De novo synthesis mechanism of polychlorinated dibenzofurans from polycyclic aromatic hydrocarbons and the characteristic isomers of polychlorinated naphthalenes

Article abstract of DOI:10.1021/es980857k

Polychlorinated dibenzofurans (PCDFs) and polychlorinated naphthalenes (PCNs) are known to be emitted from municipal waste incinerators (MWIs) with polychlorinated dibenzo-p-dioxins (PCDDs). Two formation paths for PCDD/Fs could mainly work, which are condensation of the precursors such as chlorophenols and 'de novo' formation from carbon. However the correlation between the chemical structure of carbon and the resulting PCDD/Fs still remains unknown. In this study, the PCDD/Fs formation from polycyclic aromatic hydrocarbons (PAHs) and CuCl was examined at 400 under 10% O₂. Coronene among the PAHs characteristically gave 1,2,8,9-T4CDF and the derivatives. These isomers clearly indicate that chlorination causes the cleavage of the C-C bonds in a coronene molecule and also that oxygen is easily incorporated from its outside to form 1,2,8,9-T4CDF. The symmetrical preformed structures in the coronene molecule enabled to amplify the de novo formation of the isomer. PCNs are also formed directly from these PAHs. Since there have been few reports on the formation mechanism of PCNs, this study will be a first step to know the whole formation paths. We also define the de novo synthesis as the breakdown reaction of a carbon matrix, since the word has been used without the precise definition.

Full text of DOI:10.1021/es980857k

Environ. Sci. Technol. 1999, 33, 1038-1043
De Novo Synthesis Mechanism of
Polychlorinated Dibenzofurans from
Polycyclic Aromatic Hydrocarbons
and the Characteristic Isomers of
Polychlorinated Naphthalenes
F . I I N O , * , † T . I M A G A W A , ‡
M . T A K E U C H I , ‡ A N D M . S A D A K A T A †
Department of Chemical System Engineering, University of
Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, Japan 113-8656, and
National Institute for Resources and Environment,
16-3 Onogawa, Tsukuba, Ibaraki, Japan 305-8569
FIGURE 1. Polycyclic aromatic hydrocarbons examined in this study.
The numbers in the parentheses are those of prestructures possibly
to give 1,2,8,9-tetrachlorinated dibenzofurans. The shadowed
combination is one of those preformed structures.
Polychlorinated dibenzofurans (PCDFs) and polychlorinated
naphthalenes (PCNs) are known to be emitted from
municipal waste incinerators (MWIs) with polychlorinated
dibenzo-p-dioxins (PCDDs). Two formation paths for PCDD/
Fs could mainly work, which are condensation of the
precursors such as chlorophenols and “de novo” formation
from carbon. However the correlation between the
chemical structure of carbon and the resulting PCDD/Fs
still remains unknown. In this study, the PCDD/Fs formation
from polycyclic aromatic hydrocarbons (PAHs) and CuCl
copper and chlorine, it was suggested by K. Hell et al. (10)
that 99% of PCDFs and 75-90% of PCDDs are probably
directly released from preformed structures in the amorphous
carbon matrix, e.g. biphenyls. There have been, however,
very few reports mentioning the relation between the
structures of starting materials and the resulting isomers of
PCDD/ Fs. Therefore, the details on the de novo synthesis
formation such as chlorination and incorporation of oxygen
still remain unclear.
On the other hand, polychlorinated naphthalenes (PCNs)
were found to be contained in fly ash collected from
municipal waste incinerators (MWIs) by G. A. Eiceman et al.
(11) in 1979. K. Ballschmiter et al. (12) reported in 1983 that
chloroethylenes work as precursors to form PCNs among
PCDD/ Fs, polychlorinated biphenyls (PCBs), and other
chlorinated compounds under a pyrolytic condition. In 1984,
the concentrations of PCNs in biological tissue were deter-
mined, and the results ranged from 3 to 62 ng/ g-lipid of
PCNs (13). The 2,3,7,8-tetrachlorodibenzo-p-dioxin toxic
equivalent factors (TEF) of some PCNs isomers were first
proposed by A. Hanberg et al. (14) in 1990. One of the highest
value was determined to be 0.003 for 1,2,3,4,5,6,7-heptachlo-
ronaphthalene (HpCN).
However, all 75 isomers of PCNs had not been available
at that time. A lack of information on the analysis of the
PCNs isomers has prevented us not only from knowing the
precise concentration of PCNs but also from discussing the
formation mechanism in view of homologues and isomers.
Williams et al. (15) reported the relative retention time data
of important hexa- and heptachloronaphthelene isomers.
Imagawa et al. (15-20) identified all di- through hexachlo-
ronaphthalenes using synthesized and commercial standards.
Now the isomer specific analysis is possible for di- through
heptachloronaphthalenes.
In this study, the isomer patterns of PCDFs and PCNs
from polycyclic aromatic hydrocarbons (PAHs) which are
possible carbon sources present in soot were investigated.
The more specific structure of the PAHs than the other carbon
materials could clearly show the de novo formation mech-
anisms such as chlorination and incorporation of oxygen to
PCDFs. We supposed in this study that PAHs could be one
of the possible major carbon sources for PCDFs in MWIs,
although they have not been paid much attention. PCDDs
were not characteristically formed from the PAHs. The
formation paths of PCNs in the MWIs should be also clarified,
was examined at 400 under 10% O . Coronene among the
2
PAHs characteristically gave 1,2,8,9-T4CDF and the
derivatives. These isomers clearly indicate that chlorination
causes the cleavage of the C-C bonds in a coronene
molecule and also that oxygen is easily incorporated from
its outside to form 1,2,8,9-T4CDF. The symmetrical
preformed structures in the coronene molecule enabled
to amplify the de novo formation of the isomer. PCNs are
also formed directly from these PAHs. Since there have
been few reports on the formation mechanism of PCNs, this
study will be a first step to know the whole formation
paths. We also define the de novo synthesis as the breakdown
reaction of a carbon matrix, since the word has been
used without the precise definition.
Introduction
The emission of polychlorinated dibenzo-p-dioxins (PCDDs)
and polychlorinated dibenzofurans (PCDFs) from municipal
waste incinerators (MWIs) was reported by K. Olie et al. (1)
in 1977. This report was widely paid attention to, because
some isomers of PCDD/ Fs had been known to be very toxic.
L. Stieglitz et al. (2) and H. Hagenmaier et al. (3) showed that
PCDD/ Fs are formed not only from precursors such as
chlorophenols but also from particulate organic carbon
present in fly ash and inorganic chloride. This formation
path is so-called “de novo” synthesis. Since then some studies
(4-9) have been carried out to elucidate the formation
mechanism of PCDD/ Fs from a carbon matrix.
By means of the identification of the products formed
from 12C- and 13C-carbon as starting materials with inorganic
* Corresponding author phone: +81-298-58-8227, fax: +81-298-
58-8209; e-mail: fukuya@prcom1.nire.go.jp.
† University of Tokyo.
‡ National Institute for Resources and Environment.
9
1 0 3 8 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 7, 1999
10.1021/es980857k CCC: $18.00
1999 Am erican Chem ical Society
Published on Web 02/24/1999

FIGURE 2. T4CDF isomers from graphite and 5% CuCl with dibenzopentacene, perylene, benzo[ghi]perylene, coronene and ovalene.
because not only some of the isomers are toxic but also PCNs
might be a good indicator to show where the de novo
formation occurs in the MWIs.
The PAHs for this experiment were selected in terms of
their structural differences, so that some of them are
structurally related each other and this selection is not
necessarily in accordance with the contents of the PAHs in
fly ash (21).
Experimental Section
The PAHs used in this experiment are perylene (C20H12),
benzo[ghi]perylene (C22H12), coronene (C24H12), 1,2;8,9-
dibenzopentacene (C30H18), and ovalene (C32H14) as shown
in Figure 1. Some of these PAHs are known to be practically
emitted from the MWIs (21). Graphite particles were sieved
FIGURE 3. The examples of the formation mechanisms of T4CDF
to a particle size < 125 µm and mechanically mixed with 0.1
directly formed from coronene. (a) The incorporation of oxygen
wt % of the PAHs and 5 wt % copper(I) chloride. The mixture
occurs from the outside of coronene to result in 1,2,8,9-T4CDF and
of 1 g was placed in a quartz tube reactor heated at 400 °C
for 2 h. The products were collected in an ice-cooled water
(b) form the inside to form 1,4,6,9-T4CDF.
a SP-2331 column owing to the interfering peaks at the same
retention times.
trap and a florisil trap. The inlet oxygen into the reactor was
controlled at 10% by N2 and air. The total flow rate was 200
mL/ min.
PCNs. Columns of DB-5 or DB-1701 (30 m, 0.25 mm i.d.,
0.25 µm film thickness) were used for analysis of PCNs
isomers. The column oven temperature was programmed
from 160 to 300 °C (280 °C in the case of DB-1701) at a rate
of 4/ min. PCN homologues were determined by selective
ion monitoring at m/ z 196 and 198, 230 and 232, 264 and
266, 300 and 302, and 334 and 336 for di- (DiCNs), tri-
(TriCNs), tetra- (TeCNs), penta- (PeCNs), and hexachlo-
ronaphthalenes (HxCNs), respectively. The PCNs isomers
were identified comparing the retention time with synthe-
sized standards (18, 20). The other details on the analysis of
PCNs are described with the elution order and retention
indices in the previous paper (23).
The products in the traps and residues in the reactor were
extracted with ethyl acetate. The extracted samples were
separately concentrated with a rotary evaporator to a few
milliliters and were cleaned up with chromatographic
columns filled with acidified, alkaline, and neutral silica gel.
After the sample volumes were finally reduced to 200-300
µL, the samples were injected into GC-MS Hitachi M-80B.
PCDD/Fs. A SP-2331 column (60 m, 0.25 mm i.d., 0.25 µm
film thickness) was used for analysis of PCDFs isomers. The
column oven temperature was programmed from 200 to 300
°C at a rate of 4/ min. The identification of PCDFs isomers
was carried out according to the retention times investigated
by J. J. Ryan et al. (22). The PCDFs formed were identified
from T4CDFs to O8CDF. However, 1,2,4,7-, 1,3,6,7-, 1,3,4,6-,
and 1,2,4,8-T4CDF in T4CDFs and 1,2,4,7,8-, 1,2,3,7,9-, and
1,2,3,6,7-P5CDF in P5CDFs could not be identified even by
Results and Discussion
PCDD/Fs. The isomer patterns of tetrachlorodibenzofurans
(T4CDFs) from graphite and 5% copper(I) chloride with the
9
VOL. 33, NO. 7, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 3 9

FIGURE 4. P5CDF isomers from graphite and 5% CuCl with the PAHs used in this experiment.
PAHs at 400 °C are shown in Figure 2. The isomer patterns
at 300 and 500 °C were similar to this pattern. There are two
characteristic isomers observed, which are 1,2,8,9-T4CDF
(64%) from coronene and 2,4,6,7-T4CDF (44%) from perylene.
These isomers were reproducibly formed. Even 0.1 wt % of
the PAHs in graphite provided these higher yields. These
characteristic isomers were not formed from the control case
of graphite and 5% copper(I) chloride (24).
but also fly ash which contains some PAHs, although this is
one of the numerous de novo formation paths that could
occur. The three characteristic isomers from perylene,
2,4,6,7-, 2,3,4,6-, and 3,4,6,7-T4CDFs, need more investigation
to mention the formation mechanisms in detail.
In the isomer patterns of P5CDFs (Figure 4) and H6CDFs
(not shown here), it is clear that the derivative isomers from
1,2,8,9-T4CDF are predominantly formed. Those isomers are
1,2,4,8,9- and 1,2,3,8,9-P5CDF and 1,2,4,6,8,9- and 1,2,3,4,8,9-
H6CDF which could be chlorinated after the formation of
1,2,8,9-T4CDF.
PCNs. The TriCNs isomer patterns from the PAHs by DB-5
are described in Figure 5. Some of the most characteristic
isomers are 1,2,3-TriCN from 1,2;8,9-dibenzopentacene,
1,4,5-TriCN from perylene, and 1,2,8-TriCN from benzo[ghi]-
perylene.
The TeCNs isomer patterns by DB-5 are shown in Figure
6, except for 1,2,6,7-TeCN which was determined by DB-
1701. It is clear that the yields of 1,2,7,8-TeCN from benzo-
[ghi]perylene, coronene, and ovalene are drastically higher
than those from 1,2;8,9-dibenzopentacene and perylene.
Perylene gave 1,2,4,8- and 1,2,5,8-TeCNs which are thought
to be derived from 1,4,5-TriCN. From 1,2;8,9-dibenzopen-
tacene, the analysis by a DB-1701 column made it clear that
the yield of 1,2,6,7-TeCN is higher than the other 1,2,3,7- and
1,2,3,4-TeCNs which coelute with 1,2,6,7-TeCN on DB-5.
The PeCNs isomer patterns are shown in Figure 7.
1,2,3,6,7-PeCN is one of the most remarkable isomers formed
from 1,2;8,9-dibenzopentacene. Perylene gave the charac-
teristic isomer of 1,2,3,4,5-PeCN. The three isomers of
1,2,4,7,8-, 1,2,3,4,5-, and 1,2,3,7,8-PeCNs are characteristically
formed from benzo[ghi]perylene. Coronene and ovalene gave
the common characteristic isomers of 1,2,4,7,8- and 1,2,3,7,8-
PeCNs.
The schematic example of the possible main paths from
perylene is illustrated in Figure 8. There are two combinations
of the preformed structures in a perylene molecule directly
to give 1,8-DiCN (from the shadowed part) and 1,2,3,4,5-
PeCN (from the striped part), respectively. However, the yield
of 1,8-DiCN was not higher than the other DiCNs. Therefore,
it is considered that 1,4,5-TriCN is formed directly from
3-monochloroperylene, which means that chlorination oc-
1,2,8,9-T4CDF was also formed from benzo[ghi]perylene
(14%) and ovalene (14%), while 1,2;8,9-dibenzopentacene
and perylene gave only 0.3% and 1%, respectively. This
difference of the 1,2,8,9-T4CDF yields is considered to be
derived from the structures of the starting materials shown
in Figure 1. The shadowed parts in benzo[ghi]perylene,
coronene, and ovalene (Figure 1) could provide the 1,2,8,9-
T4CDF on the assumption that chlorination occurs after the
cleavage of a C-C bond. In addition, oxygen needs to be
incorporated from the outside of the molecule in order to
form 1,2,8,9-T4CDF (Figure 3). The incorporation of oxygen
from the inside of the molecule results in 1,4,6,9-T4CDF.
The precise relative intensity of 1,4,6,9-T4CDF cannot be
determined because the peak of 1,4,6,9-T4CDF overlapped
with 2,3,6,8- and 1,2,4,9-T4CDF using whichever column of
DB-5 or SP-2331. However, the relative intensity of these
three isomers is only 16% so that 1,2,8,9-T4CDF formation
is more facilitated than 1,4,6,9-T4CDF. It means that oxygen
is relatively easily incorporated from the outside of the
molecule. The difference of the yields which is equivalent to
the relative intensities of the isomers possibly depends on
the symmetry of the molecules. The numbers in the
parentheses in Figure 1 indicate how many preformed
structures each PAH has. The preformed structure containing
a biphenyl can provide the 1,2,8,9-T4CDF. There are no
preformed structures in perylene and 1,2;8,9-dibenzopen-
tacene, whereas benzo[ghi]perylene and ovalene have one
and two, respectively. Coronene which gave the highest yield
of 1,2,8,9-T4CDF has six combinations of the preformed
structure in itself. Consequently, it can be concluded that
the structural features of the symmetrical preformed struc-
tures of a coronene molecule amplified one of the de novo
formation paths of the PCDFs isomers formed from the PAHs.
This path could be also applied in the case of not only graphite
9
1 0 4 0 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 7, 1999

FIGURE 5. TriCNs isomers from graphite and 5% CuCl with the PAHs used in this experiment.
FIGURE 6. TeCNs isomers from graphite and 5% CuCl with the PAHs used in this experiment.
FIGURE 7. PeCNs isomers from graphite and 5% CuCl with the PAHs used in this experiment.
curs before its breakdown. The yield of 1,2,3,4,5,6-HxCN from
perylene is over 70% in the HxCNs homologues.
formed structures possibly to provide 1,2,8-TriCN (from the
striped part), 1,2,7,8-TeCN (from the shadowed part), 1,2,3,4,5-
PeCN (the white part), and 1,2,3,4,5,8-HxCN as shown in
Benzo[ghi]perylene has four combinations of the pre-
9
VOL. 33, NO. 7, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 4 1

FIGURE 8. An example of possible main reaction paths to the resulting PCNs from perylene.
FIGURE 9. An example of possible main reaction paths to the resulting PCNs from benzo[ghi]perylene.
Figure 9. The yield of 1,2,3,4,5,8-HxCN was small, although
the direct formation from benzo[ghi]perylene and chlorina-
tion of 1,2,3,4,5-PeCN seem to give the high yield of the HxCN.
This is probably because the direct formation of 1,2,3,4,5,8-
HxCN has to occur not from the edge of the molecule but
from the inner part of the molecule to form the highly
chlorinated. The most characteristic isomer in HxCNs from
benzo[ghi]perylene is 1,2,3,4,5,6-HxCN which can be formed
from the breakdown of 7-monochlorobenzo[ghi]perylene
(from the white part).
Imagawa et al. (20) investigated the PCNs isomer patterns
in fly ash samples from stoker incinerators and fluidized bed
incinerators for municipal waste. It was found that 1,2,6,7-
TeCN, 1,2,3,6,7-PeCN, and 1,2,3,4,6,7- and 1,2,3,5,6,7-HxCNs
are dominant in the fly ash samples from stoker incinerators.
1,2;8,9-Dibenzopentacene showed the more similar isomer
pattern than the other PAHs used in this study. The structure
of 1,2;8,9-dibenzopentacene is relatively linear compared with
the other PAHs, which might cause chlorination on the lateral
positions. On the other hand, 1,2,8-TriCN, 1,2,3,4- and 1,2,4,8-
TeCNs, and 1,2,4,7,8- and 1,2,3,4,5-PeCNs are characteristi-
cally identified in the fly ash samples from fluidized bed
incinerators. Chlorine substitution occurs mainly on the
successive positions including 1,8-positions. This pattern
partially resembles those from benzo[ghi]perylene, coronene,
and ovalene, which implies the relation with relatively
“round” structure of the carbon sources in fly ash. The
contribution of PAHs to the emission of chlorinated aromatic
compounds from the MWIs should be clarified.
The good agreement between these characteristic isomers
observed in this experiment and the structurally possible
isomers from the PAHs including 1,2;8,9-dibenzopentacene,
coronene, and ovalene (These schematic formation paths
are not shown.) indicates that PCNs can be formed directly
from the PAHs. However, there are other possible precursors
or pathways for the PCNs formation. Further works on the
9
1 0 4 2 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 33, NO. 7, 1999

whole formation paths of PCNs in the MWIs need to be done
in order to know which reaction mostly contributes to its
emission. The comparison between their reaction rates also
has to be made for efficient countermeasures to prevent their
formation.
Literature Cited
(1) Olie K.; Vermeulen P. L.; Hutzinger O. Chemosphere 1977, 6,
455-459.
(2) Stieglitz L.; Vogg H. Chemosphere 1987, 16, 1917-1922.
(3) Hagenmaier H.; Kraft M.; Brunner H.; Haag R. Environ. Sci.
Technol., 1987, 21, 1080-1084.
(4) Stieglitz L.; Zwick G.; Beck J.; Roth W.; Vogg H. Chemosphere
1989, 18, 1219-1226.
(5) Stieglitz L.; Zwick G.; Beck J.; Bautz H.; Roth W. Chemosphere
1990, 20, 1953-1958.
(6) Jay K.; Stieglitz L. Chemosphere 1991, 22, 987-996.
(7) Addink R.; Olie K. Environ. Sci. Technol. 1995, 29, 1425-1435.
(8) Gullett B. K.; Bruce K. R.; Beach L. O. Chemosphere 1990, 20,
1945-1952.
(9) Gullett B. K.; Bruce K. R.; Beach L. O.; Drago A. M. Chemosphere
1992, 25, 1387-1392.
(10) Hell K.; Stieglitz L.; Zwick G.; Will R. Organohalogen Compd.
1997, 31, 492-497.
The Definition of the de Novo Synthesis. We suppose
that the PCNs formation in this study is also a kind of de
novo synthesis, although the formation mechanisms on PCNs
have not been classified.
What is the definition of the de novo synthesis? M. S.
Milligan and E. Altwicker (25) mentioned in their paper that
the formation from activated carbon “is often referred to as
de novo synthesis, for the combination of unlike starting
materials such as elemental carbon and sources of oxygen,
chlorine, and hydrogen”.
Here in this paper, the definition of the de novo synthesis
of PCDFs is the direct formation from a carbon matrix with
chlorination and incorporation of oxygen. The existence of
this path was practically implied by K. Hell et al. (10), and
it is supposed that the preformed structure to PCDFs in the
carbon matrix may be “biphenyl”.
(11) Eiceman G. A.; Clement R. E.; Karasek F. W. Anal. Chem. 1979,
51, 2343-2350.
(12) Ballschmiter K.; Zoller W.; Scholz Ch.; Nottrodt A. Chemosphere
1983, 12, 585-594.
(13) Jansson B.; Asplund L. Chemosphere 1984, 13, 33-41.
(14) Hanberg A.; Waern F.; Asplund L.; Haglund E.; Safe S. Chemo-
sphere 1990, 20, 1161-1164.
(15) Williams D. T.; Kennedy B.; LeBel G. L. Chemosphere 1993, 27(5),
795-806.
(16) Imagawa T.; Tanaka T. Pollution Control 1988, 23, 277-284 (in
Japanese).
(17) Imagawa T.; Tanaka T.; Miyazaki A. Pollution Control 1989, 24,
159-168 (in Japanese).
(18) Imagawa T.; Yamashita N.; Miyazaki A. J. Environ. Chem. 1993,
3, 221-230 (in Japanese).
In other words, the de novo synthesis could be defined
as one of the breakdown reactions of the carbon matrix. The
formation of PCDD/ Fs from the precursors such as chlo-
rophenols and chlorobenzenes is a condensation reaction.
This classification appears to be more reasonable, although
the de novo synthesis and other formation mechanisms have
not been precisely defined. We suggest here the precise
definition of de novo synthesis as the breakdown reaction
of a carbon matrix.
(19) Imagawa T.; Yamashita N. Chemosphere 1997, 35, 1195-1198.
(20) Imagawa T.; Takeuchi M. Organohalogen Compd. 1995, 23, 487-
490.
(21) Eiceman G. A.; Clement R. E.; Karasek F. W. Anal. Chem. 1979,
51, 2343-2350.
(22) Ryan J. J.; Conacher H. B. S.; Panopio L. G.; Lau B. P.-Y.; Hardy
J. A.; Masuda Y. J. Chromatogr. 1991, 541, 131-183.
(23) Kannan K.; Imagawa T.; Blankenship A. L.; Giesy J. P. Environ.
Sci. Technol. 1998, 32, 2507-2514.
(24) Iino F.; Imagawa T.; Takeuchi M.; Sadakata M. Organohalogen
Compd. 1998, 36, 89-92.
(25) Milligan M. S.; Altwicker E. Environ. Sci. Technol. 1993, 27, 1595-
1601.
In this paper, the evidences of the direct PCDFs and PCNs
formation from PAHs are shown. Therefore it is concluded
that this result is the first evidence to elucidate one of the
de novo formation mechanisms of PCDFs and PCNs. To shed
some light on the whole de novo synthesis mechanism, the
reaction on the medium size carbon matrix between these
PAHs and graphite has to be further investigated.
Acknowledgments
Received for review August 19, 1998. Revised manuscript
received December 17, 1998. Accepted January 6, 1999.
We appreciate Dr. Roland Weber, Ishikawajima-Harima
Heavy Industry Co., Ltd., Japan, for his comments.
ES980857K
9
VOL. 33, NO. 7, 1999 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 1 0 4 3