934 PCBs AND ASSOCIATED AROMATICS
transformer exploded and sprayed PCB contaminated soot
onto the walls. No PCDFs or PCDDs were detected.
When a transformer is retrofilled to 500 ppm PCB one
would expect that the rate of conversion of PCBs would be
very much less than in the case of an askarel fire. It becomes
increasingly less probable that a large enough quantity of
pyrolysis products would be produced as the concentration
of PCBs and chlorobenzenes is reduced. Indeed, the con-
taminated mineral oil fire in the University of Manitoba
resulted in no detectable concentration of PCDFs in the soot.
PCDDs, which are typically formed at an order of magnitude
less concentration than PCDFs, were not detected either.
In 1970 Vos et al.^12 showed a correlation between the
toxic effects of European PCBs and the concentration levels
of PCDFs. The major PCDF components contained in Yusho
oil were the highly toxic 2,3,7,8-TCDF and 2,3,4,7,8-penta-
CDF. The relative concentrations of the PCDF isomers pres-
ent in Yusho oil and in samples of used heat exchanger PCBs
(Kanechlor KC 400 and Mitsubishi-Monsanto T 1248) were
found to be strikingly similar (Kuratsune et al.^29 ). The overall
toxicity of the fluid may then be attributable to the presence of
small quantities of PCDF as degradation products of PCBs.
A 1981 EPRI study^105 of the equilibrium distribution of
PCBs between transformer solid materials and the liquid
dielectric showed that, for a typical 500 kVA transformer,
about 97.5% of PCBs are dissolved in the oil while only
2.5% are unevenly distributed among the paper, core steel
and Formvar wire. The majority of the PCB on the solids is
contained either adsorbed or absorbed in the paper.
An oil temperature fluctuation between 50°C and 110°C
causes a maximum shift of only 0.6% in the amount of dis-
solved PCBs when adsorption/desorption is considered
alone. The effect is much larger when multiple layers of
paper are involved because the migration of absorbed PCBs
through the capillaries of the impregnated paper causes a
gradual leaching which results in a final PCB concentration
in the bulk oil which usually exceeds the EPA definition of
a PCB fluid.
The study contains important implications regarding the
efficacy of retrofilling transformers using current technol-
ogy. The difficulties in obtaining a cost effective retrofill
solution would be avoided if technology were proven to cir-
cumvent the problem of PCB migration into the bulk oil.
Here, again, it is important to recognize clearly the differ-
ence between the retrofill of PCB contaminated transform-
ers and askarel transformers which have been impregnated
with PCBs.
In the case of a PCB contaminated mineral oil trans-
former it would be possible to reduce the PCB concentra-
tion of the working fluid by first draining the transformer
and refilling it with non-contaminated oil. Since the large
majority of PCBs are contained in solution, and are not
trapped by the porous insulation, it is possible to retrofill the
transformer to a required concentration level by repeated
washings, as necessary. The drained, PCB contaminated oil
can be disposed of by one of the several suitable methods
discussed in this section. On the other hand, even if all the
askarel fluid of a PCB transformer were to be washed outthere would, according to the EPRI report, still be about
2.5% of the original askarel contained in the core/coil assem-
bly. For a typical 235 gal. PCB transformer this amounts to
about 5.9 gal. of askarel impregnating the porous insulation.
A PCB concentration of 500 ppm in a 235 gal. transformer
amounts only to about one coffee mug in volume.
The diffusion of PCBs under the action of a strong con-
centration gradient from the interior of the paper towards the
very low concentration in the bulk retrofill fluid has been
discussed earlier. If the trapped PCBs are able to leach out
into the bulk retrofill fluid at a rate of as little as 1 ppm per
day, it would need less than 1.25 years before the fluid con-
centration had climbed above 500 ppm and the unit would
have to be regarded once again as a PCB transformer.
The molecular geometry of the PCB molecule makes
it susceptible to adsorption on carbon. Therefore, when an
askarel transformer is drained and filled with polymethylsi-
loxane (silicone) oil rather than an oil which contains mol-
ecules with the geometry or benzene rings, the only aromatic
structures present are those of the contaminant to be removed.
The capacity of the carbon for the adsorption of aromatic
species can be determined from Freundlich’s characteristic
isotherm. This can then be used to calculate the number of
filters of a given size required to reduce the residual askarel
concentration in the bulk oil to a defined lower concentra-
tion. The higher the initial concentration of PCBs remaining
after draining the transformer, the larger will be the number
of filters needed to adsorb the residual PCBs.
As more filters are used so the overall cost of the retrofill
approaches the cost of a new transformer. The rate of PCB
reduction is partly governed by the recontamination of the oil
and partly by the adsorption/desorption/leaching process itself.(2) THE DISPOSAL OF PCB CONTAMINATED
LIQUIDSDisposal options for PCB contaminated liquids are shown in
Figure 47. As stated earlier, PCB contaminated liquids are
understood to mean that the PCB is a minority constituent
of the fluid, usually at a concentration measured in parts per
million, but not necessarily less than the 500 ppm definition
of PCB fluid regulated by the EPA.
The most easily identified large volume of PCB
contaminated fluid is that used by the electrical industry
in its mineral oil transformers and the emphasis in this
paper is therefore placed upon the disposal of contaminated
dielectric oils. Contaminated fluids can be incinerated in the
same way as askarels but, in addition, the EPA has granted
approval for some companies to destroy their own low-
level oils in high efficiency boilers. The disadvantage of the
process is that it not only destroys the PCBs but also the
transformer oil.
Environmental Science and Engineering Inc. in
Gainesville Florida has conducted an analysis of the pyrol-
ysis products of dilute PCBs in a utility “High Efficiency
Boiler” and found that PCDFs and PCDDs were not detect-
able by the GC/MS method used. PCBs were destroyed withC016_003_r03.indd 934C016_003_r03.indd 934 11/18/2005 1:12:44 PM11/18/2005 1:12:44 PM