CPA/600/4-79/057
vvEPA
rtes Environmental Monitoring and Support EPA-600 4-79-057
.. jntal Protection Laboratory September 1979
Agency Research Triangle Park NC 27711
Research and Development
Technical Assistance
Document for the
Calibration of
Ambient Ozone
Monitors
-------
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
-------
TECHNICAL ASSISTANCE DOCUMENT FOR THE CALIBRATION
OF AMBIENT OZONE MONITORS
by
Richard J. Paur
Environmental Sciences Research Laboratory
and
Frank F. McElroy
Environmental Monitoring and Support Laboratory
Research Triangle Park, NC 27711
ENVIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NC 27711
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and Support
Laboratory, U.S. Environmental Protection Agency, and approved for publication.
Mention of trade names or commercial products does not constitute endorsement
or recommendation for use.
11
-------
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.
The ultraviolet photometric procedures and guidance described in this
Technical Assistance Document represent new monitoring technology which is
now available to assist monitoring agencies in accurately calibrating ambient
ozone analyzers and in assuring the continual quality of atmospheric ozone
measurements.
Thomas R. Hauser, Ph.D.
Director
Environmental Monitoring Systems Laboratory
Research Triangle Park, North Carolina
111
-------
PREFACE
Analyzers designated as reference methods for the measurement of ozone in
ambient air are based on the chemiluminescent reaction of ozone with ethylene.
Such ozone analyzers are calibrated by having the instrument sample a calibra-
tion atmosphere and adjusting the instrument so that its output indicates the
concentration of ozone present in the calibration atmosphere. Calibration
atmospheres are obtained by means of the calibration procedure prescribed in
40 CFR Part 50, Appendix D.
The D.S. Environmental Protection Agency has replaced the original calibra-
tion procedure, promulgated in 1971, with a new procedure based on ultraviolet
absorption photometry (U.S. Environmental Protection Agency 1979). The purpose
of this document is to provide the user community with information to aid in
successful adoption of the new calibration procedure.
This document is organized into three major sections. The first section
is a discussion of absorption photometry, with primary emphasis on measurement
of the transmittance of gaseous samples. Section 2 provides a step-by-step
explanation of the new photometric ozone calibration procedure. Section 3
discusses the construction of photometers and provides the reader with some of
the considerations to be addressed when designing a photometer. This last
section also describes commercially available photometric systems.
iv
-------
CONTENTS
Foreword iii
Preface iv
Figures and Tables vi
Acknowledgments vii
1. THEORY OF PHOTOMETRIC MEASUREMENTS 1-1
Physical basis of the photometry equations 1-1
Sources of error involved in the principle 1-6
2. STEP-BY-STEP DISCUSSION OF THE ULTRAVIOLET PHOTOMETRIC
CALIBRATION PROCEDURE 2-1
Principle 2-1
Applicability 2-1
Apparatus 2-2
Reagents 2-6
Procedure 2-7
3. OBTAINING A PHOTOMETER 3-1
Construction of a photometer 3-1
Commercially available photometric systems 3-6
References 4-1
Appendix
A. Ultraviolet photometric procedure for primary ozone
standards A-l
v
-------
FIGURES
Number
3-1 Simplified block diagram of Dasibi Model 1003 Ozone
Analyzer 3-8
3-2 Simplified block diagram of Columbia Scientific
Industries Corp. (CSI) Photocal 3000 Ozone Calibrator. . . 3-18
TABLE
1-1 Ozone Absorptivity 1-7
VI
-------
ACKNOWLEDGMENTS
The authors gratefully acknowledge the invaluable assistance of Michael E.
Beard, John C. Puzak, Kenneth A. Rehme, and many others who contributed informa-
tion and suggestions, reviewed the manuscript, or otherwise assisted in the
preparation of this document.
vii
-------
SECTION 1
THEORY OF PHOTOMETRIC MEASUREMENTS
A brief review of the theory of photometric measurements should help the
reader more fully appreciate the advantages and shortcomings of photometry.
PHYSICAL BASIS OF THE PHOTOMETRY EQUATIONS
Consider the transmission of light through a plane of area A which con-
tains some number, n, of absorbing molecules. If radiation of intensity I
is perpendicularly incident on the plane, and light of intensity I emerges
from the plane, the transmittance, T, of the plane is defined as
The transmittance is equal to
1 - fraction of light not transmitted
1-1
-------
1. THEORY '/Physical Basis of the Photometry Equations
1 _ area blocked by absorbing molecules
area of the plane, A
The area blocked by the absorbing molecules is equal to the product of
the area blocked by a single molecule times the number of molecules, and the
equation can be written as
T - i n x 0 tvn 1 1
T - I -- - (Eq. 1)
where 0 is the effective absorption cross-sectional area of the absorbing
2
species. The units of a are cm /molecule; a is a measure of the molecule's
capacity to block light by absorbing it. It is often useful to consider a to
be a measure of the effective size of the molecule.
The value of the effective absorption cross-sectional area depends on:
1. The nature of the absorbing species; i.e., some species absorb
light of a given wavelength (X) while others do not.
2. The wavelength of the light; i.e., the absorption of light by
a molecule is different at different wavelengths. This pro-
perty is what gives a species its characteristic spectrum.
In terms of a simple model, this means that the effective
size of a molecule is a function of the wavelength of light
used to observe it.
3. The temperature and pressure of the gas when observing spectro-
scopic transitions in which both the upper and lower states
are bound, i.e., under conditions where the observed spectrum
is a sharp line. Since the ozone spectrum in the 250 nm wave-
length region is primarily a broad continuum, the absorp-
tion cross section is only slightly dependent on temperature
and pressure (Griggs 1968) . To within the accuracy required
for measuring ozone photometrically at 254 nm, one can assume
that a is a constant over ordinary temperatures and pressures.
1-2
-------
1. THEORY/Physical Basis of the Photometry Equations
Since the sample cell in a photometer is of some finite length instead
of being a plane, Equation 1 can be rewritten as
n x a x H In
T = 1 : = 1 - — x a x £
A x H IV
(Eq. 2)
where £ is the length of the sample cell and V is the volume of the sample
cell. The term n/V is a concentration in units of number density. The fact
that the transmittance of a sample depends on the number density of the ab-
sorbing species is convenient because it allows one to correct photometric
measurements for the temperature and pressure of the sample by simple gas
law calculations.
Most chemists, air pollution technicians, and other workers making photo-
metric measurements prefer to use concentration units of atmospheres* and
absorption coefficients in units of atm cm instead of molecules/cm and
2
cm /molecule, respectively. The gas law can be written
P (atm) = n (molecuies) * T
„ , 3 (molecules K)
V (cm )
Therefore, Equation 2 can be changed to atmospheres by multiplying the number
density term by RT and dividing the absorption cross section by the same
factor. Letting
n (molecules) P (atm)
3 3
V (cm ) (cm atm)
RT
(molecules)
and
2
, . -1 -1. o (cm /molecule)
a (atm cm ) = —
RT (°m atm)
(molecule)
*A concentration of 1 atm equals the concentration of a gas at 1 atmosphere
of pressure and standard temperature of 0° C.
1-3
-------
1. THEORY/Physical Basis of the Photometry Equations
Equation 2 becomes
T=l-ciPS, = |- (Eq. 3)
o
This very useful form of the photometry equation will be utilized later.
First, consider the conditions for which the equation was derived. A basic
assumption in the derivation is that the fraction of the light beam blocked by
the absorbing molecules equals
or that the cross-sectional area of the beam that is blocked is n x a. This
assumption is valid only if the number of absorbing molecules in the sample is
so small that one molecule never "shades" another. As soon as one molecule
begins to shade another, the cross-sectional area of the beam that is blocked
by these two molecules is less than 2a.
To extend Equation 3 to the concentration and/or cell length range where
significant shading of one molecule by others occurs (i.e., the concentration
range where one normally applies photometry), the length of the absorption
volume is divided into small segments, dfc, such that there is no shading
within any small segment. The cumulative result is obtained by "adding" the
intermediate result from each segment by mathematical integration. Thus,
Equation 3 is rewritten as
I -1=1 a P
o o
In integral form, this becomes
dl
r cu t
/ — = -/ a
p as,
1-4
-------
1. THEORY'/Physical Basis of the Photometry Equations
or
lnI=-aP£+C
When P = 0, I = I ; therefore C = In I . The equation then becomes
o o
In - = - a P H
or
I -O.PSL -aP£/2. 30259 1n-a'PJl /TT, .,
T = — = e =10 =10 (Eq. 4)
o
Equation 4 is the form of the photometry equation that is probably most
familiar to the majority of workers. It is valid over wide concentration and
cell length ranges. Equation 4 is usually designated as the Bo uguer- Lambert-
Beer law or the Lambert-Beer law. The absorption coefficient, a1, is called
the absorptivity and is usually written simply as a when working in base 10.
The concentration is normally designated by the letter c, and in most common
usage Equation 4 is written as
I , ~-ac£ -2.30259ac& ,„ ...
T = — =10 = e (Eq. 5)
o
Note that Equation 5 contains five terms (I, I , a, c, and £) , all of which
can be measured. Furthermore, it is the ratio of I to I which is important
rather than the absolute values of those quantities.
Frequently the Lambert- Beer law is written in logarithm form, with the
term absorbance, A, often used to represent -log T:
A H -log T = ac£
When the absorbance (i.e., the difference between I and I ) is very small (as
is typically the case when measuring the absorbance of sub-ppm levels of
ozone) , the Lambert-Beer law can be approximated in a linear form.
1-5
-------
1. THEORY'/Physical Basis of the Photometry Equations
I -ctc£ ,
T = — = e *» 1 - ac£
o (Eq. 6)
(e~X = 1 - x for x « 1)
Note that this approximation is identical to Equation 3. The error introduced
by this approximation is not significant for ozone concentrations smaller than
~1 ppm (assuming a pathlength of <1 m), as will be shown below.
SOURCES OF ERROR INVOLVED IN THE PRINCIPLE
The relative error in the concentration measurement is related to the
relative error in the determinations of a, H, and T by the equations
dc dot . .
— = - — (assuming no error in X, or T)
C d
dc d£ , . _.
— = - —- (assuming no error in a or T)
C A*
dc dT . .
— = - .. - (assuming no error in a or £)
The total error is the sum of these contributions.
The length of the optical path through the sample can normally be measured
in a straightforward manner. The relative error associated with X, should not
exceed ±0.5%, and in most systems is probably no larger than 0.1%. The ab-
sorptivity of a given species is measured in separate experiments and is
normally available in the literature. The absorptivity of ozone at 254 nm has
been determined by several workers (Table 1-1). A review of the literature
(Hampson 1973) has placed the value of a at 308 ± 4 atm cm (to base e) or
134 ± 2 atm cm (to base 10). The absorptivity is normally given at standard
temperature and pressure (STP) (273 K and 760 torr); the value used in a given
experiment must be corrected via gas law calculations to reflect the pressure
and temperature of the sample being measured.
1-6
-------
1. THEORY/Sources of Error Involved in the Principle
TABLE 1-1. OZONE ABSORPTIVITY
Investigator(s), Year
a (atm cm , base e)
Method
Inn and Tanaka 1953
Griggs 1968
Becker et al. 1974
Hearn 1961
DeMore and Raper 1964
Clyne and Coxon 1968
306.2
303.9
310.8
308.5
310.8
313.2 (250 nm)
Manometry
Manometry
Manometry
Decomposition
stoichiometry
Decomposition
stoichiometry
GPT
To make the transmittance measurement, the absorption cell is filled
with a reference gas (zero air in the case of ozone), and the intensity of the
light passing through the cell is recorded as I . The absorption cell is then
filled with the sample, and the intensity of the light passing through the cell
is recorded as I.
The error in the measurement of T is multiplied by the term 1/lnT. For
a typical value of T of 0.99385 (0.2 ppm 0 ; 100 cm path), this term is ~160.
Thus the transmittance measurement must be accurate to 1 part in 16000 to
obtain a concentration accurate to 1% under the given conditions. Depending
on one's choice of hardware, the transmittance measurement can be made with
an accuracy of 1 part in 10 or better.
1-7
-------
SECTION 2
STEP-BY-STEP DISCUSSION OP THE ULTRAVIOLET PHOTOMETRIC
CALIBRATION PROCEDURE
This chapter provides explanation, augmentation, and general guidance to
assist the user in carrying out each step of the ultraviolet (UV) photometric
calibration procedure. The identification numbers and headings correspond
directly to those used in the procedure. The complete procedure is presented
as Appendix A and should be referred to as appropriate while reading the fol-
lowing step-by-step discussion.
1. Principle. This section discusses the principle of UV absorption
photometry upon which the calibration procedure is based. Note that the con-
centration unit in Equations 1 and 2a in Appendix A is "atmospheres." A
concentration of 1 atmosphere is equivalent to the concentration of a gas at
a pressure of 1 atmosphere and a temperature of O° C. The unit is used here
to be consistent with the corresponding unit for the absorption coefficient
— 1 —1
a, which is given in the literature in atm cm . In Equation 2b in Appendix
6
A, note that multiplication by 10 transforms the unit from atmospheres to
parts per million, the familiar dimensionless volume ratio concentration unit.
The minus sign in Equations 2a and 2b in Appendix A is needed because I should
always be less than I ; therefore, I/I is less than 1 and the natural loga-
rithm (In) of I/I is negative.
2. Applicability. Basically, the procedure is designed for calibration
of ambient 0 monitors; in many ways, it resembles calibration procedures for
SO , NO , or CO ambient monitors. There is one very important difference,
^ ^
however. Calibration procedures for SO , NO , and CO analyzers call for ex-
ternally obtained concentration standards traceable to primary standards
2-1
-------
2. STEP-BY-STEP DISCUSSION/'Applicability
(Standard Reference Materials) obtained from the National Bureau of Standards
(NBS). No Standard Reference Material is available for 0 . Therefore, any
0. calibration procedure must include a built-in means for obtaining 0 con-
centration standards based on some non-0 standard. In the photometric
calibration procedure, 0 concentration standards are related to the absorp-
tion coefficient of O at 254 nm, as discussed in Section 1. Thus the pro-
cedure is more than a calibration procedure — it is also a means for obtaining
0 concentration standards. Moreover, because the photometric procedure is
prescribed by EPA regulations (U.S. Environmental Protection Agency 1979), and
because no more authoritative O standards are available from NBS, the 0 con-
centration standards obtained by this photometric procedure are tantamount to
primary ozone standards.
Because the photometric procedure is so important in obtaining primary
O standards, the procedure itself is sometimes called a "UV standard" for
0 concentrations. This is simply an idiomatic expression referring, of
course, to the 0 concentrations obtained by means of the photometric proce-
dure.
Note that the calibration procedure may also be used to certify transfer
standards, which can then be used to calibrate ambient 0 monitors. Such
transfer standards must meet certain requirements and must be used properly.
Requirements for transfer standards and guidance in their use are given in a
companion EPA Technical Assistance Document (McElroy 1979).
3. Apparatus. The UV calibration system must include an 0 generator,
an output port or manifold, a photometer, a source of zero air, and whatever
other components are necessary to provide a stable 0 concentration output.
The common configuration shown in Figure 1 of Appendix A is the basis for the
procedural description. Another similar configuration uses a slightly pres-
surized system to eliminate the need for the pump to draw sample flow through
the photometer. These configurations have been used successfully and are
2-2
-------
2. STEP-BY-STEP DISCUSSION/Apparatus
probably near-optimum. Other variations are not precluded but would require
critical evaluation for possible adverse effects.
Variations in the configuration may be inevitable to accommodate different
photometers, various types of analyzers or transfer standards, and alternate
O generators. In particular, transfer standards containing their own
sources of O may require modification of the configuration; see McElroy
(1979) for guidance. Any variations must be evaluated carefully to insure no
adverse effects to the accuracy of the system.
0 is highly reactive and subject to losses upon contact with surfaces.
All components between the O generator and the absorption cell should be made
of glass, Teflon, or other nonreactive material. Lines and interconnections
should be kept as short as possible, and all surfaces must be very clean.
Also, systems may need to be "conditioned" (operated with O flowing at
maximum concentration) for 10 to 30 minutes initially and sometimes before
each use to minimize losses of 0 .
It is advantageous that both the output manifold and the photometer cell
be at atmospheric pressure. Hence, components, lines, connections, and vents
located between those items should be large enough to avoid any significant
pressure drop at the highest flow used. Note the vent located between the
zero air supply and the two-way valve. This vent is used to insure identical
pressure conditions in the photometer cell during both modes of the valve.
3.1 UV photometer. The photometer is discussed in detail in Section 3.
3.2 Air flow controllers. The photometric measurement of absorption is
not directly related to flow rate but may be indirectly related due to thermal
or other effects. Thus, equal flow through the photometer cell during the two
modes of the two-way valve may be important. The output of the 0 generator
is affected by flow changes, so good flow regulation for F is necessary.
2-3
-------
2. STEP-BY-STEP DISCUSSION/Apparatus
Adequate flow regulation can be obtained by means of a pressure regulator and
a needle valve or capillary, but more sophisticated flow control devices would
certainly be acceptable.
3.3 Ozone generator. The 0 generator must be very stable over short
periods to facilitate the sequential photometric measurement of I and I, and
o
to allow for stabilization of the analyzer or transfer standard connected to
the output manifold. Conventional UV-photolytic type generators are adequate
but should have line voltage and temperature regulation if possible. Mechan-
ically or electrically adjustable generators are most convenient.
Where the O generator is part of a transfer standard, special considera-
tion is needed; see McElroy (1979).
3.4 Output manifold. The output manifold serves the important function
of providing an interface between the calibration system and other devices or
systems that utilize the output 0 concentrations. The manifold must have one
or more ports for connection of such external devices or systems, and it must
be designed so that all ports of a multiport manifold provide identical
concentrations. A very important part of the output manifold is the vent,
which exhausts excess gas flow from the system and insures that the manifold
outlet port or ports are kept at atmospheric pressure for all flowrates. The
vent must thus be large enough to avoid any appreciable pressure drop and must
be located sufficiently downstream of the output port to insure that no ambient
air enters the manifold due to eddy currents, back diffusion, etc. The mani-
fold may range from an elaborate, specially fabricated multiport fixture to a
simple ordinary "T" fitting where one of the legs is used as a vent. The
manifold must be made of glass, Teflon, or some other inert material and
should be kept as clean as possible to avoid 0, loss.
3.5 Two-way valve. This valve (often called a three-way valve) is used
to conveniently switch the flow through the photometer cell from zero air (for
the I measurement) to manifold gas (for the I measurement). The valve is
2-4
-------
2. STEP-BY-STEP DISCUSSION/'Apparatus
automatically controlled by automatic photometers. For manually operated
photometers, the valve is manually controlled and may be an electric solenoid
valve controlled by a switch or a manually operated mechanical valve. In lieu
of a valve, the plumbing lines can be manually rerouted each time a new measure-
ment is needed; however, this is quite inconvenient because a number of repeti-
tive measurements of I and I are normally taken for each concentration assay.
o
3.6 Temperature indicator. This indicator is needed to measure the
temperature of the gas in the photometer cell in order to calculate a tempera-
ture correction. In most photometers, particularly those whose cell is enclosed
inside a case or housing with other electrical or electronic components, the
cell operates at a temperature somewhat above ambient room temperature. There-
fore, it is important to measure the actual temperature of the cell, not room
temperature. Ideally, the temperature of the gas inside the cell should be
determined. For thin metallic cells, however, measurement of the external cell
wall temperature may be sufficiently accurate for practical purposes. Good
results have been obtained by simply taping an ordinary certified mercury
laboratory thermometer to the cell. However, reading such a thermometer inside
enclosures or with covers in place may be difficult or impossible. Hence, a
small thermocouple or thermistor connected to an external readout device may
be the best approach. The thermocouple may be welded, or the thermistor glued,
to the cell wall or even inserted through the cell wall to measure interior
cell temperature. In locating the point of temperature sensing, attempt to
select a point which is representative of the average cell temperature. For
example, one end of the cell may be close to the source lamp and thus some-
what warmer than other parts of the cell. A thermocouple or thermistor tem-
perature indicator should be occasionally checked against a certified thermo-
meter or other temperature standard to insure accurate temperature readings.
3.7 Barometer or pressure indicator. The barometer or pressure indicator
is needed to measure the pressure of the gas in the cell in order to calculate
a pressure correction. Most photometer cells operate at atmospheric pressure
2-5
-------
2. STEP-BY-STEP DISCUSSION/Apparatus
Thus, if there are no restrictions between the cell and the output manifold,
the cell pressure should be very nearly the same as the local barometric
pressure, and a certified local barometric pressure reading can be used for
the pressure correction. If the cell pressure is different than local baro-
metric pressure, some means of accurately measuring the cell pressure (such as
a manometer, pressure gauge, or pressure transducer) is required. Any such
device should be calibrated against a suitable pressure standard to insure
accurate pressure measurements.
4. Reagents
4.1 Zero air. Zero air can be obtained either from compressed gas cyl-
inders or from ambient air. Since appreciable volumes of air are required,
there may be practical limits to the use of cylinder air. When the source of
aero air affects the output of a transfer standard, special consideration is
needed; see McElroy (1979).
As required by the procedure, zero air must be free of 0 and any other
substance that might react with 0 (e.g., NO, NO , various hydrocarbons, and
particulates). Air from any source must be purified to remove such substances.
Very dirty air may require a pre-cleaning process to remove large particles,
oil mist, liquid water, etc. The primary purification process is based on
mechanical and chemical filtering. While various schemes may be acceptable,
systems similar to the following have been used successfully: The air is
first dried with a Perma-Pure type dryer followed by a column of indicating
silica gel. The air is then irradiated with a UV lamp to generate O that
converts existing NO to NO_, and a large column of activated charcoal removes
NO , 0 , hydrocarbons, and various other substances. If desired, molecular
£ -J
seive can be included for good measure. A final particulate filter removes
particulates which can originate in the scrubber columns. The removal of
moisture may not be necessary, but fewer problems seem to be encountered when
dry air is used.
2-6
-------
2. STEP-BY-STEP DISCUSSION/Reagents
The particulars of air purification are not well known, so some experi-
mentation may be necessary to determine the appropriate size or volume for
scrubber columns. Also, the capacity of the scrubber materials may not be
accurately known, so frequent replacement or renewal of these materials is
advisable. Additional information on air purification is available (APHA
Intersociety Committee 1977, Section 20, Part I).
A very important requirement in photometer operation is the need for the
zero air supplied to the photometer during the I measurement to be obtained
from the same source as that used for generation of O . The impurities pre-
sent in zero air from different sources can significantly affect the trans-
mittance of an air sample. This requirement presents no problem if the con-
figuration shown in Figure 1 of Appendix A is used. However, there may be
problems in certifying transfer standards that have their own sources of zero
air. This situation is discussed in more detail in McElroy (1979).
5. Procedure
5.1 General operation. As noted in the procedure, a photometer used for
calibration must be dedicated exclusively to calibration service and specifi-
cally not used for ambient monitoring. The reason for this requirement is
that the photometric readings are used to assay the O standards to be used
for calibration, and thus the photometer must be intrinsically accurate. If
the photometer is used for ambient monitoring, the cell will eventually
become dirty. A dirty cell can lead to 0 losses and erratic readings, which
result in loss of accuracy. Reserving the photometer for calibration where
only clean, filtered gas passes through the cell will minimize this loss of
accuracy.
O analyzers to be calibrated are normally located at various widely
separated field sites. While a UV photometer and the photometric calibration
procedure can certainly be used at each field site to calibrate such analyzers,
2-7
-------
2. STEP-BY-STEP DISCUSSION/Procedure
a number of advantages may be realized by locating a single UV photometer at
a central laboratory where it can remain stationary, be protected from the
physical shocks of transportation, be operated by a responsible analyst under
optimum conditions, and serve as a common standard for all analyzers in a net-
work. Under this concept, the central photometer is used to certify one or
more 0 transfer standards which are carried to field sites to calibrate the
0 monitors. The advantages as well as inevitable disadvantages of this con-
cept are discussed in McElroy (1979), which also provides helpful information
and guidance in the use of O transfer standards.
5.2 Preparation. Since the accuracy of the calibration standards ob-
tained by this procedure depends entirely on the accuracy of the photometer,
it is very important to insure that the photometer is operating properly and
accurately. The fact that the photometer makes a ratio measurement (i.e.,
I/I ) rather than an absolute measurement eases this task. The checks de-
o
scribed in this section, if carried out carefully, provide reasonable con-
fidence that a photometer having the required inherent capability is operating
adequately.
A well designed and properly built photometer is a precision instrument;
once shown to operate adequately, it will likely continue to do so for some
time, particularly if the photometer is stationary and used intermittently
under ideal laboratory conditions. Thus, the performance checks may not
necessarily have to be carried out each time the photometer is to be used.
The actual frequency of the checks is a trade-off between confidence in the
photometer performance and the cost and effort to carry out the checks, and
is thus a matter of judgment. One reasonable approach suggested by the proce-
dure is to carry out the checks very frequently with a new photometer, keeping
a chronological record of each performance check (similar to a quality assurance
control chart). If the record of the photometer performance shows continued
adequacy and reliability, the frequency of the checks can be reduced and still
maintain adequate confidence in the photometer. (On the other hand, the record
2-8
-------
2. STEP-BY-STEP DISCUSSION/Procedure
may indicate the need for continued frequent verification of the system condi-
tion.) Even where the record shows excellent stability, the checks should
still be carried out at some minimum frequency (e.g., once every 3 or 4 weeks)
because the possibility of malfunction is always present. A regular schedule
of checks will avoid the risk of losing long periods of data due to photometer
malfunction.
5.2.1 Instruction manual: If the photometer is a commercially manu-
factured one, it should include an operation/instruction manual. This manual
should be studied thoroughly and its recommendations followed carefully and
completely. If questions arise or problems are encountered, assistance is
available from the manufacturer or, if necessary, from an EPA technical
assistant.
5.2.2 System check: A visual inspection of the photometer system should
be conducted to verify that the system is in order. The configuration and
plumbing connections are first compared to the flow diagram. All connections
are verified to be sound and not restricting the flow. Any obvious or possible
leaks are corrected. Cleanliness of cells, manifold, and lines is checked.
A more thorough check for leaks follows this visual inspection. One method
is to block the output ports and measure the inlet and outlet flow with an
external flowmeter. One can also measure the outlet flow and compare the
reading to the system's flowmeter reading. This also serves to check the
system's flowmeter. The two-way valve is checked to insure that it doesn't
leak. Flowrates and vents are checked for lack of back pressure. The zero
air supply components should be serviced periodically by replacing or renewing
consumables and by checking or replacing filters.
5.2.3 Linearity test: Because the required photometric measurement is
a ratio measurement, a linearity check of the photometer is a good indication
of accuracy. The photometer operates over a very narrow range of absorbance,
2-9
-------
2. STEP-BY-STEP DISCUSSION/Procedure
so nonlinearity is unlikely. Unless the manufacturer has demonstrated or
guaranteed a photometer linearity error of less than 3%, however, the user
should carry out a linearity test.
A linearity test is conducted by first generating and assaying an O. con-
centration near the upper range limit of the system (probably 0.5 or 1.0 ppm).
Then the concentration is diluted with a configuration similar to that shown
in Figure 2 of Appendix A. A flow of zero air is added to the original
generated concentration, and the mixture is passed through a mixing chamber to
insure a homogeneous concentration at the output manifold. For this test, the
dilution ratio, R, must be accurately known; hence the flowrates F and F
must be accurately measured to <2%. To help insure accurate flow measure-
ments, the two flowmeters should be of the same general type and one should
be standardized against the other. The dilution ratio is calculated as the
flow of the original concentration (F ) divided by the total flow (F +
Vs
With stable, high resolution flowmeters and careful work, R should be accurate
to <
When F has been adjusted and R has been calculated, the diluted concen-
tration is assayed with the photometer. The diluted assay (A ) is then com-
pared with the original undiluted assay (A ) by calculating the percent
linearity error according to Equation 3 in Appendix A. This linearity error
must be less than 5% and should be less than 3% for a well performing system.
Note, however, that the result is not really the true linearity error because
it includes possible errors in the flow measurements; the test serves only as
an indicator. If the linearity error exceeds 5% or is greater than expected,
2-10
-------
2. STEP-BY-STEP DISCUSSION/Procedure
the accuracy of the flow dilution should be checked and verified carefully
before the photometer is assumed to be inaccurate. The test should be carried
out several times at various dilution ratios, with an averaging technique used
to determine the final result. If any modifications to the UV system are
necessary in performing the linearity test, care should be exercised to avoid
introducing leaks or other adverse effects.
If the linearity error is excessive and cannot be attributed to flow
measurement inaccuracy, then the photometer system should be checked for:
1. dirty or contaminated cell, lines, or manifold;
2. inadequate "conditioning" of system (see paragraph 3);
3. leaking two-way valve or other leak in system;
4. contaminant in zero air;
5. nonlinear detectors in the photometer;
6. faulty electronics in the photometer.
Also, UV system nonlinearity might be indicated when a nonlinear calibration
curve is obtained for an analyzer that is expected to be linear.
5.2.4 Intercomparison: A good check on the overall accuracy of a
photometric calibration system is occasional comparison with O standards
from other (independent) organizations. Such comparisons can be either direct
or, where the photometer system is stationary, by means of transfer standards;
see McElroy (1979). If both organizations' standards agree, it is very likely
that both are accurate. Where two organizations' O standards do not agree, it
may not be readily apparent which organization's standard is inaccurate. Never-
theless, the discrepancy indicates that at least one of the systems is inaccurate
and that further investigation is needed. Where possible, comparison with an
agency engaged in quality assurance may be particularly advantageous, because
of the usually greater authority of its 0 standards.
2-11
-------
2. STEP-BY-STEP DISCUSSION/Procedure
5.2.5 Ozone losses: In spite of scrupulous cleaning and preconditioning,
some O may inevitably be lost upon contact with photometer cell walls and gas
handling components. Any significant loss of 0 must be quantitatively deter-
mined and used to correct the output concentration assay accordingly. In any
case, loss of O must not exceed 5%.
Possibly the best way to determine O loss, after exhausting all pos-
sibilities of minimization, is as follows: First, a stable O analyzer is
*3
calibrated with the UV calibration system, assuming no losses. An 0. concen-
tration is then generated and measured with the analyzer as closely as possible
to the actual inlet of the photometer cell. Similarly, the concentration is
measured as closely as possible to the outlet of the cell. Each measurement
may need to be repeated several times to obtain a reliable average. The tests
should be repeated at several different concentrations of O . The concentra-
tion at the output manifold is also measured. Some disassembly of plumbing
fittings is usually necessary for these 0, measurements, as the best locations
for the measurement points are at the inlet and outlet fittings of the photo-
meter cell. For convenience, access fittings may be permanently installed at
these points to facilitate frequent loss checks.
Since a continuous flow of O through the photometric system is required
when making the measurements, the two-way valve must be in the "sample" posi-
tion. In systems with manually controlled valves, this presents no problem.
In systems with automatically controlled valves, some means of maintaining a
continuous flow of 0, into the absorption cell is necessary.
In making these measurements, it is important to avoid shock or damage to
the photometer and to reassemble the fittings properly and check for the
absence of leaks. Also, any pressure differences at the measurement points
may cause inaccurate measurements if the analyzer is affected by pressure
changes.
2-12
-------
2. STEP-BY-STEP DISCUSSION/Procedure
For the configuration shown in Figure 1 of Appendix A, the percent O
loss is calculated as
R -h (Rin + Rout)
Percent 0 loss = x 100%
3 R
m
where R. is the O_ concentration measured at the cell inlet, R ^ is the O_
in 3 out 3
concentration measured at the cell outlet, and R is the 0_ concentration
m o
measured at the output manifold.
For other configurations, the percent loss may have to be calculated
differently. The O loss correction factor for use in Equation 4 of Appendix
A is calculated as
L = 1 - 0.01 x percent O loss
5.3 Assay of O concentrations.
5.3.1 Photometers use lamps and electronic circuits which generate
some heat, and therefore most photometers will experience some temperature rise
from room temperature. It may be important to the stability and hence accuracy
of the photometer to wait until thermal equilibrium has been reached. The
process cannot usually be hastened. Thus, the system should be turned on some-
time before it is needed to allow adequate stabilization time. Since the
stabilization time may vary considerably from one photometer to another, this
may be a consideration when selecting a portable photometer for field calibra-
tions.
5.3.2 It is very important that the photometer not drift between the I
o
and I measurements. Therefore, these two measurements should be as close to-
gether in time as possible. However, delay between the two measurements is
needed to allow the cell to flush completely before the second measurement.
Thus cell flowrate is an important consideration. About 5 cell volumes should
2-13
-------
2. STEP-BY-STEP DISCUSSION/Procedure
pass through the cell to adequately flush it. To keep the flush time down to
10 seconds or less, the cell flow needs to be >30 cell volumes/minute. With
a cell volume of 30 cm , a flow of 2 liters/minute will provide 5 cell flushes
in ~5 seconds.
5.3.3 The flowrate into the output manifold must be greater than the sum
of the flowrates of the device or devices connected to the manifold. When the
flows are adjusted, it is a good idea to check the vent to make sure that
sufficient vent flow exists to prevent entry of ambient air.
5.3.4 Similarly, F must be adjusted to exceed the photometer cell flow-
z
rate, and the F vent should also be checked for adequate flow to assure no
z
possibility of ambient air entry.
5.3.5 The photometer is first checked by allowing it to assay a zero
0 concentration (zero air). Needless to say, if the photometer does not
indicate that 1=1, something is wrong. If this happens, perhaps the most
obvious thing to check is whether the manifold is really at zero concentra-
tion. Sometimes the O generator continues to generate small amounts of O
•J -J
even when it is supposed to be off. This can be checked by supplying zero
air directly to both inlets of the two-way valve, bypassing the O generator.
If the photometer still fails to give the same reading, the fault is somewhere
in the photometer, which must be checked and corrected. A very small dif-
ference between I and I is perhaps tolerable.
Note here that many commercially available photometers operate automati-
cally. In this type of photometer, the two-way valve is an integral component
and is controlled automatically by the photometer control system. Also, auto-
matic photometers generally cycle through the entire operational sequence
continuously and output a single measurement of the I/I ratio rather than
individual I and I measurements. In fact, the output reading will probably
o
be scaled in actual concentration units rather than an I/I ratio.
2-14
-------
2. STEP-BY-STEP DISCUSSION/'Procedure
5.3.6 The O^ generator is adjusted to produce an 0 concentration as
needed. Usually, the first concentration is near the upper range limit, about
0.45 or 0.9 ppm, as appropriate. It is important to allow plenty of time for
the 0 generator to stabilize at this setting and for the entire system to
"condition" to this concentration of 0 .
5.3.7 The two-way valve is actuated to sample zero air. It is important
to allow sufficient time to thoroughly flush the photometer cell and to obtain
a stable reading. The I value is measured and recorded.
5.3.8 The two-way valve is actuated to sample the O concentration.
Again, it is important to allow sufficient time for cell flushing and for a
stable reading. The I value is measured and recorded. Automatic photometers
perform this and the previous step automatically.
5.3.9 The temperature and pressure of the sample in the photometer cell
are recorded. Because conditions inside the cell may be different than ambient
(room) conditions, one must try to measure the actual temperature and pressure.
Paragraphs 3.6 and 3.7 (above) present a discussion of these parameters and
their measurement.
5.3.10 Equation 4 in Appendix A is used to calculate the 0 concentra-
tion. An average of several photometer readings should be used, as there is
typically some variation in the readings. With automatic photometers, this is
no problem because operation is continuous and new readings are outputted
after each cycle. For manually operated photometers. Steps 5.3.7 and 5.3.8
have to be repeated several times.
Equation 4 in Appendix A translates the photometer absorption readings
for I and I into 0. concentration in parts per million. The first term is
o 3
the primary Lambert-Beer relationship. The second and third terms are for
2-15
-------
2. STEP-BY-STEP DISCUSSION/Procedure
temperature and pressure corrections, respectively. The last term is the
conversion factor to parts per million and also includes the correction for 0.
losses in the system.
Each symbol used in the equation is explained below the equation. Note
that the absorption coefficient is given for a temperature of 0° C, so the
temperature correction term is important and cannot be neglected.
The optical pathlength of the photometer cell needs to be known very
accurately (±0.5%). If the photometer's operation manual does not indicate
the exact pathlength, the cell will have to be measured; some disassembly of
the cell may be required to do this. In purchasing a photometer, it might be
wise to insist that the manufacturer provide this information.
The correction factor for 0 losses, L, is discussed under paragraph 5.2.5.
Since the photometer readings I and I are treated as a ratio (I/I ), the
o o
units (which are identical) are not important. The calculation of the natural
logarithm (symbol "In") of the I/I ratio can be carried out by means of a
table of natural logarithms or with a calculator that has natural log capability.
Many small pocket calculators have an "In" key for this purpose. Be sure to
compute the natural logarithm and not the base 10 logarithm (symbol "log")
which would give an incorrect result. However, if your calculator has base 10
(log key) logarithms but not natural logarithms (In key), Equation 4 in
Appendix A can be rewritten as
[°3]OUT = (~1/BA log Vy (T/273) (760/P) (106/L)
where 3 = 134 atm cm
2-16
-------
2. STEP-BY-STEP DISCUSSION/Procedure
Automatic photometers may automatically evaluate all or part of Equation
4 in Appendix A. Usually, the photometer evaluates the first term and the
remaining terms — corrections for temperature, pressure, and 0 losses — have
to be applied manually. Caution: some photometers may be set to evaluate the
first term of Equation 4 in Appendix A based on a temperature other than O° C,
in which case the "273" in the temperature term would have to be replaced by
the proper temperature.
Some automatic photometers use a linear approximation rather than the
logarithmic form for the first term of Equation 4 in Appendix A. For a path-
length of ~70 cm, this results in an error of ~1% at an 0 concentration of 1
ppm. For longer pathlengths, the error becomes more severe. At concentrations
of less than 0.5 ppm, the error is less than 0.5% and can probably be ignored.
But at higher concentrations or longer pathlengths, the error may become
significant, in which case the error must be determined and an appropriate
correction applied to the resultant concentration.
5.3.11 Additional O concentration standards are obtained by repeating
the appropriate steps between 5.3.6 and 5.3.10 or by Option 1, which is dis-
cussed later.
5.4 Certification of transfer standards. All transfer standards must
be certified (related) to primary O concentration standards using this cali-
bration procedure. Transfer standards that assay O concentrations generated
externally can be certified very similarly to the calibration of an O ana-
lyzer. But special arrangements must be made for transfer standards which
include their 01
McElroy (1979).
include their own source of O . All of these matters are discussed in
5.5 Calibration of ozone analyzers. The ozone analyzers designated
here are ambient monitors installed at field monitoring sites. Ambient monitors
are normally calibrated in situ without disturbing the normal sampling setup,
2-17
-------
2. STEP-BY-STEP DISCUSSION/Procedure
except for transferring the sample inlet from the ambient sampling point to
the calibration system. The steps in Section 5 of the procedure should be
followed regardless of whether the ambient monitor is being calibrated directly
by the photometric calibration procedure or by means of a transfer standard.
5.5.1 Warm-up: Before calibration, a newly installed 0 analyzer should
be operated for several hours (preferably, overnight) to permit stabilization.
Brand-new analyzers fresh from the factory may require several days of opera-
tion to fully stabilize. The photometer or transfer standard must also be
allowed an adequate warm-up and stabilization period before use, particularly
if stored or transported in cold weather.
5.5.2 In adjusting the analyzer zero, it is important that the input
concentration is truly zero (i.e., zero scrubbers recently serviced and in
good condition, and O generator off). Remember that 0 generators sometimes
generate small amounts of 0 even when the mechanical adjustment is in the
off position. Zero air obtained from the calibration system rather than from
any zero air source internal to the analyzer should be used. After a stable
reading has been achieved, the zero on the data recording device (not the ana-
lyzer's meter indicator) is adjusted. The data recording device is usually a
chart recorder, but if a data acquisition system or other means is used, zero
the analyzer according to the indication of the data system or other device.
When using a chart recorder (or other device which cannot read below zero),
an offset zero — 5% of scale is convenient — insures a "live" zero (i.e.,
the recorder can respond to either an increase or decrease in the zero base
line).
5.5.3 An O concentration standard of ~90% of the range of the analyzer
to be calibrated — ~0.45 or ~0.9 ppm for the 0.5 and 1.0 ppm ranges, respec-
tively — is generated. If the photometric calibration procedure is being used,
this concentration standard is determined according to Steps 5.3.6 through
5.3.10. When a transfer standard is used, the concentration is determined by
2-18
-------
2. STEP-BY-STEP DISCUSSION/Procedure
reference to the transfer standard's certification relationship (with appli-
cable corrections if necessary); see McElroy (1979).
5.5.4 When the analyzer reading has stabilized, the analyzer's span
control is adjusted to provide a convenient scale factor. An offset zero
(if used) must be taken into account according to Equation 5 in Appendix A.
Also, the span and zero controls may be interrelated, so if the span adjust-
ment is large, the zero adjustment (Step 5.5.2) should be repeated, which
requires that the span adjustment (Step 5.5.4) also be repeated.
5.5.5 A complete calibration includes five or more upscale calibration
points, so additional 0 concentration standards need to be generated and
sampled by the analyzer. The additional concentrations may be generated
either by adjusting the 0 generator or by Option 1 (discussed later). After
the span adjustment on the first concentration, no further adjustments to the
span control are made. The additional concentrations should be approximately
evenly spaced over the scale range of the analyzer. The analyzer response to
each concentration is recorded together with the concentration.
5.5.6 The concentrations and associated analyzer responses are plotted
on graph paper and connected with a smooth line to form the analyzer calibra-
tion curve. For linear analyzers, the line should be straight. If possible,
the line should be determined by least squares regression and plotted according
to the calculated slope and intercept. After establishment of the calibration
curve, the span control should be locked, if possible, and not adjusted until
the next calibration. The calibration curve is then used (either visually or
by calculation) to translate all subsequent analyzer measurements to concen-
tration units until the time of the next calibration.
If zero and span checks are going to be used to check the analyzer between
calibrations, a zero and span check should be carried out immediately following
2-19
-------
2. STEP-BY-STEP DISCUSSION/Procedure
the calibration. The readings thereby obtained become reference readings which
are compared to subsequent zero and span checks to detect possible analyzer
drift.
5.5.7 Option 1; Instead of adjusting the O generator to obtain various
0 concentration standards, such standards may be obtained by dilution of a
high standard with zero air. This option obviates the need for an adjustable
0 generator but requires additional apparatus to effect the dilution, as
shown in Figure 2 of Appendix A. Note that a mixing chamber is required to
insure thorough mixing of the two gas streams. Since the diluted concentra-
tion standards are calculated based on the dilution ratio, accurate flow
measurements of both the primary O generator flow (F ) and the dilution flow
(F ) are required. Also, since the total flow may be increased substantially,
verify:
1. that the zero air supply and purification system can adequately
handle the extra flow;
2. that the pressure in the manifold or in the O generator does
not increase significantly due to increased back pressure at
the manifold vent;
3. that the flow through the 0 generator (F ) does not change and
thereby cause a change in the original (undiluted) 0 concentration.
Equation 6 of Appendix A is used to calculate the diluted O concentrations.
2-20
-------
SECTION 3
OBTAINING A PHOTOMETER
This section discusses some of the considerations to be addressed if one
wishes to construct a photometer suitable for the assay of sub-ppm O concen-
trations in clean systems. It also describes commercially available photo-
metric systems.
CONSTRUCTION OF A PHOTOMETER
Successful construction of a photometer for O assay requires a clear
understanding of the parameters to be measured and the magnitude of the error
in concentration resulting from each error associated with mismeasurement of
different parameters.
Reviewing the basic photometric equation from Section 1,
I -aci -1,1 A
T = — = e ; c = —r log — = —
it should be clear that there are four parameters (I, I , a, and H) which must
o
be determined to calculate the concentration, c. The value of a is obtained
from a review of the literature and is 308 atm cm . This absorption coef-
ficient is considered accurate to ±1.5% when applied to O at 254 nm and
•J
STP (0° C and 760 torr). When assaying O at temperatures and pressures
other than STP, the temperature and pressure of the sample must also be
measured to permit proper calculation of the concentration.
3-1
-------
3. OBTAINING A PHOTOMETER/'Construction
From Section 1,
dc _ _ da
c a
Thus, a given error in a causes a multiplicative error in c of the same magni-
tude but in the opposite direction. That is, if the literature value of a is
1% too large the calculated concentration will be 1% too small. Also, a given
error in the pathlength, £, causes the same kind of error in c as does an
error in a.
The transmittance, T, is the parameter which must be measured with the
greatest precision, because the error in T is multiplied by the term 1/lnT:
m
T = e
dc dT
c TlnT
For T = 0.98, this term is -50; for T = 0.997 (100 ppb O ; 1 m path) the term
increases to ~325. Thus, for T near 0.997, one must measure T with more than
300 times the relative accuracy of the measurement of a or £ in order to keep
the error in c to the same level as that caused by the errors in a and H.
Also, note that the inaccuracy caused by errors in T depends strongly on the
concentration (i.e., value of T), whereas errors caused by mismeasurement of
a and H are constant multiplicative factors. For typical values of H of 1
and 2 m, the respective values of T (assuming c = 500 ppb) are 0.984718 and
0.969669. If the photometer can determine T with an error of 5 parts in
10 , the respective errors in c are 0.3297% and 0.1674%.
Measurement of T is related to many (often conflicting) design features.
For example, increasing £ decreases T so a given error in T causes a lesser
error in c. However, increasing Si generally leads to a longer instrument
3-2
-------
3. OBTAINING A PHOTOMETER/Construction
cycle time (due to the increased time period required to flush the longer
cell); the longer cycle time allows more electronic drift and thermal effects
so that the error in T generally increases. Attempting to flush the cells
faster also leads to greater error in T due to scintillation from turbulence
in the gas. Decreasing the cell diameter leads to a shorter flush time but
also a less favorable surface-to-volume ratio, and O losses may become a
significant problem.
The precision in the measurement of T is generally limited by the elec-
tronic and mechanical stability of the photometer. Therefore, the photometer
designer should attempt to use a rather short instrument cycle time (10 seconds
to a minute or so) in order to minimize these drifts. As an example of the
kind of problem one might encounter while examining instrumental drifts, con-
sider the effects of a detector whose sensitivity changes by 2% per month (a
large but not-unheard-of drift):
= — = °-0028% = 0.001%
month x 30 days/month x 24 hours/day hour 20 minutes
System drift due to the effects of one detector is 1 part in 10 in ~20
minutes. Clearly, if the design goal is to measure transmittance to 1 part
in 10 , the instrument cycle time will have to be much shorter than 20 minutes,
because a typical photometer includes more than one detector as well as some
amplifiers and voltage-to-frequency converters contributing to electronic
drift.
In order to obtain O measurements of the highest accuracy, it is important
that the photometric system operate at a single wavelength (i.e., the system
should be monochromatic). Although the problem is not widely discussed, many
photometric systems used to measure O concentration utilize some rather
fortuitous circumstances to achieve a (near) monochromatic photometer. The
low pressure mercury source emits radiation at a number of wavelengths other
than 254 nm. This can lead to at least two problems. The 185 run radiation
3-3
-------
3. OBTAINING A PHOTOMETER/Construction
from the source can produce O in the photometer cells (indeed, it is this
radiation which is used to produce 0 in most photolytic 0 generators).
Wavelengths which may ordinarily not produce O but are sensed by the detector
system cause errors unless the absorption coefficient for 0 is the same at
these wavelengths as at 254 nm.
In practice, one can construct a photometer which minimizes these prob-
lems. Use of a Vycor envelope on the source will filter out the 185 nm
radiation, since Vycor has a very low transmittance at 185 nm. Use of solar-
blind detectors eliminates interferences from radiation of wavelengths greater
than about 300 nm. In the 200 to 300 nm region passed by the Vycor and sensed
by the detectors, the mercury source has a dozen or so lines. Of these lines,
the 254 line is on the order of 100 times stronger than any other line. Rea-
sonably careful calculations of the effect of using a less than ideally
monochromatic system indicate that one can expect to underestimate the 0
concentration by ~0.5%. Limited attempts to measure the effect produced
results of 0.5 ± 0.5% (Paur and Ellis, unpublished data).
Another area of design which can greatly affect the precision of photo-
metric measurements involves signal processing to eliminate or reduce electrical
noise picked up in signal cables. In modern instrumentation this problem is
often addressed by use of current-to-frequency circuitry that converts the
detector current to a digital frequency very early in the signal processing
process. The digital frequency is relatively immune to extraneous noise be-
cause the counters or digital integrators only need to discriminate between
the presence or absence of a pulse rather than trying to measure the amplitude
of a d.c. signal with a precision of 1 part in 10 .
3-4
-------
3. OBTAINING A PHOTOMETER/'Construction
The intensity of the 254 nm radiation from low pressure mercury sources
is very sensitive to the temperature of the source. Even though a photometer
typically uses a reference detector to monitor the source intensity, the pre-
cision of the transmittance measurement can generally be significantly improved
by maintaining the source at a constant temperature. The temperature is typi-
cally in the 45 to 50° C range, since the source is generally some 4 to 6
times brighter when operated at 50° C than when operated at 20° C.
It is useful to design the source holder, detector holder, amplifier, and
voltage-to-frequency converters into a package that is thermostated to a con-
stant temperature. Ideally, the detectors and electronics should operate at
less than 50° C; however, solar-blind photodiode detectors typically have low
dark current even at 50° C, and almost all modern electronic components are
specified to at least 70" C.
As indicated in Section 2, the material of construction is of considerable
importance because O is relatively easily decomposed by contact with most
materials. The materials most commonly used are glass and Teflon or other
fluorocarbon materials like Kynar. Some metal surfaces such as aluminum may
eventually be passivated, but use of such materials is generally considered to
be quite detrimental to the reliability of the measurements. (Some metals,
such as hot copper turnings, make excellent O scrubbers and can be used to
destroy 0 from laboratory apparatus before discharging vent gases to the
atmosphere.)
Although the O generator is not part of the photometer, the reader is
cautioned to independently characterize the 0_ generator so that generator
instabi!
system.
instability will not be interpreted as instability in the O measurement
3-5
-------
3. OBTAINING A PHOTOMETER/'Commercially Available Systems
COMMERCIALLY AVAILABLE PHOTOMETRIC SYSTEMS
Dasibi Ozone Analyzer (Models 1003-AH and 1003-PC)
The Model 1003-AH differs from the Model 1003-PC only in that the latter
includes a self-contained O generation system; both will be referred to as
"Model 1003." It is important to distinguish among three different uses of
the Dasibi Model 1003 0 analyzer. These are discussed below.
Ambient Ozone Monitor—
This Technical Assistance Document specifically does not address the use
of the Dasibi O analyzer for measurement of ambient O .
"Transfer Standard" for Ozone Monitor Calibration—
When used as a transfer standard, the Dasibi O analyzer is periodically
certified in the laboratory and is assumed to give stable results between
certifications. The analyzer is taken to field sites and used to determine
the O concentration in a calibration system during calibration of the field
O monitor. Since the Dasibi analyzer is operating in a clean system, the
possibility of interference is significantly reduced. This document provides
limited information on use of Dasibi analyzers as transfer standards; see also
McElroy (1979).
Photometer—
The Dasibi Model 1003 O analyzer is easily modified to operate as a
photometer which measures the relative transmittances of two gas samples at
254 nm. If one gas sample is zero air and the second sample contains O in
zero air, the concentration of O in the latter sample can be readily calculated
from the measured transmittance of the sample, the known absorptivity of O at
254 nm, and knowledge of the optical pathlength through the sample. It is
this last use of the Dasibi Model 1003 O analyzer — its use as a convenient
hardware package for making photometric measurements at 254 nm — that is of
interest here.
3-6
-------
3. OBTAINING A PHOTOMETER/'Commercially Available Systems
In order to help the reader appreciate the significance that O measure-
ments carried out using a Dasibi photometer have for the photometric technique
in general, some of the important details of the Model 1003 hardware will be
described.
The Dasibi instrument uses a low pressure mercury discharge lamp as its
light source. The 185 nm 0 -producing radiation is blocked by means of a
Vycor shield around the lamp. The solar-blind vacuum photodiode detectors do
not respond to radiation of wavelengths greater than about 300 nm. The
source-detector combination thus produces an effectively monochromatic photo-
meter operating at the 254 nm Hg line.
The Dasibi is a two channel photometer system (see Figure 3-1). One
detector (sample channel) monitors the intensity of 254 nm radiation passing
through the absorption cell. A second detector (control channel) positioned
near the UV source views a fraction of the 254 nm radiation. This second
detector functions as an energy monitor and is used to compensate for changes
in source intensity.
The electrometers are current-to-frequency converters; signals from the
detectors are converted to digital signals very early in the signal processing
and are thereby protected from further degradation. The digitized signals are
fed into a patented counting system which also controls a solenoid valve
allowing either zero air or sample air to pass through the absorption cell.
During an instrument cycle, the solenoid valve directs zero air into the
absorption cell and flushes the cell ~2.5 times (~5 seconds with ~2 liters/
minute of air passing through a 66 cm cell). Two independent counters are
actuated first in a count-up mode. The control counter digitally integrates
the intensity of light entering the control channel while the sample counter
integrates the light intensity emerging from the sample cell. Both counters
are stopped simultaneously when the sample channel counter reaches a preset
span number. The solenoid valve is switched to direct sample air into the
3-7
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
0)
N
0)
c
o
oo
O
O
H
OJ
•a
-H
A
-H
M
(d
Q
M-l
O
S-l
tJl
m
•H
•a
u
o
0)
•H
in
•H
rH
i1
-H
ro
3
tn
•H
CM
3-8
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
absorption cell, which is flushed for ~5 seconds. The counters are then
activated in a count-down mode. When the control channel counter has accu-
mulated the same number of counts as it did during the zero reference part of
the cycle (i.e., when the control counter has counted down to zero), both
counters are stopped. The count remaining in the sample counter represents
the quantity I - I and is shifted into the instrument output display.
The Dasibi Model 1003 uses the linear approximation of the absorption
equation to convert the measured transmittance to a concentration:
^- = 1 - ac£
1 = 1 - I acX,
o o
I -1=1 acJl
o o
Since the Dasibi uses current-to-frequency signal processing, I can be
o
assigned an arbitrary value, and the measurement period can be adjusted to
allow the sample counter to reach the value of I . Dasibi chose a value of I
* oo
such that at 1.0 ppb O the difference between I and I would be 10 counts.
At 1.0 ppb ozone
I - I = 10 counts = I
o o
_ 10 counts
o ac£
where a = 308.3 atm cm at STP (273 K and 760 torr), base e
= 277.6 atrn'^m"1 at 300 K (27° C) and 752 torr
£ = 71 cm
c = 10 atmosphere or 1.0 ppb
Thus, I = 456,844 at STP
o
= 507,367 at 300 K and 752 torr
3-9
-------
3. OBTAINING A PHOTOMETER/ 'Commercially Available Systems
This value of I is the span number one would use for a Dasibi photometer
operated at 300 K and 752 torr, and it can be loaded into the instrument via
front panel switches. On older Dasibi Model 1003 instruments (those without
zero offset capability) , the span number can be loaded with 4-digit accuracy
(e.g., 507400); on current production Model 1003 instruments (those with zero
offset capability) , only 3-digit accuracy is obtainable (507000) . With the
front panel mode switch in the span position, this span number is displayed as
50.40 on the older instruments and 50.750 on the current instruments. The
added 00.050 on the current instruments occurs because of the zero offset
circuitry.
The actual span number to use in a Dasibi photometer for an actual set of
operating conditions is conveniently obtained by multiplying the span number
derived for STP by the appropriate ratios of the actual operating temperature
and pressure to the STP values:
Span No. (T,P) = Span No. (273 K, 760 torr) x
Span No. (300 K, 752 torr) = 45.684 x
273
= 50.736 or 50.740 for older instruments
50.750 for newer instruments
When an actual span number is loaded into the Dasibi instrument, subsequent
displayed O concentrations are accurate only to the extent that the operating
temperature and pressure remain constant. Thus a 3° C change in temperature
would introduce a 1% error, and a 7.5 torr change in pressure would have a
similar effect. In most cases, only minor fluctuations in operating conditions
occur; these are generally ignored. When significant changes occur or better
accuracy is desired, the indicated Dasibi reading can be corrected by measuring
the actual operating temperature and pressure at each measurement and using
the following equation:
3-10
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
45.684 760 T
1 3JCorr l 3JInd X Span No. X P X 273
where Span No. = span number loaded into Dasibi instrument
T = actual operating temperature (K)
P = actual operating pressure (torr)
For convenience, the STP span number (45.684) may be left in the instrument,
allowing the second term on the right of the above equation to drop out.
Use of the linear form of the absorption law in the Dasibi instrument
results in a small, readily calculable error in the displayed concentration.
A display value of 1 ppm (I - I = 10000 counts) in the Dasibi using a span
number of 50.740 corresponds to a measured transmittance of
*o " (Io " J> 507400 - 10000
I - I ~ 507400
o o
_
- °'980292
The concentration calculated from the transmittance using the exponential form
of the absorption law is 1.010 ppm; thus the linear absorption law causes an
error of 1.0% at 1 ppm. The error decreases at lower concentrations. At 0.5
ppm, the error is 0.5%.
The precision of the photometer is more than adequate for calibration of
ambient 0 monitors. One standard deviation of 10 consecutive measurements
is typically on the order of 1 ppb. This means the instrument has a precision
in the transmittance measurement of ~1 part in 50,000.
Modification of the Dasibi Model 1003 Ozone Analyzer —
The Dasibi Model 1003 0 analyzer was designed as an ambient O monitor.
~J ~J
To provide a reference gas for the reference part of the instrument cycle, the
field instrument employs a converter which, ideally, removes all of the 0 from
an ambient air sample without causing any other changes in the sample. In
order to avoid any question of effectiveness of this converter, it is removed
3-11
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
from the instrument. Zero air is then diverted from upstream of the O gen-
erator in the calibration system (self-contained in the Model 1003-PC) and
is brought at atmospheric pressure to the reference side of the solenoid
valve. It is important that the same zero air used for the generator be
provided for the reference part of the cycle: experience has shown that samples
of "zero" air from different sources may vary significantly with respect to
transmittance at 254 run.
The sample channel counter is prevented from counting below numerical
zero by a reset line which resets the counter to zero as soon as a digital
"Borrow" signal is generated by the most significant digit of the counter. In
order to display the true instrument zero, this reset line (located on the top
printed circuit board in the instrument) is disconnected. Note that the first
count below zero in an up-down counting system generates an "all 9's" display.
For numbers below numerical zero, the displayed value is greater than the
actual count by 100.000. Thus a displayed value of 99.992 equals (99.992 -
100.000) or -0.008. In current production models of Dasibi 1003 O analyzers
with zero offset capability, it is not necessary to disconnect the reset line
provided sufficient positive zero offset is dialed into the instrument via the
front panel zero offset switch.
A third modification which may be required on older Dasibi photometers
consists of moving the solenoid valve a few centimeters away from the absorp-
tion cell. When operated as manufactured, the instruments "zero" ~15 ppb
below numerical zero. The amount of this negative zero offset is weakly
dependent on the sample flowrate. Since the solenoid valve is immediately
adjacent to the absorption cell, it has been postulated that turbulence set up
in the solenoid valve is carried into the cell. The turbulence might be
different during different parts of the instrument cycle (due to switching of
the solenoid valve), and the schlieren effects would change the transmittance
of the zero air sample. (Note that the 15 ppb effect corresponds to a change
in transmittance of only 1 part in 3300.) Moving the solenoid valve a few
centimeters from the absorption cell and connecting it to the cell via a piece
3-12
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
of Teflon tubing appears to allow turbulence patterns set up by the valve to
decay before the gas reaches the cell. Making this modification causes the
instrument to zero properly (i.e., ~90 percent of the zero readings are within
±2 ppb of zero after the instrument has thoroughly warmed up). In current
production models of Dasibi 1003 0 analyzers, the solenoid valve is separated
from the absorption cell by a short distance. Thus, this third modification
may not be required to obtain an instrument zero within a few ppb of numerical
zero when using current models of the 0 analyzer.
Instrumental Sources of Errors—
The Dasibi instruments have at least two sources of error which may not
be readily apparent to the operator. The first is partial failure of the
solenoid valve. The instrument measures the relative transmittance of a
reference gas sample and an ozonized gas sample. If the solenoid valve has a
significant leak, the two gas samples are mixed to some degree and the dif-
ference in transmittance is reduced. This causes the instrument to indicate
lower concentrations than are actually present. Leak-testing the valve is
neither difficult nor time-consuming, and should be performed regularly.
A second possible source of error is the dark count from the current-to-
frequency converters. To measure the dark count, the light source is removed
from the instrument and the sample and control frequencies are displayed. The
fractional error in each channel is equal to the ratio of dark count to light
count. Unfortunately, if the bias currents in the electrometers result in a
net current of opposite polarity to that required to produce a dark count, no
count is produced and this simple test for determining the magnitude of the
dark count error does not yield the desired information. In the instruments
examined, the dark counts were negligible.
Dasibi Ozone Analyzer (Models 1008-AH and 1008-PC)
The Model 1008-AH differs from the Model 1008-PC only in that the latter
includes a self-contained O generation system; both models will be referred to
as "Model 1008."
3-13
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
The Dasibi Model 1008 is a commercial photometer designed specifically
for use as a primary calibration standard for 0 . This model can also be used
as a transfer standard for 0 monitor calibration (see previous discussion on
use of the Dasibi Model 1003 as a transfer standard). It can be used as an
ambient O monitor provided it is modified to incorporate an 0 scrubber in
the flow system. However, as of this writing, the Model 1008 has not been
designated as an equivalent method for measurement of O in the atmosphere.
The optical subsystem of the Model 1008 is the same as that used in the
Model 1003 except for the addition of two sensors, both located at the exit of
the absorption cell. An absolute pressure sensor with a range of 0 to 1300
torr and overall accuracy of 2 torr over the range of 500 to 900 torr, and a
temperature sensor with a range of 0 to 50° C and overall accuracy of better
than 0.5° C are used to measure the pressure and temperature of the sample
exiting the absorption cell.
Although the required transmittance measurement is made with the same
patented counting technique used in the Model 1003, the electronics subsystem
of the Model 1008 is different from that of the Model 1003. The Model 1008
uses a Zilog-80 (Z-80) microprocessor to evaluate the logarithmic form of the
absorption equation to convert the measured transmittance of the sample to a
concentration:
6
-In —— = ac£ or c (ppm) = - —— In —
J_ OC Xr J_
O O
The microprocessor calculates the 0, concentration by using the following
,- -^
parameters: £ and 10 (conversion factor for atmospheres to ppm), which are
stored in the software; the value of a, which can be varied via a 3-digit
thumbwheel switch; and the measured transmittance of the sample (I/I ). The
o
microprocessor then uses the parameters of standard temperature and pressure
(273 K and 760 torr), also stored in the software, and the measured temperature
and pressure of the sample to calculate the O concentration corrected for
3-14
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
temperature and pressure. Thus, the Model 1008 automatically evaluates all
terms in Equation 4 of the 0_ calibration procedure (Appendix A) except for L,
the correction factor for O losses. It is the operator's responsibility to
determine L and apply the appropriate correction.
When the Model 1008 is used as a primary calibration standard, a value of
308 must be used for the absorption coefficient, a. The Model 1008 has a
front panel T/P switch which allows for display of either the corrected or
uncorrected 0 concentration. A front panel function switch can be used to
display the sample temperature and pressure. When used in conjunction with
the T/P switch, these readings can be used to verify that the Z-80 is making
proper temperature and pressure corrections. The unit also has zero offset
capability similar to that of the Model 1003.
The Model 1008 is equipped with a "T" between the solenoid valve and the
cell inlet and another at the cell exit to facilitate the O loss determination
required in the 0 calibration procedure. On the back of the photometer are
inlet ports for zero air and sample. The zero air required for photometer
operation should be supplied at slightly positive atmospheric pressure, since
the unit does not contain a sample pump.
The Dasibi Model 1008-PC is similar to the Model 1008-AH with the addition
of a stable 0 generator, a manifold, and optional ozonator-photometer feed-
back circuitry.
Instrumental sources of error are basically the same as those discussed
for the Model 1003.
Columbia Scientific Industries Corp. (CSI) Photocal 3000 Ozone Calibrator
The CSI Photocal 3000 O calibrator is designed for two applications, as
discussed below.
3-15
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
"Transfer Standard" for Ozone Monitor Calibration—
When used as a transfer standard, the 0 calibrator is periodically
certified in the laboratory and is assumed to give stable results between
certifications. The calibrator is taken to field sites and used to supply
accurately known concentrations of O for calibration of the field O monitor.
Use of a good quality zero air source is important in calibration. Since the
CSI photometer references to the zero air source, the possibility of inter-
ference from other UV-absorbing gases is not significant. However, impurities
which react with 0_, such as NO, may cause a reduced O output or cause the
output to be noisy in O generating systems.
The operation of the calibrator as a transfer standard is essentially the
same as that described under "Photometric Calibrator," below; see also McElroy
(1979).
Photometric Calibrator—
The CSI calibrator contains an O generation system and a photometer
which measures the relative transmittances of two gas samples at 254 nm. The
0- generator produces 0. in operator-selected concentrations from 50 to 1000
ppb. The photometer taps off samples of zero air and of O in zero air. The
concentration of 0 in the latter sample can be readily calculated from the
measured transmittance of the sample, the known absorptivity of O at 254 nm,
and knowledge of the optical pathlength through the sample. Measurement in
the zero air serves as a reference. It is this use of the CSI O calibrator,
i.e., its use as a convenient hardware package for generating O_ and making
photometric measurements at 254 nm, that is of interest here. Some important
details of the CSI hardware are described below.
The instrument uses a low pressure mercury discharge lamp for its light
source. The 185 nm O -producing radiation is blocked by means of a Vycor
shield around the lamp. The solar-blind vacuum photodiode detectors do not
respond to radiation at wavelengths of greater than ~300 nm. The source-
detector combination thus produces an effectively monochromatic photometer
operating at the 254 nm Hg line.
3-16
-------
3. OBTAINING A PHOTOMETER/Commercially Available Systems
The Photocal 3000 is a two channel photometer system (see Figure 3-2).
One detector (sample channel) monitors the intensity of 254 nm radiation
passing through the absorption cell. A second detector (control channel)
views a fraction of the 254 nm radiation reflected from the main beam exiting
the lamp housing. The control channel functions as an energy monitor and is
used to compensate for changes in source intensity.
The electrometer used for the transmission measurement is a current-to-
frequency converter; the signal from the sample detector is converted to a
digital signal very early in the signal processing and is thereby protected
from further degradation. The digitized signals are fed into a microprocessor,
which also controls solenoid valves that allow either zero air or sample air
to pass through the absorption cell.
During an instrument cycle, the solenoid valves first direct zero air
into the absorption cell to flush the cell ~5 times (~10 seconds with ~2.5
liters/minute of air passing through a 82 cm cell). A counter digitally
integrates the light intensity emerging from the absorption cell, and the
information is stored in memory by the microprocessor. The valves are then
switched to direct sample air into the absorption cell and the cell is flushed
for ~10 seconds. The integrated light intensity emerging from the cell is
again measured.
The CSI photometer uses the exponential absorption equation to calculate
the concentration from the measured transmittance. An absorption coefficient
of 308 atm cm at STP (273 K and 760 torr) base e is used in the calculation.
The system dark current is automatically measured and subtracted from the
signal currents by the microprocessor.
From the Lambert-Beer law,
T -ac£
I = I e
o
3-17
-------
3. OBTAINING A PHOTOMETER/Coimercially Available Systems
to
op:
Do
C
0
N
o
o
o
o
PI
(0
u
0
4J
o
£
ft)
CO
u
CM
0)
S-l
3
CJ1
-H
CD
3-18
-------
3. OBTAINING A PHOTOMETER/Coimercially Available Systems
The concentration, c, is calculated as follows:
,
c = - ln
where c =0 concentration in ppm
I = signal level with O in the transmission path
I = signal level with no O_ in the transmission path
o j
I = system dark current
a = O absorption coefficient
I = transmission pathlength
10 = conversion factor from atmospheres to ppm
Continuous temperature and pressure measurements are made in the instrument
and used to correct the calculation for the actual temperature and pressure
of the sample.
When operating at zero O concentration, the instrument may display a
small "negative" concentration. This below-zero concentration is preceded by
a minus sign and an error "E" is indicated in the display. This negative
value should be added to the O measurement as an offset, although it is
generally too small to be of consequence.
The instrument has at least one source of error which may not be readily
apparent to the operator: partial failure of a solenoid valve. The instru-
ment measures the relative transmittance of a reference gas sample and an
ozonized gas sample. If a solenoid valve should have a significant leak, the
two gas samples are mixed to some degree and the difference in transmittance
is reduced. This causes the instrument to indicate lower concentrations than
are actually present. Leak-testing the valves is neither difficult nor time-
consuming, and should be performed regularly.
3-19
-------
REFERENCES
APHA Intersociety Committee. 1977. Methods of Air Sampling and Analysis
(second ed.). American Public Health Association, Washington, D. C.
Becker, K. H., U. Schurath, and H. Seitz. 1974. Ozone Olefin Reactions in
the Gas Phase. Rate Constants and Activation Energies. Int. J. Chem.
Kinetics, 6:725.
Clyne, M. A. A., and J. A. Coxon. 1968. Kinetic Studies of Oxy-Halogen
Radical Systems. Proc. Roy. Soc., A303:207.
DeMore, W. B., and 0. Raper. 1964. Hartley Band Extinction Coefficients of
Ozone in the Gas Phase and in Liquid Nitrogen, Carbon Monoxide, and
Argon. J. Phys. Chem., 68:412.
Griggs, M. 1968. Absorption Coefficients of Ozone in the Ultraviolet and
Visible Regions. J. Chem. Phys., 49:857.
Hampson, R. F. (ed.). 1973. Survey of Photochemical and Rate Data for Twenty-
Eight Reactions of Interest in Atmospheric Chemistry. J. Phys. Chem.
Ref. Data, 2:267.
Hearn, A. G. 1961. Absorption of Ozone in the Ultraviolet and Visible Regions
of the Spectrum. Proc. Phys. Soc. (London), 78:932.
Inn, E. C. Y. , and Y. Tanaka. 1953. Absorption Coefficient of Ozone in
the Ultraviolet and Visible Regions. J. Opt. Soc. Am., 43:870.
McElroy, F. F. 1979. Transfer Standards for the Calibration of Ambient Air
Monitoring Analyzers for Ozone. U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina.
U.S. Environmental Protection Agency. 1979. Calibration of Ozone Reference
Methods. Federal Register, 44:8221, February 8.
4-1
-------
APPENDIX A
ULTRAVIOLET PHOTOMETRIC PROCEDURE FOR PRIMARY OZONE STANDARDS*
CALIBRATION PROCEDURE
1. Principle. The calibration procedure is based on the photometric
assay of ozone (O ) concentrations in a dynamic flow system. The concentra-
tion 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 O at 254 nm, (2) the
optical path length (£) through the sample, (3) the transmittance of the
sample at a wavelength 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
o
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 absorption cell, are identical during measurement of I and I .
o
The quantities defined above are related by the Beer-Lambert absorption law,
. .. ...
Transmittance = — = e (1)
o
where: a = absorption coefficient of O, at 254 nm = 308 ± 4
atnT1 cn^at 0°C and 760 torr. (1'2' 3'4'5' 6'7)
C = O 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 O concentrations
over the required range. Each 0 concentration is determined from the measure-
ment of the transmittance (I/I ) of the sample at 254 nm with a photometer of
path length H and calculated from the equation,
c(atm) = - - (In I/I ) (2a)
ct JG o
or,
c(ppm) = - - (In I/I ) (2b)
OCX/ O
The calculated O concentrations must be corrected for 0 losses which may
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 O 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 O , the transfer standard may replace the O 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 O_ is gen-
erated 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
diameter 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 O 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 O analyzer, and it should be free of NO,
C H , and other species which react with O . 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
o
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 indications
recorded in a chronoloigcal record either in tabular form or plotted on a
graphical chart. As the performance and stability record of the photometer
A-4
-------
APPENDIX A. ULTRAVIOLET PHOTOMETRIC PROCEDURE'/Reagents
is established, the frequency of these steps may be reduced consistent 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 O- concentra-
tion near the upper range limit of the system (0.5 or 1.0 ppm), then accurately
dilute that concentration with zero air and reassay it. Repeat at several
different 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)
Al
where: E = linearity error, percent
A = assay of the original concentration
A = assay of the diluted concentration
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 O 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 O concentration.
This loss must not exceed 5%. Some guidelines for quantitatively determining
this loss are discussed in Reference 9.
5.3 Assay of O 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 O 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 O concentration from equation 4. An average of
several determinations will provide better precision.
r>
o
where: [O,] = O concentration, ppm
a = absorption coefficient of O at 254 nm = 308 atm cm at 0°C
and 760 torr
I = optical path length, cm
T = sample temperature, K
P = sample pressure, torr
L = correction factor for O losses from 5.2.5 = (1-fraction 0_
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 O levels. See the photometer instruction
5.3.11 Obtain additional O_ 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 O analyzer to sample zero air until a stable response
is obtained and adjust the O 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 O 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 O 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 O analyzer, ppm
Z = recorder response with zero air, % scale
Record the O 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 O analyzer by adjusting the 0
-3 o
source or by Option 1. For each 0 concentration standard, record the 0 con-
centration and the corresponding analyzer response.
5.5.6 Plot the 0 analyzer responses versus the corresponding O concen-
trations and draw the O 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 O, generator. A
mixing chamber between the O., generator and the output manifold is also re-
quired. The flowrate through the O generator (F ) and the dilution air flow-
rate (F ) are measured with a reliable flow or volume standard traceable to
NBS. Each O concentration generated by dilution is calculated from:
[°3]OUT = [°3]OUT (iTl (6)
A-9
-------
APPENDIX A. ULTRAVIOLET PHOTOMETRIC PROCEDURE/Procedure
i
3 OUT = diluted 0 concentration, ppm
F = flowrate through the 0 generator, liter/min
F = diluent air flowrate, liter/min
REFERENCES
1. E.C.Y. Inn and Y. Tanaka, "Absorption Coefficient of Ozone in the
Ultraviolet and Visible Regions". J. Opt. Soo. Am., 43,870 (1953).
2. A.G. Hearn, "Absorption of Ozone in the Ultraviolet and Visible
Regions of the Spectrum", Proa. Phys. Soa. (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, Vl,725 (1974).
6. M.A.A. Clyne and J.A. Coxom. "Kinetic Studies of Oxy-Halogen Radical
Systems", Proa. Roy. Soa., 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
-------
APPENDIX A. ULTRAVIOLET PHOTOMETRIC PROCEDURE/Figures
t-
z
Q
H -I
OUTPU
MANIFO
c
f
0
H
>
h
I
r
C
f
4
i
1
(
1
_l
u.
c
Q
UJ
La.
Ss
12?
0 z
< s^
K 3 CC
X Ul z
Ul N Q
_ < 5
_l Ul
Z Q
O O
H
—
£
D
*
r
ii
u
3
I
a
_j
_i
D
X
z
J
a.
)g£
' S <
>
o
U.
>
a:
LU
O
V-
z
o
u
I
- (~
0
»
t-
z
Ul *
CC
Ul
5
O
UJ
10
in
UJ
Z
o
a.
CC
0
CD
CC
Q
o
Ul
t-
UJ
G
O
X
X
UJ
f
) 1
a_
a.
Ul
II
"" Z
U
U.
CC
UJ
UJ
u.
OS
< "* O
^11
^ LU
10
>i
CO
C
O
•H
X!
•H
U
•H
a
rtJ
o
•H
s
O
•H
•P
O
co
tn
•rH
A-12
-------
APPENDIX A. ULTRAVIOLET PHOTOMETRIC PROCEDURE'/Figures
fi
0
•H
4J
a
w
>1
CO
o
a
H
K)
O
CJ
•H
a
id
u
•H
§
-------
TECHNICAL REPORT DATA
czse read [infractions on the reverse 'jciorc comvicunvi
"
EPA 600/4-79-057
3 RECIPIENT'S ACCE3SI Or* NO.
TLE A\C SL'STiTLE
;5 REPORT DATE
TECHNICAL ASSISTANCE DOCUMENT FOR THE CALIBRATION OF
AMBIENT OZONE MONITORS
6. PERFORMING ORGANIZATION CODE
September 1979
7 AUTHOR'S!
Richard J. Paur, ESRL/RTP
Frank F. McElroy, EMSL/RTP
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Quality Assurance Division
Environmental Monitoring Systems Laboratory
U.S. 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
U.S. 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 NOTES
Technical Assistance Document
16. ABSTRACT
In February, 1979, EPA revised certain parts of the ambient air pollution
monitoring regulations (40 CFR Part 50, Appendix D) to specify a new procedure for
calibration of ambient ozone analyzers. The new procedure is based on ultraviolet
(UV) absorption photometry, and specifies in detail the UV photometer, other
apparatus, and the procedure necessary for establishing quasi-primary ozone concen-
tration standards derived from the known absorption coefficient of ozone at 254 nm.
This Technical Assistance Document is intended to provide information and
assistance to State monitoring agencies and other organizations which must use the
new procedure to calibrate ambient ozone analyzers. The first section of the
document is a discussion of absorpHon photometry, with emphasis on the transmittance
measurement and measurement errors. Section 2 provides step-by-step explanatory
information and advice keyed to each paragraph of the procedure. Section 3 discusses
UV photometers specifically, their design and operational characteristics, and severa
commercially available models.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Air pollution
Measurement
Calibration
Air pollution monitoring
Ozone-analyzer
Ozone standards
UV photometry
Technical assistance
document
68A
43F
STATEMENT
RELEASE TO PUBLIC
19 ^ECURITY CLASS ( Tins Report/
UNCLASSIFIED
21 NO. OF PAGES
69
20 SECURITY CLASS (This page/
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
------- |