EPA-600/4-75-010
September 1975
Environmental Monitoring Series
HYDROCARBON MEASUREMENT DISCREPANCIES
AMONG VARIOUS ANALYZERS USING
FLAME-IONIZATION DETECTORS
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, N.C. 27711
-------
RESEARCH REPORTING SERIES'
Research reports of the Office of Research and Development,
U.S. Environmental Protection A'^ncy, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
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 quantification of environmental pollutants at the lowest
conceivably significant concentrations. It also includes studies
to determine the ambient concentrations of pollutants in the
environment and/or the variance of pollutants as a function of
time or meteorological factors.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-600/4-75-010
September i97b
HYDROCARBON MEASUREMENT DISCREPANCIES
AMONG VARIOUS ANALYZERS
USING FLAME-IONIZATION DETECTORS
by
Frank F. McElroy and Vinson L. Thompson
Environmental Monitoring Branch
Environmental Monitoring and Support Laboratory
Research Triangle Park, North Carolina 27711
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and
Support Laboratory, U.S. Environmental Protection Agency, and approved
for publication. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.
-------
TABLE OF CONTENTS
Section Page
BACKGROUND 1
ANALYZER STUDY 3
Experimental Design 4
Analysis of Results 8
SUMMARY AND CONCLUSIONS 12
REFERENCES 14
APPENDIX A: MEMO FROM G. C. ORTMAN 15
APPENDIX B: STATISTICAL ANALYSIS SYSTEM 1?
TECHNICAL REPORT DATA AND ABSTRACT 23
LIST OF TABLES
Table Page
1. Artificial Atmosphere Test 6
2. Statistical Analysis of Discrepancies between Pseudo-NMHC
Concentrations from Various Pairs of FID Analyzers .... 10
3. Cumulative Frequency Distribution of Discrepancies .... 11
4. 6 to 9 a.m. NMHC Concentrations 13
-------
HYDROCARBON MEASUREMENT DISCREPANCIES AMONG VARIOUS
ANALYZERS USING FLAME-IONIZATION DETECTORS
BACKGROUND
Efforts to measure hydrocarbon concentrations in the atmosphere
date back to the early 1950's or earlier. Continuous, routine moni-
toring of ambient total hydrocarbon concentrations by air pollution
control agencies came into common practice in the early 1960's, when
commercial total hydrocarbon analyzers became available. Virtually all
of these later analyzers employed a flame-ionization detector (FID)
wherein hydrocarbon compounds were measured by detecting their ioniza-
tion in a hydrogen flame. The response of these FID analyzers was
related not only to the concentration of the hydrocarbon being measured,
but also to the "effective carbon number" of the hydrocarbon compound.
The effective carbon number varied depending on the number of carbon
atoms in the molecule and on the type of compound (e.g., aliphatic,
aromatic, olefinic, acetylenic, etc.). Thus, without knowing which
hydrocarbon compound was being measured, the reading could not be
related to the actual concentration. Some FID technologists further
pointed out that there may also be still other problems that could cause
discrepancies in response from one flame-ionization detector to another.
But these problems were mitigated in the 1960's for several reasons:
(1) the ambient air could contain a great variety and number of indi-
vidual hydrocarbon compounds, (2) there was incomplete knowledge of the
relative importance, as pollutants, of the various hydrocarbons that
might occur in ambient air, (3) a large fraction of the total hydro-
carbons in the air was methane, which was measured relatively well by
the flame ionization analyzers (FIA), (4) no hydrocarbon standards
existed, and (5) no satisfactory alternative methodologies appeared to
be available. Over the years, the measurement of total hydrocarbons--
and later "hydrocarbons corrected for methane"--came to be defined in
terms of the response of the FID.
Interest in hydrocarbon measurements waxed as more was learned
about the role of hydrocarbons in photochemical reactions with other
pollutants, and particularly in the formation of ozone. When it
appeared that ambient methane did not taken an active part in these
1
-------
reactions, attention was focused on nonmethane hydrocarbons (NMHC),
sometimes referred to as "reactive hydrocarbons." Methods were devel-
oped to continuously measure ambient methane concentrations and these
measurements could than be subtracted from concurrent total hydrocarbon
(THC) measurements to obtain NMHC. A National Ambient Air Quality
2
Standard for hydrocarbons (actually NMHC) was promulgated. Although
this standard was "for use as a guide," it was considerably lower than
many of the NMHC concentrations that were being observed in urban
ambient air. The frequent exceeding of the standard started serious
questioning of the credibility of both the standard and of the measurem-
ent methodology.
Development of methods to measure methane and NMHC was relatively
slow. The definition of NMHC in terms of the FID and the impracti-
cability of attempting to measure individual hydrocarbons virtually
ruled out methodology not based on the FID. Early systems employing
columns of activated carbon were fraught with problems and often pro-
3
duced numerous negative readings. When it was demonstrated that gas
chromatography could be used effectively to separate and measure methane
accurately, a chromatographic method was designated as the reference
2
method. But while methane could be measured chromatographically, total
hydrocarbons could not. And although commercially available gas chro-
ma tographic-type instruments provided measurements of both methane and
total hydrocarbons (to yield a NMHC measurement) only the methane
measurement was actually made by chromatographic separation. The THC
measurement still had to be made more or less directly by the instru-
ment's FID without benefit of the gas chromatographic process. Thus
NMHC measurements were somewhat better, but measurement errors were
still large with respect to the ambient air quality standard for NMHC.
Moreover, commercial gas chromatographic analyzers for NMHC proved to be
complex, expensive, and sometimes unreliable. They also required high
levels of operator skill, careful maintenance, and high purity of
several support gases usually needed for operation. Since there was no
requirement to monitor NMHC, states were not encouraged to purchase NMHC
monitors for use in their State Implementation Plan (SIP) networks.
Recent advances in hydrocarbon measurement technology have resulted
in somewhat more accurate hydrocarbon monitors. Gas chromatographic-
2
-------
type instruments have become less complex, less costly, and more reliable.
New types of instruments use non-gas-chromatographic methods for sepa-
rating methane from total hydrocarbons to effect the NMHC measurement.
Many of these new analyzers are easier to operate and have improved
stability.
But problems remain. To date, no method has yet been developed to
remove methane from the THC in order to measure the NMHC concentration
directly. Thus, all present methods require the measurement of total
hydrocarbon and methane concentrations separately, determining the NMHC
concentration by subtraction. This necessitates use of scale ranges for
both THC and methane of at least 10 parts per million (ppm) to avoid
off-scale readings in typical urban air. Normal calibration and measure-
ment errors are increased by the necessity of making two measurements.
These errors are large with respect to typical NMHC concentrations,
which generally are below 2 or 3 ppm. Such errors can be very large
4
with respect to the 0.24 ppm NMHC standard.
Still another problem is reported by several experimenters who have
observed substantially discrepant NMHC measurements from various types
of hydrocarbon analyzers measuring identical samples of ambient air
(Appendix A). This phenomenon is apparently due to the variation in
response to different hydrocarbon substances among various FID's.
Again, this problem is relatively small when measuring THC because of
the large fraction of methane, but can become far more significant in
the NMHC measurements. Moreover, the magnitude of the discrepancies
from one FID to another changes as the proportion of various hydro-
carbons in the ambient air changes from hour to hour.
ANALYZER STUDY
This last problem of inconsistency from one FID to another appeared
to be inherent in the FID technology, and the reported discrepancies
from analyzer to analyzer were considerable. Such discrepancies or
variations would have a possibly serious impact on the NMHC reference
method and on any attempts to determine whether alternate methods are
equivalent to the reference method. To further and formally investi-
gate this problem, an experimental laboratory study was designed
wherein a number of FIA's could be observed under controlled conditions
3
-------
while making simultaneous measurements of urban ambient air. The
objectives of this experiment were (1) to confirm the existence of this
discrepancy problem under carefully controlled conditions of operation
and calibration, (2) to determine the magnitude and significance of the
problem with respect to NMHC measurements in quasi-typical urban ambient
air, and (3) to attempt to better define the nature and possible causes
of the problem. The scope of the experiment was necessarily quite
limited because of limitations on time and available analyzers.
Experimental Design
The experiment was conducted at the Durham Air Monitoring and
Demonstration Facility (DAMDF) in Durham, North Carolina, a location
that can be considered to have more or less typical urban air. Measure-
ments were collected in late October and early November, during which
time total hydrocarbon concentrations varied from very low (less than
2.0 ppm) to high (up to 8 ppm).
The objectives of the experiment implied that NMHC measurements of
identical samples of urban ambient air be collected and compared for a
variety of FIA's. However, for several important reasons to be explained
shortly, total hydrocarbon measurements were collected from the analyzers
and pseudo-NMHC values were analyzed and compared.
Discrepancies among simultaneous NMHC measurements from various
analyzers could arise from a number of sources. For hydrocarbon analyz-
ers the significant sources are: (1) differences in calibration, (2)
differences in calibration species (e.g., methane vs. ethane or propane),
(3) presence of unknown concentrations of higher hydrocarbons in cali-
bration standards, (4) differences in scale linearity, (5) differences in
response times, (6) differences in drift and stability, (7) normal
analyzer precision and repeatability errors, (8) differences in effi-
ciency of separating methane from THC, and (9) differences in response
to various higher hydrocarbon compounds and other organics. For the
results of the experiment to be meaningful, this list of variables was
reduced, as follows: Differences in calibration and calibration species
were eliminated by calibrating all test analyzers simultaneously with
the same calibration standard (supplied from a high-pressure cylinder).
Methane was used because it was least likely to cause differences in FID
response, was highly stable, and could be used to (simultaneously) cali-
4
-------
brate both THC and methane responses. The standard used was methane in
air with less than 0.1 ppm other hydrocarbons. Although the lack of
higher hydrocarbons could not be absolutely confirmed, use of the same
standard to calibrate all analyzers during the entire study would mini-
mize this possibility as a source of discrepancy. Each of the analyzers
was checked for linearity (with methane) to eliminate nonlinearity as a
source of discrepancy; each analyzer proved to be linear within the
normal calibration accuracy limits. Differences in response times were
excluded as a source of discrepancy by using relatively long averaging
times for the measurements -- all reported values were visually-deter-
mined one-hour averages (or half-hour averages in Table 1).
Discrepancies caused by differences in analyzer drift and stability
were minimized (but certainly not eliminated) by careful daily zero and
span adjustments to each analyzer, and by correcting the measurements
when significant zero drift did occur. Variation due to analyzer
precision and repeatability errors is inherent in the analyzer and could
not be completely controlled.
The remaining two sources of discrepancy -- differences in the
efficiency with which an analyzer separates methane from THC and differ-
ences in analyzer response to various higher hydrocarbon (NMHC) or other
compounds were of primary interest because these sources, more than
any others, are beyond the ability of the analyzer operator to control
or correct. Limitations in the scope of the study and in the analyzers
available permitted investigation of only the latter phenomenon --
difference in analyzer response. Accordingly, differences in methane
separation efficiency were eliminated from the experiment by using only
THC measurements from each of the analyzers. Concurrent methane measure-
ments from one of the analyzers were used to calculate pseudo-NMHC
concentrations for each analyzer. Using the same methane value for each
analyzer, it is obvious that any discrepancies in the resulting NMHC
measurement must be due solely to discrepancies in the THC measurement.
However, the discrepancies are thereby put in perspective relative to
NMHC measurements, not THC measurements; innocent-looking discrepancies
in the THC measurements may become significant discrepancies with
respect to the smaller NMHC measurements. This scheme also permitted
the use of methane for the simultaneous calibration of all analyzers,
-------
Table 1. ARTIFICIAL ATMOSPHERE TEST
Pollutant
CH4 (zero)
CH4 (span)
C2H6/CH4
C2H4/C2H2
Approximate Equivalent CH4
concentration, concentration, Bendix
ppm ppm 8201
0.15 0.15 0.15
8.05 8.05 8.05
1.97/0.9 4.84 5.1
5.1
5.0
5.0
4.97
4.97
5.0
1.9/1.9 7.6 6.4
6.4
6.4
6.4
6.4
6.4
6.4
1/2-hour
MSA
11-2
0.15
8.05
5.9
5.9
5.8
5.8
5.77
5.77
5.77
4.85
4.85
4.9
4.9
4.8
4.8
4.8
THC averages, ppm
Beck.
400-1
0.15
8.05
5.7
5.7
5.65
5.65
5.65
5.65
5.6
7.35
7.35
7.35
7.35
7.4
7.4
7.4
Beck.
400-2
0.15
8.05
5.8
5.8
5.6
5.6
5.75
5.75
5.7
7.35
7.35
7.35
7.35
7.5
7.55
7.55
Beck.
6800QC
0.15
8.05
6.3
6.3
6.2
6.2
6.22
6.22
6.1
7.3
7.3
7.3
7.3
7.45
7.45
7.45
-------
and allowed the use of THC analyzers in the experiment. The latter
was particularly important because two identical THC analyzers were
available for the study, making it possible to compare discrepancies
between like-model analyzers with those between different models.
The analyzers used for the experiment consisted of:
1. One Beckman Model 6800 Gas Chromatograph.
2. One MSA Model 11-2 Hydrocarbon Analyzer.
3. One Bendix Model 8201 Hydrocarbon Analyzer.
4. Two Beckman Model 400 Total Hydrocarbon Analyzers.
These analyzers all employed an FID detector and represented a variety
of detector designs and operating characteristics. The Beckman 6800 is
a gas chromatographic instrument presumably representative of the existing
reference method for NMHC, and provides 12 batch-analyses per hour. The
Bendix 8201 is also a batch-analysis type analyzer using gas ehromato-
graphic-colume-type separation of methane and providing 20 analyses per
hour. The MSA 11-2 is a dual-FID continuous NMHC instrument using
catalytic oxidation to separate methane. The Beckman 400's are simple,
continuous, FID type THC analyzers having no capability to separate
methane.
All analyzers were operated at the DAMDF where they were connected
to a common sample air distribution manifold and measured ambient air
during the field study. No artificial pollutant augmentation was used,
although a short side-experiment with artificial HC concentration was
also conducted. Data were recorded by strip chart recorders operating
at 2.54 cm per hour (1 inch per hour) and were averaged and tabulated
manually. Random checks on averaging and tabulation accuracy were made
periodically by an alternate operator. The analyzers were zeroed and
spanned daily, and standard linear interpolation corrections were made
when significant zero drift was observed.
The Beckman 400 analyzers were modified for the study by replacing
their bypass-type sample pressure regulators with dead-end-type regu-
lators with a small constant-bypass flow. This modification was needed
to reduce calibration gas consumption so that the entire experiment
could be completed on a single cylinder of span gas. The instruments
were later restored to their original condition to verify that the
modification caused no significant change in their performance. The
7
-------
Beckman 6800 operated erratically during part of the study and its
data were not used until the last few days of the experiment. The
MSA 11-2 did not have a THC output, but by operating it with the
methane-measuring FID off, the NMHC output yielded THC measurements.
The Bendix 8201 was used to provide the methane measurements.
Most FID analyzers, and all those used in this study, operate
with a sample pump ahead of the FID to provide sample air to the
detector under positive pressure. During calibration, the calibration
standard gas is normally inserted under positive pressure between the
pump and the analyzer. This assumes that the diaphram-type pump used
does not change the hydrocarbon concentration of the sample air. This
technique is commonly used and was used during the experiment. After-
wards, however, calibration gas was introduced into the normal sample
air intake ahead of the pump using a vent system to prevent positive
pressure. Unfortunately, differences were detected on two of the
analyzers (the 6800 and the 11-2). No explanation for this phenomenon
was determined. Appropriate corrections had to be made to the data
from the affected analyzers (about -8 percent correction for the 11-2
and about -5 percent for the 6800).
Analysis p_f_ Results
There is no doubt that differences in response to various hydro-
carbons exist among different FIA's. This phenomenon was indicated by
the short side-experiment wherein two different artificial mixtures of
hydrocarbons were measured by all five analyzers. Table 1 shows the
results. When a mixture containing 1.97 ppm ethane (CoHg) and 0.9 ppm
methane (CH4),* equivalent to 4.84 ppm-carbon (1.97 x 2 + 0.9 = 4.84),
was measured, the readings ranged from 4.97 ppm to 6.3 ppm. The two
identical Beckman 400's agreed within 0.1 ppm. A mixture containing
1.9 ppm ethylene (C2H4) and 1.9 ppm acetylene (C2H2),* equivalent to
7.6 ppm-carbon, produced measurements ranging from 4.8 to 7.55 ppm --
a maximum discrepancy of 2.75 ppm! Again the two 400's agreed
reasonably well. Interestingly, the analyzers producing the highest
and lowest readings are different for the two mixtures.
*This was the supplier's analysis and was not verified.
8
-------
Hourly average ambient THC measurements were collected from the
test analyzers from October 25 to November 6, 1974. Concurrent methane
measurements from the Bendix 8201 were also collected and used to
calculate pseudo-NMHC values for each of the hourly THC measurements.
These values were then compared statistically by computing the differ-
ences (discrepancies) for each of the ten analyzer pairs. The results
are listed in detail in Appendix B and summarized in Tables 2 and 3.
In Table 2, the "high" and "low" columns give the maximum positive
and negative discrepancies, respectively, and indicate the range of
the individual hourly discrepancies, some of which are quite large.
The standard deviations of the discrepancies for the various analyzer
pairs range from 0.217 to 0.454 ppm. This represents 2 to 4.5 percent
of the full scale range and is very large with respect to the 0.24 ppm
standard for NMHC. Eight of the ten 95 percent confidence intervals
fail to include zero (and the other two are marginal) indicating that
most of the analyzer pairs (including the identical Beckman 400's)
are providing statistically different readings even when the correlation
coefficient is high. Table 3 gives the cumulative frequency distribution
for each of the analyzer pairs.
It can be noted that the data for the pair of Beckman 400's does
not indicate significantly better performance than the other analyzer
pairs. Also, there seems to be no strong correlation with differences
noted in Table 1. These observations suggest that differences in FID
response may contribute to, but apparently are not an overwhelming
cause of discrepancies between various analyzers. If it is a cause,
it probably is of a similar order of magnitude as other causes. Other
causes not eliminated by the experiment are differences in drift and
stability, and analyzer precision and repeatibility errors.
The standard deviations shown in Table 2 apply to the differences
between various analyzer pairs and would therefore tend to overestimate
the errors associated with any one analyzer's measurement. The standard
deviations for individual analyzers cannot be determined from the study
data, but they can be roughly approximated if the assumption is made
that they are all nearly equal. Under this condition, the standard
deviation for the individual analyzers would be approximated by the
average standard deviation (0.322 ppm) divided by /2~, or about 0.23 ppm.
-------
Table 2. STATISTICAL ANALYSIS OF DISCREPANCIES BETWEEN PSEUDO-NMHC
CONCENTRATIONS FROM VARIOUS PAIRS OF FID ANALYZERS
(ppm)
Comparison
Beck.
6800
Beck.
6800
Beck.
6800
Beck.
6800
MSA
11-2
MSA
11-2
MSA
11-2
Ben.
8201
Ben.
8201
Beck.
400-1
- MSA
11-2
- Ben.
8201
- Beck.
400-1
- Beck.
400-2
- Ben.
8201
- Beck.
400-1
- Beck.
400-2
- Beck.
400-1
- Beck.
400-2
- Beck.
400-2
Number
96
96
84
94
215
181
191
181
191
181
High
0.85
1.75
0.75
0.35
2.50
0.80
1.75
0.10
0.50
1.40
Low
-0.70
-0.20
-1.80
-1.30
-0.10
-1.75
-1.25
-1.90
-1.70
-1 .30
Mean
-0
+0
-0
-0
+0
+0
-0
-0
-0
-0
.09]
.205
.139
.374
.371
.036
.136
.306
.461
.154
Std. dev.
0.217
0.454
0.323
0.277
0.389
0.248
0.308
0.353
0.360
0.294
95% Conf. int.
-0.140,
+0.113,
+0.209,
-0.431,
+0.318,
-0.001,
-0.181,
-0.359,
-0.513,
-0.198,
-0.051
+0.300
-0.068
-0.317
+0.424
+0.073
-0.092
-0.253
-0.409
-0.110
Corr
0
0
0
0
0
0
0
0
0
0
. coef.
.980
.982
.941
.957
.925
.935
.888
.934
.897
.909
Average std. dev. 0.322
-------
Table 3. CUMULATIVE FREQUENCY DISTRIBUTION OF DISCREPANCIES
(percent of total readings)
Discrepancy, ppm
Comparison 0.0 0.1 0.2 0.3 0.4 0.5 1.0
Beck. - MSA
6800 11-2 9.4 42.7 70.8 87.5 91.7 97.9 100.0
Beck. - Ben.
6800 8201 15.6 47.9 69.8 72.9 74.0 80.2 92.7
Beck. - Beck.
6800 400-1 9.5 36.9 71.4 82.1 90.5 92.9 97.6
Beck. - Beck.
6800 400-2 2.1 6.4 19.1 42.6 60.6 73.4 97.9
MSA - Ben.
11-2 8201 7.4 20.0 . 41.9 60.5 67.9 77.7 93.5
MSA - Beck.
11-2 400-1 13.8 57.5 79.0 91.2 95.0 95.6 99.4
MSA - Beck.
11-2 400-2 9.4 29.8 51.8 70.2 82.7 88.5 99.0
Ben. - Beck.
8201 400-1 19.9 44.8 57.6 68.0 74.6 81.8 95.0
Ben. - Beck.
8201 400-2 4.2 13.6 31.9 42.4 56.0 66.5 93.2
Beck. - Beck.
400-1 400-2 14.4 37.0 66.3 74.0 88.9 93.4 98.3
11
-------
On this basis, about 95 percent of the measurements would be within
± 0.46 ppm (two standardized deviations). This is 4.6 percent of
the 10 ppm full-scale range -- not unusual for air monitoring
instruments, but almost twice the 0.24-ppm standard, which is only
2.4 percent of the full-scale range. Obviously, it is difficult to
measure concentrations below 1 ppm with a scale range of 10 ppm, and
the 10-ppm range is necessary to include the normal range of ambient
THC and methane concentrations. (The highest concentrations observed
during the experiment were 8.9 ppm THC and 6.75 ppm methane.) This
leads to the further observation, shown in Table 4, that the 6 to 9 a.m.
average NMHC concentration exceeded the NMHC standard almost every day
for which data were collected -- sometimes several-fold! Although all
but the 8201 data are pseudo-NMHC data, the magnitude of the 3-hour
averages cause great concern when compared to the 0.24-ppm standard.
In normal operation, when the analyzers would be making actual
NMHC measurements, the discrepancies would be expected to be even
larger because of variation in the methane measurements and from other
variables that were controlled in this study. This suggests that the
types of methods tested are apparently not adequate to measure ambient
NMHC concentrations, particularly with respect to the Ambient Air
Quality Standard for NMHC.
SUMMARY AND CONCLUSIONS
Substantial discrepancies occur among different models of flame-
ionization hydrocarbon analyzers when measuring such compounds as
ethane, ethylene, and actylene.
When measuring ambient air, discrepancies in hourly average pseudo-
NMHC measurements exceeding 1 ppm were observed. The standard devia-
tions of these discrepancies ranged from 0.217 to 0.454 ppm and averaged
0.322 ppm. In actual use, discrepancies would tend to be larger because
of differences in methane separation efficiency and various routine
operating factors, all of which were minimized or eliminated during
this study.
Differences in response to various NMHC may contribute to the
discrepancies, but are not an overriding cause. Drift, in-
12
-------
Table 4. 6 to 9 a.m. NMHC CONCENTRATIONS
(NMHC Standard = 0.24 ppm)
Date
10/26/74
10/27/74
10/28/74
10/29/74
10/30/74
10/31/74
11/01/74
11/05/74
11/06/74
Beck.
6800
a
a
a
a
-a
0.75
0.40
-0.08
0.12
MSA
11-2
1.57
0.52
0.48
1.07
3.40
0.85
0.58
0.08
0.27
Ben.
8201
0.98
0.22
0.25
0.65
0.85
0.42
0.32
0.15
0.08
Beck.
400-1
1.55
0.32
0.45
1.38
-a
0.95
0.62
a
1.42
Beck.
400-2
1.55
0.42
0.58
1.15
-a
1.17
0.92
0.60
0.78
Avg.
1.41
0.37
0.44
1.06
2.12
0.83
0.57
0.19
0.53
aNo data available.
13
-------
stability, precision, and repeatability errors are apparently the
important causes of discrepancies.
When NMHC concentrations are determined by difference between
THC and methane measurements, it is necessary to use scale ranges of
at least 10 ppm. Thus the 5 to 10 percent (of 10-ppm full-scale range)
discrepancies commonly observed are large with respect to the normal
NMHC range of 0 to 2 ppm and extremely large with respect to the
0.24-ppm standard.
REFERENCES
1. 40 CFR Part 50, Appendix E - Reference Method for Determination of
Hydrocarbons Corrected for Methane. Federal Register. 36(84):
8198, April 30, 1971.
2. 40 CFR Part 50 - National Primary and Secondary Ambient Air Quality
Standards. Federal Register. 36J84): 8186, April 30, 1971.
3. Ortman, G. C., and V. L. Thompson. Performance of Hydrocarbon
Monitoring! Instrumentation. In: Instruments for Monitoring Air
Quality. Amecican. Society for Testing and Materials, Philadelphia,
Pa. ASTM STP 555.. 1974. 74-84.
4. Reckner, Louis R. Survey of Users of the EPA-Reference Method for
Measurement of Non-Methane Hydrocarbons in Ambient Air. Scott
Environmental Technology, Plumsteadvilie, Pa. Prepared for EPA
under Contract No. 68-02-1206. Publication No. EPA-650/4-75-008.
December 1974.
5. Ortman, G. C. memo to Franz Burmann. April 12, 1974. Included
herein as Appendix A.
6. 40 CFR Part 53 - Ambient Air Monitoring Reference and Equivalent
Methods. Federal Register. 40_(33):7044, February 18, 1975.
14
-------
APPENDIX A: MEMO FROM 6. C. ORTMAN
UW8TED STATES ENV8ROWW1EWTAL PROTECTdOW AGENCY
SUBJECT: identification of Problem (Hydrocarbons Corrected
For Methane Methodology)
FROM:
DATE: April 12, 1974
Gordon C. Ortman
TO:
Franz Burmann
THRU: Larry Purdue
Recent discrepancies in data obtained by hydrocarbon monitoring
instrumentation have been brought to my attention by Vinson Thompson.
More specificallys two commercial analyzers, viz.s the Mine Safety
Appliances (MSA) 11-2 and the Bendix 8201 reactive hydrocarbon analyz-
ers, produce markedly different measurements of the non-methane fraction
of the total hydrocarbon content of the ambient air. A side-by-side
operation of the two analyzers has generated data reflecting ratios of
hourly averaged data for "reactive" hydrocarbons of a magnitude at times
greater than four to one. The anomaly is not attributable to instrument
malfunction or improper calibration. Mr. Thompson has convinced me that
he has conscientiously followed accepted procedures in conducting the
tests. I am further advised Ralph Baumgardner of CPL has made similar
observations arrd additionally has noted that the Beekman S800 gas
chromatograph on a common sample produced total hydrocarbon less methane
values that were higher than the MSA readings and lower than the Bendix
readings.
Early last week, I hypothesized that the phenomenon was attributable
to the response characteristics of the detectors. I then visited our
library and did a rapid literature search reading abstracts of patents
dating back to the fifties and scanned the pages of a dozen or so articles
all dealing with FID's. The hypothesis was tenable. Subsequently, I
conferred with Mr. Poll of MSAS Mr. Leesberg of Bendixs and Drs. Perry
and Lauer of International Rockwell. I also recalled a discussion with
Messrs. Chapman and Villalobos of Beekman several months ago. Evidence
exists in the minds of FID technologists and in the literature that one
not only could but should find a disparity in the response characteristics
of FID's having even very minor differences in design configuration or
the same detector using different flow parameters. What has not been
appreciated is the degree to which these differences can be manifested.
My explanation for the lack of cognizance of the depth of the problem by
the scientific world in general and the air pollution investigator in
particular is that typically applicable reported research involved gas
chromatographic analyzers equipped with FID's. Of necessity GC column
variables, etc., dictated compound by compound calibration that tended
to minimize detector peculiarities. I do not know of any in depth
comparative studies of different total hydrocarbon analyzers used in
ambient air analyses where methane was subtracted. Failure to subtract
the methane would tend to mask the degree of nonuniformity of response
to other taydraearbeinis. This is tnne by virtus of th© fact it is the
predominant hydrocarbon in the atmosphere and methane is generally used
EPA Form 1320-6 (Rav. 6-72}
15
-------
to calibrate air monitoring analyzers, therewith insuring uniformity of
response to it.
In summary, it is my conviction that whereas FID's probably respond
reasonably uniformily to paraffins they vary significantly in their
response to olefins, acetylenes, aromatics and other organics such as
the alcohols, ketones and aldehydes. Present instrumentation does not
allow legitimate comparison of total hydrocarbon less methane data
collected by gas chromatographs designed to produce those data, such as
the Beckman 6800, Bendix 8200, and Byron 200 series or the MSA 11-2
and the Bendix 8201.
To add to the complexity of the above stated problem is the
recognized fact that the photochemical reactivity of different hydro-
carbons varies tremendously.
It is my belief EPA must face up to the stated problem. There are
solutions. But irrespective of the solution, I can not foresee any being
free of arbitration. Until a solution is agreed upon, and backed by
sound research, perpetuation of any numerical concentration limits for
hydrocarbons corrected for methane now appears inane. Within this past
week I have seen one analyzer produce readings that averaged for one
hour a value of 0.22 ppm carbon and another analyzer producing a value
in excess of an average of 1 ppm for the same time period for the same
air sample.
16
-------
APPENDIX B
17
-------
STATISTICAL
ANALYST
SYSTEM
_0
B
S
0
_ A_
T
,E .
T
I.
M
E
F
c
K
6
9
-M.
S
A_ .
. .5
£
N
(^
F
C _
K
. 1..
_£
r
C.
K
_ 2 .
C
H
.4
-P
r
c
E
8
3
E
C
7
B
6.
8
M
S
B
6
.8
8-
E
N
B
6-
8
3
1
B
.£.. :
8
e
4. ..
M
c
n
r
M
M ..
S
3
1 ...
M ..
S
B
. 2-
8
N
B
...1-
B
N
B
-»
i
B V
2
159 1027 3
160 1027 4
161_ 1C27 E
162 1027 6
IE 3 JC?7 7
164 1027 8
1CS 1 H77 q
166 1027 . 1C
16J 10'T- 11
168 1027 12
169 1 077 ' 1 ~?
170 102J 14
172 1027. 16
173 1027 17
174 1027 18
175 1027 1*
176 1027 20
177 1P?7 71
178 1027 22
£ 180 1028 0
iai i n?a ±
182 1028 2
1S3 1C28 J
184 1028 4
!«<; ic?" *
186 1028 6
187 1fI7fl 7
188 1028 8
IRQ i n? R *$
19C 1028 10
iq-j 1P7R 11
192 1028 12
jqj 102" 1 7
' 194 1028 14
19 5 1P7S 1^.
196 1023 16
197 71178 17
198 1028 18
199 1P78 13
200 1028 20
2Q.1 _1CZ8_ 2X
202 1028 22
2C3. 1028_ 23
204 1029 0
20 5. 102 9 1
' 20E 1C29 2
2C7 1D79 3
2C3 1029 4
210 1029 6
711 1P?<3 7
3.35 3.1 3. EC 3.70
3.55 3.1 3.60 2. EC
3.90 3.7 4.CC 4. 1C
4.90 4.5 4.50 4 .5 C
7. nn >.-? 7-7n 7^.an
2.3C 2.1 2.30 2.30
7-n* iTq 2-?ri 7 .in
2.15 2.0 2.10 2.20
2-rm 2-P 2>rC 7^20
1.95 1.9 1.90 2.10
I.""5 1.7 i-on 2-10
1.95 1.9 1.90 2.2C
3. C5 2.6 2.80 3.20
3.15 2*7 3.20 3.10
3.50 3.0 3.5e 3fl«"t
3.80 3.4 3.80 4.00
3.25 3.C 3.10 3.0C
3-15 2.7 2.90 3.00
2.30 2.6 2.70 2.70
?. <;n 7.7 7.70 7-*n
2.50 2.3 2.40 2.60
7.70 2.B Zi7C 2tB0
2.50 2.3 2.40 2.5G
2-70 2.1 2.PP 2-r7O
2.80 2.7 2.80 2.80
2.8C 2.5 2.70 2.90
?- 7P ?»** ' ** " 2 «7 *"*
2.4C 2.2 2.20 2.3C
2.05 1.9 1.90 2.10
2-P5 1.9 1.90 J^PO
2.05 1.9 , 2.00 2.10
2.3C 2.0 2.30 2.70
3.25 7.3 3.20 3. 20
3. i c 7.7 i.rin ?.aji
2.80 2.5 2.80 2.80
2.80 2.G 7.80 2 .SJL
2; 8C '.E 2.3C 2.9C
.3.50 3.C T.TP &.DC
4. OS 3.7 3.80 3.80
3.35 1.1 3.2.0 JJ.1C
3. -50 3.^ 3.50 3.3C
5.5C 4.5 E.2C ? .2 C
5.45 S.t; E.7C 5.7C
S.ir 4.7 C,.7CI 5.FC
2.55
2.35.
-2*65
3.15
t .IS
4.75
2.15
2.C5
1.85
1.85
1.85
1.75
1.75
1.85
2 .-!£-
2.25
r.a-
2.95
2.65
2.45
2.25
2.25
2.^5
2.25
2.45
2.25
2.25
2.05
1.95
1.75
1.75
1.75
1.85
2.15
2.3-5
2.25
-2-.3J5.
2.25
3. 55
S..D5
2.95
3.1 ;
3.«?E
.80
.70
.Bi, -
.75
.15
.25
.30
.20
.20
.10
. 90
.90
1.15
.95
.75
.60
. 70
.55
. 45
.25
r 7K
.25
- 35
.35
.55
. 45
.35
.20
.30
.30
.30
t-4C
.45
l.PC
1.10
. An
.55
.55
. 95
.40
.55
1.4C
1-1 =
.55
.25
.35
.EE
.35
-.25
.05
.C5 .
.15
iS
.15
.05
.45
.45
.55
.35
.35
.25
' .35
.25
.05
.15
.05
-i 5
.25
^25
.25
.25
.15
.05
.15
.15
.15
.15
.15
.65
.25
.2.5.
.25
-45
.05
-.05"
.25
.35
.3E
-.55.
-.C5
.75
.95
.75
.55 -
.85
.35
-.25
.15
.25
.35
.25
.15
...-15-
.05
.35
1. 20
.95
.45
.45
.45
.15
. 35
.15
.35
.45
25
.15
15
.15
.15
.25
.45
1.05
1.05
.55
.55
.75-
.15
.55
1.C5
1-2^
1.15
.65
.95
.95
.55
-.25
.25
...25-
.35
.35
.35
75
.35
l»O5~
.85
1.2-C
1.15
.35
.45
.45
.35
.4 5
.25
.35
.75
.65
.45
.25
.25
.35
.35
.45
.85
-JU.2.5
1.05
.55
.55
.65
"t.45
.15
.n?
.35
2.05
.15
.65
1.55-
.25 -.15
.45 -.05
.25 .05
.20 -.10
»?0 .30
.40 .40
.10 .10
.20 .00
.15 .05
.05 .05
.05 .05
". . ;£5 ,(J5
. '"' ^05 .05
'IB j.25"
.45 -.05
~' "" 50 05
.40 .00
.40 .30
.25 .15,
xSS ,,25
.20 .10
20 20
.20 : .10
.20 .10
.10 .00
.30 .00
.30 .10
.20 .20
.20 .20
.15' .05
.15 .15
15 15
.15 .05
.30 .00
.45 .05
.45 .05
.30 .00
.30 .00
.C0 .20
.35 .25
.30 .00
1.05 .35
1.25 .35
.45 -.25
-.J5_
.05
-.35
-.20
.10
.40
00
.00
.05
-.05
4.15
-.15
-t»15
-Us
15
.05 .
1MB * OB
-:.20
' IB
.10
oo
-.10
.00
.00
.00
.20
-.10
00
.10
-.. OS
-.05
05
-.05
^.40
.05
.00
-Op
-.10
.50
.25
- 71
.20
. nn
-.65
1.75
-.25
-»AO_
-.50
.20
-.30
.00
.00
DO
-.20
. 30
-.10
.00
20
:/.oo
' ~~ 20
-.50
jB5
-.40
ID
"^. T_Q
00
-.ID
-.10
in
-.10
-.20
00
.00
.00
00
-.10
J 20
-.30
50
-.40
30
-.30
-.30
30
-.10
in
-.30
-.70
-.90
-.70
-.60
-.40
-.40
20
.00
10
-.20
20
-.20
r»n
-20
ID
-.30
Ffl
-.40
E5
-.60
20
.otr
' 30
20
-.30
30
-.2C
*20
-.10
.50
-.40
vn
-.10
2D
-.20
10
-.20
-.70
-yp
-.40
7C
-.30
-.40
1 00
-.10
Tnn
-.10
-1.7C
.50
-.70
-.BP
-.?_ I
.1
_ n ^
-.1
2
.0
*- 1
.0
J
-.1
-'2 !
*; ;;"^
£-*. - -1
2 '^ . H
JL - - ' -
.1 - :
i "
= '* 'I
-.2 - " -^
i - - H
-'\ , *--1
.0
-:l 1
?
-.1
-:2 -s<<
-.1
_.i
-.4
.0
.0
.0
-.1
7
.0
.1
.2
- r»
-1.0
1.4
-I '
-------
SYSTEM
0
_ A
T
_ EL..
T
I
M
E
3
E
r*
. _ K _ K _. .3_
£ S -.
J A N ...
D
E
C
K
. 1.
p
E
_Q_
K
2
C. .
H
. 4
_
r
.-C
B
£
3_ ...
3
-£
3
_
n
1
3
k
8
o
2
M
S
3
r
N
M
S
_
B
J.
t»
s
_
Q
2
* 8
£^
N
P
-1
3
r
N
___
E
-.2-
q
1
. ._
3
2 . ..
106 1025 22 3.5C 3.2C 3.3 3.SP 2.3T 1
107 1026 23 3.25 2.5C 3.C 3. P.C 2 .£ S
108 1QJC 0 4.70 3.7T 2.75.
1C9 1030 1 E.fS 5.5T 3.7F
IIP JC3fl 2 5.55 5.EO JJ .75
111 1030 3 8.90 6.75
JL12__1P_3_0 4 7.5C 7.4_5
113 1030 5 8.15 8.2T 6.45
T m i n *? n £ 7m nc 7 P r R » ^ *?
115 1C3C 7 8.50 S.CC 5.25
llfi isy?- fi 7.E5 5.55 4.4E
117 1030 9 7.10 5.55 4.45
118 lD3f) IP 2. EC 2.4C 7-75
119 1030 11 2.30 2.00 1.35
170 1030 17
121 1030 13 2.40 1.70 1.7 1.80 1.65
122 1030 14 2.4fl 1.7P 1.6 1 . 8n 1.C5
123 103C 15 2.40 1.7C 1.7 1.7C 1.65
124 1030 1$ 2.00 2.50 1.80 1.8 1.8C 1^75
125 1030 17 2.85 2.70 2.4C 7.5 2.5C 1.95
_g 1Z6. 103_D 18 2.35__2.JSG_2.5-C _2. 6 -Z..EG L.95
127 1032 13 5.75 4. 9C 4.CE 5.C 5.5C 2.35
178 10.30 2C 5*45 5» GU 4*10 5*3 5.4C 7. 55
129 1030 21 5.75 5-3-5 4.20 5.4 5.75 2.65
130 1030 77 E.CC 5.55 4. SO 5.8 5.95 2 .* 5
131 1030 23 5. CO 4.9C 4. EC 5.4 5. 25 2.95
132 1Q3J C 3.05 3.15 2.50 3.5 3.30 2^'S
133 1031 1 2.30 2.70 2.3C 2.7 2.3P 1.95
134 1C31 2 2.10 2.30 2. DC 2.4 2. EC l.SS
135 1031 3 2.40 2.50 2.2C 2.4 2.SC 1.95
13S 1C71 4 2.55_2.7C 2_.4C 7_,«J_ 3j 1C 2j2c
137 1C31 E 3.CS 2.35 2.F-C '.9 3.35 2.25
138 1031 Z 3.T5 3. C5 2.SC 3.2 3.3E.2.2E
139 1C31 7 3.05 3.15 2. EC 3.1 3.4E Z.C5
141 1C31 9 2. CO 2.T5 1.9C 2.1 2.4C 1.75
142 103i ID
143 1C31 11 1.7C 1.95 1.7C 7.T 2 . 2f 1.75
144 1CT1 17 1.7C l.SE 1.7G 1.7 2.2ia.^T
14? 1C31 13 LOT 1.35 1.7." I.E 7.rr 1.65
14£ 1C31 14 1,3C l.£E 1.7C 1..7 2.J.C J..E.5.
147 IC'l 15 1.<3C 1.9E 1.90 1.8 2. ?P 1 .F 5
146 1C31 1C r.10.2.15 1.9C 7.1 3. 4f l.=5
lrr 1C'1 13 4.9'" 4.4C "'.-" 4.7 4.'"" -.15
151 1CT1 19 3.3C 3. 8C 3. 1C 1.9 4.3C 2.13
15? 1031 2C 3-^C 3. =C .2.8C 3.5 3.3C 2..C'
1 SI 1031 2° r.SC '» r" "'.3" ~.~ 7.3" ~.lr
IT: IC'l 33 3.47 3.33 ?.?" ". ~ 3.E" 3 .1 t
1E= 1C2~ C 3. OS 2.7C 3.0 3. 1C 2.4E
157 IP;- 1 2. 3C ?.FC -.7 7.7C 2.'E
IE? IC0? 2 2. EC 7.T 3.3 2.50 2 .1 C
.EE
.EC
1.95
2. 3D
1. 3C
1.70
.75
3.25
2.65
.45
.75
.75
.75
-?5 ~TT
.90 .7E
3.40 ?.5E
7.9E '.45
3. 1C 2.70
T.I c r.70
2.0 5 1.95
.an .sr
.15 .75
^2J . 45
.45 .55
.30 .45
.8C .7C
fl p on
l.CO 1.10
.-J_5 ..65.
.25 .30
-.05 .2C
.05 .20
-.05 .20
_.! r ».2C
.2 5 . 3r,
.4 5 A5C.
'.7 2 "". 'r
l.S f 1.F5
1.4: i.:r
1 .4 c 7 . TC
1.2 3 1.1"
.F.C
.45
.45
.35
.25
^95
1.75
. 75
2.15
1.75
..75
.75
i. 1C
1.10
.15
.05
Tn?
.05
.05
.45
..55
i.es
1.55
1-G5
1.C5
-75
.35
.15
.25
.15
.35
. ..35
w cc
.JJ5-
.15
-.rs
.cs
.C5
.C5
.15
" .35
. 95
.75
,.,.
.SE
.25
. 25
a ^ r
,(,r ,gc
3D .35
.05 .15
-.05 -15
.05 .CS
*cs .as
.55 .55
. 65 .£.5
2.65 3.15
2.75 7,35.
2.75 3.10
_2_..35. .3-OC
2.45 2.30
J..25 I.E 5
.75 .95
.55 . .75
.45 .85
_.45 ..85
.65 1.10
..95 J_.1_Q.
1.05 1.4C
.. .85 i.nn
.35 .65
.25 .45
«_C5 .56
-.05 .35
..C5 .45
.IE .55
.45 1.75
r".cc 2.7r
1.7E 2.15
1.4E 1.7E
1. Tr 1 ." 5
1 . 1 r 1.F5
1.3- 1.4-
.45 .65
.3E .3=
.IE ..3" 5
.15
.05
.85
.45
.40
.45.-
.10
-.40
-.2a
-.10
.10
-CO
-.10
-.C5
-.25
-J.5
-.25
--.US
-.05
_-.C5
.4C
.CC
-.1C
- .1C
.CE
?0
.45 .35
.35 .25
1.75 .7=
J..3S .15
1.55 .35
l.CO -.40
.55 -.45
.00 -.40
.20 .CO
.45 .15
.45 15
.45 -.05
-in ,4o
.10 -.10
.CC -.30
.CC .^QQ
-.1C .00
.10 . .10
.10 .10
.2C .20
.7C -.1C
7G .DC
.6C -.1C
.20
.35
.25
.25
.00
.15
-.25
^.25
-.60
-.40
-.30
"*t\
-.40
-^S5-
-.40
-.50
-.40
. 3C
-.30
-1.3Q
.CC
-.50
-.30
.45
-.20
-.2C
.30 .20
.35 .25
i.nc
1.05
J..Q5
-.05
.00
2.50
2-.ln
1.55
.20
.30
.70 .70
.70 r«0
.70 .70
.70 .70
.30 .20
.30 .20
.90 -.10
.90 . 30
1.15 -.05
1,05 .25
.90 -.50
,<:«; -.55
.40 .00
.30 .10
-7fl T00
.35 .05
»45 -.15
.55 .05
.30 -.20
.15 -.05
.25 -.C5
,15 -15
.15 .25
.15 .IS
.15 .15
. ..25- ..C5_
.3 c . 25-
.70 -.10
.80 .10 -
55 . 05
.70 . 20
.55 -.15
.35 .15
.2C .10
.30 .30
.EO
.25
.go
.70
.20
.20
-.60
.40
-.40
-.35
-.20
-.30
-.40
-.m
-.30
-.35
-.35
-.25
-^36
-.15
-.25
.45-
-.50
-.20
35
.3C
-.25
-.05
.10
.10
_01
-.1
.0
.0
-.1
.1
-1.0
1.2
-1.2
-1.3
-1.4
-.4
-.4
-.2
-.3
.P
-.5
-.2
-.3
.1
.0
-.2
-.8
^.-7
.6
-.5
-.7
-.1
.0
-. 3r
-.10
-.10
.10
.CO
.00
-.10
10'
-1.50
1.50
-1.55
-1.35
-1.25
.80
-.60
-.60
-.75
-.75
-.85
=.55-
-.50
-.50
-.51
-.30
^~AC
-.40
-IvSO
1.4C
-1.2C
.90
1.-QC
-.80
-.4C
-.1C
-.2C-
-.70
.00
-.10
.00
'00
.00
00
-.50
.10
-.35
05
.15
..20
-.20
-.20
-.40
.4n
-.45
.lc
-.35
-.30
-.20
-.51
-.40
-.4C
-L.3Q
7"
-.30 ..-
Tp
CO
-.10
-.2C . .
.00
-.20 - -
-------
STATISTICAL
A N A L Y
T S
SYSTEM
E
E 3 B ~
C T C ET E E
C A__I K... . .A._ .3 . ...C C C C._. .. .
8TH6SCKKH EM
SEC8AN1?<» 8 S
212 1C29 8 1.90 1.5C 5.5 1.50 3.75 1.E5
213 1C 7.9 9 3.25 2.7C T.r 3.rr -.35 .90
215 1029 11
216 1029 12
217 1C29 13 1.85 1.70 1.75 .10
218 1029 11 1.85 1.8G 1.65 .20
219 1C79 IE 1.85 1.70 1.55 .20
220 1C7.9 16 2.G5 1.9£ i..7E ._ . .3.C
221 1C29 17 2. ED 2. 3C 1.35 .75
. - 222 1029 18 3i 15 7. 70 7._05 1.10
223 1C79 19 3.25 2. 7C 1.95 1.3C
221 1C29 20 5.15 3. SC 2. 75 3.10
225 1079 21 5.75 1.10 2.85 2.90
_226 1C2S 22 . - C..10_A.. 3T_. . 2.35 - -2-J.5
227 1029 23 7. 1C 5. CC 3.55 3.15
^ZS^JLICI C_ 3.6C. ^t£D 3.-C£_-3_t7 -3..3.C 2 .25 . . l._35. JLJuS
229 1101 1 3. 05 3.15 2. SO 3.2 3.5C 2.75 . 8C .90
231 1101 3 3.05 3.05 2. 5C 3.1 3.30 2.75 .30 .80
. 23.2 -HIM -.4 .2.^55. .?.Jr.O_2.A£. 2.7 3..3C 2.15 ... ...JtC .. -.5 5
o 233 11C1 5 2. 1C 2.50 ?.3C 2.5 2.70 2.15 .25 .4 e
73» iip-r" F; ?.?,*, ?.rr< ?.^r ?.7 s.nr; ?*nr . s;n .S.K
235 11C1 7 2. 30 2. ED 2. 3C 2.7 2.90 1.95 .35 .?5
.736. 1101....8. ,?».?n 7.10 7.1T 7.1 ?.8P 1.15 ^2E.. .15
237 11C1 9 2. 1C 2.25 2.CC 2.2 2. EC 1.35 .25 .10
239 11C1 11 1.70 1.95 1. 3C. 1.9 2 .CC 1.75 -.05 .20
210 1101 12 1.7C 1.85 1.7^ 1.8 2.CC 1.S5 .05 ,2_C
211 11C1 13 1.7C 1.85.1.7C 1.7 2 .1C 1.S5 .0? .20
213 1101 15 1.75
211 uni is i.ao 1.95 i, st .I-..S. 2^.1" 1..6.5- . ,.ir-_Jj.c
215 1101 17 1.70 1.15 -l.rr 1.T 1.75 1.95 2.75 2.20
215 11C1 13 5.30 5.15 1. 3C 5.7 6. CD 2.25_ 3.55 .3..2.C
217 11C1 19 1.90 5.10 1.C7 5.5 5 .CC 2.35 ?.S5 2.75
218 U£i 20 2. 3S 1.5.!: 3.5f.1.5 I .1 f ...2^75. .l.SC. .2.30
219 11C1 ^1 3.70 7.90 3. IT T.f 1.CC 7.15 1. 5C 1.75
251 11C1 23 3.^5 J.50 ?.£" 3.' 3.~0 2.13! . 3C 1.35
!
n
N
.35
..15
-.05
.15
.05
..15
.15-
.. .6.5 .
.75
1.25
1.25
1-35
1.35
.35
.33-
.25
.15
...45
.35
.15
.C5
.D5
^5
1.65 2
-2..C.5. S
l.f?5 3
.95 1
^5 1
.£5 1
S
r
c
i
.65
.25.
.1*
.95
.35
.55
.35
.75
.35
.15
.05
.15
.35
.1;
.15
.7*
.35
a
C
1.25
.65
.25
1.55
1.25
1.05
1.15
.5^
-.35
.95
.75
.'S
.15
2.80
3-..7.S
2.£S
2.3,5
1.85
1.3 '
1.C5
B
8
M
S
.. .£0
-.1C
.CC
-.15
-.20
-.05
-.30.
-.15
-.25
-.15
JK
.55
-.2C
-.2C
-.15
6
.-£
E
N
&a
.15
.25
.55
..15
.10
-..15
.00
.10
-.1C
.00
.00
.on
1.1C
l.-SC
.90
- - -.-30
.60
.25
B
9
3
1
-.15
-.25-
-.05
--.15-
-.10
05
-'.20
-.10
-20
-.Iff
.CO
^00-
.10
-ID
-.SO
.-.65
.1C
1 c
-.15
B M
8
E.
B E
2 N
. . . . .1C.
.55
.20
.15
.05
.15
rl*
.30
.55
i .as
1.65
.80
2.10
-.20 ,50
-.15 .55
-.25 .55
-.75 .30
-.30 .30
-.35 .20
-.60 .30
* 60 » 30
-.50 .25
-.30' .15
-.3D .15
^.1C .15
--in -ip
-.05 .55
-.?n i -i«;
-.10 1.10
-.3D .30
.25 .70
-.15 .70
tt... M_
S S
B B
1 .2
-.60 .ID
.25 .25
-ID .1C
-il" -.20
-.05 -.35
-.05 -.25
.10 -.10
on *^n
-.10 -.30
«00 .10
.05 -.35
.05 -.05
.ns -.is
.15 -.25
r n -> m 7 ^
.15 .45
-.15 -.60
-.10 .10
.30 -.10
.00 .30
B
.. F
N
B
-i.on
-.30 -
- .10 -
.70 -
-.60 -
c n
-.60 -
-.20 -
? n
-,10 -
.30
-.20 -
-.10 -
-i n
.00 -
1-1"
.00
-.70 -1
-1.10 -1
-1.50 -1
-.50 -
-.70 -
B
p
N
IT
2
.30
JJO-
.90
.80
~-3O
.40
Kfl
.60
m70
.60
.20
.40
.11
.-6Q-
.15
.70
.00
.90
^ 40
.40
q
1
3
2
- .00
. IP
-.30
-.20
.6P
-.20
.30
-.20
-.40
-.10
-.40
-Tin
.60
-.45
.30
.50
60
-.40
lO
.30
-------
TATISTICAL ANALYSIS SYSTEM
0.
B
S
D
A
T
E
T
-X
H
£
B
e
c
if. _
6
3
_ M-_
S
A
B- 3_
t £
_ B. _.C - -C- -C..
E K K H
N L Z 4.
_B
E
C - -. - .- _
E H 8
.3 S N
8
_-E
C
2_
B
. E
C
2
8
£.
3
H
S.
8
6
9
B
E
- H
B
E
8
»-
B
1
B
K
8
_
B
2 -
H
_a
E
M
M.
S
_ _^
B
1 -_
W
S
:_-
B
-2
B
2-
53 11C5
Et 1105
55 11CE
56 1105
57 IIJTE
58 1105
t;g 1105
EC 1105
K.I nn^
62 1105
K7 HIT;
61 lies
S5 1105
E6 1105
K7 nrr;
68 1105
P.I nn^
70 1105
71 11P^
72 1105
73 1025
71 1025
76 1025
77 102*
78 1025
73 1C25
SO 1C25
RI IP?";
82 102E
AT in?«;
81 1026
a1; IP?F
85 1H2E
37 1T7F
88 1C26
3r 1C2S
31 1C26
92 1C2C
3<4 1CCC
95 1526
9<- 1C2E
27 1C2E
93 1T2E
22 1"£
icr irzc
1" "£.
1T2 1T26
1-r ^-1
1
5
E
7
o
9
1C
11
L2
13
1_4
15
1 F
17
1 1
19
'p
21
? ?
23
^3
11
16
L7
18
1 9
20
22
73
C
1
2
4
S
E
7
8
ir
n
12
1 7
11
- r-
1°
X.5 l.SC 1.7 1.7 2.10
1.6 l.EC 1.7 1.8 2.15
.1.5 1.7D 1-7 2.2E.
1.6 1.85 1.3 2.35
1..7 1.85 1.8 2-3E
1.7 1.7C 1.7
1-7(1 1,7
1. £ 1 .70. 1 -7
1.6 1.70 1.7 1.8 2.00
1-c i-7r 1-7 1-7 i-qn
1.7 1.85 1.8 1.8 2.00
1.6 1.85 1.8 2.30
1» 5 li-7C 1.7 2rPP
1.5 1.70 1.6 1.90
1-4 i-7n i -K 7.7 i,«»c
1.5 1.70 1.7 1.7 1.90
1.5 l-7n 1.7 1.7 1.80
1.6 1.35 1.8 1.8 1.30
2.15 1.9 1.9 2.30
2«lr 1.9 l.Q 2.3P
2.15 1.9 2.0 2.10
Z.6C r*."> ?-»* 2. EC
3. Of 2.5 3.0 3.CC
2.7C *>.7 "".7- 2 7 n
3.35 ?.9 3.5 2.7C
3.15 2.f 2.9 3. PC
2.30 2.5 2.8 3.CC
7- qr t ,5 ^.p ''."C
3.15 '.7 3.C 3.2C
r.OF 7.S ->.9_ T.-rr,
3.5C ".0 3.S '.50
3.7C 3.1 '.7 3.70
1.7* 7T1 1.3 tt-c. C
3.8C 7.5 3.7 3. EC
2. uc ?.i 2.2 :.rr
2.15 1.9 2.C r.lC
2.15 "." 1.3 2 .1:
2.1T T.^, 1.3 ".2?
2-Cf 1.9 1.2 2.2C
T. 2E ? .? 7 .C > .SC
^. -.- --.-,r ^r,
-..-- -.: '.z 2-ic
1.F.5
i.es
1.75
1.75
l.ES
1.65
1.65
1.65
1.65
1-55
1.65
7 -c«;
1.65
1 E5
1.75
1.75
1 7S
1.75
1JS
1.8E
1-.SS-
1.95
2.D5
2.05
1-9*!
-.05
?-rc
2.35
2. £5
2. £5
2.CS
1.35
1.33
1.7S
i. :"
^
~>* ' -
-.lr -.t5
-.C5 -.C5
.-..IE -.C5
-.15 .1C
.CS .20
.C; .05
--C-E -C5
-.cr .cs
-.pr .05
.05 ,2C
-.05 .20
.15 *O5
-.15 .05
- .2 5 ,05
-.15 .C5
.15 .05
-.15 .10
.1C
.1C
.85
1.20
-.85-
1.1C
1.30
1.10
.75
.75
.80
.95
Iol5
7 _«n
1.75
.15
,^1C.
.3C
.3C
.3C.
.r c.
i.it
.7;
,05
.05
-.CS
.05
.15
.05
.05
.05
.05
.15
.15
.15
-.05
.05
.05
.05
.IE
.15
.65
.15
.95
-.7S
.15
.35
.15
.35
,1^
.15
.1C -
.55
-,***-
1.15
.15
. .IS -
.05
.IS
B r* c
.1 £
.IE
.7E
7 r
a 7 r
.-C.5 .15
.15 .50
.En
.60
.70
.15 .35
.OS -.25
.15 .35
.25 -»»5-
.65
35
.25
.05 -25
.05 .25
.05 .15
.05 .15
.15 .55
.25 .65
75 ».85
1.15 1.15
-.85 .85
1.55 .75
1.15 1.45
.85 .95
,55 ,75
.75 .95
.65 .85
.75 l.l"
.85 .95
1*15 1.15
1.65 1.15
.25 .75
.7*, -T1
.15 .25
.CS .2S
-.C5 .35
.C5 -IS
.IE .55
1C or
.35 .75
.85 .3C
-!£
.00
-.10
-.25
-.15
.00
-.10
-.15
-.IS
-.25
-.20
-.70
20
-.25
-..2_-^2 -.^60 -.13
-.1 -.2 -.55 -.10
--1 -.6S .00
-.2 -.75 .C5
=-.1 --65 .OS-
.C .00
.00
-.1 -.2 -.10 .00
.1 T! .30 .00
-.1 -.1 -.30 .05
-.1 --2 »40 .95
-.2 -.70 .05
.2 .50 .00
-.1 -.10 ..10
.2 .3 .50 .10
-.2 ~.2 -.40 .00
-.2 -.2 -.30 .05
.25
25-
.25
.55
.40
.15
.55
.65
.40
.30
.15
.15
.50
.60
.60
.85
.30
.30
.25
.15
.25
.15
.35
.35
.5 C
-.in -»sn
-.20 -.55 -
-.55
-.50
-.^>5O
-.10 -.30 -
.05 -.15
.05 -.25
-.45
.3 0
-.2C
.00 .20 -
.00 -.20
.05 -.05
.25 -.15
j->5 r .is
.15 -.25 -
.05 .05 -
.GO -*00 *
-.15 .55 -
.15 .15
.25 .15 -
i20 eOO
.00 -.20 -
.05 -.05
.15 -.05 -
.15 .25
.10 .00 -
10 ^«2-O-
.00 .00 -
.10 .30 -
.20 .10 -
.15 -rm -
.15 .05 -
.25 .05
.35 -.05
25 -=-^.5-
.15 -.35
?" »6O
.25 .35 -
-.15 -.15 -
.30 ,OO--
r-n
.1
.1
.0
, )
, i
.0
oO
.0
.0
0
ol
j 3
.5
.6
yx
.1
.3
. 5
.3
T
.4
, «;
.6
.2
.1
, i
.1
, i
.1
. i
.0
ol
.5
***-
i
-o^S -o35
-o55
--.55 -.
-.30 -0-20
-.20 -o2C
-.50
-,30
-»»2C- o20
-.12 -.13
** « QQ '^»QC
-«5D -.00
~vC" flC
-oSO .DO
r*o 00
.20 .80
-.50 -.10
-.SO -.20
-.50 -.20
.70 ^P
-.50 -»1C
-.60 .OC
t.lO -- '0
.03 o20
.-.20 --10
-.70 --tf
-.20 -.10
-r.30 -.1C
0^0 o'JP
-.50 -.EC
0 g n 0 3 c
.00 .10
-.5C ."C i
-------
STATISTICAL
ANA L Y S I
S Y S T E H
rue,
B
E B B
D T C E E
0 AIK M B C C
B TUB'S EK K
5 EE8 AMI 2
1 110E 0 l.E 1.8E 1.7 1.8 2.00
2 11C6 1 1.5 1.35 1.7 1.8 2.0O
3 1106 2 1.6 1.35 1-.8 1.8 2.00
4 11C£ 3 1.7 1.9E 1.8 1.9 2. DC
5 1106 4 1.7 1.9f 1.9 1.9 2.10
6 11CE 5 1.8 2.QE 2.C 2 .C, 2.2C
7 1106 5 1.8 2.05 1.9 3.0 2.60
. 8 11O6 7 2.1 2.15 2-0 3.9 2.7D
9 1106 8 2-0 2.15 1.3 2.9 2.60
in HOC 9 ~
11 1106 10
12 HOG 11
. 13 11C6 12
14 tine *3
15 1106 14
ift line i« .
17 1106 16
1 «' 11 PK 1~7
19 1106 13
K 21 1106 20
7f 1TTG 71
23 1106 22
24 line 73
25 1104 C
2E 11T4 1
27 11C4 ?
2 8^1X04 3
29 1104 4
30 1104 S
31 1104 6
32 1104 7
33 11C4 8
34 11C4 9
35 1104 10
3S 1104 11 1.85 1.8 1.8 2.0C
37 1104 12 1.7 1.70 1.7 1.8 1.30
38 1104 13 1.7 1-7(1 '1.7 .1.8 1 .' .7 1.9 2.1G
46 11C1 '1 " . r 1.7C 1.7 " .' *~
47 urt " I.E i.7c i.r i.c 2.1;
48 11C4 ~ 3_ 1. ^ 1.7" " ' 1.7 ~ .* c
49 urs c i.r i.c" :.~ 1.7 r.ir
:c nr: i i. r i.c- i.- :.i5
n i:r: > i. - L.S" i .7 :.;-
-? urs - :- ' i.-.r -. .? - .
c
a.
4
1.65
1V7_5
1.75
Ij^flS.
1.85
l*3i
1.S5
1.85
,
1.65
1.65
1.65
1.6E
l.GS
JI..E5
1.65
1 G£
i:||
A/A*rtt
B
C
£
8
-.C5
-.15
-.15
-.15
-.05
.15
M
S
.20
.1C
.-1C.
.10
_»1£L
.20
.30
B
N
.05
-.05.
.05
-C 5
.05
-OS
.05
.05
_. ..B B
E E
C... JCL
1 2
."if ."35
.05 .25.
.05 .25
__.05_.15_
.05 .25
1.15 .75
1.05 .75
- -
*P =
,C5
^r g
.15
C.5.
.ir
- .r, r.
- .i'
_ - 1 -
-?r-
.05
.ns
.05
r ?P
.20
.20
^21X
.OS
..C£
.C5
. C5
-.C5
-.Ci
.15
.05
.05
.05
.05
-.05
.05
.C5
~ .r 5
.r. r
' 1* 35
.15 .25
.15 .20
...IS -.2C
.25 .30
-.15 .3C
.15 .45
.25 ..45
.25 .45
.45
.15 .52
.£5 . 50
.V* .50
.sc
.55
.55
B
B 6
6 8
8
B
M .. .£.
S N
-.25 -.1
-.35 -.2.
-.25 -.2
-.25. -..1
-.25 -.2
-.75 .2
-.25 -.1
-.15 .1
B B
6 6
B 8_
1 2
-.2 -.40
-. * SO
-.2 -.40
-.2 1.30.
-.2 -.40
? on
-1.2 -.80
-.9 -.60
M
S
B
N
.15
.15.
.05
.05
..15
.25
H n
.. J- -B_ . B.
1 2
.05 -.15
-.-OS -.15
.05 -.15
.05 -.15
>-.. 95 .-.55
lTTi .55
-r.75 -.%5
-- -,
n . n
E E
u *
-
1 2
-.1 -.30
--1 -T 30
.0 -.20
.0 -.20
-1.1 -.70
1.9 .70
-1.0 -.70
+
B - ,-l
I
R
2
-.20
-.20
-.20
i -9rT' . ,.^v. -.;'-'
' ' ' ,V a-.
S'1:.^*"*' \'<"riJ'?S."' ''
. -:/3"'"v . . - . jMt^^rff^ "*
"- ' -. - -*- -» -^ T^*.'*%"*-" - tj t;e»
r- -
^'-"^'TT^f^^
: . V -' - - ' - " »»«. T -.-. . --
._ _
* ' ' ' .
-
"'' ,-'/ '" '
, - .- ' .;--
' ' " «
..--..- -
. . :. ..:.,-, .-..«.. -,:'-_ r- -;...:» '--v^ '
.00 .0
.00 .0
-» 1 c . C
-.05 .1
-.05 .1
--.15. . ^C
-.1C -.1
-.10 .1
-.1C .0
-^20 .i
-.1G -.2
-.1C -.1
-.1C -. 2
-.1C -.2
-.1 -.20
1 ?n
-.1 -.15
-^.1_=.1S_
-.1 -.15
-.1 -.25
.0 -.30
- -.2. -.40 -
-.3 -.50
-.5E
-.2 -.55
~~»2 "£ 5-
-.2 -.65
-.05
-.70
-.70
rnt;
.00
rOP
.00
.15
.15
- .05
.15
.OC
..CD
.10
-*10-
-.1C
.OD
-.10
-.1C
-n«; .1 K
-.10 -.20
-.10 -.15
.05 .00
-.05 -.10
.05 .10
.05 -.25
^--.05^.25
-.20 -.40
.40
-.10 -.45
-..00 ^.45-
-.10 -.55
-.55
-.eo
-.EC
^ft_
-.1 -.20
" tl .21?
-.1 -.15
-.1 ,is
-.2 -.25
.0 -.15
-.1 -.40
.2 -.40
-.2 -.40
-.2 -.55
1 .55
'.0 -.45
-.55.
-.50
-.50 -
"' ''-»»1 - " '-
-.10
-in
.-.05 . .
.05
'-.OS
-.30
.20
-.20
-.35
.45
-.45
-------
TECHNICAL REPORT DATA
(Please read Inunctions on the reverse before completing)
RfPORT NO.
EPA-600/4-75-010
4. TITLE AND SUBTITLE
Hydrocarbon Measurement Discrepancies Among Various
Analyzers Using Flame-ionization Detectors
5. REPORT DATE
September 1975
6. PERFORMING ORGANIZATION CODE
3. RECIPIENT'S ACCESSION-NO.
7. AUTHOR(S)
Frank F. McElroy
Vinson- L. Thompson
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring and Support Laboratory
Office of Research and Development
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
1HA326, ROAP 22ACK
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Final Tnhnuse
14. SPONSORING AGENCY CODE
Same as Above
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Several experimenters have observed substantially discrepant nonmethane hydrocarbon
measurements from various types of hydrocarbon analyzers measuring identical samples
of ambient air. To formally investigate this problem, an experimental laboratory
study wherein a number of flame-ionization analyzers were observed under controlled
conditions while making simultaneous measurements of urban ambient air was conducted.
Substantial discrepancies were found to occur among different models of analyzers.
The standard deviations of these discrepancies averaged 0.322 part per million,
which is extremely large with respect to the ambient air quality standard (guide)
for nonmethane hydrocarbons, 0.24 part per million.
17.
a.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Air pollution
Hydrocarbons
h.lDENTIFIERS/OPEN ENDED TERMS
Flame-ionization detector
Nonmethane hydrocarbons
c. COSATI l-icld/C.roup
13B
1'i. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (This Kcpoft/
Unclassified
21. NO. Of- PAGES
26
2O. SECURITY CLASS (This page)
22. PRICE
Unclassified
EPA Form 2220-1 (9-73)
23
------- |