ENVIRONMENTAL PROTECTION AGENCY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
Analytical Quality Control Laboratory
Cincinnati, Ohio
-------
METHOD RESEARCH STUDY 3, DEMAND ANALYSES
An Evaluation of Analytical Methods for Water and Wastewater
1971
J. A. Winter
ENVIRONMENTAL PROTECTION AGENCY
National Environmental Research Center
Analytical Quality Control Laboratory
Cincinnati, Ohio 45268
-------
Table of Contents
Page
ACKNOWLEDGEMENTS v
PARTICIPATING LABORATORIES vi
SUMMARY viii
INTRODUCTION 1
DESCRIPTION OF THE STUDY 2
Test Design 2
Preparation of Samples and Reporting of Results 2
True Values 3
Analytical Methods 4
Glossary of Terms 5
RESULTS 7
Raw Data ^. 7
TREATMENT OF DATA 7
Statistical Summary 7
Rejection of Outliers 8
DISCUSSION 9
Chemical Oxygen Demand (COD) 9
Biochemical Oxygen Demand (BOD) 12
Total Organic Carbon (TOC) 15
CONCLUSIONS 20
REFERENCES 23
APPENDIX 25
-------
ACKNOWLEDGEMENTS
The author acknowledges the technical assistance and guidance
of Mr. Elmo C. Julian, Analytical Quality Control Laboratory, National
Environmental Research Center, Cincinnati, in the design, development
and operation of the statistical computer programs used in this study.
v
-------
PARTICIPATING LABORATORIES
Eighty-six analysts in 58 laboratories took part in Method Research Study 3,
Demand Analyses. Of the 58 laboratories, 40 were non-EPA. The participating
laboratories were:
EPA Laboratories
Advanced Waste Treatment Plant
D. C. Pollution Control Plant
Washington, D. C.
Advanced Waste Treatment
Research Laboratory
Cincinnati, Ohio
Alaska Water Laboratory
College, Alaska
Baton Rouge Field Facility
Baton Rouge, Louisiana
California-Nevada Basins Sub-Region
Alameda, California
Cincinnati Field Investigations Center
Cincinnati, Ohio
Edison Water Quality Laboratory
Edison, New Jersey
Illinois District Office
Chicago, Illinois
Indiana District Office
Evansville, Indiana
Missouri Basin Laboratory
Kansas City, Missouri
National Environmental Research Center
Corvallis, Oregon
New England Basin Office
Needham Heights, Massachusetts
Pomona Pilot Plant
Pomona Water Reclamation Plant
Pomona, California
Solid Waste Research Division
Cincinnati, Ohio
R. S. Kerr Water Research Center
Research
Ada, Oklahoma
R. S. Kerr Water Research Center
Technical Programs
Ada, Oklahoma
Southeast Water Laboratory
Athens, Georgia
Wheeling Field Station
Wheeling, West Virginia
Non-EPA Laboratories
California State Dept. of Public Health
Sanitation and Radiation Laboratory
Berkeley, California
California State Dept. of Public Health
Southern California Laboratory
Los Angeles, California
California Dept. of Water Resources
Sacramento, California
City of Chicago Dept. of Water § Sewers
Water Purification Laboratory
Chicago, Illinois
Clark County Sanitation Districts
Las Vegas, Nevada
Colorado Springs Waste Water Treatment
Colorado Springs, Colorado
Commonwealth of Massachusetts
Lawrence Experimental Station
Lawrence, Massachusetts
E. I. du Pont de Nemours
Belle, West Virginia
Co., Inc.
Environmental Quality Control Commission
Portland, Oregon
FMC Corporation
Inorganics Division
South Charleston, West Virginia
vi.
-------
FMC Corporation
Organic Chemicals Division
Nitro, West Virginia
FMC Corporation
Viscose Division
Nitro, West Virginia
Houston City Health Department
Houston, Texas
Hyperion Treatment Plant
Playa del Rey, California
Idaho State Health Department
Boise, Idaho
Indiana State Board of Health
Indianapolis, Indiana
International Paper Company
Springhill, Louisiana
Interstate Sanitation Commission
New York, N. Y.
Kansas City Water Department
Kansas City, Missouri
Los Angeles Dept. of Water
and Power
Los Angeles, California
Metropolitan Denver Sewage Disposal
Commerce City, Colorado
Maryland State Dept. of
Water Resources
Annapolis, Maryland
Metropolitan Sanitary District
of Greater Chicago
Cicero, Illinois
Miami Conservancy District
Dayton, Ohio
Middlesex County Sewerage Authority
Sayreville, New Jersey
Monsanto Chemical Company
Organic Chemicals Division
Nitro, West Virginia
North Carolina Dept. of Water
and Air Resources
Raleigh, North Carolina
Ohio Dept. of Health
Columbus, Ohio
Pennsylvania Dept. of Health
and Welfare
Harrisburg, Pennsylvania
Santee County Water District
Santee, California
Sewerage and Water Board
New Orleans, Louisiana
South Carolina Pollution Control
Authority
Columbia, South Carolina
South Tahoe Public Utilities District
South Lake Tahoe, California
Texas A £ M University
Environmental Engineering Division
College Station, Texas
Union Carbide Corporation
Olefins Division
South Charleston, West Virginia
Union Carbide Corporation
Organics Division
Charleston, West Virginia
Water § Air Resources Commission
Dover, Delaware
Water Pollution Control Commission
Olympia, Washington
West Virginia Dept. of Natural
Resources
Charleston, West Virginia
Virginia State Water Control
Board
Richmond, Virginia
-------
METHOD RESEARCH STUDY 3, DEMAND ANALYSES
An Evaluation of Analytical Methods for Water and Wastewater
SUMMARY
The Analytical Quality Control Laboratory of National Environmental
Research Center, Environmental Protection Agency, conducted interlaboratory
research studies on selected chemical methods of analysis for chemical
oxygen demand(COD), total organic carbon (TOC) and for biochemical oxygen
demand (BOD). Sample concentrates were prepared at low (natural water)
levels and at higher (municipal waste) levels for each constituent.
Analysts added an aliquot of each concentrate to distilled water for COD
and TOC analyses and to a natural water of their choice for BOD analyses.
Single analyses were made on the distilled and natural water samples
with and without added increments. Recoveries were compared. The bias
of the method and, where possible, the interference of natural water
samples and the relative precision of each analyst and laboratory were
determined. A statistical summary of this data on page will, shows the
precision and accuracy values which may be expected in routine work.
in/z/z.
-------
A Statistical Summary of Method Research Study 3, Demand Analyses
Chemical Oxygen Demand Analyses of Distilled Water Samples
True Value, mg COD/liter 12.3 270
Mean of Recoveries by
Difference, mg COD/liter 12.34 257.4
Accuracy as % Relative
Error (Bias) 0.3 -4.7
Standard Deviation,
mg COD/liter 4.15 17.76
Relative Deviation, % 33.6 6.9
Range, mg COD/liter 25.3 104
Biochemical Oxygen Demand Analyses of
Distilled/Seeded and Natural Water Samples
True Value, mg BOD/liter 2.2 194
Mean of Recoveries by
Difference, mg BOD/liter 2.12 175
Accuracy as % Relative
Error (Bias) -3.7 -9.8
Standard Deviation
mg BOD/liter 0.7 26.
Relative Deviation, % 33.2 15.0
Range, mg BOD/liter 5.5 118
Total Organic Carbon Analyses of Distilled Water Samples
True Value, mg TOC/liter 4.9 107
Mean of Recoveries by
Difference, mg TOC/liter 5.65 108.1
Accuracy as % Relative
Error (Bias) 15.3 1.0
Standard Deviation,
mg TOC/liter 1.89 6.
Relative Deviation, % 33.5 5.6
Range, mg TOC/liter 10.2 33
-------
METHOD RESEARCH STUDY 3, DEMAND ANALYSES
An Evaluation of Analytical Methods for Water and Wastewater
INTRODUCTION
The Office of Research and Monitoring, EPA, gathers water and air
quality data, and makes noise and solid wastes measurements, to determine
compliance with established environmental standards, to provide information
for planning natural resources development, to determine the effectiveness
of pollution abatement procedures and to assist in research activities. As
a help in achieving these goals, EPA Administrator, William D. Ruckelshaus,
recently established National Environmental Research Centers (NERC) at
Cincinnati, Ohio, Corvallis, Oregon, and Research Triangle Park, North
Carolina.
In a large measure the success of these environmental protection
efforts rests upon the reliability of the information provided by the data
collection activities. Therefore, the Analytical Quality Control Labora-
tory (AQCL) was established as part of the National Environmental Research
Center, Cincinnati, to insure the reliability of physical, chemical, biological
and microbiological water quality data generated and, when necessary, to
insure the legal defensibility of all such environmental quality information
collected by the Agency.
The Method and Performance Evaluation Activity of AQCL conducts evaluative
interlaboratory research studies of analytical procedures used in the Research
and Monitoring Office, EPA. In this study the demand parameters, chemical
oxygen demand (COD), total organic carbon (TOC), and biochemical oxygen
demand (BOD)} were tested to measure the accuracy and precision of the
selected methods used in participating laboratories. The evaluation of
results also permits a judgment of the relative capabilities of these labora-
tories performing these analyses.
-------
DESCRIPTION OF STUDY
Test Design
A simple test design was used in this study of demand parameters. Each
sample was prepared as a stable concentration in a sealed glass ampul.
When an aliquot of the concentrate was diluted to volume, constituents were
present at levels found in natural waters or sewage.
An aliquot from each ampul was diluted to volume with distilled water for
COD and TOG analysis and with distilled/seeded or natural water for BOD
analysis.
Preparation of Samples and Reporting of Results
Two water sample concentrates were prepared for Method Study 3 by
dissolving weighed amounts of reagent-grade chemicals in ASTM reagent-grade
distilled water to produce accurately-known concentrations of COD, BOD and
TOC.
The concentrates were preserved by steam sterilization and were checked
for stability by repeated analyses over a period of three months. These
analyses established that the solutions were stable and verified the concen-
tration of each constituent. Further confirmation of the true values was
obtained from determinations by an independent referee laboratory. These
calculated true values verified by analyses are shown in Table 1.
-------
Table 1. True Values for Demand Parameters
When diluted in distilled water according to instructions, the water
samples contained the following concentrations of constituents, mg/liter:
Parameter
Organic Carbon '
Chemical Oxygen Demand *
31
Biochemical Oxygen Demand '
Sample 1
4.85
12.3
2.2
Sample 2
107
270
194
The calculated theoretical levels of organic carbon produced
in these samples by dissolving the weighed amounts of high
purity reagents in distilled water.
21
The calculated theoretical COD values obtained by complete
oxidation of the samples to carbon dioxide, water, and
ammonia. In actual practice, the theoretical values are
very difficult to attain.
31
'The concentrations given for BOD are theoretical demands
based on reported values in Standard Methods using river
water as seed (1). These values may or may not be repro-
ducible in a specific laboratory because of natural
variability of the seed organisms from sample to sample.
Each analyst was instructed to dilute a separate 5.0 ml aliquot of each
concentrate to one liter with ammonia-free water or with river, lake or
estuarine water. Distilled water was used as the diluent for the COD and
TOC tests to avoid the extreme variability in data which might arise from
the inhibitive or exhibitive effects of the organic waste content of individual
waters. For COD and TOC tests, a result was obtained for distilled water plus
the increment. Natural water was used as the seed source and/or diluent for
the BOD test. Recovery of the BOD increment in natural water was determined
by difference from the natural water BOD. The sample concentrates were
shipped to participating laboratories in November, 1969 with detailed
instructions for analysis and the reporting of results within thirty days.
-------
Analytical Methods
The analyses were performed according to the FWPCA Methods for Chemical
Analysis of Water and Wastes, November, 1969. The methods used for these
analyses were referenced as follows:
Parameter
Chemical Oxygen Demand
Basic Reference
Standard Methods for the Examination
of Water and Wastewater, 12th ed.,
APHA, Inc., N. Y., 1965, 510-514.
Book of ASTM Standards, Part 23, 1969
Water; Atmospheric Analysis, pp. 243-246.
Biochemical Oxygen Demand
Standard Methods for the Examination
of Water and Wastewater, 12th ed.,
APHA, Inc., N. Y., 1965, 415-421.
Book of ASTM Standards, Part 23
1969, Water; Atmospheric Analysis,
pp. 715-723.
Total Organic Carbon
Book of ASTM Standards, Part 23, 1969
Water; Atmospheric Analysis, pp. 826-830.
-------
Glossary of Terms
The statistical measurements used in this report are defined as follows:
as % Relative Error (Bias). The signed difference between mean value
and the true value, expressed as a percent of the true value.
X - X~
R. E. = . - x 100
true
Confidence Limit (95%) . The range of values within which a single analysis
will be included, 95% of the time.
C. L. = X ± t -J-
where t = value from t table, a = standard deviation and
n = number of samples.
Mean (X) . The arithmetic mean of reported values, the average.
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 in mg/liter between lowest and highest reported values.
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 the deviation (precision) of a set of data
to the size of n so that the deviations for differing levels of a parameter
can be compared fairly.
R. D. = 100 -SL
X
Skewness (k) . A pure number, positive or negative, which indicates the lack
of symmetry in a distribution, For example, k is positive if the distri-
bution tails to the right and negative if the distribution tails to the
left.
2 (X, -X)3
* - — V-
na
-------
Standard Deviation (a). The most widely used measure of dispersion of a
set of data, a is equal to the square root of the variance and with
normal distribution indicates the deviation of 68% of the values
around the mean, while 1.96a indicates the deviation of 95% of the
values around the mean. The standard deviation, a, is the measure
of the deviation of the universe. However, in most experimental
work with limited sampling and in this study only an estimated
standard deviation, s is measurable. The calculation differs in that
n-l rather than n is used as the denominator. In this study and in
further studies, s and s% not a and a^ will be used to estimate the
deviation of the data. They will be referred to as the standard
deviation and variance respectively.
(I
n
n
t-test. The difference in analyzed and true value expressed as ratio
over the standard 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.
X . .
. , n - true value
t-value =
Standard Deviation (s)
True Value. Those amounts actually added in sample preparation. These
are not based on analyses, the latter being used only for verification.
2 2
Variance (CT ), (s ). The average of the squares of the deviations of a
group of numbers from their average, X.
C2 - (Z X±)2 S X2 - (Z 5
rc 0 n"
-^ s2 = n - 1
-------
RESULTS
Raw Data
Direct copies of the computer printout showing all test results reported
by participating laboratories are given by parameter in the APPENDIX.
TREATMENT OF DATA
Statistical Summary
Complete statistical summaries are given in pages 10 thru 19. Each
parameter is discussed in turn with data displayed and statistically
evaluated for each concentration. For ease in presentation and to prevent
round-off errors, five decimal places were carried in all measurements,
however, the number of significant figures is only equal to the number
reported for the increment.
A program described by Larsen (3) was modified for an IBM 1130 computer
and measurement of accuracy and presentation of ranked data were added.
This summary display named COLST provides all of the statistical measure-
ments necessary for evaluation of the data as a direct copy of the computer
printout. With the exception of accuracy, all measurements (number of values,
true value, mean, median, accuracy, range, variance, standard deviation,
95% confidence limit, relative deviation (coefficient of variation) and
skewness} are based on all data received, without rejection. Because the
inclusion of questionable extreme values will result in unreasonable values
for accuracy, the accuracy values for COD, BOD and TOG are based on retained
data, that is, the data remaining after rejection of outliers using the
t-test at the 99% level. In addition to the statistical measurements, all
data are ranked in ascending order and presented in a histogram, using
n = Js.d. cell divisions. Each X in the histogram represents one analytical
-------
8
result for up to 15 values/cell. When more than 15 values occur per cell
only 15 X's are printed and the number of values actually included is indicated
by the number at the base of each cell. The distribution of X values charac-
terizes the method as used on these natural water samples.
Although it is not possible to calculate a BOD result independently
of the seed material, it is necessary to use a true value to obtain some
measure of accuracy. In this study BOD values were calculated from a
table of these values for river water given in Standard Methods (1).
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 accuracy measure. These values
were probably caused by gross instrumental, chemical or human error. The
extreme values 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. The data points rejected are
indicated with a capital letter, "R", after the values in the Figures and in
the data tabulations in the Appendix. A greater spread of data around the
true value causes rejection of fewer outliers. As the standard deviation
of the method increases in the denominator of the t-test, the calculated t
value grows smaller and there are fewer extreme values rejected as outliers.
-------
DISCUSSION
Chemical Oxygen Demand (Dichromate Oxygen Demand)
COD Level, 12.3 mg/liter
The evaluation of the COD test at low level of organic demand is
shown on page 10. The COD method had only .3% bias. The data distri-
bution was normal but a number of laboratories had some difficulty as
indicated by a standard deviation of 4.15 mg/liter at this low level.
There is a 95% probability that analysis of a COD sample containing
12 mg/liter will vary 8.13 mg/liter or 66% about the mean.
COD Level, 270 mg/liter
The statistical summary on page 11 shows COD test results at the
200-300 mg/liter level expected in municipal sewage. The accuracy is
expressed as a negative bias of 4.7%, equivalent to 13 mg/liter at the
270 mg/liter level. The precision of the COD method was improved at this
270 mg/liter level to a standard deviation of 17.8 mg/liter. This is
equivalent to a 35 mg/liter variation about the mean at the 95% confidence
interval.
Summary of COD Analyses
The COD test showed a relatively low bias of 0.3% and 4.7% at levels
of 12 and 270 mgCOD/liter levels respectively. However, the COD test was
less precise at a low demand level, showing a 33% relative deviation in
analysis of 12 mg/liter level as compared with a 7% deviation about the
mean in analysis of the 270 mg/liter level.
-------
10
METHOD 6 PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
COD, SAMPLE 1
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
12.3
N
TRUE VAL-
MEAN
MEDIAN
ACCURACY
86
12.3
12.34066
12.00000
0.33062
RANGE
25.30000
17.22591
COEF. VAR.
VARIANCE 17.22591 SKEWNESS
STD. DEV. 4.15041 NO. OF CELLS
CONF. LIM. 8.13480 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
0.33631
0.85772
9
DATA IN ASCENDING ORDER
MIDPOINT FREQ
1.2R
4.0
5.8
6.2
6.3
6.7
6.8
7.3
7.4
7.8
8.0
8.1
9.4
9.6
9.8
9.9
10.0
10.1
10.2
10,2
10.4
10.7
10.7
10.7
10.8
10.8
10.9
11.0
11.2
11.3
11.3
11.3
11.3
11.4
11.4
11.4
11.5
11.6
11.6 '
11.7
11.7
12.0
12.0
12.0
12.0
12.2
12.2
12.2
12.2
12.3
12.3
12.4
12.5
12.5
12.5
12.5
12.5
12.6
12.6
12.7
12.8
12.9
13.0
13.1
13.2
13.4
13.9
14.1
14.3
14.6
14.8
14.8
15.2
15.4
16.6
16.9
17.2
17.4
17.8
18.9
19.4
21.7
22.8
22.9
24. OR
26. 5R
1.2000 1
4.3625 2
7.5249 9
10.6875 37
13.8499 25
17.0124 5
20.1749 3
23.3374 3
26.4999 1
HISTOGRAM
X
XX
xxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxx
XXX
XXX
X
REJECTED DATA.
-------
11
METHOD & PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
COD, SAMPLE 2
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
N 82
TRUE VAL. 270.
MEAN 257.37805
MEDIAN 261.00006
ACCURACY -4.67481
270.
RANGE
VARIANCE
STD. DEV.
CONF. LIM.
104.00001 COEF. VAR. 0.06900
315.39721 SKEWNESS -1.14857
17.75942 NO. OF CELLS 9
34.80847 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER
MIDPOINT FREQ
198.R
205.R
208.R
215.R
220.R
221.R
222.R
237.
238.
239.
240.
243.
244.
245.
246.
248.
250.
250.
251.
252.
253.
253.
253.
253.
254.
255.
255:
256.
257.
257.
257.
257.
258.
258.
259.
259,
260,
260,
260<
260,
261,
261,
261,
262,
262.
263,
263.
263.
263.
264,
264,
264,
264,
264,
264,
264,
265,
265,
265.
266.
267.
267,
267,
267,
267,
268,
268.
269,
269.
269,
270.
272.
274.
274.
275.
276.
277-
279.
279.
286.
289.
302.
198.
211,
224.
237,
250.
263.
276,
289,
302,
0000
0000
0000
0000
0000
0000
0000
0000
0000
1
3
3
5
16
42
9
2
1
HISTOGRAM
X
XXX
XXX
XXXXX
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxx
XX
X
REJECTED DATA.
-------
12
Biochemical Oxygen Demand (BOD)
BOD Level, 2.2 mg/liter
. ->
The BOD test had negligible bias at this low level common to many
natural waters. In this study, analysts were instructed to bring the
aliquot up to volume with distilled water and/or a natural water or
domestic sewage in sufficient amounts to provide ample seed. Although
the seventy-three analysts used as many different sources of natural water
and/or sewage as seed, the data on page 13 showed a negative bias of only
3.75%. The standard deviation of 0.7 mg/liter was 33% of the level tested.
This is the same relative deviation as was obtained with the TOC test on
the same sample. At this low 2 mg/liter level, BOD data will deviate
1.4 mg/liter about the mean at 95% confidence interval.
BOD Level, 194 mg/liter
At this higher level of organic loading; which is similar to a municipal
sewage, the BOD test results on page 14 show a negative bias of 9.8%. The
more
standard deviation of 26 mg/liter for BOD is ES3 than that of the COD or
TOC tests at this level. The BOD results can be expected to deviate 52
mg/liter about the mean, with a 95% confidence.
Summary of BOD Analyses
Although the BOD test is non-standard because of the inability to control
the most important "reagent" in the test, the microorganisms used as seed
material, the results of this study indicate that the test is capable of
reasonable accuracy and precision when a replicate synthetic non-toxic sample
is the substrate. However, in routine analyses of environmental samples, one
always needs to verify the accuracy of the BOD result because of the varying
ability of different seeds to fully oxidize the sample and the possible toxic
effect of the' substrate tested.
-------
13
METHOD £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
BOD, SAMPLE 1
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM SEEDED WATER
INCREMENT =
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
74
2.2
2.11756
2.10000
-3.74713
2.2
RANGE
VARIANCE
STD. DEV.
CONF. LIM.
5.50000 COEF. VAR.
0.49406 SKEWNESS
0.70290 NO. OF CELLS
1.37768 (95 PCT)
0.33193
1.80488
8
PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER
0.3R
1.0
1.0
1.0
1.1
1.2
1.3
1.4
1.5
1.5
1.6
1.6
1.6
1.6
1.6
1.7
1.8
1.8
1.8
1.8
1.8
1.9
1.9
1.9
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.1
2.1
2.1
2.1
2.1
2.2
2.2
2.2
2.2
2.2
2.2-
2.2
2.3
2.3
2.3
2.3
2.3
2.3
2.3
2.3
2.4
2.4
2.4
2.4
2.5
2.5
2.5
2.5
2.6
2.6
2.6
2.6
2.7
2.8
3.0
3.2
3.5
3.7
5.8R
MIDPOINT FREQ
0.3000
1.0857
1.8714
2.6571
3.4428
4.2285
5.0142
5.7999
1
7
39
23
3
0
0
1
HISTOGRAM
XXXXXXX
XXXXXXXXXXXXXXX
xxxxxxxxxxxxxxx
XXX
R = REJECTED DATA.
-------
14
MtTHOD £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
BOD, SAMPLE 2
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM SEEDED WATER
INCREMENT =
N
TRUE VAL.
MEA.M
MEDIAN
ACCURACY
73
194.
175.00003
179.00003
-9.79381
194.
RANGE
VARIANCE
STD. DEV.
CONF. LIM.
118.00001 COEF. VAR. 0.14975
686.80566 SKEWNESS -0.49208
26.20697 NO. OF CELLS 8
51.36566 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER
MIDPOINT FREO
107.R
118.R
120.R
121.R
123.R
130.
132.
135.
142.
145.
146.
150.
152.
154.
155.
155.
159.
159.
160.
164.
165.
165.
167.
168.
169.
169.
170.
171.
171.
172.
172.
173.
174.
174.
175.
175.
179.
179.
180.
180.
181.
182.
183.
183.
183.
185.
185.
185.
188.
189.
190.
191,
191.
192.
192.
193.
194.
197.
197-
198.
200.
200.
201.
203.
205.
205.
205.
205.
207.
218.
223.
224.
225.
107.
123.
140.
157.
174.
191.
208.
224.
0000
8571
7142
5714
4285
2856
1427
9998
1
6
4
11
20
18
9
4
HISTOGRAM
X
xxxxxx
xxxx
xxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxxxxxxxx
xxxxxxxxx
xxxx
R = REJECTED DATA.
-------
15
Total Organic Carbon (TOG) by Combustion-Infrared Analysis
The instrument for these TOG analyses is a single or dual furnace
with air pump, purification train, flow controls, non-dispersive type
infrared stream analyzer sensitized specifically for carbon dioxide, and
a recorder (4).
TOG Level, S mg/liter
At this low level of organic carbon, the TOC method results on
page 18 show a significant 15% positive bias, that is, the average
result was 15% higher than the true value. Furthermore, results at
this level deviate 33% about the mean TOC result of 5.6 mg/liter. At
the 95% confidence level, these low TOC values can be expected to deviate
3.7 mg/liter about the mean.
TOC Level, 100 mg/liter
At the higher municipal waste level of 100 mg TOC/liter, the TOC
results on page 19 showed greatly improved accuracy, with the bias
reduced to 1%. The standard deviation was reduced also to 6.0% of the
level tested. At the 100 mg/liter level, TOC values can be expected to
deviate 11-12 mg/liter about the mean values with a 95% probability.
Summary of TOC Analyses
The Total Organic Carbon test as performed on a Dow/Beckman Carbon
Analyzer or similar instrument had limited precision and accuracy at the
low (5 mg/liter) level. When samples contain organic carbon at levels
equivalent to municipal wastes, accuracy was improved to a 1% positive
bias and the relative standard deviation was reduced to 5-6% of the
level tested.
-------
16
Since the same technique and the same microsyringes are used to inject
the samples with low and high levels of organic matter, the same systematic
error should exist for both, with the exception that the sample containing
the higher level of organic carbon should have the added variability of
dilution. Although the specific cause of this imprecision at low levels
cannot be isolated here, it is most probably one or more of the following
factors:
1) Differences in the range settings used by the analysts for TOG
analyses, e.g., 1-30 mg, 1-100 mg or 1-1000 mg/liter at full range.
2) Differences in the volume of sample injected, i.e., use of a 20 yl,
40 yl or 100 yl syringe.
3) Individual differences in injection techniques.
4) Variable use of dilution technique to reduce high level TOC sample
before analysis.
5) Variable performance of needles, combustion systems and detectors
in the carbon analyzer instruments.
Because the EPA method research studies are intended to evaluate
analytical methods as they are used routinely in the field, little guidance
was given on technique other than furnishing the written analytical method
and informing the analyst of the range of the samples. Using basic instruc-
tions, the analyst did the measurement to the best of his ability.
There are chances for error in the TOC test which are unique among the
three oxygen demand tests. For example, injection of 40 yl volume sample
containing a 10 mg/liter level of TOC really involves a measurement on only
0.4 yg of carbon. A very slight contamination by dust, lint, cellulose, etc.
is enough to add a relatively large positive error. Similarly, small
inaccuracies in volumetric measurement could cause detectable error because
the error would be magnified by the large factor necessary to convert a
-------
17
yg/liter measurement to mg/liter in the final value.
We conclude that extreme care must be used operating the carbon analyzer
(Dow-Beckman type). Despite the availability of instrumentation and a
relatively good precision for the method reported in a single laboratory
study, data from this group of laboratories shows a significant increase
in variability. Apparently there is a lack of uniformity in the techniques
used by these laboratories. Use of automatic injection devices or syringes
should increase reproducibility.
-------
XcTHOO £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
TOC, SAMPLE 1
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
27
4.9
5.64814
5.00000
15.26828
4.9
RANGE
VARIANCE
STD. DEV-
CONF. LIM.
10.20000 COEF. VAR. 0.33522
3.58489 SKEWNESS 3.28807
1.89338 NO. OF CELLS 5
3.71102 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER
3.8
4.0-
4.2
4.3
4.5
4.7
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.4
5.7
6.0
6.0
6.3
6.4
6.5
6.5
7.0
7.2
14.0R
MIDPOINT FREO
3.8000
6.3500
8.8999
11.4499
13.9999
16
10
0
0
1
HISTUGRAM
xxxxxxxxxxxxxxx
xxxxxxxxxx
R = REJECTED DATA,
-------
19
METHOD £ PERFORMANCE EVALUATIONf AQCL
METHOD STUDY 3, DEMAND ANALYSES
TOC, SAMPLE 2
STATISTICS, ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
N
TRUE VAL.
MEAN
MEDIAN
ACCURACY
26
107.
108.07693
107.50001
1.00646
107.
RANGE
VARIANCE
STD. DEV.
CONF. LIM.
33.00000 COEF. VAR. 0.05551
35.99380 SKEWNESS 0.74660
5.99948 NO. OF CELLS 5
11.75898 (95 PCT)
PCT RELATIVE ERROR, RETAINED DATA
DATA IN ASCENDING ORDER
MIDPOINT FREQ
94.
100.
102.
103.
104.
104.
105.
106.
106.
107.
107.
107.
107.
108.
110.
110.
110.
110.
110.
110.
110.
110.
112.
114.
117.
127.R
94.0000
102.2500
110.5000
118.7500
127.0000
1
8
15
1
1
HISTOGRAM
XXXXXXXX
XXXXXXXXXXXXXXX
X
X
R = REJECTED DATA.
-------
20
CONCLUSIONS
The Office of Research and Monitoring, EPA, gave careful consideration
to the methods for measuring the oxygen demand of materials in waters,
before selection of methods for the EPA manual, Methods for Chemical Analysis
of Water and Wastes., 1971.
Three oxygen demand methods were selected for use in EPA: the Chemical
Oxygen Demand (COD) Test, the Biochemical Oxygen Demand (BOD) Test, and the
Total Organic Carbon (TOC) Test. BOD was included because of its historic
use; however, the EPA manual points out that the extreme differences in the
microorganisms used a seed and the variable biodegradability of chemical
substances, makes the BOD a non-standard test (5). ASTM's Subcommittee
D-19 has recommended withdrawal of the BOD test as a standard test because
of these same difficulties in its use (6). However, these variables also
make the BOD Test unique. It is the only test of the three that measures
the amount of oxygen used by selected microorganisms in stabilizing the
substances under test.
This study was designed to measure the effectiveness of the three
oxygen demand methods in degrading a test substance. The oxidation used
differed for each of the three methods. By using a test substrate which
was completely degradable for all three methods, the study measured
variability in the methods and analysts rather than variability in
degradation of the sample.
In this study the three methods of measuring oxygen demand had
similar accuracy and precision. The COD Test was most accurate at the
low level with a 0.3% bias and the TOC Test was most accurate at the
high level with a 1% bias.
-------
21
Although true values for oxygen demand ranged from 2-12 mg/liter
for COD, BOD and TOG analyses on the same low level sample, the relative
deviation of the three tests was an identical 33%, indicating that these
deviations were caused by imprecision in the methods not differences
in the 2-12 mg/liter level of oxygen demand present. At the higher level
of oxygen demand, the TOC and the COD tests were similar with 5.6% and
6.9% relative deviation, respectively.
The similar precision statements for COD, BOD and TOC at the low
level (1-10 mg/liter) support the contention that a major factor in
precision of demand analyses is the level tested. Regardless of the
method, good precision is difficult to achieve. At the higher levels,
all methods improved, with COD and TOC showing both increased precision
and increased accuracy. This study has also shown that when controls are
exerted over methodology and some uniformity is followed in the seed usage,
the BOD results on a replicate biodegradable sample are reasonable.
However, the accurate and precise performance of the BOD method in this
study was dependent on the uniformly good response of the wide variety of
BOD seed materials. The good response was itself dependent on the easy
biodegradability of the glucose-glutamic acid substrate tested. If the
sample substrate was more resistant to bacterial oxidation, lower and more
variable BOD values would have been recorded while the COD and TOC values
most probably would have remain the same as reported here.
The TOC test is a relatively new instrumental method of measuring
organic carbon. In this study it had a significant bias at lowest level
tested. Further study of the TOC method and increased standardization of
techniques should result in greater accuracy and precision, especially at
the lower levels of oxygen demand (less than 10 mg/liter).
-------
Ratios between COD, BOD and/or TOG can be developed easily for these
sets of data on a single sample. However, a ratio is not reported because
it is not applicable to any other sample. It is not possible to measure
oxygen demand using one of these parameters, and by use of a precalculated
ratio to determine the relative value for either of the other two parameters
for another sample.
-------
23
REFERENCES
1. Standard Methods for the Examination of Water and Wastewater, 12th ed.,
APHA, Inc., N. Y. , 1965, 419.
2. FWPCA Methods for Chemical Analysis of Water and Wastes , November, 1969.
Analytical Quality Control Laboratory, Division of Water Quality Research,
FWPCA.
3. Larsen, K. E. 1969. The Summarization of Data. J". <&a1. Technol.,
Vol. 1, No. 1, 1968.
4. Van Hall, C. E., J. Safranko and V. Stenger, "Rapid Combustion Method
for Determination of Organic Substances in Aqueous Solutions," Analytical
Chemistry, 55, 1963, 315-319.
5. Methods for Chemical Analysis of Water and Wastes, Environmental Protection
Agency, National Environmental Research Center, Analytical Quality Control
Laboratory, Cincinnati, Ohio, 1971.
6. Annual Book of ASTM Standards, Part 23, Water: Atmospheric Analysis, 1970,
p. 712, American. Society for Testing and Materials, Philadelphia, Pa.
-------
25
APPENDIX
Raw Data Summary
-------
27
METHOD £ PERFORMANCE EVALUATION, AOCL
METHOD STUDY 3, DEMAND ANALYSES
COD» SAMPLE 1
ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
12.3
101
101
102
103
106
106
107
107
109
109
109
110
110
110
112
113
115
117
120
121
122
123
123
124
124
127
128
128
130
131
133
137
138
140
140
148
148
149
150
151
152
152
153
1
2
1
1
1
2
1
2
1
2
3
1
2
3
1
2
1
1
1
1
1
1
2
1
2
1
1
2
1
1
1
1
1
1
2
1
2
1
1
1
1
2
1
10.1
11.3
11.4
6.3
7.8
8.0
11.3
12.3
22.8
22.9
13.4
12.9
11.2
11.6
16.6
11.6
18.9
12.2
8.1
12.6
6.8
12.4
15.4
12.5
12.5
11.3
14.6
13.0
11.7
11.0
12.3
5.8
12.5
10.7
10.9
14.1
17.8
13.1
11.5
14.8
14.8
15.2
11.3
NUMBER
OF
LAB/ANALYST
153
153
153
153
153
154
155
156
157
158
160
160
160
160
160
161
162
163
163
164
165
166
167
168
169
170
170
170
170
171
171
171
172
173
175
176
177
179
180
180
180
181
187
2
3
4
b
6
1
1
1
1
1
1
2
3
4
5
1
1
1
2
1
1
1
1
1
1
1
2
3
4
1
2
3
1
1
1
1
1
1
1
2
3
1
1
INCREMENT
RECOVERY
BY LAB.
11.4
11.4
10.7
10.4
10.8
13.9
17.4
7.3
12.2
12.6
6.7
17.2
10.8
9.8
9.6
4.0
1.2R
13.2
12.7
10.2
26. 5R
12.0
10,0
21.7
12.8
14.3
16.9
11.7
12.0
6.2
12.0
24. OR
9,4
7.4
10.7
19.4
9.9
12.5
12.2
12.5
12.2
12,0
10.2
R = REJECTED DATA.
-------
28
METHOD £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
COD, SAMPLE 2
ALL DATA, "ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT =
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
270.
NUMBER OF
LAB/ANALYST
INCREMENT
RECOVERY
BY LAB.
101
101
102
103
106
106
107
107
109
110
110
110
112
113
117
120
121
122
123
123
124
124
127
128
128
130
131
133
136
138
140
140
140
148
148
150
151
152
152
153
153
153
153
1
2
1
1
1
2
1
2
3
1
2
3
1
2
1
1
1
1
1
2
1
2
1
1
2
1
1
1
1
1
1
2
3
1
2
1
1
1
2
1
2
3
4
253.
248.
267-
257.
205. R
208. R
270.
265.
256.
222. R
264.
220. R
272.
240.
274.
260.
264.
289.
260.
238.
246.
260.
239.
261.
267.
265.
264.
258.
266.
263.
253.
257.
237.
269.
277.
260.
261.
264.
264.
252.
244.
254.
253.
153
153
154
155
156
157
158
160
160
160
160
160
161
162
163
163
164
165
166
167
168
169
170
170
170
170
171
171
171
172
173
175
177
179
180
180
ISO
181
187
5
6
1
1
1
1
1
1
2
3
4
5
1
1
1
2
1
1
1
1
1
1
1
2
3
4
1
2
3
1
1
1
1
1
1
2
3
1
1
259.
263.
262.
275.
215. R
264.
262.
198. R
274.
250.
276.
279.
258.
250.
255.
255.
253.
286.
263.
268.
257.
257.
302.
251.
269.
267.
263.
265.
264.
245.
261.
221. R
259.
243.
268.
267.
279.
269.
267.
R = REJECTED DATA.
-------
29
METHOD £ PERFORMANCE EVALUATION AQCL
METHOD STUDY 3, DEMAND ANALYSES
BODt SAMPLE 1
ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM
INCREMENT =
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
2.2
101
101
102
103
106
106
107
107
110
112
113
113
115
117
121
122
123
123
124
124
128
128
130
133
136
137
140
140
140
148
148
149
151
152
152
152
154
155
156
157
158
160
160
1
2
1
1
1
2
1
2
3
1
1
2
1
1
1
1
1
2
1
2
1
2
1
1
1
1
1
2
3
1
2
1
1
1
2
3
1
1
1
1
1
1
2
2.0
2.0
2.8
2.3
2.0
1.0
2.3
1.9
2.1
2.4
2.2
2.2
3.7
2.6
2.2
1.8
-2.6
2.5
2.0
2.0
1.6
1.0
2.0
1.1
1.6
0.3R
2.7
2.1
2.0
1.4
1.5
2.0
3.5
2.4
2.1
1.8
2.2
1.6
2.0
2.6
1.7
2.1
2.0
SEEDED WATER
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
1.9
1.8
1.9
2.5
1.3
1.0
2.3
3.2
2.2
2.4
2.0
1.2
1.8
2.2
1.8
2.3
2.3
2.3
1.6
3.0
2.5
5.8R
1.6
2.4
1.5
2.3
2.3
2.1
2.2
2.5
2.6
160
160
160
161
163
163
164
165
166
167
168
169
170
170
170
170
171
171
171
172
173
175
176
177
179
180
180
180
183
187
188
3
4
5
1
1
2
1
1
1
I
1
1
1
2
3
4
1
2
3
1
1
1
1
1
1
1
2
3
1
1
1
R = REJECTED DATA.
-------
30
METHOD £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
BOD, SAMPLE 2
ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM
NUMBER
LAB/ANALYST
INCREMENT =
OF INCREMENT
194,
RECOVERY
BY LAB.
SEEDED WATER
NUMBER OF
LAB/ANALYST
101
101
102
103
106
106
107
107
110
112
113
113
117
120
121
122
123
123
124
124
127
128
128
130
133
136
138
140
140
140
148
148
150
151
152
152
152
154
155
156
157
158
160
1
2
1
1
1
2
1
2
3
1
I
2
1
1
1
1
1
2
1
2
1
1
2
1
I
1
1
1
2
3
1
2
1
1
1
2
3
1
1
1
1
1
1
190.
183.
169.
123. R
159.
145.
189.
185.
205.
198.
200.
172.
142.
152.
183.
175.
150.
191.
207.
192.
224.
218.
205.
165.
107. R
155.
223.
205.
225.
121. R
185.
181.
180.
169.
170.
171.
185.
155.
168.
171.
200.
174.
191.
INCREMENT
RECOVERY
BY LAB.
160
160
160
160
161
163
163
164
166
167
168
169
170
170
170
170
171
171
171
172
173
176
177
179
180
180
180
183
187
188
2
3
4
5
1
1
2
1
1
1
1
1
1
2
3
4
1
2
3
1
1
1
1
1
1
2
3
1
1
I
197.
193.
194.
179.
130.
197.
188.
159.
182.
154.
120.
146.
175.
179.
203.
172.
205.
180.
118.
167.
135.
165.
201.
132.
164.
173.
160.
192.
174.
183.
R
R =•REJECTED DATA.
-------
31
METHOD £ PERFORMANCE EVALUATIONt AQCL
METHOD STUDY 3, DEMAND ANALYSES
TOC, SAMPLE 1
ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT = 4.9
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
101 1 6.5
101 2 6.3
102 1 5.0
103 1 6.4
106 1 5.0
106 2 5.0
107 1 5.0
107 2 5.0
109 1 5.0
109 2 6.0
109 3 6.0
110 1 4.3
112 1 4.5
115 1 14.OR
120 1 7.0
121 1 5.0
122 1 5.7
123 1 4.2
123 2 5.0
127 1 4.7
140 1 3.8
151 1 5.0
158 I 4.0
163 1 7.2
163 2 6.5
179 1 5.4
183 1 5.0
R = REJECTED DATA.
-------
METHOD £ PERFORMANCE EVALUATION, AQCL
METHOD STUDY 3, DEMAND ANALYSES
TOC, SAMPLE 2
ALL DATA, ALL LABORATORIES
MANUAL PROCEDURE
RECOVERY OF INCREMENT FROM DISTILLED WATER
INCREMENT = 107.
NUMBER OF INCREMENT
LAB/ANALYST RECOVERY
BY LAB.
101 1 108.
101 2 106.
102 1 110.
103 1 110.
106 1 107.
106 2 107.
107 1 112.
107 2 110.
109 1 110.
109 2 110.
109 3 114.
110 1 103.
112 1 107.
120 1 127.R
121 1 106.
122 1 107.
123 1 104.
123 2 100.
127 1 110.
140 1 102.
151 1 94.
158 1 117.
163 1 105.
163 2 104.
179 1 110.
183 1 110.
R = REJECTED DATA.
------- |