JUN 24 1975
EPA-650/4-75-008
December 1974
Environmental Monitoring Series
W:j:i
-------
EPA-650/4-75-008
SURVEY OF USERS OF THE EPA
REFERENCE METHOD
FOR MEASUREMENT
OF NON-METHANE
HYDROCARBONS
IN AMBIENT AIR
by
Louis R. Reckner
Scott Environmental Technology
Plumsteadville, Pennsylvania 18949
Contract No. 68-02-1206
ROAP No. 26AAF
Program Element No. 1HA327
EPA Project Officer: John H. Margeson
Quality Assurance and Environmental Monitoring Laboratory
National Environmental Research Center
Research Triangle Park, N. C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
WASHINGTON, D.C. 20460
December 1974
-------
EPA REVIEW NOTICE
This report has been reviewed by the National Environmental Research
Center - Research Triangle Park, Office of Research and Development,
EPA, and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the Environmental
Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U .S . Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
1. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2. ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved
methods and instrumentation for the identification and quantification of
environmental pollutants at the lowest conceivably significant concentra-
tions. 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 for sale through the National
Technical Information Service, Springfield, Virginia 22161.
11
-------
iii
TABLE OF CONTENTS
Page
1.0 INTRODUCTION 1
2.0 IDENTIFICATION OF USERS OF NON-METHANE HC ANALYZERS 2
3.0 PREPARATION FOR ON-SITE SURVEY 3
3.1 SELECTION OF USERS FOR ON-SITE EVALUATION 3
3.2 DEVELOPMENT OF FORMAT FOR ON-SITE SURVEY 3
3.3 UNKNOWNS FOR ANALYSIS BY USERS . 4
3.4 PRELIMINARY SURVEYS 8
4.0 RESULTS OF ON-SITE SURVEY 9
5.0 DATA ANALYSIS 13
5.1 ZERO ERROR 13
5.2 SPAN ERROR 16
5.3 INSTRUMENT RESPONSE TO HIGHER HYDROCARBONS 19
5.4 PRECISION OF METHANE AND THC DATA 19
6.0 DISCUSSION OF RESULTS 23
6.1 AREAS FOR IMPROVED TECHNIQUES 23
6.2 INSTRUMENT CAPABILITY 26
6.3 PERFORMANCE OF OTHER TYPES OF NMHC INSTRUMENTS 28
7.0 CONCLUSIONS AND RECOMMENDATIONS 29
7.1 CONCLUSIONS 29
7.2 RECOMMENDATIONS 29
APPENDIX
Table A-l - Users of Non-Methane Hydrocarbon Analyzers A-l
Figure A-l - Non-Methane Hydrocarbon Analyzer On-Site
Evaluation Format A-6
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
iv
SET 1385 06 1274
ABSTRACT
Scott Environmental Technology, Inc. performed a survey of users
of the EPA Reference Method for measurement of non-methane hydrocarbons in
ambient air which resulted in the compilation of a list of 188 NMHC analyzers
operated by seventy organizations. Field evaluations were performed on
instruments operated by sixteen of the users. The evaluations were per-
formed in the East, Midwest and Far West and included state and local air
pollution control agencies as well as private consulting firms.
The accuracy of the NMHC data being obtained by the sixteen users
of the reference method was determined by presenting a series of five gas
mixtures in high-pressure cylinders for analysis by each operator. The results
for the. mixture containing NMHC at a concentration close to the 0.24 ppm - C
ambient air standard showed that substantial errors existed in current NMHC
data. The errors are summarized below:
Error Range Number of Users
~0-10% 1
10-20% 3
20-50% 2
50-100% 4
> 100% 6
Detailed information regarding instrument operating conditions and operator
techniques was also recorded at each user location.
An analysis of the data showed that the inaccuracies in current
data make it impossible to determine whether ambient air quality is in
compliance with the standard. The major factors contributing to data
errors were:
1. Failure of operators to understand and/or follow the instrument
manufacturers' operating instructions and the reference method
procedures for NMHC as published in the Federal Register.
2. Span gases containing unknown amounts of higher hydrocarbons.
3. Span gases not in air.
A. Span gases incorrectly analyzed for methane.
5. Zero errors due to sampling system contamination and lack of
adequate checkout procedures.
6. Excessive instrument zero and span drift during unattended
operation.
•f\ ; SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-1-
SET 1385 06 1274
1.0 INTRODUCTION
This report covers work performed by Scott Environmental Technology,
Inc. under EPA Contract No. 68-02-1206, "Survey of Users of the EPA Reference
Method for Measurement of Non-Methane Hydrocarbons in Ambient Air." The
program involved the identification of organizations operating non-methane
hydrocarbon (NMHC) analyzers and the on-site evaluation of fifteen typical
instruments. The objective was to determine the accuracy and reliability of
NMHC data being recorded by users following the principles of the EPA
reference method and to recommend improvements in technique which would
aid in producing improved data.
An extensive survey of organizations engaged in air monitoring
resulted in a compilation of 188 NMHC analyzers operated by seventy organizations,
Field evaluations were performed on instruments operated by sixteen of these
organizations. The evaluations were performed in the East, Midwest and
Far West and included state and local air pollution control agencies as
well as private consulting and testing companies. Instruments from five
manufacturers were evaluated.
The accuracy of the NMHC data being obtained by the sixteen users
of the reference method was determined by presenting a series of five Scott
gas mixtures in high-pressure cylinders for analysis by each operator.
Detailed information regarding instrument operating conditions and operator
techniques was also recorded at each user location. Since the intest was
to evaluate the method as being used by the operator, precautions were taken
not to improve the user's technique until after all information had been
recorded.
The data obtained for the test cylinders was analyzed for accuracy
and precision, and the magnitude of various individual error sources was
determined. Emphasis was placed on the quality of NMHC data in the area of
the 0.24 ppm-C ambient air standard level. General problems with the
instrumentation and techniques are discussed, and recommendations for
improving the data accuracy are presented.
'' " ^-V !
° V SCOTT ENVIRONMENTAL TECHNOLOGY, INC
-------
SET 1385 06 1274
2.0 IDENTIFICATION OF USERS OF NON-METHANE HC ANALYZERS
The first step in the program was the identification of organizations
performing non-methane hydrocarbon (NMHC) analysis of ambient air according
to the Reference Method for Determination of Hydrocarbons Corrected for
Methane as published in the Federal Register, Vol. 36, No. 228, pp 22394-6.
This method specifies the use of gas chromatographic type flame ionization
hydrocarbon analyzers for the semicontinuous analysis of ambient air. It
was anticipated that the several manufacturers of NMHC analyzers would
cooperate in making the names of purchasers of their instruments available
to Scott. However, only one company made such a list available. Other
companies stated that company policy did not permit the release of purchasers'
names. Appeals to high management levels to waive the policy were unsuccessful.
Thus, alternate approaches were required to obtain the names of users.
The list of users was compiled through contact with organizations
listed in the 1973 Directory of Governmental Air Pollution Agencies, Regional
Air Pollution Control Directory, 1972 Air Pollution Consultants Guide and
the Air Pollution Directory and Laboratory Guide. These publications were
obtained from the Air Pollution Control Association, American Chemical
Society and the EPA Project Officer. Additional contacts were made with
organizations known to Scott and organizations suggested by the primary
contacts.
The list of users obtained as a result of these efforts is presented
in Appendix Table A-l. The location, type of organization and instrument
manufacturer and model number are also included. The list contains a total
of 188 instruments operated by 70 users.
The AID instrument is a portable unit designed for short term use.
As such it is in a separate class from the other instruments shown, but it was
included in the survey because it follows the same operating principles. Some
of the other instruments have been discontinued or sold in very limited quantities.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC
-------
-3-
SET 1385 06 1274
3.0 PREPARATION FOR ON-SITE SURVEY
3.1 SELECTION OF USERS FOR ON-SITE EVALUATION
The program plan called for the on-site evaluation of fifteen
instruments used for monitoring ambient air by the EPA Reference Method for
Determination of Hydrocarbons Corrected for Methane. The instruments were
selected from the list of users given in Table A-l. The considerations
in selecting the instruments were:
1. Instrument Manufacturer - At least one instrument should
be chosen from each manufacturer with emphasis on those which
represented the major share in use.
2. Type of Organization - The various organization types
included state and local air pollution control agencies,
and private consulting and testing companies.
3. Location of User - The users should be located in various
areas of the United States.
The instruments evaluated in the field program are discussed in
Section 4.0. Numerous revisions were made from the initial proposed list
of users because certain instruments selected originally were not in
operation, had been incorrectly identified by the user, or the user was
not interested in participating in the program. When substitutions were
necessary, every effort was made to locate a similar instrument and
organization type in the geographical area.
3.2 DEVELOPMENT OF PROCEDURE FOR ON-SITE SURVEY
The primary purpose of the on-site survey was to determine whether
non-methane hydrocarbon analyzers in actual use in ambient air monitoring
were being operated in accordance with the EPA Reference Method, manu-
facturers' instructions and good laboratory practice; and to determine the
accuracy of the data being collected by the users. A check list was
developed to assure that the Scott chemist performing the evaluation would
acquire a complete picture of the instrument operating practices and performance.
A copy of this check list is included in Appendix, pages A-6 to A-ll.
The information areas covered in the evaluation included:
•f S| ) SCOTT ENVIRONMENTAL TECHNOLOGY, INC
-------
-4-
SET 1385 06 1274
1. Manufacturer's name and model number.
2. Organization category of user and use type and frequency.
3. Training and experience of instrument operator.
4. Operator's technique as compared to manufacturer's instructions,
published method and good laboratory practice.
5. Instrument history including type and frequency of problems
and preventive maintenance program.
6. Instrument operating conditions.
7. Calibration technique and source of calibration gases.
8. Evaluation of instrument accuracy, precision and drift.
9. User's opinion of analyzer's strong and weak points with
recommendations for improvement.
10. Evaluator's opinion of analyzer's strong and weak points
with recommendations for improvement.
Two additional steps were added to assure that the field survey
would result in a realistic evaluation of in-use practices. First, in order
to obtain frank responses from the users, each user was assured that complete
anonymity would be maintained throughout the study. Appendix Table A-l, while
identifying all users, maintains the anonymity of those who participated in
the field survey. Secondly, so as not to bias the results the Scott chemist
performing the evaluation was careful to avoid any comment or suggestion
regarding any phase of instrument operation until after all evaluation data
had been recorded.
3.3 UNKNOWNS FOR ANALYSIS BY USERS
The key in the plan to determine the accuracy of data collected
by the users was the presentation of several gas mixtures to the users for
analysis. The composition of these mixtures would be known to Scott, but
identified to the users merely as mixtures of hydrocarbons in air. The
concentration data obtained by the users would then provide a measure of the
accuracy of the ambient air data being collected by the users.
In order to obtain a complete picture of the users' data and to
isolate specific potential sources of error, it was concluded that five test
mixtures were needed. These included:
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-5-
SET 1385 06 1274
1. Methane in hydrocarbon-free air to check the users' span gases.
2. Hydrocarbon-free air to check the instrument zero points.
3. Methane at a typical ambient level + higher hydrocarbons at
approximately the ambient standard (0.24 ppm - Carbon).
4. Methane as in 3 + higher hydrocarbons at typically high ambient
levels (~ 10 times the ambient standard).
5. Higher hydrocarbons at the same level as 4 but without methane.
The last mixture was added to explore the instrument response factor
for higher hydrocarbons versus methane. It was known that flame ionization
detectors generally yield a lower response to higher hydrocarbons in air
than methane in air on a carbon basis. What was not known was whether the
response ratio varied from instrument to instrument. This could be determined
by comparing the users' data for mixture 5 (higher hydrocarbons only) to that
for mixture 1 (methane only).
It was originally planned to present these mixtures to the users
in push-button cans. These cans, which Scott has filled and sold over the
last ten years, are readily carried in a brief case and had been shown to be
stable containers for Cn to C, saturates and olefins. They have proved
1 o
especially useful in gas chromatographic work where a relatively small sample
is required. Scott's normal procedure for preparing can mixtures is to first
prepare the mixture in a high-pressure cylinder and then to fill the cans
from the cylinder. Scott has developed can filling procedures which assure
that the can concentrations are the same as the cylinder concentration.
A test program was carried out to determine if the can mixtures
would be sufficintly stable to satisfy the needs of the field evaluation program.
Three high-pressure cylinders, corresponding to mixtures 3, 4 and 5 above
were prepared and a dozen cans were filled with each mixture. The cylinders
and several sample cans were analyzed for methane and total hydrocarbons over
a thirty-day period. The can concentrations of methane were stable through-
out. The total hydrocarbons in the cans showed a small but measurable (<0.10 ppm)
increase with time. It was concluded that small concentrations of higher
hydrocarbons were being emitted by the can gaskets. However, it was believed
that analysis of each can before and after use in the field would adequately
define their THC concentration for field evaluation. The methane would be no
problem.
i SCOTT ENVIRONMENTAL TECHNOLOGY, INC
j
-------
-6-
SET 1385 06 1274
At this time Scott became aware of potential limitations in the
use of the cans for checking the NMHC analyzers. EPA personnel suggested
that the sample flow system used in the analyzers was different than the
ordinary gas chromatograph so that sample flow rate was critical. This
potential problem was investigated by testing the cans and cylinders at a
nearby user location. It was found that small but significant variations
occurred when successive samples were injected from the cans. The cylinder
gave consistent results. The cans were then outfitted with a flow control
valve and a second field evaluation was made. Consistent results were
obtained for methane using a can, but the total hydrocarbon results still
showed somewhat greater variation than desirable. It was also noted that
the total hydrocarbon from the cans increased slightly as the can pressure
decreased. The NMHC analyzers inject samples from a flowing stream, and
there are restrictions downstream of the sampling loops. The flow rate thus
affects the volume in the loop at the time of injection. The variations in
THC results even with flow control may have been related to system contam-
ination problems found later in the field program.
Because of the potential problems with the cans, it was decided
to abandon plans for their use in favor of the high-pressure cylinders.
While not nearly as convenient to transport as the cans, the cylinders
were considered necessary to meet the program objectives. The same set of
cylinders was used for all field evaluations.
The cylinders which had been made up with higher hydrocarbons were
analyzed for individual hydrocarbons by gas chromatography. Calibration stan-
dards for the GC analysis were prepared by injecting known amounts of each C_
to C, hydrocarbon into calibrated glass flasks and pressurizing the flasks to a
known pressure in the vicinity of 0.5 atmosphere. This procedure is used by
Scott for analyzing Close Tolerance Analyzed Gas Mixtures which it sells to
government and industry. Methane was determined against Scott cylinder standards
which are checked periodically using flask standards. Isopentane and n-hexane
were calculated from the average carbon response to the C~ to C, saturates.
The composition of the cylinder mixtures is given in Table 3-1. The
cylinders were also analyzed for THC using a flame ionization detector (FID)
continuous hydrocarbon analyzer calibrated with methane. The data show that
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-7-
SET 1385 06 1274
the higher hydrocarbons were -30% lower than found by GC. This difference
in response is discussed earlier in this section. The NMHC instruments in
use in the field would be expected to give the same reading as the FID rather
than the GC because they measure THC in air using a methane calibration standard,
TABLE 3-1 CONCENTRATIONS OF INDIVIDUAL HYDROCARBONS
IN STANDARD CYLINDERS, ppm-C
Compound B-1441 B-1442 B-1AA3
Methane
Ethane
Ethylene
Propane
n-Butane
Isopentane
n-Hexane
Total HC
Total NMHC
THC Reading (FID)
NMHC*
*THC(FID) - Methane(GC)
The cylinders were analyzed for methane and THC several times
prior to the field program, in the middle; of the program and after its
completion. All differences were within experimental error (1 to 2%), so
it can be concluded that the cylinder mixtures were stable throughout the
program. The important point is that the NMHC analyzers read THC by FID
and the Scott FID showed no change in THC for the entire test neriod.
1.45
0.074
0.044
0.051
0.084
0.070 .
0.024
1.797
0.347
1.68
0.23
1.45
0.58
0.44
0.69
1.44
0.70
0.24
5.54
4.09
4.35
2.90
0.00
0.60
0.46
0.69
1.44
0.70
0.24
4.13
4.13
2.86
2.86
! SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-8-
SET 1385 06 1274
3.4 PRELIMINARY SURVEYS
Two preliminary evaluations were carried out to determine the
adequacy of the data format and evaluation procedures. They gave the
Scott evaluator an opportunity to gain experience with the test routine
and estimate the time required to complete each test.
The results of the preliminary tests showed that a significant
THC response (-0.5 ppm-C) was obtained for the Scott zero gas. The gas
had previously been a.aslyzed against a primary standard and no hydrocarbons
haa been found. The zero sir was? checked by passing it through a heated
catalyst bed and than into a FID total hydrocarbon analyzer. This reading
was compared to that obtained when the catalyst bed was by-passed and
found to be identical. The performance of the catalyst bed was checked
by treating gas from Cylinder B-1442 (methane + higher hydrocarbons) in
a similar manner. The gas read 4.35 ppni directly from the cylinder. After
passage through the catalyst bed the FID gave the same reading as it did
for the zero gas with and without catalyst bed exposure. This was firm
evidence that the zero gas was clean, and that the readings obtained in the
field were in error.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-9-
SET 1385 06 1274
4.0 RESULTS OF ON-SITE SURVEY
The field evaluation of non-methane hydrocarbon analyzers was
performed from January to April, 1974. A total of sixteen users were
included in the evaluation. General information regarding the sixteen
users is presented in Table 4-1. The users included six in the Middle
Atlantic States, five in the Midwest and five in California. Each geographical
area was represented by at least one user of each type: state air pollution
agency, local (city or county) air pollution agency and private company
(consulting and testing laboratory). In order to maintain anonymity of
the users, as discussed earlier, the numbers were assigned to the users on
a random basis.
Data relating to the concentration range in use, span gas composition
and concentration and gas supplier are shown in Table 4-2. The concentration
data for the five Scott gas mixtures obtained by each of the users are given
in Table 4-3. It should be noted that all concentrations are based on peak
heights compared to the electronic zero point. This is equivalent to the
manner in which the data are recorded and calculated by the data systems in
use. It also represents the appearance of the recorded peaks when the
instrument is operated in the normal barographic mode. During the actual
tests the instruments were operated in the chromatographic mode so that the
continuous output from the detector was recorded. This permitted an inspection
of the data for zero shifts, peak shape, etc. which was quite useful for
data interpretation.
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
TABLE 4-1 INFORMATION ON NON-METHANE HYDROCARBON ANALYZERS EVALUATED IN USERS' FACILITIES
VI
n
O
m
Z
O
rn
n
X
O
P"
O
O
o
Scott Evaluation
User
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Instrument Type of User
Mfg. /Model No. Organization
Beckman -
Beckman -
Beckman -
Beckman -
Bendix -
Beckman -
Beckman -
Beckman -
Beckman -
Bendix -
6800
6800
6800
6800
8201
6800
6800
6800
6800
8201
MSA - 2472
AID - 514
Beckman -
Beckman -
Bendix -
6800
6800
8200
Byron - 230
State
Local
State
State
Local
Private
Local
Local
Private
State
Private
Local
Local
Local
Private
Private
Experience General
of Operator Laboratory
(yrs) Condition
2
2
_
15
12
2
2-1/2
1/2
23
24
1-1/2
2-1/2
10
3-1/2
2
2
Excellent
Good
Excellent
Good
Good
Fair
Good
Good
Fair
Good
Good
Good
Good
Fair
Fair
Good
of
Operator's
Technique
Very
Good
Good
Very
Good
Poor
Good
Good
Fair
Poor
Very
Poor
Very
Good
Poor
Fair
Good
Good
to
to
to
to
to
Very Good
Fair
Very Good
Good
Fair
Good
to
Fair
Good
to
to
Fair
Good
Users Opinion
of Instrument
No opinion
Good
Fair
to Good
No opinion
Very
Good
Good
Good
Good
Good
Fair
Fair
Good
Good
Poor
Good
Good
to Good
W
H
UJ
00
Ul
NJ
O
-------
-11-
SET 1385 06 1274
TABLE 4-2 RANGE AND SPAN GAS DATA ON NON-METHANE HC ANALYZERS
Span Gas
User
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Range in Use
Mfg. /Model ppm
Beckman - 6800
Beckman - 6800
Beckman - 6800
Beckman - 6800
Bendix - 8201
Beckman - 6800
Beckman - 6800
Beckman - 6800
Beckman - 6800
Bendix - 8201
MSA - 2472
AID - 514
Beckman - 6800
Beckman - 6800
Bendix - 8200
Byron - 230
0-20
0-20
0-20
0-100
0-20
0-100
0-20
0-50
0-7
0-10
0-5
0-8
0-20
0-10 (1)
0-20
0-10
CH4
ppm
4.2
8.5
7.5
44.8
5.0
82.5
15.3
20.1
5.2
3.0
2.1
5.2
7.8
75.2
9.4
2.4
THC
PPm
12.2
15.4
7.5
51.0
5.2
85.0
15.3
20.1
5.2
3.0(3)
2.1
5.2
7.8
75.2
9.4(2)
2.4
Gas
Supplier (s)
M&G Scientific
Air Products
Union Carbide
Airco & Liq. Carbonic
Liq. Carbonic
Matheson
Specialty Gas Lab.
Liq. Carbonic
Air Products & Scott
Linde Gas
Matheson & Scott
Matheson
Liq. Carbonic & Air
Liq. Carbonic
Matheson
Products
Matheson & Air Products
(1) Attenuation changed for cylinder analysis
(2) Span gas in nitrogen
(3) Span gas in argon
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
TABLE 4-3 RESPONSE OF NON-METHANE HC ANALYZERS
TO SCOTT CALIBRATION GASES
8'
___
n
O
m
Z
MRONMENTAI
L TECHNOLOGY, INC.
User
No.
Scott
1
2
3
4
5
6
7
8
9
10
11
12
D-305 B-1394 B-1441 B-1442 B-1443
CH^ THC
ppm ppm
0.00 0.00
0.22 0.49
-0.53 0.35
0.12 0.58
0.00 0.29
0.07 0.34
0.00 0.50
0.00 0.00
0.00 0.61
0.30 0.84
0.35 0.42
0.06 0.10
0.00 1.36
13 0.11 0.24
14
15
16
Mean*
0.00 0.10
-0.20 5.52
-0.12 0.72
0.02 0.47
Median* 0.00 0.42
CH^ THC NMHC
J>pm ppm ppm
4.85 4.92 0.07
4.54 4.91 0.37
4.63 4.77 0.14
4.43 4.44 0.01
5.21 5.87 0.66
4.55 4.56 0.01
4.71 4.97 0.26
5.19 4.93 -0.26
4.26 4.38 0.12
4.73 9.21 4.48
4.68 4.47 -0.20
5.88 5.16 -0.72
4.62 4.61 -0.01
4.38 4.43 0.05
4.14 21.80 17.66
4.65 4.14 -0.51
4.75 4.74 0.00
4.63 4.61 0.05
CH4 THC MNHC
ppm ppm ppm
1.45 1.68 0.23
CH4 THC NMHC
ppm ppm Ppm
1.45 4.35 2.90
1.45 1.84 0.39 1.45 4.51 3.06
0.93 1.59 0.66
1.35 1.74 0.39
1.52 2.08 0.56
1.37 1.56 0.19
1.50 1.63 0.13
1.41 1.39 -0.02
1.31 1.56 0.25
1.57 1.84 0.27
1.50 3.24 1.74
1.42 1.47 0.05
1.53 2.57 1.04
1.43 1.61 0.18
1.30 1.46 0.16
1.18 5.52 4.34
1.25 1.27 0.02
1.38 1.69 0.31
1.42 1.60 0.22
0.94 4.17 2.23
1.37 3.78 2.41
1.49 5.19 3.70
1.37 3.74 2.37
1.49 4.55 3.06
1.28 4.37 3.09
1.32 3.99 2.67
1.58 4.65 3.07
1.48 7.75 6.27
1.43 4.28 2.85
1.54 4.29 2.75
1.45 4.23 2.78
1.31 4.10 2.79
1.24 18.40 17.16
1.27 3.25 1.98
1.38 4.22 2.77
1.40 4.26 2.79
CH4 THC NMHC w
ppm ppm ppm ^
0.00 2.86 2.86 £
Ln
0.17 3.02 2.85 °
-0.50 2.70 2.70 K
0.14 2.61 2.61 **
0.00 3.41 3.41
0.11 2.41 2.30
0.00 3.09 3.09
0.00 2.90 2.90
0.00 2.78 2.78
0.27 3.34 3.07
0.29 5.09 4.80
0.00 2.98 2.98
0.00 3.18 3.18
0.08 2.90 2.82
0.00 2.66 2.66
-0.23 11.30 11.30
-0.09 2.02 2.02
0.01 2.86 2.81
0.00 2.90 2.84
* Not Including THC and NMHC for Instruments 10 and 15 which were erroneous because span gas was not in air.
-------
-13-
SET 1385 06 1274
5.0 DATA ANALYSIS
This section presents an analysis of the data obtained by the
users for the five Scott gas mixtures. Attempts are made to estimate the
contribution to the overall error of individual error sources. The implica-
tions of this analysis in terms of obtaining data to meet air monitoring
requirements are discussed in the following section.
The mean and median values for all of the users shewn in Table 4-3
are in reasonably good agreement with the Scott values for each of the cylinders.
However, the large dispersion among the data from individual users clearly
indicates generally poor accuracy especially near the 0.24 ppm NMHC air quality
standard. The range of errors for Cylinder B-1441, which contained 0.23 ppm
NMHC, is summarized in the following table.
SUMMARY OF USER ERRORS FOR NMHC IN CYLINDER B-1441
Error Range Number of Users
0-10% 1
10-20% 3
20-50% 2
50-100% 4
> 100% 6
Ten out of the sixteen users obtained data in error by greater than 50% of
the true value and six of the ten had data in error by more than 100%. These
overall errors resulted from a combination of individual error sources. Each
is explored in detail in the following sub-sections.
5.1 ZERO ERROR
Zero errors for methane and total hydrocarbons are indicated by
the values obtained for the Scott zero gas (Cylinder D-305). The zero error
for methane can also be observed in the data for Cylinder B-1443 which con-
tained no methane. The methane values for the zero gas recorded by the users
ranged from -0.53 to 0.35 ppm. Six of the users recorded the true value of
0.00 ppm. An examination of the recorder traces showed that all of the non-
zero readings were caused by baseline shifts from the automatic electronic
zero as opposed to real methane peaks which have the distinctive shape
characteristic of gas chromatographic peaks.
Instructions for adjusting the instrument zero are included in
operating manuals. However, many of the users do not recognize the importance
of checking the zero frequently, and the procedure for doing so is sufficiently
complex to tax the skill of a typical operator. The methane zero error
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
-14-
SET 1385 06 1274
contributes an error to the NMHC value which is equal in value but with an
opposite sign. Thus, even if no additional error sources were present, ten
of the sixteen users would have reported NMHC data for cylinder B-1441 which
was in error by greater than 20% (0.06 ppm).
The THC values obtained for the zero gas ranged from 0.00 ppm to
5.52 ppm, and seven of the sixteen values were 0.50 ppm or greater. All
of the peaks appeared to be real and not due to zero shift as was the case
with the methane zero data. Since the zero gas had been demonstrated to
be free of hydrocarbons as discussed in Section 3.4 , and since the values
obtained by the users covered a wide range it became reasonably clear that
the gas was most likely being contaminated as it flowed through the analyzer.
When this type of contamination occurs in a typical GC system, the level of
contamination increases with the sample's residence time in the system. This
was checked at User 8 where a zero gas THC reading of 0.6 ppm was found at
normal sampling rates. When the zero gas flow through the sampling system
was stopped one minute before injection, subsequent sample analysis showed
approximately 1.2 ppm THC. When the sample flow was stopped entirely and
gas in the loop was injected a THC concentration in excess of 1.5 ppm was
obtained.
The THC values for Cylinder B-1441 (0.23 ppm NMHC) are plotted
against the THC readings for the zero gas in Figure 5-1. Data for the two
users who did not calibrate with methane in air are not included. It can
be seen that there was a strong tendency for the users showing higher zero
THC values to obtain higher THC values for B-1441. Errors in THC span gas
values would also affect the data. This probably accounts for the outlier
points for Users 4 and 16 as these two users also reported outlier values
for the Scott span gas (Cylinder B-1394).
It thus appears that zero errors for both methane and THC are a
major source of error in the data for Cylinder B-1441 obtained by the various
users. It is our opinion that the manufacturers' claim that zero gas is not
needed for this type of analyzer is not valid. They base their claim on
the fact that all hydrocarbons are removed from the carrier by the catalytic
purifier and the automatic zeroing systems provide a true zero point. Our
data indicate that these features do not necessarily detect problems which
result in erroneous zero readings and subsequent inaccurate aerometric data.
j SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
r>
O
m
z
I
o
O
z
n
I
o
Q
o
g
<
Q
1.4
1.3
1.2
1.1
0.9
0.8
0,7
0.6
0.5
0.4
0.3
0.2
O.t
0.0
FIGURE 5-1
COMPARISON OF USER THC DATA
FOR CYLINDER B-1441 TO D-305 (ZERO GAS)
PI
H
03
Ui
o
VJ1
SCOTT
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6
THC DATA FOR CYLINDER B-1441, ppm
-------
-16-
SET 1385 06 1274
5.2 SPAN ERROR
The span error for methane can be estimated from the methane data
for the Scott span gas (Cylinder B-1394). First, the data must be adjusted
for any CIfy zero error due to baseline shift. Since this is a constant value,
it affects the actual instrument response at the span point as well as in
the analysis of B-1394. The corrected value is obtained from the following
equation:
CA = RA ~ R0 X C_
A — r— S
Cs-Ro
where:
C. = corrected concentration of A
A
R. = read concentration of A
A
R = read concentration of zero gas
C_ = stated concentration of span gas
O
Methane data for Cylinder B-1394 corrected in this manner are shown
in Table 5.1. The adjusted values show that four of the fifteen users
obtained results within ±5% of the Scott value and twelve of fifteen obtained
results within ±10%. The users values averaged several percent lower than
the Scott value. Exhaustive reanalysis of Cylinder B-1394 has convinced us
that the stated value is accurate to ±1%.
It would be desirable to have a smaller error in the span gas
concentrations than estimated above for fifteen users. However, even with
a 10% error the effect on the NMHC data at the 0.23 ppm level would be
minimal because the errors are compensated to a considerable degree in the
subtraction process (THC - City). Assuming that no higher hydrocarbons are
present in the span gas, a 10% error at the span level would result in a 10% error
in the NMHC values or 0.02 at the 0.24 ppm level. Thus, span gas CH, error
4
is not a major contributor to errors in data for Cylinder B-1441.
Span errors for THC can likewise be estimated from THC data for
B-1394. Despite the fact that the published reference method clearly states
that the span gas shall be in air, Users 10 and 15 had span gases in argon and
nitrogen, respectively. This led to gross inaccuracies at all concentration
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-17-
SET 1385 06 1274
TABLE 5-1 ADJUSTED USER METHANE DATA
FOR CYLINDER B-1394
User No.
Scott
1
2
3
4
5
6
7
8
10
11
12
13
14
15
16
Mean
Median
Read CH,
ppm
4.85
4.54
4.63
4.43
5.21
4.55
4.71
5.19
4.26
4.73
4.68
5.88
4.62
4.38
4.14
4.65
4.75
4.63
Adjusted CH,
ppm
4.85
4.57
4.86
4.39
5.21
4.55
4.71
5.19
4.26
4.97
4.77
5.88
4.60
4.38
4.25
4.55
4.73
4.60
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
-18-
SET 1385 06 1274
levels so this data has been omitted from all data analyses. A further
shortcoming in following the reference procedure was several users use
of span mixtures containing significant amounts of higher hydrocarbons
instead of methane alone. Finally, almost all of the users disregarded
the method specification that the span gas should correspond to 80% of
full scale. While some of these factors do not necessarily lead to invalid
data, they showed that the majority of users were either unfamiliar with
or disregarded the prescribed reference method for non-methane hydrocarbons.
It is not practical to adjust the THC values for Cylinder B-1394
for zero error because the error might well vary with concentration. An
examination of the reported data shows that only three users obtained
values differing from the Scott value by more than 10%. A look at the
NMHC values for this mixture can be more revealing as to potential errors.
It can be seen that half of the users obtained NMHC values within ±0.1 ppm
of the Scott value, but the remaining values showed very large variations
from the Scott value. The negative values for five users are most likely
the result of the presence of higher hydrocarbons in the span gas. In each
of these cases the span gases were labeled as having identical methane and
THC. Of the users obtaining high positive NMHC values, User 1 is brought
into the proper range by correcting the CH4 value for zero error, User 4's error
appears to be due to an overstating of the higher hydrocarbons in the span
gas and User 6's error may be due to THC zero error.
The NMHC data for Cylinder B-1394 show that substantial errors
are caused by inconsistencies between the labeled CH4 and THC concentrations
of the span gas. That is, if both CH4 and THC values are in error by the
same amount, only a small error results, but if one is correct and the other
in error or one value is high and the other lovj very large inaccuracies
occur in ambient air data. It is interesting to note that the users who
obtained the best NMHC data for Cylinder B-1394 also obtained relatively
accurate NMHC data for Cylinder B-1441.
| SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-19-
SET 1385 06 1274
5.3 INSTRUMENT RESPONSE TO HIGHER HYDROCARBONS
It is known that flame ionization detectors yield a greater
response to methane in air than to a like number of carbon atoms of higher
hydrocarbons in air. The difference is approximately 30%, that is 1.0 ppm
carbon of higher hydrocarbons gives the same response as 0.7 ppm methane.
The difference can vary depending on fuel and sample flow rates, detector
configuration, etc. In the user evaluation, efforts were made to determine
whether this relative response varied from instrument to instrument. The
various error sources discussed previously tended to mask possible response
variations. The best estimate can be obtained by comparing the THC data
for Cylinder B-1394 (essentially all methane) to that for Cylinder B-1443
(all higher hydrocarbons). The users THC data for these two mixtures are
plotted in Figure 5-2. Other error sources caused a wide variation from
user to user for each mixture, but for each user the errors should be similar
for the two mixtures. If this assumption is true, then the data in Figure 5-2
should fall on a straight line unless there are variations in response. In
fact the data points fall within approximately ±10% of the mean slope of all
points. This indicates that there may well be response variations but the
error should not be expected to exceed 10% of the THC value. A firmer conclusion
cannot be drawn because of larger errors from other sources. The data do confirm,
however, that higher hydrocarbons yield a lesser response than methane.
5.4 PRECISION OF METHANE AND THC DATA
The precision of the data from each of the users was determined
from results obtained for five consecutive injections of the user's span
gas. The precision was calculated according to the following equation:
Precision = Std. Deviation(95%) 2S = 2~v/S(Xi - X)2
V n - 1
Precision is expressed in ppm and thus related to the concentration
of the span gas. Another user to user comparison can be obtained from the
coefficient of variation which is expressed in percent. This is calculated
as:
Coefficient of Variation (CV) = Standard Deviation(o) X 100
Mean (X)
''] SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
n
O
m
O
m
n
O
O
O
z
n
3.4
3.3
3.2
3.1
3.0
7 Q
CO Z'y
cr
Q 2.8
>~ -7 -7
o 2.7
o:
o
^ 2.6
<
h-
^ 2.5
2.4
2.3
2.2
2.1
2.0
o
FIGURE 5-2
COMPARISON OF USER THC DATA
FOR CYLINDER B-1394 (METHANE) AND
CYLINDER B-1443 (HIGHER HYDROCARBONS)
CO
pi
OJ
oo
N3
M
O
I
^ ..» .
4.0 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.
THC DATA FOR CYLINDER B-1394, ppm
-------
-21-
SET 1385 06 1274
The precision and coefficient of variation for each instrument are presented
in Table 5-5. Errors due to precision are small compared to errors discussed
previously. In addition, these errors are random so they would have little
effect on hourly averages. The previous errors were systematic in nature
so they would have a direct effect on all data points.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-22-
SET 1385 06 1274
TABLE 5-5 PRECISION OF METHANE AND THC DATA
Methane
Total Hydrocarbons
User No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Cone.
PPm
4.2
8.5
7.5
44.8
5.0
82.5
15.3
20.1
5.2
3.0
2.1
5.2
7.8
75.2
9.4
2.4
Precision
ppm
0.34
0.10
0.11
0.07
0.04
2.82
0.53
0.14
0.07
0.06
0.13
0.80
0.10
0.25
0.05
0.10
Coeff .of
Variation Cone.
%
4.0
0.6
0.7
0.1
4.3
1.7
1.7
0.3
0.7
1.0
3.2
7.7
0.6
0.2
0.3
2.0
ppm
12.2
15.4
7.5
51.0
5.2
85.0
15.3
20.1
5.2
3.0
2.1
5.2
7.8
75.2
9.4
2.4
Precision
ppm
0.15
0.09
0.18
0.19
0.25
2.80
0.31
0.28
0.16
0.20
0.12
0.11
0.10
0.95
0.14
0.09
Coeff .of
Variation
%
0.6
0.3
1.2
0.2
2.4
1.6
1.0
0.7
1.5
3.3
3.0
1.1
0.6
0.6
0.7
1.9
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
-24-
SET 1385 06 1274
by each operator on a regular and frequent schedule. If the operator could
not correct any deficiencies shown in the performance tests, he would either
summon another person in the organization with greater expertise or the
manufacturer's field service representative.
Zero error was shown to be a significant source of error in NMHC
data. This was believed due to both inaccuracies in the electronic zero
setting and in the case of THC, contamination within the system. The error
due to the zero setting can be eliminated by following the manufacturer's
instructions for adjustment. However, the contamination problem may be
more difficult to solve. It seems likely that this contamination results
primarily from adsorption of hydrocarbons on sampling system surfaces during
high concentration periods followed by desorption during low concentration
periods rather than from hydrocarbons emitted directly by the system
components. Some users reported that overnight flushing with clean air or
nitrogen reduced this error.
Zero error can be identified by frequent performance checks with
true zero THC air. If the response to zero air was not within certain
limits, corrective action would be required before valid data could be
reported.
Span error results primarily from span gases not in air, non-
methane hydrocarbons present in span gas mixtures and methane incorrectly
analyzed by the supplier. As discussed previously the error in span gas
methane analysis produces a far smaller error in NMHC data than do the
presence of other hydrocarbons or the incorrect oxygen content in the span
gas. This error could be markedly reduced if all span gases furnished by
suppliers were analyzed for both methane and THC and referenced against
standards similar to those now provided by the National Bureau of Standards
for propane.
Two further checks could determine whether valid data was being
recorded. First, each user should have available a cylinder containing
1.5 to 2.0 ppm methane plus 0.24 ppm-C of higher hydrocarbons. We believe
that the NMHC in such a mixture would be very stable over a long period of
time. If the user would analyze this mixture with his instrument after
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
-25-
SET 1385 06 1274
performing the zero and span checks, he could determine the true accuracy
for NMHC at the 0.24 ppm (ambient standard) level. If the NMHC reading
was not within specified limits further corrective action would be necessary
to report valid data. We believe that the mixture suggested above would be
stable for an extended period of time. Obviously the supplier would have to
reference each cylinder against a primary standard.
A second check would involve user participation in a cross-reference
service. In this service subscribers are supplied with unknown mixtures
for analysis every three months. Cylinders shipped to all subscribers are
filled from a single large high-pressure container and are checked for
uniformity. The results reported by the subscribers are statistically
analyzed and reports of all data are sent to each subscriber. Cross-
reference services have proved to be a valuable tool to organizations
measuring mobile source emissions, but attempts to interest ambient NMHC
instrument users in a similar service have been unsuccessful.
In analyzing the Scott gas mixtures, the users treated them in
the same manner as their span gases. That is, they were introduced through
the span gas inlet port at flow rates comparable to those used for the span
gas. While this procedure was most desirable for detecting the errors
discussed previously, it made it impossible to evaluate other potential
errors such as those due to different flow rates between air sample and span
gas and contamination in sample pump or external sampling line. To have
investigated these factors for all users would have required excessive gas
volumes and time. Therefore, this part of the study was limited to a few
users where one or more Scott gases were introduced into the air sampling
line. The results obtained were similar to those recorded when the gas
was introduced in the normal manner. We have concluded that any errors
were much smaller than those previously discussed. Any potential errors
could be detected by the user through a performance test involving intro-
duction of a gas mixture through the sampling line and a comparison to data
for the same gas introduced as in the normal span procedure.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-26-
SET 1385 06 1274
In summation, we believe that the accuracy of the NMHC data
currently being recorded can be greatly improved by providing the operators
with simplified routine checkout procedures. These procedures would
include specified performance tests which would show data accuracy and
indicate whether repairs by an instrument specialist were required.
Currently, most operators are confused by the complex instrument operating
instructions. As a result the routine calibration and maintenance now
performed by most users does little to either assure valid data or determine
its accuracy.
6.2 INSTRUMENT CAPABILITY
The accuracy which should be expected from NMHC analyzers in
current use when optimum user operating techniques are carried out will
now be considered. In obtaining user data for the Scott cylinders, all
users first spanned the instruments and performed other standard maintenance.
Thus, the following discussion, which covers items such as instrument
drift, involves errors separate from those identified previously.
The reference method for determination of hydrocarbons corrected
for methane (Federal Register, Vol. 36, No. 228, pp 22394-6) contains suggested
performance specifications for NMHC analyzers. The performance items which
most affect NMHC data accuracy are zero drift and span drift. The speci-
fications state that the zero and span drifts shall not exceed 1 percent
of full scale per 24 hours. Our survey showed that instruments are typically
operated on the 0-20 ppm range. Thus a 1% drift would be equivalent to 0.2
ppm. Since the zero and span drifts could be in opposite directions and could
result in a net positive error for methane and a net negative error for total
hydrocarbons, the NMHC data at 1.5 CH +0.24 NMHC could be in error by more
than 0.4 ppm-C after 24 hours even if the instrument met the suggested
standards. This is, of course, an extreme and probably unlikely case. A
more realistic drift effect can be ascertained from an examination of actual
user records.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-27-
SET 1385 06 1274
Various users performed span checks and other routine maintenance
at varying frequencies ranging from daily to weekly. Even where daily
checkouts were scheduled, none were performed over the weekend, so a three-day
unattended operating period occurred. Some idea of the real effect of
drift was obtained from data supplied by User 3 for daily span values. The
practice was to analyze several samples of span gas and then adjust the
instrument gain to produce a specified peak voltage for each component if
necessary. The data shows that for the month of December 1973, the drift
was as high as 0.4 ppm (2%) for one day and 0.8 ppm (4%) over three-day
weekends. The methane drift was usually downward and it was greater than
the THC drift which was usually upward. The net effect produced apparent
NMHC values ranging from -0.3 to +1.1 ppm for the span gas which contained
7.5 ppm methane and 0.0 ppm NMHC. The average error was +0.5 ppm NMHC. i'ne
error at the ambient level being measured cannot be defined because the zero and
span drifts were not checked separately. Neverthless, it is clear that this instru-
ment, which had been in use for approximately two months, had drift in excess of
the 1%/day given in the suggested specifications. It is also clear that NMHC data
collected twenty-four hours after a calibration check would be subject to sub-
stantial error in addition to the error determined using the cylinder mixtures.
The minimal operational period listed in the suggested performance
specifications is three days. While the Scott field evaluation program did
not include sufficient observations to define data accuracy after various
intervals of unattended operation, there is considerable evidence which
indicates that after twenty-four hours of unattended operation following
optimum calibration procedures, confidence in the accuracy of the NMHC
data would not be adequate to determine compliance with the 0.24 ppm ambient
standard. That is, the error in NMHC due to twenty-four hour zero and span
drift could in many cases be equal to or greater than the standard value
itself.
The performance tests suggested in Section 6.1 would permit the
users to determine data accuracy after various periods of unattended operation
prior to instrument adjustment. While this would not improve the accuracy
of prior data it would show whether or not it was valid as well as indicate
the frequency of checkout required to assure a specified data accuracy.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
-28-
SET 1385 06 1274
Other suggestions which would alleviate present instrument short-
comings include scheduled checkout just prior to the 6 a.m. to 9 a.m. period
to which the standard applies, and automatic injection of zero and span
gases at hourly intervals. This latter feature is currently available in
other instruments used for aerometric and source emission measurements.
In some systems automatic electronic adjustments for zero and span drift
are made. In order to determine the need for these features, the degree
of data accuracy required to determine compliance with the standards must
first be defined. Then the capability of the various NMHC analyzers on
the market for providing the required accuracy must be determined through
performance tests. If available instruments do not provide sufficient
accuracy, the addition of automatic zero and span gas injection systems would
probably increase accuracy to the required level.
6.3 PERFORMANCE OF OTHER TYPES OF NMHC INSTRUMENTS
In a few cases users had identified instrumentation other than
that of the GC type as meeting the reference method. The usual method
involved the use of charcoal scrubbers to remove non-methane hydrocarbons
from the air stream. Two continuous analyzers of this type were evaluated
briefly in the field program. In both cases the results showed completely
unsatisfactory performance. The large amount of instrument noise and
drift made it impossible to determine when equilibrium to the Scott gases
had been reached. The operators appeared to be aware that the instruments
were obsolete and unreliable, and they were doing little to improve
instrument operation.
A number of users reported that they were determining NMHC with a
combination of a continuous FID analyzer for THC with a separate GC for
methane on a periodic basis. None of these combinations were evaluated.
SCOTT ENVIRONMENTAL TECHNOLOGY, INC
-------
-29-
SET 1385 06 1274
7.0 CONCLUSIONS AND RECOMMENDATIONS
7.1 CONCLUSIONS
The information gathered in the field evaluation of non-methane
hydrocarbon (NMHC) analyzers shows that ambient air NMHC data being collected
by the large majority of organizations is subject to substantial errors,
especially at low concentrations in the vicinity of the ambient air standard
of 0.24 ppm-C. The inaccuracies in the data make it impossible to determine
whether the ambient air quality is in compliance with the standard.
The major factors contributing to these errors include:
1. Failure of operators to understand and/or follow the instrument
manufacturers' operating instructions and the reference method
procedures for NMHC as published in the Federal Register.
2. Span gases containing unknown amounts of higher hydrocarbons.
3. Span gases not in air.
4. Span gases incorrectly analyzed for methane.
5. Zero errors due to sampling system contamination and lack of
adequate checkout procedures.
6. Excessive instrument zero and span drift during unattended
operation.
Despite the above shortcomings, the NMHC data from the gas chrcna-
tographic type analyzers specified in the reference method appear to be more
accurate than NMHC data obtained by other types of instruments and procedures
in general use. Thus, improvements in operating techniques and in performance
of existing instruments would seem to be the best approach to achieving the
desired data quality. At the same time the full capability of recently
developed NMHC instruments should be determined, and their potential for
providing the necessary data accuracy should be evaluated at an early date.
7.2 RECOMMENDATIONS
In order to improve NMHC data accuracy to an acceptable level, we
offer the following recommendations:
1. Develop simplified operating procedures which can be under-
stood and followed by operators without extensive training.
The procedures should include regularly scheduled performance
tests to define data accuracy and indicate when more extensive
adjustments are necessary.
SCOTT ENVIRONMENTAL TECHNOIOGY, INC.
-------
-30-
SET 1385 06 1274
2. Develop reference standards for methane, THC and zero air. Specify
that all calibration gases be analyzed for methane and THC
through comparison to the reference standards.
3. Define instrument performance required to provide data
sufficiently accurate to determine compliance with ambient
air standards; revise reference method to include these
performance standards.
4. Evaluate capability of existing instruments to meet revised
performance standards; explore improvements in any areas not
meeting standards.
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
-31-
SET 1385 06 1274
8.0 ACKNOWLEDGEMENTS
We wish to express our thanks to the Project Officer, Mr. John
H. Margeson and to Dr. John B. Clements, Chief of the Methods Standardization
Branch, Quality Assurance and Environmental Monitoring Laboratory for their
assistance in the planning and performance of the survey of users of NMHC
analyzers.
We are also deeply indebted to those sixteen users who voluntarily
participated in the field evaluation. These organizations and individuals
impressed us by their fine cooperation and their deep interest in the quality
of their data and means for improving it. Although their names cannot be
listed here because of guarantees of user anonymity, it must be recognized
that without their willingness to cooperate, Scott would have been unable to
perform the major portions of the program.
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
APPENDIX
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-l
SET 1385 06 1274
TABLE A-l USERS OF NON-METHANE HYDROCARBON ANALYZERS
FOR AMBIENT MONITORING
(Gas Chromatographic Type per E.P.A. Reference Method)
Type of Instrument
Nn.
Name
Mobile County
Board of Health
Jefferson County
Board of Health
Huntsville Air
Pollution Control
Dept.
Alabama Dept. of
Public Health
Location
Alabama
Mobile
Birmingham
Huntsville
Montgomery
Arizona
Union
Beckman Bendix Bendix Byron Carbide
6800 8200 8201 200 3020 MSA AID
Phelps Dodge Corp. Phoenix
Arizona Div.of Air
Pollution Control Phoenix
Calif.State Air
Resource Board
Calif.Div. of
Highways
San Diego APCD
Orange Co. APCD
Ventura Co. APCD
Environmental
Systems Laboratory
Bay Area APCD
Univ. of Calif.
Datronics Systems
Corporation
Kern Co. APCD
Metronics
Associates Inc.
Santa Barbara APCD
California
Sacramento
Sacramento
San Diego
Anaheim
Ventura
2
3
1
1
Sunnyvale 1
SanFrancisco 13
Riverside
Panorama City 1
Bakersfield 3
Palo Alto
Santa Barbara 1
<$>
SCOU ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-2
SET 1385 06 1274
TABLE A-l USERS OF NON-METHANE HYDROCARBON ANALYZERS
FOR AMBIENT MONITORING
(Gas Chromatographic Type per E.P.A. Reference Method)
(contd)
Type of Instrument/Model No.
Name
Materials &
Research Lab.
Nat'l Ctr for
Atmospheric Res.
Location
Sacramento
Colorado
Boulder
Connecticut
Union
Beckman Bendix Bendix Byron Carbide
6800 8200 8201 200 3020 MSA AID
Air Compliance Dept.
of Environmental
Protection Hartford
Florida
Florida Dept. of
Pollution Control Tallahassee
Hillsborough Co.
Poll.Cont.Comra.
Palm Beach Co.
Health Dept.
Tampa
Riviera Bea.
Hawaii
Honolulu
Illinois
Air Resources Inc. Palatine
Springfield
Chicago
Northbrook
Hawaii State
Dept. of Health
Illinois Div. of
Air Poll.Cont.
NALCO Chemical
Corp.
Industrial Bio-
Test Labs.
Chicago Dept. of
Environmental Cont.Chicago
1
1
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-3
SET 1385 06 1274
TABLE A-l USERS OF NON-METHANE HYDROCARBON ANALYZERS
FOR AMBIENT MONITORING
(Gas Chromatographic Type per E.P.A. Reference Method)
(contd)
Type of Instrument/Model No.
Name
Evansville Air
Pollution Control
Department
State Dept. of
Health
Linn Co.
Dept. of Health
NUS Corp.
Bait.Co. Dept.
of Health
Geomet Inc.
Green Assoc.Inc.
Location
Indiana
Evansville
Iowa
DesMoines
Cedar Rapids
Maryland
Rockville
Towson
Rockville
Towson
Beckman
6800
Union
Bendix Bendix Byron Carbide
8200 8201 200 3020 MSA AID
3
1
Environmental
Research & Tech.
GM Tech.Ctr.
Massachusetts
Cambridge
Michigan
Warren
Mississippi
Mississippi Air &
Water Pollution Control
Commission Jackson
Missouri
Midwest Res.Inst. KansasCity
Missouri Air Jefferson
Conservation Comm. City
St.Louis Co.
Health Dept. Clayton
Monsanto Enviro-Chem.
Systems Inc. St.Louis
Kansas City Air
Poll.Cont.Div.
1
8
1
1
KansasCity
SCOn ENVIRONMENTAL TECHNOIOGY, INC.
-------
A-4
SET 1385 06 1274
TABLE A-l USERS OF NON-METHANE HYDROCARBON ANALYZERS
FOR AMBIENT MONITORING
(Gas Chromatographic Type per E.P.A. Reference Method)
(contd)
Type of Instrument /Model No.
Name
Environmental
Testing Inc.
Univ. of NC
Location
Beckman
6800
Union
Bendix Bendix Byron Carbide
8200 8201 200 3020 MSA AID
North Carolina
Charlotte 1
Chapel Hill
New Jersey
Trenton
Exxon Res. & Eng. Linden
Foster D.Snell,
NJ Bur. of Air
Pollution Cont.
Inc.
Florham Park
New York
NY State Dept. of
Environmental
Conservation
Erie Co.
Dept. of Health
Nassau Co.
Health Dept.
Union Carbide
Herron Testing
Labs., Inc.
General Elec. Co.
Oregon Dept. of
Env i r onmen t a1
Quality
Albany
Buffalo
Minneola
Ohio
Cleveland
Cleveland
Cleveland
Oregon
Portland
1
1
6
1
1
1
1
SCOn ENVIRONMENTAL TECHNOLOGY. INC.
-------
A-5
SET 1385 06 1274
TABLE A-l USERS OF NON-METHANE HYDROCARBON ANALYZERS
FOR AMBIENT MONITORING
(Gas Chromatographic Type per E.P.A. Reference Method)
(contd)
Type of Instrument /Model No.
Location
Name
Phila.Dept. of
Public Health
Betz Environmental Plymouth
Engineering Meeting
General Elec.Co. Valley Forge
Union
Beckman Bendix Bendix Byron Carbide
6800 8200 8201 200 3020 MSA AID
Pennsylvania
Philadelphia 1
17
Environmental
Sciences Inc.
Charleston County
Health Dept.
Texas Air Poll.
Cont. Service
City of Dallas
Health Dept.
Univ.of Salt
Lake City
Richmond Research
Labs.
Wash.State Univ.
Union Carbide Co.
W.Va.Air Poll.
Control Comm.
Wise.Dept. of
Natural Res.
Total
Pittsburgh
South Carolina
Charleston 4
Texas
Austin 5
Dallas
Utah
Salt Lake
City
Virginia
Richmond
Washington
Pullman 1
West Virginia
S.Charleston
Charleston
Wisconsin
Madison 4
109
9
1
1
1
8 18
14
31
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-6
SET 1385 06 1274
FIGURE A-l NON-METHANE HYDROCARBON ANALYZER
ON-SITE EVALUATION CHECK LIST
Laboratory Name: Code #: Director's Name:
Address: Operators' Name:
Telephone:
1.0 Non-Methane Hydrocarbon Analyzer
1.1 Manufacturer:_ Mfg. No .:_
1.2 Model No.: Date of Purchase:
1.3 Type of organization using this instrument: Federal agency
State agency , Local air pollution agency , University
Private Contractor , Other . If other, explain:
1.4 Frequency and type of use:
1.5 Training level of operator: Education:
Experience:____ .
1.6 Extent to which the operator follows the recommendations of the
instrument manufacturer and good laboratory procedure in his work:
1.7 General Laboratory Condition:
1.8 Other instruments used in the laboratory for air monitoring
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-7
SET 1385 06 1274
2.0 Performance of Instrument
2.1 Frequency and nature of breakdowns:
2.2 Parts repaired or replaced:_
2.3 Serviced by: Laboratory personnel , and/or Manufacturer's
representative .
2.4 Preventive maintenance program consists of:
2.5 Estimated service maintenance costsj
3.0 Instrument Operation
3.1 Physical condition of the instrument:_
3.2 Type of carrier gas (N2, H2> He, or Air) cc/min
3.3 General Operating Conditions:
Combustion gas: Oxygen ( ), Air ( ). Rate: cc/min.
Fuel type: Rate: cc/min Temp.
Frequency of sampling; Sample size:_
Frequency of calibration;
Operating cycle:
Calibration gas used; Range used:_
Zero gas used;
Calibration technique:
Sample flow rate on bypass:_
Sampling Rate:
r
SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-8
SET 1385 06 1274
3.4 Type of regulatory
Name:
Vendor:
3.5 Gas vendor:_ ___
3.6 Please attach a calibration curve with data points.
3.7 Instrument Checks:
3.7.1 Check user's instrument accuracy at ambient levels equivalent to
standard and ambient levels typically found in urban areas. Meth-
ane at typical ambient level + C2 to C& saturates and ethylene at
approximately 0.24 ppm-C:
Results: - —
Comments:
3.7.2 Check methane at typical ambient level + C^ to C& saturates and
ethylene at approximately 2.4 ppm-C:
Results:
Comments:
3.7.3 Relative response of user's instruments to higher hydrocarbons
versus methane using C2 - C saturates and ethylene at approximately
2.4 ppm-C:
Results: ——
; SCOTT ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-9
SET 1385 06 1274
Comments:
3.7.4 Initial check of user's zero gas for contamination with both
methane and other hydrocarbons using Scott's hydrocarbon-free
air:
Results:
Comments:
3.7.5 Check user's span gas using methane in hydrocarbon-free air, at
instrument span inlet and sample inlet.
Results:
Comments:
3.7.6 The precision will be determined using the user's span gas to get
a degree of agreement between repeated measurements of the same
concentration. This will be expressed as the average standard
deviation of the single results from the mean using an average
of five injections.
Results:
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
A-10
SET 1385 06 1274
Comments:
3.7.7 The linearity of the recorder will be measured by using a single
source to drive the recorder in equal millivolt increments from
zero input to full scale response.
Results: ...
Comments:
3.7.8 Noise determination will be obtained by operating the instrument
at attenuation settings normally used for ambient monitoring. Noise
will be expressed as percent of full scale. It will also be noted
if the instrument is used in a lower special analysis range.
Results:
Comments:
3.7.9 Zero drift determination will be the changes noted in instrument
output over a 16-hour period (overnight) of unadjusted continuous
operation. Zero drift will be expressed in percent of full scale.
Results: _
Comments:
SCOTT ENVIRONMENTAL TECHNOIOGY, INC.
-------
A-ll
SET 1385 06 1274
4.0 Lab opinion of the analyzer:
4.1 Strong points:
4.2 Weak points:
4.3 What would the laboratory personnel like to see changed or improved
on their analyzer? . .—__
5.0 Evaluator's opinion of the analyzer:
5.1 Strong points:
5.2 Weak points:
5.3 How could the system be changed or improved to give better
results? __^_
NOTE: The anonymity of the reporting laboratory shall be maintained throughout
this study.
SCOn ENVIRONMENTAL TECHNOLOGY, INC.
-------
TECHNICAL REPORT DATA
EPA 650/4-75-008
3. RECIPIENT'S ACCESSIOr*NO.
Survey of Users of the EPA - Reference Method
for Measurement of Non-Methane Hydrocarbons in
Ambient Air
5. REPORT DATE
1974
6. PERFORMING ORGANIZATION CODC
7. AUT-C.fM.il
Louis R. Reckner
S. PERFORMING ORGANIZATION REPORT NC
9. PEF-.FCf.;.'!NG ORGANISATION NAME ATJD ADDRESS
Scott Environmental Technology, Inc.
Plumsteadville, PA 18949
10. PROGRAM ELEMENT NO.
1HA327
11. CONTRACT. GRANT NO.
68-02-1206
12.
AGE\'CV
AND ADDRESS
13. TYPE OF REPORT AND PER.OD COVE HIP
Office of Research and Development
U.S. Environmental Protection Agency
Washington, D.C. 20460
14. SPONSORING AGENCY CODE
15. surr LL:.-.L\TARY NOTES
IG. ADSI r.^,.1 scot-t performed a survey of users of the EPA Reference Method for measurement
of non-methane hydrocarbons in ambient air which resulted in the compilation of a list
of 188 NMHC analyzers operated by 70 organizations. Field evaluations were performed
on instruments operated by 16 of the users..
The accuracy of the NMHC data being obtained by the 16 users of the reference method
was determined by presenting a series of 5 gas mixtures in high-pressure cylinders for
analysis by each operator. The results for the mixture containing NMHC at a concentra-
tion close to the 0.24 ppm-C ambient air standard showed that substantial errors existec
in current NMHC data. The errors are summarized below:
Error Range Number of Users
0-10% 1
10-20% 3
20-50% 2
50-100% 4
> 100% 6
An analysis of the data showed that the inaccuracies in current data make it
impossible to determine whether ambient air quality is in compliance with the standard.
The major factors contributing to data errors are discussed, and recommendations for
improving data quality are presented.
17.
KEY WORDS AND DCCUV.ENT ANALYSIS
DESCRIPTORS
(XIDENTIFIERS-OPEN ENDED TERVS
C. COSA11 i K'J
Air Pollution
Methodology
Measurement
Field Evaluation
EPA Reference Method
Non-Methane Hydrocarbon
13B
7C
Unlimited
19. SECURITY CLASS (i'tos
Unclassified
21. NO. L
42
20. SECURITY CLASS (
Unclassified
22. PRICE
EPA Form 2220-1 (9-731
------- |