Formation and decomposition of chloroaromatic compounds in chlorine-containing benzene/oxygen flames

Article abstract of DOI:10.1016/S0045-6535(00)00216-2

Premixed chlorine-containing, fuel-rich, low-pressure benzene/oxygen flames were analysed for the formation of (oxygenated) chloroaromatic compounds and their radicals by means of the condensation/radical-scavenging method (Hausmann, M., Homann, K.-H., 1995. Ber. Busenges. Phys. Chem. 99, 853-862). Several chlorinated organic compounds (methyl chloride, t-butyl chloride, chlorobenzene, chloroform) were used as additives within a maximum concentration of 10% of total fuel. Product identification and quantification were performed by GC/MS. The extent of formation of chloroaromatic compounds in these flames was largest in the cases of chlorobenzene and chloroform as additives. For chlorobenzene, 12 different chloroaromatics could be analysed in between C₇H₇Cl and C₁₂H₉Cl. Their formation is mainly due to conversion of initial chlorobenzene into substituted or oxidised derivatives, or growth products. Additional chlorination of aromatics is shown to be of minor importance in chlorobenzene-containing flames. Three isomeric (o/m/p) scavenging products could be identified for the chlorophenyl radical. In the chloroform case, 15 chloroaromatics could be analysed in between C₆H₅Cl and C₁₄H₉Cl. The weak C-Cl bond in chloroform is responsible for the high extent of chloroaromatics formation, either by Cl abstraction from the additive or by chlorination reactions via Cl radicals. Additionally, specific pathways to (di)chloroaromatics and chlorinated fulvene-type structures are outlined via CHCl₂ and CCl₂ radicals.

Full text of DOI:10.1016/S0045-6535(00)00216-2

Chemosphere 42 (2001) 439±447
Formation and decomposition of chloroaromatic compounds
in chlorine-containing benzene/oxygen ¯ames
*
J. Burfeindt, K.-H. Homann
Institut fur Physikalische Chemie, Technische Universitat Darmstadt, Petersenstrasse 20, D-64287, Darmstadt, Germany
Abstract
Premixed chlorine-containing, fuel-rich, low-pressure benzene/oxygen ¯ames were analysed for the formation of
(oxygenated) chloroaromatic compounds and their radicals by means of the condensation/radical-scavenging method
(Hausmann, M., Homann, K.-H., 1995. Ber. Busenges. Phys. Chem. 99, 853±862). Several chlorinated organic com-
pounds (methyl chloride, t-butyl chloride, chlorobenzene, chloroform) were used as additives within a maximum
concentration of 10% of total fuel. Product identi®cation and quanti®cation were performed by GC/MS. The extent of
formation of chloroaromatic compounds in these ¯ames was largest in the cases of chlorobenzene and chloroform as
additives. For chlorobenzene, 12 di€erent chloroaromatics could be analysed in between C7H7Cl and C12H9Cl. Their
formation is mainly due to conversion of initial chlorobenzene into substituted or oxidised derivatives, or growth
products. Additional chlorination of aromatics is shown to be of minor importance in chlorobenzene-containing ¯ames.
Three isomeric (o/m/p) scavenging products could be identi®ed for the chlorophenyl radical. In the chloroform case, 15
chloroaromatics could be analysed in between C6H5Cl and C14H9Cl. The weak C±Cl bond in chloroform is responsible
for the high extent of chloroaromatics formation, either by Cl abstraction from the additive or by chlorination reactions
via Cl radicals. Additionally, speci®c pathways to (di)chloroaromatics and chlorinated fulvene-type structures are
outlined via CHCl2 and CCl2 radicals. Ó 2001 Elsevier Science Ltd. All rights reserved.
Keywords: Radical scavenging; Product identi®cation; PAH; Chlorophenols; Chlorobenzene; Chloroform
1. Introduction
hardly ever been focused on the formation of chloro-
aromatics. The number of investigations on the homo-
geneous gas phase formation and decomposition of
chloroaromatic compounds is considerably less. The
main part of the experimental work has been performed
under conditions of pyrolysis (Cullis and Manton, 1958;
Fields and Meyerson, 1966; Louw et al., 1973, 1984;
Rao and Skinner, 1984, 1988; Ritter et al., 1990; Kern
et al., 1992; Mulholland et al., 1993, 1996; Rouzet et al.,
1999), hydrogenolysis (Manion et al., 1988; Cui et al.,
1989; Manion and Louw, 1990a, b) and low-temperature
oxidation (Mulder and Louw, 1985, 1986; Born et al.,
1993), while there are only a few data obtained from
homogeneous combustion (Eklund et al., 1987; Peder-
The in¯uence of chlorine on hydrocarbon combus-
tion has been extensively investigated with regard to
inhibition e€ects (Westbrook, 1982), acceleration of
polycyclic aromatic hydrocarbon (PAH) growth,
(Frenklach, 1990) and soot formation (McKinnon and
Howard, 1990). A lot of work has also been done on
combustion and pyrolysis of chlorinated aliphatic com-
pounds, either pure or in mixture with non-chlorinated
fuels, which is not to be reviewed here, and which has
* Corresponding author. Tel.: +49-6151-164149; fax: +49-
6151-164717.

sen, 1991; Pedersen and Kallman, 1991).
Laminar, premixed, low-pressure ¯ames are widely
used as model ¯ames to study the hydrocarbon
E-mail address: homann@pc2.chemie.tu-darmstadt.de (K.-
H. Homann).
0045-6535/01/$ - see front matter Ó 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 5 - 6 5 3 5 ( 0 0 ) 0 0 2 1 6 - 2

440
J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
combustion. The use of benzene as fuel in laminar
¯ames has been recently studied in detail employing
various analytical methods (Bittner and Howard, 1981;
Hausmann et al., 1992; Ahrens et al., 1994; Bachmann et
al., 1995, 1996; Canic, 1999), including condensation of
the ¯ame gases with a radical scavenger (dimethyl di-
sul®de) (Hausmann et al., 1992; Canic, 1999), and has
also been subjected to modelling (Lindstedt and Skevis,
1994). The scavenging technique has been applied to
several laminar ¯ames employing aliphatic (Hausmann
et al., 1992; Hausmann and Homann, 1995; 1997; Canic,
1999) as well as aromatic (Hausmann et al., 1992;
Griesheimer, 1997; Griesheimer and Homann, 1998;
Canic, 1999) fuels, the latter ones producing much
higher intermediate yields of monocyclic aromatics and
PAHs. To get a deeper insight into the formation and
decomposition of (oxygenated) chloroaromatic com-
pounds within ¯ame conditions, the condensation/radi-
cal-scavenging method (Hausmann and Homann, 1995)
was applied in the present work to investigate laminar,
fuel-rich benzene ¯ames containing variable amounts
(<10% of total fuel) of chlorinated organic additives.
O ratio was adjusted by a subsequent dilution with a
second oxygen ¯ow. The additives (except for gaseous
CH3Cl) were vaporised in a similar manner using argon
as the carrier gas in a two-stage vaporiser/condenser.
Burner and sampling system (Hausmann and Ho-
mann, 1995). The ¯ames were stabilised on a water-
cooled sintered-disk burner. The ¯ame gases were
sampled through a small ori®ce (d ˆ 0.8 mm) at the tip
of a conical quartz probe, allowing them to expand as a
3
jet beam into a vacuum of <10 mbar. They were
condensed simultaneously with an excess of the gaseous
radical scavenger dimethyl disul®de (DMDS) in a
spherical cold trap at liquid-nitrogen temperature.
Radicals were transferred into stable methylthio com-
pounds via the scavenging reaction (Hausmann and
Homann, 1995)
H3C±S±S±CH3 ‡ áR ! R±SCH3 ‡ áSCH3:
ꢀ1†
For a detailed characterisation of the method, see
Hausmann and Homann (1995).
Analysis. Stable, condensable ¯ame compounds, and
scavenging products from radicals were identi®ed and
quanti®ed by GC/MS (DANI GC 6500/Finnigan MAT
ITD 700) from the lique®ed sample within a detection
limit corresponding to a mole fraction of 0.1 ppm in the
¯ame. The applied method was not suitable for the
analysis of the main combustion products (CO, H2,
CO2, H2O, C2H2, HCl), for small hydrocarbons (C1±
C3), and small radicals (H, O, OH, HO2, CH3). The
upper limit of the analytical mass range was about 300
a.u.
2. Experimental
Flame parameters. Benzene was burned as the main
fuel in a laminar, fuel-rich ¯ame along with di€erent
chlorinated organic additives (Table 1), the latter being
present to a maximum extent of 10 (mol-)% of total fuel
(2% of unburned gas). Except for the 5% chlorobenzene
case, the C/Cl ratio of the investigated ¯ames was set to
a constant value of about 60, in order to evaluate
comparatively their propensities to chloroaromatics
formation. The slightly sooting ¯ames (C/O ˆ 0.80,
u ˆ 2.0) were stabilised under reduced pressure (27
mbar) applying an unburned gas ¯ow velocity of vu ˆ 45
cm s 1. A pure benzene ¯ame was analysed additionally
for comparison within the conditions described above.
For further characterisation of the investigated ¯ames,
see Table 1.
3. Results and discussion
3.1. Mono-chlorinated aliphatic additives
Methyl chloride (CH3Cl) and t-butyl chloride (t-
C4H9Cl) were used as additives with initial amounts of
10% of fuel. Table 2 gives the extent of formation
of chloroaromatics in the investigated ¯ames by means
of the sum of their maximum concentrations, i.e., their
maximum mole fractions xmax in the ¯ame, and the
Preparation of fuel mixture. A ¯ow of oxygen was
saturated with benzene in a three-stage vaporiser/con-
denser system (Canic, 1999) at t ˆ 52°C. The desired C/
Table 1
Flame parameters (for all ¯ames: C=O ˆ 0:80 ꢀu ˆ 2:0†; p ˆ 27 mbar; mu ˆ 45 cm s 1; xꢀAr† ˆ ꢀ11 Æ 1†%)
Additive (% of
total fuel)
Initial mole fraction
of additive x0/ppm
H/Cl
C/Cl
H/C
CH3Cl (10)
t-C4H9Cl (10)
C6H5Cl (5)
C6H5Cl (10)
CHCl3 (3.3)
None
20200
19200
9360
18300
6370
0
57
63
119
59
59
±
55
58
120
60
59
±
1.04
1.09
0.99
0.98
1.00
1.00

J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
441
Table 2
Sum of maximum concentrations (mole fractions)
nated benzene ¯ame. The exclusive formation of benzyl
chloride can be attributed to chloromethylation of
benzene by CH2Cl. Whereas Cl abstraction by H radi-
cals to yield methyl has been reported to be the most
ecient step of CH3Cl degradation in ¯ames, chlorom-
ethyl is formed to a somewhat lower extent (Karra et al.,
1988). The concentration pro®le of CH2Cl could not be
determined in the present work due to superimposition
of the radical scavenger upon the CH2Cl scavenging
product during gas chromatography. In the t-C4H9Cl-
containing ¯ame, three isomeric C4H8Cl radicals were
detected as scavenging products within maximum con-
centrations of 98, 9.6, and 4.4 ppm, respectively, while
C4H9 was de®nitely absent. They are assigned to be the
primary H abstraction product from t-C4H9Cl and two
isomeric radicals, which are formed due to stabilisation
of the initial primary radical by either internal CH3 or Cl
shift. The decomposition of C4H8Cl leading to propene
and CH2Cl can be assumed as a reasonable pathway for
chloromethyl formation in the butyl chloride-containing
¯ame, as no further chlorine-containing aliphatic radi-
cals could be detected (C4H8Cl may alternatively de-
compose forming isobutene and Cl radical).
P
xmax of
chloroaromatic compounds formed in chlorine-containing
benzene ¯amesa
P
Additive (% of
total fuel)
xmax/ppm
Number of detect-
able chlorinated
aromatics
CH3Cl (10)
t-C4H9Cl (10)
C6H5Cl (10)
CHCl3 (3.3)
243
181
4
4
212b
12b
489
15
a For ¯ame parameters, see Table 1.
b Excluding initial chlorobenzene.
number of di€erent chloroaromatic compounds. Com-
paring the entries for CH3Cl and t-C4H9Cl, the pro-
pensity to aromatic chlorination is similar in both cases.
The narrow product spectrum is dominated by chloro-
benzene (cf. Table 3), which is the chlorination product
of the main fuel compound. The slight di€erence in
P
xmax for the two additives gets an even lesser weight if
one takes into account that chlorobenzene is formed
earlier in the t-C4H9Cl-containing ¯ame and therefore
exhibits a broader pro®le shape along with a lower
maximum concentration. Additional experiments car-
ried out with benzyl chloride as the additive (1% of total
fuel) also indicated a similar extent of aromatic chlori-
nation reactions.
3.2. Chlorobenzene as additive
Overview. Chlorobenzene was used as the additive
within amounts of 5% or 10%, respectively, of total fuel.
As it can be seen from Table 2, for the latter case, the
spectrum of detectable chloroaromatics is noticeably
wider than in the ¯ames containing the alkyl chloride
additives, whereas the sum of maximum concentrations
is similar in each case. The chloroaromatics identi®ed
from the benzene/chlorobenzene ¯ames are presented in
Three chloroaromatics in addition to chlorobenzene
could be detected in the benzene ¯ames containing
CH3Cl and t-C4H9Cl (Table 3). As non-chlorinated
phenol is the most abundant aromatic by-product from
benzene ¯ames, chlorophenols can be expected to be
formed either by oxidation of chlorobenzene or by
chlorination of phenol. Chloronaphthalenes represent
the chlorination products of the most abundant PAH
from the pure benzene ¯ame. However, the extent of
chlorination is rather low: only about 0.5% of phenol or
1% of naphthalene, respectively, are formed as chlori-
nated derivatives. Whereas methyl chlorobenzenes (o/m/
p chlorotoluenes) could not be detected, the high extent
of benzyl chloride formation is somewhat surprising. Its
concentration pro®le is strongly correlated with that of
the benzyl radical, which is assigned to be formed from
benzyl chloride by C±Cl bond cleavage as compared to
its considerably lower concentration in the non-chlori-
Table 4 along with their maximum concentrations xmax
.
To illustrate the extent of chloroaromatics formation,
the percentages of the chlorinated compounds Ar±Cl in
the total maximum concentrations of both, chlorinated
and non-chlorinated aromatics (Ar±Cl + Ar±H), are
given additionally in Table 4. The relative abundances of
the respective o/m/p isomers are given in the left part of
Table 5.
It should be noticed, that the detected chloroaro-
matics mainly represent the mono-chlorinated deriva-
tives of the most abundant non-chlorinated aromatics
from pure fuel-rich benzene ¯ames. The relative
Table 3
Maximum concentrations xmax of chloroaromatic compounds formed in benzene/CH3Cl and benzene/t-C4H9Cl ¯amesa
Compound
Additive: 10% CH3Cl xmax/ppm
Additive: 10% t-C4H9Cl xmax/ppm
Chlorobenzene
Chlorophenols
227
5.7
2.7
7.8
159
8.0
2.4
Chloronaphthalenes
Benzyl chloride
11.2
a For ¯ame parameters, see Table 1.

442
J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
Table 4
Maximum concentrations xmax of chloroaromatic compounds formed in benzene/chlorobenzene ¯amesa
Additive: 5% C6H5Cl
Additive: 10% C6H5Cl
Compound
xmax/ppmb
Percent of total xmax
(ArH + ArCl)
xmax/ppm
Percent of total xmax
(ArH + ArCl)
Chlorotoluenes
Benzyl chloride
11.0
2.1
6.4
±
22.1
6.9
ꢁ13
±
Dichlorobenzenes
Chlorophenols
Chlorocresols
n.d.
33.0
t.
±
0.4
±
4.2
±
99.3
1.0
9.7
2.4
11.1
6.4
16.1
5.3
28.7
1.7
0.7
Ethynyl chlorobenzenes
Chlorostyrenes
27.4
t.
5.2
±
53.0
3.6
Butadiynyl chlorobenzenesc
Chloronaphthalenes
Chlorobiphenyls
ꢁ2.5
2.2
ꢁ9
2.4
16.5
±
5.6
4.9
6.6
13.5
0.7
Chloroacenaphthalenesc
n.d.
n.d.
Chlorophenyls
±
0.9
a For ¯ame parameters, see Table 1.
b t ˆ traces, n.d. ˆ not detected.
c Hypothesized structure.
Table 5
Relative abundance of o/m/p isomers of substituted chlorobenzenes formed in benzene/chlorobenzene and benzene/chloroform ¯amesa
Compound
Additive: 10% C6H5Cl o/m/p
Additive: 3.3% CHCl3 o/m/p
Chlorotoluenes
Dichlorobenzenes
Chlorophenols
29/52/19
9/84/7
40/40/20
100/0/0
51/49b
31/46/23
0/100/0
±
26/74b
Ethynyl chlorobenzenes
Chlorostyrenes
30/44/26
28/55/17
36/44/20
22/52/26
33/47/20c
Butadiynyl chlorobenzenes
Chlorobiphenyls
Chlorophenyls
25/50/25
±
a For ¯ame parameters, see Table 1.
b m and p isomers not separated.
c Data from ¯ame with C/O ˆ 0.65.
abundances of chlorophenols and alkylated chloro-
benzenes compared to their respective non-chlorinated
counterparts are similar to or slightly higher than the
initial amount of chlorobenzene in the total fuel (cf.
columns 3 and 5 in Table 4). In contrast, chlorinated
PAHs as chloronaphthalenes and chloroacenaphtha-
lenes, and the chlorophenyl radical are formed to a
considerably lower extent. Whereas the maximum con-
centrations of indene and naphthalene are similar in
benzene ¯ames, chlorinated indenes could not be de-
tected in quanti®able amounts in the chlorobenzene-
containing ¯ames. The concentration of benzyl chloride,
which is the only detectable molecule containing sp3
carbon-bound chlorine, is lowest in the benzene/chlo-
robenzene ¯ame compared to the other additives pre-
sented in Table 1. Mono-chlorinated biphenyls reach
unexpectedly high concentrations compared to non-
chlorinated biphenyl, while dichlorinated biphenyls
could not be detected. Dichlorobenzenes, representing
the only detectable compounds containing more than
one chlorine atom, are formed only to a very small ex-
tent by chlorination of chlorobenzene.
The concentration pro®les of chlorinated and non-
chlorinated molecules of the same carbon structure are
usually shaped similarly, showing identical locations of
the concentration maxima, the chlorinated molecules
sometimes being formed and destroyed slightly faster
than their non-chlorinated counterparts. Some pro®les
of chlorinated and non-chlorinated aromatics from the
benzene/5% chlorobenzene ¯ame are compared in Fig. 1.
Fuel decomposition. The fuel decay pro®les from the
benzene ¯ame containing 5% of chlorobenzene are
compared in Fig. 2 along with the temperature pro®le.
As it can be seen, the decomposition of the chlorinated
fuel is somewhat delayed in comparison to benzene.
Thus, chlorobenzene becomes enriched in the ¯ame up
to a maximum extent of about 9% at h ꢁ 7 mm instead
of the initial 5% compared to benzene. This behaviour

J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
443
the ways of formation are the same for both, chlorinated
and the respective non-chlorinated compounds: methy-
lation of (chloro)benzene (Ritter et al., 1990), reaction of
(chloro)phenyl with acetylene (Bittner and Howard,
1981; Frenklach, 1990; McKinnon and Howard, 1990)
or butadiene, and a combination of benzene with
(chloro)benzene or (chloro)phenyl (Louw et al., 1973;
Manion et al., 1988; Ritter et al., 1990; Mulholland et al.,
1993; Rouzet et al.,1999) represent the favourite routes
to the respective carbon-substituted (chloro)benzenes.
Due to the enrichment of chlorobenzene in the ¯ame, the
alkyl-substituted chlorobenzenes appear to be formed
slightly earlier compared to their non-chlorinated
counterparts, and reach relative abundances, which may
be somewhat higher than expected (assuming a value
comparable to the initial amount of chlorobenzene, i.e.,
5% or 10%). Even the high concentrations of chlorobi-
phenyls can be explained by the enrichment of chloro-
benzene. Based on the maximum intermediate
percentage of approximately 9% chlorobenzene in the
¯ame containing initially 5% of chlorobenzene and fur-
thermore taking into account that chlorobiphenyls are
formed by a combination of both fuel compounds,
benzene and chlorobenzene, one can easily con®rm the
experimentally observed relative abundance of chloro-
biphenyls (16% for the 5% chlorobenzene case) com-
pared to biphenyl.
Fig. 1. Concentration pro®les of chlorinated and non-chlori-
nated substituted benzenes in the benzene/chlorobenzene (95/5)
¯ame. For ¯ame parameters, see Table 1.
Chlorostyrenes were detected in rather small con-
centrations compared to styrene. This fact can be ex-
plained by considering additional pathways to styrene
formation, which do not start from benzene or phenyl,
but from aliphatic building blocks, e.g., C3, C4, and C5
fragments (Cole et al., 1984; Canic, 1999). As chlori-
nated aliphatic hydrocarbons could not be detected in
the present work, the chlorostyrenes might be the by-
products of the chlorophenyl + acetylene pathway,
which mainly yields ethynyl chlorobenzenes (Frenklach,
1990; McKinnon and Howard, 1990).
Fig. 2. Decay of fuel and temperature pro®le in the benzene/
chlorobenzene (95/5) ¯ame. For ¯ame parameters, see Table 1.
re¯ects the high stability of the aromatic C±Cl bond
within the early oxidation zone. Whereas H abstraction
from the aromatic ring is the main initiation reaction for
benzene (and chlorobenzene) decomposition in ¯ames
(Bittner and Howard, 1981; Venkat et al., 1982; Mc-
Kinnon and Howard, 1990; Lindstedt and Skevis, 1994),
Cl abstraction by oxygen-containing or Cl radicals can
be neglected due to unfavourable thermodynamics. Cl
removal by H atoms, which has been reported not to
occur via an abstraction mechanism but via an addition/
elimination sequence (Cui et al., 1989), becomes signif-
icant in the later oxidation zone. The latter reaction
therefore leads to the breakdown of the chlorobenzene
concentration at h P 7.5 mm, in addition to thermal C±
Cl bond cleavage (Rao and Skinner, 1988; Frenklach,
1990; McKinnon and Howard, 1990; Kern et al., 1992).
Formation of carbon-substituted chlorobenzenes. It
can be concluded from a comparison of the concentra-
tion pro®les of carbon-substituted chlorobenzenes
(chlorotoluenes, ethynyl chlorobenzenes, butadiynyl
chlorobenzenes, and chlorobiphenyls; cf. Fig. 1) to those
of the corresponding non-chlorinated molecules, that
Formation of chlorophenols. Due to the coincidence of
the concentration pro®les of phenol and chlorophenols
(cf. Fig. 1), one can conclude that chlorophenols are
formed from chlorobenzene by the same pathways as
phenol is formed from benzene. The relative amount of
chlorophenols compared to phenol is nearly indepen-
dent of the reaction time and is always slightly lower
than the initial fuel ratio (cf. Table 4). Assuming that O
or OH addition to the aromatic ring is the main pathway
to phenols in benzene ¯ames (Bittner and Howard, 1981;
Lindstedt and Skevis, 1994), the formation of chlor-
ophenols is somewhat suppressed in comparison to
phenol due to the ipso substitution of chlorobenzene
leading to non-chlorinated phenol (Mulder and Louw,
1986).
Formation of chlorinated PAHs. Chlorinated PAHs,
i.e., chloronaphthalenes and chloroacenaphthalenes, are
formed only to a small extent. Despite the enriched

444
J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
ton, 1958; Fields and Meyerson, 1966; Louw et al.,
1973).
General conclusions. The very weak formation of di-
chlorobenzenes (cf. Table 4), which can only be due to
chlorination of chlorobenzene, indicates that chlorina-
tion of aromatics in general is not an important route to
chloroaromatics in chlorobenzene-containing ¯ames.
Concluding from the comparison of the yield of chlo-
robenzene formed from benzene in the alkyl chloride-
containing ¯ames (cf. Table 3) with the dichlorobenzene
yield obtained in the corresponding benzene/chloro-
benzene ¯ame, the extent of chlorination reactions is
signi®cantly lower (by 2 orders of magnitude) in the
latter case. Therefore, the chloroaromatics, which were
analysed in the chlorobenzene-containing ¯ames are
assigned to be formed mainly starting from chloroben-
zene or chlorophenyl radicals, i.e., from the initial
chlorinated fuel compound.
Fig. 3. Concentration pro®les of chlorinated and non-chlori-
nated PAHs, and of phenyl radicals in the benzene/chloroben-
zene (90/10) ¯ame. For ¯ame parameters, see Table 1.
chlorobenzene concentration, in the benzene/10% chlo-
robenzene ¯ame, about 5% of naphthalene and only
about 1.5% of acenaphthalene are formed as chlorinated
derivatives, while larger chlorine-containing PAHs could
not be detected (cf. Table 4; for concentration pro®les
see Fig. 3). Therefore, the chlorinated PAHs are as-
signed to be mainly formed as growth products starting
from initial chlorobenzene, via e.g., ethynyl chloro-
benzenes. The decrease in relative abundance of the
chlorinated PAHs along with increasing molecular
weight is not only due to the reaction probabilities, but
also due to dechlorination reactions, i.e., exchange of Cl
by H radicals (Cui et al., 1989) and thermal C±Cl bond
®ssion (Frenklach, 1990; McKinnon and Howard,
1990), both reactions becoming faster with increasing
temperature and therefore with an increasing size of the
PAH structure.
3.3. Chloroform as an additive
Overview. As it can be seen from Table 2, the addi-
tion of chloroform to a benzene ¯ame causes the largest
concentration of aromatic chlorination products and the
widest spectrum of detectable chloroaromatics in the
series of additives presented in this paper. The detected
chloroaromatics from the benzene/chloroform ¯ame are
presented in Table 6 along with their maximum con-
centrations and their relative abundances compared to
the respective non-chlorinated compounds. The relative
abundances of the o/m/p isomers are given in the right
Table 6
Maximum concentrations xmax of chloroaromatic compounds
formed in the benzene/chloroform ¯amea
Fuel radicals: phenyl and chlorophenyls. The concen-
tration pro®les of the primary fuel radicals (phenyl and
chlorophenyls) from the benzene/10% chlorobenzene
¯ame are also compared in Fig. 3. The maximum con-
centration of chlorophenyls is less than 1% of the phenyl
concentration. Furthermore, the pro®le of the chlor-
ophenyl radicals is shifted to lower reaction times
compared to phenyl. This observation supports the as-
sumption that phenyl is produced to a considerably high
extent by the decomposition of PAHs (Hausmann and
Homann, 1997). Also taking into account that the
consumption of benzene is nearly completed at h ꢁ
9 mm, the formation of phenyl from benzene by C±H
bond ®ssion is of minor importance. In contrast, C±Cl
bond cleavage is the main decomposition pathway for
chlorinated PAHs as well as for chlorobenzene in the
high-temperature oxidation zone (Rao and Skinner,
1984; Frenklach, 1990; McKinnon and Howard, 1990;
Kern et al., 1992). Thus, the small amount of chlor-
ophenyls which could be detected in benzene/chloro-
benzene ¯ames can only be formed by H abstraction
from the chlorinated fuel compound (Cullis and Man-
Compound
xmax/ppm
Percent of total
xmax (ArH + ArCl)
Chlorobenzene
406
11.1
4.4
±
±
6-Chlorofulveneb
Chlorotoluenes
Benzyl chloride
ꢁ1.2
±
16.5
1.2
o-Dichlorobenzene
6,6-Dichlorofulveneb
Chlorophenols
±
0.8
±
18.0
11.9
1.8
1.1
1.1
1.3
0.8
3.2
±
Ethynyl chlorobenzenes
m-Chlorostyrene
Chloroindenesb
1.7
9.3
Chloronaphthalenes
Chloromethyl naphtha-
lenes
1.2
Chlorobiphenyls
Chloroacenaphthalenesb
2.3
1.5
1.1
2.5
1.7
3.6
C
14H9Cl, e.g., Chlor-
ophenanthrenesb
a For ¯ame parameters, see Table 1.
b Hypothesized structure.

J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
445
carbene, which reacts neither with benzene nor with aryl
disulphides (Searles and Wann, 1965; Field and Banks,
1975 ), but with alkyl disulphides, at least at elevated
temperatures (Searles and Wann, 1965). Nevertheless, its
recombination product, tetrachloroethylene, could be
detected in the present work. Tetrachloroethylene has
been reported to be the major intermediate in chloro-
form pyrolysis (Taylor et al., 1991), being mainly formed
by the recombination of dichlorocarbene. The presence
of CCl2 in the benzene/chloroform ¯ame is additionally
indicated by the analysis of the chloroaromatic by-
products, as described below.
Formation of chlorinated benzenes and their isomers.
Reaction (4) is assigned to be the main pathway to
chlorobenzene formation, because of its remarkably
high exothermicity and the high (initial) concentrations
of the chlorine-donating additive. In contrast, the en-
dothermal chlorination of benzene with Cl as well as the
recombination of phenyl and Cl can be rejected as ef-
fective routes to chlorobenzene. It is interesting to note,
that, on the one hand, o-dichlorobenzene is the only
detectable dichlorobenzene isomer in this ¯ame, but on
the other hand, its maximum concentration is about two
times the sum of dichlorobenzenes in the corresponding
benzene/chlorobenzene ¯ame. The formation of chlo-
robenzene and o-dichlorobenzene is accompanied by
two isomers, which are assigned to be 6-chlorofulvene
and 6,6-dichlorofulvene, respectively. The reaction of
cyclopentadiene, which is the most abundant C5 species
in fuel-rich benzene ¯ames, with dichlorocarbene is
known to yield chlorobenzene as the main product,
forming 6-chloro and 6,6-dichlorofulvene as by-prod-
ucts (Busby et al., 1971; dÕAmore and Bergman, 1971).
Due to the low reactivity of CCl2 towards the radical
scavenger, one cannot clearly decide whether the major
origin of these fulvene-type structures is the gas phase or
the lique®ed sample. Nevertheless, deducing from a
comparison of the concentration pro®les of 6,6-dichlo-
rofulvene and o-dichlorobenzene, the latter isomer is
strongly assumed to be produced from the former by an
internal rearrangement under ¯ame conditions, there-
fore, only giving rise to the formation of the ortho iso-
mer.
Fig. 4. Concentration pro®les of chloroform, tetrachloroeth-
ylene, and the radicals Cl, CHCl2, and CCl3 in the benzene/
chloroform (96.7/3.3) ¯ame. For ¯ame parameters, see Table 1.
part of Table 5. Chlorobenzene, however, widely dom-
inates the spectrum of chloroaromatic by-products.
Decomposition of the additive. Chloroform, in con-
trast to chlorobenzene, is consumed much more rapidly
in benzene ¯ames. Its decay pro®le is given in Fig. 4
along with the concentration pro®les of dichloromethyl
(CHCl2), trichloromethyl (CCl3), Cl radical, and tetra-
chloroethylene. Dichloromethyl is about 10 times more
abundant than trichloromethyl. The former can be
produced from chloroform by several ways:
CHCl3 ! CHCl2 ‡ Cl
ꢀ2†
ꢀ3†
CHCl3 ‡ H ! CHCl2 ‡ HCl
CHCl3 ‡ R ! CHCl2 ‡ RCl
ꢀR ˆ alkyl or aryl†
ꢀ4†
The latter reaction, being exothermic by about 100 kJ
1
mol for R ˆ aryl, is assumed to be a major chlorina-
tion reaction in the chloroform-containing benzene
¯ame. Trichloromethyl is mainly formed by H abstrac-
tion from chloroform by various radicals (H, O, OH,
Cl). Compared to all the other additives, the chlorine
radical reaches remarkably high concentrations in the
¯ame containing chloroform, even at low distances from
the burner. Due to experimental reasons, its pro®le can
be reliably determined only up to h ꢁ 8 mm. Thus, the
chlorine radical must also be taken into account as a
possible chlorinating agent in this ¯ame.
Pathways to chloromethyl aromatics. Concluding
from the high benzyl chloride concentration and from
the detectability of chloromethyl naphthalenes, the ad-
dition of chloroform to benzene ¯ames also provides
ecient pathways to chloromethyl-substituted aromat-
ics. Two reasonable routes might contribute to the for-
mation of these compounds: (a) reaction of benzene (or
naphthalene) with dichloromethyl, leading to the re-
spective chloromethyl compound after an internal rear-
rangement by the splitting of Cl, and (b) Cl abstraction
by the benzyl radical according to reaction (4).
The main initiation step of chloroform pyrolysis has
been reported to be the a,a elimination of hydrogen
chloride to form dichlorocarbene (Schug et al., 1979;
Taylor et al., 1991), being faster than the C±Cl bond
®ssion to form dichloromethyl according to reaction (2).
Dichlorocarbene, however, could not be detected as a
scavenging product in the chloroform-containing ben-
zene ¯ame. CCl2 is well known to be a rather unreactive
Formation of substituted chlorobenzenes. According
to Tables 4 and 6, chlorophenols and carbon-substituted

446
J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
chlorobenzenes were detected in much lower concen-
trations in the benzene/chloroform ¯ame than in the
corresponding chlorobenzene-containing ¯ame. The
relative abundance of the chlorinated compounds com-
pared to the non-chlorinated molecules is typically
about 1%, while it has been at least about 10% in the
corresponding chlorobenzene case. The concentration of
chlorobiphenyls is again increased by a factor of 2, due
to the statistical reasons described above. One cannot
clearly decide whether these products are formed as
consecutive products starting from chlorobenzene, or as
chlorination products starting from the respective non-
chlorinated aromatics. However, the abstraction of
chlorine from chloroform by the non-chlorinated aryl
radical according to reaction (4) seems to be the most
probable route to these chlorinated compounds, which is
a completely di€erent pathway compared to the chlo-
robenzene case. In the chloroform-containing ¯ame,
even small amounts of chloroindenes became detectable
(also ꢁ1% of indene), which could not be clearly iden-
ti®ed in the benzene/chlorobenzene ¯ame, and which are
therefore presumably formed also by reaction (4), i.e.,
chlorination of indenyl.
able to add acetylene to form chloroacenaphthalenes),
whereas the reaction with ¯uorene yields 9-chlorophe-
nanthrene (Busby et al., 1971) as the major product (see
reaction scheme in Fig. 5). Second, the high concentra-
tion of free chlorine radical in the zone of beginning
PAH growth in the benzene/chloroform ¯ame may also
contribute to the formation of chlorinated PAHs, either
by recombination with PAH radicals or by chlorination
of the respective stable PAHs.
4. Conclusions
Several chlorinated organic compounds (CH3Cl, t-
C4H9Cl, C6H5Cl, CHCl3) were employed as additives to
fuel-rich benzene ¯ames in order to investigate their
ability to form (oxygenated) chloroaromatic compounds
under homogeneous ¯ame conditions. The mono-chlo-
rinated aliphatic additives are decomposed quickly and
give rise to only a small extent of chlorination. In con-
trast, chlorobenzene, containing aromatic-bound chlo-
rine, is much more stable within ¯ame conditions, even
more than benzene at low and moderate temperatures
(<1500 K). Consequently, several mono±chloroaromat-
ics became detectable, which are assigned to be consec-
utive products of chlorobenzene, whereas chlorination
reactions can be neglected as possible pathways to
chloroaromatics.
Routes to chlorinated PAHs. For the chlorinated
PAHs, it was found that the relative abundance of these
compounds compared to the respective non-chlorinated
PAHs is similar in the ¯ames with both additives,
chloroform and chlorobenzene. In the chloroform-con-
taining ¯ame, chlorinated PAHs with 14 carbon atoms
could be detected in addition to those from the benzene/
chlorobenzene ¯ames. The most probable structure is
represented by chlorophenanthrenes. These rather sur-
prising observations can be reasonably explained by the
following pathways to chlorinated PAHs in the chloro-
form-containing benzene ¯ame: ®rst, dichlorocarbene is
known to react with ®ve-membered, cyclopentadiene-
type rings to form mono-chlorinated ring expansion
products (dÕAmore and Bergman, 1971). For example,
the gas phase reaction with indene mainly leads to 2-
chloronaphthalene (Busby et al., 1971) (the latter being
By adding chloroform to a benzene ¯ame, the largest
yield and the widest spectrum of chloroaromatic com-
pounds was observed in the presented series of additives.
The triply chlorinated additive, which is again less stable
than benzene, o€ers several pathways to aromatic
chlorination. Even dichlorinated products have been
detected in higher concentrations than in the corre-
sponding chlorobenzene-containing ¯ame, which is due
to reaction pathways involving CCl2 and CHCl2 radi-
cals. It has to be noticed, however, that only a small
amount of the chlorine-containing fuel compound is
converted into chloroaromatic by-products in these fuel-
Fig. 5. Formation of chlorinated PAHs in the benzene/chloroform (96.7/3.3) ¯ame.

J. Burfeindt, K.-H. Homann / Chemosphere 42 (2001) 439±447
447
Hausmann, M., Hebgen, P., Homann, K.-H., 1992. Symp.
(Int.) Combust. 24, 793±801.
rich benzene ¯ames. The maximum yield is about 1% for
t-butyl chloride, 2% for chlorobenzene, and 3% for
chloroform. As chlorinated aliphatic compounds could
be detected only as traces, one can conclude that the
main part of the initial chlorinated fuel is decomposed
rapidly by forming hydrogen chloride, which is too inert
to give rise to any chlorohydrocarbon formation.
Hausmann, M., Homann, K.-H., 1995. Ber. Bunsenges. Phys.
Chem. 99, 853±862.
Hausmann, M., Homann, K.-H., 1997. Ber. Bunsenges. Phys.
Chem. 101, 651±667.
Karra, S.B., Gutman, D., Senkan, S.M., 1988. Combust. Sci.
Technol. 60, 45±62.
Kern, R.D., Xie, K., Chen, H., 1992. Combust. Sci. Technol.
85, 77±86.
Acknowledgements
Lindstedt, R.P., Skevis, G., 1994. Combust. Flame 99, 551±561.
Louw, R., Dijks, J.H.M., Mulder, P., 1984. Recl. Trav. Chim.
Pays-Bas 103, 271±275.
The authors gratefully acknowledge the ®nancial
support by the Volkswagen-Stiftung and the Fonds der
Chemischen Industrie.
Louw, R., Rothuizen, J.W., Wegman, R.C.C., 1973. J. Chem.
Soc. Perkin Trans. II, 1635±1640.
Manion, J.A., Dijks, J.H.M., Mulder, P., Louw, R., 1988. Recl.
Trav. Chim. Pays-Bas 107, 434±439.
Manion, J.A., Louw, R., 1990a. J. Chem. Soc. Perkin Trans II,
551±557.
References
Manion, J.A., Louw, R., 1990b. J. Phys. Chem. 94, 4127±4134.
McKinnon, J.T., Howard, J.B., 1990. Combust. Sci. Technol.
74, 175±197.
Ahrens, J., Kovacs, R., Shafranovskii, E.A., 1994. Ber. Buns-
enges. Phys. Chem. 98, 265±268.
Bachmann, M., Wiese, W., Homann, K.-H., 1995. Combust.
Flame 101, 548±550.
Mulder, P., Louw, R., 1986. J. Chem. Soc. Perkin Trans. II,
1541±1550.
Bachmann, M., Wiese, W., Homann, K.-H., 1996. Symp. (Int.)
Combust. 26, 2259±2267.
Mulder, P., Louw, R., 1985. J. Chem. Soc. Perkin Trans. II,
1135±1141.
Bittner, J.D., Howard, J.B., 1981. Symp. (Int.) Combust. 18,
1105±1116.
Mulholland, J.A., Akki, U., Yang, Y., 1996. In: Proceedings of
the International Conference on Incineration and Thermal
Treatment Technology, University of California, Irvine,
pp. 439±444.
Born, J.G.P., van der Wart, H.W.A., Mulder, P., Louw, R.,
1993. Recl. Trav. Chim. Pays-Bas 112, 262±270.
Busby, R.E., Iqbal, M., Langston, R.J., Parrick, J., Shaw,
C.j.G., 1971. J. Chem. Soc. Chem. Commun., 1293±1294.
Mulholland, J.A., Saro®m, A.F., Sosothikul, P., La¯eur, A.L.,
1993. Combust. Flame 92, 161±177.
Pedersen, J.R., 1991. Chemosphere 22, 929±937.

Canic, T., 1999. Thesis, Technische Universitat Darmstadt.
Cole, J.A., Bittner, J.D., Longwell, J.P., Howard, J.B., 1984.
Combust. Flame 56, 51±70.

Pedersen, J.R., Kallman, B., 1991. Chemosphere 22, 67±76.
Rao, V.S., Skinner, G.B., 1984. J. Phys. Chem. 88, 5990±5995.
Rao, V.S., Skinner, G.B., 1988. J. Phys. Chem. 92, 2442±2448.
Ritter, E.R., Bozzelli, J.W., Dean, A.M., 1990. J. Phys. Chem.
94, 2493±2504.
Cui, J.P., He, Y.Z., Tsang, W., 1989. J. Phys. Chem. 93, 724±
727.
Cullis, C.F., Manton, J.E., 1958. J. Chem. Soc. Faraday Trans.
54, 381±389.
Rouzet, G., Schwartz, D., Gadiou, R., Delfosse, L., 1999.
Presented at the Joint Meeting of the British, German, and
French Sections of the Combustion Institute, Nancy,
France.
dÕAmore, M.B., Bergman, R.G., 1971. J. Chem. Soc. Chem.
Commun., 461±462.

Eklund, G., Pedersen, J.R., Stromberg, B., 1987. Chemosphere
16, 161±166.
Schug, K.P., Wagner, H.G., Zabel, F., 1979. Ber. Bunsenges.
Phys. Chem. 83, 167±175.
Field, L., Banks, C.H., 1975. J. Org. Chem. 40, 2774±2779.
Fields, E.K., Meyerson, S., 1966. J. Am. Chem. Soc. 88, 3388±
3389.
Searles, S., Wann, R.E., 1965. Tetrahedron Lett. 33, 2899±2902.
Taylor, P.H., Dellinger, B., Tirey, D.A., 1991. Int. J. Chem.
Kin. 23, 1051±1074.
Frenklach, M., 1990. Combust. Sci. Technol. 74, 283±296.
Griesheimer, J., 1997. Thesis, Technische Hochschule Darms-
tadt.
Venkat, C., Brezinsky, K., Glassman, I., 1982. Symp. (Int.)
Combust. 19, 143±152.
Westbrook, C.K., 1982. Symp. (Int.) Combust. 19, 127±141.
Griesheimer, J., Homann, K.-H., 1998. Symp. (Int.) Combust.
27, 1753±1759.