EPA-600/2-76-055
March 1976
Environmental Protection Technology Series
EVALUATION OF
ELECTROSTATIC AUGMENTATION 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
E PA 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 policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/2-76-055
March 1976
EVALUATION OF ELECTROSTATIC AUGMENTATION
FOR FINE PARTICLE CONTROL
by
D. W. Cooper and M. T. Rei
GCA Corporation
Burlington Road
Bedford, Massachusetts 01730
Contract No. 68-02-1316, Task 7
ROAP No. 21ADL-029
Program Element No. 1AB012
EPA Project Officer: D. 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
-------
ABSTRACT
This is a review of electrostatic augmentation of fine particulate con-
trol devices: the addition of electrical forces to scrubbing and filtra-
tion and the enhancement of electrostatic precipitation. The major elec-
trostatic force equations are presented and evaluated for some reasonable
values of particle and collector charge and geometry. A bibliography on
electrostatic augmentation is given. The following programs in electro-
static augmentation of filters, scrubbers, electrostatic precipitators
are analyzed: an investigation of fiber beds to capture particles elec-
trostatically, research in the area of dust/fabric electrostatic effects,
work done to assess the utility of electric fields applied across filters
or generated within filters, research being undertaken to further the
development of a collector using oppositely charged particles and drop-
lets, investigation and development of a charged droplet scrubber (which
accelerates droplets electrostatically and uses them to transfer charge
to particles for electrostatic precipitation), an analysis of various
polarities and configurations for charged droplet scrubbing of charged
particles, experiments and analysis directed at the use of nuclear radi-
ation to charge particles for electrostatic precipitation, study of
various possible configurations and uses for the "electric curtain," and
the improvement of particle charging by theoretical and experimental re-
search in connection with precharging chamber's. Other research in elec-
trostatic augmentation, especially on filters, is discussed briefly.
Analysis of two other possible systems is presented: an electrostatically
augmented cyclone, and a foam scrubber which uses particle precharging.
A cost/benefit method for setting research priorities is developed which
takes into account the expected applicability of the results, their
iii
-------
expected probability of and time to fruition, and the estimated costs of
implementation. The following areas of research are emphasized for con-
sideration: ionic mobility and mean thermal speed determination; tri-
boelectrification of droplets, fibers, and bed packings; use of electro-
static scattering to diminish aerosol concentration fluctuations in gen-
eration; detailed experimental determination of the efficiency of collec-
tion by charged droplets; the relations between charge and wettability and
particle rebound and adhesion; charge transfer from droplet to particle;
precharging as a means for improving collection efficiency of conventional
collection devices; trade-offs in charged droplet scrubbing; the cleaning
of open-structure electrified filters; wetted versus dry surfaces in
various forms of electrostatic interaction and precipitation; difficult
control problems and their electrical characteristics; cost/benefit anal-
yses of charged droplet scrubbing and of the addition of particle charging
devices to enhance conventional collection through electrostatic scattering.
The appendices discuss the intrinsic power requirements for dust removal,
the conditions under which insulator particles act as though conductive,
some notes on exponential penetration formulas, and a simple method for
calculating electrostatic collection efficiency of several geometries
and several types of electrostatic interaction.
iv
-------
CONTENTS
Section
Abstract
Figures vi
Tables x
Acknowledgments xiii
Sections
I Conclusions 1
II Recommendations 3
III Introduction 5
IV Technical Overview 11
V Electrostatic Augmentation of Fabric or Bed Filtration 57
VI Electrostatic Augmentation of Scrubbers 131
VII Electrostatic Augmentation of Precipitation Devices 197
VIII Other Electrostatic Devices 243
IX Setting Priorities 265
X Some Research Possibilities 271
Appendixes
A Intrinsic Power Requirements for Dust Removal 279
B Insulator Particles Can Behave as Conductors 285
C Notes on Exponential Penetration Formulae 289
D Approximate Calculation of Collection Efficiency for
Central-Force Collector 293
-------
FIGURES
No. Page
1 Some Forces Which Can be Exerted on Aerosol Particles 7
2 Particle Collection in a Channel and Particle Collection
by an Obstacle Within a Flow 12
3 Trajectories of Charged Particles Carried by Viscous Flow
Past an Isolated, Charged, Dielectric Cylinder in a
Polarizing Field Parallel to the Flow Direction l 29
4 Penetration Versus Electrostatic Scattering Parameter 35
5 Theoretically Calculated Migration Velocities for Four
Electrostatic Mechanisms Versus Particle Diameter 38
6 Schematic of Test Apparatus for Study of Removal of
Charged Submicron Particles by Fiber Beds 58
7 Schematic of Experimental Setup for Study of Removal of
Charged Fly Ash by Fiber Beds 66
8 Electrostatic Capture of Particles by Polypropylene
Fiber Bed 68
9 Aerosol Removal by a 6-Inch Polypropylene Bed 70
10 Aerosol Removal by a 3-Inch Polypropylene Bed 70
11 Outlet Loading Versus Free Area. Woven Dacron Nylon Bags,
Fly Ash Filtration at 3 grains/ft3 and 3 fpm 93
12 Performance of "Real" Filter in the Absence and Presence
of External Electric Field 99
13 Penetration Load Curve for Electric Filter 101
14 Diagram of Filter Construction 103
vi
-------
FIGURES (continued)
No. Page
15 Changes in the Electric Field Between a Pair of Insulated
Wires Due to the Deposition of Charged Particles 104
16 Experimental Apparatus for Studying the Effect of an
Electric Field on the Trajectories of Dust Particles 105
17 Deposition Efficiency of a Filter Under Different
Operational Conditions 110
18 Dependence of Deposition Efficiency on Applied Voltage 110
19 Experimental Results With Coal Dust 111
20 Experimental Results With Quartz Dust 111
21 The Increase of Pressure Drop at Constant Airflow in
Relation to the Amount of Dust on the Filter 112
22 Methods of Applying a d.c. Voltage to Electrically Con-
ductive Fibrous Filters 125
23 Analytical Model of a Dielectric Fiber Mat Filter 126
24 Effect of Charged Condition of Metal Grids on Electric
Field Distribution in a Dielectric Fiber Mat Filter 126
25 Schematic Diagram of Electrostatic Droplet Scrubber 132
26 Calculated Particle Collection Efficiencies for a Single
200-u Diameter Droplet With a 100-cm/sec Undisturbed
Fluid Velocity 136
27 Size Distribution of Water Spray Droplets 139
28 Size Distributions of Dioctylphthalate Aerosol Particles
at Electrostatic Droplet Spray Scrubber Inlet and Outlet 140
29 Particle Collection Efficiency of Electrostatic Spray
Droplet Scrubber as Function of Particle Size 140
30 Geometry and Coordinates of the Two Sphere System 144
31 Single Particle Collection Efficiency - Inertial and
ElectrosLatic Effects 145
vii
-------
FIGURES (continued)
No.
32 Collection Efficiency in Potential Flow as Function of iji
for Various KE, Computed by Nielsen (Solid Lines) and
by George (Dashed Lines) 146
33 Theoretical Overall Collection Efficiency for Scrubber
as a Function of Inertial Parameter, \jh; ES = 0.0 150
\,
34 Collection Efficiency Versus ip for ES = 0.1 151
\,
35 Collection Efficiency Versus i>2 for ES = 1.0, 10.0 152
36 Electrostatic Parameter Versus Particle Size for Several
Flow Velocities From 7.63 cm/s to 3770 cm/s 154
37 Charged Droplet Scrubber 161
38 Droplet and Particle Dimensions at Distance of Closest
Approach 171
39 Plot of Function G(a) Related to Particle Drift Time 173
40 Induced Charging of Spherical Particles 176
41 Functional Dependence of Collision Effectiveness Prob-
ability on Characteristic Charge, Q 178
42 Theoretical and Measured Collection of Positively Charged
Aerosol Particles Upon Negatively Charged Drops as a
Function of Drop Charging Voltage 189
43 Theoretical and Measured Particle Collection for Pre-
cipitation of Positively Charged Aerosol Particles by
Positively Charged Drops as a Function of Drop Charging
Voltage 189
44 Schematic Diagram of Gamma Ray Precipitator and Auxiliary
Equipment 199
45 Estimated Collection Efficiencies for Gamma Ray Precipita-
tor and Electrostatic Precipicator 206
46 Estimated Collection Efficiencies for Gamma Ray Precipita-
tor and Electrostatic Precipitator Utilizing Altered
Assumptions 208
viii
-------
FIGURES (continued)
No. Page
47 Electric Curtain 214
48 Electric Curtain Connected so as to Provide a Traveling
Wave Electric Field Moving Toward Bottom of Rods 215
49 Horizontal Rail Structure to Support Liquid Scrubber Drops
to Increase Interaction Time With Gas Flow From Which it is
Necessary to Remove Gases Such as S0_ by Absorption or
Chemical Interaction With Scrubber Drops 216
50 Electric Curtain Schematic With Coordinates 222
51 Field and Diffusional Charging of Small Particles 233
52 Model for Mathematical Treatment of Charging Rate 234
53 Comparison of Theories and Hewitt's Experimental Data for
0.28 Micron Diameter Particles and Medium Electric Field
Intensity, E = 3.6 kV/cm 238
54 Scnematically Drawn Electrostatically-Augmented Cyclone 244
55 Efficiency Versus Particle Diameter for Cyclone With and
Without Electrostatic Augmentation 251
56 Schematic of Possible Electrostatically-Augmented Sieve
Plate Scrubber 258
57 Calculated Penetrations at 0.1, 0.3, 1.0, 3.0, 10 urn
and Linear Interpolations 262
58 Model for Particle Collection by Obstacles 289
59 Geometry for Approximate Calculation of Collector Effi-
ciency for Central Forces 294
ix
-------
TABLES
No. Page
--- *-*
1 Approximate Maximum Charge 21
2 Forces, Velocities for Conducting Sphere Charged and
Precipitated at E = 10 kV/cm 22
3 Forces and Terminal Velocities for Particles Charged at
E = 33 esu, Collected in the Presence of Uncharged
Conducting Spheres 24
4 Forces (Dynes) for Charged Collector With Uncharged
Particle 25
5 Terminal Velocities (cm/s) for Combinations Given in
Table 4 26
6 Description of Collection Regimes for a Single Collector,
A Charged Particle, and an External Field, Based on Work
of Hochrainer et al. and Zebel 28
7 Bibliography of Electrostatic Augmentation 42
8 Information Utilized to Calculate Theoretical Efficiency 61
9 Experimentally Observed and Theoretical Fiber Bed
Efficiencies 61
10 Calculated Theoretical Efficiencies With Several Col-
lection Mechanisms 64
11 Aerosol Deposition in a 6-Inch Polypropylene Bed 71
12 Aerosol Deposition in a 3-Inch Polypropylene Bed 71
13 Aerosol Deposition in a 6-Inch Stainless Steel Bed 71
14a Parameters Associated With the Study of Electrostatic
Capture of Particles by Fiber Beds 76
-------
TABLES (continued)
No.
14b Parameters Associated With the Study of Electrostatic
Capture of Particles by Fiber Beds 77
15 Pressure Drop Dependence on Porosity Factor 90
16 Parameters Associated With the Study of Electrostatic
Effects in Fabric Filtration 95
17 Theoretical and Experimental Efficiency for Coal Dust
and Quartz Dust at 10 cm/s Face Velocity 117
18 Theoretical and Experimental Efficiency for Coal Dust
and Quartz Dust at 80 cm/s Face Velocity 117
19 Parameters Associated With the Study of the Electrostatic
Spray Scrubber 133
20 Efficiencies Calculated for Various Electrostatic and
Inertial Parameters 149
21 Parameters Associated With the Study of the Charged
Droplet Scrubber 159
22 Three Stage CDS Performance Data-United Sterra Talc -
1.8 pm Mean Size 162
23 Distances of Closest Approach, D, for Particles and Drop-
lets Under Assumptions Stated in Text 174
24 Summary of Basic Configurations for Collecting Submicron
Particles 186
25 Experimentally Determined Efficiencies for Three Charged
Droplet Scrubber Configurations 190
26 Parameters Associated With the Study of Systems of
Charged Droplets and Particulate 192
27 Corrected Version of MRl's Table 205
28 Results for Altered Assumptions 207
29 Approximate GRP Effective Migration Velocities
(Experimental) 210
xi
-------
TABLES (continued).
No. Page
30 Experimental Calculated Average Charge for GRP 211
31 Gamma-Ray Precipitator Advantages and Disadvantages 212
32 Parameters Associated With the Study of the Electric
Curtain as a Device for the Control and Removal of
Particulate Materials 220
33 Parameters Associated With the Study of Precharging
Chambers 236
34 Design Parameters for a High Efficiency Cyclone of
0.472 m3/Sec (1000 cfm) 245
35 Calculated Theoretical Migration Velocity and Corresponding
Efficiency for the Hign Efficiency Cyclone With and Without
Electrostatic Augmentation 252
36 Calculated Theoretical Migration Velocity for Inertial
and Electrical Forces and Predicted Efficiency Due to the
Combination of Forces for a Cyclone of Twice the Original
Cyclone Dimensions 255
37 Particle Parameters Used to Estimate Space Charge Deposition
in Bubbles 26°
38 Penetration of Space Charge Scrubber (Assuming 0.5 sec
Residence Time, 3 kV/cm Charging Field) 260
39 Widely-Used Control Devices and Particle Removal Mechanisms 275
40 Some Substances and Their Resistivities 286
xii
-------
ACKNOWLEDGMENTS
The authors appreciate the help of Eugene F. Mallove, who wrote most of
the material on the gamma-ray precipitator, and of Benjamin Kincannon,
xvho contributed substantially to the section on setting priorities.
xiii
-------
SECTION I
CONCLUSIONS
Electrostatic forces can be appreciably stronger than other forces oper-
ating on fine particles within control devices under some feasible con-
ditions o Electrical forces most likely to be useful are: Coulombic
attraction or repulsion between charged particle and charged collector,
attraction between uncharged particle and charged collector, repulsion
among highly charged particles. The relatively high collection efficien-
cies of high porosity plastic fiber beds may be due to contact charging
and charge buildup due to the deposition of charged particles. Super-
imposing an electric field across a filter medium will increase collection
of charged particles; the use of electrets presents problems in cleaning,
and the electrets become neutralized by the adhesion of charged particles
in their vicinity; conductive fibers coated with nonconducting material
would allow the use of electrical attraction without producing a current
drain in neutralizing the particles and without requiring the formation
of filter cake for high efficiencies. (The Japanese researchers have
demonstrated the increase in collection efficiency which can be obtained
with several different types of such electrified filters.) In some cases,
cleaning electrostatic filters may remain a major problem. Scrubbers
using droplets and particles charged oppositely have been shown to pro-
duce higher collection efficiencies than the same configurations without
charging. Electrically accelerating spray droplets does not seem prefer-
able to accelerating them hydraulically or pneumatically, but there may
be advantages in charging particles via charged droplets. Charged drop-
let scrubbers probably would have power and size requirements between
those for conventional scrubbers and those for electrostatic precipitators.
-------
The times associated with particle collection and with droplet dissipa-
tion in a charged droplet scrubber, along with the residence time, can
be useful for the analysis of such systems, as shown by the work at MIT
by Melcher and Sachar; the residence time and the droplet dissipation
time must be much larger than the characteristic collection time for
nearly complete collection to occur. The work done on charging particles
with radioactive materials by a reactor has produced an interesting elec-
trostatic precipitator of doubtful practicality. The "electric curtain"
has a number of possible control configurations, but its performance has
yet to be tested and some of the proposed uses are questionable. Aug-
menting cyclones electrostatically does not seem promising, but pre-
charging particles before they enter foam or packed bed scrubbers does.
A rational methodology has been formulated for setting priorities for
such electrostatic augmentation research.
-------
SECTION II
RECOMMENDATIONS
In general, applications of electrostatics can be expected to increase
collection efficiencies of control devices by adding another collection
force to those already present. Continued investigation of electrostatic
augmentation of control devices and of improved electrostatic precipitator
operating conditions and parameters should lead to improvements in such
particulate control methods.
-------
SECTION III
INTRODUCTION
This is a review and compilation of information available on devices
which utilize electrostatic augmentation for the collection of fine
particles. Electrostatic augmentation is defined here as methods for
increasing the role of electrical forces. The primary questions this
document seeks to answer are:
What has been done?
What is planned?
What has been omitted?
What should be done to bring about practical new control
technologies employing electrostatic augmentation?
What has been done has generally been published as reports or journal
articles, and what is planned has generally been outlined in program
projections. Recommendations for future work involve gauging the gap
between where we are and where we wish to be. We have attempted to
gauge this gap after having defined the current state of technology and
its needs.
SIGNIFICANCE OF FINE PARTICLES
Fine particulates are those smaller than about 3 Mm in diameter. They
are of concern because they persist in the atmosphere without settling
out rapidly and they penetrate man's natural defenses and lodge themselves
in the lung. They are much more harmful than their mass concentrations
-------
would suggest: "Total weight is an inadequate measure of particulate
pollution and its effects. Particles in the 0.1 - 1pm range generally
have a much greater impact on public health, visibility, and cloud nuclea-
o
tion when compared with the same weight of larger particulates." Particles
in the 0.1 to 1 pro range also are the most difficult to collect. Their
impact and the difficulty of controlling them have made fine particles
the focus of intense pollution technology interest in recent years.
ATTRACTIVENESS OF ELECTROSTATICS
It is more than just a play on words to say that electrostatics are an
attractive means for trying to control fine particle emissions. As
Figure 1 shows, electrical forces are much stronger than gravitational
forces, thermal forces, and adhesion forces in the 0.1 to 1 pm range.
The electrical forces for particles in this size range are often greater
than those which are readily obtainable through inertial methods, as well,
such as cyclones or scrubbers. (Inertial acceleration would be on the
order of the square of the gas velocity divided by the collector dimension,
_2
only ten times greater than gravity for a 10 cm/s velocity and a 10 cm
collecting body.) As will be seen, there are various types of electrical
interaction which can produce attractive or repulsive forces on particulate
material. Moreover, these different types of forces can be used in very
different kinds of control devices.
-------
10
10
I07
u
K
o
u.
tr
o
o
U
O
or
o
u.
u.
O
cr
10'
io:
to:
I01
10'
10"
10"*
10
io
to'
-Electric Force-E = IlkV/cm ond
moximum surface CMorge
Sound force- I40d8
dhesion
II i i
I0"2 10"' 10° 10' IOJ IO5 10*
PARTICLE SIZE (u )
Figure 1. Some forces which can be exerted on aerosol particles
The following projects have served as the principal objects of study for
this review:
ELECTROSTATIC AUGMENTATION OF FABRIC OR BED FILTRATION
Electrostatic fiber beds
Electrostatic effects in fabric filtration
Ambient fields across filter media
ELECTROSTATIC AUGMENTATION OF SCRUBBERS
Electrically accelerated droplets
Oppositely charged droplets and particles
Systems of charged drops and particles
-------
ELECTROSTATIC AUGMENTATION OF PRECIPITATION DEVICES
Gamma-ray prccipitator
Electric curtain and AC fields
Precharging chambers
Our report summarizes and analyzes these studies as well as others noted
below. These summaries are intended to indicate where we are and to
suggest where we should be headed with respect to the electrostatic aug-
mentation of fine particle control devices. As indicated by the table
of contents, the sequence in which we have presented this material is:
first, a technical overview of the subject of electrostatic forces and
aerosols; next, summary and evaluation of studies related to filtration,
scrubbing, and electrostatic precipitator collection; a short discussion
of the methods by which priorities for this kind of work might be set;
finally, recommendations for further work.
-------
REFERENCES
1. Lippman, M. Respirable Dust Sampling. Am Ind Hyg Asso J. 31:138,
1970.
2. Friedlander, S.K. Small Particles in Air Pose A Big Control Problem.
Environ Sci Technol. 7:1115-1118. 1973.
3. Whitby, K.T., and B.Y.H. Liu. The Electrical Behavior of Aerosols.
In: Aerosol Science, Davies, C.N. (ed.). New York. Academic Press,
1966.
-------
SECTION IV
TECHNICAL OVERVIEW
In this section we present a general model for particle collection de-
vices, indicate some of the forces at work in collectors, emphasize and
discuss electrostatic forces, and give a bibliography pertaining to the
augmentation of particulate pollution control devices through the ad-
dition of electrostatic collection mechanisms.
BASIC CONCEPTS OF PARTICLE COLLECTION
Two basic types of collector flow geometry exist, and these are indicated
in Figures 2a and 2b. Both rely on producing a component of aero-
sol particle motion perpendicular to the motion of the gas in which the
particle is borne. In one type, the particles are collected on the walls
of a channel. In the second, the particles are collected on an obstacle,
such as a sphere or cylinder, around which the gas flows.
For laminar flow of an aerosol of initially uniform concentration having
an average gas velocity v in a rectangular channel of length and width
L and W, the efficiency of collection (the fraction of the entering parti-
cles which are captured) will be
e = w LA? W ( s 1)
where w is the component of the particle velocity which is perpendicular
to the gas velocity. In this case e will have a maximum value of 1, com-
plete particle collection. The penetration Pn is defined as 1 minus the
efficiency. If several such channels were used in series and if they are
so positioned, or if the particles so behave, as to produce a random
11
-------
CHANNEL COLLECTOR
PARTICLE
-* v
w
W
L
(a)
FLOW BOUNDARY
PARTICLE
OBSTACLE
COLLECTOR
(b)
Figure 2. Particle collection in a channel and particle collection
by an obstacle within a flow
12
-------
distribution of particles after collection has occurred in each channel,
then the penetration through N such channels would be
Pn = (1 -
where . is the collection efficiency of each channel. For identical
channels this becomes
Pn = (1 - e)N
and for an infinite number of channels, or for "completely turbulent"
flow this is
Pn = exp (- Ne)
by the mathematical equality
exp (-Ne) = lira (1 - e) .
The exponential form is very significant. In the laminar flow case for
a single element, Pn = 1 - £ can go to zero and does so as e becomes 1.
In the second case, many elements or turbulent flow, Pn « 1 only if
Ne » 1. For both situations, the following clearly aid collection: high
particle collection velocity, low gas velocity (unless collection velocity
is strongly increasing with gas velocity), long and narrow channels. This
exponential relationship holds for the case where the collection occurs on
an obstacle, too.
Usually, the collection efficiency of an obstacle in a stream is given by
e = n A /A
' c f
where TJ is the "single collector collection efficiency factor" (or a
similar term), A is the collector cross-sectional area perpendicular
c
to the flow of the gas and A, is the cross-sectional area of the flow
associated with the collector. (For one collector Af would be the total
area of the flow, for two collectors in a plane perpendicular to the flow
each Af would be half the flow cross section, etc.) For a series of such
collectors, again the penetration would be
13
-------
Pn= (1 -
if they acted independently. As the number of such collectors in series
increases, this becomes the familiar exponential form for collection
efficiency:
Pn = exp (-N T) Ac/Af)
where N is the number of such collectors past which the gas flows, on the
average. Collection of a uniform aerosol in laminar flow on an obstacle can
be formulated similarly to collection in a channel by using
e = w L-'/'v W*
where L* is the "effective length" of the obstacle in the direction par-
allel to the mean gas velocity v and W* similarly is the effective dis-
tance between collectors in planes perpendicular to the mean gas velocity.
The analogy is imperfect (L* may well often be a function of v), but it
is true that increasing w and decreasing W* are expected to improve
collection. An equivalent common form is
= w A/Q
where A is the collector area normal to the migration velocity w and Q is
the volume rate of flow past the collector. Again, it may be that w = w(Q).
An approximate but very general formula for the collection efficiencies
of a great variety of control devices can be obtained as follows. Consider
a volume V in which particles are approaching the collection surface A at
a perpendicular velocity component w (deposition velocity) during the infi-
nitesimal period dt. If the concentration is n, and^uniform, then the number
of particles hitting (assumed captured) the surface will be nAw dt and the
change in concentration in this volume will be just this number of par-
ticles divided by the volume. This yields the following expression for
the instantaneous change in concentration:
- dn = n w A dt/V
14
-------
In turbulent flow the aerosol can be considered nearly uniform in many
cases, although its concentration changes with time, so this equation can
be integrated to give
n/n = exp (-w At/V).
Here, n is the initial concentration. This expression can be recast
o
into other forms by making some of the following substitutions:
t = V/Q
where Q is the volume flow rate of aerosol,
Q = A u
x oo
where A is the device inlet cross-sectional area and u is the inlet
o o
face velocity. The resulting forms are
n/n = exp (-w A/Q)
n/n = exp (-w A/u A )
This analysis can be linked to the usual analysis for scrubbers and
filters by using for A the total cross-sectional area of the obstacles
and for w the product of the single particle efficiency and the dif-
ference between the particle velocity (u ) and velocity, if any, of the col-
lecting object (u ), so that
w A = TI (u - u )A .
1 p c' c
Collection systems will be highly efficient only if the argument of the
exponential function above is much larger than 1 in magnitude.
Expressions of the type exp (-w A/Q) are widely used in electrostatic
precipitation analyses (e.g., White ) and allow convenient and fairly
curate comparisons among control devices.
A more mathematical discussion of this is presented in Appendix C.
15
-------
COLLECTION MECHANISMS
This work focuses on the role of electrical forces, of which there are a
great number. In general, these forces will be effective when they pro-
duce a terminal velocity component perpendicular to the gas velocity
which is large enough to make w A/Q £ 1. In general, these forces will be
significant when they are larger than or of the same magnitude as other
collection "forces." Electrical forces will be significant when the ter-
minal particle velocities they produce (w) are larger than or comparable
to the particle flux velocities (number of particles collected per unit
area and unit time divided by concentration) produced by other collection
mechanisms. The following mechanisms of particle collection may also be
taking place while electrostatic mechanisms are at work:
« Impaction
o Interception
o Diffusion
e Sedimentation
Diffusiophoresis
Photophoresis
» Thermophoresis
t> Acoustical migration
2
More information on them is available in the literature (e.g., Fuchs ).
For control devices in which one or more of these is important, the
electrostatic contribution should be compared with them.
ELECTRICAL FORCES
The total electrical force on a particle will be the vector sum of the
various electrical forces which are acting. Strauss listed the following:
16
-------
Coulombic force between charged particle and charged
collector
Image force between charged particle and uncharged collector
Image force between uncharged particle and charged collector
Coulombic repulsion of particles charged to the same polarity
(space charge)
and to this list can be added:
Charged or uncharged particle with charged or uncharged
collector in a superimposed electric field (analyzed in
depth by Zebel)
Charged particle moving in a magnetic field (Lorentz force).
We reiterate that these forces are summed; thus the first three forces
will be acting when a single charged particle is in the presence of a
charged collector. Often, however, one force (such as the Coulombic)
will be orders of magnitude larger than the others.
The two major types of electrostatic force significant in particle col-
lection are the Coulomb force a charged particle is subjected to in an
electric field and the induced charge (dipole) force, which operates in
an inhomogeneous electric field. The Coulomb force is given by the
equation
F = qE
where
q = particle charge, coul
E = electric field, v - m" , N - coul .
2
The force due to induced polarization is
F = Xg Vpgrad (E2)
where
x = (3 Tt/8)(e - l)/(e + 2) for a sphere
j_i \? Y*
p = dielectric constant of particle
V = particle volume.
P
17
-------
Because the gradient of a homogeneous field is zero, this second force
is only operative in inhomogeneous fields, such as those close to a
charged collector. Image forces are a class of induced polarization
force.
Electrostatic augmentation of collection efficiency of particulate control
devices can be expected to rely primarily on either or both of these
electrostatic forces, but we shall discuss a number of others as well.
As noted, the collection efficiency (or the penetration) can be calculated
knowing the migration velocity for the particles and knowing the collec-
tion area and the volume flow rate. We shall calculate the migration ve-
locities, w, for the particles under a variety of situations, but making
the following assumptions:
the particle has been charged to saturation in a corona
discharge with a field of 10 kV/cm = 33.3 esu,
the field is evaluated for its magnitude right at the
collector surface,
e the particle is assumed to have reached its terminal
velocity under the calculated force,
the particle is in air at 20 C and 1 atm pressure,
o the fluid resistance is given by Stokes law with the
Cunningham correction.2
From these calculations one should be able to make estimates of collec-
tion efficiency for various specific situations.
18
-------
Charge Acquired By Particles
The charge, q , acquired by a particle of diameter d >0.5 Urn in an
l
electric field, E, in the presence of an ionic concentration, N., is:
q = n e =
P P
1 + 2 ep - r
e 4- 2
P
E dp2"
4
" C/te "
1 + t/t
e
P
e
where n = number of charges
electronic charge
particle dielectric constant
charging time constant.
The value for t depends upon the ion mobility and concentration as
I e
follows:
t = I/ire N Z.
e I
where e = electronic charge, 4.8 x 10 stat-coul
-3
N. = ion concentration, cm
Z. = ionic mobility
= 2.2 (cra/s)/(v/cm) = 660 esu
t = 1.0 x 10 /N, seconds.
White notes that Cochet extended the applicability of these two equa-
tions down to d >_ 0.04 pm by replacing the first bracket in the first
equation with
19
-------
- 1
i -4
where A = 0.1 x 10 era.
-4
At d = 0.25 x 10 cm, e = 4, the replacement changes the value in
brackets from 2.0 to 4.5, a difference which may be significant for
some applications. Cochet's equation seems to eliminate the need for
a separate expression for the diffusive contribution to submicron
particle charging.
Many particles behave as though they were conductors (e -> °°) during
the time scales of interest (see Appendix A), so that for many calcu-
lations (e - l)/(e + 2) = 1 is appropriate. Furthermore, control
device designs usually assure sufficiently long residence times for
(t/t£)/(l + t/t ) = 1.0. Under such conditions:
q = n e = (1 + ZX'/d )2 + 2(1 + 2X'/d )~1 Ed 2/4
P P L P PJLPJ
or, even (d £ 1 urn)
P
q = 3 E d 2/4.
P P
The above equations enable determination of the particle charge with
several different degrees of sophistication. Table 1 has been calcu-
lated using the next-to-last equation above, based upon Cochet's.
20
-------
Table 1. APPROXIMATE MAXIMUM CHARGE (COC11ET EQUATION)
(e »1, E = 10 kV/cm = 33.3 esu)
P
Particle diameter, d
(ym) p
0.1
0.3
1.0
3.0
10.0
Particle charge, q
(esu) P
_9
8.05 x 10
-8
2.80 x 10
_7
2.59 x 10
2.26 x 10"6
2.50 x 10"5
Electrons, n
P
16.8
58.3
539.6
4.7 x 103
5.2 x 104
Coulomb Force - Charged Particle In a Field
The force, F , on a charged particle in an electric field, E, is
The charge on a spherical, conducting sphere is approximately (d > 1 u,m)
q = 3 E d /4.
P P
Assuming the charging field and the precipitating field are both near
the field at which air breaks down E - 10,000 V/cm -33 stat-volt/cm,
then
o 3E2dp2/4
can be evaluated, as it has in Table 2, to give approximate forces and mi-
gration velocities for conducting spheres. This table has these quantities
calculated & la Cochet, too. Note the difference for d < 1 ^m, for which
the Cochet equation results should be used.
21
-------
Table 2. FORCES, VELOCITIES FOR CONDUCTING SPHERE CHARGED
AND PRECIPITATED AT E = 10 kV/cm
Particle
diameter
0.1
0.3
1.0
3.0
10.0
Field Charging
Force
(dynes)
8.17 x 10~8
7.35 x 10~7
8.17 x 10"6
7.35 x 10~5
8.17 x 10~4
Velocity
(cm/s)
13.7
22.4
55.7
151.0
487.0
Field Charging (Cochet)
Force
(dynes)
2.68 x 10~7
9.32 x 10~7
8.62 x 10~6
7.52 x 10~5
8.33 x 10~4
Velocity
(cm/s)
45.0
28.4
58.8
155.0
496.0
The velocities are calculated here by
w = FB
where B
B
C
C
X
mobility, (cm/s)/dyne, s/g
C/3irud
= Cunningham slip correction
1 + 2 £- (1.257 + 0.400 e
a
-1.10 d /2X.
P j
p -4
0.0653 x 10 cm, mean free path of gas.
The equation comes from Davies (1945) and values have been taken from
a table by Gussman (1971).
Charged Particle With Uncharged Collector
The charge on the particle, q , produces an altered charge distribution
in the collector and a net attraction between the two, F , having a
magnitude given by Kraemer and Johnstone (1955);^
22
-------
for a conducting spherical collector and (see Lundgren and Whitby )
by:
2 2
F = q /d
m P c
for a conducting cylinder. (F must be multiplied by (e - l)/(e + 1)
for a dielectric cylinder.)
Assuming a particle charging field E = 10 kV/cm = 33 stat-volt/cm, and
sufficient charging time, we obtain the force and terminal velocity
values shown in Table 3 for conducting spherical collectors (e.g.,
water drops). (The Cochet charging equation has been used.)
This force is -(q /q ) times the value of the Coulomb force for a
charged collector (q^) and a charged particle, which means it is
usually much less than the Coulomb force (usually q >>q ).
Uncharged Particle, Charged Collector
The image force concept is really just a method for calculating the force
on a body due to an inhomogeneous electrical field, given by
F = Kg Vpgrad (E2)
as noted.
For a conducting sphere
23
-------
Table 3. FORCES AND TERMINAL VELOCITIES FOR PARTICLES CHARGED AT
E = 33 esu, COLLECTED IN THE PRESENCE OF UNCHARGED
CONDUCTING SPHERESa
Particle
diameter
d , urn
P
0.1
0.3
1.0
3.0
10.0
Forces (dynes) due to collector of diameter, d£ , urn
d = 0.1 urn
c
2.59 x 10~6
2.8 x 10~5
2.68 x 10~3
0.204
25.0
d = 1.0 |im
2.59 x 10~8
2.8 x 10~7
2.68 x 10~5
2.04 x 10~3
0.250
d =10. (am
c
2.59 x ]/T10
2.8 x 10~9
2.68 x 10~7
2.04 x 10~5
2.50 x 10~3
d = 100. urn
c
2.59 x 10~12
2.8 x 10"11
2.68 x 10~9
2.04 x 10~7
2.50 x 10"5
d = 300. urn
2.88 x 10~13
3.11 x 10~12
2.98 x 10~10
2.27 x 10~8
2.78 x 10~6
Velocities (cm/s)
0.1
0.3
1.0
3.0
10.0
435.
855.
1.83 x 104
-3 x 104 b
-3 x 104 b
4.35
8.55
183
4.20 x 103
.3 x 104 b
0.0435
0.0855
1.83
42.0
1.49 x 103
4.35 x 10 4
-4
8.55 x 10
0.0183
0.420
14.9
4.83 x 10~5
_5
9.49 x 10
2.03 x 10~3
0.0467
1.66
See text for assumptions.
b
Derivations not applicable to sonic velocities.
-------
becomes
F =^ d' T(E2)
T t e. u v V" /
1 lo p
Generally, we are dealing with spherical or cylindrical collectors for
_ 2
which V (E ) has only a radial component, 2 EdE/dr. If E is -33 stat-
volts/cm, an approximate upper limit, and if (as is also generally true)
dE/dr - 2E/d where d is the collector diameter, then
t- C
"l6d 2E(2E/dc>
*
-------
Table 5. TERMINAL VELOCITIES (cm/s) FOR COMBINATIONS GIVEN IN TABLE 4
Particle size
0.1 ym
0.3
1.0
3.0
10.0
Collector diameter
1 urn
4.5
22.1
183.
1494
A
1.60 x 10
10 pra
0.45
2.21
18.3
149.4
3
1.60 x 10
100 urn
0.045
0.221
1.83
14.94
160.
This force can be conveniently compared with the Coulomb force, F =
by using the approximate equation for q for a conducting sphere.
q = 3 Ed /4 for d > 1 u
P P P
so
and
F = 3 E d /4
c p
2 d
I_ _E
3 dc
Thus, the image force becomes increasingly weak compared with the Coulomb
force, as the particles become smaller with respect to the collectors.
The exact equations are available in the work by Kraemer and Johnstone
(1955),6 from which most treatments of this derive, but they require
similar approximations to characterize a given situation, as the force is
a function of distance and some "typical" distance must be chosen.
These analyses indicate image forces are weak compared to Coulomb forces
in most practical situations.
26
-------
Collection on a Cylinder in the Presence of an External Field
Two dimensionless parameters dominate the analysis of collection in
8 9
an external field (Zebel, 1965; Hochrainer et al., 1969 ):
G = E q B / V ,
00 ^ ' CO*
the ratio of the velocity of the particles of charge q in the external
P
field E to the free stream gas velocity V , and
H = 2(Qc/L)qpB / V 0
indicates repulsion. G > 0 indicates the field force and the flow are
aligned for positively charged particles.
For uncharged particles, the dimensionless parameter
F =
takes the place of H and scales the particle-collector force due to
induced charges.
A qualitative description of various field and attraction/repulsion
combinations is presented in Table 6. The most efficient system is
one with the particles and collectors oppositely charged and the
27
-------
electric field such chat particles are impelled in the direction of
Clow. This is G > 0, H < 0, |H| > |F|, and for it, Zebel (1969)10 cal-
culated a single fiber efficiency:
H = - irH/(l + G).
This assumes negligible diffusive flux. The same formula applies for
G < 0, if H < 0 too. A problem with this efficiency formula is that it
predicts better efficiency as G -» 0 in the case G > 0 which is described
as the case where "field direction facilitates separation." If the field
helps collection, it should help it more when strong than when weak.
Table 6. DESCRIPTION OF COLLECTION REGIMES FOR A SINGLE COLLECTOR,
A CHARGED PARTICLE, AND AN EXTERNAL FIELD, BASED ON WORK
OF HOCHRAINER ET AL.9 AND ZEBEL10
External field force
parallel to flow
External field force
antiparallel to flow
Particle-collector
attraction (H < 0)
Highest efficiency,
collection on
whole collector (a)
Collection on rear
of collector (d)
Neutral conditions (H = 0)
Collection on front
of collector (b)
No collection (e)
Particle-collector
repulsion (H > 0)
Little or no
collection (c)
No collection (f)
11 8
Figure 3 is taken from Davies (1973), based upon Zebel (1965).
When H = 0, and both the particle and the collector are uncharged, the
Q
inductive force group F comes into play. Zebel (1965) shows the effi-
ciency n to be
28
-------
(o) H= -05
(b)H=0
(c) H = 0 5
to
vO
(d)H=-OI
(e) H = 0
(f) H = 05
Figure 3. Trajectories of charged particles carried by viscous flow past an isolated,
charged, dielectric cylinder in a polarizing field parallel to the flow
direction (Zebel)12
-------
n = F/2 for F <. /2
for potential (ideal) flow, and
n = F/2 for F « 1
for viscous flow, where
F = F(2.002 - In Ref),
a group modified for low Reynolds number (Re ) flow a la Lamb.
Q
For an uncharged collector and a charged particle, Zebel (1965) derived,
for potential flow,
n = 1 + n' -1 <. n' <. 1
n = 2 /n"1 n1 >. 1
where
The corresponding formula for viscous flow is ri = n (G > 0).
The values for the single collector efficiency, n, can be connected with
the effective migration velocities, w, by noting that the penetration
expressions for the two approaches are
-wA/Q -T)A /A
_ CO
Pn = e = e
30
-------
where A = total collection area
A = total collector cross-sectional area perpendicular
to flow
A = collection volume face area perpendicular to flow
Q = V A = volume flow rate
o °*
V = free stream velocity upstream from collection volume.
00
Thus
7 ' "V*
CO
From these expressions for single element collection efficiency for
cylindrical collectors, we can draw the following conclusions:
1. If neither the particles nor the collectors are
charged, then the imposition of an external
electric field will enhance collection.
2. If both the particles and the collectors are
charged, then the imposition of an electric
field will reduce collection if the external
field produces a force which retards particle
motion in the direction of the free stream
velocity, and this field will enhance collec-
tion if it accelerates the particles in the
direction of the free stream velocity.
3. If the particles are charged, but the collector
uncharged, the collection efficiency increases
with the applied field.
Lorentz Force
A charged body in motion transversely with respect to a magnetic field
experiences a force, F , given by the vector cross product:
31
-------
P. = q v x H/c
L p
where v = particle velocity, cm/s
H = magnetic field, oersted
c = speed of light, 3 x 10 cm/s.
(This "Lorentz Force" is not, strictly, an electrostatic one.) Zebel
12
(1968) compared this force with the Coulomb force by noting that al-
though magnetic fields of H - 6 x 10 oersted are possible compared
with electric fields E - 60 stat-volt/cm, the ratio of the Coulomb
force, F , to the Lorentz force,
c
F /F_ = c
c L
is dominated by the ratio (v/c) so that
FT = (H/E) (v/c) F
JL C
FT < (6000/60) (3 x 10A / 3 x 1010) F
L - c
FT < 10 F
L c
Conclusion
F « F even when the particles are moving at the velocity of sound.
L c
Thus the Lorentz force is an unlikely candidate for the enhancement of
particulate collection efficiency because it is so much under then FC-
Space Charge Precipitation (Electrostatic Scattering)
An aerosol containing charged particles will expand due to mutual par-
ticle repulsion if the particles are charged to the same polarity. If
32
-------
the particles have different polarities, expansion will be retarded
but agglomeration will be enhanced. A simple analysis of this phcnora-
2
enon is presented in the book by Fuchs (1964). Either case results in
a decrease of the number concentration, c, of the particulate matter.
Faith et al. (1967) published results of their theoretical and experi-
mental investigations into the use of the charge on an aerosol exposed
to a corona as the driving force for deposition of the aerosol. Their
2
calculations indicated that the expression given by Fuchs (1964) and
others for the decay with time of particle concentration in a station-
ary medium is almost exactly correct (to within 10 percent) for turbu-
14
lent flow and plug flow. Wilson (1947) had demonstrated that this
expression is also correct for perfect stirring. The fraction of par-
ticle concentration at time t in comparison with that at t = 0 is:
Pn = c/c = 1/(1 + 4irBq 2c t) = 1/(1 + a)
o p o
where B = particle mobility (mechanical), cm/s-dyne
q = particle charge, stat-coul
_3
c = particle concentration, era
t = residence time, s.
Two limiting cases are easily identified:
for a «1,
Pn = (1 - 4irBq 2c t)
for a »1,
2
Pn = l/4TiBq c t.
P o
33
-------
The latter case is the one desired for efficient particle removal.
From this last equation, it is clear that high charge and high concen-
trations favor electrostatic scattering. If the penetration is multi-
plied by the initial concentration, c , to obtain the final concentration,
c, the c factor cancels out in this last expression, indicating that the
outlet concentrations will be independent of the inlet concentrations, as
long as the electrostatic scattering parameter, a, is much greater than
one. Thus it could be designed to meet certain outlet concentration
restrictions.
The residence time is just the length of the collector divided by the
flow velocity, or the volume of the collector divided by the volume
13
flow rate. Adapting an example given by Faith et al. (1967), a smoke
3 3
having a concentration of 2.3 g/m (1 gr/ft ) of 1 urn spherical par-
ticles of unit density charged in a corona at 5 kv/cm to an average of
261 electronic chaiges per particle would give a value of a = 4.6, thus
a penetration of 1/(1 + 4.6) = 0.18, a collection efficiency of 82 per-
cent. Figure 4 gives the fractional penetration, c/c , for values of
a between 1 and 50. Values of a other than that calculated in this ex-
ample can be approximated by noting that the charge will increase with
approximately the square of the particle diameter, but the number con-
centration, c , will decrease with the cube of the particle radius, and
the mobility is roughly proportional to the inverse of the particle
diameter. From which, we can conclude that for particles having the
same density and being > 1 ym in diameter, it is nearly true that the
penetration will be proportional to the mass concentration but indepen-
dent of the mean particle size. (The Cochet correction for particle
charging predicts somewhat of an increase in collection for particles
with mean sizes much less than 1 urn.)
This analysis produces the rather surprising conclusion that 0.47 m /s
(1000 cfm) of a fully charged aerosol having a concentration about
34
-------
1.0
0.00
0.5
0.3
0.50
070
u>
Ul
0.2
tr
H
UJ
O.I
080
o
z
UJ
(J
090 uJ
0.05
095
0.03
0.97
0.02
35 10 20 30 50
ELECTROSTATIC SCATTERING PARAMETER.a
I I I 10.98
100
Figure 4. Penetration versus electrostatic scattering parameter
-------
2.3 g/m could be reduced in concentration hy the factor 0.2 by passing
through a duct 0.7 m (2.3 ft) square by 1 m (3.1 ft) long, as long as
the material was not recntrained.
3 3
The tests made by Faith et al. were at only 3 x 10 cm /s (6.4 cfm),
very low in comparison with commercial flow rates. The problems with
such a device would be little different from those encountered with
the conventional electrostatic precipitators, however. Three such prob-
lems are the achievement of complete particle charging, the prevention
of particle reentrainment, and the periodic cleaning of the surfaces.
The main advantage such a space charge precollector would have would be
ease of retro-fit; a secondary advantage would be that it could be
arranged so that the major part of the collection (the major part of the
residence time) occurs at the collector to which it is attached.
Such a configuration might be attractive as a precollector used ahead
of another control device or even as a complete control device itself.
Charged Particle in Space Charge Field
An aerosol which has all its particles of the same charge, q , can be
collected on a conducting sphere due to the mutual repulsion of the
particles, the force being (Kraemer and Johnstone, 1955):
np (,/6) dc3/r2
which is the same as the attractive force of that same aerosol concen-
tration n having opposite polarity, -q , and filling the spherical
collector volume (u/6) d
c
Again approximating r by d /2 gives
F = 4 q 2 n (ir/6) d
s Mp p c
36
-------
An aerosol having a mass mean diameter of 3 pm, a particle density of
3 33
1 g/cm , and a mass concentration of 1 g/ra (0.44 gr/ft ) has
4-3 Q
n = 7.1 x 10 cm . Such an aerosol would have FS = 7.6 x 10"y dyn for
P _3
d = 100 pm and a terminal velocity of 15.6 x 10 cm/s. This force
C 3
is -(q /q )(n d ) times the Coulomb force. In some cases, it could
contribute appreciably to collection.
Figure 5 is a graphical presentation of theoretically calculated
migration velocities corresponding to the force which would be generated
on each specific size particle, generated by various electrostatic mech-
anisms. These migration velocities vary between 10~4 and 102 cm/sec,
with large variations due to both different electrostatic phenomena and
the size of the particles for a given electrostatic phenomena.
The simplest case and the one producing the largest forces (highest
migration velocities) is that of Coulomb force on a charged particle
in an electric field. The charge on the particles was chosen to be
the saturation charge obtained in a breakdown field, calculated using
the Cochet equation, which considers the contribution of diffusion in
particle charging. The initial migration velocity is higher for the
0.1 pm diameter particle than for the 0.3 pm diameter particle, and
then increases steadily with increasing particle size. The dip in the
curve at 0.3 pm corresponds roughly to a minimum, after which diffu-
sional effects enhance particle charging to a greater extent as the
particle diameter decreases. The migration velocities possible with
Coulomb forces are by far the greatest of any considered for any given
particle size.
The figure displays the migration velocities produced for charged particles
in the presence of a 100 pm diameter uncharged spherical conductor. The
particles are assumed to be charged as previously predicted using the
Cochet equation, and the force is calculated as the force produced by
the altered charge distribution in the collector, which was produced via
37
-------
CE
O
I03
I0
10
e
u
t iou
o
3 .
> .3
10'
10
-3 .
10'
'4
#
/
X
x
X
erx\
\
\
X
X
X
'
o I COULOMB FORCE-CHARGED PARTICLE IN
A FIELD.
A 2 CHARGED PARTICLE WITH UNCHARGED 100/j m
COLLECTOR. I
Q 3 IMAGE FORCE-CHARGED lOO^t m COLLECTOR
WITH UNCHARGED PARTICLE.
X 4 CHARGED PARTICLE IN SPACE CHARGE FIELD.
I I
10
PARTICLE DIAMETER,
30
100
Figure 5. Theoretically calculated migration velocities for
four electrostatic mechanisms versus particle
diameter
38
-------
the charges on the particle. The forces increase steadily with increas-
ing particle diameter and display a tendency towards a minimum or level-
ing in the smaller particle size range for which calculations were made.
A third force is that which a charged collector (100 urn diameter col-
lecting sphere) produces on an uncharged particle due to the inhomogeneous
electric field away from the collector. These forces increase steadily
with increasing particle diameter throughout the calculated range, and
are second in magnitude only to the Coulomb force. These migration
velocities are nearly equivalent to those produced via Coulomb attrac-
tion for the larger particles; however, they tend to decrease more
rapidly than the Coulomb forces with decreasing particle size, making
them much less desirable than Coulomb forces for very fine particles.
The fourth case corresponds to the migration velocity produced by the
force exerted on a charged particle by an uncharged conducting 100 urn
sphere, due to the space charge created by the other charged particles.
Further explanation of this phenomena occurs earlier in the text. The
assumptions utilized to make this calculation were that the aerosol had
a mass mean diameter equivalent to that plotted on the curve, and that
the particle density was 1 g/cm and the particle mass concentration
3
was 1 g/m . This, of course, means that each particle diameter corres-
ponds to a different aerosol of different particle size distribution
and number concentration. While these assumptions limit the applica-
bility of the analysis somewhat more than in the previous cases, it is
felt that the results are sufficiently valid for comparison with the
other three cases. It can be noted that the migration velocity actually
decreases with increasing particle size, showing a strong tendency to
level out over 1 ym. The migration velocities are small overall, with
the higher values for the very fine particles being somewhat biased by
the assumptions used in the method of calculation.
39
-------
We can summarize these calculations by noting the following relative
ordering of the forces (for a 0.3 urn particle and the already noted
assumptions), ranked from largest to smallest:
Coulomb force between charged particle
and charged collector
Force between charged collector and
uncharged particle
Force due to mutual charged particle
repulsion toward collector
Force between charged particle and
uncharged collector.
Increased collection efficiency can be expected by increasing any of
these forces.
CONCLUSIONS
Electrostatic forces can be appreciable, especially in the particle
size range of 0.1 to 1 um, generally the most difficult particles to
collect, yet particles which have health and visibility impacts out
of proportion to their mass concentrations. Generally, the Coulombic
forces (charged particle in an electric field) predominate over image
forces (either particle or collector charged, but not both). Mutual re-
pulsion of charged particles can also be significant for highly charged,
highly concentrated aerosols. With the simple exponential equation used
here and an estimate of the migration velocity w (calculated from the
equations above), order-of-magnitude calculations of penetration should
be relatively easy to make to judge the probable impact of various al-
ternatives for electrostatically augmenting fine particle control equip-
ment. We end this section with a bibliography concerning such electro-
static augmentation.
40
-------
ELECTROSTATIC AUGMENTATION BIBLIOGRAPHY
Table 7 is a selected bibliography related to electrostatic augmentation.
A few books on electrostatics and on aerosols in general have been listed,
but primarily the bibliography is specific to the important aspects of
electrostatic augmentation of control devices, aspects ranging from the
charging of particles to their removal after collection. This biblio-
graphy should be useful to those who wish to study or apply electro-
statics to the problem of particulate pollution control.
41
-------
Table 7. BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
GENERAL ELECTRICITY AND MAGNETISM REFERENCES
BoCtchcr, Carl Johan Friedrich. Theory of Electric Polarization.
Amsterdam, Houston, Elscvier Pub. Co. 1952.
CNRS (Centre National De La Recherche Scientifique). La Physique des
Forces Elcctrostatiques et Leurs Applications (Grenoble). 1960.
Frederick, E.R. The AFC Static Generator and Evaluator. Am
Dyest Rep. 31-33. July 15, 1968.
Harnwell, Gaylord P. Principles of Electricity and Electromagnetism.
McGraw-Hill Book Company, Inc., New York. 1949.
Harper, W.R. Contact and Fractional Electrification. Oxford
University Press, Ely House, London. 1.
Jackson, John David. Classical Electrodynamics. John Wiley & Sons,
Inc., New York. 1962.
Klyarfel'd, B. N. Investigations into Electrical Discharges in Gases.
(Trans, from Russian) MacMillan Company, New York. 1964.
Llewellyn-Jones, F. lonization and Breakdown in Gases. John Wiley &
Sons, Inc., New York. 1957.
Loeb, L. B. Static Electrification. Berlin, Springer-Verlag. 1958.
Loeb, Leonard B, Electrical Coronas, Their Basic Physical Mechanisms,
Berkeley, Univ. of Calif. Press. 1965.
Panofsky, W. K. H., and Melba Phillips. Classical Electricity and
Magnetism. Addison-Wesley Publishing Co., Inc., Reading, Mass. 1955.
Samaras, Demetrios G. Theory of Ion Flow Dynamics. Prentice-Hall, Inc.,
Englewood Cliffs, New Jersey. 1962.
Smythe, William R. Static and Dynamic Electricity. McGraw-Hill Book
Company, Inc., New York. 1950.
Conference on Static Electrification, 2nd, London, 1967. Proceedings of
the Conference organized by The Institute of Physics and The Physical So-
ciety - Static Electrification Group. London, May 1967. Institute of
Physics and Physical Society Conference Series No. 4.
Stratton, Julius Adams. Electromagnetic Theory. McGraw-Hill Brook
Company, Inc., New York. 1941.
GENERAL AEROSOL TECHNOLOGY REFERENCES
Davies, C. N. Aerosol Science. Academic Press, London and New York. 1966.
Davies, C. N. Air Filtration. Academic Press, London and New York. 1973.
Fuchs, N. A. The Mechanics of Aerosols. Pergamon Press - The MacMillan
Company, New York. 1964.
42
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Crccn, H. L., and W. R. Lane. Particulatc Clouds: Dusts, Smokes, and
Mists. Spon, England. 1964.
PARTICLE CHARGING MECHANISMS
Balabanov, E. M. Particle Charge in the Electric Field of a Corona
Discharge in a Very Dusty Medium. Elcctr Tcchnol USSR. (English
Translation) 1:109-196, February 1965.
Barreto, E., et al. Production and Neutralization of a Charged Aerosol
by Corona Fields. J Geophys Res. 70:1303-1310. March 15, 1965.
Bricard, J., and J. Pradel. Electric Charge and Radioactivity of
Naturally Occurring Aerosols. Chap. IV of Aerosol Science. C. N. Davies.
Academic Press, New York. 1966.
Chow, H. Y., and T. T. Mercer. Charges on Droplets Produced by Atomiza-
tion of Solutions. Am Ind Hyg Assoc. 32:247-255. 1971.
Cho, A. Y. H. Contact Charging of Micron-Sized Particles in Intense
Electric Fields. J Appl Phys. 35:2561-2564. September 1964.
Chubb, J. N. Experimental Studies of Airborne Particle Behavior in
Corona Discharge Fields. Inst. Elec. Engrs. Colloq. on Elec. Precipi-
tators. February 1965.
Cochet, R. Charging Laws of Submicron Particles. Colloq Int
C N R S. 102:331-338. 1960
Cooper, D. W., and P. C. Reist. Neutralizing Charged Aerosols with
Radioactive Sources. J Colloid Interface Sci. 45:17-26. 1973.
Cooperman, P. Theory of Space-Charge Limited Currents with Application
to Electrostatic Precipitation. Trans Am Inst Electr Eng. (Comm.
Electron.). 79:47-50, Part 1. 1960
Cooperman, P. Dust Space Charge in Electrical Precipitation. IEEE
Trans Cotnmun Electron. 82:324-326. July 1963.
Dodd, E. E. The Statistics of Liquid Spray and Dust Electrification
by the Hopper and Laby Method. J Appl Phys. 24:73. 1953.
Einbinder, Harvey. Generalized Equations for the lonization of Solid
Particles. J Chem Phys. 26:948-953. No. 4. April 1957.
Fuchs, N. A., and Sutugin, A. G. High-Dispersed Aerosols, in Topics
in Current Aerosol Research, II, ed. by G.M. Hidy and J.R. Brock,
Pergamon Press, New York. 1973.
Gentry, J. W. Charging of Aerosol by Unipolar Diffusion of Ions,
J Aerosol Sci. 3:65-76. 1972.
Gill, E. W, B., et al. The Electrification of Liquid Drops, Proc Phys
Soc London. 65B:546-551. July 1952.
43
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Gunn, R. Diffusion Charging of Atmospheric Droplets by Ions, and the
Resulting Combination Coefficients. J Meteorol. 11:339-347. 1954.
Gunn, R. The Statistical Electrification of Aerosols by Ionic Diffusion.
J Colloid Sci. 10:107-119. 1955.
Gunn, R. Electrification of Aerosols by Ionic Diffusion. Am Journal
Phys 25:542-546. November 1957.
Harper, W. R. Contact and Frictional Electrification. Clarendon Press,
Oxford. 1967.
Hewitt, G. W. The Charging of Small Particles for Electrostatic
Precipitation, AIE Trans. 76:300, 1957.
Hignett, E. T. Particle Charging in Electrostatic Precipitation.
Inst. Elec. Engrs. Colloq. on Electrostatic Precipitators. 1965.
Hignett, E. T. The Rate of Charging of Fly Ash Particles in an ESP,
CERL Lab Note RD-L-N 118-65, November 1965.
Hignett, E. T. Particle-Charge Magnitudes in Electrostatic Precipitation.
Proc Inst Electr Eng. 114:1325-1328. September 1967.
Keefe, D., P. J. Nolan, and T. A. Rich. Charge Equilibrium in Aerosols
According to the Boltzmann Law. Proc R Ir Acad Sect B. 60:27-45. 1959.
Kunkel, W. B. The Static Electrification of Dust Particles on Dispersion
into a Cloud. J Appl Phys. 21:820. 1950
Kunkel, W. B. Charge Distribution in Coarse Aerosols as a Function of
Time. J Appl Phys. 21:833-837, August 1950.
Liu, B. Y. H. Aerosol Research-University of Minnesota Particle Tech-
nology Laboratory. J Air Pollut Control Assoc. 18:594-695, No. 10. 1965
Liu, B. Y. H., et al. Diffusion Charging of Aerosol Particles at Low
Pressure. J Appl Phys. 38:1592-1597. March 1967.
Liu, B. Y. H., et al. Particle Charging at Low Pressures. J Colloid
Interface Sci. 23:367-378. 1967.
Liu, B. Y. H. Effect of Pressure and Electric Field on the Charging of
Aerosol Particles. Univ. of Minn. Particle Lab Publ. 119. October 1967.
Liu, B. Y. H., et al. On the Theory of Charging of Aerosol Particles
in an Electric Field, j Appl Phys No. 3. 39:1396-1402. February
1968.
Liu, B. Y. H., et al. A Pulse-Charging Pulse-Precipitating Electro-
static Aerosol Sampler. Anal Chem. 40:843-847. April 1968.
Mirzabekyan, G. Z. The Charging of Conducting Spherical Particle with
a Radius of the Order of a Free Path Length for Ions in Air. Sov
Phys Tech Phys No. 7. 11:935-941. (English Translation) January 1967.
44
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Murphy, A. T., et al. Theoretical Analysis of Effects of Electric
Field on Charging of Fine Particles. Trans Am Inst Electr Eng
(Comm. Electronics). 79:318-326, Part 1. 1959.
Natanson, G. L. Theory of Charging Submicroscopic Aerosol Particles
as a Result of Capturing Gas Ions. Sov Phys Tech Phys. (English
Translation). 5:538+. 1960.
Pauthenier, M., et al. The Charging of Dielectric Particles by an
Ionized Field - a Theoretical Difficulty. Compt. Rend. 262B. 1249-1252.
May 9, 1966.
Penny, G. W., and P. L. Smith. The Charging of Nonsphcrical Particles
in a Corona Discharge. AIEE Conference Paper presented at Winter
General Meeting, New York, February 1961.
Schnitzler, H. On the Speed of Charging of Dust in a Corona Field.
Staub. 16:221-236. 1955.
Schutz, A. The Electrical Charging of Aerosols. Staub. 27:24-31, No. 12.
December 1967.
Smith, P. L., et al. Charging of Nonspherical Particles in the Corona
Discharge. Trans Am Inst Electr Eng (Comm,, Electronics), Part 1.
80:340-346. 1961.
Wajsfelner, R., G. Madelaine, J. Delhaye, J. Bricard, and B. Liu.
Influence of Conductivity on Latex Aerosol Charge. J Aerosol Sci.
1:3-7. 1970.
Whitby, K. T., and B. Y. H. Liu. Polystyrene Aerosols - Electrical
Charge and Residue Size Distribution. Atmos Environ. 2:103-116. 1967.
Whitby, K. T., B. Y. H. Liu, and C. M. Paterson. Charging and Decay
of Monodisperse Aerosols in the Presence of Unipolar Ion Sources.
J Colloid Sci. 20:585-601. 1965.
White, H. J. The Role of Corona Discharge in the Electrical Precipi-
tation Process. Electr Eng. 71:67. 1952.
ELECTRICAL DEPOSITION MECHANISMS
Censor, D., and Z. Levin. Electrostatic Interaction of Axisymmetric
Liquid and Solid Aerosols. Atmos Environ. 8:905-915. No. 9,
September 1974.
Davies, C. N. Electrical Forces in Filters. Chapter 6 in Air Filtration.
New York. Academic Press. 1973.
45
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Dawkins, G. A. Electrostatic Effects in the Deposition of Aerosols on
Cylindrical Shapes. Technical Report No. 15. Engineering Experiment
Station University of Illinois.
Dennis, Richard. Collection Efficiency as a Function of Particle Size,
Shape, and Density: Theory and Experience. J Air Pollut Control Assoc.
24:1156-1164. December 1974
Foster, W. W. Deposition of Unipolar Charged Aerosol Particles by
Mutual Repulsion. Br J Appl Phys. 10:206-213, May 1959.
Frederick, Edward R. Some Effects of Electrostatic Charges in Fabric
Filtration. J Air Pollut Control Assoc. 24:1164-1168. No. 12.
December 1974.
Gensler, William G. A Study of Individual Particle Trajectories in
Electrostatic Precipitation. Ph.D. Thesis, Carnegie-Mellon Univ.,
Pittsburgh, Pa. 1968.
Gensler, W. G. Experimental Method of Particle Trajectory Analysis
in Electrostatic Precipitation. IEEE Indus, and Gen. Applied Group
Third Conf. Rec. 41-56. September 1968.
George, H. F., and G. W. Poehlein. Capture of Aerosol Particles by
Spherical Collectors: Electrostatic, Inertial, Interception, and
Viscous Effects. Environ Sci Technol. 8:46. 1974.
Gillespie, T. The Role of Electric Forces in the Filtration of
Aerosols by Fibrous Filters. J Colloid Sci. 10:299-314. 1955.
Goyer G. G., R. Gruen, and V. K. LaMer. Filtration of Mono-
disperse Electrically Charged Aerosols. J Phys Chem.
58:137-142. 1954.
Hochrainer, D., G. M. Hidy, and G. Zebel. Creeping Motion of a Charged
Particle Around a Cylinder in an Electric Field. J Colloid Interface
Sci. 30:553-567. 1969.
Kirsch, A. A. and N. A. Fuchs. Studies on Fibrous Aerosol Filters III.
Diffusional Deposition of Aerosols in Fibrous Filters. Ann Occup Hyg.
11:299-304. 1968.
Kraemer, H. F., and H. F. Johnstone. Collection of Aerosol Particles in
Presence of Electrostatic Fields. Ind Eng Chem. 47:2426. 1955.
Correction in Ind Eng Chera. 48:812. 1956.
LaMer, V. K. Filtration of Monodisperse Electrically Charged Aerosols.
U.S. AEC Report No. NYO-514. Columbia University. 1952.
Lentert, Gerhard. Theoretical and Experimental Studies on Electrostatic
Filter Plates. Ph.D. Thesis, 136 pages, in German Eidgenossiche Tech-
nische Hochschule, Zurich, Switzerland. 1971.
46
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Lowe, H. J., et al. The Physics of Electrostatic Precipitation.
Bri J Appl Phys No. 2. p. S40-S47.
Lundgren, D. A., and K. T. Whitby. Effect of Particle Electrostatic
Charge on Filtration by Fibrous Filters. Ind Eng Chem Process Des
Dev. 4:345. 1965.
McKelvey, J. M. Polymer Processing. John Wiley & Sons, Inc., New York.
1962.
Natanson, G. Deposition of Aerosols by Electrostatic Attraction Upon
a Cylinder Around Which They are Flowing. Dokl Akad Nauk SSSR.
112:696-699. 1957.
Penney, G. W. Weakness in the Conventional Theory of Electrostatic
Precipitation, Am. Soc. Mech. Engrs. Preprint No. 67-WA/APC-l.
November 1967.
Penney, G. W. Some Problems in the Application o£ the Deutsch Equation
to Industrial Electrostatic Precipitation. J Air Pollut Control Assoc.
19:596-600. No. 8. August 1969.
Ranz, W. E., and J. B. Wong. Impact ion of Dust and Smoke Particles.
Ind Eng Chem. 44:1371-1381. No. 2. 1952.
Reese, Oliver, W. Numerical Method and FORTRAN Program for the Solution
of an Axisymmetric Electrostatic Collector Design Problem. NASA, Lewis
Research Center. NASA-TN-D-6959, NASA STAR:N72-31185 //. September 1972.
Rose, H. E., and A. J. Wood. An Introduction to Electrostatic Precipita-
tion in Theory and Practice. Constable and Co. Ltd. 10-12 Orange Street
W.C. 2. London, 1966.
Soo, S. L. Dynamics of Charged Suspensions, Topics in Current Aerosol
Research, II, ed. by G. M. Hidy and J. R. Brock, Pergamon Press,
New York. 1971.
Stenhouse, J. I. T. The Influence of Electrostatic Forces in Fibrous
Filtration. Filtr Sep. 11:25. February 1974.
Walkenhorst, W. Reflections and Research on the Filtration of Dust from
Gases with Special Consideration of Electrical Forces. J Aerosol Sci.
1:225. 1970.
Wasan, D. T., S. K. Sood, R. Davies, and A. Lieberman. Aerosol Trans-
port: Particle Charge and Reentrainment Effects. J Colloid Interface
Sci. 43:144-149. 1973.
Whitby, K. T. and B. Y. H. Liu. The Electrical Behavior of Aerosols.
in Davies, C.N., Editor, Aerosol Science, Academic Press, London and New
York. 1966.
Woodland, P.C., and E. E. Ziegler. Static Dust Collection on Plastics.
Mod Plast. 28. (9). 95-106, 169-178, May 1951.
47
-------
Table 7 ''continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Yoshioka, N., 11. Gmi, M. Hnttori, and I. Tamori. Effect of Electrostatic
Force in the Filtration Efficiency of Aerosols. Kanaga Kogaku. Chem
Eng Japan. 32:815-820. 1968.
Zebel, G. Deposition of Aerosol Flowing Past a Cylindrical Fibre in a
Uniform Electric Field. J Colloid Sci. 20:522. 1965.
Zebel, G. Capture of Small Drops Falling in Electric Fields. J Colloid
Interface Sci. 27:294. 1968.
Zebel, G. The Use of Electrical and Magnetic Forces to Separate and
Classify Aerosol Particles. Staub 28:1-4 No. 7. (English Translation).
1968.
ELECTROSTATIC AUGMENTATION
a. Filters
Darby, K., et al. Electrostatic and Fabric Gas Cleaners. Power Works
Eng. 62:33-41. No. 731. May 1967.
Frederick, E. R. How Dust Filter Selection Depends on Electrostatics.
Chemical Engineering. June 26, 1961.
Frederick, G. R. Some Effects of Electrostatic Charges in Fabric Filtra-
tion. J Air Pollut Control Assoc. 24:1164. 1974.
Havlicek, V. The Improvement of the Efficiency of Fibrous Dielectric
Filters'by Application of an External Electric Field. Int J Air
Water Pollution. 4:225-236. 1961.
linoya, K., and K. Makino. Application of Electric Field Effects to Dust
Collection Filters. J Aerosol Sci. 5:357-372. 1974.
Inculet, I.E., et al. A Two-Stage Concentric Geometry Electrostatic Pre-
cipitator with Electrified Media. Paper 69-2. Air Pollution Control As-
sociation 62nd Annual Meeting, St. Paul, Minn. 1969.
Kirsch, A. A. The Influence of an External Electric Field on the
Deposition of Aerosols in Fibrous Filters. J Aerosol Sci. 3:25. 1972.
Mazumder, M. K., and K. T. Thomas. Improvement of the Efficiency of
Particulate Filter by Superimposed Electrostatic Forces. Filtr
Sep. 4:25. 1967.
Van Orman, W. T., and H. A. Endres. Self-Charging Electrostatic Air
Filters. ASHVE Journal Sci. 157-163. 1952.
Zebel, G. Improving the Separation Efficiency of Fiber Filters by
Electrical Fields. Staub (English Translation). 26:18-22. No. 7. 1960.
48
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Zebol, G. Aerosol Deposition on a Single Fiber Under the Influence of
Electrical Forces. Staub (English Translation). 29:21-27. No. 2. 1967.
b. Scrubbers
APCA Technical Coord. Comm. TA-5. Specifications Required for Design of
Electrostatic or Combination Mechanical-Electrostatic Collectors for
Fly Ash Collection from Boiler Gases. J Air Pollut Control Assoc.
8:249. 1958.
Barth, W. Problem of Combining Electric and Mechanical Dust Collecting
Plants. Staub 24:441-444. November 1964.
Blohm, H. Development of Electrostatic Filters, Theory of Electric
Filters, Smoke Filters, Steam Filters, Comb. Filters. Braunkohle 10:481-
483. 1958.
Frauenfelder, A. Experience Gained with a New Scrubber Precipitator
Combination. Krupp Tech Rev. 22:125-126. November 1964.
Mukhlenov, J. P., et al. Increase in Efficiency of Wet Collection of
Highly Disperse Dust by Preliminary Particle Charging. Chem Ind.
p. 3704. May 1968.
Pilat, M. J., S. A. Jaasund, and L. E. Sparks. Collection of Aerosol
Particles by Electrostatic Droplet Spray Scrubbers. Environ Sci and
Technology. 4:360. 1974
Walker, A. B. Enhanced Scrubbing of Black Liquor Boiler Fume by
Electric Preagglomeration Pilot Plant Study. 1963. J Air Pollut
Control Assoc. 13:622-627. December 1963.
c. Cyclones
Cederholm, C. Collection of Dust From Refuse Incinerators in Electrical
Precipitators Provided with Multicyclone After-Collectors. Proc. Int.
Clean Air Congr. (London). 1966.
Heinrich, R. F. Study of the Suitability of Using Electrofliters with,
Preceding, or Following Mechanical Collectors. Mitt Ver Gross-
kesselbesitzer No. 68, 322-332. 1960.
Maartman, S. Collection of Dust from Oil-Fired Boilers in Multi-Cyclones
and Electrostatic Precipitators. Proc Int Clear Air Cong.
Part 1, 131-133. 1966.
Molyneux, F. Electrostatic Cyclone Separator. Chem Process Eng.
44:517-519. 1963.
Pegg, E. H. R. Dust Control by Combined Centrifugal and Electrostatic
Principles. Heat Vent. 44:78-80. May 1947.
Soo, S. L., et al. An Electro-Aerodynamic Precipitator, Paper No. 68-104
Air Pollution Control Association Annual Meeting. 1968.
49
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Walker, A. B., et al. Application of Mechanical Collector in Combina-
tion with Electrostatic Precipitators. Blast Furn Steel Plant 47:
622-624. June 1959.
d. Other
Faith, L. , et al., Particle Precipitation by Space Charge in Tubular
Flow. Ind Eng Chem Fundam. 6:519. 1967.
Javorsky, B. Fume Control and Gas Cleaning with an Individual Scale
Foam Bed Scrubber. Filtr Sep. 10:21. 1973.
MISCELLANEOUS
Abbott, C. G., J. E. Dye, and D. J. Sartor. An Electrostatic Cloud
Droplet Probe. J Appl Meteorol. 11:1092-1100. 1960.
Adams, N. G., and D. Smith. Studies of Microparticle Impact Phenomena
Leading to the Development of a Highly Sensitive Micrometeoroid Detector,
Planet. Space Sci Rev. 19:195-204. 1971.
Aliev, G. M. A., et al. Methods of Intensifying Operation of Electric
Filters for Extracting High-Resistance Magnesite Dust. Ogneupory.
32:11-18. No. 9. September 1967.
Allander, C., et al. The Effect of Particle Size Distribution on Effi-
ciency in Electrostatic Precipitators. Staub 17:738-745. 1957.
Batel, Wilhelm. The Electrostatic Precipitator : Development State and
Trends. Staub. 33:133-140. No. 3. March 1973.
Benson, J. R. , and M. Corn. Costs of Air Cleaning with Electrostatic
Precipitators at TVA Steam Power Plants. J Air Pollut Control
Assoc. 24:339-340. No. 4. April 1974.
Bickelhaupt, Roy E. Electrical Volume Conduction in Fly Ash. J. Air
Pollut Control Assoc. 24:251-255. No. 3. March 1974.
Bohm. J. Delay in Charging of Particles in an Electro-Precipitator.
Staub 28:270-273. No. 7. July 1968.
Caplan, Knowlton, J. Needed Research in Fabric Filtration. J Air Pollut
Control Assoc. 24:1194-1197. No. 12. December 1974.
Cohen, E. Research on the Electrostatic Generation and Acceleration of
Submicron-Size Particles. Space Technology Laboratories, Inc. Redondo
Beach, California.
Cooperman, P. Boundary Layer Effects in Electrostatic Precipitation.
Paper No. 66-124. Air Pollution Control Association 59th Annual Meeting.
San Francisco, California. June 1966.
Cooperman, P. A New Formula for the Efficiency of Electrostatic
Pre-
50
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Dalmon, J., and D. Tidy. A Comparison of Chemical Additives as Aids to
the Electrostatic Precipitation of Fly Ash. Atmos Environ. 6:721. 1972.
Darby, K., and D. 0. Heinrich. Conditioning of Boiler Flue
Gases for Improving Efficiency of Electrofilters. Staub 26:12.
November 1966.
Doyle, A., D. R. Moffett, and B. Vonnegut. Behavior of Evaporative
Electrically Charged Droplets. J of Colloid Science. 19:136-143. 1964.
Frauenfelder, A. Overcoming Special Problems in Electrostatic
Precipitation. Filtr Sep. 11:52-57. No. 1. January/February
1974.
Frisch, N. W., and D. W. Coy. Commentary on "Hot" versus "Enlarged"
Electrostatic Precipitation of Fly Ash: A Cost-Effectiveness Study.
J Air Pollut Control Assoc. 24:872-875. No. 9. September 1974.
Fry, F. A. Charge Distribution on Polystyrene Aerosols and Deposition
in the Human Nose. J Aerosol Sci. 1:135-146. 1970.
Henniker, J. Triboelectricity in Polymers. Nature. 196:474. November 3,
1962.
Jarzebski, S., and M. Chojnowski. Influence of Shapes of Corona Electrodes
on Density of Emission Currents in Electrostatic Dust Collectors. Atmos
Environ. 8:29. 1974.
Junge, C. E. The Size Distribution and Aging of Natural Aerosols as
Determined from Electrical and Optical Data on the Atmosphere. J
Meteorol. 12:13-25. 1955
Kallard, T. Electret Devices for Air Pollution Control. Optosonic
Press. New York. 1972.
Roller, L. R., and H. A. Fremont. Negative Wire Corona at High Temp-
erature and Pressure. J Appl Phys. 21:741. 1950.
Liesegang, D. Effect of Gas Temperature on Separation Efficiency and
Design of Electrostatic Precipitator Installations. Staub 28:403-405.
No. 10. October 1968.
Lowe, H. J., et al. The Precipitation of Difficult Dusts. Inst. Elec.
Engrs. Colloq. on Electrostatic Precipitators. February 1965.
Luethi, J. E. Basis for Electrostatic Precipitation of High Resistivity
Particles. Ph.D. Thesis. Ill p. in German. Switzerland. NASA STAR
N69-15042//. 1967.
Masuda, S. Statistical Observations on Efficiencies of Electrostatic
Precipitators. Staub (English Translation) 26:6-11. No. 11. 1965.
Min, K., et al. Particle Transport and Heat Transfer in Gas-Solid Sus-
pension Flow Under the Influence of an Electric Field. Nuclear Sci
and Eng. 26:534-546. 1966.
51
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Murphy, P. V., F. J. Holly, and W. Bernhard. Elcctrcts as Blood Com-
patible Prosthetic Material, from: Elcctrcts and Related Electrostatic
Charge Storage Phenomena. Edited by L. M. Baxt and M. M. Periman.
The Electrochemical Society Inc. Library of Congress Card No. 68-23768.
Oglesby, S. Jr., and G. B. Nichols. A Manual of Electrostatic
Precipitator Technology. Part 2; Application Areas. Southern
Research Institute. Birmingham, Alabama. NASA STAR: N71-2-21812.
August 25, 1970.
Ramadan, 0. E. Electrohydrodynamic Secondary Flow. Ph.D. Thesis.
Illinois University. Urbana, Electrostatically Induced Secondary Flow.
NASA STAR: N71-14229.
Robinson, M. Electric Wind Turbulence in Electrostatic Precipitation.
J Air Pollut Control Assoc. 17:605-606. 1967.
Robinson, M. A Modified Deutsch Efficiency Equation for Electrostatic
Precipitation. Atmos Environ. 1:193-204. May 1967.
Robinson, M. Turbulent Gas Flow and Electrostatic Precipitation. J
Air Pollut Control Assoc. 18:235-239. April 1968.
Robinson, M. Effect of Polarity on Particle Concentration Profiles in
Electrostatic Precipitators. J Air Pollut Control Assoc. 18:688-690.
1968.
Robinson, M. Electrostatic Precipitation in Air Pollution Control -
Part 1, ed by Wernher Strauss. Wiley Interscicnce, New York. 1971.
Rose, H. E., and A. J. Wood. An Introduction to Electrostatic Precipita-
tion Theory and Practice, 2nd Edition. Constable & Co. London. 1966.
Roure, L., and G. Madelaine. Contribution to the Study of the
Electrostatic Precipitation of Aerosols. French, English Summary.
p. 61. NASA STAR: N67-33029. 1966.
Seanor, D. A. Triboelectrification of Polymers. In: Electrical Proper-
ties of Polymers, Frisch, K. C. and A. V. Patsis (ed.). Westport, Con-
necticut, Technomic Publishing Co., Inc.a 1972. p. 37-58.
Selzler, D. R., and W. D. Watson, Jr. "Hot" versus "Enlarged" Elec-
trostatic Precipitation of Fly Ash: A Cost Effectiveness Study. J
Air Pollut Control Assoc. 24:115-121, No. 2. February 1974.
Shale, C. C., and G. E. Fasching. Operating Characteristics of a High
Temperature Electrostatic Precipitator. U.S. Bureau of Mines. Kept.
Invest. RI-7276. NASA STAR: N70-11373//. July 1966.
Shashova, Victor E. Static Electricity in Polymers. I. Theory and
Measurement. J Polymer Sci. 32:65-85. 1958.
Shulepov, Yu V., and S. S. Dukhin. In Reference to the Theory of
Electrical Coagulation of Spherical Aerosol Particles. Translated from
Kolloidn Zh. 24:749-751. No. 6. USAF FRD-MT-63-70. AD 434-958.
1962.
52
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
Simon, D. J., and R. Michclier. Study of the Influence of a Magnetic
Field on the Behavior of Electrostatic Separators. United Kingdom Atomic
Energy Authority. NASA STAR: N72-20620//. July 1970.
Skinner, S. M. Diffusion, Static Charges, and Conduction of Electricity
in Nonmetallic Solids by a Single Charge Carrier. I. Electric Charges
in Plastics and Insulating Materials. J of Applied Physics. 26:498-508.
No. 5. May 1955.
Southern Research Institute. Selected Bibliography of Electrostatic
Precipitator Literature. Southern Research Institute (Birmingham, Alabama)
NAPCA Contract CPA 22-69-73 Report. March 1970.
Strindehag, 0. M. Liquid Surface Electrostatic Precipitator. Rev Sci
Instrum. 58:95-99. January 1967.
Thomas, J. B., and E. Wong. Experimental Study of DC Corona at High
Temperatures and Pressures. J Apply Phys. 29:1226. 1958.
Thomas, J. W., and D. Rimberg. A Simple Method for Measuring the
Average Charge on a Monodisperse Aerosol. Staub 27:18-22. 1967.
Vonnegut. B., D. R. Moffet, P. M. Sliney, and A. W. Doyle. Research on
Electrical Phenomena Associated with Aerosols. Final Report to Army
Chemical Corps. A. D. Little, Inc. Cambridge, Mass.
Voorhoeve, R. J. H. Electrostatic Aspects of Aerosol Filtration 11:
Investigation of the Literature and of the Applicability of Electro-
statics in the Filtration of Aerosols. In Dutch, English summary.
NASA STAR: N66-37782//. May 1966.
Whitby, K. T., and W. E. Clark. Electric Aerosol Particle Counting and
Size Distribution Measuring System for the 0.015 to 1 urn size Range, Tel-
lus, 18:573-586. 1966.
Whitby, K. T., and C. M. Peterson. Electrical Neutralization and Par-
ticle Size Measurement of Dye Aerosols. Ind Eng Chem Fundam. 4:66-72.
1965.
Whitby, K. T. Generator for Producing High Concentrations of Small Ions.
Rev Sci Instrum. 32:1351-1355. No. 12.
White, H. J. Chemical and Physical Particle Conductivity Factors in
Electrical Precipitation. Chem Eng Prog. 52:244. 1956.
White, H. J. Modern Electrical Precipitation. Ind Eng Chera
47:932. 1955.
White, H. J. Fifty Years of Electrostatic Precipitation. Paper No.
57-35 presented at Golden Jubilee Meeting of Air Pollution Control Assoc.
St. Louis, Missouri. June 1957.
53
-------
Table 7 (continued). BIBLIOGRAPHY OF ELECTROSTATIC AUGMENTATION
White, H. J. Industrial Electrostatic Precipitation. Pergamon Press.
New York. 1963.
White, H. J. Resistivity Problems in Electrostatic Precipitation.
J Air Pollut Control Assoc. 24:313. 1974.
54
-------
REFERENCES
1. White, H. J. Industrial Electrostatic Precipitation. Pergaraon
New York. 1963.
2. Fuchs, N. A. The Mechanics of Aerosols. Pergamon Press-The
MacMillan Company. New York. 1964.
3. Strauss, W. Industrial Gas Cleaning. Pergamon, New York. 1966.
4. Davies, C. N. Definitive Equations for the Fluid Resistance
of Spheres. The Proceedings of the Physical Society. 57:4
No. 322 (1 July 1945).
5. Gussman, R. A. Tables for Use in Aerosol Physics. BGI, Inc.
Ualtham, Mass. 1971.
6. Kraemer, H. F. and H. F. Johnstone. Collection of Aerosol Par-
ticles in Presence of Electrostatic Fields. Ind Eng Chem
47:2426. Correction in Ind Eng Chem. 48:812. 1956.
7. Lundgren, D. A. and K. T. Whitby. Effect of Particle Elec-
trostatic Charge on Filtration by Fibrous Filters. Ind Eng
Chem Process Des Dev. 4:345. 1965.
8. Zebel, G. Deposition of Aerosol Flowing Past a Cylindrical
Fibre in a Uniform Electric Field. J Colloid Sci. 20:522.
1965.
9. Hochrainer, D., G. M. Hidy, and G. Zebel. Creeping Motion of
a Charged Particle Around a Cylinder in an Electric Field.
J Colloid Interface Sci. 30:553-567. 1969.
10. Zebel, G. Capture of Small Drops Falling in Electric Fields.
J Colloid Interface Sci. 27:294. 1968.
11. Davies, C. N. Electrical Forces in Filters. Ch. 6 in Air
Filtration by C. N. Davies. Academic Press, N.Y. 1973.
12. Zebel, G. Aerosol Deposition on a Single Fiber Under the
Influence of Electrical Forces. Staub 29:21-27, Noa 2.
February 1969.
13. Faith, L. et al. Particle Precipitation by Space Charge in
Tubular Flow. Ind Eng Chem Fundam. 6:519. 1967.
14. Wilson, I. B. The Deposition of Charged Particles in Tubes
With Reference to the Retention of Therapeutic Aerosols in the
Human Lung. J Colloid Sci. 2:271-276. 1947.
55
-------
SECTION V
ELECTROSTATIC AUGMENTATION OF FABRIC OR BED FILTRATION
FIBER BEDS TO CAPTURE PARTICLES ELECTROSTATICALLY1
The collection device consists of a loosely woven fiber bed used to fil-
ter previously charged fine particles. A relatively clean gas stream such
as downstream from a conventional ESP, would have the particles positively
charged by conventional saturation corona charging. The gas stream would
then pass through a loosely woven fiber filter of polypropylene, teflon,
or stainless steel, on which the particles are trapped. Cleaning would
be achieved by use of water sprays. (See Figure 6.)
Goals of the Study
The long-range goal is to provide a reliable means for removing submicron
particles from industrial waste gases. Short-term goals include fully
defining the mechanism responsible for fine particle removal, and evalua-
tion of the capacity of the mechanism for removal of very fine fly ash
generated in the combustion of high ash, low sulfur coal.
Methods of Study
Theoretical Definition of the envelope of conditions within which the
fibers significantly remove submicron particles was to be determined.
A mathematical model of the phenomena was to be developed as a follow-up
to the proposed experimental work, and such a model has been developed.
It is based upon the charging of the fiber bed by the fine particles
57
-------
VELOCITY
oo
SAMPLE
MEASUREMENT
(±1
7
SAMPLE
X
\
FIBER BED
AEROSOL GENERATOR
RELATIVE
HUMIDITY
MEASUREMENT
OPTICAL LOCATION OF
FIBER BED
FAN
Figure 6. Schematic of test apparatus for study of removal of charged submicron particles
by fiber beds ^
-------
collected thereon. Unequal collection within the fiber bed produces an
electric field gradient, proposed as a significant collection mechanism.
Self-consistent solutions to the equations of bed charging, electric field,
and charge leakage were obtained. It is reported that the bed became
charged even when no particles were being collected, suggesting contact
charging (triboelectrification) is a major electrical phenomenon, which
would cast doubt on the applicability of the theoretical model developed
thus far.
It has been proposed that image forces may cause the observed removal
of charged particles by the stainless steel fiber bed. Using an equation
29
for single fiber efficiency proposed by Natanson, the overall efficiency
was then computed, and they claimed that the calculated efficiencies agree
well with the theoretical efficiency expected due to image forces. It
was noted that the observed bed efficiency at 50 fpm was 18.6 percent
and that the theoretically calculated efficiency for these conditions was
18 percent.
Although the hypothesis that image forces were at work is not crucial to
the value of the investigation, it deserves closer scrutiny. As we will
show, there were minor calculational errors and a significant error (neglect
of the Cunningham slip correction) in the use of Natanson's equation for a
charged particle and an uncharged sphere. Once these are corrected, the
agreement with experimental results at 50 fpm (0.254 m/s) becomes worse.
There is some question about the validity of the Natanson expression as
well. It is actually derived for the case of a point charge approaching
a plane, rather than for a sphere approaching a cylinder. It predicts an
efficiency which depends upon an exponent containing the cube root of the
ratio of the migration velocity to the free stream velocity, rather than
an exponent with a linear dependence on this ratio. A linear dependence
is predicted by Natanson's own equations for charged-particle-to-charged-
cylinder efficiencies and uncharged-particlc-to-charged-cylinder efficien-
cies. A linear, rather than cube root dependence, is also predicted by
59
-------
the form e we have used. The cube root expression leads to the
anomalous result that small image forces produce much greater collection
than larger image or Coulomb forces (the charged cylinder cases).
Our own calculations using the equations and data as appears in Table 8
reveals that the theoretical overall bed efficiency due to image forces
at 50 fpm would be 59 percent.
It can readily be seen that the theoretically calculated efficiencies
utilizing the equation as given* is different from their own calculated
results and rather different from their measured efficiencies.
In the single fiber efficiency equation of Natanson, as used , the
Cunningham slip correction factor was neglected. Recalculating the
efficiencies, utilizing the Cunningham correction factor in the single
fiber efficiency equation yields results in column 5, Table 9. We now
have corrected the theoretical efficiency of the stainless steel fiber
bed, according to Natanson's corrected (by addition of Cunningham slip
correction factor) equation for single fiber collection efficiency for
a charged particle and uncharged cylinder.
If it is correct to assume image forces to be the collection mechanism,
and if Natanson's equation for single fiber efficiency is correct, then
we would expect that the overall collection efficiency would vary with
the cube root of the ratio of the face velocities. The experimental
data for the stainless steel fiber pad indicates a drop in efficiency
from 18 percent at 50 fpm, to zero at 350 fpm. If image forces were the
dominant mechanism, then we would expect the efficiency to drop only to
9 percent, not zero. It would then appear that there is a discrepancy
between experimentally observed efficiencies, calculated theoretical
efficiencies, and results expected from theory. This leaves open the
possibility that the theoretical efficiency equation is wrong, or the
assumed particle capture mechanism is wrong, or both.
60
-------
Table 8. INFORMATION UTILIZED TO CALCULATE THEORETICAL EFFICIENCY1
E = 1 - a'
// / N
ct = (4Mec
e = 0 l
c K+l
- L
e /4 TT
.
V d
1/3
(Reference 29)
-4
d = fiber diameter = 0.03 cm = 3 x 10 m
d = particle diameter = 0.22 x 10 cm
e = charge on a particle = 3.2 x 10 coul
-12
Q = permitivity of free space = 8.85 x 10
e = bed porosity = 0.9
K = collector dielectric constant
VQ = fluid velocity = 0.25 m/sec (50 ft/min)
L = bed depth ~ 15 cm.
From this, one calculates an individual fiber collection efficiency of
0.0059 at 500 ft/min and an overall bed efficiency of 18 percent at
500 ft/min and 34 percent at 50 ft/min.
Table 9. EXPERIMENTALLY OBSERVED AND THEORETICAL
FIBER BED EFFICIENCIES
Face
velocity
m/sec
0.254
1.778
2.54
Experimentally
observed
% efficiency^
18.6
0
a
Calculated
% efficiency-'-
34
b
18
GCA calculated
7o efficiency
with uncorrected
equations
59
b
34
GCA calculated
% efficiency
with corrected
equations
66
39
b
Not measured.
Not calculated.
61
-------
Natanson's equation for single fiber collection efficiency is:
Vf M dpy
1/3
and is questionable due to the cube root. The analysis in Appendix D
-2 30
and the d. dependence for the image force indicate a square root
rather than a cube root. If we calculate the image force on a 0.22 micron
particle in a conducting cylinder, we get FT = 4.1 x 10 dynes, corre-
sponding to a migration velocity of 2.05 x 10~ cm/s. Using an equation
for overall efficiency of the form:
E = 1 - e
-w A/Q
where A is the surface area of the collector (filter), Q is the volume
throughput, and w is the particle migration velocity (force times par-
ticle mobility); we get an overall filter collection efficiency for the
previously calculated migration velocity of less than 0.0002 at 50 fps
(0.254 m/s). This is in close agreement with the efficiency calculated
using the Natanson equation without the cube root factor (0.0003), but
is much less than the square root (0.017), much less less than the cube
root (0.067). If we use Natanson's equation for the single fiber collec-
tion efficiency for a charged particle and charged cylinder to determine
the migration velocity required to achieve 66 percent overall collection
efficiency (which the cube root form of the image force equation pre-
dicted) we get:
1 -
E = 1 - e"°
a = 70.7n for the filter in question
n = Tiw/V
o
0.66 = 0.34 - e-70'7 * w>°'254
w = 0.0012 m/sec = 0.12 cm/sec.
62
-------
Thus Nntanson's equations for Coulombic attraction require w = 0.12 cm/s
to obtain 66 percent efficiency. Then if w = F B, the force required
would be
F = 2.47 x 10 dynes.
If we use the similar, simple collection efficiency equation for the
same conditions we get:
E = 1 - e
0.34 = e-" "6/0-219
w = 0.0013 m/sec = 0.13 cm/sec.
If w = F B, then the force required is
F = 2.68 x 10~9 dynes.
We calculated, using Natanson's image force equations, that F is
-13
4.1 x 10 dynes. These forces, producing supposedly identical effi-
ciencies are unreasonably different. A summary of the results appear
in Table 10.
Natanson's efficiency equation (cube root form) for the charged particle/
uncharged cylinder situation yields forces, and consequently migration
velocities, which are far lower than those calculated for the same over-
all efficiency in the charged particle/charged cylinder case (Coulomb
forces). It is this discrepancy that leads us to believe that Natanson's
equation for the charged particle/uncharged cylinder is in error for
the given conditions. It is further believed that the image forces
have been shown to be insignificant, in that they are too small to
cause any effective capture of particles in question.
63
-------
Table 10. CALCULATED THEORETICAL EFFICIENCIES WITH
SEVERAL COLLECTION MECHANISMS
Theory and equation utilized
Charged particle and uncharged
cylinder Natanson Equation
Charged particle and uncharged
cylinder Deutsch Equation
Charged particle and charged
cylinder Natanson Equation
Charged particle and charged
cylinder Deutsch Equation
Efficiency
7»
66
0.0002
66a
66a
Force
dynes
4.1 x lO-iS
4.1 x 10"13
b
2.5 x 10"9
b
2.7 x 10"9
Migration
sec
2.05 x lO-5
2.05 x 10'5
0.12b
0.13b
Assumed.
Calculated from efficiency
Experimental - The envelope of conditions within which submicron parti-
cles are significantly iemoved is to be measured directly when possible.
It is, however, anticipated that most of the electrical effects within
the bed will have to be determined by inference, since direct measurement
will normally not be possible.
Measurements will be taken of: particle resistivity, pad resistivity,
air velocity, pad thickness, charge level on particles, dust loading,
and particle size.
Utilizing information obtained in the first part of the study, an
experimental rig will be sized for about 500 cfm (0.24 m3/s) of gas
at temperatures typical of both hot- and cold-side ESPs. Aerosol fly
ash will be introduced with a plasma torch equipped with a solids feeder,
64
-------
with which particle size range is to be controlled at levels simulating
loadings and particle size distributions typical of gases downstream
from an ESP. Simulation of typical boiler off gas composition will be
done synthetically via the addition of SCL and CCL. See Figure 7 for
a schematic of the apparatus. The parameters to be explored include the
following (quoted from a document, Battelle Northwest Laboratories (BNWL)):1
"Particle size - This will not be considered a prime variable.
All tests will emphasize the removal of sub-
micron particles in the 0.1-to-l.O-micron-size
range as the larger particles are removed with
reasonable effectiveness by existing elec-
trostatic precipitators.
Air velocity - Varied from 50 to 500 feet per minute.
Pad resistivity - Will emphasize the use of pad materials of high
resistivity which can tolerate temperatures
greater than 240 F. At present, teflon is the
prime candidate, but at least one other material
will probably be investigated.
Pad thickness - Will load on 6-, 12-, 18-inch beds.
Particle resistixity - Coals will be examined that exhibit ash resist-
9 13
ivities in the range of 10 to 10 ohm-cm.
o
Ash loading - Will be varied from 5 to 50 mg/m . Ash on pad
will be calculated from gas sample data.
Charge level on the
particles - Will obtain saturation charge in existing
charger. Mobility spectra will be measured,
but charge level will not be a parameter to be
studied.
SC>2 level - SC>2 concentration in the gas will be varied
from 500 to 3000 ppm independent of the sulfur
content of the coal types investigated."
65
-------
AIR IN
VELOCITY
PROBE
SAMPLERS ELECTRIC
MASS RESISTIVITY
PLASMA TORCH
AEROSOL
GENERATOR
tt
/
CHARGER
MI
p
ll" '
jfa \jl
c
b
L
SO;
CO-
SAMPLE
TO EXHAUST
ORAIN
Figure 7. Schematic of experimental setup for study of removal of charged fly
ash by fiber beds
-------
"It is planned that for each experiment the following data will be
obtained
Removal efficiency as a function of time
Dust resistivity
Air flow rate
Total charge level of the fly ash
Gas composition"
Experiments will also be conducted to determine pressure drop and removal
efficiency as a function of the pad dust loading.
Results
It is expected that a process for economically removing submicron par-
ticles will be made available for plant designers. The pilot study is
expected to lead to a feasible approach for removing submicron particles
from power plant stack gases, A back-up device for use with presently
installed ESPs will be designed to help meet more stringent emission
regulations and to help control the more difficult to precipitate, high
resistivity fly ash which is generated from low sulfur coal.
Experimental results are available from the progress report (called
there "Appendix A"), and from a previous privately funded study per-
formed for Intalco Aluminum Corp0 Results from the Intalco study led
to the discovery that the loosely woven polymer fiber bed was an
effective collector of charged fine particles. Extremely high effi-
ciencies were determined for two different fine particulates at three
different face velocities; see Figure 8 for results.
Results from the present study (those in its Appendix A) indicate that
18 test runs have been completed in which, bed velocity, dust loading,
67
-------
1 - 1 I I 1 I I
© Cryolite Fume
E= Fractional Removal Eff
100 230
VELOCITY, FEET PER MINUTE
1000
Figure 8,
Electrostatic capture of particles by polypropylene
fiber bed10
68
-------
bed thickness, and bed chemical composition were variables. Results
are shown graphically in Figures 9 and 10, and are tabulated in Tables
11, 12, and 13. It can be seen that bed efficiency is a function of:
bed velocity, with a maximum occurring at 0.76 m/s (150 fpro); aerosol
loading to a small extent; the bed thickness; and the chemical composi-
tion of the bed.
Conclusions
To date it has been concluded that it is likely that there were losses
of initially deposited solids from the 3-inch beds, particularly at
higher velocities. The possibility that a threshold velocity exists
where shear forces predominate adhesive forces will be further evaluated.
Charge effects are concluded to be very important as determined via the
difference in collection efficiency between the stainless steel bed and
the polypropylene bed. It is believed that image forces are the only
significant contributor to increased deposition of charged particles
with the more highly conductive stainless steel bed. The efficiency
of the stainless steel bed is "18 percent, which is approximately that
anticipated for removal by image forces developed in conducting fibers
by the charged participates."
Evaluation
Suitability of Goals - The goals of this research are particularly per-
tinent to present and future requirements for particulate removal equip-
ment. Increasing attention toward collection of fine particles (< 3 ^m
diameter), due to increasing awareness of the more harmful health
aspects associated with fine particles versus larger particles, has
created a need for more efficient means of removal of this fine par-
ticulate. As presently available means of removal of fine particulate
are high in both initial investment capital and operating costs, the
69
-------
o
*«
u.
u.
LU
Q
LU
CO
100
90
80
70
60
50
40
20
1 I I
NOMINAL
CONCENTRATION
A 6 MG/M3
O 25 MG/M3 _
D 60 MG/M3
I
0 50 100 150 2CO 250 300 350 400 450
Figure 9. Aerosol removal by a 6-inch polypropylene
bed1
O
UJ
CO
) 50 100 150 200 250 300 350 400 450
VELOCITY THROUGH BED, FT/MIN
Figure 10. Aerosol removal by a 3-inch polypropylene
bed1
70
-------
Table 11. AEROSOL DEPOSITION IN A 6-INCH POLYPROPYLENE BED
Bed velocity
50 ft/min
50
50
150
150
150
350
350
350
Dust
concentration
9 mg/m
26
56
7
23
53
10
28
74
Overall
efficiency
90.8%
97.9
95.1
99.3
91.8
85.6
67.3
61.7
62.8
Bed efficiency
79.7%
85.5
70.6
98.7
87.0
77.8
51.4
38.5
35.5
Table 12. AEROSOL DEPOSITION IN A 3-INCH POLYPROPYLENE BED1
Bed
velocity
50 ft/min
50
150
150
350
350
Dust
concentration
14 mg/m
30
10
21
6
28
Overall
efficiency
78.6%
82.9
76.3
80
24.3
37.9
Bed
efficiency
5 %
17.7
36.7
48
11
10.4
AP bed
0.01" H20
0.01"
0.11"
0.20"
0.33"
0.33"
Table 13. AEROSOL DEPOSITION IN A 6-INCH STAINLESS STEEL BED
Bed velocity
50 ft/min
350
350
Dust
concentration
14 mg/m
7
70
Overall
efficiency
85.2%
42
47
Bed efficiency
18.6%
0
0
71
-------
goals of producing a fine particle collector of low pressure drop and
moderate cost, as a back-up on existing facilities, is certainly suit-
able for today's needs.
The alternate goal of producing a particulate control device which effi-
ciently removes high resistivity fly ash is also very timely. Recent
evaluations of the U.S. energy requirements have projected increased
reliance on coal as a fuel for fossil-fueled power plants. Demands for
more electricity, coupled with increasingly stringent environmental
regulation at power plants, and high oil prices, will force these fossil-
fueled plants to burn lower sulfur content coals. These lower sulfur
content coals are normally associated with the sub-bituminous and
lignite grades of coal of the West which suffer from high ash and low
Btu content. One obvious effect of burning high-ash, low-Btu coal is the
increased particulate generated per Btu fired; while a more subtle effect
is the generation of a high resistivity fly ash which is not efficiently
removed in present cold-side precipitators. In view of the projected
future demand for low-sulfur coal, the goal of producing a back-up col-
lection device which efficiently removes high resistivity fly ash is
extremely important.
Suitability of methods for these goals - Next is discussed the suitability
of the theoretical and experimental approach as proposed. Although a
mathematical model is mentioned as part of the study, it has not been
completed, and will therefore not appear in this section.
Analysis of theoretical approach - The majority of the experimental work
to date has been with the polypropylene fiber beds, primarily the 6-inch
thick fiber bed. The observed high efficiencies may be reasonably ex-
plained if we consider the polypropylene fiber bed to have a net negative
charge, thereby exerting a Coulombic attraction on the positively charged
particles.
72
-------
Choosing the highest observed efficiency , 98.7 percent, for the 6-inch
thick polypropylene fiber bed, we can calculate the particle migration
velocity, for a 0.22 micron particle, required to achieve this high
efficiency. Solving for migration velocity, w, in the equation:
E = i - e-w A/Q
we get:
0.987 = 1 - e-wl76/0'66
0.013 = e'266'7 w
-4.34 = -266.7 w
w = 0.0163 m/sec = 1.63 cm/sec
since w = F B,
F = 3.26 x 10~8 dynes
To determine the strength of the electric field at the surface of the
polypropylene fiber required to generate the above calculated force, we
solve the equation:
I = F/qp
_9
which, for q = 9.6 x 10 stat-coulotnbs (saturation charge), is
E = 3.4 statvolts/cm = 1.02 kV/cm
The electric field strength at the surface of the polymer fiber corres-
r\
ponds to a surface charge density of 3.4 stat-coulombs/cm It seems,
therefore, that if a charge density of 3.4 statcoulombs/cm' could be
generated on the surface of the polypropylene fibers, the efficiency
of removal of 0.22 micron particles under the aforementioned conditions
73
-------
would then be the observed efficiency of 98.7 percent. We must now
consider a mechanism which is capable of generating a surface charge
2
of 3.4 statcoulombs/cm , without any external application of electrical
energy. Triboelectrification is one such mechanism. Charging of
dielectric fiber filters by an air stream was observed by Van Orman
and Endres to be the principal collection mechanisms of filters com-
posed of highly insulating polymers. Charge densities for polymeric
2 2
materials quoted in the literature are: 1-10 statcoulombs/cm , 1-20
f\ *j O /
statcoulombs/cm , and 1.2-8.4 statcoulombs/cm . (Both references 2
and 4 are referring specifically to charge densities produced by
triboelectrification.) These charge densities would be sufficient.
Triboelectrification of polymers has been recognized by the plastics
industry, classically due to its nuisance effect. Plastics processing
techniques often produce inherent static charges due to triboelectrifica-
tion. Friction during calendering and contact during molding give rise
to charge transfers between polymer materials and process equipment.
The finished plastic product attracts dust due to its inherent static
charge, degrading its appearance, or requiring frequent cleaning which
often leads to greater charging. Charges built up during processing can
lead to sparking which may pose a serious safety hazard. It is mostly
for these reasons that interest in plastic triboelectrification exists.
Plastics, being highly resistive materials, appear to have the ability
to hold charges of either sign at close proximity, without neutralizing
11 4 5
each other. ' ' Therefore, it is possible for a polymer surface to be
highly charged, while exhibiting little or no net charge. This ability
of polymers to hold different sign charges in close proximity makes it
fiber beds includes tests with a 3000 cfm capacity control device as
a polymer surface.
There is, however, a considerable tendency for polymers to show a bias
towards a net overall charge of one sign. Numerous triboclectric series
74
-------
have appeared throughout the literature, which series contain poly-
mers. ' ' ' ' In all of the series containing Teflon, it was listed
as the last material on the negative end, with polyethylene occurring
just ahead of Teflon. This seems to indicate that these polymers, and
polypropylene, since it has a very similar chemical structure, have a
high propensity to acquire an overall negative charge (they are electro-
9
negative) due to triboelectrification. A similar statement by Frederick,
"those that are quite electronegative like the polyolefins, and espe-
cially, 'Teflon,'" indicates that the polyolefin in question, poly-
propylene, is highly electronegative.
In view of the evidence supporting the possibility of triboelectrifica-
tion of polypropylene to the required sign and surface charge density to
explain the observed experimental results, we suggest further investiga-
tion of the charging characteristics of the polymer fiber filter. Also
since "Teflon" exhibits an even higher propensity towards negative charge
acquisition, in theory, it would be interesting to compare the performance
of the "Teflon" fiber bed with the polypropylene fiber bed, for identical
conditions.
Analysis of experimental approach - The experimental work performed to
determine the phenomena causing removal of charged fine particles in
fiber beds includes tests with a 3000 cfm capacity control device as
previously described here in the first section. Testing procedure
involves the generation of a submicron particulate, charging the par-
ticles in a corona charger, determining the resistivity of the particles
and the particle size distribution upstream of the corona charger, samp-
ling the particulate material upstream and downstream of corona charger
and downstream from the fiber bed filter, and measuring overall charge
flux upstream and downstream of the bed.
Table 14a and 14b were formulated to assist in evaluation of the experi-
mental approach. Important parameters which may affect the operation
75
-------
Table 14a. PARAMETERS ASSOCIATED WITH THE STUDY OF ELECTROSTATIC
CAPTURE OF PARTICLES BY FIBER BEDS1
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Re.)
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
up to 3000 cfm (1.4 m3/s)
50-350 fpm (25 - 175 cm/s)
7
K (known)
M (to be measured)
~ ambient
ambient ?
schematic indicates it will be meas-
ured, however no mention was made in
the results or in the text
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
measured with 8-stage Andersen impac-
tor for submicron particles only
NH4C1
M
K
particles are expected to receive a
saturation charge total charge
flux of gas stream will be measured
6-74 mg/m
CHARGING SECTION
Type of charging
Ion current
Electric field
Geometry
corona
12.5 ma
26 kV
parallel plates and wires three in
a line
COLLECTOR
Chemical composition
polypropylene, stainless steel, teflon
76
-------
Table 14a (Continued).
PARAMETERS ASSOCIATED WITH THE STUDY OF
ELECTROSTATIC CAPTURE OF PARTICLES BY
FIBER BEDS1
Parameter
Resistivity
Dielectric constant
Charge
Voltage
Particulate loading
Efficiency
Geometry
Internal configuration
External configuration
CLEANING PROCESS
Method
Effect on efficiency
COMMENTS
STAGE OF DEVELOPMENT
Magnitude, description, method of
measurement or control, etc.
K
K
M
will not be measured, however it
will be determined by inference
from data
M
M
,
6-inch, 3-inch thick beds 4 ft x 2.33
ft; 3.0 x 10-* m fiber diameter;
0.9 porosity
6 ft diameter, 12 ft tall fiberglass
chamber
?
?
Pilot scale apparatus
Table 14b. PARAMETERS ASSOCIATED WITH THE STUDY OF ELECTROSTATIC
CAPTURE OF PARTICLES BY FIBER BEDS L
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Ref)
up to 500 cfm (0.24 m3/s)
50-500 fpm (25 - 250 cm/s)
77
-------
Table 14b (continued),
PARAMETERS ASSOCIATED WITH THE STUDY OF
ELECTROSTATIC CAPTURE OF PARTICLES BY
FIBER BEDS1
Parameter
Magnitude, description, method of
measurement or control, etc.
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
N.A.
M
typical of hot- or cold-side
precipitators
ambient ?
we anticipate lower percent RH than in
typical power plant off-gases due to
combustion of lesser amounts of coal
BNWL indicates RH will be measured
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
submicron particles only will be
generated
particles will be formed in a plasma
torch by burning coal; shape of par-
ticles is unspecified
fly ash from various coals
9 13
10 to 10 ohm
total charge level on fly ash will be
measured
2
5 to 50 mg/ra
CHARGING SECTION
Type of charging
Ion concentration
Electric field
Geometry
corona
parallel plates and wires - three in
a line
78
-------
Table 14b (Continued). PARAMETERS ASSOCIATED WITH THE STUDY OF
ELECTROSTATIC CAPTURE OF PARTICLES BY
FIBER BEDS1
Parameter
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage E
Loading
Efficiency
Geometry
Internal
External
CLEANING PROCESS
Method
Efficiency degradation
COMMENTS
STAGE OF DEVELOPMENT
Magnitude, description, method of
measurement or control, etc.
polypropylene, teflon, stainless steel
high can be determined for each
fiber bed material
K
M
inferred self-induced electric fields
M
M
6-inch, 12-inch, 18-inch bed filters
8 ffc2 or 4 £t2
6 ft diameter; 12 ft tall fiberglass
chamber
liquid spray
?
C02 and S0£ will be added to the gas
stream, downstream from plasma torch.
Addition of S02 downstream from coal
combustion may not generate any 803
therefore conditioning of fly ash will
not be at the same level as might be
expected from the level of S02 in the
gas stream.
proposal and lab scale
KEY: ? = Uncertain, unspecified
N.A. = Not applicable
M = To be measured
K = Known
79
-------
of the collection device are listed under generalized headings for each
of the important aspects of the control device under anticipated operat-
ing conditions.
Most important aspects of the gas have been adequately covered in the
BNWL experimental design with the possible exception of relative humidity.
Although Figure 6 indicates the provision for a measurement, there is no
mention of such measurements in the texts. ' it is anticipated that
relative humidity may be an important factor, especially concerning the
ability of the fiber bed to attract and hold a charge at elevated
temperatures.
Although the measurement of1static pressure was not considered in the
text, it appears obvious from the total flow through the system and the
size of the system that the static pressure will be very close to the
ambient pressure. Since the static pressure is required for the correc-
tion of pitot readings, it would presumably have been measured during
the velocity traverse, if it were found to be significant. Effects of
pressure may have to be considered, however, if this device is ever pro-
posed for installation in a significantly pressurized (positive or
negative) gas stream, since charging levels of particles and the col-
12
lector will be somewhat dependent upon this parameter.
The particle size distribution is measured with 8-stage Andersen Im-
pactors, upstream from the corona charger. It would appear that sampling
downstream from the corona charger would be more likely to yield results
corresponding to the actual particle size distribution of the particles
which the fiber bed sees, providing that the charged particles do not
alter the deposition mechanism in the impactor. Sampling upstream of
the corona charger allows room for doubt about the size distribution of
the particles after passing through the corona charger, where agglomera-
tion may occur. The shape of the particles may be important, especially
concerning maximum surface charge capacity; however, the shape of the
80
-------
particles will not be determined. Measurement of overall charge flux
of the particles both upstream and downstream of the bed along with
overall particle loading and size distribution does make it possible to
determine the approximate charge on particles without knowing their
actual shape.
Charging of the particles is accomplished with a conventional corona
charger, where the voltage is set, and the subsequent current may be
determined. The objective is to obtain a saturation charge on the
particles.
The collector is a variable in this series of tests, where the chemical
composition (and therefore resistivity, and dielectric constant), ex-
ternal geometry, and possibly internal geometry are parameters to be
studied. Most important parameters relating to the collector have been
given or will be measured but some exceptions are: dielectric constant,
which was not mentioned in the text, but which will need to be known to
solve the single fiber efficiency equation appearing in the text; the
voltage at the collector, which will not be measured, but will be inferred
from experimentally observed collection efficiencies; internal geometry
of the bed, which is adequately described for only the stainless steel
bed though it is likely that the fiber diameter and porosity of the two
polymer beds are somewhat different. This missing information will be
required to do a thorough theoretical analysis of the collection
efficiency.
Cleaning of the filter bed has been considered, and tests will be made
to determine the pressure drop across the filter, and consequently
loading at which the filter should be cleaned. It was stated that,
"relatively long duration runs will be made during which the pressure
drop and removal efficiency will be obtained as a function of the pad
dust loading." These runs will provide information about the efficiency
degradation associated with the cleaning of the filter; however, runs
81
-------
spanning actual cleaning cycles will be required to define totally the
problem of efficiency degradation with cleaning.
The proposed follow-up study, in which charged submicron fly ash removal
is to be investigated, is similar to the initial experimental apparatus,
see Figure 7, and the parameters to be investigated are handled simi-
larly, with one significant difference: the study concerning the re-
moval of charged fly ash by fiber beds seeks to determine the suitability
of the proposed collection system for collection of very specific par-
ticulate matter in a very specific environment, offgas from a low-sulfur-
coal-fired boiler. Fly ash from combustion of low sulfur coal usually
has a resistivity above the critical level of 10 ohms/cm, which causes
back corona, and consequent poor precipitator performance. An average
pulverized-coal-fired boiler generates 6 ppm SO- per each percentage
13
point of sulfur in the coal. Low-sulfur and or high-ash-content coal
may require additional SO- to lower their resistivity to within the
"7
acceptable limits, 10 -10-*-" ohms/cm, for electrostatic precipitation.
The proposed addition of SO- to the test apparatus to simulate actual
boiler offgases accounts for the effect that S02 undoubtedly will have
on the particles and possibly the collector. Since the resistivity
of the particles is largely determined by the SO,, concentration in the
off-gases, the particulate produced with the proposed plasma torch
arrangement may have excessively high resistivity, not at all represen-
tative of fly ash generated from the same coal under normal firing con-
ditions. The presence of SO- at natural levels may also have a notice-
able effect upon particle collection mechanisms due to its effect upon
the charge leakage from the filter pad. It appears that to simulate
coal-fired boiler offgases, especially to study electrostatic effects,
the addition of approximately 1/100 SO per SO- is required. The
study apparatus also calls for the addition of representative amounts
of CO^ to help simulate boiler offgases, which appears to be sound.
There is, however, no mention of the addition of water, which normally
32
-------
comprises approximately 10 percent of boiler off-gases. This water
will have a significant effect on suppressing corona current and raising
sparkover voltage. The effects of combined water and SO- may prove to
have some synergistic effect far greater than expected from the more
addition of S02 and H_0 effects.
Applicability to Pollution Control - The applicability to pollution con-
trol of the proposed collection device is considered, with particular
emphasis on the control of emissions from burning low sulfur coal.
Prospects of method - The proposed system of utilizing loosely woven
fiber pads to capture submicron particles at a low pressure drop has
thus shown very promising results. Efficiencies as high as 98.5 percent
were reported during the initial Intalco study and one 98.7 percent
efficiency was reported during the initial phase of the present study.
Corresponding pressure drops appear to be below 1-inch of water, making
this devi.ce truly remarkable when compared to the pressure drop required
to attain similar levels of efficiency with conventional equipment.
Scrubbers, if designed to achieve similar efficiencies, would require a
much higher pressure drop. The pressure drop required for a venturi
scrubber which is 80 percent efficient on removal of 1 micron sized par-
ticles would range from 20 inches to over 50 inches of water,14 and
scrubber pressure drops increase much greater than linearly for increased
efficiency. A precipitator designed to meet these high submicron ef-
ficiencies would result in a rather sizeable construction cost, due
to the high size requirements of the collection surface. The high col-
lection efficiency at low pressure drop of the present system looks very
promising.
Status of the method - The original work done on a pilot plant sized
scrubber of 3000 cfm (1.4 m /s) nominal capacity, attained high coll*
tion efficiencies for submicron particulate consisting of aluminum
83
-------
reduction pot off-gas. In this initial study there were reported effi-
ciencies of greater than 95 percent with submicron particulate of one
very specific type, at the rated throughput capacity of the unit,
3000 cfm (1.4 m /s) . This volume f
of approximately 320 fpm (1.6 m/s).
3000 cfm (1.4 m /s). This volume flow rate corresponds to face velocity
Results of subsequent experimental investigations associated with the
present study, see Tables 11, 12, and 13, indicate lower efficiencies at
the same face velocities utilizing NH.C1 submicron aerosol. Data in-
dicate a maximum collection efficiency occurring at some intermediate
velocity, contrary to the expected inversely proportional relationship
between velocity and efficiency. Also noted is the lack of collection
by the stainless steel fiber bed; it exhibited nearly zero efficiency
under nominal operating conditions, suggesting negligible impaction.
Further investigation of the relationship of face velocity and collection
efficiency is scheduled, with the idea that there is some threshold
velocity at which shear forces dominate over adhesion forces. This may
explain the reason for the observed maximum efficiency at an inter-
mediate velocity.
Work on the second phase of the task directed towards collection of high
resistivity fly ash has not yet started. Work on this phase will likely
not even begin until the initial phase of this study is completed.
Implications - Preliminary data from initial experimental work indicate
that the fiber bed filters are capable of removing submicron particles
from an industrial offgas, very efficiently, at low energy penalties, with
reasonably priced equipment. If the fiber beds prove capable of handling
various resistivity particles under conditions of temperature and gas
compositions typical of industrial offgases, then a significant air
pollution control device of unique capabilities will have been developed.
Methods of cleaning must be developed which do not appreciably impair
efficiency or the promise of this device will go unfulfilled.
84
-------
ELECTROSTATIC EFFECTS IN FABRIC FILTRATION31
The study does not involve designing or evaluating any new specific
? 2
device. A small lab scale fabric filter containing 0.124 m^ (1.3 ft )
of filter area will be used to study the effects of various parameters
upon performance.
Goals of Study
The goals of the study are summarized in these statements quoted from
the proposal, 31 "a comprehensive investigation of electrostatic effects
in particulate filtration ... sufficiently broad to explain static pa-
rameters of filter media and' particulate needed for establishing a
reliable engineering approach to the design of commercial baghouses."
"In addition to the determination of the electrostatic properties of a
variety of fabrics and particulates, the overall program would be ex-
pected to determine the electrostatic role in particle-to-particle and
particle-to-fabric adhesion as well as electrostatic effects in the
spatial relationship or arrangement of particles and fibers as the par-
ticles are deposited."
"The overall object of the project is to provide a clear understanding
of the relationship existing between solid particulates and filter media
as a function of such important parameters as solids entrained gas flow
rate, particulate loading, pressure drop, aerosol chemistry, air-to-
cloth ratio, cloth permeability, fabric construction (including fiber
surface chemistry, fiber type, fiber size, together with yarn and weave
characteristics) and all other defineable fabric and particulate prop-
erties including, especially, the electrostatic properties of media and
of particulates."
85
-------
Methods of Study
The study will rely heavily upon the use of experimental results to
corroborate theory. Both lab and field investigations will be employed
using a small fabric filter.
Theoretical - The theoretical aspects of the study would appear to center
on the idea that electrostatics play a major role in fabric filtration;
however, the roles of various parameters affecting electrostatics are
little understood. Again quoting from the text:
"The following statement (6) summarizes the current status of the infor-
mation available on electrostatic involvement in the filtration process:
'While electrostatics undoubtedly plays a role in the capture and reten-
tion of dust particles by a fabric filter, the evidence is inadequate
to evaluate this mechanism quantitatively. According to Frederick (1961),
electrostatics not only may assist filtration by providing an attractive
force betx^een the dust and fabric, but also may affect particle agglomera-
tion, fabric cleanability, and collection efficiency. He attributes the
generation of charge to frictional effects, staging that the polarity,
charge intensity, and charge dissipation rate of both the dust and filter
media, and their relation to each other can enhance or hinder the filter-
ing process. He cites qualitative differences only. For example, fabric
A may be better than fabric B on dust X, while fabric B is better than A
on dust Y. He gives a "triboelectric" series for a number of filter
fabrics that may be useful as a guide to selecting fabrics with desirable
electrostatic properties. This is a fertile field for further
investigations. "'
The only very specific theory to be investigated is explained in the
following quote from the proposal:
86
-------
"One of the first studies of Part I will concern the theory that charged
particles deposited in a 'nonionized1 electrostatic field will produce
a relatively porous dust layer. By 'nonionizcd': we mean a field having
no molecular ions (i.e. , no corona) . Under these conditions, slight
irregularities tend to concentrate the field. The charged particles
follow the field lines so that particles tend to deposit onto these
irregularities. This further concentrates the field. Thus particles
tend to deposit on top of particles forming a porous deposit. In the
absence of these electrical effects, particles are carried by the gas
into the spaces between particles thus tending to plug any existing
porosity. Preliminary tests indicate that this may be a very large
effect and so this will be one of the first mechanisms to be studied, in
addition to the usual measurements of gas flow and pressure drop, this
will require measurement of the charge on particles and measurement of
the electric field throughout the filter region. Also microscopic
examinations of dust deposits will be made."
Experimental - The experimental work will involve two distinct phases,
which are generally summarized by the following statements:
"Phase I will be a basic study of the various electrostatic mechanisms
which can influence the filtration of dust. Initially the work will be
directed to conditions which produce a dust layer that is permeable to
gas flow but impermeable to particulates. Another important factor is
the adhesion of the particulates to fabrics and the separation of the
collected dust from the fabric during a cleaning cycle. Instrumentation
will also be developed to make the necessary electrostatic measure-
ments in Phase II.
"In Phase II, bag filters will be tested under both laboratory and
industrial conditions. The work under these two phases will be closely
coordinated so that each phase will take advantage of developments in
the other phase."
87
-------
"Phase II operations will involve practical filtration evaluations
carried out in two distinct but complimentary methods. In Phase II-A a
conventional laboratory or bench scale filter system would serve to
relate the filtration characteristics of various media with selected
redispersed dusts. In Phase II-B, a portable filter system similar in
size to that used in Phase II-A would be designed and used to receive
and study gas-entrained particulates at operating plant sites."
Results, Attained or Expected
To date there have been no results reported, as this study is apparently
in its beginning stages. The expected results would obviously corre-
spond to the objectives which have been set forth for this research
work. It is expected that the significant parameters concerning the
particulate, fabric, and operation, will be determined to the extent
that they could be used in a systematic scheme for the design of a bag-
house facility.
Conclusions
We have drawn no conclusions from the limited x%rork to date. The experi-
mental work thus far has not proceeded beyond the initial formulation
and set-up of apparatus.
Evaluation
Suitability of Goals - The goals set forth in the proposal address
legitimate shortcomings of fabric filtration technology. In view of the
growing role of fabric filtration as an air pollution control device, it
is obviously beneficial to generate any technical information which would
contribute to the use of a more scientific approach to baghouse design.
88
-------
The emphasis on the filter cake is quite appropriate, because the buildup
of the cake provides the conditions for high efficiency filtration in
industrial filtration, but is also responsible for much of the pressure
drop, thus power consumption (Wilder an(* Dennis ). If the cake can be
made to perform at high efficiency with lower pressure drop due to the
intelligent application of electrostatics, this could be a major contri-
bution to the technology of fabric filtration.
Suitability of Methods for These Goals - Utilizing available information,
we will analyze the theoretical and experimental aspects of the proposed
study.
Analysis of theoretical approach - Two basic fabric filter performance
parameters will be investigated: filter pressure drop and collection
efficiency. Addressing filter pressure drop, the proposal refers to an
equation used to predict the resistance to flow of a granular bed and
proposes that electrostatics will lead to particle agglomeration, thus
changing the void fraction of the dust layer and decreasing the resis-
tance. An equation describing the pressure drop is:
A P(t) =
P
A
_E
V
P
e
3
where:
k = Kozeny-Carman coefficient , 25/6
2
g = acceleration due to gravity, 980 ctn/s
u = air viscosity, g/cm-s
V = average filtration velocity, cm/s = total flow
rate divided by effective filter area
A
y2 = surface area to volume ratio of the dust particles, cm
P
89
-------
e = porosity (void fraction) of the dust layer, dimensionless
p = true density of the dust, g/cm
P 3
C. = dust loading to the filter, g/cm
t = elapsed time of filter operation at above loading,
2
. = pressure drop, dynes/cm .
If the electric fields have the effect of increasing the porosity only,
then the equation predicts a decreased pressure drop. Table 15 shows
3
values of the porosity factor (1 - e)/e versus the porosity, from which
it is clear that (on a percentage basis) a small change in porosity can
produce a large change in pressure drop. If the electric field also
effectively increased the average size of the particles as deposited
(causing them to agglomerate), the ratio of the surface area to volume
(A /V ) would be expected to decrease, thus further lowering the pressure
drop. Unfortunately, both factors which tend to lower the pressure drop
are expected to lower efficiency, so that whether or not this is a fruit-
ful approach will depend greatly on the details -and magnitudes of these
effects.
Table 15. PRESSURE DROP DEPENDENCE ON POROSITY FACTOR
Poros ity
e
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0.95
0.99
Porosity factor
(1 - e)/e3
25.9
9.38
4.00
1.85
0.875
0.391
0.137
0.058
0.010
90
-------
There are several complicating factors in attempting to predict the
pressure drop across a granular layer. One factor is that the particles
arc not monodispcrscd, and thus smaller particles may lodge between large
particles, causing a high pressure drop. Another factor is that the
particles may be stacked in different arrangements, and thus the granu-
lar structure could be altered by electrostatic forces.
An important factor that should be considered is that the pressure drop
across a fabric filter is not simply a function of a granular deposit
but is also effected by interactions with the fabric.
Since it is not possible to predict the value of the expression
K =
P
p
A
_£
V
L P.
referred to as the specific cake resistance, this value is usually
determined experimentally by measuring C., V, t and Ap. Draemel
conducted an investigation with three dusts and 123 fabrics and reported
the following results:
"K values (specific cake resistance) with a given dust are
dependent on the structure of the underlying fabric. The
deep channel-like pores, formed by more rounder yarns, can
lead to significant deposition of dust under velocity con-
ditions of an order of magnitude or more greater than the
average face velocity of the fabric. Deposition at local
increased velocity would tend to increase dust packing den-
sity and thus increase K."
"(Dusts subject to cake collapse phenomena imply pressure
and/or velocity dependence on dust packing density.) Very
shallow pores and a smooth fabric surface with no project-
ing fibers can be very efficient in particle retention but
lead to a completely unsupported dust layer which has a
characteristically high K value and is subject to cake
collapse as pressure increase. Projecting fibers appear
to support a more porous dust cake (lower K. values), less
subject to cake collapse. The dense projecting fibers
-------
found with napped fabrics may tend to produce nonlinear
Ap versus t response, indicacing a deviation from the cake
lav/ type of filtration behavior normally seen with a
woven fabric. K values with a given dust may vary con-
siderably as a function of fabric even though efficiency
remains relatively constant for the same dust/fabric
comb inat ions. "16
If the effect on filter pressure drop of electrostatics is to be deter-
mined, then this effect must be separated from the effects caused by any
fabric variations, or else the effect of fabric variations must be demon-
strated to be an indirect effect acting through electrostatic forces.
A second objective of the program appears to be to determine the effects
of electrostatic forces on 'fabric filter efficiency. There have been a
good number of investigations of the effects of electrostatics or
single fiber efficiencies. Generally the collection efficiency of a
single fiber has been shown to improve under the influence of electro-
static forces. However, fabric filters operate at much higher efficiency
and the reasons for particle penetration may differ considerably from
those involved in single fiber experiments. GCA and other investigators
have found that a large part of the emissions from a fabric filter may
occur during the cleaning process or immediately thereafter and that
variations in pulse jet or mechanical shake cleaning can cause large
changes in filter efficiency. Particle penetration appears to be a
combination of seepage (successive reentrainment), direct penetration,
and dust that is loosened during cleaning. Electrostatics should af-
fect seepage and direct penetration, although the magnitude of this
effect on a high efficiency fabric filter has not been demonstrated.
If electrostatic forces are used to decrease penetration during clean-
ing then quite likely the cleanability of the filter would suffer.
I Q 1 fi
Figure 11 is an analysis by Dennis of data presented by Draemel.
This figure shows a single fabric-dust combination, Dacron-flyash, and
the effect of free area (a function of yarn size, weave, average pore
size) on the outlet concentration. Again as with filter pressure drop,
92
-------
VO
u>
O.OOI
0.01
FREE AREA
Figure 11.
Outlet loading versus free area. Woven Dacron nylon bags, fly ash
filtration at 3 grains/ft3 and 3 fpm18
-------
the question arises as to whether the result is a direct effect of
fabric properties and cleaning parameters or an indirect effect acting
through electrostatic forces or some combination of direct and indirect
effects.
The above discussion is intended to point out some of the problems and
pitfalls that may be encountered in an investigation of electrostatics
and fabric filter performance. Electrostatics and fabric filtration is
an area with large data deficiencies that should be investigated.
Whether the proposed study shows that electrostatics are or are not an
important factor, the results will be useful in that they should provide
a functional understanding of the factors affecting fabric filter
performance.
Analysis of experimental approach The proposed experimental work to be
o 1
performed was lacking as to specifics of how various parameters will be
measured or controlled. Table 16 contains all of the important parameters
which pertain to the study, for each of the components and concepts. It
appears that the parameters concerning the gas which are of importance in
this study are either controlled or will be measured.
The same is not true for the particles however, where important para-
meter such as size and resistivity are not mentioned in the proposal. It
may be that the apparently missing parameters will be determined as a
matter of routine (they may well be known in advance) and were thus not
mentioned in the text. Whatever the circumstances these parameters should
be covered as they may be important concerning electrostatic effects.
Charging of the particles will be that charge that is naturally acquired
via redispersing dust. There will be no direct application of electrical
energy involved, as in corona charging, therefore the majority of the
parameters do not apply.
94
-------
Table 16. PARAMETERS ASSOCIATED WITH THE STUDY OF ELECTROSTATIC
EFFECTS IN FABRIC FILTRATION31
Parameter
Magnitude, description, method of meas-
urement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Ref)
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
controlled - will be varied
controlled - will be varied
? (unspecified)
N.A. (not applicable)
measured continuously and recorded
controlled air stream which may be
treated to simulate industrial
off-gases
ambient
controlled
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
various types of particles will be
used
probably to be measured
probably to be measured
assumed to contain natural charges of
different polarity - will be measured
as total charge of a specific volume
of gas at known concentration
uniform predetermined loadings will be
employed
CHARGING SECTION
Type of charging
Ions
Electric field
Geometry
only the charge formed naturally during
the industrial generation or laboratory
redispersal of particles
N.A.
N.A.
N.A.
95
-------
Table 16 (continued). PARAMETERS ASSOCIATED WITH TIE STUDY OF
ELECTROSTATIC EFFECTS IN FABRIC FILTRATION
31
Parameter
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage, electric field
Particulate loading
Geometry
Internal configuration
External configuration
Collection efficiency
CLEANING PROCESS
Method
Effect on efficiency
COMMENTS
STAGE OF DEVELOPMENT
Magnitude, description, method of
measurement or control, etc.
various bag materials will be used and
the composition will be known
not mentioned specifically; however, it
seems that the properties of the filter
media will be known for any selected bag
filter media.
not mentioned specifically; however,
it seems that the properties of the
filter media will be known for any
selected bag filter media
inferred from field
electric field due to particles will be
measured - none will be otherwise induced
in the bag
M (to be measured)
3-inch diameter bag with 1-1/3 square
feet of surface area - 31 inches high
M
shaking, pulse jet,
reverse jet
?
effect of cleaning on Ap to be
investigated
proposal for laboratory scale study
All of the parameters concerning the collector will be looked at in
detail. Not mentioned specifically are the fabric resistivity and
dielectric constant; however, it is likely to be determined under the
96
-------
statement, "all other properties including especially the electrostatic
properties of media and of the participates."
Conventional methods of cleaning fabric filters were discussed as was
the effect of cleaning on the pressure drop. Both of these aspects
are to be adequately covered; however, efficiency degradation with
cleaning was not mentioned as a parameter to be studied. The effi-
ciency degradation with cleaning may be important.
Applicability to Pollution Control - The objectives of this proposed
study, to determine the electrostatic contribution to fabric filtration,
would obviously be of great interest to designers of commercial baghouse
facilities. Since current design of fabric filter installations relies
heavily upon past experience rather than laboratory data, formulation
of a more systematic scientific approach to baghouse design would be
an important tool for the designer.
AMBIENT FIELDS ACROSS FILTER MEDIA
The basic source for the material to follow is an evaluation and sum-
mary made by Midwest Research Institute (MRT),19 based upon the liter-
9n
ature and upon a document by Rao et al.
Goals
The MRI study had as its goal "to evaluate the use of electric fields
in fabric filters as a means of controlling fine particulate emissions
from industrial sources." Systems MRI investigated involved combining
with fabric filtration the following:
external fields
internal fields
electrets
97
-------
Methods
20
Theoretical - MRI summarizes one external field study: "Rao et al.,
extended Zebel's theory by including the effect of the closeness of
fibers on the deposition of charged particles by the use of a three
cylinder model. Rao et alc assumed potential flow in their model and
corrected the velocity and electrostatic potentials by the method of
images when the distances between the cylinders is small." The theory
predicted decreased deposition on fibers as porosity decreased, which
is in agreement with experimental data, for example those of linoya and
27
Makino. MRI noted that most filtration theories do not yet take into
account most of the factors known from experience to be important:
"The air-to-cloth ratio, cleaning mechanism, temperature, humidity,
weave pattern, fabric weight, gas flow rate and filter fabric 'surface1
characteristics appear to be the most important engineering parameters."
MRI also discussed the theoretical work of Ziekman with respect to
electrets. Ziekman calculated the electric field in the vicinity of a
cylinder in a square lattice array, using the field due to the cylin-
drical dipole and those of its eight nearest neighbors. His flow mod-
el was of the Kuwabara-Happel type. Efficiencies were calculated using
computer modelling of trajectories by Ziekman, who found, as expected,
higher efficiencies for highly charged particles and for low Reynolds
numbers. MRI emphasized correctly that collection of charged particles
would rapidly reduce the electret field and regeneration of the
electrets would be difficult.
9 9
Experimental - MRI referenced work by Walkenhorst, L discussed sepa-
23 >
rately here, and by Kirsch. Kirsch used monodisperse aerosols and
deliberately kept the filter loading very low. Figure 12, from the
MRI report, shows the improvement measured by Kirsch, penetration
decreasing as field intensity increased and as flow velocity decreased,
24 25
as expected. MRI cited work by Dennis and by Silverman on commerci
filters having electrostatic augmentation and concluded; "The general
98
-------
\f>
VO
Z
o
»
^r
PENETR/
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-
CT^- D
Q^
^^^
^^^^
PARTICLE DIAMETER-0 I2um
FIELD INTENSITY RANGE - IO.O- I0.9kv/cm
O-NO FIELD
D-FIELD
L 1 1 1 1 I
10 15 20
AEROSOL VELOCITY Cm/s
25
30
Figure 12. Performance of "real" filter in the absence and presence of
external electric field (MRI)1^
-------
experiences reported in these references are (1) electrification improves
collection efficiency for very light loadings of submicron aerosols; and
(2) penetration of fine aerosols is relatively high (40 to 50 percent),
depending on the flow velocities and porosity of the filter media."
As for electrets, Ziekman ^ also carried out experimental work with
electret fibers about 23 ym in diameter for 0.7 ym dioctylphthalate
aerosol particles and found, according to the MRI report, that penetra-
tion was initially 1 percent for filters made from electret fibers
and about 80 to 90 percent for filters made from regular fibers. Unfor-
tunately, penetration increases rapidly with loading as the electrets
have their fields cancelled by the collected particles, as shown in
Figure 15 from the MRI report (after Ziekman). MRI estimates that only
3 3
0.2 ft of gas having a grain loading of 1 gr/ft could pass through'
2
1 ft of area of such filter material before it needed regeneration
(this means 6 cm of gas flow path length cleaned for a concentration of
2.3 g/m ). Although, as Davies notes, electret behavior has been
extremely useful in the development of personnel respirator filters,
the problem of regeneration seems overwhelming for their use as
industrial source control devices.
Results
Figures 12 and 13 give the results from the experimental work of
23 21
Kirsch and Ziekman. They have been described above.
Conclusions
External electric fields can increase the collection efficiency of
filter media, as seen in Figure 12. The effect becomes less as porosity
20
decreases, according to theory by ^ao et al. The use of electret
material for fibers has the serious disadvantage that the captured par-
ticulate material, if charged, will deposit so as to cancel the electret
fields, leading to a severe degradation of collection efficiency.
100
-------
10'
rl
c
o
4»
o
10-2
10'
r3
10'
,-2
i i i
10'
rl
icr
Aerosol Load, mg
I I t I I i
I t I I I I
10'
102
Figure 13* Penetration load curve for electret filter (MRI)
14
-------
Evaluation
The goal of augmcntating filtration efficiencies electrostatically is
suitable, clearly. The theory developed by Rao et al. of Notre Dame
shows what experiments confirm: increased porosity increases the single
fiber collection efficiency for electrified filter materials. Theories
which are to handle practical problems in industrial filtration should
deal with filters which are heavily loaded as well as with cleaning
cycles and fabric surface characteristics. Porous filter media using
electrostatics have bright prospects because of their lower pressure
drop but there must be a way for these filters to be cleaned to prevent
blockage and/or to maintain electrical fields.
Summary
A limited evaluation of the above research has been presented here. The
Rao et al. theory seems to explain the effect of porosity on single
27
fiber efficiency, as noted by Iinoya and Makino in their experiments.
External fields, internal fields, and electret fields all have similar
possibilities and problems: the hope of efficiency collection at lower
pressure drop and the problem of removing the collected material or
the collected charge or both.
INSULATED WIRE FILTER BED
22
This work was reported by Walkenhorst, whose description follow^:
"The construction of the filter is shown in Figure [14 ]. A frame measur-
2
ing 5 x 5 cm inside and 0.5 mm thick carries two windings of wire,
insulated with varnish; the diameter of the wire is 0.08 mm and the
spacing is 0.5 mm. ... the surface of the wire is rendered water
repellent. The windings are indicated in the upper diagram of Figure [14 ],
one of them being shown with broken lines. 600 V is applied between
the windings and the polarity is reversed periodically. Very good
insulation is necessary to avoid leakage at high relative humidities."
102
-------
"To make the filter, 10 of the frames were mounted in scries with a
distance of 0.5 mm between each; this made the horizontal and vertical
distances the same between all the wires. Each layer was opposite its
neighbor. In the lower part of Figure [14], to make this clear, the posi-
tive and negative polarities, at a given moment, are shown by solid and
22
open circles, respectively."
Goals of the Study
This study was performed to determine the effect of inhomogeneous
electric fields on the capture of particulate. Efficiencies were
determined for the wire filter arrangement previously described, meas-
ured under varying electrical states to try to determine the best possi-
ble removal efficiency.
V4470a
Figure 14. Diagram of filter construction26
103
-------
Methods of Study
The study was performed by Walkenhorst in two parts, in which both
theoretical and experimental analyses were utilized.
Theoretical - The theoretical method of study was closely linked with
experimental results. Theories concerning the anticipated electric
field strengths and direction were individually checked with a simple
model to be described. Figure 15 illustrates the theoretical electric
fields anticipated with the indicated arrangement of wires and collected
charged particulate. The theories were then substantiated utilizing a
simple experimental technique, the results of which are in good agree-
ment xvith the theory.
b)
c)
a
o
f)
0 0
Figure 15. Changes in the electric field between a pair of insulated
wires due to the deposition of charged particles (equal
numbers of each polarity)22
104
-------
Figure 15 illustrates the theoretical approach behind the first series
of tests utilizing the simple model illustrated in Figure 16. Case (a)
demonstrates the electric field generated between two insulated wires,
across which an electric potential is applied. Since the wires have
curved surfaces, and the distance between the wires is small, an in-
homogeneous electric field is the result. In an inhomogeneous electric
field, particle collection can occur in two ways, via coulomb attraction
with charged particles, and via induced dipole attraction on charged
and/or uncharged particles.
Cover
0 8 sieve
JO'
Wires
Electric -field
Membrone filter
Figure 16. Experimental apparatus for studying the effect of an
electric field on the trajectories of dust particles
105
-------
Case (b) illustrates the complete neutralization of the electric field
which could occur due to the deposition of charged particles of a par-
ticulate which contains no net charge, in other words an equal amount
of positive and negative charged.
Case (c) illustrates the field resulting from removing the applied
potential across the wires in case (b). The charged particles would
now create a field equal and opposite in direction to the field in
case (a). If we were now to apply a potential across the wires in the
opposite direction as was previously applied we would have the result
of doubling the strength of the field in case (c) corresponding to
case (d) .
If charged particles are again collected by the wires, they would even-
tually have the effect of neutralizing the previously oppositely
charged particulate, at which point we would duplicate the field in
case (c), illustrated in case (e). However, the source of the field
in case (e) would be the potential applied across the wires, not the
charge on the collected particles.
Finally, if more particles were collected by the wires, we would again
neutralize the field generated by the applied potential and we would
return to the case where the case of complete field neutralization,
case (f), which corresponds to the original case (b). It was then
postulated that an insulated wire filter could be operated under the
aforementioned principles to remove naturally generated charged and/or
uncharged aerosols, with parameters concerning applied potential, geom-
etry, particle charges, and potential reversal frequency to be deter-
mined by experiment under given conditions.
It was recognized that the majority of naturally generated aerosols
display a tendency towards an overall net charge and that the simplistic
model discussed in Figure 15 would not as such strictly apply since at
no time would the field be entirely neutralized due to the excess of
106
-------
charge of one sign over another. This docs not however alter the basic
mechanism from which such a filter would operate, only the voltage
reversal frequency will require readjustment.
Experimental The experimental approach consisted of two distinct phases.
The first phase of the experimental work consisted of attempts to
corroborate the theory put forth concerning electric field configura-
tions anticipated with the wire arrangement shown in Figure 15. An
experimental apparatus was constructed, shown schematically in Figure 16,
which tested the field strength qualitatively via a photographic tech-
nique, described as follows:
"A suitable experimental method has been described previously (Walken-
horst, 1962). The present setup is shown in Figure [16]. The wire, or
several wires which are insulated from one another and between which
the electric field is established, are placed across a tube of 20-cm
diameter. At a distance of 2.5 mm below the wires there is a membrane
filter on which dust is collected in the same distribution of concen-
tration as it has after passing the wire. A distance of 2.5 mm suffices
to visualize undisturbed parallel flow, 2.5 mm being 25 times the wire
diameter of 0.1 mm. A 'picture1 of the obstacle in the flow is obtained
on the filter which shows clearly what is going on and can be evaluated
quantitatively. To prevent coarse dust from falling into the tube the
upper end is covered above an entry zone and there is a wire gauze with
0.8 mm opening on top of the tube. This helps by preventing uncontrolled
air currents in the tube and ensuring laminar flow. Using coal dust
and a white membrane filter a visible 'picture' is obtained."22
The second phase of the experimental work consisted of the construction
of the filter previously described, shown schematically in Figure 16.
This filter was then tested for particulate collection efficiency while
varying parameters of gas velocity, relative humidity, field strength
and duration, and field reversal frequency. Details concerning the
107
-------
actual filter arc given above. The experimental procedure is as follows
in this quote from the text:
"The finished filter was mounccd in a holder through which air could be
blown. Some experiments were doen with a finely powdered bore dust
(Ruhr sandstone, 95 percent < 5 pirn) . A tyndallometer was used to indi-
cate concentration. In most experiments the dust was generated by a
Wright apparatus, coal and quartz dust were used. To adjust the rela-
tive humidity an atomizer is used which may rise it up to saturation.
A rotameter was used to measure the air flow through the filter, and the
pressure drop across it was measured within 0.1 mm of water with an
inclined manometer.
In order to determine the deposition, samples of air were taken with
thermal precipitators up and .downstream of the filter. Particle size
and number were counted with a light microscope. The fractional
22
deposition could thus be ascertained down to 0.5 ym."
Results
The results of the initial experimental investigations with the appa-
ratus in Figure 16 were a series of photographs illustrating bands of
various widths associated with the distance surrounding the collecting
wire pair in which particles were captured. The appearance of a wide
fringe around the thin wire indicates that there were far fewer par-
ticles collected in that light area. The width of the fringe well
beyond the wire width is a measure of the effectiveness of the electric
field versus only the wire as an obstacle causing deposition. Thus by
varying the field strength, duration, and direction, the theories pre-
viously discussed were corroborated.
The results of the second phase of the experimental work, that with the
actual .filter, were given in graphical form, illustrating the variations
108
-------
in filter efficiency with the parameters which were experimentally
varied. Figures 17, 18, 19, 20, and 21 follow as they appeared in the
original text. Figure 17 verifies the previous theoretical considerations,
displaying the deposition efficiency with a tyndallometer under the
different operating conditions listed. As expected, the very high
porosity filter is not effective when used without any applied voltage.
Following the operation of the filter without voltage for 10 minutes,
the approximate theoretical maximum voltage of 600 volts was applied to
the filter with the polarity reversed every 10 seconds to simulate the
best possible collection conditions. Under these conditions the effi-
ciency was measured at up to 97 percent for a time of 60 minutes; again
the filter operated more efficiently with the applied voltage, as expected.
The filter was again operated with no applied voltage, relying on the
charge of the previously collected particles to generate some inter-
mediate strength field, uhich would now mean efficiencies somewhere
between the initial run without applied voltage and the subsequent run
with applied voltage. As can be seen this is again the case, as the
I
efficiency varied between 20 and 30 percent deposition. However, the
increasing efficiency with time would not be the expected result; the
existing field should be slowly neutralized as more particles are col-
lected. The 600 volts are again applied across Ihe fibers without
reversing polarity, resulting in the overall high collection efficiency
of about 95 percent for 24 minutes. The effect of running without
polarity or field reversal would be a deposition of particles on the
filter, strictly charge separated for the entire 24 minute run. This
should have resulted in a strong residual electric field, due to the
build up of charge strictly by sign. This appears to be the case when
the external voltage is removed and the efficiency decreases from
83 percent to 70 percent in 15 minutes, as would be expected.
A significant parameter concerning the operation of their filter is the
magnitude of the applied potential. A theoretical calculation, per-
formed in the text, predicted that the maximum applied voltage before
109
-------
100
9C
80
70
60
50
40
30
20
1C
"
-
"-
"
\Vithouf
- volroge
S ^
~ \A
' ! ' '
;
o
v
£
O
M
t
u
c
s
J
S<
£
c
' £
O
U- O
Of
o*
o
o
>
^
o
JC
>
y
!
1
§
ID
^ Without voitoge
^Xx° Q.
^ -0
Filter ' 5>5cm2
10 double layers of wires 008mm4
Distance oport : 05mm
Flow velocity lOcm/sec
Relative humidity <50-45%
Temperature 27-32°C
&p ^OZmmWS.
Drill dust <5^im
Dust concentration from.
90 to 115 mg/m1
(Tyndollometer)
i ' I 1 1 1 I 1 1 1 1 I I I
0 2 -i o 8 10 0 2
0 2 4 6 6 10 12 I-. 16 IS 2022 2-1 262S
Time, mm
Figure 17. Deposition efficiency of a filter under different
operational conditions^
Filter
10 double lo,?rs of wires
Oistonce opart
Flow veloc.ty
Relative humidity
Terperotu'e
Ap
Drill dust
Oust concentration fiom
100 to 250 mg/m1
Tyndoiiome:er)
800
1000
Voltage, reversal every 10 rev
Figure 18. Dependence of deposition efficiency on applied voltage
22
110
-------
100
90
GO
70
60
50
40
30-
20
10
Coal dust
(I) x lOcm/sec RH 34%
(2) lOcm/scc RH95V.
(3) o 10 cm/sec RH 95% field not reversed
(4) & 80cm/sec RH 45%
I
I
I
10 20 30 40
Particle diameter, ^i
50
60
Figure 19. Experimental results with coal dust
22
100
90
e:
7C
6C
so
40
JO
20
10
ObO'U dust
(I) » 13cm/sec RH 43V.
(2) 20 cm/sec RH 57-707.
(3) o 80 cm/sec RH 50-70%
J_
I
i
I
10 20 30 40
Porticle diomefer, /i.m
60
Figure 20. Experimental results with quartz dust
22
111
-------
3lOmq/m^ 5GOnng/m^ GI7mq/m^
I l i i
3 Or- f>
/
Fibrous filter /
1
0 iCO 163200 300335 400 503 MS
Dust, g/m2
i i '
40
mm
50
Figure 21. The increase of pressure drop at constant airflow in
relation to the amount of dust on the filter 22
corona onset was 577 volts. Figure 18, percentate deposition versus
applied voltage, shows a leveling of the curve at about 600 volts, after
which only a slight increase in deposition is observed. This is in
excellent agreement with the theory, especially considering the allow-
able tolerances in the construction of the filter, and the likely
resulting irregularities in the geometry and therefore electric field.
Figures 19 and 20 show the results of a series of tests run under the
stated conditions, with the experimental approach previously discussed,
for two different types of dust, coal and quartz. Variations in face
velocity and relative humidity for coal and quartz dust yielded expected
results of decreasing deposition with increasing face velocity and
relative humidity. These curves of percentage deposition versus
particle diameter display the usual decrease in deposition with de-
creasing particle size, however the dependence is not as pronounced
as usual for this type of curve except for the highest velocity
(80 cm/sec) case. Although the percent relative humidity adversely
affects the percent deposition, the effect appears to be relatively small.
112
-------
There are several aspects of a filter's operation which are important
in determining its performance. The most important of these are the
filter's collection efficiency and pressure drop. For most industrial
applications the loading capability and cleanability are also very im-
portant. Figure 21 illustrates pressure drop, at a constant airflow,
versus filter dust loading for the electrified insulated fiber filter
and a fibrous filter. While both filters display the expected rise in
pressure drop versus dust loading, the conventional type fibrous filter
shows a dramatically higher rate of increase of pressure drop with dust
loading. This indicates that the filter under study would be highly
advantageous from an energy per unit volume of gas filtered basis.
Conclusions
The inhomogeneous electric field generated between the wires in the
filter caused effective participate removal in the filter. Operating
the filter without any applied voltage resulted in efficiencies at or
slightly greater than 10 percent, while an applied voltage raised the
efficiency to as high as 97 percent.
It is necessary to reverse the field (current in the wires) periodically,
to achieve the maximum removal efficiency, as can be noted by comparison
of case (2) and (3) in Figure 19. It may also be noted by looking at
case (1) and (2) in Figure 19, that relative humidity adversely affects
particle removal efficiency. Figure 18 illustrates the proportional
relationship between removal efficiency and applied voltage which reached
a virtual maximum between 500 to 600 volts. Higher voltages did not
result in increased particulate removal efficiency, in agreement with
the theoretical corona onset potential of 577 volts.
The pressure drop for a given flow velocity was found to be very low,
(see Figure 21), and increased slowly with filter loading. A conventional
fiber filter displays a much larger increase in pressure drop with
increased filter dust loading, giving the new filter an advantage.
113
-------
Evaluation
Suitability of Goals - The practical goals of this study were to develop
a fiber filter which would efficienctly remove particulate matter via an
electrostatic capture mechanism. The filter was to be an efficient par-
ticle trap due mainly to the electrical capture mechanisms as opposed
to mechanical capture mechanisms normally associated with filtration.
This allows for the construction of a low pressure drop filter. Such
a low pressure drop filter would not require large energy expenditures
to force the gas through the filter, as is normally the case in filtration.
Fine particle control was not considered separately in the Walkenhorst
study; however, it is implied that the filter in question would be more
efficient capturing fine particles than a high pressure drop analog, if
the filter were sufficiently deep to create a similar pressure drop to
a conventional filter. In other words, increasing the size of the filter
to the point where the low pressure drop advantage is lost would result
in an efficient remover of particulate of all sizes when compared to a
similar conventional filter.
The goal of developing a more energy-efficient particulate removal fil-
ter is obviously important, but there is a second aspect which will re-
quire study: cleaning and reusing such filters. The goals of Walkenhorst
were limited in that this aspect was not studied in depth.
Suitability of Methods for These Goals - The methodology utilized to
achieve the goals of this study will be analyzed, with particular em-
phasis upon the theoretical aspects of the study. (Detailed information
concerning the experimental aspects were not published and the study was
completed some years ago, making it of limited value to suggest improve-
ments in experimental techniques).
Analysis of Theoretical Approach - The theoretical analysis, as previously
outlined, concerning the generation of inhomogeneous electric fields and
114
-------
the use of these fields to capture particles appears to be sound, qual-
itatively. A quantitative theoretical analysis of the filter efficiency
due to electrical forces was neglected. There was, however, an analysis
of the field strength and field gradient for assumed conditions, which
analysis bears upon the theoretical capacity of the type of filter in
question.
The filter in question is expected to remove particulate from an air
stream via two electrical mechanisms. The first mechanisms, thought to
be the most important for uncharged or only slightly naturally charged
aerosols, is the deposition of particles due to induced dipole attraction
in an inhomogeneous electric field. The second mechanism, which could
be important if sufficient natural charge exists on the particulate, is
capture by coulombic attraction. Let us look more closely at the ex-
pected efficiency of the proposed filter for each of the two mechanisms.
If we examine the induced dipole mechanism, we can determine the
electrical force (F ) on a 1 micron diameter particle using the equation:
Fr = 2 X_ V E grad E
E i; p
where V = particle volume
E = electric field and
assuming a spherical particle,
e,, - 1
K
which for a conductive particle is simply
* = 37T/8
115
-------
because the dielectric constant, e , tends to infinity. E is the average
electric field, which in this case is the average of E and E . or
max mm
2.32 x 104 V/cm, and grad E is 10.26 x 105 V/cm2. These become 77.3 stat
volt/cm and 3.42 x 1CP stat volt/cm2. Both values are quoted from the
22
text for the 0.1 mm diameter wire and 0.5 mm space between the two
wires.
(We were not able to verify the value for the average gradient due to
09
an apparent error in equation (3) of the Walkenhorst text for E . ,
min
which generated a negative number in the natural log group, making it
impossible to solve.) Solving the equation for the force acting on an
uncharged 1 ^m diameter particle in an inhomogeneous field we get:
F_ = 2 X_ V E grad E
r. r. p
F = 2 (37T/8) (47T/3) (0.5 x 10"4)3 (77.4) (3.42 x 103)
F = 3.27 x 10"7 dynes
which corresponds to a migration velocity:
w = FB
w = 2.22 cm/s.
In order to calculate the coulomb attraction on each size particle, it
is necessary to make some assumptions concerning the naturally occurring
charge on the particles. We have chosen the number of elemental charges
on each particle by using the values quoted by Walkenhorst in the text,
and extrapolating to obtain estimates for the 0.25 vm and 1.0 urn parti-
cles. In this way we had hoped to be consistent in our calculations with
what Walkenhorst apparently expected the particulate charge distribution
116
-------
to be like. The actual values are tabulated in Table 17. (It must be
noted, however, that our own lab experience with the Wright apparatus
for rcdispersing dusts leads us to expect much more highly charged
particles than indicated in Table 18. In view of the lack of any measure-
ment of the charge on the particles in the study, we cannot make any
accurate estimate of the charge concentration of the particles.)
Table 17. THEORETICAL AND EXPERIMENTAL EFFICIENCY FOR COAL
DUST AND QUARTZ DUST AT 10 cm/s FACE VELOCITY
Particle
diameter
(microns)
0.25
1.00
2.00
3.00
4.00
Assumed
number of
elementary
charges
1
2
10
200
400
Theoretical
efficiency
percent
' 90.05
96.70
99.99
100.00
100.00
Observed
efficiency
(coal dust)
@ 34% R.H.
97
98
100
100
100
Observed
efficiency
(quartz dust)
@ 43% R.H.
92
96
99
100
100
Table 18. THEORETICAL AND EXPERIMENTAL EFFICIENCY FOR COAL
DUST AND QUARTZ DUST AT 80 cm/s FACE VELOCITY
Particle
diameter
(microns)
0.25
1.00
2.00
3.00
4.00
Assumed
number of
elementary
charges
1
2
10
200
400
Theoretical
efficiency
percent
25.10
34.80
77.70
99.50
99.98
Observed
efficiency
(coal dust)
@ 34% R.H.
35
78
86
90
92
Observed
efficiency
(quartz dust)
@ 50-70% R.H.
56
70
91
99
100
117
-------
If we assume a net charge on the 1.0 micron diameter particles to be
equal to two elementary charges per particle, we can then solve for the
Coulomb force on this particle. The coulomb force is given in the
equation:
F = qE
F = (9.6 x 10 stat coulombs) (77.4 stat volts/cm)
8
F = 7.A3 x 10 dynes
and the migration velocity is:
w = FB
w = (7.43 x 10"8) (6.8 x 106)
w = 0.52 cm/s.
The overall efficiency can be approximated from the sum of the forces,
and thus migration velocities previously determined, using the equation;
E = 1 - exp ( - Wfc A/Q)
we can solve for the efficiency (using MKS units):
E = 1 - exp ( - (2.2 + 0.5) (10~2) (3.17 x 10~2) / 2.5 x 10~4)
E = 1 - exp ( - 3.4236)
E = 0.967.
118
-------
This value Cor the efficiency of the filter is in excellent agreement
wjth the experimental results. It was stated that "efficiency reached
values up to 97 percent" for the previously stated conditions, also it
can be seen in Figure 19, for case (1), that the collection efficiency
at 1 micron is approximately 98 percent.
Table 17 contains the results of calculated theoretical efficiencies ex-
pected from the combination of induced dipole and Coulomb forces, and
the corresponding experimental results obtained from Figures 19 and 20,
case (1). It can be seen that the theory agrees very well with the
experimental results; however, the values for efficiency taken from the
curves are only approximations. Because we are looking at a very narrow
portion of the curve, these approximate values are difficult to obtain
with much accuracy. Table 18 is a comparison of theoretically calculated
efficiency for coal and quartz dust at a face velocity of 80 cm/s, and
the experimentally observed results for the tests at 80 cm/s. The as-
sumed charge on the particles corresponds to those values given in Table 17,
It should be noted that the predicted efficiency for the 0.25 micron
diameter particle (Tables 17 and 18) is lower than the observed experi-
mental efficiency, especially with the coal dust. This may be due to
the misplacement of the points on the curve which correspond to the
0.5 micron diameter particles, since in the text referring to measurement
capability, it was stated that "fractional deposition could thus be ascer-
tained down to 0.5 micron." Aside from this inconsistency, the theoreti-
cal efficiencies correspond well to the observed results.
Another possible source of conflicting results, especially at the lower
particle size range, may have been the somewhat arbitrary value for the
overall net charge which we assigned to the particles to determine the
Coulomb force. It was assumed that the overall net charge on each par-
ticle would correspond to the values and ranges quoted in the text. It
may well be, however, that these values are generally too low for the
particulate used in the experimental procedure. The coal and quartz dust
119
-------
utilized in the experiment were redispcrsed aerosols which are typically
highly charged. Because the smaller particles tend to have only a small
amount of net charge (we used one elemental charge per each 0.25 micron
particle and two elemental charges per 1.0 micron particle), the error
in the predicted efficiencies would tend to be very great if the net
charge were off by only one or two elemental charges.
Examination of Table 18, which corresponds to the conditions of Figures 19
and 20, case (4), reveals a more distinct tendency for the experimental
results to be higher than the predicted theoretical results. This is
as would be expected if the net charge assigned for each particle were
indeed low, or if some collection mechanisms have incorrectly been
assumed to be negligible.
It is important to note that in all of our theoretical calculations, we
have been dealing with expressions which require the use of the equiva-
lent aerodynamic particle diameter. While never clearly stated in the
text, it would appear from the use of the optical microscope that the
author is dealing with the optically measured particle diameter. Again,
since we are dealing with small disagreements between theory and results,
and since small differences in particle diamete^ have a large effect
upon the theoretically calculated force, migration velocity, and ulti-
mately efficiency, the discrepancy between theory and experimental re-
sults is within limits placed by observational errors.
Analysis of Experimental Approach As mentioned, the analysis of the
experimental approach will of necessity be brief. The first phase of
the experimental work dealt with use of the apparatus displayed in
Figure 16 to verify the proposed theoretical collection mechanisms.
Results of these tests were previously covered, and may be summarized
by stating that the theory is in very good agreement with the experi-
mental results. The functioning of the test apparatus and the experi-
mental procedures have little bearing upon the objectives of this evalu-
ation, and thus will not be discussed further. Details of the
120
-------
cxpcrinicntnl apparatus and procedure were not given in this article, and
were referenced so that they could be studied by any interested persons.
The experimental techniques utilized were sound for their chosen use,
with some reservations pertaining to details concerning the number of
particles counted; however, without more detailed information we must
assume the techniques were properly applied. The use of thermal pre-
cipitators, which are somewhat particle size dependent, may be question-
able considering the altering of the particle size distribution expected
after passage through the filter. Also, it has been our experience at
GCA that the Wright apparatus produces a highly charged particulate,
which may have seriously affected the collection on the charged filter.
The charge on the particles was not measured and may have thus been
underestimated.
Applicability to Pollution Control - The electrified insulated wire
filter appears to have potential application as a pollution control
device.
Prospects of the Method - The positive aspects of this device include
its apparently high efficiency at very low pressure drop and the ability
to function efficiently at high humidity. Possible negative aspects of
the device deal with its capacity to handle efficiently aerosols of
higher grain loadings. The experimental results were done at a rela-
tively low, but not uncommon, grain loading of approximately 0.25 grains
3
per cubic foot (approximately 0.5 g/m ).
Status of the Method - The filter appears to be capable of efficiently
removing particles at a high face velocity, when compared to a fabric
filter, namely 10 to 20 cm/s (20 to 40 fpm) versus 1 to 1.5 cm/s (2 to
3 fpm). The filter also appears to be capable of accepting a somewhat
higher loading than some fabric filters. The higher face velocity and
loading capabilities would indicate that a smaller sized unit would be
121
-------
capable of handling a similar capacity of offgas, compared to conven-
tional fabric filters. This would translate into an initial capital
cost advantage due to the physically smaller facility, neglecting the
probable cost difference of filter media. For a given grain loading
and gas throughput, a smaller filter capacity would require more fre-
quent cleaning. It is in the question of filter cleaning that the
greatest potential problem for the insulated fiber filter arises. It
is difficult to assess the filter cleanability since no effort was
directed towards this goal; however, it is conceivable that increased
costs associated with filter cleaning requirements could offset the
previously mentioned potential savings.
Implications - If the filter'could be demonstrated to operate effi-
ciently at higher grain loadings, and is capable of being easily
cleaned without serious efficiency degradation, then the device has
many potential applications to pollution control.
RELATED STUDIES
Filter Electric Fields; Applied and Intrinsic
27
Recently, the Japanese scientists linoya and Mal-.ino (1974) published
a summary of their theoretical and experimental work concerning the
following:
Collection due to the natural charge existing on a fiber.
9 Collection on conductive fibers with applied voltage.
Collection on dielectric fibers with applied voltage.
Summary - In general they found higher efficiency could be achieved at
lower pressure drops with electrified filters than with nonelectrified
filters. The work was done with relatively light filter loadings, well
below those for which a filter cake is formed. No work on cleaning was
reported.
122
-------
NaLurally charged filters - Having shown earlier Chat collection by
gjass fiber filters was nearly equivalent to collection'by synthetic
28
fiber filters treated with anti-static coatings, linoya and Kimura
used the collection efficiency of the glass fiber filters as the ref-
erence point from which they measured the natural charge densities of
synthetic filters. For cotton, nylon, teflon, and some other fibers
they measured "characteristic numbers" ~ 10 . The definition of the
characteristic number, k , is
z
k = Pe Z/Z
z G
where Pe is the Peclet number, (length)(velocity)/(diffusion coefficient),
and Z is the ratio of the minimum experimental collection efficiency for
a single fiber to the interception regime theoretical value; the sub-
script G is for glass. Their assumption is that the differences in dis-
agreement with theory are due to a known factor (Pe) and the electrical
"characteristic number." Values - 10 indicate these fibers are captur-
ing material much more efficiently than comparable glass fibers. liaoya
7 ft
and Kimura used the formula for the capacitance (charge per unit voltage)
of an isolated fiber to convert the q that they infer from the k into a
Z
"natural electrostatic potential." Increasing natural electrostatic po-
tentials were found for the series: vinyl, cotton, nylon, teflon.
Electrically conductive fibers Two different graphite-packed fibrous
filters were constructed and tested by linoya and Makino,^' a single-
stage type and a double-stage type. (See Figure 22.) They were tested
for collection efficiency as a function of applied voltage, and the data
were presented in terms of the migration velocity, w, using the expres-
sion for penetration:
Pn = e- A/Q
123
-------
in which A is the collector surface area and Q is the volume flow rate,
as usual. For the single-stage device, they measured increasing effi-
ciencies (greater migration velocities) with increasing voltages, regard-
less of voltage polarity. For the double-stage device, they measured
greater migration velocities for smaller fiber volume fractions (presum-
ably due to less electrostatic mutual interference), smaller distances
between the two stages (more intense electrical fields), and lower face
velocities, this latter perhaps indicating that other collection mecha-
nisms (diffusion?) were important or that the complete mixing implicit
in the exponential expression did not occur. For single fiber effi-
ciency they calculated from their data the correlation for the increased
efficiency due to voltage V:
An = 1.4 x LIT14 V2/(l - E)3/2 u1/2
where An = increase in efficiency of a single fiber
u = face velocity, m/s
e = void fraction of filter
V = voltage, volts.
The correlation was obtained for particles from 0.8 /wn to 1.4 jxm in
diameter and for fibers 7.0 ^/m and 9.0 fjun. in diameter. For filters
with 99.5 percent open volume, the single fiber efficiency was 30 times
that of a conventional fiber.
124
-------
Gropni
;
Mstoge 7
TtT
i
le fiber
l
1
mesh ]|
filter';
/ "I
n
Aerosol inlet
Insulotmg matenol
v,
i
4
n
.i
ttfftf
X
X
Double
stoge
.
x
X
X
X
X
X
X
X
a
-
v
1
s
Aerosol inlet
V,
7
Figure 22. Methods of applying a d.c. voltage to electrically
conductive fibrous filters^?
Dielectric fibrous filter The models for this type of filtration are
shown in Figure 23. The voltage is zero at the wall. At the front and
rear faces of the filter, metal grids allow the imposition of voltage by
either putting the front face at one voltage and the rear at another
(case I) or putting adjacent wires at alternate voltages. Figure 24
shows the dimensionless electric fields versus dimensionless distance
for both cases. For case I, the field is nearly homogeneous across the
filter, and for case II it is most heterogeneous at the faces and decays
rapidly within the filter. The case II configuration is very similar to
99
that studied by Walkcnhorst which we have described above. linoya and
Makino give an equation based on the polarization of the fibers by the field
that predicts the increase in collection efficiency due to the field,
corresponding to case I. (They did not go into much detail about case II
because they had noted that the collection efficiency would not be ex-
pected to be as great as for case I, but this may be offset by the greater
ease of cleaning for a filter which causes most of the deposit on its face.)
They evaluated an empirical constant for the equation, using data they ob-
tained in the case I model. Again, much higher single particle efficiencies
125
-------
\
Aerosol flow direction
Figure 23- Analytical model of a dielectric fiber mat filter
27
c -
u U
6
0 5
0 «
0 3
O2
0 I
0
O v
Z ^
On center line
0 2 « 6 8 10
Nondimensionol Oistonce from front metol grid,
N (-)
Figure Ik. Effect of charged condition of metal grids on electric
field distribution in a dielectric fiber mat filter
126
-------
are predicted with the addition of electrostatic forces. AC or DC fields
could be used, as they pointed out.
Evaluation
27
The work by linoya and Makino quantified several important collection
enhancement methods employing electrostatics with filtration. If eco-
nomical disposable fiber filters can be developed, then electrical
effects could1 be used to improve the efficiency/pressure drop charac-
teristics of a filter without worrying about the cleanability of such
a filter. It may be possible to use electrostatic effects briefly at
the beginning of a filtering cycle to accelerate the formation of the
filter cake by increasing the "clean" collection efficiency of the
filter. At present, however, such methods seem far from being ready
for commercial application.
127
-------
REFERENCES
1. Study of Electrostatic Capture of Particles by Fiber Beds.
Bottcllc Memorial Institute, Pacific Northwest Laboratories,
P.O. Box 999, Richland, Washington. U.S. Environmental Pro-
tection Agency, Raleigh, North Carolina. December 10, 1973.
2. Harper, W.R. Contact and Frictional Electrification. Oxford,
Clarendon Press, 1967.
3. Murphy, P.V., F.J. Holly, and William Bernhard. Electrets as
Blood Compatible Prosthetic Material. In: Electrets and
Related Electrostatic Charge Storage Phenomena, Baxt, Lawrence
M. Richmond, Virginia and Martin M. Perlman. Saint-Jean Quebec,
The Electrochemical Society, Inc., 1968.
4. Seanor, Donald A. Triboelectrification of Polymers. In: Elec-
trical Properties of Polymers, Dr. K.C. Frisch, University of
Detroit, Detroit, Michigan and Dr. Angelos V. Patsis, State
University College, New Paltz, New York, Technomic Publishing
Co., Inc. p. 37.
5. Woodland, P.C., and E.E. Ziegler. Static Dust Collection On
Plastics. J Modern Plastics. 28:(9), 95-106, 169-178, 1951.
6. Henniker, J. Triboelectricity in Polymers. J Nature. 195:474,
November 3, 1962.
7. Shashoua, Victor E. Static Electricity in Polymers. I. Theory
and Measurement. J Polym Sci. 33:65-85, 1958.
8. Frederick, E.R. How Dust Filter Selection Depends on Electro-
statics. J Chem Eng. June 26, 1961.
9. Frederick, Edward R. Some Effects of Electrostatic Charges In
Fabric Filtration. J Air Pollut Control Assoc. 24:(2), 1164,
December 1974.
10. Postma, A.K., and B.M. Johnson. Electrostatic Scrubbing of Sub-
micron Particles from Aluminum Reduction Pot Offgas. Intalco,
September 1971.
11. Van Orman, W.T., and II.A. Endres. Self-Charging Electrostatic
Air Filters. The American Society of Heating and Ventilating
Engineers Journal (Heating, Piping and Air Conditioning). 157,
January 1952.
128
-------
12. White, H.J. Industrial Electrostatic Precipitation. New York,
Fergammon, 1963.
13. Morris, E.B. Condition Flyash with Synthetic SO.,. J Power.
July 1974.
14. Calvert, S., J. Goldshmid, D. Leith, and D. Mehta. Scrubber Hand-
book, Vol. I. A.P.T., Inc., Riverside, California. Report Numbers
NTIS PB-213016 and EPA-R2-72-1182. U.S. Environmental Protection
Agency, Raleigh, North Carolina. July 1972.
15. Billings, C.E., and J.E. Wilder. Handbook of Fabric Filter Tech-
nology. Volume 1, p. 2-159. Fabric Filter Systems Study. GCA/
Technology Division. Department A, Clearinghouse, U.S. Department
of Commerce, Springfield, Virginia 22151. Report Number GCA-TR-
70-17-G, APTD-0690, Contract CPA-22-69-38, PB-200-648. December
1970.
16. Draemel, D.C. Relationship Between Fabric Structure and Filtra-
tion Performance in Dust Filtration. Control Systems Laboratory,
U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. Report Number EPA-R2-73-288. July 1973.
17. Dennis, R., and J.E. Wilder. Fabric Filter Cleaning Mechanisms
Kinetics Study. GCA Corporation. Contracts EHS-D-71-19 and
68-02-0268. Final Report in Press.
18. Dennis, R. Collection Efficiency as a Function of Particle Size,
Shape and Density: Theory and Experience. J Air Pollut Control
Assoc. 24:1156-1164.
19. Schrag, M.P., and L.J. Shannon. Evaluation of Electric Field
Fabric Filtration. Office of Research and Development, Research
Triangle Park, North Carolina. Contract Number 68-02-1324.
Draft Final Report. March 1974.
20. Rao, K.S., et al. Collection of Dust by Fabric Filtration in an
Electrostatic Field. Department of Mechanical and Aerospace
Engineering, University of Notre Dame. EPA Grant Number
AP-01303-01. 1973.
21. Ziekman, P. Aerosol Filt ation by Electrified Fibrous Filter
Mats, III. Chemisch Laboratorium RVO-TNO, Netherlands. Report
Number 1970-16. 1970.
22. Walkenhorst, W. Reflections and Research on the Filtration of
Dust from Gases with Special Consideration of Electrical Forces.
J Aerosol Sci. 1:225, 1970.
129
-------
23. Kirsch, A.A. The Influence of an External Electric Field on the
Deposition of Aerosols in Fibrous Filters. J Aerosol Sci. 3:25.
24. Dennis, R., ct al. Evaluation of the Electro-PL and Electro-
Klean Dust Collectors. Harvard University. USAEC Report NYO 4614.
July 1958.
25. Silverman, L. et al. Performance of the Model K Electro-Polar
Filter. Harvard University. USAEC Report NYO. July 1954.
26. Davies, C.N. Air Filtration. London and New York, Academic Press,
1973.
27. linoya, K., and K. Makino. Application of Electric Field Effects
to Dust Collection Filters. J Aerosol Sci. 5:357-372, 1974.
28. linoya, K., and N. Kimura. J Chem Eng Jpn. 29:547, 1965.
29. Natanson, G. Deposition of Aerosols by Electrostatic Attraction
Upon a Cylinder Around Which They Are Flowing. Dokl Akad Nauk
(USSR). 112:696-699, 1957.
30. Lundgren, D.A., and K.T. Whitby. Effect of Particle Electrostatic
Charge on Filtration by Fibrous Filters. Ind Eng Chem Process
Des Dev. 4:345. 1965.
31. Penney, G.W., and E.R. Frederick. Electrostatic Effects in Fabric
Filtration. Proposal to the U.S. Environmental Protection Agency,
Washington, D.C. October 1973.
130
-------
SECTION VI
ELECTROSTATIC AUGMENTATION OF SCRUBBERS
OPPOSITELY CHARGED DROPLETS AND PARTICLES
This work is being done under Dr. Pilat of the University of Washington.
The scrubber consists of two spray chambers, the first being counter-
current flow and the second chamber cocurrent. The scrubber is elec-
trostatically augmented by charging of the droplets and the particulates
to opposite polarities using inductive charging and corona charging,
respectively. The scrubber configuration, with aerosol charging chamber,
is shown schematically in Figure 25. The study, to date, has utilized
-2 3
two different sized units, one of 6.61 x 10 m /s (140 cfm) capacity,
3
and one of 0.472 m /s (1000 cfm) capacity, both constructed of 1/4-inch
lucite to allow visual observation of the internal operation of the
scrubber. More details concerning the physical dimensions and con-
struction of the two scrubber units appear in Table 19.
Goals of the Study
This study is being performed to develop an efficient fine particle col-
lection device, suitable for application to industrial sources. It is
under development.
131
-------
U)
NJ
Mist
Eliminator
Outlet
Water
Outlet
Water
Outlet
Aerosol
Generator
Blower
Aerosol
Aging
Chamber
Corona
Charger
Figure 25. Schematic diagram of electrostatic droplet scrubber
-------
Table 19. PARAMETERS ASSOCIATED WITH THE STUDY OF THE
ELECTROSTATIC SPRAY SCRUBUER
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Re )
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
1AO cfm (0.066 mJ/s)
measured
?
known (K)
measured
ambient
ambient
controlled at 100%
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
measured
spherical
DOP
K
K
measured
measured - 0.15 gr/acf (0.34 g/ra )
CHARGING SECTION
Particles
Type of charging
Ion current
Electric field
Geometry
Water Droplets
Type of charging
Ion current
Electric field
Geometry
corona
2.2 mA
27,000 volts of electric potential
rectangular duct
induction
2.2 mA
5 kV power supply
nozzle spray
133
-------
Table 19 (comtinucd). PARAMETERS ASSOCIATED WITH THE STUDY OF THE
ELECTROSTATIC SPRAY SCRUBBER
Parameter
Magnitude, description, method of
measurement or control, etc.
COLLECTOR
Scrubber
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage, E
Efficiency
Internal geometry
External geometry
Water Droplets
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage
Efficiency
Internal geometry
External geometry
lucite
K
K
M
13 spray nozzles, Fogjet 7N4
45 inches high by 20 inches diameter
cylinder, co-current
water droplets
K
K
measured - 5.6 x 10 coul/gm
7
DOP - 30% uncharged, 85% charged
spherical drops
See scrubber
CLEANING PROCESS
Method
Efficiency degradation
N.A.
N.A.
134
-------
Table 19 (continued)
PARAMETERS ASSOCIATED WITH THE STUDY OF THE
ELECTROSTATIC SPRAY SCRUBBER
Parameter
Magnitude, description, method of
measurement or control, etc.
COMMENTS
The water consumption rate is 15.7
gallons/1000 acf (~ 2 liters/m3)
1000 cfm unit had been constructed and
some tests had been run.
STAGE OF DEVELOPMENT
Lab scale unit of 140 acfm constructed
and tested.
Larger 1000 acfm unit had been con-
structed and limited testing had been
done.at time of review.
135
-------
Methods of Study
Theoretical - In the work reported thus far, Pilat et al. used a simple
exponential (Kleinschmidt or Deutsch-type) model to estimate the dif-
ference in collection efficiency due to the addition of charge effects.
Figure 26 is from the paper by Pilat et al., based in turn on the work
2
done by Sparks, in which collection efficiency was gotten from particle
trajectories calculated by numerical integration of the particle equations
of motion, considering diffusion, electrostatics, and particle inertia.
The droplets were assumed to be 200 pm in diameter, moving at 100 cm/s,
and the particles were either uncharged or carried charge equivalent to
that induced by corona charging in an electric field of 1 kV/cm. This
figure indicates that the minimum collection efficiency for the charged
aerosol should be the maximum efficiency for the uncharged aerosol under
the conditions considered, which was confirmed experimentally, but the
measured efficiencies were much less than predicted by the simple model
for both cases, especially for the smaller particle sizes. The model
for the charged aerosol predicts that collection efficiency should in-
crease as particle size decreases, and exactly the opposite was measured.
300
«
I
200-
porticles(-)
and droplets (»)
100r
01
Figure 26. Calculated particle collection efficiencies
for a single 200-p diameter droplet with a
100-cm/scc undisturbed fluid velocity^-
136
-------
Experimental - The experiments involved two different capacity lucitc
scrubbers, the smaller unit having been used first in several preliminary
investigations, the Larger unit being built for subsequent analysis.
The smaller 0.066 m /s (140 cfm) unit was equipped with a corona charger
for the particles, and an inductive charger for the droplets. A fine
aerosol of OOP was generated by injecting it into an electrically heated
aluminum tube, and by exposing it to negative corona charging in the
inlet duct. The particulate was sampled at the inlet and outlet simul-
taneously, with Mark III University of Washington Source Test Cascade
Impactors.
The water droplets were charged positively by induction, and experiments
were run both with and without this charging. The spray nozzles were
Spraying Systems Fogjet 7N4 nozzle tips with an overall water flow of
6 x 10 m /s (1.0 gal/min). The liquid to gas volume ratio was
-433
21 x 10 m /m (15.7 gal/1000 acf). This overall water flow was
varied to determine the most effective rate.
The size distribution of the uater droplets was measured optically with
a Zeiss particle size counter, after collection on glass slides smeared
with petroleum 3elly. (A correction factor of 1.26 was used to correct
the flattened diameter to the actual droplet diameter.) The distribu-
tion was determined for both charged and uncharged droplets.
The overall electrostatic charge on the particulate was measured with
a glass fiber filter held in a glass holder, which collected the par-
ticles or droplets isokinetically, and a charge measuring circuit.
The droplet charge was measured with a droplet collector that was packed
with aluminum shavings and connected to a microammeter. Charge on the
droplets was determined by monitoring the current and sampling time
and weighing the amount of water collected. This yielded charge, per
unit mass for the water.
137
-------
Table 19 contains the majority of the information we considered to be
useful in evaluating an electrically augmented particulate collection
device. Properties and parameters concerning the various components of
the particulate laden gas and the collecting device are listed under
their respective headings.
The majority of the parameters that are important were determined by
2
Sparks et al., either by being known beforehand or by direct measure-
ment. The gas velocity is not stated; however, the references to iso-
kinetic sampling indicate that the velocity was measured. The gas
Reynolds number was not given in the original text, but could likely be
determined with the information available to the researchers.
The parameters concerning the particulate are measured or can be deter-
mined from available information in the literature concerning DOP. All
of the pertinent parameters concerning particle charging are covered,
with the exception of specific dimensions of the duct in which the
charging occurs.
The collector parameters are well covered, with the complication that the
collector in a scrubber is really the water droplets. Table 19 contains
available information for the collector as water droplets, and as the
lucite scrubber unit itself. "Methods of Cleaning" does not really apply
to the scrubber as it does in a precipitation or filtration device.
Results
The charge on the particles was measured at 5.3 x 10 coulombs per gram
in the first scrubber chamber. The droplets were found to contain,
typically, 5.6 x 10 coulombs per gram of water.
138
-------
Figure 27 is a log-probability plot of the spray droplet size distribu-
tion, for both charged and uncharged droplets, at the same nozzle pres-
sure. The geometric mean diameter (or number median diameter) droplet
is approximately 50 microns, with a range from under 20 microns to nearly
150 microns, and a geometric standard deviation - 1.8.
400
"D?
§ 200
1
i \OQ
| 80
5 60
5
$ 40
Q.
jS
20
t i i \ \ \ i \ \ \ \ \
Spraying systems 7M4 nozzle tip
103 psig water pressure
Olgallons /minute
droplets uncharged
^droplets charged
: ^
t»
£»
a
a.
t»
A.
i i 1111
i
-
-
-------
1000
moss conceniroton less fen siotefl diameter
( qr
Figure 28. Size distributions of dioctvlphthalate
aerosol particles at electrostatic drop-
let spray scrubber inlet and outlet^
100
5
£ 80
60-
I 40
20-
ttrrStt
Orople's and
porhcles
chorged
oppos.teiy
/lOOOocf
Mean droc diometer by
numoer = 50 microns
Drop geometric stand dev = 19
2 -JoSiO 2 468 10
PcrllClC fliOmelOr (microns)
Figure 29. Particle collection efficiency of
electrostatic spray droplet scrub-
ber as function of particle size1
140
-------
Conclusions
The particles were not necessarily charged to the theoretical saturation
charge. If we calculate the charge on the observed mass median diameter
particle of 0.4 micron, using the results of the particle charge meas-
_q
uring data, we get a charge of 5.25 x 10 statcoulombs (11 electrons)
per 0.4 micron diameter particle. This is even less than the saturation
charge calculated for a 0.3 micron diameter particle in a 10 kV/cm field,
2.8 x 10~8 stat-coul (958 electrons).
The average charge on the water droplets was calculated in a similar
manner to that used for the particles and x^as found to be 1.1 x 10
stat-coul (2.3 x 10J electrons) per 50 yra diameter droplet.
The particle size distribution is altered after passing through the
scrubber, due to the dependence of collection efficiency on particle size.
The particle size distribution was measured with an impactor, the per-
formance of which may be affected by the charge on the particles. The
charge on the particles has not normally been considered to interfere
with impactor performance, and we have assumed that the results are valid,
but we have some reservations concerning electrostatic effects. The
electrostatic spray droplet scrubber was found to have significantly
greater particle collection efficiency, especially at the lower particle
sizes, when operated with particles charged versus uncharged. From this
we could conclude that the electrostatic augmentation does increase par-
ticle collection efficiency; we also conclude that the electrostatic
spray scrubber is a more efficient particle collector per unit of
energy expended than another type of scrubber, as we shall discuss.
The power requirements have been determined for the electrostatic spray
scrubber. The scrubber consumes energy due to: pressure drop across the
scrubber, pressurizing the water for spraying, and charging the aerosol
141
-------
and droplets. The total energy demand of the system is 1.27 kW/(m /s)
(0.80 lip/1000 cfm) , of which nearly 70 percent is for pressurizing the
water for spraying. This level of power consumption is considerably
lower than that calculated for a conventional Vcnturi scrubber of
3
similar efficiency. Using values for pressure drop given by Calvert for
two Venturi scrubbers with f values of 0.25 and 0.5, we calculated a
o o
power consumption 39.4 kW/(m /s) and 9.8 kW/(ra /s) respectively (24.9 and
6.2 hp/1000 acfm). Even this calculation is conservative, since the con-
ventional scrubbers chosen are only 50 percent efficient at 0.4 urn diameter
particle collection versus over 90 percent efficiency for these particles
with the electrostatic spray scrubber. The overall energy consumption is
extremely low for the stated efficiency.
Evaluation
Suitability of Goals - Efficient fine particle collection is becoming
increasingly important. Many industries with existing particulate con-
trols are not effectively collecting the fine particles, which are usually
the most objectional, considering health. As this study is seeking
to produce a device that is generally applicable to industrial par-
ticulate control, this goal is very appropriate. Scrubbers can be
used on liquid aerosols (where filters are not suitable) and on high-
resistivity materials (where electrostatic precipitators may not
work).
Suitability of Methods for These Goals - The suitability of both the
experimental and theoretical approach utilized in this study will be
analyzed, utilizing the available information at this time.
Analysis of theoretical approach - Pilat cited theoretical calculations
2
done by Sparks, who solved numerically the particle equations of motion
for charged and uncharged particles. Figure 26 is taken from Pilat's
paper, and it shows the theoretical single droplet collection efficiency
142
-------
for a 200 ym droplet and the particle size range shown. The flow
velocity was assumed 100 cm/s and both charged and uncharged particles.
The Sparks calculations were done for charge levels of 6.6 x 10~ coul/g
on the droplets, an order of magnitude larger than the charge noted in
the experiment by Pilat.
In general the measured increase in the scrubber overall particle collec-
tion efficiency due to electrostatic charges agrees with the trend of the
theoretically calculated single droplet collection efficiencies shown in
1 4
Figure 26. Pilat's analysis and that of George and Poehlein follow.
For a 1 pm diameter particle, theoretical calculations indicate target
efficiency of 0.01 for the uncharged condition and 1.6 for the charged
condition. Pilat applied the equation for efficiency:
E = 1 - e-
u , 4 HL
where f = and
j Kb
H = the distance the droplets travel with respect to the gas, cm
L/G = the liquid to gas volume ratio
R = the droplet radius, cm
o
Pilat assumed H was 1 ft, L/G was 15.7 gal/1000 ft and R was 25 ym.
This gave overall particle collection efficiency increasing from 17.4 per-
cent for uncharged particles to near 100 percent when charged, which dif-
fers from the measured results of 80 percent for uncharged particles and
97 percent for charged particles.
4
George and Poehlein's analysis was done as follows. Trajectories of
spherical particles approaching a spherical collector were solved by
numerical methods for various collection mechanisms. The target
efficiency,
143
-------
n =
4 Y
lim
D
is che ratio of the area containing all captured particles to the
cross-sectional area of the collectors. Y, , is the greatest stream-
lim to
line offset distance for which the particle trajectory intersects the
collector surface.
geometry and coordinates of the two sphere system.
D is the collector diameter. Figure 30 shows the
'LUIO ST8C1MLINE5
o UCCTROSTiTlC COULCCTIOK
COLLCCT'ON «T INTERCC'IPOK
--O INEflTIM. IUP4CTION
Figure 30. Geometry and coordinates of the
two sphere system
The equation of motion is expressed in vector form as
»>» dv
F + F. + F =m
g e s dt
144
-------
where1 F = the gravity force
O
F = the electrostatic force
e
F = the fluid resistance force
the assumptions for this system are sticking efficiency of 1.0, par-
ticle and collector are conductors, and d « D .
P c
The target efficiency, n, versus inertial impaction parameter ty is graph-
ically presented at various electrostatic parameters (ES) in Figure 31,
2
u p_ v_ a_
where
C p V d
pep
18 p D
ES =
3ir 2 e y V D (D + d )2
o o p c p
0.001 0.01
Figure 31. Single particle collection efficiency
inertial and electrostatic effects
145
-------
A similar computation of the n, ij> relation done by Nielsen, does not
wholly agree with that of George and Poehlein. Figure 33 is the com-
parison of the two computer results.
IDOp
- 10
- L
0 1
10.0 100
Figure 32. Collection efficiency in potential flow as
function of ij> for various Kg, computed by
Nielsen (solid lines) and by George
(dashed lines)
Note that
3ir
u V D d
c o c p
is nearly identical to ES, and Nielsen indicated that George and Poehlein
actually used 1C, rather than ES.
Although Nielsen's results do not agree with those of George and
Poehlein, there are feu substantial discrepancies between them. The
work done by Nielsen seems to have been the carrying out of the goals
of George and Poehlein in somewhat more detail. Both will clearly
be somewhat incorrect for Reynolds flow numbers greater than about
10 because of the formation of eddies in the wake of the droplets,
146
-------
giving a very different flow profile in the lee of the drop than that
used by either set of calculations. This is a general problem: for
Reynolds numbers of interest, between 1 and 100, neither viscous flow
nor potential flow really suits, and almost all theoretical work has
assumed, for understandable reasons, that one or both of these flow
models is appropriate. This problem of flow model makes Nielsen's im-
provements on the work of George and Poehlein less significant than
they seem at first.
MRI's evaluation was based on George and Poehlein"s report; together
with Pilat's experimental conditions, it is used as the groundwork of
the following analysis.
In Figure 31, n = 1.0 represents target efficiency when inertial forces
are large. Pilat's measurement of overall particle collection efficiency
in Figure 29 shows E approaches 95 percent when particle size increases.
From the exponential relation
we can express
i (1 - E)
f =
-n
For n =? 1, E =? 0.95 and f = 3.0.
This value differs greatly from that calculated from f =
j RG
assuming H = 1 ft, L/G =15.7 gal/1000 ft, and R = 25 urn, which yields
f = 37. The difference may arise from other parameters in the experimental
system.
147
-------
The overall collection efficiency at different electrostatic parameters
versus /i(j calcul cited from n in Figure 31 and in Table 20 are graphically
shown in Figures 33, 34, and 35. We have predicted efficiencies using
3.0 rather than 37 for f for the unknown factors of this specific system.
We have used /i|> because it is almost proportional to particle diameter.
The electrostatic effect becomes distinct as particle size decreases.
The electrostatic parameter, ES, George and Poehlein introduced is a
function of electrostatic charges, particle size, and velocity of the
bulk gas stream toward the collector. It plays an important role in
affecting target efficiency for particle diameter less than 5 urn.
Using the operating parameters from Pilat's experiment, we calculate the
ES and d relation at different V . For CGS system of units
p o J
4 C Q Q
ES = Ljc
3iryd (D + d )2
p c p'
where d = aerosol particle diameter (cm)
P
-4
D = collector diameter (cm). (50 x 10 cm is
assumed in Pilat's report for water droplet.)
C = Cunningham correction factor 1 + 0.17 x 10 /d
P
Q = electrostatic charge on water droplet (Pilat's
report gives 5.6 x 10"? coul/g)
Qc = 5.6 x 10"7 x | TT (25 x 10"4)3 x 3 x 109 stat-coul
Q = electrostatic charge on aerosol particle
91 I 7 9 "2
«P -«' + r > +
-------
Table 20. EFFICIENCIES CALCULATED FOR
JNERTIAL PARAMETERS3
VARIOUS ELECTROSTATIC AND
Inertial
parameter
*
0.001
0.01
0.04
0.05
0.09
0.1
0.18
0.19
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1.0
1.5
1.7
2.0
3.0
5.0
7.0
10.0
vfy
0.032
0.1
0.2
0.224
0.3
0.316
0.426
0.436
0.45
0.548
0.63
0.707
0.77
0.837
0.89
1.0
1.225
1.3
1.414
1.732
2.236
2.645
3.162
ES = 0
n
0.1
0.13
0.23
0.33
0.4
0.46
0.5
0.56
0 . 65
0.75
0.79
0.85
0.9
E %
0
0
0
0
0
0
25.92
32.29
49.84
62.84
69.88
74.84
77.69
81.36
85.77
89.46
90.65
92.19
93.28
ES = 0.1
n
0.36
0.35
0.28
0.25
0.3
0.38
0.5
0.55
0.6
0.7
0.8
0.9
E %
66.04
65.01
56.83
52.76
59.34
68.02
77.69
80.80
83.47
87.75
90.93
93.28
ES = 1.0
n
4
4
3.8
3.6
3.4
3.2
2.7
2.4
2.2
2.0
1.9
1.8
1.7
1.5
1.4
1.3
E %
100
100
99.999
99.998
99.996
99.993
99.97
99.93
99.86
99.75
99.665
99.548
99.39
98.89
98.50
97.98
ES = 10
n
50
50
45
44
41
40
34
30
28.5
27
20
15
E %
100
100
100
100
100
100
100
100
100
100
100
100
From E = 1.00 -
e~fn, f = 3.0.
149
-------
999
99.8
99.5
99
UJ
(J
cr
LU
a.
| 98
UJ
o
u.
u.
UJ
z 95
o
I-
o
UJ
_J
O 90
o
80
50
NOTE
-------
99.9
998 -
99.5
uj
o
oc
UJ
a.
98
LJ
O
u.
u.
Ul
o 95
o
UJ
_)
o 90
80
50
NOTE-ES = O.I
0.5
I
INERTIAL PARAMETER, V?
1.5
O.I
0.2
0.5
1.0
UJ
o
z
o
50
10
DC
l-
LJ
z
UJ
Q.
20
50
100
Figure 34. Collection efficiency versus ^ for ES = 0.1
151
-------
>9?999
99999
99 99 -
z
LJ
O
(T
LJ
Q.
z
LJ
99 9
o
UJ
o
o
99.0
90
0001
0.01
o
a.
UJ
Q.
1.0
0.5 I I.S 2 2.5
INERTIAL PARAMETER,^i
3 3.5
10.0
Figure 35. Collection efficiency versus ^ for ES = 1.0, 10.0
152
-------
Figure 36 reveals that ES increases sharply as the particle size becomes
smaller than a certain value, d , which is about 0.2 urn in this case.
c»
We may also solve for f by calculating H for the system that Pilat
5 2
developed. The nozzle pressure was 7.1 x 10 N/ra (103 psi); by apply-
2
ing Ap = 0.5 P v for ideal flow to estimate the droplet velocity v ,
l"''**k'^^V *t^^ / A * V JK *
V0 = J T757^ = 37.7 m/s = 3770 cm/s
The mean gas velocity in chamber 1 was 17.8 cm/s and in chamber 2,
58 cm/s. The droplet settling velocity was about 8 cm/s. The distance
travelled by the droplet with respect to the gas flow has as its upper
limit the particle stopping distance (initial velocity times particle
relaxation time, v r*) plus the product of the settling velocity and the
residence time. The particle relaxation time can be approximated by
(24/55)r, where r is the usual value from Stokes law calculations and
Q
the (24/55) factor comes from Ingebo's work with accelerating droplets
at Reynolds numbers greater than 1. The initial droplet velocity
(ignoring the small correction for mean gas flow rate) gives a stopping
distance of 12.7 cm. The total residence time was 10.53 s, thus the
settling distance was 85.64 cm. Recall the assumed value of 30.5 cm
(1 ft) in the Pilat analysis. Considering the assumptions involved, this
difference in the value assigned to H is not large.
Analysis of experimental approach - The experimental work with the two-
stage lucite scrubbers was done in a similar fashion for the two units,
using 0.006 m /s (140 cfm) and 0.33 m3/s (700 cfm) capacity. The
153
-------
001
0020406081.0 14 1820
PARTICLE DIAMETER, dp,Mm
Figure 36. Electrostatic parameter versus particle size
for several flow velocities from 7.63 cm/s
to 3770 cm/s
154
-------
measurements of the particle and droplet size distributions with the
cascade impactors and the Zeiss particle counter appear to be adequate.
As previously mentioned, there remains open the question of the effect
of particle charge on impactor data.
The use of 1/4-inch lucite (polymethylmethacrylate), a highly insulating
polymer, for the construction of the entire scrubber assembly may have
biased the overall results by collecting charged particles on the dry
plastic previous to the scrubber chamber. It is not unusual for polymer
surfaces to acquire a static charge which remains immobile, which could
have aided in precipitating the aerosol after the inlet sampling port.
Although the effect may be minimal, insufficient data concerning geometry
and size of the inlet ducting and sampling location are available to
rule out this possibility. Dry plastic has been shown to collect aerosol
in localized areas having voltages as high as 10 kV (Stein et al.).
Applicability to Pollution Control - The electrostatic spray scrubbers is
being studied for its eventual use as an industrial air pollution control
device, and as such should have broad applicability, especially where
scrubbers would be suitable if their efficiencies were sufficiently high.
Prospects of method - The measured efficiencies obtained for small par-
ticles were very high, making the prospects of the method look encouraging.
An overall power consumption of 600 watts (0.8 hp) per 1000 cfm (0.47
m /s) is rather low when the fina particle collection efficiency is
considered. The higher than usual water consumption rate of approximately
66 x 10 m (15 gallons) per 28.3 m3 (1000 ft3), is somewhat more than
used by conventional scrubbers, which use approximately 33 x 10~3 m3
3 3
(8.4 gallons) per 28.3 m (1000 ft ). The potential high efficiency for
small particles makes this device attractive, but somewhat high water
consumption, with its resulting treatment costs, would be a disadvantage
in industrial application.
155
-------
Status of method - The scrubber was, at the time of this review, in its
early stages, having been successfully scaled up from a 0.066 m Is
o
(140 cfm) unit to a 0.33 m /s (700 cfm) unit with no noticeable drop in
efficiency. We understand that there are plans for further scaling-up
of the unit and for testing.
Implications - If similar fine particle collection efficiencies can be
achieved with larger units treating industrial offgases, then the
electrostatic spray scrubber would be an attractive air pollution control
alternative. Industries presently utilizing scrubbers, because their
emissions are not amenable to collection with other devices, might find
the electrostatic spray scrubber a cost-effective option that meets in-
creasingly stringent emission standards.
ELECTRICALLY ACCELERATED DROPLETS
The material presented here was derived from progress reports by Lear
8 9
and Krieve and from a paper by Lear, Krieve and Cohen. The "charged
droplet scrubber" (CDS) they are developing is designed to use electro-
static forces to accelerate droplets from a spray and to produce increased
collection efficiency either from electrostatic capture by the droplets
or from electrostatic capture after the droplets have transferred elec-
tric charge to the particles.
Goals
The goal of the work was to investigate the collection mechanisms in-
volved in the charged droplet scrubber and to test the collection ef-
ficiency of a CDS.
Methods of Study
Both theoretical and experimental investigations were carried out.
156
-------
g
Theoretical - Lear et al. presented a concise description of their
theoretical approach in their recent paper, which we quote:
"The model assumed is one in which a relatively large droplet
is introduced into the carrier gas within which a small par-
ticle is at rest. The droplet moves at a drift velocity U
which is assumed constant for purposes of the derivation. As
the droplet moves within the gas, a "wake" flow field is gen-
erated which gives rise to accelerations on the particle, and
which, if sufficiently strong, can sweep the particle out of
the direct path of the droplet.
"As the droplet moves through the gas, it sweeps out a vol-
ume equal to its path length times it projected area. Par-
ticles within this volume x^hich are not swept out by aero-
dynamic forces as the droplet moves along its trajectory are
collected on the droplet by agglomeration.
"Particles within a concentric cylinder of radius S + D may
remain within this cylinder as the droplet passes. If a
particle passes with its center within a distance D of the
droplet surface, it is assumed to have interacted with the
droplet strongly enough to be collected by induced charging.
Particles originally residing x^ithin a concentric cylinder
of radius Z ... will remain in the interaction cylinder.
A particle starting from radius Z will follow a grazing
trajectory ... and this radius defines an interaction
boundary.
"The analysis given in the present work is in terms of a
collection efficiency which is consistent with-common
usage. The basis of its definition is the cross section
of the complete interaction cylinder.
"The portion of this efficiency due to induced charging
depends on an impact parameter defined by
"Induced charging impact parameters were calculated in two
ways ... The dashed line shows values of A for which corona
breakdown will occur at the surface of a spherical particle.
The droplet is assumed to be spherical and charged to the
Rayleigh limit. The surrounding medium is air at standard
conditions. The electric field enhancement is caused by
the induced polarization of the particle.
157
-------
"If the droplet surface charge is at the Rayleigh limit,
then a field perturbation at the surface may cause a
Raylcigh-type or corona breakdown. A quantity of charge
is transferred to the particle, neutralizing the field per-
turbation. The particle charge was calculated, and the
resulting drift velocity of the particle in a field of
5 kv/cm was calculated assuming Stokes law drag. ...
Larger impact parameters result in smaller particle charges,
thus longer drift times.
"Droplet collection efficiencies were obtained by solving
the full equations of motion of a particle in a Stokes flow
field surrounding the droplet. Again, Stokes law drag was
assumed on the particle. The analysis was programmed for
a computer. The collision effectiveness probability was
found to depend on three parameters, physically correspond-
ing to droplet velocity, droplet surface charge, and induced
charging impact parameter."
More details on the theoretical work will be presented in a subsequent
discussion section.
Experimental - Although some work was done to characterize the spray
size and charge parameters, the major focus of the experimental work
was testing the collection efficiency of the device. Table 21 contains
most of the major parameters which we felt were pertinent. The CDS, as
all scrubbers considered, did not fit the format of the table particu-
larly well. The concept of the droplets being the collectors in the
case of a scrubber, and the droplets having been charged rather than
the particles should be kept in mind when reviewing this table. Fig-
Q
ure 37 is from the article by Lear et al., and it describes the experi-
mental apparatus. The scrubbing water flowed through a long insulated
tube to a nozzle which was kept at a potential of about 40 kV. The
spray is propelled from the nozzle toward the walls by electrostatic
forces, achieving velocities ^ 30 m/s. Table 22 is based on the same
publication and gives some operating conditions for the CDS. Photo-
graphic analysis indicated the spray had droplets in the range from 120
to 180 u-tn number modal diameter and 300 to 400 um mass mean diameter.
The number concentration was reported to be about 42/cm with a standard
158
-------
Table 21. PARAMETERS ASSOCIATED WITH THE STUDY OF THE
CHARGED DROPLET SCRUBBER
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Re,.)
Flow geometry
Pressure drop
Temperature
Pressure
Relative humidity
0.47 m3/s (1000 acfm)
1.5 m/s (300 fpm)
Duct with a cross-sectional area of
0.33 m2
Negligible?
24-81 °C
Ambient
Saturated
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
CHARGING SECTION
Type of charging
Ions
Electric field
Geometry
1.8 pm "mean size"
Assumed spherical
Talc
K (known)
K
No charge was placed on the particles
0.002-0.2 g/m3 (0.001-0.01 grains/scf)
Particles are thought to be charged by
corona breakdown at the droplet surface
N.A.
N.A.
N.A.
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
Water
Conductivity of 400-700 umho/cm
80
159
-------
Table 21 (continued). PARAMETERS ASSOCIATED WITH THE STUDY
OF THE CHARGED DROPLET SCRUBBER
Parameter
Magnitude, description, method of
measurement or control, etc.
Charge
Voltage, E field
Particulate loading
Efficiency
Geometry
Internal configuration
External configuration
Theoretically the water droplets are
charged to the Rayleigh or corona limit
5.6 kV/cm between the wall and elec-
trode - applied voltage of 40 kV at
6 milliamps
N.A.
96.4-99.94 percent
8 cm electrode to wall spacing
300-400 utn mass mean diameter,
120-180 |j.m modal diameter
droplets
CLEANING PROCESS
Method
Effect on efficiency
N.A.
N.A.
COMMENTS
Scrubbing water flow of 1.5 liter/min
Wall wash flex; of 4.5 liters/min
STAGE OF DEVELOPMENT
A 0.47 m /s (1000 acfm) model has been
built and tested. A 14 m3/s (30,000
acfm) pilot scale scrubber has been
built
160
-------
HIGH VOLTAGE
ISOLATION TUCING
COLLECTOR PLATE
LEAKAGJ CU?RENT
(-157. OF E!.EC7PODc
CURREtJT)
SCR'JDDED G/S
DISCHATCC
TO AT.MOS=>KJ.RE
CHARGED PROf-LET
SPRAY PATTERN^
FEED WATc? I'UET
(-O.2 GPM/MEIE" OF
ELcCTRCCE L::;CTH)
DC POWER SUPPLY
H30 WATTS/1000 SCFM)
\»
H
1 "
~ V
- $
1
X.
ELECTRODE
+ (40 KV)
\/~~\
..''
-1
' 1
'N
M
8
w
"i
s ^
:i
1
s
r^-.'.-;,-v.^\-.-',->-:'i
fevv^r^^i;-^^?
FEEO THROUGH
INSULATOR
OS x 1C'3
AMP/METER
OF ELECTRODE
GAS HOW
INSULATOR
HOUSING
WATER/DUST
SLURRY
CARRY-OFF
SCPUCCING V/ATER
SLURRY DISCHARGE
TO SETTLING POND
OUST LADEN
GAS FLOW
(-6 FT/SEC)
Figure 37. Charged droplet scrubber'
-------
Table 22. THREE STAGE CDS PERFORMANCE DATA-UNITED SIERRA TALC - 1.8 urn MEAN SIZE
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
Gas
temp.
°C
61
61
61
61
61
21
21
24.3
24.3
23.8
81.5
81.5
Gas
velocity
m/sec
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.22
1.22
2.13
1.22
1.22
Collector
spacing,
m
0.15
0.15
0.15
0.15
0.15
0.10
0.10
0.10
0.10
0.10
0.15
0.15
Spray
nozzle
voltage,
kV
41
41
41
-50
-50.5
42.5
42.5
20
30
30
43
50
Collector
current,
ma
3.0
3.1
3.1
6.0
6.3
6.0
6.0
1.0
2.5
2.3
3.5
4.8
Inlet
loading,
g/m3
0.796
0.795
0.796
0.796
0.796
1.60
1.60
2.69
2.69
0.40
0.704
0.704
Specific
power
W/nT/hr
0.147
0.153
0.153
0.365
0.388
0.470
0.470
0.037
0.139
0.072
0.182
0.294
Specific
water
flow,
-K/rn3
0.093
0.093
0.093
0.093
0.093
0.150
0.150
0.158
0.158
0.088
0.111
0.108
Scrubbing
efficiency,
(percent)
97.6
98.7
99.5
99.7
99.8
99.91
99.93
99.59
99.94
97.63
97.38
96.40
N5
-------
deviation of 30/cm . The efficiency of the three-stage model of the CDS
was tested using a "1.8 urn mean size" talc, 40 percent of which by weight
was less than 2 um diameter. The results as reported by Lear, et al.,^
are shown in Table 22. It is not clear what fraction of this collection
is done by droplet-particle contact and what fraction is done by charge
transfer to the particles and their eventual collection on the plates by
9
electrostatic forces. Lear, et al. reported that the ratio of plate
area to volume flow was significant in affecting efficiency, as was the
ratio of water flow rate to air flow rate. More information would have
been obtained if their work incorporated a factorial design and an
j
analysis of variance. The power consumption values of ~ 0.1 W/m /hr =
3
360 W/m /s = 0.2 hp/1000 cfm are attractive, considering the collection
efficiencies obtained.
Conclusions
Q
The investigators concluded that "charged droplet scrubbers are im-
portant devices for control of particulates in the 0.1 to 1.0 um range,"
and that "scrubbing efficiencies of 30 to 70 percent per stage have been
demonstrated in the submicron particulate size range." It is clear from
their data that they have obtained at least 99 percent collection ef-
ficiency for 2 ym particles during their tests on talc under some of
their conditions. Whether this CDS is a more economical alternative
than electrostatic precipitators or fabric filters or conventional high
energy scrubbers remains to be ascertained.
Evaluation
Suitability of Goals - It is generally recognized that a major drawback
of scrubbers is their power consumption when high efficiencies are de-
sired on fine particulates. It is also recognized that a major drawback
of electrostatic precipitators is their size and associate construction
costs when high efficiencies are desired on fine particulates. The goal
163
-------
of using charged droplets for particle control make sense from both per-
spectives; the electrostatic forces between particle and droplet can
reduce the velocity required of the droplets to collect fine particles
and thus, perhaps, reduce power consumption; the relatively high sur-
face area to volume achieved by using droplets (rather than walls) as
collection sites makes charged droplets an attractive possibility for
electrostatic precipitation. On the other hand, as indicated below and
in the appendix on power consumption, there seems to be no power advan-
tage to accelerating the droplets electrically rather than with fluid
pressure.
Suitability of Methods - We discuss here somewhat more than the
"suitability" of the experimental and theoretical methods.
There are several ways by which, in principle, the charged droplet
scrubber might prove advantageous:
1. The droplets might be accelerated to unusually high speeds
or be accelerated more efficiently in terms of power.
2. The droplets might capture particles with substantially
greater efficiency than is usual.
3. The charged droplets might impart more charge to par-
ticles or might impart the same charge more efficiently.
4. The charged droplet device might be more readily clean-
able than either a typical scrubber or an electrostatic
precipitator or both.
Using the material available to us, we have tried to analyze each of
these potential advantages.
Change transfer in the CDS - The sequence of events for charging of par-
ticles by the charged droplets would be as follows:
1. The droplets leave the spray nozzle initially with a
charge which corresponds to a voltage on the droplet
that matches the nozzle voltage, as both are conduc-
tors. (This is an idealization.)
164
-------
2. The droplets may-lose charge due to ionization of the
air (corona discharge) and perhaps due to the break-up
of the drops due to electrostatic repulsion at the sur-
face. The final charge will be near the lesser of two
limits: the corona breakdown field or the Rayleigh
limit. (This is assuming it started above the lesser
of one of these limits.)
3. When a particle comes sufficiently close (a distance
D between their surfaces, in the TRW notation) to the
droplet, some of the charge from the droplet will be
transferred to the particle. If the droplet and par-
ticle were connected by a conducting wire, this would
bring both to the same potential; this would seem a
reasonable upper limit to assume for the charge ac-
quired by the particle.
4. The particle thus charged would then migrate to the
walls of the collector. (D is chosen so that the
charge transferred is sufficient to cause collection.)
The CDS droplet size distribution was determined to be approximately
log normal, with a number median diameter of 173 ^im and a geometric stan-
dard deviation of a = 1.86. The modal value was 118 urn. The number
O
mean diameter was 210 urn. We will use a value of 200 urn here to char-
acterize the droplets.
One test with the CDS was performed with a nozzle voltage of 46 kV and
another at 31 kV, so we have chosen 40 kV as an approximate value to
characterize the CDS. The formula for the voltage V of a sphere having
charge Q and radius R, is
v =
-------
The charge acquired by the spherical droplet is expected to be 1.33 esu
9
or 2.78 x 10 elementary charges, if it reaches the voltage o£ the
nozzle.
The Rayleigh limit is the maximum amount of charge a sphere can hold be-
fore the charge repulsion overcomes surface tension (7) and disrupts the
sphere. This charge Q is given by
R
1/2 3/2
QR = 1/2 3/Z
where
Q = charge, esu
K
7 = surface tension, dynes/cm
D, = drop diameterj cm.
d
The surface tension for water is 73 dynes/cm. For 200 urn diameter, the
predicte
charges.
_2 g
predicted maximum charge is thus 6.1 x 10 esu or 1.26 x 10 elementary
Air will break dox^n and produce a corona discharge when the voltage
gradient (electric, field) between two parallel plates reaches 30 kV/cm.
o
It is noted that for spheres a correction factor, Peek's, should
1/2
be used, which increases this field by a factor of (1 4- 0.54/(R.) ),
where R. is the droplet radius in centimeters. Peek's correction becomes
d
6.4 for a 200 um diameter droplet; this would mean a corona discharge
field of 1.9 x 105 volts/cm or 640 sV/cm. The field at the surface
E (stat volts/cm) is
Es = VRd
_2
which means that the upper limit on Q becomes 6.4 x 10 esu or
8
1.33 x 10 elementary charges. This limit is very close to the Rayleigh
limit just calculated, so that the 200 um diameter is just about the
size for which the two limits (Rayleigh, corona) are equal. Larger
166
-------
droplets will have their charge limit set by the Raylcigh criterion.
Smaller droplets will have it set by the corona discharge criterion.
For R, « 0.3 cm, the two limits are very close to each other.
Having obtained an upper limit on the droplet charge, we calculated the
amount a droplet might transfer to a particle, if the particle is brought
to the same voltage as the droplet. Using the subscript p to denote
particle (and d to denote droplet), and equating the voltages after the
passage of charge:
V = V
P
which means a 1 um diameter particle would be charged to about 1/200
the level of the 200 urn droplets or about 1.3 x 10 elementary charges.
Such a charge would then produce a field of about 6.4 x 10 stat V/cm or
19.2 x 10^ kV/cm. If Peek's correction is still applicable for particles
as small as 1 urn, then the air breakdown field for a 1 um particle be-
comes 55 times the usual 30 kV/cm, or 5.5 x 10 stat V/cm and ao
corona discharge would occur. The charge would then be 1.38 x 10"^ esu
4
or 2.86 x 10 electronic units, an unusually large charge. The satura-
tion charge acquired by a particle In a corona discharge is such that
the field at the particle surface matches the field in which it is being
charged, thus the droplet would have charged the particle to 55 times
(at maximum) what it might be charged in a corona of 30 kV/cm. The
problem with this analysis is that the droplet may just act as a typical
corona discharge source and charge the particle so that their fields are
equal rather than their potentials, in which case the factor by which
the droplet charging improved upon the usual corona charging would only
be the droplet Peck correction, a factor of 6.4 or less for 200 um drop-
lets. The process may also have a significant rate limitation, unless
the charge transfer is nearly instantaneous.
167
-------
This approximate analysis indicates that the charged droplet method may
be able to increase the particle charge an order of magnitude. This was
not done in as elaborate a manner as by Lear and Krieve, but the assump-
tions are more readily seen and the results are presentable as equations
of closed form.
Droplet velocity production in the CDS - Is there an energy advantage
to producing high velocity droplets with electrostatic forces rather than
using fluid pressure? If the droplets start with the same velocity from
each type of nozzle, then the answer is simple: the kinetic energy per
droplet is the same in each case, and if the number of droplets produced
per unit time is the same, then the power used is identical. Actually,
the droplet steadily decelerates after it leaves a pressure nozzle, but
from the electrostatic nozzle the droplet tends toward an equilibrium
velocity given by electrical forces and fluid forces, this equilibrium
velocity itself changing as the field decreases away from the electro-
static nozzle. The average velocity v over the distance L traveled by
the drop is given by the integration along the path length, of which ds
is the infinitessimal:
L
v = (1/L) I v(s) ds
o
This will be true for both types of nozzles.
The work energy expended, W, by the droplet must come from the nozzle in
both cases and is just the integral of the resistance force over the
distance:
L
F(s) ds
168
-------
For accelerating drops with Reynolds numbers, Re, in the general range
unity to a hundred, Ingebo found that the resistance force term is
approximately (55/24) that of Stokes law, giving for the work energy
L
W = (55/24) 3« u D, f v(s) ds.
By comparing the two integrals, we can see that the same average velocity
is going to require the same energy; thus, there is no inherent advantage
from energy considerations to the charged nozzle compared with the pres-
surized nozzle.
Capture by droplets in the CDS - Most of the mechanisms for particle
capture by drops will be the same for the CDS as for conventional scrub-
bers: impaction, interception, diffusion, diffusiophoresis (under some
conditions). No claims have been made by the developers of the CDS
that the droplet size distribution is especially good for scrubbing
efficiently by such mechanisms, and we have shown that there is no in-
herent advantage to producing the droplet velocities by electrostatic
forces rather than by pressurized spraying. This leaves only electrical
phenomena as possible sources of enhanced collection efficiency. The
o
reports on the chargeu droplet scrubber have emphasized the mechanism
of collection which involves the transfer of charge to the particles by
the droplets followed by the electrostatic precipitation of the particles
onto the walls of the CDS. Their analysis has also included Coulomb and
dipole forces between droplets and particles. We will discuss these
briefly.
Central to their analysis is the concept of "distance of closest ap-
proach," labeled D, which is the farthest distance from the droplet sur-
face to the particle center which still achieves particle collection,
either on the' drop Or by electrostatic precipitation on the scrubber
walls. Another concept used is the "collision effectiveness probability,"
163
-------
labeled p, which is the ratio of the cross-sectional area of the flow
(assumed parallel and directed at the droplet at a great distance from
the droplet) cleaned to the cross-sectional area of the sphere having a
radius equal to the droplet radius plus the distance of closest approach
(Rd + D). Figure 38 shows the various distance dimensions.
These definitions are analogous to those used in the usual treatments
of other capture mechanisms: for impaction, as an example, the distance
of closest approach is D = 0 and the term "single target efficiency"
(T}) is used instead of "collision effectiveness parameter." The volume
of gas cleaned by each particle will be the integrated product
L
2
pit (R + D) ds
which is just the volume swept out by the effective collecting area.
If the mechanism of cleaning were only impaction, then the volume
cleaned by the droplets would be:
R. ds
a
Because impaction is often the predominant mechanism for collecting par-
ticles larger than about 1 ^m, it is worthwhile to compare these volumes.
It is clear that for the CDS to be substantially better as a practical
2
collector, the product of its augmented cross-sectional area, (R, + D) ,
and its collision effectiveness parameter, p, should be substantially
2
greater than TJ T R for conventional scrubbers.
The distance of closest approach augments the geometrical radius of
the droplet with respect to collection. If this increased size is
appreciable, and if the collision probability parameter is no less
170
-------
Figure 38. Droplet and particle dimensions at
distance of closest approach
171
-------
Chan it is due to conventional collection mechanisms, then increased
collection should take place, roughly in proportion to the square of
the ratio of the augmented radius to the geometrical radius, in com-
parison to the conventional situation.
g
Their analysis obtains D as follows. A dimensionless drift time
function, T, is defined as;
T= 0.164 EC-
-------
10'
10
10
-1
1 0
/
/
10
,0-3 2 468 1Q-2 2 468 )()-l 2 468
1.0
O
10
-1
Figure 39. Plot of function G(a) related to particle drift time
8
173
-------
Figure 39 comes from reference 8 and shows G(a) versus a.
We have derived Table 23 from the material above for various particle
diameters (2Rp) assuming, Rd = 60 urn, EQ = 2.3 x 105 V/cm, tQ = 15.
From this table, it is clear that for fine particles (D < 3 pro) the
droplet effective size (R + D) has not been made substantially greater
than its geometric radius.
Table 23. DISTANCES OF CLOSEST APPROACH, D, FOR PARTICLES AND DROPLETS
UNDER ASSUMPTIONS STATED IN TEXT
D
P
(urn)
0.5
1
3
10
T
0.0039
0.0157
0.141
1.57
a
0.0039
0.0151
0.110
0.617
Rd
(urn)
60
60
60
60
D
(um)
0.234
0.906
6.6
37.02
R , 4- D
d
(urn)
60.23
60.906
66.6
97.02
Q
An earlier set of computations for D made under somewhat different as-
sumptions also supports the conclusion that the effective size of the
droplets is not substantially increased with respect to the capture of
particles smaller than a few microns. We quote here the description of
that calculation (note that their system of units is MKS rather than cgs)
"A model for the induced, or corona, charging of particulate by charged
water droplets has been analyzed. The premise of the model is that
the surface electrostatic field on a particle can exceed the breakdown
strength of the medium in which the particle and droplet reside. When
174
-------
breakdown occurs, charge will be transferred between the particle and
droplet. The particle will assume charge of the same sign as the
droplet.
"The condition for induced charging of spherical particulate with
charged water droplet is:
(1)
where: E = surface electrostatic field on the particle
e_ = dielectric constant of the particle material
E = surface electrostatic field on the droplet
s *
S = radius of the droplet
D = separation distance between the droplet surface and
the center of the particle
EQ = breakdown strength of the medium for planar electrodes
R = radius of the particle
C = Peek's correction constant
C = 0.054 for R in meters
"The geometric model for Equation (1) is shown in Figure ^38J and a
solution to the equation is shown in Figure |40J. It was assumed in the
analysis that the electrostatic field on the droplet surface, E , cor-
S
responds to the Rayleigh Limit and the dielectric constant of the
particle material, e , was 5."
"The quantity, D - R, is the maximum distance between the droplet and
particle surfaces at which charge exchange can occur. The values of
the separation distance as a function of particle radius for the various
droplet sizes to the right of the peak value are only approximate
because of the deviation from uniform induced field on the particle
from geometric effect. The curves will decay to zero at a faster rate
in this region because of this effect."
175
-------
a
in
o-
<*
LJ
O C3
< a-
P M
oo
O
O
< C3
g a-
Q_
Q.
-U
op
tf
, s =
= 30 i.
mic»
crons
60
X"
icron
ons
nn
\
l\
!
10
-I
10'
10'
PARTICLE RADIUS,
10'
Figure ^0. Induced charging of spherical particles
-------
The effective size increase of the 120 (am droplet is negligible for
R < 1 urn.
To compare the collision effectiveness probability (p) with the target
efficiency expected from impaction, we have added an impaction effi-
ciency curve to a figure presented by TRW as part of their parametric
computer study of the charged droplet collection process. Figure 41
is taken from their work with the exception of the circles and triangles
which we have added. The ordinates have the particle radius, a char-
acteristic charge parameter, and the ratio A = a = D/R,; the abscissa
is both the collision effectiveness probability, p, for the combined
iset of forces studied by TRW and single target efficiency, r\, for
impaction alone (estimated by from the curves of May and Clifford
for impaction on spheres, neglecting the D contribution to the droplet
radius). The nominal values used here were: drift time of 1 second,
droplet diameter of 120 |_im, electric field of 5 kV/cm, initial droplet
3
field of 230 kV/cm, particle density 2.15 g/cm and air at STP. To
calculate the impaction parameter, we took both 1 m/s and 10 m/s as
droplet velocities, as indicated. The calculations show that the col-
lision effectiveness probability is not much different from the im-
paction efficiency for droplets traveling at 10 m/s (triangles) but
considerably better than from droplets traveling at 1 m/s (circles).
A problem with their calculation is that they have a calculated
velocity for the droplet at the high end of the range of the measured
Q
velocities. The parameter which characterizes droplet velocity is:
U = 2 eQ EQ E Rd/3|a (MRS)
or
U = EQ E Rd/6itu (cgs)
177
-------
1.5
<7>
u
5
oj
.6
.2
10
1 I 1 i 1 I I I I i
5 Q x 10'15 (COULOMB)
L b c
C
n
,'
(
O
J
^
o
)
A
/
6_
O
^
A
A
A /
L/
O
A
A
A
i
/
V'
0 I
A 1
/
/
MPACTIO
M PACT 10
A
y
I/
N ONL
N ONL
A
1
/
/
«'. 1 m/
Y, 10 m
/-
-
-
-
's
/s -
"
- 10
-1
10
-2
2 % 5 10 20 40 60 60
COLLISION EFFECTIVENESS PROBASILITY
90 95 98%
Figure 41. Functional dependence of collision effectiveness probability
on characteristic charge, Q
178
-------
which assumes Stokes law drag on the droplets. Their equations imply
velocities of 2.24 x 10 m/s for the droplets, more than 10 times greater
than those they measured.
If the velocity used in our calculation for impaction is made to be
o
2 x 10 m/s to match that implied by the theoretical parameters,
then the 10 m/s impaction curve shifts parallel to the radius axis
1/2
(the ordinate) by nearly a factor of (20) , making the impaction
efficiencies higher than the collision probabilities. In comparison,
5 2
a nozzle at a pressure of 6.9 x 10 N/m (100 psig) produces a flow
with approximate (potential flow) initial velocity for the droplets of
37 m/s, the collection efficiency curve for which would nearly match
the curves given for collision probability by their analysis.
Estimates of the collection efficiency of the CDS can be made using
.he formula for penetration (one minus efficiency):
3 Q^ T
Pn = exp (- T, -
Dd
where: rj = single target efficiency
Q , = volume rate of flow of droplet material
Q = volume race of flow for gas
O
L = droplet path length
*
D, = ratio of mean cube diameter to mean square diameter.
The efficiency will be high only when the argument of the exponent
is large compared with one. Typical values for the volume flow ratio
-4 *
were 0./Q ~ 10 ; the D, can be estimated by D, which was 200 |itn,
and the path length of the droplets for the prototype CDS was ~10 cm.
The product of the factors other than target efficiency (collision
effectiveness parameter) becomes:
* Dd
179
-------
In order for the argument of the exponent to be one or greater, the
single target efficiency must be greater than or on the order of 10.
The experimental collection efficiencies correspond to an exponential
function with an argument substantially greater than one. This
calculation would suggest that the droplets had single target effi-
ciencies of 1000 percent, much greater than they would have had due to
impaction alone. In a sense, this is true, but in another sense it is as
misleading as it would be to ascribe a very high collection efficiency
to a few drops injected into a conventional electrostatic precipitator.
Collection would primarily occur on the walls, not on the drops.
To summarize, we conclude the following:
1. There is no intrinsic power advantage to accelerating the
droplets electrostatically rather than by the use of
pressurized liquid.
2. Theoretical analysis does not indicate why this charged
droplet scrubber should capture significantly more
fine particles than an uncharged droplet scrubber of
the same power consumption. On the other hand, the
experimental results corresponded to exceptionally
high droplet-particle interaction efficiencies,
casting doubt on modeling the system as a scrubber rather
than as an electrostatic precipitator.
3. The use of droplets to transfer charge to particles may
produce as much as an order of magnitude increase in
particle charging as does conventional corona charging
methods, because che field in the immediate vicinity of
the droplets is higher (Peek's correction) than the
typical breakdown field. The kinetics of the charge
transfer and probability of approach still could negate
the possible charge increase.
4. Uetted-wall precipitation has certain advantages with
respect to cleaning, preventing reentrainment and over-
coming high resistivity in comparison to drywall pre-
cipitation. Liquid waste disposal, however, is more
difficult than dry waste disposal.
180
-------
Applicability to Pollution Control - The charged droplet scrubber studied
9
has shown that it can produce 99 percent collection efficiencies at about
2 ym particle diameter. These results suggest it could be an important
method of fine particle control.
Propects of method - If the promise of such a device were fulfilled,
it would be a hybrid of scrubbing and electrostatic precipitation
which would have lower power usage than a scrubber of identical effi-
ciency and smaller size than electrostatic prccipitators with the
same efficiency. It presents problems in the handling of the water-
borne solids, as do other scrubbers, and in the safety of dealing with
tans of thousands of volts of electricity in the presence of water spray.
Status of method - A pilot plant unit has been constructed and tested
at: 0.47 m /s (1000 cfm) and plans are under way to test a full-scale
model.
Implications - This could lower the cost of control of fine particles,
and it will place added emphasis on the successful handling of scrubber
waste water.
Summary
A charged droplet scrubber has been investigated which uses electro-
static forces to accelerate charged droplets and uses these droplets
to collect particles either directly on the droplets or indirectly
by transferring charge to the particles and collecting them on the
walls of the scrubber, which act as electrodes. It is not clear from
theoretical analysis why this should be substantially superior to
either high-energy scrubbing or electrostatic precipitation, but it
is plausible that charged droplet scrubbing would tend to be less
energy-consuming than high energy Vcntun scrubbers and smaller in
volume than electrostatic prccipitators, at the same efficiency. The
181
-------
next steps in the development of the device should be tests which allow
direct comparison between the charged droplet scrubber and the other
control devices named; data from full-scale operation would also be
very useful.
SYSTEMS OF CHARGED DROPLETS AND PARTICULATES
The basis for the discussion which follows is a report by Melcher
and Sachar entitled "Charged Droplet Scrubbing of Submicron Particulate, "
portions of which are included in a Ph.D. thesis in electrical engineer-
ing for MIT by Sachar. The report describes characteristic times for
various droplet and particle behavior; reviews the literature, including
patents, related to the concept of charged droplet scrubbing broadly
defined; gives theory and experimental confirmation of the theory for
the behavior of charged submicron particle aerosols, charged sprays
much larger than a micron, and the interactions between the two when
present together. The implications of the research are also discussed.
Goals
Objectives delineated in their report were: to analyze what was "unique
in the use of drops and electric fields in collecting fine particles
by providing a classification based upon the fundamental mechanisms
for the electrically induced collection of particles on drops," to com-
plete a literature review on this topic, to perform "experiments that
can be used to test knowledge of the electromechanical dynamics of (a)
systems of charged submicron particles; (b) charged supermicron systems
of droplets and (c) systems involving both charged droplets and charged
fine particles in charged droplet scrubbing configurations," these
experiments to be done (by implication from their being contrasted with
earlier work) with (a) "sufficiently high charged-drop densities to
be of industrial interest and (b) experimental parameters carefully
enough controlled so that comparisons could be made between theoretical
models and experimental results."
182
-------
Methods of Study
Theoretical - Although in a number of instances the relevant equations
are solved in detail and evaluated, much of the work hinges on the tech-
nique of investigating characteristic times, a variety of analysis of
scale. This kind of analysis allows one to draw conclusions based on
models which attempt to incorporate the physical mechanisms involved
and to obtain approximate values for their magnitudes using charac-
teristic dimensions, velocities, etc. Thus, for example, without spec-
ifying the shape of a control device, the gas residence tine, t . is
res
just the ratio of the device volume (V) to the gas volume flow rate
(Q ), in consistent units. The following times are important to the
o
Melcher and Sacher analysis:
Precipitator collection time, t , the electrostatic pre-
cipitator plate spacing divided^§y the charged particle
migration velocity;
"Particle self-discharge or self-precipitation time," t ,
roughly the average interparticle distance divided by a
the particle velocity produced by the Coulomb force be-
tween the particles at that distance (velocity being force
times particle mobility);
"Drop-particle collection time or time for precipitation
of particles due to space-charge of drops," t_, roughly
the average distance between droplets divided by the par-
ticle velocity produced by the Coulomb force between
particles and droplets at that distance;
i "Drop self-discharge or self-precipitation time," tR,
roughly the average distance between drops divided
by the drop velocity produced by the Coulo.nb force
between drops at that distance.
In our notation, these times are given by the following formulas':
183
-------
o
t = l/4« q B N
a MP P P
t = 1/4, q q B
ll B^ NJ
iv. d d d
where the subscripts d and p stand for droplet and particle and the
quantities q, B, N are charge (esu) , mobility (cm/dyne-s), and number
concencration (cm~ ).
It is interesting to note, as Melcher and Sachar do, that these times
are characteristic for the rates of change of number concentration of
particles (t ,t ) or droplets (t ) whether the Coulomb force is at-
3 C K
tractive or repulsive for particle-particle, particle-droplet or
droplet-droplet interactions. This idea can be overstated however,
because there is a significant difference between a particle number
or droplet concentration which decreases due to mutual repulsion to
the walls of the control device and to the decrease accompanying agglom-
eration. . .one mechanism removes mass from the air stream while the
other does not.
The use of characteristic times has an advantage in allowing easy com-
parison with "reasonable" system residence times and the disadvantage
that it is not the way most of the practitioners in air pollution con-
trol, especially those using electrostatics, have formulated the
problem. For scrubbers and electrostatic precipitators, the formula-
tion for penetration of a control system with turbulent flow is:
p o
in which formula the first ratio is that of the outlet particle con-
centration to the inlet, and the argument of the exponential function is
a particle collection velocity times the collection area divided by
184
-------
the volume flow rate of the gas. This argument can be made to be the
ratio of a characteristic time for cleaning, t , and the residence time
in which case we have
Pn = exp (- trca/tc)
and
wA/Qg = (V/Qg)/tc
in which V is the device volume. For a device of fixed volume and
volume flow rate, a decrease in t corresponds to an increase in the
product of the effective migration velocity and the effective collection
area, hence is desirable. Where drops are used as collectors they
must stay in the collection region long enough to do the cleaning,
meaning t >t is required; achieving this can become a problem, because
R c
the highly charged droplets which produce the highest individual drop
collection efficiencies are also those which are most rapidly lost to
the walls of the control device, for unipolar drops, or which coalesce
with other highly charged droplets and become neutralized and less
effective, where bipolar drops are used.
An overview of the results of the Melcher and Sachar analysis is pre-
sented in Table 24, which is an adaptation of one of theirs. It con-
tains a descriptive designation of the system, details about particle
and drop charge, whether there is a substantial net electric field due
to the particles or drops or imposed upon them, and the characteristic
time important in the analysis of its collection efficiency.
Experimental Methods
The experimental work was subdivided to study the behavior of submicron
particles, then "supermicron" drops, then the two together.
185
-------
Table 24. SUMMARY OF BASIC CONFIGURATIONS FOR COLLECTING SUBMICRON
PARTICLES. (BASED UPON TABLE BY MELCHER AND SACHAR).
System
Inert Lai
scrubber
Electrostatic
precipitator
Space-charge
precipitator
Self-
agglomerator
Charged droplet
scrubber
Charged droplet
scrubber
Charged droplet
scrubber
Charged drop
precipitator
Electric
incrtial
scrubber
Electro-
f luidized
and electro-
packed beds
Particle
charge
None
Unipolar
Unipolar
Bipolar
Unipolar
(+ or -)
Unipolar
(+ or -)
Unipolar
(+ or -)
Unipolar
None
None
Unipolar or
bipolar
Drop charge
None
-
-
-
Unipolar
(- or +)
Unipolar
(- or +)
Bipolar
Unipolar
a) unipolar
b) bipolar
None or
bipolar
Ambient field
None
Imposed
Self
None
Self
% s < Nd V
None
(NP s = Nd «a>
None
Self
(NP 1p < Nd V
Imposed or
"self"
imposed
Imposed
Characteristic
times
(deleted)
t
pc
a
t
a
V CR
t = t
c a
V CR
fcc« CR
(deleted)
t (q, based on
c d
"half-charge"
induced on
spherical col-
lector's
hemispheres)
186
-------
In the submicron particle study, the particles were generated with a
condensation generator: an atomizer followed by a heating and con-
oensing section, with the central core of nearly monodispcrse dioctyl-
phthalate (OOP) particles used for experimentation. Their particle
size was measured with the optical owl to find high order Tyndall spectra
and with a polarizer/analyzer to size the smallest particles by their
polarization ratio at 90 scattering angle. Concentration was meas-
ured by extinction measurements. The particles were charged with a
corona discharge and the charge was measured by using the method of tan-
1 7
g;ents on data from a parallel-walled precipitator (see Fuchs 's book
for details). The calculated values of the characteristic time for
self-precipitation, t , were calculated from the measurements and then
cL
were used in the solutions of the equations for penetration obtained
by assuming slug flow and Poiseuille flow profiles. Where the pene-
tration was roughly one-half, the computed solutions differed from each
other by about 20 percent and the data were about 20 percent outside the
range of the two calculations, for laminar flow. Three experiments
with turbulent flow gave evon better agreement with the turbulent flow,
parfect mixing, version of the penetration equations. Experiments
ware also conducted with an aerosol made bipolar by the mixture of two
unipolar aerosols. As with the other experiments, penetration was
inferred from measurements of current due to particle transport versus
distance traversed. The bipolar aerosols were reported to confirm
the t analysis, although the agreement was not as good as had been
3
obtained between theory and experiment for the unipolar case.
For the droplet-droplet studies, vibrating multiple orifices were used,
as: done by Berglund and Liu for example. The droplets were charged
b> induction by having this generator attached to high voltage sources
operated at the same or opposite polarities. Useful discussion is given
ccncerning this and other types of charged droplet generators. Droplet
velocity in a long tube was measured, as was gas velocity. The charge
could be calculated. Predictions using droplet characteristic time t
K
187
-------
agreed fairly well with experimental results. For self-discharge, the
charge of the drops was inferred by their deflection by electrodes; the
results were 6 to 8 times different from the predictions made by their
model, the droplets discharging each other more slowly than expected.
The droplet-particle studies focused on three systems: charged par-
ticles collected by oppositely charged droplets, charged particles col-
lected by bipolar droplets, charged particles repelled to the walls by
charged droplets with the same polarity. The model used to compare with
tests predicted droplet collection efficiency in a manner similar to
that for which charge acquired in corona discharges is calculated (as
14
done recently by Smith and McDonald), assuming a uniform external
electric field and a viscuous flow model. This flow was reported to
give results little different from those for other flow models.
Results
In general, the measured collection efficiencies of the various con-
figurations verified the time scaling approach and the efficiency equa-
tions associated with that approach. Figures 42 and 43 from the work of
Melcher and Sachar shows the agreement between the theoretical ratio
of outlet concentration to inlet concentration (n /n. ) for their test
aerosol. The measured values are graphed against the drop charging volt-
age, the specific configuration being the precipitation of positive par-
ticles by negative drops and, in Figure 43, by positive drops. Table 25
adapted from one of their tables. For three different kinds of precipi-
tator it gives the measured efficiencies with no charging, with charging
only the particles and with charging both particles and droplets.
188
-------
O.JO
i r
i r
e.u
o.io
JL_ff_J 1 1 L
10 to
«... -I J-
1» JOO ISO JOO
Figure 42. Theoretical and measured collection of positively charged
aerosol particles upon negatively charged drops as a
function of drop charging voltagel6
I I I 1 I > ri
10 10 X) 40 M - nil, 100 1» 200 . 3» 100
Figure A3. Theoretical and measured particle collection for pre-
cipitation of positively charged aerosol particles by
positively charged drops as a function of drop charg-
ing voltagel6
189
-------
Table 25. EXPERIMENTALLY DETERMINED EFFICIENCIES FOR THREE CHARGED
DROPLET SCRUBBER CONFIGURATIONS16
Conditions
Scrubber configuration used
Unipolar drops
and unipolar
aerosol, oppositely
charged
Bipolar drops
and unipolar
aerosol
Unipolar drops
and unipolar
aerosol, same
sign charge
No charge
on drops
or aerosol
Charge on
aerosol
only
Charged
drops,
charged
aerosol
257.
87%
95%
257.
86.57,
927o
257o
857o
95%
Conclusions
Based on their experimental and theoretical work, Melcher and Sachar
concluded that the various possible types of charged droplet scrubbers
would not have residence times (thus volumes, thus capital costs) smaller
than electrostatic precipitators with the same efficiency. They con-
cluded that electrostatic augmentation would increase the collection
efficiency of spray scrubbers. Charged droplet scrubbers thus form "a
class of devices x^ith the capital investment and operating cost profile
of the wet scrubber but a particle removal efficiency approaching thau
of the electrostatic precipitator."
Evaluation
Suitability of Goals - As noted elsewhere, the idea of using elec-
trical forces to enhance collection efficiencies of spray scrubbers is
one which seems promising. The work done by Melcher and Sachar aimed at
190
-------
defining the various combinations of droplet and aerosol charge, deter-
mining the efficiencies which could result and looking for the under-
Lying similarities. This rather fundamental approach is quite useful
for unifying the analysis of control device ideas which seem quite dis-
similar.
Suitability of Methods
Theoretical Approach - The emphasis on dinensionless groups (the ratios
of characteristic times) is particularly appropriate for work which
seeks to be applicable to a wide variety of device configurations which
differ appreciably from simple models, and for which the would-be
analyst has only sketchy information. A minor problem is that the use
of characteristic times differs from the usual methods of analysis:
taose who work with electrostatic precipitators are used to a similar
aaproach in terms of migration velocity and those who work with scrub-
bers are more familiar with droplet collection efficiency. In general,
Mtilcher and Sachar have used equations describing droplet and particle
charging and motion which are at or near the state-of-the-art and have
us.ed them at a level of detail appropriate to the degree of information
a\ailable about the parameters which enter into the equations.
Experimental Approach - As can be seen from the information in Table 26,
Melcher and Sachar have measured or defined the important variables
in their experimental systems. The methods used to measure particle
and droplet size and charge showed an understanding of the problems of
such measurements and an awareness of the current methods in aerosol
technology.
Applicability to Pollution Control - Because the work done was an
academic investigation rather than the testing of a specific control
device, we will not try to evaluate the applicability to pollution con-
trol of the devices except to paraphrase the authors' conclusions that
191
-------
Table 26. PARAMETERS ASSOCIATED WITH THE STUDY OF SYSTEMS
OF CHARGED DROPLETS AND PAJITICULATE
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Re )
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
Variable
(1) 0.5 m/s (see comments)
(2) M
(3) 10-50 m/s
K
K
?
ambient
ambient
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
(1) 1-1.0 micron
(2) monodisperse
spherical
(1) OOP
(2) ?
(3) ?
K
K
(1) unipolar measured
(2) bipolar
(3) A - no charging
B - bipolar charging
M
CHARGING SECTION
Type of charging
Ion concentration
Electric field
Geometry
impact charger
Variable, M
Variable, M
K
192
-------
Table 26 (continued). PARAMETERS ASSOCIATED WITH THE STUDY OF SYSTEMS
OF CHARGED DROPLETS AND PAKTICULATF.
Parameter
Magnitude, description, method of
measurement or control, etc.
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage, electric field
Particulate loading
Geometry
Internal configuration
External configuration
Collection efficiency
water
K
K
(1) unipolar using induction
charging variable
(2) bipolar-charged and re-
charged by induction
charging
(3) induction charging - measured
(1) no ambient
(2) no ambient field utilized
(3) ambient field
(1) 5-25 vim
(2) 5-25 pm
(3) 2.5-10 urn
M
CLEANING PROCESS
Method
Effect on efficiency
N.A.
?
COMMENTS: Three types of systems are covered:
(1) Unipolar particles and oppositely charged unipolar
particulate.
(2) Bipolar particles and particulate and ambient electric
field.
(3) Electrically driven impact scrubbing and agglomeration
through particle polarization.
STAGE OF DEVELOPMENT: Final Report - Phase I
193
-------
the operating characteristics of the charged droplet scrubbers can be
expected to fall between those of electrostatic prccipitators and con-
ventional high-energy scrubbers in terms of efficiency, power consump-
tion, and capitalization.
Summary
The MIT researchers have categorized the configurations for charged
droplet scrubbing, shown the similarity of the time constants
involved for the various configurations, measured particle and droplet
concentration changes under well-defined conditions and verified that
these time constants can be used in mathematical models which predict
measured collection efficiencies rather well. They concluded that charged
droplet scrubbers have performances which lie between those typical
of electrostatic precipitators and spray scrubbers, which may mean they
will be optimal for certain control problems.
194
-------
REFERENCES
1. Pilat, M. J., S. A. Jaasund, and L. E. Sparks. Collection of Aero-
sol Particles by Electrostatic Droplet Spray Scrubbers. Environ
Sci & Technol. 4:360, 1974.
2. Sparks, L. E. The Effect of Scrubber Operating and Design Parameters
on the Collection of Particulate Air Pollutants. Ph.D. dissertation,
(Civil Engineering), University of Washington, 1971.
3. Calvert, S. Engineering Design of Wet Scrubbers. J Air Pollut
Contr Assoc. 24:929, 1974.
4,, George, H. F. and G. W. Poehlein. Capture of Aerosol Particles by
Spherical Collectors: Electrostatic, Inertial, Interception, and
Viscous Effects. Environ Sci & Technol. 8:46, 1974.
5. Nielsen, K. A. Correspondence on "Capture of Aerosol Particles by
Spherical Collectors." Environ Sci & Technol. 8"767-769, 1974.
6. Mid-West Research Institute, Evaluation of Electrostatic Droplet
Scrubber. Contract No. 68-02-1324, T.O. No. 16, for Control Sys-
tems Laboratory, Office of Research and Development. Environ-
mental Protection Agency.
7. Stein, R. L., W. H. Ryback, and A. W. Sparks. Deposition of Aerosol
in a Plastic Chamber. J Colloid Interface Sci. 42:441-446, 1973.
8. Lear, C. W. and W. F. Krieve. Progress Reports for EPA on Contract
No. 68-02-1345. Application of Charged Droplet Scrubbing to Fine
Particle Control.
9. Lear, C. W., W. F. Krieve, and E. Cohen. Charged Droplet Scrubbing
for Fine Particle Control. J Air Pollut Contr Assoc. 25:184-189,
1975.
10. Ingebo, R. Drag Coefficients for Droplets and Solid Spheres in
Clouds Accelerating in Airstreams. NACA Technical Note 3762, 1956.
11. May, K. R. and R. Clifford. The Impaction of Aerosol Particles on
Cylinders, Spheres, Ribbons, and Discs. Ann Occup Hyg. 10:83-95,
1967.
12. Fuchs, N. A. Mechanics of Aerosols. Pergamon, New York, 1964.
13. Bcrglund, W. B., and B. Y. H. Liu. Generation of Monodisperse Aero-
sol Standards. Environ Sci & Technol. 7:147-153, 1973.
195
-------
14. Smith, W. B. and J. R. McDonald. Calculation of the Charging Rate
of Fine Particles by Unipolar Ions. J Air Pollut Contr Assoc.
25:168, 1975.
15. Melcher, J. R. and K. S. Sachar. Charged Droplet Technology for
Removal of Particulates from Industrial Gases. Final Report.
EPA Contract No. 68-02-0018, August 1, 1971.
16. Melcher, J. R. and K. S. Sachar. Charged Droplet Scrubbing of Sub-
micron Particulate. Draft Final Report. EPA Contract No. 68-02-0250,
July 1974.
196
-------
SECTION VII
ELECTROSTATIC AUGMENTATION OF PRECIPITATION DEVICES
GAMMA-RAY PRECIPITATOR
The gamma-ray precipitator prototype differs fundamentally from a
conventional electrostatic precipitator only in its mechanism of charge
production. The experimental gamma-ray precipitator employed a Co
y-ray source rather than a corona discharge. Any advantages as an air
cleaning device for the y~ray precipitator must involve an advantage in
the charging of the particulates. In both the conventional electrostatic
precipitator and the y-ray device the mechanism of particulate deposition
by electrostatic field removal is the same. Furthermore, the primary
charging mechanisms of diffusion charging and field charging are at work
in both devices.
Goals
The Pennsylvania State University (PSU) Department of Nuclear Engineering
undertook both an experimental and theoretical study of the gamma-ray
precipitator. Their goals were to measure the prototype device's
collection efficiency with respect to particulate size and to explain
the performance by a theoretical model. Midwest Research Institute
(MRI) undertook an evaluation of the PSU group's work which amounted
to a recheck of calculations and a general discussion of the practical
utility or lack thereof of the Y-ray precipitator. The overall con-
clusions of the two groups were at varLance: MRI was negative in its draft
final report; PSU found the concept promising.
197
-------
Methods of Study
Experimental - The PSU group performed an experimental study of the
collection efficiency of the /-ray precipitator in a pilot plant opera-
tion. The basic precipitator flow system consisted of tiro concentric
stove pipes, 20.32 cm i.d. and 12.7 cm i.d. (8 inches and 5 inches),
which served as collection electrodes. The vertically-oriented assembly
(see Figure 44) directed incoming ash-laden gases from a coal-fired fur-
nace down the central pipe. At the bottom of the assembly in the irra-
diation zone the gases were forced to make a 180° turn to enter the
annular space between the two pipe electrodes. The annular space near
the turn-around was strongly irradiated by 7-rays from surrounding
pencils of Co . This was the basic site of air ionization and much
of the deposition. To a decreasing extent the remaining length of
annular space above the irradiation zone served to collect electro-
statically the charged particulates.
The basic experimental data collected in the studies were six-stage
impactor samples taken of the inlet and outlet gases. The weight-
percent impactor data was adjusted to a number-percent scale by assump-
2
tion of 2 gm/cm as a reasonable particulate material density. Particle
collection efficiencies were obtained from these adjusted results.
Experimentally variable parameters included the annular space air
velocity (generally 152 cm/s, or 5 ft/sec), irradiation dose rate,
electrode potential, and electrode polarity. Total weight-percent
collection efficiencies were taken at varying electrode potentials,
dose rates, and alternate changes in electrode polarity. Weight of ash
deposited on test patches along the annular collection zone indicated
the anticipated fall-off in collection along the precipitator length.
Theoretical - Schultz et al. predicted the collection efficiency varia-
tion with particle size. The basic model input was the assumed ion
198
-------
Firnoce
Shed
I
Building Wall
Blower
Dampers
t Cast Iron
/I Assembly
CHO
Power Supply
Support
Structure
Water Level.
lonization-
Prccipilator
Chamber
Pool
Sampling Ports
/I Insulators
y
I-Beam
I
Standoff
Insulators
Lead Jacket
Cobalt Source
Platform
Figure 44. Schematic diagram of gamma ray prccipitator and auxiliary
equipment
199
-------
concentration which they claimed by "crude measurement" to be as high as
10 ions/cm near the electrodes. Their unpublished theoretical ioniza-
tion calculations indicated that at least 10 ions/cm would be achieved
in the irradiated zone. The particle drift velocity is linearly related
to the saturation charge of individual particles due to field charging
and the attainable diffusion charge. Theoretical efficiencies were com-
puted directly from the Deutsch equation by combining drift velocity and
the collection area to volume flow ratio (A/V) in the exponent.
Schultz et al. also discussed mechanisms of unipolar ion production
which may be involved in the 7~ray precipitator. The 7-rays interacting
with the electrode walls and the air molecules ultimately produce both
positively and negatively charged ions. In the view of the authors,
photoemission of electrons at the walls combined with high ion concen-
tration produces space charge separation between the electrodes suffi-
cient to inhibit recombination and strongly charge the particles.
The MRI evaluation accepted the validity of the Deutsch equation effi-
ciency approach but focused its attention on two difficult particle
sizes, 0.01 and 0.1 urn. The evaluation was seriously flawed in that
it made the mistaken assumption of Cunningham slip correction approxi-
mately equal to 1.0 for these very fine particles.
Results and Conclusions
In their paper, Schultz et al. compared their experimental data with
the efficiencies they would predict for an electrostatic precipitator
under comparable but not identical conditions. Although this comparison
was favorable for the GRP, it is not conclusive because the two cases
were not identical and because the comparison is between experimental
data for one system and theory for the other. In addition, as indicated
below, our calculations of the particle charge achieved, using these data,
indicates inferior rather than superior charging by the GRP compared with
the ESP.
200
-------
The MRI analysis indicated only marginal advantages in charging for the
GRP. Our recomputations still support this conclusion. Furthermore,
MRI concluded that the special problems connected with the handling of
radioactive materials probably outweigh any advantages under present
circumstances.
Evaluation
Certain shortcomings in the theoretical analysis of the GRP are cor-
rected in the material which follows and a short analysis of the experi-
mental data is presented, both of which support the conclusion that the
GRP has no substantial inherent advantages for particle charging com-
pared with the conventional corona discharge method now in use. The
economics of obtaining and handling radioactive material will vary
greatly from installation to installation and would have to be analyzed
by each potential user. For those without special access and expertise,
we believe the problems and cost to be more substantial than those en-
countered with corona charging. The studies done thus far by Dieter and
2 1
Schultz in 1971 and Schultz, et al. in 1973 have not demonstrated how
the collected material could be removed from the collector economically,
a major drawback.
Suitability of Goals - The major problems in electrostatic precipita-
tion technology can be inferred from the focus of most of the recent
work (e.g., Symposium on Electrostatic Precipitators, Pensacola Beach,
Florida, October 1974, sponsored by EPA): adhesion, high resistivity,
high temperature, sncakage, reliability, gas flow distribution, elec-
trostatic augmentation of control devices. Still, the possibility that
radioactive wastes could be put to good use was one worth investigating.
Bci ausc the primary difference between the GRP and the ESP is the method
of particle charging, this should have been studied in detail, although
the cxperjmcnts do allow some conclusions to be drawn concerning parti-
cle charging. There is value in demonstraLing, as has been done, that a
technique docs work; what remains is a determination of its practicality.
201
-------
Suitability of Methods - Experimental and theoretical methods are dis-
cussed here.
Theoretical - Leipunskii et al. point out that there are more than
10 types of elementary processes of interaction of 7-rays with matter.
For the 1.17 Mev and 1.33 Mev 7-rays emitted by Co , only three pro-
cesses occur with significant probability:
(1) Photo-electric absorption - 7-ray completely
absorbed by K shell
electron
(2) Compton scattering
(3) Pair production - minimum of 1.022 Mev 7-ray
required to produce pair
Either the entire energy of the 7-ray or partial energy is transmitted
to electrons in each of the three processes. These primary electrons
are called:
(1) Photo-electrons
(2) Compton electrons
(3) Electron-positron pairs
Part of the energy lost to electrons in these primary events may
ultimately be recovered as radiation in the form of Bretnsstrahlung
(radiation due to acceleration of charged particles) and annihilation
radiation (positron-electron combination). The secondary radiation is
of negligible importance compared to the primary 7-ray flux.
It is of great importance to the efficiency calculations to determine
an average theoretical ion concentration in the annular space of the
irradiation zone. Precise analysis is difficult because primary elec-
trons emitted at the electrodes may be heavily involved. However, a
lower bound on the average ion concentration can be obtained from the
published range of dose rates to the annulus.
202
-------
Dose (as contrasted with the term absorbed dose for biological systems)
describes the ionizing effect of y-rays on air. Dose is in units of
Roentgens. One Roentgen (R) is the dose required to produce one
CGS unit of charge (of'either sign) in 1 cm of air at 0°C, 760 mm Hg.
9 ^
So 1.0 R = 2.08 x 10 ion pairs/cm . Since mechanisms of energy
absorption of y-rays depend on the energy of the y-rays, different
fluxes are required to produce the same dose for different 7 energies.
Schultz et al. indicate the dose rate range employed in the expert-
o C
ments was 10 R/hr to 1.5 x 10 R/hr. Using the conversion factor for
Roentgens to ion pairs, the ion pair production rate is calculated.
Dose rate
(R/hr)
IO3
io4
5
1.5 x 10^
Dose rate
(R/sec)
2.78 x IO"1
2.78
41.7
Q, production rate of
(lons/cm-s)
5.78 x IO8
5.78 x IO9
in
8.67 x 10iU
ions
An equilibrium ion concentration, nro, can be obtained from the relation
given by Cooper and Reist:
f\ *\
Where ex = 3.6 x 10~ cm /s, the recombination coefficient.
The results for the various dose rates are:
Dose rate
(R/hr)
IO3
1.5 x IO
n ion pairs/cm
2.53 x IO
8.01 x IO
3.10 x IO
7
8
203
-------
The calculated n values compare favorably with the theoretical result
7 1
of a minimum of 10 ion pairs/cm mentioned by the PSU group. However,
the figures fall short of their approximate measurement of 10 ion
pairs/cm .
Table 2 and Figure 2 of the MRI draft report have been redone (Table 27,
Figure 45) with the sole alteration of use of proper slip corrections.
That investigation assumed C - 1.0 for both 0.01 urn and 0.1 urn parti-
cles whereas the proper figures are approximately 17 and 2.6. The
correct drift velocities were thus 17 and 2.6 times those calculated.
The increased drift velocities yield substantially increased calculated
efficiencies for both GRP and ESP. Efficiencies for 0.01 um particles
are now shown to be greater than for 0.1 pm particles for both devices.
A further adjustment of the table and chart is made in Table 28 and
Figure 46, based on somewhat different assumptions. First, a higher
stack temperature of 400°K instead of 300°K was used throughout -
this change increases diffusion charge achieved in the nominal 1-second
charging period. Second, instead of MRI's assumed 10 ion pairs/cm ,
8 3
a typical value of n^ = 5 x 10 ion pairs/cm for a corona discharge
ESP is assumed. White states that this is a typical ESP value.
Even though our theoretical calculations indicate a maximum of
83 9
3.10 x 10 ion pairs/cm for the GRP, we assume here n= 10 ,
O
roughly the logarithmic mean of 3.10 x 10 and the experimental re-
I rj O 93
suit of ~ 10 ions/cm . The assumption of 10 ions/cm was also
made in the MRI report.
The results obtained using more generous assumptions for the ESP-GRP
comparison show very marginal collection efficiency superiority for
the GRP. The theoretical model for calculating collection efficiencies
is sound, but there is still the uncertainty about attainable unipolar
ion concentration in the y-ray precipitator. PSU's impactor measure-
ments lack sufficient resolution in the sub-micron region to indicate
what collection efficiencies were achieved for 0.1 urn particles.
204
-------
Table 27. CORRECTED VERSION OF MRl's TABLE
Particle
radius
(nm)
0.1
0.01
Charge per particle
(elementary charges/
ESP
4.11
4.11
4.11
4.11
0.11
0.11
0.11
0.11
GRP
12.2
12.2
12.2
12.2
0.81
0.81
0.81
0.81
Drift velocity
(cm/sec)
ESP
1.54
1.54
1.54
1.54
2.71
2.71
2.71
2.71
GRP
4.57
4.57
4.57
4.57
20.0
20.0
20.0
20.0
A/V
(sec/cm)
0.08
0.39
0.59
0.79
0.08
0.39
0.59
0.79
Collection efficiency
a)
ESP
11.6
45.2
59.7
70.4
19.5
65.2
79.8
88.2
GRP
30.6
83.2
93.3
97.3
79.8
99.96
99.99
fa 100.0
to
o
-------
GRP = GAMMA-RAY
PRECIPITATOR
ESP = ELECTROSTATIC
PRECIPITATOR
01 0.2 0.3 04 05
A /V , sec /cm
06
0.7
0.8
Figure 45. Estimated collection efficiencies for gamma ray
precipitator and electrostatic precipitator
206
-------
Table 28. RESULTS FOR ALTERED ASSUMPTIONS
Particle
radius
(urn)
0.1
0.01
Charge per particle
(elementary charges)
ESP
6.13
6.13
6.13
6.13
0.343
0.343
0.343
0.343
GRP
6.96
6.96
6.96
6.96
0.423
0.423
0.423
0.423
Drift velocity
(cm/sec)
ESP
2.30
2.30
2.30
2.30
8.46
8.46
8.46
8.46
CRP
2.61
2.61
2.61
2.61
10.5
10.5
10.5
10.5
A/V
sec/cm)
0.08
0.39
0.59
0.79
0.08
0.39
0.59
0.79
Collection efficiency
(%)
ESP
16.8
59.2
74.3
83.7
49.2
96.3
99.3
99.9
GRP
18.8
63.9
78.6
87.3
56.8
98.3
99.8
X. 100.0
-------
ESP= ELECTROSTATIC
PRECIPITATOR
GRP = GAMMA-RAY
PRECIPITATOR
3 .4 .5
A/V , sec /cm
.8
Figure 46. Estimated collection efficiencies for gamma ray precipitator
and electrostatic precipitator utilizing altered assumptions
208
-------
One possible mode of particle charging not explored in the published
reports of cither PSU or MRI is direct y-ray charging of the particles.
The prospect of direct 7-ray particle charging might have altered the
ESP-GRP comparison favorably toward the new device. Our calculations
showed that direct 7-ray charging will be unimportant and will not
alter the conclusions of the diffusion and field charging calculations.
Experimental - Hie experimental data obtained by the PSU group can be
used to estimate the charge levels actually achieved by the GRP, arid
these turn out to be no better than the usual ESP charge levels. To
cbtain the charge levels, one first calculates the effective migration
velocity (w, cm/s) from data on the penetration (Pn = 1 - efficiency)
2
and collection area (A, cm ) and volume flow rate (V), using the
Eieutsch equation:
v -wA/V
Pn = e
- In Pn = w(A/V)
w = - (V/A) In Pn
From Figure 13 of Schultz et al., E = 1 - Pn was 0.997 at 5.4 ym,
0.965 at 1.8 urn, 0.90 at 1.1 urn, and 0.78 at < 0.4 urn. The volume flow
can be estimated from the statement that most of the tests were done at
5 ft/s and were in the 4 to 15 ft/s range; the pipes used were 8 inches
i.d. (20.3 cm) and 5 inches i.d. (12.7 cm), which is a cross-sectional
2 2
area of about (it/4)(413 - 161) cm = 198 cm , thus a volume flow of
(L52 cm/s)(198 cm2) = 3.02 x 104 cm3/s [64.0 ft3/min]. The collector
length seems to have been 12 ft (3.6 m). Collector area was thus
A = n(20.3 + 12.7)(360) cm2 = 3.73 x 104 cm . This means:
V/A = 0.81 cm/s = 0.027 ft/s.
209
-------
The ratio A/V was about 1.2 sec/cm = 37 sec/ft. The tests were done
with a much larger A/V than usually used with ESP's, values close to
those here in Tables 27 and 28.
From the penetration value we obtain the effective migration velocities
given in Table 29.
The migration velocities can, in turn, be used to calculate approximate
particle charge, using for the electric field E - 20 kV/1.5 inches or
5.2 kV/cm or 17 stat-volt/cm. Then
W = qpE B
Table 29. APPROXIMATE GRP EFFECTIVE MIGRATION
VELOCITIES (EXPERD1ENTAL)
Particle size
(nm)
5.4
1.8
1.1
~ 0.4
Migration velocities
(cm/s) (ft/s)
4.70 0.155
2.72 0.089
1.85 0.061
1.22 0.041
where q is particle charge (elementary charge = 4.8 x 10 esu) and
B is particle mobility (C/3irud). This gives the experimental average
charges shown in Table 30.
These charge values are somewhat lower (2-1/2 x) than the corona charging
done by Penney and Lynch at 2.3 kV/cm and by Hewitt at 0.6 kV/cm, as
reported by Whitby and Peterson (1965). Thus, the charging produced
by the y-ray precipitator does not seem to have been superior to that
achieved by conventional corona charging methods, even at the low end
of their electric field range.
210
-------
Table 30. EXPERIMENTAL CALCULATED AVERAGE CHARGE FOR CRP
Particle diameter
0.4 um
1 .1 um
1.8 um
5 .4 um
Particle charge
(esu)
3.40 x 10"9
17.7 x 10"9
4.47 x 10"8
25.0 x 10"8
(elementary charges)
7.1
36.8
93.2
528.
In summary, the experimental evidence does not demonstrate that the
gamma-ray precipitator produced superior particle charging in compari-
son with the usual electrostatic precipitator method of corona dis-
charge electrification. The theory for the charge production and
equilibrium for the device is still incomplete, but the available
theory indicates particle charging only marginally superior to that
oi the corona technique. Table 31 summarizes advantages and disadvan-
tages of the GRP, from the MRI draft. Because tnere are no substantial
inherent advantages with respect to particle charging and particle
collection, economic considerations will determine the relative utility
oi the GRP with respect to the ESP. At present, the problems of
shielding and safety seem to outweigh the advantages of reduced
electrical power consumption and no maintenance of corona discharge
wires. Particularly problematic would seem the removal from the GRP
of the particulate material collected therein. The GRP is not at
present a practical replacement for the ESP.
Applicability to Pollution Control - Prospects, status, implications are
treated next.
Prospects of method - At best, the method would give collection
efficiencies comparable to those of electrostatic precipitators.
Energy consumption would be less than that of ESP technology, but
special material handling techniques would be needed for the radio-
active sources and for removal of the collected material.
211
-------
Table 31. GAMMA-RAY PRECIPITATOR ADVANTAGES AND DISADVANTAGES
Advantages
Disadvantages
1. Possibly greater collec-
tion efficiency for
< 0.2 um particles
2. Low power consumption
3. Eliminates fragile corona
wires - rapping easy
4. Useful storage for
utility-produced
radioactive wastes
1. Little efficiency improve-
ment with added techni-
cal and economic
problems
j
2. Expensive on a per ft
treated basis
3. Radiation hazards,
licensing problems
4. Massive shielding required
5. Complex maintenance and
operation
Status of method - Collection efficiencies have been demonstrated
at a nuclear reactor site using a small home-type coal furnace.
Implications of rrethod - If practical, it would provide a use for
radioactive wastes and reduce the energy consumption of the particle
charging aspect of electrostatic precipitation. It would increase
radiation hazards, especially if used in numerous small applications.
Summary
The gamma-ray precipitator does not charge particles to appreciably
higher charge levels than normally achieved by electrostatic precipi-
tators. The collection mechanisms in the two types of devices are
the same. The only advantage of the gamma-ray precipitator is that
it charges particles with radioactive sources rather than with a
corona discharge, saving electrical energy, but this is also its
primary disadvantage, as the handling of radioactive materials poses
formidable problems. Only if it can be shown to be prospectively a
212
-------
more economical alternative than electrostatic precipitator would
there be a rationale from an air pollution control perspective to
Investigate if further.
AC FIELD "ELECTRIC CURTAIN"
This discussion is based on information from C.D. Hendricks (Illinois
University), who has had extensive experience in a related field, mass
spectroscopy, and some of the concepts were germinated by that experience.
Goals
Dr. Hendricks"s work involves theoretical and experimental research on
four related devices:
1. An electrostatic curtain to be used downstream from an
electrostatic precipitator to use electrical repulsion
of the particles that penetrated the precipitator to
keep these particles from flowing through a charged
planar assembly of rods. See Figure 47.
2. A planar assembly of rods used instead of the collect-
ing plates of an electrostatic precipitator, with a
travelling wave electric field parallel to the surface
of the assembly to move the particles into a collect-
ing zone. See Figure 48.
3. An electric curtain placed so as to keep scrubber drop-
lets in the flow longer than they would if the curtain
were not there, to enhance scrubbing. See Figure 49.
4. A device for sampling which would use the size-
selective aspects of the electric curtain field to
collect size-fractionated particle samples. (Because
of the very limited information about this idea, it
will not be considered here further.)
Methods of Study
Theoretical Methods - The equations studied are those for a particle under
electrical and other forces, such as gravity, in a medium which is
213
-------
N
L
E
T
I IT
FLOW
PARTICLE
TRAP
AC GENERATOR
Figure 47. Electric curtain
-------
N>
»-
Ul
HIGH
VOLTAGE A.C.
PLUS D.C.
DELTA-WYE
30 TRANSFORMER
HIGH
VOLTAGE D.C.
SUPPLY
CONNECTED IN THE
SAME WAY AS
OPPOSITE SIDE
Figure 48. Electric curtain connected so as to provide a traveling wave
electric field moving toward bottom of rods (after Hendricks)
-------
POSSIBLE TRAJECTORIES OF
M
DUCT
INLET-
RAILS ENERGIZED BY HIGH VOLTAGE A.C.
.^-DUCT
> OUTLET
Figure 49. "Horizontal rail structure to support liquid scrubber drops
to increase interaction time with gas flow from which it is
necessary to remove gases such as S02 by absorption or
chemical interaction with scrubber drops" (Hendricks)
-------
resisting the particle motion according to Stokes law of fluid resis-
tance. This is a second order ordinary differential equation for
position as a function of time. Using dots to indicate derivatives
with respect to time, these equations would be:
m x* = F + q E cos cot - 3rtuxd
x p x p
for one-dimensional motion of a particle having diameter d with an
external force F and electrical field E cos cut. Details of the
x x
approach to solving such equations to be used bv Hendricks and co-workers
were not presented, but computer solutions seem straightforward and they
will use "computer simulation." Features to be studied, presumably theo-
retically as well as experimentally, included the effect of the following
on efficiency:
« Particle size
Particle charge
e Gas flow velocities
Voltage
Geometry
Thaoretical analysis very similar to that described in the preceding
paragraph will be done for the case where the rods are used as the
plates of an electrostatic precipitator. Results already obtained by
Masuda may prove useful, although it is expected that for the problem
of many particles rather than one, such analysis may not be applicable.
A similar type of analysis seems planned for the scrubber geometry,
as well.
Experimental! Methods - Each of the three applications for the electric
curtain are to be tried out in test models. The following variables
were explicitly listed to be tested for their effect on efficiency:
217
-------
Particle size
Particle charge
t Gas flow velocity
o Gas composition
Voltages, AC and DC
Geometry
Scrubber droplet sizes
Flow patterns
Power consumption
Laminar and turbulent flow
Methods of reentraimnent elimination
This is a very complete list. The group is experienced in the genera-
tion of uniformly sized particles and in charge measurement, two areas
in which expertise is very useful. The actual methods to be used to
accomplish the above were not elaborated in their proposal.
Results
If the program is successful, the effects of the various parameters
listed in the Methods sections will be quantified by computer simula-
tion and by experiment. The experimental set-up would involve an
2
electric curtain of about 25 m , so that the results would be nearer
a pilot scale device than a bench top or laboratory scale, and if any
of the applications studied appeared feasible, it would be possible to
scale up to a demonstration unit much more readily than if a lab scale
device were being used.
218
-------
Conclusions
The results would probably be sufficient evidence to remove from con-
sideration those configurations which did not prove feasible. As
there are no results yet, no conclusions can be drawn.
Evaluation
Suitability of Goals - Each of the proposed electric curtain applica-
tions has a goal which is suitable to air pollution control:
1. The elimination of the particles xvhich escape electrostatic
precipitators.
2. The reduction or elimination of particle reentrainment
due to rapping used to clean solid surface electrostatic
precipitator electrodes.
3. The formation of something analogous to a fluidized bed
using water droplets for scrubbers. Water is about the
cheapest material imaginable for such applications. (Arc-
ing may turn out to be a serious hazard and drawback,
however.)
iSuitability of Methods - As presented, we find nothing to object to in
the proposed methods of study of the possible applications of the elec-
tric curtain. As Table 32 shows, there are a number of unspecified
aspects with regard to the particles and with regard to their charging.
The particle characteristics will probably be known or measured, how-
over, and the particle charging, though of interest, is of less concern
than the particle charge, which is to be measured.
Applicability to Pollution Control - There are several aspects which
nake its applicability questionable.
Prospects - Although the study is to be a very thorough one, one which
should generate a substantial amount of useful information, each of the
proposed applications has drawbacks which make it a doubtful prospect
219
-------
Table 32. PARAMETERS ASSOCIATED WITH THE STUDY OF THE ELECTRIC CUR-
TAIN AS A DEVICE FOR THE CONTROL AND REMOVAL OF PARTICU-
LATE MATERIALS
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume flow rate
Face velocity
Reynolds number (Ref)
Flow geometry
Pressure drop (Ap)
Temperature
Pressure
Relative humidity
M (to be measured)
M (will be one of the variables in
testing the curtain)
laminar and turbulent flows will
be measured
several types
it is noted that Ap is to be
minimized utilizing expanded
ducting
M
approximately ambient
gas composition is to be con-
trolled; no specific gases men-
tioned
PARTICLES
Size
Shape
Chemical composition
Resistivity
Dielectric constant
Charge
Concentration
M
9
M
M
CHARGING SECTION
Type of charging
Ions
Electric field
Geometry
charging of particles is by
corona-type in the preceding ESP
220
-------
Table 32 (continued). PARAMETERS ASSOCIATED WITH THE STUDY OF ELEC-
TRIC CURTAIN AS A DEVICE FOR THE CONTROL AND
REMOVAL OF PARTICULATE MATERIALS
Parameter
Magnitude, description, method of
measurement or control, etc.
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
Charge
Voltage electric field
Particulate loading
Geometry
Internal configuration
metal rods
low
detailed information is given for
a small sized unit
varied loadings will be employed-
AC + DC
M
planar structure, composed of
cylindrical rods placed at dif-
ferent angles to flow in the duct
CLEANING PROCESS
Method
Effect on efficiency
trapping; electrically induced
flow
will be measured in the form of
reentrair.ment study
COMMENTS: Proposed device would have three possible applications:
(1) supplemental device to follow a conventional ESP
(2) replacement for the collection plates in a conventional
ESP
(3) support and containment system for liquid drops in a
scrubber
STAGE OF DEVELOPMENT: Bench-scale testing
221
-------
for pollution control. The electrostatic curtain may be prohibitively
large and expensive as an add-on to a precipitator. The replacement
of the flat electrodes in the electrostatic precipitator with a bank
of rods may well not prevent very substantial reentraintnent, and the
use of electric fields to support scrubbing droplets against a gas
flow is problematic. We shall elaborate on each of these points next.
Using Figure 50, we can set down the equations governing the field
from an electric curtain and gain some insight into its operation. We
obtain the field by taking the gradient of the electrical potential
-V ->-
function (E = v $) once the potential function $ has been obtained from
the appropriate differential equation.
The electrical field distribution can be obtained from the solution
of Laplace's equation for the electrical potential:
32/ay2
z) o = o
DUCT
FLOW
ELECTRIC
CURTAIN
VU.y)
V=0 AT DUCT SURFACES
Figure 50. Electric curtain schematic with coordinates
222
-------
Let be a separable function:
0 = X(x) Y(y) Z(z)
and use primes (') to indicate differentiation.
Then:
X"/X = -a2
Y"/Y = -B2
2 22
Z"/Z = -T = a + B .
2 2
Choosing a and 3 to be positive gives:
±iax ±i0y ± '-2 ' "2
where a and 3 are arbitrary.
If we assume that we have a potential V on'a plane surface, V(x,y),
which surface is at z = 0 and if we assume that the walls are at zero
potential at x = 0, x = a and y = 0, y = b, as in Figure 50, then
for z < 0 (upstream):
X « sin a x
Y a sin $ y
-7 +72
Z « e
satisfy the boundary conditions (including Z -> 0 as z « 0). The full
solution will be:
223
-------
4> (x,y,z) = , A sin (a x) sin (3 Y) e~ZYnra
n,m=l nm n m
where a = mi/a
n
3 = imr/b
m
[2, 2 ^ 2 ..2
Y = irJn /a + m /D
HTTl *
and the A are obtained by solving:
nm
nj=lAnmSin (onx)
a double Fourier series for V(x,y). (For the electric curtain,
V(x,y) might be approximated by adjacent strips at +V.)
The spatial derivative of the potential in z direction is our major
-v
interest, because this is the z-component of E:
E
and here it is exactly,
sin
The first term of this series gives a contribution to the field of
E = A/a2 + 1/b2 exp (- ir //a2 + 1/b2 z)
z v
which (for a = b) is a field that falls off as e~ 1TZ'a from a
maximum at z = 0. (The other components fall off more rapidly in the
224
-------
same e manner.) Thus the field will extend out upstream to dis-
tances on the order of a or b.
If the whole curtain were at one voltage V, then the equation for
becomes:
co 4v /- Y z\
* (x,y,z) = £ sin(mi) sin(mir) sin(ct x) sin($ y) exp I mn )
n,m=l L n m '
nran
and the z-component of the field would be the same series with (-^nm)
multiplying each terra. The major component of this field would be:
. 4V A/a2 + 1/b2 exp (- IT /I/a2 + 1/b2 z)
i.e., a field on the order of the voltage divided by the smaller dimen-
sion, a or b. By using closely spaced rods at alternate potentials, one
achieves fields on the order of the voltage divided by the larger of
the two dimensions, the rod spacing or the rod diameter. Thus the
rod configuration with different voltages will generally give more
intense fields than a curtain with each of its elements at the same
voltage. Having different voltages would produce a net motion toward
one of the rods for a given charged particle, if the voltages were not
chosen to be alternating, as they have been.
It is not intuitively clear, however, that the net result of alternat-
ing fields would be a force away from the electric curtain, which is
needed if the device is to produce a particle motion opposing the gas
flow, or opposing gravity (where scrubbing drops are to be suspended).
To sum up: the DC field would have an approximate magnitude equal to
the voltage divided by the smallest curtain dimension (the rod diam-
eter in the immediate vicinity of the rods, the curtain width or
height at distances comparable to either) and the AC field does not
225
-------
seem able to produce a net force on a charged particle to push it
away from the curtain.
The goal of the electric curtain is to hold up those particles which
have not been caught in an electrostatic precipitator. This means
*
that the curtain must give the particles a migration velocity w
*
(w = q EB) which is greater than the mean velocity of the gas in the
vicinity of the electric curtain:
* * , *
w > v = Q/A
where Q = volume rate of flow,
*
A = cross-sectional area at electric curtain.
The formula for penetration of an electrostatic precipitator has been
given as:
-w A/Q
Pn = e
where w = particle migration velocity,
A = collecting surface area.
Particles which penetrate the precipitator with a one percent penetra-
tion or greater, targets of the electric curtain, are characterized by:
w A/Q < 4.6.
We can combine the inequalities to form the fol]owing two relation-
ships:
(Q/A*) < w* < 4.6 (Q/A)
A/4.6 < A*
226
-------
where we have made the assumption that the field at the electric
curtain is as strong as it is in the precipitator (and it may well be
less, as the designers want to prevent corona in the electric curtain).
Thus the cross-sectional area at the curtain must be at least a fifth
as large as the total surface area of the electrostatic precipitator
plates. A major cost for electrostatic precipitators is construction
cost because of their size. Even if the electric curtain can produce
fields comparable in intensity to those of an electrostatic precipi-
tator, it would seem to require a fairly large additional section in
which the flow is subjected to an expansion so that the gas flow
velocities reach the particle migration velocities. Precisely for
those particles for which the electrostatic precipitator is least
efficient will the electric curtain also have the most difficulty. If
the migration velocity is 10 cm/s (20 ft/min), the curtain will have
22 3
an open area of about 10 m (109 ft ) for every m /s (2120 cfm) volume
flow rate. This is a relatively large structure for such flows, and
its cost would be expected to be relatively high.
We consider next its application as a scrubber. If, as proposed, the
curtain uses rods 2.5 cm in diameter at a voltage of 30 kV, then the
maximum field will be 24 kV/cm at the surface (80 stat-volt/cra) which
would be above the breakdown field for corona discharge, generally.
Assuming a field of 10 kV/cm and water droplets 100 urn in diameter we
can calculate the droplet charge necessary to have the electric field
offset gravity for d = 100 ym:
q = mg/E
= (103 kg/m3) (Ti/6) (10"4 m)3 (9.8 m/s2)/(104 V/cm)
= 5.1 x 10~13 coul
= 3.2 x 10 elementary charges.
227
-------
This charge could easily be put on the droplets using a corona or an
inductive nozzle method. The sedimentation velocity in the absence
of the field would be about 25 cm/s. Equivalently, the droplets could
be supported against a flow velocity ~ 25 cm/s.
Another consideration is whether the electrical force will be sufficient
to stop the droplets, assuming they start with the velocity of 25 cm/s.
Using a Stokes law approximation to the drag force and assuming an
electric field which is homogenous over the distance of interest, it
will take 0.75 cm to stop 100 pm diameter water drops. This distance
will vary inversely with the force, linearly with the velocity and with
the square of the droplet size. The electric field, even if strong right
near the curtain, will have to extend for at least such a distance
upstream into the flow, if the curtain is used across the flow. If the
curtain is used parallel, the force field would have to extend most of
the way through the duct, and, as we show, it is expected to fall off
more rapidly than e"Z , where z is distance perpendicular to the
curtain and a is a dimension characteristic of the curtain.
What is the magnitude of the force parallel to the curtain in the
traveling wave mode? This is crucial to the analysis of its use in
both situations in which the surface is to be placeu parallel to the
flow. At least a dimensionless group for this electric force and the
fluid resistance should be derived and evaluated.
If the curtain is used perpendicular to the flow, then it may well
present an appreciable pressure resistance, especially in a scrubber
mode, where it will be desired to get relatively large velocities
between air and held-up droplets in order to give them appreciable
efficiency for capturing the particulate material by impaction,
usually the predominant capture mechanism in such cases. This resist-
ance may be a significant power drain.
228
-------
]f the droplets are to be repelled by the rods, then these two will
have the same polarity. Then a choice must be made: if the particles
also are charged to the same sign, droplet capture efficiency will be
:educed; if the particles have the opposite sign, then they will
collect on the rods as well as on the drops and this may create clean-
ing, arcing problems. The decision is not trivial. In general, it
seems quite difficult to work with high voltages in a spray environ-
ment without getting short-circuiting.
The idea of making something like a fluidized bed using electrical
forces rather than gravitation and water rather than solid collectors
is an interesting one and deserves to be explored.
Finally, let us look at the idea of using the curtain in place of the
collection plates of an electrostatic precipitator. The major advantage
would be to overcome the difficulty of cleaning the precipitator plates
once they have collected appreciable particulate material. This is
reflected in such problems as back-corona and reentrainment, often
especially problematic in the collection of high resistivity dusts.
Where will the dust actually collect? The high potential at the
corona wires will produce a field toward the curtain and toward the
walls behind the curtain. Although it is conceivable that the fields
can be arranged to make the dust collection bin the lowest potential
surface, there can be expected to be particle capture by the rods and
the walls behind the rods due to particle inertia, local fields, etc.
Tie traveling wave cannot be made to have an amplitude which overwhelms
t.ie DC level or there will be no net motion from the corona to the
curtain. Once again, an important question will be the magnitude of
the electric force produced by the traveling wave in comparison with
fluid forces. The method may be feasible, but there are evidently
several aspects which will pose problems. Once the particles reach the
collection area they will tend to repel other particles unless the
229
-------
charge is conducted away, which is precisely the same problem as the
precipitator plates usually have. The curtain will present an aero-
dynamically rough surface to the flow, inducing turbulence which will
be especially strong in the vicinity of the collection zone (the
rods). The space between the rods and the walls of the device will
have to be kept much smaller than the distance between the two
curtains, or a substantial amount of "sneakage" will occur, the flow
of particulate material in areas without significant collecting fields.
Status - Although the electric curtain has been fabricated, no experi-
mental results have been reported in any formal publication.
Implications - If the electric curtain placed immediately downstream
from an electrostatic precipitator is shown to have promise, it could
be implemented as a retro-fit, although a bulky one. If either the
scrubber configuration or the electrode configuration were successful,
they would probably be made part of new installations rather than
existing ones.
Summary
There are three proposed control uses of the planar assembly of charged
rods referred to as the electric curtain:
o A screen to remove particles downstream from an electro-
static precipitator.
» A support to keep scrubber droplets retarded with respect
to particle-laden air flow.
A replacement for the flat collecting electrodes now used
in electrostatic precipitators.
The proposed theoretical and experimental work will consider the major
relevant aspects of the problems: particle size, charge, composition;
gas flow, temperature, and composition; collector geometry voltages and
electrical characteristics.
230
-------
The electrostatic curtain's prospects arc modest. Used downstream
from a precipitator, it will be trying to control particles by
electrical forces in competition with viscous forces, and the particles
will be those for which a similar process in the electrostatic precipi-
':ator has been insufficient. As a scrubber modification, it should be
able to retard droplet motion and thus improve collection efficiency,
but the droplets and the curtain will produce a greater pressure drop
than did the droplets without this modification; electrical insulation
nay prove very difficult. As a collecting electrode for a precipitator,
It may reduce reentrainment and sneakage, although it may do just the
opposite, depending upon the details of the geometry, the electrical
field and the ratio of the fluid resistance forces to the electrical
iorce parallel to the curtain, and the induced turbulence.
It will be worthwhile to test the various possible uses, however, and
&uch tests will be more informative if theoretical analyses are carried
out as well and used to guide the testing.
FINE PARTICLE CHARGING DEVELOPMENT
A paper detailing the theoretical methods and results has just been pub-
g
lished by Smith and McPonald (1975). We have been informed that the
literature review, the development of a new charging theory, its compari-
son with work by other investigators, and the assembling of experimental
test equipment have been completed at Southern Research Institute (SRI).
Goals
The aim is to increase understanding of particle charging and to improve
particle collection in electrostatic precipitators by raising the average
charge per particle and thus raising the migration velocity. Quoting
fcom a summary of the SRI work:
Z31
-------
"This joint theoretical-experimental study has three important objectives:
(1) to develop an adequate theory for charging of fine particles in a
unipolar ion field with an applied electric field; (2) to supplement
the existing experimental data on fine particle charging; and (3) to de-
sign and build a pilot scale charging device to investigate the tech-
nical and economic feasibility of improved collection of high resistivity
dust by using a precharging section in conjunction with a high field,
low current density, precipitator.
"Aerosols will be generated having diameters from 0.01 to 10 urn, and the
charging rate measured for a variety of charging conditions. Variables
are: particle diameter and dielectric constant, electric field intensity,
ion polarity, and gas constituency and temperature. Theoretical studies
will be performed in an effort to adequately describe the experimental
results."
Methods
Theoretical - The SRI scientists began by reviewing the literature in
which two major types of charging theory have evolved: diffusion charg-
ing, for which the driving mechanism is the ion gradient between the
particle surface and the gas, and field charging, for which the driving
force is the applied electric field. (In Section IV we presented the
results of a field charging equation due to Cochet.) Figure 51 shows the
electronic charges picked up by particles due to the two mechanisms,
based upon White's book, cited by them. (The product of the ion concen-
7 ^
tration and the time was assumed to be 10 s/cm and the particles
were assumed conductive.) For particles with diameters c. 10" ym, the
contribution of diffusion charging can rival that of field charging in
typical coronas. Moreover, there is an interaction between the two me-
chanisms, so the resulting rate of charging is not just the sum of the
two. Improved equations for charging were the goal of the theoretical
Q
analysis. Figure 52 is from the work of Smith and McDonald. For a spe-
cific particle size (0.92 \im diameter) and applied electric field
232
-------
1000
500
FIELD CHARGING, 3000 V/cm
cr
UJ
2
UJ
_J
UJ
UJ
ID
01
<
I
o
100
50
DIFFUSION
CHARGING
0
5
FIELD CHARGING, lOOOV/cm
0.02 O.I 02 1.0 2.0
PARTICLE DIAMETER>/im
10 20
Figure 51. Field and diffusional charging of
small particles-*
233
-------
Z AXIS
0 = 0
B-TT/2
C0=900kV/m
a =0.46/tm
n=!60
n =285
Figure 52. Model for mathematical treatment of
charging rate
234
-------
(9 kV/cm), they have calculated the field surrounding the conductive par-
ticle after while it is acquiring charge (here, 160 electronic units of
a saturation value of 285 units). The dotted line is the locus of points
Eor which the resultant electric field has a zero radial component. Their
analysis breaks the charging process up into three regions one (8 < 9 ) for
o
which a field charging equation applies, one (9 < ir/2) for which diffusion
charging applies and one of which a hybrid equation applies. They used a
computer program to calculate the number of charges as a function of time
ior a given ambient ion concentration.
Lxperimental - The material available to us at this time do not indicate
details of the work, but clearly particle charging is to be measured as
a function of particle size, which means using aerosol sizing techniques
with, possibly, the generation of monodisperse aerosols in the size range
of interest. The particle dielectric constant is to be varied; it would
be known from the chemical composition of the material (for example,
dioctylphthalate, OOP, with a dielectric constant of 5.1 and metal fumes
with infinite dielectric constants). Knowing the electric field intensity
involves the knowledge of corona voltage and geometry; the ion polarity
will be known from the corona current. Finally, they will measure gas
composition (suggesting the addition of water and perhaps SO.), gas
temperature, and presumably, pressure.
Table 33 contains what we infer about the experimental plans. N.A. is
used for "not applicable."
Results
o
Theoretical - Figure 53 comes from the publication by Smith and McDonald
or SRI. It compares the model they developed with various theories and
9
with the experimental data obtained by Hewitt. * The data are closer to
the SRI theory than they are either to a theory proposed by Liu and
or field charging or diffusion charging or the sum of field and diffusion
charging. In general, though, the difference between this SRI theory
235
-------
Table 33. PARAMETERS ASSOCIATED WITH THE STUDY OF PRECHARG1NG
CHAMBERS
Parameter
Magnitude, description, method of
measurement or control, etc.
GAS
Volume rate
Velocities
Reynolds number
Geometry
Pressure drop
Temperature
Pressure
Relative humidity
M (to be measured)
K (to be known)
K
K
N.A.
M
M
M
PARTICLES
Size
Shape
Chemical composition
Dielectric constant
Charge
Concentration
0.01 to 10 urn
K
K
K
M
M
CHARGING SECTION
Type
Ion current
Voltage
Geometry
Corona
M
M
K
COLLECTOR
Chemical composition
Resistivity
Dielectric constant
N.A.
N.A.
N.A.
236
-------
Table 33 (continued). PARAMETERS ASSOCIATED WITH THE STUDY OF
PREC11ARGING CIIAMBliRS
Parameter
CHARGE
Voltage, electric field
Loading
Geometry
Internal
External
Efficiency
CLEANING
Method
Efficiency degradation
Magnitude, description, method of
measurement or control, etc.
M
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
N.A.
COMMENTS: Study of charging rates.
STAGE OF DEVELOPMENT: Theory complete, experimental set-up in progre9s.
and the theory which sums diffusion and field charging was less than
about 25 percent in the amount of predicted charge at any time. A num-
ber of other comparisons were made by the investigators, besides that
shown in Figure 53. The SRI analysis was generally better than all those
to which it was compared, for Hewitt's data.
Experimental - When the experimental part of this program is completed.
rates of charging for particles will have been measured and compared with
theory, with the following treated as variables:
o Particle diameter
o Particle dielectric constant
o Electric field intensity
o Ion polarity
e Gas composition
» Gas temperature
237
-------
n SRI
A LIU AND YEN
O FIELD ONLY
-f DIFFUSION ONLY
O FIELD + DIFFUSION
X EXPERIMENTAL
en
LJ
O
o:
u.
o
or
LJ
00
0246 8 10
ION CONCENTRATION x TIME, NUMBER -SEC/M3 xlO13
Figure 53. Comparison of theories and Hewitt's experi-
mental data for 0.28 micron diameter par-
ticles and medium electric field intensity,
E = 3.6 kV/cm
238
-------
ConclusJons
Tlic work by Smith and McDonald on charging theory produced results within
25 percent of Hewitt's data over the range for which it was available
and was generally the most accurate theory of those compared. The theory
was roughly 25 percent higher than the combination of diffusion and
field charging, which calculations can be done without a computer; this
is an improvement, although often the combining of field and diffusion
equations will be sufficiently accurate for design purposes, we believe.
The experimental work will serve to check the charging theory against
a new set of data and may also lead to the development of improved corona
charging section which could be used upstream from a high field, low-
current electrostatic precipitator.
Evaluation
Suitability of Goals - The collection efficiency of an electrostatic
precipitator for a given particle size can be modeled by using an expo-
nential expression with the negative product of the migration velocity and
surface area divided by volume rate of flow as its argument. As the
particle charge increases, the migration velocity will increase,and
the efficiency would be expected to increase. Two other factors
influence the collection efficiency appreciably: the reentrainment of
material which has been captured and the sneakage of some of the flow
through regions of low electric field. A detailed model is presented
by Gooch and Francis (1975) which incorporates these effects. If the
sneakage and reentrainment effects are predominant in the penetration of
a given precipitator, then increasing the effective migration velocity
may not be much help. In general, increasing the particle charge would
soem advantageous.
II: the corona power is used more efficiently in charging particles in
some precharging chamber, then power savings might be expected beyond
any savings due to lowered penetration for existing devices or lower
239
-------
construction costs for new operations with such a precharging chamber.
As noted in Appendix C, however, the work per particle precipitated in-
creases with the magnitude of the applied force, so that high-intensity
field methods will inherently require more power per volume of gas treated,
other things being equal; this may not be much of a practical problem,
however, because a power consumption in electrostatic precipitators is
low compared with scrubbers.
Suitability of Methods - We analyze next the methods used and proposed.
Theoretical Approach - It remains to be seen whether or not the improved
theory of particle charging will improve corona charging technology.
For the relatively strong fields near a corona, the field charging aspect
of the new theory predominates and this is the same as existing theory.
Still, one of the terms in the rate equation is new and may offer in-
sights, and much can be said in behalf of more accurate theories even
when they do not change, qualitatively, our understanding of the processes
involved. The new theory clearly improved upon the existing ones.
Experimental Approach - From what we infer about the experimentation
to be done, the important variables seem to be measured, controlled,
or known. There is a value to having Hewitt's data confirmed. Further-
more, innovations in the design of the charging chamber may have control
impact.
Applicability to Pollution Control - The theoretical analysis will allow
a more accurate prediction of the particle charging. The experimental
work may lead to improved charging. These would both be applicable to
pollution control by electrostatic precipitation or the electrostatic
augmentation of other types of control device.
Prospects - A now type of charging chamber might be suitable for retrofit
as well as for use in new installations. Where insufficient charging
is a problem, this could be a solution.
-------
Status - The theoretical framework for particle charging is completed and
the experimental equipment is being assembled.
Implications - Improved charging could decrease installation size at the
same collection efficiency thus realizing construction cost savings for
the electrostatic precipitator, for which construction costs are major.
Ihe theory developed in this project seems more accurate than previous
theories. In practice, the use of a simpler theory, field charging
plus diffusion charging, may be as accurate as the accuracy of the
various parameters needed to calculate it justifies. The correct values
for mean ionic mobility and thermal speed seem subject to dispute,
although those used here are the conventional ones. The experimental
work is aimed at investigating the variables relevent to particle charg-
ing and may lead to an improved charging device as well as a confirma-
tion of the data obtained by Hewitt with which the theory was compared.
241
-------
REFERENCES
1. Schultz, M. A., M. E. Crotzcr and W. R. Kiiapick. Collection of
Particulate Matter From Smokestacks Using Gamma - Ray lonization.
Nucl Technol. 15:38, 1973.
2. Dickter, W. and M. A. Schultz. Investigation of a Device Using
Radiation to Charge and Collect Particulate Matter. Nucl
Technol. 12:243, 1971.
3. Leipunskii, et al. The Preparation of Gamma Quanta in Matter.
Translated from Russian. Pergamon Press, 1965.
4. Cooper, D. W. and P. C. Reist. Neutralizing Charged Aerosols
With Radioactive Sources. J Colloid and Interface Sci.
45:17-26, 1973.
5. White, H. J. Industrial Electrostatic Precipitation. Pergamon
Press, New York, 1963.
6. Whitby, K. T. and C. M. Peterson. Electrical Neutralization
and Particle Size Measurement of Dye Aerosols. Ind Eng Chem
Fund, 4:66-72, 1965.
7. Jackson, J. D. Classical Electrodynamics. John Wiley & Sons,
Inc., New York, 1962.
8. Smith, W. B. and J. R. McDonald. Calculation of the Charging
Rate of Fine Particles by Unipolar Ions. J Air Pollut Contr
Assoc. 25:168-172, 1975.
9. Hewitt, G. W. The Charging of Small Parcicles for Electro-
static Precipitation. A1EE Trans. 76:300, 1957.
10. Gooch, J. P. and H. L. Francis. A Theoretically Based Ma-
thematical Model for Calculation of Electrostatic Precipitator
Performance. J Air Pollut Contr Assoc. 25:108-113, 1975.
11. Masuda, S. Charming Spectacle of Charged Particles. Kagaku
Asahi. 103-117, May 1972.
12. Liu, B. Y. H. and H. C. Yeh. On the Theory of Charging of
Aerosol Particles in an Electric Field. J Appl Phys. 39:1592,
1968.
242
-------
SECTION VIII
OTHER ELECTROSTATIC DEVICES
ELECTROSTATIC CYCLONE
lo extend the work done by Molyneux on his proposed combination of
electrostatic and centrifugal collecting mechanisms, a high efficiency
3
cyclone was designed by us to handle 0.472 m /s (1000 cfm) at an effi-
ciency of 50 percent removal of 3 p.m size particles. This basic design
was then modified by the addition of corona charging wires, making cne
body of the cyclone the equivalent of the collector plates in a conven-
tional electrostatic precipitator. The proposed electrostatically aug-
mented cyclone is shown schematically in Figure 54, with the actual di-
mensions given in Table 34.
Goals of the Study
The high efficiency cyclone was designed for the purpose of determining
the potential for improvement in collection efficiencies versus particle
aerodynamic diameter when electrostatic forces are applied within the
cyclone.
Methods of Study
Theoretical - The use of electrostatic forces in a cyclone was demon-
strated by Molyneux in which the particles were charged by corona dis-
charge as in a conventional electrostatic precipitator, and the cyclone
body itself acted as the collecting electrode. The proposed cyclone was
243
-------
Figure 54. Schematically drawn electrostatically-augmented cyclone
244
-------
Table 34. DESICN PARAMETERS FOR A HIGH EFFICIENCY
CYCLONE OF 0.472 m3/sec (1000 cfm)
dpc = 3 ^m '
a = 17.3 cm
b = 6.9 cm
D = 34.5 cm
D = 17.3 cm
e
S = 17.3 cm
h = 51.8 cm
H = 138.0 cm
B = 12.9 cm
AH = 6.38
4 2
AP = 5.98 x 10 dynes/cm or 24 in. HO
2 i
A = Area = 8237.9 cm
C 3
Volume = 81538 cm
V = 3954 cm/sec
t = 0.173 sec
ret 5 3
Q = 4.72 x 10 cm /sec
3
de-signed for 0.47 m" /sec (1000 cfm) using the parameters suggested by
2
Stairmand for a high efficiency cyclone. The actual dimensions were
2
derived from Lapple's equation for the cyclone particle cut size:
d = 3
PC - " V 27TP
h V1/2
Eb \
V Ne/
g /
where:
d =50 percent cut diameter
pc
u - viscosity of gas
ft
b = cyclone inlet dimension
245
-------
p = density of the particle
V = velocity of gas
O
Ne = effective number of turns the gas makes in the cyclone,
using a 3 micron particle diameter cut size the equation was solved for
b, the width of the gas inlet. This value of b was then used with
Stairmand's suggested ratios to determine the remaining cyclone dimen-
sions. Pressure drop was calculated from the Shepherd and Lapple
2
equation:
where K = 16 for a cyclone with a standard tangential inlet, which
gives pressure drop in inlet velocity heads, AH. This AH was converted
to pressure drop using the equation:
V 2 p AH
AP = -8 - g
where AP = pressure drop
V = velocity of gas
5
p = density of the gas.
The interior collecting surface of the cyclone was determined using the
equation:
A = irDh 4- 4 (D + B)
C 2.
(H - h) +
/D-BV
\ 2 /
and the volume of the cyclone was determined using traditional formulas
of solid geometry. The residence time was calculated by dividing the
cyclone volume by the volume throughput. All of the aforementioned
parameters are listed in Table 34. The efficiency of the cyclone was
2
determined using Sproull's equation:
246
-------
n = 1 - exp I-
The migration velocity, w} was calculated from Sproull's equation:
j 2 w 2
d p VT
p p L
w = *- K
r 18 u D
g
for the migration velocity of the particles due to the motion of the
carrier gas. The velocity of the particles due to electrostatic attrac-
tion was determined using the equation:
2 2
3 E dn
P_
3ir u d
8 P
where:
E = electric field strength
d = diameter of the particle
C = Cunningham slip correction factor
y = viscosity of the gas
o
Utilizing this cyclone design, the addition of four corona wires within
the cyclone was investigated to determine the necessary parameters: wire
diameter, corona starting voltage, initial field strength, corona cur-
rent, corona voltage, field at cyclone wall, and power consumption.
The corona wire diameter 0.2 cm, was chosen as a likely wire size from
3
irformation given by White. The critical voltage gradient at the wire
surface for corona onset in air (the initial field -trength) was deter-
mined from White's equation for E in kV/cm:
o
E = 30 f 6 (1 + 0.30 /6/a)
247
-------
where f is the roughness factor, chosen to be 0.5, and 6 is the relative
air density, chosen to be 1.0. The corona starting voltage, the minimum
voltage at which corona will occur, is determined by the equation:
V = a E In -
o o a
where a is the corona wire diameter, and c is the cylinder diameter,
0.1 cm and 5 cm, respectively. The current density, j was determined
using equation:
j = NQ e Z EQ
where:
N = number density of ions
e = unit electron charge
Z = ion mobility, 660 esu .
The factor N , ion density adjacent to the cyclone surface, was determined
utilizing a time charging constant, t , which was chosen to be one-tenth
t T*p C
of the particle residence time. If t = - , then substituting into
the equation:
fco N e Z IT
o
yields:
N -- _
o t e Z it
res
Multiplying the current density at the surface of the cyclone by the area
of the cyclone, A , yields the total corona current i . Dividing the
total corona current by the total length of the four corona wires gives
the current per unit of length for the corona wires, i.
248
-------
The applied potcnlial across the wLrcs was then determined using the
3
equation:
V = V^ + a E_
r
J 1 + (2i/Z) (c2/EQ2 d2 j-
, Jl + (2i/2L) c/E d
1 - In ±^t -
\
2 2 2)
Z.
The electric field at the cyclone wall was determined utilizing the
3
equation:
Ecw
and the anticipated electrical power consumption of the electrostatic
augmentation was determined by:
P = E 1 .
All of the equations utilized are for the geometry associated with a con-
centric corona wire within a cylinder. While the geometry of the cyclone
is not strictly a cylinder, it was felt that these equations would give
reasonable estimates of the corona parameters with the chosen corona wire
arrangement, shown schematically in Figure 54.
Experimental - Since this was strictly a feasibility analysis, it in-
volved only a theoretical evaluation with no experimental work.
Results
The resulting parameters concerning the electrostatic augmentation of the
proposed cyclone are as follows:
d = 0.1 cm
c = 5.0 cm
249
-------
EQ = 29 kV/cm
V = 11 kV
o
A O
j = 1.8 x 10 stataraps/cm
it = 49 mA
L = 414 cm
w
i = 3.5 statamps/cra
V = 51 kV
Ecw = 10 kV/cm
P = 2.5 kW-
The resulting applied voltage of 51 kV j.s much greater than the corona
starting voltage of 11 kV therefore assuring corona production. There
is an upper limit on applied voltage to produce corona, beyond which
3
sparkover occurs. It would appear from information presented by White
that the upper limit for the cyclone in question is approximately 50 kV
or nearly equal to the calculated applied voltage. Since sparkover is
a function of voltage waveform as well as applied voltage, and is also
often tolerated to a controlled extent in practice to take advantage of
using a higher voltage, it was felt that the calculated applied voltage
would be very near the actual voltage used in practice.
The efficiencies of the cyclone, both with and without electrostatic
augmentation, for various sized particles and their associated migration
velocities are listed in Table 35, and depicted graphically in Figure 55.
w is the migration velocity due to inertia, w is the migration
velocity due to electrical forces, and w is the sum of these tvro
migration velocities.
250
-------
NJ
in
O.I
CYCLONE WITHOUT ELECTROSTATIC /
AUGUMENTATION '
CYCLONE WiTH ELECTROSTATIC /
AUGMENTATION
0.3
03 I
PARTICLE DIAMETER IN MICRONS
10
Figure 55. Efficiency versus particle diameter for cyclone with and without electrostatic
augmentation
-------
Table 35. CALCULATED THEORETICAL MIGRATION VELOCITY AND
CORRESPONDING EFFICIENCY FOR THE HIGH EFFICIENCY
CYCLONE WITH AND WITHOUT ELECTROSTATIC AUGMENTATION
Particle
diameter
in microns
0.1
0.3
0.5
1.0
3.0
5.0
10.0
Migration velocity cm/sec
w
0.014
0.126
0.350
1.40
12.6
35.0
140.0
w
e
13.72
22.42
31.85
55.74
151.12
247.3
487.0
Wt
13.74
22.55
32.19
57.14
163.70
282.3
627.0
Efficiency % due to
Inertial
forces
0.02
0.22
0.61
2.41
19.74
45.61
91.25
Inertial &
electrical
forces
21.26
32.45
42.88
62.99
94.21
99.26
99.99
The total energy consumption of the cyclone was calculated from the flow
rate times the pressure drop and the electrical energy consumption of
the corona charging. It was found that the energy consumption due to
mechanical losses, the pressure drop, was 2.8 kilowatts, and the elec-
trical charging required 2.5 kilowatts. Therefore, the total energy
consumption x^ould be 5.3 kilowatts for an electrically augmented
cyclone with a 50 percent cut diameter of 0.7 um.
Conclusions
The increase in efficiency expected from the addition of electrostatic
augmentation is substantial, and as such may seem attractive. However,
the energy requirements are fairly high to achieve the efficiencies
stated in Table 35. The energy rr-quirements for a venturi type scrubber
of efficiency similar to the cyclone with electrostatic augmentation
yielded an energy consumption of 2.45 kilowatts per 1000 cfm (0.47 m /sec),
considerably lower than that for the electrostatically augmented cyclone.
252
-------
A similar comparison between the same capacity commercial cyclone rated
as a high eificicncy unit, reveals a much lower pressure drop, as ex-
pected, with a correspondingly higher 50 percent cut diameter particle.
It would appear that the cyclone is simply not an inherently efficient
device for removing small particles, because it relies upon the movement
of the entire gas stream to be cleaned at high velocities. The accelera-
tion of the entire gas stream, compared to the normally small fraction
of the particles which need to be accelerated, obviously entails a much
higher energy expenditure. It can be noted from the results in Table 35
that the efficiency attributable to the electrostatic augmentation alone
is greater than that attributable strictly to mechanical forces, with the
exception of the 10 |im particles. Since the mechanical energy require-
ments are actually somewhat higher than the electrical energy requirements,
i. is obvious that the electrostatic mechanism is a more energy-efficient
removal mechanism.
Ii: should also be noted that there is a basic conflict between the
cyclonic inertial removal mechanism and the electrostatic attraction
removal mechanism. Increased gas stream velocity tends to increase the
efficiency of cyclones due to the higher inertia of the particles, which
are removed by impaction. Increased gas stream velocity in electrostatic
piecipitation means a decrease in residence time, which decreases both
particle charging and removal efficiency. Obviously, then, there is a
basic conflict between the two mechanisms described which would lead
to poor overall energy efficient particulate removal.
It is interesting to note that the pressure drop across a cyclone is
related to the 50 percent cut diameter by the relationship:
Ap.-p
pc
253
-------
and as such rises dramatically as the particle diameter decreases and
vice versa. (Particle cut diameter is related to the square root of
the gas velocity, and pressure drop to the square of the velocity.)
We also have the relationship between the particle migration velocity
due to inertial forces and the diameter of the cyclone which is, -
assuming the other dimensions change proportionately:
5
from which we may deduce that the migration velocity decreases dramat-
ically with an increase in the cyclone's physical dimensions. This
is especially important when dealing with fine particles. It should be
further noted that the relationship between migration velocity due to
electrical forces and the cyclone diameter is:
W2 Dl
that is, as the cyclone diameter increases, the migration velocity de-
creases linearly. The equation for efficiency is
E = 1 - e - WA/Q .
The area increases with increasing cyclone diameter squared if all
dimensions change proportionately and the expression relating the two is:
A2
therefore the parameter wA/Q changes in proportion to inverse cubed
diameter for geometrically similar cyclones operated at the same flow rate,
using inertial forces, and linearly with diameter, using electrostatic
forces. The efficiency of the cyclone increases substantially due to the
254
-------
electrostatic augmentation as the dimensions are increased, and the
efficiency due to inertial forces decreases dramatically with an increase
in cyclone dimensions. For fine particles the migration velocity due to
inertial forces is negligible, so the overall electrostatic cyclone
efficiency increases with increasing dimensions.
Table 36 contains the results of the migration velocity due to both
mechanical and electrical forces, and the efficiency, for a cyclone of
double the dimensions of that in Table 35. These values can be compared
t.o those in Table 35, to note the decrease in migration velocity and
l:he overall increase in collection efficiency. In theory, as we in-
crease the size of the cyclone and hold the volume throughput constant,
the cyclone approaches a conventional electrostatic precipitator. There
is a basic conflict between efficient cyclone design and efficient
electrostatic precipitator design; we conclude it does not seem promising
to try to augment a cyclone electrostatically, in the above fashion. Use
of space charge repulsion due to having the particles highly charged (to
the same polarity) has not been ruled out, however.
Table 36. CALCULATED THEORETICAL MIGRATION VELOCITY FOR INERTIAL
AND ELECTRICAL FORCES AND PREDICTED EFFICIENCY DUE TO
THE COMBINATION OF FORCES FOR A CYCLONE OF TWICE THE
ORIGINAL CYCLONE DIMENSIONS
.Particle
diameter
in microns
0.1
0.3
0.5
1.0
3.0
5.0
10.0
w
r
(cm/sec)
4.4 x 10~4
3.9 x 10~3
1.1 x 10~2
4.4 x 10~2
0.394
1.090
4.375
w
e
(cm/sec)
6.86
11.21
15.93
27.87
75.56
123.65
247.88
Wt
(cm/sec)
6.86
11.25
15.94
27.91
75.95
124.74
247.88
Efficiency
(%)
38.00
54.40
67.10
85.80
99.50
99.98
99.99
255
-------
Evaluation
Analysis of Theoretical Approach - The initial phase of the study involved
designing a cyclone in sufficient detail to yield results close to those
expected if the cyclone were built and tested.
The second phase of the analysis, the addition of corona charging wires
within the main body of the cyclone involved a less precise or more
approximate approach than that utilized in designing the cyclone. A
major source of possible error in the analysis would occur from the use
of the wire in a cylinder model from which the equations utilized were
derived. Since the cyclone as proposed, with four wires increasing in
proximity to themselves and the cyclone walls to follow the taper of
the cyclone, is not strictly a wire in a cylinder we would not expect
the estimates concerning corona currents, voltages, etc. to be highly
accurate. It seems probable that the approximations which were made
with respect to cyclone geometry are well within the limits of uncer-
tainties of the overall analysis, and that the results as presented here
would approximate the results which would have been obtained with a much
more rigorous analysis.
Analysis of Experimental Approach - As previously stated, there was no
experimental work done.
Prospects of Method - In view of the conclusions drawn from the analysis,
it would appear that there is little or no prospect in pursuing the above
electrostatic augmentation of cyclones. It is highly unlikely that
existing cyclones would be readily amenable to the economical retro-
fitting of the necessary electrical equipment. Secondly, it has been
shown that it would be more advantageous to go directly to a precipitator
design if higher efficiencies than those readily attainable with a
cyclone are required. It appears that the prospects of the method are
severely limited, although space charge precipitation may aid in collection
(see next evaluation).
256
-------
Status of the McLhod - The idea of electrostatically augmenting a cyclone
lias not received widespread acceptance as an emission control device.
The article by Molyncaux describes an electrostatically augmented
cyclone for use on diescl truck exhausts, and beyond this reference
and possible application we have not found significant mention of this
type of device. It would appear that this is a novel device having very
little if any experimental work done to evaluate its performance.
Implications - The electrostatically augmented cyclone suffers from the
combination of two competing collection mechanisms, and does not have any
clearly advantageous area of application. As such, it seems that there
is little likelihood of the further pursuit of this device by people
seeking improved particulate control devices.
ELECTROSTATICALLY AUGMENTED SIEVE PLATE SCRUBBER
1'he Scrubber Handbook contains calculations for the collection efficiency
to be expected for a sieve plate scrubber used to control highly charged
particulate emissions, such as those which would result if the particulate
material passed through a particle charging section upstream from the
scrubber. A schematic of such a system is shown in Figure 56.
Goals
We will estimate the collection efficiency of such a system due to the
electrostatic factor and compare this with the collection efficiency due
4
to inertial impaction, as calculated in the Scrubber Handbook.
Methods and Results
The charge per particle is calculated by the equation for q given in
Section IV. The charging field is 3 kV/cm to coincide with one used in
4
the Scrubber Handbook analysis. The number concentrations are functions
of time; the initial concentration is a parameter. As noted this means
that the penetration, Pn, is given by:
257
-------
SCRUDQING LIQUOR <-
EXHAUST
EMISSIONS
SOURCE
AIR EXHAUST
SIEVE PLATE SCRUBBER
SCRUBBING LIQUOR
INLET
PARTICLE
CHARGING
SECTION
Figure 56. Schematic of possible electrostatically-augmented
sieve plate scrubber
Pn = n/n =
4- 4irBq n t)
For large values of the denominator, the penetntion becomes, effectively,
inversely proportional to the initial number concentration, the residence
time in the bubble and the square of the particle charge. Note that the
only factor which relates to the bubble from the sieve plate is the resi-
dence time, the height of the foam divided by the mean bubble velocity.
It is worth noting that the us« of electrostatic scattering (space charge
repulsion) tends to produce an emission concentration which is independent
of the inlet concentration. As the denominator becomes large (thus small
penetration) the preceding expression can be re-written as
n = 1/ATTBq t
258
-------
showing that the outlet concentration will just be a function of particle
charge and mobility and the residence time in the system.
The Scrubber Handbook indicates that a typical bubble velocity is about
20 cm/s and a typical foam layer height is 10 cm, so the residence time,
t, would be about 1/2 second. Table 37 has the factor ATT Bq 2 calculated
P
from the particle charge and mobility. This factor has the units which
2
are the inverse of the number concentration; l/47TBq is the initial con-
centration which gives a penetration of 0.33 in 0.5 second, the bubble
residence time.
Ln Table 38 we list the penetrations calculated from the above for number
O c
concentrations from 10 to 10 in decade steps. This is also the space
charge effect penetration expected for other scrubbers with 0.5 sec
residence time.
£ substantial collection efficiency is achieved for, say, 1 ^m particles
st 10 cm concentration, but it should be noted that for unit density
particles this would be a mass concentration of 5.2 g/m (2.3 gr/ft3),
higher than industrial emissions often are.
The penetration values can be compared with the penetration or (effi-
ciency) expected for particle collection in the bubble by diffusion and
by impaction, obtaining the following parameters from the Scrubber
Handbook as appropriate for such a scrubber:
Diffusional collection was calculated using the Scrubber Handbook^
equation
n/n = e
o
259
-------
Table 37. PARTICLE PARAMETERS USED TO ESTIMATE SPACE
CHARGE DEPOSITION IN BUBBLES
Particle
diameter, d
(jm)
.1
.3
1.0
3.0
10. 0
Particle
a
mobility , B
(cm/dyne-s)
1.68 x 108
1.95 x 107
5.86 x 106
1.95 x 106
5.86 x 105
Particle
, b
charge , qp
(esu)
2.41 x 10"9
0.84 x 10"8
0.77 x 10"7
0.68 x 10"6
0.75 x 10~5
Factor
2
47TBq
3P
(cm /s)
1.23 x 10"8
1.13 x 10"8
4.37 x 10"7
1.13 x 10"5
4.14 x 10"4
From tables by R. A. Gussman, BGI, Inc., Waltham, Mass.
See Table 1. Here, the charging field is 3 kV/cm = 10 esu.
Table 38. PENETRATION OF SPACE CHARGE SCRUBBER (ASSUMING
0.5 sec RESIDENCE TIME, 3 kV/cm CHARGING FIELD)
Particle
diameter (jitn)
0.1
0.3
1.0
3.0
10.0
n = 105/cm3
o
0.999
0.999
0.979
0.639
0.046
106/cm3
0.994
0.994
0.821
0.150
0.0048
107/cm3
0.942
0.946
0.314
0.017
0.0005
108/ctn3
0.619
0.639
0.044
0.0018
5 x 10"5
260
-------
where h = foam thickness = 10 cm
2
D = parlicle dif fusivity, cm /s
r = bubble diameter, 0.33 cm
v = bubble velocity, 20 cm/s .
The calculated penetrations were 0.944, 0.976 and 0.988 for 0.1, 0.3,
and 1.0 |im particles, respectively, all too high to show conveniently
on Figure 57.
Itnpactive collection was calculated using the Scrubber Handbook, equation
/ -
n/n = e
o
2
where a = 40 F =10
F = foam volume concentration, 0.5
₯ = itnpaction parameter, based upon orifice velocity (1220 cm/s)
and bubble diameter
The results of the calculation of penetration for impactive collection
alone are also sho-vrn in Figure 57.
Conclusions
From the foregoing analysis, there are realistic levels of particle
ccncentration, charge and mobility for which the mechanism of space
charge repulsion (also called electrostatic scattering) is more effec-
tive than impact ion, typically the predominant mechanism for scrubbing.
At high concentrations and fine particle sizes this difference increases,
and these are conditions for which improved scrubbing would be advantageous,
261
-------
o
z:
UJ
o
u.
u_
LJ
I
z.
o
CC
I-
UJ
z.
UJ
QL
SPACE CHARGE
IMPACTION
0.001
= !05cm~3
0.01 -
O.
0.3 I 3
PARTICLE DIAMETER,
Figure 57. Calculated penetrations at 0.1, 0.3, 1.0, 3.0, 10 urn
and linear interpolations
Evaluation
Electrostatic augmentation of seive-plate scrubbers is of possible utility
because it would improve collection efficiency lor fine particles and
would do so most when the concentrations are high. If a simple and
reliable means for charging the particles upstream from the scrubber can
be employed, this might be an attractive way to augment scrubber collec-
tion efficiency with relatively small/power input. It would seem worth-
while to test this at the bench scale and the pilot scale.
262
-------
Summary
Charging the emissions before they enter a sieve-plate scrubber can be
expected to increase their collcctibility, due to space charge repulsion,
and under some reasonable circumstances this effect could outweigh impac-
tion. Experimental investigation of this would seem worthwhile. (This
should also be true of a packed bed scrubber.)
REFERENCES
1. Molyneux, F. Electrostatic Cyclone Separator. Chem Process
Eng. 44:517-519, 1963.
2. Leith, D. and D. Mehta. Cyclone Performance and Design.
Atmos Environ. 7:527-549, 1973.
3. White, H. J. Industrial Electrostatic Precipitation. Pergamon,
New York, 1963.
4. Calvert, S., J. Goldschmid, D. Leith, and D. Mehta. Scrubber
Handbook. Office of Air Programs, Environmental Protection
Agency. Research Triangle Park, N.C. 27711, 1972.
263
-------
SECTION IX
SETTING PRIORITIES
This section presents a model for setting priorities concerning future
work in the area of electrostatic augmentation. We have attempted to
incorporate essential elements necessary for rational decision-making
into the model formulation, even though not all the information required
by the model is available.
ASSUMPTION
One basic assumption is made. The model seeks solutions which have the
minimum cost in comparison to benefits. Factors such as uncertainty are
introduced through the application of probability or discount factors
to costs or benefits.
DEFINITIONS
Advantage
The advantage of electrostatic augmentation is defined as the difference
between the mass of fine particulates removed with the use of an aug-
mented control device and the mass of fine particulates removed by use
of a particulate control device with no electrostatic augmentation, both
having the same mass input. Both of these removal factors are expressed
in the common removal, efficiency fashion - in terms of percentage reduc-
tion. For each application of a given device, the advantage factor is
expressed as follows:
265
-------
A.. = (P - p )
ij ac nc ij
where A.. = advantage of augmentation device i for application j
P = removal efficiency factor expressed as a percentage
ac = augmented control
nc = nonaugmented control.
It is assumed that where P ^ P (A <_ o) , the augmentation technique
nc 3.c
is rejected with no further consideration.
Cost
Two types of cost are considered, capital cost and annualized cost.
Capital cost is defined as the amount of monetary outlay required for
the purchase and installation of a given augmentation device. Capital
costs are significant in that they represent the significance of com-
mitment on the part of the purchaser.
Annualized costs are the yearly expenses associated with operating a
given augmentation device. Included in the annualized cost are fixed
costs (depreciation, insurance, finance charges) and variable costs
(maintenance, power requirements, etc.). Annualized costs are important
since they increase the users' costs of production and thereby influence
the price charged for the good or service produced.
k. = capital cost of augmentation device i
. = annualized cost of augmentation device i
o. =
Applicability
*
The costs of the various electrostatic augmentation devices are not
directly comparable for two main reasons. First, the devices vary as
266
-------
to their likelihood of reaching commercial application. Secondly, they
differ in the scope of their potential applications. Thus, factors whach
take account of these two areas must be derived if objective comparisons
among the various devices are to be made.
Likelihood of Application - The likelihood of commercial application is
related to the stage of development of the augmentation device. It is
assumed that there are four stages of development. For each development
stage the probability of success, the chance Lhat the device will become
commercially available, and the number of years until commercial avail-
ability are estimated. It should be noted that these two items are
assumed to be independent of one another. For example, a given device
may have a high likelihood of becoming commercially available because
it is based on sound theoretical and practical grounds yet the time
needed to work out production kinks and the like may lengthen its
development time.
The four development stages, from the least to the most advance stage,
are listed below.
o Research
e Pilot
e Demonstration
e Commercially available
The probability, or likelihood, of success is designated as follows:
£. = probability of augmentation device i becoming
commercially available.
The number of years until an augmentation device becomes commercially
available is taken into consideration by the application of the standard
*
discounting technique to the advantage factor A..
267
-------
P.V. of A = AI/(! +
where P.V. of A. = present value of the advantage factor
A. = advantage factor of augmented device i
t = number of years until the augmentation device
becomes commercially available
r = rate of discount reflecting the opportunity
cost involved in waiting for augmentation
device i to become commercially available.
(The inflation-corrected interest rate can
can be used as an approximation of the
"true" discount rate).
Potential Applicability - The potential applicability of a given elec-
trostatic augmentation device must also be taken into consideration. It
is assumed that devices with a widespread potential applicability are
preferable to those devices with a limited number of possible applica-
tions. The potential applicability is expressed as follows:
a. - g. M.
where a. = mass emissions for which augmentation device i
would apply.
g. = likelihood that device i would have its
assumed application.
M. = mass emissions from all control devices for which
device i is applicable.
Cost Normalized for Benefits
The information previously discussed is used to normalize costs in order
to make possible objective comparisons among the various electrostatic
augmentation devices. Costs are normalized as 'follows:
268
-------
where K.. = normalized capital cost of augmentation device i
= reciprocal of likelihood factor
k. = nominal capital cost of augmentation device i
-\ = reciprocal of discounted advantage factor for
I augmentation device i
= reciprocal of the applicability factor for
i device i .
As probability, applicability, and the advantage factor increase,
normalized capital cost declines. Increases in normalized capital cost
are caused by increases in the nominal capital cost, the time until
commercial availability and the discount rate.
Treatment oE Annualized Costs
Once the normaliz°d capital cost has been calculated, annualized cost
can be determined This is accomplished as follows:
Fixed Costs - Fixed costs include depreciation, insurance, and finance
charges. Once the lifetime of the augmentation device is known, the
straight line depreciation technique can be applied to determine annual
depreciation charges. Finance charges, if any, can be amortized in the
same manner. Insurance charges are computed as a percentage of total
capital investment.
Variable Costs - Variable costs include labor, power requirements, and
naintennnce. For these costs, estimates from the manufacturer may be
the best source of information.
269
-------
Summntion - The summation of fixed and variable costs is the
cost, o^ This can be expressed in terms of per unit benefit, just as
was done for the capital cost k., by replacing k. with o. in the
preceding equation.
CONCLUSIONS
A method for setting priorities for research and development with respect
to control device technologies such as electrostatic augmentation has
been formulated. The devices with the lowest annualized cost per benefit
should be given highest priority. This model requires the following
inputs:
Efficiency of the augmented device and the unaugmented
device, if any.
o Capital cost of the augmented device.
c Annualized cost of the augmented device.
« Likelihood the device will ever become commercially available.
e Estimated time until commercially available (related to
stage of development).
6 Mass emissions of sources for which the device would be
applicable.
o Likelihood it would be applied to such sources once available.
o Inflation-adjusted interest rate.
This model is an attempt to build a methodology for setting priorities
for investing in research and development. It can be no better than its
assumptions and its data, and some of the information it requires is not
readily available or may always be a matter of judgment. The fact that
such a model exists may provide impetus for the collection of the kind
of information it requires. Even x^ith incomplete information, it may
allow ordcr-of-magnitude cost/benefit estimates which will show certain
investments to be clearly more advantageous than others.
270
-------
SECTION X
SOME RESEARCH POSSIBILITIES
INTRODUCTION
In this section, we will point out some areas relating to electrostatics
and aerosols which might fruitfully be investigated further.
RESEARCH POSSIBILITIES BY RESEARCH CATEGORY
One categorization of research in air pollution control is:
o Fundamentals
o Unit mechanisms
e Control systems
9 Control systems applications
e Comparison of control systems
In reviewing the work done with respect to electrostatic augmentation,
we have found possible research topics in each of these categories.
Fundamentals
e What values are correct for mean ionic thermal speed and
mobility ?
How do these vary as functions of gas composition?
What factors govern rate of charging and equilibrium charge
level on:
scrubber droplets,
271
-------
- filter materials such as teflon, polypropylene, and
others at the extremes of the triboclectric series,
- bed packing materials?
Unit Mechanisms
Can a constant-concentration aerosol generator be built using
the space charge effect to dampen concentration variations ?
Recall:
Pn = n/nQ = 1/(1 + 4 TT qp2 B not)
n = l/47rqp2 B t,
2
for 4irqBnt>>! .
What is the experimental collection efficiency of charged
drops when inertial forces are negligible? (Recall that the
electrostatic droplet scrubber collection efficiency de-
creased with particle size even though the predicted col-
lection efficiencies increased.)
How does particle charge interact with wettability in
scrubbing? (It has been argued that particle charge com-
pletely dominates poor wettability.)
« Can particles be more readily charged through contact with
or close approach to charged droplets than in a comparable
conventional corona discharge?
0 How can one design a trouble-free pre-charging section to
be used to enhance collection in scrubbers, cyclones,
packed beds, etc., through space charge repulsion (electro-
static scattering)1' Must field charging be used or would
diffusion charging suffice to increase the collection of
i the fine particle fraction?
272
-------
Control Systems
Can precharging the particles produce marked improvements
in fine particle collection in foam scrubbers or packed
beds?
What are the operating differences between charged droplet
scrubbers of the same polarity and of opposite polarity
with respect to the aerosol to be collected?
What are the trade-offs between efficiency (as function of
particle size), power consumption, and residence time
(thus capital costs) for charged droplet scrubbers?
e How can open-structure filters be cleaned once their
efficiency has been enchanced electrostatically?
e Can disposable filters have their fine particle collection
efficiency substantially enhanced by superimposing an
electric field parallel to the air flow?
e How does the use of wetted surfaces affect cleaning and dust
resistance problems in electrostatic precipitators?
c How can sneakage and reentrainment be minimized in electro-
static precipitators?
Applications
Which applications areas (such as acid mist) are most dif-
ficult to achieving high efficiency control9 What arc the
electrical characteristics of the aerosols and gases in-
volved? How might electrostatic augmentation be achieved?
Comparisons
What are the cost/benefit factors in the use of charged
droplet scrubbers?
What would be the cost/benefit comparison for adding pre-
charging sections to augment control devices by adding to
the particle collection through space charge repulsion?
(This will depend upon the cost of such a system, the
aerosol concentration and size distribution, precharger
design, and the system residence time, including the time
spent in ducting after the precharger.)
273
-------
CONTROL SYSTEMS
The particle pollution control systems all have the following features:
some expenditure of resources (materials, labor, power) is used to
remove particulate material from a gas stream and transport this ma-
terial elsewhere. The studies discussed in this document have had
the following general approaches to electrostatic augmentation as
a means to lessen resource expenditures:
improve particle charging (raise levels, charge inexpensively)
add electrostatic iorces to augment collection by impaction,
interception, diffusion, sedimentation, diffusiophoresis, etc.
use electrostatic forces to change system geometry (increased
porosity in filter cake through electrostatic repulsion,
increased residence time for scrubbing droplets with the
electrostatic curtain).
By comparing such approaches with the mechanisms operating in the con-
ventional control methods, one can note some attractive possibilities.
Table 39 lists widely-used particle collection devices and the pri-
mary and secondary mechanisms they employ to achieve gas/particle
separation. We will discuss some possible applications of electro-
statics device-by-device.
Settling chambers are normally used for the coarsest aerosols, pro-
ducing gas/particle separation by the settling of particles under
the force of gravity. This is favored by large particles and long
residence times. Because large particles will take high charges
and because the residence times are long, precharging the particles
before they enter the settling chamber may appreciably increase col-
lection efficiency, due to the space charge (electrostatic scattering)
effect. The chamber should be grounded in any such application.
274
-------
Table 39. WIDELY-USED CONTROL DEVICES AND PARTICLE REMOVAL MECHANISMS
Removal
mechanisms
sedimentation
centrifugation
impaction
interception
diffusion
electrostatic
diffusio-
phores is
thermophoresis
photophoresis
vaporization
combustion
turbulent
deposition
sonic migration,
oscillation
Control devices
60
C )-i
H 0)
!"H . ***!
[_l C2
W CO
oi x:
to u
P
a
c
o
,-H
y
0
P
J-i
0)
M
C
H
a,
E
H
P
S
Q)
.S
0)
en
S
P
T3
Q)
,*
O T>
(0 0)
P
P
P
V-i
(U
4-J
^-1
fa
P
P
P
S
u
u o
nj 4J
4J P)
W AJ
O i-l
!-< D-
U *r-J
U 0
-------
Cyclones use high velocities and relatively short residence times.
We have analyzed one possible method of adding electrostatics to
enhance collection (Section VI ). The relative influence of electro-
static forces would increase for cyclones with slower gas velocities,
other things being equal, so that electrostatic augmentation, through
the use of an applied electric field within the cyclone or through
space charge precipitation, would have a greater impact on cyclones
with fairly large cut diameters rather than those with substantial
efficiencies in the fine particle range.
Impingers also use relatively large gas velocities, velocities it is
difficult to match with electrostatic forces.
Packed beds generally use face velocities substantially slower than
those in venturi scrubbers or cyclones, so that the addition of
electrostatic collection through space charge precipitation might be
quite advantageous. A conductive system (water-washed, for example)
would allow collection of precharged particles without the build-up
of a field opposing such deposition. It is hard to superimpose a
strong external field across the bed, however, because of the typical
dimensions involved.
Filters which build up a charge may achieve enhanced collection by
Coulomb attraction of oppositely charged particles. Even uncharged
particles, as shown, will be appreciably attracted by the charged
fibers (see Section III discussion on image force between charged
collector and uncharged particle.) Penney and Frederick at Carnegie-
Mellon are, as discussed, also following up the intriguing possibi-
lity that particle charge may be used to lox^er filter cake resis-
tance without lessening collection efficiency. The low face velo-
cities ('»- 1 cm/s) suggest electrostatic forces can have important
contributions to collection. The large structures and long resi-
dence times in fabric filtration also suggest that precharging the
276
-------
particulatc matter could produce substantial space charge precipi-
tation within the baghouse, a possibility worth pursuing. (Again,
the structure should be grounded.) Because the cleaning and resis-
tance characteristics of the dust/fabric system are so important
and because the electrostatic forces are known to be important in
adhesion and cake formation, work in this area, such as that by
Penney and Frederick, is quite promising.
Electrostatic prec ipitation could be furthered by improved particle
charging (note the possibility of droplet-particle charge transfer
and the questions surrounding ionic mobility and mean thermal speed)
and by methods for preventing the flow of gas through areas of low
electric field strength. The work on particle charging might well
be coordinated with the work on resistivity, because the addition
of charge to highly resistive particles may only exacerbate the
problems posed by such an aerosol.
Spray scrubbers (including venturi scrubbers) should be able to
improve their performance by the application of electrostatic forces
between the droplets and the particles as well as the particle-
particle repulsion due to space charge. Design of such systems
should take into accourt the time scales emphasized by Melcher and
Sachar (see Section V), to assure that the collecting droplets
are present for times which are long compared to the characteristic
time for particle-droplet collection. At present, mutual repulsion
by droplets and particles of the same sign seems promising, as does
the collection of particles by droplets of opposite sign (as inves-
tigated by Pilat and his colleagues at the University of Washington).
In passing, it can be observed that combustion and flame behavior
has been found to be sensitive to electrostatic,fields, so that
there may even be a role for electrostatics in enhanced incineration.
277
-------
Finally, as we have noted in Section VIII, the fo.im scrubber would be
a logical candidate for electrostatic augmentation by precharging the
particulate matter and using space charge repulsion. The relatively
slow flow velocities and long residence times would help and the im-
provement should be substantial in the fine particle fraction, where
much current interest centers (see Section VI for more details).
In general, it would be advantageous for those conducting experimental
investigations into electrostatic augmentation to indicate the ranges
and values of the parameters listed in Table 33, where applicable.
This would allow other researchers to analyze the results more readily
and would assure that important parameters are not being unmeasured
or unrecorded.
SUMMARY
Possible courses of investigation have been outlined in this section,
along with some data reporting suggestions. The section on setting prio-
rities might be useful in judging between the possibilities presented here,
even if the estimates which are entered into the methodology are only
rough approximations. It is hoped that this work will stimulate the
research into applying £n inherently energy-efficient means, electro-
static forces, in removing undesirable particulate material from the
air.
278
-------
APPENDIX A
INTRINSIC POWER REQUIREtffiNTS FOR DUST REMOVAL
The work done (W) in removing a particle from a gas stream is the inte-
gral of the fluid resistance force (F ) during the particle motion and
the path length (ds):
f S2
W = / * F (s)ds .
S r
The power (P) is a similar integral, involving the force and the veloc-
ity, v(s):
fS2
= J s Fr(s)v(s)ds .
In both integrals, the total path length is L = S~ - S1. For times
fi
which are long compared to the particle relaxation time (= 3.6 x 10 s
for 1 um diameter particle), the particle velocity is the terminal ve-
locity, given by:
v(s) = F C/3iryd
e p
in those instances where Stokes law applies (particle Reynolds number
much smaller than one),
" For a very different approach using thermodynamics, see: Soo, S.L.,
Environ. Sci. and Tcchnol. ]_:(>3 (1973).
279
-------
where F = applied force = F
e r
d = particle diameter
P = fluid viscosity
Cunningham corre
at STP for air.)
C = Cunningham correction factor (= 1 + 0.16 x 10 cm/d
The force may be constant or it may change with position. If it is
governed by a power law,
csn,
the analysis is simplified (n = 0 is the constant force situation)
The work and power integrals become:
W = c(S2n+1 - S/Cn + 1), n t - 1
f °2 2
P = J c (3iryd_/C)v (s)ds
P = f c2 (C/3iryd )F2(s)ds
> .3- D
P = (C/37rydp) c2 s - S /(2n + 1) , n
'n the homogenous force field, n = 0, these integrals reduce to:
280
-------
W = F S = (3Trud v/C)S
r P
P = F v = (3Tryd v2/C) .
r P
2
We can average over S and v by using the definition of the average of
a quantity x for the particles, which is
r x
/ max er \A
= J x f(x)dx
*» i
mm
where f(x) is the fraction of particles having the value in the range
x to x + dx.
The work in removing particles always increases as the mean particle-
to-collector distance, S increases. The integrals for power show a
very marked increase in power consumption as the particle mean squared
velocity, v^, increases.
A simple formula for penetration, Pn, the ratio of outlet to inlet
particle concentrations for a control device, is
Pn =
where A = collection area perpendicular to particle migration
velocity, v,
Q = volume rate of flow.
In a homogeneous force field (or for a field adequately represented
by using an average force):
v = P^(C/3irud )**
P
281
-------
This can be substituted into the exponential expression to give the
power required to achieve a given penetration:
-p
Pn = e
For 95 percent efficiency (Pn = 0.05), one needs
u
(A/Q) = 3
or P = 9(Q/A)2/(C/3irUd )
P
The ratio of flow rate to collectirg area, Q/A, can be concerted to the
ratio of control device volume to collecting area times the average lin-
ear- gas velocity divided by the flow path length:
Q/A = (V/A) (U/L) .
This is also just (V/A)/t ; for the residence time, t . For an
J res' ' res
electrostatic precipitator with V/A = 25 cm = 0.25 m, and t =1 sec,
1TGS
the power needed for Pn = 0.05 for one 1 pm particle would be
P = 9(0.25 m/s)2/(6.8 x 109 s/kg)
P = 8.3 x 10~11(kg - m/s2)(m/s)
P = 8.3 x 10~U W -
3 12 3
A particle concentration of 1 g/m would be 1.9 x 10 /m number con-
centration if the mass mean diameter of the aerosol were 1 pm and the
3
particle density that of water. At a 4.7 m /s (10,000 cfm) flow rate,
282
-------
3
the prccipitator volume would be 4.7 m (1 second residence time) and
the total power consumption:
P = (8.3 x 10 11 W)(1.9 x 1012/m3)(4.7 m3)
= 741 W.
Soo (1973) indicates typical actual power consumptions about five times
this value for precipitators and 50 times this value for high-energy
scrubbers. Our theoretical value, though low, is more than an order
of magnitude higher than Soo's theoretical values and seems an improve-
ment on his approach.
When the device volume, V, the mass concentration, m, and the mass mean
diameter, d 3, are used in the expression for power, one obtains
(9 Q2/A2)(3Tryd~/C).
From this equation, it is clear that the following factors will increase
intrinsic power requirements:
1. Increased mass concentration for the same
size distribution;
2. Decreased mean size for the same mass
concentration;
3. Increased volume for the same ratio of
flow rate to collection area; and
4. Increased flow rate for the same geometry.
Because penetration is related to the square root of power through an
exponential relationship (given above) , to change the penetration from
283
-------
-1 -3
e (0.37) to e (0.05), for example, can be done by increasing the
ratio of the collecting area to the flow by a factor of 3, or the power
per particle by a factor of 9.
This appendix is just a preliminary analysis of the intrinsic power
consumption question. The power so estimated is expected to be less
than that usually expended by control devices because the analysis
assumes that all power goes to the collection of the particulate mate-
rial, although in fact much of it may go to gas/collector flow resis-
tance as well. This kind of analysis should indicate minimum power
requirements and should suggest ways in which control devices in the
future can approach these minimum values. Electrostatic augmentation
of control device efficiency is attractive because the electrostatic
forces are applied directly to the particles, rather than indirectly
as is done with methods which rely on particle inertia, such as scrubbers
or cyclones.
284
-------
APPENDIX B
INSULATOR PARTICLES CAN BEHAVE AS CONDUCTORS
Fuchs notes that experimental work with oil and mercury droplets
showed that the oil droplets behaved as though they were conductors
(X -* 1). The following calculation indicates why this is so. An un-
charged cubical particle (L x L x L) aligned with an electric field (E)
perpendicular to one set of faces would, if a conductor, have charges
+q and -q migrate to the faces to offset the electric field's potential
difference, EL = V The current would be
I = V/R = 2q/At
i
where R = resistance
V = voltage difference
At = time to isach equilibrium.
An insulator is just a poor conductor, requiring a much larger At than
the conductor. The resistance, R, is given in general by
R = pL/A = pL/L2 = p/L
where p = material resistivity
A = cross-sectional area .
285
-------
Then,
At = 2q R/v
= 2q p/LV
The final charge will be such that
V = 2q/L
so,
At = p.
The time is equal to the resistivity, p, which is given in ohm-cm or
sec, because
1 sec = 9 x 10 ohm-cm.
A list of some substances and their resistivities (thus, charge equilib-
ration times) is given in Table 40.
Table 40. SOME SUBSTANCES AND THEIR RESISTIVITIES
Substance
Glass
NaCl
Si
C
Cu
Resistivity, p
(ohm-cm)
~io13
-io9
~io4
- io-2
-io-6
Charge equilibrium time
(s)
io1
io-3
io-8
io-14
io-18
286
-------
_3
Particles travelling ~ 1 cm/s would spend ~ 10 s in the vicinity of
_3
a collector of L ~ 10 yin, and 10 s would be sufficient time for polari-
zation for particles less resistive than pure Nad, so that such particles
would act as conductors.
Particles with adsorbed water will act as conductors, too, even if
they are made of highly insulating material. Thus, most particles
will be charged collected as though they were conductive.
REFERENCES
1. Fuchs, N. A. Mechanics of Aerosols. Pergamon, New York. 1964.
2. Jackson, J. D. Classical Electrodynamics. John Wiley & Sons,
Inc. New York. 1962.
287
-------
APPENDIX C
NOTES ON EXPONENTIAL PENETRATION FORMULAE
A generalized collection configuration is shox^n in Figure 58. Col-
lecting surfaces each having a total area, A , and an area normal to
S
flow, A , are present in a collector which itself has cross-sectional
area, A , and mean face velocity, v (also called the free stream
velocity), which velocity is assumed to have the same mean throughout
the collection device, for simplicity.
PARTICLE
w
vo « \1/ COLLECTORS Ao
Figure 58. Model for particle collection by obstacles
Two frequently-used forms of penetration equations are
_ - -w(ZA-)/v A
Tn = e s x = e s o o
239
-------
where Q is the volume flow rate, A the total collector surface, and
s
_ -n n A L
Pn = e c c
where n A is the cross-sectional area of collectors per unit volume
c c
and L is the length (parallel to the mean flow) of the collection sec-
tion, n is the single collector efficiency.
The first expression can be obtained by equating the change in number
concentration in a volume to the number per volume reaching surface,
A , at perpendicular velocity, w:
S
dn
V = W + V
P g
w = velocity with respect
to gas
v = gas velocity.
O
Because
vdA = <$> (V-v)dV
S
by Gauss's Theorem (dV is the volume element), and because
V-v = 0
g
for incompressible gas flow (Mach «1), then
290
-------
dn
1 r -
= - - p ru/
where the migration velocity is perpendicular to the collecting surface.
Thus,
-w(ZA )t/V
n = e s
n = e-w(ZAs)/(>.
The other formula is derived similarly, assuming that each collector
removes nA /A of the particles approaching it from a great distance.
The connection between these formulae would be
w(ZA )/v A = nn A L
s o o c c
)L/A L
C 0
or
(w/v
O S O CO
Tlius
n = (W/VO)(EAS/ZAC),
291
-------
APPENDIX D
APPROXIMATE CALCULATION OF COLLECTION EFFICIENCY FOR
CENTRAL-FORCE COLLECTOR
In Figure 59 is given the geometry for the following discussion of col-
lection of an aerosol particle by a collector which produces a central
force given by the equation:
_ , , n
Fr = k/r
in which r is the radial distance from the center of the collector, k
is a constant of proportionality, and n is the exponent associated with
the force (Coulomb force would have n=l for a collecting cylinder and
n=2 for a collecting sphere). The collector has radius R. The net
flux into an imaginary surface at R* is just the integrated product of
the net migration velocity of the particle under the force F with the
concentration at that surface and the surface area. We can define R*
such that it is the distance at which this integrated product equals
the product of the free-stream velocity, v , and the free stream con-
centration, N , and the geometrical cross-section of this imaginary
("Gaussian") surface for a spherical collector:
2 2
v N n RVc = w* N 4« R*
o o
We also know that, for negligible inertia, the migration velocity
is the product of F (r = R*) and particle mobility, B:
w* = Bk/(R*)n
293
-------
Figure 59. Geometry for approximate calculation of collector
efficiency for central forces
The single target efficiency is defined as the ratio of the cross-
sectional area swept clean by a collector to the geometrical cross-
sectional area:
E = (R*/R)'
s
for spheres and
E = (R*/R)'
s
for collecting cylinders. Analysts of such problems, such as Kraemer and
Johnstone, usually try to get collection efficiency in terms of the
ratio of the value of the central force at the surface of the collector,
F , to the drag force on the particle, F , at the free stream velocity
R I)
of the gas. For a collecting sphere, and assuming that N = N (an im-
portant assumption), the migration velocity at R* can be put in terms
294
-------
of the free stream velocity, allowing the formulation of an expre&sion
for R*:
w* = v (N /N)/4 = v /4
O 0 0
R* » (fik/w*)1/n = (4Bk/vJ1/n .
Noting that
and
= FD
we have, for a spherical collector
E = (R*/R)2
s
= [(4 Bk/vQ Rn)1/nJ
For cylindrical collectors, the ratio of surface area to cross-sectional
area is IT rather than 4 and the exponent becomes 1/n rather than 2/n.
The efficiencies thus calculated match those of Kraemer and Johns tone
for the Coulomb force (for which N = N exactly) with spherical and
cylindrical collectors and have Lhc same exponent and nearly the same
coefficient as the efficiency expressions gLvcfi by Kraemer and
Johns tone for image force collection on spheres and cylinders. The Kraemer
and Johnstonc calculations required assuming specific flow velocity
profiles, whereas this approach clearly does not.
295
-------
Actually, the collection efficiency approach masks the physics: the
size of the collector is not really determining the collection due to
electrostatic interaction...instead it is the size of R* which indicates
how far from the center of the collector is the central force effective
in cleaning:
R*«
-------
TECHNICAL REPORT DATA
/ricase read l> usiictiiin\ o>i the n imc bLjuie Li
REPORT NO
EPA-600/2..76-055
3 RECIIMtNT'S ACCbSSIOONO
1ITLL *NU SUCT'lTLE
Evaluation of Electrostatic Augmentation for Fine
Particle Control
0 REPORT DATE
March 1976
0 PERFORMING ORGANIZATION CODE
ALH HOR(S)
D.W. Cooper and M.T. Rei
8 PERFORMING ORGANIZATION REPORT NO
GCA-TR-75-34-G
. PERFORMING ORGANIZATION NAME AND ADDRESS
GCA Corporation
Burlington Road
Bedford, Massachusetts 07130
10 PROGRAM ELEMENT NO
1AB012; ROAP 2IADL-Q29
11 CONTRACT/GRANT MO
68-02-1316, Task 7
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; 10/74-10/75
14 SPONSORING AGENCY CODE
EPA-ORD
15
Ext 2925.
^v NOTES
pro]ect officer for this report is D.C.Drehmel, Mail Drop 61,
16 ABSTRACT ,p|ie rgp0r(; reviews electrostatic augmentation of control devices for fine
particulate- the addition of electrical forces to scrubbing and filtration and the
enhancement of electrostatic precipitation. It gives the major electrostatic force
equations and their evaluation for some reasonable values of particle and collector
charge and geometry. It includes a bibliography on electrostatic augmentation. It
analyzes the following programs on electrostatic augmentation of filters, scrubbers,
and electrostatic precipitators: fiber beds used to capture particles electrostatically,
dust/fabric electrostatic effects, electric fields applied across filters or generated
within filters, a collector using oppositely charged particles and droplets, a charged
droplet scrubber (accelerates droplets electrostatically and uses them to transfer
charge to particles for electrostatic precipitation), various polarities and configu-
rations for charged droplet scrubbing of charged particles, nuclear radiation used to
charge particles for electrostatic precipitation, various configurations and uses for
an 'electric curtain,' and improvement of particle-charging in connection with pre-
charging chambers. Other research in electrostatic augmentation, especially on
filters, is discussed briefly. Analysis of two other possible systems is presented:
an electrostatically augmented cyclone and a foam scrubber that uses particle pre-
c h arging.
17
KEY WORDS AND DOCUMENT ANALYSIS
DCSCRIPTOTS
Air Pollution
Electrostatics
Dust
Electrostatic
Precipitation
Electrostatic
Dust Filters
Scrubbers
Fibers
Fabrics
Nuclear Radiation
Cyclone Separators
Foam
b IDENTIFIERS/OPEN ENDED TERMS
Air Pollution Control
Stationary Sources
Fine Particulate
Electrostatic Augmen-
tation
Charged Droplets
c COSATi Field/Group
13B 13K
20C
11G HE
13H
09C
18H,20H
07A
18 UI3I HIIJUTiON S.PATF.MLNT
Unlimited
10 SfcCUHHY CLASS (1 his Hi
Unclassified
?n .>
EPA tor. \ 2?20-l (3 ,'3)
297
------- |