PCBs AND ASSOCIATED AROMATICS 943
than the original toxic, water insoluble, stable compounds.
The sodium phenolate product can be acidified and safely
separated from the aqueous wash solution containing dis-
solved inorganic salts.
Brunelle and Singleton^148 have reported the reaction
between potassium hydroxide, polyethylene glycol (PEG)
and PCBs. It was found that the complete removal of up to
1% PCBs from transformer oil could be achieved in less than
two hours at 100°C. The advantage of the system over pro-
cess which involve alkali metal organometallic reagents is
that it is not moisture or air sensitive.
The Franklin Institute in Philadelphia, Pennsylvania
has patented the “NaPEG” process for the detoxification of
transformer oils (Lee^149 ). The reagent used is derived from
the reaction of alkali metal with a polyglycol such as poly-
ethyleneglycol with an average molecular weight of about- The preparation can be described in the following
equations:
Na PEG NaO (PEG) H
NaO (PEG) O NaPEG reagent2
2+→ +
+→—The NaPEG agent, like the Vertac reagent and the GE reagent,
is both air and water stable and therefore has a long shelf life
without special handling precautions. Dechlorination of PCBs
in transformer oils can be achieved by heating the reagent and
oil at temperatures between 40°C and 180°C depending upon
the particular reagent chosen and the degree of chlorination
of the PCB. The reaction evolves hydrogen gas and pro-
duces polyhydroxylated biphenyls and sodium chloride. The
effectiveness of the system has been proven on a large scale
(Garland^150 ) by the Philadelphia Electric Company (PECO)
who have subsequently reused the decontaminated oil pro-
duced in a pilot plant. It was observed that air or oxygen was
necessary to make the reagent active towards PCBs in insu-
lating oils and that the use of pure oxygen enhanced the rate
of dehalogenation by a factor of five. This is in contrast to
Brunelle and Singleton’s work described above since they
found that the presence or absence of air (oxygen) had no
effect on the rate or outcome of the reaction. Also, in order
to preserve the oxidative stability of the oil the reactions
were recommended^148 to be conducted under an inert atmo-
sphere. Webber and Wilson’s work^62 on the inhibition of
PCB dechlorination in oxidation inhibited oils suggests that
the cause of the NaPEG reagent inactivity may be that the
DBPC oxidation inhibitor oxidatively dimerizes to produce
the observed yellow solution of stilbenequinone which then
acts as an efficient scavenger of the radicals produced in the
dechlorination steps. The formation of phenoxy radicals in
Webber and Wilson’s work was indicated by ESR spectros-
copy. The presence of peroxy species in the NaPEG process
was also observed.
The structures of naphthalene, anthracene and biphenyl
radical anions with alkali metal gagenions have been inter-
preted by the alkali metal spin density distributions in the
aromatic systems (West^151 ). Webber et al.^152 have utilizedthe chemistry of this organometallic series of complexes to
produce a preferred process for the PCB decontamination of
insulating oil which is unaffected by the presence of oxida-
tion inhibitors and which can be used at moderately low tem-
peratures in a continuous system to yield a dielectric quality,
decontaminated product.
Tiernan et al.^153 at Wright State University have reported
on the use of KPEG reagent on PCDD/PCDF contami-
nated soil at hazardous waste sites. A similar paper has
also been given by DesRosiers^154. A 40 foot mobile trailer
containing a 2700 gallon stainless steel reactor was used
with 600 gallons of KPEG to treat 1400–2000 gallons of
waste material. The mixture was heated to 150°C for 1.5
h. Under optimal conditions the chemistry allows com-
plete dechlorination. A practical destruction efficiency of
99% is observed at reaction temperatures above 70°C
with reaction times of 15 minutes to 1 hour. Interestingly,
bioassays of residues from the reaction were negative.
The cost of the KPEG treatment has been estimated to be
approximately $24.00/gallon i.e., about 10% of the cost of
on-site incineration.UNCONTROLLED REACTIONSThe extreme toxicity of PCDFs and PCDDs discussed ear-
lier makes it imperative that uncontrolled side reactions in
processes used to chlorinate PCBs should be investigated
and steps taken to avoid the generation of concentrations of
compounds of concern.
Uncontrolled reactions during the large scale prepara-
tion of trichlorophenol from tetrachlorobenzene and sodium
hydroxide in ethylene glycol have resulted in runaway con-
ditions on at least six occasions and resulted in the formation
of PCDDs. The Seveso incident, which occurred in 1976,
is the most recent and is probably the best documented.
Ethylene glycol used as solvent in the reaction can undergo
an exothermic base catalyzed polymerization above 180°C
to produce substantial quantities of TCDD from the dimer-
ization of chlorophenol.
If the reaction conditions in a process to destroy askarels
can yield chlorophenols as a side reaction product then the
potential exists for the generation of PCDFs and/or PCDDs.
The reaction mechanisms which explain the formation of the
cyclized products is described in an earlier section.
The dimerization of chlorophenates is a bimolecular
reaction and therefore the products formed should be highly
dependent upon the chlorophenate concentrations. The par-
ticular PCDD isomers formed in the system, and their quan-
tities, will depend upon the relative kinetics of alternative
reaction routes.
The pyrolysis of PCDPEs follows two competitive reac-
tion pathways, viz., dechlorination or ring closure to PCDFs.
The cyclization of pre-dioxins is a bimolecular reaction
caused by heating. The final concentration of PCDDs in
a heated chlorobenzene system should ultimately depend
upon the concentration of polychlorinated phenols which are
formed at an intermediate stage.C016_003_r03.indd 943C016_003_r03.indd 943 11/18/2005 1:12:45 PM11/18/2005 1:12:45 PM