EPA-600 /8-90-053
June 1990
RECOMMENDED OPERATING PROCEDURE NO. 45s
ANALYSIS OF NITROUS OXIDE FROM
COMBUSTION SOURCES
EPA Project Officer; Judith S. Ford
Quality Assurance Manager
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
Air and Energy Engineering Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park NC 27711
-------
Disclaimer
This recommended operating procedure has been prepared for the sole use of
the Air and Energy Engineering Research Laboratory, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, and may not be
specifically applicable to the activities of other organizations.
ACKNOWLEDGMENTS
Assisting in the preparation of this procedure were Jeffrey V, Ryan and
Richard D. Rhinehart, Acurex Corporation, Research Triangle Park, NC, under
EPA Contract 68-02-4701 for on-site technical support to the Air and Energy
Engineering Research Laboratory (AEERL), Research Triangle Park, NC; and
Constance V. Wall, Research Triangle Institute, Research Triangle Park, NC,
under EPA Contract 68-02-4291 for Quality Assurance (QA) support to AEERL.
William P. Linak, a chemical engineer in the Combustion Research Branch of
AEERL, is the Work Assignment Manager for the nitrous oxide developmental
work performed by Acurex Corporation. Judith S. Ford, QA Manager for AEERL,
is the Project Officer for the QA contract with Research Triangle Institute.
ii
-------
CONTENTS
Section Page
Disclaimer 1!
Acknowledgments 11
g Ur6 S 1v
Tables 1v
1 Introduction 1
1.1 Scope ......a...1
1.2 Limitations 1
1.3 Definitions 1
2 Start-Up 3
2.1 Apparatus 3
2.1.1 Equipment Needed 3
2.1.2 Maintenance ........................................... 4
2.1.3 Theory of Gas Chromatography and Detection by
Electron Capture 4
2.2 Interferences 5
3 Operation 7
3.1 Samples/Sampling Procedure 7
3.2 Analytical System 7
3.3 GC Operation 10
3.3.1 Standards 10
3.3.2 Precalibration 10
4 Troubleshooting 12
4.1 Multipoint Calibration 12
4.1.1 Calibration for Grab Sample Analysis 12
4.1.2 Calibration for On-Line Analysis 13
4.2 linearity Check .. 14
4.3 Method Precision/Accuracy 18
5 Data Reductlon ..«.«««•.a...*......*..........*.*...IS
5.1 Calculations 19
5.2 Data Management 22
6 Quality Assurance/Quality Control...... 25
6.1 QC Checks 25
6.2 QC Controls 26
7 Bibliography 27
lit
-------
FIGURES
No. Page
1 Schematic of analytical system with precolumn backflush 9
2 ECD response * 0.5-200 ppnt range 15
3 ECD response; expanded view of lower range 16
4 Standard deviation for calibrations with P-5 and P-10 carriers... 20
5 First-order regression: HP calibration with P-5 carrier ........ 23
TABLES
No. Page
1 Results of first-order linear regression (log model) 21
2 Results of second-order (quadratic) linear regression
(log model) 24
iv
-------
SECTION 1
INTRODUCTION
1.1 SCOPE
This procedure Is used to determine the concentration of nitrous oxide
(N2O) in gaseous post-combustion process streams. Any gaseous stream
occurring after the combustion process may be considered a post-combustion
process stream. Gas streams in ducts, fluidized beds, and stacks or flues
are examples. Concentrations determined with this procedure are based on
real-time (on-line) analysis or delayed analysis over time (grab sample).
This procedure is applicable to gas streams where N2O concentrations
range from 0.5 to 200 parts per million (ppm). The upper and lower limits
can be extended by changing the sample loop volume; sensitivity limits are
determined by the minimum detectable concentration of the standards.
1.2 LIMITATIONS
Recommended operating procedures (ROPs) describe non-routine or
experimental research operations where some judgment in application may be
warranted. ROPs may not be applicable to activities conducted by other
research groups, and should not be used In place of standard operating
procedures. Use of ROPs must be accompanied by an understanding of their
purpose and scope. Questions should be directed to the preparers or
project personnel listed in the Acknowledgments.
1.3 DEFINITIONS
• Accuracy - The degree of agreement between an average measurement (X)
and an accepted reference or true value (T) expressed as a percentage
of the reference or true value; a data quality indicator. Accuracy
includes a combination of random and systematic error or bias
components which are due to sampling and analytical operations.
1
-------
• Calibrate - To determine, by measurement or comparison with another
standard, the correct value of each scale reading on a meter or other
device. The levels of the calibration standards should bracket the
range of planned measurements.
• Calibration Standard - A substance or reference material used to
calibrate an instrument.
• Method Detection Limit (HDL) - The concentration corresponding to five
times (5X) the background noise level of the measurement.
• Minimum Quantification Limit (MQL) - The concentration corresponding
to 10 times (10X) the background noise level of the measurement, or
the lowest calibration standard.
• Precision - The degree of variation among individual measurements of
the same property, usually obtained under similar conditions; a data
quality indicator. Precision is usually expressed as standard
deviation, variance, or range, in either absolute or relative terms.
• Baseline Drift - The change in instrument output over a stated time
period (usually 8 hours) of calibrated operation, when the initial
input concentration 1s zero; usually expressed as a percentage of the
full scale response.
2
-------
SECTION 2
START-UP
2.1 APPARATUS
2.1.1 Equipment Needed
When equipment is first received, refer to the operation manuals for
specific guidance in start-up and operating procedures for the gas
chromatograph and the detector.
Gas Chromatograph (GC): Temperature- and time-programmable; capable
of operating in a 50 to 200°C range; sample inlet equipped with dual mass
flow controllers and a 10-port switching valve.
Precolumn: 1.8 m x 0.32 cm (6 ft x 0.125 in.) O.D. stainless steel,
packed with 100/120 mesh support (e.g., Hayesep D by Alltech Associates,
Inc., Deerfield, IL).
Analytical Column: 3.7 m x 0.32 cm (12 ft x 0.125 in.) O.D. stainless
steel, packed with 80/100 mesh support (e.g., Porapak Super Q by Alltech
Associates, Inc., Deerfield, IL).
Detector: Electron capture with 63N1 constant current cell, capable
of operating at 330°C.
Carrier Gas: Argon/methane mixture specifically prepared for electron
capture detector (ECD) analysis; 90-95% Ar, 5-101 CH4.
Data Acquisition System: Computerized peak measurement system
tailored to gas chromatographic data.
3
-------
Certified Calibration Gases: Four compressed gas cylinders of
certified concentrations are required for calibrating the selected working
range. Suggested working ranges are 0.5-5, 5-40, and 40-200 ppm. Nitrogen
is the required balance gas for each cylinder.
Quality Control Standards: Three cylinders containing a known NgO
concentration which approximates the midpoint of each suggested working
range*
Moisture Removal System: Refrigeration dryer or similar apparatus
capable of condensing and removing moisture from sample stream. Desiccant
trap containing P2O5 to remove remaining moisture.
Gastlght Syringe: 5 cm3.
2.1.2 Maintenance
Gas Chromatograph: Replace oxygen traps and molecular sieve at least
every 6 months.
2.1.3 Theory of Gas Chromatography and Detection by Electron Capture
Gas chromatography (GC) refers to the separation technique where the
constituents of a gaseous mixture (mobile phase) are transported through a
heated column containing an adsorbing solid support (stationary phase). A
GC system will generally consist of a pressurized cylinder of carrier gas
equipped with a pressure regulator, sample injection system, analytical
column, detector, and data recording system. A sample of unknown
concentration is transported through the system by the carrier gas. To
eliminate reaction with the sample or any component of the system, the
carrier must be inert. The type of detection system used in the analysis
will dictate the choice of carrier.
Separation is based on the attraction each constituent has for the
stationary phase. A constituent's attraction for the stationary phase
4
-------
determines the rate at which 11 passes through the column. Constituents
having the least attraction will be adsorbed and eluted from the stationary
phase more rapidly than those with a greater attraction.
As constituents elute from the analytical column, the detector
produces an electrical signal proportional to the constituent
concentration. The electron capture detector (ECD) uses two electrodes to
produce this signal. One electrode is positively polarized and the other
is treated with 63Ni. The 63Ni ionizes the carrier gas as 1t passes
between the electrodes.
The 63Ni emits high-energy electrons as it decays. These electrons
produce large amounts of low-energy secondary electrons in the carrier gas
which are collected by the other electrode. Sample constituents which have
an attraction for secondary electrons cause a decrease in steady state
current that 1s proportional to constituent concentration.
2.2 INTERFERENCES
Elimination or control of possible interferences to an analytical
procedure is the responsibility of both sampling and analytical teams.
Although the analyst must demonstrate on a daily basis that the analytical
system 1s free from contaminants or residuals from previous analyses,
certain precautions must also be observed by those persons responsible for
obtaining the field sample. Known interferences to the measurement of N2O
in post-combustion gas streams include particulate matter, moisture, sulfur
dioxide (SO2), and halogenated organic compounds. Particulates must be
removed during the initial sample extraction and before the gas enters the
sample container (grab sample analysis) or the sample line (on-Hne
analysis). An in-line filter may be used for particulate removal.
Moisture and SO2 are two components of combustion flue gas which have
been shown to influence N2O generation 1n sample containers. Moisture Is
removed by pulling the sample through a non-react1ve drying system such as
5
-------
a refrigeration condenser or a desiccant trap containing a solid indicating
agent. A refrigeration dryer and a column of P2O5 are the recommended
drying systems. Use of these systems alone or in series will depend upon
the percentage of moisture in the combustion gas.
CAUTION
Because of possible reaction with nitrogen oxides, silica gel should not be
used as the Indicating agent.
Interference from SO2, residual moisture, and organics is controlled
by a precolumn backflush system. The principal theory in this process is
that, after the constituent of interest elutes from the precolumn, the
effects of the remaining constituents are controlled by flushing the column
with clean carrier gas. For example, the expected constituents in their
respective order of elution include O2, CO2, ^0, H2O, SO2, and some
halogenated compounds. Since N2O will come off the column before SO2,
backflushing the precolumn with carrier gas immediately following NgO
elution will protect the analytical system from the adverse effects of SO2.
The results of other studies have shown that halogenated compounds, such as
fluorinated hydrocarbons and chlorinated hydrocarbons, will contaminate the
ECD if they are not controlled. This contamination is usually seen as a
long-term baseline upset. Again, the backflush system would eliminate
organic interference because these compounds would typically come off the
column after N2O.
6
-------
SECTION 3
OPERATION
3.1 SAMPLES/SAMPLING PROCEDURE
A sampling procedure which reduces the moisture and SO2 levels 1n an
extracted sample Is under development. A detailed discussion of steps
involved Is beyond the scope of this procedure. The minimum sampling
requirement for the analytical procedure is that the sample must pass
through a moisture removal system before it reaches the analyzer or the
sample container.
3.2 ANALYTICAL SYSTEM
The gas chromatograph (GC) must be capable of operating between 50 and
200°C. Since this procedure uses a precolumn backflush to minimize the
effect of expected interferences, 1t may be considered a two-column
procedure Involving a 3.7 m x 0.32 cm (12 ft x 0.125 1n.) O.D. stainless
steel analytical column and a 1.8 m x 0.32 cm (6 ft x 0.125 in.) O.D.
stainless steel precolumn. The columns must be preconditioned at 100°C-
200°C before their Initial use and subjected to an overnight conditioning
at 75°C before each day's analysis. During analysis, the oven temperature
should be 50°C.
The electron capture detector (ECD) with a 63Ni constant current cell
is operated at 300-350°C. Although the use of nickel as the energy source
allows for operation at up to 400°C, it is recommended that the detector
temperature be maintained in the indicated range since detector life 1s
reduced at higher temperatures. The carrier gas, a mixture specially
prepared for ECD analysis, contains argon and methane. The component ratio
is 90-95% argon and 5-10% methane. The carrier gas Initially flows through
a 5 angstrom (A) molecular sieve and an oxygen scrubber before 1t passes
through the analytical system.
7
-------
A schematic of the analytical system, including the precolumn
backflush, is shown in Figure 1. A 10-port valve is used to direct the
sample and the carrier gas, as one stream or as separate streams, through
the analytical system. When the valve is in the "off" position, as seen in
Figure 1(a), a sample coming into the 1 cm3 fixed volume sample loop is
circulated through the loop. At the same time, two streams of the carrier
gas are circulated through the precolumn and the analytical column. The
sample loop circulation is controlled by valve-ports 1, 2, 3,and 4. Flow
for the precolumn backflush is controlled by ports 6, 8, 9, and 10. The
gases exiting the sample loop and the precolumn are directed to vent.
Ports 5 and 7 are used to direct the carrier flow to the analytical column.
Gas exiting the analytical column is passed to the detector.
In the "on" position (Figure lb), the bridges connecting the ports are
realigned so that (1) the carrier gas stream flowing to the analytical
column and the sample (from the sample loop) are combined to make up the
mobile phase which passes through the precolumn, (2) the gas exiting the
precolumn is directed to the analytical column and the detector, and (3)
the second carrier gas stream (backflush) flows to vent. ECD response to
calibration standards and samples is translated by a digital integrator.
The analysis of field samples will basically follow the steps outlined
for instrument calibration in Section 4. The one exception is that, during
analysis, the sample is initially injected onto the precolumn instead of
the analytical column. After instrument calibration, the following steps
are used to analyze grab samples:
A. Adapt a septa fitting to the inlet of the sample loop.
B. Flush a 5 cm3 syringe twice by f1111ng the syringe with a sample
of combustion gas.
C. Inject a third 5 cm3 sample into the sample loop.
D. At 0.1 minute, activate the 10-port valve so that the sample is
loaded onto the precolumn. N2O will elute to the analytical
column in ~3.5 minutes.
E. At 3.5 minutes, turn the valve to the "off" position so that the
precolumn is backflushed with carrier gas.
-------
f
FROM
CARRIER
SOURCE
la) Gas Sample with Backflush: Valve Off
Figure 1. Schematic of analytical system with precoluran backflush.
9
-------
F. N2O will elute from the analytical column at "6,5 minutes. The
total run time for analysis is ~8 minutes.
G. Introduce another 5 cm3 sample into the sample loop. Repeat
steps C-F until three data points (peak areas) are obtained for
the sample.
H. Calculate the average peak area. Calculate the N2O concentration
according to equation (4) 1n section 4.2.
For on-Hne analysis, a stainless steel tee is used to connect the GC
Inlet to the main sample line from the combustion source. The sample flows
from the source, through the sample conditioning system (moisture removal),
and then directly to the GC inlet. Using this method of direct injection,
instead of syringe injection, on-line analysis is accomplished by following
the preceding steps D-H for analysis of grab samples.
3.3 GC OPERATION
3.3.1 Standards
A. Four compressed gas cylinders of certified concentrations are
required for calibrating the selected working range. Suggested
working ranges are 0.5-5, 5-40, and 40-200 ppm. Nitrogen is the
required balance gas for each cylinder. Outfit each cylinder
with appropriate pressure regulators. These gases will be used
to conduct the multipoint calibration.
B. Three commercially prepared compressed gas cylinders certified to
contain N2O concentrations approximately equal to the midpoint
of each suggested working range are required. These gases will
be used to conduct quality control (QC) checks during analysis.
C. Verify the stability of the certified concentration of each
cylinder of gas.
3.3.2 Precalibration
A. With the GC power off, open the main valve of the carrier gas
cylinder and set the pressure to 482.36 kPa (70 ps1g).
B. Adjust the carrier gas flow to approximately 30 cm3/m1n.
C. Turn the power on.
10
-------
Set the oven temperature to 75°C, and allow the temperature to
stabilize.
Condition the analytical column and the precolumn overnight at
75°C. Both columns should be conditioned overnight prior to each
day's analysis.
11
-------
SECTION 4
TROUBLESHOOTING
4.1 MULTIPOINT CALIBRATION
The multipoint calibration must be performed for initial start-ups if
the analytical system is subjected to a full shutdown, or if the carrier
gas cylinder is changed. The calibration must complement the type of
sample to be analyzed. If grab samples are analyzed, the calibration must
be by syringe injection. For on-line analysis, calibration must be
accomplished through direct injection from compressed gas cylinders.
Calibration data points are generated with standards of known
concentration. The calibration concentrations must bracket the expected
sample concentrations. The following procedures must be conducted prior to
analysis of samples.
CAUTION
Do not connect pressurized gas cylinders directly to the gas chromatograph.
Use a sample valve to avoid damaging the Instrument.
4,1.1 Calibration for Grab Sample Analysis
A. Remove the end cap from the ECD inlet and connect the detector to
the GC.
B. Set the detector temperature to approximately 20°C below the
maximum operating temperature (e.g., 330°C 1f maximum temperature
is 350°C).
C. Allow the analytical column and the precolumn to cool to 50°C.
D. Access the data acquisition system.
E. Verify the stability of the column temperatures.
F. Connect a septa to the outlet of the sample line that 1s attached
to the pressure regulator of the gas cylinder containing the
lowest concentration calibration standard, or transfer a portion
of the standard to a clean sample container (bomb) equipped with a
septa. To transfer the standard:
12
-------
1. Purge a bomb with the lowest concentration of standard for
sufficient time to allow at least 10 liters of gas to flow
through the container.
2. Close the valve on the bomb inlet, turn off the cylinder
at the regulator outlet, and allow the sample container to
equilibrate to ambient pressure.
3. Close the valve on the bomb outlet.
4. Repeat steps 1-3 with the remaining standards.
G. Using a 5 cm3 gastight syringe, extract a sample of the lowest
concentration calibration standard. Introduce the sample into
the GC inlet. Repeat the injection to ensure that dead space in
the syringe tip does not dilute the sample. Allow the sample to
reach equilibrium 1n the sample loop.
H. Inject the sample into the sample,loop. Set the GC to run, and
load the sample onto the analytical column,
I. After approximately 3.5 minutes, return the GC to the load
(backflush) position.
J. NgO will elute from the column in approximately 6.5 minutes.
Measure the ECD response (as integrated peak area) to the
calibration standard.
K. Repeat steps G through J until three data points are obtained.
L. Calculate the average integrated peak area (PA).
M. Repeat steps G through L for the remaining standards.
N. To derive the calibration curve, follow steps A-D outlined 1n
section 4.2.
4.1.2 Calibration for On-Line Analysis
A. Follow steps A through E in the procedure for grab sample
analysis (section 4.1.1).
B. Adjust the regulator for the cylinder containing the lowest
concentration of N2O to 34.47 kPa (5 psig).
C. Using a Teflon sample line, connect the cylinder to the GC at the
sample loop inlet.
D. Open the regulator valve and allow the sample loop to fill. Close
the valve, and allow the sample to reach equilibrium. Load the
sample onto the analytical column.
13
-------
E. Measure the ECO response (as integrated peak area) to the
calibration standard. Repeat the measurement and obtain three
data points.
F. Calculate the average integrated peak area (PA).
G. Repeat steps B through F for the remaining standards.
H. To derive the calibration curve, follow steps A-D in
section 4.2.
4.2 LINEARITY CHECK
Figure 2 shows that the ECD output is a non-linear function of
concentration. The relationship between the two variables, concentration
and average peak area, exhibits a bow at the lower end of the curve.
Figure 3 shows an expanded view of the lower end of the curve. To minimize
and control relative error, it is recommended that a log transformation be
performed on the original variables. Refer to section 5.1 for a detailed
discussion of transformed curves. The transformed data are used to derive
a least-squares linear regression. The equation for the curve would be of
the form;
ln(y) = 1n(a) + b[ln(x)] (1)
where: In(a) = the intercept of the calibration curve, and
b = the slope of the calibration curve.
A. Calculate a linear regression curve using the logarithm of each
data point with y = In (average peak area) and
x = In (concentration).
B. Verify the acceptability of each calibration curve by back-
calculating the concentration of each calibration standard.
Apply the log of the average peak area (y) to the regression,
where:
ln(x) « Tn(y) - ln(a) (2)
b
14
-------
«
4J
G
9
O
U
« n
Q) c
u o
m «h
-------
120
110
100
90
80
70
60
50
40
30
20
10
0
~
Figure 3. ECD response: expanded view of lower range.
-------
(4)
Calculate the percent difference (D) between the true N?0
concentration and the concentration derived from the calibration
curve through back-calculation, where:
If |Dj i 10% for each standard, the calibration is acceptable.
Repeat the calibration if |D| is >101.
Verify the relationship between the calibration curve and the three
QC standards by introducing the standards to the GC in same manner
that the instrument was calibrated.
Obtain duplicate data points for each standard and calculate the
average peak area.
Calculate the concentration of each QC standard by applying the
average peak area to the equations given in step B.
Use equation 5 from step C to calculate the percent difference
between the true concentration and the concentration derived from
the calibration curve. The |D| must be 10%.
Use the QC standards to verify the integrity of the calibration
after 1 hour of analysis during an on-line routine and after every
third sample when analyzing grab samples.
Verify the effects of the sample conditioning system by introducing
the QC standards at the combustion source.
1, Disconnect the sample line at the combustion source outlet
from the Inlet of the sample conditioning system.
2. Connect the QC cylinder to the conditioning system. Push the
sample through all components of the system as if it were a
sample originating from the combustion source.
Repeat the calculation in steps F through H. If ]D| is *15%, the
sample conditioning system has not compromised the N2O
concentration.
D = [(^Ocurve ~ ^2^true)/^Otrue] x
(5)
17
-------
NOTE
If N2O concentration ts compromised by the sample conditioning system,
the source of contamination must be eliminated. When the contamination
problem cannot be resolved, the multipoint calibration must be conducted
with all standards passing through the conditioning system.
4.3 METHOD PRECISION/ACCURACY
Data accuracy 1s primarily a function of instrument calibration.
Obtain a measure of method accuracy from performance audits and quality
control checks (section 6.1). If the analytical system is out of control
(error exceeds ±15% of the true concentration) and problems cannot be
resolved through corrective action, recalibrate the analyzer. Data taken
between the previous calibration and the corrective calibration should be
considered suspect.
Precision is a function of analyzer stability. Obtain estimates of
precision from the analytical system response to dally span and QC checks,
using equation 7 (section 6.1). The error due to analyzed drift should not
exceed 15% of the true value of the QC standard for the selected range.
18
-------
SECTION 5
DATA REDUCTION
5.1 CALCULATIONS
The response of the ECD is a non-linear function of concentration. To
evaluate the best fit of the data, full-scale calibrations were conducted
on two analyzers manufactured by Hewlett Packard (HP) and Shimadzu (S).
Two types of carrier gas, P-5 and P-10, were used with each analyzer. A
review of the data indicates that the precision of triplicate data points
is an Increasing function of concentration. Figure 4 shows that standard
deviation is approximately proportional to concentration. This suggests
that a log transformation may be appropriate in order to have variability
constant over the range of interest.
The usual approach to calibration, a least-squares regression on the
original scales (ppm and peak area), would be appropriate If the objective
is to minimize error on an absolute basis and if variability were
independent of concentration. In such a case, error could be reported as
less than a certain concentration, which could also be reported as a
percentage of full scale. A problem with this type of error control and
estimate is that the estimated error for low concentrations will be very
large relative to the measured or true concentration.
By performing a linear regression in terms of log units (i.e.,
transforming the data), the operator would be able to control and report
error on a relative basis. Relative error improves as transformed data are
applied to higher order regressions. Beyond the second order (quadratic),
however, the improvement is not statistically significant.
During field studies, N2O concentrations will be determined through a
first-order regression. An example calculation using the HP/P-5
calibration follows Table 1. For final reporting, the results may be
compared to a second-order regression.
19
-------
S - Shlmadzu
HP - Hewlett Packard
~
A
¦©>
O
& +
o
0 20
~
EP
~
A
J L_
O
X
_L
J L
120
S. P—5
S, P-10
80 100
*2° (ppm)
O HP, P—5
140
160
180
200
HP, P-10
Figure 4, Standard deviation for calibrations with P-5 and P-10 carriers.
-------
TABLE 1. RESULTS OF FIRST-ORDER LINEAR REGRESSION (LOG MODEL)
Instrument
Error (%)
Carrier
In (a)
b
1-5 ppm
5-40 ppm
40-200 ppm
HP/P-5
9.5328
0.8828
4.3
4.2
2.7
HP/P-10
9.6289
0.8747
3,7
4.2
2.6
S/P-5
9.2185
0.9195
4.1
4.7
3.0
S/P-10
9.2123
0.9222
4.4
4.7
3.0
Equation:
ln(y) - In (a) + b[ln(x)]
)nW , '"M - '"(*)
x ¦ RH
Example Calculation
True concentration = 0.514 ppm
Average response = 7874 area units
7874
9.5328
Vo.
8828
Error =
x = (0.5704)1-1328
x = 0.529
0.529 - 0.514
0.514
x 100
Error =3.0%
21
-------
The regression coefficients and the magnitudes of the relative error
(average) from the HP and Shlmadzu calibrations are given In Table 1.
Figure 5 shows the f1t of the first-order regression for the HP with a P-5
carrier gas. Table 2 shows the regression coefficients and average
relative errors for a second-order regression.
DATA MANAGEMENT
A. Validation - Recheck all data prior to reporting. Validation must be
conducted by the project or test leader.
B. Reporting - Report all data on standard forms signed by the test
operator and data validator. Include estimates of accuracy (percent
error) and precision (standard deviation) based on daily calibration
and QC checks and the results of performance audits. Report the
number of data points used to calculate the precision and accuracy.
Also include an explanation of any missing or suspect data (i.e.,
outliers).
22
-------
15
14
13
12
11
10
9
8
7
13
jef
X
2
ln(N20 ppra)
Figure 5. First-order regression: HP calibration with P-5 carrier.
-------
t
TABLE 2. RESULTS OF SECOND-ORDER (QUADRATIC) LINEAR REGRESSION (LOG MODEL)
Instrument
Error (%)
Carrier
1n(a)
b
c i
-5 ppm
5-40 ppm
40-200 ppm
HP/P-5
9.5144
0.9369
-0.0111
3.0
0.5
1.0
HP/P-10
9.6112
0.9264
-0.0110
1.2
0.6
2.0
S/P-5
9.1979
0.9801
-0.0129
1.7
0.8
S/P-10
9.1912
0.9842
-0.01318
2.0
0.6
1.7
Equation: 1n(y) = ln(a) + b[ln(x)] + c f [1 n (x) ]2|
24
-------
SECTION 6
QUALITY ASSURANCE/QUALITY CONTROL
6.1 QC CHECKS
During a series of tests, analyze a QC sample from a cylinder
containing a verified concentration that is approximately 50% of the full-
scale of the working range. The QC check should be conducted every hour
during on-line analysis and after every third sample (I.e., after nine
measurements for triplicate analysis} during grab sample analysis.
Calculate the cylinder concentration from the Instrument response and the
appropriate calibration curve. Compare the calculated value to the
cylinder's true value and report the error as a percentage of the true
value, according to:
N2°calc. " N2°true x ioo (6)
terror
N2°true
Errors greater than ±15% are unacceptable and indicate the need for
corrective action. Gas cylinders containing less than 689.25 kPa
(100 psig) must not be used as calibration or QC standards.
The triplicate and duplicate data points generated by the calibration
standards and QC samples should be used to calculate precision by:
_n
n ^ x
1=1
1
n
1 x.
Li=1 1
where: n = number of data points
Xi = individual data points.
/n(n-l)
(?)
25
-------
6.2 QC CONTROLS
A. Documentation
1. Laboratory Notebook - Keep all test results, calibration data,
and quality control data in a bound laboratory notebook. Sign
and date the notebook at the time of data entry.
2. Instrument Notebook - Assign a logbook to the GC. Use the
logbook to maintain a record of all calibrations, maintenance,
and repairs.
3. Certlf1cat1on/Ver1f1cation - In a secure area, maintain a file of
manufacturer's certifications and laboratory verification of
standards.
4. Instrument Manuals - Keep operator's manuals for all components
of the analytical system available and easily accessible.
B. Raw Data - Maintain all measurement data (storage diskettes,
printouts) on file 1n a secure area.
C. Internal/External Audits - Audit the GC performance by challenging the
system with verified concentrations of N2O unknown to the
system operator. If the system fails the audit with error greater
than ±15%, it Is out of control; take corrective action. Maintain a
file of performance audit results and corrective actions.
26
-------
SECTION 7
BIBLIOGRAPHY
Cicerone, R.J., et al. Atmospheric N2O: Measurements to Determine Its Sources,
Sinks, and Variations, Journal of Geophysical Research, Vol. 83, No. C6, 3042-
3050, June 20, 1978.
Grob, R.L, Chromatographic Analysis of the Environment. New York: Marcel
Dekker, Inc., 1983.
Lanier, W.L. and S.B. Robinson. EPA Workshop on N2O Emission from Combustion.
EPA-600/8-86-G35 (NTIS PB87-113742).September 1986.
U.S. Environmental Protection Agency. Quality Assurance Handbook for Air
Pollution Measurement Systems; Volume III, Stationary Source Specific Methods.
EPA-600/4-77-Q27b (NTIS PB80-112303), 1977.
Willard, H.H., et al. Instrumental Methods of Analysis. California: Wadworth
Publishing Company, 1981.
27
-------
TECHNICAL REPORT DATA
(Please tend Insltnctions on the reverse before comp!
1. REPORT NO. 2.
EPA-600/8-90-0 53
: PB90-2385A2 j
v_ J
4. TITLE AND SUBTITLE
Recommended Operating Procedure No. 45; Analysis
of Nitrous Oxide from Combustion Sources
5. REPORT DATE
June 1990
6. PERFORMING ORGANIZATION COPE
7. AUTHOR(S)
Judith S. Ford
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
See Block 12
10, PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-4701 (Acurex), and
68-02-4291 (RTI)
12. SPONSORING agency name and address
EPA, Office of Research and Development
Air and Energy Engineering Research Laboratory
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT ANC PERIOD COVEREO
Final; 9/89-5/90
14. SPONSORING AGENCY CODE
EPA/600/13
is.supplementary notes AEERL project officer is Judith S. Ford, Mail Drop 49, 919/541-
ft Vvv«
16-^~I"n*c-T*'Tbe recommended operating procedure (HOP) has been prepared for use in
research activities conducted by EPA's Air and Energy Engineering Research Labora-
tory (AEERL). The procedure applies to the measurement of nitrous oxide (N20) in
dry gas samples extracted from gas streams where N20 concentrations range from
0. 5 to 200 pprn. N20 concentrations are interpreted by an electron capture detector
(ECD). The EC'.D uses a 63Ni constant current cell. The upper limits of this procedure
can be extended by changing the sample loop pressure to reduce the volume of sample
in the look or by diluting the sample, ROPs describe non- routine or experimental
research operations where some judgment in application may be warranted. ROPs
may not be applicable to activities conducted by other research groups, and should
not be used in place of standard operating procedures. Use of ROPs must be accom-
panied by an understanding of their purpose and scope. <2==*
17, KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. cosati Field/Group
Pollution
Analyzing
Nitrogen Oxide (N20)
Combustion
Pollution Control
Stationary Sources
Operating Procedures
13B
14 B
07B
21B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF.PAGES
/ 32
20. SECURITY CLASS (This page)
Unclassified
22, PRICE
fcPA Form 2220-1 (9-73)
1
------- |