United States
Environmental Protection
Agency
Environmental Monitoring and
Support Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S4-84-039 June 1984
SEPA Project Summary
EPA Method Study 22 Method
612-Chlorinated Hydrocarbons
Jack R. Hall, J. Richard Florance, Dennis L Strother, and Marlene N. Wass
An interlaboratory study in which 20
laboratories participated was conducted
to provide precision and accuracy
statements for the proposed EPA
Method 612-Chlorinated Hydrocarbons
for measuring concentrations of the
Category 3 chemicals hexachloroethane,
hexachlorobutadiene, 2-chloronaphtha-
lene, 1 -2, dichlorobenzene, 1,3-dichloro-
benzene, 1,4-dichlorobenzene, 1,2,4-
trichlorobenzene, hexachlorobenzene,
and hexachlorocyclopentadiene in muni-
cipal and industrial aqueous discharges.
Hexachlorocyclopentadiene was elimi-
nated from the study because of its
instability in the solvent used to pre-
pare sample concentrates.
The study design was based on
Youden's plan for collaborative tests of
analytical methods. Three Youden pair
samples of the test compounds were
spiked into six types of test waters and
then analyzed. The test waters were
distilled water, tap water, a surface
water, and three different industrial
wastewater effluents. The resulting
data were statistically analyzed using
the computer program entitled "Inter-
laboratory Method Validation Study"
(IMVS). Using the mean recovery for
each subject compound, the mean
recoveries for the method were in the
range of 64 to 90%. Overall precision
was in the range of 26 to 41%, and
single-analyst precision was in the
range of 16 to 24%. In general, mean
recoveries, overall standard deviation
(S), and the single-analyst standard
deviations (SR) were directly propor-
tional to the true concentration levels.
In all cases, there was no evidence of a
statistically significant effect on accuracy
or precision due to water type.
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
EPA first promulgated guidelines
establishing test procedures for the
analysis of pollutants in 1973, following
the passage of the Federal Water Pollution
Control Act in 1972 by Congress. Pursuant
to the amendment and publication of
these guidelines, EPA entered into a
Settlement Agreement—the Consent
Decree—which required the study and, if
necessary, regulation of 65 "priority"
pollutants and classes of pollutants of
known or suspected toxicity to the biota.
Subsequently, Congress passed the
Clean Water Act of 1977, mandating the
control of toxic pollutants discharged into
ambient waters by industry.
In order to facilitate the implementation
of the Clean Water Act, EPA selected for
initial study 129 specific toxic pollutants,
113 organic arid 16 inorganic. The
organic pollutants were divided into 12
categories based on their chemical
structure. Analytical methods were
developed by EPA for these 12 categories
through in-house and contracted research
and may eventually be required for the
monitoring of the 113 toxic pollutants in
industrial wastewater effluents, as
specified by the Clean Water Act of 1977.
This report describes the interlaboratory
study of Method 612 Chlorinated Hydro-
carbons, which is proposed for the
Category 3 chemicals: hexachloroethane,
hexachlorobutadiene, 2-chloronaphtha-
lene, 1,2-dichlorobenzene, 1,3-dichloro-
benzene, 1,4-dichlorobenzene, 1,2,4-
trichlorobenzene, hexachlorobenzene,
and hexachlorocyclopentadiene. Hexa-
chlorocyclopentadiene, one of the category
-------
compounds, was dropped from the
method study because of its instability in
the solvent used to prepare sample
concentrates. The primary objective of
the study was to characterize the behavior
of Method 612 in terms of accuracy,
overall precision, single-analyst precision,
and effect of water type on accuracy and
precision. The study was conducted with
the cooperation of 20 participating
laboratories under auspices of the
Environmental Monitoring and Support
Laboratory (EMSLJ-Cincinnati.
The data were collected from the 20
laboratories according to Youden's colla-
borative testing design. Formal statistical
techniques compatible with the Youden
design were used to identify outliers,
estimate the method's accuracy and
precision, and test for the effect of water
type. The formal statistical analyses were
carried out using the IMVS computer
program. The information obtained from
the statistical analyses was summarized
and reduced to a descriptive form for the
purpose of interpretation and presentation.
Method 612 was developed by IT
Enviroscience under a contract with the
Physical and Chemical Methods Branch,
EMSL-Cincinnati. It requires extraction of
the pollutants with methylene chloride,
solvent exchange, concentration by
Kuderna-Danish, extract cleanup on
activated Florisil®, and subsequent gas
chromatographic (GC) analysis using
electron capture (EC) detection.
Procedure
The interlaboratory study design was
based on Youden's plan for collaborative
evaluation of precision and accuracy for
analytical methods. According to Youden's
design, samples are analyzed in pairs. A
Youden pair consists of two samples of
similar, but distinctly different concentra-
tions. The analyst is requested to perform
only a single analysis for each sample and
report only one value as in routine use of
the method.
Selection of Laboratories
Of the 20 participating laboratories, 19
were selected as the result of competitive
bidding after their technical capabilities
and experience in trace organic analyses
of wastewater had been evaluated. The
twentieth laboratory was a volunteer
from within the EPA.
Preparation of Ampuls
All starting materials were reagent-
grade quality or better. Distilled-in-glass
methyl ethyl ketone was the solvent.
Separate stock solutions for each of the
eight chlorinated hydrocarbons were
prepared by dissolving a precisely weighed
amount of the compound into Class A
volumetric glassware containing the
solvent. Appropriate volumes of the stock
solutions were mixed and diluted to
volume in 2000-mL volumetric flasks.
The flasks were refrigerated overnight at
4°C. The following day, approximately 3-
mL of the refrigerated solution was
transferred into 5-mL glass ampuls using
an all-glass or Teflon® delivery system.
After the ampuls were cooled in a freezer
at -30°C for three hours, they were sealed
by a professional glass blower using the
pull-and-twist technique.
True Value and Stability of
Concentrates
An important segment of this study
was to verify the true values and stability
of the chlorinated hydrocarbons in the
methyl ethyl ketone concentrates before
they were used in the study. To achieve
this verification, three replicate ampuls
were randomly selected from each
concentrate level batch and analyzed by
triplicate injection into a gas chromato-
graph equipped with a recording integra-
tor. To check stability these analyses
were carried out at 0, 45, and 90 days
after the ampuls were sealed and before
the study was started. All concentrate
values determined experimentally were
within instrumental error or the calculated
true values.
Preliminary Study
Previous EPA method studies have
shown that more realistic results can be
obtained if all participants thoroughly
understand the analytical and sample
handling procedures before undertaking
the full study. To familiarize the analyst
with these procedures, each of the 20
laboratories was sent a low-level Youden
pair of sample concentrates (different
from those to be used in the actual study)
for spiking distilled water, along with
instructions, a copy of the method, and
data report sheets.
The results of these analyses were
collected, statistically analyzed, and
discussed with the laboratories' repre-
sentatives in a one-day conference
meeting at EMSL-Cincinnati. The meeting
also allowed discussion of analytical
problems and clarification of any method-
ology procedures.
Actual Interlaboratory Study
A summary of the test design using
Youden's nonreplicate technique based
on pairs with slightly dissimilar analyte
concentrations is given below:
• Twenty laboratories were sent three
Youden pairs in,sealed glass ampuls
containing various levels of the
eight chlorinated hydrocarbons in
methyl ethyl ketone.
• When an analyst was ready to start
the analyses, the ampuls were
opened and aliquots were diluted to
volume in the appropriate water
types according to instructions.
• Each sample (ampul) was analyzed
only once. ;
• The six water types were analyzed
with and without spiking, and the
added level of constituent was
determined by difference and reported
as fig/L in each water sample.
• The three levels of chlorinated
hydrocarbons used in the study
were within the working range of
the method and represented the
range of levels one would'normally
expect to encounter in application of
the method to actual samples.
Description and Distribution of
Samples
The individual laboratories provided
their own samples of laboratory-distilled
water, tap water, and a local surface
water. The source for each surface water
is listed in the final report.
The wastewater effluent samples
representative of the industries of
concern were collected by the prime
contractor as grab samples in 55-gallon
stainless steel drums which had been
precleaned with acetone, methylene '
chloride, and distilled water. The unfiltered
wastewater samples were mixed and
transferred into one-quart bottles using
an all-Teflon system and stored in a
refrigerator at 4°C until shipment to the
laboratories. Each laboratory was sent 36
ampul concentrates (six sets of three
Youden pairs), seven one-quart bottles
each of the three industrial effluent types,
instructions, and data report sheets. The
wastewater samples were packed in ice
in coolers and sent by air freight to
minimize sample change and assure
comparability of the wastewaters from
laboratory to laboratory.
Analysis and Reporting
A water spiking technique involving a
water-soluble solvent concentrate (methyl
ethyl ketone is approximately 28%
soluble in water) was used in this study.
Each analyst was instructed to add
separate 2.0-mL aliquots of each concen-
trate to the bottle containing approximately
-------
1 L of water or wastewater. The spiked
sample was stirred for 15 minutes and
then handled as a routine sample in the
method. The results of the method's
measurement for each constituent (in
micrograms per liter of water) were then
corrected by subtracting any blank
sample reading and reported on data
sheets by ampul and water type. Method
612 was utilized by the participating
laboratories with no reported deviations.
The data were reviewed for complete-
ness and abnormal results, and were
then entered into the computer for
statistical treatment. Any laboratory
reporting unusually high or low data was
requested to review them for errors in
calculations, but was not told how its
reported data varied from the true values.
Treatment of Data
The objective of this interlaboratory
study was to obtain information about the
accuracy and precision associated with
measurements generated by Method
612. This objective was met through the
use of statistical analysis techniques
designed to extract and summarize the
relevant information about accuracy and
precision from the data reported by the
participating laboratories. The statistical
techniques were similar to the techniques
recommended in the ASTM Standard
Practice D2777-77.
The algorithms required to perform the
statistical analyses were integrated into
the IMVS system of computer programs.
The analyses performed by IMVS included
several tests for the rejection of outliers
(laboratories and individual data points);
summary statistics by concentration level
for mean recovery (accuracy); overall and
single-analyst standard deviation (preci-
• sion); determination of the linear relation-
ship between mean recovery and concen-
tration level; determination of the linear
relationship between the precision
statistics and mean recovery; and a test
for the effect of water type on accuracy
and precision.
Results and Discussion
The IMVS computer program was
designed to output the raw data in tabular
form and compile summary statistics
including: number of data points; true
value; mean recovery; accuracy as
percent relative error; overall standard
deviation; overall percent relative standard
deviation; single-analyst standard devia-
tion; and single-analyst percent relative
standard deviation. The statistical analyses
performed by-the IMVS program included
the determination of the linear relationship
between both the overall (S) and single-
analyst (SR) precision statistics and mean
recovery along With accuracy statements
based on the determination of the linear
relationship between mean recovery (X)
and concentration level. The results of
the regression analyses indicate apparent
linear relationships for each of the
above cases.
For all data for the eight compounds,
25% of the raw data were rejected as
determined by laboratory ranking and
individual outlier tests. The data rejection
was found to be nonuniform among
laboratories. More than 82% of the
rejected data were generated by eight of
the 20 participating laboratories, and for
one, 98% of its total raw data were
rejected as outliers. Nine laboratories'
raw data were rejected for between one
and four of the compounds studied in all
water types.
Regression equations for single-analyst
precision, overall precision, and accuracy
are presented in Table 1. Mean recoveries
of the subject compounds were in the
range of 64 to 90%. Overall precision was
in the range of 26 to 41%, and single-
analyst precision was in the range of 16
to 24%.
Based on the IMVS test for the effect of
water type on precision and accuracy,
there was no statistical significance
between distilled water and the corres-
ponding wastewater for any of the
associated parameters during EPA Method
Study 22.
Conclusions and
Recommendations
Based on the results of EPA Method
Study 22, EPA Method 612 is a viable
analytical method for measuring trace
concentrations of the eight Category 3
chemicals used. As a result of the
collaborative study conducted in the
IMVS data analysis, the following conclu-
sions and recommendations can be made
concerning Method 612.
• The accuracy of the method could be
expressed as a linear function of the
true concentration. In the majority of
equations, the slope represents the
percent recovery attributable to the
method.
• The precision of the method could be
expressed as a linear function of the
mean recovery. In the majority of
equations, the slope represents the
relative standard deviation attribu-
table to the method, both as single-
analyst and overall standard devia-
tions.
• The average mean recovery of the
six concentrations in sixwatertypes
for each compound compared well
with data generated on distilled
water industrial effluents during the
development of this method.
• No significant difference in method
performance was attributable to the
water type from which the analysis
was performed.
• One laboratory had trouble with the
extract concentration step using the
Kuderna-Danish apparatus and
some laboratories had to use peak
height measurements for quantita-
tion because occasional interference
peaks in wastewater created faulty
integration when using recording
integrators; ' - ' -
• When using the method, the analyst
should develop recovery and precision
data for the wastewater being
analyzed. These data should be
based on the concentration levels
determined or expected.
• In future interlaboratory studies,
very detailed instruction should be
given to the participating labora-
tories to ensure labeling of each
chromatogram. In this study it was
very difficult to interpret much of the
raw chromatographic data because
of inadequate labeling. Other points
to be emphasized in future studies
are that (a) blanks and spiked sam-
ples must be analyzed at the same
sensitivity and (b) calculations and
record keeping should be uniform or
consistent to aid in data interpreta-
tion.
-------
Table 1. EPA Method Validation Study ,
Regression Equations for Accuracy and Precision
Water Type
Applicable Cone. Range
Distilled Water
Single-Analyst Precision
Overall Precision
Accuracy
Tap Water
Single-Analyst Precision
Overall Precision
Accuracy
Surface Water
Single-Analyst Precision
Overall Precision
Accuracy
Waste Water J
Single-Analyst Precision
Overall Precision
Accuracy
Waste Water 2
Single-Analyst Precision
Overall Precision
Accuracy
Waste Water 3
Single-Analyst Precision
Overall Precision
Accuracy
Water Type
Applicable Cone. Range
Distilled Water
Single-Analyst Precision
Overall Precision
Accuracy
Tap Water
Single-Analyst Precision
Overall Precision
Accuracy
Surface Water
Single-Analyst Precision
Overall Precision
Accuracy
Waste Water 1
Single-Analyst Precision
Overall Precision
Accuracy
Waste Water 2
Single-Analyst Precision
Overall Precision
Accuracy
Wests Water 3
Single-Analyst Precision
Overall Precision
Accuracy
X = Mean Recovery
Hexachloroethane
(1.O2 - 14.80)
SR = 0.23X + 0.07
S =O.36X-0.00
X = 0.740-0.02
SR=0.33X-0.10
S = 0.34X - 0.04
X =0.780-0.09
SR = 0.17X + 0.51
S = 0.46X + 0.07
X =0.830 + 0.06
SR = 0.24X - 0.05
S = 0.26X + 0.07
X = 0.80C + 0.05
SR = 0.25X + 0.03
S = 0.45X - O.09
X =0.730-0.03
SR = 0.20X - 0.05
S = 0.27X - 0.03
X =0.810-0.13
Regression Equations
7, 3-Dichlorobenzene
(20.4O - 238.00}
SR = 0.21X - 1.03
S = 0.49X - 3.98
X =0.720 + 0.87
SR = 0.23X + 2.91
S =0.35X + 2.65
X =0.790 + 4.02
SR = 0. 17X - 0. 16
S =0.26X- 1.04
X =0.750 + 0.55
SR = O.J6X-0.47
S =0.34X + 0.03
X =0.780+4.47
SR = 0.1 9X + 2.61
S =0.39X + 0.22
X =O.75C+1.34
SR = 0. 15X + 0.58
S = 0.32X - 0. 13
X =0.710 + 2.52
Hexachlorobutadiene
(3.12 -36.80)
SR = 0. 18X + 0.08
S = 0.53X - 0. 12
X =0.610 + 0.03
SR = 0.29X - 0.23
S = 0.42X + 0. 14
X =0.640 + 0.18
SR = 0.30X + 0.05
S = 0.39X - 0.06
X =0.620 + 0.12
SR =0.25X + 0.21
S = 0.37X + 0.02
X =0.660 + 0.08
SR = 0.25X + 0.44
S = 0.39X + 0.39
X =0.630 + 0.39
SR = 0. 18X + 0. 14
S = 0.38X - 0.06
X =0.710 + 0.31
for Compounds 1-4
2-Chloronaphthalene
(19. 10 - 268.00)
SR =0.28X - 1.17
S =0.38X-1.39
X =0.750 + 3.21
SR = 0. 16X + 0.25
S =0.26X-0.71
X =0.740+3.41
SR = 0. 13X + 0.92
S =0. 18X + 1.20
X =0.770 + 3.67
SR = 0.12X+ 14.77
S =0.24X+ 8.18
X =0.740 + 11.64
SR = 0.1 2X + 8.51
S = 0.27X + 2.91
X =0.720 + 7.64
SR = 0. 14X + 2.92
S =0.26X + 2.79
X =0.740+2.55
1
1,2-Dichlorobenzene
(29.80-356.00)
SR = 0.22X - 2.95
S = 0.41 X- 3.92
X =0.850 + 0.70
SR = 0.17X+ 10.94
S =0.31X+ 5.35
X =0.870 + 6.70
SR = 0.30X - 4.00
S =O.32X-2.18
X = 0.81 C + 0.44
SR = 0. 13X + 4.87
S =0.26X + 4.26
X =0.850-0.52
SR = 0.22X + 1 .60
S =0.34X-0.63
X =0.830+3.41
SR = 0.21 X + 2.78
S =0.32X + 0,77
X = 0.870 + 5.53
for Accuracy and Precision for Compounds 5 - 8
1,4-Dichlorobenzene
123.00 - 324.00)
SR = 0. 16X - 0.48
S = 0.35X - 0.57
X =0.720+2.80
SR = 0.20X + 2.80
S = 0.26X + 5.88
X =0.750 + 4.15
SR=0.17X + 5.13
S = 0.42X + 0.22
X =0.730 + 5.56
SR = 0.24X - 0.81
S = 0.25X + 4.30
X =0.770+2.06
SR = 0.19X+1.11
S = 0.34X + 1.80
X =0.780 + 2.25
SR = 0.23X - 1.94
S = 0.34X - 3.58
X =0.780 + 1.72
1 ,2,4-Trichlorobenzene
(15.10-216.00)
SR = 0.23X - 0.44
S =0.40X-1.37
X =0.760+0.98
SR = 0.1 5X + 0.60
S =0.30X + 0.72
X =0.680 + 1.97
SR = 0.1 6X + 0.75
S = 0.43X + 0.42
X =0.740+2.40
SR = 0.39X - 1.97
S = 0.42X - 1.05
X =0.750+0.70
SR = 0.26X - 0.81
S =0.31X + 0.33
X = 0.800 - 0.21
SR = 0.15X+ 1.52
S =0.34X-0.71
X =0.780+2.11
Hexachlorobenzene
(1.29 - 14.90)
SR = 0. 14X + 0.07
S =0.36X-0.19
X =0.870-0.02
SR = 0.22X - 0.08
S = Q.32X - 0. 14
X = 0.920 - 0.08
SR = 0.08X + 0. 13
S = 0.22X - 0.01
X =0.970-0.06
SR = 0.11X + 0.19
S =0. 18X + 0. 16
X =0.870 + 0.17
SR = 0.25X -0.18
S =O.39X + 0.16
X = 0.990 + 0.09
SR = 0.24X -0.04
S = 0.38X + 0.03
X = 0.830 + 0.07
C = True Value for the Concentration
-------
-------
Jack R. Hall, J. Richard Florence, Dennis L. Strother, and Marlene N. Wass are
with IT Enviroscience, Knoxville, TN37919.
Edward L. Berg is the EPA Project Officer (see below).
The complete report, entitled"EPA Method Study 22, Method 612—Chlorinated
Hydrocarbons," (Order No. PB 84-187 772; Cost: $13.00, subject to change)
will be available only from:
National Technical Information Service
5285 Port Royal 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
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
EMSU0158933
U.S. GOVERNMENT PRINTING OFFICE: 1984-759-102/0978
------- |