&EPA
            United States
            Environmental Protection
            Agency
           Environmental Monitoring and Support
           Laboratory
           Research Triangle Park NC 27711
EPA-600 4-79-056
September 1979
            Research and Development
Transfer Standards
for  Calibration of
Air  Monitoring
Analyzers for Ozone

Technical Assistance
Document

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and  application of en-
vironmental technology.  Elimination  of traditional grouping  was  consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1    Environmental  Health Effects Research
      2   Environmental  Protection Technology
      3.   Ecological Research
      4.   Environmental  Monitoring
      5.   Socioeconomic Environmental Studies
      6.   Scientific  and Technical Assessment Reports (STAR)
      7.   Interagency Energy-Environment Research and Development
      8.   "Special" Reports
      9.   Miscellaneous Reports

This report has been assigned to the ENVIRONMENTAL MONITORING  series.
This series describes research conducted to develop new or  improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations.  It also includes
studies to determine the ambient concentrations of pollutants  in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
 This document is available to the public through the National Technical Informa-
 tion Service, Springfield, Virginia 22161.

-------
                  TRANSFER STANDARDS
FOR THE CALIBRATION OF AMBIENT AIR MONITORING ANALYZERS
                       FOR OZONE
             Technical Assistance Document
                          by

                   Frank F. McElroy
               Quality Assurance Branch
    Environmental Monitoring and Support Laboratory
     Research Triangle Park, North Carolina  27711
    ENVIRONENTAL MONITORING AND SUPPORT LABORATORY
          OFFICE OF RESEARCH AND DEVELOPMENT
         U.S. ENVIRONMENTAL PROTECTION AGENCY
     RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711

-------
                                  DISCLAIMER

     This document has been reviewed by the Environmental Monitoring and
Support Laboratory, U.S. Environmental Protection Agency, and approved for
publication.  Mention of trade names or commercial products does not consti-
tute endorsement or recommendation for use.
                                       ii

-------
                                   FOREWORD

     Measurement and monitoring research efforts are designed to anticipate
potential environmental problems, to support regulatory actions by developing
an in-depth understanding of the nature and processes that impact health and
the ecology, to provide innovative means of monitoring compliance with regula-
tions and to evaluate the effectiveness of health and environmental protection
efforts through the monitoring of long-term trends.   The Environmental Moni-
toring Systems Laboratory, Research Triangle Park, North Carolina has respon-
sibility 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 and Radiation, the Office of Toxic
Substances and the Office of Enforcement.

     This Technical Assistance Document defines, specifies, and formalizes the
certification of ozone transfer standards for calibrating ambient ozone
analyzers.  The procedures and guidance in this document provide monitoring
agencies additional flexibility and specific benefits in designing and imple-
menting an effective quality assurance program for their ambient ozone moni-
toring .
                                                Thomas R. Hauser, Ph.D.
                                                       Director
                                     Environmental Monitoring Systems Laboratory
                                       Research Triangle Park, North Carolina
                                     111

-------
                                    PREFACE

     Ultraviolet  (UV) photometry currently appears to be the most accurate
technique for assaying ozone calibration atmospheres in the sub-ppm concentra-
tion range to obtain primary ozone standards.  Accordingly, the U.S. Environ-
mental Protection Agency has adopted UV photometry as the prescribed procedure
for the calibration of reference methods to measure ozone in the atmosphere.

     To minimize the immediate impact of the new calibration procedure, as
well as to permit flexibility to capitalize on the potential advantages of
transfer standards, the new calibration procedure specifically allows the use
of transfer standards for calibrating ambient ozone monitors.  Such transfer
standards, however, must be suitably referenced to a UV primary ozone standard.

     The U.S. Environmental Protection Agency believes that the use of ozone
transfer standards can provide monitoring agencies with a number of worth-while
advantages.  However, their use entails a considerable degree of technical ex-
pertise in selecting, qualifying, certifying, using, and maintaining the trans-
fer standards.  This document offers users technical guidance in these areas.

     For the experienced user, this document will hopefully help to formalize
and standardize the qualification, certification, and use of transfer standards.
For the less experienced user, it should serve as a learning aid and an informa-
tion source.  The document is quite comprehensive and oriented more toward new
users, being primarily a reference document.  As it is designed for quick and
easy referral, there is some repetition.

     The accurate, long-term measurement of ozone concentrations in ambient air
is not an easy task.  Error processes are relentless; the infiltration and
magnitude of errors must constantly be held to a minimum to realize quality
                                       v

-------
data.  Hopefully, attention to the information and guidance given in this
document will help toward that end.

     I regret that the data represented in the figures are pure fiction.  I
would have greatly preferred to use figures generated from real data, but such
data were not available at the time.  When such data become available, a supple-
ment or revision will be considered to include them in the document.  Comments
or criticism from readers and users are welcome; any changes resulting from
such comments will also be included in any subsequent revision or supplement.
                                      vi

-------
                                   ABSTRACT

     On February 8, 1979  (Federal Register, 44:8221-8233), the U.S. Environ-
mental Protection Agency amended Appendix D of Title 40, Code of Federal Regu-
lations (CFR), Part 50, to prescribe a new calibration procedure for the cali-
bration of reference methods for measuring ozone in the atmosphere.  The new
procedure is based on the use of ultraviolet  (UV) photometry as a primary stan-
dard for ozone.  The new calibration procedure specifically allows the use of
transfer standards for the calibration of ambient ozone monitors, provided
such transfer standards are adequately referenced to a primary UV ozone stan-
dard.
                                             t

     This document is intended as a reference aid to help users select ozone
transfer standards and reference them to a primary UV standard.  It first
defines ozone transfer standards and then discusses their purpose and role in
calibrating ambient ozone analyzers.  The various advantages and disadvantages
of ozone transfer standards are pointed out to help users determine whether to
use a transfer standard or the UV procedure directly.  Several different types
of ozone transfer standards are described, including analytical instruments
(chemiluminescence and UV analyzers), manual analytical procedures (potassium
iodide and gas phase titration procedures), and ozone generation devices.

     The major part of the document is devoted to the procedures necessary to
establish the authority of ozone transfer standards:  qualification,  certifi-
cation, and periodic recertification.  Qualification consists of demonstrating
that a candidate transfer standard is sufficiently stable (repeatable) to be
useful as a transfer standard.  Repeatability is necessary over a range of
variables such as temperature, line voltage, barometric pressure, elapsed
time, operator adjustments, or other conditions, any of which may be encountered
during use of the transfer standard.  Tests and possible compensation techniques
                                      VI1

-------
for several such common variables are provided.  Detailed certification proce-
dures are also provided together with the quantitative specifications that the
transfer standard must meet to achieve certification.  Finally, the periodic
recertification procedure and recertification specifications necessary to
maintain continuous certification of the transfer standard are given.

     For convenience, the UV primary ozone standard procedure from 40 CFR Part
50 is reproduced in Appendix A.  Other appendices give more specific guidance
for the qualification and certification of several common and practical types
of transfer standards — the BAKl manual procedure, two gas phase titration
procedures, ozone generators, and ozone analyzers.
                                     viii

-------
                                   CONTENTS

Foreword	    iii
Preface 	      v
Abstract	     vi
Figures	     xi
Acknowledgments 	    xii

     1.  WHAT IS A TRANSFER STANDARD?	    1-1
            Definition	    1-1
            Purpose	    1-1
            How Transfer Standards Are Used	    1-2
     2.  WHY USE AN OZONE TRANSFER STANDARD?	    2-1
            Ozone Concentration Standards 	    2-1
            Advantages of Ozone Transfer Standards	    2-2
            Disadvantages of Ozone Transfer Standards 	    2-4
     3.  PRESCRIBED ULTRAVIOLET PROCEDURE FOR PRIMARY OZONE STANDARDS .    3-1
            Description	    3-1
     4.  TYPES OF TRANSFER STANDARDS FOR OZONE	    4-1
            Analytical Instruments	    4-1
            Manual Analytical Procedures	    4-5
            Generation Devices	    4-9
     5.  ESTABLISHING THE AUTHORITY OF OZONE TRANSFER STANDARDS ....    5-1
            Transfer Standard Specifications	    5-1
            Comparing Transfer Standards to an Ultraviolet Primary
              Ozone Standard	    5-1
            Qualification 	    5-6
            Qualification Tests 	    5-9
            Certification 	    5-20
            Re certification	    5-27
                                      IX

-------
                             CONTENTS (continued)
     6.   SPECIFICATIONS FOR OZONE TRANSFER STANDARDS	    6-1
            Definition	    6-1
            Apparatus	    6-1
            Documentation 	    6-1
            Qualification 	    6-2
            Certification 	    6-2
            Recertification 	    6-4
            Transfer Standards Having Fewer Than Five Discrete Outputs.    6-5

References	    7-1
Appendices	    A-l

     A.   Ultraviolet Photometric Procedure for Primary Ozone Standards.    A-l
     B.   Certification of the Boric Acid Potassium Iodide 	    B-l
     C.   Certification of the Gas Phase Titration with Excess
           Nitrogen Oxide Procedure as Transfer Standard	    C-l
     D.   Certification of the Gas Phase Titration with Excess
           Ozone Procedure as Transfer Standard 	    D-l
     E.   Certification of an Ozone Generator as a Transfer Standard . .    E-l
     F.   Certification of an Ozone Analyzer as a Transfer Standard. . .    F-l

-------
                                    FIGURES
Number                                                                    Page

 5-1     Sample comparison set-up of an 0  generation-type transfer
           standard to a UV primary standard (Procedure 1)	   5-3
 5-2     Sample comparison set-up of an 0  generation-type transfer
           standard to a UV primary standard (Procedures 2 and 3)  ...   5-5
 5-3     Example of temperature qualification test results showing up
           dependence on temperature	   5-12
 5-4     Example of a temperature dependence quantitatively defined
           as a correction factor	   5-13
                                             t
 5-5     Example of a defined barometric pressure dependence	   5-15
 5-6     Example of a comparison  (regression slope and intercept)  of
           a transfer standard to a primary 0  standard 	   5-22
 5-7     Example of a preliminary calibration relationship for an
           adjustable 0  device 	   5-23
 5-8     Example of a transfer standard certification relationship. . .   5-24
 5-9     A "family" of preliminary calibration curves 	   5-26
 5-10    Example of a chart showing recertification slope data for
           a transfer standard	   5-28
 5-11    Example of a chart showing recertification variability of slope
           and intercept for a transfer standard	   5-29
 A-l     Schematic diagram of a typical UV photometric calibration
           system	   A-12
 A-2     Schematic diagram of a typical UV photometric calibration
           system (Option 1)	   A-13
 B-l     Schematic diagram of a typical BAKI calibration system ....   B-20
 B-2     Components of a KI sampling train	   B-21
 B-3     Schematic diagram of a typical BAKI calibration system
           (Option 1)	   B-22
 C-l     Schematic diagram of a typical GPT system	   C-18
 D-l     Schematic diagram of a typical GPT system	   D-17
                                     XI

-------
                                ACKNOWLEDGMENTS

     The author wishes to acknowledge the generous and invaluable assistance
of Michael Beard, Dr. John Clements, Dr. Richard Paur, Larry Purdue, Jack Puzak,
Kenneth Rehme, and the many others who reviewed or otherwise helped in the pre-
paration of this document.
                                      xa i

-------
                                   SECTION 1
                         WHAT IS A TRANSFER STANDARD?

DEFINITION

     In ambient air monitoring work, the very low pollutant concentration
standards needed to calibrate ambient monitors are often difficult or impossible
to contain in a movable form.  These concentration standards must therefore be
generated in situ in some sort of flowing system.  When the monitor to be
calibrated is located at a remote monitoring site, it is often more convenient
to use a transfer standard rather than a primary standard calibration system.
A transfer standard is defined as a transportable device or apparatus which,
together with associated operational procedures, is capable of accurately
reproducing pollutant concentration standards or of producing accurate assays
of pollutant concentrations which are quantitatively related to an authorita-
tive master standard.

     Transfer standards may be used for many different purposes.  In this
document, however, the discussion of transfer standards for ozone (0 ) applies
only to the final standard used to calibrate an air monitoring analyzer (usually
under field conditions).

PURPOSE

     The primary functions of a transfer standard are to duplicate and distri-
bute concentration standards to places where comparability to a primary stan-
dard is required.  Ideally, only one primary standard exists for each entity
to be measured.  This standard is often maintained by the National Bureau of
Standards (NBS).  The NBS then makes available a limited number of secondary
standards that are carefully compared and certified against the primary standard.
                                      1-1

-------
1.  WHAT IS A TRANSFER STANDARD?/Purpose

These special transfer standards are known as Standard Reference Materials
(SRM) .   The limited number and expense of SRM's generally make them impractical
for the numerous routine applications for standards.  The SRM's are normally
reserved as local standards and referred to as "local primary standards".
These local primary standards are then used to certify local (secondary)
transfer standards — often called "working standards" — for routine use.
Occasionally, additional intermediate standards are used.  With each additional
stage,  the number of standards available is multiplied.  Each standard is
traceable through a chain of "higher" standards to the primary standard.  How-
ever, each lower standard in the chain must assume a somewhat lower accuracy
and hence less authority than the preceding standard.

     In the less-than-ideal realm of concentration standards for ambient air
pollutants, this scheme is complicated by many factors.  Not the least of
these factors is that available SRM1s may be in a form requiring an accurate
application technique (e.g., a permeation device must be used in an accurate
dilution apparatus to obtain a concentration standard), or that SRM* s may not
be available at all, as is the case for 0 .  In these situations, the technique
or procedure used to generate a local primary standard dynamically becomes
critically important.  Nevertheless, the function of a transfer standard re-
mains the same:  to transfer the authority of a pollutant standard from a locally
generated primary standard to a remote point where it is used to calibrate an
air monitoring analyzer.

HOW TRANSFER STANDARDS ARE USED

     In use, a transfer standard is first precisely related to an SRM or local
primary standard by careful comparison.  Then it is transported to the site of
the analyzer to be calibrated and used to calibrate the analyzer.  For some
types of transfer standards used in air monitoring, and particularly for 0
transfer standards, it is highly desirable to recompare the transfer standard
to the local primary standard following its return from field use; this recom-
parison serves as a supplemental check on reliability.

                                     1-2

-------
1.  WHAT IS A TRANSFER STANDARD?/Use

     While this concept is relatively simple, the actual use of transfer stan-
dards for 0  is not so simple.  Because of the nature of 0  , transfer standards
must be capable of accurately reproducing standard concentrations in a flowing
system.  Ozone transfer standards are complex systems consisting of devices or
equipment that generate or assay 0  concentrations.  Consequently, their certi-
fication and use must be in accordance with prescribed procedures that are speci-
alized to each specific type of transfer standard.  Section 4 describes several
different types of 0  transfer standards and provides general information on
the use of each.  Sections 5 and 6 provide additional information and specifica-
tions on certification of 0  transfer standards.

     Due to the complexity of 0  transfer standards, there is always some
degree of doubt or lack of confidence in their reliability.  Therefore,  a major
                                            t
part of the use of an 0  transfer standard is the need to qualify it, i.e., to
determine and prove that it has adequate stability (repeatability)  and reliabil-
ity under conditions of use.  General procedures for determining repeatability
prior to use, and for assuring reasonable reliability during use are presented
in Section 5.  Appendices B, C, D, E, and F contain more specific information
on the qualification of several commonly used types of transfer standards for 0 .
                                      1-3

-------
                                   SECTION 2

                      WHY USE AN OZONE TRANSFER STANDARD?
OZONE CONCENTRATION STANDARDS
     The reactivity and instability of 0  precludes the storage of 0  concentra-
tion standards for any practical length of time.  It also precludes direct certi-
fication of 0  concentrations as SRM's.  Moreover, there is no available SRM
which can be readily and directly adapted to the generation of 0  standards
analogous to permeation devices for sulfur dioxide (SO ) and nitrogen dioxide
(NO ) .  While an 0  generating device may someday achieve SRM status, the pre-
sent situation requires that 0  standards be generated and certified locally,
based on a related chemical or physical primary standard.  Dynamic generation
of 0  concentrations is relatively easy with a source of ultraviolet (UV) radia-
tion.  But accurately certifying an 0  concentration as a local primary stan-
dard, requires assay of the concentration by a comprehensively specified analy-
tical procedure which must be performed each time a standard is needed.

     Until recently, the analytical procedure prescribed by the U.S. Environ-
mental Protection Agency (EPA) (under Title 40, Code of Federal Regulations
(CFR) , Part 50, Appendix D) for certifying local primary 0  concentrations has
been a wet chemical technique based on spectrophotometric analysis of iodine
generated by 0  in neutral buffered potassium iodide (NBKI) and referenced to
an arsenious oxide primary standard.  EPA has amended these regulations by
regulations by replacing the NBKI technique with a better technique based on
absorption of UV radiation and referenced to the well-established absorption
coefficient of 0  at a wavelength of 254 nm (EPA 1979).  This new UV technique
is described briefly in Section 3 and appears in complete form in Appendix A
of this document.
                                      2-1

-------
2.  WHY USE AN OZONE TRANSFER STANDARD?/Advantages
ADVANTAGES OF OZONE TRANSFER STANDARDS
     We are assuming here that 0  concentration standards are needed to calibrate
an 0  analyzer for ambient monitoring.  Usually, a number of such analyzers need
to be calibrated, and they are located at various field sites separated by ap-
preciable distances.  Also, these analyzers presumably require recalibration
at periodic intervals.  Consequently, a large number of 0  standards will be
required at various times and places.  Ozone standards may also be needed to
check the span or precision of these analyzers between calibrations.

     As noted earlier, primary 0  standards* must be obtained at the time and
                                «J
place of use by analytically assaying dynamically-generated 0  concentrations.
The primary UV standard procedure could be used at each field site each time
an 0  analyzer required calibration, but the availability of 0  transfer stan-
dards provides an alternative:  the primary UV standard procedure can be used
at a fixed location to certify one or more 0  transfer standards which can then
be transported to the various field sites and used to calibrate the 0  analyzers.
The choice of whether to use transfer standards or not is up to the monitoring
agency, and may depend on a number of factors such as number and location of
ambient 0, analyzers to be calibrated, equipment and expertise available, fre-
quency of calibration, etc.  Some possible advantages from using transfer stan-
dards are described below.
Singularity

     By using transfer standards, all 0  analyzer calibrations in a network
can be related to a single UV photometer.  All measurements in the network are
*From here on, the word "local" has been dropped from "local primary ozone stan-
 dard".  This is largely for convenience, but it also reflects the fact that at
 present local primary 0  standards are also absolute primary C>3 standards by
 default, since no higher 0  standards are available from NBS.

                                     2-2

-------
2.  WHY USE AN O20NE TRANSFER STANDARD?/'Advantages

then directly related to a single common standardt, which can be  verified by
intercomparison with other UV standards (see below) more easily and more fre-
quently than multiple UV photometers could be.  Concern about variations or
discrepancies among  multiple UV photometers is then eliminated.   If the  common
UV photometer should ever prove to be significantly inaccurate on some occasion,
a single common correction factor could possibly be applied to all  network data
(provided that historical details of the photometer pc  jrmance are quantita-
tively determinate).

Primary Standard  Uncompromised

     The use of transfer standards allows the primary standard equipment and
procedures to be  used at a fi-;ecl laboratory location w!vrre the conditions of
use can r*> carefully controlled.  Neither tir.-j' equipment nor procedures need to
be comp:    sed for field use  and there is r>^ i isk of damage to sensitive
equipment curing  transport.  1T.ider the controlled cond -:~ions and  fixed location,
variability in the generated ~r:*nary stands* i;, will b   educed, providing
better accuracy and uniformity   ::ong all 0   i.-' r-zers    the network.

Economy
     Tr
the pr •'.
(see 5 •
based o"
.-•-hc:r~
at J i
fie
pri.  ..
:er standards may .    .ess
/  standard procedure   TV
n 4)  offer an ager ,.
is available budge"
            die 0  analyzer
                                       oens.
                 or ecup.
           >n-f-.inue"   e.
           .n>  rd >-• ;>  pm-.-
                                              c  -'c
                                               expert.
                                    lo:  0  i
                                  ^7en be
                                  .  .  o p.
 quipment required for
  transfer standards
  transfer standards
 situation.   Spare or
'•; transfer standards
 brations may be quali-
  to avoid the cost of
 / standards through
 gencies,   There is
"No •?
                                   u c   to r £<-...>• eitix   -:o 0  co.' centrati^r
                                  ^    ,-e  ... ^  ;&i.  to   i.ain '

-------
2.  WHY USE AN OZONE TRANSFER STANDARD?/Advantages

also the possibility of purchasing periodic transfer standard certification
services from commercial laboratories.

Practicality

     Transfer standards are generally more rugged and easily portable than
primary standard equipment.  They can be designed to be more adaptable to a
variety of applications and to be insensitive to various field or transporta-
tion conditions, i.e., they can be optimized for field use.  They may be
easier to use, require less operator training, and be less subject to operator
error during use.

Intercomp arisen
     Because of the lack of an NBS SRM for 0 , there is always some doubt as
to the accuracy of a locally generated 0  standard.  Transfer standards can be
used conveniently to intercompare primary standards among various local, State,
and Federal agencies to assure accuracy and confidence.
Convenience
      " • " " ~                       ,

     Transfer standards may be more convenient to use than the primary 0
standard procedure.  Multiple transfer standards can be used simultaneously at
different locations.  The operation of transfer standards may be simpler or
more convenient.  Commercial transfer standards tend to be more readily avail-
able than equipment for the primary 0  standard procedure.  Also, transfer
standards are often more flexible or adaptable to calibration of various types
or models of 0  analyzers.

DISADVANTAGES OF OZONE TRANSFER STANDARDS

     As might be expected, the use of transfer standards is not without some
disadvantages.  Since a choice is available between using or not using transfer

                                      2-4

-------
2.  WHY USE AN OZONE TRANSFER STANDARD?'/Disadvantages

standards, the disadvantages must be weighed carefully against the advantages
to determine whether their use is appropriate or cost-effective.  Some of the
more important disadvantages are described below.

Qualification

     Before a device or procedure can be used as a transfer standard, it must
be tested and shown to have adequate performance and reliability.  It must be
precisely repeatable over reasonable periods of time and over the range of
conditions encountered during field use and during transport.  Qualification
of a transfer standard may require a series of initial tests to determine
reliability.  This problem is addressed in more detail in Section 5.

Certification of Accuracy

     Transfer standards have no authority until they are related to a primary
standard by critical comparison.  Moreover, they must be recompared to the
primary standard periodically to retain their certification.  The time and
effort necessary to certify and recertify a transfer standard against a primary
standard depends on the type or nature of the transfer standard, and may vary
considerably from one type to another.  Sections 5 and 6 address this problem
in more detail.

Reliability

     Most transfer standards for 0  involve complex apparatus or procedures
or both.  As a result, there is always some possibility of error, malfunction,
drift/ or some other cause ffy~ loss of repeatability.  Continual tests and
checks are necessary to verify and assure the continued accuracy ?nd integrity
of the transfer standards.  Additional information on quality assurance pro-
cedures tailored to the particular type of transfer standard used is contained
in Sections 5 and 6.
                                      2-5

-------
2.  WHY USE AN OZONE TRANSFER STANDARD?/Disadvantages

Loss of Accuracy

     Use of a transfer standard in lieu of a primary standard will necessarily
increase the possible error by some degree.  However, an adequate quality
assurance procedure will keep the loss of accuracy within reasonable limits.
                                      2-6

-------
                                   SECTION 3
         PRESCRIBED ULTRAVIOLET PROCEDURE FOR PRIMARY OZONE STANDARDS

     The procedure prescribed by EPA regulations for obtaining dynamic primary
0  concentration standards is specified in Part 50 of Title 40, CFR Appendix
D (EPA 1979).  This section of the regulation prescribes a new assay technique
based on absorption of UV radiation, replacing the previous iodometric technique.
The new UV technique is reproduced in Appendix A to this document.

DESCRIPTION
     The UV technique requires a stable 0  generator, a UV photometer, and a
source of clean, dry, pollutant-free air.  A flowing (dynamic) system is set
up in which clean air is passed through the 0_ generator at a constant flow
rate and discharged into a multiport manifold.  The 0  concentration in the
manifold is assayed by the photometer and is available for calibration of 0
analyzers or certification or transfer standards.   (To certify some types of
transfer standards, some modifications in this system may be required; see
Section 5.)  After the air flow rate is adjusted, the 0  generator is adjusted
to provide the approximate 0  concentration desired.  The UV photometer is
then used to measure the UV absorption of the generated concentration at a
wavelength of 254 nm.  This transmittance measurement, together with the well-
established absorption coefficient of 0  at that wavelength and various instru-
ment parameters, is used to calculate the 0  concentration by means of the
Beer-Lambert absorption law.  The -"-curacy of the photometer is critically im-
portant to this technique; however, certain commercial and laboratory photo-
meters have been shown to be adequate to the task.  Additional information and
guidance on the use of this technique and on selecting a suitable photometer is
given in the following sections.

-------
                                   SECTION 4
                     TYPES OF TRANSFER STANDARDS FOR OZONE

     Several different types of devices or techniques can be considered for
use as transfer standards for 0 .  They can be loosely grouped into three
general categories:  analytical instruments, manual analytical procedures, and
generation devices.  Within these categories are a variety of techniques and
devices having widely different degrees of precision, reliability, portability,
economy, and convenience.  No single technique or device is necessarily best
for all situations.  An agency should select an appropriate type of transfer
standard based on a complete evaluation of its situation with respect to
available funds, available personnel and expertise, equipment on hand, location
and distance to field sites, modes of transportation used, number and calibra-
tion frequency of analyzers to be calibrated or spanned, etc.

     A discussion of the three categories follows, together with examples of
some techniques and devices which are currently available and have been used
as transfer standards for 0 .  Of course, any device or technique to be used
must first qualify as an acceptable 0  transfer standard by demonstrating ade-
quate repeatability.  Then, the transfer standard must be certified by relating
it to a primary standard.  Refer to Sections 5 and 6 for detailed information
on the actual tests and procedures used to qualify, certify, and establish the
reliability and accuracy of 0  transfer standards.

ANALYTICAL INSTRUMENTS

     Transfer standards which fall into this category employ an instrumental
technique to assay stable, flowing 0  concentrations.  The analytical instru-
ment must be capable of measuring 0  concentrations adequately over the con-
centration range of interest, using a well-defined chemical or physical property

                                      4-1

-------
4.  TYPES OF TRANSFER STANDARDS/Analytical Instruments

of 0 .   Examples of transfer standards using an instrumental technique would
include almost any commercial 0  analyzer designed for ambient air monitoring,
or any other 0  analyzer that can measure 0  concentrations suitably in the
appropriate concentration range.

     As with all 0  transfer standards, the analytical instrument to be used
must be related to a primary standard once it has been qualified.  This process
is accomplished by simply allowing the analytical instrument to sample or ex-
tract, from the output manifold, a portion of the primary standard 0  concen-
tration obtained by the prescribed UV technique (see Section 3).   Most analy-
tical instruments provide a measurement of 0  over a range of concentrations.
The entire analytical range can be certified by comparing the response of the
instrument to a series of different 0  concentrations over the range.  (See
"Certification" in Sections 5 and 6.)  The certification relationship between
the analytical instrument and the primary 0  standard is then expressed math-
ematically and by a continuous plot of average instrument response versus pri-
mary standard 0  concentration, as shown in Figure 5-8 (pg 5-24).

     Alternatively, the analytical instrument can be related to a primary
standard at only one or two 0  concentrations rather than over the entire
analytical range, thereby reducing the time required for certification.  When
subsequently used at a field site, the 0  concentration must be adjusted until
the transfer standard indicates exactly the same response as it did during
comparison to the primary standard.  Once this identical concentration is
obtained, it can be quantitatively diluted with zero air to produce lower con-
centrations.  However, this dilution technique may reduce the accuracy of the
field concentrations because of the errors associated with the flow measure-
ments and because of the increased complexity of the field calibration system
due to the additional dilution and mixing apparatus.
     The analytical instrument itself can only assay existing 0  concentrations.
When used as a transfer standard to calibrate an 0  monitor at a field site,
some additional means must be provided to generate the 0  concentrations:

                                     4-2

-------
4.  TYPES OF TRANSFER STANDARDS/Analytical Instruments

usually, a UV 0  generator, an air pump, and an ambient air scrubber to provide
clean zero air.  The 0  generator must produce very stable concentrations of
0  (preferably less than ± 2% change per hour).

     It is debatable whether or not the clean air and 0  generator components
should be considered a part of the transfer standard.  Ideally, an 0, transfer
                                                    /                3
standard should be self-contained such that it can completely reproduce 0
standards.  To be self-contained, the 0  generation components must be an in-
tegral part of the transfer standard.  An added advantage of this concept is
that the 0  generation components can be inspected, tested, and serviced when-
ever the analytical instrument is recertified.  On the other hand, the authority
of the transfer standard is clearly contained in the certification of the analy-
tical instrument.  The 0  generation components could thus be considered in-
cidental to the use of the transfer standard.' Some advantages may therefore
be obtained by equipping each field site with its own 0  generation system —
which might also double as a zero-and-span system — and transporting only the
analytical instrument from site to site.

     General information on qualification and certification of analytical in-
struments is contained in Section 5, and the specifications in Section 6.
More specific information for qualifying and certifying an 0  analyzer as a
transfer standard may be found in Appendix F.

Chemiluminescenee Analyzers

     Chemiluminescence analyzers may be used as transfer standards, but they
tend to have several practical disadvantages.  Most such analyzers intended
for ambient monitoring are designed for continuous operation, and may experi-
ence difficulty during intermittent operation.  Most require from 1 to 4 hours
to warm up and stabilize before highly repeatable performance can be expected.
The need for an external supply of ethylene and the potential safety hazards
of frequent connection and disconnection of the ethylene source may constitute
serious drawbacks.  Commonly available commercial analyzers tend to be heavy,

                                      4-3

-------
4.  TYPES OF TRANSFER STANDARDS/Analytical Instruments

not readily portable, and not designed for rapid and frequent start-up and
shutdown.  Although not particularly delicate, the electronics, photomultiplier,
cell assembly, and coolers are not intended to resist the frequent mechanical
shocks that might be encountered in transit.  Sensitivity to temperature, line
voltage, and pressure variations would have to be checked carefully.  Also, the
cost of chemiluminescence analyzers tends to be quite high.

     On the positive side, there is at least one commercially available chemi-
luminescence 0  analyzer designed for portable, intermittent operation.  It is
fairly light weight, battery or line powered, has a self-contained ethylene
supply, and claims to have good temperature regulation.  Other portable analy-
zers may be available in the future.  Such portable units would alleviate many
of the problems mentioned above and could be seriously considered as candidate
transfer standards.

Ultraviolet Analyzers

     As of this writing, at least one UV analyzer in the ambient concentration
range is commercially available.  Others are likely in the future.  One model
includes a self-contained 0  generator.  In general, UV analyzers appear to
                                 t
be better suited for use as transfer standards than chemiluminescence analyzers.
The UV units are considerably simpler to set up and operate, do not require
ethylene, and are more tolerent of intermittent operation.  They are usually
self-contained and lighter in weight than the chemiluminescence units.  Al-
though not designed specifically for portability, the UV analyzers may be trans-
ported readily and relatively safely if given careful handling; the UV optical
systems in these analyzers, however, must not, be subjected to rough handling.
The warm-up and stabilization time of the UV analyzers is normally less than
an hour.  They are likely to be sensitive to temperature and barometric pres-
sure changes and may require corrections for those parameters.  Line voltage
sensitivity should also be checked carefully.
                                     4-4

-------
4.  TYPES OF TRANSFER STANDARDS/Analytical Instruments

     When used as a transfer standard, a UV analyzer may appear to function
very much the same as the UV photometer in the 0  primary standard procedure
(see Appendix A).  It is therefore easy to confuse the two.  The distinction
between the two is important and should be clearly understood.  When a UV photo-
meter is used to generate primary standard 0  concentrations, it must meet all
of the specifications prescribed in the 0  primary standard procedure (see
Appendix A), and the concentration measurements are referenced to the absorp-
tion coefficient of 0 .  In contrast, an 0  analyzer used as a transfer stan-
dard can only certify an 0  concentration by previous comparison to a primary
standard; reference to the 0  absorption coefficient is used only indirectly,
although it may serve as an internal check of the transfer standard's reliabil-
ity  (see Section 5).
MANUAL ANALYTICAL PROCEDURES

     Transfer standards in this category consist of a combination of laboratory
apparatus and a prescribed procedure that is manually executed to assay 0  con-
centrations.  The procedure must be capable of measuring 0  concentrations
adequately over the concentration range of interest, using a well-defined chemi-
cal or physical property of 0  .  Since the apparatus, the execution of the pro-
cedure, and possibly even the procedure itself may vary somewhat from user to
user, each user must qualify and certify a procedural transfer standard in his
own use-situation.  Examples of transfer standards using manual analytical pro-
cedures include the Boric Acid Potassium Iodide  (BAKI) technique, the Gas Phase
Titration  (GPT) with excess nitrogen oxide  (NO) technique, and the GPT with
excess 0  technique.  Any other analytical procedure capable of measuring 0
concentrations adequately in the appropriate concentration range could also be
considered a possible transfer standard.  Since only 0  should be present in
the atmosphere to be analyzed, interferences are not normally a problem.
     At one time, procedures based on the BAKI and GPT techniques were
considered candidate procedures for obtaining primary 0  concentration stan-
dards  (EPA 1976).  A BAKI procedure was even given that status on a temporary

                                     4-5

-------
4.  TYPES OF TRANSFER STANDARDS/Manual Analytical Procedures

basis (EPA 1979).  For this reason the two GPT procedures which were previously
published by EPA (1976) and the BAKI procedure were all designed to provide
measurements of 0  concentration based on a specified non-0, standard.  For
BAKI, the standard is potassium iodate primary standard, and for GPT the stan-
dard is an NO SRM.

     When these techniques are used as transfer standards, they are used to
transfer the authority of a UV primary standard, though only in a relative
sense.  The techniques' capability to provide 0  concentration assays based
on an independent standard is not used as a primary part of the transfer stan-
dard function.  It is important to understand this distinction between the use
of these techniques as transfer standards and their use as independent (primary)
standards.  Except for a temporary exception for BAKI, 0  monitors in the field
must be calibrated by either direct or traceable reference to a primary stan-
dard based on UV absorption as specified in the regulations (EPA 1979) and not
by reference to one of the independent standards associated with these tech-
                                                          \
niques.  Each of the procedures based on BAKI and GPT has now been revised to
reflect its use as a transfer standard rather than as a primary standard.  These
suggested, revised procedures, together with specific guidance on qualification
and certification, are included in Appendices B, C. and D.

     Once a procedural manual analytical transfer standard has been qualified
for use, it is certified.  The certification is similar to the certification
of an analytical instrument transfer standard:  The procedure must be related
to the primary standard by direct comparison.  Primary 0  standards over the
range of interest are assayed according to the procedure and compared (see
Section 5).  Then the certification relationship between the analytical pro-
cedure and the primary ozone standard is expressed both mathematically and by
a continuous plot of the concentrations indicated by the procedure versus the
primary ozone concentrations (see Figure 5-8, pg 5-24).  Since manual analytical
procedures may have considerable variation in response to a fixed 0  concentra-
                                                                   
-------
4.  TYPES OF TRANSFER STANDARDS/Manual Analytical Procedures

assays must be specified as a part of the transfer standard procedure and
must then be carried out during use of the procedure.

     When the procedure is subsequently used at a field site to calibrate an
0  analyzer, the certification relationship is used to determine the "certified"
(standard) 0  concentration from the concentration indicated by the procedure.
Even though the indicated concentration is derived from the procedure's indepen-
dent standard, the certified concentration to be used for the analyzer calibra-
tion must be determined from the certification relationship.  Because of the
possible variability noted above, it may be advisable to carry out multiple
assays on each 0  concentration at the field site and average them to achieve
accurate results.
                                             /
Boric Acid Potassium Iodide Procedure

     Following expiration of the 18-month period, during which time the BAKI
procedure is approved for independent certification of primary 0  concentra-
tions  (EPA 1979), this technique may be useful as a transfer standard.   Even
during the 18-month period the procedure would be more reliable if used as a
transfer standard instead of a primary standard.  Successful use of the pro-
cedure requires•thorough familiarization with the procedure and good laboratory
technique on the part of the user; careful attention to the procedural instruc-
tions is important to control variability.  If variability proves to be a
problem, multiple samples may be required of each assay concentration.

     Nevertheless, use of the BAKI procedure is quite inexpensive and may be
advantageous for an agency that has the necessary equipment, is familiar with
the procedure or able to train operators, and is willing to spend the time
required.  The agency, though, must be able to demonstrate acceptable measure-
ment variability.  A suggested BAKI procedure incorporating the appropriate
restrictions and steps to qualify and certify the procedure as a transfer stan-
dard is provided in Appendix B.
                                      4-7

-------
4.  TYPES OF TRANSFER STANDARDS/Manual Analytical Procedures

     As with instrumental transfer standards, the BAKI procedure can only assay
existing 0  concentrations.  Hence, an 0  generator system is needed at the
field site to produce 0  concentrations that are assayed by the procedure and
then used to calibrate the 0  monitor.  Whether or not this C>3 generator system
should be considered a part of the procedure is discussed briefly above under
"Analytical Instruments".
Gas Phase Titration Procedures

     There are two separate GPT techniques:  in one, an CL concentration is
titrated with an excess of NO; in the other, an NO concentration is titrated
with excess 0 .  Suggested procedures and guidance for qualifying, certifying,
and using both techniques as transfer standards are presented in Appendices C
and D, respectively.  As the two procedures are similar they are discussed to-
gether here.

     Both GPT procedures require the user to be thoroughly familiar with the
procedure and the GPT apparatus.  Careful attention to the procedural steps
and specifications is essential for good results.  The need for an NO cylinder
standard may affect the portability of the system.
                                t

     The GPT-NO procedure requires the use of an NO analyzer to measure the
titrated NO.  This technique is an approved and widely used technique for the
calibration of NO  analyzers.  Many agencies are familiar with the technique,
have the required equipment, and may be able to use the same equipment for
both NO  and 0  calibrations conveniently at sites which have an NO analyzer.
       *L      .5

     The excess 0  GPT procedure does not require the calibrated NO analyzer,
and can therefore be used at sites which do not have an NO analyzer.  The
excess 0  procedure is somewhat more critical with respect to flow dynamics
than the excess NO technique, and particular care should be taken to insure
that all procedural specifications are met.
                                     4-8

-------
4.  TYPES OF TRANSFER STANDARDS/Manual Analytical Procedures

     Because of the rather specific flow dynamics requirements of the GPT
procedures, and the intimate relationship of the 0  generator to the GPT
system, the 0  generator is usually considered an integral part of the GPT
system and not readily separable.  Thus, the GPT procedures, though basically
analytical techniques, can assay only those 0  concentrations generated within
the GPT system.  The GPT system provides assayed 0  concentrations convenient
for calibrating field 0  analyzers at an output manifold.  Certifying GPT
systems is therefore very similar to certifying 0  generation devices (see
Section 5 for details).
GENERATION DEVICES

     Transfer standards in this category are simply devices that generate
accurate 0_ concentrations without having any capability to assay the gen-
erated concentration output.  The accuracy of these devices depends entirely
on their inherent generation stability and reproducibility under changing con-
ditions of use.  The most common example of a generation device is the UV
(photolytic) 0  generator.  Other types of 0  generators capable of generating
              j                             J
reproducible 0  concentrations in the appropriate range may also be suitable.

     Ozone generator transfer standards consist of an air pump, an air scrub-
ber to provide zero air, a flow control system, a means to generate 0  within
the flowing zero air stream, and an output manifold at which the 0  concentra-
tion standard is available for calibration or spanning of field 0  analyzers.
The pump and air scrubber components could be considered ancillary to the 0
generator system, but ideally these components should be included as integral
to the system, since the quality of the zero air and the flow regulation are
important to the accuracy of the generation system.  If possible, the genera-
tion system should be completely self-contained.

     Most generation devices have means to adjust the 0  output concentration
over a considerable range for convenient calibration of field 0  analyzers.
This adjustment normally has a dial or scale associated with it that can be

                                     4-9

-------
4.  TYPES OF TRANSFER STANDARDS/Generation Devices
related to primary 0  standards by a relationship as shown in Figure 5-7  (pg
5-23) .  In some generation systems, the output concentration is varied by chang-
ing the flow rate, which produces a curved (nonlinear) relationship to the
primary 0  standards unless the concentration is plotted versus the reciprocal
of flowrate.  Fixed or discretely adjustable 0  generators produce only one or
a few fixed 0  concentrations that may have to be diluted at the field location
to calibrate an 0  analyzer.

     In concept, it would seem that generation devices are well suited as
transfer standards for 0 .  They are often of relatively simple design, easy
to use, moderate in cost, fairly rugged, commercially available, and relatively
immune to operator error.  In practice, the devices fall somewhat short of the
ideal, but nevertheless warrant serious consideration as transfer standards.
As noted above, the devices may have no assay capability and depend entirely
on their own inherent stability.  Consequently, their sensitivity to changing
conditions from certification lab to field site and their stability with time
must be checked carefully and frequently (see Section 5).   Mechanical and elec-
trical integrity and flow stability are also important.  Since most generation
devices are sensitive to pressure changes (altitude) and often to temperature,
corrections for these variables may be necessary.  A modest but important warm-
up period is usual before the generation device has a stable, repeatable output.
With some devices, a restabilization period may be required after each adjust-
ment of the output concentration.

     Generation devices are likely to have lamps, glassware, and various elec-
tronic components that require reasonably careful handling.  Those devices
that require dilution of a fixed 0  concentration must have suitable apparatus
and flowmeters to effect accurate dilution, putting an extra burden of accuracy
on the operator.  Since the generation devices provide their own 0  output con-
centrations, a slightly modified procedure is necessary to certify them against
a primary UV 0  standard (see Section 5).
                                      4-10

-------
4.  TYPES OF TRANSFER STANDARDS/Generation Devices

     General information on the qualification and certification of 0  generators
is contained in Section 5, the specifications in Section 6.  More specific in-
formation for qualifying and certifying an 0  generator as a transfer standard
may be found in Appendix E.
                                     4-11

-------
                                   SECTION 5

            ESTABLISHING THE AUTHORITY OF OZONE TRANSFER STANDARDS
     As noted in previous chapters, the primary purpose of an 0  transfer stan-
dard is to transfer the accuracy of a primary 0  concentration standard from
one place and time to another.  Because a transfer standard has initially (by
definition) no authority of its own, its authority must first be established.
The essence of establishing the authority of the transfer standard is to
establish a high probability or confidence that 0  concentration standards ob-
tained by means of the transfer standard, under a variety of operational condi-
tions, are very nearly as accurate as primary 0  standards.  This confidence
is established first by determining that the transfer standard has adequate
reproducibility to qualify it as a useful transfer standard, then by certifying
the transfer standard by relating it to a UV primary standard, and finally by
periodically recertifying it by reverifying its accuracy and stability.
TRANSFER STANDARD SPECIFICATIONS

     Section 6 specifies the formal requirements a transfer standard must meet
to be certified for use in calibrating ambient 0  analyzers.  The material in
Section 5 explains those requirements more fully, as well as the actual techni-
ques used for qualifying and certifying transfer standards.

COMPARING TRANSFER STANDARDS TO AN ULTRAVIOLET PRIMARY OZONE STANDARD
     Basic to the qualification and certification of any 0  transfer standard
is the need to compare the output  (either a concentration assay or an 0  con-
centration) of the transfer standard to a primary 0  standard, so that relation-
                                                   •J
ships such as shown in Figures 5-7 and 5-8  (pp 5-23 and 5-24) can be determined.
Exactly how such a comparison is carried out depends on whether the transfer
standard is of the assay-type or the 0 -generation type.
                                      5-1

-------
5.  ESTABLISHING AUTHORITY/Comparing to Primary Standard

Assay-TyPe Transfer Standards

     For transfer standards which provide an assay of an externally generated
0  concentration (BAKI, 0  analyzer), the transfer standard is simply connected
to the output manifold shown in Figures 1 and 2 of Appendix A.  Make sure that
the UV calibration apparatus can supply sufficient flow for both the photometer
and the transfer standard.  The output of the transfer standard is an indicated
concentration, which can be compared directly to the primary standard concentra-
tion obtained from the UV calibration system.

Ozone-Generation Type Transfer Standards

     Transfer standards that generate 0  concentrations themselves include C>3
generators and may include those assay procedures which have an integral source
of 0   (such as GPT).  In comparing a generation-type transfer standard to a UV
primary 0  standard, it obviously cannot be simply connected to the output mani-
fold shown in Figures 1 and 2 of Appendix A; some alternate procedure is neces-
sary.  Described below are three alternate procedures that may be used to com-
pare a generation-type transfer standard to a UV primary 0  standard.  They are
listed in order of preference.

     Other procedures are not necessarily precluded, but due concern for accu-
racy should be exercised.  In designing or selecting a transfer standard, pre-
ference should be given to transfer standard configurations that allow direct
comparison with the photometer, as described in the first procedure.  Thus, a
transfer standard should have an air supply capable of providing sufficient
zero air for the photometer reference cycle without adversely affecting the
generated 0  atmospheres.
Procedure 1:  Assay by Ultraviolet—
     The UV procedure for obtaining
3 and Appendix A) is basically an analytical technique using a UV photometer.
The UV procedure for obtaining primary 0  standards (described in Section
                                      5-2

-------
5.  ESTABLISHING AUTHORITY/Comparing to Primary  Standard

The photometer can be used to assay  the output concentration of a generation-
type transfer standard.  To do this,  the photometer must be disconnected from
its own 03 generation and output system and connected to the transfer standard
output  (see Figure 5-1).  Care must  be exercised to disturb the UV photometer
as little as possible from its normal configuration, and to ensure that the
output flow of the transfer standard exceeds the flow demand of the UV photo-
meter .
r


ZERO
AIR
L—













1
fn OUTPUT
_^. °» „ MANIFOLD
Ft OW 3 I " ' " ' T - -- %/r-»i-r
— uu«» 	 -1 —— "' - »" VcNI
CONTROLLER GENERATOR ' 	 1 | 	 '
i

J

FLOW TVUO.WAY 	
CONTROLLER | VALVE i
VENT I
UV PHOTOMETER f f
OPTICS |
: SOURCE I
o i
i
i
i
SIGNAL , , _. 1
i ^ FLOW I
PROCESSING IFLOWMETER PUMP " "'^ EXHAUST!


Figure 5-1.  Sample comparison set-up of an 0  generation-type transfer stan-
             dard to a UV primary standard (Procedure 1).
                                     5-3

-------
5.  ESTABLISHING AUTHORITY/Comparing to Primary Standard

     A significant problem arises with this procedure, however.  In order to
accurately measure I  for the transmittance (I/I ) measurement, the UV photo-
                    o                           o
meter must be able to sample zero air from the same source as that used for
the generation of the 0, concentrations (see Appendix A).  If the zero air
supply of the transfer standard (or GPT system) is capable of providing suf-
ficient additional zero air for the photometer, it may be tapped and connected
to the two-way valve as shown in Figure 1 of Appendix A.  Care must be exercised
in the process to ensure that the transfer standard is not adversely affected.
If the zero air supply cannot provide sufficient additional zero air, or cannot
be readily tapped, then Procedure 2 or 3 below must be used to compare the
transfer standard to the UV primary standard.

Procedure 2:  Comparison by Calibrated Ozone Analyzer—
     In its normal configuration, the UV primary 0  standard system produces
assayed 0  concentration standards.  Such standards can be used to calibrate
an ordinary ambient 0_ analyzer as specified in Appendix A.  The calibrated
0  analyzer is then immediately used to assay the output concentrations of
the generation-type transfer standard (see Figure 5-2).  In this situation the
0  analyzer can be used as a temporary or short-term transfer standard without
the otherwise required qualification tests described later in this section,
                                  t
since the 0  analyzer is used immediately after calibration.  Accordingly, it
is used under the same conditions as those during calibration, and it is not
shut down or moved between calibration and use.  Nevertheless, it is important
that the analyzer be stable between calibration and subsequent use to assay
the transfer standard output.  The analyzer calibration should therefore be
rechecked at several concentrations against the primary standard after the
transfer standard is certified.  Also, the entire process of calibration, use,
and calibration recheck of the 0  analyzer should be repeated each time a
generation-type transfer standard is certified or recertified.
Procedure 3:  Comparison by Uncalibrated Ozone Analyzer—
     It is possible to avoid the calibration and calibration recheck steps in
the previous procedure by using an uncalibrated 0  analyzer to compare the

                                     5-4

-------
 5.   ESTABLISHING AUTHORITY/Comparing to Primary  Standard
               ZERO
                AIR

FLOW
CONTROLLER

Fo


°3
GENERATOR
                                                      OUTPUT
                                                     MANIFOLD
                                                                       VENT
                                  03
                               ANALYZER
   o
                                                  OUTPUT
                                                  MANIFOLD
                                                           EXTRA OUTLETS CAPPED
                                                             WHEN NOT IN USE
                                                            VENT
                                                                    SOURCE
                                                                      o
                    SIGNAL
                  PROCESSING
                 ELECTRONICS
Cl nUUuCTPR

'p

FLOW
CONTROLLER


PUMP
                                                      EXHAUST
Figure  5-2.
Sample  comparison set-up  of an O  generation-type transfer  stan-
dard to a  UV primary standard (Procedures  2  and 3).
                                       5-5

-------
5.  ESTABLISHING AUTHORITY/Comparing to Primary Standard

transfer standard to a primary standard.  The primary 0  standard, the transfer
standard output, or both, are adjusted until each is producing exactly the same
output concentration.  This equivalence point is determined by switching an un-
calibrated 0  analyzer back and forth between the primary standard and the trans-
fer standard  (see Figure 5-2).  Of course, the 0  analyzer must not drift or
change its sensitivity during this process, and sufficient time must be allow-
ed for a stable reading on each source.  When there is no doubt that the analyzer
response is exactly the same for both the primary standard and the transfer stan-
dard, the transfer standard output is comparable to the UV primary standard at
that concentration.  This process must be repeated for each concentration at
which the transfer standard is to be compared.

     This procedure is of obvious advantage for transfer standards which have
only a single or a few fixed outputs.  It could also be used to rapidly verify
the accuracy of transfer standards between their periodic recertifications.

QUALIFICATION

     The first step in establishing the authority of a candidate transfer
standard is to prove that it qualifies for use as a transfer standard.  In
other words, can the output (either an actual 0  concentration or a concentra-
tion assay, depending on the type) of the candidate transfer standard be trusted
under the changing conditions of use that might be encountered in field use.
A transfer standard must be assumed unacceptable until it can be conclusively
demonstrated to be acceptable.

     The primary requirement of a transfer standard is repeatability — repeat-
ability under the stress of variable conditions that may change between certi-
fication and use.  A candidate transfer standard is qualified by proving that
it is repeatable over an appropriate range for each variable likely to change
between the time and place of certification and the time and place of use.  Ac-
cording to the specifications in Section 6, the repeatability must be within
± 4% or ± 4 ppb, whichever is greater, for each condition or variable that may
change between the point of certification and the point of use.
                                     5-6

-------
5.  ESTABLISHING AUTHORITY/Qualification

     Selecting the conditions that are likely to vary and that may affect the
repeatability of the device or procedure is largely a matter of intelligent,
informed judgement.  To a large extent, the variables will depend on the
nature of the device or procedure; for some candidate transfer standards, the
variables to be considered may be quite numerous.  It is the user's responsibil-
ity to determine all of the conditions to be considered in the demonstration
of repeatability before a candidate transfer standard can be considered
qualifed for use as a transfer standard.  Common conditions likely to affect
a wide variety of types of transfer standards include such items as ambient
temperature, line voltage, barometric pressure, elapsed time, physical shock,
etc.  These variables are discussed individually later in this section.
Conditions not likely to affect the transfer standard can usually be eliminated
from consideration.  The user must, however, be constantly alert for the
                                             t
unusual situation where an unexpected condition may significantly affect the
repeatability of a transfer standard.

     Note that a transfer standard does not necessarily need to be constant
with respect to these variables, only repeatable or predictable.   While it
is certainly desirable that a device or procedure be insensitive to any given
variable, it may still qualify as a transfer standard if it is repeatable with
respect to the variable.  For example, it may be difficult to find or design a
generation-type transfer standard device that is insensitive to barometric
pressure.  However, if it is repeatable with respect to barometric pressure,
the relationship can be quantitatively defined by a curve or table.  At the
time of use, the local barometric pressure must be measured and the curve or
table used to "correct" the transfer standard's indicated output.  This technique
is acceptable for one or perhaps two variables.  But beyond two variables, the
difficulties of determining and specifying the relationship to the variables
may become impractical.  Fortunately, sensitivity to most variables can be
reasonably controlled.

     Demonstration of repeatability for a candidate transfer standard normally
requires testing for each condition that could or may affect it.  Typical

                                      5-7

-------
5.  ESTABLISHING AUTHORITY/Qualification

tests for common conditions are discussed below.  Again, intelligent judgement
                                                                               i
is required to determine what conditions to test and the extent of testing re-
quired to qualify the device or procedure.  For qualification of procedural
candidates such as BAKI or GPT, testing may be minimal, provided the user is
adequately trained, uses good laboratory technique, and uses a specific appara-
tus and set of supplies (see Appendices B, C, D,).  For commercially available
transfer standard devices, some or all of the testing may be carried out by
the manufacturer, thereby reducing the burden on the user.  In some cases it
may be possible to judiciously substitute design rationale for actual testing.
For example, a device whose power supply is designed to be highly regulated
electronically may not require specific line voltage tests.  However, such
situations should be viewed with considerable skepticism.

     The preceding discussion brings up the further question of whether candi-
date transfer standards must be tested individually or whether they can be
qualifed by type, model, or agency.  In the case of procedural candidates such
as BAKI and GPT, it seems clear that each user must qualify them in his own
laboratory/use situation, since these procedures have a number of potential
variables.  The procedure should be tested with respect to conditions to which
it might be sensitive, such as those discussed below, while attempting to hold
all other conditions  (source of chemicals, apparatus, laboratory technique,
operator, etc.) as constant and uniform as possible throughout the testing
process as well as during subsequent certification and use of the procedure.

     The units of commercially produced transfer standard devices are designed
and manufactured to be identical and should therefore have very similar charac-
teristics.  The manufacturer could carry out the necessary qualification tests
on representative samples, sparing the user the burden of testing each unit he
buys or the cost of paying the manufacturer to test each unit individually.
Under this concept, it would certainly be appropriate to require the manufac-
turer to guarantee that each unit meet appropriate performance specifications.
However, the user should assume a skeptical attitude, in view of manufacturing
                                     5-8

-------
5.  ESTABLISHING AUTHORITY/Qualification

tolerances and possible defective components, and carry out at least some
minimal tests to verify that each unit is acceptable.

     In the case of unique devices assembled by users, testing for all perti-
nent conditions which could or might affect the device are normally required.

QUALIFICATION TESTS

     Some of the more common conditions likely to be encountered or to change
while using transfer standards and that may often affect the repeatability of
the device or procedure are discussed below.  Also discussed are ways or
approaches to test for sensitivity to the condition.  As noted previously, the
exact conditions or variables that must be considered depend on the specific
nature of the device or procedure.  The user (or manufacturer, etc.)  should
determine the conditions for each case on an intelligent judgemental basis
derived from a complete understanding of the operation of the device or pro-
cedure and supported by appropriate rationale.  Specific recommendations for
several common transfer standards — BAKI, GPT with excess NO, GPT with excess
0 , 0  generators, and 0  analyzers — are given in Appendices B,  C, D,  E, and
 J   J                  O
F, respectively-

     Once the conditions to be considered have been determined, the objective
of the qualification tests is either a or b:

     (a) to demonstrate that the candidate transfer standard's output
         is not affected by more than ± 4% or ± 4 ppb (whichever is
         greater) by the condition over a range likely to be encountered
         during use of the device or procedure;
     (b) to demonstrate that the candidate transfer standard's output
         is repeatable within ± 4% or ± 4 ppb  (whichever is greater)
         as the variable is changed over a range likely to be encountered
         during use, and to quantify the relationship between the output
         and the variable.
                                     5-9

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests

Temperature

     Changes in ambient temperature are likely to occur from place  to place  and
from one time to another.  Temperature changes are very likely to affect  almost
all types of transfer standards unless appropriate means are used to avoid ad-
verse effects.  Temperature affects transfer standards in many ways:  changes
in the action of components, changes in chemical reactions or rates of  reaction,
volume changes of gases, electronic drift, variable warm-up time, etc.  The
most important effects may well be  (1) changes in the output of generation
devices,  (2) changes in the sensitivity of 0  assay systems, and  (3) changes
in the volume of air flows which must be measured accurately.

     Temperature effects can be minimized in several ways.  The simplest  way
might be to restrict the use of the transfer standard to a temperature  range
over which the effects are within the specification.  This restriction  may be
the only practical approach for some candidates, but it may also preclude use
of such a transfer standard in too many situations.  Transfer standard  devices
may be made insensitive to temperature changes by design, such as thermostatic
regulation of sensitive components or of the entire device, or by temperature
compensation.

     Temperature effects on air flow measurement can be minimized by the  use
of mass flowmeters, which do not measure volume, or by the regulation of  gas
temperatures.  In another approach, ordinary ideal-gas-law corrections  could
be made manually to adjust to measured volumetric flowrates.  However,  when
using orifice control or measurement devices such as critical orifices  and
rotameters, be sure to use an appropriate correction formula.
                               r\
     Testing a candidate transfer standard for sensitivity to temperature is
facilitated by the use of a controlled temperature chamber.  However, success-
ful temperature tests can be carried out in many ordinary laboratories  where
the temperature can be manually controlled by adjusting thermostats, blocking
air vents or outlets, opening doors or windows, or using supplemental heaters

                                     5-10

-------
5.  ESTABLISHING AUTHORITY/Qualification  Tests

or air conditioners.  A reasonable  temperature  range  would be  20  to  30°C (68
to 86°F).  Broader temperature  ranges  could  be  used if  appropriate.

     The candidate transfer  standard is tested  by  comparing its output  to  a
stable concentration reference.   This  reference should  ideally be a  UV  photo-
meter system as described in Appendix  A.   The reference could  also be another
transfer standard known to be very  repeatable and, in particular,  very  insensi-
tive to temperature changes.  Still, it would be best to locate the  reference
outside of the variable temperature test  area.   The candidate  transfer  standard
should be tested at several  different  points over  the temperature  range, in-
cluding the extremes, and at several different  concentrations.  Be sure  to
allow sufficient time for the device or any  instruments or equipment associated
with the transfer standard to equilibrate each  time the temperature  is  changed.
The test results should be plotted  in  a fashion similar to the example  shown
in Figure 5-3.

     If the candidate transfer  standard has  a significant  temperature depen-
dence, additional test points at  various  concentrations  and temperatures should
be taken to define the relationship between  output and  temperature accurately.
Furthermore, if the candidate turns out to have a  dependence on more than one
condition or variable, tests must be carried out over the  range of both vari-
ables simultaneously to determine any  interdependence between the two variables.
Once the test data are acquired,  they  should be analyzed to determine if some
general formula or curve can be derived  (either analytically or empirically)
to predict the correct 0  concentration at any  temperature in the range  (see
Figure 5-4).  The correction formula or curve must be accurate within ±  4% or
± 4 ppb, whichever is greater.  If  two or more  variables are involved,  a family
of curves may be required; unless the  relationship is rather simple, this  situa-
tion may prove impractical in actual use.
                                      5-11

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests







o.
Q.
U
z
o
o
o





Y
.55

.50 •
.45 t

.40 -
.35 -
t
.30


.25

.20 t
.15
.10 [
.05


SPECIFICATION

— +.018 •
i n n * %
J 	 LJ LJ
— .018

n n — +-013 + d <•/
— -.013




i n D — +-008 + 1 "
j u u_ OQ8 _4^

] D D- +'004 +4rrb
J LJ LJ 	 . U(J4 t HP"
(.II.....IIIIIIII1
20 22 24 26 28 30 32 X
                                 TEMPERATURE. DEC. C
 Figure  5-3.  Example of temperature qualification test results  showing no depen-
             dence on temperature.
 Line Voltage

     Line voltage is very likely to vary from place to place  and  from one time
 to another.  Good electrical or electronic design of the transfer standard
 should avoid sensitivity to line voltage variations, but poorly designed equip-
 ment can easily be affected.  In addition, line voltage sensitivity may appear
 only as long-time thermal drift, a rather subtle effect.
     Aside from adequate design, line voltage effects  can be minimized by the
addition of an outboard line voltage regulator.  However, such devices may dis-
tort the line voltage waveform, thereby adversely affecting some  types of
                                     5-12

-------
5.
ESTABLISHING AUTHORITY/Qualification  Tests
            20
            1
                      24
26     28     30
TEMPERATURE, DEG. C
32
Figure 5-4.  Example  of  a  temperature dependence  quantitatively defined as a
             correction  factor.
equipment.  If such regulators  are  used,  it  is  important  that the  same regulator
is used during both certification and  use of the  transfer standard.  Restriction
of the transfer standard  to a line  voltage range  in which the effects are  in-
significant is another alternative, but that would require monitoring the  vol-
tage during use and may preclude use at some sites.

     Testing for line voltage sensitivity can be  carried  out along the same
lines as described for temperature  testing.   The  line  voltage can  be varied by
means of a variable voltage transformer  ("Variac") and measured by an accurate
ac voltmeter.  Do not use electronic "dimmer" controls which operate on  a  de-
layed-conduction principle, as  such devices  cause drastic waveform distortion.
                                     5-13

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests

A line voltage range of 105 to 125 volts should adequately cover the vast
majority of line voltages available in the U.S.  If the transfer standard is
used when powered by a small power generator, it should be checked for fre-
quency dependence.

Barometric Pressure/Altitude

     Since 0  concentrations are gaseous in nature, all transfer standards will
probably have some basic or inherent sensitivity to change in barometric pres-
sure.  Unfortunately, it is rather difficult to minimize barometric pressure
effects by design.  Air pressures can be regulated mechanically against an
absolute reference, but most such schemes are not practical when working with
0  concentrations because of restrictions to inert materials such as glass or
Teflon.  In the case of BAKI and GPT, the effect is limited primarily to the
measurement of flowrates, which were discussed briefly under temperature effects
and are applicable to barometric pressure changes as well.  At a constant alti-
tude, normal day-to-day variation in barometric pressure is only a few percent.
If the use of the transfer standard can be restricted to altitudes within a
hundred meters of the certification altitude, it may be acceptable to neglect
the barometric effect entirely.  However, if the use of a transfer standard is
necessary at altitudes significantly different than the calibration altitude,
then pressure effects cannot generally be ignored.

     Although not readily preventable, pressure effects are likely to be re-
peatable.  As a result, barometric pressure may be the variable most likely
to be handled by the defined-relationship approach discussed previously in
connection with temperature effects.  The technique is very similar to the
technique used to determine a temperature relationship; hopefully, a unique
quantitative relationship will result, such as that illustrated in Figure 5-5.
Remember that in any work with CL concentrations at altitudes significantly
above sea level, the concentration units must be clearly understood.  The vol-
ume ratio concentration units must be clearly understood.  The volume ratio
concentration units (ppm, ppb, etc.) are independent of pressure, while density

                                      5-14

-------
5.  ESTABLISHING AUTHORITY/Qualification  Tests

units such as yg/m3 are related to pressure.   (However,  the  iug/m3 unit defined
and used by EPA is "corrected" to 1.01 kPa  (760 mm Hg) and 25°C and is there-
fore related to ppm by a constant.)






DC
O
II
Ik
z
O
d
(11
cc
IT
O
O






Y
1.35 •
1.30 -

1.25 -

1.20 -

1.15 •

1.10 •

1.05



1.00 •
.95 -
.90 •
.85


PARABOLIC CURVE
Y •> A + BX +• CXX
A -» 3.068399

B -> - .005191
C •» .000003
s.
\gi CORR. COEF. ^ .999912
^^.^
^^El '
^^--^^
— ~-~d
^Cj^^Hw_
^*^^^~~^*^.
"*~~~~~^~— _
- a



. . i i i • i * . * . i . i . . i •
560 600 640 680 720 760 800 X

     Figure 5-5.  Example of a defined barometric pressure dependence.

     Testing with respect to barometric pressure may be difficult.  The use of
a variable pressure chamber is the best approach, but few laboratories have
access to such facilities.  It is conceivable that various pressures could be
obtained in a manifold setup, but construction of such an apparatus is dif-
ficult and of questionable validity.  The use of a mobile laboratory vehicle
which can be driven to various altitudes to conduct tests may offer the most
feasible solution.  Some types of transfer standards may not require pressure
tests because their pressure sensitivity is well known.  For BAKI and GPT, the
                                     5-15

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests

flow-measurement problem constitutes the only pressure effect.  Some  assay-type
devices  (such as a UV analyzer) are clearly related directly to gas density,
where a  simple ideal-gas-law correction can be applied.  Pressure  tests  are
not needed for these types.  For commercially-produced devices, the manufacturer
would be expected to carry out the necessary qualification tests and  to  offer
the devices as type-approved, at least with respect to pressure effects.

     As  a final note of encouragement, automatic compensation for  barometric
pressure is rapidly becoming economically feasible for some types  of  0   trans-
fer standards by the incorporation of microprocessor technology.   At  least two
manufacturers have used this approach in commercially available instruments.

Elapsed  Time

     As  the elapsed time between certification and use increases,  the confi-
dence in the repeatability decreases.  As a result, periodic recertification
is needed.  Some types of 0  generation devices have a definite loss of  output
 (decay)  with time.  This decay is usually associated with use-time or on-time
rather than total elapsed time.  Since the decay rate tends to be  quantifiable,
it can be accomodated with the defined-relationship mechanism discussed  in
connection with temperature effects:  the transfer standard is equipped  with
an hours meter or another me-time and a series of tests over a sufficient time
period can then be used to determine the decay rate.  During use,  a correction
to the output is applied based on the number of hours of on-time since the last
certification.

     Another approach is to recertify such a transfer standard often enough so
that the error due to decay never exceeds the ± 4% or ± 4 ppb specification.

Variability

     The preciseness of the relationship between a transfer standard  and a
primary  0  standard is dependent on the variability of the transfer standard.

                                     5-16

-------
5.  ESTABLISHING AUTHORITY/Qualification  Tests

Variability reduces confidence in the  accuracy  of  a  certified  transfer stan-
dard.  A high degree of variability may be  cause for disqualifying a device
or procedure for use as a transfer standard, or for  selecting  one with lower
variability.  Although the certification  procedure in Section  6 includes a
test for variability, more extensive tests  for  variability may be necessary
to qualify a transfer standard because the  certification test  is for variabil-
ity in the slope of the certification  relationship and not for individual
point variability.  Furthermore, variability may be  due to changes in condi-
tions not encountered during certification.

     Many different types of transfer  standards may  have excessive variability
for a variety of reasons.  Qualification  variability testing is perhaps most
needed to test for the effect of a variety  of non-specific or  non-quantitative
variables that cannot be tested individually.  For example, qualification
variability tests for BAKI and GPT could  include the use of various operators,
various sources of chemicals and water, minor variations or substitutions of
apparatus and components, etc.  These  tests might  be conveniently combined
with tests for the relocation and operator  adjustments described below.  When-
ever increased variability can be assigned  to a specific cause, corrective
actions or restrictions can be and should be applied to reduce the variability.

     Qualification testing for individual-point variability, unlike the certi-
fication variability test, should be carried out on  a single-point basis.  A
series of at least 6 single-point comparisons should be made between the candi-
date transfer standard and a UV reference at each  of at least  two fixed concen-
trations — one low concentration  (less than 0.1 ppm) and one high concentration
(over 80% of the upper range limit).   These comparisons should be made over a
variety of conditions and situations and  over a number of days.  For each con-
centration, verify that all 0  concentration measurements determined by the
UV primary standard are very nearly equal.  Then calculate the average of the
6 (or more) concentrations indicated by the transfer standard, using the fol-
lowing equation:
                                     5-17

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests
                                        1?
                                  Ave = — 2, y-
where n = number of comparisons
     y. =0  concentration indicated by the transfer standard
     Determine the difference between each concentration indicated by the trans-
fer standard and the average concentration (y. - Ave).  Each difference must be
less than ± 5% of the average (for concentrations over 0.1 ppm) or less than
± 5 ppb  (for concentrations less than 0.1 ppm).

     For this test, the acceptable limits are ± 5% or ± 5 ppb rather than ± 4%
or ± 4 ppb, because the test is for general variability, which may derive from
a number of non-identifiable causes.  Under these circumstances slightly wider
limits than those allowed for the other qualification tests are acceptable.

     One technique that can reduce variability and improve accuracy is repeti-
tion and averaging.  For example, the variability of assay procedures can be
reduced by assaying each concentration several times and averaging the results.
Of course, if this technique is used, it becomes a necessary part of the trans-
fer standard procedure and must be carried out each time the transfer standard
is used and certified.

Relocation

     A transfer standard obviously needs to maintain repeatability after being
moved and possibly encountering mechanical shocks, jolts, and stress.  Any
electrical or thermal stress incident to turning the device or equipment on
and off frequently is also of concern, as is consideration of orientation or
set-up factors.
                                     5-18

-------
5.  ESTABLISHING AUTHORITY/Qualification  Tests

     Tests for these conditions, while perhaps  not particularly  quantitative,
should include actually moving the  candidate  device  or  equipment to different
locations and comparing the output  each time  it is returned.   Tests could also
include mild shock or drop tests, or  tests  for  any set-up  factors which can be
specifically identified, e.g., physical orientation,  removal of  covers, any
set-up variations.  Any cause-and-effeet-relationship discovered should be
investigated completely.  The tests may be  conveniently combined or included
with those discussed previously for variability.

Operator Adjustments

     Those transfer standard devices  whose  output is  to be related to an
operator adjustment  (such as an adjustable  0  generator) should  be tested
for repeatability with respect to the adjustment.  Mechanical  adjustments
might need to be tested for play, backlash, hysteresis, slippage, and resolution.
Other types of adjustments may require tests  for analogous aspects.  If possible,
specific tests should be used.  For example,  approaching a given  setting from
both above and below the setting might be appropriate for testing play or
hysteresis.  If specific tests cannot be  designed, then simple repeatability
tests at several different settings should  be carried out.

Maifunctions

     The usefulness of a transfer standard  is dependent on the degree of con-
fidence that can be put on its ability to reproduce  0  standards.  While any
device is subject to occasional malfunctions, frequent  malfunctions would cer-
tainly compromise the purpose of a  transfer standard.   Of particular concern
are non-obvious type malfunctions that can  cause a significant error of which
the operator is unaware.  While no  specific tests for malfunctions are normally
used, the tests described above for the other conditions need  to be repeated
periodically to check for non-obvious malfunctions.   After a malfunction has
been corrected, the transfer standard must  be recertified.
                                      5-19

-------
5.  ESTABLISHING AUTHORITY/Qualification Tests

Other Conditions

     Any other condition that might affect a candidate device or procedure
or that might cause change between the point of certification and the point
of use should be tested.

CERTIFICATION
     The accuracy of a transfer standard is established by (1) relating the
output to a primary 0  standard and (2) demonstrating that the repeatability
of the transfer standard is within the limits specified in Section 6.
     A good transfer standard is precisely repeatable.  Its accuracy, however,
is entirely relative and comes strictly by certification to a primary UV 0
standard.  Certification establishes a precise quantitative relationship be-
tween the output of the transfer standard and a series of primary 0  concentra-
tions.  The primary 0  concentrations must be obtained by means of the UV cali-
bration procedure reproduced in Appendix A.  Note that the BAKI procedure, which
is a temporary alternate to the UV procedure to provide monitoring agencies
with a transition period for implementing the new UV procedure, is not approved
for certifying transfer standards.

     After a transfer standard has been shown to meet the qualification require-
ments discussed earlier in this section, the transfer standard must be certi-
fied before it can be used.  The prescribed formal certification procedure and
specifications are set forth in Section 6, but are explained in more detail
below.  Refer to Section 6 while reading the explanations; the certification
procedure step numbers correspond.
Procedure for Certification

     5.1  Certification requires the averaging of 6 comparisons between the
transfer standard and a UV primary 0  standard system.  Each comparison must

                                     5-20

-------
ESTABLISHING AUTHORITY /Certifica ti on

cover the full range of 03  concentrations  and is  to  be  carried  out  on  a  different
dard to a primary standard, refer to  the portion  of  this  section  entitled
"Comparing Transfer Standards to an Ultraviolet Primary Ozone Standard".

     5.2  Each comparison must  consist  of  6  or more  individual  comparison points,
including 0 and  (90 ± 5)% of the upper  range limit of the transfer  standard.  The
other points must be approximately evenly  spaced  between  these  points.   For each
comparison, the  slope and intercept is  computed by a least squares  linear regres-
sion.  The result should be similar to  the example shown  in Figure  5-6.  Most
assay-type transfer standards will be linear and  the linear regression can be
calculated directly.  However,  for non-linear transfer  standards  or generation-
type transfer standards where the output is  related  to  a  control  setting or an
adjustable parameter, a preliminary calibration relationship such as shown in
Figure 5-7 is required.  Note that the  curve shown in Figure 5-7  may have a con-
siderable zero offset and may be nonlinear.   This preliminary calibration should
also include any necessary  correction formulas for defined variables.  A smooth
curve fitting the points in Figure 5-7  should be  drawn  or calculated.  There are
no specific requirements on the form, number of points, linearity,  or frequency
of repetition for this preliminary calibration.   However,  excessive inaccuracy
in this relationship will show  up as  variability  in  the certification comparison
and may cause failure of the certification specifications.  During  the certifica-
tion comparisons, the preliminary calibration relationship (see Figure 5-7) is
used to obtain the indicated 0  concentration used in the linear  regression cal-
culations of Figure 5-6.   (Note that  Figure  5-6 should  be linear  even though
Figure 5-7 is nonlinear.)

     5.3  When 6 comparisons as shown in Figure 5-6  have  been completed  compute
the average slope  (in) from  the  6 individual  slopes  (nO ,  and the  average inter-
cept (I) from the 6 individual  intercepts,  (I.).
     5.4  Compute the relative standard deviation  (s ) of the 6 slopes  (m.)
     he quantity S
the formulas given.
and the quantity  S   defined by Equation 4  for the  6  intercepts  (1.^)  using
                                      5-21

-------
5.  ESTABLISHING AUTHORITY/Certification
                              	1	,	
                                 .3      .4      .5
                              UV STANDARD 03 CONC., PPM
                                               PRIMARY UV STD.
                                               TRANSFER STD.
                                               SERIAL NO.
                                               DATE
                                               NAME


                                               -I	
.2
                            .6
 Figure  5-6.   Example of a comparison (regression slope and intercept) of a
              transfer standard to a primary 0  standard.

      5.5   Compare s  to the 3.7% specification, and compare s_ to the 1.5
                    m                                         -l
 specification.   If either of these specifications is exceeded, it indicates
 that the  transfer standard has too much variability, and corrective action
 must be taken to reduce the variability before the transfer standard may be
 certified.  (Excessive variability in the UV primary standard is possible, al-
 though  it is  much less likely.)

      5.6   If  the specifications are met, the certification relationship for
 the  transfer  standard consists of the average slope (m) and the average inter-
 cept (I)  and  can be plotted as shown in Figure 5-8.  Note that the qualification
 restrictions  must also be included with the certification relationship and
                                      5-22

-------
5.  ESTABLISHING AUTHORITY/Certification
    Ul
    cc
    DC
    UJ
    z
    LU
                                                             Barometric Pressure Correction:
                                                             l03l ' = |03l x KB/760
                                                             I03)' = Corrected 03 cone.
                                                             (03) » Indicated 03 cone.
                                                             8 - Barometric pressure (mm Hy)
                                                  PRIMARY UV STD.
                                                  TRANSFER STD.
                                                  SERIAL IMO.
                                                  DATE
                                                  NAME
                            .2      .3      .4       .5      .6
                            O3 CONC. INDICATED BY TRANS. STD., PPM
Figure  5-7.   Example of  a preliminary calibration relationship for an adjust-
              able 0  device.

should  be shown on a plot of the certification  relationship as illustrated  in
Figure  5-8.   When the transfer standard is subsequently used, the standard  0
concentration is calculated from Equation 5.

Transfer  Standards Having a Defined Dependence  on Some Variable

     Certification of transfer standards having a defined dependence on a
variable  such as barometric pressure or temperature is complicated somewhat by
the need  to  take the variable into account.  During certification, the variable
must be measured accurately,  ideally with the same measuring instrument that
will be used during subsequent utilization of the transfer standard.  A cor-
rection for  effect of the variable must then be included, via the preliminary
                                       5-23

-------
5.  ESTABLISHING AUTHORITY/Certification
               O, STD= -±- (Indicated 03 Cone. -X)
                                                                COMPARISON DATES:
                                                                  1	
                                                                  2. 	
                                                                  3. 	
                                                                  4. 	
                                                                  5.
                                                                  6..
                                                                Y=mX+I
                                                                 T=
                                                PRIMARY UV STD.
                                                TRANSFER STD.
                                                SERIAL NO.
                                                DATE
                                                NAME
                                 .3
.4
                                                .5
.6
                                UV STANDARD O3 CONG, PPM
Figure  5-8.   Example of a transfer standard certification  relationship (average
              of 6 comparisons).

calibration,  during the certification comparisons.  The  final certification
relationship  must also clearly identify the applicable calibration and correc-
tion associated with the transfer standard.

      If the effect of the variable is limited to the measurement of gaseous
flow rates, the appropriate corrections should be  applied  to the flowrates
during  the certification comparisons, and an ordinary comparison similar to
Figures 5-6 or 5-7 will result.  Where a simple linear dependence on a variable
exists, a suitable reference level should be defined  (e.g., 1.01 kPa  (760 mm
Hg)  for barometric pressure);  then the appropriate mathematical correction
(formula)  to  correct the output from the reference level to any other level
                                       5-24

-------
5.  ESTABLISHING AUTHORITY/Certification

within a reasonable range should be  specified.  An  inverse correction is
applied to the transfer standard output during establishment of the prelimi-
nary calibration  (see Figure 5-7) to normalize the  output to the reference
level.  The preliminary calibration  relationship is then plotted with the
normalized data.

     Other approaches can also be used.  For example, if the linearity of the
transfer standard is not affected by the variable,  then the certification re-
lationship can include a "correction factor" relationship as illustrated in
Figures 5-4 and 5-5  (pp 5-13 and 5-16).  Such a correction factor relationship
can be determined either by changing the variable during the preliminary cali-
brations or by calculation based on  data obtained during qualification tests.
Another technique is to use a preliminary calibration consisting of a family of
curves as illustrated by Figure 5-9.  Here, interpolation can be used to obtain
the indicated 0  concentration at values of the variable falling between the
lines.

     In any case, the technique used must be completely described in writing,
verified to make sure it is accurate, and clearly understood by all users of
the transfer standard.

Use

     After certification, when the transfer standard is used to reproduce 0
standards, the certification relationship  (such as  illustrated in Figure 5-8)
is used to determine the certified 0 concentration from the concentration in-
dicated by assay or variable setting.  In using transfer standards, it is good
practice to try to minimize any change in variables even though the transfer
standard may be insensitive to them.  This insensitivity is particularly true
of transfer standards where there are numerous variables that may not have been
included in qualification tests.  Special effort should be made to use the same
standards, reagents, apparatus, technique, etc. to  the greatest possible extent
during certification and use.  Of course, for any transfer standard which has

                                     5-25

-------
 5.   ESTABLISHING AUTHORITY/Certification
   1
   CO
   to
   Q
   111
   S
   u
   o
   z
   8
   o"
                                                       .6
                              UV STANDARD O3CONC., PPM
          Figure  5-9.  A "family" of preliminary  calibration  curves.

a defined relationship to some variable, that variable must be  accurately mea-
sured and the output of  the transfer standard must be corrected accordingly.

Traceability

     Ozone standards obtained from a transfer standard will always be  somewhat
less accurate than primary 03 standards because of the inevitable variability
in the certification process.  Consequently, it is good practice to always
certify a transfer standard directly against primary 0  standards obtained by
the UV calibration procedure.  Certification of a transfer standard against
another transfer standard  is discouraged, but could be appropriate for some
                                     5-26

-------
5.  ESTABLISHING AUTHORITY/Certification

limited purposes if the variability of both  transfer  standards  is very  low.
The use of more than one intermediate transfer  standard  should  be avoided.

RECERTIFICATION

     While the principle accuracy of a transfer standard is established during
certification, the confidence  in that accuracy  is maintained by continual re-
verification to demonstrate  stability.  The  objective is to show, to the
greatest extent possible, that the transfer  standard  did not change signifi-
cantly between certification and use.  However,  if the UV reference system is
located in a laboratory, the recertification process  may always take place under
nearly identical conditions  (temperature, line  voltage,  barometric pressure,
etc.).  Therefore, ocasional repetition of the  qualification tests discussed
earlier in this section is an  important and  indispensable supplement to re-
certification, though not specifically required by Section 6.

     The procedure and specifications for recertification are prescribed ex-
plicitly in Section 6, but are explained in  more detail  below.  Paragraph num-
bers refer to the corresponding paragraph of the recertification procedures in
Section 6.

Procedure for Recertification

     A certified transfer standard must be recertified at least twice per
calendar quarter to maintain continuous certification.   A transfer standard
which looses its certification may cause the loss of  ambient 0  measurements
made with ambient monitors that were calibrated with  the transfer standard.
Consequently, more frequent  recertification  schedule  will reduce the magnitude
and risk of any such loss.   More frequent recertification may also provide bet-
ter accuracy, particularly for transfer standards that show slow but steady
change  (drift) over long periods of time.
                                      5-21

-------
5.  ESTABLISHING AUTHORITY/Recertification

     6.1  The first step in the recertification procedure is to carry out a
comparison to a UV primary standard as  specified in step 5.2 (see also Figure
5-6, pg 5-22).

     6.3  To maintain certification, the  linear regression slope of the new com-
parison (m) must be within ± 5% of the  average slope of the current certifica-
tion relationship  (m)  (i.e., the average  slope of the last 6 comparisons).  Thus,
m must be within the interval  . 95 m S m < 1.05 m.   A convenient way to monitor
the performance of a transfer  standard  is to plot each new slope on a chart for-
mat such as shown in Figure 5-10.
                 SLOPE,M

                  1.06

                  1.04
   1.02
ui  1.00
CL
o
W   .98

    .96

    .94
                                         ffi+5%
                                      _T
                                                        fii-5%
                      0   20   40   60  80  100   120  140  160  180  200
                                         DAYS
Figure 5-10.  Example of a chart  showing  recertification slope data for a
              transfer standard.
     6.3  If the new slope is within the  ±  5%  specification,  then a new
average slope  (m) and a new average intercept  (I)  are calculated using the new
comparison and the 5 most recent previous comparisons.   Thus  m and I are run-
ning or moving averages always based on the 6  most recent comparisons.  The
new m and the new ± 5% limits can also be plotted  on the chart shown in Figure
5-10.
                                     5-28

-------
5.  ESTABLISHING AUTHORITY/Recertification

     6.4  New values  for the relative standard deviation of the slopes  (s )
                                                                          m
and the quantity  (s^  are calculated based on the new comparison and the 5
most recent previous  comparisons.   The formulas are given in step 5.4.  These
parameters can also be monitored with a chart format similar to Figure 5-11.
                 VARIABILITY
                   4.0
                    3.5
                    3.0
                    Z5

                    2.0
                    1.6
                cr
                01
                °-   1.0
                   0.5

                   0.0
                                                         sm max
•E
 max
                      0   20  40  60   80   100  120  140  160  180 200
                                         DAYS
Figure 5-11.  Example  of a chart showing recertification variability of slope
              and  intercept for a transfer standard.

     6.5  The new  s  and s_ must again meet the respective 3.7% and 1.5 speci-
                   m      I
fications given  in step 5.5.  If all specifications are met, then a new certi-
fication relationship  (based on the updated m and I) is established according
to step 5.6 and  illustrated by Figure 5-8 (pg 5-24).

     6.6  If a certified transfer standard fails to meet one of the recerti-
fication specifications, it looses its certification.  Recertification then
requires 6 new comparisons according to the entire certification procedure
starting at step 5.1.   This failure could be due to a malfunction, which
obviously should be  corrected before repeating the certification procedure.
(If a transfer standard has been repaired or serviced in a way which could
                                      5-29

-------
5.  ESTABLISHING AUTHORITY/Recertification

affect its output, the complete certification procedure must also be repeated.)
Another possible cause for failure of the recertification specifications might
be a change in the preliminary calibration  (Figure 5-7, pg 5-23), which should
then be re-established before the certification is repeated.

Cross Checks

     Frequently, an assay-type transfer standard is used with an 0  generator
for which a preliminary calibration such as shown in Figure 5-7 (pg 5-23) is
available (even though the 0  generator is not certified as a transfer standard),
In such a case, any large discrepancy between the two could serve as a warning
that the transfer standard may need recertification.

Recertification Tests

     Normally, the characteristics of a transfer standard would not be expected
to change profoundly or suddenly.  For example, a transfer standard that is
not initially sensitive to line voltage changes is not likely to become so
after a period of use.  Malfunctions are a major exception:  malfunctions in
line voltage regulation or temperature regulation, or other variable control
components can easily render a transfer standard sensitive to a variable at
any time.  Furthermore, the malfunction may not be obvious to the operator and
could go undetected for some time.  Even a recertification may not disclose
such a malfunction.  In consequence, some of the qualification tests described
                                                                  ,T~;?
earlier should be repeated on some periodic basis.  Such tests may be more
cursory than the original tests, but are nevertheless important.  Other tech-
niques include warning lights or operation indicators on components that are
critical to regulatory functions and that might otherwise provide no indication
of malfunction.  A user should always be somewhat skeptical that a transfer
standard is operating properly.
                                     5-30

-------
                                   SECTION 6
                  SPECIFICATIONS FOR OZONE TRANSFER STANDARDS
1.  DEFINITION

     An 0  transfer standard is a transportable device or apparatus together
with associated operational procedures and documentation that is capable of
either accurately reproducing 0  concentration standards or of producing accu-
rate assays of 0  concentrations that are quantitatively related to an authori-
tative master standard.
2.  APPARATUS
     An 0  transfer standard should include all basic equipment, materials,
and supplies  (but not necessarily incidental items) required to carry out its
function.
3.  DOCUMENTATION

     The following comprehensive documentation of an 0  transfer standard is
required:

     3.1  A complete listing and description of all equipment, materials,
and supplies necessary or incidental to the use of the transfer standard;

     3.2  A complete and detailed operational procedure for using the transfer
standard, including all operational steps, specifications, quality control
checks, etc.;
                                      6-1

-------
6.   SPECIFICATIONS/Documentation

     3.3  Test data, rationale, evidence, and other information indicating
that the transfer standard meets the qualification requirements given below;

     3.4  The current certification relationship information  (slope and in-
tercept) as described in step 5.6 and applicable to current use of the transfer
standard, together with any corrections or restrictions in the operating con-
ditions  (temperature, line voltage, barometric pressure, etc.); and

     3.5  A logbook including a complete chronological record of all certifica-
tion and recertification data, as described under "Certification" and "Recerti-
fication" above, as well as all 0  analyzer calibrations carried out with the
transfer standard.

4.  QUALIFICATION

     An 0  transfer standard must meet the general requirements for quali-
fication as described in Section 5; the transfer standard output should not
vary by more than ± 4% or ± 4 ppb  (whichever is greater) from its indicated
value over a stated range of any of the conditions to which it might be sensi-
tive.  Documentation of conformance to this requirement shall be provided as
                                f
required by step 3.3.

5.  CERTIFICATION

     Prior to use, an 0  transfer standard must be certified by establishing
a quantitative certification relationship between the transfer standard and the
primary 0  concentrations obtained by the UV calibration procedure as specified
in Appendix D of 40 CFR, Part 50 (EPA 1979).  The certification procedure fol-
lows:

     5.1  The certification relationship shall consist of the average of 6
individual comparisons of the transfer standard to the primary UV 0  standard
system.  Each comparison must be carried out on a different day.

                                     6-2

-------
6.  SPECIFICATIONS /Certification

     5.2  Each comparison  shall consist  of at least  6  comparison points  at
concentrations evenly  spaced  over the  concentration  range  of  the transfer
standard, including  0  and  (90 ± 5)% of the upper range limit.  For the 6 or
more comparison points of  each comparison, compute the slope  and intercept by
the least squares linear regression of the transfer  standard  output  (either a
generated 0^ concentration or a concentration assay) and the  UV primary  0
standards .

     5.3  For the 6  comparisons,  compute the  average slope (m) :
m
                                     1
                                   = —  2,   m-
                                       i=l
 and the  average intercept (I) :
                                        6
                                 I = -  y  I,                          (Eq. 2)
 where m.  and I.  are the individual slopes and intercepts,  respectively, of
        i       i
 each comparison  regression.

     5.4   Compute the relative standard deviation of the  6 slopes,  (s  ):
                                                                      (Eq. 3)
                       m         '               ~  *
 and the quantity S  defined in Equation 4 for the 6 intercepts:
                 Sj-^/i I   I  (I^ -± ( I  l.ri             (Eq.  4)
                                       6-3

-------
6.   SPECIFICATIONS/Certification
     5.5  The value of s  must be £ 3.7%, and S_ must be ^ 1.5.
                        m                      i

     5.6  If the transfer standard meets the specifications of 5.5, compute
the certification relationship as:

                Std. 0  cone. = - (Indicated 0  cone. - I)            (Eq. 5)
                      J         m             J
6.  RECERTIFICATION

     To maintain continuous certification, an 0  transfer standard must be
recertified as follows.  In general, 0_ transfer standards should be recerti-
fied at least twice per calendard quarter; a transfer standard which remains
at a fixed monitoring site may be recertified once per quarter if it is suf-
ficiently stable to avoid loss of certification over that time period  (see
step 6.6).

     6.1  At the time of recertification, carry out a comparison of the trans-
fer standard to the UV primary standard as prescribed in step 5.2.

     6.2  The slope of the new comparison must be within the interval m ± 0.05 m.

     6.3  If the transfer standard meets the specification in 6.2, compute a
new m and a new I as prescribed in step 5.3 using the 6 most recent comparisons
(running averages).

     6.4  Compute a new s  and s  as prescribed in step 5.4 using the 6 most
recent comparisons.
     6.5  If the new s  and s  meet the specifications given in 5.5 compute a
new certification relationship as prescribed in step 5.6 using the updated m
and I.
                                      6-4

-------
6.  SPECIFICATIONS/Certification

     6.6  If the transfer  standard fails any of the recertification specifica-
tions, it looses its  certification.   Recertification then requires  repeating
all the initial certification steps  (steps  5.1 — 5.6.)

7.  TRANSFER STANDARDS  HAVING FEWER THAN FIVE DISCRETE  OUTPUTS

     Generation-type  transfer standards having discrete or fixed outputs that
are too few in number to meet the requirements of step  5.2 may be certified as
follows:

     7.1  If the transfer  standard is to be used for calibration, it must in-
corporate an integral dilution system to provide capability for  at  least 6
calibration concentrations (including zero).   The complete transfer standard
system, including  the dilution system, should then be certified  as  specified
under  subsections  5 and 6  above.

     7.2  If the transfer  standard is to be used for purposes other than
calibration where  discrete outputs are acceptable (e.g.,  audits,  span checks),
the transfer standard may  be certified at each discrete output by substituting
single point comparison, d., for  the slopes m., and ignoring all steps pertain-
ing to intercepts. Calculate d (Equation 1), s  (Equation 3) and substitute
them for in and s  , respectively.   If s, is < 3.7%, compute the certified dis-
                m                      d
crete  output as
                   Std. 0   cone.  = —  (Indicated 0  cone.)             (Eq. 6)
                         3         d
Recertification is carried out similarly.
                                      6-5

-------
                                  REFERENCES

U.S. Environmental Protection Agency.  October 6, 1976.  Measurement of Photo-
     chemical Oxidants in the Atmosphere:  Calibration of Reference Methods.
     Federal Register, 41:44049.

U.S. Environmental Protection Agency.  February 8, 1979.  Amendments to Title
     40, Code of Federal Regulations, Part 50  (Appendix D), Measurement Princi-
     ple and Calibration Procedure for the Measurement of Ozone in the Atmo-
     sphere.  Federal Register, 44:8821.
                                      7-1

-------
                                  APPENDIX A

        ULTRAVIOLET PHOTOMETRIC PROCEDURE FOR PRIMARY OZONE STANDARDS*
CALIBRATION PROCEDURE
     1.  Principle.  The calibration procedure is based on the photometric
assay of ozone  (0 ) concentrations in a dynamic flow system.  The concentration
of 0  in an absorption cell is determined from a measurement of the amount of
254 nm light absorbed by the sample.  This determination requires knowledge
of (1) the absorption coefficient  (a) of 0  at 254 nm, (2) the optical path
length (£) through the sample,  (3) the transmittance of the sample at a wave-
length of 254 nm, and  (4) the temperature (T) and pressure (P) of the sample.
The transmittance is defined as the ratio I/I , where I is the intensity of
light which passes through the cell and is sensed by the detector when the cell
contains an 0_ sample and I  is the intensity of light which passes through
the cell and is sensed by the detector when the cell contains zero air.  It is
assumed that all conditions of the system, except for the contents of the ab-
sorption cell, are identical during measurement of I and I .  The quantities
defined above are related by the Beer-Lambert absorption law,

                                          I     -aC£                  ....
                          Transmittance = —  = e                      (1)
                                           o

where: a = absorption coefficient of 0  at 254 nm = 308 ± 4
           atm"1 cnf^t 0°C and 760 torr. U'2,3,4,5,6,7)
       C = 0  concentration in atmospheres
       H = optical path length in cm
*Extracted from the Code of Federal Regulations, Title 40, Part 50, Appendix D,
 as amended February 8, 1979  (Federal Register, 44:8221-8233).
                                      A-l

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Principle
     In practice, a stable 0  generator is used to produce 0  concentrations
over the required range.  Each 0  concentration is determined from the measure-
ment of the transmittance (I/I ) of the sample at 254 run with a photometer of
                              o
path length H and calculated from the equation,

                            c(atm) = -^ (In I/IQ)                    (2a>
or,
                           c(ppm) = - ij- (In I/IQ)                  (2b)
The calculated 0  concentrations must be corrected for 0  losses which may
                •5                                       «5
occur in the photometer and for the temperature and pressure of the sample.

     2.  Applicability.  This procedure is applicable to the calibration of
ambient air 0  analyzers, either directly or by means of a transfer standard
certified by this procedure.  Transfer standards must meet the requirements and
specifications set forth in Reference 8.
     3.  Apparatus.  A complete UV calibration system consists of an ozone
generator, an output port or manifold, a photometer, an appropriate source of
zero air, and other components as necessary.  The configuration must provide a
stable ozone concentration at the system output and allow the photometer to
accurately assay the output concentration to the precision specified for the
photometer (3.1).  Figure 1 shows a commonly used configuration and serves to
illustrate the calibration procedure which follows.  Other configurations may
require appropriate variations in the procedural steps.  All connections be-
tween components in the calibration system downstream of the 0  generator
should be of glass, Teflon, or other relatively inert material.  Additional
information regarding the assembly of a UV photometric calibration apparatus
is given in Reference 9.  For certification of transfer standards which provide
their own source of 0 , the transfer standard may replace the 0  generator and
possibly other components shown in Figure 1; see Reference 8 for guidance.
                                      A-2

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC  PROCEDURE/Apparatus

     3.1  UV photometer.   The photometer consists of a low-pressure  mercury
discharge lamp,  (optional)  collimation optics, an absorption cell, a detector,
and signal-processing  electronics,  as illustrated in Figure  1.   It must be
capable of measuring the transmittance,  I/I ,  at a wavelength of 254 nm with
sufficient precision such  that the  standard deviation of the concentration
measurements does  not  exceed the greater of 0.005 ppm or 3%  of  the concentra-
tion.  Because the low-pressure mercury lamp radiates at several wavelengths,
the photometer must incorporate suitable means to assure that no 0   is generat-
ed in the cell by  the  lamp, and that  at least 99.5% of the radiation sensed by
the detector is  254 nm radiation.   (This can be readily achieved by  prudent
selection of optical filter and detector response characteristics.)   The
length of the light path through the  absorption cell must be known with an
accuracy of at least 99.5%.  In addition, the cell and associated plumbing
must be designed to minimize loss of  0  from contact with cell  walls  and gas
handling components.   See  Reference 9 for additional information.

     3.2  Air flow controllers.  Devices capable of regulating  air flows as
necessary to meet  the  output stability and photometer precision requirements.

     3.3  Ozone  generator.   Device  capable of generating stable levels of 0
over the required  concentration range.

     3.4  Output manifold.   The output manifold should be constructed of glass,
Teflon, or other relatively inert material, and should be of sufficient dia-
meter  to insure  a  negligible pressure drop at the photometer connection and
other output ports. The system must  have a vent designed to insure  atmospheric
pressure in the  manifold and to prevent ambient air from entering the manifold.

     3.5  Two-way  valve.   Manual or automatic valve, or other means  to switch
the photometer flow between zero air  and the 0  concentration.
                                      A-3

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Apparatus

     3.6  Temperature indicator.  Accurate to ±1°C.

     3.7  Barometer or pressure indicator.  Accurate to ±2 torr.

     4.  Reagents.

     4.1   Zero air.  The zero air must be free of contaminants which would
 cause  a detectable response from the 0  analyzer, and it should be  free of
 NO,  C  H    and other species which react with 0 .  A procedure for generating
 suitable zero air is given in Reference 9.  As shown in Figure 1, the zero
 air  supplied to the photometer cell for the I  reference measurement must be
 derived from the same source as the zero air used for generation of the ozone
 concentration to be assayed  (I measurement).  When using the photometer to
 certify a  transfer standard having its own source of ozone, see Reference 8
 for  guidance on meeting this requirement.

     5.  Procedure.

     5.1   General operation.  The calibration photometer must be dedicated
 exclusively to use as a calibration standard.  It should always be  used with
 clean, filtered calibration gases, and never used for ambient air sampling.
 Consideration should be given to locating the calibration photometer in a
 clean  laboratory where it can be stationary, protected from physical shock,
 operated by a responsible analyst, and used as a common standard for all field
 calibrations via transfer standards.

     5.2  Preparation.  Proper operation of the photometer is of critical im-
 portance to the accuracy of this procedure.  The following steps will help
 to verify proper operation.  The steps are not necessarily required prior to
 each use of the photometer.  Upon initial operation of the photometer, these
 steps  should be carried out frequently, with all quantitative results or in-
 dications recorded in a chronological record either in tabular form or plot-
 ted  on a graphical chart.  As the performance and stability record  of the

                                      A-4

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure

photometer  is  established,  the frequency of these steps may be reduced consis-
tent with the  documented stability of the photometer.

     5.2.1   Instruction manual:   Carry out all set-up  and adjustment procedures
or  checks as described in the operation or instruction manual associated with
the photometer.

     5.2.2   System check:   Check the photometer system for integrity,  leaks,
cleanliness, proper flowrates, etc.   Service or replace filters and zero air
scrubbers or other consumable materials, as necessary.

     5.2.3   Linearity:  Verify that the photometer manufacturer has adequately
established that the linearity error of the photometer is less than 3%, or test
                                              »
the linearity  by dilution as follows:  Generate and assay an 0 concentration
near the upper range limit of the system (0.5 or 1.0 ppm) ,  then accurately di-
lute that concentration with zero air and reassay it.   Repeat at several dif-
ferent dilution ratios.  Compare the assay of the original concentration with
the assay of the diluted concentration divided by the  dilution ratio,  as follows:

                                  A  - A /R
                              E =        —  x 100%                   (3)
 where:   E  = linearity error, percent
         A  = assay of the original concentration
         A  = assay of the diluted concentration
          4Lt
         R  = dilution ratio = flow of original concentration divided by  the
              total flow

 The linearity error must be less than 5%.   Since the accuracy of the measured
 flowrates will affect the linearity error as measured this way,  the test is
 not necessarily conclusive.  Additional information on verifying linearity is
 contained in Reference 9.
                                       A-5

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure

     5.2.4  Intercomparison:  When possible, the photometer should be occasion-
ally intercompared, either directly or via transfer standards, with calibration
photometers used by other agencies or laboratories.

     5.2.5  Ozone losses:  Some portion of the 0  may be lost upon contact with
the photometer cell walls and gas handling components.  The magnitude of this
loss must be determined and used to correct the calculated 0. concentration.
This loss must not exceed 5%.  Some guidelines for quantitatively determining
this loss are discussed in Reference 9.

     5.3  Assay of 0  concentrations.

     5.3.1  Allow the photometer system to warm up and stabilize.

     5.3.2  Verify that the flowrate through the. photometer absorption cell,
F  , allows the cell to be flushed in a reasonably short period of time (2 liter/
min is a typical flow).  The precision of the measurements is inversely related
to the time required for flushing, since the photometer drift error increases
with time.

     5.3.3  Insure that the flowrate into the output manifold is at least
1  liter/min greater than the total flowrate required by the photometer and any
other flow demand connected to the manifold.
      5.3.4  Insure that the flowrate of zero air, F , is at least 1 liter/min
                                                   Z
greater than the flowrate required by the photometer.
     5.3.5  With zero air flowing in the output manifold, actuate the two-way
valve to allow the photometer to sample first the manifold zero air, then F  .
                                                                           Z
The two photometer readings must be equal  (1=1).

     NOTE:  In some commercially available photometers, the operation of the
two-way valve and various other operations in section 5.3 may be carried out
automatically by the photometer.
                                      A-6

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/ Procedure

     5.3.6  Adjust  the  0  generator to produce an 0  concentration as needed.

     5.3.7  Actuate the two-way valve to allow the photometer  to sample zero
air until the  absorption cell is thoroughly flushed and record the stable mea-
sured value of I  .
                o

     5.3.8  Actuate the two-way valve to allow the photometer  to sample the
ozone concentration until  the absorption cell is  thoroughly  flushed and record
the stable measured value  of I.

     5.3.9  Record  the  temperature and pressure of the  sample  in the photometer
absorption cell.   (See  Reference 9 for guidance.)
      5.3.10  Calculate the 0  concentration from equation 4.  An average of
 several  determinations will provide better precision.
                                      f '
 where:   [0  ]     = 0  concentration,  ppm
          j OUT    j                                             _i   _i
         a        = absorption coefficient of 0  at 254  nm =  308 atm   cm   at 0°C
                   and 760 torr
         &        = optical path length,  cm
         T        = sample temperature,  K
         P        = sample pressure,  torr
         L        = correction factor for 0  losses from 5.2.5  =  (1-fraction 03
                   lost).

      NOTE:   Some commercial photometers may automatically evaluate  all or part
 of  equation 4.   It is the operator's responsibility to verify that  all of the
 information required for equation 4 is obtained,  either automatically by the
 photometer  or manually.   For "automatic" photometers which evaluate the  first
                                       A-7

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure

term of equation 4 based on a linear approximation, a manual correction may
be required, particularly at higher (
manual and Reference 9 for guidance.
be required, particularly at higher 0  levels.  See the photometer instruction
     5.3.11  Obtain additional 0  concentration standards as necessary by re-
peating steps 5.3.6 to 5.3.10 or by Option 1.

     5.4  Certification of transfer standards.  A transfer standard is certi-
fied by relating the output of the transfer standard to one or more ozone stan-
dards as determined according to section 5.3.  The exact procedure varies de-
pending on the nature and design of the transfer standard.  Consult Reference
8 for guidance.

     5.5  Calibration of ozone analyzers.  Ozone analyzers are calibrated
as follows, using ozone standards obtained directly according to section 5.3
or by means of a certified transfer standard.

     5.5.1  Allow sufficient time for the 0  analyzer and the photometer or
transfer standard to warm up and stabilize.

     5.5.2  Allow the 0  analyzer to sample zero air until a stable response
is obtained and adjust the 0  analyzer's zero control.  Offsetting the analyzer's
zero adjustment to + 5% of scale is recommended to facilitate observing negative
zero drift.  Record the stable zero air response as "Z".
     5.5.3  Generate an 0  concentration standard of approximately 80% of the
desired upper range limit  (URL) of the 0  analyzer.  Allow the 0  analyzer to
sample this 0  concentration standard until a stable response is obtained.
     5.5.4  Adjust the 0  analyzer's span control to obtain a convenient re-
corder response as indicated below:
                                      A-8

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure
                                              [°3]OUT
               recorder response (% scale)  = (	— x 100)  + Z      (5)
                                                URL
where:  URL = upper range limit of the 0  analyzer,  ppm
        Z   = recorder response with zero air,  % scale

Record the 0  concentration and the corresponding analyzer response.   If  sub-
stantial adjustment of the span control is necessary,  recheck the  zero and
span adjustments  by repeating steps 5.5.2 to 5.5.4.

     5.5.5  Generate several other 0  concentration  standards (at  least 5 others
are recommended)  over the scale range of the 0   analyzer by adjusting  the 0
source or by Option 1.  For each 0  concentration standard,  record the 0  con-
                                   •3                                    -J
centration and  the corresponding analyzer response.
      5.5.6   Plot the 0  analyzer responses versus the corresponding 0  concen-
 trations  and draw the 0  analyzer's calibration curve or calculate the appro-
 priate  response factor.

      5.5.7  Option 1;  The various 0  concentrations required in  steps 5.3.11
 and  5.5.5 may be obtained by dilution of the 0  concentration generated in
 steps 5.3.6 and 5.5.3.  With this option, accurate flow measurements are re-
 quired.   The dynamic calibration system may be modified as shown in Figure  2
 to allow  for dilution air to be metered in downstream of the 0  generator.  A
 mixing  chamber between the 0  generator and the output manifold  is also re-
 quired.   The flowrate through the 0  generator (F ) and the dilution air flow-
 rate (F ) are measured with a reliable flow or volume standard traceable to
 NBS.  Each  0  concentration generated by dilution is calculated  from:
                                       A-9

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure

             i
where:   [°3]OUT = diluted 0  concentration, ppm
        F       = flowrate through the 03 generator,  liter/min
        F       = diluent air flowrate, liter/min
         D

REFERENCES FOR APPENDIX A

      1.  E.C.Y. Inn and Y. Tanaka, "Absorption Coefficient of Ozone  in  the
         Ultraviolet and Visible Regions".  J. Opt. Soc. Am., 43,870 (1953).

      2.  A.G. Hearn, "Absorption of Ozone in the Ultraviolet and Visible
         Regions of the Spectrum", Proc. Phys. Soc.  (London), 78,932 (1961).

      3.  W.B. DeMore and O. Raper, "Hartley Band Extinction Coefficients
         of  Ozone in the Gas Phase and in Liquid Nitrogen, Carbon Monoxide,
         and Argon", J. Phys. Chem., 68,412  (1964).

      4.  M.  Griggs, "Absorption Coefficients of Ozone in the Ultraviolet
         and Visible Regions", J. Chem. Phys., 49,857  (1968).

      5.  K.H. Becker, U. Schurath, and H. Seitz, "Ozone Olefin Reactions in
         the Gas Phase.  1.  Rate Constants and Activation Energies", Int'l. J.
         Chem. Kinetics, VI,725  (1974).

      6.  M.A.A. Clyne and J.A. Coxom. "Kinetic Studies of Oxy-Halogen Radical
         Systems", Proc. Roy. Soc., A303,207  (1968).

      7.  J-W. Simons, R.J. Paur, H.A. Webster, and E.J. Bair, "Ozone Ultra-
         violet Photolysis.  VI.  The Ultraviolet Spectrum", J. Chem. Phys.,
         59, 1203 (1973).

      8.  "Transfer Standards for Calibration of Ambient Air Monitoring  Analy-
         zers for Ozone", EPA publication available from EPA, Department E
         (MD-77), Research Triangle Park, North Carolina 27711.
                                     A-10

-------
APPENDIX A.  ULTRAVIOLET PHOTOMETRIC PROCEDURE/References

     9.  "Technical Assistance Document for the Calibration of Ambient Ozone
         Monitors", EPA publication available from EPA, Department E (MD-77),
         Research Triangle Park, North Carolina  27711.
                                      A-ll

-------
       O
                                       GENERATOR
                                                                            OUTPUT
                                                                           MANIFOLD
                                                                         VENT
                                                                 I  I
                                                      VENT
                                           I.'
                                                                                     EXTRA OUTLETS CAPPED
                                                                                       WHEN NOT IN USE
                                                                      TO INLET OF ANALYZER
                                                                       UNDER CALIBRATION
                                                               TWO-WAY
                                                                VALVE
r
M
to
UV PHOTOMETER
                              DETECTOR
                                                                                            OPTICS
                                                             ABSORPTION CELL
                                SIGNAL
                              PROCESSING
                             ELECTRONICS
                                                                  SOURCE
                                                                    O

FLOWMETcH
^

FLOW
CONTROLLER


PUMP
                                                                 EXHAUST
                                                                                     s
                                                                                     I
                                                                                     s
                                                                                     I
                                                                                     s
                                                                                     O
                                                                                     S
                                                                                     3
§
8
ti
W
                                                                                                                        a
                                                                                                                        s
                                                                                                                        H
                                                                                                                        rn
                                                                                                                        in
                    Figure 1.   Schematic diagram  of a typical UV photometric calibration system.

-------
                                                                               VENT
                                                                                                  I
                                                                                                  3
                                                                                                  I

                                                                                                  I
                                                                                                  i
                                                                                                  3
                                                                                                  c
                                                                                                  S
                                                                                                  to
Figure 2,  Schematic diagram of a typical UV photometric calibration system  (Option 1)

-------
                                  APPENDIX B

          CERTIFICATION OF THE BORIC ACID POTASSIUM IODIDE PROCEDURE
                             AS A TRANSFER STANARD
     Following the guidance in Sections 4 and 5 and the specifications in
Section 6, this appendix attempts to provide more specific instructions for
certifying the Boric Acid Potassium Iodide  (BAKI) procedure as a transfer
standard.  The recommended version of the BAKI procedure appears at the end
of this appendix.  Because of the extensive and intimate involvement of the
user with the procedure, the user-procedure combination must be certified,
rather than the procedure itself; each user or user-agency must certify the
procedure under its own set of circumstances.

PRELIMINARY REQUIREMENTS

     Perhaps the first and surely one of the most important requirements is
the skill of the analyst, who must have adequate training in chemical labora-
tory technique, be thoroughly familiar with the BAKI procedure, and follow
the procedure completely and accurately.  Some period of practice is often
necessary to achieve consistent results.  If problems are encountered, the
analyst should seek assistance.  The success of the method is quite dependent
on the skill and ability of the analyst.

     The apparatus should be assembled and checked for compliance with specifi-
cations, cleanliness, absence of leaks, and proper operation.  Once assembled,
the apparatus and components should be dedicated exclusively to use as a trans-
fer standard, and the same apparatus  (as far as possible) should be used when-
ever the transfer standard is used or certified.  To the greatest extent prac-
tical, reagents should be common (i.e., from the same container or lot) through-
out the qualification, certification, and use of the procedure as a transfer
standard.
                                     B-l

-------
APPENDIX B.  BAKI PROCEDURE/Preliminary Requirements

     An important requirement is a stable source of 0  to be assayed by  the
transfer standard procedure, also an attendant source of zero air for the  0
generator.  Since the concentration of 0  generated by most C>3 generators
varies with flow, some means of reasonable flow regulation for the zero  air
is usually needed.  Some helpful guidance for these items may be found in
paragraphs 3.3, 3.4, and 4.1 of Reference 3.  Decide whether or not the  0
generation/zero air system is to be an integral part of the transfer standard
 (see pg 4-2).  If not, arrangements must be made for an adequate 0  generator/
zero air supply at each site where the transfer standard will be used.

     The 0  generator will also need an output manifold which meets the  re-
quirements of paragraph 3.4 of Appendix A.  The manifold may be as simple  as
a TEE where one of the legs serves as a vent.

     Access to a UV calibration system as described in Appendix A is required
for certification of the transfer standard and is also recommended for the
qualification tests.  Comparing the output  (indicated concentration) of  the
BAKI procedure to the UV primary 0  standard is easy, as the procedure simply
assays the output manifold of the UV primary 0  standard.

     Review any operation information or instructions provided by the manu-
facturer of the spectrophotometer to become familiar with its operation.

     Review the documentation requirements specified in paragraph 3 of Section
6 and complete item 3.1.

QUALIFICATION

     The next step is to qualify the procedure by demonstrating that it  is
repeatable to within the specifications in paragraph 4 of Section 6 (± 4%
or ± 4 ppb, whichever is greater).  The variables likely to affect the
BAKI procedure are similar to those discussed generally in Section 5 and are
discussed more specifically below.  Refer to Section 5 for additional guidance

                                     B-2

-------
APPENDIX B.  BAKI PROCEDURE/Qualification

on each of the qualification tests.   A preliminary calibration relationship
(as shown in Figure  5-7, pg 5-23)  is  not necessary for the  BAKI procedure,
as it is linear  and  provides a direct output indication of  concentration.  To
minimize potential sources  of variability,  the apparatus, reagents,  supplies,
techniques, procedures,  and conditions should be kept as constant  as possible
from qualification to  certification to use  of the procedure.   Therefore, these
items should be  identified  in detail  as part of the documentation  requirements
of paragraph 3 of Section 6.

QUALIFICATION TESTS

Temperature

                                             /
     Over normal ranges  of  room temperatures,  effects of temperature change on
the BAKI procedure should be minimal.  The  most important effect is  on the
measurement of the air volume sampled.  As  noted in step 4.5.4.1,  this air
sample volume must be  corrected to reference conditions (25°C  and  760 torr).
This correction  depends  on  the type of flowmeter used.   For a  flowmeter which
measures actual  volume (wet test meter, dry gas meter,  bubble  meter, etc.)
the volume measurement is corrected by using the ideal gas  law.  Wet volume
meters  (wet test meters, bubble meter) may  require water vapor corrections.
Orifice type flowmeters  (rotameters,  calibrated hypodermic  needles)  require
an orifice correction.  Flowmeters which measure mass flow  should  require no
correction.   (All flowmeters, of course, require accurate calibration.)  If
you are unsure of the  volume measurement correction, consult the flowmeter
instruction manual or  other authoritative reference.

     Temperature may affect the spectrophotometer slightly  but probably not
enough to be significant over reasonable room temperatures. The flow control-
lers and 0  generator  are likely to be affected but any such effect  is accounted
for in the procedure.
                                      B-3

-------
APPENDIX B.  BAKI PROCEDURE/Qualification Tests

Line Voltage

     Any line-powered components, such as the spectrophotometer or flowmeter,
could be affected by changes in line voltage.  Well-designed equipment will
show little or no sensitivity to voltage changes over a range of 105 to 125.
Look for voltage regulating capability and voltage sensitivity specifications
associated with the equipment.  Or, use external line voltage regulators.  If
there is any doubt about line voltage sensitivity effects, then line voltage
tests should be carried out as described in Section 5.

Barometric Pressure/Altitude

     The only effect of normal changes in pressure on the BAKI procedure will
be in measurement of the sample volume flowrate.  (The 0  generator will likely
be affected but, again, the procedure accounts for that.)  If use of the trans-
fer standard can be restricted to altitudes within a range of about 100 meters
 (300 feet), the range of pressure variations is only about 2 or 3% (about 1%
for altitude variations and about 1 to 2% for normal barometric pressure
changes).  Under these conditions, the effect could be ignored.  However,
since the sampled volume must be corrected to standard conditions, the cor-
                                 t
rection for pressure should be included in that correction.  As with tempera-
ture effects, the pressure correction depends on the type of flowmeter used.
Consult the flowmeter manual or other reference to determine the proper cor-
rection formula.

Elapsed Time

     Normally, elapsed time has no significant effect on the BAKI procedure
as long as the apparatus is kept clean and in good working order, and the
reagents are prepared fresh as specified.  Flowmeters normally require periodic
recalibration,
                                     B-4

-------
APPENDIX B.  BAKI PROCEDURE'/Qualification Tests

Variability

     General variability  can  sometimes  be a problem with BAKI.  As  noted
previously, changes  in  the apparatus, reagents,  supplies, procedure, etc.,
should be avoided as much as  possible to minimize  potential variation from
those sources.  When a  component,  reagent,  or procedure  is changed, a recerti-
fication comparison  should be carried out to see if the  variability or the
certification relationship is affected.   For any significant  change which is
to be a variable or  option  (i.e.,  sometimes used,  sometimes not used), then
a variability test as described in Section 5 chould be carried out to see if
the BAKI procedure still  qualifies when the variable is  exercised.  In this
case, judgment  is needed  to decide whether the change would or might cause
additional variability  in the procedure.

     As an example,  consider  an agency  which wishes to qualify several analysts
and several sets of  apparatus.   Must each analyst  be tested,  qualified, and
certified with  each  set of apparatus?   For more than a very few analysts and
sets of apparatus, this becomes impractical.   In this case, the agency should
insure that each analyst  is adequately  trained,  thoroughly familiar with the
procedure, and  can make accurate temperature and pressure corrections to the
measured flowrates.  Then, if each analyst is qualified  with  at least one set
of apparatus, and each  apparatus is qualified with at least one analyst, ade-
quate results can reasonably  be expected from any  analyst-apparatus combination.

Relocation

     Relocation tests help to establish the ruggedness of the BAKI  apparatus.
Move the equipment about  as it might be moved during subsequent use to see if
any malfunctions, leaks,  variability, or other problems  are observed which
might make use  of the procedure questionable or inconvenient.  (Make sure that
the wavelength  and other  adjustments on the spectrophotometer can be locked  so
they don't change when  it is  moved.)  If possible, include relocation in the
variability tests to measure  the effect of relocation on variability.

                                     B-5

-------
APPENDIX B.  BAKI PROCEDURE/Qualification Tests

Malfunctions

     There is no special test for malfunctions.  During other tests, be
observant for any leaks or malfunctions that occur or other characteristic
weaknesses in the apparatus or equipment which could cause unreliability.
Understand the design and operation of the components so that you can be alert
to nonobvious types of malfunctions.

     In summary, the primary qualification tests for most users of the BAKI
procedure are variability and relocation tests.

CERTIFICATION

     When the BAKI procedure has been shown to meet the qualification require-
ments, certify the procedure as specified in Section 6.  Additional guidance
is contained under "Certification" in Section 5.  If any volume or flow correc-
tion formulas are needed, use them during the certification procedure and
clearly specify the conditions at the time of the certification relationship.
Be sure that the specified correction formulas are accurate.  Note any special
operating instructions, operating restrictions, limits, or other pertinent in-
formation.  It is very important to complete the documentation requirements
specified in paragraph 3 of Section 6.
USE
     In using the BAKI procedure as a transfer standard, review the guidance
under "Use" in Section 5.  Recertify the transfer standard as required by
Section 6  (with further explanation of recertification in Section 5).  Consider
occasional cross checks of the transfer standard to other 0  standards, and
occasionally repeat the qualification tests to be sure the transfer standard
is maintaining adequate reliability.
                                      B-6

-------
APPENDIX B.  BAKI PROCEDURE'/Recommended Version

RECOMMENDED VERSION  OF BORIC ACID POTASSIUM IODIDE TRANSFER STANDARD PROCEDURE

     This procedure  can be used as a transfer standard for calibration of 0
analyzers in accordance with the guidance and specifications set forth in
Sections 5 and 6.

     1.  Principle.   This  procedure    is based upon the reaction between
ozone  (0 ) and potassium iodide (KI)  to release iodine (I )  according to  the
                        (2)
stoichiometric equation   :

                         03 + 2I~ 4- 2H+ -*• I2 + H20 + 02               (Eq. B-l)

The  stoichiometry is such  that the amount of I  released is  equivalent to the
                                              t
amount of 0  absorbed.  Ozone is absorbed in a 0.1 M boric acid  (H,BO ) solution
containing 1% KI, and the  I  released reacts with excess iodide  ion  (I  )  to form
                            £t
triiodide ion  (I-) which is measured spectrophotometrically  at a wavelength of
352  nm.  The output  of a stable 0  generator is assayed in this  manner, and the
generator is immediately used to calibrate the 0  analyzer.   The 0   generator
must be used immediately after calibration and without physical  movement, and
it is recalibrated prior to each use.  Alternatively,  the 0   analyzer may be
calibrated by assaying the 0  concentrations using the prescribed procedure
while simultaneously measuring the corresponding 0  analyzer responses.   Ozone
concentration standards may also be generated by an optional dilution technique.
With this option, the highest 0  concentration standard is assayed using  the
prescribed procedure.  The additional 0  concentration standards required are
then obtained by dilution.

     2.  Apparatus.   Figures B-l and B-2 illustrate a typical BAKI 03 transfer
standard system  and  show the suggested configuration of the  components listed
below.  All connections between components downstream of the C>3  generator
should be of glass,  Teflon, or other relatively inert material.
                                      B-7

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version — Apparatus

     2.1  Air flow controller.  Device capable of maintaining a constant air
flowrate through the 0  generator within ± 2%.

     2.2  Air flowmeter.  Calibrated flowmeter capable of measuring and moni-
toring the air flowrate through the 0  generator within ± 2%.

     2.3  Ozone generator.  Device capable of generating stable levels of 03
over the required concentration range.

     2.4  Output manifold.  The output manifold should be constructed of glass,
Teflon, or other relatively inert material and should be of sufficient diameter
to insure a negligible pressure drop at the analyzer connection.  The-system
must have a vent designed to insure atmospheric pressure in the manifold and
to prevent ambient air from entering the manifold.

     2.5  Impingers.  All glass impingers with the specifications indicated
in Figure B-2 are recommended.  The impingers may be purchased from most major
glassware suppliers.  Two impingers connected in series are used to insure
complete collection of the sample.

     2.6  Air pump and flow controller.  Any pump and flow control device capable
of maintaining a constant flowrate of 0.4 to 0.6 1/min through the impingers may
be used.  A critical orifice as described by Lodge et al.    is recommended.
The orifice should be protected against moisture and particulate matter with a
membrane filter or moisture trap containing Drierite, silica gel, or glass wool.
The air pump must be capable of maintaining a pressure differential of at least
0.6 to 0.7 atmospheres across the critical orifice.  Alternatively, a needle
valve could be used with the pump to adjust the flow through the impingers.  A
flowmeter is then recommended to monitor the flow.  The needle valve-flowmeter
combination should be protected against moisture and particulate matter with a
membrane filter or moisture trap.
                                      B-8

-------
APPENDIX B.  BAKI PROCEDURE'/Recommended Version — Apparatus

     2.7  Thermometer.  Accurate  to  ± 1%°C.

     2.8  Barometer.  Accurate  to ±  2 torr.

     2.9  Volumetric  flasks  (Class A).   25,  100, 200,  1000-ml.

     2.10  Pipets  (Class  A).   1,  5,  10, 15,  20, and 25-ml volumetric;  1-ml or
10-ml graduated.

     2.11  Spectrophotometer.   Capable of measuring absorbance  at  352  nm with
an absolute accuracy  of ± 1%  and  linear reponse over the  range  of  0 to 1.0 ab-
sorbance units.  The  photometric  accuracy may be checked  using  optical glass fil-
ters which have  certified absorbance values  at specified  wavelengths.   Matched
1-cm or 2-cm cells  should be  used for all absorbance determinations.

     3.  Reagents.

     3.1  Zero air.   The  zero air must be free of contaminants  which will cause
a detectable response on  the  0  analyzer or  which might react with 1%  BAKI.
Air meeting this requirement  may  be  obtained by:  (1)  passing it through silica
gel for drying;  (2) treating  it with 0  to convert any nitric oxide  (NO) to
nitrogen dioxide (NO  );  (3) passing  it through activated  charcoal  (6-14 mesh)
                    £
and molecular sieve  (6-16 mesh, type 4A)  to  remove any NO2,  hydrocarbons, and
traces of water  vapor; and (4)  passing it through a 2-y filter  to  remove any
particulate matter.

     3.2  Boric  acid  (H BO3).   ACS reagent grade.

     3.3  Potassium iodide (KI).   ACS reagent grade.

     3.4  Hydrogen peroxide  (HO).   ACS reagent grade, 3 or 30%.
                               2, £t

     3.5  Potassium iodate (KIO ).  ACS reagent grade, certified  0.1  N.

                                       B-9

-------
APPENDIX B.  BAKI PROCEDURE/REcommended Version - Reagents

     3.6  Sulfur acid  (H SO ).  ACS reagent grade, 95 to 98%.

     3.7  Distilled water.  Used for preparation of all reagents.

     3.8  Absorbing reagent.  Dissolve 6.2 g of boric acid  (H B0_) in approxi-
mately 750 ml of distilled water in an amber 1000-ml volumetric flask.  The
flask may be heated gently to speed dissolution of the H BO  , but the solution
must then be cooled to room temperature or below before proceeding with the
reagent preparation.   (While the H BO  solution is cooling, prepare the hydro-
gen peroxide  (HO) solution according to the directions in step 3.9.)  When
               ^£ ^
the H BO  solution has cooled, add 10 g of potassium iodide  (KI) to the H_BO
solution and dissolve.  Add 1 ml of 0.0021% HO  solution (see step 3.9) and
                                             & £t
mix thoroughly.  Within 5 minutes after adding the peroxide, dilute to volume
with distilled water, mix, and determine the absorbance of this BAKI solution
at 352 nm against distilled water as the reference.  The pH of the BAKI solution
should be 5.1 ± 0.2.
     Set the absorbing solution aside for 2 hours and then redetermine the
absorbance at 352 nm against distilled water as the reference.  If the resul-
tant absorbance from this second determination is at least 0.008 absorbance
units/cm greater than the first determination, the absorbing reagent is ready
for use.  If no increase or an increase of less than 0.008 absorbance units/cm
is observed, the KI reagent probably contains an excessive amount of a reducing
contaminant and must be discarded.  In this event, prepare fresh absorbing re-
agent using a different number lot of KI.  If an unacceptable absorbing reagent
results from different lots of KI, test the possibility of contamination in the
H.BO  by using a different numbered lot of H BO .
  33                                       «J  *5

     3.9  Hydrogen peroxide solution (0.0021%).  Using a graduated-pipet,
add 0.7 ml of 30% or 7.0 ml of 3% hydrogen peroxide  (HO ) to approximately
                                                      £ ^
200 ml of distilled water in a 500-ml volumetric flask, dilute to volume with
distilled water, and mix thoroughly.  To prepare the 0.0021% solution, pipet
5 ml of the above solution into 50 ml of distilled water in a 100-ml volumetric

                                     B-10

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version -Reagents

flask, dilute to volume with distilled water,  and mix thoroughly.   This  0.0021%
HO  solution must be prepared fresh each time a fresh batch of  absorbing
reagent is prepared.  Therefore,  the remaining contents of both  volumetric
flasks should be discarded  after  treatment of  the BAKI absorbing reagent (see
step 3.8).

     3.10 Standard potassium iodate  solution  (0.1 N) .   Use a commercial  standard
solution of potassium iodate (KIO )  having a certified normality.

     3.11  Sulfuric acid  (1 N).   Dilute 28 ml  of concentration  (95  to 98%)
sulfuric acid  (H SO ) to  volume in a 1000-ml volumetric flask.

4.  Procedure.
     4.1  Assemble  an  0   generation system such  as shown in Figure B-l.

     4.2  Assemble  the KI sampling train  such  as shown in Figure B-2.  All con-
nections between the various  components must be  leak tight and may be made using
grease-free ball joint fittings,  heat-shrinkable Teflon tubing or Teflon tube
fittings.  The  connection to  the  0  output manifold should be made using 6 mm
 (1/4 in.) Teflon tubing  not to exceed 1.5 meters in length.

     4.3  Calibrate all  flowmeters and critical  orifices under the conditions
of use against  a reliable flow or volume  standard such as an NBS traceable
bubble flowmeter or wet  test  meter.   Correct all volumetric flowrates to 25°C
and 760 torr as follows:
                                   P  - P
                                    S   HO
                         F  =  F
                               s                V73
                                                S
where:  F    = flowrate  corrected to reference conditions  (25°C  and  760 torr),
         R    n ,  .
               1/min
        F    = flowrate  at sampling conditions, 1/min
         O
        P    = barometric  pressure at sampling conditions,  torr
                                      B-ll

-------
APPENDIX B.  BAKI PROCEDURE'/Recommended Version — Procedure
         HO = vapor pressure of HO at T , torr, for wet volume
               standard  (for a dry standard, P  n = 0.)
                                               2
        T    = temperature at sampling conditions, °C
     4.4  Potassium iodide calibration curve.

     4.4.1  Prepare iodine standards, fresh when needed, as follows:

     4.4.1.1  Accurately pipet 10 ml of 0.1 N standard potassium iodate  (KlO  )
solution into a 100-ml volumetric flask containing approximately 50 ml of dis-
tilled water.  Add 1 g of potassium iodide (KI) and 5 ml of 1 N sulfuric acid
(H2SO ), dilute to volume with distilled water, and mix thoroughly.

     4.4.1.2  Immediately before use, pipet 10 ml of the iodine (I ) solution
                                                                  £*
prepared in step 4.4.1.1 above into a 100-ml volumetric flask and dilute to
volume with absorbing reagent.  Then further dilute this solution by pipetting
10 ml of it into a 200-ml volumetric flask and diluting it to volume with ab-
sorbing reagent.

     4.4.1.3  In turn, pipet 5, 10, 15, 20, and 25 ml aliquots of the final
I  solution prepared in step 4.4.1.2 above into a series of 25-ml volumetric
flasks.  Dilute each to volume with absorbing reagent and mix thoroughly.  To
prevent "L  losses by volatilization, the flasks should remain stoppered until
absorbance measurements are made.  Absorbance measurements (see step 4.4.2)
should be taken within 20 minutes after preparation of the I  standards.
                                                            •b

     4.4.2  Determine the absorbance of each I  standard at 352 nm.  Also mea-
sure the absorbance of a sample of unexposed absorbing reagent.  Determine the
net absorbance of each I  standard as

                                   /  sample  \    / unexposed \
                 net absorbance =  ^absorbancej -  I  reagent  j         (Eq. B-3)
                                                  ^absorbance /
                                      B-12

-------
APPENDIX B.  BAKI PROCEDURE '/Re commended Version - Procedure

     4.4.3  For each 1   standard,  calculate  the net absorbance/cm as
                      net absorbance/cm = net
where: b = s^ectrophotometer  cell path  length,  cm

     4.4.4  For each I2 standard, calculate  the I2  concentration  in mole/1 as

                    1 eq I2      -1 Inole  I2     10    10     10   V
     [I2]i:"= \10. X 1 eq KIO, X  2 eg  I,
                •j           j         2t
where:  eq = equivalent
or,
                            [I2]i = NKIO  x Vi x 10~                   (Eq' B~5b)

where:   [I ] . = concentration of each  I  standard, mole I /liter
          & 1                          £                 £
        N
         KIO  = normality of KIO   (from step 3.10), equivalent/liter
            •3                   ~3
        V.    = volume of I  solution  (from step 4.4.1.3) = 5, 10, 15, 20,
         1      or 25 ml
     4.4.5  Plot net absorbance/cm  (y-axis) versus the mole I~/l  (x-axis) for
each I  standard and draw the KI calibration curve.  Calculate the slope of the
                   —1   —1
curve in liter mole   cm   and record as S  .  The value of the slope should be
                                          c
26,000 ± 780.  If the slope is not within this range, and the photometric ac-
curacy of the spectrophotometer meets the specifications given in 2.11, repeat
the procedure using freshly prepared I_ standards.  If the slope is still not
within the specified range, repeat the procedure using a different lot of certi-
fied 0.1 N KIO  to prepare the I  standards.

     NOTE:  Preparation of a new KI calibration curve may not be necessary
each time the BAKI procedure is used, if adequate repeatability can be establi-
shed with less frequent KI calibration.
                                     B-13

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version — Procedure
     4.5  Calibration of the 0  generator.

     4.5.1  Adjust the air flow through the 0  generator to the desired  flow-
rate and record as F .  At all times the air flow through the generator  must
be greater than the total flow required by the sampling systems, to assure ex-
haust flow at the vent.
     4.5.2  With the 0  generator off, flush the system with zero air for at
least 15 minutes to remove residual 0  .  Pipet 10 ml of absorbing reagent into
each of 2 impingers and connect them into the sampling train as shown in Figure
B-2.  Draw air from the output manifold of the 0  calibration system through the
sampling train at 0.4 to 0.6 1/min for 10 minutes.  Immediately transfer the
exposed solutions to clean spectrophotometer cells.  Determine the net absor-
bance  (sample absorbance - unexposed reagent absorbance) of each solution at
352 nm within 3 minutes.  Add the net absorbances of the two solutions to ob-
tain the total net absorbance.  Calculate the indicated 0  concentration  (sys-
tem blank) as equivalent 0  concentration according to 4.5.4.  If the system
blank is greater than 0.005 ppm 0 , continue flushing the 0  generation system
for an additional 30 minutes and redetermine the system blank.  If the system
blank is still greater than 0.005 ppm 0 , the zero air probably contains traces
of an oxidizing contaminant, and the activated charcoal and molecular sieve
(see step 3.1) should be replaced.
     4.5.3  Adjust the 0  generator to generate an 0  concentration in the range
of interest and allow the system to equilibrate for about 15 minutes.  The un-
calibrated 0  analyzer to be calibrated can conveniently be used to indicate
the stability of the 0  generator output.  When the 0  generator output has
stabilized, pipet 10 ml of absorbing reagent into each impinger.  Draw 0  from
the output manifold of the 0  calibration system through the sampling train at
0.4 to 0.6 1/min.  Use a sample time of between 10 and 30 minutes such that
a total net absorbance between 0.1 and 1.0 absorbance units is obtained.   (At
an 0  concentration of 0.1 ppm and a sampling rate of 0.5 1/min, a total net
absorbance of 0.1 absorbance units should be obtained if a sampling time of

                                      B-14

-------
APPENDIX B.  BAKI PROCEDURE/Recommended  Version — Procedure

20 minutes and 1-cm spectrophotometer  cells  are used.)   Immediately after col-
lection, transfer the exposed  solutions  to clean  spectrophotometer cells.  De-
termine the net absorbance  (sample  absorbance  - unexposed reagent absorbance)
of each solution at 352 nm within 3 minutes.   Add the net absorbances of the
two solutions to obtain the  total net  absorbance.

     4.5.4  Calculation of indicated 0  concentration.

     4.5.4.1  Calculate the  total volume of  air sampled, corrected to reference
conditions of 25°C and 760 torr  as
                                 VR = FR x ts                         (Eq. B-6)

where:  V  = volume  of  air sampled,  corrected to  reference conditions, liter
        F  = sampling flowrate  corrected to  reference conditions, 1/min
         R
        t  = sampling time, min
         S'

      4.5.4.2   Calculate the I  released in moles  as
                              £t

                                total net absorbance  x Q.01            (Eq. B-7)
                      mole I2 =  	___	
                                          c

where:  total  net absorbance =  sum of net absorbances for the  two solutions
         0.01  = volume of absorbing reagent in each impinger,  liter
         S    = slope of KI calibration curve, liter mole    cm 1
          c
         b    = spectrophotometer cell path length, cm
      4.5.4.3  Calculate the yl of 0  absorbed as

                                  1 mole 0    24.47 1 0    106  yl  QS
                yl 00 = mole I  x T	-	— x 	-	r	 x —z-r	  (Eq. B-8a)
                    3         21 mole !„    mole 0        1  03
                                      B-15

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version — Procedure
or,
                        pi 0  = mole I_ x 24.47 x 106                 (Eq. B-8b)
                            j         £
     4.5.4.4  Calculate the indicated 0  concentration in ppm as
                                         pi 0
                                [°3]OUT ' —
     4.5.5  Repeat steps 4.5.3 and 4.5.4 at least one more time at the same
0  generator setting.  Average the two (or more) determinations and record
the average along with the 0  generator setting.

     4.5.6  Adjust the 0  generator to obtain other 0  concentrations over
the desired range.  Determine each indicated 0  concentration using the pro-
cedure given above.  Five or more 0  concentrations are recommended.  Plot the
indicated 0  concentrations versus the corresponding 0  generator settings and
draw the 0  generator calibration curve.

     4.6  Certification as a transfer standard.
     4.6.1  Carry out appropriate qualifications tests as discussed earlier
in this appendix.

     4.6.2  Certify the procedure against a primary 0  standard as described
in Sections 5 and 6 of this report.  Prepare a certification relationship curve
relating indicated 0  concentrations to the prima:
similar to Figure 5-8 (pg 5-24) of this document.
relating indicated 0  concentrations to the primary standard 0  concentrations,
                    J                                         J
     4.6.3  Recertify the procedure as appropriate in accordance with Sections
5 and 6 of this report.
     4.7  Calibration of the 0  analyzer.
                                     B-16

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version — Procedure

     4.7.1  Allow sufficient  time for the 0  analyzer to warm-up and stabilize.

     4.7.2  Allow the 0   analyzer to sample zero air until a stable  response
is obtained and adjust the  0   analyzer's zero control.   Offsetting the  analyzer's
zero adjustment to + 5% of  scale  is  recommended to facilitate observing negative
zero drift.  Record the stable zero  air response as "Z".
     4.7.3  Using the  03  generator,  the same F ,  the 0  generator  calibration
curve, and the certification relationship obtained in step  4.6.2,  generate a
certified 0  concentration near 80%  of the desired upper range  limit  (URL) of
the 0  analyzer.

     4.7.4  Allow the  0   analyzer to sample this  0  concentration  until a
                       •J                           O
stable response  is  obtained.  Adjust the analyzer's span control to obtain a
convenient recorder response as indicated below:

                                              [°3]CERT
              recorder response (% scale)  = (—rrr-	 x 100)  +  Z      (Eq. B-10)
                                                UKXj

where:   [0 1     =  certified 0., concentration at  the output manifold, ppm
          3 CERT              3
        URL      =  upper  range limit of the 0  analyzer,  ppm
        Z        =  recorder response with zero air, % scale

Record the certified 0 concentration and the 03  analyzer response.   If sub-
stantial adjustment of the span control is necessary, recheck the  zero and span
adjustments by repeating  steps 4.6.2 through 4.7.4.

     4.7.5  Generate several other 0  concentrations (at least  5 others are
recommended) over the  scale range of the 0  analyzer by adjusting  the 03 gen-
erator settings  (preferably the same settings as  used in step 4.5) or by Option
1.  For each 0   concentration, allow for a stable analyzer  response,  then record
the response and the corresponding certified 0  concentration from the certi-
fication relationship.
                                      B-17

-------
APPENDIX B.  BAKI PROCEDURE/Recotamended Version — Procedure
     4.7.6  Plot the 0  analyzer responses versus  the  corresponding certified
0  concentrations and draw the 0  analyzer's calibration  curve  or  calculate
the appropriate response factor.
     4.7.7  Option 1;  The various 0  concentrations required  in  step  4.7.5 may
be obtained by dilution of the certified 0  concentration generated  in 4.7.3.
With this option, accurate flow measurements are required.  The dynamic cali-
bration system must be modified as shown in Figure B-3 to allow for  dilution
air to be metered in downstream of the 0  generator.  A mixing chamber between
the 0  generator and the output manifold is also required.  The flowrate through
the 0  generator  (F ) and the dilution air flowrate  (F ) are measured  with a
reliable flow or volume standard traceable to NBS.  The highest 0 concentra-
tion standard required (80% URL) is assayed according to the procedure in step
4.5.  Each 0  concentration generated by dilution is calculated from:

                        [°3]CERT *  l°3]CERT m = diluted 0  concentration, ppm
          j CERT            3
        F        = flowrate through the 0  generator, 1/min
        F        = diluent air flowrate, 1/min

NOTE:  Direct calibration of the 0  analyzer may also be accomplished by ..^assay-
ing the 0  concentrations using the procedure in step 4.5 while simultaneously
measuring the corresponding 0  analyzer responses as specified in step 4.7.
                                      B-18

-------
APPENDIX B.  BAKI PROCEDURE/Recommended Version — References

References for Appendix B

1.   D.L. Flamm.  Analysis of Ozone at Low Concentrations  with  Boric Acid
     Buffered Potassium lodiate.   Environ.  Sci. Technol.,  10:978,  1977.

2.   B.E. Saltzman  and N. Gilbert.   lodometric Microdetermination  of Organic
     Oxidants and Ozone. Anal.  Chem., 31:1914, 1959.

3.   J.P. Lodge, Jr.,  J.B. Pate,  B.E.  Ammons,  and G.A.  Swanson.  The Use of
     Hypodermic Needles as Critical Orifices in Air Sampling.   J.  Air Pollut.
     Control Assoc.,  16:197,  1966.
                                       B-19

-------
                  AIR
	 ^
1

FLOW
CONTROLLER



FLOWMETER

FO

03
GENERATOR


CO
                                                           OUTPUT
                                                          MANIFOLD
                                    VENT
                                    EXTRA OUTLETS CAPPED
                                      WHEN NOT IN USE
                                                                       TO INLET OF ANALYZER
                                                                        UNDER CALIBRATION
                                                         TO INLET OF
                                                       Kl SAMPLING TRAIN
                                                                                                                   I
                                                                                                                   to
                        Figure B-l.  Schematic diagram of a typical BAKI calibration system.

-------
APPENDIX B.   BAKI  PROCEDURE/Recommended Version  - Figures
                      30ml

                      20ml

                      IQml
                                              (T
                                          30ml
• IMPINGERS
          "lOrn!

           10ml
                                                               JJ
                                                           MEMBRANE
                                                            FILTER
                                                                         HYPODERMIC
                                                                           NEEDLE
                                        RUBBER
                                        SEPTUM
                                                                                       TO AIR
                                                                                        PUMP
CRITICAL ORIFICE FLOW CONTROL
                      S mm I.D.*j  |—

<
(
I
I
S
/•







—*






\ 1
\

30
25
20
15
10
5
\J~-
     .J_1
     •T1
                                       10mm O.D.
                        t
                    INSIDE   '25mm1
                  CLEARANCE   Q0
                   3 TO S mm
                                     20/40 CONCENTRIC WITH
                                    'OUTER PIECE AND NOZZLE
                                     GRADUATIONS AT 5 ml
                                    -INTERVALS, ALL THE
                                     WAY AROUND
      NOZZLE 1.0. EXACTLY

      dm AT 12 in. HgO VACUUM'
      PIECES SHOULD BE INTER.
     * CHANGEABLE, MAINTAINING
      NOZZLE CENTERING AND
      CLEARANCE TO BOTTOM
      INSIDE SURFACE
                  ALL-GLASS MIDGET IMPINGER
                  (THIS IS A COMMERCIALLY STOCKED ITEM).
                                                                          NEEDLE VALVE
                                                                           FLOWMETER
                                 ALTERNATE FLOW CONTROL
                      Figure B-2.   Components of a  KI  sampling train.
                                                B-21

-------
                                                                                                                  1
                                                                                                                  to
to
ZERO /~^ ~
AIR W
1
-*.


FLOW
CONTROLLER

FLOW
CONTROLLER




CI niAlfUlCTCD
rLUWIVIt I tK

FLOWMETER

^•D
PQ


03
GENERATOR


k"
s
R
!
(6
8
r -^ 1
<^N ) a
MIXING §*
CHAMBER a-
                                                           OUTPUT
                                                          MANIFOLD
                                   VENT
EXTRA OUTLETS CAPPED       I
  WHEN NOT IN USE          f
                    TO INLET OF
                  Kl SAMPLING TRAIN
                                                                        TO INLET OF ANALYZER
                                                                         UNDER CALIBRATION
                                                                                                                   §
                  Figure  B-3.   Schematic  diagram of a  typical BAKI calibration system (Option 1)

-------
                                  APPENDIX C
 CERTIFICATION OF THE GAS PHASE TITRATION WITH EXCESS NITRIC OXIDE PROCEDURE
                            AS A TRANSFER STANDARD

     Following the guidance in Sections 4 and 5 and the specifications in
Section 6, this appendix attempts to provide more specific instructions for
certifying the gas phase titration with excess nitric oxide (GPT-NO) procedure
as a transfer standard.  The recommended version of the GPT-NO procedure appears
at the end of this appendix.  Because of the extensive and intimate involvement
of the user with the procedure and often with the assembly of the apparatus,
the user-procedure combination must be certified, rather than the procedure
itself.  Hence, each user or user-agency must certify the procedure under its
own set of circumstances.

PRELIMINARY REQUIREMENTS

     Laboratory tests show that the GPT-NO apparatus and equipment is of criti-
cal importance to the success of the GPT-NO procedure.  The GPT-NO system must
be assembled with great care and attention to the requirements and specifica-
tions in the procedure and by an analyst who thoroughly understands the pro-
cedure.  The assembled system, or a commercially obtained GPT system, should
be carefully and completely checked for absence of leaks, compliance with the
dynamic specification and other specifications, cleanliness, and proper opera-
tion.  Once assembled, the apparatus and components should be dedicated ex-
clusively to use as a transfer standard, and the same apparatus  (as far as
possible) and the same flowrates should be used whenever the transfer standard
is used or certified.  The NO cylinder standard used should be considered a
part of the transfer standard, and the procedure should be recertified when-
ever a new NO cylinder is needed.
                                     C-l

-------
APPENDIX C:  GPT-NO PROCEDURE/Preliminary Requirements

     Also important is the skill of the analyst, who must have adequate train-
ing in laboratory technique, be thoroughly familiar with the GPT-NO procedure,
and able to follow the procedure completely and accurately.  Some period of
practice is often necessary to achieve consistent results.  If problems are
encountered, the analyst should seek assistance.
     The procedure requires a stable source of 0  and an attendant source of
zero air for the 0  generator.  Since the concentration of 0  generated by
                  «J                                         -3
most 0  generators varies with flow, some means of reasonable flow regulation
for the zero air is usually needed.  Some helpful guidance for these items may
be found in paragraphs 3.3, 3.4, and 4.1 of Reference 2.
     The GPT-NO system will also need an output manifold which meets the re-
quirements of step 3.4 of Appendix A.  The manifold may be as simple as a TEE
where one of the legs serves as a vent.  The GPT-NO procedure requires an on-
site NO analyzer, but it need not be a part of the transfer standard.

     Access to a UV calibration system as described in Appendix A is required
for certification of the transfer standard and is also recommended for the
qualification tests.  Since the 0  generator is an integral part of the GPT-NO
system, compare the output (indicated concentration) of the GPT-NO procedure
to the UV primary 0, standard by using one of the three methods described in
Section 5.

     Although the NO cylinder standard is not required to be NBS-traceable,
use of an NBS-traceable NO standard provides a valuable advantage as a cross
check of the accuracy of the transfer standard.  A substantial discrepancy
between the indicated 0  concentration based on the traceable NO standard and
the UV 0  standard would indicate a serious problem with one of the systems
which must be resolved before the transfer standard can be relied upon.
                                    C-2

-------
APPENDIX C:  GPT-NO PROCEDURE/Preliminary  Requirements

     Review any operation information  or instructions provided by the manu-
facturer of the GPT system, the NO  analyzer,  and  other  components to become
familiar with their operation.

     Review the documentation requirements specified in paragraph 3 of Section
6 and complete item 3.1.

QUALIFICATION

     The next step is to qualify the procedure by demonstrating that it is
repeatable to within the specifications in paragraph 4  of Section 6 (± 4%
or ± 4 ppb, whichever is greater) .  The variables likely to affect the GPT-NO
procedure are similar to those discussed generally in Section 5 and are dis-
cussed more specifically below.  Refer to  Section 3 for additional guidance on
each of the qualification tests.  A preliminary calibration relationship (as
shown in Figure 5-7, pg 5-23 is not necessary for the GPT-NO procedure, as
it is linear and provides a direct  output  of  indicated  concentration.   To min-
imize potential sources of variability, the apparatus,  NO standard, flowrates,
techniques, procedures, and conditions should be  kept as constant as possible
from qualification to certification to use of the procedure.  Therefore, these
items should be identified in detail as part  of the documentation requirements
of paragraph 3 of Section 6.

QUALIFICATION TESTS

Temperature

     Over normal ranges of room temperatures, effects of temperature change
on the GPT-NO procedure should be minimal. The most important effect is on
the measurement of the gas volume flow rates. As required by the procedure,
gas flow rates must be corrected to reference conditions  (25°C and 760 torr).
This correction depends on the type of flowmeter  used.  For a flowmeter which
measures actual volume  (wet test meter, dry gas meter,  bubble meter, etc.),

                                      C-3

-------
APPENDIX C:  GPT-NO PROCEDURE/Qualification Tests

the volume measurement is corrected by using the ideal gas law.  Wet volume
meters  (wet test meters, bubble meters) may require water vapor corrections.
Orifice type flowmeters  (rotameters, calibrated hypodermic needles) require an
orifice correction.  Flowmeters which measure mass flow should require no cor-
rection.   (All flowmeters, of course, require accurate calibration.)  If you
are unsure of the volume measurement correction, consult the flowmeter instruc-
tion manual or other authoritative reference.

     The flow controllers and 0  generator may be affected by temperature
changes, but any such effect is accounted for in the procedure.

Line Voltage

     Any line-powered components, such as flowmeters could be affected by
changes in line voltage.  Well designed equipment will show little or no sensi-
tivity to voltage changes over a range of 105 to 125.  Look for voltage regula-
ting capability and voltage sensitivity specifications associated with the
equipment.  Or, use external line voltage regulators.  If there is any doubt
about line voltage sensitivity effects, then line voltage tests should be
carried out as described in Section 5.

Barometric Pressure/Altitude

     The only effect of normal changes in pressure on the GPT-NO procedure
will be in measurement of the gas volume flowrate.  (The 0  generator will
likely be affected but, again, the procedure accounts for that.)  If use of
the transfer standard can be restricted to altitudes within a range of about
100 meters (300 feet), the range of pressure variations is only about 2 or 3%
(about 1% of altitude variations and about 1 to 2% for normal barometric pres-
sure changes).  Under these conditions, the effect could be ignored.  However,
since the flowrates must be corrected to standard conditions, the correction
for pressure should be included in that correction.  As with temperature ef-
fects, the pressure correction depends on the type of flowmeter used.  Consult

                                     C-4

-------
APPENDIX C:  GPT-NO PROCEDURE'/Qualification Tests

the flowmeter manual or other  reference to determine the proper correction
formula.

Elapsed Time

     Normally, elapsed time  has  no significant  effect on the  GPT-NO procedure
as long as the apparatus  is  kept clean and in good working  order.  Flowmeters
normally require periodic recalibration.

Variability

     General variability  can sometimes be  a problem with GPT-NO.  As noted
previously, changes in the apparatus,  NO standard,  flowrates, procedure, etc.,
should be avoided  as much as possible  to minimize potential variation from
those sources.  When a component,  NO standard,  or procedure is  changed, a re-
certification comparison  should  be carried out  to see if the  variability or the
certification relationship is  affected.  For any  significant  change which is
to be a variable or option (i.e.,  sometimes used,  sometimes not used), then
a variability test as  described  in Section 5 should be carried out to see if
the GPT-NO procedure still qualifies when  the variable is exercised.  In this
case, judgment is  needed  to  decide whether the  change would or might cause
additional variability in the  procedure.

     As an example, consider an  agency wishing  to qualify several analysts
and several sets of GPT-NO apparatus.   Must each  analyst be tested, qualified
and certified with each set  of apparatus?   For  more than a very few analysts
and sets of apparatus, this  becomes impractical.   In this case, the agency
should insure that each analyst  is adequately trained, thoroughly familiar
with the procedure, and can  make accurate  temperature and pressure corrections
to the measured flowrates.   Then,  if each  analyst is qualified  with at least
one set of apparatus,  and each apparatus is qualified with  at least one analyst,
adequate results can reasonably  be expected from  any analyst-apparatus combi-
nation.

                                     C-5

-------
APPENDIX C:  GPT-NO PROCEDURE'/Qualification Tests

Relocation

     Relocation tests help to establish the ruggedness of the GPT-NO apparatus.
Move the equipment about as it might be moved during subsequent use to see if
any malfunctions, leaks, variability, or other problems are observed which might
make use of the procedure questionable or inconvenient.  If possible, include
relocation in the variability tests to measure the effect of relocation on vari-
ability.

Maifunctions

     There is no special test for malfunctions.  During other tests, be obser-
vant for any leaks or malfunctions that occur or other characteristic weakness
in the apparatus or equipment which could cause unreliability.  Understand the
design and operation of the components so that you can be alert to non-obvious
types of malfunctions.

     In summary, the primary qualification tests for most users of the BAKI
procedure are variability tests, relocation tests, and verification that each
operator can make the proper temperature, pressure, and other corrections to
                                 t
the flow measurement.

CERTIFICATION

     When the GPT-NO procedure has been shown to meet the qualification re-
quirements, certify the procedure as specified in Section 6.  Additional guid-
ance is contained under "Certification" in Section 5.  If any volume or flow
correction formulas are needed, use them during the certification procedure
and clearly specify the conditions at the time of the certification relation-
ship.  Be sure that the specified correction formulas are accurate.  Note any
special operating instructions, operating restrictions, limits, or other per-
tinent information.  It is very important to complete the documentation require-
ments specified in paragraph 3 of Section 6 .

                                      C-6

-------
APPENDIX C:  GPT-NO PROCEDURE/Use
USE
     In using the GPT-NO procedure  as  a transfer  standard,  review the guid-
ance under "Use" in Section  5.   Recertify the  transfer standard  as required
by Section 6  (with further explanation of recertification in  Section 5).  Con-
sider occasional cross  checks of the transfer  standard to other  0  standards,
and occasionally repeat the  qualification tests to  be  sure  the transfer stan-
dard is maintaining adequate reliability.

RECOMMENDED VERSION OF  GAS PHASE TITRATION WITH EXCESS NITRIC OXIDE TRANSFER
STANDARD PROCEDURE
     This procedure  can  be  used as  a transfer  standard  for calibration of 0
analyzers in accordance  with the guidance  and  specifications set forth in Sec-
tions 5 and 6.
      1.  Principle.
      1.1  The procedure  is  based upon the  rapid gas phase  reaction between
 nitric oxide  (NO)  and  ozone (0  )  as  described  by the  following equation   :

                              NO + 0  -> NO  +  0                      (Eq. C-l)

 When  0  is  added  to  excess  NO in a dynamic system, the  decrease  in NO concen-
 tration is  equivalent  to the concentration of  0  added.  The  NO  is obtained
 from  a standard NO cylinder and 0 is produced by a  stable 03 generator.  A
 chemiluminescence NO analyzer is used to measure the  change in NO concentration.
 The concentration of 0  added may be varied to obtain calibration concentra-
 tions over  the range desired.   The dynamic system is  designed to produce  local-
 ly high concentrations of NO and 0  in the reaction  chamber,  with subsequent
 dilution, to  insure  complete 0   reaction with  relatively small chamber volumes.
                                     C-7

-------
APPENDIX C:  GPT-NO PROCEDURE'/Recommended Version — Principle
     1.2  This procedure may be used either to calibrate an 0  analyzer direct-
ly, or  (more commonly) to calibrate an 0  generator which is used immediately
after calibration and without physical movement to calibrate the 0  analyzer.
This 0  generator must be recalibrated prior to each use.

     1.3  When this procedure is used directly to calibrate an 0  analyzer, the
0  concentration is first established by gas phase titration (GPT), then the
NO flow is diverted to allow the 0  concentration to pass to the output mani-
fold.
     2.  Apparatus.  Figure C-l, a schematic of a typical GPT-NO apparatus,
shows the suggested configuration of the components listed below.  All connec-
tions between components in the calibration system downstream from the 0  gene-
rator should be glass or Teflon.  Additional information regarding the assembly
of a GPT-NO calibration apparatus is given in Reference 2.

     2.1  Air flow controllers.  Devices capable of maintaining constant air
flow within ±2%.

     2.2  Nitric oxide flow controller.  A device capable of maintaining con-
stant NO flow within ± 2%.  Component parts in contact with the NO must be of
a non-reactive material.

     2.3  Air flowmeters.  Properly calibrated flowmeters capable of measuring
and monitoring air flows within ± 2%.

     2.4  Nitric oxide flowmeter.  A properly calibrated flowmeter capable of
measuring and monitoring NO flows within ± 2%.  (Rotameters have been reported
to operate unreliably when measuring low NO flows and are not recommended.)

     2.5  Pressure regulator for standard nitric oxide cylinder.  This regula-
tor must have non-reactive internal parts and a suitable delivery pressure.
                                     C-8

-------
APPENDIX C:  GPT-NO PROCEDURE'/Recommended Version — Apparatus

     2.6  Ozone generator.   Capable of generating stable levels  of 0   over
the range and flow rates  required (see step 4).

     2.7  Reaction chamber.  A glass chamber for the quantitative  reaction of
0  with excess NO.  The chamber should be of sufficient volume  (V   )  such that
 •3                                                               RC
the residence time  (t  ) is  as  specified in step  4.   For practical  reasons, t
                     K.                                                     J^
should be less than 2  minutes.

     2.8  Mixing  chamber.   A glass  chamber of proper design  to provide thorough
mixing of reaction products and diluent air.  The residence  time is not critical
when the dynamic  parameter  specifications given  in step 4 are met.

     2.9  Output  manifold.   The output manifold  should be constructed of glass
or Teflon of sufficient diameter to insure a minimum pressure drop at the ana-
lyzer connection.  The system  must  have a vent designed to insure  atmospheric
pressure at the manifold  and to prevent ambient  air from entering  the manifold.

     2.10  Chemiluminescence nitric oxide analyzer.   The NO  channel of any
chemiluminescence NO/NO/NO analyzer which essentially meets the  performance
                       <£•    X
requirements for  reference  methods  for NO  (Federal Register, 41:52693, Decem-
ber 1, 1976) may  be used; it is used only as a relative indicator  and is not
considered an integral part of the  transfer standard.

     3.  Reagents.

     3.1  Nitric  oxide concentration standard.   Compressed gas cylinder contain-
ing 50 to 100 ppm NO in N .  Although this NO standard need  not  be NBS-traceable,
                          ^
a useful cross check of the transfer standard's  accuracy is  obtained  if this
NO standard is traceable  to an NBS  NO Standard Reference Material  (SRM 1683 or
SRM 1684) or an NO  Standard Reference Material  (SRM 1629).  With  a traceable
NO standard, the  transfer standard's indicated 0^ concentration  should agree
with the UV standard within -  5% to + 15% (most  GPT-NO systems have a positive
bias).  If it does not, a problem with either the transfer standard or the UV
standard is indicated.
                                      C-9

-------
APPENDIX C:  GPT-NO PROCEDURE/Recommended Version — Reagents
     3.2  Zero air.  Air, free of contaminants which will cause a detectable
response on the NO or 0  analyzer or which might react with either NO or  0
in the gas phase titration.  A procedure for generating zero air is given in
Reference 2.
     4.  Dynamic parameter specifications.
     4.1  The residence time  (t ) in the reaction chamber and the gas flows
                               R
 (F  and F  )  (see Figure C-l) must be adjusted according to the following
relationships:
           PR =
                                   x tR > 3.75 ppm-min
                           [NO]RC =  [NO]
                                                "NO
                                                             (Eq. C-2)

                                                             (Eg. 03)
                           STD  F0 + PNO
                                      RC
                                   n + F
                                   0    NO
                                           < 2 min
                                                         (Eq. C-4)
where:
 R

[NO]

[NO]
F.
            RC
    = dynamic specification, determined empirically, to insure
      complete reaction of the available 0 ,  ppm-min
    = NO concentration in the reaction chamber, ppm
         NO
        V.
         RC
    = residence time in the reaction chamber, min
    = concentration of the undiluted NO standard, ppm
STD
    = NO flow, scmVmin
                                Q
    =0  generator air flow, scnr/min
    = volume of the reaction chamber, scm3
     4.2  These parameters may be selected according to the following sequence:
     4.2.1  Determine F , the total flow required at the output manifold
(F  = analyzer(s) demand plus 10 to 50% excess).
                                     C-10

-------
APPENDIX C:  GPT-NO PROCEDURE'/Recommended Version - Dynamic parameter

             sped fications





     4.2.2  Establish  [NO]QUT  as  the  highest NO concentration  (ppm) which will


be required at the output manifold.   [N°1OUT should be  approximately  equivalent


to 90% of the upper range limit (URL)  of the 0  concentration  range to be cov-


ered.
     4.2.3  Determine F    as
                       NO



                                     [NO]OUT X FT                     (Eq. 0-5)

                                NO -    [NO]STD




     4.2.4  Select  a convenient or available V  .   Intially,  a trial V   may be
                                               RC                     RC

 selected to be  in the range of approximately 200 to 500  scm3.
      4.2.5   Compute  F  as
                                 [NO]     x F   x V

                         T?  -        bJ-U    MU     L   -P
                         F0 ~ V - 3?75           FNO
      4.2.6   Compute t  as
                      R
                                          VRC                         (Eq. 07)
                                       F0 + FNO
 Verify  that t   > 2  minutes.   If not,  select a reaction chamber with a smaller
 V-
      4.2.7   Compute F  as






                               FD = FT - F0 - FNO





 where:   F  = diluent air flow, scm3/min





      4.2.8   If F  turns out to be impractical for the desired system,  select


 a reaction  chamber having a different VRC and recompute FQ and FD>   For a more
                                      Oil

-------
APPENDIX C:  GPT-NO PROCEDURE'/Recommended Version — Dynamic parameter
             sped ficati ons
detailed discussion of these requirements, and other related considerations as
well as example calculations, refer to Reference 2.

     5.  Procedure.

     5.1  Assemble a dynamic transfer standard system such as shown in Fig-
ure C-l.

     5.2  Establish the dynamic parameters as indicated in step 4.

     5.3  Insure that all flowmeters are properly calibrated under the condi-
tions of use against a reliable standard such as a soap-bubble meter or wet
test meter traceable to NBS.  All volumetric flowrates should be corrected
to 25°C and 760 torr.
     5.4  Precautions must be taken to remove 0  and other contaminants from
the NO pressure regulator and delivery system prior to the start of the pro-
cedure to avoid any conversion of the standard NO to NO .  Failure to do so
can cause significant errors.  This problem may be minimized by  (1) carefully
evacuating the regulator, when possible, after the regulator has been connected
to the cylinder and before opening the cylinder valve; (2) thoroughly flushing
the regulator and delivery system with NO after opening the cylinder valve; (3)
not removing the regulator from the cylinder between calibrations unless abso-
lutely necessary.  Further discussion of these procedures is given in Reference
2.
     5.5  Adjust the diluent air and 0  generator air flows to obtain the flows
determined in step 4.2.  The total air flow must exceed the total demand of the
analyzer(s) or UV certification photometer connected to the output manifold to
insure that no ambient air is pulled into the manifold vent.  Record F  , F ,
and F   used during certification and use approximately the same F_, F  , and
     NO                                                           0   D
F   during field use.
 NO

                                     C-12

-------
APPENDIX C:  GPT-NO PROCEDURE /Recommended Version - Procedure
     5.6  Calibration of  the NO analyzer.

     5.6.1  Allow sufficient time  for  the NO analyzer to warm up  and  stabilize.
     5.6.2  Allow the NO  analyzer  to sample  zero air until  a  stable NO response
is obtained, and make the proper zero  adjustments.
     5.6.3  Adjust F   to obtain an NO output concentration of approximately
80% of the upper range limit  (URL) of  the  NO range.   The exact concentration
is calculated from

                                           '     '
where:   [NO]    = diluted NO  concentration at the output manifold, ppm

Sample this NO concentration  until  the NO analyzer response has stabilized.
Adjust the NO span  control  to obtain  a convenient recorder response as indicat
ed below:
                   recorder  response  (% scale) = — — — — x 100       (Eq. C-10)
                                                   URL

where:  URL = upper range  limit  of the NO analyzer, ppm

Record the NO concentration  and  the analyzer response.  The NO analyzer should
be spanned to the same range as  that  of the 0  analyzer to be calibrated.  If
substantial adjustment of  the span control is necessary, it may be necessary to
recheck the zero and span  adjustments by repeating steps 5.6.2 and 5.6.3.

     5.6.4  Generate several additional NO concentrations  (at least five are
suggested to verify linearity) by decreasing F   or increasing FD>  For each
NO concentration, calculate  the  exact NO concentration generated using Equation
C-9 and record the analyzer  response.  Plot the analyzer response versus the cal-
culated NO concentration and draw the NO calibration  curve.  This plot should

                                      013

-------
APPENDIX C:  GPT-NO PROCEDURE/Recommended Version — Procedure

be linear.  For subsequent calibrations, this curve may be verified with a
two-point calibration.
     5.7  Calibration of the 0  generator.

     5.7.1  Adjust F  . F   . and F  as determined in step 4.2 and used in step
                    0   NO       D
5.5; generate an NO concentration near 90% of the URL of the NO range.  Using
the NO analyzer and calibration curve obtained in step 5.6.4, measure and re-
cord this NO concentration as  [NO]   . .
                                  ong

     5.7.2  Adjust the 0  generator to generate sufficient 0  to produce a
decrease in NO concentration equivalent to approximately 80% of the URL.  The
0  concentration must not exceed 90% of the available NO concentration.

     5.7.3  Calculate the indicated 0  concentration as
                  I03]OUT =  ([N°]orig

whs IT s *   [0  1
          3 OUT  = indicated 0  concentration at the output manifold when
                   FNQ = 0, ppm  ^
         [NO]   .  = original NO concentration, ppm
         [NO]     = NO concentration remaining after addition of 0 , ppm

Record  [0 ]    and the corresponding generator settings.

     5.7.4  Adjust the 0  generator settings to obtain other 0  concentrations
over the desired range, using Equation C-ll to calculate the indicated 0  con-
centrations (5 or more 0  concentrations are recommended.)  Plot the indicated
0  concentrations versus the corresponding 0  generator settings and draw the
 J                                          -J
0  generator calibration curve.  Record the 0  generator settings used during
certification, and use approximately the same setting during field use.  The 0.
                                      C-14

-------
APPENDIX C:  GPT-NO PROCEDURE/Recommended Version - Procedure

generator calibration curve  should replicate  fairly closely each  time  the trans-
fer standard is used; substantial  deviations  may indicate  a problem with the
transfer standard.

     5.8  Certification  as a transfer  standard.

     5.8.1  Carry out appropriate  qualification  tests  as discussed  earlier in
this appendix.

     5.8.2  Certify the  procedure  against a primary 0  standard as  described
in Sections 5 and 6.  Prepare a  certification relationship  curve relating in-
dicated 0  concentrations to the primary  standard 0 concentrations, similar
to Figure 5-8 on page 5-24.

     5.8.3  Recertify the procedure as appropriate in  accordance with Sections
5 and 6.
     5.9  Calibration of  the  0   analyzer.

     5.9.1  Allow  sufficient  time  for  the  0   analyzer to warm up and stabilize.

     5.9.2  Allow  the 0   analyzer  to sample  zero air until a stable response
is obtained and adjust the  0  analyzer's zero control.  Offsetting the analyzer's
zero adjustment to ± 5% of  scale is recommended to  facilitate observing negative
zero drift.  Record the stable  zero air response as "Z".
     5.9.3  Using the  0   generator as  calibrated  above,  generate a certified
concentration near  80% of the  URL of the  03  analyzer.  FQ, FD/ FNQ, and the C>3
generator setting should be  as close as possible  to  their values during certi-
fication.  Use the  0   generator calibration  curve and  the certification rela-
tionship curve from step 5.8.2 to establish  the certified concentration,  [°3JCERT-
                                     C-15

-------
APPENDIX C:  GPT-NO PROCEDURE'/Recommended Version -Procedure
     5.9.4  Allow the 0  analyzer to sample this 0  concentration until a
stable response is obtained.  Adjust the analyzer's span control to obtain
a convenient recorder response as indicated below:

                                                5 CFRT
              recorder response  (% scale) = (—„„    x 100) + Z      (Eq. C-12)
                                               URL

where:   [0 ]     = certified 0  concentration at the output manifold, ppm
          3 CERT              3
        URL      = upper range limit of the 0  analyzer, ppm
        Z        = recorder response with zero air, % scale

Record the certified 0  concentration and the analyzer response.  If substantial
adjustment of the span control is necessary, recheck the zero and span adjust-
ments by repeating steps 5.8.2 and 5.8.4.

     5.9.5  Generate several other 0  concentrations (at least 5 others are
recommended) over the scale range of the 0  analyzer by adjusting the 0  gen-
erator setting  (preferably the same settings as used in step 5.7.4).  For each
0  concentration, allow for a stable analyzer response, then record the response
and the corresponding certified 0  concentration from the certification rela-
tionship.
     5.9.6  Plot the 0  analyzer responses versus the corresponding certified
0  concentrations and draw the 0  analyzer's calibration curve or calculate the
appropriate response factor.

     OPTIONS:  If either of the optional procedures below are elected, the same
procedure must be used during qualification, certification, and field use.

     5.9.7  Option 1; The various 0  concentrations required in step 5.9.5
may be obtained by dilution of the certified 0  concentration generated in
step 5.9.3.  In this case, F  is increased to various values to decrease the
0  output concentration, which is calculated as
                                     C-16

-------
APPENDIX C:  GPT-NO PROCEDURE/Recommended Version - Procedure
                                        F  + F
                     [°3]OUT = t03]OUT (FQ + F?>

where:   t°3JouT = diluted °3 concentration, ppm
        F^      = the  new diluent air flow, scm3/min

Since only one 0  generator setting is used, the  generator  need be  calibrated
only at that setting.   Or,  [°.JOUT inay be obtained from Equation C-ll and the
certification relationship without actually calibrating the 0  generator, if
the setting is not  changed from that used in steps 5.7.2 and 5.7.3.  Because F
must be increased substantially to obtain low 0   concentrations, the original
F  should be as small  as  possible (consistent with step 5.5).

     5.9.8  Option  2;   The 0  analyzer may be' calibrated "directly" by GPT-NO
without intermediate calibration of the 0  generator and without changing F .
Under this option,  the various 0  concentrations  required for calibration in
steps 5.9.4 and 5.9.5  are obtained by adjusting the 0  generator settings.  For
each such adjustment,  the certified 0  concentration is first established by
GPT-NO using Equation  C-ll and the certification  relationship.  Then the NO flow
is diverted to allow the  0  to be delivered to the output manifold  and sampled
by the 0  analyzer.

References for Appendix C

1.   K.A. Rehme, B.E.  Martin, and J.A.  Hodgeson.   Tentative Method  for the
     Calibration of Nitric Oxide, Nitrogen Dioxide, and Ozone Analyzers by Gas
     Phase Titration.   EPA-R2-73-246, U.S. Environmental Protection Agency,
     Research Triangle Park, North Carolina.  March, 1974.   pp. 17

2.   E.G. Ellis.  Technical Assistance Document  for the Chemiluminescence
     Measurement of Nitrogen Dioxide.  EPA-600/4-75-003, U.S. Environmental
     Protection Agency, Research Triangle Park, North Carolina.
                                      C-17

-------
O

M
00



cH
ZERO
AIR
1




r
>













>•

FLOW — 3>
CONTROLLER FLOWMETtR






now PQ^ °3
1 LUW ... FLOWMETER .... -^. 	 £ \i v\
CONIROLLER "-uwiwtitit UtNtHAIOH (^ VHC J
REACTION
LIIAMbtK

A
|VNO

>l
d—
>s



NO
STD.






. 	 „, . _LJ|..|_UT-r-, \/ft| \«7
I VALVc.
(] 1 ~^ FLOW R OWMETFR . ( ,' )
\J CONIHULLtH -wwrirtt "C/ ^
.„ 	 	 	 .„„__ ,_ , - ..... - f

OUTPUT
MANIFOLD
VtNl ^—r^ ' |-
w ^ >|
§
13
O
s»
6
J/F° §
f t5!

X" ^"\
V* J
MIXING
CHAMBER






_.FT
^^^

1
8
§
a,
8
Wj
0)
t-i.
o
S3
i
a
EXTRA OUTLETS CAPPED TO INLET OF S
WHEN NOT IN USE ANALYZER $
UNDER
CALIBRATION
                             Figure C-l.  Schematic diagram of a typical GPT system.

-------
                                  APPENDIX D
          CERTIFICATION OF THE GAS PHASE TITRATION WITH EXCESS OZONE
                       PROCEDURE AS A TRANSFER STANDARD

     Following the guidance in Sections 4 and 5 and the specifications in Sec-
tion 6, this appendix attempts to provide more specific instructions for certi-
fying the gas phase titration with excess ozone (GPT-0 ) procedure as a transfer
standard.  The recommended version of the GPT-0  procedure appears at the end
of this-appendix.  Because of the extensive and intimate involvement of the
user with the procedure and often with the assembly of the apparatus, the user-
procedure combination must be certified, rather than the procedure itself.   Hence,
each user or user-agency must certify the procedure under its own set of circum-
stances.
PRELIMINARY REQUIREMENTS

     Laboratory tests show that the GPT-0  apparatus and equipment is of criti-
cal importance to the success of the GPT-0  procedure.  The GPT-0  system must
be assembled with great care and attention to the requirements and specifica-
tions in the procedure and by an analyst who thoroughly understands the proce-
dure.  The assembled system, or a commercially obtained GPT-0  system, should
be carefully and completely checked for absence of leaks, compliance with speci-
fications, cleanliness, and proper operation.  Once assembled, the apparatus
and components should be dedicated exclusively to use as a transfer standard,
and the same apparatus (as far as possible) should be used whenever the trans-
fer standard is used or certified.  The NO cylinder standard used should be
NBS-traceable and should be considered a part of the transfer standard; the
procedure should be recertified whenever a new NO cylinder is used.

     Also important is the skill of the analyst, who must have adequate train-
ing in laboratory technique, be thoroughly familiar with the GPT-03 procedure,

                                      D-l

-------
APPENDIX D.  GPT-0  PROCEDURE/Preliminary Requirements

and follow the procedure completely and accurately.  Some period of practice
is often necessary to achieve consistent results.  If problems are encountered,
the analyst should seek assistance.

     The procedure requires a stable source of 0  and an attendant source of
zero air for the 0  generator.  Since the concentration of 0  generated by most
0  generators varies with flow, some means of reasonable flow regulation for
the zero air is usually needed.  Some helpful guidance for these items may be
found in paragraphs 3.3, 3.4, and 4.1 of Reference 2.

     The GPT-0  system will also need an output manifold which meets the require-
ments of step 3.4 of Appendix A.  The manifold may be as simple as a TEE where
one of the legs serves as a vent.

     The dynamics of the procedure (various flowrates and residence time of the
reaction chamber) are somewhat critical to the procedure and must be optimized
to avoid both incomplete reaction of NO as well as secondary reaction of NO
with 0 .  Initial flowrates are established by theoretical calculations, but
the final values must be determined emperically for each system.  The flowrates
are optimized by measurement to detect residual NO and loss of NO , or alterna-
tively by carefully determining the GPT/UV 0  ratio.  In either case, thorough
understanding of the GPT reactions and careful measurements are necessary.

     Access to a UV calibration system as described in Appendix A is required
for certification of the transfer standard and is also recommended for the quali-
fication tests.  Since the 0  generator is an integral part of the GPT-0  sys-
tem, compare the output (indicated concentration) of the GPT-0  procedure to
the UV primary 0  standard by using one of the three methods described in Sec-
tion 5.
     Review any operation information or instructions provided by the manufac-
     : of the GPT-0  s;
with their operation.
turer of the GPT-0  system components and other components to become familiar
                                      D-2

-------
APPENDIX D.  GPT-03 PROCEDURE/Preliminary Requirements

     Review the documentation requirements specified in paragraph 3 of Section
6 and complete item 3.1.

QUALIFICATION

     The next step is  to  qualify the GPT-0  procedure by demonstrating that
it is repeatable  to within the specifications  in paragraph  4 of Section 6
 (± 4% or ± 4 ppb, whichever is greater) .   The  variables likely to affect the
procedure are similar  to  those discussed  generally  in Section 5 and are dis-
cussed more specifically  below.   Refer to Section 5 for additional  guidance
on each of the qualification tests.   A preliminary  calibration relationship
 (as shown in Figure 5-7,  pg 5-23 is  not necessary for the GPT-0  procedure,
as it is linear and provides a direct output of  indicated concentration.  To min-
imize potential sources of variability, the  apparatus,  NO standard,  techniques,
procedures, and conditions should be kept as constant as possible from quali-
fication to certification to use of  the procedure.   Therefore, these items
should be identified in detail as part of the  documentation requirements of
paragraph 3 of Section 6.

QUALIFICATION TESTS

Temperature

     Over normal  ranges of room temperatures,  effects of temperature change
on the GPT-0  procedure should be minimal.  The  most important effect  is on
the measurement of the gas volume flow rates.  As required by the procedure,
gas flow rates must be corrected to  reference  conditions (25°C and  760  torr).
This correction depends on the type  of flowmeter used.   For a flowmeter which
measures actual volume (wet test meter, dry  gas  meter,  bubble meter, etc.), the
volume measurement is  corrected by using  the ideal  gas law.  Wet  volume meters
 (wet test meters, bubble  meters)  may require water  vapor corrections.   Orifice
type flowmeters  (rotameters, calibrated hypodermic  needles) require an orifice^
correction.  Flowmeters which measure mass flow  should require no correction.

                                       D-3

-------
APPENDIX D.  GPT-0  PROCEDURE/Qualification Tests

 (All flowmeters, of course, require accurate calibration.)  If you are unsure
of the volume measurement correction, consult the flowmeter instruction manual
or other authoritative reference.

     The flow controllers and 0  generator may be affected by temperature
changes, but any such effect is accounted for in the procedure.

Line Voltage

     Any line-powered components, such as flowmeters, could be affected by
changes in line voltage.  Well designed equipment will show little or no sensi-
tivity to voltage changes over a range of 105 to 125.  Look for voltage regu-
lating capability and voltage sensitivity specifications associated with the
equipment.  Or, use external line voltage regulators.  If there is any doubt
about line voltage sensitivity effects, then line voltage tests should be car-
ried out as described in Section 5.

Barometric Pressure/Altitude

     The only effect of normal changes in pressure on the GPT-0  procedures
will be in measurement of the gas volume flowrates.   (The 0, generator will
likely be affected but, again, the procedure accounts for that.)  If use of
the transfer standard can be restricted to altitudes within a range of about
100 meters (300 feet), the range of pressure variations is only about 2 or 3%
 (about 1% for altitude variations and about 1 to 2% for normal barometric
pressure changes).  Under these conditions, the effect could be ignored.
However, since the flowrates must be corrected to standard conditions, the
correction for pressure should be included in that correction.  As with tempera-
ture effects, the pressure correction depends on the type of flowmeter used.
Consult the flowmeter manual or other reference to determine the proper cor-
rection formula.
                                     D-4

-------
APPENDIX D.  GPT-03 PROCEDURE/Qualification Tests

Elapsed Time

     Normally, elapsed  time  has  no significant effect on the  GPT-0  procedure
as long as the apparatus  is  kept clean and in good working  order.  Flowmeters
normally require periodic recalibration.

Variability

     General variability  can sometimes be  a problem with GPT-0  .  As noted
previously, changes in  the apparatus,  NO standard,  procedure, etc., should be
avoided as much as possible  to minimize potential variation from those sources.
When a component, NO  standard, or procedure is changed,  a recertification com-
parison should be carried out to see  if the variability  or  the  certification
relationship is affected. For any significant change which is  to be a variable
or option  (i.e., sometimes used, sometimes not used),  then  a variability test
as described in Section 5 should be carried out to see if the GPT-0  procedure
still qualifies when  the  variable is  exercised.   In this  case,  judgment is need-
ed to decide whether  the  change  would or might cause  additional variability in
the procedure.

     As an example, consider an  agency wishing to qualify several analysts
and several sets of GPT-0 apparatus.   Must each  analyst be tested, qualified
and certified with each set  of apparatus?   For more than a  very few analysts
and sets of apparatus,  this  process becomes impractical.  In this case, the
agency should insure  that each analyst is  adequately  trained, thoroughly
familiar with the procedure, and can  make  accurate temperature  and pressure
corrections to the measured  flowrates. Then,  if  each analyst is qualified
with at least one set of  apparatus, and each apparatus is qualified with at
least one analyst, adequate  results can be reasonably expected  from any analyst-
apparatus combination.
                                      D-5

-------
APPENDIX D.  GPT-0  PROCEDURE'/Qualification Tests

Relocation

     Relocation tests help to establish the ruggedness of the GPT-0  apparatus.
Move the equipment about as it might be moved during subsequent use to see if
any malfunctions, leaks, variability, or other problems are observed which
might make use of the procedure questionable or inconvenient.  If possible,
include relocation in the variability tests to measure the effect of relocation
on variability.

Malfunctions

     There is no special test for malfunctions.  During other tests, be obser-
vant for any leaks or malfunctions that occur or other characteristic weakness
in the apparatus or equipment which could cause unreliability.  Understand the
design and operation of the components so that you can be alert to non-obvious
types of malfunctions.

     In summary, the primary qualification tests for most users of the BAKI
procedure are variability tests, relocation tests, and verification that each
operator can make the proper temperature, pressure, and other corrections to
the flow measurements.

CERTIFICATION

     When the GPT-0  procedure has been shown to meet the qualification require-
ments, certify the procedure as specified in Section 6.  Additional guidance is
contained under "Certification" in Section 5.  If any volume or flow correction
formulas are needed, use them during the certification procedure and clearly
specify the conditions at the time of the certification relationship.  Be
sure that the specified correction formulas are accurate.  Note any special
operating instructions, operating restrictions, limits, or other pertinent in-
formation.  It is very important to complete the documentation requirements
specified in paragraph 3 of Section 6.
                                      D-6

-------
APPENDIX D.  GPT-0  PROCEDURE/Use
USE
     In using the  GPT-03  procedure as a transfer standard,  review the guidance
under "Use" in  Section 5.   Recertify the transfer standard  as required by
Section 6  (with further explanation of recertification in Section 5).   Consider
occasional cross checks of the transfer standard to other 0  standards, and
occasionally repeat  the qualification tests to be sure the  transfer  standard
is maintaining  adequate reliability.

RECOMMENDED VERSION  OF GAS PHASE TITRATION WITH EXCESS OZONE  TRANSFER STANDARD
PROCEDURE
     This procedure can be used as a transfer standard for  calibration of 0
analyzers in  accordance with the guidance and 'specifications set forth in Sec-
tions 5 and 6.
     1.  Principle.

     1.1   The procedure is based upon the rapid gas  phase  reaction between ozone
 (0  ) and nitric  oxide (NO) in accordance with the  following equation    :

                               NO + 0  -> NO  + 0                      (Eq. D-l)
                                     j     £.     £

When NO is added to  0  in a dynamic system,  the decrease in 0  response observed
on  an uncalibrated 0  analyzer is equivalent to the  concentration of  NO added.
By  measuring this decrease in response and the initial response, the  03 concen-
tration can be determined.  Additional 0  calibration concentrations  are gen-
erated by  a dilution technique.   The GPT system is used under  predetermined
flow conditions  to insure that the reaction  of NO  is complete  and that  further
reaction of the  resultant nitrogen dioxide (N02) with residual 03 is  negligible.

     2.  Apparatus.   Figure D-l, a schematic of a  typical  GPT  apparatus, shows
the suggested configuration of the components listed below.  All connections

                                      D-7

-------
APPENDIX D.  GPT-0  PROCEDURE/Recommended Version — Apparatus

between components in the calibration system downstream from the 03 generator
should be glass or Teflon.  Additional information regarding the assembly of
a GPT calibration apparatus is given in Reference 2.

     2.1  Air flow controllers.  Devices capable of maintaining constant air
flow within ± 2%.

     2.2  Nitric oxide flow controller.  A device capable of maintaining con-
stant NO flow within ± 2%.  Component parts in contact with the NO must be of
a non-reactive material.

     2.3  Air flowmeters.  Properly calibrated flowmeters capable of measuring
and monitoring air flows within ± 2%.

     2.4  Nitric oxide flowmeter.  A properly calibrated flowmeter capable of
measuring and monitoring NO flows within ± 2%.  (Rotameters have been reported
to operate unreliably when measuring low NO flows and are not recommended.)

     2.5  Pressure regulator for standard NO cylinder.  This regulator must
have non-reactive internal parts and a suitable delivery pressure.
                                 t

     2.6  Ozone generator.  Capable of generating a stable level of 0  at the
flow rates required (see step 4).

     2.7  Reaction chamber.  A glass chamber for the quantitative reaction of
NO with excess 0 .  The chamber should be of sufficient volume (V  ) such that
                •3                                                RC
the residence time (t ) is as specified in step 4.
                     R

     2.8  Mixing chamber.  A glass chamber of proper design to provide thorough
mixing of reaction products and diluent air.  The residence time is not critical
when the other specifications given in step 4 are met.
                                     D-8

-------
APPENDIX D.  GPT-03 PROCEDURE '/Recommended Version — Apparatus

     2.9  Output manifold.   The output manifold should be constructed of glass
or Teflon of sufficient diameter to insure a minimum pressure drop at the analyz-
er connection.  The system  must have a vent designed to insure atmospheric
pressure at the manifold and to prevent ambient air from entering the manifold.

     3 .  Reagents .

     3.1  Nitric oxide concentration standard.   Cylinder containing 50 to 100
ppm NO in N2-  The cylinder should be traceable to an NBS NO in N  Standard
Reference Material  (SRM 1683 or SRM 1684)  or NO2 Standard Reference Material
 (SRM 1629).  Traceability is important because  the [0 ]    /[O 1    ratio  is
useful for optimizing the system flowrates (step 4.3)  and as a check  of  the
transfer standard's accuracy.  The cylinder (working standard)  should be re-
certified on a regular basis as determined by the local quality control  pro-
gram.

     3.2  Zero Air.   Air, free of contaminants  which will cause a detectable
response on the 0  analyzer or which might react with either NO or 0   in the
gas phase titration.   A procedure for generating zero air is given in Reference
2.

     4.  Dynamic parameter  specifications.

     4.1  Prior to qualification and certification of the apparatus and  pro-
cedure, the residence time  (tD) in the reaction chamber and the gas flows
                              t\
 (F  and F   , see  Figure D-l) must be optimized.  Initially, they  should  be
  0      NO
adjusted according to the following relationships:

                         P«  =  10,] __ x t_ = 2.0 ppm-min                (Eg. D-2)
                          R     j RC    K
                          rn i    = ro l
                          l°3JRC   LU3JOUT
                                 t  =
                                                                      (Eq. D-4)
                                      D-9

-------
APPENDIX D.  GPT-0  PROCEDURE/Recommended Version — Dynamic parameter

             specifications





where:  P       = dynamic specification, determined empirically, to insure  com-

                  plete reaction of NO, ppm-min


        [0 ]    =0-, concentration (approximate) in the reaction chamber, ppm
          3 RC     3

        t       = residence time in the reaction chamber, min
         R

        [0 ]    = 80% URL concentration of 0   (approximate) at the output
          «j UU X      , -, _ _                  -j
                  manifold, ppm


        F       = total flow at the output manifold, scm3/min


        F       = 0  generator air flow, scm3/min


        FM_     = NO flow, scm3/min
         NO

        V       = volume of the reaction chamber, scm3
         RC




     4.2   These parameters may be selected according to the following sequence:
     4.2.1  Determine F , the total flow required at the output manifold  (F  =


output demand plus 10 to 50% excess).
     4.2.2  Determine [0 ]    as the 80% URL concentration required at the out-


put manifold.
     4.2.3  Determine F   as
                                  0.8 x
                             NO



where:  [NO]    = concentration of the undiluted NO standard, ppm
            O J.JJ





     4.2.4  Select a convenient or available reaction chamber volume.  Initial-


ly, a trial V   may be selected to be in the range of approximately 300 to


1500 scm3.
     4.2.5  Compute F  as
                               no l     x F  x v
                              ,  3JOUT    T    RC

                        F0 - /	2TO	FNO
                                      D-10

-------
APPENDIX D.  GPT-0   PROCEDURE/Recommended Version - Dynamic parameter
             sped fi ca ti ons
     4.2.6  Compute  t  as
                      R
                                         V
                                   R ~ F  + F	                       (E
-------
APPENDIX D.  GPT-0  PROCEDURE /Recommended Version — Dynamic parameter
             sped ficati ons

     NOTE:  The  [0 ]  ,r/[0-J ty ratio can also be affected by an inaccurate NO
cylinder standard, inaccurate flow measurements, leaks, and possibly other fac-
tors.  Errors from these factors must thus be eliminated before the flows are
adjusted based on that ratio.  If possible, measurements with an NO/NO  analy-
zer should be used to establish optimum flowrates to avoid both incomplete NO
reaction and secondary reaction of NO  with 0 .
                                     £*       J

     5.  Procedure.

     5.1  Assemble a dynamic transfer standard system such as shown in Figure
D-l.

     5.2  Establish the dynamic flow conditions as indicated in step 4.

     5.3  Insure that all flowmeters are properly calibrated under the condi-
tions of use against a reliable standard such as a soap-bubble meter or wet
test meter traceable to NBS.  All volumetric flowrates should be corrected to
25°C and 760 torr.
     5.4  Precautions must be taken to remove 0  and other contaminants from
the NO pressure regulator and delivery system prior to the start of calibration
to avoid any conversion of the standard NO to NO«.  Failure to do so can cause
significant errors in calibration.  This problem may be minimized by (1) care-
fully evacuating the regulator, when possible, after the regulator has been
connected to the cylinder and before opening the cylinder valve; (2) thoroughly
flushing the regulator and delivery system with NO after opening the cylinder
valve; and  (3) not removing the regulator from the cylinder between calibrations
unless absolutely necessary.  Further discussion of these procedures is given
in Reference 2.
     5.5  Certification as a transfer standard.
                                     D-12

-------
APPENDIX D.  GPT-03 PROCEDURE/Recommended Version - Procedure

     5.5.1  Carry out  appropriate qualification tests as discussed  earlier
in this appendix, using the indicated QS concentration ([O.j]    )  from Equation
D-10 for qualification tests.
     5.5.2  Certify the  procedure against a primary 0   standard as described
in Sections 5 and  6.   Prepare a certification relationship  curve relating in-
dicated 03 concentrations  ([03]QUT from Equation  D-10,  step 5.6.6) to the pri-
mary standard 03 concentrations,  similar to Figure  5-7, pg  5-23).  Record the
gas flows  (FQ, FD/ FNQ) , and the 03 generator setting required for the 80% URL
0  concentration.

     5.5.3  Recertify  the  procedure as  appropriate  in accordance with Sections
5 and 6.
     5.6  Calibration  of  the  0_  analyzer.

     5.6.1  Allow sufficient  time  for  the  0  analyzer to warm-up and stabilize.
Adjust the diluent air and  0   generator  air  flows to the same values as used
during certification.   To insure that  no ambient air is pulled into the manifold
vent, the total air  flow  must exceed the demand of the analyzer under calibra-
tion or the UV certification  photometer  connected to the output manifold.  Allow
the 0  analyzer to sample zero air until a stable response is obtained and ad-
just the 0  analyzer's zero control.   Offsetting the analyzer's zero adjustment
to + 5% of scale is  recommended  to facilitate observing negative zero drift.
Record the stable zero air  response  as "Z".

     5.6.2  Adjust the 0  generator  to generate an 03 concentration of approxi-
mately 80% of the URL  using the  same generator setting as during certification.
When the response has  stabilized,  record the response as IQ.  If the analyzer
response is offscale,  adjust  the diluent air flow  (FD) until an on-scale re-
sponse is obtained and measure the new flow.
                                     D-13

-------
APPENDIX D.  GPT-0  PROCEDURE'/Recommended Version — Procedure

     5.6.3  Turn the NO flow on and adjust to the same value as used during
certification.  Provided the 0  generator output has not changed significantly,
the 0  analyzer response should decrease by 75 to 80% of its original value.
When the resultant response has stabilized, record the response as I.
     5.6.4  Measure the NO flow and record as F  .
     5.6.5  Calculate the exact NO concentration from
                            [H01 .
                                                                        .  D-9)
where:   [NO] = diluted NO concentration, ppm
     5.6.6  Calculate the indicated 0  concentration from
                         - zl '' - " - z)
where:   [0 ]    = indicated 0  concentration, ppm
                                                                     (Eq'
        I       = original 0  analyzer response, % chart
        I       = resultant 0  analyzer response aft
        Z       = stable zero air response, % chart
I       = resultant 0  analyzer response after addition of NO,  % chart
     5.6.7  Remove the NO flow.  The 0  analyzer response should return to
its original value.
     5.6.8  Determine the certified 0  concentration from the certification
relationship obtained in step 5.5.2 and adjust the 0  analyzer's span control
to obtain a convenient recorder response as indicated below:
                                     D-14

-------
APPENDIX D.  GPT-03 PROCEDURE /Recommended Version - Procedure

                                                3 PFRT
              recorder response (% scale)  = ( —  —* x x 100)  + Z     (Eq.  D-ll)
                                                URL

where:   [03JCERT = certified 03 concentration,  ppm
        URL       = upper range limit of the 0   analyzer,  ppm
        Z         = recorder response with zero  air, % scale.

If substantial  adjustment of the span control  (more than  ±  15%)  is necessary,
make the span adjustment and repeat steps 5.6.1 through 5.6.8.   If F  had to be
adjusted to  give an on-scale 0  analyzer response,  readjust to the initial
value prior  to  repeating these steps.   After the final span adjustment, record
the certified 0  concentration and the analyzer response.

     5.6.9   Calculate the indicated 0  concentration produced by the 0  genera-
tor from
                           [°3]GEN * C°3]OUT         >

where:   [0_]__.T = 0_  concentration produced by the  0  generator, ppm
           3  GEN    o                                 o

      5.6.10   Generate several other 0  concentrations  (at  least four are sug-
gested)  by increasing F ,  the diluent air flow.   Calculate the diluted 0  con-
centrations  from
                                                 F
                           [0 1 '    = [0 ]     ( - - — )                 (Eq. D-13)
                           L 3JOUT   l 3JGEN        ,
where:   t^QUT = diluted indicated 03 concentration,  ppm
                                                Q
         F'       = the new diluent air flow,  scm /min
          D

      5.6.11  Allow the 0  analyzer to sample each diluted QS concentration un-
til a stable response is obtained.  For each concentration,  determine the certi-
fied  0   concentration from the certification relationship and record the ana-
lyzer response and the corresponding certified DS concentration.

                                       D-15

-------
APPENDICES D.  CPT-0  PROCEDURE/Recommended Version —Procedure

     5.6.12  Plot the 0  analyzer responses versus the corresponding certified
0  concentrations and draw the 0  analyzer's calibration curve or calculate  the
 j                              ,5
appropriate response factor.

References for Appendix D

1.   K.A. Rehme, B.E. Martin, and J.A. Hodgeson.  Tentative Method for the
     Calibration of Nitric Oxide, Nitrogen Dioxide, and Ozone Analyzers by
     Gas Phase Titration.  EPA-R2-73-246, U.S. Environmental Protection Agency,
     Research Triangle Park, North Carolina.  March, 1974.  pp. 17

2.   B.C. Ellis.  Technical Assistance Document for the Chemiluminescence
     Measurement of Nitrogen Dioxide.  EPA-600/4-75-003.  U.S. Environmental
     Protection Agency, Research Triangle Park, North Carolina.
                                     D-16

-------
                      ZERO
                      AIR
                                  FLOW
                               CONTROLLER
                                  FLOW
                               CONTROLLER
O
                           NO
                           STD.
v
,_^
J
FLOW
CONTROLLER


FLOWMETER
                                                                        lip
                                             VENT
                                                                      OUTPUT
                                                                     MANIFOLD
TIT
f
P
                                                            EXTRA OUTLETS CAPPED   TO INLET OF
                                                             WHEN NOT IN USE       ANALYZER
                                                                                 UNDER CALIBRATION
                                                                                                                          tf
                                                                                                                          I
                                     Figure D-l.   Schematic diagram of  a typical  GPT system.

-------
                                   APPENDIX E
          CERTIFICATION OF  AN OZONE GENERATOR AS A TRANSFER STANDARD

     Following the  guidance in Sections  4  and 5  and the  specifications in
Section 6, this appendix attempts  to provide more specific  instructions for
certifying an 03 generator  as a transfer standard.

     We will assume,  here,  that the 0 generator is of the  common UV lamp
type.  We will also assume  that it has a means to adjust the 0  concentra-
tion over a convenient range without changing the flowrate, but that the
device has no means to assay the output  0   concentration.

PRELIMINARY REQUIREMENTS

     The first requirement  is a source of  zero air.  A means of reasonable
flow regulation for the zero air is also needed,  since the  concentration of
0  generated by the UV lamp varies with  flow.  Some commercial 0  generation
systems may include some or all of the required  zero air source and flow com-
ponents.  Otherwise,  the zero air  subsystem must be provided by the user.
Scrubbed ambient air  is preferable to cylinder zero air  as  the latter may vary
in oxygen content or  impurity level from one cylinder to another.  Some helpful
guidance may be found in Section 2 of this document.  Decide whether or not the
zero air subsystem  is to be an integral  part of  the transfer standard (see
pg 4-2).  If not, the flow  regulation and  flow monitoring components at least
should be integral.   Be aware that differences in zero air  from one zero air
system to another may affect the repeatability of the O3 generator.

     The 0  generator also  needs an output manifold which meets the require-
ments specified in  paragraph 3.4 of Appendix A.   The manifold may be as simple
as TEE where one of the legs serves as a vent.

                                      E-l

-------
APPENDIX E.  OZONE GENERATOR/Preliminary Requirements

     Access to a UV calibration system as described in Appendix A is required
for certification of the transfer standard and is also recommended for the
qualification tests.  Review the 0  alternate procedures for comparing the
output of an 0 -generation-type transfer standard to a UV primary 0  standard
(see Section 5) and decide which procedure to use.

     Review any operation information or instructions provided by the manufac-
turer of the 0  generator to become familiar with its operation.

     Review the documentation requirements specified in paragraph 3 of Section
6 and complete item 3.1.

QUALIFICATION

     The next step is to qualify the transfer standard by demonstrating that
it is repeatable to within the specifications given in paragraph 4 of Section
6  (± 4% or 4 ppb, whichever is greater).  The variables likely to affect an 0_
generator are the same as those discussed generally in Section 5, and specifi-
cally below.  Refer to Section 5 for additional guidance on each of the qualifi-
cation tests.

     Before starting qualification tests, first prepare a tentative preliminary
calibration relationship as shown in Figure 5-7 (pg 5-23), where the 0  out-
put concentration is related to the 0  adjustment setting.  Although you may
want to prepare a more complete preliminary calibration relationship after
qualification, this tentative relationship is necessary to carry out the quali-
fication tests.  Prepare the relationship as shown in Figure 5-7 by plotting
the transfer standard's output concentration as measured by the UV reference
system at various 0  settings as discussed in Section 5 under "Comparing
Transfer Standards to an Ultraviolet Primary Ozone Standard".  Note the tempera-
ture, barometric pressure, line voltage, and other pertinent conditions.  During
the qualification tests, use this preliminary relationship to determine each
"indicated output" from the transfer standard for a given setting of the con-
centration adjustment (i.e., sleeve setting or current setting).
                                     E-2

-------
APPENDIX E.  OZONE  GENERATOR/'Qualification Tests

QUALIFICATION TESTS

Temperature

     If possible, select an 0   generator which has temperature  regulation,
preferably one which has a temperature indicator or other warning  device to
indicate whether the temperature regulator is working properly.  Keep in mind
that such systems will require a warm-up period before the temperature stabi-
lizes.
     Select a temperature range over which the  0   generator  is to be quali-
fied.  Twenty to  30°C (68 to 86°F)  might be appropriate, or  possibly 15 to 30°C
(59 to 86°F).   To show how the indicated output of the generator varies as the
temperature changes,  test the generator over this  temperature range at several
0_ concentrations as  suggested by Figure 5-3 (pg 5-12) .  Be  sure (1) that the
proper temperature and pressure corrections are made  to the  UV standard, (2)
that the 0  generator is  allowed to equilibrate each  time the temperature is
changed, and  (3)  that the 0  setting is repeated precisely for each different
temperature.  If  the  manufacturer has tested the 0 generator  (or one like it) ,
only enough tests are needed to show that the generator is operating properly
and meets the specifications.

     If the 0   generator's actual output concentration does  not vary more
than ± 4% over  the entire temperature range, then  it  is qualified over that
temperature range.  If it does not  meet those specifications, the following
options are available:

     (a)  determine if the generator has a malfunction or inadequacy,
          attempt to  correct it, and then retest it;
     (b)  reduce  the  temperature range to a range  over which the generator
          does  meet the specifications (this may inconveniently restrict the
          subsequent  use  of the generator); or
                                       E-3

-------
APPENDIX E.  OZONE GENERATOR/Qualification Tests

      (c)  attempt to determine, either analytically or empirically, the
          temperature-output relationship such as illustrated in Figure  5-4
           (pg 5-13.  When this relationship is used to calculate a "cor-
          rected" indicated output, the 0  generator should meet the specifi-
          cations.  If so, this correction formula becomes a necessary and
          integral part of the transfer standard and must be included with
          the preliminary calibration relationship  (see Figure 5-7, pg 5-23)
          and ultimately with the certification relationship (see Figure 5-8,
          pg 5-24).

Line Voltage

     Line voltage tests are conducted in a fashion analogous to the temperature
tests, substituting line voltage variation for temperature variation.  A range
of 105 to 125 volts should be appropriate.  A well designed 0  generator should
show little or no sensitivity to line voltages over this range.  If the genera-
tor does not meet the specifications, you have the same three options as with
temperature, plus a fourth option of adding an external voltage regulator.

Barometric Pressure/Altitude
     Virtually all UV 0  generators are sensitive to pressure changes unless
they are specifically designed to compensate for pressure effects.  If use of
the 0  generator can be restricted to altitudes within a range of about 100
meters (300 feet), the range of pressure variations is only .about 2 or 3%
(about 1% for altitude variations and about 1 to 2% for normal barometric
pressure changes). Under these conditions an uncpmpensated 0  generator opera-
ting at ambient pressure can be expected to meet the qualification specifica-
tions.  Pressure changes could then be ignored and no tests would be necessary.

     Where a larger altitude range is needed (or for applications requiring
more accuracy), the 0  generator's sensitivity to pressure changes must be
compensated, either by design or by defining the pressure-output relationship
and developing a correction formula (see Figure 5-5, pg 5-16).
                                     E-4

-------
APPENDIX E.  OZONE  GENERATOR/Qualification Tests

     Testing for pressure sensitivity is most practically carried out by
moving both the 03  generator and the UV standard system to various altitudes.
In doing this, be sure (1)  that the proper temperature and pressure corrections
are made to the UV  standard, (2)  that the 0  setting is repeated precisely for
each new pressure,  and (3)  that other variables (temperature,  line voltage,  etc.)
are controlled so that they do not affect the generator.   From the test  data,
determine an appropriate correction formula and include it with the prelimi-
nary calibration relationship and then ultimately with the certification rela-
tionship .

Elapsed Time

     The output of  a UV 0  generator is likely to decrease somewhat with
usage time.  Comparisons with a UV reference over a period of  time  are neces-
sary to determine the rate of decay.  Then,  either the transfer  standard  can
be recertified often enough so that it stays within specifications between
certifications, or  a correction factor based on elapsed time can be determined
and used with the preliminary calibration relationship.

Variability

     General variability (variability other than that associated with a  speci-
fic known variable) is not likely to be a problem with an 0 generator unless
it is very poorly designed.  If any of the other qualification tests show vari-
ability not strongly correlated with a specific variable, then a test for gen-
eral variability would be needed.  If the generator meets the  specifications
for the other tests, no specific test for general variability  is required.

Relocation

     Recolation tests help to establish the ruggedness of the  C>3 generator.
During the course of the elapsed-time or other tests, move the C>3 generator
                                      E-5

-------
APPENDIX E.  OZONE GENERATOR/Qualification Tests

about as it might be moved during subsequent use to see if any malfunctions,
variability, or other dependability problems are observed that might make use
of the generator questionable or inconvenient.

Operator Adjustments

     Test the 0  generator for repeatability of 0_ concentration setting.
               J                                 -J
Check each of several settings repeatedly, approaching sometimes from a higher
setting and sometimes from a lower setting.  Intersperse the various settings.
Note if a stabilization period is required for each new setting, or if any
other observable perculiarities are evident.  A well-designed 0_ generator
is not likely to fail this test, but the information obtained will help to
achieve better precision.

Maifunctions

     There is no special test for malfunctions.  During other tests, be obser-
vant for any malfunctions that occur or other characteristic weaknesses in the
generator that could cause unreliability.  Understand the design and operation
of the generator so that you can be alert to non-obvious types of malfunctions
such as failure of temperature, line voltage, flow or other regulation mecha-
nisms .

Other Conditions
     The tests described above should cover the performance variables for most
     m UV-type 0  generators.  However, new types of gei
generator designs may require additional special tests.
common UV-type 0  generators.  However, new types of generators or unusual
CERTIFICATION

     Before conducting the certification tests, decide whether a new prelim-
inary calibration relationship (see Figure 5-7, pg 5-23) should be prepared.

                                     E-6

-------
APPENDIX E.  OZONE GENERATOR/Certification

A new preliminary calibration relationship is  advisable if:

      (a)  the transfer  standard needs  one  or more  correction  formulas for
          defined-relationship variables;
      (b)  the original  relationship was  rough,  inaccurate, or incomplete;
      (c)  the original  relationship indicates  output  concentrations more
          than  30% different  than the  UV standard; or
      (d)  the qualification tests provided other information  to suggest
          that  a new  relationship should be prepared.

If a new preliminary  calibration relationship  is needed, prepare it carefully
and accurately, including  enough points  to define  it  precisely over the entire
operating range.  Ozone generators that  are non-linear  or have appreciable
variability need additional comparisons  to define  the relationship precisely.
If any correction formulas are needed, clearly  specify  the conditions at the
time the preliminary  calibration relationship  is established, and be sure that
the specified correction formula is accurate.   If  necessary, various qualifi-
cation tests should be  repeated to verify  that  the corrections are accurate.
Note or clarify any special operating  instructions, operating restrictions,
limits, or other pertinent information.

     When the preliminary  calibration  relationship has  been prepared, proceed
to certify the  transfer standard as specified  in Section 6.  Additional guid-
ance is contained under "Certification"  in Section 5.   It is very important to
complete the documentation requirements  specified  in  paragraph 3 of Section 6.

USE

     In using the 0   generator as a transfer standard,  review the guidance
under "Use" in  Section  5.   Recertify the transfer  standard as required by  Sec-
tion 6  (with further  explanation of recertification in  Section 5).  Consider
occasional cross checks of the transfer  standard to other 03  standards,  and
occasionally repeat the qualification  tests to be  sure  the transfer standard
is maintaining  adequate reliability.
                                     E-7

-------
                                  APPENDIX F
           CERTIFICATION OF AN OZONE ANALYZER AS A TRANSFER STANDARD

     Following the guidance in Sections 4 and 5 and the specifications in
Section 6, this appendix attempts to provide more specific instructions for
certifying an 03 analyzer as a transfer standard.

     The analyzer does not have to be a UV-type analyzer; any type of 0
analyzer may be considered for use as a transfer standard.  However, UV-type
analyzers may be more readily portable or have other practical advantages over
other types of 0  analyzers.  We will assume, here, that the analyzer is to
be certified over a range of 0  concentrations, e.g., 0 to 1 ppm.

     An analyzer used as a transfer standard should receive special treatment
consistent with its authoritative status:  careful handling and storage, extra
frequent maintenance service and maintenance checks, operation by a fully com-
petent operator, etc.  In particular, the analyzer should not be used for
ambient monitoring between uses as a transfer standard, as dust or dirt build-
up in the cell and other operational degradation may occur.  Where it is
necessary to use an analyzer that has had previous service as an ambient
monitor, the analyzer should be thoroughly cleaned and reconditioned prior to
certification as a transfer standard.

PRELIMINARY REQUIREMENTS

     First, a stable source of 0  to be assayed by the analyzer/transfer
standard, and an attendant source of zero air for the C>3 generator are
required.  Some analyzers have internal or associated 0  generators that
may be used; otherwise, an 0  generator/zero air system must be provided by
the user.  In addition, some means of reasonable flow regulation  for the zero

                                     F-l

-------
APPENDIX F.  OZONE ANALYZER/Preliminary Requirements
air is usually needed, since the concentration of 0  generated by most 0
generators varies with flow.  Some helpful guidance may be found in Section
2 of this document.  Decide whether or not the 0. generation/zero air system
is to be an integral part of the transfer standard (see pg 4-2).  If not, ar-
rangements must be made for an adequate 0, generator/zero air supply at each
site where the transfer standard will be used.
     The 0  generator will also need an output manifold meeting the require-
ments specified in paragraph 3.4 of Appendix A.  The manifold may be as simple
as a TEE where one of the legs serves as a vent.

     Access to a UV calibration system as described in Appendix A is required
for certification of the transfer standard and is also recommended for the
qualification tests.  Comparing the output (indicated concentration) of the
0  analyzer to the UV primary 0  standard is easy, as the analyzer is simply
 •J                             J
connected to the output manifold of the UV primary 0  standard.

     Review any operation information or instructions provided by the manufac-
turer of the 0  analyzer to become familiar with its operation.

     Review the documentation requirements specified in paragraph 3 of Section
6 and complete item 3.1.

QUALIFICATION

     The next step is to qualify the standard by demonstrating that it is
repeatable to within the specifications given in Section 6  (± 4% or ± 4 ppb,
whichever is greater).  The variables likely to affect an 0  analyzer are
normally the same as those discussed generally in Section 5 and more specifi-
cally below.  Refer to Section 5 for additional guidance on each of the qualifi-
cation tests.
                                     F-2

-------
APPENDIX F.  OZONE ANALYZER/Qualification

     Generally, a preliminary  calibration  relationship  (as  shown in Figure 5-7
pg 5-23) is not necessary for  03  analyzers,  as most  analyzers are linear and
provide a direct output  indication  of  concentration.  The zero and span of the
analyzer should be adjusted  for approximate  calibration  (with respect to the
UV standard) over the desired  concentration  range.   The  final "calibration" of
the analyzer is the certification relationship.

     An output-indicating device  such  as a chart recorder or digital meter is
a helpful accessory.  But, if  such  an  output indicator is used, it should be
permanently associated with  the transfer standard analyzer  and employed during
qualification, certification,  and use  of the analyzer as a  transfer standard.

QUALIFICATION TESTS

Temperature

     If possible, select an  0  analyzer that has good temperature regulation
(or compensation), preferably  one with a temperature indicator or other warning
device to indicate whether the temperature regulator is working properly.
Keep in mind that temperature  regulation systems require a warm-up period
before the temperature stabilizes.

     Select a temperature range over which the 0  analyzer  is to be qualified.
Twenty to 30°C  (68 to 86°F)  might be 'appropriate, or possibly 15 to 30°C
(59 to 86°F).  Test the  0  analyzer over this temperature range at several
0  concentrations as suggested by Figure 5-3 (pg 5-12) to depict how the
indicated output of the  analyzer  varies as the temperature  changes.  Be sure
(1) that the proper temperature and pressure corrections are made to the UV
standard, (2) that the 0 analyzer  is  allowed to equilibrate each time the
temperature is changed,  and  (3) that the 0^  analyzer's span is not adjusted
between each different temperature  (adjustment of other  parameters to nominal
values is permitted).  If the  manufacturer has tested the 0 analyzer  (or one
like it), only enough tests  are needed to  show that  the  analyzer is operating
properly and meets the specifications.
                                     F-3

-------
APPENDIX F.  OZONE ANALYZER/Qualification Tests


     If the 0  analyzer's indicated concentration does not vary more  than

± 4% or ± 4 ppb over the entire temperature range, then  it is qualified  over

that temperature range.  If it does not meet those specifications, the fol-

lowing options are available:


     (a)  determine that the analyzer has a malfunction  or inadequacy,
          attempt to correct it, and then retest it;

     (b)  reduce the temperature range to a range over which the analyzer
          does meet the specifications (this may inconveniently restrict
          the subsequent use of the analyzer); or

     (c)  attempt to determine, either analytically or empirically,
          the temperature-output relationship such as illustrated in
          Figure 5-4 (see pg 5-13).  When this relationship is used
          to calculate a "corrected" indicated output reading, the
          analyzer should meet the specifications.  If so, this cor-
          rection formula becomes a necessary and integral part of the
          transfer standard and must be included with the ultimate
          certification relationship (see Figure 5-8, pg 5-24).


Line Voltage


     Line voltage tests are conducted in a fashion analogous to the temperature

tests, substituting line voltage variation for temperature variation.  A range

of 105 to 125 volts should be appropriate.  A well-designed 0  analyzer  should

show little or no sensitivity to line voltage changes over this range.   If the

analyzer does not meet the specifications, the same three options as with tem-

perature are available, plus a fourth option of adding an external voltage re-

gulator.                                       :?~:;b^"


Barometric Pressure/Altitude


     An 0  analyzer is likely to be sensitive to pressure changes unless it

is specifically designed to compensate for pressure effects.  If the use of

the transfer standard can be restricted to altitudes within a range of about

100 meters (300 feet), the range of pressure variations  is only about 2  or 3%

(about 1% for altitude variations and about 1 to 2% for  normal barometric

                                     F-4

-------
APPENDIX F.  OZONE ANALYZER/Qualification  Tests

pressure changes).   Under  these  conditions an uncompensated 0   analyzer
which operates at ambient  pressure  could be expected to meet the qualification
specifications.  However,  the  analyzer should be  tested over a  normal range of
barometric pressure  to be  sure:   Test  the  analyzer  on various days when the
barometric pressure  is different and plot  the results as shown  in Figure 5-3
(pg 5-12).

     Where a larger  altitude range  is  needed (or  for applications requiring
more accuracy) the 0 analyzer's sensitivity to pressure changes may have
to be compensated, either  by design or by  defining  the pressure-output relation-
ship and developing  a correction formula (see Figure 5-5, pg 5-16).

     Testing for pressure  sensitivity  is most practically carried out by
moving both the  0  analyzer and  the UV standard system to various altitudes.
In doing this, be sure  (1)  that  the proper temperature and  pressure corrections
are made to the  UV standard,  (2) that  the  0  analyzer is not adjusted between
each different pressure, and  (3) that  other variables (temperature, line voltage,
etc.) are controlled so that they do not affect the analyzer.   From the test
data, determine  an appropriate correction  formula and include it with the
certification relationship.

Elapsed Time

     Ozone analyzers operated  exclusively  as transfer standards are not likely
to change much with  elapsed time.  The specified  recertification frequency
should be sufficient to compensate  for any long-term response changes in the
analyzer.

Variability

     General variability  (variability  other than  that associated with a  specific
known variable)  is not likely  to be a  problem with  an 0  analyzer  unless  it
                                      F-5

-------
APPENDIX F.  OZONE ANALYZER/Qualification Tests

is very poorly designed.  If any of the other qualification tests show variabil-
ity not strongly correlated with a specific variable, then a test for general
variability would be needed.  Otherwise, if the analyzer meets the specifica-
tions for the other tests, no specific test for general variability is required.

Relocation

     Relocation tests help to establish the ruggedness of the 0  analyzer.
During the course of the barometric pressure or other tests, move the transfer
standard about as it might be moved during subsequent use to see if any malfunc-
tions, variability, or other problems are observed that might make use of the
analyzer questionable or inconvenient.  Make sure that the zero, span, and
other adjustments can be locked so they don't change when the analyzer is
moved.

Operator Adjustments

     Test the analyzer for repeatability with respect to any operator settings
such as flow or gas pressure.  Note if a stabilization period is required
after an adjustment.  Note any other observable peculiarities.  A well-designed
0  analyzer is not likely to fail this test, but the information obtained
will help to achieve better precision.

Maifunctions

     There is no special test for malfunctions.  During other tests, be obser-
vant for any malfunctions that occur or other characteristic weaknesses in the
analyzer that could cause unreliability.  Understand the design and operation
of the analyzer so that you can be alert to non-obvious types of malfunctions
such as failure of temperature, line voltage, flow or other regulation mecha-
nisms.
                                     F-6

-------
APPENDIX F.  OZONE ANALYZER/'Qualification  Tests

Other Conditions

     The tests described above  should cover  the  performance  variables  for most
common types of 0  analyzers.   However,  new  types  of  analyzers  or  unusual
analyzer designs may  require  additional  special  tests.

CERTIFICATION

     Before conducting  the  certification tests,  make  any zero,  span, or other
adjustments to the analyzer as  necessary so  that the  analyzer readings are
close to the 0  concentrations  obtained  from the UV primary  standard.

     When the adjustments are complete,  lock the adjustments, record their
values, and proceed to  certify  the  transfer  standard  as specified  in Section
6.  Additional guidance is  contained  under "Certification" in Section 5.  If
any correction formulas are needed, use  them during the certification procedure
and clearly specify the conditions  at the  time of  the certification relation-
ship.  Be sure that the specified correction formulas are accurate.  Note any
special operating instructions,  operating  restrictions, limits, or other
pertinent information.   It  is very  important to  complete the documentation
requirements specified  in paragraph 3 of Section 6.
USE
      In  using the  0   analyzer as  a transfer  standard,  review the guidance
under "Use"  in Section 5.   Make any zero,  flow, or other adjustments except
span  adjustment to the values recorded at  the  time of  certification.  Recertify
the transfer standard as  required by Section 6 (further explanation in  Section
5).   Consider occasional  cross checks of the transfer  standard  to  other 0
standards, and occasionally repeat the qualification tests  to be sure the
transfer standard  is  maintaining  adequate  reliability.
                                     F-7

-------
                                   TECHNICAL REPORT DATA
                            i/\Vj5V rccj f'.^j-.icnons on liie jvrcn'c. nciorc completing
  EPA-600/4-79-056
                                                          3. RECIPIENT'S ACCESSION-NO.
TRANSFER^STANDARDS FOR CALIBRATION OF AIR MONITORING
ANALYZERS FOR OZONE - TECHNICAL ASSISTANCE  DOCUMENT
             5. REPORT DATE
              September 1979
             6. PERFORMING ORGANIZATION CODE
                                                          8. PERFORMING ORGANIZATION REPORT NO.
 Frank  F.  McElroy
  PE ^f=OS%.!l\o ORGANIZATION NAME AND ADDRESS
 Duality  Assurance Division
 Environmental  Monitoring Systems Laboratory
 LS.  Environmental  Protection Agency
 Research Triangle Park, NC  27711
             10. PROGRAM ELEMENT NO.

             1AD800
             11. CONTRACT/GRANT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
Environmental  Monitoring Systems Laboratory
Office  of Research and Development
LS.  Environmental Protection Agency
Research  Triangle Park, NC  27711
             13. TYPE OF REPORT AND PERIOD COVERED
             Final            	   	
             14. SPONSORING AGENCY CODE
             EPA 600/08
 15. SUPPLEMENTARY .MOTES
Technical  Assistance Document
 16. At
       vfEPA regulations (40 CFR Part 50, Appendix  D)  allow the use of transfer standards
for  the  calibration of ambient ozone monitors.  This  document is  a reference aid to help
users  select ozone transfer standards and reference them  to a primary UV standard.  It
defines  ozone transfer standards, explains their purpose  and role in calibrating ambienl
ozone  analyzers, and discusses various advantages  and jdisadvantages of their use.
Several  different types of ozone transfer standards are described, including analytical
instruments, manual analytical procedures and ozone generation devices.
     The  major part of the document is devoted to the  procedures necessary to establish
the  authority of ozone transfer standards:  qualification, certification, and recertifi-
cation.   Qualification tests to determine repeatability over a range of variables such
as temperature,  line voltage, barometric pressure, elapsed  time,  operator adjustments,
or other conditions are discussed, and possible compensation  techniques are provided.
Detailed certification procedures are also provided together  with the quantitative spec-
ifications  that  the transfer standard must meet to achieve  and  maintain certification.
      Appendix A  contains the UV primary ozone standard  procedure.   Other appendices
give  more specific guidance for the qualification and certification of several  common
and practical  types of transfer standards—the BAKI manual  procedure, two gas phase
titration procedures, ozone generators, and ozone analyzers.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b. IDENTIFIERS/OPEN ENDED TEFTMS
                             COSATI Field/Group
Air  pollution
Measurement
 alibration
Air  pollution monitoring
 Ozone_analyzer
 Ozone standards
 Transfer standards
 Technical  assistance
     document
68A
43F
           ;:,• STATEMENT
19. SECURITY CLASS f This Report/
 UNCLASSIFIED
                                                                        21. iMO. OF PAGES
                                                                           150
RELEASE  TO  PUBLIC
20. SECURITY CLASS (Thispage)

 UNCLASSIFIED
                                                                        22. PRICE
EPA Form 2220-1 (9-73)

-------