Reactions of diphenyl ether with chlorine and bromine atoms around 750 K - Relevance for gas-phase 'dioxin' formation

Article abstract of DOI:10.1002/(SICI)1099-0690(199901)1999:1<261::AID-EJOC261>3.0.CO;2-I

The title reactions have been studied to scrutinize rate data recently inferred for the two reverse steps - reaction of phenoxy radicals with chlorobenzene and bromobenzene which were at variance with commonly accepted model values. Both with chlorine and bromine atoms, splitting to halobenzene and phenoxy radical was found to occur in competition with abstraction of o-, m-, p-hydrogen atoms. On this basis, the displacements of Cl and Br from the benzene ring by phenoxy radicals must have activation energies above 20 kcal/mol, and are therefore slow. As a consequence, formation of 'dioxins' from halogenated phenols, in (slow) combustion, should proceed by combination of two (halo)phenoxy radicals rather than by displacement of (ortho-)halogen in a halophenol molecule.

Full text of DOI:10.1002/(SICI)1099-0690(199901)1999:1<261::AID-EJOC261>3.0.CO;2-I

FULL PAPER
Reactions of Diphenyl Ether with Chlorine and Bromine Atoms Around
750 K – Relevance for Gas-Phase “Dioxin” Formation
Izabela Wiater[a,b] and Robert Louw*[a]
Keywords: Dioxins / Gas-phase reactions / Diphenyl ether / Hydrogen abstraction / ipso Substitution
The title reactions have been studied to scrutinize rate data
this basis, the displacements of Cl and Br from the benzene
recently inferred for the two reverse steps – reaction of ring by phenoxy radicals must have activation energies
phenoxy radicals with chlorobenzene and bromobenzene –
which were at variance with commonly accepted model
values. Both with chlorine and bromine atoms, splitting to
halobenzene and phenoxy radical was found to occur in
competition with abstraction of o-, m-, p-hydrogen atoms. On
above 20 kcal/mol, and are therefore slow. As
a
consequence, formation of “dioxins” from halogenated
phenols, in (slow) combustion, should proceed by
combination of two (halo)phenoxy radicals rather than by
displacement of (ortho-)halogen in a halophenol molecule.
Introduction
S&T’s notation). PD was believed to react rapidly to a
(chlorinated) dioxin D by loss of HCl (Reaction 2).
The origin and formation mechanisms of polychlorinated
dibenzo-p-dioxins and furans (“dioxins”, PCDD/F) con-
tinue to be heavily debated and researched. Their demon- PD Ǟ D ϩ HCl
P• ϩ P Ǟ PD ϩ Cl
(1)
(2)
strated high toxicity especially to some test animals has led
Step 1, with an estimated activation energy of 26 kcal/
mol, should give perceptible rates only at higher tempera-
tures, but then the phenoxy radicals will decompose. As
a result, the “S&T” gas-phase model can only account for
low dioxin levels throughout.
to stringent regulations in parts of Europe, the USA, and
in Japan. Meanwhile, the real dioxin deposition is much
bigger than the estimated total emission from known
[7]
[1]
sources and the major pathways in dioxin formation are
still ill-defined.
Since in practice dioxin emissions exceed the levels pre-
dicted by S&T, the process was reinvestigated by Sidhu and
co-workers.[ They studied the thermal decomposition of
Combustion processes, especially those of Municipal
Solid Waste, are an important source of PCDD/F.[2] Since
8]
it became known that the levels of PCDD/F increased con-
2
,4,6-trichloro- and 2,4,6-tribromophenol in excess of air in
siderably upon the passage of the primary combustion gases
through Air Pollution Control devices (ESP, baghouse)[3]
a flow reactor. The experiments were done over a tempera-
ture range of 300Ϫ800°C with minimal contribution of sur-
face reactions. Maximum yields of PCDD were obtained
around 600°C. In order to adjust the “S&T” model to those
observations they decreased the activation energy of Reac-
tion 1 to 19.5 kcal/mol for chlorinated phenols, and to 8.8
kcal/mol for the corresponding brominated case (to account
for a factor 500 higher value). On this basis they suggested
that gas-phase formation Ϫ by the S&T mechanism Ϫ
could make a significant contribution to the observed
PCDD/F yields in full-scale incinerators.
Both studies ignored an alternative, viz. condensation by
combination of two (chloro)phenoxy radicals. It has been
known for a long time that thermal conversions of phenol
lead to dibenzofuran, and closer investigations have shown
that this involves the combination of two (unchlorinated)
much research effort has been devoted to unravel pertinent
mechanisms. Two distinct pathways have been proposed: (i)
So-called “de novo” synthesis from carbonaceous materials
on solid surfaces (ashes) in a sequence of steps catalyzed by
[2,4,5]
metals, on a long time scale (up to hours).
(ii) Forma-
tion from “precursors”, primarily (chloro)phenols. Model
experiments have shown that ashes can create especially
PCDD efficiently in the temperature region of 300Ϫ400°C
[5]
within seconds. However, little PCDF is found, meaning
that this precursor scenario is too simple. Other precursors
may play a part, too.
The gas-phase formation has already been discussed by
[6]
Shaub and Tsang (S&T) in 1983 with trichlorophenol as
an example; a model mechanism involving 13 gas-phase
(radical) reactions was proposed. Displacement of an ortho-
8
phenoxy radicals with an overall rate constant of 10
chlorine in the (chloro)phenol P by a (trichloro)phenoxy
Ϫ1 Ϫ1
[9,10]
•

s
or higher.
a hydrogen atom in ortho position react analogously. In
mixed” situations, with both H and Cl present in ortho
Chlorinated phenols still containing
radical P (Reaction 1) was advanced as a key step, leading
to the corresponding ortho-hydroxydiphenyl ether (PD in
“
[
a]
position, as in o-Cl-phenol, the appropriate DCDFs are
Center for Chemistry and the Environment, Leiden Institute of
Chemistry, Gorlaeus Laboratories, Leiden University,
P. O. Box 9502, NL-200 RA Leiden, The Netherlands
E-mail: R.Louw@chem.leidenuniv.nl
On leave (through TEMPUS grant IMG-97-Pl-2135) from:
Jagiellonian University, Faculty of Chemistry, Cracow, Poland
formed, but also (Cl-free) dibenzodioxin, the “S&T”-like
[9]
product. Whatever the details of the respective pathways,
[b]
it is reasonable to accept radical/radical combination as a
key step in both types of reaction, formation of dibenzofu-
Eur. J. Org. Chem. 1999, 261Ϫ265

WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1434Ϫ193X/99/0101Ϫ0261 $ 17.50ϩ.50/0
261

I. Wiater, R. Louw
FULL PAPER
ran Ϫ with formal loss of the elements of water Ϫ and of
dibenzodioxin, with net overall loss of two chlorine atoms.
•
•
P ϩ P Ǟ “dioxin”
(3)
Returning to the example of 2,4,6-TCP[8] it will be clear
that the result of a competition between Reactions 1 and 3
as key steps Ϫ with greatly different rate constants Ϫ is
•
governed by the (actual) P /P ratio. Depending on the tem-
perature, for Reaction 1 to “win”, the concentration of
(chloro)phenol should be at least some five powers of ten
larger than that of the chlorophenoxy radical. Useful (ex-
perimental) data on this are lacking, however.
In order to check the kinetics of the S&T model, Louw
[11]
and Grotheer
have recently investigated the co-combus-
tion of phenol and chlorobenzene, to obtain rate param-
eters for Reaction 4, the simplest analogue of Reaction 1.
PhO ϩ PhCl Ǟ PhOPh ϩ Cl•
•
(4)
Figure 1. Enthalpy diagram for Reaction 4 and related processes
The direct reaction product, diphenyl ether (PhOPh), is
[11][12]
(
energy data in kcal/mol); TS (a) according to ref.
; TS level
[
8]
more stable than phenolic compounds like PD, and can be (b): modeled value of Sidhu et al. for 2,4,6-Cl C H O(H)
3
6
2
easily determined. Moreover, the products from Reaction 4
can be distinguished from those of Reaction 3. Since
unchlorinated phenol was used as a source of phenoxy rad-
icals the only one “dioxin” product was dibenzofuran (Re-
action 5).
•
•
PhO ϩ PhO Ǟ DF
(5)
(6)
PhO ϩ PhBr Ǟ PhOPh ϩ Br•
•
Later on the study had been extended to bromoben-
zene.[ Since HBr, produced from the slow combustion of
PhBr, can affect the overall rate coefficient of the DF for-
mation from phenoxy radicals the competition between
chloro- and bromobenzene was investigated as well. Re-
markably Reaction 6 with PhBr was not much faster than
Reaction 4 with PhCl. The activation energy of Reaction 4
was derived to be 24.5 kcal/mol (corresponding well with
E estimated in the S&T model). For Reaction 6 E was
12]
a
a
found to be 21.4 kcal/mol, much higher than the 8.8 kcal/
[8]
mol proposed by Sidhu and co-workers to comply with
the “S&T” model. This let Grotheer and Louw conclude
that in both studies Reaction 1 was insignificant compared
to Reaction 3 in the gas-phase formation of dioxin.
The energy diagrams (Figures 1 and 2) are helpful in a (energy data in kcal/mol); TS (a) according to ref.[ ; TS (b) this
further analysis of the possible importance of the parent
Figure 2. Enthalpy diagram for Reaction 6 and related processes
12]
work; TS level (c): modeled value of Sidhu et al.[ for 2,4,6-
8]
3 6 2
Br C H O(H)
Reactions 4 and 6 of the S&T model.
For the reaction of PhOPh with Cl or Br there are two
•
•
possibilities: (i) splitting, by ipso addition, to phenoxy rad- thermicity of the H abstraction together with the 21.4 kcal/
icals and chlorobenzene (Reaction Ϫ4) or bromobenzene mol barrier for Reaction 6 would again suggest pathways
(Reaction Ϫ6); (ii) formation of halogenated phenoxyben- (i) and (ii) to be competitive.
zenes, by H abstraction.
The gas-phase chlorination of benzene derivatives (ArH)
[13]
If the data of Sidhu et al. apply, chlorine atoms would has been studied by Kooyman and co-workers.
split the ether rather than abstract hydrogen; bromine mal process proceeds by H abstraction, Cl ϩ ArH Ǟ HCl
atoms and PhOPh should give bromobenzene and phenoxy ϩ Ar as key step, the latter species giving ArCl by reaction
radicals exclusively, with no chance for hydrogen abstrac- with Cl (or another appropriate Cl donor).
tion. With the data of Grotheer and Louw, however, Cl
can follow both pathways (i) and (ii). With Br the endo- dibenzofuran
This ther-
•
•
[14]
With PhOPh
2
•
the reported main products are m- and p-ClC H OPh and
6
4
•
[15]
(the o-phenoxyphenyl radical cyclizes in-
262
Eur. J. Org. Chem. 1999, 261Ϫ265

Reactions of Diphenyl Ether with Chlorine and Bromine Atoms Around 750 K
FULL PAPER
stead of giving o-ClC H OPh). Formation of chloroben- runs of PhOPh/PhCN indeed gave negligible outputs of DF
6
4
zene, if any, was not mentioned. Gas-phase bromination is and of course no chlorobenzene.
[13]
known to follow the same mechanism,
diphenyl ether has not been reported on.
but in this case
At lower to trace levels, a large number of other products
are also formed. These include typical products from the
In order to obtain a better insight we have re-investigated perchloro C /C mix, e.g. hexachlorobutadiene, hexachloro-
1
2
the gas-phase chlorination of PhOPh with emphasis on the propene, and hexachlorobenzene. Trichloromethylbenzene
possible importance of Reaction Ϫ4. In order to prove or illustrates the intermediacy of CCl radicals. The most likely
3
disprove the overall high barrier for Reaction 6 Ϫ depicted explanation for formation of this compound is ipso-substi-
•
as TS1 in Figure 2 for reasons to follow Ϫ the gas-phase tution of PhOPh to give PhO and PhCCl . Another ex-
3
reaction of PhOPh with bromine has also been studied. For ample of ipso-substitution is the observation of two isomers
proper comparison, conditions (temperature, dwell times, of PhOC H Ph, with abundance of about 5% on those of
6
4
concentrations of arenes) have been designed close to those m- and p-ClC H OPh. It is very likely that the correspond-
6
4
[11,12]
of the “forward” slow combustion.
The desired tem- ing m-(p-)PhOC H radicals will have reacted with PhOPh,
6 4
peratures are quite high for halogenation, however; reac- with ipso displacement of PhO as the productive channel.
tions involving Br proceeded well at 500°C, but Cl then Note that with CCl /C Cl as chlorination agent for ben-
2
2
4
2
6
•
is too aggressive; in the latter case C Cl , smoothly generat- zene derivatives ArH, part of the aryl radicals Ar generated
2
6
[14]
[14]
ing Cl atoms
has been employed.
will react with ArH to create biaryls ArϪAr.
With
PhOPh the corresponding products, isomers of PhOC H -
6
4
C H OPh were too involatile to be seen in our GC analysis.
6
4
Results and Discussion
The ipso product PhOC H Ph is a proper manifestation of
6
4
this type of reaction.
Splitting (Reaction Ϫ4) implies formation of phenoxy
(i) Chlorination
radicals next to chlorobenzene. Their fate is not yet sure.
With hexachloroethane as a source of chlorine atoms and
•
•
The formation of phenol by recombination of PhO and H
tetrachloromethane as chlorine donor to reactive (aryl) rad-
icals if formed, a series of runs have been conducted at tem-
peratures of 460Ϫ480°C. At 500°C the product contained
quite a lot of sooty material Ϫ in accordance with earlier
radicals is fast (k ϭ 2·1011

Ϫ1 Ϫ1[17]
s
), but the concentration
of atomic hydrogen will be extremely low. In fact no phenol
could be detected, showing that other pathways for acquir-
ing hydrogen are negligible as well.
[14]
observations on this chlorination agent.
collected in Table 1.
Key data are
Condensation of two phenoxy radicals (Reaction 5) leads
•
to DF, the same product as that from o-PhOC H cycliza-
The formation of C Cl (ca. 10% on C Cl at 460°C, ca.
6
4
2
4
2
6
•
tion. PhO can also become chlorinated before conden-
sation.[ At least two isomers of DCDF were produced,
but at a very low level (less than 1% on DF, each). MCDF
was not seen at all. One of the products with low abundance
was identified by the MS search library as 2-chlorobenzoyl
chloride. Without detailing possible mechanistic pathways,
we suggest this to be a product stemming from reaction
2
0% at 480°C) shows that (atomic) chlorine is generated;
18]
the identification of m- and p-ClC H OPh together with
6
4
comparable amounts of dibenzofuran demonstrates that
overall) chlorination by H-abstraction does occur. Our
product ratios m-/p- ϭ 4Ϫ5, and [DF] ഠ [m- ϩ p-], resemble
(
[15]
those of Engelsma
for gas-phase chlorination with Cl in
2
the same temperature range. Interestingly, the formation of
chlorobenzene, at substantial levels (about one-half of DF)
between phenoxy and CCl radicals.
3
•
•
shows that on a per-site basis, splitting of PhOPh by Cl ,
In sum, the exact fate of PhO is not known; but should
Reaction Ϫ4, is as fast as H abstraction. At this point it is DF have been formed with a high efficiency, the maximum
•
•
important to note that PhOPh itself is thermally very stable. contribution via PhO ϩ PhO combination would be 0.5
The Arrhenius parameters for bond fission to phenyl and that of PhCl, or ca. 25% of the DF output.
Ϫ1 Ϫ1
phenoxy radicals (log A ϭ 15.5 
s
; Ea ϭ 75 kcal/
The output of HCl is several times larger than the
[16]
Ϫ6 Ϫ1
mol ) entail k Յ 10
s
at the temperatures used. Blank amount of Cl which must have become free in the formation
Table 1. Chlorination/chlorinolysis of diphenyl ether[a]
Exp.
no.
Temp.
[°C]
Inflow[b]
6
Outflow[c]
C
2
Cl
PhCN
C
2
Cl
6
C
Cl
2 4
PhCl
DF
m-/p-ClC
6
H
4
OPh
HCl
C1
C2
C3
C4
C5
C6
C7
460
460
480
480
480
480
480
4.05
3.00
4.05
4.05
4.05
3.00
3.00
10.1
Ϫ
10.1
10.1
10.1
Ϫ
3.80
2.82
3.13
3.42
3.01
2.55
2.59
0.36
0.28
0.84
0.87
0.77
0.75
0.68
0.062
0.067
0.47
0.48
0.50
0.53
0.44
0.15
0.16
0.89
0.93
0.86
0.97
0.83
0.23/0.047
0.27/0.066
0.69/0.18
0.72/0.19
0.62/0.16
0.73/0.19
0.68/0.18
1.56
1.40
5.8
6.5
6.7
4.6
Ϫ
4.4
[
a]
[b]
Run times: 60 min (exp. C1, C3ϪC5) or 85 min (C2, C6, C7); residence time 150 ± 2 s; flow data in mmol/h. Ϫ
2
Other inflows:
263
[c]
N
ϭ 270; PhOPh ϭ 20.1; CCl
4 4
ϭ 20.2 mmol/h. Ϫ Recoveries of PhOPh 75Ϫ92%; PhCN > 95%; CCl > 80%.
Eur. J. Org. Chem. 1999, 261Ϫ265

I. Wiater, R. Louw
FULL PAPER
Table 2. Bromination/brominolysis of diphenyl ether at 500°C[a]
Exp.
no.
Inflow
Diluent
Outflow
[b]
PhOPh
Br
2
PhBr
DF
m-/p-BrC
6
H
4
OPh
HBr
B1
B2
B3
B4
B5
B6
24.7
24.7
24.7
32.1
32.1
16.05
17.4
17.4
17.4
4.6
4.6
4.7
4.5
4.6
4.6
2.3
0.27
0.32
0.31
0.40
0.45
0.17
1.59
1.70
1.63
2.27
2.39
0.94
0.82/0.68
0.95/0.82
0.93/0.80
1.51/0.50
1.62/0.53
0.56/0.27
6.6
6.6
6.5
6.5
6.8
3.6
4.6
2.3
[a]
Residence time 153±2 s; flows in mmol/h. Ϫ [b] In exp. B1ϪB3 benzonitrile was used, in exp. B4ϪB6 p-difluorobenzene.
of C Cl . This merely shows that CCl is also contributing output. Altogether, the safest way to discuss the compe-
2
4
4
to “mineral” chlorine. Likewise the amount of hydrogen in tition between ether splitting by halogen atoms and H ab-
HCl is larger than that removed, displaced, in the phenyl straction is to take one “isomer”: ortho, with its straightfor-
ether products of Table 1. Unspecified condensation reac- ward cyclization to DF. The average ratio of PhBr/DF is
tions (usual in pyrolytic reactions) must be responsible for 0.18, per site it is 0.36 (these values can be up to 10% larger
•
this.
In experiments C1 and C3ϪC5 benzonitrile was co-fed
when the production of DF from PhO is substantial).
•
So, splitting of PhOPh by Br is nearly by a factor of 2
with PhOPh. In those runs the three isomers of ClC H CN slower than hydrogen abstraction from the ortho-position.
6
4
were produced; the ClC H CN output indicated that PhCN
6
4
was 5 times less reactive than PhOPh Ϫ in accordance with
earlier observation.[13]
(iii) Mechanistic Aspects
As can be seen from the ratios of DF/m-/p-ClC H OPh
Table 1), PhCN has no influence on the investigated reac-
tions.
6
4
The per-site rate difference for splitting and abstraction
of ortho-H by Br mentioned above implies that, at 500°C,
the free enthalpy of activation for splitting is Ϫ at most Ϫ
(
•
1
.6 kcal/mol higher than that for abstraction. Possible en-
tropy differences between the two rate-determining steps
(ii) Bromination
may make ∆E somewhat different, but inspection of Figure
a
PhOPh was admixed with benzonitrile or p-difluoroben- 2 shows that our result is close to that of Grotheer and
zene, and Br served as the source of Br atoms. Additional Louw for the forward Reaction 6, substantiating that the
2
experiments were done to check whether benzonitrile can (overall) barrier for displacement of Br from PhBr by PhO•
react with bromine under the prevailing conditions to give is well over 20 kcal/mol. In no way would Sidhu’s value,
bromobenzene. No PhBr was found. PhOPh was 4Ϫ5 times proposed to cover rates of “dioxin” formation from 2,4,6-
more reactive than PhCN. The pattern of isomeric bromo- tribromophenol/phenoxy apply here: Bromine atoms
benzonitriles was similar to that reported in the litera- should then have reacted with PhOPh exclusively by split-
[15]
ture.
In exp. B1ϪB3 the o/m/p ratio of BrC H CN was ting Ϫ and with a much higher rate than for H abstraction
6 4
1
ture.
8:49:33 compared with 23:49:28 at the same tempera- (thence, substitution to ArBr), which should also have left
[15]
benzonitrile untouched in the PhOPh/PhCN competition.
Bond strengths can, will, depend on structure and on (de-
As can be seen in Table 2, the m-/p- ratios of BrC H OPh
6
4
are smaller than in the case of chlorination, but still the grees of) substitution; but differences in bond strengths be-
main organic product was DF, and as in the chlorination tween RϪX and RϪY vary relatively little, certainly if X
yields of DF are about equal to the sum of m- and p- and Y are of like polar character, as in the case for electro-
BrC H OPh. The reason for such small m-/p- ratios (1.2 negative groups Br, Cl, and OR. So Figure 2 will expectedly
6
4
and 3 for the mix with PhCN and C H F , respectively) not change much when applied to the case of tribromo-
6
4 2
may be that some residual Br causes electrophilic “after”- phenoxy/phenol, and the (higher than expected) rate for
2
[
8]
bromination, which will involve the p-position only. Fur- “dioxin” formation must have been due to the (faster)
thermore, if the conversion of Br2 nears completion, the ArO ϩ ArO combinations.
HBr/Br ratio becomes high. Then m- and p-PhOC H rad-
•
•
In Figure 2, TS2 represents the (exothermal) loss of Br•
2
6
4
icals will also react with HBr, regenerating PhOPh. Such from the ipso-adduct radical. Its level will be far below that
possible influences on the output of m- and p-BrC H OPh of TS1. Being kinetically insignificant for the forward Reac-
6
4
•
are much smaller, if present at all, in case of the o-PhOC H4 tion 6, it merely means that ipso addition of Br to PhOPh
6
radical, which unimolecularly reacts to DF.
As has been already discussed in the chlorination case, Cl case (Figure 1), for endothermal loss of Cl from the
some of the DF may have been produced from PhO . In adduct species vice versa, will approach that for PhO ad-
the bromination case the maximal DF yield from 2 PhO
is highly reversible. Note that the analogous barrier in the
•
•
•
•
dition to PhCl. (Both activation barriers have arbitrarily
(0.5 times that of PhBr) comprises only 10% of the total been sketched at the same level.) As a consequence, ipso
264
Eur. J. Org. Chem. 1999, 261Ϫ265

Reactions of Diphenyl Ether with Chlorine and Bromine Atoms Around 750 K
FULL PAPER
•
addition of Cl to PhOPh is expected to be reversible to a liquid samples were injected at a split ratio of 1:15. Absolute con-
centrations were deduced from the peak areas, after calibration by
limited extent only.
injecting standard mixtures of known composition. A Hewlett
Anyway, in the chlorine case our observed product ratios
Packard 5890 GC-MS was used to identify unknown products.
PhCl/DF Ϫ as discussed in section (i) Ϫ entail nearly equal
Outflows of HCl or HBr were quantified by means of a Mettler
rates (activation barriers) for splitting and H abstraction.
DL25 automatic titrator. Exit non-condensable gases were analysed
This is in full accord with the observed E for the forward
Reaction 4 as reported by Grotheer and Louw.
Sidhu’s value of 19.5 kcal/mol inferred from the slow com-
a
using a Packard series 428 GC equipped with FID detector, Car-
boplot 007 column and a methanizer, calibrated by independent
injections of a standard gas mixture. The reaction of chlorine
[11]
Should
bustion of 2,4,6-Cl C H OH apply here, splitting would (atoms) was studied at 460Ϫ480°C by using diphenyl ether ad-
3
6
2
•
[14]
have been by far the major process. In our view, ArO ϩ mixed with tetrachloromethane and hexachloroethane
ArO combination, etc. must have been at least as impor- dence times of about 2 min. Experiments involving bromine atoms
tant as the S&T-type ArO ϩ chloroarene displacements.
As mentioned above for the bromine case, bond strengths
may vary, but the overall endothermicity for the latter type
and resi-
•
•
were carried out at 500°C, with comparable residence times. Br
2
was fed in from an impinger by a calibrated flow of nitrogen. Here
PhOPh was diluted with benzonitrile or p-difluorobenzene, both
less reactive in gas-phase halogenation than diphenyl ether.
[13,14]
•
of reaction (17 kcal/mol for our base case PhO ϩ PhCl)
Purity of Reagents: Tank N 99.99% was supplied by Hoekloos in
2
should not change drastically upon introducing (chlorine)
substituents Ϫ let alone that displacement becomes about
thermoneutral, or even exothermal, as advanced by Bozzeli
To the best of our knowledge there are no known
experimental data to support such a proposal.
a standard cylinder. Diphenyl ether (Merck, pro synthesis), tetra-
chloromethane (J. T. Baker, > 99%), hexachloroethane (Fluka,
pure), benzonitrile (Merck, pro synthesis), p-difluorobenzene
Fluka, pure) were checked by GC to be of adequate (> 99%) purity
and used as such. Reference compounds such as chlorobenzene,
bromobenzene, and dibenzofuran were high-grade (> 99.9) com-
[19]
et al.
(
•
Our results on the PhCl(Br) ϩ PhO systems, and vice
versa, imply that PhO has only a low reactivity in addition mercial products.
•
to benzene derivatives, even in the nearly neutral displace-
ment of Br. The reasons for that are not yet clear, but we
[1]
L. P. Brzuzy, R.A. Hites, Environ. Sci. Technol. 1996, 30,
1797Ϫ1804.
2]
wish to point out that Ϫ while phenoxy-type radicals are of
[
•
K. Tuppurainen, I. Halonen, P. Ruokojarvi, J. Tarhanen, J.
Ruuskanen, Chemosphere 1998, 36, 1493Ϫ1511 and references
cited there.
3]
key importance in antioxidant action Ϫ next to ArO (self
or cross) radical combinations, and “surprisingly fast”
[
•
•
[20]
M. S. Millingan, E. Altwicker, Environ. Sci. Technol. 1993, 27,
ArO ϩ ArЈOH Ǟ ArOH ϩ ArЈO hydrogen transfers,
1
595Ϫ1601 and references cited there.
•
we are unaware of other examples of ArO ϩ molecule reac-
tions, such as with CϭC double bonds.
[4]
R. Addink, K. Olie, Environ. Sci. Technol. 1995, 29, 1425Ϫ1435.
L. Stieglitz, G. Zwick, J. Beck, W. Roth, H. Vogg, Chemosphere
[5]
1
989, 18, 1219Ϫ1226.
[
6]
W. M. Shaub, W. Tsang., Environ. Sci. Technol. 1983, 17,
721Ϫ730.
7]
[
C. Y. Lin, M. C. Lin, J. Phys. Chem. 1986, 90, 425Ϫ431.
S. S. Sidhu, L. Maqsud, B. Dellinger, G. Mascolo, Combust.
Flame 1995, 100, 11Ϫ20.
Experimental Section
[8]
All experiments were conducted in a cylindrical, stirred-tanktype
flow reactor of 680 ml volume placed in an electrically heated oven.
The temperature was controlled by a proportional regulator and
monitored with chromel-alumel thermocouples displayed on digital
thermometers (Therma 1, type ST-861Ϫ107). The upper and lower
ends of the oven were insulated by quartz wool. For further details
[9]
J. G. P. Born, Ph. D. Thesis (in English), Leiden University,
1
992.
[10]
I. W. C. E. Arends, R. Louw, P. Mulder, J. Phys. Chem. 1993,
97, 7914Ϫ7925.
[
11]
H. H. Grotheer, R. Louw, 26th Symp. ( Int.) on Combustion,
Naples, Italy 1996, p. 2405Ϫ2411.
12]
[
[12]
H. H. Grotheer, R. Louw, Combust. Sci. Technol. 1998, 134,
31Ϫ49.
on this type of setup see ref.
Gas flows were regulated by needle
valves and measured with capillary flow meters. Liquid (organic)
reactants were introduced by means of a calibrated motorized syr-
inge pump (B. Braun, Melsungen, Perfusor VI type 871222/0) via
a gas-tight rubber septum and vaporized into the gas flow before
entering the reactor. The entrance and exit tubes of the reactor
were heated by wrapped heating tape to prevent condensation of
the less volatile organics. Condensable products were collected in
two cold traps. The first one contained aqueous potassium hydrox-
ide and toluene, and was cooled with ice. The second trap, with
toluene, was cooled by acetone with liquid nitrogen. For experi-
ments C1ϪC7 bromobenzene was used as an external standard and
for B1ϪB6 chlorobenzene. Aromatic compounds were quantified
with a Hewlett Packard 5890A gas chromatograph with FID using
a CP-SIL5-CB column (50 m ϫ 0.32 mm ID). 1-µL portions of
[13]
E. C. Kooyman, Advances in Free-Radical Chemistry, vol. 1,
chapter 4 (Ed.: G. H. Williams), Logos Press, London, 1965.
W. Dorrepaal, R. Louw, Int. J. Chem. Kinet. 1978, 10, 249Ϫ275.
J. W. Engelsma, E. C. Kooyman, Recl. Trav. Chim. Pays-Bas
[14]
[15]
1
961, 80, 526Ϫ532.
[16]
I. W. C. E. Arends, Ph. D. Thesis (in English), Leiden Univer-
sity, 1993.
17]
[
M. Pecullan, K. Brezinsky, I. Glassman, J. Phys. Chem. 1997,
1
01, 3305Ϫ3316 and references cited there.
[18]
P. M. Sommeling, P. Mulder, R. Louw, Chemosphere 1994, 29,
2015Ϫ2018.
19]
[
J. W. Bozzeli, Y. G. Wu, E. R. Ritter, Chemosphere 1991, 23,
1221Ϫ1232.
20]
[
M. Foti, K. U. Ingold, J. Lusztyk, J. Am. Chem. Soc. 1994,
1
16, 9440Ϫ9447.
Received August 14, 1998
[O98382]
Eur. J. Org. Chem. 1999, 261Ϫ265
265