent 5971) with a 60 m
J&W DB dioxin column and single-ion monitoring (SIM) was
used for the PCDD/ F analysis. To ensure unambiguous
The participation of preexisting C-Cl bonds in the soot
matrix versus chloride in the fly ash for PCDD/ F formation
9
2 0 9 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

TABLE 2. Experimental Matrix and Measured Yields of PCDDs/Fs from de Novo Experiments on in Situ Formed Soot Depositsa
experimental matrix
measured yields
expt
no.
[Cl2]
(ppm)
gas
phase
soot amtb
(mg)
PCDD amt
(pmol/min)
PCDF amt
(pmol/min)
deposit
surface
#ClPCDD
#ClPCDF
1a
1b
1c
1d
1e
1f
2
3
4
5
6
Cl-soot
Cl-soot
Cl-soot
Cl-soot
Cl-soot
Cl-soot
Cl-soot
Cl-soot
Cl-soot
soot
EX
EX
EX
EX
EX
EX
EX
EX
EX
EX
none
200
200
200
200
200
200
0
200
0
200
0
air
air
air
air
air
air
air
N2
N2
air
N2
413
203
44
99
291
83
125
103
88
166
NA
9
10
2
92
43
9
38
107
63
47
41
7
7.7
6.5
7.5
7.2
7.2
7.0
7.3
7.6
7.3
6.8
ND
7.5
5.6
7.4
7.1
6.7
6.5
7.1
7.2
6.7
5.8
ND
8
20
18
8.0
5.4
1.1
9
31
BDL
Cl-soot
BDL
a
BDL ) below the detection lim it, ND ) no data, and NA ) not analyzed. The experim ents were conducted in random run order to avoid
b
system atical errors; thus, the experim ental num ber is not equal to the run order. Am ount of soot detected in the M23 train during the sooting
phase of the EFR prior to the de novo experim ent.
in this study (examples of two average profiles are presented
in Figure 1). Thus, chlorination of the carbon macromolecular
matrix appears to be critical for the release of the PCDD/ F
carbon backbone structure from the carbon matrix, a reaction
step closely related with the definition of de novo formation.
The highly chlorinated structures are sufficiently resistant to
the oxidation process to be released from the carbon matrix
intact.
Role of Fly Ash and in Situ Form ed Soot for de Novo
Form ation of PCDDs/Fs. Two carbon sources were available
for formation of PCDD/ F when a fly ash/ soot mixture was
present in the VR (experiments 1-5). However, the amount
present as fly ash carbon (1.0 wt %) was estimated as being
between 20 and 100 times less than the amount of soot
carbon. Consequently it was not a surprising result that more
than 98% of the PCDDs/ Fs detected were found to originate
FIGURE 1. Relative PCDD/F homologue profile formed de novo in
an air/200 ppm Cl2 environment with or without preexisting C-Cl
bonds in the soot matrix.
from the soot carbon in an experiment conducted with a
was further studied by generating a CH4 flame soot doped
labeled [13C]soot/ native [12C]fly ash mixture. The relative
with 100 ppm isotopically labeled 37Cl2. EX ash was simul-
contribution of each carbon source (soot and fly ash) to the
taneously fed into the VR as a catalyst and a native chloride
total PCDD/ F formation was determined by a direct com-
(35Cl) source. De novo formation of PCDDs/ Fs from the 37Cl
parison of the GC/ MS ion signals corresponding to the
soot/ 35Cl ash deposit was studied by exposing the soot/ ash
characteristic peaks of each mono- to octacongener within
deposit to an air environment and sampling for 60 min. A
each sample.
homologue profile dominated by the tetrachlorinated and
The role of fly ash was demonstrated by comparison of
higher chlorinated homologues resulted, with consistently
two experiments done with and without fly ash added during
1-2 Cl atoms in each congener originating from the fly ash-
the sooting phase. Nondetectable levels of PCDDs/ Fs were
bound chloride (i.e., 35Cl); this finding is in agreement with
found in an experiment conducted on a matrix consisting of
previous work (12, 20) designed to study the role of various
solely a Cl-laden soot deposit (no added fly ash) produced
chlorine species involved in PCDD/ F formation. The majority
with a sooting CH4 flame doped with 200 ppm Cl2 during the
of the chlorine was derived from the 37Cl-chlorinated soot
preconditioning sooting phase (experiment 6). However,
deposit, consistently nearly 78% and 85% of all chlorine in
significant levels of PCDDs/ Fs were detected in an identical
the tri- to octa-CDFs and tri- to octa-CDDs, respectively.
experiment (experiment 4) performed with the presence of
The consistent participation of the native and labeled
fly ash, showing the importance of fly ash as a catalyst for
chlorine within the PCDD/ F products in these three experi-
PCDD/ F formation. Thus, in the absence of catalytic fly ash,
ments suggests that the same predominant chlorination
the presence of chlorinated and/ or nonchlorinated aromatic
pathway occurs despite the differences in experimental
carbon structures in the soot deposit and the addition of
conditions. One possible pathway for 37Cl in the PCDD/ F
gas-phase Cl2 are not sufficient to form detectable amounts
molecule could be via formation of additional metal chloride
of PCDDs/ Fs, a result in agreement with a previous study
(11). It should be noted that a new vertical quartz reactor
(Me37Cl) complexes in the soot/ ash matrix, which have been
shown in other studies to be very active and necessary for
tube was used for the nonash experiment. Significant
de novo formation of PCDDs/ Fs and low-temperature carbon
formation has been observed in a cleaned reactor (previously
oxidation (3, 21-23). Another probable pathway is that the
exposed to fly ash), without the subsequent addition of fly
ash, due to catalytic residues on the quartz tube wall (12).
labeled gas-phase chlorine directly forms new, chlorinated
carbon (C-37Cl) bonds in the soot/ ash carbon matrix.
Im portance of the Soot Structure for de Novo Form ation
The amounts of DF and DD (i.e., unchlorinated diben-
zofurans and dibenzo-p-dioxin structures) formed were
analyzed in a select number of samples from the EFR. Levels
below the detection limit were observed in all samples,
consistent with the observed declining homologue pattern
from tri- to monochlorinated DD/ DF for all samples analyzed
of PCDDs/Fs. Table 2 presents six experiments (experiments
1a-1f) conducted at the same de novo condition (air/ 200
ppm Cl2) but with a large variation in PCDD/ F yields. The
amount of soot formed during each of the six sooting phases
varied widely between 100 and 400 mg due to the unstable
sooting properties of the CH4 flame (16). Thus, the amount
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 0 9 9

FIGURE 3. Statistical distributions between the possible combina-
tions of PCDDs formed by C12 and C carbon structures.
6
FIGURE 2. Relative PCDD/F homologue distributions detected from
the different carbon sources present on the prechlorinated [13C]-
[12C]soot/ash deposit.
of carbon present for de novo formation of PCDDs/ Fs in the
VR differs significantly between each experiment. A linear
relationship (not shown) between the amount of soot and
the amount of PCDDs/ Fs is observed, suggesting that the
amount of carbon present is one important factor controlling
the rate of de novo formation of PCDDs/ Fs. This finding is
supported by other work (5) showing PCDD/ F formation
proportional to fly ash carbon content.
Further de novo experiments were conducted on a
sequential [13C]soot/ ash, [12C]soot/ ash deposit to investigate
the relative importance of the two de novo pathways, namely,
via direct release of the PCDD/ F carbon backbone structure
during catalyzed soot/ carbon oxidation, and via condensa-
tion reactions of smaller carbon fragments (<C12) also formed
during the low-temperature carbon oxidation process. The
deposit was generated by 10 min of 13CH4/ 200 ppm Cl2 sooting
combustion (Φ ) 0.86) followed by 10 min of combustion
with a sooting 12CH4/ 200 ppm Cl2 flame (Φ ) 0.86). EX ash
was simultaneously fed into the VR during the combustion
to promote catalyzed reactions. On the basis of previous
results, the activity of the fly ash carbon to form PCDDs/ Fs
is assumed to be negligible compared to that of the soot
carbon. De novo formation of PCDDs/ Fs from the [13C]-
[12C]soot/ ash deposit was sampled in an air environment for
30 min to investigate the role of the two carbons. The entire
experiment, i.e., the sooting phase and the following 30 min
de novo experiment, was repeated, resulting in acceptable
variation in terms of both homologue patterns and total
formation. The two-run averages of the relative homologue
profiles of the PCDF and PCDD are presented in Figure 2;
the mono- to trichlorinated DD/ DF homologues were
excluded from the figure due to their relatively insignificant
levels.
FIGURE 4. Changes in de novo formation rate on a Cl-soot/ash
deposit with time.
The observed PCDD distribution is almost identical to
the expected statistical distribution presented in Figure 3,
demonstrating that C6 structures are as important for the de
novo formation of PCDDs as larger aromatic carbon struc-
tures (gC12), such as PAHs, in the soot matrix. The ap-
proximately equivalent distribution among the [12C12]- and
[13C12]PCDFs and the [12C12]-, [12C613C6]- and [13C12]PCDDs
observed in Figure 2 shows that the two carbon sources (13C
and 12C) present in the soot deposit are equally active for de
novo formation of PCDDs/ Fs, consistent with others’ findings
(9). The equal 12C and 13C participation occurs despite the
slight difference in relative age of the soot (i.e., the 13C soot
was 10 min older than the 12C soot at the start of the sampling).
Change in PCDD/F de Novo Form ation with Tim e.
Experiments to study the change in de novo PCDD/ F for-
mation with time (memory effect) were conducted in the
EFR over a period of 34 h. A 13CH4/ 200 ppm Cl2 sooting flame
was operated for 10 min, simultaneously while feeding EX
ash, to supply a [13C]soot/ ash layer on the wall of the VR. The
gas phase was changed to N2, and a 2 h sample was taken
directly after the flame combustion was terminated. The soot
deposit in the VR was further exposed to a constant flow of
N2 (11 L/ min) over the next 24 h, after which four samples
were taken: a 2 h sample in a N2 atmosphere, two samples
for 4 h in air, and a 2 h sample during the operation of a
nonsooting 12CH4 flame (Φ ) 0.67). The relative yields for the
five samples taken during the 34 h are shown in Figure 4. The
levels declined only by about a factor of 2 after 26 h in N2,
and only a slight increase was noticed when the environment
was changed to air. Even 34 h after sooting, the persistently
high rate of formation suggests that the de novo PCDD/ F
formation pathway may remain significant in applied condi-
tions beyond a day after sooting incidences. Moreover, de
novo formation from the [13C]soot/ ash matrix was not found
to be affected by the presence of flame radicals in the gas
phase since the PCDD/ F yields detected with a native CH4
lean flame present were in the same range as those of
formation in the N2 or air environments as shown in Figure
4. Furthermore, no detectable amounts of [12C]PCDDs/ Fs
were observed, showing that the pathway for PCDDs/ Fs at
this condition was via de novo synthesis involving the 13C
soot and not the fuel lean CH4 flame products or the fly ash
carbon.
All PCDFs detected at the two conditions were either
12
13
completely
C
or completely C12, which indicates that
12
PCDFs are formed via direct release from larger carbon
structures, such as PAHs present in the soot/ ash matrix
(Figure 2). The lack of mixed [13Cx12Cy]PCDFs indicates that
these compounds are not formed via condensation/ recom-
bination reactions of smaller carbon fragments. The PCDDs,
on the other hand, were found to contain both 12C and 13C
in equal proportions (13C612C6, Figure 2). Additionally,
equivalent yields of 13C and 12C polychlorinated benzenes
(PCBz’s) and PCPh’s were observed, with all compounds
containing only a single isotope of carbon. The lack of PCDD
congeners with mixed-carbon-isotope rings (e.g., 13C312C3)
and the absence of mixed-isotope PCBz and PCPh structures
indicate that the [13C612C6]PCDDs are formed via a combina-
tion of whole C6 aromatic compounds, and not by a
combination of smaller carbon fragments such as C3. Due
to the unlikelihood of homogeneous reactions under these
low-temperature conditions, these C6 aromatics are likely
volatilized from the soot structure, only to adsorb and react
on surfaces to form PCDDs.
9
2 1 0 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 7, 2004

geners under different combustion conditions is discussed
in further detail in a related research paper (12).
Acknowledgments
Special thanks are due to Dennis Tabor at the U.S. Envi-
ronmental Protection Agency (EPA) for all his help with the
chemical analysis and fruitful discussions. This research was
performed at the U.S. EPA and supported in part by the
appointment of Evalena Wikstro¨ m to the Postdoctoral
Research Program at the National Risk Management Research
Laboratory, administered by the Oak Ridge Institute for
Science and Education (ORISE) through Interagency Agree-
ment BW89938167 between the U.S. Department of Energy
and the U.S. EPA.
FIGURE 5. Formation of PCDDs/Fs at three conditions, solely de
novo on a soot/ash deposit, during moderate sooting (Φ ) 0.86) CH
4
combustion, and at fuel-rich CH combustion (Φ ) 1.03).
4
The importance of PCDD/ F formation via de novo syn-
thesis compared to other formation pathways, such as for-
mation via gas-phase carbon precursors, remains unknown.
Figure 5 presents the PCDD/ F homologue profiles detected
at three experimental conditions: (1) de novo formation in
air via a soot deposit, (2) during operation of a moderately
sooting CH4/ Cl2 flame (Φ ) 0.86), and (3) with CH4/ Cl2 at a
high-sooting, poor combustion condition (Φ ) 1.03). Ash
was present in the VR in all three experiments to provide
catalytic activity. A clear shift in the homologue profile toward
the lower chlorinated DD/ DF homologues was observed
when a flame was present compared to the de novo formation
experiments on the deposits alone, a shift even more
pronounced during poor combustion conditions. The shift
may be partially due to the reported dechlorination effect of
water and the conversion of Cl2 to HCl in the presence of the
flame; however, this does not explain the significant difference
observed between the two flame experiments at relatively
equivalent water concentrations. The C:Cl ratio will be higher
during poor combustion because of higher soot concentra-
tions at constant Cl2 addition. However, it may be likely that
an additional or more prominent formation pathway occurs
during flame combustion in the presence of soot, and
especially during poor combustion, contributing to the for-
mation of mono- and dichlorinated DDs/ DFs. These results
are consistent with others’ findings during low combustion
efficiency conditions in various solid fuel combustors (13,
24, 25).
The total PCDD/ F yields during the moderately sooting
flame (condition 2 above) and in the de novo experiment on
the soot deposit (condition 1) were of the same magnitude
(90 vs 60 pmol/ min, respectively). At the higher sooting
condition, the total formation increased significantly (500
pmol/ min); as noted above, this increase was predominately
due to the increase in mono- and dichlorinated DDs/ DFs.
These experiments suggest that de novo synthesis involving
in situ formed solid carbon matrixes (soot) is an important
formation pathway of PCDDs/ Fs during sooting combustion
processes. The emissions from the resulting carbon/ ash
deposits can persist with little diminution for more than 30
h. The potential for additional pathways leading to the
significant increase in yield of the lower chlorinated con-
Literature Cited
(1) Vogg, H.; Stieglitz, L. Chemosphere 1986, 15, 1373.
(2) Ryan, S. P.; Altwicker, E. R. Chemosphere 2000, 40, 1009.
(3) Stieglitz, L. Environ. Eng. Sci. 1998, 15, 5.
(4) Weber, R.; Sakurai, T. Organohalogen Compd. 2002, 59 49.
(5) Wikstro¨ m, E.; Ryan, S.; Touati, A.; Gullett, B. K. Environ. Sci.
Technol. 2003, 37, 1962.
(6) Weber, R.; Iino, F.; Imagawa, T.; Takeuchi, M.; Sakurai, T.;
Sadakata, M. Chemosphere 2001, 44, 1429.
(7) Iino, F.; Imagawa, T.; Takeuchi, M.; Sadakata, M. Environ. Sci.
Technol. 1999, 33, 1038.
(8) Schoonenboom, M. H.; Olie, K. Environ. Sci. Technol. 1995, 29,
2005.
(9) Hell, K.; Stieglitz, L.; Dinjus, E. Environ. Sci. Technol. 2001, 35,
3892.
(10) Milligan, M. S.; Altwicker, E. R. Environ. Sci. Technol. 1995, 29,
1353.
(11) Lee, C. W.; Kilgroe, J. D.; Raghunathan, K. Environ. Eng. Sci.
1998, 15, 71.
(12) Wikstro¨ m, E.; Ryan, S. P.; Touati, A.; Tabor, D.; Gullett, B. K.
Environ. Sci. Technol., submitted for publication.
(13) Zimmermann, R.; Blumenstock, M.; Heger, H. J.; Schramm, K.
W.; Kettrup, A. Environ. Sci. Technol. 2001, 35, 1019.
(14) Gullett, B. K.; Dunn, J. E.; Raghunathan, K. Environ. Sci. Technol.
2000, 34, 282.
(15) Gullett, B. K.; Touati, A.; Lee, C. W. Environ. Sci. Technol. 2000,
34, 2069.
(16) Turns, S. R. An introduction to combustion: Concepts and
applications; McGraw-Hill: New York, 1996; Chapter 9.
(17) Test Method 0023A; SW-856 (NTIS PB88-239223); Office of Solid
Waste and Energy Response, U.S. EPA: Washington, DC, 1996.
(18) Puri, B. R. Chem. Phys. Carbon 1970, 6, 191.
(19) Stieglitz, L.; Bautz, H.; Zwick, G.; Will, R. Organohalogen Compd.
1996, 27, 5.
(20) Wikstro¨m, E.; Ryan, S.; Touati, A.; Telfer, M.; Gullett, B. K. Environ.
Sci. Technol. 2003, 37, 1108.
(21) Mul, G.; Kapteijn, F.; Moulijn, J. A. Appl. Catal., B 1997, 12, 33.
(22) Weber, P.; Dinjus, E.; Stieglitz, L. Chemosphere 2001, 42, 579.
(23) Addink, R.; Altwicker, E. R. Chemosphere 2001, 44, 1361.
(24) Wikstro¨ m, E.; Tysklind, M.; Marklund, S. Environ. Sci. Technol.
1999, 33, 4263.
(25) Blumenstock, M.; Zimmerman, R.; Schramm, K.-W.; Kettrup,
A. Chemosphere 2000, 40, 987.
Received for review June 6, 2003. Revised manuscript re-
ceived December 16, 2003. Accepted December 30, 2003.
ES034564P
9
VOL. 38, NO. 7, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 2 1 0 1