EPA-650/4-74-046
SEPTEMBER 1974
Environmental Monitoring Series
-------
EPA-64 650/4-74-046
COLLABORATIVE TEST
OF THE TGS-ANSA METHOD
FOR MEASUREMENT OF NITROGEN DIOXIDE
IN AMBIENT AIR
by
Paul C. Constant, Jr., Michael C. Sharp,
and George W . Scheil
Midwest Research Institute
425 Volker Blvd.
Kansas City, Missouri 64110
Contract No. 68-02-1363
Program Element No. 1HA327
ROAPNo. 26AAF
Project Officer: John H . Margeson
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON. D.C. 20460
-------
This report has been reviewed by the Environmental Protection Agency
and approved for publication. Approval does.n,ot signify that the con-
tents necessarily reflect the views and policies of the Agency, nor does
mention of trade names or commercial products constitute endorsement
or recommendation for use.
Publication No. EPA-650/4-74-046
11
-------
FOREWORD
This program, "Collaborative Testing of Methods for Measurement of
N02 in Ambient Air," is being conducted under the Environmental Protec-
tion Agency (EPA) Contract No. 68-02-1363, which is Midwest Research
Institute's (MRI's) Project No. 3823-C. The program is concerned with
the evaluation of the following four methods with regard to their pre-
cision and accuracy:
1. Sodium-Arsenite,
2. TGS-ANSA,
3. Continuous-Saltzman, and
4. Cherailuminescence.
The collaborative study covered by this two-volume report is of
the TGS-ANSA procedures, which is a tentative manual method. In summary,
MRI's responsibility was to develop an N02, ambient-air sampling system
for use with the four methods, provide the test site and facilities
thereon where the collaborative tests would be conducted, select the
collaborators with regard to the program, prepare a plan of test for
the collaborative test, schedule testing, coordinate the test, retrieve
field data and results from the collaborators' analysis of their sam-
ples, statistically analyze their results, and report its findings to
EPA. The 10 collaborators who participated in the TGS-ANSA collaborative
test are:
Commonwealth of Kentucky
Department for Natural Resources and Environmental Protection
Division of Air Pollution
Frankfort, Kentucky 40601
Shell Development Company
P.O. Box 481
Houston, Texas 77001
111
-------
City of Philadelphia Air Management Services Laboratory
1501 East Lycoming Street '
Philadelphia, Pennsylvania 19124
North Ohio Valley Air Authority
814 Adams Street
Steubenville, Ohio 43952
State of New York Department of Health
Division of Laboratories and Research
New Schotland Avenue
Albany, New York 12201
Kennecott Copper Corporation
Utah Copper Division
P.O. Box 11299
Salt Lake City, Utah 84111
San Bernardino County Air Pollution Control District
172 West Third Street
San Bernardino, California 92415
Mecklenburg County Department of Public Health
1200 Blythe Boulevard
Charlotte, North Carolina 28203
Air and Industrial Hygiene Laboratory
State of California Health and Welfare Agency
Department of Health
2151 Berkeley Way
Berkeley, California 94704
National Bureau of Standards
B 326 Chemistry Building
Washington, D.C. 20234
This volume, Volume 1 of the report of test, summarizes MRI's and
the collaborators' activities. It describes the development of the NC>2,
ambient-air sampling system, which covers the general concept of the
system, design considerations, the design of the system and the system
checkout. Following this, there are discussions on the test site, the
selection of collaborators, the formal statistical design including the
presentation of factors and parameters that were considered, the col-
laborators' field sampling at the test site, the collaborators' analysis
iv
-------
of their samples—both test and standard samples—MRI's statistical
analyses of the collaborators' results, conclusions and recommendations.
Appendices contain a copy of the tentative, TGS-ANSA method, informa-
tion on the permeation tubes prepared for this program by the National
Bureau of Standards, written communiques with collaborators, instruc-
tions for collaborators, and MRI's field, operational, and data-log
sheets.
Volume 2 of this report of tests contains only the collaborators'
field data sheets for the four-run, 240-test sample, collaborative test.
The following individuals of the collaborating organizations are
acknowledged for their excellent work in the TGS-ANSA collaborative
test:
California Department of Health
Mr. Kenneth Smith, field sampling and laboratory analysis
City of Philadelphia
Mr. Donald Kutys, field sampling and laboratory analysis
Kennecott Copper Corporation
Mr. Lynn Hutchinson, field sampling and laboratory analysis
Kentucky Division of Air Control
Mr. Joe Andrews, field sampling and laboratory analysis
Mecklenburg County Department of Public Health (North Carolina)
Mr. James T. Ward, field sampling and laboratory analysis
National Bureau of Standards
Mr. Bob Deardorff, field sampling and laboratory analysis
North Ohio Valley Air Authority
Mr. Dan Zorbini, field sampling and laboratory analysis
San Bernardino County Air Pollution Control District
Dr. C. Kenneth Wilcox, field sampling and laboratory analysis
Shell Development Company
Mr. W. T. Shebs, field sampling and laboratory analysis
State of New York Department of Health
Ms. Barbara Kladatos, field sampling and laboratory analysis
-------
Special acknowledgements are made to the National Bureau of Stan-
dards and to Mr. Ernest E. Hughes and Dr. John K. Taylor of NBS who
provided the N0£ permeation tubes for this collaborative test; and to
Dr. John B. Clements, Chief, Methods Standardization Branch, National
Environmental Research Center, Environmental Protection Agency, and
Mr. John H. Margeson, Government Project Officer, Methods Standardiza-
tion Branch for their valuable suggestions in planning and design.
This MRI collaborative program is being conducted under the
management and technical supervision of Mr. Paul C. Constant, Jr.,
Head, Environmental Measurements Section of MRI's Physical Sciences
Division, who is the Program Manager. Those who contributed to this
test are: development of the N02, ambient-air sampling system -
Dr. Chatten Cowherd, Jr., Mr. Fred Bergman, Mr. Emtie Baladi, and
Mr. Wallace Yocum; experimental design and statistical analysis -
Mr. Michael C. Sharp; and preparation and operation of test facili-
ties - Dr. George W. Scheil, Mr. John LaShelle, Mr. Robert Stultz,
Mr. Bob Kamerman, and Mr. Kevin Cline.
Approved for:
H. M. Hubbard, Director
Physical Sciences Division
17 January 1975
vi
-------
CONTENTS
Page
Foreword m
List of Figures ix
List of Tables xi
Summary ...... ..... xii
Introduction 1
N02> Ambient-Air Sampling System 3
General Concept 3
Design Factors 4
System Design 7
System Checkout 19
Ambient Levels of NO and N02 19
Subsystems and Units 20
System Operation 21
Test Site 23
Selection of Collaborators 29
Statistical Design 33
General Considerations and Comments 33
The Formal Design 35
Collaborators' Field Sampling 37
vii
-------
CONTENTS (Concluded)
Page
Analyses of Samples 41
Analyses Performed by the Collaborators 41
Collaborators' Results 43
Statistical Analyses of Collaborators' Results 45
Biases 47
Precision 49
Summary and Discussion of Statistical Analysis 51
Lower Detectable Limit (LDL) 53
Conclusions 55
Recommendations 57
Appendix A - Tentative Method for the Determination of Nitrogen
Dioxide in the Atmosphere (TGS-ANSA Procedure) ... 59
Appendix B - Data on the Permeation Tubes Used as the Source of
the Spiked Levels of N02 73
Appendix C - Calibration of the Venturi and Dry-Gas Meter 75
Appendix D - Written Communications with Potential Collaborators . 79
Appendix E - Instructions for Collaborators NOo Collaborative
Test: Method TGS-ANSA Procedure 85
Appendix F - N02» Ambient-Air Sampling System Operation Data:
Test Log Sheets and Test Data Sheets 97
Appendix G - Collaborators1 Comments Ill
Appendix H - Analysis of Variance Including Collaborator I .... 117
viii
-------
FIGURES
Figure Page
1 N02, Ambient-Air Sampling System Concept 5
2 Final Design of the N02, Ambient-Air Sampling System ... 8
3 Annotated Photographs of the NC>2, Ambient-Air Sampling
System in Operation 9
4 Ambient-Air Stream Splitter 11
5 Photographs of the N02 Bleed-In Unit—Assembled and
Diassembled 14
6 Schematic Drawing of the NO- Permeation Tube Assembly . . 15
7 Schematic Drawing and Photographs of the Diffuser .... 17
8 Schematic Drawing and Photographs of the Sampling
Manifold 18
9 Collaborative Test Site: MRI's Field Station 24
10 Photographs of the Test Facility 25
11 Collaborators' Sampling Areas at the Test Site 26
l
12 Photograph of Field Personnel of the N02 Collaborative
Test (TGS-ANSA Method); MRI Field Station; April 29 -
May 3, 1974 38
ix
-------
FIGURES (Concluded)
Figure Page
13 Nitrogen Dioxide Data Sheets--TGS-ANSA Method 40
14 Collaborator Percent Bias Versus Level of N02 50
C-l Venturi and Dry-Gas Meter Calibration System 77
-------
TABLES
Table Page
1 Sodlum-Arsenite Collaborative Test Schedule 37
2 Collaborator Results from Collaborative Test Using the
TGS-ANSA Method 42
3 R" Versus NC>2 Level 47
4 Analysis of Variance (Response = Bias) 47
3 '
5 Average Bias (ug/m ) Per N02 Level 48
6 Collaborator Biases (ug/m3) (All N02 Levels) 48
7 Collaborator Percent Bias Per N02 Level 49
8 Components of Variance (All N02 Levels) 49
F-l Run No. 1 Test Data 99
F-2 Run No. 2 Test Data 100
F-3 Run No. 3 Test Data 101
F-4 Run No. 4 Test Data 102
xi
-------
SUMMARY
A collaborative test was conducted by MRI in the Greater Kansas
City Area during the week of 29 April 1974. Ten organizations partici-
pated in this test of the "Tentative Method for the Determination of
Nitrogen Dioxide in the Atmosphere (TGS-ANSA Procedure)." All collab-
orators sampled simultaneously from the N02» ambient-air sampling sys-
tem that was developed by Midwest Research Institute specifically for
this collaborative test program. For each of the four 24-hr runs (each
of a different average N02 level: 65.6, 117, 223, and 315 jig/mS), each
collaborator drew six samples simultaneously; four from the N02~spiked
section and two from the unspiked (ambient-air) section of the sampling
system. Each collaborator was given, for analysis with his test samples,
three standard samples (two N02 and one blank) that were prepared by
MRI.
The N02 challenge (spike) levels were obtained from permeation tubes
that were developed by the National Bureau of Standards.
The collaborators analyzed the test and the standard samples at
their home laboratories, and submitted their results to MRI. MRI checked
the collaborators' calculated results and found no gross errors. The
collaborators' results were then statistically analyzed.
The collaborators sampled from both the spiked and unspiked lines of
the N02» ambient-air sampling system, providing two sets of collaborators'
results. The two sets of results were used to determine true values of the
levels of NO2 that comprised the challenges to the collaborators' sampling
trains. In addition, for both sets of results, there was an analysis of
variances made to determine biases and components of variances—the vari-
ances of repeated observations and variances between collaborators.
xii
-------
The bias to the NO2 determinations was relatively small—approxi-
mately 5%—and it was independent of the NO2 level.
Three methods of estimating the lower detectable limit (LDL) were
used. From the results of these calculations, it is reasonable to state
that the LDL within a collaborator is probably £ 9 p,g/m3, and the LDL
from a set of collaborators £ 15 |j,g/m3.
The major conclusions that can be drawn from the results of this
collaborative test are:
1* The N02, ambient-air sampling system developed by MRI is an ef-
fective system for use in collaborative testing of manual methods such
as the TGS-ANSA procedure.
2. The "Tentative Method for the Determination of Nitrogen Dioxide
in the Atmosphere (TGS-ANSA Procedure)" is adequately written for those
knowledgeable of sampling and analysis techniques as presented therein.
3. If the tentative TGS-ANSA procedure as given in Appendix A of
this report is followed by people knowledgeable of the sampling and analy-
sis techniques given therein, then such a person will obtain results with
an average bias of 9.5 jig/m3 over the range 50-300 jig/m3. The precisions
can be estimated from the within laboratory standard deviation (ae) of
7.5 vig/m3, and the collaborator standard deviation (ac) of 8.8 yg/m3.
Based upon the conclusions that have been drawn from the results of
this collaborative test, it is recommended that:
1. The same N02 sampling system be used in the evaluation of the
remaining N0£ methods to be tested;
2. The data sets to be obtained from the subsequent methods to be
evaluated be based on experimental designs, test procedures and sampling
system operational procedures as similar as possible to those of the TGS-
ANSA collaborative test so that comparisons of the methods are based on
similar criteria; and
3. No further statistical analysis be made of the results from the
TGS-ANSA method until the results from the other three methods are ob-
tained.
xiii
-------
INTRODUCTION
The Methods Standardization Branch, National Environmental Research
Center of the Environmental Protection Agency (EPA) is engaged in a pro-
gram to evaluate four methods for measuring N(>2 in ambient air. Midwest
Research Institute (MRI) is working for EPA under Contract No. 68-02-
1363 to provide EPA data on the precision and bias of the following four
methods: sodium-arsenite and TGS-ANSA, which are manual methods, and
continuous-Saltzman and. chemiluminescent, which are instrumental methods.
To achieve this objective, a collaborative testing program Is
being conducted that will assess interlaboratory as well as intralab-
oratory testing. In summary, MRI in the execution of this program,
selects the collaborators, provides sampling locations and facilities
thereon, orients the collaborators relative to the program, prepares
a plan of test for each method tested, schedules testing, coordinates
the collaborative tests, retrieves field data and results of the col-
laborators' chemical analyses of their field samples, statistically
analyzes results received from the collaborators, and reports results
of the program to EPA.
These activities were performed by MRI on its second test under-
taken on the contract. The method investigated was the "Tentative
Method for the Determination of Nitrogen Dioxide in the Atmosphere
(TGS-ANSA Procedure)," dated April 1974. A copy of the write-up of this
method is given in Appendix A.
The program was initiated on 30 June 1973, and this collaborative
test took place at MRI's field station in Kansas City, Missouri, during
29 April - 3 May 1974, with 10 different collaborators. The interim
period was devoted to the preparation for this test and conduction of
the first collaborative test, which covered the sodium-arsenite pro-
cedure. A major task of the preparation activity was the development of
a precise NC^, ambient-air sampling system that could be housed indoors
and be suitable for all four methods.
-------
The two major phases of the test program were sampling and analysis.
The sampling phase covered the field test where the collaborators col-
lected their samples from the ambient-air sampling system. The analysis
phase covered the chemical analyses of field samples by the collabora-
tors and the statistical analyses of their results by MRI. After the
field test, the collaborators returned to their home laboratories where
they analyzed their samples and reported their results to MRI. Then MRI
performed its statistical analysis and prepared this report of the TGS-
ANSA collaborative test.
This report covers the collaborative test of the tentative TGS-ANSA
method in the following order: the second section discusses the N02,
ambient-air sampling system MRI developed for this program, covering the
general concept of the system, the design considerations, the system
design, and the system checkout. The third section describes the test
site and the facilities that were used at this site. The fourth discusses
how the collaborators were selected and who they are. The fifth section
presents the factors and parameters that were considered in the formal
experimental design as well as the formal design. The sixth section
summarizes the test activities during the collaborative test. The seventh
section discusses the analyses that were performed by the collaborators.
The collaborators' results are presented in this section as well as the
analysis MRI conducted. The eighth section discusses the statistical
analysis of the collaborators' results and presents the results from
this analysis, which includes biases and components of variance. The
ninth and tenth sections present conclusions and recommendations, re-
spectively. The appendices contain a copy of the tentative TGS-ANSA Pro-
cedure method, data on the permeation tubes that were used as the source
of NO2 in the spiked section of the sampling system, information con-
cerning the calibration of the venturi and dry-gas meter, copies of
written communiques MRI had with the collaborators, a copy of the test
instructions that were given to the collaborators, the N02» ambient-air
sampling system's operational data, collaborators' comments, results of
MRI's analyses, and additional statistical-analysis information.
-------
N02, AMBIENT-AIR SAMPLING SYSTEM
GENERAL CONCEPT
Primary requirements for the evaluation of an ambient-air method
by on-site collaborative testing are: (1) that all collaborators sample
the same air, (2) that the samples be representative of ambient air, and
(3) that the concentration of N02 in the samples be accurately known and
controllable over the region of interest. The first requirement can be
met by using a manifold system with each collaborator taking samples
from a common stream of air. The second and third requirements are some-
what antagonistic to one another and not as easily solved.
Ideally, these requirements can be met by obtaining actual ambient
samples over a wide range of concentration. However, this approach would
require that each level be obtained at a different location with the
additional requirement of fortuitous weather conditions, since weather
conditions have a strong effect on ambient N02 concentrations. An addi-
tional problem with this approach is that no accepted primary reference
method exists for the analysis of N02 in ambient air.
However, gravimetrically calibrated NO^permeation tubes are avail-
able which generate a stable, precise rate of release of high purity
NC-2 over a period of a few years. By using a set of these tubes different
levels of N02 can be generated by adding the N02 from the permeation
tubes to a stream of air with a known flowrate. Since the test conditions
must relate to actual ambient-air conditions, the N02 from the permeation
tube can be added as a known addition or spike to the ambient air stream.
The method under test should show a difference in concentration between
samples of ambient and spiked air equal to the spike level. To ensure
that the NC-2 concentration of the spiked sample does not exceed the maxi-
mum level of interest--350 pg/ra3—and to allow control of the spiked
air N02 level over a reasonably broad range, the average ambient levels
must be well below the lowest NO? concentration to be tested, in this
o *•
case 50 yg/m .
-------
To achieve this, the following system is used: outdoor ambient
air is drawn into the sampling system through a single tube, as shown
in Figure 1. The air is divided downstream into two sections—spiked
and unspiked. A controlled flow of ambient air at a specific value
exists in the spiked section. A comparable ambient-air flow exists in
the unspiked section, but the latter is uncontrolled. Temperature-
controlled permeation tubes provide the source of N02 which is injected
into the spiked section at a desired level. The N02 is then thoroughly
mixed with the ambient air in a mixing unit—a diffuser. The mixture
is then equilibrated before it reaches the sampling station where the
collaborators sample from identical ports—subjected to the same gas
flow (spiked plus ambient). A continuous monitor is attached to moni-
tor the gas at the spiked and unspiked sampling levels to monitor the
integrity of the spike. The collaborators sample ambient air simul-
taneously at an identical sampling manifold that is at a similar
location in the unspiked section. The gas in both sections is then
exhausted to the outdoors.
DESIGN FACTORS
The design of the N02, ambient-air sampling system was based on
the following factors:
1. The flowrate of each of the four methods to be tested is ap-
proximately 0.2 liter/min, with a maximum of 1 liter/min for some of
the instruments that would be used in the instrumental methods.
2. The sampling period of each manual method is 24 hr; each in-
strumental method is preferably 24 hr, but could be less.
3. NC>2 permeation tubes whose rates are approximately 1 ug/min,
which are furnished by the government, are the source for the spiked
levels of N(>2. These tubes are to be operated at 25.1°C + 0.2°C.
4. The number of collaborators for each collaborative test is to
be 10.
5. The maximum number of samples taken simultaneously by each
collaborator during each run of a manual method is four spiked and two
ambient (unspiked) for a total of six samples/collaborator/run. The
multiplicity of samples per run—both spiked and unspiked—is to pro-
vide replicates.
-------
AMBIENT AIR INTAKE
CENTRIFUGAL
BLOWER
ROOF
a
LLJ
BAFFLE-
SPLITTER
SAMPLING
MANIFOLD
SAMPLE
DRAW-OFF
LVENTURI
LlCTCn
1
EQUILIBRATION
SECTION -x.
PURGE LINE
CONTROL
VALVE
EXHAUST
RECORDER
SAMPLE
DRAW-OFF
RECORDER
SAMPLING
MANIFOLD
PERMEATION
TUBES
CARRIER
GAS
EXHAUST
PURGE LINE
Figure 1. N02, ambient-air sampling system concept
-------
6. The N02 range of concern is 50-350 WS/m3, which is representa-
tive of ambient conditions.
7. There are four different NC^ spiked levels: high, low, and
two medium. Each level is maintained throughout the run's period, within
the accuracy of the system.
8. The test period is to be no more than 6 days, which is based
upon the concensus of potential collaborators surveyed.
9. The overall N0£ sampling system accuracy is to be 5% or better.
10. The flow control in the spiked section is to be 2% or better.
11. Flow parameters of the spiked section are to be measured.
12. One NC>2/NO chemiluminescent device, switched between spiked
and unspiked sampling manifolds (or stations), is to be used as a moni-
toring instrument.
13. Only one person from each collaborators' organization will be
needed in the field for each method.
14. There is turbulent flow in the spiked section between the point
of injection of the spiked levels of N02 and the diffuser to provide
mixing of the spiked N02 with the ambient air. The diffuser insures
proper mixing. Up to 20% of the stream in each section--spiked and
ambient air—can be sampled to (a) insure that there is capacity in
the main stream to provide each collaborator with his needs in case
there is a problem with one or more collaborators drawing an excess
amount, and (b) allow the quantity of spiked flow to be drawn from the
center of the spiked line where there is assurance of equilibration.
There is to be a minimum amount of adsorption of the spiked N02 on sur-
faces, from its source to and including the sampling manifold. By using
Teflon or glass as the material in which the gases come in contact and
by maintaining a high gas flowrate, which allows for extremely short
residence times, adsorbtivity of NO2 on surfaces and reaction with water
vapor and other losses are insignificant.
15. Each section--spiked and unspiked--is to be similar, including
material and geometric aspects.
16. Each section is to be under positive pressure so that no un-
wanted air will be pulled into the system in case there were a leak.
-------
17. Collaborator's equipment size, configuration and power require-
ments must be met.
18. Environmental effects on operation of sampling system must be
considered.
SYSTEM DESIGN
The final design of the N(>2» ambient-air system is shown in a gen-
eral schematic form in Figure 2. Annotated photographs of this opera-
tional system are given in Figure 3.
The input to the system is located outdoors about 2 m above ground
level and approximately 30 m from the building. A valve at the intake of
the 2-in. aluminum tubing provides resistance to the flow of ambient air
to keep the Model 8700 DMP "Tornado" blower at a stable rpm, and to serve
as a gross flow control. A Variac inside the building serves as an opera-
tional flow control. The blower is located at the input end of the system
to provide positive pressure in the system. It is located outdoors to
keep out the intensive noise it generates and is housed as shown in
photographs 1 and 3 of Figure 3 to protect it from the elements.
The line from the blower to the splitter is 2-in. diameter, alumi-
num pipe. It is sufficiently long to serve as a trap for any excess
moisture and to bring the ambient air to room temperature. The splitter
is also made of aluminum. This splitter, shown in Figure 4, reduces
large-scale turbulence from the blower and divides the ambient air
stream between the spiked and unspiked 1-in. diameter, aluminum lines.
A controlled flow goes to the venturi where the air flow in the spiked
line is continuously measured and recorded. This flow is determined by
the following equation:
method sampling rate (number of samples x number
_, ... . . of collaborators + monitor number + purge number)
Flow in liters/min = — —*—e , ,.
percent flow drawn through sampling manifold
0.2 liters/min x (4 samples x 10 collaborators +
_ 1 N0/N02 monitor + purge-line flow)
percent flow drawn through sampling manifold
= 0.2 (4 x 10 + 1 + 4) _
0.15
-------
oo
Diffuser Equilibration 45-Port
Purge Control
Line Valves
NO2 Permeation
Tube System
Monitoring Points:
1. Flow Temperature
2. Flow Pressure at Input
to Flow Meter
3. Ambient Air Flow
4 . Flow Temperature 10. NO2 & NO
5. Pressure Drop of Venturi & Temperature 11. NO2 & NO
of Pressure Transducer
6. Carrier Gas Flow
7. NO2 Flow Temperature
8. Port. Pressure
9.. Port Pressure
Notes:
1. Component within Dashed Area
Made of Teflon
2. Piping Out Side Dashed Area made
of Aluminum
3. Venturis made of Stainless Steel
4. Spiked & Unspiked Lines Symmetric with
Respect to Geometry & Material
Figure 2. Final design of the N02» ambient-air sampling system
-------
Icrl Vmlurl (tj. NO;
Figure 3 - Annotated Photographs of the NC^, Arabient-Air Sampling System in Operation
-------
Ambient Air
Figure 4. Ambient-air stream splitter
-------
The monitor number and the purge number are flows attributable to the
continuous monitor and the purge line of the system, respectively. The
flow on each line—the spiked and the unspiked--is turbulent—Reynolds
No. > 2,100--with the Reynolds number being
_ Q _ Q liter/min x 1.000 cm3/liter
0.785 x 0.15 cm2/sec x D sec x 60 sec/min
1.000 Q 1.000 x 60
7.065 D 7.065 x 2.1 H»uuu<
Since the spiked and unspiked sections are identical except that
the spiked section also contains the monitoring points 1, 2, 3, 4, and
5 identified in Figure 2 and the NC>2 permeation tube system, only the
spiked section will be discussed.
From the splitter, the spiked line connects to a Singer AL-175 dry-
gas meter, which is made by the American Meter Company. (See photograph
9 of Figure 3.) This flow meter has a pressure drop of 10 mm of water
and is temperature-compensated. Thus, only the gas pressure is measured
to correct the flow readings to obtain the true flowrate of ambient air
delivered during a test run. This flowrate is determined hourly by mea-
suring the time required for a known quantity of air to pass through the
meter.
The output of the flow meter is connected, as shown in photographs
7 and 9 of Figure 3, to a stainless steel venturi, which was designed
for a flow of 60 liters/min. This venturi is used as a general flow con-
trol device, and provides a continuous record of flowrate using a strain
gage pressure transducer and thermocouples—see point 5 of Figure 3(A).
Both the pressure drop of the venturi and the temperature of the pressure
transducer are recorded on analog recorders. Control of the flowrate is
handled by monitoring the venturi pressure drop. When the value deviates
from a reference value, 60 liters/min, the flowrate can be changed ap-
propriately by making an appropriate adjustment of the Variac control
to the blower.
The flow temperature measurement (point 4 of Figure 3(A)) is actually
the gas-flow temperature at the output of the gas meter and at the input
to the venturi, since those two units are physically close together (about
12.5 cm apart). Tests have shown that the temperature at this point is
identical with the temperature at the gas flow meter inlet. The gas tem-
perature at this point is normally within 5°C of room temperature. This
temperature measurement is used to obtain accurate gas-flow values.
12
-------
To provide more accuracy, the thermocouples at points 2 and 4 of
Figure 3(A) were replaced for this test by a 0-50°C bimetallic dial
thermometer that is located at point 4 of Figure 3(A).
The output of the venturi is a few centimeters from the input of
the N02 bleed-in unit as shown in photograph 9 of Figure 3. These two
units are connected by 10-in. diameter, aluminum tubing. From the input
of the N02 bleed-in unit through the sampling manifold, the system is
made of Teflon.
The N02 bleed-in unit, as shown in Figure 3(A) and photographs 7
and 9 of Figure 3, receives ambient air from the venturi and a level of
N0~ (a spike) from the N02-perraeation tube assembly (see Figures (A) and
(B), and photographs 8 a-c of Figure 3). Detailed photographs of this
bleed-in unit are given in Figure 5. Photograph 1 of Figure 5 is a
closeup showing the assembled Teflon unit with its metal holding/mount-
ing plates. The gas stream, or ambient air, enters the opening to the
right and passes through the unit, mixing with the spiked level of N02
which exits through the tapered smaller tubing shown as concentric to
the output of the bleed-in unit at the left of photograph 1 of Figure 5.
The vertical tube of this bleed-in unit accepts the N02 gas from the
permeation tube assembly. This spiked gas flows downward through this
tube, which is inside the unit (see photograph 2 of Figure 5), and after
a short run, mixes with the ambient air as stated before.
The N02-permeation system is shown in Figure 6 and photographs 8
a-c of Figure 3. Details of the system are given in the captions of these
photographs. The nitrogen carrier gas is used to flush the N02 into the
system. It is passed through a charcoal and soda-lime scrubber before it
is delivered to the N02-perraeation tubes. Also, the flow is set by means
of control values and rotameters. This flow is monitored during system
operation. The carrier gas is then fed into four separate branches to
achieve different levels of N02. (More detail on the permeation tubes
and their arrangements in the branches is given in Appendix B.) The N02~
permeation tubes* are arranged in these four different branches to provide
N02 spike levels of approximately 50, 100, 200, and 300 ug/m^. Branch 1
has four permeation tubes, branch 2 has five permeation tubes, branch 3
has two permeation tubes, and branch 4 has two permeation tubes. An ASTM
calibrated thermometer (0.1°C or better accuracy) is an integral part of
"Operation Characteristics of N02 Permeation Device,11 by Harry L. Rook,
Ernie E. Hughes of NBS, Washington, D.C., and Robert S. Fuerst and
John Margeson of EPA, Research Triangle Park, North Carolina. A paper
presented before the Division of Environmental Chemistry, American
Chemical Society, Los Angeles, California, 31 March - 5 April 1974.
13
-------
Photo 1 - Detail of NC>2 Bleed-In Unit with Vertical
Tube from Permeation Assembly, Chamber with Central
Tapered Pickup Tube and Stainless Steel Mounting
Components.
Photo 2 - Close-Up Showing Machined Chamber with
Pickup for Bleed-In in Place.
Figure 5 - Photographs of the NCL Bleed-In Unit-
Assembled and Disassembled
14
-------
Nitrogen Gas
Shutoff Valve
Charcoal & Soda Lime Filter
Control Valves
Rotameters
NC>2 Permeation
Tube Holders
1 Thermometers
^ - Temperature Control led
Water Bath
III!
Control Valves
to NO2 Bleed In Unit on
Spiked Line (See Figure 3)
Figure 6. Schematic drawing of the N02 permeation tube assembly
15
-------
each permeation tube branch. Each set of permeation tubes is enclosed
in a glass tube which has an inlet for the nitrogen carrier gas and an
outlet for the nitrogen carrier gas/NC^ mixture. These NC>2 permeation
tube, enclosure units are immersed in a temperature-controlled, water
bath for operation at 25.1°C. If the temperature of this bath were to
vary more than 0.2°C, a correction would be made from the following
relationship:
Log r = 0.034857 (273.12 + T) - 10.29198
where T = temperature in °C of the permeation tube environment, and
r = the permeation rate.
Flow meters of the permeation tube assembly that measure the
nitrogen flow were calibrated by the manufacture to 1% accuracy.
Thermometers that were used to measure the gas temperature in the
permeation tube holders are ASTM type that are accurate to within 0.1°C.
The permeation tubes used were calibrated by the National Bureau of
Standards and checked by EPA. (See Appendix B.) The entire permeation
assembly from the tube holders to the pickup fitting, where the spiked
gas enters the main gas stream, was checked for leaks with Snoop and
found to be airtight.
The ambient air and the NOo flow from the bleed-in unit to the dif-
fuser where they are well mixed. The diffuser is a few centimeters down-
stream from the bleed-in unit, as shown in photograph 9 of Figure 3. At
the diffuser, shown schematically and in the photograph in Figure 7, the
gases enter the diffuser through the Teflon tubing (Section A of the
schematic drawing of Figure 6), pass into the spiraled tube and through
its angled holes into the space outside the tube. The flow continues
through the holes in the prescreen block, item D of the schematic, and
then through a series of Teflon screens, item E.
The homogeneous mixture passes through an equilibration section that
is Teflon tubing 1.0 m long. This section of tubing provides the final
equilibrated concentration. This tubing is connected to the input of the
sampling manifold. (See photographs 10 and 11 of Figure 3.)
The 45-port sampling manifold is constructed of Teflon except for
its metal plates which are entirely external. Photographs 1-3 of Figure
8, which show external and internal views of the sampling manifold and
a schematic drawing, describe the operation of the manifold. The stream
of the homogeneous mixture of ambient air and a spiked level of N0~
flows through the bottom portion of the manifold, into the exhaust line.
16
-------
Photo 1 - Top View Diffuser Components: Housing, End
Sections, Spiraler Tube, Teflon Screens, Retaining
Rings.
Photo 2 - External View of Diffuser.
Him
rrR.
7M
a
7L.
TIL
Exploded cross section of all-Teflon diffuser with inlet (A),
end section (B), spiraler tube with angled holes (C),
prescreen block with holes (D), five sets of fine mesh
Teflon screen and retaining blocks (E), end section (F),
exit (G) and diffuser housing (H). Double cross-hatched
end plates are stainless steel.
Figure 7. Schematic Drawing and photographs
of the diffuser
H
-------
Photo 1 - Sampling Manifold External View.
Photo 2 - Internal View (Right Component is Inverted
in this Photo).
00
Cross section of all-TFE Teflon manifold with pickup tube (A), mixing impeller (B), main chamber flow
spreader (C), exits to collaborator ports (D), channel to exhaust manifold (E), and manifold exhaust
duct (F). Gas not captured by pickup assembly exhausts at left side of manifold base. Double cross-
hatched assembly plates at top, middle, and bottom are stainless steel.
Photo 3 - Internal View of Manifold Pickup Section
Showing Flow-Spiraling Impeller.
Figure 8. Schematic drawing and photographs of the sampling manifold
-------
Section A of the manifold is in the pickup tube through which flows the
total volume of gas sampled by the collaborators. The inlet to this
pickup tube is located such that this volume is drawn from the central
portion of the main stream. The sampled volume flows past a mixing im-
peller (B) and then into the main chamber of (C) of the manifold. In
this chamber, the flow is spread evenly to the 45 symmetrically located
exit channels (D). The gas in the main chamber that is not drawn through
the exit channels to the collaborator ports flows out the exhaust duct
or purge line which has a control valve. Both exhaust lines from the
manifold join downstream to form a common exhaust tube, which also con-
tains a flow control valve.
One port of each sampling manifold is used to monitor the pressure
in the sampling manifold to determine if it remains positive (see
schematic drawing of Figure 3). Another port of each manifold is used
to monitor the N(>2 and NO levels being sampled by the collaborators. A
Bendix Model 8101 B chemiluminescence NO-N02-NOX analyzer is used for
this and is switched between the spiked and unspiked manifolds. (See
photograph 12 of Figure 3.)
SYSTEM CHECKOUT
Readying the system for the collaborative test comprised three
principal areas of activity: (1) determination of levels of NO and N02,
both ambient and inside the building; (2) checkout of the sampling sys-
tem, including monitoring devices and test instrumentation; and (3)
checkout of the sampling system as an operational system. These three
areas are discussed below.
Ambient Levels of NO and NO2
Ambient levels of nitrogen oxides at the test site were generally
low, but there were considerable variations at these levels. Since the
test site is located in a rural area south of Kansas City where there is
very little industry, the primary factors that influence NOX levels at
the site are wind speed and wind direction.
During tests of NOX levels using MRI's Bendix Model 8101 B cherai-
luminescence NO-N02-NOX analyzer for 24-hr monitoring, the lowest levels
were found when the wind was from the south. Both NO and N02 seldom ex-
ceeded 20 p,g/nr*. Periods of more than 1-hr duration were measured when
readings were indistinguishable from the purified zero gas used to cali-
brate the analyzer.
19
-------
With northerly winds, N(>2 levels were generally between 30 and 50
and NO levels were approximately 10 pg/m^. As expected, the ambient
levels followed an inverse relation with respect to wind speed. The high-
est daily readings were coincident with the morning and evening rush
hours. These peak levels generally began at about 7:00 a.m. and again at
5:00 p.m. and lasted between 2 to 4 hr.
The highest recorded levels of NO occurred under calm wind con-
ditions when the light vehicular traffic in the vicinity of the test
station generated levels in excess of 100 g,g/m3. NO levels did not exceed
N02 levels at this site.
Over a 24-hr period, average N0£ levels were 10-50 ug/m^, and NO
levels were of the order 10-20 ug/m-*. During any 24-hr period, maximum
NC>2 levels were generally several times higher than the minimum levels.
Thus, while NC>2 levels at the test site are lower than those at urban,
industrial locations, the N0£ levels do exhibit the variability found
under normal ambient conditions. Indoor readings were similar but did
not show the sudden changes often found when monitoring outdoor levels.
Subsystems and Units
The venturi and dry-gas meter were calibrated using a 1.0 ft^/min
wet-test meter. Information concerning the calibration is given in
Appendix C. The entire system was prepared for the test by bringing all
components to normal operating conditions several days prior to the test
and running the system continuously in this mode until the beginning of
the test. Water addition to the constant-temperature bath was the only
maintenance required. The temperature variation of the permeation-tube
bath during this time was less than 0.1°C. A check of NOX levels in the
cylinders of prepurified nitrogen carrier gas found no N02 and 40 ug/nr
NO.
The Bendix NOX Analyzer was checked at MRI by a Bendix field rep-
resentative. The difference in spiked and unspiked readings of the Bendix
analyzer agreed within 10% of the calculated spike levels at all four
levels used for the test. The instrument was stable and reliable when
operated continuously at the levels found during normal testing. Checks
with calibration gases reveal that the catalytic converter efficiency
does fall off sharply above 400 ug/m .
The symmetry of the sampling ports was checked in two ways. The
primary way was that the pressure drop at each port was measured under
the normal load of 200 cc/min. This test showed that all ports gave a
pressure drop of 1.5 cm of water + 0.5 cm. Such a pressure drop should
have no effect on normal sample flows and the flowrate from the ports
should be identical to that obtained by pulling free room air into the
sampling trains.
20
-------
A second way was to connect the N(>2 monitor to ports of the spiked
and unspiked sampling manifolds and measure the level of N0£ in micro-
grams per cubic meter. This was done in two ways: the system under a
load, e.g., a spiked level of approximately 350 ug/m^; and an unloaded
condition where just ambient air was passed through each section—
spiked and unspiked—of the NO2 sampling system. In both cases, the N02
monitor showed no variation between four ports spaced equally around
the manifold.
System Operation
Identical materials and dimensions are used on the spiked and un-
spiked sections of the N02 sampling system. Handling and treatment of
all components were also identical.
Flowrates of the spiked and unspiked sides were within 10% of each
other with all dampers open. In normal operation, the exhaust dampers
are adjusted to give a positive pressure of 2-4 mm water at the sample
manifolds. Once set, this pressure is stable.
The rise and fall times to equilibrium in response to changes in
a spike level were checked. Rise time was less than 15 min and fall time
was less than 5 min (when permeation tubes were disconnected). The fall
time is essentially that of the analyzer response time, allowing for the
purge time of the sample lines. The rise time is longer than the fall
time because of the increased pressure against which the carrier gas
stream must work when a set of permeation tubes are connected. Some flow
reversal in the permeation-tube holders occurs after connection.
Since the response times were essentially limited by flowrates and
instrument response, no observable adsorption effects were noted. Checks
of NOX levels found at the sampling ports agreed, within normal accuracy
limits, with those measured outside the building. At the 50 jig/m^ level
both readings were within 5 pg/m (0.5% of full scale), which is within
the accuracy of the monitoring instrument. Thus, the unspiked samples
at the 'sample ports accurately reflect ambient levels and the sampling
system may be considered to be inert with respect to
21
-------
TEST SITE
The general criteria one would use in selecting a site include the
ambient level of N02 and variation thereof, general meteorological and
climatological conditions, work facilities for the collaborators (ade-
quate space, facilities, housing, etc.), cooperation of the organization
furnishing the site, logistic aspects, and local lodging accommodations.
The levels of N02 required are those representative of ambient N02
conditions, which are in the range of a few micrograms per cubic meter
to 350 jjbg/m^. These levels could be achieved at one site with a low
level of N02 by spiking the ambient air with various levels of N02 in
a manifold sampling system.
MRI's field station (see Figure 9) which is located in a rural area
south of Kansas City, meets all the criteria and was selected as the
test site. The N02, ambient-air sampling station is housed in building 3
shown in Figure 9. The input to the sampling system is located outside
the building near the roadway (see photograph 3 of Figure 3).
The test facilities are described in conjunction with the sampling
system (see pages 7-19). Photographs of the facilities are given in
Figure 10. Photograph 1 shows the circular tables that house the sam-
pling manifolds and the collaborators' sampling trains. Each table--
spiked and unspiked--has a multiplicity of AC power receptacles, with
each collaborator having its own branch of outlets. Each branch has its
own circuit breaker and branch indicator (see photograph 2 of Figure 10).
This arrangement is to protect other collaborators in case one collabora-
tor has a power failure due to faulty equipment.
Photograph 3 of Figure 10 gives a close-up view of some of the col-
laborators' trains positioned in their table areas (see Figure 11).
Photograph 4 of Figure 10 gives a view of part of the bulletin board
where test instructions and general information was posted.
23
-------
NJ
-P"
DERAMUS FIELD STATION
Figure 9. Collaborative test site: MRI's field station
-------
Photo 1 - Unspiked Table In the Forefront; Spiked Table in the Rear.
Photo 2 - AC Power Control and Visual Indicator;
the Boarded Windows to Keep Out EM Radiation.
ro
Photo 3 - Collaborators' Sampling Trains on Spiked Table; Spiked Sampling
Manifold on Shelf above Table.
Photo 4 - Collaborators' Bulletin Board with Test
Requirements, Gereral Information, etc.
Figure 10. Photographs of the test facility
-------
SPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS *
* For unspiked manifold—collaborator areas marked in reverse
order--counter clockwise.
Figure 11. Collaborators' sampling areas at the test site
26
-------
Photographs 1, 2, and 3 of Figure 10 show that the windows on the
north side of the building were boarded to keep electromagnetic radia-
tion from entering the building. With this blockage and a temperature
control system in the building, the 25.1°C permeation bath was able to
be maintained at that temperature throughout the four 24-hr runs with
no detectable deviation from the 25.1°C temperature, except for a few
hours when the deviation was 0.1°C.
27
-------
SELECTION OF COLLABORATORS
A principal activity was to compile a list of potential collabora-
tors and from this list select 10 to perform the testing according to
the tentative TGS-ANSA method. Information was obtained from EPA (names
and addresses of 150 organizations) and from MRI's files to compile a
list of nearly 200 potential collaborators.
A letter was sent to 162 organizations seeking their desire to par-
ticipate as a volunteer collaborator on this test and evaluation pro-
gram. Attached to this letter was a "Collaborator Form" to be completed
which surveyed their experience with the four methods, methods they had
used, equipment they could make available for the tests, acceptable length
of test period, etc. A second letter was sent to those who expressed in-
terest in the TGS-ANSA method after a test date was selected. A copy of
these letters and the collaborator form are given in Appendix D.
A majority of the responses indicated the desire that a test period
for a method be no more than 6 days.
Ten organizations were selected for the TGS-ANSA collaborative test
from those organizations that responded in the affirmative to partici-
pate in the test. The selection was based upon the following criteria:
1. Willingness to participate on a volunteer basis,
2. Technical capabilities,
3. Related past experience,
4. Availability,
5. Ability to furnish sampling equipment, instruments, and mate-
rials required to perform the test strictly according to the method, and
6. Type of organization (industrial, educational, governmental—
local, state, federal—etc.).
29
-------
The information needed to make the selection based on the above criteria
was obtained from the collaborator forms that were returned, and from
subsequent telephone conversations with the candidate collaborators.
The 10 organizations selected as collaborators for the TGS-ANSA
collaborative test were:
Kentucky Division of Air Pollution
311 East Main Street
Frankfort, Kentucky 40601
502-564-3382
(Ms. Diana Dunker, Mr. Joe Andrews!*!')
Shell Development Company
P.O. Box 481
Houston, Texas 77001
713-667-5661
(Mr. J. H. Bradley, Mr. W. T. Shebsiti/)
City of Philadelphia Air Management Services Laboratory
1501 East Lycoming Street
Philadelphia, Pennsylvania 19124
215-288-5117
(Mr. Donald Kutysiii/)
North Ohio Valley Air Authority
814 Adams Street
Steubenville, Ohio 43952
614-282-3098
(Mr. Frank G. Norris, Mr. Dan Zorbiniia!/)
State of New York
Department of Health
Division of Laboratories and Research
New Schotland Avenue
Albany, New York 12201
518-457-3118
(Ms. Barbara Kladatos1!2/)
\J These individuals performed the sampling at the field site.
2J These individuals performed the analyses of the samples.
30
-------
Kennecott Copper Corporation
Utah Copper Division
P.O. Box 11299
Salt Lake City, Utah 84111
810-322-1533
(Dr. Robert J. Heaney, Mr. Lynn A. Hutchinson— »•=•')
San Bernardino County Air Pollution Control District
172 West Third Street
San Bernardino, California 92415
714-383-1661
(Mr. Mark Villalabbos, Dr. C. Kenneth
Mecklenburg County Department of Public Health
1200 Blythe Boulevard
Charlotte, North Carolina 28203
704-374-2607
(Mr. James T. Wardlii/)
Air and Industrial Hygiene Laboratory
State of California Health and Welfare Agency
Department of Health
2151 Berkeley Way
Berkeley, California 94704
415-843-7900
(Mr. Emil R. de Vera, Mr. Kenneth
National Bureau of Standards
B 326 Chemistry Building
Washington, D.C. 20234
301-921-2886
(Dr. John K. Taylor, Mr. Bob Deardorf fill/)
These organizations will be referred to as Collaborators A through
J, without defining which is A, B, etc., to allow the organization data
to remain anonymous.
31
-------
STATISTICAL DESIGN
GENERAL CONSIDERATIONS AND COMMENTS
The purpose of this collaborative test was to determine the preci-
sion and bias of the TGS-ANSA method. A major element of the collabora-
tive test was to have an experimental design that would allow this purpose
to be met. Considerations that formed the bases of this design, which is
given later in this section in a formal manner, are:
1. Challenge (spike) levels of N02,
2. Ambient levels of N02,
3. True values of N02,
4. Sampling time of a run,
5. Test period of the method,
6. Number of collaborators,
7. Number of samples per run,
8. Interferences,
9. Adsorptivity,
10. Sampling ports, and
11. Instrumentation.
Challenge (spike) level of NO2 is an experimental design variate.
Four levels of challenge were selected based upon the normal range of
values found in ambient air on a 24-hr average basis: one low level on
the order of 50 ug/m3; two medium levels, one near 100 ug/nr and the
second near 200 p,g/nr; and one high level of approximately 300
33
-------
A challenge level should be steady state, or continuous at a specific
level, plus or minus acceptable deviations—less than + 2%.
Ambient levels should be lower than the lowest challenge (spike)
level (~ 50 p,g/m ). Since the ambient levels are the actual ambient levels
of NO2 at the test site, those levels present during the time of testing
may vary from this criteria (see pages 19 and 20) . The ambient levels will
be mixed with the challenge levels to provide the spiked challenges. There
will be just ambient challenges which are identical with the ambient por-
tion of the spiked challenges. The collaborators will sample both spiked
and ambient challenges simultaneously.
For a run, the true value of NO2 sampled by the collaborators will
be taken as the NO2 spiked level generated by the permeation tube as-
sembly plus the average value of the ambient challenges sampled by the 10
collaborators. The error involved here adds to the overall error in the
analysis.
The TGS-ANSA method requires a sampling period to be 24 hr.
From the survey for volunteer collaborators, it was determined that
6 days would be the limit for a test period of a method. Thus consid-
ering this period, the mandatory 24-hr sampling period (or a run) travel
time, and orientation, set-up and switch-over time (time in between runs),
four runs would be the maximum possible.
Ten collaborators were deemed to be sufficient to obtain a cross-
section of the population of the type organizations that would be involved
in sampling N02» be within acceptable project costs, and provide statisti-
cal significance with the results.
Replicate samples are desirable and generally needed. In this test,
replication is constrained by the test period and the duration of a run,
and thus any replicates must be of the nature of simultaneous sampling
by collaborators using as near identical trains as possible. This type
replication, in turn, has constraints, which include principally the
number of collaborators, space limitations at the test site, size of
the NC>2 sampling system and cost limitations. Naturally some of these
are interdependent.
An important consideration for the TGS-ANSA method is that of in-
terferences. These factors will vary depending upon geographic location,
time of year, etc. The interference consideration was not included in the
experimental design, since work had shown that the method did not have any
known interferences.*
See Section 3 of Appendix A.
34
-------
Adsorptivity is of concern because of the possibility of error in
the N02 level received by the collaborators' sampling devices in contrast
to the known level of the challenge—from both the standpoints of in-
creasing and decreasing the challenge level from run to run. Teflon mate-
rial was used from the NC^ bleed-in port through the sampling manifold
to minimize if not eliminate the adsorptivity factor. For further
assurance, prior to commencing a run, the challenge could be run for a
sufficiently long period so that all surfaces exposed would have reached
a state of equilibrium with the new concentration. Both aspects were
covered; Teflon was used in the construction and sufficiently long
challenges were made to the system prior to commencing a run.
The port-to-port effect did not need to be incorporated in the ex-
perimental design because results of the evaluation of the NC>2, ambient-
air sampling system indicated that all ports were identical (see first
paragraph of page 21).
The major considerations with regard to instrumentation for the TGS-
ANSA collaborative test were: (1) MRI would only instruct the collabora-
tors that they are to use the sampling equipment and calibration speci-
fied in the method writeup, and (2) MRI's monitoring instrumentation and
test instrumentation used in the calculation of the NC^, ambient-air
system was sufficiently reliable and accurate. In both cases, all •re-
quirements were met.
THE FORMAL DESIGN
The N02 data are collected according to a two-way analysis of
variance model with the analysis of primary interest being estimation
of the components of variance and the bias. Thus, the mean square errors
can be constructed.
Specifically, we have:
Xljk = n + Ct + Lj + CLi:j + ek(ij)
where (j, = overall mean,
Ci = ith collaborator (1=1,... 10; C± is a random
factor),
LJ = jth level of N02 (j = 1, . . . 4; Lj is a fixed factor),
CL = collaborator-level interaction,
ek(ijj = error term (k = ! 4 v ij)»*
xiik = k*"*1 response observed by ith collaborator on j*-*1 level.
* V means "for every."
35
-------
The expected mean squares are:
E (MSC) = ae2 + 16 ac2, E (MSL) = afi2 + 4 CTCL2 + *0* aL2, E (MSCL)
= „ 2
= CTe + 4
and
E (MSe) = ae2
Although the F-tests for significant effects were performed for com-
pleteness, the primary object is the components of variance analysis; in
particular, the components ae (variance of repeated observations) and
ac (variance between collaborator means).
Let
NJS = spiked level j of NOgi
Nijk = kth observation by i" collaborator at level j of N02,
A.J = average of collaborators' observations of ambient NO2 at
level j,
then
Xijk = iJkth response = (N^k - A.j) - NJS.
Therefore, X^^ is the ijkth deviation from the true (bias) subject
only to the error in using A.J as the true ambient level. It would, of
course, be desirable to have no error at all in the true values, but the
above method of estimating bias is preferable to no estimate at all. Thus,
we produced mean square error (per collaborator, if necessary) estimates
in addition to variance estimates.
In any analysis of variance the homeoscedastic assumption must be
validated. It was found that the measurement error was relatively uniform
throughout the experiment, i.e., not a function of the level of N02«
Therefore, no data transformation was made (see page 46 for detailed dis-
cussion).
Outliers were deleted before analysis, and the frequency of them
noted. One collaborator produced useless results (see Appendix H). There-
fore, the results are reported using nine collaborators.
* Collaborator I was deleted from the actual analysis (see page 41), so
E (MSL) = ae2 + 4 acij2 + 36L2.
36
-------
COLLABORATORS' FIELD SAMPLING
The collaborative test took place at the MRI Deramus Field Station
during 29 April - 3 May 1974. The 10 collaborators (see Figure 12),
started the test at 0830, 29 April, with an orientation. The N(>2,
ambient-air sampling system they used was shown and explained to them.
The written instructions that comprise Appendix D 'were given to and
discussed with the collaborators. After this orientation period the
collaborators set up their sampling trains in preparation for the
first run. They were ready to start sampling at 11:20 a.m., 2 hr ahead
of schedule. The actual schedule of the four runs that took place is
given in Table 1. All 10 collaborators cleared the site by 1500, Friday
afternoon, 3 May.
Table 1. TGS-ANSA COLLABORATIVE TEST SCHEDULE
NO Spiked Level Date/Time
Run ( g/m ) Started Completed
1 65.6 4-29-74 at 1120 4-30-74 at 1120
2 117 4-30-74 at 1207 5-1-74 at 1207
3 223 5-1-74 at 1239 5-2-74 at 1239
4 315 5-2-74 at 1304 5-3-74 at 1304
Each run was 24 hr in duration. The collaborators were at the site
from 2 to 3 hr prior to the start of the run, or completion of a run.
They stayed through the change over to the subsequent run and from 1 to
2 hr thereafter.
During the test MRI personnel observed that all collaborators fol-
lowed the sampling procedures given in the method writeup, and their equip-
ment met the requirements presented therein.
The collaborators prepared their absorbing solutions at their home
laboratories, to minimize biasing.
37
-------
Front Row: Bob Deardorff, Lynn Hutinson, DonZorbini, Kenneth Wilcox,
James Ward, Barbara Kladatos, Paul Constant _!_/ Donald Kutys
Back Row: Bob Stuttz _L/, George Scheil -I/, Kenneth Smith, Fred
Bergman _L/, Joe Andrews, John Margeson _2_/ W.T. Shebs
-L/ MRI personnel.
-2/ EPA Project Monitor.
Figure 12. Photograph of field personnel of the NC>2 collaborative test
(TGS-ANSA Method); MRI field station; 29 April - 3 May 1974
test
38
-------
Field data were recorded in duplicate by the collaborators on data
sheets designed by MRI for this test (see Figure 13). A copy of this
data sheet was collected from each collaborator after the completion of
a run, and before a subsequent run would be started. The collaborators
retained their copies for recording subsequent analysis work at their
home laboratories.
During 23-25 April—just prior to the start of the test on Monday,
29 April—MRI prepared standard samples. These samples were drawn from
the spiked line (first day) and unspiked line (second day) into absorb-
ing solutions which were contained in 200-ml impingers. The solutions
from all impingers were mixed for each day to provide a homogeneous
sample. Individual samples were then prepared for the collaborators and
MRI. Each collaborator was given two N(>2 sample and one blank (absorbing
solution only) to be analyzed at his home laboratory along with his
test samples. MRI followed the method in preparing for sampling and
sampling with the exception that the trains were scaled upward.
The samples taken by the collaborators, as well as MRI's standard
samples given to them, were either taken with them when they returned
home, or shipped to them by MRI. The samples were shipped via Air Mail
on different days to insure against loss of all samples of a collabora-
tor in case a shipment was lost or destroyed.
MRI had a laboratory supervisor who was in charge of the NC^j
ambient-air system operation. He was on duty from 0800 to 1700 each day,
which was the period of run starts and completions. He was available
anytime during the 24-hr runs, if any problems arose, as was the pro-
gram manager.
There was a technician on duty throughout each run at all times
during the test. These people monitored the sampling system operation,
recording operational data and general observations. A general log book
was kept as well as the log sheet for operational data. Copies of these
log sheets are given in Appendix F.
39
-------
MRI DERAMUS FIELD STATION - K.C. MO.
NITROGEN DIOXIDE DATA SHEET - SODIUM ARSENITE METHOD
Sampling
Collaborator Run Number
Sampled by Sampling Port Number
Sampling Train No.
Bubbler Identification No.
Rotameter Description (make, model, etc.,)
Rotameter reading at start at finish type of ball
Start:Date Time Finish:Date Time _
q
Sampling duration (min) Sample flow rate (cnr/min) _
q
Total air volume sampled (m )
Remarks:
Analysis
Date of Analysis
Analyzed by: Person Organization
Standardization plot slope (absorbance units/ug N02/ml)
Absorbance of sample against blank (540 nm) Aliquot (ml)
N0£ Concentration
Remarks:
Figure 13. Nitrogen dioxide data sheets--TGS-ANSA method
40
-------
ANALYSES OF SAMPLES
This section discusses the analyses performed by the collaborators
and by MRI. The collaborators' analyses were of the samples they took
during the test from both the spiked and unspiked lines, and the stand-
ard samples prepared by MRI and given to them. For each of the four test
runs in the field, each collaborator had four spiked samples and two
unspiked samples. In addition, each collaborator had three standard
samples--one blank and two N02» System operational data logged by MRI
during the four tests are given in Appendix F.
ANALYSES PERFORMED BY THE COLLABORATORS
The collaborators performed the analyses of their samples accord-
ing to the procedures given in the TGS-ANSA method, with one exception:
Collaborator I reported an extremely low slope for his calibration curve
and his results were low by approximately a factor of 10. Further informa-
tion was then obtained which revealed that the error was apparently
caused by contaminated methanol, which was in violation of the method
specifications. The collaborator reported that the dye did not dissolve
and had to be filtered before use. Previous use of this methanol had not
exhibited this behavior. A sample of the methanol was subsequently sent
to MRI. The liquid was clear and colorless with a small amount of white
gelatinous precipitate. The methanol was in a fresh 5-gal. metal can,
not used for other purposes. Apparently ANSA prepared before the collab-
orative test was taken from the relatively uncontaminated top portion of
the can, while the methanol used later was from the more contaminated
bottom of the can.
In all cases, the collaborators' representative who performed the
field sampling also performed the analyses of the samples at the col- <
laborators* home laboratory. The collaborators' comments on the test are
given in Appendix G.
41
-------
Table 2. COLLABORATOR RESULTS FROM COLLABORATIVE TEST USING THE TCS-ANSA METHOD
*•
ro
•
Run 1
Spike = 94 ug/m3 Spiked
Unsplked
Run 2
Spike = 197 ug/m3 Spiked
Unspikcd
Run 3
Spike = 54.3 ug/m3 Spiked
Unsplked
Run 4
Spike = 302 ug/m3 Spiked
Unsplked
Standards (ug/ml)
(Mill's values = 0.004)
= 0.64
= Blank
A
102.40
110.86
107. 10
104.67
16.18
16.91
225.95
224.89
209. 18
212.75
19.45
20.51
76.42
63.17
59.85
57.31
5.84
5.11
307.69
327.52
298.30
301.99
9.82
9.73
0.025
0.645
0
MRI
105
113
108
105
19.1
19.8
229
229
211
216
22.3
23.6
79
67
62
60
9.0
7.2
308
328
298
302
10.9
10.3
_i_
127
110
99
128
28
25
189
194
201
192
26
24
MRI
127.0
112.2
101.4
134.1
27.1
24.7
190
196
202
200
25.2
23.5
ZB2' 27.2
50
58
58
12
16
252.
211
272
288
17
17
0.
0.
< 0.
48.1
57
56
11.8
14.7
, 254
' 211
274
285
14.3
14.2
05
55
02
C MRI
115 112
108., 106
30^ 26
120 118
29 24.3
32 28.8
197 195
203 200
188 184
207 203
25 24.6
30 25.6
73 69
62 58
55 60
65 60
15 10.5
21 13.9
288 284
222* 21.6
295 293
291 288"
15 10.4
11 12.7
•
0.05
0.66
. 0
D
127.3
119.5
122.5
123.9
27.5
27.8
234.3
225.5
227.0
230.9
27.7
27.7
67.2
66.3
64.4
64.2
14.3
14.6
330.7
327.7
323.2
327.3
12.6
13.2
0.047
0.672
0.015
MRI
127
120
122
124
27.5
27.8
235
226
227
231
27.8
27.7
67
66
64
64
14.3
14.7
331
328
323
328
12.6
13.2
E
116.0
117.1
127.4
114.8
25.3
25.2
213.0
218.4
219.0
. 216.2
25.2
24.2
60.6
61.0
63.2
60.2
15.0
14.5
315.8
314.4
315.1
312.5
11.9
12.9
0.048
0.641
0.016
MRI
116
117
128
115
25.3
25.2
213 .
218
219
216
25.2
24.7
61
61
63
60
15.0
14.5
316
314
315
313
11.9
12.9
F
108.5
109.9
115.9
85.1
26.1
27.6
195.9
196.8
195.7
175.3
24.5
24.1
61.1
59.0
58.3
56.0
13.2
13.5
274.7
270.5
289.1
267.4
15.7
16.5
0.
0.
0.
MRI
109
111
117
86
26.3
27.7
197
198
199
176
24.7
24.2
61
59
59
56
13.1
13.6
276
"2 .
291
269
15.9
16.4
073
63
032
c
116.1
116.1
117.8
122.6
29.9
43.1
218.8
216.5
225.2
220.8
27.5
29.1
52.3
59.4
50.8
59.4
14.0
12.5
298.9
298.8
306.6
308.4
l*.l
15.1
0.048
0.596
U
MRI
115
118
117
121
29.6
42.6
217
214
223
219
27.2
28.8
52
59
60
59
13.8
12.4
296
296
304
305
15.9
14.9
H MRI
i"8. 11 106
55. 982-/ 55
102.17 100
S9.5&2/ SB
17.42 16.8
17.73 17.3
214.30 210
227.10 224
233.34 228
214.33 211
22.74 22.7
17.23 16.9
44.55 43.5
50.99 49.9
40.28 39.6
44.09 43.1
4.71 4.8
4.71 4.9
301.04 294
296.43 290
307.95 301
307.87 300
6.06 5.8
6.12 5.9
0.039
0.684
0
u>
3.19
10.72
4.68
3.90
3.90
3.23
23
21
6.47
19.41
1.92
2.15
14.47
13.97
3.23
4.31
1.01
5.39
51.77
39.00
47.6
42.09
3.90
2.27
0.075
0.531
0.016
MRI
3.4
11.5
5.0
4.6
4.2
3.5
24.2
22.5
7.0
20.8
2.1
2.3
15.5
15.0
3.5
10.4
1.0
6.1
55.6
41.8
50.9
45.2
9.4
2.4
_,
107.5
111.7
107.5
113.6
22.3
29.8
213.5
183.1
207.9
213.6
30.2
30.5
58.6
67.0
66.0
64.8
16.7
19.1
295.2
305.6
292.9
291.8
18.6
18.6
C.06
0.73
0.02
MRI
108
112
106
114
22.2
29.7
212
183
206
213
29.7
29.7
62
65
65
64
16.9
19.9
294
305
292
290
19.3
19.1
\l All Collaborator I results deleted from subsequent analyses.
21 These points deleted as outliers In subsequent enalyses.
-------
COLLABORATORS' RESULTS
The one primary set of results the collaborators furnished MRI is
the results of their chemical analysis of their samples. These data,
which include calibration and absorbance data relating to these results,
as well as raw field data that resulted from sampling in the field, are
given in volume 2 of this report.
For convenience, these final results of the collaborators analyses
are summarized in Table 2. There are two sections to Table 2: the top
portion that gives measurements from the test samples taken by the col-
laborators; and the bottom portion that gives the results of the col-
laborators' analyses of the standard samples given them. Column 1 of
the top portion of Table 2 gives the N02 level that is mixed with the
ambient air to form the run's challenge. It does not include the quantity
of NO2 that was present in the ambient air (see page 46 for information
on the "True Value")* Column 2 and succeeding even columns provide the
NO- values the collaborators measured during each run. There are six
values per run; two are- from the two samples taken from the ambient air
(unspiked) manifold; and 'four are from the four samples taken from the
spiked manifold (each collaborator pulled all six of his 24-hr samples
simultaneously). Column 5 and succeeding odd columns provide the values
MRI obtained when it checked the collaborators' calculations.
MRI's check of the collaborators' results was a gross overall
check to determine if there were major errors due to, for example, mis-
placement of the decimal point. Minor difference could be attributed
to the reading of the collaborators' calibration curves.
43
-------
STATISTICAL ANALYSIS OF COLLABORATORS' RESULTS
Our analysis of variance model for the whole experiment is;
Xijk = n + CJL + Lj + 0,^
where p, = overall mean,
Ci = ith collaborator, 1=1, . . ., 9,-'
LJ = jth N02 level, j = 1, . . .,4,
CLjj = collaborator-level interaction,
ek(ii) = residual error in k measurement in the ijth cell,
k = 1 4 V ij, and
Xi>k = ijkth bias, i.e., ijkth determination - true value.
The NOo level is a fixed factor, but the collaborators are con-
sidered to be a random factor; i.e., the nine collaborators used in
the experiment are considered a sample drawn from a population of pos-
sible collaborators. Thus, the expected mean squares (EMS) are:
Term EMS
C ae2 + 16 ac2
L ae2 + 4 aCL2 + 36 aL2
CL ae2 + 4 a^2
e ae2
y Ten collaborators participated in the test, but one produced use-
less results (see discussion, page 41). See Appendix H for a sta-
tistical analysis including all 10 collaborators.
45
-------
The true value is the spiked level of N0£ plus the ambient N02
the ambient level, of course, being unknown. Therefore, a true value
is taken as the spiked amount plus the average ambient determination
of all collaborators. For example, the true level one N0£ is taken
as 116.9 ug/ra because 94.1 ug/m^ were spiked and the average ambient
reading from the nine collaborators was 22.8
An individual response is a bias (collaborators' reading of the
spiked line minus the true value). For example, collaborator A's first
reading of the spiked line at level one was 102.4 pg/m3, so ^m =
102.4 - 116.9 =-14.5 pg/m3, etc. Since subtracting a constant is
merely coding the data, the components of variance are unaffected by
the use of biases rather than spiked readings as the response.
Six observations (out of 144) were discarded as outliers (via
the Dixon test, a = 0.05). The outliers were not associated with par-
ticular collaborators or particular N02 levels. All the outliers were,
however, too low. The presence of six artificial observations means
that the error degrees of freedom are 102 (rather than 108).
The assumption of homeoscedasticity was checked by computing the
cell ranges (R^) and E for each N02 level (see Table 3). The statistic
R" is proportional to a, so that if TL is related in some monotonic
fashion to the N(>2 level (L), a data transformation would be indicated.
Although TO. is not uniform, it is not a monotonic function of L, either,
and the ratio R/L is less stable than R itself* In general, if "E (i.e.,
a) is not independent of L, a data transformation is indicated. In prac-
tice, however, one must find some reasonably simple relationship between
R and L that is "more constant" than the original R itself. TOT: instance,
R/L or (R/L)1/2 or (R/L)2 or log ^R/L) or (R/log L) or (log R/lpg L),
etc., should be more stable than R (without reference to L) in order for
a data transformation to be indicated. Tor the N0£ data in this test,
such functions are no more stable than R itself, which is not surprising
since 1 is essentially constant for the top three N02 levels. Therefore,
no data transformation was applied. However, ae at the lowest NO 2 level
is probably less than <7e elsewhere, and the ce estimates per level were
computed.
46
-------
Table 3. R VERSUS N02 LEVEL
R L (u.g/m3) K/L (%)
9.3 65.6 (L3) 14.3
13.2 116.9 (LI) 11.3
15.8 222.6 (L2) 7.1
15.5 314.7 (L4) 4.9
A complete analysis of variance was performed (see Table 4). The
discussion of biases will be presented, and the components of variance
analysis and discussion will follow.
Table 4. ANALYSIS OF VARIANCE (RESPONSE = BIAS)
Source df j>S MS F
C 8 10,439.11 1,304.89 23.34
L 3 3,100.45 1,033.48 2.83
CL 24 8,760.67 365.03 6.53
e 102 5,702.18 55.90
BIASES
Collaborators differ significantly in their average bias, and the
calibration curves (bias versus level) for the various collaborators
are significantly nonparallel. However, bias depends slightly, if at
all, on the N02 level (F (3,24) = 2.83, 0.10 < a < 0.05).
3
In general, collaborators read -9.5 ug/m too low, i.e., about
5% below the true value (see Tables 4 and 5). The collaborators can be
divided into four groups (via the Fischer method): two collaborators
reading 21-23 iig/nr too low, three collaborators reading 11-13 pg/m3
too low, three collaborators reading 2-6 \ig/w? too low, and one col-
laborator reading 6 p,g/m3 too high. One collaborator exhibits no bias,
i.e., produces an average reading not statistically different from zero.
47
-------
Table 5. AVERAGE BIAS (pg/m3) PER N02 LEVEL
(p,g/m3) Average Bias Percent true
65.6 -6.4 -10
116.9 -3.9 -3
222.6 -12.0 -5
314.7 -15.7 -5
Overall average
180.0 -9.5 -5
Table 6. COLLABORATOR BIASES (pg/m3) (ALL N02 LEVELS)
Collaborator Average bias Percent true Group^'
A -5.6 -3 C
B -20.9 -12 A
C -12.9 -7 B
D 6.4 4 D
E -2.2 -1 £f
F -22.5 -13 A
G -5.7 -3 C
H -11.0 -6 B
J -11.2 -6 B
a/ Members of a group do not differ significantly in (average) bias.
b/ - Not statistically different from zero.
Although, the CL interactions term is significant, almost all of
the reason for the significance is due to two collaborators (see Table
7 and Figure 14). Collaborators H and G showed a much larger negative
bias at the 66 p,g/m3 N02 level than they had at the other N02 levels.
All other collaborators exhibited a relatively uniform percentage error
for all N02 levels. Thus, in general, a calibration curve is relatively
linear over the range of NO, examined.
48
-------
65.6 ug/m
-2
-16
-3
0
-7
-11
-15
-31
-2
3 116.9 ug/m3
-9
-1
-2
+5
+2
-10
+1
-10
-6
222.6 ug/m3
-2
-13
-11
+3
-3
-14
-1
0
-8
314.7 ug/m3
-2
-14
-7
44
0
-12
-4
-4
-6
Table 7. COLLABORATOR PERCENT BIAS PER N02 LEVEL
Percent Bias
Collaborator
A
B
C
D
E
F
G
H
J
PRECISION
n
The components of variance ae (dispersion of repeated measurements
within a collaborator) and ac^ (dispersion between collaborator averages)
are shown in Table 8.
Table 8. COMPONENTS OF VARIANCE (ALL NO2 LEVELS)
Absolute standard Relative standard error
error (ue/m3) (% true)
ae 7.48 4.2
ar 8.84 4.9
ae2 + oc2 11.58 6.4
The two errors ae and ac are approximately the same size. On the
average, a collaborator will read between + 2 ae of his average "most"
95%) of the time; a set of collaborators will read between + 2
O 2 + a^2 Of their average most of the time, etc.
49
-------
10,-
-CIO
60 80 100 120 140 160 180 200 220 240 260 280 300 320
N02(/ig/m3)
Figure 14. Collaborator percent bias versus level of N(>2
-------
SUMMARY AND DISCUSSION OF STATISTICAL ANALYSIS
1. There is a general negative bias Co the NO2 measurements, but it
is relatively small (approximately 5%) and nearly independent of the N0£
level. Various collaborators differ in the amount of bias shown but this
variability is also relatively small (o~c ~ 5% true value).
2. The measurement errors are essentially uniform for all collab-
orators. Also, CJe is constant for the three higher NO2 levels, but some-
what smaller at the lowest NO2 level. In general, oe is about 4% of the
true value, but is about 7% of the true value at the lowest N0£ level.
3. Although two collaborators exhibited nonlinear calibration
curves due to much larger biases at the lowest N02 level than elsewhere,
in general the calibration curves are relatively linear.
4. Two points regarding the usefulness of the estimates should be
mentioned:
a. A replicate in this experiment is not a genuine replicate,
since all duplicate readings were taken simultaneously. Physically, the
fact that the collaborators did not have to "gear up" to produce a repli-
cate probably depresses the ae estimates. Mathematically, the measure-
ment errors are perhaps dependent, i.e., correlated due to the way the
test was performed. The effect of such dependence on the CTe estimates can
be in either direction; a positive correlation inflates cre and vice versa.
b. The higher the N02 level, the more reliable the relative
bias estimate, since the unknown ambient becomes an increasingly smaller
fraction of the NO2 amount.
51
-------
LOWER DETECTABLE LIMIT (LDL)
Two meanings of LDL are used in the following discussion: (a) the
smallest value of NO2 that can be reliably identified as existing (i.e.,
positive) when the method is used by _a collaborator (a "pure" LDL) and
(b) the smallest reliable value of NO2 determined by a set of collabora-
tors using the method (a "practical" LDL).
Three methods of estimating the LDL were used. The first method is
based upon the set of blank readings, and falls, more or less, into the
category of a rule of thumb. The average blank value was doubled to esti-
mate the LDL; this results in a value of 2 pg/m3.
The second method of 'estimating the LDL used the ambient readings of
the collaborators obtained during the actual experiment. Using the within-
collaborator standard deviation (oe) of 1.86 pg/m3, an estimated "pure"
LDL of about 3.7 p,g/m3 is calculated. Incorporating the collaborator-
collaborator standard deviation (ac = 4.69 pg/m3, V CTe^ + ac2 = 5.05
lig/m3), an estimated "practical" LDL of about 10 pg/m3 is obtained. Note
that there is no way of incorporating the bias into these estimates, since
the true ambient levels are not known.
The third method of estimating the LDL used the collaborators' cali-
bration curves* themselves by computing the standard errors of the esti-
mate (s.e.'s) of the calibration lines. To estimate the pure LDL, the
average s.e. was doubled; to estimate the practical LDL, the s.e. from all
collaborators combined was doubled. This yields a pure LDL estimate of
4 p,g/m , and a practical LDL estimate of 10 pg/m3. Note that in this data
set a bias estimate is possible, and in fact the intercept of the com-
bined calibration curve is + 4.6 pg/m3 (rather than zero). Thus, the pure
LDL is about 9 pg/m , and the practical LDL is 15 pg/m3, when the bias is
accounted for.
The three sets of results seem to agree well. It seems reasonable to
state that the pure LDL is probably £ 9 pg/m , and the practical LDL
=s 15 pg/m3.
The only available calibration curves were from collaborators C, E, F,
and G.
-------
CONCLUSIONS
The major conclusions that can be drawn from the results of this
collaborative test are:
1. The N02, ambient-air sampling system developed by MRI is an
effective system for use in collaborative testing of manual methods
such as the TGS-ANSA procedure.
2. The "Tentative Method for the Determination of Nitrogen Dioxide
in the Atmosphere (TGS-ANSA Procedure)" is adequately written for those
knowledgeable of sampling and analysis techniques as presented therein.
3. If the tentative TGS-ANSA procedure as given in Appendix A of
this report is followed by people knowledgeable of the sampling and analy-
sis techniques given therein, then such persons will obtain results with
an average bias of -9.5 yg/m^ over the range 50-300 ug/m3. The precisions
can be estimated from the within laboratory standard deviation (ae) of
7.5 p,g/m3, and the collaborator standard deviation (ac) of 8.8
55
-------
RECOMMENDATIONS
Based upon Che conclusions that have been drawn from the results
of this collaborative test, it is recommended that:
1. The same N02 sampling system be used in the evaluation of the
remaining N02 methods to be tested;
2. The data sets to be obtained from the subsequent methods to
be evaluated be based on experimental designs, test procedures and sam-
pling system operational procedures as similar as possible to those of
the TGS-ANSA collaborative test so that comparisons of the methods are
based on similar criteria; and
3. No further analysis be made of the results from the TGS-ANSA
method until the results from the other three methods are obtained.
57
-------
APPENDIX A
TENTATIVE METHOD FOR THE DETERMINATION OF NITROGEN
DIOXIDE IN THE ATMOSPHERE (TGS-ANSA PROCEDURE)
59
-------
ENVIRONMENTAL PROTECTION AGENCY
METHODS STANDARDIZATION BRANCH
QUALITY ASSURANCE AND ENVIRONMENTAL MONITORING LABORATORY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
APRIL -1974
TENTATIVE KETHOD FOR THE DETERMINATION OF NITROGEN DIOXIDE
IN THE ATMOSPHERE (TGS-ANSA)9.
A tentative rrethod is one which has been carefully drafted from
available experimental information, reviewed editorially within
the Methods Standardization Bran.ch and has undergone extensive
laboratory evaluation. The method is still under investigation
and therefore is subject tc'revision.
60
-------
1. Principle and Applicability
1.1 Nitrogen dioxide is collected by bubbling air through a
solution of triethanolamine, o-irethoxyphenol and sodium metabisulfite.
The nitrite ion produced during sampling is determined colornetrically
by reacting the exposed absorbing reagent with sulfam'lamide and 8-
anilino-1-naphthalenesulfonic acid, airaonium salt.
1.2 The method is applicabl£ to collections of 24-hour samples in
the field and subsequent analysis in the laboratory.
2. Range and Sensitivity
2.1 The range of the analysis is 0.025 to 4.0 yg NOZ/ml. Beer's
law is obeyed throughout this range. With 50 ml of absorbing reagent and a
sampling rate of 200 on /min for 24-hours, the range of the method is 20 to
7
/uu ug/m" nitrogen dioxide.
2.2 A concentration of 0.025 ug NOl/ml will produce an absorbance
of approximately 0.025 using 1 cm cells.
3. Interferences
3
3.1 At a nitrogen dioxide concentration of 100 yg/m the following
3
pollutants, at the levels indicated, do not interfere: ammonia, 205- yg/m;
carbon monoxide, 154,000 ug/m ; formaldehyde, 750 ug/m ; nitric oxide, 734
3 3 ' 3 3
ug/m ; phenol, 150 ug/m ; ozone, 400 yg/m and sulfur dioxide, 439 ug/m .
3.2 • A temperature of 40°C during collection of sample had no effect
on recovery.
. 61
-------
-2-
4. Precision and Accuracy
4.1 Precision and Accuracy
4.1.1 On making measurements from standard nitrogen dioxide atmos-
pheres, prepared by using perrceation devices, a relative standard deviation
of T\ and a collection efficiency of 932 were determined throughout the
range of the rcethod.
4.2 Stability
4.2.1 The absorbing reagent is stable for 3 v/eeks before sampling and
the collected samples are stable for 3 v/eeks after sampling.
5. Apparatus
5.1 Sampling. A diagram of a suggested sampling apparatus is shown
in Figure l.; '
5.1.1 Probe. Teflon, polypropylene, or glass tube with a polypropylene
or glass funnel -at the end.
5.1.2 Absorption tube. Polypropylene tubes 164 x 32 mm. equipped with
polypropylene two-port closures. Rubber stoppers cause high and varying blank
values and should not be used. A glass tu6e restricted orifice is used to
disperse the gas. The tube, approximately 8 mm O.D.-6 mm. I.O., should be
152 mm long with the end drawn out to 0.3-0.6 mm. 1.0. The tube should be
• *.
positioned so as to allow a clearance of 6 rrm from the bottom of the absorbsr.
• -5.1.3 Tloisture trap.Polypropylene tube equipped with a two port closuro.
The entrance port of the closure is fitted with tubing that extends to the
bottom of the trap. The unit is loosely packed with glass wool to prevent
moisture entrainn;cnt.
62
-------
-3-
5.1.4 Kerr.brane Filter, 0.8-2.0 microns porosity.
5.1.5 Flow Control Device. Any device capable of maintaining a
constant flow through the sampling solution between 180-220 cm /min.
2
A typical flow control device is a 27 gauge hypodermic needle, three-
eights inch long, ("ost 27 gauge needles will give flow rates in this
range.) The device used should be protected from particulate matter.
A membrane filter is suggested. 'Change filter after collecting 10 samples.
5.1.6 Air Purr.p. Capable of maintaining a pressure differential of at
least 0.6-0.7 of an atmosphere across the flow control device. This value
includes the minimum useful differential, 0.53* ' atmospheres, plus a safety-
factor to allow for variations in atmospheric pressure.
5.1.7- Calibration Equipment. Flowmeter for1 measuring airflows up to
275 cmj/nnn. within +_2%, stopwatch, and a precision wet test meter (1 liter/
revolution). •
5.2 Analysis
5.2.1 Volumetric Flasks. Two each 250, 1000 ml; three each 200; 7 each
100 ml; one 500 ml.
5.2.2 Pipets, volumetric. One each, 2, 3, 9, 10, 20 and 50" ml; seven 5
5.2.3 Pipsts serological, graduated in 1/10 ml divisions. One each 1,
5 ml. •
5.2.4 Test Tubes. Each approximately 20 x 150 mm.
5.2.5 Spectrophotorceter. Capable of measuring absorbance at 550 r,m.
5.2.6 Graduated cylinder. One each 50 ml.
-------
-4-
6. Reagents
6.1 Sampling
6.1.1 Triethanolamine [N^HjOtO^]. Reagent grade.
6.1.2 o-'iethoxyphenol (o-CHjCCgH^CH). Also known by its trivial
nar.e, guaiacol. Reagent grada. belting point 27-28°C. (Caution: Technical
grade naterial will not irs3t this specification and should not be used).
6.1.3 Sodium "etabisulfite (Na-SgOg).' ACS reagent grade. •
6.1.4 Absorbing Reagent - Dissolve 20g of triethanolamine, 0.5g of
o-rcethoxyphenol, and 0.250g of sodium rcetabisulfita consecutively in 500 rr.l
of distilled water. Dilute to one 1-iter. with distilled water. Mix thoroucli-
ly. The solution should be colorless. This solution is stable for thrsa -./'j-iks
if protected from light.
6.2 Analysis
6.2.1 -Hydrogen Peroxide (HgO-). ACS reagent grade, 30%.
6.2.2 Sulfanilamide [4-(H2N)CgH4S02NH2]. Melting point 165-167°C.
6.2.3 8-Anilino-l-naphthalenesulfonic acid Ammonium salt (ANSA) (8-C5H5;:;i
l-C10H6SOj!Hj). Minimum analysis, 983.
6.2.4 Sodium Nitrite, [NaN02]. ACS reagent grade. Assay of 97" Naf;02
or greater.
6.2.5 Methanol, absolute [CH.^OH]. ACS reagent grade.
• .
6.2.6 Hydrochloric acid, [HC1].' Concentrated. ACS reagent grade.
6.2.7 Hydrogen Peroxide Solution. Dilute 0.2 ml of 30fj hydrogen poroxi;!o
to 250 ml with distilled water. This solution can be used for a month if pro-
tected from light and refrigerated.
64
-------
-5-
6.2.8 Sulf anil snide Solution (2.0:; in 4N.HC1). Dissolve 2.0g of
sulfam'lanids in 33 nl of concentrated HC1 and-dilute to 100 ml with
distilled water. Mix. This solution can be used for two weeks, if
refrigerated.
6.2.9 A:;SA Solution. (0.1S W/v). Dissolve O.lg ANSA in 50 ml absolve
irethanol. Dilute to 100 ml with absolute irethanol in a voluxetric flask.
Mix. Keep stoppered, when not in use, to minimize evaporative losses. Pre-
pare fresh daily. (CAUTIC'.'; Older reagent may result in lower absorbance).
6.2.10 Standard Nitrite Solution. Dissolve sufficient desiccated
sodium nitrite and dilute with distilled water to 1000 ml so a solution
containing 1COO ug.'iol/ml is obtained. The amount of NaNO« to use 1s
calculated as follows:
6 = L500 x 1QO
where
G = Amount of NaNOg, grams.
1.500 = Gravimetric factor in converting NOg Into NaNOg
A = Assay, percent
7. Procedures
7.1 Sampling. Assemble the sampling apparatus, as shov/n in Figure 1.
Components upstream from th-3 absorption tube may be connected, v/here required,
with teflon or polypropylene tubing; glass tubing with dry ball joints; or
glass tubing with butt-to-butt joints with tygon, teflon or polypropylene.
Add exactly 50 ml of absorbing reagent to the calibrated absorption tube
65
-------
-6-
(8.1.3). Disconnect funnel, insert calibrated flowrceter (8.1.1) into
the end of the probe and treasure flow before sampling. Denote as F^.
If flow rate before sen.pling is not between 180-220 cm /min replace the
flc1./ controlling device and/or check the system for leaks. Start sanoling
only after obtaining an initial flow rate in this range. Sample for 24-
hours and rraasure the flow after sampling by again inserting a calibrated
flowrvster into the probe, after removing the funnel. Denote as Fp.
7.2 Analysis. Replace any water lost by evaporation during sampling
by adding distilled water up to the calibrated mark on the absorption tu^5.
Mix well. Pipet 5 ml of the collected sample into a test tube, add 0.5 rnl
of tho peroxide solution and mix vigorously for approximately 15 seconds.
Add 2.7 ml of sulfanilamide solution and mix vigorously for about 30 seco"r.:.
Then pi pet 3 ml of the ANSA solution, mix vigorously for about 30 seconds.
The ANSA must be-added within 6'minutes of mixing the sulfanilamide solution.
(CAUTIC'J: Longer time intervals will result in lowered absorbance values).
•
Prepare a blank in the same manner using 5 ml of unexposed absorbing solution.
The absorbance of the blank should be approximately the same as the y-inter-
cept in the calibration curve (Section 8.2). Determine absorbance at 550 nr.i
with distilled water in the reference cell using 1 cm cells. The color ccr.
be read anytime from 1 to 40 minutes after addition of the ANSA. Read
yg "Ol/:il fro i the calibration curve (Section 8.2).
7.3 Snectrophoto.72ter cells must be rinsed thoroughly with distilled
water, acetone, and dried, otherwise a film will build up on the cell walls.
8. Calibration and Efficiencies
8.1 Sarpli'iO
66
-------
-7-
8.1.1 Calibration of Flo-.-.reter. (See Figure 2). Using a wet test
3
rater and a stopwatch, determine the rates of air flow (cm /min) thrcu^
the flc-..::ster at a ninir'un of four different ball positions. Plot ball
position versus flew rate.
8.1.2 Flow Control Device. The flow control device results in a
constant rate of air flcv; through the absorbing solution and is dcter™ir.:d
in 7.1.
8.1.3 Calibration of Absorption Tube. Calibrate the polypropylena
absorption tube, (Section 5.1.2) by first pipating in 50 ml of water or
'absorbing reagent. Scribe the level of the meniscus with a sharp objact,
go over the area with a-felt-tip marking pen, and rub off the excess.
8.2 'Calibratieh Curve. Dilute 5.0 ml of the 1000 yg NOl/ml sol'j--r/
to 250 ml with absorbing reagent. This solution contains 20 pg fl
Dilute 5.0 "ml of the 20 ug NOZ/ml standard to 200 ml with absorbing
This solution contains 0.50 pg FiO^/ml. Prepare calibration standards by
pipeting the indicated volume of the standard into volumetric flasks and
diluting to the nark with absorbing reagent.
Final Concentration
Volura of Standard 'Volurr.a ml ug NOZ/nl
10 ml of 0.50 -,g .'lO^/r.l 100 0.05
20 n:l of- 0.50 ^g ;,'o;/,al 100 0.10
2 .il of 20 yg :iO;/!-:l 200 0.20
Use 'J.50 ug/ral Sta;:riard Directly — Q.50
5 nil of 20 ]:g NO^/^l Standard ' 100 ' 1.00
0 r.l of 20 :.•] !:o;/n:l Standard' 100 l.FO
67
-------
-8-
Run standards, plus a blank, as instructed in 7.2. Plot absorbance vs
ug N'Ol/rcl. A straight line should be obtained with a slope of approximately
0.5 absorbance units/'jg fiOl/ml, and a y-intercept (i.e., zero pg NOZ/ml) of
approximately 0.01 absorbance units. The absorbance is linear up to a con-
centration of 4.0 ug NOl/rcl, absorbance of 1.9. Therefore, if samples
exceed the abscrbance of the highest calibration standard and the above
absorbance is within the ra'nre of the spectrometer, the calibration curve
can be extended by including higher concentration standards. If a higher
absorbance range is not available, samples must be diluted with absorbing
reagent until the absorbance is within'the range of th« highest standard.
c
8.3 Efficiencies.. An overall average-efficiency of 93S was obtained
f»T™ test ?f"??pHprps havina a nitrogen dioxide concentration of 20 to 7CG
yg/m .
9. Calculation
9.1 Sampling
9.1.1 Calculate volume of air sampled.
V = Fl * F2 x T x 10"6
2
V = Volume of air sampled m . •
Fj = Measured flow rate before sampling, cm/min.
F2 = Measured flow rate after siampling, cm /nin.
T = Time of sampling, min.
-6 ^ ^
10" = Conversion of cm to m .
9.1.2 Uncornjcted Volume. The volume of air sampled is not corrected
to S.T.P., because of the unccrtainity associated with 24-hour av
68
-------
-9-
terperjjture and pressure values.
9.2 Calculate the concentration of nitrogen dioxide as yg :;02/.~.
vg i:o2/n3 = ('Q :;o;/m) x so
V X (0.93)
50 = Volure of absorbing reagent used in sampling, ml.
3
V = Volu./s of air se~oled,.m .
0.93 = Overall efficiency of irethod.
9.2.1 If desired, concentration of nitrogen dioxide may be calcin;
as pc*7! "iOo.
ppm = (ug::02/r,3) X 5.32 X 10~4
10. References
1. Nuiu, J. u., f-'uerst, K. G., Meeker, J. K., Guyer, n., Sav/icki, E.
"A Twenty-Four Hour Method for the Collection and Manual Colorir.etric
Analysis of Nitrogen Dioxide. Presented at the 165th ACS National
Meeting in Dallas, Texas, April 8-13, 1973.
2. Lodge, J. P., Jr., Page, J. B.,-Arcmons, B. E., Swanson, G. A. "The
Use of Hypodermic Needles as Critical Orifices in Air Sampling."
J.A.P.C.A. 16_, 197-200 (1966).
69
-------
Inverted
funnel
BUBBL^I
Figure 1. Sampling train.
-------
WET
TEST
METER
RATE
CONTROL
VALVE
PUMP
OPEN TO
ATMOSPHERE
FLOWMETER
FIGURE 2
-------
APPENDIX B
DATA ON THE PERMEATION TUBES USED AS
THE SOURCE OF THE SPIKED LEVELS OF N02
73
-------
Branch
4
4
4
3
3
3
1
1
1
1
1
2
2
2
2
2
2
Number
29-2
34-12
35-13
29-4
35-8
35-16
29-3
28-10
34-3
34-13
34-6
34-1
34-10
Rate of NO? (ue/min)
1.210
1.770
1.990
1.210
1.434
1.597
1.345
1.160
1.195
1.275
1.548
1.226
1.138
As shown in Figure 6 of the text (p. 15), there were four branches
to the NC>2 permeation tube assembly. Each branch contained a set of
permeation tubes as follows:
Permeation Tube+ 1 S Branch
( g/min) ( g/min)
0.001
0.002
2.980
0.003
0.001
3.200
0.001
0.002
0.002
0.002
5.536
0.002
0.002
0.001
0.003
0.001
6.382
a_/ The sum of the NC>2 generated by each permeation tube in the branch.
Permeation rates for the above tubes were determined by the National
Bureau of Standards and validated by the Methods Standardization Branch
(MSB) of EPA at 25.1°C before they were given to MRI for use on the
collaborative test.
The combinations of branches used for the four runs of the TGS-
ANSA collaborative test are:
Run No. Date Branches Used
1 April 29-30 1
2 April 30 - May 1 1, 3 and 4
3 May 1-2 3
4 May 2-3 1, 2, 3 and 4
74
-------
APPENDIX C
CALIBRATION OF THE VENTURI AND DRY-GAS METER
75
-------
o
The venturi and dry-gas meter were calibrated using a 1.0-ft /rev,
wet-test meter, as shown in Figure C-l. The wet-test meter is connected
between the splitter and the dry-gas meter. A bubbler is used before
the wet-test meter to saturate the air with water. The air flow then
proceeds through the venturi to the N02 bleed-in as it does in normal
operation.
Since the saturated air coming from the wet -test meter is not dried
before going into the dry-gas meter, no correction for water vapor pres-
sure is necessary and only the normal corrections for temperature and
pressure are used. The flowrate of the wet-test meter (to stp) is:
F 294
Flowstp = Flow(meter reading) x — x —
where T = temperature of wet -test meter + 273, and
P = Patm + pressure of test meter manometer.
The venturi flowrate is dependent on both temperature and pressure.
Therefore Flowgt is corrected to venturi conditions.
760 T2
Flowventuri = Flowstp x ^ x ^
where T~ = temperature of gas stream + 273, and
p2 = patm + p(gas stram).
The dry-gas meter is temperature compensated, so only pressure
corrections are made for its readings and a temperature base of 21° C is
used for calibration. Thus the true flowrate of the dry-gas meter (Fm)
is
„ „, 760
F = Flowstp x —
where P = Patm + P(gas stream)'
The correction factor f to convert Fm, measured dry-gas meter
flowrate, to true flowrate is then
76
-------
Splitter
L
Temp
Wet Test Readin9
Meter
Spiked
Line Bubbler
t
1 NO2 Bleed-in
*t>
Dry Gas Venfuri
Meter
Pressure
Reading
Figure C-l. Venturi and dry-gas meter calibration system
77
-------
The venturi and dry-gas meter were calibrated at three flowrates;
50, 55, and 60 liters/min. Normal system flowrates are 55-60 liters/rain.
The calibration factor for the dry-gas meter is constant at the calibra-
tion flows (+ 0.27.). The average value of flow from seven determinations
is used in calculating true flowrates of the system. The plot of venturi
P versus flowrate follows a straight line over the range used in cali-
bration. From the slope and intercept of the line flowrates were calcu-
lated.
78
-------
APPENDIX D
WRITTEN COMMUNICATIONS WITH POTENTIAL COLLABORATORS
79
-------
MIDWEST RESEARCH INSTITUTE
425 Volker Boulevard
Kansas City, Missouri 64110
Telephone (816) 561-0202
Dear Sir:
Your name has been given to Midwest Research Institute (MRI) by the
Environmental Protection Agency (EPA), as having expressed an interest
in becoming a voluntary collaborator in an N0« Testing Program, to be
sponsored by EPA. I am writing to confirm this expression of interest
by your organization.
The objective of this program is to determine the reliability and bias
of four methods for measuring NC^-ambient air. MRI has the responsibili-
ties for organizing the program, furnishing the test facilities, co-
ordinating the testing, analyzing the results of the collaborators, and
reporting the findings to EPA.
The sodium-arsenite method—the first method to be evaluated--will be
tested in Kansas City, Missouri, during the first part of January 1974.
Ten collaborators are needed for each of the four methods. A tentative
test schedule of the other three methods is given on the attached "Col-
laborator Form."
A writeup of the sodium-arsenite method is enclosed, and provides the
information needed for a collaborator to perform the testing and analyze
the samples he takes.
Each collaborator organization, once chosen, will be reimbursed for travel,
subsistence, lodging, and miscellaneous expenses (e.g., shipment of equipment
and local travel), for the employee sent to Kansas City to perform the
testing. Each collaborator will need to furnish the sampling apparatus
called for in the writeup. For each of the two manual methods, sodium-
arsenite and TGS-ANSA, six sampling trains will be needed. Cases will be
furnished in which to ship field samples to the collaborator's laboratory
for analysis.
We would appreciate your completing and returning to us the Collaborator
Form before November 15.
Sincerely,
Paul C. Constant, Jr., Head
Environmental Measurements Section
Enclosures: (1) Collaborator Form
(2) Tentative Method for the Determination
of Nitrogen Dioxide in the Atmosphere
(Sodium-Arsenite Method)
80
-------
COLLABORATOR FORM
1. Methods to Test (Check the ones in which you want to participate
as a collaborator) :
Sodium Arsenite TGS-ANSA Procedure
{ | Continuous Saltzman li Chemiluminescent
2. Equipment Available for Test:
Could you furnish six trains for:
Sodium Arsenite : | | yes | | no
TGS-ANSA: Q yes Q no
Have you a Colorimetric (Continuous Saltzman) Ambient N0£ Monitor?
LJ yes [I no Make _ Model _
Have you a Chemiluminescent Ambient N02 Monitor that you would use?
Q yes Q no Make _ Model _
3. Test Period (Each Method) ;
Period acceptable (calendar days) :
6 days 10 days 13 days
4. Methods You Have Used;
LJ Sodium Arsenite, [_] TGS-ANSA, [__] Continuous Saltzman,
[_) Chemiluminescent , j_J Others : _
5. Remarks:
6. Company:
Address:
Person to Contact:
Telephone Number;
81
-------
MIDWEST RESEARCH INSTITUTE
425 Volker Boulevard
Kansas City. Missouri 64110
Telephone (816) 561-0202
21 February 1974
AIR MAIL
Name
Company
Address
City, State ZIP
Dear Mr. Name:
Last November you responded to a survey of ours indicating your interest
in participating in an EFA-sponsored NC<2 Collaborative test using the TGS-
ANSA Method. If you still desire to be considered as a candidate to par-
ticipate in this test, would you:
1. Let me know before 8 March 1974;
2. Review the enclosed write up, "Tentative Method for the
Determination of Nitrogen Dioxide in the Atmosphere (TGS-ANSA)," dated
February 1974;
3. Use this TGS-ANSA technique exactly as presented to obtain
five samples of N02 in the ambient atmosphere in your locality; and
4. Send me a copy of the results of your analysis of these five
samples by 29 March 1974? (These results will be used to assist us in the
selection of ten collaborators.)
The NC-2 collaborative test using the TGS-ANSA method is tentatively
scheduled to take place in Kansas City, Missouri, starting the morning of
29 April and ending the afternoon of 3 May. The test will be indoors
using a sampling system developed by MRI for this collaborative testing.
(This system was used quite successfully for the sodium-arsenite method on
a similar 5-day test in January.)
82
-------
Mr. Name
21 February 1974
Page 2
The sampling system that will be used in this test is shown in diagram
form in Figure 1, an enclosure of this letter. Each collaborator will
have four of his trains connected to the sampling manifold of the spiked
line (bottom line of Figure 1) and two of his trains connected to the
sampling manifold in the unspiked line. Thus, each collaborator is re-
quired to furnish a minimum of six of the trains that are specified in
the enclosed write up.
The test schedule will require each collaborator to sample simultaneously
at six ports of the sampling system during each of the four 24-hour runs
that will constitute the 5-day test, to provide him 24 individual samples.
In addition to these, he will be given several standard samples to analyze
at his home laboratory along with the 24 collected. Wooden cases in which
samples can be shipped to your home laboratory will be furnished on a loan
basis.
As stated in our initial letter to your organization, each collaborator
will be reimbursed for travel, subsistence, and lodging for the employee
it sends to Kansas City to perform the field sampling, as well as local
travel and miscellaneous expenses such as cost of shipping field equipment
to be used on site for the sampling.
We will be in touch with you as soon as the ten collaborators are selected
for this TGS-ANSA method. If you have any questions, please contact me.
Very truly yours,
Paul C. Constant, Jr., Head
Environmental Measurements Section
PCC:cdn
Enclosures:
1. "Tentative Method for the Determination of Nitrogen
Dioxide in the Atmosphere (TGS-ANSA)"
2. Figure 1 — N02 Ambient Air Sampling System
83
-------
APPENDIX E
INSTRUCTIONS FOR COLLABORATORS N02 COLLABORATIVE
TEST; METHOD TGS-ANSA PROCEDURE
85
-------
INSTRUCTIONS FOR COLLABORATORS
N02 COLLABORATIVE TEST: TGS-ANSA METHOD
GENERAL INFORMATION
1. Calibration, sampling, analysis, etc. should be done ex-
plicitly as stated in the April 1974 write-up furnished you on "Tenta-
tive Method for the Determination of Nitrogen Dioxide in the Atmosphere
(TGS-ANSA)."
2. Each collaborator will have an area and a set of sampling
ports at both the spiked-samp ling-manifold table and at the unspiked
table—see accompanying figure and table below:
Collaborator Spiked Table Unspiked Table
I.D. Name M§a Ports Area Ports
1 Joe Andrews l 1-4 l 41-44
2 W. (Bill) T. Shebs 2 5'8 2 37"40
3 Donald Kutys 3 10"13 3 33~36
4 Dan Zorbini 4 14'17 4 29-32
5 Ms. Barbara Kladatos 5 18'21 5 25-28
6 Lynn Hutchinson 6 25-28 6 17-20
7 Ken Wilcox 7 29-32 7 13-16
8 James Ward 8 33-36 8 9-12
9 Kenneth Smith 9 37-40 9 5-8
10 Bob Deardorff 10 41-44 10 1-4
3. For each run each collaborator will have six sampling trains
running simultaneously: four in his area on the spiked table and two in his
area on the unspiked table. Each train is to be attached to a separate port
as specified in the above table.
4. Nitrogen Dioxide Data Sheets will be provided—a copy of one
is attached. All information on sampling should be filled in during the
period of the run. The bubbler identification number will be made by MRI's
laboratory supervisor. He will have labels to affix to your samples. The
coding is run number (1-4), followed by port number (1-45), followed by
sampling table designation (S or U for spiked or unspiked), and terminated
in collaborator number (1-10); e.g., 1-6-S-2 for run 1, port 6, spiked table
86
-------
and collaborator 2 (W. T. Shebs of Shell Development Company). Analysis
information is to be filled in at your home laboratory.
5. A copy of the data sheet for each run should be given to MRl's
laboratory supervisor after each run is done.
6. Each collaborator should work independently of each other
collaborator.
7. Shipping containers for the samples (absorbing tubes) will be
available on a loan basis.
8. On the spiked sampling table, a separate power circuit (110-V,
60-cycle, four-outlet strip) is to be used by each collaborator for his
four trains. On the unspiked line one strip will be shared by two collab-
orators. These strips are under the table tops near the periphery of the
tops.
9. MRI will provide each collaborator with two standard samples
(each of a different level of N02). This will raise the total number of
samples to be analyzed by each collaborator to 26: six from each of four
runs plus the two standard samples.
10. Each collaborator should analyze his samples at his home
laboratory according to the tentative method write-up identified in Item 1
above. Results should be recorded on the Nitrogen Dioxide Data Sheets used
in the field.
11. A copy of the completed data sheet on each sample as well
as all calibration data and a complete description of the rotometer and wet-
test meter used must be furnished MRI. All this information is needed by
MRI within 1 month after completion of the field test.
TEST INSTRUCTIONS
1. Prepare your six sampling trains. Place four in your area on
the spiked sampling table and two in your area on the unspiked sampling
table.
2. Prepare a data sheet for each sampling train.
3. Upon notification of "Start testing," from the MRI laboratory
supervisor, connect your trains to the proper ports and start your sampling
according to the procedure in the April 1974 write-up on "Tentative
87
-------
Method for Determination of Nitrogen Dioxide in the Atmosphere (TGS-ANSA)."
4. Upon notification, "Stop testing," from the MRI laboratory
supervisor, terminate test according to the procedure in the method write-up.
/
NOTES
88
-------
SPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS *
89
-------
UNSPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS *
90
-------
COLLABORATORS
Mr. Joe Andrews, Chief
Air Quality Section
Technical Services Program
Commonwealth of Kentucky
Department for Natural Resources
and Environmental Protection
Division of Air Pollution
Frankfort, Kentucky 40601
(502) 564-3382
Mr. W. (Dill) T. Shebe
Shell Development Company
Post Office Box 481
Houston, Texas 77001
'713) 667-5661
Mr. Donald Kutys
City of Philadelphia Air Manage-
ment Services Laboratory
1501 East Lycoming Street
Philadelphia, Pennsylvania 19124
(215) 288-5117
Mr. Dan Zorbini, Control Chemist
North Ohio Valley Air Authority
814 Adams Street
Steubenville, Ohio 43952
(614) 282-3908
Ms. Barbara Kladatos
Si nior Chemist (Air Pollution)
State of New York
Department of Health
Division of Laboratories & Research
New Scotland Avenue
Albany, New York 12201
(518) 457-3118
Mr. Lynn A. Hutchinson
Kennecott Copper Corporation
Utah Copper Division
Post Office JJox 11299
Salt Lake City, Utah 84111
(810) 322-1533
Mr. Ken Wilcox
San Bernardino County Air
Pollution Control District
172 West Third Street
San Bernardino, California 92415
(714) 383-1661
Mr. James T. Ward, Air Hygienist
Air Pollution Control Section
Environmental Health Division
Mecklenburg County Depament of
Public Health
1200 Blythe Boulevard
Charlotte, North Carolina 28203
(704) 374-2607
Mr. Kenneth Smith
Air and Industrial Hygiene Laboratory
State of California Health & Welfare
Agency
Department of Health
2151 Berkeley Way
Berkeley California 94704
(415) 843-7900
Mr. Bob Deardorff
National Bureau of Standards
B 326 Chemistry Building
Washington, D.C. 20234
(301) 921-2886
91
-------
MRI DERAMUS FIELD STATION - KANSAS CITY, MISSOURI
NITROGEN DIOXIDE DATA SHEET - TGS-ANSA METHOD
Sampling
Collaborator
Sampled by
Run Number
Sampling Port Number
Sampling Train No.
Bubbler Identification No.
Rotameter Description (make, model, etc.)
Rotameter Reading At Start
Pump Vacuum At Start
Start :Date _________ Time
Sampling Duration (min)
Total Air Volume Sampled (TO?)
Remarks:
At Finish
At Finish
Finish:Date
Type of Ball
Time
Sample Flow Rate (cm-*/min)
Analyzed by: Person
Analysis
Date of Analysis
Organization
Standardization Plot Slope (absorbance units/M-g N02/ml) ______
Absorbance of Sample (550 nm) _______________ Absorbance of Blank
N02 Concentration (p,g/m^) _______________ Aliquot (ml) _____
Remarks:
92
-------
Kansas City
International
Airport
(KG)
N
KANSAS CITY, MISSOURI
87th St.
Driving distance
from KCI to
Ramada Inn:
Approximately
35 mi.
RAMADA INN
Grandview, Missouri
93
-------
RAMADA INN
HOLIDAY
INN
FIELD STATION
94
-------
Robinson Pike Rd.
To Grandview
Ul
COLLABORATIVE
TEST SITE-BLDG.3
~\
Bldg
Bldg.5
Bldg. 6
DERAMUS FIELD STATION
MIDWEST RESEARCH INSTITUTE
GRANDVIEW, MISSOURI
-------
APPENDIX F
N02, AMBIENT-AIR SAMPLING SYSTEM OPERATION DATA:
TEST LOG SHEETS AND TEST DATA SHEETS
97
-------
System operational data are summarized in Tables F-l through F-4.
Readings were taken hourly during the tests and are recorded on the log
sheets which are located after Table F-4. The first nine columns of
Tables F-l through F-4 list various readings used in calculating flow-
rates and spike levels, which are given in columns 10-14. The last six
columns list various ambient air conditions at the test site. The
venturi and meter flowrates (columns 10-11) are calculated from the
calibration equations in Appendix C. Due to the temperature compensation
of the dry-gas meter and the above-ambient pressure of the gas stream
at these instruments, the flowrates are calculated at 21°C and 760 mm
Hg. The readings of the two devices are then averaged (column 12) and
the average flowrate is corrected to the temperature and pressure at
the sampling ports (column 13). Some of the methods being evaluated
with this system are not corrected for temperature and pressure. How-
ever, if the spike levels are not calculated at the temperature and
pressure existing at the manifold ports, a significant degree of un-
certainty enters into any subsequent use of the spike level. The spike
level (column 14) is determined from the permeation rates of the per-
meation tubes used in each test.
98
-------
Table F-l. RUN NO. 1 TEST DATA
\0
\0
Calculated flowrates and aplke levela
NO2 Sampling ayatem data
Boon Barometric Flow Flow
Date temperature pressure pressure temp.
Time CO (mm Hi) (mm Ha) CM
4-29-74
1120 22.5
1230 22.0
1330 21.0
1430 22.0
1530 -23.0
1630 22.0
1730 22.0
1830 22.0
1930 22.0
2030 22.0
2130 22.0
2230 22.0
2330 22.0
4-30-74
0030 21.0
0130 22.0
0230 22.5
0330 22.0
0430 22.0
0530 22.0
0630 22.0
0730 22.0
0830 22.5
0930 22.0
1030 22.5
1120 23.0
Average
£/ Temperature and
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
741
742
743
743
743
pressure at
8
B
8
g
8
8
8
a
8
8
8
8
8
8
8
8
8
8
8
8
8
a
8
8
8
aampllng
22.0
21.4
21.0
21.2
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
21.5
21.0
21.5
21.5
21.5
21.5
21.5
21.5
21.5
22.0
22.3
22.4
22.4
ports.
Flow-
meter
56.1
56.9
57.0
57.3
57.2
57.6
57.5
57.5
58.1
57.4
57.1
58.5
56.9
54.6
57.2
57.4
56.2
56.7
56.2
55.0
57.5
56.7
56.5
56.1
56.0
Venturl
»2
reading flowrate
(mm H.20) (cc/mlnl
241
248
247
253
251
253
254
255
257
234*'
237*'
267
248
235
255
251
242
243
240
239
255
252
248
245
244
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
Permeation
tube temp.
25.0
24.9
24.9
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
24.95
24.95
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
Venturl
Meter
to 21* to zi"
+760 mm Hg +760 on Hg
(1/mln) (t/mln)
56.2
57.2
57.1
57. g
57.4
57.7
57.8
57.9
58.1
K.&
55.7*'
59.4
57.1
55.6
58.0
57.5
56.4
56.5
56.2
56.0
58.0
57.6
57.1
56.8
56.6
57.1
56.5
57.3
57.4
57.7
57.6
58.0
57.9
58.0
58.5
.
58.9
57.3
55.0
57.6
57.8
56.6
57.1
56.6
55.4
57.9
57.2
57.0
56.6
56.5
57.1
Average £lovrate
to 21*
+760 mm Hg
(f/mln)
56.3
57.2
57.2
57.7
57.5
57.8
57.8
57.9
58.3
55.3*'
55.7*'
59.1
57.2
55.3
57.8
57.6
56.5
56.8
56.4
55.7
57.9
57.4
57.1
56.7
56.6
57.1
Amblenti'
(1/mtn)
58.0
58. B
58.7
59.3
59.2
59.5
59.5
59.6
60.0
56.9
57.3
60.8
58.8
56.7
59.4
59.2
58.0
58.4
57.9
57.2
59.5
59.0
58.6
58.3
58.1
58.7
Spike
level
ambient^
95.2
93.9
94.0
93.1
93.2
92.7
92.7
92.5
91.9
97.0
96.3
90.7
93.9
97.3
92.9
93.2
95.1
94.5
95.3
96.4
92.7
93.6
94.1
94.7
94.9
94.1
NO
back-
ground
(ug/m3)
20
20
10
10
10
10
10
20
15
10
10
10
10
10
10
30
10
10
10
10
10
30
30
40
30
16
NOj
Out-
emdltll
back- aoor wind
ground temp. speed
(ug/m3) CC> (a/sec)
90
80
10
20
10
10
10
20
15
10
20
50
50
50
50
60
SO
40
30
20
40
40
40
50
40
36
15.5
15.5
17.0
17.5
19.0
19.5
19.0
19.5
17.0
16.0
15.5
15.0
14.5
14.0
14.5
14.O
13.5
12.5
11.0
10.5
11.0
13.9
16.0
20.6
19 .O
4
4
6
6
7
7
5
0
0
0
0
0
o
0
0
o
7
5
6
IDS
Relative
Wind bumldlty
direction (X)
SEE
E
E
HE
HE
SE
SB
HE
E
HE
mi
H
HW
N
II
HW
HW
NW
H
HE
100
1OO
9S
90
85
85
85
80
80
84
89
95
1OO
90
95
95
89
90
84
76
77
70
70
61
53
FloH variation during thla perlod»ettlmated average flowrate.
-------
Table F-2. RUN NO. 2 TEST DATA
Calculated flowratea and spike levela
o
o
K>2 Sampling system data
Time
4-30-74
1207
1300
1400
1500
1600
1700
1800
1900
2OOO
2100
2200
2300
2400
5-1-74
0100
0200
O300
0400
0300
0600
0700
0800
0900
1000
1100
1207
Room
CO
22.0
22.5
23.5
23.5
22.0
23.0
23.0
23.5
22.0
22.0
22.0
22.0
22.0
22.0
22.
22.
22.
22.
22.
23.0
23.0
22.5
22.0
22.0
22.0
Barometric Flow Flow
(mi Ha) (mm Ha) Ccl
743
743
743
743
742
742
742
742
742
742
742
742
742
742
742
742
742
742
742
742
743
743
742
740
741
22.0
22.5
22.5
22.8
22.8
22.8
22.8
22.5
22.1
22.1
21.5
21.5
21.5
21.5
21.5
22.0
22.0
21.5
21.5
22.0
22.0
22.5
22.3
22.5
22.4
Flow
(Jt/min)
57.0
56.8
56.7
56.2
56.2
56.5
57.4
56.2
56.5
56.1
55.3
55.2
56.2
58.8
58.6
58.0
57.6
57.9
56.6
56.2
56.0
55.7
55.8
55.2
55.2
Venturi
(mm H20)
253
249
250
246
249
248
253
246
230
244
238
236
247
261
271
268
264
264
254
244
243
240
242
236
236
"2
(ec/mln)
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
600
" I'cT*'
23.0
25.0
25.05
25.02
2S.O
25.05
25.05
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25. O
25.0
25.O
25.0
25.0
25.0
25.0
25.0
25.0
25.0
Flowrate
Venturi
to 21*
+760 mm HR
q/mln)
57.8
57.2
57.3
56.8
57.1
57.0
57.6
56.8
57.3
56.6
56.0
55.8
57.1
58.8
60.0
59.6
59.1
59.2
58.0
56.6
56.8
56.1
56.3
55.4
55.5
Meter
to 21*
+760 mm Hs
(Jt/mln)
57.5
57.3
57.2
56.7
56.6
56.9
57.9
56.6
56.9
56.5
55.7
55.6
56.6
59.3
39.1
58.5
58.1
58.4
57.1
56.6
56.5
56.2
56.2
55.5
55.6
Average flotrrate
To 21*
+760 mm Kg
57.7
57.3
57.3
56.8
56.9
57.0
57.7
56.7
57.1
56.6
55.9
53.7
56.9
59.0
59.5
59.0
58.6
58.8
57.5
56.6
56.7
56.2
56.3
55.5
55.5
Ambient^
q/rnin)
59.2
58.9
58.9
58.4
58.6
58.7
59.5
58.4
58.8
58.2
57.3
37.1
58.4
60.6
61.1
60.6
60.2
60.3
59.0
58.2
58.2
57.8
57.9
57.2
57.2
Spike
level
. Tua/m^T
196.0
197.0
196.9
198.5
197.9
197.6
195.0
198.6
197.4
195.4
202.3
202.9
198.7
191.5
189.9
191.3
192.7
192.4
196.6
199.2
199.4
200.8
200.2
202.5
202.6
NO
back-
date3)
10
10
0
0
0
0
0
0
O
10
0
0
0
0
0
0
0
0
0
0
40
20
10
0
0
N02
back-
(J|/m3)
20
40
20
20
20
20
30
40
60
110
65
60
45
30
20
40
10
0
10
40
9O
70
30
10
10
Ambient conditions
Out-
door
Wind
CO (m/»ee)
20.0
20.5
21.0
19.0
19.1
20.5
20.5
19.5
16.0
11.5
11.0
10.4
9.0
9.0
8.0
8.S
8.5
7.5
6.5
9.0
15.0
18.8
19.1
24.5
19.2
4
4
6
5
4
3
3
2
0
0
0
0
0
0
0
0
0
O
0
0
0
2
3
7
8
Wind
direction
HE
NE
N
IW
NE
N
N
N
N
N
N
N
H
N
H
N
N
N
N
N
S
SE
SE
S
S
Relative
humidity
51
50
45
43
45
44
42
41
48
62
63
65
70
67
70
65
65
70
70
71
82
68
60
52
52
Average
57.3
57.0
57.2
58.7
197.3
mf Temperature and preasure at sampling porta.
-------
Table F-3. BIIN NO. 3 TEST DATA
IK>2 sampling system data
Koom
Date temperature
Tine CC»
5-1-74
1239
1330
1430
1530
1630
1730
1830
1930
2030
2130
2230
2330
5-2-74
0030
0130
0230
0330
0430
0530
0630
0730
0830
0930
1030
1130
1239
Average
22.0
24.5
25. S
26.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
21.0
22.0
21.5
22.0
22.0
22.0
21.9
preaaure
(m Hat
741
740
740
739
739
739
739
738
738
738
738
738
738
738
737
735
734
734
732
732
732
732
732
732
732
preasure temp.
(mm H.I CO
a
e
a
a
a
a
a
8
8
8
8
8
8
8
8
8
8
8
22.0
23.8
25.0
25.7
22.5
22.2
22.0
22.0
21.5
21.5
21.5
21.5
21.5
21.5
21.5
21.5
21.5
21.5
21.5
21.5
22.0
22.3
22.6
22.5
22.5
F Ion-
meter
U/mln)
54.3
53.4
54.8
56.2
57.6
56.7
57.5
57.2
55.3
37.2
58.5
58.4
59.1
57.4
57.5
57.9
38.0
56.0
56.5
55.8
55.0
56.1
55.1
55.9
55.8
Venturl
reading
(mm H20)
230
226
236
249
258
250
253
235
241
255
262
264
274
264
263
258
259
245
249
242
238
243
239
245
243
"2
f lowrate
(cc/mln)
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
220
200
200
200
200
Permeation
tube temp.
25.0
25.1
25.1
25.3
25.1
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
24.9
24.9
25.0
25.0
25.0
24.9
24.9
25.0
25.0
25.0
25.0
25.0
Calculated flowrates and spike levels
Flovrate
Venturl
to 21*
+760 mn Hg
(t/mln)
54.9
54.0
55.0
56.3
5B.O
57.1
57.5
57.7
56.1
57.8
58.6
58.9
60.1
58.9
58.7
57.9
57.9
56.2
56.6
55.7
55.2
55.7
55.2
55.9
55.7
56.9
Meter
to 21*
+760 mn Hg
54.7
53.7
55.1
56.4
57. 8
56.9
57.7
57. 3
55.4
57.3
58.7
58.6
59.3
57.5
57.6
57.8
57.8
55.8
56.2
55.5
54.7
55.8
54.8
55.6
55.5
56.5
Ambient conditions
Average flowrate
To 21*
+760 m Hg
54.8
53.8
55.0
56.4
57.9
57.0
57.6
57.5
55.8
57.6
58.6
58.7
59.7
58.2
58.1
57.9
57.9
56.0
56.4
55.6
54.9
55.7
55.0
55.7
55.6
56.7
AmblenL±'
U/mln)
56.4
55.
57.
58.
59.
58.
59.5
59.4
57.5
59.4
60.5
60.6
61. S
60.0
60.0
59. »
60.0
58.1
58.6
57.1
57.2
58.1
57.4
58.2
58.0
58.8
Spike
level
ambient!'
56.6
57.1
55.7
54.2
53.3
54.2
53.6
33.7
55.4
53.7
52.7
52.7
51.9
53.1
53.1
53.2
53. 1
54.9
54.4
55.1
55.7
54.9
55.6
54.8
55.0
54.3
NO
back-
ground
(un/m3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
20
10
0
0
0
1
W>2
back-
ground
(uB/m3)
0
0
0
0
10
10
20
20
30
20
0
0
50
10
30
20
0
0
10
10
50
70
30
20
10
17
Out-
UI_jl
door "nn*
temp. speed
CO (m/aec)
20.4
21.0
21.1
20.9
21.5
20
20
17.5
15.5
14.5
13.3
12.5
11.5
11.0
11.0
9.5
9.5
9.0
7.5
10.5
14.2
16.5
18
20.5
23.9
10
9
10
8
6
7
6
2
2
0
3
3
4
0
0
4
0
0
0
0
0
0
0
3
2
Wind
direction
S
SE
S
st
SE
SE
SE
SE
E
E
S
SE
E
SE
S
E
E
E
E
E
E
NE
E
E
S
Relative
humidity
»>
48
36
39
42
47
45
45
47
50
38
61
65
61
61
76
61
61
61
88
88
88
76
78
79
75
£/ Toper* t ure and pressure at samp Ling ports.
-------
Table F-4. BUN NO. 4 TEST DATA
Calculated flowrates and spike levels
N02 sampling syst
Date
Time
1304
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
2400
5-3-74
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
Room
temperature
CC)
21.1
22.2
22.1
22.1
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
22.0
21.3
21.0
22.0
22.0
22.5
22.5
22.5
22.9
22.1
22.1
22.0
21.2
Barometric
pressure
732
731
731
731
731
731
731
731
731
731
731
731
731
733
735
736
736
736
737
740
740
741
741
741
740
Flov
pressure
(mm Hn)
8
8
B
8
B
8
8
B
8
8
8
8
8
8
B
8
B
8
B
8
8
B
B
8
8
Flow
temp.
na_
22.5
22.5
22.5
22.5
22.5
22.5
22.0
22.0
22.0
22.0
21.
21.
21.
21.
21.
21.
21.
21.
21.5
22.0
22.0
22.0
22.2
22.1
21.8
em data
Flow-
rate
meter
U/min)
57.7
57.4
57.3
57.5
56.8
56.0
57.6
56.7
57.5
57.4
56.2
55.9
57.0
57.5
58.1
59.4
57.0
55.0
53.8
54.9
56.1
56.4
57.3
58.0
57.6
Flowrate
Venturl
pressure
reading
(ma H20)
257
259
251
255
249
238
250
254
256
254
240
236
253
256
267
270
264
244
240
238
24 S
250
255
263
256
N2
carrier
f lowrate
(cc/mln)
800
800
800
800
BOO
800
800
800
BOO
800
800
800
800
800
800
800
800
BOO
800
BOO
800
800
BOO
BOO
800
Pe meat Ion
tube tenqi.
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
24.95
25.0
25.0
25.0
25.0
25.0
25.0
Venturl
to 21*
+760 DID Kg
U/min)
57.3
57.5
56.5
57.0
56.3
55.0
56.5
57.0
57.2
57.0
55.4
54.9
57.0
57.3
59.0
59.4
58.7
56.3
55.9
55.8
56.6
57.3
57.9
58.9
58.0
Meter
to 21*
+760 DID Kg
U/mln)
57.4
57.0
56.9
57.1
• 56.4
55.6
57.2
56.3
57.1
57.0
55.8
55.5
56.6
57.3
58.0
59.4
57.0
55.0
53.9
55.1
56.4
56.8
57.7
58.4
57.9
Averase
To 21*
+760 mo Kg
57.4
57.3
56.7
57.1
56.4
55.3
56.9
56.7
57.2
57.0
55.6
55.2
56.8
57.4
58.
59.
57.
55.
54.
55.
56.
57.0
57.8
58.6
57.9
f lowrate
Ambient^
U/min)
59.9
59.8
59.3
59.6
58.9
57. B
59.3
59.1
59.6
59.5
57.9
57.5
59.1
59.6
60.6
61.5
59.8
57.5
56.7
57.2
58.2
58.7
59.5
60.3
59.7
Spike
level
ambient^'
298.4
298.5
301.3
299.3
303.2
308.9
301.1
302.1
299.4
300.3
308.2
310.4
301.9
299.7
294.8
290.8
298.5
310.2
314.9
312.3
306.7
304.2
300.2
296.0
299.3
NO
back-
ground
(ua/m3)
0
0
0
0
0
0
0
0
0
0
0
S
0
0
0
0
0
0
0
0
o
0
0
0
0
tK>2
back-
ground
(utt/n3)
10
0
0
0
0
30
20
20
50
45
50
60
20
40
10
0
0
0
10
20
10
10
10
20
20
Ambient
Out-
door
temp.
i^L
22.9
22.3
20.5
21.0
21.5
21.5
20.5
18.5
16.5
16.5
16.5
15.5
14.5
14.5
13.0
12.0
1O.O
9.5
9.5
12
20.6
19.0
21.5
17.0
16.4
conditions
Wind
speed
(m/sec)
6
5
4
6
6
5
6
3
2
0
0
0
0
14
11
7
a
11
9
7
9
6
10
8
6
Wind
direction
S
S
S
SE
SE
SE
SSE
SSE
SSE
SE
SE
SE
S
N
N
N
N
"
N
N
N
N
N
H
N
Belatlve
humidity
66
71
71
64
64
64
67
7S
84
84
84
84
84
80
89
84
82
71
70
76
70
70
52
48
40
Average
57.0
56.8
56.0
59.1
302.4
18
Temperature and pressure at sampling ports.
-------
LOG SHEET
Sidel
1 2 34 5 6 78 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
DATE V-.2
RECORDER 1 - PAPER
rk
y
IX
OPERATION
J\
x/
V
RECORDER 2 - PAPER
Of
vX
\x
OPERATION
C •*-
y
V
V
vX
.X
-------
LOG SHEET
Side 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
TIME
/ffc
n
11 10
f*
/.'Jo
tti'.lc
NO2 ANALYZER - PAPER
QU
OK
J
v/
\/
u/
IX
OPERATION
OK
J
y
v/
IX
VAC PRESSURE
21
il'
ir
O2 PRESSURE
So
to
20
.'.J
QO
21)
-3.0
UNSPIKEIjNO
1-2.
tl
ll
, I
/I
II
13
n
NO2
It
1 1
tl.
'5
£••//
/3
'1
/y
SPIKED -NO
II
13
II
I/
1 1
1 1
(i
n
II
13.
'Y
17
30
l/l'
zz
57
!-"•*
3ST
VENTURI MANOMETER - HI
538
5395
LO
ft*
£8*5
2*3
^77
J&.l
3S.6
VENTURI
INDOOR TEMP.
21
2.1
9-}
JJ1-
:».$•
22-
^25
23
GAS FLOW TEMP.
ffZ
22-
7/5
2/
91 f
GAS FLOW MANOMETER - HI
Irri
-fit
v4«H?
'/"'
Y1U
LO
•fto
IT!
371
57 S
T7./
73
37_?
2>0
53i
097
EXHAUST - U MANIFOLD
OK
y
S MANIFOLD
c>\
y
v/
CHECK SAMPLE TRAINS
tx
iX
3EP
59.1
INITIALS
S9.6
ST. 1
Sin
J'c
firs
-------
LOG SHEET
Sidel
3 4 5 6 7 8 9 10 11 12 13 14 15 16
17
18
19
20
21
22 23 24 25
DATE
TIME
Wat
fr
TM'.oo
<"
1 00 7
.oo
: 07
OUTDOOR TEMP.
20.0
XI
tfo.S
l\
'I
WET BULB
-S3
£-5"
51
SH
55
•fO
V?
DRY BULB
63
67
10
•fV
REL. HUMIDITY
57
ft
••s
7o
BAR. PRESSURE
7V.307V.ag
7Y3
WIND SPEED
/.r
2.1
/.S
0
0
0
.£
o
Ul
WIND DIRECTION
ve
IN/
N
M
POWER LIGHTS
J
y
N/
LEAKS PERMEATION
ASSEMBLY
y
y
J
y
BATH TEMP.
35
ROTAMETERS - 1
71
'— \
6'
f-Iyl
7*-
7
loo
h
i*
SI
too
N2 PRESSURE
7/0
700
rxo
9/O
RECORDER I - PAPER
\/
v/
OPERATION
J
RECORDER 2 - PAPER
V
v
OPERATION
I/"
i/
ux
V
Is
-------
LOG SHEET
Side 2
10 11 12 13-.
14
15
16
17
18
19
20
21
22 23 24 25
TIME
loo
Ifr-8*
/ 00
3 'Ofl
y <><»
6 eo
NO, ANALYZER - PAPER
y
vX
tX
OPERATION
v/
v/
J
VAC PRESSURE
f*
3S.5
21.5
v.\
O2 PRESSURE
PO
-I 0
-.*
20
20
Ao
10
10
UNSPIKED - NO
II
lo
lo
la
I0
10
/O
10
/o
10
NO2
>Y
IT
12
/a.
m
J
/v
17
II
/I
SPIKED - NO
(1
II
It
II
ii
1 1
II
ii
n
II
II
NO2
36
3S.5
10
3?
3?
o
OS
VENTURI MANOMETER - HI
SJ.v
53.3
57."
37
LO
18.5
IS.g
aat
2y.f 2X?
3-7.1
VENTURI *P
INDOOR TEMP.
225
2.3
i v
22.
32.4
t-3
33.
aa.
G*S FLOW TEMP-.
ATI
'a/.*
jfQ
tv
GAS FLOW MANOMETER - HI
ff.7
yj.7
18.S
LO
W 17,1
17?
17. T
37.6
•J77
-- -7
CAS PLOW PRC55UHC iuFr
7(0
&L\
.as?,
V/
$08.7 &
EXHAUST - U MANIFOLD
IT
v/
\*S
y
S MANIFOLD
y
v/
X
U
Is
CHECK SAMPLE TRAINS
t/
IS
v/
y
U/
v/
v/
INITIALS
s-..-/;
i- -
6 *
4-1.5
ROIt
CO-1/
OTTC,
- a
rf-
/&>.?
R:
-------
LOG SHEET
Sidel
8 9 10 11 12 13 14 15 16 17 18 19
20
21
22
23
24 25
DATE
TIME
m\&
Jii
OUTDOOR TEMP.
20'
3I.D3U
2L
(3.5
:=U>
WET BULB
55
53.
=>'
o
•*/
DRY BULB
7O
^JL
sa
5-7
REL. HUMIDITY
37
MS
4-
7?
BAR. PRESSURE
71 of
7-/.OC
73.5
7)77
75.21
J.30
WIND SPEED
4.1
7.3
3/1
o
o
1.0
WIND DIRECTION
SE
A/E
POWER LIGHTS
y
v/
v/
LEAKS PERMEATION
ASSEMBLY
is
v/
XX'
s/'
vX
BATH TEMP.
JK/
?r
2fj 2S-.9
ROTAMETERS -*C
100
(Ob
f4'
N2 PRESSURE
320
3oo
370
'2
"7/0
5 A'
RECORDER 1 - PAPER
IS
V
y
-y
v/
X
OPERATION
J
V
RECORDER 2 - PAPER
V
OPERATION
I/
v/
-------
LOG SHEET
Side 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
TIME
e •»„
•» V:
i.lA
?.*>
la1.??
NO2 ANALYZER - PAPER
V/
IX
OPERATION
y
I/
I/
\x
VAC PRESSURE
ar.r
20'
X."
rX"
25-. 3
O2 PRESSURE
9-0
ao
9.0
10
UNSPIKED - NO
/J
10
10
JO
I'J
10
/O
/O
/a
10
10
NO2
to
SO
10
It
lo
//';;
/ J
/7
13
SPIKED - NO
/o
10
10
n
1 _.
/o
10
i)
/a
/o
/a
II
10
to
NO2
17
ft
17
17
'1
n
/7
n
/7
1-7
t?
17
o
OB
VENTURI MANOMETER - HI
£H
£r>£&
OAf3.8
SJ.O
LO
Z?£
IB!:
35.2.
28.9
2Z.7
VENTURI AP
INDOOR TEMP.
20.
li.o
22.
it,
GAS FLOW TEMP.
O1.
?£2
ihT
?S PL-OW PREODUHg CL> Ff
7SS
^
533373
EXHAUST - U MANIFOLD
V/
S MANIFOLD
V
IX
CHECK SAMPLE TRAINS
V
INITIALS
f .- J fg. (J FT
Kcs
Kr'j
KCS
r-'-
PC/c
f" /.
91 .M
r* -
58.1
.7
&^v/^
-------
LOG SHEET
Sidel
10
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
DATE
TIME
if: --
U'.n
At..
// -J
OUTDOOR TEMP.
21
IRn
I'-r
Kj&iS
>\ jr./
IZ
[73
WET BULB
fra-
12-
-57
Cd
DRY BULB
67
67
67
7o
REL. HUMIDITY
7/
71
O
81'
76
-70
£7,2-
BAR. PRESSURE
73.15"
77.1*
n 10
73.08
71 oi
77, <
7V.J*
71.
WIND SPEED
t.7
/•Y
a./
a. I
-S
23
.0
WIND DIRECTION
Sfe
5£
sse
TV
,v
A/
A/
o
\o
POWER LIGHTS
y
vX
x/
y
tfeAKS PERMEATION
ASSEMBLY
J
y
y
y
y
Y
vi
BATH TEMP.
X.O
PS
5-.' •
for-'*
I-IAI
*< ?r
ROTAMETERS - 1
?7
?7
97
?7
r7
<&-
77
77
77
99
7?
11
Yi,'-
OS
n
9*
i1;
•h
17
77
N2 PRESSURE
3*10
RECORDER 1 - PAPER
\.
--ft
OPERATION
J
I/
J
RECORDER 2 - PAPER
OPERATION
IS
v/
y
I
L/
-*- .n Uj
-------
LOG SHEET
Side 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
TIME
/JJ
''.'53
/.'J
NO2 ANALYZER - PAPER
\y
tX
OPERATION
I/
VAC PRESSURE
.Xo
Oj PRESSURE
30. /
20 1
J J
UNSPIKED - NO
10
10
10
O
lo
/ 0
10
/o
NO2
U
O
/J
/O
\\
•'0
II
SPIKED - NO
10
/o
10
/i
If
I/
lo
i"
n
n
It
H
U
II
/O
NO2
17
V7
ft
So
H1
V5
5-0
*- VENTURI MANOMETER - HI
£17
53.?
SJ.(, »,3
an
C'J-
52,7
•Tf-7
£T3./
LO
27-1
It. I
.27 •»
WO
VENTURI
INDOOR TEMP.
2/.I
23.3
22
22.1
2 i.
•21-
32. J
ai.a.
GAS FLOW TEMP.
22.*
225
23.
ZZ
•33. .a.
GAS FLOW MANOMETER - HI
WS
mZLZ*
«*.(.
"»87
LO
7.
37,7
ITT
Tf.T
57 C.
377
•7,7
***Cu FT
HT7
(,11.
973
•y
ni
EXHAUST - U MANIFOLD
v/
S MANIFOLD
v/
T7
CHECK SAMPLE TRAINS
N/
J
V/
i/
INITIALS ^
5^- •-,.,/;
9. f"
\'c:
:i,r f?,s
5f./ «.£ ^1 5«jiU
il,-
59.
<-. t
^.Ac.S'K-
4'.?
6' i
-------
APPENDIX G
COLLABORATORS' COMMENTS
111
-------
During the orientation of the collaborators at the field site, they
were asked to provide MRI with comments on the method writeup, problem
areas, etc. Eight of the 10 responded with comments which are given be-
low by collaborator code.
Collaborator B
The scatter in the data is considerable for some of the runs. I ran
a few duplicates, as indicated, which gave a nearly identical pattern in
the scatter. It would seem to me that the problem was in the sampling end
of the test.
Several explanations have occurred to me. A leak in the sampling
system at the point of attachment to the N(>2 system would give low re-
sults. In addition, if the scrubbing action of the bubblers was inadequate,
this would also give a low result. Both seem to me to be possible explana-
tions for the apparently low values for some of the samples.
I reran the calibration curve. There was a slight change over the old
one, and the values for the new curve were used in the calculations.
Collaborator C
Absorbing solution in samples Sets 3 and 4 had mold growing in the
liquid when received.
Collaborator D
The water lost by evaporation during sampling was replaced in the
absorption tubes before they were packed for shipment. After shipping, 10
of the tubes had lost some liquid. The tubes were sealed in plastic bags
which had the liquid in droplets all over the inside. The caps on the
tubes seemed tight and there were no visible cracks in any tube. I could
not tell if the liquid had evapprated out of the tubes because of heat,
or if it had spilled out.
Two of the low volume tubes in the first days run were brought up to
50 ml using a buret to measure the amount of water needed. Analysis showed
these two tubes had lower concentrations than their companion tubes. Re-
calculating the concentrations, taking into account how much the solutions
112
-------
had been diluted, increased the concentrations to agree with their com-
panion solutions. Therefore, the remaining low volume tubes were analyzed
without making them up to 50 ml. All the low volume samples are desig-
nated on the data sheets.
Collaborator F
Analyses; Duplicate aliquots were run on the 24-hr samples and
triplicates on the MRI samples. The calibration curve was run in tripli-
cate. Although the calibration curve at low NC^ concentrations is not
linear, the variations at these points are within the limits of repli-
cate analyses.
For the 24-hr samples, the net sample absorbance ranged from 0.034
to 0.784 absorbance units. Based on 24 duplicate analyses the pooled
standard deviation of the mean for duplicate analysis is 0.0038 absorbance
units. This represents 0.0075 ug NC^/ml of absorber solution, or for
samples of approximately 0.3 nH of air, about 1.3 ug NC>2/tn of air. For
the spiked samples, this represents a possible error of 1-4% due to the
color development and absorbance measurement. For the unspiked samples,
this could account for errors of 10-20%.
The spiked samples from Port 8 are lower than the other replicates
in each case. For Runs 1 and 2 the differences are substantial, only
marginal in Run 3, and insignificant in Run 4. The maximum relative error
associated with the air flow measurements would be about 2%. The maximum
error expected in the analysis (including instrumental error) would be
about 3 ug/nr^ or about 3% for Run 1 and 1.5% for Run 2. Thus, for Runs 1
and 2, the samples from Fort 8 are obviously outliers. These are almost
certainly due to a leak in the sample train, probably at the connection
to the sample port. All other runs show good agreement, well within the
expected errors.
Comments; We have not encountered any significant difficulties in
applying this method and we plan to use it as a cross-check with the con-
tinuous chemiluminescence detector. The drawback to the method is the
limited shelf-life of the prepared solutions, plus the difficulty in ob-
taining £-methoxyphenol of the specified grade.
113
-------
Collaborator G
The samples collected and unknowns given were analyzed in strict ac-
cordance with the method. The only modification made in the procedure was
in the calibration of the rotameter used due to the unavailability of the
1 liter wet test meter required. A bubblemeter was used instead.
The overall method was simple and straightforward. No difficulty was
encountered in the analysis.
Collaborator H
All samples were run twice to verify results. Slopes of standard
curves were 0.44 and 0.38. This is not the 0.5 slope suggested by the
method. Analyses were performed exactly as outlined in the procedure.
Certain problems were experienced in sampling and analysis:
Sampling; Evaporation during sampling resulted in a loss of almost
50% of the fluid in the sampling tubes. This problem was most pronounced
on Run No. 1. It was not noticeable in the sodium-arsenite testing.
Calibration; Calibration of flow meters at our laboratory causes
a slight variation from flow determined at the test site. A difference
of 3,000 ft in elevation and different humidity might be the cause.
Analysis; Odd sized aliquots, such as 2.7 ml of sulfanilamide solu-
tion provide means where error might enter in. Standard pipet sizes of
2 or 3 ml would be found in all laboratories, and could be added in a
more rapid, accurate and reproducible manner. A change in solution
strength would accomplish this change.
It is interesting to note fritted glass bubblers were not allowed
because of expense, and "They are not needed." Yet in the method of
Continuous Saltzman, paragraph 5, it is noted that NC^, "Is somewhat
difficult to absorb."
Collaborator I
You will notice the slope is less than it should be. I tried a number
of times but got approximately the same slope (must be the dye).*
Subsequent communications have revealed that the poor results were
caused by contaminated methane1. See page 41.
114
-------
Collaborator J
Calibration; Since we did not have a 1-liter wet test meter and the
volume to be measured was small, the Fisher-Porter Company rotometer was
calibrated with a 250-ml soap bubble meter.
Observations; Upon opening the sample tubes on 14 May prior to
analysis, it was noted there was white, flocculent matter in all the
tubes of varying amounts. Absorbing reagent used in the test was found
to be clear. Tubes taken to the test site but not used were filled with
absorbing reagent and observed for clarity--no flocculence was noted.
The tubes had not been exposed to light. The absorbing reagent was pre-
pared on 26 April and was 17 days old at the time of analysis. All
chemicals used were of the purity specified by the method.
115
-------
APPENDIX H
ANALYSIS OF VARIANCE INCLUDING COLLABORATOR I
117
-------
One collaborator (I) produced useless results, but the analysis of
variance with (I) included is presented for completeness. Since (I) ob-
served virtually no N02» the average bias with (I) included depends on the
level of N02« Also, of course, the average bias is much larger with (I)
included (-24.6 yig/m^ versus -9.5 \tf>/ra~). The collaborator variance is
30 times as large when (I) is included.
ANALYSIS OF VARIANCE ((I) INCLUDED)
Source df
Total 153
C 9
L 3
CL 27
Error 114
Collaborator
Level
CL
Error
!SS MS F EMS
465,837.33
339,435.23 37,715.03 703+ ae2 + 16 ac2
23,651.13 7,883.84 147+ ae2 + 4 acc2 + 40 ac2
96,637.49 3,579.17 66+ ae2 + 4 acc2
6,113.08 53.62 - afi2
COMPONENTS OF VARIANCE
a2 (ug2/m6) a (us/in3)
2,353.84 48.52
107.62 10.37
881.39 29.69
53.62 7.32
118
-------
TECHNICAL REPORT DATA
(Plcatc read IUU.-IICIIOIH .HI the nvene ficjurc completing)
1 REPORT NO 2. 3. RFCI
EPA-650/4-74-046
4 TITLE AND SUBTITLE 5. REPO
Collaborative Test of the TGS-ANSA Method for Measure- Sep
ment of Nitrogen Dioxide in Ambient Air B.PERF
7 AUTHORIS) 8. PERF
Paul C. Constant, Jr., Michael C. Sharp, George W.
Scheil
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PHt'
Midwest Research Institute 1 H
425 Volker Boulevard " CON
Kansas City, Missouri 64110 68-
12. SPONSORING AGENCY NAME AND ADDRESS 13. TYP
Office of Research and Development . ..
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
IS. SUPPLEMENTARY NOTES
"IFNT'S ACCESSION-NO.
RT DATE
tember 1974
ORMING ORGANIZATION CODE
ORMING ORGANIZATION REPORT NO
GRAM ELEMENT NO.
A 327
TRACT/GRANT NO.
02-1363
E OF REPORT AND PERIOD COVERED
MSORING AGENCY CODE
16. ABSTRACT
A report on the collaborative test, by 10 organizations, of the "Tentative Method
for the Determination of Nitrogen Dioxide in the Atmosphere '(TGS-ANSA Procedure)" to
determine the precision and bias of the method. The report covers the NO., ambient-air
sampling system, test site, selection of collaborators, statistical design, collab-
orators' field sampling, their analysis of samples, statistical analysis of collab-
orators' results, conclusions and recommendations.
17. KEY WORDS AND DOCUMENT ANALYSIS
j DESCRIPTORS h.lDCNTIFIERS/OPEN ENDE
Air Pollution N02 - ambient air
Statistical Analysis sampling sy
Nitrogen Dioxide Collaborative Tes
Chemical analysis N09 permeation de
Design TGS-ANSA Method
Colorimetric
T-i DISmibUTION STATtMCNT 19 SECURITY CLASS (Ihn 1
Unclassified
Unlimited . » SECURITY CLASS ,!»,.,
Unclassified
D TERMS C. COSATI l-wlil/driiup
13B
stem 7B
t
vice •
(i-po't) 21. NO. O* PAG fc3
WKf) 32. PRICC
EPA Form 1Z?0-I (9-73)
119
------- |