r/EF>A
                 United States
                 Environmental Protection

                 Agency
                              EPA-600A-8I-015

                              March 1931
Research and


Development

TECHNICAL ASSISTANCE DOCUMENT FOR
THE CALIBRATION AND OPERATION OF
AUTOMATED AMBIENT NON-METHANE ORGANIC
COMPOUND ANALYZERS          . >
IEPA
leoo/
U-
 81-015
                          \
                                               US
                 Prepared for

                 OFFICE OF AIR QUALITY PLANNING AND STANDARDS
                                      US-
                                                  **
                                                     »
                 Prepared  by


                 Environmental Monitoring Systems
                 Laboratory
                 Research Triangle Park NC 27711

-------

-------
                                 DISCLAIMER

     This report has been reviewed by  the Environmental Monitoring  Systems
Laboratory, U.S. Environmental Protection Agency,  and  approved  for
publication.  Mention of trade names or commercial  products  does  not
constitute endorsement or recommendation for  use.

-------

-------
                                                                    EPA 600/4-81-015
                                                                    March 1981
              TECHNICAL ASSISTANCE DOCUMENT FOR THE
              CALIBRATION AND OPERATION OF AUTOMATED
       AMBIENT NONMETHANE ORGANIC COMPOUND ANALYZERS
                                       by
          Frederick W. Sexton
         Raymond M. Michie, Jr.
Environmental Quality Assurance Department
        Research Triangle Institute
     Research Triangle Park, NC 27709
         Frank F. McElroy
       Vinson L. Thompson
  Methods Standardization Branch
U.S. Environmental Protection Agency
  Research Triangle Park, NC 27711
                        Contract Nos. 68-02-3222, 68-02-3431
                         EPA Project Officer: Frank F. McElroy
                             Quality Assurance Division
                      Environmental Monitoring Systems Laboratory
                         U.S. Environmental Protection Agency
                           Research Triangle Park, NC 27711
                            This document was prepared
                                in cooperation with
                            the Research Triangle Institute
                          Research Triangle Park, NC 27709
          U.S. EPA Headquarters Library
          1PnnD Ma" code 3201
          1200 Pennsylvania Avenue NW
            Washi     DC 20460
                      Environmental Monitoring Systems Laboratory
                         Office of Research and Development
                         U.S. Environmental Protection Agency
                           Research Triangle Park, NC 27711

-------

-------
                                  FOREWORD

     Measurement and monitoring research  efforts  are  designed to anticipate
potential environmental problems, to  support  regulatory actions by develop-
ing an in-depth understanding of  the  nature and processes  that impact health
and the ecology, to provide  innovative  means  of monitoring compliance with
regulations, and to evaluate the  effectiveness of health and environmental
protection efforts through the monitoring of  long-term trends.  The
Environmental Monitoring Systems  Laboratory,  Research Triangle Park, North
Carolina, has responsibility for:   assessment of  environmental monitoring
technology and systems; implementation  of Agency-wide quality assurance
programs for air pollution measurement  systems; and supplying technical
support to other groups in the Agency,  including  the  Office of Air, Noise
s.nd Radiation, the Office of Toxic  Substances, and the Office of
Enforcement.

     Although monitoring of non-methane organic compounds  (NMOC) is not
required, per se, by present compliance monitoring regulations, NMOC
monitoring data are vitally  important in some areas for planning hydrocarbon
control strategy necessary to achieve the National Ambient Air Quality
Standard for ozone.  Prepared at  the  request  of  the Office of Air Quality
Planning and Standards (Air, Noise  and  Radiation), this document provides
technical information and quality assurance procedures which should prove
highly beneficial to control agencies making  ambient  NMOC measurements with
automated NMOC analyzers.
                                           Thomas R.  Hauser, Ph.D.
                                                   Director
                                 Environmental Monitoring Systems Laboratory
                                    Research Triangle Park, North Carolina
                                     iii
U.S. EPA Headquarters Library
      Mail code 3201
1200 Pennsylvania Avenue NW
   Washington DC 20460

-------
iv

-------
                                   PREFACE

     Although this document details problems  and  difficulties  associated
with collecting continuous non-methane  organic  compounds  (NMOC)  monitoring
cata, it does so to accent the need for  improved  quality  control and better
use of good laboratory  procedures  while  operating automated NMOC analyzers.
Comments contained herein are not  intended  to reflect  invidiously on the
analyzer manufacturers, authors  of instruction  manuals,  analyzer operators,
or authors of any documentation  evaluating  NMOC analyzers.
                                       v

-------
                                  ABSTRACT

     This technical assistance document  is  intended  to  serve  as  a supplement
to manufacturers' instruction manuals  for automated,  ambient,  non-methane
organic compound analyzers.  It addresses augmented  setup,  calibration,
operation, and maintenance procedures  that  have  been  developed for the
purpose of producing non-methane organic compound  data  suitable  for use  in
the Empirical Kinetic Modeling Approach  and other  dispersion  models used to
predict ozone concentrations.  The  document initially discusses  common
aspects of these procedures  in general terms for hydrocarbon  analyzers as a
whole and then specifically  addresses  their application to  analyzers
manufactured by The Bendix Corporation,  Mine Safety  Appliances Company,  and
Beckman Instruments, Inc.  Comments  also address analyzers  manufacturered by
Byron Instruments, Inc., and Meloy  Laboratories, Inc.
                                     vi

-------
                                   CONTENTS

foreword	,	ill
Pre face	   V
Abstract	  VI
Figures	   X
Tables	  xl
Abbreviations and Symbols	.	XI i
Acknowledgments	xii 1

L.   INTRODUCTION	   1
     1.1  Hydrocarbon chemistry  and measurement	   1
     1.2  Continuous and discrete  measurements	   3
     1.3  Purpose	   3
:».   GENERAL OPERATING PROCEDURES	   6
     2.1  Introduct ion	. . ..	   6
          2.1.1  Site selection	   6
          2.1.2  Analyzer  environment  control	   6
          2.1.3  Ambient air  sampling  system	   6
     2.2  Installation of  NMOC Analyzers	   7
          2.2.1  Unpac king	   7
          2,2.2  Analyzer  instruct ion  manual	   8
          2.2.3  Recordkeeping	   8
          2.2.4  Setup.	  15
     2 .3  Operat ion	  16
          2.3.1  Start-up	  16
          2.3.2  Periodic  calibrations and  zero/span	  17
                 checks	  17
          2.3.3  Operational  checks	  17
          2 .3 .4  Preventive maintenance	  17
     2 .4  Troubleshooting	  18
          2,4.1  Flow measurements	  18
          2.4.2  Leak-checks	  19
          2.4.3  Sample pump  checks	  21
3 .   CALIBRATION STANDARDS, EQUIPMENT, AND  PROCEDURES	  22
     3.1  Introduction	  22
     3.2  Zero air standards	  23
     3 .3  Span gas standards	  24
     3.4  Effects of moisture	  26
     3.5  Calibration gas  dilution  system	  26
     3 .6  Flow measurements	  26
     3.7  General multipoint  calibration procedure	  28
     3.8  Level 1 zero and  span  check	  30
     3.9  Level 2 zero and  span  check	  31
                                     VII
                                                      U.S. EPA Headquarters Library
                                                            Wai! COGS 3201
                                                      120C Pennsylvania Avenue NW
                                                        W&snington DC 20460

-------
4.   AUGMENTED PROCEDURES FOR THE MSA 11-2 CONTINUOUS
     NMOC/METHANE MONITOR	  32
     4.1  Principles of operation	  32
     4.2  System description	  32
     4.3  Modifications	  35
     4.4  Installation	  37
     4. 5  Operat ion	  38
          4.5.1  Start-up	  38
          4.5.2  Periodic calibrations and zero/span checks	  38
          4.5.3  Operat ional checks	  40
          4.5.4  Moisture trap check	  40
          4.5.5  Cutter temperature	  40
     4.6  Cal ibrat ion	  42
          4.6.1  Multipoint calibration	  42
          4.6.2  Level 1 zero/span check	  47
     4.7  Troubleshooting.	  49
          4.7.1  Preliminary investigat ions	  49
          4.7.2  Sample pump system investigation	  51
          4.7.3  Moisture trap investigation	  52
          4.7.4  Hydrogen system investigation	  52
          4.7.5  FID response investigation	  53
          4.7.6  Sample flow rate investigation	  56
          4.7.7  Burner air flow rate investigation	  56
5.   AUGMENTED PROCEDURES FOR THE BENDIX 8201 AND 8202
     REACTIVE HYDROCARBON ANALYZER	  58
     5.1  Principles of operation	  58
     5.2  System description	  58
     5.3  Modifications	  59
     5.4  Installation	  61
     5 .5  Operat ion	  62
          5.5.1  Start-up	  62
          5.5.2  Periodic calibrations and zero/span checks	  62
          5.5.3  Operat ional checks	 •  63
          5.5.4  Routine chromatograms	  63
     5 .6  Cal ibrat ion	  63
          5.6.1  Multipoint calibration	  63
          5.6.2  Level 1 zero and span check	  68
     5.7  Troubleshooting	  69
          5.7.1  Preliminary investigations	  69
          5.7.2  Chromatogran investigation	  70
          5.7.3  Sample and back-flush valve leak-check	  80
6.   AUGMENTED PROCEDURES FOR THE BECKMAN 6800 AIR QUALITY
     CHROMATOGRAPH (Cfy, TOC, NMOC)	  82
     6.1  Principles of operation	  82
     6 .2  System description	  82
     6.3  Modifications	  84
     6.4  Installation	  85
     6 .5  Operat ion	  86
          6.5.1  Start-up	  86
          6.5.2  Periodic calibrations and zero/span checks	  88
          6.5.3  Operat ional checks	  88
                                   vlii

-------
          6.5.4  Routine chromatograms	  90
     6.6  Calibration	  93
          6.6.1  Multipoint calibration	  93
          6.6.2  Level 1 zero and span check	 100
     6.7  Troubleshooting	 100
          6.7.1  Preliminary investigations	 101
7.   COMMENTS ON THE BYRON 233 THC/CH4/CO ANALYZER	 102
     7.1  Principle of operation	 102
     7.2  General comments	 102
a.   GENERAL COMMENT ON THE MELOY HC500-2C FID HYDROCARBON
     ANALYZER	 104
9.   References	 105

-------
                                  FIGURES
Number
Page
  1   Leak-check apparatus	   20

  2   Diagram of a zero air system	   24

  3   Diagram of a dilution system	   27

  4   Recommended water trap system  	   37

  5   CH4 channel's response curves - effects of
      variations in H2 flow rate	   55

  6   TOC channel's response curves - effects of
      variations in H2 flow rate	   55

  7   Chromatogram from a. Bendix 8202 analyzer showing
      proper analyzer ope rat ion	   72

  8   Chromatogram from a Bendix 8202 analyzer showing
      improper gate timing and baseline shift	   72

  9   Chromatograms showing analyzer malfunctions	   76

-------
                                   TABLES
Number
                                                                  Page
  3

  4
  8
 10
 11
 12
 13
      Summary of problems associated with  gathering NMOC
      data with automated analyzers	   4

      Summary of recommendations to reduce  the  effect  of
      problems lisced in Table 1	   4
Form for recording analyzer  information,
10
Typical form for recording routine operational data
from an NMOC analyzer	  11
  5   Form for recording calibration data	  12

  6   Form for recording zero and span data	  13

  7   Form for recording a routine maintenance schedule	  14
Listing of methane and propane NBS/SRMs available
in compressed gas cylinders	 23
      Nominal operating specifications  for the MSA  11-2 NMOC
      Analyzer	
                                                                  39
Form for recording routine operational data  from the MSA
11-2 NMOC analyzer	41

Form for recording routine operational data  from the Bendix
8201 or 8202 NMOC analyzer	 64

Comparator card 1 (Component card J-10) functions
and potentiometer controls	 74
Form for recording routine operational data  from the
Beckman 6800 NMOC analyzer	
                                                                        89
                                     Xi

-------
                         ABBREVIATIONS AND  SYMBOLS
ABBREVIATIONS

CRM
DAS
DVM
SKMA
EPA
FID
kPa
LOX
NAAQS
NBS
NMHC
NMOC
psig
ppmC
ppinV
RH
SBFM
SRM
TAD
TNMHC
TOG
URL
url
WTM

SYMBOLS

CH4
C2H4
C3H8
C02
H2
H20
NO 2
03
RCHO
R2CO
SO 2
certified reference material
data acquisition system
digital voltmeter
empirical kinetic modeling approach
U.S. Environmental Protection Agency
flame ionization detector
kilo-Pascals
liquid oxygen
National Ambient Air Quality Standard
National Bureau of Standards
non-methane hydrocarbon
non—methane organic compound
pounds per square inch
parts per million as carbon
parts per million by volume
reactive hydrocarbon
soap bubble flow meter
standard reference material
technical assistance document
total non-methane hydrocarbon
total organic compound
upper range limit, analyzer
upper range limit, recording device
wet test meter
methane
ethylene
propane
benzene
carbon dioxide
hydrogen
water
nitrogen dioxide
ozone
aldehyde
ketone
sulfur dioxide
                                   xii

-------
                              ACKNOWLEDGMENTS

     The authors of this document gratefully acknowledge  the  generous  assis-
tance of Mr. A. Poli of Mine Safety  Appliance  Company,  Mr.  R.  Villalobos of
Beckman Instruments, Inc., Mr. G. Funderbunk of  the  Bendix  Corporation,  and
Mr. B. Behr of Byron Instruments, Inc.,  who all  provided  technical  details
and suggestions, many of which would have been unavailable  elsewhere.

     Also, we acknowledge  the invaluable contributions  of Dr.  C.  Eaton,  Dr.
J. Sickles, Mr. A. Gaskill, and Ms.  S. Powell  of  the Research  Triangle
Institute; Dr. H. Richter, Dr. D.  Mage, Dr. B.  Dimitriades,  and  Messrs. L.
?urdue, G. Ortmon, and W.  Lonneman of EPA; and others who technically
reviewed the manuscript and helped make  the preparation and distribution of
:his document possible.
                                    xiii

-------

-------
                                  SECTION  1

                                INTRODUCTION
1.1  HYDROCARBON CHEMISTRY AND MEASUREMENT

     Organic compounds containing  only  the  elements  carbon  and  hydrogen are
called hydrocarbons.  These elements  can be  formed  into  open  chain  aliphatic
molecules (alkanes, alkenes,  alkynes) or closed  ring molecules  (alicycli.cs,
aromatics).  Hydrocarbon compounds  are  frequently classified  as  paraffins,
olefins, or aromatics.  Paraffins  are aliphatic  compounds  like  propane
(C3Hg) which contain no unsaturated  (multiple  bonded)  sites.  Olefins
are also aliphatic compounds, but  contain one  or more  multiple  bonds,  as  in
ethylene (0284). Aromatics are unsaturated  ring-like structures  in  which
the benzene ring (CgHg) is the parent structure.

     Methane (Cfy) is the simplest  and  most  abundant hydrocarbon in the
ambient air, and at common ambient  concentrations it is  not harmful to  human
health.  Attention is directed primarily at  the  non-methane hydrocarbon
compounds because these compounds  are involved in photo-chemical chain
reactions, resulting in an accumulation of  photo-chemical  oxidants  in  the
atmosphere.  Methane is not sufficiently reactive to participate appreciably
in these photo-chemical reactions  and can therefore  be subtracted from  the
total hydrocarbon concentration  to  obtain a  more accurate measure of
reactive hydrocarbon concentrations.  (A detailed discussion  of  the
hydrocarbon/oxidant relationship is  provided in  references  1  and 2.)

     The concentration resulting from the subtraction  of the  methane
concentration from total hydrocarbon  concentration has been referred to in
the past as reactive hydrocarbon (RH),  non-methane hydrocarbon  (NMHC),  and
total non-methane hydrocarbon (TNMHC).  A more precise reference is
non-methane organic compounds (NMOC)  because it  includes oxygenated organics
such as aldehydes (RCHO) and  ketones  (R2CO)  that also  contribute to the
buildup of photochemical oxidants.   In  addition, the flame  ionization
detector (FID) used to measure these  compounds exhibits  some  sensitivity  to
oxygenated compounds.  Therefore,  the designation NMOC is  recommended when
referring to organic compounds detectable by flame  ionization that
contribute to the buildup of  photochemical  oxidants.

     Concentrations of gaseous hydrocarbons  have historically been  reported
in various units such as parts per  million  as  carbon (ppmC),  micrograms per
cubic meter (Ug/m^) as methane,  parts per million by volume as methane

-------
(ppmV as CH^), and others.  To  eliminate  Che  confusion  of  using  several
units, and to normalize the response of the FID  to  different  organics,
hydrocarbon concentrations should be reported as  ppmC.   This  unit  is
obtained by multiplying the ppmV concentration of the hydrocarbon  compound
by the number of carbon atoms in that compound.   For example,  3.0  ppraV of
propane (C^Hg) is equal to 9.0  ppmC.

     A National Ambient Air Quality Standard  (NAAQS) for hydrocarbons
corrected for methane was promulgated by  the  Environmental Protection
Agency in 1971 (3).  This NAAQS was unique among  the standards promulgated
at that time because it was not based upon direct harmful  effects  to human
health or welfare caused by NMOC.  Rather, the standard was based  upon  the
role NMOC plays as a precursor  to the formation  of  photochemical oxidants
such as ozone.  The standard was intended to  serve  as a guide for  attaining
the ozone NAAQS (through control of NMOC  emissions) and not as a standard
for compliance.  More recently, photochemical models have  been developed
which more accurately describe  the relationship  between NMOC  concentrations
and subsequent peak ozone levels.  Consequently,  the original  NAAQS  for NMOC
is no longer used.

     Historically, ambient hydrocarbon concentrations have been  measured  by
an FID.  In the continuous (automated) NMOC analyzer, the  sample is  injected
into a stream of hydrogen gas or zero air, and directed to a  burner  where
combustion occurs.  The resulting ionization  of  the hydrocarbons in  the
sample creates an electric current roughly proportional to the carbon
concentration of the sample.  The FID is  used because it responds  to a wide
range of hydrocarbon compounds; is relatively insensitive  to  non-
hydrocarbons; has a wide, linear dynamic  range;  and is  suitable  for
continuous field operations.

     Although the detector responds linearly  to  varying concentrations of
individual hydrocarbons, it does not respond  uniformly  from one  hydrocarbon
species to another.  For example, the response of a methane-calibrated FID
to a sample of 9.0 ppmC of methane would  be 9.0  ppmC, whereas  its  response
to a sample of 9.0 ppmC of propane may be only approximately  6.3 ppmC.  This
non-uniform per carbon response characteristic is apparently  related to the
presence of the oxygen in the sample air.  The FID's response  to different
compounds is much more uniform  when the compounds are separated  from the  air
sample by a chromatographic column, and are injected into  the  FIB  via an
inert carrier gas.

     In continuous NMOC analyzers, which  have no  columns to separate the
hydrocarbons from the air, or in GC analyzers which use an air carrier, this
non-uniform response to different organic compounds is  a problem when
attempts are made to measure the concentrations  of  unknown organic
compounds.  One method to help  compensate for this  effect  is  the judicious
choice of the compound used for calibration.  Accordingly,  propane is
recommended for the calibration of automated  NMOC analyzers.   Section 3.1
discusses this recommendation in more detail.  Further  information is also
contained in reference 4.

-------
:. .2  CONTINUOUS AND DISCRETE  MEASUREMENTS

     Current analytical methods  for  obtaining NMOC measurements of ambient
air can be classified  as  either  discrete or continuous.   Discrete (manual)
analytical methods use a  gas  chromatograph  with  flame ionization detection
(GC-FID) to quantify organic  species in the sample.  Because such systems
.ire not easily transported,  ambient  samples are  collected in inert bags or
in stainless steel canisters,  and  are  then  transported to the chromatograph.
The sophisticated nature  of  the  GC system permits  quantification of many
individual organic compounds  in  the  sample.  The concentrations of each
organic species can be summed  under  the categories of paraffins less
uethane, olefins, and  aromatics.   The  total of these three categories, plus
,iny unidentified organics, is  the  NMOC concentration.  Further guidance on
"sum of species" analysis  is  given in  reference  5.

     Continuous methods use  automated  analyzers  with FIDs to provide NMOC
concentrations that are generally  reported  as hourly averages.  Such
.inalyzers use either gas  chromatographic columns to separate methane from
i:he more complex organics, or  catalytic oxidation  to oxidize all organics
except methane.  The result  is that  methane and  total organic compounds
CTOC) are quantified by the  FID  separately.  Subtracting methane from TOC
(TOC - CH4) gives the NMOC concentration.

     Automated NMOC analyzers  suffer from a number of inherent technical
problems that limit the quality  of data they provide.  Several
evaluations (6,7,8) of NMOC  analyzers  present a  clearer  assessment of these
problems, and recommend potential  solutions.  Table 1 summarizes the
problems and Table 2 summarizes  the  recommendations.

     During the summer and fall  of 1979, an evaluation (9) of NMOC analyzers
<*as conducted to further  study the quantitative  effects  of the problems with
continuous NMOC monitoring.   This  evaluation pointed out that careful
attention to set-up, calibration,  operation, and maintenance enabled a group
of commercial NMOC analyzers  to  produce measurements above about 0.5 ppmC
which could be used in photochemical oxidation models.

L.3  PURPOSE

     The purpose of this  document  is to supplement the instruction manuals
of continuous monitoring,  ambient  level, NMOC analyzers. The document should
3e used in conjunction with  the manufacturer's instruction manual  so that
careful attention is directed  to critical procedures throughout set-up,
calibration, operation, and maintenance of  the NMOC analyzer.  Most  of the
procedures recommended here  were developed  and evaluated during the  1979
itfMOC analyzer evaluation  (9).  The document  initially addresses these
procedures in general  terms  for  "conventional" NMOC analyzers as a whole and
:hen specifically addresses  their  application to the following analyzers:
MSA 11-2, Bendix 8201 and 8202,  and  the Beckman  6800. Comments are  also
made on the Byron 233 and the  Meloy  500-2C.   Note  that the TAD is  limited to

-------
           TABLE 1.  SUMMARY OF PROBLEMS ASSOCIATED WITH  GATHERING
                        NMOC DATA WITH AUTOMATED  ANALYZERS
1.  Contaminants may be present in compressed gas cylinders containing calibration gases.

2.  Compressed gas cylinders of calibration gases sometimes contain the standard in a nitrogen or argon back-
    ground.  When no oxygen  is btended with these gases, FID sensitivity is altered.

3.  The assay of calibration gases contained in compressed gas cylinders (as received from the supplier) is
    sometimes incorrect.

4.  There are wide differences in the per-carbon response to different NMOC species.

5.  FID analyzers require hydrogen, which presents a potential operational hazard.

6.  The NMOC concentration  is obtained by subtraction of two relatively large and nearly equal numbers
    (TOC-CH4 = NMOC) and thus is subject to large, relative errors.

7.  NMOC analyzers may exhibit excessive zero and span drift during unattended operation.

8.  The complex design of some NMOC analyzers creates unique problems that are generally not experienced in
    other pollutant analyzers.  Meticulous setup, calibration, and operation procedures (which are analyzer-
    specific) are difficult to understand and follow.
      TABLE 2.  SUMMARY OF  RECOMMENDATIONS TO REDUCE THE EFFECT OF
                                PROBLEMS LISTED IN TABLE 1
1.   Calibration gases should be checked to determine the concentration of contaminants.

2.   Calibration concentrations should be obtained by dynamic dilution of a pollutant standard with zero grade
    air containing oxygen. The dilution ratio should be sufficiently high (~ 100:1) to ensure the calibration
    sample contains 20.9 ± 0.3 % oxygen.

3.   All calibration standards contained in compressed gas cylinders should be traceable to Standard Reference
    Materials from the National Bureau of Standards.

4.   The NMOC response should be calibrated to a propane standard.

5.   The operator should use documented procedures for hydrogen safety.

6.   Alt channels should be properly calibrated.

7.   The FIDs should be operated in accordance with instructions supplied by the manufacturer and this
    document.

8.   The training of qualified operators should be augmented with  a Technical Assistance Document, which
    details calibration and operation procedures for NMOC analyzers.

-------
those "conventional" analyzers which measure TOG and CH& separately and
derive the NMOC reading by subtraction.  No effort has been made to address
other types of NMOC anlyzers, such as direct-reading chromatographs.

-------

-------
                                  SECTION  2
                        GENERAL OPERATING  PROCEDURES
2.1  INTRODUCTION

     The complex operation of NMOC  analyzers  necessitates  the  need  for close
attention to maintenance  schedules  and  calibration  procedures  in  order to
obtain data suitable  for  photochemical  modeling.  Three  separate
studies (6,7,8) have  documented  these needs,  as well  as  the  need  for  less
complex procedures  to verify analyzer operations.   The  procedures discussed
in this document have been selected  to  address these  needs and to direct
attention to the importance of  rigid quality  control  procedures.

     This section  is  intentionally  general  so  procedures  pertinent  to NMOC
analyzers as a whole  can  be collectively  addressed.   More  specific
information is contained  within  the  individual analyzer  sections  located  in
the latter half of  this document.   DO NOT OVERLOOK  THIS  GENERAL PROCEDURES
SECTION.  Many suggestions detailed here  must  be  incorporated  into  the
monitoring program  to obtain suitable NMOC  data.  These  suggestions are not
detailed in the specific  analyzer sections.

2.1.1  Site Selection

     Proper selection of  the monitoring site  has  been carefully detailed  in
references 4, 10, and 11.

2.1.2  Analyzer Environment Control

     Control of the station environment surrounding the  NMOC analyzer is
recommended.  In a  recent evaluation of NMOC  analyzers  (9),  controllers
failed (for several days) to maintain the temperature in  the monitoring
station within the  specified range.  The  temperature  hovered at 38* C
(100* F) and resulted in  poor analyzer  stability, excessive  noise,  and
erratic responses  from many of  the  analyzers.  To reduce  these potential
problems, the temperature range  for operating  NMOC  analyzers should be
controlled between  20 and 30° C  (68 and 86° F), the temperature range in
which equivalency  testing of other  automated methods  is  currently conducted.
When ambient temperatures exceed about  29*  C  (85* F)  in  areas  of  high
humidity, the sample  manifold should be monitored for moisture buildup.
Analyzers without moisture drop-out traps may  be damaged by  moisture;
therefore, increase the station  temperature to within 5  to 7°  C (9  and 13*
F) of ambient temperature if excessive  moisture accumulates  in the  station
manifold.

-------
     Power  supplied  to  the  analyzer  should  be  between 105 and 125 volts at
60 Hz.  The normal voltage  range  supplied  to the  analyzer,  as well as  the
presence of power surges, can  be  determined by periodically monitoring the
voltage supplied to  the  analyzer  with  a  voltmeter.   Typically, voltage
monitoring  should be performed  when  the  site is  initially set up and at the
beginning of heating and cooling  seasons.   Elaborate systems to stabilize
the electrical  supply and continually  record its  voltage are expensive and
may be unnecessary if preliminary voltmeter results  indicate the voltage
range is within specifications  during  periods  of  sudden high power demand.

     Another factor  affecting  the analyzer's environment is dust.
Collection of dust on electrical  components can  cause shorts to occur  in
electrical circuits;  therefore, gently blow dust  off electrical components
every six months, or more frequently if  necessary.   Use a dry gas such as
nitrogen or compressed  air.

2.1.3  Ambient Air Sampling System

     A conventional,  ambient air  sampling  system  is  detailed in section
2.0.2 of EPA's QA Handbook  (10);  therefore, only  brief  comments are made
here.

     1.    Accumulation  of dirt  and moisture in the  sample manifold will lead
          to poor analyzer performance.  Regularly  clean the manifold  when
          particulate or high moisture buildup is detected, and position air
          conditioning  vents away from the  manifold.

     2.    To ensure  that the manifold  blower is operating,  strips of paper
          can be suspended in  front  of the  blower's  exhaust.  The motion of
          the paper  strips provides  a  visual check  that the blower is
          operational.  The blower should provide a  sample  flow of between
          85 and 140  L/min  (3  and 5  ft3/min) through the system.

     3.    Ensure that the sample  ports on the  manifold  are  pointed towards
          the ceiling to reduce the  potential  for moisture  accumulation in
          the analyzer's sample line.

     4.    Periodically  inspect  the manifold for cracks  or leaks and replace
          as necessary  to prevent  sample air contamination  by room air.

2.2  INSTALLATION OF  NMOC ANALYZERS
      NOTE:  THE ANALYZER'S INSTRUCTION MANUAL AND  THIS  TAD
      SHOULD BE READ PRIOR TO SUPPLYING POWER TO THE ANALYZER.
2.2.1  Unpacking

     Owners of new (and used) NMOC  analyzers  can  avoid  extended  downtime  by
applying a few common sense procedures during  unpacking.   Upon  receiving  an
analyzer, ensure that the instruction manual  and  Manufacturer's  Final  Data

-------
Sheet are included.   If  they  are  not  included,  immediately notify the
Manufacturer.  Then carefully  inspect  the  shipping  container and note the
presence and location of any  damage.   Remove  the  analyzer from the box (but
do not apply power to the  analyzer) and  determine  if  the  exterior of the
analyzer is damaged in the  same general  area  as  the shipping container.
Thoroughly inspect the exterior of  the analyzer  for scratched surfaces, bent
or broken knobs, cracked rotameters,  and cracked meter  covers.  Inspect the
Interior for electronic  cards  which may  have  come  unseated.   Continue
Inspection for broken cards,  loose  components,  loose  connections, or other
items which appear abnormal.

     If the analyzer  is  obviously damaged,  a  shipping claim  should immedi-
ately be prepared and directed to both the  shipper  and  the manufacturer.   If
analyzer damage  is not obvious but  the shipping  container shows damage,
schedule ample time during  start-up for  a  thorough  equipment  evaluation.
Leak checks and  other evaluations discussed later  should  be  performed on
analyzers suspected of having  been  damaged  during  shipping.   The
complexities of  the NMOC analyzer necessitate proper  performance of all
components to insure correct  operation.

2.2.2  Analyzer  Instruction Manual

     A complex analyzer  obviously will require  a thorough and sometimes
complex instruction manual  to  describe its  maintenance  and operational
procedures.  An  approach  to aid the operator  in  understanding the
instruction manual and recalling  analyzer  operations  follows.

     1.   Read the instruction manual  briefly and become  familiar with the
          analyzer and the  manual.  Highlight all  references  to maintenance
          procedures and  schedules, routine operational  recordings (such  as
          normal pressure  gauge readings or rotameter readings), safety
          notes, and actions  which  can invalidate  the analyzer warranty.

     2.   Several days later,  reread  the manual  and fill  in  highlighted data
          on appropriate  forms.   (Examples  of forms are  given in section
          2.2.3J

     3.   Summarize power-up  and  power-down operations  and calibration pro-
          cedures on a step-by-step basis.  A brief summary  which should  be
          prominently displayed in  the analyzer  area  can  be   obtained from
          the manual and/or this  TAD.  Some instruction manuals contain
          condensed start-up  sheets that will serve this  purpose if a clear
          understanding  of  the analyzer  operation has first  been obtained.

2.2.3  Recordkeeping

     To aid in the operation  of an  NMOC  analyzer,  records and documents
should be organized and  kept  close  at  hand.  Maintaining  awareness of

-------

operating parameters and maintenance  schedules  is  important  to  the  contin-
uing, accurate operation of an NMOC analyzer.   Detailed descriptions  of
recordkeeping for air monitoring  programs  are contained in section  2.0.3  of
the EPA QA Handbook (10).

     An analyzer logbook is a typical  recordkeeping  system.   The  logbook
contains important records which  enable the operator  to recognize  patterns
in the analyzer's behavior and anticipate  potential  difficulties.  Such a
logbook can be contained in a ring binder  and should  be located at  the
monitoring site near the analyzer.  Typical sections  which should  be
contained in the logbook follow.

     1.   Instruction manual.  Punch  the manual (or  a copy of it)  to  fit  the
          rings and secure it into the binder.

     2.   Analyzer information.   Identify  the manufacturer,  the analyzer's
          name and model number,  pollutant monitored,  serial  number,  date
          purchased and cost, salesperson's name and  phone number,  service-
          person's name and phone number,  and recorder make,  model  and
          serial number (Table 3).

     3.   Routine check sheet.  During routine  checks, record operational
          information (Table 4).

     4.   Calibration data - Identify  the  analyzer make, model  and  serial
          number; name of operator and date of  calibration;  pollutant
          concentration, analyzer responses,  and the  resulting  calibration
          curve (Table 5).

     5.   Zero and span data.  Identify the analyzer  make, model  and  serial
          number, the date, zero  response, span response, span  concentra-
          tion, slope and intercept (Table 6).

     6.   Maintenance.  Identify  the  analyzer make, model and serial
          number, the maintenance required, when it  is required,  and  when it
          is performed (Table 7).

     7.   Comments.  Record all comments relating  to  the analyzer,  such as
          manufacturer communications, analyzer malfunctions,  etc.  Record
          the analyzer make, model and serial number  at the  top of  eacj^ page
          and date each comment.  The  first page should be the  Manufac-
          turer's Final Data Sheet.

-------
     TABLE 3.  FORM FOR RECORDING ANALYZER INFORMATION
Manufacturer




Model




Pollutant




Serial Number




Date Purchased




Cost                 $.




Salesperson's Name




Saleperson's Address




Salesperson's Phone No.   .




Serviceperson's Name




Serviceperson's Address   .




Serviceperson's Phone No..




Recorder Make




Recorder Model




Recorder S/N




Instrument Output
                                 10

-------

TABLE 4. TYPICAL FORM FOR RECORDING ROUTINE OPERATIONAL DATA
FROM AN NMOC ANALYZER
1'
*
>
i
1
!
1
l
x"
4
1
I
1
1
i
*
S
1
|
i
g
I
5
8
|
S*
I
n
f
g
s
5*
I
\J
li\
l
!
I
1
1
1
|
5"
8
S
5
§
i
5'
8
II
g
g
1


























































































































































































































































































































































































































































































































































































































































































































































































































































































11

-------
             TABLE 5. FORM FOR RECORDING CALIBRATION DATA
ANALYZER MAKE AND MODEL




CODE NUMBER 	




PC LLUTANT	
DATE OF CALIBRATION



DERIVED CURVES	
SERIAL NUMBER




RANGE, ppmC	



OPERATOR	
          Concentration, ppmC



Unadjusted Data
         Analyzer Response—DAS Voltage
     NMOC
CHy
TOC
Adjusted Data
                                   12

-------
          TABLE 6. FORM FOR RECORDING ZERO AND SPAN DATA




ANALYZER MAKE AND MODEL 	

POLLUT
Data


















ANT RANGE, ppmft

Zaro
Response


















Span
Rasoonsa


















. Span
Concentration


















Slopa


















Intareapt


















                                13

-------
             TABLE 7.  FORM FOR RECORDING A ROUTINE
                     MAINTENANCE SCHEDULE
ANALYZER MAKE AND MODEL:

SERIAL NUMBER: _	
       Function
Frequency
                                                  Date Performed
                                14

-------
2.2.4  Setup

     After generally understanding how  the  analyzer works,  the  operator  is
prepared to follow  the manufacturer's instructions  for  setting  up  the  unit.
During the setup of any analyzer which  uses hydrogen gas, carry out  a  leak
check of the hydrogen flow system.   (See  section 2.4.2  for  details  in
leak-checking flow  systemsJ

     The hydrogen (H2) gas for  the FID  requires extreme caution during
handling due to its highly flammable nature.   Colorless and odorless H2
gas leaking into an unventilated area is very  dangerous and is  potentially
explosive.  The following precautions summarize safety  considerations  when
using high pressure H2 cylinders.  Additional  information on H2 cylinder
safety, as well as safety when  using electrolytic hydrogen  generators,  is
contained in reference 13.

     1.   Store H2 cylinders  in a secured,  well ventilated, non-flammable
          area, preferably outdoors  and out of direct sunlight.  Secure  the
          cylinder to a wall  prior to installing the pressure  regulator. Do
          not store H2 cylinders beside cylinders containing corrosive or
          highly oxidizing materials.

     2.   The H*2 regulator should be dedicated to H2 gas only.   The
          output stage should contain a pressure relief valve.   A  flow
          restrictor must be  connected  to the  exit of the regulator  to limit
          gas flow  in the event of a break  in  the H2 tubing.  If the
          output pressure of  a  new cylinder is less than the typical charge
          pressure  (~ 2200 psig) , leak-check the cylinder valve while  at
          full clockwise and  full counter-clockwise position.

     3.   Use only clean or new 3 mm (1/8 in)  diameter  stainless steel
          tubing to supply H2 to the analyzer.  Thoroughly  leak-check  the
          hydrogen line by pressurizing the line up to  the  analyzer.   Turn
          off the H2 cylinder valve  and observe the outlet  pressure.
          There should be no  decay in pressure after 30 minutes.   If there
          is a leak, pressurize the  line  and search for the leak using a
          liquid leak detector.

     4.   When ordering hydrogen cylinders, specify TOC concentration  to be
          less than 0.1 ppmC.

     5.   A particulate filter  (in a metal  holder) can  be located  on the
          outlet of the regulator to ensure that no particulates enter  the
          analyzer, and subsequently, the FID.  Carefully leak-check each
          filter holder after the filter  is changed.
                                      15

-------
After Che hydrogen  system  is  determined  to be safe, recorder and gas
connect ions can be made  to  the  analyzer.   A terminal strip for connecting
recorder leads is usually  located  on  the  back of the analyzer.  Follow
instruction manual  directions  for  identifying each channel and the voltage
outputs  from each.  Care should be exercised by using signal cable that
contains no splices and  is  as  short as  possible.  Follow the recorder
manufacturer's suggested procedures in  installing the recorder.  Signal
cable should be connected  to  recorders  and analyzers with approved
connectors that are either  soldered or  crimped to the signal cable.  Do not
wap bare cable around a connector.
     Gas connections  can now  be made  to  the inlets located on the rear of
Che analyzer.  All  sample  gas  tubing  should be composed  of Teflon® and be
connected to the analyzer  and  sample  manifold with either stainless steel,
Teflon®, or all glass connectors.   Ensure that the tubing has not been
previously used.  Specific  size tubing  is addressed in the analyzer sections
of this  document.   An all-Teflon®  particulate filter, with a pore size of 5
microns  or less, should  be  installed  in  the sample line.  The sample line
should be as short  as possible,  not exceeding 3m (10 ft.).

2.3  OPERATION

2.3.1  Start-up
      REMINDER:  SECTION  2  IS  INTENTIONALLY GENERAL.   MORE
      SPECIFIC  INFORMATION  IS  CONTAINED IN THE  ANALYZER
      SECTIONS  TO AID  THE OPERATOR DURING START-UP.
     By this time, the operator  should  have  a  basic  understanding of the
•analyzer's operation, should  be  assured that the hydrogen flow system does
not contain leaks, and should have  a  set  of  summary  sheets  which will aid in
operating the  analyzer.   It  is  important  to  achieve  these goals prior to
starting up the analyzer  because  the  complexity  of  the  system may disguise
malfunctions and  lead the operator  toward improper  repairs  or unsafe
conditions.

     After initial settings  and  adjustments  are  complete, supply power to
the analyzer,  allow the support  gases to  flow  into  the  unit,  and attempt to
ignite the burners.  Do not  be  concerned  with  flameouts during the first 15
minutes of operation.  If after  this  period  flameouts continue,  allow the
flame to extinguish.  Allow  the  analyzer  to  warm up  for twenty-four hours to
ansure that heated elements  are  up  to operating  conditions  and attempt
ignition if the burner(s)  is  not  already  operating.   If ignition
difficulties persist, consult troubleshooting.   While the analyzer is
stabilizing, sec  up the recorder(s) as  per the manufacturer's suggestions.
Perform maintenance on the recorder to  verify  operation and ensure that  the
recorders will not be responsible for  faulty data.
                                     16

-------
      If  the  analyzer  is  new,  the  flow  rates  and  pressure  gauges  should have
already  been correctly adjusted by  the manufacturer.   Do  not  adjust  pressure
gauges or  flow  controls  if  minor  discrepancies exist  between  your  readings
and the  Manufacturer's Final  Data Sheet.   Major  discrepancies,  such  as no
gas flow into the FID, should obviously be investigated.   If  the analyzer  is
not new, adjustments  should be made  to bring  the flow rates and  gauge
settings into the ranges  recommended in each  analyzer section of this
document or the  instruction manual.

2.3.2  Periodic  Calibrations  And  Zero/Span Checks

     An  initial  multipoint  calibration is  recommended to  establish the cali-
bration  curve and determine response linearity.  Periodic zero and span
checks are thereafter necessary to  update  the curve,  with less  frequent
multipoint calibrations  to  verify response linearity.  The frequency of zero
and span checks  and multipoint calibrations  should  be based on the
analyzer's performance.   Analyzer performance should  be monitored  using a
control chart which documents  analyzer response  to  standards, thus allowing
calculation of  response  variation to standards.  Section  2.0.9 of  reference
10 contains additional information  on calibration of  automated  analyzers.
Appendix H of reference  14 discusses application of the control  chart.
Although these discussions  are directed to analyzers  that monitor  ambient
air for criteria pollutants,  application of  this strategy to  NMOC  anlayzers
is recommended.  Further  details  on calibrations are  contained  in  section 3
and within each  specific  analyzer section.

2.3.3  Operational Checks

     After the NMOC analyzer  is calibrated and on-line, routine  checks of
operational parameters are  recommended.  These checks are part  of  the
quality control  program  and include recording pressure gauge, rotameter,  and
pyrometer readings.  The  gauge and  meter readings should  be recorded
immediately following a multipoint  calibration and  should be  designated as
the reference readings.   Readings taken during subsequent checks should be
compared to the  reference readings  to determine  if  operational  parameters
have shifted.   (These data  can be recorded on routine check sheets similar
to Table 4.  Specific forms for other analyzers  are included  in  each
analyzer section of this  document.)  The sensitive  nature of  NMOC  analyzers
requires investigations of  any substantial drift in operational  parameters.
Assistance in such investigations is given in the troubleshooting  sections
of the instruction manual and  this  document.

2.3.4  Preventive Maintenance

     The complex pneumatic  systems  and controls  contained within NMOC  analy-
zers necessitate thorough maintenance procedures and  schedules.  Adherence
to the manufacturer's maintenance schedules  is required to obtain  reliable
monitoring data.  The importance  of maintenance  performed by  experienced
                                      17

-------
     personnel cannot be overemphasized.  The manufacturer's  instruction
     manual. details procedures and schedules.  Comments on  areas  of  primary
     importance follow,

     1.    It is strongly recommended that maintenance of NMOC  analyzers  be
          performed by a competent, qualified person.  This person should
          "specialize" in NMOC analyzer operation, maintenance, and  repair
          and should be prepared to devote a substantial portion  of  time to
          such matters.

     2.    The hydrogen flow systems should be leak-checked  during start-up
          and at least once every 180 days of operation.  (A leak-check
          procedure is discussed in section 2.4.2.)

     3.    The analyzer's sample pump should be evaluated every 180 days  of
          operation to ensure that it has the capacity to maintain necessary
          flows and pressures.  (Sample pump evaluation is  discussed  in
          section 2.4.3.)

     4.    Check for loose or dusty electronic cards every 30 days of
          operation.

     5.    Check the particulate filter in the sample line at least every
          month and replace when it becomes visibly soiled.

     6.    Evaluate chromatograms weekly.  Analyzers that contain  chromato-
          graphic columns are usually designed to output chromatograms.  A
          chromatogram is a valuable troubleshooting tool because it  is  a
          "fingerprint" of the analyzer's operation.  A chroraatogram  should
          be obtained while the analyzer is sampling the methane  and  propane
          standards, so gate timing and sensitivity can be  evaluated  and
          compared to previous chromatograms.  These evaluations  can
          indicate a forthcoming malfunction.  (The specific analyzer
          sections contain additional information on chromatograms.)

2.4  TROUBLESHOOTING

2.4.1  Flow Measurements

     The sensitivity of NMOC analyzers requires proper flow rates throughout
the analyzer.  If pressure gauge readings of flow systems begin to decay,
calibration curves begin to drift in one direction, or balance procedures
cannot be performed, flow systems may be developing leaks or obstructions.
When any of these potential problems cannot be corrected by adjusting
pressure gauges to reference readings obtained during the last multipoint
calibration, flow rates throughout the analyzer should be measured and
compared to flow rates obtained during the last multipoint  calibration.  To
                                      18

-------
enable comparison, all  flow  rates must  be  corrected  to  standard temperature
and pressure (25* and 760 mm Hg).   Details  on  flow rate corrections  are
contained in section 2.1.2 of  reference 10.

2.4.2  Leak-checks

     If measurements indicate  reduced  flow rates,  a  leak or obstruction may
be present in the flow  system.  This section discusses  a leak-check
procedure and identifies apparatus  that  can aid  in verifying or locating a
leak.  The test apparatus addressed can be  fabricated  from materials readily
available to most monitoring programs;  however,  it may  not be as sensitive
as a leak-check device  produced by  the  analyzer  manufacturer.
       CAUTION:  INDISCRIMINATE APPLICATION  OF  LEAK  DETECTOR
       SOLUTIONS CAN BE HARMFUL TO ELECTRONIC COMPONENTS.
       SINCE LOW FLOW RATES IN NMOC ANALYZERS REDUCE THE  EFFEC-
       TIVENESS OF SUCH SOLUTIONS, THE USE OF LEAK DETECTOR
       SOLUTIONS IS RECOMMENDED ONLY UNDER SPECIAL CIRCUMSTANCES,
       AS DETAILED LATER IN THIS SECTION.
     1.    Locate all gauges and flow systems on  the  instruction  manual's
          flow chart and highlight each  system with  different  colors.   (Red
          for H2» green for carrier, etc.)

     2.    Construct a leak-check apparatus  as shown  in  Figure  1  using  a
          pressure source (compressed air or nitrogen containing < 0.1 pptnC
          TOC),  pressure regulator, toggle  valve,  a  340 kPa  (50  psig)
          pressure gauge graduated in 7 kPa (1 psig) increments, flexible
          tubing, and an exit cap.

     3.    Leak-check the apparatus by applying pressure,  closing the  toggle
          valve, disconnecting the pressure source,  and  observing the
          pressure gauge reading for 30 minutes.   If there  is  no decay in
          the pressure reading, the apparatus contains  no  leaks.  If  a leak
          develops, locate it with a liquid leak detector  and  repair  it.
          Recheck for leaks.

     4.    Connect the leak-check apparatus  to the  inlet  of  the analyzer's
          flow system to be leak-checked (either a span gas  inlet,  carrier
          gas inlet, support air inlet, hydrogen gas inlet,  or sample
          inlet).  Ensure that the sample pump is  not in  the system to be
          checked.  If it is, bypass it with tubing.

     5.    Cap the exit(s) of the system to  be leak-checked.  Most  systems
          exit into the FID; therefore, locate the flow system's last
          fitting from the FID and connect  the exit  cap.  AVOID  TOUCHING THE
          THREADS OF THE FID AND FITTING.
                                      19

-------
  Pressure Regulator
                                           Pressure Gauge
                                             Flexible
                                             Tubing
Exit Cap
                    Toggle Valve
Pressure
Source
                            Figure 1.  Leak-check apparatus.
           Apply a pressure level  to  the  flow system which will not  exceed
           maximum allowable pressures  in the analyzer and close the  toggle
           valve on the apparatus.  The  instruction manual or pressure  gauge
           in-line will indicate maximum  allowable pressures.

           Monitor the pressure gauge  located on the apparatus.  If  after 30
           minutes there is no decrease  in the pressure reading, the  flow
           system contains no leaks.   If  pressure decrease occurs, relocate
           the apparatus and pressurize one-half of the analyzer's flow
           system.  Continue moving the apparatus until a general area  of the
           flow system containing  the  leak is identified.  Carefully  apply
           the leak detector solution  to  the  fittings and tubing if
           necessary. Foaming or bubbles  indicate the location of the  leak.
           Inspect the fitting or  valve and  replace or repair as necessary.

           After repairing the leak,  recheck the entire system to verify
           that all fittings or valves are functioning properly with  no
           leaks.
                                       20

-------
2.4.3  Sample Pump Checks

     Once every 180 days of operation, determine  if the sample  pump  can
supply the necessary head pressure recommended in the manufacturer's
instruction manual.  Perform the check by connecting a high pressure  gauge
to the outlet of the pump and capping the exit from the gauge.   Start  the
pump and record maximum pressure output.  A noticeable decrease  in head
pressure from the last sample pump check may indicate that the  pump  needs
service.
                                      21

-------
                                  SECTION 3

              CALIBRATION  STANDARDS,  EQUIPMENT,  AND  PROCEDURES
3.1  INTRODUCTION

     As noted  in other  sections  of  this  TAD,  the response of an FID analyzer
may vary considerably from  one organic compound  to  another.   Thus,  the
choice of the  calibration species has  a  direct bearing on resulting
analyzer calibration.   Historically, continuous  ambient  NMOC analyzers have
been calibrated largely with methane-in-air  standards.  Methane was used
primarily because  it is chemically  simple  (C=l,  and  ppmC = ppmV),  it is
readily available,  .and  it can be used  to calibrate  both  CH^  and TOC
measurement channels (from  which NMOC measurements  are derived by
subtraction) .-

     But propane-in-air standards have also  been used  to calibrate  automated
NMOC analyzers because  (the contention is) propane  allows a  more accurate
measurement of the  NMOC concentration  in ambient air.   This  contention is
supported by the following  facts and reasons:

     1.   Propane  is more representative (than methane)  of the hydrocarbon
          mix used  to calibrate  hydrocarbon  analyzers  that  collected the
          data base from which the  EKMA  model was developed.

     2.   In a recent study (9), automated NMOC  analyzers measured
          individual NMOC compounds commonly  found  in  ambient  air.   Response
          to propane was found to be approximately  equal to  the average
          response  to all compounds measured.  Response  to methane  was found
          to be 20  Co 30% greater than the average  response  to all  compounds
          measured.

     3.   In a recent study (9), analysis of  ambient  air by  automated NMOC
          analyzers calibrated to propane produced  responses that  agreed
          quite well with sum-of-species GC-FID  results.  Agreement between
          these methods was also quite good when sampling synthetic organic
          mixtures.

     4.   Direct reading, automated NMOC analyzers  (that do  not produce
          CH^ or TOC output) are generally calibrated  to propane
          standards (they obviously would not respond  to methane standards).
                                       22

-------
     For these  reasons,  EPA  recommends  chat  continuous  NMOC data, obtained
for use in dispersion models  for  predicting  ozone  concentrations, should be
referenced to propane standards.  CE^ and TOO  responses should be balanced
and referenced  to methane  standards.  Discussions  on  calibrations and
standards are contained  in this section.

3.2  ZERO AIR STANDARDS

     The first  step  in calibrating  an NMOC  analyzer  is  to obtain the
response to a hydrocarbon-free matrix.   Ideally,  this matrix should
approximate "clean"  ambient  air;  that is, air  containing only nitrogen,
oxygen, carbon  dioxide,  and  water vapor.  Since  the  analyzer's response to
this matrix is  its baseline  or "zero" response,  the matrix is frequently
referred to as  "zero air".

     Zero air is actually  not totally free  of  all  organic (or other)
compounds, but  is usually  considered adequate  if  the  TOG concentration is
less than 0.1 ppmC.  It  is important that zero air contain ambient
quantities of oxygen, but  carbon  dioxide and water vapor may also be
present.  The effects of oxygen content  variations have been studied (9) and
determined to have a significant  effect  on  FID response.  Moisture and
carbon dioxide  effects on  FID response  are  not as  well  understood; therefore,
the recommended zero air standard should contain  20.9 _+ 0.3% oxygen in a
nitrogen balance with TOG  concentration  <0.1 ppmC.

     The recommended method  for preparing zero air is by scrubbing and
oxidizing unwanted compounds  from ambient air.  Such  cleanup systems are
usually large and contained  in the  laboratory  but  may also be small and
contained in the analyzer  itself.  Zero  air  is also  available from specialty
gas manufacturers who compress the  air  and  store  it  in  cylinders; however,
this source of  zero  air  is not recommended  due to  possible variable oxygen
concentrations, economic considerations, and frequent bottle changes.

     A typical  cleanup system recommended  for  producing zero air for NMOC
analyzers is shown in Figure  2.   Ambient air should be  drawn from the
outside of the  building  and  filtered before  being  compressed.  Use of
outside air is  more  likely to provide true  ambient concentrations of oxygen
and nitrogen in the  zero air matrix.  The dryer  removes moisture before the
air enters the  scrubber  unit.  As the air passes  through the scrubber unit,
compounds such  as nitrogen dioxide  (N02), sulfur  dioxide (802), ozone
(03), etc., are removed.   The air then  enters  the  heated catalytic
oxidizer where  organics  and  carbon  monoxide  (CO)  are  converted into carbon
dioxide (۩2)   and water (1^0).   Since many laboratories prepare zero
air, different  detailed  designs are acceptable if  the resulting product is
determined to be within  the  specifications  previously mentioned.  (Note:
All pressure regulators  should be LOX cleaned  and  contain non-permeable,
metal diaphragms.  Do not  use a heat less air dryer in the zero air system
because it will alter oxygen  levels in  the  sample.)
                                      23

-------
Ambient
Air In
, n
U
i
Panic
Filter

ulate
Oil- Lass
Compressor



Storage/Ballast
Tank
P
P
-C
ressure
egutator
                                                             Scrubber
                                                              Unit
              To Dilution
           "*" System
                       Figure 2.  Diagram of a zero air system.
     Cleanup  systems  contained within NMOC analyzers may use  only  a  heated
catalyst to oxidize organics.  Care should be taken to ensure that the
heater and catalyst are  properly maintained.  This internal system should  be
used only for  level 2 zero checks (section 3.9) unless it has been verified
against an external cleanup system.  Many monitoring sites normally  contain
an external zero  air  system so a common zero matrix can be delivered to  all
analyzers at  once.

     As mentioned  before,  zero air compressed in cylinders can also  be
utilized to zero  the  analyzer, although this is not recommended.   If
compressed gas  is  used,  ensure that it has the correct oxygen content  and
that it is delivered  to  the analyzer's inlet at atmospheric pressure.
Failure to supply  a vent  between the cylinder and the analyzer will
pressurize the  sample system and could result in improper flows or damage  to
i.nternal flow controls.  (See the discussion on calibration gas dilution
systems in section 3.5.)

3.3 SPAN GAS  STANDARDS

     As noted  earlier, both methane and propane standards are required  to
calibrate an  automated analyzer that produces CH^, TOC and NMOC outputs.
These standards  are available from several commercial manufacturers  in

                                                   U.S. EPA Headquarters Library
                                                         Mai! code 3201
                                                   1200 Pennsylvania Avenue NW
                                        24             Washington DC 20460

-------
 compressed  gas  cylinders  at  a  variety of concentrations.  Standards must be
 traceable to  a  Standard Reference Material  (SRM) from the National Bureau of
 Standards (NBS) or  to  a commercially certified reference material (CRH).  The
 procedure to  certify a cylinder of gas as traceable to an NBS/SRM or CRM can
 be  performed  by the gas manufacturer or the user.  (Traceablity protocol is
 addressed in  reference 15 or in section 2.0.7 of reference 10.)  Table 8
 lists several concentrations of methane and propane SRM's that are available
 from NBS; however, only one SRM is required for traceability purposes.
 Reference 16  addresses NBS/SRM's; consult gas manufacturers about the
 availability  of CRM's.  Methane and propane standards can be combined in a
 single cylinder with an ambient air balance.  Such a gas cylinder containing
 both standards may be used for spot checks between calibrations, but will not
 permit the CH4 and TOC channels to be balanced.

      Inconsistent results from NMOC analyzers have been partly attributed to
 contaminated  calibration standards.  For this reason, the methane and propane
 standards should not only be referenced to an NBS/SRM or CRM every three to
 six months, but they should also be analyzed yearly to ensure that the
 concentrations of organic contaminants  remain less  than 0.1 ppmC.   For
 example, the methane standard should contain less than 0.1 ppmC NMOC, and the
 propane standard should contain less than 0.1 ppmC organics other  than
 propane.  Both standards must contain ambient levels of oxygen in a nitrogen
background if no ambient level diluent  air is to  be blended with the
 standards.
        TABLE 8.  LISTING OF METHANE AND PROPANE NBS/SRMs AVAILABLE
                         IN COMPRESSED GAS CYLINDERS
SRM
1658A
1659A
1660A
GAS
CH4/Air
CH4/Air
CH4/C3H8/Air
ppmV
0.951
9.43
4.10 (CH4),
0.976 (C3Hg)
ppmC
0.951
9.43
4.10 (CH4),
2.928 (C3Hg)
        1665

        1666

        1667

        1668

        1669
C3Hg/Air

C3H8/Air

C3H8/Air

C3H8/Air

C3H8/Air
  3

 10

 50

100

500
   9

  30

 150

 300

1500
                                       25

-------
3.4  EFFECTS OF MOISTURE

     The effects of moisture  on  automated NMOC analyzers is not well known.
Elecommendations are therefore  not  available.   Preliminary investigations
indicate that  analyzers  addressed  in this document will not be affected by
moisture if the precautions  in each  analyzer  section are followed.

3.5  CALIBRATION GAS  DILUTION  SYSTEM

      There are basically  two  methods to  deliver calibration standards to
automated analyzers in the  field:  by dynamic  dilution of a single high
concentration  standard with  zero  air, and by  using compressed gas cylinders
containing calibration concentrations. The dynamic dilution approach is
recommended for calibrating  automated NMOC analyzers because this approach
offers greater stability of  the high concentration standard in a cylinder,
greater control of oxygen  content  in the  sample, greater ease in delivering a
variety of accurate calibration standards,  and a lower cost to produce
different concentration  levels.   However, accurate flow measurements are
required.

     A dynamic dilution  system, as shown  in Figure 3,  can be used to dilute a
high concentration standard  with  zero air.  This system requires flow
controllers or needle valves  to control diluent  and pollutant flows; a flow
meter to measure the  flow  rates;  stainless  steel,  Teflon*,  or glass tubing to
direct the flows; a glass  mixing  flask to blend  the two flows; and a glass
nanifold or tee for sampling  the  standards.   The manifold must be vented to
t:he atmosphere to prevent  pressure buildup in  the sample line connected to
t:he analyzer's SAMPLE port  (not CALIBRATE port)!  To prevent contamination of
line calibration standard by  the room atmosphere, all unused sample ports on
i:he manifold must be  sealed,  and  the flow to  the manifold must be 20 to 50%
greater than the analyzer's  sample demand.    To  maintain oxygen at 20.9 +
0.3%, the diluent to  pollutant flow  rate  ratio should  be greater than 100:1.

     Care must be taken  to ensure  that the  system contains  no leaks that can
introduce error into  flow measurements or allow  contaminants to enter the
system.  Leaks may be detected by  applying  a  low level of pressure to the
system and submerging the glassware  and tubing in  a bath, or by using a
liquid leak detector  on  the  fittings.  Care must also  be taken to ensure that
i:he system is  clean and maintained to prevent  contamination.  Use only new
Cubing and glassware  to  construct  a  dilution  system.

3.6  FLOW MEASUREMENTS

     Preparing a known pollutant  concentration by dynamic dilution requires
accurate measurement of both pollutant and diluent  flow rates.  Flow rates
must be corrected to  standard  temperature and  pressure (25*C and 760 mm Hg)
unless two measurements  taken  with the same flowmeter  are to be ratioed.  For
                                       26

-------
                       Figure 3. Diagram of a dilution system.
corrections, atmospheric  pressure  and  temperature in the flow measurement
area must also be measured.   The  flow  measurement device must be traceable to
an NBS standard.  Two  devices  commonly used  to  measure flows in the field are
soap bubble flowaeters  (SBFM)  and  wet  test meters (WTM).  Details on
calibrating flow measurement  devices are  given  in section 2.1.2 of reference
10.  Comments concerning  areas  of  specific concern follow.

     1.   When using an SBFM,  select a bubble tower that will permit a flow
          measurement  for at  least 15  seconds.   That is, use a 10-cm-*
          tower for flow  rates  less  than  40  cm^/min, use a lOO-cnP tower
          for flow rates  between 40  and 400  cm^/min and use a 1000-cnH
          tower for flow  rates  between 400 and  4000 cm^/min.

     2.   During SBFM measurements,  allow only  one bubble to traverse the
          column while  measuring  the time interval.  The entire calibrated
          portion of the  column should be used  during the measurement.  If
          the bubble bursts before it  traverses the column, check to see if
          the column is clean (clean regularly  with hot, soapy water and
          rinse well)  and is  not  in  hot or cold drafts.  A standard soap
          solution designed^ for flow measurements can reduce the difficulties
          with bubble  flows.
                                        27

-------

     3.   Wet test meters should generally  be  used  within a flow range of
          1/3 to 3 neter revolutions  per minute.  A one  liter  per revolution
          WIM, for example, should be  used  within the  range of 1/3 to 3
          liters per minute.  WTM's are available in many different
          capacities, covering a wide  range of applicable flows.

     4.   WTM calibration factors vary with the  flow rate through the meter.
          When WTM's are used to measure dynamic  flows  (rather than
          incremental volumes), the ideal calibration  (capable of 1%
          accuracy) should relate measured  flow rate to  actual flow rate
          (both adjusted to some common set  of reference  conditions).
          However, if meters using factors  are operated  within the range of
          1/3 to 3 revolutions per minute,  an  accuracy of 2 to 3% can be ex-
          pected.

     5.   Flows through WTM's should  be measured by complete meter
          revolutions only.  Additionally,  the number of  timed,  complete
          revolutions should be chosen so the  timed interval is  not less
          than 20 seconds.

3.7  GENERAL MULTIPOINT CALIBRATION PROCEDURE

     The actual procedure by which an  NMOC  analyzer calibration  is
referenced to propane varies, depending on  the design of  the analyzer,  and
specifically, the number and type of  output channels,  the number of zero and
:jpan controls available, and whether  the TOC-CH^ subtraction is
implemented on signals before or after they are  processed by the span
control circuits.

     The following generalized procedure illustrates the  normal  sequence of
steps that are recommended for calibrating  an  NMOC  analyzer.  This general
procedure must be individually adapted to the  particular  design  of each NMOC
analyzer.  For example, in its normal  configuration, the  MSA Model 11-2
analyzer has no TOC output channel.   Therefore the  internal TOC  reading
necessary to obtain the NMOC reading  must be balanced  to  the CH^ channel
to provide a zero NMOC reading when sampling methane.  More explicit
information on the exact calibration  procedure for  each  analyzer is con-
tained in the analyzer-specific sections of the  TAD and  in the operation
aanual for each analyzer.

     1.   Sample zero air and adjust  all channels to read 0.0  ppmC.

     2.   Sample a concentration of the methane  standard  equivalent to  about
          85% of full scale.  Adjust  the span  controls of both the CH4  and
          the TOC channels' outputs.   Make  sure  the TOC  response equals the
          CH4 response so that subtraction  of  the CH4  response from Che
          TOC response will yield an  NMOC response  of  0.0 ppmC.   (This
          balance may be accomplished  in certain models  by adjusting the
                                      28

-------
          hydrogen  or  sample flow rate to the FID.  Analyzers  without  a TOC
          output must  be  balanced to ensure that the NMOC  reads  0.0  ppmC.)
          If  substantial  adjustments are necessary, resample  zero  air  and
          adjust the zero responses as necessary, then repeat  the  balance.

          Sample at  least two additional concentrations of the methane
          standard  approximately equally spaced over the scale range.
          Record the Clfy  and TOC responses.  Least squares regression  of
          the  analyzer responses (dependent variable, Y) against the cali-
          bration concentration (independent variable, X)  should result in  a
          correlation  coefficient of 0.9996 or better if the  channel is
          linear.

          Sample a  concentration of the propane standard equivalent  to about
          85%  of full  scale.  The CH^ channel's response should  equal
          zero. Analyzers with an independent NMOC channel span  pot  may be
          adjusted  as  necessary to make the NMOC channel's response  equal to
          the  propane  concentration.  Do not adjust the TOC channel's
          response.  If substantial adjustments are necessary, resample the
          zero  standard and  the methane standard as before.
     5.  Sample  at  least  two additional concentrations  of  the  propane
         standard approximately equally spaced over the  scale  range,  and
         record  the  NMOC  response.  Least squares regression,  as  before,
         should  result  in a correlation coefficient of  0.9996  or  better if
         each  channel  is  linear.

NOTE: WHEN THE NMOC  CHANNEL IS CALIBRATED WITH PROPANE,  THE  SUM OF THE CH4
AND NMOC CONCENTRATIONS MEASURED WILL NOT EQUAL THE MEASURED TOC  CONCEN-
TRATION BECAUSE  OF /THE  DIFFERENCE IN FID RESPONSE BETWEEN  METHANE AND
PROPANE.

     After linearity is established, generate the regression equations that
define the curves for  each channel and post the calibration  curve, date, and
channel identification  on the appropriate recorder.  All calibration  data and
curves can be  entered  on  a form similar to Table 5 and  filed in the cali-
bration section  of  the  analyzer's log book.  See section 2.0.9.1.2 of
reference 10 for more  detail on calibration theory.

    For analyzers which have no direct NMOC output, calculate  NMOC response
using the following  equation:
  NMOC
  response
f TOC response
  to sample
                       TOC response
                         to zero
               TOC slope to methane
CH4 response	CH4 response
 to sample     to zero    )   TOC slope to methane
                          CH4 slope to methane
                       TOC slope to propane
                                       29

-------
An example calculation using  actual  calibration curves and ambient data
responses from an NMOC analyzer  follows.

     Methane calibration curves
          CH4 response « 9.658X  +  0.2256
          TOC response * 9.731X  -  0.0359

     Propane calibration curves
          TOC response - 6.536X  •*•  0.111

     Analyzer responses to  ambient  air  following calibration
          CH4 channel * 22.1% Chart
          TOC channel « 41.92 Chart

Applying the data to the NMOC equation:
  NMOC
response
41.9 + .0359   22.1 -  .2256
   9.731
9.658
9.731
6.536
                                          3.04  ppmC
                                        (2)
3.8  LEVEL 1 ZERO AND SPAN  CHECK

     The purpose of the  level  1 zero  and  span  check (see  section 2.0.9 of
reference 10) is to periodically correct  the most  recent  calibration curve
for zero and span drift.  Once  linearity  is established  in a multipoint
calibration, the calibration needs  to be  checked with  only two  points;
usually at zero and 85%  full scale.   The  standards and the procedure used
iuring the multipoint calibration should  be used,  except  that during the
level 1 check, only one  upscale concentration  is necessary.

    All NMOC measurements obtained  from the analyzer should  be  converted to
ppmC via the most recent calibration  curve or  zero and span  check.
Accordingly, adjustments to the zero  and  span  pots need be made only if
considerable drift has occurred.  (See section 2.0.9.1.3  of  reference 10 for
details on drift.)  For  example, Phase II of the EPA's Equivalency
Designation Program specifies  that  zero adjustments  of analyzers are needed
if the zero response drifts more than _+ 3% of  full scale  from the  intercept
of the most recent multipoint  calibration curve.   Span adjustments  are to be
aade only if the slope of the  new calibration  curve  drifts approximately
nore than +7% from the slope of the most  recent multipoint curve.
     To calculate zero drift on an  NMOC analyzer operating in the 0 to 10
ppmC range, subtract the level 1 zero response from  the intercept of the
nost recent multipoint curve.  (Response  and intercept must  be  in the same
anits.)  The difference  is the zero drift, which should be greater  than
«-0.3 ppmC,  or +3.0% full scale, to warrant a zero  pot  adjustment.
                                      30

-------
     To calculate  span  drift,  first  determine  the  slope  of the Level 1
calibration curve  using  the  equation:
Slope of  level  1
calibration curve
Level I span response - Level 1 zero response
    Level 1 span concentration
(3)
(Be sure to report  response  and concentration  in  units  identical  to those
used in the multipoint  calibration  curve.)   Calculate  span drift  using the
equation:
     % span drift *
                      Slope of  level  1
                      calibration  curve
                     Slope of multipoint
                      calibration curve
                       Slope  of multipoint  calibration  curve
                                         X  100   <4)
The drift should be greater than  about  +7%  to  warrant  a span pot  adjustment,
(See section 2.0.9.1.3 of  reference  10  for  more  details.)

NOTE:  BEFORE POTENTIOMETERS ARE  ADJUSTED,  CHECK ALL OPERATIONAL  PARAMETER
READINGS AND ADJUST THOSE  THAT HAVE  DRIFTED FROM READINGS  TAKEN AFTER THE
LAST MULTIPOINT CALIBRATION.  NMOC ANALYZERS ARE VERY  SENSITIVE TO CHANGES
IN OPERATIONAL PARAMETERS  (FLOW RATES,  ETC).
3.9  LEVEL 2 ZERO AND  SPAN  CHECK

     A level 2 zero and  span check may  be  used  between  the  regular level 1
zero and span checks as  an  unofficial test  of analyzer  response.   This  level
2 check may use non-traceable  or  uncertified  standards  which  may  enter  the
analyzer at a point other than the normal  sample  inlet  (e.g.,  span inlet).
No adjustments are made  to  the analyzer based on  a level  2  check,  which is
primarily a check of repeatability and  operation  rather than  of absolute
response.  If used, the  level  2 check should  be carried out  immediately
following a regular level 1 check or multipoint calibration  to obtain
reference readings.  At  subsequent checks,  if the readings  have changed
appreciably, then a level 1 zero  and span  check or a  multipoint calibration
should be carried out  to determine if the  analyzer requires  adjustment.
(See section 2.0.9.1.3 of reference  10  for  more details.)
                                       31

-------
                                  SECTION 4

                   AUGMENTED  PROCEDURES  FOR  THE  MSA 11-2
                      CONTINUOUS  NMOC/METHANE  MONITOR
     NOTE - READING  SECTIONS  1,  2  AND  3  BEFORE  SECTION 4 IS
     STRONGLY RECOMMENDED.  INFORMATION  WHICH IS  CRITICAL TO
     THE UNDERSTANDING AND  SUBSEQUENT  ACCURATE  OPERATION OF
     THIS ANALYZER IS DETAILED  IN  THOSE  SECTIONS  AND WILL NOT
     BE REPEATED HERE.
4.1  PRINCIPLES OF OPERATION

     The MSA 11-2 continuously  splits  an  ambient  sample into two aliquots
for passage into separate FIDs.   One  aliquot  is directed  to FID #1 where TOC
concentration  is measured,  and  the  second aliquot  is  directed through a
heated cutter  (Hopcalite® is used to  oxidize  all  hydrocarbons other than
methane to C02 and H20) prior to  entering FID #2,  where the methane
concentration  is measured.  The CH4 response  is then  electronically
subtracted from the TOC response  to produce the NMOC  response.   The NMOC and
CH4 responses  are then directed to  the  recorder outputs.

4.2  SYSTEM DESCRIPTION

     A description of the analyzer  with photographs  is  given in the analy-
zer's instruction manual.   To avoid redundancy, this  text  will  not address
location of components, but rather, emphasize precautions  when  using
critical components.

4.2.1  Flame lonization Detectors

     The analyzer contains  two  FIDs that  enable monitoring NMOC and CH4
concentrations simultaneously.  The response  from  the  two  FIDs  mugt be
balanced to a methane-in-air standard.

4.2.2  Hydrogen Pressure Gauges

     The flow of hydrogen fuel  to each  FID is controlled by a separate pres-
sure regulator, gauge, and  restrictor.  Increasing the  pressure reading on
                                      32

-------
 Che  hydrogen  gauge  will  increase the flow of hydrogen to the burner and
 affect  the  sensitivity of  that  FID.

 4.2.3   ^gnitor  Switches  (Flame-out  Indicators)

     Two  push button  controls  located on the front panel'control the flow of
 hydrogen  to each  FID.  When  the button is depressed,  hydrogen will flow to
 the  FID then  ignite.   When the  FID  ignites,  the light in the button will
 turn off.   During subsequent  flameouts,  a solenoid will shut off hydrogen
 flow and  the  light  in  the  button will turn on.   The flameout indicator is
 controlled  by a thermistor located  above the flame in the FID.  If the
 sensitivity of  the  thermistor  is not properly adjusted, the light in the
 button  may  not  show true  flame  status (i.e., the indicator may be off while
 the  flame is out).  Adjustments can  be made  on  the Dual Flame-Out Card to
 correct this condition.   If  the status of the flame is in doubt, observe
 analyzer  response first  to zero air  and then to a sample known to contain
methane.  If the  Clfy channel's  responses are the same to both samples, the
 CH4  channel's flame is out.   If the  response varies,  the flame is burning.
 Check the NMOC  channel's  flame  status with zero air and a sample known to
 contain NMOCs (propane).

4.2.4   Air Pressure Gauge

     Burner airflow rate,  condensate trap vent  flow rate,  sample vent flow
 rate, and sample  flow  rate to the FID are all related to the air pressure
 gauge reading on  the  front panel.  Variations in this reading indicate a
 potential malfunction  which  requires troubleshooting  (section 4.7.1.).

 4.2.5   Back—Pressure Regulator

     Located in the oven  are  two non-adjustable back-pressure regulators
 (one in older models)  that control sample flows to the FIDs by controlling
 the  pressure applied to  the  two sample capillaries.  The pressure gauges
 connected to these  regulators and the flow rates that vent from the
 regulators  should be checked  regularly because  variations  in these
 parameters  indicate a  potential malfunction  that requires  troubleshooting
 (section 4.7.1) .

4.2.6   Heated Oven

     The FID burners,  associated electronics, and pneumatics are contained
 in a heated oven maintained  at  50 to 55" C.   Small fluctuations in the room
 temperature therefore  do not  affect  these critical components.  Opening the
 oven door will  obviously cause  a loss of heat that may affect the analyzer
 response; therefore,  it  is good practice to  perform measurements and adjust-
ments on equipment  located inside the oven several hours prior to a multi-
                                      33

-------
point calibration or  level  1  zero  and  span check.   If opening the oven door
is necessary, close it as soon  as  possible to  minimize the thermal upset.

4.2.7  Oxidizer

     The oxidizer contains  palladium  and  Hopcalite® heated to between 340 to
350" C.  At this temperature, all  hydrocarbons  and  carbon monoxide are
removed, leaving a source of  hydrocarbon-free  air.   This air is primarily
used for burner air in the  FIDs but can  also be  tapped and delivered to the
FXDs as a zero  air standard.

     The use of oxidizer air  as a  zero  air standard is not recommended dur-
ing multipoint calibrations or  level  1  zero  and  span checks if an external
zero air source is used to  dilute  the  span gas  standard.  Oxidizer air can
3e used in level 2 zero and span checks.

i.2.8  Cutter

     The cutter contains MSA  brand Hopcalite®  that  is heated to between 240
and 260" C.  At this  temperature,  all hydrocarbons  except methane are
oxidized, allowing CH^ to pass  to  FID #2.   Use  only MSA brand Hopcalite®
if the catalyst ever  requires replacement.  Other catalysts may not  perform
properly at the temperature attainable  by the  cutter.  (See section  4.7.1.8
for adjustments.)

4.2.9  Sample Filter/Moisture Trap

     Compression of sample  air  by  the  analyzer  pump almost always results in
moisture condensation in the  compressed  sample  air.  It  is imperative that
none of this liquid water be  allowed  to  pass into the analyzer.  Liquid
water entering the analyzer may damage  the cutter or oxidizer, may result in
clogged capillaries or orifices, and may  damage  other flow control
components.

     On the rear of older analyzers is  a  single  Balston condensate trap for
removing water  from the sampler.   This  water removal system should be up-
graded because  it is  susceptible to malfunction.  Analyzers with this system
may have been damaged by water  during previous  operation and should  there-
fore be carefully inspected for damage  in the  cutter, oxidizer, FID, and
connecting tubing.

     On the rear of new analyzers  is  a  water removal system consisting of
two Balston filters and accessories.  This system provides much greater
protection against water entry  into the  analyzer  and is  discussed in detail
in section 4.3.2.
                                      34

-------
4.3  MODIFICATIONS

4.3.1  Introduction

     The original MSA 11-2 T/HC/Methane monitor was manufactured  during  the
early 1970s to analyze ambient air continuously for Clfy,  TOG,  and NMOC
concentrations.  Modifications to the original design have occurred  over  the
years, and an updated model called the MSA  11-2 Continuous NMOC/Methane
Monitor is now available.  The principles of operation  and detection have
not changed from the earlier model, but the  following modifications  have
been made:

 1.  A more powerful sample pump manufactured by Air Dimensions Incorporated
     has replaced the smaller Thomas pump.

 2.  The pump is no longer mounted on the analyzer chassis.

 3.  A 10-micron Hoke particulate filter has been added to the exit  of the
     oxidizer.

 4.  The moisture drop-out trap has been modified.

 5.  An activated charcoal filter has been  added to the inlet  of  the
     cutter.

 6.  Electronic balance adjustments have been added to  enable  balancing  the
     responses  from the FIDs more closely.

 7.  The range change card has been modified to allow for ranges  of  10 or 20
     ppmC full  scale.

 8.  The oxidizer volume has been increased.

 9.  The cutter has been moved from the interior of the oven to the  rear  of
     the analyzer chassis.

10.  A back-pressure regulator and gauge have been added  to the TOC  sample
     line.

11.  A modification has been made to the zero pot to enable zero  offset  to
     observe negative zero drift.

     Modifications 1 through 6 should be made to older  model 11-2 analyzers
to obtain adequate NMOC data.  The remaining modifications are less  critical
and may not be needed.  The manufacturer suggests employing a  factory-
trained serviceperson to update the unit if  the operator  is not very
familiar with this work.
                                      35

-------
4.3.2  Moisture Trap

     It is imperative  that  no  liquid water be allowed to enter the analyzer.
For maximum protection  against  liquid  water,  using a trap system as shown  in
Figure 4 is strongly recommended.   The filter (5 micron) removes par-
ticulates from the sample air  and  helps  to avoid clogging of subsequent flow
control components.  The  first  condensate trap, consisting of a collection
bowl, Balston filter,  and drain valve, collects most of the liquid water.
The valve is normally  closed,  but  is manually opened to drain the water
before the trap fills  and water is drawn over to the second trap.  In
draining this trap, a  small  amount of liquid  water should be left in the
trap to humidify dry calibration gases during calibration or zero/span
checks.  Under conditions of high  anbient humidity, the trap may fill  in
only a day or two and must  be  watched  closely and drained frequently.  More
convenient operation can be  obtained by replacing this first Balston trap
with an automatic trap  such  as  the Wilkerson  M20-02-000 or equivalent.  The
Wilkerson trap has a built-in  float valve that drains the trap automatically
as needed, but always  retains  some liquid water in the trap.

     The second water  trap  consists of a collection bowl, Balston filter,
and a vent controlled by a  back-pressure regulator.  This trap provides
secondary protection and collects  any water carryover from the first trap.
The continuous venc flow prevents  any  water buildup in this trap, which
should never show more  than  a  partial  fogging.

     Beyond the second  trap  is  a valve which  expands the sample air (by
about 4 psig), thereby  lowering its dew point and avoiding any subsequent
condensation.  This valve and  its  connection  to the trap should rise
vertically from the trap to  discourage water  droplets from approaching the
valve.  Finally, a section  of  transparent tubing (a rotameter with no ball
works fine) follows the valve.   This tubing provides a. final verification
that no liquid moisture is  entering the  analyzer.  Any sign of condensation
or fogging in this area indicates  a malfunction in the trap system.  See
section 4.5.4 for procedures to set up and adjust the system and see 4.7 for
troubleshooting procedures.
                                       36
U.S. EPA Heac.•'' :-j^. 3201
1200 "'em-^K ,.-,,- Avenue NW
   Wf:S!•;„ gior, DC  20460

-------
Balston or
Wilkerson
Trap
_ »



j


Baiston )
Trap L




                                                                                    To Analyzer
                                                                                    Inspection
                                                                                    Tube
                                                                                    Expansion
                                                                                    Valve
                   Filter
Pump
                  Drain   Valve
                                                                     Vent
                                                                            Back Pressure
                                                                            Regulator
                                  Figure 4.  Recommended water trap system.
             4.4  INSTALLATION

                   The manufacturer's  instruction manual addresses  procedures to install
             the analyzer.  Additional  comments  follow.

                   1.    Before supplying power to the analyzer, ensure  that  all electronic
                        component cards  are  properly seated by attempting  to push the
                        cards snugly into the connectors.  Be particularly certain that
                        the electrometer box located inside the oven  is  properly seated.

                   2.    Minimize the distance between the analyzer  and  the sample mani-
                        fold.  Use clean,  3  mm (1/8 in) diameter  inert  tubing to make the
                        sample line.   The sample  line length should not  exceed 3 m
                        (10 ft).

                   3.    Leak-checks should be made on the hydrogen  flow system of old and
                        new analyzers.   The  leak-check procedure  outlined  in section 2.4.2
                        can be used.
                                                    37
I

-------
      CAUTION:  ONLY EXPERIENCED, QUALIFIED  OPERATORS  SHOULD
      PERFORM THE LEAK-CHECK,  AVOID  FINGER  CONTACT WITH THE
      FITTING THREADS DUE TO POSSIBLE TRANSFER OF  OILS THAT
      WILL TEMPORARILY INCREASE THE BASELINE RESPONSE.
^.5  OPERATION

4.5.1  Start-up

     The MSA 11-2 analyzer has  two  selectable  modes  of  operation.   Channel 2
can be operated to output either TOC  (non-subtracting)  or NMOC (sub-
tracting), depending on the  location  of  a  jumper  on  the electrometer card
(see MSA Instruction Manual).   For  most  NMOC monitoring applications,  the
ItfMOC mode is much more advantageous than the TOC  mode;  therefore,  the  NMOC
'subtracting) mode is strongly  recommended.  Calibration and  zero  and  span
procedures in this TAD address  only the  CH4/NMOC  mode of operation.   The
iJJMOC mode is selected by placing the  jumper in the "S"  position.

     With the ON/OFF switch  in  the  OFF  position,  plug the analyzer's power
cord into a power source that has been determined to be stable (see  section
1.1.2 for details).  Turn the switch  to  ON, light the FID burners,  and moni-
:or oxidizer, cutter, and oven  temperatures.   If  temperatures  do not rise,
or if they rise beyond suggested levels, turn  the analyzer off and  consult
:roubleshooting procedures.  If the temperature rises smoothly,  allow the
analyzer to warm up for 24 hours before  recording the readings from the
pyrometer and pressure gauges.  With  new analyzers,  compare pyrometer  and
pressure gauge data to those supplied on the Manufacturer's Final  Data
Sheet,  If disagreements exist, notify the manufacturer and ascertain  if  the
difference is significant.   If  the  analyzer has been used before,  adjust
flows, temperatures, and pressures  to the  values  used previously.   If
previous settings are unknown,  select trial settings within the  ranges
specified in Table 9.   Flow measurement details  and other procedures  for
setting up or checking out component  operation are discussed  in  this section,

4.5.2  Periodic Calibrations and Zero/Span Checks

     An initial multipoint calibration  is  recommended to verify  linearity in
the calibration curve, and periodic level  1 zero  and span checks are there-
after necessary to update this  curve. The frequency of level  1  zero/span
checks and subsequent multipoint calibrations  should be based  on the
analyzer's performance.  Performance  can be monitored through  use  of a con-
trol chart which documents responses  to  standards and ultimately enables
calculation of variation in  response  to  standards (drift). (See section
2.0.9 of reference 10 for details on  calibration  and level 1  and level 2
zero and span checks.  See section  H  of  reference 14 for details on  control
charts.  See section 4.6 of  this TAD  for specific calibration  procedures.)
                                       38

-------
                TABLE 9. NOMINAL OPERATING SPECIFICAT1ONS
                          FOR THE MSA 11-2 ANALYZER
               Parameter
          Specifications
Front Panei

 H2 pressure reading (per gauge):

 Air pressure reading:

 Oven temperature:

 Oxidizer temperature:

 Cutter temperature:


Oven

 Back-pressure regulator gauge(s) reading:

 Back-pressure regulator vent flow rate:

 Sample flow rate to FID No. 1:

 Sample flow rate to FID No. 2:

 H2 flow rate to FID No. 1:

 H2 flow rate to FiD No. 2:

 Burner  air flow rate to FID No.  1:

 Burner  air flow rate to FID No.  2:


Back Panel
48 to 62 kPa {7 to 9 psig)

138to152kPa(20to22psig)

50 to 55° C (122 to 131° F)

340 to 350° C (644 to 662° F)

240 to 260° C (464 to 500° F)
17 to 21 kPa (2.5 to 3.0 psig)

500 ± 100cm3/min

12 to 14cm3/min

Equal to sample flow rate to FID No. t

15 to 20 cm3/min

- Equal to H2 flow rate to FID No. 1

200 ± 50 cm3/min

200 ± 50 cm3/min
 Moisture trap flow rates:
Set up and adjust as per section 4.5.4.
                                       39

-------
4.5.3  Operational Checks

     The first six parameters  listed  on  Table  9  should  be checked at least
once a week and more frequently  if necessary.  These  readings,  along with
analyzer responses, zero and span pot  settings,  and other routine checks
:hat may be necessary can be recorded  on Table 10.  (Routine monitoring may
tot require that unadjusted responses  to standards be recorded;  however,
special cases may warrant such data.)  If any  of these  parameters shows
substantial change from the last check,  or  if  drift out  of the  specified
range has occurred, recheck the  analyzer calibration  with a level 1 zero and
span check and make the necessary adjustments.
      NOTE:  FREQUENTLY CHECK THE MOISTURE TRAP  FOR  PROPER
      OPERATION DURING PERIODS OF HIGH RELATIVE  HUMIDITY.
4.5.4  Moisture TrapCheck

     If the analyzer uses the double-bowl moisture  trap  previously
described, use the following procedure  to adjust vent  flow  at  least  once  a
week, or more frequently if water  removal is  inefficient.

 I.  Adjust the expansion valve  to  full open.

 2.  Adjust the back-pressure regulator on  the  second  trap  for zero  vent
     flow.

 3.  Adjust the air pressure regulator  on the  front  panel of  the  analyzer
     to about 32 psig  (220 kPa).   If the sample pump cannot produce  32  psig
     here, see section 4.7.2.

 4.  Adjust the back pressure regulator on  the  second  Balston  trap to obtain
     30 psig (207 kPa) on the front panel gauge.

 5.  Adjust the expansion valve  to  obtain 26  psig (179 kPa) on the front
     panel gauge.

 6.  Readjust the front panel pressure  to the  correct  operating pressure,
     20-22 psig (139-152 kPa).

4.5.5  Cutter Temperature

     The cutter temperature should  be checked  and adjusted  during  start-up
and approximately every 6 months of operation  thereafter.   It  should also be
checked and adjusted if:  (1) the  cutter is replaced,  (2) the  cutter
temperature changes, (3) the analyzer shows abnormal span drift, or  (4) the
CH4 channel responds to non-methane compounds.  (See section 4.7.1.8 for
details.)
                                      40

-------
HI

X
cc
IL

<
O
°i
•I
Q-
cc <
ow

2s
cc

cc
O
u.

S
cc
o
o
^>

111
_l
CD

J
*

f
I

r




1
2
i


I
2
!

1
1
\
1.

I
2
1


}
i
i



<
4


j
1
i
M
<
r
«~
1
1


3
}
1
1
E
II
;-
i
1


i
1
I
i
s

1
if
!
!
!
i
*

t
*
1
f
1
5*
J
if
|
a'
I
a*
j
\
|
8'
1
5*
I
5*
|
5*













































































































































































































































































































































































































































































































1
1
























































































































































































































































































































                     41

-------
i.6  CALIBRATION

     This section details specific  procedures  for  multipoint  calibrations
and zero and span checks of the MSA  11-2.  The  analyzer's  instruction manual
nay contain deviations from these procedures,  but  the  step-by-step proce-
dures that follow are currently recommended by  the EPA.  These procedures
cover the CHa/NMOC mode of operation only  (section 4.5.1)!

4.6.1  Multipoint Calibration

 1.  If this is an initial calibration, or if  the  cutter  temperature is  due
     for checking (section 4.5.5),  it  should be  checked and  adjusted prior
     to the calibration according to the procedure given  in  section 4.7.1.8.

 2.  To conserve calibration gas during calibration, shut  off  the  vent  flow
     from the moisture drop-out trap located on  the back of  the  analyzer.
     Quickly check the back-pressure regulator gauges  located  inside the
     oven before and after vent flow is shut off.   If  the  pressure reading
     varies before and after shutting  off  the moisture  trap  vent,  trouble-
     shoot the back-pressure regulators located  in the  oven.

 3.  Place the range selector switch in the range  most  frequently  used
     during anbienC monitoring.

 4.  Disconnect the analyzer's sample  line from  the ambient  manifold and
     reconnect it to the calibration system's manifold.  Cap  the port on the
     ambient manifold and any open  ports on the  calibration  system's sample
     manifold.

 5.  Supply an atmosphere of the zero  standard  to  the manifold at  a flow
     rate that is 20 to 50% greater  than the analyzer's sample flow demand.
     DO NOT PRESSURIZE THE ANALYZER'S  SAMPLE INLET.  The test  atmosphere
     must contain an ambient level of  oxygen and must  not  contain  more  than
     0.1 ppmC TOC.  See section 3.2  for details.

 6.  Adjust the analyzer's zero controls for both  channels to  the  desired
     baseline responses.  A 5% of full scale positive  offset  on  the
     recording device is recommended to observe  any negative  drift.  Use
     either the analyzer's zero pots or the recorder's  controls  to obtain
     the offset.  Ensure that the responses from both  channels are equal
     before recording the responses  on forms similar to Tables 5 and 10.

 7.  Supply an atmosphere of methane standard  to the calibration manifold at
     a flow rate that is 20 to 50% greater than  the analyzer's sample flow
     demand.  The methane concentration should be  between  70 and 90% full
     scale.

 8.  Adjust span pot #2 (which controls CH4 response from  FID  #2)  to
     provide the analyzer response calculated  as follows:
                                      42

-------
Response
'Sample concentrat ion
I        URL
         X url  + Zero offset
(5)
where

Response • Response of  the  recording device measuring  the
           analyzer output  in  recording  device  units.

Sample concentration =  Concentration of  the calibration  standard
                        delivered  to the  analyzer  in  ppmC.

URL * The upper range limit of the analyzer in  ppmC.

url ™ The upper range limit of the recording  device  in recording device
      units .
Zero offset
   The amount  the recording device response is set above the
   zero  baseline  while the analyzer is measuring the zero
   calibration standard (in recording device units).
NOTE:  The response, url, and zero offset must  be  in  identical  units.
For example, if a strip chart recorder  is the  recording  device, units
will be % Chart and the url may equal 100% chart.   If  a  digital
voltmeter (DVM) is the recording device, units  may be  millivolts or
volts and the url may be, for example,  1000 mV.   If a  data  acquisition
system (DAS) is the recording device, units may be ppraC  and the url
will equal 10 ppmC.  If the front panel meter  on the  analyzer  is the
recording device, units will be ppmC and the url may  be  either  10 or 20
ppmC.
For example,

     methane span concentration
                           8.0 ppmC,
     analyzer range being monitored *  10  ppmC

     recorder range being used      =  100% chart

     recorder offset                «  5%  chart
Response
   8_.Q ppmC
   10 ppmC
X 100%)
                                 +  5% - 85%  chart
 (6)
Therefore, the analyzer's span potentiometer  should  be  adjusted  until
the analyzer output  is equal  to 85% chart  on  the  recorder  while
sampling the 8.0 ppmC methane standard.
                                 43

-------
     If step 8 results in a span  pot  setting  greater  than 300,  decrease
     pressure to H2 gauge #2 until  the  Clfy response has  increased to a
     point where the span pot can be  reduced  to  about 250.   If  the span pot
     setting is less than 200,  increase  pressure  to H2 gauge #2 until the
     CH4 response has decreased to  a  point where  the  span pot can be
     increased to about 250.  DO  NOT  EXCEED 9.0 psig  H2-
        NOTE:  IF THE FID RESPONSE  INCREASES WHEN  INCREASING
        HYDROGEN PRESSURE, THE BURNER  IS OPERATING ON  THE  WRONG
        SIDE OF ITS PEAK RESPONSE CURVE.  CONTINUE INCREASING
        HYDROGEN PRESSURE UNTIL RESPONSE STARTS  TO DECREASE  AND
        THE CORRECT SPAN POT SETTING IS OBTAINED.
        (SEE SECTION 4.7.5 IF THE PROPER RESPONSE  IS UNOBTAIN-
        ABLE OR IF DIFFICULTIES DEVELOP.)
10,   Set span pot #1 NMOC to the same dial setting  (not  response)  as  span
     pot #2.

11.   Repeat steps 5, 6, 7, 8, 9, and 10 if span  pot  adjustments  were
     necessary or if hydrogen pressure gauge #2  was  adjusted.

12.   Continue sampling the methane standard.  For the  FIDs  to  be balanced,
     the NMOC response must be within +1.0% of the  zero  response obtained in
     step 6.  If the FIDs are balanced, go to step  14.

13.   Balance detectors:  If the FIDs are not balanced:

     (a) Check that both span pots are dialed to the same setting.

     (b) Locate the range change board (inside the  electrometer  assembly
         located in  the oven) and rotate the balance  controls of channels  1
         and 2 (accessible through holes in the  electrometer cover) to  their
         maximum clockwise position.  Make adjustments quickly to  minimize
         heat loss!
          NOTE:  ONLY THE 1980 11-2 MODELS HAVE ELECTROMETER
          BALANCE CONTROLS.  USERS OF THE OLDER MODEL MUST
          UPDATE THEIR ANALYZER TO INCLUDE THIS MODIFICATION.

-------
14.
 (c) Record both hydrogen  pressure  gauge  readings.   Adjust  hydrogen
     pressure gauge #1 until  the NMOC  response  to  the methane  standard
     equals the earlier NMOC  response  to  the  zero  standard  (step 6).

     Do not adjust the hydrogen pressure  outside the 34-62  kPa (5 to
     9 psig) range.   Increase hydrogen pressure to  decrease NMOC
     response, or vice versa.  See  the note in  step 9.

 (d) If balance is achieved,  repeat  steps  5 through 13.

 fe) If balance is not achieved by  step c, adjust  the NMOC  hydrogen
     pressure (gauge  #1} or both hydrogen  pressures, if  necessary, to
     get as close to  balance  (NMOC  response within +_ 1%  of  NMOC zero
     response) as possible.   Do not  exceed the  34-62 kPa (5-9  psig)
     hydrogen pressure range. Then,  using  the balance controls on the
     range change board located inside the electrometer  (accessible
     through holes in the  electrometer cover),  rotate the channel #2
     balance control  counter-clockwise to  increase  the NMOC response,
     or rotate the channel #1 balance  control counter-clock wise to
     decrease the NMOC response.  Repeat  steps  5 through 12.   If
     balance cannot be achieved, see section  5.7.5.

After the FIDs are balanced,  record  responses on Tables  5 and  10.
Determine the CIfy channel's response to two or  more additional con-
centrations of the methane standard  that  are  spaced  approximately
equally over the analyzer  range.  (The NMOC channel's response should
remain equal to the earlier response to zero  standard.)   Record the
CH4 and NMOC channel's responses (from the recording device) on
Table 5.  Using a calculator, perform  a least squares linear regression
of the CH4 channel responses  (to methane  and  zero  standards) and the
corresponding calibration  concentrations.  The  calibration  concentra-
tions should be in units of ppmC and should be  entered into the cal-
culator as the independent variable  X.  The CH^ channel's response
should be in units of the  recording  device and  should be entered as the
dependent variable Y.  A correlation coefficient (r) of  0.9996 or
better verifies that  the CH^  response  is  linear.   (If the response is
not linear, plot the  data  and determine if an error has  been made in
data entry or in determination of calibration concentration.)   Obtain
the slope and intercept of the regression  and record the equation in
the following form:
     CH
  ;, Response * CH, Slope x Methane  concentration  +  CH,  Intercept    (7)
                                      45

-------
     where

     CH4 Response = Analyzer's CH4 channel  reading  in  recording  device
                    units (see note  following  step  8).

     CH4 Slope * Regression slope in recording device  units  per  ppmC.

     Methane concentration » Calibrated methane concentration  in ppmC.

         Intercept = Regression intercept  in recording  device  units.
Post the CH^ channel's multipoint calibration curve  equation  on  the
analyzer's recording device and also on Tables  5 and 10.

15.  Supply an atmosphere of propane standard to the calibration manifold  at
     a flow rate that is 20 to 50% greater than the  analyzer's sample  flow
     demand.  The propane concentration should be between 70  and 90% of  full
     scale.

16.  Adjust span pot #1 (which controls NMOC response  from  FID #1) to  pro-
     vide the desired analyzer response.  Equation 5 in  step  8 can be  used
     here.

17.  If an adjustment is made in step 16, recheck the  NMOC  channel's
     response to the zero standard and adjust zero pot #1 if  necessary.
     Record the stable zero responses from the NMOC  and  CH^ channels.
     Sample the propane standard, and again record the stable responses  to
     propane. (Cfy response should be equal to the earlier  response  to
     zero air.)

18.  Determine the NMOC channel's response to two or more additional concen-
     trations of the propane standard that are spaced  approximately  equally
     over the analyzer range.  (The CH^ channel's response  should remain
     equal to the earlier response to zero standard.)  Record the NMOC and
     CH4 channel's responses (from the recording device) on Table 5.
     Using a calculator, perform a least squares linear  regression of  the
     NMOC channel's response (to propane and zero standards)  and the
     corresponding propane calibration concentrations.   The calibration  con-
     centrations should be in units of ppmC and should be entered into the
     calculator as the independent variable X.  The  NMOC channel's response
     should be in units of the recording device and  entered as the dependent
     variable Y.  A correlation coefficient (r) of 0.9996 or  better verifies
     that the NMOC response is linear.  (If the response is not  linear,  plot
     the data and determine if an error has been made  in data entry  or in
     determination of calibration concentration.)  Obtain the slope  and
     intercept of the regression and record the equation in the  following
     form:
                                      46

-------
     NMOC  Response *  NMOC  Slope  x NMOC  Concentration •*• NMOC Intercept   (8)
     where

     NMOC Response * Analyzer's  NMOC  channel  reading in recording device
                   units  (see  note  following  step 8).

     NMOC Slope * Regression  slope  in recording  device units per ppmC.

     NMOC Concentration » Calibrated  NMOC concentration in ppmC.

     NMOC Intercept » Regression intercept in recording device units.

     Post the NMOC channel's multipoint  calibration  curve equation on the
     analyzer's recording device and  on  Tables 5 and 10.  Record all zero
     and span pot readings  in  Table 10.

19.  Disconnect the analyzer's sample line from  the  calibration system's
     manifold and reconnect  it to the ambient manifold.

20.  Follow the procedure in  section  4.5.4 to reset  the moisture trap vent
     flow rate.

21.  Record readings  from the  first seven analyzer parameters listed in
     Table 10.  These readings will now  be the reference readings until
     another multipoint calibration is performed.   Be  certain that the
     station logbook, strip charts, etc.,  are properly annotated, dated, and
     initialed.

22.  If  level 2 zero  and  span  checks  are to be used, carry out such a check
     now to establish reference  zero  and span readings for subsequent com-
     parison.  (See section 2.0.9.1.3 of reference 10  for definitions and
     additional information on level  1 and 2  zero  and  span checks.)

4.6.2  Level 1 Zero/Span  Check

     The following level  1 zero/span  check (2-point  calibration) is similar
to a multipoint calibration except  only  the zero air standard and one up-
scale standard are used.  These  responses  are then used to calculate the new
calibration curve which is used  to  correct  subsequent  monitoring, data until
another level 1 zero/span check  or  multipoint calibration is performed.   As
part of an overall quality assurance  program,  the  operator may want to
record unadjusted responses to standards  before  making any changes to opera-
tional parameters or zero and  span  potentiometers.

1.   Disconnect the analyzer's sample line from  the  ambient manifold and
     reconnect it to  the  calibration  system's output sampling manifold.  Cap
     the sample port  on the  ambient manifold  and any open ports on the cali-
     bration system's sample manifold.
                                      47

-------
 2.   To  conserve calibration gas during calibration, shut off  the vent  flow
     from the moisture trap located on the back of the analyzer.  Quickly
     check the back-pressure regulator gauges located inside the oven
     immediately before and after the moisture trap vent is shut off.   If
     the pressure reading varies before and after shutting off the moisture
     trap vent,  troubleshoot the back-pressure regulators located in the
     oven.

 3.   Supply an atmosphere of the zero standard to the calibration manifold
     at  a flow rate of 20 to 50% greater than the analyzer's sample flow
     demand.   DO NOT PRESSURIZE THE ANALYZER'S SAMPLE INLET.

 4.   After  obtaining a stable zero reading, determine if zero drift is
     greater  than about _+ 3,0% (see section 3.8 for zero drift calculation).

     If  necessary,  adjust the analyzer's zero pots to obtain the correct
     responses.   Record the zero responses on forms similar to Tables
     6 and  10.

 5.   Supply an atmosphere of methane standard to the calibration manifold at
     a rate 20 to 50% greater than the analyzer's sample flow demand.   The
     methane  concentration should be between 70 and 90% full scale.

 6.   After obtaining a stable Clfy reading, use the span drift equation
     from section 3.8 (Equation 4) to calculate the amount of span drift on
     the CH^  channel.  If the drift is greater than about _+_ 7.0%, perform
     steps  8  through 21 in the multipoint calibration procedure in section
     4.6.1.  (Linearity assessment is not necessary in steps 14 and 18.)

 7.   If  span  drift  is within specifications, check to see that the NMOC
     channel's response is within _+_ 1.0% of the NMOC channel's earlier
     response when  sampling the zero standard.  If so, the FIDs are
     balanced.  If  the FIDs are not balanced, perform steps 8 through 21 in
     the multipoint calibration procedure, section 4.6.1.  (Linearity
     assessment  is  not necessary in steps 14 and 18.)

 8.   If  the span drift is less than about +_ 7.0% and the FIDs  are balanced,
     record the response and determine the slope and intercept of the Cfy
     channel's level 1 calibration curve.  Post the level 1 calibration
     curve equation on the analyzer's recording device and on Tables 6  and
     10.

 9.   Supply an atmosphere of propane standard to the calibration manifold at
     a rate 20 to 50% greater than the analyzer's sample flow demand.   The
     propane  concentration should be between 70 and 90% of full scale.

10.   After obtaining a stable span reading, use the span drift equation from
     section 3.8 (Equation 4) to calculate the amount of span drift on  the
                                      48

-------
     NMOC channel.  If  Che drift  is  greater  than  about _+  7.0%,  adjust  span
     pot #1 to provide  the proper  analyzer response.   (Use  Equation 5  in
     step 8, section 4.6.1,  to  calculate  response.)

11.  Record the responses and determine the  slope and  intercept  of the NMOC
     channel's level 1  calibration curve.  Post  the NMOC  channel's level 1
     calibration curve  equation on the analyzer's recording device and in
     Tables 6 and 10.   Record all  zero and span  pot readings on Table  10.

12.  Reconnect the sample line  to  the ambient manifold.

13.  Reopen the moisture trap vent and adjust vent  flow  according to the
     procedure given in section 4.5.4.

14.  If level 2 zero and span checks  are  to  be used,  carry  out  such a check
     now to establish reference zero  and  span readings for  subsequent  com-
     parison.

4.7  TROUBLESHOOTING

     The manufacturer's instruction  manual covers this subject  under the
same heading.  Additional comments follow.

4.7.1  Preliminary Investigations

     Operational problems such  as  slow response  to  standards (response time
greater than 10 seconds), inability  to balance FID  responses, long stabili-
zation periods (greater than 5 minutes),  frequent flame-outs, or low sensi-
tivity indicate the potential need for in-depth  troubleshooting.  Prior to
undertaking in-depth troubleshooting, a preliminary investigation is recom-
mended to rule out variations in  operational parameters.

 1.  Check that all pressure gauges  on the analyzer are  set to  the specifi-
     cations stated on  the Manufacturer's Final  Data  Sheet  (or  to the  speci-
     fications stated in Table  9  if  a manufacturer's  data sheet  is not
     available).  If the analyzer  has been operating  properly for several
     months, ensure that the pressure gauges are  set  to  the readings
     recorded during the most recent  multipoint  calibration when the FIDs
     were balanced.

 2.  Inspect the hydrogen source  and  hydrogen filters  (located  outside the
     analyzer) for proper operation.  If  neither  hydrogen pressure gauge on
     the analyzer can be adjusted  to  its  proper  setting,  see section 4.7.4.

 3.  Inspect the sample filter  and replace it if  dirty.   Replace the sample
     line leading to the analyzer  if  particulate  buildup-is noticeable.
                                       49

-------
Perform the moisture trap setup  procedure  outlined  in section 4.5.4.
If the sample pump cannot produce  32  psig  head  pressure during the pro-
cedure, see section 4.7.2.

Determine if the sample vent  flow  rates  through the back-pressure regu-
lators (located in the oven)  are within  specifications.

(a)  Open the door and connect tubing to a back-pressure regulator
     vent.

(b)  Run the tubing outside  the  oven  and tape  the door closed.  Do not
     crimp the tubing.

(c)  Connect a flow meter to  the tubing  and measure the flow rate.  The
     rate should be approximately  500 cm^/min.

(d)  If the flow rate from either  vent differs  by more than 5% from the
     flow rate measured during the most  recent  period of normal opera-
     tion, see section 4.7.2.

Measure the sample flow rates  into the FIDs.

(a)  Disconnect a sample  line  from the base of  the  FID.

(b)  Connect an adapter to the nut of the  sample line.

(c)  Connect tubing to the adapter, run  the tubing  out of the oven, and
     tape the oven door closed.  Do not  crimp  the tubing.

(d)  Connect a flow meter to  the tubing  and measure the flow rate.  The
     rate should be between  10 and 15 cm^/min.

(e)  If either sample flow rate  differs  by more than 5% from the flow
     rate measured during the most recent  period of normal operation,
     see section 4.7.6.

Measure the burner air flow  rates  into the FIDs.

(a)  Disconnect a burner  air  line  from the side of  the FID.

(b)  Connect an adapter to the nut on the  burner air line.

(c)  Connect tubing to the adapter, run  the tubing  out of the oven, and
     tape the oven door closed.  Do not  crimp  the tubing.

(d)  Connect a flow meter to  the tubing.

(e)  Measure :he  flow rate.   The rate should be approximately
     200 + 50 cra3/min.
                                 50

-------
      (f)   If  etcher  burner  air  flow rate  differs  by more than 5% from the
           flow  rate  measured  during the most  recent period  of normal
           operation, see  section 4.7.7.

 8.   Perform  the  cutter temperature adjustment  procedure.

      (a)   Locate  the "cutter  temperature  card"  in the  top  of the
           analyzer.   (See  instruction  manual  for  location  and detailed
           photograph of the card.)

      (b)   Sample  a methane  standard until a stable methane  response is
           obtained.

      (c)   Slowly  increase cutter temperature  (wait 30  minutes after each
           adjustment) until a decrease in methane response  is noticed.

      (d)   Allow the  response  to  stablize  (1 hour) and  record the temperature
           as TI .

      (e)   Slowly  reduce the cutter  temperature  and observe  the increasing
          methane response.

      (f)   When  methane response  cannot be increased by a further cutter
           temperature decrease,  allow  the response and temperature to
           stabilize  and record  the  temperature  as T£.

      (g)   The difference between T2 and Tj should be at  least 7* C
           (10*  F).   If not, decrease T2 until the difference is  obtained.
           No change  in methane  response should  occur.

4.7.2  Sample Pump System Investigation

     The sample pump supplies the sample, burner  air,  moisture trap vent,
and sample vent to specific areas in the  analyzer.   Deterioration of the
pump, blockages in flow systems,  or leaks in  flow systems can alter the flow
rate of one or more  of these  critical  flows.  If  preliminary investigations
identify a significant change in the sample vent  flow  rate  (flow rate
differs by more than 5% from  the  flow  rate measured during  the most recent
period of  normal  operation),  or  the proper AIR  pressure  reading  cannot be
obtained or the sample pump cannot  produce at least 32 psig  during the
moisture trap setup,  carry  out  the  following  investigation.

 1.   Inspect the  sample filter  and  intake to  the  pump  for obstructions.

      (a)   Clean or replace  as necessary.

 2.   Inspect the  tubing between  the pump  outlet and AIR  pressure gauge for
      an obvious leak (hissing will  be  detected).

      (a)   Repair or  replace the  leaky  element.
                                      51

-------
 3.  If adjustments were made  in  steps  1  or 2 ,  remeasure the sample
     vent flow rate (4.7.1.5).   If  flow rates  are not within 400 to
     600 cra
     (a)  Adjust  the  flow  control  valve located inside the analyzer (see
          flow diagran  in  the  instruction  manual!"!

     (b)  If the  flows  cannot  be  properly  adjusted,  ensure that the sample
          pump is  functional before  replacing  the  back-pressure regulator.

 4.  Perform the  moisture  trap set up  procedure (4.5.4).

     (a)  If the  pump pressure cannot  produce  32 psig, clean or replace the
          pump diaphragm.

4.7,3  Moisture Trap  Investigation

     If the second moisture trap  begins accumulating water,  immediately
investigate the system.

 1.  Ensure that  the  first moisture  trap is  neither  flooded  nor installed in
     reverse.  (Observe the arrows located on  the  top of  the trap for flow
     direction. )

 2.  Perform the  setup  procedure  outlined  in section 4.5.4.

 3.  Perform troubleshooting steps in  section  4.7.2.

4.7.4  Hydrogen System  Investigation

     If the hydrogen  pressure  gauge  readings begin  to drift, investigate the
system.

 1.  Ensure that  the hydrogen  source is not  depleted and  that  no leaks or
     obstructions exist in the tubing  delivering the gas  to  the analyzer.

 2.  Measure the  hydrogen  flow rate  to the FID.

     (a)  Disconnect a hydrogen line from  the  side of the distribution
          manifold.

     (b)  Connect an adapter to the  fitting  located  on the side of the dis-
          tribution manifold.

     (c)  Connect tubing to the adapter, run the tubing out  of the oven, and
          tape the oven door closed.   Do not crimp the tubing.

     (d)  Connect a flow meter to the  tubing.

     (e)  Depress the ignitor  button on the  front panel to allow hydrogen
          gas to  flow through  the flow meter.
                                                         U.S. EPA Headquarters Library
                                                                Mai! code 3201
                                                         1200 Penn?y>vani& Avenue NW
                                      52                   Washington  DC 20460

-------
      (f)  Measure  the  flow rate.   The rate  should be approximately
          15  to  20  cm3/min.

      (g)  If  the flow  rate differs by more  than  5% from the flow rate
          measured  during  the most recent period of normal  operation, see
          step 3 below.

 3.   Inspect  the hydrogen  restrictors in  the  fittings connected to the dis-
      tribution manifold  for debris and clean  or  replace as  necessary.
      (Clean or replace both fittings  if one needs cleaning  or replacing.)

      (a)  Remove the hydrogen line fitting  from  the exit  side of the distri-
          bution block.

      (b)  Soak the  fittings in  a  solvent  (methyl alcohol)  for several hours
          (ultrasonic bath can  hasten the cleaning).

      (c)  Connect  a clean  air pressure line to  the outlet  of the fitting and
          force  air through the fitting (in reverse) while  the fitting is
          submerged in solvent.

      (d)  After  unrestricted  flow is  obtained through the  fitting, remove
          the fitting  from the  solvent and  continue to pass air through it
          until  all solvent is  evaporated.

      (e)  Install the cleaned restrictor  and  measure the hydrogen flow.

 4.   Determine if  any other part  of the hydrogen flow system contains
      obstructions.  If so,  replace or attempt to clean by  reverse flushing
      with methyl alcohol.

 5.   Leak test the  hydrogen flow  system and determine if  a leak is present.
      (Use the procedure  detailed  in section 2.4.2.)

 6.   Check the hydrogen  gauge for possible  malfunctions.

4.7.5  FID Response Investigation

      If the FID  response increases with increasing hydrogen pressure, or the
FIDs  cannot be balanced, or the factory-specified pressure  or flow settings
do not produce satisfactory analyzer  operation,  and no flow abnormalities
exist, new settings must be established.  The complex interrelationship
between hydrogen pressure, hydrogen flow, and FID response  for the two
burners makes this  difficult  to do by trial and  error. A  better approach is
to determine  the complete  pressure-flow and flow-response  relationships for
each  FID over a  wide range, and then  use  these  relationships to pick a
viable operating point.
                                     53

-------
 1.  Measure hydrogen flow races  into  the  FIDs  at  several  different  hydrogen
     pressure settings.  Cover the  5 to  9  psig  range  of  the  gauges  and plot
     flow rates against pressure  gauge readings.

 2.  Measure both sample flow rates into the FIDs.

 3.  Reset the moisture trap vent flow to  ensure that  it  is  properly
     adjusted (see 4.5.4).

 4.  Place the jumper on the electrometer  card  in  the  N  (non-subtract ing)
     position to permit recording of the TOG and Cfy.  responses.   Adjust
     the two electrometer balance controls to their maximum  clockwise
     position and the two span controls  to their maximum  position.   Allow
     the analyzer to re-establish thermal  equilibrium.

 5.  Disconnect the analyzer's sample  line from the ambient  manifold and
     reconnect it to the calibration system's manifold.

 6.  Supply an atmosphere of methane standard to the  calibration  manifold  at
     a rate 20 to 50% greater than  the analyzer's  sample  flow  demand.

     The methane concentration should be between 70 and 90%  full  scale.

 7.  Reduce hydrogen pressure on gauges #1 and #2  until each gauge reads 5.0
     psig (35 kPa) .  (If a flame-out occurs, increase  H2 pressure to the
     point where the flame can be sustained.) Record  the hydrogen gauge
     readings, along with the corresponding response  (H2 #1  controls TOG
     response and H2 #2 controls  CH4 response).

 8.  Increase both hydrogen pressure gauge readings in intervals  of  0.5 to
     1.0 psi.  Record stable responses from both channels  to each hydrogen
     pressure reading.  Continue  increasing hydrogen  pressure  up  to  9.0 psig
     (62 kPa).  Operating the burners  at high hydrogen flow  rates for
     extended periods may be detrimental to the burner components; there-
     fore, return hydrogen pressure to 7.0 psig after  obtaining the  last
     response.

 9.  Repeat step 8 while sampling the zero air standard.

10.  Using the hydrogen pressure-flow relationship from step 1, calculate
     the hydrogen to sample flow  rate ratios for each hydrogen setting.
     Plot the hydrogen to sample  flow rate ratios on the abscissa, and the
     corresponding analyzer responses to methane and  zero  on the  ordinate
     for each FID.  NOTE:  Analyzer response to methane should be the  net
     response.  Net response is the actual methane response  less  the zero
     response at each hydrogen pressure setting.   See  Figures  5 and  6.
                                     54

-------
                                  NET RESPONSE
         1.0    1.2    1.4    1.S    1.8

         H. TO SAMPLE FLOW RATE RATIO
 Figure 5.  CH4 channel's response curves—effects of variations
           in H2 flow rate. Sample flow rate @ 10.2 cm3/min,
           net response » response to constant methane concen-
           tration less response to  zero air.
100 -

 80 •

 80 .

 70 •

 60 •

 SO .

 40 •

 30 •

 20 •

 10 •

  0 .
                              NET RESPONSE
                              ZERO RESPONSE
         1.0
               1.2
                      1.4
                           1.8
                            1.8
                                       2.0
         H2 TO SAMPLE FLOW RATE RATIO
Figure
6.  TOO channel's response curves—effects of variations
    in H2  flow rate. Sample flow rate @ 10.2 cm^/min,
    net response = response to constant methane concen-
    tration less response to zero air.
                            55

-------
11.  If either  response  curve  contains  a peak,  adjust the hydrogen pressures
     to selected values  which  produce  approximately equal CH4 and TOC
     responses.  Preferably, the  resulting  hydrogen setting should be near
     or somewhat to  the  right  of  the  peak response hydrogen setting.

12.  If the response curves  from  step  10 do not  contain a peak,  or if the
     peak is contained on  a  broad plateau,  adjust  the hydrogen pressures to
     selected values which produce  approximately equal CH^ and TOC
     responses, preferably at  hydrogen  to sample ratios of about 1.0 to 1.4.

13.  Reconnect  the sample  line to the  ambient manifold and place the jumper
     on the electrometer card  in  the  S  (subtracting)  position to permit
     recording  of NMOC and CH^ responses.

14.  Recheck analyzer operation to  see  if proper balance and calibration can
     now be achieved.  If  not, pick new operating  points from the response
     curves, adjust  parameters, and recheck analyzer  operation.

4.7.6  Sample Flow Rate  Investigation

     If either  sample flow rate to  the  FID  (4.7.1.6)  differs by more than 5%
from the flow rate measured  during  the  most recent  period of normal  opera-
tion and the sample  pump system has been eliminated as the problem (section
4.7.21, carry out the following investigation.

 1.  Ensure that both sample capillaries produce the  same flow when  con-
     nected to  the same  outlet of the distribution manifold.  If flows  are
     not within +_ 0.2 cm^/min  of  each other, clean both in an ultrasonic
     bath of methyl  alcohol, and  then recheck flow rates.  If differences
     still exist, order  new, matched  capillaries from the manufacturer.

 2.  Determine  if any other  part  of the sample  flow system contains  obstruc-
     tions.  (Measure flow rates  from both  sample  ports of the distribution
     manifold.  They  should be  equal.   If not, measure the flow rate  entering
     and exiting the cutter.   If  these  flow rates  are not equal,  the cutter
     is clogged.  Replace  the  cutter  and check  flow rates.)
     Leak test the  sample  flow  system.
     section 2.4.2.)
(Use the procedure detailed in
4.7.7  Burner Air Flow Rate  Investigation

     If either burner air  flow  rate  to  the  FID  (4.7.1.7)  differs  by more
than 5% from the flow rate measured  during  the  most  recent  period of normal
operation, and if the sample  pump  system has  been  eliminated  as  the problem
(4.7.2), carry out the following investigation.
                                       56

-------
Inspect the burner  air  restrictors  in  the  distribution manifold for
debris and clean or  replace  as  necessary.   (If one needs  cleaning or
replacing, clean or  replace  both  restrictors.)

(a)  Remove the burner  air line  from the exit  side of the distribution
     block.

(b)  Soak the  fittings  in  a  solvent (methyl  alcohol)  for  several hours
     (an ultrasonic  bath can hasten the cleaning).

(c)  Connect a clean air pressure line to  the  outlet  of the fitting and
     force air through  the fitting  (in reverse)  while the fitting is
     submerged in solvent.

(d)  After free flow is obtained  through the restrictor,  remove the
     fitting from the solvent  and continue to  pass air through it until
     all solvent has evaporated.

(e)  Install the cleaned restrictor and measure the burner air flow as
     before.

Determine if any other  part  of the burner  air  flow system contains
obstructions.  If so, replace  or  attempt to  clean by reverse flushing
with methyl alcohol.  (If  any  obstruction  is located  in the oxidizer,
replace or repack the unit.)

Leak test the  burner air system and determine  if a leak is present.
(Use the procedure detailed  in  section 2.4.2.)
                                 57

-------
                                  SECTION  5

                  AUGMENTED  PROCEDURES  FOR  THE  BENDIX 8201
                   AND  8202  REACTIVE  HYDROCARBON ANALYZER
     NOTE: READING  SECTIONS,  1,  2,  AND  3  BEFORE  SECTION 5  IS  STRONGLY
     RECOMMENDED.   INFORMATION  WHICH  IS CRITICAL TO  THE UNDERSTANDING
     AND SUBSEQUENT ACCURATE  OPERATION  OF THIS ANALYZER IS DETAILED IN
     THOSE SECTIONS AND WILL  NOT BE REPEATED  HERE.
The manufacturer has prepared  a  relatively  detailed  instruction manual  which
can guide users during  installation  and  operation  of the  model  8201  and 8202
inalyzers.  The thoroughness of  the  manual  will  therefore  allow this
section to be brief.  As stated  in the  introduction,  the  purpose of  this TAD
Is to supplement the analyzer's  instruction manual by augmenting critical
installation and operation  procedures.

5.1  PRINCIPLES OF OPERATION

The analyzer's  instruction  manual contains  a  discussion on  this subject in
;he "Introduction".  In the discussion  the  term  "mechanical separating
device" is used in place of chromatographic column;  and the  "detector" cell
is a hydrogen flame  ionization detector  (FID).   Both  analyzer models have
essentially the same design, so  a brief  summary  of their  operation follows.

Ihe analyzers contain one FID  and one sample  loop.   At  the  start of  the 200
second cycle, sample is directed from the sample loop into  the  mechanical
separating device.   The methane  component of  the sample emerges first  and is
injected  into the FID for measurement.   As  the separating device is  back-
flushed,  another sample from the sample  loop  is  directed  straight  to the FID
for measurement of TOC.  Internal subtraction of the  CH4  and TOC measure-
ments results in the NMOC reading.

5.2  SYSTEM DESCRIPTION

     A detailed description of the analyzer is given  in the analyzer's
instruction manual.  This TAD  will therefore  not address  location of com-
ponents,  but will emphasize the  uses and precautions  of those which  are
critical.
                                      58

-------
5.2.1  Flame  lonization Detector

     The analyzer contains  one  FID.  The  separate  responses  of the Clfy. and
TOC channels  must be  balanced to  a methane-in*air  standard.

5.2.2  Hydrogen Pressure Gauge

     The flow of hydrogen  fuel  to  the  FID is  controlled  by a pressure regu-
lator, gauge, and restrictor.   Increasing the pressure  reading on the
hydrogen gauge will  increase the  flow  of  hydrogen  to  the  burner and affect
the sensitivity of  the FID.

5.2.3  Flameout Indicator

     If a condition exists  which  causes the hydrogen  flame in the FID to
extinguish, the analyzer will either automatically attempt reignition or
will shut off flow of hydrogen  to  the  FID.  The  action  is  determined by the
position of toggle  switch  S7 located on the front  panel.   In the event of a
flame-out, the flame-out indicator will illuminate.   If  the  status of the
flame is in doubt,  observe  analyzer response  first to zero air and then to a
sample known  to contain hydrocarbon.   If  the  TOC channel's responses are the
same to both  samples, the  flame is out.

5.2.4  Sample/Back-Flush Valves

     The 0-rings of these  two valves are  susceptible  to  wear and thus
leakage.  Replace 0-rings  in both  valves  every 6 months,  or  sooner if leaks
develop.  The 8202 manual contains details on specific  leak-check proce-
dures .  A procedure  for leak-checking  these valves, and  the  sample flow
system, is detailed in section  5.7.3 of this  TAD.

5.2.5  Columns

     To prevent column contamination,  the columns  should  be  capped whenever
they are removed from the analyzer or  whenever the analyzer  is powered-down
for an extended period of  time.

5.3  MODIFICATIONS

     The model 8201 analyzer has been  slightly modified  to produce the model
8202.  The principle of operation  and  detection  is unchanged.   Differences
between the two models follow.

5.3.1  Relocation of Timer

     The MANUAL CYCLE POSITION  TIMER and  the  CH4,  THC and  ZERO FUNCTION
LIGHTS have been moved from inside the 8201 analyzer  to  the  front  panel  of
the 8202.
                                      59

-------
5.3.2  Instruction Manual

     The analyzer instruction manual  has  been  rewritten.

5.3.3  Calibration Procedure

     The calibration procedure  for  the  8202  calls  for  a zero air standard
and three methane/ethane  in air  span  standards.   (See  section 5.6 for the
EPA recommended calibration procedure.)   The sequence  of  events  during cali-
bration, zero/span checks, and  electronic timing has been changed.

5.3,4  "C ali forn i a" Mod i fie at ion

     At monitoring sites  where  significant concentrations of "heavy"  hydro-
carbons exist, a problem  may arise  with the  Bendix  analyzer  which causes  the
TOC reading to be too low.  The  cause of  the low TOC reading is  that  during
the TOC measurement, the  methane and  heavy hydrocarbons may  separate  due  to
the chromatographic action of the surfaces of  the  lines and  components with
which the sample comes in contact.  This  causes the normal TOC peak on the
FID to separate into 2 distinct  peaks,  one for methane  and  light compounds
and the other for the heavier compounds.  The  height of these separate peaks
is less than the height of a single combined peak.  Since the analyzer
detects only peak height  and not peak area,  the resulting TOC reading is
reduced by the separation.

     This TOC peak separation may be  worse with some analyzers than with
others.  A test for the effect  is to  run  a chromatogram (see section  5.7.2)
using a mixture of methane and  a heavier  compound  such  as ra-xylene.  If
separation is occurring,  it will show up  as  a  separation  of  the  TOC peak
into separate peaks in the chromatogram.

     A modification to the analyzer to  reduce  the  effect  of  this TOC  separa-
tion has been developed by Bendix for certain  California  monitoring
agencies.  The modification may  become  standard for new model 8202  analy-
zers, and Bendix will likely offer  a  kit  to modify  existing  instruments.
The modification consists of a  new  chromatographic  column,  a new capillary,
and some plumbing changes.  It  also includes a change  in  the operation of
the analyzer such that the TOC  reading  is obtained  first  and the CH^
reading last.  This means that the  valve  and gate  timing  must be changed.
For analyzers incorporating this modification, appropriate changes  must be
made to the instructions  in section 5.7.2.   Consult the Bendix Corporation
for further details on this modification.
                                      60

-------
5.4  INSTALLATION

     The analyzer's instruction manual contains a discussion on  the  subject
under the same heading.  Additional comments follow.

5.4.1  General Comments

 1.  Before supplying power to the analyzer, ensure that all electronic
     cards are properly seated by attempting to push the cards snugly  into
     the connectors .

 2.  Minimize the distance between the analyzer and the sample manifold.
     Use clean, 3 mm (1/8 in) diameter inert tubing to make the  sample  line.
     The sample line length should not exceed 3 m (10 ft).

 3.  Leak-checks should be made on the hydrogen flow system of old and  new
     analyzers.  See the instruction manual for details or use the procedure
     in section 2.4.2 of the TAD.
          CAUTION:  ONLY EXPERIENCED, QUALIFIED OPERATORS SHOULD
          PERFORM THE LEAK CHECK.  AVOID FINGER CONTACT WITH THE
          FITTING THREADS DUE TO POSSIBLE TRANSFER OF OILS THAT WILL
          TEMPORARILY INCREASE THE BASELINE RESPONSE.
 4.   Hydrogen gas, burner air, and carrier air should not contain TOC con-
     taminant concentrations that can not be "bucked-out" by the auto zero.
     (Perform a level 1 zero and span check after replacing any of these
     support gases.)

 5.   If zero air is used to dilute span gases, it must not contain more than
     0.1 ppmC TOC concentration.

 6.   Connect a recorder to the NMOC output located in the rear panel of the
     analyzer.  If desired, recorders can also be connected to the analy-
     zer's TOC and CH4 outputs.  A temporary, multiple speed recorder
     capable of operating at ~16 in per hour is necessary for recording
     chromatograms from the analyzer's front panel jack.

 7.   Do not connect calibration gas to the CALIBRATE port.

 8.   The needle valves located inside the analyzer (NDLV 1) and oven (NDLV 2
     and 3) can be accessed through small holes drilled through the analyzer
     cover and oven.  Knobs on the valves can be removed, leaving the stem
     to be adjusted by a long screwdriver.  This modification enables flow
     rate adjustments without removing the analyzer cover and oven top,
     which eliminates heat loss and subsequent flow rate instabilitly during
     adjustments.
                                      61

-------
"'.4.2  Sample Pump Installation

     The instruction manual  specifies  installing  the sample pump outlet to
t.he SAMPLE port on the rear  of the  analyzer.   In  this  configuration, the
iiample passes through the  pump before  analysis  by the  analyzer.   However, a
very simple modification allows  the pump  to  draw  the sample through the
isample loop in the analyzer  such  that  the  sample  does  not  pass  through the
pump first.  The modification consists  of  connecting the  inlet  of the sample
pump to the CALIBRATE port of the  analyzer and  turning the front panel
selector switch to the calibrate  position.   This  switch must  be  left in the
CALIBRATE position during  normal  operation.   Sample  air will  thus be drawn
Into the SAMPLE VENT port  and out  the  CALIBRATE port to the pump.  In this
mode, the flow is adjusted with  the sample needle valve (006718-2), as the
back pressure regulator on the rear panel  will have  no effect.

J.5  OPERATION

5.5.1  Start-Up

     The analyzer's instruction manual  covers  this discussion under "Opera-
:ion".  Additional comments  follow.

 1.  After supplying power and support  gases  to the  analyzer, allow the
     analyzer to operate for several hours to  enable temperatures and flow
     rates to stabilize; then measure  flow rates.  Use good laboratory pro-
     cedures in measuring  flow rates.   (See  section  3.6 in this  TAD for
     guidance during flow measurements.)

 2.  The model 8201 instruction manual  specifies  normal settings or ranges
     for most of the operational  parameters.   If  a Manufacturer's Final Data
     Sheet (Bendix's Factory Test  Data  Sheet)  is  not  available with the
     analyzer, set the parameters  as specified  in  the  instruction manual,  or
     to previously used values.

5.5.2  Periodic Calibrations and  Zero/Span Checks

     An initial multipoint calibration  is  recommended  to verify  linearity  in
the calibration curve.  Periodic  level  1 zero  and  span checks are thereafter
necessary to update the calibration  curve.   The frequency  of  level 1 zero
and span checks and subsequent multipoint  calibrations should be based on
the analyzer's performance.  Performance can  be monitored through the use  of
a control chart which documents responses  to  standards and ultimately
enables calculation of variation  in  response  to standards  (drift).  (See
section 2.0.9 of reference 10 for  details  on  calibrations  and level 1 and
level 2 zero and span checks.  See  section H of reference  14  for details  on
control charts.  See section 5.6 of this TAD for  calibration  procedures.)
                                      62

-------
5.5.3  Operational  Checks

     The  analyzer's operational  parameters  should  be checked at  least once a
week, or  more  frequently  if  necessary.   Parameters that  should be routinely
checked and recorded  on Table  11 are  pressure  gauge readings,  analyzer
responses, and other  routine checks that  may be  necessary.   (Routine
monitoring may not  require that  unadjusted  responses to  standards be
recorded; however,  special cases may  warrent such  data.)   If any of these
parameters shows substantial change from the last  check,  recheck the
analyzer  calibration  with a  level 1 zero  and span  check  and make the
necessary adjustments.

5.5.4  Routine Chromatograms

     A chromatogram (Bendix  uses the  term "spectrum")  should be  evaluated
during start-up and on a weekly  basis thereafter to determine  if analyzer
parameters have shifted outside  the limits  of  proper operation.   Critical
parameters, such as gate timing,  FID  sensitivity,  and  flow  rates can be
investigated through  careful evaluation  of  the chromatogram.  See section
5.7.2 for procedures  to obtain a chromatogram, methods to evaluate the
chromatogram, and adjustments  to correct  improper  analyzer  operation.

5.6  CALIBRATION

     This section details specific procedures  for  multipoint calibrations
and zero and span checks of  the  Bendix 8201 and 8202.  The  analyzer's
instruction manual may contain deviations from these procedures, but the
step-by-step procedures that follow are  currently  recommended  by the EPA.

5.6.1  Multipoint Calibration

 1.  Disconnect the analyzer's sample line  from the ambient sample manifold
     and reconnect  it to the calibration  system's  manifold.  Cap the port on
     the ambient manifold, and on  any open  ports on the  calibration system's
     manifold.

 2.  Place the analyzer's range  selector  switch in the desired concentration
     range.  (Calibrate the  analyzer on  the range  normally  used  during
     ambient sampling.)  Ensure  that  the  TIMER and VALVE  toggle  switches are
     in the AUTO position.

 3.  Supply an atmosphere of methane-in-air to the calibration manifold at  a
     flow rate that is 20 to 50% greater  than  the  analyzer's sample flow
     demand.  DO NOT  PRESSURIZE  THE ANALYZER'S SAMPLE  INLET.

 4.  Run a chromatogram (at  least once a  week) and  evaluate it  for proper
     analyzer operation (gate  timing, flow  rates,  etc; see  section 5.7.2 for
                                        63

-------
TABLE 11. FORM FOR RECORDING ROUTINE OPERATIONAL DATA FROM THE
BENDIX 8201 OR 8202 NMOC ANALYZER
jl
]
i
1
s"
•
1
I
I
1
i
s
i
i
i
i
i
t
**
!
si
i
1
5
I
|
5*
2
ll\
I
4
1
!
1
i
I
5'
I
i
5
1
1
5
§
|
S*
8
1







































































































































































































































































.


















































































































































































































































































































































































































































































































































































































64

-------
procedures).
calibration.
Allow the analyzer to stabilize before continuing
Supply  an  atmosphere  of  the  zero  standard  to  the sample manifold at a
flow rate  20  to  50% greater  than  the  analyzer's  sample flow demand.  DO
NOT PRESSURIZE THE ANALYZER'S  SAMPLE  INLET.   The test  atmosphere must
contain an  ambient level  of  oxygen  and  must not  contain more than 0.1
ppmC TOC.   (See  section  3.2  for details.)
  NOTE:  THE ANALYZER'S AUTO  ZERO  WILL  CORRECT FOR MINOR
  CONTAMINATION  IN  THE HYDROGEN, BURNER AIR,  AND CARRIER AIR,
  BUT WILL NOT CORRECT FOR  ZERO GAS  CONTAMINATION.
  DISCREPANCY BETWEEN THE AUTO-ZERO  LEVEL  AND THE ZERO-AIR
  RESPONSE MAY INDICATE CONTAMINATED ZERO  AIR.
Adjust the TOC zero  pot, 0*4  zero  pot,  and  the  NMOC zero pot to
obtain the desired baseline.   (A 5%  full  scale  offset  is recommended to
observe negative drift.)  The  THC, CH4,  NMOC  sequence  must  be
followed if subsequent  adjustments are  necessary.   Ensure that  readings
from all channels are equal before recording  them  on Tables 5 and 11.

Supply an atmosphere of the methane  standard  to the sample  manifold at
a flow rate 20 to 50% greater  than the  analyzer's  sample flow demand,
DO NOT PRESSURIZE THE ANALYZER'S SAMPLE  INLET.   Methane concentration
should be between 70 and 90%  full  scale.

Adjust the front panel CH4 and TOC span  pots  to provide the analyzer
response calculated  as  follows:
    Response  «
                  Sample concentration
                          „   ,           ....
                          X url + zero offset
                                                      (9)
where

Response = Response of  the  recording device  measuring  the  analyzer
           output  in recording device  units .

Sample Concentration *  Concentration of  the  calibration  standard
                        delivered  to the  analyzer  in  ppmC.

URL * The upper range limit of the analyzer  in  ppmC.

url * The upper range limit of the recording  device  in recording device
      units .
                                   65

-------
     Zero offset = The amount the recording  device  response is set above
                   the zero baseline while the  analyzer  is  measuring the
                   zero calibration standard  (in  recording  device units).

     NOTE:  The response, url, and zero offset  must  be  in identical units.
     For example, if a strip chart recorder  is  the  recording device, units
     will be in percentages of chart and  the  url  may equal  100% chart.  If a
     digital volt meter (DVM) is the recording  device, units may be in
     millivolts or volts and the url may  be,  for  example, 1000 mV. If a Data
     Acquisition System (DAS) is the recording  device, units may be in ppmC
     and the url will equal 10 ppmC.   If  the  front  panel  meter on the
     analyzer is the recording device, units  will be in  ppmC and the url may
     be 10 ppmC.

     Record the stable responses on Tables 5  and  11,

     If either span pot is adjusted, recheck  the  zero and then the methane
     responses (Steps 5 through 9).  Ensure  that  all channels' responses to
     the zero standard are within +_ 1.0%  of  zero, that the  CH4 and THC
     responses to the methane standard are within _+  1.0%  of the actual con-
     centration, and that the NMOC response  to  the methane  standard is
     within + 1.0% of the NMOC response to the  zero  standard.
10.  Determine the CH4 and TOC channels'  responses  to  two  or more addi-
     tional concentrations of the methane  standard  that  are  approximately
     equally spaced over the analyzer  range.   (The  NMOC  channel's response
     should remain equal to the earlier  zero  standard  response.)   Record the
     CH4 and TOC channels' responses (from the recording device)  on Table
     11.  Using a calculator, perform  a  least  squares  linear regression of
     the CH4 channel responses (to methane and zero standards)  and the
     corresponding calibration concentrations.   The calibration
     concentrations should be in units of  ppmC and  should  be entered into
     the calculator as the independant variable X.   The  CH4  channel's
     responses should be in units of the  recording  device  and should be
     entered as the dependent variable Y.   A  correlation coefficient (r) of
     0.9996 or better verifies the CH4 response as  linear.  (If the
     response is not linear, plot the  data and determine if  an  error has
     been made in data entry or in determination of a  calibration concen-
     tration.)  Obtain the slope and intercept of the  regression  and record
     the equation in the following form:
     CH, Response
CH. Slope X Methane Concentration + CH,  Intercept
                                                                         (10)
                                        66
                                     U.S. EPA Headquarters Library
                                           Me;- code 3201
                                     1200 Pern?*. v--n,?. Avenue NW
                                        Wasningzon DC 20460

-------
     where
     CH4 Response
Analyzer's CH4 channel reading  in  recording device
units (see note following  step  8).
     CH4 Slope » Regression  slope  in  recording  device  units  per  ppmC.

     Methane Concentration * Calibrated methane  concentration  in ppmC.

     CH^ Intercept « Regression  intercept  in  recording device  units.

     Post the CH4 channel's  calibration curve equation on  the  analyzer's
     recording device and also on  Tables 5 and  11.  Also,  using  the  above
     equation and the TOG responses,  determine  the TOC calibration curve  and
     post it accordingly.

11.  Supply an atmosphere of the propane standard to the sample  manifold  at
     a flow rate 20 to 502 greater than the analyzer's  sample  flow demand.
     DO NOT PRESSURIZE THE ANALYZER'S SAMPLE  INLET.  Propane concentration
     should be between 70 and 90%  full scale.

12.  Adjust only the NMOC span pot to provide the analyzer response  cal-
     culated from equation in 9  step  8.  If the  NMOC span  pot  is adjusted,
     recheck the NMOC response to  the zero standard and adjust  the zero pot
     if necessary.  Recheck the  NMOC  response to the propane standard.
     Record the stable responses on Tables 5  and 11.

13.  Determine the NMOC channel's  response to two or more  additional concen-
     trations of the propane standard that are  spaced  approximately  equally
     over the analyzer range.  The CH4 channel's response  should remain
     equal to the earlier response to zero standard.   Record the stable
     responses (from the recording device) on Table 5.  Using  a  calculator,
     perform a least squares linear regression  of the  NMOC channel's
     response (to propane and zero standards) and the  corresponding  propane
     calibration concentrations.   The calibration concentrations should be
     in units of ppmC and should be entered into the calculator  as the
     independent variable X.  The  channel's response should  be  in units of
     the recording device and should  be entered  as the dependent variable Y.
     A correlation coefficient (r) of 0.9996  or  better  verifies  that the
     response is linear,  (if the  response is not linear,  plot  the data and
     determine if an error has been made in data entry or  in determination
     of calibration concentration.)  Obtain the  slope  and  intercept  of  the
     regression and record the equation in the  following form:
     NMOC Response = NMOC Slope X NMOC Concentration +•  NMOC  Intercept    (11)
                                        67

-------
     where
     NMOC Response
Analyzer's NMOC channel's  reading  in  recording device
units (see note following  step  8).
     NMOC Slope = Regression slope  in  recording  device  units  per ppmC.

     NMOC Concentration a Calibrated NMOC  concentration in ppmC.

     NMOC Intercept » Regression  intercept  in  recording device  units.

     Post the NMOC channel's multipoint  calibration  curve equation on  the
     analyzer's recording device  and on  Tables 5 and 11.

14.  Record all operational parameter  readings.   These  readings will now
     become the reference readings  until another multipoint  calibration is
     performed.  Be certain that  the station  logbook,  strip  charts, etc.,
     are properly annotated, dated, and  initialed.

15.  If level 2 zero and  span  checks are to be used, carry out  such a check
     now to establish reference  zero and span  readings  for subsequent  com-
     parison.  (See section 2.0.9.1.3  of reference  10  for definitions  and
     additional information for  level  1  and 2  zero  and  span  checks.)

5.6.2  Level 1 Zero and Span Check

     The following level  1 zero/span procedure (2-point calibration) is
similar to a multipoint calibration except  that  only the  zero air standard
and one upscale standard  are used.  These  responses  are then  used to cal-
culate the new calibration curves,  which are used to correct  subsequent
monitoring data until another  level 1  zero/span  check  or  multipoint calibra-
tion is performed.

 1.  Follow the procedures in  section  5.6.1, modified  as  follows.

     (a)  Record responses on  Tables 6 and 11.

     (b)  After step 5, determine the  zero drift in  the CH4,  TOG, and  NMOC
          channels (see section  3.8).

          If drift exceeds about  _+  3%, perform step  6  and make  the adjust-
          ments.  If drift is  less  than  about  _+_ 3%,  skip  step 6.

     (c)  After step 7, determine the  span drift in  the CH4  and TOG
          channels (see Equation  4  in  section  3.8).   If drift exceeds  about
          + 7%, make the  adjustments in  steps  8  and  9.
                                         68

-------
     (d)  Omit step  11, but determine the  slope  and  intercept  of  the  CH4
          and TOG channel's level  1 calibration  curve.

          Post the level  1 calibration  curves  on Tables  6  and  11  and  on the
          analyzer's  recording device.

     (e)  After step  II,  determine the  span  drift  in the NMOC  channel (see
          Equation 4  in section  3.8).   If  drift  exceeds  about  + 7%, make  the
          adjustment  in step  12.

     (f)  Omit step  13, but determine the  slope  and  intercept  of  the  NMOC
          channel's  level  1 calibration curve.   Post  the level 1  calibration
          curve on Tables  6 and  11 and  on  the  analyzer's recording device.

5.7  TROUBLESHOOTING

     The manufacturer's instruction manual covers  this  subject under  the
same heading.  Additional  comments follow.

5.7.1  Preliminary Investigations

     Operational problems, such  as slow response to  standards, inability  to
balance the TOG and CH4 channels'  responses,  long  stabilization periods,
frequent flame-outs,  or low sensitivity indicate the potential need  for
in-depth troubleshooting.  Prior to undertaking  in-depth troubleshooting
procedures, a preliminary  investigation is recommended  to  rule out varia-
tions in operational  parameters.

 1.  Ensure that all  pressure gauges on the  analyzer are set  to the  specifi-
     cations stated on the Manufacturer's  Final  Data Sheet.   (If  a manu-
     facturer's data  sheet is not  available,  use the instruction  manual's
     settings.)  If  the analyzer has been  operating  properly  for  several
     months, ensure  that  pressure  gauges are  set to  the  readings  recorded
     during the most  recent multipoint  calibration when  analyzer  operation
     was satisfactory.

 2.  Inspect the hydrogen  source and hydrogen  filters (located outside the
     analyzer) for proper  operation.

 3.  Inspect the sample filter and replace if  dirty.  Replace  the sample
     line leading- to  the  analyzer  if particulate buildup is noticeable.

 4.  Run a chromatogram as explained in section  5.7.2.   If the chromatogram
     has changed since the last multipoint calibration,  measure all  flow
     rates in the analyzer and adjust to the  proper  specifications.   Signi-
     ficant changes  in flow may  indicate a leak  or an obstruction.  If flow
     rates are within specifications, the  gate timing may  need adjustment.
                                        69

-------
5.7.2  Chromat og r am Inve s tig at io n

5.7.2.1  Procedures for Obtaining a Chromatogram—
       NOTE:  IF YOUR ANALYZER INCORPORATES THE "CALIFORNIA"
       MODIFICATION, SEE SECTION 5.3.4.	
 1.   Allow the analyzer to sample an atmosphere containing a tnethane-in-air
     standard.  Ensure that the atmosphere contains ambient levels  of
     oxygen and that it is delivered at atmospheric pressure to  the  SAMPLE
     inlet of the analyzer (section 3.5).

 2.   Place the METER selector switch on the front panel to the ELECTROMETER
     position.

 3.   Connect a strip chart recorder to the analyzer's  front panel.  Select a
     fast chart speed that will allow easy identification of time on the
     strip chart.

 4.   Allow the analyzer to operate for about 10 minutes.  Identify  the
     repetitive measurement cycles which last 200 seconds.  Each cycle
     should contain an automatic zero, CH^ peak, automatic zero, and TOC
     peak.  The strip chart recording of the cycle is  a chromatogram.   (See
     Figure 7 for a chromatogram depicting proper analyzer operation.)

 5.   Place the METER selector switch to the TIMER POSITION.

 6.   Advance the strip chart to a point where the pen  can rest on a timing
     line of the chart.  Stop the chart advance motor.

 7.   Observe the FRONT PANEL METER.  Immediately after the meter indicates
     completion of a 200-second cycle (the needle will pass full scale  then
     drop to zero), simultaneously place the METER selector to ELECTROMETER
     and engage the strip chart advance motor.

 8.   Observe the CH^, TOC, and zero function lights.   When the ZERO  light
     illuminates, flick the chart recorder pen to mark the chart.  When the
     light goes off, flick the pen again.

 9.   When the CH4 light illuminates, flick the pen, and repeat when  the
     light goes off.

10.   Perform similar actions when the auto zero and THC lights illuminate
     and go off.
                                        70

-------
11.  Remove Che chart, measure the number of  seconds  from  the  start  of the
     measurement cycle to each flick, and record  the  times  on  the  chart.

12.  For increased accuracy, repeat  steps 6 through 11  several times and
     average the ON/OFF times for each gate.  Also for  increased  accuracy,
     the same procedure can be repeated with  a  stopwatch  instead  of  a strip
     chart.  With the METER in the TIMER position, begin  the watch when the
     cycle begins and call out times  to an assistant  when  the  function
     lights illuminate and go off.   A digital watch with  split time
     functions would be very helpful  in performing this procedure.

5.7.2.2  Chromatogram Evaluation—

     Chromatograms depicting proper  and improper  analyzer  operation  are
shown in Figures 7 and 8.  The Chromatogram depicting proper operation shows
auto zero gates ON and OFF just prior to the  Cfy  and  TOC  peaks, the  peak
gates ON just prior to elution of the peak, and the peak  gates OFF just
after the response returns to the baseline.

     Gate timing is critical because  when a peak  gate is  ON, the  signal from
the FID is collected until the gate  goes off.   This signal  is  referenced  to
the auto zero signal and is eventually directed out the back panel to the
recorder.  If gate ON/OFF activity occurs during  peak emergence,  as  in
Figure 8, some of the peak may be missed.  This results in  an  incorrect
analyzer response for the component  measured.   A  procedure  to  correct
improper timing is discussed in 5.7.2.3, below.

     Another characteristic of proper operation is a  stable baseline
response.  The auto zero response prior to the  CH^ peak must be equal to
the auto zero response prior to the  TOC peak.   In Figure  7, the responses
during auto zeros remain the same, whereas in Figure  8, the responses during
auto zeros are different.  Variable  baselines can be  caused by differences
in carrier flow rates when passing through the  analytical  column and when
bypassing this column.  The variable  carrier  flow rate  will alter  the
hydrogen-to-carrier ratio, and thus  alter the sensitivity  of the  FID.  To
prevent a change in sensitivity, the  carrier  flow rate  bypassing  the column
during the TOC measurement must be balanced to  the carrier  flow rate passing
through the column during the CH4 measurement.  A procedure to correct
improper carrier flow rates is discussed in 5.7.2.3,  below.

5.7.2.3  Adjustments to Correct Improper Baseline Shift on  Gate Timing—

     If evaluation of a Chromatogram  indicates  improper carrier flow rates
(discussed in 5.7.2.2) perform the following  steps:

 1.  Place the front panel VALVE 1 and 2 switches in  the  OFF position and
     activate the AUTO H2 SHUT OFF.
                                        71

-------
              End m*Mur«m«it cyd« 9 200 MC
            - TOC «•«<>«« 147 we
Auto ztro off
9113MC	,
•TOC gat* one 119 Me
 Auto zero on 9 106 sec
                                                                      ' TOC Peak
           *. CH4 gitt off © 46 MC
Auto zero off
                                                                      ,CH4P«ak
             ' CH4 git* on 9 25 MC
              Auto zero on 9 16 MC

              Begin meaiurament cycle


           Figure 7.  Chromatogram from a  Bendix 8202 analyzer
                      showing proper analyzer operation.
                                                                       •TOC fmk
                                               TOC 9*U on ifttr ittit of p«k
Improper
bcMlinc
riirft
                           CH4 gM* off bwfor* end of p«k
                                                          CH4Ptik
          Figure 8.  Chromatogram from a Bendix 8202 analyzer showing
                     improper gate timing and baseline shift.
                                         72

-------
 2.  Connect a flow measuring device to the FID VENT on  the  back panel  and
     measure the flow rate.  Record this as the column flow  rate.   (The
     manufacturer may call this the "total column  flow rate".)

 3.  Adjust the AIR CARRIER pressure regulator to  obtain the column flow
     rate recorded on the Manufacturer's Final Data Sheet.

 4.  Connect a rotameter or mass  flow meter to the FID VENT  (if  not already
     connected) and record the flow rate reading.

 5.  Place the VALVE 2 switch in  the ON position and adjust  the  restrictor
     on Valve 2 (located inside the analyzer oven  off port 4)  to obtain the
     same flow rate as in step 4.  The carrier flow bypassing  the  column
     should now be equal to the carrier flow through the column.

 6.  Activate the AUTO IGNITION switch, place both VALVE 1 and 2 switches  in
     the AUTO position, and obtain a chromatogram.

 7.  If auto zero responses are not equal, adjust  the restrictor on Valve  2
     until auto zero responses are equal.

 8.  Activate the AUTO H2 SHUT OFF and place VALVE 1 in  the  OFF  position
     and VALVE 2 in the ON position.  Measure the  FID VENT flow  rate.   If
     this column bypass flow rate is not within +  1% of  the  column flow rate
     obtained in step 3, troubleshoot the  flow systems for leaks,
     obstructions, and column contamination.

     If evaluation of a chromatogram indicated improper  gate timing
(discussed in 5.7.2.2), perform the following steps:

 1.  Operators of the 8201 model  should refer to section 5.4.4 TIMING
     TROUBLESHOOTING in their instruction manual and Figure  7-16 COMPARATOR
     CARD SCHEMATIC DIAGRAM.  Operators of the 8202 model should refer  to
     section 5.5 CARD ADJUSTMENTS in their instruction manual  and  Figures
     5-7, 5-8 and 5-9.  Actions recommended in the instruction manual may be
     necessary if the following steps are  unsuccessful in correcting
     improper gate timing.

 2.  Convert the ON/OFF times (obtained in the TAD section 5.7.2.1) for each
     improperly timed gate from seconds to timer-dial settings.   The
     timer-dial (part #R-60) is a 10-turn, 200-second potentiometer that
     facilitates checking or setting the analyzer's gates.   The  timer-dial
     is mounted inside the 8201 analyzer beside the oven and is  labeled
     TIMER POSITION.  On the 8202 model, the timer-dial  is mounted on the
     front panel and is labeled MANUAL CYCLE POSITION TIMER.  For  this
     discussion, the R-60 dial will be referred to in both models  as the
     timer-dial.
                                        73

-------
In Figure  8,  the improperly timed  City  gate goes off at X  seconds.  To
convert  seconds to timer-dial units, use the following equation:
       Y  a or
       *       X 1000 timer-dial  units
      200 sec
y timer-dial  units
(12)
Verify  that  the City gate indicator  goes  off at y timer-dial  units.
Place Valves 1 and 2 in the OFF  position and the TIMER toggle  switch in
MANUAL.   Increasing the timer-dial  to  past y units should cause  the
City gate  indicator to go off.

The correct  gate OFF time  for the City  peak should be several seconds
after the  peak returns to  the baseline.   Estimate the number of  seconds
that should  be added to X  to produce the correct OFF time.   Convert
this corrected time from seconds to timer-dial units and rotate  the
timer-dial  to the corrected units.  The  City gate indicator should be
off.

Locate  Comparator Card #1  (component card J-10) inside the analyzer and
rotate  pot R-76 until the  City gate  indicator just comes on.  Check
gate operation by rotating the time dial through y units.

Reset the  analyzer controls to obtain  a  chromatogram and verify  that
the City off  gate is properly adjusted.

After the  operator is confident  performing timing adjustments, the
verifying  chromatogram need not  be  obtained until all timing adjust-
ments are  completed.  For  example,  to  correct  the TOC ON gate  problem
in Figure  8,  the Auto Zero ON and OFF  gates may need to be adjusted to
allow sufficient time for TOC gate  adjustment.  Follow the sequence of
events  discussed in steps  2 through 5, with modifications to fit the
specific  timing adjustment.  Table  12  lists the specific potentiometers
that will  alter the different functions.
      TABLE 12. COMPARATOR* CARD 1 (COMPONENT CARD J-10)
            FUNCTIONS AND POTENTIOMETER CONTROLS
              Function
          CH4 gate OFF
          TOC gate ON
          TOC gate OFF
          CH4 Auto Zero ON
          CH4 Auto Zero OFF
          TOC Auto Zero ON
          TOC Auto Zero OFF
  Potentiometer Number

        R-68
        R.76
        H-54
        R-64
        fl-40
        R-52
        R-29
        R-38
    'Comparator card 2 (component card J-11) is identical to card 1 (J-10) but control* valve 1 and 2
     ON/OFF timing and the actual timer cycle. Sea illustrations of these cardi in the instruction
     manual for details.
                                    74

-------
5,7.2.4   Chromatograms Showing Analyzer Malfunctions—
     The following section identifies specific chromatogram patterns, shown
in Figure 9, that indicate analyzer malfunctions.  The  information  is
reproduced from the Bendix 8202 Reactive Hydrocarbon Analyzer Operation  and
Service Manual.  Permission for reproduction was granted by the  Bendix
Corporation, Environmental and Process Instruments Division, Lewisburg,  West
Virginia  24901.

 A.  SYMPTOM:  Constant basline drift in one direction, then starts  in other
               direction.
     PROBABLE CAUSE

     Column has stopped bleeding and has
     started to clean.

     Dirty carrier or dryer.  Dirty dryer
     on H2 and/or air on FID.

     Dirty FID cell.
REMEDIAL ACTION

Allow time for column to
complete its cleanup.

Change carrier bottle and/or
dryers.

Clean FID cell.
 B.  SYMPTOM:  Blip in baseline that varies with flow rate.  The higher the
               flow, the larger the blip.
     PROBABLE CAUSE

     Dust in the FID cell.

     Condensate in cell vent.

     Varying pressure on FID vent
     SYMPTOM:  No peaks.

     PROBABLE CAUSE

     Electrometer off.


     No carrier gas flow.


     No flame.
REMEDIAL ACTION

Clean the cell.

Clean the vent.

Check for partial blockage.
Also, room pressure may be
changing due to doors or
windows being opened.
REMEDIAL ACTION

Check jet potential voltage.
Turn power ON.

Check carrier gas supply and
carrier shut-off switch.

Check H2 and air flows and
light flame.
                                        75

-------
                            B
               SYMPTOM
Constant baseline drift in
one direction, then startJ
in othar direction.
             SYMPTOM

Blip in baseline that varies
with flow rate. The higher
the flow the larger the blip.
           SYMPTOM


      No peeks
                                                          t_
        SYMPTOM
Little or very poor
sensitivity (normal
elution time).
                                                                              H
          1    SYMPTOM
Poor sensitivity (longer
elution time).
             SYMPTOM


      Baseline irregular
        I   SYMPTOM

Baseline appears as a
sinewave type trace.
         SYMPTOM

Stepping baseline
               SYMPTOM



   Negative peaks
            [SYMPTOM
           SYMPTOM
   Irregular or noisy
   baseline
Distorted peaks
         SYMPTOM

Constant baseline
drift in one
direction
                 Figure 9.  Chromatograms showing analyzer  malfunctions.
                                                   76

-------
    Defective recorder.


    Leak at column connections.

    Recorder not connected correctly.
    No voltage applied to detector
    cell on ionization detector.
    Cable disconnected from FID.

    Column temperature too low to elute
    peaks.
Refer to recorder manual for
troubleshooting procedure.

Tighten connections.

Refer to schematic of
recorder hookup.  Be sure no
jumpers are grounding signal

Check electrometer voltage
output.  Refer to
electrometer schematic.

Connect the cable.

Raise column temperature to
proper level.
    SYMPTOM:  Little or very poor sensitivity (elution time normal).

    PROBABLE CAUSE                            REMEDIAL ACTION
    Incorrect sample size.

    Electrometer OFF (FID).

    No sample flow.
Clean sample chamber.

Turn electrometer ON.

Check pump operation and
STANDARD SAMPLE solenoid
(SOLV 3) valve operation.
E.  SYMPTOM:  Poor sensitivity (longer elution time).
    PROBABLE CAUSE

    Analyzer temperature too low.


    Carrier flow too low.

    Leak downstream of injection valve.

    SYMPTOM:  Baseline irregular.

    PROBABLE CAUSE

    Column bleeding.
REMEDIAL ACTION

Check and reset to correct
temperature.

Increase flow rate.

Locate and repair leak.



REMEDIAL ACTION

Replace defective column
with a properly conditioned
column.
                                       77

-------
G.
    FID cell contaminated.

    Bad electrical ground.


    Carrier gas regulation poor,
Defective electrometer.

Unstable oven temperature.



H2 and/or air regulation poor.



SYMPTOM:  Baseline appears  as a sinewave

PROBABLE CAUSE

Poor temperature control.
    Hydrogen flow too high.

    Dryers dirty.

    Leak at jet.

    Low tank pressure on carrier gas—
    bottle pressure should be over  10
    psig.

    Defective carrier gas  flow  regulator.

    Defective H2 and/or air  regulators.
 Clean the FID cell.

 Check the analyzer for
 proper grounds.

 Check carrier gas flow
 control and regulator.  Be
 sure upstream pressure is
 high enough to operate
 regulator.  Pressure should
 be at least 10 psig.

 Check electrometer.

 Check to see if temperature
 controller is operating
 properly.

 Check flow rates to be sure
 of proper flows and regula-
 tion.

type trace.

 REMEDIAL ACTION

 Check the temperature control
 for proper operation.  Check
 seal around door for leaks.
 Be sure there is no obstruc-
 tion to air flow in the oven.

 Lower the hydrogen flow.

 Replace the dryers.

 Locate and repair leak.

 Replace tank.
                                          Replace  the  flow  regulator.

                                          Clean  or  replace  regulator.
                                       78

-------
H.  SYMPTOM:  Stepping baseline.

    PROBABLE CAUSE

    Defective recorder.

I.  SYMPTOM:  Negative peaks.

    PROBABLE CAUSE

    Signal reverse circuit energized.


    Recorder leads reversed.


J.  SYMPTOM:  Irregular or noisy baseline

    PROBABLE CAUSE

    Loose leads.


    Bad ground connections.



    Carrier gas leak.

    Contaminated columns .


    Contaminated carrier gas.



    Dirty switch or switches.
REMEDIAL ACTION
Refer to the recorder manual.
REMEDIAL ACTION

Refer to Control Module
manual (if applicable).

Reverse the recorder leads on
the rear hookup panel.
REMEDIAL ACTION

Tighten all leads from power
supply to cell.

Tighten all ground
connections to ensure earth
potential.

Locate and repair leak.

Replace or recondition
columns.

Replace carrier gas bottle
and replace dryers and
filters .

Locate the dirty switch and
clean with suitable solvent.
    Recorder slidewire dirty.

    H2 and/or air flow rates too
    high or too low.

    Detector insulators dirty.

    Contaminated FID cell.
Clean with suitable solvent

Adjust flow rates.


Clean insulators.

Clean FID cell.

-------
     Detector coaxial cable defective.

     Loose signal leads.


 K.   SYMPTOM:  Distorted peaks.

     PROBABLE CAUSE

     Flame burning above jet into ignitor
     or collector electrode area.
Replace the coaxial cable.

Check all connections to be
sure they are tight.
REMEDIAL ACTION

Hydrogen and air flows too
high.  Reduce flows.
 L.   SYMPTOM:   Constant baseline drift.

     PROBABLE  CAUSE

     Column bleed.


    •Dirty carrier.


     Dirty FID cell.

     Defective electrometer.


     Detector  temperature increasing.


     Leak in column tubing.

     Column temperature too high for
     the column.


5.7.3  Sample  and Back-Flush Valve Leak Check
REMEDIAL ACTION

Replace or recondition
columns.

Replace carrier or dryer, and
filters.

Clean FID cell.

Repair or replace electro-
meter .

Allow time for detector tem-
perature to stabilize.

Locate and repair leak.

Lower the column temperature
or choose another type of
column.
 1.  A pressure test apparatus described in section 2.4.2 can be used.  The
     analyzer's instruction manual also addresses leak-checks.

 2.  Select the CALIBRATE position on the front panel toggle switch.

 3.  Cap the SAMPLE VENT.

 4.  Pressurize the CALIBRATE inlet port.  With VALVE 1 in the ON position,
     pressure decay indicates a possible leak in the sample valve at 0-rings
                                        30

-------
    to ports 3, 5, 4, and 6.  With VALVE  1  in  the OFF  position,  pressure
    decay indicates a possible  leak  in the  sample valve  at  0-rings  to  ports
    3, 2, I, and 6.

5.  Remove the cap and pressure  source.

6.  Cap the FID vent and the back-flush vent.

7.  Pressurize the AIR CARRIER  inlet  port.  When VALVE 2 is in the  ON
    position (and VALVE 1 is known to be  leak  proof),  pressure decay
    indicates a possible leak in the  back-flush valve  at 0-rings to ports
    3, 5, 4, and 6.  With VALVE  2 in  the  OFF position,  pressure  decay
    indicates a possible leak in the  back-flush valve  at 0-rings to ports
    3, 2, 1, and 6.

-------
                                  SECTION  6

                 AUGMENTED PROCEDURES FOR THE  BECKMAN  6800
                AIR QUALITY CHROMATOGRAPH (CH4,  TOC, NMOC)
  NOTE: READING SECTIONS  1, 2, AND  3  BEFORE  SECTION  6  IS  STRONGLY
  RECOMMENDED.  INFORMATION WHICH  IS  CRITICAL  TO  THE UNDERSTANDING
  AND SUBSEQUENT ACCURATE OPERATION OF  THIS  ANALYZER IS DETAILED IN
  THOSE SECTIONS AND WILL NOT BE REPEATED HERE.
     The manufacturer has prepared  a  relatively  detailed  instruction  manual
which can guide users during  installation  and  operation of  the  model  6800
analyzer.  The thoroughness of the  manual  will therefore  allow  this  section
to be brief.  As stated  in  the introduction,  the purpose  of this  TAD  is  to
supplement the analyzer's instruction manual by  augmenting  critical
installation and operation  procedures.

6.1  PRINCIPLES OF OPERATION

     The analyzer's  instruction manual  contains  a discussion of this  subject
under "Instrument Theory".  A brief summary of the CH4, TOC,  and  NMOC
component analysis follows.

     A measured amount of sample  is injected directly  into  the  FID to
measure TOC concentrations.   Following  the TOC measurement,  a second,
separate sample is injected into  a  stripper column to  separate  CH4 and  CO
from C02 and heavy organics.  A second  column  then separates CH4  from CO
prior to injection into  the FID,  where  the CH4 concentration is measured.
The CH4 concentration can then be externally  subtracted from the  TOC  con-
centration to obtain a measure of the NMOC concentration.   (Modifications
are available to enable  internal  subtraction of  CH4 from  TOC to obtain  a
direct NMOC output.)

6.2  SYSTEM DESCRIPTION

     A detailed description of the  analyzer is given in the analzer's
instruction manual.  This TAD will  therefore not  address  location of  com-
ponents, but will emphasize the uses  and precautions of those which  are
critical.
                                      82

-------
6.2.1  Flame  lonization  Detector

     The  analyzer  contains  one  FID.   Response from the CH^ and TOC
channels  must be balanced  to  a  methane-in-air standard.  If zero air is
necessary to  dilute  span gas  to the  required  calibration concentrations, the
CH4 and TOC channels must  first be balanced  to the zero air standard, and
then to the methane-in-air  standard.

6.2.2  Hydrogen Pressure Gauge

     The  flow of hydrogen  fuel  to the FID is  controlled by a pressure
regulator, gauge,  and  restrictor.  Increasing the  pressure reading on the
hydrogen  gauge will  increase  the  flow of hydrogen  to the FID and affect the
sensitivity of the burner.  Ensure that  the hydrogen reading does not vary
from day  to day.   If the reading does vary, measure the flow rate and take
appropriate corrective actions.
    CAUTION:  ENSURE THAT  ALL  TUBING  CARRYING HYDROGEN GAS IS LEAK
    PROOF.  SEE  INSTRUCTION  MANUAL, PAGE  12.	      	
6.2.3  Flame-out Indicator

     In the event of  a  flame-out,  the  flame-out  indicator will illuminate.
If the status of the  flame  is  in doubt,  observe  analyzer  response first to
zero air, and then to a sample known to  contain  methane.   If the TOC
channel's responses are  the  same to both samples,  the  flame is out.

6.2.4  Slider Valves

     The slider valve is  a  stainless steel  valve which contains a Teflon®
slider.  The ON/OFF positions  of the slider dictate  where sample is
directed.  Friction of  the  slider  against  the  stainless  steel causes wear of
both the slider and the  steel,  ultimately  resulting  in a  leak.  Frequent
maintenance of these valves  is strongly  recommended. The  slider should be
replaced every 6 months,  or  sooner  if  a  leak develops  or  if the slider
begins to open and close  sluggishly.   When  the slider  is  replaced, the
stainless steel valve must be  lapped to  ensure a smooth  seat for the slider.
During reassembly, the  valve connectors  must be  tightened uniformly and only
to the point where no sample leakage occurs.  Use  of a vacuum applied to  a
valve port is recommended to aid in determining  when the  valve is leak-
proof and  thus sufficiently secure.

6.2.5  Columns

     To prevent column  contamination,  all  columns  should  be capped whenever
they are removed from the analyzer or  whenever the analyzer is powered-down
                                       83

-------
for an extended period  of  time.   Failure  to  do so will allow the molecular
sieve to absorb C(>2.  Subsequent  operation results in CO2 outgassing
which will show up  as a noisy, erratic  baseline response.

6.3  MODIFICATIONS

     The model 6800  analyzer has  been  slightly modified over the years.  The
principles of operation and detection  have not changed, but the following
modifications have  been adopted.
6.3.1  Stripper Column
                                                                          CO
     To improve  stability  of  the molecular  sieve column (which separates
from CH^), the stripper  column  that  preceeds  the molecular sieve column
has been modified.  The  original packing material,  Poropak,  appeared to
break down and allow  passage  of volatile  fragments  to  the sieve column,
resulting in deactivation.  To  correct  this,  a  combination column containing
silica gel and an  inert  Teflon® support/partition liquid  has replaced the
Poropak column.

6.3.2  Hydrogen  Pressure Regulator

     To eliminate  diffusion of  atmospheric  water vapor into  the hydrogen
carrier stream,  the original  permeable  diaphragm located  in  the hydrogen
carrier gas regulator has  been  replaced with  a  stainless  steel diaphram.
Also, the body of  the regulator has  been changed to brass.

6.3.3  Scrubber

     The scrubber.  (683526) upstream  of  the  molecular sieve column is not
necessary and has  been removed.  Subsequent resizing of the  capillary to
adjust pressures and  flows must be performed  when the  scrubber is removed.

6.3.4  Sample Pump

     In the normal configuration of  the sample  pump, the  sample passes
through the pump before  analysis by  the analyzer.  However,  a very simple
modification allows the  pump  to draw the sample through the  sample loops in
the analyzer such  that the sample does  not  pass through the  pump first.  The
modification consists of disconnecting  the  pump from its  normal connection
to the 3-way solenoid valve,  and reconnecting it in series with the line
going from Valve B to the  SAMPLE VENT.  The pump inlet is connected to port
6 of Valve B, and  the outlet  is connected to  the SAMPLE VENT.  The bypass is
removed from the pump, and a  needle  valve,  installed between the pump inlet
and Valve B, is  used  to  control the  flow rate.   The sample inlet for the
analyzer then becomes the  port  on the 3-way solenoid valve to which the pump
was originally connected.
                                        84

-------
6.3.5  Pressure Regulator

     The service  air,  air  carrier,  hydrogen  fuel,  and  burner  air  pressure
regulators have been replaced by  improved  regulators  (865774).

6.3.6  Catalytic  Converter  Board

     The catalytic converter board  installed  in  analyzers  before  February
1973 contains a 4.7-ohm resistor  (Rl)  rated  at 2 W.   This  should  be  replaced
with a 5-ohm 5-W  resistor.

6.3.7  Oxidation  Catalyst

     The oxidation catalyst, which  scrubs  air to be used  for  combustion and
sample carrier, has been replaced.  The original nickel oxide catalyst  has
been replaced with palladium-on-alumina.

6.4  INSTALLATION

     The analyzer's instruction manual contains  a  discussion  of this  subject
under the same heading.  Additional comments  and a specific start-up  proce-
dure follow.

6.4.1  General Comments

 1.  Before supplying  power to the  analyzer,  ensure that  all  electronic com-
     ponent cards are  properly seated  by attempting to push the cards  snugly
     into the connectors.

 2.  Minimize the distance  between  the analyzer  and the sample  manifold.
     Use clean 3 mm (1/8-in) diameter  inert tubing as  the  sample  line.   The
     sample line  length should not  exceed  3 m (10  ft).

 3.  Leak-checks  should be made on  the hydrogen, air carrier,  and  sample
     flow systems of old and new  analyzers.   See sections  3.1.4 through
     3.1.6 in the instruction manual for details.  A  leak-check apparatus  is
     discussed in section 2.4.2 of  this TAD.
    CAUTION:  ONLY EXPERIENCED, QUALIFIED OPERATORS  SHOULD
    PERFORM THE LEAK-CHECK.  AVOID FINGER CONTACT WITH THE
    FITTING THREADS DUE TO POSSIBLE TRANSFER OF OILS THAT WILL
    TEMPORARILY INCREASE THE BASELINE RESPONSE.
 4.  Hydrogen gas, burner air, and carrier air should  not contain  TOC  con-
     taminant concentrations that cannot be "bucked-out" by  the  auto zero.
     Perform a level 1 zero and span check after replacing any of  these
     support gases.
                                        85

-------
 5.  Zero air used to dilute span gases must  trot  contain  more  than 0.1  pptnC
     TOG concentration.

 6.  Connect recorders to the Clfy and TOC outputs  (and  NMOC  output if
     available) located on the rear panel connector.  Also,  a  temporary
     recorder for the front panel jack (to record  chromatograms)  is useful.

 7.  Calibration gases should enter the analyzer  through  the SAMPLE inlet
     port.  Do not connect calibration gas to  the  CALIBRATE  port.


-------
5.
(d)  Turn on Valve B switch and observe slider movement  and  illumi-
     nation of switch indicator.

(e)  Move Manual/Auto switch to Auto, Valves A and  B  should  deactivate
     and indicators should go out.

(f)  Return valve switches A and B  to off  position  and Manual/Auto
     switch to Manual.

Leak-check the air and sample flow  systems  as described  in  section
3.1.5 and 3.1.6 of the analyzer's instruction manual.
     WARNING:  CHECK THE CONDITION OF THE WATER  DRAIN  TUBING  BE-
     TWEEN THE FID AND OVEN EXTERIOR FOR HARDENING AND BRITTLENESS.
     LEAKS IN THIS DRAIN MAY LEAD TO AN EXPLOSION BY ALLOWING A
     BUILDUP OF HYDROGEN IN THE OVEN DURING REPEATED IGNITION
     ATTEMPTS.
 6.  Establish correct gas flows.

     (a)  Turn off H2 fuel at the pressure  control.

     (b)  Connect a flow meter to the hydrogen  carrier  and  fuel  line  after
          disconnecting from FID.   (See Figure  5-9  in the instruction
          manual.)

     (c)  Adjust H2 Carrier pressure control to obtain  25 cm^/rain  and
          record gauge read ing.  A  typical  reading  is between  12 and  20
          psig.

     (d)  Adjust H2 Fuel pressure control to obtain  a total  flow (carrier
          plus fuel! of 30 to 31 cm^/min, and record gauge  reading.   A
          typical gauge reading  is  between  5 and  11  psig.

     (e)  Disconnect flow meter  and reconnect H2  carrier and  fuel  line.

     (f)  Connect a flow meter to the air carrier line  after disconnecting
          from the FID.

     (g)  Adjust air carrier pressure control to  obtain a carrier
          flow of 33 to 35 cm-Vmin  and record the gauge reading.
          A typical gauge reading is 18-26  psig.  Disconnect the
          flow meter and reconnect  the tubing.
                                        87

-------
 7.  Ignite detector burner.

     (a)  Place Flame-Out Override  switch  to ON.

     (b)  Place the Manual Range Select  switch  to _1£ and  the  Monitor Select
          switch to Ampl. Out.  Note the meter  response while proceeding
          with the following  steps.

     (c)  Reduce the Burner Air pressure gauge  setting by approximately 50%
          and momentarily C3  to 5 seconds)  raise the  Flame Ignition switch.
          The meter response  should drive  upscale momentarily upon  ignition
          of the flame, then  settle to some upscale  level,  indicating  flame
          ignition.

     (d)  Repeat the above step until  flame ignites,  and  then reset  burner
          air to recorded setting.

     (e)  Turn off Flame-Out  Override  switch.

 8.  Allow system to stabilize 12 to 24 hours prior  to performing any
     calibrations.

6.5.2  Periodic CaLibrations  and Zero/Span Checks

     An initial multipoint calibration is  recommended to  verify  linearity  in
the calibration curve.  Periodic level 1 zero and span checks are thereafter
necessary to update the calibration curve.  The  frequency of  level  1 zero
and span checks and subsequent multipoint  calibrations should be based  on
the analyzer's performance. Performance can be monitored  through the use of
a control chart which documents responses  to standards, and ultimately
enables calculation of variation in response to  standards (drift).   (See
section 2.0.9 of reference 10 for details  on calibrations and level  I and
level 2 zero and span checks.  See  section H of  reference 14  for details on
control charts.  See section  6.6 of this TAD for specific calibration
procedures.)

6.5.3  Operational Checks

     The analyzer's operational parameters should be  checked  at  least once  a
week, or more frequently if necessary.  Parameters that should be routinely
checked and recorded on Table 13 are pressure gauge  readings, analyzer
responses, and other routine  checks that may be  necessary.  (Routine moni-
toring may not require that unadjusted responses to  standards be recorded;
however, special cases may warrant  such data.)   If any of these  parameters
shows substantial change from the last check, or if  drift out of the speci-
fied range has occurred, recheck the analyzer calibration with a level  1
zero and span check and make  necessary corrections.
                                        88

-------
uj
I
S
O
oc
<
i
-i
<
O OC
P£i
2>
UJ 7
0. <
°?
UJ
p
°5
2£" O
«1
ii
cc|
§1
tr Ul
= ffl
tr
u.
K
LL
«
UJ
_l
<








i
j



i
i




\
1


1
i




{
1





•
•


£

.*
<

£

i
*

;-

i


|
3

j


i

i

!
l
i
g
s
1
!
!
•|
|
g
§
1
\
S
5"
i
1
5*
§
|
if
§
|
5'
I































































































































































































































































































































































































































































































































































































































































































































89

-------
5.5.4  Routine Chromatograms

     A chromatogrant should be evaluated during  start-up,  and  on  a weekly
oasis thereafter, to determine if analyzer parameters have  shifted outside
the limits of proper operation.  Critical parameters, such  as  gate timing,
FID sensitivity, and flow rates can be investigated through a  careful
evaluation of the chrotnatogram.  The procedure  for obtaining  a chromatogram
and setting gate timing follows.

 1.  Momentarily close RESET switch to place  system in  reset  condition.

 2.  Set OPERATIONAL STATUS switches as follows,

     (a)  Move Manual/Auto Switch to Manual.

     (b)  Move Calibrate/Operate Switch to Operate.

     (c)  Move Manual ATT/Auto ATT Switch to  Manual ATT.

 3.  Turn off front panel switches designated Valve A,  Valve B.

 4.  Allow the analyzer to sample an atmosphere containing  a methane-in-air
     standard at about 8 ppmC plus about 10 ppm or higher CO.  Ensure  that
     the atmosphere contains ambient levels of  oxygen and that it  is
     delivered at atmospheric pressure to the SAMPLE inlet  of  the  analyzer
     (section 3.5).

 5.  Move Monitor Select Switch to AmpI. Out.

 6.  Momentarily close Start switch and turn  Manual Range Select  switch
     to 1.

 7.  Turn on Auto Zero switch until meter reading drops to  zero,  then  turn
     switch off.  If zero is not obtained, refer to section 6.5  in the
     analyzer's instruction manual.

 8.  Turn Manual Range Select switch to 100 and Chromatogram Attenuator
     switch to 1 for a span gas between 8 and 10 ppm.

 9.  Turn on chromatograph recorder chart drive.  Select  a  chart  speed that
     permits adequate peak separation.  The minimum chart speed  should be
     40 on/hr (16 in/hr).

10.  One or more manual test Chromatograms, using the recorder connected  to
     the front panel jack, must now be run to determine TOC peak  emergence
     time.  To inject the sample, turn on the Valve B switch for  25 seconds.
     The recorder should register a flat-topped peak.   If the  peak is off-
     scale, turn the chromatogram attenuator  switch to  the next higher
                                        90

-------
     setting, wait at least 30 seconds, then repeat  actuation  of Valve  B
     switch.  Continue until desired onscale peak is obtained.  When  onscale
     peak height is obtained, note and record settings used on Manual Range
     Select and Chromatogram Attenuator switches.  Also note the times  of
     the start and end of the TOG peak.
       NOTE:  ALLOW AT LEAST 30 SECONDS BETWEEN SUCCESSIVE  SAMPLE
       INJECTIONS WITH VALVE B TO ENSURE COMPLETE PURGING OF  THE
       SAMPLE LOOP.

11.   A series of two-component manual test chromatograms must  now  be  run  to
     determine the methane and carbon monoxide peak  emergence  times  and
     Valve A actuation time.  Both CH£ and CO components must  be  run,  even
     through some users may not require the CO measurement output.   It has
     been determined that column degradation can cause overlapping of the CO
     and CH^ components.  Therefore, the CO may act  as an  interferent  for
     the CH4 measurements.  It is not sufficient to  remove,  turn off,  or
     deprogram the CO component card to eliminate the interferent  response.

     (a)  In the first trial run, when the recorder  pen  is on  a major  time
          division on the chart, turn on Valve A.  After 75  seconds,  turn
          off Valve A switch and wait for the methane and  carbon monoxide
          peaks to register on the recorder chart.   If peaks are offscale,
          turn the chromatograptn attenuator switch to the  next higher
          setting and repeat.  Note the height of the methane  peak.

     (b)  In the next trial, turn on Valve A for 65  seconds  and note  the
          height of the methane peak on the resulting chromatogram.

     (c)  Continue trial runs, each time using a shorter energization  period
          (10-second decrement) for Valve A, until a chromatogran  is
          obtained showing decreased amplitude for the methane peak.   (Be
          sure to take into consideration any changes in the setting  on the
          Crhomatogram Attenuator and Manual Range Select  Switches).  A
          decreased height of methane peak indicates premature de-energi-
          zation of Valve A, with a resultant loss of a portion of the
          sample.  Therefore, the correct energization period  for  Valve A is
          about 6 to 10 seconds longer than that used in the last  trial run.
          For most instruments, duration of the required energization  period
          is between 40 and 75 seconds.
                                        91

-------
     (d)  When the desired chromatogram  is  obtained,  note and record the
          settings used on Manual Range  Select  and  Chromatogram Attenuator
          Switches.  Also note  the  start  and  end  times  of the CH^ peak.

12.   Using the test data obtained in  the  preceding  steps, program the time-
     related functions for the  Valve  Board,  auto  zero board,  and individual
     component boards following the procedures  given  in Section 3.0 of the
     analyzer's instruction manual.
       NOTE:  PRIOR TO REMOVING OR  INSTALLING  THE  BOARDS  FOR
       PROGRAMMING, TURN OFF THE ELECTRICAL  POWER  SWITCH  LOCATED
       ON THE FRONT PANEL.  IF THE  SWITCH  IS NOT PROMPTLY RESTORED
       TO THE ON POSITION, THE FAILSAFE  FEATURES WILL  TURN THE
       ANALYZER OFF.
13.  Timing Circuit Board for Valves A  and  B.

     (a)  Program che Valve B Function,  TOC sample  injection "on" at  001
          second, and "off" at 025  seconds.

     (b)  Program the Valve A, Sample-Injection/Stripper  Function "on" at
          030 seconds.  The required "off"  time  is  determined  from the ener-
          gization time determined  in step  11  plus  30  seconds.   For example,
          if the duration of the required energization period  for Valve A
          has been determined to be 65  seconds,  the corresponding "off" time
          would be 095 seconds, with respect to  time "zero".

14.  Component board assigned to TOC.
     (a) Program the TOC component  gate  window  to  include  the  full  TOC peak
         on the TOC test chromatogram.   The  typical  timing is  "on"  at 006
         seconds; "off" at 020 seconds.  Leave  the  window  "on" as  long as
         possible to allow detection of  heavy organics.
     (b) Set range selector switch  S3  for  the  10  ppm range  position.
                                        92
U.S. EPA Headquarters Library
      Mai! code 3201
1200 Pennsyvanie Avenue NW
   Washington DC  20460

-------
15.  Component board  assigned  to methane.

     (a)  Program the "on"  and  "off  times  to  include  the  full  methane peak
          on the methane manual  test  chromatogram (step 11).   Remember to
          add 30 seconds to  the  times determined  in  step 11.

     (b)  Set Range Select  switch  S3  for  the  10 ppm  range.

16.  Auto Zero Board.

     (a)  Program the first  auto-zero "A"  to  occur 12  to 15  seconds prior to
          the emergence of  the methane  peak on a  stable portion of the base-
          line.  Duration of the auto zero  period need only  be 4 or 5
          seconds.  Check the manual  test  chromatogram (step 11) to verify
          flat baseline in  the  region,  to  be  used for  auto zeroing.

     (b)  Program the second  auto-zero  "B"  near the  end of the  cycle
          (292-298 seconds)  where  the baseline is flat (preparatory to the
          TOC peak on the next cycle).  The auto-zero  function must not be
          programmed  to coincide with an  elution  of  a  peak.

17.  Run additional chromatograms  to verify the various time  settings, if
     necessary.

6.6  CALIBRATION

     This section details specific  procedures  for multipoint  calibrations
and zero and span checks of  the  Beckman 6800.  The analyzer's instruction
manual may contain deviations  from  these  procedures, but the  step-by-step
procedures that follow are  currently  recommended  by  the EPA.

6.6.1  Mult ipoint Galibrat ion

 1.  Disconnect the analyzer's  sample line  from the  ambient  sample manifold
     and reconnect it to the calibration  system's manifold.   Cap the port on
     the ambient manifold,  and  on  any open  ports  on  the calibration system's
     manifold.

 2.  Place the analyzer's range  selector  switch in the desired concentration
     range.  (Calibrate the  analyzer on the range normally used during
     ambient sampling.)

 3.  Supply an atmosphere of methane  standard  to  the calibration manifold at
     a flow rate 20 to 50%  greater  than the analyzer's sample flow demand.
     DO NOT PRESSURIZE THE  ANALYZER'S SAMPLE  INLET.
                                        93

-------
    Run a chroraatogram  (at  least  once  a  week)  and  evaluate it  for proper
    analyzer operation  (gate timing,  flow  rates, etc.;  see section 6.5.3
    for procedures).  Investigate  unusual  behavior and  consult trouble-
    shooting for procedures to correct the  problems.  Allow the analyzer to
    stablize before continuing calibration.

    Supply an atmosphere of the zero standard  to the  sample manifold  at a
    flow rate 20 to 50% greater than the analyzer's  sample flow demand.
    DO NOT PRESSURIZE THE ANALYZER'S SAMPLE  INLET.   The  test atmosphere
    must contain an ambient level  of oxygen  and must  not  contain more than
    0.1 ppmC TOC.  (See section 3.2 for details.)
      NOTE:  THE ANALYZER'S AUTO ZERO WILL  CORRECT  FOR  MINOR
      CONTAMINATION IN THE HYDROGEN AND  BURNER AIR  BUT  WILL  NOT
      CORRECT FOR ZERO GAS CONTAMINATION.   DISCREPANCY  BETWEEN
      THE AUTO-ZERO LEVEL AND THE ZERO AIR  RESPONSE  MAY INDICATE
      CONTAMINATED ZERO AIR.
6.  Set Operational status switches  as  follows.

    (a)  Move Single/Continuous  to Cone inuous.
    (b)  Move Manual/Auto to "Auto.
    (c)  Move CaUJbirat: e/Operate  to Operate.
    (d)  Move ATT/Auto ATT to Auto ATT.

    Momentarily actuate the start switch  and  allow  the  analyzer  to  complete
    3 complete cycles in the automatic mode,  or  to  cycle  until a stable
    response is obtained on the  memory  recorder  outputs  for  each component.
    Record the stable responses  on Tables  5 and  13.   If  the  responses  are
    not within + 1% of each other and zero, ensure  that  the  TOC  and  Cfy
    channel ele~tronics are balanced before investigating  contamination  in
    the cylinder, lines, and regulators.   (Be certain all  regulators con-
    tain metallic diaphragms and are LOX  cleaned.)

    If the analyzer does not contain an NMOC  subtraction  card, proceed to
    step 8.

-------
7.  Zero the NMOC response by the  following procedure:

    (a)  Set operational status switches as follows.

         Move Manual/Auto to Manual.
         Move Calibrate/Operate to Calibrate.

    (b)  On the TOC component board, place the OFF/MANUAL/AUTO switch on
         MANUAL.  Make certain this  switch on the  CH^  component  board is
         in either the OFF or AUTO position.

    (c)  Place the MONITOR SELECT  switch at the memory position  assigned to
         TOC (usually MEMORY 1 position).  Adjust  CALIBRATE control  for a
         reading of exactly zero on  the front panel meter.   Leave the
         CALIBRATE control at this setting until the zero procedure  is
         completed.

    (d)  On the NMOC subtraction board, remove the pin from the  position
         marked SAME RANGE or LOWER  RANGE, thus removing  the CH^ memory
         input from the subtraction  board.  Place  the  AUTO/MANUAL switch
         on MANUAL.  Insert the pin  into the connection marked READOUT
         MEMORY 5 POSITION.  (With a five-component analysis system,  the
         pin may be used only during initial zeroing of the subtraction
         circuit.  The pin must be removed prior to programming  of the
         component board assigned  to the fifth component.)

    (e)  Place MONITOR SELECT switch at MEMORY 5 position.   Adjust
         potentiometer R14 on the  subtraction board for a zero reading on
         the front panel meter.

    (f)  Replace pin in the SAME RANGE or LOWER RANGE  position.

    (g)  Repeat step 6 and obtain  the stable NMOC  response  to the zero
         standard.  Record the response on Tables  5 and 13.

8.  Supply an atmosphere of the methane standard to the calibration mani-
    fold at a flow rate 20 to 50%  greater than the analyzer's sample  flow
    demand.  DO NOT PRESSURIZE THE ANALYZER'S SAMPLE INLET.    The methane
    concentration should be between  70 and 90% full scale.

9.  Set Operational status switches  as follows.

    (a)  Move Single/Continuous to Continuous .
    (b)  Move Manual/Auto to Auto.
    (c)  Move Calibrate/Operate to Operate.
    (d)  Move Manual ATT/Auto ATT  to Auto ATT.
                                       95

-------
LO.   Momentarily actuate the start switch and allow chromatograph  to  com-
     plete 3 complete cycles in the automatic mode, or  to  cycle  until  a
     stable response is obtained on the memory recorder outputs  for each
     component.  If the recorder output is off scale for a  given  component,
     turn the respective potentiometer (R12) two turns  counterclockwise to
     decrease memory gain and repeat until an onscale response is  obtained,
11.  Calibrate the
                       channel as follows.
     (a)   Momentarily actuate the reset switch.

     (b)   Set operational status switches as follows.

          Move Manual/Auto to Manual .
          Move Calibrate/Operate to Calibrate .
          Move Manual ATT/ Auto ATT to Auto ATT.

     (c)   Turn Manual range select to position 1 and the CH^ component
          board switch to the manual (center) position.

     (d)   Adjust the calibrate control to simulate the memory  recorder
          response of the CH^ component which was obtained with  the
          methane span gas.

     (e)   Adjust the CH4 component board potentiometer (R12) to  yield the
          correct memory recorder response as calculated in the  following
          equation:
                       Sample concentration  .,    .            ..,
          Response  *  	«-—===	  X  url  +  zero offset    ,
                             UKL                                       \
                                                                        13)
          where

          Response * Response of the recording device measuring  the analyzer
                     output in recording device units.

          Sample Concentration » Concentration of the calibration standard
                                 delivered to the analyzer in ppmC.

          URL = The upper range limit of the analyzer in ppmC.

          url = The upper range limit of the recording device in recording
                device units .

          Zero offset = The amount the recording device response is set
                        above the zero baseline while the analyzer is
                        measuring the zero calibration standard  (in
                        recording device units).
                                        96

-------
12.
     NOTE;  The response, url,  and  zero offset must  be  in  identical
     units.  For example, if a  strip chart  recorder  is  the  recording
     device, units will be  percentages of  chart  and  the url may equal
     1002 chart.  If a digital  voltmeter (DVM) is  the recording device,
     units may be millivolts or volts and  the url  may be,  for  example,
     1000 mV.  If a Data Acquisition System (DAS)  is the recording
     device, units may be ppmC  and  the url  will  equal 10 ppmC.   If the
     front panel meter on the analyzer is  the recording device, units
     will be ppmC and the url may be 10 ppmC.

(f)  Return the CH4 component board switch  to the  auto  position.

Calibrate the TOC component as  follows.

(a)  Momentarily actuate the reset  switch.

(b)  Set operational status switches as follows.

     Move Manual/Auto to Manual.
     (c)
     Move C al ibrate/Operate to Calibrate.
     Move Manual ATT/Auto "ATT to Auto ATT.

     Turn Manual range select to position _1_ and  the  TOC component
     board switch to the manual (center) position.
13.
14.
(d)  Adjust the calibrate control  to simulate  the memory recorder
     response of the TOC component which  was obtained  with the methane
     span gas.

(e)  Adjust the TOC component board potentiometer  (R12)  to yield the
     correct memory recorder response  as  calculated  in Equation 13.
     The CH4 and TOC responses to  the  methane  standard must  be within
     +1.0% of each other and the actual concentration  of the test
     atmosphere.

(f)  Return the TOC component board switch  to  the  auto position.

Return all switches to the normal  automated "run"  position.   Momentar-
ily actuate the start switch and allow the  chromatograph to  complete 2
automated cycles to verify that adjustments were correct.

Determine the Qfy and TOC channel's response to  two  or more  addition-
al concentrations of the methane standard,  spaced  approximately equally
over the analyzer range.  (The NMOC channel's  response should remain
equal to the earlier response for  zero standard.)  Record  the CH4 and
TOC channel's responses (from the  recording device)  on Table 5.  Using
a calculator, perform a least squares  linear regression of the
                                        97

-------
     channel responses (to methane and zero  standards)  and  the  corresponding
     calibration concentrations.  The calibration concentrations  should be
     in units of ppraC and entered into the calculator  as  the  independent
     variable X.  The CH4 channel's response  should  be  in units of the
     recording device and entered as the dependent variable Y.  A
     correlation coefficient (r) of 0.9996 or better verifies  the CH4
     response is linear.  (If the response is not linear,  plot  the data and
     determine if an error has been made in data entry  or in determination
     of calibration concentration.)  Obtain the slope  and intercept  of the
     regression and record the equation in the following  form:
15,
L6.
CH4 Response « CH4 Slope X  Methane  Concentration +  City Intercept  (14)

where

CH4 Response = Analyzer's City  channel  reading  in recording device
                 units  (see  note  following  step  11).

CH4 Slope » Regression  slope  in recording device units per ppmC.

Methane Concentration = Calibrated  methane  concentration in ppraC.

CH4 Intercept = Regression  intercept  in  recording device units.

Post the CH4 channel's  calibration  curve equation on  the analyzer's
recording device and also on  Tables 5  and 13.  Also,  using the above
equation and the TOC responses, determine the  TOC calibration curve,
and post accordingly.

Supply an atmosphere of the  propane standard  to  the calibration mani-
fold at a flow rate 20  to 50%  greater  than  the analyzer's  sample flow
demand.  DO NOT PRESSURIZE  THE ANALYZER'S SAMPLE INLET.   The propane
concentration should be between 70  and 90%  full  scale.

Set Operational status  switches as  follows.

(a)  Move Single/Continuous  to Continuous.
(b)  Move Manual/Auto to "Auto.
(c)  Move C alib rate/Ope r at e  to Operate.
(d)  Move Manual ATT/Auto ATT to-Auto  ATT.

Momentarily actuate the start  switch  and allow chromatograph to com-
plete 3 complete cycles in  the automatic mode, or to  cycle until a
stable response is obtained  on the  memory recorder  outputs for each
component.  The TOC response  should be about  30% less than the actual
propane concentration,  and  the City  response should  be equal to the
earlier zero response.  Record the  stable responses on Tables 5 and 13.
(If the analyzer contains an  NMOC subtraction  board,  record the stable
NMOC responses and disregard  the  TOC  response.)
                                        98

-------
17.  Determine  the TOG  channel's  response  (or  Che NMOC channel's response if
     the analyzer has the  NMOC  subtraction  board) to  two  or more additional
     concentrations  of  the propane  standard spaced approximately equally
     over the analyzer  range.   Record  the  stable  responses (from the
     recording  device)  on  Table 5,   Using  a calculator,  perform a least
     squares linear  regression  of the  channel's response  (to propane and
     zero standards) and the  corresponding  propane calibration concentra-
     tions.  The calibration  concentrations  should be in  units of ppraC and
     entered into the calculator  as  the  independent^ variable X and the
     channel's  response should  in units  of  the  recording  device and entered
     as the dependent variable  Y.  A correlation  coefficient (r) of 0.9996
     or better  verifies that  the  response  is linear.   (If the response is
     not linear, plot the  data  and determine if an error  has been made in
     data entry or in determination  of calibration concentration.)  Obtain
     the slope  and intercept  of the  regression  and record the equation in
     the following form:
     Channel's response
                     Channel's  slope  X Propane  concentration +
                        Channel's  intercept
                                                                         (15)
18.
where

Channel's response « Analyzer  channel's  reading  in recording device
                     units  (see note  following step 11).

Channel's slope • Regression slope  in recording  device units per ppmC.

Propane concentration  =  Calibrated  propane  concentration  in ppmC.

Channel's intercept *  Regression  intercept  in recording device units.

If the analyzer contains  an NMOC  subtraction board,  enter  the NMOC
responses in the regression and post  the NMOC channel's multipoint
calibration curve equation on  the analyzer's recording device and  on
Tables 5 and 13.  If the  analyzer does not  contain an  NMOC board,  enter
the TOC responses in the  regression and determine  the  TOC  calibration
curve to propane.  Use the  slope  of this curve and Equation 1 in
section 3.7 to calculate  NMOC  responses from future  TOC and CH^
responses.

Record all operational parameter  readings.  These  readings will now
become the reference readings  until another multipoint  calibration is
performed.  Be certain that the station logbook,  strip charts, etc.,
are properly annotated, dated, and  initialed.
                                        99

-------
19.  If level 2 zero and span checks  are  to be  used,  carry out  such a check
     now to establish reference zero  and  span readings  for subsequent
     comparison.  (See section 2.0.9.1.3  of reference  10  for definitions and
     additional information in level  1  and 2 zero  and  span checks.)

5.6.2  Level 1 Zero and Span Check

     The following level 1 zero/span  procedure  (2-point calibration)  is
similar to a multipoint calibration except that only  the  zero air  standard
and one upscale standard are used-.  These responses are needed  to  calculate
the new calibration curves, which should be used to correct  subsequent moni-
:oring data until another  level 1 zero/span check  or multipoint  calibration
is  performed.

 1.  Follow the procedures in section 6.6,1 modified  as follows.

     (a)  Record responses on Tables  6  and 13.

     (b)  After step 6, determine the zero drift in the CH^  and  TOC chan-
          nels (Equation 4, section 3.8).  If drift exceeds  about  + 3X,
          determine that the CH^ and TOC channel's electronics have not
          drifted (see the instruction  manual)  and evaluate  the  zero  air
          system for possible contamination.

          If the analyzer  contains an NMOC subtraction board, determine  the
          zero drift in the NMOC channel  after  step 6.  If drift exceeds
          +3%, perform step 7 and make  adjustments.   If drift is less than
          +3%, skip step 7.

     (c)  After step 10, determine the  span drift  in  the  Clfy and TOC  chan-
          nels (section 3.8).  If drift exceeds about  + 7%,  make the  adjust-
          ments in steps 11 and 12 and  obtain stable  responses,  as in
          step 13.

     (d)  Omit step 14, but determine the slope and intercept of the  CH^
          and TOC channel's level 1 calibration curve.

          Post the level 1 calibration  curve on Tables 6  and 13  and on the
          analyzer's recording device.

     (e)  After step 16, determine the  TOC channel's  response to propane and
          calculate the new TOC slope to  propane.  If  the  analyzer contains
          an NMOC board, determine the  NMOC response  to propane  and deter-
          mine the level 1 calibration  curve.

6.7  TROUBLESHOOTING

     The manufacturer's instruction manual covers  this subject under  the
same heading.  Additional  comments follow.
                                                      U.S. EPA Headquarters Library
                                                            Mai! code 3201

-------
6.7.1  Preliminary Investigations

     Operational problems such as slow response to  standards,  inability  to
balance channels' responses, long stabilization periods,  frequent
flame-outs, or low sensitivity indicate the potential need  for  in-depth
troubleshooting.  Prior to undertaking in-depth troubleshooting  procedures,
a preliminary investigation  is recommended to rule  out variations  in  opera-
tional parameters.

 1.  Check that all pressure gauges on the analyzer  are  set  to  the  specifi-
     cations stated on the Manufacturer's Final Data Sheet  (or  to  the speci-
     fications stated in section 6.5.1.6 if a manufacturer's  data  sheet  is
     not available).  If the analyzer has been operating  properly  for
     several months, ensure  that pressure gauges  are set  to  the  readings
     recorded during the most recent multipoint calibration  when analyzer
     operation was satisfactory.

 2.  Inspect the hydrogen source and hydrogen filters {located  outside of
     the analyzer) for proper operation.

 3.  Inspect the sample filter and replace if dirty.  Replace the  sample
     line leading to the analyzer if particulate  buildup  is  noticeable.

 4.  Run a chromatogram.  If the chromatogram has changed  since  the last
     multipoint calibration, measure all flow rates  in the  analyzer and
     adjust to the proper specifications.  Significant changes  in  flow may
     indicate a leak (slider valves may need replacement  every  6 months) or
     an obstruction.  Follow guidelines in the instruction manual.

 5.  If flow rates are within specifications, the gate timing may  need
     adjustment (see section 6.5.4).

 6.  If columns are found to be deactivated, as indicated by the loss of
     separation, replace them with new columns rather than  trying  to
     reactivate the used ones.
                                        101

-------
                                 SECTION  7

             COMMENTS ON THE  BYRON  233 THC/CIty/CO ANALYZER
     The Byron model 233 analyzer  was  not  evaluated  in  the 1979 NMOC analy-
zer study (9); thus, augmented  procedures  to  supplement  the analyzer's
instruction manual are not reported here.   Conversations  with  the
manufacturer have  produced comments on general  guidance  which  are reported
here in lieu of augmented procedures.   (NOTE:   The Byron  model 404 was
evaluated in the 1979 study.)

7.1  PRINCIPLE OF  OPERATION

     The analyzer  was designed  to  fulfill  the requirements of  the Federal
Reference Method for determination of  hydrocarbons corrected for methane
(Appendix E, Part  50, Title 40  of  the  Code of Federal Regulations).  A brief
summary of the Byron 233 operation follows.

     A measured amount of sample  is  injected directly  into the FID to
measure TOC concentration.  Following  the  TOC measurement, a second,
separate sample is injected into  a stripper column to separate CH^., CO,
and C02 from the heavy organics.   A second column then  separates CH4
from CO and CC>2 prior to injection into the flame ionization detector
where the CH^ concentration is  measured.   The City concentration can then
be externally subtracted from the  TOC  concentration  to  obtain  a measure of
the NMOC concentration.

7.2  GENERAL COMMENTS

 1.   Support air  supplied to the  analyzer is used as carrier  air, FID
      burner air,  and valve actuation.   This  air  must not contain more than
      0.1 ppraC TOC or more than 0.1 ppm CO + C02-

 2.  The hydrogen  supply must be  ultra-high purity or better (contains no
     more than 0.1 ppmC TOC).

 3.  The hydrogen  flow system can  be  leak-checked as  follows.

     (a)  Remove the hydrogen restrictor from the methanizer and cap the
          tubing.

     (b)  Increase the analyzer's  hydrogen regulator  to  the maximum
          position.
                                        102

-------
    (c)  Supply hydrogen to  the  analyzer's H2  inlet  at 30 psi.   Shut  off
       •  the hydrogen supply.

    (d)  Any decay in the analyzer's hydrogen  pressure gauge  reading  in
         30 minutes indicates a  leak.

    (e)  The leak must be located  and repaired.

4.  Prior to ignition of the hydrogen flame, increase the hydrogen  pres-
    sure about 3 to 10 psig  higher than  recommended  for  normal  operation,
    After the burner has been lighted for at least  10 minutes,  reduce
    hydrogen pressure to specifications.

5.  Whenever the support air is  changed, check the  span  response.

6.  Moisture will deactivate the molecular sieve  column.  The column
    should be reactivated every  6 months.
                                       103

-------
                                SECTION 8

                  GENERAL COMMENT ON THE MELOY HC  500-2C
                         FID HYDROCARBON ANALYZER
     The Meloy HC 500-2C was included in the  1979 hydrocarbon  analyzer
evaluation (9).  After the study was completed, the manufacturer  indicated
the analyzer's detector would be redesigned.  This major change in  the
analyzer reduces the usefulness of augmented  procedures  for  operating  the
ariginal analyzer; therefore, no procedures are reported here.

     The manufacturer has been requested to supply information on  the  modi-
fied analyzer; however, the  information was not available  in time  to be
included in this TAD.  Users of this analyzer should consult the manu-
facturer for recommendations.
                                        104

-------

-------
                            SECTION 9

                            REFERENCES
Air Quality Criteria for Ozone and Other Photochemical Oxidants,
Environmental Criteria and Assessment Office, U.S. Environmental
Protection Agency, Research Triangle Park, N.C.  27711, Publication No.
SPA-600/8-78-004, April 1978,

Demerjian, K. L., Kerr, J. A., and Calvert, J. G., The Mechanism of
Photochemical Smog Formation in Advances in Environmental Science and
Technology, Volume 4, p. 1-262, John Wiley and Sons, Inc., New York,
1974.

40 CFR Part 50, National Primary and Secondary Ambient Air Quality
Standards, FR 36 (84):8186, April 30, 1971.

Guidance for Collection of Ambient Non-Methane Organic Compound (NMOC)
Data, for Use in 1982 Ozone SIP Development, and Network Design and
Siting Criteria for the NMOC__and NOV Monitors, Office of Air Quality
Planning and Standards, the U.S. Environmental Protection Agency,
Research Triangle Park, N.C., Publication No.  EPA-450/4-80-011, June
1980.

Singh, H. B., Guidance for the Collection and Use _qf^_Ambient
Hydrocarbon Species Data in Development of Ozone Control Strategies,
Office of Air Quality Planning and Standards, U.S.  Environmental
Protection Agency, Research Triangle Park, N.C., Publication No.
EPA-450/4-80-008, April 1980.

Reckner, L. R., Survey of Users of the EPA Reference Method for
Measurement of Non-Methane Hydrocarbons in Ambient Air, U.S.
Environmental Protection Agency, Research Triangle Park, N.C.,
Publication No. EPA-650/4-75-008, December 1974.

Harrison, J. W., et al., Evaluation of the EPA Reference Method for
Measurement of Non-Methane Hydrocarbons, Environmental Monitoring
Systems Laboratory, U.S. Environmental Protection Agency, Research
Triangle Park, N.C., Publication No.  EPA-600/4-77-033, June 1977.

McElroy, F. F., Thompson, V. L., Hydrocarbon Measurement Discrepancies
Among Various Analyzers Using Flame-Ionization Detectors, Environmental
                                   105

-------
     Monitoring Support Laboratory, U.S. Environmental Protection Agency,
     Research Triangle Park, N. C., Publication No. EPA 600/4-75-010,
     September 1975.

 9.  Sexton, F. W., McElroy, F. F,, Michie, R. M.,  Jr., Thompson, V. L., A
     Comparative Evaluation of Seven Automatic Ambient Non-Methane Organic
     Compound Analyzers, Environmental Monitoring Systems Laboratory, U.S.
     Environmental Protection Agency, Research Triangle Park, N,C.   (Draft
     document).

10.  Quality Assurance Handbook for Air Pollution Measurement Systems,
     Volume II - Ambient Air Specific Methods, U.S. Environmental Protection
     Agency, Research Triangle Park, N.C., Publication No.  EPA-600/4-77-
     027a, Revision 1, July 1, 1979.

11t  Site Selection for the Monitoring of Photochemical Air Pollutants, U.S.
     Environmental Protection Agency, Publication No. EPA-450/3-78-013,
     April 1978.

12.  40 CFR Part 50, Appendix E, Addenda A, Suggested Performance Specifica-
     tions for Atmospheric Analyzers for Hydrocarbons Corrected for
     Methane.

13.  Eaton, W. C., Use of the Flame^Photometric Detector Method for
     Measurement of Sulfur Dioxide in Ambient Air - A Technical Assistance
     Document, Environmental Monitoring Systems Laboratory, U.S.
     Environmental Protection Agency, Research Triangle Park, N.C.,
     Publication No.  EPA-600/4-78-024, May 1978.

14.  Quality Assurance Handbook for Air Pollution Measurement Systems,
     Volume I - Principals, U.S. Environmental Protection Agency, Research
     Triangle Park,  N.C., EPA Publication No. 600/9-76-005, January 1976.

15.  Traceability Protocol for Establishing True Concentrations of Gases
     Used for Calibrations and Audits of Air Pollution Analyzers (Protocol
     No.  2 ), June 1978, Environmental Monitoring and Support Laboratory,
     U.S.  Environmental Protection Agency, Research Triangle Park, N.C.
     27711.

16.  NBS Special Publication 260,  NBS Standard Reference Materials Catalog,
     1979-1980 Edition, National Bureau of Standards, Washington, DC 20234.
                                       106

-------
                                   TECHNICAL REPORT DATA
                            tPlcase read fuunictions on the reverse before comnltting;
        SO.
                              2.
                                                            3. RECIPIENT'S ACCESSIOWNO.
  -iTi.S A\C SUBTITLE
TECHNICAL  ASSISTANCE DOCUMENT FOR  THE  CALIBRATION AND
OPERATION  OF AUTOMATED AMBIENT NON-METHANE ORGANIC
COMPDUND ANALYZERS
                                                            5. REPORT DATE
                                                             March,  1931
                                                            6. PERFORMING ORGANIZATION CODE
7 ALTHOAtSI
Frederick W.  Sexton, Raymond A. Michie,  Jr.  (RTI)
Frank  F.  McElroy, Vinson L. Thompson  (EPA)
                                                            8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental  Quality Assurance  Department
Research  Triangle Institute
Research  Triangle Park, North Carolina 27709
                                                            10. PROGRAM ELEMENT NO.
                                                             1A9F60E10C
                                                            11. CONTRACT/GRANT NO.
                                                             Contract Ho.  68-02-3222
                                                             Task Directive No.  7
12. SPONSORING AGENCY NAME AND ADDRESS
Quality Assurance Division
Environmental  Monitoring Systems Laboratory
Offi;e of Research  and Development
U.S. Environmental  Protection Agency
     yrin TriannTo Daylc  Mnr»fi-> f.arnMna  77711
                                                            13_T,YPE,OF REPORT AND PERIOD COVERED
                                                             Final
                                                            14. SPONSORING AGENCY CODE
IS. SUi'PLEMENTARVNOTiS

Technical  Assistance Document
16. ABSTRACT
      This  technical assistance document is intended to serve as  a  supplement to manu-
factjrers1  instruction manuals for  automated, ambient, non-methane organic compound
analyzers.   It addresses augmented  set-up, calibration, operation, and maintenance
procedures  that have been developed for the purpose of producing non-methane organic
compound data suitable for use in the Empirical Kinetic Modeling Approacn and otner
dispersion  models used to predict ozone concentrations.  The document initially
discusses  common aspects of these procedures in general terms  for  hydrocarbon analyzers
as  a  whole  and then specifically addresses their application to  analyzers manufactured
by  Tne Bendix Corporation, Mine Safety Appliances Company, and Beckman Instruments,
Inc., and  Meloy Laboratories, Inc.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.IDENTIFIERS/OPEN ENDED TERMS
                                                                         c.  COSATI Field/Group
Air  pollution
Measurement
Calibration
Air  pollution monitoring
                                               Non-methane organic
                                                compound analyzers
                                               NMOC analyzers
                                               FID analyzers
                                               Hydrocat»bon analyzers
                                               Technical assistance
                                                document
   6SA
'3. ii
             STATEMENT
 RELEASE TO PUBLIC
EPA Form 2220-1 (9-73}
                                              19. SECURITY CLASS iTha Report!

                                               UNCLASSIFIED	
21. NO. OP PAGES
       122
                                               20. SECURITY CLASS I This page >

                                               UNCLASSIFIED
                                                                          22. PRICE

-------

-------