EPA-600/7-77-059
June 1977
PROCEDURES MANUAL
FOR
ELECTROSTATIC PRECIPITATOR
EVALUATION
Wallace B, Smith, Kenneth M. Gushing.
and Joseph D, McCain
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
Contract No. 68-02-2131
Technical Directive 20604
Program Element No. EHE624
EPA Project Officer: F). Bruce Harris
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, N.C. 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into seven series. These seven broad categories
were established to facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously planned to foster
technology transfer and a maximum interface in related fields. The seven 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
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the effort
funded under the 17-agency Federal Energy/Environment Research and Development
Program. These studies relate to EPA's mission to protect the public health and welfare
from adverse effects of pollutants associated with energy systems. The goal of the
Program is to assure the rapid development of domestic energy supplies in an environ-
mentally-compatible manner by providing the necessary environmental data and
control technology. Investigations include analyses of the transport of energy-related
pollutants and their health and ecological effects; assessments of, and development
of, control technologies for energy systems; and integrated assessments of a wide
range of energy-related environmental issues.
REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Approval does not signify that the contents necessarily reflect the
views and policies of the Government, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Information
Service, Springfield, Virginia 22161.
-------
ABSTRACT
The purpose of this procedures manual is to describe
methods to be used in experimentally characterizing the per-.
formance of electrostatic precipitators for pollution control,
A detailed description of the mechanical and electrical
characteristics of precipitators is given. Procedures are
described for measuring the particle size distribution, the
mass concentration of particulate matter, and the concentra-
tions of major gaseous components of the flue gas-aerosol mix-
ture. Procedures are also given for measuring the electri-
cal resistivity of the dust. A concise discussion and out-
line is presented which describes the development of a test
plan for the evaluation of a precipitator. By following this
outline useful tests may be performed which range in complex-
ity from qualitative and relatively inexpensive to rather
elaborate research programs.
111
-------
TABLE OF CONTENTS
Page No,
Abstract. „..,„..-.,.,. , . , , iii
Acknowledgements. . . , vi
1. INTRODUCTION. . . . 1
1.1 ELECTROSTATIC PRECIPITATOR INSTALLATIONS 1
1.1.1. Types of Electrostatic Precipitators.„.. 1
1.1.2 Characteristics of Typical Precipitator
Tnstallat:1 ons. . 2
1,1.3 Parameters Which Govern Electrostatic
Precipitator Operation. 4
1.2 PARTICIPATE SAMPLING FOR ELECTROSTATIC
PRECIPITATOR EVALUATION. 9
1.2.1 General Problems . , 9
1.2.2 Particulate Mass Measurements...., 12
1. 2 . 3 Particle Sizing Techniques.... 12
.1.2.4 Particulate Resistivity Measurements.,.. 13
2. TECHNICAL DISCUSSION 16
2,1 ELECTRICAL AND MECHANICAL CHARACTERIZATION OF
AN ELECTROSTATIC PRECIPITATOR 16
2.1.1 Electrical and Mechanical Design Data... 17
2.1.2 Collecting Electrode System 20
2.1.3 Discharge Electrode System .,. 22
2.1.4 Electrical Power Supplies 25
2.1.5 Rapping Systems 30
2.1.6 Dust Removal Systems.... 35
2 . 2 MASS EMISSION MEASUREMENTS 36
2.2.1 General Discussion 36
2.2.2 EPA-Type Particulate Sampling Train..... 37
2.2.3 ASTM-Type Particulate Sampling Train.... 39
2.2.4 ASME-Type Particulate Sampling Train.... 41
2.2.5 General Sampling Procedures 42
2 . 3 PARTICLE SIZE MEASUREMENT TECHNIQUES 44
2.3.1 General Discussion 44
2.3.2 Inertial Particle Sizing Devices 50
2.3.3 Optical Measurement Techniques 67
2.3.4 Ultrafine Particle Sizing Techniques.... 69
2.4 PARTICULATE RESISTIVITY MEASUREMENTS 77
2.4.1 General Discussion „ « 77
2.4.2 Laboratory Determination of Particulate
Re s i s t i v i ty 82
2,4.3 In Situ Particulate Resistivity
Measurement 84
2 . 5 PROCESS EFFLUENT GAS ANALYSIS 85
2.5.1 General Discussion 85
2.5.2 Qualitative Gas Analysis 85
2.5.3 Quantitative Gas Analysis.. ,....,,,.... 86
i v
-------
TABLE OF CONTENTS
(Cont'd.)
Page No,
3. DEVELOPMENT OF TEST PLANS FOR ELECTROSTATIC
PRECIPITATOR EVALUATION, 90
3.1 General Discussion... 90
3.2 Level A Evaluation 91
3.3 Level B Evaluation 91
3.4 Level C Evaluation ..,...,.. 95
Appendix A - AEROSOL FUNDAMENTALS, NOMENCLATURE, AND
DEFINITIONS 99
Appendix B - PARTICULATE MASS CONCENTRATION MEASURE-
MENTS. 119
Appendix C - CASCADE IMPACTOR SAMPLING TECHNIQUES 181
Appendix D - SIZE DISTRIBUTIONS OF SUBMICRON AEROSOL
PARTICLES .249
Appendix E - LABORATORY DETERMINATION OF PARTICULATE
RESISTIVITY 305
Appendix F - IN SITU PARTICULATE RESISTIVITY MEASURE-
MENTS 316
Appendix G - FEDERAL STATIONARY SOURCE PERFORMANCE
REFERENCE METHODS .349
Appendix H - FEDERAL STATIONARY SOURCE PERFORMANCE
STANDARDS 411
BIBLIOGRAPHY 417
-------
ACKNOWLEDGEMENTS
Members of the Environmental Engineering Division of
Southern Research Institute who contributed to the writing
of this report include Dr. R. E. Bickelhaupt, Head of the
Ceramics Section, Mr. W. R. Dickson, Research Chemist, and
Mr, G. B. Nichols, Head of the Environmental Engineering
Division.
We also appreciate the assistance and guidance of our
Project Officer, Mr, D. Bruce Harris.
VI
-------
1. INTRODUCTION
Many different types of measurements must be made in
order to accurately evaluate the performance of an electro-
static precipitator installed to remove suspended particulate
matter from an industrial process stream. Among the required
measurements are determinations of the compositions of the
gas and suspended particles and of the particles' electrical
resistivity, concentration, and size distribution. Also,
the precipitator geometry and operating parameters must he
recorded for prope*- interpretation of the measurements.
This document provides information and guidelines for
use in planning and conducting tests to obtain the necess^<>
data .
A brief description of electrostatic precipitators and
various evaluation methods is provided in the remainde" of
this section. In Section 2 the methods of measuring preci-
pitator operating parameters and the technical background
and procedures for flue gas and particulate characterization
are discussed. Section 3 describes the logic and procedure
to be used in developing a test plan for the evaluation of
a precipitator. The Appendices contain detailed information
on the test methods, as well as a listing of the Federal
Stationary Source Performance Standards and Federal Source
Testing Reference Methods.
T.I ELECTROSTATIC PRECIPITATOR INSTALLATIONS
1.1.1 Types of Electrostatic Precipitators
Two general types of electrostatic precipitutors are used
to control particulate emissions from stationary sources:
-------
dry and wet electrode precipitators. Within each of these
categories, precipitators may be further classified by elec-
trode geometry. Figure 1 shows a typical parallel plate
precipitator. This is by far the most common electrode
configuration for large installations. Another common
electrode geometry is a wire-pipe or cylindrical type.
Dry electrostatic precipi tators are installed in indus--
tries with widely varying gas conditionsr temperature, and
pressure. In the electric utility industry there are two
positions in the flue ducts £or locating the collectors--
either on the hot gas or cold gas side of the air preheater
Wet electrode precipitators operate in a manner similar
to hot or cold side units with the difference that a thin
film of liquid flows over the collection plates to wash off
the collected particulate. In some units, liquid is also
sprayed into the interelectrode space to provide cooling,
conditioning, or a scrubbing action. The spray is collected
with the aerosol particles and provides a secondary means
of wetting the plates. The operating temperatures are gen-
erally less than 65°C. Wet electrode precipitators are
widely used in the metals industries.
1.1.2 Char a c t e r_i_s tics of Typ i c a 1 P T e c i p i t a t o r Irtst all a t i on s
Electrostatic precipitators constructed as industrial
gas cleaning devices vary widely in size and configuration.
The differences in size depend on the type of industrial
process where they are used, the gas volume to be handled,
the gas residence time, and the desired collection efficiency.
Electrostatic precipitators are as large as 12 meters high,
-------
3S30 065
Figure 1. Typical wire-plate electrostatic precipitator.
-------
45 meters wide, and 30 meters long. Gas volume flow rates
are as much as 1500 ma/sec. For large gas volumes, however,
several precipitators are usually placed in parallel. At
some installations, the parallel precipitators are stacked
vertically to minimize the use of ground space, especially
if they are retrofitted. Figures 2 through 4 show some typi-
cal precipitator installations.
1.1.3 Parameters Which Govern ElectrostaticPrecipitator
Operation
The theoretical collection efficiency, n» of a precipi-
tator for particles of diameter D, is given by the Deutsch
equation1
n = 1 - exp (-UD A/Q) (I)
A = total effective collecting electrode area (m2),
Q = total gas flow rate (m3/sec), and
o)_. = migration velocity (m/sec) of a particle with dia-
meter D.
The migration velocity is the terminal velocity of a charged
particle in the boundary layer near the collecting electrode
and is largely a function of five variables:
WD = UD^' V' Cg' T' D)
where D = the particle diameter,
j = the current density in the interelectrode space,
V = the applied voltage (actually the electric
field is more accurate, but is not measured
directly),
Cg = the gas composition, and
T = the gas temperature.
-------
t
BOILER
UPPER
ELECTROSTATIC
PRECIPITATOR
INDICATES DIRECTION
OF GAS FLOW
NOT TO SCALE
INLET
TEST
PORTS
OUTLET
TEST PORTS
FORCED
DRAFT FANS
SECON-
DARY
AIR TO BOILER
vv
LOWER
ELECTROSTATIC
PRECIPITATOR
PRIMARY
AIR TO MILLS
PRIMARY AIR FANS
3630 050
Figure 2, Schematic diagram of n hot side ESP installed on a coal-fired
power boiler. Side view.
-------
BOILER
INLET TEST PORTS
PRIMARY AIR PREHEATER
ELECTROSTATIC
PRECIPITATOR
INDICATES DIRECTION
OF GAS FLOW
NOT TO SCALE
OUTLET TEST PORTS
SECONDARY AIR PREHEATER
363O 049
Figure 3. Schematic diagram of a hot side ESP installed on 3 coat-fired
power boiler. Top view.
-------
INDICATES GAS FLOW
NOT TO SCALE
HOT SIDE
ELECTROSTATIC
PRECIPITATOR
COAL SAMPLE *•
POINT
SUPER HEATER
BOILER
\/\A
STACK
ELECTROSTATIC
AIR PREHEATER ^ ^ PRECIPITATOR
INDUCED DRAFT FAN
Figure 4. Schematic of an ESP system when a hot side precipitutor has
been retrofitted to supplement the existing cold sidfi precipitatnr.
-------
If the particle size distribution is unknown, the Deutsch
equation is sometimes used with an empirical parameter, M ,
called the precipitation rate parameter, substituted for LJ .
For this application, ri would represent the overall mass
collection efficiency.
The six parameters A, Q, j, V, Cg, and T are the minimum
number which must be measured in addition to the electrode
design in order to diagnose poor precipitator performance.
Of course, the actual collection efficiency can be known
only after measuring the inlet and outlet mass loading con-
centrations, and a rigorous analysis requires a knowledge
of the particle size distribution.
The quantity A/Q, measured in units of m2/(m3/sec) is
the specific collection area (SCA) used to describe the ef-
fective size of a precipitator. An SCA value of 40 m2/(mVsec!
would be considered small for most applications and 120
m2/(m3/sec) would be considered large. High SCA values may
indicate a problem dust, a high efficiency design, or merely
a conservative design.
The total collection area, A, must be obtained from
the precipitator design drawings and is usually known by
the plant engineer.
An approximate value of the total gas flow rate, Q,
can be obtained by making a mass balance at known plant
operating conditions, or by inquiring of plant personnel.
During testing, however, a detailed gas velocity distribu-
tion determination by a pitot traverse across the duct or
-------
stack will give a more accurate value for total gas volume
flow rate. A well defined gas flow rate distribution is
necessary for proper isokinetic sampling during particulate
sampling tests (i ._e. , the sampling velocity is equal to the
gas velocity). Detailed velocity traverses and gas analyses
also permit the discovery of gas leaks into or out of the
precipitator.
The electric current and primary and secondary voltages
are indicated by panel meters in the precipitator control
room. In some instances secondary voltages must be measured
using voltage dividers installed by test personnel.
Experienced personnel are able to obtain qualitative
estimates of the effectiveness of particle charging, to ob-
serve resistivity problems, or to detect electrode misalign-
ment from the experimental data listed above.
1.2 PARTICULATE SAMPLING FOR ELECTROSTATIC PRECIPITATOR
EVALUATION
1.1.2 General Problems
Measurements of particle size and concentration are usually
made at both the precipitator inlet and outlet to obtain
an accurate characterization of the precipitator performance,
Gas velocities may vary from 15 m/sec to 30 m/sec, in
ducts leading toward or away from the precipitator, to as
low as 1-2 m/sec in transforms immediately upstream or down-
stream of the precipitator.
-------
If samples are taken near the precipitator proper, then
low gas velocities pose problems because of the difficulty
of sampling isokinetically at such locations. Large dia-
meter nozzles are required to sample at low gas velocities
when the normal sampling train flow rates are used.
Particulate concentrations may range from 0.001 to
10 g/m , depending upon the type of source and the efficiency
of the precipitatot. Sampling times required to obtain accurate
data vary approximately inversely with the dust concentration.
At precipitator outlets it is not uncommon to sample for
12 hours to collect a sufficient amount of dust. Sampling
times can also be discontinuous if studies are being conduc-
ted to isolate the effects of rapping reentrainment or hopper
blowoff during dust removal.
At some sources, condensation can occur within the pol--
lution control system or in the stack, and particles may
grow larger and change composition, or be created, by this
mechanism. Both the concentration and size distribution
of the dust effect the precipitator performance.
A source of difficulty in testing a wet electrostatic
precipitator is the possible presence of mists and entrained
water droplets, perhaps with suspended solids, in the gas
at the outlet sampling location. Also the particles can
act as condensation nuclei for water droplet growth, thus
causing an apparent change in the particle size distribution.
Entrained water can cause filters or substrates to become
soaked and erroneous measurements of the flue gas water
vapor content can occur if precautions are not taken.
In many instances the duct dimensions at the sampling
location can pose problems, especially if sampling from
the top of a duct is required. The duct may be up to 7
10
-------
meters in depth. A hoist must be constructed to handle the
probes required for obtaining samples near the bottom of
such a duct.
When a precipitator is sampled Immediately upstream or
downstream, especially when long probes are required, it
is important to keep the probe from coming in contact with
the corona wires as severe electrical shock can result.
Unless the probes are properly grounded, sampling down-
stream can also result in electrical shock from the charged
particles leaving the precipitator.
In a dry electrostatic precipitator the dust is collected
on vertical plates which are rapped on a regular basis.
Upon rapping the dust falls into hoppers and is removed
from the precipitator. During the rapping cycle, however,
some of the collected dust is reentrained into the gas
stream and contributes significantly to the losses. This
causes cyclical variations in the outlet dust concentration.
Air flow may also extend into the hoppers causing dust
to be blown up and out of the hopper and back into the gas
stream. This is generally localized in the lower portions
of the precipitator, and a comprehensive sample traverse
is necessary to include the higher dust loading at the
bottom of the precipitator.
In conducting source sampling for hot side precipitator
evaluations, high temperatures can cause special problems.
These include difficulties in probe handling and leak seal-
ing, filter material integrity, condensation of gaseous com-
ponents upon cooling, instrument failure, etc.
11
-------
In some instances, it is neither feasible nor economical
to test an entire system of parallel precipitators in detail.
It is reasonable to make detailed tests on a single unit
and to infer the performance of untested units in the same
system from electrical performance data.
1.2.2 Particulate Mass Measurements
Measurements of the particulate mass concentration are
made by pumping the dust laden gas through a system contain-
ing a filter and a means of measuring the volume of the gas
stream sampled. The total material collected on all surfaces
within the system is recovered and weighed. This weight,
normalized to a unit of gas volume, is the suspended par-
ticulate concentration. The samples are collected using
a prescribed traversing procedure, which, in effect, yields
an approximate integration of the average mass emission rate
past a cross-section of the duct or stack. The velocity
distribution of the gas is also measured as pa^t of the test.
The gas flow rate and velocity distribution and the par-
ticulate mass concentration are used to calculate the mass
flow rate or emission rate at the point of interest. Mea-
surements of the mass flow rate are made at both the inlet and
outlet to determine the precipitator collection efficiency.
Section 2.2 contains a summary of the methods for conducting
mass measurements. Appendix B describes in detail how to
carry out the procedures in EPA Method 5 particulate mass
measurement.
1.2.3 Particle Si zing Tech n i q ues
Measurements of particle size distribution in industrial
flue gas streams are made for several reasons. The aerosol
12
-------
must be characterized as completely as possible in o.ccler
to assess the potential for adverse health or environmental
effects; emission measurements can be useful as a process
monitor; and the aerosol particle size distribution must
be known in order to completely quantify the behavior of
a control device. Also, particle size measurements on un-
controlled sources are useful in precipitator design.
In recent years the emphasis on pollution control has
been placed on "fine" particles which are defined by the
EPA as those particles having aerodynamic diameters smaller
than 3 micrometers. Fine particles are more difficult to
control than large particles, and because they are respi-
rable, may constitute a greater hazard to health.
Several techniques must be applied if information on
particle size over a wide range of diameters is required,
or if real time data is desirable. As a general rule of
thumb, most particle sizing techniques yield accurate in-
formation over approximately a factor of ten range in par
ticle diameter. The particle size range with which this
manual is concerned is 0.01 to 10 ym. Therefore, several
techniques must be used.
Section 2.3 contains a summary of the instruments that
are available for particle sizing, and a discussion of their
applicability to specific tests. Appendices C and D contain
detailed descriptions of the procedures for measurement of
particle size distributions in flue gases.
1.2.4 Particulate Resistivity Measurements
The electrical resistivity of particulate matter, present
in the effluent gas stream, is an important factor that affects
13
-------
the operating characteristics of an electrostatic precipi-
tator. In a conventional single-stage dry electrostatic
precipitator, the corona current must flow through a pre-
viously collected dust layer to reach the grounded collec
tion electrode. This flow of current requires an electric
field (E) in the dust layer proportional to the corona cur--
rent density (j) and the particulate resi'stivity (D) as given
by
E = jp (2)
The electric field in the dust layer causes a potential dif-
ference (AV) across the dust layer proportional to the dust
layer thickness (t),
AV = Et (3)
which is reflected in the operating V-I characteristic of
the precipitator.
If the resistivity of the dust layer increases for a
given current density, the electric field in the layer will
increase proportionately. If the electric field in the inter-
stices or on the surface of the dust layer exceeds the field
strength for electrical breakdown, an electron avalanche
will occur similar to that which occurs adjacent to the corona
wire. This electrical breakdown acts as a limit on the allow-
able electrical conditions in the precipitator.
The manner in which electrical breakdown at the collec-
tion electrode limits the precipitator performance is depen-
dent upon the value of the resistivity of the dust and the
thickness of the layer. If the resistivity is in the mod-
erately high range (1011 ohm-cm) the breakdown will generally
initiate electrical sparkover between the precipitator elec-
-------
trodes; whereas if the resistivity is very high (>1012 ohm-cm)
breakdown of the dust layer will occur at a voltage too low
to propagate a spark across the interelectrode region. This
gives rise to a condition of reverse ionization or back corona
which will introduce positive ions into the interelectrode
space, reducing the electrical charge on the particles.
If the particulate resistivity is too low (<107 ohm-
cm) , particles will be discharged immediately upon touching
the collection electrode. If this occurs there will be no
electrical forces holding the collected particles to the
plate and they tend to be reentrained into the gas stream.
In view of the importance of the resistivity of the
dust layer as a prime factor in limiting the performance
of a precipitator, it is necessary to determine the resis-
tivity of the material to be collected in order to perform
a~comprehensive evaluation of a precipitator.
Techniques of measuring particle resistivity are dis-
cussed in Section 2.4 and Appendices E and F.
15
-------
2. TECHNICAL DISCUSSION
2.1 ELECTRICAL AND MECHANICAL CHARACTERIZATION OF AN
ELECTROSTATIC PRECIPITATOR
In order to perform a comprehensive analysis of the
performance of a precipitator it is necessary to obtain
information on the design and operating parameters of
the complete installation. These data are generally
used, with the test results, to make comparisons with
other precipitator evaluations or with performance pre-
dictions of theoretical or empirical models. Plant or
process operating data should also be obtained as part
of each test program., and correlated with the control
device performance. Plant data are routinely recorded by
the plant personnel and arrangements can usually be made
to obtain copies of these records or to have a member of
the test crew record the data during a test.
In the remainder of Section 2.1, the nomenclature
which is used to describe electrostatic precipitators is
defined and a number of parameters which should ideally
be measured or noted as part of a precipitator evaluation
are listed. Some of the data listed are essential to
a meaningful evaluation, while some may be difficult
or impractical to obtain and must be sacrificed. Also,
some precipitator installations may have individual pecu-
liarities which require that additional data be taken.
Although this manual is intended as a comprehensive guide.
there is no substitute for intelligent and experienced
judgement in the discrimination of essential from nonessen-
tial data.
16
-------
2.1.1 Electrical and Mechanical Design Data
The objective of a precipitator design is to combine the
component parts into an effective arrangement that results
in an optimum collection efficiency. Experience, efficiency
requirements, and economics generally dictate the best arrange-
ments. However, precipitator technology can be improved
through the study and comparison of the electrode design,
the total collecting surface area and total length of discharge
electrode in one bus section, the total area of collecting
surfaces cleaned by each rapper, the ratio of the area of
one bus section to the total, the number of power supplies,
the collecting surface area associated with one power supply,
the number of fields, the total collecting surface area com-
pared to the gas flow rate, and other parameters deemed impor-
tant during the evaluation.
For completeness, determinations of the following items
pertaining to the physical layout of the precipitator should
be made during the test. This data may be obtained by observa-
tion, from plant personnel, or from manufacturers' literature.
Number of precipitators - A single precipitator is an
arrangement of collecting surfaces and discharge electrodes
contained within one independent housing. (See Figure 5.)
Number of chambers per precipitator - A chamber is a
gas-tight longitudinal subdivision of a precipitator. A
precipitator without any internal dividing wall is a single
chamber precipitator. A precipitator with one dividing wall
is a two-chamber precipitator, etc. (See Figure 5.)
17
-------
TRANSFORMER/RECTIFIER
BUS SECTIONS
CHAMBERS
INTERNAL PARTITION
BUS SECTIONS
GAS FLOW
CHAMBERS
FIELDS
CASE 1: 1 PRECIPITATOR, 2 CHAMBERS, 12 BUS SECTIONS, 6 POWER SUPPLIES, 3 FIELDS
TRANSFORMER/RECTIFIER
BUS SECTIONS
INTERNAL PARTITION
BUS SECTIONS
CHAMBERS
GAS FLOW
CHAMBERS
FIELDS
CASE II: 1 PRECIPITATOR, 2 CHAMBERS, 12 BUS SECTIONS, 12 POWER SUPPLIES, 3 FIELDS
3630-047
Figure 5, Typical prec/pftator arrangements.
18
-------
Number of fields - A field is the physical portion of
the precipitator in the direction of gas flow that is ener-
gized by a single power supply. A field can denote an arrange-
ment of bus sections in parallel or series as long as this
arrangement is energized by a number of parallel power sup-
plies. (See Figure 5).
Number of power supplies - A power supply is a transfo*:-
mer /rectifier (T/R) arrangement with either a full or half
wave voltage output. Usually a single power supply energizes
more than one bus section.
Number of bus sections in each field - A bus section
is the smallest portion of the precipitator which can be
electrically de-energized independently. It is advantageous
from a performance standpoint to have as much sectionalization
as possible in a precipitator for two reasons. Localized
electrical problems will not disrupt the performance of a
large part of the precipitator, and since inlet sections nor-
mally operate at lower current densities than outlet sections,
these should be separated electrically.
Total numberof bus sections in the precipitator - The
total number of sections in a precipitator is equal to the
number of fields multiplied by the number of bus sections
per field.
Number of gas passages - A gas passage is the volume
of the precipitator between any two adjacent collecting surfaces,
Effective precipitator dimensions - The volume inside
the precipitator through which the flue gas passes determines
the effective height, width, and length of the device for
the collection of dust. The effective length is equal to
19
-------
the total length of the collection surface in the direction
of gas flow; the effective height is equal to the total height
of the collecting surface measured from top to bottom; and
the effective width is equal to the total number of gas passages
multiplied by the center to center spacing of the collecting
surfaces (ignoring the shape of the collecting surface).
Gas distribution devices - A gas distribution device is
any physical plate, screen, or baffle positioned at the entrance
to a precipitator to cause a change or smoothing effect in
the gas flow characteristics.
2.1.1 Collecting Electrode System
The collecting surfaces are the individual grounded
components on the surfaces of which the particulate matter
is collected. The shape of the collecting electrode is de--
signed to maximize the electric field while minimizing dust
ceentrainment.
Type of collection electrodes - Many shapes of collecting
electrodes are used in electrostatic precipitators. For
dry precipitators it is important that dust reentrainment
be minimized during rapping. In well designed precipitators
where gas velocities are kept low and gas flow uniformity
is maintained within limits of good practice, there is little
or no direct scouring of dust from the plates. All preci-
pitators have some sort of baffle arrangement on the collec-
tion plate to minimize gas velocities near the dust surface
as well as to provide stiffness to the plate. Some of these
baffle-plate designs are known as Opzel plates, Rod curtain
plates, Zig-Zag plates, V Plates, V-pocket plates, channel
plates, offset plates, shielded plates, and tulip plates
(See Figure 6).
20
-------
COLLECTING PLATES
WIRES
ROD CURTAIN
oooooooo
oooooooo
ZIG-ZAG PLATE
\AAAAAAAA DISCHARGE
vvvvvvvv \_ ELECTRODES
\AAAAAAAA
V-POCKETS
CHANNEL
OFFSET PLATES
SHIELDED PLATE
TULIP PLATE
ELECTRODES
V-PLATES
0700-12,10
Figure 6. Various types of collection electrodes.
21
-------
Active collecting surface area - This is the total pro-
jected area of the collecting surface exposed to the electro-
static field (effective length multiplied by the effective
height and number of sides).
Plate to plate distance - This is the distance between
the centers of two adjacent plates on a line perpendicular
to the gas flow. The maximum current and electric field
are strongly dependent upon the electrode spacing.
Aspect ratio - The collecting plate aspect ratio is the
effective length divided by the effective height (L/H).
2.1.3 Discharge Electrode System
This system is designed in conjunction with the collec-
ting electrode system to maximize the electric current and
field strength. The discharge electrode is also referred
to as the corona electrode, cathode, high voltage electrode,
or corona wire. Several properties of the corona or discharge
electrodes can influence precipitator efficiency and should
be included in a test report,
Type of dischargeelectrode - The discharge electrodes
are held at a high electrical potential during the precipi--
tator operation to ionize the gas and establish electric
fields for particle charging and precipitation. The elec-
trodes may be in the form of cylindrical or square wires,
barbed wire, or stamped or formed strips of metal of various
shapes. Some discharge electrode geometries are shown in
Figure 7.
The discharge electrode may be suspended individually
from an insulating superstructure, with weights attached
to the lower ends, or they may be mounted on frames, or masts.
Figure 8 shows two ways that electrodes may be mounted.
22
-------
u
0700 -12.4
Figure 7, Typical forms of discharge or corona electrodes.
23
-------
SUPPORT FRAME
RAPPER ANVIL
CORONA ELECTRODES
MAIN MAST SUPPORT
\
SUPPORT PIPES
Figure 8. Supported electrode structures.
A. Frame-type. B. Mast-type.
WIRES
3630-62
24
-------
Pischar ge e 1 ectrode dimensions -- The physical dimen-
sions of the corona electrode should be noted. The diameter,
or width of this electrode is an important factor in con-
trolling the electric field in the interelectrode space,
and the starting voltage. The effective length of the dis
charge electrode may also be of interest; it is equal to
the height of the collection electrode.
Discharge to collecting electrode spacing - The applied
voltage divided by the distance between the discharge and
collecting electrodes is the average electric field.
Numberof discharge electrode per gas passage - The
total number of electrodes (corona wires) along the length
of one gas passage may be of some interest in mathematical
modeling of the precipitator.
Total number of discharge electrodes - The number of
electrodes in one gas passage multiplied by the number of
gas passages.
Total effectivedischarge electrode length - The total
number of electrodes multiplied by the effective electrode
length.
Spacing of discharge electrodes in the direction ofgas
flow - The distance between discharge electrodes in a single
gas passage has a significant effect on the precipitator vol-
tage-current characteristic.
2.1.4 Elect r i c P ow er Supplies
Each power supply consists of four components as shown
in Figure 9: a step-up transformer, a high voltage rectifier
25
-------
.AC VOLTAGE
"INPUT
CONTROL
ELEMENT
STEP-UP
TRANSFORMER
ELECTROSTATIC
PRECIPITATOR
HIGH VOLTAGE
RECTIFIER
MANUAL
AUTOMATIC
CONTROL FEEDBACK
O7 00-10.1
Figure 9. Power supply system for modern precipitators.
26
-------
a control element, and a sensor for the control system.
A step-up transformer is required because the operating
voltages for modern precipitators range from about 20 to
100 kV. The high voltage rectifier is required to convert'
the secondary AC voltage to half or full wave voltage in
order to be compatible with the electrostatic precipitator
requirements.
One function of the control system is to vary the ampli-
tude of the unfiltered voltage waveform that is applied to
the discharge electrode system. Control can be applied to
either the primary or secondary side in the power supply,
but it is usually on the primary (low voltage) side. The
control system can be operated either manually or in one
of several automatic modes. Automatic controls are usuallv
installed. A well designed automatic control system main-
tains the voltage level at the optimum value, even when the
dust characteristics and concentration exhibit temporal fluc-
tuations .
Transformers - In many precipitators, performance is
limited by the maximum voltage or current ratings of the
power supply transformers. The transformer ratings should
be noted and reported.
Rectifiers - The rectifiers change the alternating current
to a pulsating DC current, either full or half wave. In
general, half wave power supplies allow a greater degree
of sectionalization, although the sections are not completely
independent, since control is normally associated with the
transformer primary. Full wave rectification is used where
higher average currents are desirable, as for example, where
27
-------
large dust loading or extremely fine particles lead to a large
space charge which limits the maximum current.
Number of transformer - rectifier(T/R) sets - The
general layout and number of T/R sets should be noted; i.e.,
the number per field, number per chamber, the number .per
precipitator. This should be shown in a schematic drawing
similar to Figure 5.
Spark_ rate - The spark rate is the number of times per
minute that electrical breakdown occurs between- the corona
wire and the collection electrode. A spark-rate controller
establishes the applied voltage at a point where a fixed number
of sparks per minute occur (typically 50-150 per corona section)
The sparking rate is a function of the applied voltage for
a given set of precipitator conditions. As the spark rate
increases, a greater percentage of the input power is wasted
in the spark current, and consequently less useful power
is applied to dust collection. Continued sparking to one
spot will cause erosion of the electrode, and sometimes,
mechanical failure.
Voltage-Current Characteristic - Voltage-current rela-
tionships for the primary and secondary circuits of each
T/R set for both clean and dirty plate conditions are verv
useful in interpreting precipitator behavior. Clean plate
V-I data, however, must be obtained immediately after shut
down after thorough plate rapping to remove excess dust or
from data taken during tests occurring in the process of
precipitator installation and shakedown. The dirty plate
V-I measurements usually take place after particulate emis-
sion tests because it is undesirable to disturb the precipi-
tator before the tests. When mass testing and particulate
sizing measurements are being conducted, voltage and current
readings should be recorded regularly for each power supply.
28
-------
TO VOLTMETER FOR
SECONDARY VOLTAGE
PRECIPITATOR CONTROL
PANEL PRIMARY VOLTAGE
AND CURRENT CONTROL
TO VOLTMETER FOR
SECONDARY CURRENT
TRANSFORMER
S.A. = SURGE ARRESTOR
1, SECONDARY VOLTAGE = Vn
2. SECONDARY CURRENT =
3630-051
Figure 10. Voltage divider network for measuring precipitator secondary
voltages and currents.
29
-------
Many precipitator control rooms have panel meters for
each T/R set which show the primary and secondary voltages
and current and the sparking rate. If meters are not in-
stalled on the transformer secondary, a temporary voltage
divider network can be installed on the precipitator side
of the recitifer network as shown in Figure 10. Typically,
the resistor R2 has a value of about 1x109 ohms and Rj has
a value of about 12x103 ohms. Because of the voltage drop
across R2, this resistor should be well insulated. If it
is necessary to measure the secondary current, a voltmeter
can be placed across resistor R3 in the Surge Arrester net-
work. Ra is typically 50 ohms or less. Some manufacturers
may place a current meter with very low internal impedance
across this resistor and allow all the precipitator current
to pass through the meter. The resistor is installed to
prevent isolating the power set should the meter be removed
from the circuit. To obtain accurate data in this case,
a calibrated current meter can be installed in series with
this current meter. Figure 10 shows a typical precipitator
T/R circuit but many individual installations may have their
own peculiarities.
Since many commercial precipitators do not have secondary
voltage and current meters installed, it is worthwhile for
anyone involved in precipitator evaluations to build and
calibrate several divider networks for field use.
2.1.5 Rapping Systems
Type of rappers - Differences in precipitator design
also occur in the types of rappers used to remove dust collected
on collection plates and discharge electrodes.
30
-------
Mechanical rappers may be located within the shell of
the precipitator with the rapping system consisting of ham-
mers which are lifted by a rotating shaft until they pass
over the top of the shaft and then fall under the force of
gravity and strike anvils connected to the collection plates
or discharge electrode frames. (See Figure 11.)
A second type of mechanical rapper is located outside
of the precipitator shell with the anvils externally mounted
and extending through the shell. In such a system, seals
must be provided between the anvil and shell to prevent gas
leakage. Rapping cycles can vary from a single impact to
multiple blows for each rapping sequence and the rapping
cycle can be changed by varying the frequency of rap.
Pneumatically or electromagnetically operated rappers
may be of the impact or vibratory type. These rappers are
connected to the discharge or collection electrode frame
through the precipitator shell. The impact rapper functions
by lifting a weight to a height determined by a pneumatic
or electromagnetic controller and allowing the weight to
fall against an anvil when the holding force is released.
Vibratory rappers impart vibrations to the electrodes through
rods extending into the precipitator shell. (See Figure
12) .
Different rapping systems are used because of differences
in dust types and manufacturers philosophy, as well as cost
and maintenance considerations. The plant engineer will
usually know the type of rapper and the sequence and timing
of the rapping cycle.
Rapping variables - The efficiency of a precipitator
is affected by the rapping interval, the rapping intensity,
and the length of each rapping cycle. Rapping reentrainment
31
-------
0700-12.13
Figure Tl. Mechanical type rapper.
32
-------
•ENCLOSURE
QUICK OPENING
CLAMP
GROUND CONNECTION
CLAMPS
(RAPPER RODST
CERAMIC SHAFT)
HOUSING
- CERAMIC INSULATING SHAFT
PRECIPITATOR ROOF
VIBRATION
TO
DISCHARGE
WIRE
ENCLOSURE
VIBRATOR
MOUNTING PLATE
STUFFING BOX AND GUIDE
— FLEXIBLE CONDUIT
CONDUIT FITTING
DUST LADEN
GAS AREA
CLOSURE PLATE
HIGH VOLTAGE BUSHING
RAPPER ROD ASSEMBLY,
MUST BE PLUMB
HIGH TENSION FRAME
DISCHARGE WIRES
O700-12.12
Figure 12. Typical vibratory rapper.
33
-------
contributes a significant fraction of precipitator emissions
and it is usually desirable that oarticulate sampling at
a precipitator outlet cover a time span which includes several
rapping cycles. This will help ensure a more representative
outlet mass loading. A different strategy would be used,
of course, if the contribution of rapping to the outlet dust
loading were the object of investigation.
Interval of rapping - It is desirable to know the time
interval of rapping for each electrode in each field of the
precipitator. Usually the upstream fields are rapped more
frequently than the downstream fields because dust builds
up more slowly on the outlet side.
Intensity of rapping - It is desirable to determine
with what force the electrodes are being rapped within each
field. Although this is an interesting quantity, it is
usually impossible to obtain experimentally, except as part
of elaborate research programs. Some qualitative information
may be obtained from the plant engineer or manufacturing
literature.
Durationofthe rapping cycles - It is desirable to
note the length of time of each rapping cycle for each elec-
trode system in each field of the precipitator.
Rappe r sj c ti on a 1 i z a t j^on - Precipitators are rapped field
by field with the inlet fields usually being rapped more
often, as noted in the report.
34
-------
2.1.6 Dust Removal Systems
It is important during any precipitator evaluation to
determine that the dust removal system is working properly
and to specifications. If the system is not working properly
discussions should take place with plant personnel to see
whether these problems can be fixed in a reasonable time
span. If this is not the case, then testing should be de-
layed .
Hoppers are used to collect and store dry particulate
which is removed from the electrodes. If hoppers are allowed
to overflow, the collected dust will be reentrained thereby
greatly reducing precipitator efficiency, or electrical short
circuits may occur which disable part of the precipitator
power supplies. Baffles are frequently placed in hoppers
to minimize undesirable gas flow which may lead to reentrain-
ment (hopper sweepage)-
If the precipitator system is operated with internal
pressures less than ambient, then air in-leakage through
the hopper can cause a ueentrainment of the dust from the
hoppers.
Several types of systems exist for removal of dusts
accumulated in hoppers. These include container removal,
dry vacuum, wet vacuum, screw conveyors, and scrape bottom
systems.
As much information about the dust removal system should
be obtained and reported as possible.
35
-------
2.2 MASS EMISSION MEASUREMENTS
2.2.1 General Discussion
As part of most electrostatic precipitator studies,
the particulate collection efficiency is determined experi-
mentally by making measurements of the mass concentration
and gas flow rate at the precipitator inlet and outlet.
A number of test procedures have been developed for per-
forming mass emission measurements on process streams. All
of the "standard" sampling trains are similar and basically
are composed of a nozzle, a probe, a filter, one or more
devices for monitoring gas flow, and a pump. Generally pi tot
tubes and thermocouple assemblies are also used to measure
the gas velocity and temperature. The nozzle is streamlined
to minimize flow disturbances and the diameter is chosen
for isokinetic sampling. The probe must be rigid so that
the nozzle can be positioned accurately at the selected
sampling points.
Filters of various compositions and geometries are used,
although glass fiber is most common. Any filter chosen must
be an efficient collector of submicron particles. A gas
meter and sometimes a calibrated orifice are used to measure
the gas flow rate and total volume sampled. Condensers are
used to measure the water vapor content as well as for pump
and meter protection. In some instances impinger bubblers
and liquid traps are used and the contents analyzed after
sampling for various volatile elements.
Many emissions contain substances which condense at
temperatures well above ambient to form solid or liquid par-
ticles. Care must be taken that the temperature in all parts
of the system upstream from the filter be kept at temperatures
36
-------
high enough to prevent condensation. Also, there is con-
siderable deposition on the surfaces of the nozzles, probes,
etc., and these must be thoroughly cleaned as part of normal
sampling and analysis.
Three systems which are commonly used to measure mass
emissions are described in the following paragraphs.
2.2.2 EPA-Type Particulate Sampling Train (Method 5f
Official performance testing of control devices on
stationary sources in the United States must be conducted
with the "EPA Method 5 Sampling Train" illustrated in Figure
13. A heated sample probe is used to transport the particulate
sample to a glass fiber filter which is maintained at 160°C
or more. A reverse type pitot tube is attached to the probe
to insure that isokinetic sampling conditions are maintained
during a traverse of the duct or stack. According to the
EPA method the glass fiber filter must have a penetration
value equal to or less than that of MSA 1106 BH (approximately
0.005% for a standard 0.3 urn DOP aerosol penetration test
at 5 cm/sec face velocity). Gases, vapors, and any particles
that penetrate the heated filter enter a series of impingers
or condensers that are immersed in an ice bath. They trap
the uncombined water that is present in the gas stream so
that the moisture content can be determined; they also pre-
vent the hot, humid gases from entering the gas metering
system and pump.
After the traverse, the filters are dried and the probe
is washed to remove and collect particulate matter from the
probe walls for subsequent analysis. The EPA Method 5 requires
37
-------
HEATED PROBE
\
IMPINGER TRAIN OPTIONAL:
MAY BE REPLACED BY AN
EQUIVALENT CONDENSER
HEATED
AREA FILTER HOLDER
THERMOMETER
THERMOMETERS
CHECK
VALVE
ORIFICE
*¥- rCX}r*H>0
MANOMETER
MAIN
VALVE
DRY TEST METER AIR TIGHT PUMP
VACUUM LINE
0700-14,16
Figure 13. EPA Method 5 Paniculate Sampling Train.
38
-------
that acetone be used for the pcobe wash. This creates a
problem when the long metal probes are hot from sampling,
so distilled water is often used instead of acetone. The?
probe wash liquid is collected and evaporated to dryness
so that the amount of particulate matter removed from the
probe can be weighed. This weight, and the weight of the
particulate matter on the filter, and the measured gas flow
are used to determine the mass emission rate of the source.
Approved systems of this type are currently available in
low {230-472 cm3/sec) (0.5-1 ft3/min.) and high (up to 170
LPS}(6 ft3/min.) flow rate configurations. A modification
to the EPA particulate sampling procedure (Method 17) has
been proposed to allow the use of in-stack filters (See Ap-
pendix G) .
2.2.3 ASTM-Type Particulate Sampling Train
The American Society of Testing and Materials has described
a particulate sampling train which is .illustrated in Figure
14. The main difference between this method and the EPA
Method 5 is the use of an in-stack particulate filter. With
this arrangement, a thimble-shaped filter is used to sample
high mass concentrations, and a conventional, disk-shaped,
filter is used for low mass concentrations. It is important
to heat the filter holder to insure that the filter temperature
is maintained above the dew point temperature if condensible
vapors are present in the gas stream. The advantage of this
system is that the particles are trapped before they enter
the probe and a probe wash is not required. Also, external
heating of the filter is often unnecessary. A condenser
and gas cooler are still required between the probe and the
gas metering system. The pitot tube, pump, and other parts
of the system are similar to the EPA Method 5 Sampling Train.
The thimble-filter system is not an EPA approved method but
it is often used in engineering tests to evaluate the per-
formance of a control device.
39
-------
SAMPLING
NOZZLE
GLASS FIBER THIMBLE FILTER
HOLDER AND PROBE(HEATED)
REVERSE-TYPE
PITOT TUBE
CHECK
VALVE
THERMOMETERS
ORIFICE
I
DRY TEST METER
AIR-TIGHT PUMP
07 00-14.17
Figure 14, ASTM Type Paniculate Sampling Train.
40
-------
2.2.4 • ASME-Type_Parti culate Sampling Ti:ajn
The American Society of Mechanical Engineers has
described in its Power Test Codes the use of a sampling
train to measure particulate emissions from industrial
sources. To meet the ASME specifications the participate
sampling train must have the following parts:
m
A tube or nozzle for insertion into the gas strea
and through which the sample is drawn.
A filter (thimble, flat disk, or bag type) for re-
moving the particulate. For the purpose of the Power
Test Code, 99.0% efficiency by weight is satisfactory.
A means of checking the equality of the velocity of
the gas entering the nozzle and the velocity of the
gas in the flue at the point of sampling.
A method by which the quantity of gas sampled is de-
termined .
A pump for drawing the gas stream through the nozzle,
filter, and metering device along with the necessary
tubing. It is important that the temperature of the
gas be above the dew point until after it has passed
the filter.
This sampling procedure is not very restrictive. Both
the EPA and ASTM particulate trains comply with the ASME re-
quirements .
41
-------
2.2.5 General Sampl i ng Pro cedar: es
Because the EPA Method 5 is required for all compliance
testing, it is the only one discussed in this section and
in Appendix B where the details of performing a mass emis-
sions test are described.
The general sampling procedure outlined herein is
presented in the EPA Test Method 5 "Determination of Par-
ticulate Emissions from Stationary Sources."2 Before sampling,
however, it is necessary to determine the number of sampling
points appropriate for the particular duct or stack under
consideration. EPA Test Method 1 "Sample and Velocity Tra-
verses for Stationary Sources"2 describes the computations
to determine the number of sampling points for both the velo-
city traverse and mass sampling traverse. The number of
points will depend on the size and shape of the duct. If
the velocity traverse indicates that the velocity profile
in the duct is very unstable, the number of sample points
should be increased to obtain a more accurate integrated
mass emission rate (See Figure 15).
The use of the S-Type pitot tube, and its calibration
for measuring the stack gas velocity and flow rate, is de-
scribed in the EPA Test Method 2 "Determination of Stack
Gas Velocity and Volumetric Flow Rate (Type S Pitot Tube)."2
The S-Type pitot tube is used because it is less susceptible
to clogging in high dust loading environments. It is also
advantageous to perform a temperature traverse of the duct
during the velocity traverse. This can be easily accomplished
by attaching a thermocouple temperature sensor near the end
42
-------
8 10 12
TRAVERSE POINTS
14 16
18 20
0700-11,5
Figure 15. Typical gas velocity distribution at the inlet to a precipitator.
43
-------
of the pitot. During actual mass sampling both the tem-
perature and gas velocity are monitored to allow isokinetic
sampling at each traverse point.
Mass sampling at electrostatic precipitators can pose
problems because of the dust concentration gradients across
the inlet and outlet ports. Precipitator inlet mass loadings
are usually high, while outlet loadings can be up to a fac
tor of 103 smaller. Extended sampling times are required
in many situations at precipitator outlets. Problems in
obtaining average inlet and outlet mass loadings can also
occur at sources which have cyclical operations. Outlet
mass concentrations at precipitators also fluctuate because
of the collection plate rapping cycles. The person preparing
for a sampling procedure must take all of these variables
into account when designing a test plan.
2.3 PARTICLE SIZE MEASUREMENT TECHNIQUES
2.3.1 General Discussion
Any detailed experimental program designed to evaluate
an electrostatic precipitator must include measurements on
the particle size distributions at the inlet and outlet.
These size distributions can then be used to calculate .the
precipitator collection efficiency versus particle size,
or "fractional efficiency curve".
Overall precipitator efficiency is strongly influenced
by the inlet particle size distribution. The migration velo-
city, defined in Paragraph 1.1.3, is equal to the product
44
-------
of the electric field (E) and the particle electrical mobility.
Figure 16 is a graph of electrical mobility versus particle
size for a laboratory aerosol which shows a minimum value
at approximately 0.1 to 0.3 pm diameter. Because of this
minimum in migration velocity, fine particles, which are
defined by the EPA as particles having aerodynamic diameters
smaller than 3 vim, are more difficult to collect than larger
particles. Unfortunately, fine particles also contribute
more to visible light scattering and opacity and present a
greater health hazard than do the larger particles.
At any given mass concentration, fine particle size
distributions accumulate far greater total charge than large
particle size distributions. The space charge associated
with fine particles frequently causes sparking at relatively
low voltage; this is sometimes a limiting factor in precipitator
performance. Although most of the mass emitted from a particu-
lar pollution source may consist of large particles, in gen-
eral, the largest number of particles is in the fine particle
range. Thus, high mass collection efficiency does not always
imply high number collection efficiency nor does it insure
that a particular opacity standard will be met.
An ideal particle size measurement device would be located
ir\ situ and give a real time readout of particle size distri-
butions and particle number concentration over the size range
from 0.01 ym to 10 pro diameter. At the present time, how-
ever, particle size distribution measurements are made using
several instruments which operate over limited size ranges
and do not yield instantaneous data.
Particle sizing methods may involve instruments which
are operated in~stackf or out of stack where the samples
45
-------
10
-6
rvi
i-7
1-
cc
10
,-8.
0.01
O E = 5.0 x 105 V/M
Nt = 8,0 x 1011 sec/M3
D E = 1.5 x 105 V/M
Nt = 3.2 x 1012 sec/M3
SHELLAC AEROSOL K= 3.2
O
O
"I
,0 o'
D
o o
D 0
0.1
PARTICLE DIAMETER, fim
1.0
3630-052
Figure 16. Particle mob's I sty as a function of diameter for shellac aerosol
particles charged in a positive ion field (after Cochet and
K is the dielectric constant of the aerosol.
46
-------
are taken using probes. For in--stack sampling, the sample
aerosol flow rate is usually adjusted to maintain near iso-
kinetic sampling conditions in order to avoid concentration
errors which result from under or oversampling large par-
ticles (dia. > 3 ym) which have too high an inertia to follow
the gas flow streams in the vicinity of the sampling nozzle.
Since many particulate sizing devices have size fractionation
points that are flow rate dependent, the necessity for iso-
kinetic sampling in the case of large particles can result
in undesirable compromises in obtaining data - either in
the number of points sampled or in the validity or precision
of the data for large particles.
In general, particulate concentrations within a duct
or flue are stratified to some degree with strong gradients
often found for larger particles and in some cases for small
particles. Such concentration gradients, which can be due
to inertial effects, gravitational settling, lane to lane
efficiency variations in the case of electrostatic precipi-
tators, etc., imply that multipoint (traverse) sampling must
be used.
Even the careful use of multipoint traverse techniques
will not guarantee that representative data are obtained.
The location of the sampling points during process changes
or variations in precipitator operation can lead to significant
scatter in the data. As an example, rapping losses in dry
electrostatic precip.itators tend to be confined to the lower
portions of the gas streams, and radically different results
may be obtained, depending on the magnitude of the rapping
losses, and whether single point or traverse sampling is
used. In addition, large variations in results from successive
47
-------
multipoint traverse tests can occur as a result of dif-
ferences in the location of the sampling points when the
precipitator plates are rapped. Similar effects will occur
in other instances as a result of process variations and
stratification due to settling, cyclonic flows, etc.
Figure 17 illustrates temporal concentration variations due
to rapping for two particle sizes at a point in space loca-
ted in a duct immediately downstream of a dry electrostatic
precipitator.
Choices of particulate measurement devices or methods
for individual applications are dependent on the availability
of suitable techniques which permit the required temporal
and/or spatial resolution or integration. In many instances
the properties of the particulate are subject to large changes
in not only size distribution and concentration, but also
in chemical composition (for example, emissions from the
open hearth steel making process). Different methods or
sampling devices are generally required to obtain data for
long term process averages as opposed to the isolation of
certain portions of the process in order to determine the
cause of a particular type of emission
Interferences exist which can affect most sampling
methods. Two commonly occurring problems are the condensation
or vapor phase components from the gas stream and reactions
of gas, liquid, or solid phase materials with various por
tions of the sampling systems. An example of the latter
is the formation of sulfates in appreciable (several milli-
gram) quantities on several of the commonly used glass fiber
filter media by reactions involving SO and trace constitu-
ents of the filter media. Sulfuric acid condensation in
cascade impactors and in the probes used for extractive
sampling is an example of the former.
48
-------
UPPER CURVE 1.5-3,0 urn PARTICLES
LOWER CURVE 6-12 urn PARTICLES
MAJOR RAPPING PUFFS INDICATED
BY ARROWS
TIME, minutes
3630 058
Figure 77. Relative concentrations of particles in two size ranges between
and during rapping puffs as observed at the exit of a cold side
precipitator collecting fly ash from a coal fired boiler.
49
-------
If extractive sampling is used and the sample is con-
veyed through lengthy probes and transport lines, as is the
case with several particulate sizing methods, special at-
tention must be given toward recognition, minimization, and
compensation for losses by various mechanisms in the trans
port lines. The degree of such losses can be quite large
for certain particle sizes.
In this section, individual particulate sampling systems
and procedures are discussed. These are categorized according
to the physical mechanism that is used to obtain the data:
inertial, optical, electrical, or diffusional.
2.3.2 Inertial Particle Sizing Devices
Two devices which fall into the inertial sizing category
are impactors and cyclones. In both of these devices, the
aerosol stream is constrained to follow a path of such curva-
ture that the particles tend to move radially outward toward
a collection surface because of their inertia. Subsequent
analysis of the particle size distribution may be made by
gravimetric means, quantitative chemical analysis, or micro
scopic inspection.
Particle size distribution measurements related to preci-
pitator evaluation have largely been made using cascade impac-
tors, which are effective in the size range from 0.3 to 20 ym
diameter? although, in some cases, hybrid cyclone-impactor
units, or cyclones have also been used. The particle size
distributions are normally calculated from experimental data
by relating the mass collected on various stages to the theo-
retical or calibrated size cutpoints associated with those
stage geometries.
50
-------
Cascade impactors - Because of its compact arrangement
and mechanical simplicity, the cascade impactor has gained
wide acceptance as a practical means of making particle size
measurements in flue gases. In most cases, the impactors
can be inserted directly into the duct or flue, eliminating
many condensation and sample loss problems which occur when
probes are used for extractive sampling.
Figure 18 is a schematic which illustrates the principle
of particle collection which is common to all cascade impactors.
The sample aerosol is constrained to pass through a slit
or circular hole to form a jet which is directed toward an
impaction surface. Particles which have lower momentum will
follow the air stream to lower stages where the jet velocities
are progressively higher. For each stage there is a character-
istic particle size which theoretically has a 50% probability
of striking the collection surface. This particle size,
or Dsoi is called the effective cut size for that stage.
The number of holes or jets on any one stage ranges from
one to several hundred depending on the desired jet velocity
and total volumetric flow rate. The number of jet stages
in an impactor ranges from one to about twenty for various
impactor geometries reported in the literature. Most commer-
cially available impactors have between five and ten stages.
Parameters which determine the collection efficiency
for a particular geometry are the gas viscosity,- the particle
density, the jet diameter or width, the jet-to-plate spacing,
and the velocity of the air jet.
Most modern impactor designs are based on the semi-em-
pirical theory of Ranz and Wong.1* More comprehensive theories
51
-------
TRAJECTORY OF
LAHGE PARTICLES
'TRAJECTORY.OF
SMALL PARTICLES
JET
GAS STREAMLINES
IMPACTION SURFACE
A, TYPICAL IMPACTOR JET AND COLLECTION PLATE
s 10°
o
z
LU
O
o
o
11!
O
CJ
50
U50
PARTICLE DIAMETER
B. GENERALIZED STAGE COLLECTION EFFICIENCY CURVE
3S30-O59
Figure 18. Operation principle and typical performance for a
cascade impactor.
52
-------
have been developed by Davies and Aylward5 and by Marple."
In practice, deviations from ideal behavior in actual impac-
tors dictate that they be calibrated experimentally.
A large number of experimental studies have been
published on cascade impactor design and performance in the
laboratory environment. Most of these have been reviewed
in the dissertations of Marple5 and Rao.7 Recently,
Gushing et alfl have presented calibration data on several
commercially available cascade impactors. Figures 19, 20,
21, 22, and 23 show schematics of the commercial impactor
designs which are commonly used in source testing. Table
I gi.ves a listing of the manufacturers, and some operational
information for stack sampling. The details of cascade impac
tor applications are discussed in Appendix C.
It is usually impractical to use the same impactor at
the inlet and outlet of a precipitator when making fractional
efficiency measurements because of the large difference in
particulate loading. For example, if a sampling time of
thirty minutes is adequate at the inlet, for the same im-
pactor operating conditions and the same amount of sample
collected, approximately 3000 minutes sampling time would
be required at the outlet (a collection efficiency of 99%
is assumed). Although impactor flow rates can be varied,
they cannot be adjusted enough to compensate for this dif-
ference in particulate loading without creating other
problems. Extremely high sampling rates result in particle
bounce and in scouring of impacted particles from the lower
stages of the impactor where the jet velocities become ex-
cessively high. Short sampling times may result in atypical
samples being obtained as a result of momentary fluctuations
in the particle concentration or size distribution within
the duct. Normally, a low flow rate impactor is used at
53
-------
JET STAGE (9 TOTAL)
SPACERS
GLASS FIBER
COLLECTION
SUBSTRATE
NOZZLE
INLET
BACKUP
FILTER -
PLATE
HOLDER-
CORE
07OO-14.3
Figure 19. Andersen Mark III Stack Sampler,
54
-------
NOZZLE
D
PRECOLLECT10IM
CYCLONE
JET STAGE
(7 TOTAL)
COLLECTION
PLATE
SPRING
0700-14,1
Figure 20. Modified Brink Model BMS-11 Cascade Impactor.
55
-------
NOZZLE
INLET JET STAGE NO. 1
STAGE NO. 2
STAGE NO. 3
STAGE NO. 4
STAGE NO. 5
STAGE NO, 6
STAGE NO.7
FILTER
IMPACTOR BASE
3630 046
Figure 21. MR I Model 7502 Inert ial Cascade Impact or.
5fi
-------
NOZZLE
INLET CONE
STAGE 0
STAGE 1
STAGE 2
STAGE 3
STAGE 4
STAGE 5
FILTER
SUPPORT
IMPACTOR
BASE
3630 053
Figure 22. Sierra Model 226 Cascade Impactor.
57
-------
JET STAGE 0-RING
COLLECTION PLATE
INLET
FILTER HOLDER
COLLECTION
PLATE (7 TOTAL]
\
JET STAGE
(7 TOTAL)
0700-14.2
Figure 23. University of Washington Mark III Source Test
Cascade fmpactor.
58
-------
TABLE I
COMMERCIAL CASCADE IMPACTOR SAMPLING SYSTEMS
Name
Andersen Stack Sampler
(Precollection Cyclone
Avail.)
Univ. of Washington
Mark III Source Test
Cascade Impactor
(Precollection Cyclone
Avail.)
Brink Cascade Impactor
(Precollection Cyclone
Avail.)
Nominal Flow rate
(cm /sec)
236
236
Substrates
Glass Fiber (Available from
manufacturer)
Stainless Steel Inserts,
Glass Fiber, Grease
14.2
Glass Fiber, Aluminum,
Grease
Manufacturer
Andersen 2000, Inc.
P.O. Box 20769
Atlanta, GA 30320
Pollution Control
System Corp.
321 Evergreen Bldg,
Renton, WA 98055
Monsanto EnviroChem
Systems, Inc.
St. Louis, MO 63166
Sierra Source Cascade
Impactor - Model 226
(Precollection Cyclone
Avail.)
MRI Inertial Cascade
Impactor
in
„£>
118
236
Glass Fiber (Available
from manufacturer)
Stainless Steel, Alumi-
num, Mylar, Teflon.
Optional: Gold, Silver,
Nickel
Sierra Instruments, Inc.
P.O. Box 909
Village Square
Carmel Valley, CA 93924
Meteorology Research, Inc,
Box 637
Altadena, CA 91001
-------
the inlet and a high flow rate .impactor at the outlet. The
impactors are then operated at their respective optimum flow
rates, and the sampling times are dictated by the time re-
quired to collect weighable samples on each stage without
overloading any single stage.
Seriescyclone particulate sampling techniques - Pro-
totype series cyclone sampling systems have been developed
for industrial source sampling.9 In general, series cyclones
are easy to use, trouble-free, and efficient collectors of
large quantities of size segregated particulate. Their main
drawbacks are that their size limits the number of size segre-
gated samples during each test that can be obtained as com-
pared to most commercial impactors and an accurate theory
of operation has not been developed. However/ the ability to
collect large quantities of sized material for analysis makes
these devices irreplaceable for some applications.
The Source Assessment Sampling System (SASS) incorporates
three cyclones and a back up filter.10 A schematic of this
system is illustrated in Figure 24. It is operated at a
flow rate of 3065 cm3/sec (6.5 ft*/min) with approximate
cyclone cut points of 10, 3, and 1 micrometer aerodynamic
diameter and a gas temperature of 205°C. Besides obtaining
information on the particle size distribution, this system
collects gram quantities of particles for later chemical
or biological analyses. The SASS train is large and requires
extractive sampling through a heated probe. The cyclones
are mounted in an oven to keep the air stream at stack tem-
perature or above the dew point until the particles are
collected. This system is supplied with nozzles, a probe-
pitot-tube-thermocouple assembly, cyclones, back up filter,
oven, a gas conditioning chamber, and a flow metering device
pump adapted from the Aerotherm High Volume Sampling System.
60
-------
HEATER
CONTROLLER
FILTER
GAS COOLER
OVEN
T.C.
XAD-2
CARTRIDGE
GAS
TEMPERATURE
T.C.
CONDENSATE
COLLECTOR
IMP/COOLER
TRACE ELEMENT
COLLECTOR
DRY GAS METER
ORIFICE METER
CENTRALIZED TEMPERATURE
AND PRESSURE READOUT
CONTROL MODULE
10 CFM VACUUM PUMP
1MPINGER
T.C.
3(330 Obi
Figure 24. Schematic of the Source Assessment Sampling System.
-------
Cyclone calibration details are furnished along with equa-
tions for calculating approximate cyclone outpoints for opera-
ting conditions other than those measured during the calibra-
tion procedure.
It is mandatory that the gas velocity and temperature
through the cyclones be maintained at a constant setting
while sampling, because the cyclone cut points are dependent
upon the gas flow rate and temperature. This usually means
that periods of nonisokinetic sampling may occur. Depending
on the magnitude of fluctuations in the velocity of the
sampled stream, this may or may not introduce significant
errors in the sizing process.
Southern Research Institute, under EPA sponsorship,
has designed and built a prototype three-stage series cyclone
for in-stack use. A sketch of this system is shown in Figure
25. It is designed to operate at 472 cm3/sec (1 ft3/min.).
The calibrated cut points for these cyclones are 3.0, 3..S,
and 0.6 micrometer aerodynamic at 21°C. A 47 mm Gelman filter
holder is used as a back up filter after the last cyclone.
This series cyclone system was designed for in-stack use
and requires a six inch sampling port. A sampling system
similar to that for a high flow rate impactor is usually
adequate, although a more powerful pump may be required under
some sampling situations. As with the SASS train, a constant
flow rate through the cyclone system is required to maintain
stable cyclone cut points.
Figure 26 shows a second generation EPA/Southern Re-
search series cyclone system under development which con-
tains five cyclones and a back up filter. It is a compact
system and will fit through 4 inch diameter ports. The
62
-------
TO PUMP
BACK-UP FILTER-
CYCLONE 2 •»
KU
CYCLONE 3
NOZZLE
CYCLONE 1
3630 055
Figure 25. Three Stage Series Cyclone System,
63
-------
CYCLONE 1
CYCLONE 4
CYCLONE 5
CYCLONE 2
CYCLONE 3
OUTLET
INLET NOZZLE
3630-055
Figure 26. Five Stage Series Cyclone System.
64
-------
Initial prototype was made of anodized aluminum with stain-
less steel connecting hardware. A second prototype, for
in-stack evaluation, is made of titanium.
Figure 27 shows laboratory calibrations of the five
cyclone prototype system. The cut points, at the test con-
ditions are 0.32, 0.6, 1.3, 2.6, and 7.5 ym. A continuing
research program includes studies to investigate the depen-
dence of the cyclone cut points upon the sample flow rate
and temperature so that the behavior of the cyclones at stack
conditions can be predicted more accurately.
Small cyclone systems appear to be practical alterna-
tives to cascade impactors as instruments for measuring
particle size distributions in process streams. Cyclones
offer several advantages:
Large, size segregated samples are obtained.
There are no substrates to interfere with analyses.
They are convenient and reliable to operate.
They allow long sampling times under high mass loading
conditions for a better process emission average.
They may be operated at a wide range of flow rates with-
out particle bounce or reentrainment.
On the other hand, there are some negative aspects of
cyclone systems which require further investigation:
65
-------
100
90
80
8*
>" 70
o
| 60
LL.
Si 50
I «
u
LU
;} 30
o
0 20
10
°0.2 0.3 0.40.50.6 0.8 1.0 2 34568 10 20
PARTICLE DIAMETER, micrometers
• FIRST STAGE CYCLONE
• SECOND STAGE CYCLONE
6 THIRD STAGE CYCLONE
V FOURTH STAGE CYCLONE
O FIFTH STAGE CYCLONE
3630-057
Figure 27. Laboratory Calibration for the Five Stage Series
Cyclone System. (472 cm^/sec, particle
density—1.0 gm/cm3)
66
-------
Unduly long sampling times may be required to obtain
large samples at relatively clean sources.
The existing theories do not accurately predict cyclone
performance.
Cyclone systems are bulkier than impactors and may require
larger ports for in-stack use.
As discussed above, cyclones are now used on an experi-
mental basis by the EPA and EPA contractors If current
research programs ace successful in developing a better under-
standing of cyclone behavior, they may play a very important
role in control device evaluation.
2'. 3.3 Optical Measurement Techniques
The basic operating principle for one type of optical
particle counter is illustrated in Figure 28. Light is
scattered by individual particles as they pass through a
small viewing volume, the intensity of the scattered light
being measured by a photodetector. The sizes of the particles
determine the amplitude of the scattered light pulses, and
the rate at which the pulses occur is related to the particle
concentration. Thus, a counter of this type gives both size
and concentration information. The simultaneous presence
of more than one particle in the viewing volume is interpreted
by the counter as a larger single particle. To avoid errors
arising from this effect, dilution to about 300 particles/
cm3 is generally necessary. Errors in counting rate also
occur as a result of electronics deadtime and from statistical
effects resulting from the presence of high concentrations
of sub-countable (D < 0.3 ym) particles in the sample gas
stream.11 The intensity of the scattered light depends upon
67
-------
INLET
SENSOR
CHAMBER. X \
PHOTOMULTIPL1ER
CALIBRATOR
LAMP
3630 060
Figure 28. Operating principle for an optical particle counter. Courtesy
of Climet Instruments Company.
68
-------
the viewing angle, particle index of refraction, particle
optical absorptivity? and shape, in addition to the particle
size. The schematic in Figure 28 shows a system which uti
lizes "integrated near forward" scattering. Different view-
ing angles might be chosen to optimize some aspect of the
counter performance. For example, near forward scattering
minimizes the affect of variations in the indicated particle
size with index of refraction, but for this geometry, there is
a severe loss of resolution for particle diameters near 1 urn.
Right angle, or 90° scattering smooths out the response curve,
but the intensity is more dependent on the particle index
of refraction.
Available geometries are:
Bausch & Lomb 40-1 Near Forward Scattering
Royco 220 Right Angle
Royco 245, 225 Near Forward
Climet CI-201, 208 Integrated Near Forward
Optical particle counters have not been used extensively
in stack sampling because they cannot be applied directly to
the effluent gas stream. The sample must be extracted, cool-
ed, and diluted; a procedure which requires great care to
avoid introducing serious errors into calculations of the
particle size distribution. The main advantage of optical
counters is the capability of observing emission fluctuations
in real time. After extraction, the useful particle size
range is approximately 0.3 to 1.5 vim.
2.3.4 Ultrafine Particle Sizing Techniques
There are two physical properties of ultrafine particles
(diameter < 0.5 ym) which are size dependent and which can
69
-------
be predicted with sufficient accuracy under controlled con-
ditions to be used to measure particle size. These are the
particle diffusivity and electrical mobility. Although ultra-
fine particle size distribution measurements are still in
a developmental stage, instruments are available which can
be used for this purpose, and some field measurements have
been made. A practical limitation on the lower size limit
for this type of measurement is the loss of particles by
diffusion in the sampling lines and instrumentation. These
losses are excessive for particle sizes below about 0.01 pm
where the samples are extracted from a duct and diluted
to concentrations within the capability of the sensing devices
Piffusipnal sizing - Diffusion batteries may consist
of a number of long, narrow, parallel channels, a cluster
of small bore tubes, or a series of screens. Figure 29a
shows a typical parallel channel diffusion battery, and
Figure 29b shows the aerosol penetration characteristics
of this geometry at two flow rates. The parallel plate geom-
etry is convenient because of ease of fabrication and the
availability of suitable materials, and also because sedi-
mentation can be ignored if the slots are vertical, while
additional information can be gained through settling, if
the slots are horizontal.
Sinclair12 and Breslin et al13 report success with more
compact, tube-type and screen-type arrangements in laboratory
studies and a commercial version of Sinclair's geometry is
available.* Although the screen-type diffusion battery must
be calibrated empirically, it offers convenience in cleaning
*Thermo-Systems, Inc., 2500 N. Cleveland Ave., St. Paul,
Minnesota 55113.
70
-------
0700 14.11
Figure 29a. Parallel plate diffusion battery.
20
0.01
PARTICLE DIAMETER,
0700-14.12
Figure 29b, Parallel plate diffusion battery penetration curves for
monodisperse aerosols (12 channels, 0,1 x 10 x 48 cm).
71
-------
and operation, and compact size. Figure 30 shows Sinclair's
geometry. This battery is 21 cm long, approximately 4 cm
in diameter, and weighs 0.9 kg.
Variations in the length and number of channels (tubes,
or screens) and in the aerosol flow rate are used as means
of measuring the number of particles in a selected size
range. As the aerosol moves in streamline flow through the
channels, the particles diffuse to the walls at a predictable
rate, depending on the particle size and the diffusion bat-
tery geometry. It is assumed that every particle which reaches
the battery wall will adhere, therefore, only a fraction
of the influent particles will appear at the effluent of
a battery. It is only necessary to measure the total number
concentration of particles with a condensation nuclei counter
at the inlet and outlet to the diffusion battery under a
number of conditions in order to calculate the particle
size distribution.
Diffusional measurements are less dependent upon the
aerosol parameters than the other techniques discussed and
perhaps are on a more firm basis from a theoretical standpoint.
Disadvantages of the diffusional technique are the bulk
of the diffusional batteries, although advanced technology
may alleviate this problem; the long time required to measure
a size distribution; and problems with sample conditioning
when condensible vapors are present.
Electrical particle counters - The most complete set
of experiments performed in order to determine the relation-
ship between particle size and charge were reported by Hewitt1*
72
-------
SAMPLING
PORT !TYP)
SECTION CONTAINING
SCREENS (TYPS
3630-045
Figure 30. Screen type diffusion battery. The battery is 21 cm
long, 4 cm in diameter, and contains 55, 635 mesh
stainless steel screens.
73
-------
in 1957. These experiments confirmed theoretical predictions
that there exists a unique charging rate for each particle
size if the charging region is homogeneous with respect to
space charge density and electric field.
In the course of his work, Hewitt developed a mobility
analyzer, shown in Figure 31. Charged particles enter through
the narrow annular passage A, and experience a radial force
toward the central cylinder due to the applied field. By
moving the sampling groove B, axially, or by varying the
magnitude of the applied field, the mobility of the charged
particles can be measured. In Hewitt's experimental work,
the particle size was known, and the mobility determined
the charge. If, however, the particle charge were known,
the mobility would determine the particle size. This concept
has been used by Whitby, Liu, et alL 3 at the University of
Minnesota to develop a series of Electrical Aerosol Analysers
(EAA). A ruggedized field test unit based on the earlier
University of Minnesota designs is now commercially available.*
A schematic of this system is shown in Figure 32.
The EAA has the distinct advantage of very rapid data
acquisition compared to diffusion batteries and condensation
nuclei counters (two minutes as opposed to two hours for
a single size distribution analysis.)
Disadvantages of this type of measurement system are
difficulties in predicting the particle charge, and the frac-
tion of the particles bearing a charge, with sufficient accu-
racy, and the requirement for sample dilution when making
particle size distribution measurements in flue gases.
*Thermo-Systems, Inc., 2500 N. Cleveland Ave., St. Paul,
Minnesota 55113.
74
-------
CLEAN
AIR
j
f
1
-Hi
1 1-
1
L
|
1
\
1 1
»%M
r^
B
x.
**v
'
1
1
/—
AEROSOL
GENERATOR
i
1
CHARGER
1
r
if 4
i
i
V\
V
. *
:L VARIABLE
DIRECT
_ VOLTAGE
AEROSOL
CONCENTRATION
MEASUREMENT
j
fe
AEROSOL
DISCHARGER
D7DO-14.13
Figure 31. Coaxial cylinder mobility analyser. Charged aerosol enters
at A. Sampling groove is at B in inner cylinder, which is
adjusted axiatly. After Hewitt (14).
75
-------
CONTROL MODULE
ftNflLTZER TUTPUT SIGNAL -
CYCLE START COMMANU -
CYCLE RE.5LT COMMflHD
AEROSOL FLQWMETLR HtADOiH
CH6HJER CURRENT RtAOCUT
---•CHWCER VOLTAGE READOUT
AUTOUA1IC HKTH VULJKSL CONTROL flND HEAOOUT
ALYZLR CURRENT! NEiTdu
TQT4L FLOWMET£R READCHT
~~ «-TO VACUUM HIJWP
-I* EXTERNAL
-* OfcTA
—•ACQUISITION
• SYSTEM
3630-043
Figure 32, Flow schematic and electronic block diagram of the Electrical
Aerosol Analyser. Liu and Put (15).
-------
Figures 33, 34, and 35 show typical particle size dis-
tributions and a precipitator fractional efficiency curve
measured in the field with impactors, an optical particle
counter, and a diffusional ultrafine apparatus.
The details of sampling with optical particle counters,
diffusion batteries with condensation nuclei counters, and
EAA using extractive sampling and dilution techniques are
presented in Appendix D.
2.4 PARTICULATE RESISTIVITY MEASUREMENTS
2.4.1 Gen e r a 1 D iscussion
The resistivity of the dust collected in a precipitator
has a direct influence on the efficiency of collection.
If the resistivity is greater than about 5 x 1010 ohm-cm,
the electrical field developed in the collected particulate
layer can exceed the breakdown field strength. Excessive
spark rates and back corona can occur, necessitating preci-
pitator operation at lower current densities with resul-
tant degraded performance. If the particulate resistivity
is less than about 107 ohm-cm, the electrical forces holding
the dust to the collection plates will be low. Excessive
reentrainment can occur yielding lower performance. There-
fore, to perform a complete evaluation of a precipitator
installation, it is important to determine the particulate
resistivity.
The particulate resistivity can be determined in the
laboratory or in the field. In the laboratory, the dust
sample is placed in a suitable test cell, and the current
passed through the dust layer under a given voltage is
77
-------
108
TO7
E
3
O 106
o
UJ
_i
o
105
Q.
u 104
D
u
103
1Q2
D1FFUSIONAL DATA
OPTICAL DATA
SEDIMENTATION SIZES
0.01 0.02 0.05 0.1 0,2
MINIMUM PARTICLE DIAMETER, /im
0.5
1.0
3G30-090
Figure 33, Cumulative particle number concentration. Diffusions/ and
optical data are indicated.
-IB
-------
0.10
0.01
Q
<
o
t/1
U)
0.001
0.1
1.0
10.0
DIAMETER,
100.0
3S3D 083
Figure 34. Precipitator outlet particle size distribution obtained with
an Andersen Mark IK Stack Sampler.
-------
c
o>
o
O
z
LU
u.
UJ
o
UJ
O
(J
99.98
99.9
99.8
99.5
99
98
95
90
60
301
0.05
MEASUREMENT METHOD:
ACASCADE IMPACTORS
O OPTICAL PARTICLE COUNTERS
• DIFFUSIONAL
PRECIPITATOR CHARACTERISTICS:
TEMPERATURE - 335°C
SCA • 85 M2/(M3/sec)
CURRENT DENSITY - 35 nA/CM2
0.1
0.5 1.0
PARTICLE DIAMETER,
5.0
10.0
3630-083
Figure 35. Measured fractional efficiencies for a hot side electrostatic
precipitator installed on a pulverized coal boiler. The
operating parameters are indicated.
30
-------
measured. Resistivity is calculated from these values and
known geometrical factors. For field determinations of re-
sistivity, a variety of resistivity probes are available.
These probes are inserted into the gas stream to extract
or collect the dust either electrostatically or inertially.
Resistivity can be determined from the V-I curves generated
if electrostatic collection is used or from the measured
values of current passing through the collected dust layer
under given applied voltages and geometrical factors.
Since the resistivity is determined while the dust is in
its original or natural environment, the data are referred
to as in situ resistivity values. The equipment and pro-
cedures used for measuring resistivity in the laboratory
and field have been recently reviewed.16
The question of whether laboratory or in_ situ resistivity
data should be acquired depends upon many factors. The
question is somewhat simplified when the viewpoint is restric-
ted to the evaluation of existing precipitator performance
characteristics.
First, consider a "hot side" installation operating
in the temperature range of 200°C to 400°C. It is generally
accepted that in this temperature range the resistivity is
not significantly influenced by environmental factors.
Therefore it is assumed that in situ and laboratory mea-
surements will give comparable results. However, too little
in _s_i tu_ information is available to substantiate this as-
sumption. Since high temperature probes are not presently
generally available, laboratory data are usually used.
If a "cold side" installation operating between 100°C
and 200°C is to be evaluated, the in. situ data are recom-
mended. In this temperature range the resistivity can be
81
-------
affected greatly by the environment. Furthermore in evaluating
precipitator performance, one generally desires data taken
simultaneously with the measurement of other operating char-
acteristics. Laboratory data have been successfully used
for situations in which no conditioning agents were being
used and the particulate was generated from low sulfur coals.
There is an ongoing attempt by several investigators to de-
velop a sophisticated laboratory technique which will allow
one to utilize laboratory resistivity data when evaluating
"cold side" precipitator operation.
2.4.2 Labor atory Determination of Particulate Res i_s t iyity
There are two almost identical standards available for
the determination of bulk particulate resistivity in the
laboratory. These are the ASME Power Test Code 2817 and
the APCA Informative Report No. 2.18 The preferable dust
specimen for these tests is an isokinetically obtained high
volume sample taken at the precipitator inlet. Samples used
include: hopper samples proportionately blended based on
a known precipitator efficiency and the number of fields,
samples from individual hoppers, and samples from ash stor-
age facilities. The hardware, environmental control, and
procedures recommended by the above standards are given in
detail in Appendix E. Figure 36 shows typical resistivity
data as a function of temperature and water concentration
in an air environment obtained using these standards with
exception of recommended voltage gradient. Many laboratories
use equipment and procedures significantly different from
the above standards. These variations have developed with
82
-------
10
13
1012
10"
o
E
.c
O
>
>
(-
LLJ
oc
109
O 9.1 v/o H2O
0 12.7 v/o H2O
D 4.9 v/o H20
3.2
39
102
2.8
84
183
2.4
144
291
2,0
227
440
TEMPERATURE
1.6 1DOO/T(°K|
352 °C
666 °F
3630 095
Figure 36. Typical resistivity versus temperature curves for fly ash at
three water vapor concentrations.
83
-------
respect to specific research goals and in efforts to elimi-
nate observed shortcomings. An example of an alternate pro-
cedure is also detailed in Appendix E.
2.4.3 In Situ Particulate Resistivity Measurement
An in situ determination of particulate resistivity
makes use of a point-to-plane probe which can be inserted
directly into the gas stream upstream of the precipitator.
A high voltage is applied to the point-to-plane electrode
system such that a corona is formed in the vicinity of the
point. The dust particles are charged by ions created in
the corona, and precipitated by electrical forces onto
the collection plate. Thus, the probe is intended to simu-
late the behavior of a full-scale electrostatic precipitator
and to measure the resistivity of the dust in a way that
simulates the operating condition.
In the point-to-plane technique, two methods of making
resistivity measurements on the same sample may be used.
The first is the "V-I" method. In this method, voltage-cur-
rent curves for the point-plane system are obtained before
and after the electrostatic deposition of the dust.
Alternatively, a disc the same size as the collecting
disc is lowered onto the collected sample. Increasing vol-
tages are then applied to the dust layer and the currents
recorded until the dust layer breaks down electrically and
sparkover occurs. More information can be found in the paper
1 6
by Nichols concerning the advantage and disadvantage to using
in situ resistivity probes.
Several types of resistivity probes have been developed
and are currently available. In Appendix F, full details for
the use of one particular probe are given.
84
-------
2.5 PROCESS EFFLUENT GAS ANALYSIS
2.5.1 General Discussion
In evaluating the performance of a precipitator it is
advantageous to know something of the constituents of the
effluent gas stream, excluding the particulate matter.
A gas analysis generally concerns the amount of N2,
02, CO, C02, and H20 in the effluent. These can usually
be determined using an ORSAT apparatus or one of the EPA
Stationary Source Test Methods developed for gas analysis.
Experimental studies on modifying combustion and condition-
ing the particles to modify their resistivity, and the ef-
fect this has on emissions, has resulted in the need for
measuring the amounts of several other gas constituents such
as NH3, S02, S03, NO , etc. These measurement methods are
briefly described in the following section.
2.5.2 Qualitative Gas Analysis
Interviewing key plant personnel or obtaining the proper
records should be sufficient to determine the qualitative
nature of the process gas stream. This data should include
the average fractional amounts of C02, CO, 02, N2, and H2O
in the process effluent gas stream during normal operation.
Depending on the industrial process under consideration,
there may also be measurable amounts of NH3, S02, 50^ , NO ,
J\
HF, sulfuric acid mist, or other volatile substances present.
Other qualitative information which should be gathered
includes the average gas temperature at the precipitator
inlet and outlet, and the average actual and standard volu-
metric flow rates through the precipitator.
85
-------
2.5.3 Quantitative Gas Analysis
Flue gas constituents normally specified for analysis
are Nzf Oz, CO, and CO2, and H20. In addition to these anal-
ysis, S02 and S03 concentrations are usually measured and
sometimes the NO , HF- or other vapor concentrations are
X
determined.
Oxygen. CO, and CO2 concentrations are measured with
a commercial Orsat-type apparatus. Two Orsat-type analyzers
are used to determine the oxygen content of the gas entering
and leaving the precipitator simultaneously. Comparisons
of the inlet and outlet oxygen concentration provides a
check for leakage of gas into or out of the precipitator.
Although in principle leakage can be determined from an
examination of inlet and outlet gas velocity profiles, in
practice, flow disturbances at available sampling locations
often severely limit the accuracy with which flow determina-
tion can be made. Therefore, the simultaneous inlet and
outlet oxygen determinations may be a more sensitive indica-
tor of the physical integrity of the precipitator casing.
The Environmental Protection Agency has developed several
Stationary Source Test Methods for the determination of var-
ious flue gas components. These methods, of course, are
not the only way by which the quantity of the gases can be
measured. There are many other acceptable analytical methods
developed by different testing societies such as the ASME
and ASTM. EPA Test Method Number 4 describes a procedure
to determine the H20 content of the flue gas. EPA Test
Method Number 6 can be used to determine the S02 content
of the flue gas. EPA Test Method Number 7 explains a method
for measuring the nitrogen oxide in the flue gas. The amount
of sulfuric acid mist and S02 content can be determined using
the EPA Test Method Number 8.
86
-------
Figure 37 illustrates a system not described in the
Federal Register, which has been found to be accurate and
convenient to use for measurements of SO2 and S03 concen-
trations. This is the Controlled Condensation System de-
veloped for the EPA by TRW Systems Group19 to accurately
determine S02 and SQa concentrations in process gas streams.
This procedure is applicable in high or low mass loading
environments with temperatures up to 300°C and S02 concen-
trations up to 6000 ppm.
The Controlled Condensation System is based on the
separation of SO3 as HjSQt from S02 by cooling the gas stream
below the dewpoint of H2S04f but above the H2O dewpoint.
Cooling is accomplished by a water-jacketed coil where the
H2SCU is collected. Particulate matter is collected by a
quartz filter mat inserted in the line prior to the conden-
sation coil. The particulate filter system is maintained
at a temperature of 288°C to insure that none of the H2SOH
will condense on the filter mat or filter holder.
The Controlled Condensation Coil (CCC) is a modified
Graham Condenser. The water jacket is maintained at 60°C.
This is adequate to reduce the flue gas below the dewpoint
of the HaSOn. Following the CCC are two impingers for re-
moving the S02 and H20. The SO2 scrubber is a bubbler
filled with a 3% solution of H202 in water. The water vapor
is removed by a silica gel filled impinger. A vacuum pump
with a capacity of 472 cm3/sec is recommended. The total
volume of gas sampled is measured with a dry gas meter.
The sampling rate/time is normally 135 cm3/sec for one
hour. An indication of the proper amount of sample comes
87
-------
FLUE-GAS
SAMPLE
1
I
,WALL OF FLUE
HEATED
' SAMPLING PROBE
. SO., CONDENSER
o
S02 ABSORBER
fPEROXIDE-WATER
SOLUTION)
VENT
VDR1ERITE OR
COLD TRAP
3B3O-092
Figure 37, Schematic diagram of apparatus for the collection of
by the condensation method.
33
-------
from watching the condenser coil. When the HzSCK fog has crept
about one-half to two-thirds of the way along the coil, the
sampling can cease. After sampling has been completed, the
probe and coil are rinsed with deionized HaO and the recovered
solution is analysed in the lab. The amount of H2SOn in the
condensation coil and probe can be determined by a sulfate
or H titration. Because of its simplicity and sensitivity,
the H titration is preferred. This recommended acid/base
titration uses Bromophenol Blue as the indicator, since the
endpoint of the NaOH and HaSCK titration falls near the pH
range (3-4.6) of the Bromophenol Blue color change. Complete
details can be found in the reference above.
89
-------
3. DEVELOPMENT OF TEST PLANS FOR ELECTROSTATIC PRECIPITATOR
EVALUATION
3.1 GENERAL DISCUSSION
Details of precipitator operation and particulate sam-
pling procedures are given in the preceding sections and
the appendices. This section is intended as a summary or
outline to be used in the development of a test plan. Spe-
cific references to other more detailed sections are also
included.
The scope of precipitator tests ranges from short visits
to the site where the plant personnel are interviewed and
the precipitator operating conditions are observed, to
very elaborate experimental programs where it is attempted
to completely characterize the properties of the flue gas
aerosol and the electrical and mechanical conditions of the
precipitator. Short visits may be sufficient to obtain data
for diagnosis of precipitators which are malfunctioning,
but more elaborate tests are almost always required if the
precipitator is a poor or inadequate design, or if the proper-
ties of the dust limit performance.
For the purposes of developing test plans, it is possible
to define three categories or levels of effort. These are
denoted in this document by Level A, Level B, or Level C.
A Level A test consists of the minimum effort that can be
expected to yield positive results. It is the least expen-
sive and also the most qualitative. Level C tests are com-
prehensive, expensive, and enough data is accumulated for
definitive analyses. A Level B test falls between a Level
90
-------
A and a Level B test in its expense and information
gathered. A Level B evaluation includes all of the infor-
mation obtained by a Level A evaluation and a Level C in-
cludes both Level A and Level B data. Table II is a flow
diagram showing the major considerations for each level
of effort.
3.2 LEVEL A PRECIPITATOR EVALUATION
Points to be considered in a Level A analysis are out-
lined in Table III, Information should be obtained on each
of the items listed and references are shown on the table
of the sections of this report which contain more detailed
discussion.
Attempts should be made to determine if any of the power
supplies show abnormally low readings. Any information that
is available concerning the composition of the flue gas,
the particle size distribution, and opacity should be obtain-
e<^ • I" situ and laboratory measurements of the particulate
resistivity are also included in a Level A test. Any evidence
of excessive rapping losses should be noted.
3.3 LEVEL B PRECIPITATOR EVALUATION
The decisions involved in planning for a Level B pre-
cipitator evaluation are outlined in Table IV. A Level
B evaluation includes all of the information obtained by
a Level A evaluation plus measurements of the mass concentra-
tion at the precipitator inlet and outlet and quantitative
gas analysis. Decisions must be made of the type of par-
ticulate sampling trains and gas sampling equipment to be
used.
91
-------
TABLE II
THREE LEVELS OF EFFORT FOR PRECIPITATOR EVALUATION
LEVEL
C B A
ENGINEERING AND OPERATING DATA
Precipitator Design & Operating Data
Qualitative Gas Composition
Particulate Resistivity (and approximate
size distribution)
Opacity
Any Evidence of Rapping Reentrainment
Two men on site for two days. Two man
weeks for analysis and reporting.
PARTICULATE MASS CONCENTRATION AND PRECIPITATOR
EFFICIENCY
Pre-test Site Survey, Port Installation
Inlet and Outlet Mass Sampling
Quantitative Gas Analysis
Consider Test For Rapping Losses
Five men on site for 5 days. One man month
for analysis and reporting.
PARTICLE SIZING AND PRECIPITATOR FRACTIONAL
EFFICIENCY
Inlet arid Outlet Particle Size Distribution
Ten to fifteen men on site for one to two
weeks.
Six to twenty man months for analysis,
modeling, and report writing.
92
-------
TABLE III
LEVEL A EVALUATION
DESIGN AND OPERATING INFORMATION
A. Precipitator Design and Operating Data
(From observations, interviews, records)
Mechanical Design
Collecting Electrode System
Discharge Electrode System
Electrode Rapping Systems
Electrical - T/R-Systems
Dust Removal Systems
Current Mechanical and Electrical
Condition
Current Collection Efficiency
B. Flue Gas Characterization
Qualitative
(From interviews, or
existing data)
Volumetric Flow rates
Qualitative data on frac-
tional amounts of H20,
Oj, Nj, CO, C02.
Evidence of SO2 , SO3 , H2SO,, ,
or any other substance
affecting ESP performance.
Inlet & Outlet Gas Temperatures
Particulate Resistivity
(Measured)
Laboratory Measurement -
Grab Sample
lH situ Measurement - Resis-
tivity Probe at Inlet Port
Refer to Section 2.1
Refer to Section 2.5
Refer to Section 2.4 and
Appendices E and F.
-------
TABLE IV
LEVEL B EVALUATION
PARTICULATE MASS SAMPLING 8. QUANTITATIVE GAS ANALYSTS
A. Pie-test Site Survey
(Observation, inter-
view, records)
Sampling Locations
Accessibility
Work Platforms
Number of Ports
Diameter of Ports
Laboratory Space
Typical Flue Gas Conditions
Typical Mass Concentrations
and Particle Size
Distributions*
Necessity of Substrate
Conditioning
*Qualitative
Refer to Section 1 and
Appendix A
B. Precipitator Inlet and Outlet Mass Sampling
Sampling Train Type
Flue Depth
Traversing Capability
Velocity/Temperature Traverse
Gas Temperature
Probe Heating
Filter Integrity
Isokinetic Sampling
Sampling Times
Entrained Water
Number of Samples
Process Variations
Rapping Puffs
Hopper Blow-off
Integrated or Time Averaged Sample
Refer to Section 2.2 and
Appendices B and G
C. Quantitative Gas Analysis
ORSAT Measurement of Oz ,
N-, , CO, COZ
EPA Methods Measurements
of S02 , H2SO,, , NO, S03
Alternate Method for Mea-
surement of SO2 & SO3
Other Gases Particular
to Source Under Considera-
tion.
Refer to Section 2.5 and
Appendix G
-------
If the process or emissions are thought to be cyclical,
a decision must be made whether or not to obtain represen-
tative samples during each part of the cycle. Again, per-
tinent sections of this report are referenced in the table.
3.4 LEVEL C PRECIPITATOR EVALUATION
Level A and B tests may not yield sufficient data if
the objectives of the tests include resolution of an opacity
problem or space-charge problem, or if a theoretical model
is to be used to compare with the test results. For each
of these, it is necessary to measure the particle size dis-
tribution. Therefore, in addition to the Level B evaluation
procedure, the Level C evaluation outline in Table V lists
procedures for measuring the inlet and outlet particle size
distribution and the precipitator fractional efficiency.
Some information on rapping losses can be obtained with op-
tical particle counters, as described in Section B.4, or
by sampling with impactors during and between rapping cycles,
95
-------
TABLE V
LEVEL C EVALUATION
PARTICLE SIZE DISTRIBUTION MEASUREMENTS
A. Inlet B. Outlet
Process Variations Process Variations
Rapping Puffs
Hopper Blow-off
Size Range
0.3-10.0 urn
Cascade Impactors
High Loading Low Loading
Brink Andersen
MRI
Sierra
U. of W.
Isokinetic Sampling
Jet Velocitv Limits
Nozzle Selection
Precutter Selection
Loading Limits
Substrate Integrity
Number of Samples
Traversing
Vert ical/Horizontal
Condensible Vapors
Heating/Entrained Water
Extractive Sampling
Sampling Times
Measurement Methods
0.3-10 |jm
Series Cyclones
Accuracy/Resoluti on
Sampling Times
Back Up Filters
Filter Integrity
Filter Loading
Constant Flow rate
Entrained Water
Traver sing
Isokinetic Sampling
Extractive Sampling
Probe Losses
Refer to Section 2.3.2
0.3-10 pm 0.01-0.3 ym
Optical Diffusional/EAA
Extractive Sampling
Sample Conditioning
Dilation of. Sample
Gas Composition
Condensible Vapors
Heated Lines
Probe Losses
Real Time Monitoring
Duct Pressures
Refer to Sections 2.3.3
and Appendix D
Refer to Section 2.3.2 and Appendix C
-------
REFERENCES
1. White, H. J. Industrial Electrostatic Precipitation.
Addison-Wesley Publishing Company, Inc., Reading,
Massachusetts, 1963. 376 pp.
2. "Standards of Performance for New Stationary Sources."
Federal Register. Vol. 36, No. 247, December 23, 1971.
3. Cochet, R. and J. Trillat. Charging of Submicron Parti-
cles in Electrically Ionized Fields; Measurement of the
Rate of Precipitation in a Uniform Electric Field. Compt.
Rend. Acad. Sci. 250: 2164-2166, 1960.
4. Ranz, W. E., Wong, J. B. Impaction of Dust and Smoke Par-
ticles. Industrial Engineering Chemistry. 4_4_(6) : 137-
1381, 1952.
5. Davies, C. N., and M. Aylward. Proc. Phys. Soc., 64: 889, 1951,
6. Marple, V. A. Doctoral Thesis, University of Minnesota,
Minneapolis, MN, 1970.
7. Rao, A. K. Doctoral Thesis, University of Minnesota, Minnea-
polis, MN, 1975.
8. Gushing, K. M., G. Lacey, J. D. McCain, and W. B. Smith.
Particulate Sizing Techniques for Control Device Evaluation
Cascade Impactor Calibrations. EPA 600/2-76-280, U.S.
Environmental Protection Agency, Research Triangle Park,
North Carolina, October, 1976, 94 pp, NTIS-PB-245184 ($5.75).
9. Chang, Hsing-Chi. A Parallel Multicyclone Size-Selective
Particulate Sampling Train. Amer. Ind. Hyg. Assoc. J.,
September 1974.
10. Hamersma, J. W., S. L. Reynolds, and R. F. Maddalone.
Procedures Manual for Level 1 Environmental Assessment.
EPA-600/2-76-160a, EPA, Research Triangle Park, North
Carolina, NTIS-PB-
11. Whitby, K. T. and B. Y. H. Liu. J. Colloid Interface Science,
2J>: 537, 1967.
12. Sinclair, D. Amer. Ind. Hygiene Assoc. J. , 3_6_(1): 39, 1975.
13. Breslin, A. J., S. F. Guggenheim, and A. C. George. Staub
(English Translation), _3_M8):l-5, 1971.
97
-------
14. Hewitt, G. W. The Charging of Small Particles for Elec-
trostatic Precipitation. AIEE Winter General Meeting,
New York. Paper No. 73-283, 1957.
15. Liu, B. Y. H., K. T. Whitby, and D. Y. H. Pui. A Portable
Electrical Aerosol Analyzer for Size Distribution Measure-
ments of Submicron Particles. 66th Annual Meeting of the
Air Pollution Control Association. Paper No. 73-283, 1973.
16. Nichols, Grady, B., "Techniques for Measuring Fly Ash Re-
sistivity", EPA-650/2-74-079 (August 1974).
17. "Determining the Properties of Fine Particulate Matter"
PTC 28, The American Society of Mechanical Engineers,
United Engineering Center, 345 East 47th Street, New York,
New York 10017.
18. "Information Required for Selection of Electrostatic and
Combination Fly Ash Collectors: Methods of Analysis for
Chemical, Physical and Electrical Properties of Fly Ash - In-
formation Report No. 2", APCA Journal 1_5 (6) 256-260, 1965.
19. Maddalone, R. and N. Garner. Process Measurement Procedures •
Sulfuric Acid Emissions. TRW Systems Group Document 28055-60-
04-RO-OO, February 1977.
98
-------
APPENDIX A
AEROSOL FUNDAMENTALS, NOMENCLATURE, AND DEFINITIONS
A.I GENERAL DISCUSSION,
A. 2 PARTICLE SIZE DISTRIBUTIONS 101
A.2,1 Cumulative and Differential Graphs
A, 3 NOMENCLATURE AND DEFINITIONS 110
A.3.1 Definitions of Particle Diameter
A.3.2 Mean Free Path of Gas
A.3.3 Slip Correction Factor
A.3.4 Viscosity of Gas
A.3.5 Particle-Gas Interactions
A.3.6 Cascade Irapactor Technology
99
-------
APPENDIX A
AEROSOL FUNDAMENTALS, NOMENCLATURE, AND DEFINITIONS
A.I GENERAL DISCUSSION
The particulate matter suspended in industrial gas streams
may be in the form of nearly perfect spheres, regular crystal-
line forms other than spheres, irregular or random shapes,
or as agglomerates made up from combinations of these. It
is possible to discuss particle size in terms of the volume,
surface area, projected area, projected perimeter, linear
dimensions, light scattering properties, or in terms of drag
forces in a liquid or gas (mobility). Particle sizing work
is frequently done on a statistical basis where large numbers
of particles, rather than individuals, are sampled. For this
reason the particles are normally assumed to be spherical.
This convention also makes transformation from one basis to
another more convenient.
Experimental measurements of particle size normally cannot
be made with a single instrument if the size range of interest
extends over much more than a decimal order of magnitude.
Presentations of size distributions covering broad ranges of
sizes then must include data points which may have been obtained
using different physical mechanisms. Normally the data points
are converted by calculation to the same basis and put into
tabular form or fitted with a histogram or smooth curve to
represent the particle size distribution. Frequently used
bases for particle size distributions are the relative number,
volume, surface area, or mass of particles within a size range.
The size range might be specified in terms of aerodynamic,
100
-------
Stokes, or equivalent PSL diameter. There is no standard equa-
tion for statistical distributions which can be universally
applied to describe the results given by experimental particle
size measurements. However, the log-normal distribution func-
tion has been found to be a fair approximation for some sources
of particulate and has several features which make it conven-
ient to use. For industrial sources the best procedure is
to plot the experimental points in a convenient format and
to examine the distribution in different size ranges separately,
rather than trying to characterize the entire distribution
by two or three parameters. The ready availability of inex-
pensive programmable calculators which can be used to convert
from one basis to another compensates greatly for the lack
of an analytical expression for the size distribution.
A.2 PARTICLE SIZE DISTRIBUTIONS
Figure Al shows plots of generalized unimodal particle
size distributions which will be used to graphically illustrate
the terms which are commonly used to characterize an aerosol.
Occasionally size distribution plots exhibit more than one
peak. A size distribution with two peaks would be called bimodal
Such distributions can frequently be shown to be equivalent
to the sum of two or more distributions of the types shown
in Figure Al. If a distribution is symmetric or bell shaped
when plotted along a linear abscissa, it is called a "normal"
distribution (Figure Ale). A distribution that is symmetric
or bell shaped when plotted on a logarithmic abscissa is call-
ed "log-normal" (Figure Aid),
Interpretation of the frequency or relative frequency
shown as f in Figure Al is very subtle. One is tempted to
interpret this as the amount of particulate of a given size.
This interpretation is erroneous however and would require
101
-------
a. Distribution Skewed Left
b. Distribution Skewed Right
LOG ag
c. Normal Distribution
LOG D
d. Log Normal Distribution
3630-081
Figure A1. Examples of frequency or particle size distributions. D is
the particle diameter.
102
-------
that an infinite number of. particles be present. The most
useful convention is to define f in such a way that the area
bounded by the curve (f) and vertical lines intersecting the
abscissa at any two diameters is equal to the amount of par-
ticulate in the size range indicated by the diameters selected.
f then is equal to the relative amount of particulate in a
narrow size range about a given diameter.
The median divides the area under the frequency curve
in half. For example, the mass median diameter (MMD) of a par-
ticle size distribution is the size at which 50% of the mass
consists of particles of larger diameter, and 50% of the mass
consists of particles having smaller diameters. Similar defini-
tions apply for the number median diameter (NMD) and the sur-
face median diameter (SMD).
The term "mean" is used to denote the arithmetic mean
of the distribution. In a particle size distribution the mass
mean diameter is the diameter of a particle which has the ave-
rage mass for the entire particle distribution. Again, similar
definitions hold for the surface and number mean diameters.
The mode represents the diameter which occurs most commonly
in a particle size distribution. The mode is seldom used
as a descriptive term in aerosol physics.
The geometric mean diameter is the diameter of a particle
which has the logarithmic mean for the size distribution.
This can be expressed mathematically as:
log DI + log Dz + • • • • log DN
log Dg = s (Ala)
.03
-------
or as
1/N
D = (0,0203 . . . DM J (Alb)
The standard deviation (a) and relative standard deviation
(ct) are measures of the dispersion (spread, or polydispersity)
of a set of numbers. The relative standard deviation is the
standard deviation of a distribution divided by the mean, where
o and the mean are calculated on the same basis; i.e., number,
mass, or surface area. A monodisperse aerosol has a standard
deviation and relative standard deviation of zero. For many
purposes the standard deviation is preferred because it has
the same dimensions (units) as the set of interest. In the
case of a normal distribution, 68.27% of the events fall within
one standard deviation of the mean, 95.45% within two standard
deviations, and 99.73% within three standard deviations.
A.2.1 Cumulative and Differential Graphs
Field measurements of particle size usually yield a set
of discrete data points which must be manipulated or trans-
formed to some extent before interpretation. The resultant
particle size distribution may be shown as tables, histograms,
or graphs. Graphical presentations are the conventional and
most convenient format and these can be of several forms.
Cumulative size distributions - Cumulative mass size dis-
tributions are formed by summing all the mass containing par-
ticles less than a certain diameter and plotting this mass
versus the diameter. The ordinate is specifically equal to
104
-------
M .
t=l
where M, is the amount of mass contained in the size inter-
val between D. and D. ,.
The abscissa would be equal to D.. Cumulative plots can be
made for surface area and number of particles per unit volume
in the same manner. Examples of cumulative mass and number
graphs are shown in Figures A2b and A2a, respectively, for
the effluent from a coal fired power boiler. Although cumula-
tive plots obscure some information, the median diameter and
total mass per unit volume can be obtained readily from the
curve. Because both the ordinate and abscissa extend over
several orders of magnitude, logarithmic axes are normally
used for both.
A second form of cumulative plot which is frequently used
is the cumulative percent of mass, number, or surface area
contained in particles having diameter smaller than a given
size. In this case the ordinate would be, on a mass basis:
1
Y.
Cumulative percent of mass less than size = t=l
I Mt
N
L M
t=l t
x 100%. (A21
The abscissa would be log D.. Special log-probability paper
is used for these graphs, and for log-normal distributions
the data set would lie along a straight line. For such dis-
tributions the median diameter and geometric standard deviation
can be easily obtained graphically. Figures A3a and A3b show
cumulative percent graphs for the size distribution shown in
Figure A2a and a log-normal size distribution.
Differential size distributions - Differential particle
size distribution curves are obtained from cumulative plots
105
-------
c
ai
Q Q
t$
i H
1012
u.£
O O
1010
0,01 0.1 1.0
PARTICLE DIAMETER,
a. Cumulative No. Graph
10
0.01 0,1 1.0
PARTICLE DIAMETER, fj
b. Cumulative Mass Graph
10
0.01 0.1 1.0
PARTICLE DIAMETER, /im
c. Differential No. Graph
10
10*
~^ 103
E
102
10
10"'
,-1
0.01 0.1 1.0
PARTICLE DIAMETER, ,
d. Differential Mass Graph
10
3630-089
Figure A2. A single particle size distribution presented in four ways.
The measurements were made in the effluent from a coal-
fired power boiler.
•"06
-------
a. Cumulative Percent Graph
b. Cumulative Percent Graph
(Log Normal Distribution)
PARTICLE DIAMETER, j
Figure A3. Size distributions plotted on tog probability paper.
3630-087
107
-------
by taking the average slope over a small size range as the
ordinate and the geometric mean diameter of the range as the
abscissa. If the cumulative plot were made on logarithmic
paper the frequency (slope) would be, taking finite differ-
ences :
A(log D) log D. - log D.
and the abscissa would be D = /D.D._, where the size range
of interest is bounded by D. and D. , . M. and M._, correspond
D j~ •*- D H -1-
to the cumulative masses below these sizes. Differential number
and surface area distributions can be obtained from cumulative
graphs in precisely this same way. Differential graphs show
visually the size range where the particles are concentrated
with respect to the parameter of interest. The area under
the curve in any size range is equal to the amount of mass
(number, or surface area) consisting of particles in that range,
and the total area under the curve corresponds to the entire
mass (number, or surface area) of particulate in a unit volume.
Again, because of the extent in particle size and the emphasis
on the fine particle fraction, these plots are normally made
on logarithmic scales. Figures A2c and A2d are examples of
differential graphs of particle size distributions.
Log-normal size distributions - The formation of aerosols
by different means frequently result in particle size distribu-
tions which obey the log-normal law. For log-normal particle
size distributions the geometric mean and median diameters
coincide.
The normal distribution law is, on a mass basis:
108
-------
dM
dD
-exp
2 I
2a
(A4;
The log-normal distribution law is derived from this equation
by the transformation D ->• log D
f - dM _ = 1 _ exD
r " d(log D) log a /2iT p
/log D-log
\
% }
(A5)
where a , the geometric standard deviation, is obtained by
using the transformation D -* log D in equation A4 . This dis-
tribution is symmetric when plotted along a logarithmic abscissa
and has the feature that 68.3% of the distribution lies within
one geometric standard deviation of the geometric mean on such
a plot. , . , .
Mathematically, this implies that log a = log
-log D or log D -log D,_
which 84.14% of the distribution is found
where DgA , , is the diameter below
simplified to yield:
D
a -
g
84
or
16
etc. This can be
(A6)
(A7)
(A8)
When plotted on log-probability paper, the log-normal
distribution is a straight line on any basis and is determined
completely by the knowledge of D and a • This is illustrated
in Figure A3b. Another important feature is the relatively
simple relationships among log-normal distributions of differ-
I£ Dgm' Dgs' Dgvs' and % are the Seometric mean
ent bases.
diameter of the mass, surface area, volume-surface, and number
109
-------
distribution, then:
log Dgs = log Dgm - 4.6 Iog2ag, (A9)
log D = log D - 1.151 loq 2o , and (AID)
^ gvs y gro 9
log DgN = log Dgm ~ 6'9 Io92°g-
The geometric standard deviation remains the same for all bases.
More examples of particle size distribution graphs are
given in the data reduction sections of Appendices C and D.
The following section in this Appendix lists useful definitions,
equations, and nomenclature for aerosol sampling.
A.3 NOMENCLATURE AND DEFINITIONS
A.3.1 Definition of Particle Diameter
Aerodynamic diameter, D. - The aerodynamic diameter of
a particle is the diameter of a sphere of unit density which
has the same settling velocity in the gas as the particle of
interest.
Aerodynamic impact ion diameter, Dfl, - The DftT of a particle
is an indicator of the way that a particle behaves in an iner-
tial impactor or in a control device where inertial impaction
is the primary mechanism for collection. If the particle Stokes
diameter is known, D , the DAI is equal to:
DAI = Ds/pC ' (A12)
110
-------
where p is the particle density, gm/cm3, and
C is the slip correction factor.
Stokes diameter, D - If the density of a particle is
o
known, the Stokes diameter may be used to describe particle
size. This is the diameter of a sphere having the same density
which behaves aerodynamically as the particle of interest.
For spherical particles, the Stokes number is equal to the
actual dimensions of the particle.
An average density for the particles can be obtained from
volume-weight data using a helium pycnometer if large enough
samples are available. The validity of size information based
on an average density depends upon the uniformity of the density
from particle to particle, particularly with respect to size.
Visual inspection of some size-classified samples from flue
gases sometimes shows a variation in color with size which
would seem to indicate compositional inhomogeneities.
Equivalent polystyrene latex (PSL) diameter - The intensity
of light scattered by a particle at any given angle is dependent
upon the particle size, shape, and index of refraction. It
is impractical to measure each of these parameters and the
theory for irregularly shaped particles is not well developed.
Sizes based on light scattering by single particles are there-
fore usually estimated by comparison of the intensity of scattered
light from the particle with the intensities due to a series
of calibration spheres of very precisely known size. Most
commonly these are PSL spheres.* Spinning disc and vibrating
orifice aerosol generators can be used to generate monodisperse
*Available from The Dow Chemical Company, P. 0. Box 68511,
Indianapolis, Indiana 46268.
Ill
-------
calibration aerosols of different physical properties. Because
most manufacturers of optical particle sizing instruments use
PSL spheres to calibrate their instruments, it is convenient
to define an equivalent PSL diameter as the diameter of a PSL
sphere which gives the same response with a particular optical
instrument as the particle of interest.
Equivalent volume diamet er - Certain instruments, such
as the Coulter Counter, have, as the measured size parameter,
the volumes of the individual particles. Size distributions
from such techniques are given in terms of spheres having the
same volume as the particles of interest.
A.3.2 Mean Free Path ofGas
The mean free path of a gas, which is the average distance
that molecules travel between collisions, is an important para-
meter in determining the aerodynamic behavior of particles.
For practical purposes, the mean free path is given with suf-
ficient accuracy by the following equation.
, _ ^,M 18. 3x10 j. i i n -, o,
x _ _. i j (A1J)
1.01xlO°P V J ^
>Ir\
1M /
where y is the viscosity of the gas, poise,
P is the pressure of the gas, atm,
T is the temperature, °Kelvin, and
MM is the mean molecular weight.
A.3.3 Slip CorrectionFactor
Stokes law can be applied to submicron particles if a
correction factor, C, is used.
112
-------
P 1.23+0.41 exp( °-44D) , (A14;
"
where X is the mean free path of the gas, ym, and
D is the particle diameter, ym.
The constants in equation A14 were determined empirically
for air at standard temperature and pressure, and are thus
only approximate for stack conditions. If the exponential
term is neglected, equation 14 is referred to as the Cunning-
ham correction factor.
A.3.4 Viscosity of Gas
A parameter which appears in many equations describing
flow fields, drag forces, and shear forces in liquids and gases
is the viscosity.
In order to find the viscosity of the flue gas, v, the
viscosity of the pure gas components of the flue gas must first
be found. Viscosity is a function of temperature, and the
temperature difference in different flue gases can be quite
significant. The following equations (derived from curves
fitted to viscosity data from the Handbook of Chemistry and
Physics, Chemical Rubber Company Publisher, 54 Edition, 1973-
1974, pp. F52-55), are used to find the viscosities of C02
(Mi), CO(M2), N2 {y3)f CMuO and H20(ys).
Ui = 138.494 4 0.499T -0.267 x 1CT3T2 + 0.972 x 10"7T3
y2 = 165.763 + 0.442T - 0.213 x 10~3T2
y3 = 167.086 + 0.417T - 0.139 x 10~3T2
UU = 190.187 + 0.558T - 0.336 x 10 ~3T2 + 0.139 x 1Q-5 T 3
Ms = 87.800 + 0.374T + 0.238 x 10~"4TZ
where T is the temperature of the flue gas in degrees Celsius.
113
-------
The units of p are 10~6 g/cm-sec. Next, these values of Ui
through us are used in a general viscosity equation for a
mixture of any number of components (See "A Viscosity Equa-
tion for Gas Mixtures" by C. R. Wilke, Journal of Chemical
Physics, Volume 8, Number 4, April 1950, page 517) used to
find the viscosity of the flue gas:
n
A is;
where . . is given by the equation:
and
[l + (yi/p.)'< (M./M. )**\
+ (M./M..)J h
(A16)
M
X
V
= molecular weight of a component in the mixture,
= mole fraction of a component in the mixture,
= viscosity, g/cm-sec; ui, U2, etc., refer to the
pure components at the temperature and pressure of
mixture, v is the viscosity of the mixture, and
= dimensionless constant defined above.
A.3.5 Particle-Gas Interactions
Particle relaxation time - For the purposes of this doc-
ument, the particle relaxation time, T, may be defined as the
time required for a.particle to accelerate from some initial
velocity to the velocity of the carrier gas.
p DZC
18
(A17)
114
-------
where P is the particle density, gm/cm ,
D is 'the particle Stokes diameter, cm,
C is the slip correction factor, and
lj is the gas viscosity, poise.
Part i cle stopping d istance - The particle stopping distance,
Jl, is the distance travelled by a particle as it decelerates-from
the initial velocity of the gas to zero velocity.
i = TV (A18)
where T is the relaxation time (sec), and
V is the initial velocity of the particle (cm/sec).
Stokes number - The Stokes number, 4», is the ratio of
the particle stopping distance and some characteristic dimen-
sion of the sampling system. For example, if the stopping
distance for particles of a given diameter is much smaller
than the diameter of a sampling nozzle, (vb « 1) the particles
will be sampled accurately in spite of flow disturbances due
to the nozzle design or sampling velocity. If the particle
stopping distance is comparable in magnitude to the nozzle
diameter; however, the particles may cross flow streamlines
and either enter or miss the nozzle in quantities which are
not proportional to the sample flowrate. Thus, for llj on the
order of 0.1 or greater, isokinetic sampling is required.
Particle mo b i1i ty - The ratio of the velocity of a par-
ticle to the force causing steady motion is called the mobi-
lity, b.
(A 19)
115
-------
where p is the gas viscosity, poise,
D is the particle diameter, cm, and
C is the slip correction factor.
A.3.6 Cascade Impactor Terminology
Blanj< — A blank usually refers to a controlled test run
in which the particles are removed by a prefilter. If the mea-
sured stage weight is found to change significantly and consis-
tently, the normal runs should be corrected for background.
Bounce — Bounce in this document refers to inadequate
retention of particles which strike the impaction surface.
If the particle does not adhere, it is said to bounce.
Condensation — Condensation in an impactor refers to
the coalescence of vapors either into liquid particulate in
the gas stream or on the impactor walls.
Control — A control run is a technique which is used
to confirm that variables are isolated. The control is made
up to be as similar as possible to an actual run, but it is
not run through the test situation. The control is then
examined as would an actual test run. If the experimental
variable changed significantly, the experiment is not properly
set up.
Cut-point — The cut-point of an impactor stage is the
particle diameter for which all particles of equal or greater
diameter are captured and all particles with smaller diameters
are not captured. No real impactor actually has a sharp cut-
point, but the D5o of a stage is often called its cut-point.
116
-------
DS_^ — the Db o of an impactor staqe is the particle Dia-
meter at which the device is 50 percent efficient. Fifty per-
cent of the particles of that diameter are captured and 50%
are not.
The impactor cut-point, or D5 is given by:
/18 i!> y D.
' c - „ 3 (A20)
P J
where v = Stokes inertial impaction parameter, determined
by calibration, dimensionless,
y = gas viscosity, poise,
D. = impactor jet diameter (for slot impactors, the slot
width, cm),
V. = gas velocity through impactor jet, cm/sec,
C = slip correction factor, dimensionless,
p = particle density, g/cm3.
D50(AI), aerodynamic impaction diameter, is found by setting
C and p = 1.0.
D^ „ (A), the aerodynamic diameter, is found by setting p = 1.0,
and
D50(S), Stokes diameter, is found by setting a = the actual
particle density.
Grease -- In impactor terminology, grease is a substance
which is placed on an impactor stage or substrate to serve
as an adhesive.
Isokinetic sampling — This is sampling with the bulk
fluid velocity through the impactor nozzle equal to the velo-
city in the duct. This is necessary to prevent sample bias.
117
-------
Preconditioning — Unwanted weight changes of impactor glass
fiber collection substrates may be reduced by placing a large
number of substrates inside the duct to be sampled, and pumping
filtered flue gas through them for several hours. Such a pro-
cedure is referred to as "preconditioning" the substrates.
Precutter or precollector — A collection device, often
a cyclone, which is put ahead of the impactor in order to re-
duce the first stage loading. This is necessary in some streams
because the high loading of large particulate would overload
the first stage before an acceptable sample had been gathered
on the last stages.
Re-entrainment — Re-entrainment in an impactor is the
phenomenon of particles which impacted on a given stage being
picked up by the gas stream and moving downstream to another
stage.
Stage -- A stage of an impactor is usually considered
to be the accelerating jet (or plate containing multiple jets)
and the surface on which the accelerated particles impact.
Substrate — The removable, often disposable, surface
on which impacted particles are collected. Substrates are
characteristically light and can be weighed on a microbalance.
Wall Losses - Wall losses are that portion of the parti-
cles in the gas stream which impact with and adhere to sur-
faces in the impactor other than the substrates. They should
be collected if possible and assigned to the proper stage catch.
118
-------
APPENDIX B
PARTICULATE MASS CONCENTRATION MEASUREMENTS
Page
B.I GENERAL SAMPLING EQUIPMENT .\20
B.I.I Temperature Measurement
B.I.2 Pressure Measurement
B.I.3 Velocity Measurement
B.I.4 Nozzles
B.I.5 Probes
B.I.6 Gaseous Sample Collectors
B.I.7 Particulate Sample Collector
B.I.8 Sample Flow Rate Meters
B.I.9 Total Sample Volume
B.I.10 Gas Conditioning
B.I.11 Pumps
B.I.12 Flow Control
B.2 ELECTROSTATIC PRECIPITATOR SAMPLING-PRELIMINARY
PROCEDURES 142
B.2.1 Introduction
B.2.2 Physical Sampling Location Characteristics
B.2.3 Velocity Determination
B.2.4 Stack Moisture Content
B.2.5 Molecular Weight of the Stack Gas
B.3 ELECTROSTATIC PRECIPITATOR SAMPLING FOR PARTICULATE
AND GASES 154
B.3.1 Isokinetic Sampling
B.3.2 Sampling for Effluent Gases
B.3.3 Nozzle Selection and Nomograph Setting
B.3.4 EPA Reference Method 5 Procedure
B.3.5 General Sampling Procedures
B.4 MASS CONCENTRATION DATA REDUCTION 171
B.4.1 Volume of Gas Sampled
B.4.2 Volume of H20 Vapor in Stack Gas
B.4.3 Moisture Content of Stack Gas
B.4.4 Molecular Weight of Stack Gas
B.4.5 Excess Combustion Air
B.4.6 Particulate Emissions Concentration
B.4.7 Stack Gas Volumetric Flow Rate
B.4.8 Average Isokinetic Ratio
B.4.9 Mass Collection Efficiency Calculation
.U9
-------
APPENDIX B
PARTICULATE MASS CONCENTRATION MEASUREMENTS
This appendix contains the details for conducting par-
ticulate mass concentration measurements using the EPA Ref-
erence Method 5 procedure (See Appendix G). The information
is presented in four sections. Section Bl describes the
equipment used in conducting these tests. Section B2 is
concerned with preliminary procedures prior to sampling.
Section B3 details the actual mass sampling procedures.
Section B4 deals with data analysis after testing has been
concluded. The sampling system described is depicted in Fig-
ure 13 of Section 2.2.2.
B.I GENERAL SAMPLING EQUIPMENT
B..1.1 Temperature Measurement
Several temperature measurements are required in con-
ducting a test for particulate mass loading, including the
temperature of the stack gasr the particulate filter, and
the cooled sample stream. The relative errors encountered
in temperature measurements are usually small since absolute
temperatures are used in all gas law calculations. In source
testing, dial thermometers and thermocouples are usually
used for making temperature measurements.
Two common scales are used in temperature measurement,
the Celcius scale and the absolute Kelvin scale. Conversions
for these are shown below for °C to °K and °F to °C.
°K = °C + 273°
°C = 5/9 (°F - 32)
120
-------
Mercury bulb thermometers - The mercury bulb thermometer
operates by measurement of mercury expansion with temperature
increase. This expansion is linear over the range of the
temperature scale. Glass mercury thermometers break easily
and this is a risk in source sampling.
Dial thermomet e r s - Two types of dial thermometers are
available for use in sampling. One is bimetallic and the
other is a gas bulb thermometer. The bimetallic thermometer
contains a strip of two different metals bonded together.
Because of the different thermal expansion coefficients of
the two metals, the bonded strip will deform with tempera-
ture, and depending on the configuration of the strip, this
deformation will be transferred to a dial movement which
contains a temperature scale.
Gas bulb thermometers rely on the expansion of an inert
gas with temperature. This expansion is sensed as a change
in pressure. The dial-temperature scale is actually a pres-
sure scale. Gas bulb thermometers are used for lower tempera-
ture ranges.
Thermocouples - Thermocouples are the most popular device
for measuring high temperatures. These consist of two dis-
similar wires welded together at one junction (See Figure
Bl). These two wires are joined to a third wire held at
a reference temperature. The difference between the tempera-
ture in question and the reference junction temperature causes
an electromotive force in the system which can be sensed by
^a potentiometer. Several metal pair types are available.
Generally Chromel/Alumel will be the best choice due to
resistance to oxidation. This pair is useful in the range
121
-------
INDICATOR
METAL A
METAL B
I 1
METAL C
LEADS
(LOW RESISTANCE AND
AS SHORT AS PRACTICAL!
METAL C
REFERENCE
JUNCTION
Figure B1, Thermocouple /unction.
3630-067
-------
from -184°C to 1260°C. Other common pairs are Copper/Con-
stantan (-184°C to 350°C) , Iron/Constantan (-158°C to 1010°C} .
and Platinum/Platinum 10% Rh (0°C to 1538°C).
B.I.2 Pressure Measurement
Pressure is defined as a force per unit area. Most
pressure measurements are made with local atmospheric.pres-
sure as reference. The pressure above atmospheric is con-
sidered positive, and that below negative. The absolute
pressure at a point is the atmospheric plus the pressure
differential.
An easy way to measure a low pressure is to balance a
column of liquid against the pressure. The magnitude of
the pressure can be calculated based on the measured height
of the liquid column. Devices which do this are called
manometers. In source sampling, manometers are often used
for the determination of the stack gas velocity and the
sample train flow rate. For small pressure differentials,
the manometer is often inclined to increase the sensitivity.
The inclined manometer is used to measure the stack velocity
pressure and sample stream orifice pressure differential.
It is advantageous to use manometers which have some means
of protection against accidental blow out.
Mechanical pressure gauges are also available to measure
low differential pressures commonly encountered as velocity
pressure and orifice meter heads in sampling systems. The
Magnehelic gauge manufactured by F. W. Dwyer, Mfg. in Michigan
City, Indiana is an example of such an instrument. Inclined
manometers, however, are generally more reliable and easier
to use. They are also easier to repair.
123
-------
B.I.3 Velocity Measurement
The measurement of velocity in a duct utilizes one of
the simplest devices in most sampling systems, the pitot tube
These devices are necessary because it is impossible to deter-
mine the total volumetric flow through large ducts. Only
by measuring the velocity at many points and knowing the
area of the duct can an accurate determination of duct volu-
metric flow be made. The pitot tube will not directly mea-
sure the average duct velocity but measure only the instan-
taneous velocity at the point at which it is located.
Several configurations are possible for pitot tubes.
One, the Prandtl type is shown in Figure B2. The static
pressure is measured at point W. The velocity pressure is
measured at point P. The velocity, V , then is given by
V = C
s p
ps
where V is the gas velocity, cm/sec,
S
P is the measured velocity pressure, mm Hg,
P is the measured static pressure, mm Hg,
W
C is the pitot coefficient, dimensionless,
p is the density of the stack gas, gm/cin3 .
5
This pitot tube shown in this figure is usually called a
standard pitot tube; tl
is approximately 0.99.
standard pitot tube; the C value for this configuration
124
-------
w
w^to> P
pp-pw
J:W
MANOMETER
3530-059
Figure B2. Standard pi tot tube.
.25
-------
One configuration which does not resemble the standard
pitot is the reversed or Stausscheibe (S-type) pitot tube.
The S-type pitot tube is used primarily for stack testing
because of one major advantage: it will not rapidly clog
in gases with heavy dust concentrations. An example of one
of these is shown in Figure B3. The S-type does not give
the same velocity pressure as the standard pitot tube. The
observed delta P is larger for a given velocity because the
rear part of the tube faces downstream. This P is a wake
S
pressure which is lower than the static pressure. When used
with a water manometer, the S-type pitot tube equation be-
comes
h
I 1<-"r \
= 422.67 C
/Vp \
\PsMs /
where V is the gas velocity, cm/sec,
s
AP is the velocity head (P -P ), mm Hg
P s
T is the stack gas temperature, °K,
j
P is the absolute stack gas pressure, mm Hg,
M is the molecular weight of the stack gas, gm/gm-mole,
S
and
C is the pitot tube correction factor.
For the limits of 0.025 to 25.4 cm of water velocity pressure,
C for a standard (Prandtl) pitot tube usually takes on values
of 0.98 to 1.00. The C of an S-type pitot tube usually
is between 0.83 and 0.87. Each must be calibrated before
a test, preferably in a gas stream in which the gas properties
and velocity are similar to those of the test conditions.
The calibration of a pitot tube requires a gas stream
of constant and known velocity. Thus, a wind tunnel facility
126
-------
GAS FLOW
MANOMETER
3630-068
Figure B3. S-type pi tot tube.
127
-------
should be available. However, the S-type pitot tube can
be calibrated against the standard pitot tube. This is the
procedure designated in EPA Reference Method 2.
The following equation applies in this case:
c
ptest pstd APtest
where C = Coefficient of the S-type pitot tube,
ptest
C = Coefficient of the standard pitot tube,
pstd
A = Velocity pressure measured by the S-type
test
pitot tube, and
A = Velocity pressure measured by the standard
pstd
pitot tube.
To calibrate the S-type pitot tube, the^ velocity pres-
sure is measured at the same point with both the S-type and
standard pitot tubes. Both pitot tubes must be properly
aligned in the flow field. The appropriate values are
inserted in the above equation and the coefficient for the
S-type pitot tube is calculated. If C is not known,
then a value of 0.99 should be used.
The coefficients for the S-type pitot tube should be
determined first with one leg, then with the other leg pointed
downstream. If the computed coefficients differ by more
than 0.01, the pitot tube should not be used without proper
labeling.
The determination of the average stack gas velocity,
V , is one of the greatest sources of error in stack sampling.
5
Therefore, it is recommended that the pitot tube be recali-
brated on a regular basis.
128
-------
At extremely low or high velocities the pitot method
is inaccurate and unreliable. There are several other mechanical
and electronic methods which are available, hot wire ane-
mometers, rotating vane anemometers and certain fluidic
devices.
B.1.4 Nozzles
The nozzle is considered the initial sampling sy^t^m
boundary. It removes a portion of the effluent from the
duct and delivers it to the sampling probe. The nozzle has
several restrictions in its use:
1. It should not disturb the duct gas stream flow.
2. It should not alter the particulate being sampled.
3. It should not add to the sample being collected.
4. It should be of a size allowing easy access.
For particulate sampling, the nozzle should disturb
the gas flow as little as possible or the sample will not
be representative. Any nozzle will disturb the flovv, b"t
a thin wall, sharp edged nozzle disturbs it the least. In
the case of particulates, any bends in the nozzle will cause
impingement of larger particles. The nozzle must then be
cleaned carefully and any material found within it must be
added to the total collected particulate.
B.1.5 Probes
The probe is the sampling interface between the gas
stream in the duct and the external sampling train (See
Figure B4). It is exposed internally and externally to
the flue gas at the nozzle end and ambient air at the exit
end. The probe should not alter the sample in any way.
It must be able to support itself. It must be easy to clean
and it must not add to the sample. Ideally the sample should
129
-------
GLASS
BALL
JO1NTV
GLASS
PROBE
HIGH
TEMPERATURE
TAPE
FRO.NT
FERRULE
RUBBER
STOPPER'
POWEFT
CORD
HEATING
WIRE
WELD
UNION
REAR
FERRULE
NOZZLE
REAR
FERRULE
3S30-O70
Figure B4. Probe and nozzle assembly.
-------
be delivered to the sample train at the stack temperature.
In most instances this requires that the probe be heated
to maintain the sample at stack conditions and prevent con-
densation. Structurally the probe must support itself, the
nozzle, the pitot tube, and sometimes a thermocouple. This
strength is obtained by use of a metal sheath around the
probe. The requirement of probe cleaning conflicts with
the structural needs. The surface of a glass tube is much
more easily cleaned than a metal tube, but it is more fra-
gile. In most cases glass probes over 2 meters in length
are impractical. If a metal tube is used, some material
may become trapped in the rough surface of the probe.
In the case of extremely high temperatures, the only
practical choice is to use a water-cooled, high quality,
stainless steel probe.
Glass has another advantage which should be considered.
It is for all practical purposes chemically inert. This
is not true of stainless steel especially if there are acid
gases present in the gas stream which is being sampled.
B. 1.6 Gaseous Sample Collectors
There are four main types of gas sample collection de-
vices. One type is the cold trap which condenses vapors in
the sample flow stream. Another type is that which contains
a solid adsorbent. This removes the gas from the stream
by surface adsorption. The third type is the grab sample
container. The fourth uses the principle of gas absorp-
tion by a liquid.
Gas absorbers are generally called impingers or bubblers,
Their efficiency depends on the diffusivity of the gases,
131
-------
the retention time in the devices, the bubble size, and the
gas solubility. There are four general types of impingers;
the midget impinger, the fritted glass bubbler, the modified
impinger tip bubbler, and the most commonly used type, the
Greenburg-Smith impinger.
B.I.7 Particulate Sample Collectors
Filtration is the basic method for particulate collec-
tion. There are three major types of filters available to-
day: the flat glass fiber filter, the ceramic Alundum
filter, and the glass fiber bag.
Alundum thimbles are subject to variations in their
particle retention efficiency as they are used because
of changes in their porosity as a particulate cake forms
on them. Therefore new filters should not be used when
testing relatively clean gas streams. In addition, when
the thimble is used in relatively clean gas streams, only
small amounts of particulate may be collected and weighing
accuracy suffers. There are sometimes problems with unknown
penetration characteristics of the glass fiber bags and
flat filters. In many sampling trains a small cyclone pre-
collector is used to remove larger particulate and allow
longer sampling times.
B.I.8 Sample Flow Rate Meters
The sharp edged orifice meter is a simple and accurate
method to measure instantaneous volumetric flow rate. In
source sampling it is used in conjunction with a total vol-
ume gas meter. As the gas passes through the orifice re-
striction, a pressure drop is created. The following equa-
tion is used for determining the flow rate through an orifice
132
-------
T AP
where Q = Gas flow rate, cm3/sec,
K = Proportionality factor determined by calibration,
m
T = Upstream gas temperature, °K,
Ap = Orifice meter pressure drop, mm
om L
P = Upstream absolute pressure, mm Hg, and
M = Molecular weight of gas, gm/cm-mole.
For a given orifice, K must be determined by calibration.
K is a function of Reynolds number and thus will not be
in
a constant over the entire range of flow rates. However
for a small range of Q , such as for most sampling cases,
K is a constant. Thus it is important to calibrate the
m
orifice for the range of flow rates anticipated. Generally
commercial sampling trains have orifices with delta P
of 0-25 cm HzO over the useful flow rate range. For cali-
bration, see Section B.I.9.
B.I.9 Total Sample Volume
The total volume of gas sampled must be determined in
most sampling trains. This provides the volume necessary
to calculate the particulate concentration. Dry gas meters
with capacities from 0.094-10.8 liters/sec (0.2 to 150 ft3/min)
are generally used. For sampling, the smallest dial face
division should be 5 cm3 (0.01 ft3) because the meter move-
ment is not smooth over one revolution. The dry gas meter
is calibrated using a wet test meter. In a wet test meter
the gas displaces water in a chamber and causes the rotor
133
-------
to revolve. It should be noted that the gas leaving the
wet test meter is saturated with water vapor. The setup
for calibrating a dry gas meter and sampling orifice is
shown in Figure B5. The pertinent information should be
entered on a form similar to that shown in Figure B6. The
following equation is used to determine how well the dry
gas meter performs as compared with the wet test meter. If
the ratio defined by the equation is less than 0.99 or great-
er than 1.01, then the dry gas meter should be readjusted
and recalibrated.
Vw VTd>
V, (P, + AM) (T )
a b W
where y is the ratio of accuracy of the wet test meter to
the dry gas meter,
V is the gas volume passing through the wet test
meter, cm3,
P. is the absolute barometric pressure, mm Hg,
T, is the average temperature of the gas in the dry
gas meter, °K,
T is the temperature of the gas in the wet test meter,
AM is the orifice meter pressure drop, mm Hg,
V, is the volume of gas passing through the dry gas
meter, cm3, and
9 Is the time, seconds, for sampling the gas volume,
V .
w
The data in Figure B6 can be used to calibrate the orifice
meter. In making these calculations the wet test meter is
used for the flow rate, Qm, and
V
Q^ = __ .
134
-------
f>Fi_ow
INCLINED
MANOMETEH
(Am)
3630-086
Figure B5. Set-up for calibration of dry gas meter and orifice meter.
135
-------
DRY GAS METER NO.
DATE
ORIFICE METER NO, -
WET TEST METER NO,
BAROMETRIC PRESSURE Pb -
, mm Hg CALIBRATED BY .
ORIFICE
SETTING
iiM
mm Hg
0,5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
WET TEST
METER
VOLUME
vw
crrT
DRY TEST
METER
VOLUME
Vd 3
crrr
TEMPERATURE
WET TEST
V
°K
DRY TEST
^
°K
Tdo
°K
td
°K
TIME
G
sec
AVERAGE ^
T
km
3S30-O83
Figure 86. Data sheet for calibrating dry gas meter and orifice meter.
-------
A.1.10 Gas Conditioning
Often the gas sample must be conditioned or treated
before or while it is passed through the sampling tcain
components. This is done either to preserve the sample
or to prevent damage to the sampling train. Typical gas
conditioning operations include condensing, drying, heat-
ing, and dilution,
Condensing
Condensers are used to remove water and other vapors
from the gas sample. They work on the priniciple that the
partial pressure of water vapor decreases with a decrease
in sample temperature. For example, as indicated in steam
tables, the partial pressure of water vapor at 0°C is only
0.15% of the partial pressure at 150°C. Thus the use of •
an ice bath type condenser is an effective way to remove
water vapor from a gas sample. This provides the moisture
content which must be determined in order to calculate the
molecular weight of the stack gas.
The ice bath condenser is also used in source sampling
to protect other components from damage. The deposition
of water vapor and water soluble constituents in such com-
ponents as the dry gas meter and pump can cause severe
damage.
The ice bath condenser usually consists of several wet
and dry impingers connected in series, but it may be as simple
as a piece of coiled tubing. A measured initial amount of
water is put into the impinger-type condenser to assist in
the condensation process. When the sampling train is opera-
ted a known amount of sample gas is passed through the system.
137
-------
By observing the pressure and temperature operating condi-
tions, the amount of water condensed, and the amount of gas
passing through the system, the moisture content of the gas
stream can be calculated.
Drying
In source sampling trains, gas drying is used to ac-
complish the same objectives as condensing. The drying opera-
tion is achieved using special chemicals which have a great
affinity for water vapor. One such chemical is silica gel.
The silica gel strongly adsorbs water and hence its change
in weight can be used to calculate the moisture content of
the gas stream. Indicating silica gel, which is granular
and has a bright blue color, can be obtained commercially.
As it becomes saturated with water vapor, its color changes
to a light pink. If this method is to be used to determine
the moisture content of a gas sample, care must be taken
to insure that all part.iculate matter is removed first and
there is no other major constituent in the gas stream which
may also be adsorbed by the silica gel. The silica gel re-
leases (desorbs) the adsorbed water vapor upon heating to
177°C and can be reused.
Often the condenser and drying tube are used in series
to increase water vapor collection efficiency and obtain
a high capacity for the water removed. Large mesh silica
gel (6-16) is used with a filter support backing to prevent
the possibility of entrainment of small particles which might
damage other components.
138
-------
Heating
The heating operation is used solely to preserve the gas
sample prior to passing it through the particulate collector.
This is an effective method of preventing condensation of water
vapor and high molecular weight substances. Therefore it is
common practice to heat sampling probes and particulate collec-
tors such as filters to prevent deposition by condensation.
If such a condensation process were allowed to occur, it would
cause loss of gaseous constituents from the gas sample. In
addition to causing a sampling error these materials could
be deposited in accessible areas of the sampling system and
lead to later malfunctions.
Ideally it is a good policy to try to maintain stack gas
temperatures throughout the sampling train preceding the filter.
However, high temperatures favor chemical reactions such as
oxidation of hydrocarbons in a gas stream containing appreciable
amounts of oxygen. Low temperatures, as mentioned, are conducive
to condensation of water vapor and high molecular weight hydro-
carbons. Hence a compromise is required and most probes and
heated filter boxes operate at about 121°C. Heat sensitive
sampling train components will not be affected by this tempera-
ture. Water vapor will not condense and some o'f the safety
problems involved with the handling of hot equipment will be
alleviated. Particulate compliance tests require that a gas
stream temperature no higher than 115°C be maintained prior
to particulate filtration. Any condensables which are taken
out by the probe and filter under this condition are considered
to be part of the particulate catch. A recent ruling however,
allows the use of temperature up to 120°C when testing at fossil
fuel utility boilers.
139
-------
Dilution
Addition of a dry gas can be an effective method for pre-
venting condensation. When this gas is added the sample is
diluted. Condensation is prevented because the dry gas is
capable of supporting a part of the water vapor from the gas
sample even though the temperature of the mixture is reduced.
The dry gas must be added in such a manner that the original
sample constituents are not altered. This could be a problem
with respect to particulate matter because the dilution (mix-
ing) process could cause such events as particle agglomeration,
deposition, and condensation.
B.I.11
The purpose of a pump is to pull the sampled gas through
the sampling train components. The detail of the particular
type of pump required will depend on several criteria. The
pump must provide adequate flow and pressure characteristics
and be durable and portable. The pump must be able to overcome
the pressure drop of the other sampling train components and
thereby provide the desired flowrate. It must be able to
provide a wide range of flowrates as required by isokinetic
sampling conditions. Often the head loss across the filter
increases through the sampling tests. This puts an added bur-
den on the pump which must still be able to maintain the required
sampling rate at the nozzle tip within the stack.
The pump must be leakless when it is located ahead of the
gas meter in the sampling train. If it isn't, then the metered
volume will be greater than the sampled volume and hence the
measured particulate concentration will be less than the
true particulate concentration. The EPA Method 5 sampling
train falls in this category. In many of the other sampling
trains, the pump is located after the gas meters and there-
fore no error is involved if a leak exists in the pump.
140
-------
The pump must be durable in that it is exposed to corro--
sive environment of the sample gas. During most source tests
it is in constant operation and should be a long life component.
The design should enable this component to be maintained easily;
the key components should be accessible and replaceable with
a minimum amount of time.
The need for portability becomes readily apparent when
performing source tests/ and consequently a small, lightweight
pump is desirable.
There are several types of pumps suitable for source sam-
pling trains. All are of positive displacement types which
are capable of producing relatively high vacuums K686 mm
of Hg below atmosphere pressure) and operate with a direct
linear correspondence between the flow rate and inlet pressure,
In commercial source sampling equipment, reciprocating dia-
phragm and rotary vane pumps are commonly used.
The diaphragm pump operates on the moving diaphragm princi-
ple. Gas is drawn into the chamber on a suction stroke and
pushed out on the discharge stroke. On the suction stroke
a suction valve is open, allowing gas to flow in. On the dis-
charge stroke the suction valve closes and a discharge valve
opens allowing the gas to flow out. This intermittent operation
can cause some flow fluctuation (pulsation) in the sampling
train. However, this problem can be somewhat reduced by running
two such pumps in parallel or by a specifically designed surge
chamber in the flow line. The diaphragm in these pumps is
made out of metal, rubber or plastic.
The rotary vane pump is one rotor in a casing, which is
machined eccentrically in relation to the shaft. The rotor
contains a series of movable vanes which seal against the pump
casing. The vanes are free to slide in and out of the slots
141
-------
as the rotor turns. If the pump must be leakless, then only
the fiber vane type pump with an oiler should be used. The
oiler may have to be modified so that no ambient air leaks
into the system through the oil bowl.
B.I.12 Flow Control
Flow regulation for most sampling trains is accomplished
by using a throttling valve preceding the pump. This valve
varies the vacuum the pump must work against and thereby changes
the flow rate. A more sophisticated arrangement uses two
valves. One precedes the pump and provides a coarse control
while a second one is installed in a recycle (by-pass) loop
to protect the pump and provide a fine control. This latter
arrangement is used in the EPA Method 5 particulate sampling
train. This double valve arrangement is also easier on the
pump, allowing longer pump life.
The major reguirements of the flow control valve are:
(a) it allows sensitive flow rate adjustment to meet propor-
tional sampling (isokinetic) conditions and (b) it does not
allow any leakage. Both these requirements depend upon the
valve construction. The leakage problem poses the same poten-
tial error as was discussed for the pumps. Good quality
needle valves are reguired in most source sampling applica
tions.
B.2 ELECTROSTATIC PRECIPITATOR SAMPLING-PRELIMINARY PROCEDURES
B.2.1 Introduction
Before a sample is taken, several preliminary tests must
be made to determine some of the characteristics of the sam-
pling location and the gas stream. The results of these
142
-------
preliminary tests are the basis for determining suitability
of the sampling location, the nozzle size, number of sampling
points, sampling time, and nomograph settings.
B,2.2 Physical Sampling Location Characteristics
The sampling location should be at a position where the
gas flow is sufficiently uniform that an accurate sample can
bs obtained. Eight to ten duct diameters downstream and two
duct diameters upstream from any disturbance such as bends, duct
inlets, duct outlets, or changes in diameter should give suffici-
ently uniform flow. If the flow at the sampling location is
very uniform, a minimum number of twelve sample points per
sample may be used. If the upstream and downstream diameter
requirement is not met, the flow at the sampling location is
likely to be very non-uniform and require an increase in the
number of sampling points. In such a case, Figure B7 is used
as a guide to determine the number of sample points necessary
to define the velocity profile adequately. Figure B7 is used
by reading the number of sampling points corresponding to both
the number of downstream (A) and upstream (B) diameters and
selecting the greater number of sampling points. A quick pitot
tube survey should indicate whether a sufficient number of
points has been chosen to define the velocity profile adequately.
After the number of sample points is selected, the cross sec-
tion of the duct is divided into a number of equal areas as
shown in Figure B8. In the case of round stacks, the sample
point is located such that half of the area increment repre-
sented by that point is radially on each side of the sample
point. The location of sampling points is determined as shown
in Table Bl. The area increments must be small enough to insure
that the flow at the sampling point in each area is representa-
tive of the flow in the area; however, the total number of
area increments must be limited enough so that all the points
may be sampled within a reasonable period of time,
143
-------
NUMBER OF DUCT DIAMETERS DOWNSTREAM FROM PORT
(DISTANCE A)
1.0
8
NUMBER OF DUCT DIAMETERS UPSTREAM FROM PORT
(DISTANCE B)
10
3630-066
Figure B7. Minimum number of traverse points per sample obtained
from "Distances to Disturbances", upstream and downstream.
144
-------
CROSS SECTION OF CIRCULAR DUCT
DIVIDED INTO 12 EQUAL AREAS,
SHOWING LOCATION OF TRAVERSE
POINTS.
CROSS SECTION OF RECTANGULAR DUCT
DIVIDED INTO 12 EQUAL AREAS, WITH
TRAVERSE POINTS AT CENTROID OF
EACH AREA.
3630*093
Figure B8. Examples of equal area sample points.
145
-------
TABLE Bl
Duct Traversing Length Factors
LENGTH FACTORS, K
LJ
(Fraction of stk. diam. from inside wall to traverse pt,)
Traverse
Point
Number
1
2
3
4
5
6
7
8
9
,1.0
11
12
13
74
15
16
17
18
19
20
21
22
23
24
NUMBER OF TRAVERSE POINTS ON
2468
.146 .067 .044 .033
.854 .250 .147 .105
.750 .295 .194
.933 .705 .323
.853 .677
.956 .806
.895
.967
10
.025
.082
.146
.226
.342
.658
.774
.854
.918
.975
12
.021
.067
.118
.177
.250
.355
.645
.750
.823
.882
.933
.979
14
.018
.057
.099
.146
.201
.269
.366
.634
.731
.799
.854
.901
.943
.982
A DIAMETER
16
.016
.049
.085
.125
.169
.220
.283
.375
.625
.717
.780
.831
.875
.915
.951
.984
18
.014
.044
.075
.109
.146
.188
.236
.296
.382
.618
.704
.764
.812
.854
.891
.925
.956
.986
20
.013
.039
.067
.097
.129
.165
.204
.250
.306
.388
.612
.694
.750
.796
.835
.871
.903
.933
.961
.987
22
.011
.035
.060
.087
.116
.146
.180
.218
.261
.315
.393
.607
.685
.739
.782
.820
.854
.884
.913
.940
.965
.989
24
.01.1
.032
.055
.079
.105
.132
.161
.194
.230
.272
.323
.398
.602
.677
.728
.770
.806
.839
.868
.895
.921
.945
.968
.989
146
-------
In rectangular ducts or stacks, the cross section is di-
vided into a number of equal area rectangles. The sample is
taken at the centroid of each rectangular area. These areas
should be laid off such that the ratio of the length to the
width of the elemental areas is between one and two.
The minimum number of sample points is twelve, and the
same criteria as for round ducts is used to insure that the
velocity profile is adequately defined. The equivalent dia-
meter of a square duct is approximated by 2LW/L+W where L and
W are the duct cross-sectional dimensions.
In most cases the dimensions of the ducts obtained from
construction drawings are accurate? however, the inside dimen-
sions should still be measured if feasible, particularly in
the case of horizontal ducts on the bottoms of which dust
deposits of considerable thickness are often found. The pitot
tube may be used to make this measurement but the ends should
be protected to prevent material from the back wall from clog-
ging the ends. Another critical measurement is the length
of the port extension. With these measurements, the pitot
tube is marked at the points to be sampled. If the outermost
points are less than one inch from the walls, they should be
located at one inch and noted on the sampling form. The re-
quired length of the probe (pitot tube) for each of the points
may be marked with hose clamps, tape, or other suitable mater-
ial compatible with flue gas conditions.
Stack pressure is determined with a leveled and zeroed
manometer. The pitot tubes are aligned perpendicular to the
flow stream in the stack and one of the two pitot lines is
disconnected from the console. If the stack pressure is posi-
tive gauge pressure, the manometer will show positive deflection
with the one pitot line connected to the positive side of the
manometer. If the stack pressure is negative gauge pressure,
the manometer will show positive deflection with the one pitot
147
-------
tube line conncected to the negative side of the manometer.
The stack gauge pressure (P ) in millimeters of Hg is obtained
o
by adding the stack to ambient differential pressure, APD,
to the ambient pressure (
P = AP + P
s D AMB
The preliminary flue gas temperature is obtained by a suitable
means such as a stem thermometer placed in the sampling port
or a thermocouple with an appropriate readout device.
B.2.3 Veloci ty Determina t ion
Before velocity measurements are taken, the inclined mano-
meter must be leveled and zeroed and must remain level during
sampling. The openings of the pitot tubes should be shielded
from any wind currents but not be completely closed off when
the manometer is zeroed. Correct connection of the pitot tube
lines may be checked by blowing gently on the upstream pitot
tube opening and noting the response on the manometer. The
probe is then inserted, the pitot reading noted and the pitot
tube lines are switched both on the console and on the probe.
If the manometer reading is the same as that prior to switch-
ing the pitot tube lines it is reasonably certain that there
is no significant leak in the lines. If a leak is detected,
it must be eliminated before any readings are taken.
The pitot tube lines must not be pinched or the tube
stopped up during the traverse. If fluctuations in the mano-
meter are noted, pieces of cotton or glass wool may be placed
in the pitot tube lines to dampen the fluctuation but should
not be packed too tightly. Since the pitot tube measures
pressure differences, there is no actual air flow through
148
-------
the lines. If one line is completely plugged, however, the
results will not be accurate. In some instances condensation
of water within the pitot tube lines can cause difficulties
and erroneous readings.
Once the traverse has been completed and the pressure
and temperature readings have been recorded the velocity
may be determined. Velocity may be calculated as follows:
= (422.
~ avg
where (V ) is the stack gas velocity, cm/sec,
avg
T is the average stack temperature, °K,
o
AP is average stack gas velocity head, mm Hg,
M is the molecular weight of the stack gas, wet
o
basis, gm/gm-mole,
P is the absolute stack pressure, mm Hg, and
S ~"
C is the pitot tube correction factor.
B.2.4 Stack Moisture Content
The stack moisture content is an important factor in
stack sampling. Nozzle size selection and sampling rate
are both dependent on the moisture content. A condenser
method or a determination (based on the dry bulb tempera-
ture with knowledge that saturated conditions exist in the
stack) are two ways of determining moisture content. A
wet bulb-dry bulb technique requires less equipment but
must be limited to non-acid gas streams with moisture con-
tents of less than 15% and dew points less than 52°C. The
condenser method works well for most gas streams and is rela-
tively easy to perform.
149
-------
B.2.4.1 Condenser Method
Several condenser techniques can be used to determine
stack moisture content. One such technique uses a Greenburg-
Smitb impinger approximately half full of water followed
by a straight impinger approximately half full of silica
gel. A measured volume of stack gas, usually 10 cubic feet,
is drawn through the impingers at a moderate flow rate.
The total change in weight of the impingers is the weight
of the moisture caught. The impingers should be in an ice
bath while the stack gas is drawn.
With the impinger volume increase in milliliters, v ,
the stack gas volume in cm3 (V ) corrected by the dry gas
meter correction factor, the absolute average meter tempera-
ture in °K (T ), and the meter pressure in mm of mercury
(P ), the moisture fraction (BWQ) is calculated as follows:
Vwc (1243.34 cmVml(H20))
wo
V (1243.34 cm'/ml(H20))
B.2.4.2 Saturation Method
If water droplets are present in the stack and the stack
gas temperature is below 100°C, the gas stream may be assumed
to be saturated. The moisture content is read from the satu-
ration curve on the psychrometric chart at the stack gas tem-
perature, Figure B9.
150
-------
100
o
CD
1-
H
LU
O
E
til
o.
IT
O
IT
LU
60
80
100 120 140 160
STACK GAS TEMPERATURE, °F
180
200
220
3S30-O71
Figure B9. Percent water vapor in air at saturation.
151
-------
B,2.4.3 Wet Bulb-Dry Bulb Method
In this method two thermometers are placed in the gas
stream. One is dry and the other has a wet sock over the bulb.
The temperatures are read after they stabilize. If the stack
or duct pressure is near atmospheric pressure, the percent
moisture may then be found from the psychrometric chart, Figure
BIO.
The percent water vapor by volume is found directly on
the ordinate axis. Inputs are the dry bulb temperature on
the abscissa, the wet bulb temperature, and the sloping lines
which terminate at the saturated vapor line.
When obtaining wet bulb-dry bulb readings with a sling
psychrometer, the plane of the thermometers should be perpendicu-
lar to the flow of gas. If it is parallel to the flow, the
dry bulb should be upstream of the wet bulb. The gas velocity
past the wet bulb should be from 3.7 to 9.3 meters per second.
Sufficient time must be allowed for the wet bulb temperature
to stabilize or inaccurate results will be obtained.
B.2.5 Molecular Weight of the Stack Gas
The most common method of determining the composition
of combustion effluents is the Orsat apparatus. Although flue
gases vary in composition, they normally contain CO2 , CO, Oz ,
HjO, and Nz . The Orsat analysis determines the quantities
of these components (except HaO) present by successive removal
using suitable absorbents and measurement of the volume changes
of the original sample. The Orsat analysis, as it is normally
used, measures the percentage of CQz , Qz , and CO in the sample.
The difference is largely Na. By changing the absorbents other
components may also be measured.
152
-------
30 40 50 60
70
80 90 100 110 120 130
DRY BULB TEMPERATURE, °F
140 150 160 170 180
3630-072
Figure BIO. Percent water vapor with wet and dry bulb.
153
-------
After the Orsat analysis has been performed, the molecu-
lar weight of the stack gas may be determined by the following
equation:
proportion by vol.
of component on
dry gas basis L component]
Mscof
where Mc is the molecular weight of the stack gas,
w
Bn_ is dry gas fraction of the stack gas, and
M~r is the molecular weight of each component of the
stack gas.
B.3 ELECTROSTATIC PRECIPITATOR SAMPLING FOR PARTICULATE AND CASE:
B.3.1 Isokinetic Sampling
To obtain a representative particulate sample, the sample
must be collected at a rate as nearly isokinetic as possible,
i.e., the kinetic energy of the gas stream in the stack is
equal to the kinetic energy of the gas stream through the sam-
pling nozzle. Since the composition of the two gas streams
is the same, this energy balance simplifies to: the velocity
in the stack is equal to the velocity through the nozzle.
If a particulate sample is not pulled isokinetically, inaccu-
rate results may be obtained.
Whenever an object is placed in a moving gas stream, some
disturbance of the flow patterns will occur. The purpose of
isokinetic sampling is to minimize any disturbance caused by
the sampling nozzle. A sample collected isokinetically through
a sharp-edged nozzle, should create very little disturbance.
Figure Bll illustrates this point.
1.54
-------
\\\
///
i, i, I, h ii In.
I , i , , i
ISOKINETIC
OVER ISOKINETIC
UNDER ISOKINETIC
3630-064
Figure Bit. /sok/netic flow patterns.
155
-------
Large (heavy) particles tend to travel in a straight line
and are not greatly affected by flow disturbances, whereas
small (light) particles tend to follow the flow lines. in
a gas stream with a homogeneous distribution of large and
small particles, over-isokinetic sampling will give a low particu-
late mass rate (PMR) because fewer large particles will be
caught than are representative of the flow stream from which
the gas was withdrawn. On the other hand, under-isokinetic
sampling will give a high PMR due to a greater than represen-
tative number of large particles that will be caught.
The velocity of a gas stream i.n a stack generally varies
from point to point; therefore, the tiow rate or velocity through
the sampling nozzle must be adjusted to maintain isokinetic
conditions at each sampling point. In the sampling train, de-
termination of the nozzle volume and the flow rate through
the nozzle are based on dry gas volume and flow rate measured
at approximately ambient temperatures. For this reason,
the flow rate through the orifice meter which corresponds
to the desired flow rate through the sampling nozzle must
be determined. The stack velocity as measured by the pressure
drop (AP) across the pitot tube and the velocity through
the nozzle as measured by the pressure drop (AH) across an
orifice meter at the end of the sampling train must be equal
in order to maintain isokinetic flow. To speed up this cal-
culation, two nomographs were developed by the old National
Air Pollution Control Administration (NAPCA). Through the
use of these nomographs, the proper size sampling nozzle may
be selected and the flow adjustments required to maintain
156
-------
isokinetic sampling conditions may be determined. An alter-
nate method employs the use of one of the NAPCA nomographs
or a calculator for adjusting flow rates along with a mathe-
matical method of nozzle selection. These two procedures
give equivalent results.
B.3.2 SamplingFor Effluent Gases
When effluent gases and particulates are to be sampled
simultaneously the sampling must be performed isokinetically.
It is advisable however, to choose a nozzle which will give
a low volumetric flow rate to optimize gas absorption effi-
ciency in the impinger train; 142 cm3/sec (.3 ft3/min) or
less is desirable. If only gases are to be sampled and if
they are well mixed, isokinetic sampling is not necessary.
B.3.3 Nozzle Selection and Nomograph Setting
Isokinetic sampling involves maintaining the flow rate
through the sampling nozzle such that the velocity in the
nozzle equals the velocity in the stack at the sampling point,
Obviously, the flow rate through almost any size nozzle could
be adjusted such that this velocity requirement is met.
There are certain physical limitations, however, placed on
nozzle size by the sampling equipment: pump capacity, port
diameter, filter efficiency, and the critical flow through
the Greenburg-Smith impingers. Another limiting factor in-
volves the reliability of the sample. Small nozzles can
yield less representative samples when large particles are
present. As the ratio of nozzle tip area to stack cross-sec-
tional area decreases, the chance of sampling at a point
where the flow is not representative of the flow in the
stack area that the point represents increases. The nozzle
157
-------
should also be at least a little larger than the largest
particles that might be encountered in the stack. Some
guidelines for nozzle size selection are given in the next
section.
B.3.4 EPA Reference Method 5 Procedure
Isokinetic sampling, the condition of equal velocities,
implies a mathematical relationship between the two pressure
drops, AP and AH. The pressure drop measured by the pitot
tube, AP, indicates the stack velocity and the desired velo-
city through the sampling nozzle. The pressure drop across
the orifice plate, AH, represents the flow rate of dry gas .
through the dry gas meter. This relationship has been incor-
porated into two stack sampling nomographs, Figures B12 and
B13 . To operate the nomographs, a factor which is a composite
of the test constants is obtained from one nomograph and
is used to set up the second nomograph. This factor will
be called "C". As the pressure drop across the pitot tube
(AP) changes from point to point, an updated desired value
for the orifice pressure drop (AH) is found from the second
nomograph. The flow rate may then be adjusted to give this
desired pressure drop across the orifice which establishes
isokinetic conditions.
In using the first nomograph, the following parameters
are required to determine the "C" factor which will be car-
ried over to the second nomograph:
1. Pressure drop (AH) across the orifice plate when 0.75
cubic feet per minute of dry gas is flowing and the pres-
sure and temperature are 760 mm mercury and 21.1°C
respectively is determined. Note that the calibration
factor of the orifice must be determined in the lab.
158
-------
REF 1
.3.0
-2.5
.2.0
-1.5
-1.0
A
*-__,
-50-
% HO
150
— 10D ....... M-
50
rREF 2
C
:— 2.0
:—1.5
I—1.0
^^(LB
0,5
EXAMPLE: AH = 2.7 in.
H2O
AH = 2.7 IN. H20
Tm = °°F
%H2O = 30
FIND C
10-
20-
3D-
40.
50-
DRAW LINE FROM AH TO T^F) TO OBTAIN POINT A ON REF, 1.
DRAW LINE FROM POINT A TO %H20 AND READ B ON REF. 2.
DRAW LINE FROM POINT B TO Ps/Pm. AND OBTAIN ANSWER OF
0.85 FOR C.
3630-061
Figure B12. Stack Sampling Nomograph (side 1}.
•1,2
-1.1
-1.0
.0,9
• O.B
159
-------
ORIFICE READING
AH
9 —
O ~ —
"""5
-I
6— |
—
5— |
-H
4_=
_
_z
™
3 ^^^
^D
~
~
_
2 • m
_
-
-
I
1 mmmmi
0.9 —
0.8
0.7— =
0-6—1
-=
0.5—|
-^
0.4—=
-E
0.3—3
Jjj
j[
0.2— -
I
—
~
-
-
—
0.1
Raf,
— Ref.
2,0
-
1- 1,5 c
CORRECTION
FACTOR
1 0
— 0.9
— 0.8
0.7
-
0.6
~
0.5
2500
2000
1500
1000
800
600
500
400
300
200
100
0
•—
—
^1—
— LL
- O
F*^^ V*
"~ ^
— ill
_ EC
— D
; — ;_ (-
~ CC
>— — ^ LU
~ 5 §
—
r~
|—
§_
i-
=—
SLIDING
SCALE
V
X\ /
^^\S
CUT ALONG LINES
0.001 1
K FA
AH = IN. H2o
CTOR PITOT READING -
AP -=
0.002—=
-|
0.003-1
0.004— =
0.005—1
PROBE ~
TIP DIAMETER 0.008-=
D 0.01-=
C= DIMEMSIO^LESS
T ~ OF
5
K = DIMEIMSIONLESS
D = IN.
AP = IN. H2O
p — 1.0 I
— ^ -
•" f\ Q
""" ~
=__08 0.02—=
~ ' —
5 — 0.7 0.03—|
z
0.04— =
=—0,6 =
= 0.05—=
= 0.06—=
=—0,5 -±
= 0.08-=
E~ 0-1 ""=
E— 0.4 -
_ _z
E~ 0.2—;
— -i
^—0,3 0.3—=
I 0.4— |
~ 0.5—|
~ 0,6-^
0.2 a8^
I 1,0-=
- ~
— ~-
2~1
3—=
„_=
~
^^ 0.1 ^ "™~r
6 ^^
8-E
10 —
3630-084
57J, Stack Sampling Nomograph (side 2).
160
-------
2. Percent HjO in the stack gas is determined prior to
sampling.
3. Expected temperature at the orifice must be estimated
from experience. It is usually at least ten degrees
Fahrenheit higher than ambient temperature and is the
same as the outlet temperature from the dry gas meter.
4. The ratio of absolute stack pressure to atmospheric
pressure is also required. This ratio is approximately
equal to one.
Proceed to find "C" as follows:
A. Obtain a point on the "reference one" line by connecting
the orifice pressure drop, AH, (for 0.75 cubic feet per
minute) to the temperature of the orifice plate (T J.
B. Draw a line from this point on the "reference one" line
to the percent H2O to obtain a point on the "reference
two" line.
C. Connect this point to the pressure ratio point? this
line crosses the "C" scale and gives the value required
to set up the second nomograph.
After the initial velocity traverse, and prior to sampling, set
up the second nomograph as follows:
A. Set the vertical sliding scale so that the "C" factor men-
tioned above is at the reference point.
B. From the velocity traverse, calculate the average AP. A
line from this point to the stack temperature intersects
the probe tip (nozzle) diameter scale.
.161
-------
C. From the nozzles on hand, select one near this size.
Reconstruct the line through the stack temperature and
use the selected probe tip diameter to give a reference
point on the AP scale.
D. A line from the reference point on the AP scale to the
permanent reference point on the AH scale locates the
pivot point on the K factor line. Lock the pivot bolt
that carries the clean plastic rule at this point. The
nomograph is now set up for isokinetic sampling. Note
that stack temperature is assumed to remain constant.
For each point of the sampling traverse the pressure drop
across the pitot tube, AP, is used as input to the nomograph.
The output is the desired pressure drop across the orifice,
AH, which is required to maintain isokinetic flow. If the
temperature of the stack changes appreciably during sampling,
(±5% on an absolute temperature basis) the pivot point must
be reset. Once the pivot point on the K factor line has been
set, it is suggested that the maximum and minimum AP's from
the velocity traverse be used to determine the range of AH's.
This range should lie between 0.3 and 6 inches of H20. If
it does not, a different choice of nozzle tip diameter should
be considered.
B.3.5 General Sampling Procedures
B.3.5.1 Preparing the Glassware
The glassware must be prepared and placed in the sampling
box. All of the glassware should be clean and dry. The glass
liner in the probe should be washed and the nozzle attached
to the probe. In the sampling box, the sample first passes
through a glass cyclone which has a flask attached to the
bottom to catch large particles that the cyclone separates
from the gas stream. The sample then goes through a fritted
glass filter holder. The filter holder is 6.4 cm to 10 cm in
162
-------
diameter. A glass fiber filter is placed over the fritted
surface. The filter should be preweighed to ±0.0001 grams.
From the filter the sample goes into the condenser section.
The type and number of impingers to be used in this section
is dictated by the type of sample to be taken. The different
impinger solutions and their uses are listed in the following
paragraphs. Gummed labels can be used to label the impingers
and should be placed near the top to prevent water from be-
ing absorbed from the ice slurry. The impingers should be
weighed to ±0.1 gram before and after sampling. A very light
coating of silicone lubricant should also be placed on the
ball joints connecting the glassware to insure a vacuum-tight
seal. The impinger solution volume is normally 200 ml.
B.3.5.2 Impinger Trains
The following tables give some recommended impinger (IMP)
trains.
Preparation of Impinger Solutions
80% Isopropanol: 160 ml Technical Grade isopropanol +
40 ml deionized (DI) water.
6% H202: 40 ml Reagent Grade 30% Hydrogen Peroxide + 160
ml DI water.
0.1 N NaOH: 20 ml 1 N NaOH + 180 ml DI water.
1 N NaOH: 40 grams Reagent Grade NaOH (pellet form)
dissolved in DI water made up to 1.0 liter volume.
Alkaline arsenite solution (0.500 N NaAsO2 in 2.5 N NaOH):
Dissolve 100 grams Reagent Grade NaOH (pellet form) and
32.5 grams of NaAs02 in DI water and dilute to 1.000 liter
with DI water.
0.1 N H2SO^: 10 ml 2 N Hz SO- diluted with DI water to
200 ml. Prepare 2 N HaSO^ by adding 1 volume of Reagent
Grade concentrated H2SO^ (18 M) to 17 volumes of DI water.
0.1 M zinc acetate: 22 grams Zn(C2H3O2)2. 2H2O in 1.0
liter DI water.
163
-------
I. Particulate Only
Application: Hot mix plants, ore sintering processes,
gypsum manufacturing, etc.
IMP.
NO.
1
2
3
4
CONTENTS
DI Water
DI Water
Dry
Silica Gel
POLLUTANT
OR COMPONENT
Particulate
Particulate
Water
Water
IMPINGER TIP
CONFIGURATION
Straight
Greenburg-Smith
Straight
Straight
(G-s;
II. Particulate + Sulfur Dioxide
Application: Non-ferrous smelters, Portland cement
kilns, coal-fired boilers, fluid catalytic cracking
units, sulfuric acid plants, boiler recovery stacks
(kraft paper mills).
IMP
NO.
2
3
4
5
CONTENTS
80%
isopropanol
6% H2O2
6% Hj02
Dry
Silica Gel
POLLUTANT
OR COMPONENT
Particulate,
HzSO* , SOs
S02
S02
Water
Water
IMPINGER TIP
CONFIGURATION
Straight
G-S
G-S
Straight
Straight
NOTES:
The impinger train must be purged with two
cubic feet of ambient air at the end of the
run to sweep SO; out of impinger #1 into im-
pingers 2 and 3.
The first impinger not only is effective .i r,
trapping sulfuric acid but will also trap
submicron particulate that passes through
the filter.
III. Particulate + Chlorine and/or Chlorides
Application: Pulp bleaching (effluent from), Magnesium
plants (drying of magnesium chloride).
164
-------
IMP
NO.
1
2*
3
4
5
CONTENTS
DI Water
Alkaline
Arsenite
Solution
Alkaline
Arsenite
Solution
Dry
Silica Gel
POLLUTANT
OR COMPONENT
HC1, Cl~
Particulate
HC1, C12
Particulate
HC1, C12
Water
Water
IMPINGER TIP
CONFIGURATION
G-S
G-S
G-S
Straight
Straight
*NOTE: It is important to measure the volume of this
solution as accurately as possible since the
analysis will be based on the molar quantity of
arsenite remaining.
IV. Particulate in the Presence of Hydrogen Fluoride
Application: Alumina reduction plants and phosphate
fertilizer manufacturing (acidulation process).
IMP POLLUTANT IMPINGER TIP
NO. CONTENTS OR COMPONENT CONFIGURATION
1 Water Particulates, G-S
inorganic
fluor ide
particulate
2 Water Particulates, G-S
inorganic
fluoride
particulate
3 Dry Water Straight
4 Silica Gel Water Straight
NOTE: Use filter bypass instead of fritted filter.
V. Particulate + Ammonia
Application: Effluent from manufacturing, prilling
and drying of ammonium nitrate or ammonium phosphate
fertilizer.
165
-------
VI.
IMP
NO.
1
2
3
4
5
NOTE:
CONTENTS
Water
0.1 N H2S04
0.1 N H-,SOk
Dry
Silica Gel
POLLUTANT
OR COMPONENT
Particulate
NH3
NH3
Water
Water
IMPINGER TIP
CONFIGURATION
G-S
Straight
Straight
Straight
Straight
Water is used in the first imp.; rigor in order t.^t
particulate may be measured. The contents of chis
irnpinger will be saturated witTi ammonia. Partial
recovery of ammonia from the first imp ing t- . ir« /
be accomplished by purging the impinger trriin with
approximately 2 ft3 of ambient air.
Particulate + Hydrogen Sulfide
A. Application: Carbon black plants (furnace effluent).
lime kilns (kraft paper mills).
IMP
NO.
CONTENTS
POLLUTANT
OR COMPONENT
IMPINGER TIP
CONFIGURATION
1
2
3
4
B.
IMP
NO.
1
2
3
4
5
NOTE:
0.1 M zinc
acetate
0.1 M zinc
acetate
Dry
Silica Gel
Application :
stacks .
CONTENTS
Water
0 . 1 M zinc
acetate
0.1 M zinc
acetate
Dry
Silica Gel
HaS
H2S
Water
Water
kraft pulp mills, boile
POLLUTANT
OR COMPONENT
Particulate
SO 3 , H zSOi*
HzS
H2S
Water
Water
With a conventional sampling train Hz
G-S
,q _ ,-
Straight
Straight
r recovery
IMPINGER TIP
CONFIGURATION
Straight
G-S
G-S
Straight
Straight
S and SO 2
cannot be determined simultaneously.
166
-------
B.3.5.3 Checking the Sampling Train for Leaks (Vacuum Check)
A vacuum-tight system is necessary to prevent any dilution
air from being pulled into the sampling line. After the probe
and sample box are connected and suspended from the monorail
or other support, the umbilical cord containing the sample
line, pitot tube lines and thermocouple leads may be connected.
The inlet side of the cyclone is then sealed for the vacuum
check. The pump is then started with the coarse valve closed
and bypass valve open. As the coarse valve is slowly opened
the vacuum will begin to increase. The bypass valve is slowly
closed until the vacuum reaches 381 mm gauge. At this vacuum
the flowrate through the dry gas meter should not exceed 9 cm3/
sec (.02 ft3/min). If a leak is present all connections
should be checked to eliminate the leak and the above procedure
repeated. The seal on the cyclone must be removed slowly before
the pump is turned off to prevent liquid backup in the impingers,
After the above vacuum check is performed the probe liner is
connected to the cyclone and a vacuum check on the total sys-
tem may be performed. The pump is started again and the coarse
valve is slowly opened. After the flow starts, the nozzle
tip is sealed and vacuum should start to build in the system.
When the sample line vacuum reaches 381 mm the coarse valve
should be closed. The vacuum should hold steady if there are
no leaks. If it holds for about fifteen seconds, the nozzle
tip is opened and the vacuum should drop. Then the pump is
turned off with care not to back up liquid in the impingers.
Under no circumstances should the pump be turned off while
the nozzle tip is sealed. The power line is then plugged in
and the heater turned on. The probe heater, if used, should
be connected and the probe heater turned on. The pitot tube
manometer may then be zeroed by shielding the end of the pitot
tube from any wind or disturbance. The ends should not be
plugged, however. The orifice manometer should also be zeroed.
The manometer must be level at this time and throughout the
sampling period. The pitot tube should also be checked to
167
-------
be sure the lines are hooked up properly. The man on the plat-
form can blow lightly on the upstream pitot tube while the
console operator checks the manometer displacement. Before
sampling begins, the console operator should be sure the
manometers are leveled and zeroed, the temperature indicator
is working properly, and the nomograph or calculator is pro-
perly set up.
The console operator should have decided on how long to
sample at each point. The number of sample points is determined
by generally accepted rules, e.g. velocity traverse, but the
sampling time is based on knowledge of the plant operations and
the approximate particulate loading which may be obtained from
plant personnel on a pretest. The sample collected on the
filter should be large enough to weigh accurately and be repre-
sentative of the conditions in the stack; however, care must
be exercised to prevent the filter from clogging, or in wet
stacks, the impingers from filling up. It is preferable to
sample for at least an hour. If the process is cyclic and
portions of the cycle give high particulate mass rates (PMR)
which are not upset conditions, these portions of the cycle
must be included in the sample. The sample size should be
at least 850 liters at standard conditions. If conditions
permit, some adjustments in nozzle size and sample flowrate
can often be made to satisfy time requirements. It is usually
best to sample each point for no less than three minutes as
this allows time for adjusting flows and recording data.
B.3.5.4 Sampling
When the console operator is ready and the probe and sample
box have heated sufficiently, the initial dry gas meter reading
is recorded and the sampling probe is pushed carefully into
the duct to the point nearest the back wall. This allows the
probe to cool in hot stacks as it comes out, shortening the
168
-------
time required for cooling after the sample is taken. This
also allows the use of the stack heat to help heat the probe.
The nozzle must not hit the back wall or the inside of the
port where deposited material might contaminate the sample.
If this procedure is followed, the last point sampled will
be the point nearest the port. As soon as the probe is po-
sitioned the operator should record the time and start the
run by turning on the pump, opening the coarse valve and
adjusting the bypass valve until the desired flowrate calcu-
lated from the nomograph or K factor is obtained. The data
should then be recorded. The probe crew should be notified
15 to 30 seconds before the probe is to be moved and the
signal to move to the next point is given approximately
5 seconds before that time. When the probe is repositioned,
the operator should read the new AP, use the nomograph or
calculator, adjust the flow rate, and record the data required
on sampling field data sheets. The opening at the port
should be plugged to prevent dilution or abnormal distortion
of the flow patterns in the stack. After the last point
has been sampled, the operator turns off the pump and re-
cords the meter reading. If more than one port is to be
sampled, the sampling box is transferred to the next port
and the above procedure repeated. The probe is then removed
carefully from the port so the open end of the nozzle does
not hit the port. The probe should also be kept horizontal
and the nozzle plugged as soon as possible after it is re-
moved from the stack to prevent loss of sample. In some
cases condensate will collect in the probe and if the probe
is tipped, some of the sample might be lost. The particulate
trapped in the probe can represent a significant portion
of the total sample. Normally three sample runs are taken.
Each run normally consists of two traverses. Each run is
considered as a separate sample, and the calculations are
performed for each sample. The results of the runs are
then averaged.
Before making a second run the percent isokinetic of
the first run should be checked. The nomograph or calculated
169
-------
K factor should be changed and an additional sample taken
if the percent isokinetic varies by more than ±10% from
100%. (See Data Reduction Section.)
B . 3.5.5 Sample Handling
At the end of each sample run the electrical power
is disconnected and the hot side of the sample box is opened
to allow the glassware to start cooling. The pi tot tube
lines to the probe are disconnected and the probe is removed
from the box and the ends plugged as soon as possible.
The impingers may now be removed and weighed. After the
impingers are weighed, the liquid in the impingers is placed
in clean sample bottles to be taken to the lab. Some particu-
late or condensable compounds will occasionally get past
the filter and be collected in the liquid. The sample catch
from the filter and the probe washings should be collected
and stored separately from the impinger catch. The probe
and nozzle are washed carefully and the washings collected
in a clean sample bottle to be taken to the lab, evaporated
and then weighed. The cyclone and flask and the connecting
glassware are washed and the washings added to the probe
washings. The weight of any water caught in the cyclone
must be determined and added to the impinger weight gain.
The filter is removed from the holder, folded with the parti-
culate side in and placed in an envelope to be taken to
the lab and weighed. The filter holder is washed and the
washings added to the probe wash. Usually distilled water
is used to wash the glassware, but in some instances, ace-
tone may be used. Precautions must be taken to eliminate
the possibility of tampering with, accidental destruction
of, and/or physical and chemical action on the samples.
To reduce the possibility of invalidating the results,
all components of the sample should be carefully removed
170
-------
from the sampling train and placed in sealed, nonreactive,
numbered containers. The samples are then delivered to
the laboratory for analysis. It is recommended that this
be done on the same day that the samples are taken. If
this is impractical, all samples should be placed in a carry-
ing case (preferably locked) in which they are protected
from breakage, contamination, and loss.
Each container should have a unique identification
to insure positive identification and to preclude the possi-
bility of interchange. The identification of the container
should be recorded on the analysis data sheet so it will be
associated with the sample throughout the test and analysis.
The samples should be handled only by persons associated
in some way with the task. A good general rule to follow is
"the fewer hands the better", even though a properly sealed
sample may pass through a number of hands without affecting
its integrity.
B.4 MASS CONCENTRATION DATA REDUCTION
After performing particulate mass measurements on the
inlet and outlet of an electrostatic precipitator, the data
must be reduced to obtain particulate emissions concentrations
and other pertinent parameters.
To illustrate these calculations, an example of data col-
lected on a single test will be presented along with the appro-
priate calculations to obtain the necessary results.
From the particulate source test data given in Table B2
we find:
Orifice APat7ri = 3.69 mm Hg
a. V y
171
-------
TABLE D2
EXAMPLE - PARTTCULATE SOURCE TEST DATA - ESP INLET
General
Firm - CM 6, S Public Service
Source - Coal Boiler SI
Sampling Location - ESP Inlet
Equipment
Can Meter No.-3
Sample Box No,-8
Probe No.-3 Length-305 cm Cp-.B7
Date - 3/1
Test No. -5
Witnessed
Traverse
Pt.
1
2
3
4
5
6
7
a
9
10
11
12
Totals
Averages
a/77
by - D.D,
Time
(sec)
300
300
300
3GO
300
300
300
300
300
300
30G
300
3600
.11.
Dry Gas
Meter
(xlO-'m3)
53.5
244.4
41-3.7
677.4
929.5
1088.3
1311.5
1613.4
1928.6
2137.9
2331.3
2560.1
2800
Pitot AP
mm Hg
.99
1.08
1.18
0.45
1.53
1.62
] .77
1.76
1.55
1.18
] .12
1 .08
1.28
Or if ice
mm Hq
4.52
4.89
4.86
4. 30
3.31
3.16
2.97
2.80
3.29
3.10
3.38
3.74
3.69
Filter No.
212
Probe Wash
Meter
AH Inlet
°C
23
24
24
24
25
25
25
25
26
27
28
28
25.3
Silica Gel
87
Sample No. -
Temp.
Outlet
°C
20
20
21
22
22
22
22
22
22
23
23
24
21.9
No.
KC10
Stack
Temp.
"C
145
146
150
150
145
147
151
151
153
150
147
145
14B.4
Conditions
Orifice AH - 3.74 ram Hg
Assumed %HZ0 - 8% T - 149eC
S
Assumed APB,,_-1.5 mm Hg
AVG P /P -1.0
Assumed TM-26.7^C ""A - 9.29 m2
Bar Pressure - 749 mm Hg
C-Factor - .860
Probe Heater Set - 135°C
Filter Oven Sot - 121°C
Pump
Vacuum Stack
mm Hg Pressure
-7.5 m-n Hg
Test End - 12:00 noon
Test Start-ll:00 a.m.
152. 4
152.4
152.4
177.8
177.8
152.4
152.4
152.4
127.0
152.4
177.3
177.8
Impinger Outlet
Max Temoerature
10°C
Condensate
Collected
#1 100 ml
#2 50 ml
#3 10 ml
1.69 mmHg
CO;
11.0%
11.5%
12.0%
Oz
6.0%
5.5%
5.8%
Silica Gel Mass
Charge - +46 grams
Filter anr) Probe
Wash Catch - 150 mg
-------
T = 23.6°C - 296.9°K
mavg
B.4.1 Volume of Gas Sampled
T f P
V = V » V. std(Pbar
mstd m Tm Pstd
„ 294 (749 + 3.69
. o
297 760
= 2.75 DNM3
where V is the volume of gas sampled through the dry
gas meter at standard conditions (21°C, 760
mm Hg), DNM3.
V is the volume of gas samples through the dry gas
meter at meter conditions, m3.
Tstd is 21°C, 294°K.
T is the average dry gas meter temperature,
23°C, 297°K.
P, is the uncorrected barometric pressure at the
outlet of the orifice meter, 749 mm Hg.
AH is the average pressure drop across the orifice
a ve
meter, 3.69 mm Hg,
P . , is the absolute pressure at standard conditions,
760 mm Hg.
B.4.2 Volume of HzO Vapor in Stack Gas
Water condensed in impingers = 160 ml, water absorbed
on silica gel = 46 grams.
173
-------
PH2Q R Tstd
we 1 M P
c HzO std
1«°° (6.2383x10^)(294)
18.0 760
= 0.28 m3
where V is the total volume of water in the sampled gas
W t^
at standard conditions (21 °C, 760 mm Hg) ,
m3.
V., is the total volume of liquid H20 collected in
Q
impingers and on silica gel, 206, ml.
PH 20 is the density of water at standard conditions,
1.00 g/ml.
R is the ideal gas constant, 6.2383 x 10 mm Hg
cniVgm mole-°K.
T , is the standard temperature, 21°C, 294°K.
P ,j is the standard pressure, 760 mm Hg .
Mu ^ is the molecular weight of water, gm/gm-mole.
n U
B.4.3 Moisture Content of Stack Gas:
V
we
B
wo V -l- V
mstd wc
°"28 = 0.0924 or 9.24%
2.75 + .28
where B is the mole fraction or proportion by volume of
HZ0 vapor in the stack gas.
174
-------
V is .28 m3 H20 vapor at standard conditions.
V is 2.75 m3 of dry sampled gas at standard conditions,
std
B.4.4 Molecular WeightofStack Gas
From Orsat analysis per EPA Method 3:
C02 ave = 11.5%
O2 ave = 5.8% by volume (dry basis)
CO ave = 0%
From moisture analysis per EPA Method 5:
H20 avg = 9.24%
From EPA Method 3:
% N2 (dry basis) = 100% - (% COZ + 02 + % CO)
=100% - (17.3%) = 82.7%
Mw = (1~Bwo)(BC02MC02 + B02M02 + BCOMCO + W +
BwoMH20
= (1-.0924) [(0.115) (44.0) + (0.058) (32.0) + (0.0) (28.0) +
(0.827) (28.0)] -1- (0.0924) (18.0)
=28.96 gm/gm-mole
where M is the molecular weight of the stack gas,
w
gm/gm-mole, wet basis.
175
-------
B^ is the mole fraction of the component gas.
M. is the molecular weight of the component gas.
B is the mole fraction of the water vapor in the
stack gas.
B.4.5 Excess Combustion Air (per EPA Method 3)
= (%0z) - 0.5(%CO) „ inn
0.264(%N2)'- (%0i) + 0.5(%CO)
(5.8) - 0.5(0)
0.264(82.7) - (5.8) - 0.5(0)
= 36%
where %EA is the percent excess combustion air.
%02 is the percent 02 by volume on a dry basis
%N2 is the percent N 2 by volume on a dry ba.~..- •
%CO is the percent CO by volume on a dry bar;. • "
0.264 is the ratio of 02 to N2 in air by volume.
B.4.6 Particulate Emissions Concentration
C's
™
mstd
75 = 54.55 mg/DNM3
where C' is the concentration of particulate matter
in the stack gas, mg/DNM3.
M is the total particulate mass collected on
n c
filter media, 150.0 mg.
V is the volume of stack gas sampled (volume through
ITl . j
dry gas meter), DNM 3.
176
-------
B.4.7 Stack Gas Volumetric Flow Rate (per EPA Method 2]
(V ) = K C
avg P P
(/Ac)
avg
"avg~
L PsMw J
where (V ) is the average stack gas velocity, m/sec.
avg
K is 128.83™
-- -
sec\ gm-mole- K /
C is the pitot tube coefficient.
(T } , _ is the average stack gas temperature, °K.
S avy
(/Sp) is the average of the square roots of velo-
a vg
city heads of stack gas, in HzO, deter-
mined according to EPA Methods 1 and 2.
P is the absolute stack gas pressure, mm Hg,
5
Nl is the molecular weight of the stack gas,
gm/gm-mole (wet basis) .
From Table B2:
= 1.13
C = 0.860
(T )avg = 421.7 °K
Ps - Pbar + (Ps) ' 749
= 748.25 mm Hg
MW = 28.96 gm/gm-mole
177
-------
(V ) = (128.83) (0.860) (1.1.3)
3 avg
421.7
(748.25) (28.96)
(V ) = 17.47 m/sec
5 avg
Then calculating the average stack gas volumetric flow rate:
T P
- ^fi- =;t-rt
-JDUU ^ ^j. £5WOAVS) « /rri \ r.
where Q is the average volumetric flow rate at dry standard
S 3
conditions, DNM /h
B is the mole fraction of water vapor in the stack gas,
dimensionless, and
sample point, 9.29 m3 (from the Particulate Test
A is the cross-sectional area of the stack at the
o
sample point
Data Sheet) .
Then,
Qs = (3.600 x 103) (.9076) (17.47) (9.29)
= 3.64 x 10s DNM3/h
Therefore, the particulate inlet mass emission rate can be
calculated by PMR = Q C
S S S
where PMR is the average particulate inlet mass emission
o
rate, (mg/hr calculated by the concentration
method) .
DNM3
PMRg = 3.64 x 10= ~±- 54.55 mg/DNM3
= 1.986 x IQ7 mg/hr of particulate.
178
-------
B.4.8 Average Isokinetic Ratio
•avg (VJ
5 avg
where I is the average percent isokinetic.
(V ) is the average gas velocity into the nozzle
• - entrance. '
(V } is the average gas velocity of the stack gases.
s avg
Writing the1 equations--for the nozzle velocity independent
of the velocity pressure measured in the stack.
V • •. T P
me s m
- - T P D
(V ) = n x
n'avg A 9
where ^V^ ' '^s tne avera<3e. -nozzle velocity.. .-
V is -the; volume of gas through the dry;-gas .
•• meter,, • DNM3 .(dry-). .,.-•...:•; •..-..- . •/ . -, -. •
T is the average absolute stack gas; "temperature
°K. •,' ••-.. , • ...-.-.. - ... •
P is the average absolute stack pressure, mm Hg,
o
P is the average absolute dry gas meter pres-
sure, mm Hg
(Pm '
179
-------
T is the average absolute meter temperature,
°K
M, is the average mole fraction of the dry stack
gas, (1-BWO).
A is the area of the nozzle, m .
P is the total sampling time, sec.
Then,
(2.8) 421.7 752.69 1
294 748.21 .9076
n avg (6.22 x 10 5m2) (3600 sec)
= 19.88 m/sec
(V )avg iq ftft
avg (Vs)avg 17.47 A'" Wi
To conform with the federal performance standards, the
average isokinetic rate ~L~va' must be >0.90 and <1.10. In
Q vy
this example, since Iavq does not fall within these limits,
this test would be discarded.
B.4.9 Mass Collection Efficiency Calculation
This example has shown calculations for a test at a
precipitator inlet. Similar calculations are made on data
taken at ESP outlets. After determining the inlet and outlet
mass loading concentrations, the precipitator efficiency
can be calculated from
%EFF . - ^outlet „
PMKInlet
PMRInlet - Particulate inlet mass emission rate.
)utlet
PMRO_^,_J_= Particulate outlet mass emission rate,
180
-------
APPENDIX C
CASCADE IMPACTOR SAMPLING TECHNIQUES
Page
C . 1 THE PRESURVEY 183
C . 2 EQUIPMENT SELECTION 185
C.2.1 Impactor Selection
C.2.2 Sample Trains
C.2.3 Balance Requirements
C.3 IMPACTOR SUBSTRATES. 191
C.3.1 Collection Substrates
C.3.2 Back Up Filters
C.4 PREPARATION AND SAMPLING 194
C.4.1 Substrate Preparation
C.4.2 Impactor Orientation
C.4.3 Heating the Impactor
C.4.4 Probes
C.4.5 Nozzle and Sampling Rate Selection
C.4.6 Use of Precollector Cyclones
C.4.7 Impactor Flow Measurement
C.4.8 Sampling Time
C.4.9 Readying the Impactor
C.4.10 Pre-Sample Checks
C.4.11 Taking the Sample
C.4.12 Number of Sample Points
C. 5 SAMPLING RETRIEVAL AND WEIGHING 205
C.5.1 Impactor Clean-Up
C.5.2 Drying and Weighing Substrates
C.5.3 Data Logging
C.6 QUALITY ASSURANCE. 207
C.6.1 Impactor Techniques
C.6.2 Weighing Techniques
C.6.3 General Notes »
C.6.4 Data Analysis.
C.7 USE OF COMMERCIAL IMPACTORS 213
C.7.1 Brink BMS-11 Cascade Impactor
C.7.2 Andersen Mark III Stack Sampler
C.7.3 University of Washington Mark III Impactor
C.7.4 MRI Model 1502 Cascade Impactor
C.7.5 Sierra Model 226 Cascade Impactor
181
-------
APPENDIX C (CONT'D.)
Page
C.8 DATA ANALYSIS .220
C.8.1 General Discussion
C.3.2 Calculation of Impactor Stage Dso's
C.8.3 Cumulative Particle Size Distributions
C.8.4 Differential Size Distributions
C.8.5 Confidence Limits of Reduced Data
C.8.6 Cascade Impactor Data Reduction - Sample
Calculations
182
-------
APPENDIX C
CASCADE IMPACTOR SAMPLING TECHNIQUES
This appendix contains detailed guidelines for the opera-
tion of cascade impactors to measure the particle size distri-
butions at the inlet and outlet of electrostatic precipitators.
These instructions are taken from "PROCEDURES FOR CASCADE
IMPACTOR CALIBRATION AND OPERATION IN PROCESS STREAMS",
EPA-600/2-77-004, by D. B. Harris. Minor modifications were
made to make them specifically applicable to precipitator
evaluation, and a more detailed description of data reduc-
tion techniques and data presentation formats has been added.
C.I THE PRESURVEY
The key to performing a successful fractional efficiency
evaluation is thorough planning based on a complete pretest
site survey. The survey should provide adequate information
at as low a cost as possible. Some sites will require more
information and some less. As far as is possible, the in-
formation noted during the presurvey should be measured rather
than obtained from plant records or personnel.
As the presurvey is generally conducted by one or two
men "traveling lightly," the apparatus used during the pre-
survey should be as light and compact as possible. A pre-
survey sample train is shown in Figure Cl. This system can
be built into a single, suitcase-size package, and serves
well as a presurvey sample train. The impactor which is
to be used during the main test program should normally be
used during the presurvey. This is because the suitability
of substrates and adhesives must be checked out. These
problems are discussed more fully in later sections.
183
-------
DIFFERENTIAL PRESSURE
INDICATORS
0-2 IN. H2O 0-10 IM, Hg
STATIC-IMPACT
PITOT-TUBE
STACK ATMOSPHERE
STATIC PRESSURE
PRESSURE
COLLAPSIBLE
PITOT-TUBE
NEEDLE
VALVE
PUMP
MOISTURE
TRAP
COOLING
COIL
TEMPERATURE
W/CONTRQLLER
0-500° F
OUTPUT
TO HEATING
TAPE
IMPACTOR
TEMPERATURE
SENSOR
SIGNAL
HEATING
TAPE
( ,
Figure C1. Presurvey sampling with a cascade impactor.
184
-------
In general, the presurvey work should be done using
the techniques described in this guideline. Less precision
is required, but the accuracy must be high enough to provide
useful information in designing the test program. The de-
cisions which must be made are summarized in Table Cl.
C.2 EQUIPMENT SELECTION
C.2.1 Impactor Selection
The selection of the proper impactor for a particular
test situation is primarily dependent upon the mass load-
ing of the gas stream and its effect on sampling time. There
are three major criteria to be met to match an impactor to
a particulate stream.
1. The sampling period must be long enough to provide
a reasonable averaging of any short term transient
in the stack.
2. The loading on a given impactor stage must be low
enough to prevent re-entrainment.
3. The sampling rate through the impactor must be low
enough to prevent scouring of impacted particles
by high gas velocities.
For these reasons, an impacto* with a comparatively low
sample rate must be used in a gas stream with a high mass
loading. The low sample rate allows a longer sampling time,
although in some situations it will still be undesirably
short. Conversely, in a low mass loading situation such
as a control device outlet, a high sample rate device must
be used if a significant amount of sample is to be gathered
in a reasonable amount of time.
185
-------
TABLE Cl. IMPACTOR DECISION MAKING
Item
Information Requ ired
Criteria
Itnpactor
Sampling rate
oo
Nozzle
Pre-cutter
Sampling time
Collection substrates
Number of sample points
Loading and size estimate
Loading and gas velocity
Gas velocity
SJze and loading
Loading and flow rate
Temperature and gas
composition
Velocity distribution
and duct configuration
a. If particle concentration below 5.0 um is less
than 0.46 gm/am3 (0.2 grain/acf), use high flow
rate impactor,
b. If particle concentration below 5.0 um is greater
than 0.46 gm/am3 (0.2 grain/acf), use low flow
rate impactor.
a. Fixed, near isokinetic
h. Limit so last jet velocity does not exceed:
60 m/sec greased
35 m/sec without grease
a. Near isokinetic, ±10%
b. Sharp edged; min T.4 mm ID
If pre-cutte>- loading is comparable to first stage
loading--use pre-cutter.
a. Per Figure C4
b. No stage loading greater than 10 mg
a. Use metallic foil or fiber substrates whenever
possible
b. Use adhesive coatings whenever possible
a. At least two points per station
b. At least two samples per point
-------
A cascade impactor can normally yield useful information
over a range of sample rates differing by a factor of 2 or
3. As high efficiency control devices cause the outlet
mass loading to differ from the inlet by a factor of 103 ,
the same impactor can seldom be used on both inlet and out-
let. Both high and low flow rate impactors are usually re-
quired to determine the fractional efficiency of electro-
static precipitators. Some commercial impactors are constructed
such that their stage and nozzle configurations can be altered,
and they can serve as either high or low sample rate impac-
tors. Others are fixed with respect to sample rate.
C.2.2 SampleTrains
Figure C2 is a flow diagram of a typical impactor sample
train. As shown, it is desirable to have the impactor inside
the stack with a straight nozzle. The various parts of the
sample train are discussed below.
A sampling probe leading to an impactor outside of the
duct should be used only if absolutely necessary. The probe
should be as short as possible and contain the fewest possible
bends. It is recommended that a pre-cutter cyclone be mounted
at the probe inlet to remove particles larger than approximately
ten micrometers and thus reduce line losses.
Heating system --- The criteria for heating are given
in Section C.4.3. If heating is required, the entire impac-
tor must be either wrapped in a heating tape or put in a
custom-fitted heating mantle. The temperature control should
be based on the temperature at the outlet end of the impactor.
Often the temperature is measured between the last stage
and back-up filter. The impactor temperature can be con--
trolled either manually or automatically. An automatic con-
187
-------
HEATING
JACKET
TEMPERATURE
CONTROLLER
INPUT T
OUTPUT
TO JACKET
HEAT
EXCHANGER
ICE BATH
BLEED
CALIBRATED
\
ORIFICE *-\
.1, N/
r
MANOM
ETER
VENT
MANOMETER
u
DRY GAS
METER
VACUUM
PUMP
LEGEND
@ - PRESSURE MEASUREMENT POINT
CO- TEMPERATURE MEASUREMENT POINT
3630 081
Figure C2. Typical sample train with a heated impactor.
188
-------
troller has usually been found to be worth the money by
releasing the operator for other tasks.
Flue gas conditioning — It is usually necessary to
cool and dry the flue gas before it reaches the flow measur-
ing section. Condensation in the orifice would distort the
measurement; also, it is useful to protect the equipment
from the condensate, which, in S02-containing gases, is
likely to be sulfuric acid. The type of condensers shown
are usually satisfactory. Packed bed drying columns are
commercially available. The heat exchange coil is used to
bring the gas temperature to essentially ambient so that
there will not be a significant temperature gradient across
the flow measuring devices.
Flow measurements -- At least two flow measuring devices
are used in series. Normally, a calibrated orifice is used
in conjunction with a dry gas meter, as shown. At very low
sample rates the dry gas meter may be inaccurate. The com-
monly used diaphragm-type positive displacement gas meter
becomes increasingly inaccurate at flow rates less than five
percent of rated capacity. For a typical stack sampling
gas meter this would be approximately 23.3 cm3/sec (0.05
cfm). Another calibrated orifice or a rotameter should then
be used as the second flow meter.
Vacuum pumps — The vacuum pump should usually be placed
at the end of the sample train. This is because vacuum pumps
tend to leak and all of the flow measurements must be made
upstream of any leak. The flow rate can be controlled by
using an inlet side air bleed or with a recirculating bypass.
189
-------
Pressure measurements -- Most of the pressure measurements
are made with manometers, but calibrated differential pres-
sure meters are equally acceptable. The in-stack pressure
needed is the static pressure, which is not exactly the down-
stream pressure of an S-type pitot tube. A true static pres-
sure measurement should be made. It is not necessary that
this be part of the impactor train, but it can be.
The pressure at the downstream end of the impactor,
between the last stage and the final filter, must be known.
It can be measured, but this is often inconvenient. If a
flow rate pressure drop calibration is available for the
impactor (without final filter), it is normally acceptable
to calculate the pressure drop. Correction must be made
for pressure and temperature differences between the cali-
bration conditions and the actual conditions.
The pressure at the inlet to the metering devices must
be known. In the system shown in Figure C2, the pressure
is metered ahead of the calibrated orifice and the orifice
pressure drop is used to calculate the pressure going into
the dry gas meter. The dry gas meter pressure should be
measured if there is a reason to think the procedure above
was not adequate.
Temperature measurements — It is necessary to know
the temperature at all points where the pressure is measured.
Any convenient device of known accuracy can be used to make
the measurements. The measurement in-stack can easily be
made at the probe end with a thermocouple. The temperature
at the downstream end of the impactor is made directly be-
hind the final filter and is used to control the heating
tape if one is used. If the heat exchanger in the train
190
-------
brings the gas temperature to about ambient, only one tem-
perature reading will be necessary at the flow meters. This
is usually most conveniently done at the dry gas meter, as
taps are available on the meter.
C.2.3 Balance Requirements
For the accurate weighing of collected material a balance
with a sensitivity of at least 0.05 mg is required. This
is especially true for the lower stages of the low sample
rate impactors where collection of 0.3 mg or less is not
uncommon. The balance must also be insensitive to vibration
if it is to be used in the field. It is also desirable to
have a balance with a large weighing chamber. These capa-
bilities are available in several electrobalances marketed
in the United States.
More information on selected sampling train components
can be found in Appendix B, Section B.I.
C.3 IMPACTOR SUBSTRATES
C.3.1 Collection Substrates
For reasons which have been discussed, very accurate
determinations of impactor stage catch weights are necessary.
Impactor stages are generally too heavy for the tare capacity
of field-usable precision balances. Thus, a substrate which
can be weighed on the balance is used. Generally, these
substrates are made of metal foil or glass fiber.
Glass fiber substrates -- Glass fiber substrates are
used on some commercial impactors. In addition to providing
191
-------
a lightweight impaction surface, glass fiber mats greatly
reduce re-entrainment due to particle bounce. They are su-
perior to greased metal substrates in very high temperature
applications where the greases tend to evaporate. Care must
be taken when using glass fiber substrates in streams con-
taining sulfur dioxide, however. Recent experimentation1
has shown that glass fiber materials often exhibit anomalous
weight gains due to sulfate on the substrates. Apparently,
sulfur dioxide in a gas stream can react with basic sites
on most glass fiber materials and form sulfates.
There are two approaches to alleviate this problem.
Substrates which will gain weight from sulfate uptake can
be preconditioned in the flue gas before weighing. Two to
six hours of exposure to the flue gas will suffice where
mass loadings are high and sample times are short. In the
situations where sample times are long and the collected
amount of particulate matter is small, it may be necessary
to condition the substrates for as long as twenty-four hours
to eliminate significant sulfate uptake and weight gains.
Repeated weighings to check weight gains are necessary to
confirm that the substrates can be used. Another approach
is to use a fibrous substrate which shows little weight gain
in a sulfur dioxide stream, if one can be found. It should
be noted that the particle retention characteristics of dif-
ferent fiber materials vary, and the impactor calibration
could change significantly if the substrate is changed.
Greased substrates -- "Grease" must often be used on
metal foil substrates to improve their particle retention
characteristics. This is particularly important with hard,
bouncy particles. Impactor stage velocities of 60-65
m/sec have been used on greased substrates with good results,
while particle bounce can become a problem at about half
of that rate on ungreased substrates.
192
-------
Finding a suitable grease can be difficult. The grease
should not flow at operating temperature,- and it must be
essentially nonvolatile. Gas chromatographic materials such
as polyethylene glycol 600 have exhibited more consistent
characteristics than materials such as stopcock grease.
Another class of materials which may be suitable are high
vacuum greases; Apiezon L and H in particular have performed
well at temperatures up to 120°C. The greased substrates
must be tested as blanks in filtered process gas before they
are used in the test program.
The greases are normally applied as suspensions or solu-
tions to 10-20 percent grease in toluene or benzene. The
mixture is placed on the substrate with a brush or eyedropper,
baked at 204°C for 1 to 2 hours, and then desiccated for
12 to 24 hours prior to weighing. It is important to avoid
an excess of grease. The desiccated, greased substrate
should be tacky, but not slippery, with a film thickness
about equal to the diameter of the particles which are to
be captured.
Horizontal operation of the impactors with greased sub-
strates is not recommended due to possible flow of the grease.
Care must also be taken to ensure that grease is not blown
off the substrates (which tends to occur at jet velocities
greater than 60 m/sec). To some degree, grease blow-off
can be avoided by using a light coating of grease on the
last stages. This is normally satisfactory from an adhesive
standpoint, as the last stages usually have the lightest load-
ing along with the highest jet velocity. Inspection of the
stage catches is the best way to check on this problem.
193
-------
C.3.2 Back up Filters
Back up filters are used on all impactors to collect
the material that passes the last impaction stage. Binder-
less glass fiber filter material is normally used for this
purpose in all the impactors, although the exact configura-
tion varies.
Glass fiber back up filters have the same problems as
do glass fiber substrates. Their use in process gases con-
taining sulfur dioxide is suspect, and blanks must be run
to check out the problems. Pure Teflon filters may alleviate
this problem if they can be used.
C.4 PREPARATION AND SAMPLING
C.4.1 Substrate Preparation
It is assumed that the substrates have been properly
prepared and that the necessary quality assurance steps have
been taken. The substrates should be carefully weighed and
kept in a desiccator until they are to be placed in the im-
pactor.
C.4.2 Impactor Orientation
Whenever possible, the impactor should be oriented ver-
tically to minimize gravitational effects such as flow of
grease or fall-off of collected particles. Sampling situa-
tions requiring horizontal placement will occur, and extra
care must be taken on such occasions not to bump the impac-
tor against the port during entry or removal.
194
-------
C.4.3 Heating the Impactor
All condensable vapors must be in a gaseous state until
they exit from the impactor, unless a condensate is the prime
aerosol being measured. In gas streams above 177°C, auxill-
iary heating should not be required. Below 177°C the exit
temperature of the impactor should be maintained at least
10°C above the process temperature if condensable vapors
are present. A thermocouple feedback temperature controller
has proven useful.
When condensable vapors are present, it is sometimes
necessary to heat the impactor probe to prevent any conden-
sate formed in the probe from entering the impactor and con-
taminating the substrates. Water vapor is the primary problem.
The probe temperature should be maintained above the vapor's
dewpoint.
Whether the impactor is being heated in the duct or
externally with heater tape, an allowance of 45 minutes warmun
time is recommended as a minimum to ensure that the imr/jctor
has been heated to duct or operating temperature. Thermocouple
monitoring of the impactor temperature and gas temperature
is recommended.
C. 4.4 Probes
Sampling probes leading to an impactor outside of the
duct should be used only if there is no other way. They
should be as short as possible and contain the fewest possible
bends. It is recommended that a precutter be mounted at
195
-------
the duct end of the probe to remove the large (>10 \im) par-
ticles and thus reduce line losses.
C.4.5 Nozzle and __s amp ling Ra t e Selection
It is preferable to use as large a nozzle diameter as
possible to minimize sampling errors resulting from nozzle
inlet geometry. When very small nozzles have been used
with the Brink impactor, there have been some cases in which
large amounts of material were retained in the nozzle or
the nozzle was completely blocked. It is recommended that
the inlet nozzle not be smaller than 1.4 mm, and some types
of particulate material may require a larger minimum nozzle
size. In some instances bent nozzles are necessary due to
port location and gas direction, but these should be avoided.
Problems occur in cleaning bent nozzles, and it is diffi-
cult to determine the size interval in which the deposited
material originated. If they cannot be avoided, bends should
be as smooth as possible and of minimum angle in order to mini-
mize the losses in the fine particle region.
For hard, "bouncy", particulate, the sampling rate must
be such that the last stage velocity does not exceed 60 m/sec
for greased collection surfaces or 35 m/sec for nngreased
plates if no suitable substrate can be found to limit particle
bounce. The flow rates above should not be considered the
final word on nozzle velocity. Particle bounce has been
observed at nozzle velocities as low as 10 m/sec. Some par-
ticulate materials are "sticky" and will adhere at well above
the maximum velocity for hard particles. The exposed substrate
should be visually examined for evidence of re-entrainment
and the rates adjusted accordingly.
196
-------
It is apparent that sample rate and nozzle size are
closely coupled. The requirements for isokinetic or near -
isokinetic nozzle flow sometimes impose a compromise on nozzle
selection. The general order of priorities when choosing
the sample rate is nozzle diameter fat least 1.4 mm). last
stage jet velocity and flow rate required for Jsokinetic
samplina. Selection of nozzle diameter and imnactor flow
rate combinations for achieving near-isokinetic sampling
conditions can be made from Figure C3.
If a choice must be made between undersized and over-
sized nozzles, undersized nozzles will usually result in
lower sampling errors than will oversize.
C.4.6 Use of Precollector Cyclones
In many instances the percentage (by weight) of material
with sizes larger than the first impaction stage cut point
is quite high. In such cases a precollector cyclone is neces-
sary to prevent the upper impactor stages from overloading.
A precutter should always be used for the first test. If the
weight of material obtained by the precutter is greater than
or equal to that on the first stage, the precollector should
be used in all subsequent runs. Cyclones can be obtained from
the impactor manufacturer or can be shop made. The use of
two first stages in series has also been suggested and appears
to be a valid approach; however, no data are available.
C.4.7 Impactor Flow Rate Measurement
The flow rate through an impactor must be accurately
measured in order to set the isokinetic sampling rate and
to determine the correct impactor stage cut points. Unfor-
tunately, it is usually very inconvenient, and sometimes
197
-------
l/min acfm
GAS VELOCITY
4 6 8 100
3050
3630 080
Figure C3, Nomograph for selecting nozzles for isokinetic sampling.
196
-------
impossible, to measure the impactor flow rate at the condi-
tions present in the irapactor . The gas is normally drier,
cooler, and at a lower pressure by the time the flow rate
is measured; and the flow must be corrected to impactor con- .
ditions. The use of calibrated orifices and dry gas meters
is discussed below.
Units — The equations presented are valid only if the
units of the various terms are consistent. For instance,
the pressure drop terms could be in units of mm H20 or cm
H2O or something else, but all pressure drop terms must have
the same units. The same is true for the other properties.
Note that pressure and temperature are both absolute measure-
ments.
Orifice meters -- The gas flow rate through a particular
orifice meter is related to the pressure drop across that
orifice by an equation of the form:
Q2 = C ^ /CD
where Q = volumetric flow rate at upstream conditions,
AP = pressure drop across orifice,
p = density of gas at upstream conditions, and
C = dimensional constant, (length) 5 (mass) (time) "2 (force)
Solving for the constant, C, in equation (Cl) , one ob-
tains:
(C2)
As C is a constant at all conditions, its value can be
obtained at a convenient set of conditions with a known flow
rate and used later to calculate the flow rate. Equation (C2)
199
-------
can" be rewritten to:
c c
C = -^r^ (C2a)
c
The subscript "c" indicates that these parameters were
determined during a calibration. Density and flow rate are
at upstream conditions.
Substituting equation (C2a) into equation (Cl) yields
an equation suitable for obtaining flow rates from a cali-
brated orifice:
Qc2 PC
APc
AP
m
(C3)
denotes the parameters of the gas
as it is being "measured". All are at conditions immediately
upstream of the orifice.
For use with impactors, the measured flow rate, Q ,
must be converted to a flow rate at stack conditions, Q .
s
Assuming that the stack gas was dried as well as altered
in temperature and pressure, the stack flow rate is related
to the measured flow rate by:
V1 - BWO>T^ - Qm r1 (C4)
s m
where B = water removed from flue gas, expressed as a volu-
metric fraction,
P = absolute pressure, and
T = absolute temperature.
The subscript
200
-------
At the usual conditions of relatively high temperature
and low pressure which occur during stack sampling, the flue
gas behaves very much like an ideal gas. The density of
an ideal gas can be approximated as:
a P(MW)I_ 5i
P (R) (T) '
where R = the universal gas constant, and
MW = the molecular weight of the gas.
Equations (C3), fC4), and (C5) can be combined and re-
arranged into a form which gives the pressure drop which must
exist across the calibrated orifice, AP , to obtain the re-
quired impactor flow rate, Q .
where (MVJ) = molecular weight of the stack gas at the orifice,
normally the dry molecular weight,
(MW) = molecular weight of the calibration gas.
Dry gas meter -- The dry gas meter, like the orifice,
can only directly measure the flow rate of the gas which
passes through it. This measured flow rate can be converted
to the flow rate through the impactor (which is at stack
conditions) using equation (C4a) :
W
-------
C.4.8 Sampling Time
The length of the sampling time is dictated by mass
loading and size distribution. An estimate for initial tests
can be obtained from Figure C4. Two conflicting criteria
complicate the choice of the sampling time. It is desirable
from the standpoint of minimizing weighing errors to collect
several milligrams on each stage. However, most size dis-
tributions are such that the upper stages are overloaded
and are re-entraining particles by the time the lower stages
reach a few milligrams. A rule of thumb is that no stage
should be loaded above 10 mg, but the determining factor
is whether or not re-entrainment occurs. As is discussed
later, a comparison of the relative distribution determined
by a long run with that from a shorter (about half as long)
run can be used to check on re-entrainment due to stage over-
loading .
C.4.9 Readying the Impactor
As equipment is not always cleaned as well as it should
be, the impactor should be inspected prior to use. The nozzles
must be clean, gaskets in good shape, and the interior clean.
Nozzles can be cleaned with fine wire if necessary.
After inspection, the impactor should be carefully loaded
with the preweighed stage substrates and assembled. Teflon
thread sealant tape or antiseize compound should be applied
to the threads, especially when high temperatures (>215°C)
are encountered. The thread sealant tape generally works
better and causes fewer problems but probably cannot be
successfully used at temperatures above 290°C.
If supplemental heating is required, a heating device
and temperature monitor need to be added. A thermocouple
202
-------
0.01
1000
FLUE GAS MASS LOADING
1 mg/t ,
0.023
0.23
0.01
0.1
gr/acf
2.28
22.8
228
—I—
1.0
10
100
I II
T
TTT
IT
.READ DOWN FROM MASS LOADIMG TO IMPACTOR
SAMPLE RATE. READ LEFT TO TIME REQUIRED -
TO COLLECT A 50 MG SAMPLE AT THAT SAMPLE —
RATE.
^W f* A n A rt i i— '
= LOADING x RATE x TIME
acfm 2.0 1.0 0.6 0.2 0.1 0,04 0.02 0.01
fe/min 28.3 2.83 0.283
IMPACTOR SAMPLE RATE
3630-079
Figure C4. Nomograph for sampling time selection (50 mg sample).
203
-------
mounted in the gas flow immediately after the impactor
is best for controlling impactor cut points.
The supplemental heat can be supplied with either a
heating mantle which has been made to fit the impactor or
by using heating tapes. If the tapes are to be used, a heat-
ing tape of sufficient wattage is wrapped around the impac-
tor. Glass fiber tape works well for holding the heating
tape. Insulation such as asbestos tape is then wound around
the impactor. Glass fiber tape again is used to hold the
asbestos in place and also acts as additional insulation.
The impactor can now be mounted on the appropriate probe,
taken to the sampling position, and installed in the sam-
pling system.
C.4.10 Pre-Sample Checks
Impactors are prone to leak, and they must be checked
at operating temperature for leaks. This can be done in
several ways. The nozzle can be plugged and the impactor
pressure-tested or vacuum-tested. Because impactors are
basically a series of orifices, they should have a constant
flow to pressure drop relationship. Checking the pressure
drop on various flows of filtered air will point out devia-
tions from normal operations—both leaks (external or inter-
nal) and plugged jets.
C.4.11 Taking the Sample
The impactor should be preheated for at least 45 minutes
before sampling. If supplemental heat is being used, the
impactor should be brought up to temperature outside the
duct and then allowed some time to equilibrate after insertion,
204
-------
The nozzle should not point into the flow field during this
phase. Without supplemental heat, the whole warm-up is con-
ducted within the duct, again with the nozzle pointed away
from the flow field.
A predetermined flow rate must be maintained to ensure
stable cut points. Any attempt to modulate flow to provide
isokinetic sampling will destroy the utility of the data I'/-.
changing the cut points of the individual stages. Rapid
establishment of the correct flow rate is especially impor-
tant for the short sampling times typically found at the
inlets to control devices. Capping the nozzle during preheat
in the flue is also desirable.
C.4.12 Number of Sample Points
As the velocity and particulate distributions in indus-
trial ductwork are unlikely to be ideal, a large number of
samples are often required for accurate particulate measure
ments. A velocity traverse should be run to check on the
velocity distribution. At least two points within a duct
should be sampled in each measurement plane, and at least
two samples taken at each of these points. These are the
minimum sampling efforts and are appropriate only for loca-
tions with well developed flow profiles in the absence of
significant concentration stratification. If the flow pro-
file at the inlet or outlet is uncertain due to duct con-
figuration and/or the mass loading is not uniform, the number
of samples may need to be increased for reliable results,
C.5 SAMPLE RETRIEVAL AND WEIGHING
C.5.1 Impactor Clean-up
The careful disassembly of the impactor and removal
of the collected particulate are essential to the success
205
-------
of the test program. The crucial points are to make sure
that the collected material stays where it originally im-
pacted and to remove all the particulate. After the sam-
pling run, the impactor should be carefully removed from
the duct without jarring it, removed from the probe, and
allowed to cool. Disassembly can be difficult in some cases,
particularly if the impactor was used at elevated tempera-
tures .
Typically, not all of the particulate which collects
in an impactor collects on the substrates. Some accumulates
on the interior surfaces, especially in the nozzle. By con-
vention, all of the particles collected upstream of a given
impaction stage are assigned to that stage.
The collection of these "misdirected" particles is
often troublesome. If the particles are hard and dry, they
can be brushed off into the weighing container. A No. 7
Portrait brush or its equivalent is suggested, and care must
be taken to prevent brush hairs from contaminating the sample
If the particles are sticky or wet, some type of washdown
procedure should be used. The solvent must be considerably
more volatile than the particulate matter.
C.5.2 Drying and Weighing
All of the samples must be dried to constant weight,
with 2 hour checks used to establish the uniformity of the
weights. Hard, nonvolatile particles are often dried in
a convection oven to 100°C, desiccated until cooled to room
temperature, weighed, then check weighed. Volatile mater-
ials will require some other technique using low temperature.
Whatever the technique used, constant weight of the sample
with further drying is the criteria to be met.
206
-------
C.5.3 DataLogging
Permanent records should be kept of all pertinent infor-
mation. It is generally necessary to keep records in three
places—in the lab with the balance using a bound notebook,
and using either looseleaf data forms of a bound notebook
at both the inlet and outlet of the control device. Table
C2 presents a fairly complete listing of the information
required concerning an impactor run. In addition, records
of the weighing of the catches must be kept. Notes should
be taken on any abnormalities which occur and on the ap-
parent condition of the stage catches.
C.6 QUALITY ASSURANCE
The field use of cascade impactors is a difficult task.
The accuracy required is more appropriate for a laboratory
program than for a field test. There are many places in
the operational sequence where errors can occur in spite of
a conscientious effort to do a good job. Quality assurance
attempts to discover inaccuracies before they are propagated
throughout the test program. The techniques presented in
this section are not the only ways to ensure quality data.
However, they have been used successfully in field testing
with impactors.
C.6.1 Impactor Techniques
Glass fiber substrates — As has been discussed previously,
glass fiber substrates are not without problems. Two poten-
tially serious problems are 862 uptake on the substrate and
mechanical or manual abrasion of the filter mat.
207
-------
TABLE C2. SAMPLING INFORMATION REQUIRED
Date
Time
Run Code Number
Impactor Type and Identification Number
Operator
Port Number/Sampling Location
Ambient Temperature
Ambient Pressure
Impactor In-Stack or Out-of-Stack
Impactor Orientation
Number of Traverse Points
Stack Pressure
Stack Temperature
Nozzle Diameter/Type
Probe Depth, if used
Stack Pitot Tube Delta P/Stack Gas Velocity
Desired Impactor Flow rate for Isokinetic Sampling
Metering Orifice Identification Number
Metering Orifice Delta P
Impactor Temperature
Scalping Cyclone in Use? Identification
Prefilter Identification
Postfilter Identification
Substrate Set Identification
Pressure Drop Across Impactor
Test Start/End Time: Duration of Test
Gas Meter Start/End Readings? Gas Meter Volume
Agreement Between Meter and Orifice
Volume of Condensible H2O in Flue Gas
Gas Meter Temperature
208
-------
The problem of S02 uptake on the substrate is dis-
cussed by Felix et. a 1. l The two approaches available are
to use a substrate which does not change weight in the flue
gas or to precondition the substrate in filtered flue gas
prior to weighing. Using a new glass fiber material, which
does not react with the S02/ may alter the particle reten-
tion characteristics of the impactor and change the impac--
tor's calibration. This must be checked and the data re-
ported. The use of preconditioned filter mats requires that
the glass fiber substrates be preconditioned long enough
to reduce the weight change during the expected duration
of the impactor runs to 10 percent or less of the minimum
stage weight. At the present time, this SOz reaction phe-
nomenon is not well understood, and only rough guidelines
are available. For some common glass fiber materials tested,
the saturation times were on the order of 2 to 6 hours at
the temperatures tested.
The applicability of the method chosen to overcome this
substrate problem must be tested during the presurvey and
periodically during the test runs by running blanks.
Glass fiber substrates must be handled carefully to
prevent damage and possible loss of fibers. Loose surface
fibers should be removed by shaking prior to initial weigh-
ing. After weighing, every precaution must be taken to pre-
vent the loss of any part of the substrate. One approach
which will quantify the problem of substrate abrasion is
to prepare a substrate set, load the impactor, then dis-
assemble and reweigh.
Greased metal substrates — The problems which occur with
the use of greased substrates are usually related to the
properties of the grease. A grease which has been applied
209
-------
too heavily or has a low viscosity at operating temperature
can be physically blown off the impactor stage. The grease
could also react chemically with the flue gas or be exces-
sively volatile at the operating temperature. Again, these
phenomenon must be checked during the presurvey and period-
ically during the test program.
Re-entr_a_ininent -- Re-entrainment is the phenomenon of
an impacted particle being blown off the stage on which it
was collected initially and being recollected downstream.
This can be caused by excessive jet velocities or by over-
loaded stages. The effect of re-entrainment can be serious,
because only a few large particles on a small particle stage
will considerably affect the size distribution.
One way to spot re-entrainment is to very carefully
examine the stage catches. If, for example, a low velocity
through the jets resulted in a well-defined pile of particu-
late and a high velocity sample gave a diffuse deposit,
re-entrainment should be suspected at the high sampling rate,
Microscopic examination of the lower stages and final filter
for large particles (which should have been collected up-
stream) is another way to check for re-entrainment.
Re-entrainment due to stage overloading can be detected
by running two otherwise identical tests for two different
test durations. If the two size distributions are not the
same, overloading should be suspected at the higher stage
loadings.
Impactor leaks -- Two types of leaks can occur with
impactors--internal or external. A flow rate versus pres-
sure drop check of a pressure test will pick up most leaks.
An internal leak, where part of the airstream is bypassing
210
-------
the proper flow path, will give results similar to re-en--
trainment. Leak checks must be made at operating tempera--
tur e.
General procedure — A general procedure for impactor
use, concentrating on quality assurance, has been outlined
below:
1. Prepare Impactor
a. Wash impactor. Use ultrasonic cleaner if avail-
able.
b. Visually check cleanliness. Jets must be clear,
sidewalls clean. Must be done in good lighting.
c. Obtain preweighed substrates and assemble impac-
tor.
2. Sampling
a. Assemble impactor train and heat to operating
temperature.
b. Leak check the impactor.
c. Sample with impactor.
d. Disassemble impactor, examine stage catches
and impactor walls. Note any anomalies.
3. Substrate and Re-entrainment Checks
a. Check during presurvey.
b. Check substrates if flue gas composition
changes significantly.
211
-------
C.6.2 Wei_ghing Techniques
Preci_s ion an_d calibration -- The manufacturer's direc-
tions should be followed when operating the balance. The
balance should be calibrated at least once a day. The re-
peatability of measurements should be checked by repeatedly
weighing a substrate and a test weight.
Technique -- The assembly and disassembly of an impactor
should have no effect on the substrate weights. This should
be checked by weighing up a set of substrates, assembling
them in an impactor, then disassembling and reweighing.
Any weight losses from this process should be within the
repeatability of the balance (approximately 0.02 milligram
for an electrobalance). Dry weight checks are made by
desiccating the substrate, weighing, then desiccating again
and reweighing. When the agreement is within the repeatability
of the balance, dry weight has been achieved.
C.6.3 G e n era1 Notes
Spare parts — The well equipped sampling team wj11
travel with an adequate supply of spares. Improvisation
due to an equipment failure can lead to poor quality data.
Flow meters — At least two flow meters should be used
in series. If they do not agree, the problem should be in-
vestigated.
Pumrjs — Typically, vacuum pumps in sampling trains
leak. For this reason, the flow meters should be upstream
of the pump.
212
-------
C.6.4 DateAnalysis
Final filter data -- The fine particulate information
obtained from the final filter can sometimes be misleading.
It is assumed for analysis that a stage captures everything
larger than its D , and captures nothing smaller. A real
stage misses some large particles. Under some conditions
(including but not limited to re-entrainment), large particles
will penetrate to the final filter. In this case the size
distribution will be skewed towards the small particles.
Microscopic examination of the final filter may provide an
indication of this problem. If it occurs, the best choice
in data analysis is probably to ignore the final filter on
runs where this phenomenon was encountered.
Cumulative size dataanalysis — If either a probe or
a precutter cyclone is used with an impactor, the resultant
probe losses and precutter catches must be included in cumu-
lative size analysis. Failure to do so will lead to an in-
correct cumulative distribution.
Inspection of data — After the data have been collected,
they should be examined for any inconsistencies and outliers
should be rejected before the final averaging is done.
C.7 USE OF COMMERCIAL IMPACTORS
C.7.1 Brin k BMS-11 Cascade Impactor
Additional information on the following impactors can
be found in Section 2.3.2 under Table I.
The Brink impactor is a five-stage, low sample rate,
cascade impactor, suitable for measurements in high mass
loading situations. The Brink uses a single round jet on
each of its stages.
213
-------
Sampling rate -- The usual sampling rates for the Brink
are in the range of 9-33 cm3/sec (0.02 to 0.07 acfm). The
sampling rate must be low enough to prevent re-entrainment
of particles from the lower stages. With hard, bouncy par-
ticulate, the last stage nozzle velocity must be less than
30-35 m/sec with ungreased substrates, and less than 65 in/sec
with greased substrates.
Collection substrates and adhesives — The Brink impactor
collection stage is too heavy to use without some type of
substrate insert. Foil cups are commonly preformed and
fitted into the collection cups of the Brink stages. If
grease is to be used, the top stages require about 5 or 6
drops of solution while the bottom stages normally require
only about one drop in the center of the cup. Glass fiber
substrates cut to fit the collection cups have also been
found satisfactory in many situations,
Backup filter — The Brink back up filter is normally
made of binderless glass fiber filter material. Two 2.5 cm
diameter disks of filter material are placed under the spring
in the last stage of the impactor. The filter is protected
by a Teflon 0-ring and the second filter disk acts as a
support.
Precutter cyclone -- A precutter cyclone for the Brink
is not presently commercially available.
Sampling train — The Brink uses the usual type of sam-
pling train. Orifices on the order of 0.77, 1.52, and 2.29
cm in diameter allow full coverage of its range of sampling
rates at reasonable pressure drops.
214
-------
Brink clean-up — Careful disassembly of a Brink impac-
tor is necessary for obtaining good stage weights. If a
precollector cyclone has been used, all material from the
nozzle to the outlet of the cyclone is included with the
cyclone catch. All of this material should be brushed onto
a small, tared, 2.5 x 2.5 cm aluminum foil square to be saved
for weighing. Cleaning the nozzle is also important, espe-
cially if it is a small bore nozzle-.- ---All material between
the cyclone outlet and the second stage nozzle is included
with material collected on the first collection substrate.
All appropriate walls should be brushed off, as well as around
the underside of the nozzle, where as much as 30 percent
of the sample has been found.
C.I.2 Andersen Mark III Stack Sampler
The Andersen impactor is a relatively high sample rate
impactor. The normal sample rate is about 236 cm3/sec (0.5
acfm). The Andersen is a multiple jet, round hole impactor.
Sampling rate — The nominal Andersen sampling rate
is given above. As with other cascade impactors, the flow
rate must be low enough to prevent re-entrainment of impac-
ted dust.
Collection substrates and adhesives — Andersen sub-
strates are obtained precut from the manufacturer.. The sub-
strates are glass fiber and of two types—one cut for the
odd numbered stages, one for the even. As discussed earlier,
normal Andersen substrates have a tendency to absorb SOZ
on basic sites in the substrate and therefore gain weight.
215
-------
The Andersen requires careful assembly, as over tighten-
ing will cause the substrates to stick to the metal separator
rings.
Back up filter -- The Andersen uses a 63.5 mm diameter
disk placed above the final F-stage. (This F-stage is an
option not normally included with the standard stack head.)
The filter should be cut from binderless glass fiber filter
material, such as Reeve-Angel 934AH filter paper.
Precutter cyclone — A precutter cyclone for the Andersen
is available from the manufacturer. It is necessary to have
a 6-inch diameter or larger sampling port when using the
precutter cyclone with its nozzle.
Andersen sampling train — The Andersen requires the
usual type of sample train. The pumping and metering systems
of the commercial EPA Method 5 mass sampling trains are ap-
propriately sized for use with the Andersen.
Care should be exercised never to allow a gas flow re-
versal to occur through the impactor, Material could be
blown off the collection substrate onto the underside of
the jet plate or the collection substrates could be disturbed.
A check valve or maintenance of a yery low flow while re-
moving the impactor from the duct avoids this problem.
Andersen clean-up — Cleaning an Andersen impactor is
difficult. Foils should be cut to hold the substrates, and
each foil and substrate weighed together before and after
the run. For disassembly, the foil to hold the stage 1 sub-
strate should be laid out. Next the nozzle and entrance
cone should be brushed out and onto the foil. Then the
216
-------
material on stage 0 should be brushed onto the foil. The
stage 1 filter substrate material should then be placed on
the foil and, finally, the top of the stage 1 plate, 0-ring,
and cross piece should be brushed off. Depending on how
tightly the impactor was assembled, some filter material
may stick to the 0-ring edge contacting the substrate. This
should be carefully brushed onto the appropriate foil. This
process is continued through the lower stages. Finally,
the after filter is carefully removed.
C.7.3 University of Washington Mack III jPilat)_Impactor
The Mark III impactor is a seven-stage, high flow rate
device with generally the same characteristics as the Ander-
sen. The Mark III is a round hole, multiple jet imapctor.
Sample rate — The Mark III sampling rate is on the
order of 236 cm3/sec (0.5 acfm). The flow rate must be low
enough to keep scouring of impacted particles to a minimum.
Collection substrates and adhesives — The Mark III
has often been used with supplementary foil (aluminum oc
stainless steel) substrates. These substrates require the
use of grease for easily re-entrained particles. Enough
of the grease solution is placed evenly on the substrate
to adequately cover the area under the jets. The normal
cautions on the use of greased substrates apply as discussed
in the text.
Precutter cyclone — A BCURA (British Coal Utilization
Research Association) designed precutter cyclone is avail-
able from the manufacturer.
217
-------
Hark III sampling train — As the Mark III is a high
flow rate device, its sampling train is similar to that of
the Andersen.
Mark III clean-up -- Mark III impactor clean-up is
similar to that for the Brink. Some problems have been noted
with 0-rings sticking rather tenaciously and care must be
exercised not to dislodge the sample while trying to sepa-
rate the stages.
C.7.4 Meteorology Research, Inc. (MRI) Model 1502 Cascade
Impactor
The MRI impactor is a high flow rate sampler. The body
of the instrument is constructed from quick-disconnect rings
which allow flexibility in configuration of the impactor
and a positive gas seal between stages. The impactor uses
multiple round jets in its stages.
Sampling rate — The sampling rate is nominally 235
cm3/sec (0.5 acfm) in the seven-stage configuration. Higher
flow rates have been used by removing the last stage.
Collection substrates and adhesives -- The MRI collec-
tion disc is a self-supporting foil (316 stainless steel)
which is functionally similar to the collection cup or tray
and inserts used in other impactors. The collection discs
are mass produced and normally are used only once and dis-
carded.
Grease applied as described earlier is recommended for
most applications.
218
-------
Back-up filter — The MRI impactor has a built-in
filter holder for 47 mm diameter filters. Normally, hinder--
less glass filters are used. Filter losses can be prevented
by placing tared Teflon washers on both sides of the filter
during the test.
MRIsampling train — The MRI sampling train is similar
to that of the Andersen.
MRI clean-up — The clean up of the MRI impactor is
similar to the Brink. The device is clamped in a vice and
all of the sections and nozzles are loosened with wrenches.
The wall losses are carefully brushed onto the appropriate
collection disc. Care is taken not to brush contamination
from the threads into the sample. A tared foil dish is used
to hold the back up filter. Any worn 0-rings should be re-
placed and the whole unit carefully cleaned before the next
test.
C.7.5 Sierra Model 226 Source Cascade Impactor
The Sierra impactor is a six-stage, high sample rate
cascade impactor. The Sierra instrument uses a radial-slot
design.
Sampling rate -- The Sierra impactor has a nominal sam-
pling rate 236 cm3/sec (0.5 acfm). The flow rate must be
low enough to prevent re-entrainment of particles.
Collection substrates and adhesives — Substrates for
the Sierra are obtained precut from the manufacturer. These
are glass fiber substrates and should be checked for weight
gain. Stainless steel substrates are also available and
these should normally be coated with grease as described
earlier.
219
-------
Back-up filter -- The back-up filter uses a 47 mm glass
fiber filter mat. It is supported by a screen from below.
Precutjier cyclone -- A precutter cyclone is available
from the manufacturer.
Sampling train — The sampling train for the Sierra
is similar to that of the Andersen,
Clean-up — Clean-up of the Sierra is fairly similar
to Andersen clean-up. Care should be taken to be sure the
glass fiber substrates are removed intact.
C.8 DATA ANALYSIS
C.8.1 Gen era1 D iscussion
The majority of impactor data reduction is done using
a technique referred to as the "D5.:" method. In this method
it is assumed that all of the particles caught by an impactor
stage consist of particles having diameters equal to or great-
er than the D50 of that stage, but less than the diameter
of the stage above. The mass caught by the first stage or
cyclone is considered to be equal to or greater than the
D5o of that stage, but smaller than the largest particle
diameter present in the aerosol. The largest diameter may
be determined, approximately by microscopic examination of
the stage catch, or it may be assigned some reasonable ar-
bitrary value, say 100 ym.
Particle size distributions may be presented on a dif-
ferential or a cumulative basis. When using the D5 o method,
either type of presentation may be easily employed.
220
-------
The size parameter reported can be aerodynamic diameter,
aerodynamic impaction diameter, or Stokes diameter. In all
cases, the particles are assumed to be spherical. The method
of reporting diameters depends to a large extent upon the
ultimate use of the size distribution information. For this
reason it is suggested that the data be reported in three
parallel sets: one set based on aerodynamic impaction dia-
meter, one based on aerodynamic diameter, and one based on
the Stokes diameter.
C.8.2 Calculation of Impactor Stage DSQ'S
The reduction of field data obtained with a cascade
impactor can sometimes be troublesome and time consuming
because of the computations involved. The equations below
are based on the motion of particles in the Stokes regime
for which the Reynolds number is less than 1000. Although
this is not always true for impactors, the equations are
often a good approximation. The basic equation that defines
the theoretical impaction behavior of a given stage of a
cascade impactor is
p D2V.C
,|, = P 3 tr~i}
V 18 uD. ' (Lf]
where u is the Stokes number,
p is the particle density,- gin/cm3 ,
V. is the velocity of the gas (and particles) in
the impactor jet (cm/sec),
y is the gas viscosity, poise,
D, is the diameter (width for slots) of the jet, cm,
and
C is the slip correction factor.
221
-------
If the value of \|j for 50 percent collection, 'J'sc. can
be determined, equation C7 can be inverted to give the stage
DSC for a wide range of test conditions. Historically, the
experimental values reported by Ranz and Wong2 have been
used. These are:
For round jets, ibso = 0.145, and
For rectangular jets, ip50 = 0.44.
Subsequent studies, however, have shown that there is
no universal value for &$$ and the actual value must be de-
termined by calibration for each impactor design. Several
papers and reports are available which tabulate stage con-
stants of different impactors, and outline procedures for
impactor calibration.3'4
From equation C7,
/18il>50 D.\ h
•>>« -(-^Tv
As equation (CB) is written, with the actual particle
density and the calculated slip correction factor, it defines
the Stokes diameter. If the particles are treated as if
their density, p , was 1.0, equation (C8) defines the aero-
dynamic diameter. If the slip correction factor is also
assumed to be equal to 1.0, the aerodynamic impaction dia-
meter is defined by equation (C8).
Since Cf the slip correction factor, contains D, this
equation must be solved by iteration where Dso and C are
calculated alternately.
Equation C8 may be written more conveniently in terms
of the test parameters: For round jet impactors
222
-------
(C9)
where X. is the number of jets on the stage,
P. is the absolute pressure downstream of the jet(s),
mm Hg,
P is the absolute pressure in the stack, mm Hg, and
o
Q is the sample flow rate, cm3/sec-
and for rectangular jet impactors
where W. is the jet width, cm, and
L. is the total jet length, cm.
One approach that can be used to simplify the computa-
tions is to develop curves for the impactor stage cut points
at one set of conditions — e.g., air at standard conditions
and a particle density of 1.0. Then a suitable correction
factor can be applied to these curves for the actual sampling
conditions. Unfortunately, further steps are involved in
making the correction factor simple enough to be of value.
Therefore, the use of this type of approach suffers from
some restrictions.
All of the assumptions and calculations involved in
going from equations C9 to CIO to the calibration curve can
be quite awkward, particularly in cases where different types
of sources are being sampled. The best and easiest approach
is to write or obtain a computer program based on the rigorous
223
-------
equations given initially. Such a program can calculate
impactor stage cut points, compute concentrations of par-
ticles in each size range, determine the precipitator effi-
ciency, and plot graphs.
A sophisticated computer data reduction program IB avail-
able from the EPA and also less powerful programs are avail-
able for the Hewlett Packard HP-65 and HP-25 programmable
calculators.5,'s
C.8.3 Cumulative Particle Size Distributions
Impactor data may be presented on a cumulative basis
by summing the mass on all the collection stages and back
up filter, and plotting the fraction of the mass below a
given size versus size. This is frequently done on special
log-probability paoer. Semi-log paper may be preferable
for distributions that are not log-normal.
Cumulative distributions are very easy to understand
and present the data with clarity. For this reason, they
should be presented as part of each particle sizing report.
Cumulative distributions do have a couple of disadvantages
when compared to differential distributions. An error in
stage weight will be propagated throughout a cumulative
analysis, but will be isolated by the differential approach.
Also the differential method does not involve the use of
total mass concentration or total size distribution from dia-
meters of zero to inifinity, and so is useful in comparing
instruments with overlapping but different size fractiona-
tion ranges and different stage cut points while cumulative
analysis is not.
When cumulative plots are used, the abscissa is normally
the logarithm of the particle diameter and the ordinate is
224
-------
the weight percent smaller than this size. The value o£
the ordinate at a given (D-ia), would he
k-1
E
AM.
Weight percent smaller than (D50)k = ---- ^ -- x 100%, (Cll)
AM.
i
1 = 0
where i = o corresponds to the filter,
i = k corresponds to the stage under study, and
i = K corresponds to the coarsest jet or cyclone.
This equation requires that the stages be counted from
the final filter UD. There is no (D50) , as the "o" corre
o
spends to the filter. (Dso)-i is the cut point of the last
stages, which collects mass, AM,.
C.8.4 Differential Size Distributions
Differential particle size distributions are used to
plot the relative concentration versus particle diameter.
The area under the frequency curve, between two designated
diameters, is equal to the mass of particles in that size
range. Differential curves may be obtained directly from
the reduced impactor data, or by differentiation of curves
fitted to the cumulative particle size distributions.
Many cascade impactors are designed so that the relation
ship between successive stage D5o's is logarithmic. For
this reason, and to minimize graphical scaling problems,
the differential particle size distr ibutoins are plotted
on log-log or semi-log paper with AM/A(logD) as the ordi-
nate and geometric mean of A(logD) as the abscissa. The
225
-------
mass on stage "n" is designated by
mation, the particulate mass with diameters between
M and is, in approxi
(Dso
and (D5o)n+n- Tne A(logD) associated with AM is log(D5c) ^-,
log (D5 o
n
Using these approximations, the derivative term
(ordinate) associated with stage
is:
AM/A(logo)
AM
n
Mass on Stage
n
A(LogDs o)
n
log(D5o)n+1 - log(Dsc)n
(CL2;
and the abscissa, D , is
[_(Dso)n+1 x (D50)nJ
-------
can be calculated. Generally the data are presented an
1) Cumulative Mass Versus Particle Size, 2) Cumulative
Per Cent Mass Versus Particle Size, and 3) AM/A(logD) Ver-
sus Geometric Mean Particle Size. To calculate the Upper
and Lower 90% Confidence Limits for the data, the follow-
ing equations are used:
UCL9Q = Average + ^JL5- = Average + C.I. 90 'C141
/N
LCL90 = Average - i-!±2 = Average - C.I. 90 (C15)
/N
where
Average = The average of the N values for AM/AlogD or
AN/AlogD at a particular particle size.
tso = The Student's t distribution for 90% confidence
limits with N sets of data.
a = The Standard Deviation of the N sets of data.
N = The number of data values in the average.
The average cumulative mass and cumulative per cent
graphs are obtained from AM/AlogD data by piecewise inte-
gration, after discarding outliers. The confidence interval
of the cumulative graph, at a particular size, is equal to
the square root of the sum of the squares of the confidence
intervals of each size increment of the differential graph
less than or equal to that size.
Clearly, the average values are more reliable and the
confidence limits are small when the number of data points
is large.
A determination of precipitator fractional penetration
is made by taking the ratio of the average AM/AlogD outlet
values to the average AM/AlogD inlet values at a series of
227
-------
particle sizes. The confidence limits of the calculated
penetration are given by:
and
where :
and
UCLso = Average Penetration + CIPsa = P + CIP30 (C16)
LCLsc = Average Penetration - CIPgo = P - CIP9o (C17
CIP
3 Q
_
(C.I, go, Outlet)2 _,_ P2 (C.I.go. Inlet)2
(Inlet Average)
Out 1 e t Average
Inlet Average
(Inlet Average)
(CIS
C.8.6 CASCADE IMPACTOR DATA REDUCTION—SAMPLE CALCULATION
This section contains a detailed description of the
calculations that are required to derive particle size dis-
tributions and a precipitator fractional efficiency curve
from raw impactor data. The specific example given is for
a single hypothetical test performed with an Andersen impac-
tor. The calculation procedures outlined here can be used
with any impactor, however. Normally the results of all
tests made under the same process and precipitator operating
conditions are grouped and averaged, and confidence limits
calculated as described in Section C.8.5.
This discussion is based on Table C3 which was generated
by a computer program. All of the calculations, however,
can be done on programmable calculators. In the example
shown, the data is reduced using a particle density of 1.35
gm/cm3; thus the diameters reported are Stokes diameters.
228
-------
TABLE C3
SAMPLE CAr-COLATTON - INPOT DATA AND RESULTS
Hypothetical Andersen
Irapactor Flowrate = 0.500 ACFM
Irapactor Pressure Drop = 1.5 In. of Hg
Assumed Particle Density = 1.35 gm/cu.cra.
Irapactor Temperature = 400.0 F = 204.4 C
Stack Temperature = 400.0 F = 204.4 C
Stack Pressure = 26.50 In. of Hg
Sampling Duration = 20.00 Min
Max. Particle Diameter = 100.0 Micrometers
Gas Composition (Percent) CO2 = 0.95 CO = 0.00 N2 = 76.53 O2 - 20.53 H20 = 1.00
Calc. Mass Loading ~ 8.0711E-03 gr/acf 1.4948E-0? qr/fiscf 1.8470E+01 mg/acm 3.4207E+01 mg.ATscm
Impactoc Stage
Stage Index Number
D50 (Micrometers)
Mass (Milligrams)
MG/DSCM/STAGE
Cum. Percent Of Mass Smaller Than D50
(MG/ACM) Smaller Than D50
(GR/ACF) Smaller Than 050
(GR/DSCF) Smaller Than D50
Mean Dia. (Micrometers)
Cum,
Cum.
Cum.
Geo.
DM/DLOGD (MG/DSCM)
DN/DLOGD {NO. PARTICLES/DSCM)
SI
1
10.74
0,72
4.71EHH)
86.24
1 . "59E+01
6.96E-03
1 .29E-02
3.28E+01
4.86E+00
1.95E+05
32
9
0
2
78
1
6
1
1
7
1
7
-*)5
.50
.62F,*00
.59
-45E+01
.3-5E-03
.17E-02
.03E+01
-94E+OL
.02E+08
S3
6
0
3
68
1
5
1
7
1
•>
3
.36
.53
. 47RtOD
.46
. 26E+01
. "J3E-03
-02E-02
. 96E+00
.""tE+Ol
-01E+07
34
4
4.19
0.09
5.89E-01
66.74
1.23E+01
5.39E-03
B.'JSE-OS
•;.17E+00
3.25E^OO
3.33E+07
S5
r.
2.
0.
2,
59.
1 ,
4.
8.
3.
8.
4.
22
38
49E+00
47
10E+01
BOE-03
89E-03
05E1-00
q9K(-OD
48E(-Ofl
1 ,
t
0.
32.
5 ,
2.
4,
1.
3.
\ ,
S6
d
29
43
35E+00
13
93E+QO
59E-03
80E 03
69E+00
99E+C1
16EUO
0
1
e
8
1
6
1
9
2
5
S7
7
.69
.25
. 18E+00
.23
. 52E+OQ
.64E-04
.23E-03
. 43E-01
. 98E+01
.03E+10
S3
8
0,
0.
2,
7,
I.
6
1.
4,
8
1 ,
.33
,04
.62E+01
,46
, 3BE+00
. 02E-04
, 12E-03
.74E-01
•09E-01
, 08E+10
FILTER
9
0.39
2.55E+00
2.31E-01
8.47E+00
9.74E+11
Normal or standard conditions are 2]°C and 760 ram Hg
-------
For aerodynamic, or aerodynamic impaction diameters, p
or p and Cp, are set to unity, respectively- (P is particle
density and C is the slip correction factor.)
Information obtained from the data log sheets for each
test is printed at the top of Table C3. The maximum particle
diameter is measured by examining the particles collected
on the first stage (or first cyclone) with an optical micro-
scope. Gas analysis samples are taken at the same time the
impactor is run. The mass loading is calculated from the
total mass of the particles collected by the impactor, and
listed in four different systems of units after the heading
CALC. MASS LOADING. The symbols are defined as:
GR/ACF - grains per actual cubic foot of gas at stack con-
ditions of temperature, pressure, and water content.
GR/DSCF - grains per dry standard cubic foot of gas at en-
gineering standard conditions of the gas. Engineer-
ing standard conditions are defined as 0% water
content, 76°F, and 29.92 inches of Hg.
MG/ACM - milligrams per actual cubic meter of gas at stack
conditions of temperature, pressure, and water
content.
MG/DSCM - milligrams per dry standard cubic meter of gas at
engineering standard conditions of the gas. En-
gineering standard conditions are defined as 0% water
content, 21°C and 760 mm of Hg (Torr).
The conditions at which the impactor was run to determine
stage Dso cut points. These are calculated by iterative solu-
tion of the following equations:
230
-------
14. 1 gD^ 3i|J5o .'P_ X.
O * 1 o - 1
Pp Q Po C.
'C19
C. = 1 +
1
DSC,- x
LJ5 0 •
[l.23 + 0.41 EXP(-0.44^ x 10" "'M
, n - i. L Ai J
where Dso- = stage (i) cut point (cm),
p = gas viscosity (poise),
D = stage jet diameter (cm),
P = local pressure at stage jet (mm Hg),
i
P = particle density (gm/cm3)r
Q = impactor flow rate (cm3/sec),
Po = ambient pressure at impactor inlet (mm Hg),
C. = Slip Correction Factor,
X , = gas mean free path (cm),
x. = number of holes per stage (i), and
ilJso • = Stokes number of stage (i) .
To find the pressure P at each impactor stage, the
i
following equation is used:
Ps = Po - (F.) (DP) , (C21)
i
where P0 is the ambient pressure at the impactor inlet,
is the fraction of
at each stage, and
F. is the fraction of the total impactor pressure drop
231
-------
DP is the total pressure drop across- the impactor.
The total pressure drop DP across the impac-
tor is given by the following equation.
DP = K Q2 p MM, (C22)
where K = Empirically determined constant for each impac-
tor ,
Q = Flow rate through impactor (cm3/sec),
= Gas density (gm/cm3), and
MM = Mean Molecular Weight of gas (gm/gm-mole).
To calculate the gas mean free path, X., for each im-
pactor stage, the following equation is used:
/8.31 x 107 T
__ - £ (C23)
3 m
1 1.013 x 106 P
>D '
where y is the gas viscosity (poise) ,
P is the pressure at each impactor stage (atm) ,
T is the gas temperature at the impactor
stage (°K), and
MM is the average flue gas molecular weight.
To find the viscosity of the flue gas, y, the viscosity
of the pure gas components of the flue gas must first be
found. Viscosity is a function of temperature, and the tem-
perature difference in different flue gases can be quite
significant. The following equations (derived from curves
fitted to viscosity data from the Handbook of Chemistry and
232
-------
Physics, Chemical Rubber Company Publisher, 54 Edition,
1973-1974, pp. F52-55), are used to find the viscosities
of C02(yi), C0(y2), N2(jJ3), 02(yO and H20(Ms).
Hi = 138.494 + 0.499T - 0.267 x IO"'T2 + 0,972 x 10~7T3
y2 = 165.763 + 0.442T - 0.213 x 10~3T2
y3 = 167.086 + 0.417T + 0.417 - 0.139 x 10"3T2
y* = 190.187 + 0.558T - 0.336 x 10~3T2 + 0.139 x 10~6T5
ys = 87.800 + 0.374T + 0.238 x 10"4TZ
where T is the temperature of the flue gas in degrees Celsius,
The units of y are 10 6 g/cm-sec. Next, these values of u.
through y5 are used in a general viscosity equation for a
mixture of any number of components (See "A Viscosity Equa-
tion for Gas Mixtures" by C. R. Wilke, Journal of Chemical
Physics, Volume 8, Number 4, April 1950, page 517) used to
find the viscosity of the flue gas:
n
M t^t
i-i
1 f l
L Xi
i=?n
v ^ ,, .
X A • CD . .
/ j 1 1]
j = l " J
where
-------
y = viscosity, gm/cm-sec; \i i , y2r etc. refer to the
pure components at the temperature and pressure of
the mixture, y is the viscosity of the mixture, and
-------
where absolute means the temperature and pressure are in
absolute units-degrees Rankin or degrees Kelvin for tempera
ture, and atmosphere, inches or millimeters of mercury for
pressure. For SI,
= '72 m9 v 35.31 cubic feet/cubic meter
X 0.500 ACFM
x (400 + 460)°R 29.92 in. Hg __^ = A 71
(70 + 460)°R 26.50 in. Hg (1.0 - 0.01) '
The subscripts indicate stage index numbers.
The percent of the mass of particles with diameters
smaller than the corresponding D5a is called the CUMULATIVE
PERCENT OF MASS SMALLER THAN D5„. It is the cumulative mass
at stage j divided by the total mass collected on all the
stages, and converted to a percentage:
9
XT' MASSi
rnM 5- = inJ—±_ y inn ic7~i\
CUM °j Total Mass x 1UU (L//)
For example, for S6, the cumulative percent is given by
CUM %s = MASS7 ! y^S, + MASSs
Total Mass
= 1-25 me, + 0.04 m(? + Q.39 mg 10Q = 32
-------
Note that the apparent error in the least significant figures
of the calculated percentages is due to using masses from
the computer printout which have been rounded off to two
decimal places before printing.
The cumulative mass loading of particles smaller in
diameter than the corresponding D=o in milligrams per actual
cubic meter (CUM. (MG/ACM) SMALLER THAN Dbo) for a particu-
lar stage j is given by the formula
9
E
MASS.
i
CUM. (MG/ACM) - =
j sampling duration (min)
35. 3]^ cubic feet/cubic meter ,
FLOWRATE (ACFM)
From the information at the top of the computer print-out
sheet, the flow rate is 0.500 actual cubic feet per minute
(ACFM) and the sampling duration is 20.00 minutes. There-
fore, for S4,
PTTM fMr/ar*n - s + MASS6 + MASS? + MASSe + MASS
CUM. (MG/ACM) „ 20 minutes
„ 35.31 cubic feet/ cubic meter _
x - Q.5QQ ACFM - ~
For 58, the mass of the particulate collected on the filter
is again used,
HTTM /Mr/a™/n - MASS9 „ 35.31 cubic feet/cubic meter
CUM. (MG/ACM) a - 20 minutes X 0.500 ACFM
0.39 mq 35.31 cubic feet/cubic meter
~ 20 minutes x 0.500 ACFM
= 1.38 mg/ACM
236
-------
The cumulative mass loading of particles smaller in
diameter than the corresponding Dso in grains per actual
cubic foot (CUM. (GR/ACF) SMALLER THAN D5a) for a particu
lar stage j is given by the formula
CUM. (MG/ACM) .
CUM. (GR/ACF) « 2.288 -
•3 - : — — -, — r-. - 2 - r * 1000 mq/gram
grains/cubic foot ^ ^
(C30;
For SI,
1.52 mg/ACM
CUM. (GR/ACF) 7=
-
^rams/cubic meter rag/gram
grains/cubic foot ^'^
= 6.64 x 10~" grains/ACF
The cumulative mass loading of particles smaller in
diameter than the corresponding 0,0 in grains per dry stan-
dard cubic foot (CUM. (GR/DSCF) SMALLER THAN D5 o) is calcu-
lated to show what the above cumulative would be for one
cubic foot of dry gas at 70''F and at a pressure of 29.92 .
inches of mercury. For a particular stage j,
CUM.(GR/DSCF). = CUM.(GR/ACF).
Absolute Stack Temperature Absolute Standard Pressure
Absolute Standard Temperature X Absolute Stack Pressure
(C31)
(1-Fraction of H20)
where absolute means the temperature and pressure are in
absolute units-degrees Rankin or degrees Kelvin for tempera-
ture, and atmospheres, inches or millimeters of mercury for
pressure. For Si,
237
-------
CUM. (GR/DSCFJj = 6.96 x 10 3 gr/ACF
(400 + 460)°R 29.92 in. Hg .1 = i ?q x i n~ ^
x (70 + 460)°R x 26.50 in. Hg (1.00-0.01)
gr/DSCF
The particle size distribution may be presented on a
differential basis which is the slope of the cumulative curve
Differential size distributions may be derived two ways:
1. Curves may be fitted to the graphs of cumulative
mass vs. particle size, and then the differential curves
(slope) of each test would be calculated by taking finite
differences along the ordinate and abscissa of the fitted
curve.
2. The finite differences may be taken equal to the
differences in Dso's from stage to stage (abscissa) and the
particulate mass on each stage (ordinate). This technique
was used to calculate the differential size distribution
data in Table C3, and is described in detail in the follow-
ing paragraphs.
If we define the terms:
AM . = MG/DSCM/STAGE , and
(AlogD). = logio (Dso j_1) - logi0(D50.), then
/ .... \ MG/DSCM/STAGE,
___AM _ ; _ 1 _ . _ D _
__ _ _
AlogD f . logi o (D5Q • •
Because the computer printer does not contain Greek letters,
the computer print-out sheet reads DM/DLOGD instead of
AM/ALOGD. For S6,
238
-------
AM _ __ 9.35 mg/DSCM _ g ma/DSCM
ALOGD B ~ log10{2.22) - Iog10(1.29) " ^ '7 m<3/DSCM
Note that AM/ALOGD has the dimensions of the numerator since
the denominator is dimensionless . In the calculation for SI.
a maximum particle diameter is used. For this example, MAX.
PARTICLE DIAMETER = 100.0 micrometers.
( AM \ _ _ 4.7.1 mq/DSCM _ - 4 86
\ALOGD j: ~ Iog10(100) - log, „ (10.74) ~ 4>b6
For the filter stage, the DSD is arbitrarily chosen to
be one-half of the Dso for stage eight (S8). For this example,
it is chosen to be 0.33 micrometers/2 = 0.165 micrometers.
Thus ,
AM
\ = _ 2.55 mg/DSCM
_
ALOGD logia(0.33) - log, „ (0.165)
The geometric mean diameter in micrometers (GEO. MEAN
DIA. (MICROMETERS)) for a particular stage j is given by
the formula
GEO. MEAN DIA.. = /D5a . x D5o . , (C33)
3 3 1~L
For S8,
GEO. MEAN DIA.3 = /0.33 x 0.69 micrometers
= 0.477 micrometers
As in the ALOGD calculation, we again use the maximum particle
diameter for the stage one calculation and one-half the D5d
for stage eight for the filter stage calculation.
239
-------
For SI,
GEO. MEAN DIA.i = /I0.74 x 100.0 micrometers
= 32.8 micrometers
For the filter,
GEO. MEAN DIA.g = /0.165 x 0.33 micrometers
= 0.23 micrometers
A differential number distribution /for comparison with
ultrafine data) can also be derived. Since AM. = MG/DSCM/STAGE.
is the mass per unit volume for stage j then we can define
AN. as AN. = NUMBER OF PARTICLES/DSCM/STAGE. or the number
of particles per unit volume for stage j. Now AM. and AN.
are related by the equation M. = N. x m , where m is the
average mass of the particles collected on one stage. Di-
viding both sides of the equation by m x ALOGD yields
m ^ ALOGD
Now m = p V where is the assumed particle density and
P P P P
V is the average volume of one particle on one stage. To
obtain m in milligram units when p is in grams per cubic
centimeter and V is in cubic micrometers? certain conversion
factors must be used. The complete formula, using the correct
conversion factors and the expression (4/3) (n) (D/2)3 for
V where d is the geometric mean diameter in micrometers, is:
P
240
-------
____ ___ ! ^_ ^, ... —j ls -•'i*wii t .L XJ v—. JII
P P 1 9m \ 3/\2y y 1 cubic microraeten
= (5.23599 x 10'1 °)P<33 .
Therefore ,
AN \ _ (AM/ALOGD) .
A LOGO
5.23599 x 10 10p d3
P
where AM/ALOGD is in units of mg/DSCM, p is in gin/cm3 ,
d is in microns, and AN/ALOGD is in number of particles/
DSCM. For S3,
/ AN \ _ 17.9 mq/DSCM
\ALOGD r (5.23599 x 10~10) x (1.35 gm/cc) x (7.96 microns)3
= 5.02 x 107 particles/DSCM.
For the filter stage
/ AN \ _ 8.47 mq/DSCM
\ALOGD/3 (5.23b99 x 10~10) x (1.35 gm/cc) x (0.231 microns);
= 9.72 x 1011 particles/DSCM
The test data are usually classified according to sampling lo-
cation (outlet of inlet), sampling time (day, week, etc.)
and combustion chamber or pollution control device conditions
(high or low sulfur coal for coal plants, normal or below
normal fuel consumption, normal or below normal current density
for electrostatic precipitators, etc.). When classified,
all of the data taken in a single classification are usually
averaged and plotted on appropriate graph paper.
241
-------
Curves are drawn through the discrete points. At selec-
ted particle sizes values of AM/Alogd and AN/Alogd are chosen
from these curves for averaging.
Figure C5 shows a typical average AM/Alogd outlet par-
ticle size distribution curve, from Andersen impactor data,
and Figure C6 shows the same data presented as AN/Alogd.
Figure C7 shows similar data for a hypothetical inlet
test when a Brink impactor was used. The size distributions
of Figures C5 or CV'were used to calculate the penetration/
efficiency curve of Figure C8. Confidence limits were cal-
culated for the average data as described in Section C.8.
A computer program has been written which does all of
the calculations necessary to generate plotted graphs of
the particle size distributions and fractional efficiency
curves from the raw field data. Table C4 shows a flov? dia-
gram for this program. Copies of the report describing this
computer program will soon be available from the EPA or NTIS.
242
-------
102
101
o
us
Q
en
E
cT 10°
(3
O
10
1-1
10-'
ERROR BARS INDICATE 90% CONFIDENCE INTERVAL
10-1 10 ° 101
GEOMETRIC MEAN DIAMETER, micrometers
102
3630-091
Figure C5. Hypothetical particle size distribution at an ESP outlet
determined from Andersen impactor data.
243
-------
1Q10
109
E
Z
D
a
c
u
o
_!
<
Z
108
107
106
105
ERROR BARS INDICATE 90% ' \\
CONFIDENCE INTERVAL
I
icr1 10° iQ1 io2
GEOMETRIC MEAN DIAMETER, micrometers
3530-098
Figure C6. Hypothetical particle size distribution at an ESP outlet
determined from Andersen impactor data.
244
-------
u
M
Q
"Si
E
o
_l
<
ID2
10-1
ERROR BARS INDICATE 90% CONFIDENCE INTERVAL
10° 101
GEOMETRIC MEAN DIAMETER, micrometers
102
3630-0&7
Figure C7, Hypothetical particle size distribution at an ESP inlet
determined from Brink impactor data.
245
-------
ERROR BARS INDICATE 90% CONFIDENCE INTERVAL
<
cc
1-
LU
a.
U.U 1
0.1
0.2
1
2
5
10
20
40
60
80
90
95
98
PQ
— .
—
__
—
—
—
—
_
—
"
I I I I I
(
" J 5 5
> -*•
I I I I I
?
i i i
I T
TO
V
I I I
^^^w
—
r -
—
as. as
99.9
99.8
99
98
95
90
80
60
40
20
10
5
2
y
LU
GEOMETRIC MEAN DIAMETER, micrometers
3630-O96
Figure C8. Hypothetical ESP fractional efficiency curve based on
the data presented in Figures C5 and C7,
-------
TABLE C4
PROGRAM FLOW
inlet data:
I. Impactor Program (MPPROG)
Takes testing conditions, stage weights, and impactor constants
to produce stage Die's, cumulative and cumulative % mass
concentrations < Dsc* geometric mean diameters, and mass
and number size distributions.
II. Fitting Program (SPLIN1)
Uses modified spline technique to fit cumulative mass loading
points for each plot. Stores fitting coefficients and boundary
points on file.
III. Graphing Program (GRAPH)
Produces individual run graphs with points based on stage
weights and impactor Dso's. Also superimposes plot based
on fitted data, if desired. Graphs include cum. mass load-
ing, cum. % mass loading, and mass and number size distribu-
tions.
IV. Statistical Program (STATIS)
Recalls cum. mass loading fitting coefficients to produce
avg. cum. mass loading, avg. % cum. mass loading, avg. mass
size distribution, and avg. number size distribution plots
each with 90% confidence bars.
Repeat programs I-IV for outlet data.
V. Efficiency Program (PENTRA)
Recalls avg. mass size distribution values along with 90%
confidence limits for inlet and outlet to plot percent pene-
tration and efficiency with 90% confidence bars.
247
-------
REFERENCES
1. Felix, L. G., G. I. Clinard, G. E. Lacey, and J. D. McCain.
Inertial Cascade Impactor Substrate Media For Flue Gas
Sampling. Interagency Energy-Environment Research and
Development Program Report, EPA-600/7-77-060, June 1977,
2. Ranz, W. E. and J. D. Wong,"Impaction of Dust and Smoke
Particles," Ind. Eng. Chem., 44: 1371-1381, June 1952.
3. Harris, D. B. Procedures For Cascade Impactor Calibration
And Operation In Process Streams. U.S. Environmental Pro-
tection Agency Report No. EPA-600/2-77-004, January
1977.
4. Gushing, K. M., G. E. Lacey, J. D. McCain, and W. B. Smith.
Particulate Sizing Techniques For Control Device Evalua-
tion: Cascade Impactor Calibrations. U.S. Environmental
Protection Agency Report No. EPA-600/2-76-280, October
1976.
5. Ragland, J. W., K. M. Gushing, J. D. McCain, and W. B.
Smith. HP-25 Programmable Pocket Calculator Applied To
Air Pollution Measurement Studies: Stationary Sources.
Interagency Energy-Environment Research and Development
Program Report, EPA-600/7-77-05, June 1977.
6. Ragland, J. W., K. M. Gushing, J. D. McCain, and W. B.
Smith. HP-65 Programmable Pocket Calculator Applied
To Air Pollution Measurement Studies: Stationary
Sources. U.S. Environmental Protection Agency Report
No. EPA-600/2-76-002, October 1976.
248
-------
Appendix D
SIZE DISTRIBUTIONS OF SUBMICRON AEROSOL PARTICLES
Page
D.I SYSTEM FOR EXTRACTIVE SAMPLING 250
D.I.I Line Losses
D.I.2 Condensation of Gases
D.I.3 Temperatures
D.I.4 Electrostatic Losses
D.I,5 Humidity
D.I.6 Dilution
D.I.7 Sample Extraction-Dilution System (SEDS)
D.2 SUBMICRON PARTICLE SIZING TECHNIQUES BASED ON PARTICLE
DIFFUSIVITY 263
D.2.1 Diffusion Batteries
D.2.2 Particle Concentration Indicators-
Condensation Nuclei Counters (CN)
D.2.3 Using Diffusion Batteries with Condensation
Nuclei Counters and a Sample Extraction-
Dilution System to Measure Concentrations
of Submicron Aerosols in Industrial Flue
Gases
D.2.4 Data Reduction Techniques and a Sample
Calculation
D.3 SUBMICRON PARTICLE SIZING TECHNIQUES USING THE
ELECTRICAL MOBILITY PRINCIPLE . 285
D.3.1 Electrical Mobility
D.3.2 Thermosysterns Inc. Model 3030
Electrical Aerosol Size Analyzer
D.3.3 Using the TSI Model 3030 Electrical Aerosol
Analyzer with a Sample Extraction-Dilution
System to Measure Concentration of Submicron
Aerosols in Industrial Flue Gases
D.3.4 Data Reduction Techniques and a Sample Calculation
D.4 OPTICAL PARTICLE COUNTERS. 300
249
-------
APPENDIX D
SIZE DISTRIBUTIONS OF SUBMICRON AEROSOL PARTICLES
If it is desirable to measure the precip.itator collection
efficiency for ultrafine particles, measurements of the particle
size distribution must be made at the inlet and outlet. This
appendix describes procedures for making measurements of the
particle size and concentration from 0.01 to 2 urn diameter.
D.I SYSTEM FOR EXTRACTIVE SAMPLING
When possible, in-stack sampling is preferred because it
eliminates many condensation and sample loss problems which
occur when probes are used for extractive sampling. Unfortu-
nately, existing submicron sizing techniques and instrumenta-
tion are designed for a laboratory environment and cannot be
used in-stack.
Particulate concentrations are usually extremely high in
industrial flues and vary by orders of magnitude from one in-
dustrial process to another and from the inlet side of a control
device to the outlet of the same device. Temperature, pressure.
moisture content, and the physical properties of the particulate
also vary widely from one industrial process to another. Be-
cause of this complexity and the limited useful concentration
range for particle sizing techniques, extensive sample dilution
and conditioning is required to obtain information on submicron
particles in an industrial gas stream.
D.I.I Line Losses
When extracting the sample, attention should be given to
line losses which can be a problem for particles smaller than
0.01 ym or greater than 1.5-2.0 pm diameter. Since the objec-
tive is to measure the concentration of ultrafine particles,
250
-------
there is little interest in measuring particles larger than
about 2.0 ym. Thus losses due to impaction and settling are
not significant, and isokinetic sampling is unnecessary. Dif-
fusional and electrostatic line losses are of concern, however.
For example, at a sampling rate of 1 S.pm, a sample line will
remove 0.005 urn particles by diffusion at a rate of about 8%
of the instantaneous concentration for every foot of sample
line, independent of the radius of the sample line. This pro-
blem can never be eliminated for the case of out-of-stack sampling,
but it can be minimized by (1) using short probes and as high
a flow rate as is practical, or(2) by using high, nonturbulent,
flow rates through a probe and connecting lines to a conditioning-
dilution system in which a sample is split off at the required
rate. This second technique allows one to use long probes and
change sampling points without disconnecting and reconnecting
all sampling lines. Diffusional line losses for non-turbulent
flow can be estimated from the equations for diffusion to the
walls in a circular geometry (Figure Dl). A similar discussion
of diffusional line losses has been given by Ensor and Jackson.1
D.I.2 Condensation of Gases
Another problem of concern is condensation. Elements which
are at a gaseous state at stack temperature (SO3/H2S04 in parti-
cular) can drop below their dew point and form high concentra-
tions of very small particles resulting in anomalously high
readings. In the case of S03 in the presence of H2O, a sulfuric
acid fume can be formed if temperatures fall below the acid
dew point (temperature, pressure, and concentration sensitive).
Once this fume has been formed, very high temperatures are re-
quired to re-evaporate the droplets. For this reason tempera-
tures above the dew point must be maintained throughout the
system until the gaseous S03/H2SO^ can be removed or diluted.
Two techniques appear to be useful for doing this: (1) dif-
fusion to an absorber reagent and (2) dilution of the S03/H2S
-------
100
LENGTH
A 183 cm
B 305 cm
C 183 cm
D 305 cm
DIA. FLOW RATE
0,64 cm 0.1 LPM
0,64 cm 0.5 LPM
0.64 cm 0.5 LPM
1.6 cm 10 LPM
I I I I I I I
0.01
0.1 1.0
PARTICLE DIAMETER, /jm
3630-035
Figure D1, Probe losses due to settling and diffusion for spherical particles
having a density of 2.5 gm/cm^ under conditions of laminar flow.
252
-------
a farther dilution is required to bring the aerosol to a
temperature within the operating range of the instrument.
As a result, the minimum total dilution, which is the product
of hot stage and cold stage dilutions, may be excessive for
low particulate concentration levels such as those found
at the outlet of some gas cleaning devices. Hence the con-
centration of ultrafine particles even at the minimum dilu-
tion may be below the minimum detection level of the si zing
instrument.
At low levels of S03, such as those at power plants burning
low sulfur coal, copper has been found to be a successful
S03/H2S04 absorber reagent through the formation of CuSO^.
This reagent has the particular advantage that the reaction
product is water soluble, and the absorber can easily be re-
juvenated. Activated charcoal is also a very effective absorber
of SO , even at relative high concentrations.
A
A special effort must be made to detect any particles
generating interferences and to eliminate these if possible
by conditioning the extracted sample. Condensation may be
observed by periodically checking the linearity of the dilution
system. When the dilution system is adjusted to produce a
many-fold change in dilution, the indicated concentration
should reflect an equal change in measured concentration.
Figure D2 shows a diffusional absorber/dryer for the re-
moval of water vapor from the sample stream. Figure D3 shows
a sample extraction system which includes diffusional absorbers
for high temperature use.
D.I.3 Temperatures
Consideration must also be given to reducing the gas tem-
perature to a level at which the instruments were designed to
operate. This is normally done by using a large volume of cool,
253
-------
DRiERITE
GLASS CYLINDER
SAMPLE AEROSOL
100 MESH STAINLESS
SCREEN
3630-05
Figure D2. Diffusion a/ adsorption apparatus for removal of
sample aerosol.
from
254
-------
NJ
Ul
I
TIME
AVERAGING
CHAMBER
^J
_____
SOX ABSORBERS (OPTIONAL)
PROCESS EXHAUST LINE
CHARGE NEUTRALIZER
CYCLONE
ORIFICE WITH BALL AND SOCKET
JOINTS FOR QUICK RELEASE
HfcATED INSULATED BOX
RECIRCULATED CLEAN, DRY, DILUTION AIR
FILTER BLEED NO, 2
MANOMETER
COOLING COIL
3G30 0315
PRESSURE
BALANCING
LINE
DRYER
BLEED NO. 1
Figure D3. Sample Extraction - Dilution System (S£DSj.
-------
dry dilution air but in some cases can be done by simply pulling
the sample through an ice bath condenser. It should be remem-
bered, however, that unless condensible gases have previously
been removed, it is possible to form a condensation fume
when using a simple condenser.
D.I.4 Electrostatic Losses
Electrostatic line losses can also be a problem. Due to
charges present on the particles, large static electric fields
can be established which result in particle deposition and non-
representative sampling. Line losses may be more severe if
the aerosol particles are charged. Electric fields may exist
which result in particle deposition and nonrepresentative sam-
pling. This problem is most severe if the sample lines are
made of insulating materials. Particle losses may also be a
problem in the diffusion batteries where the theory assumes
that the particles are electrically neutral and no considera-
tion is given to unknown electrical forces. Also, it is assumed
in the application of the electrical aerosol analyzer that the
sample aerosol particles initially bear no charge. Precharged
particles could acquire charges different from those of cali-
bration, causing them to exhibit different mobility vs size
characteristics. It is desirable to neutralize the particle
charge prior to entering the probe nozzle, and charge neutrali-
zation to Boltzmann Equilibrium can be accomplished by exposure
to an ion field created by radioactive materials, however, no
suitable radioactive source has been developed for in-stack appli-
cations. The approach generally taken has been to use radioac-
tive materials such as P02*° in the diluter to neutralize the
particulate after it has been cooled and before it goes to the
sizing instruments. The extent of electrostatic interferences,
however, has not been well quantified.
256
-------
D.I.5 Humidity
High humidities can alter the charging characteristics
of electrical mobility analyzers and can cause water to con-
dense on particles and change their size (similar to the
controlled process occurring in a condensation nuclei counter).
Humidity problems can be eliminated bv the use of (1) dry
dilution air or (2) diffusional dryers.
D.I.6 Dilution
In-stack concentrations generally exceed instrument
concentration limits and must be reduced to levels at which
the instrument functions properly. Large changes in the
aerosol size distribution due to coagulation are also of
concern, particularly in a diffusional configuration in-
volving long residence times, since the loss rate due to
coagulation for a given size particle rises rapidly with
increasing concentration.
Most of the dilution systems used to date involve the
measurement of two different flow rates. Several methods are
available for doing this: (1) Rotameters. (2) Orifices, (3) Ven-
turis, and (4) Mass Flowmeters. Spink3 has summarized the use
of the first three and Parry1* has described the use of several
mass flowmeters in an automated system. Rotameters are con-
venient for measuring cool dilution air flow rates but could
cause sample losses when measuring sample gas flows. They are
also sensitive to temperature, pressure, and gas composition.
Flowmetering orifices are useful for measuring flow rates of
hot gases but significant sample losses can occur for particles
larger than about 2 urn. Orifices also require, as do Venturis,
that pressure drops, etc. be monitored in order to calculate
the flow rate. Venturis have the advantage that less turbu-
lence occurs in the meter and hence size dependent losses
257
-------
should be much less than those for orifices. Venturis have
the disadvantage that the pressure tap must be taken at the
Vena Contracta making them difficult to construct for low
flow rates. Mass flowmeters are ideal but may require line
restrictions in order to throttle the flow rate.
D.I.7 Sample Extraction-Dilution Systems (SEDS)
Figure D3 is a block diagram of a sample extraction-dilu-
tion system (SEDS) developed by Southern Research Institute
under EPA Contract No. 68-02-2114. In this system a 23.3 cm3/
sec sample flow is removed from the process exhaust stream and
is pulled through a rigid probe, a flexible connector hose,
and a cyclone, into a "T" where the flow splits. The excess
flow is dumped and the desired sample flow goes into
the diluter via a calibrated orifice and an optional bank of
sulfur oxide absorbers. The cyclone, orifice, and sulfur oxide
absorber bank are housed in a heated box so that all components
of the system except the diluter can be maintained at the stack
temperature (up to 200°C) to prevent condensation. Pressure
taps for the cyclone and the orifice allow conti-nuous monitoring
of the cyclone flow rate and the orifice flow rate by reading
the pressure drop across the respective component.
Currently, charge neutralization is done in the cone of
the diluter by two 500 yC Polonium-210 strips, mounted as shown.
The sample gas enters the dilution chamber at the apex of a
perforated cone into which clean, dry air is pumped through
the perforations, creating a highly turbulent mixing zone.
Calibration data for the diluter, Figure D4, is shown in
Figures D5 and D6. At a downstream point, after adequate mixing
has occurred, the diluted sample is extracted and conveyed to
the sizing instrument. This diluted sample passes through a
diffusional dryer where any remaining moisture is removed.
The major drying action is accomplished by using dilution air
which has been passed through an ice bath condenser.
258
-------
SAMPLE
IN
SAMPLE
OUT
3630-037
Figure D4. Sample Extraction Diluter, cut-away view.
259
-------
10,000
1,000 —
cc
o
o
g
i-
3
_l
5
LLJ
cc
O BASE AT POINT 1 - COLLISON
Q BASE AT POINT 1 - SPRAYER
OBASE AT POINT 2 - COLLISON
O BASE AT POINT 2 - SPRAYER
SAMPLE ORIFICE DESIGNATIONS
100
100 1,000
CALCULATED DILUTION FACTOR
Figure D5. Calculated dilution versus true dilution for the Southern
Research Institute Ultrafine Particle Diluter, 0.092 p.m
particles.
10,000
363O038
260
-------
10,000
I-
1,000
cc
o
o
I-
D
_l
5
UJ
cc
I-
100
O BASE AT POINT 1 - COLLISON
D BASE AT POINT 1 - SPRAYER
O BASE AT POINT 2 - COLLISON
O BASE AT POINT 2 - SPRAYER
,082
SAMPLE ORIFICE DESIGNATIONS
.059 , -042 . .029 . .oVl K
.014K
10
10
100 1,000
CALCULATED DILUTION FACTOR
10,000
363O-039
Figure D6. Calculated dilution versus true dilution for the Southern
Research Institute Ultrafine Particle Diluter, 0.15 ym
particles.
261
-------
Just prior to entering the sizing instrument, the sample
passes through a plenum to damp out short term concentration
changes. The exhaust from each instrument is returned to the
diluter to reduce pressure drops across the sizing devices.
If further drying of this recycled gas is necessary, absorption
driers are placed in the instrument exhaust lines.
Changing the sample air flow and the dilution air flow
allows one to change the dilution ratio. Sample air and di-
lution air flowrates are controlled by two bleed valves on the
dilution air pump, one upstream of the pump (#1) and one down-
stream (#2). Manipulation of these valves changes the internal
pressure of the diluter which, in turn, sets the sampling rate.
As the pressure in the diluter is reduced, the sample flow rate
is increased. In practice, the operation of these valves
changes the dilution air flow only about 10% for a many-fold
change in sample flow.
McCain5 has reported some problems with an earlier pro-
totype conditioning system; growth of particles when high con-
centrations of S03 are present, pluggage of orifices, single
point sampling limitations, and diffusional losses in sampling
lines. The configuration described above allows the use of
optional SO absorber chambers, rapid replacement for plugged
X
orifices, quick positive determination of partial pluggage,
full traverse sampling capability, and decreased sample line
losses. The hot box configuration also increases the total
time available for data acquisition by decreasing the time
needed to change orifices or dilution ratios (2 minutes compared
to about 30 minutes).
Different sampling and dilution systems have been developed
and reported by Ensor and Jackson,1 Bradway and Cass,6 and Schmidt
er al.7
262
-------
It i.s clear that the most difficult problem in making size
distribution measurements of submicron particles is extracting
a representative sample from the duct and conditioning it for
compatibility with the sizing instruments.
A variety of instruments are available to characterize
the conditioned sample, and these are described in the remainder
of this appendix.
D.2 SUBMICRON PARTICLE SIZING TECHNIQUES BASED ON PARTICLE
DIFFUSIVITY
D.2.1 Diffusion Batteries
Fuchs8 has reviewed diffusional sizing work up until 1956,
while Sinclair,9'10 Breslin e_t a_l, 1 : Twomey,12 and Sansone
and WeyellJ have reported more recent work, both theoretical
and experimental.
Diffusion batteries may consist of a number of long, narrow,
parallel channels, a cluster of small bore tubes, or a series
of screens. Variations in the length and number of channels
(tubes, or screens) and in the aerosol flow rate are used as
means of measuring the number of particles in a selected size
range. As the aerosol moves in streamline flow through the
channels, the particles diffuse to the walls at a predictable
rate depending on the particle size and the diffusion battery
geometry. It is assumed that every particle which reaches the
battery wall will adhere; therefore, only a fraction of the
influent particles will appear as the effluent of a battery.
It is only necessary to measure the total number concentration
of particles at the inlet and outlet to the diffusion battery
under a number of conditions in order to calculate the particle
size distribution.
263
-------
When the Stokes diameter is used to describe particle
size, the penetration of diffusion batteries is virtually
independent of physical properties of the individual aerosol
particles.
Parallel plate geometry - The parallel plate geometry is
convenient because of ease of fabrication and the availability
of suitable materials, and also because sedimentation can be
ignored if the slots are vertical, while additional informa-
tion can be gained through settling if the slots are horizon-
tal. See Figure D7. Disadvantages of the parallel plate dif-
fusion batteries are (1) the bulk of the diffusional batteries,
and (2) the long transport time required to measure a size
distribution.
The mathematical expression for the penetration of a
rectangular slot or parallel plate diffusion battery by a
monodisperse aerosol was given in series form by Gormley and
Kennedy.! "* The coefficients were calculated and tabulated
by Twomey12 using a computer.
By varying the number of diffusion batteries in series
and the flow rate, it is possible to measure penetrations under
a variety of conditions. Using a set of diffusion batteries
such as those developed at Southern Research (four-98 channel
diffusion batteries, and a 13 channel diffusion battery) and
measuring the penetration at three different flow rates, yields
fifteen data points from which the particle size distribution
(0.01-0.2 ym diameter) can be reconstructed.
When calculating the fraction (n/no) of the aerosol which
penetrates a series of diffusion batteries, the transport time
through the diffusion batteries must be taken into account.
This transport time is about 3^-5 minutes for a 98 channel,
264
-------
CHANNEL DIMENSIONS
MULTI CHANNEL BATTERY
3630-040
Figure D7, Parallel Plate Diffusion Battery.
265
-------
parallel plate diffusion battery. Thus, with four of these
diffusion batteries in series, n, at a particular time, would
be related to n0 at a point 16-20 minutes earlier in time on
a chart recording. Making a complete measurement of a particle
size distribution requires 2-4 hours and diffusional measure-
ments are most useful on stable sources where the distribution
is constant in time when using parallel plate diffusion bat-
teries. The transport time is determined by the open air volume
of the diffusion battery and flow rate. If the same geometry
constant can be obtained with smaller open air volumes, the
transport time can be reduced. This can be accomplished by
decreasing the plate spacing, but one very quickly approaches
problems with material flatness errors, or equipment surviv^-
bility when such materials as precision ground graphite sheets
are used. These problems seem to have been eliminated with
the screen geometry, which is discussed below.
Screen geometry - Sinclair10 and Breslin et all' report success
with more compact, tube-type and screen-type arrangements in
laboratory studies.
Although the screen-type diffusion battery must be cali-
brated empirically, it offers convenience in cleaning and ope-
ration, and compact size. Figure D8 shows Sinclair's geometry.
This battery is 21 cm long, approximately 4 cm in diameter,
and weighs 0.9 kg.
This system is commercially available from Thermosystems,
Inc., St. Paul, MN 55113 as the Model 3040 Diffusion Battery.
Because of the small internal volume of the battery, the time
necessary to obtain a test on one battery is reduced by about
a factor of ten as compared to the parallel plate batteries.
This diffusion battery system allows data to be collected on
process streams where the particle concentration is somewhat
unstable (^15 minutes per cycle).
266
-------
SAMPLING
PORT (TYP)
SECTION CONTAINING
SCREENS (TYPI
3630-045
Figure D8. Screen type diffusion battery. The battery is 21 cm
long, 4 cm in diameter, and contains 55 635 mesh
stainless steel screens.
261
-------
A disadvantage of the small volume is the sensitivity
to surges in flow rate caused by the commercially available
CN counter.
D.2.2 Particle Concentration Indicators-Condensation Nuclei
Counters (CMC)
Condensation nuclei counters function on the principle
that particles act as nuclei for the condensation of water
or other condensable vapors in a supersaturated environ-
ment. This process is used to detect and count particles in
the 0.002 to 0.3 micron range (often referred to as condensa-
tion or Aitken nuclei). In condensation nuclei detectors, a
Sample is withdrawn from the gas stream, humidified, and brought
to a supersaturated condition by reducing the pressure. In
this supersaturated condition, condensation will be initiated
on all particles larger than a certain critical size and will
continue as long as the sample is supersaturated. This con-
densation process forms a homogeneous aerosol, predominantly
composed of the condensed vapor containing one drop for each
original particle whose size was greater than the critical
size appropriate to the degree of supersaturation obtained;
a greater degree of supersaturation is used to initiate growth
on smaller particles. The number of particles that are formed
is estimated from the light scattering properties of the
final aerosol.
Because of the nature of this process, measurements of
very high concentrations can be in error as a result of a lack
of correspondence between particle concentration and scattering
or attenuation of light. Additional errors can result from
depletion of the vapor available for condensation. Certain
condensation nuclei measuring techniques can also obtain infor-
mation on the size distribution of the nuclei; that is, vari-
ations in the degree of supersaturation will provide size dis-
268
-------
crimination by changing the critical size for which condensation
will occur. However, MacLauren and Junge15 have predicted that
the critical size for initiating condensation is also affected
by the volume fraction of water soluble material contained in
the original aerosol particle, so the critical size will be
uncertain unless the solubility of the aerosol particles is
known. At very high degrees of supersaturation (about 400%),
solubility effects are only minor and essentially all particles
in the original aerosol with diameters larger than 0.002 \im
will initiate the condensation process.
A continuous flow CN counter has been described by Sin-
clair16 and an absolute calibration of a CN counter has been
done by Liu.17 The theory and principles of operation of CN
counters has been described by Haberl.18'19
D.2.3 Using Diffusion Batteries With Condensation Nuclei Counters
and aSample Extraction-Dilution System To Measure Concen-
tions of Submicron Aerosols In Industrial Flue Gases
Before taking equipment into the field, a preliminary exami-
nation of the sampling site should be made. There must be ample
space for the diffusion batteries (D,B,) and the condensation
nuclei counter (CNC). The sample lines to and from the D.B.'s
should be as short as is practical. The diffusion batteries
should be placed out of direct sunlight to reduce thermal inter-
ferences .
Four models of CNC's are presently used. Two automatic
models are the General Electric (GE)* and the Environment One
*General Electric - Ordnance Systems, Electronics Systems
Division, Pittsfield, MA 01201.
269
-------
Model Rich 100 (E-l).** Small manual particle detectors are
also commercially available from Gardner Associates*** and Environ
ment One.** The valving system in the GE is mechanical, and
pressure differentials across the valves are irrelevant. How-
ever, the E-l has pneumatic valves and a pressure difference
of greater than 2 inches of water across the valves will lock
them either open or shut, thus the return lines on the E-l CNC
must be connected to the sample extraction system to prevent
instrument malfunctions. A return line is normally used on
the GE but may not be connected to the diluter in some circum-
stances. The GE creates substantial pulsation in the sample
lines which is intolerable for diffusional analysis, and an
antipulsation device consisting of two metal cylinders connected
by a small orifice may be used as a pneumatic R-C network to
damp the oscillations to an acceptable level.
Once all sample and return lines have been connected, the
equipment is turned on and allowed to warm up. The flows are
then adjusted to the proper rates. If the GE model is used, the
vacuum gage on the front panel should read 8 inches of Hg as
recommended in the operation manual. This gives a nominal
flow rate of 6 &pm and a sample supersaturation after expansion
of approximately 400%. The E-l may be adjusted,to any desired
flow rate between about 0.6 tpm and 4.2 jlpm.
The water supplies for the humidifiers are filled with
a mixture of distilled water and a wetting agent. About 0.5%
Kodak photoflow (ethylene glycol) is used in the water. This
reduces the surface tension of the water and allows the wick
to wet better and more quickly.
**Environment-One Corporation, Schenectady, NY 12301
***Gardner Associates, Schenectady, NY 12301.
270
-------
Before any data can be taken the system should be leak
checked. This is done by connecting all the D.B.'s in series,
clamping one end off and pumping several inches of mercury
vacuum. If the vacuum holds data can be taken. If there is
a leak it can be isolated by repeating the process with suc-
cessively fewer D.B.'s until the leaky one is found. The exact
location of the leak can be determined by putting a concentrated
condensation nuclei source (e.g., a burning string) close to
various parts of the D.B. When the leak is found, the CNC panel
meter will rapidly rise.
There are two methods of data collection: graphing the
CNC output on a strip chart recorder or directly writing down
the meter reading. In general, a combination of the two is
used to insure that no faulty connection exists between the
CNC and the chart recorder or that the chart recorder is mal-
functioning .
Since with either the GE or the E-l CNC the largest flow
possible is 6 £pm, a certain minimum amount of time is required
to pull the sample through the parallel plate D.B.'s. If gra-
phical techniques are used a characteristic output will be
observed. The indicated concentration is zero while the clean
air already in the D.B.'s is being exhausted. Then the output
rises to a peak and stabilizes. It is at this point that mean-
ingful data is being taken. However, if data is taken by meter
readings alone this characteristic response is very difficult
to follow. Adequate time must be allowed for transport through
the D.B.'s before each reading is taken. These transport times
have been calculated for 6 fi,pm for two types of parallel plate
D.B.'s used: Thirty seconds for the 13 channel parallel plate
and five minutes for each 98 channel parallel plate D.B. used.
Quite frequently the in-stack concentrations are unsteady and
data is normally recorded for several minutes after the system
has had time to stabilize as a double check on the validity
271
-------
of the data and to provide some time averaging of the data on
fluctuating sources. With the screen type diffusion batteries,
transport times of about 1/10 of the times given above can be
used.
If the GE CMC is used, diffusional sizing cut points are
selected by changing the number of D.B.'s, since the flow rate
is constant. However, with the E-l, the cut sizes may be ad-
justed by either changing the number of D.B.'s or the flow rate.
The flow rate is sometimes held constant because source fluctua-
tions introduce so much uncertainty that attempts to achieve
high resolution are futile.
In order to obtain a set of data, the CNC is first connect*
directly to the diluter to obtain total concentration. Data
is then taken by pulling the sample next through a single battei
two batteries, three batteries, etc., until all the available
permutations of geometry and flow rate have been used.
If only one set of equipment is available, after data
are collected at the inlet the equipment is carried to the
outlet, set up, leak checked, and the above procedure is re-
peated .
D.2.4 DataReduction Techniques and a Sample Calculation
Fuchs et al2 ° presented a technique for calculating the
particle size distribution from raw data, assuming that the
size distribution was log normal. A technique suggested by
Sinclair9 does not include this restriction. In Sinclair's
method a nomograph is prepared using the penetration for each
diffusion battery geometry and flow rate and a large number of
monodisperse particle sizes. Comparing this nomograph with
experimental penetrations, one calculates the particle size
272
-------
distribution using a "graphical stripping" process. However,
it is usually more convenient to use a "D50" technique like
that used for the reduction of cascade impactor data.
Data for a given test condition are averaged as follows
to yield a representation of the source at that condition.
All instrument readings are converted to indicated concentra-
tions by means of individual instrument calibration curves.
These values are then corrected for dilution to obtain flue
gas concentration. Since the CNC calibration curve is non-lin-
ear, multiplying by the dilution factor before converting to
concentration will yield erroneous results. Next, process
averaging is accomplished by taking appropriately weighted
averages of the data obtained through the various process cy-
cles. These weighted averages are found for each D.B. arrange-
ment, including no D.B.'s, and test condition, both inlet and
outlet.
For the sample calculation it is assumed that five para-
llel plate diffusion batteries were used, in four configura-
tions, and at three flow rates to obtain data at twelve D$0 sizes
These five diffusion batteries consisted of one 13 channel (Type
A) and four 98 channel (Type B) units. The four sampling con-
figurations were 1) one Type A, 2) one Type B, 3) two Type B
in series, and 4) four Type B in series. These diffusion bat-
teries are similar to the one depicted in Figure D7. The three
flow rates were 1, 6, and 10 liters per minute. The aerosol
was sampled from a Sample Extraction and Dilution System and
the total number of particles entering and exiting the diffusion
batteries was determined using a condensation nuclei counter
(CNC). The sample data for this experiment is shown in Table
Dl.
To calculate the particle size for 50% penetration (the
Dso) through a diffusion battery configuration, the following
equations must be used.
273
-------
The penetration of a rectangular plate diffusion battery
is given by
P = \ = [o.9104e-1'88522DY + O.C531e-21'43lDY
M'
+ 0.015e-62'317DY + 0.0068e-124-537DY] (Dl)
where D = Particle diffusivity, cmVsec,
Y = Diffusion battery flow rate-geometry constant/
sec/cm 2,
m = Number of identical batteries in series,
n0 = Diffusion battery inlet concentration, #/cm3, and
n = Diffusion battery outlet concentration, #/cm3.
The particle diffusivity, D, is given by
D = kTB , where (D2)
k = Boltzmann's Constant, gm cm2/sec2 °K,
T = Absolute Temperature, °K, and
B = Particle Mechanical Mobility, sec/gm.
The particle mechanical mobility is given by
B =/l + 2,49 (X/d) + .84 (X/d)e~(-44) (d//XJ//3Tryd (D3)
where X = gas mean free path, cm,
d = particle diameter, cm, and
y = gas viscosity, gm/sec cm.
The gas mean free path is given by
X = 3.109 x 10~ H VT/M where (D4)
274
-------
TABLE Dl
DIFFUSION BATTERY SAMPLE CALCULATION DATA
CNC
Configuration
1 Type A
1 Type B
2 Type B
4 Type B
1 Type A
1 Type B
2 Type B
4 Type B
1 Type A
1 Type B
2 Type B
4 Type B
(JL/min)
Calc.
(urn)
o Reading
(#/cm3)
Dilution
Factor
6
6
6
6
1
1
1
1
10
10
10
10
0.02
0.06
0.098
0.17
0.028
0.07
0.12
0.20
0.015
0.045
0.08
0.14
53000
43000
45000
11000
84000
33600
29000
5500
64000
62000
66500
20500
1000
500
200
200
500
500
200
200
1000
500
200
200
Actual
Concentration
(ft/cm3)
53.0
21.5
9.0
2.2
42.0
16.8
5.8
1.1
64.0
31.0
13.3
4.1
10
275
-------
P = Ambient pressure, atm,
T = Absolute temperature, °K, and
M = gas mean molecular weight, gm/gm-mole.
The approximate gas viscosity is given by
= [(.495 (T-294°K) + 182)] x 10~6 gm/cm sec, (D5!
The diffusion battery flow rate-geometry constant, Y, is
given by
4LhN , ._..
where (D6)
L = Channel length, cm,
h = Channel height, cm,
N = Number of channels,
Q = Flow rate/ cm3/sec, and
W = Channel width, cm.
See Figure D-7.
The two diffusion batteries under consideration in this
example have the following dimensions.
Type A Type B
L 45.72 cm 45.72 cm
h 10.15 cm 11.46 cm
W 0.10 cm 0.10 cm
N 13 98
Thus, Ya = 2.358 x 103 sec/cm2 and YQ = 2.006 x 10" sec/cm2
n O
for Q = 6 Jl/min.
276
-------
The manipulation of these equations allows one to calculate
the penetration of the diffusion batteries at different flow rate
and particle-size combinations. After plotting the penetration
versus particle size, the D50's can be determined.
To aid in the calculation of the particle diffusivity
and penetration of a particular D.B. arrangement, two programs
have been written for the Hewlett-Packard HP-65 Program-
mable Calculator. If a programmable calculator is not
available, the calculations can be done manually using
the equations given above. The calculator programs
and their applications are described below:
Program 1 is used to calculate the viscosity (y) and
mean free path (A) of standard air. These values are then
used to calculate the diffusivity (D) for a monodisperse
aerosol having diameter d. Given values for the flow rate-
geometry configuration (y) and the diffusivity (D), Program
2 is used to calculate the theoretical penetration (n/n<,).
Program 1; Diffusivity (Program Steps are Listed in Table
D2.)
Over the temperature range from 0-350°C, the viscosity
of dry standard air is very nearly linear. For a given
temperature y can be found, in CGS units, from:
y = [.495 (T - 294° K) + 182] x 10~ poise
Knowing the viscosity, the mean free path is given
by
X - 5.77^/T x 1CT2 cm for standard air (P is the absolute
pressure, atm.).
277
-------
From a knowledge of p and A for the carrier gas, the
diffusivity (D.) of a monodisperse aerosol having diameter
(d.) is given by the following equation:
D. = (1.46 x 10-") -^ [l + 2.49 (\ ) + .84 (| ) e'(' 44> ]
for T in °K, M in poise, ^ and d in cm.
User Instructions. Enter the program shown in Table D2.
To calculate u and A for Standard Air having 250°K £. T _<
600°K:
-1. Load storage registers with the following variables
Temperature T, °K, STO 2
Absolute Pressure, P, "Hg, STO 3
2. Start program "y, A" "B"
3. Output:
Values of y are stored in the correct storage register
for retrieval during calculation of D
To display M (units of poise) RCL 4
To display A (units of cm) RCL 5
To calculate D. for a monodisperse aerosol
4. Load storage registers with the following variables
Temperature, T, K, STO 2 Performed
Viscosity, u, poise, STO 4 in Steps
278
-------
TABLE D2
PROGRAM 1 - PARTICLE DIFFUSIVITY
HP-65 CALCULATOR PROGRAM FOR A PARALLEL PLATE DIFFUSION BATTERY
CODE
LBL
A
RCL 1
RCL 5
-r
-
4
3
5
CHS
X
i1
LM
RCL 5
RCL 1
-r
STO 8
X
•
8
4
X
1
+
RCL 8
2
4
9
2
X
+
RCL 2
X
1
4
KEYS
23
11
34 01
34 05
81
83
04
03
05
42
71
32
07
34 05
34 01
81
33 08
71
83
08
04
71
01
61
34 08
02
83
04
09
02
71
61
3402
71
01
04
CODE
6
EEX
CHS
1
9
X
RCL 4
-r
RCL 1
-f
STO 7
RTN
LBL
B
4
9
5
RCL 2
2
9
4
—
X
1
8
2
4
EEX
CHS
6
X
STO 4
RCL 3
-r
RCL 2
KEYS
06
43
42
01
09
71
3404
81
34 01
81
33 07
24
23
12
83
04
09
05
34 02
02
09
04
51
71
01
08
02
61
43
42
06
71
33 04
34 03
81
34 02
CODE
f
V
X
•
0
5
7
7
X
STO 7
RTN
KEYS
31
09
71
83
00
05
07
07
71
33 05
24
R! d R4 /i R7 P
R2 T RB 1 RS (WORK)
R3 p Rg (BLANK) Rg (BLANK)
279
-------
Mean Free Path, X, cm, STO 5 1 and 2
Diameter of the
monodisparse
aerosol, d., cm, STO 1
5. Start program "Diff" "A"
6. Output:
Displayed value is "D" in cgs units (also stored in
register 1) .
Test Problem
Find y, X, and DI for T = 75.2°F, P = 14.39 PSIA, and
d = .023 microns.
1. Convert T, P, and d. to the proper units then store
in registers 2, 3, and 1, respectively.
75.2°F = 24°C = 297°K, store in register 2
14.39 PSIA x 2.036 p|^L = 29.30 "Hg, store in
register 3
.023 ym = .023 x 10~" cm = 2.3 x 10~6 cm,
store in register 1
2. Start Program B. X is displayed (\ = 6.23 x 10~f
cm) , RCL 4 to display (\i - 1.83 x 10~k poise} .
3. Start Program A. Di is displayed (D. = 9.95
10~5cm2/sec) .
280
-------
2: Theoretical Penetration (Program Steps are Lj.sted in Table
D3.)
When the diffusivity (D.) of a monodisperse aerosol
of diameter d. is known, and the flow rate-geometry configu-
raion (y.) for the diffusional apparatus (diffusion battery)
are known, fractional penetration as a function of size
(D.) is given by
^ + 0.0531e-21-431Di*j + 0.0531e
-62.317Diy,
J + 0.068e
If several diffusional configurations are used in series
where the aerosol is allowed to remix before going through
the next configuration, the final penetration is the product
of the penetration for each, thus
If m diffusion batteries having the same y. value are
used in series, then
To calculate p. . for m identical batteries in series,
enter program in Table D3.
28].
-------
TABLE D3
PROGRAM 2 - PARTICLE PENETRATION
HP-65 CALCULATOR PROGRAM FOR A PARALLEL PLATE DIFFUSION BATTERY
CODE
LBL
A
RCL 7
RCL 6
X
STD 8
1
-
8
8
5
2
CHS
X
f"1
LN
9
1
0
4
X
RCL 8
2
1
4
3
1
CHS
X
f>
LN
0
5
KEYS
24
11
34 07
34 06
71
33 08
01
83
08
08
05
02
42
71
32
07
83
09
01
00
04
71
34 08
02
01
83
04
03
01
42
71
32
07
83
00
OS
CODE
3
1
X
-i-
RCL 8
6
2
-
3
1
7
CHS
X
f
LN
0
1
5
3
X
+
RCL 8
1
2
4
5
3
7
CHS
X
f-1
LN
0
KEYS
03
01
71
61
34 08
06
02
83
03
01
07
42
71
32
07
83
00
01
05
03
71
61
3408
01
02
04
83
05
03
07
42
71
32
07
83
00
CODE
0
6
8
X
RCL
9
9
yX
RTN
KEYS
00
06
08
71
61
34
09
35
05
24
Hi (BLANK)
R2
(BLANK)
R3 p
R4
RB
Re
(BLANK)
(BLANK)
Y
R?
RS
RB
D
(WORK)
m
282
-------
].. Load storage registers with the following variables
YJ'
Diffusivity,
Flow rate-geometry
configuration
Number of identical m,
batteries in series
cgs units,
cgs units,
integer,
STO 7
STO 6
STO 9
2. Start Program
•A"
Displayed value is the fractional penetration (P- •)
1' J
of this monodisperse aerosol of size d. through n
identical diffusion batteries in series.
Test Problems
1. Load data
y. = 2.36 x 103 in cgs units
D. = 5.39 x 10~6 in cgs units
P. ,
STO 6
STO 7
STO 9
2. Calculate P "A"
P = 9.38 x 10 l = 93.8%
Therefore the penetration of the diffusion battery Type
A at a flow rate of 6 liters/minute is 93.8% for 0.023 urn
particles.
After calculating the penetration, a graph similar
to Figure D9 will be obtained. The D50's can be determined
from this graph as indicated. The experimental data for
this sample calculation are shown in Figure Dll.
283
-------
100
90
80
70
35
Z 60
o
< 50
DC
S 40
H
LU
"• 30
20
10
0
0.
01
0,02 0.03 0.04 0.05 0,1
PARTICLE DIAMETER,
0
10
20
30
40 §
P
50 <
60
70
80
90
100
0.2
0.3 0.4 0.5
363O042
Figure D 9. Theoretical parallel plate diffusion battery penetration
curves.
284
-------
A cumulative particle size distribution is plotted using
the corrected concentration from the last column in Table Dl.
as the ordinate and the DEC'S for each configuration as the
abscissa. Differential number graphs ace obtained by dif-
ferentiation of the cumulative curve (finite differences) as
described in Sections C.8 and C.9.
D.3. SUBMICRON PARTICLE SIZING TECHNIQUES USING THE ELECTRICAL
MOBILITY PRINCIPLE
D<3'1 Electrical Mobility
If a particle charge is known, measurements of the elec-
trical mobility are sufficient to determine the particle size.
This concept has been used by Liur et aJL,'1 at the University
of Minnesota to develop a series of Electrical Aerosol Analyzers
(EAA). A schematic of this device is shown in Figure DID,
The EAA has the distinct advantage of very rapid data acquisi-
tion compared to parallel plate diffusion batteries and con-
densation nuclei counters (two minutes as opposed to two hours
for a single size distribution analysis.
D.3.2 Thermosystems Model 3030 Electrical Aerosol Size Analyzer*
The EAA is designed to size particles in the range
of 0.01 micrometer diameter to 1.0 micrometer diameter.
It can size solids and non-volatile liquids. The concen-
tration of 1 to 1000 ug/m3 limit its application to a sample
extraction-dilution system as described earlier.
*Thermosystems, Inc., St. Paul, MN 55113,
285
-------
ro
CD
.V.VC. ^?SVM
r
il
|J
ill
i
CONTROL MODULE ,. . .
POSITIVE HIGH Ofl'tt HFflD COMMAND - - - - — *• DATS
, VOLTAOl SUI""LI
...j I
COKSTMUT
j J CONTROL
i1'] — -iF
;J L^ VOLTAGE M
J DiViOtB |^
NFGflTlVL MICH J
VUl TflGf SUPPO
L
— N E — ^ HtCTP
ly.ww- •/•••' if ,n' -Ml
L
<~4 • • • &CAO&
-1- CHAP
1 AUTOMATIC Hd
3L FLOWMCTtP BLflDOUT
H tfOLTfltJt COWTROL dH^ RtflOOUl
f
TOTAL rLow L
CONTROL Q<
(8A. » VAUC I >
l--HLzJ!
TOTfiL '1
fLGwMCTER | '
1
OMITEB -, H
AEROS
CUP
COLL
f IL
"xXNM
•_rrr:_ *" TO VACUUM PUMP
N
BINT ^ ••;"•::::;:.".•;::: '
L^
H
J
i
r^
3fi30 043
Figure DID. Schematic diagram of eiectrica! aerosol analyser.
After Liu and Pui,
-------
The EAA operated in the following manner. As a vacuum
pump draws the aerosol through the analyzer (See Figure Did),
a corona generated at a high voltage wire within the charg-
ing section gives the sample a positive electrical charge.
The charged aerosol flows from the charger to the
analyzer section as an annular cylinder of aerosol surround-
ing a cone of clean air. A metal rod, to which a variable,
negative voltage can be applied, passes axially through the
center of the analyzer tube. Particles smaller than a cer-
tain size (with highest electrical mobility) are drawn to
the collecting tod when the voltage corresponding to that size
is on the rod. Larger particles pass through the analyzer
tube and are collected by a filter. The electrical charges
on these particles drain off through an electrometer, giving
a measure of current.
A step increase in rod voltage will cause particles of
a larger size to be collected by the rod with a resulting
decrease in electrometer current. This decrease in current
is related to the additional number of particles being col-
lected. A total of eleven voltage steps divide the 0.0032
to 1.0 micron size range of the instrument into ten equal
logarithmic size intervals. Different size intervals can
be programmed via an optional plug-in memory card.
The electrical aerosol analyzer can be operated either
automatically or manually. In the automatic mode, the ana-
lyzer steps through the entire size range. For size and con-
centration monitoring over an extended period of time, the
analyzer may be intermittently triggered by an external timer.
The standard readout consists of a digital display within the
control circuit module, although a chart recorder output is
287
-------
available. It is almost always advantageous to use
a strip chart recorder to record the data. This allows
the operator to identify a stable reading superimposed on
source variations and gives a permanent record of the raw
data.
D.3.3 Using The TSI Model 3030 Electrical AerosolAnalyzer
With a Sample Extraction-Dilution System To Measure
Concentrations of^ Submicron Aerosols in Industrial
Flue Gases
Once the equipment is set up as shown in Figure D3, the
flows are adjusted through the sample orifice and the dilution
air orifice, to obtain the desired dilution factor. The EAA
is placed in a manual mode and the current readings for each
channel are recorded with a strip chart recorder. Manual con-
trol allows run times of from two to five minutes in each of
the nine channels. This allows one to average out rapid source
fluctuations. At the beginning of each day, the interval cali-
bration points and flows through the EAA are checked, as de-
scribed in the instrument manual. These are periodically re-
checked throughout the day.
D.3.4 DataReduction Techniguesand a Sample Calculation
It is assumed that a Thermo-Systems Inc. Model 3030 Elec-
trical Aerosol Size Analyzer (EAA) with a 0.0032 ym to 0.360
ym range at the normal operating conditions has been used to
determine concentration ys size information in the ultrafine
size range for the effluent of a precipitator. The EAA sampled
the gas stream after the sample was extracted with a Sample
Extraction and Dilution System as described in Section D.I.7.
288
-------
TABLE D4
EAA (Model 3030) Data Reduction Form
Concentration, Cumulative Concentration, and AN /ALogD from Scan No
1
Channel
No.
3
4
M 5
CD
VC
6
7
8
9
10
11
2
Collector
Voltage
196
593
1220
2183
3515
5387
7152
8642
9647
3
D , ym
P
0.0100
0.0178
0.026
0.036
0.070
0.120
0.185
0.260
0. 360
4
D . , ym
P1
0.0133
0.0215
0.0306
0.0502
0.0917
0.149
0.219
0.306
for DF ~
5
AN/A I
4. 76x10 5
2. 33x10 5
1. 47x10 5
8. 33x10"
4.26x10"
2.47x10"
1.56x10"
1. 10x10"
6
AlogD
0.250
0.165
0.141
0.289
0.234
0.188
0.148
0.141
7 8 9 10 11 12
I,pA AI,pA AN AN EN AN /AlogD
5 5 ^
-------
The EAA was placed in a manual scan mode and the current
readings for each channel were recorded with a strip chart re-
corder. Manual control allowed run times of from two to five
minutes in each of the nine channels. This allowed the averag-
ing of rapid source fluctuations.
The theory of operation and basic equations for the EAA
have been given by Liu et al21 and calibration of the Model
3030 EAA has been done by Liu and Pui.22 The latest cali-
bration and modifications revise the initial EAA calibration
which assumed eleven size fractions from .0032 (jm to 1.0 urn,
but which was actually eleven size fractions from 0.0032 ym
to 0.36 ym (as usual in this example). The mechanics of the
data reduction is identical, however. Table D4 shows the pre-
vious calibration constants in a data reduction format. The
calibration by Liu21 suggested the use of a calibration
matrix; however, typical source fluctuations in industrial
processes generally negate any potential advantage of such
refinements. Table D4 is essentially self-explanatory. The
heading "D , ym" (Column 3) is the particle diameter in micro-
meters. A value of 0.100 \im diameter and smaller are col-
lected in the analyzer tube while larger particles pene-
trate to the current collecting filter where an electrometer
measures the total current carried by the unprecipitated par-
ticles. This current represents the charges on all particles
larger than 0.100 ym. This measured current is the basic output
of the Model 3030.
The fourth column (D ., ym) is the geometric mean diameter
of the particles represented by the current difference of
two successive steps (Channel No.'s). For example, the
difference in current for the 0.100 ym cut-off and the current
for the 0.0178 ym cut-off is the total current collected
from particles between these sizes, or rather for a mean
diameter of 0.0133 ym. The current differences are entered
in Column 8 headed " I, pA" (picoAmps).
29C
-------
B. Calculate all dilution factors (DF.).
STEP 2
Calculate current differences (A I. .) from adjacent chan-
1 r ~]
nels and average the a. products (a. = AI. • x DF.) foe the
] i i / J J
same size band for all scans taken for the same test conditions.
Calculate 90% confidence intervals for each a^. Note: the
i subscript denotes size and the j subscript denotes dilution
setting.
STEP 3
Using a. and Table D6 calculate "number concentration"
(AN ), "average cumulative concentration of all particles having
S
diameter greater than the indicated size" Z (AN ), and"AN /ALogD"
s s
for each size band for each test condition.
STEP 4
Plot "Cumulative Concentration vs. Size" for each test
condition.
STEP 5
Plot AN /ALogD (with upper and lower 90% confidence limits!
S
vs. size for each test condition.
SAMPLE CALCULATION FOLLOWING THE CALCULATION FORMAT
Table D5 contains hypothetical test data for the follow-
ing sample calculation.
291
-------
NJ
TABLE D5
EAA Current Readings (I, in picoamps and Dilution Factors)
for this Sample Calculation: Hypothetical Inlet Data
SCAN Time
1 1=30P
2 1:32
3 1:34
4 1:36
5 1:38
6 1:40
7 1:45
8 1:47
9 1:49
10 1:51
CH 3
2.869
2.835
2.641
2.859
2.866
2.866
6.477
6.580
6.377
6.390
CH 4
2.734
2.711
2.709
2.722
2.740
2.736
6.188
6.2B8
C.087
6.094
CH 5
2.519
2.495
2.500
2.522
2.530
2.531
5.716
5.818
5.620
5.614
CH
2.
2.
2.
2.
2.
2.
5.
5.
4.
4.
6
227
205
200
235
251
238
056
153
960
956
CH 7
1.362
1.344
1.340
1.368
1.381
1.378
3;m
3.233
3.021
3.006
CH a
.682
.669
.655
.676
.714
.698
1.575
1.613
1.526
1.467
CH 9
.242
.220
.21B
.226
.279
.255
.565
.510
.537
.492
CH 10
.102
.075
.081
.096
.137
.115
.243
.195
.227
.187
CH 11
.020
- .010
.001
.010
.052
.033
.053
.010
.032
.005
Dilution Factoi
255
255
255
2S5
255
253
113
113
113
113
-------
TABLE D6
EAA (Model 3030) Data Reduction Form
Concentration, Cumulative Concentration, and AN /ALogD
From Average a for Condition s
1
Channel
No.
3
4
5
6
7
8
9
10
11
2
Collector
Voltage
196
593
1220
2183
3515
5387
7152
8642
9647
3
D f um
P^
0.0100
0.0178
0.026
0.036
0.070
0.120
0.1B5
0.260
0.360
4
D . , ]im
pi
0.0133
0.0215
0.0306
0.0502
0.0917
0.149
0.219
0.306
5
AN/AI
4.76X105
2. 3 3x10 5
1.4 7x10 5
8.33x10*
4.26x10"
2.47x10"
1.56x10"
1. 10x10"
6789 10
AlogD a AN EAN AN /AlogD
y p^ s s s' y
0.250
0.165
0.141
0.289
0.234
0.188
0.148
0.141
-------
The fifth column gives the revised calibration factor
(based on the calibration by Liu and Pui22) for each of
the eight size bands. These factors are in units of particles
per cm per picoAmpere. Multiplying this size specific
current sensitivity, AN/AI, (Column 5) by the current dif-
ference, AI, (Column 8) gives the total number of particles,
AN, (Column 9) in units of particles per cm3, within this
size band (Column 4) for the diluted aerosol. To correct
for dilution and find in-stack concentrations, multiply
•Column 9 by the dilution factor (DF) and enter the result,
AN , in Column 10. Columns 6 and 12 are used for AN /ALogD
S 5
information calculated from the number distribution in Column
10. Column 11 is used for cumulative concentrations, correc-
ted for dilution to engineering standard (normal) conditions
by a dilution factor (i.e. Column 10). Engineering standard
or normal conditions are defined as dry gas at 21°C and
760 mm Hg pressure.
The basic data from the EAA is cumulative current for
each of nine channels (Column 7). One must then take the
differences of the current readings for successive channels
(Column 8) in order to find AN, etc. These I values are
multiplied by a series of constants (AN/AI. , DF.) to arrive
at AN (concentration in stack corrected to dry, standard
5
conditions). While a single scan should be made at a constant
dilution, different scans may be made at different dilutions.
To simplify the arithmetic for each test condition,
the product a- = AI • . x DF . is formed and all such inlet
1 1 r J J
(outlet) products for the same size band are averaged.
SUMMARY OF THE CALCULATION FORMAT
STEP 1
A. Calculate the average instrument reading (I) for each chan-
nel as obtained from the strip chart recording of channel cur-
rent vs. time.
294
-------
STEP 1
A. Calculate the average instrument reading (I) for each chan-
nel as obtained from the strip chart recording of channel cur-
rent vs. time. Each complete size scan (Table D5) consists
of nine instrument readings (I, Column 7 of Table D4), These
instrument readings are the average current outputs as taken
from the strip chart recordings, for each of the nine channels.
Run times were manually controlled and varied from two to ten
minutes per channel as the instrument operator sequentially
stepped through channels 3, 4, 5, . .., 11. Table D5 gives the
instrument readings used as data for the sample calculation
(10 scans, 90 average current readings).
B. Calculate all dilution factors (DF.; corrected to engineer-
ing standard (normal) conditions: 70°F (20°C) and 29.92 inches
of mercury pressure (760 mm Hg)).
STEP 2
Calculate current differences (AI. -) from adjacent chan-
nels and average the a. products for the same size band for
all scans taken at the same test condition. Calculate 90% con-
fidence intervals for each a.. (Refer to Section C.8.5),
The a. product is given by the following:
a . = AI. . x DF.
i 10 1
where i denotes the size band and j denotes the dilution value.
For channels 3-4 we have:
Scan #1: a,_, , = (.135) (255) pA
j 4 , J.
#2: a3_4/1 = (-124)(255) pA
#3: «3_4fl = (.132) (255) pA
295
-------
#9: a3_4f2 = (.290)(113) PA
#10: a3_4f2 = (.296)(113) pA
thus a3_4 = 33.179 pA; n = 10 and CI = .579.
In a similar manner we can find of. c» ^c_cf ...» oT-,n _-,-,.
Thus the mean, with upper and lower 90% confidence limits for
a-,, is given by:
"a3_4 = (33.179 + 0.579) pA
or
a_ = (33.2 + 0.6) pA
STEP 3
Using a". and Table D6 calculate "number concentration"
, "average cumulative concentration . . ." (ZANg) , and
for each size band for each test condition.
Table D7 shows these calculations for the sample data of
Table D5 . Column 7 is « as shown in Step 2. Column 8 is the
product of columns 7 and 5. Column 9 is the summation of 8
for all sizes "equal to or greater than the indicated size."
Column 10 is column 5 times column 7 divided by column 6.
STEP 4
Plot cumulative concentration vs. size for each test con-
ditions. For the sample data set of Table D5 this would be
the concentrations in Table D7 column 9 plotted against the
sizes in column 4. No errors bars are used.
296
-------
TABLE D7
EAA (Model 3030) Data Reduction Form
Concentration, Cumulative Concentration, and AN /ALogD
From Average AI for Condition Inlet
(Sample Calculation)
to
1
Channel
No.
3
4
5
6
7
8
9
10
11
2
Collector
Voltage
196
593
1220
2183
3515
5387
7152
8642
9647
3
D , ym
0.0100
0.0178
0.026
0.036
0.070
0.120
0.185
0.260
0.360
4
D . , urn
pi
0.0133
0.0215
0.0306
0.0502
0.0917
0.149
0.219
0.306
5
AN/AI
4.
2.
1.
8.
4.
2.
1.
1.
76x10 5
33xl05
47xl05
33x10"
26x10"
47x10 ^
56X1011
lOxlO1*
6 7
A log D a
IT
0.250
0.165
0.141
0.289
0.234
0.188
0.148
0.141
33.2+. 6
53.3+. 7
74.3+. 8
219. 8+. 8
174+2
114+2
35.4+. 6
21.2+.3
8
AN
s
xlO6
15.8+. 3
12.4+.2
«H
10.9+.1
18.3+.1
~
7. 41+. 09
2. 82+. 05
.552+. 009
.233+. 003
9
IAN
s
xlO6
68. 4
52. 6
40.2
29. 3
11.0
3.61
.785
.233
10
AN
s
xlO
63.
75.
77.
63.
31.
15.
3.
1.
/ALogD
G
2 + 1.1
3+1.0
5+.S
4 + . 2
*"*
7+. 4
~
0+. 3
~
73+. OG
65+. 02
-------
STEP 5
Plot ANs/ALogD with upper and lower 90% confidence limits
for each test condition.
For the sample data set of Table D5 this would be the con-
centrations in Table D7, column 10 plotted against the sizes
in column 4. The upper error bar is the value plus the 90%
confidence interval. The lower error bar is the value minus
the 90% confidence interval. For cU_4 in Table D7 we would
have o3_4 = 33.2 + 0.6
thus:
AVt /AT „ 33.2 x 4.76 x 10 , 0.6 x 4.76 x 10
ANs/ALogD = -^ ± —
= (63.2 + 1.1) x 10
6
The data shown in Column 10, Table D7 is graphically displayed
in Figure D-ll.
298
-------
102
Q
d
z
cc
t-
z
LU
o
z
o
u
cc
LU
CQ
LU
D
10L
HYPOTHETICAL DATA
• PARALLEL PLATE DIFFUSION BATTERIES
O ELECTRICAL AEROSOL ANALYSER
O
•
io-il_
10-2
ID'1 10°
PARTICLE DIAMETER, micrometers
101
3630-044
Figure D 77, Hypothetical inlet size distribution at an ESP on a
coal-fired boiler. Sample data for parallel plate
diffusion batteries and electrical aerosol analyser
are shown.
299
-------
D.4 OPTICAL PARTICLE COUNTERS
Optical parti-cle counters are a useful adjunct to the
ultrafine sizing system, giving additional data in the region
where impactors, diffusional and electrical methods have the
worst resolution.
A number of commercial optical particle counters are avail-
able which will perform adequately in the field environment
associated with precipitator evaluation. These instruments
can be obtained with several optical sensor configurations
and may usually be specified to perform in a threshold mode,
where all particles larger than a certain size are counted,
or in a window mode where a narrow range of particle sizes
is indicated. In either instance, the smallest particle
which will be detected is approximately 0.3 ym diameter.
In this respect, most optical counters size particles in
the same range as cascade impactors and cyclones.
The calibration of optical particle counters is usually
done with monodisperse plastic spheres which have a refractive
index of 1.59. Any deviations from this index of refraction,
or sphericity of the particles will cause the indicated sizes
to be in error. If high accuracy in the optical particle counte
data is important to the test, it is possible to do calibrations
in the field, using the test aerosol. McCain5 has used a sedi-
mentation technique which employs parallel plate diffusion bat-
teries as sedimentation chambers to correlate aerodynamic par-
ticle diameter with the indicated, or equivalent PSL, diameter.
Marple23 reported the development of special impaction devices
which may be used for this same purpose.
In the majority of full sampling situations, optical par-
ticle counters must be used with some type of Sample Extraction
and Dilution System as described in Section D.I. The sample
300
-------
is thus cooled and diluted to less than 300 particles per cm1.
Also it is usually necessary to place the return line back into
the dilution system to minimize pressure gradients across the
counter system.
Most commercially available systems come with real-time
analog outputs and digital outputs with one to 10 minute ac-
cumulation or integration times. Analog signals are particu-
larly useful for monitoring source variations and rapping emis-
sions.
Data accumulated by optical particle counters are given
as particles per unit volume and, after multiplication by the
proper dilution factor, may be plotted on a cumulative or dif-
ferential number basis. See paragraph C.8 for a discussion
of data plotting. Figure 35 in Section 2.3.4. shows data taken
using an optical-diffusional impactor system.
301
-------
REFERENCES
1. Ensor, D. S. and B. S. Jackson, "Evaluation of a Particu-
late Scrubber On A Coal-Fired Utility Boiler," EPA-600/2-
75-074.
2. Sem, G. J., "Submicron Particle Size Measurement of Stack
Emissions Using The Electrical Mobility Technique," pre-
sented at the Electric Power Research Institute Workshop
on Sampling, Analysis, and Monitoring of Stack Emissions,
October (1975).
3. Spink, L. K. and D. M. Considine, Process Instruments and
Controls Handbook, McGraw-Hill, 1957.
4. Parry, E. P. and R. A. Meyer, "Gas Flow Measurements in
Air Pollution Monitoring," presented at the 67th Annual
Meeting of the Air Pollution Control Association, Denver,
CO (June 1974) .
5. McCain, J. D., K. M. Gushing, and W. B. Smith, "Methods
for Determining Particulate Mass and Size Properties:
Laboratory and Field Measurements," J. Air Pollution Con-
trol Association, 12_, 1172-76 (1974) .
6. Bradway, R. M. and R. W. Cass, "Fractional Efficiency of
A Utility Boiler Baghouse, Nucla Generating Plant,"
EPA-600/2-75-013a (August 1975) .
7. Schmidt, E. W., J. A. Greseke, and J. M. Allen, "Size Dis-
tributions Of Fine Particulate Emissions From A Coal-Fired
Power Plant," Atmospheric Environment, 10, 1065-1069 (1976)
302
-------
References (Cont'd.)
8. Fuchs, N. A. The Mechanics of Aerosols, The MacMillan Co.,
60 - 5th Avenue, NY (1954), pp. 204-12.
9. Sinclair, D., "A Portable Diffusion Battery," Am. Ind.
Hygiene Assoc. J. 33(11), pp. 729-35 (1972).
10. Sinclair, D., "A Novel Form of Diffusion Battery," Am. Ind.
Hygiene Assoc. J., 36J1), pp. 39-42 (1975).
11. Breslin, A. J., S. F. Guggenheim, and A. C. George, "Compact
High Efficiency Diffusion Batteries," Staub (in English),
31(8), pp. 1-5 (1971).
12. Twomey, S. , "The Determination of Aerosol Size Distributions
From Diffusional Decay Measurements," J. Franklin Inst.,
275, pp. 121-38 (1963).
13. Sansone, E. B. and D. A. Weyel, "A Note On The Penetration
of A Circular Tube By An Aerosol With A Log-Normal Size
Distribution," Ae rosol Science , 2_, pp. 413-15 (1971).
14. Gormley, P. G. and M. Kennedy, Proc. Roy. Irish AcacL ,
52A (1949) .
15. MacLauren, E. , and C. Junge, "Relationship of Cloud Nuclei-
System to Aerosol Size Distribution and Composition,"
J_._ of Atmospheric Science, 4_, (1971) .
16. Sinclair, D., and G. S. Hoopes, "A Continuous Flow Con-
densation Nuclei Counter," Aerosol Science, 6_, pp. 1-7
1975.
17. Liu, B. Y. H., and D. Y. H. Pui, "A Sub-micron Aerosol
Standard And The Primary, Absolute Calibration Of The
Condensation Nuclei Counter," J. of Colloid and Interface
Science (1973).
303
-------
References (Cont'd.)
18. Haberl, J. B. and S. J. Fusco, "Condensation Nuclei Counters
Theory and Principles of Operation," General Electric
Technical Information Series, No. 70-POD 12 (1970).
19. Haberl, J. B., "A Mathematical Model for Cloud Chamber
Droplet Growth, " General Electric Technical Information
Series, No. 70-POD 11 (February 1970).
20. Fuchs, N. A., I. B. Stechkina, and V. I. Starasselskii,
"On The Determination Of Particle Size Distribution
On Polydisperse Aerosols By The Diffusion Method,"
Brit. J. Appl. Phys., 13, 280-81 (1962).
21. Liu, B. Y. H., K. T. Whitby, and D. Y. H. Pui, "A Portable
Electrical Aerosol Analyzer for Size Distribution Measure-
ment of Submicron Aerosols," presented at the 66th Annual
Meeting of the Air Pollution Control Association, Paper
No. 73-283 (June 1973).
22. Liu, B. Y. H., and D. Y. H. PUi, "On the Performance of
the Electrical Aerosol Analyzer," J. Aerosol Science,,
6, pp. 249-64 (1975).
23. Marple, V. A. and K. L. Rubow, "Aerodynamic Particle Size
Calibrations Of Optical Particle Counters," J. Aerosol
Science, 7, pp. 425-438 (1976).
304
-------
APPENDIX E
LABORATORY DETERMINATION OF PARTICULATE RESISTIVITY
Page
E . 1 STANDARD LABORATORY TECHNIQUE 306
E.I.I Apparatus
E.I.2 Procedure for Laboratory Resistivity Mea-
surements
E.I.3 Calculations
E. 2 AN ALTERNATE LABORATORY TECHNIQUE 312
305
-------
APPENDIX E
LABORATORY DETERMINATION OF PARTICULATE RESISTIVITY
E.I STANDARD LABORATORY TECHNIQUE
The standard technique for conducting laboratory resis-
tivity measurements is described in the American Society
of Mechanical Engineers Power Test Code 28, Determining
the Properties of Fine Particulate Matter. This code was
adopted by the Society in 1965 as a standard practice for
the determination of all the properties of fine particulate
matter which are involved in the design and evaluation of
dust-separating apparatus. The tests include such proper-
ties as terminal settling velocity distribution, particle
size, bulk electrical resistivity, water-soluble sulfate
content, bulk density, and specific surface.
The document defines bulk electrical resistivity as
the resistance to current flow, expressed in ohm-centime-
ters, through a dust sample contained in a cubic volume
one centimeter on a side when exposed to an electrical vol-
tage equivalent to 90% of the breakdown voltage of the sample,
applied uniformly across two opposite faces of the cube.
The code specifies that the property is to be determined
at 150°C (300°F) and at a humidity of 5% by volume, unless
otherwise specified.
E.I.I Apparatus
The basic resistivity cell is shown in Figure El.
It consists of a cup which contains the ash sample and serves
as the lower electrode, and an upper electrode with a guard
306
-------
MECHANICAL
GUIDE
(INSULATED)
1/32 IN,
AIR GAP
GUARD RING
1-1/8 IN. DIA. BY
1/8 IN. THICK ^
MOVABLE ELECTRODE
3/4 TO 1 IN. DIA.
BY 1/8 IN. THICK
7
DUST CUP
3 IN. ID, 5 mm DEEP
HIGH VOLTAGE SUPPLY
0700-14.22
Figure El. Resistivity cell for laboratory measurements.
307
-------
ring. To conform with the code, the high-voltage resistivity
cell must have the dimensions shown, and must use electrodes
constructed from 25-micron porosity sintered stainless
steel.
The controlled environmental conditions required for
the measurement of resistivity in the laboratory can be
achieved by an electric oven with thermostatic temperature
control and with good thermal insulation to maintain uniform
internal temperature, and a means to control humidity.
Humidity may be controlled by any one of several conventional
means, including circulation of pre-conditioned gas through
the oven, injection of a controlled amount of steam, use
of a temperature-controlled circulating water bath, or the
use of chemical solutions which control water vapor pressure.
Figure E2 illustrates a suitable setup for resistivity mea-
surements .
E.I.2 Procedure for Laboratory Resistivity Measurements
The first problem encountered in making any resistivity
measurement is obtaining an appropriate dust sample. The
prescribed procedure for PTC 28 Code assumes that samples
of gas-borne dust are taken from a duct in accordance with
the Test Code For Determining Dust Concentration in a Gas
Stream (PTC 27-1957).2 The PTC 27 Code involves isokinetic
dust sampling at various points in the duct. It is rec-
ommended that samples should not be obtained from a large
bulk of material in a hopper, silo, or similar location. If
it is necessary that samples be obtained from such a location,
procedures which will insure that the sample is representative
of the whole must be used. For any resistivity test to be perform-
ed on a bulk sample, it is necessary that a random sample be
obtained. This can be done by quartering the bulk sample to
308
-------
1. Pressure regulator
2, Constant temperature bath
3, Pump
4, Heater
5. Make-up water reservoir
6, Externally heated piping
7, PTC 28 apparatus
3, Environmental sampling port
9, Externally heated exit piping
10. Calibrated C/A thermocouple
11. Power source for oven
12. mV Potentiometer
13. Cold junction
14. Oven
15, Fritted disc
16. Environmental chamber
17. Fritted disc air bubbler
18, Bath water in overflow
19, Air flowmeter
20. Air tank 9
8
0700-1-4.23
Figure E2. Laboratory apparatus for measuring particulate bulk resistivity.
309
-------
obtain the test sample. To break up agglomerates and to
remove foreign matter, e.g., collection plate scale, the speci-
men can be passed through an 80-mesh screen.
The procedure for making the resistivity measurement accord-
ing to Power Test Code 28 follows:
1. The sample is placed in the cup of the resistivity
cell by means of a spatula. Then it is leveled by drawing a
straight edge blade held vertically, across the top of the cup.
2. The disc electrode is gently lowered onto the surface.
It should rest freely on the sample surface without binding
on any supports.
3. The resistivity cell is mounted in the environmental
chamber and equilibrium temperature and humidity are established,
The Code specifies that a temperature of 150°C (300°F) and
a humidity of 5% by volume are to be used for the test, unless
otherwise specified.
4. A low voltage is applied to the cell and then gradually
raised in a series of steps up to the point of electrical break-
down of the sample layer. Current transients will occur when
the voltage is first applied or increased across the cell.
It is necessary that these die away before recording current
and voltage readings (approximately one minute), A record of
the current-voltage characteristic of the dust is obtained.
Preferably using another sample, the above is repeated; when
another sample is not available, the sample layer should be
remixed and releveled after each run in order to break up any
spark channels that may have been formed in the dust layer.
A total of three runs should be made. The average breakdown
310
-------
voltage is then calculated. Before taking the samples to break-
down, it is necessary to determine whether the temperature and
moisture content of the sample are in equilibrium with tempera-
ture and humidity of the controlled environment. A test for
equilibrium is that the voltage-current measurements are re-
producible to within 10% when determined by two successive mea-
surements made 15 minutes apart.
5. The resistivity of the samples is then calculated in
the range of 85 to 95% of the average breakdown voltage, using
the corresponding currents from the previously recorded voltage-
current characteristics.
6. In any case, resistivity measurements at elevated tem-
peratures and controlled humidity should be made under equili-
brium temperature and moisture conditions between the gas and
the dust. The condition for equilibrium of temperature and
moisture conditions in the ash shall be determined by the require-
ment that resistivity measurements be reproducible within 10%
when determined by two successive measurements 15 minutes apart.
Further, since the particle conductivity usually depends on
the past treatment of the particles, as well as on the tempera-
ture and humidity equilibrium reached under given conditions,
it is preferable to determine the conductivity by raising the
oven temperature to the value desired rather than by first heating
above the desired temperature and then cooling to the desired
temperature. For additional information, see Reference 3.
E.I.3 Calculations
Resistivity can be calculated in the following way. First,
calculate the resistance of the dust layer, R.
-
I(amps)
311
-------
Then calculate the resistivity,
p (ohm-cm) = R(ohms) ™ (E2
The moisture content of the air in the environmental cham-
ber can be determined by weighing a tube filled with calcium
sulfate (Drierite) before and after passage of a measured volume
of air through it. The volume of dry air passed through the
tube is determined from the flow rate and the sampling time.
E.2 An Alternate Laboratory Technique
An example4 of an alternate laboratory technique is given
below. This procedure has been quite useful for the comparison
of resistivities for a large number of fly ashes over an extend-
ed temperature range. All procedural steps, equipment, calcula-
tions, etc. are identical to those stated in Section E.I. unless
otherwise specified.
1. The resistivity cell and ash specimen are thermally
equilibrated at 460°C overnight in an environment of dry air
or nitrogen.
2. After determining the resistivity under condition 1
(usually a voltage of 2 kV/cm is used), the test environment
is introduced. The environment can be an air-water mixture
or a simulated flue gas without sulfur oxides. The voltage
gradient and water concentration are selected with respect to
the specific test objectives.
312
-------
3. After the specimen "equilibrates" with the test environ-
ment, resistivity is determined, and the oven is turned off
and allowed to cool naturally, 50°C to 100°C/hour.
4. As the test specimen cools, resistivity is determined
at selected temperature intervals between 460°C and 85°C.
Figure E3 shows data acquired using this procedure.
313
-------
•TO"
1011
o
I
O
t io10
UJ
EC
108
2.8
84
183
2.4 2,0 1.6 1.2
143.5 227 352 560
290 440,5 665.5 1040
TEMPERATURE
1000/T(SK)
"C
°F
3630-077
Figure E3. Resistivity versus temperature for fly ash obtained using the
procedure outlined in Section E.2,
314
-------
References
1. The American Society of Mechanical Engineers. Determining
the Properties of Fine Particulate Matter - Power Test Code
28. United Engineering Center, 345 East 47th Street, New
York, NY 10017.
2. The American Society of Mechanical Engineers. Determine
Dust Concentration in a Gas Stream - Power Test Code 27.
United Engineering Center, 345 East 47th Street, New York,
NY 10017.
3. "Information Required for Selection of Electrostatic and
Combination Fly Ash Collectors; Methods of Analysis for
Chemical, Physical and Electrical Properties of Fly Ash - In-
formation Report No. 2", APCA Journal 15_(6) 256-260, 1965.
4. R. E. Bickelhaupt, Private Communication.
315
-------
APPENDIX F
IN_ SITU PARTICULATE RESISTIVITY MEASUREMENTS
Page
F.I PRACTICAL FACTORS IN RESISTIVITY MEASUREMENTS 319
F.I.I Selection of Sampling Sites
F.I.2 Number of Measurements Required
F . 2 A POINT-PLANE PROBE 321
F . 3 GENERAL MAINTENANCE 324
F.4 OPERATION OF THE PROBE 326
F.4.1 Pre-Field Trip Preparation
F.4.2 Operating Instructions
F.4.3 Operating Outline
F.4.4 Calculations
F.4.5 Example
316
-------
APPENDIX F
IN SITU PARTICULATE RESISTIVITY MEASUREMENTS
Point-to-plane probes for measuring particle resistivity
have been used since the early 1940's in this country.1 Two
models of these devices are shown in Figure Fl.
This section describes the use of a particular point-to-
plane resistivity probe. Similar procedures would be used
with other point-to-plane resistivity probes. This probe can
operate under the following conditions:
1. From 1 to 23 g/m3 dust loadings (0.5 to 10 grains/ft3)
2. Corrosive gas compositions.
3. Sampling ports of 10.2 cm to 15.4 cm diameter (4 inch
to 6 inch pipe size).
4. Temperatures up to 200°C (392°F).
5. Resistivities of 106 to 10ll4fl-cm.
6. Gas velocities of up to 30.5 m/sec (100 ft/sec).
This description is presented in the following sections:
practical factors in resistivity measurement; a detailed de-
scription of the probe; general maintenance instructions; and
operating instructions.
317
-------
PICOAMMETER
CONNECTION
HIGH VOLTAGE
CONNECTION
DIAL INDICATOR
PICOAMMETER
CONNECTION
MOVABLE
SHAFT
STATIONARY
POINT
GROUNDED
RING
(W
3630-099
(0700-14,24)
Figure F1. Two types of point-to-p/ane resistivity probes.
318
-------
Nichols2 has written a more general description of methods
for measuring particle resistivity, both laboratory and i_n
situ.
F.I PRACTICAL FACTORS IN RESISTIVITY MEASUREMENTS
F.I.I Selection of Sampling Sites
The first priority in the selection of a sampling site
is the location of a point in the operating system where the
conditions of the gas and the gas-borne dust particles are
representative of the environment for which the particle re-
sistivity may be of practical significance. That is, the gas
temperature, gas composition, and particle history should be
the same as that found in the precipitator. Usually the inlet
of the precipitator is selected as the point for making re-
sistivity measurements. However, sampling at several points
across the duct may be required to obtain a representative
measurement if there are large variations in temperature across
the duct.
When selecting a site for the measurements, practical
considerations must also be remembered. At the site location,'
sampling ports must exist or be installed. The normal practice
is to use 4-inch pipe for the ports. Electrical power must
be available at the site location for the operation of the
measuring equipment. In many locations, adapters will be re-
quired for mating of plant electrical outlets with the stan-
dard three-prong plugs found on most laboratory equipment.
F.I.2 Number of Measurements Required
The determination of the number of individual measurements
required to characterize the resistivity of the dust is related
319
-------
to the range of operating conditions anticipated and the varia-
bility in the process emissions. It is frequently desirable
when gathering design data for a new precipitator installation,
that the worst operating conditions be covered in the test
schedule.
The variable plant operating condition that is of greatest
concern is the flue gas temperature. Seasonal changes in the
ambient air temperature can cause the flue gas temperature
to fluctuate as much as 15°C (27°F) and the temperature profile
across the duct downstream from a rotating air heater may show
variations up to 25°C (45°F). Many cyclical processes have
emissions with large fluctuations in particle and gas concentration,
flowrate, and temperature. Changes in temperature may cause
very large variations in the dust resistivity and care must
be exercised to assure that the widest variation is anticipated
and covered in the test plan.
The day-to-day variations in characteristics of the process
or emissions may also cause significant variations in the particu-
late resistivity. This variability will show up as scatter
in the measured value of resistivity over the measurement period.
Thus it is imperative to make enough measurements at each opera-
ting condition to obtain a statistically significant value
for the resistivity.
A precipitator acts to smooth out short-term variations
in particulate resistivity. Dust layers ranging in thickness
from perhaps one centimeter on the inlet plates to some lower
value, perhaps only a millimeter, on the outlet plates build
up during several hours of collection time. The average build-
up rate on the precipitator plates is on the order of one mil-
limeter per hour, exponentially distributed through the preci-
pitator, so that the dust layer on the plates may represent
320
-------
an averaging of the instantaneous dust conditions for many
hours of operation. Therefore, it is reasonable to average
the measured values of resistivity for each temperature con-
dition to arrive at the resistivity representative of the dust
at a particular installation.
A determination of the number of measurement points re-
quired is based on the variability of the source and the ex-
perience of the technician making the measurements. Typically,
six to ten measurements, each at intervals of 10°C (18nF),
are adequate if plant conditions are reasonably constant.
The auxiliary data required when conducting tests on an
operating precipitator include:
Process material samples for proximate and ultimate analysis
Flue gas temperature and composition
Precipitator voltage-current relationships
Hopper dust samples for laboratory analysis.
F.2 A POINT PLANE PROBE
System for making in srtu resistivity measurements include
a probe for insertion in the flue, a high voltage supply, a
volt meter, an ammeter with overload protection, a temperature
indicator, and a pump. A schematic diagram of a complete sys-
tem is shown in Figure F2. A set of drawings showing the de-
tails of the design of a particular probe are shown in Figures
321
-------
HIGH VOLTAGE
SUPPLY
r —
PROBE
AMMETER
(MULTIMETER)
***
t '
-01 '
t— i
r
1
z
50P.
\AAAt
ZENER
PROTECTION
CIRCUIT
3030-076
Figure F2, Schematic diagram of probe system.
322
-------
F7-F13. The first drawing in this set gives an assembly view
of the probe (drawing no. 2620-D-10).*
The power supply for the probe is a modified Peschel Model
H20-Y (20 kVDC Neg, 4 mA) with two voltage scales (0-25 kV
and 0-2.5 kV). The ammeter which is mounted in the power supply
cabinet is a Keithley digital multimeter model 160 (sensitivity
to currents as low as 10" : * amps). The input to the multimeter
is protected from surge currents during sparkover by a zener
diode protective circuit. This circuit also contains a 10
resistor for testing the probe.
The probe is equipped for collecting the dust, making
electrical contact with the dust, measuring the thickness of
the collected dust layer, and removing the collected dust layer,
all without removing the probe from a sampling port.
The particulate sample is collected by a point plane corona
discharge cell mounted in the end of the probe. The corona
point (1203) is located 5.72 cm from the 5.2 cm diameter collec-
ting electrode (11A). The collecting electrode consists of
a guard electrode (1108) and a center disc electrode (1104)
(diameter 2.52 cm. , area 5.0 cm2). The guard electrode is
*Every item in the probe has a unique identification number con-
sisting of the drawing number and the consecutive item number:
2^20-0-17 ,03.,
drawing1 no. item
A reference to item 1703 means the third item of drawing 17.
Every sub-assembly in the probe has a unique identification
number consisting of the drawing number and an assembly letter:
2620-D-17
drawing no. assembly letter
All items in the material list bracketed with the assembly letter
will be assembled into that sub-assembly. A reference to assembly
17A means sub-assembly A as specified on drawing number 17.
323
-------
connected directly to ground. The center disc is isolated
from ground by a machinable glass insulator (1103) and is connected
to an external ammeter. A Chromel-Alumel thermocouple mounted
in the back of the guard electrode is used for measurement
of the flue gas temperature.
Electrical contact to the exposed surface of collected
dust layer can be made by lowering a disc electrode (1206)
onto the collected dust. The thickness of the layer is deter-
mined by comparing the readings of a dial indicator (1003)
connected to the movable electrode. The sliding electrode
push rod is equipped with a spring to assure that a constant
compression force will be applied to the dust layer for each
test.
The collected dust layer must be removed by removing
the probe from the flue and manually cleaning the electrodes.
F.3 GENERAL MAINTENANCE
General maintenance of the probe requires that it be period-
ically disassembled and cleaned. Instructions for maintenance
of the electrical equipment are given in the manuals supplied
by the manufacturers.
Before disassembling the probe, consult the drawings.
In order to clean the probe, first clean and remove the
shield (1601) from the point-plane corona discharge assembly.
Remove the high voltage and sliding electrode assembly (12B
and 13D) from the probe by removing the bolts (1004) on the
upper flange and the three screws (1002) holding the high vol-
tage teflon junction block to the middle bulkhead (1404) (plate
from which the corona point protrudes). Slide this assembly
out of the probe casing and clean.
324
-------
The high voltage teflon junction block (12-B) consists
oj~ two concentric cylinders (1204 and 1205). It can be dis-
assembled by removing the screws (1216) in the side of the
junction block. Separating the cylinders exposes the high
voltage connection and the sliding electrode contact. This
area should be cleaned of any accumulated dust. The graphite
contacts (1211) should be checked for good electrical contact
and for free motion of the movable electrode (1206) .
At the upper end of the high voltage and sliding electrode
assembly is the dial indicator assembly (13-D), spring assem-
bly, control mechanisms for lowering the sliding electrode,
Swagelok quick connect connector (1324), and the high voltage
connector (1312). The dial indicator mechanism has a tendency
to corrode and should be very lightly oiled. The vertical
location of the dial indicator can be adjusted by losing the
locking screw (1308) and sliding the indicator up or down.
When the probe is assembled, the dial indicator should be
adjusted to read 5.00 when the sliding electrode is in contact
with the collecting electrode. The spring assembly should
be inspected to insure that the spring operates freely. If
it doesn't, dust has probably accumulated in this assembly
and it must be disassembled and cleaned. The electrode lower-
ing control (1307) should be free to turn and move up and down
when the acme screw is not engaged. The high voltage connector
which is fabricated from teflon tubing, Swagelok connectors,
and a banana plug, should be cleaned and electrical continuity
to the sliding electrode checked.
The collecting electrode electrical connections (11A)
are accessible by removing the flange (1102) from the bottom
325
-------
of the probe casing (15A). The machinable glass insulator (1103
which isolates the center disc electrode from ground should
be cleaned. The resistance to ground from the center electrode
should be greater than 1013ohms.
After reassembly, the probe electrode alignment must be
inspected. The probe is designed to be self-aligning. In
the lowered disc position, the sliding electrode should be
parallel and in contact with the center disc electrode. The
electrode alignment can be altered slightly by adjusting the
three screws (1002) that attach the teflon junction block
(1204) to the inner bulkhead.
F.4 OPERATION OF THE PROBE
F.4.1 Pre-Field Trip Preparation
Before using the probe in the field, general maintenance
should be performed to insure .that the probe will operate pro-
perly. It is possible to bench-test the probe using the 103 fi
resistor built into the spark protector box to simulate a
collected dust layer.
Set up the probe system as described in the next section.
Lower the sliding electrode so that it makes contact with the
collecting electrode and switch the control on the spark protec-
tor to the 109 ft position. Set the power supply for an output
of 100 volts (V) and read the current (I) to the multimeter.
Calculate the resistance (R) of the resistor in the protective
box by Ohm's Law:
326
-------
This value should be 1.00 x 103 9. ± 2%. Electrical con-
nectors and instrument calibration should be checked if the
above value is not obtained.
A pre-field inspection check-list is given in Table Fl.
P.4.2 Operating Instr uct ions
At the test site, the equipment should be carefully un-
packed and inspected. The electrical instrumentation package
is not sealed to keep out moisture and must be located out
of the weather but within 3 m (10 ft) of the sampling port. .
Connect the probe to ground. This is necessary to insure
proper operation of the probe, and for operator safety. Before
inserting the probe in the sampling port, lower the sliding
electrode until it makes contact with the collecting electrode.
If the metal shield (1601) for the corona discharge cell has
been removed; replace it at this time. (Between runs it is
necessary to remove this shield to clean dust from the cell.)
Adapters should be fabricated for 6" and 4" pipe nipples.
For some sampling ports special adapting flanges must be made,
or a temporary arrangement, such as rags or other suitable
sealing material, will have to be used. However, for strongly
negative or positive pressure flues, airtight flange connectors
should be used.
Insert the probe into the flue with the holes in the cell
cover perpendicular to the gas flow and allow the cell to reach
flue temperature. Approximately thirty minutes will be required
to reach this temperature.
327
-------
YES NO
Comments:
TABLE FI
RESISTIVITY PROBE
PRE-FIELD TRIP INSPECTION CHECK LIST
1. Probe - Breakdown inspection and calibration - wit
ing - HV cable, etc.
2. Power supply - Inspect operation - general condi-
tion, wiring, calibration, etc.
3. Multimeter - Inspect operation - general condi-
tion, wiring, calibration, etc.
4. Tool box - Correct tools for in-field breakdown
repair-inventory spare parts.
5. Power cords - Continuity of extension cord and
box.
6. Tent - Covering for instruments.
7. Field cleaning kit - Rags, brushes, and dusters fc
on-site cleaning.
8. Sample containers and data sheets - Supply of bott
or plastic bags to collect ash samples-supply of d
sheets.
9. Shipping boxes - Serviceable and in condition to r
ceive rough and abusive handling. Insure that ins
ments are sufficiently padded.
10. Confirmation - Unusual conditions at test site
accounted for: flue gas temperature, gas velo-
city, flue pressure, sampling port sizes, hot/cold
weather conditions, etc.
328
-------
A layout of the operating controls on the instrument
package is shown in Figure F3. Plug the AC line from the instru-
mentation package into a 117-120 VAC line. Using their individual
power switches, turn on the multimeter and the high voltage
supply.
A suitable temperature indicator for a Chromel-Alumel
thermocouple should be connected to the thermocouple output
on the side of the probe (1506) .
After enough time has passed to allow the cell to equili-
brate at the flue gas temperature, a test may be started.
Twist the operating knob to insure the sliding electrode is
lowered down in place. With the probe cover holes oriented
perpendicular to the gas stream, unscrew the sliding electrode
control and raise it to the up position. Lock in place. Now
run a "clean-plate" V-I curve by placing the multimeter on the
100 nA scale and setting the VM switch to the high position.
Check that the slide switch on the spark protector is in the
normal position. Press the HV ON push button; the red HV light
should be on to show the high voltage supply is activated.
The use of the high voltage supply is described in the manu-
facturer's manual. Adjust the OUTPUT control through its full
range using the kV meter as a guide and make a current reading
every 1000 volts until a spark level or the maximum output vol-
tage is reached. Keep the multimeter within its range during
these measurements to prevent excessive overranging. Record
these readings on a data sheet and mark it "clean plate".
Adjust the HV output control for a current of 2 pA and rotate
the probe so the cover plate holes face into the gas stream.
A dust layer is now precipitated on the collecting elec-
trode. The proper operating current density required for the
type of dust being collected has to be experimentally deter-
mined. Thus, the first test may not be useful for obtaining
329
-------
— lEROSTaHT — —V.W.-
OFF O3>))ON LOWCSfiJH
160 DIGITAL MULTIMETER
CASE LO HI
£-2Hf'i
— OUTPUT —
-MAIN POWER-
8,
— ON H.V.—OFF —
3S30-063
Figure F3. Instrument package.
330
-------
data. The current density normally used should fall somewhere
in the range between 0.2 and 2,0 uA/cm2 . If a high resistivity
dust is encountered, reduced current densities may be required
for proper operation. Use of the V-I curves will be explained
later to indicate how the proper current for precipitation
may be found if the originally selected value proves to be
insufficient. A current of 2 uA, giving a current density
of 0.4 uA/cm2, is a good place to start the initial test.
The voltage necessary to obtain this current will be approximately
of 15,000 V. Depending on the resistivity of the dust being
collected, the grain loading/ and the current density selected,
about thirty minutes to one hour may be required to precipitate
a sample of thickness between 0.5 and 1.5 mm.
As the dust layer is being deposited, the current will
begin to decrease. This current drop may be used to estimate
the collection time. When current drops significantly, or
if an hour has passed, the test should be stopped. If too
small a sample was collected on a short time run, run longer
the next time, no matter how much the current happens to de-
crease. After a sufficient sampling time has elapsed, turn
the probe so that the holes in the cover plate are perpendicu-
lar to the gas stream. Now run a "dirty-plate" V-I curve using
the same procedure as that for the "clean-plate" V-I.
After completing the "dirty-plate" V-I, turn the high
voltage off by turning the HV output control to zero and pres-
sing the HV OFF pushbutton. Switch the slide switch on the
spark protector box to the 10sfl position. This inserts a 1C9 ft
resistor in the circuit and protects the multimeter from
an overload current when lowering the sliding electrode with
the voltage on. Set the multimeter on the 100 nA range. Turn
the voltage supply on and adjust for a 100 V output.
331
-------
Unlock and lower the sliding electrode carefully and slowly
until the acme screw is engaged. Turn the control to lower
the electrode until the multimeter indicates that electrical
contact with the dust layers has been made. Turn the control
knob one-half turn and stop. If the dust resistivity is less
than about 109 ft-cm, the multimeter should read approximately
100 nA. For high resistivity dust, smaller currents will be
obtained, the exact current depending on the thickness of the
dust layer and the resistivity. Now set the multimeter on the
1000 pA scale and switch the slide switch on the spark protector
back to the normal position. If the power supply does not indi-
cate an overload (4 mA}, "spark data" can be taken.* Increase
the voltage across the dust layer in 100 V steps. Read and
record the corresponding currents until a spark occurs across
the dust layer. This will be indicated by the voltmeter needle
jumping and an erratic reading on the multimeter.
Before starting another run the dust layer must be removed
from the electrodes by mechanically removing the dust.
Remove the probe from the flue. Remove the metal cover
from the discharge cell and clean the cell thoroughly. If the
sample is to be saved, place a sheet of paper under the- disc
to collect the sample when the operating rod is pulled back
to its up and locked position.
After cleaning, replace the metal cover on the probe.
Return the probe to the sampling port. While the probe is
*Overloads frequently occur with high carbon content samples.
The carbon particles or similar type conductors provide a con-
ducting path between the disc, allowing the full output current
of the power supply to flow. If a short is encountered, it is
impossible to obtain data for determining the resistivity of the
layer between the parallel discs.
332
-------
returning to the flue temperature, perform calculations for
the test just completed.
F.4.3 Operating Outline
1. Prepare sampling port
2. Clean and align cell
3. Lower disc and lock
4. Insert into flue, with inlet holes 90 degrees
to flow, and bolt to flange
5. Allow cell to reach flue gemperature
6. Zero dial indicator
7. Raise operating rod
8. Run "clean-plate" V-I, slide switch normal position
9. Turn inlet holes into flow
10. Apply necessary voltage to supply precipitating
current
11. After desired length of time turn probe so inlet
holes are again 90 degrees to flow. (Leave high
voltage applied so dust layer will not be shaken
off in the turning process.)
12. Run "dirty-plate" V-I
13. Lower disc, slide switch IP9 position, voltage 100
14. Record thickness of dust layer
15. Apply "spark-data" voltage 100 V steps until sparkover
occurs, slide switch normal position
16. Remove probe or remove collected dust by pressurizing
probe
17. Observe layer and save if needed
18. Clean probe and check alignment
19. Insert back into flue
20. Make calculations
333
-------
F.4.4 Calculations
A sample data sheet for a typical run is given in Figure
F4. All the information necessary for making the resistivity
calculations is given on this data sheet. The "clean" and
"dirty" plate V-I information should be graphed. The curves
for data on this sheet are shown plotted in Figure F5.
The formula for calculating the resistivity is:
p =
|^ x 5.00 cm2
I (cm)
where p = resistivity (ohm-cm) ,
R = resistance V/I (ohms) ,
AV = voltage across the dust layer (volts) ,
I = measured current (amps),
A = area of disc (5.00 cm ), and
i = thickness of dust layer (cm).
The term A/1 is called the cell factor. This factor will
remain constant for the V-I or spark calculation for each indi-
vidual run. For different dust layers it is apparent that
the cell factor will change.
F.4.5 Example
The following data were obtained from V-I curves in Figure
F5.
334
-------
SRI POINT PLANE PROBE DATA
Location - Power Plant Layer Thickness - 1.0 mm
Time • 0915 Data - 14 May 1973 Test No. - A-6 Temp. - 314*F
Conditions - Normal, full load 56 MW
Unit 1, Port 3
V-l DATA
KV | CLEAN
1
2
3
4
B
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1.0 NA
0.25 MA
0,65 MA
1,15 MA
1,8 MA
2.6 MA
3.2 MA
DIRTY
1,0 NA
0.1 fiA
0.3 ^A
0.5 nA
1.1 AiA
1.65^A
2.19 fiA
4.3 MA 2.8 MA
5.1 MA
6.2 MA
7.1 MA
8,2 MA
9,8 ^A
11.1 ,uA
12.6 /M
SPARK
3.7 #JA
4.2 M
4,8 MA
5.6 MA
6.25 MA
SPARK
SPARK DATA
V j I
50
100
150
200
250
300
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1100
1200
1300
1400
1500
1600
1700
1800
1900
2000
2100
2200
2300
2400
2500
2600
2700
2800
2900
3000
2.5 NA
5.0 NA
7.5 NA
10.0 NA
13.5 NA
17.4 NA
23.6 NA
29.0 NA
39.5 NA
55.5 NA
70.5 NA
96.7 NA
0.14 MA
0.17 MA
0.23 MA
0.36 MA
0.46 MA
0.61 MA
0.75 MA
1,0 MA
SPARK
E
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
5500
6000
6500
7000
7500
8000
8500
9000
9500
10,000
363O075
Figure F4. Sample data sheet.
335
-------
O CLEAN PLATE
A DIRTY PLATE
12 14
VOLTAGE, kV
3,330
Figure F5. V-l information.
336
-------
V = AV = 850V
I = l.OxlO"6 A
I = 1.0 mm = l.OxKT ' cm
The value V is the voltage drop across the dust layer
as interpolated from the V-I curves at a current value of 1.0x10
A. Certain considerations must be taken into account when
obtaining this voltage drop. The first is to look at the shape
of the V-I curve. There are three basic shapes that may be
encountered. The diagrams in Figures F.6. will illustrate
two of these shapes.
In Figure F (Top), the point x shows the voltage at which
electrical breakdown occurs in the dust layer. This indicates
the onset of back corona, a characteristic of a high resistivity
dust. It would be incorrect in this case to use any of the
current and voltage relationships above the point x for cal-
culating resistivity values.
This V-I curve may be used also to determine the operating
point for the next run. If the point x is located at a lower
current value than the one selected for collecting the sample
then there is a good chance that the sample was collected in
a back corona situation. If this is the case then the current
for the next run should be reduced to the value of current
that corresponds to the point x. More efficient collection
should be observed at this setting.
In Figure F6 (Bottom)/ the "dirty-plate" curve is on the
left side of the "clean-plate" curve. This is a characteristic
of either a very high or a low-resistivity dust. Since the
AV taken from the curve will have a negative value, it will
not be possible to use the V-I procedure for resistivity cal-
culations in this case. Figure F5 is the third shape and it
shows a standard curve.
337
-------
CLEAN
CLEAN
3630-074
Figure F6. Two types of clean and dirty plate V-l curves.
338
-------
The cell factor is the first calculation to be made.
For the sample case, the cell factor is 50 cm and it comes
from the term A/2, , where A is 5 cm2 and i is 0.1 cm. The next
step is to find the resistance R of the dust layer. For this
run, AV is equal to 850 V, this was taken from the V-I graph
at a current of l.OxlO"6 amps. From the relation, R = AV/I,
the resistance is found to be 0.85 x 103 ohm. By multiplying
the cell factor by the resistance, a resistivity of 4.2x10
ohm-cm is obtained. This complete calculation is:
P = A/B. x V/I
or
5 cm z 0.85 x 103 V
p 0.1 cm x 1.0 x 10'6 A
D = 4.2 x Iff0 ohm-cm
After obtaining a value for the dust resistivity from
the V-I data, a check of this value may be obtained from the
spark data information. The proper values to take from the
spark data information are the last voltage and current reading
before spark. In this case the layer broke down at 1100 volts
and the last reading before breakdown was at 1000 volts with
a current of 1.0xlO~bamps. Using the following formula the
resistivity data may be obtained:
P = A/fc x V/I
5.0 cm 1.0 x 103 V
A
P = 0.1cnil.0x 10-6
p = 5.0 x 101Q ohm-cm
339
-------
The column labeled "E" on the spark data sheet is for
the calculated electric field for the voltage applied and the
thickness of the layer. In this example, breakdown of the
layer occurred at an electric field of 10,000 volts/cm. When
a series of measurements are made, the resistivities should
be calculated not only at sparkover for each run, but also
at a fixed value of the electric field. This will eliminate
the electric field dependence when comparing runs.
340
-------
ol-o-O?17
-------
} ,
«?<-,
Figure F8. In-situ resistivity probe.
-------
__...^_____ 7-/a-
£f-4.o*J} ^{ UtA,* * >tj& tali's
, i 4 »*i«*.*s;
*h
^ *»i * oat ***** ^ "•'
i\- » ?ZL"'a-*'*>""' \
' ST/T.A.";".»««: — . J"i r
'"''"'' wve , n.
_, - B.I.,. 1HW|, Tffin, .Trrffa
t. V - -, i "T -p— ^
* -i ;0-- i^. L-
4 •/— -^ ^ rr-i-v1 *- f
s - <* . -
— k i r
uci -*fl r • --„ ^_^i,
*a/s**] it % I Tn
Af ' -<«^< ^ » h'* * ^'i r'"1"
?,?! £t" V .,,..£- |
Figure F9. In-situ resistivity probe.
Reproduced from
best available copy.
-------
c
F.
, "fr "Nf LOMG.
if t fit fOD f j iw*"-* fcj MTffLj . l*H(T
to*, ^VlWr /£ &4 A * C t- ff *• * /^C CJ ~ C
Bj EHrii'Tf fft-lCM. ,
«/t rrup " J**<(. A«T
K-Pe«-,*? HO., JJ',**-^<' "4 ts
t,m* .|t| u- lirm> J
• r«.rrAv£ WC-ILU
MT '51 •>'»UP ^*^l«*,""*
SOUTMIM ttlUKH HSTnUtl
lilMNQNAM. «tAI*M* lHOt ___
Figure FIG. In-situ resistivity probe.
-------
_
\eA /*^ aw*' f
-------
~f~T~
rt iH
L1
(IPS'!
HU
0-'
A
.,| Ii - 1
-S-
,.
oi
K5UKH
*". »!»•»**• S1J03
Figure F12 . In-situ res-i stivity probe.
-------
")/ -a -
! r;zu;
•H.1L.-TC* P*
£**> T-*"~ j^'
x
,\'1 ce,^ (.
rr
4ILJ.*I* '"
-
**).
jfv 4. *?•
\, -
__/ f^r^^
N V«l -'"
soumiw MSUKM Ktmnt
Figure F13. In-situ resistivity probe.
-------
References
White, H. J. Industrial Electrostatic Precipitation.
Addison-Wesley Publishing Company, Inc., Reading, Massa-
chusetts, 1963. 376 pp.
Nichols, Grady B. Techniques for Measuring Fly Ash Resis-
tivity. EPA 650/2-74-079 (August, 1974) . U.S. Environmental
Protection Agency, Research Triangle, Park, North Carolina.
348
-------
APPENDIX G
FEDERAL STATIONARY SOURCE PERFORMANCE REFERENCE METHODS
Page
Summary of Source Performance Reference Methods 352
Method 1 -• Sample and Velocity Traverses for
Stationary Sources (Type S Pitot Tube) 358
Method 2 - Determination of Stack Gas and Volu-
metric Flow rate 361
Method 3 - Gas Analysis for Carbon Dioxide, Excess
Air. and Dry Molecular Weight.. 354
Method 4 - Determination of Moisture in Stack
Gases. 356
Method 5 -- Determination of Particulate Emissions
from Stationary Sources 368
Method 6 - Determination of Sulfur Dioxide Emissions
from Stationary Sources 373
Method 7 - Determination of Nitrogen Oxide Emissions
from Stationary Sources. 375
Method 8 - Determination of Sulfuric Acid Mist and
Sulfur Dioxide Emissions from Stationary Sources 375
Method 9 - Visual Determination of the Opacity of
Emissions from Stationary Sources. 38]
Method 10- Determination of Carbon Monoxide Emissions
from Stationary Sources 386
Method 11- Determination of Hydrogen Sulfide Emissions
from Stationary Sources 389
Method ISA-Determination of Total Fluoride Emissions
from Stationary Sources-SPADNS Zirconium Lake
Method 392
349
-------
APPENDIX G (Cont'd.)
Method 13B-Determination of Total Fluoride Emissions
from Stationary Sources-Specific Ion Electrode
Method 401
Method 14- Determination of Fluoride Emissions from
Potroom Roof Monitors of Primary Aluminum Plants.... 408
350
-------
APPENDIX G
FEDERAL STATIONARY SOURCE PERFORMANCE REFERENCE METHODS
To evaluate the performance standards for new stationary
sources, the EPA has promulgated reference methods which speci-
fy the manner in which certain tests must be performed. These
Reference Methods can be found in the Code of Federal Regula-
tions under Title 40 -- Protection of Environment; Chapter 1 -
Environmental Protection Agency; Subchaptec C - Air Programs;
Part 60 - Standards of Performance for New Stationary Sources;
Appendix A - Reference Methods,
In the first section of this appendix these reference
methods are summarized. In the later part these reference
methods, as presented in the Code of Federal Regulations, are
reproduced verbatim. They have been corrected and amended
through May 15, 1977.
351
-------
METHOD 1
Sample and Velocity Traverses for Stationary Sources:
Procedure for selecting stack sampling site and selecting and
locating minimum number of traverse points.
Federal Register
Vol. No. Date . Page
Regulation Promulgated 36 247 12/23/71 24882
Regulation Amended 38 99 5/23/73 13562
Proposed Revised Regulation 41 111 6/08/76 23061
METHOD 2
Determination of StackGas and Volumetric Flowrate:
Procedure for determining stack gas velocity from gas density and
velocity head using Type S pitot tube. Volumetric flowrate is
calculated from gas velocity and stack cross-sectional area.
Federal Register
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24886
Proposed Revised Regulation 41 111 6/08/76 23063
METHOD 3
GasAnalysisfor Carbon Dioxide,Excess Air, and Dry Molecular Weight:
COi, CO, and 02 are determined by absorbing in appropriate reagents
(Orsat analysis).
Federal Reg_ister_
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24886
Proposed Revised Regulation 41 111 6/08/76 23069
352
-------
METHOD 4
Determination of Moisturein Stack Gases:
Stac]< gas moisture content is determined by condensation and
measuring condensed water volumetrically.
Federal Register
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24887
Proposed Revised Regulation 41 111 6/08/76 23072
METHOD 5
Determination of Particulate Emissions from Stationary Sources;
Stack gas is sampled isokinetically, particulate matter filtered
and weighed after removal of uncombined water.
Federal Register
Vol. No. Date Page
36 247 12/23/71 24888
38 99 5/23/73 13563
111 6/08/76 23076
Regulation Promulgated
Regulation Amended
Proposed Revised Regulation
41
METHOD 6
Determination ofSulfur Dioxide Emissions from Stationary Sources:
S02 is separated from any S03 and H2SOi, mist present using aqueous
isopropyl alcohol. The SO2 is passed through dilute hydrogen per-
oxide which is titrated by the barium-thoris method to determine
sulfate formation.
Federal Register
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24890
Proposed Revised Regulation 41 111 6/08/76 23083
353
-------
METHOD 7
Determination of Nitrogen Oxide Emissions from Stationary Sources:
A grab sample is drawn into an evacuated flask containing actified
hydrogen peroxide which converts NO and N02 to HNO3. Analysis is
made by the colorimetric PDS method.
Federal Register
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24891
Proposed Revised Regulation 41 111 6/08/76 23085
METHOD 8
Determination of Sulfuric Acid Mist and Sulfur Dioxide Emissions
f ron* Stationary Sources:
Regulation Promulgated
Proposed Revised Regulation
SOs and HzSOi* mist are absorbed in aqueous isopropyl alcohol and
SO2 passes through to be collected in hydrogen peroxide solution.
Each solution is titrated by the barium-thoris method to determine
sulfate formation.
Federal Register
Vol. No. Date Page
36 247 12/23/71 24893
41 111 6/08/76 23087
METHOD 9
Visual Determination of the Opacity of Emissions from'- Stationary Sourc<
Opacity of Emissions from a stationary source is determined visually
by qualified observers.
Federal Register
Vol. No. Date Page
Regulation Promulgated 36 247 12/23/71 24895
Proposed Revised Regulation 39 177 9/11/74 32857
Revised Regulation Promulgated 39 219 11/12/74 39874
354
-------
METHOD 10
Determination of Carbon Monoxide Emissions from Stationary_Sources:
Integrated or continuous qas samples from stacks are analysed for
CO content using NDIR analysis.
Federal Register
Vol. No. Date Page
Regulation Promulgated 39 47 3/08/74 9319
Regulation Amended 39 75 4/07/74 13776
METHOD 11
Determination of Hydrogen Sulfide Emissionsfrom Stationary Sources:
K2S is collected in alkaline Cd(OH)2 to form CdS, which is dissolved
in HCL and reacted with a known amount of iodine. Amount of iodine
consumed is measure of H2S present in gas.
Federal Register
Vol.
Regulation Promulgated
Regulation Amended
Proposed Revised Regulation
METHOD 12
Determination of Sulfur Dioxide Emiss
39
39
42
ions
No.
47
75
99
from
Date
3/08/74
4/17/74
5/23/77
Stationary
Page
9321
13776
26222
Sources
by Continuous Monitors:
No specific continuous monitor required; results must show acceptable
relationship to those determined by Method 6 or 8.
(RESERVED)
355
-------
METHOD 13A
Determination of Total Fluoride Emissions from Stationary • Sources -
SPADNS Zirconium Lake Method:
Gaseous and particulate fluorides are withdrawn isokinetically and
concentration determined by SPADNS method.
Federal Register
Vol. No. Date Page
Regulation Promulgated 40 152 8/06/75 33157
Regulation Amended 41 230 11/29/76 52299
METHOD 13B
Determination of Total Fluoride Emissions from Stationary Sources -
Specific Ion Electrode"Method:
Gaseous and particulate fluorides are withdrawn isokinetically and
concentration determined by a specific ion electrode.
Federal REgister
Vol. No. Date Page
Regulation Promulgated 40 152 8/06/75 33163
Regulation Amended 41 230 11/29/76 52299
METHOD 14
Determination of Fluoride Emissions from Pot Room Roof Monitors
of Primary Aluminum Plants;
A permanent sampling manifold is constructed to isokinetically with-
draw sample gas from the roof monitor. The gas sample is brought to
ground level and sampled using Method 13A or 13B for determination
of fluoride concentration.
Federal Register
Vol. No. Date Page
Regulation Promulgated 41 17 1/26/76 3829
356
-------
METHOD 15 (Proposed)
Determination_gf H25, COS, and CS2 Emissions from Stationary Sources^
A gas sample is extracted and diluted with clean dry air. An aliquot
of the diluted sample is then analysed by gas chronnatographic separa-
tion (GC) and flame photometric detection (FPD).
Federal Register
Vol. No. Date Page
Proposed Regulation 41 193 10/04/76 43870
METHOD 16 (Proposed)
Semicontinuous Determination of Sulfur Emisjj.ons from Stationary Sources;
A gas sample is extracted and diluted with clean dry air. An aliquot
of the diluted sample is then analysed for gas sulfur. Two GC/FPD
analysers are used for resolution of both high and low molecular
weight sulfur compounds.
Federal Register
Vol. No. Date Page
Proposed Regulation 41 187 9/24/76 42017
METHOD 17 (Proposed)
Determination of Particulate Emissionsfrom Stationary Sources
(In-Stack Filtration MethodlT
Particulate matter is withdrawn isokinetically anc1 collected on a
filter maintained at stack temperature. Particulate matter mass
is determined gravimetrically after removal of uncombined H2O.
Federal Register
Vol. No. Date Page
Proposed Regulation 41 187 9/24/76 42020
357
-------
METHOD 1
Sample and Velocity Traverses For Stationary Sources
1. Principle and Applicability.
1,1 Principle. A sampling &ite r.rd Ihe
number of traverse points are selected to aid
In the extraction of a. representative sample.
1.2 Applicability. TlilB method should
be applied only \vhen specified by the test
procedures for determining compliance wah
tlie New Source Pciformr.uce Standards. Un-
less otherwise speclnod, Ui'.,-. mcrhod is n-,t
Intended to apply to gas streams o'her tha:i
those emitted directly to the atmosphere
without further processing.
2. Procedure.
2.1 Selection of a sampling tlte ai«J mini-
mum number of traverse points.
2.1.1 Select a sampling site that 5s at leiint
eight stack or duct dlame:ers dou.'r.EtretEn
and two diameters upstream from auy flow
disturbance such as a bend, expansion, con-
traction, or visible flame. For rectangular
cross section, determine an equivalent dJum-
eter from the following equation:
. , ,, /(icneth) (width) \
equivalent diametcr=2[ ^—' ••—7-—- )
\ length-}-width /
equation 1-1
2.1.3 When the above sampling site
criteria can be met, the minimum number
of traverse points Is twelve (12),
2.1.3 Some sampling situations renda* the
above sampling elte criteria Impractical.
When this Is the case, chooae a convenient
sampling location and use Figure 1-1 to de-
termine the minimum number of traverse
points. Under no conditions should a sam-
pling point be selected within I Inch ot tae
•tack wall. To obtain the number of Wavers*
points for stacka or duets with a diameter
leas than 2 feet, multiply the number oi
points obtained from Figure l-l by 0.67.
3.1.4 To use Figure 1-1 first measure the
distance from the chosen sampling location
to the nearest upstream and downstream dln-
turbances. Determine the corresponding
number of traverse points for each distance
from Figure 1-1, Select the higher of the
two numbers of traverse points, or & greater
value, such that for circular stacks the num-
ber la a multiple of 4, and for rectangular
stacks the number follows the criteria of
section 222
2.2 Cross-sectional layout and location ol
traverse points.
22.1 For circular stacks locate the tra-
verse points on at least two diameters ac-
cording to Figure 1—2 and Table 1-1. Tee
traverse axes shall divide the stack cruse
MctJon Into equal parts.
For rectangular stacks divide the
cross section Into as many equal rectangular
areas as traverse points, such, that the ratio
of the length to the width of the elemental
areas Is between one and two. Locate the
traverse points at the centrold of each equal
area according to Figure 1-3.
3. References,
Determining Dust Concentration In a Qaa
Stream, A3ME Performance Teat Coda #27,
New York. N.T.. 1BB7.
DevorkLn, Howard, et a!., Air Pollution
Source Testing Manual, Air Pollution Control
District, L03 Angeles, Calif, November 1963.
Methods for Determination of Velocity,
Volume. Dust »"^ Mlat Content of Oases,
Western Precipitation Division of Joy Manu-
facturing Co., Loa Angeles, Calll. Hull at m
WP-60, 1988.
Standard Method for Sampling Stacks for
Partlculat* Matter, In; 1971 Book of ASTM
Standards. Part 23, Philadelphia, Pa. 1B71,
A3TM Designation D-292&-71.
358
-------
NUMBER OF DUCT DIAMETERS UPSTREAM'
(DISTANCE A!
2
5
'FROM POINT OF ANY TYPE OF
DISTURBANCE (BEND. EXPANSION, CONTRACTION, ETC.;
10
NUMBER OF DUCT DIAMETERS DOWNSTREAM1
(DISTANCE 61
Figure 1-1. Minimun number of traverse points.
Figure 1-2. Cross section of circular stack divided into 12 equal
areas, showing location of traverse points at centroid of each area.
o
Q
0
O
r ~"~
o
9
0
o
o
r
o
Figure 1-3. Cross section of rectangular stack divided into 12 equal
areas, with traverse points at centroid of each area.
359
-------
Table 1-1. Location of traverse points in circular stacks
(Percent of stack diair.eter froir Inside wall 10 traverse point)
Traverse
point
number
on a
diameter
1
2
3
Hunber of traverse points on a diameter
2 4 • 6 8
14.6
85.4
6.7
25.0
75.0
4 93.3
5
6
7
B
9
10
n
12
13
14
15
16
17
18
19
20
21
22
23
24
4.4
14.7
3.3
10.5
29.5 19.4
70.5
85.3
95.6
32.3
67.7
80.6
89.5
96.7
10 1 12
2.5
8.2
14.6
22.6
34,2
65.8
77.4
85.4
91,8
97.5
2.1
6.7
11.8
17,7
14
1 .8
16
1.6
5.7 4.9
9,9
14.6
25.0 20,1
35.5
64.5
75.0
25.9
36,6
63.4
82.3 73.1
88.2 73.9
53.3 85.4
97.9 90.1
94.3
98.2
3.5
12,5
16:9
22,0
28,3
37.5
62.5
71.7
78.0
83.1
87.5
91.5
95.1
98,4
18
1.4
4.4
7,5
10.9
14.6
18.8
23,6
29.6
38.2
61.8
70.4
76,4
81.2
85.4
89.1
92.5
95.6
98.6
20 22
1.3
1.1
3.9l 3.5
6.7 6.0
9.7
8.7
12.9 j 11 .6
16,5
20.4
14.6
18.0
25,0 • 21 .8
30.6 26.1
38.8
61.2
69.4
75.0
79.6
31.5
39.3
60.7
68.5
73.9
83.5 J7B.2
87.1 I 62.0
90.3
93.3
96.1
98,7
85. 4
88,4
91.3
94.0
96.5
98.9
24
1.1
.3.2
5.5
7.9
10.5
13.2
16.1
19.4
23.0
27.2
32.3
39.8
60,2
67.7
72.8
77.0
80.6
63,9
86.8
89.5
92. J
94.5
96.8
93.9
360
-------
METHOD 2
Determination Of Stack Gas Velocity And Volumetric
Flow Rate (Type S Pitot Tube)
1. Principle and applicability.
1.1 Principle, Stack gaa velocity is deter-
mined from the gas density and from meas-
urement of tho velocity head using a Type S
(Stauechelbe or revere* type) pilot tube.
1.2 Applicability. Tbla method should be
applied only when specified by the test pro-
cedures lor determining compliance with the
Nsw Source Performance Standards.
2, Apparatus.
2.1 Fltot tube—Type S (Figure 1-1), or
equivalent, with a coeQclent within ±5%
over the wording rouge.
2.2 Differential pressure gauge—Inclined
manometer, or equivalent, to measure veloc-
ity head to within 10% of the minimum
value.
2.3 Temperature gauge—Thermocouple or
equivalent attached to tba pluit tube to
measure stack temperature to within 1.5% of
the minimum absolute stack temperature.
2.4 Pressure gauge—Mercury-Blled U-tube
manometer, or equivalent, to measure stock
pressure to within 0,1 in, Hg,
2.5 Barometer—To measure atmospheric
pressure to within 0.1 In. Hg,
2,9 Gae analyzer—TO analyze gas composi-
tion for determining molecular weight.
2.7 Pltot tube—Standard type, to cali-
brate Type S pltot tube.
3. Procedure,
3,1 S«t up the apparatus S3 shown in Fig-
ure 2-1, Make sure all connections are tight
and leak tree. Measure the velocity bead and
temperature at the traverse points speclfled
by Method 1.
3.2 Measure the ctatlc pressure Ln th»
stack,
3.3 Determine the stack gaa molecular
weight by gaa analysis and appropriate cal-
culations as Indicated In Method 3,
4. Calibration.
4.1 To calibrate the pltot tube, measure
the velocity head at some point In a flowing
pas stream with both a Type S pltot tube and
a standard type pltot tube with known co-
efficient. Calibration should be done In the
laboratory »"rt the Telocity ol the flowing gaa
stream should be varied over the normal
working range. It IB recommended that the
calibration be repeated after use at each field
site.
4.2 calculate the pltot tube coefficient
using equation 2-1.
P — P /^PlKJ
"""" BllllVAp~.7 equation 2-1
where:
Cn1i« = pltot tuba coefficient of Type S
pltot tube.
d..,a=Pltoit tube coefficient of standard
type pltot tube (If unknown, use
0.99).
ap.id= Velocity head measured by stand-
ard type pltot tube.
4pi.n=s Velocity head measured by Type S
pltot tube.
4.3 Compare the coefficients of the Type S
pltot tube determined first with one leg and
then the other pointed downstream. Use the
pltot tube only If the two coefficients differ by
no more than 0,01.
5. Calculations.
Use equation 2-2 to calculate the stack gas
velocity.
wbare:
Equation 2-2
i. -Stock gas velocity, fwt per wccnd ((-P-9.).
K,=S&-48-E
trbeDthmeiiDlu
C, = Pltot tube coefficient, d
rr,).T«.' Avua^e absolutB Black gu
°
e velocity head oi slack gag, fnchoa
!=-'« K:S. 3-a.
.\i,=Muli.i;ubr
.
l of sta;
Mj = Ury mol?cnlai" weight of stack ga3 ifroui
Methods).
D.o— Proportion by volume o( waur vspnr In
the gtis stream Urom Mbttiod 4),
Pigurs 2-2 3ho*-s a sample recording sheet
lor velocity traverse data. Use the averages
In the lost cwo columns ol Figure 2-2 to de-
termine the average alack gas velocity from
Equation 2-2.
Use Equation 2-3 to calculate the stack
K»a volumetric now rate.
Equation 2 i
where;
Q,— Volumetric flaw rate, try basis, standard oiniJI-
tions, fl.'/hj.
A = Cross-»!ci1onaI area of fta>.'k, Tt.'
THd^Ateilutc toTTiperatuTfi at standard cundltimia
5*r- K.
P,,i = Abso!!jle pressure a! standard conditions, M.M
Inches Hg.
6. References.
Mart, L. 3,, Mechanical Engineers' Hand-
book, McGraw-Hill Book Co.. Inc., New York,
N.Y., 1951.
Perry. J. H., Chemical Engineers" Hand-
book, McGraw-Hill Book Co.. Inc.. New York.
N.Y,, 19BO.
Shlgehara, R. T., W. P. TodrJ, and W. S,
Smith, Significance of Errors In Stack Sam-
pling Measurements. Paper presented at the
Annual Meeting of the Air Pollution Control
Association, 3t. Louis, Mo.. June 14-10, 1970.
Standard Method for Sampling StacSa for
Paniculate Matter, In: 1ST1 Book of ASTM
Standards, Fart 23, Philadelphia. Pa, 1971,
ASTM Designation D-292B-71.
Vennard, J. K., Elementary Fluid Mechan-
ics. John Wiley & Sons. Inc., New York, H.Y,
1847,
361
-------
PIPE COUPLING^ TUBING ADAPTER
'Pigure 2-1. Pilot tube-manometer assembly.
362
-------
PLANT_
DATE
RUN NO.
STACK DIAMETER, in.
BAROMETRIC PRESSURE, in.
STATIC PRESSURE IN STACK !Pg), in. Hg.
OPERATORS
SCHEMATIC OF STACK
CROSS SECTION
Traverse point
number
Velocity head,
in. H20
Slack Temperature
AVERAGE;
Figure 2-2. Velocity traverse data.
363
-------
METHOD 3
Gas Analysis For Carbon Dioxide, Excess Air, And Dry
Molecular Weight
1. Principle and applicability.
1.1 Principle, An Integrated or grab gaa
sample la extracted from a sampling point
dnd analyzed for Its components using an
Great analyzer,
1-2 Applicability. This method should be
applied only when specified by the test pro-
cedures for determining compliance with the
New Source Performance Standards, The teat
procedure will indicate whether a grab sam-
ple or an integrated sample ta to be used.
2. Apparatus.
2,1 Grab sample (Figure 3-1).
2.1.1 Probe—Stainless steel or Pyrex1
glass, equipped wltb a filter to remove panic-
ulate matter,
2.1.2 Pump—One-way squeeze bulb, or
equivalent, to transport gas sample to
analyzer,
3.2 Integrated sample (Figure 3-Z).
2,2,1 Probe—Stainless steel or Pyrex1
glass, equipped with a filter to remove par-
ticulate matter.
2.2,2 Air-cooled condenser or equivalent—
To remove any excess moisture.
2,2,3 Needle valve—To adjust flow rate.
2.2.4 Pump—-Leak-free, diaphragm type,
or equivalent, to pull gas.
2.2 5 Bate meter—To measure a flow
range from 0 to 0,035 dm.
2.2,6 Flexible bag—Ted'.ar,' or equivalent
with a capacity of 2 to 3 cu. ft. Leak test the
bag in the laboratory before using.
2.2.7 Pilot tube—Type S, or equivalent
attached to the probe so that, the sampling
Row rats can be regulated proportional to
the Rtacfe gas velocity when velocity is vary-
ing with time or a sample traverse is
conducted.
2.3 Analysis,
2.3.1 Great analyzer, or equivalent.
3. Procedure,
3.1 Grab sampling.
3.1-1 Set up the equipment as shown In
Figure 3-1. malting sure all connections are
leafc-lree. Place the probe In the stack at a
sampling point and purge the sampling line,
3,1.3 Draw sample Into the analyzer.
3.2 Integrated sampling.
S3.1 Evacuate the Bexlble bag. Set up the
equipment as shown ha Figure 3-2 with the
bag disconnected. Place the probe In the
stack and purge the sampling line. Connect
the bag, making euro that all connect Ions are
light and that there are no leaks.
3.2.2 sample at a rate proportional to the
stack velocity.
3.3 Analysis.
3.3.1 Determine the CO,, O,, and CO con-
centrattonB as soon as possible. Make as many
passes as are necessary to give constant read-
ings. If mere than ten passes are necessary,
replace the absorbing solution.
3.3.3 For grab sampling, repeat the sam-
pling and analysis until three consecutive
samples vary no more than 0.6 percent by
volume for each component being analyzed.
3.3.3 For Integrated sampling, repeat the
analysis of the sample until three consecu-
tive analyses vary no more than 0,2 percent
by volume for each component being
4. Calculations.
. 4.1 Carbon dloiide. Average the three con-
wrutlve runs and report the result to the
nearest 0.1 % CO,.
43 Excess air. Use Equation 3-1 to calcu-
late excess air, and average the runs. Report
the result to th« neirest 0.1 % excess air.
__ ___
6:204 (',-. X,)"- (70 0,) +0..i{% CO) *
equation -'1-1
where:
%EA = Percent excess air.
^0,= Percent oxygen by volume, dry basl».
%N, = Percent nitrogen by volume, dry
basis,
ffCO — Percent carbon monoxide by TO!-
ume, dry basis,
0.384 = Ratio of oxygen to nitrogen 1n air
by volume.
1 1 Dry molecular weight. UM Equation
t-2 to calculate dry molecular weight and
nveriige the runs. Report the result to the
nf-nre-sc tenth.
-n.44(%CO,)
equation 8-2
Mp'Dry molecular weight, lb./Tb-mcl«.
%CO-=Percent carbon dloxldt by »o5um«,
dry basis,
%Or-Psrcent oxygen by »olum», dry
basis.
%Ki— Percent nitrogen by rolum*, dry
basis.
0.44 = Molecular weight of rarbon dioxidft
divided by 10O.
0 32.= Molecular weight o! oxygen divided
by 100.
0.28=MolecuIar weight o- nitrogen and
CO divided by 100,
6. References.
Altshaller, A. P., et all., Storage of Gases
and Vapors in Plastic Bags, Int. J. Air &
Water Pollution, 6:76-81, 1863.
Conner, William D., and J. S. Nader, All
Sampling with Plastic Bags, Journal of the
Americsn Industrial Hygiene AssoclAUrra.
25:291-297, May^June 1954.
Devorkln, Howard, et al,. Air Pollution
Source Testing Manual, Air Pollution Con-
trol District, Los Angeles, Calif., November
less.
1 Trade
364
-------
PROBE
FLEXIBLE TUBING
TEftIG
FILTER (GLASS WOOL)
SQUEEZE BULB
Figure 3-1. Grab-sampling (rain.
BATE METIR
VALVE
AIR-COOLED CONDENSER / PUMP
PROBE
FILTER (GLASS WOOL)
TO ANALYZER
QUICK DISCONNECT
RIGID CONTAINER
Figure 3-2. Integrated gas - sampling train.
365
-------
METHOD 4
Determination Of Moisture In Stack Gases
1. Principle and applicability,
1.1 Principle. Moisture Is removed from
the gaa stream, condensed, and determined
voUimetrlcally.
1.2 Applicability. Thla method is appli-
cable for the determination of moisture ID
staelt gas only whpn specified by test pro-
cedures for determining compliance wits New
Source Performance Stundrxrds. Thla method
does not apply when liquid droplets are pres-
ent In the c;a3 stream' and the moisture 1»
subsequently used In the determination ol
stack gaa molecular weight.
Other methods such as drying tubea, wet
bulb-dry bulb techniques, and volumetric
condensation techniques may he used.
2. Apparatus,
2.1 Probe—Stainless steel or Pyrex • gla&g
sufficiently heated to prevent condensation
and equipped with a filter to remove particu-
lar matter,
2.2 Implngers—Two midget Implngere.
each with 30 ml. capacity, or equivalent.
2.3 Ice batli container—To condensn
moisture In Implngers.
2.4 Silica gel tube (optional)—To protect
pump and dry gas meter.
3.6 Need:* valve—To regulat* gaa flow
rate.
2.6 Pump—Leak-free, diaphragm type, of
equivalent, to pull gaa through train.
2.7 Dry gas mater—To measure to within
!•£ of the total sample volume.
2.8 Rotameter—To measure a flow rang*
from 0 to 0,1 c.f.m.
2.9 Graduated cylinder—25 ml.
2.10 Barometer—Sufficient to read to
within 0.1 InchHg.
3.11 Pltot tube—Type 8, or equivalent,
attached to probe no that the sampling flow
rate can be regulated proportional to the
staek gaa velocity when velocity la varying
with time or a sample traverse IB conducied-
8. Procedure.
3.1 Place exactly 6 ml. distilled water In
each Implnger. Assemble the apparatus with-
out the probe as shown In Figure 4—1. Leak
check by plugging the inlet to the flrst 1m-
puiger and drawing a vacuum. Insure that
now through the dry gas meter Is less than
1 % of the sampling rate.
S-3 Connect tha probe and sample at &
constant rate of 0.076 o.f.m. or at a rate pro-
portional to the stack gaa velocity. Continue
sampling until the dry gaa meter registers 1
cubic loot or until visible liquid droplet are
carried over from the first implnger to the
second. Record temperature, pressure, and
dry gas meter readings aa required by Flgur*
4-2.
3.3 After collecting the sample, measure
the volume Increase to the nearest 0.6 tcJ.
4. Calculations.
4.1 Volume of water vapor collected.
=<
equation 4-3
where:
B¥o=Proportlon by volume of water vapor
In the gas stream, dlmenslonleBS.
V»« =Volume of water vapor collected
(standard conditions) , cu, ft.
Vm« =Dry gas volume through meter
i standard conditions) , cu. ft.
Bwn= Approximate volumetric proportion
of water vapor Ln the gas stream
leaving the Implngers, 0.025.
1 If liquid droplets are present In the gal
stream, assume the stream to be saturated,
determine the average stack gas temperature
by traversing according to Method 1, and
use a psychrometric chart to obtain an ap-
proximation of the molstur* percentage.
'Trade name.
equation 4—1
366
-------
5, References.
Air Pollution Engineering Manual, DsnJel-
•on, J. A, (ed.), U.S. DHEW, PHS, National
Center lor Air Pollution Control, Cincinnati,
Obio, PHS Publication No. 099-AP-40, 1887.
Devorkln, Howard, et al., All Pollution
Source Testing Manual, Air Pollution Con-
trol District, Los Angelas, Calif,, November
1963.
Methods for Determination of Velocity,
Volume, Dust and Mist Content of Oases,
Western Precipitation Division of Joy Manu-
facturing Co., Los Angeles, Calif., Bulletin
WP-60, 1968,
SILICA GEL TUBE
HEATED PROBE
FILTER' [GLASS WOOL)
ROTAMETER
ICC BATH
IDGET IMPINGERS PUMP
\
DRY GAS METER
Figure 4-1, Moisture-sampling train.
LOCATION.
TEST
COMMENTS
DATE
OPERATOR
BAROMETRIC PRESSURE
CLOCK TIME
GAS VOLUME THROUGH
METER. (Vm),
n3
ROTAMETER SETTING
fl^/min
METER TEMPERATURE.
"f
Figure 4-2. Field moisture determination.
367
-------
METHOD 5
Determination Of Particulate Emissions From Stationary
Sources
1. Principle und applicability.
1,1 Principle. Parilciilttte mailer ig with-
drawn Isokinetlcally from the source and Its
weight is determined gravimetrically after re-
moval or uncomblned water.
1,2 Applicability. Tiiis method is applica-
ble for the determination of particular emis-
sion from stationary sources only when
specified by the teat procedures for determin-
ing compliance with New Source Perform-
ance Standards.
2. Apparatus.
2.1 Sampling iraln. The design specifica-
tions of the paniculate sampling train used
by EPA (Figure 5-1) are described In APTD-
0561, Commercial models of this train are
avail able.
3.1.1 NozzJe—Stainless steel (316) with
•harp, tapered leading edge.
2.1.2 Probe—Pyrex' glass with a heating
syswm capable of maintaining a minimum
gas temperature of 250° F, at tha exit end
during sampling to prevent condensation
from occurring. Whea length limitations
(greater than about 8 ft.) are encountered at
temperatures leas than 600' P., Incoloy 825 »,
or equivalent, may be used. Probes far sam-
pling gas streams at temperatures in excess
of 600° F. must have been approved by the
Administrator.
2.1.3 Fitot tube—Type S, or equivalent,
attached to probe to monitor stack gas
Telocity,
2,1.4 Filter Holder—Pyrex < glass with
heating system capable of maintaining mini-
mum temperature of 225 • P.
2,1,5 Impingers / Condenser—Four impin-
gers connected In series with glass ball Joint
fittings. The first, third, and fourth impin-
gers are of the Greenburg-Smlth design,
modified by replacing the tip with a ',4 -inch
ID glass tube extending to one-ball Incl)
from the bottom of the flask. The second, im-
plnger 12 ot the Greenburg-Smlth design
with the standard tip. A condenser may be
used in place of the impingars provided that
the moisture content of toe stack gaa can
still be determined.
2,1.6 Metering system—Vacuum gauge,
leak-free pump, thermometers capable of
measuring temperature to within 5" F., dry
gas meter with 2% accuracy, and related
equipment, or equivalent, as required to
maintain an isoklnetie sampling rate and to
determine sample volume.
2.1.7 Barometer—To measure atmospheric
pressure to ±0.1 inches Hg.
2.2 Sample recovery.
2.2.1 Probe brush—At least as long aa
probe.
2,2.2 Qlass wash bottles—Two.
2.2.3 Qlass sample storage contalnera.
2.2.4 Graduated cylinder—250 ml.
2.3 Analysis.
3.3.1 Glass weighing dishes.
2.3.2 Desiccator.
2.3.3 Analytical balance—To measure to
±0.1 mg.
2.3.4 Trip bnl>ince—300 g. capacity, to
measure to ±0.05 g.
1 Trade name,
•Dry using Driertte' at 70' F,±10' F.
3. Reagents.
3.1 Sampling.
3.1.1 Filters—Glass fiber. M3A 1108 EH*
or equivalent., numbered for Identification
and prewelghed.
3.1.2 Silica £»!— Indicating type, 6-lfl
mesh, dried at 1"5' C. (350* F.) for 2 hours
3.1.3 Water.
3.1.4 Crushed ice,
3.2 Sample recovery
3 2 1 Acetone—Reagent grade.
3.3 Analysis,
3.3.1 Water
3.33 Desiicant—Drlerlte,1 indicating.
4, Procedure.
4,1 Sampling
4.1.1 After selecting the sampling site and
tha minimum number of sampling points,
determine the stack pressure, temperature,
moisture, and range of velocity head.
4.1,2 Preparation of collection train.
Weigh to the nearest gram approximately 200
g. of Elllca gel. Label a filter oT proper diam-
eter, desiccate* for at least 24 hours and
weigh to the nearest 0.5 rag. In a room where
the relative humidity Is less than 60%. Place
100 ml. Of water In eacb of the first two
Impingers, leave the third Implnger empty.
and place approximately 200 g. or preweighea
silica gel in the fourth impinger. Set, up the
train without the probe aa In Figure 6-1.
Leak check the sampling train at the sam-
pling site by plugging up the Inlet to the fil-
ter holder and pulling a 15 In. Hg vacuum. A
leakage rate not In excess of 0.02 o_f_m. at »
vacuum of 15 In. Hg Is acceptable. Attach
the probe and adjust the heater to provide a
gas temperature oT about 250° F. at the probe
outlet. Turn on the filter heating system.
Place crushed ice around the Impingers, Add
more ice during the run to keep the temper-
ature of the gases leaving the last Implngei
as low as possible and preferably at 70° F.,
or less. Temperatures above 70" F. may result
in damage to the dry gas meter from either
moisture condensation or excessive beat,
4.1.3 Particulate triUn operation. For each
run, record the data required 0:1 the example
sheet showi! in Figure 5-2. Take readings at
eacl) sampling point, or. least every 5 minutes.
and when significant changes In stack con-
dltlur.s necessitate additional adjustments
in flow rate. To begin sampling, position the
nozzle at tho flrbt traverse point with the
dp pointing directly into the gas stream.
Immediately start the pump and adjust the
flow to isoklnetic conditions. Sample for at
least 5 minutes at eacb traverse point; sam-
pling time must he che same lor each point
Maintain Isoklnetic sampling throughout, tho
sampling period. Nomographe are availabl*
which aid In the rapid adjustment of the
sampling rate without other compulations.
APTD--0576 details the procedure for using
those nomographs. Turn off the pump ot the
conclusion of each run and record the nnai
readings Remove the prohc nnd nozzle frv«:s
the stack and handle In ftccordnnc? with the
sample recovery process aescribed l:i section
4.2.
368
-------
IMPINGE* TRAIN OPTIONAL. MAY BE REPLACED
BY AN EQUIVALENT CONDENSER
HEATED AREA FJLTEH HOLDER / THERMOMETER CHECK
xVALVE
PROBE
REVERSE-TYPE
PITOT TUBE
IMPINGERS ICE BATH
BY-PASS VALVE
THERMOMETERS
VACUUM
GAUGE
MAIN VALVE
DRY TEST METER AIR-TIGHT
PUMP
Figure 5-1. paniculate-sampling train.
PLANT
LOCA'K?
QPEK»1DI
DAIE
Hi
B
PUfchO
S.JWPU B
0< NC
VfcTEfi BCt MO
C FaCTQ
TRAVinE PQIM
1GTJU
R
-r-
i raise
PWS5UKE
AVt^AGF
lEWfuTuEE
4^,£R|T ,^
XPE^BIU^
HBfWMHirPBISSIriSF
Ml Alt « tot
JljruK !,
!"rTM
fH »•
Wiri ,r
PK»r HFAIER SET'IiVG
SCMC^ar C 0* STAC* £*QE3 UCIION
NEfiD
ACROSS
QfflflCE
• a Hi
CAS S«Kfn.i
«"Ti
,«^^-r»a«
[ INLfl
" *!'"•'
SAWLF 5CI
TLiPEJUTtSE
"isf iipactp
"F
40
>.i.
««J
F SursS-2, Pan.ciil
369
-------
4.;-! Surni'In ri'«-»v.v\ Kvi'i;lR« rntu in mov-
Hi(5 i'le collection trulii from the test alie to
the sfiuiple recorory area to inliilmUo the
loss of coIIerLcrt >-'U>i|ite nr ilu: Ruin of
extraneous pfirtle-iirvte mutter, fof ajliie a
portion of tht* rir.L-)s>;jo usotl }u the sruupSt
recovery as a hiank for niinlyM-.. Meiiaure tlie
Toluaie of wnlcr frorn tlie Ilr.st thr<'0 1m-
plngers, then (Hck-ard. Plane the samples In
containers as follows:
Container No, I. Remove the filter from
Its holder, plftce In this container, and seal.
Container No. 2. Place loose pnrtloolat*
matter ana ncvUme wusliinfis In-m ftll
sample-exposed surfucf's prior to tlie filter
In this container and st.il. Use a razor blade,
brush, or rubber polle*;nmri ui iu«.» adher!:j({
6,3 Volunifr of wiitcr vaprir.
ir tln= BUK-&
ihe urltfu.al c
pu'.
gel i
C'anittltifT No, 3, Trni.-f
rftiiix the fourth implr.gcr i
'.:u::cr n'-a fcal. U^c: u ruij
mi airt In removing «IliL-.'i
Impluger.
4,3 Ar.ftlysi3_ Recurd the cl:ita reauireri on
tbe example sheet shown In Flgnre 6-3.
Handlf each sansplc contnl:ic-r as follows:
Container No. 1. Transfer the filter and
any loose particulate maucr from the sample
container to a tared glass weighing dish.
desiccate, and dry to a constant weight, Re-
port results to the nearest 0.5 mg.
Container No. Z. Transfer the acetone
washings to a tared beaker and eraporats to
drvness at ambleu» temperature and pres-
sure, Desiccate and drv to a constant, weight.
Rep-rc results to fhc nearest 0,5 mg,
Container No. 3. Weigh the spent silica gel
and reiiort to the nearest gram.
Use mettods and equipment which have
been fii/prov!implc (si&iitirird condltlanss,
cu, ii.
Vil = Tol«l '.i.'himc 'if liquid collsctsil !c
ImpinRers and silica gel (scb Fig-
lu-.j 5-,'n, irl.
Prtju— O-isllv of TanLer. 1 c; /ml.
Mii/i -Alulf-'i :,.r v...-iglu of w,ftt«r. 18 Ib.'
curi-iutui.M. 5'JQ* H.
P,.,= Al).>c.!u".e pressure at r*s:i
ditl'ms, 29.32 lnclic.< Hj
8 4 Moisture con'«at,
V,
d"''"""^T\?:;~
equaliou 5-3
l*n "^1* "portion by To'in.i" f.I
^ •in— v M.!]::*; i^f wati?r lit the ^as aampifi (FiutidEird
M'H lillonS), CU. »t.
^ ™iui&Vrlufr.c1 of K^3 5C-inple Lb rough the dry gas meter
(stiii-.. lard fuiiJitltmsj, cu. It.
8.5 Total partlculate weight. Determine
the total particulate catch from the sum of
the weights on the analysis data ahe«t
(Figure 5-3) ,
fi fl Concentration.
6. B.I Coucentra'-lon Ingr./s.cJ.
equation 5-4
e:
c'l^Ccnc^ntrHt.'on af putlsulnfa mailer la slack
pa.^, t'f-'s.ii-f , '5ry IntSlB.
« = Toia! air, .::nt c.1 ;Mr(lr.ulate matter coiloctwi,
nig.
1^1"= Volume of ^s5 pasrjplB through dry gas :nelar
(stand:i:\t CL-ridnions), cu. It,
equ.'ittnn .T-3
'!'*(Tifr i'i.»u • f pvrti^nifiE*1 mattar I:i ?f»
a III.
370
-------
PLANT,
DATE
RUN NO-
CONTAINER
NUMBER
1
2
TOTAL
WEIGHT OF PARTICIPATE COLLECTED,
mg
FINAL WEIGHT
;XL
TARE WEIGHT
!X
WEIGHT GAIN
FINAL
INITIAL
LIQUID COLLECTED
TOTAL VOLUME COLLECTED
VOLUME OF LIQUID
WATER COLLECTED
1MPINGER
VOLUME,
ml
SILICA GEL
WEIGHT,
9
g*| ml
CONVERT WEIGHT OF WATER TO VOLUME BY DIVIDING TOTAL WEIGHT
INCREASE BY DENSITY OF WATER. (1 g ml):
'"CREASE 9 = VOLUME WATER, ml
(1 g/ml)
Figure5-3. Analytical data.
371
-------
InnlunPrlr vnrlA'lnn.
,,..> .«
I ,,, i i " •'"~'>t
• h--p
l-|V.r.-!| , f.5,,kin
V, ,-T |;tl r,, !;,-;» .1
s n
' I!,
Mi.^-Y./ ,*; tr -VI-SL-'I! f •* 5i>", !*< )h lu.-j:?:!'!*.
Tn-'-A!'- !h- ,i\"?a-"- ilrv u ^ r-r;,'" fi*-],-.t't I'll---
i-.-' I iuM . ; .'• -I;
r-...-n,;.••.,•<•.* iv -.= !-,- .,: Mi-urn. 5in- t-M-li.~
lit.
T.-stlriir at Fedorni J-V.: -lIllcF. FHS, KCAPC,
Ti(H- C. in07.
Martin, R':hei-t M, Construction Dctiillr. of
Tsnkinetic Sjurce Sampling Equipment, En-
vlTonmcTiJsl rr'H-f-tiort Airency. APTD-5081.
Rnni. Jenrr.e J . Maintenance Callbratioji
B".C! oppT-at.'iii of is-iUinetlc Source Sntn-
rjlnig Eqillpm^ii1", En','s'ronme;ital Protectlcii
Acrcrcy, APTD-057C.
Smith. \v. S. R. T ShlE-cliEiri, nnd W. F.
Tndd. A Meihixl of Ir.ccrprvtlng Stack satn-
p"!nc Dn*^, Paper prewnted st trie 53d ATI-
nnal Moe-'!!•_• of the Air Pollution Control
Acceptable results. The following A.^ToUtmn rtt. Lci-ils. Mo.. June 14-13. 1570
Srr.lLJi. \v f . pi al . Stick Has SampMni'
Improved and S'ir.plir.ed with W«-,v E<:M!p-
m":!", nrr \ pi;>cr No C7-11Q. 1^67
Kpccil.'ftt or:-: (Mr l"i'!rerntnr T?Fting at
Fi-ri^ra: Fnnlr'es, THS. .TCATA. 1907
7Ri-jfe "ets the llnr.lt on ncceptab!e 1?oKlnettc
sampling results:
If on ~ < no ?,. the results are acceptable;
otherwise, re;ect the results and repeat
rha test.
372
-------
METHOD 6
Determination Of Sulfur Dioxide Emissions From
Stationary Sources
1. Principle and appli^nhHii:/.
1,] Principle, A gas sample is extracted
from the sampling point In the staci. The
acid mist, including sulfur trloidde, Is sep&-
rBl*d from the sulfur dioxide, me sulfur
dioxide Traction is measured hr the barium-
thorln tltmtion method.
1.3 Applicability. Thla method is appli-
cable for the determination ol sullur dioxide
emissions Irom stutiormry sources only when
specified by the teat procedures for determin-
ing compliance wlib New Source Performance
Stauflhrus.
2. A.upnraftM.
2.1 Sampling. See FlflUr? 6-!.
3.1.1 Probe—fjrei1 glnss, approximately
5 to 0 mm ID, with a b^fUltiK system to
prevent, condensation find B filtering medium
lo rciiiove paniculate matter Including auj-
tnrlc acid mist.
2.1 a MIrlpet bubbl-r -One, with glasn
wool packed ir. np ;n prevent sulfurlc add
m'.at carryover.
2 1.3 GliRa i-uol
2.1.4 Midgtl Impin^ers—T.'iree.
2.1.5 Drying lube---Hackfid with 8 to 16
mesL :ndlcat;ng-ty|je Flllca gel, or equivalent,
to dry the sample.
2.1.6 Valve—Needle valve, or equivalent,
to adjuat flow rate.
3.1.7 Pump— Lonk-free, vnriuim type.
3.1.B Rate mpy train by drawing
r>nn ambient Rlr t!-rough the system for 15
tn'nut-es,
4 1 Sample recovery. Disconnect the !ra-
pingere after purglns;. Discard the contents
if ihfl midget b-ibbler. Pour thfl Contents of
the mldee* Implngera Into B polyethylene
shipment, h.ittle. Blnse fche three midget 1m-
pinKcrs and the cuti.iectlng tubes with dis-
tilled water and add these washlnga to the
tame storage container.
4.3 Sample analysis. Transfer the contents
of tiie storage container Co a 50 ml. volu-
metric Cask. Dilute to the mark with de-
loulzed. distilled water. Pipette a 10 ml,
aliquot of this solution into a 120 ml, ErJen-
mtyer Husk. Add 40 ml, of laopropanol and
tn-i» to four drops of thorin l!irtier.tor. TUi'are
to n pinji eudpo:nt using 0.01 JV barium
pe:c:ilorate. Run a blank with each series
o' samples.
5. Calibration.
5.1 Use standard methods and equipment
which have been approved by the Adminis-
trator to calibrate the rotameter, pitot tube,
dry gas mewr, and probe heater.
373
-------
& 2 Standardize tho htirliun perchlorate
o.i:aiis.st 25 ml. of standard sulfuric acid cuii-
lainliiK 100 ml, of lsopro(inm>I,
d. Calrulatiuit*.
61 Dry gas voiume. Cnr:o..'L the p;i;»ple
vulim.c mensural by Liio dry ^as meter lo
standard conditions ("0" p. and 29.92 inches
Hg) by using equation 6-1.
°li /V ?, \
]7 71 _ / _ln-J;?.! ]
in. Hg \ Tu, / equulidii 0-1
where:
^'"•ta— Volume of go;; sample through the
dry gas uieU'r (standard coudl-
tlons), cu. ft.
¥„•= Volume ol gas sample through tha
dry gas meter (me car condi-
tions), CU- ft.
Tlta — Absolute temperature- at standard
conditions. 53(1° R.
Tm= Average dry niii meler temperature,
CR.
pt.r™ Barometric pu-isuro at ttis orlflcfl
C.2
,a= Absol-jrc pressure at standard con-
dltlous, 29.92 Inches Hg.
Sulfur dioxide concentration.
wt.crc:
Catt| — ( !"noeni rnM^ti of buHur dioxide
ni si.niirjurd concJttloii.i, dry
I/ttsK Ib./cii. ft.
7. !>5X 10-°" (.::n. verolun lactor, InctildlnB tlio
number o* griirnH per gram
equivalent or sulfur dioxide
(32 g./g.-eq.) , 453,6 e_/lb , and
1,000 ml./l., Ib.-l./g.-ml.
7t=Vn!uxne of barium perchlorate
tltrant used lor tiie (sample,
ml.
V,," V-jliin.e ftf horluin perchlurat/o
tllrant used for tho blank, ml.
A''-NV)rn),tI:l;, ,(f barlHtu perchlorntu
, ,..
V,,!L — Torul -')Iu:lon volume of sulfur
d:;i,'.;;lrr, 50 ml.
VE = Volume of sample aliquot ti-
trated, ml.
VD.,,^ Volume of gas sample through
tha dry gas meter (standard
conditions), cu. ft.. Bee Equa-
tion 6-1.
/ II, 1 \ O'.-V^N
c802=(-.o.-.xio-«-"—; ) ,-.- :~
V g.-nil./ Vu .
eqiuit.on fi-2
7. References.
Atmospheric Emissions from Su.lfu.rlc Acid
Manufacturing Processes, U.S. DHEW, PUS,
Division of Air Pollution, Public Health Serv-
ice Publication No, B99-AP-13, Cincinnati,
Ohio, 1865.
Corbett, P. P., The Determination of SO,
and SO, in Flue Oases, Journal of the Insti-
tute of Fuel, 24:231-243, 1861.
Matty, H. S. ana E. K. Dk-h!, Measuring
Flue-Gas SO, and SO.,, Power I(U:94-a7, No-
vember, 1357".
Fatten. W. F. and J. A. Brink. Ji- . New
Eqclpmoni, ar.d Techniques f^r Earr.pilng
Cliemical Process Gases, J. Air p.^ll Jihir. Con-
trol Association, 13, 162 (1863).
PflOEE (END PACKED
WITH QUARTZ OR
P»REX *OOtj
STACK WALL
SILICA GEL DRYING TUBt
MIDGET BUBBLER M!Dfi£T IMPINGERS /
GLASS WOOL ; / \ \ /
7
TYPE S PITOT TUBE
THERfflOMETER
DHV GAS METER ROTAMETER
PUMP
Figure 6-1. SOj sampling train.
374
-------
METHOD 7
Determination Of Nitrogen Oxide Emissions From
Stationary Sources
3.2 Sample rcccsi'ery.
3.2.1 Eadiurn hydroxide (1W)—Dissolve
40 g. NaOH In distilled water ana dilute to 1
liter.
32.2 Red Iltm-j5 paper.
3.2.3 Water—Delonlzed. distilled.
3.3 Ar.alys'.f;
3.3.1 Fujninv, r.ulfurlc acid—35 to IR~, by
weight free sulfur trloxlde.
3.3.2 Phenol—White eolld reagent grade.
3.3.3 Sulfurtc acid—Concentrated reagent
grade.
3.3.4 Standard solution—Dissolve 0.5495 g-
potasslum citrate (KNOa) in distilled watar
and dilute to 1 liter. For the working stand-
ard solution, dilute 10 ml. oi the resulting
solution to 100 ml. with distilled water. One
ml. of the working standard solution is
equivalent to 25 *g. nitrogen dioxide,
3.3.5 Water—UelonLzed, distilled.
3.3.8 Phenoldlsulfonlc acid solution—
Dissolve 26 g. of pure white phenol In 150 ml.
concentrated sulfurlc acid on a steam bath,
Cool, add 75 ml. fuming eulfurlc acid, and
heat at 100° C. for 2 hours, Storo In a dark,
stoppered bottle.
4. Procedure.
4.1 Sampllng,
4.1,1 Pipette 2.5 ml. of absorbing solution
into a sample flask. Insert the flask valve
stopper Into the 3ask -with the valve In tha
"purge" position. Assemble the sampling
train as shown in Figure 7-1 and place tn«
probe at the sampling point. Turn th« flask
valve and the pump valve to their "evacuate"
positions, Evacuate the flask to at least fl
inche" Hg absolute pressure. Turn the pump
valve to Its "\ent" position and turn off the
pump. Check the manometer for any fluctu-
ation Ir. the mercury level. If there is a visi-
ble change over the span of one minute,
check for loni.s. Record the mttlal volume,
temperature, and barometric pressure. Turn
the flask valve to HP "purge," pcsmon, and
then do the same with the pump valve.
Purge the probe ana the vacuum tube uslus
the squeeze bulb. If condensation occurs In
tha probe and flask valve area, heat the probe
and ivurge unrli Lhe condensation disp.ppefue.
Thrn turn the pump valve to Its "vent" posi-
tion. Turn the flask valve to Its "sample"
position and p.llow sample to enter the flask
for about 15 seconds. After collecting the
sample, turn the flask valve to its "purge"
position and disconnect the flask from the
sampling train. Shake the flask lor 6
mimifs.
4.2 Sample recovery.
4-2.1 Let the flask set for a minimum of
IS hours anil then shake the contents for 2
minutes. Connect the flask to a mercury
filled U-tube manometer, open the valve
from the nask to the manometer, and record
thB flask pressure and temperature along
with the barometric pressure. Transfer tbe
flask contents to a container Tor shipment
or to a 250 ml. beaker for analysis. Rinse the
flast with two portions of distilled water
prnvedure.
1.2 Applicability. This method Is applica-
ble lor the measurement of nitrogen oxides
from stationary tnurces only when specified
by the left procedures for determining com-
pliance with New Source Performance
Str.ndards.
2, Apparatus,
2.1 "sampling. See Figure 7-1,
2.1.1 Probe—Pyrex1 glass, heated, with
filter to remove participate matter. Heating
is unnecessary if tha probe remains dry dur-
ing the purging period.
2.1.2 Collection flask—Two-liter, Pyrex,1
round bottom with short neck and 24/40
standard taper opening, protected against
implosion or breakage.
2.1.3 Flask valve—T-bare stopcock con-
nected to a 24/40 suiadard taper Joint.
2.1.4 Temperature gaxige—DiEU-type ther-
mometer, or equivalent, .-apnble of measur-
ing 2° F. Intervale from 25* to 125' F.
2.1.5 Vacuum line—Tubing capable oi
withstanding a vacuum of 3 Inches Hg abso-
lute pressure, with "T" connection andT-bore
stopcock, or equivalent.
2,1.6 Pressure gauge—U-tube manometer,
38 Inches, with 0,1-lneh divisions, or
equivalent.
2.1,7 Pump—Capable of producing a vac-
uum of 3 Inches Hg absolute pressure.
2.1.8 Squeeze bulb—One way.
2.2 Sample recovery.
2.2.1 Pipette or dropper.
2,2.2 Glass storage containers—Cushioned
lor shipping.
2 2.3. Glass wash bottle.
2.3 Analysis.
2,3.1 Steam bath,
2.3.2 Beakers or casseroles—250 ml., one
for each sample and standard (blank),
2,3.3 Volumetric pipettes—1, 2, and 10 ml.
2.3.4 Transfer pipette—10 ml. with 0,1 ml.
divisions.
2.3.5 Volumetric flask—100 ml,, one for
each sample, and 1,000 ml. lor the standard
(blank).
2.3.6 Spectrophotometer—To measure afa-
Eorbance at 420 nrn.
2.3-7 Graduated cylinder—100 ml, with
1.0 ml. divisions.
2.3.8 Analytical balance—To measure to
o.i ing.
3. Reagents,
3.1 Sampling.
3.1.1 Absorbing solution—Add 2.8 ml, oi
concentrated H,SO, to 1 liter of distilled
water. Mix well and add 8 ml. of 3 percent
hydrogen peroxide. Prepare a fresh solution
weekly and do not expose to extreme neat or
direct sunlight.
1 Trade name.
375
-------
f//y /'J"1P VALVE
Ptjf.'P
snou^o-cisss CONE.
STANDARD TAPER. GfiOWO-GLASS
J SLEEVE NO, Z4 tB SOCKET. § NO. 52 5 '
PVRFI -t-?V ^
'Vx i
'J EKCASEVtNI
2-LITEB TOUND-BOTIW.1, SK2RTKIK.
*!1H j SLEEVE NO,2a,'«C
376
-------
to the flfimplr. For a nlnnk use i?5 ml of ab-
Borolrig solution and the some volume of dis-
tilled water as used In rinsing tl;n flnsk. Prior
to shipplui> or analysis, add sodinm hydrox-
ide (1ft) dropwlse Into both the sample and
the blank until alkaline to litmus paper
(about 25 to 35 drops In each).
4.3 Analysis.
*.3.1 If the sample hao been shipped In
K container, transfer the contents to a 250
ml. beaker using a small amount of distilled
•water. Evaporate the solution to dryuess on a
steam bath and then cool. AdJ 2 ml. phenol-
dlsulfonlc acid solution to the dried residue
and triturate thoroughly with a glass red.
Make sure itie solution contacts aSl the resi-
due. Add 1 ml. distilled -Auter and four drops
o( couceiilruicd sult'isrSc acid, Heat the solu-
tion on a stoam bath for 3 minutes with oc-
casional sllrrliis. Cool, add 20 ml. distilled
water, mix '.veil by stirring, and add concen-
trated ammonium hydroxide dropwise with
constant stirring u:itu alkaline to lltmua
paper. Transfer the solution to a 100 ml.
volumetric QasK and wash the beaker tttfea
times with 4 to 5 ml, portions of distilled
water. Dilute to the mark and jn:x thor-
oughly. H the Kiun]>''> o'ljitattiK s ii absorfilnij solution
cmd add sodium hydroside (Iff) dropwln*
until Bliailne to litmus paper (ftboul, 25 to
35 drops). Follow the Analysis procedure of
section 4.3 to collect enough data to drev, a
calibration curve of cuiicantratlou In fig. NOi
por sample veiius Hbsorbauce.
C. Calculations.
6.1 Sample volume,
,
where:
V§1 —Sample volume at standard condi-
tions (dry basis), ml.
T,,^—Absolute temperature at standard
conditions, 630' R.
Pfa™ Pressure at standard conditions,
29.92 Inches Hg.
Vr—Volume of flask acd valve, ml.
V.— Volum« of absorbing solution, 25 ml.
1 )b.
Pf=« Final absolute pressure of flask,
niches Hg.
P, —Initial absolute pressure of Oaefc,
Inches Hg.
1, — Final absolute temperature of flask..
°B.
T, — Initial absolute temperature ol flasm,
°R.
6.2 Sample concentration. Read
-------
METHOD 8
Determination Of Sulfuric Acid Mist And Sulfur Dioxide
Emissions From Stationary Sources
1. Principle and applicability.
1.1 Principle. A jrns sample Is. extnincd
!rom ft sampling point In tho stiir-v ntid the
acid mist Including sulfur trlnxide is Srpa-
rated from sulfur dlcxlde. Bru'i frfici.icr-;, on'
measured separately by thr bari'% ".-thorir.
tltr&tlon method.
1.2 Applicability. Tills method Is ^pp'.lca-
ble to determination or suiu.ric ur.id mlsi
(including sulfur trloxlrie) and 6ui.' :r diox-
ide from 5tnt!onery sources n«iy -*I-,en spe-
cified by the test procedures lor deicmunmg
compliance with the New Source Ft-rfurru-
hnee Standard^.
2. X.,'*par/]i ...*
2-1 Sampling. See Figure 6--1. Miiuy of
ihe design specifications at tftis sar.iplJng
tram are described In APTD-Q581.
2 1.1 Nozzle— St. iln'.t^-. "leul l318> wlUi
ehurp. tuptn'oci leading edyc*.
2.1.2 Prone — Fyrex ' g'l^s with & heating
system to prevent vis'U!e condensation dur-
iifi sampling.
2.1.3 Pitot tube — Type- S, or equivalent,
attached to probe to monitor stack g&a
velocity.
2.1.4 Filter-holder — Pyres ' glass.
2.1.5 Iraplngers — Four as shown !n Figure
8-1. The first ncd third tire ol :he Orecntmrg-
Smlth design with standard tip. The second
and lourch are cf the Grecnburg-Smlth de-
sign. modified by replacing the standard tip
with 0. Vi-lnch ID glnzs tube extending to
one-hall Inch from the bottom of. the 1m-
pmger flank. Slinliar collection systems.
'.-, hich huve been approved by the Adnilnls-
rrator. may '5C used.
2.1.6 Metering system — Vacuum gauge,
lenii-frea pump, thermometers capable of
measuring "empcrauire to within 5" F., dry
gaa meter with 2'-'c accuracy, and related
'jqulpment, or eqmvo-!f:;,t, as required to
maintain on isokluetlc sampling rate aud
lo determine sample voiume.
2J.7 Hnrori:«ier — Tn measure atmoap'.f tic
pressuro to ±0.1 lacli Hg.
2,3 Sample recovery.
2.2.1 Wash bartles— -Two.
Grudualed cylinders — 250 ml., EOO
Glass saiiipie storage containers.
Graduated cylinder— 250 ml.
2.2.3
nil.
a.2.3
a 2,4
£.3
2.3,1 Pipette — 25 ml., 100 on.
2.3.2 Burette — 50 nU.
2.3,3 Erlenmeysr flask — 250 ml.
2.3.4 Graduates cylinder — 100 mi.
2,3.5 Trip balance — 300 g, capacity, to
measure to ± 0,05 g,
2.3,6 Dropping bottle
solution.
3, Reagents,
3.1 Sampling.
3,1.1 Filters — Qlr.ss fiber. MSA type HOB
BH, or equivalent, of a sui,.,Q!e hiie tu flt
In the filler holder.
3.1.2 Silica gel — Indicating type, 6-18
me^h, dried at 175° C. f3BO- P.) lor 2 hours.
31.3 Water — DelonlEed, distilled.
to add Indicator
3.i 4 tsoprt'pajiol, SOTo — Mix 600 mi. Of
isoHupnm., with 200 ml. or delonh'.ed, fila-
tllU'fl wutci-.
3.1.0 Hydinycn pr'fc.Klde, a%— Dilute 100
iul. ol T07<- hvfirotcen poruside tu 1 liter ttlti:
deioiiliCil. diiiill-x:; »'. ater.
3,1. a Crushedlce.
3.2 Sample recovery.
3,2.1 Water— Di-lonlzcd, >::.-itlilcO
3.2.2
3.31 'Water— Deloaizcd, oiKtllled
3,3.2 Isopropanol.
3,3.3 Thcrin indicator — l-'o-r.ra
yla7,o)-2-naphthc-;-a. S-dlsuIftnic acid, dl-
*,(1him salt (or equivalent). Dissolve 0.20 g.
In inn ml. distilled water.
a.ti.4 Bnrlum perchlorata (O.Olff) — D!s-
aolve 1.95 g. of barium p«rchlorate [Ba
(fro^, 3 tl o; :n 200 m!. rtlrtiilsrt \.-ater and
dilute to 1 li'er with Lsopropatiol. 3tandard'.aB
with siilturlc acid.
3,3.5 Sulfuric acid standard (0.01W) —
Filrcnass or S'.andnrd'.ae to ± 0.0002 N e.salaat
0.01 -V Kr.OU -^hlch has prertously been
flr.ttadtcdlzKfl aefilast primary otsaidard pc-
tftiiekirn ncld phthalate.
4. lfv.tidv,Tf .
4,1 Sor^pnug.
4.1.1 Altf^r B43leotlug the sampling site and
tha minimum number ol sampling points,
determine the stack pressure, temperature,
moisture, and range ol velocity head.
4,1.2 Preparation of collection train,
Placa 100 ml, of 80% Isopropanol In the first
lmpliiE«r. 100 ml. of 3% hydrogen peroxide In
both the second and third Implngers, aad
about 200 g, or silica gel In f.he fourth. 1m-
plnser. liataln a portion of the reagenia Tor
use aa b Lin.fr solutions. Assemble the tr^in
without the proba as shown in Figure- 3-:
with the filter betweea the first and second
Implngcra. Leak check the sampling trair.
at tho 6acij7!!ng site by plugglug the Inlet to
tha nisi, Imptnger and pulling a 15-Inch H£
vacuum. A leakflge rate iiut ift exctes of f' u2
cjjn. at a, vacuum of 16 inches HU ii> ^-", -
capra'ole. Attach ths probe riicl turn or. tho
proba heating Eystem. Adjust the probe
hiater setting during tuii;pltng to pr. . ^at.
any visible condensation. Piauo crushed ice
around tae Implogera. Add more Ice during
the run to keep the teinijer«turc of the gases
laavlng tbe lost implnger at 70' r. or less,
4,1.3 Ti';tln operation, For each run, re-
cord. tha data required on the example sheet
shown In Figure 3-3. Take readings at each
sampling point at least every 5 minutes sind
when significant changes in stack coud;c;ius
necessitate additional adjustments In flc-n
rate. To begin eamplirtg, position the no7,z.'e
at the first traverse point with the tip pi-L0570 details
tJie procedure for us'.yig these nomographs,
' Tr,-4e name.
378
-------
At the conclusion of each run, turn off the
pump and record the finni readings. Remove
the probe from the alack and disconnect It
from the train. Drain the Ice bath and purge
the remaining part of the train by drawing
clean ambient air through the system for 16
minutes.
4.2 Sample recovery.
4.2 1 Transfer the Isopropanol from the
first Implnger to a 250 ml. graduated cylinder.
Rinse the probe, flrst Lmpinger. and all con-
necting glassware before the filter with 90%
IsopropJinol. Add the rinse sulut'.on to the
cylinder. Dilute to 350 ml. with 3D'rt Isopro-
punol. Add the niter to the solution, mix,
and transfer to a suitable storage container.
Transier the solution from the s;coud and
third Implngera to a 500 ml. graduated cyl-
inder. Rinse all glassware between the flHer
n:id silica pel Implnger with (iclonlzed, dis-
tilled water nrti add this rir.se water tn trie
cylinder Dilute to a volume o? 500 ml with
delonized, distilled water. Transler ihc EOJU-
tlon "o a suitable storage container,
4.3 Analysis.
1:1,1 Shake the container holding Iso
prepanol and the filter. If tr.e filter breaks
up, allow the fragments to settle for a few
niir.utea before removing a sample. Pipette
ft JOO ml. aliquot of sample Into e, 250 ml.
Erleiiraeyer flast and add 2 to 4 drops oT
thorin Indicator. Titrate the sample wltb
barium pcrcMoratc to ft pint end point. Make
.-•ure ",u ric-jiti ', u'.umes. Repeat the ttira-
tion with a second aliquot of sample. Shate
the container hol-Jing the concents of the
second and third impingrrs. Pipette a 25 rnL
tllquot of sample Into a 250 nil. Erlenrnever
fl-if.it. Add I'M ml. of Isoprcpanol and 2 to 4
Crops of tl-.or:i: indicator. Titrate the sample
with barium perch lorn* <• to c, pink er.d point.
Reptut tho il'ratinn v/r.)-, ^ second ahquoi of
sample. Titrate the blanks In the samo
manner as tho sanip'.KS.
5. Calibration.
5.1 Use s1-.Tiicia.rd methods and equipment
which have been approved by the Aclmirjls-
tcator to calibrate the orifice meter, plrot
tubs, dry gas mefer, and probe heater.
52 StandftidlKc the twrlwm perch!or,\:e
with 23 ml. cf standard sulturic acid con-
taining I'JiMnl i;f 1stpropanol,
6 Ca'.c^litiOTis.
6 1 Dry s;ns volume. Correct the sample
volume measured by the dry gas meter to
standard coruJltScus 170° F., 29.92 inches Hg)
by using Equation 8-1.
where:
Vi»Jia—VWurne of gas sample through tho
dry gft3 meter (standard condi-
tions) , cu. ft.
T,u—Absolute temperature at standard
conditions, 630° R.
T^—Average dry gas meter temperature,
•B.
Pj,,— Barometric pressure at the orlBcfi
meter. Inches Hg.
Oj— Concentration ol sulfurlc acid
at standard conditions, dry
basis, ib./eu. ft,
1.03 x KH— Conversion fnctor InclucUng tae
number of grams per gram
equivalent of sulfuilc acid
(49 g./g.-eq.), 453.6 g./lb,, and
1,000 ml./l., lb,-l./g.-ml.
V, —Volume of barium parchlorate
tltrant used lor tho sample,
ml.
V^-"Volume of bnrlum perchlorate
tltrant used for the black, ml.
srhere:
Cso,™ Concentration of sulfur d'.oxJde
at standard conditions, dry
basis, ib./cu, ft.
T.05X 10-*— Conversion factor including the
numbsr of grams per gram
equivalent of sulfur dioxide
(32. g./g.-eq.) 453.6 g./lb., and
1,000 ml./l.. lb.-l./g.-ml,
V,= Volume of barium perchlorate
tltrant used for tho sample,
ml.
PbK,+
\~~~T;
equation 8-1
Vo'-ump of SHE semple through tha
dry gas meter (meter condi-
tions) . en. ff
AM - Press. ire drop acrosa Uie crlQca
me'-er. Inches H_O,
13.6"=SiXici..c LTS"!'? o/ mercury.
P. ,j— Absolute pressure at standard con-
ditions, 29.02 Inches Hg,
8.2 fiul'iirlc acid concentration,
* = •!<• equation fi-2
W = Normality of barium perchlorat*
litrant, g.-eq./l.
Vlolll=*Ti.'',al solution volume of eul-
furlc acid (nrst impinger and
filter), ml.
V<—Volume of sample aliquot ti-
trated, ml.
Va,ld-= Volume of gas sample through
the dry gas meter (standard
conditions), cu, ft., see Equa-
tion 8-1.
fi.S Sullur dioxide concentration.
/Y" ,
/ V S I Jr'-n
\'«
equation S-3
Vtk— Volume of barium perchlorata
tltrant used tor the blank, ml.
W— Normality of barium perchJorate
tltrant, g,-eq./l.
7loU-> Total solution volume of sulfur
<31oslde (second and third irn-
plngers), ml.
Vj—Volume of sampie aliquot ti-
trated, ml,
V =,.,„•= Volume of eas sample ttuough
tte dry gas meter (standard
conditions), cu. ft., see Equa-
tion 8-1.
379
-------
7. References,
Atmospheric Emissions from Sulfurlo Acid
Manufacturing Processes, U.S. DHEW, PHS.
Division o! Air Pollution. Public Health Serv-
lea Publication No. 909-AP-13, Cincinnati,
Oalo, 1965.
Coroett, D. P., Tlie Determination of SO,
and SO, In Flue Oases, Journal of the Insti-
tute 01 Fuel, 24:237-243, 1061.
Martin, Robert M.. Construction Details of
Isoklnetlo Source Sampling Eo.uipment. En-
vironmental Protection Agency, Air Pollution
Control Office Publication No. APTD-0581.
Pal ton, w. F., and J. A. Brink, Jr., New
Equipment and Techniques Tor Sampling
Chemical Process Gases, J. Air Pollution Con-
trol Assoc. 13, 102 (1963).
Rom, Jeroma J., Maintenance, calibration,
and Operation of Isokir.etlo Source Sam-
pling Equlprfteut, Environmental Proi&ctlon
Agency, Air Pollution control Office Publi-
cation No. APTD-C576.
Shell Development Co, Analytical Depart-
ment, Determination ol Sulfur Dioxide and
Sulfur Trlcxlde in Slack G&ses, EmeryTill«
Method Series, 4516/59a.
THERMOMETER
CHECK
VALVE
REVERSE-TYPE
PITOT TUBE
VACUUM
LINE
VACUUM
GAUGE
MAIN VALVE
'AIR-TIGHT
PUMP
DRV TEST METER
Figure 8-1, Sulfuric acid mist sampling train.
rtV-'lr.iiuF'. I t-EAn
380
-------
METHOD 9
Visual Determination Of The Opacity Of Emissions From
Stationary Sources
Many stationary sources discharge visible
emissions Into the atmei-phere; these emis-
sions are usually In the shape of a plume.
This muthod Involves the extermination of
plume opacity by qualified observers. The
method includes procedures for the training
and certification of observers, and procedures
to be xiscd In the field for determination of
plume opacity. The apptiuar.ee of a. plume as
viewed by an observer depends upon a num-
ber of vtriablcs, some of tthlch may be con-
trollable and some of whic:i may not be
controllable in tho field. Variables which ciii
be controlled to an extent to whlnh they no
longer exert a significant Influence upon
plume appearance Include: Angle of the Ob-
server with respect to the plume; angle ol the
observer with respect to the sun: point of
observation of attached and detached steam
plume; and angle of the observer -xliti re-
spect to a plume emitted from a rectangular
stack Tilth a large length to width ratio. Tte
method Includes specific criteria applicable
to thsse variables.
Other variables which may not be control-
lable In the field are luminescence and color
contrast between the plume and the back-
ground against which the plume Is viewed.
These variables exert an Influence upon the
appearance of a plume as viewed by an ob-
server, and can affect the ability of the ob-
server to accurately assign opacity values
to ihe observed plume. Studies of the theory
of plume- opacity and field studies have dem-
onstrated that a plume Is most visible and
presents the greatest apparent opacity when
viewed against a contrasting background. It
follows from this, and is confirmed by field
trials, that the opacity of a plume, viewed
•under conditions where a contrasting back-
ground Is present can be assigned with the
greatest degree of accuracy. However, the po-
tential for a positive error Is p.lso the greatest
when a. plume Is viewed under such contrast-
ing conditions. Under conditions presenting
a less contrasting background, the apparent
opacity of a plume Is less and approaches
zero as the color and luminescence contrast
decrease toward zero. As a result, significant
negative bias and negative errors can be
made when a plume Is viewed under less
contrasting conditions. A negative bias de-
creases rather than Increases the possibility
that a plant operator will be clwd for a vio-
lation of opacity standards due to observer
error.
Studies have been undertaken to determine
the magnitude of positive errors which can
be made by qualified observers while read-
Ing plumes under contrasting conditions and
using the procedures set forth In this
method. The results of these studies (field
trials) which involve a total of 769 sets of
25 readings each are as follows:
"For a set, positive error=average opacity
determined by observers' 25 observations —
average opacity determined from transmls-
someter's 25 recordings.
(1) For black plumes (133 set? at a smoke
generator), 100 percent of the sets were
read with a, positive error > of less tban 7.5
percent opacity; 99 percent were read with
a positive error of less than 5 percent opacity.
(2j For white plumes (170 sets at a smoke
generator, 168 sets at a coal-fired power plant,
298 =ets at a sulfurlc acid plant), B9 percent
of the sets were read with a positive error of
less than 7.5 percent opacity; S5 percent were
read with a positive error of less than 5 per-
cent opMtry,
The positive observational error associated
with an average of twenty-five readings Is
therefore established. The accuracy of the
method must be taken Into account when
determining possible violations of appli-
cable opacity standards.
1. Principle and applicability,
1.1 Principle. The opacity of emissions
from stationary sources Is determined vis-
ually by a qualified observer.
1,2 Applicability. This method is appli-
cable for the determination of the opacity
of emissions from stationary sources pur-
suant to §00.11 (to) and for qualifying ob-
servers for visually determining opacity ol
emissions.
2, Procedures. The observer qualified in
accordance with paragraph 3 of this method
shall use the following procedures for vis-
ually determining the opacity of emissions:
2.1 Position, The qualified observer rhail
stand at a distance sufficient to provide a
clear view of the emissions with the sun
oriented In the 140° sector to his back. Con-
sistent with maintaining the above require-
ment, the observer shall, as much as possible,
make his observations from a position such
that his line of vision Is approximately
perpendicular to the plume direction, ar.d
when observing opacity of emissions from
rectangular outlets s'e.g roof monitors, open
bn.ghoiose.3, nonclrcular stacks), approxi-
mately perpendicular to the longer axis of
the outlet. The observer's line ol sight should
not Include more than one plume at a. time
when multiple stacks are involved, and In
any case the observer should amfce his ob-
servations with his line of sight perpendicu-
lar to the longer axis of such a set of multi-
ple stacks (e.g. stub stacks on baghousesj.
2,2 Field records. The observer shall re-
sord the name of the plant, emission loca-
tion, type facility, observer's name and
affiliation, and the date on a. field data sheet
(Figure 9-1). The- time, estimated distance
to the emission location, approximate wind
direction, estimated wind speed, description
of the sky condition (presence and color of
clouds), and plume background are recorded
on a. field data sheet at the time opacity read-
ings are Initiated and completed.
2.3 Observations. Opacity observations
shall be made at the point of greatest opai'.ty
in that portion of trie plume where con-
densed water vapor is not present. The ob-
server shall not look continuously at the
plume, but Instead shall observe the plume
momentarily at 15-second intervals.
381
-------
2.3.1 Attached steam plumes. When con-
densed water vapor la present within the
plume as It emerges from the emission out-
let, op«iclty observations shall 1)c made be-
yor.d tht! point hi the plume at which con-
densed water vapor is no longer visible. The
observer shall record the approximate dis-
tance from the emission outlet to the point
In the plume at which the obscrvatlciis are
made.
2.3.2 Detached steam plume. When water
vapor in the plume condenses and becomes
visible at a distinct distance from the emis-
sion outlet, the opacity of emissions should
be evaluated a! the emission outlet prior to
the condensation of water vapor and the for-
mation of the steam plume.
2.4 Recording observations. Opacity ob-
servations shall be recorded to the nearest 5
percent at 15-second Intervals on an ob-
servational record sheet. (See Figure 9-2 for
an example.) A minimum of 24 observations
shall be recorded. Each momentary observa-
tion recorded shall be deemed to represent
the average opacity of emissions lor a 15-
sccoiiu period.
2 5 Data Reduction. Opacity shall be de-
termined as an average of 24 consecutive
observations recorded at 15-second intervals.
Divide the observations recorded on the rec-
ord sheet Into sets of 24 consecutive obser-
vations. A set is composed of any 24 con-
secutive observations. Sets need not be con-
secutive in time and in no case shall two
sew overlap. For each set of 24 observations,
calculate the average by summing the opacity
o,' the 24 observations and dividing this sum
by 24, If an applicable standard specifies an
averaging time requiring more than 24 ob-
servfitions, calculate the average for all ob-
servations made during the specified time
period. Record the average opacity on a record
sheet, (See Figure 9-1 for an example.)
3. Qualifications and testing.
3.1 Certification requirements. To receive
certification us a qualified observer, a can-
didate must be tested and demonstrate the
ability to assign opacity readings in 5 percent
Increments to 25 different blac's plumes and
25 different white plumes, with an error
not to e.vcsed 15 percent opacity on any one
reading and an average error not to exceed
7.5 percent opacity In each category. Candi-
dates shall be tested according to the pro-
cedures described In paragraph 3.2. Smoke
generators u.srd pursuant to paragraph 3.2
shall be equipped with a .smoke meter which
meets the requirements of paragraph 3.3.
The certification shall be valid for a period
of 6 months, at which time the qualification
procedure must be repeated by any observer
in orti<=r to retain certification,
3 2 Certification procedure. The certifica-
tion te^-t consists of showing the candidate a
complete run of 50 plumes—25 black plumes
und 25 white plumes—generated by a smoke
generator. Plumes within each set of 25 blacK
and 25 white runs shall be presented in ran-
dom order. The candidate assigns an opacity
value to each plume and records his obser-
vation on a suitable form. At the completion
of each run of 50 readings, the score of the
candidate Is determined. If a candidate fails
to cmallfy, the complete run of 50 readings
must be repeated In any retest. The smoke
test may be administered p,s part of a smoke
school or training program, and may be pre-
ceded by training or familiarization runs of
the smoke generator during wli'r.h candidates
are sho\vr> black and white plumes of known
opacity.
3.3 Smoke generator specifications. An?
smoke generator u?ed for the purposes of
paragraph 3.2 shall be equipped with a smnke
meter installed to measure opacity across
the dlanet-er r>f the smoke generator st_\ck.
The smoke meter output 5hal! display in-
scack opacity ba"ed upon a pathlcng'.h equal
to th2 stack exit diameter, on a full 0 tn 100
percent chart recorder .scale, The smnke
moter optical desipn p.nd performance shall
meet tho specifications shown ia Table 9-1.
The smoke me-ier shall be calibrated as pre-
scribed in paragraph 3.3.1 prior to the con-
duct of each smako reading test. At the
completion of each test, the zero and spaa
drift shall be checked and if the drift ex-
ceeds =1 percent, opacity, the condition shall
be corrected prior to conducting any subse-
quent tf-st runs. The smoke meter shall be
demonstrated, a,t the time of Installation, to
meet the speci'isntions listed In Table 3-1.
1h;s demonstration shall be repeated fol-
lowing any subsequent repair or replacement
of t:ie photocell or associated electronic cir-
cuitry including the chart, recorder <,r output
meter, ar every 6 mouths, wh.chever occurs
ilr.st.
TABLE 9-1 SMOKE METER DESIGN- AKD
FEKFORMANEE SPECUTCATIONS
Parameter: Specification
a. Llgh; source Incandescent lamp
operated, at nominal
rated voltage.
b. Spectral response Photopic sdaylight
of photocell. spectral response" of
ihe human eye—
relerer.ee 4 3).
c. Angle of view 15" maximum total
arigle,
d. An^l? of projec- 15" maximum total
tion. angle.
e. Calibration error- ±3',i opacity, maxl-
nv.im.
f. Zero and spa.1 ±rl "t- opacity, 30
drift. minuses.
g. Response time— 5.'o seconds.
3 n, 1 Calibration. The smoke .lie'or Is
calibrated after Allowing a '.-.inhnu:!'. cf 30
r-'l.m"^ wa'.'mup b;: .V.tr-r-if.fe'v pixducS'tg
simulated opacity of 0 perorrit and ion per-
cent. VV'hrn stable reF;'rr.Pi fit 0 ne-oent or
lOu pcrrf.-t t? rote:;, tho env:ki> meter 1- ad-
Justed to prsduce an output ;>f 0 percent or
ICO pc.-::.-n:. ar Apprnpr'/iU' This cal.'ir.tiion
sltall be repented until "table j percent and
300 per-en.-, re.idin?5 are prijJncerl without
adjustment. Simulated 0 perernr £n:l :CiO
percent opsci;;. values nay be precluded by
alternately s-.vitchinj th; power to '.!~a- li^h;
sourcB 0:1 and o!T v hllc .he smoke generator
is n-!t prcd'.i'ir-j siro-'e.
332 Snohe meter e\'alu.T:r:i. The smoke
moter c*e.jkcu and pcrf^rrri^n^3 are 10 be
ev".:>'a:ed as follows:
3 3 2.1 Light source. Verify from mar.u-
farturer'r, (lata and from •,•<••!:?.(;? ineisure-
ments r.iade at the lamp, is installed, that
the l:imp ;^ operated \\cith!n —s percent of
the norr.inal rawd \al"age.
3.3.2.2 Spectral rospcnse af photocell.
Verify from manufacturer's data that the
photocell has a photopic response: ! e,, the
spectral sensltK'iTv of 'he fell shall closely
approximate the standard ^pnctr^il-Juminos-
itr curve for photonic "Islon which L5 refer-
enced in ;b) of Tible 9-1.
382
-------
3.32.3 Angle of view. Check construction
geometry to ensure that ihe tccal angle or
v,c-,v o! the smoke plurr.o, as seen by [he
photocell, does not exceed 15'. Tho total
angle of view may be calculated from: » = 2
taiv; ci-2L, where 5 —total angle of view;
d = £he sum of the photocell diameter — the
diameter of the li-T.lting aperture; and
L=:-te distance from the pho:oreH to the
:'.m!:!:ig aperture. The lirr.i'.Uip aperture is
the pc;nt in trie path between the photocell
tha sinokc phtme v here
of
v!ev- ;s most restricted. In srr.oka pcnrr.'Ho;
Emnke me'.ers this is normally an or.fice
piav
33.24 Angle of projection. Check con-
struction geometry to ensure that the total
angle of projection of the lamp on the
smoke plume docs not excead. 15". The total
angle ol projection may be calculated from;
f—2 tan-' ci 2L. where 0— total angle of pro-
jection; (1= the sum of the length, ol the
lamp filament 4- the diameter of the limiting
aperture; and L= the distance from the lamp
to the limiting aperture.
3.3.2.5 Calibrailoii error. Using neutral-
density filters of >:nowii opacity, check the
error between the actual response ar.d the
theoretical linear response of the smoke
meter. This check Is accomplished by first
calibrating the smoke meter according to
3.3,1 and then Inserting a series of three
neutral-density filters ol nominal opacity of
20, 50, and 75 percent In the smoke meter
pathlength. Filters calltaarted within ±2 per-
cent shall be used. Care should be taken
when Inserting the filters to prevent stray
light from affecting the meter. Make a total
of five nonconsccutlve readings for each
filter. The maximum errcr on any one read-
ing shall be 3 percent opacity.
3.3.2.6 Zero and span drift, Determine
the zero and span drift by calibrating and
operating the smoke generator In a normal
manner over » 1-hour period. The drift IE
measured by checking the zero and span at
the end of this period.
3.3.2.7 Response time. Determine the re-
sponse time by producng the series of five
simulated 0 percent and 100 percent opacity
values and observing the time required to
reach stable response. Opacity values of 0
percent and 100 percent may be slm-Jlated
by alternately switching the power to the
light Eourcs on and on while the smoke
generator is not operating.
4, References.
4.1 Air Pollution Control District Rules
and Regulations, I,oa Angeles County Air
Pollution Control District, Regulation IV,
Prohibitions. Rule 60.
4.2 WelEbuid, Melvln I., Field Operations
and Enforcement Manual for Air, U.3, Envi-
ronmental Protection Agency,.Research Tri-
angle Park, N.C., APTD-1100, August 1972.
pp. 4.1-4,36.
4.3 Condon, E. D., and Odishaar, H., Hand-
book oj Physics, McGraw-HlU Co., N.T.. N.T.,
1B58, Table 3.1, p. S-52.
RCCC" Cr VISl'-'-L fETi'.
LOCA'IO'!
o:-i:rM:p. '.^.27 ION
Dfs'.-j'.ce :s S's-rharge
Birectlcr frc- Clsc^3rg^
HeigH of Cbservttlon "O'.nt
E'CCP.0'.'.C DESCPPTIO'i
'.-l 5=eed
1O CC'OIr;CSS (c'car,
CvCrCiSt, X C'OlC'Si CtC.)
nir_t.'-ce V'i -.Irle
CI',:' INi;""-7;:?i
^t ,_, -'•<••- I o^.-^>
"•-~^r Start-r.'d rs..:r, I t'
rc2 v.-.s''ia5 pot ii cc-p"
? ?va 1 J.11icn i.?5 ride-.
383
-------
COMPLY
LO"4T I"!1;
TEST .'K'MSiT
DATE
FIGURE 9-2 OBSERVATION RECORD
(Continued)
OBSERVER
TYPE FACILlT-
PAGE OF
POI'iT OF Ff'ISSlO'lS
Hr.
Mi--;.
33
3'
32
33
V.
35
35
37
38
39
40
<•> i
^2
43
i'1
45
46
47
48
49
50
51
5?
53
54
55
56
57
SB
59
0
Se
!s
coif
2)
s
;"-b
(ch«ckV
Att^chco
PLUME
eDDlir^ tl e)
) Det3cneri
J
: COMMENTS
t
385
-------
COMPLY
LO"4T I"!1;
TEST .'K'MSiT
DATE
FIGURE 9-2 OBSERVATION RECORD
(Continued)
OBSERVER
TYPE FACILlT-
PAGE OF
POI'iT OF Ff'ISSlO'lS
Hr.
Mi--;.
33
3'
32
33
V.
35
35
37
38
39
40
<•> i
^2
43
i'1
45
46
47
48
49
50
51
5?
53
54
55
56
57
SB
59
0
Se
!s
coif
2)
s
;"-b
(ch«ckV
Att^chco
PLUME
eDDlir^ tl e)
) Det3cneri
J
: COMMENTS
t
385
-------
METHOD 10
Determination Of Carbon Monoxide Emissions From
Stationary Sources
1- Principle and
1.1 Principle. An uusgnited or continuous
gas sample Is extr&c'.eii from t, sampling point
a:id analyzed for cr.rbon monoxide (CO) con-
tent using a Tjufl-tyye nondispersive Infra-
red ar.aly-er fKDJR) or equivalent,
1.2 Applicability. This method is appli-
cable for the determination of carbo?i mon-
oxide emissions froJr. stationary sources only
when specified by the test procedures lor
determining compliance \v!th new source
performance standards. The test procedure
will Indicate u-hether fl continuous or an
integrated sample is to be used.
2. Range and sensitivity,
2.1 B.jnne. 0 to 1,000 ppm.
2.2 Sensitivity. Minimum detectable con-
centration is 20 ppai for a 0 to 1,000 ppm
span.
3. Interferences- Any substance having a
strong absorption of Infrared energy will
interfere to tome extent, For example, dis-
crimination ratios for water (H.D) and car-
bon dioxide (CO.) are 3-5 percent H,O per
7 ppm CO and ib percent CO, per 10 ppm
CO, respectively, for devices measuring In the
1,500 to 3,000 ppra range. For devices meas-
uring in tlie 0 to 200 ppm range, interference
ratios can be as high as 3.5 percent H..O per
25 ppm CO and 10 percent CO, per 50 ppm
CO. The vise of silica gel and ascarlte traps
will alleviate the major Interference prob-
lems. The measured gas volume must be
corrected If these traps are used.
4. Precision and arcuracy.
4.1 Precision. The precision of most NDIR
analyzers is approximately ±2 percent or
span.
4,2 Accuracy. The accuracy or most NDIR
analyzers is approximately ±5 percent ol
span after calibration,
5. Apparatus.
6,1 Continuous sample (Figure 10-1).
S.l.l Probe. Stainless steel or sheathed
Pyrex i glass, equipped with a filter to remove
particulars matter.
5.1.2 Air-cooled condenser or equivalent,
To remove any excess moisture.
5.2 Integrated sample (Figure 10-2).
. 5.2.1 Probe. Stainless steel or sheathed
Pyrex glass, equipped with a filter to remove
participate matter.
5.2.2 Air-coolefl. condenser or equivalent.
To remove any excess moisture.
5.2.3 Valve. Needle valve, or equivalent, to
to adjust flow rate.
5.2,4 Pump. Leafc-free diaphragm type, or
equivalent, to transport gas.
5.2.5 flare meter. Rotameter. or equivalent,
to measure a flow range from 0 to 1.0 liter
per mln. (0.035 cfm).
5,2.6 Flexible bag. Tedlar, or equivalent,
•With a capacity of 60 to 90 liters (S to 3 ft').
Leak-test the bag In the laboratory before
using by evacuating bag xvlth a pur.ip fol-
lowed by a ciry gas meter. When evacuation
is complete, there should be co flow through
the meter.
5.2.7 PHOf tube. Type S, cr equivalent, at-
tached to tte probe so that the sanpling
rate can bn regulated proportional to the
stach gas veloi;l\ when velocity is vr.rn^g
with the time or a sample traverse
ducted.
5,3 Analysis (Fig-jre 10-3).
5.3.1 Carbon monoxide ar.alyzfr. No
sivc infrared spectrometer, cr equivalent.
This instrunient should be domoi.strayed,
preferably by the manufacturer, to meet, or
exceed manufacturer's specifications and
t]JO?e described in this method.
5.3,2 Drying t!;be. To contain approxi-
mately 20G 2 Of sUica gel.
5.3.3 Ca'Mraticn ga£- Refer to paragraph
6.:.
5,3,4 Filter. As recommended by N
manufacturer.
6.3.5 CO. removal turje. To contain approxi-
mately 500" g of atcorlte.
5.3.6 Ice water bath. For ascarlte and silica
gel tubes.
5.3.7 Valve. Needle valve, or equivalent, to
adjust flov.- rate
5.3.8 Rate meter. Rotameter or equivalent
to measure gas flow rate of 0 to 1.0 Ht«r per
min. (0,035 Cfm) through NDIR.
5,3.9 Recorder (optional). To provide per-
manent record of NDEE readings.
6. Reagents,
1 Mention of trade names or specific prod-
ucts does not constitute endorsement by the
Environmental Protection Agency.
386
-------
6 1 Cc/itrcJiort gases, Krxv.vn concentration
of CO In nltrorea (N^) for ins'.rumrnt span,
prepurLfled grade of N- for zero, aixi two addi-
lior.al concentrations corresponding approxi-
mately to ce percent ar.d in pr-rrcr.t ?pan. The
sp-in eor.cerv it Ion shal1. net exceed 1.5 Times
xl-.s iopllca-..s source pt-rrcrmnuce .standard.
The call >ration irase^ Pball Lie certified by
the .'rruiuf&cturer :o be within ±2 percent
of thv specified co.ioeiitralion.
6.2 Silica gel. Indicating type, 6 to 16 mesh.
dried at 175' C (347° F) for 2 Hours.
0.3 Azcaritc. Commercially available.
7. Procedure.
7.1 £cmpjin the stack velocity.
CCX >"v::i.ent of UIL- gr.5 mn.y be determined
by 1.,-u,..' tin? Met.iod 3 integrated sample
pruciHi-ircs (KG 1"R 2-136C), or by weighing
the ii :ur:te CO. rctr.uval lu'«e and conipuc-
ing CO.j concentraiion iru:a the gas volume
sanipled and the wei^h*. y^a.n of the tube,
7.2 CO ATitili,3i3. Arscmtle the apparatus a3
shnwn in Figure 10-3, calibrate the instru-
ment, anfl peri'or.'n other required operaticr.s
aj described In paragraph 8. Purge analyzer
w:tn K_. prior to introduction ol each sample.
Direct the sample stream through the instru-
ment for the test penad, recording the read-
ings. Check the zero and span aga:n after the
test to assure rhat any drift or malfunction
is detected. Record the sample data on Table
10-1.
a. Calibration. Assemble the apparatus ac-
cording to Figure 10-3. Generally an instru-
ment requires a '.varm-up period before sta-
bility is obtained. Follow the manxifacturer's
instructions lor specific procedure. Allow a
minimum time of one lioar for warm-up.
During this time check the sample cor.cl-
tior.ing apparatus, I.e.. filter, condenser, dry-
ing tube, and CO2 removal tube, to ensure
that each component Is in good operating
condition. Zero and calibrate the Insirutne^
according to the manufacturer's procedures
using, respectively, nitrogen and the calibra-
tion gases.
oxide. Calculate the concentration of carbos
equation 10-1
C"co,..cl=connentriition of CO in stack, ppm b.v volume (dry b.inis).
('cri_s,J]E = cpnrrntration of CO measured by NDIli r1r,i]yz'--r, pprn by volume (dry
f (.o- = vi Inn *• fr.irijrii of COj in farr|>!e,
di\:d from Orsat iinrJj-sifl
MSA LIRA Infrared Gas and, Uqula
Analyzer Instruction Bosk, Mine Safety
Appliances Co.. Technical Products Di-
vision, Pittsburgh. Pa.
Mcdsls 215A, 315A, and 415A Infrrired
Analyzers, Beckmau Instruments, Inc.,
Deckman Instructions 1535-B, Fuller-
ten, Calif., October 1967.
Continuous CO Monitoring System,
Model A5611, Intertech Corp,, Princeton,
U.VOH Infrared Gas Analyzers, BendlJ
Corp., Ronceverte, West Virginia,
TABLE 10-1.—Field data
Location .
Teat ____
Comments:
Operator ,
Clock time
Rotarrteter setting, liters per minute
(cubic fest per minute)
387
-------
ADDENDA
A. Performance Specifications for ND1R Carbon Monoiide Anclyzcra.
Range (minimum) O-lOC-Oppm.
'Output (minimum) 0-lCm.V.
Minimum detectable sensitivity 20 ppm,
Rise time, 90 percent (maximum) 30 seconds.
Fall time, 90 percent (maximum) 30 seconds,
Zero drift (maximum) 10% In 8 hours.
Span arm (maximum) 10^, In 8 hours.
Precision (minimum) ± 2 % of full scale.
Noise (maximum) ± 1 To o! lull scale.
Linearity (maximum deviation) 2% o! full scale.
Interference rejection ratio CO.—1000 to 1, H^O—500 to 1.
B. Definitions of Performance Specifica-
tions.
Range—The minimum and maximum
measurement limits.
Output—Electrical signal wWch Is propor-
tional to the measurement; Intended for con-
nection to readout or data processing devices.
Usually expressed as millivolts or mlillamps
full scale at a given Impedance.
Full scale—The maximum measuring limit
for a given range.
Minimum detectable tensiiivity—The
smallest amount of input concentration that
can be detected aa tha concentration ap-
proaches zero.
Accuracy—The degree of agreement be-
tween a measured value and the true value;
usually expressed as -±_ percent of full scale,
Time to 3D percent response—The time In-
terval from a step change In the Input con-
centration at the Instrument inlet to a read-
Ing of 90 percent of the ultimate recorded
concentration,
Rise Time (90 percent]—The Interval be-
tween initial response time and time to 90
percent response after a step Increase In the
inlet concentration.
Foil Time (90 percent)—The Interval be-
tween Initial response time and time to 90
percent response after a step decrease In the
Inlet concentration.
Zero Drift—The change In Instrument out-
put over a stated time period, usually 21
hours, or unad]usted continuous operation
when the input concentration Is zero; usually
expressed as percent full scale.
Spun Drift—The change In Instrument out-
put over a stated time period, usually 24
hours, of unadjusted continuous operation
when the Input concentration Is a stated
•upscale value; usually expressed as percent
full scale.
Precision—The degres of agreement be-
tween repeated measurements or the same
concentration, expressed as the average de-
viation of the single results from the mean.
Noise—Spontaneous deviations from a
mean output not caused by Input concen-
tration changes.
Linearity—The maximum deviation be-
tween an actual Instrument reading and the
reading predicted by a straight line drawn
between upper and lower calibration points.
388
-------
METHOD 11
Determination Of Hydroge'n Sulfide Emissions From
Stationary Sources
1, Principle and applicability,
1,1 Principle. Hydrogen sulflcle (HjSJi 13
collected from the source In a series or midget
Implngcrs and reacted with alkaline cad-
mium hydroxide [Cd(OH),l to form cad-
mium suiflde (CdS). The precipitated CdS
Is then dissolved In hydrochloric acid and
absorbed In ft known volume of Iodine solu-
tion. The iodine consumed Is a measure of
the HaS content of the gaa. An Imptngar con.
talnlng hydrogen peroxide Is Included to re-
move SO, as an Interfering species.
l.a Applicability. This method Is applica-
ble for the determination of hydrogen sml-
fido emissions from stationary sources only
wten specified by the test procedures Tor
determining compliance with the new source
performance standards.
2. Apparatus.
2.1 Sampling train.
2.1.1 Sampling line—6- to 7-mm (Ji-lach)
Teflon ' tubing to connect sampling train to
sampling valve, with provisions for heating
to prevent condensation, A pressure reduc-
ing vatre prior to the Teflon sampling line
may be required depending on sampling
stream pressure.
2.1.2 Iriipingers—Five midget implngers,
each with 30-ml capacity, or equivalent.
2.1.3 Ice bath container—To maintain ab-
sorbing solution at a constant temperature,
2.1.4 Silica, gel drying lube—To protect
pump tuid cry gas meter.
2.1.5 Needle valve, or equivalent—Stainless
steel or other corrosion reslKtanr material, to
adjust gas Eow rate.
2.1,6 Pump—Leak free, diaphragm typo, or
equivalent, to transport tra5- (^ot required
If sampling stream ur.der positive pressure.)
2.1.7 Dry gas meter—Sufficiently accurate
to measure sample volume to within 1 per-
cent.
2.1.B Rate meter—notameter, or equivalent,
to measure a flow rate of 0 to 3 liters per
minute (O.l fts/mln),
2,1.9 Graduated cylinder—25 ml.
2.1.10 Barometer—To measure atmospheric
pressure within ~2.5 ram (0.1 In,) Hg.
2.2 Sample Recovery.
2.2." Sample container—500-ml glass-stop-
pered Iodine flask,
2.2.2 Pipette—50-ml volumetric type.
2,2.3 Beakers—250 ml.
2,2.4 Waafi dottle—QlaS9.
2.3 Analysis.
2,3.1 Flask—500-ml glass-stoppered iodine
flask.
2,3,2 Burette—One 50 ml.
~2.3.2 Finsk—125-mI conical.
3. Reagents.
3.1 Sampling.
3.1.1 Absorbing solution—Cadmium hy-
droxlcie (Cd(OH),)—Mix 4.3 g cadmium sul-
fate hydrate (3 CdSO4.8HaO) and 0.3 g of
sodium hydroxide (NaOH) In 1 liter of dis-
tilled wnter (H/3). Mix well,
1 Mention of trade names or specific prod-
ucts does not constitute endcrsement by the
Environmental Protection Agency.
Note: Tne cadmium hydroxide formed In
this mixture will precipitate as a '.vhlte sus-
pension. Therefore, this solution must be
thoroughly mixed Before using to ensure an
even distribution of the cadmium hydroxide,
1.1,2 liyil-rcigcTi peroxide, 3 percent—Dilute
30 p~rcenf hydrogen peroxide to 3 percent
as needed. Prepare fresh dally;
3.2 Sample recovery.
3.2.1 Hydrochloric acid solution (HCl), to
percent by it-tight—mix 230 :i-.l ot concen-
trated HC1 (specific jri-avlty 1.19) raid 770 tr.l
of distilled H2O.
3.2.2 Iodine sol.i*. ion, 0,1 N—Dissolve 24 r,
potn.isiu.rn iodide (KI) in 30 ml of distilled
H.O In a 1-llter graduated cylinder, Weigh
12.7 g of resuMlmed iodine (IJ luto a wclgl:-
inj bottle and add to the potassium iodide
solution. Shake the mixture until the lodrr.e
Is completely dissolved. Slowly dilute the so-
lution to 1 liter with distilled H.O, with
skirling. Filter the solution, it cloudy, and
store 111 ft brcv.Ti £l"'*s-,-3toppcred bottle.
32,3 Standard iodine solution, 0,01 N—Di-
lute 100 ml of the 0,1 N iodine solution lr. a
volumetric flask to 1 liter with distilled
wp.ter.
Standardize dally as follows: Pipette 2j a\l
of tV.o 0.01 ,V Iodine solution Into a 125-ir.i
co: leal aask. Titrate v.-lth standard 0,01 ;V
thlosulf.ite solution (see paragraph 3,3,2) un-
til the solution is a light y.llcsw. Add a few
drops of the starch solution and coniHiue
titrating until the blue color Just disap-
pears rram tlia results of thl.H titratlon, c-il-
cuUite the exnct normality of tlic loriine
sol!.u":n (see paragraph 5.1).
3,2.1 Dls'tl'."d,dcionized u-aier.
3.3 Analysts.
3.3.1 Sodium Ihios-ulfaic solution, slandsfd
0,1 N—For each liter of solution, dLssohe
24.B g of sodium thlosulfate (KA,SjO, 5R.O)
In distilled water arid add 0 01 g of anhydrous
sndtuni cBrbG'iate (KajCO,^, and 0,4 ml of
chloroform tCIICL,) to" strhilize. Mix thor-
tntghly by EhakLn^ or by aentu-g \MCII intro-
Kt-',\ for approximately 15 minutes, and store
111 a gl.iss-stopporccl giasa bottle,
Standardize frequently as follows: Wei^h
into a 500-ml volumetric flnsk about 2 g of
potassium dlchromate (K.Cr O.) weighed
to the nearest milligram arid "dilute to the
500-ml mark with distilled H.,O. TJst- dl-
chromate v/hlch has been crystallized from
distilled '.vater and oven-dried at 182°C ta
109*0 s3oO'F to 390'F). Dissolve approxi-
mately 3 g of potasbium lodlcle (KI) in 50 nil
of disuUled water In a glass-stoppered, 500-ml
conical flash, tiieu add 5 ml of 20-percenl
hydrochloric acid solution, Pipette 50 ml of
the dlchromate solution Into this mixture,
Gem LI y swirl the solution once and allow It
to stand in the dark for 5 minutes. Dilute
the solution with lOO.to 200 ml of distilled
water, washing down the sides of the P.aslc
with part of the water. Swirl the solution
slowly'aiid titrate with the thlosulfate solu-
tion until the solution Is light yellow. Add
4 ml of starch solution and continue with a
389
-------
slow titration with the thlosuUate until the
bright blue color has disappeared and only
the pale green color of the chromic 1cm re-
mains. From this titration, calculate the ex-
act normality of the sodium thloaulfate solu-
tion (see paragraph 5.2),
3.3-2 Sodium thiosulfate solution, standard
0.01 N—Pipette 100 ml of the standard 0.1 N
thiosulfate solution Into a volumetric flast
and dilute to one liter with distilled water.
3.3.3 Starch, indicator solution—Suspend
10 g of soluble starch In 100 ml of distilled
water and add 15 g ol potassium hydroxide
pellets. Stir until dissolved, dilute wilh 900
ml of distilled water, and let stand 1 hour.
Neutralize the alkali with, concentrated hy-
drochloric actd, using an indicator paper
similar to Alkacld test ribbon, then add 2 ml
of glacial acetic acid as a preservative.
Test lor decomposition by titrating 4 ml of
starch solution in 200 ml of distilled water
with 0.01 N Iodine solution. If more than 4
drops ol the 0.01 N loilne solution are re-
quired to obtain the blue color, make up a
fresh starch solution.
4. Procedure,
4.1 Sampling.
4.1.1 Assemble the sampling train as shown
la Figure 11-1, connecting the five midget
implngers In series. Place 15 ml of 3 percent
Hydrogen peroxide In the first Lmpinger. Place
15 ml of the absorbing solution In each of
the next three implngers, leaving the fifth
dry. Place crushed ice around the Implngers.
Add more Ice during the run to keep the
temperature of the gases leaving the last
impinger at about 20°C (70'F), or less, •
1.1.2 Purge the connecting line between
the sampling valve and the first Impinger,
Connect the sample line to the train. Record
the initial reading on the dry gas meter aa
shown in Table 11-1.
11-1.—FislA data
Location Comments:
Test
Date
Operator
Barometric pressure..
Clock
cin.e
G&a volume
Ihrcueh
TjietT (VJ,
liters tcub'.c
feel)
Ttotiuiieter
irui'.n- Feet
por :Tnnuie>
Meter
to*npftr iti'"9
• C t° Kj
4.1.3 Open the flow control vnlve and ad-
just tb% sampling rate to 1.13 liters per
minute (0.04 Mm). Read the meter temper-
ature and record on Table 11-1.
4.1.4 Continue sampling a minimum of 10
minutes. If the yellow color of cadmium sul-
fide is visible in the third impinger, analysis
should confirm that the applicable standard
.has been exceeded. At the and of the sample
'time, close the Cow control valve and read
'the fl-jal meter volume and temperature,
4.1.5 Disconnect tlie impinger train from
the sampling line. Purge the train with clean
ambient air for 15 minutes to ensure that all
JH,S Is removed from the hydrogen peroxide,
:Cap the open ends and movo to the sample
clean-up area.
4,2 Sample recovery.
4.2.1 Pipette 50 ir.l of O.Ol .V iodine solution
Imo a 250-ml beaker. Add 50 ml of 10 percent
HC1 to the solution. Mix well.
42.2 Discard the contents of the hydrogen
peroxide Implnger. Carefully transfer the con-
tents of the remaining four impingers to a
500-ml iodine flask,
i 4.2.3 Rinse the iour absorbing Impingera
and connecting glassware with three portions
;of ihc acidified lodl;ie solution. Use the en-
tire 100 ml of acidified Iodine fnr this pur-
'pose. Immediately after pouring :ha acidified
'lodins into an Impinger, stopper It , nci shake
for a tew moments before transferring the
rinse to the Iodine a.isk. Do not transfer any
rinse portion from one impinger to another;
transfer It directly to the iodine flask. Once
.acidified Iodine solution has been poured Into
:any (rlassware containing cadratum sulfids
sample, the container must be tightly stop-
pered at all times except when adding more
solution, and this must be done as quickly
and carefully as possible. After adding any
acidified Iodine solution to the Iodine flask,
allow a few minutes for absorption cf the II..S
into the iodine before adding any further
rinses.
4.2,4 Follow this rinse with two more rlnsea
using distilled water. Add the distilled water
rinses to the iodine flask. Stopper the flask
and shake veil. AHow about 30 minutes for
absorption of the H^S Into the Iodine, then
Complete the analysis titration.
Caution: Keep the Iodine flask stoppered
except when adding sample or Htrant.
4.2.5 Prepare a blank in an lodir.e fiasfc
using 45 ml of the absorbing solution, 50 nil
of 0.01 N Iodine solution, and 50 ml of 10
percent HC1. Stopper the flask, shake well
and analyze with the samples.
4-3 Analysis.
Nats: This analysis titration should be
conducted at the sampling location In order
to prevent Inss of iodine from the sample,
Titration should never be mads In direct
sunlight,
4,3.1 Titrate the solution In the flask wits
0,01 N sodium thiosulfate solution until the
solution is light yellow. Add 4 ml of the
st-srch Indicator solution and. continue
titrating until the blue color ]ust disappears.
4.3.2 Tl-'fti; the blanks In the same man-
ner aa the samplea.
5. Cak>-'alions.
5.1 Nonna'nty of the standard iodine solution.
equation 11-1
390
-------
where;
.V/ — normality of iodine, g-cq;'litcr.
F,= volume of iodine used, nil.
A'r~»orni:ility of .-odium thiosulfate, g-eq/liler.
Yf = volume of sodium thk:unlfute used, nil.
5.2 \onriatitif of the standard thwsuifate suiuiian,
VT e<'i>r).
5.3 7)r.v oas volume. Corrcrt the sample volume inc-^-nircd 1,'v Ihc dry gus n.> '.o: 10
Btaudard condilions [21°C(70"F)] and 760 mm <20.02 ini'itis) linfb^ u^ing c(|t.iiti»n 11-3.
V —
"- Ptl equation 11-3
where:
V'milj, = volume at standard conditions of gas s^'iinple through the dry gni meter,
sj.indrtrd liter- (?cf).
1"B= volitnir of gas sample through ihe dr-.- g.i-i muter fuictcr ccii(liiions\ liters
(cu. ft.).
T.,0 mg-'g)
~ (1,000 uil;l ){2Hj3 oci/uiole)
V; = vo!urnp «jf stniidaid iodine solution, ml.
A'; = !inrm;!.Iity of standard iodine solution, g-cq/liter.
I'"T= volume of standard sodium tliiosulfafe solution, ml.
A"r=iinrmiiljty ot iiondurd sodium ihiosulfate solution, g-eq,;liter,
Vmj|rJ = dry gas volume al stundurd conditions, liters.
where (English units):
-
_________
G. References.
6.1 petcrmination of Hydrogen Sulfide, Ammoniacal Ctdinium Cdloridc Method
API Method 772-54. In: Manual on Disposal of Refinery Wastes, Vol. V: Sampling
and Analysis of Waste Gasea and Paniculate Alatter, American Petroleum Institute
\Va=hmgvon, D.C., 1954. '
6.2 Tentative Method for Determination of Hydrogen Sulfide and Mcrcanran Sulfur
m Natural Gas, Naturnl Gas Processors Association, Tulsa Oklahoma KG PA Publi-
cation No. 2285-65, 1905.
391
-------
METHOD 13A
Determination Of Total Fluoride Emissions Froni
Stationary Sources-SPADNS Zirconium Lake Method
1. Principle and Applicability.
1,1 Principle. Gaseous and particular
fluorides are withdrawn isoklnetlcally from
the source usinj a sampling train. The-fluo-
rides are, collected In the impinger water and
on the filter of the sampling train. The
weight of total fluorides In the train Is de-
termined by the SPADNS Zirconium Lake
colormietrlc method,
1.2 Applicabiliti/. This method Ls applica-
ble for the determination of fluoride emis-
sions from stationary sources only when
specified by the test procedures for deter-
mining compliance with new source per-
formance standards. Fluorocarbons, such as
Freons, are not quantitatively collected or
measured by this procedure.
2. Range and Sensitivity.
The SPADNS Zirconium Lake analytical
method covers the r.inge from 0-1.4 tig/ml
fluoride. Sensitivity has not been determined.
3. Interferences.
During the laboratory analysis, aJumlnum
In exr-ess of 300 mg/liter and silicon dioxide
Sr. excess of 300 ^g/Hter will prevent com-
plete recovery of fluoride. Chloride will distil!
over and Interfere with the SPADNS Zirconi-
um Lake color reaction. 1' chloride Ion is
present, use cf Specific. Ion Electrode (Method
13B) Is recommended; otherwise a chloride
determination Is required and 5 mg of silver
sulfate (see section 7.3,8) must be added for
each mg of chloride to prevent chloride In-
terference. If sulfurlc acid Is carried over In
the distillation, It will cause a positive inter-
ference. To avoid sulfurlc acid carryover, It
is Important to stop distillation at 175°C.
4, Precision, Accuracy and Stability.
4.1 Analysis. A relative standard devia-
tion of 3 percent was obtained from twenty
replicate intralabcratory determinations on
stack emission samples with a. concentration
range of 39 to 360 mg/}. A phosphate rock
standard which was analyzed by this pro-
cedure contained a certified value of 3,84
percent. The average of five determinations
was 3.88 percent fluoride.
•4,2 Stability. The color obtained when
the sample and colorlmctric reagent are
mixed Is stable for approximately two hours.
After formation of the color, the absorbances
of the sample and standard solutions should
be measured at the same temperature. A 3°C
temperature difference between sample and
standard solutions will produce an error of
approximately 0.005 mg F/litcr.
6. Apparatus.
5,1 Sample train. See Figure 13A-1; It is
similar to the Method 5 train except for the
Interehangeablllty of the position of the fil-
ter. Commercial models of this tram are
available. However, If one desires to build his
own, complete construction details are de-
scribed in APTD-05ai; for changes from the
APTD-OSfll document &nd for allowable
modifications to Figure 13A-1, see the fol-
lowing subsections.
The operating and maintenance procedures
,for the sampling train are described In
APTD-0576. since correct usage Is Important
in obtaining valid results, all users should
read the APTU-C578 document and adopt
the operating and maintenance procedures
outlined in It, unless otherwise specified
herein,
5.1,1 Probe nozzle—Stainless steel (316)
with sharp, tapered leading edga. The angle
of taper shall be £30' and the taper shall
be on the outside to preserve a constant
Internal diameter. The probe nozzle shall be
of the button-hcolc or elbow design, unless
otherwise specified by the Administrator. The
wall thickness of the nozzle shall be less than
or equal to that of 20 gauge tubing, i.e.,
0.165 cm (C.065 In ) and the distance from
the tip of the nozzle to the first bend or
point of disturbance shall be at least two
times the outside no?zle diameter. The nozzle
shall be constructed from seamless stainless
steel tubing. Other configurations and con-
Btrjction material may be used with approval
from the Administrator.
A range or sl/es suitable for isokinetic
sampling should be available, e.g., 032 cm
Us in.) up to 1.27 cm (% In.) (or larger If
higher volume sampling trains are used) In-
side diameter (ID) nozzles In Increments of
0.16 cm (i/k In.). Each nozzle shall be cali-
brated according to the procedures outlined
In the calibration section.
fi.1,2 Probe liner—Boroslllcate glass or
stainless steel (316). When the niter IB lo-
cated immediately after the probe, a probe
nesting system may be used to prevent filter
plugging resulting from moisture condensa-
tion. The temperature in the probe shall not
exceed 120 i 14'C (248 ± 25°P),
5.1.3 PItot tube—Type S, or other device
approved by the Administrator, attached to
probe to allow constant monitoring of the
stack gas velocity. The face openings of the
pltot tube and the probe nozzle shall be
adjacent and parallel to-each other, not
necessarily on the same plane, during sam-
pling. The free space between the nozzle and
pltot tube shall be at least 1,8 cm (0.75 In.),
The free space shall be set based on a 1.3 cm
(05 Jr..) ID nozzle, which Is tlie largest v.ze
nozzle used.
The pltot lube nrjst also meet the criteria
specified in Method 2 and be calibrated nc-
cording to the procedure In the calibration
section of that method.
5,1.4 Differential pressure gauge—In-
clined manometer c.ipadle of measuring ve-
locity head to within lO'/t of the minimum
measured value. Below a differential pressure
of 1.3 mm (O.Oo in.) water gauge, micro-
manometers with sensitivities of 0.013 mm
(0.0005 in.) should be used. However, mlcro-
manometers are not easily adaptable to field
conditions and are net easy to use with pul-
sating flow. Thus, other methods or devices
acceptable to the Administrator may be
used when conditions warrant,
S.I .5 Filter holder—Borosilicate glass with
a glMs frit filter support and a slllcone rub-
ber gasket. Other materials of construction
may be used with approval from the Ad-
ministrator, e.g.. If probe liner is stainless
steel, then filter holder may be stainless steel.
The holder design shall provide a positive
seal against leakage from the outside or
around the filter.
392
-------
TEMPERATURE
SENSOR ,
—r-"^. I ^ PROBE
PnOTTUBE
j( STACK WALL | "LO'CATtON
FHOBE L'
«4=
/
REVERSE-TYPE
PITDTTUBE
"^ifi
ORIf ICE MANOMETER |J
^
PITQTMANOHETSR
3A 1, F L,c.fidc
CONWECTIKGTUHE
12'mm ID
T24 CO 'V
THERMOMETER TIP MUST EXTEND BELOW
THE LIQUID1 LEVEL
WITH 110/30
124/40
HEATING
MANTLE
CONDENSER
250ml
VOLUMETRIC
FLASK
Figure 13A-2, Fluoride Distillation Apparatus
393
-------
5.1.6 Filler heating system—When mcls-
ture condensation Is a problem, any Jiaatlng
system capable of maintaining a temperature
around the filter bolder during sampling of
no greater than JZO-14'C (248i:25"F).
A temperature gauge capable of measuring
temperature to within 3'C (5.4"PJ "shall 1)e
Installed so that \vhen the filter heater Is
used, the temperature around the filter
holder can be regulated and monitored dur-
ing sampling. Heating Ej-stems other llian
the one shown in AFTD-05S1 may be used.
5.1.7 Implngers—Four implnger* con-
nected as shown in Figure 13A-1 with ground
glass (or equivalent), vacuum tight fittings.
The first, third, nnd fourth Implngers are
of the Greenrjurg-Smltli design, modified by
replacing the tip with a i;i cm ('4 in.)
Inside diameter glass tube extending to 1'4
cm ('a in.) from the bottom of the flask.
The second Imptnger Is of the Greenburg-
Smlth. design with the standard tip,
5.1,8 Metering system—Vacuum gauge,
leak-free pump, thermometers capable of
measuring temperature to within 3»C
(—5=F), dry gas metsr witli 2", accuracy at
the required sampling rate, and related
equipment, or equivalent, as required to
maintain an IsoKinetlc sampling rate and
to determine sample volume. When the
metering system Is used In conjunction with
a, pilot tube, tbe system shall enable cteoKs
of Isoklnetic tales.
5.19 Barometer—Mercary, aneroid, or
other barometers capable of measuring at-
mospheric pressure to within 2.5 mm Hg
(0.1 In. Hg}. In many cases, the barometric
reading may be obtained from a nearby
weather bureau station, In which case the
station value shall be requested and an ad-
justment for elevation differences shall be
applied at a rate ol minus 2.5 mm Hg (0.1
In. Hg) per 30 m (100 ft) elevation increase.
5-i f ample recovery.
5.2.1 rrcbe liner and probe nozzle
brushes—Nylon bristles with stainless steel
wire handles. Tlie probe brusli shall have
extensions, at least as long as the probe, of
•stainless steel, teflon, or similarly Inert mate-
rial. Boil) brushes shall be properly sized and
shaped to brush out the probe liner and'
nozzlo,
5,2.2 Gloss wasa bottles—Two.
5.2.3 S.imple storage containers—Wldo
moutb, high density polyethylene bottles,
1 liter.
5 2,4 Plastic storage containers—Air tight
containers of sufficient volume to store silica
6*1.
5.25 Graduated cylinder—250 ml.
5.2.6 Funnel and rubber policeman—to
aid in transfer of silica gel to container; not
necessary if silica gel is weighed in the field.
5.3 Analysis.
5.3,1 Distillation, apparatus—Glass distil-
lation apparatus assembled as shown In Flg-
UTS 13A-2,
5.3 2 Hot plate—Capable of heating to
500' C.
5.3.3 Electric muffle furnace—Capable of
heating to 600° C.
5.3,4 Crucibles—Nickel, 75 to 100 ml ca-
pacity.
5,3.5 Beaker, 1500 ml,
6.3-6 Volumetric flask—50 ml.
5.3.7 Erlenmeyer flask or plastic bottle—
500 ml,
5.3.8 Constant temperature bath—Capa-
ble of maintaining a constant temperature of
il.O' C lu the range of room temperature.
6,3,9 Balance—300 g capacity to measure
to ±0.5 g.
5.3.10 Spectrophotometer — Instrument
capable of measuring absorbance at 570 nm
and providing at least a 1 cm light path.
5.3,11 Spectrophotometer cells—I cm.
6, Reags-nts
6.1 Sampling.
6.1.1 Filters—Whatman No. 1 filters or
equivalent, sized to fit filter holder.
6.1.2 Silica gel—Indicating type, 6-13
mesh, If previously used, dry at 1755 C
(350= F) lor 2 hours. New si)lea gel may be
used as received.
• 6.1.3 Water—Distilled,
S.I,4 Crushed Ice.
6,1.5 Stopcock greaae—Acetone Insoluble,
heat stable silicone grease. This Is not neces-
sary if screw-on connectors with teflon
sleeves, or similar, are used.
6,2 Sample recovery.
6,2.1 Water—Distilled frcra snme con-
iBtner as 6,1,3.
6,3 Analysis.
6.3,1 Calcium oxide (CaO)—-Certified
e--ade containing 0,005 percent fluoride or
less.
5,3.2 Phenolphthflleln Indicator—0.1 per-
cent In 1:1 ethanol-water mixture.
8,3,3 Silver sulrate (Ag,.SO,)— ACS re-
agent grade, or equivalent.
6,3.4 Sodium hydroxide (NaOH)— Pellets,
ACS reagent grade, or enulvalent.
6.3.5 Sulfuric ti;;fi SHLSO,)—Concen-
trated, ACS rengctiL grade, or equivalent.
S3.6 Filters—Wiiatman No. 5-J1, or equiv-
alent,
6.S.7 HydrodiK-rJC field (HC11 —Coi.cec-
traiptl, ACS reaji! :u €'r;idel or equivalent,
6,3.8 Water—D^-uilcd, Jrcsn samb con-
tainer as 6,1.3,
630 Ecrlir.m fluoride—Standard Folutlon.
Disnclre 0.22:0 g oi scdlum fluoride lu 1
liter of distilled water. Dilute 100 ml of this
solution to 1 liter with distilled weter. One
milli'iiwr of the solu'ion co-mains 0.01 mg
of r.uorlde.
3.3.10 SFADN'5 E.->:utlori--[4.5d;hydroxy-
3-,p-sulfopl!enrlak:-))-2,7-naphthalcne - cll-
suli'on;c acid trisodium sal:]. Dissolve 0.960
-.(lit) g of SPADNS reagent In 500 ml dis-
unetJ water. This solution Is siabl; for at
;e?,st one month. i[ stored in a well-sealed
bot'Je pr.V.ecte-d from sunl^ht.
B.3.11 ReCerener solution—Afld 10 n;! of
SPADES solution ,6fi,:0i to 100 ml c!5c;lled
water Enid nclc;:y v ;ih a ?oluiion prepared by
diluting 7 n^l of concerlrated HC1 tc 10 n\i
'.van dislillcu MTiter. Th's sclurion Is used to
set ilie speetrophotDmeter zero poir.t, and
should be prepared dsiiy.
C.3.12 5PADXS Mined Reagent—Dissolve
0135 ^0.005 g of zirconyl chloride cctahy-
clrati fZrOCl5.8H ()>. m 25 ml dis'jllcd water.
Add 350 nil of concentrated HC1 and dilute to
500 ml wllh distilled water. Mix equal vol-
umes of this solir.iGr, and 51-ADN'S solution
to form a single reagent. T'nii reaaeiu :s
stable for at least two months,
7. Procedure.
NOTE; The- fusion and dist'.llatSr.n steps of
this procedure will liot br required, if It fan
be shov.-n to the satisfaction of the Adminis-
trator ;hat the samples contain only water-
soluble fluorides,
7.1 Sampling. The campling fhall be con-
ducted 1)7 competr-n: personnel experienced
with tbb test procedure.
7.1.1 Pretest preparation. All train com-
ponents shall be maintained and calibrated
according to the procedure described In
APTD-0576, unless otherwise specified herein.
WeiEh approximately 200-300 g or silica gel
in air ticbi containers to the neare.it 0.5 g.
Record the total weight, both silica gel and
container, on the container, More silica g-fll
aiay be used but care should be taken during
sampling that It Is not entrained and carried
ont from the Implnger. As an alternative, the
silica gel may be weigl-.ed directly In the 1m-
plngor or Its sampling holder just prior to
the train assembly.
394
-------
' 7.1.2 Preliminary determinations, select
the sampling site and the minimum number
of sp.mplfng point? according to Method I or
as specified by the Administrator. Determine
the stack pressure, temperature, and the
range of velocity heads using Method 2 and
moisture content using Approximation Meth-
od 4 or Its alternatives for the purpose of
miking isokinetic sampling rate calculations,
Estimates may be used. However, final results
will be based on actual measurements made
during the test.
Selec:. a nozzle size based on the range of
velocity heads s'jch that it is not necessary
to change the nozzle size in order to ir.air.-
talu Jsoklnetic sampling rntes, During the
run, do no; change the nozzle size, Lnsurc
that the differential pressure gauge is capable
of measuring the minimum velocity head
value- to within 10^, or as spccified'by the
Administrator.
Select, a suitable probe liner and probe
length such that all traverse points call he
sampled. Consider sampling 'rom opposite
sides lor large stacks to reduce the length of
probes,
Select a total samplin? liin: greater e laUfr is based on an ap-
proximate- average sampling rate. Note also
that the minimum total sample volume Is
corrected to standard conditions.
It is recommended that a hall-integral or
Integral number of minutes be sampled at
each point in order to avoid timekeeping
errors,
In some circumstances, e.g. batch cycles, it
may be necessary to sample for shorter times
it the traverse points and to obtain smaller
»as sample volumes. In these cases, the Ad-
ministrator's approval must first be obtained.
7.1.3 Preparation of collection train. Dur-
ing preparation and assembly of the sp.rr.-
pling train, keep all openings where contami-
nation can occur covered until just prior to
assembly or until sampling is about to begin,
Place 100 ml of water in each of the firs;
two implngers, leave the third Implr.ger
empty, and place approximately 200-300 g
or more, IT necessary, of prewcighed silica,
gel in the fourth implnger. Record the weight
of the silica gel and container on the data
Bheet. Place the empty container in a clean
place for later u^e in the sample recovery.
Place a filter in the Slter holder. Bs sure
that the filter Is properly centered and the
gasket properly placed so as to not allow the
sample Has stream to circumvent the fllrer,
Cheek filter for tears alter assembly Is com-
pleted.
When glass liners are used, install selected
nt>27,le using a Viton A O-ring; the V:ton A
O-ring is installed as a sea] where the nozzle
Is connected to a glass liner. See APTD-0576
for details. When metal liners are used. In-
stall the nozzle as above or by a leek free
direct mechanical connection. Mark the
probe with heat resistant tape or bv some
other method to denote the proper distance
into the stack or duct lor each sampling
point,
Unless otherwise specified by Use Admin-
istrator, attach a. temperature probe to the
metal sheath of the sampling probe so that
the censor extends beyond the probe tip and
does not touch any metal, It.s position should
be about 1.9 to 2.54 cm (0.75 to 1 111.) from
the pitot tube and probe nozzle to avoid
interference with the gas flpw._
tlic train sa iiio.m in Figure
13A-J with the niter between the third _and
;ourih. impiit^er.i. Alternatively, tisc niter
may be placed between the prjbe a.nd the
lirst Uup'inger. A niter heating system may
bo used to prevent rnni?"-Ui'c condensation,
but the temperature around the Iiller "aplcler
5,:vnll not exceed 120=14 C (24S-i;i!5'F).
((Now: Whatman Mo. 1 niter decomposes at
150'C (300'Pij.J Record niter location on
the da-tu sheet.
Place crushed ice around, tin? imnlngers,
7.U Leak check procedure—Alter the
sampling train has been asicr.iblecl, turn o.i
and sec {it applicable) she probe and filter
heating systemts) to reach a temperature
sufficient to avoid condensation in the probe.
Allow time for tiic temperature to stabilize:.
Leak check the train at the suppling site !;v
plugging the nozzle and pulling a 33U mm Hg
(15 i"). He) vnuuuin. A leakage rate in ex-
cr.i.5 of V e. 01" the average sampling rats or
0.00057 m'.. mln. (0.02 ci'mj, whichever Is, less,
IS unacceptable,
Tnc follow] 115 ienk check instructions fur
the str.ipllng lr:rn rtc«:ribej In APTD-0.373
aiid ATTD-OSdl i:i.iy be iitipful Start the
pump with by-pu^s vulve fully open and
coarse adjust valve completely closed. Par-
tially open the coir.-o adjust valve and slowly
close the by-pass valve until 330 mm Hg '15
in. HKS vac'iram is reached. Do not reverse
direction of by-p.vs valve. This will cause
water to bac~k up into the filter holder. If
330 mm H£ 115 in. H^l is exceeded, cither
leak check at lliis hi§!ier vucuu:n or end the
leak check as described below and start over.
When the leak check Is completed, first
slO'.vly remove the plug from the inlet to the
probe or filter holder and immediately turn
off the vacuum pump. Tills prevents the
water in the Implngers from beiiig forced
backward into '.he filter holder (if placed
before the lrnplr..;er=.) aud silica gel from
bting entrained backward into the third
impinger.
Leak checks s'nr.!l be eoii£i:etc3 as described
whenever ti:e tram is cllreisgciLjcd, 6-6- *or
silica t'el or niter chance? d-jring the test,
prior to each test run, and at the completion
cf each test run. If leaks are found to be In
excess of tlie acceptable rate, the test will be
corislclerccl invalid. To reduce lost time dve
to leakage occurrences, it is recommended
that leak checks be conducted between port
changes.
7.1.5 Par;iculate train operation—During
the sampling run, an isokinctlc sampling rate
within 10 <;'t, or as specified by the Adminis-
trator, of true isokinetic shall be maintained.
For e.ic.'i run, record the data required 0:1
the e:.Li;r.pli; data sheet E:IO-,MI i:i Figure 13A-
3. EC .Hire *.a rei.oui t;ie linti.-l dry gp.s r.ieter
reaiiriL; Kfccura ire dry {^r.s m;.t£r i-cuJiu j.< at
tlie boi-.uiiini; aad eud of each ^kinpUr.t; time
,nereme;,t, v.'nea cl-au^c-5 n: how races are
m;.c!e, and whea sainpime := nftltecl. Take
ulner uata po;i;t reti;jings L't lc.;v,t once a"
e3.ch s'.'.rriplo point during each lime incie-
metit and adclirional readings when 51^:111!-
cai.t changes ('2(1',,. variition in velocity head
readings) neaesslcate udtilcional adjustrne:iti
111 fio-,\ I'iLe. Be sure to level r.nd zero the
m.iuc,n:cti:r.
Clear: the portholes prior to the test run to
minimize chn.iice of sampling deK'Ositttl
m.-.tcrial. To begin sampling, remove the
iio7Ele cap, verify (if applicable) that the
probe he.iT.ter Is working and. filter heater is
up to temperature, and that the pKot tube
and probe are properly positioned. Position
tho noi-,;,le at the first traverse point with the
tip pointing directly into the gas strep.m. Im-
mediately start the pump and adjust the
ilow to iioklnetic conditions. Nomogrtiplis are
available for sampling trains using type S
pltot tubes wltli 0.85^0.02 coefflclents (Cp),
395
-------
and u'.'ien samp!;i:i> in ,-dr or a stack pas with
couH'alent density (molecular weight M-i,
equal to 29:-_-4>. which aid In the rapid ad-
justment of the Isoklnctic sampli.-ii* ra'e
without fxrc'sive roaipntrvtloni. APU'b-0578
details the procedure for us;r.g these nomo-
graphs. If Cs and Mi ar? outside the above
stated ranges, do net use '.he nomof>-rs.p>i
un'.e«s appj-opirate steps are taken to com-
pensate for the deviations.
When the stack Is vndcr significant r. <• :•:>.-
tive pressure .'height or impinger stem>, :ake
c?re to close the rn-"-«c udjusi vf-.lve before
inserting the p^obe Into the r.ack to a^.'iri
v,-a*er backing Into the- filler I'ol'icr. If nst:?^-
sary, the pump may ha turned •;:; deposited matcr;al.
During i-'o fsi mil. make periodic nci;;_iL;t-
mc:n:s to kec-p the probt- nr.d :if ippliutblr1 >
filter temperatures at their proper values. Atct
more ice and, If necessary, salt to the lea
otxih, tc* maintain a temperature of loss lha'i
20°C (68"F1 at the imnlnger/slllcft gel cmHet.
to avoid excessive moisture If^MS. AlFO, pa-
riodicallv check the level and zero o[ the
manometer.
If the pressure drop across the filter bo-
comes high enough to make i&okinetlc sarr-
piing difficult 10 maintain, the filter may be
replaced In the middl of a sample run. It is
recommended that another complete filter
assembly be used rather than attempting to
change the fllter Itself. After the new niter or
filter assembly Is Installed conduct a leak
:heck. The final emission results shall ba
based on the summation of all fllter catches.
A single train shall be used for the entire
sample run, except for filter and silica gel
changes. However, If approved by the Admin-
istrator, two or more trains may be used for
a single test run when there are two or mote
ducts or sampling ports. The flnal emission
results shall be based an the total of all
sampling train catches.
At the end of the sample run, turn off the
pump, remove the probe and nozzle Irom
the stack, and record the final dry gas meter
reading. Perform a leak check.' Calculate
percent l=oklnetlc (see calculation section)
to determine whether another test run
should be made. If there Is difficulty in main-
taining isaklnetlc rates due to source con-
ditions, consult with the Administrator for
possible variance on the Isoklnetlc rates.
7.2 Sample recovery. Proper cleanup pro-
cedure begins as soon as the probe Is re-
moved trcm the stack at the end o£ the
sampling period.
When the probe can. be safely handled,
wipe off all external partlculate matter near
the tip of the probe nozzle and place a cap
over It to keep from losing part of the
sample. Do not cap off the probe tip tightly
while the sampling train IB cooling down, as
this would create a vacuum In the niter
holder, thus drawing water Irom the 1m-
pingers Into trie fllter.
Before moving the sample train to the
cleanup site, remove the probe from the
sample tialn, wipe o3 the sillcone grease, and
cap the open outlet of the probe. Be careful
not to lose any conSensate, If present. Wipe
off the silicons grease from the fllter Inlet
where the probe was fastened and cap It.
1 With acceptability of the test run to be
based on the same criterion as In 7.1.4,
Remove tile umbilical cord from the last
Impingcr and cap the Irnplnger. After wip-
ing off the slllconc grease, cap off the filter
hojder outlet and Irnplnger inlet. Ground
glass stoppers, pln?tlc caps, or serum caps
may be used to close these openings.
Transfer the probe and fllter-Implnger as-
sembly to the cleanup area. This area should
be clean and protected from the wind so that
the chancrs of contaminating or losing the
sample will be minimized.
Inspect the train prior to and during dis-
assembly and note any abnormal conditions.
Using a graduated cylinder, measure and re-
cord the volume of tho water In the first
three Implrigers, to the nearest ml; any con-
ciensate in the probe should be Included In
Ihls determination. Treat the samples ag
follows:
7,2,1 Container No. 1. Transfer the 1m-
pinger water from the graduated cylinder to
this container. Add the fllter to this con-
tainer. Wash all sample exposed surfaces,
Including the probe tip, probe, first three
implngers, impinger connectors, filter holder,
and graduated cylinder thoroughly with dis-
tilled water. Wash each component three
separate times with water and clean the
probe and nozzle with brushes. A maximum
wash of SCO ml is used, and the washings are
added to the sample container which must
be made of polyethylene,
12,2 Container No. 2. Transfer the silica
gel from the fourth Impinger to this con-
tainer and seal.
7.3 Analysis. Treat the contents of each
sample container as described below.
7,3.1 Container No. 1.
7.3.1.1 Filter this container's contents. In-
cluding the Whatman No. 1 fllter, through
Whatman No. 541 fllter paper, or equivalent
into a 1500 ml beaker. Note: If filtrate volume
exceeds 900 ml make nitrate basic \vl"n
NaOH to phenolphthaleln and evaporate to
less than 900 ml,
7.3.1.2 Place the Whatman No. 541 fllter
containing the Insoluble matter (including
the Whatman No. 1 filter) In a nickel cruci-
ble, add a tew ml of water and macerate the
filter with a glass rod.
Add 100 ms* CaO to the crucible and mix
the contents thoroughly to form a slurry.
Add 3. couple of drops of phenolphthaleln
indicator. The Indicator will turn red in a
basic medium. The slurry should remain
basic during the evaporation of the water
or fluoride ion will be lost. If the Indicator
turns colorless during the evaporation, an
acidic condition Is Indicated. If this happens
add CaO until the color turns red again.
Place the crucible In a hood under Infra-
red lamps or on a hot plats at low heat. Evap-
orate the water completely.
After evaporation of the water, place the
crucible on a hot plate under a hood and
slowly Increase the temperature until the
paper chars. It may take several hcairs for
complete charring of the. fllter to occur.
Place the crucible In a cold mufflo furnace
and gradually (to prevent smolclng) Increase
the temperature to SOO'C, and maintain un-
til the contents are reduced to an ash. Re-
move the crucible from the furnace and allow
It to cool.
T.3.1.3 Add approximately 4 g of crushed
NaOH to the crucible and mix. Return the
crucible to the muffle furnace, and fuse the
sample for 10 minutes at 600°C.
Remove the sample from the furnace and
cool to ambient temperature. Using several
rinsings of warm distilled water transfer the
contents of the crucible to the beaker con-
taining the filtrate from container No. l
(7.3.1), To assure complete sample removal,
396
-------
rinse dually with two 20 ml portions of 25
percent (v/v) sullurlc acid and carefully add
to the beaker. Mix well and. transfer to one-
liter volumetric flask. Dilute to volume with
distilled water and mix thoroughly. Allow
any undlssolved solids to settle.
7.3.2 Container No. 2. Weigh the spent
silica gel and report to the nearest 0.5 g.
7.3,3 Adjustment of sold/water ratio In
distillation flask—(Utilize a protective shield
when carrying out this procedure.) Place 400
m] of distilled water In the distilling flask
and add 200 ml ol concentrated HJ5O,. Cau-
tion: Observe standard precautions when
mixing the HjSO, by slowly adding the acid
to the flask with constant swirling. Add some
soft glass beads and several small pieces of
broken glass tubing and assemble the ap-
paratus as shown in Figure 13A-2. Heat the
flask until It reaches a temperature of 175 "C
to adjust the acid/water ratio for subsequent
distillations. Discard the distillate.
7.3.4 Distillation—Cool the contents of
the distillation flask to below 80 C. Pipette
an aliquot of sample containing less than 0.6
mg P directly Into the distilling flask and add
distilled water to make a total volume of 220
ml added to the distilling flask. [For an es-
timate of what size aliquot does not exceed
0.6 mg F, select an aliquot of the solution
and treat as described In Section 7.3.6. This
will give an approximation of the fluoride
content, but only an approximation since
interfering Ions have not been removed by
the distillation step.]
Place a 250 ml volumetric flask at the con-
denser exit. Now cegln distillation and grad-
ually Increase the heat and collect all the
aistlllatlon up to 175=C, Caution: Heating
the solution above 17S°C win cause sxilfurle
acid to dlatlU over.
The- acid In the distilling flask can be used
until there Is carryover of Interferences or
poor fluoride recovery. An occasional check of
fluoride recovery with standard solutions Is
advised. The acid should be changed when-
ever there Is less than 90 percent recovery
or blank values are higher than 0.1 ug/ml.
Note: If the sample contains chloride, add
5 mg Ag.JSO, to the flask for every mg of
chloride. "Gradually Increase the heat and
collect all the distillate up to 175° C. Do not
exceed 176°C.
7.3.5 Determination of Concentration—
Bring the distillate In the 250 ml volumetric
flask to the mark with distilled water and
mix thoroughly. Pipette a suitable aliquot
from the distillate (containing 10 fjg to 40
ng fluoride) and dilute to 50 ml with dis-
tilled water. Add 10 ml of SPADNS Mixed Rea-
gent (see Section 6.3.12) and mix thoroughly.
After mixing, place the sample In a con-
stant temperature bath containing the stand-
ard solution for thirty minutes before read-
ing the absorbance with the spectropho-
tometer.
- Set the spectrophotometer to zero absorb-
ance at 670 urn with reference solution
(6,3.11), and check the spectrophotometer
calibration with the standard solution. De-
termine the absorbance of the samples and
determine the concentration from the call-.'
bration curve. If the concentration does not
fall within the range of the calibration curve,
repeat the procedure using a different size
aliquot.
8, Calibration,
Maintain a laboratory log of all calibrations,
8.1 Sampling Train.
8.1.1 Probe nozzle—Using a micrometer,
measure the Inside diameter of the nozzle
to the nearest 0,025 mm (0,001 In.). Make
3 separate measurements using different
diameters each time and obtain the average
of the measurements. The difference between
the high and low mimaers shall not exceed
O.lmm (0.004111.).
When nozzles become nicked, dented, or
corroded, they shall be reshaped, sharpened,
and recalibrated before use.
Each nozzle shall be permanently and
uniquely Identified,
8.1.2 Pltot tube—The pitot tube shall be
calibrated according to the procedure out-
lined in Method 2.
8.1.3 Dry gas meter and orifice meter.
Both meters shall be calibrated according to
the procedure outlined in AFTD-0576. When
diaphragm pumps with by-pass valves are
used, check for proper metering system de-
sign by calibrating the dry gas meter at an
additional flow rate of 0.0057 mVmin. (0.2
cfm) with the by-pass valve fully opened
and then with it fully closed. If there la mere
than ±2 percent difference In flow rates
when compared to the fully closed position
of the by-pass valve, the system Is not de-
signed properly and must be corrected.
8.1.4 Probe heater calibration—The probe
heating system shall be calibrated according
to the procedure contained In APTD-0576,
Probes constructed according to APTD-0581
need not he calibrated If the calibration
curves in AFTD-0576 are used.
8.1.5 Temperature gauges—Calibrate dial
and liquid filled bulb thermometers against
mercury-ln-glaas thermometers. Thermo-
couples need not be calibrated. For other
devices, checS with the Administrator.
8.2 Analytical Apparatus. Spectrophotom-
eter. Prepare the blank standard by adding
10 ml of SPADNS mixed reagent to 50 ml of
distilled water. Accurately prepare a series
of standards Irom the standard fluoride solu-
tion (see Section 6.3.9) by diluting 2, 4, 6,
8, 10, 12, and 14 ml volumes to 100 ml with
distilled water. Pipette 50 ml from each solu-
tion and transfer to a 100 ml beaker. Then
add 10 ml of SPADNS mixed reagent to each.
These standards will contain 0, 10, 20, 30,
4"0', 50, BO, and 70 Ag of fluoride (0—1.4 ^g/ml)
respectively.
After mixing, place the reference standards
and reference solution In a constant tem-
perature bath for thirty minutes before read-
ing the absorbance with the spccuoptiotom-
eter. All samples should be adjusted to this
same temperature before analyzing. Since
a 3°C temperature difference betweei: samples
and standards wt!l produce an error of ap-
pro?;imately 0,005 rng I'/liter, care murt be
taken to sea that samples and standard? are
at nearly Identical temperatures when ab-
fcorbance's are recorded.
With the spectrophotometer at 570 r.m,
use the reference solution (see section G.3.11)
to set the absorbance to sero-
Determine the absorbance of the stand-
ards. Prepare a calibration curve by plotting
its, F/50 ml versus absorbance on linear graph
paper. A standard curve should be prepared
initially and thereafter whenever the
SPADNS m'.-xed reagent is newly made. A'so,
a calibration standard should be run with
each set of samples and If it differs from the
calibration curve by ±2 percent, o new
standard curve should be prepared,
9. Cdlculitinns.
Carry out calculations, retaining at least
one extra decimal figure beyond that of the
acquired data. Round off figures after final
calculation,
9.1 Nomenclature.
At.— Aliquot of distillate taken for color
development, ml.
jin=Cross sectional area of nozzle, m1 (ft5).
jli=Allquot of total sample added to still,
ml.
B,,,=Water vaoor In the gas stream, propor-
tion by volume.
397
-------
'Ci— Concentration of fluoride In stack gas,
mg-'m', corrected to standard conditions
of 20' C, V6D mm Hg (68° F, 29.92 In. Hg)
on dry basis.
Fi = Total weight ol fluoride 1n sample, mp,
pgF~ Concentration from the calibration
curve, pg.
7=Fercent of isokinetlc sampling.
mn = Total amount of paniculate matter
collected, me.
jWic = Molecular weight ol water, 18 g'g-mole
(IB ib/lb-mo!e).
171. = Mass of resldxie of acetone after evap-
oration, mg,
Pi^, — Barometric pressure at trie sampling
site, mm Hg (In. Hg).
P, = Absolute stack gss pressure, mm Hg (In.
Hg).
P.irf = Standai-d absolute pressure, 760 mm
Hg (20.02 In. Hi;).
.R = Ideal gas coi^tan", C.ODJ36 mm H.«-:ii;-
°K-g-;r,o!e (21.83 in. Hg-it'/"R-lb-niolei.
IV, = Absolute Everage dry gaj merer tem-
perature (see rig. 13A-3), "K. (CK).
T< = Ab30;UVe average £ta;k gas lempeiaturc
(see fig. ISA-Si, °K ('Hi.
T,i4 = Sta:!dnrd absolute temperature, 293'
K (528' Rt.
Vc— Volume or acetone blank, ml.
V*-i =VoU'jr.e of accvor.e used in was!-,, ml.
Vd=: Volume of distillate collected, m'-
7n=Total volume of liquid collected in ini-
plngers and silica gel, ml, Volume of -.'.-arc:
in silica gel eeu.il? ?i:ic.i gel v/esght in-
crease in grams times 1 ml/gram. Volume
of liquid coliecc^d l:i napinger etiuals rn;.\i
volume nnnus sruriai volume.
V1!.^ Volume of :^*c sr.mple as me.'isu;Jd 'ay
dry gas meier, dcm !dcf).
Vr,•»!:• — Volume c-r g-s sample measu.vcl Dy
the dry gas me'.er corrected 10 3l?.!idarcl
canditlo:is, dscm id-ifi).
V. .1,1 .,. = Volume of watsr vapor in t'r.e eas
sample curreccecl to standard cor.thriar!?,
scrn (&ct'i.
V'r—Total volume of sample, ml.
v.=atack g.is velocity, calculated by Methyl
2, Equation 2-7 u.-i'ng dat.i obttlijed ;":T>;II
Melh'jd 5, m rec (ft.'i-ec).
W*,, = V»'eight of rcfid\jc i:. acetone wa-h. nig.
^H = Average p: no-urc- ci.iTcrer.rial acr.^s-s t:ic
orifice (see fig. 13A-3). merer, nun H.O
(In. H 0).
pa~Deriri.ty of acetone, mg.'ml (reo IB'DC'. r-ii
botvle 1.
p,, •• Density cl ws.tcr, 1 g, n-.l (O.OL'2^0 !fc"'
ml).
e = Total sampV.r? tline. m:n.
]3.6 = Spcciric gvavi:y or mercury.
60 = Sec/m:n.
100 = Conversion to percent.
fi 2 Averse dry g:^"i meter tempeiaturc
and average orifice presv.;re drop. Sec data
sheet (fl£. 13A-3K
9.3 Dry gas volume. Correct the san-.ple
volume measured by the dry gas meter to
standard conditions '20'~ C, 760 mm H.7 168°
F, 29.92 Inches Hg!] by using equ&ttcn
13A-1.
T'
,
l',.,J
where:
K = 0.3855 »K/mm Hg for metric unlta.
= 17.65 'R/ln. Hg for English units.
9.4 Volume of wa,ter vapor.
where:
X = O.OD134 mVml for metric units.
= 0.0472 ftVml for English units.
9.5 Moisture content.
-, 7- - p ~-
.' / tt i fid
equation l.l A-l
equation l^A-3
If the liquid droplets are present tn the
gas sLi-eain assume the stream to be saturated
and use a ppychrometrlc chsrt to obtain an
approximation of the moisture percentage.
9,6 Concentration.
9 0.1 Calculate the amount ol fluoride In
the sample according to Equation 13A— 1.
equation 13A-4
where :
3.6.2 Cor.ceniration of fluoride in stack
£fis. Determine ihe concentration of fluoride
In the stack g&3 according to Equation 13A-S.
C.=
equation 13A-5
where:
ftv-'m».
398
-------
9.7 Isokinetlc variation.
9.7.1 Calculations from raw data.
100 T, lKV:r-L(Vm/T^ (Pt
equation 13A-6
it = 0.00346 mm Hg-rnVmI-°K for metric
unltg.
= 0,00267 la. Hg-ftVml.=R for English
units.
0.7.2 Calculations from Intermediate v&l-
uei-
__
(\-Bat)
r, I-
A* = 1,323 for metric units.
= 0.0044 for English uniis,
9 8 Acceptable resiilca. The following
ranee sets the li.Tiit on acceptable isokineiic
sampling results:
H 90 percent •.
I •«-.!=« .O.-.T. ".I , „'»„ I" Hi ! -cu, I .... ^ ,,, ,„ ,%..- I
L^jj^^^-Ll'^ZLJ-^121 Jli-°^^^S^!- -"^f- j—-'. ':1^ l'^-L
399
-------
Reference Method 13A Is amended
as follows:
(a) In section 3,. -the phrase "300
fig/Uter" is corrected to read "300 mg/
liter" and the parenthetical phrase "(see
section 7.3.6)" is corrected to read "tsee
section 7.3.4)".
(b) Section 5.1,5 Is revised to read as
follows:
B.1.5 Filter bolder—If located between the
probe and first ImpLngw, borosiiScata glass
with a, 20 Biesh Etalfijess steel screen Altar
support and a sUlcone rubber gasket; neither
a glass frit Biter support nor a sintered metal
filter support ma? be used It the filter Is in
front of tba Implngeta. II located between
the third and fourth implngere, borosUicate
glass with a glasa frit filter support and a
elllcone rubber gasket. Other materials ot
construction may be used with approval from
the Administrator, e.g., U probe liner Is stain-
less steel, th&n filter bolder may be atalrjesa
steel. The holder design shall provide a posi-
tive seal against leakage from the outside or
around the filter.
(c) Section 7.1.3 is amended by re-
vising the first two sentences of the sixth
paragraph to read as follows:
7,1.3 Preparation oJ collection train. • • •
Assemble the train ELI shown In Figure
13A-I with the filter between, the third and
fourth, ImpInRers. Alternatively, the filter
may be placed between the prob« and first
Implnger If a 20 mean stainless steel screen
13 used for the Slter support. « • •
• • • • •
(d) In section 7.3.4, the reference In
the first paragraph to "section 7.3.6" is
corrected to read "section 7,3.5".
400
-------
METHOD 13B
Determination Of Total Fluoride Emissions From
Stationary Sources-Specific Ion Electrode Method
1. Principle and Applicability,
I 1 Principle. Gaseous and pur'Jcul.ite flu-
orides ave withdrawn isokinecically from the
.-cure* using a sampling traUi. The fluorides
are collected in the implnger wa'er and on
tile filter of the sampling (rain. The weigh!
of tvt?.l fluorides In the trp.ln Is determined
bv the specific Ion electrode method.
1 2 Applicability. Tills method Is ap-
p't'at!e for the determination of fluorias
ci-ilj.sicns from stationary sources cr.ly '.vhe'i
specified by the test procedures -'or det«r-
niiir.rij: compliance with new source pcr-
lovinancs standards, Fluorocarhons such as
Treons, are not quantitatively collected or
ir.cav.ired by this procedure.
2. Range and Sensitivity.
The fluoride specific Ion electrode analyti-
cal method covers the range of 0.02-2,000 ^g
F irii: however, measurements of less than
0.1 -g F, ml require extra care. Sensitivity has
r_r_ bten determined.
3. interferences.
During the laboratory analysis, aiumlnv.m
in e>:CESs a: 300 mg.'lltsr und silicon dioxide
t:i excess of 300 ag.'liter \vV.l prevent complete
recovery of fluoride.
4. Precision, Accuracy and Stability,
Tr.e accuracyTrf-ftunrJde electrode measure*
merits has been reported by various re-
iearchers to be In the range of 1-5 percent In
i ccr.cBntration range of 0.04 to 80 mgil. A
change In the temperature of the sample will
ciifan|;e Che elGctrooc rvipo,'..-^; a change of
1°C will produce a 1.5 percent relative error
In Hie measurea,ent. Lajlc of stability in the
electrometer usacl to ir.casurs EMF can intro-
duce error- A:i error of 1 m'-lli.'olt la ihe EMF
measurement produces a relative error of 4
percent regardless of the absolute concen-
tration beniE; measured.
5. Apparatus,
5.1 Samp.e train. See Figure 13A-1
(Method ISA); it is similar to Uie Methcd 5
tram cxrept for the intercIuingeablHty of
the position of the Slter. Con-.mercial mcdels
o* this train are available. However, If one
desires to build his own, comp,cte construc-
tion details are described in APTD-G531; fcr;
chanEcs from the APTD-0581 document "and
for allowable modifications to Figure 13A-1,
see the following ,'ubsections.
The operating ,inil mr.inten^nce procedures
for the sampling train are described lu
APTD-Q57S, Since correct usjirre Is Impor-
tant !ri obtaining valid results, all usrrs
shouM rrad the APTD-0576 document ;u:d
fldopt the operating and malnten.-nce pro-
cedures ontl'.ned In it, unless otherwise spec-
ified herein.
5.1.1 Probe nor.zle—Stainlsss stec-1 (310)
with sharp, tapered leading eclgs. The angle.
of taper EUall be £30' and the taper shall be";
on the outside to preserve a constant inter-.
nil cinrni'ter. The probe rspzr.Ie shall be of,
the Sutton-hooK or elbow design, unless
otherwise specified by the Administrator.
The Will thickness of the nozzle shall be
less than cr equal to that of 20 gauge tub-
ing, i.e., 0,165 cm (O.OS5 in,) ani the distance
from the tip of the nozzle to the ilrst bend
or point o;' diaurbaace shall be at least two
times the outside r.uw.le diameter. The 1102-
7,lt shall te contracted ircm sjamless staln-
Isso steel tubing. Other configurations and
construction oiaterial may be used with ap-
pro', al from vhc Administrator.
A range or sizes suitable for Isokinetlc
sampling should be available, e.g., 0.32 cm
(|,8 in.) up to 1,27 cm (',-, in.) (or larger if
higher volume sampling trains lire used)
Inside diameter (ID) noz'-les In Inc-ements
of 0.16 cm O'i,; in.). Each noazle shall be
calibrated according to the procedures OUL-
lined in the calibration section.
5.1,2 Probs liner—Borosilicate glass or
stainless steel (316). When the filter Is lo-
cated immediately after the probe, a probe
heating system may fc« used T.O prevent filter
plugging resulting from mctsture conden-
sation. The temperature In the probe shall
not exceed 120±14-'C (249±2C'F|,
5.1.3 Pltot tube—Typc s. or other device
approved bv the Administrator, attached to
probe to allow constant monitoring of the
stack gas velocity. The face openings of the
pitot tube and the probe nozzle shall be ad-
jacent and parallel to each other, nor neces-
sarily on the same plane, during EampHv.g.
The free space between the :ioz?le and pitot
tube shall be at least 1.9 cm (0.75 ln.1. The
free space shall be set based on n 1.3 cm
(0.5 In.) ID nozzle, whicli Is the largest size
nozzle used.
The pitot tube must also meet the criteria
specified In Method 2 and be calibrated ac-
cording to the procedure In the calibration
section of that method.
5.1-4 Differential pressure gauge—In-
clined manometer capable of measuring
velocity head to within 10 nercent cf the
minimum measured value. Below a differen-
tial pressure of 1.3 mm (0.05 In.) water
gauge, mtcromanameters with sensitivities
of 0.013 mm (0.0005 in.) should be used.
However, mlcromanometRK are not easllv
adaptable to field conditions and are not
easy to use with pulsating flow. Thus, other
methods or devices acceptable to the Ad-
ministrator may be used when conditions
warrant.
5.1.5 Filter holder—Borosllicate glass
with a glass frit filter support and a slllcone
rubber gasket. Other materials of construc-
tion may ne used with approval from the
Administrator, e %. if probe lino1- is stain-
less steel, then filter holder miy be stainless
steel. The holder de.-ign shall provide a. posi-
tive seal aga'.nst leakage from the outside
or around the filter,
5.1.6 Filter heating system—When mois-
ture condensation 1? a problem, any heating
system capable of maintaining a temperature
around the filter holder during sampling of
no greater than 120il4'C (2.18 ±25T). A
temperature gauge capable of measuring tem-
perature to '.vHJiin. 3°C (5.4'F) shall be in-
stalled so that when the filter neater is used,
the temperature around the niter holder can
be regulated and monitored durlne samo'.Uu.
401
-------
Healing systems other than the one shown
In APTD-0581 may be used.
5.1.7 Ir.ipingers—Four impinjiets con-
nected as shown in Figure 13A-] with grou.icl
glass cor equivalent), vacuum tij-'ht. fittings.
The first, third, and fourth impir.gers Arc of
the Greenburg-Smith de,=ign. modified £>;/ re-
placing the tip with a 1-4 cm (Vj in.) insida
diameter glass tube extending to l',4 cm (i.j
in.) from the bottom af ths flask. The si>:or.cl
Implnger is of the Greenburg-Smith design
with the standard tip.
5.1-3 Metering system—Vacuum gauge,
leak-free pump, thermometers capable of
measuring te-Tipernrure to within 3"C
(~5°F), dry gas meter with 2 percent ac-
curacy at the required sampling rue. and
related equipment, or eq-j!vr.lenC
-------
results will be bis eel on actual measure-
ments made during the test.
Select a nozzle size based on the range of
velocity heads such that It Is not necessary
to change the nozzle size in order to maintain
Isoklnetic sampling rates. During the run, do
not change the nozele size. Ensure that the
differential pressure gauge Is capable of
measuring the minimum velocity head value
to within 10 percent, or as specified by the
Administrator,
Select & suitable probe liner and probe
length such that all traverse points can be
sampled. Consider sampling from opposite
sides for large stacks to reduce the length of
probes.
Select a total sampling time greater than,
or equal to the minimum total sampling
time specified In the test procedures Tor the
specific Industry such that the sampling time
per point is not less than 2 mln. or select
some greater time interval as specified Sy
the Administrator, and such that the sample
volume that will be taken will exceed the re-
quired minimum total gas sample volume
specified In the test procedures lor the spe-
cific industry. The latter Is based on an ap-
proximate average sampling rate. Note also
that the minimum tot&l sample volume is
corrected to standard conditions.
It Is recommended that a hair-integral or
Integral number of minutes be sampled at
each point In order to avoid timekeeping
errors.
In some circumstances, e.g. batch cycles. It
may be necessary to sample for shorter U^ies
at the traverse points and to obtain, smaller
gas sample volumes. In these cases, the Ad-
ministrator's approval must first be obtained.
7.1.3 Preparation or collection train. Bur-
Ing preparation and assembly or the sampling
train, Keep all openings where contamination
can occur covered until Just.prior to assembly
or until sampling is about to begin.
Place 100 ml of water in each of the first
two ircpliigers. leave the third Impmger
empty, fled place approximately 200-300 g or
more, if necessary, of preweighed SIIJCR gel in
the fourth implnger. Record the weight of
the silica gel and container en the dnta sheet.
Place the empty container tn a clean place
for later u?e In the sample recovery.
Place a filter in the filter holder. Be sure
that the filter Is properly centered and the
gasket properly placed so as to not allow the
sample gas stream to circumvent the filter.
Check filter for tears after assembly is com-
pleted.
When ijlass liners are used, install selected
nozzle using a Vlton A O-rlng; tlie Vlton A
O-ring Is installed as a seal where the nozzle-
is connected to a glass liner. See APTD-0576
for details, when metal liners are used, In-
stall the no?ale as above or by n leak free
direct mechanical connection. Mark the probe
with heat resistant tape or by some other
method to denote the proper distance Into
the stack or duct for each sampling point,
Unless otherwise specified by the Admin-
istrate.-, attach a temperature probe to the
metal sheath of the sampling probe so that
the sensor extends beyond the probe tip and
noes not touch any metal- Its position should
be about 1,9 to 2.54 cm (Q.75 to 1 In.) from
.the pilot tube and probe nozzle to avoid In-
terference with the gas flow.
Assemble the train as shown In Figure
13A-1 (Method I3A1 with the filter between
the third and fourth impingers. Alterna-
tively, the nicer may be placed between the
probe ami first Iraplr.ger. A filter heating sys-
tem may be used to prevent moisture con-
densation, but the temperature around the
filter holder shall not exceed 1200~14"C
(248-25 Fj. ((Note: Whatman No. 1 filter
decomposes at 150 C (300'-F)).] Record
filter location on the data sheet.
Place crushed ice around the impingers.
7.1.4 Leak check procedure—After the
.sampling train has been assembled, turn on
and set (if apolicab'.e) the probe and filter
heating svbiernfs) to reach a temperature
sufficient to a\oid condensation in the probe.
Allow time lor the temperature to stabilize.
Leak check the train at the sampling site by
plugging the nozzle and pulling a 330 mm
Hg (15 In. Hj) vacuum, A leakage rate in ex-
cess of 4" of the average sampling rate of
0.0057 mvniin. (0,02 eftrO, whichever is less,
is unacceptable.
The following leak check Instruction for
the sampling train described in APTD-0576
and APTD-0631 may be helpful. Start the
pump with by-pass valve fully open and
coarse adjust valve completely closed, Par-
tially open the coarse adjust valve and slow-
ly close the by-pass valve until 380 mm Hg
(15 In. Hgi vacuum is reached. Do not re-
verse direction oJ by-pass valve. This will
cause water to back up into the filter holder.
If 380 mm Hg (15 In. Hg) is exceeded, either
leak check at this higher vacuum or end the
leak check as described below and start over.
When the leak check is completed, first
slowly remove (he plug frcni the iniet to the
probe or filter holder and immediately turn
off tlie vacuum pump. This prevents tlie
water in the Implngetd from being forced
backward Into the Slter holder ).
a;id when sampling in air or a stack gas with
equivalent density (molecular weight, Mj,
equal to 28±4), wliich aid m the rapid ad-
justment of the isokinetic sampling rate
without excessive computations. APTD-0576
details the procedure lor using these nomo-
graphs. II Co and Ma are outside the above
stated ranges, do not use the nomograph un-
403
-------
less appropriate slops ai e liken to compen-
sate for the deviation*.
When the stael; is under significant neg-
ative pressure (height of Impinger stem),
take care to close the coarse adjust valve
before inserting the probe Into the stack to
avoid water backing into the flli;r holder. If
necessary, the pump may be turned on with
the coarse adjust valve closcS.
When the probe is in pcf.ltlon, block off
the openings around the probe and porthole
to prevent unrepresentative dilution or the
gas stream.
TrBveri? tbe stack eras0 section, as re-
quires by Method 1 or as specified by the Ad-
ministrator, being careful not to bump rhe
probe nozzle into the start walls When
sampling near the walls or when removing
or inserting the probe through the pcrt-
hcles to minimize chance or exlraet.nj da-
posited material.
During the tr;-t run, msiKe periodic atljvsr-
ments to keep the probe and (ir applicable)
nlt?r temperatures at their proper values.
Add more ice and, if necesrary, sal: to the
ice bath, to maintain a temperature of )crs
than 2CTC fSa-F) at the Implnger. s!I!ca'gel
outlet, to avoid excessive rnoMurc losse*.
Also, periodically check the level and zero
of the manometer.
If the pressure drcp across the filter be-
comes high enough to make Isoltinetlc sam-
pling difficult to maintain, the filter may be
replaced In the midst or a sample run. it i.s
recommended that another complete filter as-
sembly be used rather than attempting to
change the filter itself. ATur the new i^ter
or filter assembly Is Installed, conduct a.
leak check. The final emission results shall
be based on the summation of all filter
catches,
A single train shall be used for the entire
sample run, except lor filter and sili~3 gel
changes. However, if approved hy the Admin-
istrator, two or more trains may bo used for
a single test run when there are two or more
ducts or sampling ports. The anal emission
results shall be based on the total of all
sampling train catches.
At the end of the sample run, turn off the
pump, remove the probe and nozzle from
the stack, and record the flnnl dry gas meter
reading. Perform a leak check.1 Calculate
percent Isokinetic (see calculation section) to
determine whether another test run rhonld
be made. If there Is difficulty in maintainiui
isoklnetic rates due to source conditions, cun-
sult with the Administrator for possible
variance on the is-oklnetic rates
7.2 Sample recovery. Proper cleanup pro-
cedure/ begins as soon as the probe is re-
moved from the stack at the end of the
sampling period.
When the probe can be safely handled,
wipe off all external partlculate matter near
the tip of the probe nozzle and plp.cc u c ,p
over It to keep from losing part of the sam-
ple. Do not Cap oft the probo tt;> li^ht,;-
while the sampling train is cooling down,
as this would create a vacuum in the filter
holder, thus drawlns water from tile ira-
plngers into the filter.
Before moving the sample train to th?
cleanup site, remove the probe from r.lie
samp'.e train, wipe off the siiicone grease,,
and cap the open outlet of the probe. lie
careful not to loss any conclenjate, it prl2t v.'here the probe was fa.-;;c:itd
nnd cao it. Remove the umbilical cord 'ro;u
the last iniplnger ana cap tae imphiger. Af:_r
wiping oil His silictn'-e greiss:, c^p cr, :;te
filter holder outlet and implnger ;!"•.'.,-••.
1 With acceptability of the test run to be
baaed on the same criterion as In 7.1.4.
Ground floes stoppers, pU-.: ".e caps, rr ?<.-. .•:.>.
cap- ma" b£ used to close these opening".
Tr-an:f-.,r the probe and filior-irr.ninai:1 ,v -
serr.bly to the cleanup area. Thii area jliou', '•
'ue c.can ar,u protected from the wird ji ;•• t
the chances of eontnnii'Kitini Or lcs:-'.t-, c---
sarr.ple will be minimized.
Inspect the train prior to p-r.ct duri'Vj di -
asLombly and note any abnormal condu:oiii,
TJsu;g a graduated cylinder, rnea-jure ar.ci rfi-
ccrd the volume of the \v;Uer in the flr.-i
three nnnmgers, to the nearssv ml: &:.y coi:-
dfnsdte in the probe should oe in:"-"ac;--u >*;
this detft'i-iuiation, Treat the saaipis^ ai
follows;
7,2.1 Container No. 1. Transfer the Irr.-
pinj:er wa-.er from the graduated cylinder
t"> tliis container. Add the filter to th-,5
container. Wash nil sample exposed sur-
face.1;, including the probe tip, probe, first
three Impwgera, Smplnger connectors, fi'.tcr
holder, and graduated cylinder thoroughly
with distilled water. Wash each component
t'nrse separate times with water and eleiM
the prtibe titid nozzle with brushes. A max-
Imvmi wash of 500 ml Is used, and the wash-
Ings are added to the simple container
which must be made of polyethylene
7.2.2 Container No. 2. Transfer the silica
gel from tbe fourth Implnger to this con-
tainer and seal.
7.3 Analysis. Treat the contents of each
sanip'.e container as described below.
".3.1 Container No. 1,
1.3 1.1 Filter this container's contents, !n-
chidlne the Whatman Mo 1 filter, through
Whfviman No. 541 filter paper, or equivalent
Into a 1500 ml beaK
-------
to volume w:th distilled water and mix
thoroughly. Allow any undissolved solids to
settle.
732 Container No. 2, Weigh the spent
silica gel and report to the nearest 0-5 g.
7,3-3 Adjustment of acid/water ratio In
distillation flask—(Utilize a protective shield
when carrying out this procedure). Place 400
ml of distilled water In the distilling flask.
artd &dd 200 ml of concentrated H.SO,. Cau-
tion: Observe standard precautions when
mixing the H.SO, by slowly adding the acid
to the S;vsk with constant swirling. Add some
roll glass beads and several small pieces of
broken glass tubing and assem'nle the ap-
p.irncus as shown In Figure 13A-2. Heat the
fl?sk until it reaches a temperature of 175°c
to ."djust the acid/water ratio lor subsequent
distillations. Discard tile distillate.
7,3.4 Distillation—Coo] ti~e rontcrits ot
the d.stillatlon flask to below 80 C.- Pipette
nu aliquot of sample containing less
than 0.5 mg F directly into the distl'lin^
flask and add distilled water to make a total
volume of 220 ml added to the distilling
flask. [For an estimate of what size aliquot
does not exceed 0.6 mg F, select an aliquot
cf the solution and treat ns described in
Section 7.3.6. This will give an approxima-
tion of the fluoride content, but only an ap-
proximation since Interfering ions have not
been removed by the distillation step.)
Place a 250 ml volumetric flask at the con-
denser exit. Now begin dlstlllaUcn and
gradually increase the heat and collect all the
dis-.illate up to 175'C. Caution: Heating the
solution above 175°C will cause sulfurle acid
to distill over.
The acid in the distilling flasK can be
used until there is carryover of interferences
or poor fluoride recovery. An occasional
check of fluoride recovery with standard
solutions Is advised. The acid should
be changed whenever there is less than 90
percent recovery or blank values are higher
than 0.1 »g/ml.
7.3,5 Determination of concentration—
Bring the distillate In the 250 ml volumetric
flask to the mark with distilled water and
mix thoroughly. Pipette a 25 ml aliquot from
the distillate. Add an equal volume of TISAB
and mix. The sample should be at the
same temperature as the calibration stand-
ards when measurements are made, IJ
ambient lab temperatxire fluctuates more
than ±2°C from the temperature at which
the calibration standards were measured,
condition samples and. standards In a con-
stant temperature bath measurement. Stir
the sample with a magnetic stlrrer during
measurement to minimize Electrode response
time. If the stlrrer generates enough heat to
"chang* solution temperature, place a piece
of insulating material such as cork
between the stlrrer and the beaker. Dilute
samples (below 10-* M fluoride Jon content)
should be held In polyethylene or poly-
propylene beakers during measurement.
Insert the fluoride and reference electrodes
Into the solution. When a steady millivolt
reading Is obtained, record It. This may take
several minutes. Determine concentration
from the calibration curve. Between elec-
trode measurements, aoak the fluoride sens-
ing electrode In distilled water for 30 seconds
and then remove and blot dry.
8 Calibration,
Maintain a laboratory log of all-
calibrations,
8.1 Sampling Train.
B.l.i Probe nozzle—Using a micrometer,
measure the inside diameter of the nozzle
to the nearest 0.025 mm (0.001 in.]. Make
3 separate measurements using different
diameters each time and obtain the average
of the measurements. The difference between
the high and low numbers shall not exceed
0,1 mm (O.ODi in.).
When nozzles become nicked, dented, or
'corroded, they shall be reshaped, sharpened,
and recalibrated before use.
Each nozzle shall be permanently and
uniquely Identified.
8.1.2 Pitot tu'je—The pltot tube shall be
calibrated tccordlng to the procedure out-
lined In Method 2.
8.1.3 Dry gas meter and orifice meter.
Both meters shall be calibrated according to
the procedure outlined in APTD-0576, When
diaphragm pumps with By-pass valves are
used, check for proper metering system
design by calibrating the dry gas meter at an
additional flow rate of 0.0057 mVmln. (0.2
efra) with the by-pass valve fully opened
and then with it fully closed, K there is
more than ±2 percent difference In flow
rates v.-hen compared to the fully closed posi-
tion of the by-pass valve, the system Is not
desicned properly and must be corrected.
B.1.4 Probe heater calibration—The probe
neating system shall be calibrated according
to the procedure contained In APTD-0576-
Probes constructed according to APTD-0581
need not be calibrated If the calibration
curves in APTD-0576 are used,
8.1.5 Temperature gauges—Calibrate dial
and liquid filled bulb thermometers against
mercury-ln-plass thermometers. Thermo-
couples need not be calibrated. For other
devices, check with the Administrator.
8.2 Analytical Apparatus.
8.2.1 Fluoride Electrode—Prepare fluoride
standardizing solutions by serial dilution of
the 0.1 M fluoride standard solution. Pipette
10 ml of 0.1 M NaF into a 100 ml volumetric
flask and make up to the mark with distilled
water for a 10-s M standard solution. Use 10
ml of 10-1 M solution to make a 10-r' M solu-
tion in the same manner. Repeat for 10-" and
10-a M solutions.
Pipette 50 ml of each'standard into a sep-
arate beaker. Add SO ml of TISAB to each
beaker. Place the electrode in the most dilute
Standard solution. When a steady millivolt
reading Is obtained, plot the value on the
linear axis of semi-log graph paper versus
concentration, on the log axis. Plot the
nominal value for concentration o{ the
standard on the log axis, e.g., when 50 ml of
10-s M standard Is diluted with BO ml TISAB,
the concentration Is still designated "10-' M".
Between measurements soak the fluoride
sensing electrode In distilled water for 30
seconds, and then remove and blot dry.
Analyze the standards going from dilute to
concentrated standards. A straight-line cali-
bration curve will be obtained, with nominal
concentrations of 10-s, 10'', 10-', 10-% 10-'
fluoride molarlty on the log axis plotted
versus electrode potential (In millivolts) on
the linear seftle.
Calibrate the fluoride electrode dally, and
check It hourly. Prepare fresh fluoride stand-
ardizing solutions dally of 10-s M or less.
Store fluoride standardizing solutions In
polyethylene or polypropylene containers.
(Note: Certain specific Ion meters have been
designed Fpeclflcally for fluoride electrode
use and give a direct readout of fluoride Ion
concentration. These meters m?.y be used In
lieu of calibration curves for fluoride meas-
urements over narrow concentration ranges.
Calibrate the meter according to manufac-
turer's Instructions.)
9, Calculations,
Carry out calculations, retaining at least
one extra decimal figure beyond that of the
405
-------
acquired data. Round off figures after flnal
calculation,
9.1 Nomenclature.
A* = CTOSS sectional area of nozzle, m' (IV).
Ai = Aliquot ol to til sample added to still,
ml.
Bvi = Water vapor in the gas stream, propor-
tion by volume.
C. = Concentratloa of fluoride In stack gas,
mg/m3, corrected to standard conditions
of 20° C, 760 mm Hg fS8< P, 29.92 in. Hg)
on dry basis.
Ft=Total weight of fluoride in sample, mg.
/ = Percent of isokinetic sampling.
M = Concentration of fluoride from calibra-
tion curve, molanty.
mn = Total amount of p&rtlculate matter
collected, mg.
JH1 * = Molecular weight of water, 18 g/g-mole
(IB Ib/lb-mole),
7n.0 = Mass of residue of acetone after evap-
oration, mg.
Pnar = Barometric pressure at the sampling
sice, mm Hg (In. Hg).
P. = Absolute stack gas pressure, mm Hg (In.
. Hg).
P,-« = Standard absolute pressure, 760 mm
Hg (29.92 in. Hg).
R — Ideal gas constant.. 0,06236 mm Hg-m'/
•K-g-mole 121,83 in. Hg-ftV'R-1'o-mole).
7"™ = Absolute average dry gas meter tem-
perature (see Sg. 13A-31, 'K (°R).
Ti = Absolute average stack gas temperature
(see flg. 13A-3), =K f'R).
TV a ^Standard absolute temperature, 293'
K (528* R).
Vo — Volume of acetone blank, ml.
Va* = Volume of acetone used In wash, ml.
Vd = Vo!ume of distillate collected, ml.
Vir=Total volume of liquid collected In 1m-
plngers and silica gel, ml. Volume of water
In silica gel equals silica gel weight In-
. crease In grams times 1 ml/gram. Volume
of liquid, collected in Implnger equals flnal
volume minus Initial volume.
Vm — Volume of gas sample as measured by
dry gas meter, dcm (dcf).
Vm!. id) = Volume of gas sample measured by
the dry gas meter corrected to standard
conditions, dscra (dsof).
V^i.tfl, —Volume of water vapor In the gas
sample corrected to standard conditions,
scm (scf).
l'i=Total volume of sample, ml.
u/=Stack gas velocity, calculated by Method
2, Equation 2-7 using data obtained from
Method 5, m/sec (ft/sec).
W. = %Velght of residue In acetone wash, mg.
Atf = Averago pressure differential across the
orlflce (see fig, 13A-3), meter, mm H=O
(In. H--O).
pn=Denslty of acetone, mj/ml (see label on
bottle) , "
p, =-• Density or water, 1 g/ml (0.00220 lb/
ml).
O = Total sampling time, mln.
13.6 — Specific gravity of mercury.
100 — Corivorslou to percent.
9.2 Average dry gas meter temperature
and average oriflce pressure drop. See data
sheet (Flgv.re 13A-3 of Method 13A).
9.3 Dry gas volume. Use Section S.3 of
Method 13A.
9.1 Volume of Water Vapor. Use Section
D.4 of Method 13A.
9.5 Moisture Content. Use Section 9.5 of
Method 13A.
9 6 Concentration
9.6.1 Calculate the amount of fluoride in
the sample according to equation 13B--1.
Vt
F.-K — iV.) (.if)
At
where:
K :il9 mg/ir.l.
9.B.2 Concentration
gas. UFC Section fl.6,2
of
of
fluoride
Method
In stack
ISA.
9.7 Isoklnetlc variation. Use Section B.7
or Method ISA,
9.8 Acceptable reiiilts. Use Section 5.8 of
Method 13A.
10. Refcrencrs.
Bellack, Ervln, "Simplified Fluoride Distil -
lation Method." Journal ol the American
Water Works Astociaiion =50: 530-6 (1956).
MacLeod, Kathryn E , and Howard L. Crist,
"Comparison of the SPADNS — Zirconium
L?.ke and Specific Ion Electrode Methods of
Fluoride Determination In Stack Emission
Samples." Analytical Chc'-littry 45: 1272-1273
(1973).
Martin, Robert M. "Construction Details of
Isoklnetic Source Sampling Equipment,'*
Environmental Protection Agency, Air Pol-
lution Control Office Publication No. APTD-
0581.
1973 Annual Book of ASTM standards. Far:
23. Designation: D 1179-72.
Pom, Jerornt- J., "Maintenance, Ca', 'oration.
and Operation of Isoklnetic Source Sampling
Equipment," Environmental Protection
Agency, Air Pollution Control Office Publica-
tion No. APTD-0576.
Standard. Methods for the EiaminntioTi of
Water o''d Waiie Water, published Jointly by
American Public Health Association, Ameri-
can Water Works Association and Water Pol-
lution Control Federation, 13tb Edition
(1S71).
406
-------
Reference Method 138 13 amended
as follows:
(a) In the third line of section 3, the
phrase "300,ug/liter" is corrected to read
"300 mg/liter".
Cb> Section 5.1.5 is revised to read &a
follows:
5.1.5 Filter holder—a.' located between tha
probe and first Implngar, bom-ilUcate glass
with a, 20 mesh stainless steel screen filter
support and a silicons rubber gasket; neither
a glass frit filter support nor a sintered metal
Biter support may be used If the filter Is In
front of the Implnfere. If located between
the third and fourth tmpingera, borosilJcate
glass with a glass frit alter support and a
silicons rubber gaslcet. Other materials of
construction may be used with approval from
the Administrator, e.g., li probe llnsr Is stain-
less steel, then filter holder may be stainless
steel. The holder design shall provide a posi-
tive seal against loakaga from tha outside or
around the filter.
(c) Section 7.1.3 Is amended by revis-
ing the first two sentences of the sixth
paragraph to read as follows:
7.1.3 Preparation or collection train. • • •
Assemble the train as shown In Figure
13A-1 (Method 13A) with the fllter between
the third and fourth Imp'.ngers. Alterna-
tively, the filter may be placed between the
probe the first Irnplnger If a 20 mesh stain-
less steel screen la used for the niter sup-
port. • • •
(d) In section 7.3.4, the reference In
the first paragraph to "section 7.3.6" is
corrected to read "section 7,3.5".
407
-------
METHOD 14
Determination Of Fluoride Emissions From Potroom Roof
Monitors Of Primary Aluminum Plants
1. Principle and applicability.
1.1 Principle. Gaseous and participate
fluoride roof monitor emissions arc drawn
.Into a permanent sampling manifold through.
several large nozzles. The sample is trans-
ported frcrn the sampling manifold to ground
level through a duct. The gas In. the duct Is
sampled •using Method 13A or 13B—DETER-
MINATION OF TOTAL FLUOHIDE EMIS-
SIONS FROM STATIONARY SOURCES. Ef-
fluent velocity and volumetric flow rate are
determined with anemometers permanently
located la the roof monitor.
1.2 Applicability. This method Is applica-
ble for the determination of fluoride emis-
sions from stationary sources only when
specified by the test procedures for deter-
mining compliance with new source perform-
ance standards.
2. Apparatus,
2.1 Velocity measurement apparatus.
2.1,1 Anemometers, Vane or propeller
anemometers with a velocity measuring
threshold as low as 15 meters/minute ana a
range up to at least 600 meters/minute. Each
anemometer shall generate an electrical sig-
nal which caa be calibrated to the velocity
measured by the anemometer. Anemometers
shall be able to withstand dusty and corro-
sive atmospheres.
One anemometer shall be installed for
every 85 meters of rool monitor length. If
the roof monitor length divided by 85 meters
is not a whole number, round the fraction
to the nearest whole number to determine
the number of anemometers needed. UBB one
anemometer for any roof monitor less than
85 meters long. Permanently mount the
anemometers at the center ol each equal
length along the roof monitor. One ane-mom-
eter shall be installed ia the same section
of the roof monitor that contains the sam-
pling manifold (see section 2.2.1). Wake a
velocity traverse of the width of the roof
monitor where an anemometer is to be-placed.
This traverse may be made with any suit-
able low velocity measuring device, and shall
be made during normal process operating
conditions. Install the anemometer at a point
of average velocity along this traverse.
2.1.2 Recorders. Recorders equipped with
signal transducers for converting the electrl-
caJ signal from each anemometer to a con-
tinuous recording of air flow velocity, or to
an Integrated measure of volumetric flow,
For the purpose of recording velocity, "con-
tinuous" shall mean one readout per 15-
nilnute or shorter time interval, A constant
amount of time shall elapse between read-
ings. Volumetric Sow rate may be determined i
by an electrical count of anemometer revo-
lutions. The recorders or counters shall per-
mit identification of the velocities or flow
rate measured by each 'Individual anemom-
eter.
2.2 Roof monitor air sampling system.
2 2,1 Sampling ductwork. The" manifold
system arid connecting duct shall be per-
manently installed to draw an air sample
from the roof monitor to ground level, A
U t fcJCA'f t= *!,>;<-:
* 5 '*' _J ', !,__ C 35 0 02S Dlft
typical Installation of duct for drawing a
samp:e frcm a roof monitor to ground level
is shown In Figure 14-1. A plan ot a manl-
fcld system that is located in a roof monitor-
Is shown in Figure H-2. These drawings rep-
resent a typical Installation for a generalized
roof monitor. The dimensions on these fig-
ures may be altered slightly to make the
manifold system fit into a particular roof
monitor, but the general configuration shall
be Ic'.lcwed. There shall be eight nozzles, each
having a diameter of 0.40 to 0.50 meters. The
lencrh of the manifold system from fie first
nozzle to the eighth shall he 35 meters or
•eight percent of the length of the roof moni-
tor, whichever is greater. The duct leading
from the rocf monitor manifold shall be
round with a diameter of 0.30 to 0.40 meters.
As shown in Figure 14—2, each of the sample
lesjs of the manifold shall have a device, such
as n blast gate or valve, to enable adjustment
of flaw into each sample nozzle.
Locate the manifold along the length of
cho rcof monitor so that it lies near the
:ti!:tee::ti3n of the roof monitor. If the design
of a particular roof monitor makes this 1m-
porslble, the rr.aiillold may be located else-
v.-hcre along the roof monitor, but avr.id
;oc&tt:;g tl'e manifold near tb-5 cr.rts of the
rcof monitor or In a peefion where the
aH'mirmm reduction Tint arrangement Is not
tycic-,1 of the ren of the potroom. Center the
simple nozzles in. the throat of the rent
monitor. (See Fiaure 14-1.) Construct all
408
-------
'sample-exposed surfaces within the nozzles,
marUold and sample diiet of 316 stainless
Eteel. Aluminum may 'tie used If a new duct-
work system IK conditioned with fluoride-
laden roof monitor air for a period of six
v.'esks prior ID l::!tial testing. Other materials
cf construction may be used If It Is demon-
strated through comparative testing that
ttere 1' no loss of flue-Ides in the system. All
connections in the ductwork shall be leak.
free.
Locate two sample ports In a vertical sec-
tion of the duct between the roof monitor
a.id exhaust fan. The sample ports shall be at
least 10 duct diameters downstream and
two diameters upstream from any flow dis-
turbance such as a bend or contraction. The
two sample ports shall be situated 90° apart.
One of the sample ports shall be situated so
that the duct can be traversed In the plane
of the nearest upstream duct bend,
2.2,2 Exhcust fan. An industrial fan or
blower to be attached ta the sample duct
at ground level. (See Figure 14-1.) This ex-
haust fan shall have a maximum capacity
such that a large enough volume of air can
be pulled through the ductworK to main-
tain an isoklnetic sampling rate In all the
sample nozzles for all flow rates normally en-
countered In the roof monitor.
The exhaust fan volumetric flow rate Ehall
be adjustable so that the roof monitor air
can be drawn Isoklnetically Into the sample
nozzles, This control of flow may be achieved
by a damper on the Inlet to the exhauster or
by any other workable method.
"2.3 Temperature measurement apparatus,
2.3.1 Thermocouple. Installed In the roof
monitor near the sample duct.
•2.3.2 Signal transducer. Transducer to
change the thermocouple voltage output to
a temperature readout.
2.3.3 Thermocouple wire. To reach from
roof monitor to signal transducer and •
recorder.
3.3.4 Sampling train. Use the train de- •
scribed in Methods 13A and 13B—Determi-
nation of total fluoride emissions from sta-
tionary sources,
3. Reagents.
3.1 Sampling and analysis. Use reagents
described In Method 13A or 13B—Determi-
nation of total fluoride emissions from sta-
tionary sources.
4. Calibration.
4.1 Propeller anemometer. Calibrate the
anemometers so that their electrical signal
outnut corresponds to the velocity or volu-
metric flow they are measuring. Calibrate
according to manufacturer's instructions.
4.2 Mcnijold intake- nozzles. Adjust the ex-
haust fan to draw a volumetric flow rate
tracer to Equation 14-1) such that the en-
trance velocity Into each manifold nozzle
approximates the average efSuent velocity In
the roof monitor. Measure the velocity of the
air entering ea'-h noz7l« by inserting an S
ty^e pltot tube Into a 2-5 cm or less diameter
hale (see Figure 14-2) located In the mani-
fold between each blast gate (or valve) and
n O7.ila. The pltot tube tip shall be extended
Into the center of the manifold. Take care
t'j Insure that there is no leakage around the
pitot probe which could affect the Indicated
velocity In the manifold leg, If the velocity
o; ?.lr being drawn into each nozzle Is not
the same, open or close eacb. blast gate (or
valve) until the velocity In each nozzle Is the
same. Fasten each, blast gate (or valve) BO
that It will remain in this position and close
the pltot port hotes. This calibration shall be
performed whan the manifold system IB In-
stalled. (Note: It la recommended that this
calibration b« repeated at least once a year.)
S. Procedure.
5.1 Roof monitor velocity determination.
5.1.1 Velocity value for setting isokinetic
flow. During the 24 hours preceding a test
run, determine the velocity Indicated by the
propeller anemometer In the section of roof
monitor containing the sampling manifold.
Velocity readings shall be taken every 15
minutes or at shorter equal time Intervals.
Calculate the average velocity for the 24-hour
•period.
5.1.2 Velocity determination during a test
run. During the actual test run, record the
velocity or volume readings of each propeller
anemometer In the roof monitor. Velocity
readings shall be taken for each anemometer
every 15 minutes or at shorter equal time
intervals (or continuously) .
5,2 Temperature recording. Record the
temperature of the roof monitor every two
hours during the test run.
5,3 Sampling.
5,3.1 Preliminary air flow in duet. During
the 24 hours preceding the test, turn on the
exhaust fan and draw roof monitor air
through the manifold duct to condition the
ductwork. Adjust the fan to draw a volu-
metric flow through the duct such that the
velocity of gas entering the manifold nozzles
approximates the average velocity of the air
leaving the roof monitor.
5.3.2 Isokinetic sample rate adjustment.
Adjust the fan so that the volumetric flow
rate In the duct is such that air enters Into
the manifold sample nozzles at a velocity
equal to the 24-hour average velocity deter-
mined under 5.1.1. Equation 14-1 gives the
correct stream velocity which la needed In the
duct at the sample ports In order for sample
gas to be drawn Isokinetlcally Into the mani-
fold nozzles. Perform a pltot traverse of the
duct at the sample ports to determine II the
correct average velocity In the duct has been
achieved. Perform the pltot determination
according to Method 2. Make this determina-
tion before the start of a test run. The fan
setting need not be changed during the run.
8 (/M!
1 minute
__,___
where:
Fj=de3ired velocity in duct at sample
ports, meter/sec,
Dn=dlameter of a root monitor manifold
nozzle, meters.
D*=dlameter of duct at sample port,
meters,
V™=a.verage velocity of the air stream In
the roof monitor, meters/minute, as
determined under section 5.1.1.
5.3,3 Sample train operation. Sample the
duct using the standard fluoride train and
methods described In Methods 13A and 13B —
Determination of total fluoride emissions
from stationary sources. Select sample trav-
erse points according to Method 1. If a se-
lected sampling point Is less than one Inch
from the stack wall, adjust the location of
that point to one- inch away from the wall.
5.3,4 Each test run shall last eight hours
or more. It a question exists concerning the
representativeness of an eight-hour test, a
longer test period up to 24 hours may be se-
lected. Conduct each run during a period
when all normal operations are performed
underneath the sampling manifold. I.e. tap-
ping, anode changes, maintenance, and other
normal duties. All pots In the potroom shall
be operated In a normal manner during the
teat period.
5.3,5 Sample recovery. Same as Method
13A or 13B — Determination of total fluoride
emissions from stationary sources.
5.4 Analysis. Same as Method 13A or 13B —
Determination of total fluoride emissions
from stationary sources.
409
-------
6- Calculations.
6,1 IsoTtinetic sampling test. Calculate the
mean velocity measured during each sam-
pling run by the anemometer in the section
of the roof monitor containing the sampling
manifold. If the mean -velocity recorded dur-
ing a particular test Tun does not lall within
±20 percent of the mean velocity established
according to 5.3.2, repeat the run.
6.2 Average velocity of roof monitor gases
Calculate the average roof monitor velocltj
using all the velocity or volumetric flow read.
Ings from section 5.1,2.
6.3 Roof monitor temperature. Calculate
the mean value of the temperatures recorded
In section 5,2.
6.4 Concentration of fluorides in roof -moni-
tor air in mg F/m>. This Is given by Equation
13A-5 in Method 13A—Determination of
total fluoride emissions from stationary
sources.
6.5 Average volumetric flow from roof Is
given by Equation 14-2.
_ V«i (A) (Mi) Pm (294'K)
*m~ (Tm -f 273°) (760mmHg)
where:
Qn.=average volumetric flow from roof
monitor at standard conditions on
a dry basis, mVmin,
A=roof monitor open ares, m*.
V m j^ average velocity of air In the roof
monitor, meters/minute, from sec-
tion 6.2,
Pm=»tmospherlc pressure, mm Hg.
rr,i=roof monitor temperature, 'C, from
section 6.3.
M« = mole fraction of dry gas, which Is
100 — 100 (flu,)
given by Mt~ —
B«tf=l3 the proportion by volume of water
vapor la the gas stream, from
Equation 13A-3, Method 13A—De-
termination of total fluoride emis-
sions from stationary sources.
[33 FB 24877, Dec. 23, 1971, as amended at
38 FR 13S62, May 23. 1973; 39 FB 9319,
Mar. 8, 1974; 39 FR 13776, Apr. 17, 1S74; 39
FB 20794, June 14, 1974; 39 FB 39874, Nov.
12, 1974; 40 FB 33157. Aug. 6, 1975: 41 FR
3828, Jan. 26, 1976]
410
-------
APPENDIX H
FEDERAL STATIONARY SOURCE PERFORMANCE STANDARDS
Since December 1971 the Environmental Protection Agency
has promulgated source performance standards for twenty indus-
trial air pollution sources. These regulations have been pre-
sented in numerous issues of the Federal Register and can now
be found in the Code of Federal Regulations under Title 40 -
Protection of Environment; Chapter 1 - Environmental Protec-
tion Agency; Subchapter C - Air Programs; Part 60 - Standards
of Performance For New Stationary Sources.
This appendix presents a summary of these regulations as
found in the Code of Federal Regulations. This summary is an
updated listing as found originally in the following reference;
L. S. Chaput. 1976. Federal Standards of Performance
For New Stationary Sources Of Air Pollution - A
Summary of Regulations. Journal of the_Air_Pol-
lution Control Association. Vol. 26, No- 11;
1055-1060.
411
-------
Source Afecled
category lacilily
Subpart D:
Steam generators Coal fired boilers
(>250 million Btu/hr)
Promulgated
12/23/71 (38 FH 24876)
Revised
7/26/72f37FR14877t
10/15/73I38FR28566) Oil fired boilers
6/14/74I39FR20790)
1/16/75140FR2803)
10/6/75I40FR46250)
1 1/22/76(41 FR51398)
1/31/77I42FR5936) Gas fired boiiers
Subpart E:
Incinerators Incinerators
(>50 tons/day)
Promulgated
12/23/71 (36 FR 24876)
Revised
6/1 4/74 (39 FH 20790)
Subpart F.-
Portland cement plants Kiln
Promulgated
12/23/71 (36 FR 24876) Clinker cooler
Revised
. 6/14/74 (39 FH 20730) Fugitive
11/12/74 (39 FR 39874) Emyiss,on }
10/6/75 (40 FR 46250)
Subpart G:
Nitric acid plants Process equipment
Promulgated
12/23/71 (36FR2V.C6)
Revised
5/23/73(38FR13562!
10/1S/73!38FR285661
6/14/74(39FR20790)
10/6/75<40FR46250l
Subpart H:
SuHuric acid plants Process equipment
Promulgated
12/23/71 (36 FH 24876)
Revised
5/23/73O8FR13562)
10/15/73!38FR28566>
6/14/74C39FR20790!
10/6/75(40FR462SQ)
Polluian)
Participate
Opacity
SO2
NOX
(except lignite
and coal
refuse)
Particuiate
Opacity
SO2
NO,
Particulale
Opacity
NO,
Participate
Ptrticulale
Opacity
Participate
Opacity
Opacity
Opacity
NOX
SO2
Acid mist
Opacity
Emijsion level
0.10 lb/10« Btu
2C%
1.2 !b/10«Blu
OJOIb/106Btu
0,10 lb/10B Btu
20%; 40% 2min/hr
0.80 !b/10«E3tu
0.30 lb/10* Btu
0.1 Oib/ 1Q6 Btu
20%
0.20 lb/ 10s Btu
O.BO gr/dscf corrected
to 12% CO
0.3C Ib/ton
20%
0.10 Ib/ton
10%
10%
10%
3.0 Ib/ton
4.0 Ib/ton
0,15 ib/lon
10%
Monitoring
requirement
No requirement
Continuous
Continuous
Continuous
No requirement
Continuous
Continuous
Continuous
No requirement
No requirement
Continuous
No requirement
No requirement
No requirement
No requirement
Mo requirement
No requirement
No requirement
Continuous
Continuous
No requirement
No requirement
412
-------
Scarce
celegory
Subparl t:
Asphalt concrete plants
Promulgated
3/8/74 (39 FR9308)
Revised
10/6/75 (40 FR 46250)
Subpsrt J:
Petroleum relineries
Promulgated
3/B/74(39FR9308)
Revised
10/6/75 (40 FR A 6250)
AHecied
l.cimy Pollutant
Dryers; screening and Particular
weighing systems;
storage, transfer, and Opacity
loading systems; and
dust bEndling equipment
Catalytic cracker Participate
Opacity
CO
Fuel gas SC>2
combination
Emission level
0,C4gr/dscf
(90 rtig/dscm)
20%
1.0lb/1DOOIb
3Q% (3 rnin. exemplion)
0.05%
0,1 grHjS/dscl
(230 mg/dscm)
Monilorina
requirement
No requirement
No requirement
No requirement
Continuous
Continuous
Continuous
Subpart K:
Storage vessels for
petroleum liquids
Promulgated
3/8/74 (39 FR 9308)
4/17/74 (39 FR 13776)
6/14/74 (39 FR 20790)
Storage tanks
>40,OCU gal. capacity
Hydrocarbons
For vapor pressure
78-57Q mm Hg, equip
with floating rool,
vapor recovery system,
or equivalent; for
vapor pressure >570
mm Hg, equip with
vapor recovery system
or equivalent
No requirement
Subparl L:
Secondary lead
smelleis
Promulgated
3/8/74 (39 FR 9308)
Revised
4/17/74 (39 FR 13776)
10/6/75 (40 FR 46250)
Subpart M:
Secondary brass end
bronze plants
Promulgated
3/8/74 (39 FH 9308)
Revised
10/6/75 (40 FH4625D)
Subpart N:
Iron and sleel plants
Reverberalory and
blastfurnaces
Pot lurnaces
Reverberator/
furnace
Blast and
electric furnaces
Basic oxygen
process furnace
Paniculate 0.022 gr/dscf
(SO mg/dscm)
Opacity 20%
Opacity 10%
Paniculate 0.022gr/dscf
(50 mg/dscm)
Opacity 20%
Opacity 10%
Parliculate 0.022 gr/dscf
(50 mg/dscm)
No requirement
No requirement
No requirement
No requirement
No requirement
No requirement
No requirement
3/8/74 (39 FR 9308)
-------
Source
category
AHected
tacility
PoltMartl
Emission l
Monitoring
Subpart 0:
Sewage treatment
plants
Promulgated
3/8/74 (39 FR 9308)
Revised
M17/74 (39 FR 13776)
5/3/74 (39 FR 15396)
10/6/75 (40 FR 46250)
Sludge incinerators
Participate 1.30 Ib/ion
Opacity 20%
Mass or volume
of sludge
No requirement
Subpart Pi
Primary copper
smellers
Promulgated
1/15/76(41 FR2331)
Revised
2/26/76(41 FH8346)
Subpart Q:
Primary zinc
smelters
. Promulgated
1/15/76(41 FR2331)
Subpart R.:
Primary lead
smellers
Promulgated
1/15/76(41 FR 2331)
Subpart S:
Primary aluminum
reduction plants
Piomulgaied
1/26/76(41 FR3825)
Dryer
Roaster, smelling
furnace,* copper
converter
"Reverberaiory furnaces
that process high-
impurity feed materials
are exempt from
SOz standard
Sintering machine
Roaster
Blast or reverberatory
furnace, sintering
machine discharge end
Sintering machine,
electric smelting
furnace, converter
Potroom group
(a) Soderberg
plant
(b) Prebake
plan)
Anode bake plants
Particulate
Opacity
SO2
Opacity
Particulate
Opacity
SO2
Opacity
Particutate
Opacity
' S02
Opacity
(a) Total
fluorides
Opacity
(b) Total
fluorides
Opacity
Total fluorides
Opacity
0.022 gr/dscf
(50 mg/dscm)
20%
0.065%
20%
0.022 gr/dscf
(50 mg/dscm)
20%
0.065%
20%
0.022 gr/dscf
(50 mg/dscm)
20%
0.065%
20%
2.0 Ib/ton
10%
1.9 Ib/ton
10%
0.1 Ib/ton
20%
No requirement
Continuous
Continuous
No requirement
No requirement
Continuous
Continuous
No requirement
No requirement
Continuous
Continuous
No requirement
No requirement
No requirement
No requiremant
No requirement
No requirement
No requirement
414
-------
Source
cplcgory
Subparl T:
Phosphate fertilizer
plants
Promulgated
8/6/75 (40 FR 33152)
Subpart U:
Subpart V:
Subparl W:
Subpart X:
Subparl Y:
Coal preparation
plants
Pfomulqsted
1/15/76(41 FR2232)
AHcclcd . .
f«,hly
Wo* process To'.^l fluorides 0,03 Ib/lcn
phosphoric acid
Superphosphonc acid Total flue-rides 0.01 Ib/lon
Diammonium Total fluorides 0.06 Ib/ton
phosphate
Triple super- Total fluorides 0.2 Ib/ton
phosphate
Granular triple Tolal fluorides 5.0x10"*
superphosphate Ib/hr/ton
Thermal dryer Participate 0.031 gr/dscf
J0.07Q g/dscm)
Opacity 20%
Pneumatic coal Participate 0.01 8 gr/dscf
cleaning equipment (0.040 g/dsem)
Opacity 10%
Processing and Conveying Opacity 20%
equipment, storage
systems, transfer and
loading systems
Mon.tofinq
requirement
Tolal press jry
diop across
process
scrubbing
system
Tola! pressure
drop across
process
scrubbing
system
Total pressure
drop across
process
scrubbing
system
Total pressure
drop across
process
scrubbing
system
Total pressure
drop across
process
scrubbing
system
Temperature
Scrubber
pressure loss
Water pressure
No requirement
No requirement
No requirement
No requirement
Subpart Z:
Ferroalloy production
facilities
Promufgafed
5/4/76(41 FR1S497)
Revised
5/20/76(41 FR 20659)
Electric submerged
arc furnaces
Participate
0.99 Ib/Mw-hr
(0.45 kg/Mw-hr!
("high silicon alloys'')
0.51 Ib/Mw-hr
(0.23 kg/Mw-hr)
(chrome and
manganese alloys)
No visible emiss'Ons
may escape lurnacs
capture system
No requirement
Flowrate
monitoring
in hood
415
-------
Source
category
Aflocted
tecil.ty
Emission Icvei
Ferroalloy production
facilities (cent.)
Opacity
CO
Dust handMng equipment Opacity
No visible emission
may escape tapping
system for >4Q% of
each lapping period
15=13
20% volume basis
10%
Fbwralo
monitoring
in hooa
No requ:re:ren!
No requirerren;
Subpart AA:
Iron and steel
plants
Promulgated
9/23/75 (40 FR 43050)
Electric arc furnaces
Paniculate
Opacity
(a) contro!
device
(b) shop roof
Dus, handl.ng equipment Opacity
OOQ52gr/dscf
(12 mg/dscm)
3%
0, except
20%—charging
40%—tapping
10%
No
Continuous
Flowrate
monitoring
in capture hcod
Pressure
mon'toring
in DSE system
No requirement
416
-------
BIBLIOGRAPHY
Allen, T. 1975. Particle Size Measurement. John Wiley
and Sons, New York, New York. 454 pages.
Bickelhaupt, R. E. 1974. Sodium Conditioning To Reduce
Fly Ash Resistivity. EPA-650/2-74-092. U.S. Environ-
mental Protection Agency, Washington, DC 20460.
15 pages, NTIS--PB 236 922/$3.25.
Bird, A. N., Jr. 1973. Summary of Techniques Used for
Measuring Concentration and Size of Fine Particulate
Emission. Memorandum Prepared for the Working Group
for Stationary Source Air Pollution Control Technology.
Southern Research Institute, Birmingham, Alabama. 14
pages.
Blake, D. E. 1975. Symposium On Particulate Control In
Energy Processes. EPA-600/7-76-010. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina
2771.1. 584 pages.
Brenchley, D. L., G. D. Turley, and R. F. Yarmac. 1974.
Industrial Source Sampling. Ann Arbor Science Publish-
ers, Inc., Ann Arbor, Michigan. 484 pages.
Davies, C. N. 1966. Aerosol Science. Academic Press,
New York, New York 10003. 468 pages.
417
-------
Bibliography (Cont'cl.)
Drebmel, D. C. 1974. Seminar On Electrostatic Precipitators
and Fine Particulate. September 1973. EPA-650/2-74-
081. U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina 27711. 131 pages.
NTIS-PB 236 676/$4.75.
Electrostatic Precipitators - Special Report. 1975. Journal
of the Air Pollution Control Association. Volume 25,
No. 2, February, 1975.
Feazel, C. E. 1975. Symposium On Electrostatic Precipitators
For The Control Of Fine Particulates. EPA-650/2-75-
016. U.S. Environmental Protection Agency, Washington,
DC. 501 pages. NTIS-PB 240440/$12.00.
Fuchs, N. A. 1964. The Mechanics of Aerosols. The MacMillan
Company, New York, New York. 408 pages.
Gooch, J. P., J. R. McDonald, and S. Oglesby, Jr. 1975.
A Mathematical Model Of Electrostatic Precipitators.
EPA-650/2-75-037. U.S. Environmental Protection Agency,
Washington, DC. 162 pages. NTIS-PB 246188/$6.25.
Gooch, J. P. and J. D. McCain. 1975. Particulate Collection
Efficiency On A Wet Electrostatic Precipitator. EPA-
650/2-75-033. U.S. Environmental Protection Agency,
Washington, DC. 60 pages. NTIS-PB 244173/$4.25.
Gooch, J. P., G. H. Marchant, Jr., and L. G. Felix. 1976.
Particulate Collection Efficiency Measurements On An
Electrostatic Precipitator Installed On A Pulp Mill
Recovery Boiler. EPA-600/2-76-141. U.S. Environmental
Protection Agency, Research Triangle Park, North Caro-
lina 27711. 121 pages. NTIS-PB 255 297/$5.50.
418
-------
Bibliography (Cont'd.)
Gooch, J. P. and A. H. Dean. 1976. Wet Electrostatic Pre-
cipitator System Study. EPA-600/2-76-142. U.S. Environ-
mental Protection Agency, Research Triangle Park, North
Carolina 27711. 204 pages. NTIS-PB 257 128/$7.75.
Katz, J, J. R. Zarfoss, H. C. Dorhrmann, R. L. Bump, E.
L. Coe, Jr., and C. L. Martzoloff. 1975. Information
Required For the Selection and Application of Electrosta-
tic Precipitators for the Collection of Dry Particulate
Material. Journal of the Air Pollution Control Associa-
tion, Vol. 25, No. 4: 362-368.
Liu, B. Y. H. 1976. Fine Particles: Aerosol Generation,
Measurement, Sampling, and Analysis. Academic Press,
New York, New York 837 pages.
Marple, V. A. and K. Willeke. 1976. Impactor Design.
Atmospheric Environment. Vol. 10: 891-896.
McCain, J. D., K. M. Gushing, and A. N. Bird, Jr. 1973.
Field Measurements Of Particle Size Distribution With
Inertial Sizing Devices. EPA-650/2-73-035. U.S.
Environmental Protection Agency, Washington, DC 20460.
52 pages. NTIS-PB 226292/$.75.
McCain, J.D., K. M. Gushing, and W. B. Smith. 1974. Methods
For Determining Particulate Mass and Size Properties:
Laboratory and Field Measurements. Journal of the
Air Pollution Control Association, Vol. 24, No. 12:
1172-1176.
419
-------
Bibliography (Cont'd.)
Nichols, G. B. 1973. A Discussion Of The Problems Encountered
And Approaches To Their Solutions For Electrostatic
Collection Of High Resistivity Fly Ash. American Power
Conference, May 10, 1973. 20 pages.
Nichols, G. B. 1974. Techniques For Measuring Fly Ash
Resistivity. EPA-650/2-74-079. U.S. Environmental
Protection Agency, Washington, DC. 49 pages. NTIS-
PB 244140/$3.75.
Nichols, G. B. 1975. Particulate Emission Control From
Pulp Mill Recovery Boilers With Electrostatic Precipi-
tators. Southern Research Institute, Birmingham, Alabama.
17 pages.
Nichols, G. B, and J. D. McCain. 1975. Particulate Collection
Efficiency Measurements On Three Electrostatic Precipi-
tators. EPA-600/2-75-056. U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina 27711.
84 pages. NTIS-PB 248220/55.00.
Noll, K. E. and W. T. Davis. 1976. Power Generation:
Air Pollution Monitoring and Control. Ann Arbor Pub-
lishing Co., Ann Arbor, Michigan. 555 pages.
Oglesby, S. Jr., and G. B. Nichols. 1970. A Manual Of
Electrostatic Precipitator Technology. Prepared Under
Contract CPA 22-69-73 For The National Air Pollution
Control Administration, Cincinnati, Ohio. 875 pages.
Oglesby, S. Jr. 1975. Survey Of Information On Fine Par-
ticle Control. EPRI 259/Final Report. Southern Research
Institute, Birmingham, Alabama. Prepared For Electric
Power Research Institute, Palo Alto, California. 247
pages.
420
-------
Bibliography (Cont'd.)
Ragland, J. W., D. H. Pontius, and W. B. Smith. 1976.
Design, Construct, and Test A field Usable Prototype
System For Sizing Particles Smaller Than 0.5 m Dia-
meter-Special Summary Report (SORI-EAS-76-094) Southern
Research Institute, Birmingham, Alabama. 65 pages.
Rose, H. E. and A. J. Wood, 1966. An Introduction To Elec-
trostatic Precipitation In Theory And Practice. Con-
stable and Company, London, England. 212 pages.
Sem, G. J. , e_t al. 1971. State of the Art: 1971. Instrumen-
tation For Measurement of Particulate Emissions From
Combustion Sources. Volume I - PB 202665, Volume II
PB 202666, Volume III - PB 223 393,Volume IV - PB -
231 919. Thermosysterns, Inc., St. Paul, Minnesota.
605 pages.
Smith, W. B., K. M. Gushing, and J. D. McCain. 1974. Par-
ticle Sizing Techniques For Control Device Evaluation.
EPA-650/2-74-102. U.S. Environmental Protection Agency,
Washington, DC. 127 pages. NTIS-PB 240670/55.75.
Smith, W. B. , K. M. Gushing, G. E. Lacey, and J. D. McCain.
1975. Particle Sizing Techniques For Control Device
Evaluation. EPA-650/2-74-102a. U.S. Environmental
Protection Agency, Washington, DC. 132 pages. NTIS-
PB 245 184/$5.75.
421
-------
Bibliography (Cont'd.)
Smith, F. H. 1974. The Effects Of Nozzle Design And Sampling
Techniques On Aerosol Measurements. EPA-650/2-74-076.
U.S. Environmental Protection Agency, Washington, DC.
89 pages.
Southern Research Institute. 1975. Proceedings Of The
Workshop On Sampling, Analysis, And Monitoring Of Stack
Emissions. EPRI Special Report 41. Prepared for the
Electric Power Research Institute, Palo Alto, Califor-
nia. 346 pages.
Spencer, H. W., III. 1976. Electrostatic Precipitation:
Relationship Between Resistivity, Particle Size, and
Sparkover. EPA-60Q/2-76-144. U.S. Environmental Pro-
tection Agency, Research Triangle Park, North Carolina
27711. 68 pages. NTIS-PB 257 130/$4.50.
Stern, A. C. 1968. Air Pollution. Academic Press, New
York, New York. 644 pages.
Strauss, W. 1975. Industrial Gas Cleaning. Pergamon Press,
New York, New York.
Texas Air Control Board. 1975. Compliance Sampling Manual.
Texas Air Control Board, Source Sampling Section,
Austin, Texas 78758.
*
U.S.-U.S.S.R. Working Group-Stationary Source Air Pollution
Control Technology. 1974. Proceedings of the Symposium
on Control of Fine Particulate Emissions from Industrial
Sources. San Francisco, California, January 15-18,
1974.
422
-------
Bibliography (Cont'd.)
Varga, J. 1976. Control Of Reclamation (Sinter) Plant
Emissions Using Electrostatic Precipitators. EPA-
600/2-76-002. U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina 27711, 81 pages.
NTIS-PB 249 505/S5.00.
White, H. J. 1963. Industrial Electrostatic Precipitation,
Addison-Wesley Publishing Co., Inc., Reading, Massachu-
setts. 376 pages.
423
-------
TECHNICAL REPORT DATA
fftcasc read laaruciiuns on the reverse btforu complcnnpj
1 =iE=CRT NO.
EPA-600/7-77-05D
4 TITLE AND SUBTITLE
Procedures Manual for Electrostatic Preeipitator
Evaluation
7. AL'THOR(S)
Wallace B. Smith. Kenneth M. Gushing, and
Joseph D. McCain
3.1
PB 269 698
F-DATE-----
June 1977
6. PERFORMING ORGANIZATION CODE
SORT.--EAS-77-3 10
8. PERFORMING ORGANIZATION' REPORT M
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Southern Research Institute
2000 Ninth Avenue, South
Birmingham, Alabama 35205
10. PROGRAM E<_EM£NT NO.
EHE624
11. CONTRACT,GHAN/T NO.
63-02-2131
Technical Directive 20B')4
12. SPONSORING AGENCY NAME AND ADDRESS
EPA. Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT ANi D PERIOD COvEPv
TD Final; 11/76-3/77
14. SPONSORING AGENCY CODE
EPA/600/13
is. SUPPLEMENTARY NOTES IERL-RTP project officer for this
Drop S2, 919/519-8411 Ext 2557,
report is D. Brace Harris , Mall
16. ABSTRACT-.
"•The purpose of this procedures manual is to describe methods
to be used in experimentally characterizing the performance of electro-
static precipitators for pollution control. A detailed description
of the mechanical and electrical characteristics of precipitators is
•given. Procedures are described for measuring the particle size dis-
tribution, the mass concentration of particulate matter, and the con-
centrations of major gaseous components of the flue gas-aerosol mix-
ture. Procedures are also given for measuring the electrical resis-
tivity of the dust. A concise discussion and outline is presented
which describes the development of a test plan for the evaluation of
a precipitator. By following this outline useful tests may be perform-
ed which range in complexity from qualitative and relatively inexpensive
to rather elaborate research programs.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
(3.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Gioui;
Air Pollution Aerosols
Electrostatic Precip- Dust
itators Resistivity
E valuation Tes ts
Measurement
Flue Gases
Air Pollution Coatrol
Stationary Sources
Size Distribution
Mass Concentration
13 B
14 B
21B
07D
11G
13, DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report/
Unclassified
21. I
I 2O. SECURITY CLASS (This page/
I Unclassified
22. PHiC
PC A
EPA Form 2220-1 (9-73)
------- |