EPA-650/4-75-019
April 1975
Environmental Monitoring Series
EVALUATION OF EFFECTS
OF NO, CO2, AND SAMPLING FLOW RATE
ON ARSENITE PROCEDURE
FOR MEASUREMENT OF NO2
IN AMBIENT AIR
U.S. Environmental Protection Agency
Office of Research and Development
National Environmental Research Center
Research Triangle Park, N. C. 27711
-------
EPA-650/4-75-019
EVALUATION OF EFFECTS
OF IMOf CO2, AND SAMPLING FLOW RATE
ON ARSENITE PROCEDURE FOR MEASUREMENT
OF NO2 IN AMBIENT AIR
by
Michael £. Beard, Jack C. Suggs,
and John H Margcson
Quality Assurance and Environmental Monitoring Laboratory
Program Element No. 1HA327
ROAP No. 26AAF
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
RESEARCH TRIANGLE PARK. NORTH CAROLINA 27711
April 1975
-------
EPA REVIEW NOTICE
This report has been reviewed by the Office of Research and Development,
EPA, and approved for publication. Mention of trade names or commercial
products does not 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:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOM1C 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
concentrations. It also includes studies to determine the ambiont concen-
trations of pollutants in the environment and/or the variance of pollutants
as .1 lunclion of time or meteorological factors.
Copies of this report are available free of charge to Federal employees,
current contractors and grantees, and nonprofit orgamzations-as supplies
permit-from the Air Pollution Technical Information Center, Environmental
Protection Agency, Research Triangle Park, North Carolina 27711; or, for
a fee, from the National Technical Information Service, 5285 Port Royal,
Springfield, Virginia 22161.
Publication No. EPA-650/4-75-019
11
-------
ACKNOWLEDGMENT
The authors would like to thank Ms. E. Carol Ellis for her
meticulous efforts in maintaining calibrations for the N0?
permeation devices used in this study.
MI
-------
CONTENTS
Page
ABSTRACT v
CONCLUSIONS vii
Sections
I INTRODUCTION 1
II EXPERIMENTAL 2
III RESULTS AND DISCUSSION 8
IV REFERENCES 21
TECHNICAL REPORT DATA SHEET 23
LIST OF FIGURES
Figure
1 N02, NO, and ttL Atmospheric Generation System 4
2 N02, NO, and CU2 Study Format 10
LIST OF TABLES
Table
1 Effect of Sampling Flow Rate on Recovery of N02
in the Arsenite Procedure 8
2 N02i NO, and C02 Concentrations Selected for Study ... n
3 Recovery of N02 by the Arsenite Procedure at Generated
N02, NO, and C02 Levels 13
4 Analysis of Variance and Test for Linearity 14
5 Average Values of Method Response 16
6 Analysis of Variance of Bias 17
7 Average Bias 18
IV
-------
ABSTRACT
The arsenite method for measurement of NOo In ambient
air was investigated to quantify the effect of sampling flow
rate and of NO and C02 concentration on method response. NO and
COp were previously identified as positive and negative inter-
ferents, respectively, in the method.
3
The results show that flow rates of 220 to 270 cm /min had
no effect un the method response; higher flow rates decreased
tne method response. The flow rate range over which the method
response is unaffected is considered adequate for ambient sampling.
Atmospheres containing N02, C02, and NO were sampled with
the arsenite method in a 3x3x3 factorial experiment with five ob-
servations per cell. The concentrations were: N02 - 50, 150, and
250 yg/m3; NO - 50, 180, and 310 ug/m3; C02 - 200, 350, and 500 ppm.
A statistical analysis of the resultant data shows that:
1. The method response is linearly related to changes
in N02 level, as expected. Changes in levels of NO or C02
significantly change the slope of this linear relationship.
2. Over all concentrations, the method has an average
positive bias of 9.9 ug/m3. The 95 percent confidence interval
for this bias is +7.5 to 12.2 yg/m3.
3. The method response is related to the NO, C02, and
N02 concentration by
y = 4.36 + [1.12 + 0.0004 (NO - C02)]N02
-------
over the concentration ranges cited above where:
3
y = method response in pg/m
3
NO = NO concentration in yg/m
C02 = C02 concentration in1 ppm
NOp = N02 concentration in pg/m
The average bias introduced into the method by NO and C02
interference is small and does not necessitate applying a correction
to data obtained with the method within the concentration range des-
cribed.
VI
-------
CONCLUSIONS
A. Sampling Flow Rate
Nitrogen dioxide recoveries by the arsenite method using flow rates
3
of 220 and 270 on /min are the same; however, the recovery decreases
15 to 20 percent when the sample flow rate is increased from 270 to
380 cm /min. Flow rates above 270 cm^/min should be avoided.
U. N02, NO, and C02 Experiment
The following conclusions are valid for the range of values
specified in the experiment.
1. Method response is linear in relationship to changes
in N02 levels, but changes in levels of NO or C02
significantly change the slope of the linear
relationship.
2. Average method response changes significantly from
level to level of C02 or NO, but the amount of
change due to changes in C02 levels is not the same
for each level of NO.
•
3. On the average, the method response is significantly
higher than the actual N02 level by 9.89 ug/m .
4. The best linear equation of method response as a
function of N02> NO, and C02 is
Y(pg/m3) = 4.36 + [1.12 + .0004 (NO yg/m3 - C02 ppm)] N02
VII
-------
C. Utility of the Method
The average bias introduced into the method by NO and
COp interference is small and does not necessitate applying a
correction to data obtained with the method within the concentration
range described.
VIII
-------
EVALUATION OF EFFECTS
OF NO, CO2, AND SAMPLING FLOW RATE
ON ARSENITE PROCEDURE FOR MEASUREMENT
OF NO2 IN AMBIENT AIR
I. INTRODUCTION
The arsenite procedure for the determination of nitrogen
dioxide (NOo) in ambient air has been evaluated earlier by EPAJ
This evaluation showed the arsemte procedure to have a constant
collection efficiency of 82.2 ±4.5 percent. The evaluation also
identified nitric oxide (NO) and carbon dioxide (C02) as positive
and negative interferents, respectively. Sampling flow rate was also
shown to affect the method response. In order to determine the utility
of the arsenite method, the effect of these parameters needed to be
defined. This work reports our efforts to quantify the effect of
sampling flow rate and of NO and C02 concentrations on the arsenite
method.
-------
II. EXPERIMENTAL
A. General
The arsenite procedure used for this phase of the evaluation is
described in Appendix A of the original reportJ the method
consists of drawing ambient air through a glass tube having a
restricted orifice immersed in 50 ml of a solution containing 0.1
N NaOH and 0.1 percent w/w NaAs02- The N02 in ambient air is converted
to nitrite ion. The concentration of nitrite is then determined
colorimetrically by formation of a purple azo-dye.
The effect of sampling flow rate was measured by sampling
test atmospheres at various sampling rates. The effect of NO and
C02 concentrations was tested by sampling from test atmospheres
containing various amounts of N02> NO, and C02. Each of the three
test concentrations was held constant over the 24-hr sampling
period. In each case, the response of the method to each variable
was measured.
B. Test Atmosphere Generation
1. Nitrogen Dioxide
Test atmospheres containing known amounts of N0~ were
generated by diluting the effluent from gravimetrically calibrated
NOp-permeation devices with various measured volumes of purified air.
This procedure has been described by O'Keeffe and Ortman,^ and
Scaringelli e_t aj_.3»4 The permeation devices were made by the
Microchemical Analysis Section of the National Bureau of Standards
(NBS) and were calibrated frequently between sampling periods.
The stability of permeation rates from these devices with respect
to time has been well established.5
-------
The temperature of the devices was controlled by a water-
jacketed condenser which was maintained at 25.0 ±1° C by a
Forma Temp Jr. constant temperature bath. This apparatus is shown
in Figure 1. The N02 was flushed from the condenser by a flow of
100 cm /min dry N2. The permeation devices had rates of 1.062 ±0.001
and 0.836 ±0.001 ug/min.
Purified air was obtained by passing compressed (House) air
through silica gel for drying, treating with ozone to convert any
NO to NOp, and by passing through activated charcoal (6-14 mesh)
and molecular sieve (6-16 mesh, type 4A) to remove NOo and hydro-
carbons. Caroon dioxide was removed by passing the air through a
trap containing approximately 1 kg of Ascarite (8-20 mesh asbestos
particles impregnated with NaOH).
2. Nitric Oxide
Nitric oxide was added to the atmosphere by means of a "T"
connection in the NOo system as shown in Figure 1. A Kjeldahl trap
following the "T" insured mixing of the NO with the test atmosphere.
A cylinder of NO in N2 was analyzed by gas phase titration with 0^
as described in the Federal Register** and found to contain
92.4 ±3.1 ppm NO (113,700 pg/m3). The NO concentration determined
uy the gas phase titration was verified by comparing, on a chenri-
luminescent rfO-N02-NOx monitor, the N02 produced during the titration
with the output of an N02-permeation tube. The N02 concentrations
from these two sources agreed within 2 percent. The cylinder was also
-------
FLOW/METER, 0 TO 20 liters/mm
TEFLON TUBING
GLASS CONDENSER
TEFLON TUBING
FLOWMETER
0 TO 200 cm3/mm
DRY NITROGEN
3ft 1/4 in.
COPPER TUBING
PURIFIED
DILUTION AIR
THERMOMETER fJI PE™"f*™N
GLASS JOINT
TYGON TUBING
WATER
CIRCULATION
.. PUMP
VENT TO HOOD
NITRIC OXIDE
CARBON DIOXIDE
KJELDAHL MIXING
BULB
CONSTANT-TEMP. BATH
25.0 ± 0.10 C
LASS MAN! FOLD
TEFLON STOPCOCf
CONNECTOR (FOR SAMPLING TRAIN)
Figure 1. N02-NO-C02 atmospheric generation system.
-------
analyzed for NO, impurity using the triethanolamine-guaiacol-
sulfite (TGS) manual method^'8 and found to contain 1.6 ±0.19
ppm NOg (2181 yg/m3). Because of the N02 impurity in the NO cylinder
it was necessary to calculate the exact NOg and NO concentrations
using:
^x^-+ 2181 1= pg N02/m3 (1)
m
and
113,700 = ug NO/m3 (2)
v/here :
P.R. = permeation rate of the N02 device (s), ug/min
X = total dilution air flow rate, a/min
Y = NO flow rate, n/min
The portion of the N02 coming from the NO cylinder was approximately
O.b to 7.0 percent of the total N02 in the test atmosphere.
3. Carbon Dioxide
Carbon dioxide was added to the test atmosphere by means
of a "T" connection, as in the addition of NO (See III. B. 2). The C02
was supplied from NBS Standard Reference Material Ib74. These
C02 cylinders contained 7.01 to 7.03 ±0.07 mole percent
(70,100 to 70,300 ppm) C02- The C02 concentration of each cylinder
was verified by Orsat analysis.9 The cylinders were also
checked for NO and N02 impurities by means of a chemi luminescent
NO-NO, -NO monitor. None was found. The C0? concentration in each
£ A C,
-------
test atmosphere was calculated using:
II2 x c _ cl
r1 A **f*f\ ^* (*{\
""Total LU2 LU2
where:
Frn = flow rate from CO, cylinder, n/min
wUrt £
Fy t -j = combined flow rates in manifold, £/min
Crr = concentration of CO, added to test atmosphere, ppm
IsUo b
I
C co = concentration of C02 in test atmosphere, ppm
As stated in III. B. 1., C02 was removed from the purified
air by means of an Ascarite column. The purified air was tested
for C02 by samplino the air with a bubbler containing a
solution. No CC2 was detected.
C. Sampling
Samples were collected in quintuplicate by attaching five
sampling tubes to a corrmon manifold. The flow rate for each tube
was measured before and after sample collection as directed by the
method. The total flow rate into the common manifold was also
measured immediately before and after sampling and was compared with
the sum of the individual flows to insure that there were no leaks
in the system. Samples with a final flow more than 10 percent
different from the initial flow were rejected.
-------
U. Flow Control
1. N0-C02 Experiments
The samples collected for the NO-COo experiments were
collected at a rate of approximately 200 cm /min by using a 27-gauge
hypodermic needle as a critical orifice as suggested by the method.
A (iast Model 0211 oil less vacuum pump was used to maintain a
pressure drop across the orifice of approximately 0.6-0.7 atmos-
pheres. The total sampling time was about 20 hours.
2. Sampling Flow Rate Experiments
3 3
Flows of approximately 200 cm /min and 380 cm /min were
obtained by using 27 and 26 gauge hypodermic needles, respectively,
as critical orifices in the manner described above. A flow of
approximately 270 cm /mm was obtained by slightly crimping a
26 gauge needle until the desired flow rate was obtained. Again,
the total sampling time was about 20 hours.
E. Analysis
After sampling was completed the tubes were disconnected
from the manifold. Water lost by evaporation during the sampling
was replaced, and an aliquot of the sample was analyzed as described
in the method. A Beckman Model "B" Spectrophotometer was used
for the absorbance measurements. A standard curve of wgN02/ml versus
absorbance was determined for each experimental run.
-------
I. RESULTS AND DISCUSSION
A. Sampling Flow Rate
In the previous evaluation, 1 the results of a ruggedness
test show that increasing the flow rate from the value specified
3 3
in the method write-up, 180-220 cm /min, to 300 cm /min produces
a 14.4 percent decrease in the method response.
To quantify the effect of flow rate on method response,
test atmospheres containing approximately 60 and 700 ug N02/m
were sampled at flow rates of approximately 220, 270, and 380 cm3/min.
The results are given in Table 1. The recoveries at flow rates
o
of about 270 cm^/min (106.7 and 107.2 percent) are essentially the
same as the recoveries found at 220 cm3/min (109.9 and 105.7 percent).
At a flow rate of 380 cm3/min the recovery was 89 percent which is a
decrease of 19 percent. Thus, a decrease in recovery occurred between
270 and 380 cm3/min.
Table 1. EFFECT OF SAMPLING FLOU RATE ON RECOVERY
OF N02 IN THE ARSENITE PROCEDURE
NC2 generated,
ug/m3
51.9
60.3
666
63.3
671
63.5
663
Sampling flow rate,
cm /min
381.7
388.6
382.6
267.6
267.6
224.6
224.5
Percent recovery
(N02 analyzed/N02 generated) x 100
89.3
89.7
89.2
106.7
107.2
109.9
105.7
-------
Since a flow rate of 270 cm /min is considerably above the
upper limit specified in the method, no tightening of the flow
rate specifications is required.
It should be noted that the data in Table 1 show some N02
recoveries greater than 100 percent. This is due to removal of ambient
C0£ by the Ascarite scrubber such that the C02 concentration during
these experiments was considerably less than the ambient C02 concen-
tration present when the N02 collection efficiency (82 percent) was deter-
mined. The error in the method caused by determining the collection
efficiency in the presence of ambient levels of C02 is somewhere
between 0 and 3 percent. This statement is based on the results of
a collaborative test of the arsenite method,10 carried out in
ambient air, which shows that the bias of the whole method is only
-3 percent. Thus, the collaborative test indicates that any error
in the collection efficiency as a result of the above "C02 effect"
is quite small and does not justify redetermination of the
collection efficiency.
A more detailed description of the effect of C02 on the
method is given in the following sections.
B. NO and C02
1. Design
A series of experiments was conducted to quantify the
effect of NO and C02 concentrations on the recovery of N02 by the
arsenite method. Test atmospheres containing combinations of one
of three levels, at a constant concentration of each substance, were
sampled according to the method. Three levels were chosen to demon-
strate the effect of each material and to show if the effects were
-------
linear or curvilinear over the range of interest. This plan results
in 27 experiments and a format for this study is shown in Figure 2.
Each NO, C02, and NOp concentration was held constant
during an experiment, rather than varying the concentration with
time as would be the case in ambient air, because it would have
been difficult to accurately control the concentration under the
latter conditions. It is believed that the constant-concentration
N02 concentration, NO concentration, CO concentration,
ug/m3
ppm
NO, (150)-
(250)-
0(50
NO(180
0(310]
-------
conditions represent a reasonable approximation to ambient conditions,
and therefore, provide a useful means for evaluating interferences
to the method.
Nitric oxide and N02 levels for the experiments were derived from
examination of EPA NOp network data collected by chemiluminescence and
continuous colorimetric procedures. Carbon dioxide levels were derived
from the literature and from unpublished EPA data collected by a non-
12
dispersive infrared (NDIR) method.
Maximum, mid-range, and minimum concentrations were chosen
for each of the three parameters and their nominal values are
given in Table 2. The mid-range values were chosen to approximate
average ambient NO and C02 concentrations and the ambient air
standard for N(L (0.05 ppm'. Upper and lower levels were set
approximately at the maximum and minimum 24-hour averages for
the NO and N02 concentrations. The minimum C02 value is lower
than ordinarily found in ambient air and was chosen to allow the
ambient air average to be the mid-range value. Also, minimum
values for NO and N02 were set at low concentrations rather than
Table 2. N02> NO AND C02 CONCENTRATIONS
SELECTED FOR STUDY
NO
N02
Minimum
50 ug/m
(0.04 ppm)
V
50 gg/m3
(0.03 ppm)
200 ppm
Mid-range
180 ug/nf
(0.15 ppm)
150 pg/m
(0.08 ppm)
350 ppm
Maximum
310 ng/m3
(0.25 ppm)
250 ^g/m
(0.13 ppm)
500 ppm
11
-------
their minimum ambient levels of 0 ug/m since a concentration of
zero would obviously have no effect on the method.
The experiments were conducted in random order with each
atmosphere being sampled in quintuplicate. The data are given in
Table 3. Column 1 of Table 3 lists the experiment number, and
column 2 shows the random order in which the experiments were
conducted. Columns 3 through 5 show the actual generated values
of N02, NO, and CO,,. Columns 6 through 10 show the method
response of the arsenite procedure to the generated values.
2. Analysis and Discussion of Results
a. Analysis of Variance
The data in Table 3 were generated according to a
completely randomized 3x3x3 factorial design with five repeat samples
per treatment combination. The sources of variation in the method
response were identified in the analysis of variance, Table 4.
Equal spacing of the generated N02 levels facilitated investigations
about the linearity of the method response. For example, the two
degrees of freedom for NOg in line one of Table 4 were divided
into two single degree of freedom components for testing whether
the method response is linear or quadratic in relation to changing
NOo levels. The deviations of the acutal generated levels of NOp,
NO, and C02 from those given in Figure 2 are minimal. This is
common for an experiment of this type and does not seriously
affect the conclusions drawn from the analysis of variance. How-
ever, any linear relationships shown to be significant in the
analysis of variance were determined by regressing the method
response (Y) onto the actual generated N02 values rather than the
nominal levels.
12
-------
Table 3. RECOVERY OF N02 BY ARSENITE PROCEDURE AT GENERATED N02, NO, AND C02 LEVELS
Experiment
number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Random
order
number
24
9
11
25
3
5
18
16
15
26
14
13
17
1
4
20
27
7
23
6
22
2
8
21
12
19
10
Generated
N02, NO, C02,
ug/m3 wg/m3 pg/m3
48.0 49.5 194.9
47.9 51.3 354.9
47.7 49.1 509.7
48.3 182.7 204.0
47.6 153.4 362.3
48.3 183.6 516.9
48.2 315.9 203.7
48.0 306.0 350.7
Method response, pg/m^
N02(l)
54.0
50.1
47.1
60.8
52.1
49.6
58.9
58.8
48.4 307.8 521.3 ' 53.9
145.0 49.5 201.4 : 156.2
N02(2)
51.9
47.6
46.8
59.4
51.2
48.8
56.6
56.5
54.4
146.7
146.1 49.9 351.0 155.1 ! 145.2
146.0 48.8 505.0 • 141.2 , 140.3
145.6 184.3 203.1
144.5 179.9 347.9
145.7 182.7 510.0
146.7 314.5 203.6
145.4 307.9 349.5
145.6 307.1 512.1
243.5 50.3 203.9
242.0 49.5 357.6
243.3 50.2 511.6
243.2 179.5 201.0
242.8 180.6 354.9
244.2 182.4 < 521.1
240.8 317.0 202.8
240.9 311.7 356.2
241.8 308.8 499.3
165.6 • 156.5
179.1 177.7
141.8 137.3
170.4
165.0
146.7
165.4
153.7
142.9
260.5 ' 241.6
254.5
231.4
275.7
262.2
239.5
286.3
263.3
268.1
229.2
219.3
260.0
233.0
242.6
267.3
247.4
257.6
N02(3)
53.9
49.4
48.7
57.3
56.3
50.8
57.7
57.1
53.7
156.1
153.7
141.9
173.8
173.5
142.0
171.0
163.6
151.7
257.9
250.3
229.0
280.3
260.1
241.7
285.7
259.9
286.0
N02(4)
49.5
49.5
47.5
57.7
52.9
49.7
53.3
59.6
51.6
154.0
150.5
139.3
162.8
179.2
133.6
171.4
161.8
145.1
255.3
242.2
228.7
276.7
253.4
225.3
271.6
251.1
252.4
N02(5)
54.1
5A.4
46.8
77.7
51.5
51.4
61.0
58.3
51.0
160.3
153.5
138.8
167.0
178.1
142.3
175.2
167.5
152.8
266.2
259.9
231.5
283.3
259.8
240.3
290.8
263.4
257.1
-------
Table 4. ANALYSIS OF VARIANCE AND TEST FOR LINEARITY
Line
1
2
3
4
5
6
7
8
9
10
11
Source
N02
Linear
Quadratic
co2
NO
N02 * C02
Linear x C02
Quadratic x C02
N02 x NO
Linear x NO
Quadratic x NO
12 ; co2 x NO
13 ' N02 x C02 x NO
14
ERROR
TOTAL
D.F.
2
1
1
2
2
4
2
2
4
2
2
4
8
108
134
Sum of squares
915474.98
915365.02
109.95
7846.80
4369.60
2783.34
2658.35
Mean square
457737.49
915365.02
109.95
3923.40
2184.80
695.84
1329.18
124.98 ' 62.49
1382.19 345.55
1216.78 608.39
165.41 82.71
965.61 241.40
1691 ?1 211.40
4525.67 41.90
939039.41
F
10923.38
21846.42
2.62 N.S.
93.62
52.14
16.60
31.72
1.49 N.S.
8.25
14.50
1.97 N.S.
5.67
5.04
N.S. = not significant at the u = 0.05 level.
From the second line in Table 4, the method response is
significantly linear in fit over the range of generated NOp values
and is given by
Y = 4.41 + 1.04 (N02) (4)
From lines seven and ten in Table 4, this relationship is
shown to remain linear but changes significantly in slope as the
generated levels of C02 or NO change. The relationships between
method response and generated N02 for each level of C02 are:
(200 ppm C02) Y = 4.48 + 1.09 (N02)
(350 ppm C02) Y = 7.45 + 1.03 (N02)
(500 ppm CO^) Y = 1.20 + 0.99 (N02)
-------
As the level of C02 increases, the slope decreases and the
slopes are significantly different according to line seven of
Table 4. The relationships between measured N02 and generated
N02 for each level of NO are:
(50 pg/m3 NO) Y = 3.24 ••- 0.99 (N02)
(180 ug/m3 NO) Y = 7.24 + 1.03 (N02)
(310 ug/m3 NO) Y = 2.54 + 1.09 (N02)
As the level of generated NO is increased the slope increases and
the slopes are significantly different according to line ten of
Table 4. The fact that the coefficients are the same in both
sets of relationships but appear in reverse order for increasing
NO levels as compared to increasing COp levels is merely coinci-
dental.
One must be careful not to extrapolate the method response
to NO^ values near the lower minimum detectable limit of 9 pg/m . °
The values of the generated N02 levels are all some distance from
this value and a straight line not going through zero proves to
oe the best fitting line. The explanation may be that the true
relation between method response and generated N02 is curved
near zero out this curvature is slight in the range within which thp
NO, is being generated.
The effect of different levels of NO and C02 on the metnod
response at three different levels of N02 may be seen in Table 5.
For instance, as the nominal CO, level is increased from 200 ppm
to 350 ppm (75 percent increase), the average method response decreases
by 7 percent, as seen in the lower margin. A further increase in the
C02 level from 350 ppm to 500 ppm causes a further drop of 7 percent in
15
-------
Table 5. AVERAGE VALUES OF METHOD RESPONSE3
NjO.ng/m3
N02,vg/m3 \^
50
150
250
NO x C02
2, ppm
200
50
52.68
154.66
256.30
154.55
180
62.58
165.14
275.20
167.64
163.90
310
57.20
171.68
280.34
169.51
350
50
50.20
151.60
247.22
149.67
180
52.80
177.52
2b3 70
161 34
156.7?
310
58. 06
162.32
257.02
159.13
500
50
47.78
140.30
227.98
138.69
180
50.06
139.40
237.88
142.45
145.38
310
52.92
147.84
264.24
155.00
53.84
156.61
255.54
Grand mean
155.26
*Each cell is the average of 5 values, and the right-hand and lower margins are the average of 45 method response
values. The NO x CO- line represents the average of 15 values.
-------
the average method response. However, for each level of NO, the
decrease in average method response due to increasing C02 is not
the same as seen in the NO x (X^ margin. As the nominal NO level
3 3
is increased from 50 yg/m to 180 yg/m (overt 300 percent), the average
method response increases by 6 percent. A further increase in the NO
3 3
level from 180 yg/m to 310 yg/m (67 percent increase) results in a 3
percent increase in the average method response. These results are not
i
readily seen in Table 5 without some additional averaging.
b. Bias the Precision
The overall bias of the arsenite method, calculated by
taking the difference (method response - generated NOo) and aver-
, £,
aging over all 135 values, is +9.88 yg/nr. The 95 percent confidence
3 3
interval for this bias is (7.52 yg/m , 12.24 yg/m ) indicating
that this bias is real and significantly different from zero.
However, the bias does not remain constant but is affected sig-
nificantly by changes in NO and C02 as seen in the analysis of
variance of this difference summarized in Table 6. As seen in
Table 7, more than tripling the NO from 50 yg/m to 180 ug/m
Table 6. ANALYSIS OF VARIANCE OF BIAS
Source
NO
co2
NO x C02
Error
Total
D.F.
2
2
4
126
134
Sum of squares
4553.68
8053.30
977.41
12376.92
25961.32
Mean square
2276.84
4026.65
244.35
98.22
F
23.173
40.99a
2.49a
Significant at the a = 0.05 level
17
-------
Table 7. AVERAGE BIAS
(yg/m3)3
NO, yg/m3
50
180
310
C02, ppm
200
9.05
21.94
24.27
18.42
350
4.34
15.79
14.36
11.5
500
-6.98
-3.62
9.73
-.29
2.14
11.37
16.12
9.88
Each cell is the average of 15 values, and the
margins represent the average of 45 values.
increases the overall bias five-fold from 2.14 yg/m3 to 11.37 yg/m3.
A further increase of NO to 310 yg/m increases the bias from
3 3
11.37 yg/m to 16.12 yg/m (approximately 42 percent). A significant
decrease in bias is observed in going from level to level of
increasing COp.
The smallest average positive bias is 4.34 pg/m and this
3
occurs when NO is at approximately 50 yg/m and C02 is at 350 ppm.
The smallest negative bias is -3.62 yg/m and this occurs for NO at
3 3
180 yg/m and C02 at 500 yg/m . Both of these values are not
significantly different from zero at the a = 0.05 significance level.
This indicates that the method response is not significantly affected
o
by these level combinations. The cell average of -6.98 yg/m is also
not significantly different from zero. All other cell averages in
Table 7 are different from zero since their absolute value exceeds the
upper 95 percent confidence limit of 7.09 yg/m3 for averages of 15 values,
o
The largest bias (+24.27 yg/m ) occurs at low C02 levels (approximately
18
-------
200 ppm) and high NO levels (approximately 310 yg/m ). This bias
would most likely occur only rarely in practice because its occurrence
would require maintaining these extremes of (XL and NO concentration
over a 24-hour period.
The precision of the method is a measure of the closeness
of two method responses each determined by the same operator using
the same set of instruments under the same environmental conditions
(constant N02, NO, and C02). If two method responses, i.e.,
replicates, differ by more than 1.96^2 V41-90 = 17.96 yg/m , we
must suspect operator problems, instrument failure, or unstable
environmental conditions. If a value of N02 is generated in the
range specified by the experiment and the method response is
evaluated at this level, then no matter what values of COo or NO
are present (just as long as they too are in the range specified
by the experiment), the difference (bias) must exceed
(1.96 V 98.22 = 19.42) yg/m3 to be declared significant at the
a = 0.05 significance level.
c. Response Surface and Prediction
To best describe the performance of the method under known
conditions a linear regression of the method response onto the
actual generated values of N02, NO, and C02> including all their
respective squares and cross products, was performed. Using
backward elimination, all variables that did not account for a
significant portion of the total variation in the method response
were discarded. The final equation is
Y(yg/m3) = 4.36 + [1.12 + 0.0004 (NO yg/m3 - C02 ppm)]N02 yg/m3 (5)
19
-------
which accounts for 99.02 percent of the total variation in the method
response (Y). The remaining 0.98 percent is attributed to the residual
error variance of 70.36 yg/rrr* wi th 131 degrees of freedom. This
equation is known as a response surface. A response in this case
explains how the method will respond to known concentrations of
NOo in the presence of known concentrations of COo and NO but
only for the range specified by the experiment. That is, if we
know the levels of N02, C02,and NO we can predict what the method
response will be.
20
-------
IV. REFERENCES
1. Beard, M.E. and J.H. Margeson. An Evaluation of the Arsenite Pro-
cedure for Determination of Nitrogen Dioxide in Ambient Air.
Environmental Protection Agency, Research Triangle Park, N.C.
Publication No. EPA-650/4-74-048. November 1974.
2. O'Keefe, A.E. and G.C. Ortman. Primary Standards for Trace Gas
Analysis. Analytical Chemistry. 3JJ:760, 1966.
3. Scaringelli, P.P., S.A. Frey, and B.E. Saltzman. Evaluation of
the Teflon Permeation Tubes for Use with Sulfur Dioxide. Journal
of the American Industrial Hygiene Association, 28^:260, 1967.
4. Scaringelli, P.P., A.E. O'Keeffe, E. Rosenburg, and J.P. Bell.
Preparation of Known Concentrations of Gases and Vapors with
Permeation Devices Calibrated Gravimetrically. Analytical
Chemistry. 4£(8):871, 1970.
5. Rook, H.L., R.G. Fuerst, and J.H. Margeson. Progress Report: EPA-
NBS Study to Determine the Feasibility of Using NO? Permeation
Devices as Standards. December 1972 - January 1973. Environmental
Protection Agency, Research Triangle Park, N.C.
6. Title 40 - Protection of Environment. Candidate Reference Method
for Determination of Nitrogen Dioxide. Federal Register. 38(110),
June 8, 1973.
7. Mulik, J, R. Fuerst, M. Guyer, J. Meeker, and E. Sawicki. Develop-
ment and Optimization of 24-Hour Manual Method for the Collection
and Colorimetric Analysis of Atmospheric N02- International
Journal of Environmental Analytical Chemistry. 5_, 1974.
8. Fuerst, R.G., and J.H. Margeson. An Evaluation of the TGS-ANSA
Procedure for the Determination of Nitrogen Dioxide in Ambient Air.
Environmental Protection Agency, Research Triangle Park, N.C.
Publication No. EPA-650/4-74-047. November 1974.
9. Burrell Manual for Gas Analysts. 7th Ed. Pittsburgh, Burrell
Corporation, 1951.
10. Constant, P.C., C. Sharp, and G.W. Scheil. Collaborative Testing
of Methods for Measurement of N02 in Ambient Air. Vol. 1 - Report
of Testing. Midwest Research Institute, Kansas City, Mo. Prepared
for Environmental Protection Agency, Research Triangle Park, N.C.
under Contract No. 68-02-1363. Publication No. EPA-650/4-74-019-a.
June 1974.
21
-------
11. Clarke, John F, A Meteorological Analysis of Carbon Dioxide Con-
centrations Measured at a Rural Location. Atmospheric Environment.
3:375-383, 1969.
12. Ortman, Gordon. Personal communication. Environmental Protection
Agency, Research Triangle Park, N.C.
13. Title 40 - Protection of Environment. National Primary and Secondary
Ambient A1r Quality Standards. Federal Register. 36(84):8186-8187,
April 30, 1971.
22
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-650/4-75-019
3. RECIPIENT'S ACCESSIOI*NO.
4 TITLE AND SUBTITLE
Evaluation of Effects of NO, C02» and Sampling Flow Rate
on Arsenite Procedure for Measurement of NG*2 in Ambient
Air
5 REPORT DATE
/April 1975
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
M.E. Beard, J. Suggs, and J. Margeson
8. PERFORMING ORGANIZATION REPORT NO
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Protection Agency, NERC
Quality Assurance & Environmental Monitoring Laboratory
Methods Standardization & Performance Evaluation Branch
Research Triangle Park, N. C. 27711
10 PROGRAM ELEMENT NO
1HA327
11 CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Office of Research and Development
Washington, D. C. 20460
13 TYPE OF REPORT AND PERIOD COVERfcO
Final
14 SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
arscnite method for measurement of N02 in ambient air was investigated
to quantify the effect of sampling flow rate and NO and C02 concentration on method
response. NO and C02 were previously identified as positive and negative interferents
in the method.
The results show that flow rates of 220 to 270 cm3/min had no effect on the methoi
response; higher flow rates decreased the method response. The flow rate range over
which the method response is unaffected is considered adequate for ambient sampling.
Atmospheres containing N02, C02, and NO were sampled with the arsenite method in
a 3x3x3 factorial experiment with five observations per cell. The concentrations were
N02--50, 150, and 250 yg/m3; NO--50, 180, and 310 ug/m3; C02--200, 350, and 500 ppm. A
statistical analysis of the resultant data shows that: (1) The method response is lin-
early related to changes in N02 level, as expected. Changes in levels of NO or C02 sig
nificantly change the slope of this linear relationship. (2) The method has an average
positive bias of 9.9 yg/m3 over all concentrations. The 95 percent confidence interval
for this bias is +7.5 to 12.2 yg/m3, and 3). The method response is related to the NO,
C02, and N02 concentration by y=4. 36+1. 12+0. 0004 (NO-C02) N02- Over the concentration
ranges cited above where: y = method response in yg/m3, NO = NO concentration in yg/m3
C02 = C02 concentration in ppm, and N02 = N02 concentration in yg/m3.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b IDENTIFIERS/OPEN ENDED TERMS
c COSATI Field/Group
Measurement methods
Arsenite procedure
Nitrogen dioxide
Nitric oxide
Carbon dioxide
18. DISTRIBUTION STATEMENT
Unlimited
19 SECURITY CLASS (This Report)
Unclassified
21 NO OF PAGES
30
20 SECURITY CLASS (Thispage)
Unclassified
22 PRICE
EPA Form 2220-1 (9-73)
23
------- |