EPA-600/2-77-077
April 1977
Environmental Protection Technology Series
DESIGN, DEVELOPMENT, AND
DEMONSTRATION OF A FINE
PARTICIPATE MEASURING DEVICE
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental Protection
Agency, have been grouped into five series. These five 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 five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
Ths report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumenta-
tion, equipment, and methodology to repair or prevent environmental degradation from point
and non-point sources of pollution. This work provides the new or improved technology
recuired for the control and treatment of pollution sources to meet environmental quality
standards.
EPA REVIEW NOTICE
Th s report has been reviewed by the U.S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily reflect the views and
policy of the Agency, 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.
-------
EPA-600/2-77-077
April 1977
DESIGN, DEVELOPMENT, AND
DEMONSTRATION OF A FINE
PARTICULATE MEASURING DEVICE
by
Pedro Lilienfeld, Daniel P. Anderson,
and Douglas W. Cooper
GCA/Technology Division
Burlington Road
Bedford, Massachusetts 01730
Contract No. 68-02-1341
ROAP No. 21ADL-018
Program Element No. 1AB012
EPA Project Officer: William B. Kuykendal
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
-------
ABSTRACT
This report presents the design, development, and testing of a fine
particulate source monitoring instrument for the real-time measurement
of the mass concentration as a function of aerodynamic particle size.
This report also includes a literature review and selection of the oper-
ating principle on which the instrument is based.
The device described in this report size-segregates the particulates
by means of inertial jet-to-plate impaction on a continuously moving
substrate and determines the collected mass by the method of beta radi-
ation attenuation.
The collection-detection system consists of a seven-impaction stage
cascaded configuration for direct insertion into a stack, with beta
mass sensing at each collection stage. Although the initial objective
of this program was to develop an instrument compatible with operation
in a stack environment at temperatures up to 260°C (500 F), this objec-
tive could not be met as it was determined that the beta detectors
failed to operate satisfactorily when exposed to such temperatures.
The instrument was nevertheless completed and tested at room tempera-
ture conditions within a test tunnel with satisfactory results.
This report was submitted in fulfillment of Project No. 1-359, Contract
No. 68-02-1314 by the GCA/Technology Division under the sponsorship of
the Environmental Protection Agency. Work was completed as of
31 August 1975.
111
-------
CONTENTS
No. Page
Abstract iii
List of Figures vi
List of Tables viii
Acknowledgments ix
Sections
I Introduction 1
II Conclusions 4
III Recommendations 6
IV Literature Review 7
V Instrument Development 67
VI Instrument Testing 98
VII References 114
-------
FIGURES
No.
1 The Size of Airborne Contaminants (Mine Safety 16
Appliances Company)
2 Product of Light-Scattering Cross-Section and Measured 18
Concentration for Atmospheric Aerosol. From Pueschel
and Noll (1967). Visibility Inversely Proportional
to b.
3 "Respirable" Dust Mass Measurement Sampling Criteria 19
(Lippman, 1970)
4 Generalized Aerosol Assessment System Schematic 23
5 Three Types of Aerosol Assessment Systems in General Use 34
6 Generalized Elutriator/Centrifugator 43
7 Circular Nozzle Impaction Efficiency as a Function 48
of q. (Mitchell and Pilcher, 1959)
8 Particle Cut-Off Diameter for a Circular Nozzle 50
as a Function of Jet Diameter
9 Assembled Seven-Stage Collector-Sensor 71
10 Single Modular Impaction-Cassette Assembly 72
11 Cassette, Impaction Nozzle and Collector Inlet Sections 73
12 Substrate Drive and 90° Rotation Mechanisms 75
13 Experimental Quench Circuit (First Version) 86
14 Simplified Quenching Circuit 88
15 Log Computation Module 92
VI
-------
FIGURES (continued)
No. Page
16 Strip Chart Recording of Dust Collection Run With Test 93
"Breadboard"
17 Front View of Control Console 97
18 Comparison of GCA In-Stack Beta Impactor Data With „ 110
Corresponding Brink Curve (Concentration = 0.267 g/m )
19 Comparison of GCA In-Stack Beta Impactor Data With 111
Corresponding Brink Curve (Concentration = 0.955 g/m )
20 Relative Centerline Concentration Versus Distance Down- 112
stream From Obstruction in EPA Test Tunnel (Data Points
From Figure 2 of Reference 99)
vii
-------
TABLES
No. Page
1 Books Concerning the Behavior of Aerosols and Their 9
Analysis
2 Periodicals Containing Articles Concerning the Behavior 12
of Aerosols and Their Analysis
3 Efficiency of Dust-Arresting Equipment for Various Sizes 20
of Dust (Ball and Griffiths, 1970)
4 Summary of Types of Size-Analysis Techniques 29
5 Recently Developed Particle Sizing Techniques 32
6 Various Parameters of Aerosol Concentration and Techniques 36
by Which They May Be Determined
7 Sensor Principles Affected by Particle Properties 38
(Dorsey and Burckle, 1971)
8 Methods of Particle Size Discrimination 39
9 Comparison of Mass Sensing Techniques 52
10 Z/M Ratios of Selected Elements and Compounds Typical 61
of Particulate Emissions from Coal-Fired Steam Power
Plants
lla Design Sheet for Industrial and Commercial Facilities 99
lib Design Sheet for Industrial and Commercial Facilities 101
lie Design Sheet for Industrial and Commercial Facilities 103
viii
-------
ACKNOWLEDGMENTS
The useful contributions of James Congdon and numerous other members of
the GCA/Technology Division staff are gratefully acknowledged. The
cooperation and assistance provided by Mr. Charles Gooding and other
RTI personnel, as well as by Mr. Bill Kuykendal of the EPA Industrial
Environmental Research Laboratory are acknowledged with gratitude.
IX
-------
SECTION I
INTRODUCTION
The objective of this program was the development of a fine particulate
mass sizing sampler designed to measure the mass concentration of par-
ticulates as a function of size, upstream and downstream of control de-
vices in coal-fired steam power plants.
Mass-size monitoring of particulates contained in stack gases is a
challenging problem for the instrumentation scientist. In-stack samp-
ling and sizing of particulates has, in general, been performed by
extracting a sample, under manually adjusted isokinetic flow conditions,
and collecting the particles by filtration or by impaction. The col-
lected material is then evaluated gravimetrically; i.e., by weighing
before and after collection. The average mass concentration as a func-
tion of size is then calculated on the basis of the total volume of air
sampled during the collection period. This method is tedious, inef-
ficient and incompatible with the continuous recording, automatic data
transmission and processing that are required to handle the increasing
volume of information resulting from intensive air pollution monitoring
programs. Furthermore, these manual methods do not provide real-time
information, nor the temporal resolution required to determine emis-
sion variability both from the point of view of total mass emission
as well as size distribution.
Until recently the only methods available for recording-type particu-
late monitoring were based on such indirect sensing principles as mea-
suring certain parameters related to either particle area or particle
-------
number concentration. These techniques are applicable to mass measure-
ments only when particle density is known. In addition, if light scat-
tering or absorption are used, variations in particle size, shape, and
index of refraction, such as are found in the stack environment, can
introduce serious measurement errors. A recently developed technique,
based on piezoelectric mass detection, utilizes either electrostatic
precipitation or inertial impaction to deposit particulates on a quartz
crystal whose oscillation frequency is a function of the collected mass.
Operational difficulties, and the basic incompatibility of this tech-
nique with prolonged unattended operation, have restricted its use to
certain laboratory applications.
The preferred solution to the problem of automated mass measurement of
particulates from stack gases is based on beta radiation absorption
sensing of the material collected on a suitable substrate. This tech-
nique was recommended as a result of a review program undertaken for
the Environmental Protection Agency. This recommendation has been con-
firmed and supported independently by the experience acquired by the
recent application of beta absorption to a variety of particulate air
pollution sensing instruments and the experimental use of such instru-
mentation under laboratory and field conditions.
/
The basic instrumental technique of the stack-gas particulate mass-
sizing recording monitor is based on the principle of combined inertial
impaction-beta attenuation mass sensing. Aerosol mass monitoring by beta-
radiation absorption has been used before by various workers in the
2-12
field of air pollution and industrial hygiene. These investigators
applied beta attenuation to the measurement of particulate matter sus- •
pended in a column of air, collected on filter media, or on a metallized
foil strip by electro-precipitation. Direct measurement of aerosols by
attenuation of betas by particles suspended in a column of air has
proven to be impractical, except for extremely high concentrations
3
(greater than 50 g/m ), due to the overwhelming predominance of the
-------
beta-attenuation due to the air molecules, and the effect of air density
fluctuations associated with in-stack temperature and pressure varia-
tions. The filter collection/beta-absorption approach has been applied
to a number of instruments for the automated measurement and recording
Q
of aerosol mass concentration, in ambient environments.
The instrument described in this report is based on particulate size sep-
aration and measurement by means of a multiple-stage cascade impaction
configuration with a continuously advancing substrate, combined with beta
attenuation sensing at each stage. This collection-sensing head was de-
signed for direct in-stack operation, rather than the usual sample extrac-
tion approach. This program represents a further application of the beta
absorption concept in the direction of an instrument specifically designed
for in-stack measurements of the size distribution of particulates by mass
to be used for the quantitative evaluation of the performance of control
methods for power plant particulate emissions.
-------
SECTION II
CONCLUSIONS
The development of a recording particulate mass-sizing instrument for
in-stack particulate measurements required a very substantial effort
whose difficulties exceeded original estimates. The principal problem
encountered during the execution of this program was associated with
the failure of the present state-of-the-art technology of beta radia-
tion detectors capable of operating at temperatures of the order of
200 C or above. Several promising avenues of solution of this problem
were explored but were not carried to fruition as the required efforts
exceeded the scope and intent of this program.
The feasibility of the concept of multiple cascaded impaction stages
with continuously advancing collection substrates combined with beta
radiation mass sensing at each stage, with near real-time readout of a
signal proportional to fractional stage mass concentration was dem-
onstrated. The principal objection to this general approach, without
consideration of the problem of high temperature beta detection, re-
sides in the intrinsic complexity of the resulting system which could
probably be simplified only to a relatively minor degree. This approach
should thus be considered as a useful research tool but its full appli-
cation must be contingent on the development and future availability of
beta detectors compatible with the above-mentioned temperatures.
The results of the test series performed at the EPA facilities with the
system at room temperature were quite encouraging considering that no
-------
prior optimization testing had been performed and the operational
idiosyncracies of this rather complex measurement system were largely
unknown before these tests.
-------
SECTION III
RECOMMENDATIONS
Although the feasibility of the technique of cascaded collection on
moving impaction substrates has been proven within this program, it is
felt that the complexity of the mechanical and electronic systems is
somewhat excessive to warrant the pursuit of this particular approach
beyond the research prototype completed within this program.
The problem of high temperature beta detection warrants further efforts
as a solution would provide an extremely useful tool for the mass mea-
surement, sensing, and sizing of in-stack particulates using several
alternative collection and sampling techniques. Two general approaches
should be attacked: the development of an electro-chemically stable
high temperature geiger detector, presumably to be operated with ex-
ternal quenching; i.e., electronic pulse quenching, circuitry for which
has been developed within the performance of the program reported
herein; and the development of a high temperature compatible scintil-
lation detection-fiber optic beta detection scheme using a CaF2 crystal as
a scintillator, and quartz fiber optics to carry the light pulse signal
to the exterior of the stack.
Two promising collection and sampling techniques should be explored
contingent on the development of a high temperature-compatible beta
detector: cascaded virtual impaction with fixed, discardable filters
as developed under another EPA contract; and total mass concentration
sensing by means of beta attenuation sensing through a small column of
gas before and after a continuous flow miniature electrostatic
precipitator.
-------
SECTION IV
LITERATURE REVIEW
INTRODUCTION AND SUMMARY
The characterization of particulates suspended in stack gases has been
receiving increasing attention as the need to establish more specific
emission regulation standards has evolved, at the same time as a better
quantitative assessment of the effectiveness of control methods and de-
vices has become an imperative requirement.
In-stack sampling for particulate mass emission determinations has
been, and continues to be performed in general by extracting a sample
and collecting the particles externally to the stack by filtration.
Gravimetric techniques are then used to evaluate the mass of collected
material. Sizing of particulates is usually accomplished by means of
selective inertial impaction, preferably within the stack itself to
preserve the pristine size distribution of the aerosol to be assessed.
A number of variations of these two basic techniques have evolved, but
their intrinsic disadvantages of tediousness, lack of time resolution,
and the requirements for laboratory facilities for analysis and/or
weighing of the collected samples remains a common denominator limit-
ing their use and applicability as well as the information detail re-
sulting from such measurements.
The present document is an attempt to review the open literature perti-
nent to the problem of measuring particulates carried by stack efflu-
ents, with special emphasis on techniques and methods compatible with
continuous or continual, automated operation and electrically recorded
-------
information. This review is introduced by an overview of the litera-
ture sources available to perform such a review, both published books
as well as scientific and technical journals. The next section of this
document pertains to a selective summary of the general field of the
present state of aerosol analysis technology, again with specific atten-
tion to automated methods and techniques compatible with the measure-
ment of particulates in stack gases. The discussion then turns to the
reasons why, and the methods with which particle aerodynamic diameter
is measured. Techniques for the measurement of particle mass concen-
tration and the basis for selecting mass as a primary parameter for
the assessment of particulate emissions are treated. The document con-
cludes with a brief review of the particular characteristics of and
the problems associated with stack gas aerosol measurements.
The present document is not intended as an all-inclusive, totally ex-
haustive review of published literature on the subject but as a useful
compilation of work performed, techniques assessed, and definitions re-
viewed in the context of stack gas particulate emission characteriza-
tion. Furthermore, it is intended to serve as background source for
the selection of techniques to be incorporated in the design of a fine-
particulate measuring device, a development program within which the
present document has been compiled.
LITERATURE ON AEROSOLS
An extensive literature has grown concerning the behavior and analysis
of suspended particulates, as is indicated by the number of books
listed in Table 1, derived from Books in Print and from previous litera-
ture reviews. Articles on aerosols appear in a wide variety of period-
icals because aerosols play significant roles in diverse areas of con-
cern, exhibiting behavior that requires the expertise of many disci-
plines to understand. A list of the periodicals in which work relat-
ing to aerosols appears is given in Table 2, a selection from those
-------
Table 1. BOOKS CONCERNING THE BEHAVIOR OF AEROSOLS
AND THEIR ANALYSIS
Allen, T. (1968): Particle Size Measurement, Chapman & Hall, London.
American Conference of Governmental Industrial Hygienists (1972): Air
Sampling Instruments, 4th Ed.
Cadle, R. D. (1953): Particle Size Determination, Interscience, New
York.
Dalla Valle, J. M. (1948): Micromeritics, 2nd Ed., Pitman Publ.
Corp., New York.
Davies, C. N. (1967): Aerosol Science, Academic Press, New York.
Dennis, R., Ed. (in press): Handbook on Aerosols, U.S. Atomic Energy
Commission, Monograph Series.
Englund, H. (1971) : Proceedings of the Second International Clean
Air Conference, Academic Press, New York.
Fuchs, N. A. (1964): Mechanics of Aerosols, Pergamon Press, New York.
Fuchs, N. A. (1971): Collection of Aerosol Abstracts, Vol. 1,
Viniti, Moscow.
Fuchs, N. A. and Sutugin, A. G. (1971): Highly Dispersed Aerosols,
Pergamon Press, Oxford.
Green, H. L. and Lane, W. R. (1964) : Particulate Clouds; Dusts,
Smokes, and Mists, Nostrand Co., Princeton, N. J.
Herdan, G. (1960): Small Particle Statistics, Academis Press, New
York.
Hidy, G. M.: International Reviews in the Physics & Chemistry of
Aerosols, 3 Vols., Pergamon Press, Oxford.
Hidy, G. M. and Brock, J. R. (1971): The Dynamics of Aerocolloidal
Systems, Pergamon Press, Oxford.
International Atomic Energy Agency (1967): Assessment of Airborne
Radioactivity, Proceedings of Symposium, Vienna, 3-5 July 1967,
IAEA, Vienna.
Intersociety Committee for Ambient Air Sampling and Analysis, 4 Vols.,
American Public Health Association, New York.
-------
Table 1 (continued). BOOKS CONCERNING THE BEHAVIOR OF AEROSOLS
AND THEIR ANALYSIS
Irani, R. and Callis, C. F. (1963): Particle Size; Measurement,
Interpretation, and Application, Wiley, New York.
Kerker, M. (1969) : Scattering of Light and Other Electromagnetic
Radiation, Academic Press, New York.
Kuhn, W. E., et al., Eds. (1963): Ultrafine Particles, Wiley, New York
Lapple, C. E. (1956): Fluid and Particle Mechanics, Univ. of Delaware,
Newark, Del.
Ledbetter, J. (1971): Air Pollution, Vol. 1, Analysis, Marcell
Dekker, New York.
Leithe, W. (1970): Analysis of Air Pollutants, Ann Arbor Science
Publ., Ann Arbor, Mich.
Magill, P., et al., Eds. (1956): Air Pollution Handbook, McGraw-Hill,
New York.
Mancy, K. H. (1971) : Instrumental Analysis of Air Pollutants, Ann
Arbor Science Publ., Ann Arbor, Mich.
McCrone, W. C., et al. (1971): Particle Atlas, Ann Arbor Science
Publ., Ann Arbor, Mich.
Mednikov, E. P. (1965) : Acoustic Coagulation and Precipitation of
Aerosols, Plenum Press, New York.
Mercer, T. T., et al., Eds. (1971): Assessment of Airborne Particles:
Fundamentals, Applications & Implications to Inhalation Toxicology,
C. C. Thomas, Springfield, 111.
Orr, C. (1966): Particulate Technology, The Macmillan Company,
New York.
Orr, C. (1971): Filtration - Principles and Techniques, Marcel Dekker,
New York.
Pazar, C. (1970) : Air and Gas Clean-up Equipment, Noyes Data Corp.,
Park Ridge, New Jersey.
Richardson, E. G., Ed. (1960): Aerodynamic Capture of Particles,
Pergamon Press, New York.
10
-------
Table 1 (continued). BOOKS CONCERNING THE BEHAVIOR OF AEROSOLS
AND THEIR ANALYSIS
Rossano, A. T., Jr. and Cooper, H. B. (1970): Source Testing Handbook
for Air Pollution Control, Environ. Sci. Serv., Wilton, Conn.
Ruzer, L. S. (1968): Radioactive Aerosols, Bureau of Standards,
Moscow.
Sanders, P. A. (1970): Principles of Aerosol Technology, D. Van
Nostrand, Reinhold, New York.
Scorer, R. S. (1968): Air Pollution, Pergamon Press, New York.
Sheperd, H. R. (1960): Aerosols; Science and Technology, Wiley,
New York.
Silverman, L., Billings, C. E., and First, M. W. (1971): Particle
Size Measurement in Industrial Hygiene, U. S. Atomic Energy
Commission, Monograph Series, Academic Press, New York.
Sittig, M. (1968): Air Pollution Control, Noyes Data Corp., Park
Ridge, New Jersey.
Soo, S. L. (1967): Fluid Dynamics of Multi-Phase Media, Blaisdell,
Waltham, Mass.
Sproull, W. T. (1970): Air Pollution and Its Control, Exposition
Press, Jericho, New York.
Spurny, K. (1966) : Aerosols: Physical Chemistry and Applications,
Gordon and Breach, New York.
Starkman, E. S., Ed. (1971): Combustion-Generated Air Pollution,
Plenum Press, New York.
Stern, A. C., Ed. (1968): Air Pollution; A Comprehensive Treatise,
3 Vols., Academic Press, New York.
Strauss, W. (1966): Industrial Gas Cleaning, Pergamon Press, New York.
Strauss, W. (1970): Air Pollution Control, Wiley, New York.
White, H. J. (1963): Industrial Electrostatic Precipitation, Pergamon
Press, New York.
Zimon, A. D. (1969): Adhesion of Dust and Powder, Plenum Press,
New York.
11
-------
Table 2. PERIODICALS CONTAINING ARTICLES CONCERNING THE
BEHAVIOR OF AEROSOLS AND THEIR ANALYSIS
A.I.A.A Journal
Air Pollution Titles
Air Quality Control Digest
American Industrial Hygiene Association Journal
Annals of Occupational Hygiene
Applied Optics
Arch. Met. Geoph., Biokl., Ser. A.
Archives of Environmental Health
Atmospheric Environment
Beitra'ge zur Phyzik der Atmosphere
Colloid Journal of the U.S.S.R. (English Translation of Kolloid
Zhurnal)
Contamination Control
Environment
Environmental Science and Technology
Icarus
Industrial Engineering Chemistry Fundamentals
I.S.A. Transactions (Instrument Society of America)
Journal of Aerosol Science
Journal of the Air Pollution Control Association
Journal of Applied Meteorology
Journal of Applied Physics
Journal of the Atmospheric Sciences
Journal of Atmospheric and Terrestrial Physics
Journal of Colloid and Interface Science
Journal of Environmental Health
Journal of Environmental Sciences
Journal of Geophysical Research
Journal of the Meteorological Society of Japan
Journal of Meteorology
Journal of Occupational Medicine
Journal of Physics D - Applied Physics
Journal of deRecherches Atmospheriques
Journal of Scientific Instruments
Journal of the Royal Meteorological Society
Kogai to Taisaku (Journal of Pollution Control)
Kolloidhyi Zhurnal (Colloid Journal)
Kolloid-Zeitschrift and Zeitschrift Fuer Polymere
Kuki Seijo (Journal of Japan Air Cleaning Association)
Pollution- Abstracts
Pollution Atmospherique
Pollution Engineering
Powder Technology
12
-------
Table 2 (continued). PERIODICALS CONTAINING ARTICLES CONCERNING
THE BEHAVIOR OF AEROSOLS AND THEIR ANALYSIS
Proceedings of the Royal Irish Academy
Quarterly Journal of the Royal Meteorological Society
Review of Scientific Instruments
Smokeless Air
Staub
tellus
13
-------
used as the basis of Air Pollution Abstracts by the Air Pollution
Technical Information Center (APTIC) at Research Triangle Park, North
Carolina. The number and diversity of sources of information on aero-
sols, combined with the rapidity in growth of the field, mean that any
review is probably incomplete, although reviewers with experience in
this field are aware of this and take pains to overcome these diffi-
culties.
Hodkinson (1967)53 listed and briefly discussed about 40 books and con-
ference proceedings on aerosol fundamentals, aerosol measurement, and
aerosols in the atmosphere; most of his list is included in the one
given here. In the recent book by Strauss (1970),90 H. M. Englund re-
viewed "the literature of air pollution," giving a paragraph of descrip-
tion on each of a dozen books in a section on the analysis of gases
and particles, and a valuable description of periodicals pertaining to
air pollution. Another extensive listing of periodicals in the field
is available in Green and Lane (1964).^9
These books and periodicals form the major part of the literature re-
sources of those working on problems relating to aerosols. Some, by
no means all, of these sources have been used in the preparation of the
material which follows on methods and techniques for the measurement of
stack gas particulates.
REVIEW OF METHODS FOR THE MEASUREMENT OF AIRBORNE PARTICULATES
Aerosols and Their Definition
Aerosols, as defined by Fuchs (1964)^ and as meant by workers in aerosol
science and technology, are "disperse systems with a gas-phase medium
and a solid or liquid disperse phase." Aerosols are gasborne solids or
droplets whose size range is approximately 10 cm to 10 cm (Fuchs,
1964),41 but most of the interest in this field and most of the research
14
-------
-6 -2 -5
done is on particles of sizes from 10 to 10 cm or possibly 10 to
-3
10 cm. The size unit frequently used in aerosol science is the mi-
crometer (ym, 10 m, 10 cm) and the sizes of primary interest are
from tenths (and recently hundredths) of micrometers to tens of microm-
eters. Figure 1 lists many aerosols with their size ranges, from which
it becomes apparent that most, by no means all, aerosols are character-
ized by sizes of the order of 1 micrometer. Those particles that pene-
trate to human lungs are the fraction of the inhaled aerosol that is
below a few micrometers in diameter (Lippman, 1970)°^ and the mass de-
posited there per particle is greatest for the largest size that do
penetrate, two reasons for the importance of studying particle sizes on
the order of 1 micrometer.
Because liquids and solids evaporate or sublimate and condense depend-
ing upon the ambient vapor pressure of their constituents in comparison
with the particulate equilibrium vapor pressure (which is a function of
temperature, pressure, particle size, and particle charge), molecular
species may be in the gaseous or particulate state depending upon the
conditions under which they are observed.
The definition of "particulate matter" has been the subject of consider-
able controversy (Crandall, 1971),30 but in any case particle size distri-
butions should be measured under conditions as closely as possible re-
sembling the conditions under which the information is to be applied.
Under other conditions, the particles may be larger or smaller or may
have changed into the vapor phase. If behavior in the atmosphere of
stack emissions is the concern, then these emissions should be measured
in the atmosphere, under the appropriate conditions of temperature,
pressure, residence time, and dilution, if such can be specified. If
behavior in the stack is the concern, then the particle size distribu-
tions should be measured under the conditions existing in the stack,
especially if the retention effectiveness of control methods is to be
evaluated as a function of particle size.
15
-------
H
J001
H.ONH
O,/
N,CO,
1
Diame
-
Tobacco
Mosaic
Virus •
er ol Gas Mo
CONV1NTIONS
Rjnfe ol Sim
Small Rtnge- Average
»
Ooubtlul Values
T
0.0005 0.001
Not
Toba
Necrc
Virus
3
ecule:
nal Impuritie
Zinc Ox
"CO
sis Vires &
Protein
p —
rbon Black
Tobac
__ Majinesiui
Silver Iodide
Combustion
• In
Me
Am
deFi
:oSm
Aerosol?
uiet Outdoor
lallurcical Du
mnium Chlo
Alk
mes
*e .
Ail Smoke
i Oxide Smoke
1 •'
Rosin Smoke
Nuclei
Sea Sa
(Enamels)
I Nuclei
PAR
»ir
land
Fumes
Smeller Duit
ideft
uliFu
1
Pigi
RE
TICLE
mes
Fog
LFurt
a
Foundry Oust
Flour Mill Cust
nes Sprayed Zinc Dust
,S
Sulluric Acid
Condensed
Zinc Dust
r~
isecticide Dus
Bacteria
nenls (Fla
Spray Dried
FERENCE j
SIZES )
I
SIZE (.Micro
~l
ulphide Ore, Pulps (
list
Ce
s '
«),
Milk
icrcei
«)
ment Dust
Pulv
lanl Sporej
Pollens
Si
Fly Ash
Human Ha
....'.
.tahtS
Gro
irflo
irized
cues
iand
rOiar
> 200
«sl
ind Limestoni
otion
Coal
'ailings
Washed
neter
Visible
R
Founi
oE»«
100 65 48 351 28 11
n Drops
rySand
0.005 0.01 0.05 0.1 0.5 1 5 10 50 100 500 1,000 5,000 10,000
Figure 1. The size of airborne contaminants
(Mine Safety Appliances Company)
16
-------
Aerosols and Their Size
The behavior of aerosols is strongly dependent on size. Different par-
ticle sizes will characteristically be produced by various aerosol gen-
eration mechanisms; the smallest particles come from chemical reactions
such as combustion and from condensation of vapors into droplets or
solids (mists or "fumes"); somewhat larger particles come from the atom-
ization (breaking up) of liquids by rapidly flowing gases, such as the
atomization of ocean water into sprays which then dry to produce NaCl
aerosol particles; and the largest particles generally come from the
solid-solid interaction of grinding. If these particles are produced
by man-made causes, such as by industrial processes, the type of control
equipment suitable for removing them depends, in a large part, on their
size distribution. (See Table 3.) The finest particles are most effi-
ciently removed using large baghouses containing fabric filters; the
coarser aerosols are removed by electrostatic precipitators, spray scrub-
bers, centrifugal collectors, and settling chambers, sometimes used in
combination. Knowledge of particle size is necessary to predict not
only behavior in emission control devices but also behavior in the
atmosphere; all of the following aspects of particle behavior are size-
dependent: dispersion through turbulent mixing, sedimentation due to
gravity, impaction on objects due to flow curvature, nucleation of rain,
scavenging by rain, reaction with gases, coagulation with other par-
ticles, reduction of visibility, electrical charging and discharging.
Particle behavior with respect to deposition in the lungs of man and
animals is also size-dependent: capture by the upper respiratory tract
and deposition due to sedimentation, diffusion, impaction, and electro-
static attraction are all influenced by particle size. To illustrate
this role of size, Figure 2 gives the contribution to visibility degrada-
tion as a function of particle size for a particle size distribution sim-
ilar to that found in the atmosphere, as computed by Pueschel and Noll
(1967).81 Figure 3 shows the penetration versus size curve adopted as a
standard model of the human lung for "respirable" sampling (Lippman, 1970).65
17
-------
-6
10 r-
- t
b(r) cm"
IQ
-8
10
bTOT • 7.6-10" cm
i.3-io"rcm*'
0
Ql
0.2
0.3
0.4
0.5 0.6
0.7
Figure 2. Product of light-scattering cross-section and measured con-
centration for atmospheric aerosol. From Pueschel and Noll
(1967).81 Visibility inversely proportional to b.
18
-------
z
g
o
u
DC
QC
_
O
UJ
O
•z
<
cr
u
z
.UJ
Q.
Z 4 6 8 10
DIAMETER UNIT DENSITY SPHERE, (microns)
Figure 3. "Respirable" dust mass measurement sampling criteria.
(Lippman, 1970)65
19
-------
Table 3.
EFFICIENCY OF DUST-ARRESTING EQUIPMENT FOR
VARIOUS SIZES OF DUST (BALL AND GRIFFITHS, 1970)18
Inertial collector
Medium efficiency cyclone
High-efficiency cyclone
Tubular cyclones
Self-induced spray deduster
Spray tower
Venturi scrubber (medium
energy)
Venturi scrubber (high
energy)
Fabric filter (shaker type)
Elect ro-precipitator
Irrigated electro-
precipitator
Percentage efficiency at
50 ym
95
94
96
100
100
99
100
100
99
99
99
5 ym
16
27
73
89
94
94
99
99
99
99
98
1 ym
3
8
27
40
48
55
97
98
99
86
92
Sampling Considerations
The analysis of an aerosol, such as the analysis of particles in stack
gases, usually involves the determination of certain properties charac-
teristic of the individual particles plus the determination of the con-
centration, in suitable units, of the aerosol. Because the analysis is
to be used to predict the behavior of the aerosol, the particle proper-
ties measured and the concentration determined will depend on what as-
pect of their behavior is considered important. Important particle
characteristics are: size, shape, composition (e.g., density, refrac-
tive index, chemical constituents), phase (liquid or solid), and elec-
trical charge. The characterization of an aerosol is frequently based
on a total concentration measurement (in units per unit volume such as
number, optical cross-sectional area, or mass) as well as a detailed
20
-------
definition of the particle characteristic of greatest importance, such as
particle size. If changes with time in particle or aerosol cloud char-
acteristics are significant for the application for which the analysis is
made, then the analysis technique should be able to provide such informa-
tion. Real-time aerosol assessment presents a challenging problem to the
instrumentation scientist, and no fully satisfactory and generalized in-
strumental solution is available at present.
The most important information about a measuring instrument or technique
is the parameter it purports to measure. Other important characteristics
include the measurement principle, range, resolution, reproducibility, and
accuracy and, finally, cost (materials, effort). A minimum requirement of
an aerosol analysis method is that it provide the concentration in some
units (number per volume, mass per volume); additional information is
obtained if the method provides the concentration as a function of par-
ticle size. Even better characterization is obtained if the parameters
are given as a function of time. Selecting an aerosol analysis instrument
requires primarily the selection of what concentration parameter is of
significance and what kind of size is to be measured.
A number of classifications of aerosol assessment devices have been made.
Friedlander (1971)^ classified a variety of particle analysis methods,
comparing their size resolution, time resolution, chemical composition
determination, and concentration parameter measured. Fuchs (1968)^ clas-
sified aerosol analysis methods into those which required capturing the
particles before analysis and those which made the analysis "in-situ;" he
further subdivided the methods on the basis of the physical methods used
to precipitate the particles (if precipitated), the physical property used
as an indicator of size, the manner in which this physical property was mea-
sured, and the property used as the basis of concentration; these last three
were implicit subdivisions under his general headings of "microscopic,"
"macroscopic," and "other" methods of "in-situ" analysis. Lapple (1968)60
classified aerosol analysis schemes on the basis of the property of the
particles which was used to discriminate by size, the method by which this
21
-------
property was used (for instance, the optical cross-section of a particle
might be sensed by extinction or scattering methods), and variations of the
technique used to discriminate this property; he provided examples of equip-
ment (by manufacturer's name) and the range of particle diameters mea-
surable by the method. Lieberman (1968)62 also made a preliminary separation
of analysis methods into those which required collection and those which
measured "in-situ," and then further subdivided (not very systematically)
on the basis of the physical principles used either to size-classify or to
measure concentration. Davies (1970)^ noted 201 recent studies on par-
ticle size analysis, listed by application.
In the present document, particle measurement techniques are classified on
the basis of the measured particle size parameter and/or particle concentra-
tion parameters.
Aerosol assessment systems which measure particle size and concentration
can, and have, been built around a great variety of principles of particle
sizing and sensing. They can be generalized, however, as shown in Figure 4.
A sample must be selected (minimizing bias) and transported (minimizing
changes in concentration or size distribution); the size distribution may
then be determined either by sorting the particles on the basis of their
size and then sensing the fractions in the various size intervals into
which they have been sorted, or by sensing the particles individually and
sorting the signals from the sensor into intervals representing size inter-
vals, totalling the signals in each size interval. Finally, the results of
the size classification and concentration measurement must be indicated,
preferably in a convenient recordable format.
Aerosols, such as particulates in stack gases, generally change with time
and are rarely homogeneous. Samples taken that have small volumes compared
to the volume over which significant changes take place are called "point"
samples. Samples whose duration is short compared to the time over which
significant changes take place are called "instantaneous" samples. Those
samples which do not meet these spatial or temporal criteria are called
22
-------
SAMPLE
TRANSPORT
SORT
PARTICLES
SENSE
SORT
SIGNALS
1
INDICATE
Figure 4. Generalized aerosol assessment system schematic
23
-------
"integrated" samples'. Integrated samples provide less detail, but the
integration they perform is often useful. What is generally unacceptable
is sampling bias; i.e., the preferential selection of particles of certain
characteristics. A major source of sampling bias in aerosol assessment
comes from anisokinetic sampling, the condition in which the velocity of
the gas at the inlet of the sampling probe does not match the velocity of
the gas in the gas,stream being sampled. There have been some attempts to
provide correction factors for anisokinetic samples (Davies, 1968;31 Sehmel,
1972),^ but isokinetic sampling is to be performed where possible.
Once the sampling process has occurred, this fraction of the total flow
must be transported to the sorting/sensing region of the aerosol assess-
ment device. A distinction is widely made between "in-situ" and extraction
methods of measurement, the latter involving the transport of the sample
away from the environment of its source. Before comparing the "in-situ"
method with the extraction method, it seems advisable to clarify the con-
cept of "in-situ" sensing. Generally, "in-situ" measurements (in the context
of stack testing) are those made without transporting the sample from its
original environment. In-stack transmissometers are a widely-accepted
example. The desirability of "in-situ" measurements is due to the as-
sumption that they do not alter the environment within which the sample is
studied. Of course, even a transmissometer makes some change in the aerosol
environment, the presence of light where there normally is little. A
question arises whether a measurement made with a device immersed in the
gas flow but which draws the sample through short conduits rather than al-
lowing it to flow with the source stream is an "in-situ" measurement. We
will use that term for such a measurement system, however, recognizing this
possible objection. A practical definition of "in-situ" recognizes the
root of the word ("in the place") and the reason for its desirability: the
negligible change in the environment to which the sample is exposed. Thus,
the transmissometer and any device that senses within the stack (rather
than outside the stack) can be considered "in-situ" methods of measurement.
24
-------
Most stack sampling devices used for routine measurements are based on the
sample extraction approach; i.e., an isokinetic inlet nozzle followed by a
duct, commonly called a "probe" whose length is determined by the cross-
sectional dimensions of the stack at the point of sampling. The particle-
laden gas flows through this probe, and the particulate matter is col-
lected on filters or other devices on the outside of the stack wall, through
which the probe passes. This procedure is specified for mass emission sam-
pling by the Environmental Protection Agency. This method has several draw-
backs that are, although acceptable for manual operation, not compatible
with the meaningful operation of an automated device, and in particular for
a particle sizing instrument. These problem areas are: (1) particle losses
along the probe or, extraction duct, and (2) condensation within the ex-
terior section of the probe and at the collector exposed to lower tempera-
tures than the stack environment (or conversely the need for heating of
these elements, not only above the water dewpoint, but also above the acid
mist condensation temperature).
The first problem (i.e., particulate losses within the probe) is caused by
the following mechanisms:
1. impaction in tube bends
2. sedimentation due to gravity
3. electrostatic scattering due to mutual repulsion by
charged particles
4. turbulent deposition
5. electrostatic precipitation due to image charges
6. diffusional loss to the tube wall
7. thermal precipitation due to temperature difference
between gas and tube wall.
All of these loss mechanisms are dependent on particle size, thus changing
not only the total concentration but also the shape of the size distribution.
Such losses in stack tests have been shown to occur (Bird, et al., 1973).21
Further changes in the size distribution can come from:
25
-------
8. coagulation, enhanced by shear in the gas velocity
profile and by vapor condensation
9. re-entrainment
10. condensation.
The approximate magnitude of such influences can be estimated from formulas
available in the literature (Fuchs, 1964;1*1 Green and Lane, 1964).^9 In any
case these losses are unacceptable for an automated-recording type in-
strument for the obvious reason that the fraction deposited in the probe
is not incorporated in the measurement. The EPA specified procedure for
manual sampling includes the requirement for probe washout and addition of
the mass of particulates retained in the probe to the collected mass on
the filter. Such a procedure must be ruled out for an automated measure-
ment system. No practical method has yet been found to eliminate the depo-
sition of particulates in probes for size distributions with mass median
diameters of 10 to 20 ym and standard deviations between 2 and 5, typical
of stack environments, and for probe lengths of the order of, or in excess of
4 meters required to sample representatively a stack with a flow duct area
of up to about 50 m (approximately 500 square feet). The only method by
which dust deposition can be reduced, albeit not eliminated, is by in-
creasing the flow velocity to very high values (of the order of 20 m/s or
more) to produce re-entrainment of particles collected on the probe walls.
This procedure may be marginally acceptable for total mass measurements,
but not for accurate sizing purposes since particle agglomeration and
clustering can be expected in any situation where re-entrainment plays a
significant role.
No line losses are acceptable unless they constitute a known fraction of
the inlet loading, a highly unlikely condition indeed. Obviously the con-
clusion is that since the probe length cannot be minimized if the
collection-sensing device is exterior to the stack, the only rational
solution is to place the device within the stack duct to be sampled. This
method also resolves in large measure a second problem: the condensation
of water and acid mist at the measurement site. By operating the sampler
26
-------
wholly at the source temperature, condensation can be precluded. At
typical stack gas temperatures of 150°C to 260°C (300°F to 500°F) all the
water is in the vapor phase and if the collection-detection head is ex-
posed to the same environment, no condensation will occur.
The measurement of particulate mass and especially of the size distribution
by mass of stack emissions can only be determined meaningfully at two loca-
tions: (1) within the stack itself, or (2) downstream from the stack exit
(stack plume). The first of these approaches can provide information on
the particulate emissions as they are generated, and within the high tem-
perature environment of the stack itself. The second approach would pro-
vide a measure of the particulates as they will eventually disperse in the
atmosphere. Both of these approaches have their merit.
Sample extraction, by lengthy probes and requiring thermal control, is an
approach that we believe is an unacceptable compromise between the two al-
ternatives presented above, and it is especially incompatible with repre-
sentative size measurements.
Method of Particle Assessment
Once a sample of the aerosol has been withdrawn and transported, the
aerosol assessment system obtains a concentration measurement by sensing
the particles and may obtain more information by sorting the particles
according to some property before sensing them or by sorting the signals
produced in sensing them. A summary of methods for sorting particles is
contained in Tables 4 and 5. A summary of methods for sensing particles is
in Table 6. There is little overlap between physical mechanisms used to
sort particles and those used to sense them. In sorting particles, it is
desirable to segregate the particles on the basis of some characteristic,
whereas in sensing them it is desirable to leave them unaffected by the
sensing process. Sorting the particles involves the danger of changing
them and thus biasing the sensing of the particles, but sorting them is
27
-------
necessary when the desired physical basis for size discrimination (e.g.,
aerodynamic diameter) is different from the property used to obtain con-
centration (e.g., mass).
Mechanisms for size classifying particles, independently of sensing them,
are shown in Tables 4 and 5, along with examples of the use of these
mechanisms and the "size" given by such techniques. Table 4 is adapted
from Lapple (1968).60
Most of the techniques for particle size discrimination which have emerged
since Lapple's review are variations on themes presented there. Some
significant new developments are presented in Table 5, a supplement to
Table 4.
An aerosol assessment system must have some way of sensing the particles,
either as individuals or as an aggregate. Three kinds of general
approaches for measuring particle size distributions can be identified:
1. Those which use one property of the particles to sort them into
size intervals and another property to measure their concen-
tration.
2. Those which measure only one property of the particles and
either:
a. measure this property particle-by-particle and sort
the signals from these measurements to get a size
distribution as well as a concentration, or
b. measure the aggregate of the property for a number
of particles to get an average concentration.
These possibilities are shown schematically in Figure 5. Which method
is chosen will depend on what information is desired and upon the re-
lative difficulty of sorting particles or sorting signals to get size
classification.
28
-------
Table 4. SUMMARY OF TYPES OF SIZE-ANALYSIS TECHNIQUES
K>
VO
Size-discriminating property
Type
Geometric
Mechanical
or Dynamic
(in fluids)
-
Character of
mechanism
Physical barrier
Inertia
Terminal settling
velocity (in
gravity or centri-
fugal fields)
(possibilities of
using electrical
fields exist)
Diffusion
Technique and variations
Sieving
Wet
Dry
Ultrafiltration
Impaction on surface
Pressure pulse (sonic)
Elutriation
Sedimentation
Single
fractionation
Series
fractionation
Layer
methods
(differential;
Suspension
methods
Liquid
Gas
Liquid
Gas
Liquid
Gas
Differential-
liquid
Integral-
liquid
Integral-gas
Example of common
apparatus3
Ro-Top (Tyler) , Pulverit
Shaker, Sonic- Sifter
(Allen-Bradley) ; electro-
formed sieves (Buckbee
Hears)
Membranes, Millipore,
Gelman
Greenburg-Smith; May,
Anderson; Cassella;
Battelle
IITRI
Schone
Roller (G); Federal
Pneumatic (C) ; Bahco (C)
Andrews
Houltain "Infrasizer";
Gonell; Aminco
Palo Travis (G) ; Werner
(G), MSA-Whitby (G, C) ;
Kaye (C)
Micromerigraph (Sharpies)
(C), Conifuge (C)
Pipette (Andreasen) (G) ;
hydrometer (G) ; diver
(G) ; Komack (C) ; turbi-
dimeter (G, C) ; differen-
tial manometer; gamma-
ray absorption
Sedimentation balance
(Oden, Cohn, Sartorius)
(G) ; manometer (Weigner)
(G); pendulum (G)
Aerosol spectrometer
(Goetz) (C); sedimenta-
tion channel (G)
Decantation (in liquid only) Cunnings
Photo-sedinentation (photography of particle streak) Carey & Stainnand
Particle di»placeoent (randoa walk) Millkon cell
Particle deposition Diffusion battery
Range of
applicable particle
diameter,8
microns
5-1000
0.01-5
0.1-100
10-1000(7)
5-100
1-100 (G)
0.02-10(C)
0. 002-1 (UC)
1-100(€)
1-100{G)
0.01-1
-------
Table 4 (continued). SUMMARY OF TYPES OF SIZE-ANALYSIS TECHNIQUES
u>
o
Size-discriminating property
Type
Optical
Electrical
Character of
mechanism
Imaging1"
Transmission
(spectral)
. Scattering
Diffraction
Resistance
Capacitance
Charge
Technique and variations
Light microscopy
Ultramicroscopy (gives mean size only)
Electron microscopy
Extinction measured as function of wavelength
Single particle
count
Right angle (90°)
Angular
Forward
Polarization
Macroscopic (gives mean size only)
Light
X-ray
Laser (holography-reconstruction of diffraction
pattern)
Alteration of current flow by particles
Potential pulse due to particle deposition
Tribo-charging
Induction charging
Corona charging
Example of common
apparatus3
American Optical; Bausch
& Lomb; Leitz; Nikon
Reichert; Tiyodo;
Unitron; Vickers; Wild;
Zeiss, etc. Flying-Spot
counters
Hitachi; Norelco; RCA;
Siemens
Bailery
Royco counter (gas,
liquid); O'Konski
Owl (monodisperse
aerosols)
Sinclair-Phoenix (aerosol
concentration)
Stat Volt Co.
Coulter count
Drozln and LaHer; Mercer
Range of
applicable particle
diameter ,a
microns
0.2-100
0.005-1
0.002-15
0.1-2(?)
0.2-50
(higher with
microwave)
1-100
-------
Table 4 (continued). SUMMARY OF TYPES OF SIZE-ANALYSIS TECHNIQUES
Size-discriminating property
Type
Magnetic
Thermal
Physico-
chemical
Character of
mechanism
Applicable to
magnetic
materials only
Particle deposition
Condensation
Technique
Particle migration in
pulses
Particle migration in
Growth of nuclei with
and variations
magnetic field; magnetic
thermal gradient
controlled supersaturation
Example of common
apparatus3
Cassella;
GE condensation nuclei
counter
Range of
applicable particle
diameter,8
microns
0.1-10(7)
0.01-0.1(7)
Items in parenthesis have following significance: C = in centrifugal field; G = in gravity field; DC = ultracentrifuge.
Replicas may be used in place of particles that might evaporate or be destroyed during measurement; e.g., in electron microscopy to
avoid the effect of vacuum or electron beams, and in magnesium oxide film method for measuring size of drops.
-------
Table 5. RECENTLY DEVELOPED PARTICLE SIZING TECHNIQUES
Size discriminating
property
Dynamic
OJ
Electromagnetic
Mechanism
employed
Inertia
Terminal
velocity
Diffusion
Transmission/
extinction
Scattering
Technique
Impaction into stag-
nation zone without
obstacle
Sound pulse from
impact
Impaction on rotat-
ing rod
Hindered settling
versus pressure
Electrical mobility
Laser Doppler shift
Quenching of laser
by particles in
cavity
Single particle ex-
tinction measure-
ments
Forward scattering
of x-rays by single
particles
Example
Cascade centripeter
(Hounam and Sherwood,
1965)
Microdynamometer
(Benarie and Quetier,
1970)
Roto-rod sampler
Giant particle
sampler (Noll, 1970)
(Turner and Fayed,
1970)
Mobility analyzer
(Whitby and Clark,
1968)
(Hinds and Reist,
1973)
(Schleusener, 1968)
(Knollenberg, 1971)
(Brusset and Donati,
1969)
-------
Table 5 (continued). RECENTLY DEVELOPED PARTICLE SIZING TECHNIQUES
u>
Size discriminating
property
Electrical
Thermal
Physico-chemical
Mechanism
employed
Charge
Evaporative
cooling
Combustion
Technique
Charge measurements
on particles charged
by impact
Hot-wire anemometry
Measurement of light
from single particle
combustion
Example
(Abbott, Dye, Sartor,
1972)
(Adams and Smith,
1971)
(Keiley, 1960)
(Goldschmidt, 1965)
Scintillation
counter
(Binek and Dohnalova,
1967)
-------
SORT
PARTICLES
1
SENSE
SENSE
1
SENSE
SORT
SIGNALS
INDICATE
INDICATE
INDICATE
Figure 5. Three types of aerosol assessment systems in general use
Analysis of the signals from particle sensors generally produces as data
one of the following:
1. A total concentration obtained by summing the signals.
2. A number concentration based on counting the number of
signal pulses.
3. A size distribution obtained by subdividing the signal
pulses into size intervals (usually on the basis of
amplitude) and counting the number (or totalling the
"quantity") contained in each size interval.
The processing of electrical signals is easier and thus has received much
more attention than the sorting of particles, and the techniques for the
former are more sophisticated and less prone to error than the latter.
Thus, it would be advantageous to sort signals, rather than sort par-
ticles, if the signals from the sensor give the size information desired.
Unfortunately, the sensor signals often are not unequivocably related to
particle size. The scattering of light by particles can be converted
34
-------
into a convenient electrical signal by suitable optics and a photo-
multiplier tube, but the signal amplitude is not an unambiguous function
of size, for example. Generally, the "size" of interest is related to
the motion of the particle in the gas (aerodynamic diameter for sedimen-
tation and impaction and other motion in an acceleration field, mobility
for diffusion and electric migration). In general, however, particle
dynamics do not provide a direct sensing mechanism. (One exception to
this is the Doppler shift method of measuring the mobility of diffusing,
laser-illuminated particles (Hinds and Reist, 1973).)51 To obtain distri-
butions based on a size related to motion, it is generally necessary to
sort the particles dynamically before sensing them, using size discrim-
inating methods such as sedimentation elutriators, aerosol centrifuges,
diffusion batteries, impactors, and electrical mobility analyzers.
Table 6 lists a number of concentration sensing techniques and the quan-
tities (concentration units) they yield.
In their brief review of "particulate emission and process monitors,"
Dorsey and Burckle (1971) summarized much information about particle
sensors in a tabular form, and Table 7 is taken from that reference.
The particle properties are self-explanatory. The sensing techniques
are mass determination by gravimetric techniques, light scattering or
absorption in a variety of realizations, gamma ray absorption (Holze and
Demmrich, 1969),55 beta absorption, electrical charge transfer from
charged particles to a grounded collecting surface, and the absorption
of acoustical energy. This table indicates why certain methods cannot
yield true mass concentrations, if such are desired.
Much more could be written about the various ways of measuring particle
size and sensing concentration, including the possible combinations of
these two to produce a variety of instruments for aerosol assessment.
Rather than attempt an encyclopedic exposition of these possible mea-
surement systems, the authors now turn to answering the following
35
-------
Table 6. VARIOUS PARAMETERS OF AEROSOL CONCENTRATION AND TECHNIQUES
BY WHICH THEY MAY.BE.DETERMINED .
Concentration
Available techniques
Number
Diameter
Area
Settling
velocity
Volume
Optical
activity
Mass:
A. Total mass,
composition
unknown
B. Mass, com-
position
known
Counting of images... optical microscopy, electron
midroscopy, holography, etc.
Counting of electrical pulses from photomultipliers
sensing light scattering from single particles
Counting of sonic pulses
Counting of electrical pulses from collection of
charged particles
Counting of scintillation pulses produced by combus-
tion of particles
Coulter counting
Counting of hot-wire anemometry pulses
Measurement of total scattered light from condensa-
tion of nuclei
Small ion deposition on particles (see Duwel, 1968)
Absorption of gas on particle surfaces
Contact electrification
Pressure drop in dust-measuring nozzle (see Duwel,
1968)
Coulter counting (electrical resistance)
Transmissometer
Nephelometer
Total scattering meters
LIDAR
Opacity matching
Weighing
Beta absorption
Mechanical resonance
Impact momentum - energy sensing
Coulter counting, integrated
Gas discharge obstruction counting, integrated
Gamma absorption
36
-------
Table 6 (continued). VARIOUS PARAMETERS OF AEROSOL CONCENTRATION AND
TECHNIQUES BY WHICH THEY MAY BE DETERMINED
Concentration
Available techniques
Other
Chemical methods: conductivity, colorimetry, calori-
metry, fluorimetry, atomic absorp-
tion, neutron activation analysis,
etc.
Scintillation pulse height, integrated
Acoustic absorption
Electrical charge transfer:
Ionic charging
Impact momentum charging
Tribo-electric charging
Filter resistance (see Duwel, 1968)37
37
-------
Table 7. SENSOR PRINCIPLES AFFECTED BY PARTICLE PROPERTIES3
(DORSEY AND BURCKLE, 1971)
Particle Property
Size
Shape
Density
Color
Resistivity
Composition
Mass
0
0
0
0
0
0
Light
+
+
+
+
0
0
GcHHIHcl
0
0
0
0
0
+
Beta
0
0
0
0
0
0
Electrical
+
+
+
0
+
0
Acoustical
+
+
+
0
+
0
Particle properties that may have a significant effect on measure-
ment accuracy; (+) indicates a major influence is to be expected,
while (0) is indicative of little or no effect.
questions as guides to the method or methods which will yield the most
promising solution for source aerosol analysis;
Which of the various possible size parameters is of most importance?
How should it be measured? and
Which of the various concentration measurements is desired?
should it be measured?
How
SIZING ON THE BASIS OF AERODYNAMIC DIAMETER
General Considerations
Based on the review on aerosol analysis given above, Table 8 lists those
size parameters that can be discriminated by present day aerosol analysis
equipment.
The linear dimension of a particle would be important as to how it
behaved with regard to sieving and with regard to the mechanism of inter-
ception, sometimes important in filtration.
The mobility, a parameter related to aerodynamic size, is important in
the prediction of particle behavior in the presence of electric fields
and in predicting particle diffusion.
38
-------
Table 8. METHODS OF PARTICLE SIZE DISCRIMINATION
Size parameter
Some appropriate methods of discrimination
Linear dimension
Mobility equivalent
diameter
Aerodynamic
equivalent
diameter
Optical equivalent
diameter
Volume
Other
Sieving
Microscopy
Diffusion battery
Electrostatic mobility analyzer
Laser doppler shift
Millikan cell
Impactor
Aerosol centrifuge
Sedimentation cell
Elutriator
Photo-counter
Single-particle transmissometer
Coulter counter
Gas discharge counter
Contact or impact charging
Sonic analysis of impact
Deposition in thermal field
Scintillation analysis
Evaporative cooling of hot-wire anemometer
Note: These methods either perform the size selection before
sensing or analyze the signals from the sensor, as
discussed before.
39
-------
The aerodynamic diameter indicates particle behavior under the force of
gravity and under inertial acceleration in fluid flow fields, such as
acceleration caused by curvature of the flow; the mechanisms of impac-
tion and sedimentation and, indirectly, diffusion, depend on the aero-
dynamic diameter and these mechanisms are quite important in filtration
and in lung deposition. Furthermore, the behavior of particles in such
control devices as cyclones, Venturi scrubbers, electrostatic precipita-
tors, filters, and settling chambers also depends directly or indirectly
upon aerodynamic diameter. The aerodynamic diameter comes into play in
the dispersion of particles in the atmosphere, controlling the sedimen-
tation of particles, their capture by obstacles in a wind, and their
scavenging by rain.
The optical equivalent diameter (or light scattering cross-section) is
significant in the effect of particles on visibility.
Other measures of particle size have special contexts that are less
directly related to particle behavior and are of interest for special
and more restrictive particle studies.
In sampling stationary sources, the goals are the design and evaluation
of control devices or the determination that they are not necessary.
These devices are being installed primarily to see that humans, other
animals, and plants are not injured by the effluent. Of the size
parameters listed above, the mobility and the aerodynamic diameter are
the two most important to the questions of control device effectiveness
and aerosol transport and deposition (on plants, in lungs). Although
a good case can be made for the importance of particle mobility (which
is indirectly related to aerodynamic diameter), a generally better case
can be made for the aerodynamic diameter as the most useful size
parameter in this context, especially as this parameter has been chosen
to be the basis of the measurements for "respirable dust" (Lippman, 1970).65
40
-------
The aerodynamic diameter of a particle of arbitrary shape is that
diameter which would give the same terminal settling velocity for a
sphere of unit density.
Mathematically ,
Vs = [
po
where v = terminal velocity
p = unit density
D = aerodynamic diameter
a
C(D ) = Cunningham correction factor
Si
g = gravitational acceleration
Vi = coefficient of viscosity of the gas.
The term in brackets, the "relaxation time" of the particle, governs the
behavior of particles falling due to gravity (sedimentation) or crossing
the streamlines of a gas to strike an obstacle deflecting the gas
(impact ion) .
To reiterate, the aerodynamic diameter is significant in particle response
in a number of control device techniques, in particle transport and re-
sidence time in the atmosphere, in deposition on plants, and in deposition
in lungs.
Selection of Method to Measure Aerodynamic Diameters
Aerosol elutriators and centrifuges separate particles on the basis of
aerodynamic diameter (equivalently, sedimentation velocity) by passing
them through a channel at right angles to a field which produces a velo-
city toward the walls of the channel. In elutriators, the channel is
generally formed by, parallel plates having relatively narrow spacing,
and the field is that of gravity. In aerosol centrifuges, the channels
41
-------
are various: the annular region between rotating cylinders, a spiraling
duct on the surface of a rotating cone, a spiraling duct on a rotating
disk, etc.; the field being the result of centrifugation. To analyze
the deposit in elutriators or centrifuges requires a certain amount of
disassembly, generally. Furthermore, even if the device is operated
at a constant flow rate (and constant rotation for the centrifuges),
the particle deposit is a complicated function of position and par-
ticle aerodynamic diameter (StSber and Zessack, 1964). ^ The flow rate
(or rotation rate) must be varied if it is desired to get a particle
size distribution using a single sensor located somewhere along the
deposit, and this data may require considerable analysis before yield-
ing a particle size distribution. Aerosol centrifuges (Goetz, 1957 J1*6
Stober and Flachsbart, 1969)^® have been found to have significant
secondary flows (Hochrainer, 1972),52 further complicating the analysis
problem, and attempts to correct for the secondary flow (Gerber, 1971)1*5
have been found inadequate so far (Stober and Boose, 1973).^7
The deposition in centrifuges and elutriators is a function of both
particle size and position, making the measurement of what is deposited
an unattractive method for particle sizing. Instead of deposition,
penetration might be measured to give size information, and one could
design a device consisting of a series ("cascade") of centrifugal
elements or elutriators, the penetrations of which could give particle
size data. As we show next, such a device would have size separation
characteristics inferior to those of impactors.
Consider the generalized elutriator or aerosol centrifuge geometry
depicted in Figure 6.
42
-------
i
V
s -»
u -»
g
< L >
Figure 6. Generalized elutriator/centrifugator
For parallel plates spaced h apart and having length L, the fraction e
of aerosol particles having settling velocity v that are caught when
S
the average flow velocity is u is given by (Fuchs, 1964):41
O
e = v L/u h , e £ 1. (2)
s g
For elutriators, under laminar flow conditions and ignoring the
Cunningham slip correction term, this velocity is proportional to the
square of the aerodynamic diameter and we can write
e = k Da2 (3)
where k is a coefficient which takes into account the other variables at
a particular set of operating conditions. Efficiency curves for impac-
tors and elutriators go from values of 0 percent efficiency to 100 per-
cent efficiency as particle size increases. The particle size interval
over which the change takes place is a measure of the ability of the
device to discriminate on the basis of size. One can define the rela-
tive standard deviation, RSD, of such efficiency curves in terms of the
particle diameter corresponding to 84 percent efficiendy, D and the
diameter corresponding to 50 percent efficiency, D :
RSD= (D84-V/D50-
43
-------
The smaller the RSD, the better is the size selectivity of the instrument.
It can be shown that elutriators and cyclones that operate with laminar
flow and have the aerosol uniformly distributed across their inlets have
RSD = 0.30. Elutriators and cyclones that operate with turbulent flow
have even larger RSD's; from data on such cyclones given by Liethe and
Mehta (1973), it can be estimated that RSD ~ 4. These are considerably
larger than RSD's characteristic of impactors designed by Lundgren (1967),67
Mercer and Chow (1968),71 Mercer and Stafford (1969),72 and Cooper and
Spielman (1973),29 RSD <_ 0.2.
The gradualness of the efficiency curves for elutriators and cyclones is
not always a drawback to their use, as they can be used to ..model the
response of the human respiratory system (Lippman, 1970).65 For the de-
termination of size distributions, however, this lack of sharpness of
cut is undesirable.
The remaining methods on the list of devices which size classify accord-
ing to aerodynamic diameter are impactors and sedimentometers. Sedimen-
tometers operate by following the change in concentration with time of
a sample held in a special chamber or by following the buildup of depo-
sit in time under such conditions. Sedimentation measurements for
aerosols present a number of difficulties, one of the most significant
of which is convection in the sample cell, a problem which would only
be aggravated by working with samples at fluctuating temperatures, such
as stack samples taken "in situ." The rate of settling is proportional
to the square of the aerodynamic diameter, approximately, which means
that particles having a range of 10:1 in aerodynamic diameter will have
a range of 100:1 of times characteristic of the settling in the chamber.
To sample a. volume of air to obtain a sufficient mass of aerosol on
which the sensor can give significant measurements may require an un-
wieldy, large settling chamber or a series of sampling/waiting-to-settle
cycles, accompanied by an averaging of the changes in concentration
(or deposit) with time. Sedimentation chambers do not seem likely
44
-------
prospects for aerodynamic size classification for source monitoring
or control device testing.
Collection and Sizing by Inertial Impaction
This analysis of the potential ways to discriminate on the basis of
size for a source test device has eliminated or nearly eliminated all
alternatives except for impactors.
Literature recently published on the adaptation of impaction to source
testing indicates that this conclusion is shared by other investigators.
Pilat, Ensor, and Bosch (1970, 1971)80 have described a six-stage, multi-
jet cascade impactor for stack sampling. Bird, McCain, and Harris
(1973)21 concluded that inertial techniques (impactors and cyclones) were
most appropriate for sizing stack effluents and tested the following:
Modified Brink impactor, Andersen stack sampler impactor, University
of Washington Source Test Impactor, Environmental Research Corp. TAG
Sampler, modified Battelle CIS-6 Impactor, and a McCrone Associates
Parallel Cyclone Sampler. They found some difficulties in using these
for in-stack sampling, generally problems of particle rebound and re-
entrainment, but these problems would occur whether the same impactors
were used in-stack or after sample extraction through probes. Bird,
et al. (1973),21 reported a significant loss of particles (especially
those greater than 2 microns diameter), due to the use of probes and
sampling lines, in their comparison of size distributions from in-stack
and out-of-stack sampling.
The feasibility of using a cascade impactor placed inside the source
stack has been amply demonstrated by the nearly 20 years of such ex-
perimental experience reviewed by Brink, et al. (1972).23 This article
discusses in-stack monitoring in a variety of situations with the Brink
(1958)22 adiabatic impactor manufactured by Monsanto. The normal cut-
offs are 2.5, 1.5, 1.0, 0.5, and 0.25 microns. For some applications,
45
-------
stages with larger cutoffs were added, primarily to keep the smaller
cut-off stages from overloading. One of the interesting bits of ex-
perience was that grounding the impactor helped reduce reentrainment
of particles; reentrainment was most severe at low humidities. By using
special gasketing, the cascade impactor worked on metal fume at 1500 F,
at which temperature the 316 stainless steel maintained its integrity.
(It is not mentioned whether or not there was a viscous coating that
withstood these temperatures; however, although such a coating may not
have been necessary because the small size of fume particles would, in
general, inhibit their reentrainment.)
Downs and Strom (1972)35 successfully adapted a Brink cascade impactor
to sample industrial stacks up to temperatures of 450 F.
Carpenter and Brenchley (1972)26 also concluded that there would be real
utility in an instrument that measured mass concentration as a function
of aerodynamic diameter. To do this, they built a cascade impactor
which impacted the particles onto piezoelectric crystal oscillators. .
Their work is discussed in a following section of this document.
The inertial impactor operates on the principle that particles in a
moving stream will impact upon a perpendicular surface provided that
the inertia of the particles is sufficient to overcome the drag exerted
by the air stream deflected by the impaction surface. The impaction
process can be characterized by the impaction parameter ty which is
defined by the following expression:
C p v d2
where p = particle density
v = jet velocity
d = particle diameter
46
-------
y = gas coefficient of viscosity
D = jet diameter
C = Cunningham correction factor defined empirically as:
C = 1 + 2.492 X/d + 0.84 X/d exp (-0.435 d/X) (6)
where X is the molecular mean free path.
Two general impaction configurations have been used in practical instru-
mentation and have been evaluated by researchers in this field, rectan-
gular and round impaction jets. From Mercer and Chow (1968),71 Mercer
and Stafford (1969),72 and Marple (1970),68 it becomes apparent that for
sizing purposes the round cross-sectional jet is to be preferred because
of sharper cut-off characteristics. In addition, the typical value of
Acf. of Equation 5 for round jets is about 0.3, whereas for rectangular
jets it is about 0.45, implying higher jet velocities for a given par-
ticle cut-off size for the rectangular configuration. Finally, the
round jet geometry is generally more compatible with automatic sensing
of the mass deposited, particularly by means of beta radiation absorption
(Lilienfeld, 1970;63 Lilienfeld and Dulchinos, 1972).10°
Mitchell and Pilcher (1959)7t* determined that for round jets a near-
optimum particle size cut-off sharpness is found for an S/D ratio of
0.375. (S is the jet-to-impaction surface gap.) At 50 percent im-
paction efficii
see Figure 7).
1/2
paction efficiency, ^ was found to be equal to 0.084 (i|> = 0.29;
Marple (1970)68 concludes his exhaustive analysis of impactors as follows:
"Application of the above studies on the impaction efficiency curves
indicates that a round impactor should be one with throat length of
about one jet diameter, a jet-to-plate distance of 1/2 jet diameter,
and operated at a Reynolds number of approximately 3000. Such an im-
pactor should have a fairly sharp cut-off characteristic, is reasonably
47
-------
1.0
0.9
0.8
0.7
P" 0.6
>•
o
z
UJ
o 0.5
u.
u.
LJ
0.4)-
0.3
0.2
O.I
S/D = 0.375
I
I
0.5 0.6 0.7
«Y^PV
/2
10 O.I 0.2 0.3 0.4
IMPACTION PARAMETER
Figure 7. Circular nozzle impaction efficiency as a
function of ij). (Mitchell and Pilcher, 1959)
48
-------
insensitive to small variations in the jet-to-plate distance which un-
doubtedly would occur in actual application, and has dimensions that
are reasonable from a manufacturing standpoint."
For particles larger than about 0.1 ym under STP conditions, Equation (6)
can be approximated to
C = 1 + 2.492 X/d . (7)
Replacing Equation (7) in (5) and solving for d, we obtain:
*
-2.492 X + ((2.492 X)2 + 72" D
- l - - _ - (8)
Figure 8 is a plot of d „ versus D, the jet diameter, for several flow
rates. The nozzle pressure drop Ap is assumed to be essentially equal
to the velocity pressure:
Ap = (l/2)p v2 , (9)
cl
where p is the air density and thus X, the molecular mean free path at
3.
the impaction point, would be increased following the relationship:
P + Ap
X = X — (10)
° Po
where X = the mean free path at standard conditions
o
p = the pressure at standard conditions.
The curves of Figure 8 are based on Equations (8), (9), and (10).
The converging behavior of the curves of Figure 8 (dashed lines) for
the smaller jet diameters, is associated with critical flow conditions
49
-------
cc
UJ
UJ
o
(K
O
2
t
3«
8
fi
tu
O
I
UJ
o
t-
oe
2
O.I
g/cm3 (spheres)
« 0.084
« 2xlO"5 kg mr| sec"'
T » 450 *K
DASHED LINES INDICATE
SONIC FLOW REGIME
j_
j_
j_
I
2345
JET DIAMETER -D.- MILLIMETER
Figure 8. Particle cut-off diameter for a circular nozzle
as a function of jet diameter
50
-------
(e.g., sonic velocity) at the jet nozzle. This critical flow regime
may be used to maintain a constant flow rate through the system by
designing the last impaction stage for such limiting conditions as are
achieved when the ratio of downstream to upstream pressures across the
critical nozzle is equal or less than 0.53. Under such conditions the
flow rate Qcrit becomes (Lilienfeld, 1970) t63
where A = the jet area
k = the ratio of specific heats (1.4 for air)
p = upstream pressure
p = upstream density
Under stack gas conditions p can be expected to vary as an inverse
function of temperature, and thus Q . depends on the stack effluent
.temperature. This dependence, however, is not too severe since Q
varies as the 1/2 power of the temperature. The particle cut-off size
for each impaction stage remains essentially unaffected by temperature
changes when using a critical orifice as flow control. This results
from a fortuitous combination of factors: the critical flow is pro-
portional to 1/2 power of temperature, and the particle cut-off size
is inversely proportional to the 1/2 power of the flow rate (for a
constant geometry) , thus the particle cut-off diameter is inversely
proportional to the 1/4 power of temperature as related to the change
in critical flow. However, since the coefficient of viscosity of air
varies approximately as the 2/3 power of temperature, and the particle
cut-off size varies approximately proportional to about the 1/2 power
of temperature due to the change in viscosity. Combining the two
opposing effects, the particle cut-off size varies only as 1/12 power of
temperature in this configuration, and can thus be neglected.
51
-------
THE MEASUREMENT OF MASS CONCENTRATION
General Considerations
At present both the standards for air quality and the standards for
source emissions are based on mass concentration of particulates per
unit volume of gas (air), as specified in the Federal Register (EPA,
1971).38 Mass concentration measurements are thus the generally accepted
particulate emission parameter for source testing (Sem, et al., 1971)1
and, therefore, the sensing principle to be'used must respond to particle
mass rather than some other particle property.
The criteria of applicability and usefulness of any mass sensing tech-
nique to the solution of this problem are: (1) ability to provide an
electrical output signal that can be related to collected mass; (2) com-
patibility with automated and unattended operation, as well as con-
tinuous or continual recording; (3) compatibility with the environmental
conditions prevailing within the stack; and (4) operational reliability.
Table 9 summarizes the application of these criteria to the three mass
sensing techniques mentioned above. This tabulation indicates that from
all points of view, beta attenuation appears as the preferred method of
mass sensing.
Table 9. COMPARISON OF MASS SENSING TECHNIQUES
Technique
Beta attenuation
Resonance sensing
Gravimetry
Criteria
(1)
Good
Good
Imprac-
tical
(2)
Good
Fair to
imprac-
tical
Imprac-
tical
(3)
Fair to
good
Fair
Fair to
imprac-
tical
(4)
Good
Fair
Fair to
low
52
-------
Methods used heretofore have been based either on the pragmatic approach
of determining the weight increment of a filter sample, or on the mea-
surement of a variety of physical or chemical properties with varying
degree of correlation with the parameter of interest, the mass concen-
tration of airborne particles. Such indirect techniques are based on
light scattering, light transmission, electric charge, flow resistance
of filters, acoustic effects, thermal transfer, thermal emission, etc.,
as discussed above. The applicability of any of these indirect methods
is limited to the measurement of aerosols within specific constraints of
size, concentration, index of refraction, phase, composition, etc.
Light scattering, the most commonly used indirect method of measurement
of suspended particulates, suffers from its inherent dependence on index
of refraction and its extreme sensitivity to size characteristics.
Nephelometry, single particle counting and lidar techniques, have shown
only limited applicability to the problem of routine surveillance of air-
borne particulates, and although their fast response makes them valuable
detection tools, for accurate and generally accepted mass concentration
measurements, reliance is still placed on gravimetric determinations.
This gravimetric technique is inherently incompatible with automation,
continuous unattended operation, recording and data transmission. Further-
more, for ambient monitoring, filter weighing implies a logistics problem
of daily removal of loaded filters, their replacement by clean ones and
the initial and final weighing in addition to the storage and/or heating
procedures required to reduce errors associated with water adsorption.
Methods based on the chemical evaluation of collected samples fulfill
their specific purposes but are not useful as routine measurement
techniques.
Two approaches appear at present as uniquely adapted to the measurement
of the mass concentration of airborne particles: radioactive techniques
and mechanical resonance. The latter method is based on the measurement
of mass by the change of the natural resonance frequency of vibrating
53
-------
mechanical systems. In practice, particles are collected on a resonating
body, such as a taut wire (Cast, 1970)^ or a quartz crystal (Olin, et al.,
1971),77 and the resulting mass increment is indicated by a decrease of the
resonant frequency of the vibrating system.
In a comprehensive report to the Environmental Protection Agency,
Sem, et al (1971)1 evaluated the available methods for measuring particle
mass concentrations and concluded that the method most near full develop-
ment as a continuous and automatic stack sampling instrument was the beta
absorption method. They noted that the piezoelectric crystal method was
also very promising, although its development had not progressed as far.
Besides these two "very promising" techniques, they listed a number as
"promising," but needing considerable development: gravimetric, sensing
of impact momentum through vibration, sensing of impact through flexible
beam capacitance change. Several methods which did not measure particle
mass were evaluated as having potential usefulness in certain situations:
contact charging (Konitest), ion capture (a measure of surface area), and
light transmission and scattering (to sense changes in effluxes, to be
followed by mass concentration measurements). Sem, et al. (1971)*
emphasized that the instruments that do not actually measure mass may
give readings well correlated with mass concentration as long as the
particle size distribution and composition do not change, but cannot be
relied upon when these do change, and such instances often are of the
greatest interest, as for example in the evaluation of a control device
or a process change.
Piezoelectric Mass Sensing
The decrease in frequency of the resonant oscillation of a piezoelectric
crystal when mass is added to the crystal has been the basis of a number
of mass monitoring instruments, and some of the experience with such de-
vices reported in the literature is reviewed here. Extremely high mass
sensitivities can be achieved by this technique, in theory limited only
by the stability of the resonator. Efficient and indiscriminate particle
54
-------
collecting, as well as the requirement for subsequent particle removal,
are inherent problems of this method whose complete solution has so far
limited its range of applicability.
Chuan (1970)27 built and tested a piezoelectric quartz crystal oscillator
which he .combined with an impactor to measure particle masses as small
as 5 x 10 g.
A "piezoelectric microbalance" for aerosol mass determination was re-
ported by Olin and Sem (1971).77 They found it highly sensitive to mass
in microgram amounts and greater and they gave the mathematical relation-
ship for the sensitivity (frequency change per mass change). The con-
ditions which they specified for accurate operation were:
1. The deposit should be thin in comparison with the crystal
thickness.
2. The deposit should be uniform.
3. The force of adhesion between particle and surface (paral-
lel to the surface) should be greater than the particle
mass times the particle acceleration due to the vibrations
of the crystal surface, and, for linearity
4. The change in frequency should be small in comparison with
the resonant frequency of the clean crystal.
Olin, et al. (1971)77 built and tested a piezoelectric sensor. They used
it with an electrostatic sampling device to try to provide a uniform sample
on the crystal face. Although they reported successful operation when
sampling tobacco smoke and auto exhaust, they noted a number of problems:
1. In general, for humidities less than 25 percent relative
humidity (r.h.), adhesion to the crystal was poor, re-
sulting in artificially low concentration measurements.
2. For larger particles, sand which had diameters about 10
micrometers, adhesion was poor (at 25 percent r.h., the
only value of relative humidity reported).
55
-------
3. "Both the 1.5 and 5.0 MHz crystals stopped oscillating
when the accumulated particle deposit reached about 400
micrograms."
4. Once the oscillations stopped, the crystal had to be
cleaned, an operation reported to take a couple of
minutes.
On the basis of their work, they concluded that the piezoelectric technique
had a wide range of applications, including automobile exhaust evaluation
and measurement of outdoor aerosols (for which cleaning was necessary after
50 to 100 micrograms of material had been collected), but as for the mo-
nitoring of source emissions, they concluded the method requires that "the
sample is properly conditioned, perhaps by cooling and/or diluting." Of
course, cooling introduces the problems associated with condensation, and
diluting introduces, often, particle size biasing, as well as adding to
the complexity of the instrumentation. <
Carpenter and Brenchley (1972)26 combined the size classifying abilities of
a cascade impactor with the mass sensing of the piezoelectric crystal
frequently-shift techniques. At a flow rate of 0.5 liter per minute,
the four-stage device had particle aerodynamic size cutoffs at 18.9,
12.6, 6.3, and 2.5 micrometers; they did not indicate why these par-
ticular size cutoffs were chosen, so it is not clear whether the re-
sults for this instrument would be applicable to such an instrument
with flow rate and/or slit dimensions tailored to the analysis of dust
in the "respirable" range, which would involve cutoffs about a factor
of five smaller. They used a series of circular-geometry jets at a jet
spacing to jet diameter ratio of 0.375, which ratio was noted above as
having been recommended by previous investigators of round jets. They
had to calibrate the sensors on each stage separately, since the mass
change to frequency change ratio is proportional to the mass sensing
area of the crystal. The standard deviations for the calibration were
12 percent or less. Although Carpenter and Brenchley26 did not present
size distributions for their test aerosols, the particles used to
56
-------
calibrate the device were atomized from a 0.1 percent uranine solutions,
and experience with such atomizers indicates that most of the particles
were smaller than 1 micrometer in diameter. Thus, although the size
cutoffs tested were several micrometers and larger, these calibration
tests indicated only the response to mass by the crystals, but not the
ability or inability of particles larger than a few micrometers to ad-
here to the oscillating crystals.
In summary, although extremely sensitive, piezoelectric mass detection,
the most important embodiment of the resonance frequency method,
suffers from a number of operational difficulties that are more or less
inherent in that technique: (1) incompatibility with prolonged con-
tinuous or continual operation (i.e., without crystal replacement,
cleaning, rezeroing, etc.); (2) the need for intimate particle contact
with the quartz crystal (problems with particle "bounce" or "levitation"
above the collection surface due to the high-frequency crystal vibration);
(3) the limited mass that can be collected before crystal oscillation is
quenched; and (4) the appearance of occasional "glitches" or negative
mass readings, as a result of internal crystal structure shifts or
lattice reorientation.
Beta Radiation Absorption Mass Sensing
Radioactive techniques are, in general, based on the absorption or
scattering of alpha, beta, or gamma radiation of airborne particles
collected on a variety of substrates, by means of filtration, inertial
or centrifugal separation, thermal or electrostatic precipitation, etc.
Absorption of gamma radiation has not found any significant application
to particulate detection mainly because of measurement insensitivity
resulting from the high penetrating power of this type of radiation.
The scattering of gamma rays, however, appears as a promising technique
for the assessment of specific materials in dust mixtures, such as the
determination of the relative proportions of carbon and rock dust in a
57
-------
sample of coal mine dust (Martin and Stewart, 1968), or the relative
amounts of fly ash and carbon resulting from an incomplete combustion
process (Dresia, et al., 1962).36 The disadvantages of these gamma scat-
tering techniques is again the requirement for large amounts of sample,
between tens and hundreds of grams, due to the relatively weak inter-
action cross-section of gamma rays and matter.
Alpha emissions, helium nuclei produced by nuclear decay processes,
have very weak penetrating power and are rapidly absorbed by matter.
Furthermore, the absorption of alpha particles, which is intrinsically
species dependent, is characterized by a step-like, sharply descending
attenuation curve, restricting any practical application of this method
to a narrow range of absorber mass per unit area. Nevertheless, an
airborne dust monitor, utilizing a filter tape and an alpha source-
detector configuration, has recently been awarded a U.S. patent (Babich,
et al., 1971)16 issued to five Soviet inventors. Whether such an instru-
ment has any practical value remains to be determined.
Beta radiation, the nuclear emission of an electron as a result of the
spontaneous radioactive conversion of a neutron into a proton, presents
some unique properties of particular value to the measurement of col-
lected aerosol particulates. Beta scattering has been utilized in
applications similar to those of gamma scattering; i.e., the determi-
nation of fly ash/carbon ratios, or to ascertain the combustible versus
incombustible fraction of coal mine dust (Martin and Stewart, 1968).69
The main application of beta radiation in the field of airborne par-
ticulate pollution, however, is based on its absorption properties.
Beta absorption has been studied extensively in the past, and its
application has ranged from gas densitometry to thickness gaging
(Shumilovskii and Milttser, 1964;85 Brownell, 1961 ;2k Dempsey and Polishu,
1966). 33 A brief review of some basic and relevant concepts follows.
58
-------
The electrons emitted by the beta process have energies ranging from
zero to some well-defined value E . The governing absorption mech-
max
anism is dependent on the energy of the beta electrons. In their
interaction with matter, beta particles are absorbed either by ionizing
collisions or by radiative processes (e.g., bremsstrahlung). For low-
14
energy betas such as those generated by C, where ionization predo-
minates as an absorption mechanism the attenuation of beta particles
is a function of the electron density of the absorber. The electron
density in matter can be expressed by the relationship:
P = P N Z (12)
AM
where p = the electron density of the absorber
p - the mass density of the absorber
N = Avogadro's number
&
Z = the atomic number of the absorber
M = the mass number of the absorber
From the above equation, it is apparent that for a constant Z/M ratio,
the absorption of low-energy betas is only a function of the mass
density.
Except for hydrogen (Z/M = 1), the ratio of atomic number to mass number
for all other naturally occurring elements varies from about 0.4 to 0.5,
and for most elements associated with airborne particulates, this ratio
remains between narrower limits (0.45 to 0.5). Table 10 lists a few
selected elements and compounds - typical of stack gas particulates -
and their respective Z/M ratios. As a consequence the penetration of
low-energy beta radiation depends almost exclusively on the area density
(thickness times density) of the absorber and the maximum beta energy of
the impinging electrons, and is essentially independent of the chemical
composition of the absorbing matter (Lilienfeld and Doyle, 1973;6lt Horn,
1968;2 Dresia, et al. , 1964;3 Nader and Allen, I960;1* Denzel and Horn,
1966;5 Aurand and Bosch, 1967;6 Benarie and Loverdo, 1967;'7 Lilienfeld,
59
-------
1971,8 Herling, 1971;9 Izmailov, 1961;10 Bulba and Silverman, 1965;11
Lilienfeld and Dulchinos, 1971),12 with the exception of hydrogen whose
mass contribution to aerosol particles is vanishingly small.
Thus, Equation (12) can be expressed as
Pe = kp
(13)
and thus low-energy beta attenuation is essentially a function of the
mass/area of absorbing matter.
As a beam of beta particles traverses a substance, the radiation flux
is attenuated, following the exponential law
1=1 exp (-y 6)
o m
(14)
Table 10. Z/M RATIOS OF SELECTED ELEMENTS AND COMPOUNDS
TYPICAL OF PARTICIPATE EMISSIONS FROM COAL-
FIRED STEAM POWER PLANTS
Element or Compound
Calcium
Carbon
Oxygen
Silicon
Sodium
Potassium
Si02
A1203
Fe2°3
CaO
so3
MgO
Z/M
0.499
0.5
0.5
0.498
0.478
0.486
0.499
0.490
0.476
0.499
0.499
0.496
60
-------
where I - the radiation flux incident on the absorbing material
o
I = the transmitted particle flux
6 = the absorber mass per unit area
p = the mass absorption coefficient defined by the approxi-
m mate empirical relationship (Shumilovskii and Milttser,
1964)
p = 0.022 E~4/3 (15)
m max '
where E is the maximum beta energy of the isotope, in MeV.
max
Equation (15) indicates that the sensitivity of a measurement based on
beta absorption increases with decreasing maximum energy. The choice
of a specific radioactive isotope for beta absorption sensing is usually
based on a compromise between measurement sensitivity requirements and
the range or depth of penetration of the beta radiation. In addition,
the half -life or decay constant, interfering presence of concomitant
gamma radiation, as well as chemical and physical stability, cost and
availability are important practical considerations bearing on the se-
lection of radioactive isotope for a specific application.
The sensitivity of a beta-absorption measurement is limited by the random
fluctuations of the rate at which beta particles are ejected from the
atomic nucleus. As in any process subject to statistical errors, the sen-
sitivity or resolution of a beta-attenuation measurement increases with
the total number of events counted. Poisson statistics have been found
to be applicable to nuclear decay processes. Thus, the standard deviation
a of a given beta count N can be expressed as :
a = N (16)
By differentiation of Equation (14) with respect to 6, and replacing I by
N, we find that
61
-------
m
which can be expressed as
A6 = — = ^— (18)
^m y
m
Where A6 is the minimum detectable mass per unit area increment, or the
one-sigma error of the beta attenuation measurement of 6.
The relative measurement error due to counting statistics, E can be ob-
tained from Equations (14) and (18)
E • .
r 6
y -vN y
m m
i —
6-v/N y 6
E has a minimum when y 6=2, which defines the condition for obtaining
the minimum relative statistical error; i.e., the maximum signal-to-noise
ratio in a beta sensing system (Lilienfeld, 1971).8 Since statistical
errors usually predominate in instrumentation based on beta-attenuation
measurements, the above derived criterion provides a practical design
guideline for the choice of the radioactive isotope (y ) and the collec-
tion substrate (6).
The detected beta count N is equivalent to the product of the average
beta rate f and the counting time T. For a dust sampling system, the two-
sigma error limit (95 percent confidence of a given measurement) of the
measured mass concentration can be obtained from Equation (18) as
c = 2A
min(2o)
62
-------
where A is the collection area, Q is the volumetric flow rate, and t is
the dust sampling time. Equation (20) can be expressed in terms of the
flow velocity v at the collection site as:
m
Equation (21) is applicable for step-like particulate sampling; i.e., a
continual cyclic operation wherein each cycle is composed of an initial
beta counting period T, followed by a particulate sampling period t, which
in turn is followed by another, final beta counting period T.
For a collection-detection configuration utilizing a continuously advancing
substrate, Equation (21) becomes (Lilienfeld, 1971):8
c = 2 sb ( .
Cmin(2o) (22)
where s = the substrate transport velocity
b = the width of collection area
T = the time constant of the averaging circuit of the
beta pulse rate sensing circuit
The selection of the radioactive isotope to be used for a given application
is based on a compromise between measurement sensitivity and beta-penetration
range. Measurement sensitivity indicates the need for a large value of the
y , the absorption coefficient, thus a low-beta energy. Low energy, however,
implies reduced range or penetration. The inevitable additive presence of
source self-absorption, detector window, air gap, and collection substrate
absorption dictate a minimum value for the range of the order of 5 to
2
10 mg/cm . As a result, the use of such low-energy beta emitters, as
nickel-63 is usually eliminated. The most commonly selected isotope is
63
-------
carbon-14 whose properties are particularly compatible with particulate
mass sensing. It has a maximum beta energy of 0.156 MeV and decays into a
14
stable nuclide ( N). An important consideration is the half-life of the
isotope, which should be long as compared with typical utilization periods
14
of the order of 10 years. C with a half-life of about 5700 years, is
ideally suited for this purpose, obviating periodic adjustments or routine
14
source replacement. Other favorable characteristics of C are handling
safety, relative freedom from licensing requirements, price, and availability.
GENERAL CONCLUSIONS
The conclusions reached as a result of the foregoing literature review will
now be summarized, and a resulting instrumental concept will be outlined
incorporating the design and operational guidelines based on those
conclusions.
Reviewing the objective, the fundamental purpose is to evolve a technique
capable of representatively assessing the particle size distribution by
mass of particulates in effluent stream, and to perform such a measurement
automatically with a recorded output. The following combination of tech-
niques is recommended as a result of the review of the present state-of-
the-art .
• The collection of particulates should be performed by
means of inertial impaction.
• Size segregation of particulates should be performed
by means of series or cascade impaction.
• Round jet impaction is to be preferred because of
superior size selectivity and compatibility with mass
sensing techniques.
• Mass sensing of the particulates collected by impac-
tion should be performed by beta-radiation attenuation.
A size-selective aerosol monitor, however, faces a difficult challenge
when used for in-stack monitoring. Temperatures significantly higher
64
-------
than room temperature, fluctuations in flow velocity and particle concen-
tration, as well as spatial inhomogeneities of both these quantities,
combine with moisture and corrosive gas to present an inhospitable environ-
ment for accurate determination of particle mass concentration as a func-
tion of aerodynamic diameter. Any design of such an instrument should
take these conditions into consideration.
As a result of their thorough investigation into the question of "particle
sizing techniques for control device evaluation," Bird, et al. (1973)21
concluded that inertial separation devices covering the range from 0.2 to
5 micrometers and used in conjunction with sensitive mass concentration
sensors were the best candidates for such application. Their conclusions,
however, also include the statement: "Based on simultaneous size deter-
mination with the in-stack and out-of-the-stack sampling arrangements, it
seems almost certain that a signififant quantity of particles are lost
in the probe and sampling lines, particularly in sizes larger than about
2 ym."
Thus, in addition to the specific techniques listed above the following
important conclusion is reached with respect to the location of the par-
ticulate collector-sensor: for representative mass concentration measure-
ments and especially to size the aerosol by mass (i.e., mass versus aero-
dynamic diameter), the extraction of the sample flow through ducts (probes)
for collection and measurement at typical distances of several meters
from the sampler inlet is unacceptable because of largely uncontrollable
line losses and alteration of the original size distribution.
In conclusion, there is a consensus that it is desirable to measure par-
ticulates from source emissions with a device that discriminates particle
size on the basis of aerodynamic diameter and measures particle mass con-
centration. The most promising approach to these goals is an in-stack
device based on impaction to achieve size discrimination and using beta
absorption to give mass concentration readings automatically and in a
65
-------
form compatible with convenient data recording, transmission, and analysis.
Such an instrument would provide an extremely valuable tool for the quan-
titative assessment and evaluation of emission control measures and de-
vices under a variety of conditions, and would lead to their full charact-
terization and eventual optimization.
66
-------
SECTION V
INSTRUMENT DEVELOPMENT
Based on the conclusions reached within the preceding Literature Review,
an instrument concept was evolved incorporating the basic principles of
inertial size segregation by impaction, mass sensing by beta radiation
attenuation, and direct sampling from the stack environment. The design,
development, fabrication and testing phases of a prototype instrument are
described in the following sections of this document.
OVERALL DESCRIPTION
The multi-stage in-stack particulate mass monitor system described in this
report consists of the following:
1. A cascade seven-stage modular collector-sensor section
2. A support and substrate driving structure
3. Electronic control and recording console
4. Pump.
The first section consists of seven modular impaction-sensing assemblies
each of which contains a circular nozzle jet impaction stage, a thin foil-
film moving substrate cartridge, and two beta radiation source-detector
pairs, one of which serving as a reference and the other sensing the col-
lected mass.
The second element of the system consists of a support member for the
collection-detection head, and it contains the stepping motors gears and
couplings to drive the substrates at all the collection stages.
67
-------
The console contains all electronic circuitry required for signal processing
and conditioning, the multichannel strip chart recorder, the substrate
stepping motor drive circuitry, and the power supplies.
The pump originally contemplated during the development effort consisted of
a small air-ejector driven by a compressor. Actual laboratory tests, how-
ever were performed with a rotary vane pump.
The principle of operation for the instrument can be described as follows.
Particles entering the collection-detection inlet nozzle are collected on
successive cascaded impaction stages for size segregation. Particles are
collected at each stage by impaction on a adhesive-coated continuously
advancing substrate consisting of a thin film. The resulting collection
is thus a trace or band whose width is approximately equal to each of the
impaction orifice nozzles. The collection density of this trace clearly
depends on three principal parameters: the mass concentration of the par-
ticulate size fraction collected at a given stage, the width of the trace,
and the rate of substrate advance. Mass sensing is performed by means of
beta radiation at each stage using one source detector-pair to sense the
beta transmission of the clean substrate before it passes under the im-
paction nozzle, and another source-detector pair to sense the beta trans-
mission thourgh the soiled portion of the substrate; i.e., after passing
under the impaction nozzle.
The pulse train signals from each detector of each of the seven stages are
fed to the control console by means of individual coaxial cables and each
pair of signals (i.e., from each stage) is processed to provide a signal
proportional to the particulate concentration at each sizing stage. These
signals are continuously recorded on the strip chart recorder; one trace for
each collection-sensing stage.
68
-------
MECHANICAL DESIGN
Design Criteria for Impaction Stages
From the Literature Review study (see preceding section of this report)
evolved within this program the design of the impaction stages was based
on the following criteria:
• impactor throat length = jet diameter
• impactor jet-to-plate distance = 1/2 jet diameter
• impactor jet cross section, circular.
These criteria were obtained from the exhaustive impactor study performed
by Marple (see Marple 1970 of above Literature Review). The additional
recommendation for a Reynolds number of approximately 3000 could only be
adhered to within the first three stages of the cascade impactor. In
general this latter criterion can only be applied in a very limited manner
to cascade impactors because of the common flow rate of all stages. Never-
theless, the highest Reynolds numbers only exist at the last stages where
smaller particles are involved and where reentrainment was thus less likely
to occur.
Design of Collection-Detection Head
Since the collection-detection in-stack head constituted a centrally im-
portant element of the measuring system especially careful attention was
given to these design details in order to optimize the operation, reli-
ability and provide adequate operational flexibility within the limitations
imposed by dimensional and environmental constraints. The design of the
substrate advance mechanism was particularly critical and several al-
ternative variations were carefully considered. The design of the foil
advance mechanism was modified somewhat with respect to the originally con-
sidered approach, in order to provide a more efficient use of the entire
substrate width (38 mm).
69
-------
This approach consisted of the incorporation of a lateral displacement
feature whereby when the substrate film reaches one end the entire cassette
would move forward and the foil is then transported in the opposite
direction. This arrangement permited six dust traces to be collected on
one strip of foil before it had to be cleaned or replaced. This mechanism
requires three drive shafts which are coupled to each of the stages. In
practical operation each of these cassettes is inserted completely into
each stage compartment and then slowly and automatically moves back such
that when the operator removes these cassettes at the end of a measurement
run, they are all on the near side of their insertion opening. Various
operational difficulties precluded the actual use of this lateral motion
feature during the final testing progress and only unidirectional substrate
translation was used during these tests.
A Mesur-Matic stepping motor was driven by a programmable circuit which
provides a predetermined number of pulses to the stepping motor, at a
selected rate. At the completion of the preselected number of pulses, the
direction of the motor is automatically reversed. Translational motion of
the cassette occurs at the same time by means of a separate stepping motor.
At the completion of five parallel collection traces the driving circuit
automatically stops the stepping motor. An electromechanical pulse counter
provides running information on the length of substrate that has been
transported, thus permitting the interuption of a run, and its subsequent
resumption, if so desired, without changing the substrate cassettes before
the end of the normal substrate excursion.
Figure 9 depicts the collection-detection head assembly in the 90° folded
position to be used for typical vertical stack sampling. This picture
shows the seven-stage head without the cover plate, and without the inlet
nozzle. The detector cables are not plugged in.
Figure 10 shows separately one of the modular stages (side view) with the
inlet side of the conical impaction nozzle at the top, and the substrate
transport cassette at the bottom.
70
-------
Figure 9. Assembled seven-stage collector-sensor
7!
-------
Figure 10. Single modular impaction-cassette assembly
-------
Figure 11 depicts three sections of the system: at the left, the inlet
section; at the center, one of the modular stages (view from inlet end)
showing the three drive shaft interlocks, the central one for the lateral
cassette translation, and the two side ones ofr the normal substrate trans-
port, one for each direction of motion; and at right, a loaded cassette
with the stainless steel foil substrate.
Figure 12 shows a detail of the articulated interface between the support
probe and the collection-detection head.
IMPACTION SUBSTRATE ASPECTS
Two principal aspects of impaction were investigated during this develop-
ment program: the substrate material; i.e., the foil or film on which col-
lection was performed, and the coating to be applied to ensure particle
retention. The substrate film had to fulfill the following conditions:
compatibility with high temperature operation (i.e., up to 260°C if pos-
sible) ; mechanical strength and flexibility compatible with continuous
motion and transport; and sufficiently low mass per unit area to permit a
transmission of at least 25 percent of the impinging radiation from the
selected Carbon-14 source isotope.
The following were considered for the selection of a suitable collection-
impaction tape substrate material: teflon filter medium (Millipore),
ceramic paper, and metallic foils. Both commercially available teflon
and ceramic paper tapes were partially acceptable, their common dis-
advantage being their excessive thickness resulting in excessive beta at-
tenuation, and reduced useful tape length per roll. Metallic films (e.g.,
stainless steel) appeared quite promising and although their requirement
for adhesive coating and cost considerations are somewhat negative char-
acteristics, the initial selection was a stainless steel film with a 5 vim
2
thickness corresponding to a mass/area of about 4 mg/cm which resulted in
a Carbon-14 beta transmission of the order of 35 percent (i.e., 65 percent
attenuation). For reasons of operational convenience, however, during the
laboratory tests performed around room temperature (see Section VI of this
report) mylar substrate films were used.
73
-------
Figure 11. Cassette, impaction nozzle and collector inlet sections
-------
Figure 12. Substrate drive and 90 rotation mechanisms
7r>
-------
Observations at GCA have indicated that the interaction between the im-
pacting particles and the substrate adhesive coating is more complex than
usually assumed. The retention of particles by actual inelastic penetra-
tion into the coating probably plays only a secondary or partially con-
trolling role, since the depth of penetration into a coating even as fluid
as vaseline, at the velocities to be considered in this application are .
considerably less than the particle diameter, thus precluding the collection
of further dust, after a monolayer had been deposited. Since this is
obviously not the case in most situations of this type a different mechanism
must predominate to retain the particles impacting on a viscous surface.
Experimental observations at GCA have indeed confirmed the presence of such
a mechanism: capillary migration of the fluid substrate, or particle
"wetting." This phenomenon gives rise to an upward migration of the viscous
fluid substrate tending to provide a thin sheath of adhesive above the par-
ticulates as they are collected. The upward capillary motion or the
wetting rate, is a function of a number of variables, such as viscosity,
surface tension and thickness of the surface coating, and to a certain
extent depends on particle characteristics such as size, and surface
properties. The controlling particle retention criterion, that emerges
from this mechanism is that capillary migration rate be equal or exceed the
rate of particle deposition in order to maintain a wet or adhesive upper
layer. This criterion provides a high limit for the coating viscosity. At
the other extreme, if the viscosity becomes too low, the dynamic pressure
exerted by the air jet results in a net outflow of the material radially
outwards from the axis jet; i.e., cratering of the adhesive coating ensues.
For room temperature conditions, such mixtures as petroleum jelly and
paraffin oil (3/1 ratio) appears to provide the adequate combination of
viscosity and surface tension for the collection of most particulates. At
higher temperatures (> 40°C) however, that type of coating becomes totally
fluid; i.e., the viscosity and surface tension decrease drastically re-
sulting in the above mentioned radially outward displacement of the coating,
especially at jet velocities in excess of about 30 or 40 meters per second.
76
-------
Within this program various substrate coatings were investigated using a
GCA respirable dust monitor, a device which employs impaction for dust col-
lection and beta attenuation for mass sensing. Normally, this instrument
is operated with a 3/1 petroleum jelly to paraffin oil coating. Comparisons
were made between two instruments, one operated with the usual coating, the
other operated with various other substrate coatings. Texaco high tempera-
ture grease, with or without a small amount (less than 10 percent by weight)
of motor oil added to lower viscosity, and which was applied to the mlyar
substrate by solutionizing with benzene which evaporates leaving a uniform
coating, gave results as good as the usual coating. Unfortunately, tempera-
ture challenge showed severe oxidation at 500°F, and some evidence of lateral
flow with weight loss at 400°F to an extent that precluded its use at these
temperatures. Tests with mixtures of Dow silicone oil (e.g., 100 centistoke)
and Dow High Vacuum Grease gave results almost as good as the normal
petroleum jelly/paraffin oil mixture at room temperature. It is felt that
refinement of the exact mixing ratio (around 1:1) and improvements in ap-
plication to ensure homogeneous coating should provide good results. Tem-
perature challenge to 500°F showed no evidence of lateral flow with less
than 1 percent weight loss after 30 minutes indicating its potential use-
fulness for high temperature use.
It should be mentioned that Dow High Vacuum Grease by itself can be used
as an impaction substrate coating only for light particle loadings and
that the addition of the silicone oil is necessary to lower the viscosity
sufficiently to allow wetting for heavier particle loadings.
PUMP SELECTION AND FLOW CONTROL ASPECTS
The problem of condensation of low boiling point gaseous cprnpoundslextracted
from the stack effluent at elevated temperatures that are exposed to a
significant decrease in temperature before reaching the pump, had to be
considered. One possible solution was based on maintaining the temperature
of the gas emerging from the support probe above the condensation point of
most of such condensibles; e.g., 120 to 150°C by means of a heated duct
77
-------
upstream of the pump and by using a pump impervious to temperatures of up
to about 200°C. This approach would have required the use of a metal
bellows type flexible duct from the exit of the support probe to the pump
inlet, wrapped with an electric heating tape, and the pump required for
this application would have been a high temperature welded bellows positive
pressure pump which uses stainless steel valves with teflon gaskets capable
of operation up to about 230°C, and commercially available.
An alternative solution to the high temperature duct-pump was pursued and
adopted for this program. This method consists of using an air ejector
powered by a small compressor (or high pressure supply if available at the
test site). This air ejector can be mounted directly on the exhaust of
the support probe emerging from the stack. A high pressure air line from
the compressor provides the flow to operate the air ejector, whose exhaust
can be ducted away or conceivably back into the stack. Two air ejectors
manufactured by Air Vac Engineering Co., Inc., Milford, Connecticut were
evaluated (Models TD-260 and AV-191) and both were found to be capable of
providing the required suction flow rate and vacuum for this application,
of about 55 kN/m2 absolute pressure (14 in Hg, or 16 in Hg vacuum) and
300 cm3/sec (18 liters/min). Air ejector model AV-191 required a supply
flow of 2.12 x 10-3 m3/sec (4.5 scfm) at 310 kN/m2 (45 psig), model TD-260
required 3.8 x 10~3 m3/sec (8 scfm) at the same pressure. Thus the former
was found to be more efficient for this application.
The above outlined method to obtain the necessary sampling flowrate through
the system had the notable advantage that the hot gases sampled are not
brought in contact with any moving parts and no deterioration of the flow
system is thus to be expected. Another advantage of this approach is its
lower cost than the metal bellows pump solution.
The two air ejectors mentioned above were tested in combination with a
small Sears air compressor acquired on this program. It was found that the
Model AV-191 ejector provided the flow-vacuum regime of 18 liters/min at
about 16 in Hg vacuum estimated to be required for the in-stack system (at
standard temperature and pressure), with the Sears compressor.
78
-------
Flow rate control was to be achieved by operating the last impaction stage
under critical (i.e., Mach 1) conditions. It was considered that this
relatively high jet velocity would not result in significant particle bounce
at that stage because of the small particle dimensions of the fraction re-
maining after passage through the preceding six stages.
The use of critical or sonic flow conditions to maintain a constant flow
was discussed within the preceding Literature Review section.
BETA RADIATION SOURCES
The preliminary selection of the radioactive isotope to be used as beta
radiation source for this instrument was made within the Literature Review
(Section IV of this document). The use of Carbon-14 within the present
application presented a number of advantages over other beta emitters and
the only question remaining before a final selection could be made was the
aspect of high temperature compatibility.
Special 100 microcurie carbon-14 beta radiation sources were prepared for
this program by New England Nuclear, North Billerica, Massachusetts using
high temperature epoxy, polyimide kapton, cured under vacuum at about
150°C. This latter technique is required to prevent trapping air within
the sealed source and the concomitant problem associated with expansion of
that gas in a high temperature environment.
The first two sources so manufactured were tested by placing them in an oven
and bringing the temperature up to an initial 240°C, after which they were
left in the oven for over 48 hours, by which time the temperature had
drifted up to 270°C. The beta output was then compared to the pre-bake
level with the result that the count rate had varied less than 2 percent
between the two conditions.
Subsequently, fifteen sources of similar design were temperature challenged
up to 260°C. Prior to the temperature testing, a careful microscope ;
79
-------
examination of the sealing window covering the radioactive material was
made so that any thermo-mechanical stresses or construction imperfections
could be later ascertained. No change in physical appearance was seen
under the microscope. Leakage tests were negative confirming that the
temperature cycling had not degraded the physical integrity of these
sources. Fast temperature rate of change (about 60°C/min), which would
tend to provoke thermal stress damage, was used to exaggerate anticipated
operating conditions where the large mass of the impactor would result in
lower temperature change rate.
One source was tested to destruction to determine maximum operating tem-
perature. At a temperature of 400°C (750°F) the source was destroyed,
that is, severe C-14 leakage was observed.
The beta sources were physically incorporated at the end of small stainless
steel rod whose dimensions are: length, 25 mm; and diameter, 2 mm. The
active source area at one end of this rod has a diameter of approximately
1 mm.
BETA RADIATION DETECTION
The problem of sensing beta radiation within the above described multi-stage
collector-sensor system at elevated temperatures proved to be insurmountable
within this development program, at least insofar as a stable, long-term,
reliable detector operation could not be achieved at temperatures approach-
ing 200°C (400°F). Some degradation was observed even at lower temperatures.
Based on verbal information provided by a manufacturer* it was assumed, at
the inception of this program, that the manufacture of geiger-mliller
detectors for low energy beta radiation detection capable of operation at
260°C was well within the state-of-the-art. This proved not to be the
case.
*
LND, Inc., Oceanside, New York.
80
-------
End-window miniature geiger detectors especially designed by LND were
subjected to high temperature testing at GCA to determine their ability to
withstand the required environment of up to 260°C (500°F). Sixteen
detectors of this type were placed in an oven after establishing their room
temperature operating performance with a standard source test configuration.
The temperature rise and fall were carefully controlled in order not to
exceed a temperature rate of change of about 5°C/min and thus not to impose
excessive differential expansion (or contraction) stresses on the various
bonding surfaces of the detectors.
The results of these tests were discouraging. All detectors exhibited
significant decreases in the plateau voltage (i.e., the operating voltage
range) and significant increases in the count rates even when the applied
voltage was reduced to accommodate for the decreased plateau level. After
several high temperature test cycles the plateau voltages had fallen
typically from central values of 750V down to 500V with count rate in-
creases of as much as 50 percent. Such changes were obviously unacceptable
and furthermore implied a serious gradual deterioration of the detectors
leading to their complete breakdown, which was actually observed in two of
the units at the completion of this test series.
The manufacturer of the geiger detectors was contacted and following a
visit to GCA by one of their principals, the detectors were returned to the
factory. LND promised to resolve the problem and once more supply GCA with
improved detectors. The problem of detector degradation appeared to be
associated with the gradual depletion of the halogen quenching component of
the gas filling, due to surface absorption to the cathode as well as to the
insulation material separating the two electrodes. A possible solution to
this problem appeared to be appropriate surface treatment of these absorb-
ing areas by careful decontamination, vacuum treatment, passivation and
other methods.
Another manufacturer claiming to have developed high temperature geiger
tubes was contacted (Harshaw Chemical Co., Solon, Ohio) in order to search
for an alternative source for such devices.
81
-------
After the detectors were returned to the manufacturer (LND) and following a
rather lengthy waiting period, several detectors were received and tested
up to 200°C, a few of these appeared to withstand that temperature at
least for a few hours, while the majority of them, however, exhibited rapid
failure modes similar to those observed during the initial tests. Several
months had been lost in this fruitless effort and the decision was reached
to abandon the approach of using self-quenching geiger tubes for this high
temperature application (no progress could be made with alternate manu-
facturers) . As mentioned above, the problem of high temperature induced
degradation of these devices was obviously associated with the decomposition,
adsorption, and gradual depletion of the halogen quenching gas additive used
in these tubes to achieve rapid ion recombination after a beta pulse is
generated enabling the detector to sense a subsequent beta event. This
halogen quenching gas has a very long lifetime at, and around, room
temperature, but apparently reacts with the inner tube surfaces at elevated
temperatures, with the concomitant inability of the tube to restore itself
after each beta pulse.
Various alternative approaches were considered; among these the most inter-
esting were: (1) scintillation detection in the high temperature environ-
ment, with light signal transmission to the stack exterior by means of
fiber optics, and detection by individual photo-multiplier tubes; and
(2) use of geiger detectors without internal quenching agent, using ex-
ternal electronic pulse quenching. This latter approach appeared more
promising and feasible than the scintillator method which entailed a quantum
jump in system complexity and cost.
Electronic quenching of geiger detectors is aimed at extinguishing the
gaseous discharge, resulting from each detected beta event, by rapidly
interrupting the high voltage applied to the tube, followed by a rapid
recovery of that potential to enable the tube to detect a subsequent event.
The total time required from the initiation of the first pulse until a
second one can be detected by the tube is called the dead time, which is
always slightly less than the actual resolving time, which incorporates
82
-------
the.delay caused by constraints imposed by the electronic triggering cir-
cuit. 93 Electronic, or external, quenching had been applied in the recent
past mainly to shorten the dead time of detectors already using self-
quenching gases and to extend their lifetime.94,95 Plateau slope and
length improvements have been reported as a result of such circuits.96 one
specific application, however, was found in the Iiterature97 wherein a
geiger-type tube was developed for operation at elevated temperatures (as
high as 540 C or 1000°F) using external or electronic quenching without an
internal quenching additive. The quenching circuit used with that device
was obtained from the General Electric Co.98 xhe above described device,
however, operated with dead times of about 1 ms, about 20 times longer than
the dead time required for the in-stack analyzer. It was decided to at-
tempt to utilize the general approach of the G.E. circuit for the present
application, but to modify it in order to achieve a reduction in the dead
time, to permit operation with a lengthy coaxial cable required between
the in-stack head and the control console, and to replace some of the out-
dated components of the original circuit with better ones from the point of
view of speed of signal processing, cost and packaging simplicity. Several
circuit versions were developed in order to optimize the operation of non-
self quenching geiger tubes. Several of these detectors were obtained from
a local source,* and an extensive series of tests was performed with dif-
ferent variations of circuitry and detector design.
The problem of transferring the detected pulse from the tube to the circuit
and to return the quenching pulse from the circuit to the detector through
a length of coaxial cable interconnecting these elements without excessive
time response degradation had to be resolved as the circuit was being de-
veloped. It appeared that a successful solution to that problem is obtained
by subdividing the high resistance detector anode resistor (total of about
10 M ohm) into two elements, one connected directly to the anode and the
other at the high voltage power supply output.
*
TGM, Inc., Waltham, Massachusetts.
83
-------
Operating at room temperature and by proper selection of circuit components,
operating high voltage, high voltage pulse downward excursion amplitude,
etc., detection dead times of 45 to 50 ys were obtained routinely, values
that are essentially equal or better than the self-quenched tube dead-
time, using a miniature geiger detector without the halogen gas additive,
supplied by TGM. This organization supplied GCA with several miniature
tubes filled with the usual noble gas mixture (argon and neon) at dif-
ferent pressures ranging from 8000 to 53,000 N/m2 (60 to 400 torr) at room
temperature. The higher pressure filled tubes provided the more reliable
and faster operation at room temperature, however, 400 torr is approxi-
mately the highest permissible pressure at room temperature because it
entails a pressure of about 710 torr at the maximum stack temperature of
260 C. The tube internal pressure must remain below or exceed only very
slightly the outside pressure in order to prevent window rupture.
These detectors, without halogen quenching gas additive, were temperature
tested up to 400°F under actual operation with the external quenching
circuit mentioned above. These tubes appeared to operate well only for a
short period of time (1 to 2 hours) after which they failed irreversibly
by exhibiting completely unstable operation. This behavior was quite
similar to that observed with the self-quenching type of tubes previously
tested and rejected.
After a series of in-depth discussions with the supplier of these detectors
it was concluded that these "quenchless" tubes may have had trace amounts
of unwanted quenching gas as a result of the manner in which they had been
fabricated and that the circuit developed for their quenching relied on
the presence of this spurious contamination. Once this trace amount was
depleted by heating the tubes ceased to operate.
A new batch of detectors was made by TGM taking special care in insuring
gas purity. These detectors were filled with pure neon. Upon receipt of
these new tubes it was found that the quenching circuit was totally in-
operative, confirming the hypothesis that the initial batch of tubes con-
tributed their own quenching to the circuit operation. It was determined
84
-------
that one of the problems was that a beta event which occurred during the
voltage reapplication period (following an initial beta event) could
initiate a discharge but would not be detected by the cricuit and would
thus not be quenched leaving the tube in a continuous discharge condition.
The reason for this oversight in the design of the circuit was that the
tubes with which the circuit were made to work had some self-quenching
properties due to residual quench gas, as mentioned above. Furthermore,
it was found that the inherent dead time of the^circuit was about 1 milli-
second when not accelerated by intrinsic detector self-quenching.and
limited to such long dead times when used in combination with a significant
cable capacitance (about 200 pF) required to interconnect the tube with
the electronic circuitry.
To achieve both rapid quenching response as well as to prevent the pos-
sibility of "latching on" when two beta events follow each other within
the system dead time a completely redesigned circuit was gradually evolved
(see Figure 13). Typical dead times of 70 microseconds were obtained
under operating conditions with this circuit. Several versions of the
quenchless type geiger detector were made by TGM and tested. The tubes
exhibiting the most reliable operation were constructed by following
a series of steps whose objective was to minimize contamination of
the filling gas and the internal surfaces. After the parts; i.e., elec-
trodes, ceramic insulator, mica window, etc., were cleaned, they were
assembled in air at elevated temperature. The electrodes, which are
20 percent chromium-iron, become oxidized thereby providing a satisfactory
surface for the gas-metal sealing bond. An additional processing step
added in these tubes is a low pressure ion bombardment to remove this oxide
after assembly. Ordinarily in self-quenched halogen tubes, the oxide is
not removed so as to passivate the electrode surfaces thereby inhibiting
chemical depletion of the quenching gas. The tubes made for GCA were
cleaned by ion sputtering after which they were baked out at 350°C for
1-1/2 hours under vacuum and then filled from a pre-mixed cylinder with a
mixture of neon and 0.5 percent argon after which they were sealed.
85
-------
oo
IO/32/IOOK
R4 -HH.V.
/SA< 1
260 "=
to
300V DC
0+5V
|
Figure 13. Experimental quench circuit (first version)
-------
A series of tests were performed to determine the behavior of these
specially constructed detectors in conjunction with the experimental cir-
cuit described above, which afforded the functional flexibility required
to optimize the operation of these tubes.
The main test series consisted of a number of 2-hour heating cycles such
that the tube under test would be taken alternatively from room tempera-
ture up to 400°F and down again after 2 hours, permitting it to cool off
again after which the cycle was repeated. This test was continued for a
total of 32 hours of which about 20 hours were at the elevated tempera-
ture of 400°F. Although the tubes under test exhibited a slight change in
the operating voltage range (from an initial range or platteau of 280 to 330
volts to a final one of 250 to 295 volts), these tests indicated that the
tubes under scrutiny were not "perfect" and that a finite lifetime could
be foreseen. Furthermore, the "survival" of these detectors under cyclic
temperature conditions appeared to depend on their continuous electrical
operation; i.e., these tubes had to be run continuously to preserve their
viability.
Subsequently several advances were made in the design of the electronic
quenching circuit providing both much improved circuit performance and
reduced component count. The circuit resulting from this development (see
Figure 14) demonstrates that all tubes which operate properly as counters
also work well as relaxation oscillators at moderate (0.5 mA) currents so
that considerable circuit simplification can be achieved'. This oscillation
eliminates the need for a d.c. current sense as it will retrigger the
monostable if the tube does not quench properly.
The input transistor, Q3, is run with grounded base with an emitter current
of 0.27 mA to provide 95 ohm termination to the cable to which the tube
is connected. This stage provides a voltage gain of about 30. Q2 and Q3,
operating as emitter followers, provide the current required to drive the
emitter of Q4 (about 700 mA). The circuit is essentially a one-shot with
an input sensitivity of about -20 mv, and an output fall time of less than
200 ns.
87
-------
oo
oo
1.8 K
-O + 15V
;- IN9I4
O.OOI
I5K
0>
6.8K
2N2907
02
D29F7
TISI29
O + HV
Rs
.0.
03
][
50/A/if
56K
0.001
-I5V
TO CABLE
AND TUBE
040N5
*+•
-V QUENCH
(-5OV OR AS
DESIRED)
IN9I4
Figure 14. Simplified quenching circuit
-------
This circuit has several attractive features, including the ability to
operate over a wide range of quench voltage both positive and negative
without modification. It also has a very rapid response and fall time,
totaling less than 250 ns for a -50 mV input pulse. When it was employed
with tubes which operate "normally" with acceptable results, it provided
no fundamental improvement over the preceding (Figure 13) more complex
circuit. It also failed to operate properly with tubes which do not work
with the more complex circuit. The conclusion drawn from this situation
was that for satisfactory GM tube performance to be realized at high tem-
perature, circuit improvements were no longer required; subsequent im-
provements had to involve the tube per se. All detectors became inoperative
after a maximum of 25 hours of high temperature operation regardless of the
circuit configuration employed. The difficulty encountered in obtaining
long-term high temperature operation was complicated by the fact that the
operating voltage requirement decreases continuously throughout the useful
life of the detector.
At that point in the program it became obvious that the development of a
high temperature beta detection technique based on electronically quenched
GM tubes had to be considered to be rather fundamental and thus outside
the scope of this instrument development program. An optimized quenching
circuit design had been achieved, but a thorough investigation and develop-
ment activity would have to be devoted to the evolution of a stable and
reliable quenchless geiger tube capable of operating at up to 260 C with-
out exhibiting internal electro-chemical degradation over typical periods
of 10 or 20 hours.
The program efforts, by mutual agreement between the EPA and GCA, were then
redirected at a laboratory test demonstration of the overall system operating
at near-room temperatures and using the standard self-quenching type de-
tector tubes.
89
-------
SIGNAL PROCESSING, RECORDING AND CONTROL CIRCUITRY
The principle of mass concentration detection incorporated in this in-
strument and the method of signal processing will now be discussed in
detail. Collection on a continuously moving substrate results in a deposit
mass/area, 6, of:
6 =
where C is the mass concentration of particulates collected at a given
stage
Q is the flowrate
s is the substrate advance velocity
b is the width of the collection trace.
Since the mass/area sensed by beta radiation attentuation is derived from
an exponential absorption process:
f
a = — £n -- (V-2)
Where y is the characteristic absorption coefficient for the beta
emission of the source isotope (0.262 cm2 /rag) for
Carbon-14 betas)
fr and ff are the detected beta rates for the reference and
mass sensors, respectively.
By combining the two preceding equations one obtains:
C-*j.n£
The method of implementing, this function within the multi-stage instrument
under scrutiny was to feed the detected beta pulse train from each detector
of the stage pairs into a frequency-to-voltage converter module whose out-
put is directly proportional to the incoming beta rate (i.e., fr and ff for
90
-------
each stage). The output voltages of each pair of these converters was then
fed into a common log-ratio module whose output voltage Eout is equal to:
Eout - ki loe (v-4)
where kj is an adjustable constant
Er and Ef are the output voltages of the two frequency-to-
voltage converters corresponding to the beta pulses
frequencies fr and ff , respectively.
Thus each collection stage consisted of an independent circuit as described
above from which the individual output signals Eout where obtained and
recorded on a multi-trace strip chart recorder.
It becomes obvious that Equation (V-4) is the electronic analog of the mass
concentration beta detection Equation (V-3) for a moving substrate. Thus,
C = k2 Eout (V'5)
where k2 is a combined proportionality constant incorporating all
the geometrical, substrate transport, flowrate and
absorption coefficient parameters, as well as the con-
version factor from base-ten to natural logarithms.
Therefore, the amplitude of each of the recorded traces is a direct repre-
sentation of the mass concentration of particulates detected at each im-
paction stage.
Figure 15 depicts the schematic diagram of the signal processing circuit
described above. The constant k£ is adjusted electronically for each stage
as a result of the calibration of each collection stage.
Figure 16 is shown to illustrate the typical format of the output signal
from a single stage. This strip chart recording of Eout was obtained
*
Model 8M-3D, MFE Corp., Salem, New Hampshire.
91
-------
STM ECO NO
VO
FREQ TO VOLTAGE
CONVERTER
tISV
O
COMO
FREQ IN
-ISV
O M I
TRIMQ
TELEDYNE .
PHIL BRICK *47O2
TRIMO
TELEDYNE .
PHIL BRICK *47OS
DATE | CHKD|APPROVED
u
| IM
R20
1 — vvv
SK
R2I
IM
R22
IM
n
I
i
i
i
LOG RATIO MODLE
0 / VOLT/DEC
-OKADJ
M3 -is o-
~°I2
-01,
ANALOG DEVICES '756N
R7M~RESISTOR\
REC OUT
-06^
MUX OUT
/77
*I5V
I
/?
1^1 '
I I
^
-I
-I
MOUNTED ON
CONTROL PANEL
iLUN I KUL rANt.Li
SET TO ZERO 756N OUTPUT
FOR EQUAL INPUTS
LIST OF MATERIAL
I
SPECIFICATION
1FTEM
*>•
GCA/TECHNOLOGY DIVISION
K*t>. ttDKm, MUMCHUXCTO OITM / MOT* «17.1?wow
LOG COMPUTATION MODULE
IO75I8
Figure 15. Log computation module
-------
VOLT
v:
60 ~
28cm
J^C.^-lSUBTRAT
Figure 16. Strip chart recording of dust collection run with test "breadboard"
-------
from a dust collection run, impacting test road dust with a 1 mm diameter
nozzle onto a glass-fiber tape moving at a constant speed of 28 cm/sec.
The actual concentration of the dust was not determined during this test
since it was performed by feeding the dust directly from the outlet nozzle
of a Wright dust feeder into the impaction nozzle using the air pressure
of the feeder to provide the flow through the impactor. The recording of
Figure 16 consists of four different sections representing an equal number
of different operating conditions: From left to right: (a) the short
straight recording was obtained with zero input to the frequency-to-voltage
converters (no signal from either beta detector); (b) the next section was
obtained with beta pulse trains to both converters at the same time as the
collection substrate filter tape was advancing at 28 cm/sec; (c) the third
section (near the level 3 on the chart) was obtained by turning on the
dust feeder and running it at a constant feed rate for about 3 minutes and
then shutting it off again; and (d) stopping the substrate tape advance but
feeding the beta pulses into their respective converters.
As can be seen (a) represents the electronic zero of the system and con-
tains no discernible inherent noise at the sensitivity setting of the
recorder. Condition (b) shows the sum of the noise contributions resulting
from statistical beta rate fluctuations and substrate tape thickness in-
homogeneities. Section (c) shows the signal due to dust collection with
the superposition of the noise contributions present in condition (b).
Section (d) solely represents the contribution from the beta radiation
statistical fluctuations, and thus represents the ultimate practical noise
limit of such a system. The amplitude of this noise as well as that of the
fluctuations in section (b) of the recording depend on the time constant of
the electronic signal processing circuitry, in this case about 13 seconds,
and the beta count rate (about 1500 sec"1) which is more or less typical
within this application.
An important design aspect was the electrical connection of the individual
beta detectors to the above discussed electronic signal processing cir-
cuitry (i.e., to the outside of the stack). A single coaxial cable is
94
-------
required to carry both the pulse signal from each detector to the subsequent
circuitry, as well as to supply the detectors with the high voltage they
require for their operation (about 750 V d.c.). This coaxial cable was
selected on the basis of the original high-temperature objective; i.e.,
capability to withstand up to 260°C. A miniature (1/8 inch outside
diameter) teflon jacketed coaxial cable with teflon insulation was selected
for this application (type RG-196A/U) with typical lengths of 5 meters re-
sulting in a shunt capacitance of about 200 pf.
To permit the transmission of the beta pulses to the detection circuit
through these long cables the anode resistor had to be connected directly
to the high voltage electrode of the geiger detector; i.e., leaving a
negligible capacitance to ground between the detector and the anode re-
sistor. As a consequence, an additional component had to be exposed to the
high temperature environment; i.e., the anode resistor (one for each
detector). Such resistors are commercially available for operation up to
275°C, with a typical resistance value of several megohms, and are made
of an oxide resistance film on ceramic core.
The substrate advance drive for all the collection stages consist of a
stepping motor with integral gearing manufactured by Mesur-Matic Electronics
Corporation. This device, in addition to its conveniently low speed output,
has a two-ended shaft design, such that by using a one-way clutch configura-
tion, only one motor was required for the two directions of substrate
motion. The low speed characteristics, furthermore obviates the need for
the somewhat complex reducing gear that had been originally designed into
the upper section of the in-stack head assembly.
The Mesur-Matic stepping motor is driven by a programmable circuit which
provides a predetermined number of pulses to the stepping motor, at a
selected rate. At the completion of the preselected number of pulses, the
direction of the motor is automatically reversed, engaging the clutch at
the opposite shaft end and thus driving the substrates in the opposite
direction. Translational motion of the cassette occurs at the same time by
95
-------
means of a separate stepping motor. At the completion of five parallel
collection traces the driving circuit automatically stops the stepping
motor. An electromechanical pulse counter provides running information on
the length of substrate that has been transported, thus permitting the
interruption of a run, and its subsequent resumption, if so desired, with-
out changing the substrate cassettes before the end of the normal substrate
excursion.
The method of substrate advance incorporating motion reversal has as a con-
sequence the need for electrical inversion of the two signals of each stage
whenever the substrate direction is changed. This function is performed
by electronic gating. These gates are in turn controlled by a circuit
that controls the advance stepping motors. When a predetermined number of
advance steps have been performed the motion is reversed at the same time
as the lateral tape translation is performed. The need to reverse the in-
put connections to the logarithmic ratio modules when direction reversal
takes place arises from the fact that the detector-source pair that func-
tions as reference for one direction becomes the mass sensor in the op-
posite direction, and in order to preserve a positive output signal at the
recording end, the input to the log-ratio converters must be reversed.
Figure 17 shows a front view of the control console. The uppermost section
contains the 8-channel strip chart recorder with its individual controls
(usually adjusted once), and the chart advance speed selector (2.5, 25, and
500 mm/min) to be used in conjunction with the impaction substrate advance
speed selector. The second compartment from the top contains the substrate
indexing driver with the thumb wheel decade selector (0 to 9999) of the
number of advance pulses, the rate selector (0.7, 5, and 45 steps/second
corresponding to 0.42, 3, and 27 cm/min of substrate advance speed), and
start, stop and manual advance controls and indicators. The third compart-
ment contains the electronic signal conditioning, computation, and drive
circuitry for the seven stages of detection with individual calibration
adjustment controls, time constant selector (1, 3, and 10 seconds), and
digital stack temperature display. The bottom section contains all power
supplies with individual switches.
96
-------
-
Figure 17. Front view of control console
97
-------
SECTION VI
INSTRUMENT TESTING
Following the decision to abandon the high-temperature operation objective
of the original program, as a result of the failure to obtain beta detectors
compatible with such elevated temperatures, system test alternatives were
pursued: field testing of a source whose effluent stream temperature did
not exceed about 65 C (150 F), and/or testing at the particulate aerodynamic
test facility at the EPA Research Triangle, N. C.
SURVEY OF FIELD TEST FACILITIES
The first alternative was pursued very diligently by researching the Plant
Approval listings of industrial sources in the State of Massachusetts Public
Health Department archives. The resulting survey is listed in detail in
Table 11. It has been deemed useful to incorporate this rather exhaustive
listing in this report for future information purposes.
Although a significant number of the listed organizations were contacted
with respect to the location of an adequate test site for the instrument
developed under this program, only two companies agreed, in principle, to
allow such testing to be performed — Rosenfeld Co., a cement manufacturer,
and Johns-Manvilie, an insulation material plant. Because of its geo-
graphical proximity, the latter organization was considered as an attrac-
tive candidate as potential test site. GCA personnel visited Johns-Manvilie
(North Billerica, Mass.) and assessed this facility from the point of view
of compatibility with the operation of the multistage-beta-impactor.
98
-------
Table lla. DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
VO
VO
Company name
Am. Optical - Lens Plant
Bay State Abrasives
Benevento Sand & Gravel
Crown Wire & Cable
General Tire & Rubber
Johns-Manville
Lee Lima •
N. Attileboro Foundry
Iyer Rubber
Borden Chemical
Ideal Tape
ITT Surprenant Div.
Kessell & Horse
Matlack
Monsanto
New England Foundry
Revere Copper and Brass
Texas Instruments
U.S. Gypsum
Wyman-Gordon
Avon Sale
GAF
General Tire & Rubber
G&U
Nat. Polychemical
Trow Corp.
Bandy & Haroan
Approval
year
1975
1973
1975
1975
1975
1975
1975
1975
1975
1974
1974
1974
1974
1974
1974
1974
1974
1974
1974
1974
Company address
14 Mechanic, Southbridge
Union St., Westboro
Rt. 62, North Wilmington
37 Cashman, Taunton
General St., Reading
High St. , North Billerica
Marble St . , Lee
103 Clayton, Dorchester
10 Railroad St., Andover
Old Clark Rd, N. Andover
1400 Middlesex, Lowell
172 Sterling St., Clinton
242 Canterbury, Worcester
•Beldertown
190 Grochnal, Indian Orchard
10 Embankment, Lawrence
24 North Front, New Bedford
34 Forest, Attleboro
200 Terminal, Char lest own
105 Madison, Worcester
1973 E. Spring, Avon
1973
1973
1973
60 Carver, Mill is
70 Guichon , Lawrence
Summer Rd, Boxboro
1973 i Wilmington
1973 River Rd, E. Drenfield
1972 . Frank Massberg Dr., Attleboro
Company
phone No.
366-4431
658-4762
823-1731
944-1540
663-3401
(413) 243-9953
825-6500
475-5300
686-9591
752-1901
(413) 788-6911
688-1811
999-5601
241-9100
756-5111
587-4180
376-2661
682-5121
226-1000
Person
contacted
Dan James
Way land
Mydlack
Process operation
Spraying gold paint
Machining of grinding wheels
Stone crushing
Silo loading
Dispersion of resins and pigments
Manuf. troughs
Limestone grinding & screening
Cast iron foundry
Rubber covering of rolls
Chemical transfer
Rubber chopper
Compounding powders
Unloading cement truck
Cement transfer
Chemical drying
Nonferrous foundry
Nonferrous melting
Abrasive sandblast
Crushing & grinding
Grinding
Bafflng rubber
Production of asphalt roofing
Coating
Coating with vinyl plastics
Flash drying
Rock crushing
Gaseous chlorination of molten gold
-------
Table lla (continued). DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
o
o
Conpany name
Lawrence Ready-Mix Concrete
Lynn Sand & Stone
W. R. Grace
Washington Mill Abrasive
Neponser Brass Foundry
New England Concrete
Burnett Bros.
Wrenthan Sand & Gravel
Rosenfeld
N.E. Concrete Pipe
Old Colony Crushed Stone
Whittemore Perlite
Mass Brokes Stone
Fletcher Granite
W. R. Grace-Dewey & Almy
Chemical Division
Monsanto
Borden Chemical
Approval
year
1972
1972
1971
1971
Company address
Gifford St., Falmouth
30 Danvers, Swampscott
College Rd, Easthampton
20 Main, North Grafton
115 Boston, Dorchester
Cambridge
1 Chemical, Everett
103 Foster, Peabody
Company
phone No.
595-0820
839-4426
436-2563
969-0220
325-8600
384-3138
470-0317
893-0489
251-4031
876-1400
397-5010
531-2222
Person
contacted
Groleau
San LaRousa
Bob Barton
Gus Sheehy
Roy LaBoug
Process operation
Rock crushing
Dry solids mixing
-------
Table lib. DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
Company name
Am. Optical - Lens Plant
Bay State Abrasives
Benevcnto Sand & Gravel
Crown Wire & Cable
General Tire & Rubber
Johns-Manville
Lee Lima
N. Attleboro Foundry
Tyer Rubber
Borden Chemical
Ideal Tape
ITT Surprenant Dlv.
Kesscll f. Morse
Matlack
Monsanto
New England Foundry
Revere Copper and Brass
Texas Instruments
U.S. Gypsum
Wyman-Gordon
Avon Snle
CAF
General Tire & Rubber
G&W
Nat. Polychemlcal
Trow Corp.
Handy & Harman
System
flow,
scfm
1,275
11.7 k
22.5 k
7,650
4,500
5 k
15 k
10 k
4 k
8 k
5,500
6,600
900
1,200
16 k
12 k
85 k
1,500
2,500
11,800
4,500
16,400
15-20 k
4 k
1,534
48 k
18 k
Temperature
inlet/outlet,
0F
70
70
Ambient
50/75
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
Ambient
80 - 90
93
110
Ambient
140
70
70
Ambient
130/120
140
165/150
Ambient
106
Concentration
Inlet/outlet,
gr/ft3
0.012/0.0055
14.4/0.144
10/0.05
0/0.0198
10/0.02
0.25/0.003
0/0.03
48/0.0043
700/0.0035
5.5/0.055
4.67/0.0047
3/0.003
13.12/0.018
6.9/0.14
0.11/0.016
0.058
0.095/0.0005
0/0.0015
10/0.01
0.207/0.023
6/0.06
2.5/0.002
0.023/0.014
0.0036/0.0002
40.33/0.04
10.6/0.03
1.5/0.04
Z
< 10 um
30
2
100
100
33
10
34
75
30
95
90
99
30
100
Discharge
height,
£t
22
15
30
25
45
30
5
10
16
40
35 - 40
64
36
87
24
12
10
39
Stack
outlet
dimensions
1.33 ft dia.
2 ft x 3 Tt
4 ft fll.-i.
20 in. dhi.
16 in. dia.
30 in. dia.
16-5/8 in.x20-l/a in.
0.92 ft x 1.08 ft
13 in. x 14-3/8 in.
19.5 ft2
19 In. dia.
5.1 ft dia.
23 in. x 30 in.
3 ft dia.
20 In. dia.
1.0 in. dia.
40 in. x 32 in.
3 ft dia.
-------
Table lib (continued). DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
O
N3
Company name
Lawrence Ready-Mix Concrete
Lynn Sand & Stone
W. R. Grace
Washington Mill Abrasive
Neponset Brass Foundry
New England Concrete
Burnett Bros.
Wrenthan Sand & Gravel
Rosenfeld
N. E. Concrete Pipe
Old Colony Crushed Stone
Whlttetnore Perltte
Mass Brokes Stone
Fletcher Granite
W. R. Grace-Dewey & Almy
Chemical Division
Monsanto
Borden Chemical
System
flow,
scfm
500
80 k
5 k
6,280
Temperature
inlet/outlet,
oF
Ambient
Ambient
80
60
Concentration
inlet/outlet,
gr/ft3
3.3/0.0033
5/0.03
0.035/0.0035
0.1/0.0001
. %
< 10 lira
28.7
Discharge
height
ft
60
20
40
fi
Stack
outlet
dimensions
2 ft*
4.1 ft x 4.1 ft
2 ft dia.
13 in. x 3fi in.
-------
Table lie. DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
Company name
Am. Optical - Lens Plant
Bay State Abrastves
Benevento Sand & Gravel
Crown Wire & Cable
General Tire & Rubber
Johns-Manville
Lee Lima
N. Attleboro Foundry
Tyer Rubber
Borden Chemical
Ideal Tape
ITT Surprenant Div.
Kesseli & Morse
Matlack
Monsanto
New. England Foundry
Revere Copper and Brass
Texas Instruments
U.S. Gypsum
Wyman-Gordon
Avon Sale
GAF
General Tire & Rubber
G&W
Nat. Polychemical
Trow Corp.
Handy & Harman
Date
site
contacted
N. C.
6/16/75
6/16/75
6/16/75
6/16/75
6/11/75
N. C.
N. C.
N. C.
N. C.
6/11/75
6/11/75
6/11/75
N. C.
N. C.
6/11/75
6/11/75
Site response
No - see letter
No
No
Maybe - send letter
No stack
Only hot stacks
Corporate offices only
All stack temp. > 200°F
Wet collectors
Not manufacturing any more
Why site was not contacted
More liquid than particulate
Concentration too high or low
Concentration too high or low
10 percent < 10 um
34 percent < 10 um
Low flow
Concentration too low
Concentration too high or low
-------
Table lie (continued). DESIGN SHEET FOR INDUSTRIAL AND COMMERCIAL FACILITIES
Company name
Date
sice
contacted
Site response
Why site was not contacted
Lawrence Ready-Mix Concrete
Lynn Sand & Stone
W. R. Grace
Washington Mill Abrasive
Neponset Brass Foundry
New England Concrete
Burnett Bros.
Wrenthan Sand & Gravel
Rosenfeld
H. E. Concrete Pipe
Old Colony Crushed Stone
Whittemore Perlite
Mass Brokes Stone
Fletcher Granite
W. R. Grace-Dewey & Almy
Chemical Division
Monsanto
Borden Chemical
No
No
No stack
No
Maybe
Oil combustion only
Phone out - out of business?
Too hot (350-500°F)
No
-------
The particulates to be sampled are generated by the sanding of sheets of
insulation (Marinite). The sanding operation is intermittent and large
variations in dust loading were anticipated. The approximate particulate
concentration was calculated from the estimated mass collected by the bag-
house of 9000 kg per 6-1/2 hours of operation with the flow rate through
the baghouse of 17 m3/sec. The calculated upstream concentration is
3 3
9.97 grains/ft or 22.83 grams/m . Since the process varies considerably,
it was considered possible to sample when the concentration is lower.
There were three sampling locations which were considered. One was lo-
cated inside the work area requiring approximately 20 feet of scaffolding.
This location had the advantage of being inside so the equipment would be
protected. The other locations were on the roof. One of them was located
in a horizontal section leading to the cyclone with the advantage of easy
access. The other, located in a short section between the cyclone and the
baghouse, had the advantage of a lower particulate concentration. None
of the sampling locations was found to be ideal from set up considerations
because of problems of supporting the sampling head in the duct. Since the
effort required to adapt the instrument to any of the above sampling sites
was considered excessive, the decision was reached to concentrate all ef-
forts on the second alternative mentioned above; i.e, testing at the EPA
facilities. It should be noted that the personnel at Johns-Manvilie were
exceptionally cooperative and helpful during the survey visit performed by
GCA personnel.
INSTRUMENT TESTING AT THE EPA
An intensive effort was devoted to prepare the instrument for a test se-
quence scheduled for mid-August 1975. The entire electronics subsystem
was debugged, tested and repaired as necessary, as well as adjusted to
provide a proper output for the multichannel strip-chart recorder. A
number of minor, but time-consuming problems had to be overcome within
this effort, such as detector cable cross-talk interference, component
"infant mortality," etc.
105
-------
The other major aspect that had to be resolved was the problem of impaction
substrate transport and coating. When all seven cassettes were loaded it
was found that the resulting friction prevented normal translational motion
of the substrate tape. It was found necessary to replace the existing
helical gears, which provided the 90 drive from the support member to the
collector array, by a set of worm gears with a transmission ratio of 1 to
10. This implied that the stepping motor had to be driven 10 times as fast
as before to accomplish the same translational speed. Consequently, the
motor drive circuit had to be modified.
The substrate tape used on the instrument for the room temperature tests
performed at EPA was 0.036 cm thick Mylar film coated with petroleum jelly
as particle adhesive. The substrate tape is coated before it is rolled up
on the supply reels, and the desired coating thickness is obtained by the
adjustment of the tension of the tape as it is transported to the take-up
reel. Coating uniformity was improved by using a spreader pad at the
place where the tape unrolls.
Upon arrival at EPA it was found that the instrument had been shaken up
considerably in shipment. Although some cards had been jarred from their
sockets and many other elements were loose, the instrument was put into
operational condition with little difficulty.
3
On the first full day of testing, fly ash was used at about 0.1 g/m .
Initially, it was found that the instrument would not respond adequately
in the continuously advancing substrate mode due to a high level of "noise"
caused by the unevenness of the coating. As a result of this difficulty,
a regime was established for making stationary impactions, then pulling
the film under the mass sensing detector by advancing the tape a specific
amount with the extraction pump inoperative. A sampling time of about 5
minutes was found to give the most satisfactory results.
3
Next, high level tests were run (1 g/mm ); again continuous operation gave
very marginal results. Therefore, stationary impaction was again used
106
-------
during this initial effort. Following this set of tests with fly ash as
planned, the system was changed over to Fe.O,. While this was being done,
the collection films were removed and recoated. The effort to obtain sat-
isfactory operation of the wind tunnel with iron oxide particles proved
quite unsuccessful due to the inability of the dust feeder to handle this
material.
When testing was begun again using fly ash, it was found that the newly
coated substrate tapes gave a much lower noise level (about an order of
magnitude) than those which had been used before. This fact, in combina-
tion with the increase in circuit gain which was also made during the shut
down, allowed the instrument to respond satisfactorily in the continuous
mode with a reasonable signal to noise ratio even at somewhat reduced con-
3
centrations (0.3 to 0.4 gm/m ). The drastic improvement in operation re-
alized after substrate film recoating was attributed to the achievement of
a very significantly improved coating smoothness. It was theorized that
during shipment from Bedford, Mass, to the Research Triangle, N.C., suc-
cessive melting and solidification of the petroleum jelly coating on the
rolled up substrate film (the instrument was shipped during the beginning
of the month of August) resulted in large scale inhomogeneities of the
coating with ridges and troughs causing the unacceptably high noise level
observed during the initial tests when operating the system with continuous
substrate motion. The mass concentration sensing calibration constants
for the recording trace for each of the seven channels, based on their
collection geometries, the operating flow rate, the substrate advance
speed used during these tests (0.42 cm/min), etc., are listed below for
the two operating modes; i.e., stationary and moving substrate.
It is apparent that the sensitivity increases as the impaction orifice (and
particle size) gets progressively smaller. This effect is greater for the
stationary impaction than for the continuously moving impaction substrate.
107
-------
Stage
1
2
3
4
5
6
7
Stationary
5 minute
impaction
7.7 mg/M div
4.2 mg/M3 div
2.1 mg/M3 div
1.04 mg/M3 div
0.57 mg/M div
0.325 mg/M3 div
0.184 mg/M3 div
Continuous impaction
with 0.7 step/sec
tape drive rate
4.6 mg/M3 div
3.4 mg/M3 div
2.4 mg/M3 div
1.7 mg/M div
1.3 mg/M div
1.1 mg/M div
0.71 mg/M3 div
Since the particle distribution of fly ash (and many other particle types)
is weighted quite markedly towards the larger particles, response falls off
in the smaller stages despite the indicated increase in sensitivity. Due
to this fact, the stationary impaction scheme normally will show more
stages with an observable response than the continuous scheme, since the
smallest stages may have insufficient sensitivity to give an indication in
the continuous mode for low mass concentrations.
The instrument was run at a flowrate of 9 liters/minute and the correspond-
ing particle cut-off sizes for room temperature operation were: 6.5, 4.0,
2.4, 1.44, 0.90, 0.57 and 0.35 micrometers, aerodynamic diameter. The
tests were run in conjunction with the Research Triangle Institute, con-
tracted by the EPA to operate the test facilities at the Research Triangle
Park. The results of these tests, wherein other instruments were evaluated
as well, are presented and discussed within a pertinent report issued by
99
RTI. These data and accompanying discussion will not be repeated here
as they are available as part of that report. Some of the conclusions and
relevant data of that report, however, will be reviewed and reexamined here.
Reference will be made to Figures 1, 2, 12 and 13 of the RTI report, as
well as to the accompanying text.
108
-------
Figures 12 and 13 of the referenced report are plots of the size distribu-
tion as determined by a reference Brink-type cascade impactor and the data
points obtained during the normal continuous substrate motion operation of
the GCA instrument. The two graphs were obtained at two different mass
3
concentrations, 0.267 and 0.955 g/m . These two plots are shown in
Figures 18 and 19 of the present report, the circles representing the
original data as presented in the RTI report. It is obvious that from
the observations made by RTI of the very large and unexpected effect of
upstream obstructions in the test tunnel on downstream mass concentrations,
the measurements performed with the GCA seven-stage instrument were totally
distorted and that any conclusions reached without taking into consideration
the obstruction effect are, at best, questionable. As indicated in the
RTI report the inlet of the GCA device was only about 2m downstream of the
Brink impactor since the latter was inserted in port D and the former in
port F of the test tunnel (see Figure 1 of the RTI report). At the op-
erating tunnel velocity of 9.1 m/s, used for the GCA versus Brink tests,
Figure 2 of the referenced report shows that the relative centerline con-
centration in the tunnel dropped to about 28 percent of the upstream or
unobstructed concentration value for an obstruction 3m upstream of a sensor.
Figure 20 of the present report is a replot of Figure 2 of the RTI document
from which it is possible to find the extrapolated concentration fraction
for an obstruction 2m upstream of the GCA instrument. This extrapolation
is in all probability quite accurate since the three actual data points used
fall very close to a straight line. It is thus found from Figure 20 that
for the actual test conditions under which the GCA instrument was operated
the mass concentration entering this instrument was probably only about
19 percent of the concentration measured by the reference Brink impactor.
Thus, it is felt that all the data points of Figures 12 and 13 of the RTI
report labeled "GCA data" should be multiplied by 5.3 (i.e., 1/0.19) in
order to compensate for the large obstruction effect on the centerline mass
concentration. When this correction is applied to the points of the RTI
Figures 12 and 13, as shown by the crosses of Figures 18 and 19 of the
present report, the agreement between the Brink curves and the GCA data
3
becomes quite good, especially for the 0.955 g/m concentration test. The
109
-------
10'
o>
s
01
o
o
•o
O
i 10
CD
CO
5
U
CL
Id
-2
10
10"
O.I
O GCA DATA
x CORRECTED
GCA DATA
1.0 , 10
PARTICLE DIAMETER, Dg*o,Mm
100
Figure 18. Comparison of GCA in-stack beta impactor data with corre-
sponding Brink curve (concentration = 0.267
110
-------
J§ io-
0<
o>
o
o
•o
•^
o
h-
:D
OJ
a:
H
(O
o
a:
Lul
10
-2
10
-3
O GCA DA IA
x CORRECTED
GCA DATA
1.0 TO
PARTICLE DIAMETER, Dgeo.^m
100
Figure 19. Comparison of GCA in-stack beta impactor data with corre-
sponding Brink curve (concentration = 0.955
111
-------
H1
M
N3
o
z
UJ
o
z
o
o
UJ
>
-------
problem associated with the first collection stage, mentioned in the RTI
report remains significant but far less staggering than the uncorrected
data indicated.
It is felt that in light of the fact that these were the first series of
tests to which this instrument was subjected, its performance was quite
acceptable and its potential performance should be considered promising.
113
-------
SECTION VII
REFERENCES
1. Sem, G. J., et al., "State of the Art: 1971. Instrumentation for
Measurement of Particulate Emissions from Combustion Sources,"
Report to EPA on Contrct No. CPA-70-23 (1971).
2. Horn, W., "Process for Continuous Gravimetric Determination of the
Concentration of Dust-Like Emissions," Staub (English), 2&_ No. 9,
20-24 (1968).
3. Dresia, H., Fischotter, P., and Felden, G., "Kontinuierliches Messen
des Staubgenaltes in Luft and Abgasen mit Betastrahlen," V.D.I. -
Z., 106 No. 24, 1191-1195 (1964).
4. Nader, J. S. and Allen, D. R., "A Mass Loading and Radioactivity
Analyzer for Atmospheric Particulates," Amer. Indus. Hyg. Assoc.
J., 21, 300-307 (1960).
5. Denzel, P. and Horn, W., "Ein Empfindliches Messverfahren zur
Kontiriuierlichen Bestimmung der Gravimetrischen Konzentration von
Staubformigen Emissionen," Electrotechnische A. PT. A., 87^ No. 9,
311 (1966).
6. Aurand, K. and Bosch, J., "Recording Counter for the Continuous
Determination of the Concentration of Pulverulent Air Pollutants,"
Staub (English), 27/10), 21-24 (1967).
7. Benarie, M. and Loverdo, A., "Pesee Automatique des Pollutants
Atmospheriques par Jauge Beta," Mesures, Regulation et Automisme,
^2 No. 9, 90-95 (1967).
8. Lilienfeld, P., "Beta Absorption Mass Monitoring of Particulates,
A Review," Paper No. 71-1031, Joint Conference on Sensing of Environ-
mental Pollutants, Palo Alto, Calif. (1971).
9. Herling, R., "A Beta Gauge and Filter Collection System for the
Determination of Automobile Particulate Emissions," Paper No.
71-1032, Joint Conference on Sensing of Environmental Pollutants,
Palo Alto, Calif. (1971).
114
-------
10. Izmailov, G. A., "Measuring the Gravimetric Concentration of Dust
in the Air Using Beta Radiation," Industrial Laboratory (English
Trans.), 27_. 40-43 (1961).
11. Bulba, E. and Silverman, L., "A Mass Recording Stack Monitoring
System for Particulates," Paper No. 65-141, 58th Annual Meeting
of the Air Pollution Control Assoc., Toronto, Canada (1965).
12. Lilienfeld, P. and Dulchinos, J., "Vehicle Emission Mass Monitor,"
Final Report, Contract No. 68-02-0209 (1971).
13. Abbott, C. E., Dye, J. E., and Sartor, D. T., "An Electrostatic
Cloud Droplet Probe," J. Appl. Meteorol., 11, 1092-1100, (1972).
14. Adams, N. G. and Smith, D., "Studies of Microparticle Impact
Phenomena Leading to the Development of a Highly Sensitive Micro-
meteoroid Detector," Planet. Space Sci., 19, 195-204 (1971).
15. Anonymous, "Air Quality Criteria for Particulate Matter," U.S. Dept.
H.E.W., NAPCA Publ. AP-49 (1969).
16. Babich, V. G., et al., "Instrument for Determining the Weight and
Active Concentration of Aerosols," U.S. Patent No. 3,558,884
(Jan. 26, 1971).
17. Bahco, A. P., "Particle Detector," Brit. Pat. Abst., £(49) (1969).
18. Ball, D. F. and Griffiths, D. F., "Measurement of Dust Concentration
and Size in Ducts and Chimneys," Environ. Health, 78, 561-566 (1970).
19. Benarie, M. and Quetier, J. P., "Etude d'une Methode Microdynamome-
trique Pour la Mesure de la Concentration des Poussieres," J. Aerosol
Sci., I, 77-109 (1970).
20. Binek, B., et.al., "Using the Scintillation Spectrometer for Aerosols
in Research and Industry," Staub (English), 27_ No. 9, 1 (1967).
21. Bird, A. N., Jr., McCain, J. D., and Harris, D. B., "Particulate
Sizing Techniques for Control Device Evaluation," Paper 73-282,
A.P.C.A. Meeting Chicago, 111. (1973).
22. Brink, J. A., "Cascade Impactor for Adiabatic Measurements," Indus. &
Engrg. Chem., _50, 645-648 (1958).
23. Brink, J. A., Kennedy, E. D., and Yu, H. S., "Particle Size Measure-
ments with Cascade Impactors," 65th Annual Meeting, A.I.Ch.E.,
New York (1972).
115
-------
24. Brownell, L. E., "Radiation Uses in Science and Industry," Univer-
sity of Michigan, 30 (USAEC) (1961).
25. Brusset, H. and Donati, J. R., "Small Angle X-Ray Scattering for
the Determination of Particle Size Distribution," J. Appl.
Crystallogr., 2, 55-64 (1969).
26. Carpenter, T. E. and Brenchley, D. L., "A Piezo-Electric Cascade
Impactor for Aerosol Monitoring," Am. Ind. Hyg. Assoc. J., 33,
503-510 (1972).
27. Chuan, R. L., "An Instrument for the Direct Measurement of Par-
ticulate Mass," J. Aerosol Sci., _!, 111-114 (1970).
28. Coenen, W., "Bin Neues Messverfahren Zur Beurteilung Fibrogener
Staube am Arbeitsplatz," Staub (German), 33 No. 3, 99-103 (1973)
29. Cooper, D. W. and Spielman, L. A., "A New Particle Size Classifier:
Variable-Slit Impactor with Photo-Counting," ASME Symposium on
Flow Studies in Air and Water Pollution, Georgia Inst. of Technol.,
Atlanta, Ga. (June 1973).
30. Crandall, W. A., "Determining Concentration and Nature of Particulate
Matter in Stack Gases," ASME Winter Annual Meeting, Washington, D.C.
(Nov. 1971).
31. Davies, C. N., "The Entry of Aerosols into Sampling Tubes and
Heads," Brit. J. Appl. Phys., Ser. 2, !_, 921-932 (1968).
32. Davies, R., "Particle Size Analysis," Indus. Engrg. Chem., 62,
87-93 (1970).
33. Dempsey, J. C. and Polishuk, P., Radio Isotopes for Aerospace,
Part 2, Plenum Press, New York (1966).
34. Dorsey, J. A. and Burckle, J. 0., "Particulate Emissions and Process
Monitors," Chem. Eng. Prog., 67_, 92-96 (1971).
35. Downs, W. and Strom, S.S., "New Particle Size Measuring Probe -
Application to Aerosol Collector and Emissions Evaluation,"
J. Engineering for Power, 117-126 (1972).
36. Dresia, H., Fredrick, W., and Vogel, J., "Untersuchungen zur
Bestimmung des Aschengehaltes in Kohle mit 3-und y~ Strahlen,"
Brennstoff-Chemie, 43_ No. 5, 27-30 (1962).
37. DUwel, L., "Latest State of Development of Control Instruments for
the Continuous Monitoring of Dust Emissions," Staub, ^8_ No. 3,
42-51 (1968).
116
-------
38. Environmental Protection Agency, "Standards of Performance for New
Stationary Sources," Federal Register, 36^ No. 247, 24876 (Dec. 23,
1971).
39. Environmental Protection Agency, "Appendix B: Reference Method for
the Determination of Suspended Particulates in the Atmosphere,"
Federal Register, 36. No. 84, 8191 (Apr. 30, 1971).
40. Friedlander, S. K., "The Characterization of Aerosols Distributed
with Respect to Size and Chemical Composition - II," J. Aerosol
Sci., ^, 331-340 (1971).
41. Fuchs, N. A., The Mechanics of Aerosols, The Macmillan Company,
New York (1964).
42. Fuchs, N. A., "An Attempt at a Classification of Aerosol Literature,"
Staub (English), 28_ No. 9, 1-6 (1968).
43. Cast, T., "Dust Samplers with Mass-Proportional Indication of
Recording," Staub, 20(8), 266-272 (1960).
44. Cast, T., "Acoustical Feedback as Aid in the Determination of Dust
Concentrations by Means of an Oscillating Ribbon," Staub, 30 No. 6,
1-5 (1970).
45. Gerber, H. E., "On the Performance of the Goetz Aerosol Spectro-
meter," Atmos. Environ., 5^ 1009-1031 (1971).
46. Goetz, A., "An Instrument for the Quantitative Separation and Size
Classification of Airborne Particulate Matter Down to 0.2 Microns,"
Geofis. Pura Appl., ^6_, 49-69 (1957).
47. Goldschmidt, V. W., "Measurement of Aerosol Concentrations with a
Hot Wire Anemometer," J. Colloid Sci., ^0, 617-634 (1965).
48. Goldschmidt, V. W. and Householder, M. K., "Hot Wire Anemometer as
an Aerosol Droplet Size Sampler," Atmos. Environ., 3^ 643 (1969).
49. Green, H. L. and Lane, W. R., Particulate Clouds; Dusts, Smokes,
and Mists, D. Van Nostrand Co., Princeton, N.J. (1964).
50. Gussman, R. A. and Gordon, D., "Notes on the Modification and Use
of a Cascade Impactor for Sampling in Ducts," Am. Ind. Hyg. Assoc.
J., 27_, 252-255 (1966).
51. Hinds, W. and Reist, P. C., "Aerosol Measurement by Laser Doppler
Spectroscopy," J. Aerosol Sci., 3, 501-527 (1973).
117
-------
52. Hochrainer, D., "On the Reliability of Measurements with the Goetz
Aerosol Centrifuge," Atmos. Environ., 6^, 699 (1972).
53. Hodkinson, J. R., "The Literature on Aerosols," Powder Technol.,
I, 51-53 (1967).
54. Hoffman, K. P. and Mohnen, V., "The Operation of the Acoustic Par-
ticle Counter," Staub, 28 No. 9, 15-19 (1968).
55. Holzhe, J. and Demmrich, H. Neue Hue He, 14_, 74 (1969).
56. Horvath, H. and Rossano, A. T., "Technique for Measuring Dust
Collector Efficiency as a Function of Particle Size," J. Air Pollut.
Control Assn., ^0, 244 (1970).
57. Hounam, R. F. and Sherwood, R. J., "The Cascade Centripeter: A
Device for Determining the Concentration and Size Distribution of
Aerosols," Am. Ind. Hyg. Assn. J., 2£, 122-131 (1965).
58. Keiley, D. P., "An Airborne Cloud Drop-Size Distribution Meter,"
J. Meteorol., 17_, 349 (1960).
59. Rnollenberg, R. G., "Particle Size Measurements from Aircraft Using
Electro-Optical Techniques," Proc. Tech. Prog., Electro-Optical
Systems Design Conf., West Anaheim, Calif. (1971).
60. Lapple, C. E., "Particle-Size Analysis and Analyzers," Chem. Engr.,
149 (20 May 1968).
61. Leith, D. and Mehta, D., "Cyclone Performance and Design," Atmos.
Environ.,. 7_, 527-549 (1973).
62. Lieberman, A., "Particle Detection and Monitoring - Current State
of the Art," Powder Technol., I, 317 (1967/68).
63. Lilienfeld, P., "Beta Absorption - Impactor Aerosol Mass Monitor,"
Am. Indus. Hyg. Assn. J., 31., 722 (1970).
64. Lilienfeld, P. and Doyle, A., "Aerosol Particle Monitor," U.S.
Patent No. 3,711,707 (Jan. 16, 1973).
65. Lippmann, M., "Respirable Dust Sampling," Am. Indus. Hyg. Assn. J.,
31, 138 (1970).
66. Liu, B. Y. H., Marple, V. A., and Yazdani, "Comparative Size
Measurements of Monodisperse Liquid Aerosols by Electrical and
Optical Methods," Environ. Sci. and Technol., 3, 381-386 (1969).
118
-------
67. Lundgren, D. A., "An Aerosol Sampler for Determination of Particle
Concentration as a Function of Size and Time," J. Air Pollution
Control Assoc., Yl_, 225 (1967).
68. Marple, V. A., "A Fundamental Study of Inertial Impactors," Thesis,
University of Minnesota, Minneapolis, Minn. (1970).
69. Martin., J. W. and Stewart, R. F., "Determination of the Incom-
bustible Content of Mine Dust by a Nuclear Method," U.S. Dept. of
Interior, Bureau of Mines Report No. 7193 (1968).
70. McShane, W. P. and Bulba, E., "Automatic Stack Monitoring of a Basic
Oxygen Furnace," J. Air Pollut. Control Assoc., 18, 216 (1968).
71. Mercer, T. T. and Chow, H. Y., "Impaction from Rectangular Jets,"
J. Colloid & Interface Sci., 27^ 75-83 (1968).
72. Mercer, T. T. and Stafford, R. G., "Impaction from Round Jets,"
Annals of Occup. Hyg., 12_, 1 (1969).
73. Mitchell, R. I. and Engdahl, R. B., "A Survey for Improved Methods
for the Measurement of Particulate Concentration in Flowing Gas
Streams," J. Air Pollut. Control Assoc., 558-563 (1963).
74. Mitchell, R. I. and Pilcher, J. M., "Improved Cascade Impactor for
Measuring Aerosol Particle Sizes," Indus. & Engrg. Chem., 51,
1039-1042 (1959).
75. Nader, J. S., "Developments in Sampling and Analysis Instrumentation
for Stationary Sources," J. Air Pollut. Control Assoc., 23, 587-591
(1973).
76. Noll, K. E., "A Rotary Inertial Impactor for Sampling Giant Par-
ticles in the Atmosphere," Atmos. Environ., 4_, 9 (1970).
77. Olin, J. G. and Sem, G. J., "Piezoelectric Microbalance for Monitoring
the Mass Concentration of Suspended Particles," Atmos. Environ.,
.5, 653-668 (1971).
78. Olin, J. G., Sem, G. J., and Christenson, D. L., "Piezoelectric
Aerosol Mass Monitor," Amer. Indus. Hyg. Assoc. J., 32, 209-220
(1971).
79. Pilat, M. J., Ensor, D. S. and Bosch, J. C., "Source Test Cascade
Impactor," Atmos. Environ, 4^ 671-679 (1971).
80. Pilat, M. J., Ensor, D. S., and Bosch, J. C., "Cascade Impactor for
Sizing Particulates in Emission Sources," Amer. Indus. Hyg. Assoc. J.,
32, 508-511 (1971).
119
-------
81. Pueschel, R. F. and Noll, K. E., "Visibility and Aerosol Size
Frequency Distribution," J. Appl. Meteorol., 6^ 1045 (1967).
82. Schleusener, S. A., "Automatic High-Speed Particle Counting Using
a Gas Laser," Rev. Sci. Instr., 39U2), 1916 (1968).
83. Schultz, A., "Einer Anbrdnung zur Registrierenden Kontaktelektrischen
Staubmessung," Staub, 24, 359-363 (German) (1964).
84. Sehmel, G. A., "Particle Sampling Bias Introduced by Anisokinetic
Sampling and Deposition Within the Sampling Line," Amer. Indus. Hyg.
Assoc. J., 3a, 758-771 (1970).
85. Shumilovskii, N. N. and Milttser, L. V., Radioactive Isotopes in
Instrumentation and Control. The MacMillan Company, New York (1964).
86. Sto'ber, W., "A Note on the Aerodynamic Diameter and Mobility of Non-
Spherical Aerosol Particles," J. Aerosol Sci., 2^ 453-456 (1971).
87. Sto'ber, W. and Boose, C., "Developing Flow and Particle Deposition
in Horizontal Elutriators and Semi-Dispersive Aerosol Centrifuges,"
Atmos. Environ., 7_, 119-130 (1973).
88. Sttfber, W. and Flachsbart, H., "Size-Separating Precipitation of
Aerosols in a Spinning Spiral Duct," Environ. Sci. Technol., 3^
1280-1296 (1969).
89. Sto'ber, W. and Zessack, J., "Zur Theorie Einer Konischen Aerosol-
zentrifuge," Staub, 24, 295-305 (1964).
90. Strauss, W., Air Pollution Control, Wiley, New York (1970). .
91. Turner, G. A. and Fayed, M. E., "Sedimentation Rates of Fine Powders
in Air at Different Pressures," Fine Particle Soc. Seminar, Chicago,
111. (June 1970).
92. Whitby, K. T. and Clark, W. E., "Electric Aerosol Particle Counting
and Size Distribution Measuring System for the 0.015 to 1 ym Size
Range," Tellus, 18_, 573-586 (1966).
93. O'Kelley, G. D., "Detection and Measurement of Nuclear Radiation,"
NAS-NS3105, April 1962.
94. Ettinger, K. W. and Zastawny, A., "A New Fast Quenching Circuit for
Self-Quenched Geiger-MUller Counters," J. Sci. Instrum. Vol. 42,
p. 125 (1965).
95. Eckey, H., "A Quenching Circuit of Very Short Rise Time for Geiger-
MUller Counters," J. Sci. Instrum., Series 2, Vol. 2, p. 1121 (1969).
120
-------
96. Karasik, A. S. and Prokofiev, 0. N., "Solid-State Construction
for Discharge Quenching in Geiger Counters," Pribori i Tekhnika
Eksperimenta, No. 4, pp. 61-63 (July-August 1970).
97. Glascock, H. H., Jr. and Webster, H. F., "A Flame Sensor for High-
Temperature Ambients," IEEE Transactions on Electron Devices,
Vol. ED-18, No. 5, p. 330 (1971).
98. Jernakoff, G., private communication to P. Lilienfeld, August 6,
1974.
99. Gooding, C. H., "Wind Tunnel Evaluation of Particle Sizing Instru-
ments," Report No. EPA-600/2-76-073, March 1976.
100. Lilienfeld, P. and Dulchinos, J. "Portable Instantaneous Mass
Monitor for Coal Mine Dust." Amer Ind Hyg Assoc J. 33:136. 1972.
121
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing}
1. REPORT NO.
EPA-600/2-77-077
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Design, Development, and Demonstration of a Fine
Particulate Measuring Device
5. REPORT DATE
April 1977
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Pedro Lilienfeld, Daniel P. Anderson, and
Douglas W. Cooper
8. PERFORMING ORGANIZATION REPORT NO.
GCA-TR-76-27-G
9. PERFORMING ORGANIZATION NAME AND ADDRESS
GCA/Technology Division
Burlington Road
Bedford, Massachusetts 01730
10..PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-018
11. CONTRACT/GRANT NO.
68-02-1341
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 AND PERIOD COVERED
Final; 7/73-8/75
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES
project officer for this report is William B. Kuykendal,
Mail Drop 62, 919/549-8411 Ext 2557.
is. ABSTRACT
report describes the design, development, and testing of a fine particu-
late source monitoring instrument for real-time measurement of mass concentration
as a function of aerodynamic particle size. It includes a literature review and selection
of the operating principle on which the instrument is based. The described device size
segregates particulates using inertial jet-to-plate impaction on a continuously moving
substrate, and determines collected mass using beta radiation attenuation. The
collection-detection system consists of a 7-impaction-stage cascaded configuration
for direct insertion into a stack, with beta mass sensing at each collection stage.
The program's initial objective — to develop an instrument that could be operated at
up to 260 C (500 F) — cowld not be met: the beta detectors failed to operate satisfac-
torily when exposed to such temperatures. However, the instrument was completed
and tested satisfactorily in a test tunnel at room temperature.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Fines
Measurement
Instruments
Weight (Mass)
Beta Particles
Air Pollution Control
Stationary Sources
Particulate
Mass Concentration
Inertial Impaction
13B
14B
13. DISTRIBUTION STATEMEN1
Unlimited
19. SECURITY CLASS (ThisReport)'
Unclassified
21. NO. OF PAGES
132
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
123
------- |