United States
Environmental Protection
Agency
Environmental Monitoring Systems
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S4-84-086 Dec. 1984
vvEPA Project Summary
-¥
Stability Evaluation of Sulfur
Dioxide, Nitric Oxide and Carbon
Monoxide Gases in Cylinders
R. C. Shores, F. Smith, and D. J. von Lehmden
The purpose of this project was to
assess the stability of certain reactive
gases used to calibrate/audit air quality
analyzers as a function of time. The
objective of the assessment was to
determine if EPA's present guidelines
suggesting recertification every 6
months could be relaxed to some longer
time period. The reactive gases were
contained in aluminum and steel cyl-
inders and include: carbon monoxide in
air (CO/air); carbon monoxide in nitro-
gen (CO/N2); nitric oxide in nitrogen
(NO/IM2); and sulfur dioxide in nitrogen
(SO2/N2). The assessment included
stability data for a total of 316 com-
pressed gas cylinders, 300 aluminum
and 16 steel. The 300 aluminum cyl-
inders consisted of: 37 CO/air; 86
CO/N2; 103 NO/N2; and 74 SO2/N2.
Time intervals covered by the stability
data ranged from 10 to 58 months, with
the average time period being in excess
of 12 months.
Stability of each pollutant category
(e.g., CO/air) was assessed by linear
regression analysis of the measured
concentration versus time in monthsfor
each cylinder. The average percent
change of the median concentration
within each pollutant category was
calculated for time periods of 6, 12,18
and 24 months. The criterion for judging
a pollutant category to be stable over a
specific time period was an average
change of ±1 percent or less over that
time period.
Based on the results of the analysis it
was recommended that a recertification
period of 18 months be adopted for
reactive gases in aluminum cylinders as
follows: CO/air at concentrations be-
tween 5 and 45 ppm; CO/N2at concen-
trations between 5 and 5,000 ppm;
NO/IM2 at concentrations between 10
and 150 ppm; and SO2/N2 at concen-
trations between Sand 375 ppm. It was
also recommended that the present
suggested recertification period of 6
months be retained for all other reactive
gas/cylinder type combinations.
This Project Summary was developed
by EPA's Environmental Monitoring
Systems Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
Introduction
A model used to estimate the optimum
time interval for recertification of a pollu-
tant concentration in a cylinder would
ideally include the following terms:
• the indicated stability of that pollutant
at a specified concentration and in a
particular cylinder type as a function of
time;
• the precision, and potential for bias, of
the certification process;
• the costs associated with recertifica-
tion; and
• the costs associated with generating
environmental data of poor quality
(when used as a calibration or audit
standard) and the cost associated with
rejecting good quality data as bad
(when used as an audit standard).
-------
Due to limited resources this assess-
ment of an acceptable recertif ication time
period was limited to consideration of the
first two items listed above. The stability
data contained in this report were evalu-
ated to determine if the recertification
period stated in Section 1.2.5 of EPA
Protocol No. 2 could be extended beyond
six months for carbon monoxide (CO),
nitric oxide (NO), and sulfur dioxide (S02)
gases contained in aluminum steel cylin-
ders. Traceability Protocol No. 2 provides
procedures for establishing the true con-
centration of reactive gases using
National Bureau of Standards (NBS)
gaseous Standard Reference Materials
(SRMs) or EPA/NBS-approved commer-
cial Certified Reference Materials (CRMs).
Data Collection
Reactive gas stability data were solic-
ited from federal, state, and local agen-
cies, SLAMS Ambient Air Audit Centers,
and selected specialty gas manufacturers.
Also included in these data are reactive
gas cylinders analyzed by the EPA-
EMSL/QAD standards laboratory. Data
were requested on reactive gas cylinders
that had been analyzed over a time period
of twelve months or longer.
Stability data were received for 37
CO/air, 86 CO/N2, 103 NO/N2, and 74
S02/N2 for a total of 300 compressed
gases in aluminum cylinders. Also, in-
formation on 1 5 CO/air and 13 CO/N2 in
steel cylinders was received. The analysis
period covered by the reported data
ranged from 10 to 58 months.
Procedures
Each pollutant/balance gas category
(CO/air, CO/N2, NO/N2, SO2/N2) was
divided into concentration ranges and
cylinder construction (aluminum or steel)
for analysis. The concentration ranges
were selected according to natural gaps
in the concentration distribution of the
stability data and to allow for assuming
constant variances over these relatively
small concentration intervals.
Stability of the reactive gases was
assessed by calculating a linear regres-
sion equation of concentration versus
time in months from the initial analysis
for each cylinder. The slope of the resul-
tant regression equation was in units of
ppm change per month. When the slope
was multiplied by a number of months,
the product equaled a change in pollutant
concentration (ppm) for that specific
number of months. Also calculated for
each pollutant category werethe average
slope, the standard deviation of the
slopes, and, based upon the median
concentration in each concentration inter-
val, an average percent change for 6, 12,
18, and 24 months, and the standard
deviation of the percent change for an 18
month period (Table 1). Within a specific
pollutantjsategory, the average percent
change (P) was used as a measure of
stability. Two times the standard deviation
of the percent change (sp) represented the
approximate 95 percent probability limits
for the percent change of individual
pollutant cylinders in the category. The
pooled standard deviation for analysis
results for a pollutant cylinder category
was used to estimate the range of error
that can be expected for the pollutant
concentration, based on one analysis.
The following variables and formulas
were used to calculate the stability of the
reactive gases:
X = Median concentration of a con-
centration range or pollutant
category, ppm.
m = Average slope for all cylinders in
a concentration range ppm
change/month.
sm = Standard deviation of the slope
calculated for each concentra-
tion range, ppm change/month.
T = Time since the original analysis,
months.
P = 100(m)(T)/X, average percent
change for a median concentra-
tion over a specified time period,
percent.
sp = 100(sm)(T)/X, standard deviation
of percent change at the median
concentration, over time period
T, percent.
±2 sp = 95% probability limits for the
percent change of a reactive gas
concentration, percent.
sy = standard deviation of the meas-
ured concentrationsfor a specific
cylinder, ppm.
SyP = [((n,-1) s?, + - + (nk-1) Sykt/fn, +
— + nk-k)]1/2, pooled standard
deviation of the measured con-
centrations across cylinders
within a pollutant cylinder cate-
gory, ppm.
Results
Results of the analysis of stability data
for thirty-seven aluminum cylinders con-
taining carbon monoxide with a balance
gas of air are presented here as an
Table 1. Summary of Reactive Gases Contained in Aluminum and Steel Cylinders*
Concentration Median, X Slope, m
Pollutant Number of Range Concentration (ppm/month}
Category Standards (ppm) (ppm} ^ I
Average
Percent Change, P
Sp
118 mo)
6 mo 12 mo 18 mo 24 mo
Aluminum Cylinders
CO/Air
CO/Air
co//v2
CO/N2
NO/N2
NO/N,
SOz//V2
S02//V2
CO/Air
CO/N,
CO/N2
10
22
16
17
16
26
7
25
71
10
21
33
10
5-11
17-45
5-15
19-46
100-1000
2500-5000
10-22
43-55
80-113
9-22
45-56
78-110
152-375
9-11
19-43
1000
7
30
10
30
500
3750
16
49
97
15
51
94
264
-0.004
0.000
+0.001
+0.006
-0.112
+0.009
-0.003
-0.025
-0.009
-0.003
+0.007
-0.039
+0.061
0.013
0.017
0.010
0.018
0.268
4,922
0.014
0.070
0.112
0.020
0.070
0.115
0.103
-0.3
0.0
+0.0
+0.1
-0.1
0.0
-0.1
-0.3
-0.1
-0.1
0.0
-0.2
+0.1
-O.7
O.O
-1.0
0.0
-1.4
0.0
+0.1 +0.2 +0.2
+0.2 +0.4 +0.5
-0.3 -0.4 -0.5
0.0 0.0 0.0
-0.2 -0.3 -0.5
-0.6 -0.9 -1.2
-0.1 -0.2 -0.2
-0.2 -0.4 -0.5
+0.2 +0.2 +0.3
-0.5 -0.7 -1.0
+0.3 +0.4 +0.6
Steel Cylinders
10 -0.028 0.020 -1.7 -3.4 -5.0 -6.7
31
1000
-0.032 0.029 -0.6 -1.2 -1.9 -2.5
0.153 0.095 -0.1 -0.2 -0.3 -0.4
3.3
1.0
1.8
1.1
1.0
2.4
1.6
2.6
2.1
2.4
2.5
2.2
0.7
3.6
1.7
0.2
"See Procedures section for definition of terms.
-------
example. The same type of information is
contained in the subject report for the
other reactive gased included in this
study. This stability data (cylinder number,
date of initial analysis, initial concentra-
tion, time in months (T) since the initial
analysis, and the analyzed concentration
in ppm are contained in Table 2. Also
contained in this table are the linear
regression coefficients: slope (m), inter-
cept (b) calculated from the analyzed
concentrations versus time in months
since the initial analysis, the standard
deviation (sy) of the y axis values (meas-
ured CO concentrations), and summary
statistics for each category. The CO/air
data were divided into two concentration
ranges, 5 to 11 and 17 to 45 ppm for
analysis.
Concentration Range 5 to
11 ppm
For CO/air in aluminum cylinders,
there were 10 cylinders with measured
concentrations ranging from 5 to 11 ppm
(see Table 2). The 10 calculated regres-
sion equation slopes included seven
negative, two positive and one zero
values. The average slope was m = -0.004
ppm change/month and the standard
deviation of the slopes about that average
was sm = 0.013 ppm change/month (see
summary statistics in Table 3). A Stu-
dent's t-test indicated that average slopes
of 0.004 ppm change/month or larger
could be expected to occur more than 20
percent of the time due to chance causes
alone. Thus, there was no reason to reject
the hypothesis that the mean slope was
zero.
The approximate 95 percent probability
interval for the uncertainty (excluding
bias) of a CO/air concentration estimated
by one analysis is ±2 syp = 0.48 ppm or
±10 percent at a concentration of 5 ppm.
Concentration Range 17 to
45 ppm
There were 22 cylinders with measured
CO/air concentrations ranging from 17
to 45 ppm (see Table 2).
The calculated regression equation
slopes include thirteen negative, eight
positive and one zero values. The average
slope was m = -0.000 ppm change/month
and the standard deviation of the slopes
Table 2. Stability Data for CO/Air in Aluminum Cylinders*
Analysis No.
Cylinder No.
BAL695-8
BAL1180-9
CAL6857-1
CC3384-11
BAL1193-10
AAL1053-S
AAL420-6
AAL1 508-3
AAL17W-4
AAL459-7
Initial A natysis
Date
05-13-80
01 -24-80
05-01-81
05-01-81
01-24-80
11-04-80
11-04-80
11-04-80
11-04-80
11-04-80
Cone.
(ppm)
5.1
5.6
9.2
9.7
W.I
10.6
10.6
10.7
10.7
10.7
Time
T
(months)
10
9
4
17
9
10
10
10
10
10
Cone.
(ppmj
5.6
5.6
8.7
9.7
10.1
10.0
10.1
10.1
10.2
10.1
Time
T
(months)
18
19
10
19
23
23
23
23
23
Cone.
(ppm)
5.4
5.4
9.0
10.0
10.3
10.3
10.3
10.3
10.3
Subsequent Analysis
Time
T
(months)
40
16
40
Time Time
Cone. T Cone. T Cone.
(ppm) (months) (ppm) (months) (ppm)
5.5
9.2 19 9.6 24 9.2
10.0
Regression
Coefficients
m
(ppm/mo)
+0.018
-0.003
+0.018
0.000
-0.003
-0.011
-0.012
-0.016
-0.016
-0.016
b
(ppm)
5.20
5.58
8.93
9.70
10.10
10.42
10.46
10.54
10.58
10.54
Standard
Deviation
Sy
(ppm)
0.25
0.10
0.29
0.00
0.06
0.30
0.25
0.31
0.27
0.31
Summary Statistics: m = -0.004 ppm change/month; sm = 0.013 ppm change/month; s,p = 0.24 ppm
CC10574-32
CC15074-19
BAL1 187-27
CC4222-18
CAL6997-16
BAL1 172-28
SGCC27593-12
BALI 190-29
SGCC27603-53
SGCC27774-13
AAL2644-21
AAL785-23
AAL1 707-25
AAL4338-22
AAL 1 990-24
SGAL3235-14
CC15150-31
BALI 001 -26
BAL1181-17
SGAL639-15
BAL1 171 -30
AAL5920-20
09-01-80
05-01-81
05-13-80
12-01-81
05-01-81
01-24-80
01-13-82
01-24-80
01-13-82
01-13-82
11-04-80
11-04-80
11-04-80
11-04-80
11-04-80
01-29-82
09-01-80
05-13-80
05-01-81
02-03-82
01-24-80
06-11-81
17.9
18.0
18.1
18.2
18.9
19.6
20.1
29.6
29.0
40.2
40.6
40.8
40.8
40.9
41.2
41.8
41.9
42.8
43.1
43.8
44.9
45.1
8
17
10
10
4
9
2
9
2
2
10
10
10
10
10
10
8
10
4
11
9
9
18.0
17.9
18.5
18.3
19.2
19.6
20.2
29.7
29.0
40.4
40.4
40.5
40.6
40.6
40.9
41.7
41.8
42.8
43.2
43.8
45.1
45.1
25
18
10
19
5
19
5
5
23
23
23
23
23
25
18
10
19
19
17.9
18.3
18.8
19.4
20.1
29.4
29.1
40.4
40.5
40.6
40.7
40.7
41.0
41.7
42.6
42.2
44.2
44.7
16
40
8
40
8
8
19.4
19.4
20.2
29.7
29.0
40.3
19
14
14
14
19.9
20.6
29.3
40.4
24
16
16
16
19.0
20.6
29.4
40.6
16
43.3
19
43.7
24
43.5
-0.001
-0.006
+0.012
+0.010
+0.018
-0.006
+O.O34
+0.001
+0.025
+0.015
-0.004
-0.008
-0.004
-0.008
-0.008
-0.010
-0.008
-0.011
+0.026
0.000
-0.038
-0.021
17.94
18.00
18.19
18.20
18.98
19.60
20.05
29.58
28.95
40.27
40.54
40.72
40.74
40.82
41.12
41.80
41.88
42.83
42.86
43.80
45.09
45.17
0.06
0.07
0.20
0.07
0.41
0.12
0.24
0.14
0.17
0.13
0.10
0.15
0.10
0.15
0.15
0.07
0.10
0.12
0.52
0.00
0.47
0.23
Summary Statistics/ m = 0.000 ppm change/month: sm = 0.017 ppm change/month: syp = 0.33 ppm
"See Procedures section for definition of terms.
-------
about the average was sm = 0.017 ppm
change/month (see summary statistics
in Table 2). Based on m = 0.000 ppm
change/month there was no reason to
reject the hypothesis that the true mean
slope was zero.
The approximate 95 percent probability
interval for the uncertainty (excluding
bias) of a CO/air concentration estimated
by one analysis is ±2 syp = 0.66 ppm or
±2.2 percent at the median concentration
of 30 ppm.
Discussion
For reactive gases in aluminum cylin-
ders the average percent change (P),
based on the average slope (m), was 1
percent or less for all pollutant categories
for an eighteen month period. For this
reason, an eighteen month recertifieation
time period is being recommended for
reactive gases contained in aluminum
cylinders. The average percent change
and the 95 percent probability interval of
the percent change have been summar-
ized in Table 3.
The average percent change for CO/air
and CO/N2 gases contained in steel
cylinders approaches or is less than 1
percent for the 6 month period only. The
average percent change and the 95
percent probability limits for the percent
change for CO contained in steel cylinders
for an 18 month period have been calcu-
lated in Table 3.
Conclusions and
Recommendations
Data contained in this report, indicate
that reactive gases (CO/air, CO/Na,
NO/Na, S02/N2> contained in aluminum
cylinders, will remain stable for at least
eighteen months. However, it is desirable
that a pollutant gas being used as a
calibration or audit standard be analyzed
at least twice within its useful lifetime. A
pollutant gas user should take into
account the consumption rate of the gas
and plan for a second analysis before the
cylinder is empty. This second analysis
will provide assurance of the stability of
the standard over its period of use.
It is recommended that the present
requirement for reanalysis of reactive
gases every six months be extended to
eighteen months for the following:
• Carbon monoxide in air in aluminum
cylinders at concentrations be-
7'able 3. Average Percent Change and Probability Limits of Reactive Gases Contained in
Aluminum and Steel Cylinders for Eighteen Months
Concentration
Pollutant/Balance
Gas
Range
(ppm)
Median
fppm)
P±2sf*
(Percent)
Aluminum Cylinders
CO/Air
CO /Air
CO/N,
CO/Nz
NO/Ni
SOi/Nt
SO2//V2
SOi/N*
SO2//V2
CO/Air
CO/Nt
5 to 11
17 to 45
5 to 15
14 to 46
100 to 1000
2500 to 5000
10 to 22
43 to 55
80 to 113
9 to 22
45 to 56
78 to 110
152 to 375
7
30
10
30
500
3750
16
49
97
15
51
94
264
Steel Cylinders
9 to 11
JO to 43
WOO
10
31
1000
-1.0 ±6.6
0.0 ±2.0
+0.2 ±3.6
+0.4 ± 2.2
+0.4 ± 2.0
-0.0 ±4.8
0.3 ± 3.2
-0.9 ± 5.2
-0.2 ± 4.2
-0.4 ±5.0
-0.2 ± 5.0
+0.7 ±4.4
-0.4 ± 1.4
-5.0 ± 7.2
-1.9 ±3.4
-0.3 ± 0.4
'P = A verage percent change at the nominal concentration over a specified time period, percent.
sf = Standard deviation of the percent change at the median concentration over the time period,
percent.
tween5 ppm and45 ppm. NOTE: To
ensure uncertainties of less than
10 percent in measured concentra-
tions near 5 ppm, it may be neces-
sary to use the average of two
Protocol No. 2 analyses.
• Carbon monoxide in nitrogen in alumi-
num cylinders at concentrations
between 5 ppm and 5,000 ppm.
NOTE: To ensure uncertainties of
less than 10 percent in measured
concentrations near 5 ppm, it may
be necessary to use the average of
two Protocol No. 2 analyses.
• Nitric oxide in nitrogen in aluminum
cylinders at concentrations be-
tween 10 ppm and 150 ppm.
• Sulfur dioxide in nitrogen in aluminum
cylinders at concentrations be-
tween 9 ppm and 375 ppm.
The data contained in this report,
indicates that CO/air and CO/N2 con-
tained in steel cylinders will remain stable
for no more than six months. Even though
this conclusion was based upon a rela-
tively small data set, it is recommended
that the present requirement for reanal-
ysis every six months remain unchanged
for the following:
• Carbon monoxide in air contained in
steel cylinders.
• Carbon monoxide in nitrogen contained
in steel cylinders.
Current recertifieation requirements
should remain unchanged for reactive
gases and cylinder types not considered
in this report.
. S. GOVERNMENT PRINTING OFFICE;1985/559-l 11/10750
-------
R. C. Shores andF. Smith are with Research Triangle Institute, Research Triangle
Park, NC 27709.
DarrylJ. von Lehmden is the EPA Project Officer (see belowj.
The complete report, entitled "Stability Evaluation of Sulfur Dioxide, Nitric Oxide
and Carbon Monoxide Gases in Cylinders," (Order No. PB 85-122 646; Cost:
$20.00, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, V'A 221'61
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
------- |