884 PCBs AND ASSOCIATED AROMATICS
hydroxyl anions the radiolysis reaction requires very much
larger irradiation doses and hence very much longer reac-
tion times. Under these circumstances the concentration of
hydroxyl radicals which are produced is similar to the con-
centration of PCB radicals and hydroxylation occurs to give
a polychlorinated phenol. Intramolecular cyclization of the
hydroxylated intermediates can occur as discussed earlier to
yield PCDDs.
As a method for the PCB decontamination of oils the
process is impractical but it serves to illustrate how the
reaction conditions may induce rather than reduce a potential
hazard.
The analysis of polychlorinated non-oxygenated toxic
compounds is made extremely difficult because of the lack
of suitable standards. The presence of such classes of com-
pounds as PCQs, PCBPs, PCPYs and PCCYs has neverthe-
less been established when askarel has been degraded under
anaerobic conditions.
Sodium dispersions are effective in the dechlorination
of PCBs and were first applied by Japanese workers in 1973
who used it to decontaminate kerosene extracts of PCB
contaminated sewage sludge. The method was applied in
the decontamination of electrical insulating oils by Webber
et al.,^60 who found that even high concentrations of askarels
would react to give a non-toxic polyphenyl sludge. The
mechanism was investigated by Pilgrim and Webber^61 and
Webber and Wilson^.^62 It was postulated that chlorinated
biphenyl radical anions are produced by interaction of
the PCB molecule in solution with metallic sodium in sus-
pension. The radical anion eliminates chlorine as chloride
to form a chlorobiphenyl radical which, in turn, abstracts
available hydrogen from oil components to yield a partially
dechlorinated PCB. The reaction continues in the presence
of a stoichiometric excess of sodium until total dechlorina-
tion is achieved. The reaction sequence is discussed in more
detail in the section below on PCB disposal methods.
In the absence of a readily available quantity of
abstractable hydrogen the biphenyl radicals tend to couple
and form quaterphenyls and higher polyphenyls. This reac-
tion is relatively less likely to occur that the dechlorination
reaction when an excess of sodium is present and conse-
quently the quaterphenyls which are formed either have few
remaining chlorine atoms on the rings or are completely
dechlorinated.
Quaterphenyls and hexaphenyls are slightly soluble in
mineral insulating oil and become totally dechlorinated.
The sludge is predominantly hexa- and polyphenyl com-
pounds. Work done by Brown and Webber^62 on the NMR
of poly merized biphenyls has indicated that electrical con-
ductivity through the polymer chains is dependent upon the
extent of electron orbital overlap between the ring constitu-
ents of the polymer. Thus, when a small excess of alkali
metal reagent is maintained in contact with insulating oil
for a long time, in order to minimize the use of reagent for
example, polyphenyls are produced which are dissolved in
the oil. The power factor of the processed oil then increases
dramatically to 10% (ASTM D924) compared with
that of useful oil of 0.1%. An increase in power factorcorresponds to an increase in the concentration of polar
constituents in the oil.
The concentration of polyphenyls in the oil and the extent
of their chlorination was not investigated, but, since PCQs
are not expected to be more toxic than PCBs, and because
they are only produced in trace quantities, the practical limi-
tation to their presence lies in the power factor of the product
oil rather than their toxicity.
In 1970 Vos 79 (Vos et al., 1970) showed a correlation
between the toxic effects of European PCBs and the con-
centration levels of PCDFs. The major PCDF components
contained in Yusho oil were the highly toxic 2,3,7,8-tetra-
CDF and 2,3,4,7,8-penta-CDF. The relative concentrations
of the PCDF isomers present in Yusho oil and in two sam-
ples of used heat exchanger PCBs (Kanechlor KC 400 and
Mitsubishi-Monsanto T 1248) were found to be strikingly
similar^74 (Nagayama et al., 1976). This fact underscores the
findings of other workers in this field that there is a connec-
tion between the toxicity of degraded PCB fluids, including
tri-/tetrachlorinated benzene/PCB blends, and the concentra-
tion of PCDFs. The overall toxicity of the fluid may then
be attributable, as Vos indicates, to the two particular PCDF
isomers mentioned earlier.
In addition to the formation of PCDFs and PCDDs from
PCBz/PCB fluids there are also several classes of polychlo-
rinated aromatic hydrocarbons which may be contained as
impurities or degradation products of PCBs. These classes
are shown in Figure 22 and include:- polychlorinated terphenyls (PCTs),
- polychlorinated quaterphenyls (PCQs),
- polychlorinated quaterphenyl ethers (PCQEs),
- polychlorinated naphthalenes (PCNs),
- polychlorinated biphenylenes (PCPs),
- polychlorinated pyrenes (PCPYs) and
- polychlorinated chrysenes (PCCYs).
Polychlorinated quaterphenyls were found as contami-
nants of Yusho and Yu-Cheng oils and for a long time were
an unrecognized interference in the quantitation of PCB con-
tamination in these incidents. Studies 88 (Hori et al., 1982)
show that the toxic effects of PCQs are similar to PCBs.
Polychlorinated quaterphenyl ethers (PCQEs) have
been identified in Yusho and Yu-Cheng oils. This class of
compounds does not seem to have been tested for toxicity
but structure/activity relationships do not predict that they
should be particularly toxic. Nevertheless, cyclization to a
phenyl substituted dibenzofuran possibly could yield a toxic
product. The likelihood of the cyclization reaction provides
a measure for the concern to be accorded the presence of
ether precursors.
Polychlorinated naphthalenes (PCNs) have been char-
acterized in commercial PCBs and may have resulted from
the presence of naphthalene as an impurity in the biphenyl
raw material used in the manufacturing process. Typical
concentrations in PCB fluids are less than 600 ppm. Results
obtained by 89 (Kimbrough, 1981); 90 (Goldstein, 1980)C016_003_r03.indd 884C016_003_r03.indd 884 11/18/2005 1:12:33 PM11/18/2005 1:12:33 PM