EPA-600/2-76-001
January 1976
Environmental Protection Technology Series
EVALUATION OF
SONICS FOR FINE PARTICLE CONTROL
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research-and Development,
U.S. Environmental Protection "Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed
to develop and demonstrate instrumentation, equipment and
methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U. S. Environmental Protection
Agency, and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
This document is available to the public through the National
Technical Information Service, Springfield, Virginia 22161.
-------
EPA-600/2-76. 001
EVALUATION OF SONICS
FOR FINE PARTICLE CONTROL
by
R. Hegarty and L. J. Shannon
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
Contract No. 68-02-1324, Task 24
ROAP No. 21ADL-029
Program Element No. 1AB012
EPA Project Officer: Dennis C. Drehmel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
January 1976
-------
TABLE OF CONTENTS
Page
List of Figures iv
List of Tables. . v
Acknowledgements vi
Summary 1
Introduction 4
Literature Search „ 5
Theoretical Aspects of Sonic Agglomeration 6
Laboratory, Pilot-Scale and Full Scale Studies of Sonic
Agglomeration 22
Laboratory or Small-Scale Studies of Sonic Agglomeration 22
Studies of Sonic Agglomeration
Summary of Laboratory, Pilot-Scale and Full-Scale Data 32
Comparative Analysis of Sonic Agglomerators and Conventional Control
Systems .33
Conclusions ........ 40
References • • • • • 61
Appendix A - Supplementary Reading List 43
iii
-------
LIST OF FIGURES
No. Pag
1 Orthokinetic Collision of Aerosol Particles. The "Aggrega-
tion" Volume for a Single Approach of Particles is Shown
by the Dotted Line 9
2 Fraction of Particles of Unit Density of Various Radii
Vibrating with Gas at Sound Intensities Between 1 and
100 kc/sec Amplitude 11
3 Flow Lines Around a Streamlined Spherical Particle for
Re = 1 14
4 Hysteresis in the Flow Around an Aerosol Particle; (a) Near
the Center Line, (b) Far from the Center Line 14
5 Diagram Showing How Self-Centering (a) and (b) of Aerosol
Particles Occurs in a Sonic Field 15
6 Diagram of Overall Mechanism Postulated for Sonic
Agglomeration 18
7 Acoustic Coagulation Rate of a Paraffin Fog as a Function of
Sound Intensity (f = 10 kc/sec) . 23
8 Degree of Agglomeration of Tobacco Coke Particles as a Func-
tion of the Product Agtres (f = 10 kc/sec) 25
9 Degree of Agglomeration of Titanium Tetrachloride Particles
as a Function of the Quantity J1'2 (f = 1.6 kc/sec) .... 25
10 Sound Level Required for 50% Agglomeration of a Water Fog
for Different Times and Frequencies 26
11 Acoustic Coagulation Efficiency of Gas Black (pc = 0.35 atm,
f = 2.1 kc/sec) as Affected by Adding a Water Fog ..... 26
12 Collection Efficiency of 93.5-95.0 Scale Furnace Black as a
Function of Time and Exposure (Cyclone Collector) 31
13 Performance of Sonic Carbon Black Collection System (Three
Cyclones in Series) 31
14 Predicted Sound Intensities as a Function of Agglomeration
Index 35
15 Correlation of Data on Agglomeration and Sound Intensity. . . 35
Iv
-------
LIST OF TABLES
No. Page
1 Summary of Industrial Tests with Sonic Agglomerators 28
2 Optimum Frequencies for Sonic Agglomeration, Industrial
Tests 30
3 Estimated Specific Energy Requirements for Example Sonic
Agglomerator 37
-------
ACKNOWLEDGEMENTS
The work'presented in this report was performed by Midwest Research In-
stitute for the Control Systems Laboratory as Task Order No. 24 on Con-
tract No. 68-02-1324. The work was performed by Dr. R. Hegarty, Staff
Associate, and Dr. L. J. Shannon, Head, Environmental Systems Section,
Physical Sciences Division.
vi
-------
SUMMARY
The aim of this investigation was to analyze and evaluate sonic agglomera-
tion as an approach to improving our ability to control fine particulate
pollutants (i.e., particles in the 0.1 to 3.0 um diameter range) emitted
from stationary sources.
Acoustic fields have frequently been cited as an effective means to
rapidly agglomerate particles and droplets to a sufficient size whereby
they can be readily and efficiently collected in a conventional control
device. Particle agglomeration in an acoustic field occurs by a complex
process which is not well understood. Existing theories provide a
reasonably detailed description of the mechanisms assumed to occur in
the agglomeration process, but the equations describing the postulated
mechanisms have not and perhaps cannot be integrated into a single tract-
able mathematical expression. A simpler approach using a modified form
of Smoluchowski's equation for thermal agglomeration has been advanced by
several investigators. The equation describing the agglomeration process
based on this approach is
n
where Ka is the acoustic agglomeration coefficient and t is the resi-
dence time. The acoustic agglomeration coefficient can be related to
several parameters that influence the postulated mechanisms underlying
the agglomeration process.
Mednikov—' has developed the most complete model for acoustic agglomera-
tion. His model subdivides the aerosol into its small particle concen-
tration n2 and its large particle concentration n^ (assumed to be
constant with time). Mednikov's model leads to the result that the
-------
acoustic agglomeration coefficient is proportional to the quantity
1/2}
J where r is the radius of the large particles and J is
the sound intensity. Combining Mednikov's model with the relationship ob-
tained from the modified Smoluchowski equation results in the prediction
that acoustic agglomeration rates are dependent upon the quantity
nlrl
Experimental studies at laboratory, pilot-scale and full-scale level pro-
vide data which confirm most of the parametric dependencies predicted by
the above theoretical approach. In addition, the experimental work has
shown that:
• Sound intensities of 150 to 160 db (0.1 to 1.0 w/cm^) are required
for industrial applications.
. Residence times of 5 to 10 sec are tenable for industrial applications.
. Sonic agglomeration rate is strongly affected by the concentration
by weight of particles and efficiency of agglomeration decreases
rapidly at concentrations below 2 to 5 g/m3 (1 to 2 grains/ft^).
• Highly polydisperse aerosols are easier to agglomerate than a low
polydisperse aerosol (i.e., if particles are all of nearly same size,
acoustic agglomeration is not effective.
. Injection of atomized water into low concentration aerosols seems
to improve sonic agglomeration.
• Physical properties of aerosol particles have comparatively little
effect on acoustic agglomeration.
• Viscosity, temperature, and pressure of gases are generally about the
same in most aerosols and these factors are not important as a rule.
• Agglomerates formed in sonic fields often do not have a high degree
of stability and they may be broken up in the collection device,
especially in cyclones.
• Explosive aerosols can be safely agglomerated by acoustic fields.
-------
The utilization of sonic agglomeration for industrial gas cleaning re-
quires the coupling of an agglomeration chamber to a particulate collec-
tion system. Cyclones, electrostatic precipitators, and fabric filters
have been used or proposed as the collection system with cyclone systems
being the prevalent system in previous applications. For a sonic ag-
glomerator to be of significant help in the collection of fine particles,
the ratio of the final to initial mean particle radius or diameter must
be in the range of 5 to 20. The lower end of the range would correspond
to the use of a high efficiency system as the collector (i.e., electro-
static precipitator, fabric filter, venturi scrubber) while the upper
side would reflect the use of a low efficiency system as the collector
(i.e., cyclone, spray scrubber). To achieve a growth ratio of 5 to 20
in equipment of reasonable length and diameter, sound intensities of
160 db (1.0 w/cnr) or greater will be necessary. Specific energy con-
sumptions of the order of 10 to 20 hp/1,000 cfm appear to be necessary
as a direct result.
Because of the high energy requirements, the application of sonic agglomera-
tion to industrial gas cleaning problems involving predominantly fine par-
ticles does not appear promising;.
-------
INTRODUCTION
Numerous investigations have shown that aerosol particles behave in a
substantially different manner in an acoustic field than in an unperturbed
medium or in a turbulent stream. Soon after an aerosol is exposed to a
suitable acoustic field, particulate agglomeration is noted and particulate
formations are seen moving about in a complex manner. Successful agglomera-
tion of particulates by acoustic fields in laboratory experiments has led
to the development of sonic agglomerators for use in industrial gas clean-
ing applications. However, sonic agglomerators have not enjoyed a wide-
spread utilization, especially in the United States.
Sonic agglomeration occurs by a complex process which is not well under-
stood and existing theory provides only nominal guidelines for the design
of commercial systems. The present study was undertaken to review (1). the
current state of knowledge of sonic agglomeration on both the theoretical
and experimental level, (2) previous industrial experience with sonic
agglomeration, and (3) recent advances in hardware for sonic agglomerators.
The results of this review were then to be used to assess the utility of
sonic agglomeration as a means of enhancing the collection of fine parti-
culate pollutants.
The following sections of this report present the results of the literature
search conducted as part of the task, a discussion of the theoretical
aspects of sonic agglomeration, an overview of the results of laboratory-
and industrial-scale applications of sonic agglomeration, a comparative
analysis of performance, energy requirements, etc., between sonic ag-
glomerators and other control systems, and conclusions and recommendations.
-------
LITERATURE SEARCH
An extensive literature search was conducted as the initial stage in the
task. Various scientific abstracts (e.g., Chemical Abstracts. Engineering
Abstracts, Air Pollution Abstracts) and selected technical journals were
searched for relevant publications. Patent literature was also scanned.
Approximately 500 references dealing with the general subject of aerosol
interaction with acoustic fields were identified. The literature search
clearly indicated that detailed work on sonic agglomeration is predominantly
of foreign origin. In addition to the reference list delineating the
specific papers cited in the teixt, a supplementary reading list is presented
in Appendix A.
-------
THEORETICAL ASPECTS OF SONIC AGGLOMERATION
Several different effects have been postulated as being responsible for
particle agglomeration resulting from acoustic fields including (1) hydro-
dynamic forces between the particles and (2) additional collisions due to
the different vibrational amplitudes of different sized particles. Although
no comprehesnive theory is available, Mednikov— has presented the most
thorough discussion of sonic agglomeration. Mednikov's general approach
will be used as the basis for our discussion of the theoretical aspects
of sonic agglomeration.
Acoustic coagulation (i.e., sonic agglomeration) involves the response of
aerosol particles suspended in a gaseous medium to forces arising from an
impressed sound field. In general we are concerned with particles above
0.1 um in diameter. Below 0.1 urn, Brownian agglomeration is a more rapid
and efficient means of particle agglomeration. The speed with which aerosol
particles react to forces is obviously important. The rate at which the
equilibrium state of a particle-medium system is restored, as well as the
sensitivity of the particle to changes in the forces acting on it, is deter-
mined by a parameter called the relaxation time T of the particle, given
by
where p is the density of the (assumed spherical) particle, T\ is the
dynamic viscosity of the medium, and r is the radius of the particle.
The smaller the relaxation time of the particle, the more rapidly the
particle takes on a new velocity relative to the medium. Normally, the
relaxation time is very small for aerosols undergoing coagulation. For
example, a particle with r = 0.1 um has a relaxation time T *** 10"^ sec,
while a particle of radius r = 10 um has a relaxation time T <** 10~^ sec.
-------
From the above, it can be seen tltiat if the particles suspended in a
vibrating gas are quite small, they are carried along with the motion,
while if the particles are quite large, they will remain virtually
stationary in the vibrating medium. One measure of the degree to which
any specific particle is carried along with the motion of the medium is
the ratio of the amplitude of displacement (or velocity) of the particle
undergoing vibration to that of the vibrating medium. This ratio, called
the degree (or coefficient) of entrainment of the particle, is given by
Mednikovi' as
(2)
where u> is the angular velocity equal to 2TTf and f is the frequency.
The values of u will range between zero and one. For very large par-
ticles, Up *» 0 as the relaxation time T becomes very large, while
for very small particles, up <« 1 and the aerosol particles will follow
the vibration of the medium. Another parameter that is useful is the
flow-around factor Ug , given by:
(3)
(1 + 0,2^)1/2 '
Note that the entrainment factor Up and the flow-around factor ug obey
the following relation:
up2 + ug2 = 1 . (4)
Employing the concepts of relaxation time and entrainment factor (or
flow-around factor), we can begin a discussion of the way in which
acoustic coagulation of aerosols is envisioned to work. Agglomeration
of particles requires collisions between the particles. According to
many investigators, it appears that the so-called orthokinetic interac-
tion between aerosol particles is the primary mechanisms accounting for
particle collisions and subsequent coagulation in acoustic fields.
Basically, orthokinetic interaction occurs between two particles of dif-
ferent dimensions (or densities) which are moving at different velocities.
Because of their relative motion, they will, in the course of this motion,
approach or separate from each other. This interaction occurs during
vibratory, drift and pulsational motion of the particles. However, the
orthokinetic interaction which occurs when the particles are executing
vibratory motion is by far the most important in acoustic coagulation.
-------
Many authors postulate that orthokinetic collision between two vibratory
particles becomes possible when the smaller particle gets into the
"aggregation volume" of the larger particles (see Figure 1). This volume
is generally considered to consist of a cylinder with a radius equal to
the sum of the radii of the two particles.
(5)
and an altitude equal to twice the difference between the amplitude of
the displacements that the particles undergo during vibration.
ha = 2 (A-A ) . (6)
P2 Pi
If Ag is the amplitude of the displacement of the gaseous medium, then
we can employ the concept of the particle entrainment factor and write:
V ' *«
and thus
ha = 2 (n - up ) Ag . (8)
2 1
To complete the picture, the aggregation volume would have hemispheres
at the ends. According to Mednikov, however, this picture of the aggrega-
tion volume needs modification. First, it neglects the phase differences
between the vibrations of the interacting particles, and second, it ignores
flow in the medium around the particles, which results in the particles
being deflected somewhat to the side as they approach each other.
To satisfy the first objection, Mednikov proposes that the altitude of
the aggregation volume be written as:
-------
Figure 1. Orthokinetic collision of aerosol particles. The "ag-
gregation" volume for a single approach of particles
is shown by the dotted line.
-------
ha = 2 u12 Ag (9)
where
- 2
is the degree of relative motion of the particles. Examination of Eqs.
(2), (3) and (10) indicates that there is an optimum frequency of sound
at which u-^ will be a maximum. If the derivative of Eq. (10) is
taken with respect to cu and set equal to zero, the following equation
for the optimum frequency is obtained
9T]
£opt •
r2
where r2rr = i-"~ *-s *-^e reduced radius of the smaller particle.
<,
Brandt, Freund, and Hiedemanniz' have developed a similar relationship
for single particles by neglecting the buoyancy effect of the gas on the
particle and assuming that Stokes" law can be applied to the relative
motion of the particle through the gas. The ratio of the amplitude of
particle vibration to the amplitude of vibration of the gas was found to
be
*g f l"TTd2fP I2 1 1/2 <13>
" + 1
where d = particle diameter
C = Cunningham correction factor.
For constant particle density and gas viscosity, Eq. (13) simplifies to
A ,
T
A
g [kdf
10
-------
For particles of unit density vibrating in air, the relative amplitude
has been plotted (Figure 2) for frequencies varying from 1 to 100 kc/sec.
For the highest frequencies (50 and 100 kc/sec) the curves can only be
taken as approximations, as the assumptions regarding the Stokes' law
region no longer hold.
o>
c
u
p
1.0 10
Radius of Particles, p.
20
Figure 2. Fraction of particles of unit density of various radii
vibrating with gas at sound intensities between 1
and 100 kc/sec amplitude. The broken line at 0.8
(80%) indicates the size above which nearly all
particles are considered to vibrate with the gas.—'
Figure 2 shows how, until a certain particle size, the aerosol particles
swing along with the vibrations in the gas. This size can be called the
critical particle size for the particular frequency, while a critical
frequency for a particular particle size can be similarly calculated.
The critical particle size is indicated by the point where the curves
enter the steep gradient after their initial shallow decrease, which
occurs at approximately the 80% value of Ap/AK shown by the broken line
in Figure 2. Substituting 0.8 for Ap/Ag in Eqs. (13) or (14) gives
the critical particle diameter dc
2.2
(15)
or
de'f
4 x
m2 /
sec
(16)
11
-------
Equations (15) and (16) indicate that as a first approximation, a unique
relation exists between the critical particle size and the frequency of
vibration of the gas, which for particles less than 7 um occurs at
frequencies greater than 1 kc/sec and extends into the ultrasonic region.
These equations enable the critical frequency or particle sizes to be
readily calculated. Frequencies of tens to hundreds of kilocycles are
indicated for fine particles. However, frequencies this high attenuate
so rapidly with distance from the sound source in gases that the aerosol
flowing through the chamber will encounter no acoustic fi.eld for the
greater part of the time. Thus, lower frequencies would have to be used
in any actual application.
The second objection to the simplified picture of the aggregation volume
concerns the trajectory of the smaller particle as it approaches the
larger particle. If the smaller particle were so small that it was of
size comparable to the gas molecules, it would follow the motion of the
gas along its original line of flow and the possibility of its being
captured by the larger particle is equal to zero, since all the flow
lines go around the larger particle without intersecting it. However,
actual aerosol particles have inertia, so that the smaller particle is
deflected from its original flow line. In this case, the aggregation
volume is of the form shown by the dotted line in Figure 1. Mednikov
defines the capture coefficient of particles, e , by the following
relation:
rr
(17)
where y is as given in Figure 1.
In effect, e is the reduced radius of the aggregation volume. Mednikov
derives estimates for the minimum and maximum values of this reduced radius;
r2rr
and
12
-------
where eT =11 +
L V k-kcr
2
emax - r\T*L + r2rr)
\ /
0.75 log 2 k\
16 - 3 Re
:4 + 12 R.,'
and Uo^ and T2 are the relative velocity and relation time of the
small particle, and Re is the Reynolds number for the flow. Thus, the
aggregation volume turns out to be considerably less than the volume of
a cylinder of radius r^ + r~ .
Although this concept of a largo particle sweeping its aggregation volume
clear of smaller particles by colliding and coagulating with them is the
ultimate means by which sonic agglomeration works, it alone cannot ex-
plain the extent to which acoustic coagulation has been observed to occur.
For once the aggregation volume has been cleared of small particles (dur-
ing one period of vibration), no more collisions or coagulation would be
possible. Thus, there must be one or more mechanisms which account for
particles refilling the aggregation volume and allowing coagulation to
continue. In actuality, two primary and two secondary mechanism for re-
filling the aggregation volume have been suggested.
The first primary mechanism accounting for the refilling of the aggrega-
tion volume is called the parakinetic interaction between aerosol par-
ticles. This type of interaction is a result of the distortion of the
fields of flow around the particles during relative motion, and is due to
the difference in dimensions or densities. Mednikov shows that the flow
lines around aerosol particles in a sonic field are asymmetric in the
direction of flow (see Figure 3), so that a particle moving toward another
one during one-half cycle of vibration will follow a different flow line
than it will as it moves away during the other half of vibration. Figure
4 shows what would happen to a gaseous particle vibrating near an aerosol
13
-------
Figure 3. Flow lines around a streamlined spherical particle
for Kg = 1.
3x
X
L
33'
Figure 4. Hysteresis in the flow around an aerosol particle;
(a) near the center line, (b) far from the center
line.
14
-------
particle. In (a), a particle of the medium is vibrating close to the
center line of the aerosol particle (x-axis) and far to the left of the
y-axis. During one-half cycle of the vibrations, the particle moves from
Point 1 to Point 2 along one flow line. When the motion reverses it-
self, however, the particle moves from Point 2 to Point 3 along another
flow line and thus returns to a position closer to the center line. In
like manner, a particle far from the center line and near the y-axis will
also move along two different flow lines during one complete cycle, but
in this case, it will find itself further from the center line. An
aerosol particle would behave similarly when vibrating near a larger par-
ticle, except that its inertia would cause it to deflect slightly from
the original flow lines. Mednikov has termed this tendency of a aerosol
particle to move toward the center line "self-centering" while the op-
posite tendency is called "self-decentering."— Figure 5 shows the ef-
fect of inertia on these tendencies. Indications are that the parakinetic
interaction between aerosol particles is extremely large. As a result
of the self-centering phenomenon, orthokinetic collision is not limited
to vibrating particles initially in the aggregation volume, but is also
possible for particles that are far outside. In fact, according to
Mednikov, the parakinetic effect: is the primary interactional effect out
to a distance of 40 to 60 particle radii from the coagulation center
(i.e., aggregation volume).
3<*
3'cxQ
Figure 5. Diagram showing how self-centering (a) and
(b) of aerosol particles occurs in a
sonic field.
15
-------
The second primary mechanism accounting for particles entering the
aggregation volume is the so-called attractional interaction between
aerosol particles. This interaction, like the parakinetic interaction,
results from the mutual distortion of the fields of flow around the par-
ticles, but in this case, the interaction is due not to the asymmetry of
the flow lines but to the asymmetry of the velocity fields around the
particles. Mednikov shows that attraction occurs only for particles of
different sizes, a fact not realized for many years. The reason for this
attraction is that, as two particles vibrate, the particle trailing
during one-half cycle enters a "dead" zone behind the leading particle.
In this dead zone, the medium is moving more slowly than it is in front
of the leading particle. Thus, the trailing particle has a greater rela-
tive speed with respect to the medium than does the leading particle,
and the two particles approach. During the second-half cycle, the two
particles exchange roles, and again they approach each other because of
the asymmetry of the velocity fields.
It is noted by Mednikov that the attractional approach rate of aerosol
particles is strongly dependent on the frequency. For each size particle,
there is an optimum frequency at which the attractional approach rate
of the particles is a maximum. The optimum frequency decreases as the
particle size increases, a trend also noticed in the orthokinetic inter-
action mechanism. Both the parakinetic and attractional interactions
can easily account for the rapid refilling of the aggregation volume and
continued coagulation observed in a sonic field. As was already mentioned,
the parakinetic effect is capable of bringing particles into the aggrega-
tion zone from as far away as 40 to 60 particle radii from the coagula-
tion center. Significant attractional interaction effects seem to occur
at almost twice this distance (i.e., 100 to 120 radii).
However, eventual collision and coagulation is not restricted to particles
within the extended region of parakinetic and attractional interaction.
Two secondary mechanisms can be isolated which can account for particles
coming close to a large particle (coagulation center) so that the primary
mechanisms can take over. The first such mechanism is that of particle
drift. Aerosol particles can experience drift in an acoustic field due
to a number of causes. Particles are found to drift due to the action
of the acoustic wind, which is a translational acoustic streaming in the
medium. The drift velocity of a particle due to this acoustic wind is
usually greater than the drift velocity due to any other cause and is
found to be proportional to J/Tj where J is the sound intensity, defined
as the flux of acoustic energy passing through 1 cm2 of the sound wave
front per second. The most commonly used unit of sound intensity is
16
-------
watts per square centimeter. Another definition which will be useful
is that of the sound level L , expressed in decibels (db). The sound
level is related to the sound intensity by the equation L = 10 log^o J/JO ,
where J0 = 10" "> w/cnr. Although it is not certain, a commonly accepted
opinion is that the acoustic wind is a result of the radiation pressure
gradient set up by absorption of the sound waves in the medium.
Four other causes of particle drift in an acoustic field can be identified.
Particle drift might be due to the difference in phase of the vibrations
of the particles in the medium. This drift is somewhat retarded by the
periodic viscosity changes in the medium, and is the predominant type of
drift in an undistorted traveling wave and does not exist in a standing
wave field. The maximum value of this drift velocity is about 0.5 cm/sec.
Particle drift due to wave form distortion is the predominant type of
drift in a distorted traveling wave or a distorted standing wave with
large particles (r > 5 urn). The drift velocity can reach quite high values,
on the order of 10 cm/sec or more, if the amount of distortion and the
asymmetry are quite large. Particle drift due to acoustic radiation
pressure is significant only for large particles in an undistorted stand-
ing wave. For very large particles and high frequencies, the velocity
can reach very high values, such as 10 to 40 cm/sec. The last cause of
particle drift is the asymmetry in the vibrations of the medium in a
standing wave field with small sized particles. This drift velocity
might reach a value of 0.5 cm/sec. The drift velocity of aerosol par-
ticles is strongly dependent on the particle size and the frequency. In
addition, the drift velocity of a particle due to any of these causes is
proportional to the sound intensity J .
The other secondary mechanism by which aerosol particles might approach
a coagulation center results from the interaction of particles with the
pulsations of the vibrating medium. With the concentration gradient set
up by the clearing of the aggregation volume after each cycle of vibration,
distant particles might approach the coagulation center in a purely dif-
fusional way. In addition, particles of different sizes will have dif-
ferent velocities due to the turbulence of the medium.
The overall mechanism which has been postulated for sonic agglomeration
can be summarized with the aid of Figure 6. The basic mechanisms account-
ing for acoustic coagulation, as described in the preceding paragraphs,
are depicted in Figure 6. The actual particle collisions and resulting
coagulation take place in the aggregation volume, where the orthokinetic
interaction mechanism is dominant. In this volume, the aerosol particles
take part in the vibrations of the medium to differing degrees, and the
resulting collisions account for the coagulation noticed in aerosols under-
going sonic treatment. It should be noted that the actual size of the
17
-------
Orthokinetic Interaction
(Aggregation Volume)
Parakinetic
Interaction
Flowline
Distortion
I
Drift
1. Acoustic Wind
2. Phase Difference
& Viscosity Change
3. Waveform Distortion
4. Radiation Pressure
5. Asymmetric Vibration
of Medium
Attractional
Interaction
Actual Collisions &
Coagulation Mechanism
Primary Refilling
Mechanisms
Velocity Field
Distortion
1
Pulsations
Secondary Refilling
Mechanisms
]. Diffusion
2. Turbulent
Figure 6. Diagram of overall mechanism postulated for sonic agglomeration.
18
-------
aggregation volume is a function of the relative sizes of the interacting
particles, and thus two small particles of different sizes approaching
the same large particle (coagulation center) would experience different
aggregation volumes.
St. Clair—' objected to orthokinetic interaction as the main coagulation
mechanism because he could see no way for additional particles to refill
the aggregation volume once it had been swept clean by the large particle.
Mednikov and others refute this objection, however, through the use of
the two primary mechanisms of parakinetic interaction and attractional
interaction. In the parakinetic interaction, once a small particle ap-
proaches to within about 50 particle radii of the coagulation center, the
distortion of the flow lines around the interacting particles causes the
small particle to enter the aggregation volume. Likewise, in the attrac-
tional interaction, the distortion of the velocity fields around the par-
ticles will draw a particle into the aggregation volume after it has ap-
proached to within 100 particle radii of the coagulation center. Thus,
these two primary mechanisms account for the refilling of the aggrega-
tion volume once the particles approach to within a certain distance of
the coagulation center.
There are also two secondary mechanisms which can account for particles
coming close enough to the aggregation volume for the primary mechanisms
to take over. Particle drift can be caused by a number of factors, the
most important of which is the acoustic wind. The turbulence of the
medium is also important as the particles take part in turbulent pulsa-
tions. Thus, particles very far from a coagulation center can eventually
enter the aggregation volume and be coagulated. It has been tacitly
assumed in the discussion of these mechanisms that the flow velocity is
relatively low. For very high velocities, inertial effects would greatly
complicate matters and the above mechanisms would have to be modified.
Equations describing the postulated processes involved in sonic agglomera-
tion have not and perhaps cannot .be integrated into a single tractable
mathematical expression. A simpler approach using a modified form of
Smoluchowski's equation for thermal agglomeration has been advanced by
several authors including Mednikov. According to Smoluchowski the normal
coagulation rate of an aerosol will obey the following relation:
(20)
19
-------
where n is the particle count concentration, K is the coagulation
constant, and t is the time. Considering the orthokinetic interaction
process as the main collision and coagulation mechanism, it seems reason-
able to say, first, that the number of large particles, ni = n^o » re-
mains a constant, and second, that the number of small particles, n2 ,
is much larger than n^ (n2 » ni). In this case, Eq. (20) becomes:
a?
If the particle count concentration (essentially n2 as a result of
preceding assumption) at t = 0 is no , the solution of Eq. (21) is:
n = noe" . (22)
If the assumption is made that the primary and secondary filling mechanisms
can keep the aggregation volume filled with small particles in such a
way that the rate of coagulation is dependent on the speed of the ortho-
kinetic interactions, then one might say that a state of reaction control
had been achieved. If this is the case, the acoustic coagulation con-
stant Ka must be proportional to the number of large particles
n!0
Likewise, the coagulation constant should also be proportional to the
volume of the aggregation zone. If the aggregation volume is taken to be
a cylinder with radius e , as given by Eq. (17), and altitude ha , given
by Eq. (9), then Ka can be written as:
(2 uu Ag) (24)
where p is a filling factor which must be determined experimentally.
The acoustic coagulation constant Ka can now be written:
Ka = 2npner2uA (25)
20
-------
Ag is the amplitude of the displacement of the medium (i.e., gas molecules)
and can be related to the sound intensity J by the following equation:
Ag =: 3- (26)
where Pe and Cp are the density of the medium and the speed of sound
6 o
in the medium, respectively. Considering Eqs. (25) and (26), it can be
seen that:
Ka ~ J1/2 (27)
Inspection of Eqs. (22), (25), (26), and (27) shows that sonic agglomera-
tion will be strongly influenced by frequency, particle size distribu-
tions, sound intensity, particle number density, and exposure time. Specifi-
cally, the acoustic agglomeration rate will be improved by increasing the
quantity n^r^2jl/2tres . Experimental evidence exists which confirms
most of the parametric dependencies predicted by the above-mentioned
equations. The results of laboratory and field studies of sonic agglomera-
tion are discussed in the next section.
21
-------
LABORATORY, PILOT-SCALE AND FULL SCALE STUDIES
OF SONIC AGGLOMERATION
Both laboratory and field studies of sonic agglomeration have shown that
aerosols can be agglomerated in acoustic fields under appropriate condi-
tions . An overview of the results of work in both areas is presented
next.
LABORATORY OR SMALL-SCALE STUDIES OF SONIC AGGLOMERATION
Some experimental evidence exists which directly supports the microscopic
coagulation mechanisms which have been discussed. For instance,
^Schlichtingi' provided pictures confirming the parakinetic interaction
between aerosol particles. Gorbachev and Sevemii > ' studied the ele-
mentary interaction between two water droplets suspended on glass threads
and proved that an attractional interaction mechanism was occurring between
the particles. Most experimental studies, however, cannot provide direct
evidence of microscopic mechanisms, but rather do demonstrate parametric
dependencies.
Nearly all experimental data available show excellent agreement with Eq.
(22) (i.e., n = nQe'^a*-). One such verification was made by Brandt^' in
experiments on the acoustic coagulation of a paraffin oil fog. Some of
his results are shown in Figure 7. The radii of fog droplets used by
Brandt were in the range 0.2 to 1.9 urn and the concentration was 15 to 20
g/nP. The curves given in Figure 7 yield the following values for the
coagulation coefficient.
J = 0.0067 w/cm2, Kfl = 0.43
J = 0.06 w/cm2, Ka = 0.92
J = 0.11 w/cm2, Ka = 1.28
22
-------
l.Or-
n0 0.1
0
J= 0.0066 W/cm2
J=0.06 W/cm2
J=0.11 W/cm2
I
0
2
t, sec
J
4
Figure 7. Acoustic coagulation rate of a paraffin fog as a func-
tion of sound intesity (f = 10 kc/sec).
23
-------
Experiments by Brandt and Freund**' and Neumann and Nortoni' show that the
coagulation coefficient is a function of the square root of sound in-
tensity as predicted by Eq. (27) (see Figures 8 and 9).* It has also
been noted, that sound absorption in a medium increases as the intensity
increases so that, from an economic standpoint, it is not recommended
that very high sound intensities be used in industrial aerosol coagula-
tion. Most often, the sound intensity used is of the order of 0.1 w/crn^
to 1 w/cnr.
It has been shown experimentally that the frequency of the sound used
has a very substantial effect on the acoustic coagulation process. In
fact, numerous experiments have shown that each aerosol has an optimum
frequency at which acoustic coagulation proceeds in the most efficient
manner. The optimum frequency is not a function of either sound in-
tensity or the treatment time (see Figure 10), but is a function solely
of the properties of the aerosol as predicted by Eq. (12).
Experiments also have shown that the efficiency of the sonic agglomera-
tion process decreases rapidly at mass concentrations below 2 to 5 g/m .
On the other hand, very high concentrations (> 100 g/m^) are also un-
desirable since there is an exceedingly large decrease in sound intensity
with distance from the sound source at such high concentrations, and this
reduces the coagulation rate in the aerosol.
Experiments have also shown that an increase in the viscosity, tempera-
ture, or pressure of the gas all have a negative effect on the coagula-
tion rate in the aerosol. It should be noted, however, that a high tem-
perature does not prevent the process from occurring; it merely slows it
down. This was supported by Stokes—' who noted that the cooling and
dehumidification (which reduces the viscosity) of the aerosol aided ag-
glomeration. Boucher!!/ made a claim that increased viscosity enhances
agglomeration, but this is not supported by experimental evidence.
The addition of atomized water to a low concentration aerosol is generally
found to increase the efficiency of agglomeration. Figure 11 presents
the results of addition of water to a gas black aerosol. The addition of
water apparently increases the efficiency of agglomeration and subsequent
collection of particles in the cyclone collector.^/ There is some ques-
tion as to whether the water drops promote agglomeration, act more as a
scrubbing media, or both. If the water drops promote agglomeration, they
would do so by increasing the number of coagulation centers.
* In Figures 8 and 9, the index of agglomeration, I , is the ratio of
the final to the initial mean particle diameter.
24
-------
CN
c
o
_o
O)
X
0)
I
_L
40 80 120
160 200
i m-sec
240 280
Figure 8.
Degree of agglomeration of tobacco coke particles as a function
of the product Agtres (f = 10 kc/sec).
D
O
8
CN
-o
"o
i
J>
en
o
x
A
O
3.75 sec
3.13 sec
2.25 sec
1.67 sec
1.25 sec
0.88 sec
Figure 9.
0123
j1/2(J-W/cm2)
Degree of agglomeration of titanium tetrachloride particles as
a function of the quantity J1'2 (f = 1.6 kc/sec).
25
-------
=S 155
0)
>
Q>
3
o
145
135
0 1000 3000 5000
Frequency, cps
Figure 10. Sound level required for 50% agglomeration of a
water fog for different times and frequencies.
100
c
4)
U
X
U
c
0)
c
^o
•4-
o
o
U
80
60
40
O Dry Form
A Water Fog Added
0 0.2 0.4 0.6 0.8 1.0 1.2
Treatment Time, sec
Figure 11. Acoustic coagulation efficiency of gas black (PC = 0.35
attn, f = 2.Ike/sec) as affected by adding a water fog'.
26
-------
Reference 17 discusses the results of a series of tests to determine the
performance of a sonic agglomerator-cyclone collector system currently
being developed by the Braxton Corporation. Since the agglomeration rate
should be enhanced by both a higher concentration and wider size range of
the inlet particulate, the Braxton system is designed to allow the admis-
sion of steam and/or water droplets. The collector efficiencies of the
Braxton device using suspended cupola dust and different combinations of
sound, steam and water ranged from 81 to 977». These efficiencies are
based on using a cyclone after the sonic agglomerator equal in collec-
tion efficiency to the cyclone used for sampling during these tests.
The correlation coefficients for the small particle reduction for all
test runs with sound, water and steam individually and neglecting other
operating parameters were 0.68, 0.65, and 0.06, respectively. The
coefficients for sound and water were statistically significant while
that for steam was not. A multiple regression analysis of those tests
using cupola dust showed that the use of steam is not significant, but
that the use of water sprays and the sonic generator are highly signifi-
cant in reducing the mass of fine particles. The regression analysis
also showed that the sonic generator has less influence on fine particle
reduction than does water sprays when using resuspended cupola dust.
A possible scrubbing action by the water droplets is indicated.
FIELD STUDIES OF SONIC AGGLOMERATION
Sonic agglomeration chambers in conjunction with cyclone collectors or
other control devices have been used in industrial gas cleaning applica-
tions on a random basis for many years, especially in Europe and Russia.
Table I presents a summary of some applications. The applications in
Table 1 date to the 1950 to 1960 time period.
The test data from industrial systems show essentially the same parametric
dependencies as noted in laboratory work. The exponential nature of the
process (i.e., n = noe'^a*1) is confirmed as well as the existence of optimum
frequencies. Table 2 presents a compilation of data on optimum frequencies
required for acoustic agglomeration of different aerosols. The data in
Table 2 show that the optimum frequency varies inversely with particle
size in agreement with the theoretical prediction given in Eq. (12).
Performance variations with changes in inlet grain loadings, particle
size, and water addition are shown in Table 1 and Figures 12 and 13.
Analyses of industrial applications indicate that the sound generators
usually operate in the 1 to 4 kc/sec range and sound intensities in ex-
cess of 130 db are necessary (130 db = 10~3 w/cra^).* Contact times of
about 4 sec have generally been used in industrial systems.
* Intensities of 150 to 160 db are common.
27
-------
Table 1. SUMMARY OF INDUSTRIAL TESTS WITH SONIC AGGLOMERATORS
,
H
_4
O
W
O
Ui
a
<
Gas furnace black
Gas furnace black
Aggregated gas
black
Atomized carbon
black
Hard coal black
Sulfuric acid fog
Natural sulfuric
acid fog
Dilute sulfuric
acid fog
Aerosol and Gas Stream Properties
to
3 -a
•H C 0>
•O O *J '*•
n T4 O Q M
c£ s^ -LJ U 5) X
g WAJx-s 3 X-* U ^~
0)3. ^fjCCO AJ O> &0 0> » 6 •
JJ U -J \^ Q. *J DO"
lj C B v-' ^ ~-
RJ O X 0) O
o. o a H >
0.03-0.07 1.2-12.6 40 1,700-2,000
0.03-0.07 1.2-2.1 40 1,700-2,000
0.5-15 0.5-2.5 - 600
0.1-0.2 26 82 45
0.5-1.0 0.5-2.4 80-90 90-100
0.5-5.0 5-40 180 1,700
0.25-2.5 1 50 40,000
2.5-50, pre- 0.5-1.2 20 1,800
dominant
m
§
•r-l
W /-»
c a
0)
1 £
•a x
c
a u
o
o» ^
a.
H
Experimental direct
flow 1.1 dia. x 6.6
The same, with water
addition
Experimental reverse
flow with water
0.5 dia.
Experimental, rising
stream 0.29 dia. x
1.9
Experimental, reverse,
flow, 0.2 dia. x 2.5
The same, 0.6 dia. x
6
Industrial, composite
flow, 2.4 dia. x
10.5 (2 sets)
Experimental, reverse
flow, 0.64 dia x
AEelomera t ion Char
c
o
•H
to
tkl
O
O)
a.
H
Dynamic, radial
The same
Dynamic, axial
Static with
pump- of f
Dynamic , axia 1
The same
Dynamic, radial
Dynamic, axial
nber
u
01 r-
co U
u u to
Jd t* '-^
N_, £
>-) O O
U4 Cfl U
>s X *-• U-l 4J
U 4J J H
4 0.5-1.0 4.5
2-4 0.5-1.0 1.2
3 0.1 10
4.6 1.0 7
3.6 0.10-0.14 3-4
2.15 0.1 3
2.25 0.1 4
1-2 0.1 7
Collector System
P
T3 ^
a
> c
O P1^
CJ O
« v.
C 3
U O
O U "~
a. b u
H &* 3
Two cyclones 40
1.3 in dia.
(in series)
The same 8-32
One or four 58-72
cyclones (in
parallel)
Two cyclones (30)—
and a glass
cloth filter
(in series)
Cyclone 0.15m 68-74
dia. (81)-
Multicyclones 84
(in parallel)
Two cyclones
(in parallel)
Four cyclones 69-72
(in parallel)
u ?•£
O *"
E •*
0 -0
sc c
ffl 3
C i-
o
u ~
Ol -^
— 3
83-90
99
95
99.98
(97>5'
87 (97)^
99.6-99.9
90
78-82
7.5
-------
Table 1. (Concluded)
41
£
O
ra
O
41
Zinc oxide sub-
limate from
roasting zinc
ore
Zinc oxide sub-
limate from
ing
Zinc oxide sub-
limate from
brass melt-
ing
Coke gas (tar)
Cracking gas (con-
densate)
Cracking gas (con-
densate)
Open-hearth furnace
smoke
Carbide furnace
smoke
Carbide furnace
smoke
Aerosol and Gas Strem Properties
CO
3 -O
3 § 2 ~
£ ~ i 2 Si
6 OJAJ^-t 3^-* M^.
41 a. n .en j-> o Hfi
-* s_, u as e a o g
o c -H ~- b 41
•••4M 41 41 00 41 » 6 -
" t> IX ^ CX *J 3 O"
™ S ^ B ^ -1 v'
CO O X 41 O
Ok UAH >
0.5-5.0, 1-2 40-100 1,600
predomi-
nant 2.5
0.5-4.0 0.5-20 50-350 1,300-2,160
0.4-0.6 10 400 7,000
0.5-5.0, 30-70 40-60 1,300-2,100
predomi-
nant 2.5
0.5-5.0 5-70 35 1,200
0.5-5.0, 6-15 40 12,000
predomi-
nant 3.0-
3.5
2.5 (557.) 2 150 5,000
0. 5- 15, pre- 0.25-2.8 120 500
nominant
0.5
The same 0.25-2.8 120 500
Agglomeration Chamber
CO U
c o
O to
^ ^
m ^* o
e s -*
41 C -^
e - a
f* O u tw
ox -^
•ox- o
B >M e
CO U O 41
O 3
CX O. 41
E- H" £
Experimental, reverse Dynamic, axial 3-3.5
flow, 0.75 dia. x
10
The same, 1.0 dia. x Dynamic, radial 3-9
9
The same, 0.7 dia. x Dynamic, axial 0.7
10
The same, 0.5-0.64 The. same 4
dia. x 9
The same, 0.5 dia. x The same 4
9
Industrial, reverse The same 3.5
flow, 1.6 dia. x
11 (2 sets)
Experlmetnal, reverse Dynamic 2.2
flow with water
addition
Experimental, reverse Static 7-10
flow
The same, with water Static 10 5
addition (5 g/m3)
Collector System
/-^ ^ ^^
u c)
0 > C
0 M 2 *~
"c ""' o; *o
•-> o o °* S
4. I
>% x-s oj «J eo «
*J -^
9 .
B •»*
£ t=^
Sel
CO 3
C V3
U JS
b W
S. 3
94-98
90-95
99.8
99-99.8
*
97.5-99.3
95
90.7
94
86
a_/ Result without cloth filter, in parentheses.
b/ Result from water addition, in parentheses.
-------
Table 2. OPTIMUM FREQUENCIES FOR SONIC AGGLOMERATION J:/ INDUSTRIAL TESTS
Aerosol
Furnace gas black
Aggregated carbon black
Zinc oxide sublimate
Coke gas fog
Cracking gas fog
Dilute sulfuric acid
Particle
Radius^/
r (lira)
0.03-0.07
0.5-15.0
0.5-5.0
2.5
0.5-5.0
2.5
0.5-5.0
3.0
2.5-50
Optimum
Frequency
fopt (kc/sec)
3.5-4.0
3.0
3.0-3.5
3.5-4.0
3.5-4.0
1.0-2.0
a/ The numerator gives the maximum radii, while the denominator gives
the predominant radius.
30
-------
to
Q
z
o
u
ULJ
to
U
<
O
u
20 40 60 80 100
COLLECTION EFFICIENCY %
Figure 12. Collection efficiency of 93.5-95.0 scale furnace black as
a function of time and exposure (cyclone collector) ..IP./
LLJ
O
u
O
u
100
u 90
Z
LLJ
U
80
70
60
i i
j
'
i r i i
Grain Loadings as Noted
i i
i i i i
0 2 4 6 8 10 12
APPROX. INTENSITY-EXPOSURE TIME PRODUCT
(WATTS/CM2 -SEC),
Figure 13. Performance of sonic carbon black collection system (three
cyclones in series).—'
31
-------
SUMMARY OF LABORATORY, PILOT-SCALE AND FULL-SCALE DATA
Testing of sonic agglomerators at the laboratory, pilot-scale and full-
scale plant level have resulted in the following observations:
1. Frequency has a strong effect on sonic agglomeration and an optimum
frequency exists for various particle size?..
2. Optimum frequency increases with decreasing particle size.
3. Sound intensity has a significant effect on sonic agglomeration and
the particle growth rate varies with the square root of the sound inten-
sity.
4. Sound intensities of 150 to 160 db have been used in previous industrial
applications.
5. Sonic agglomeration rate is strongly affected by the concentration by
weight of particles and efficiency of agglomeration decreases rapidly at
concentrations below 2 to 5 g/nr (1 to 2 grains/ft^).
6. Polydisperse aerosols are easier to agglomerate than relatively mono-
disperse aerosols (i.e., if particles are all of nearly same size, acoustic
agglomeration is not effective).
7. Injection of atomized water into low concentration aerosols seems to
improve sonic agglomeration. Improvement from water injection is due pri-
marily to the fact that additional coagulation centers are available.
8. Reduction of turbulence in the agglomeration chamber generally re-
sults in a decrease in the agglomeration rate.
9. Physical properties of aerosol particles have comparatively little
effect on acoustic agglomeration.
10. Viscosity, temperature, and pressure of the gas are generally not
important.
11. Agglomerates formed in sonic fields often do not have a high degree
of stability and they may be broken up in the collection device, especially
in cyclones.
32
-------
COMPARATIVE ANALYSIS OF SONIC AGGLOMERATORS
AND CONVENTIONAL CONTROL SYSTEMS
The utilization of sonic agglomeration for industrial gas cleaning re-
quires the coupling of an agglomeration chamber to a particulate collec-
tion system. Cyclones, electrostatic precipitators, and fabric filters
have been used or proposed as the collection system with cyclone systems
being the prevalent system in the past. For a sonic agglomeration system
to be competitive with a conventional control device, it must offer im-
proved collection efficiency at reasonable cost or reduced cost for the
same degree of efficiency. .Analysis of the performance and energy require-
ments of sonic systems is clearly in order, and since our interest is
mainly in the collection of fine particles, our comparative analysis will
be confined to that size range.
Increasing the size of submicron particles by sonic agglomeration to a
size where they can be readily collected by inertial methods requires a
relatively long treatment time, a considerable expenditure of energy or
both. If the acoustic coagulation coefficient is known, the time required
to achieve a given increase in particle size can be estimated from Eq.
(28) which is a rearrangement; of Eq. (22).
1/3
r
T,
(28)
1/2
Recalling that Ka is proportional to J ' , Eq. (28) can be written as
,1/3
r
r.
(29)
If it is assumed that the density of the agglomerate is the same as the
original particulate, Eq. (29) becomes
33
-------
(30)
Figure 14 presents a graphical depiction of Eq. (30) while Figure 15
presents the experimental data of Neumann and Nortoni' in a slightly
different form than Figure 9.
For a sonic agglomerator to be of significant help in the collection of
fine particles,* the ratio of the final to initial mean particle radius
or diameter must be in the range of 5 to 20. The lower end of the range
would correspond to the use of a high efficiency system as the collector
(i.e., ESP, fabric filter, venturi) while the upper range would reflect
the use of a low efficiency system as the collector (i.e., cyclone,
spray scrubber). To achieve a growth ratio of 5 to 20, Eq. (30) predicts
that the product of t0 J^" would range from 5 to 9. Assuming the ex-
perimental data in Figure 15 can be used for illustrative purposes, growth
ratios of 5 to 20 would require that the product of to J vary from
6 to 20 in real systems.
At first glance one might assume that either t0 or J could be varied to
achieve the desired product of to and J . However, sound absorption in
the medium increases rapidly as the sound intensity increases and it is
not economical to use very high sound intensities. As noted previously
sound intensities of 150 to 160 db (0.1 to 1.0 w/crn^) were common in
previous applications. With the preceding values of J , the exposure
time would have to be in the range of 6 to 200 sec. Exposure times up to
10 sec have been used, but clearly the 200 sec is not feasible. It is
clear that sound intensities of 1.0 w/cnr or greater will be needed
if fine particles are to be agglomerated in residence times suitable for
industrial applications.
The preceding discussion regarding Eq. (30) and Figure 14 did not consider
the efficiencies of the sound source system and other equipment. When one
makes allowances for system inefficiencies, the energy or exposure time
requirements are more severe as indicated by actual data in Figure 15.
Equation 31, developed by Mednikov relates the specific energy required
to agglomerate an aerosol to the efficiencies of various components of a
sonic agglomerator.
* For the purposes of this discussion, fine particles will be defined
as those in the size range of 0.1 to 3 urn in diameter.
34
-------
4
0)
4/1
CN
CM
10 i-
8
0
10
15
20
Figure 14. Predicted sound intensities as a function of agglomeration
index.
CN
E
u
>
0)
if
X
•-
c
Q>
c
X
J
p
18
16
14
12
10
8
6
4
-
-
-
-
/
/
//
- //
/ / /
- V'
///
'7
/
/
/
• Limits of scatter
of data
Grain loading held constant
at approx. 2.5 grains/ft^
Frequemcy = 1.5 kc/s
T5 Cl> Aerosol
i I
2 4 6 8 10 12 14
Index of Agglomeration, d2/d1
Figure 15. Correlation of data on agglomeration and sound intensity.-/
35
-------
J to
(0.36) escceuf H0
E =
o
where E = specific energy, kw-hr/1,000 m
J = sound intensity, w/cnr
to = exposure time, sec
H = height of agglomeration chamber, m
eg = acoustic efficiency of sound source
ec = overall efficiency of compressor during sound source
euf = utilization factor of agglomeration chamber (i.e., factor
to account for losses in chamber)
The acoustic efficiency of the sound source (es) and the compressor effici-
ency (ec) can be estimated for various types of equipment, but euf must
be determined by experiment for a given chamber.
Equation 31 can be used to estimate specific energy requirements if assump-
tions are made regarding the various efficiency terms and the dimensions
of the acoustic chamber. Example calculations are presented in Table 3
and are based on the following assumptions :
a. J = 150, 160, 165, and 170 db
b. gas flow 1,000 m3/min (35,300 cfm)
c. d2/d1 = 2, 5, 10, 15, 20
d. ec = 0.65, es = 0.5, euf = 0.7*
e. cylindrical chamber with diameters of 2 and 4 m.
Several observations on the data in Table 3 are in order. First, the
specific energy requirements are dependent only upon sound intensity and
chamber height (actually chamber volume). Second, at the low level of
sound intensity (i.e., 150 db) very long chambers are required to achieve
* Data on these efficiences are not generally available. Mednikov cites a
value of cc = 0.65, and Ref. 15 mentions that dynamic sirens of recent
design have efficiences (i.e., es) of about 50%. The value of euf = 0.7
is an arbitrary assumption.
36
-------
Table 3. ESTIMATED SPECIFIC ENERGY REQUIREMENTS FOR EXAMPLE SONIC AGGLOMERATOR
J
db w/cn^
150 0.1
16C -1.0
165 3.2
170 10
d2/d!
2
5
10
15
20
- - 2
5
10
15
20
2
5
10
15
20
2
5
10
15
20
J t0
2
5
7
8
9
2
5
7
8
9
2
5
7
8
9
2
5
7
8
9
to
(sec)
6.6
15
21.6
25.3
28.1
2
5
7
8
9
1.2
2.7
3.9
4.5
5.0
0.7
1.5
2.2
2.5
2.8
D
(ra)
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
2
4
H
(m)
35
9
80
20
115
29
134
34
149
37
. 11
2.8
25.5
6.4
36.6
9.2
43.9
10.7
47.8
12
6.4
1.6
14.3
3.6
20.7
5.2
24
6
26
6.6
35
0.9
8
2
11.7
2.9
13.3
3.3
14.9
3.8
(0.36) esec€uf H
2.9
0.74
6.6
1.7
9.4
2.4
10.4
2.8
12.3
3.0
0.90
0.23
2.1
0.52
3.0
0.75
3.6
0.88
3.9
0.98
0.53
0.13
1.2
0.3
1.7
0.4
2.0
0.5
2.1
0.54
0.28
0.07
0.66
0.16
1.0
0.24
1.1
0.27
1.2
0.3
J
E (0.36) s
kw-hr/1,000 m3
0.23
0.89
0.23
0.89
0.23
0.89
0.23
0.89
0.23
0.89
2.3
9.1
2.3
9.1
2.3
9.1
2.3
9.1
2.3
9.1
7.2
29.5
7.2
29.5
7.2
29.5
7.2
29.5.
7.2
29.5
25
94.6
25
94.6
25
94.6
25
94.6
25
94.6
to
£ e c H
hp/1,000 cfm
0.5
2.1
0.5
2.1
0.5
2.1
0.5
2.1
0.5
2.1
5.2
21
5.2
21
5.2
21
5.2
21
5.2
21
16.4
67.2
16:4
67.2
16.4
67.2
16.4
67.2
16.4
67.2
57
215
57
215
57
215
57
215
57
215
-------
necessary particle growth. In fact, at all ratios of d£/di exceeding
2, the length of the chamber is more than 60 ft. Third, more acceptable
chamber lengths can be achieved at higher sound intensities, but the specific
energy requirements increase rapidly with sound intensities. With regard
to the last point, the data in Table 3 suggest that sound intensities of
160 db or greater will be necessary to grow fine particles to an acceptable
size in realistic equipment sizes. At the high growth ratios (i.e.,
d2/d^ > 5), a 4-m diameter chamber would be necessary to maintain a reason-
able chamber length. Specific energy values' of approximately 5 hp/1,000
cfm are required for moderate growth ratios (d£/d^ < 5) in a 2-m diameter
column of reasonable length, while energy values in excess of 20 hp/1,000
cfm are needed for high growth ratios (d2/d^ > 5) in a 4-m diameter chamber
of reasonable length.
Apart from increasing the cost of equipment, very long chambers are unde-
sirable from the standpoint of increasing acoustic losses in the chamber.
The longer the sound chamber the greater the attenuation of sound in the
chamber. This means that the acoustic coagulation coefficient decreases
as the chamber is made longer. As noted previously sound absorption by the
gas phase also increases as sound intensity increases. The key conclusion
emerging from the example calculations and the influence of chamber length
and sound intensity on efficiency is that chamber diameters of 4 m or larger
and sound intensities of 160 db or greater are likely to be needed to growth
fine particles to the requisite size levels for relatively easy collection.
Quite large energy consumptions (i.e., of the order of 20 hp/1,000 cfm)
appear to be necessary as a direct result.
By way of comparison with the estimated values of energy consumption,
Mednikov reports specific energy consumptions on actual industrial systems
of 0.65 to 20 kw-hr/1,000 m3 of gas (1.5 to 46 hp/1,000 cfm). However, it
must be noted that the reported values for industrial applications are in
some cases for aerosols that did not contain a high percentage of fine
particles. Mednikov states that minimum total energy requirements of 4 to
10 hp/1,000 cfm should be achievable with optimum design..!/ The minimum
values quoted by Mednikov are for a "typical" industrial aerosol and pre-
sumably represent the energy needed to grow the particles to a sufficient
size to be removed by inertial devices. For an aerosol stream containing
a high percentage of fine particles, it seems reasonable to assume that
the upper figure of 10 hp/1,000 cfm represents the likely minimum energy
requirement even if an ESP or fabric filter is used as the collector
rather than an inertial device.
Measures which might be taken to improve the efficiency of the sonic
agglomerator and decrease the energy usage include:
38
-------
1. Addition of water drops to increase number of agglomeration centers;
2. Treating aerosol with a number of frequencies; and
3. Optimizing chamber designs.
However, even if the above steps are taken1 and are successful, energy
consumption is still likely to be moderately high. For example, specific
energies calculated from Eq. (31) assuming that es = cc = euf = 1.0 are
1.2 and 4.7 hp/1,000 cfm for chamber diameters of 2 and 4 m and a sound
intensity of 160 db. At 165 db the corresponding numbers are 4 and 15 hp/
1,000 cfm.
Since the extent to which fine particles are removed from a gas stream
depends both on the amount of growth in the agglomerator and the efficiency
with which particles are collected by the control device coupled to the
agglomerator, the total energy required is the sum of that for the ag-
glomerator and the control ssystem. Typical energy usages for conventional
control systems are:
1. Cyclones--0.5 to 5 hp/1,,000 cfm
2. Fabric filters--2 to 3 hp/1,000 cfm
3. Electrostatic precipital:or--l to 1.5 hp/1,000 cfm
4. Venturi scrubber--10 to 20 hp/1,000 cfm
Even under ideal conditions, a sonic agglomerator would add an energy
burden equivalent to that oi: the collection device that must be coupled
to it. In order to reduce energy consumption in the agglomeration chamber
(i.e., reduce the requisite growth ratio d2/dj^), high efficiency col-
lectors such as an ESP, fabric filter, or venturi scrubber should be used.
It seems likely that the hig;h efficiency control devices alone might do
nearly as good of a cleaning; job as would a combination sonic agglomerator-
high efficiency collector. The utility of a sonic agglomerator as an aid
to controlling fine particulates seems doubtful at best.
39
-------
CONCLUSIONS
The application of sonic agglomeration to industrial gas cleaning problems
involving fine particles does not appear promising. The energy require-
ments of current sonic agglomerators coupled to an inertial separator
are very high, of the order of 10 to 20 hp/1,000 cfm, if a gas stream
containing predominantly fine particulates is to be controlled. Replace-
ment of an inertial collector by a high efficiency ESP, fabric filter,
or venturi scrubber might reduce the energy requirements to 5 to 10 hp/
1,000 cfm. However, if it is necessary to use a high-efficiency col-
lector, the utility of a sonic agglomerator is suspect unless considera-
tions other than energy consumption are important (i.e., a need exists
to control a highly toxic stream composed of fine particles).
40
-------
REFERENCES
1. Mednikov, E. P., "Acou:;tic Coagulation and Precipitation of Aerosols,"
Consultants Bureau, New York (1965).
2. St. Clair, H. W., "Theory and Basic Principles of the Sonic Smoke
Flocculator," Interdepartmental Committee on Air Pollution,
Washington, B.C., Air Pollution Proc., pp. 382-387, U.S. Technical
Conference, Washington, D.C. (1950).
3. Schlichting, H., "Berechnung Ebener periodischer Grenzschichtstrominger,"
Phys. Z., 33_(8):327-335 (1932).
4. Gorbachev, S. V., and A. B. Severnyi, "The Problem of the Effect of
Sound Waves on Fog Drops," Zn. Fiz. Khim.. £(4):536-545 (1936).
5. Gorbachev, S. W., and A. B. Severnyi, "Zur Frage der Bevegung ernes
schweren Tropfens in akustischen Felde," Kollord - Z., 73.(2):145-
154 (1935).
6. Brandt, 0., "Uber das Verhalten von Schwebstoffen in schwingenden
Gasen bei Schall-und Ultraschall-frequenzen," Kolloid-Z.. 7^(3):
272-278 (1936).
7. Inoue, I., "Sonic Agglomeation Chamber," Kagaku Kogaku, 18.(4): 180-186
(1954).
8. Brandt, 0., and H. Freund, "Uber die Aggregation von Aerosolenmittele
Schallawellen," Z. Phys., 94^(5-6): 348-355 (1935).
9. Neumann, E. P., and J. L. Norton, "Application of Sonic Energy to
Commercial Aerosol Collection Problems," Chem. Eng. Progr., Symp.
Sec. lt 47J1) :4-10 (1951).
10. Soderberg, C. R., Jr., "Industrial Applications of Sonic Energy," Iron
Steel Engr.. 29(2):87-94 (1952).
41
-------
11. Nord, M., "Sonic Precipitation of Smoke, Fumes and Dust Particles,"
Chem. Eng., pp. 116-119, October 1950.
12. Stokes, C. A., "Sonic Agglomeration of Carbon Black Aerosols," Chem.
Engr. Prog.. 46(8):423-432 (1950).
13. Boucher, R. M. G., "Ultrasonics in Processing," Chem. Eng., pp. 83-
100, October 1961.
14. Brandt, 0., et al., Z Kolloid, 77, p. 103 (1936). .
15. Strauss, W., Industrial Gas Cleaning, Pergamon Press, Oxford,
pp. 215-232 (1966).
16. Stokes, C. A., and J. E. Vivian, "Application of Sonic Energy in
the Process Industries, Chem. Engr. Prog. Symp. Series 1, 47^(1):
11-21 (1951).
17. Dennis, R., et al., "Braxton Sonic Agglomerator Evaluation," EPA
Report EPA-650/2-74-036, May 1974.
42
-------
APPENDIX A.
SUPPLEMENTARY READING LIST
THEORY
Pshenai-Severin, S. V., "The Coagulation of Aerosol Particles in an
Acoustic Field Under the Influence of Oseen's Hydrodynamic Forces,"
Dokl. Akad. Nauk SSSR. JL2J>(4) : 775-778, 1 April 1959.
Gudemchuk, V. A., B. F. Podoshevnikov, and B. D. Tartakovskii, "On the
Role of Turbulence in the Phenomenon of Acoustic Coagulation of Aerosols,"
Akust, Zh.. .5(2):246 (1959).
Richardson, E. G., "Behavior of Aerosols in Acoustic and Turbulent Fields,"
Acustica, 2/4): 141-147 (19'52).
Hirata, M., "The Sonic Agglomeration Apparatus. Sound Generator and1
Agglomeration Tower," Chen.. Eng. (Japan), JjJ., pp. 180-186 (1954).
Fryar, R. M., "An Investigation of Aerosol Coagulation by Means of Air
Jet Generated Ultrasonic Vibrations," Purdue University, Lafayette,
Indiana, Thesis (Ph.D.), Ann Arbor, Michigan, University Microfilsm,
Inc., 183 pages (1950).
Buravov, L. I., and 0. K. Eknadiosyants, "On the Behavior of Aerosol
Particles in an Acoustic Field," Akust. Zh. (USSR). _7(4):492-493 (1961).
St. Glair, H., "Agglomeration of Smoke, Fog, or Dust Particles by Sonic
Waves," Bureau of Mines, College Park, Maryland, Ind. Eng. Chem., 4^,
pp. 2434-2438 (1949).
Shirokova, N. L., "Coagulation of Aerosols," Fiz. Tekh. Moshch. Ul'trazvuka.
3_, pp. 641-480 (1970).
Mednikov, Y. P., "Theory of Acoustic Coagulation of Aerosols," Doklady
Adad. Nauk SSR. I§3_(2) :382-385 (1968).
43
-------
Timoshenko, V. I., "Aggregation of Aerosol Particles in a Sound Field
Under the Conditions of Stokes Law Flow," Akust. Zh. (USSR). 11(2):222-
225, April-June 1965.
Andrzejewski, R., "Fundamentals of Coagulation Theory," Air Conserv., 4(3):
34-48 (1970).
Shirokova, N. L., and 0. K. Eknadiosyants, "Interaction of Aerosol Particles
in an Acoustic Field," Akust. Zh (USSR). LLp) :409-411 (1965).
PATENTS
Stokes, Charles A., "Recovery of Aerosol Solids," U.S. 2,720,939,
18 October 1955.
Horsley, C. B., and G. C. Seavey, "Apparatus for Agglomerating Aerosols
by Sound Waves," U.S. 2,535,679, 26 December 1950.
Coleman, W. E., and T. F. Reed, "Apparatus for Stripping Entrained Solids
from Blast-Furnace Gases," U.S. 2,949,166, 16 August 1960.
EXPERIMENTAL TESTING
Tsetlin, V. M., V. F. Denisov, and S. A. Tsedilin, "Effect of Various
Factors on Coagulation of Nonferrous Metallurgy Dusts in a Sound
Field," Sb. Tr. Cos. Nauchn.-Issled. Inst. Tsvetn. Met., 19^:595-607 (1962).
Varlamov, M. L., E. L. Krichevskaya, G. A. Manakin, A. A. Ennan, L. M.
Kozakova, and L. S. Zbrozhek, "Acoustical Coalescence of Aerosols Fomred
in Chemical Industries," Primenenie Ul'traakustikik Issled. Veshchestva,
Moscow, Sb., 12j 199-204 (1960).
Danser, H. W., and E. P. Neumann, "Industrial Sonic Agglomeration and
Collection Systems," Ind. Eng. Chem.. 41:2439-2442 (1949).
Neumann, E. P., C. R. Soderberg, Jr., and A. A. Fowle, "Design, Application,
Performance, and Limitations of Sonic Type Flocculators and Collectors,"
Air Pollution, Proc. U.S. Tech. Conf.. Washington, D.C., pp. 388-393,
3-5 May 1950 (L. C. McCabe, editor).
Varlanov. M. L., E. L. Krichevskaya, A. A. Ennan, L. M. Vaspol'skaya,
G. A. Manakin, and R. A. Georgalin, "Acoustic Coagulation of Mist
Containing Fluorine Compounds," J. Appl. Chem. USSR, 41^(12) :2494-2499,
December 1968.
Crawford, A. E., "Sonics in Chemical Processing," Brit. Chem. Eng.. 1,133,
7 July 1956.
-------
Podoshevnikov, B. F., "Changes In the Disperse Composition of a Dioctyl
Phthalate Vapor Due to Coagulation in A Sound Field," Primenenie
Ul'traukustike k Issled. Veshchestva, 15:137-150 (1961).
Haunori, D., "Sonic Dust Collector Making Use of Noise of a Crusher,"
Japan Pat. Sho 3p., 24 April 1971.
Boucher, R. M. G., "Acoustic Purification of Gases," Genie Chim., 7j8,
14-28 (1957).
Podoshevnikov., B. F., "Acoustic Coagulation of a Highly Dispersed Aerosol,"
Vestn. Tekhn. i Ekon. Inform. Nauchn Issled. Inst. Tekim-Ekon. Issled.
Cos. Kim. Soveta Min. SSSR po Khim, 4_:31-37 (1960).
Mikhal'kov, P. V., and T. P. Ivanov, "Increasing the Effectiveness of Gas
Separation," Novosti, Neft.! i Gaz. Tekhn. Gas. Delo, (4):6-10 (1961).
Monchevskii-Rovinskii, B. S.j, "Data on Sonic Coagulation of Aerosols in
Poland," Primenenie Ul'traakustiki k Issled. Veshchestra, No. 17, 75-
82 (1963).
Shkol'nikova, and V. P. Makar'ev, "Acoustic Coagulation and Capture of
Mercury Aerosol from Technological Gas of Mercury Production," Tr. Proekt.
Nauch.-Issled. Inst. "Gipronikel," ,3JJ.:60-71 (1966).
45
-------
TECHNICAL REPORT DATA
• (Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-001
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Evaluation of Sonics for Fine Particle Control
5. REPORT DATE
January 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
R. Hegarty and L. J. Shannon
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Midwest Research Institute
425 Volker Boulevard
Kansas City, Missouri 64110
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-029
11. CONTRACT/GRANT NO.
68-02-1324, Task 24
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND.PERIOD COVERED
Task Final; 8-9/74
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT,
The report gives results of an evaluation of the potential of sonic agglomer-
ation as a means of improving capabilities for controlling fine particle emissions.
Available theoretical and experimental information indicates that sonic agglomerator!
can increase the mean particle size of aerosols; however, the energy requirements
are quite high when the gas stream contains predominantly fine particulates. Even
under ideal conditions, energy consumption would range from 1 to 15 hp/1000 cfm.
These ideal energy levels are not very competitive with other devices capable of
removing fine particulates , especially when a high efficiency control system is
required as a collector in order to minimize energy consumption in the sonic agglo-
merator.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDEDTERMS
c. COSATl Field/Group
Air Pollution
Acoustics
Agglomeration
Fines
Aerosols
Air Pollution Control
Stationary Sources
Sonic Agglomeration
Fine Particulate
13B
20A
07D
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
52
20. SECURITY CLASS (This pagej
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
46
------- |