EPA-650/4-75-013
February 1975
Environmental Monitoring Series
.•Xv.v.v.v.v.1
!:-:!:!:!:::!:^
li;l:lsiS^iS^ii^^l^^^
-------
EPA-650/4-75-013
COLLABORATIVE TEST
OF THE CHEMILUMINESCENT METHOD
FOR MEASUREMENT OF N02
IN AMBIENT AIR
by
Paul C. Constant, Jr., Michael C. Sharp
and George W. Scheil
Midwest Research Institute
425 Volkcr Boulevard
Kansas City , Missouri 64110
Contract No. 68-02-1363
ROAP No. 26AAF
Program Element No . 1HA327
EPA Project Officer: John H . Margeson
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, North Carolina 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D. C. 20460
February 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify tbit the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
I. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL MONITORING
series. This series describes research conducted to develop new or
improved methods and instrumentation for the identification and quanti-
fication of environmental pollutants at the lowest conceivably significant
ccm«.ontrations. It also includes studies to determine the ambient concen-
trations of pollutants in the environment and/or the variance of pollutants
as a I unction of time or meteorological factors.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/4-75-OJ3
11
-------
FOREWORD
This program, "Collaborative Testing of Methods for Measurement of
NO, in Ambient Air," is being conducted under the Environmental Protection
Agency Contract No. 68-02-1363, which is Midwest Research Institute's Proj-
ect No. 3823-C. The program is concerned with the evaluation of the fol-
lowing four methods with regard to their precision and accuracy:
1. Sodium-Arsenite,
2. TGS-ANSA,
3. Continuous-Colorimetric, and
4. Chemiluminescent.
The collaborative study covered by this report is of the Chemilumi-
nescent Procedure, which is a tentative instrumental method. In summary,
MRl's responsibility was to develop an IK^, 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 re-
sults from the collaborators, statistically analyze their results, and re-
port its findings to EPA. The 10 collaborators who participated in the
chemiluminescent collaborative test are:
Mr. Roger Ellard
Montreal Urban Community
Department of Air Purification and
Food Inspection
1125 Ontario Street East
Montreal 132, Quebec, Canada
Ms. Carol Ellis
National Environmental Research Center
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
iii
-------
Mr. William Find lay
Air Pollution Control Directorate
Environmental Health Centre
Tunney's Pasture, Ottawa, Ontario, Canada
Mr. Jim Harris
Air Pollution Control District
of Jefferson County
400 Reynolds Building
2500 South Third Street
Louisville, Kentucky 46208
Mr. John Higuchi
Air Pollution Control District
County of Los Angeles
434 South San Pedro Street
Los Angeles, California 90013
Dr. Melvin Ken Muir
Kennecott Copper Corporation
P.O. Box 11299
Salt Lake City, Utah 94111
Mr. Robert Richardson
Air Quality Evaluation Division
Measurement and Analysis Programs
Texas Air Control Board
8520 Shoal Creek Boulevard
Austin, Texas 78758
Mr. Vinson L. Thompson
National Environmental Research Center
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Mr. Jon Zimmer
Institute of Gas Technology
3424 South State Street
IIT Center
Chicago, Illinois 60616
Mr. Bruce DaRos
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
iv
-------
This report of test summarizes MRl's and the collaborators' activi-
ties. It describes the development of the NC^, ambient-air sampling sys-
tem, 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 for-
mal statistical design including the presentation of factors and parameters
that were considered, the collaborators' field sampling at the test site,
the collaborators' analysis results, MRl's statistical analyses of the col-
laborators' results, conclusions, and recommendations. Appendices contain
a copy of the tentative Chemiluminescent Method, information on the per-
meation tubes prepared for this program by the National Bureau of Standards,
written communiques with collaborators, instructions for collaborators, and
MRl's field, operational, and data log sheets.
Those individuals named above with the collaborating organizations are
acknowledged for their excellent work in the chemiluminescent collaborative
test.
Special acknowledgements are made to the National Bureau of Standards
and to Mr. Ernest E. Hughes and Dr. John K. Taylor of NBS who provided the
NO permeation tubes for this collaborative test; and to Or. John B. Clements,
Chief, Methods Standardization and Performance Evaluation Branch, National
Environmental Research Center, Environmental Protection Agency, and
Mr. John H. Margeson, Government Project Officer, Methods Standardization
and Performance Evaluation Branch for their valuable suggestions in planning
and design.
This MRI collaborative program was conducted under the management and
technical supervision of Mr. Paul C. Constant, Jr., Head, Environmental
Measurements Section of MRl's Physical Sciences Division, who was the pro-
gram manager. Those who contributed to this test are: development of the
N02 ambient-air sampling system - Dr. Chatten Cowherd, Jr., Mr. Fred Bergman,
Mr. Emile Baladi, and Mr. Wallace Yocum; experimental design and statistical
analysis - Mr. Michael C. Sharp; and preparation and operation of test facil-
ities - Dr. George W. Scheil, Mr. John LaShelle, Mr. Donald Gushing, and
Mr. Edward Cartwright, Jr.
Approved for:
MIDWEST RESEARCH INSTITUTE
Xr>_H. M. Hubbar'df Director
Physical Sciences Division
September 22, 1975
v
-------
CONTENTS
Page
Foreword iii
List of Figures ix
List of Tables x
Summary 1
Introduction 3
NO,, Ambient-Air Sampling System 5
General Concept 5
Design Factors 7
System Design 8
System Checkout 19
Ambient Levels of NO and N02 22
Subsystems and Units 22
System Operation 23
Test Site 25
Selection of Collaborators 31
Statistical Design 35
General Considerations and Comments 35
The Design 37
Collaborators' Field Sampling 39
vii
-------
CONTENTS (Concluded)
Page
Collaborators' Sampling Results 43
Statistical Analysis of Collaborators' Results 49
Analysis of All Spiked Readings 49
'Analysis of Spiked Ambient Readings 55
Summary Discussion 57
Lower Detectable Limit (LDL) 59
Conclusions 61
Recommendations 63
Appendix A - Tentative Method for the Continuous Measurement of
Nitrogen Dioxide in the Atmosphere (Chemiluminescent
Procedure) 65
Appendix B - Data on the Permeation Tubes Used as the Source of
the Spiked Levels of N02 87
Appendix C - Calibration of the Venturi and Dry-Gas Meter 89
Appendix D - Written Communications with Potential Collaborators. . . 93
Appendix E - Instructions for Collaborators NCv, Collaborative Test:
Chemiluminescent Procedure 101
Appendix F - N02> Sampling System Data: Test Log Sheets with
Field Operational Data Ill
Appendix G - Field Data 121
viii
-------
FIGURES
No. Page
1 N02, Ambient-Air Sampling System Concept ........... 6
2 Final Design of the N02, Ambient-Air Sampling System ..... 9
3 Annotated Photographs of the NC^, Ambient-Air Sampling
System in Operation ..................... 11
4
5
6
7
8
9
10
11
Photographs of the NO, Bleed-In Unit—Assembled and
Schematic Drawing of the N02 Permeation Tube Assembly
Schematic Drawing and Photographs of the Sampling Manifold . .
Collaborative Test Site: MRl's Field Station
14
16
18
?0
21
?6
?7
28
12 Photograph of Field Personnel of the N0« Collaborative Test
of the Chemiluminescent Procedure, MRI Field Station,
September 23 to 27, 1974 .................. 40
13 Collaborator -Level Interaction ................ 54
ix
-------
TABLES
No. Page
1 Collaborative Test Schedule 41
2 Average Results of Collaborators from their Sampling NO.
at Level 1 (48.6 ug/m3) 44
3 Average Results of Collaborators from their Sampling NO-
at Level 2 (186 ug/m3) 45
4 Average Results of Collaborators from their Sampling N02
at Level 3 (292 ug/m3) 46
5 Average Results of Collaborators from their Sampling N02
at Level 4 (102 ug/m3) 47
6 Analyses of Variance, Spiked Readings (ug/m3) 50
7 Collaborator Average Estimates Spiked Readings (ug/m3). ... 51
8 Collaborator Rank Orders (Ascending) Per N02 Level 52
9 Components of Variance, Spiked Readings (ug/m3) 52
10 Analysis of Variance Biases (ug/m3) 56
11 Collaborator (Average) Bias Versus N02 Level 56
G-l Level 1 Test Data 123
G-2 Level 2 Test Data 124
G-3 Level 3 Test Data 125
G-4 Level 4 Test Data 126
-------
SUMMARY
A collaborative test was conducted by Midwest Research Institute (MRI)
in the Greater Kansas City Area during September 23 to 27, 1974. Ten col-
laborators participated in this test of the "Tentative Method for the Con-
tinuous Measurement of Nitrogen in the Atmosphere (Chemiluminescent Proce-
dure}." All collaborators sampled from the N(>2i ambient-air sampling system
that was developed by MRI specifically for this collaborative test program.
For each of the four test days, a different average NO2 challenge (spike)
was used: 48.6, 102, 186, and 292 ug/m-*. These levels were obtained from
permeation tubes that were developed by the National Bureau of Standards
(NBS).
The collaborators sampled from both the spiked and unspiked (ambient)
lines of the NC^, ambient-air sampling system, providing three sets of
collaborator results per test day. The first set of data per 24-hr period
(a test day) comprised results where all 10 collaborators sampled from the
spiked line for approximately 14 hr (1800 to 0800). The second set of
data per test day comprised results collected for approximately 3 hr and
20 min (0930 to 1250) by the collaborators who were divided into two groups
of five collaborators each, with one group sampling from the spiked line
while the other group sampled from the unspiked line. The third set of
data per test day comprised results collected for approximately 3 hr and
40 min (1250 to 1630), with the two groups interchanging sampling lines.
These 12 sets of results were used for determining the bias and precision
of the collaborators' measurements.
In general, the measurement error (ae) is around 6% and the collabora-
tor variability Jff2 + a\ about 14% for the N02 concentrations examined.
These values (particularly cre) do depend on the N(>2 level.
The average bias is low (about -5%), and in general, stable versus
NC>2 level. However, one collaborator had very large biases, and another
collaborator had unstable biases (as a function of NOn level). Neverthe-
less , it is fair to say that for most collaborators (8 out of 10) using
the Chemiluminescent Method, the bias is small and well balanced.
-------
In summary then, collaborators using the Chemiluminescent Procedure
produce reproducible values with little bias on the average.
Two 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 £ 13 ug/m^, and the LDL from a set
of collaborators is s 22 ug/m^.
The major conclusion that can be drawn from the results of this col-
laborative test is:
The "Tentative Method for the Determination of Nitrogen Dioxide in
the Atmosphere (Chemiluminescent Procedure)" is adequately written for
those knowledgeable in sampling and analysis techniques presented therein.
Based upon the conclusion that has been drawn from the results of
this collaborative test, it is recommended that no further collaborative
testing and analysis be done on any of the four methods tested--sodium-
arsenite, TGS-ANSA, continuous colorimetric, and chemiluminescent.
-------
INTRODUCTION
The Methods Standardization and Performance Evaluation Branch,
National Environmental Research Center of the Environmental Protection
Agency (EPA), is engaged in a program to evaluate four methods for measur-
ing N02 in ambient air. Midwest Research Institute worked 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-colorimetric and chemiluminescent, which are instru-
mental methods.
To achieve this objective, a collaborative testing program was con-
ducted that will assess interlaboratory as well as intralaboratory testing.
In summary, MRI in the execution of this program selected the collaborators,
provided sampling locations and facilities thereon, oriented the collabora-
tors relative to the program, prepared a plan of test for each method tested,
scheduled testing, coordinated the collaborative tests, retrieved field data
and results of the collaborators' analyses, statistically analyzed results
received from the collaborators, and reported results of the program to EPA.
This report discusses the activities performed by MRI during the
fourth and final test undertaken on the contract. The method investigated
was the "Tentative Method for the Determination of Nitrogen Dioxide in the
Atmosphere (Chemiluminescent Procedure)," dated August 1974. A copy of
the write-up of this method is given in Appendix A.
The program was initiated on June 30, 1973, and this collaborative
test took place at MRl's field station in Kansas City, Missouri, during
September 23 through 27, 1974, with 10 different collaborators. The in-
terim period was devoted to the preparation for this test and conduction
of the first three collaborative tests, which covered the sodium-arsenite,
TGS-ANSA, and continuous colorimetric procedures. A major task of the
preparation activity was the development of a precise IK^, ambient-air sam-
pling 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 obtained
continuous analog N02 readings from the ambient-air sampling system. The
analysis phase covered the calculation of average N(>2 levels from the re-
corder charts and the statistical analyses of their results by MRI. After
the field test, the collaborators returned to their home laboratories
where they analyzed their recorder charts and reported their results to
MRI. Then MRI performed its statistical analysis and prepared this report
of the chemiluminescent collaborative test.
This report covers the collaborative test of the tentative Chemilumi-
nescent 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 section
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 sec-
tion summarizes the test activities during the collaborative test. The
seventh section discusses the analyses that were performed by the collabo-
rators. The collaborators' results are presented in this section along
with MRl's test data. The eight 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 sec-
tions present conclusions and recommendations, respectively.
The appendices contain a copy of the tentative Chemiluminescent Pro-
cedure, data on the permeation tubes that were used as the source of NO?
in the spiked section of the sampling system, information concerning 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 opera-
tional data, results of MRl'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: (a) that all collaborators sample the
same air; (b) that the samples be representative of ambient air; and (c)
that the concentration of NCL in the samples be accurately known and con-
trollable 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 somewhat an-
tagonistic to each other 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 addi-
tional requirement of fortuitous weather conditions, since weather condi-
tions have a strong effect on ambient N02 concentrations. An additional
problem with this approach is that no accepted primary reference method
exists for the analysis of NC>2 in ambient air.
However, gravimetrically calibrated N02 permeation tubes are avail-
able which generate a stable, precise rate of release of high purity N02
over a period of a few years. By using a set of these tubes, different
levels of NOo can be generated by adding the N0« from the permeation tubes
to a stream of air of a known flow rate. 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
N02 concentration of the spiked sample does not exceed the maximum level
of interest--350 ug/ra^—and to allow control of the spiked air N02 level
over a reasonably broad range, the average ambient levels must be well be-
low the lowest NO- concentration to be tested, in this case, 50 ug/nH.
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
-------
AMBIENT AIR INTAKE
CENTRIFUGAL
BLOWER
ROOF
z
_J
o
BAFFLE-
SPLITTER
PURGE LINE
SAMPLE
DRAV/-OFFp
EQUILIBRATION ,AMPI F
VENTURI SECT|ON >AMPLE
METER \
DRAW-OFF
RECORDER
0
CONTROL
VALVE ^-f
EXHAUST
RECORDER
SAMPLING
MANIFOLD
PERMEATION
TUBES
CARRIER
GAS
EXHAUST
PURGE LINE
Figure 1. NC^t ambient-air sampling system concept
-------
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 NC^ which is injected into the spiked section
at a desired level. The NO. 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 the spiked and unspiked sampling manifolds to moni-
tor the integrity of the spike. The collaborators sample ambient air si-
multaneously at an identical sampling manifold that is at a similar loca-
tion in the unspiked section. The gas in both sections is then exhausted
to the outdoors.
DESIGN FACTORS
The design of the NC^, ambient-air sampling system was based on the
following factors:
1. The flow rate of each of the four methods to be tested is approx-
imately 0.2 liter/min, with a maximum of 1 liter/min for some of the in-
struments that would be used in the instrumental methods.
2. The sampling period of each manual method is 24 hr; each instru-
mental method is preferably 24 hr, but could be less.
3. N02 permeation tubes whose rates are approximately 1 ug/min,
which are furnished by the government, are the source for the spiked lev-
els of N02- These tubes are to be operated at 25.1 + 0.2°C.
4. The number of collaborators for each collaborative test is to be
10.
5. The N02 range of concern is 50 to 350 ug/nr*, which is representa-
tive of ambient conditions.
6. There are four different N02 spiked levels: high, low, and two
medium. Each level is maintained throughout the run's period, within the
accuracy of the system.
7. The test period is to be no more than 6 days, which is based upon
the consensus of potential collaborators surveyed.
8. The overall N02 sampling system accuracy is to be 5% or better.
9. The flow control in the spiked section is to be 27, or better.
-------
10. Flow parameters of the spiked section are to be measured.
11. One NCL/NO chemiluminescent device, switched between spiked and
unspiked sampling manifolds (or stations), is to be used as a monitoring
instrument.
12. Only one person from each collaborator's organization will be
needed in the field for each method.
13. There is turbulent flow in the spiked section between the point
of injection of the spiked levels of N02 and the diffuser to provide mix-
ing of the spiked IK^ with the ambient air. The diffuser ensures proper
mixing. Up to 20% of the stream in each section--spiked and ambient air —
can be sampled to (a) ensure 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 quan-
tity of spiked flow to be drawn from the center of the spiked line where
there is assurance of equilibrium. There is to be a minimum amount of ad-
sorption of the spiked N02 on surfaces, 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 flow rate, which
allows for extremely short residence times, adsorptivity of NOo on sur-
faces and reaction with water vapor and other losses are insignificant.
14. Each section—spiked and unspiked--is to be similar, including
material and geometric aspects.
15. Each section is to be under positive pressure so that no unwanted
air will be pulled into the system in case there were a leak.
16. Collaborator's equipment size, configuration and power require-
ments must be met.
17. Environmental effects on operation of sampling system must be
considered.
SYSTEM DESIGN
The final design of the N02, ambient-air system is shown in a general
schematic form in Figure 2. Annotated photographs of this operational sys-
tem are given in Figure 3.
The input to the system is located outdoors about 2 m above ground
level and approximately 10 m from the building. A 1/4-in. orifice in 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
-------
vO
Monitoring Points:
1. Flow Temperature
2. Flow Pressure at Input
to Flow Meter
3. Ambient Air Flow
4. Flow Tempcro'ure
6. Carrier Gas Flow
7. N©2 Flow Temperature
8. Port Pressure
9. Port Pressure
10. NO2 & NO
5. Pressure Drop oF Venturi & Temperature II. NC>2 & NO
oF Pressure Transducer
Notes:
1. Component within DcsHed Area
Made of Teflon
2. Piping Out Sice Davied Area made
of Alumimn
3. Venturis made of Stcinless Steel
4. Spiked & Unspiked Lines Symmetric with
Respect to Geomc'ry & Material
Figure 2. Final design of the N02* ambient-air sampling system.
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
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 sys-
tem 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 aluminum
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 airstream 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 mea-
sured and recorded. This flow is determined by the following equation:
method sampling rate (number of samples x number
Flow in liter s/min = °£ collaborators + monitor number + purge number)
percent flow drawn through sampling manifold
0.2 liters/min x (4 samples x 10 collaborators +
_ _ 1 NO/N02 monitor + purge- line flow) _
percent flow drawn through sampling manifold
= 0.2 (4 x 10 + 1 + 4) = 9 = 6Q
0.15 0.15
The monitor number and the purge number are flows attributable to the con-
tinuous monitor and the purge line of the system, respectively. The flow
on each line — the spiked and unspiked --is turbulent- -Reynolds No. > 2,100
--with the Reynolds number being
R _ Q _ Q liter/min x 1.000 cm3/liter _
LVD 0.785 x 0.15 cm2/sec x D sec x 60 sec/min
* 60 4,000.
7.065 D 7.065 x 2.1
Since the spiked and unspiked sections are identical except that the
spiked section also contains the monitoring points 1, 2, 3, 4, and 5 iden-
tified in Figure 2 and the N02 permeation -tube system, only the spiked sec-
tion will be discussed.
From the splitter, the spiked line connects to a Singer AL-U5 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 flow rate of ambient air
13
-------
Ambient Air
Figure 4. Ambient-air stream splitter.
-------
delivered during a test run. This flow rate 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/rain. This venturi is used as a general flow-control de-
vice, and provides a continuous record of flow rate 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 trans-
ducer are recorded on analog recorders. Control of the flow rate is han-
dled by monitoring the venturi pressure drop. When the value deviates
from a reference value, 60 liters/min, the flow rate can be changed appro-
priately 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 0.5°C of room temperature. This
temperature measurement is used to obtain accurate gas-flow values.
To provide more accuracy, the thermocouples at Points 2 and 4 of Fig-
ure 3(A) were replaced for this test by a 0 to 50°C bimetallic dial ther-
mometer 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 1.0-in. diameter aluminum tubing. From the input of the
NO. 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 N02
(a spike) from the N02~permeation 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 hoIding/mounting plates. The gas
stream, or ambient air, enters the opening to the right and passes through
the unit, mixing with the spiked level of NOo 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 NOo gas from the permeation-tube assembly. This
spiked gas flows downward through this tube, which is inside the unit (see
15
-------
Photo 1 - Detail of N02 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 N02 bleed-in unit-
assembled and disassembled.
16
-------
Photograph 2 of Figure 5), and after a short run, mixes with the ambient
air as stated before.
The N0?-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 NO, into the
system. It is passed through a charcoal and soda-lime scrubber before it
is delivered to the NO? permeation tubes. Also, the flow is set by means
of control valves 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 pro-
vide K02 spike levels of approximately 50, 100, 200 and 300 ug/m3. 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
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 out-
let for the nitrogen carrier gas/N02 mixture. These N02 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 nitro-
gen flow were calibrated by the manufacturer to 17, accuracy. Thermometers
that were used to measure the gas temperature in the permeation-tube hold-
ers 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 hold-
ers to the pickup fitting, where the spiked gas enters the main gas stream,
was checked for leaks with Snoop and "found to be airtight.
JY "Operation Characteristics of N02 Permeation Device," 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, March 31 to
April 5, 1974.
17
-------
Nitrogen Gas
Shutoff Valve
Charcoal & Soda Lime Filter
Control Valves
1
2
) (
! n
n n
J Rotamel
il -«
«0
<^
IJ
J| _ NC>2 Permeation
Tube Holders
•Thermometers
^-Temperature Controlled
Water Bath
Control Valves
to NC>2 Bleed In Unit on
Spiked Line (See Figure 3)
Figure 6. Schematic drawing of the N02 permeation tube assembly.
18
-------
The ambient air and the NC^ spike flow from the bleed-in unit to the
diffuser where they are well mixed. The diffuser is a few centimeters
downstream 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 through 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
N02 flows through the bottom portion of the manifold into the exhaust line.
Section A of the manifold is in the pickup tube through which flows the
total volume of gas sampled by the collaborators. The inlet of this pick-
up tube is located such that this volume is drawn from the central portion
of the main stream. The sampled volume flows past a mixing impeller (B)
and then into the main chamber (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 port 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 contains 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 NC>2 and NO levels being sampled by the collaborators. A Bendix Model
8101 B chemiluminescence NO-lTC^-NOjj 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 prin-
cipal areas of activity: (a) determination of levels of NO and N02» both
ambient and inside the building; (b) checkout of the sampling system, in-
cluding monitoring devices and test instrumentation; and (c) checkout of
19
-------
c
1-1
(B
en
r>
o :r
Ml (B
ft 01
=T rt
(D H-
o
Q.
H- O.
M> l-t
H, B>
C S
0) H-
(D P
H 00
*
RJ
a
0.
f
rt
O
(W
co
Photo 1 - Top View Diffuser Components: Housing, End
Sections, Spiraler Tube, Teflon Screens, Retaining
Rings.
Photo 2 - External View of Diffuser.
BHL
Jffl
V.
///
HL
'///,
'///
///
'///
'///,
///
rr-r-r
'/A
D
''//,
C_? t_J
O 0
C © ©
oeoe
J-
L
j
i
'//
B
//
j
ffl~ IIIINU
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.
Y/////////////////////////////,
-------
Photo 1 - Sampling Manifold External View.
Photo 2 - Internal View (Right Component is Inverted
in this Photo).
ro
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.
-------
the sampling system as an operational system. These three areas are dis-
cussed below.
Ambient Levels of NO and N(>2
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 MRl's Bendix Model 8101 B chemilumi-
nescence 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 exceeded
20 pg/m . Periods of more than 1-hr duration were measured when readings
were indistinguishable from the purified zero gas used to calibrate the
analyzer.
With northerly winds, N02 levels were generally between 30 and 50
Ug/m3 and NO levels were approximately 10 ug/m3. 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 conditions
when the light vehicular traffic in the vicinity of the test station gener-
ated levels in excess of 100 pg/m3. NO levels did not exceed N02 levels
at this site.
Over a 24-hr period, average NO, levels were 10 to 50 pg/m3, and NO
levels were of the order 10 to 20 pg/m3. During any 24-hr period, maximum
N02 levels were generally several times higher than the minimum levels.
Thus, while N02 levels at the test site are lower than those at urban, in-
dustrial locations, the N02 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 ft3/min
wet-test meter. Information concerning the calibration is given in Ap-
pendix 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
22
-------
maintenance required. The temperature variation of the permeation -tube
bath during this time was less than 0.1°C. A check of NO levels in the
cylinders of prepurified nitrogen carrier gas found no NOo and 40 ug/m^
NO.
The Bendix NO analyzer was checked at MRI by a Bendix field repre-
sentative. The difference in spiked and unspiked readings of the Bendix
analyzer agreed within 10% of the calculated spike levels at all four lev-
els used for the test. The instrument was stable and reliable when oper-
ated continuously at the levels found during normal testing. Checks with
calibration gases reveal that the catalytic converter efficiency does fall
off sharply above 400
The symmetry of the sampling ports was checked in two ways. The pri-
mary 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 pres-
sure drop of 1.5 cm of water +0.5 cm. Such a pressure drop should have
no effect on normal sample flows and the flow rates from the ports should
be identical to that obtained by pulling free room air into the sampling
trains. However, since analyzers for this test required flow rates greater
than 200 cc/min, larger Teflon tube connections were provided on each mani-
fold. These lines were capable of supplying more than 2 liters/min with-
out developing a pressure drop of 1.5 cm of water.
A second way was to connect the NC>2 monitor to ports of the spiked
and unspiked sampling manifolds and measure the level of N02 in micrograms
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 NOo sampling system. In both cases, the NO 2 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 NOo sampling system. Handling and treatment of all
components were also identical.
Flow rates 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 to 4 mm water at the sample mani-
folds. 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
23
-------
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 flow rates 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 ug/m^ level
both readings were within 5 ug/nr* (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 sys-
tem may be considered to be inert with respect to NO?.
24
-------
TEST SITE
The general criteria one would use in selecting a site include the
ambient level of NC^ and variations 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 NC^ required are those representative of ambient N02
conditions, which are in the range of a few micrograms per cubic meter to
350 ug/nr*. These levels could be achieved at one site with a low level
of NO, by spiking the ambient air with various levels of NC>2 in a mani-
fold sampling system.
MRl'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 NO-, 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. Photographs of the facilities are given in Figure 10. Photograph
1 shows the circular tables that house the sampling manifolds (Photograph
2) and the collaborators' sampling trains. Each table—spiked and unspiked
--has a multiplicity of AC power receptacles, with each collaborator hav-
ing its own branch outlets. (Each branch has its own circuit breaker and
branch indicator.) This arrangement is to protect other collaborators in
case one collaborator has a power failure due to faulty equipment.
Photograph 1 of Figure 10 gives a close-up view of some of the col-
laborators' trains positioned in their table areas (see Figure 11). Since
each collaborator had only one instrument, Teflon or polypropylene tubing
was run through the 2-in. pipe, which spanned the two test tables, to the
other manifold (see Photograph 1 of Figure 10). This allowed the instru-
ments to remain in one place during tests and yet sample from either the
spiked or unspiked line by simply switching lines.
25
-------
to
DERAMUS FIELD STATION
Figure 9. Collaborative test site: MRI's field station.
-------
Photo 1 - Test Tables with Collaborators'
Instruments
Photo 2 - Close-Up of Sampling Manifold and
Collaborators' Lines Attached to It.
Figure 10. Test facilities and collaborators' instruments,
27
-------
SPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS*
* Unspiked manifold Layout similar (see Appendix F).
Figure 11. Collaborators' sampling areas at the test site
28
-------
The windows on the north side of the building were boarded to keep
electromagnetic radiation 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, ex-
cept for a few hours when the deviation was 0.1°C.
29
-------
SELECTION OF COLLABORATORS
A principal activity was to compile a list of potential collaborators
and from this list select 10 to perform the testing according to the ten-
tative Chemiluminescent Procedure. Information was obtained from EPA
(names and addresses of 150 organizations) and from MRl'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 program.
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 inter-
est in the Chemiluminescent Procedure after a test date was selected.
These collaborators were requested to submit certain information as well
as to do some ambient sampling according to the procedure and submit their
results for our evaluation. After the 10 collaborators were selected,
they were apprised by letter of their selection and given pertinent infor-
mation about the test. A copy of these letters and the collaborators 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.
Nine organizations* were selected for the Chemiluminescent collabo-
rative test from those organizations that responded in the affirmative to
participate in the test. The selection was based upon the following cri-
teria:
1. Willingness to participate on a volunteer basis.
2. Technical capabilities.
One organization provided two collaborators, each of whom had his own
equipment and worked entirely independent of each other.
31
-------
3. Related past experience.
4. Availability.
5. Ability to furnish sampling equipment, instruments, and materials
required to perform the test strictly according to the method.
6. Type of organization (industrial, educational, governmental--
local, state, federal—etc.).
The information needed to make the selection based on the above cri-
teria was obtained from the collaborator forms that were returned, and
from subsequent telephone conversations with the candidate collaborators.
The organizations selected as collaborators for the Chemiluminescent
collaborative test were:
Air Quality Evaluation Division
Measurements and Analysis Programs
Texas Air Control Board
8520 Shoal Creek Boulevard
Austin, Texas 78758
Mr. Robert Richardson
Institute of Gas Technology
3424 South State Street
IIT Center
Chicago, Illinois 60616
Mr. Jon Zimmer
Air Pollution Control Directorate
Environmental Health Centre
Tunney's Pasture, Ottawa, Ontario, Canada
Mr. William Find lay
Montreal Urban Community
Department of Air Purification
and Food Inspection
1125 Ontario Street East
Montreal 132, Quebec, Canada
Mr. Roger Ellard
32
-------
Air Pollution Control District
of Jefferson County
400 Reynolds Building
2500 South Third Street
Louisville, Kentucky 46208
Mr. Jim Harris
Air Pollution Control District
County of Los Angeles
434 South San Pedro Street
Los Angeles, California 90013
Mr. John Higuchi
Kennecott Copper Corporation
P.O. Box 11299
Salt Lake City, Utah 94111
Dr. Melvin Ken Muir
National Environmental Research Center
Environmental Protection Agency
Research Triangle Park
North Carolina 27711
Ms. Carol Ellis
Mr. Vinson L. Thompson
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Mr. Bruce DaRos
These organizations will be referred to as Collaborators A through J,
without specifying which is A, B, etc., to allow the organizational data
to remain anonymous.
33
-------
STATISTICAL DESIGN
GENERAL CONSIDERATIONS AND COMMENTS
The purpose of this collaborative test was to determine the precision
and bias of the Chemiluminescent Procedure. A major element of the col-
laborative 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 NO.,
2. Ambient levels of NO-,
3. True values of NO,,
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 N02 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/m3 and the second near
2QO ug/m3; and one high level of approximately 300 ug/m3. A challenge
35
-------
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 (approximately 50 ug/m^) . Since the ambient levels are the actual
ambient levels of N02 at the test site, those levels present during the
time of testing may vary from these criteria. 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 unspiked challenges.
For a run, the true value of N0£ sampled by the collaborators will be
taken as the N0« spiked level generated by the permeation-tube assembly
plus the average value of the ambient challenges sampled at the same time.
The error involved here adds to the overall error in the analysis.
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 statis-
tical significance with the results.
Adsorptivity is of concern because of the possibility of error in the
NO, level received by the collaborators' sampling devices in contrast to
the known level of the challenge- -from both the standpoints of increasing
and decreasing the challenge level from run to run. Teflon material was
used from the NOo 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 equilib-
rium 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 exper-
imental design because results of the evaluation of the NC^, ambient-air
sampling system indicated that all ports were identical.
The major considerations with regard to instrumentation for the Chem-
iluminescent collaborative test were: (a) MRI would only instruct the
collaborators that they are to use the sampling equipment and calibration
specified in the method write-up; and (b) MRl's monitoring instrumentation
and test instrumentation used in the calculation of the N02, ambient-air
system was sufficiently reliable and accurate. In both cases, all require-
ments were met.
36
-------
THE DESIGN
Since some spiked readings were being taken throughout the test, but
ambient readings were only sometimes obtained, there were two experimental
designs used.
One statistical model applies to all the spiked readings, but does
not incorporate any ambient observations. In other words, all-precisions
of interest were estimated, but no accuracies (biases) were determined.
This analysis of variance model is:
Xijk4 = H + C± + t. + Lk + eje(ijk)
where u = overall mean
C^ = ith collaborator, i = 1, .... 10
t. = jth hour, j = 1 20
1^ = kth N02 level, k = 1, . . . , A
measurement error in Jfcth reading in ijkth cell, i = I for
every ijk.
Since the N0« level may change from hour to hour, there are no repli-
cates in this framework (£ = one always). Also, some cells are missing
altogether (because sometimes a collaborator was on the ambient line and
did not get a spiked reading). Each collaborator measured the spiked line
only (a maximum of) 17 out of the 20 experimental hours. So all effects
have to be "adjusted" for this sample imbalance.
Therefore, the general analysis of variance was performed.* (In
practice, four such analyses were performed (one per level) because it
turned out that the repeatability of the method depended on the N02 level.)
The second experimental design model describes the data set of 6 to 7
hr/run when both ambient and spiked readings were taken (by different col-
laborators of course).
Since the hourly variation in ambient N02 is significant, a "true"
value was constructed for each of the 6 hr (per level). That is, for each
hour, the true value was estimated as the spiked NO, amount plus the
* See Appendix E for a discussion of this general analysis.
37
-------
average ambient reading in that hour. During these 6 hr, one-half of the
collaborators measured the spiked line and the other half observed ambient.
The roles ambient versus spiked are switched after 3 hr. An individual
response is a bias, i.e., the collaborator's reading minus the true value.
Thus, the data framework becomes three responses per collaborator* per
level.
Formally, the experimental design model is:
Xijk = » + Ci + Lj + ^ij + 6k(ij)
where u = overall mean
Ci = ith collaborator, i = 1, . . . ,9
Lj = jth level, j = 1, .... 4
CLji = collaborator level interaction
e v = kth measurement error in ijth cell, k = 1, . . . , 3 for
k(ii) ...
v •" every ij*
X.., = ijk bias (collaborator reading - true reading).
1JK
C-l not included due to anomalous results,
38
-------
COLLABORATORS' FIELD SAMPLING
The collaborative test took place at the MRI Deramus Field Station
during September 23 to 27, 1974. The 10 collaborators (see Figure 12)
started the test at 0830, September 23, with an orientation. The N02,
ambient-air sampling system they used was shown and explained to them.
The written instructions that comprise Appendix F were given to and dis-
cussed with the collaborators. After this orientation period the collab-
orators set up their equipment in preparation for the first run. The ac-
tual schedule of the four runs that took place is given in Table 1. All
10 collaborators cleared the site by 1900, Friday, September 27.
All collaborators sampled from the spiked line during the A runs.
During the B and C runs, the collaborators were divided into two groups
with one group sampling on the spiked line while the other group sampled
from the unspiked line. At 1250 each day all collaborators then switched
sample manifolds—those that had been on spiked went to unspiked and vice
versa. Also, the group that was on the spiked line for Run 1-B changed to
the unspiked line for 2-B, so that each group began every other B and C
run set on the same manifold.
During the test MRI personnel observed that all collaborators followed
the sampling procedures given in the method write-up. All collaborators
except two determined the nitric oxide start concentration of their span
gas cylinder prior to the test at their home laboratories. Two collabo-
rators could not obtain a gas cylinder in time and uncalibrated cylinders
were supplied by MRI on the 1st day of the test. These collaborators used
an arbitrary nitric oxide concentration during the test and then determined
the true value after returning to their home laboratories. Their field
data were adjusted accordingly.
Zero and span checks were made by all collaborators during the tests
at 0800 and 1630 each day.
The collaborators recorded all pertinent sampling data on their re-
corder charts. The calculations of the NO- levels from the recorder chart
readings were made after returning to their home laboratories.
39
-------
Front Row: Paul Constant ,3.' Melvin Ken Muir, Jim Harris, Carol Ellis,
Roger Ellard, William Findlay
Back Row: John Higuchi, Bruce DaRos,—' Vinson Thompson, Robert
Richardson, George Scheil,§/ Jon Zimmer
a/ MRI personnel.
Figure 12. Photograph of field personnel of the N02 collaborative test
of the chemiluminescent procedure, MRI Field Station,
September 23 to 27, 1974.
40
-------
Table 1. COLLABORATIVE TEST SCHEDULE
NO spike level Date/time
Level Run (ug/nr) Started Completed
1 A 48.6 9-23-74 at 1800 9-24-74 at 0800
B 48.6 9-24-74 at 0930 9-24-74 at 1250
C 48.6 9-24-74 at 1250 9-24-74 at 1630
2 A 186 9-24-74 at 1800 9-25-74 at 0800
B 186 9-25-74 at 0930 9-25-74 at 1250
C 186 9-25-74 at 1250 9-25-74 at 1630
3 A 292 9-25-74 at 1800 9-26-74 at 0800
B 292 9-26-74 at 0930 9-26-74 at 1250
C 292 9-26-74 at 1250 9-26-74 at 1630
4 A 102 9-26-74 at 1800 9-27-74 at 0800
B 102 9-27-74 at 0930 9-27-74 at 1250
C 102 9-27-74 at 1250 9-27-74 at 1630
41
-------
MRI had a laboratory supervisor who was in charge of the N02,
ambient-air system operation. He was on duty from 0800 to 1800 each day,
which was the period of run starts and completions. He was available any-
time during the 24-hr runs, as was the program manager, if any problems
arose.
There was a technician on duty throughout each run at all times dur-
ing the test. These people monitored the sampling system operation, re-
cording operational data and general observations. A general log book
was kept as well as log sheets for operational data. Copies of these log
sheets are given in Appendix F.
42
-------
COLLABORATORS' SAMPLING RESULTS
Each collaborator's sampling instrumentation included an analog re-
corder on which all his sampling data were recorded. Each collaborator
calculated 1-hr averages during the A runs and 1/2-hr averages during the
B and C runs, from his analog-sampling recordings by a method of his own
choosing. These results were submitted to MRI along with his calibration
data. The averages ot the collaborators are tabularized by N02 spiked lev-
el* in Tables 2 through 5, with Table 2 comprising Level 1 (48.6 ug/nr* of
N02> results, Table 3 comprising Level 2 (186 ug/nr* of IKk) results,
Table 4 comprising Level 3 (292 ug/nr* of IKL) results, and Table 5 com-
prising Level 4 (102 ug/m^ of IK^) results. For Runs B and C shown in
Tables 2 through 5, two consecutive half-hour average values were averaged
arithmetically to provide 1-hr averages. The 1-hr averages were the val-
ues used in the statistical analysis.
Each table presents the results of the 10 collaborators for a test
day. Explanatory notes are given at the bottom of each table.
MRI checked the collaborators' results for any gross overall error,
e.g., misplacement of the decimal point. Minor deviations were attributed
to the reading of the analog charts.
The NO., sampling system data along with calculated flow rates and
spike levels of the system, and data on ambient test conditions are given
in Appendix G.
The level value of N02 is that generated by the permeation tubes.
43
-------
Table 2. AVERAGE RESULTS OF COLLABORATORS FROM THEIR SAMPLING N02 AT LEVEL 1 (48.6 ug/m3)-/
Collaborators
Run
A
B
C
a/
b/
c/
d/
Date
9-23-74
9-23-74
9-23-74
9-23-74
9-23-74
9-23-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9_24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
This is the
60 ug/m3.
Time
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
2400-1000
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0930-1000
1000-1030
1030-1100
1100-1130
1130-1200
1200-1230
1230-1300
1300-1330
1330-1400
1400-1430
1430-1500
1500-1530
1530-1600
1600-1630
A B
56 61
66 66
94 103
113 117
94 84
66 70
66 66
66 66
66 70
66 66
66 61
66 66
66 70
75 80
gc.d/ l4£j
-------
Table 3. AVERAGE RESULTS OF COLLABORATORS FROM THEIR SAMPLING ND2 AT LEVEL 2 (186 ug/m3)!/
-P-
Ln
Collaborators
Run Date
A 9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-24-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
B 9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
C 9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
Time
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
2400-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0930-1000
1000-1030
1030-1100
1100-1130
1130-1200
1200-1230
1230-1300
1300-1330
1330-1400
1400-1430
1430-1500
1500-1530
1530-1600
1600-1630
A
197
197
197
197
197
197
197
197
197
197
197
197
197
207
207c/
216
216
216
216
216
216
19k/
9k/
9k/
ok/
ok/
ok/
ok^/
B
197
202
197
197
197
197
202
197
197
197
197
202
202
207
193c/
197
202
202
202
197
197
14k/
9b/
5k/
sk/
ok/
ok/
ok^/
c
184
188
188
188
186
190
190
186
186
188
186
192
188
195
195c/
205
205
201
201
199
205
24k/
17k/
13k/
iik/
iik/
iik/
llb.fi/
D
165
169
167
165
165
165
167
165
167
165
162
164
165
167
184£/
178
190
190
180
177
173
34k/
30k/
22k/
22k/
21k/
2lk/
Ub.fi/
E
169
164
164
165
156
162
160
160
156
160
156
162
156
158
171c/
188
180
180
180
182
194
28k/
28k/
24k/
19k/
19k/
19k/
19bjc/
F
154
150
148
148
147
147
147
145
145
147
145
150
150
154
15b.c/
13k/
isk/
13k/
13k/
13k/
isk/
.
_
_
_
_
_
-
G
195
195
194
194
192
195
194
188
188
192
188
197
197
197
15b.c/
19k/
24k/
24k/
22k/
isk/
24k/
199
197
197
197
197
197
197£/
H
197
197
197
197
197
197
197
197
197
197
197
201
199
207
9k*£/
15k/
19k/
19k/
19k/
19k/
19k/
207
197
197
197
197
197
197£/
I
167
162
162
162
160
160
160
158
158
158
158
164
164
167
15k»c/
17k/
19k/
19k/
19k/
19k/
2lk/
178
169
165
164
164
164
J
—
177
177
177
177
177
177
177
177
177
177
182
173
-
6b.c/
6k/
6k/
9b/
nk/
2ik/
2Lk/
194
224
188
188
188
188
a/ This is the spiked value—the statistically determined average true value of NO2 (which includes ambient NC>2) is 203
Ug/nr*.
b/ From unspiked samples—all other results are spiked samples.
c/ These values not used in the statistical analysis.
-------
Table 4. AVERAGE RESULTS OF COLLABORATORS FROM THEIR SAMPLING N02 AT LEVEL 3 (292 ug/m3)£/
Collaborators
Run Date
A 9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-25-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
B 9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
C 9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
Time
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
2400-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0930-1000
1000-1030
1030-1100
1100-1130
1130-1200
1200-1230
1230-1300
1300-1330
1330-1400
1400-1430
1430-1500
1500-1530
1530-1600
1600-1630
A
320
329
338
338
320
320
320
320
320
320
320
320
320
329
19ki£/
gb/
gb/
gb/
gb/
gb/
ok/
320
320
320
320
320
320
320£/
B
—
-
310
310
296
296
296
291
291
291
291
291
296
301
Q D » C/
9e/
gb/
5k/
sk/
ok/
ok/
296
296
291
291
291
291
296£/
C
286
288
299
293
280
276
280
276
276
276
276
278
280
284
2lk*£/
19k/
17k/
17k/
13k/
iik/
iik/
284
284
280
280
280
280
280£/
D
282
308
382
391
368
385
382
376
376
363
368
368
368
368
15bic/
17b/
isk/
iik/
Ilk/
ak/
ak/
258
254
254
259
259
258
261£/
E
258
244
254
248
246
242
231
244
252
254
254
254
256
258
47k/*
47k/
43k/
43k/
36k/
38k/
374
357
368
338
297
286
283£/
F
239
244
242
233
231
229
229
229
229
229
231
233
233
28k/
28k/
28k/
22k/
19k/
13k/
256
256
259
244
259
261
26 L£/
G
274
282
288
286
274
272
272
269
271
271
267
272
274
278
24 L£/
239
237
237
235
233
235
i9k/
isk/
isk/
isk/
isk/
isk/
H
318
320
327
325
312
308
306
301
301
303
301
301
308
316
27 8£/
280
282
278
276
278
278
gb/
gb/
gb/
gb/
6k/
8b/
I
267
269
274
272
263
261
261
259
259
261
259
259
261
263
320£/
320
323
320
320
310
310
ok/
ok/
ok/
ok/
ok/
Qb/
\J^
ok^c/
J
282
282
295
295
282
280
282
282
282
282
282
282
282
282
282£/
295
295
295
*• J J
295
295
295
28k/
28k/
28k/
28k/
28k/
3^
a/ This is the spiked value— the statistically determined average true value of NO
308 ug/m3.
b/ From unspiked samples — all other results are spiked samples.
£/ These values not used in the statistical analysis.
(which includes ambient NO ) is
2
-------
Table 5. AVERAGE RESULTS OF COLLABORATORS FROM THEIR SAMPLING N02 AT LEVEL 4 (102 ug/m3)£/
Collaborators
Run Date
A 9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-26-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
B 9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
C 9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
9-27-74
Time
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
2300-2400
2400-0100
0100-0200
0200-0300
0300-0400
0400-0500
0500-0600
0600-0700
0700-0800
0930-1000
1000-1030
1030-1100
1100-1130
1130-1200
1200-1230
1230-1300
1300-1330
1330-1400
1400-1430
1430-1500
1500-1530
1530-1600
1600-1630
A
94£/
132
122
113
113
113
113
113
113
113
113
113
113
113
U3c/
103
113
122
113
113
113
ok/
ok/
ok/
ok/
ok/
ok/
oki£/
B
84^/
117
108
108
103
103
103
99
99
103
103
99
99
103
94£/
94
103
108
103
103
103
ok/
ok/
ok/
ok/
sk/
sk/
5bjc/
C
88£/
116
107
107
102
103
102
100
100
103
100
98
98
105
ioo£/
100
102
111
102
102
100
ilk/
iik/
iik/
iik/
sk/
iik/
nb.c/
D
113S/
137
133
128
128
126
137
137
132
128
126
122
120
122
103£/
103
109
115
107
109
103
28k/
28k/
sok/
28k/
24k/
28k/
28ki£/
E
was/
107
102
'94
90
92
83
83
86
94
88
90
88
94
86c/
88
81
92
94
92
96
19k/
2lk/
2ik/
22k/
17k/
17k/
19k*£/
F
81c/
102
94
92
86
86
86
86
86
90
86
84
86
88
gb.C/
6k/
9k/
isk/
6k/
6k/
6k/
84
84
86
84
84
86
86£/
G
86c/
111
103
103
94
94
100
98
98
102
98
94
98
94
/|b.c/
ok/
4k/
9b/
zk/
2k/
ok/
98
96
100
102
100
100
10 2£/
H
92£/
118
103
103
103
103
-
-
90
92
94
94
94
94
—
_
.
—
—
-
—
_
—
—
—
-
I
94£/
113
100
100
94
98
94
94
94
98
94
92
92
94
9b.c/
ak/
9k/
19k/
9k/
sk/
sk/
96
96
96
98
96
96
96£/
J
103c/
122
116
116
113
113
113
113
113
116
113
103
103
113
6b.c/
6k/
8k/
sk/
9k/
9k/
gk/
102
105
105
105
105
105
107£/
a_/ This is the spiked value—the statistically determined average true value of NC>2 (which includes ambient N02> is
113' ug/m3.
b/ From unspiked samples—all other results are spiked samples.
£/ These values not used in the statistical analysis.
-------
STATISTICAL ANALYSIS OF COLLABORATORS' RESULTS
The analysis of the spiked readings and Che analysis of the ambient
readings will be discussed separately. A summary discussion will follow.
ANALYSIS OF ALL SPIKED READINGS
Recall that the experimental design model for this set is:
XijU - » + Ci + 'j + Lk + e*(ijk)
where u = overall mean
Ct = ith collaborator, 1 = 1, . . . , 10
t. = jth hour, j = 1, . . . , 20
Lk = kth N02 level, k = 1, . . . , 4
= measurement error in i reading in ijk cell, i> - 1 for
. .,
every ijk
= ijk£tn response (spiked reading).
The results of the analysis of variance of the spiked readings are
shown in Tables 6 through 8.
Table 6 shows the basic analyses of variance themselves. Note that
a separate analysis of variance was done for each level. This was nec-
essary because the variability within a collaborator (described by cre)
was not the same at all N02 levels (see also Table 9).*
* An assumption (the homoscedastic assumption) of all analysis of vari-
ance models is that a is uniform.
49
-------
Table 6. ANALYSES OF VARIANCE, SPIKED READINGS (ng/m3)
Source
Level 1 *
Total
Collaborators, adjusted
Time (t) , adjusted
Error (e)
Level 2
Total
C, adjusted
t, adjusted
e
Level 3
Total
C, adjusted
t, adjusted
e
Level 4
Total
C, adjusted
t, adjusted
e
df
8
13
104
165
9
19
135
167
9
19
137
155
9
19
115
SS
674,178
10,723
24,868
1,147
5,483,037
47,869
3,858
1,945
14,038,750
227,039
11,194
50,302
1,677,020
17,687
3,446
1,987
MS
1,340
1,913
11.03
5,319
203
14.41
25,227
589
367.17
1,965
181
17.28
F
121+
173+
369+
14.1
68.71
1.60
133+
10.5
* Collaborator F deleted as an outlier at level 1.
50
-------
Table 7. COLLABORATOR AVERAGE ESTIMATES* SPIKED READINGS (jig/m3)
Collaborator
A
B
C
D
E
F
G
H
I
J
Level 1
(avg 79)
70
72
73
89
65
174**
68
64
55
60
Level 2
(avg 181)
199
197
189
166
162
149
193
197
161
180
Level 3
(ave 287)
330
303
289
366
258
225
280
316
269
290
Level 4
(avg 103)
115
104
103
125
91
88
99
98
96
111
* The sample averages per level per collaborator are not unbiased estimates of the true collaborator
means, because a collaborator missed some hours during a day. The sample average differences be-
tween collaborators are not unbiased estimates of the true collaborator differences because dif-
ferent collaborators missed different hours. The values shown in the table are the unbiased esti-
mates (produced by the general analysis of variances).
** Not included in the analyses of variance.
-------
Table 8. COLLABORATOR RANK ORDERS (ASCENDING) PER N02 LEVEL
Ol
N>
Collaborator
F'
A
B
G
H
C
I
D
E
J
Source
"e
ac
at
/ 2 2
J" + o
e c
LI
10
6
7
5
3
8
1
9
4
2
Table 9.
Level 1
(ava 79)
3.32
9.74
14.54
10.29
L2
1
10
8
7
9
6
2
4
3
5
COMPONENTS OF VARIANCE, SPIKED READINGS
Level 2 Level 3
(ave 181) (ave 287)
3.80 19.16
16.29 35.26
4.34 4.71
16.73 40.13
L3
1
9
7
4
8
5
3
10
2
6
(Pg/o3)
Level 4
(ave 103)
4.16
9.87
4.05
10.71
L4
1
9
7
5
4
6
3
10
2
8
Average
(avg 163)
10.12
20.62
8.20
22.97
-------
All the F-values in Table 6 are significant, i.e., at all levels of
NC>2 the collaborator averages are separated and a significant variability
in NOn exists in time.
Since the C and t effects are significant, it is desirable to quan-
titatively describe the differences between collaborators. Table 7 dis-
plays the averages for each N0« level. The average N0« value per level
is listed in this table in order to give a rough idea of the relative
separation of collaborators (even though the average value is not a
"true").
Since the collaborator (mean) differences are often quite large, at
least some of the collaborators must be biased significantly. Note that
Collaborator F produced a very strange reading at Level 1, and was re-
jected as an outlier (despite the fact that the true NO, values are un-
known). Collaborator F was not deleted from the other NC^ levels.
Also, the order of the collaborators seems to vary quite a bit from
level to level. The Kendall concordance (a coefficient of agreement
equal to one when order is perfectly preserved) is only 0.53. Therefore,
it is reasonable to suppose that a significant collaborator-level inter-
action exists (see Figure 13).
The components of variance are shown in Table 9. Recall that:
a = standard deviation within a collaborator,
a = standard deviation between collaborator means,
o~t = standard deviation between hourly means.
Although we do not know the exact NO. values per level, it is surely
true that in descending order the levels are L4, Ll, L3, L2; approximately
spanning the range 80 to 300 ug/nr*. Therefore, Table 9 indicates that CTC
is proportional to the N02 level. The ae component varies with the N02
level, but not in a simple way.
Also, a is larger than a ; i.e., the variability between collabora-
tors is greater than the variability within a collaborator.
As a point of interest, note that crt also varies but is not propor-
tional to the NOn level; that is, the ambient fluctuations were not con-
sistent from day to day.
53
-------
40
30
20
10
<
CQ
z o
LU
U
0£
-10
-20
-30
-40
I
I
E&l
LI
L4 L2
NO2 LEVEL
L3
Figure 13. Collaborator-level interaction.
54
-------
ANALYSIS OF SPIKED AMBIENT READINGS
The experimental design model for this data set is:
Xijk = u + C. + L. + CL.j + ek(ij)
where u = overall mean
Ci = ith collaborator, C = 1, . . . ,9*
L. = jth N02 level, i = 1, . . . , 4
CL.. = collaborator level interaction
e = k measurement error in ij cell, k = 1, . . . , 3 for
kUjJ every ij
X.., = ijkth bias (collaborator reading - true value).
1JK
Recall that a true value for each hour is constructed from the aver-
age ambient observation during that hour. Since only half the collabora-
tors were measuring the ambient at a given time, their average value has
this source of error in it.
The analysis of variance is shown in Table 10. All the F-values in
Table 10 are highly significant. Therefore, the bias does differ between
collaborators, does depend on the N02 level, and the individual collabora-
tor's bias curves are not parallel.
The biases of Collaborator F varied from -24% to +125% at various
NO levels. These values are far larger than those exhibited by any
other collaborator, so Collaborator F was deleted from the analysis.
This complicated situation is displayed in Figure 13 and Table 11.
The overall average bias is -8 ug/m , or about -5%. The absolute
bias is the most at the highest N02 level, but the percentage bias is most
at the lower N02 levels.
Although the CL interaction term is significant, this is mostly due
to collaborator D, who exhibited biases ranging from -11 to +38% at vari-
ous N02 levels. In general, the collaborators exhibited farily stable
* One cell (three observations) was missing; these data were replaced
by the usual criteria of minimizing residual sum of squares.
55
-------
Table 10. ANALYSIS OF VARIANCE BIASES (ug/m3)
in
Source
L
C
CL
e
3
8
23
69
SS
1,070.74
18,015.17
19,045.43
733.33
MS
356.91
2,251.90
828.06
10.63
F
33.58
211.88
77.91
Table 11. COLLABORATOR (AVERAGE) BIAS VERSUS N02 LEVEL
Collaborator LI
A 0
B -1
G -4
H -17
C -2
I -16
D 23
E -7
J -13
( 60)8-/
( -2)
( -7)
(-28)
( -3)
(-27)
( 38)
(-12)
(-22)
L2
12
-4
-4
-3
-1
-34
-23
-20
-7
(203)a/
( 6)
( -2)
( -2)
( -2)
( -D
(-17)
(-11)
(-10)
( -3)
L3
14
-14
-32
7
-25
-50
53
-50
-15
(308)a/
( 5)
( -5)
(-10)
( 2)
( -8)
(-16)
( 17)
(-16)
( -5)
L4
3
-7
-16
(113)8-/
( 3)
( -6)
(-14)
missing
8
-19
-3
-20
-10
( 7)
(-17)
( -3)
(-18)
( -9)
Avg
7
-7
-14
-4
-5
-30
13
-24
-11
U63)a/
( 4)
( -4)
( -9)
( -2)
( -3)
(-18)
( 8)
(-15)
( -3)
Average
-4 ( -7)
-9 ( -5)
-12
( -4)
-8 ( -7)
-8 ( -5)
a/ True value in micrograms per cubic meter.
b/ Numbers in parentheses are percents.
-------
biases over the range of N02 examined, and therefore averaging the biases
was justified.
SUMMARY DISCUSSION
In general, the measurement error (o*e) is around 6% and the collabora
tor variability \la% + cr| about 14% for the N0~ concentrations examined.
These values (particularly ac) do depend on the N02 level.
The average bias is low (about -5%), and in general, stable versus
NOo level. However, one collaborator had very large biases, and another
collaborator had unstable biases (as a function of N02 level). Neverthe-
less, it is fair to say that for most collaborators (8 out of 10) using
the Chemiluminescent Procedure, the bias is small and well balanced.
In summary then, collaborators using the Chemiluminescent Procedure
produce reproducible values with little bias, but the level differences
between collaborators are often sizeable.
57
-------
LOWER DETECTABLE LIMIT (LDL)
Two meanings of LDL are used in the following discussion: (a) the
smallest value of N0_ 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 NOo estimate from a set of collaborators using
the method (a "practical" LDL).
Two methods of estimating the LDL were used. The first method used
was the ambient readings obtained during the actual experiment, while the
second method uses the collaborators' calibration curves.*
The ambient readings furnish estimates of crg (standard deviation within
a collaborator) and CTC (standard deviation between collaborators) that allow
estimation of the LDL's, although there is no way to incorporate bias** into
these estimates. Using a& = 4.69 ug/m3 and CTC = 6.80 ug/m3, results in:
o
Estimated pure LDL = 9 ug/m
Estimated practical LDL = 16 ug/m3.
The calibration curves do allow estimation of biases in addition to
components of variance. Using the average S.E. of individual calibration
curves results in a pure LDL estimate of 13 ug/m3, of which 5 ug/m3 is
bias. Using the whole data set, one arrives at an estimated practical
LDL of 22 ug/m3, of which 5 ug/m3 is bias.
The two sets of results agree fairly well. It seems reasonable to
state that the pure LDL is probably £ 13 ug/m3, and the practical LDL
« 22 pg/m3.
* The only calibration data available were from Collaborators B and H.
** The intercept on the calibration curve.
59
-------
CONCLUSIONS
The major conclusions that can be drawn from the results of this col-
laborative test are:
1. The N02» ambient-air sampling system developed by MRI is an ef-
fective system for use in collaborative testing of manual and instrumental
methods such as those tested on this program.
2. The "Tentative Method for the Determination of Nitrogen Dioxide
in the Atmosphere (Chemiluminescent Procedure)" is adequately written for
those knowledgeable in sampling and analysis techniques as presented
therein.
3. If the tentative Chemiluminescent Procedure as given in Appendix
A of this report is followed by people knowledgeable in the sampling and
analysis techniques given therein, then such persons will obtain results
with an average bias of -8 jig/m3 over the range 80 to 300 ug/m3. On the
average, the within laboratory standard deviation (
-------
RECOMMENDATIONS
Based upon the conclusions that have been drawn from the results of
this collaborative test, it is recommended that no further collaborative
testing and analysis be done on any of the four methods tested—sodium-
arsenite, TGS-ANSA, continuous colorimetric and chemiluminescent.
63
-------
APPENDIX A
TENTATIVE METHOD FOR THE CONTINUOUS MEASUREMENT OF
NITROGEN DIOXIDE IN THE ATMOSPHERE
(CHEMILUMINESCENT PROCEDURE)
65
-------
ENVIRONMENTAL PROTECTION AGENCY
METHODS STANDARDIZATION AND
PERFORMANCE EVALUATION BRANCH
QUALITY ASSURANCE AND ENVIRONMENTAL MONITORING LABORATORY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
August 1974
TENTATIVE METHOD FOR THE CONTINUOUS MEASUREMENT OF
NITROGEN DIOXIDE IN THE ATMOSPHERE (CHEMILUMINESCEMT PROCEDURE)3
A tentative method is one which has been carefully drafted from
available experimental information, reviewed editorially within
the Methods Standardization and Performance Evaluation Branch/
QAEKL, and has undergone extensive laboratory evaluation. The
method is still undor investigation and, therefore, is subject
to revision.
67
-------
1 . Principle and Acpliccbili ty
1.1 Atmospheric concentrations of nitrogen dioxide (i-!C>2) are
measursd indirectly by the chtrni luminescent reaction of nitric
(1 2
oxida (NO) with ozone (0,) at reduced or near atmospheric pressures/ '
** ($}
NO* is first thermally convertedv ' (reduced) to NO before it is
reacted with O^. A photomultiplier tube detector measures the intensity
of the light from the resulting chemi luminescent reaction. Since
ambient air usually contains both NOg and NO simultaneously, soma
rceans nust be provided for distinguishing between the NO concentration
already in the sample and the NO concentration resulting from the
reduction of NO* tc NO. Ambient air is drawn directly into the
detector assembly to dc.terr.ir': t^G NO ccnccr.tr*tion. The f!0x ccr-
centralion (NO + f^) is measured when the air is passed through the
thermal converter and tnen into the detector. The NOg concentration
is then determined electronically by subtracting the NO detector
response (or concentration) from the NO detector response. The
A
measurement is accomplished in approximately one minute.
1.2 The procedure is applicable to the continuous meas-
urement of NOp in ambient air.
2 . Raryie and Lowe*- Detectable limjt_
2.1 Analyzers with full scale ranges from U to 376
(0 to 0.2 ppm) to 0 to 18,800 vg/pi3 N02 (0 to 10 ppm) are available.
The fill I scale N00 range recrrsjiiended for normal ambient air
68
-------
measurements is the most sensitive range on the analyzer not to
exceed a sensitivity of 0 to 940 pg/m3 (0 to 0.5 ppm). Separate
range selectors should be made available for NO, NO^ and NOX since
NO concentrations sonetimes exceed 1 ppm.
^
3
2.2 The lower detectable limit is approximately 20 ug/m
(0.01 ppm) at the 0 tc 940 ug/m (0-0.5 ppm) range.
3. Interferences
3.1 The chemiluminescent detection of NO with 03 is not
subject to interference from any of the common air pollutants such as
0,, NO,, CO, NH, and S09. ' The elimination of possible hydrocarbon
*3 L j £
c arhievpH hy rnoans nf a red sharp-cut optic::!
3.2 Any compounds which will' be converted to NO in the
thermal converter will interfere with the measurement of the NOX
(NO + NO?) concentration. The principal compound of concern is
ammonia; however, it is not an interferent at converter temperatures
below 300° C. Unstable nitrogen compounds, such as peroxy-
acetylnitrate (PAN) and organic nitrites decompose thermally to
form NO. However, their ambient concentrations are usually so lew
that their interference can be disregarded.
4. Apparatus
4.1 Nitrogen Dioxide Analyzer. The analyzer should be of
the cf.fcmiluminescent type with a linear response over the desired
69
-------
range. The analyzer should meet or exceed the specifications
described in Appendix A. See Figure 1 for a general schematic of
a typical chemiluminescent analyzer.
5. Procedure
5.1 Ambient air is drawn into the analyzer with a sample
pump or with the vacuum pump used in the detector. Ozonized
oxygen or air at a constant flow is drawn into the detector
assembly to react with the sample. For exact operating procedures
refer to the analyzer instruction manual.
6. Calibration
6.1 principle. Tiie calibration teCimiquc is based upcr,
application of the rapid gas phase reaction between NO and 03 to
(2 3}
produce a stoichiometric quantity of NO. '
"1
NO + 03 -*• N02 + 02 k = 1.0 x 107 liter mole"1 sec
The quantitative nature of the reaction is used in a manner such
that, once the concentration of one component is specified, the
concentrations of the other two are determined. The stoichiometry
of the reaction holds only when NO is present in excess.
Atmospheres of known 0^ concentration are produced
from an ozone generator, the output of which has been calibrated by
ioJo-.etry (r.eutral buffered K! procedure).^ ' The NO content of an
NO in 1^2 cylinder (50-100 ppm) is then assayed by gas phase
70
-------
titrction with the 03. Analysis of the NO cylinder is necessary
because the nominal NO concentration of some manufacturers has been
found to be inaccurate. The amount of Og added in the titration is
equivalent to the amount of NO consumed (or N02 produced), thus the
NO content of the cylinder can be determined.
Once the cylinder is assayed it can be used throughout
its life to prepare known concentrations of NO. In addition to
generating standard NO concentrations for calibrating the NO response
of an analyzer, the assayed NO cylinder is also used to generate
kno^n N02 concentrations by a gas phase titratior, with 0^. To
calibrate the M^ response of a chemiluminescent analyzer, 0.. is
added in incregents tc excess NO. The incremental decrease in NO
concentration (observed by noting the NO detector response of the
analyzer) is equivalent to the N02 produced in the titration. Since
the NOp is produced continually in the titration, a dynamic calibration
results.
6.2 Apparatus. A schematic drawing of the gas phase titration
apparatus is snown in Figure 2.
6.2.1 Pressure regulator for NO cylinder. A regulator to fit
the NO cylinder having internal parts of stainless steel with a
Teflon or Kel-F seat and a delivery pressure of 30 psi. Note: This
regulator as well as needle valve and delivery lines should be
thoroughly flushed after connecting to the NO cylinder. Flushing
can be accomplished by adjusting the delivery pressure to near
maximum and venting the gas for scvtral minutes.
71
-------
6.2.2 Nsedle valves. Stainless steel parts are required in
controlling the NO flow. Component parts are not critical for control
of air flow.
6.2.3 Flowmeters
6.2.3.1 NO flow. A flowmeter capable of monitoring NO flows
o 3
between 0-100 cm /min. A 25 cm capacity soap bubble flowmeter is
required for measuring absolute NO flows.
6.2.3.2 Air flow. A flowrceter capable of monitoring flows
between 0 and 10 fc/min. A large soap bubble meter or wet test meter
is required for calibration of the flow meter and for making absolute
flow measurements in this range.
6.2.4 Capillary res>Li ~ii.tur~. GlaiS Of Sia'InlcSS Steel Cup~!11ui~y
of the proper I.D. and length to maintain a 1:10 to 1:15 ratio of
air flow through the generator to total air flow. A Teflon or stainless
steel needle valve can be used in place of a capillary restrictor
provided that a stable flow through the generator is maintained.
6.2.5 Calibrated Ozone Generator, Capable of producing 0-
over the range 0 to 1 ppm at a total air flow rate of 2 to 10 ?./min.
One such ozone source consists of a quartz tube into which ozone-free air
is introduced and then irradiated with a stable low-pressure mercury
r-j o\
lamp/ ' ' The lamp's level of irradiation is controlled by an adjustable
aluminum sleeve which fits around tha lamp. Ozone concentrations
are varied by adjustment of this sleeve. At a fixed level of air flow
and irradiation, ozone is produced at a constant rate.
72
-------
6.2.6 Reaction Chamber and Mixing Bulb. Kjeldahl connecting
bulbs with volumes of approximately 250 cm should be used.
6.2.7 Sample Manifold. A multiport all-glass manifold should
be used. All connections in the calibration system should be glass
or Teflon.
6.3 Reagents
6.3.1 Nitric Oxide. Cylinder containing approximately 50 to
100 ppm NO in N2 with less than 1 ppm NOo. In ordering cylinders
from commercial vendors, specify the use of 02 free N2. This will
minimize N02 formation.
6.3.2 Clean Air Source. Compressed (house) or cylinder air
containing no more than 0.005 ppm of NO, N02, and 03, or reactive
hydrocarbons. The air is cleaned by passing it through silica gel
for drying, treating with ozone to convert any NO to N02> and finally
by passing through activated charcoal (6-14 mesh), and molecular
sieve (6-16 mesh, type 4A), to remove any N02 or hydrocarbons.
Regardless of the source of air, compressed (house) or cylinder air,
it should be purified as described above.
6.4 Procedure. The gas phase titration apparatus shown in
Figure 2 is used to assay the NO cylinder and also to generate N02
for analyzer calibration. NO from the cylinder is diluted with a
constant flow of clean air (with the ozone source set at zero 03
concentration) to-produce an NO concentration at the sample manifold
of approximately 1 ppm. The ozone source is then adjusted to produce
73
-------
incremental increases in 03 concentration which result in equivalent
incremental decreases in NO concentration down to approximately
0.05 ppm NO. During this titration NO- is generated in amounts
equivalent to the NO consumed by the 03.
Upstream of the ozone generator, the air stream is split
such that the capillary restrictor allows only 1/10 to 1/15 of the total
air flow to go through the ozone generator, with the remainder bypassing
the generator. The air stream is split in order to permit low flows
through the ozone generator to create locally high concentrations of
D.J and NO in the reaction chamber. At these high concentrations and
low flows, the reaction chamber provides a residence time long enough
for quantitative reaction to occur. The N0-N02 mixture is combined
with tnp nynasspri air in the mixinn chambpr. The turbulent flow
created assures a homogeneous sample at the sample manifold.
6.4.1 Analysis of NO cylinder by Gas Phase Titration (G.P.T.).
Allow the chemiTumi nascent analyzer to sample clean air until a
stable response is obtained. After the response has stabilized, make
proper zero adjustment. Determine the NO concentration in the cylinder
as follows. Uith the NO flow off, set the clean air flow at approxi-
mately 2 to 10 fc/min. (the actual flow depends on the operating
characteristics of the ozone generator, Section 6.2.5). Measure and
record the absolute air flow, FQ. Adjust the NO flow to generate
approximately 1.0 ppm NO at the manifold. (Use the NO cylinder concentration
provided by the manufacturer.) Measure and record the absolute NO flow,
F.,n. Adjust the span control for MO response on a 0 to 1 ppm range so
74
-------
that the NO response reads exactly the NO concentration generated. After
the NO response has stabilized, record the reading and then add approxi-
mately 0.1 ppm 0., by adjustment of the calibrated ozone source. (The
ozone source should be stable before beginning the gas phase titration.)
Allow the NO response to stabilize and record the reading. (Readings
may be taken directly from the instrument readout or from a properly
spanned recorder.) Adjust the ozone source to obtain approximately
0.2 ppm 03 and allow the NO response to stabilize. Continue this
procedure until up to 0.8 to 1.0 ppm 03 has been added in a stepwise
fashion. A minimum of six ozone additions are required. Remcasure
FQ and FNQ to insure that the flows have not changed during the titration.
6.4.1.1 Calculation. Plot the 03 concentration added (y-axis)
versus the resultant NO concentration in ppm (x-axis). A linear
curve should be obtained. (See Figure 3, for example.) The 03 con-
centration at the y-axis intercept, C'Q, is the 03 equivalent to the
initial diluted NO concentration. The cylinder NO concentration is
calculated as follows:
, Fo * c'o
m"~^~
where C = cylinder NO concentration, ppm
NQ
= measured NO flow, cm /min.
C'Q = 03 equivalence coint, ppm
FO = total clean air flow, cm /min.
75
-------
6.4.2 N02 Impurity in the NO Cylinder. The concentration of
N02 in the NO cylinder must be determined because the error in
disregarding this impurity can be significant. Consider, for example,
a 100 ppm NO cylinder with only 1 ppm of N02> When used to generate
1 ppm NO for G.P.T., the NO., is diluted to 0.01 ppm. Although this
is a small amount of N02, the error it would introduce relative to
typical measured N02 concentrations, 0.03 to 0.05 ppm, would be
greater than 20%. The N02 impurity in the cylinder must be determined
by an independent method which has no appreciable interference from
NO. The triethanolamine-quaiacol-sulfite (T6S) manual method^ ' is
recommended. However, any other method which measures N02 directly
and has no interference from NO may be used.
To determine the amount of N02 impurity in the NO
cylinder, sample a diluted stream from the cylinder. The exact dilution
will depend on the N02 concentration in the cylinder and the method
chosen for analysis. Generally, dilutions of 1 to 100 or 1 to 50 will
be adequate. Convert the N02 analyzed to ppm. Then calculate the
N02 impurity using:
)A x FT
'NO
where C..Q = N07 concentration in cylinder, ppm
(NO ) = f^2 concentration analyzed, ppm
Total flow at manifold, cm /min.
from NO cylinder, cm /ruin.
NO
76
-------
6.4.3 N02 Converter Efficiency. The efficiency of the N02
converter should be constant and linear over the range of
operation. The NO* converter efficiency may be determined in the
following manner. While the chemiluminescent analyzer is sampling
clean air, adjust the zero controls for the NO and NO responses to
A
read zero. Dilute the assayed cylinder of NO with clean air to
produce approximately 1.0 ppm NO and adjust the span controls for the
NO and NOX responses to read exactly the NO concentration generated.
The exact NO concentration is calculated as follows:
(NO) =
FT
wnere (NU) - NO concentration generated, ppm
CMA = cylinder NO concentration, ppm
NU
3
FNO = N0 flow> cm 'mi"n
FT = total flow at manifold, cm /min
Note: If N02 impurity is present in the NO cylinder, adjust the
span control for NO response to include the N0« impurity concentra-
J\ t
tion assuming a converter efficiency of 1.
Observe and record the initial NO and NOX readings. Next,
add ozone to the NO stream to reduce the NO response to 0.9 ppm.
Observe and record the final NO and NOX readings. Repeat the above
procedure adding ozone such that the NO response Is decreased to 0.8,
77
-------
0.7, 0.6 and 0.5 ppm successively. Calculate the concentration of
N0? generated by G.P.T. for each stepv/ise addition of ozone, the
total NOp concentration generated, and the concentration of M02
converted to NO as follows:
N02 G.P.r., ppm = NOInitial, ppm " NOFinal, ppm
N02 Generated, ppm = N02 G.P.T., ppm + N02 Impurity, ppm
2 Converted, ppm ~ °x Final, ppm " Final, ppm
Plot the N02 concentration converted (y-axis) versus
the total N02 concentration generated (x-axis). The slope of this
plot is the converter efficiency. A straight line indicates a
constant converter efficiency over the 0 to 0.5 ppm range. If
the converter efficiency is less than 0.90 (90£) or if it is not
constant (i.e. nonlinear curve), the converter should be replaced
or rejuvinated.
6.4.3.1 Frequency of Determination. The converter efficiency
should be determined monthly when the instrument calibration is
checked (Section 6.4.4.2).
5.4.4 N02 Calibration (0 to 0.5 ppm range).
6.4.4.1 Procedure. While sampling zero air from the gas phase
titration apparatus, zero the NO, NOX and N02 detector responses.
Adjust the NO flow rate to generate approximately 1.0 ppm NO by
78
-------
flow dilution (Section 6.4.3) and adjust the span controls to read
the exact NO concentration generated. Any N02 impurity in the
NO cylinder should be included (with an appropriate correction for
dilution and converter efficiency) in the NO span adjustment.
Adjust the ozone source to add enough CL to decrease
the NO response to approximately 0.50 ppm. This results in the
generation of an equivalent amount of N02 which is used to span the
N02 response. Allow the response of the instrument to stabilize and
adjust the span control to give a direct readout of the generated
N02. including any N02 impurity if present. Continue to decrease
the 03 concentration so as to produce N02 concentrations of approxi-
mately 0.40, 0.20, 0.10 and 0.05 ppm. Using this procedure the
N02 generated can be determined by the following equation:
N02 Generated = NOInitial " NOFinal * N02 Impurity, ppm
where N02 Qeneratec| = Tota1 concentration of N02 generated, ppm
^Initial = *n1tia^ concentration of NO, ppm
NOp. , = Concentration of NO after 03 addition, ppm
Plot the total N02 concentration generated versus analyzer
N02 response. A straight line should be obtained.
6.4.4.2 Frequency of Calibration. A complete calibration
should be made once a month. Analyzers should be zeroed and
79
-------
spanned at 80 ±5% of full scale- daily, then adjusted for any drift
that is detected. If the drift is more than ±15%, the analyzer should
be recalibrated.
7. Calculations
7.1 Determine the N02 concentration directly from the N02
calibration curve (Section 6.4.4.1).
7.2 Nitrogen dioxide concentrations in ppm can be converted
to vg/m3 at 25° C and 760 mmHg as follows:
(N02), vg/m = (N02), ppm x 1880 x
8. Bibliography
1. A. Fontijn, A.J. Sabadell , and R.J. Ronco, "Homogeneous
Chemi luminescent Measurement of Nitric Oxide with Ozone," Anal . Chan. .
42_, 575 (1970).
2'. J.A. Hodgeson, R.E. Baumgardner, B.E. Martin, and
K.A. Rehme, "Stoichiometry in the Neutral lodometric Procedure for
Ozone by Gas-Phase Titration with Nitric Oxide," Anal. Chem.. 43_,
1123 (1971).
3. K.A. Rehtne, B.E. Martin, and J.A. Hodgeson, "Tentative
Method for the Calibration of NUric Oxide, Nitrogen Dioxide and
Ozone Analyzers by Gas Phase Titration," Environmental Protection
Agency, Research Triangle Park, N.C., EPA Publication
No. EPA-R2-73-246, March 1974.
80
-------
4. D.H. Stedman, E.E. Daby, F. Stuhl, and H. Niki, "Analysis
of Ozone and Nitric Oxide by a Cheroiluminescent Method in Laboratory
and Atmospheric Studies of Photochemical Smog," J. Air Poll. Control
Assoc., 22_, 260 (1972).
5. J.A. Hodgeson, K.A. Rehrne, B.E. Martin, and'R.K. Stevens,
"Measurements for Atmospheric Oxides of Nitrogen and Ammonia by
Chemiluminescence," Preprint, Prepared at 1972 APCA Meeting, Miami,
Florida, June 1972, Paper No. 72-12.
6. Environmental Protection Agency, "Part 50—National
Primary and Secondary Ambient Air Quality Standards," Federal Register.
Vol. 36., No. 228, pp. 22384-22397, November 25, 1971.
7. J.A. Hodgeson, R.K. Stevens, and B.E. Martin, "A Stable
Ozone Source Applicable as a Secondary Standard for Calibration of
Atmospheric Monitors," ISA Transactions. ]j_, 161 (1972).
8. NBS Technical Note, No. 585, pp. 11-25 (January 1972).
Available through: Superintendent of Documents, Government Printing
Office, Washington, O.C. 20402. Price 70£.
9. "An Evaluation of the TGS-ANSA Method for Measurement
of NOg," Fuerst, R.G., and J. H. Margeson. Copies of this document,
which includes a copy of the method write-up, can be obtained from:
Methods Standardization Branch, Environmental Protection Agency,
Research Triangle Park, North Carolina 27711.
81
-------
APPENDIX A
Performance Specifications for Atrcospheric-Chemiluminescent-
NO Analyzers
Parameter Um'ta
Rangeb ppm
Noise ppm
Lower Detectable Limit ppm
Zero Drift, 12-, 24-Hour ppm
Span Drift, 24-Hour ppm
Lag Time minutes
Rise Time, 9555 minutes
Fall Time, 95* minutes
Specification
multiple
0-0.005
0.01
±0.02
±0.02
0.5
1.0
1.0
aTo convert from ppm to yg/m at 25° C and 760 mm, multiply by 1880.
bNo performance test required. All other performance specifications
are tested on instrument operating in the 0-0.5 ppm range.
82
-------
APPENDIX B
Definitions of Performance Specifications
Range - Minimum and maximum concentrations which the system shall be
capable of measuring.
Noise - Spontaneous, short duration deviations in the analyzer cutout
about the mean output, which are not caused by input concentration
changes.
Lower Detectable Limit - The minimum pollutant concentration which
produces a signal of twice the noise level.
Zero Drift - The change in analyzer response to zero pollutant
concentration over 12- and 24-hour periods of continuous unadjusted
operation.
Span Drm - The change in analyzer response to an upscale pollutant
concentration over a 24-hour period of continuous unadjusted operation,
Lag Time - The time interval between a step change in input concen-
tration at the analyzer inlet and the first observable corresponding
change in the analyzer response that is equal to twice the noise in
the instrument output.
Rise Time - The time interval between initial response and 95% of
final response after a step increase in input concentration.
Fall Time - The time interval between initial response and 95% of
final response after a step decrease in input concentration.
83
-------
oo
r
TIMER
\
0 ^
•wSij'i
—_ r^jnl
°3 f
3ENERATOF
n^
2
OPTICAL FILTER
s
t
LHWIUtK -^
\
... ^
\
^ '
—
— •
—
V] =J=- SAMPLE
INLET
N02 - NO
CONVERTER
'SOLENOID
VALVES
DETECTOR
HOUSING TEMPER-
ATURE CONTROL
HIGH VOLT
ftGE POWER
SUPPLY
n
TTjl
PHOTO-MULTIPLIER
TUBE
/ / / / /T7/./ /
EXHAUST
PUMP
THEPMO ELECTRIC
COOLED HOUSING
HO
TREND
N0?
TREHD
»
TREND
NO NO.
GROUND
Figure 1. Automated NO, NOg, NO chemiluminescent analyzer.
-------
NEEDLE
VALVE
00
Figure 2. Flow scheme for calibration of NO, N02 and NOX monitor by gas phase titration.
-------
00
CL
a.
c
o
01
u
o
co
o
Equivalence Point, C1
NO Concentration, ppm
Figure 3. Gas phase titration of NO with 03,
-------
APPENDIX B
DATA ON THE PERMEATION TUBES USED AS THE
SOURCE OF THE SPIKED LEVELS OF NO.
87
-------
There were four branches to the NO permeation tube assembly. Each
branch contained a set of permeation tubes as follows:
Branch NO *
Branch Number (u,g/min) (ug/min)
1
1
1
1
1 5.325
2
2
2
2
2
2 6.382
3
3
3 3.200
4
4
4 2.980
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 col-
laborative test.
The combinations of branches used for the four runs of the cherai-
lurainescent collaborative test are:
Level Date Branches used
1 September 23 to 24 4
2 September 24 to 25 1, 3, and 4
3 September 25 to 26 1, 2, 3, and 4
4 September 26 to 27 3 and 4
Permeation
Number
35-8
35-16
34-7
28-10
34-3
34-13
34-6
34-1
34-10
35-13
29-4
29-2
34-12
tube
Rate of N02
(iig/min)
1.434
1.597
1.134
1.160
1.195
1.275
1.548
1.226
1.138
1.990
1.210
1.210
1.770
* The sum of the NO. generated by each permeation tube in the branch.
88
-------
APPENDIX C
CALIBRATION OF THE VENTURI AND DRY-GAS METER
89
-------
The venturi and dry-gas meter were calibrated using a 1.0-ft /rev.,
wet-test meter, as shown in Figure 9 of the text. 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 NCL bleed-in as it does in normal
operation (see Figure 3 of the text).
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 flow rate of the wet-test meter (to stp) is:
P 294
Flow = Flow (meter reading) x —— x —-
where T = temperature of wet-test meter + 273, and
P = P + pressure of test meter manometer.
atm
The venturi flow rate is dependent on both temperature and
pressure. Therefore Flow is corrected to venturi conditions.
stp
760 To
Flow = Flow x x „-,
venturi stp P_ 294
where !„ = temperature of gas stream + 273, and
2 atm (gas stream)
90
-------
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 flow rate of the dry-gas meter (F ) is
_. 760
F = Flow x
ra stp P
where P, = P _ + P, . * .
3 atm (gas stream)
The correction factor f to convert f , measured dry-gas meter
flow rate, to true flow rate is then
f =|m .
*m
The venturi and dry-gas meter were calibrated at three flow rates;
50, 55, and 60 liters/rain. Normal system flow rates are 55 to 60 liters/
min. The calibration factor for the dry-gas meter is constant at the cali-
bration flows (+ 0.27.). The average value of flow from seven determina-
tions is used in calculating true flow rates of the system. The plot of
venturi &P versus flow rate follows a straight line over the range used
in calibration. From the slope and intercept of the line flow rates were
calculated.
91
-------
APPENDIX D
WRITTEN COMMUNICATIONS WITH
POTENTIAL COLLABORATORS
93
-------
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 NO- Testing Program, to be
sponsored by EPA. 1 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 chc
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, discs 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)
94
-------
COLLABORATOR FORM
1. Methods to Teat (Clicck the ones in which you want to participate
as a collaborator) :
li Sodium Arsenite |J TGS-ANSA Procedire
| _ I Continuous Saltzman | | Chemi luminescent
2. Equipment Available for Test:
Could you Curnish six trains for:
Sodium Arson ite : I I yes [_] no
TGS-ANSA: Q] yes Q no
Hnvc you n Colorimctric (Continuous Saltzman) Ambient N02 Monitor?
yes Q] no Make _ Model _
Have you a Chemilumlnesc'ent Ambient N02 Monitor that you would use?
Q] yes Q] no Make _ Model _
3. Teat Period (Each Method) :
Period acceptable (calendar days) :
[] 6 days Q 10 days
4. Methods You Have Uspd :
Q Sodium Arsenitc, (_J TGS-ANSA,
| I Chnmiluriinoscent , | | Others: _
13 days
Continuous Saltzman,
5. Remarks :
6. Company:
Address:
Person to Contact:
Telephone Number;
95
-------
5 August 1974
MIDWEST RESEARCH INSTITUTE
425 Volker Boulevnr
Kansas City, Missouri 64 r
Telephone (816) 561-0?C
Name
Pollution Control Agency
City of Toledo
26 Main Street
Toledo, Ohio 43605
Dear Name:
You have indicated your interest in participating in an EPA-Sponsored N02 col-
laborative test using the Chemiluminescent method. If you desire to be con-
sidered as a candidate to participate in this test, would you let me know by
19 August, providing me also with the following information: (1) Whether or
not you could provide instrumentation at the test site in Kansas City for your
sampling; and (2) If so, the type instrument along with its flow intake rate,
and the bench area it will occupy.
We would appreciate your reviewing the enclosed preliminary draft, "Tentative
Method for the Determination of Nitrogen Dioxide in the Atmosphere (Chemilum-
incscent Procedure)." After your review, would you use this method to obtain
a few sample measurements of your ambient atmosphere and then send me a copy
of your results by 26 August. These results will assist us in the selection
of ten collaborators.
The 1-week nitrogen dioxide test using the enclosed method is tentatively
scheduled to take place in Kansas City, Missouri, with sampling starting on
Monday, 23 September. The test will be indoors using a sampling system
developed by MRI for this collaborative program. This system, which was
used quite successfully for three other methods, is shown in diagram form
in the attached figure.
Each collaborator will be reimbursed for travel, subsistence and lodging for
the employee it sends to Kansas City to perform this field sampling, as well
as for local travel and miscellaneous expenses such as the cost of shipping
the instrumentation to be used on site for the sampling. The cost of all ma-
terials to be used in the operation of the sampling instruments such as cali-
bration systems, chemicals, etc., are not included and are to be borne by the
collaborative organization.
We will be in touch with you as soon as the ten collaborators are selected for
this test. If you have any questions, please contact me.
Very truly yours,
Paul C. Constant, Jr., Head
Environmental Measurements Section
PCC:cdn
Enclosures
96
-------
9 September 1974
This confirms our selection of your organization as one of the 10 collabor-
ators for the EPA-Sponsored Nitrogen Dioxide Collaborative Test Using The
Chemiluminescent Method, and presents information about this test.
A copy of the "Tentative Method for the Determination of Nitrogen Dioxide
In the Atmosphere (Chemiluminescent Procedure)," dated August 1974, is
enclosed for your study and retention. Please destroy any previous write-
up we may have given you on the Chemiluminescent Procedure.
The test schedule is given in Table 1. The starting date is Monday, 23
September 1974. We will meet in the lobby of the Ramada Inn (See Figure
1, upper right-hand portion) which is located at 87th Street and Highway
1-435. From there we will go to the field site which is several miles
from the motel. At the site there will be an orientation program of the
facilities prior to your preparing your instruments for sampling.
The schedule in Table I is based on the assumptions that all equipment of
the collaborators will be on-site early Monday morning, and that all goej»
well during the week.
The sampling by each collaborator must be performed according to the
attached write up.
The sampling system that will be used in this collaborative test is shown
in diagram form in Figure 2, an enclosure to this letter. Each collabora-
tor will attach his instrument to ports of the spiked and unspiked mani-
folds according to a specific experimental design pattern.
There will be two runs per each of four 24-hr periods. The duration of one
run will be approximately 14 hr; the other will be 7 hr. For the 14-hr run,
all collaborators will sample from the spiked manifold. For the 7-hr runs,
each collaborator will sample from each manifold—spiked and unspiked—for
approximately 3-1/2 hr.
97
-------
MIDWEST RESEARCH INSTITUTE
Page 2
9 September 1974
The hours 0800-1630 on Monday, 23 September are for collaborator prepara-
tion. From then on testing will be according to the following schedule:
Time
1630-1800
1800-0800
0800-0930
0930-1250
1250-1310
1310-1630
Activity
Calibrate
Sample
Calibrate
Sample
Switch Lines
Sample
Spiked Line Unspiked Line
1 thru 10
4, 6, 7, 8, 10 1, 2, 3, 5, 9
1, 2, 3, 5, 9 4, 6, 7, 8, 10
Enclosed is a Teflon adaptor. One of these adaptors constitutes a port of
the sampling manifolds. It is the type of port to which you will attach
your instrument. The larger-diameter end is the one to which an instrument
is to be attached. This adaptor is being sent to you, as an enclosure, so
that you will know what size tubing will be required for the connection of
your instrument to the sampling manifolds. You will need at least 30 feet
of this tubing.
The switching of your instrument between the spiked and unspiked lines will
be done manually.
Each collaborator will be reimbursed for travel, subsistence, and lodging
for the employee it send to Kansas City to perform the field work as a
collaborator, as well as miscellaneous expenses such as costs for shipping
the field equipment you will use on site for the sampling. Please keep
receipts such as airline tickets and equipment shipment invoices. Mr. Fred
Damon, MRI Administrative Officer, will be contacting you to make contrac-
tural arrangements. If you have any questions you may contact him at (81,6)
561-0202.
Reservations will be made by MRI for each collaborator at the Ramada Inn.
These lodging charges will be billed directly to MRI, therefore, you will
need not consider this expensed Also, if you are not driving to Kansas
City, but rather arrive by plane, MRI will provide local transportation.
Upon your arrival at the Kansas City International Airport call Econo-Car-
Rent-A-Car Company at 464-5656 and you will be provided a car. As is the
case of the motel, these charges will be billed directly to MRI.
If you are not driving and bringing your equipment with you, I suggest that
you ship it a couple of weeks before the 23rd of September so that it will
be at MRI before 20 September. For your convenience, you could send it to
yourself, % Mr. Paul C. Constant, Jr., Midwest Research Institute, 425
Volker Boulevard, Kansas City, Missouri 64110, COD or prepaid.
98
-------
MIDWEST RESEARCH INSTITUTE
Page 3 9 September 1974
To help us with preparations at this end, I would like the following infor-
mation if you have not already given it to us:
1. The name of the person(s) from your organization who will be
coming to Kansas City.
2. The mode of transportation the person will use to come to
Kansas City (in case of airline, the airline and flight number), and the
time of his arrival.
3. How your equipment will be sent to Kansas City and when it
is expected to arrive.
Very truly yours,
Paul C. Constant, Jr., Head
Environmental Measurements Section
PCCrcdn
Enclosures:
1. "Tentative Method for the Determination of Nitrogen Dioxide
in the Atmosphere."
2. Table I — Test Schedule
3. Figure 1 — Map: Ramade Inn to Field Station
4. Map — Kansas City International Airport to Ramada Inn
5. Map — Deramus Field Station
6. Figure 2 — Nitrogen Dioxide Ambient-Air Sampling System Concept
7. Teflon Adaptor
99
-------
APPENDIX E.
INSTRUCTIONS FOR COLLABORATORS NO, COLLABORATIVE
TEST; CHEMILUMINESCENT PROCEDURE
101
-------
INSTRUCTIONS FOR COLLABORATORS
N02 COLLABORATIVE TEST: CHEMILUMINESCENT PROCEDURE
General Information
1. Calibration, sampling, analysis, etc., should be done explicitly
as stated in the August 1974 write-up furnished you on the "Tentative Method
for the Determination of Nitrogen Dixoide in the Atmosphere," (Chemilumines-
cent Procedure).
2. Each collaborator will have an area for his analyzer on one or
the other table. Use of connecting tubing to the opposite table will enable
alternate sampling of the spiked and unspiked manifolds by each collaborator.
3. For each run, each collaborator will connect his analyzer to
either the unspiked manifold or the spiked manifold, according to the instruc-
tions given before the run. Heavier tubing, for those collaborators whose
instruments require higher flow rates is available in several areas of both
the spiked and unspiked side. Please ask the MRI person in charge concerning
use of the heavier tubing.
4. Collaborators are urged to make ample explanatory notes on their
analyzer charts to coordinate information and aid in data reduction. All per-
tinent data should appear on the charts.
5. Each collaborator should work independently of other collabora-
tors.
6. On the spiked table, a separate power circuit (120 V 60-cycle,
4-outlet strip) is to be used by each collaborator for his analyzer. On the
unspiked side one strip will be shared by two collaborators where necessary.
These strips are under the table tops near the periphery of the tops.
Test Instructions
1. Preparation by collaborators will precede sampling runs.
2. Analyzer calibration will be included in preparation time, not
to be done during the run.
3. Upon notification of "start testing" from the MRI person on
site, collaborators will mark the appropriate place on their analyzer charts.
It is expected that the analyzer will be operating and the connecting tubing
in proper hook-up arrangement prior to the start signal.
102
-------
4. Note all pertinent data on the analyzer charts as the run pro-
gresses .
5. Upon notification by the MRI person in charge, indicate the end
of the run by marking on the analyzer chart.
General Schedule
There will be two runs per each of four 24-hr periods. The dura-
tion of one run will be approximately 14 hr; the other will be 7 hr. For the
14-hr runs, all collaborators will sample from the spiked manifold. For the
7-hr runs, each collaborator will sample from each manifold--spiked and un-
spiked—for approximately 3^ hr.
The hours 0800-1630 on Monday 23 September are for collaborator prep-
aration. From then on testing will be according to the following schedule:
Time Activity Spiked Line Unspiked Line
1630-1800 Calibrate
1800-0800 Sample 1 thru 10
0800-0930 Calibrate
0930-1250 Sample 4, 6, 7, 8, 10 1, 2, 3, 5, 9
1250-1310 Switch Lines
1310-1630 Sample 1, 2, 3, 5, 9 4, 6, 7, 8, 10
The numbers given in the spiked and unspiked columns are the collaborators'
I.D. numbers. (See attached table for your number.)
103
-------
COLLABORATOR ANALYZER LOCATION AREAS
AND PRINCIPLE PORT ASSIGNMENTS
Collaborator
Spiked Table
I.D. No.*
Name
Area
Robert Richardson
Jon Zimmer
William Findlay
Roger Ellard
Jim Harris
John Higuchi
Melvin Ken Muir
Carol Ellis
Vinson L. Thompson
Midwest Research Institute
Ports
1-6
7-12
13-18
26-31
38-43
Unspiked Table
Area Ports
32-37
26-31
7-12
1-6
38-43
NOTE: Assignments of secondary ports (on the opposite table from the analyzer
location) will be made prior to the start of each test.
* I.D.'s removed from this report to keep collaborators anonymous.
104
-------
SPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS*
105
-------
UNSPIKED
SAMPLING MANIFOLD
COLLABORATOR
AREAS*
106
-------
87»li St.
RAMADA INN
I- 435
Red Bridge Rd.
112th St.
•
N
i.
TRUMAN
CORNERS
FIELD STATION
Figure 1 -- !ln.p: Ilar.iacla Inn to Field Station
107
-------
Kansas City
International
Airport
(KCI)
N
KANSAS CITY, MISSOURI
87th St.
I - 435
Driving distance
from KCI to
Ramao'a Inn:
Approximately
35 mi.
RAMADA INN
Grandview, Missouri i
I
I
_J
I
[_
Map -- Kansas City International Airport to Ramada Inn
108
-------
Robinson Pike Rd.
To Grondview >•
COLLABORATIVE
TEST SITE-BLDG.3
DERAMUS FIELD STATION
MIDWEST RESEARCH INSTITUTE
GRANDVIEW, MISSOURI
109
-------
APPENDIX F.
N02, SAMPLING SYSTEM DATA: TEST LOG SHEETS
WITH FIELD OPERATIONAL DATA
111
-------
LOG SHEET
Side 1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
DATE * 3 Se/» f
TIME
fio
tf*e*»at
OlOOoV
of co
cfcco
7700
OUTDOOR TEMP.
ff
59"
5?
r?
62,
67
cc
WET BULB
f/
TO
5'
HI
/I
57
53
5-5
Tl
53
si
DRY BULB
16
57
5f
60
REL. HUMIDITY
ft
70
70
11.
10
11
15-
47
BAR. PRESSURE
mm
7*3
a?.*
W-IT
THXlt-i
7YJ
7V/
WIND SPEED
/. 9
/.d
0
WIND DIRECTION
S
W/
(AJ
HJ
NJ POWER LIGHTS
u-
LEAKS PERMEATION
ASSEMBLY
BATH TEMP.
ts.l
J-fl
**"/
K.I
T-6-1
25,
ROTAMETERS -"
te
fg
n
•ti
12
9?
N2 PRESSURE
/MO
110
Iflo
RECORDER 2 - PAPER
IS
OPERATION
tX
-------
rt
LOG SHEET
Side 2
10
n
12
13
14
15
16
17
18
19
20
21
22
23
24
25
TIME
H'1
2000
00°
\ooo
/Sao
NO2 ANALYZER - PAPER
OPERATION
VAC PRESSURE
14"
ar
O2 PRESSURE
10
*<»
a*
00
10
1.0
10
Zo
UNSPIKED - NO
4.5
NO2
2. •
SPIKED - NO
. r
NO2
7.,?
7.0
VENTURI MANOMETER - HI
S)L
stf
5-32
rw
LO
' VENTURI AP
US',
*
-------
LOG SHEET
Side 1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 25
DATE
15
TIME
/goo
rfoo A00o
»/<*>
0/00
OtOO >}00
OUTDOOR TEMP.
57
57
ft
f5
a
7?
77
7*
WET BULB
53
5}
Ti
53
r,/
DRY BULB
57
57
51
•7?
77
REL. HUMIDITY
77
7?
gx
11
27
BAR. PRESSURE
Iff
7V/
7V/
fit
Til
74!
7HI
7VO,J-7»-o
WIND SPEED
3.0
/fr
i. 2
0.9
I-2-
o
/.z.
• X
WIND DIRECTION
4
U/
POWER LIGHTS
tx
y
v/
y
LEAKS PERMEATION
ASSEMBLY
J
BATH TEMP.
ROTAMETERS - \
tf.t
Xf.o
"6<°
25-0
75:
/oo
too
(00
f DO
91
/oo
Tf
100
IVDD
!*>-
|UO
00
(QU
n
97
97
97
9?
111
n
1*
16
96
77
9-f
N2 PRESSURE
575-
910
590
160
MoO
3 to
00
RECORDER I - PAPER
OPERATION
RECORDER 2 - PAPER
OPERATION
J
y
I/
-------
LOG SHEET
Side 2
10
11
12
13
M
15
16
17
18
19
20
21
22
23
24
25
TIME
/l££
o«ico
9600
o'jo
/too
Qoo
I'ioo
NO2 ANALYZER - PAPER
y
J
J
y
y
OPERATION
VAC PRESSURE
if
ir
25"
IS
O2 PRESSURE
ID
1x0
Ift
10
HO
UNSPIKED - NO
X. •
*.»"
NO2
SPIKED - NO
1.0
NO2
*».*•
*0.0
lo
2/0
Z-f
xl
22-
xt.1
VENTURI MANOMETER - HI
sisrts
ttf
rsf
537
flo
LO
a £3
H.I-
261
VENTURI
27Z
272
277
^77
^77
276
266
372
Z69
INDOOR TEMP.
az
i-a.
a*
11
it
n
XI
S1.S
GAS FLOW TEMP.
a/.f
22
A*.
-U
1.1
VI
U
12.0
33.
43
XS
GAS FLOW MANOMETER - HI
vrf
H67
tit 6
LO
34*
ur
V-r
GAS FLOW PRESSURE
in
1-20
ISO
//v
nl
Hi
EXHAUST - U MANIFOLD
IS
I/
y
y
J
S MANIFOLD
CHECK SAMPLE TRAINS
7
J
J
y
y
INITIALS
57Sff.fSt.lf *** f?
S 1-
ff./
-------
LOG SHEET
Sidel
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
DATE
r J =
7'*
<~
Stf
Tf
TIME
1*00
0/00
0200
O^Oo
foo
llfO
/fcgi
tjoa
OUTDOOR TEMP.
15
73
£3
51
C.C.
70
7V
/T*
/7s
77
WET BULB
rf
y?
5-f
51
66
DRY BULB
7*
7*
7S
SI/
7S>
79
77
REL. HUMIDITY
37
37
40
SI
c/7
81
If
41
6%
62-
5-1.
BAR. PRESSURE
737
717
736
736,
1%
73?
7J2.
73^
WIND SPEED
t.o
WIND DIRECTION
vf
A/'
-• POWER LIGHTS
y
y
LEAKS PERMEATION
ASSEMBLY
BATH TEMP.
IS.I
as./
IS-'
;?£'
If.f
ROTAMETERS - 1
17
97
ft
IV
OIL
7%
9*
ft
7V-
7*3
93
17
n
11
11
11
Sffc
If
N2 PRESSURE
IK
1*
1*
iK
faOD
ffao
RECORDER 1 - PAPER
OPERATION
RECORDER 2 - PAPER
V
i/
OPERATION
J
-------
LOG SHEET
Side 2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
TIME
woo a**0 a3»o
ojeoouo
fl^oo
d(/oo
ejem
aao
Ifo*
NO2 ANALYZER - PAPER
y
y
OPERATION
y
y
VAC PRESSURE
I*
15
15
as
02 PRESSURE
la
10
10
UNSPIKED - NO
NO2
SPIKED - NO
?'*>
?>*•*
NO2
31
s/.r
32
93
31
3/
ii
M
1\
32,
3/
VENTURI MANOMETER - HI
f3«fj»
5*7
59-1
5-70
927
ft?
LO
^(,7 ifcV
3*9 261
VENTURI AP
261
ZW36J
366
361
262.
262
25?
59
257
is*
INDOOR TEMP.
13
23
13
il-
11-
11-
25
53
GAS FLOW TEMP.
13
at
11
a-l-
J7.2
GAS FLOW MANOMETER - HI
VI
Hit-
LO
X?
GAS FLOW PRESSURE
//*
//S
//*
V/7
//y
irt
//V
EXHAUST - U MANIFOLD
y
y
S MANIFOLD
y
CHECK SAMPLE TRAINS
577
INITIALS
ft-
58.1-
St.!
51 1
51.
, 5
*«.* ZTr ?O
£00
it**
-------
LOG SHEET
Side 1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24 25
DATE
2 .1
7Y
TIME
9to
!**•
0\tP
0J.4D
oloo
0500
0700
f/00
IJoo
if"
OUTDOOR TEMP.
Tf
LI
00
Jf
58
(12
4.
7V
WET BULB
W
$1
51
5?
It
60
62
DRY BULB
I,?-
Co
51
70
76
REL. HUMIDITY
ft
7'
61
81
61
BAR. PRESSURE
7*9
7*?
12-8
72-g
111
72 1
729
7X1
72 «?
727
WIND SPEED
Z.I
4.0
/.i
1.0
/.r
WIND DIRECTION
5
POWER LIGHTS
LEAKS PERMEATION
ASSEMBLY
y
BATH TEMP.
if.)
3S.I
ICI
ROTAMETERS - 1
/Ol
foo
too
too
IffO
flflT
too
190
11
r?
N2 PRESSURE
ll.it
J575
IHoO
RECORDER 1 - PAPER
OPERATION
RECORDER 2 - PAPER
J
OPERATION
J
y
-------
LOG SHEET
Side 2
10
11
12
13
M
15
16
17
18
19
20
21
22
23
24
25
TIME
fto
13*0.
rt*MtOoiOb
02.00
if oo
6700
llofl
NO2 ANALYZER - PAPER
J
OPERATION
J
y
VAC PRESSURE
*»"
25
if
if
O2 PRESSURE
0.0
1J
2o
10
UNSPIKED - NO
NO2
h
y
SPIKED - NO
NO2
/I
/3
13-
10-
IV
/i
tf.5
VENTURI MANOMETER - HI
f*9 &
ttt
$27 $27
LO
wl
2U
VENTURI AP
259
35?
S6I
261
257
357
355
263
INDOOR TEMP.
a. a.
>•*
22 .r
2-56
U..6
22
GAS FLOW TEMP.
^^-
a-J-
GAS FLOW MANOMETER - HI
it/
1ft
<*?/
191
y-?/
LO
17*
V*
37
"570
GAS FLOW PRESSURE
III
ii 6
110
EXHAUST - U MANIFOLD
S MANIFOLD
y
CHECK SAMPLE TRAINS
S
5*?
ITo
INITIALS
-
9*
&•'
59. /
57.
-------
APPENDIX G
FIELD DATA
121
-------
The first eight columns of Tables G-l through G-4 list various
readings used in calculating flow rates and spike levels, which are
given in columns 9 through 13. The last six columns list various ambient-
air conditions at the test site. The venturi and meter flow rates
(columns 9 and 10) 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 flow rates are calculated at 21°C and 760 mm Hg. The readings of
the two devices are then averaged (column 11) and the sampling ports
(column 12). Some of the methods being evaluated with this system are
not corrected for temperature and pressure. However, if the spike
levels are not calculated at the temperature and pressure existing at
the manifold ports, a significant degree of uncertainty enters into
any subsequent use of the spike level. The spike level (column 13)
is determined from the permeation rates of the permeation tubes
used in each test.
122
-------
TableG-l- LEVEL 1 TEST DATA
N02 Samp]
Flow
Bar Flow Temp -
Date Preia Preaa. Meter
Time flna ml (mm ml CC1
9-23-74
1800 744
1900 743
2000 743
2100 743
2200 743
2300 744
2400 744
9-24-74
0100 744
0200 744
0300 744
0400 744
0500 744
0600 744
0700 744
0800 743
0900 742
1000 742
1100 740
1200 742
1300 742
1400 742
1500 742
1600 742
1700 741
22.0
22.0
22 0
22.0
21 5
21 5
21.5
21 5
21.5
21.5
21.0
22.0
22 0
22.0
21.5
22.0
22 0
22.0
22 0
22 0
22.0
22.5
22.0
22 0
Lug Syoten
Plow
Rate
Meter
ft/mini
SB.l
58 4
58 4
58.6
59.0
.
60.7
59.6
59 3
59.0
59.2
59 2
59 0
58 7
58 4
58 4
57 9
58 4
58 6
58 4
58.8
59 0
58.9
59.0
Data
Venturl
Preaaure
Beading
(an H20)
265
265
265
269
269
269
272
272
274
272
274
274
274
270
267
265
262
269
270
267
271
271
271
271
N2
Cerrler
Flowrate
fee/mini
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
200
(
Plm
Venturl
Permeation to 21*
Tuba Temp + 760 m HI
25. 59 4
25 59 3
25. 59 3
25 59 8
25. 59 9
25. 59.9
25. 60 3
25. 60.3
25. 60 6
25. 60 3
25 60.7
25 6O.5
25 60 5
25. 60 1
25. 59 6
25. 59.2
25 58.8
25. 59.5
25. 59.8
25. 59 4
25. 59 9
25 59 8
25.1 59 9
25 1 59 8
Meter Averaae
to 21* to 21*
•f 760 OB HI + 760 mm Kg
U/nlnl (l/mtnl
58 7 59 0
58 9 59 1
58.9 59 1
59.1 59 5
59.! 59 7
59 9
61.3 60 8
60. 60 3
59 60.2
59. 60 0
59 60 2
59. 60 1
59 60.0
59. 59.7
58. 59 3
58 59.0
58 58.6
58. 59 1
59 59 4
58. 59 2
59.3 59.6
59.5 59 6
59.4 59 6
59.4 59 6
Plowrate Snlke HO
Level Back-
Aablenel/ AnblentV ground
(1/mlnl (ii«/B3l (an/in1)
60.5 49 o 0
60.7 4, o o
60.7 49 o o
61 0 48 7 0
61.2 4gt 0
61.3 4B , o
62.2 47 7 0
*>•' 48.2 0
61.6 48 2 0
61 4 48 0
61.5 4g D
61.6 48 0
61. 4g a
61- 48 0
'0 48 0
60 49. 0
60. 49 0
60 48 0
61 4g 0
60 48 0
>1 48 0
61 48 4 o
" 48 5 5
61 48 4 10
KOJ
Back-
ground
(UK/n3)
0
0
30
50
50
10
10
0
10
0
10
0
10
10
20
10
0
0
0
0
0
S
0
0
Amble
Out-
door
Temp.
12i_
64
60
57
55
55
55
55
55
55
54
53
54
55
55
57
60
62
62
67
66
63
58
60
59
at Condltlona
Wind -
gpeed
fta/aeel
5
0
0
3
4
3
0
5
4
0
0
0
0
3
2
4
3
11
10
6
12
9
7
-
Wind Relative
Dlrec- Rumldltv
tioo (11
S 43
S 58
S 66
S 70
g 70
g 70
S 76
S 65
S 70
g 81
g 75
S 65
S 76
S 76
S 77
S 73
W 65
v 50
W 56
W 48
u 64
S 72
S 68
S 67
•/ Teap«r«Cur« end preiiure mt imp Ling porti.
-------
Table G-2. LEVEL 2 TEST DATA
Calculated Flovratei and Splka L«v«U
mi Sanollni Syatem Dita
Flo*
Bar. Flov Temp.-
Date Frees Prean. Meter
Time (m En) (on Ha> CO
9-24-74
1800 741
1900 741
2000 741
2100 741
2200 741
2300 741
2400 741
9-25-74
0100 741
0200 741
0300 741
0400 741
0500 741
0600 741
0700 741
0800 740
0>OO 740
1000 740
11OO 740
1200 740
1300 740
1400 719
1500 738
1600 711
1700 717
21.5
22.0
22.0
21.5
21.5
21.0
21 0
21.0
21.5
21.0
21.0
21.0
21.0
21.0
21.0
22.0
22.5
23.0
23 0
23.0
23 0
23 0
23.0
23.0
Flow- Venturi
Bate Preaaure
Meter Reading
(Jt/mln) (mn mo)
59 1 272
59.1 272
59 0 272
59 0 272
59 5 272
59.6 272
59.6 272
59.5 272
59.6 277
59.6 277
59.7 277
59.7 277
59.8 275
59.6 275
57.7 269
57. S 266
58.2 265
58.1 262
59.1 272
58.6 269
58.4 267
58.4 264
58 5 264
58.1 264
"2
Carrier
Flowrnta
(cc/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
Permeation
Tube Temp.
i^2
25 1
25 1
25 1
25 1
25.0
25.0
25 0
25.0
25.0
25.0
25.0
25.0
25.0
25.0
25 0
25
25.
25
25.
25
25
25
25.
25
Fin
Venturi
to 11'
+ 760 am Bg
61 4
59.5 59
59.4 59
59 4 59
7 61 5
7 61.4
7 61 4
59.9 60.0 61 6
60.0 60
60.0 60
L 61 6
1 61 6
59.9 60 0 61 6
60.0 to. 3 o2.o
60 0 60 4 61.9
60.1 604 62.0
60.1 60 4 62.0
Spike
Level
Ambient^/
(u»/m3)
186
185
186
186
185
185
185
185
184
184
184
184
60 2 60 4 61 9 184
60 0 60.
58.0 58
57.8 58
58 5 58
58. 58
59 59
58. 59
58. 58
58. 58
58 58
58.2 58
61 8
60 5
60 3
60.6
60.4
61 6
61 1
60 9
60 7
60.8 .
60 6
184
188
189
188
189
185
186
187
188
187
188
»0 M>2
Back- Back-
ground ground
i&Jsb. iEB/pJl
.
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 0
0 10
0 0
0 0
0 0
10
20
15
15
5
0 5
5 0
-
Out-
door Kind
Temp Speed
CO (m/aec)
57 7
57 5
55 6
55 0
55 6
56 3
55 4
55 0
54 3
53 3
S3 5
52 0
52 0
52 "0
58 4
57 1
63 1
67 1
72 0
75 1
77 0
78 0
75 4
75 0
tJiod Relative
Dlrec- Humidity
tiop
-------
Table G-3. LEVEL 3 TEST DATA
10
Ul
HD2 San
Flow
Sar. Plov Tenp.-
Pata Preaa. Freee Mater
Time (mm Bg.1 (na BA) C*CI
9-25-74
1800 737
1900 737
2000 737
2100 736
2200 736
2300 116
2400 736
9-26-74
0100 716
0200 736
0300 736
0400 736
0500 736
0600 716
0700 71*
0800 734
0900 734
1000 714
1100 730
1200 731
1300 732
1400 731
1500 710
1600 730
1700 710
23 0
23.0
22 5
22 0
22 0
22.0
22.0
22.0
22.0
22.0
22.0
22 0
22.0
22.0
21.
21.
21.
22.
22.
22 0
22.0
22.5
23.0
23.0
Dllna Svitc
Plow
Kate
Meter
U/rnlnl
58 1
58.1
37 7
37 9
58 2
3G.1
58 4
58.8
58.8
58.7
59.2
59 1
38.5
58 6
58.1
58.1
58.1
57.9
58.0
58.1
58 2
58.0
58.1
58.1
n Date
Venturl
Preaauro
Reading
(mn B0>
264
264
258
258
264
264
264
264
271
261
269
269
263
265
26J
262
262
259
258
239
257
262
258
258
»1
Carrier
Flowrate
(cc/mlnl
800
800
800
BOO
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
800
Calculated Flovratee and Spike
Levell
Venturl Mater Avaraee Flowrate
Formation to 21* to 21* to 21*
Tube Temp + 760 mm % + 760 nm Hg -I- 760 on Kg
t'Cl (I/mini H/mlnl (i/mtnl
23.1 58 58.2 38 3
25 1 58. 58 2 58.3
25.1 57 57. 57.8
25.1 57 57. 57 9
25.1 58 38 58.4
23. 58 58. 58 5
25 58 58. 58 5
25. 58 58. 58.7
25. 59. 58. 59.1
25. 59. 58. 59 0
25. 59 58. 58.9
25. 59. 59 59.2
25. 58. 58. 58.6
25. 58. 58. 58 7
25. 58 57. 58 2
25. 58. 57. 58 1
25. 58. 57. 58.1
25. 57. 57. 57.5
25. 57 57. 57.7
25. 57 57 57 7
25 37. 37. 57.6
25. 57. 57. 57.7
25. 57. 57. 57.4
25. 57. 57. 57.4
Ambient*/
tl/mlnl
60.
60
59
60
60
60.
60
60.8
61.3
61
61.
61.
60.
60
60
60 3
60.3
60.0
60.1
60 1
60 1
60 3
60.2
60.2
Spike
Level
Ambient*/
(UH/P3)
292
292
295
294
292
292
291
290
288
288
289
288
291
290
293
291
293
294
294
293
294
293
293
293
ID
Back-
ground
(lia/m3)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5
0
10
1
M>2
Back-
0
0
0
15
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
0
o
0
1
Ambient Condition*
Out-
door Hind
Temp Speed
t*C> (m/eecl
75 0
75 0
73 0
64 0
63 0
62 0
61 0
61 5
58 4
58
57
57
56
36
59
62
62
66
70
74
75
79
79
77
Hind Ulacln
Urec- Bumldltr
tlon til
S 37
S 37
S 35
8 60
8 69
8 84
8 89
8 67
8 89
8 89
8 9*
8 94
8 94
8 94
V 100
V 84
v a*
V SO
SV 68
8 65
81 62
SI 50
S SO
8 52
•/ Jfrj.f injure end praiaon
pllng port*
-------
Table G-*. LEVEL 6 TEST DATA
Calculated Ployratea and Spike Levels
W>2 Sampling Syaeem Data
Pate
Tine
9-26-74
1800
1900
2000
2100
2200
2300
I-1 2400
ro
<* 9-27-74
0100
0200
0300
0400
0500
0600
0700
0800
0900
1000
1100
1200
1300
1400
1500
1600
1630
Flow
Bar Flow Temp -
Preaa Preai Meter
(mm Hal (m BO f'Cl
729 22 5
729 22.5
729 22 5
729 22 0
729 22 0
729 22 0
729 22 0
729 22 0
729 22 0
728 22 0
728 22 0
728 22 0
728 22 0
728 22 0
728 22 0
728 21 5
728 21 5
728 22 0
728 22 5
728 22 5
727 22 5
727 22 5
727 22 5
727 22 5
Flow- Venturl
rate Preesure
Meter Reading
(1/minl tan HOI
SB 0 259
58 1 259
58 1 255
58 0 259
58 2 261
58 5 261
58 5 261
58 6 262
58 8 262
58 5 262
58 5 264
58 6 264
58 S 264
58 7 260
57 7 258
58 0 257
57 7 257
57 5 255
57 7 255
57 3 253
SB 0 259
58 1 259
58 1 258
57 9 258
N2
Carrier
Flovrate
(cc/mln)
200
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
400
Permeation
lube Temp
C£i
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
25 1
Flowratea
Venturl
to 21*
+ 760 on Hg +
t*/1°) -
57 4
57 4
56 9
57 4
57 7
57 7
57 7
57 8
57 8
57 7
58 0
58 0
58 0
57 5
57 3
57 2
57 2
56 9
56 8
56 6
57 2
57 2
57 1
57 1
Hater Average
to 21* to 21
760 mm Hj + 760 mm
U /mini H/mln
57 4 57 4
57 57 4
57 57 2
57 57 4
57 57 7
57 57 8
57 57 8
58 0 57 9
58 2 58 0
57 9 57,8
57 9 57 9
58 0 58 0
57 9 57 9
58 1 57 8
57 1 57 2
57 4 57 3
57 1 57 2
56 9 56 9
57 1 56 9
56 7 56 6
57 1 57 2
57 4 57 3
57 4 57 2
57 2 57 1
Flovrate Solke
Level
Hg Ambient*' Ambient*'
!_ (1/minl (in/up)
K/
60 1 53 0°-'
60 2 102
59 102
60 102
60 102
60 102
60 102
60 101
60 101
60 101
60 101
60 101
60 101
60 101
59 103
59 102
59 103
59 103
59 103
59 103
60 102
60 102
60 102
60 0 102
ND
Back-
ground
(M/m3)
10
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Ambient
K>2 Out-
Back- door
ground Tea?
(ut/m3) CO
0 75
20 74
10 68
0 66
0 65
0 63
0 62
0 62
0 60
0 59
10 59
10 58
0 58
0 58
0 62
0 63
0 65
20 65
5 69
5 70
5 74
5 75
5 76
5 76
4
Condition!
Hind
Speed
(m/aee)
4
3
0
3
5
5
0
3
2
5
7
10
8
9
7
4
S
S
Hind Relative
Dlrec- Humidity
tlon_ Bi_
S 58
S 58
S 71
S 75
S 80
S 84
S 89
S 89
S 89
S 89
S 89
S 94
S 94
S 94
U 84
U 84
U 80
S 75
S 67
SZ 68
SB 61
SU 58
SU 59
SU 59
Thsa..perature tad pressure at tamp ling pores
lolclal iplka level Incorrect—changed to proper level at 1820
------- |