198
M.K. Cieplik et al. / Chemosphere 40 (2000) 195±199
quite substantial degrees of conversion with still well-
controlled levels of the target reactant species. The
output of DBF comprises ca. 2.5% (mol) on the ap-
parent converted phenol.
To compare with product levels in absence of acety-
lene, a temperature in between those of experiments (II)
and (III) was chosen. In experiment (IV) only traces of
acetylene were formed as a result of the decomposition
of aromatics. Phenol shows an expected degree of con-
version. The DBF output is in an absolute sense higher
than in both experiments with acetylene, (II) and (III),
and it comprises ca. 4% (mol) on converted phenol. On
the whole, in experiments (I)±(IV) both BF and DBD
were below the detection limit, which is equivalent to ca.
Scheme 3. Reaction of phenyl radicals with oxygen and
acetylene.
1
1
1994); logꢀka; M
s
7:9 (Fahr and Stein, 1988).
Note that even if the rate constants of the two reactions
were equal, 10% (105 ppm), of oxygen in a combustor
would make reaction with acetylene even at 100 ppm
practically impossible. Hence it is not surprising that
phenylacetylene could not be detected in the liquid
samples.
5
10 mmol/h.
To learn more about the fate of DBF ± expectedly of
about the same reactivity as benzene ± experiment (V)
was conducted, using benzene/DBF as the substrate.
Both benzene and DBF reacted to a limited degree. The
origin of the small amount of phenol can be either from
DBF or benzene. Interestingly BF is now positively
identi®ed at levels of ca. 2% (mol) on the converted
DBF.
A likely conclusion will hold for analogous aryl-type
radicals, such as that postulated in Scheme 1 (BF minus
hydrogen atom).
In our experiments (I)±(IV) BF, if present at all, was
just at the detection limit. If it were an intermediate for
DBF, this would imply that its reactivity is very high,
much higher than that of phenol, which is unrealistic.
Rather, the reverse is true: BF is a degradation product
of DBF. This is substantiated by the results of experi-
ment (V).
4. Discussion
The key steps of the Huang±Buekens model imply
altogether three acetylene molecules to add to aromatic
species in order to form ®nal products. Logically this
pathway should dramatically gain in importance with
increasing acetylene concentration. In our case, the ¯ow
of acetylene was ca. 1% (mol) of the total gaseous ¯ow
and also about equal to that of benzene. This is a value
by far exceeding real concentrations in exhaust gases
(ppm level, Lemieux and Ryan, 1999). Consequently, if
formation of dioxins by reactions as depicted in Schemes
1 and 2 would play even a limited role in real combus-
tion, it should have been clearly observable under our
conditions. However, our data show that acetylene has
no promotive eect at all on the formation of DBF as
compared to the runs without acetylene.
In summary, we conclude that even in the presence of
acetylene, DBF arises via recombination of two phenoxy
radicals, as shown in the simpli®ed Scheme 4. The
1
1
overall rate constant is ca. 108:3
M
s
at 530°C
(Grotheer and Louw, 1998). In this respect, the Huang±
Buekens model is too simple, and it underestimates the
pseudo-equilibrium concentration of phenoxy radicals.
In slow combustion, species like hydroperoxy radicals
(HOO) are always important (Scheme 5) Dryer and
Sawyer (1997) and their participation will lead to the
increase of the concentration of phenoxy radicals. Be-
cause of the bimolecular nature of the reaction as de-
picted in Scheme 4, a 30-fold increase in the phenoxy
radical concentration would enhance the rate of DBF
formation by three powers of ten.
Moreover dibenzo-p-dioxin has not been observed in
any case, underscoring the unimportance of such mul-
tiple acetylene additions. Furthermore o-benzoquinone
(BQ), the proposed intermediate (Scheme 2) has been
shown to be thermally unstable with respect to loss of
CO (Schraa et al., 1994).
While the Huang±Buekens mechanism cannot be
valid under the conditions of slow, homogenous com-
bustion with excess of oxygen, one may argue that reac-
tions as depicted in Scheme 1 can be signi®cant in real
combustion, e.g. in oxygen-depleted ¯amefronts or
It is known that phenyl radicals react with O2 irre-
versibly, with a rate constant close to the collision fre-
quency (Frank et al., 1994; Sommeling et al., 1993).
Also, phenyl radicals react with acetylene, to give phe-
nylacetylene (Frenklach et al., 1985; Fahr and Stein,
1988). Both reactions are depicted in Scheme 3.
Overall rate parameters for both reactions at
1
1) 8.8 (Frank et al.,
Scheme 4. Recombination of phenoxy radicals.
T 530°C, are: log (ko,M
s