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

-------