&EPA
United Stales
Environmental Protection
Agency
Industrial Environmental Research
Labor.r
Research Triangle Park NC 2771 1
EPA 600 7-78-142a
September 1978
Electrostatic Effects in
Fabric Filtration:
Volume I. Fields, Fabrics,
and Particles
(Annotated Data)
Interagency
Energy/Environment
R&D Program Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range of energy-related environ-
mental issues.
EPA REVIEW NOTICE
This report has been reviewed by the participating Federal Agencies, and approved
for publication. Mention of trade names or commercial products does not con
stitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service. Springfield. Virginia 22161.
-------
EPA-600/7-78-142a
September 1978
Electrostatic Effects in Fabric Filtration
Volume I. Fields, Fabrics, and Particles
(Annotated Data)
by
Gaylord W. Penney
Carnegie-Mellon University
Schenley Park
Pittsburgh, Pennsylvania 15213
Grant No. R803020
Program Element No. EHE624
EPA Project Officer: James H. Turner
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
Volume I examines the effect of particle charge and electric fields on the
filtration of dust by fabrics. Both frictional charging and charging by
corona are studied. Charged particles and an electric field driving particles
toward the fabric can greatly reduce the initial pressure drop and also give
some reduction in the steady state pressure drop. Efficiency of collection is
also increased. Frictional charging is ususally erratic and corona charging
limited by dust resistivity. Thus further work is required to develop practi-
cal devices.
In long term or equilibrium conditions, nodular deposits attached to fi-
bers may be of major importance. These nodular deposits may act like check
valves giving a high pressure drop in the forward or filtering directions, and
a low pressure drop in the reverse or cleaning direction. Thus these nodules
are difficult to remove by reverse air flow. These nodular deposits, as well
as electrostatic fields between adjacent fibers of a filter, appear to be
electrostatic in nature. This appears to be an important area for further
research.
11
-------
FOREWORD
The United States Environmental Protection Agency (EPA) awarded a re-
search grant to Carnegie-Mellon University for basic work on electrostatic
effects in fabric filtration. The research was performed by coinvestigators
(E.R. Frederick and G.W. Penney) and is reported in two volumes: one by each
investigator.
Because of the exploratory nature of the research a non-standard format
was developed for the reports: an annotated data book. The reports describe
a series of individual, but related experiments each designed to investigate
a specific facet in electrostatics effects. Although major conclusions and
recommendations are included at the beginning of each volume, there are
separate conclusions following most of the individual experiments. In some
cases the experiments were inconclusive, but provided data and discussion for
future work.
The authors brought considerable experience to their work and have pre-
sented many speculations not conclusively supported by the experimental data.
Finding such support should provide the basis for a number of future research
projects. It is expected that the data and discussions presented in these
two volumes will serve to help initiate such projects as well as to give the
fabric filtration community explicit knowledge not previously obtainable.
John K. Burchard
Director
Industrial Environmental Research
Laboratory - RTP
iii
-------
ACKNOWLEDGEMENTS
The co-investigators are particularly grateful to Dr. James H. Turner,
the EPA Project Officer, for his thoughtful considerations and technical
support during the course of the investigative program. Others, too numerous
to mention here, have contributed significantly to the.success of the study
through their cooperation in providing commercial and experimental filter
media and in supplying problem dusts. The extraordinary assistance given by
Messrs. John W. Brooks and Robert W. MacWilliams is acknowledged with sincere
appreciation.
Section 14 of Volume I was given as paper 77-51.6 at the Toronto meeting
of the APCA. It is reproduced by permission of Harold Englund, Editor, Journal
of the Air Pollution Control Association.
The project could not have functioned effectively without the dedication
and independent contributions of the graduate student Robert Lembach and the
technicians Ron Feigel, Kirk Ludington, David Richey, John Wieczorkowski, and
Brent van Zandt.
iv
-------
CONTENTS
Abstract ii
Foreword ! ! ill
Acknowledgements . . . iv
Figures vi
Tables viii
1. INTRODUCTION 1
2. SUMMARY 4
3. FRICTIONAL CHARGING OF DUST 6
4. POTENTIALS AND FIELDS DUE ELECTRICALLY CHARGED AEROSOLS . . 13
5. MEASUREMENT OF NET CHARGE ON AN AEROSOL 15
6. MEASUREMENT OF POSITIVE AND NEGATIVE FRACTIONS 17
7. VISUAL EVALUATION OF ELECTROSTATIC FIELD EFFECTS DUE
TO FRICTIONAL CHARGE ON DUST 21
8. PRELIMINARY COMPARISON OF ELECTROSTATIC (ES)
AND NON-ELECTROSTATIC (NES) FILTERS 27
9. CONVEYING A CHARGED AEROSOL 31
10. CHARGE CARRYING ABILITY (CCA) AS A MEANS FOR
MEASURING PARTICULATES IN CASES 37
11. COMPARATIVE MEASUREMENTS OF PRESSURE DROP IN
FABRIC FILTERS 40
12. ELECTROSTATIC AUGMENTATION OF EFFICIENCY OF
FABRIC FILTERS 57
13. MEASUREMENT OF AGGREGATION OR AGGLOMERATION
OF DUST DISLODGED FROM A FILTER 64
14. NODULAR DEPOSITS IN FABRIC FILTERS 69
15. DEMONSTRATION OF ELECTRIC FIELD BETWEEN ADJACENT
FIBERS OF A FILTER 79
ADDENDUM 81
ELECTRIC FIELDS IN FABRIC FILTERS DUE TO
NATURAL FIBER PROPERTIES 81
Fabric Contact Potentials as Measured
by the Kelvin Method 81
Fabric Contact Potentials as Measured
by the Radioactive Source Method 83
High Resistivity Resin Particles in Fabric Filters 85
SECTION 16 CONCLUSIONS TO VOLUME I 37
-------
FIGURES
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Cross Section of Vortex Chamber for Charging Dust
by Friction
Dust Feed System for Vortex Charger
Photograph of Device for Charging Dust by Impingement
Cross Sections of Absolute Filter as Mounted for
Measuring Net Charge on an Aerosol
Parallel Plates for Separating Positive and Negative
Fractions of Dust
Foil Electrode Used to Measure Current to Cardboard Funnel . .
Photograph of Assembled Filter
Photograph With Lower Section Removed to Expose Uneven Deposit
on the Filter
Close Up View of the Dust Deposit on a Grounded Metal Plate . .
Photograph of Dust Deposit After a Test in Which the Fabric
Had an Antistatic Treatment So That the Filter Surface
Was an Equipotential
Another Comparison of NES with ES Filtering
Comparison of NES with ES Filtering
Apparatus for Passing Charged Dust Through an Insulating
Pipe and Measuring any Change in Charge
Diagramatic Sketch of Apparatus for Measuring Charge
Carrying Ability (CCA)
Block Diagram of Apparatus for Comparison of ES
and NES Filters
Cross-Section of Filter
Pressure Drop vs. Time - Test B-l
Pressure Drop vs. Time - Test B-2 . . .
Pressure Drop vs. Time - Test B-3 ....
Filter As Seen Through the Window
Pagi
9
10
11
16
18
20
22
24
25
26
28
30
33
38
41
42
44
45
50
51
vi
-------
FIGURES (CONT.)
Number Page
21 Close-Up View of Dust 52
22 Photograph Showing Both Filters Early in the Test 53
23 Electrostatic Augmentation of Efficiency of Clean
Darvan Fabric 58
24 Electrostatic Augmentation of Efficiency of Clean
Wool-Acrylic Fabrics 59
25 Electrostatic Augmentation of Efficiency of a Clean
Woven Fabric 60
26 Electrostatic Augmentation of Efficiency of a Clean
Woven Glass Fabric 61
27 Electrostatic Augmentation of Efficiency of a Clean
Polyethylene Fabric 62
28 Apparatus for Measuring Aggregation of Dislodged Dust .... 65
29 Nodules Hanging From Fibers 70
30 Inlet Dust Particles 70
31 Rock Dust Nodule and Particle Makeup 72
32 Silica Dust Nodule and Particle Makeup 72
33 Adhesion Force vs. Particle Size for Four Materials 75
34 A Model of Nodule Action 75
35 Forward/Reverse Pressure Drops of Rock Dust and Nomex Felt . 76
36 Opening of Nodules Subjected to a Reverse Air Flow 75
37 - Negatively Charged Particles Attracted to Particular
Fibers 80
vii
-------
TABLES
Number Page
1 Conversion Factors 12
2 Various Dusts Charged by Impingement Against a
Tungsten-Carbide Target 12
3 Currents Measured Using the First Test Section 34
4 Results of Tests with Second Test Section 36
5 Conditions for Test B-l 46
6 Conditions for Test B-2 47
7 Conditions for Test B-3 49
8 Filtering with Variable Upward Air Velocity 67
9 Measurement of Forward and Reverse Pressure Drop:
The Conclusion of Test #23 68
10 Contact Potential (Volts) at 60-80% RH 83
11 Contact Potential (Volts) at 20-30% RH 84
viii
-------
SECTION 1
INTRODUCTION
The thesis that electrostatics, developing normally in particulate col-
lection by fabric media, are very important in controlling the filtration
process has long been espoused by E.R. Frederick.1 He has been interested
particularly in relating the relative triboelectric properties of dusts and
fabrics and then utilizing these data with other characteristics to specify
filter fabrics for optimal performance. Other investigators have applied
fields across the thickness of a filter in order to increase the efficiency of
dust collection. Discussions with Frederick stimulated G.W. Penney's interest
in possible electrostatic effects. He became interested in the possibility of
depositing the dust on the surface of the fabric and depositing this dust in a
relatively porous layer or filter cake. In this concept the filter becomes a
support for the collected dust and a means for catching any dust that escapes
the electrostatic process.
As described in two patents2'^ dust particles can be deposited in a porous
manner on the filter fabric under the following conditions:
1. The particles are electrically charged.
2. There is an electric field driving the charged particles toward the
filter surface.
3. The dust and fabric have sufficient conductivity so that the surface
is an equipotential.
The porous deposit is obtained because charged particles tend to follow
the lines of force of the field. These lines of force tend to concentrate at
any projection. Thus the dust deposit is greater at a projection.
Any slight initial projection then tends to grow.
be deposited in a fiber like manner.
In this way dust can
As described in the patents, the dust is charged by corona. However
corona charging requires that resistivity be of the order of 2 x 1010 ohm-cm
or less. This is easy to do in the laboratory but it turns out that in most
filter-applications the resistivity is materially above this 2 x 1010 ohm-cm
limit.
Edward R. Frederick, "How Dust Filter Selection Depends on Electro-
statics," Chem. Engr. 68, 107 (1961).
2Gaylord W. Penney, U.S. Patent No. 3,910,779 Electrostatic Dust Filter,
October 7, 1975.
3Gaylord W. Penney, U.S. Patent No. 3,966,435 Electrostatic Dust Filter,
June 23, 1976.
-------
Another possibility exists: namely, that the dust is generated in some
process such that the particles are naturally charged. In such a case and
with proper dimensions the electric field could be produced by the space
charge of the dust itself provided that the dust loading is reasonably con-
stant. This would be particularly attractive since it would not require any
auxiliary equipment. A grant from the Environmental Protection Agency was
obtained to study this and other possible electrostatic effects.
In this grant, Frederick and Penney are coinvestigators. Penney has been
primarily concerned with small scale tests to isolate and evaluate particular
electrostatic phenomena which may be of importance in fabric filtration. This
work is reported in Volume I. Frederick has employed Bench Scale experimental
filtration tests and attempted to use the resulting data, with information
from supporting evaluations, to relate electrostatic effects with performance.
This work is reported in Volume II.
GENERAL CHARACTER OF THE WORK
There has been much speculation but few quantitative tests regarding
electrostatic effects in fabric filtration. Thus our work is mainly explora-
tory in character. A major part of the work has been to determine the order
of magnitude of the effects and develop methods of measurement. Considerable
time may be required to develop a test. Then if the effect being studied does
not appear to have commercial application only a limited number of tests ap-
pear to be economically justified.
Electrostatic effects can be divided into two classes - Class I Effects
due to large scale fields acting on charged particles. Particles can be
charged either by friction or by corona. The average or net charge can be
readily measured. Also approximate measurements can be made of the fraction
of the particles that are charged to a given polarity. Electric fields can be
produced by potentials applied to electrodes or by the space charge of a cloud
of particles predominantly charged to one polarity. These electric fields can
be measured. Class II Triboelectric Effects: whenever two surfaces having
different work functions come in contact, there tends to be an exchange of
charge between the two surfaces so that one becomes positive and the other
negative. A common example is the rubbing together of two dielectrics.
Triboelectric effects are notoriously erratic because the exchange of
charge depends on the difference in work functions of the two surfaces in con-
tact. The work function of a surface depends on surface layers of adsorbed
molecules as well as the base material. Two sides of a crystal will in
general have different work functions. Thus with crystalline particles it is
possible for two identical particles to approach so that opposite crystal
faces come in contact. This causes an exchange of charge so that as the par-
ticles separate they carry opposite charges.
These triboelectric effects are difficult to control and measure. Static
charges are commonly observed with insulators at low humidities where the
charges can be retained for long times. However, with particles, low con-
ductivity is not necessary. Kameinski14 has measured the charging of particles
by impingement against surfaces. He has shown that metal particles can be
-------
strongly charged by impingement against a different metal, e.g., tungsten
particles were strongly charged when they impinged against a stainless steel
target. Thus, with aerosols, triboelectric effects are almost impossible to
avoid or eliminate.
The first half of the work under Volume I of the Grant was primarily de-
voted to Class I effects largely because the parameters could be measured and
important effects demonstrated. However as the work progressed it appeared
that commercial applications were severely limited by the high resistivity of
most industrial dusts. This prevents effective charging by corona. Also it
appeared unlikely that many industrial dusts would be found where the particles
carried like charges.
Furthermore it appeared that triboelectric effects are very important in
fabric filtrations. Nodules of dust appear to be held together and to fibers
by electrostatic forces. Also individual particles seem to be held together
and to fibers by electrostatic forces. Thus as the work progressed more work
was directed to these Class II, or triboelectric, effects. However, the
tests are more difficult because the effects are difficult to measure. Two
sides of a 1 pm particle may have quite different work functions and as yet
there are no satisfactory methods for measuring the work functions of these
small areas.
**P. Kameinski, "Triboelectrification of Aerosol Particles by Impingement,'
Ph.D. Thesis, University of Minnesota (1974).
-------
SECTION 2
SUMMARY
EFFECTS OF ELECTRIC FIELDS ON CHARGED PARTICLES
The combination of charged particles and electric fields can both reduce
the pressure and increase the efficiency of a fabric filter. In the labora-
tory particles can be charged by friction or by a corona discharge. If then an
electric field is provided to drive particles toward the fabric, the particles
are deposited primarily on the fabric rather than in the fabric. Furthermore,
the dust deposit for filter cake tends to be of a porous nature, giving a low
pressure drop.
In one form, silica particles are blown against a tungsten carbide target,
giving the particles a negative charge. If then this charged aerosol is blown
inside of a filter bag, the space charge due to the charge on particles gives
an electric field driving particles towards the bag.
In another form particles are charged by passing the aerosol through a
corona discharge. With either form of charging, the field can be provided
either by space charge or by electrodes connected to some voltage source.
In laboratory tests under favorable conditions a marked reduction in
pressure drop has been obtained by this method. In our tests an electro-
static (ES) filter and a nonelectrostatic (NES) filter have been operated si-
multaneously to compare the two methods of operation. The greatest reduction
in pressure drop in the ES filter occurs in the first cycle of operation.
The filters were cleaned by pulses of reverse air. After many cycles of
operation the pressure drop gradually increases, however, the ES filter main-
taines a distinctly lower pressure drop than the NES filter.
This combination of charged particles and electric field can increase
efficiency as well as reduce pressure drop. Even some relatively porous,
woven fabrics in the new or clean condition would give 99% efficiency in the
removal of submicron particles in the ES mode as compared to 80% efficiency in
the NES or usual mode of operation.
While a decrease in pressure drop and increase in efficiency can be ob-
tained in the laboratory, they cannot as yet be obtained in most industrial
situations. Friction charging of the dust is usually erratic and satisfactory
corona charging requires that the dust have a resistivity of 2 x 1010 fi cm or
less, but most dust under industrial conditions will have a resistivity of
10n j? cm or more. Furthermore, even if the dust can be charged, there is
another limit of the order of 101U ft cm above which conductivity of the col-
lected dust cannot carry away the charge on the arriving dust.
-------
EFFECTS DUE TO THE SEPARATION OF CHARGE BETWEEN CONTACTING PARTICLES
It was found that in operating filters the pressure drop in the forward
or filtering direction may be several times the pressure drop in the reverse
or cleaning direction. This can be traced to dust formations or "nodules"
attached to fibers in such a way that they can move relative to the fabric and
thus act as crude check valves. These nodules appear to be held together by
electrostatic forces associated with differences of work functions or contact
potentials. It appears that this type of electrostatic force between adjacent
particles may be very inportant in the normal operation of filters. Such
forces may be responsible for dust formations which are difficult to remove
from filters.
CONCLUSIONS
In the laboratory where the resistivity of dust can be decreased by
raising the relative humidity, electrostatic forces can materially reduce
pressure drop and increase efficiency of fabric filters. However, more work
is required to accomplish this in typical industrial situations.
Electrostatic forces between adjacent surfaces appear to be important in
holding particles to each other and to fibers. This appears to be an impor-
tant area for fundamental studies.
-------
SECTION 3
FRICTZONAL CHARGING OF DUST
BACKGROUND
As discussed in U.S. Patents 3,910,77 and 3,966,435, a porous dust depo-
sit can be produced by charged dust particles in the presence of an electric
field driving these particles toward the filter surface. Electric fields are
usually produced by electrodes at high potential. However, fields can also be
produced by the space charge of the particles themselves. As will be dis-
cussed, the dust particles must have charges of like sign and there must be
the proper relationship between dust density and filter dimensions. If then
the dust could be properly charged by friction, the porous dust deposit could
be obtained with all filter surfaces at ground potential and without requiring
any high voltage sources or electrodes. The high potentials would be produced
by the dust itself.
According to the triboelectric series viewpoint, if particles from one
end of the series rub against a material from the other end of the tribo-
electric series all particles should carry electrical charges of the same
sign. Experiments with friction charging of materials such as fabrics sug-
gests that a vigorous rubbing should give the strongest charge.
This suggests that, if particles can be made to rub vigorously against
a material of widely different triboelectric properties, all particles should
be charged to the same polarity.
PURPOSE
The primary purpose of the test was to obtain an aerosol in which all
particles had strong charges of the same sign. This charged aerosol could
then be used for filter tests. Another purpose was to obtain information on
the frictional charging of dust as an aid in locating industrial situations in
which the particles might carry unipolar charges.
FIRST APPARATUS FOR CHARGING DUST
It was first assumed that a large amount of rubbing of a particle against
a surface would produce the maximum charge. So the first apparatus consisted
of a vortex chamber into which particles could be injected and, by centrifugal
force, be thrown to the outside where they would rub against the chamber wall
or liner. The liner of the vortex chamber could be changed to test the effect
of triboelectrically different materials.
-------
A sketch of the vortex chamber is shown in Figure 1 and a diagramatic
sketch of the assembled apparatus is shown in Figure 2.
TEST RESULTS
The measurements of dust coming out of this vortex chamber seemed to be
completely erratic. The average charge was small and variable. Various li-
nings and dust were tried as well as means for feeding dust into the cylinder
but no consistent readings were obtained.
Methods for measuring positively and negatively charged fractions of the
dust will be discussed later, but the conclusion was the dust leaving Che vor-
tex chamber was a mixture of positively and negatively charged dust regardless
of the positions in a triboelectric series of the dust and vortex chamber
liner.
CONCLUDING OBSERVATION
The liner of the vortex chamber usually developed holes due to abrasion.
With a copper liner silica dust coming out of the chamber seemed to have a
slight copper color. Thus it might be that the particles were being contami-
nated by materials worn off from the vortex liner.
CONCLUSION
Instead of intensive rubbing we should try a single impingement of par-
ticles against a target. The target should be a hard material to minimize
wear. This leads to the impingement charging apparatus.
SECOND DEVICE FOR CHARGING DUST
In order for a particle to make a single contact with the target material,
dust was blown through a nozzle so that it would impinge against a target.
Most dusts have a wide range of particle sizes. In order to insure that the
fine particles strike the surface, a high velocity (20,000 ft/min or more) is
desirable.* A photograph of the apparatus is shown in Figure 3. A fluidized
bed dust feeder is supplied with compressed air at 5 to 15 Ibs. pressure. A
hose connects this feeder to a nozzle of 1/16" to 1/8" diameter. A tungsten
carbide target is mounted with its surface oriented about 45° with respect to
the jet. Initially the tungsten carbide target was mounted on insulation and
with a lead which could be connected to a Keithley Electrometer.
*0ur tests attempted to duplicate conditions in industry and so used the
units common in industry. Conversion factors for converting to metric units
are given in Table 1.
-------
TEST RESULTS
Initially the current to the tungsten carbide target was assumed to be
equal and opposite to the charge carried away by the dust. This would be true
if the dust was uncharged before touching the target. But it was found that
there could be considerable charging of dust in passing through the nozzle.
Thus the current to the target was not a true measure of the charge on the
dust as it left the target. For this reason the charge on the dust is
measured by collecting the dust in a filter and the charge evaluated by the
current imparted to the filter.
The results of impinging various dusts against a tungsten carbide target
are shown in Table 2. The dusts tested are samples submitted by various fil-
ter users. These were merely exploratory tests. With the fly ash, the dust
loading and also the charge density were both high, but we did not make tests
to determine if the charge was proportional to dust loading. The charge
density of 1.75 x 10"1* was about the upper limit that our apparatus was in-
tended to handle. Thus it is possible that this might be space charge limited.
-------
Compressed Air
Entronce Slit
Exit Tube
Phenol ic Cap
Copper Cylinder
Tube Copped at End
CROSS SECTION OF CHARGER
Fig. 1 - Cross Section of Vortex Chamber for Charging Dust by Friction
-------
Charged Dust
Out
Friction
Charger
This Tube
Capped
TJ
,^-Flow Meter
— Dusl Reservoir
Orifice
L
-*- AP-—
For Determining
Flow Rate
DUST FEED SYSTEM
I | Compressed
[1 Air In
Figure 2 - Dust Feed System for Vortex Charger
-------
Figure 3 - Photograph of Device for Charging Dust by Impingement
11
-------
TABLE 1. CONVERSION FACTORS
To Convert From:
°F
ft.
ft. 2
ft.3
ft./min.
ft.3/min. (CFM)
inches
oz.
grains
grains /ft.3
To:
°C
meters
meters^
meters^
meters/sec.
meters3/sec.
meters
prams
grams
grams /m.
Multiply By:
| (°F - 32)
0.305
0.0929
0.0283
5.08 x 10"3
4.72 x 10~4
2.54 x 10"2
6.45
0.0647
2.288
TABLE
PTPINGEMENT
2. VARIOUS DUSTS CHARGED BY
AGAINST A TUNGSTEN-CARBIDE TARGET
Type of Dust
Electric Furnace Dust, U-P
P.K-Pico Resin
Ferromolybdenum
Fly Ash
Potato Dust
Dust Loading
grains/ft.3
1.7
19.6
3.4
32.0
6.2
Acrylic Polymer Dust 6.4
Molybdenum Dust 3.1
Rock Dust From Asphalt Plant 3.2
Silica 8.5
Charge Density
coulombs /m3
+1.23 x 10~6
<1.7 x 10~9
-1.58 x 10~6
-1.75 x 10~A
+1.75 to ,
+15.8 x 10
Variable Sign
-7 x 10~6
+1.05 x 10~6
-3.5 x 10~5
12
-------
SECTION 4
POTENTIALS AND FIELDS DUE ELECTRICALLY CHARGED AEROSOLS
Potentials due to an electric charge distributed throughout a volume are
given by Poisson's Equation:
v2 v = - £- (i)
o
The charge density p (coulombs/meter3) is assumed to be uniform through-
out the volume.
e0 = 8.85 x 10- 12 Far ads /meter.
V = Potential (volts) at a point having coordinates x,y,z.
1. Potentials between infinite parallel grounded planes separated by a dis-
stance H (meters) .
This is a one dimensional problem so the Eq. 1 reduces to:
- ~
e
Solving and using the boundary conditions that
V = 0 at x = 0 and at x = H gives :
V = + 2^ (Hx - x2) (3)
o
The electric field is:
g.g- CH-2*>
2. Field inside a long cylinder of radius R:
Transforming Eq. 1 to cylindrical coordinates and solving gives:
_
dr 2e
o
2e 4e
o o
(R2 - r2) (6)
13
-------
where r = radius of a point within the cylinder at which the potential is
"V."
EXAMPLE
p = 10~5 coulombs/m3. This is a medium charge density in our work. In
a cylindrical bag R = 10 cm gives a field at the bag surface of 565 volts/cm.
In our work 10-l* is a relatively high charge density and in the 10 cm radius
bag gives a field of 5650 volts/cm. This is a high field. Thus in bags of
reasonalbe size it is possible to obtain fields of the order of 3000 volts/cm
which in our tests give good results.
Initially the potential produced between parallel plates or at the center
of a long cylinder was used to measure the charge density.
14
-------
SECTION 5
MEASUREMENT OF NET CHARGE ON AN AEROSOL
In previous work with atmospheric air two methods were used to measure
the net charge carried by the particles in a unit volume of gas.
1. Measure the potential at the center of a long cylindrical grounded
tube. From Poisson's equation calculate the charge density required
to produce that potential.
2. Pass the charged aerosol through an '"Absolute" or "HEPA" filter. The
filter collects the particles and their charge. The current to the
filter is measured with an electrometer. An absolute filter arranged
for measuring charge is shown in Figure 4.
The absolute filter is useful for measuring very dilute aerosols. It
does not have the dust holding capacity to handle a high dust loading except
for very short periods of time. For this reason, initially the charge on dust
was evaluated by measuring the potential in some defined volume. Later it was
realized that, for our purposes, the charge collected by a fabric filter could
be used to measure the charge on the particles. A fabric filter is mounted on
a plastic frame with the fabric supported against a metal screen. The current
collected by the screen is measured by an electrometer and the rate of air
flow through the filter is measured using a calibrated orifice.
15
-------
12"
8
"
8"
///////////////
A,R SEAL
AIR SEAL
ELECTRICAL
CONTACT
COPPER
HOUSING
HEPA
FILTER
AIR
SEALS
ELECTROMETER
CONNECTION
^POLYSTYRENE
INSULATORS
Figure 4 - Cross Sections of Absolute Filter as Mounted for Measuring Net Charge on an Aerosol
-------
SECTION 6
MEASUREMENT OF POSITIVE AND NEGATIVE FRACTIONS
BACKGROUND
As has been mentioned, the measurement of the average or net charge on an
aerosol is relatively simple, but the determination of the fraction that is
positive and the fraction that is negative is not so simple.
With fine (0.3 urn diam.) oil droplets, the positive and negative fractions
can readily be determined by passing the aerosol between positive and negative
plates with the aerosol at a velocity of 50 to 300 ft./min. But with our test
dusts there are many difficulties.
PURPOSE OF TESTS
These tests attempted to attract negatively charged particles to the posi-
tive electrodes and positively charged particles to the negative electrodes and
so determine the fraction of the particles charged to a given polarity.
APPARATUS
The first apparatus consisted of the friction or impingement charger fol-
lowed by a tapered transition section leading to 3 parallel steel plates with
plastic side walls. Each plate was 6" wide by 12" long. A potential was
applied to the center plate. The outer plates were grounded through a
Keithley Electrometer. This apparatus is shown in Figure 5. Alternately a
potential could be applied to the outside plates and the current to the center
plate measured.
TEST RESULTS
Silica dust was blown against the target using about 5 lb./in.2 pressure.
A potential of -4.5 KV was applied to the side plates and the current to the
center plate measured. The current from the target was +2 x 10~8 amps and
from the center plate -2.5 x 10~8. This would indicate that the dust had some
initial negative charge before it hit the target or that there was some
bouncing of dust between plates. Visually the dust on the center plate was
estimated to be 10 times as great as that on the side plates. Of this 10% on
the outer plates there was no way to tell whether it was initially positively
charged so that it was deposited on the negative plates or whether it was
first precipitated on the positive plate and then changed its charge and hopped
to the negative plate. Also some dust may have been deposited by inertia due
to the high velocity at which particles left the tungsten carbide target.
17
-------
Paper Funnel
Plates
Dust entered
here at 1 CFM
Plexiglas
Steel
Figure 5 - Parallel Plates for Separating Positive
and Negative Fractions of Dust
18
-------
In these tests some dust was deposited on the cardboard funnel. The
measurement of current to the cardboard funnel is illustrated in Figure 6. In
a typical measurement the current from the tungsten carbide target was
+4 x 10~7 amps and from the cardboard funnel -3 x 10~7 A. This indicates
that the high velocity and turbulence in the aerosol leaving the target throws
dust against surfaces so that the electric fields are unable to adequately
control the motion of the particles.
Other tests were made with varying plate spacings and potentials. How-
ever in all cases some of the silica dust would deposit on negative surfaces
even though we were attempting to charge this dust negatively.
DISCUSSION
Out test dust contains a wide range of particle sizes. In charging the
dust by impingement, a high velocity jet must be used to insure that the fine
particles strike the target. The particles leave the target in somewhat random
directions. The high velocity air jet causes violent turbulence. Due to the
combination of air turbulence and inertial effects, the motion of charged par-
ticles can only partially be controlled by the electric field. The tendency
of particles to bounce between plates increases with the square of the electric
field. This tends to limit the field strength which can be used. Due to
these various effects the current to a given plate may not be a true measure
of a given polarity of charge in the aerosol. For example a positive plate
attracts negative particles but due to inertia and turbulence some positive
particles may strike the positive plate.
Thus we do not have any simple means for accurately measuring the fraction
of the particles charged to a given polarity. Our judgment is that for the
silica dust impinged on the tungsten carbide target between 90% and 100% of the
particles are negatively charged.
19
-------
Charger
ro
o
PVC
Connector
Cardboard
Funnel
Foil
Cover
L
Verticle Plates
Plexiglas Adapter
to Plates
Figure 6 - Foil Electrode Used to Measure Current to Cardboard Funnel
-------
SECTION 7
VISUAL EVALUATION OF ELECTROSTATIC FIELD EFFECTS
DUE TO FRICTIONAL CHARGE ON DUST
BACKGROUND
Tests show that dust, particularly silica, blown against a tungsten car-
bide target acquires a strong electric charge. Theory indicates that a few
grains per cubic foot of such charged dust can create strong electric fields
tending to drive these charged particles to the containing walls. With a
spherical container having a uniform wall, the deposition would be the same
for either grounded metal, or an insulating wall. By using a grounded filter
with other walls insulated, the field effect can be concentrated at the
filter.
PURPOSE
The purpose of this test was to permit a visual inspection of the dust
deposit on a filter when dust was charged by friction and when there was a
space charge field at the filter surface.
APPARATUS
A flat fabric filter supported against a grounded metal screen was
mounted across the face of a rectangular acrylic box. This box formed the
outlet or clean air side of the filter. The inlet side of the box was an
acrylic box 2" in depth. The filter was 6" x 24". Theoretically much of the
dust could be electrostatically precipitated near the entrance. For this
reason the outlet box was divided into three sections so that the airflow and
pressure drop could be monitored for each of the three sections. The assem-
bled apparatus is shown in the photograph of Figure 7. The chamber containing
the impingement charger is shown at the left. The acrylic box was held to-
gether by clamps so that it could be readily opened for inspection.
The filterused was polypropylene fabric without antistatic treatment.
To evaluate the potential drop through the fabric, there were three metal
probe's mounted on the face of the fabric. The one at the center was grounded
to the screen so that one could observe the electrostatic precipitation of
dust at a surface where there could be no filtration. The two probes near
the ends of the filter were simply glued to the face of the fabric with an
insulated lead going to an electrostatic voltmeter so that the voltage drop
through the fabric could be measured.
21
-------
Figure 7 - Photograph of Assembled Filter
22
-------
RESULTS
After running a short time, the lower box of the filter assembly was re-
moved. A photograph of the lower side of the filter is shown as Figure 8.
The heavy deposit on the center grounded probe can be seen at the center of
the filter. The two probes near the ends of the fabric operated at about
2000 volts which was the potential drop through the fabric. Thus the pott-n-
tial of the center grounded probe was 2000 volts below the surrounding fabric.
This higher field strength at this surface caused a heavy deposit of dust.
Close-up photographs of this heavy deposit are shown as Figure 9. In this
figure, a rectangular pattern of smaller deposits can be seen at the lower
left corner of the photograph. The spacing between these deposits corresponds
to the mesh size of the screen supporting the fabric. At each point where
wires of the screen cross there is better contact with the fabric; accordingly
dust deposited at these positions.
In these tests no difference in the thickness of deposit or pressure drop
could be observed along the length of the filter.
In this first test, the fabric had no antistatic treatment and the
probes indicated that the potential drop through the fabric was about 2000
volts. In the next test the fabric had an antistatic treatment. The photo-
graph of Figure 10 shows the deposit of dust in this second test. In this
case the deposit is distributed over both the fabric and the metal plates. In
this case the fibers are relatively conducting and so any projecting fiber
collects a dust deposit.
CONCLUSIONS
The electrostatic effects were obtained even though all metal parts
(except the two measuring probes) were grounded. The particles were charged
by impinging against a grounded surface. The space charge of the cloud of
charged dust particles could create strong electrostatic fields tending to
drive the charged particles toward grounded surfaces. A strong electrostatic
deposition of particles was demonstrated. The electrostatic velocity of the
particles relative to the air appeared to be greater than the air velocity
through the filter (3 ft./min.).
23
-------
f
Figure 8 - Photograph With Lower Section Removed
to Expose Uneven Deposit on the Filter
24
-------
Figure 9 - Close Up View of the Dust Deposit
on a Grounded Metal Plate
25
-------
. ,: •• .- £r
- - ••• «Q -. . -..-^
:-, s -
-->» p^. .
*
Figure 10 - Photograph of Dust Deposit After a Test in
Which the Fabric Had an Antistatic Treatment
So That the Filter Surface Was an Equipotential
26
-------
SECTION7 8
PRELIMINARY COMPARISON OF ELECTROSTATIC (ES)
AND NON-ELECTROSTATIC (NES) FILTERS
PURPOSE
To study the possibility of reducing the pressure drop of a filter by
electrostatic means for situations where the dust is charged by friction and
to compare frictional charging and corona charging of dust.
APPARATUS
These tests used the plastic boxes and flat filter used for the visual
demonstration. This was shown in Figures 1 and 8. This apparatus used a
fabric stretched and glued to a rectangular plastic frame. This frame was
then clamped between the two plastic boxes using Mortite as a gasket. At
filtering pressures this provides a convenient and effective seal. The fabric
was treated with 5% Downy solution which gave a resistivity of about 2 x 109
ohms per square. In order to control the electric field at the fabric surface,
a copper plate 6" x 24" was mounted parallel to the filter surface. For the
ES tests it was 2" from the fabric with a potential applied to it. For the
NES filter, the copper plate was spaced V' from the fabric and parallel to it.
THE TESTS
Tests were made by first making a test at one condition. The filter was
then cleaned using a household vacuum cleaner which appeared to return the
fabric to its original condition; then making a test for another condition.
Thus three tests were made in series using the same piece of cloth.
The results of one series of tests are plotted in Figure 11. Pressure
drop is plotted against running time. For curve //I there was no impingement
of the dust and the grounded copper plate was "j" from the fabric. There is
always some charging of a dust in a feeder but for curve #1 the electrostatic
effects were a minimum.
For curves 2 and 3 the copper plate was 2" from the fabric with 8 KV
applied between the plate and the fabric. The average field was 1.6 KV/cm.
although the field at the fabric would be increased due to the space charge of
the dust. For curve //2 the dust was charged by impingement and for curve #3
the dust was charged by corona. All tests used pulverized silica as the test
dust.
27
-------
20 40 60
RUNNING TIME : MINUTES
Figure 11 - Another Comparison of NES with ES Filtering
1. No Electric Field and Uncharged Dust (NES) Filtering
2. Dust charged by impingement against a tungsten carbide
target. Average electric field of 1.6 KV/CM ES filtering.
3. Dust charged by corona. Applied field of 1.6 KV/CM
average ES filtering.
2P
-------
The filters were cleaned by reverse air pulses. A small reservoir (ap-
proximately 75 in.2) was filled with compressed air to a controlled pressure
(5 to 15 Ib.) and then released by a quick opening, electrically operated,
valve. This pulse of compressed air was released into the clean air side of
the filter through a V D. copper tube. Each cleaning pulse was called a
"pop" because of the sound. In test #1 the filter was given a cleaning pulse
after 20 minutes and in tests #2 and #3 after 30 minutes.
The results of a repeat test are shown in Figure 12. In this case the
filters were given a cleaning pulse after about 2 hours. Apparently the dust
feed was a little slower resulting in a slower rise of pressure.
DISCUSSION
These tests have the weakness that the three tests were run in series.
Our fluidized bed feeder had to be used on order to supply the aerosol under
pressure to the nozzle used to impinge the dust. The feed rate of this feeder
was not constant and so the density of dust fed may not have been the same.
Also if there was any variable which was not recognized that variable might
be different. For these reasons it was felt that the comparative tests should
be run simultaneously. Various tests were made with smaller filters trying to
compare a nonelectrostatic (NES) filter with an electrostatic (ES) filter.
All of these tests were made in the spring, summer and early fall when the
relative humidity was not low. Thus the dust had a resistivity below 1011
ohm-cm. All tests showed a marked reduction in pressure drop for the ES fil-
ter with impingement charging of silica dust. However, the impingement
charging did not appear very constant. Thus it was concluded that to compare
the performance of ES and NES filters, corona charging was more consistent
and so should be used. Also tests should be much longer to see if the lower
pressure drop could be maintained. The longer tests are described in Section
11 of this report.
29
-------
O
CO
(ft
o>
I
o.
o
0>
tn
O)
it
3-
2-
I 2
Running Time: Hours
I
3
Comparison of pressure drops for three electrostatic conditions*
(T) X No charging of dust and no applied f ielde
(?) O Dust charged by impingement. Additional field of
8 kV across 5 cm.
(D A Dust charged by corona. Additional field of 8 kV
across 5 cm.
Filter? cleaned by « reverse air pulse after approximately 2 hours.
2
Woven fabric with napned surface, permeability 30 CFM/ft at 0.5 inches H20.
Silica dust, 5 grains per cubic foot*
Air to cloth ratio 3.7
Figure 12 - Comparison of NES with ES Filtering
30
-------
SECTION 9
CONVEYING A CHARGED AEROSOL
THEORETICAL BACKGROUND
Metal Ducts
If in industry some process generates an aerosol in which substantially
all particles carry the same sign of charge, special precautions would be re-
quired to convey this aerosol to a baghouse without loss of charge. The
mutual repulsion of electric charges of like sign creates potentials which are
described by Poisson's equation (Eq. 1 of Section 4). With a long cylindrical
duct of 0.3 meters (0.98 ft.) diameter, a charge density of 10"14 coulombs/m3
gives a field of 8.5 x 105 volts/m or 8500 volts/cm. Experience in electro-
static precipitators indicates that at this voltage gradient irregularities
at the wall will cause corona which will act to neutralize the charge on the
dust. A charge density of 10"1* coulombs/m3 can readily occur with dust
loadings of a few gr/ft3 if all particles carry the same sign of charge.
Thus ducts would have to be small. Furthermore this space charge field tends
to precipitate dust on the duct walls and the small ducts required would pre-
sent a large surface for dust collection.
Insulating Ducts
It might seem that charged dusts could be conveyed through insulating
ducts without loss of charge. However, experience indicates that loss of
charge is more probable with insulating ducts than with metal ducts. In our
situation there will be a dust generator at one end and a filter at the other
and both will normally be grounded. The insulating duct acts to concentrate
the field at any grounded object and cause corona there which tends to dis-
charge the dust.
The usual textbook treatment of electric fields discusses the case for
alternating currents and voltages. In this A.C. case the voltage gradient
through a dielectric varies inversely as the dielectric constant. Thus for
the A.C. case the voltage gradient through a solid dielectric is lower than
the gradient in air.
However with unipolar charged aerosols, the reverse is true. Charged
particles will, initially, be driven by the electric field toward any con-
taining wall. If the wall is an insulator it will acquire a charge until the
field perpendicular to the wall is neutralized. Thus, in steady state, the
field perpendicular to a perfect insulator will be zero. The field then be-
comes one-dimensional along the axis of the duct. This acts to produce high
fields at the ends of an insulating duct, and especially high fields if any
grounded conductor should puncture the insulating wall.
31
-------
PURPOSE OF TESTS
The tests were intended to explore this worst case where some grounded
point projects into an insulated duct, and to explore the kinds of effects
that would have to be considered if a charged aerosol were to be conveyed over
any distance.
APPARATUS
A DOP smoke was used because it is a very convenient and controllable
source of fine particles. A diagramatic sketch of the apparatus is shown in
Figure 13. The rates of flow of the incoming DOP and air are controlled in-
dependently and measured by orifices. These two streams combine and enter the
charger. The corona charger consists of a 2" length of .012" D tungsten wire
mounted along the axis of a 2" I.D. copper pipe. This charger connects to the
test section which is an 8" length of 1%" I.D. PVC pipe such as used for
plumbing. The test section leads to an absolute filter. The collected cur-
rent is measured by an electrometer. The filter is followed by another ori-
fice and fan. The details of the mounting for the absolute filter were shown
in Figure 4. The filter is mounted between polystyrene insulating plates to
insure that all of the collected charge flows to the electrometer. This
assembly is mounted inside a grounded metal housing to shield the filter from
stray fields and currents.
The first test section had four points projecting V into the PVC pipe.
Any one of these points could be connected by a clip lead to a Keithley
electrometer and ground.
The second test section had one grounded point consisting of 0.12" D
wire going radially to the axis of the pipe and then extending %" axially and
pointed up stream.
TEST RESULTS
Currents measured using the first test section have four probes or
points are shown in Table 3. The current collected by the absolute filter
was measured only when the current to probe A was being measured. Note that
only one probe was grounded at any one time. Probe A was nearest the corona
charger and so would be at a lower voltage point in aerosol stream. For 5%
DOP the current to Probe A was only 1% of the current to the absolute filter.
Currents to the other probes increased with distance from the charging sec-
tion.
For 20% DOP the situation is very different. All probes give approxi-
mately the same current and the current to the filter is reversed in sign.
This indicates that the corona from a grounded point can not only neutralize
the original charge but can actually reverse the net charge in an aerosol.
This reversal led to the second test section which was intended to accentuate
this tendency to reverse the charge.
32
-------
CO
OOP
AIR
FILTER
CHARGER
TEST
SECTION
-I2KV
8"-
\
/
8
FAN
I
PVC
TEST SECTION
TEST SECTION 2
Figure 13 - Apparatus for Passing Charged Dust Through an Insulating
Pipe and Measuring any Change in Charge
-------
TABLE 3
CURRENTS MEASURED USING THE FIRST TEST SECTION
Relative Absolute
Concentration Current Current Current Current Filter
of OOP Probe A Probe B Probe C Probe D Current*
5%
10%
20%
-2.5 x 10"9a -5.0 x 10~8a -1.1 x 10~7a -1.6 x 10~7a -2.5 x 10~7a
-4.2 x 10~7a -4.8 x 10~7a -5.5 x 10~7a -5.6 x 10~7a -1.6 x 10~7a
-1.1 x 10~6a -1.1 K 10~6a -1.1 x 10~6a -1.2 x 10~6a +5.0 x 10~8a
Measured when the current to Probe A was being measured.
-------
The currents obtained using the second test section are listed in Table
4. For 5% DOP the current to the probe is four times the current to the fil-
ter. The current to the filter first decreases as the percentage DOP in-
creases, and at 12.5% DOP the sign of the current to the filter reverses. The
current reversal increases as the percentage DOP increases. This again indi-
cates the possibility that a grounded point can not only neutralize the charge
on an aerosol but can also reverse the sign of the net charge.
The charge density of the aerosol leaving the corona charger is calculated
assuming that there is no loss of charge except that to the probe and filter.
Note that up to 12%% DOP the charge density is proportional to the percentage
of DOP. Above 12%% DOP the charge density does not increase with particle
density. These currents only increase due to the rate of gas flow.
This failure of the charge density to increase with the percentage of DOP
illustrates the difficulty called "corona quenching" which occurs in the
electrostatic precipitation of high particle densities. The charge density of
the particles being charged reduces the corona current. Thus at any given gas
velocity and ionizer voltage there is an upper limit to the charge that can be
imparted to an aerosol. At this limit, as the number of particles per unit
volume increases the average charge per particle must decrease.
DISCUSSION
These tests were only intended to indicate the difficulties if it should
be necessary to convey a high charge density aerosol for any distance. The
loss of charge due to corona can be controlled by using small metal ducts or
by subdividing a larger duct. However the large area required provides a
large surface where dust will be deposited by the electric field. This would
be another troublesome problem.
35
-------
TABLE 4
RESULTS OF TESTS WITH SECOND TEST SECTION
Calculated Charp.e
%DOP
5
10
12 1/2
15
17 1/2
20
25
30
Air
(cfm)
10
10
10
10
10
10
10
10
DOP
(cfn)
0
1
1
1
1
2
2
3
.5
.0
.25
.5
.75
.0
.5
.0
Current
To Probe
(amperes)
-2
-5
-7
-8
-1
-1
-1
-1
.0 x
.4 x
.0 x
.0 x
.0 x
.2 x
.5 x
.6 x
io-7
_7
10 '
_7
10 '
_7
10 '
-6
10 °
-6
10
-6
10 °
..
10~6
Current
To Filter
(amperes)
-5
-1
+5
+4
+1
+1
+2
+3
.0 x
.0 x
.0 x
.0 x
.5 x
.7 x
.5 x
.0 x
io"8
-8
10 °
q
10
-8
10 °
_7
10 '
-7
10
-7
10 '
-7
10
Density Leav:
Charger
< coulombs/in'
0.504
1.06
1.3
1.4
1.53
1.81
2.11
2.11
x 10'
x 10'
x 10'
x 10"
x 10'
x 10"
x 10'
x 10'
-4
-4
-------
SECTION 10
CHARGE CARRYING ABILITY (CCA) AS A MEANS FOR
MEASURING PARTICULATES IN GASES
INTRODUCTION
To measure particulates by the CCA method the aerosol is passed through a
corona discharge to impart an electrical charge to the particles. The aerosol
is then passed through an absolute filter which collects the particles and
their charges. An electrometer is used to measure the current due to the
collected charge. This current then serves as a measure of the collected par-
ticles. This gives an instantaneous and continuous measure of particle den-
sity. For particles larger than 0.5 um diameter, the charge is proportional
to the square of the diameter. Below 0.3 pm diameter the charge per particle
is proportional to the first power of the diameter. Thus the CCA method is
especially sensitive to fine particles.
There is a very serious limitation to the use of the CCA method. The
aerosol must be charged by passing through a corona discharge. If the resis-
tivity of the dust is greater than about 2 x 1010 ohm-cm back corona may occur
and prevent proper charging of the particles. In the tests reporte in Section
12 three aerosols were used, kerosene smoke, tobacco smoke, and welding smoke.
The kerosene smoke and the tobacco smoke both have a sufficiently low resis-
tivity. Welding smoke could cause trouble at humidities of 15% or lower, but
we made these tests at higher humidities where resistivity was not a problem.
The other method which we considered was light scattering. This is
widely used for measuring 0.3 Pro diameter particles. However for particles
small as compared to the wave length of light the light scattered varies as
the sixth power of the diameter. Thus CCA is much more sensitive than light
scattering for measuring particles smaller than 0.2 um diameter.
APPARATUS AND TEST PROCEDURE
The general arrangement is shown in Figure 14. The air enters through
IV PVC pipe which serves as insulation for the high voltage lead. The
corona-wire is .012" D tungsten and extends about IV inside the copper tube.
The 2" I.D. copper tube serves as the grounded electrode for the corona and
for collecting air ions. Under some conditions there will be a significant
number of air ions carried downstream from the corona. If not removed they
would cause a false reading. These air ions have a high mobility so that
they can be removed by a weak field which removes only a negligible fraction
of the dust particles. This is done by applying 75 volts to electrode 8 which
is 3/8" D by 6" long.
37
-------
I
Figure 14 - Diagraroatic Sketch of Apparatus for
Measuring Charge Carrying Ability (CCA)
-------
The construction of the charge collecting filter was shown in Figure 4.
It consists of a high efficiency (HEPA or Absolute) filter mounted between
polystyrene insulating plates. A current measuring lead is fastened to the
filter frame. Under most conditions the filter media would have adequate con-
ductivity to carry the minute currents involved. However to facilitate cur-
rent conduction, conducting paint is used to paint stripes from the wire to
the ridges or folds of the filter. A Keithley Electrometer is used to measure
the current collected from the particles.
The corona wire is operated at a voltage to give about 15 microamperes
corona current. The electrode for removing air ions is operated at 75 volts.
On an average fall day at Carnegie-Mellon University the CCA of the air
is of the order of 5 x 10~8 coulombs/m3. The instrument can measure up to
10-I* coulombs/m3 without diluting the incoming air. For our tests the air
flow was 3 cfm. The manufacturer's rating of the filter was 30 cfm at 1.0"
water pressure drop. At this low flow the filter has a very high dust holding
capacity and so a long life. We have used this method to measure down to
5 x 10~10 coulombs/m3, i.e., 1% of the normal particulates in Pittsburgh air.
It appears possible to go appreciably below this.
The ionizer should not be operated below 3 cfm or there will be excessive
collection in the ionizer. For continuous monitoring of high density aerosols,
the aerosol should be diluted so that only a low density aerosol is passed
through the filter. This is desirable in order to prolong the life of the
filter.
DISCUSSION
This method for measuring fine particulates is very useful in a labora-
tory where the operator knows how to recognize and avoid troubles due to back
corona and will clean the corona chamber when needed. The method can be used
to measure a wide range of dust densities and is much more sensitive than
light scattering for particles below 0.3 urn diameter. More work is needed to
develop a method that is suitable for general use.
39
-------
SECTION 11
COMPARATIVE MEASUREMENTS OF PRESSURE
DROP IN FABRIC FILTERS
PURPOSE
This series of tests was devised to compare the pressure drop of a con-
ventional filter with one in which the dust is electrically charged and in
which there is an electric field to drive this charged dust toward the fabric
surface.
METHOD
A dust stream is divided into two equal parts. One stream goes to a con-
ventional filter with no charging of the dust nor applied voltage; this is
called the non-electrostatic (NES) filter. The other dust stream is passed
through a corona charger and to a filter where an applied voltage produces an
electric field perpendicular to the filter surface; this is called the
electrostatic (ES) filter. The pressure drops of the two filters are then
compared over the test period.
In our experience corona charging is more consistent than impingement
charging. Accordingly corona charging was used in these comparative tests.
APPARATUS
A block diagram of the apparatus is shown in Figure 15. Dust from the
feeder (1) goes to a venturi (2) where a compressed air jet disperses the
dust. The dust stream then divides into two equal parts. The left hand
stream goes through a corona charger (3). Next there is a heater (4) which
by heating the air can raise the resistivity of the dust. This heating sec-
tion was used only in the second test (B-2). It was removed for the other
tests. The dust then goes to a filter (5) the details of which are shown in
Figure 16. From the filter the cleaned air goes through a measuring orifice
(6) and to a fan (7) whose speed can be adjusted to maintain the desired air
flow. The right hand dust stream goes directly to the NES filter (8), an
orifice (9), and fan (10).
Recording pressure meters, (14) and (15), record the pressure drop across
each filter. Both filters are cleaned by identical and simultaneous bursts of
compressed air (called "Pops" because of the accompanying sound). A pressure
switch (11) triggers the air valves (12) and (13) which release air from two
identical compressed air chambers.
40
-------
15
14
1 DUST FEEDER
2 VENTURI
3 IONIZER
4 HEATER
5 ELECTROSTATIC FILTER
6,9 MEASURING ORIFICES
7,10 FANS
8 NON-ELECTROSTATIC FILTER
11 PRESSURE SWITCH
12,13 REVERSE AIR VALVES
14,15 PRESSURE RECORDERS
Figure 15 - Block Diagram of Apparatus
for Comparison of ES and
NES Filters
-------
E^S
PLEXIGLAS
GROUNDED SCREEN
FABRIC FILTER
4"Dja. x 6" LONG
FOR TEST B 3
6" Die. x 12" LONG
FOR TEST B1 8 B2
STEEL PIPE
6" LDio.
Figure 16 - Cross-Section of Filter
42
-------
THE TEST RESULTS
Test B-l
The results of the first test are shown in Figure 17. Pressure drop vs.
time is plotted for each filter. In this case the pressure drop for the NES
filter is always higher than that for the ES filter. This test was run at
room humidity (40 to 50%) at which the resistivity was below 2 x 1010 ohm-cm.
The test dust was ground silica rock. The air-to-cloth ratio was 6.
Test B-2
This test was made to explore two questions; a) the possibility of opera-
ting at a higher air-to-cloth ration, and b) the effect of higher resistivity
of the dust. Pressure drop vs. time is plotted in Figure 18. For the first
26 hours both filters were operated at room humidity (45 to 50%) which re-
sulted in a dust resistivity below 2 x 1010 ohm-cm. For the first 45 hours
both filters were operated at an air-to-cloth ratio of 12.
Numerals on the chart refer to the following changes in operation:
1. At this point (hour 4) there were several cleaning cycles (Pops).
2. At hour 26, heater 4 was activated to increase the temperature of the
air going to the ES filter to 113°F. This corres-ponds to a dust re-
sistivity of about 7 x 1011 ohm-cm.
3. At hour 31-% temperature was further increased to raise the ES filter
air temperature to 150° - RH 6% - or a dust resistivity of 1013
ohm-cm.
4. At hour 41-% the NES filter was given many quickly repeated cleaning
cycles (Pops) thinking that the filter might clean to a lower pres-
sure drop. But the record shows that after 2 cycles a "pop" only
reduced the pressure drop from 10" to 8".
5. At hour 45 the air-to-cloth ratio of the NES filter was reduced to 6.
The ES filter remained at 12. At this condition the pressure drop
overlap is shown by the double cross hatching. Remember that both
filters are "popped" simultaneously whenever the NES filter reaches
10" pressure drop. The lower line of the double cross hatched area
is the pressure drop of the NES filter just after a "pop". The
charge density of the aerosol going to the ES filter averaged about
2 x 10~5 coulombs/m3- However the feed rate was somewhat variable
resulting in fluctuations in the charge density.
.The parameters for tests B-l and B-2 are given in Tables 5 and 6. The
dust loadings are determined by weighing the dust collected for the entire
period. A small stream of filtered air is passed through a Gelman filter for
the entire period and this gives the dust loading leaving the filter. This
in combination with the entering dust loading gives the average efficiency.
In tests B-l and B-2 the attempt is to have the dust loading entering
the NES filter equal to that entering the corona charger for the ES filter.
The corona charger removes some dust. Thus the dust entering the ES filter
43
-------
NON-ELECTROSTATIC FILTER
PRESSURE ENVELOPE
ELECTROSTATIC FILTER
PRESSURE ENVELOPE
GENERALLY 50r6 CYCLES PER HOUR
i I i i i i I i i i i I i i i i I
15 20
TIME (hours)
25
30
35
Figure 17 - Pressure Drop vs. Time - Test B-l
-------
NON-ELECTROSTATIC FILTER
PRESSURE ENVELOPE
ELECTROSTATIC FILTER
PRESSURE ENVELOPE
CYCLES/HOUR—INCREASING TO 20 CYCLES/HOUR
2
3
2O 25
TIME (hours)
Figure 18 - Pressure Drop vs. Time - Test B-2
-------
TABLE 5
CONDITIONS FOR TEST B-l
Type of dust - Silica
Fabric data
a. AFC 40/300 #99 Orion
b. Permeability - 29 cfm at 1/2" HZO
c. Resistivity - 2.2 x 1010 «/D
Relative humidity - 37%
2
Filters are cylindrical with cloth area = 0.96 ft
Air/cloth - 6 cfm/ft2
Air to each filter = 5.75 cfm
Field voltage = 7.5 Kv with 2.5 cm spacing
Charge density
a. Field 2.98 x 10~5coul/m3
b. No field 7.45 x 10~7 coul/m3
Forward to reverse pressure drop ratio at 10 cfm
a. Field - 2 to 1
b. No field - 3.7 to 1
Dust Loading
a. Field 4.2 to 4.3 gr/ft3
b. No field 2.8 to 3.2 gr/ft3
Reverse air "pops" applied for cleaning - 20 Ib gauge for each
Pressure transducer showed peak reverse pressures of 6" H_0 during
cleaning.
46
-------
TABLE 6
CONDITIONS FOR TEST B-2
Type of dust - Silica
Fabric data
a. AFC 40/300 #99 Orion
b. Permeability - 22.5 cfm at 1/2" H20
c. Resistivity - 2 x 10 n/D
Relative humidity at ionizer 47%
2
Filters are cylindrical with cloth area = 0.96 ft
Air/cloth = 6 cfm/ft2
Air to each filter =5.75 cfm
Field voltage = 7.5 Kv with 2.5 cm spacing
Charge density
a. Field 2.6 x 10~5coul/m3
b. No field 3 x I0"6coul/m3
Dust densities
a. Field 1.6 to 2.10 gr/ft3
b. No field 2.8 to 3.2 gr/ft3
Reverse air "pops" applied for cleaning - 50 Ib gauge for each
47
-------
is less than that entering the NES filter.
Test B-3
In this test the position of the corona charger was changed. The charger
for test B-3 was immediately after the feeder so that the air entering both
filters was charged. This was done so that the dust loading for both filters
would be the same. Thus there was a slight field at the NES filter due to the
space charge of the dust. The calculated field at the NES filter was 100
volts/cm. This compares with an applied field of 3000 V/cm in the ES filter.
Thus the NES had a slight electrostatic effect but apparently not enough to
seriously affect the pressure drop.
The primary purpose of Test B-3 was to compare the first operating cycle
of ES and NES filters, starting with clean cloth.
In this test each filter was operated without cleaning until it had
reached 10" pressure drop. In this case the filters were shorter, 6" long
instead of 12" long, but otherwise the same as in test B-l. This test differs
from test B-2 in that the filters in test B-2 are 12" long also in test B-2
there was a heater between the corona charger and the ES filter. There was no
heater in tests B-l and B-3. The test parameters are tabulated in Table 7 and
pressure drops plotted as a function of time in Figure 19.
Another modification in Test B-3 was that 3" x 3" windows were cut in the
enclosing 6" D steel cylinders. These windows were provided with removable
covers. Thus by removing the covers the dust deposit could be observed and
photographed at various times during the test.
In this test it appeared that in the ES filter the dust deposited electro-
statically on the lower part of the cylinder leaving some cloth relatively
clean. This appears to account for the low pressure drop of the ES filter for
the first 14 hours of operation. We cannot explain why this effect was so
much more pronounced in this test than in the previous two tests.
A photograph, Figure 20, taken through the window midway through the test
shows how the top left side of the filter is relatively clean. At the lower
right hand corner some dust appears to have fallen off. A close-up photograph
of the lower part of the dust deposit is shown in Figure 21. Note that the
silica dust forms fiber-like deposits. Early in the test the two filters were
removed so that a photograph could be taken to compare the two types of dust
deposit. This photograph is shown in Figure 22. The ES filter is on the left.
The upper half of the filter is almost clean. The NES filter is on the right.
Here the dust deposit appears relatively uniform and comparatively compact.
A SECOND RESISTIVITY LIMIT
When electrically charged dust is collected on a filter, the charge on
the deposited dust must be conducted through the filter cake to the fabric
and to ground. Thus far it has been assumed that the dust is sufficiently
48
-------
TABLE 7
CONDITIONS FOR TEST B-3
Silica dust fed using geartooth dust feeder and Venturi nozzle
with 10 Ib pressure applied.
2
Cloth area = .5 ft
Air/cloth = 6 cfm/ft2
Cylindrical filters used with removable window
Field spacing = 2.5 cm
Applied field voltage = 8.5 Kv
Fabric is AFC //115 50% wool 50% acrylic. Fabric resistivity is
5.9 x 1010
Fabric permeability = 130 cfm @ .5" H_0.
Test data
Field No field
Dust collected by
filter 971.2 g 208.7 g
Dust loading 4.50 gr/ft3 4.64 gr/ft3
Deposit depth .815 cm* .235 cm
Deposit weight 65.4 g 33.6 g
3 3
Deposit density .173 g/cm .307 g/cm
Time required for fil-
ter to reach 10" H_0
Weight efficiency 99.69%
ter to reach 10" H20 18.0 hr 3.75 hr
^Estimated average depth
49
-------
5ILICA DUST
WOOL ACRYLIC FELTED
B A ES -FILTER
A ° NES- FILTER
10
12
14
16
18
TIME(Hours)
Figure 19 - Pressure Drop vs. Time - Test B-3
-------
I
Figure 20 - Filter As Seen Through the Window
51
-------
Figure 21 - Close-Up View of Dust
52
-------
Figure 22 - Photograph Showing Both Filters Early in the Test
53
-------
conducting so that the voltage drop or gradient in the dust is small. However
there will be some resistivity at which this will not be true.
Example or Analysis
Assume that the charged aerosol has a charge density of q (coulombs/m3)
and the dust has a resistivity p (ohm-cm). The filtering velocity (air/cloth
ratio) is V (cm/sec). Since it is customary to specify resistivity in ohms-cm
we will use centimeter rather than meter in the calculation.
The current per cm2 is:
i = q/106 x V = qV • 10~6 A/cm2
(7)
The resistance of a cm2 of filter cake of thickness t is pt. The voltage drop
in the dust layer is pt • qV • 10~6 volts. The voltage gradient in the dust
layer is
E = P-t-q^MO-6 = p.q.v.10-6 volts/cm (8)
Assume q = 10~5 coulombs/m3.
Assume V = 3.04 cm/sec(air cloth = 6 ft/min).
If p = 10ll* fl cm:
E = 101*1' 10~5'3.04-10~6 = 3.04 x 103 volts/cm.
This is approximately equal to the field that can be applied at the fil-
ter surface. Thus the dust surface can no longer be considered as an equi-
potential and in theory there would be no tendency to produce a porous dust
layer.
The resistivity of dust depends on the contact between individual par-
ticles. As measured for electrostatic precipitation the dust is reasonably
compact. However when dust is deposited on a filter in the porous form,
microscopic examination indicates that the particles often have a chain-like
formation with apparently poor contact between adjacent particles. Thus the
resistance of a filter cake may be consideralby higher than would be indicated
by the conventional resistivity as determined for ESPs.
Note that the electric gradient in the dust layer (Eq. 8) is proportional
to the product of resistivity by charge density. The value of 10~* coulombs/
m3 is fairly representative but it could be as high as 10"1*. On the other
hand with low dust loadings the charge density could be quite low. Thus the
critical resistivity is dependent on the charge density and the air/cloth
ratio.
In test B-2 the charge density averaged about 2 x 10~5 coulombs/m3. At
the end of the test the charged aerosol was heated so as to give a dust re-
sistivity of 1013 ohm-cm. The test is far from conclusive but it indicated
that at 1013 pressure drop was not as low as for a lower resistivity dust.
54
-------
This is in reasonable agreement with the above calculation. In fact it is a
better agreement than we expected. We had expected that the porous nature of
the dust would cause a very large increase in resistivity. Thus it had been
considered probable that at a resistivity of 1013 the ES filter would show no
advantage.
Charging some dust positively and some dust negatively, mixing these two
and applying a potential between parallel filter surfaces is described in U.S.
Patent #3,966,435. In this case the positive dust would be drawn to the
negative dust to the positive filter. In case the dust is naturally charged
with positive and negative polarities only the voltage between filters is
needed.
A company was interested in this possibility and supported tests to see
if it could be applied to one of their problems. It was found that in that
particular case, there was the necessary mixture of + and - particles. At
room temperature and a high humidity (> 50%) the dust had a low resistivity
and a marked reduction in pressure drop was obtained. However in their appli-
cation the dust was at a temperature such that the resistivity was greater
than 1013 n cm. When this resistivity was duplicated in the laboratory the
pressure drop of the ES filter was essentially equal to the NES filter. No
further work was done.
DISCUSSION AND CONCLUSIONS
The primary purpose of these tests was to compare an ES filter with a NES
filter. Since static is dependent on the atmosphere and since the dust may be
variable, the filters being compared were operated simultaneously, using fil-
ters cut from the same piece of cloth, and operated from the same dust stream.
In all cases the ES filter showed a distinctly lower pressure drop than the
NES filter.
There are important differences between the different tests. In test B-3
the ES filter operated at a very low pressure drop for the first 14 hours.
The other tests did not exhibit this low initial rate of rise of pressure. We
can only speculate as to the cause. Apparently the low rate of rise of pres-
sure was due to the fact that the upper part of the filter remained almost
clean for this 14 hour period. Observations made through the removable
3" x 3" section indicated that the dust was being electrostatically precipi-
tated on the lower part of the filter. However we do not know why this did
not happen in the other tests.
'These tests indicate that, under the proper conditions, a significant
reduction in pressure drop can be obtained, if the particles are charged, and
if there is a field driving the particles toward the filter. However we do
not know of any industrial situation in which the particulates have the neces-
sary charge as the dust originates. Also most dusts have too high resistivity
for satisfactory charging by corona. So there is no obvious application in
industry. If work on precipitation should develop a means for charging high
resistivity dust that would open up new possibilities both in precipitation
and filtration.
55
-------
While the possibility of reducing pressure drop in a filter appears to be
established, many questions would have to be investigated before a reliable
process could be developed. We are not investigating these questions further
because the work does not appear to be economically justified until there is
some probability of industrial application.
56
-------
SECTION 12
ELECTROSTATIC AUGMENTATION OF EFFICIENCY
OF FABRIC FILTERS
INTRODUCTION
In Section 11 it was shown that electrically charged dust in combination
with an electric field driving particles toward the filter can materially re-
duce the pressure drop. Since this deposit is more porous it raises the
question that the filter may be less efficient.
PURPOSE
To measure the effect of particle charge and electric field on the effi-
ciency of collection of fine particles.
THE TESTS
In preliminary tests, tobacco smoke, atmospheric air, kerosene smoke,
and smoke generated by an electric arc were all tried as test aerosols. All
gave similar increases in efficiency but the electric arc smoke seemed to be
the most reproducible and so was used for the tests. Some preliminary tests
were made using a filter having a previously deposited filtercake of other
dusts. However a clean fabric filter was subject to less variation and was
more definable and so each test started with a clean fabric. Charge carrying
ability is especially sensitive to fine particles and the test can be made
quickly and so was used in these tests.
A dilute smoke was used and the tests were made rapidly so that there
would be a minimum formation of filter cake and consequent change in efficiency
during the test.
The filter used was 4" diameter by 6" long inside a 6" diameter steel
tube as shown in Figure 16. Each test started with a new piece of fabric.
' First efficiency was measured as a function of face velocity with no
applied field and no corona charging. This curve is labeled "A". Next the
test was repeated with an applied field. This curve is labeled "B". Finally
the test was repeated with the aerosol charged by corona and with an applied
field driving the charged particles toward the filter (curve "C").
Comparative efficiencies are shown for different fabrics in Figures 23,
24, 25, 26 and 27. With charged particles and an electric field the efficiency
is 99% or better at 2 ft/min through the fabric. For the #92 Darvan the
57
-------
lOOr
90
80
o
Si 70
o
iZ
u_
UJ
60
UJ
o
(T
UJ
Q_
50
99.9 99.6
B
FABRIC -*92 DARVAN
PERMEABILITY-54
8
10
12
VELOCITY THROUGH FABRic(ft/min)
Figure 23 - Electrostatic Augmentation of Efficiency
of Clean Darvan Fabric
A - No field no charging
B - Electric field, no charging
C - Electric field and charged dust
58
-------
100
99.9
96.9
90
\
h
80
o
uj 70
o
u_
u.
UJ
LJ
O
IT
UJ
60
50
B
FABRIC - *75 WOOL ACRYLIC
PERMEABILITY-100
40
246 8 10
VELOCITY THROUGH FABRIC (ft/min)
Figure 24 - Electrostatic Augmentation of Efficiency
of Clean Wool-Acrylic Fabrics
12
59
-------
100,-
90
o
u.
LL.
UJ
UJ
U
80
70
60
UJ
°- 50
FABRIC- SI473-L WOVEN
PERMEABILITY -75
40
«M
1
2 4 6 8 10
VELOCITY THROUGH FABRIC (ft/min)
Figure 25 - Electrostatic Augmentation of Efficiency
of a Clean Woven Fabric
12
60
-------
100
90
o
z
tul
O
u_
LJ
80
-a a—a a s
c
-a 9 9.6%
70
UJ
o
or
LJ
Q. 60
1
1
2 4 6 8 10
VELOCITY THROUGH FABRIC (ft/min.)
12
Figure 26 - Electrostatic Augmentation of Efficiency
of a Clean Woven Glass Fabric
61
-------
lOOr
99.6 -^
-C
90
80
570
UJ
o
ul
u. 60
UJ
UJ
o
£ 50
Q_
\
FABRIC *I8POLYETHYLENE
40
I
246 8 10
VELOCITY THROUGH FABRIC (ft/min)
Figure 27 - Electrostatic Augmentation of Efficiency
of a Clean Polyethylene Fabric
12
62
-------
efficiency is 99.9% at 2 ft/min and 99.6% at 10 ft/min. For all other
fabrics the efficiency falls off more rapidly with velocity. The specifi-
cations of the various fabrics are given in Volume 2 of this report.
In all cases charging the particles combined with the electric field gave
a marked increase in efficiency. In all cases the electric field alone gives
some increase in efficiency. It may be that the particles as generated in the
arc carry a small charge.
There seemed to be little correlation between efficiency of the clean
cloth and the permeability of the cloth.
DISCUSSION
Theoretically the mobility of the fine charged particles should be about
10~3 cm2/volt sec. The electric field perpendicular to the fabric can seldom
exceed 5 Kv/cm. Thus the velocity of the particles relative to the air could
be of the order of 5 cm/sec or 10 ft/min. Within the fabric local air velo-
cities will be higher than the face velocity and electric fields relatively
low. Thus it appears that electric forces should be most effective in
capturing particles on the tips of the fibers or nap and this seems to agree
with microscopic examination of the dust on the fabric.
The large increase in efficiency of a clean fabric achieved by electro-
static means shows that this is an effect which should be investigated fur-
ther. The poor correlation between permeability and efficiency for the clean
cloth is another area for further study.
63
-------
SECTION 13
MEASUREMENT OF AGGREGATION OR AGGLOMERATION
OF DUST DISLODGED FROM A FILTER
BACKGROUND
In bag filters dust dislodged from a filter must fall downward to the
hopper. In a pulse jet system the dislodged dust must fall against upward
flowing air. For dust to fall satisfactorily into the hopper it must be dis-
lodged as aggregations or agglomerations and not in the form of the original
individual particles. So a test is needed to measure the ability of dust to
fall into the hopper when it is dislodged from the bag.
It has been proposed that the agglomeration of dust is accentuated if
the dust and fabric are widely different on a triboelectric series. The
falling of dust provides a measure of the degree of aggregation or agglomera-
tion of the collected dust.
PURPOSE
The primary purpose of the test was to measure the ability of dust dis-
lodged from the filter to fall against a controlled upward air flow. Other
objectives are to measure pressure drops and to compare filter performance
with triboelectric properties of dust and fabrics.
APPARATUS
The apparatus utilizes two 4" x 4" filters mounted on opposite sides of
a 1" x 4" vertical duct. An adjustable speed fan recirculates the aerosol
through this 1" x 4" duct. This vertical velocity can be controlled by ad-
justing the speed of the fan. Air is drawn off through the filters by another
adjustable speed fan. Air from a dust feeder in an amount equal to that drawn
off through the filters is injected into the upward air stream below the fil-
ters. In this way the air velocity through the filter (air-to-cloth ratio)
can be adjusted independently of the upward air velocity past the filter.
In a pulse jet system the upward air velocity varies from top to bottom
of the bag and also varies with the design of the baghouse. In this test
apparatus, any combination of position and design can be represented since
upward air velocity and air-to-cloth ratio can be controlled independently.
The arrangement of the apparatus is shown in Figure 28. There is a
settling chamber above the filter, called "Bin A," which collects dust,
which is dislodged from the filter and carried upward by the air stream.
64
-------
ON
Ui
TO ORIFICE
8 FAN
BIN A
V FILTERS4"X4"
DUCT
- . ii -i ii
I x 4
DUST FED HERE
-» I PVC
/ \
..
1 ~*
*- 3 PVC
4" ID. — ^
BINB
\^_ _—--' — •*
i> ^
FAN
^V
n
/ /
/ /
Figure 28 - Apparatus for Measuring Aggregation of Dislodged Dust
-------
There is another settling chamber below the filter called "Bin B." If dust
dislodged from the filter is able to fall against the upward air velocity, it
will be collected in Bin B.
RESULTS OF TESTS
The upward air velocity was varied from 200 to 600 ft/min. In all cases
the filtering velocity (air-to-cloth ratio) was 6 cfm/ft2. The results are
listed in Table 8.
In six tests there was a cleaning pulse every minute. In the other tests
a reverse air pulse was released whenever the pressure drop reached 4" of
water. In this case there are two significant numbers, the number of cleaning
cycles per hour and the pressure drop after a cleaning pulse. Since the
initial cleaning periods may be much longer, the cleaning cycles per hour are
given for the first two hours and for the remaining time.
From a triboelectric standpoint, silica and the dust from the molybednum
plant are our most negative dusts and Portland cement the most positive dust.
Wool-nylon is our most positive fabric and Teflon a very negative fabric.
These tests did not demonstrate any significant effect due to relative tribo-
electric properties of dust and fabric. However, the fabrics varied in other
characteristics. Thus much more elaborate tests would be required to reach
any definite conclusion.
Another test included in the series is the ratio of the forward to re-
verse pressure drop. At the end of each filtering test, the following sequence
of measurements was made.
Measure forward pressure drops for 2, 6 and 8 ft/min in the forward fil-
tering direction. Pass air in the reverse direction to attempt to blow off
the filter and measure pressure drops for 2, 6 and 8 ft/min. Repeat this
sequence for a total of 3 forward and 3 reverse flows. A typical set of
pressure drops is shown in Table 9 for test 26. Presumably dust is dislodged
only during reverse flow. So the comparative ratio can be the pressure drop
1R/2F or 2R/3F. There is usually little difference between these two ratios.
In Table 9 the last ratio 2R/3F is tabulated.
66
-------
TADI.E 8. FILTERINO WITH VARIABLE UPWARD AIR VELOCITY
Ov
•-4
Tent
16
24
23
25
26
28
30
29
31
33
34
36
37
20
22
1-'
18
14
21
38
Upward
Air
Velocity
Ft /Mln
200
240
240
240
240
240
240
240
240
240
240
240
240
400
400
400
400
400
600
600
Dust
S
C
C
C
C
C
C
C
C
C
C
n
R
c
c
s
e;
S
c
c
Fabric
UN
1711
IPI
T
T
n
D
n
n
0
D
n
DT
WN
Wl
"N
'fll
T
UN
0
Room
R
H
14Z
20Z
23Z
21Z
20*
21Z
25Z
20Z
333!
18Z
50Z
IR*
20Z
15=
137!
16Z
207
22X
18Z
60Z
* Percent of dint th.it Is nble to
Oust Tvno
S -
C -
I! -
r.lllcn
("orient
llro-m I'prro Moly
Z Of
Hunt
In
Anin D
56
75
58
76
60
67
59
59
48
60
78
53
45
73
47
30
37
39
28
26
Cleaning Cvcl e
T
I P.emnlnlnp,
0-2 llr. Time
1 3
45 160
1 C.C. Every Mln.
1 C.C. Every Vln.
38 59
45 110
21 46
1 C.C. Every Mln.
1 C.C. Kverv Mln.
1 C.C. F.vcrv Mln.
18
12 25
.5 1
1.5 2.25
1 C.C. Kverv Mln.
.5 3.2
2.5 5
2.5 5.3
3.5 3.6
15 25
Prensure
Drop
After
" " 2
.6
3.4
1.5-4.5
1.0-3.0
1.0-3.25
1.0-3.5
1.0-3.0
0-3.5
1.0-3.5
1.0-5.0
1.0-3.0
4.0
4.0
2.0
1.5-2.5
.8
1.5
1.6
2.0-2.5
1.0-2.0
mint
l.ondtnn
r.rnlna
ft3
_
3.0
2.1
3.2
2.0
4.0
3.7
3.25
3.4
3.5
3.')
1.2
7
.2
1.5
-
.00
_
.302
_
Duration
Of
Tent
Hours
5.25
4.5
5.0
4.0
5.0
5.0
5.0
4.0
6.0
4.0
3.5
3.75
4.0
5.0
5.5
4.0
4.5
3.5
10.0
4.0
Pntio
l-'orwnrd
reverse
Air
Pressure
1.82
4.5
4.75
2.15
2.0
3.1
2.7
2.75
2.8
3.0
2.5
2.6
2.0
J.33
2.13
1.92
3.9
1 .07
3.0
2.1
fill nnninit stnteil im<-'.ir
-------
TABLE 9. MEASUREMENT OF FORWARD AND REVERSE PRESSURE
D90P: THE CONCLUSION OF TEST «3
Face Velocity
CFM
FT2
Pressure Drops (inches
of H20)
IF 1R 27 23 3F 3R
2
6
8
1.1 .4 .7
4.5 .9 3.35
7.4 1.3 6.3 1.
Ratio 2F/1R - p| •
at 8 ft/min.
35 .6 .3
85 3.1 .8
25 5.9 1.2
4.85 3F/2R - p
4.7
F - Air flow In the forward or filtering direction
R • Air flow in the reverse or cleaning direction
68
-------
SECTION 14
NODULAR DEPOSITS IN FABRIC FILTERS
INTRODUCTION
One phenomenon in fabric filtration is the appearance of agglomerations
or clumps of dust suspended from one or more fibers. The most obvious examples
of these nodules or 'dingleberries' are those which hang away from the filter
surface. Two examples are shown in Figure 29. One apparent effect of these
deposits is the extra weight added to the fibers that hold them. Without
closer inspection, these nodules appear to be just an interesting anomaly.
But this discussion will show that deleterious effects can arise if these
formations are left on the filter. The persistence of the nodules even after
many cleaning cycles relates to their particle size makeup and the mechanical
holding action of the fibers. For nodules held close to the surface of the
filter, a 'check valve' effect can arise giving a high pressure drop in the
filtering direction but low pressure drop in the reverse or cleaning direction.
High speed movies show this effect and give a possible clue to the mechanism
of formation of nodular deposits.
NODULE COMPOSITION
An important property of nodular deposits is their particle size makeup
compared with the particles present in the inlet dust stream. Does the nodule
just mirror the inlet dust in terms of particle size distribution? To answer
this question, many nodules were examined in the Scanning Electron Microscope
(SEM). These nodules were formed during laboratory scale filter tests. Dif-
ferent types of fabrics and cleaning cycles were employed. Two types of dust
were used and with both, nodular deposits eventually formed. The first type
of dust used was an asphalt plant rock dust composed primarily of silicon,
aluminum, and iron. The second type of dust used was silica. Samples of each
dust used are given in Figure 30.
A rock dust nodule that formed on a Noraex felt filter is shown in Figure
31. A view of the individual particles in the nodule is shown below. This
nodule was removed from the filter after 46 hours of filtration. The air/
cloth ratio was 6 cfm/ft2 over the entire filtration period. The dust loading
was 2.5 grains/ft3 for the first 22.5 hours, 5 grains/ft3 for the remainder of
the test. The cleaning method used was a reverse pulse of air created by the
release of compressed air from a tank through a solenoid valve triggered by a
differential pressure switch. It is important to note that the particle size
found in the nodule is much smaller than the maximum particle size seen in the
rock dust (Figure 30). And although many of the larger particles in the inlet
dust stream settle out before reaching the filter surface, the maximum par-
ticle size found in the nodule is much smaller than the largest particle
69
-------
(a)
(a)
(b)
Fig. 29 - Nodules hanging from
fibers: (a).Rock dust,
(b) Silica oust.
(b)
Fig. 30 - Inlet dust particles:
(a) Asphalt plant rock
dust, (b) Silica dust.
70
-------
reaching the filter. Many 30-40 micron particles reach the filter but were
not found in the nodule.
A silica nodule found on an acrylic fiber after 52 hours of filtration is
shown in Figure 32. The individual particles are shown below. The filtration
cycle consisted of a dust loading of 3.1 grains/ft3 at an air/cloth ratio of
6 cfm/ft2- The fabric used was a 50/50 wool/acrylic felt with a permeability
of 130 cfm at 0.5 inches H20. The filter was cleaned with a reverse pulse of
air when the pressure drop reached approximately 4.5 inches H20. The maximum
particle size found in the nodule is about 10 tricrons. Again, as in the first
case, only the smaller particles of the inlet dust stream are found in the
nodule. This has been typical in every nodule observed even through filtra-
tion conditions were modified.
The persistence of these small particle formations even after repeated
cleaning cycles is consistent with past studies of dust adhesion. Probst5
has shown that in an electrically deposited dust layer (dust charged and pre-
cipitated) , the smaller particles will exhibit a higher adhesive force than
the larger particles. Figure 33 shows this relationship for four different
dusts. The particles were separated into different sizes, deposited electri-
cally, and the force required to remove the dust layer measured with a verti-
cal centrifuge. A later study by NTiedrae showed that randomly deposited
particles not using an electric field to aid deposition can exhibit adhesion
of the same order of magnitude as the electrically deposited sample. In the
randomly deposited sample, adhesion again decreased rapidly with increasing
particle size. The adhesion of 10 micron silica was 8-15 gm/cm2, but only
0.5 gm/cm2 for silica particles in the 37-44 micron range.
While the cleaning cycles remove the less adherent deposit, the very
adhesive particles continue to stick to fibers or to each other. As such,
nodule formation may be a random process. Of the many particles that touch
fibers or other particles, only the adhesive particles remain after the
cleaning cycle. It should be noted that the randomly deposited sample allowed
each particle to arrive and orient itself prior to touching another particle.
The opposite of this would be to pour a quantity of dust onto a metal plate
where most of it would slide off easily if the plate were tilted. Nodules are
very adherent and the long time required to form the larger nodules suggests
this random formation process where certain particles and aggregations of
particles remain even after repeated cleaning cycles.
THE MECHANICAL HOLDING ACTION OF FIBERS
The nodules which are easily observable are those which hang away from
the fabric surface as in Figure 29. Most of these nodules are held by single
5R.E. Probst, "Dipole Phenomena and Particle Size Effects in Electro-
static Precipitation," Ph.D. Thesis, Carnegie Institute of Technology,
Pittsburgh, PA (1962).
6J.M. Niedra, "Electrical Effects in the Adhesion of Powders," Ph.D.
Thesis, Carnegie Institute of Technology, Pittsburgh, PA (1964).
71
-------
(a)
(a)
PM&-W Kt$5f!
MM?mi iti
&".*2AVft*
^ *>&
(b)
Fig. 31 - Rock dust nodule and
particle makeup:
(a) Nodule and nomex
fiber, (b) Particles
in the nodule.
EvLi^3Bpi^
L-wiv
(b)
Fig. 32 - Silica dust nodule and
particle makeup: (a) Nodule
and acrylic fiber, (b) Particles
in the nodule.
72
-------
fibers. The nature of the holding action will be discussed in this section.
A direct look at the site of the nodular formations gives a clue to the
reason for their ability to hang from the filter surface. The holding of the
nodule to the fiber is best described by the geometry of the spot on the fiber
where it is located. Nodules can also be held by more than one fiber. This
becomes important later in describing the check valve effect. Small nodules,
about 10 fiber diameters wide (about 200 microns), are often spherical, in-
dicative of an abrading process whereby the less adherent particles and pro-
truding edges are removed during successive cleaning cycles while the nodule
itself remains intact.
Nodules, when observed after a long filtration cycle and many cleaning
cycles, are found to encompass kinks or irregularities on the fibers. Several
examples are mentioned here. A silica nodule taken from the same filter as
the nodule in Figure 32 was examined in the SEM. The acrylic fiber on which
it was found entered the nodule at one point but appeared to divert from this
path before exiting the nodule. By pressing a pointed object to the nodule,
it fractured revealing the underlying structure of the fiber. The fiber had
a bend or kink in the area where the nodule once surrounded it. The nodule,
when fractured, split into several pieces which fell away from the fiber with
little of the original nodule still clinging to the fiber.
A second silica nodule formed on the same type of fiber indicated how an
irregularity may accompany a nodule. When viewed in the SEM, a short piece of
fiber was found exiting the nodule along with a fiber which appeared to pass
directly through the nodule. But, after fracturing the nodule, two separate
fibers were found entering the nodule. The short fiber exiting the nodule was
the termination of the fiber entering from the filter side of the nodule. The
other fiber was not connected to the filter when the nodule was removed from
the filter. Another example of a fiber irregularity had the fiber twisted
almost 180 degrees at the center of a silica nodule.
Whether these sites are the actual formation areas of the nodules is un-
certain since modules have been observed to have the capability of sliding on
a fiber, remaining intact. This would explain the appearance of nodules on
fibers which are quite a distance from the filter, as in Figure 29. The
nodule could slide until stopped mechanically by a kink or irregularity.
The cleaning cycle mentioned previously would also tend to cause nodules
to slide away from the filter surface. The air pressure pulse causes the
fabric to accelerate outward providing a snap action. Some nodules would con-
tinue to remain after the cleaning cycle and would be found hanging relatively
far from the fabric as in Figure 29. While the single fiber nodule is common,
large nonspherical multiple fiber nodules also appear. The cleaning technique
presumably affects the presence or absence of the multiple fiber nodules.
THE CHECK VALVE EFFECT
From an engineering standpoint, the results presented so far do not des-
cribe the dynamic filtration effects of nodular deposits. This section will
show how nodular deposits exhibit themselves in terms of pressure drop across
73
-------
a filter. A key observation is the appearance of the check valve effect in
filters having these deposits. Comparison of forward and reverse flow measure-
ments and direct observation of the dust cake movement under reverse air flow
complete the examination of the valve effect.
The nodules which hang away from the fabric surface as in Figure 29 do
not affect the overall pressure drop characteristics of the filter. But the
nodules which are close to the fabric surface can influence the forward and
reverse pressure drops of the filter. As used here, the forward pressure drop
is the pressure drop in the normal filtering direction. The reverse pressure
drop is the drop in the cleaning direction. A model of nodule action during
the application of forward and reverse air flows is presented in Figure 34.
The best way to observe this type of action is by applying forward and
reverse air flows and, while this is occurring, recording the movement of the
dust cake with a movie camera. This was accomplished using the rock dust and
nomex filter mentioned previously. With the dust cake facing up, both forward
and reverse air flow pressure drops were measured at air/cloth ratio between
3 and 11 cfm/ft2. The plot of pressure vs. air/cloth ratio for these two
cases is given in Figure 35. The reverse flow pressure drop is considerably
less than the forward flow drop. A sequence of photos made from the vertical
and the dust layer is shown with no applied flow. In the second and third
photos, a reverse air flow is applied for increasing air/cloth ratios. The
last photo is at an air/cloth ratio of approximately 6 cfm/ft2. The opening
of the crack in the center of the pictures is a visual example of the check
valve effect. The filter cloth was securely held and its motion under this
reverse flow is negligible. While this same type of movement has been docu-
mented for the smaller type of nodular deposit previously discussed, this
sequence of photos gives the best visual picture of the effect and the analogy
to a valve opening and closing under forward and reverse air flows.
The check valve effect, exhibited by differences in forward and reverse
pressure drops, can appear soon after filtering begins even though nodules
which are large enough to be easily seen may not become apparent until much
later. An example of this early presence of the check valve effect is shown
in the pressure drop characteristics of Figure 35. This is a time history of
forward and reverse pressure drops at various times in a filtration cycle for
two fabrics filtering the same dust. The difference in forward and reverse
pressure drops is seen at hour 16 and increases considerably by hour 46. By
hour 52, the ratio of forward pressure drop to reverse pressure drop is about
5 to 1 even though the reverse pulse cleaning was employed.
A THEORY OF FORMATION
While the effects of nodular deposits have been indicated and examples of
nodules presented, the cause of this type of formation is unclear at this time.
In this section, evidence will be presented which indicates how such a deposit
might form. It would be best to observe one fiber over a filtration period
and watch to see if a nodule formed. An easier approach is to observe the
break-up of the dust cake during a cleaning pulse. It was stated previously
that nodules could slide along a fiber. This tends to indicate lower adhesion
74
-------
V SILICA
D FLYASH
O ZINC
A BRONZE
40 80 120
PARTICLE SIZE
200
Fig. 33 -
Adhesion Force vs.
Particle Size for
Four Materials
(After Probst5)
Fig. 36 - Opening of nodules
subjected to a
reverse air flow.
From no flow (top
photograph), the
air/cloth ratio is
increased to about
6 cfm/fr (bottom
photograph).
DUST CAKE
( "~*~\ REVERSE FLOW
FORWARD FLOW
Figure 34 - A Model of Nodule Action
-------
I I I I
| ) I UHWAW) I I OW
O Kl VI KM I I OW
4 6 8 10
AIR/CLOTH RATIO (CFM/FT2)
i. 0
o
/•40
20
I AIINIV
UAHVAN
WOOL
NYi ON
• I OKWANI) MOW
[") RLVlHSt t IOW
FORWARD FlOW
1(1 VI 1(1,1 M OW
(o
HOUR 16
369
AIR/CLOTH RATIO (CFM/FT2)
(a)
(b)
0 O
40
20
HOUR 46
I. _.J J_.
369
AIK/CLG1 It HA1IO (CFM/FT
60
(c)
AIH/CLOTH RATIO (CfM/FT11)
(d)
Fig. 35 - Forward/reverse pressure drops of rock dust and nomex felt,
(a), and time history of forward/reverse pressure drops for
two fabrics, (b, c, and d).
76
-------
between the particle and fiber than particle to particle in these tests. If
this is the case, it would seem reasonable to expect creation of a nodule if a
dust layer would fracture while surrounding a fiber leaving a core on which
other particles may deposit.
Several dust cake break-ups were investigated to ascertain whether nodule
formation could be related to the reverse pulse cleaning technique. One par-
ticular example of the effect of a reverse pulse was for a nomex felt filter
filtering the rock dust shown in Figure 30. The filtration cycle was the same
as that for the nomex felt/rock dust case presented in the first section. The
test was stopped after 22.5 hours. A high speed camera recorded the action of
the dust cake during the reverse pulse. The time between individual frames
was about 1.55 milleseconds which was adequate for observing the dust cake
break-up. A particular sequence of frames began just before the pulse arrived
and continued until the dust, flying off the filter, obscured everyting.
The high speed moving picture can be described as follows - the first 4
frames (6.2 milliseconds) show the initial effects of the pulse. As the
pressure wave struck the filter, the gas was forced through small cracks in
the dust cake giving a geyser-like appearance. Then, after 6 milliseconds, as
the pressure increased, these areas could no longer release the gas fast enough
and the cake fractured. The fractured pieces opened up to release the gas but
still remained on the fabric surface which had moved very little up to that
point. This opening up of the cake at this stage was very similar to the mo-
del for nodule movement under reverse flow given in Figure 34. As the pres-
sure kept increasing, these pieces opened up more in an effort to release the
gas. At the end of the photographs, the fabric began to move. This fabric
movement can be compared to a test in which a clean filter of the same material
received the same cleaning pulse. For the clean filter, the maximum excursion
was reached 12.5 milleseconds after arrival of the pressure pulse. The maxi-
mum distance travelled by the filter was 13 mm and 85% of this distance was
covered between 9.5 and 12.5 milleseconds. So most of the fabric movement
occurred in a relatively short time span. Although the movement of the dust
laden filter was not measured during the sequence above, comparison to the
movement of the filter gives a rough estimate of when the fabric begins to
move rapidly neglecting the inertial dust loading. With this approximation,
the release of the dust cake from the fabric surface corresponds to the time
the fabric begins to move. The photographs show this to be a reasonable
assumption.
The two major events seen in the reverse pulse application were 1) the
fracturing of the dust cake before removal from the surface which may be an
indication of a nodule forming mechanism, and 2) the snap action of the fab-
ric removing the fractured pieces of dust cake from the surface.
CONCLUSIONS
Nodules have been shown to be composed of the small particles of the in-
let dust stream. This is consistent with past studies showing small particles
exhibiting higher adhesion than large particles. The smaller nodules often
encompass kinks or irregularities in the fibers which can provide a
77
-------
mechanical constraint preventing removal from the fiber. Both small and large
multiple fiber nodules evidence themselves by the check valve effect producing
a difference in forward and reverse pressure drops at a constant air/cloth
ratio. The break-up of the dust cake prior to release from the fabric during
a reverse pulse suggests one method by which nodules may form.
78
-------
SECTION 15
DEMONSTRATION OF ELECTRIC FIELD
BETWEEN ADJACENT FIBERS OF A FILTER
INTRODUCTION
A student, Arthur E. Uber III, made tests to explore possible fields be-
tween adjacent fibers of a filter. For an aerosol he dissolved food colorings
in alchohol. This solution was dispersed into air using a nozzle giving a
fine spray. As the alcohol evaporates this left submicron particles of red
and blue coloring. An aerosol of one color, e.g., red, could be charged
negatively by corona and passed through a filter. Then a blue aerosol could
be charged positively and be passed through the same filter. The filter
could then be examined under the microscope. With many filters red (negative)
and blue (positive) particles tended to deposit on different positions, some-
times different portions of the same fiber. The most striking cases were
obtained using a resin-treated gas-mask filter (Hansen filter). A photomicro-
graph showing three fibers of such a filter in which red particles had been
deposited is shown in Figure 37. This is a striking example of particles of
one sign continuing to deposit on one location.
DISCUSSION
The work demonstrates the existence of electric fields between adjacent
fibers but it raises many questions. To resolve these questions would require
a much more extensive investigation.
The conventional explanation for the Hansen filter is that static charges
exist on resin surfaces and produce the fields between fibers. Davies7 points
out that any reasonable calculation of life time of a static charge on resin
cannot account for the observed life time of the filters. Another possible
explanation is that there are contact potentials between adjacent fibers.
Many of the phenomena cited to support the static charge theory can also be
used to support the contact potential theory. An investigation of the contact
potential theory would involve the work function or contact potential of the
particulates collected as well as the fibers.
This work was not carried further because these fields between fibers
deposit dust within the fabric where it is difficult to remove. We were
looking for means for depositing dust on the fabric rather than within fabric.
Furthermore the questions to be answered are very involved so that it appears
that a major effort would be required to obtain meaningful results. However
this does appear to be an important area for basic research.
7C.N. Davies, Air Filtration. Academic Press, New York (1973) p. 108.
79
-------
Figure 37 - Negatively Charged Particles Attracted to Particular Fibers
80
-------
ADDENDUM
ELECTRIC FIELDS IN FABRIC FILTERS DUE
TO NATURAL FIBER PROPERTIES
by
Robert F. Lembach
Electric fields between adjacent fibers in fabric filters can arise from
the retention of static charge or the existence of contact potential dif-
ferences (CPD's) between fibers. Both depend only upon natural fiber proper-
ties eliminating the need for an external voltage source. To retain static
charge for long periods of time, a fiber or dust must be highly insulating.
The Hansen type filter consisting of highly insulating resin particles im-
planted in a wool filter is one example. With static charge placed on the
resin particles, nonuniform fields are produced in the filter which aid in
attracting incoming particles to yield a high efficiency filter. A CPD be-
tween adjacent fibers can produce a similar effect. But whereas static charge
decays with time, the CPD does not unless the surfaces of the fibers are
altered or covered to produce a new surface.
The following sections examine both techniques of producing electric
fields in fabric filters to aid collection. The CPD of several fabric ma-
terials was measured by two methods. The resistivity of a resin compound was
measured to determine the time static charge would remain. Finally, the na-
ture of the particle deposition in the resin type filter was observed.
FABRIC CONTACT POTENTIALS AS MEASURED BY THE KELVIN METHOD
In 1898, Lord Kelvin8 introduced a technique for measuring CPD's of
metals. As an example, when two parallel plates, one zinc and one copper are
electrically connected, there exists an attractive force between them. For
closely spaced plates, this force increases as the inverse square of the dis-
tance separating them. CPD's of metals are typically on the order of 1 volt
or less and can be found in reference texts. A change in surface characteris-
tics by polishing or scratching can produce a change in contact potential. As
a result, this contact potential is often referred to as the surface potential.
There is a desire to determine if forces arising from CPD's are important
in dust collection apparatus, specifically fabric filters where a 1 volt dif-
ference over a 10 ym spacing may give a field of 1 kilovolt/cm. It is neces-
sary to know if CPD's exist for commonly used fabrics and dusts. The tribo-
electric technique of rubbing two materials together, then separating them to
8Kelvin, Lord, "Contact Electricity of Metals," Phil. Mag., 46:82-120
(1S98).
81
-------
find one charged positively, one negatively, indicates the possibility of
CPD's existing between fabric filter materials. Two recent papers9'1" suggest
that a direct relationship exists between the CPD and the charge transferred
upon contact. But just as metal surface potentials can be altered by polishing
or scratching, a technique which did not disturb the fabric surface would be
advisable.
The purpose of these tests was to acquire basic information concering
contact potentials of dusts and fabrics. Another reason was to see if a tri-
boelectric series, determined by bring two materials in contact and separating
them could be manifested by a nondestructive test such as the Kelvin method.
The apparatus, similar to that used by Klingler11, consisted of two electri-
cally connected electrodes (5x5 cm), an oscillating reference and the sample,
forming a parallel plate capacitor. A review of this method is described
in 12. Assuming an electrically linear system, the charge q(x) on a plate is
related to the contact potential difference Vcpd by the capacitance C(x)
q(x) = C(x)
The capacitance is only a function of x due to the mechanical constraint on
the oscillating electrode. The time varying current i(t) flowing in the cir-
cuit is therefore,
i(t) = dq/dt = d(C(x)Vcpd)/dt = C(x)(dVcpd/dt)+Vcpd(dC(x)/dx)(dx/dt)
Since the contact potential difference is constant,
i(t) = Vcpd(dC(x)/dx)(dx/dt)
For this geometry, and by neglecting fringing fields, the capacitance is re-
lated to the separation x between plates, the plate area A, and the permit-
tivity e as
C(x) = eA/x
Constraining the oscillating electrode as
x(t) = XQ + xjsinut
9Davies, D.K. , "Charge Generation on Dielectric Surfaces," Brit. J. Appl.
Phys. (J. Phys. D.), 2(2) : 1533-1537 (1969).
10Garton, C.G. , "Charge Transfer from Metal to Dielectric by Contact
Potential," Appl. Phys. (J. Phys. D.), 7:1814-1823 (1974).
ier, E.H. , "The Adhesion of Dust as Related to Electrostatic
Precipitators," Doctoral Dissertation, Carnegie-Mellon University, Pittsburgh,
PA (1961).
12Surplice, N.A. and R.J. D'Arcy, "A Critique of the Kelvin Method of
Measuring Work Functions," Sci. Instr. (J. Phys. E.), 3:477-482 (1970).
82
-------
where x0 and Xl are constants and u> is the radian frequency of oscillation
yields the current for the system model to be
i(t) = (-VcpdeAx1coso)t)/(x0 + xjsinut)2
which was found to be in good agreement with the experimental current waveform.
Determining the CPD involves insertion of a series voltage source in the cir-
cuit and adjusting the voltage until i(t) = 0. At this point the voltage of
the external source equals the CPD between the oscillating electrode and
sample.
Tests were performed at room temperature and 60-80% RH. The high humidity
was necessary due to inherent limitations on allowable sample resistivity. A
vacuum deposited gold oscillating electrode was used as the zero volt
reference. Table 10 shows CPD's relative to this reference for several dusts
and fabrics. In terms of charge transfer between two contacting materials,
the material with the higher relative potential tends to acquire electrons and
become negatively charged. Of note was the silica dust which tends to charge
negatively in impingement tests. It has a high potential consistent with the
impingement results. There appeared to be no definitive relationship between
fabric samples. However, since typical filter operations operate at a lower
RH, a means of testing higher resistivity materials was developed.
TABLE 10. CONTACT POTENTIAL (VOLTS) AT 60-80% RH
Sample Contact Potential*
Silica dust +0.18 to -0.27
Cement dust -0.26 to -0.52
Wool dust -0.73 to -0.90
60/40 wool/nylon fabric #97 -0.69 to -0.96
85/25 wool/nylon fabric #96 -0.81
Polyester (Dacron) fabric #79 -0.65 to -1.08
Acrylic (Darlan) fabric #125 -0.83 to -1.12
*Contact potential with respect to gold reference (zero volts).
FABRIC CONTACT POTENTIALS AS MEASURED BY THE RADIOACTIVE SOURCE METHOD
The current waveform i(t) derived in the last section assumes that the
sample has a conductivity high enough to allow electrical charges to move
through it at a rate much faster than the oscillating electrode frequency.
This rate is associated with e/o, the ratio of sample permittivity to sample
conductivity. For example, a sample with a conductivity of 10~10 (fl-ra)-1 and re-
lative dielectric constant e_ of 3(e=ereo.eo=8-85xl° F/m) yields an e/o of
about 0.3 seconds. To insure instantaneous changes in i(t) with changes in
83
-------
external voltage, eui/o should be « 1. To eliminate this frequency dependence,
an alternate scheme to measure the CPD was devised.
This technique employed a radioactive source (226Ra, 1.3 pCuries) to ionize
air in the space between the reference electrode and sample. The electric
field due to the CPD causes ions to move preferentially and this current is
again measured by a series ammeter. As before, a series voltage source is
adjusted until the current reaches a null to give the CPD.
The purpose of these experiments was to acquire basic information on dust
and fabrics at a lower RH. The technique employed an ionizing agent and a
fixed reference electrode. The apparatus is the same as in the previous sec-
tion except that the reference electrode is stationary and the radioactive
source is placed near the periphery of the reference electrode.
Table 11 shows results for several fabrics. Where possible, both the
Kelvin method and the radioactive source method were used to determine the
CPD. The numerical difference between the two methods used was typically less
than 100 mV. A range of potentials is listed since typically four samples of
each fabric were tested. Where only a small amount of cloth was available, a
single sample was used.
TABLE 11. CONTACT POTENTIAL (VOLTS) AT 20-30% RH
Sample
Kelvin
Method
Radioactive
Source Method
A 2nd gold sample
Olefin (polyprop) #1
Aramid (Nomex) #2
Acrylic (Dralon T.) #3
Aramid (Nomex) #4
Acrylic (Dralon T.) #5
Polyester (Kodel) #6
Acrylic (Zefran) #7
Acylic (Orion) #8
Polyester/Acrylic
(50/50 Dacron/Orlon) #9
Polyester #19
Polyester (Dacron) #20
75/25 Wool/Nylon #96
Aramid/Glass
(87/13 Nomex/Glass) #113
singed side
+0.082 to +0.031
-0.714 to -0.819
-0.649 to -0.738
+0.064 to +0.022
-0.782
-0.800
-0.803
-0.913
-0.845
-0.763
-0.942
-0.819
-0.855
-0.839
-0.745
-0.767 to -0.850
-0.719 to -0.822
continued
84
-------
TABLE 11. (Continued)
Sample
Kelvin
Method
Radioactive
Source Method
Aramid/Glass
(87/13 Nomex/Glass) #113
plain side
Acrylic (Darlan) #125
-0.573 to -0.720
-0.635 to -0.741
-0.682 to -0.782
-0.707 to -0.800
Underlying these tests was the desire to find materials whose CPD was
large enough to produce large electric fields (several KV/cm) in a filter
composed of these materials. As seen in Table 11, the fabric CPD's were in a
narrow range of -0.6 to -0.9 volts relative to the gold reference. Because of
this, it would be difficult to predict the charge transferred upon contact of
two of these materials. This is consistent with the difficulties encountered
in trying to reproduce triboelectric series although certain trends can be
observed. Since these tests did not find fabric materials with large dif-
ferences in average surface potential, they were stopped at this point.
HIGH RESISTIVITY RESIN PARTICLES IN FABRIC FILTERS
The Hansen wool/resin filter has been treated in several publica-
tions13* ll*. Operation is believed to be due to the highly insulating proper-
ties of the resin particles. Their ability to hold static charge along with
the relatively conducting wool fibers would create nonuniform electric fields
in the filter to aid deposition of both charged and uncharged particles.
Tests were conducted to measure the resistivity of a commercial zinc
resinate compound found in a zinc resinate (ZnR) /wool/acrylic filter which
exhibited the presence of electric fields15. Resistivity of the powdered
ZnR compound was determined by two methods. First, a dust cup was placed in
parallel with an electrostatic voltmeter. A voltage source was applied to
this parallel combination, then removed. The decay of the voltage with time
was recorded. Knowing the meter capacitance, the ZnR resistance was estimated.
Given the dust cup geometry, the resistivity was determined. A second method
employed a point to plane corona discharge onto a 2 mm thick ZnR deposit. A
metal plate connected to an electrometer was brought over the charged dust
layer and a voltage measured. The apparatus was calibrated against a plate at
a known voltage thereby giving the ratio of measured voltage to the actual
voltage at the dust layer surface. Again, the decay of the dust layer voltage
was recorded and a resistivity calculated.
13Davies, C.N. (1973), p. 171.
l*Rossano, A.T. and L. Silverman, "Electrostatic Effects in Fiber Filters
for Aerosols," Heating and Vent., 51(5): 102 (1954).
ISpenney, 6.W. and A.E. Uber, "Collection of Electrically Charged Par
tides in Filters," J. APCA, 26(1) :58 (1976).
85
-------
Resistivity of the ZnR powder ranged from 1016-1019 fl-cm when measured
over a one week period at about 25% RH. Assuming a relative dielectric con-
stant of 3, the relaxation time for static charges would be roughly 1 hour to
30 days.
The next tests examined three filters, a ZnR/wool/acrylic, a wool/
acrylic, and a ZnR/stainless steel filter under similar filtration conditions.
The last filter was created from a stainless steel filter dipped in a ZnR
solution and allowed to dry. Flexing the coated fibers fractured the ZnR
coating to produce fibers partially covered with ZnR crystals much like the
commercial ZnR/wool/acrylic filter.
Small red particles, typically 0.3 vim diameter when viewed with a scanning
electron microscope, were generated by spraying a solution of these particles
dissolved in a solvent and allowing the solvent to evaporate prior to entry
into the filters. An observable deposit was formed in typically 8-10 hours at
a filtration velocity of 6 ft/min. Particles were charged negatively by
corona before entering the filter under test.
The appearance of the deposits was similar in the ZnR/wool/acrylic and
ZnR/stainless steel filters and resembled a recent APCA cover picture16- In
the ZnR/stainless steel filter, tree-like deposits grew out from the fibers.
Few particles resided on the ZnR particles themselves indicating the presence
of a repelling negative charge. No similar type of deposit was discovered in
the wool/acrylic filter without ZnR treatment when run continuously for 24
hours.
Since several time constants are generally required to completely dis-
charge a sample, the ZnR would require, assuming the 1019 Si-cm resistivity,
almost 5 months time. It appears as if this resistivity is high enough to
produce the long lasting high efficiency seen in this type of filter.
Davies17 cites a resistivity of 1021 fi-cm for the best resin compounds which
could explain some of the long lasting effects of the Hansen type filter.
Evidence of internal electric fields was discovered in both the ZnR/wool/
acrylic and ZnR/stainless steel filter. In the latter case, negatively
charged particles deposited on the conducting fibers and not on the ZnR par-
ticles themselves. If the ZnR particles can be charged negatively tribo-
electrically as is done to renew the filter, it may indicate a material with
a high relative surface potential along with its high resistivity. This
additional property could explain Davies1 remark that the Hansen type filter
performs efficiently much longer than expected (i.e., > 2 years). It is
longer than expected since static charges are presumed to have discharged
long before that time.
16Penney and Uber (1976).
17Davies, C.N. (1973), p. 171.
86
-------
SECTION 16
CONCLUSIONS TO VOLUME I
1. Aerosol particles striking a surface usually acquire a frictional charge.
Often the polarity of the charge is erratic. For silica particles
striking a tungsten carbide surface 90 to 100% were charged negatively.
2. Field effects produced by frictionally charged particles can influence
the nature of the filter cake.
3. Corona charging of the dust can be used to produce a more porous dust
deposit than is produced by uncharged dust. The reduction in pressure
drop can be quite significant. However it is not yet commercially
feasible.
4. The collection efficiency of clean fabric filters can be materially
increased by electrostatic effects.
5. The size of aggregations of dust formed during filter cleaning vary with
type of dust, manner of deposition, type of fabric, and relative humidity.
6. Dust collected on a filter may form nodules attached to individual fibers.
These nodules may act as "check valves" during forward air flow, but pro-
vide low resistance to reverse air flow. In these nodules, the inter-
particle adhesion may be sufficient so that these nodules are not removed
in normal cleaning cycles.
7. Electric fields may exist between adjacent fibers of a filter.
87
-------
TECHNICAL REPORT DATA
(Please read liiitructioiis on the reverse before completing)
1. REPORT NO.,
EPA-600/7-78-142a
2.
3. RECIPIENT'S ACCESSION NO.
4. T.TLE AND SUBTITLE Electrostatic Effects in Fabric Filtra-
tion: Volume I. Fields, Fabrics, and Particles
(Annotated Data)
5. REPORT DATE
September 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gay lord W. Penney
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Carnegie-Mellon University
Schenley Park
Pittsburgh, Pennsylvania 15213
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
Grant R803020
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES JERL-RTP project officer is James H. Turner, Mail Drop 61, 919/
541-2925.
16. ABSTRACT,
The report examines the effect of particle charge and electric fields on the
filtration of dust by fabrics. Both frictional charging and charging by corona are stu-
died. Charged particles and an electric field driving particles toward the fabric can
greatly reduce the initial pressure drop and also give some reduction in the steady
state pressure drop. Collection efficiency is also increased. Frictional charging is
unusually erratic and corona charging limited by dust resistivity. Thus further work
is required to develop practical devices. In long term or equilibrium conditions,
nodular deposits attached to fibers may be of major importance. These nodular depo-
sits may act like check valves, giving a high pressure drop in the forward (filtering)
direction, and a low pressure drop in the reverse (cleaning) direction. Thus these
nodules are difficult to remove by reverse air flow. These nodular deposits, as well
as electrostatic fields between adjacent fibers of a filter, appear to be electrostatic.
This appears to be an important area for further research.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Held/Group
Pollution
Dust
Filtration
Fabrics
Electrostatics
Electric Fields
Friction
Coronas
Pollution Control
Stationary Sources
Particulates
Fabric Filtration
Particle Charge
13B
11G
HE
20C
14B
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (Till* Report)
Unclassified
21. NO. OF PAGES
96
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
88
------- |