Gel Permeation Chromatography
In The GC/MS Analysis Of Organics In Sludges
Robert H. Wise, Dolloff F. Bishop, Robert T. Williams,
and Barry M. Austern
Municipal Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
One method for measurement of priority organics in sludges
consists of sequential base/neutral and acid extractions with
methylene chloride using a homogenization-centrifugation tech-
nique; gel permeation chromatograph ;(GPC) for removal of high-
molecular-weight interferences from both extracts; and GC/MS iden-
tification and quantitation of the organics in the GPC fractions.
Removal of interferences from the sludge extracts by semi-automated
GPC produced an analyzable, but relatively contaminated, large-
molecule (phthalate) fraction and a "clean" small-molecule fraction
from base/neutral extracts and one "clean" phenol fraction from the
acid extracts. The GPC clean-up removed between 48-65 percent by
weight of the interferences from base/ neutral extracts of primary
sludges and about 35 percent from base/ neutral extracts of
activated sludges. GC/MS analyses confirmed low amounts of
interferences in the small molecule GPC fraction of the base/-
-------
neutral extract and in the phenol GPC fraction of the acid extract.
The study also revealed that interferences in the extracts did not
significantly alter the GPC elution position of representative
organics as compared to their GPC elution position in pure
methylene chloride.
Recovery studies on spiked sludge generally revealed satis-
factory recoveries for most of the 21 representative organics.
Representative pesticides, however, partitioned into both base/
neutral GPC fractions necessitating GC/MS analyses of both frac-
tions for maximum pesticide recovery. Representative phenols
exhibited erratic and sometimes low recoveries during the recovery
studies. Significant amounts of the weak acid phenols were
extracted inappropriately into the base/neutral extract. The
study suggests that the main problem in analysis of phenols is in
the extraction process.
-------
Gel Permeation Chromatography in the GC/MS
Analysis of Organics in Sludges
Robert H. Wise, Dolloff F. Bishop, Robert T. Williams,
and Barry M. Austern
U.S. Environmental Protection Agency, Municipal Environmental
Research Laboratory, Cincinnati, Ohio 45268
Introduction
The U.S. EPA in a 1976 Consent Decree (1) established, for
regulation, a list of 129 priority pollutants. The list includes
83 semivolatile (extractable) organics. Measurement of these
• ' • • • I i : • . . •
semivolatile organics usually employs GC/MS methods (2) using
liquid/liquid extraction of the organics from the environmental
sample before the GC/MS analysis. In complex environmental samples
the extractable background organics can interfere with conventional
extraction techniques and with the subsequent GC/MS analyses of the
organics of interest.
Municipal wastewater sludges which contain a very complex
background of extractable organics (3) are important environmental
samples. The high organic content of municipal sludges prevents
-------
efficient conventional extraction of organics. While continuous
liquid-liquid extraction, microextraction, and extractive steam
distillation techniques may have potential for extraction of
sludges, homogenization-centrifugation and modified Soxhlet tech-
niques have demonstrated efficient extraction capabilities (3). The
homogenization-centrifugation technique was adopted as the extrac-
tion step in the EPA's interim procedures (4) for the analysis of
sludges. The large amounts of organics extracted by the method
necessitate separation and clean-up of the extract.
As further background, the principal classes of organic inter-
ferences (5) extracted from municipal sludge samples are:
. Lipids
Fatty acids
i- I > • I :
. Saturated hydrocarbons
In the sludge samples the large amounts of extractable inter-
ferences overwhelm both the GC and the mass spectrometer. These
interferences must, therefore, be decreased in the extract before
injection into the GC/MS system in order to permit analysis. The
interference of total organics with the GC/MS analysis of an
extract can be reduced by removal of some of the interferences and
by separation of the extract into multiple fractions such that the
total amount of organics "in any fraction permits satisfactory
-------
GC/MS analysis.
Three conventional methods are available for this reduction and
separation:
. Acid/base separation
. Polarity separation (silica gel chromatography,
etc.)
• Molecular size separation (gel permeation
chromatography)
Acid/base separation is the fundamental approach behind the
Agency's basic GC/MS methodology for wastewaters (2). In this
basic methodology, base/neutral extraction - followed by acid
i i
extraction - divides the amount of extracted interferences between
acid and base extracts, separates the base/neutral organics from
the acids (phenols), and thus tends to equalize the interferences
in each extract. Polarity separation, with silica gel or florisil,
and a sequence of eluting solvents of increasing polarity, is
applied to organic extracts to separate the extract into multiple
fractions (such that the total organics in any fraction does not
prevent GC/MS analyses).
Molecular size separation removes the large lipids, large fatty
acids, and large hydrocarbons as a discard from the extract. These
-------
materials apparently thermally decompose in the GC system and
create very comolex GC chromatograms. Heavy loads of these
materials "also reduce the life of the GC columns and increase mass
spectrometer down-time. The molecular size separation also sepa-
rates extracts into multiple fractions.
The above techniques for separation of extractable organics in
sludge have been incorporated into an "Interim Method for the
Measurement of Organic Priority Pollutants in Sludges" (4) with two
alternative approaches for the separation of extractable organics.
One alternative consists of a base/neutral extraction followed by
an acid extraction, where both extractions use methylene chloride
as solvent and the homogenization-centrifugation technique. Gel
permeation chromatography (GPC) is applied for removal of high
• I . :
molecular, weight interferences in both base/neutral and acid
fractions, and finally GC/MS identification and quantitation of the
extractable organics in each GPC fraction is carried out. The other
alternative approach employs silica gel or florisil chromatography
of the base/neutral extract rather than gel permeation chroma-
tography. The silica gel or florisil chromatography separates the
extract into four fractions suitable' for GC/MS analyses. Both
alternatives analyze the-pesticide and PCB subclass within the
other base/neutrals. Thus each provides a consolidated analytical
method for the extractables.
-------
The objectives of this work were to evaluate the capability of
gel permeation to remove and separate the interfering materials, to
develop a semi-automated procedure of gel permeation chromatography
and to evaluate the effectiveness of the Interim Method with the gel
permeation separation in the analysis of priority organics in
municipal sludges.
Organic Interferences
To gain a perspective on the organic interferences in munici-
pal sludges, methylene chloride extracts were obtained by homoge-
nization-centrifugation extraction of municipal sludges from the
Cincinnati Metropolitan Sewer District's Mill Creek Sewage Treat-
ment Plant. The extracts were extracted as either base/neutrals
i i
(B/N) at pH 11 (followed by acids (A) at pH 2) or as acid/neutrals
(A/N) at pH 2 (followed as bases (B) at pH 11). Aliquots of the
extracts were air dried in a solvent hood at ambient temperatures
(22°C) for 24 hours to remove the methylene chloride and then
weighed to estimate the amount of semivolatile organics extracted
from the sludge by the homogenization-centrifugation technique.
The weighings (Table 1) revealed that more than 25% of the
organics based upon dry weight of solids were extracted from raw
primary sludge, about 18% from combined primary/secondary sludge
-------
and 10% or less from activated sludge. These weights of extracted
organics clearly indicate the large amounts of material in munici-
pal sludge extracts that act as interferences in the subsequent
analysis (GC/MS) of parts per billion levels of individual or-
ganics. The amounts extracted also clearly reveal that acid
neutral extraction followed by base extraction places most of the
mass of extracted organics in the acid/neutral extract whereas
performing the base/neutral extraction .first followed by acid
extraction more evenly distributes the extracted organics between
the base/neutral and acid extracts. In addition, the "clean" base
extract in the acid/neutral followed by base extraction sequence
should contain only two extractable priority organics. Thus the
sequence of base/neutral extraction followed by the acid extraction
results in significantly fewer background interferences in the
•' i i
extract containing most of the priority organics.
Our permeation chromatography for reduction and separation of
interferences uses methylene chloride as the eluting solvent. In
o
this work, 60A Styragel of 37/75-ji particle size was selected for
column packing to theoretically provide the best reduction of
interferences. The column and system (described later) was sized
to process extracts from 50 ml samples of primary sludges.
-------
The separation of the interferences by gel permeation chroma-
tography was initially evaluated using UV absorbance and refractive
index measurements as relative indicators of the amounts of organic
material in the eluent from the GPC column. The smoothed (ideal-
ized) curve of either refractive index or ultraviolet absorption as
a function of GPC eluent volume (Figure 1) revealed the typical
spreading of the organics in sludge extracts when chromatographed.
Using a trial-and-error approach with GC/MS analyses, the EPA's
Interim Method had recommended two fractions to aid in the
separation of much of the interferences found in the base/neutral
extracts.
Since our semi-automated gel permeation procedure was to be
I • H I . I .
used in a planned study of toxic substances which involved the
spiking of 21 selected organics (including six phthalates), two GPC
eluent fractions were established for the base/neutral extract to
provide a phthalate fraction FI and a fraction ^2 containing the
organic molecules smaller than the smallest phthalate, dimethyl
phthalate. The location of the fraction positions in the eluent
profile (Figure 1), established using di-n-octyl phthalate and
dimethyl phthalate standards, produced a highly contaminated FI
fraction and a "clean" F2 fraction from sludge extracts. The
organics larger than the largest phthalate, di-n-octyl phthalate
-------
were discarded. A fraction F3 was similarly established for the
acid (phenol) extract using a mixture of the eleven priority
phenols. The refractive index indicated that the F3 fraction was
relatively "clean."
Ambient-temperature drying and subsequent weighing of the F]_
and F£ GPC fractions (Table 1 and Figure 2) confirmed the presence
of significant amount of interferences in the phthalate (Fj)
fraction and the very low amounts of interferences in the clean F2
fraction. The relative amounts of organics in each fraction, as the
bar graph (Figure 2) reveals, support the use of the ultraviolet or
refractive index measurements as simple tools for evaluating the
GPC separation process.
Experiments were also performed in order to evaluate differing
column packing materials in the gel permeation chromatography.
This evaluation was done by passing 5-ml aliquots of combined
base/neutral-acid extracts of Cincinnati Mill Creek primary sludge
through two different sizes of Styragel. Appropriate f\ and F2
fractions, .based upon calibration with the pure phthalate standards
for each, size of Styragel, were air dried and then weighed to
determine the amounts of material in each fraction.
-------
0
The results (Table 2) confirmed that the 60A Styragel produced
fractions with t^e least amount of interferences. An interesting
sequential GPC treatment of the extract, in which the aliquot was
o o
separated first by 60A Styragel, followed by 200A Styragel,
revealed even further reductions of the interferences. This
sequential approach with the two Styragel columns in series
however, was not evaluated for its impact on recoveries of priority
organics. It has not been determined whether or not the observed
reduction in background organics is simply due to the doubling of
the column length.
Semi-Automated GPC Procedure
Based on the above, a semi-automated gel permeation procedure
i . t i ,i . , , .1
was developed for which fractions Fj and F2 (Figure 1) are collected
separately for each base/neutral extract, while a fraction F3 is
collected for the corresponding acid (phenol) extract. Under no
circumstance should absolute retention volumes shown in Figure 1 be
accepted purely on face value; instead, they should be regarded as
relative retention volumes. These warnings also apply to the
retention values (i.e., "cut points") quoted in the two procedures
given below since the correct absolute values will vary from column
to column.
-------
In our work with municipal sludge, cleanup of concentrated
methylene chloride extracts was accomplished by means of pumped gel
permeation chromatography (GPC) under semi-automated microproces-
sor control. For all of this work, the equipment consisted of the
following:
o
Column: Styragel, 20-mm ID x 122 cm long, 60 A pore
size, 37/75-ju particle size. These columns
were furnished by Waters Associates under
their catalog No. 40966.
Pump: Altex Scientific, Model No. 101A, semi-
preparative, solvent metering system. Pump
capacity = 28 ml/min.
Detector: Altex Scientific, Model No. 153, with 254-nm
UV source and 8-jjl semi-preparative flowcells
(2-mm pathlengths)
Microprocessor/controller: Altex Scientific, Model
No. 420, Microprocessor System Controller,
with extended memory.
Injector: Altex Scientific, catalog No. 201-56, sample
-------
injection valve, Tefzel, with 10-ml sample
loop.
Recorder: Linear Instruments, Model No. 385, 10-inch
recorder.
Effluent Switching Valve: Teflon slider valve, 3-way
with 0.060" ports.
Supplemental Pressure Gauge with connecting Tee: U.S.
Gauge, 0-200 psi, stainless steel. Installed
as a "downstream" monitoring device between
column and detector.
• f
Flowrate was typically 16.8 ml/min. of pure methylene chloride
(MCB's "OmniSolv"). Recorder chart speed was 0.50 cm/min.
The GPC system was calibrated under manual control with
standard solutions of select, individual, Consent Decree toxic
organics. In carrying out all exploratory calibrations, this study
relied heavily on relative retention data furnished by outside
contractors; this permitted significant savings of time and chemi-
cals.
-------
Two microprocessor programs were developed for the base/
neutral and acid extracts. Program number I was based on our
knowing that (a) the first class of toxics to elute includes the
phthalate esters and (b) the phenols should not be in this base/
neutral extract, but rather should be in the subsequent acid
extract. Program number I therefore is designed to yield a forecut
for discard, a "first" fraction (FI) containing all the phthalates
(plus other higher-molecular-weight neutral toxics), and a "second"
fraction (F2) containing the remaining (i.e., non-phthalate) neu-
tral and basic toxics. Program number II is based on a knowledge
that prior to elution of the phenols, everything can be discarded
since all the non-acidic toxics of interest should already have
shown up in the preceding B/N extract.
The overall GPC system was sized to handle the extract from a
50-ml sample of sludge. All fractions of interest from the above
programs were concentrated with Kuderna-Danish evaporators to 3 ml,
then subjected to GC/MS/DS analysis. The concentration to 3 ml
should be carefully determined for GC/MS quantitation but the
actual final volume may be varied to prevent precipitation of
organics in the final extract fraction.
The detailed GPC procedures used were as follows:
-------
Program No. I - With a methylene chloride flowrate of 16.80
ml/min. and a stabilized recorder baseline, the injector loop is
filled (by suction, not pressure) with a maximum of 7 ml of the
base/neutral extract; if at all possible, each extract should be
preconcentrated to a volume of 7 ml, or less so that the GPC'
clean-up can be accomplished without overloading the sample loop or
having to carry out repetitive sample injections.
The base/neutral extract is injected at what is arbitrarily
designated as time t0; for the following 9.0 min. (151 ml), the
column effluent is diverted to waste. At time tg.Qj collection of
fraction no. 1 (Fi) is started by means of the effluent switching
valve; this is continued until FI = 102.5 ml. At time 15.1 min.
(tl5.l)» the effluent switching valve is again actuated to begin
collecting fraction no. 2 (F2); this is continued until time t24.o
at which point F£ = 149.5 ml.
The effluent is again diverted to waste, and pumping is
continued until the recorder baseline has once more stabilized.
During this column-flushing period, the injector loop is also
flushed with clean CH2C12; then it can be loaded with the next
sample.
Program No. II - Flowrate of CH2C12 is again 16.80 ml/min. Also
-------
a maximum of 7 ml of (acid) extract is loaded into the injector loop
while the recorder baseline is stabilizing.
The acid extract is injected at time t0, and the column effluent
is diverted to waste for a total elapsed time of 12.5 min
(210 ml). At time ti2.5> collection of fraction no 3 (F3) is
started; this is continued until t24.o at which point FS = 193.2 ml
Once again, the column effluent is diverted to waste until the
recorder baseline has stabilized. At that point, the column is
ready to receive the next sample, or it may be shut down if desired.
At least as frequently as every other day, accuracy of column
flowrate (i.e., pumping accuracy) should be checked. In general,
significant errors in column flowrate will cause greater deviations
in separation reliability than any other single factor.
As an additional cross-check, column calibrations should also
be repeated at least once every two or three months. Our experience
has shown that calibration drift is not noticeable unless one of two
things occurs:
1) Column flowrate has changed significantly, or
-------
2) The GPC column is nearing "exhaustion" due to
irreversible column contamination.
GPC Separation of the Priority Organics
In order to evaluate the impact of the extracted interfering
organics. on the GPC separation of the priority organics independ-
ently of extraction variability, a primary sludge containing 6%
solids was extracted using the homogenization-centrifugation tech-
-vu^A^-'
nique, and aliquots of the base/neutral and acid extracts.then
"spiked" with a selected priority semivolatile organics as shown in
Table III and IV. Spiked and unspiked aliquots were separated by
the GPC procedure. The proper fractions were collected, concen-
trated and then analyzed using capillary GC/MS methods. Spiked GPC
solvent samples were also carried through the GPC and GC/MS
procedures.
The samples in this phase of the work and for all subsequent
work were analyzed by gas chromatography/mass spectrometry (GC/MS)
using a Finnigan 4000 GC/MS with an INCOS 2300 data system. The
chromatographic column-a J and W. SE-54, fused silica capillary
(30 m x 0.32 mm)-handled all the B/N and phenolic priority organics
satisfactorily. Earlier versions of the column were plagued by
progressive fragility and breakage problems. However, newer
columns are sturdy and deteriorate chromatographically before they
-------
become fragile. A typical column has lasted at least 4-6 months.
The gas chromatograph, after a grob-type splitless injection,
was programmed from 60-270°C at 5°/min. The mass spectrometer,
tuned to a satisfactory bis (pentafluorophenyl) phenylphospine
spectrum (6) was scanned from 40-450 amu, using 1-sec. scans.
Quantitation by the data system was performed using an internal
standard of perdeuteroanthracene (D^gA) added to each sample.
Response factors were determined by the injections of standard
mixtures of the compounds to be determined.
The GPC separation of the selected extractable organics, spiked
into the base neutral (F^ + F£) and acid (F3) extracts fractions,
are shown in Tables III and IV. The GC/MS chromatograms also
i . r i • I •'. ' '
confirmed that the F£ and F3 GPC fractions were "clean" with little
or modest amounts of background organics. The chromatograms
revealed that the GPC fraction Fj contained substantial amounts of
background organics. All six phthalates, isophorone, the two
dinitrotoluenes, and N-nitrosodipropylamine separated into GPC
fraction Fj (large molecules). All sixteen PAH's, the 4-chlor-
ophenyl phenyl and 4-bromophenyl phenyl ethers, all six chlorinated
aromatics, the three chlorinated aliphatics, nitrobenzene, and N-
nitrosodiphenylamine separated chiefly into GPC Fraction F2 (Table
III). As expected, all the phenols separated into the F3 fraction
(Table IV). The pesticides, PCB's, the two benzidines, and all
-------
other extractable priority organics not referenced above were not
tested. The above results and subsequent field tests revealed that
the background organics in the Cincinnati sludge extracts did not
significantly alter the elution position of the selected organics
in the GPC separation as compared to the elution position of these
same selected organics in methylene chloride alone.
Methods Recoveries
The overall utility of the analytical method with GPC clean-up
and separation, plus capillary GC/MS for the final analyses was
tested using a mixture of twenty-one representative priority
organics spiked into organic free water and also into primary and
activated sludges from our Cincinnati pilot plant.
• i . ' • .
These Cincinnati sludges were used because the GPC/GC/MS method
is employed for priority pollutant research samples obtained from
the EPA pilot plant in Cincinnati. The overall method (Figure 3) is
similar to the Agency Interim Method (4) but is adapted for semi-
o
automation using our available laboratory equipment and the 60A
Styragel resin (i.e. S-X3 Bio Beads are used in the Interim Method).
The sludges were extracted three times with three 80-ml
aliquots of methylene chloride (first at pH >_ 11 and then at pH <_ 2)
using homogenization-centrifugation; and were then subjected to the
semi- automated GPC separation. Each of the three GPC fractions
-------
, F2 and F3) and the discarded forecuts were all analyzed for the
twenty one spiked organics (Table V and VI) by the previously
described capillary GC/MS procedure. The GC/MS analyses were
applied to all fractions as well as discards to evaluate the
distribution of each organic during extraction and GPC separation.
Measurable amounts of the selected organics were not found in any
discard.
In Table V the normalized distribution of the selected organic
revealed that better than 90% of the measured (GC/MS) amounts of
phthalates, polynuclear aromatic hydrocarbons, and Arochlor 1254
are found in a single GPC fraction. The total overall recoveries
(Table VI) of the phthalates, representative polynuclear aromatic
hydrocarbons and the Arochlor 1254 were generally satisfactory and
r • .1 . i > I
comparable to those recently published (7) for the Interim Method
(4).
The smallest phthalates, dimethylphthalate and (to a lesser
degree) diethylphthalate exhibited low or very poor overall re-
coveries in the organic-free water and activated sludge samples.
The recoveries were satisfactory in the primary sludge samples.
Since spiking of phthalates directly into sludge extracts after
extraction did not reveal recovery problems (Table III) for the
small phthalates, the poor recoveries observed for the complete
-------
method in the organic free water and activated sludge samples are
probably associated with significant saponification losses of the
small phthalate esters during extraction at pH 11.
Similar losses have been observed (8) for extractions of
wastewaters at pH 11 by the standard EPA method (2). The presence
of substantial amounts of fatty acids in primary sludges is likely
to be responsible for the observed better recoveries of the small
phthalates from the primary sludge. Since repetitive monitoring of
sample pH during extraction was not performed, it is only postu-
lated, however, that the fatty acids in the primary sludge consume
enough base to reduce the saponification. Further work is needed to
clarify the observed results more fully.
['•'.'
The three representative phenols exhibited erratic, and some-
times low, overall recoveries (Table VI). They tend to distribute
into more than one GPC fraction (Table V). The very weak acid
phenols were found both in the acid F3 fraction and inappropriately
in the base/neutral F£ fraction. Pentachlorophenol, however, as a
relatively strong acid was usually found appropriately in the F3
(acid) fraction. Finally the recoveries of the weak acid phenols
spiked directly into the acid extracts (Table IV) were superior to
the same phenols spiked into the sludge before extraction. Thus the
extraction process appears to be the main problem for the weak acid
phenols.
-------
The selected pesticides while generally exhibiting satis-
factory overall recoveries (Table VI) would at times distribute
substantially into both the small molecule (F2) and the phthalate
(FI) fractions. Indeed, the multi component toxaphene distributed
into each GPC base/neutral fraction (Fj and F2) approximately
equally. Clearly maximum recoveries of the selected pesticides
required GC/MS analyses of both GPC base/neutral fractions.
Discussion
The EPA provides two alternatives in the "Interim Method for
the Measurement of Priority Organic Pollutants in Sludges." The
determination of whether the polarity (silica gel or flourisil)
separation of the GPC separation of the base/neutral extract is the
1 . i : ; • |
more appropriate for use has not been fully evaluated. Rigorous
evaluation of the two alternatives is best achieved by measuring,
for the same sludge samples, the weights of total organics in the
various GC/MS fractions produced by each alternative, while also
determining total recoveries (accuracy) and precision of the
measurements of priority organics by each alternative using GC/MS.
While such a definitive study has not been performed, the
current work does provide insight into the utility of the al-
ternatives of the Interim Method. Both alternatives have similar
-------
limitations in their common extraction approach and in their common
phenols analyses. The sequence of the base/neutral extraction
followed by the acid extraction, required to distribute the total
amounts of interferences nearly equally into the two extracts,
probably causes saponification losses of the small phthalate
esters. The difficulties in the extraction and recovery of
priority phenols also occurs as a common limitation in the method.
The GPC alternative for the separation and reduction of
interferences in the base/neutral extract provides only two base/-
neutral fractions for GC/MS analyses compared to the four fractions
obtained by polarity separation. The GPC approach obviously
requires fewer GC/MS injections if fractions are not combined
before GC/MS analysis. While definitive studies to determine the
actual reduction in organic interferences and the accuracy and
precision of the two alternatives on the same sludges have not been
performed, intuitive assessment indicates that the GPC separation
and clean-up should produce greater reduction in the background
organics. The polarity separation should only irreversibly remove
the most polar organics from each extract fractions, but such polar
organics tend to be preferentially excluded from the solvent during
the solvent extraction step.
The distribution and nature of the background organics and the
distribution of the priority pollutants into the various fractions,
-------
however, can impact recoveries of the individual priority organics.
The GPC separation of the base/neutral extract from Cincinnati
sludges produced a very clean F2 fraction with more than 90 percent
of the extracted background organics separated from the priority
organics in the $2 fraction. The ?2 fraction also contained the
major number of the tested neutral priority organics. In contrast,
the FI fraction, which includes the six phthalates and a few
additional priority organics, contained substantial amounts of
background organics but was still analyzable by capillary GC/MS
procedures. The GPC F]_ fraction, especially for primary sludges,
thus may contain more background organics than any separate
fraction from the polarity separation.
The collection of fractions in any separation and clean-up
' ' 1,1 ; , • j ,
procedure unfortunately may permit the distribution of organics
into more than one fraction. The results of GPC separation reveal
that most of the tested priority organics distributed desirably,
usually with 90% or better efficiency, into a single fraction. The
two single component pesticides, lindane and heptachor, distributed
with approximately an 80 percent efficiency .into the "clean" F£
fraction. The multi-component toxaphene, which appeared approxi-
mately equally in both fractions, obviously requires GC/MS analysis
of both base/neutral fractions for reasonable recoveries. The
observed distribution of phenols into the F2 and F3 fraction
-------
occurred because of extraction difficulties and is not related to
the GPC separation.
In contrast the four fractions from the polarity separation
alternative increase the chances for distribution of organics into
more than one fraction. Unpublished results (9) reveal that the
distribution of the specific organics into more than one fraction
is more of a problem in the polarity separation alternative.
Finally, automation of GPC separation with long term reuse of
the packed column is simple and effective. The single solvent,
methylene chloride, used in the GPC separation is the same as the
extraction solvent and has a low boiling point for efficient
Kuderna-Danish concentration. In contrast, the polarily separation
requires changing of solvents (4'solvent mixtures)'for the sep-
aration with one solvent, hexane, having a relatively high boiling
point. The polarity separation also requires fresh deactivated
silica gel or florisil not more than 5 days old for each separation.
Such a procedure is not easily automated.
Summary
The Agency's "Interim Method for Measurement of Organic Pri-
ority Pollutants in Sludges," includes two alternatives, gel
permeation chromatography and polarity (with either silica gel or
-------
fliorisil) chromatography for separation and clean-up of the
base/neutral extract before GC/MS analysis for organics. This study
evaluated the effectiveness of the GPC alternative for reducing the
background organics in the extract fractions before GC/MS analysis,
and also developed a semi-automated GPC procedure.for efficient
laboratory operation.
Removal of interferences from the sludge extracts by semi-
automated GPC produced an analyzable, but relatively contaminated,
large-molecule (phthalate) fraction and a "clean" small-molecule
fraction from base/neutral extracts, as well as one "clean" phenol
fraction from the acid extracts. The GPC clean-up permitted the
discard of 48-65 percent by weight of the interferences from
base/neutral extracts of primary sludges and about 35 percent from
base/neutral extracts of activated sludges. GC/MS analyses con-
firmed low amounts of interferences in the small-molecule GPC
fraction of the base/neutral extract and in the phenol GPC fraction
of the acid extract. The study also revealed that interferences in
the extracts did not significantly alter the GPC elution positions
of representative organics as compared to their GPC elution
positions for standard mixtures prepared in pure methylene chlo-
ride.
Recovery studies on spiked sludges generally revealed satis-
-------
factory recoveries for most of the 21 representative organics.
Representative pesticides, however, partitioned into both base/-
neutral GPC fractions, thus necessitating GC/MS analyses of both
fractions for maximum pesticide recovery. Recovery of represent-
ative phenols was erratic and sometimes low during the evaluation.
In addition, significant amounts of the weak acid phenols were
extracted undesirably into the base/neutral extract. The study
suggests that the main problem in analysis of phenols is in the
extraction process.
Acknowledgements
We thank Stephen Billets and James E. Longbottom for their
support and review of this work.
Literature Cited
(1) Natural Resources Defense Council (NRDC) et al. vs. Train, 8
ERC 2120 (DDC 1976).
(2) Federal Register, 44 (233) December 3, 1979, "Guidelines
Establishing Test Procedures for Analysis of Pollutants;
Proposed Regulations," pp 69526-69558.
-------
(3) Warner, J.S.; Jungclaus, G.A.; Engel, T.M.; Riggin, R.M.;
Chuang, C.C.; "Analytical Procedures for Determining Organic
Priority Pollutants in Municipal Sludge," EPA-600/2-80-030,
Municipal Environmental Research Laboratory, U.S. EPA; Cin-
cinnati, OH, March 1980.
(4) "Interim Methods for the Measurement of Organic Priority
Pollutants in Sludge," U.S. EPA, Environmental Monitoring and
Support Laboratory: Cincinnati, OH, August 1981.
(5) Bishop, D.F.; "GC/MS Methodology for Priority Organics in
Municipal Wastewater Treatment," EPA-600/2-80-196, Municipal
Environmental Research Laboratory, U.S. EPA: Cincinnati, OH,
November 1980.
(6) Budde, W.L.; Eichelberger, J.W.; "Manual for Organic Analysis
Using Gas Chromatography-Mass Spectrometry;" EPA-600/8-79-
006, Cincinnati, OH, March 1979.
(7) Lopez-Avila, V.; Haile, C.L.; Goodard, P.R.; Malone, L.S.;
Northcutt, R.V.; Rose, D.R.; Robson, R.L.; In "Advances in the
Identification and Analysis of Organic Pollutants in Water,"
Vol. II; Keith, L.H.; Ed. Ann Arbor Publishers: Ann Arbor, MI.
1981; pp 793-828.
-------
(8) Longbottom, J.E.; Environmental Monitoring and Support Lab-
oratory: U.S. EPA; Cincinnati, OH, personal communication,
1982.
(9) Haile, C.L.; and Lopez-Avila, V.; "Development of Analytical
Test Procedures for the Measurement of Organic Priority Pol-
lutants in Sludge;" Final Report in Preparation, EPA Contract
68-03-2695; Environmental monitoring and Support Laboratory:
Cincinnati, OH.
-------
TRIGLYCERIDES
FATTY ACIDS
ALIPHATIC HYDROCARBONS
PRIORITY PHENOLS
PRIORITY PHTHALATES
D
z
m
a
UJ
IT
10
uJ
U
O
z
o
50-ML SLUDGE SAMPLE
400-ML COLUMN
80 A STYRAGEL
403
GPC VOLUME. ML
Figure 1. GPC separation of organics.
-------
80
\
BASE/NEUTRAL EXTRACT
| | PRIMARY 8LUOOE
{::j:::) ACTIVATED SLUDGE
FORECUT CALCULATED BY DIFFERENCE
ATYPICAL IDEALIZED UV OR
REFRACTIVE INDEX PROFILE
181 264
QPC VOLUME. ML
Figure 2.' Amounts of extracted organics in the GPC base/neutral fractions.
-------
Sludge
(50 ml)
Adjust to pH > 11
with 6H NaOH
Extract 3 times with 80 ml
CHzd2 by homogenization/
centrifugation
Adjust to pH < 2
with 6N HC1
Extract 3 times with 80 ml
Ct^Cl? by homogenization-
centrifugation
Dry with
Kuderna-Oanish concentration
to 7 ml. Cleaning by semi-
automated GPC into fraction
Kuderna-Oanish concentration
to 3 ml. Analyze for phenols
by GC/MS with SE-54 WCOT column
Dry with
Kuderna-Oanish concentration
to 7 ml. Separate by semi-
automated GPC into fractions
FI and Fj
Concentrate f\ and Fo
r (-
to 3 ml. Capillary
column GC/MS analysis
with SE-54 WCOT column
Figure 3. Scheme for analysis of extractable organics in sludges.
-------
Table I. Extractable Organics Fran Municipal Sludges. .
Uelght of Extractable
Organics, g/1
Sludge, X of Extracted 8/N
Percent Extraction CH2C12 GPC Separation X Sol Ids Sol Ids Removed
Extracted by GPC
36
Sludge Type Solids
activated, 2.0
1/6/80
activated, 2.0
1/6/80
activated, 1.5
1/6/80
activated, 1.5
1/6/80
primary, 1.5
1/6/80
-
primary, 3.5
8/13/79
primary, 4.5
3/20/79
primary, 4.5
3/20/79
primary/
activated, 5.0
11/15/78
Sequence
B/N
A
Combined
A/N •
8
Combined
8/N
A
Combined
A/N
8
Combined-
8/N
A
Combined
A/N
B/N
A/N
A/N
8
Combined
Extract
0.62
0.53
1.15
1.02
0.08
1.10
0.78
0.73
1.51
1.22
.10
1.32
3.47
2:11
5.58
9.22
10.50
15.98.
8.59
0.23
8.82
Fj F2
0.34 .055
0.46 .055
1.79 .029
3.21 0.44
F]+F2 • 5.56
10
37
26"
34
48
65
35
18
a B/N is base/neutral extraction at pH 11; A 1s acid extraction at pH 2 after B/N
extraction. A/N is acid/neutral extraction at pH 2; 8 Is base extraction at pH 11 after A/N
extraction. & Since the base extraction after acid/neutral extraction removes very little
additional extracted material, the percent solids extracted can be estimated using only the
weight of the acid/neutral extract.
-------
Table II. Selection of GPC Packing3
Total Solute in
Phthalate Fraction b Small Molecule Fraction
Styragel Packing f\, mg F2, mg
60A 138 11.4
200A 197 12.3
60A + 200A C 67.8 5.8
a 5-ml aliquots of combined base/neutral - acid extract from a Cincinnati Millcreek Primary
Sludge were used in study, b Phthalate fraction is based upon elution volumes for largest
(di-n-octylphthalate) and smallest (dimethyphthalate) phthalates. Small-molecule fraction
contains all elutable molecules smaller than dimethylphthalate. c These special FI and F2
fractions generated by the two columns in series.
-------
Table V. Fraction of Organics in Each GPC Fraction*
Distilled Water Primary Sludge
Meanb Heanb
Fl
dimethyl phthalate
diethyl phthalate .948
di-n-butyl phthalate .946
butyl benzyl phthalate .968
di-n-octyl phthalate .975
bis(2-ethylhexyl)phthalate .968
naphthalene
acenaphthene
fluorene
phenanthrene
anthracene
fluoranthene • -,
chrysene
pyrene
phenol
2,4-dimethylphenol
pentachlorophenol
lindane .186
heptachlor .257
toxaphene .543
arochlor 1254
F*
-
.052
.052
.032
.025
.032
1.000
.986
.980
.980
.950
.970
-
.970
.667
1.000
-
.805
.731
.458
1.000
'3
-
-
.004
-
.001
.001
-
.015
.021
.020
.050
.031
-
.031
.334
-
1.000
.010
.013
-
.
"
.935
.905
.904
.919
.952
.946
.007
.005
' -
.004
.002
,-_
-
-
.001
-
-
.033
.188
.528
_
F*
.005
.028
.043
.040
.044
.042
.909
.906
.931
.916
.926
.907,
-
.907
.879
.869
.274
.956
.812
.315
1.000
F3
.059
.052
.052
.040
.004
.012
.083
.089
.069
.080
.073
.093
-
.093
.119
.131
.726
.011
-
.157
_
Activated Sludge
Meanb
"
.999
.915
.889
.962
.974
.903
-
-
.001
.003
.005
.002
-
.002
.005
-
-
.157
.229
.526
_
F*
.001
.033
.035
.022
.025
.028
.974
.947
.950
.928
.945
.936
-
.940
.467
-
-
.808
.753
.475
1.000
'3
-
.052
.076
.016
.004
.070
.026
.054
.049
.069
.050
.062
-
.059
.528
-
1.000
.036
.019
-
.
4 F] is the phthalate fraction of the base/neutral extract; Fj is the small molecule fraction
of the base neutral extract; F3 is the phenol fraction of the acid extract. & The mean is the
average weight fraction of the specific organic found in each GPC fraction by GC/MS analysis; the
mean is normalized and is usually based upon 4 determinations in distilled water, 6 determinations
in primary sludge, and 8 determinations in activated sludge.
-------
Table VI. Percent Recoveries
Priority Organic0
dimethyl phthalate
diethyl phthalate
di-n -butyl phthalate
butylbenzyl phthalate
df-n-octyl phthalate
bi s(2-ethylhexyl ) phthalate
naphthalene
acenaphthene
fluorene
phenanthrene
anthracene
Muoranthene
chrysene
pyrene •
phenol
2,4-dimethylphenol
pentachlorophenol
lindane
heptachlor
toxaphened
arochlord 1254
4 The mean is usually an
water, 6 determinations in the
Spiked
Distilled
HsO
mean3 . Sb
-
43 31
100 42
79 47
112 57
122 69
82 15
74 26
73 22
77 23
86 20
77 15
94 35
82 14
56
25 18 .
38
80 25
71 28
120 38
71 11
Primary
Sludge
mean4 Sb
64
60
96
74
72
45
47
73
76
76
78
105
111
108
56
25
152
62
35
130
95
average percent recovery of four
primary sludge and
11
10
52
12
35
29
26
19
19
17
18
29
64
t t
34
6
17
113
20
21
44
44
determinations
Activated
Sludge
Mean4 Sb
3 4
30 14
100 48
75 34
85 37
137 67
60 18
72 14
74 16
73 21
78 23
78 23
t 88 42
80 21
39 20
-
78 43
91 57
89 57
91 25
69 28
in distilled
8 determinations in the activated sludge.
0 S is an estimate of the standard deviation of the S where
tions. c Except for Toxaphene
with 1500 jig/1 of each organic
of Toxaphene and Arochlor 1254
and Arochlor 1254,
. d The sludges and
the sludges
n is the number
of determina-
and distilled water were spiked
distilled water were spiked
with 4500 jig/1
------- |