PA:R4:72"i°n0-,3o Environmental Monitoring Series
September 1972
Analyses for Mercury in Water
A PRELIMINARY STUDY OF METHODS
National Environmental Research Center
Office of Research and Monitoring
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-R4-72-003
September 1972
ANALYSES FOR MERCURY IN WATER
A PRELIMINARY STUDY OF METHODS
John A. Winter and Harold A. Clements
Method and Performance Evaluation Activity
The Analytical Quality Control Laboratory
1014 Broadway, Cincinnati, Ohio 45202
Program Element 1H1327
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
FOREWORD
To find, through research, the means to protect, preserve, and
improve. our environment, we need a focus that accents the interplay
among the components of our physical envIronment -- the air, water,
and land. The missions of the National Environmental Research
Centers -- in Cincinnati, Research Triangle Park, N.C., Corvallis,
Oregon, and Las Vegas, Nevada -- provide this focus. The research
and monitoring activities at these centers reflect multidisciplinary
approaches to environmental problems; they provide for the study of
the effects of environmental contamination on man and the ecological
cycle and the search for systems that prevent contamination and
recover valuable resources.
Man and his supporting envelope of air, water, and land must
be protected from the multiple adverse effects of pesticides, radia-
tion, noise, and other forms of pollution as well as poor management
of solid waste. These separate pollution problems can receive inter-
related solutions through the framework of our research programs - -
programs directed to one goal -- a clean livable environment.
This publication of the National Environmental Research Center,
Cincinnati, entitled: Analyeee for Mercury in Water, A Preliminary
Method Study reports the results of a broad study of methods currently
in use in the United States and Canada for analysis of mercury in
waters. Federal agencies, states, municipalities, universities,
private laboratories, and industry should find this comparative study
of methods of analysis for mercury of vital importance in their efforts
at monitoring and controlling mercury pollution in the environment.
ANDREW W. BREIDENBACH, Ph.D.
Director, National Environmental
Research Center, Cincinnati
t.14.
-------
PREFACE
The Office of Research and Monitoring, EPA, coordinates the
collection of water quality data to determine compliance with water
quality standards, to provide information for planning of water
resources development, to determine the effectiveness of pollution
abatement procedures, and to assist in research activities. In a
large measure, the success of the pollution control program rests
upon the reliability of the information provided by the data
collection activities.
The Analytical Quality Control Laboratory, NERC, Cincinnati,
is responsible for insuring the reliability of physical, chemical,
biological, and microbiological data generated in the water programs
of EPA. Within the Analytical Quality Control Laboratory, the
Method and Performance Evaluation (M PE) Activity conducts the inter-
laboratory studies of analytical procedures to:
1) assist in the selection of EPA methods,
2) evaluate the selected methods of analyses, and
3) measure the performance of EPA analysts and laboratories.
This report describes one study in the series conducted by the
MfjPE Activity.
7,
-------
Staff of the
Method and Performance Evaluation Activity, Analytical Quality Control Laboratory
John A. Winter, Chief
Harold A. Clements, Senior Chemist
Guy F. Simes, Chemist
Everett L. Barnett, Chemist
Betty J. Smith, Secretary
VVt
-------
Table of Contents
Page
FOREWORD .
PREFACE . v
Staff, Method and Performance Evaluation Activity vii
PARTICIPATING LABORATORIES
SUMMARY
INTRODUCTION i.
DESCRIPTION OF THE STUDY 2
Sample Design 2
Preparation of Samples and Reporting of Results 2
True Values 3
Study Plan and Analytical Method 3
RESULTS 6
Raw Data 6
TREAThIENT OF DATA 16
Statistical Sunuuary 16
Rejection of Outliers 16
DISCUSSION 34
CONCLUSIONS 36
APPENDIX A. EPA Method for Mercury in Water, April 1972 39
APPENDIX B. Details on Other Method (non-EPA) 51
GLOSSARY OF TERMS
ix
-------
PARTICIPATING LABORATORIES
Forty-two analysts in 42 laboratories reported results of the analyses
of the mercury reference samples. Of the 42 laboratories, 32 were non-EPA
laboratories. The participating laboratories were:
U.S. Environmental Protection Agency Laboratories
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Cincinnati, Ohio
Analytical Quality Control Laboratory
Trace Metals Analyses Group
National Environmental Research Center
Cincinnati, Ohio
National Field Investigations Center
Office of Enforcement General Counsel
Cincinnati, Ohio
Edison Water Quality Research Laboratory
National Environmental Research Center
Edison, New Jersey
Indiana District Office
Region V
Evansville, Indiana
National Water Quality Laboratory
Duluth, Minnesota
Pacific Northwest Environmental Research Laboratory
National Environmental Research Center
Corvallis, Oregon
Southeast Environmental Research Laboratory
Nat jonal Environmental Research Center
Athens, Georgia
Surveillance F Analysis Division
Technical Support Branch
Region IX
Alameda, California
Wheeling Field Station
Region III
Wheeling, West Virginia
x
-------
Non-EPA Laboratories
Bohna Engineering & Research, Inc. New Hampshire Water Supply & Pollution
San Francisco, California Control Commission
Pesticide Surveillance Laboratory
Deer Island Treatment Plant Concord, New Hampshire
Winthrop, Massachusetts
North Carolina Department of Water
Diamond Shamrock Chemical Company fj Air Resources
Painesville, Ohio Raleigh, North Carolina
Diamond Shamrock Chemical Company North Carolina State Board of Health
Sheffield, Alabama Laboratory Division
Raleigh, North Carolina
Division of Water Resources
Charleston, West Virginia Oregon State Department of Environmental
Quality
E. I. du Pont de Nemours Co. (Inc.) Portland, Oregon
Savannah River Laboratory
Aiken, South Carolina South Carolina Pollution Control
Authority
E. I. du Pont de Nemours Co. (Inc.) Columbia, South Carolina
Savannah River Plant, Lab. Division
Aiken, South Carolina T. W. Beak Consultants Limited
Toronto, Canada
Georgia Kraft Research Company
Rome, Georgia Texas A & M University
Agriculture Analytical Service
Georgia-Pacific Corporation College Station, Texas
Bellingham Division
Bellinghain, Washington Tonis River Chemical Corporation
Toms River, New Jersey
Georgia Water Quality Control Board
Atlanta, Georgia U. S. Geological Survey
Denver, Colorado
Instrumentation Laboratory, Inc.
Lexington, Massachusetts U. S. Geological Survey
Ocala, Florida
International Paper Company
Mobile, Alabama Union Oil Company
Los Angeles Refinery
Massachusetts Division of Fish & Game Wilmington, California
Westborough, Massachusetts
University of Illinois
Milwaukee Department of Public Works College of Agriculture
Milwaukee, Wisconsin Urbana, Illinois
Monsanto Corporation University of Missouri
Sauget, Illinois Environmental Trace Substances Center
Columbia, Missouri
National Council of Paper Industry
for Air & Stream Pollution Water Pollution Research Laboratory
Gainesville, Florida Stevenage, England
Water Quality Control Laboratory Westvaco Research Center
Stanford University Laurel, Maryland
Stanford, California
X2 -
-------
ANALYSES FOR MERCURY IN WATER, A PRELIMINARY METHOD STUDY
SUMMARY
The Analytical Quality Control Laboratory, National Environmental Research
Center--Cincinnati, of the U.S. Environmental Protection Agency (EPA), conducted
a preliminary interlaboratory study on analysis of organic and inorganic mercury
by the cold vapor technique.
Sample concentrates were prepared at four mercury levels and furnished to
the analysts with the known values. Analysts added an aliquot of each concentrate
to distilled water and to a surface water of their choice. Single analyses
were made on the distilled and on the surface water samples with and without
the known increments, using the EPA Method and another of their choice (Other
Method). Recoveries were compared. The accuracy and precision of the methods
were noted as were interferences from natural water samples.
In this comparative study of the EPA Method and Other Method, the
EPA Method was equal or superior to the Other Method in accuracy and precision
in 80% of all tests. More significantly, In those samples containing organic
mercury with or without inorganic mercury, the EPA Method was equal or superior
to the Other Method. This study indicates that if water samples contain mercury
in organic form, a method involving a vigorous digestion step such as that in
the EPA Method must be used to obtain a good recovery. On pages xiv and xv a
statistical summary of this data shows the precision and accuracy measurements
obtained in this study.
XVZ4
-------
Statistical Summary
Recovery of Inorganic Mercury (RgCl 2 ) from Distilled Waters
Sample 1
Sample 2
EPA Other
Method Method
EPA Other
Method Method
True Value, i.ig/liter 0.34 0.34
Mean Recovery, pg/liter 0.341 0.409
Accuracy as %
Relative Error (Bias) 0.3 20.2
Standard Deviation, pg/liter 0.07 0.16
Relative Deviation, % 19.5 40.0
Range, pg/liter 0.32 0.80
4.2 4.2
3.94 3.86
-63 -8.0
0.54 0.77
13.6 19.9
2.5 3.6
Recovery of InorganIc Mercury (HgC1 2 ) from Natural Waters
Sample 1
Sample 2
EPA Other
Method Method
EPA Other
Method Method
True Value, pg/liter 0.34 0.34
Mean Recovery, pg/liter
(By Difference) 0.351 0.369
Accuracy as %
Relative Error (Bias) 3.2 8.5
Standard Deviation, pg/liter 0.07 0.12
Relative Deviation, % 19.8 31.5
Range, pg/liter 0.25 0.40
4.2 4.2
3.99 4.34
-5.1 3.3
0.66 0.60
16.6 13.8
3.3 2.2
x v
-------
Statistical Summary
Recovery of Organic and Inorganic Mercury (Phenyl Mercuric Acetate-HgC1 )
from Distilled waters
Sample 3
Organic Inorganic Mercury
Sample 4
Organic Mercury Alone
EPA Other
Method Method
EPA Other
Method Method
True Value, pg/liter 6.3 6.3
Mean Recovery, pg/liter 6.10 5.23
Accuracy as %
Relative Error -3.1 -17.0
Standard Deviation, pg/liter 0.68 1.45
Relative Deviation, % 11.2 27.8
Range, pg/liter 4.0 7.0
4.2 4.2
4.02 3.02
-4.2 -28.0
0.80 1.53
19.8 50.7
4.5 4.8
Recovery of Organic and Inorganic Mercury (Phenyl Mercuric Acetate-HgC1 2 )
from Natural Waters
Sample 3
Organic Inorganic Mercury
Sample 4
Organic Mercury Alone
EPA Other
Method Method
EPA Other
Method Method
True Value, pg/liter 6.3 6.3
Mean Recovery, pg/liter 5.61 6.15
(By Difference)
Accuracy as %
Relative Error —10.9 -2.4
Standard Deviation, pg/liter 1.27 1.06
Relative Deviation, % 22.7 17.2
Range, pg/liter 5.3 3.8
4.2 4.2
3.83 3.30
-8.7 -21.4
1.07 1.48
27.9 45.0
5.3 4.1
x v
-------
1
INTRODUCTION
This preliminary study on methods of analysis for mercury in water
was conducted to develop further information on the ability of the
analytical methods to measure total mercury, i.e., organic plus inorganic
mercury.
It differs from formal method studies in that the analysts received
the true value data with the samples. The main purpose of the study was
to make a preliminary comparison of methods for mercury analysis performed
with known concentrations of organic and inorganic mercury added to distilled
and natural water samples. Analysts used different digestion reagents,
different volumes of reagents and varied the times and temperatures and
separation techniques. The mercury concentration was detected and measured
by cold vapor technique using an atomic absorption spectrophotometer, Coleman
mercury meter or UV/visible spectrophotometer. One laboratory used an
emission spectrographic procedure.
It should be pointed out that the Other Method(s) used in this study
are really the methods of choice of the laboratories participating in the
study. As these methods are those in routine use, familiarity and confidence
in the procedures would tend to favor the Other Method over the proposed
EPA procedure in some laboratories.
-------
2
DESCRIPTION OF STUDY
Sample Design
In this comparison of methods for mercury analyses, the test design was
as follows:
1. Each sample was prepared as a stable concentrate in a sealed glass
ampul.
2. When an aliquot of the concentrate was diluted to volume, the
mercury was present at levels found in waters and wastewater.
Several levels of concentration were tested to cover the range
of levels found in surface waters and wastewater.
3. The sample concentrates containing known increments of organic
and/or inorganic forms of mercury were added to distilled water
and to a surface water of choice and the samples analyzed. The
surface waters were analyzed with and without increments and the
added level determined by difference.
4. Recoveries in distilled and surface waters were compared. Precision,
accuracy and interferences were noted.
Preparation of Samples and Reporting of Results
Four water sample concentrates were prepared by dissolving weighed
amounts of reagent-grade chemicals in ASTM reagent-grade distilled water
to produce accurately-known concentrations of organic and inorganic mercury.
The concentrates were preserved with 0.15% nitric acid, and were checked
for stability by repeated analyses over a period of three months. These
analyses verified the concentration and stability of each solution. Further
confirmation of the true values was achieved with analyses by an
Independent referee laboratory. The true values are shown in Table 1.
-------
3
Table 1
True Values for Parameters*
When diluted in distilled water according to instructions, the water
samples contained the following concentrations of mercury:
Sample
Mercury Form
Concentration, i.ig/liter
1
Mercuric Chloride
0.34
2
Mercuric Chloride
4.2
3
Mercuric Chloride
+ Phenyl Mercuric
Acetate
3.15 + 3.15
4
Phenyl Mercuric Acetate
4.2
*The concentration given are those calculated and added. These are not based
on analyses which are used for verification only.
Since dilution of the concentrates renders the acid ineffective as a
preservative, the analyst was instructed to add 1.5 ml of redistilled acid/liter
in sample make-up.
Study Plan and Analytical Method
The mercury samples were announced in a September 8, 1970 memorandum.
Cards were included with the invitation with which laboratories could request
samples. On September 28 and 29, 1970, the first series of samples were
mailed to requesting laboratories. Since these samples were announced again
through the quarterly AQCL Newsletter and by word of mouth, requests for
samples were received over the next 12 months. Each requesting laboratory
was sent a set(s) of four ampuls, instructions for sample preparation,
duplicated report sheets and a statement of true values. Because many labora-
tories are interested in efficiency of different digestion procedures used to
convert organic mercury prior to analysis, the Analytical Quality Control
Laboratory requested that data be returned from all laboratories for
preparation of an informal report to all participants. A cut-off date of
-------
5
September 30, 1971 was set and all data returned to M PE by that date are
included in this report.
The data obtained using the basIc EPA Method (acid-.permanganate-persulfate
oxidation) are combined. Because the many laboratories used almost as many
method variations, there was not sufficient data to statistically treat these
method results separately. Therefore, all data obtained by methods other
than the EPA Method are combined as “Other Method”. However, anyone who
wishes, can retrieve any segment of data obtained using an other method and
make a separate comparison with recoveries by the EPA Method. Each method
is specified as clearly as possible without identifying the laboratory. The
number of significant figures used are equal to the true values which contain
two figures, or to one significant figure if so reported by the laboratory.
Any reported “less than” values were accepted but could not be used in the
computerized evaluation.
-------
6
RESULTS
Raw Data
The test results are reported by sample in Tables 2, 3, 4, and 5, using a
numbering of the laboratories and analysts in the order in which the data were
received. For distilled water samples, a simple listing is given of labora-
tories and analysts with their reported values. For the natural water samples,
values determined by difference were reported by each analyst and because spiking
995 ml of natural water with 5 ml of sample concentrate causes a 0.5% dilution
change, the reported values for natural waters were reduced by this amount
whenever significant. To avoid premature round-off errors in computer calcula-
tions four digits were carried in the statistical tables. However, a maximum
of two significant figures was used for all final values based on the number of
significant figures in the increment values . When values containing only one
significant figure were returned by an analyzing laboratory, these figures are
reported as received.
It should be noted that the provisional EPA Method as used in this study
included acid and permanganate plus potassium persulfate to improve oxidation,
but no heating step. However, later data have shown that heating is required
to recover methyl mercury chloride from natural waters.
The approved EPA Method now includes an acid-permanganate-persulfate
oxidation procedure and heat, prior to reduction, to assure conversion of all
organic mercury compounds to ionic form before AA measurement. It is given in
Appendix A and is further described in: Kopp, J. F., Longbottom, M. C., and
Lobring, L. B., “ ‘Cold Vapor’ Method for the Determination of Mercury,” JAWWA,
Vol. 64, No. 1, January, 1972, pp. 20-25. However, the heating step was not
needed to recover 100% of the particular inorganic and organic mercury compounds
used in this study .
-------
Because the analytical methods, other than EPA, varied in the kinds and
concentration of oxidizing reagent, time and temperature of oxidation, etc.,
each laboratory was asked to describe its method in detail or to furnish a
complete published reference to it. This information is given in Appendix B.
7
-------
8
TABLE 2
Average Recoveries of Increment; 0.34 pg of Inorganic Mercury/liter, (Sample 1)
Concentration, pg/liter
Laboratory
Code
Distilled
Water
Natural
Water
Description
of
Other Method
EPA
Method
Other
Method
EPA
Method
Other
Method
1
2
<0.5
0.3
0.48
Ag wire deposition
Sulfuric, nitric, and
perchloric acid reflux
3
4
0.52
0.2
0.39
0.2
Acid permanganate
CdS pad
5
0.4
0.3
Cu wire deposition
6
0.33
0.34
0.35
Permanganate
7
8
0.40
0.42
0.3
0.36
0.4
Acid permanganate
Acid permanganate
(overnight)
9
10
<0.5
0.3
0.5
<0.5
0.3
Acid permanganate
11
1)0.25
2)0.45
0.35
l)Acid persulfate
at 55 C
2)Acid persulfate
at 75 C
12
13
0.3
0.4
0.5
0.3
0.4
0.3
Acid permanganate
14
15
0.38
0.31
0.32
Acid + heat
16
17
0.4
0.25
0.3
0.3
18
19
0.30
.33
0.30
Acid perinanganate
20
0.56
Acid, perinanganate,
and persulfate with
heating
21
22
<0.5
<0.5
<0.5
<0.5
Acid perinangante
23
24
0.35
0.35
0.55
25
26
0.38
0.33
0.34
-------
9
TABLE 2 contd.
Average Recoveries of Increment; 0.34 pg of Inorganic Mercury/liter, (Sample 1)
Concentrat Ion, pg/i Iter
Laboratory
Distilled Water
Natural Water
Description
of
EPA
Other
EPA Other
Code
Method
Method
Method Method
Other Method
27
0.35
0.35
28
1)0.6
2)0.8
l)Acid permanganate (DOW)
2)Acid permanganate +
KBr (Modified DOW)
29
0.4
30
0.32
0.32
Acid permanganate —
31
1)0.40
2)0.60
1)0.40
2)0.60
l)Acid permanganate (DOW)
2)Acid permanganate +
KBr (Modified DOW)
32
0.2
0 (Sewage Sample)
33
.28
34
l)
-------
10
TABLE 3
Average Recoveries of Increment; 4.2 iig of Inorganic Mercury/liter, (Sample 2)
ConcentratIon, pg/liter
Laboratory
Distilled
Water
Natural
Water
Description
of
EPA
Other
EPA
Other
Code
Method
Method
Method
Method
Other Method
1
4.0
4.3
Silver wire
deposition
2
4.4
Sulfuric, nitric,
and perchioric
acid reflux
3
4.2
3.9
Acid permanganate
4
4
3
CdSpad
5
4.5
4.4
Copper wire
deposition
6
3.86
3.99
3.90
Permanganate
7
3.81
4.2
4.1
Acid permanganate
8
4.0
4.0
Acid permanganate
(overnight)
9
4.5
4.5
10
3.8
5.1
3.0
Acid permanganate
11
1)3.55
2)4.30
4.90
l)Acid persulfate
at 55 C
l)Acid persulfate
at 75 C
12
3.6
3.5
3.9
(Tap
4.0
Water)
Acid permanganate
13
4.5
4.2
14
4.4
15
2.0
2.0
Acid + heat
16
4.3
4.0
17
3.4
3.9
18
4.1
Acid permanganate
19
4.3
3.7
20
4.1
Acid, permanganate,
and persulfate
21
3.8
3.2
Acid permanganate
22
4.2
4.0
23
3.50
3.94
25
3.1
4.1
-------
11
TABLE 3 contd.
Average Recoveries of Increment; 4.2 g of Inorganic Mercury/liter, (Sample 2)
Concentration, pg/lIter
Laboratory
Distilled Water
Natural
Water
Description
of
EPA
Other
EPA
Other
Code
Method
Method
Method
Method
Other
Method
26 4.2
27 4.30 4.30
28 1)3.5 l)Acid permanganate (DOW)
2)4.1 2)Acid perinanganate +
KBr (Modified DOW)
29 4.3
30 4.3 4.1 Acid permanganate
31 1)4.1 1)4.6 l)Acid perinanganate (DOW)
2)4.1 2)4.3 2)Acid permanganate +
KBr (Modified DOW)
32 4.2 2.1 (Sewage Sample)
33 4.2
34 1)4.4 1)4.3 1)4.4 1)4.6 Acid permanganate
2)4.2 2)4.6 2)5.4 2)5.2 Acid permanganate
(Tap Water)
35 4.1
36 4.1 Stannous chloride
reduction
37 3.9
38 3.13 1)3.08 1)Acid permanganate
2)2.72 2)Acid only
3)3.69 3)Acid only
39 1.5 Acid permanganate
40 4.0 4.4 Acid, peroxide,
and permanganate
41
4.6
Acid permanganate
42
4.4
Modified DOW
-------
12
TABLE 4
Average Recoveries of Increment; 6.3 pg of Inorganic and Organic Mercury/liter
CSample 3)
Concentration, pg/liter
Laboratory
Distilled
Water
Natural
Water
Description
of
EPA
Other
EPA
Other
Code
Method
Method
Method
Method
Other Method
1
6.0
5.1
Ag wire deposition
2
6.7
Sulfuric, nitric,
and perchioric acid
ref lux
3
6.8
5.0
4
4
6
CdSpad
5
7.0
7.0
Cu wire deposition
6
6.20
3.57
5.96
Permanganate
7
6.0
4.6
6.6
Permanganate
8
4.8
4.8
Acid permanganate
(overnight)
9
7.0
7.0
10
5.8
4.1
4.4
Permanganate
11
1)5.40
2)6.20
5.40
l)Acid persulfate
at 55 C
2)Acid persulfate
at 75 C
12
5.6
6.4
6.3
(Tap
6.4
Water)
Acid permanganate
13
6.2
4.2
14
6.2
15
6.9
5.7
Acid + heat
16
6.4
5.4
17
5.8
5.6
18
6.3
Acid pernianganate
19
6.1
5.5
20
6.4
Acid, perinanganate,
and persulfate
21
6.0
3.7
Acid permanganate
22
6.5
5.8
23
5.44
5.13
25
6.0
6.8
-------
1-3
TABLE 4 contd.
Average Recoveries of Increment; 6.3 jig of Inorganic and Organic Mercury/liter
(Sample 3)
Concentration, jig/liter
Laboratory
Distilled
Water
Natural
Water
Description
of
EPA
Other
EPA
Other
Code
Method
Method
Method
Method
Other Method
26 6.1
27 6.50 6.50
28 1)3.0 l)Acid permanganate (DOW)
2)6.1 2)Acid permanganate +
KBr (Modified DOW)
29 8.0
30 5.6 6.2 Acid permanganate
31 1)5.8 1)4.4 1)5.6 1)Acid permanganate (DOW)
2)8.7 2)6.9 2)8.6 2)Acid, permanganate +
KBr (Modified DOW)
32 6.2 1.7 (Sewage Sample)
33 6.0
34 1)4.0 1)3.6 1)5.9 1)6.0 l)Acid permanganate
2)5.2 2)5.4 2)7.0 2)6.3 2)Acid permanganate
35
6.1
36
5.9
37
38
6.4
5.33
Acid
39
40
1.7
4.9
5.4
Acid
permanganate
41
42
6.2
5.9
3.8
Acid perinanganate
Modified DOW
-------
14
TABLE 5
Average Recoveries of Increment; 4.2 ig of Organic Mercury/liter, (Sample 4)
ConcentratIon, pg/liter
Laboratory
Code
Distilled
Water
Natural
Water
Description
of
Other Method
EPA
Method
Other
Method
EPA
Method
Other
Method
1
2
4.2
0.8
4.8
Ag wire deposition
Sulfuric, nitric and
perchioric acid ref lux
3
4
4.3
4
4.1
4
CdSpad
5
4.4
4.2
Cu wire deposition
6
4.30
0.60
3.90
Permanganate
7 4.3 3.0 4.4 Acid permanganate
8 2.3 3.2 Acid permanganate
(overnight)
9 5.5 5.5
10 4.2 0.7 3.5 Acid per,nanganate
11 1)3.70 3.20 1)Acid persulfate
at 55 C
2)4.65 2)Acid persulfate
at 75 C
12 4.2 4.2 4.3 4.6 Acid pennanganate
(Tap Water)
13 4.6 3.8
14 3.9
15 4,7 3.9 Acid and heat
16 4.2 2.9
17 4.0 4.0
18 4.2 Acid permanganate
19 4.1 3.6
20 4.4 Acid, permanganate,
and persulfate with
heating
21 4.1 0.6 Acid permanganate
22 4.3 4.0
23 3.50 3.31
25 4.3 4.5
-------
15
TABLE 5 contd.
Average Recoveries of Increment; 4.2 ig of Organic Mercury/liter, (Sample 4)
Concentration, pg/liter
Laboratory
Code
Distilled
Water
Natural
Water
Description
of
Other Method
EPA
Method
Other
Method
EPA
Method
Other
Method
26 4.3
27 4.10 4.10
28 1)0.7 1)Acid perinanganate (DOW)
2)3.7 2)Acid perinanganate +
KBr (Modified DOW)
29 5.0
30 4.1 4.4 Acid permnanganate
31 1)1.0 1)0.6 1)Acid permanganate (DOw)
2)4.6 2)4.1 2)Acid pernianganate +
KBr (Modified DOW)
32 3.8 0.2 (Sewage Sample)
33 4.2 Acid permnanganate
34 1)1.0 1)<1.0 1)4.4 1)3.7 1)Acid permnanganate
2)3.0 2) 2.9 2)4.6 2)4.7 2)Acid permnanganate
(Tap Water)
35
4.2
36
3.3
37
4.2
38
2.58
1)2.81
2)2.64
3)2.89
1)Acid permanganate
2)Acid
3)Acid
39
2.8
Acid permanganate
40
3.9
0.7
Acid, peroxide,
and permanganate
41
42
4.1
5.4
0.7
Acid permnanganate
Modified DOW
-------
16
TREATMENT OF DATA
Statistical Summary
Complete statistical summaries are given in pages 18 thru 33. Each sample
is discussed in turn with the data displayed by type of water sample and method
of analysis. A statistical evaluation is provided for each concentration with
each set of sample-test conditions.
In addition to the statistical measurements, all data are ranked in ascending
order and are presented in a histogram using ’iT cell divisions. Each X in the
histogram represents one analytical result for 1-15 values/cell. When more than
15 values occur per cell, only 15 X’s are printed and the number of values
included in the cell is printed at the base of the cell.
With the exception of accuracy, the statistical parameters: number of values,
true values, mean, range, variance, standard deviation, 95% confidence limit,
and relative deviation (coefficient of variation) are based on all of the data
received, without rejection. Accuracy is based on retained data, that is the
data remaining after rejection of outliers.
Rejection of Outliers
To determine the accuracy of each method, it was necessary to remove
those extreme values which had only a small chance of validity and which
would make a significant change in the reported accuracy. These values
were probably caused by gross instrumental, chemical or human error and
were rejected by applying the two-tail t—test to all values at a 99% probability
level; that is, with a 99 to 1 assurance that the data rejected were invalid
and should be rejected. Each rejected value is indicated in the statistical
sununaries by laboratory and analyst with a capital letter “R”, after the
value.
-------
17
A greater spread of data round the true value causes rejection of fewer
outliers because a larger standard deviation in the denominator of the t-test
reduces the calculated t value and fewer extreme values are rejected as outliers.
Contrariwise, with better accuracy and precision, the t—test is more powerful and
more outliers are rejected. However, the data rejected in either case are true
outliers and laboratories reporting such values should carefully review procedures
for the cause of inaccuracy. A glossary of the statistical terms used in this
study is included at the end of this report on pages 57 and 58.
-------
.18
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL 1 INCREMENT = 0.34 pg/liter, INORGANIC MERCURY
N 21 RANGE 0.32000 COEF. VAR. 0.19541
TRUE VAL. 0.34 VARIANCE 0.00443 SKEWNESS 0.45532
MEAN 0.34095 STD. 0EV. 0.06662 NO. OF CELLS 4
MEDIAN 0.33000 CONF. LIM. 0.13058 (95 PCI)
ACCURACY 0.28003 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
0.20 0.2000 2 XX
0.25 0.3066 12 XXXXXXXXXXXX
0.28 0.4133 6 XXXXXX
0.30 0.5199 1 X
0.30
0.30
0.31
0.31
0.32
0.33
0.33
0.35
O • 35
0.35
0.38
0.38
o • 40
0.40
0.40
0.40
0.52
R REJECTED DATA
-------
19
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 1
INCREMENT =
Q•3I ig/Hter, INORGANIC MERCURY
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
13
0.34
0.35076
0.34000
3.16737
RANGE 0.25000 COEF. VAR.
VARIANCE 0.00480 SKEWNESS
STD. 0EV. 0.06933 NO. OF CELLS
CONF. LIM. 0.13590 (95 PCT)
PCI RELATIVE ERROR, RETAINED DATA
0.19167
1.92644
3
DATA IN ASCENDING ORDER
0.30
0.30
0.30
0.30
0.30
0.32
0.34
0.35
0.35
0.36
0.39
0.40
0.55
MIDPOINT FREQ
HISTOGRAM
0. 3000
0. 4250
0. 5 500
10 XXXXXXXXXX
2 XX
1 X
R REJECTED DATA
-------
20
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL 1 INCREMENT = Q•3L 1 iig/liter, INORGANIC MERCURY
N 24 RANGE 0.80000 COEF. VAR. 0.40046
TRUE VAL. 0.34 VARIANCE 0.02679 SKEWNESS —0.01854
MEAN 0.40874 STD. DEY. 0.16368 NO. OF CELLS 4
MEDIAN 0.40000 CONF. LIM. 0.32082 (95 PCI)
ACCURACY 20.22049 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
0.00 0.0000 1 X
0.20 0.2666 12 XXXXXXXXXXXX
0.25 0.5333 10 XXXXXXXXXX
0.30 0.8000 1 X
0.30
0.30
0.31
0.32
0.33
0.34
0.36
0.40
0.40
0.42
0.45
0.48
0.50
0.50
0.50
0.56
0.59
O • 60
0.60
0.80R
R REJECTED DATA
-------
METHOD & PERFORMANCE EVALUATION, AQCL 21
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 1. INCREMENT 0.34 pg/lIter, INORGANIC MERCURY
N 8 RANGE 0.39999 COEF. VAR. 0.31540
TRUE VAL. 0.34 VARIANCE 0.01352 SKEWNESS 0.67914
MEAN 0.36874 SID. 0EV. 0.11630 NO. OF CELLS 2
MEDIAN 0.37500 CONF. LIM. 0.22795 (95 PCI)
ACCURACY 8.45584 PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
0.20 0.2000 4 XXXX
0.30 0.6000 4 XXXX
0.30
0.35
0.40
0.40
0.40
0.60
R REJECTED DATA
-------
22
METHOD & PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL 2
INCREMENT =
1 .2 Mg/liter, INORGANIC MERCURY
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
28
4.2
3. 93535
4.15000
—6.30110
RANGE 2.50000 COEF. VAR.
VARIANCE 0.28664 SKEWNESS
STD. DEV. 0.53539 NO. OF CELLS
CONF. LIM. 1.04937 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
0.13604
—1.92003
5
DATA IN ASCENDING ORDER
MIDPOINT FREO
H I STOGRAM
2.OR
3.1
3.1
3.4
3.5
3.6
3.8
3.8
3.8
3.8
3.9
4.0
4.1
4.1
4.2
4.2
4.2
4.2
4.2
4.2
4.3
4.3
4.3
4.3
4.3
4.4
4.5
4.5
2. 0000
2. 6250
3. 2500
3. 8750
4. 5000
I K
0
4 XXXX
9 XXXXXXXXX
14 XXXXXXXXXXXXXX
R REJECTED DATA
-------
METHOD & PERFORMANCE EVALUATION, AQCL 23
PRELIMINARY STUDY, ANALYSES FUR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 2 INCREMENT = LF.2 ig/Hter, INORGANIC MERCURY
N 18 RANGE 3.30000 COEF. VAR. 0.16576
TRUE VAL. 4.2 VARIANCE 0.43646 SKEWNESS —0.95187
MEAN 3.98555 STD. 0EV. 0.66065 NO. OF CELLS 4
MEDIAN 4.00000 CONF. LZM. 1.29487 (95 PCI)
ACCURACY —5.10587 PCI RELATIVE ERROR, RETAINED DATA
DATA TN ASCENDING ORDER MIDPOINT FR&I HISTOGRA
2. IR 2.1000 1 X
3.0 3.2000 2 XX
3.7 4.3000 14 XXXXXXXXXXXXXX
3.9 5.3999 1 X
3.9
3.9
3.9
3.9
4.0
4.0
4.1
4.1
4.2
4.3
4.4
4.4
4.5
5.4
R REJECTED DATA
-------
24
METHOD & PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM DiSTILLED WATER
AMPUL 2
INCREMENT =
ti,2 ig/Hter, INORGANIC MERCURY
N
TRUE VAL.
ME AN
MEDIAN
ACCURACY
29
4.2
3.86309
4.10000
-8.02143
RANGE 3.59999 COEF. VAR.
VARIANCE 0.58928 SKEWNESS
STD. 0EV. 0.76764 ND. OF CELLS
CONF. LIM. 1.50459 (95 PCT)
PCI RELATIVE ERROR, RETAINED DATA
0.19871
—1 .480 13
5
DATA IN ASCENDING ORDER
MIDPOINT FREQ
HIS TOGi AM
1.5R
1.5000
1
X
2.OR
2.4000
2
XX
2.7
3.2999
6
XXXXXX
3.0
4. 1999
19
XXXXXXXXXXXXXXX
3.2
5.0999
1
X
3.5
3.5
3.5
3.6
3.9
4.0
4.0
4.0
4. 1
4.1
4.1
4. 1
4. 1
4.1
4.2
4.3
4.3
4.3
4.4
4.4
4.5
4.6
4.6
5.1
R REJECTED DATA
-------
METHOD PERFORMANCE EVALUATION, AQCL 25
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 2 INCREMENT = 14.2 ug/liter, INORGANIC MERCURY
N 10 RANGE 2.20000 COEF. VAR. 0.13790
TRUE VAL. 4.2 VARIANCE 0.35822 SKEWNESS —0.89451
MEAN 4.33999 SID. DEV. 0.59851 NO. OF CELLS 3
MEDIAN 4.40000 CONF. LIM. 1.17309 (95 PCI)
ACCURACY 3.33327 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MiDPOINT FREQ HISTOGRAM
3.0 3.0000 1 X
4.0 4.1000 7 XXXXXXX
4.0 ¶.1999 2 XX
4.3
4.4
4.4
4.6
4.6
4 1 19
5.2
R c EJECTED DATA
-------
26
METHOD PERFORMANCE EVALUATION, AQCI
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL 3 INCREMENT = 6.3 iig/liter, ORGANIC + INORGANIC MERCURY
N 28 RANGE 4.00000 COEF. VAR. 0.11151
TRUE VAL. 6.3 VARIANCE 0.46351 SKEWNESS —0.25732
MEAN 6.10499 STD. DEV. 0.68081 NO. OF CELLS
MEDIAN 6.10000 CONF. LIM. 1.33440 (95 PCI)
ACCURACY —3.09527 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOG (AM
4.OR 4.0000 1 X
5.2 5.0000 2 XX
5.4 6.0000 21 XXXXXXXXXXXXXXX
5.6 7.0000 3 XXX
5.6 8.0000 1 X
5.8
5.8
5.9
6.0
6.0
6.0
6.0
6.0
6.1
6.1
6.1
6.2
6.2
6.2
6.2
6.4
6.4
6.5
6.5
6.8
6.9
7.0
8.0
R REJECTED DATA
-------
27
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 3
INCREMENT =
6.3 iig/liter, ORGANIC + INORGANIC MERCURY
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
18
6.3
5.61055
5. 8 5000
—10. 94361
DATA IN ASCENDING ORDER
1.7R
4.2
4.4
5.0
5.1
5.4
5.5
5.6
5.8
5.9
5.9
6.2
6.3
6.5
6.6
6.8
7.0
7.0
MIDPOINT FREQ
H ISTOGRAM
RANGE
5.30000
COEF.
VAR.
0.22676
VARIANCE
1.61873
SKEWNESS
—1.62836
Sb. DEV.
1.27229
NO. 0F
CELLS
4
CONF. LIM.
2.49369
(95
PCI)
PCT RELATIVE
ERROR, RETAINED
DATA
1. 7000
3.4666
5. 2333
7. 0000
1 X
1 X
9 XXXXXXXXX
7 XXXXXXX
R R±JECTED DATA
-------
28
METHOD i PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL 3 INCREMENT = 6.3 pg/lJter, ORGANIC + INORGANIC MERCURY
N 29 RANGE 7.00000 COEF. VAR. 0.27779
TRUE VAL. 6.3 VARIANCE 2.11163 SKEWNESS —0.14497
MEAN 5.23102 STO. 0EV. 1.45314 NO. OF CELLS 5
MEDIAN 5.40000 CONF. LIM. 2.84816 (95 PCI)
ACCURACY —16.96777 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
1.IR 1.7000 1 x
3.0 3.4499 7 XXXXXXX
3.5 5.1999 11 XXXXXXXX XXX
3.6 6.9499 9 XXXXXXXXX
3.7 8.6999 1 X
3.8
4.0
4.1
4.4
4.6
4.8
4.9
5.1
5.3
5.4
5.4
5.7
5.8
5.9
6.1
6.2
6.2
6.3
6.4
6.4
6.7
6.9
7.0
8.7
R REJECTED DATA
-------
29
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY iN WATER
ALL DATA, ALL lABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 3 INCREMENT 6.3 xg/1iter, ORGANIC + INORGANIC MERCURY
N 10 RANGE 3.79999 COEF. VAR. 0.17229
TRUE VAL. 6.3 VARIANCE 1.12277 SKEWNESS 1.16776
MEAN 6.14999 STD. DEV. 1.05961 NO. OF CELLS 3
MEDIAN 6.00000 CONF. LIM. 2.07683 (95 PCI)
ACCURACY —2.38097 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
4.8 4.8000 4 XXXX
5.4 6.6999 5 XXXXX
5.4 8.6000 1 X
5.6
6.0
6.0
6.3
6.4
7.0
8.6
R REJECTED DATA
-------
30
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
EPA METHOD
RECOVERY OF INCREMENT FROM DISTILLED WATER
AMPUL Ii
INCREMENT =
li.2 iig/liter, ORGANIC MERCURY
DATA IN ASCENDING ORDER
MIDPOINT FREQ
HISTOGRAM
1 • OR
2.5
3.0
3.3
3.5
3.8
4.0
4.1
4.1
4.1
4.1
4.1
4.2
4.2
4.2
4.2
4.2
4.2
4.2
4.3
4.3
4.3
4.3
4.3
4.3
4.6
4.7
5.0
5.5
1. 0000
2.1250
3. 2500
4.3750
5. 5000
xxxx
xxxxxxxxxxxxxxx
xx
N
29
RANGE
4.50000
COEF.
VAR.
0.19789
TRUE VAL.
4.2
VARIANCE
0.63394
SKEWNESS
—1.97527
MEAN
4.02344
STD. DEV.
0.79620
NO. OF
CELLS
5
MEDIAN
4.20000
CONF. LIM.
1.56056
(95
PCI)
ACCURACY
—4.20372
PCI RELATIVE
ERROR, RETAINED
DATA
1 X
1 X
4
21
2
R REJECTED DATA
-------
3]-
METHOD & PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORtES
EPA METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 4
INCREMENT =
4.2 ug/liter, ORGANIC MERCURY
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
18
4.2
3.83388
4.00000
—8.71698
RANGE 5.30000 COEF. VAR.
VARIANCE 1. 14055 SKEWNESS
SID. DEV. 1.06796 NO. OF CELLS
CONF. LIM. 2.09321 (95 PCI)
PCI RELATIVE ERRUR, RETAINED DATA
0. 2 7855
—2. 14235
4
DATA IN ASCENDING ORDER
0. 2R
2.9
3.3
3.5
3.6
3.8
3.9
3.9
4.0
6.0
4.1
4.1
4. 3
4.4
4.4
4.5
4.6
5.5
MIDPOINT FREQ
0. 2000
1.966b
3. 7333
5.5000
HISTOGRAM
1 X
0
16 XXXXXXXXXXXXXXX
1 X
R REJECTED DATA
-------
32
METHOD £ PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WATER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM DISTI LIED WATER
AMPUL 4 INCREMENT = 4.2 tg/1iter, ORGANIC MERCURY
N 28 RANGE 4.80000 COEF. VAR. 0.50731
TRUE VAL. 4.2 VARIANCE 2.35456 SKEWNESS —0.46796
MEAN 3.02464 STD. DEV. 1.53445 NO. OF CELLS 5
MEDIAN 3.35000 CONF. LIM. 3.00753 (95 PCT)
ACCURACY —27.98476 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
0.6 0.6000 7 XXXXXXX
0.6 1.7999 1 X
0.7 3.0000 6 XXXXXX
0.7 4.1999 13 XXXXXXxXXXXXx
0.7 5.3999 1 X
0.8
1.0
2.3
2.6
2.8
2.8
2.8
2.9
3.0
3.7
3.7
3.9
3.9
4.0
4.2
4.2
4.4
4.4
4.4
4.6
4.6
4.8
5.4
R REJECTED DATA
-------
33
METHOD PERFORMANCE EVALUATION, AQCL
PRELIMINARY STUDY, ANALYSES FOR MERCURY IN WAlER
ALL DATA, ALL LABORATORIES
OTHER METHOD
RECOVERY OF INCREMENT FROM NATURAL WATER
AMPUL 14 INCREMENT = 14.2 ig/1iter, ORGANIC MERCURY
N 10 RANGE 4.10000 COEF. VAR. 0.44969
TRUE VAL. 4.2 VARIANCE 2.20222 SKEWNESS —1.09921
MEAN 3.29999 STD. 0EV. 1.48398 NO. OF CELLS 3
MEDIAN 3.85000 CONF. LIM. 2.90861 (95 PCI)
ACCURACY —21.42860 PCI RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER MIDPOINT FREQ HISTOGRAM
0.6 0.6000 2 XX
0.7 2.6500 2 XX
3.2 4.6999 6 XXXXXX
3.2
3.7
4.0
4.1
4.2
4.6
4.7
R REJECTED DATA
-------
34
DISCUSSION
0.34 pg/liter, InorganIc Mercury
At the 0.34 pg inorganic mercury/liter level, the EPA Method was more
accurate and more precise in both distilled water and natural water. Accuracies
of 0.3% positive bias for EPA Method and 20% positIve bias for the Other Method
were found in distilled water and 3% and 8.5% positive bias In the natural waters
for these two methods, respectively. Differences in precision were similar
in distilled and natural waters with 20% and 40% relative deviation for the
EPA and Other Method in distilled water, and 20% and 32% relative deviation
in natural waters for the two methods, respectively.
4.2 iig/liter, Inorganic Mercury
At this 4.2 pg/liter level, the EPA and Other Method showed low bias
values of 6% and 8% in distilled water and -5% and +3% in natural waters.
The precision values were similar for the two methods with relative deviations
of 14% and 20% in distilled water for the EPA and Other Method, respectively,
and 17% and 14% in natural waters for the two methods, respectively.
6.3 pg/liter of Organic and Inorganic Mercury
This higher level of mercury made up of equal amounts of organic and
inorganic mercury showed negative bias of 3% and 17% in distilled water for
the EPA and Other Methods, respectively. However, in natural water, the
negative bias values were reversed with 11% for the EPA Method and 2% for
the Other Method.
Likewise, precision as relative deviation was 11% for the EPA Method
and 28% for the Other Method in distilled water. In natural waters the
EPA Method showed a 23% deviation while the Other Method showed a 17% deviation.
4.2 pg/liter, Organic Mercury
The presence of only organic mercury as phenyl mercuric acetate in the
fourth sample provided a rigorous test of the oxidation step in the methods
-------
35
since no mercury was measurable without conversion. In distilled water, the
results showed a negative bias of 4% for the EPA Method and 28% for the Other
Method, and relative deviations of 20% and 51% for the EPA and Other Methods,
respectively.
In the natural water samples the results showed negative bias of 9% for
the EPA Method and 21% for the Other Method, and relative deviations of 28%
and 45% for the EPA and Other Methods, respectively.
This study was not intended to demonstrate the superiority of the EPA
Method for total mercury over all other methods. However, the complete
variety of methods applied in this study has forced an evaluation only of
the EPA Method and other methods as a group. The reader is urged to compare
recoveries by each of the methods in Tables 2, 3, 4, and 5. He will note
that almost all methods used did recover inorganic mercury rather completely.
However, there was a divergence of results from the Other Method, with
samples containing organic mercury, Into those which could and those which could
not recover the added levels. In Table 5 reporting recoveries of 4.2 pg/liter
of organic mercury, the copper wire deposition method of laboratory 5, the
acid permanganate of laboratory 12 and the acid permanganate and persulfate
of laboratory 20 obtained very good recovery from distilled water. Laboratories
5 and 12 repeated their good recovery of organic mercury in natural waters and
were joined by laboratories 31 and 34.
Table 5 shows examples of poor recovery from distilled water samples where
laboratories 1, 6, 10, 21, and 42 had minimal recovery with their own method
and very good recovery with the EPA Method.
-------
36
CONCLUS rONS
This preliminary study on measurement of organic and/or inorganic mercury
in distilled and natural waters showed that problems of level of mercury and
sample substrate (distilled vs. natural water) were not as significant as the
form of the mercury tested.
At the 0.34 - 4.2 pg/liter level of Inorganic mercury with one exception
the EPA and Other Method showed similar levels of accuracy and precision for
both the distilled and natural water samples containing 0.34 or 4.2 pg of
inorganic mercury/liter. As expected, when the mercury level was less than
1 pg/liter, the results showed a greater imprecision than at the 4.2 pg/liter
level.
At the 6.3 pg/liter level of organic/inorganic mercury containing equal
amounts of organic and inorganic mercury, the two methods reversed accuracies
and precision in distilled and natural waters. The EPA Method had better
accuracy and precision with the distilled water sample, -3% and 11% respectively
than in the natural waters where it had a -11% bias and 23% deviation respectively.
Conversely, the Other Method showed a negative 17% bias and 28% deviation in
distilled water and a negative 2% bias and 17% deviation in the natural waters.
Despite the similar accuracy and precision values shown by the two methods
with Sample 3 which contained equal amounts of organic and inorganic mercury,
the methods differed significantly with Sample 4 which contained only organic
mercury. The EPA Method had a -4% bias in distilled water and a 9% bias in
natural water while the Other Method had a -28% bias in distilled water and
a -21% in natural water. Likewise, the relative deviation of the EPA Method
was 20 and 28% in the distilled and natural water while the Other Method had
a relative deviation of 51 and 45% respectively for the distilled and natural
water samples.
-------
37
The performances of the EPA Method and Other Method in recovery of organic
and inorganic mercury from distIlled and natural water samples have shown that:
1. The EPA Method was more precise with three of the four inorganic
mercury samples and three of the four of the samples containing
organic mercury.
2. Both methods showed a limited and ‘variable bias with the two inorganic
mercury samples and a larger and consistently negative bias with the
samples containing organic mercury.
If water samples contain mercury in organic form, a method involving a
vigorous digestion step such as that in the EPA Method should be used to
insure good recovery.
-------
39
APPENDIX A
EPA Method for Mercury in Water, April 1972
(Note: Heating step 3 8.1, was not included
in EPA Method evaluated in this study)
-------
April, 1972
41
MERCURY IN WATER
(Cold Vanor Techniaue)
STORET NO.
Total: 71900
Dissolved: 71890
1. Scope and Application
1.1 This method is applicable to surface waters, saline waters, waste-
waters, effluents, and domestic sewage.
1.2 In addition to inorganic forms of mercury, organic mercurials may
also be present in an effluent or surface water sample. These
organomercury compounds will not respond to the flanieless atomic
absorption technique unless they are first broken down and converted
to mercuric ions. Potassium permanganate oxidizes many of these
compounds but recent studies have shown that a number of organic
mercurials, including phenyl m rcuric acetate and methyl mercuric
chloride, are only partially oxidized by this reagent. Potassium
persulfate has been found to give approximately 100% recovery when
used as the oxidant with these compounds. Therefore, a persulfate
oxidation step following the addition of the permanganate has been
included to insure that organomercury compounds, if present, will be
oxidized to the mercuric ion before measurement. A heat step is
required for methyl mercuric chloride when present in or spiked to a
natural system. For distilled water the heat step is not necessary.
1.3 The range of the method may be varied through instrument and/or
recorder expansion. Using a 100 ml sample, a detection limit of 0.2 i tg
Hg/l can be achieved; concentrations below this level should be reported
as <0.2 (see Appendix 11.2).
-------
42
2. Summary of Method
2.1 The fiameless AA procedure is a physical method based on the absorption
of radiation at 253.7 nm by mercury vapor. The mercury is
reduced to the elemental state and aerated from solution in a closed
system. The mercury vapor passes through a cell positioned in the
light path of an atomic absorption spectrophotometer. Absorbance
(peak height) is measured as a function of mercury concentration and
recorded in the usual manner.
3. Sample Handling and Preservation
3.1 Until more conclusive data are obtained, samples should be preserved
by acidification with nitric acid to a pH of 2 or lower immediately
at the time of coliection . If only dissolved mercury is to be
determined, the sample should be filtered before the acid is added.
For total mercury the filtration is omitted.
4. Interference
4.1 Possible interference from sulfide is eliminated by the addition of
potassium permanganate. Concentrations as high as 20 mg/i of sulfide
as sodium sulfide do not interfere with the recovery of added in-
organic mercury from distilled water.
4.2 Copper has also been reported to interfere; however, copper concen-
trations as high as 10 mg/i had no effect on the recovery of mercury
from spiked samples.
4.3 Sea waters, brines and industrial effluents high in chlorides require
additional permanganate (as much as 25 ml). During the oxidation
step chlorides are converted to free chlorine which will also absorb
radiation at 253 nm. Care must be taken to assure that free chlorine
is absent before the mercury is reduced and swept into the cell. This
may be accomplished by using an excess of hydroxylainine sulfate
-------
43
reagent (25 ml). In addition, the dead air space in the BOD bottle
must be purged before the addition of stannous sulfate. Both in-
organic and organic mercury spikes have been quantitatively recovered
from sea water using this technique.
4.4 Interference from certain volatile organic materials which will
absorb at this wavelength is also possible. A preliminary run with-
out reagents should determine if this type of interference is present
(see Appendix 11.1).
5. Apparatus
5.1 Atomic Absorption Spectrophotometer*: Any atomic absorption unit
having an open sample presentation area in which to mount the
absorption cell is suitable. Instrument settings recommended by the
particular manufacturer should be followed.
5.2 Mercury Hollow Cathode Lamp: Westinghouse WL - 22847, argon filled,
or equivalent.
5.3 Recorder: Any multi-range variable speed recorder that is compatible
with the UV detection system is suitable.
5.4 Absorption Cell: Standard spectrophotometer cells 10 cm long, having
quartz end windows may be used. Suitable cells may be constructed from
plexiglass tubing, 1” O.D. x 4-1/2”. The ends are ground perpendicular
to the longitudinal axis and quartz windows (1” diameter x 1/16”
thickness) are cemented in place. Gas inlet and outlet ports (also
of plexiglass but 1/4” O.D.) are attached approximately 1/2” from
each end. The cell is strapped to a burner for support and aligned
in the light beam by use of two 2” by 2” cards. One inch diameter
*Instrwnents designed specifically for the measurement of mercury using
the cold vapor technique are commercially available and may be substituted
for the atomic absorption spectrophotometer.
-------
44
holes are cut in the middle of each card; the cards are then placed
over each end of the cell. The cell is then positioned and adjusted
vertically and horizontally to give the maximum transmittance.
5.5 Air Pump: Any peristaltic pump capable of delivering 1 liter of air
per minute may be used. A Masterflex pump with electronic speed
control has been found to be satisfactory.
5.6 Flowmeter: Capable of measuring an air flow of 1 liter per minute.
5.7 Aeration Tubing: A straight glass frit having a coarse porosity.
Tygon tubing is used for passage of the mercury vapor from the sample
bottle to the absorption cell and return.
5.8 Drying Tube: 6” x 3/4” diameter tube containing 20 grams of magnesium
perchlorate (see Note 1.) The apparatus is assembled as shown in the
accompanying diagram.
NOTE 1: In place of the magnesium perchlorate drying tube, a small
reading lamp with 60W bulb may be used to prevent condensation of
moisture inside the cell. The lamp is positioned to shine on the
absorption cell maintaining the air temperature in the cell about
10°C above ambient.
6. Reagents
6.1 Sulfuric Acid, Conc: Reagent grade
6.1.1 Sulfuric acid, 1.0 N: Dilute 28.0 ml of conc. sulfuric acid
to 1.0 liter.
6.1.2 Sulfuric acid, 0.5 N: Dilute 14.0 ml of conc. sulfuric acid
to 1.0 liter.
6.2 Nitric Acid, Conc: Reagent grade of low mercury content.
NOTE 2: If a high reagent blank is obtained, it may be necessary to
distill the nitric acid.
-------
45
6.3 Stannous Sulfate: Add 25 g stannous sulfate to 250 ml of 0.5 N
sulfuric acid. This mixture is a suspension and should be stirred
continously during use.
NOTE 3: Stannous chloride and hydroxylamine hydrochloride may also
be used.
6.4 Sodium Chloride-Hydroxylamine Sulfate Solution: Dissolve 12 grams
of sodium chloride and 12 grams of hydroxylamine sulfate in distilled
water and dilute to 100.0 ml.
6.5 Potassium Permanganate: 5% solution, w/v. Dissolve 5 grams of
potassium permanganate in 100 ml of distilled water.
6.6 Potassium Persulfate: 5% solution, w/v. Dissolve S grams of potassium
persulfate in 100 ml of distilled water.
6.7 Stock Mercury Solution: Dissolve 0.1354 grams of mercuric chloride in
75 ml of distilled water. Add 10 ml of concentrated nitric acid and
adjust the volume to 100.0 ml. 1 ml = 1 mg Hg.
6.8 Working Mercury Solution: Make successive dilutions of the stock
mercury solution to obtain a working standard containing 0.1 ig per
ml. This working standard and the dilutions of the stock mercury
solution should be prepared fresh daily. Acidity of the working
standard should be maintained at 0.15% nitric acid. This acid should
be added to the flask as needed before the addition of the aliquot.
7. Calibration
7.1 Transfer 0, 0.5, 1.0, 2.0, 5.0 and 10.0 ml aliquots of the working
mercury solution containing 0 to 1.0 ig of mercury to a series of
300 ml BOD bottles. Add enough distilled water to each bottle to make
a total volume of 100 ml. Mix thoroughly and add 5 ml of concentrated
sulfuric acid and 2.5 ml of nitric acid to each bottle.
-------
46
Add 15 ml of KMnO 4 solution to each bottle and allow to stand at
least 15 minutes. Add 8 ml of potassium persulfate to each bottle
and heat for 2 hours in a water bath maintained at 95°C. Cool
and add 6 ml of sodium chloride - hydroxylainine sulfate solution
to reduce the excess permanganate. When the solution has been
decolorized wait 30 seconds, add 5 ml of the stannous sulfate
solution and immediately attach the bottle to the aeration
apparatus forming a closed system. At this point the sample is
allowed to stand quietly without manual agitation. The circulating
pump, which has previously been adjusted to a rate of 1 liter per
minute, is allowed to run continuously.
NOTE 4: An open system where the mercury vapor is passed through
the absorption cell only once may be used instead of the closed
system.
The absorbance will increase and reach maximum within 30 seconds. As
soon as the recorder pen levels off, approximately 1 minute, open the
bypass valve and continue the aeration until the absorbance returns to
its minimum value (see Note 5). Close the bypass valve, remove the
stopper and frit from the BOD bottle and continue the aeration.
Proceed with the standards and construct a standard curve by plotting
peak height versus micrograms of mercury.
NOTE 5: Because of the toxic nature of mercury vapor precaution must
be taken to avoid its inhalation. Therefore, a bypass has been
included in the system to either vent the mercury vapor into an exhaust
hood or pass the vapor through some absorbing media, such as:
a) equal volumes of 0.1 N KMnO 4 and 10% H 2 S0 4
b) 0.25% iodine in a 3% KI solution
-------
47
A specially treated charcoal that will adsorb mercury vapor is
also available from Barixebey and Cheney, E. 8th Ave. and North
Cassidy St., Columbus, Ohio 43219, Cat. #580-13 or #580-22.
8. Procedure
8.1 Transfer 100 ml or an aliquot diluted to 100 ml containing not more
than l• 0 i.’g of mercury to a 300 ml BOD bottle. Add 5 ml of sulfuric
acid and 2.5 ml of nitric acid mixing after each addition. Add 15
ml of potassium permanganate solution to each sample bottle. For
sewage samples additional perinanganate may be required. Shake and
add additional portions of potassium permanganate solution if
necessary until the purple color persists for at least 15 minutes.
Add 8 ml of potassium persulfate to each bottle and heat for 2 hours
in a water bath at 95°C. Cool and add 6 ml of sodium chloride -
hydroxylamine sulfate to reduce the excess permangaixate. After a
delay of at least 30 seconds add 5 ml of stannous sulfate and
immediately attach the bottle to the aeration apparatus. Continue as
described under Calibration.
9. Calculation
9.1 Determine the peak height of the unknown from the chart and read the
mercury value from the standard curve.
9.2 Calculate the mercury concentration in the sample by the formula:
igHg/l = g Hg x 1000
in aliquot volume of aliquot
9.3 Report mercury concentrations as follows:
Below 0.2 ugh, <0.2; between 1 and 10 ugh, one decimal; above 10 g/l,
whole numbers.
10. Precision and Accuracy
10.1 Using an Ohio River composite sample with a background mercury con-
centration of 0.35 ug/l, spiked with concentrations of 1, 3 and 4
-------
48
pg/i, the standard deviations were ±0.14, ±0.10 and ±0.08,
respectively. Standard deviation at the 0.35 level was ±0.16.
Percent recoveries at the three levels were 89, 87, and 87%,
respectively.
11. Appendix
11.1 While the possibility of absorption from certain organic substances
actually being present in the sample does exist, the AQC Laboratory
has not encountered such samples. This is mentioned only to caution
the analyst of the possibility. A simple correction that may be used
is as follows:
If an interference has been found to be present (4.4), the sample
should be analyzed both by using the regular procedure and again under
oxidizing conditions only, that is without the reducing reagents.
The true mercury value can then be obtained by subtracting the two
values.
11.2 If additional sensitivity is required, a 200 nil sample with recorder
expansion may be used provided the instrument does not produce
undue noise. Using a Coleman MAS—SO with a drying tube of magnesium
perchiorate and a variable recorder, 2 niv was set to read full scale.
With these conditions, and distilled water solutions of mercuric
chloride at concentrations of 0.15, 0.10, 0.05 and 0.25 ug/l the
standard deviations were ±0.027, ±0.006, ±0.01 and ±0.004. Percent
recoveries at these levels were 107, 83, 84 and 96%, respectively.
-------
49
References
1. Wallace, R.A., Fulkerson, W., Shults, W.D., and Lyon, W.S., “Mercury
in the Environment - The Human Element”, Oak Ridge National Laboratory,
O1 NL - NSF - EP - 1. January, 1971, Page 31.
2. Hatch, W.R.,, and Ott, W.L., “Determination of Sub-Microgram Quantities
of Mercury by Atomic Absorption Spectrophotometry”, Anal. Chem. 40,
2085 (1968).
3. Brandenberger, H. and Bader, H., “The Determination of Nanogram Levels of
Mercury in Solution by a Flameless Atomic Absorption Technique”, Atomic
Absorption Newsletter, 6, 101 (1967).
4. Brandenberger, H. and
Atomic Absorption II.
7, 53 (1968).
Bader, H., “The Determination of Mercury by Flameless
A Static Vapor Method”, Atomic Absorption Newsletter,
SAMPLE SOLUTION 1P4 BOO BOTTLE
FIGURE I . APPARATUS FOR FLAMELESS MERCURY
DETERMINATION
SCRUBBER
CONTAINING
A MERCURY
ABSORBING MEDIA
ABSORPTION CELL
-------
51
APPENDIX B
Details on the Other Method (non-EPA)
Used for Mercury Analyses in this Study
-------
53
Details of the Other Method (non— EPA) Used for Mercury Analysis in this Study
Laboratory Digestion Steps Used in
Designation Other Method, Details, and Reference
1 SIlver- flre Method:
Fl shman, J. 3., Anal. Chem.,
42, 1462 (1970).
2 To sainple,add 25 ml 18 N sulfuric acid, 20 ml 7 N
nitric acid and 1 ml sodium molybdate solution
(2% W/V). Heat for one hour. Cool, add 20 ml (1 + 1)
nitric acid-perchioric acid. Boil vigorously till
white fumes appear. Heat for another 10 minutes,
cool, add 10 ml water, boil again for 10 minutes, cool
bring to volume.
Source: Massachusetts Division of Fisheries and Game
Field Headquarters
Westborough, Massachusetts 01581
4 Mercury is precipitated from the sample by cadmium
sulfide followed by emission spectrographic arc
excitation of the mercury in the precipitate.
Powdered copper metal is added to samples suspected
of containing organo-mercury compounds to reduce the
mercury to the metallic form.
Source: E. I. du Pont de Nemours Co. (Inc.)
Savannah River Plant
Aiken, South Carolina 29801
5 Electrodeposition of mercury onto copper wire coil
as described in: Brandenberger and Bader, Atomic
Absorption Newsletter, 6, 101, (1967).
6 KMnO 4 oxidant for five minutes. No other reference.
7 EPA method without persulfate.
8 To a 100 nil sample, add 10 ml 18 N sulfuric acid,
5 ml 7 N nitric, 2 ml of 4% potassium permanganate.
Let stand overnight.
10 To a 100 ml sample, add 10 ml 18 N sulfuric acid,
5 ml 7 N nitric acid and 5 ml of 4% potassium
permanganate. Mix; then neutralize.
-------
54
Details on the Other Method (non-EPA) Used for Mercury Analysis in this Study contd.
Laboratory Digestion Steps Used in
Designation Other Method, Details, and Reference -
11 To a 100 ml sample, add 2-1/2 ml acid-persulfate
(2:1 V/W). Digest on shaking water bath at 55 C
for one hour. Cool to R. T., add 5% W/V potassium
permanganate in 1 ml increments until color holds
for 10 minutes.
12 Nitric acid-permanganate digestion described in:
Roif, A. C., Russell, F. R., and Wilkinson, N. T.,
The Anal Bt, 80, 523, (1955).
15 To a 100 ml sample add 25 ml nitric acid, heat
gently for 30 minutes
Source: Texas A M University
Agricultural Analytical Services
College Station, Texas 77843
18 Samples digested at 50 - 60 C with sulfuric acid
and oxidize with perinanganate.
Source: Uthe, J. F., Armstrong, F. A. J., and
Stainton, M. P., J. of the Fisheries
Research Board of Ccxiada , 27, 805, (1970).
20 EPA procedure modified: sulfuric acid: nitric acid
(20: 10 ml), 10 ml potassium persulfate, 15 ml
potassium permanganate, heat for four hours at 50 C.
Source: Surveillance and Analysis Division, EPA
Technical Support Branch
620 Central Avenue
Alameda, California 94501
21 Sulfuric acid-potas ium permanganate digestion.
No further reference.
28 To a 50 ml sample add 1 ml of 1:4 sulfuric acid
and 1 nil of 4% potassium permanganate. Heat to
boiling for few seconds, allow to cool.
Source: Dow Method CAS-AM-70.l3, June 22, 1970,
revised.
Chlorine Institute Pamphlet M1R-l04,
No. 3, (1970).
-------
55
Details on the Other Method (non- EPA) Used for Mercury Analysis in this Study contd.
Laboratory Digestion Steps Used in
Designation Other Method, Details, and Reference
30 Add 10 ml 7 N nitric acid, 25 ml 18 N sulfuric
acid and cool. Add 5% potassIum permanganate
until pink color persists.
Source: Hatch, 1V . R. and Ott, L L., Anal jttcal
Chin tr j, 40, 2085 (1968).
31 a) Dow procedure as In 28.
B) Dow procedure modIfied by addition of one ml
of 10% potassIum bromide after the acid and
permanganate. Sample allowed to set for one
minute before boiling.
Source: Diamond Shamrock Chemical Company
Sheffield, Alabama 35660
34 To a 100 ml sample add 10 ml of 18 N sulfuric acid,
5 ml of 7 N nitric acid and 5 ml of 4% W/V potassium
permanganate. Modification from Hatch and Ott,
reference 30.
36 To a 5.0 ml sample add 0.5 ml of 5% stannous chloride
in 1 N hydrochloric acid.
Source: Applications Laboratory
Instrumentation Laboratory, Inc.
Lexington, Massachusetts 02173
38 a) To a 10 ml diluted sample add 10 ml sulfuric acid
and 2 ml 5% potassium permanganate. Digest 15 minutes
at 90 C then cool.
b) To a 5 ml, diluted sample add 10 nil nitric acid,
digest for 1-1/2 hours at 75 C and 1-1/2 hours at
90 C, cool.
Measurement using a Beckman DB-G spectrophotonieter
and a 62.5 cm cell.
Source: University of Illinois
Agronomy Department
Urbana-Champaign Campus
Urbana, Illinois 61801
-------
56
Details on the Other Method (non-EPA) Used for Mercury Analysis in this Study contd.
Laboratory Digestion Steps Used in
Designation Other Method, Details, and Reference
39 To a one-liter sample, add 10 ml of 1 + 1 sulfuric acid
and 5 ml of aqueous 2% W/V potassium permanganate.
Mix well and let stand 20 hours at room temperature.
Source: Omang, S., Anal. Chim Acta, 53, 415 (1971).
40 Add 20-30 ml sulfuric acid to sample and ref lux on a
steam bath with agitation for 30 minutes. Cool to room
temperature. Add 50% hydrogen peroxide in 1/2 - 1 ml
portions with vigorous mixing. After decomposition is
complete, add 5 ml portions of permanganate slowly until
color persists for 15 minutes.
Source: Kimura, Y. and Miller, V., Anal. Chim Acta,
27, 325—331 (1962).
41 To a 100 ml sample add potassium permanganate until
a pink color is maintained. 20 ml of 2:1 sulfuric and
nitric acid added and sample digested for 24 hours.
Source: Water Quality Control Research Laboratory
Stanford University
Civil Engineering Department
Palo Alto, California 94305
42 Modified DOW procedure.
To a 50 ml sample add 1 ml 4% KMriO 4 , boil 2 seconds
Source: Westvaco Corporation
Westvaco Research Center
Laurel, Maryland 20810
-------
57
GLOSSARY OF TERMS
The statistical measurements used in method study reports of the Analytical
Quality Control Laboratory, are defined as follows:
Accuracy as % Relative Error (Bias). The signed difference between mean
value and the true value, expressed as a percent of the true value.
x -r
R. E., % = true x 100
true
Confidence Limit (95%). The range of values within which a single analysis
will be included, 95% of the time.
95% C. L. = X ± l.96a
Mean. The arithmetic mean of reported values, the average.
n
Median. Middle value of all data ranked in ascending order. If there
are two middle values, the mean of these values.
n. The number of sets of values or analysts reported in a study.
Range. The difference between the lowest and highest values reported
for a sample.
Relative Deviation (Coefficient of Variation). The_ratio of the standard
deviation, a, of a set of numbers to their mean, X, expressed as percent.
It is an attempt to relate deviation (precision) of a set of data to the
size of the numbers so that deviations between levels of values can be
compared.
R. D. = l0O_2
x
Skewneaa (ki. A pure number, positive or negative, which indicates the
lack of symmetry in a distribution. For example, k is positive if the
distribution tails to the right and negative if the distribution tails
to the left.
E (X.
3
na
-------
58
Standard Deviation (a), Cs). The most widely used measure of dispersion of a
set of data. It is equal to the square root of the variance and indicates
the deviation of 68% of the values around the mean, and l.96a, the devia-
tion of 95% of the values around the mean, a, the standard deviation,
is the measure of the deviation of the universe. However, in most experi-
mental work with limited sampling and in this study only an estimated
standard deviation (s) is measurable. This differs in calculation in
that n-i rather than n is used as the denominator. In this study and in
further studies, a and 2 not a and a 2 will be used to measure deviation
of the data. They will be referred to as the standard deviation and
variance respectively.
a = Ex2
8 =
n
z
x
X) 2
t—teat. The difference in analyzed and true value expressed as ratio over
the 8tandard deviation. The value obtained is compared with critical
values in a table. If the calculated t-value exceeds the theoretical
t-value, the analyzed value is probably not from the same population as
the rest of the data and can be rejected.
- true value
t-value = Stand d Deviation (a)
Variance (a 2 ), ( 2), The average of the squares of the deviations of a group
of numbers from their average, L
E X - (Z X) 2
2
a
2
B
n
E
X (
X) 2
uSw ERJ IsT, INTu r,ncLIm_ 759-551/1065
------- |