PCBs AND ASSOCIATED AROMATICS 923
method served to remove interferences from the PCBs con-
tained in a pentane fraction. A procedure was developed for
the PCB extraction of milk by Noren and Westöö (1968)
using ethyl ether and petroleum ether, centrifuging and
removing the separated phases.
The Yusho incident (1968) occurred shortly after Jensen’s
report on PCBs (1966). The incident prompted the US Food
and Drug Administration (FDA) to initiate a national survey
to determine the exact extent and levels to which PCBs
might have made their way into the food chain by indi-
rect use of PCB contaminated animal feed, industrial and
environmental sources, and the use of PCB-containing paper
food-packaging materials. The survey resulted in a ‘Notice of
Proposed Rule Making’ to limit the levels of PCBs in foods
in 1972.
The survey found that several accidents involving PCBs
had contaminated animal feed and subsequently the poultry
and eggs intended for human consumption. Other parts of
the survey indicated that the use of PCB-containing coat-
ings on the inner walls of grain silos had been responsible
for PCB residues in milk derived from dairy cows which fed
on the grain stored in such silos. The FDA concluded that
it would be in the best interest to limit the ways in which
PCBs might enter the food chain as well as limit the levels
of PCBs in food.
Since part of the problem with PCBs was the vulner-
ability of food and feed commodities to direct contamination
through accidental causes, the EPA issued rules governing
the continued deployment of PCBs in certain industrial
applications. These regulatory controls were made under
the Toxic Substances Control Act of 1976 and proposed the
discontinued use of PCBs in heat transfer systems in plants
manufacturing or processing food, drugs and cosmetics. An
interagency alert notice [1979] was then issued by the EPA
to urge voluntary compliance in removal of equipment con-
taining PCBs and replacement with non-PCB units to pre-
vent food contamination.
On February 24, 1969 the San Francisco Chronicle car-
ried a major feature on work reported by R. Risebrough
(1968) and that of S. Jensen (1966). Jensen had reported the
identification of PCBs while analyzing for DDE.
The early analytical method applied to the quantitation of
PCBs in oils uses a calibration technique developed by Webb
and McCall in 1973. The method is based upon the detec-
tor response comparison obtained for a commercial PCB
mixture with that of the sample and uses packed column gas
chromatography with, typically, an electron capture detector
(see ASTM method D-4095). Slight variations in the Aroclor
manufacturing process, or chemical alteration of the analyte
PCB, causes problems with the quantitation. Reasonable
results can be achieved, however, in the analysis of insulating
oils where the PCB have not been degraded. More sophis-
ticated analytical methods have been published by Albro
et al.^22 ; Erickson^23 ; Leo et al.^24 ; and Stalling et al.^25
The USEPA methods manual, SW-846, of test pro-
cedures is commonly used to quantitate PCBs in environ-
mental samples. The manual describes both the method of
sample cleanup as well as of analysis. The protocols aregiven in SW-846 as method 8080 using packed columns or
8081 with capillary columns. Each sample is spiked with
surrogate compounds and extracted using methods 3510 or
3520 for liquids or methods 3540, 3541 or 3550 for solids.
Cleanup methods use silica gel adsorption (method 3630) sul-
phuric acid/potassium permanganate destruction of interfer-
ing compounds, including pesticides (method 3665) and/or
sulfur cleanup by agitation with powdered copper, elemental
mercury or tetrabutyl ammonium sulfite (method 3660).
The electron capture detector used in the gas chromato-
graph is extremely sensitive towards compounds which con-
tain halogen atoms, such as the chlorine of PCBs. However,
the detector is not totally insensitive to other elements such
as oxygen and sulfur. Indeed, the sensitivity of the gas chro-
matographic quantitation is very dependent upon the level of
interferences present.
The sample chromatograms derived from the electron
capture detector in the gas chromatograph are interpreted
according to the pattern of peaks and their retention times
relative to known standard mixtures. In the case of PCB
contamination, the Aroclor products, such as Aroclor 1016,
1242, etc., each has a characteristic “fingerprint”.
Aroclor 1016 was derived from the vacuum distilla-
tion of Aroclor 1242 which removed the less volatile, more
highly chlorinated congeners of the mixture. The PCB con-
gener composition of Aroclors 1016 and 1242 are given in
Table 47.
Packed columns for GC analysis do not have enough
separating power for individual components of a PCB mix-
ture to yield anything but a series of envelopes. The pattern
formed by the envelopes tend to be characteristic of the PCB
mixture and therefore the pattern, taken as a whole, is usu-
ally recognizable as PCB and may be quantitated as such
by comparison with the envelopes produced by a standard
Aroclor mixture.
The pattern produced on packed columns by some sam-
ples shows that the electron capture detector was respond-
ing to components which, in all probability, contained a
highly electronegative i.e., electron attracting, element, such
as chlorine. The pattern of envelopes in the chromatogram
may not be distinguishable as any standard Aroclor prod-
uct. Since there is often no reason to suspect that any other
chlorinated compounds similar to PCBs may be present in
a sample the origin of the chlorine is usually an sometimes
erroneously attributed to PCB.
Attempts are sometimes made to remove interferences
from the oil to allow an unequivocal identification of PCB
by gas chromatography/mass spectroscopy (GC/MS). The
mass selective detector in place of the electron capture detec-
tor is typically used as an identification tool for individual
compounds and compound types. The response obtained in
the machine is compared with a computer stored library of
known compound responses. The closest matched response
at the gas chromatographic retention time of the species
is usually evidence enough to establish the identity of the
component of interest.
The GC/MS technique applied to complex matrices such
as oils tends to swamp the detector with naturally occurringC016_003_r03.indd 923C016_003_r03.indd 923 11/18/2005 1:12:42 PM11/18/2005 1:12:42 PM