United States
Environmental Protection
Agency
Environmental Monitoring and
Support Laboratory
Cincinnati, OH 45268
Research and Development
EPA-600/S4-84-007 Mar. 1984
&EPA Project Summary
Screening Methods for PAH
Priority Pollutants in Wastewater
Ralph Riggin and Paul E. Strup
A screening method using a simplified
group test procedure for determining
polynuclear aromatic hydrocarbons
(PAH) priority pollutants in wastewaters
was developed. This screening proce-
dure, using total ultraviolet absorbance,
is designed to serve as an indicator of
pollutant levels and can be employed in
conjunction with the more specific
procedure. Method 610, which deter-
mines the levels of individual PAH
priority pollutants. Aqueous sources
used to validate this screening proce-
dure were distilled water, effluents
from a coking operation and a refinery,
and the secondary effluent from a
publicly owned treatment works (POTW).
Accuracy of the screening procedure
was estimated by comparing the results
obtained using this screening procedure
and Method 610- The screening proce-
dure exhibits a slight positive bias
(+30%) in the four effluents tested.
This Project Summary was developed
by EPA's Environmental Monitoring
and Support Laboratory, Cincinnati,
Ohio 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, EPA is required to promulgate
guidelines establishing test procedures
for the analysis of pollutants. The Clean
Water Act Amendments of 1977 empha-
size the control of toxic pollutants and
declare the 65 "priority" pollutants and
classes of pollutants to be toxic under
Section 307(a). This full report is one of a
series that investigates the analytical
behavior of selected priority pollutants and
suggests a suitable test procedure for
their measurements.
These 16 PAH's were: acenaphthene;
benzo(g,h,i)perylene; fluorene; phenan-
threne, dibenzo(a,h)anthracene; ideno(1,
2,3-c,d)pyrene; acenaphthalene, anthra-
cene, benzo(a)anthracene, benzo(a)pyrene;
benzo(b)fluoranthene; benzo(k)fluoran-
thene; chrysene; fluoranthene; naphtha-
lene; and pyrene.
The objective of this program was to
develop a simplified group test procedure
or "screening method" for determining
PAH priority pollutants in aqueous
samples. This method is designed to
serve as an indicator of pollutant level
which can be employed in conjunction
with the more specific Method 610,
previously developed by the EPA for
determining 16 individual PAH priority
pollutants in industrial wastewater.
The use of such methods will be
especially useful for screening samples
to determine whether the pollutant levels
are relatively high or low.
The requirements which such screening
methods are expected to meet are listed:
(1) Quantification shall be positively
correlated to that produced by the
refined Section 304(h) Clean Water
Act methods (as the sum of the
individual pollutants).
(2) The methods shall be able to stand
alone as reliable indicators of
pollutant level.
(3) The cost shall be nor more than
one-half of that using the refined
304(h) gas chromatographic (GC)
or combined gas chromatography/
mass spectrometry (GC/MS) meth-
ods.
(4) Response shall be class-specific,
producing a very low frequency of
false responses.
-------
(5) Responses need not be produced
for individual pollutants. Quantifi-
cation may be based on a unified
response of the category.
(6) Quantification shall produce a
numerical value (rather than a
range).
(7) Detection limit shall be at least 1
/"9/L.
(8) Recovery shall be 70% or greater
for the priority pollutants in the
category.
(9) The relative precision (as relative
standard deviation) of replicate
determinations shall be less than
50%.
(10) Ruggedness shall be maximized.
(11) The cost of the instrumentation re-
quired for the methods shall not
exceed $10,000.
The major objective of this program
was to select, evaluate , and validate the
method which best meets the require-
ments listed above for the priority
pollutant PAH's. To accomplish this
objective, a literature survey was conducted
to formulate the most logical approach to
meet these goals. Three candidate
approaches (thin-layer chromatography,
(TLC), fluorescence and Ultra-Violet (UV)
absorbance) were selected for laboratory
evaluation.
Screening Methods Evaluation
The simplest method and one ultimately
selected as the PAH screening method is
the measurement of total UV absorbance.
This method avoids the problem of
different responses for individual PAH's
at a specific wavelength measurement.
For this method, a very inexpensive
instrument is required, consisting of a
deuterium lamp as a UV light source
(—200-350 nm), a bandpass filter which
blocks wavelengths below —240 nm to
remove contributions from the solvent, a
cuvette to contain the isolated PAH
extract, and a photodetector.
The approaches used for fluorescence
measurements were similar to those
used for absorption measurements (i.e.,
total fluorescence using broad bandpass
filters). Advantages of fluorometry over
UV spectrometry and higher selectivity
(nonfluorescent potential interfering
compounds are not detected). Adisadvan-
tage is that the sensitivity for naphthalene
will be lower than for the PAH's because
it absorbs and fluoresces at lower
wavelengths. In addition to total fluores-
cence measurement, a spectrofluorome-
ter can be used to obtain measurements at
several specific wavelengths, to partially
overcome the variability in response
among the various PAH's.
2
TLC methods using very narrow chan-
nels on a silica gel TLC plate have been
proposed. PAH's are eluted as a single
band and the length of the band is related
to the total PAH concentration. This
procedure is actually the TLC equivalent
of the total UV measurement method
described. The method employs a single
wavelength UV source as the visualization
technique, and incorporates its own
cleanup step, thus requiring only
sample extraction and concentration. The
variability in response between various
PAH's is expected to be quite large, thus
making quantification more difficult.
All three approaches were evaluated.
The UV absorbance method was selected
as the best of the three and the results
obtained are described below.
Evaluation of UV Detection
The determination of total PAH concen-
tration by monitoring UV absorbance was
accomplished using two approaches. The
first approach, the multiple wavelength
method, involves monitoring the absorb-
ance at two or more specified wavelengths
and performing a weighted average
calculation to determine total PAH
concentration. This approach is useful
only if two wavelengths are found such
that each of the components of interest
absorb strongly at one of the wavelengths
but not at both wavelengths. This
approach becomes difficult as the number
of components of interest increases.
The second approach, called the total
UV measurement approach, involves
making a single UVabsorbance measure-
ment over a wide wavelength region, over
which the average molar absorptivities
of the components of interest are
approximately equal.
In order to evaluate which of these UV
methods would be most appropriate for
the PAH's, the molar absorptivity of the
PAH's over specified wavelength regions
were calculated from literature data. The
average molar absorptivities of the PAH's
over narrow (10 nm) wavelength regions
vary over two orders of magnitude. The
molar absorptivities over certain wave-
length regions (e.g., 240-320) vary by a
factor of 15. Since the most responsive
components are the high molecular
weight components (e.g., dibenzo(a,h)-
anthracene), the variation in response on
a weight basis is substantially less.
In view of the high variation in
response for PAH's over narrow wavelength
ranges, and the greater complexity of the
multiple wavelength approach, the total
UV approach was selected for laboratory
evaluation.
Two optical filters were evaluated for
use in the total UV measurement. The
first filter, Corning 7-54, is a bandpass
filter transmitting UV radiation in the
240-380 nm region, whereas the second
filter. Corning 9-54 is a cutoff filter
transmitting UV radiation above 230 nm.
The apparent molar absorptivities and
relative response (on a weight basis) for
various PAH's using these two filters was
obtained. Examination of the data reveals
that a smaller compound-to-compound
variation in response is obtained using
the bandpass filter than using the cutoff
filter. Sensitivity for the various compounds
was somewhat better using the bandpass
filter (except in the case of naphthalene
which gave a slightly higher response
using the cutoff filter). Consequently, the
bandpass filter was chosen for use in all
later experiments.
The linearity in response using the total
UV approach was evaluated using a
solution containing equimolar concentra-
tions of all 16 priority pollutant PAH's in
cyclohexane. The plot of total PAH
concentration versus absorbance indicates
that the response is nearly linear over the
4-1 20 /jM range or approximately 0.2-24
/ug/mL (using an average molecular
weight of 200 daltons).
The response data for individual PAH's
as a function of concentration using the
total UV approach was obtained. The re-
sponse for many of the compounds is not
linear over a wide range of concentra-
tions. It is important to recognize that the
concentrations of the individual compo-
nents in this experiment are approximate-
ly 20 times greater than in the previous
experiment where good linearity was ob-
tained. This result implies that the total
UV determination should be conducted
within a rather narrow absorbance region
(—0.05-0.35 absorbance units) so that
nonlinear response does not bias the
data. Dilution of sample extracts will be
required to achieve this criterion in some
cases.
The compound used for instrument
calibration in the total UV method should
have a molar absorptivity close to the
average molar absorptivity of the 16
priority pollutant PAH's and should also
exhibit good linearity so that a degree of
bias introduced into the method can be
minimized. The reference compound
should also be readily available in pure
form, inexpensive, and noncarcinogenic.
Initially, anthracene was chosen as the
reference compound for calibration
purposes, since it appeared to best meet
these requirements. However, examina-
tion of the response data led to the
conclusion that fluoranthene would be a
-------
better calibration standard, since it gave
more nearly linear response over the
desired absorbance range. Fluoranthene
has the advantages that: 1) it is not a
carcinogen; 2) it is readily available in
pure form at low cost; and 3) its total UV
response is near the median for the
priority pollutant PAH's. Consequently,
all further measurements using the total
UV approach were made using fluoran-
thene as the calibration standard. The
calibration curve was generally constructed
over fluoranthene concentrations of 5 x
10~6to 10~4M°r 1 to 20//g/mL Results
were calculated as fluoranthene equiva-
lents by comparing the UV absorbance for
the sample extract to the fluoranthene
calibration curve.
Evaluation of Column Cleanup
Approaches
Silica gel and alumina adsorbents were
evaluated in terms of ability to selectively
remove interfering compounds. The
results obtained using these two adsorb-
ents are described below.
Silica Gel Column Cleanup
Silica gel columns were prepared using
the procedure described in Method 610.
Initial experiments involved eluting the
ST INTF 1 and ST INTF 2 interference
solutions, and collecting and analyzing
the "PAH fraction." Concentrations of ST
INTF 1 and ST INTF 2 were 100/ug/mL of
each compound. ST INTF 1 contains
nonpolar compounds (e.g. benzene PCB's
and ST INTF 2 contains polar components
(e.g., nitrogen heterocyclics).
Because dioctyl phthalate (OOP) is a
widespread contaminant in wastewater,
this compound was examined separately
at a concentration level of 1CT3 M. PAH's
were analyzed at a concentration level of
1CT5 M per compound. All samples were
applied to the silica gel column at a
volume of 1.0 mL.
Three chromatographic fractions were
obtained for the samples listed above
(applied separately) as follows:
• Fraction 1 - 25 mL of pentane
• Fraction 2 - 25 ml of 40% CH2CI2 in
pentane
• Fraction 3 - 25 mL of CH2CI2.
The UV absorbance of each of these
fractions at 230 nm was determined.
The UV absorbance data indicated that
silica gel readily separates the polar
compounds (ST INTF 2 and OOP) from the
PAH's since UV absorbance was noted
only in the methylene chloride fraction
(Fraction 3) for these compounds. However,
some of the nonpolar compounds were
found to elute in the PAH fraction
(Fraction 2). All of the PAH's were found to
elute in Fraction 2, as expected.
Additional experiments were conducted
to determine which specific nonpolar
compounds were eluting in the PAH frac-
tion and to determine the elution volumes.
If these compounds were eluting near the
initial elution volume where PAH fraction
collection began, then by addition of a
larger amount of pentane these com-
pounds could be separated from the PAH
fraction.
Elution of the ST INTF 1 solution was
performed with fraction collection at 10 mL,
10 mL, and 5 mL of pentane followed by 5
mL, 10 mL, and 10 mL, or 40% CH2CI2 in
pentane. These fractions were concen-
trated by Kuderna-Danish evaporation on a
water bath at 50°C and analyzed by GC.
Compound identification was based
on retention time and recoveries were
determined by comparing the peak areas
obtained for the components in the
fractions to the peak areas obtained for the
ST INTF 1 stock solution.
Results of these analyses, indicate
that trimethylbenzene and 2,4,2',4'-
tetrachlorobiphenyl elute at the very end
of the PAH fraction. Therefore, it is not
possible to separate these compounds
from PAH's using silica gel. Since PCB's
absorb UV radiation and are frequently
present in environmental samples, this
finding represents a serious limitation of
the silica gel cleanup procedure when UV
absorbance is used as the determinative
procedure for PAH's.
Alumina Column Cleanup
The selectivity of alumina adsorption
chromatography columns for PAH's was
evaluated. Elution volumes for ST INTF 1
compounds were determined using 10%
CH2CI2 in pentane as the eluent. Elution
volumes for PAH's ST INTF 2, and OOP
were determined used 100% CH2Cl2 as
the eluent. The resulting fractions were
analyzed by UV spectrometry as described
earlier.
Results of these analyses indicated
that the elution volume for ST INTF 1
compounds was from 5 to 15 mL using
the 10% CH2CI2 in pentane eluent while
PAH's were completely retained on the
column for the total elution volume of 50
mL When using 100% CH2CI2 as the
eluent, PAH's showed two maxima. One
maximum occurred at an elution volume
betweeen 15 and 25 mL and the second
maximum between 35 and 45 mL.
However, there was still some absorbance
indicated for the tenth 5-mL fraction
PAH's. OOP and ST INTF 2 compounds
appeared to be totally retained on the
alumina column with an elution of 50 mL
of CH2CI2.
Because PAH's were partially retained
after eluting with 50 ml of CH2CI2, the
above experiments were repeated using
an elution volume of 60 mL of CH2CI2.
Results of these experiments indicated
that PAH's were eluted at a total elution
volume to 55 mL and OOP and ST INTF 2
compounds were completely retained.
Elution with 30 mL of diethyl ether was
necessary to elute the OOP and ST INTF 2
compounds from alumina.
GC analysis was conducted on the later
eluting PAH fractions to determine which
compounds were being retained. Results
fo these analyses indicated the PAH's in
the second maximum were primarily
benzo(b)fluoranthene, indeno(1,2,3-cd)
pyrene and dibenzo(a,h)anthracene with
a small amount of benzo(a)pyrene.
On the basis of these results, alumina
was selected as the absorbent of choice
since both polar and nonpolar (e.g., PCB)
components are separated from the
PAH's using this approach.
Validation of Screening
Method for PAH's
Based on the results presented above, a
screening method for determining PAH's
in wastewater was selected. This method,
consists of the following steps:
(1) Extraction of the sample (1 literjwith
a single aliquot (150 mL) of CH2CI2.
Only 100 mL of the CH2CI2 is
collected so as to reduce susceptibil-
ity to emulsion problems.
(2) Drying (sodium sulfate) and Kuderna-
Danish concentration of the extract.
(3) Alumina cleanup.
(4) Measurement of total UV absorbance
relative to a fluoranthene calibration
standard, using a filter photometer.
The total UV determinative procedure
was judged to be superior to either the
fluorescence or "channel" TLC methods
because of great variability in response
betweeen the PAH compounds of interest
for the latter methods. Bias (inaccuracy)
of ± 1000% of greater may be observed
for the latter methods, depending on the
specific PAH compounds present. By
comparison, the total UV method bias
should not exceed ± 100%, based on the
response ratios determined for the
priority pollutant PAH's.
The proposed method was initially
applied to reagent water and two waste-
waters, spiked at various levels with
equimolar concentrations of all 16 PAH's.
In all cases, reproducibility and recovery
were excellent, and therefore, the
proposed method was considered suitable
for detailed validation.
-------
Method validation consisted of: 1)
determination of the method detection
limit (MDL) for the screening method
using the protocol; and 2) comparison of
analytical results obtained for relevant
water samples, unspiked at various
levels, using the PAH screening method
and the Method 610.
Method Detection Limit
(MDL) Determination
The MDL for the PAH screening
method was determined according to the
EPA protocol. The spiked aliquots contained
equimolar concentrations of all 16
priority pollutant PAH's. The process
blank value (reagent water process blank)
was subtracted from all the values prior to
calculating the MDL, since this calculation
procedure is recommended in the analy-
tical method.
The MDL data for reagent water and
three relevant water samples are presented
in Table 1. The MDL values for the various
water types are essentially the same
(approximately 4 /ug/L). These results
seem reasonable, based on a reproducible
process blank level of 2.5 /ug/L and the
MDL of 4 /ug/L, which is considered to be
satisfactory compared to alternative
methods, such as Method 610. Although
Method 610 detects less than 4 /ug/L of
most individual PAH's the detection limit
in terms of total PAH will be higher
(poorer) when a variety of PAH's are
present in a given sample.
Comparison of Method 610
and PAH Screening Method
Aliquots of distilled water, POTW
secondary effluent, a final effluent from a
coking operation, and a refinery effluent
were extracted and cleaned up according
to the procedure outlined above. Three
aliquots of each sample unspiked and
spiked at each of two concentration levels
were extracted. The extracts were
analyzed by the PAH screening method
and then exchanged into acetonitrile and
analyzed by the high performance liquid
chromatography approach described in
Method 610. One extract of each sample
(unspiked) was analyzed by GC/MS to
confirm component identities.
The results of the Method 610 analyses
for the three samples indicate that in
most cases, with the exception of benzo(k)
fluoranthene, the recovery was greater
than 75% at both the high and low spike
levels. The process blanks contained no
detectable level of any of the 16 priority
pollutant PAH compounds.
Table 1. Method Detection Limits for PAH Screening Method in Terms of Fluoranthene
Equivalencies
Matrix
Reagent Water
Coking Effluent
Refinery Effluent
Spike
Level
5.5
5.5
5.5
Average
Value
(Uff/Ll
5.8»
4.5"
7.4b
Standard
Deviation
fU.Q/U
1.1
1.2
1.6
MDL
(ua/LI
3.5
3.1
4.9
Percent
Recovery
105
82
135
a Spiked with an equimolar mixture of all 16 priority pollutant PAH's.
b Blank corrected values.
Table 2. Comparison of Screening Method to Method 610 for Various Water Samples
Sample
PAH Spike Level
Average Percent Recovery*
Screening Method
Method11 610
Reagent Water
Reagent Water
POTW
POTW
Coking Effluent
Coking Effluent
Refinery Effluent
Refinery Effluent
18.1
3.6
36.2
3.6
36,2
3.6
36.2
3.6
96+ 19
99 + 8
76 + 3
92 + 12
78 + 4
97+ 15
84 ±8
135 ± 32
77 ±8
76 ±6
90 + 9
75 ±6
87 ±3
61+6
85 ±3
43 + 12
a Three replicates at each spike level.
b In terms of fluoranthene equivalents.
The values obtained using the screen-
ing method are compared to the Method
610 results in Table 2 for all four water
types. The values for the screening
method are consistently higher than the
Method 610 results. This positive bias
results from the use of fluoranthene as
the calibration standard. Since the molar
absorptivity of fluoranthene is slightly
less than the average molar absorptivity
of the 16 PAH compounds, the absorb-
ance of the PAH mixed standard used to
spike the samples is greater than the
fluoranthene standard. This factor is
demonstrated by comparing the actual
spike level (first column of Table 2) to the
spike level in terms of "fluoranthene
equivalents."
While this bias can be significantly
reduced by selecting a different PAH
spiking standard, this procedure would
not be effective in a real-world situation
(i.e., when the levels of individual PAH
compounds are not known). In practice,
the amount of bias will not be known and
will vary from sample to sample, within
the relative sensitivity factors (approxi-
mately ± 100% relative to fluoranthene)
of the various PAH compounds (last
column of Table 10 in the full report). In
certain cases where the major PAH
species is known, the amount of bias can
be minimized by calibrating with that
particular PAH species.
The background level of the distilled
water sample was approximately 2.5
ug/L while the POTW, coking effluent,
and refinery effluent samples had slightly
higher backgrounds due to the presence
of PAH. The refinery effluent gave a back-
ground level of 11.8/ug/L. In practice, the
screening method blank consistently
gives a background level of approximately
2.0-2.5 /ug/L. The method description
suggests that results be corrected for
method blank in order to reduce the de-
gree of positive bias for the method. This
a mount of background may be due to con-
tamination of the sample with trace
amounts of UV absorbing material, such
as phthalates, following the cleanup step.
GC/MS analysis of the wastewater
extracts confirms that traces of phthalates
(representing less than 5 /ug/L in the
wastewater sample) are present in the ex-
tracts. Trace quantities of phenanthrene
and naphthalene were identified in the
refinery effluent and the POTW effluent.
Methylthiobenzene was also identified in
the refinery effluent. Napthalene, methyl-
naphthalenes, dimethyl naphthalenes,
phenanthrene, and pyrene were detected
in the POTW secondary effluent.
The PAH screening method results and
Method 610 results were subjected to lin-
ear regression analysis to determine the
degree of correlation between the two
methods. Screening method data were
entered as the "Y" values and Method
610 were entered as the "X" values. The
slope, intercept, correlation coefficient,
and R-squared (residual) values were cal-
culated for each water type as well as for
the data set as a whole.
-------
The slope values were approximately
1.3, with the exception of reagent water,
which is indicative of a positive bias
(—30%) for the screening method. The
reasons for this bias are discussed above.
The intercept values were approximately
3 jug/L, which is approximately equal to
the MDL. This result indicates that the re-
lative bias in the method is constant over
the concentration range tested (i.e., there
is not a large fixed bias inherent in the
method). The correlation coefficients
were somewhat better in wastewater
than in distilled water, and are con-
sidered to be satisfactory in all cases. The
R-squared values were always > 95%, in-
dicating that the data points fall relatively
close to the regression line.
It is important to recognize that these
correlation data are rigorously valid only
for the particular spiking scheme shown.
If other combinations of PAH's are pres-
ent, a different slope will be obtained for
the regression line, although the other
parameters (e.g., correlation coefficient,
intercept) will not be greatly affected.
These results indicate that the PAH
screening method is precise and positive-
ly correlated with Method 610, which is
currently in use for the determination of
priority pollutant PAH's in wastewater.
An important consideration in the evalua-
tion of these two methods is the ana-
lytical cost, since one of the primary ob-
jectives in the development of screening
methods is to reduce the analytical costs
associated with monitoring pollutants.
Summary
The PAH screening method requires
considerably less time to conduct than
does Method 610. The sample extraction
step involves a single extraction step with
a fixed proportion (2/3) of the extract be-
ing recovered. This process reduces the
problems with emulsions substantially
since the emulsive interface does not
need to be recovered. In addition, the in-
strument calibration, sample analysis,
and data reduction steps are accom-
plished much more rapidly for the screen-
ing method, since onlyasinglenumerical
value (total PAH concentration) rather
than 16 individual concentration values
is obtained, and the lengthy chromato-
graphic analysis is eliminated. In addi-
tion, the time required to prepare refer-
ence standard solutions is greatly re-
duced. Only a single standard (fluoran-
thene) is required.
The PAH screening method utilizes
considerably less expensive equipment
than does Method 610 ($5,000 vs.
$25,000) and requires fewer materials
and less trained personnel.
A potential use for the total UV ap-
proach, employed in the PAH screening
method, is analysis of samples prior to
GC/MS or high performance liquid
chromatography determination. For
example, the total UV approach could be
used to screen sample extracts generated
by Method 610(prior to high performance
liquid chromatography analysis) to selec-
tively eliminate low level samples and
thereby reduce analytical costs.
Ralph Riggin and Paul Strup are with Battelle-Columbus Laboratories, Columbus,
OH 43201.
Stephen Billets was the EPA Project Officer (see below for present contact).
The complete report, entitled "Screening Methods for PAH Priority Pollutants in
Wastewater, "(Order No. PB 84-132 992; Cost: $11.50, subject to change) will
be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA22161
Telephone: 703-487-4650
For information contact Denis L. Foerst at:
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
U.S. GOVERNMENT PRINTING OFFICE: 1984-759-102/876
------- |