EPA-600/4-84-052
June 1984
EPA METHOD STUDY 21,
METHOD 611--HALOETHERS
by
Carl R. McMillin, Roger C. Gable, Joseph M. Kyne,
Richard P. Quill, Arthur D. Snyder, and James A. Thomas
MONSANTO COMPANY
Dayton Laboratory
Dayton, Ohio 45407
CONTRACT NO. 68-03-2633
Project Officer
Raymond Wesselman
Quality Assurance Branch
Environmental Monitoring and Support Laboratory
Cincinnati, Ohio 45268
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
DISCLAIMER
The information in this document has been funded wholly or in part
by the United States Environmental Protection Agency under con-
tract 68-03-2633 to Monsanto Research Corporation. It is been
subject to the Agency's peer administrative review, and it has
been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement
or recommendation for use.
ii
-------
FOREWORD
Environmental measurements are required to determine the quality
of ambient waters and the character of waste effluents. The En-
vironmental Monitoring and Support Laboratory (EMSL)-Cincinnati
research responsibilities are to:
- Develop and evaluate techniques to measure the presence and
concentration of physical, chemical, and radiological pollut-
ants in water, wastewater, bottom sediments, and solid waste.
• Investigate methods for the concentration, recovery, and
identification of viruses, bacteria, and other microorganisms
in water.
• Conduct studies to determine the responses of aquatic organ-
isms to water quality.
• Conduct an Agency-wide quality assurance program to assure
standardization and quality control of systems for monitoring
water and wastewater.
This publication reports the results of EPA's interlaboratory
method study for bis(2-chloroisopropyl)ether (BCIPE), his(2-chloro-
ethyl)ether (BCEE),bis(2-chloroethoxy)methane (BCEXM), 4-chloro-
phenyl phenyl ether (CPPE), 4-bromophenyl phenyl ether (BPPE).
Federal agencies, states, municipalities, universities, private
laboratories, and industry should find this interlaboratory study
useful in monitoring and controlling pollution in the environment.
Robert L. Booth, Director
Environmental Monitoring and Support Laboratory
iii
-------
ABSTRACT
This report describes the interlaboratory study of an analytical
method which detects haloethers in water. EPA Method 611 -- Halo-
ethers, consists of a liquid/liquid extraction using methylene
chloride, an evaporation step using Kuderna-Danish (K-D) evapora-
tors, a cleanup procedure using Florisil sorbent, another K-D
evaporation of the fraction from the Florisil column, and subse-
quent analysis by gas chromatography (GC) using a halide-specific
detector. The six concentrations (three Youden pairs) of spiking
solutions used in this study contained BCIPE, BCEE, and BCEXM,
CPEE, and BPPE. Six water types were used in the study: distil-
led, tap, surface, and three different industrial wastewaters.
Statistical analyses and conclusions in this report are based on
analytical data obtained by 20 collaborating laboratories.
Participating laboratories were selected based upon technical
evaluation of proposals and upon the analytical results of pre-
study samples. The data*obtained from the interlaboratory study
were analyzed employing EPA's series of computer programs known as
the Interlaboratory Method Validation Study (IMVS) system, which
basically implements ASTM Standard D 2777. The statistical
analyses included tests for the rejection of outliers, estimation
of mean recovery (accuracy), estimation of single-analyst and
overall precision, and tests for the effects of water type on
accuracy and precision.
This report was submitted in fulfillment of Contract 68-03-2633
by Monsanto Company under the sponsorship of the U.S. Environmen-
tal Protection Agency and covers a period from January 1979 to
March 1980.
iv
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables vii
Acknowledgements ix
1. Introduction 1
2. Conclusions 3
3. Recommendations 8
4. Description of Study 9
Selection of participating laboratories 9
Study design 10
5. Statistical Treatment of Data 16
Rejection of outliers 16
Statistical summaries 20
Regression analysis of basic statistics 29
Comparison of accuracy and precision
across water types 31
6. Results and Discussion 39
Accuracy 39
Precision 43
Effects of water types 51
Responses to questionnaire 51
References 59
Appendices
A. Test method - haloethers - Method 611 60
B. Raw data 69
C. Effect of water type 85
D. Other MC findings 91
v
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
age
94
94
97
97
98
98
99
99
101
101
102
102
106
FIGURES
Chroraatogram of wastewater blank extraction
(after Florisil cleanup)
Chromatogram of spiked wastewater extraction. .
Chromatogram of wastewater D extract before
Florisil cleanup
Chromatogram of spiked wastewater D extract
after Florisil cleanup
Chromatogram of wastewater E extract before
before Florisil cleanup
Chromatogram of spiked wastewater E extract
after Florisil cleanup
Chromatogram of wastewater F extract before
Florisil cleanup
Chromatogram of spiked wastewater F extract
after Florisil cleanup
Chromatogram of wastewater G extract before
Florisil cleanup
Chromatogram of spiked wastewater G extract
after Florisil cleanup
Chromatogram of wastewater H extract before
Florisil cleanup
Chromatogram of spiked wastewater H extract
after Florisil cleanup
Hall response at various reactor temperatures .
vi
-------
TABLES
Number Page
1 Regression Equations for Accuracy and
Precision 5
2 Revised Regression Equations for
Wastewater 2 7
3 Participating Laboratories 11
4 Spiking Solution Concentrations 14
5 Youden Laboratory Ranking Procedure for
4-Chlorophenyl Phenyl Ether in Water 3. . . . 19
6 Critical Values for Thompson's T (Two-Sided
Test) when Standard Deviation is Calculated
from the Same Samples 21
7 Results of Tests for Individual Outliers
(4-Chlorophenyl Phenyl Ether in Water 3). . . 21
8 Statistical Summary for £/S(2-Chloroisopropyl)-
Ether Analysis by Water Type 24
9 Statistical Summary for BIS(2-Chloroethyl)Ether
Analyses by Water Type 2b
10 Statistical Summary for £IS(2-Chloroethoxy)-
Methane by Water Type 26
11 Statistical Summary for 4-Chlorophenyl Phenyl
Ether Analyses by Water Type 27
12 Statistical Summary for 4-Bromophenyl Phenyl
Ether Analyses by Water Type 28
13 Method 611 Accuracy (%) 40
14 Method 611 Precision (%RSD) 45
15 Method 611 Precision (% RSD-SA) 48
vii
-------
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
TABLES (continued)
Pa9e
Summary of the Test for Difference Across Water
Types 52
Laboratory Analytical Conditions (Ordered in
Detector Groupings) 55
Raw Data for BIS(2-Chloroisopropyl)Ether by
Water Type 70
Raw Data for BIS(2-Chloroethyl)Ether by Water
Type 73
Raw Data for BIS(2-Chloroethoxv)Methane by Water
Type "! 76
Raw Data for 4-Chlorophenyl Phenyl Ether by Water
Type 79
Raw Data for 4-Bromophenyl Phenyl Ether by Water
Water Type 82-
Effect of Water Type on BIS(2-Chloroisopropyl
Analysys 86
Effect on Water Type on BIS(2-Chloroethyl)Ether
Analysis 87
Effect of Water Type on BIS(2-Chloroethoxy)Methane
Analysis 88
Effect of Water Type on 4-Chlorophenyl Phenyl Ether
Analysis 89
Effect of Water Type on 4-Bromophenyl Phenyl Ether
Analysis 90
Replication Data 93
Chromatographic Conditions for Replication
Analyses 93
Summary of Wastewater Inferferences 103
Interfering Compounds 104
vi i i
-------
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the hard work and cooperation
of the staff of the Quality Assurance Branch, EMSL-Cincinnati,
who assisted in the study. They especially acknowledge the excel-
lent technical assistance, guidance, and understanding of Raymond
Wesselman of EMSL-Cincinnati. Also acknowledged is the work of
Dr. Thomas Bishop at Battelle Columbus Laboratories, Columbus, Ohio,
for statistical analysis of the data under contract 68-03-2624.
ix
-------
SECTION 1
INTRODUCTION
The EPA's analytical laboratories gather water quality data to pro-
vide information on water resources, to assist research activities,
and to evaluate pollution abatement activities. The success of
the Agency's pollution control activities, particularly when legal
action is involved, depends upon the reliability of the data pro-
vided by the laboratories.
Under provisions of the Clean Water Act, the EPA promulgates guide-
lines 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 clas-
ses of pollutants to be toxic under Section 307(a) cf 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 measurement. The priority pollutants
analyzed by Method 611 in this report are the study haloethers:
BCIPE, BCEE, BCEXM, CPPE and BPPE.
EMSL-Cincinnati develops analytical methods and conducts a quality
assurance program for water laboratories to maximize the reliabil-
ity and legal defensibility of water quality information collected
by EPA laboratories. This responsibility is assigned to the Qual-
ity Assurance Branch (QAB) which conducts interlaboratory studies
on the methods in order to generate precision and accuracy data.
This report presents the results of interlaboratory study 21, con-
ducted for the USEPA by the prime contractor; Monsanto Company
(MC).
1
-------
Monsanto Company conducted the study in three phases. Phase I in-
volved the analysis of the prestudy samples by 20 participating
laboratories. Two samples were analyzed for each of the five halo-
ethers. A medium concentration sample was analyzed in distilled
water supplied by the participating laboratories and a low level
sample was analyzed in a wastewater sample supplied by MC. The
objective of Phase 1 was to familiarize laboratories with Method
611 and to identify potential problems associated with the analyt-
ical methodology. A short report, including the data obtained and
any potential problems encountered, was received from each subcon-
tracting laboratory by MC at the completion of Phase I.
Phase II consisted of a prestudy conference held at EMSL-Cinci-
natti, on May 16, 1979 to which each subcontracting laboratory
sent at least one participant. The prestudy conference was de-
signed to examine the results of Phase I and to discuss any pro-
blems encountered in the methodology.
Phase III was the formal interlaboratory study. Five haloethers
were analyzed at six concentrations (three Youden pairs) in six
different water matrices. Each participating laboratory supplied
its own reagent grade water, tap water and surface water. MC sup-
plied the three industrial wastewaters. In addition, the partici-
pating laboratories performed analyses of all water blanks with no
spiked compounds. Each participating laboratory then issued a re-
port to Monsanto Company containing all data obtained, copies cf
all chromatograms, and comments.
The final step in the study was a statistical analysis of data by
Battelle Columbus Laboratories, Columbus, Ohio, under contract
68-03-2624 employing U.S. EPA's IMVS computer programs.
2
-------
SECTION 2
CONCLUSIONS
The object of this study was to characterize the performance of
Method 611 in terms of accuracy, overall precision, single-analyst
precision and the effect of water types on accuracy and precision.
Through statistical analyses of 3,600 analytical values, estimates
of accuracy and precision were made and expressed as regression
equations, which are shown in Table 1. One measure of the perform-
ance of the method is that 16.3% of the analytical values were re-
jected as outliers. Of these, 6.1% were rejected through applica-
tion of Youden's laboratory ranking procedure and 10.2% were re-
jected employing the Thompson-!'-test.
The accuracy of the method is obtained by comparing the mean re-
covery to the true values of the concentration. It is expressed
as percent recovery and ranges from 56% to 85% for all five ana-
lytes in all six water.
The overall standard deviation indicates the precision associated
with measurements generated by a group of laboratories. The per-
cent relative standard deviation (% RSD) ranges from 32% to 53%.
The single-analyst standard deviation indicates the precision
associated within a single laboratory. The percent relative
standard deviation for a single analyst (% RSD-SA) ranges from
15% to 31%.
A statistical comparison of the effect of water type was performed
indicating a statistically significant difference for six of the
3
-------
analyte/water matrix combinations. Of these six cases, a practi-
cal significant difference was established only for 4-chlorophenyl
phenyl ether in wastewater 2. This combination also exhibited the
lowest accuracy and highest precision (lowest % RSD and RSD-SA)
values of all 30 analyte/water pairs.
In general, the slopes of the regression equations presented in
Table 1 provided an excellent fit to the data especially in the
middle and high Youden concentrations pairs. Recovery and pre-
cision data at the lowest concentrations suffered from a lack of
detection sensitivity and from the presence of background inter-
ferences in the blank samples. The fit of the regression equations
for accuracy and precision reinforces the assumption that percent
recovery is independent of concentration and that absolute recovery
is a linear function of the analyte concentration.
A detailed examination of the data indicated a background inter-
ference problem for wastewater 2 where the recoveries for the low
Youden pairs were 541% and 442% for BCEE, and 46% and 287% for
4-CPPE. For this reason a new set of linear regression equations
were generated omitting the low Youden pair data. The revised re-
gression equations for BCEE and 4-CPPE in wastewater 2 are pre-
sented in Table 2.
In preliminary studies by the prime contractor, it was found that
each of the haloether compounds responded best to the Hall 310 de-
tector at different temperatures. It was therefore necessary to
find a compromise reactor temperature that gave good response for
all five compounds. The newer Hall detectors use metallic reactor
tubes rather than quartz tubes. These metallic tubes are said to
have a catalytic effect which eliminates the need for such criti-
cal temperature selection.
4
-------
TABLE 1. REGRESSION EQUATIONS FOR ACCURACY AND PRECISION
WATER TYPE B IS(2-CHLOROISOPROPYL )ETH IMS(2-CHL0R0E THYL )E THER B1S(2-CMLOROE THOXY )ME THAN 4-CHLOROPHENYL PHENYL ETH
APPLICABLE CONC. RANGE (2.40 - 624.011) (1.40 - b0?.()0) (1.00 - b?H.OO) (fi.bO - 489.00}
UISFILLEO WATER
SINGLE-ANALYST PRECISION SR = 0.20X * 1.0b SR = 0.19X + 0.2H SR = 0.20X + 0.15 SR = O.IHX ~ 2.13
UVrHALL PRECISION S = 0.36X + 0.79 S = 0.3SX ~ 0.36 S = 0.33X ~ 0.11 S = 0.41X * O.bb
ACCURACY X = 0.85C ~ 1.6/ X = 0.81C + 0.54 X = 0./1C ~ 0.13 X = U.U^C + 1.4/
TAP WATER
SINGLE-ANALYST PRECISION SR = O.lbX + 0.03 SR = 0.18X + 0.25 SR = 0.21X + 0,21 SR - 0.17X ~ 1.22
OVERALL PRECISION S = 0.36X + O.bb S = 0.40X + 0.18 S = 0.38X + 0.69 S = 0.39X ~ 0./8
ACCURACY X = 0.7HC + 0.99 X - 0.72C + 0.48 X - 0.67C ~ 0.69 X = 0./5C + 0.63
SURFACE WATER
SINGLE-ANALYST PRECISION SR = 0.29X + 0.77 SR = 0.2/X - 0.06 SR = 0.29X - 0.08 SR = 0.22X ~ 0.83
OVERALL PRECISION S = 0.47X ~ 0.23 S = O.bOX ~ 0.U9 S = 0.53X ~ 0.47 S = 0.42X + 0.14
ACCURACY X = 0.//C + 0.42 X * O.b/C ~ 0.39 X = 0.60C 0./4 X = 0.6/C ~ 1.14
WASTE WATER 1
SINGLE-ANALYST PRECISION
OVERALL PRECISION
ACCORACY
WASTE WATtK 2
SINGLE-ANALYST PRECISION
OVERALL PRECISION
ACCURACY
WASTE WATER 3
SINGLE-ANALYST PRECISION
OVERALL PRECISION
ACCURACY
SR = 0.24X ~ 0.15
S = 0.40X + 1.93
X * 0.73C ~ 2.00
SR = 0.26X + 0.07
S = 0.41X + 0.06
X -- 0.69C * 0.25
SR = 0.23X + 0.43
S = 0.48X + 0.54
X - 0.69C ~ 0.69
SR = 0.25X + 0.78
S = 0.43X + 0.40
X - 0.6SC ~ 0.97
SR = 0.29X ~ 0.09
S = 0.52X ~ 1.00
X » 0.83C ~ 1.66
SR = O.lbX ~ 2.26
S = 0.3SX * 4.12
X = 0.72C ~ 7.77
SR = 0.22X + 1.37
S = 0.34X ~ 2.10
X « 0.71C ~ 2.33
SR = O.lbX Mb.99
S = 0.32X ~17.01
X * 0.S6C *20.40
SR * 0.28X ~ 0.22
S « 0.42X ~ 0.33
X = 0.80C ~ 0.39
SR = 0.23X + 0.04
S = 0.41X + 0.06
X = 0./2C ~ 0.14
SR = 0.26X ~ 0.18
S * 0.36X ~ O./O
X = 0.67C ~ 0.97
SR = 0.28X ~ 0.89
S = 0.38X ~ 0.97
X = 0.69C ~ 1.51
X = MEAN RECOVERY
C •- TRUE VALUE FOR THE CONCENTRATION
-------
TABLE 1 (continued)
WATER TYPt 4-HROMOPHFNYL PHENYL E THE
APPLICABLE CONC. RANGE (2.80 - 626.00)
DISTILLED WATER
SINGLE-ANALYST PRECISION SR = 0.25X + 0.21
OVERALL PRECISION S = 0.47X + 0.37
ACCURACY X = (UibC ~ 2.bb
TAP WATER
SINGLE-ANALYST PRECISION SR = 0.22X + U.33
UVERALL PRECISION S = 0.47X + 0.62
ACCURACY X = 0.82C ~ 1.87
SURFACE WATER
SINGLE-ANALYST PRECISION SR = 0.27X + 0.b9
OVERALL PRECISION S = 0.49X ~ 0.47
ACCURACY X = 0.7HC ~ 2.10
WASTE WATER 1
SINGLE-ANALYST PRECISION SR = 0.30X + 0.33
OVERALL PRECISION S = 0.48X + 0.61
ACCURACY X * 0.77C + 2.16
WASTE WATER 2
SINGLE-ANALYST PRECISION SR = 0.29X ~ 1.26
OVERALL PRECISION S = O.blX + 0.45
ACCURACY X = O.H1C ~ 2.30
WASTE WATER 3
SINGLE-ANALYST PRECISION SR = 0.31X + 0.13
OVERALL PRECISION S = 0.4/X + 0.22
ACCURACY X = 0.79C + 1.68
X = MEAN RECOVERY
C = 1 RUE VALUE FUR THE CONCENTRATION
-------
TABLE 2. REVISED LINEAR REGRESSION EQUATIONS FOR WASTEWATER 2
BCEL
4-CPPE
Single-Analyst Precision
Overall Precision
Accuracy
SR = 0.10X - 3.47 SR = 0.11X + 6.00
S = 0.39X - 2.26 S - 0.49X - 8.98
X = 0.72C + 8.61 X - 0.69C + 9.75
A preliminary study of eight different effluent/wastewaters was
conducted to determine the effectiveness of the method in elimi-
nating potential interferences to the analysis of haloethers and
identifying some of the remaining interferences. This study
vividly demonstrated the effectiveness of the Florisil cleanup
in removing potential interferences. In many cases, very large
potential interferences observed in samples after extraction and
concentration were totally eliminated by the Florisil cleanup. A
gas chromatographic/mass spectrometric (GC/MS) analysis of samples
after cleanup also identified a number of compounds which were not
observed in the halide-specific detector chromatograms. Compounds
identified in the wastewaters as posing interference problems in-
cluded a nonpriority pollutant haloether and a cyclic chlorinated
hydrocarbon. Large quantities of nonhalogen containing hydrocar-
bons were also found to give a response, though greatly reduced,
with the halide-specific detectors. Minor changes in chromato-
graphic conditions were generally able to separate potential inter-
ferences from the haloethers even with the worst case wastewater
used. In final discharge waters, it is not anticipated that inter-
ferences will pose a significant problem in the analysis of halo-
ethers with this method.
In general, the most sensitive portion of the method is the
Kuderna-Damsh concentration step. It requires some analyst care
and experience to conduct this concentration step in a reproduci-
ble manner.
7
-------
SECTION 3
RECOMMENDATIONS
Method 611 is recommended for the analysis of haloethers in munici-
pal and industrial wastewaters. The matrix effects are significant
only at low concentration levels.
Care should be taken in the Florisil cleanup and K-D concentration
steps. Analyst care and experience is required to conduct the con-
centration step in a reproducible manner.
Special care should be taken to break the emulsions developing in
the extraction step of the analysis to prevent loss of analyte.
8
-------
SECTION 4
DESCRIPTION OF STUDY
SELECTION OF PARTICIPATING LABORATORIES
As prime contractor, Monsanto Company sent requests for quotation
(RFQ) to approximately 150 laboratories which had been identified
as potential subcontractors for this interlaboratory study. The
RFQ contained a Scope of Work, a description of the projected tim-
ing of the required analyses, and a copy of the analytical method.
The detailed writeup for Method 611 as published by EPA is pre-
sented in Appendix A of this report. Interested laboratories were
asked to respond to the RFQ by providing the following information
on:
* Facilities available at the laboratory, including all
instrumentation to be used for the study.
• Previous experience in carrying out the types of analyses
specified in the Scope of Work for the compounds of
interest.
* Handling procedures for working with hazardous and poten-
tially hazardous chemicals.
* The organization and managerial structure of the laboratory,
identifying those personnel involved in managing this study.
* The analyst involved in the analyses to be per-
formed, including his/her experience.
9
-------
• Quality control/quality assurance procedures and good
laboratory practices followed by the laboratory.
Approximately 30 proposals were received in response to the RFQ.
The proposals received were ranked, and the 20 most qualified
laboratories were selected for participation. Table 3 lists the
participating laboratories for the Method 611 interlaboratory
study. Throughout this report, data provided by these laborator-
ies will be identified only by an anonymous code number.
STUDY DESIGN
Two preliminary samples were sent to the participating labora-
tories. One was supplied at a medium level to be analyzed in
distilled water to assure that the method could be properly imple-
mented. The second sample was at a low level to be spiked into a
liter of wastewater that was supplied. This sample was to find -
the method problems under adverse conditions.
The analysts from these laboratories met in Cincinnati on May 16,
1979 to discuss the procedures and potential problems. The dis-
cussion included elements dealing with problems in achieving low
enough detection limits, the necessity of Florisil standardization
per the Federal Register method, reactor tube composition, and
other questions as to which elements of the study were fixed and
which elements could be optimized by the individual laboratories.
About a month after the prestudy conference, agreement was reached
by the EPA and MC concerning which components were to be rigidly
fixed. The method study samples and wastewaters were then sent to
the laboratories.
Rigidly set conditions included the specifications of column pack-
ing material, again excluding EC detectors as not being halogen
specific, and specifying that the Federal Register method be
1C
-------
TABLE 3. PARTICIPATING LABORATORIES
Analytical Development Co.
1875 Willow Park Way
Monument, Colorado 80132
Analytical Research Laboratories,
Inc.
160 Taylor Street
Monrovia, California 91016
Battelle (Columbus Laboratories)
505 King Avenue
Columbus, Ohio 43201
CDM Environmental Sciences
Division
6132 West Fond du Lac Avenue
Milwaukee, Wisconsin 53218
Environmental Research and
Technology
2625 Lowingate Road
Westlake Village,
California 91361
Environmental Research and
Technology
696 Virginia Road
Concord, Massachusetts 01742
Environmental Research Group
117 North First
Ann Arbor, Michigan 48103
Environmental Science and
Engineering
P.O. Box 13454
Gainesville, Florida 32604
Finnigan Institute
11750 Chesterdale Road
Cincinnati, Ohio 45246
Hydrosciences, Incorporated
363 Oik Hood Road
Westwood, New Jersey 07675
New Mexico Scientific
Laboratory Systems
700 Camino de Salud
Albuquerque, New Mexico 87106
Orlando Laboratories, Inc.
90 West Jersey Street
Orlando, Florida 32856
Raltek Science Services
33 01 Kinsman Boulevard
Madison, Wisconsin 53707
Southwest Research Institute
8500 Culebra Read
San Antonio, Texas 78284
Spectrix Corporation
7408 Fannin
Houston, Texas 77054
Texas Instruments, Inc.
P.O. Box 5621-MS 949
Dallas, Texas 75265
Versar, Inc.
6621 Electronic Drive
Springfield, Virginia 22151
Water and Air Research, Inc.
P. O. Box 52329
Jacksonville, Florida 32201
West Coast Technical Service
17605 Fabrica Way
Cerritos, California 90701
Wilson Laboratories
528 North 9th Street
Salina, Kansas 67401
11
-------
followed in detail. The temperature program, flow rates, and sol-
vent compositions were given as guidelines only since a variety of
detectors were being used.
Suggested guidelines were also given for parameters such as fur-
nace temperature on the Hall electrolytic conductivity detectors.
MC has used a Hall 310 detector for many analyses and it was
found the sensitivity of the detector varied significantly as a
function of the temperature and differed from compound to compound.
Only two of the 20 laboratories, however, used this detector. Be-
cause a wide variety of detectors were used in the study, a great
deal of flexibility was allowed for individual optimization of the
instrument conditions described within the scope of the Federal
Register method.
Three industrial wastewaters were selected for the interlaboratory
study. Each wastewater was obtained from a different chemical
company which either produces haloethers or had the potential of
haloether byproducts in the production of other chemicals. Waste-
waters #1 and #2 were raw effluents before treatment, and waste-
water #3 was diluted effluent destined for deep well injection.
These were selected as worst case examples to evaluate the method
in the presence of the types of interferences which might be ex-
pected by NPDES permit holders analyzing for haloethers. The final
treated discharge waters would generally have lower levels of these
interferences. Wastewater #2 contained the most significant quan-
tities of interfering compounds.
Each wastewater was thoroughly mixed, filtered, and dispensed in
one liter Weaton bottles equipped with Teflon lid liners. Each of
the participating laboratories was sent seven one liter bottles of
each wastewater. Six of these bottles were each spiked with one
of the six spiking solutions, while the seventh bottle served as
the blank.
12
-------
In addition to the three wastewaters, each laboratory supplied
its own tap water, reagent grade water, and surface water samples.
After spiking, each laboratory had 42 different samples (including
blanks), for analysis.
The study design was based on Youden's non-replicate plan [1] for
collaborative evaluation of precision and accuracy for analyti-
cal methods. According to Youden's design, samples are analyzed
in pairs, each sample of a pair with slightly different concentra-
tions of the constituents. The analyst is directed to do a single
analysis and report one value for each sample. Analyses in reagent
grade water evaluated the proficiency of the analyst to use the
method on a sample free of interferences; analyses in the other
waters were intended to reveal the effects of interferences on the
method.
Six spiking solutions were made such that three different concen-
tration ranges were each represented by two different solutions
(Youden pairs). Solution numbers 1 and 5 had haloether concen-
trations near the minimum detectable limits. Solution numbers 2
and 6 had concentrations around 100 ppb. Solution numbers 3 and 4
contained haloethers at levels about five times the intermediate
levels. Table 4 shows the individual haloether concentrations
for each spiking solution.
A problem with CPFE occurred when the solutions were made. CPPE
was purchased in sealed glass ampules containing 10, 20, or 50 mg
of CPPE. It was assumed that the weight listed on the label was a
precise weight, so the contents were simply rinsed into the spiking
solutions. Initial analyses of the solutions showed considerable
discrepancy between the "theoretical" concentrations and the ana-
lyzed values for CPPE. Subsequent ampules were checked for actual
weights, and it was found that they normally contained much more
than the stated weight. A check with the supplier confirmed that
13
-------
TABLE 4. SPIKING SOLUTION CONCENTRATIONS
Con contra t ion, pg/mL,
Solution
BCIPE
BCEE
BCEXM
CPPE
BPPE
#1
3.0
1.4
1.4
14.5
2.8
#2
132
108
106
94
145
#3
486
602
398
489
552
#4
624
402
528
424
626
#5
2.4
1. 6
1.0
6.6
3.8
#6
92
87
126
120
116
the stated weights were intended as approximations showing the
minimum weight. The collaborating laboratories were informed of
this potential source of error when making their own standards.
The initial analytical values determined in triplicate by direct
injection were used as the true values for CPPE for this study.
The spiking solutions were heat sealed in 5-mL hard glass ampules.
Each ampule contained 1.5 mL of solution, of which 1.0 mL was used
to spike a liter of water. To prevent loss of analyte or acetone
solvent, the ampules were cooled under a liquid nitrogen stream
while being sealed. The ampules were refrigerated until used or
were sent to the subcontract laboratories. Each laboratory was
then sent a set of spiking solutions containing six ampules of
each of the six solutions.
In the methods development phase of this program, stability studies
were conducted by MC to select the optimum solvent for preparation
of the spiking solutions. Acetone was selected as the solvent and
it was determined that the haloethers were stable up to 90 days.
The spiking solutions for the validation study were prepared well
in advance of the time they would be used, and the turnaround
time of the participating laboratories was long; therefore, it
14
-------
was decided to analyze all the solutions at about the same time
the 20 laboratories were using them in addition to the shorter
term (90 day) stability tests. In this way, it could be assured
that haloether degradation or losses would not contribute to vari-
ations in data from the various laboratories. The additional ana-
lyses for stability of the spiking solutions were conducted 230
days after the solutions were made. Even after 230 days, the
spiking solutions were stable.
MC interacted with all of the laboratories involved in the study.
This interaction varied from verbal discussions of the potential
problems of analysis to sending a MC employee to five laboratories
to assist in troubleshooting. The major difficulties encountered
were interfacing the Hall detectors to a wide variety of gas chro-
matographs and lack of proper sensitivity.
At the conclusion of the study, a questionnaire was sent to each
of the participating laboratories requesting information on the
operating conditions used for the analyses, problems encountered
with the method, and any other variables associated with the con-
duct of the method, for example, how emulsions were broken in the
methylene chloride separations. Comparisons of the detector type
to the quality of results obtained showed little correlation.
Initial feelings were that the Hall 700A detector would produce
superior results to the Hall 700 and Hall 310 detectors, but this
was not found to be significant. Users of detectors other than
the Hall models said their detectors had nonlinear responses. The
data generated, however, were similar in quality to the Hall detec-
tor data.
The raw data reported by the 20 laboratories are presented in Appen-
dix B of this report. The values reported have been corrected for
the blank values. The asterisked pieces of data were rejected as
outliers for further statistical analysis. Details of the methods
for detection of outliers are presented in Section 5 of this report.
15
-------
SECTION 5
STATISTICAL TREATMENT OF DATA
Data obtained from the interlaboratory study were subjected to
statistical analyses by the Battelle Columbus Laboratories,
Columbus, Ohio, under EPA Contract 68-03-2624. The analyses were
performed employing EPA's Interlaboratory Method Validation Study
(IMVS) system [2] of computer programs which was designed to im-
plement ASTM procedure D2777, "Standard Practice for Determination
of Precision and Bias of Methods of Committee D-19 on Water" [3].
The analyses conducted using the IMVS system included tests for
the rejection of outliers (both whole laboratories for a water type
and individual data points), estimation of mean recovery (accuracy),
estimation of single-analyst and overall precision, and tests for
the effects of water type on accuracy and precision.
REJECTION OF OUTLIERS
An outlying observation, or "outlier," is a data point that
appears to deviate markedly from other members of the sample in
which it occurs. Outlying data points are very commonly encount-
ered during interlaboratory test programs. If they are not re-
moved, they can result in a distortion of the accuracy and preci-
sion statistics which characterize the analytical method. These
outlying points cannot be removed inaiscriminantly, however, be-
cause they may represent an extreme manifestation of the random
variability inherent in the method.
16
-------
ASTM procedure E178-80, "Standard Practice for Dealing with
Outlying Observations," [4] and ASTM procedure D2777-77 [3 J
present explicit statistical rules and methods for identifi-
cation of outliers.
Data from outlying laboratories for a particular water type were
rejected employing Youden's laboratory ranking test procedure [3,
5J at the 5% level of significance. Data remaining after the
laboratory ranking procedure were subjected to individual outlier
tests. After zero, missing, "less than" and "nondetect" data
were rejected as outliers, the average and standard deviation for
the remaining data were calculated. The remaining data were
examined for additional outliers employing the outlier rejection
test constructed by Thompson [6J. Data rejected as outliers for
this study are identified by an asterisk in the tables of raw
data shown in Appendix B.
Youden's Laboratory Ranking Procedure
Using the data for each water type, Youden's laboratory ranking
test [3, 5] was performed at the 5% level of significance. The
Youden laboratory ranking procedure requires a complete set of
data from each laboratory within each water type. Missing data
from laboratory "i" for water type "j" were replaced by the
following procedure. Letting X- denote the reported measurenien
x j )>
from laboratory "i" for water type "j" and concentration level
it is assumed that
Yi
X •, = p . • C, J • L- • c • (1 )
ljk k i i]k v
where and Yj are fixed parameters which determine the effect
of water type "j;" L^ is the systematic error due to laboratory
"i," and r. . is the random intralaboratory error.
1 j K
-------
Taking natural logarithms, it follows that
*n Xijk = £n Pj + Yj An Ck + £n Li -f £n ^ijk (2)
which is a linear regression model with dependent variable £n
and independent variable £n C^. (Details and justification for
this model are discussed in the section "Comparison of Accuracy
and Precision Across Water Types.")
The natural logarithms of the individual laboratory's data were
regressed against the natural logarithms of the true concentra-
tion levels for the six ampuls in each water type. The predicted
values for 2n~X.were obtained from the regression equation, and
X J Jv
the missing values for X •, were estimated by £ .. = exp(£n^X. • , ).
^ ljk J ljk ijk'
(For complete details of this procedure, see Reference 2.)
An example of the use of Youden's laboratory ranking procedure
is presented in Table 5, where the rankings of the values for
4-chlorophenyl phenyl ether in water 3 are listed for each labora-
tory and for ampuls 1 through 6. For 20 laboratories and 6 ampuls,
the upper and lower critical limits of the sums of the rankings
are 104 and 22. If the sum of the rankings of any laboratory
equals or exceeds 104, or is equal to or less than 22, that labora-
tory's data are rejected for all determinations for that analyte
(4-chlorophenyl phenyl ether) in that water (water 3). From
Table 5 it is apparent that the data from laboratories 2, 11, and
13 must be rejected. The estimated missing data were then removed
from the data sets.
Test for Individual Outliers
The data remaining after rejection of zero, missing, "less than,"
and "nondetect" data were subjected to an individual outlier test
based on calculation of the average value, X, for each ampul and
the standard deviation of the remaining values.
18
-------
TABLE 5. YOUDEN LABORATORY RANKING PROCEDURE FOR
4-CHLOROPHENYL PHENYL ETHER IN WATER 3
Labor-
atory
Rankinq
values
Cumulative
number
Ampul 1
Ampul 2
Ampul 3
Ampul 4
Ampul 5
AjtiduI 6
score
1
9
9
14
11
9.5
10
62 .5
16*
2
4
1
1
2
5
3
3
13
10
10
7
13
12
65
4
3
3
11
4
9.5
4
34.5
5
10
14
12
10
11
14
71
6
15
19
15
18
17
8
92
7
19.5
17
2
1
19.5
1
60
8
19.5
8
9
12
19.5
13
81
9
11
4
4
8
8
5
40
10
14
11
18
17
16
6
82
11
17
20
19
15
18
17
106a
12
18
18
13
13
12
20
94
13
2
2
3
3
2
2
14a
14
5
6
5
6
7
9
38
15
8
15
16.5
19
4
18
80.5
16
7
5
6
9
6
7
40
17
1
16
8
14
1
15
55
18
6
7
7
5
3
11
39
19
12
13
20
20
14
19
98
20
16
12
16.5
16
15
16
91.5
Laboratories rejected versus upper and lower criteria of 104 and 22.
The criterion for rejection of individual outliers is based on
calculation of Thompson's T-value [3,6].
In these calculations the mean recovery, X, is given by
n
* = n Xi
n 1 = 1 1
(3)
and the standard deviation, s, is given by
s =
Vi
n
E
i = l
(x-x
19
-------
where = individual analyses
n = number of retained analyses
values in the ampul set
The Thompson's T-test is defined as
Xe"*
rn
where is the retained X^ value farthest away from the mean (X)
of the set of retained data. The data point may be rejected if
the value of T calculated exceeds critical values for T (two-sided
test 5% significance level) as presented in Table 6. If the
extreme value is rejected as an outlier, the test is repeated for
the next most extreme value among the remaining data until the
value being tested passes the test.
Table 7 summarizes calculations to examine suspect data points for
4-chlorophenyl phenyl ether in water 3 by the T-test for outliers.
Four additional data points are identified as outliers. Of the
original 120 data points for 4-chlorophenyl phenyl ether in water 3
(20 laboratories x 6 ampuls), all data points for laboratories 2,
11, and 13 were rejected on the basis of Youden's laboratory rank-
ing procedure (total of 18 points), and four additional data points
were found to be outliers based on Thompson's T-test (for a total
of the 22 data points). These same outlier tests were applied for
all five analytes in the six water matrices. All outlier data
points are marked with an asterisk in Appendix B.
STATISTICAL SUMMARIES
After the outlier rejection tests were performed, the following
summary statistics were calculated employing the remaining data
for each ampul (single analyte, single concentration, single water
matrix):
20
-------
TABLE 6. CRITICAL VALUES FOR THOMPSON'S T (TWO-SIDED TEST) WHEN
STANDARD DEVIATION IS CALCULATED FROM THE SAME SAMPLES
Number of 5%
observations, significance
n level
3
1.15
4
1.48
5
1.71
6
1.89
7
2.02
8
2.13
9
2.21
10
2.29
11
2 .36
12
2.41
13
2 .46
14
2.51
15
2.55
16
2 .58
17
2.62
18
2.65
19
2.68
20
2.71
TABLE 7. RESULTS OF TESTS FOR INDIVIDUAL OUTLIERS
(4-CHLOROPHENYL PHENYL ETHER IN WATER 3)
Extreme
value
Mean
Standard
deviation
Calculated T
n ' n
• , #
Nunber
of point®
Critical
T
Ampul
laboratory
X
e
X
t
•
h
Tc
Decision
1
17
40.05
11.64
8.*7
1.17
IS
2.55
Reject
4
7
en.40
319.03
170.02
2 91
17
2.62
Reject
5
17
34.OS
7.73
1.67
3.43
IS
2.5S
lleject
6
7
213.40
88 S3
46.70
2.67
17
2.62
Reject
21
-------
* Number of retained data points, n
* Mean recovery of retained data, X
* Accuracy as a percent of relative error, % RE
* Overall absolute standard deviation, S
* Percent relative overall standard deviation, % RSD
* Absolute single-analyst standard deviation, SR
* Percent relative single-analyst standard deviation,
% RSD-SA
All of these statistics, except the single-analyst absolute and
relative standard deviations, were calculated using the retained
data for each ampul. The basic statistical formulas used for
these calculations are given below, where , X2,..., Xn denote
the values for the n retained data points for a given ampul.
Mean Recovery (X)
n
X = - Z X (3
n 1 N
1
i=l
Accuracy as a % Relative Error:
% RE = x - true value
/0 true value
Overal1 Standard Deviation:
S =
Tn-l
n
£
i = l
(X- - X
(4
and
Percent Relative Overall Standard Deviation:
% RSD - | x 100 (7
22
-------
The overall standard deviation, S, indicates the precision assoc-
iated with measurements generated by a group of laboratories.
This represents the broad variation in the data collected in a
collaborative study. A measure of how well an individual analyst
can expect to perform in his own laboratory is another important
measure of precision. This single-analyst precision, denoted by
SR, is measured by
The Youden-pair design employed in this study permits the calcu-
lation of single-analyst precision without duplicate measurements
on the same sample and helps to avoid the well-intentioned mani-
pulation of data that can occur when laboratories make duplicate
analyses.
The percent relative standard deviation for the single-analyst
precision is calculated by
where X* is the average of the two mean recoveries corresponding
to the two ampuls defining the particular Youden pair. These sum-
mary statistics are presented in Tables 8 through 12 for each of
the five haloether compounds in the six water matrices.
(8)
where m = number of retained Youden-paired observations
th
= difference between observations in the i pair
D = average of values
% RSD-SA = — x 100
X*
(9)
23
-------
TABLE 8. STATISTICAL SUMMARY FOR BIS(2-CHLOROISOFROPYL)-
ETHER ANALYSIS BY WATER TYPE
WATER 1 WATER 2 WATER 3 WATER 4 WATER 5 WATER 6
LOW YOUDEN PAIR
1 5
1
b
1 b
1 b
1 5
1 5
NUMBER Oh DATA POINTS
13 lb
14
11
16 15
15 14
7 9
16 17
TRUE CONC (C) UG/L
3.00 2.40
3.00
2.40
3.00 2.40
3.00 2.40
3.00 2.40
3.00 2.40
MEAN RECOVERY (X)
4.01 3.83
3.18
2.97
3.10 2.04
4.03 3.87
2.84 4.47
2.64 2.41
ACCURAC Y(XREL ERROR)
33.62 59.72
6.05
23.83
3.37 -14.92
34.31 61.34
-5.19 86.30
-11.96 0.51
OVEKALL STD DEV (S)
2.24 2.IS
1.90
1.43
2.17 0.97
3.65 3.41
1.97 4.53
1.33 1.45
OVEKALL REL STD DEV. X
56.00 55.97
59.71
48.20
69.93 47.28
90.66 88.UU
69.36 101.21
60.39 60.06
SINGLE STD DEV, (SR)
1.83
0.49
l.bl
1.09
1.16
0.93
ANALYST REL DEV, X
46.55
15.94
58. 74
27.63
31.82
36.61
MEDIUM YOUDEN PAIR
2 6
2
6
2 6
2 6
2 6
2 6
NUMBER Of DATA POINTS
18 18
18
17
20 20
19 19
19 18
19 20
TRUE CONC (C) UG/L
132.00 92.00
132.00
92.00
132.00 92.00
132.00 92.00
132.00 92.00
132.00 92.00
MEAN RECOVERY (X)
107.94 82.75
100.77
69.13
91.41 76.11
98.47 69.46
112.69 80.46
103.99 74.94
ACCURACY(XREL ERROR)
-18.23 -10.05
-23-66
-24.86
-30.75 -17.27
-25.40 -24.50
-14.63 -12.54
-21.22 -18.54
OVERALL STD DEV (S)
47.49 26.06
38.13
22.08
37.94 39.43
42.36 26.17
53.69 43.20
50.63 34.50
OVERALL REL STD DEV, X
43.99 31.48
37.84
31.94
41.51 51.80
43.02 37.67
47.64 53.70
48.69 46.04
SINGLE STD DEV, (SR)
18.48
15.19
31.63
17.78
31.21
29.49
ANALYST REL DEV, X
19.39
17.88
37.77
21.18
32.32
32.96
HIGH YOUDEN PAIR
3 4
3
4
3 4
3 4
3 4
3 4
NUMBER OF DATA POINTS
19 18
19
18
20 19
20 2U
20 19
19 19
TRUE CONC (C) UG/L
486.00 624.00
486.00
624.00
486.00 624.00
486.00 624.00
486.00 624.00
486.00 624.00
MEAN RECOVERY (X)
448.24 501.OH
394.51
520.16
381.44 477.27
353.76 476.86
463.92 464.09
380.90 527.38
ACCURACY(XREL ERROR)
-7.77 -19.70
-18.82
-16.64
-21.51 -23.51
-27.21 -23.58
-4.54 -25.63
-21.63 -15.48
OVERALL STD DEV (S)
165.25 165.92
158.27
176.94
181.35 194.58
163.45 190.36
265.40 185.78
144.41 199.61
OVERALL REL STD DEV, X
36.87 33.11
40.12
34.02
47.54 40.77
46.20 39.92
57.21 40.03
37.91 37.85
SINGLE STD DEV, (SR)
102.39
56.35
93.91
110.30
123.28
106.50
ANALYST REL DEV, X
21.57
12.32
21.87
26.56
26.57
23.45
WATER LEGEND
1 - DISTILLED WATER
2 - TAP WATER
3 - SURFACE WATER
4 - WASTE WATER 1
t b - WASH WATER 2
' 6 - WASTE WATER 3
-------
TABLE 9. STATISTICAL SUMMARY FOR Bf S( 2-CI1LOROETHYL) ETHER ANALYSES BY WATER TYPE
WATER 1 WATER 2 WATIR 3 WATER 4 WATER 5 WAH R 6
LOW YO'JUEN PAIR
1
b
1
b
1
S
1
5
1
5
1
b
NUM8ER UK OAIA POINTS
12
IS
12
13
16
14
14
13
13
12
16
15
TRU: CONC (C) UCi/L
1.10
1.60
1.40
1 .t>0
1.40
1.60
1.40
1.00
1.40
1.60
1.40
1.60
MEAN RECOVERY (X)
1.24
2.40
l.M
1.4/
1.4H
1.28
1.24
1.32
H.9R
8.6 7
1.24
1.19
ACCURACY(7RlL ERROR)
-li. n
50.29
lb.12
-8.32
5.80
-20.2/
-11.48
-17.64
541.21
442.08
-11.47
-25.58
OVERALL SI'J UEV (b)
U.b4
2.19
I). 76
0.83
1.18
0.48
0.5 7
0.60
7.84
6.62
0.74
0.39
overall ri l stu oev, t
43.93
90.H8
4b.9b
bb.H9
79.77
37.39
4b.91
4b.21
87.31
76.37
59.53
32.80
S I N(a l Sll) DEV. (SRJ
0.64
0. S3
0.32
0.40
3.60
0.32
ANALYST Rl L DEV. %
35.00
34.S3
22.89
31.34
40. 76
26.74
MEDIUM YUUUEN PAIR
2
6
2
6
2
6
2
6
2
6
2
6
NUMHFR l)f DATA POINTS
17
lb
18
18
20
19
17
16
18
17
18
19
TRUE CONC (C) UG/L
108.UU
87.00
108.00
87.00
108.00
87.00
108.00
87.00
108.00
87.00
108.00
87.00
MEAN RECOVERY (X)
79.40
73.70
76.64
63.83
70.90
5b.52
70.04
5 7.24
90.25
69.03
76.46
58.13
ACCURACY(*RFl ERROR)
-2b.48
-lb.29
-29.04
-26.b3
-34.3b
-36.19
-3b.lb
-34.21
-16.43
-20.66
-29.20
-33.19
OVERALL STU DEV (S)
33. bO
10.96
32.9b
24.b2
32.20
25.99
30.39
20.37
36.21
23.65
31.07
20.89
OVERALL REL STD DEV, %
42.2U
14.87
42.99
38.42
46.41
46.82
4 3.39
35.58
40.12
34.27
40.64
35.93
SINGLE STO OEV, (SR)
15.45
14.19
16.40
13.22
11.73
15.41
ANALYST REL OEV, %
20. iy
20.20
26.02
20.77
14.73
22.90
HIGH YUUUEN PAIR
3
4
3
4
3
4
3
4
3
4
3
4
N'JMBER Of UATA POINTS
18
17
18
18
20
20
18
18
18
17
19
19
TRUE CONC (C) UG/L
602.00
402.00
602.00
402.00
602.00
402.00
602.00
402.00
602.00
402.00
602.00
402.00
MEAN RECOVERY (X)
490.23
321.82
416.22
302.8/
40/.8b
297.82
409.67
307.51
438.52
298.20
426.51
327.25
ACCURALY(XREL ERROR)
-18.bi
-19.94
-JO.Mb
-24.bb
-32.2b
-25.92
-31.95
-23.51
-27.16
-25.82
-29.15
-18.60
OVERALL SID OEV (S)
163.b6
99.74
179.4b
lib.OH
219.2b
Ibl.45
202.21
108.56
1H1.73
104.51
186.83
142.60
OVERALL REL STO UEV, %
33.36
3U.99
43.11
38.00
53.76
50.8b
49.36
35.30
41.44
35.05
43.80
43.57
SINGLE STU 01 V, (SR}
77.42
ML 61
97.90
111).SI
*6.H2
MO.97
ANALYST REL DEV, %
19.07
16.30
27.75
30.82
18.14
24.14
WAIER LEGENO
1 - IJI ST 11.LEO WATER
2 - TAP WAIER
3 - SUKKAU WATER
4 - WASH WAUR 1
'j - WASH WATER 2
6 - to AS 11 WA T E R 3
-------
TABLE 10. STATISTICAL. SUMMARY FOR HI S ( 2-CHI.OROETHOXY ) METHANE BY WATER TYPE
WATER 1 WATER 2 WATTR 3 WATER 4 WATER 5 WATFR f>
low YU'jOEN PAIR
1
b
1
S
1
b
I
b
1
b
I
b
NUMHLR OF DATA POINTS
10
11)
U
13
13
14
16
12
11
12
15
14
TRIJF. CONC (C) UG/L
1.40
l.uo
1.40
1.00
1.40
1.00
1.40
1.00
1.40
1.00
1 .40
1.00
Ml AN RKUVERY (X)
l.US
0.87
1.26
1.54
1.09
1.69
2.11
0.99
3.25
3.07
1.7/
1.71
AtCl RAL Y(-tRF.L ERROR)
-24./I
-12.80
-9.68
b4.U0
-22.47
b9.3b
74.2U
-U.92
132.27
206.83
26.1U
71.07
OVERALL Si!) DEV (S)
0.b4
0.33
0. /6
1.89
0.63
2.22
3.29
U. 76
3.29
3.06
1.02
1.63
UVi MAI 1. Rl t. STU DEV, X
bl.3b
cc
r^,
60.38
122.87
b/.Hb
139.48
134.79
76.63
101.21
99.*>6
b 7.59
9b. 04
SINIil E STO DEV, (SR)
0.35
0.61
0.31
0.83
2.07
0.64
analyst KEL DEV. 1
3b,26
36.07
23.11
4K.36
66.65
36.67
MEDIUM YUUUEN PAIR
2
6
2
6
2
6
2
6
2
6
2
6
NlJMtS! R U* DA IA POINTS
lb
1/
18
IB
19
19
17
16
18
17
18
19
true conc (o ug/l
106.00
126.00
106.00
126.UO
106.00
126.00
106.00
126.00
106.00
126.00
106.00
126.00
MEAN RECOVERY (X)
7b. 40
88.96
70.23
90.00
61.38
79.80
69.2b
77.9/
74.98
97.b 6
71.7b
8b.63
ACCURACY(%RtL ERROR)
-27.92
-29.40
-33.75
-28.b7
-42.09
-36.67
-34.67
-38.12
-29.2 7
-22.57
-32.31
-32.04
overall STD DEV (S)
21.93
24.91
2b.89
28.8b
26.22
42.94
1 /. 79
29.b7
24.64
2b.73
24.30
27.30
OVERALL REL STD DEV, %
28.71
28.UO
36.87
32.0b
42.71
63.81
2b.69
37.92
32.86
26.36
33.86
31.89
SINGLE STD DEV, (SR)
17.74
23.21
25.90
16.64
19.03
19.42
anai yst rfl nrv, %
21.45
28.97
36. 70
22.61
22.06
24.68
HIGH YOUDEN PAIR
3
4
3
4
3
4
3
4
3
4
3
4
NUMBER OF DATA POINTS
17
16
18
17
19
19
18
18
19
19
19
19
TRUE CONC (C) UG/L
398.00
528.UO
398.UO
528.UO
398.00
528.00
398.00
528.00
398.00
528.00
398.00
528.00
MEAN RECOVFRY (X)
294.18
360.72
270.47
374.52
254.29
359.95
253.63
378.58
268.32
391.16
251.82
391.97
ACCURACY(%RI I IRROR)
-26.08
-31.68
-32.04
-29.07
-36.11
-31.83
-35.30
-28.30
-32.58
-2b.86
-36.73
-2b.7b
OVERALL STU DEV (S)
9b.bl
143.96
114.b6
134.69
127.81
174.93
120.98
158.43
99.92
171.44
104.77
161.40
OVERALL Rt 1. STD DI V, %
32.47
39.91
4?. 3b
3b. 96
bU.26
48.60
4 7.72
11.8b
3 7.24
43.80
41.61
41.18
SINGlE STD DEV, (SR)
62.92
46.10
bb.75
78.91
81.02
CD
X
T
ANALYST REL DEV, I
19.21
14.29
21.41
24.97
24.bb
27.7b
WATER LhUENO
1 - DISTILLED WATER
2 - TAP WATER
3 - SlJ'UACl WAHR
4 - WASTE WAIER 1
b - WASH WAILR 2
I; - WA'Wl WAHR }.
-------
TABLE 11. STATISTICAL SUMMARY FOR 4-CHLOROPHENYL PHENYL ETHER ANALYSES RY WATER TYPE
WATER I WATER 2 WA1ER 3 WAIER 4 WATER b WAIER 6
LOW YUUOEN PAN
1
b
1 b
1
b
1
b
1 b
1
5
NUMBER OF DATA POINTS
12
13
14 14
14
14
16
14
W 15
17
13
TRUi r.ONC (C) UG/l.
14.50
6. GO
14.bO 6.6U
14.bO
6.60
14.bO
6.60
14.bO 6.60
14.bO
6.60
MIAN RECOVERY (X)
I2.y l
/.bJ
10.81 S./4
9.61
b.85
8.99
5.47
21.14 2b.56
9.59
6.46
Alt JRACY( IRE L ERROR)
-10.9b
14.1b
-25.46 -13.06
-33.73
-11.43
-38.00
-17.11
4b.76 287.20
-33.85
-2.12
OVERAll STD DEV (5)
6.96
3. 31
b.37 2.99
4. bO
2.49
b .2/
2.46
20.23 32.61
6.00
2.85
OVERALL REL STD OEV, t
b3.89
43.89
49.69 52.19
46.85
42.bb
58.57
44.92
95.70 127.60
62.57
44.17
SINGLE STD OEV, {SR)
3.94
2.60
2.52
2.58
19.67
3.08
ANALYST KEL DEV, T
38.53
31.47
32. b6
35.63
84.24
38.37
MEDIUM YUUiJtN PAIR
2
6
2 6
2
6
2
6
2 6
2
6
N'JMBER 0^ DATA POINTS
lb
17
16 17
17
16
18
17
19 17
16
17
TRUfc CONC (C) UG/L
94.00
120.00
94.00 120.00
94.00
120.00
94.00
120.00
94.00 120.00
94.00
120.00
Ml AN REMOVE RY (X)
77.55
105.75
69.83 94.42
66.37
81.0b
70.79
79.61
79.21 84.58
69.42
80.44
ACCURACY(iREL ERROR)
-17.bO
-11.87
-25.71 -21.32
-29.39
-32.46
-24.69
-33.66
-lb.73 -29.b2
-26.15
-32.97
OVERALL STD OEV (S)
3b. 33
30.2b
19.33 31.43
20.9b
3b.03
29. S6
29.63
29.93 33.0b
22.36
30.98
OVERALL REL STD DEV, %
46.84
28.61
27.67 33.29
31.56
43.23
41.76
37.22
37.78 39.07
32.20
38.51
SINGLE STD OEV, (SR)
21.01
20.18
18.90
22.96
23.67
26.1b
ANALYST REL OEV, %
22.93
24.57
2b.64
30.53
28.91
34.90
HIGH YOMUEN PAIR
3
4
3 4
3
4
3
4
3 4
3
4
NiJMKER Of DATA POINTS
17
16
17 17
17
16
18
18
19 19
17
17
TRUE CUNC (I) OG/L
489.00
424.00
489.00 424.00
489.00
424.UO
489.00
424.00
489.00 424.00
489.00
424.00
MEAN RECOVERY (X)
412.77
337.79
369.14 324.91
34 7.88
288.13
288.86
278.96
3b4.17 304.10
324.18
340.48
AC CUR AC Y {%RE I. ERROR)
-15.59
-20.33
-24.bl -23.3/
-2H.86
-32.04
-40.93
-34.21
-27.5/ -28.28
-33.71
-19.70
OVERALL MO OEV (S)
160.38
161.92
169.19 161.97
11> 1.08
134.70
131.02
128.89
165.0b 142.01
119.35
145.88
OVIRAEL REL STD OEV, %
38.85
47.94
45.83 49.8S
43.43
46.75
45.36
46.20
46.60 46.70
36.82
42.84
SINGLE STD DEV, (SR)
60.08
42.2b
63.61
60.64
77.01
75.37
ANALYST REL OEV, 1
16.01
12.1/
20.00
21.36
23.40
22.68
WATER LEGEND
1 - UI b> TILLEO WATEK
2 - TAP WAIf R
3 - SlNlALf WATTR
4 - WAS f f UAKR 1
b - W;v>TE rfAIEK d
h - UA'.Ti: WAM.K S
-------
1ABLE 12. STATISTICAL SUMMARY FOR 4-BROMOPHENYL PHENYL ETHER ANALYSES BY WATER TYPE
WAT t" W 1 WATER 2 WATER 3 WATER 4 WATER 6 WATER 6
N>
CO
LOW YOUOLN PAIR
1
6
1
5
1 6
1
5
1 6
1
6
NUMBER Of OATA POINTS
13
13
13
13
1/ 17
1/
lb
15 13
17
15
HUE CONC (I) UU/L
2.80
3.80
2.80
3.80
2.80 3.80
2.80
3.80
2.80 3.80
2.80
3.80
HI AN HfCUVI RY (X)
4.62
6.37
4.14
6.06
4.31 5.02
4.34
5.00
3.52 7.32
4.01
4.44
ACCURACY( ;RtL ERROR)
64.92
67.56
47.7/
32./ 9
53.84 32.16
54.94
31.63
26.76 92.69
43.13
16.95
UVl RAl L SID 01.V (S)
2.69
3.06
2.36
3.08
2.88 2.52
2.66
3.10
2.01 5.19
2.22
2.18
UVtKALl KLL STO [)EV, I
68.21
48.10
5/.06
61.04
6b.78 50.20
61.08
62.10
57.14 70.88
55.41
49.15
SINGlt SID DEV. (SR)
1.
,66
1.
31
1.83
1.
71
2.83
1,
,42
ANALYST RIL lU'V, %
28.
.45
28.
63
39.14
36.
69
52.21
33,
.67
Ml OllIM YOUUEN ^AIR
2
6
2
6
2 6
2
6
2 6
2
6
NUM8ER OF UATA POI NTS
18
19
19
19
20 20
2U
18
19 16
18
20
I RUE CONC (I) Oti/L
146.00
116.00
145.00
116.00
146.00 116.00
146.00
116.00
146.00 116.00
145.00
116.00
MEAN RECOVERY (X)
127.2/
100.21
120.92
92.68
11/.09 93.80
128.84
92.73
139.29 88.26
119.29
96.82
ACCURACY ("REL ERROR)
-12.23
-13.61
-16.61
-20.19
-19.25 -14.14
-11.15
-20.06
-3.94 -23.91
-17.73
-16.63
OVIRAIL STO OEV (S)
68.12
47.27
58.63
46.06
66.47 48.56
64.10
44.37
72.24 36.82
56.46
50.02
OVtRAl.L KLL SID 1)1. V, t
63.52
47.17
48.49
48.6 7
4 7 .38 61 . 77
49.76
47.85
51.87 40.69
47.33
51.66
SINGLE STO UEV, (SR)
40,
.52
36.
30
37.03
43.
,80
41.71
41,
.08
ANALYST REL OEV, X
35.
.63
34.
,01
35.12
39.
.63
36.66
38
.01
HliiH YOUDt'N PA Ik 3 4 3 4 3 4 3 4 3 4 3 4
NUMBER UAIA KHN1S 14 1H 14 18 ly 2U 20 2U 19 18 18 20
TKUE CONt (C) UG/L 652.00 626.UO 562.00 626.00 562.00 626.00 562.00 626.00 552.UO 626.UO 552.00 626.00
rll AN Kt.COVL^Y (X) 487.29 493.91 465.45 532.26 426.42 489.42 408.19 426. 32 436 .61 400. 71 365.32 553.24
ACTURACYUREL ERROR) -11.72 -21.10 -17.49 -14.97 -22.93 -21.82 -26.05 -31.90 -20.90 -35.99 -33.82 -11.62
OVtRAI I. STO orv (S) 206.47 233. 38 213. 1 1 245.99 194.?f> 261.9/ 189.9K 213.89 263.HI 1/3.96 162.07 256.46
OVtRAl.L REL SIL) OEV, * 42.3/ 4/.26 46. 79 46.22 45.66 53.53 46.64 60.17 60.42 43.41 44.36 46.36
6 I N'jlE
ANALYST
STn ofv, (s«)
REL OEV. i
74.00
16.08
51.96
10.62
88.72
19.40
86.86
20.82
98.49
23.52
110.61
24.H6
WML* LEGEND
1 - DIM ill.LI) WAUR
2 - IAP WA11R
3 - SUKI ACt WM( R
4 - WAS It VJrt 11 R 1
6 - WAS It WA11 R 2
0 - W\s;t WAI1R J
-------
REGRESSION ANALYSIS OF BASIC STATISTICS
Systematic relationships frequently exist between the mean re-
covery statistics and the true concentration levels across ampuls,
and between the precision statistics and the mean recovery statis-
tics. Given a plot of precision values versus concentration
levels, a smooth curve drawn through the points can show that the
precision is found to (1) be constant and not vary with level;
(2) vary directly with level in a linear manner; or (3) vary with
level in a curvilinear fashion.
In order to derive statements for method accuracy and precision,
the basic statistics were regressed assuming linear relationships,
fitting the data to a line using weighted least-squares. The
weights were chosen to be inversely related to the true concentra-
tion in the case of accuracy and inversely related to the mean
recovery in the case of precision. The inverse weightings were
employed to moderate the influence of the high Youden-pair data.
The results of the regression analyses are discussed below.
Statements of Method Accuracy
The accuracy of Method 611 is characterized by comparing the mean
recovery of the analyLe, X, to the true concentration level of the
compound, C, in the water sample. In order to obtain a mathemati-
cal expression for this relationship, a regression line of the form
X = aC + b (10 )
was fitted to the data by regression techniques.
The true concentration values often vary over a wide range. In
such cases, the mean recovery statistics associated with the larg-
er concentration values tend to dominate the fitted regression line
29
-------
producing relatively larger errors in the estimates of mean recov-
ery at the lower concentration values. In order to eliminate this
problem, a weighted least squares technique was used to fit the
mean recovery data to the true concentration values. The weighted
least squares technique was performed by dividing both sides of
Equation (10) by C resulting in Equation (11)
(=)
£ = a + b (±| (11)
which can then be converted to the desired relationships by
multiplying through by C, giving:
X = aC + b (12)
These equations were presented earlier in Table 1.
If the intercept "b" associated with the fitted line is negligible
(i.e., essentially zero), then the slope "a" provides a unique
value which represents the percent recovery over all of the concen-
tration levels.
Statements of Method Precision
The precision of Method 611 is characterized by comparing the
overall and single-analyst standard deviations to the mean re-
covery, X. The IMVS program conducts these calculations via
matrix algebra, where a weighted least-squares linear regression
of S and SR versus X is conducted with weights chosen to be in-
versely proportional to the square of the mean recovery (see
page 108 of Reference 2 for details). This method is equivalent
to that suggested by Britton [7] where the linear regressions for
30
-------
S and SR versus C are achieved by using the customary least-
squares procedure to fit the equation:
§ = C + d i (13)
In this study, however, the regression was conducted versus X as
follows:
| = c + d ~ (14)
X X
which is then converted by multiplying through by X to yield the
linear relationships
S = aX + b (15)-
and
SR = cX + d (16)
These equations also were presented earlier in Tables 1 and 2.
If the intercepts, b and d, are negligible, then the slopes, a
and c, are good approximations of the overall and single-analyst
percent relative standard deviations, respectively. These, in
turn, are measures of the method precision.
COMPARISON OF ACCURACY AND PRECISION ACROSS WATER TYPES
It is possible that the accuracy and precision of Method 611
depend on the water type analyzed. The summary statistics X, S,
and SR are calculated separately for each concentration level
within each water type. They can be compared across water types
in order to obtain information about the effects of water type on
31
-------
accuracy and precision. However, the use of these summary statis-
tics in this manner has several disadvantages. First, it is
cumbersome because there are 36 mean recovery statistics (X) (six
ampuls x six waters), 36 overall precision statistics (S), and
18 single-analyst precision statistics (SR) calculated for each
compound. Comparison of these statistics across concentration
levels and across water types becomes unwieldy. Second, the
statistical properties of this type of comparison procedure are
difficult to determine. Finally, due to variation associated with
X, S, and SR, comparisons based on these statistics can lead to
inconsistent conclusions about the effect of water type. For
example, distilled water may appear to produce a significantly
lower value than drinking water for the precision statistic S at
a high concentration, but a significantly higher value for S at
a low concentration.
An alternative approach [2], has been developed to test for the
effects of water type. This alternative approach is based on the
concept of summarizing the average effect of water type across
concentration levels rather than studying the local effects at
each concentration level. If significant differences are estab-
lished by this alternative technique, then the summary statistics
can be used for further local analysis.
The test for the effect of water type is calculated using the
following statistical model. If X-denotes the measurement
1J K
reported by laboratory "i," for water type "j," and ampul "k,"
then
X- - p • • C, J • L- • t •, (1
ijk Kj k l ljk x
where i = 1,2,..., n
j - 1,2, 6
k = 1,2 6
32
-------
Model components pj and Vj are fixed parameters that determine
the effect of water type j on the behavior of the observed
measurements — The parameter is the true concentration
level associated with ampul "k." The model component is a
random factor which accounts for the systematic error associated
with laboratory "i." The model component r, ¦ ¦, is the random factor
1 j K
that accounts for the intralaboratory error.
The model is designed to approximate the global behavior of the
data. The multiplicative structure was chosen because of two
important properties. First, it allows for a possible curvilinear
relationship between the data (X- •, ) and the true concentration
1J K
level (C^) through the use of the exponent Vj on C^. This makes
the model more flexible in comparison to straight-line models.
Second, as will be noted below, an inherent increasing relation-
ship exists between the variability in the data and the concen-
tration level in this model. This property is important
because it is typical of interlaboratory data collected under
conditions where the true concentration levels vary widely.
Accuracy is related directly to the mean recovery or expected
value of the measurements (X^
data modeled by Equation 1 is
value of the measurements (X. ¦. ). The expected value for the
1 j K
E = • ckYi ¦ E
Precision is related to the variability in the measurements (X--,
1 J K.
The variance of the data modeled by Equation 1 is
Var(Xljk) =
Pj ' CkVj
2
Var(Li • jk), (18)
which is an increasing function of C^. (See Reference 2 for a
complete discussion of this model.)
33
-------
The accuracy and precision of Method 611 depend upon water type
through Equations 17 and 18 and the parameters p^ and Yj- If Pj
and Yj vary with j (i.e., vary across water type), then the
accuracy and precision of the method also vary across water type.
To determine if these parameters do vary across water type and to
compare their values, they must be estimated from the laboratory
data using regression techniques. Equation 1 represents the basic
model. However, taking natural logarithms of both sides of Equa-
tion 1, the following straight line regression model is obtained.
• 2n Xijk = £n + vj 2n Ck + 2n Li + £n cijk (2)
The parameter £n pj is the intercept, and Yj is "the slope of the
regression line associated with water type "j." It is assumed that
£n L- is normally distributed with mean 0 and variance a2, that
•1 Li
&n c • is normally distributed with mean 0 and variance o 2, and
1 J £
that the £n L- and £n z>¦, terms are independent,
l i j K.
Based on Equation 2, the comparison of water types reduces to the
comparison of straight lines. Distilled water is viewed as a
control, and each of the remaining lines is compared directly to
the line for distilled water.
Using the data on the log-log scale and regression techniques, the
parameter £n p_j (and hence p^) and Yj can be estimated. These
estimates are then used to test the null hypothesis that there is
no effect due to water type. The formal null and alternative
hypothesis, HQ and H^, respectively are given by:
H - £n p. - £n p, =0 and Y- * V, - 0 for j -2 (19)
0 f j r 1 lj'l J
H ¦ £n p. - £n p, f 0 and/or y. " Yi t 0 f°r some j = 2 (20)
A J 1 J 1
34
-------
The null hypothesis (HQ) is tested against the alternative hypoth-
esis (H^) using an F-statistic. The probability of obtaining the
value of an F-statistic as large as the value which was actually
observed, Prob(F > F OBS), is calculated under the assumption that
Hq is true. HQ is rejected in favor of if Prob(F > F OBS) is
less than 0.05.
If Hq is not rejected, then there is no evidence in the data that
the Pj vary with "j" or that the Vj vary with "j." Therefore,
there is no evidence of an effect due to water type on the accuracy
or precision of the method. If H^ is rejected, then some linear
combination of the differences (£n p^ - £n p ) and - y^) is
statistically different from zero. However, this does not
guarantee there will be a statistically significant direct effect
attributable to any specific water type since the overall F test
can be overly sensitive to minor systematic effects common to
several water types. The effect due to water type is judged to
be statistically significant only if one of the differences,
(£n Pj - £n p^) and/or (y^ - y^), is statistically different
from zero. This is determined by checking the simultaneous 95%
confidence intervals which are constructed for each of these dif-
ferences. Each true difference can be stated to lie within its
respective confidence interval with 95% confidence. If zero is
contained within the confidence interval, then there is no
evidence that the corresponding difference is significantly dif-
ferent from zero.
If at least one of the confidence intervals for the differences
(Zn p^ - £n p^) or (y^ - y-^) fails to include zero, then the stat-
istical significance of the effect due to water type has been
established. However, establishment of a statistically signifi-
cant effect due to water type does not necessarily mean that the
effect is of practical importance. Practical importance is
related to the size and interpretation of the differences.
35
-------
The interpretation of the differences involves comparing the mean
recovery and standard deviation for each water type to the mean
recovery and standard deviation obtained for distilled water.
These comparisons are made on a relative basis. The mean recovery
for water type "j," given by Equation 17, is compared to that for
distilled water (j = 1) on a relative basis by
E = V'j " Vl)
E(Xiik)
Ck
E(L. " enk
(21)
(The ratio of the standard deviations would be equivalent to
Equation 21; therefore, the interpretation of the effect on
precision is the same as that for the effect on mean recovery.)
The ratio in Equation 21 is a measure of the relative difference
in mean recovery between water type "j" and distilled water. It is
composed of two parts (a) p/p,, which is independent of the true
^ ( v • ~V )
concentration level (i.e., the constant bias), and (b) "1 /
which depends on the true concentration level (i.e., the concen-
tration dependent bias). If (v j - Yj_) is zero, then the relative
difference in mean recovery is P^/P^' which is independent of con-
centration level C^. Then the mean recovery of water type "j" is
x 100 percent of the mean recovery for distilled water. If
(y- - y,) is not zero, then the mean recovery of water type "j"
J ! y - Y 1
is ([p /Pil'C, J ^ ) x 100% of that for distilled water, and
J 1 K
therefore depends on the true concentration level .
To illustrate these points, consider the following example. Sup-
pose that a significant F-value has been obtained, and the confid-
ence intervals for all of the differences contain zero except for
water type 5. For water type 5, the point estimate for (J2n -
j2n p^) is -0.38, and the confidence interval for (J2n p^ - J2n p^) is
(-0.69, -0.07). The point estimate for (y^ - y^) is -0.07, and the
36
-------
confidence interval for - y^) is (-0.04, 0.18). In this case,
a statistically significant effect due to water type has been es-
tablished that involves only water type 5. The practical signific-
ance of this effect is judged by considering Equation 21. The
ratio of mean recoveries for water type 5 and distilled water is
given by
E(Xi5k) _ P5 (y5 ~ Yl)
E(Kilk) " Pik (
and the ratio of the standard deviations is given by
/Var(Xllk) E, (t5 - Vl,
Vvar(Xlli() <23>
Because the confidence interval for contains z^ro, this
difference is assumed to be insignificant and is set to zero.
Therefore, Equations 22 and 23 reduce to P^/P^- The point estimate
for (£n p^ - £n p^) was -0.38. Therefore, the point estimate for
P5/Pl is 0.68, and the mean recovery for water type 5 is estimated
to be 68% of the mean recovery for distilled water. Similarly,
the standard deviation for the data for water type 5 is estimated,
to be 68% of the standard deviation for distilled water. Since
the 95% confidence interval for (£:i p^ ~ Pj) was (~0.69, -0.07),
any value in the interval (0.50, 0.93) is a reasonable estimate for
P^/Pl* and the mean recovery (standard deviation) for water type 5
can be claimed to be from 50% to 93% of the mean recovery (standard
deviation) for distilled water. The practical significance of the
effect due to water type 5 would depend on the importance of a mean
recovery (standard deviation) that is between 50% and 93% of the
mean recovery (standard deviation) observed for distilled water.
37
-------
The comparison of accuracy and precision across water types just
discussed, is based on the assumption that Equation (1) approxi-
mately models the data. It is clear that in practical monitoring
programs of this type, such models cannot model the data complete-
ly in every case. This analysis, therefore, is viewed as a
screening procedure which identifies those cases where differences
in water types are likely to be present. A more detailed, local
analysis can then be pursued using the basic summary statistics
for precision and accuracy.
Results of the accuracy and precision comparison among the waters
in the study are presented in Appendix C.
38
-------
SECTION 6
RESULTS ANL DISCUSSION
The objective of this study was to characterize the performance of
Method 611 in terms of accuracy, overall precision, single-analyst
precision, and the effect of water type on accuracy and precision.
One measure of the performance of the method is that 16.3% of the
3600 analytical values were rejected as outliers. Of the 16.3%
outliers, 6.1% were re]ected through application of Youden's lab-
oratory ranking procedure and 10.2% were rejected employing the
Thompson T-test.
ACCURACY
The accuracy of Method 611 is obtained by comparing the mean
recovery, x, to the true values of concentration in yg/L. In
Tables 8 through 12, individual values of accuracy as percent
relative error are listed for each analyte, in each water matrix,
and at each of the six concentration levels in that water matrix
(three Youden pairs). This results in 180 separate values for
accuracy. The linear regression of mean recovery, x, versus true
concentration level, c, provides values representing the percent
recovery over all of the concentration levels. This reduces the
separate values for accuracy to 30, one value for each of five
analytes in each of six waters. Table 13 presents the percent
recovery for each compound in water types as measured by the
slopes of the linear equations for recovery presented earlier
in Table 1. In Table 13, the linear regression slopes are com-
pared to percent recoveries calculated from the average of the
39
-------
TABLE 13. METHOD 611 ACCURACY (%)
BCIPE
BCEE
. b
Water
Water 1
Water 2
Water 3
Wdtei 4
Water 5
Water 6
Average
all waters
Slopi
78
77
73
83
80
79
Moan Recovery by Youden Pair
Low Medium High
. b
147
115
94
148
141
94
123
86
76
79
75
86
80
80
Slope'
86
82
78
75
85
81
81
81
72
67
69
72
72
72
Mean Recovery by Youden Pair
Low Medium High
119
103
93
8b
592
82
179
79
72
65
65
81
69
72
(continued)
81
72
71
12
74
76
74
-------
TABLE 13 (continued)
BCEXM CPPE
Mean Recovery by Youden Pair^ Mean Recovery by Youden Pair*3
_ Walter Slope I.ow Medium High Slopea Low Med i um High
Water 1 71 81 71 71 82 101 85 82
Water 2 67 122 09 69 75 81 72 76
Water 3 60 118 62 66 67 77 69 70
Water 4 65 96 64 68 65 72 71 62
Water 5 71 270 74 71 56 267 77 72
Water 6 67 149 68 69 69 92 70 73
Average 67 139 68 69 69 115 74 73
all waters
(continued)
-------
TABLE 13 (continued)
BPPE Average All Anaiytes
Water
Slope2
Mean Recovery by Youden
Low Medium
Pair*5
Hiqh
Slope3
Mean Recovery by Youden Pair'
Low Fledium High
Water
1
85
166
87
84
81
123
82
81
Water
2
72
140
82
84
73
112
74
77
Water
3
78
141
81
78
70
105
71
73
Water
4
n i
/ i
143
34
76
70
109
72
71
Water
5
81
159
86
74
73
280
80
73
Water
6
79
130
83
77
73
109
74
75
Average
79
147
84
79
73
140
76
75
all waters
3
Percent accuracy from slope of regression equations (Table 1).
^Mean percent recovery for Youden pairs (Tables 8-12).
-------
quotients x:c (presented in Tables 8 through 12) for all three
Youden pairs individually.
The validity of using the slope to estimate the percent recovery
depends up on the negligible magnitude for the intercept of the
linear regression equation. From Table 13, it is evident that the
linear regression slope agrees extremely well with the average re-
coveries from the medium and high Youden pair concentrations.
Examination of Table 13 reveals that recoveries of the analytes in
the low Youden pair samples often exceeded 100%. This could be
attributed to difficulty in correcting for background interferences
in the blank analyses. This cause is suggested for the high recov-
eries of BCEE and 4-CPPE in wastewater 2 (eg 582% and 267%, respec-
tively). In these cases the intercepts of the regressions equa-
tions for accuracy in wastewater 2 were 7.77 for BCEE and 20.40
for 4-CPPE. Both values cannot be viewed as being insignificant.
Because of these large intercepts, new linear regression equations
were calculated for these two analytes in wastewater 2, as pre-
sented earlier in Table 2.
Therefore, it is evident that the driving force in determining the
veracity of the use of the regression equation slope as the percent
recovery throughout the concentration range studied is the low
Youden pair values. Based upon the excellent agreement between
the linear regression equation slopes and the average recoveries
for the higher Youden pair samples, the values presented in Table
13 and earlier in Tables 1 and 2 are considered to be representa-
tive of the accuracy of Method 611.
PRECISION
The overall and single-analyst precisions of Method 611 were
determined as percent relative standard deviations for each
analyte, water type, and concentration level. As presented in
Tables 8 through 12, 180 individual values of overall percent
43
-------
relative standard deviation and 90 individual values of single-
analyst percent relative standard deviation result. The linear
regression of standard deviation, s, versus mean recovery, x, pro-
vides values of percent relative standard deviation over all the
concentration ranges. This reduces the separate measures of pre-
cision to 30, one value for each of five analytes in each of six
water-types. Tables 14 and lb present the percent relative stand-
ard deviations as measured by the slopes of the linear regression
equations presented earlier in Table 1 for the overall and the
single-analyst precision, respectively. These values are compared
to the averages of the percent relative standard deviations pre-
sented in Tables 8 through 12 for all three individual Youden pairs.
From Tables 14 and 15, it is evident that the % RSD and % RSD-SA
precision values obtained from the slopes of the linear regression
equations presented in Table 1 agree very closely with the average
precision values for the middle and high Youden pairs presented in
Tables 8 through 12. This agreement offers support to the preci-
sion value obtained via the linear regression process.
The poor precision (high % RSD and % RSD-SA) values demonstrated
in the low Youden pair samples in Tables 8 through 12 are attribu-
ted to background interferences in the water matrices for the low
ccncentration range of the haloetherc (1.0 to 3.8 pg/L).
Two questionable precision values are evident on examination of
Tables 14 and IS. The first is the value of 32% RSD for chloro-
phenyl phenyl ether in water 5. In this case, the average % RSD
from the middle and high Youden pe,ir is 43%. The second question-
able value also occurs for chlorophenyl phenyl ether in water 5
where the linear regression slope gives a value of 15% RSD-SA com-
pared to an average of 26% for the middle and high Youden pairs
and an average of 26% for the middle and high Youden pairs and an
average of 46% for all Youden pairs. These unusual values can be
attributed t.o interences in the wastewater 2 matrix.
44
-------
TABLE 14. METHOD 611 PRECISION (% RSD)
BCIPE BCEE
Water
Slope3
Mean %
Low
RSD by Youden
Pai r^
Slopea
Mean %
Low
RSD by Youden
_ . b
Pair
Medium
Hiqh
Medi urn
Hiqh
Water
1
36
56
38
35
35
67
29
32
Water
2
36
54
35
37
40
52
41
41
Water
3
47
59
47
44
50
59
46
52
Water
4
40
89
40
43
41
46
39
42
Wa ter
5
52
85
51
49
35
82
37
38
Water
6
42
55
47
38
41
46
38
44
Average
42
66
43
41
40
59
38
42
all waters
(continued)
-------
TABLE 14 (continued)
BCEXM
CPPE
Water
Water 1
Water 2
Water 3
W a t e r 4
Water 5
Water 6
Average
all waters
Mean % RSI) by Youden Pair
Low Medium High
S lope'
33
38
53
33
34
36
39
Slope'
Mean % RSD by Youden Pair^*
Low Med i um H i cjh
45
92
99
106
100
76
86
28
34
48
30
33
34
36
39
49
43
41
41
42
41
39
42
* ">
H J
32
38
39
59
51
45
52
112
53
62
38
30
37
39
38
35
36
(continued)
43
48
45
4G
47
40
45
-------
TABLE 14 (continued)
BPPE Average All Analytes
Water
Slopea
Mean %
Low
RSD by Youden
Medium
Pairb
Hiqh
Slope3
Mean
Low
% RSD by Youden Pair*3
Medium High
Water
1
47
53
50
45
38
56
37
38
Water
2
47
59
49
47
40
62
38
42
Water
3
49
58
50
50
48
64
46
48
Water
4
48
62
49
48
42
71
40
45
Water
5
51
64
46
52
41
89
40
45
Water
6
47
52
49
45
41
56
40
42
Average
48
58
49
48
42
66
40
44
all waters
a% RSD from slope of regression equations (Table 1).
^Mean % RSD for individual Youden pairs (Tables 8-12).
-------
TABLE
15, METHOD 611
PRECISION
(% RSD-SA)
BC1PE
BCRE
Water
Slope3
Mean %
Low
RSD-SA by Youden
Medium
Pair'3
H i qh
Slope8
Mean %
Low
t
, RSD-SA by Youden Pair
Medium Hiqh
Water
i
20
47
19
22
19
35
20
19
Water
2
15
36
18
12
18
35
20
16
Water
3
29
59
38
22
27
23
26
28
Water
4
24
28
9 1
i . A.
27
26
31
21
31
Water
5
29
32
32
27
15
41
15
18
Water
6
28
37
33
23
23
27
23
24
Average
all waters
24
37
27
22
21
32
21
(continued)
23
-------
TABLE 15 (continued)
BCEXM
CPPE
Water
Mean % RSD-SA by Youden Pair
Slope Low Medium High
Slope'
Mean % RSD-SA by Youden Pair
Low Medium High
Water 1
Water 2
Water 3
Water 4
Water 5
Water 6
Average
all waters
20
21
29
24
22
26
24
36
36
23
48
66
37
39
21
29
37
23
22
25
26
19
14
21
25
25
28
22
18
17
22
25
15
28
21
39
31
33
36
04
38
44
23
25
26
31
29
35
28
(continued)
16
12
20
21
23
23
19
-------
TABLE IS (continued)
KITE
all waters
. b
Water
Slope3
Low
Medium
High
Water
1
25
28
36
15
Water
2
22
29
34
11
Water
3
27
39
35
19
Water
4
30
3/
40
21
Water
5
29
52
37
24
Water
6
31
34
38
24
Average
27
37
37
19
a% RSD-SA from slope of regression equations (Table 1).
^Mean % RSD-SA for individual Youden pairs (Tables 8-12).
Average All Analy tes
Mean % RSD-SA by Youden Pair^
Slope3 Low Medium High
29
37
24
18
19
29
25
13
27
35
32
22
26
36
27
22
22
55
27
23
27
35
31
24
24
38
28
20
-------
EFFECTS OF WATER TYPES
The comparison of accuracy and precision across water types is
summarized in Table 16, where the observed F values and the prob-
ability of exceeding the F values are entered for each of the
seven analytes.
For every analyte except 4-bromophenyl phenyl ether, the F-test
suggests a statistically significant effect due to water type
(P[F>observed F]<0.05). The null hypothesis test indicates that
a statistically significant effect has been established at the 95%
confidence limit for the following analyte - water combinations:
bis(2-chloroisopropyl)ether in waters 3 and 6; £is(2-chloroethoxy)
methane in waters 5 and 6; and 4-chlorophenyl phenyl ether in water
5. These effects are indicated since zero is not contained within
the confidence limits for (2npj - and/or (\j - y^) for the
above analyte-water combinations.
After examination of several factors including final regression
equations for all waters and the absolute values of the point es-
timates, the only instance in which a practical significance is
evident is for 4-chlorophenyl phenyl ether in wastewater 2. This
analyte-water combination coincides with that which exhibited the
lowest accuracy (Tables 1 and 13) and the largest discrepancies in
precision (% RSD and % RSD-SA) between the linear regression equa-
tion slopes and the averages of the precision values (see Tables
14 and 15).
RESPONSES TO QUESTIONNAIRE
One of the goals of this study was to conduct the interlaboratory
study of the haloether method in a manner consistent with how it-
would eventually be used. A number of decisions were made both
prior to the study and after the prestudy conference concerning
51
-------
TABLE 16. SUMMARY OF THE TEST FOR DIFFERENCE ACROSS WATER TYPES
~ Statistical Practical
F test significance significance
statietically established established
significant by the 95% by the 95%
Compound
Observed
F-value
PlF>observed F|
at the
5% level?
confidence
limit?
Waters
confidence
limit? Waters
6is(2»Chloroisopropyl)ether
2.73
0.0027
Yes
Yes
3.6
No
6is(2-Chloroethyl)ether
20.42
0.0000
Yes
Yes
5
No
&ift(2*Chloroethoxy)methane
CD
O
4
0.0000
Yes
Yes
5,6
No
4-Chlorophenyl phenyl ether
4.87
0.0000
Yes
Yes
5
Yes 5
4-Bro*ophenyl phenyl ether
0.32
0.975
No
•
-
-
-------
which variables would be controlled and which variables would be
allowed to contributate to a wider distribution of results. The
method requires the use of halide-specific detectors. The use of
Hall, Coulson, and Dohrmann detectors was recommended. The EC
detector used by one laboratory was not recommended due to a lack
of specificity. Examination of the data from that laboratory
showed no acceptable values for BCEXM, but all values for BCIPE
and 4CPPE were acceptable. The values for BCEE in four of the six
waters for BPPE in five of the waters were acceptable employing
the EC detector. It was also decided not to supply the partici-
pating laboratories with standards or calibration solutions.
Rigidly controlled, however, was the use of only one specified
column packing material, although the labs were required to pur-
chase it themselves from any supplier they desired.
The method, as published in the Federal Register, had a recom-
mended temperature program. One compound (2-chlorcethyl vinyl
ether), was dropped from the study late in the method's develop-
ment due to its high volatility, which led to low and variable
recovery. The Federal Register temperature program was not opti-
mized to take advantage of this change in compounds. Because of
the wide diversity of gas chromatographs and detectors being used,
the laboratories were allowed to optimize the temperature program-
ming for their equipment.
Other operational parameters were also suggested as a results of
MC studies, for example, hydrogen flow rate, furnace tempera-
tures, electrolyte composition and flow rate, and so forth.
However, they were not mandated and were not even applicable fci
some of the detectors being used. The individual laboratories
were to start with the method conditions and optimize for their
particular instrumentation. It was found after the method study
that there were some other laboratory practices which differed
from laboratory to laboratory which were not specified in the
method. The most striking was the variety of means which were
53
-------
used to eliminate emulsion problems. Almost every laboratory had
a different method to solve this problem. Some of these methods
may be analytically superior to others. It was felt, however,
that any of this wide array of methods used would probably be ac-
ceptable. The more questionable approaches were "heat gun and
glass wool" and "gentle extraction" but the labs using these meth-
ods did not report significantly different data compared to the
other labs.
Responses to these questionnaries are presented in Table 17,
ordered in detector groupings.
In general, any of the acceptable detectors were shewn to be
capable of generating high quality results. It was initially
believed that the use of a Hall 700A detector would be a signif-
icant advantage. This was not prcven to be the case since the
data are fairly randomly distributed.
Some of the laboratories that participated in the method valida-
tion study reported interference problems. The exact causes for
these variances have not been determined, but several possibilities
exist. Sample preparation techniques (especially the Florisil
cleanup) could cause varying amounts and numbers of interferences
in the final concentrate. The many GC/detector parameters could
cause varying peak separations. Finally, there is always the pos-
sibility of accidental introduction of external interfering con-
taminates .
In the cleanup step of Method 611, Florisil seems to be most ef-
fective in removing compounds that elute during the last 70% of a
GC analysis. Therefore, BCEXM, CPPE, and BPPF are easy to quantify
even at low concentrations. Florisil was less effective on early
eluting compounds, thus causing some interferences with BCIPE and
BCEE . However, even with these two haloethers, the concentration
54
-------
TABLE 17. LABORATORY ANALYTICAL CONDITIONS
(ORDERED IN DETECTOR GROUPINGS)
LP
LP
Labor
CO
lory
JO
Ha 11 TOO
Hall 700
Ha 11 700
Hall 700
Mall 700
Hal1 700
Hal I TOO
IU11 700
Hall 700A
Hall 7©OA
Hall 700ft
Halt 700ft
Hall 70OA
Hall 700A
Hall 110b
Hall )I0
foulton
'4 S 7-I008
Coulson
tMttrwno
(X - 20 ( w I ton
Itrrtron
c«pt Uf f
Good
dill Cn Tr^imlurr
_po ml % rhfoail n^riph |>» oyi •(/
21
21
I)
18
23
It
2
5
9
IS
Var tan 1400
lf(U«n 55
H-F (auto)
$7 10
Perk in-Kl«er
rertin-EI**r
900
M-P 5750
H-t SS40A
Varlan 3700
Trirar SfcO
Tracor 560
Tracer 560
Tracor 560
Tracor S60
Tracor SiO
Trtcor 222
Tracor HT-2
Trftcor 222
H-P (auto)
57|0
Pirk|n-(lwr
9000
H-r 5> IOA
60" - 2 lain
6*/am «o 210*
100" - 4 m,n
Ib'/min to 230*
100* - 4 all*
H*/min to 2*0*
Standard
75* - 4 "in
I 2"/ain to 200*
60" • 2 am
8"/ain to 2 JO*
standard
Standard
60* - 4 mm
l(>*/ain to 230*
Standard
100" - 4 ain
l6"/aln to 220*
70* • 2 ain
8*/aln to 230*
»0" -
8*/ain to 210*
St anrtard
100* - c. oin
16*/aln to 200*
60* - 2 »in
7 Wain to
ISO*
Standard
S t onclcr cj
Standard
Hf I nw Hydimfen
| I nw IIim
rate. r •» t » .
_»L/»tn aiy*in_
40
30
40
40
40
20
25
30
25
40
40
30
30
30
50
60
50
20
20
80
35
SB
40
rutnac*
traperilurt,
•C
Cltct rolyt*
tlrrtrolftff flow rata,
co«|>oait ion n
"»oo
875
0'5
900
895
880
900
900
900
800
875
900
820
820
925
910
860
820
F t hano I
n-Propartol
75/25
tthanol/H|0
I topropyl
• at r r
75/25
f t hanol/HfO
75/25
Cthanol/H|0
90/10
tthanol/HgO
75/25
Cthanol/HjO
Fropanol
n-Propanol
n-Propanol
95/5
tthanol/H|0
Itopropyl
n-Propanol
75/25
f t hanol/MjO
w«(«r
0 )
0 4
0 5
0 5
0 5
3 5
0 4
0 5
1
1 I
0 6
0 3
0 5
4
15
0 3
70\ Act Ic
acid in H,0
l< ont i »»*>»••! I
-------
TABLE 17 (continued)
laboratory el iatnat irvq rcmrrntrttion by wtilfwaifr InierJrifnrr drlf( t«r inilytr on Riirlinr forwent ral ion
rodr twli ion ft oi)lrai nialf i pnks srnt 11 ivilj Moristl none loo luv Othtr
11 7\ Mac I ¦ ¦ a
12 Continue*!* 2 a a Hiqh
eatraction boiler
m+motf
t 2. 3 *
1S Oentr i luq*
2 Gin* woo! x a i
trpir•lory
I utint 1
I) W*,SO, 1,7,3 ¦ ¦
10 CentI* 2, 3 * a ¦
• ¦Iract ion
I Claaa voot Too slow at no probltai
bS*C
10 N«rSO« Tooh 30 oin I. 2 * t
with lnaul
4 Separator^ ? a
f urvt* |
It Claaa wool 2 ¦
* »•»»«
• Whipping 2 M I
with wire
17 2 no prablm
7 Heat gun t, 2 a i
Ulasa wool
14 Crnlrifugt 2 a a i tuaping,
add chip*
on telvrnt
•schan9t
19 tta,S0« Too alow a a a a a
3
5 Glata wool 2 Tap a Buaplng,
add chips
o*i solvent
fnchan9*
* St irr Imj 2. J a a
rent rIfuge
20 Cl*4i wool Loogrr at 1 a
t raster at m e
It not I nu
-------
TABLE 17 (continued)
t/i
Labor
< odr
I
I
lory
20
linear
rnpontr CiMbntlow ¦>thod
No <10 (tlerature variation in the
\ran* lr» llnri
Tm
T«a
>1 r*q/aL
Tat
Tea
Mo, alaoi
Tea
Mo
T««
Lin«ir plot
Sinqlt point
"*• concent rat ion
Linear plot
Slnqla point
Linear plot
1100 rx}
1 irwii
flectrolyte flow
flectrolyte coopoaitlon
Reain bed
Coatlnq of HI t(4>r
In f * I, (In ¦)' fluctuation electrolyte
Tea
>10 nq
9o, but
cloae
• In a ~ Bi
r ¦
Slnqle point
flo»» * transfer Una
co for low
concent rat ion*
Additional cleanup netSod for
iperial probleaa
Wore effective cleam^ needed
laiie taction li»tta. Stop
at ?oo#c Check sr-?J)0 va.
I .000
Oae Floriail on atandards for
better analyala
Better cleanup
fIiainate Plroiai1
cc/m
BOT2 Blanka indicate no reaponae given In queationnoirv
*Tt*e average reaulta of the aia eater aaa^lea waa determined for each laboratory for each co^>ound in ea« h
apihinq solution ThI¦ qtvea 30 averaqea for each laboratory (S rna^onndi a t> apiMnq aolutmnO Thu
col«aw> ahiMi how aarty of thoae values lina each laboratory were wilhin of I he true value
^U*ed 700* for I and S
' tlif "atandard" proqra* furnished to the laboratories 100* for 4 atimlev Ifc* per ainute to
ond bold at JJO* tor 4 ainules This qive* a faster than the en4mfi|e pnl< I i Oteil in (he
ferieral Rfjiilrr on hereafter 1. I9'9
-------
could usually be determined by either subtracting blank interfer-
ence values or changing the GC program to effect better separation
For more details including chromatograms and specific interfering
compounds see Appendix D where other MC findings concerning Method
611 are presented.
58
-------
REFERENCES
1. Youden, W. J., Statistical Techniques for Collaborative Tests,
Association of Official Analytical Chemists, Inc., Washington,
DC, 1969, 64 pp.
2. Outler, E. C. and McCreary, J. H., Interlaboratory Method
Validation Study: Program Documentation, Battelle Columbus
Laboratories, 1982.
3. ASTM D 2777-77, 1980 Annual Book of ASTM Standards, Part 31,
pp. 16-28. American Society for Testing and Materials,
Philadelphia, PA.
4. ASTM E 178-80, 1980 Annual Book of ASTM Standards, Part 41,
pp. 206-231, American Society for Testing and Materials,
Philadelphia, PA.
5. Youden, W. J., "Statistical Manual of the AOAC," The Associa-
tion of Official Analytical Chemists, Washington, DC, 1975.
6. Thompson, W. R., "On a Criterion for the Rejection of Observa-
tions and the Distribution of the Ratio of the Deviation to
the Sample Standard Deviations", The Annals of Mathematical
Statistics, AASTA 6 (1935), pp. 214-219.
7. Britton, P. W., "Statistical Basis for Laboratory Performance
Evaluation Limits," presented at the 142nd Joint: Statistical
Meeting, Cincinnati, OH, August 17, 1982.
59
-------
APPENDIX A
TEST METHOD - HALOETHERS-METHOD 611
60
-------
&EPA
{
United Stales Environmental Monitoring and
Environmental Protection Support Laboratory
Agency Cincinnati OH 45268
Research and Development
Test Method
Haloethers —
Method 611
1. Scope and Application
1.1 This method covers the
determination of certain haloethers
The following parameters can be
determined by this method
Parameter STORET No CAS No
BiS(2-chloroethyl) ether 34273 11 1 *44-4
B!S(2-chioroethoxy) methane 34278 11 1 *91 * 1
Bis(2-chloroisopropylJ ether 34283 108-60-1
4-Bromophenyl phenyl ether 34636 101*55-3
4'Chlorophenyl phenyl ether 34641 7005-72-3
1.2 This is a gas chromatographic
{GC) method apphcab e to the
determination of the compounds listed
above in municipal and industrial
discharges as provided under 40 CFR
136 1 When this meihod rs used to
analyze unfamiliar samples for any
or all of the compounds above,
compound identifications should be
supported by at least one additional
qualitative technique This method
describes analytical conditions for a
second GC column that can be used
10 confirm measurements made
with the primary column Method 625
provides gas chromatograph/mass
spectrometer (GC/MS) conditions
appropriate for the Qualitative and
quantitative confirmation of
results for all of the parameters
listed above, using the extract
from this method.
1.3 The method detection limit
(MDL. defined in Section 14 1)'" for
each parameter is listed in Table 1
The MDL for a specific wastewater
may differ from that listed, depending
upon the nature of interferences in
the sample matrix
1.4 The sample extraction and
concentration steps in this method are
essentially the same as m methods
606. 608. 609. and 612 Thus a
single sample may be exuacteo to
measure the parameters included m
the scope of each of these methods
When cleanup is required, the
concentration levels must be high
enough to permit selecting aliQuots.
as necessary, to apply appropriate
cleanup procedures The analyst is
allowed the latitude, under Gas
Chromatography (Section 12). to
select chromatographic conditions
appropriate for the simultaneous
measurement of combinations of
these parameters
1.6 Any modification of this method,
beyond those expressly permitted
shall be considered as major
modifications subject to application
and approval of alternate test
procedures under 40 CFfl 136 4
and 136 5
1 .$ -This method is restricted to
use by or under tho supervision of
analysts experienced in the use of
611-1
61 Jubf1982
-------
gas chromatography and in the
interpretation of gas chromatograms
Each analyst must demonstrate the
ability to generate acceptable results
with this method using the procedure
described in Section 8 2
2. Summary of Method
2 1 A measured volume of sample,
approximately one-liter, tt solvent
extracted wrth methylene chloride
using a separatory funnel The
methylene chloride extract »s dried
and exchanged to hexane during
concentration to a volume of 10 ml or
less GC conditions are described
which permit the separation and
measurement of the compounds in
the extract usmg a haboe
specific detector
2 2 The method provides a Florisil
column cleanup procedure to aid in
the elimination of interferences that
may be encountered
3. Interferences
3 1 Method interferences may be
caused by contaminants in solvents,
reagents glassware, and other
sample processing hardware that lead
to discrete artifacts and/or elevated
baselines in gas chromatograms All
of these materials must be routinely
demonsvated to be free from
intertefences under the conditions of
the analyses by running laboratory
reagent blanks as described m
Section 8 5
3 11 Glassware must be
scrupulously cleaned Clean all
glassware as soon as possible after
use by rmsmg with the last solvent
useo in it This should be followed by
detergent washing With hot water,
and r.nses with tap water and reagent
water It should then be drained dry.
and heated m a muffle furnace at
400:C for !5 to 30 minutes Some
thermally stable materials, such as
PCBs may not be eliminated by this
treatment Solvent rinses with
acetone and pesticide quality hexane
may be substituted for the muffle
furnace heating Volumetric ware
should not be heated m a muffle
furnace After drying and cooling,
glassware should be sealed and
stored m a clean environment to
prevent any accumulation of dust or
other contaminants Store inverted or
capped with aluminum foil
3 12 The use of high purity
reagents and solvents helps to
minimize interference problems.
Purification of solvents by distillation
in all-glass systems may be required
3 2 Matrix interferences may be
caused by contaminants that are
coextracted from the sample The
extent of matrix interferences will
vary considerably from source to
Source. depending upon the nature
and diversity of the industrial complex
or municipality being sampled The
cleanup procedures in Section 11 can
be used to overcome many of these
interferences, but unique samples
may require additional cleanup
approaches to achieve the MDL listed
in Table 1
3 3 Dichlorobenzenes are known to
coelute wrth haloethers under some
gas chromatographic conditions ff
these materials are present together
in a sample. »t may be necessary to
analyze the extract wtth two different
column packings to completely resolve
all of the compounds
4. Safety
4.t The toxicity or carcinogenicity of
each reagent used in this method has
not been precisely defined, however,
each chemical compound should be
treated as a potential health hazard
From this viewpoint, exposure to
these chemicals must be reduced to
the lowest possible level by whatever
means available The laboratory is
responsible for mamtainirg a current
awareness file of OSHA regulations
regarding the safe handling of the
chemicals specified m this method
A reference file of material data
handling sheets should also be made
available to all personnel involved in
the chemical analysis Additional
references to laboratory safety are
available and have been identified'4*4'
for the information of the analyst.
5. Apparatus and Materials
5.1 Sampling equipment, for
discrete or composite sampling
5.1.1 Grab sample bottle * Amber
glass. one-liter or one-quart volume,
fitted with screw caps lined with
Teflon Foil may be subSMuted tor
Teflon if the sample is not corrosive
If amber bottles are not available,
protect samples from light The
container must be washed, rinsed
with acetone or methylene chloride,
and dried before use to minimize
contamination
5.12 Automatic sampler (optional) •
Must incorporate glass sample
containers for the collection of a
minimum of 250 mL Sample
containers must be kept refrigerated
at 4®C and protected from light during
compositing If the sampler uses a
peristaltic pump, a minimum length of
compressible silicone rubber tubing
may be used Before use. however,
the compressible tubing should be
thoroughly rinsed with methanol,
followed by repeated rinsings with
distilled water to minimize the
potential for contamination of the
sample An integrating flow mete' is
required to collect flow proportional
composites
6 2 Glassware (AH specifications are
suggested Catalog numbers are
included for illustration only)
5.2.1 Separatory funnel • 2000-mL.
with Teflon stopcock
5.2 2 Orying column •
Chromatographic column 400 mm
long x 19 mm ID. with coarse frit
5 2 3 Chromatographic column ~
400-mm long x 19 mm ID glass with
coarse fritted plate on bottom and
Teflon stopcock (Kontes K-420540-
0224 or equivalent)
5 2 4 Concentrator tube. Kude^na-
Danish - 10-mL graduated (Kontes
K-570050-1025 or equivalent)
Calibration must be checked at the
volumes employed m the test Ground
glass stopper is used to prevent
evaporation of extracts
5 2 5 Evaporative flask. Kuderna
Danish - 500 mL (Kontes K-570001-
0500 or equivalent) Attach to
concentrator tube with springs
5.2.6 Snyder column. Kuderna-
Danish - Three-ball macro (Kontes
K-503000-0121 or equivalent)
5.2.7 Vials - Amber glass. 10- to
15- mL capacity, with Teflon-lined
screwcap
6.3 Boilmg chips • Approximately
10/40 mesh Heat to 400 C for 30
minutes or SoxhJet extract with
methylene chloride
6.4 Water bath • Heated, with
concentric ring cover, capable of
temperature control (±2°C) The bath
should be used m a hood
6.6 Balance - Analytical, capable of
accurately weighing 0 0001 g
5.6 Gas chromatograph • An
analytical system complete with
temperature programmable gas
chromatograph suitable for on-cotumn
injection and all required accessories
including syringes, analytical columns
gases, detector, and strip-chart
recorder A data system is recom-
mended for measuring peak areas
5.6.1 Column 1 - 1.B m long x 2
mm ID pyrex glass, packed with
Supaicopon. (100/120 mesh) coated
611-2
62
Jy*Y 1982
-------
with 3% SP-1000 or equivalent This
column vvas used jo develop the
method performance statements in
Section 14 Guidelines for the use of
alternate column packings are
provided m Section 12 1
5 6 2 Column 2 ¦ 1 8 m long x 2
mm ID pyre* glass packed with
Tena«-GC (60 80 mesh) or
equivalent
5 6 3 Detector ¦ Hai'de specific
electrolytic conductivity or
microcouiometr»c These detectors
have proven effective m the analysis
of wastewaters for the parameters
listed in the scope of this method The
HaM conductivity detector was used to
develop the method performance
statements in Section 14 Guidelines
for the use of alternate detectors are
provided m Section 12 1 Although
less se'ect've an electron capture
detector is an acceptable alternative
6. Reagents
6 1 Reagent water • Reagent water
is defined as a water in which an
mterferent is not observed at the
MDl of each parameter of interest
6 2 Sodium thiosulfate • (ACS)
Granuiar
6.3 Acetone methanol methylene
chloriae. hexane and petroleum ether
(boiimg range 30 to 60:C) ~ Pesticide
quality or equivalent
6.4 Sodium sulfate - (ACS)
Granular, anhydrous Purify by
heating at 400:C for four hours in
a shallow tray
6 5 Flons'l - PR Grade (60/100
mesh), purchase activated at 1250SF
and store m the dark m glass
containe' with glass stoppers or foil-
lined screw caps Before use activate
each batch overnight at 130rC in a
foil-covered glass container
6 6 Ethyl ether • Nanograde.
redistilled m glass, if necessary
6 6 1 Must be free of peroxides as
indicated by EM Laboratories Quant
test strips (Available from Scientific
Products Co . Cat No PI 126-8. and
other suppliers )
6 6 2 Procedures recommended for
removal of peroxides are provided
with the test strips After cleanup 20
mL ethyl alcohol preservative must be
added to each liter of ether
6 7 Stock standard solutions (1 00
*/g i/l) • Stock standard solutions can
be prepared from pure standard
materials or purchased as certified
solutions
6 7.1 Prepare stock standard
solutions by accurately weighing
about 0 0100 g of pure material
Dissolve the material in pesticide
quality acetone and dilute to volume
in a 10-ml volumetric flask Larger
volumes can be used at the
convenience of the analyst If
compound purity is certified at 96%
or greater, the weight can be used
without correction to calculate the
concentration of the stock standard
Commercially prepared stock
standards can be used at any
concentration if they are certified
by the manufacturer or by an
independent source
6 7.2 Transfer the stock standard
solutions into Teflon sealed screw-cap
bottles Store at 4CC and protect from
tight Stock standard solutions should
be checked frequently for signs of
degradation or evapcration. especially
just prior to preparing calibration
standards from them Quality control
check standards that can t>e used to
determine the accuracy of calibration
standards will be available for the
U.S. Environmental Pro:ection
Agency. Environmental Monitor ng
and Support Laboratory. Cincinnati.
Ohio 45268
6 7 3 Stock standard solutions must
be replaced after six months, or
sooner if comparison with checK
standards indicate a problem
7. Calibration
7.1 Establish gas chromatographic
operating parameters to produce
retention times equivalent to those
listed m Table 1 The GC chromato-
graphic system may be calibrated
using the external standard technique
(Section 7 2) or the internal standard
technique (Section 7 3)
7.2 External standard calibration
procedure
7.2.1 Prepare calibration standards
at a minimum of three concentration
levels for each parameter of interest
by add-ng volumes of one or more
stock standards to a volumetric flask
and diluting to volume with hexane
One of the external standards should
be at a concentration near, but above,
the MDL and the otfer concentra-
tions should correspond to the
expected range of ccncemrations
found in real samples or should
define the working range of the
detector
7.2.2 Using injections of 2 to 5 pi of
each calibration standard tabulate
peak height or area responses against
the mass injected The results can be
used to prepare a ca-ibration curve for
each compound Alternatively, if the
ratio of response to amount injected
(calibration factor) is a constant over
the working range (< 10% relative
standard deviation. RSD). linearity
through the origin can be assumed
and the average ratio or calibration
factor can be used in place of a
calibration curve
7.2.3 The working calibration curve
or calibration factor must be verified
on each working day by the
measurement of one or more
calibration standards If the response
for any parameter var«es fror^ the
predicted response by more than
±10%. the test must be repealed
using a fresh calibration standard
Alternatively a new calibration curve
or calibration factor must be prepared
for that compound
7.3 Internal standard calibrator
procedure To use this approach the
analyst m-jst select one or more
internal standards that are simiia'm
analytical behavior to the compounds
of interest The analyst must further
demonstrate that the measurement of
the internal standard is not affected
by method or matrix interferences
Because of these limitations, no-
mternal standard can be suggested
thai is applicable to all samples
7.3 i Prepare calibration standards
at a minimum of three concentration
levels for each parameter of interest
by adding volumes of one or more
stock standards to a volumetric flask
To each calibration standard, add a
known constant amount of one or
more internal standards, and dilute to
volume with hexane One of the
standards should be at a concentra-
tion near, but above, the MDL ana
the other concentrations should
Correspond to the expected range of
concentrations found m real samples
or should define the working range of
the detector
7.3.2 Using injections of 2 to 5 pi
of each calibration standard, tabulate
peak height or area responses aga.nst
concentration for each compound and
internal standard and calculate
response factors (P.F) for each
compound using equation 1
Eq 1 RF = (AsC„)/(A„Ci)
where
A, = Response for the parameter to
be measured
A* = Response for the internal
standard
C,» = Concentration of the internal
standard. ^ L)
Ct s Concentration of the parameter
to be measured.
6n-3
63
July J 982
-------
if the RF value over (he working range
is a constant « 10% RSDl the RF can
be assumed to be invariant and the
average RF can be used for calcula-
tions Alternatively the results can
be used to plot a calibration curve of
response ratios. A, A*, vs RF
7 3 3 The working calibration curve
or RF must be verified on each
working day bv t^e measurement of
one or more calibration standards If
the response for any parameter varies
from the predated response by more
than z 10% the test rnust be repeated
using a fresh calibration standard
Alternatively a new calibrate curve
must be prepared for that compound
7 4 Before using any cleanup
procedure the analyst must process a
series of caiibrafon standards through
the procedure to validate elution
patterns and the absence of
interferences from the reagents
7 5 The cleanup procedure m
Section 1 1 utilizes Flonsil column
Chromatography Flonsil from different
batches or sources may vary in
adsorption capacity To standardize
the amount o* FloriS'l which i$ used,
the use of launc aod value'7' is
suggested The referenced procedure
determines the adsorption from
hexane solution of lauric acid (mg) per
gfam Fiorisil The amount of Flonsil to
be usee for each column -s calculated
by dividing 1 10 by this ratio and
mult-plying by 20 g
8. Quality Control
8 1 Each laboratory that uses this
me:hod is required to operate a formal
auai'?v control program Tne minimum
requirements of this program consist
o* an initial demonstration of
laboratcy capability and the analysis
o' sp'*eo samples as a continuing
chec* or. performance The laboratory
is required to maintain performance
records to define the Quality of data
that is generated Ongoing
performance checks must be
compared wan established
performance criteria to determine if
the results of analyses are withm
accuracy and precision limits expected
of the method
8 11 Before performing any
analyses the analyst must
demonstrate the ability to generate
acceptable accuracy and precision
with this method This ab-lity is
established as described m Section
8 2
8 12 In recognition of the rapid
advances that are occurring m
Chromatography, the analyst is
permitted certain options to improve
the separations or lower the cost of
measurements Each time such
modifications are made to the method
the analyst is required to repeat the
procedure in Section 6 2
8 13 The laboratory must spike and
analyze a minimum of 10% of all
samples to monitor continuing
laboratory performance This
procedure i$ described in Section 8 4
8 2 To establish the ability to
generate acceptable accuracy and
precision. the analyst musi perform
the following operations
8 2 1 Select a representative spike
concentration for each compound to
be measured Using stock standards.
prepare a quality control check sample
concentrate in acetone 1000 tmies
more concentrated than the selected
concentrations Quality control check
sample concentrates appropriate for
use with this method, will be available
from the U S Environmental
Protection Agency. Environmental
Monitoring and Support laboratory.
Cincinnati. Ohio 45268
8 2.2 Using a pipet. add 1 00 mL of
the check sample concentrate to each
of a minimum of four 1000-mL
aliouots of reagent water A
representative wastewater may be
used m place of the reagent water,
but one or more additional ahquots
must be analyzed to determine
background levels, and the spike level
must exceed twice the background
level for the test to be valid Analyze
the ahquots according to the method
beginning in Section 10
82 3 Calculate the average percent
recovery. (Rf. and the standard
deviation of the percent recovery (si.
for the results Wastewater back-
ground corrections must be made
before R and s calculations are
performed
8 2 4 Usmg Table 2. note the
average recovery (X) and standard
deviation (p) expected for each method
parameter Compare these to the
calculated values for R and s If s >
2p or iX-R > 2p. review potential
problem areas and repeat the test
8 2 5 The U S Environmental
Protection Agency plans to
establish performance criteria for
R and % based upon the results of
mterlaboratory testing When they
become available, these criteria must
be met before any samples may be
analyzed
8 3 The analyst must calculate
method performance crrtenj and
define the performance of the
laboratory for each spike
concentration and parameter being
measured
8 3 1 Calculate upper and lower
control limits for method performance
Upper Control Limit (UCl) - R - 3 s
Lower Control Limit (ICL) R — 3 s
where R and s are calculated as m
Section 8 2 3 The UCL and LCL can
be used to construct control charts'*'
that are useful m observing trends m
performance The control limits above
must be replaced by method per-
formance criteria as they become
available from the U S Environmental
Protection Agency
8 3 2 The laboratory must develop
and maintain separate accuracy
statements of laboratory performance
for wastewater samples An accuracy
Statement for the method is def-ned
as R r s The accuracy statement
should be developed by tne analysis of
four aliquots of wastewater as
described m Section 8 2 2. followed
by the calculation of R and s
Alternately, the analyst may use four
wastewater data points gathered
through the requirement for
continuing quality control in Section
8 4 The accuracy statements should
be updated regularly8
8.4 The laboratory is required to
collect a portion of their samples in
duplicate to monitor spike recoveries
The frequency of spiked sample
analysis must be at least 10% of an
samples or one sample per month,
whichever is greater One aliquot of
the sample must be spiked and
analyzed as described m Section 8 2
If the recovery for a particular
parameter does not fall within the
control limits for method performance,
the results reported for that parameter
in all samples processed as part of
the same set must be qualified as
described in Section 1 3 3 The
laboratory should monitor the
frequency of data so qualified to
ensure that it remains at or below 5%
8 5 Before processing any samples,
the analyst should demonstrate
through the analysis of a one-liter
aliquot of reagent water, that all
glassware and reagent interferences
are under control Each time a set of
samples is extracted or there >s a
change m reagents, a laboratory
reagent blank should be processed as
a safeguard against laboratory
contamination
8 6 It is recommended that the
laboratory adopt additional quality
$11-4
Jyfy 1962
-------
assurance practices for use with this
method The specific practices that
are most productive depend upon the
needs of the laboratory and the nature
o< the samples Field duplicates may
be anaiwed to monitor the precision
of the sampling technique When
doubt exists ovef the identification of
a peak or. the chromatogram.
conf»rmatory techniques such as GC
with a dissimilar column, specific
element detector or mass spec-
trometer must be used Whenever
possible the laboratory should
perform analysis of standard
reference materials and part/c/pate
in relevant performance evaluation
studies
9. Sample Collection,
Preservation, and Handling
9 1 Grab samples mus: be collected
m glass containers Conventional
sampling practices 9 should be
followed. e*cepi that the bottle must
not be prewas^ed with sample before
co lect.on Compos:te samples should
be collected m refrigerated glass
containers in accordance with the
requirements of the program
Automatic sampling equipment must
be a free as possible of Tygon and
other potential sources of
contamination
9.2 The samples mus: be iced or
refngcated at 4*C from the time of
collection until extraction Fill the
sample bottles and. if residual
chiorme is present, add 80 mg of
sodium thiosu'fate per each liter of
water U S Environmental Protection
Agency methods 330 4 and 330 5
may be used to measure the residual
chiorme '0' F»eid test kits are available
for this purpose
9 3 All samples must be extracted
withm 7 days and completely analyzed
withm 40 days of extraction
10 Sample Extraction
101 Mark the water meniscus on
the side of the sample bottle for later
determination of sample volume Pour
the entire sample into a two-liter
separatory funnel
10.2 Add 60 mL methylene chloride
to the sample bottle, seal, and shake
30 seconds to nnse the inner walls
Transfer the solvent to the separatory
funnel and extract the sample by
Shading the funnel for two minutes
with periodic venting to release
excess pressure Allow the organic
layer to separate from the water
phase for a minimum of 10 minutes
If the emulsion interface between
layers is more than one-third the
volume of the solvent layer, the
analyst must employ mechanical
techniques to complete the phase
separation The optimum technique
depends upon the sample, but may
include Stirring, filtration of the
emulsion through glass wool,
centnfugation. or other physical
methods Collect the methylene
chloride extract in a 250-mL
Erlenmeyer flask
10 3 Add a second 60-mL volume
of methylene chloride to the sample
bottle and repeat the extraction
procedure a second time, combining
the extracts in the Erlenmeyer flask
Perform a third extraction in the same
manner
10 4 Assemble a Kudema-Oanish
(K-D) concentrator by attaching a 10-
ml concentrator tube to a 500-mL
evaporative flask Other concen-
tration devices or techniques may
be used in place of the K-D if the
requirements of Section 8 2 are met
10 5 Po ur the combined extract
through a drying column containing
about 10 cm of anhydrous sodium
sulfate, and collect tne extract in the
K-D concentrator Rinse the
Erlenmeyer flask and column with 20
to 30 mL of methylene chloride to
complete the quantitative transfer.
10.6 Add one or two clean boiling
chips to the evaporative flask and
attach a three-bail Snyder column
Prewet the Snyder column by adding
about 1 mL methylene chloride to the
top Place the K-D apparatus on a hot
water bath (60: to 65CC) so that
the concentrator tube is partially
immersed in the hot water, and the
entire lower rounded surface of the
flask is bathed with hot vapor Adjust
the vertical position of the apparatus
and the water temperature as
required to complete the
concentration in 15 to 20 minutes At
the proper rate of distillation the balls
of the column will actively chatter but
the chambers will not flood with
condensed solvent When the
apparent volume of liquid reaches 1
mL. remove the K-D apparatus and
allow it to dram and cool for at least
10 minutes
NOTE Some of the haloethers are
very volatile and significant losses will
occur m concentration steps if care is
not exercised It ts important to
mamtam a constant gentle
evaporation rate and not to allow the
liquid volume to fall below 1 to 2 mL
before removing the K«D from the hot
water b^th
10.7 Momentarily remove the
Snyder column, add 50 mL of hexane
and a new boiling chip and replace
the column Raise the temperature
of the water bath to 85 to 90' C
Concentrate the extract as m Section
10 6 except use hexane to prewet the
column When the apparent volume of
liquid reaches 1 to 2 mL. remove the
K*D and allow n to dram and cool at
least 10 minutes Remove the Snyder
column and rinse the flask and its
lower joint into the concentrator tube
with 1 to 2 mL of hexane A 5-mL
Syringe is recommended for this
operation Stooper the concentrator
lube and store refrigerated if further
processing will not be performed
immediately tf the extracts w»il be
Stored longer than two days, they
should be transferred to Teflon-sealed
screw-cap bottles
10.8 Determine the original sarspie
volume by refilling the sample bottle
to the mark and transferring the water
to a 1000-mL graduated cylinder
Record the sample volume to the
nearest 5 mL
11. Cleanup and Separation
11.1 Cleanup procedures may not
be necessary for a relatively clean
sample matrix The cleanup procedure
recommended m this method has been
used for the analysis of various clean
waters and industrial effluents if
particular circumstances demand the
use of an alternative cleanup pro-
cedure. the analyst must determine
the elution profile and demonstrate
that the recovery of each compound
of interest is no less than 65°
-------
imate'y 5 ml/mm and collect the
eluate in 8 500-mL K-D flask
equipped with a 10-mL concentrator
tube Th»s fraction should contain all
of the haioethers
112 4 Concentrate the fraction by
K-D as »n Section 10 6 except prewet
the Snyder column with hexane
When the apparatus is cool, remove
the column and rmse the flask and
its lower joint into the concentrator
lube with hexane Adjust the volume
lo 10 ml Analyze by GC (Section 1 2 }
12 Gas Chromatography
12 1 Table 1 summarizes the
recommended operating conditions for
the gas chromatograph This table
includes retention times and MDL
that were obtained under these
conditions Examples of the
parameter separations achieved
b\ these columns are shown in
Figu'es 1 and 2 Other packed
columns, chromatographic conditions,
or detectors may be used if the
requirements of Section 8 2 are met
Capillary (open-tubular) columns may
also be used if the relative standard
deviations of responses for replicate
injections are demonstrated to be less
than 6°o and the requirements of
Section 8 2 are met
12 2 Calibrate the system daily as
describeo m Section 7
12 3 if the internal standard
approach is being used the analyst
must not add the internal standard to
sample extracts until immediately
before injection into the instrument
Mix thoroughly
12 4 Inject 2 to 5 pi. of the sample
extract usmcj the solvent-flush
technique " Smaller (1 0 j/L) volumes
car be in,ected if automatic devices
are employed Record the extract
volume to the nearest 0 1 mL and the
volume injected to the nearest 0 05
l/L and the resulting peak size in area
or peak height units
12.5 The width of the retention time
window used to make identifications
Should be based upon measurements
of actual retention time variations of
standards over the course of a day
Three times the standard deviation of
a retention lime for a compound can
be used to calculate a suggested
window size, however, the experience
of the analyst should weigh heavily m
1he interpretation of chromatograms
12.6 If the response for the peak
exceeds the working range of the
system, dilute the extract and
reanalyze
12 7 H the measurement of the peak
response is prevented by the presence
Of interferences, further cleanup is
required
13. Calculations
13 1 Determine the concentration of
individual compounds m the sample
13 1.1 If the external standard
calibration procedure g)
Ve = Volume of water extracted, in
liters
13.2 Report results m micrograms
per liter without correction for
recovery data When duplicate and
spiked samples are analyzed, report
all data obtained with the sample
results
13-3 For samples processed as part
of a set where the laboratory spiked
sample recovery falls outside of the
control limits in Section 3 4. data for
the affected parameters must be
labeled as suspect
14. Method Performance
14.1 The method detection limit
(MDl) is defined as the minimum
concentration of a substance that can
be measured and reported with 99%
confidence that the value »s above
lero'1' The MDL concentrations listed
in Table 1 were obtained using
reagent water Similar results were
achieved using representative
wastewaters
14 2 This method has been tested
for linearity of recovery from spiked
reagent water and has been
demonstrated to be applicable for the
concentration range from 4X MDL to
1000 » MDL""
14 3 In a single laboratory
(Monsanto Research Center), using
spiked wastewater samples, the
average recoveries presented in Table
2 were obtained'2' Each sp'ked
sample was analyzed m tnpicate on
three separate occasions The
standard deviation of the percent
recovery is also included in Tab e 2
14 4 The U S Environmental
Protection Agency is in the p'ocess of
conducting an interlaboratory method
Study to fully define the performance
of this method
References
1 See Appendix A
2 "Determination of Haioethers
m Industrial ana Munic-pai
Wastewaters " Report to' EPA
Contract 68-03-2633 tin
preparation)
3 ASTM Annual Book of
Standards Part 31. D 3694
"Standard Practice for
Preparation of Sample
Containers and for
Preservation." American
Society for Testing anc
Materials Philadeiph.a PA.
p 679 1980
4. "Carcinogens - Working With
Carcinogens." Department of
Health, Education, and
Welfare. Public Health Service.
Center for Disease Control.
National Institute for
Occupational Safety and
Health. Publication No
77-206. Aug 1977
5 "OSHA Safety and Health
Standards. General industry "
(29CFR1910). Occupational
Safety and Health
Administration. OSHA 2206.
(Revised. January 1976'
6 "Safety »n Academic Chemistry
Laboratories." American
Chemical Society Publication
Committee on Chemical
Safety. 3rd Edition. 1979
611-6
66 Juty1992
-------
7 Mills. P A . "Variation of
Florisil Activity Simple Method
for Measuring Absorbent
Capacity end Its Use in
Standardizing Flonsil
Columns. Journal of the
Association of Official
Analytical Chemists 57. 29
(1968)
8 "Handbook of Analytical
Quality Control in Water and
Wastewater Laboratories.'
EPA-600 4-79-019, US
Environmental Protection
Agency. Environmental
Monitoring and Support
Laboratory. Cincinnati. Ohio
45268. March 1979
9 ASTM Annual Book of
Standards Part 31. D 3370.
"Standard Practice for
Sampling Water." American
Society for Testing and
Materials Phiiade'phia PA.
p 76. 1980
10 "Methods 330 4 (Titrimetric.
DP0-FAS! and 330 5
(Spectrophotometry. DPD) for
Chlorine. Total Residual."
Methods for Chemical Analysis
of Water and Wastes EPA
600-4 '79-020. U S
Environmental Protection
Agency Environmental
Monitoring and Support
Laboratory. Cincinnati. Ohio
45268. March 1979
11 Burke. J A , "Gas
Chromatography for Pesticide
Residue Analysis. Some
Practical Aspects Journal of
the Association of Official
Analytical Chemists, 48.
1037 (1965)
12 "EPA Method Validation Study
21 Method 611 (Haloethers),"
Report lor EPA Contract 68-
03-2633 (In Preparation)
Tab/a 1. Chromatographic Conditions and Method Dataction Limits
Retention Time Method
(mm ! Detection Limit
Parameter Column 1 Column 2 (vg/L)
BisQ-chlorotsopropyl) ether 8 4 9 7 0 8
8isf2 chtoroethy0 ether 9.3 9 7 03
Bis(2'Chloroethoxy} methane 73 7 70 0 05
4'Chlorophenyl phenyl ether 79 4 75 0 3 9
4 Bromophenyl phenyl ether 21.2 76.2 2 3
Column 1 conditions Supelcoport (700/720 mesh) coated with 3% SP-7000
packed m 7.8 m long x 2 mm ID glass column with helium earner gas at a flow
rate of 40 mL/min Column temperature 60*C for 2 min after injection then
program at 8'C/mm to 230°C and hold for 4 mm Under these conditions
the retention time for Aldnn is 22 6 min
Column 2 conditions Tenax-GC (60/80 mesh) packed in a 7.8 m long x 2mm
ID glass column with helium earner gas at 40 mi/mm flow rate Column
temperature 750*C for 4 mm after injection then program at 7 6^0/min to
370'C Under these conditions the retention time for Aldnn is 78 4 min
Table 2 Single Operator Accuracy and Precision
Average
Standard
Spike
Number
Percent
Deviation
Range
of
Matrix
Parameter
Recovery
%
(ug/l)
Analyses
Types
Bis(2 • chloroethyl/ether
59
4.5
97
27
3
Bis(2-chloroethoxy)methane
62
5 3
738
27
3
BiS(2-chloroisopropy/)ether
67
iO
$4
27
4-Bromophenyl phenyl ether
78
3 5
14
27
3
4'Chlorophenyl phenyl ether
73
4 5
30
27
3
6117
Juby 1992
-------
Column 3% SP-1000 on Supalcopon
Program 60°C 2 mmutas 8°/minuta to 230°C.
Datactor. Hall alactrolytic conductivity
8 10 12 14 16 18 20 22 24
Ratantion time, minutas
Figur# ^ . Gas chromatogram of haloathars
Column: 7ana* CC
Program; 150°C.-4 minutas 16°/minuta to 310°C.
Oatactor. Hall alactrolytic conductivity
< £
I *
20
24
0
4
Ratantton lima. minutas j
$11-8
Figure 2. Gat chromatogram of haloathars.
68 Jury T982
-------
APPENDIX B
RAW DATA
69
-------
TABLE 18.
RAW DATA FOR
AM^l'l. NO:
It
-------
6
7
8
9
in
1 1
11
13
14
lb
16
1/
18
19
20
TABLED LO (continued)
DISTILLED WATEK
2 6
132.UO 92.00
TAP WATtK
2 6
132. 00 92.01)
SURFACE WATKH
2 6
132.01) 92.00
wash: WAHH 1
6
132.00 92.00
WASTE WATt.K 2
2 6
132.00 92.00
WASTE WATKK 3
2 6
132.00 92.00
120.00 73.UO
146.00 93.00
60.00 78.00
132.00 94.00
13b.00 80.UO
143.00 112.00
197.30 111.20
94.11 55.60
128.70 112.50
94.50 7/. 10
11.80 221.00*
26.00* 36.00*
111.90 86.70
76.70 90.30
2b.00 18.20
180.00 123.00
82.90 52.40
100.00 87.00
* 93. bO
97.00 52.00
120.00
HI. 00
146.00
89.00
117. UU
H4.00
120.00
89.00
103.OU
94.60
280.00*
300.00*
157.30
91.10
1U0.90
54.10
103.40
77.80
7 o • 80
76. lu
128.UU
161.00*
24.UU
26.00
124.10
79.60
106.00
HU.3U
20.40
2b. 10
192.UO*
129.00*
89.95
61.bO
113.90
71.70
107.20
54.30
34.00
4 J. 01)
101).00 79.00
lbO.OO 92.00
107.00 74.00
120. Oil 94.00
108.00 7b.70
109.00 160.00
72.00 167.bO
96.13 51.05
57.70 91.HO
87.70 69.30
23.8U 74.60
28.00 2.20
93.00 89.20
130.00 88.10
31.00 20.40
138.00 79.00
76.22 71.20
14b.bO 70.40
102.20 20.80
53.00 b1.00
88.00 52.00
159.00 91.00
9b.00 b4.00
109.00 94.00
124.00 82.40
158.30 97.00
299.00* 9 7.70
97.19 25.62
90.90 77.20
83.20 73.10
17b.00 1670.OU*
24.00 51.00
89.40 85.70
122.00 91.20
lb.10 lb.20
125.00 91.00
61.73 3b.bO
117.90 97.10
86.bO 63.00
bO.00 46.00
130.00 88.00
167.00 81.00
136.00 100.00
125.00 82.00
13 b.0 0 98. 70
211.00 190.00
420.90* 339.20*
103.70 51.52
69.90 136.40
80.30 60.60
164.00 37.70
38.00 57.00
123.70 99.30
129.00 104.00
14.80 23.4U
192.00 103.00
89.13 10.40
139.20 94.30
60.40 *
33.00 31.00
110.00 75.00
149.00 91.00
116.00 58.00
125.00 84.00
96.60 67.70
210.00 164.00
206.80 4 7.90
94.11 55.60
69.40 108.40
90.00 72.50
142.00 141.UO
48.00 53.00
107.10 10b.00
6b.6U 77.40
18.90 21.20
* 73.00
45.28 36.80
111.50 52.10
121.60 74.30
49.00 41.00
-------
TABLE IB (cont.i nuod )
DlSTIl.U 11 WATf'K
AMPUL NO: 3 4
MIJL CONC: 486.00 624.00
LAB NUMBER
1
470.00
610.00
2
5/5.00
5/1.00
3
464.00
446.00
4
520.00
5/0.00
5
517.00
513.00
6
340.00
510.00
7
775.70
832.00
o
349.50
507.40
9
275.50
328.30
10
392.00
469.00
11
680.00
244.00
12
1U6.U0*
46.00
1 *
393.80
572.00
14
4/9.00
565.00
15
75.110
81.00
16
400.00
589.00
1/
519.60
bl5.70
18
392.90
596.00
19
66/.60
20
230.00
400.00
TAP WATER
3 4
4Mb.no 624.no
SURFACE WATER
3 4
486.00 624.00
WASTE WATER 1
3 4
486.00 624.00
WASTE WATER 2
3 4
486.00 624.00
WASTE WATER 3
3 4
486.00 624.00
430.00
540.00
653.00
714.00
371.00
390.00
500.00
610.00
402.00
553.00
420.00
650.00
652.40
791.20
3/8.60
4b/.60
307.10
506.90
462.00
464.00
588.00
6 lf>. 00
284.00
393.00
452.70
623.00
385.00
592.00
63.90
95.00
856.00*
915.00*
440.40
547.00
414.80
650.20
110.80
*
180.00
160.00
250.00 54U.00
6b1.00 75 7.00
348.tin "b30.no
450.00 700.00
438.00 589.00
610.00 540.00
6/4.20 130/.00*
361.20 •118.21)
307.00 430.70
37I.00 441.00
206.00 594.00
197.UO 379.00
42b.30 643.00
422.00 bf>O.U0
89.00 48.80
452.00 582.00
555.60 389.00
b81.30 584.bU
30.20 41.80
210.00 220.00
380.00 470.00
642.00 610.00
33b.00 306.00
48').00 640.00
43X.00 477.00
180.00 'j40.00
743.30 858.10
313.SO 462.20
280.50 364.40
310.00 466.00
194.00 446.00
213.00 129.00
40/.10 596.10
410.UO 4/8.00
110.00 100.80
329.00 605.00
487.60 552.80
3/6.80 504.80
86.5U 741.00
160.00 190.00
430.00 560.00
670.00 711.00
394.00 512.00
470.00 600.00
506.00 673.00
550.00 770.00
1133.00 1232.00*
391.10 462.30
374.30 303.20
398.00 474.00
1120.00 515.00
20/.00 243.00
520.40 549.20
361.00 395.00
93.00 62.90
425.00 593.00
302.50 436.10
510.20 456.50
283.00 381.60
140.00 120.00
430.00 590.00
661.00 /b/.OO
420.00 652.00
490.00 700.00
421.00 533.00
545.00 640.00
1019.00* 1448.00*
419.20 481.7Q
352.00 421.20
366.00 485.00
571.00 614.00
218.00 395.00
521.10 605.80
319.00 415.00
105.00 65.00
431.00 571.00
247.10 651.30
258.30 340.00
292.40 923.20
170.0'J 180.00
-------
TABLE 19.
RAW DATA FOR HIS(2-CHLOROETHYL)ETHER BY WATER TYPE
UISULtlU WATEK TAP WATEK SUKFACt WAU'R WASTF WATER 1 WASTL WATFK 2 WASTE WATFK 3
AMPUL NU: 1 b 1 b 1 b 1 b 1 b I b
TKiiE CONC: 1.40 1.60 1.40 1.60 1.40 1.60 1 .40 1 .60 1.4(J l.bO 1.40 1.60
LA 13 N'JMBLK
1
1.3U
l.bO
1.40
1.40
1.30
1 .SO
1.20
1.20
13.00
14.00
l.SO
1.60
2
1.10
1.20
1.10
1.10
1.10
1 .so
1 .00
1.40
1.10
0.00*
1.30
1.60
j
1.10
1..10
2.00
1.60
0. BO
1.20
1.10
1.10
0.00*
0.00*
1.20
0.80
4
0.60
0.70
1.1U
I), /o
10.00*
0.60
1.10
1.30
0.00*
o.oo*
1.00
1.00
b
1.83
1.89
1.13
2.68
1 .48
1.61
2.63
1.13
22.60
24.30
1.01
1.29
6
0. JU
u.oo*
O.UU*
0.40
0. 10
0.4U
0.4U
U.1U
6.0U
9.00
0.40
0.30
/
0.00*
0.00*
o.ou*
U.OO*
3.10
U.OO*
0.00*
O.OO*
0.00*
0.80
0.00*
1.20
8
0.00*
0.8b
0.00*
¦J.00*
0.00*
0.00*
1.1/
U.OO*
2.4U
1.38
0.1/
0.00*
9
o.oo*
(1.00*
n,no*
¦.1.00*
0.00*
o.ou*
0.00*
O.OO*
0.00*
0.00*
0.00*
0.00*
10
0.90
1 - bb
0.90
1.4/
U. 69
U.BH
U.61
1.37
0.00*
0.00*
1.03
0.90
11
0.00*
0.70*
3.20
2.10
0.70
2.2!)
3.40*
20.40*
8.40
13.20
0.10
0.80
12
0.40*
0.13*
1.90
U.bH
(1.09
1.00
O.bS
0.40
1.90
4.30
0.27*
0.13*
13
1.10
1.73
1.1/
1.40
1.09
1.32
1.49
2.20
1.10
1.90
1.25
1.18
14
0.00*
2.21
1.64
1.64
1.62
1.7b
1.30
1.63
13.90
10.40
1.92
1.79
lb
7.90*
6.90
8.80*
6.00*
9.00*
6. SO*
9.10*
8.90*
6.10
6.30
6.00*
8.60*
lb
2.40
2.60
3.?:)*
l./O*
1.10
l.bO
1.70
2.10
22.00
9.30
2.SO
1.20
17
3.70*
7.72
4.60*
9.23*
4.23
8.10*
S .63*
66.30*
0.00*
0.00*
2.8S
4.38*
18
1.40
4.20
2. 90
3.30
3.60
3.40*
1.80
1.90
1.80
0.0 J*
1.30
1.60
19
1.40
0. 70
0.00*
J. 00*
1.60
1.20
1.30
1.30
16.10
9.20
1.40
1.40
20
1.40
1.00
1.00
J. BO
0.90
1.30
1.00*
U. 90*
0. HO*
0.60*
0.90
1.20
-------
TABI.K 10 (continued)
DISTILLED WATER TAP WATER SURF At; [ WA1EK WASTt WAH.K 1 WAbU WAll.W 2 WASTE WATER 3
AMPJL W: 26 2 62 6 2 (.2626
Hot. CONl: 108.00 87.00 108.01) 87.00 108.00 87.00 10X.00 87.00 10>'-.00 87.UO 108.00 87.00
LAb NUM8EK
1
72.00
6b.00
97.00
74.00
79.00
69.00
69.00
43.00
110.00
77.00
81.00
63.00
2
116.UU
72.00
114.UO
66.00
129.00
66.00
126.00
73.00
126.no
78.00
126.00
78.00
3
57.OU
7b.00
113.00
84.00
89.00
b7.00
8 7.00
5 7.00
110.00
91.00
93.00
54.00
4
lib.00
82.00
98.00
I 7.00
110.00
86.00
90.00
72.00
120.00
84.00
105.00
88.00
5
99.2H
66.30
72. bO
74.60
82.20
6/.10
94.80
69.30
94.00
79.30
76.70
b9.90
b
85.00
70.00
93.00
100.00
53.00
84.00
73.00
51.00
76.00
70.00
79.00
5 7.00
7
133.40
91.bO
121.bO
8b.00
65.60
153.00*
241.10*
155.40*
167.30
121.50
137.50
72.10
H
/6.X4
52.73
82.39
62.84
78.78
66.91
7.62
24.64
82.35
51.44
76.24
52.73
y
61.80
73.20
b6.30
58.30
44.70
25.70
40.80*
22.90*
17.70*
20.40*
40.70
59.00
l'j
60.80
69.60
58.60
62.90
63.20
64.20
58. 70
65.80
60.80
55,90
64,60
61.60
ll
2b.70*
2b.00*
34.20
59.40
Ib.bO
45.90
5 7.20
298.00*
55. 40
47.40
38.70
bl .20
l?
20.00*
34.00*
18.00
18.00
19.00
1.50
22.00
40.00
38.00
38.00
34.00*
36.00*
13
93.70
89.50
100.30
83.bO
94.70
81.30
90.20
9 7.40
110.80
84.90
108.00
100.00
14
56.00
74.90
81.50
68.6-0
97.90
75.10
91.40
75.40
111.00
93.30
56.60
64.20
lb
40. 10
2b.40*
4b.20
22.20
3b.40
28.90
20.80
18.40
30.60
25.40
61.20
16.90
IB
133.00
94.00
136.00*
93.00*
109.00
72.00
89.00
72.00
125.00
70.00
•
60.00
17
64.11
63.00
77.0b
H4.70
71.88
6 7 .80
60.48
41.90
78.7 7
46.90
42.19
36.80
18
80.40
68.20
105.00
57.90
H8.40
59.30
78.40
61.50
88.90
59.40
88. 10
53.90
19
*
72.20
81.10*
34.70*
7b.80
18.10
74.50
53.50
39.60
*
82,80
65.10
20
6.00
8.40*
12.00
10.00
16.00
9.00
16.00*
12.00*
11.00*
9.50*
19.00
11.00
-------
TABLE 19 (continued)
DISTILLED WATER TAP WATER SURFACE WATER WAbTE WATER 1 WASTE WATER 2 WASTE. WATER 3
AMPUL NO: 3 4 3 4 3 4 3 4 3 4 j 4
TKi'E CONC: 602.00 40?.00 60?.(10 402.UU 602.00 402.00 602.00 402.UO 602.UO 402.00 6U2.0U 402.00
tAK NUMitfR
1
SB0.00
400.00
510.00
340.00
330.00
330.00
450.00
290.00
490.00
360.00
510.00
380.00
2
bb 1.00
425.00
687.00
454.00
589.00
460.00
59 '.00
421.00
601.00
423.00
569.00
427.00
J
bb 1.1)0
278.00
47b.00
271.00
392.00
340.00
43U.00
221.00
432.00
348.UO
479.00
4/2.00
4
660.00
350.00
680.00
366.00
600.00
370.00
640.00
385.00
560.00
380.00
560.00
390.UO
5
617.00
30/.00
444.00
329.00
643.00
379.00
622.00
300.00
655.00
383.00
512.00
335.00
b
270.00
170.00
300.00
200.UO
600.00
180.00
310.00
190.00
390.00
250.00
390.00
190.UO
7
677.80
41b./O
616.40
426.30
652.60
666.60
696.70
467.50
843.00
798.60*
823.00
581.00
H
4S6.00
332.10
486.90
294.20
453.00
288.60
392.10
323.30
444.40
304.50
508.20
331./U
9
23/.90
146.10
2S5.30
246.60
310.80
161.90
264.80*
154.90*
382.20*
95.70*
328.70
144.8U
10
674.00
290.00
541.UO
323.00
403.00
303.01)
368.00
285.00
476.00
323.00
441.00
303.00
i i
160.uu*
163.00*
i 3 7.00
166.00
72. uu
18?.00
4.40
200.00
127.00
226.00
i01.00
268.00
l?
123.00*
27.00*
316.00
228.00
208.00
285.00
218.00
106.00
181.00
132.00
238.00*
223.00
13
•180.70
379.70
603.40
•123.10
496.00
436.40
609.30
399.40
599.30
361.60
6U9.00
418.90
14
52U.UO
367.00
473.UO
417.UU
568.00
391.00
467.00
322.00
466.00
290.00
452.UU
284.OU
15
76.00
89.50
70.10
68.9U
38.00
72.30
82.00
92.00
110.00
43.10
92.UO
61.UO
16
4b4.00
3bH.00
876.00*
b03.00*
569.00
359.00
440.00
356.00
501.00
346.00
519.00
326.00
17
69b.00
4O0.9U
639.40
413.20
769.90
314.30
704.20
367.90
391.40
451.00
367.60
409.60
18
472.30
3:j2 .00
4 78. St!
3H6.30
640.80
360./O
438.60
363.10
466. 10
229.bO
438.70
280.40
19
b8l.b0
*
68.90*
*
32.00
2 7.80
114.80
438.90
260.10
??8.60
330.60
549.30
20
3/0.00
410.00
HO. 00
100.00
91.00
70.00
90.00*
71.00*
53.00*
51.00*
7 j .00
66.00
-------
TABLE 20. RAW DATA FOR [1] S ( 2-CHLOROETHOXY )METIIANE BY WATER TYPE
l)isr:i I ri) WATFK TAP WA'tK SL'RIACt WA11K WASTI WAT^R 1 WASTE UATEK 2 WASTE WATER 3
AMPUL NO: lSlblblblSlb
1*11 tUNC: 1.40 l.i)l) l.JO 1.00 1.40 1.00 1.40 1.00 1.40 1.00 1 .40 1.00
LAB NUMBER
1
1.10
0.90
0.70
O.HO
1.10
0.80
1.20
1.30
2.20
1.80
2.10
o
o
2
0. SO
0.60
o.so
0.70
O.bO
0.00
o.so
O.bO
0.00*
0.70
0.40
0.70
3
O.bO
0.60
l.bO
0.80
0.70
O.bO
1.10
0.40
0.00*
0.00*
2.00
1.00
4
1.00
l.'jO
0.70
O.bO
1.40
0. 30
1 . SO
0.90
1.40
1.40
1.10
1.10
b
1. %
0.94
1.4?
1.00
1.88
1.00
1.40
1.78
1.34
8.SO
0.88
1.20
b
0.40*
0.00*
0.00*
0.10*
0.30
0.30
0.40
0.00*
0.00*
0.00*
0.30
0.20
7
0.00*
0.00*
2.60
1.30
0.00*
0.00*
0.20
0.00*
0.00*
O.bO
0.00*
0.00*
8
O.b?
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
0.69
0.00*
0.00*
0.00*
9
0.00*
0.00*
0.00*
n , no*
1) [III*
0 00*
0.00*
0.00*
o.oo*
o.no*
0.00*
0.00*
10
0.93
0.79
0.88
0. H 1
0. b2
0. 18
0.38
0.2b
l.OO
0.82
1.24
0.2b
11
0.00*
o.oo*
0.00*
0.00*
2.00
0.00*
17.20*
0.20*
8.00
4.00
2.60
o.oo*
12
0.30*
0.14*
0.43
0. b3
0.11
0.74
0.4 b
0.18
1.60
1.40
1.50
0.78
13
1.38
0.9b
1.0b
0. 70
1. 18
0.81
2.00
1.00
I. b9
1.10
1.90
1.82
14
0.00*
1.18
1. b3
0.89
1.42
0.89
1.08
0.70
o.oo*
O.OO*
1.33
0.02
IS
11.00*
8.30*
9.20*
6.00
8.30*
7.10
11 .00
12.00*
8.30
b .40
4.20
4.30
lb
1.90
0.90
2.60
0.40
1.10
o./o
1.30
O.bO
O.OO*
0.00*
2.80
12.10*
i;
3.38*
3.bb*
b. 10*
b.b3
3.83*
0.43
0.9b
2.78
0.9b
2.00
2.63
4.98
18
H.bO*
b. 10*
12./')*
12./O*
b. 30*
12.60*
H.bO
8.40*
8.70
9.20
14.<)0*
4. bO
19
0. 70
0.30
o.oo*
0.00*
1.90
1.10
1 .00
0.90
bO.40*
0.00*
l.bO
l.bO
20
3/.00*
39.00*
4. 30*
0.00*
b .00*
3.20*
4.20*
4.20*
2.90*
b.20*
b. 10*
b. 90*
-------
TABf.F 70 (continued)
niSTIILEO WATTU TAP WATCH SURFACE WATER WASTE WATER 1 WAS U. WATER 2 WASTE WATER 3
AMPUL NO; 2 6 2 6 ? 6 2 6 2 6 2 6
fK'.JL l.UM: 106.(JO 126.00 106.00 126.'10 106.00 126.00 106.00 126.00 106.00 126.00 106.00 126.00
LAH NlMKER
1
76.00
68.00
83.00
110.00
59.00
93.00
72.00
76.00
79.00
85.00
63.00
62.00
2
92.00
102.00
102.00
92.00
96.00
108.00
94.00
108.UO
9S.00
115.00
101.00
118.00
3
44.00
90.UO
64.00
78.00
72.00
HI).00
70.00
68.00
HS .00
105.00
82.00
68.00
4
9 7.00
114.00
90.00
114.00
102.00
126.00
85,00
105.00
100.00
116.00
98.00
120.00
S
81. 50
94.60
62.81)
104.00
69.20
92.60
83.20
95.50
88.90
112.00
61.50
90.70
6
40.00*
60.00*
31.00*
64.00*
3 7.00
70.0')
66.00
81.00
74.00
97.00
61.00
80.00
7
118.7U
113.40
103.60
109.20
48.60
191.7U
225.UO*
195.90*
16/.20*
135.90
91.10
37.30
8
64.22
80.52
68.22
80.03
69.23
67.G2
71.55
31.58
75.63
89.98
64.22
BO.52
9
72.50
66.70
64.20
60.41}
41./0
52.90
38.70*
40.10*
42.70
117.00
116.30
114.9U
10
5?.40
87.20
48.60
84. SO
si .so
fll 7n
SO. .10
CT
X
o
46.90
67.90
56.30
Hi. 50
11
o.oo*
70.00
18.4U
1 Jb.00
J. 30
11.40
83.10
453.00*
64.80
112.OU
48.30
110.00
IX
21.00*
52.00*
19.00
60.00
21.00
3.10
27.00
6S.OO
30.00
76.00
41.00
82.00
13
9b.7U
121.90
112.60
121.10
93.4 u
110.60
93.40
143.60
112.90
132.3U
114.30
144.10
1-1
4/.2U
98.10
63.10
85.91)
76.00
99.60
68.70
91.40
93.80
109.00
4 7.30
95.30
lb
HO.SO
32.60
63. HO
21.10
4,3.60
41.00
49.90
21.30
66.70
38.80
75.80
45.10
10
102.OU
130.00
94.00
131.00
71 .00
88.00
69.00
85.00
105.00
86.00
*
80.00
1/
54.03
6H.40
63. 30
89. 40
46.09
88.70
44.25
55.30
60.97
237.50*
13.17
51.50
1*
68.30
1 Ob.80
7/.9U
8 7.60
93.70
8b. 70
74.00
80.10
94.40
63.70
67.10
88.30
19
*
69.10
75.70
66. 70
71.90
26.60
75.90
57.00
33.90
•
7(J.10
77.70
20
460.oo*
310.00*
180.00*
160.00*
280.00*
180.00*
290.00*
190.00*
180.00*
130.00*
240.00*
210.00
-------
DIMILUU WATl R
TAP WATfK
AMPIJL NO:
3
4
3
4
IKL'E CO'«C :
398.00
528.00
398.00
528.00
LAtf NUMtJEK
1
300.00
340.00
310.00
340.00
2
302.00
470.00
340.00
3/8.00
3
305.00
253.00
143.00
233.00
4
410.00
4/0.00
360.00
480.00
5
4J8.00
410.00
2 79.00
409.00
6
10.oo*
120.00*
80.00*
150.00*
/
436.80
483.bO
388.60
459.90
8
289.90
443.70
308.60
39 3.80
9
148.10
228.10
170.60
263.50
lu
2b6.00
342.00
269.00
323.00
1 1
152.00
69.00
162.00
1/2.00
12
98.00*
39.00*
224.00
336.00
13
313.90
541.60
399.40
572.20
14
339.00
425.00
296.00
46/.00
15
98.00
5 3.50
83.10
65.10
16
24b.00
346.00
521.00
559.00
1/
311.60
428.10
242.60
4/1.00
18
29 3.30
4f,H.10
292.50
'144 .40
19
362.50
*
/9.00
*
20
380.00*
520.00*
500.00*
600.00*
20 (continued)
bUiUACE WAU.U
3 4
198.00 528.00
WAS If WATFK 1
3 4
398.UO 528.DO
WAS ^E WATEK 2
3 4
398.00 528.00
WASTE WATEK 3
3 4
398.1)0 528.00
170.00
254.00
2/0.00
400.00
364.00
160.00
402.30
294. 20
193.50
233.00
8.30
168.00
35 3.10
393.0'J
9 b. 00
28?.HO
333.40
4 36.50
21.20
610.00*
290.00
44b.00
43b.00
bOO.00
502.00
130.00
727.60
139.00
2/9./O
338.00
180.00
3bO.00
581.40
44b.00
63.10
352.00
329.60
4b9.30
30.30
1000.00*
270.00
363.00
238.00
380.00
334.00
70.00
502.90
260.80
184.30*
219.00
97.10
20/.00
369.bO
JO/.00
63.00
213.00
306.10
288.1)0
/b.00
400.00*
350.00
531.00
2b9.00
bbO.OO
381.00
130.00
610.10
402.10
205.80*
315.00
297.00
Mb-00
580.20
402.00
68.40
370.00
42b.20
4/4.50
533.90
910.00*
290.00
342.00
315.00
400.00
349.00
130.00
435.30
321.60
288.20
262.00
151.00
149.00
38b.20
270.00
83.00
271.00
162.20
314.yu
1/8.60
460.00*
410.00
599.00
402.00
580.00
4/4.00
220.00
838.40
411.10
1 79.80
332.00
336.00
217.00
515.00
303.00
120.1U
434.00
484.10
338.bO
243.60
580.00*
310.00
328.00
240.00
390.00
312.00
140.00
4/2./O
336.20
247.20
255.00
131.00
224.00
372.30
205.00
38.00
219.00
144.10
249.70
170.40
4/0.00*
430.00
460.00
396.00
520.00
408.00
160.00
755.70
426.80
1 1 ) tn
327.00
440.00
65.00
499.30
322.00
110.50
34/.00
513.80
3/0.50
562.30
950.00*
-------
1
2
3
4
S
6
7
8
9
1U
12
13
14
lb
lb
1/
18
19
20
TABLfc: 21. RAW DATA FOR 4-CHIiOROI,llENYL FIIFNYL ETHF.R BY WATER TYPF
DlSTlLLfl) WATFK TAP W4TEK SUKFAU. WAItK WASH! WATER 1 WASH WATfK 2 WASTE
1 b 1 b 1 b 1 S 1 S 1
14.bO 6.60 14. bO 6.60 14.Si) 6.60 14.SO 6.60 14.SO 6.60 14.SO
9.80
6.40
12.00
6.30
12.00
b.80
12.00
b.10
19.00
14.00
13.00
14.00
6.90
lb.00
b. 20
lb.00*
H. 30*
lb.00*
8.30*
16.00
7.SO
14.00*
b. 70
4.30
11.10
S. -)0
7.00
s.oo
6. SO
2. SO
0.00*
0.00*
11.80
lb. 20
8.70
11.bO
6.00
lb.SO
b.80
1/.00
9.20
28.00
23.00
IS.00
20.30
7.99
11.bO
/.Mb
1 1 .90
S./4
14.20
4. HO
16.40
12.00
12.80
1.30*
0.00*
1.70
2.10
4.40
2.10
6.00
1S.U0*
70.00
10/.00
8.10
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
0.80
o.oo*
0.30
0.00*
0.60
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
0.00*
6.07
9.11
1.97
IS,
lj , < 40
! S. 60
7 jtn
11.00
fc. 90
12.10
/.60
0.00*
0.00*
0.00*
b.04
3.9b
b.02
?'.M
S.22
2.SI
1.67
2.7b
3.3/
0.00*
2.00
0.00*
0.00*
l.bO
0.60
3. SO*
1.70*
b.bO
6.70
53.00
42.40
1.90
2.ho*
1.20*
18.00
b. 2 0
1.10
b. 10
7.20
3.00
18.00
4.80
11.00
20.HO*
13.bO*
19.90*
14.HO*
18.00*
13.90*
19.40*
13. JO*
19.30*
14.20*
23.00*
0.00*
12.30
lb.20
7.40
13.H0
6.99
7. b 7
3.64
17.10
12.80
14.80
13.00
10.00
10.bO
8.90
12.20
8.60
lb. 00
6.60
4.90
/. 10
8.00
23.bO
11.bO
28.10*
7.30*
12.SO
7.10
16.30
7.20
24. SO
7.40
18.40
4b.03*
39.23*
44.3b*
42.33*
40.0b*
34.0b*
36.OH*
23.93*
7.46
96.32
21.00
10.00
11.HO
16.70
12.10
13.10
1 1.90
13.30
10.10
14.H0
9.80
9.10
21 .HO
b. 90
0.00*
o.oo*
9.90
4.20
3. SO
3.30
S6.40
27.SO
8.70
0.40
1.30
6.00
3. JO
3.90
4.10
b. 20
4.10
4.00
2.60
4.90
-------
TABLE 21 (continued)
Distil I.El) WATI-'R
AMPUl NU: 2 6
IKLU. CONC : 94.00 120.00
LAH NUM3FK
1
74.00
93.00
2
143.00
lbtf.00
i
40.00
95.00
4
9/.00
11H.00
S
JO. 30
103.00
6
33.0;)*
66.00*
7
114.00
123.40
8
59.60
89.21
u
12b. 60
lib.80
10
68. ko
lib.00
11
18.bO
91.10
12
24.00*
42.00*
li
110.HO*
169.70*
14
34.00
126.00
IS
100.20
88.30
H.
114.00
161.00
W
60.8U
b6.70
18
h4 .0.)
106.40
19
*
116.90
20
Jb.00
41.00
TAP WATEK
2 6
94.00 120.00
bUH'r ACE WATEN
2 6
91.00 120.00
WASTE WATEK 1
2 6
94.00 120.00
WASTK WATEK 2
2 6
94.00 120.00
WASTE WATEK 3
2 6
94.00 120.00
83.00
100.00
141.00*
1 b4.00
63.00
88.00
88.00
lib.00
6b. 50
112.00
49.00
132.00
82.70
10b.10
6b.43
84.23
102.40
103.20
73.20
118.00
46. 70
38.40
26.00
88.00
129.10*
164.20*
74.30
93.10
69.60
45.00
120.00*
183.00*
60.70
60.60
97.80
11/.bO
64.60*
72.10*
/o.oo
bl .00
74.00
96.00
138.00*
lbH.OO*
6 / .00
8/.00
104.00
12b.00
b/.OO
7H.30
32.00
103.00
44.10
213.40*
/b.b6
79.73
93.60
119.00
64. bO
111.00
11.80*
35.00*
J4.00
6.60
106.bO*
161.30*
88.70
101.00
bb. 60
32.20
89.00
110.00
47.39
74.40
83.80
9b.40
S 7.10
24.10
61.00
b4.00
71.00 91.00
140.00* 158.00*
S9.00 b9.00
81.00 109.00
86.00 101.00
4-j. 00 49.00
141.10 12(>.2U
54.36 30.79
!??.r;0 76.40
64.bO lib.00
68.10 bib. 00*
IB.00 9b.00
lib. /O* 186.10*
73.80 101.00
56.10 93.20
79.00 108.00
30.74 46.40
60.90 4b.90
101.20 60.l,0
6?.00 46.00
88.00
100.00
141.00
160.00
81.00
93.00
100.00
112.00
b9. 80
92.20
74.00
113.00
58. 30
69.00
78.99
68.19
86.00
329.70*
6/.20
86.10
36. 30
20.90
34.00
73.00
123.70*
174.00*
82.30
110.00
120.20
63.10
89.00
109.00
44.51
53.90
84.70
63.50
39. 70
*
130.00
41.00
74.00
92.00
145.00*
159.00*
76.00
68.00
100.00
120.00
b7.40
83.70
76.00
100.00
115.40
22.50
59.60
89.21
214.10*
239.70*
70.20
108.00
27.70
40.00
7b.00
117.00
lb3.60*
163.30*
b 7.30
125.00
83.00
38.20
96.00
30.29
49.70
81.50
72.00
69.10
93.20
48.00
53.00
-------
TABl.K 2\ (continued)
UlSUl.LEl) VJATTAP WATtK SURfACE WATEK WASTE WATtK 1 WASTE WATEK 2 WASTE WAfEK 3
AMPUL NO: 3 4 3 1 3 4 3 4 3 4 3 4
iKUb UJNC: 489.00 424.00 489.00 424.00 489.00 424.00 489.00 424.00 489.00 424.00 489.00 424.00
LAB NtlMtUK
1
2
3
4
b
6
7
8
9
10
11
12
13
14
lb
lb
1/
18
19
20
440.00
703.00
441.00
bbO.OO
484.00
120.00*
645.70
340.80
388.60
169.00
2/2.00
84.00'
588.40'
450.00
100.00
'HI.00
4 7 0.bO
4 73.80
4//. /()
120.00
330.00
712.00
294.00
480.00
309.00
130.00*
480.00
37b.20
187.60
16 7.00
207.00
55.00*
bOO.1)0*
420.00
63.00
407.00
294.30
498.70
*
180.00
460.00
7b0.00
418.00
vio.oo
407.00
170.00
561.90
417.90
411.00
169.00
240.00
211.00
682.SO*
373.00
lbO.OO
8 lb.00*
360.20
496.30
112.30*
150.00
390.00
6/0.00
280.00
MJIJ.UO
3/9.00
70.00
403.00
329.40
482.40
144.00
202.00
244.00
o47.00*
394.00
83.10
575.00*
351.00
451.bC
*
150.00
260.00
768.00'
411.00
398.00
380.00
2bO.00
594.30
421.20
545.30
lh'J.ftO
129.00'
208.00
b91.201
4bb.00
210.00
462.(10
436.30
4b8.90
33.90
210.00
340.00
678.00*
389.00
490.00
346.00
130.00
813.40*
12b.70
384.10
16 3.00
19b.00*
310.00
606.00*
398.00
84.20
3b7.00
226.70
4b0.00
16.40
180.00
390.00
712.00*
31 / .00
395.00
424.00
130.00
581.90
261.50
360.20
160.00
41.bO
269.00
599.70*
389.00
110.00
312.00
262.50
334.60
211.20
230.00
340.00
662.00*
262.00
560.00
335.00
190.00
472.90
254.50
322.60
174.00
145.00
13b.00
600.90*
361.00
b3. 20
340.00
228.30
314.00
14 3.80
400.00
390.00
683.00
345.00
450.00
654.00
230.00
441.10
404.90
b/9.90
183.00
291.00
228.00
621.50*
209.00
200.00
41/.00
123.10
414.60
35S.60
130.00
390.00
621.00
363.00
530.00
356.00
130.00
410.30
248.10
349.40
1/1.00
166.00
244.00
599.00*
173.00
210.40
450.00
216.40
210.30
419.00
120.00
380.00
740.00*
38R.00
428.00
435.00
230.00
854.30*
439.90
587.90
186.00
291.00
198.00
613.90*
373.00
151.00
370.00
208.40
385.60
239.20
220.00
390.00
682.00*
540.(10
510.00
3dH.no
160.00
883.90*
351.00
552.80
164.00
208.00
262.00
509.70*
384.00
10b.80
358.00
3S9.60
341.70
552.30
lbO.OO
-------
TABLE* 22.
RAW DATA FOR 4-BROMOPHENYL PHENYL ETHER BY WATER TYPE
l) I b TIL1.10 WAT TAP WAT t k SIJRfACF WATlH WASTf WATFR 1 WASTE! WATER 2 WASTE WATER 3
AMPUL NO: lOlblSlSlblS
CUV!: 2.80 3.HO ?.HO 3.HO 2.80 3.HO 2.80 3.HO 2.HO 3.HO 2.80 3.HO
I Art NUMBf K
1
1.90
3.60
?.bO
3.70
1.70
4.00
2.90
3.30
3.10
4.00
3.10
4.30
2
2.40
4.30
1.80
2.30
KHO
4./0
2.00
2.70
2.00
3. 7fl
0.00*
3.30
3
4.00
3.90
6.00
0.10
4.30
4.80
3.70
2.00
0.00*
0.00*
8.00
6.40
4
3.60
7.00
3.40
4.20
b. 60
4. SO
8.20
8.20
12.00*
12.00
4.00
4.80
O
6. bb
7. 2b
3. bO
6.82
b, 4 3
b.19
4.41
b. U
13.00*
30.20*
4.08
6.86
b
0.00*
0.00*
0.00*
0.90
2.20
0.40
/.UO
2.00
2.00
4.00
2.30
2.30
7
0.00*
0.00*
0.00*
u.oo*
0.001
0.00*
1.20
0.00*
1.8/
0.00*
0.80
o.oo*
8
O.OO*
0.00*
0.00*
0.00*
0.00*
6.29
0.00*
0.00*
0.00*
0,(10*
O.OU*
0.00*
9
0.00*
0.00*
0.00*
0.00*
o.oo*
0.00*
o.oo*
0.00*
0.00*
0.00*
0.00*
0.00*
10
2.44
2.Hb
2.6b
2.47
2.9H
2.4b
o.oo*
2.06
2.17
2.07
2.17
1.98
11
0.00*
0.00*
0.00*
0.00*
11.60
7.00
8.60
b/.10*
4.30
220.00*
4.30
0.00
1?
1). llj
0.42
1.40
4 .20
0.9b
2./0
1.30
2.00
1.10
0.7b
1.60
0.48
13
3.43
10.20
2.HI
B.40
3.06
8.34
3.86
9.10
3.94
8.39
3.94
8.60
14
0.(10*
9. bb
3. HO
S. 2 1
3.3b
5.40
K 3b
4.04
6.74
16.10
6.74
4.09
lb
8.00
9.00
a. uo
10.00
10.10
9.00
4.00
8.30
6.00
b.bO
6.00
6.30
lb
b. 70
7.00
(3.20*
9.60*
3.00
3.60
1.00
b.SO
7./0
13.00
7.70
4.80
17
/.4b
22.2H*
0.23
21.13*
6.6 b
16.63*
6.13
10.00
J./O
14.38
3.70
17.98*
1H
y. 30
8.10
8.80
10.00
4.10
9. bO
9.00
9.90
1.20
8.00
1.20
0.00*
19
b.OJ
9.50
0.1)0*
0.00*
3.40
4. bO
1.10
2.00
4.30
0.00*
4.30
0.00
21)
0.80*
1.30*
2.HO
km;)
3.00
1.90
2.00
2.10
2.70
1.80
2.70
1.90
-------
TABLE 22 (continued)
UlSMU.f.O WATF.K TAP WATt'K SUWFACF! WATER WASU WAIfcK 1 WAS IE WATtK 2 WAbTE WATtR 3
AMPUl NO: 2626262b2626
TRUE CONC: 145.OU 116.00 14b.00 116.00 14b.00 116.00 14b.00 116.00 14b.00 116.00 145.00 116.00
LAB NUMBER
1
2
3
4
b
7
8
9
10
1 i
12
13
14
lb
16
1/
18
19
2U
120.OU
192.00
6b. 00
lib.00
139.00
7 I. UO
245.80
9 5. y 3
266.40
lyn.oo
69.00
37.00
134.70
79.10
bU.bO
204.00
86.04
102.30
*
16.00*
8b. 00
113.00
98.00
104.00
U 3.00
70.00
166.40
57.90
64.30
200.00
116.00
39.00
128.20
112.00
41.10
149.00
7.43
95.90
143.70
23.00'
130.00
184.00
93.00
14b.00
112.00
8b.00
1Kb.90
100.HU
257.90
21b.00
74.00
3b.00
149.10
88.20
40.10
196.UO*
63.81
122.10
133.SO
83.UO
110.00
113.00
89.00
128.00
125.00
96.00
161.40
47.66
48.60
? 10.00
49.00
84.00
112.bO
88.SO
3'J.10
166.00*
43.20
8H.90
89. 1U
45.00
120.00
203.00
103.00
155.00
9 7.30
46.00
103.80
102.SO
246.bO
189.00
68.00
45.00
120.10
131.00
38.90
161.00
43.12
12b.00
119.bO
120.00
100.00
117.00
89.00
148.00
100.00
80.00
197.60
60.43
82.30
188.00
50.00
2.20
113.30
106.00
28.90
126.00
82.30
88.70
32.20
84.00
120.00
197.00
89.00
134.00
lb 1~UO
118.00
286.60
77.47
226.20
189.00
99.10
26.00
139.80
101.00
109.00
128.00
23.89
8b.20
186.bO
90.00
96.00
1 17.01)
61.00
118.00
103.00
66.00
271.80*
153.40
44.bO
205.00
4 36.00*
89.00
134.00
93.70
28. bO
112.00
39.30
96.30
61.40
bl.00
120.00
213.00
120.00
145.00
159.00*
100.00
2/2.30
118.80
262.80
191.00
247.00
SO.00
146.80
162.00
b6.10
140.00
37.40
143."0
69.40
bl.00
89.00
128.00
89.00
118.00
139.00*
108.00
2?b.80*
46.14
366.70'
86.50
31.40
88.00
129.20
132.00
48.30
126100
45.60
109.00
i
38.00
130.00
216.00
110.00
148.00
91.80
120.00
23b.10
95.93
4?H.60*
181.00
63.00
101.00
183.90
98.50
42.10
*
25.18
114.20
120.50
71.00
94.00
129.00
67.00
125.00
83.70
81.00
20.20
57.90
195.60
194.00
83.00
112.00
154.00
162.00
48.30
101.00
43.80
S9.10
84.80
41.00
-------
TABLE 22 (continued)
OISTILI.ED WATER fAP WATER SURFACE WATER WAbTE WATER 1 WASTE WATER 2 WASTE WATER 3
AMPUL NO: 3 4 3 4 3 4 3 4 3 4 3 4
IR'.ll CUNC: bbX.OO 626.00 552.00 626.00 5b2.00 626.UO bb2.nO 626.00 5b?.DO 626.00 552.00 626.00
LAB NUMBER
1
550.00
550.00
bOO.OO
560.00
330.00
470.00
490.00
550.00
420.00
510.00
480.00
600.UO
2
526.00
694.00
554.00
74 7.00
b34.00
673.00
547.00
691.00
bbl.00
621.00
bl9.00
773.00
J
4M7.no
397.00
47b.00
380.00
483.00
562.00
333.00
176.00
361.00
4S0.00
431.00
558.00
4
blU.OO
b40.00
690.00
760.00
640.00
6 70.00
b75.00
610.00
bOO.OO
610.00
690.00
770.00
b
528.Oo
415.00
44 7.00
530.0!)
40b.00
466.00
4/2.00
470.00
586.00*
bbO.OU*
413.00
463.UO
b
130.00
180.00
160.00
200.00
320.00
170.00
1bO .00
240.00
200.00
230.00
200.00
240.00
7
927.SO
874.30
835.10
820.70
87b.50
1142.00
899.bO
9/3.90
1101.00
1841.00*
10/8.00*
1019.00
6
405. 1(1
467.10
439.20
506.90
476.bO
413.80
229.10
328.00
441.90
312.20
35 7.90
340.90
9
804.4U
776.40
897.HO
1089.00
1289.00*
692.40
460.90
471.70
1033.00
603.70
1340.00*
1133.00
10
410.00
46b.00
388.00
448.00
411.00
4b 1.00
48/ 00
4/b.O'J
448 00
4/0.00
439.OU
464.00
n
33B.OO
338.00
282.00
320.00
113.00
310.00
606.00
237.00
296.UO
302.UO
99.00
343.00
12
138.00
63.00
519.00
536.OU
394.00
626.00
405.00
258.00
315.00
182.00
328.00
556.00
13
499.30
679.30
572.80
651.00
S26.00
706.00
b7b.30
637.10
539.80
627.30
509.30
541.90
14
b72.00
609.00
505.00
613.00
583.00
bH4.00
4S5.00
S32.00
396.00
40b.00
429.OU
520.00
lb
lbU.OO
39.00
121.00
63.00
160.1)0
39.10
9b.00
bb. 10
100.00
73.80
16U.OO
bb. UO
16
538.00
733.00
1179.OU*
99b.00*
bb8.00
681.00
376.00
565.00
376.UO
480.00
493.OU
626.00
1/
b7U.bO
447.70
4 27.bO
538.60
4S2.S0
309.30
218.00
297.20
HI.20
229.10
9S.40
573.20
18
464.80
622.60
485.20
607.SO
S04.90
b/2.70
330.10
410.90
b22.10
3/2.00
439.6U
4/2.20
19
710.(10
*
17S.00
#
bb. bO
80.10
240.00
297.60
440.bO
bb4./O
282.50
766.60
20
41.00*
91.00*
180.00
210.00
2bO.OO
170.00
210.00
240.00
160.00
180.00
210.00
240.UO
-------
APPENDIX C
EFFECT OF WATER TYPE
8^
-------
-------
TABLE 24. FFFFCT OF WATER TYPF ON HI S ( 2-CHLOROFTHYL) L'THKR ANALYSIS
** P(JINF tSTlMATLS **
U 1ST ILU.U WATER SI DPI :C.AKMA( 1) - .ySSOS
WAlt R 1 NTTRCEPT( WATER-t) i S TILLEO) SLOPE(WATER-D1 ST ILLED)
2
.0*30
-.0148
3
-.2110
.0011
4
2/OS
.0133
5
l.bS30
-.2H41
6
-.3299
.0453
SCURCE
** ANALYSIS OF VARIANCE **
OF SUM OF SQUARES MEAN SQUARE
cc
REG(OISTILLEO)
REG{WAJER/UISTILLED)
f- uunu
1
10
bb,'
267/.86120
62.7/06/
i/4.2/S10
26//.H6120
6.2//0/
. 3;;7 j 6
F PROB
20.42 .0000
TOTAL
578 29l4.«J0698
TABLE OF 9S% CONFIOENCE INTERVALS FOR THE DIFFERENCES BETWEEN INTERCEPTS ANO THE DIFFERENCES BETWEEN SLOPES
WATER
INTERCEPT(WATER-DlSTILLED)
ESTIMATE INTERVAL
SLOPE(WATER-01 STILLED)
ESI IMATE INlERVAl
2
.0430 (
-.4113 f
.49/2)
-.014H (
-.1116 ,
.0820)
3
-.2110 (
-.6446 ,
,222b)
.0011 (
-.0914 ,
.0442)
4
-.2/05 (
-. /1 b 4 ,
.1/44)
.0133 (
-.0823 ,
.1089)
b
1.SS.10 (
1.04/1
2.ooa;i)
-.2H4 1 (
-.3B14 ,
-. 1 H68)
b
-.32'W (
-./'HH ,
.1001)
.04'. 3 (
-.04/7 ,
. i;w2)
NOTE : IF ZERO IS CONTAINED WITHIN A (,|VEN CONFIDENCE INTERVAL THIN THERE IS NO STATISTICAL SI TjNir ICANLl BETWEEN
1)1 SFII Lin WATi 4 AND rill CORRESPOND IM) WASH WAtl.K I ll-t HI; ASSOC I A1t. D PAKAME11 R(INII KCE»5/SLOPE).
IHf M !)P( AND [NM.KCU'I EST I MAT! S I HUM THIS ANALYSIS AUi Hi) I lill SAME AS THOSE OKIAINEO I' ROM T: IF PRECISION
AN.) AMUPAI.f -UMU.SSIONS PI RIORMI !) 1ARIIIR.
-------
TABLE 2 b. EFFECT OF WATER TYPE ON BIS(2-CHLOROETHOXY)METHANE ANALYSIS
** POINT ESTIMATES **
01S IILI.H) WA 11 R SL DPI : UAMfAt 1) - .97/08
WAT t R INTERCEPT(*ATFR-0 I ST ILLFU) SLOPE (WAT ER -D I ST ILLEO)
2
. 18S4
-.0324
3
.1 120
- - OS?1
4
. lb 48
-.030/
S
.898/
-.1419
6
.4922
-.0820
SOURCE
** ANAl YS1S OF VARIANCF **
L)F SUM OF SQUARES MEAN SQUARE F PR08
oc
REG(OISTILLED) 1 2700.34836 2700.3483b
HEG(WATE"-8S4)
-.0324 (
-.1370 ,
.0722)
3
.1120 (
-.3/b2 ,
.bTJ3)
-. Ob21 (
-. 1b40 ,
.0449)
4
. IS'18 (
-..no3 ,
.f>398)
-.0307 (
-.1330 ,
.0/16)
S
. Hl)H 7 (
.3W ,
1.40/2)
-.1419 (
-.24/4 ,
-.0364)
b
.4422 (
.0102 ,
.9/42)
-.0820 (
-.1832 ,
J)IM2)
NO! F : IF /FRO Is CONTAINED WJ'UlN A filVI N 10N(- I PENCE IN T t KVAl THF.N THI RE IS NO SIAflSIICAl SIGNIFICANCE UElVUfN
:)ISTii.lei) wmer anij nif f.imm s.m.'ndiMi wasii. wauk mm Iml assouaieu ''arami. h ¦«{ inm.kqpi/m.os": ).
IHL SLOPE AM) 1NILR:;IPT ESU^AIES f ' I i 1S ANALYSIS ARl. NO' I Hi. SAMI AS mov. OP. T AI \ I D nuiM Till. prk;;mon
AND AU.iJKnlY Kl.'MM.SSIONS PHfURMKl) I ARI iFR.
-------
TAB Lb; 26. EFFECT OF WATKR TYPE ON 4-CHr.OROniENYI, PHKNYl, KTHER ANALISIS
** POINT 1ST IMA Tl S **
UISTILLII) WATt K SLOPt :GAMMA( 1) - .'J8157
WATER 1NTEWCEP F(WATER-01 Sf ILLEU) SLOPE(WATER-DISTILLEO)
2 -.1713 .0204
3 -.0894 -.iiiyy
4 1(>08 -.0071
5 .498/ -.1949
6 1895 .022/
SOIKCE
** ANALYSIS OF VARIANCE **
DF SUM OF SQUARES MEAN SQUARE
CO
reg(oisi;lled) i 1253.54013
w E." b {W A T E ^ / D ! S T! LLf.D) 1U IS.5/14/
ERROR 554 177.0S007
1253.54913
1. j b 1 1 b
. 31958
PROS
4.8/ .0000
TOTAL
bbb 144ft.1/06/
TAHLE OF 95% CONFIDENCE INTtKVALS FUR THE DIFFERENCES BETWEEN INTERCEPTS AND THE DIFFERENCES BETWEEN SLOPES
WATER
2
J
4
5
U
INIf RCEPT (WAll R-DI SI I I.LI 0)
ESMMAIE INTUVAL
-.1/13 {
-.0894 {
- . 1 f»08 (
.998/ (
-.HNS (
-. 8890
-.K()h7
-J'.<»9 1
.2'j/y
-. 9038
.S464)
.02/8)
.54/6)
i.b'Mb)
.s;''4k)
SI OPE (WAIt R-OIST lllh))
ESI IMA TL IN1LRVAL
.0254
,018V
.00/1
,1949
.022/
-. 1 24(>
-.lb92
-.15^4
-.3415
-.12/1
.1/54)
.1315)
.1412)
-.0483)
.1/24)
NOlt: [f /NO ! S (JUNTA I NF U WITHIN A GIVEN C9NIIDFNCI INHRVAI THEN I ML RE IS NO S T AI I S I I CAl. SIGNIFICANCE HLTW1.LN
n:sii!M:) wa i t r and r mi cokri spomiin!. wash. waim< for rut assui iatcm rarami t i k ( inti kcl im/slooi ).
T ¦ 'L SIUi'l AND IMMURf ESIIMATFS I I - IS ANALYSIS AR i. NOT IM! SAW AS T"ir>c OKTAINF.) \ R:JM THE PRECISION
A:«U Al UWACY RE oR '.SSI UNS PI K''OkMI [) t IL •<.
-------
TABLE 27. EFFECT OF WATER TYPE ON l-RROMOPHENYL PHENYL ETHER ANALYSIS
** POINT ESTIMATES **
UISTlLLEl) WATER SLOPE :tiAMMA( 1 ) = .8/998
WAT I Kt I NTt RCEPT (WATt K-iM ST I LLF U) SLOPE(WATER-DlST1LLED)
2 -.04*1 ,0171
3 .oOZb -.0106
4 .0239 -.11114
b .093b -.018b
6 -.1010 .020b
ANALYSIS OF VARIANCE **
DF SUM OF SQUARES MEAN SQUARE F PROB
SOURCE
REG(OISTILLFD) 1 2099.01021 2099.01021
REliiwATEK/MSTILLED) 10 .9/ 749 .097 7b
FRRUR 6U2
182.07048
. <02'lh
.32 .9750
total
613 2282.00818
TABLE OF 9bX CONFIDENCE INTERVALS FOR THE DIFFERENCES BETWEEN INTERCEPTS AND THE DIFFERENCES BETWEEN SLOPES
WATER
INIERCEP T(WATLR-01SIILLLD)
LSTlMAIt I NIERVAL
SLOPE(WAIER-D1ST1LLEU)
ESI IMA IL INTERVAL
.0451
,002S
,02 J9
(W3S
.1010
b696
4921
.4/32
4234
f.t)l9
4794)
.49/0)
b21 1)
6103)
3999)
.0171
.0106
0114
018S
020b
-.0894
-.1124
-.1134
-.1239
-.0824
.1236)
.0912)
.0906)
.0869)
.1214)
NU TF: IF ^ERO lb CONTAINED WITHIN A CWVtN CONFIDENCE INTERVAL THF N THERE IS NO STATISTICAL SIGNIFICANCE BETWEEN
Li I STILLED WA1ER AND 1 ME CORRESPOND I Nb WASTE WATER FOR THE ASSOCIATED PARAMF T I R ( I N TF.RCI'.P T / SLOP!. ).
TU| M.-iPE AND INTERCEPT ESTIMATES FROM :HIs ANALYSIS ARE NO' Till SAME AS THDSE OH TAINI- 0 FROM thL PRECIS!
ANO ACCURACY REPRESS IONS PERFORMED EARLIER.
-------
APPENDIX D
OTHER MC FINDINGS
91
-------
Recovery and Reproducibility Studies (Worst Case Basis)
The industrial effluent selected (Wastewater D) contained several
compounds (after extraction) which interfered with the haloether
analysis. This effluent/wastewater was a "before treatment" sam-
ple which would be expected to have high levels of interferences
present which would more typically be present in much lower con-
centrations in discharge waters. Ampul #5 of the method valida-
tion study was selected for spiking the wastewater. This solution
contained the lowest haloether concentrations used in that study.
This would therefore be a worse case study considering both con-
centration and interferences.
Eight 1-liter portions of the selected wastewater were used in this
experiment. Seven of these portions were each spiked with 1 mL of
Ampul #5 (see Table 28 for concentrations). The eighth portion
was the blank. The samples were extracted and concentrated as
described in the Federal Register method. They were analyzed using
a Hail Model 310 detector (see Table 29 for conditions) with 5-pL
manual injections. (The Model 310 detector is the least sensitive
of the Hall detectors.) After they were analyzed; all samples
were cleaned with Florisil, reconcentrated, and were again ana-
lyzed using the same conditions and injections as before. Figure
1 is the chroir.atogram of the Florisil cleaned blank/ and Figure 2
is one of the spiked sample chromatograms.
The final relative standard deviations for the seven replicate
samples in Table 28 are quite small, considering the severity of
the experimental design. The wastewater had enough interfering
compounds lo truly test the Florisil cleanup. More importantly,
the low haloether concentrations (ranging from 1 ppb to 6.6 ppb
of wastewater) are near the present detection limits for this
procedure. Small analyte losses during handling represent large
relative losses, while at higher concentrations these same small
losses would represent small relative losses. Table 28 also shows
92
-------
TABLE 28.
REPLICATION DATA
Comnound
Spiking
solution
concentration,
pg/L
7 Samples
before cleanup
7
after
Samples
cleanuD
k Average
RSD , recovery,
% %
RSDb,
°/
/o
Average
recovery,
%
BCI PL
2.4
NDC
—
NDC
BCEE
1.6
6.5 106d
9.3
91d
BCEXM
1.0
5.2 89
20.9
75
CPPE
6.6
6.4 88d
17.3
63
BPPE
3.8
19.3 93
20.6
78
aCleanup with Florisil.
^RSD = relative standard deviation.
Q
None detected due to obscuring interference peak.
dHad some interference - values were determined by
subtracting interference.
TAELE 29. CHROMATOGRAPHIC CONDITIONS FOR REPLICATION ANALYSES
I tem
Description
Chromatography
Hewlett-Parkara Model 571OA
Injection port temperature
250°C
Auxiliary transfer line
250°C
Column gas flow
30 mL/min UHP-grade helium
Column
1.8 m x 2.1 mm, glass, packed with 3%
SP-1000 on Supelcoport (100/120 mesh)
Column program
100°C for 4 min, programmed at
16°C/min to 230°C
Hall 310 furnace
850 ° C
Hall electrolyte flow
0.5 mL/min, 50/50 ethyl alcohol/water
Hydrogen flow
70 mL/min
Conductivity range (pmho)
1
Attenuation
4
93
-------
o 1.5 3 4*5 6 7 5 5 i0-5 12 U"5
Tine (min)
Figure 1. Chromatogram of wastewater blank extrac-
tion (after Florisil cleanup).
M
in
7.5
Figure 2. Chromatogram of spiked wastewater extraction.
9m
-------
that Florisil cleanup and the additional Kuderna-Danish concentra-
tion increase the analytical variation and decrease the percent
recovery as would be expected from additional sampling handling.
The chromatograms in Figures 1 and 2 show the relationship of the
haloether peaks to the interfering peaks which remained after
Florisil cleanup. The BCIPE peak is completely obscured by the
interfering compounds at 3.25 minutes. The additive effect of the
BCEE spiking can be seen on the interference peak at 4.31 minutes.
The rest of the haloethers are sufficiently separated from inter-
ferences to allow normal quantitation.
An average final recovery of 75-90% and a relative standard devia-
tion of 10-20% for low ppb samples in a worse case wastewater or
effluent demonstrates the high recovery and good reproducibility
which could be expected in end-use applications of the method.
Clean-up Effectiveness and Interference Identi fication
Eight wastewaters including the three used for the method valida-
tion study (1, 2, and 3) were analyzed by Method 611 using both
electrolytic conductivity and mass spectroscopic detection, before
and after Florisil cleanup. Chromatograms of these additional
wastewaters are shown on the following pages, along with discus-
sions about interferences in each wastewater. The chromatographic
conditions were the same as in Table 29. The spiking solution used
in these wastewaters was Ampul #5 (low level haloethers) of the
method validation study.
Figure 3 shows the chromatogram of the blank extraction of waste-
water D before Florisil cleanup. Figure 4 shows the spiked sample
of the same wastewater cleaned with Florisil. The Florisil
removed an interference which eluted at 11.6 minutes; this allowed
95
-------
quantification of CPPE. BCEWM and BPPE have no interference prob-
lems. BCEE could be determined by blank subtraction. Only BCIPE
remains undetectable at low concentrations due to interference from
bis(chloromethyl)ethyl ether.
Figure 5 shows the chromatogram of the blank extraction of waste-
water E before Florisil cleanup. Figure 6 shows a spiked sample
of the same wastewater cleaned with Florisil. The interferences
shown in Figure 5 obviously would prevent analyses of all five
haloethers at low concentrations. Figure 6 shows that all five
haloethers at low concentrations. Figure 6 shows that three of
the haloethers (BCIPE, CPPE, BPPE) could easily be quantified by
blank subtraction. Only the BCEE spike remains obscured by
"interference"; GC/MS analysis showed this "interference" to be
BCEE already in the wastewater.
Figure 7 shows the chromatogram of the blank extraction of waste-
water F before Florisil cleanup. Figure 8 shows the spiked sample
of the same wastewater cleaned with Florisil. Many of the
wastewater compounds were removed, but only BCEE could easily be
quantified. Three haloethers (BCIPE, BCEXM, and BPPE) show as
shoulders on interference peaks. This suggests the possibility
that a different GC program could provide enough separation to
quantify these three haloethers. Only CPPE is completely masked
by an interference.
Figure 9 shows the chromatogram of the blank extraction of waste-
water G before Florisil cleanup. Figure 10 shows the spiked
sample of the same wastewater cleaned with Florisil. Florisil
removed most of the wastewater compounds, especially those eluting
after five minutes. All five haloethers could be quantified, with
only two of them (BCIPE, BCEE) needing blank subtraction.
Figures 11 and 12 show the chromatograms produced by wastewater H
extracts. This wastewater is similar to wastewater G m that ail
96
-------
5
3
4.5
I.
6
7,5
9
12
TIME (MIN.)
Figure 3. Chromatogram of wastewater D extract
before Fiorisil cleanup.
CO
a.
1,5
3
4,5
0
7.5
9
10.5
12
TIME (MIN.)
Figure 4. Chromatogram of spiked wastewater D
extract after Fiorisil cleanup.
97
-------
4 6 8 10
TIME (MIN.)
12
U
Figure 5. Chromatogram of wastewater E extract
before Florisil cleanup.
CC CI
c"!
4 6 3 10
TIME (MIN.)
14
Figure 6. Chromatogram of spiked wastewater E
extract after Florisil cleanup.
98
-------
lit ^
2 4 6 a 10 12
TIME (MIN.)
14
16
Figure 7. Chromatogram of wastewater F extract
before Florisil cleanup.
6 8 10
TIME (MIN.)
Figure 8. Chromatogram of spiked wastewater F
extract after Florisil cleanup.
99
-------
five haloethers could be quantified, with only BCIPE and BCEE
needing blank subtraction.
Based on the results obtained with the five wastewaters discussed
above, and the three wastewaters used in the method validation
study, the Florisil seems to be most effective in removing com-
pounds that elute during the last 70% of a GC analysis. There-
fore, BCEXM, CPPE, and BPPE are easy to quantify even at low
concentrations. Florisil was less effective on early eluting
compounds, thus causing some interference with BCIPE and BCEE.
However, even with these two haloethers, the concentration could
usually be determined by either subtracting blank interference
values or changing the GC program to effect better separation.
Table 30 summarizes the effectiveness of the Florisil cleanup for
each haloether in each of the eight wastewaters. This table
divides the degree of interference into three classes: where the
interference completely obscures the haloether peak, where the
interference necessitates variation in the analytical technique
to quantify the haloether concentration, and where there are no
remaining interferences. It should be pointed out that these
three classes are applicable only to low haloether concentrations.
The interferences normally become insignificant for moderate and
high haloether concentrations where less sensitive detector set-
tings are used. A summary of interferences which are removed by
Florisil, and those interferences which are not removed, is pre-
sented in Table 31.
Study of Furnace Temperature Effect
The GC/Hall detector system has many parameters that affect the
sensitivity of the system for a given compound. These parameters
are interrelated and each compound responds best to a different
set of parameters. Since there is such a strong interrelation-
ship among the various conditions and compounds, no single param-
eter can be optimized independently of the other parameters.
100
-------
10 12 14
TIME (MIN.
Figure 9.
Chromatogram of wastewater G extract
before Florisil cleanup.
— +
Cj
— CQ
I
A
CO
A-JL
4 6 8 in
TIME (MIN.)
r- in
U*1 Ul
U K
JX CL
d a.
u a
12 14
Figure 10. Chromatogram of spiked wastewater G
extract after Florisil cleanup.
131
-------
¦
8 10 12
14
TIME (MIN.)
Figure 11. Chromatogram of wastewater H extract
before Florisil cleanup.
4 6 8 10
TIME (MIN.)
Figure 12. Chromatogram of spiked wastewater H
extract after Florisil cleanup.
102
-------
TABLE 30. SUMMARY OF WASTEWATER INTERFERENCES3
BCIPE
BCEE
BCEXM
CPPE
BPEE
Completely obscured*3
1
0
0
1
0
Q
Partial interference
3
4
2
1
2
No interference
4
4
6
6
6
The numbers in this table represent the number of
wastewaters in which a given haloether falls into
the described class.
Low haloether concentrations are not detected due
to larger interference peaks.
CAaditional analytical steps needed to quantify
haloether concentration.
Haloether" peak has baseline separation from
interference peaks.
103
-------
TAELE 31. INTERFERING COMPOUNDS IDENTIFIED BY GC/MSa
Haloether
RTb
Potential interference
removed bv Florisil
Potential interference
not removed by Florisil
Name
RTb
Name
RTb
BCIPE
3.4
Hexane
0.7
Cyclohexane
2.0
Methyl cyclohexane
2.3
Dibromo propane
3. C
Cg-cycl opentane
3.3
Tetrachloro
propane
3.4
Bis(chloromethyl)
ethyl ether
3.7
C2"pentane
3.9
BCEE
4.3
Benzene0
5.0
Toluene
6.5
BCEXH
7.8
Phenyl acetate
6.3
Phenol^
9.7
Butanol
7.3
C2-Alkyl benzenes
9.0-9.6
Dichloropropanol
8.2
Methyl thioethyl
benzene
8.4
Benzothiazole
9.3
CPPE
11.6
Dipropylene
c
Styrene
11.1
glycol methyl
Dioxothiocane
11.3
ester
10.3
Dimethyl malonate
11.9
Chloro ethoxy-
tolyloxy ethane
11.9
BPPE
12.3
Tripropylene
Tnthiolane
12.3
glycol
12.4
Trithiane
14.0
Hexanol
13.6
Tetrapropylene
glycol
14.2
aNot all interfering compounds could be identified.
^Retention time in minutes.
CSeen by GC/MS and may or may not be seen by GC/Hall.
^Partially removed.
104
-------
The delicate balance among the parameters was demonstrated during
a study of reactor temperatures. The GC/Hall conditions listed
in Table 29, were used except that the furnace temperature was
varied during the study. One-pL injections of Ampul #5 (low con-
centration) were made as the furnace temperature was increased
from 8G0°C to 880°C. The detector response was then recorded for
each compound at each furnace temperature. The results are shown
in Figure 13. It should be stressed that Figure 13 has true mean-
ing only for the exact parameters and instrumentation that produced
the data. Figure 13 is presented as a graphic illustration of
the way each compound reacts differently to a given temperature.
Even a slight change in GC/Hall conditions would drastically
change the presentation of Figure 13. Some of the factors affect-
ing response at various furnace temperatures follow:
• Transfer line temperatures
• GC column temperature program
• Electrolyte flow rate
• Electrolyte composition
• Hydrogen flow rate
Finally, the above data were generated using a Tracor Hall model
310, using a quartz reactor tube. Since the quartz tube is non-
catalytic, compound response is very sensitive to temperature.
The newer Tracor Hall detectors use metallic reactor tubes which
contribute a catalytic effect to the reactor. This makes the re-
actions much less temperature sensitive and broadens the range of
acceptable temperatures.
105)
-------
4
t * 1 1 r
too 110 120 130 140 150
Temperature (*C)
• 60
170
ISO
Figure 13. Hall response at various reactor temperatures.
106
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse be/ore completing/
1. REPORT NO.
EPA-600/4-84-052.
4. TITLE AND SUBTITLE
EPA METHOD STUDY 21, METHOD 611—
HALOETHERS
3. RECIPIENT'S ACCESSION NO
PB84 20 5939
5. REPORT DATF
June 1984 ¦
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
C. R. McMillin, R. C. Gable, J. M. Kyne, R. P. Quill,
A. D. Snyder, and J. A. Thomas
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Monsanto Company
1515 Nicholas Road
Dayton, OH 45407
10 PROGRAM FlEMENT NO.
CBL1A
11 CONTRACT/GRANT NO.
68-03-2633
12. SPONSORING AGENCY NAME AND ADDRESS
Quality Assurance Branch, EMSL-Cincinnati
U.S. Environmental Protection Agency
26 W. St. Clair Street
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 1-79 to 3-80
14. SPONSORING AGENCY CODE
EPA 600/06
15. SUPPLEMENTARY NOTES
Described herein are the experimental design and the results of an
interlaboratory study of an analytical method to detect haloethers in water.
The method, tPA Method 611 - Haloethers, consisted of a liquid/liquid
extraction using methylene chloride, an evaporation step using Kuderna-Danish
(K-D) evaporators, a cleanup procedure using Florisil sorbent, another K-D
evaporation of the fraction from the Florisil column, and subsequent analysis
oy gas chromatography using a halide-specific detector. The six concentration
(three Youden pairs) of spiking solutions used in this study contained bis
(2-chloroisopropyl) ether, bis(2-chloroethyl) ether, bis(2-chloroethyoxy)
methane, 4-chlorophenyl phenyl ether, and 4-bromophenyl phenyl ether. Six
water types were used in the study: distilled, tap, surface, and three
different industrial wastewaters. Statistical analysis and conclusions in
this report are based on analytical data obtained by 20 collaborating
laboratories.
17. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSATi l-jeld/'Group
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Rfportj
Unclassified
"207SECURlfY CLASS (this page,
Unclassified
21. NO. OF PAGES
117
22. PRICE
1PA Form 2520-1 (R«y. 4-77) previous edition is obsolete -j
------- |