United States
      Environmental Protection
      Agency
                                                  Environmental Monitoring and Sup
                                                  Laboratory
                                                  Cincinnati OH 45268
      Research and Development
                                                  EPA-600/S4-82-023  June 1982
      Project Summary
\ ''
      Determination  of
      Pesticides  and  PCBs in
      Industrial and Municipal
      rWe*£tewaters
    r ^ v'
?j   \iilbhnD.MillarandRichardE.Thomas

'
       Steps in the procedure for the
      analysis of 25 chlorinated pesticides
      and polychlorinated biphenyls were
      studied. Two gas chromatographic
      columns and two detectors, electron
      capture and  Hall electrolytic con-
      ductivity, were evaluated. Extractions
      were performed with two solvents—
      dichloromethane and 15% dichloro-
      methane in hexane—at three pHs to
      determine extraction efficiencies. The
      effects of storage for seven days, in
      the presence  of residual chlorine, at
      two temperatures were determined.
      Florisil and alumina were compared as
      adsorbents for the clean up of extracts.
      Recoveries of the substances from
      clean  water and  wastewater were
      measured, and  assessments  of ac-
      curacy and precision were made.
       The method is satisfactory for the
      analysis of clean waters and waste-
      waters having a relatively low back-
      ground of interferences.  However, it
      does not work well against medium to
      high levels of background interfer-
      ences produced by substances that
      are electron  capture  sensitive, es-
      pecially halogenated ones. Use of the
      Hall detector  is  indicated when non-
      halogenated electron capture sensitive
      interferences are a problem, even
      though some loss in sensitivity will
                                                  occur. When  halogenated interfer-
                                                  ences are overwhelming, altered gas
                                                  chromatography conditions and
                                                  columns, such as temperature pro-
                                                  gramming and columns which produce
                                                  better resolution than the ones studied
                                                  in this work, will be required.
                                                    This  Project Summary was devel-
                                                  oped by EPA's Environmental Moni-
                                                  toring and Support Laboratory,
                                                  Cincinnati, OH. to announce key
                                                  findings of the research project that is
                                                  fully documented in a separate report
                                                  of the  same title (see Project Report
                                                  ordering information at back).


                                                  Introduction
                                                    Under provisions of the Clean Water
                                                  Act, the Environmental  Protection
                                                  Agency is required to promulgate
                                                  guidelines establishing test procedures
                                                  for the analysis of pollutants. The Clean
                                                  Water Act Amendments of  1977
                                                  emphasize the control of toxic pollutants
                                                  and declare the 65 "priority" pollutants
                                                  and classes of pollutants  to be toxic
                                                  under Section  307(a) of the Act. This
                                                  report is one of a series that investigates
                                                  the analytical behavior of selected
                                                  priority pollutants and suggests a
                                                  suitable test procedure for their mea-
                                                  surement.

-------
  The substances studied in this effort
were:
1.  a-BHC
2.  0-BHC
3.  <5-BHC
4.  Heptachlor epoxide
5.  DDE (p,p'(
6.  ODD (p,p')
7.  DDT (p.p')
8.  Endosulfan sulfate
9.  x-BHC
10. Heptachlor
11. Aldrin
12. Endosulfan I
13. Dieldrin
14. Endrin
15. Endosulfan II
16. Endrin  aldehyde
17. Chlordane
18. Toxaphene
19. PCB-1016
20. PCB-1221
21. PCB-1232
22. PCB-1242
23. PCB-1248
24. PCB-1254
25. PCB-1260
  The study was conducted  in two
phases. In Phase I, work was conducted
with clean  water and was intended to
provide information that would give
direction to Phase II work, conducted on
actual wastewaters, and  serve  as  a
comparison base for the information
developed.
  The ultimate objective was to develop
a method,  having a maximum  of ap-
plicability, that could be published in the
Federal Register and carried out in most
analytical laboratories.
   Other objectives included the devel-
opment of accuracy  and precision
information, procedure variations as
required by  a particular wastewater,
and the maximization of using common
sample treatment  steps for all the
categories  of priority pollutants.

Phase I

Gas Chromatography
  The  25  substances can  be gas
chromatographed very well. Sensitivity
quantities  producing a response 10X
noise level for  the  electron capture
detector range from 1 to 380 picograms.
Sensitivity quantities for the Hall
electrolytic conductivity detector are
roughly 10 to 100 times greater than
this. The 16 single compound pesticides
can be fully resolved if chromatographed
in two groups of eight each. The PCBs
are best chromatographed individually,
but several combinations can be ade-
quately chromatographed for some
analytical purposes.
  The primary gas chromatography
column used in this work, 1.5% SP2250
+ 1.95% SP 240} on 100/120 mesh
Supelcoport, is preferred on a general
use basis over 3% OV-1 on the same
solid  support.  In  limited situations,
however, it is  possible that either the
OV-1  column  or  columns with other
liquid phases, such as OV-17,  may
perform as well  or better than  the
primary column.
  A comparison of detectors showed
that  the Hall  electrolytic conductivity
detector is more discriminating than the
electron capture detector, and it should
be used as the primary detector, or used
as an  auxiliary detector when non-
halogenated electron capture sensitive
materials present interference problems.

Extraction Study
  The extraction study was initiated to
determine  the recoveries of the 25
substances of interest from clean water
at pH 2, 7 and 10 using  15% dichloro-
methane (DCM) in hexane and 100%
DCM as the extracting solvents.
  The water used in the extraction study
is a naturally buffered well water
obtained from the Southwest Research
Institute supply line  prior to chlorina-
tion.  Samples of this water were
collected and transported to the labora-
tory in empty solvent bottles which had
previously contained  Burdick and
Jackson solvents. The  pH of the  un-
treated water is close to 8. Adjustment
to the required pHs was accomplished
by adding strong acid or base.
  Extraction of  the pesticides from
clean water at parts per trillion  con-
centrations when the pH was 2,7, or 10,
was generally 80% or better using 1 5%
DCM  in hexane or pure DCM as the
extracting  solvent. When heptachlor
and aldrin were used, however, extrac-
tion percentages were generally 50 to
75%. Occasionally, one of the solvents
at a particular pH gave a more efficient
extraction than did the other solvent-pH
combinations,  although the improve-
ments  were  not great enough to
demand the particular combination be
adhered to exclusively. Under the same
extraction conditions, PCB-1242, PCB-
1248, PCB-1254, and PCB-1260 at low
parts per billion levels were extracted
with 90% or better recoveries. At similar
levels, PCB-1016, PCB-1221, and PCB-
1232 were extracted with adequate
efficiency, but  the recoveries recorded
were occasionally as  low as 60%.
  No clear advantages have been noted
for either solvent studied in this work. In
the interest  of commonality, either
solvent may be  chosen without  an
intolerable  loss of efficiency occurring
during sample extraction. The pH of the
water at the time of extraction is not of
substantial concern.  However,  if the
substance being sought is designated
beforehand, a slight advantage may be
available in some instances by consult-
ing the statistical analysis data for
combined pH-splvent effects on the
extraction.of that particular substance.

Preservation Study
  The preservation study was conducted
to determine the effects of a 90-day
storage period, at various conditions, on
the recovery of the 25 substances of
interest from dosed water samples. Each
sample consisted  of one quart of water
dosed with one of the eleven groups, as
in the extraction study. Two replicates
for each of twelve conditions of pH,
temperature and residual chlorine were
prepared for each group in the following
model (Table 1):
  The 2 ppm residual chlorine level was
obtained where required, by adding 160
mLof Mallinckrodt sodium hypochlorite
analytical reagent (5% minimum avail-
able Cl). Storage containers were one-
quart, flint glass, round, narrow-mouth
bottles closed with aluminum foil-lined
caps. A sample was prepared by filling
the bottle about two-thirds full with pH
adjusted water, adding 160 mL sodium
hypochlorite when required,  swirling
vigorously,  adding the 100  mL dosing
solution into the vortex, and combining
the remainder of the pH adjusted water
with the  solution  in the bottle while it
was still swirling. The bottle was capped
immediately and stored in a closed
cardboard box at 4°C or 24°C. Care was
taken not to slosh the bottle contents
onto the aluminum lining the cap after
closure.
  The  well  water was dosed  with a
quantity of substance inversely related
to the detector response for the sub-
stance. Thus, a-BHC was spiked at the
20 parts per trillion level and toxaphene
at the 4 ppb level. Other substances
were dosed at levels in between these
extremes.  Extraction was  performed
with  15%  DCM  in hexane and with
100% DCM at pHs of 2. 7 and 10. The
most serious problem encountered was
the  presence  of  interfering peaks in
chromatograms of blanks. The varying
magnitude of these peaks  made sub-
traction of a constant blank  usually  not

-------
possible.  Occasionally, work had to be
repeated.
  At the concentrations used, some
conditions of the preservation study
produced  pronounced effects  on  the
substances. Deleterious effects were
found to occur with about half of the 25
substances at some of the 12 combina-
tions of pH, temperature, and residual
chlorine levels studied However, 23 of
the substances were stored in the dark
for seven days at 4°C, under neutral or
acidic  pH,  and in the presence  and
absence  of two ppb  residual  chlorine
with the only losses suffered being of a
tolerable  magnitude. Aldrin sustained a
15 to 20% loss at acid or neutral pHs in
the presence of chlorine. Endosulfan
sulfate losses were 90% or more under
all conditions of storage at 230 parts per
trillion concentration in clean water. In
the storage and recovery tests involving
wastewater, however, endosulfan
sulfate was dosed at the  15  ppb
concentration and no large losses were
observed.
  Storage of samples before extraction
should be at a neutral to an acidic pH
and at 4°C. Storage periods of up to
seven  days' duration do not  produce
significant decreases in concentrations,
except for  endosulfan sulfate  in con-
centrations of parts per trillion.  For
determinations at these levels, samples
should be extracted immediately after
sampling. If aldrin is to be determined,
residual chlorine, if present, should be
eliminated before storage of samples.

Liquid-Solid Column
Chromatography
  Nearly 100%  recoveries  where
achieved during column chromatogra-
phy of substances  applied  to fully
activated Florisil  PR or alumina mixed
with  10%  by  weight water.   Elution
patterns  of the substances from  the
Florisil were more desirable than from
the alumina. In application studies with
wastewaters, the Florisil was  slightly
superior to the alumina in clean up of
extracts.

Florisil
  The  Florisil  clean  up column was
prepared  by gently packing Florisil PR,
which was taken directly from storage
at 130°Cfora minimum of 24 hours, toa
height of 10 cm (about 21 grams)  in a
400mm x  25mm (OD) glass  column
containing  a  course fritted disc  and
fitted with a Teflon stopcock. A layer of
sodium sulfate 1-2 cm in height was
added. The column was prewetted with
Table 1 .Experimental Model for Preservation Study

                          4°C
                         24° C
                0 ppm Cl
2 ppm Cl
0 ppm Cl
2 ppm Cl
pH2
pH7
pH 10
2
2
2
2
2
2
2
2
2
2
2
2
60 mL of hexane and drained to the top
of the sodium sulfate. Next,  a 5 mL
volume of hexane containing a dose of
one  of  the groups was put on the
column, which was again drained to the
top of the sodium sulfate. Elution was
then performed using 200 mL volumes
of 6, 15 and 50% ethyl ether in hexane,
with each 200  mL elution  fraction
received in a Kuderna-Damsh flask.
Each fraction was concentrated to a 10
mL final volume  and an appropriate
volume  was injected into the gas
chromatograph.

Alumina
  Add 10mLwaterto90gramsalumina
(Woelm N-Super I),  as purchased in a
jar. Cap the jar, shake until the mixture
is flowing smoothly, then let stand 24
hours before using. Slurry 22 grams of
the prepared alumina in enough hexane
so it can be  poured into a  400mm x
25mm OD glass column with a coarse
fritted disc and a Teflon stopcock. Settle
the alumina with gentle tapping on the
glass column and add a layer of sodium
sulfate  1-2  cm in  height.  Wash  the
column with 20 mL hexane and add the
dosed hexane which should  be about 5
mL in volume. Successively elute with
40 mL hexane, 110 mL hexane, 100 mL
50% ethyl ether containing 2% ethyl
alcohol  in  hexane,  and receive each
fraction  in a Kuderna-Danish flask.
Concentrate the fractions to final
volumes of lOmLandinjectappropnate
quantities into the gas chromatograph.

Evaluation of Results
  Recoveries  from both columns are
excellent. Elution times with the alumina
column are shorter than those with the
Florisil column. However, the single
compound pesticides,  chlordane, and
toxaphene, are more frequently found in
two fractions when using  a alumina
column  than when using  a  Florisil
column. Thus, processing and analyzing
additional fractions,  as would  be the
case with alumina, would likely be more
time consuming'than  would be  the
longer elution times with Florisil. Since
more components are recovered  near
the 100%  level  in  the first fraction
        (Florisil), this advantage appears to
        more  than  offset the less significant
        advantages of the alumina. Further-
        more, Florisil may be stored as received
        in a 13P°C oven for indefinite periods of
        time and still be usable. The alumina as
        received must be cautiously deactivated
        to Activity IV (10% water by weight) and
        then protected with  care until  used.
        How long Activity IV alumina may be
        held before use is not known. However,
        it appears  to be a  reasonably  long
        period, perhaps two weeks or longer.

        Phase II

        Wastewater Application
          With the  assistance and approval of
        the  project officer,  five wastewaters
        were  procured  and  analyzed.  The
        sources of these  wastewaters and the
        results of the analyses are given below.
        All samples were put into one-gallon
        glass  bottles which  had previously
        contained Burdick and Jackson "distilled
        in glass" solvents, suitable for pesticide
        residue analysis.  Sample sizes ranged
        from 20 to 24 gallons. All samples were
        shipped by air, unrefrigerated. Time in
        shipment was five days for Wastewater
        1  and about two days for the others.
        Upon  receipt  at  Southwest Research
        Institute, the samples were pooled in a
        25-gallon stainless steel vessel, where
        the pH was adjusted to 3-4, as required,
        with concentrated sulfuric  acid.'The
        samples  were then  returned to the
        bottles  in which they were shipped and
        stored at 4°C until used.

        Analysis of  Wastewaters
          In order to develop improvements
        upon the method and to provide base
        data for the  dosing and  recovery
        experiments, as well as the  accuracy
        and precision  evaluations to follow,
        each wastewater  was analyzed in
        triplicate for each substance of interest
        in this program. As an arbitrary working
        device  and  to allow for small  errors
        attendant to  injection, flow control,
        temperature control,  and so on, the
        coincidence of retention times was
        considered for identification purposes
        to be anything within 0.03 minutes.

-------
Dosing of Wastewaters and
Recovery Analyses
  In order to accumulate data  for
accuracy and precision evaluations,
each group was  put into six, one-liter
portions of each wastewater, one group
at a time. Three of these dosed samples
were immediately analyzed while three
were stored in the dark at 4°C for seven
days in sealed, glass containers. The
dosed  quantities were keyed to  the
specific response of the substance to
electron capture detection,  ranging
from 1 yug/L for the substances giving
the highest response, to 200 fjg/L for
toxaphene, except where interfering
background  required increase of  the
quantity. Except where noted to  the
contrary, extraction, clean up and GC
analysis of the extracts were as given in
earlier sections of this .report.

Wastewater 1
  This wastewater was  sampled by
personnel of the Surveillance and
Analysis  Division, USEPA, Athens,
Georgia and was identified there as Hl-
003 The pH as received was between
7.5 and 8. The wastewater was light
yellow to light orange in color, free from
solids and had a mild odor, reminiscent
of pine oil with a sweet overtone.  The
plant  producing  the sample manufac-
tures  toxaphene and  other organic
materials, and it was understood that
this effluent was"derived from all the
plant  operations,  and not from  the
toxaphene  unit.  This was due  to the
toxaphene plant's effluent treating  unit
being  "down" on the day of sampling,
and as a result no effluent was being
discharged from  it at that time.
  As a preliminary step to the analysis
of Wastewater 1, extracts were cleaned
up  with Florisil  and alumina. Of the
three fractions taken from each column,
alumina  produced the cleaner initial
two fractions but a much dirtier third
fraction. Since  six single compound
pesticides were determined in the third
fraction from the alumina, but only two
in the third fraction were determined
from the Florisil, Florisil was the better
overall choice and  was subsequently
used in all work with Wastewater 1. A
comparison with standards rules out all
of  the pollutants  of  interest in  this
program, except heptachlor and endo-
sulfan II and PCB-1242. The presence of
PCB-1242  is more questionable than
the other two substances, since only six
of the nine peaks being followed in the
analysis of this substance were present,
and the amount present, as indicated by
Table 2.    Ranges of Recoveries, Wastewater 1
                                      Ranges of Recoveries, %, avg.

Single compound pesticides
Chlordane (6 peaks}
Toxaphene (8 peaks)
PCB-1 254 (7 peaks)
PCB-1 01 6 (8 peaks)
PCB-1 260 (11 peaks)
PCB-1221 (7 peaks)
PCB-1 232(10 peaks)
PCB-1242 (9 peaks)
PCB-1 248 (7 peaks)
Zero
90 -
88 -
99 -
93 -
86 -
91 -
98 -
89 -
92 -
98 -
Time
110
106
105
98
97
102
117
134
101
101
7 Days' Storage
90 -
93 -
96 -
102 -
98 -
97 -
97 -
92 -
93 -
100 -
99
109
106
105
106
112
101
96
97
103
each peak, was not a constant value.
This is strong,  but not conclusive
evidence that PCB-1242 is not present.
However,  if the substance  present is
PCB-1242, the amount indicated is less
than two ppb, and the other two sub-
stances are at 20 and  40 parts per
trillion, respectively
  The dosing and recovery experiments
for Wastewater 1  went smoothly. It is
estimated  that the dosing  quantities
could have been reduced to 25% of the
quantities  used with  analytical  results
remaining  satisfactory in most instances.
The greatest difficulties occurred with
interferences for  one or  two   early
eluting  peaks  of PCB-1221  and  PCB-
1232.  If these peaks had been elimi-
nated  from the analyses, adequate
peaks would have remained for analyti-
cal purposes, and the ranges given for
these  substances would  have  been
narrowed. Recovery ranges were  as
follows in  Table 2.

Wastewater 2
  Wastewater 2 was sampled by per-
sonnel  of the Eastern  District Office,
USEPA, Westlake, Ohio  No information
was  supplied as to the  source  or
character  of this sample. The sample as
received was amber in color, free from
solids, had a urinelike odor, and had a
pH of 4.
  Despite  the absence of substances of
interest, the substances which  were
present provided many interferences in
the spiking  and recovery experiments.
Due to the interference  problems,
especially from the substances eluting
from the GC column during thefirstfour
minutes,  a considerable effort was
made  to  identify  the source of these
interferences, with the hope that such
identification would suggest approaches
for their  removal. The  presence of
elemental sulfur  was strongly sus-
pected. Treatment of first fractions of
extracts with elemental mercury began,
but the results were erratic. Careful
drying  of  the  extracts,  addition of a
mercuric salt, addition of acid-washed
copper,  and prolonging the agitation
period were without consistently bene-
ficial effects.
  During concentration of the extracts
in the  Kuderna-Danish apparatus, a
deposit of yellowish needle crystals and
a gummy residue were noted. Infrared
spectroscopy examination  suggested
that pentachlorophenol might have
been present.
  GC/MS analysis of the concentrated
extract indicated  the  presence of
naphthalene, methyl napthalene, two
dichlormated compounds,  2-methyl-
thiobenzothiazole, phthalates, dioctyl
adipate, anthracene, and two  hydro-
carbons. No pesticides were found.
  The difficulties with this wastewater
were so great that dosing and recovery
experiments were not  completed  for
several  of the pesticides and PCBs. An
erratic response to clean up, especially
with mercury,  produced inconsistent
background values in many instances
The greatest difficulties occurred within
the first four minutes after injection. It
was not  unusual  to experience no
reduction  in the  interferences  in un-
dosed  wastewater  replicates treated
with mercury, and yet, find  substantial
reductions in interferences in the dosed
replicates,  or vice versa An impasse
was reached and work with this waste-
water was set aside so attention could
be given to the  other wastewaters.

Wastewater  3
   Wastewater  3 was  procured by
Southwest Research Institute-Houston
personnel, and as received, had a mildly
bad odor, a cloudy appearance, and a pH
near  7. It was  the final  chlorinated
effluent from the Northside Treatment
plant  in  Houston,  which treats  both
residential sewage and  industrial plant
effluents
                                  4

-------
  The  analyses of three, one-liter
replicates suggests the presence of DDE
at 22 parts per trillion, dieldrin at 38
parts per trillion and chlordane at about
600 parts per trillion. Although only two
peaks  are shown  as matching with
chlordane, they are the major peaks for
the substance. The lesser peaks might
not be  detectable in this matrix at this
level. Five of the ten peaks matched the
standard for PCB-1232, but the amount
present, as indicated by each peak, was
quite variable. This is strong evidence
against its presence, and therefore,
PCB-1232 was not considered to be
present. The difficulties with this
wastewater included emulsion forma-
tion and erratic response  to clean up
with mercury. The  latter difficulty,
although not  as pronounced  as with
Wastewater 2, persisted throughout the
work with this sample. As with Waste-
water 2, the  undosed replicates were
more erratic as a rule than the dosed
ones. As a consequence, the back-
ground values were from replicates not
exposed to mercury, and the dosed
replicates all   received clean  up with
mercury. In  some  instances, this
resulted in recovery percentages which
were perhaps too  low  by a few per-
centage points Aldrin, DDE, DDT, and
toxaphene suffered the highest losses
during  the  seven-day storage period,
with  losses ranging from  12  to 17%.
Recovery ranges  were as follows  in
Table 3.

Wastewater 4
  Wastewater 4 was  procured by
Southwest Research Institute,  and was
from one of the influent streams in  a
highly  industrialized  area  of  the city,
flowing to  the Northside Treatment
Plant in Houston. As  received, this
sample  had a rank sewage odor, some
solid material, a cloudy appearance, and
a pH near 7
  Flonsil was used  to clean  up the
extracts, and was followed by agitation
of the  first fractions with mercury.
Emulsions formed as with Wastewater
3 and were treated as reported for that
wastewater.
  Treatment  with mercury produced
somewhat erratic reductions  of inter-
ferences Six  replicates were treated.
The  last two were  subjected  to an
increased amount of violent agitation
with  the vortex tube agitator  as com-
pared to the first four. This resulted in
more  uniform  and  lower analytical
values  in the  data derived from first
fractions. Comparison  of retention
 Table3.     Ranges of Recoveries, Wastewater 3
                                      Ranges of Recoveries, %, avg

Single compound pesticides
Chlordane (6 peaks)
Toxaphene (8 peaks)
PCB-1 254 17 peaks)
PCB-1 01 6 (8 peaks)
PCB-1 260 (11 peaks)
PCB-1 221 (7 peaks)
PCB-1 232 (10 peaks)
PCB-1 242 (9 peaks)
PCB-1 248 (7 peaks)
Zero Time
79-102
78-89
89-99
89-94
92-99
95-146
86-97
81-171
79-87
83-91
7 days' Storage
83-106
98-105
83-88
91-95
92-98
88-92
91-93
85-108
109-114
91-96
times  with standards indicates the
presence of DDE at 31 parts per trillion,
heptachlor at 119  parts per trillion,
dieldrin at 32  parts per trillion,  and
chlordane at 1.4 ppb.
  As mentioned earlier, Wastewater 4
was quite similar to Wastewater 3, so
for this and other reasons, only certain
substances were dosed and recovered.
This was performed at zero time, only.
There  were no  unexpected problems
met, and  mercury clean  up was  less
erratic than with Wastewater  3  and
with earlier undosed replicates of this
wastewater. Treatment  of  an extract
with BioBeads SX-2 was without bene-
ficial effects.

Wastewater 5
  This  sample was  taken by Southern
Research Institute personnel, through
arrangements  made by the EPA, from
the final effluent stream from a pesticide
manufacturing plant in Memphis,
Tennessee. As received,  this effluent
was clear, had a chemical odor, and a pH
around 3.
  Emulsification  was  not a problem.
Column clean  up was carried out with
Florisil after a  preliminary comparison
with alumina  showed no advantage
resulting from the use  of  alumina.
When this wastewater was analyzed, it
was found that Florisil  afforded  a
slightly better, although  inadequate,
clean up  than  did alumina. In another
experiment, the first fraction from
Florisil  of a dosed sample extract was
exchanged  into DCM and applied to a
22mm x 31cm  column of BioBeads SX-
2. During  elution of this  column with
DCM, several fractions were collected
and analyzed.  There was no improve-
ment in the separation of the interfering
materials from  the dosed pesticides.
  This  wastewater  had  a very high
background of  substances which were
electron capture  sensitive compared to
the other four wastewaters. However,
most of the substances producing peaks
are not substances of interest on this
program. Comparison of retention times
with standards indicates the presence
of heptachlor at  6.6 ppb. GC/MS
examination of the first  and second
fractions  of an extract indicated that
nearly all  of the substances are
chlorinated but only the  presence of
heptachlor was confirmed.
  This wastewater represented a seri-
ous challenge to this program. It was an
effluent from a pesticide manufacturing
plant, and it contained numerous non-
priority, chlorinated pollutants which
interfered with the analyses for many of
the priority  pollutants  which  are  the
subject of this  program.
  The ability of concentrated  sulfunc
acid and fuming sulfuric acid (7.5%) to
extract interfering  materials from  a
hexane solution of the firstfraction from
Flonsil was tested. Concentrated sulfuric
acid removed a small, but insignificant,
amount of these materials, and  the
fuming acid was no better.
  By operating the GC on a temperature
program, advantages were realized with
some  of the substances, for example,
single compound pesticides and  chlor-
dane  This technique  was applied to
analyze recovered  doses of  y-BHC,
heptachlor, aldrin, endosulfan I, dieldrin,
endrin, endosulfan II, endrin aldehyde,
and chlordane  from this wastewater at
zero time. The temperature was pro-
grammed from 160° to 200°C at 4° per
minute. Only dieldrin defied determina-
tion  completely.  Recoveries  for  the
other  components  ranged  from  62 to
107%. Endrin determination gave more
reasonable  values  under  isothermal
conditions that programmed conditions.
Peak  height measurements produced
the more reasonable recovery values.
Recoveries ranged from 57 to 109%. It is
noted that  the  57%  value  for the

-------
heptachlor component  of  chlordane
checks fairly well with the 62% recovery
obtained for heptachlor.
  Dosing and recovery were carried out
at zero time only.  PCB-1242 and PCB-
1254 were spiked into  the same
samples and  measured  using  the
procedure  in ASTM Method  D3304-77.
In this  method,  alkaline hydrolysis
followed by extraction with concen-
trated  sulfuric acid is  performed to
reduce  non-PCB  interferences. Some
interfering  peaks  remained  in  the
treated fraction, especially in the early
part of the chromatogram. Three such
replicates were dosed, extracted,
treated, and analyzed. The  data show
that peaks  1 and  5 in PCB-1242 were
obscured by interferences to the extent
they could not be read.  Recoveries of
other  components ranged from 22 to
85%. The  poor recoveries indicated by
the other  PCB-1242 peaks, except for
peak 6, are probably the result of poor
quantitation,  as a consequence of
interferences,  rather than poor extrac-
tion. This view is supported by the fact
that the indicated recoveries for PCB-
1254  are much higher and  uniform in
values.
  The other PCBs should give compar-
able results. Those which  elute early
(PCB-1016, PCB-1221, PCB-1232,
PCB-1248) will be analyzed with greater
difficulty and less precision and accu-
racy,  probably,  than the  other  late
eluter, PCB-1260, in Wastewater 5.

Discussion
  In Phase I,  endosulfan sulfate was
found to disappear almost entirely from
dosed clean water replicates stored for
seven days at  4° and 24°C.  In Phase II,
no similar inordinate disappearance of
endosulfan sulfate  occurred  in  the
replicates  of Wastewaters  1, 2 and 3
during storage at  4°C.
  Only a few of the 25 substances were
found in the five wastewaters analyzed.
Endrin, which was tentatively identified
in Wastewater 5, was the only one of
the substances present  above the 10
parts per billion level. Its concentration
was estimated at  32 ppb.
  Recovery of the  substances dosed
into the five wastewaters at the 1 to 20
ppb level was  satisfactory to very good
with respect to accuracy and precision
for Wastewaters  1, 3 and  4. Gas
chromatographic analysis to determine
the recoveries from the Wastewaters 2
and 5 could not be  performed in some
instances because of interferences that
were not sufficiently removed by clean-
up columns,  agitation  with  mercury,
alkaline hydrolysis and/or extraction
with sulfuric acid. It is very probable that
Wastewater 2 would have been analyzed
with greater success  with  the Hall
detector than with the electron capture
detector. Since Wastewater 5 was from
a pesticide plant, the limited analytical
success with this wastewater is indica-
tive of the limitations imposed on the
method by  high concentrations of
chlorinated  substances that are not
among the 25 substances of interest in
this program but that have extraction
and  chromatography characteristics
similar to  them. Therefore, neither the
Hall detector nor the electron capture
detector is completely  satisfactory m
this situation.  Better gas  chromatog-
raphy resolution or separation of such
interferences before gas chromatog-
raphy is required.
  John D. Millar and Richard E. Thomas are with Southwest Research Institute,
    San Antonio. TX 78229.
  James E. Longbottom is the EPA Project Officer (see below).
  The complete report, entitled "Determination of Pesticides  and PCBs in
    Industrial and Municipal Wastewaters," (Order No. PB 82-214 222; Cost:
    $9.00, subject to change) will be available only from:
          National Technical Information Service
          5285 Port Roy ill Road
          Springfield, VA. 22161
          Telephone:  703-487-4650
  The EPA Project Officer can be contacted at:
          Environmental Monitoring and Support Laboratory
          U.S. Environmental Protection Agency
          Cincinnati,  OH 45268
                                                                                         *USGPO: 1982 — 559-092/3424

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
    PS   0000339
    U  S  ENVIR  PROTECTION AGENCY
    REGION  5 LI8RAHY
    230  S DEARBORN STREET
    CHICAGO  IL  60604

-------