HEW
United States
Department of Health,
Education, and Welfare
Public Health Service, Center for Disease Control DHEW(IMIOSH)Put).. No. 78-125
National Institute for Occupational Safety and Health
Cincinnati, Ohio 45226
EPA
United States Office of Energy, Minerals, and Industry
Environmental Protection Office of Research and Development
Agency Washington, D.C. 20460
EPA-600/7-77-147
December 1977
DEVELOPMENT AND
FABRICATION OF A
PROTOTYPE FIBROUS
AEROSOL MONITOR (FAM)
Interagency
Energy-Environment
Research and Development
Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5. Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects, assessments of, and development of, control technologies for energy
systems, and integrated assessments of a wide range of energy-related environ-
mental issues
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
DEVELOPMENT AND FABRICATION OF A PROTOTYPE
FIBROUS AEROSOL MONITOR (FAM)
Pedro Lilienfeld
Paul B. Elterman
GCA CORPORATION
GCA/TECHNOLOGY DIVISION
Bedford, Massachusetts
Contract No. 210-76-0110
U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
Public Health Service
National Institute for Occupational Safety and Health
Division of Physical Sciences and Engineering
Cincinnati, Ohio 45226
December 1977
-------
DISCLAIMER
The contents of this report are reproduced herein as received from
the contractor.
The opinions, findings, and conclusions expressed herein are not
necessarily those of the National Institute for Occupational Safety
and Health, nor does mention of company names or products constitute
endorsement by the National Institute for Occupational Safety and
Health.
NIOSH Project Officer: Paul Baron
Principal Investigator: Pedro Lilienfeld
Partial funding of this project was provided
by the Environmental Protection Agency under
the Energy/Environmental R and D Program.
DHEW (NIOSH) Publication No. 78-125
ii
-------
ABSTRACT
This report describes a program whose objective was to develop, design, fab-
ricate and laboratory-test two prototype instruments capable of real-time
selective detection and measurement of airborne fibrous-shaped particles.
The operation of this fibrous aerosol monitor (FAM) is based on the rotation
imparted to the acicular particles by means of a rotating quadrupole electric
field of the order of 3000 V/cm. This field induces a dipole charge separa-
tion on the fibers and aligns them with the applied field. The selective
detection of the fibers is effected by synchronous detection of the resulting
modulation of the light scattered from a continuous-wave helium-neon laser
beam by the rotating particles.
The theory of operation, design of the electric-optical detection configuration
and the electronic signal processing method are discussed. By means of a scat-
tering pulse sharpness discrimination technique, fibers whose length exceed a
selectable value are detected and their number concentration digitally dis-
played. Counting times of 1 to 1000 minutes permit the measurement of fiber
concentrations between 0.001 and about 30 cm~3. A concentration of the order
of one fiber per cubic centimeter can be assessed within a relative standard
deviation of 10 percent with a counting period of 10 minutes.
The FAM is a portable, battery-powered instrument capable of operating con-
tinuously for about 4 hours between battery charges. Preliminary tests with
crocidolite and crysotile asbestos as well as with glass fibers were performed
in the laboratory and the results indicate good correlation with the standard
NIOSH phase contrast microscopy filter counting method.
111
-------
CONTENTS
Abstract iii
Acknowledgments viii
Introduction 1
Theory 3
Principle of Operation 3
Electric Field-Induced Fiber Alignment 4
Light Scattering by Fibrous-Shaped Particles 11
Single Fiber Detection Statistics and Flow Considerations ..... 20
Rectilinear Aerodynamic Behavior of Airborne Fibers 22
Prototype Instrument Development -27
Laboratory Test System- • 27
Prototype Design 33
Prototype Aerosol Testing .... 49
Prototype Instrument Instruction Manual 55
Specifications -55
Operating Procedures ..... 56
Electronic Testing and Adjustment Procedures 59
Optical Alignment Procedure and Cleaning 61
Safety Aspects 63
References 64
Conclusions and Recommendations • 67
Appendix A 69
v
-------
FIGURES
1. Phase lag between the fiber axis and rotating electric field vector
as a function of fiber length-to-diameter ratio and field intensity
for to = 2-rr • 400 sec"1, n = 1.8 x lO"4 poise, and e = 5 8
2. Quadrupole field configuration 10
3. Light scattering-electric field geometry , ... 12
4. Pulse shape as a function of fiber length 14
5. Detection pulse shape for a 4 ym long fiber (X = 0.6328 ym) and
gating intervals for the ratio method of fiber length
determination 17
6. Detection pulse shape for a 6 ym long fiber (A = 0.6328 ym) and
gating intervals for the ratio method of fiber length
determination 18
7. Ratio signal as a function of fiber length 19
8. Sedimentation of ellipsoids, i) axis parallel to direction of
motion; ii) axis normal to direction of motion 25
9. Schematic of initial version of the circuitry used to generate
rotating electric field 28
10. Flow test system used to observe air flow streamlines 30
11. Hydrolized bed aerosol generator 34
12. Schematic of optical and flow systems of FAM 35
13. Optical and flow elements of the FAM 36
14. FAM quadrature field gen and driver board 40
15. FAM overall wiring layout 41
16. Overall block diagram of electronic subsystem of the FAM 43
17. FAM gating and processing board 44
18. FAM counter and display boards 46
19. Battery output voltage as a function of FAM operating time 48
20. View of bottom of FAM enclosure with panel removed 50
21. Top view of FAM with electronic circuitry card cage in the lifted-up
position 51
22. FAM test data 53
23. Top view of FAM . 57
24. Top views of quadrature field generator-driver and of gating-
processing boards 60
25. Top view of counter-display board 62
vi
-------
TABLES
1. Comparison of fiber lengths as a function of nebulizer operation
generation time ....,,...,,..,... 32
2. Comparison of fiber lengths as a function of nebulizer operation
time 32
3. Test data of FAM « . , , 54
vii
-------
ACKNOWLEDGMENTS
We wish to acknowledge the crucial contributions to the
execution of this development program of M. Yagjian, D.
Anderson, R. Stern, and A. Armstrong of the GCA/Technology
Division, as well as the seminal participation of Dr. D.
Cooper of Harvard University.
Vlll
-------
INTRODUCTION
The presence of fibrous-shaped airborne particles, in particular asbestos dust,
has been found to result in several serious human respiratory ailments among
which asbestosis and lung cancer are salient, O»2,3) jn addition, fibers of other
substances such as glass, talc and textiles are suspected of being causatory
agents in similar pulmonary problems. The commonality of the use of asbestos
as a construction and insulation material has resulted in its pervasive pre-
sence and in frequently unsuspected contamination of both enclosed spaces as
well as open air, environments.
The pathological effects of such dusts have been found to be correlated and
directly associated with the retention of fibrous-shaped particles with lengths
typically exceeding about 5 ym.
The high correlation of cause and effect between the exposure of asbestos in-
dustry workers to these fibers and their respiratory pathology has resulted in
stringent regulations4 in terms of 8-hour shift exposure concentra-
tions of asbestos fibers, and a continuous trend in reducing the threshold
limit values for such concentrations has been pursued and will probably continue
in the foreseable future.
The detection and measurement of the concentration of airborne fibers is ob-
viously a crucial requirement in any industrial protection program, in the
assessment of potential dangers, and for the enforcement of regulations.
The only method, presently accepted, for the determination of the concentra-
tion of airborne fibers; e.g. asbestos in its various forms, consists of
manual counting and sizing of such fibers by phase-contrast microscopy of sam-
ples collected by filtration on membrane media.^' This method is relatively
imprecise since it is subjected to human interpretation and concomitant errors.
In addition, the process involves significant and in many cases intolerable
time delays before the information becomes available and can be fed back for
the implementation of corrective and/or protective measures. Although inte-
grated shift exposure will, at least for the time being, rely on the above-
mentioned method of assessment, an urgent requirement exists for an instrument
capable of providing real-time measurements of the concentration of airborne
particles of fibrous (or acicular) shape. No such monitoring device, compat-
ible with field use, portability, and the capability to size-selectively count
fibers has been available to date, although several conceptually feasible
solutions have been and continue to be explored.^9'1°)
-------
Within this background, the GCA/Technology Division endeavored to develop a
practical, field-worthy instrument for the on-the-spot selective detection,
counting and sizing of discrete acicular-shaped particles, even in the pre-
sence of nonfibrous particle concentrations several orders of magnitude higher
than those of the particles of interest. The present report describes the de-
velopment program pursued under a NIOSH contract whose objective was to de-
velop and fabricate two identical prototype instruments to detect exclusively
fibrous aerosols and to measure the number concentrations of fibers in the
range of 0.1 to 20 cm"3, with diameters exceeding 0.1 ym and lengths exceeding
5 ym.
Further requirements and specification objectives were: relative standard de-
viation of a measurement within 10 percent for a constant concentration of
five fibers/cm3 and a measurement time Df 3 minutes or less; battery and/or
line operation with a battery supply compatible with a 4-hour continuous opera-
tion without external power input; maximum weight and overall volume of 13.6 kg
and 0.057 m3, respectively.
The program included a theoretical effort whose objectives were to relate the
physical dimensions of the fibers to the parameters sensed by the instrument,
and to characterize and optimize the operational parameters of the system.
The principle on which the Fibrous Aerosol Monitor (henceforth called FAM) is
based is a result of the combination of electrostatic and light scattering
properties'of fibrous-shaped particles. A continuously flowing sample stream
of air is subjected to a high intensity rotating electric field and concurrent
illumination by a parallel beam of monochromatic light from a continuous wave
gas laser. Fibers present in the sample stream are aligned by dipole charge
separation with the electric field whose lines of force are perpendicular
to the illuminating light beam. The ensuing fiber rotation is synchronous
with the applied field and results in a pulsed modulation of the light scattered
within the plane of fiber rotation. Individual fibers are thus detected in
the form of pulse train bursts whoge frequency and phase (with respect to the
driving field frequency) are known. Phase sensitive electronic amplification
is used in order to improve the signal-to-noise ratio in the presence of noise
associated with nonfibrous-shaped particles and other sources of spurious light
reaching the detector.
The theoretical work was pursued concurrently with the optical, electronic and
mechanical design effort whose objactive was the development and fabrication
of two field-worthy prototype instruments. Laboratory testing of the electro-
optical system played an important role in the experimental facets of this
development. Among the peripheral, although vitally important, efforts of this
program was the design and development of a reliable fiber aerosol generator
without which the success of this intensive instrument development program would
have been rather questionable.
-------
THEORY
PRINCIPLE OF OPERATION
The FAM bases its operation on'a two-step sensing procedure: (a) inducing the
fibers to rotate rapidly by the application of a rotating high-intensity
electric field, and (b) detecting the light scattering signature associated with
these fibers as they are illuminated by a monochromatic light beam.
Because the frequency ofi1 the light scattering pulses thus produced is known,
synchronous or lock-in electronic signal detection can be used in order to
greatly enhance the. signal-to-noise ratio. Each individual fiber detected in
this manner generates a pulse train as it rotates describing a-helical.trajectory
resulting from the combined-effects of the rectilinear air flow and the per-
pendicular field-induced rotation. Fiber length discrimination is achieved by
sensing pulse sharpness, since longer fibers produce narrower; pulses than
shorter ones. The detailed description and the theory underlying this length
discrimination method will be presented further ahead in this report.
The technique of induced fiber rotation"and concurrent detection of the result-
ing light scattering signature results in a powerful discriminatory method for
the selective counting of airborne fibers even in the concomitant presence of
other, nonacicular, particles in concentrations exceeding those of fibers by
factors of up to 106.
The overall technique required the combination and solution, for the first time
within a practical instrument concept, of such phenomena as electrostatic
polarizability, field-induced rotation against Stokes-type air viscosity forces
and randomization by Brownian bombardment, within a laminar fully-developed
parabolic flow profile regime, detection of 90° light scattering pulse trains
and the characterization and discrimination of the pulse shape of these fiber-
associated trains.
I , .- : ' . ...
The general idea of electric field-induced alignment of elongated particles such
as long molecular chains, rod-like crystals, filamentous viruses, etc., in
liquid'suspensions, and characterizing their properties-by observing the bulk
lightr-scattering behavior of the suspension has been applied and studied during
the last 40 years.(11-17) Particle polarizability, light scattering properties,
rotational diffusion constants, determination of the diameter distributions of rod-
like particles, and in general characterization of such particles in liquid
suspensions,have been objectives of such investigations.- The experimental ap-
proach used in the majority of these studies'consisted of exposing a sample
of the colloidal'suspension under scrutiny to a bipolar electric field; i.e.,
two parallel plates with either a d.c. or an a.c. applied potential. Sample
illumination by an incandescent source and bulk light scattering detection by
means of photomultipliers were typical. Polarization and relaxation properties
3
-------
have thus been investigated using the sudden application and interruption of
d.c. fields, whereas the use of a.c. fields has been applied to similar ob-
jectives as well as to more detailed characterizations made possible by the
use of sophisticated techniques such as synchronous detection of the light
scattering signal.
Within this background, the idea of applying the fiber alignment-light scatter-
ing sensing approach to the selective detection of airborne fibers was developed
at the GCA/Technology Division. Initially, as stimulated by the above cited
liquid suspension studies, the objective of detecting bulk light scattering
oscillations produced by the application to a volume of air of an alternating
field between two electrodes was considered. This approach was used, indeed,
to initially (i.e., prior to this instrument development program) establish
the feasibility of the overall method, but questions arose as to the ability
of this approach to provide useful quantitative information on fiber number con-
centration and size. The two-electrodes a.c. approach produces fiber alignment
when the field maxima occur, and relies on randomization of fiber orientation
as the field intensity passes through zero; this results in a fiber motion
whose amplitude is an inverse function of the field frequency; i.e., shorter
time between field maxima reduces the deviation associated with fiber randomi-
zation due to Brownian bombardment. It was decided that a more controlled
fiber rotation, that could be extended to higher frequencies, could be achieved
by the use of a rotating electric field within which the fibers would be under
the continuous influence of a field of constant intensity.
The requirement for discrete fiber detection, as opposed to the mere detection
of the presence of fibers, imposed far more stringent demands on the method of
optical sensing and electronic signal processing, as well as on the flow regime
and dimensional constraints of the system.
The subsequent sections will present a somewhat more detailed treatment of the
various central aspects of this technique. The scope of this presentation,
however, will be limited to the essentials involved with specific emphasis on
the aspects most specific to this instrument design.
ELECTRIC FIELD-INDUCED FIBER ALIGNMENT
Fiber Orientation and Rotational Dynamics
The orientation or alignment of fibers; i.e., elongated or acicular particles,
suspended in a gaseous or liquid matrix can be accomplished by several means;
e.g., by the application of electric fields, magnetic fields or by flow-shear
forces. Although magnetic fields have been used for the orientation of certain
types of asbestos fibers,(-18'19) the use of an electric field presented a num-
ber of advantages within the present development, such as: more universal ap-
plicability to all types of fibers, without restrictions imposed by their mag-
netic properties, and greater simplicity and reduced power and weight require-
ments for the generation of relatively intense fields. Flow-shear alignment
is limited to fiber axis orientations parallel to the direction of flow under
laminar conditions, and is thus not applicable to the generation of more com-
plex motions such as fiber rotation.
-------
Fiber alignment with an applied electric field implies the generation of a
torque resulting from the interaction of the external field and the field-
induced dipole moment. This moment is associated with the electric polariza-
tion of the elongated particle. In dielectric (i.e., nonconducting) materials,
polarization is associated principally with electronic and atomic charge dis-
tortion and orientation. Conductive materials or semiconductors exhibit polar-
ization mainly due to field-induced charge migration also called space charge
polarization. This latter mechanism probably plays a predominant role in the
creation of the dipole moment and the resulting torque in an applied electric
field for fibrous-shaped aerosol particles. The observed behavior of fibers
subjected to a rotational electric field, during the experimental phase of this
program was consistent with the assumption that these particles are electric-
ally conductive. This observation confirms Fuchs'^20^ contention that ..." in
practice, aerosol particles can be considered conducting...," although prelim-
inary empirical evidence gathered during the closing phases of this program
indicates that this assumption may not apply to all types of fibrous particles
under all environmental conditions (see section on Prototype Testing).
The degree to which an acicular particle can be aligned with an applied elec-
tric field, in the absence of turbulence, depends on the opposing effects of
the orienting force; i.e., the interaction of the field and the dipole moment,
and the randomizing forces resulting from Brownian molecular bombardment. The
torque resulting from shear flow associated with laminar regimes will be
neglected in this treatment mainly because within the configuration under
scrutiny the fibers are sensed very near to the flow axis where the radial
flow velocity gradient is negligible.
The basic equation of motion governing the rotation of elongated particles is
given by:(2o)
de/dt = B T = en (1)
U)
where 6 is angular measure of orientation
B is angular or rotation mobility
03
T is the torque = d^/d9, and
6
tt is the change in field energy caused by particle polarization.
The general equation of the angular mobility for a prolate ellipsoid is:
3
~232 - 1
/P2"^"!
—
u i.
Jin (3 + A
\i -\\ Q
•> -1-/ P
i ^
(2)
*For reasons of simplicity and consistency with Fuchs' treatment(2°'these equa-
tions are based on the c.g.s. and electrostatic system of units.
-------
where n is the coefficient of viscosity of the gas
a is fiber radius (d/2), and
3 is length to diameter ratio of the particle (L/d).
For large values of 3 (i.e., 3 > 10) equation (2) can be replaced by:
B = 3(2Jln(23) -
(3)
For conducting fibers the mean torque on the elongated particle is obtained by
differentiation of the expression:
n -
\l — —
cos2e sinV
(4)
where V is particle volume,
E is the electric field intensity, and
X, and XT are defined as:
Xl
- 1
= Jin (3 + /32 - 1 j -
- 2(32 -
Thus the torque is equal to:
Jin (3 + /B2 - 1)
- 1
(5)
(6)
T = VE2 cos 6 sin 9 [ — - —
X
From equations (1), (3) and (7) we obtain:
(7)
3 (2 £n(23) - 1) VE2 . / 1 1\
q ~2~ S111 2Q 7" ~ 7~
2irr|L3 z \xl X2 /
\ '
Trd2
which for large values of 3, and since V = —— L, can be reduced to:
(2 Jin (23) -1)
(in (23) - 1)
6
(8)
(9)
-------
From equation (9) we can write the expression for the phase angle 6 between the
electric field vector and the axis of the prolate ellipsoid (i.e., the fiber
axis) :
sin 29 = 161CQ (*° 23 - 1) (1Q)
3 E2(2 £n 2g - 1)
Since the maximum torque occurs at 8 = 45 , if the phase lag between the fiber
axis and the field vector exceeds that angle the particle can no longer follow
the field rotation and its rotation falls out of synchronism with the applied
field frequency.
If the same assumption is made as for the above case; i.e., 3 > 10, the cor-
responding equation for dielectric fibers, whose dielectric constant is e is:
sin 26 =
3E2 (2 £n 2B - 1) (e - I)
Equations (10) and (11) are plotted on Figure 1 for to = 2-rr x 400 sec"1 which
was the frequency selected for the final prototype version. A discussion of
the rationale underlying the selection of that particular field frequency will
be presented in a subsequent section of this report. Equation (11) was plotted
using e = 5 (typical for asbestos) . Upon inspection of the curves depicted on
Figure 1 it becomes apparent that the behavior of conductive and that of non-
conductive fibers is very different with respect to the dependence of phase an-
gle on the length-to-diameter ratio of the fibers. Conductive acicular par-
ticles rotate with a phase angle essentially independent of the value of L/d,
whereas this angle is strongly dependent on that ratio for nonconductive
fibers. In addition, dielectric fibers are far more difficult to align with
the field than conductive ones; e.g., for a field intensity of 5000 V/cm
(16.67 stat volts/cm for equations (10) and (11)) and for an L/d of 20 the
phase angle is more than 100 times larger for the nonconductive particles than
for the conductive case. Similarly, for a field of about 3000 V/cm (10 stat
volts/cm) fibers with an L/d exceeding about 19 cannot be aligned unless they
are conductive. The relevant experimental observations that indicate that the
behavior of the fibers appears to be more closely approximated by the conductive
model will be presented further ahead. In this context it should be considered
that small fiber resistivities many decades higher than those commonly accepted
as typical of conductors still result in predominantly conductive behavior at
frequencies of the order of 400 hertz. In fact, because of the exceedingly
small capacitance of a fiber having the typical dimensions of the order of
1 ym diameter and 10 ym length (~10~19 farads), resistivities as high as of
the order of 1013 ohms/square result in charge transport time constants of the
order of 10~5 sec; i.e., considerably shorter than the period of the 400 Hz
field frequency. Some questions remain, however, whether a clearly defined
classification of many fibers into either conductive or dielectric is possible,
and the existence of transitional and nonohmic (nonlinear resistivity) condi-
tions may have to be considered for a more rigorous characterization of the
behavior of asbestos, glass and other types of fibers.
-------
100
50
20
10
O c
u.o
0.2
O.I
0.05
0.02
0.01
45° LIMIT
3000
4000
5000
10
Figure 1.
20
30
40
50
60
70
60
90
100
Phase lag between the fiber axis and rotating electric field
vector as a function of fiber length-to-diameter ratio and
field intensity for o> = ZTT • 400 sec"1, n = 1.8 x 10~4 poise,
and e = 5 (equations 10 and 11). Nearly horizontal curves
are for conductive fibers, curves in upper left are for
dielectric fibers.
8
-------
Quadrupole Rotational Electric Field
In order to impart a continuously rotating component to the fiber motion as it
passes through the detection region of the FAM a quadrupole field configura-
tion was selected. The electrodes consist of four equally-spaced sectors with
cylindrical symmetry. For flow-dynamic reasons (i.e., in order to prevent
localized turbulence and eddy formation) these electrodes were designed in the
form of conductive foils attached to the inner surface of a glass cylinder.
In theory these electrodes could be exterior to the insulating support cylinder,
however, this latter solution was found to be electrically unstable because of
its sensitivity to any contamination of the inner dielectric surface. The
rigorous mathematical characterization of the electric field intensity dis-
tribution throughout the volume enclosed by the quadrupole both as a function
of radial distance and time, requires a rather complicated treatment which is
beyond the scope of this work.
An approximate solution to the field intensity at the axis of quadrupole;
i.e. , the region of interest in this case, can be obtained by taking only
the first term of the series expansion:
V(r,6) = I. — i-rr (C cos n6 + D sin n6)rn (12)
i TTria11 n n
n=l
where Cn = 2(V± + V2) sin 4r and Dn = (V± - V2) (1 - cos nir)
and V-, and V7 are the peak amplitudes with respect to ground of the potentials
between the two pairs of opposing electrodes; i.e., one-half of the peak am-
plitude of the applied potentials, a is the radius of the electrodes and r is
the radial distance. Figure 2 illustrates these parameters and the quadrupole
configuration. Of the series expressed by equation (12) only the first term
contributes significantly to the field near the axis of the system, the second
term becomes zero and the third is already of negligible magnitude compared with
the first one and since in this case V, = V2 = V, the approximate expression for
the axial field becomes :
'i-H <»>
The coefficient 4/ir represents the electric field enhancement at the axis of
the quadrupole due to the focusing effect of the concave shape of the field
electrodes; i.e., the field at the axis is A/TT times higher than that result-
ing from a uniform field (parallel electrode configuration) . This focusing
field may tend to contribute to the stability of the axial motion of fibers
passing through the sensing volume.
Equation (13) is a reasonably good approximation of the field magnitude at the
system sensing volume both for the axis of the quadrupole as well as a function
of time (or 6) assuming sinusoidally varying applied potentials.
-------
-V,
Figure 2. Quadrupole field configuration
LIGHT SCATTERING BY FIBROUS-SHAPED PARTICLES
General Discussion
Light scattering equations in the Mie region; i.e., when the dimensions of the
scatterer are of the same order as those of the illuminating wavelength, are
notoriously complex even for spheres. The only other particle shape that has
been treated extensively in the literature from the point of view of scatter-
ing properties are cylinders,(22~27) whose scattering functions are even more com-
plicated than those of spheres as two additional angular dependences must be in-
cluded; i.e., the angle between the direction of illumination and'the cylinder
axis and the 'angle between this axis and the direction of observation. This
problem becomes more .tractable, however, for specific limiting cases. For
the present configuration, the axis of the scattering cylinder rotates in a
plane normal to the direction of the illuminating beam; i.e., the case of per-
pendicular incidence of light on the cylinder axis which is equivalent to a
constant, maximum illumination oh the scatterer.(28) A further simplification
results-from the choice of a scattering angle of observation of 90° from the
forward direction; i.e., the detector axis lies in the plane of cylinder rotation.
The primary assumption being made with respect to fiber shape is that it can
be approximated by a cylinder. The scattering analysis of irregularly-shaped
fibers is totally intractable and their detection can only be approached in a
totally empirical manner as will be discussed further on within this document.
On the other hand it must be considered that several types of fibrous'aerosol
particles, such as amosite and crocidolite asbestos, and glass fibers are in
most cases straight cylinders whose light scattering behavior'is well modeled
by the cylindrical assumption.
10
-------
90C
Scattering of Cylinders With Perpendicular Illumination
The choice of the 90 scattering angle was based on simplicity of instrument
design and signal processing. For this angle the modulation of the scattered
light is nearly 100 percent because the projected fiber area viewed by the de-
tector undergoes the largest excursion during each semirotation. This can best
be understood when considering the other two limiting scattering angles of 0°
and 180 where the modulation due to the fiber rotation goes to zero; i.e., the
fiber axis would always be viewed perpendicularly and no fiber-selective signal
would be detected.
Figure 3 depicts the essential geometrical configuration of the relevant scat-
tering and illuminating angles, and the relative position of the orienting
quadrupole field electrodes.
ELECTRIC FIEL
PLATES
INCIDENT LIGHT
Figure 3. Light scattering-electric field geometry
11
-------
The fibers scatter a maximum in the direction perpendicular to their axis, thus
for every complete fiber rotation two light scattering maxima or peaks are
detected within the configuration of Figure 3. This angular dependence of the
light scattered by a cylinder can be compared with a rotating search-light beam,
or with the emission of a pulsar, in that the 'detected signal is very sharply
peaked at the maxima and nearly zero at all other angles (or times, for a uni-
formly rotating fiber). In fact, it is the degree of this signal pulse sharp-
ness that serves as the basis for one of the methods of fiber length discrimina-
tion used on the FAM. The general scattering function to be applied in this
case is: (22)
sin kLZ/2
T
where Z is the angle between the scattering detection axis and the plane per-
pendicular to the fiber gxi&. For the geometry illustrated in Figure 3, Z can
be expressed in terms of .0 and $ by the relationship:
sin Z = sin 6 sin (15)
Assuming <(> to be nearly 90 , to a close approximation equatidn (14) becomes
The scattering intensity or detected intensity is defined by:
I = IQ S2/k2r2 (17)
Replacing equation (16) in equation (17), and since k = 2TT/X we obtain:
sin irL9/X\2 m2
where T(<|>) is a complex function of , m, A and af.
The parameters of the above function are:
af = fiber radius
L = fiber length
m = fiber index of ' refraction at X
X = wavelength of illumination
0 = fiber axis angle in the plane of rotation
(6 = 0 is defined as broadside viewing of fiber axis)23
<|> = scattering angle (90° for the FAM)
r = distance between the fiber and the detector.
Note: Van de Hulst's notationC22) was reversed for consistency with the nota-
tion previously used in this report.
12
-------
For fibers whose diameter is larger than about 1/3 A (approximately 0.2 ym for
the He-Ne laser the scattering intensity increases as decreases, but for
smaller diameter fibers T2() becomes nearly independent of c|>.(23>210 Thus
the use of forward scattering optics would not substantially improve the ability
of the FAM to detect thin fibers but would result in added complexity of signal
detection and interpretation. The function T() will not be further analyzed
because within the method selected for fiber length discrimination, the con-
tribution from that function will be cancelled out as will be shown forthwith.
Assuming all parameters of equation (18) to be constant with the exceptions of
fiber length L, and fiber angle 0, the intensity of the detected scattering
signal in the vicinity of the maximum amplitude point of 0 = 0 is shown in
Figure 4 for four different fiber lengths. It becomes apparent from this graph
that as fiber length increases (all other parameters remaining invariant) pulse
sharpness and amplitude increase concurrently. It further becomes obvious that
for fibers longer than about 5 ym most if not all of the scattering signal is
contained within a very narrow angle of rotation of less than about ± 6° around
the perpendicular or face-on position of observation, thus producing the above-
mentioned search-light-like pulses.
Pulse Width Detection Method
One of the primary objectives of this development was to detect individual
fibers and size-discriminate them in terms of length. During the early phases
of this program several alternative methods for fiber geometry discrimination
were investigated such as the detection of the phase angle between the applied
field and the fiber axis, the ratio of the detected signal between side view-
ing and end-on viewing, sensing of the two polarization components, etc. It
was concluded that all of the methods had serious drawbacks in that they pro-
vided mainly information on the length-to-diameter ratio of the fibers and/or
the diameter, and in the case of the phase angle detection method this approach
was only applicable to nonconductive fibers (see preceding discussion in this re-
port) , the method of end view to side view intensity ratio which is approxim-
ately proportional to the ratio L2A2 requires a very precise determination of
the signal baseline (to within better than 1 percent), which is very difficult
to achieve in the presence of nonfibrous particles in the sensing volume. Fur-
thermore this method would be extremely sensitive to deviations of fiber shape
from a perfect cylinder. The primary length discrimination method selected for
incorporation in the FAM was based on the analysis of the angular scattering
intensity distribution of each fiber as it undergoes its synchronous rotation.
As mentioned before, two pulses of scattered light are observed for each com-
plete fiber rotation, and a number of these pulses is detected as the fiber
rotates through the sensing volume. The measurement of the widths or the sharp-
ness of these pulses was considered to uniquely characterize their lengths as
discussed within the preceding section on the light scattering behavior of
acicular particles. The method of pulse width sensing has the advantages of:
(1) independence from the electrical properties, the index of refraction (and
the diameter of the detected fibers), (2) independence from the absolute in-
tensity values both of the illuminating laser beam as well as the detector
sensitivity, (3) relative simplicity of signal processing.
13
-------
120
Figure 4. Pulse shape as a function of fiber length
14
-------
Basically this method consists of sensing the detected pulse width at some
specific fraction of its peak amplitude, thus normalizing each measurement
against that peak value. Two main approaches were considered in order to im-
plement this technique: sensing the ratio between the peak amplitude and the
signal average, and sensing the ratio of the averages of the signals contained
between two different rotation angle ranges. The first approach can be ex-
pressed mathematically as (from equation (18)):
Vak (9 = 0) _ __^ (19)
average (6 < 9 < 2ir)
(20)
which for L > 1 ym and X = 0.628 ym is essentially equal to:
average
peak 4irL
I
Equation (21) indicates that this ratio is proportional to fiber length, X
being constant. The disadvantage of implementing this method, however, is that
true pulse peak detection suffers from poor noise immunity. Thus the alternate
method of pulse sharpness detection was pursued and incorporated in the final
version of the prototype FAM. This technique consists of integrating each pulse
train from a fiber using two different gate widths corresponding to two different
intervals of 9. The first interval extends from the pulse center (9 = 0) to a
time T-i corresponding to an angle 0]_. The second gating interval extends from
T-L to 2Ti, or 26-, . In practice, this integration is performed symmetrically on
both sides of the pulse center as illustrated in Figures 5 and 6. If the ampli-
tudes of the pulses are proportional to the scattering intensities observed by
the detector, then, from equation (18) the ratio, R, of the two pulse integra-
tions is:
r
Jo
/sin
"
d9
R = — (22)
where a = TrLO/X
61 = WT1'
to = dG/dt
15
-------
9, d»gr««»
Figure 5. Detection pulse shape for a 4 pm long fiber (X = 0.6328 ym)
and gating intervals for the ratio method of fiber length
determination
16
-------
40
36
32
w 28
z
ui
I 24
£
ui
t-
w 20
16
12
AREA 8
AREA A + AREA C
CHANNEL 2
CHANNEL 2
L-
AREA C
-8 -2R.-6 -4 -^. -2
0, 4
«, 25,
Figure 6. Detection pulse shape for a 6 ym long fiber (X = 0.6328 ym)
and gating intervals for the ratio method of fiber length
determination.
17
-------
From equation (22) it is evident that R is only a function of L for a constant
value of X, the wavelength of illumination. In practice R is obtained by
averaging electronically over each complete pulse train associated with a dis-
crete fiber passage through the sensing volume. By the appropriate choice of
9i and the discrimination level of R the desired fiber length discrimination
is achieved.
From equation (18), for a fiber 5 ym in length, illuminated by X = 0.632 ym
radiation, the scattering intensity drops to one half of the peak value when
6 = ± 3.2°. Figure 7 is a plot of equation (22), obtained by numerical inte-
gration for 9-^ = 3.2°. As can be seen from that plot R varies quite rapidly
as a function of fiber length and is thus a sensitive discrimination parameter.
For example, as illustrated on Figures 5 and 6 and from Figure 7, with L = 4 ym,
R = 2.4 and for L = 6 ym, R = 7.4. Thus, for the discrimination fiber length
of 5 ym, the selection of 9^ = + 3.2° appears as an optimal'solution. Although
for L > 8 ym some degree of ambiguity exists in the relationship between R and L
(see Figure 8) for that value of 9^, the discrimination against L < 5 ym re-
mains obviously unaffected.
Light Source Considerations
The light source required for the illumination of the fibers must fulfill
several requirements which will now be discussed.
From the geometric point of view the beam must be nondivergent and well collim-
ated in order to provide a well-defined sensing volume of cylindrical shape.
Any significant departure from parallelism over-the detection length would re-
sulr in occasionally incomplete pulse trains when a fiber is moving near the
edge of the beam. The second requirement is that the beam be polarized and that
the radiation electric vector be positioned perpendicularly to the scattering
plane defined by the incident beam and the observed scattering direction. The
use of polarized light source further permits the use of Brewster-angle windows
on the sensing-flow duct of the FAM,- thus minimizing spurious reflections at
such windows.
The third requirement imposed on,the source is monochromaticity. This require-
ment arises from the method used for length discrimination; i.e., pulse sharp-
ness detection. If a polychromatic source is used (e.g., an incandescent lamp)
a continuum of pulse widths would be produced and the sharply defined pulses
associated with a single-valued illuminating wavelength could not be observed.
*"
The source wavelength is not critical, although the,lower the value of X, the
higher the intensity of the light scattered by a Mie-type scatterer. If the
main source of background noise is associated with Mie-scattering from non-
fibrous shaped dust particles (i.e., a reasonable approximation to this case)
then, as a first order approximation, shorter wavelengths result in higher
signal-to-noise ratios.
The light intensity within the detection volume determines, among other param-
eters, the signal-to-noise ratio because the signal amplitude is directly.
proportional to the intensity of illumination, whereas the noise (being mainly
18
-------
100
90
80
70
60
30
40
30
20
o
10
9
7
6
S
SATIN* WIDTHS'
CHANWtL I' -3.2* to 3.2*
CHANNCL 2- -6.4* to -3.2* M4 3,2* to C4» .
J L
_L
6 7 • 9 10
FIBER LENGTH , L
20
4O SO
too
Figure 7. Ratio signal as a function of fiber length
19
-------
of the shot noise type) is proportional to the square root of the light intens-
ity. Thus, beam expansion would result not only in an increase in fiber
detection coincidence errors but also in diminished signal-to-noise ratios.
Last but not least, cost, availability, size, weight and power consumption
were important selection criteria imposing stringent limitations on candidate
light sources.
Based on all the above considerations and, in addition, because continuous wave
illumination (as opposed to pulsed operation) was required to prevent inter-
ference with the scattering modulation associated with fiber rotation, a 2 mW
helium-neon polarized laser was selected for this application.
SINGLE FIBER DETECTION STATISTICS AND FLOW CONSIDERATIONS
Counting Statistics
Each fiber as it passes through the detection volume produces a pulse train
consisting of a specific number of pulses whose number (per individual fiber)
is given by:
n = 2ft = 2fxv (23)
where f is the applied field frequency,
T is the particle residence time in the sensing volume,
x is the length of sensing volume, and
v is the flow velocity within the sensing volume.
L
Each of these pulse trains must be detected separately in order to sense dis-
crete fibers, and thus the detection volume must be sufficiently small such
that for the highest number concentration to be measured the probability of
more than a single fiber being present in this volume remains negligible.
The probability P that N events (or fibers) are present within the sensing
volume at the same time can be expressed by:
(c VN -c v
P(N) = — e U s (24)
N:
where C is the number concentration of fibers, and
V is the sensing volume.
N=°°
The total probability of particle coincidence is the sum Z P(N). As a use-
N=2
ful practical rule, however, if the sensing volume is smaller than the one-half
of the inverse of the maximum particle concentration the total coincidence error
is less than about 10 percent.
20
-------
The statistical accuracy of the counting process follows reasonably well a
Poison-type distribution and thus the relative standard deviation e of the
fiber count can be calculated from:
= ±— x 100 (25)
where c is the total fiber count over a sampling period T and thus
TCV
c = -7- (26)
The program objectives were stipulated in terms of e , T and C and thus V and T
became the variable design parameters of the FAM. r S
Flow Considerations
One of the primary requirements for the selective detection of fibrous-shaped
particles is that the flow within the sensing region be totally devoid of tur-
bulence, and in addition the flow stream must be parallel to the illuminating
light beam; i.e., the axis of the sensing volume. Furthermore it is preferable
that the velocity of all streamlines within the sensing volume be the same and
that no acceleration or deceleration of the flow velocity occur within this
volume. These conditions can be met with a fully developed parabolic laminar
flow profile if the sensing volume is constrained to a small radius (as com-
pared to the duct radius) centered on the flow axis. The requirement for
laminar flow conditions within the sensing volume results from two detection
aspects: (a) fiber motion must be rectilinear in the direction of flow and
purely rotational in the perpendicular plane in order to preserve the signal
pulse shape, and (b) fibers cannot be allowed to drift in and out of the sensing
volume thus producing incomplete pulse trains. From the experimental evidence
found during the initial phases of the FAM laboratory investigation, it was
found that a low Reynolds number (< 500) based on the duct diameter, is by it-
self an insufficient criterion for flow stability. Flow obstructions, discon-
tinuities and short inlet lengths can produce unacceptable instabilities that
are reflected in signal breakup and distortions.
RECTILINEAR AERODYNAMIC BEHAVIOR OF AIRBORNE FIBERS
General Theory for Ellipsoids of Revolution
The theoretical aspects of the rectilinear aerodynamic behavior of fibrous-
shaped particles were explored within this program in order to provide guide-
lines for the use and selection of respirable versus nonrespirable size separa-
tion methods. The rotational aerodynamic behavior of fibers was treated in a
preceding section of this report as related to the motion imparted to the fibers
by the applied rotational electric field.
In general, the dynamic behavior of all aerosol particles is governed by their
aerodynamic properties; i.e., their motion is the resultant of opposing forces:
a driving force such as gravitational acceleration, and the viscous resistance
21
-------
of the medium within which the particle moves. In this context, highly useful
characterization is that of aerodynamic equivalent diameter or Stokes diameter,
defined as that diameter of a sphere of unit density whose settling velocity
equals that of the particle under consideration. In the case of acicular par-
ticles this settling velocity depends on particle orientation with respect to
the direction of motion.
The treatment to be followed on aerodynamic characterization and that of the
preceding section on the electrodynamic behavior are based exclusively on the
assumption that the fiber motion is in the Stokes regime and no attempts are
made to incorporate slip corrections (Cunningham) in these calculations since
no adequate treatment of such regime has been developed for particles other
than spherical. It is, however, to be expected that for fiber diameters of
less than about 0.5 ym, especially when the fiber motion is parallel to its
axis, the aerodynamic properties will no longer be correctly characterized by
a purely Stokes treatment. In this context it is worth mentioning that the
FAM may be developed into a tool for the experimental investigation of slip
correction factors for acicular particles.
The Stokes regime can be characterized by the condition wherein the resistance
to motion within the fluid is purely viscous; i.e., the Reynolds number is
typically less than 0.5, resulting in a symmetrical distortion of the fluid
upstream and downstream of the particle, without wake asymmetry. The drag
equation for spherical particles is well known under these conditions. Even
cylindrically shaped fibers require a somewhat more elaborate treatment. A
mathematical approximation to an ellipsoid of revolution appears as a reasonable
approximation, " which has also been applied to model the deposition of fibers
in the human respiratory system.29 Two extreme cases can be recognized:
(1) motion of the ellipsoid (or fiber) along its axis of revolution, and (2)
motion perpendicular to the axis of revolution. It should be "considered that
the gravitational settling motion of any particle with three mutually perpen-
dicular planes of symmetry (such as an ellipsoid of revolution) will be invariant
during its descent, maintaining its initial orientation through its fall tra-
jectory. In practice, however, asbestos fibers for example may not always be
perfectly straight and the above-mentioned rule can be violated.
As noted above, acicular particles can be approximated by ellipsoids of revolu-
tion falling under the action of gravity in either of two attitudes described
under (1) and (2), or any intermediate angle with respect to the direction of
motion. In general, an acicular particle falling with its axis vertical will
have a higher terminal velocity than the case when its axis is normal to the
direction of motion. Intermediate angles will exhibit intermediate velocities.
The difference between cases (1) and (2) are, however, not as significant as
intuition would indicate. Two equations represent the two extreme axis-to-
motion angles defined above.
For case (1): D = - _ - (2?)
9 fi2_i Q
arc cosh (3) -- ^~
22
-------
and for case (2) : D = ^-^ (28)
— arc cosh (B) H
(l-B-2)3/2 1-fT2
where D is the diameter of the sphere having the same settling velocity.
Equations (28) and (29) are plotted on Figure 8,(2°) curve i) refers to the ellip-
soid of revolution moving in the direction of its axis of rotation, and curve
ii) refers to the case when the motion is at right angles to the axis.
Once the value of D has been determined for a given particle, and provided that
the particles fall vertically (no lateral glide) and do not change their orien-
tation during their vertical motion, D can be replaced in the classical Stokes
equation for the diameter of a spherical particle:
(M - M )g
V = —2 -S— (29)
s 3ir nD
where V is the settling velocity
S
M and M are the masses of the particle and the displaced gas,
IT O
respectively (M is usually negligible)
O
g is the acceleration of gravity
D is the diameter of the spherical equivalent particle
T] is the gas viscosity
An an example, the fiber dimensions will be assumed to be;
diameter = 0.1 ym
O r~ f\
length =5 ym i
Based on a density of asbestos fiber of 2.6 g/cm3, the viscosity for air at
standard conditions of 1.8 x 10~4 poise, the determination of settling velocity
is performed as follows:
Case i):
Referring to Figure 8, curve i), we find a value of D/d = 8 for B = 50. There-
fore D = 0.8 ym.
The particle mass is:
M = T O.I2 x 10~8 x 5 x i(T4 x 2.6 = 1.02 x KT13g
p 4
Replacing M in equation (29) we obtain V,^ = 7.4 x IQ"4 cm/sec.
23
-------
100 r
500 200 100 50 20 10 5
/3 (CASE i i )
2 I 0.5 0.2 O.I 0.05 0.020.01 0.005 0.003 0.001
D/L 5 -
0.2 -
O.I
0.001 0.002 O.OO5 0.0« 0.02 0.05 O.I 0.2 0.5
I 2 5 10 20
f (CASE i )
50 100 200 500 1000
Figure 8. Sedimentation of ellipsoids, i) axis parallel to direction of
motion; ii) axis normal to direction of motion.20
-------
Case ii): in this case g. = --
ii B1
Referring to Figure 8, curve ii), the value of D/L = 0.27 is found for &±. =
0.02, thus D = 1.35 ym, thus Vs(±±) = 4.4 x 10"4 cm/sec.
One of the important conclusions to be reached from the preceding calculations
is that the settling velocity of fibers does not vary more than about a factor
of 2 whether the fiber falls with its axis aligned vertically or horizontally
(this also applies to larger particles than the above example). Since in most
real cases deviation from true ellipsoidal shape will preclude perfect vertical
or horizontal trajectories, a good practical rule for the calculation of set-
tling velocities of fibers is to average the results obtained for the two ex-
treme particle orientations. For the above example, the average would be
5.9 x 10-tt cm/sec.
Another aspect to be considered in the context of the aerodynamic behavior of
acicular particles is their orientation within a flow field as a result of
shear forces exerted by the moving fluid. Although, as noted previously, an
acicular particle (rigorously an ellipsoid) tends to fall without rotation in
a gravitational field, in the case of a laminar flow through a duct, these par-
ticles tend to align with their long axis with the direction of flow. This is
a result of the transverse velocity gradient in a fully developed laminar flow
(i.e., with a parabolic velocity profile) which causes a torque on the elongated
particle until alignment with the direction of flow is attained. As applied to
the FAM, this means that fibers arriving at the sensing volume will, most
probably, be aligned with the flow direction before being oriented within the
quadrupole field.
Respirable Fiber Segregation
The penetration of fibrous-shaped particles into the lower respiratory passages
is determined by their aerodynamic properties.^=U Therefore methods for fiber
size selection or precollection are preferably based on these properties. Al-
though, as demonstrated within the preceding analysis, the effect of fiber
orientation on their aerodynamic behavior is limited to factors of the order
of 2, it can be concluded that such a size selection should be performed within
the laminar flow regime in order to fully define the orientation of the fiber
prior to their aerodynamic segregation. Thus, centrifugal flow fields, espe-
cially those associated with cyclone collectors do not fulfill this condition.
Furthermore these devices, due to their high velocity gradients may lead to
fiber break-up and reentrainment. The lack of fiber size segregation defini-
tion of cyclones has been reported(30) indicating that a sampling flow range
as large as 1.1 to 1.9 liters/min resulted in approximations to the ACGIH
Respirable Dust Curve. Horizontal or vertical elutriation can, however, be
effected within the laminar flow regime. Horizontal elutriation, although
used as a standard method of precollection for nonfibrous dusts such as coal
mine dust, has the potential disadvantage of particle reentrainment from the.
collection surfaces. Vertical elutriation appears as the most compatible
solution for the retention of nonrespirable fibers whose aerodynamic size ex-
ceed a minimum particle cut-off. Laminar flow conditions are usually encoun-
tered in such devices; e.g., the 7.4 liter/min elutriator proposed as a stan-
dard method for cotton dust sampling operates at a Reynolds number of approx-
imately 100 with a particle cut-off Stokes diameter of 15 ym.(31>32) The
25
-------
particular cut-off size for fibrous aerosols such as asbestos remains to be de-
termined. The flow laminarity within a vertical elutriator results in a para-
bolic velocity profile with a continuum of velocities ranging from zero at the
wall of the duct to twice the average velocity at the axis. It is thus inevi-
table that the particle cut-off characteristic of such a device will be rather
broad, a fact which is probably compatible with the retention characteristics
of the human respiratory system.
26
-------
PROTOTYPE INSTRUMENT DEVELOPMENT
LABORATORY TEST SYSTEM
Experimental "Breadboard"
One of the earliest tasks within this instrument development program was to
set up an experimental system which could be used as a test "breadboard". This
set-up was then to be modified and optimized in order to determine empirically
the design parameters to be implemented in the final prototype instrument design
and fabrication. Several of these "breadboards" were assembled and tested dur-
ing the initial phases of the program and as a result of the difficulties en-
countered and the experience gained from the operation of these test devices a
preferred design emerged.
The experimental breadboard consisted of the optical illumination train, the
flow tube-quadrupole assembly, the light scattering detector, and the auxilliary
circuitry for the generation of the rotational field and for signal detection.
The final optical and flow-detection breadboard configuration was essentially
identical to that of the two identical prototype instruments delivered at the
completion of the program and will thus be discussed in detail within the sec-
tion of this report describing the final instrument design. The intermediate
developmental steps are obviously only of anecdotal interest and will not be
detailed within this report. The only basic difference between the final bread-
board version and the completed instruments was that the former was supported
by a large bench-top optical table with magnetic latches whereas the latter
used a special honeycomb-type structural panel for the support of the optical
components.
Among the developmental steps of more than passing interest, however, it is
worth mentioning the following ones. The quadrupole electrode configuration
underwent a series of design modifications dictated by flow stability considera-
tions (see preceding section of this report) and by surface contamination
problems when these electrodes were placed on the exterior surface of the
transparent flow duct. The optical system underwent a number of evolutionary
modifications: midway through the program it was considered that a laser beam
expander followed by a stop and re-imaging optics would be required to provide
a squared-off beam at the sensing volume. The need for these elements was
later obviated by the method of pulse width detection, and in addition, as it
was established that maximum beam intensity resulted in optimum signal-to-noise
ratios.
The quadrupole field generator circuit initially employed for the early testing
efforts consisted of the elements depicted in Figure 9. The oscillator was a
Hewlett Packard Model 200 CD and the two amplifiers were part of a Realistic
27
-------
DUAL AMPLIFIERS
VARIABLE
FREQUENCY
OSCILLATOR
I J
PHASE SHIFTING
NETWORKS
STEP UP
TRANSFORMERS
Figure 9. Schematic of initial version of the circuitry used to generate
rotating electric field
28
-------
Model SA-10 stereo unit with an output power of about 1 watt per channel. Two
power transformers were used as step-up transformers and the final field poten-
tial thus obtained was obtained 1800 V peak-to-peak. The two complementary
RC circuits were used to produce the desired 90° phase shift between the two
voltages. The final version of the field generating circuit will be described
further ahead in this report.
The signal detection system used with the laboratory breadboard consisted of an
integral photomultiplier-high voltage supply-dynode divider to be discussed in
more detail within the instrument description, followed by Princeton Applied
Research HR-8 Lock-in amplifier and/or a direct oscilloscope display using one
of the quadrupole waveforms as horizontal trigger, in order to discern the
pulse trains associated with the detection of discrete fibers. Final testing
and system optimization were, however, performed using the complete advanced
signal processing circuit developed for the portable prototypes.
Flow Tests
Early during the developmental efforts it had been determined, as mentioned
previously, that flow laminarity was a condition s-ine qua non for the success-
ful operation of the FAM. In order to achieve fully developed flow laminarity
the following conditions had to be fulfilled in addition to the usual low Rey-
nolds number criterion (less than about 1000 for a duct consisting of several
sections with some degree of interfacial discontinuity): (a) inlet diameter
comparable to duct diameter, (b) duct length (from inlet to sensing volume) to
duct diameter ratio exceeding about 12 for the selected diameter of 1 cm and
flow rate of 2 liters/minute, (c) absence of any obstructing elements in the
flow duct upstream of the sensing region, and (d) absence of any measurable
leaks at any point within the flow-sensing duct leading to flow distortions.
The above requirements and the actual flow laminarity were confirmed experi-
mentally using the smoke injection method depicted in Figure 10. The test
aerosol (cigarette smoke) was injected axially with a hypodermic needle into
a 1 cm ID glass tube and a high intensity lamp was used to axially illuminate
the entire volume of the tube. At low flow velocities entering through the
right-angle inlet bend, the injected smoke was observed as a thin straight
line traversing the length of the tube parallel to the duct axis. When using
an inlet diameter of about 0.48 cm (i.e., less than one-half the flow duct
diameter), onset of turbulence was clearly observed even at a duct Reynolds
number of only about 130. By increasing the inlet diameter to 1 cm laminarity
at the sensing region was preserved up to a Reynolds number of 400 corresponding
to a flowrate of about 3 liters/minute through the 1 cm diameter flow tube;
i.e., well in excess of the 2 1pm design objective.
Fibrous Aerosol Generation
The successful development of the FAM was contingent on the reliable laboratory
generation of fibrous-shaped aerosol particles. It had been assumed, perhaps
naively, that the stable and repeatable production of such aerosols with dis-
crete nonagglomerated fibers at typical concentrations of a few fibers per
cubic centimeter was an easily achievable objective. As it became apparent
that these conditions of generation were extremely difficult to obtain with
the commonly available methods of aerosolization, an increasing and sustained
29
-------
TO PUMP
i >
OJ
o
SMOKE INJECTING CAPILLARY
AIR
INLET
§r Im
\
'/Z&* ^
p
GLASS TUBE— 7
r
A
OUT
GLASS WINDOW
Figure 10.
Flow test system used to observe air flow streamlines.
The glass tube was 25 cm in length with a 1 cm I.D.
LIGHT
-------
effort had to be expended in order to develop the required methodology and
instrumentation required to produce a useful fibrous test aerosol. The
successful development of a simple and at the same time reliable generator
should be considered an extremely useful byproduct of this program.
Initially all tests were performed using aerosol generation by atomization or
nebulization from a hydrosol suspension of asbestos in water. The use of a
Collison Generator resulted in frequent and usually rapid clogging of the small
orifices of that device, and consequently a glass De-Vilbiss No. 40 pharmaceu-
tical nebulizer was used for the initial experimental effort.
Freshly prepared 0.01 percent (by weight) aqueous crocidolite hydrosols were
thus aerosolized to provide straight test fibers. When this aerosol was in-
troduced into the laboratory breadboard of the FAM and the detected signals
were displayed on an oscilloscope it was consistently observed that the number
concentration of long fibers, as characterized by pulse trains consisting
of narrow pulses of large amplitude, decreased noticeably as a function of
time (of the order of a few minutes) elapsed from the start of the generation
process. Concurrently, an increase in the number of broad pulse trains,
associated with short fibers, was observed. Eventually a point was reached
when the signal resulting from these short fibers no longer consisted of
separate pul'se trains typical of discrete particles but became a continuous
background signal. These observations, repeated at several different times,
strongly indicate that the aerosol fiber generation process by nebulization
gradually breaks up the longer fibers either at the point of generation; i.e.,
the nozzle, or upon impact of the larger droplets against the nebulizer wall.
To confirm these observations the experiment was repeated by collecting a sam-
ple of the fibrous aerosol on a 37 mm type AA Millipore filter. Three consecu-
tive filter samples were thus obtained, one during the first 10 minutes of
generation, one during the second 10 minutes, and the last during the third
10 minutes. Fiber counts were made for the first and last filters thus ob-
tained, following the NIOSH method of phase contrast microscopy, with the
modification that both fibers with length L less than 5 ym but with aspect ra-
tios greater than 3:1 and exceptionally long fibers were also counted and
categorized. This entire experiment was repeated twice. The first time an
equal number of viewing fields from the first and last filter were compared.
The second time the total number of fibers counted on the first and last filter
was kept constant. These results are summarized in Tables 1 and 2 and although
the fiber counts were statistically insufficient, these results can be consid-
ered as qualitative confirmation of the results obtained with the FAM breadboard.
It became evident from these tests that the FAM even in that partially complete
state of development was an extremely useful tool for the detection and real-
time analysis of fibrous aerosols and provided some heretofore unsuspected
information on generally-accepted methods of generation. These tests also sug-
gested that the development and testing of the FAM system required a more
reliable and stable method of fibrous aerosol generation capable of providing
unagglomerated particles at concentrations of up to 20 or 30 cm"3.
31
-------
Table 1. Comparison of fiber lengths as a function of nebulizer operation
generation time. A total of 60 fields were observed for each
filter.
time period
after start-up
Initial 10 minutes
Final 10 minutes
Number of fibers
L < 5 ym
41
-31
5 ym < L < 37 ym
16
8
37 ym
2
0
< L
Table 2. Comparison of fiber lengths as a function of nebulizer
operation time. A total of 51 fibers were observed
for each filter.
Time period
after start-up
Initial 10 minutes
Final 10 minutes
Number of fibers
L < 5 ym
36
42
5 ym < L < 37 ym
13
8
37 ym
2
1
< L
Fiber break-up in the nebulizer furthermore prevented the generation of,any
fibers whose length exceeded about 8 ym as indicated by microscopy of collected
samples, although the suspension itself (before nebulization was started) was
observed to contain numerous fibers exceeding 40 ym in length. Other methods
of generation were tested, among them the spinning disc technique, but signifi-
cant fiber agglomeration was generally observed. Before the FAM operating-
parameters could be adjusted and optimized it was imperative to generate reliable
signals associated with discrete straight fibers. The gating angles or phase
angles of the instrument could only be adjusted to their proper values with that
type of fibrous aerosol. Thus the-reliable generation of such discrete, fibers
of length greater than 5 ym had to be achieved.
Two methods of dry generation of fibrous aerosol-were considered at that time:
the first was the technique described by Timbrell(33) and others, utilizing a
plug of asbestos continuously.fed into a rotating blade; the second method dis-
cussed by Spurny(3lt) -involved the use of a fluidized asbestos (or fibrous dust
in general) bed subjected to low frequency vibrations.
In order to .evaluate the applicability of the first method to the test require-
ments of the FAM, a visit was made to Dr. David Hemenway at the Civil Engineer-
ing Department of the University of Vermont. A rotating blade device obtained
in England was being utilized by-the above-mentioned group in Vermont." This
device consists of an asbestos plyg slowly fed into a rotating blade in the
standard manner used by Timbrell, '^' and Ortiz.^35'
The asbestos (chrysotile UICC) from the generator was fed into a vertical elu-
triator made from acrylic and of dimensions of approximately 0.3mxlmxlm.
The output of this elutriator was then fed into a second acrylic chamber of
approximate dimensions 0.6 m * imx im high containing a fan. This chamber
was used to damp out any variations in generation rate and to further break up
the fibers.
32
-------
The output of this generator was sampled and collected on a Nuclepore filter
and subsequently coated with a thickness of a few Angstroms of platinum for
observation with a scanning electron microscope.
The asbestos aerosol thus generated was found to be inadequate for the FAM
testing requirements because of excessive fiber agglomeration. It was then
decided to investigate the use of the fluidized bed approach because of its
inherent simplicity. A small vibrating table was constructed as depicted in
Figure 11. A vibrational amplitude of 0.75 mm peak-to-peak defined by the
eccentricity of the drive shaft was established. The vibrational frequency
was made variable from 0 to 40 Hz by adjustment of the voltage applied to the
small d.c. drive motor. The fluidized bed consists of a 1.2 cm diameter filter
on which a layer of about 0.5 cm thickness of the fibrous-material to be dis-
persed is placed. The flow rate through the bed was adjusted to result in an
upward velocity of about 0.5 to 1.5 cm/sec, corresponding to the settling
velocity of unity density spheres of 12 ym and 22 ym in diameter, respectively.
As shown in Figure 11, the flow containing the fibers was diluted with either
room or filtered air in order to obtain the total flow rate of about 2 liters/
min required for the operation of the FAM. Most of the work was performed at
a vibration frequency of the generator of 35 Hz and a flow rate of 50 cm3/min
(0.7 cm/sec).
Fiber Reference Counting
The concentration of airborne asbestos fibers during the FAM tests was determined
by the NIOSH method(7' as described in a manual prepared by the NIOSH Division
of Training. Salient features of this method are that Millipore type AA mem-
brane filters, used to collect the aerosol, are cleared for transmission optical
microscopy by dissolving a filter wedge with a solution of dimethyl phthalate
and diethyl oxalate on a glass slide under a cover slip, and examining the now
transparent filter with a phase-contrast microscope at approximately 450X
magnification to determine the number of fibers greater than 5 ym in length.
Sizing of the fibers is accomplished by the use of a Porton reticule, which is
simply a rectangle with a series of circles the areas of which double from one
circle to the next larger circle, that is scribed on one of the ocular lenses
of the binocular microscope. The Porton reticule, in turn, is calibrated by
comparing the circles with a stage micrometer, which is a microscopy slide that
has 0.01 mm (10 ym) lines etched on it. For some of the comparison tests between
the FAM and the NIOSH sampling method, fibers were characterized as less than
5 ym in length, or greater than 12.5 ym in length. All fibers with an aspect
ratio (length to width ratio) of less than 3:1 were not counted.
The microscope used for fiber counting was an American Optical Microstar 10 out-
fitted with an Ortho-Illuminator light source, a phase condenser, green filter,
and 40X phase objective. Photomicrographs were attainable by use of a Polaroid
camera attachment to the microscope.
PROTOTYPE DESIGN
Optical System
Figure 12 is a schematic illustrating the optical configuration of the final
version of the FAM prototype. Figure 13 is a photograph of the actual components
33
-------
-DILUTION AIR
U)
I —»r0 FAM
PLEXIGLASS ASBESTOS HOLDER
BEARING SHAFT
ECCENTRIC SHAFT
FLEXIBLE
COUPLING
MOTOR
II
Figure 11. Hydrolized bed aerosol generator
-------
LIGHT
TRAP
PRESSURE
DROP 6AU6E
LASER BEAM
HI6H VOLTAGE
QUAORUPOLE
ELECTROOES
LAMINAR AIR FLOW
SECTION
OETECTOR
ORIFICE
/CENTRIFUGAL
r ' BLOWER
BREWSTER
ANGLE
WINOOW
ADJ.
MIRRORS
APERTURE
AJR INTAKE
AIR EXHAUST
Figure 12. Schematic of optical and flow systems of FAM
-------
ELECTRONIC
CARD CAGE
FLOW TUBE
BREWSTER *
ANGLE
LIGHT TRAP
-CENTRIFUGAL
BLOWER
SIDE WINDOW
PHOTOMULTIPLJER
• »Ht PRESSURE
DROP
* A?/ > GAUGE
ADJUSTABLE
MIRRORS
ENTRANCE
BREWSTER
ANGLE
WINDOW
SENSING SECTION
LASER MOUNTING
BLOCKS
Figure 13. Optical and flow elements of the FAM
-------
and their layout. This latter view is of the underside of the top panel of the
instrument on which the optical and flow components, as well as the electronic
card cage are mounted. This panel is made of an aluminum honeycomb sandwich
with an overall thickness of 1 cm with the honeycomb enclosed between two alu-
minum sheets whose thickness is 0.5 mm each. The weight per unit area of this
panel is 0.42 g/cm2, or a total weight of about 0.7 kg which should be compared
with a 4.5 kg aluminum jig plate with equivalent structural rigidity required
for the optical system. The use of the honey-comb panel was dictated by the
contractual overall instrument weight limitation of 13.6 kg (30 Ibs) , which
would otherwise have been exceeded had a solid plate been used.
The helium-neon laser selected for the FAM was a Hughes* model 3222H-P laser
with a larger than 500:1 polarization, with a minimum CW output power of 2 mW
at 632.8 nm in the TEMQO mode, having a beam diameter of 0.63 mm (1/e2) and
divergence of 1.3 milliradians, and a weight of only about 0.2 kg.
This laser was attached to the panel by means of two mounting block-clamps with
rubber lining,, as shown in Figure 13.
The two 45 mirrors are kinematic-mounting orthogonal adjustment front surface
mirrors which are used to fold the laser beam by 180° and to align the beam with
optical axis of the flow-sensing tube thus permitting a rigid mounting of both
the laser and the detection tube.
The optical elements of the flow-detection tube are the entrance Brewster-angle
(56°) glass window, and the exit Brewster-angle (57°) Corning ultraviolet trans-
mitting glass type 7-54 with a thickness of 3 mm which constitutes the light
trap for the primary beam. Conventional light traps generally employ multiple
reflections of the unwanted beam by partially absorbing surfaces; at each re-
flection the beam intensity is reduced but scattering from surface imperfections
and contamination usually produces spurious background radiation. To minimize
these effects the above-mentioned light trap was devised making use of the high
degree of polarization of the laser beam. This trap was composed of a single
glass filter placed at its Brewster-angle to minimize reflection. Virtually
complete light absorption is achieved by selecting the above-mentioned material
whose transmission at 632.8 nm is negligible.
The light scattering photomultiplier tube detector used on the FAM is a 9-stage
side-on type with a multialkali photocathode with a relative responsivity of
about 26 percent, a quantum efficiency of about 2.2 percent and an absolute
responsivity of about 5 x 105 A/W at the He-Ne wavelength of 632.8 nm. This
photomultiplier (RCA type 4840) is packaged within an integrated assembly which
includes a 12 volt d.c. to-high voltage converter, regulation circuit, dynode
voltage divider network and an electrostatic/magnetic shield. This photo-
multiplier assembly is attached to the flow-sensing duct as shown in Figure 13.
A 0.8 mm longitudinal slit within that mounting limits the angular scattering
dispersion to about 1.5° in order to provide the necessary angular selectivity
required to preserve the fiber scattering pulse sharpness information. The
detection volume is viewed by the detector through the gap between two adjacent
*
Hughes Aircraft Co., Industrial Products Div., Carlsbad, California.
37
-------
electrodes of the rotating field quadrupole. These electrodes are thin copper
foils cemented to the inside of a length of Pyrex tubing which serves both as
an insulating support for the quadrupole as well as a transparent element for
the scattered light in its path to the detector. All other optical support
elements including the flow-sensing duct are made of black anodized aluminum.
Sampling-Flow System
The air flow system of the FAM is depicted in Figure 12. The air intake is
a short section of vertical tubing with a diameter of 1.27 cm equal to that
of the flow-sensing tube which forms a 90° angle with the inlet duct. A
20 cm long laminar flow development section follows that bend, followed in turn
by the sensing section whose inner diameter matches the rest of the tube in
order to prevent any flow discontinuities. The length of this Pyrex section
is 6.5 cm of which 2.5 cm constitutes the detection length. An additional
4.5 cm from the end of detection region is allowed to preserve flow straightness
before the exit 90° bend leading'into the centrifugal blower through a flow
constriction across which a pressure drop gauge (Dwyer Minihelic 0 to 0.5 in. H20
range) is connected to provide sampling flow rate indication. For the rated
flow rate of 2 liters/min the indicated pressure drops were 300 newtons/m2 (1.2 in.
H20) and 450 newtons/m2 (1.8 in. H20) for the two prototype instruments.
The air driver is a miniature centrifugal blower powered by a brushless d.c.
motor (Brailsford mod. TB1-1L, 5 to 18 volts operation). This blower is a
low pressure head device of low power consumption (approximately 0.5 watts)
compatible with the generation of a pulsation-free 2 liter/min flowrate through
an essentially open duct system whose major flow resistance consists of the
flow measuring orifice discussed above.
In addition to the sampling-flow system described above the FAM was provided
with an optional, external accessory for insertion into its inlet, consisting
of a vertical elutriator designed to provide an approximate 50 percent aero-
dynamic particle cut-off of 15 ym. Vertical elutriation is, as discussed in
a preceding section of this report, the preferred method of inertial size
selection of fibrous-shaped aerosols, and is compatible with the low pressure
capabilities of the selected air driver.
With the 2 liter/min flow rate and an inner duct diameter of 1.27 cm (1/2 inch)
the axial flow velocity at the sensing volume is about 52.6 cm/sec (based on
a fully developed laminar flow parabolic profile) which combined with a sensing
length of 2.54 cm results in a fiber residence time of about 48 milliseconds.
The volume of the sensing regibn is difficult to establish accurately because
the effective laser beam diameter at the detection region is not known exactly.
It is estimated that this diameter is about 0.7 mm, on the basis of the manu-
facturer specified beam divergence and on the geometry of the system. It should
be understood that the effective beam diameter refers to that dimension within
which fibers receive sufficient illumination to exceed the overall noise level
associated with the entire beam width. The actual effective sampling flow rate
is thus estimated at about 10 cm3/min with respect to the total flow rate of
2 liters/min passing through the duct.
*
Brailsford & Co., Rye, New York.
38
-------
The flow-detection configuration of the FAM, wherein the illuminating beam
diameter is a small fraction of the flow duct diameter was based on two prin-
cipal reasons: (a) the need for a uniform flow velocity, within a laminar
regime, within the illumination region precluding radially variable residence
times, and (b) the requirement of a nearly uniform electric field where the
sensing diameter is small with respect to the interelectrode spacing.
The sensing volume is, based on the above considerations, about 10~2 cm3, which
when replaced in equation (25) indicates a 1.6 percent coincidence probability for
the simultaneous presence of 2 fibers within the detection volume for a fiber
concentration of 20 cm~3.
Quadrupole Electric Field Generator
The rotating field at the sensing region of the FAM is generated by two sinu-
soidal voltages, 90°-phase shifted with respect to each other and applied to
each of the opposing electrode pairs of the quadrupole. The two sine voltages
are generated by the circuitry shown on the schematic of Figure 14, and the
high-voltage transformers of the schematic of Figure 15 (upper right corner).
Referring to Figure 14, a quadrupole oscillator generates two 90° phase-shifted
400 Hz sinewaves which are then fed into a class A/B power amplifier circuit
with feedback. Each of the two sinewaves is amplified separately and the out-
put of -the power amplifiers are fed into the primaries of two separate high-
voltage transformers (Figure 15) with 1:300 turn ratios with grounded secondary
center taps to insure proper potential balance at the quadrupole. The output
voltage of each of the two transformers which is applied to the quadrupole
electrodes is about 3000 V peak (6000 V peak to peak) which results in a nearly
constant axial field intensity of 3000 V/cm (based on equation (13)). The
power amplifier is required to provide the necessary primary drive to overcome
the transformer losses along, since there is no significant output load except
for the small capacitances and inductances associated with the high voltage leads
and the quadrupole electrodes themselves.
The selection of the operating field parameters of 3000 V and 400 Hz for the
rotating field generator were based on several considerations which will now be
discussed.
It was concluded from the initial phases of the program that higher field in-
tensities were to be preferred because of the improved definition in the re-
sulting fiber orientation. It was determined both theoretically and experimen-
tally that fiber jitter as a result of Brownian bombardment decreases with in-
creasing electric field intensity. In addition it was considered that more
intense fields would facilitate rapid fiber charge separation and the concurrent
generation of the required dipole moment. Practical limitations imposed by the
experimental transformers used on the FAM established an upper operating limit
of 3000 V both from the power drive requirements as well as from the insulation
capabilities of these devices.
The frequency of 400 Hz was, as in the case of potential, a result of theoret-
ical and practical considerations. Lower rotational frequencies can be more
easily achieved because of the decreased viscous drag. Operation at higher
frequencies, however, provides a better signal-to-noise ratio since that ratio
follows approximately a I/frequency characteristic. The 400 Hz value was
39
-------
•p-
o
ft ELD AMPLITUDE.
ADJUSTMENT
^/vV
Z.1VI
100^4 13-14
TIP 3Z
^
L_
TIP 31
^
£5
• FIELD MONITOR TEST POINT VF|
-M5Y
SIN °-h&/ "^
BOTTOM VIEW ~'SVl
FREQUENCY DEVICE 5 IMC.
z. QUADRATURE GENERATOR
CLA5S A/B
DUAL POWER
AMPLIFIER
v^a
FIELD AMPLITUDE
ADJUST ME.NT
PM TUBE
DECOUPUNS
RESISTOR
»-— F\ELO MONITOR
/*. *^ 3'
' YZ ° »
I—-^i
54,
» TEST REF.
TEST POfflT
•»• ISV.
-ISV 3
OV. ^
SK
,.AjVA
•^yv\
^ i ^
t«
• >
VR •?
~2_
r'2
-50
3^--
FM TUBE
SA\N CONTROL
OCA/TtONODGT DM3ON >•*
F
-------
L
1
Figure 15. FAM overall wiring layout (note: the two capacitors across the ±15 V supply are 3.3 \if
not 33 yf)
-------
arrived at as a compromise between these conflicting effects as well as based
on the frequency limitations imposed by the particular transformer design.
Signal Processing System
The signal processing system of the FAM was designed to fulfill two principal
objectives: selective detection of discrete fibers and fiber length
discrimination.
For the description of the signal processing circuitry reference will be made
to the system block diagram of Figure 16 and the circuit diagram of Figure 17.
The signal detected by the photomultiplier consists of discrete 800 Hz pulse
trains resulting from the fiber rotation. Although the fibers rotate at a rate
of 400 revolutions per second, each complete rotation results in two scattering
pulses each of which is produced as the fiber axis becomes perpendicular to the
detection axis. Since the residence time, within the detection region, of an
individual fiber is about 48 msec, as mentioned previously, a typical pulse
train consists of about 38 to 40 discrete pulses at the frequency of 800 Hz. The
output from the photomultiplier is preamplified and buffered to provide a low
impedance signal which is then fed into two phase-sensitive or synchronous de-
tection circuits whose theory has been treated by a number of authors.C36"41)
A signal inversion circuit is provided in order to subtract the signal present
when the fiber is viewed end-on from the signal pulse obtained when the fiber
axis is viewed side on (signal peak). This subtraction provides cancellation
of unwanted asynchronous signals such as d.c. shifts, noise, etc. The reference
signals for these gating controls are derived from the electric field generator
circuit (see Figures 16 and 17). The primary gated signal is then fed into two
similar synchronous gating channels labeled PSD No. 1 and PSD No. 2 on Figure 16
corresponding to the "side signal" and "central signal" channels, respectively
of the schematic of Figure 17. These two gated channels provide two separate
outputs required to implement the pulsewidth detection method described in a
preceding section of this report. The electronic implementation of this method
consists of sampling the signal contained within two fiber angular position
ranges, corresponding to the central portion of the scattering pulse and the two
symmetric side wings of this pulse, and integrating the signal over the 40 or so
pulses associated with each passing fiber. The reference gating pulses are
generated with the basic 400 Hz reference signal by frequency doubling and by
generating two pulses of different width centered on the peak of detected signal
and subtracting these two square pulses to obtain the central gate and the two
side gates (see lines CG and SG of Figure 17).
It should be understood that the peak of the detected signal, for negligible vis-
cous friction-associated phase lag, occurs at a phase angle of about 45° with
respect to the peaks of the two high-voltage sinewaves applied to the field qua-
drupole. This 45° angle results from the fact that the optical detection axis
passes between the gap formed by two of the adjacent field electrodes.
The output from the two synchronous detection amplifier stages representing the
central and side parts of the detected pulses, after low-pass filtering, are
then fed into integrator circuits to provide a broad signal pulse with a dura-
tion comparable with the fiber residence time. These two signals (i.e., repre-
senting the central and side parts of the original pulse trains) are then fed
42
-------
PHOTOMULTIPLIER
DETECTOR
ROTATING FIELD
QUADRUPOLE
400 Hz
FIELD GENERATOR
AND
DRIVER
GATING CONTROL
GENERATOR
ANALOG
DIVIDER
PSD
PSD #2
INTEGRATOR
AND ASSOCIATED
LOGIC
COUNTER
AND
CONTROLS
Figure 16. Overall block diagram of electronic subsystem of the FAM
-------
SIGNAL DURATION
TIMER
Figure 17. FAM gating and processing board
-------
into an analog division or ratio module whose output is proportional to the
ratio of the central pulse signal to the pulse wing contribution; i.e., an
output signal related to fiber length as defined by equation (22). This ratio
signal is entered into a comparator whose reference voltage can be adjusted
on the instrument panel to select signals exceeding a selected level to be
correlated with fiber length.
The output signal from the central part of the pulse signal is also fed into
another comparator in order to accept only signals whose amplitude exceeds a
selected minimum. This selection is performed to protect against low signal
or noise associated ambiguities of the ratio operation. The two comparator
outputs are connected to an AND gate which provides an output pulse only if
both of the above-stipulated signal conditions are met; i.e. , ratio and amplitude
above predetermined values.
The resulting pulse is then integrated to perform a time-dependent selection;
i.e., only signals that persist for a period comparable to the fiber residence
time will be processed. A third adjustable comparator performs this last
signal selection.
A ratio criterion override selector has been incorporated on the plug-in board
such that for irregularly shaped fibers the ratio criterion can be bypassed
leaving only the amplitude discrimination as a basis for fiber acceptance.
The output from the last comparator (see Figure 17) consists of negative-going
square pulses, one for each fiber, with a pulsewidth of the order of 25 msec.
The above described circuitry actually applies four different acceptance tests
to every detected signal and only if all of these criteria are met is that signal
train registered as a fiber count: (a) the signal pulses must be synchronous
in frequency and phase with the internal reference waveform, (b) the average
amplitude of the pulses of a train must exceed a preset minimum, (c) the
average ratio of the central to the side portions of the pulses must exceed a
preset minimum, and (d) the duration of the pulse train must exceed a preset
minimum.
Counting-Display Circuitry
The output pulses from the time duration comparator, representing discrete fibers,
are then fed into the counter and display circuit. The schematic of this section
is shown in Figure 18. This circuitry consists of a 8.738 kHz oscillator fol-
lowed by a 2^ ' frequency divider to provide a basic 1-minute counting gate.
The display, a liquid-crystal readout, is driven by a 32 Hz waveform. When
power is applied the unit automatically resets itself and displays one zero to
the left and all zeros to the right of the decimal point. The decimal point
position is controlled by the position of the count time selector switch and
leading zeros are suppressed. Counting periods of 1, 10, 100 and 1000 minutes
are selectable by the panel mounted switch. The display provides counts- of up
to 199,999 fibers. When the "start" button is pushed and released the counter
initiates the counting operation with dynamic display. A "C" in the left-most
digit position indicates it is in the counting mode and the decimal point is
blanked out. At the completion of the counting period the unit stops counting
input pulses, the "C" disappears and the decimal point appears in the proper
45
-------
L/QVIO CKfSTAL DI5PLfir
T/STOP FLIP FL
Figure 18. FAM counter and display boards
-------
position, providing a readout directly in fibers/cm3. The volume of air passing
through the sensing region for a 1-minute sample is equal to about 10 cm3/min,
thus by displaying the fiber count accumulated over 1 minute and by automatic
division by 10 the display indicates directly the concentration in fibers/cm3.
Thus, for the 1-minute count period the display resolution is 0.1 fibers/cm3,
for the 10-minute period it is 0.01 cm~3, for 100 minutes it is 0.001 cnT3 and
for 1000 minutes it becomes 0.0001 cm~3.
Power Supplies and Power Consumption
The basic power source for the FAM is a set of four rechargeable gelled elec-
trolyte batteries Globe type GC660 connected in a series-parallel combination
to provide a nominal output of 12 volts with a total capacity of 12 ampere-
hours. This type of battery is very reliable, relatively inexpensive and can
be subjected to numerous charge-discharge cycles without noticeable degradation
(up to 1000 cycles can be expected if the battery is not allowed to discharge
completely during each cycle). These batteries are completely sealed and can
be operated in any position. Their charge loss under no-load conditions is of
the order of 2 to 3 percent per month at room temperature. Extended storage
at discharged conditions does not result in either cell reversal or loss of
ability to accept recharge.
A Globe charger type GR6 122-000 CDM-12 volts was supplied as a separate item
with each of the two prototype FAM's. This charger provides a variable current
depending on the charge condition of the batteries. When operating the FAM the
charger provides almost the entire replacement current to the system permitting
nearly indefinite operation from the 115 volt a.c. line.
The nominal 12 volts from the batteries (in practice about 13.5 volts for a
fully-charged condition) is supplied to the following elements of the FAM
(see Figure 15): the laser power supply, a d.c.-to-d.c. converter with a high
voltage output required for the operation of the 2 mW He-Ne laser; a d.c.-to-d.c.
low voltage converter with an output of +_ 15 volts for the operation of all the
electronic circuitry, the input power to the photomultiplier assembly, and for
the generation of lower voltages such as +; 5 volts for various circuit modulesi
and an adjustable voltage for the blower motor. All power supplies were selected
on the basis of maximum operating efficiency in order to minimize the battery
energy requirements. The contractual objectives called for a continuous battery
operation of at least 4 hours between charges. This objective was met and some-
what exceeded as can be concluded from Figure 19 which indicates that the FAM
operated continuously for 5 hours with its initial battery charge, during which
the voltage dropped from about 13.3 V to 10.8 volts.
The power requirements of the various elements of the FAM prototype including
the intermediate power required by the voltage converters are tabulated below.
The total current drawn from the batteries averages about 2.8 amperes, which
over a 4-hour operating period corresponds to a capacity of 11.2 ampere-hours
or an energy of 480,000 joules (134 watt-hours).
Globe Battery, Milwaukee, Wisconsin.
47
-------
-p-
00
V)
MINIMUM OPERATING-]
VOLTAGE
6
TJME (hours)
Figure 19. Battery output voltage as a function of FAM operating time
-------
Table
He-Ne laser and d.c. to d.c. power supply combination 24.0 watts
Photomultiplier-power supply assembly 1.5 watts
Blower 0.25 watts
Electronic circuitry 7.75 watts
Total power 33>5 watts
Instrument Packaging Aspects
The contractual stipulations imposed the following limits on the overall instru-
ment: (a) maximum weight without the charger but including the batteries of
13.6 kg (30 Ibs), and (b) maximum volume, excluding charger of 56,600 cm3
(56.6 liters or 2 cubic feet). The first of these two stipulations imposed
stringent design constraints since one of the major weight contributors; i.e.,
the batteries (4.8 kg), could not be reduced given the power requirements and
operating time specifications. The most significant weight reduction was
achieved through the use of the honeycomb panel discussed in more detail in
another section of this report. The final weight of the FAM (excluding the
1.5 kg cover) was about 13.5 kg.
The volume of the instrument without the separable cover was 26,000 cm3, and
with the cover 43,000 cm3, both well within the maximum objective. The outside
dimension of the enclosure are: 53.3 cm in length, 43 cm in depth and 11.4 cm
in height for the basic instrument and 19 cm in height with the cover.
As can be seen in Figure 13 all optical and flow elements, the electronic
circuitry card cage assembly, the controls and display, the blower and flow
readout are mounted under the top honeycomb panel, which can be separated
from the rest of the instrument as an integrated assembly. The batteries,
laser supply, and the low voltage supply are mounted directly to the bottom of
the metal enclosure, a drawn aluminum case. The top panel is mounted on this
enclosure with a foam rubber gasket to prevent access of dust into the interior
of the FAM (see Figure 20). The card cage for the electronic circuitry contains
three plug-in cards with wire-wrap wiring. The entire card cage can be lifted
up into a locked position above the panel for easy access to the electronic
circuitry, as shown in Figure 21.
PROTOTYPE AEROSOL TESTING
Aerosol tests of the FAM both during the developmental stages as well as during
the. final evaluation of the deliverable prototype instruments consisted mainly
of a series of fiber count comparisons between the FAM output information and
optical microscope counts of filter collections of the aerosol downstream of
the instrument. Microscopy was performed following the NIOSH phase contrast
method as discussed in a preceding section of this document. The test were per-
formed by introducing the aerosol, as generated by the device of Figure 11, into
the inlet of the FAM using as much dilution air as required to establish a
2 liter/min flowrate. For these tests an external pump was used placing the
reference filter directly downstream of the FAM and bypassing the internal blower
49
-------
Ul
o
.
Figure 20. View of bottom of FAM enclosure with panel removed
-------
Figure 21. Top view of FAM with electronic circuitry card cage in the lifted-up position
-------
in order to prevent errors associated with fiber losses within that air driver.
Since the reference membrane filter was operated in line rather than in its
customary manner; i.e., as an open face collector, it was found necessary to
incorporate a 13 cm long flow expansion and diffusion section upstream of the
filter using a series of plastic rings with an inner diameter of about 3.2 cm.
This configuration ensured essentially uniform particle deposition across the
filter face, precluding the direct impaction observed when a short (< 2 cm)
distance between the constricted filter holder inlet and the filter was first
used producing a highly nonuniform deposition.
Typical test durations were in the range of 10 to 100 minutes depending on fiber
concentration. The duration of the tests was based on counting statistics re-
quirements of the reference filter collection as well as for the FAM. Concen-
trations in the range of 0.4 to 40 fibers/cm3 as determined by the phase contrast
microscopy method were generated using three types of fibrous aerosols: UICC
Crocidolite, Code 110 fiber glass and UICC Chrysotile. Tests with the former
two materials were performed with the signal ratio threshold setting of 3 (i.e.,
ratio of 3 between the central part of the pulse with respect to the sides of
the pulse) for a central gating width of 4; 4° and side gates of +_ (4° to 8°) .
Although initial testing had been performed for + 3° and + (3° to 6°) it was
considered that the finite degree of pulse jitter observed during those tests
justified the slight widening of the gating widths. The ratio discrimination
was bypassed for Chrysotile because of the extreme irregularity and curving of
these fibers which invalidated the ratio sensing.
The amplitude discrimination was set to a dial reading of 0.5 for all three
aerosols and the tests were performed as described above. The results of this
series of these aerosol tests are shown in Figure 22. The two data points
below 0.5 fibers/cm3 are somewhat questionable because the reference filter
counts were very low resulting in potentially large statistical errors. It
appears reasonable that for a statistically more accurate reference count these
two data points would have been close to the same FAM/filter ratio as the data
points between 1 and 30 fibers/cm3 which fell reasonably close to an average
level of 0.5. The two data points above 30 mg/m3 were at concentrations exceed-
ing the design objective of the FAM; i.e., 20 fibers/cm3, and the response of
the instrument was obviously inadequate as a result of two concurrent factors,
coincidence count losses and the presence of a large continuum background of
short (< 5 ym long) fibers whose generation was concomitant with the production
of these high concentrations.
The results of these tests can be interpreted as follows. For a pulse ratio
discrimination setting of 3 and an amplitude setting of 0.5 the FAM-determined
concentration was about 1/2 that indicated by the reference filter. Time
limitations did not permit to rerun these tests with different ratio and/or sen-
sitivity settings in order to bring the results of the two methods into agree-
ment. A small pulse ratio adjustment; e.g., from 3 to about 2.5 would have
shifted the FAM count output to a FAM/filter count ratio of about 1. The prin-
cipal feature of the data points for concentrations below 30 fibers/cm3 is that
there is essentially a constant relationship between the FAM counts and the
reference filter results. The precise final setting for the threshold controls
must, of course, be obtained by a more extensive testing program and the above
reported results are to be considered only as preliminary, although very support-
ive information on the performance of the FAM.
52
-------
10
~l—I—I—I I I 1
1—I I I 11
z
o
o
u
0.5
0.2
1 0.1
0.05
0.02
Plot of FAM count of fiber concentration divided by filter
count of fiber concentration. Plot is a function of filter
count concentration. Sensitivity setting » 0.5; beam
diameter measured to be 0.6 mm.
o ^
o
A
OUICC CROCIDOLITE ^RATIO =3
Q CODE 110 FIBER GLASS R = 3
AUICC CHRYSOTILE' RATIO DISCRIM. BYPASSED
O
A
O.OL
0.5 1.0 5.0 10
FIBERS/cm3, FILTER COUNT
50
100
Figure 22. FAM test data
53
-------
It should also be considered that the conversion of FAM counts over a given
sampling period to fiber concentration involves a parameter whose exact value,
although constant, is not known; i.e., the effective cross sectional area of
the sensing volume, an uncertainty that has been discussed previously. Therefore
the FAM count duration required to achieve the agreement between the reference
method and the FAM results was only an approximation based on an estimate of the
effective laser beam cross section at the sensing region.
Table 3 lists all the relevant parameters of the tests plotted in Figure 22.
In addition to the tests reported above several qualitative tests were performed
to determine the ability of the FAM to discriminate against nonfibrous aerosols.
Two types of such aerosols were generated, cigarette smoke for small, submicro-
meter particles, and 12 pm diameter plastic spheres to determine the effect of
large particles. Mass concentrations were estimated at several mg/m for both
types of aerosol. Neither one produced any counts on the FAM. Arizona road dust,
however, generated FAM counts at very high concentrations. Microscope analysis
of filter collections of this latter type of dust showed a number of elongated
particles with length-to-diameter ratios exceeding 2.5 or 3 to which the ob-
served counts were attributed. These FAM counts, however, only became signifi-
cant at very high concentrations of dust, possibly in excess of 10 or 20 mg/m^
when uninterrupted streams of such elongated particles was present in the
detection volume.
Table 3. Test data of FAM.
FAM and filter count versus concentration ratio 3,
sensitivity 0.5 (75 mV), gate 8°
Run
Date
Duration
(min)
FAM
count
Filter
count
Fibers
Fields
FAM
concentration
(cm'3)
Filter
concentrat ion
(cnT3)
UTCC crocidolite
1
2
3
4
5
6
7
4/22
4/22
4/28
4/28
4/28
4/29
4/29
10
20
57
40
32
60
63
385
329
1076
770
812
884
485
50
51
159
76
101
95
50
64
88
10
68
76
53
83
4.3
1.81
2.1
2.1
2.84
1.65
0.9
11.7
4.32
41.8
4.19
6.22
4.5
1.4
Fiber glass code 110
8
9
10
11
5/2
5/3
5/3
5/4
17
8
78
100
738
895
705
524
96
102
33
40
96
67
147
122
4.85
12.5
1.01
0.585
8.82
28.6
0.43
0.49
UICC chrysotile ratio discrimination bypassed
12
13
14
15
5/9
5/16
5/17
5/18
15
35
85
35
960
676
407
961
112
48
50
65
31
60
55
25
7.07
2.16
0.53
3.06
36
3.43
1.59
11.1
54
-------
PROTOTYPE INSTRUMENT INSTRUCTION MANUAL
SPECIFICATIONS
• Total fiber concentration range: 0.0001 to 20 fibers/cm3
• Minimum detectable fiber length: 2 micrometer (estimated)
• Minimum detectable fiber diameter: 0.2 micrometer (estimated)
• Counting periods: 1, 10, 100, and 1000 minutes
• Concentration resolution: 0.1 cm"3 for 1 minute period
0.01 cm"3 for 10 minute period
0.001 cm"3 for 100 minute period
0.0001 cm"3 for 1000 minute period
• Maximum count rate: 20 sec *
• Maximum total count: 199,999
• Display: 7' segment, 1.27 cm high, liquid crystal, 6 digit.
• Ratio threshold control range: 0 to 10
• Intensity threshold control range: 0 to 10 (0 to 500 mV)
• Output pulse signal: + 15 volts, 20 msec, 3.5 mA max, (1 TTL gate)
• Flow rate (nominal): 2 liters/minute (adjustable 1.5 to 2.5 liters/min)
• Light source: 2 mW He-Ne laser (X = 632.8 nm)-
• Minimum duration of continuous operation with initially fully charged
batteries: 4 hours.
• Battery voltage (nominal): 12 volts
• Operating current (typical): 2.8 amps
• Weight (without cover): 13.5 kg (30 Ibs)
• Weight (with cover): 15.2 kg (33.5 Ibs)
• Dimensions: 53.3 cm length, 43 cm depth and 11.4 cm height (without
cover), 19 cm height (with cover).
55
-------
OPERATING PROCEDURES
Reference will be made to the photograph of Figure 23. The numbers in parenthe-
sis refer to those called out on that figure.
Identification of Top Panel Component Placement
(1) ON-OFF power switch
(2) START-RESET pushbutton switch
(3) Sampling period selector
(4) Sensitivity threshold control
(5) Ratio threshold control
(6) Count-concentration display
(7) Pulse output connector
(8) Laser ON/OFF momentary pushbutton switch
(9) Battery charger receptacle.
(10) Flow inlet
(11) Flow exhaust
(12) Flow readout gauge
(13) Flow readout zero adjustment screw
(14) Flow rate adjustment control.
Start-Up and Measurement Procedure
Set controls (4) and (5) to 0.5 and 2.5 respectively, or to those set-
tings that have been established by independent comparison measure-
ments. An approximate 5 ym length discrimination is obtained with the
above settings.
Check flow rate gauge zero reading. If necessary correct with screw-
driver adjustment (13).
Place switch (1) in the ON position, this applies power to all
elements of the FAM.
Select the desired sampling period with selector (3).
Allow about a 1-minute warm-up period before proceeding.
Observe flow rate gauge (12), its reading should be between 0.1 and
0.2 inches of water (340 to 680 newtons/m2). The calibrated 2 liter/
minute reading is marked on the rear of the instrument. If the ap-
propriate flow reading is not indicated (after a maximum warm-up
period of 10 minutes), adjust with a small screwdriver the flow rate
to the correct value with control (14) . Correct flowrate will only
be obtained if there are no flow constrictions either at the inlet
(10) or at the exhaust (11) ports.
56
-------
Ul
Figure 23. Top view of FAM
-------
Observe digital display (6). Its readout should indicate 0.0
(for a 1-minute sampling time selector (3) position), 0.00
(for 10-minutes), etc.
Fiber concentration measurement is initiated by pushing down
switch (2) and then releasing it (pushing down resets display
and releasing starts count sequence). A letter C should appear
on the left side of the display (6) indicating that the in-
strument is in the counting mode; at the same time the decimal
point should not be displayed.
Fiber count is displayed dynamically; i.e., as a fiber is detected
the display will increment by one count.
At the completion of the selected sampling time, the letter C on
the left side of the display should disappear and the decimal
point should reappear in the proper place to read out directly
in the units of fibers/cm3. The concentration value remains dis-
played until either the FAM is shut off or it is reset. If the
accumulated fiber count is consistently less than about 10 (e.g.,
1 fiber/cm3 for a 1-minute period) the sampling time should be
switched to a longer period in order to achieve an adequate count-
ing accuracy.
The RESET-START switch (2) can be operated during a count period,
thus starting a new count sequence.
To check system noise, if necessary, or to determine battery charge
condition, the laser power push button switch (8) can be depressed
thus shutting off the laser; power to the laser is restored when
releasing switch (8).
Output pulses for recording purposes are provided at the BNC-type
connector (7). Each pulse corresponds to a single fiber.
When not counting fibers it is desirable to shutroff FAM to con-
serve battery power.
Battery Charging
A separate battery charger has been provided with the FAM. The instrument can
be operated either with the internal battery power or using the external charger
to replace most of the power required by the FAM.
To connect charger to the FAM use the following sequence: shut-off FAM power
(1) ; plug in cable from charger into the FAM receptable (19) ; plug charger line
cord into 115 V a.c. outlet. When disconnecting charger from FAM follow reverse
procedure, making sure that the line cord is unplugged from wall outlet before
removing charger plug from FAM. To run charger at full current, place its own
selector switch to CYCLE, especially when operating FAM and when normal charging
current is to be applied to batteries.
To maintain full charge over extended periods of time when not using the FAM,
place the charger selector to FLOAT.
58
-------
ELECTRONIC TESTING AND ADJUSTMENT PROCEDURES
Signal Detection Circuitry
Reference is made to Figures 17 and 24.
1. Ground the signal input located at the micro-connector on the
Signal Board. A 800 Hz square wave will be present at Vy,
the output of the synchronous rectifier (pin 6 of IC3) . Trim
T^3 until the p-p amplitude of this square wave equals zero
volts.
2. A d.c. offset at Vy may be present. This should be nulled by
T
12'
3. Adjust TIQ until V <|> 1, available at pin 6 of the AD434B analog
divider, equals -10 millivolts.
4. Adjust TI;L until V 4> 2, pin 9 of the analog divider, equals +10
millivolts.
5. Input a 100 mv p-p, 800 Hz, square wave and trim T,, until Vg =
1 volt p-p (IC2, pin 6).
6. Change the input to a periodic square burst signal with a 35 msec
burst duration. The square burst frequency should be 800 Hz with
a coherent, 400 Hz reference signal available. The reference
signal should be input to pin 56 of the Field Board and the slide
switch positioned as to conceal the white dot.
7. Set the ratio knobpot to 1.0 and adjust T until a count just
begins to register.
8. Reset the slide switch on the Field Board so it displays the white
dot and ground the signal input. Adjust T.., until Vft = -4v.
Gating Circuitry
1. Adjust T-^ until the duration of the period pulse present at Z6
pin 1 equals 7 y sec.
2. Repeat this procedure for T2 available at Z6 pin 2.
3. Set the duration of the pulse at Z6 pin 5 to approximately 35°
(240 y sec) by trimming Ty.
4. T-j is used to adjust the duration of delay pulse, which multiplies
the frequency of the gating function. It is important that this
delay represents 90° of the field frequency t 1 90°
x = b p
Hz 360°
sec) .
59
-------
Figure 24. Top views of quadrature field generator-driver and of gating-
processing boards
60
-------
Adjust T3 until the falling edge of this delay pulse, available at
Z6 pin 4, meets this condition.
At this point the desired signal gating duration must be set. We will assume
the center gate (CG) width to be 8° (55.5 y sec) making the side gate (SG)
widths each 4° (27.8 p sec). If the desired sampling differs from 8° the
procedure remains the same; providing the appropriate symmetry is maintained.
5. Tc directly sets the center gate duration. Adjust this for an
8° pulse at Z6 pin 9.
6. Adjust Tg to yield a 16° (111 y sec) wide gate (WG) duration
(twice that of CG) available at Z6 pin 8.
7- The side gate (SG) sampling signal is formed by subtracting the
center gate from the wide gate. The CG, located at Z6 pin 10,
should be centered by T^ yielding a symmetrical SG signal of
two 4° pulses.
The gating system alignment should now be complete. Final adjustment synchro-
nizing the gate phase position with the actual signal phase must be performed
when an actual signal is available from the fibrous sample. At that time T?
should be used to make this adjustment.
Counter-Display Circuit
Reference is made to components on Figures 18 and 25.
1. The sampling period time base is adjusted by means of T15.
This adjustment determines the sampling time time periods
which have been adjusted to 1, 10, 100 and 1000 minutes.
If it is found necessary to deviate from these values by a
constant factor, T15 should be used.
2. The pulse period during which the counting circuit does
not respond to any new fiber can be adjusted by means
of T16, whose factory setting is for 20 msec pulsewidths.
OPTICAL ALIGNMENT PROCEDURE AND CLEANING
Refer to Figures 13 and 23.
1. Remove top FAM panel from case by removing the eight outermost
screws (Figure 23).
2. Carefully place this panel upside-down on top of the case.
3. Rotate the laser about its axis to a position minimizing beam
reflection from the Brewster angle window at the light input
end of the aerosol duct assembly (ADA).
4. Remove the Brewster angle light trap.
61
-------
Figure 25. Top view of counter-display board
62
-------
5. Carefully replace main panel on the carrying case and tighten
the eight screws gently.
6. Remove both side access panels of carrying case.
7. Place FAM with panel facing down.
8. Insert alignment aperture, supplied with FAM, about 1 cm
into beam exit end of ADA,
9. Place power switch (1) to ON. CAUTION: do not look into the
laser directly at any time.
10. Adjust the mirror closest to the laser until the beam goes
through the center of the entrance aperture.
11. Adjust the other mirror closest to the flow tube until the beam
goes through the center of the exit (alignment) aperture by
observing the laser beam spot reflected by a dark surface.
Repeat steps 7 and 8 until beam is aligned with both apertures.
12. Remove alignment aperture and replace Brewster angle light trap.
Be sure both 0-rings are properly in place.
13. Replace side access panels.
NOTE: If main panel hold-down screws are loosened or tightened
realignment may be necessary.
The access of coarse dust or dirt to the FAM inlet should be prevented as this
material would tend to deposit on the window and light trap internal surfaces.
Occasional cleaning of the interior of the flow-sensing duct should be performed,
especially if coarse dust contamination is suspected, by removing the end win-
dows as described in steps 6, 7 and 13 of the above described alignment procedure,
and blowing out the dust from the duct and cleaning the windows with lens tissue.
SAFETY ASPECTS
Two areas of the FAM should be considered from the point of view of operator
safety: the laser and the high voltage supplies.
As mentioned before the laser beam should never be viewed directly or by re-
flection on a reflecting surface without protection glasses. The beam can be
viewed by reflection from a dull black surface such as felt or flat black
optical paint.
There are two potentially accessible high voltage circuits in the FAM: the qua-
drupole field potentials generated at the high voltage secondaries, and the
laser high voltage drive. Caution should be exerted in not touching any of the
four terminals at the quadrupole block, nor to separate the small cable con-
nector between the laser power supply and the laser when power is on.
63
-------
REFERENCES
1. Becklake, M.R., "Asbestos-Related Diseases of the Lung and Other Organs:
Their Epidemiology and Implications for Clinical Practice," Am Rev
Respir Disease, 114:187. 1976.
2. Bruckman, L., R.A. Rubino, and B. Christine. "Asbestos and Mesothelioma
Incidence in Connecticut," Air Pollut Contr Assoc J. 27:121. 1977.
3. Woitowitz, H.J., and H. Valentin. "Arbeitsmedizinische Untersuchungen zur
Gesundheits gefahrdung durch Asbesthaltige Staube," Staub. 36:112. 1976.
4. Federal Register Sec. 1910.93a added at 36 FR 23207, December 7, 1971, as
emergency temporary standard, issued as permanent standard at 37 FR 11318,
June 7, 1972, effective July 7, 1972, redesignated Sec. 1910.1001 at
40 FR 23072, May 28, 1975.
5. Edwards, G.H. and J.R. Lynch. The Method Used by the U.S. Public Health
Source for Enumeration of Asbestos Dust on Membrane Filters. Ann Occup
Hyg. 11:1-6. 1968.
6. Keenan, R.G. and J.R. Lynch. Techniques for the Detection, Identification
and Analysis of Fibers. AIHA J. 587- September-October 1970.
7. Addingley, C.G. Asbestos Dust and Its Measurements. Ann Occup Hyg.
9:73-82. 1966.
8. Sampling and Evaluating Airborne Asbestos Dust, Division of Training,
NIOSH, U.S. Department of HEW, PHS, Cincinnati, Ohio 45202. April 1974.
9. Davies, R., R. Karuhn, J. Graf, and J. Stockham. The Rapid Counting and
Sizing of Fibers in a Mixture Using an IITRI-modified Coulter-Counter,
Am Ind Hyg Assoc J. 36:825. 1975.
10. Tolles, W.M., R.A. Sanders, and G.W. Fritz. Dielectric Response of Ani-
sotropic Polarized Particles Observed With Microwaves: A New Method for
Characterizing the Properties of Nonspherical Particles in Suspension.
Appl Phys J. 45:3777. 1974.
11. Wallach, M.L., and H. Benoit. Light Scattering by Poly-L-Benzyl Glutamate
Solutions Subject to an Electric Field. Polymer Sci J. 45:41. 1962.
12. Schwarz, G., M. Seito, and H.P. Schwan. On the Orientation of Nonspherical
Particles in an Alternating Electric Field. Chem Phys J. 43:3562. 1965.
64
-------
13. Jennings, B.R. and H.G. Jerard. Light Scattering Study of Tobacco Mosaic
Virus Solutions When Subjected to Electric Fields. Chem Phys J. 44:1291.
1966.
14. Kielich, S. Orientation of Colloid Particles in Laser Optical Fields and
Its Effect on Light Scattering of Colloids. J Colloid Interface Sci.
28:214. 1968.
15. Jennings, B.R. and H. Plummer. A Study of Light Scattered by Hectorite
Solutions When Subjected to Electric Fields. J Colloid Interface Sci.
27:377. 1968.
16. Wallach, M.L. and H. Benoit. Light Scattering of Polar Chain Molecules
Subjected to Electrical Field. J Polym Sci. 4:491. 1966.
17. Stoylov, S.P. and S. Sokerov. Transient Electric Light Scattering.
J Colloid Interface Sci. 24:235. 1967-
18. Timbrell, V. Desired Characteristics of Fibers for Biological Experiments,
Inst Occup and Environ Health Conference. Fibres for Biological Experi-
ments. Quebec, Canada. October 29, 1973.
19. Timbrell, V. Alignment of Respirable Asbestos Fibres by Magnetic Fields.
Ann. Occup Hyg J. 18:299. 1975.
20. Fuchs, N.A. Mechanics of Aerosols. New York, Pergamon Press. 1964.
21. Smythe, W.R. Static and Dynamic Electricity, 2nd Edition. McGraw-Hill,
Inc. 1950.
22. Van de Hulst, H.C. Light Scattering by Small Particles. J. Wiley and
Sons, Inc. p. 304. 1957.
23. Farone, W.A., M. Kerker and Matijevic. Scattering of Infinite Cylinders
at Perpendicular Incidence. In: Electromagnetic Scattering. M. Kerker
(ed.). Pergamon Press, Macmillan Company. 1963.
24. Larkin, B.K. and S.W. Churchill. Scattering and Absorption of Electro-
magnetic Radiation by Infinite Cylinders. J Opt Soc Am. 49:188. 1957.
25. Birkhoff, R.D., et al. Light Scattering From Micron-Size Fibers.
Opt Soc Am J. 67:564. 1977.
26. Stoimenova, M.V. Electric Light Scattering by Cylinder-Symmetrical
Particles. Colloid Interface Sci J. 53:42. 1975.
27. Lattimer, P. Light Scattering by Ellipsoids. Colloid Interface Sci J.
53:102. 1975.
28. Lundberg, J.L. Light Scattering From Large Fibers at Normal Incidence.
Colloid Interface Sci J. 29:565. 1969.
65
-------
29. Harris, R.L., Jr. and D.A. Fraser. A Model for Deposition of Fibers in
the Human Respiratory System. Am Ind Hyg J. 37:73. 1976.
30. Fairchild, C.I., L.W. Ortiz, H.J. Ettinger, and M.I. Tillery. "Aerosol
Research and Development Related to Health Hazard Analysis." Los Alamos
Scientific Laboratory Report No. LA-6277-PR, March 1976.
31. Anderson, D.P., M.C. Aughlin, H.E. Ayer, and G.A. Carson. Identification
of the Quality and Quantity of (Biologically Significant) Cotton Mill Dust.
J Occup Med. 15:302. 1973.
32. Lynch, Jr. Air Sampling for Cotton Dust. Transactions, National Confer-
ence on Cotton Dust and Health, University of North Carolina, Chapel Hill,
1971.
33.- Timbrell, V., A.W. Hyett and J.W. Skidmore. A Simple Dispersion for
Generating Dust Clouds From Standard Reference Samples of Asbestos.
Ann Occup Hyg J. 11:273. 1968.
34. Spurny, K. , C. Boose, and D. Hochrainer. Zur Zerstaubung von Asbestfasern
in einem Fliessbett-Aerosolgenerator. Staub. 35:440. 1975.
35. Ortiz, L.W., H.E. Black, and J.R. Coulter. A Modified Fibrous Aerosol
Generator. Los Alamos, Scientific Laboratory, Report No. LA-UR-76-1866, 1976
36. Blair, D.P. and P.H. Sydenham. Phase Sensitive Detection as a Means to
Recover Signal Buried in Noise. J Phys E. 8:621. 1975.
37. Moore, R.D. Lock-In Amplifiers for Signals Buried in Noise. Electronics.
June 8, 1962.
38. Diamond, J.M. A Double Phase-Sensitive Detector for Bridge Balancing.
IEEE Spectrum, p. 62, June 1969.
39. Cole, J.B. and R. M. Duffy. Phase Sensitive Detector and Reference
Generator for Use in Third Derivative Locking of the Frequency of a Laser
to a Saturated Absorption Feature. J Phys E. 7:1019. 1974.
40. Letzer, S.G. Explore the Lock-In Amplifier. Electron Des. 21:104.
October 11, 1974.
41. Lock in the Devil, Feed Him to the Cat, or Take Him for the Last Ride in
a Boxcar. Tek Talk, Vol. 6, No. 1, American Dynamics Corporation,
Cambridge, Massachusetts.
66
-------
CONCLUSIONS AND RECOMMENDATIONS
The development of the FAM, described in detail in the foregoing document,
represents, in the view of the authors, a significant contribution to the
state-ol-the-art of aerosol measurement technology and consequently to the
methodology available for the detection and assessment of potentially health-
injurious airborne fibers.
Although the instrument resulting from this development program is a fully op-
erational and portable device, it should be considered as the prototype it was
intended to be, and that both as a result of governmental agency as well as
further manufacturer conducted testing, additional operational performance
characterization and criteria are expected, and should serve as a basis for
further evolution of this promising technique.
Since the presently used method of fibrous-aerosol monitoring is tedious, rela-
tively complex. It is conceivable that this new technidue may
provide a viable alternative compatible with both short-term as well as a time-
averaged monitoring and concurrent recording and/or data transmission. At the
least, the FAM will serve as an heretofore unavailable tool for the rapid iden-
tification and assessment of sites and conditions associated with excessive
fiber concentrations, for the determination of their real-time fluctuations
and their correlations with processes and control measures.
Although the FAM development program entailed an intensive laboratory testing
activity directed at defining and optimizing the instrument design and opera-
tional criteria, the time and funding constraints limited the extent and depth
of such activities especially in the characterization of the instrument re-
sponse for a range of different fibrous aerosols under a variety of concomitant
conditions. It is thus recommended that the following areas be pursued in
order to more fully establish performance and limitations of the instrument as
well as to provide the directions required to extend and improve this technique.
1. Investigation Into the effects of, and required measurement
discrimination techniques against the presence of large con-
centrations of potentially interfering particles such as plate-
lets, short (less than 3:1 aspect ratio) elongated particles, and
most importantly, irregularly or complex shaped fibers.
2. Determination of the effects of environmental conditions such as
humidity, temperature, external electrical interference, shock
and vibration, and the long-term exposure to high dust concentra-
tions, in the operation of the FAM.
67
-------
3. Investigation of design additions and/or modifications in order to
both sense and alarm system deviations from operating conditions
leading to possible measurement errors as well as to implement cor-
rective measures to compensate for the effect of such conditions.
68
-------
APPENDIX A
ASSEMBLY DRAWINGS
69
& U S Government Printing Office 1978 — 757-141/6772, Region No 5-II
-------
DUCT ^~5EMBLN L.IWOUT
J-375-IQI9
-------
K $ CONTROL PWME
O DEC. 76
J-375-1O32
-------
J-375-1023
------- |