EPA-600/2-76-132
May 1976
Environmental Protection Technology Series
ELECTROSTATIC CAPTURE OF
FINE PARTICLES IN FIBER BEDS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
E PA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use. .
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-132
May 1976
ELECTROSTATIC CAPTURE
OF FINE PARTICLES
IN FIBER BEDS
by
D. L. Reid and L. M. Browne
Battelle Northwest
P.O. Box 999
Richland, Washington 99352
Grant No. R801581-02-2
ROAPNo. 21ADL-035
Program Element No. 1AB012
EPA Project Officer: Dennis C. Drehmel
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
LEGAL NOTICE
This report was prepared by Battelle as an account of sponsored research
activities. Neither Sponsor nor Battelle nor any person acting on behalf of
either: (a) Makes any warranty or representation, express or implied, with
respect to the accuracy, completeness, or usefulness of the information
contained in this report, or that the use of any information, apparatus, process,
or composition disclosed in this report may not infringe privately owned
rights; or (b) Assumes any liabilities with respect to the use of, or for damages
resulting from the use of, any information,'apparatus, process, or composition
disclosed in this report.
-------
ABSTRACT
The emission of submicron sized particles is the most difficult to
effectively control with current state-of-the-art technology. Experi-
ments at Battelle-Northwest led to the discovery that charged particles
could be efficiently captured by a thick fiber bed having a high bed-
volume to fiber-volume ratio and a very low pressure drop. Interim
results are presented for the removal efficiency as a function of par-
ticle resistivity, superficial gas velocity, bed depth and aerosol con-
centration. In general, for bed pressure drops less than 0.5 in HgO
efficiencies greater than 90 percent were obtained at 300 ft/min face
velocity over the particle size range of 0.06 - 0.7 ym AMD and a particle
o 12
resistivity range of 10 - 10 ohm-cm. The most important variable was
the bed face velocity with the collection efficiency varying inversely
with velocity. The bed depth and volume void fraction were other sig-
nificant variables bearing on the efficiency of the system.
Emperical equations indicate a 100 percent collection efficiency
might be attainable near or somewhat below a bed face velocity of 50
ft/min. However, a more sensitive measurement method would be required
for experimental confirmation.
111
-------
ACKNOWLEDGMENTS
The authors wish to acknowledge the assistance of co-workers D. A.
Nelson for conducting the tests, D. L. Lessor for discussions on the
mathematical modeling of the phenomenon,and the technical guidance and
suggestions of J. A. Abbott and D. C. Drehmel of the Environmental
Protection Agency.
iv
-------
SUMMARY AND CONCLUSIONS
The efficient removal of charged submicron particles from flowing
gas streams with a bed consisting of fibers having a high dielectric con-
stant was demonstrated experimentally. Although not all of the facets
of operation for optimization of the concept have been explored in detail,
the tests conducted to date suggest many practical applications for rela-
tively economical particle control not attainable with presently available
concepts. Based on mass measurements, the emperical equations developed
from the data collected to date indicate that a 100 percent collection
efficiency at some face velocity below 50 ft/min might be attainable but
would be extremely difficult to prove experimentally. Other observations
from the information accumulated to date were as follows:
1- Velocity was the prime variable with respect to efficiency. For the
6 in. thick bed, the efficiency decreased linearly with increasing
velocity up to about 200 ft/min bed face velocity. The minimal data
beyond this velocity are suspect and need redefinition. For the
12 in. thick bed the linear relationship existed up to the highest
flow rate tested of 320 ft/min. Both produced efficiencies above 90
percent at the above-stated flow rates for the submicron particles
used.
2. Bed thickness apparently had a pronounced effect on collection effi-
ciency. However, bed thickness and fiber volume density are dependent
variables as evidenced by the compressed 3 in. bed which showed a
higher efficiency than the 6 in. bed having a slightly higher void
fraction. Although not directly comparable due to void fraction dif-
ferences the 12 in. filter showed about a 10 and 2 percent higher
efficiency than the 6 in. bed at 300 and 100 ft/min face velocity.
This difference in the slope of the curves suggests a deeper pene-
tration with increasing velocity which appears reasonable providing all
other conditions are equal. Thus the ratio of the efficiencies should
increase with velocity for equivalent particle characteristics.
-------
3. Particle Size. Although not a prime variable, there was no detectable
influence on efficiency over the range tested -- ^0.06 to 0.83 ym AMD.
The stated AMD's are thought to be high due to the uncertainty of
extrapolation from about 90 to 50 percent of the log normal plot.
4. Pad Resistivity. There was no discernable difference between polypro-
pylene or teflon fibers with respect to efficiency and apparent field
charge strength and shape. The principal criteria for a fiber are
that it be electrically non-conducting and have a significantly larger
diameter than the submicron particles. A 6 in. thick stainless steel
demister pad produced very low collection efficiencies — somewhat
higher than that expected for image force collection.
5. Particle Resistivity. The majority of the data indicated there was
no significant trend in collection efficiency over the range of
8 12
approximately 10 - 10 ohm-cm particle resistivity. Since one test
g
series showed a slightly lower efficiency for the 10 ohm-cm NHLC1 and
the model suggests that efficiency should decrease with decreasing
particle resistivity, particles with a much lower resistivity should
be studied to better describe any differences and establish the lower
boundary limit, if one exists.
6. Dust Concentrations. Particle concentrations ranging from 10 to 250
mg/M were used without any significant differences except for the
apparent higher field charge for the higher concentrations. Since
measurement reproducibility increased and length of test time varied
inversely with concentration, most of the new wind tunnel tests were
3
made between 50-100 mg/M dust loading.
7. Particle Charge Level. The fraction of uncharged particles challeng-
ing the filter is apparently more pertinent to the success of the
concept than the charge level of the particles at least if near satur-
ation charge is obtained. As would be expected, the bed had no
affinity for uncharged particles at the higher velocities and showed
about 8 percent retention at 50 ft/min. The instrumentation used
was not sensitive enough to provide the refinement required to accur-
ately define the fraction of uncharged particles downstream of the
charge section. Consequently, valid information is not available on
VI
-------
the ratio of charged to uncharged particles reaching the filters.
However, the efficiencies obtained indicate that the uncharged par-
ticles must have been a very small fraction of the total particle
load. Faraday cage measurements indicated both near and supersatur-
ation charge of the particles. The variability of the data is
believed to be a function of the measurement of small currents in the
presence of relatively high electrical fields and very little credence
is given to the data.
8. Humidity. No attempts were made to study humidity of the aerosol as
a variable. The maximum relative humidity observed during the experi-
ments was 50 percent and generally hovered between 30-40 percent.
Consequently, no effects generally applicable to humidity were dis-
cernable. For one test series with Na^O particles, room humidity in
excess of 45 percent produced a liquid NaOH deposit in the charge
section and fiber bed resulting in degraded particle charging and bed
efficiency. The principal cause for the loss in collection efficiency,
particle charging or bed penetration or a combination of both was
indeterminant. Evidence of shorting in the charge section as well as
a lower field charge suggests that inadequate particle charging might
have been the predominant factor.
vii
-------
CONTENTS
ABSTRACT iii
ACKNOWLEDGMENTS iv
SUMMARY AND CONCLUSIONS V
CONTENTS viii
INTRODUCTION 1
BACKGROUND 2
METHODS OF APPROACH 3
EXPERIMENTAL EQUIPMENT 6
TEST SYSTEM 6
CORONA CHARGING SECTION 6
FIBER BED SECTION 8
EXPERIMENTAL CONDITIONS 10
FIBER BED VOID FRACTION 10
PARTICLE GENERATION AND CHARACTERISTICS 11
SAMPLING 12
TEST PROCEDURE 13
RESULTS AND DISCUSSION 14
SPRAY-DROP SYSTEM TESTS 14
Bed Configuration 14
System Efficiency for NH^Cl, Na20 and MgO .... 15
NEW WIND TUNNEL TESTS 18
POLYPROPYLENE BED - SIX INCH THICK 18
TWELVE-INCH THICK POLYPROPYLENE FILTERS 20
AEROSOL DECONTAMINATION FACTOR 23
THREE-INCH THICK FIBER BED 25
SIX-INCH THICK STAINLESS STEEL BED 27
BED LOADING RUNS 28
PRESSURE DROP INCREASE 28
ELECTRICAL PROPERTIES OF BED 30
CONVERSION FACTORS 32
vm
-------
FIGURES
1. Wind Tunnel Schematic 5
2. Schematic Charge Section Without Can 7
3. Fiber Bed Schematic 9
4. Decontamination Factor - 12" Fiber Bed Old System .... 16
5. System Efficiency vs. Flowrate, 6" Polypropylene Bed,
0.96 Void Fraction 19
6. System Efficiency vs. Flowrate, 12" Thick Polypropylene,
Graded Bed Void Fraction 21
7. Fiber Bed Decontamination Factors 24
8. Decontamination Factor for Fiber Bed 26
9. Efficiency vs. Loading - 6" Bed 29
10. Photos of Fiber Bed 31
TABLES
I. Particle Properties of Generated Aerosols 12
II. Emperical Equations of Efficiency 17
III. Emperical Equations of Efficiency 22
IV. Decontamination Factor Emperical Equations 23
IX
-------
INTRODUCTION
The cost/benefit ratio for the control of participate emissions has
increased significantly with the decreasing control limits and for the
more resistant submicron particulates there is a large imbalance which
will probably increase as attention becomes focused on the submicron
species once the "boulders" are under control. Consequently, there is a
great need for control technology that will meet the atmospheric disposal
limits and yet have a marked influence on lowering both the capital invest-
ment and operating expense. It is recognized by all that as the particle
resistivity increases and the size decreases, the efficiency of the elec-
trostatic precipitator (ESP) is significantly degraded. The trend to burn
Western low sulfur coal to reduce the sulfur emission has initiated a large
effort in gas-particle conditioning in an effort to offset the ESP loss of
efficiency due to the higher particle resistivity. Present-day mechanical
devices control the "rocks" reasonably well but are found wanting for the
smaller submicron particulates. Bag filters are the most effective present
technology for filtration of submicron particulates, but have a very sig-
nificant drawback in the limited face velocity and high pressure drop.
Consequently, for high volume discharge gases the filter area required is
extremely large producing high capital and operating costs. The problem
initiated a myriad of research programs aimed at improving or developing
a system or device to meet the emission standards and somewhat secondarily
for more economically controlling emissions. A significant portion of the
work was oriented toward the use of electrical charges or electrical pro-
perties for enhancement of filtration efficiencies. The novel concept
described utilizes the electrostatic properties of particles and fibers
of high dielectric strength to effectively remove submicron and other par-
ticulates from waste or process gas streams.
-------
BACKGROUND
During the course of experiments based on charged water spray-droplet
scrubbing, it was discovered that a dry demister pad of a non-conducting
fiber was acting as an effective filter for submicron particulates of
unipolar charge and produced significantly greater efficiencies than
observed for the charged particle-charged spray drop tests. This obser-
vation, by A. K. Postma and W. K. Winegardner,* initiated a few confirma-
tory tests that reduced the phenomenon to practice and led to an
Environmental Protection Agency (EPA)** grant to explore the boundary
conditions of the variables thought to be of importance. Funding was
later obtained from the Battelle Energy Program office (BEP) for a phase
of the work specifically directed toward fly ash and for continued mathe-
matical modeling of the observed phenomenon which will be issued in a
separate report. At this point in time, a logical equation-of-state has
been developed by D. L. Lessor but an equation providing apriori predic-
tions is not yet completed.
Availability and cost of a new system dictated the use of the
spray-drop system for the initial parametric tests. The variability of
the data and unexplained observations highlighted the inadequacies of the
system for the more refined tests addressing parametric boundary condi-
tions. However, the data were sufficiently encouraging to warrant a new
wind tunnel that would provide a unidirectional, constant velocity flow
between the principal sampling points. This discussion of interim infor-
mation relates, principally, to the experimental data obtained with the
new wind tunnel and addresses only the interim experimental tests assigned
to the EPA grant program.
* Patent pending.
** Work performed under EPA Grant No. R-801581-02-2.
-------
METHODS OF APPROACH
The collection mechanism involves the deposition of charged particles
by self-induced electric fields within a bed of electrically non-conducting
fibers. Since the electrical field in the bed is induced by molecular ion
or charged particle deposition, efficient operation of the system depends
on the development of electrical fields of the same polarity sufficient
to overcome the coulombic repulsion of the fiber within times smaller than
the residence time of the particle within the bed. From this criteria,
the following variables might be expected to play an important role:
Particle Size - influences mobility
Air Velocity - residence time of particles in the
vicinity of a fiber
Pad Resistivity - charge leakage rate and field
strengths and distribution
Pad Thickness - target area residence time
Dust Surface Coverage - flow geometry, charge leakage rate
Particle Resistivity - charge leakage for continuous coating
and particle charge level
Charge Level of Particles - mobility of particles in field
Total Charge Interception
Rate - maximum field in bed due to particle
deposition
Fiber Density - mean free path between particle and
fiber-localized residence time
Humidity - " influence on bed conductivity for
hygroscopic particles
A complete block experimental plan involving the ten parameters listed
would require more than a practical number of tests for proof of the con-
cept. Therefore the program was designed to explore ranges of most of the
variables listed. The boundary conditions of the parameters to be explored
include the following:
Particle Size - not to be considered a prime variable. All tests will
involve submicron particles near the 0.1 to 0.3 ym size range, which
corresponds to minimum mobility and maximum challenge to the system.
The removal efficiency would be expected to be higher for particles
above 1 ym.
-------
Air Velocity - to be varied from 50 to 400 ft/min
superficial bed velocity.
Pad Resistivity - Available demisters of stainless steel, polypropylene
and teflon will be used. The stainless steel will show the effect of
pad conductivity. Teflon fibers would allow operation at temperatures
near 300°F and should be examined for application.
Pad Thickness - Thickness of 3, 6 and 12 inch beds will be tested for
comparative efficiencies.
Particle Resistivity - Highly important, generation of particles having
resistivities of 10^, 10^, 10'' ohm-cm will bevsought -- the highest
resistivity reflecting a severe challenge to electrostatic precipita-
tors.
Particle Charge Level - not considered as a primary parameter for
study at this time. Saturation or near saturation charge appears
advantageous and the charger will be operated at a voltage to produce
same.
Dust Loading - Particle concentrations will be varied from 10 to 100
milligrams per cubic meter.
The data for each experiment included the removal efficiency, pres-
sure drop, relative humidity, air flow rate, and dust loading. The par-
ticle size and particle resistivity was determined for each generating
condition thought to alter these characteristic properties of the particle.
The electrical properties of the bed, e.g. field charge level, would be
dependent primarily on charge deposition and leakage rate through some
finite resistance for the ungrounded bed. Consequently, some indication
of the relative field charge level and distribution within the bed was
obtained during "steady state" conditions by inference from current mea-
surements. These measurements were considered as important input for
support of a mathematical model and possibly desirable for scaleup with a
design engineering model.
-------
en
AEROSOL DISPERSION
GENERATOR PLATE
n
WIND TUNNEL SCHEMATIC
SAMPLING POINTS
EXHAUST
SUPPORTS CHARGE SECTION FIBER BED
FIGURE 1. Hind Tunnel Schematic
-------
EXPERIMENTAL EQUIPMENT
The initial investigation of the observed phenomenon was conducted in
the available charged spray-drop test apparatus which experimentally left
much to be desired for the more definitive parametric study attempting to
define the influencing variables and the related boundary conditions. A
new wind tunnel system was designed to provide a unidirectional uniform
velocity from particle generation point to the last downstream sampling
point.
TEST SYSTEM
Figure 1 is a schematic drawing of the system which consists of a
7 foot long 24 inch diameter pipe, a short corona charging section with an
effective cross-sectional charging area equal to that of the duct, a
3 foot long section between the charger and fiber bed to preclude any
direct field charge effects from the charge section, and a bed frame for
the knitted fiber mats. The inlets and outlets to the bed were 24* inches
diameter. The bed dimension was made 28 inches by 28 inches to extend the
experimental run time before edge leakage became the predominant source of
downstream particles. The frame was made with lucite and was electrically
isolated from the rest of the system as was the charging section. The
system was built in sections for future alteration of configurations and
distances between principal components and was held together with elec-
trically insulated external straps. A dispersion plate was placed in the
12 inch diameter inlet at the transition point to produce the required
uniform velocity profile at the three principal sampling points at posi-
tions 1, 2 and 3.
CORONA CHARGING SECTION
The charging section of the system, shown schematically in Figure 2,
was designed for rather easy alteration of the electrode-to-ground spacing
to vary the particle deposition on the ground rods. Half-inch diameter
rods rather than plates were used for the ground in an effort to reduce
deposition, to minimize the sparkover rate late in the runs and to extend
run time without particle charging degradation. The last electrode wire
* Although EPA policy requires the use of metric units in all its publica-
tions, non-metric units are used in this report for simplicity and clarity.
Readers more familiar with the metric system units are asked to use the
conversion factors on page 32 .
6
-------
TOP VIEW
. 0
o
o
o
•
o
o
o
o
•
0
o
o
o
•
0
o
o
o
•
o
o
o
o
•
o
o
o
o
•
o
o
o
o
I
(
GROUND RODS
20 mil ELECTRODES
DUCT DIAMETER
.WEIGHTS
FIGURE 2. Schematic Charge Section Without Can
-------
was placed downstream of the last rod in an effort to charge any reentrained
uncharged particles arising from the ground rods. Whether or not this was
effective has not yet been determined since repairs to the measuring instru-
mentation could not be made prior to the end of the test series. However,
it is doubtful any small differences could have been determined under the
9 12 3
high particle loads of 10 - 10 part/M due to instrument sensitivity.
The voltage to the electrodes was applied through 30 meg-ohm resistors
which acted as electrical chokes limiting the current to any one electrode
under sparkover conditions or under a direct short to ground preventing
degrading of corona from the other electrodes. This electrical choke
tended to reduce the sparkover late in the run and increase run time. The
effective field depth of about 6 in. gave a particle residence time of
about 100 milliseconds at 300 ft/min face velocity — more than adequate
to reach near saturation charge.
FIBER BED SECTION
The frame holding the fiber bed was made of 1/2 inch thick lucite for
visibility and for its high dielectric properties. The bed surface area
extended 2 inches beyond the duct diameter to eliminate early edge leakage
along the framing where the fiber density is at a minimum as was observed
in experiments using the spray-drop test system. This apparently delayed
leakage until a significant load had accumulated on the front face and
the increase in the pressure drop diverted some particles to the region
of a lower pressure drop. From visual observation this edge leakage, which
was accentuated by the support wires, appeared to lower the efficiencies
for some of the last tests in a run series just prior to bed cleaning. The
observed "frame" leakage is strictly a function of the experimental bed
construction and could be designed out of a bed made specifically for elec-
trostatic filtration. Also, under industrial applications the bed would
normally be cleaned before the pressure drop increased sufficiently to
alter the flow pattern and increase the velocity through a much smaller
bed fraction.
The bed was supported within the frame by a teflon coated wire grid
with about a 4 inch wire spacing (Figure 3). The used demister pads were
disassembled and the individual knitted mats layered into the supporting
8
-------
US-2
FLOW DIRECTION
LUCITE FRAME
FIBER PADS
2> SHIELDED WIRES
3
COPPER SCREENS
2" SQUARE
FIBER MAT
SUPPORT WIRES
TEFLON COATED
(BOTH FACES AND Ml DOLE)
FIGURE 3. Fiber Bed Schematic
-------
frame. Since the maximum width of the mats was about 14 inches, alter-
nate layers were juxtapositioned at 90° angles to each other to preclude
gaps or any major lower density regions that might "short circuit" the
bed. This method of forming the bed also provided a means for varying
the void volume fraction -- that space unoccupied by fibers.
EXPERIMENTAL CONDITIONS
FIBER BED VOID FRACTION
The void fraction of the fiber bed was defined as that volume of the
bed not occupied by the fiber volume and was determined from the bulk
weight of the fiber mats, density of the fiber per unit length and the
fiber diameter.
For this series of tests the 6 inch bed was packed to a void frac-
tion (VF) of MD.96. For the 12 inch bed, the flow direction of the frame
was reversed and the first 6 inches then packed to a VF of -^0.975. The
lower fiber density was used in the first 6 inches to increase particle
penetration depth and increase the load to bed pressure drop ratio. The
lower density in the first 6 inches accomplished both goals, but appeared
to be detrimental with respect to efficiencies in the later stages of a
continuous run series and to the length of operating time before break-
through. This was true particularly for MgO which normally showed a
greater penetration depth "than NH^Cl or Na20. Breakthrough of the bed is
defined as the appearance of agglomerates on the sample downstream from
the fiber bed which were not observed in the previous samples. A con-
ducting copper wire surrounded the fiber bed and small screens (Figure 3)
were placed on the front and back faces to intermittently measure the
current to ground as an indication of the bed electrical conditions during
operation.
10
-------
PARTICLE GENERATION AND CHARACTERISTICS
Three particles were used in this series of experiments to highlight
any change in efficiency with particle resistivity — namely ammonium
chloride, magnesium oxide and sodium oxide. The NhLCl particle was gen-
erated by air sparging of HC1 and NH^OH solutions with interception of the
gas streams at the inlet to the system. The particle generation rate was
controlled by varying both the concentration of the solutions and the
sparge air flow rate. The MgO and Na20 particles were generated by metal
vaporization and oxidation. The metal was heated inductively in an
enclosed boat. An inert gas forced the metal vapors out the end of the
tube where ignition to the oxide occurred. The generation rate was con-
trolled principally by the temperature but also to a small degree by the
inert gas flow. It was found that the aerodynamic mean diameter (AMD) was
rather simply altered by varying the flow rate of the inert carrier gas.
No special attempts were made to obtain a specific particle size or
develop a technique for generating a monodisperse aerosol since this was
not a specific part of the program. Since the technique was available for
varying the particle size, some tests were made with MgO particles having
an AMD of less than 0.1 ym.
An eight-stage Anderson sampler was inserted upstream of the charge
section to obtain the aerodynamic mean diameter for each generation in
which the conditions were altered. It should be recognized that the last
stage of the available sampler had an effective cutoff diameter of about
0.4 ym for the isokinetic flow conditions. Since about 90 percent of the
particle mass was collected on the filter, extrapolation of a log normal
plot to obtain the AMD was a questionable practice although it was the
only approach available at the time. Consequently, the stated AMD's are
thought to be high. The apparent particle sizes and particle properties,
shown in Table I, are the predominant observations for all generations.
The mass median diameter (MMD) was calculated assuming a unit particle
density and is related to the AMD by MMD = ~- where p is the particle
density.
11
-------
TABLE I
PARTICLE PROPERTIES OF GENERATED AEROSOLS
AMD MMD Resistivity
Particle in Microns in Microns in Ohm-Cm
NH4C1 0.25, 0.38 0.16-0.24 <^108
Na~0 0.25, 0.3, 0.4, 0.17-0.56
-------
weight bias. The fiber glass filters created an opposite problem since
some sloughing off or loss of fibers generally resulted when removing the
paper from the holder. This weight loss was reduced from about 0.15 mg
to 0.04 mg by carefully brushing off the edges of the paper before weighing
and inserting into the sampler. However, since these were used only for
the higher flow rates at high air loadings, the fractional influence on
the efficiency was around .002.
TEST PROCEDURE
The aerosol generator was started along with the elapsed generation
timer and brought up to the desired generation rate as quickly as possible,
the conditions for which had been previously, roughly established. Sam-
ples were collected at three points; namely, (1) upstream of the charger,
(2) between the charger and bed and (3) downstream of the bed. The
average volume flow rates and superficial bed velocities for any setting
were based on the pi tot tube profiles of the duct at the flow rates of
interest. Velocity profiles at the three sampling points were not signif-
icantly different. Consequently, the same sampling flow rate was used
for the three mass sampling points. The high particle and field charges
at the second sampling point produced some odd results that defied inter-
pretation and precluded the use of the data for determining the bed effi-
ciency. For example, indicated bed efficiencies were 10 and 92 percent
for two successive runs under identical conditions with both showing about
98 percent total efficiency. Consequently, the results reported are for
the total efficiency of the system. More recent tests have shown a marked
improvement in the comparative bed and system efficiencies — within 1 to
2 percent with some being identical. The only change made to the system
for these tests was an increase in the distance between the electrodes and
ground rods of the charger which reduced the current to one-half of that
observed with the original 'configuration. This further indicated the
problem was associated with sampling in a high electrical field since a
major reduction in the field downstream of the charger ameliorated the
discrepancies markedly.
13
-------
Generally, sampling was continuous until the end of the day or until
apparent breakthrough (whichever came first) except during adjustment of
generation rate and flow. Low aerosol concentration runs generally
required two days of operation. When this occurred, the system was shut
down over night. Prior to particle generation the second day, the charger
was turned on to impose a field charge on the bed and a reentrainment test
made at 300 ft/min flow. There were no apparent degrading
effects even though the bed charge was at zero over night. For a series
of measurements, the flow rate was varied from low to high and back to
low or in the reverse order. Generally, two consecutive measurements
were made before changing the flow and/or concentration.
It should be recognized that the collection efficiencies noted were
based on mass measurements for submicron particles and comparative effi-
ciencies would be expected to be higher for aerosols having a mass median
diameter greater than 1 ym.
RESULTS AND DISCUSSION
With some modification of the spray-drop test system, most of its
disadvantages encountered for submicron particulate studies were eliminated
for the larger resuspended fly ash aerosol. The previous major problem of
resuspended uncharged particles in the reducing transition duct between the
charge section and the large verticle spray chamber was eliminated through
a short coupling duct equal in diameter to that entering the charger. A
few tests with submicron particulates were made following the revision to
give more perspective to the original data and possibly indicate any test
bed bias. These results are presented along with those obtained with the
new wind tunnel system.
SPRAY-DROP SYSTEM TESTS
Bed Configuration
Since earlier work had shown predominent front face loading, the
lower density 6 inch thick teflon demister [(VF) ^0.975] was placed in
front of the 6 inch thick polypropylene bed (VF ^0.96) to form a 12 inch
14
-------
2
thick bed of about 8 ft area. From earlier visual observations the depth
of particle penetration into the bed apparently was directly related to
particle resistivity with the more highly resistive MgO depositing more
uniformly throughout the bed depth than the NH^Cl which was deposited
predominantly in the first few inches of the lower void-fraction bed with
NagO about intermediate between MgO and NH.C1. This somewhat graded fiber
density altered the loading characteristics by producing a more uniform
deposition throughout the first 8 inches of the bed without any apparent
detrimental effects on efficiency. This extrapolates to a higher bed
load for an equal pressure drop across the bed for the lower resistivity
particles. For example, under similar NH,C1 loadings, less than maximum,
the 12 inch polypropylene bed of uniform fiber density showed an increase
in pressure drop across the bed at 150 ft/min face velocity from 0.15 to
1.2 inches of f^O while the teflon-polypropylene combination changed from
^ 0.10 to 0.23 inches HgO. In subsequent tests with the new system in
which the loading was taken to breakthrough, an earlier breakthrough was
observed for the MgO particles using the graded density configuration.
This suggests that the use of a graded density bed might reduce the fre-
quency of cleaning without any sacrifice in the efficiency if matched with
the particle resistivity. Particle size may also exhibit a similar effect
but was not observed over the particle size range tested - <1 ym.
System Efficiency for NH^Cl, NapO and MgO
The system efficiency and aerosol decontamination factor for the
three types of particles are shown in Figure 4. Comparison of this data
with that previously generated showed that some of the inherent problems
of the system were corrected by the revisions, especially at the higher
flow rates where the efficiency was increased from about 75 to 90 percent
at 310 ft/min flow. Also the linear relationship between
velocity and efficiency was observed for the first time over the flow
range of 50-300 ft/min. Some leakage along the bed frame was still
observed particularly for the higher flow rates late in a run and probably
accounts for some of the scatter in the data as well as a lower indicated
efficiency.
15
-------
1000
100
o;
o
o
o
UJ
O
10
BED
1.0
Es= SYSTEM EFFICIENCY
O NfyCI- 0.26pm AMD
• MgO-0.15pm AMD
a Na20-0.25pm AMD
AIR CONCENTRATION~100 mg/m3
IxlO5) V"1-63
= 0.93
99.5
99
98 —
95 ~
90
80 H
1
1
o
CM
ft
1.0 5
Q.
O
CXL
O
00
LU
on
Q_
O
UJ
CQ
a
o
0.01
30 100 1000
SUPERFICIAL BED VELOCITY, FT/M1N
DECONTAMINATION FACTOR -12" FIBER BED
OLD SYSTEM
FIGURE 4.
16
-------
As would be expected, the log of the decontamination factor varied
linearly with a power of the velocity. Since there was no apparent dis-
tinguishing trend with respect to particle resistivity or size, the data
were considered as from a single population for the least squares fit,
shown by the curve, which showed relatively good agreement with a corre-
lation coefficient of 0.93. The pressure drop shown in Figure 4 is for a
clean bed. For these tests the pressure drop increase at the end of each
series of measurements averaged about 20 percent.
The equations for the least squares fit of the data for each particle
are shown in Table II along with the coefficient of determination.
TABLE II
Emperical Equations of Efficiency
12" Thick Bed - Spray-Drop Test System
.2
Particle Equation
NH4C1 Es = 101. 05-0. 041V 0.996
Na20 ES = 100. 96-0. 02V 0.85
MgO ES = 101. 68-0. 035V 0.96
ES = System Efficiency
V = Bed Superficial Velocity - ft/min.
The poor correlation coefficient for Na20 reflects the low efficiency
obtained at 310 ft/min face velocity ascribed to the high leakage along
the frame.
17
-------
NEW WIND TUNNEL TESTS
Polypropylene Bed - Six Inch Thick
The void fraction for this bed was approximately 0.96. Figure 5
shows the average efficiency as a function of velocity as well as the
sample 2a limits. For both, the efficiency up to 200 ft/min filter face
velocity is approximately linear and then appears to change to some other
function or to a linear function with a different slope. Sufficient data
are not available to describe the point of change and/or any change in
the relationship between 200-300 ft/min. The single NH,C1 measurement at
250 ft/min was the last measurement taken before bed cleaning and is
thought to be low due to observed breakthrough which, from visual observa-
tion, was assigned to migration of the particles along th-e lucite frame.
This short circuiting was observed only after the pressure drop increase
in the original flow regime of the bed diverted some of the flow to the
outer edges. The supporting wire grid work appeared to accentuate the
diversion of particles to the bed fringes as evidenced by the predominant
earlier particle migration along the wires. This leakage along the bed
frame would be eliminated with a fiber bed designed for this specific
application. The equations developed from a linear regression fit of the
data, 200 ft/min velocity and below, are shown in Table III and represented
by the solid line in Figure 5.
The total efficiency measured for Na^O was significantly lower than
that for NH^Cl which has a lower resistivity than Na«0. Since this was
contradictory to the developed theory which predicts a decreasing effi-
ciency with decreasing particle resistivity due to a lower space charge,
an investigation was initiated to shed some light on the observation.
During examination of all operating data it was noted that the current
flow in the charger during this series of measurements increased with
time and was uncharacteristically very high during the last few measure-
ments at 200 and 300 ft/min. Examination of the charging section
showed that a buildup of Na^O on the terminal weights of the
18
-------
100
90
E »
Q_
O.
¥ 100
X
AMM ONI UA/1 CHLORIDE
0.37M AMD
RES ISTIVITY~2xl08 ohm-cm
1 1 L
SINGLE * \
MEAS
L
>•
CO
90
80
70
SAMPLES
SODIUM OXIDE
0.3p AMD
RES ISTIVITY~8xlo9 ohm-cm
I I I
100 200 300
BED FACE VELOCITY, FT/MIN
SYSTEM EFFICIENCY VS FLOW/RATE
6" POLYPROPYLENE BED
0.96 VOID FRACTION
FIGURE 5.
-------
electrodes, which were nearly touching the ground rods, created a more
conductive path to ground through absorption of moisture and subsequent
formation of the more conducting NaOH. Thus this effectively reduced the
corona field and apparently the particle charging efficiency resulting in
an anomalously low system efficiency at least near the end of the series.
Another indicator of this problem was the necessity to reduce the power
supply voltage gradually with time to prevent automatic cutoff by the
power supply current limit switch. The moral of this sequel is to never
use NapO test aerosols at high air humidity (> 40 percent) when the elec-
trode weights are in close proximity to an electrical ground that produces
a direct short -- principally a geometric design problem.
Twelve-Inch Thick Polypropylene Filters
The first six inches of this bed were packed to a void fraction of
0.975 with the last six inches at * 0.96 principally to extend the run
time for NFLC1 aerosols of the higher concentrations. Since there was no
discernable effect of particle concentration on system efficiency over the
range of 6-148 mg/M , only higher concentrations were used in this series
of tests in an effort to reduce the run time and maintain reasonable repro-
ducibility. This was accomplished to some degree but again sensitivity
of the measurements produced a higher variability than desired. However,
the frequency of unexplained "fliers" was significantly lower.
The efficiency for NH^Cl, MgO and Na20 is plotted against flow rate
in Figure 6. As shown by the least squares fit curves, the efficiency
decreased linearly with increasing velocity at least up to 300 ft/min
flow rate. The Na«0 data points are the average of all
measurements at a single velocity. The 95 percent confidence limits are
also shown. The MgO and NH,C1 data are duplicate measurements. Conse-
quently, all data are plotted and no limits shown. The low point at
100 ft/min for MgO appears to be one of those "fliers" that occur now and
then without any apparent explainable reason. However, with a statisti-
cally significant number of tests, these generally fall out as "outliers".
Consequently, it was not used in the linear regression analysis. Although
20
-------
100
90
80
AMMONIUM CHLORIDE
0.3|j AMD
RESISTIVITY~2xl08 ohm-cm
100
cr
LU
Q_
6 90
z:
O
on
>-
1/1
80 -
100
90
80
SODIUM OXIDE
0.6u AMD
RESIST!VITY~8xl09 ohm-cm
MAGNESIUM OXIDE
0.0/M AMD
RESISTIVITY -1012 ohm-cm
100
200
FLOW/RATE, FT/MIN
300
400
SYSTEM EFFICIENCY VS FLOWRATE
12" THICK POLYPROPYLENE
GRADED BED VOID FRACTION
FIGURE 6.
21
-------
this value was not used, the data indicate an efficiency less than 100 per-
cent at zero flow -- not very probable. This suggested that these last
two measurements at 100 ft/min (^ 50 cms/sec) in the run series flow were
probably on the low side due to edge leakage.
The average efficiency for NH^Cl was lower than that of MgO or Na20
by about 2 to 5 percent over the range of 100-300 ft/min
face velocity. This is the first data that might confirm the model's pre-
diction that efficiency should decrease with decreasing resistivity. Since
the bulk of the data does not indicate a like trend, final judgment is
reserved until a statistically significant number of measurements are com-
pleted. Also bearing on the reservation is the lack of any discernable
difference in the bed field charge, inferred from current measurements,
between the three aerosols used. Thus, it is not unlikely that signifi-
cantly lower efficiencies would be observed only for a much more conducting
particle and the threshold for this effect would occur at a lower resis-
o
tivity than the 10 ohm-cm of NH4C1. The linear equations for the lines,
Figure 6, are shown in Table III.
TABLE III
Emperical Equations of Efficiency
Particle 6" Bed - < 200 ft/min 12" Bed - < 300 ft/min
Na20 E = 101.5 - 0.065V E = 100.91 - 0.0116V
Mgo E = 99.8 - 0.0116V
NH4C1 E = 100.513 - 0.028V E = 101.286 - 0.0283V
Where E = Efficiency in Percent and,
V = Superficial Bed Velocity - ft/min.
22
-------
Aerosol Decontamination Factor
The calculated decontamination factors for both the six inch and
twelve inch beds are plotted as a function of velocity in Figure 7. The
data approximate the expected relationship with the log of the decontami-
nation factory varying linearly with the power of the velocity. For the
six inch bed, all data were considered to be from the same population for
the least squares fit which is represented by the solid line. This
appears to be a reasonable assumption since the coefficient of determina-
tion was 0.9 and all data fell within the 2a limits of sampling and
measurement errors. Consequently, the fitted equation probably repre-
sents a reasonable average for the experimental conditions.
For the twelve inch bed, the log-log plot of the data suggests a
real, although small, difference in the bed efficiency for NH.C1 particles
compared to that for Na^O and MgO and were segregated for calculation of
the linear regression fit of the data while the Na«0 and MgO data were
lumped together. The apparently anomalously low values for NH.C1 and MgO
(100 ft/min - 50 cms/sec) were the last samples taken in each series when
the operating log noted leakage along the bed edges. Since it is unreal-
istic, based on the preponderance of data, that the efficiency should
decrease with decreasing velocity, the bed leakage was assigned as the
major cause for the low values which were not included in the curve fit-
ting process. The pressure drop curves shown in Figure 7 are for the pre-
viously described void fractions in a clean state. The equations for the
fitted curves are contained in Table IV.
TABLE IV
Decontamination Factor Empirical Equations
(New Wind Tunnel)
Fiber Bed
Six Inch
V.F. - 0.956
Twelve Inch
V.F. - 6" at 0.975
and 6" at 0.956
Type
NH4C1
NH4C1
MgO
Na20
Parti cal
Size
* 0.25
* 0.06
•u 0.25
D, = (9.5xl04)V"1<72
' ?
r^ = 0.9
Df M2.12xl07)V-2.52
r*
23
0.99
8xl0
0.97
= (7.8xl06)V'2'21
-------
1000
500
'- 100
.
o
o
o
o
LU
O
10
s* EFFICIENCY
o NH4CI
A Na20
• MgO
BED . POLYPROPYLENE
_ AP - 12" BED
Na,0 >12" BED - POLYPROPYLENE
I
99.8 -
99.5 -
99 -
98 -
LU
95 -
90 •
80 -
50 -
50 100 1000
BED SUPERFICIAL VELOCITY, ft/min
FIBER BED DECONTAMINATION FACTORS
FIGURE 7.
24
1.0
0.5 o
on
0
t/1
LU
o:
o_
0.1
o
-------
Three-Inch Thick Fiber Bed
Since the efficiency appeared to be dependent upon fiber density as
well as the demonstrated dependence on thickness, a run series was made
with the NH^Cl aerosol using a three inch thick bed packed to a void frac-
tion of 0.946. In this test series it was planned to obtain successive
measurements at one velocity until breakthrough at the bed edges was
apparent. In an effort to increase the number of consecutive measurements
at the high aerosol concentrations used, a folded fiber mat was inserted
at the bed-frame interface to bring the fiber density at that point nearer
to that in the rest of the bed. This produced a marked improvement in
total run time before edge breakthrough was indicated at a much higher
pressure drop than in previous tests. Under normal industrial applica-
tions ', cleaning of the bed would precede the more than an order of magni-
tude increase in the bed AP observed in the lab tests. Five successive
measurements were made for all flowrates except 100 ft/min where 20 meas-
urements were made. The particle size for this series of tests hovered
o
around 0.25 ym AMD at a mass concentration near 75 mg/M .
The equation developed from the least squares fit of the efficiency
versus velocity was
Es = 102.8 - 0.0514 V
where
ES = system efficiency in percent and,
V = bed face velocity in ft/min and
r2 = 0.986.
The plot of the decontamination factor vs. velocity (Figure 8) shows the
average values for each velocity along with the 95 percent confidence
levels, The equation developed from the least squares fit was
Df ^ (3.87 x 107) V"2'74
where
Df = i c , dimensionless and
25
-------
X
Dp = 1428
AT 50ft/min
ES = 99.95
Dp = 2000
(3.87xl07)V~2'74
oc
o
CJ
UJ
o
/
Es = SYSTEM EFFICIENCY
FORNH4CI -0.25 Mm AMD .
f
AP - FIBER BED
i I i
99.9
99.8
99.5
99 -|
98 -
95
90
1.0
0.5
a.
o
CO
00
LU
0.1 £
a
LU
CO
100 1000
BED SUPERFICIAL VELOCITY, ft/min
DECONTAMINATION FACTOR FOR FIBER BED
3 INCH POLYPROPYLENE - 0.946 VOID FRACTION
0.05
0.01
FIGURE 8.
26
-------
V = bed face velocity in ft/min and,
r2 = 0.94
The slope of the regression line was highly influenced by the D* of
2000 at 50 ft/min face velocity. The relatively poor fit suggests either
that the sensitivity of the measurements was insufficient to accurately
determine the efficiency at the lowest velocity giving a high value or
that the log of the decontamination factor is not linearly related to a
power of the velocity at lower flowrates. Thus similar tests at several
velocities between 25 and 100 ft/min will be required to describe the
relationship. However, some measurement method other than mass will be
required to provide adequate sensitivity at the sampling point downstream
of the bed at these lower velocities.
Comparatively, this thinner 3 in. bed with a lower void fraction
exhibited an efficiency about equal to the 6 and 12 in. beds at 100
ft/min flow rate and was about 5 percent higher and lower than the 6 and
12 in. beds at 300 ft/min flow rate. At 50 ft/min, measurements indi-
cated a 100 percent efficiency for the first six hours of continuous
operation. The initial pressure drop across this bed was slightly higher
and lower than the 6 and 12 in. beds.
This test illustrates that for low flow rates and low aerosol load
the thinner bed would have a practical application. The dependency of
efficiency on both the fiber bed thickness and density was demonstrated
but not in sufficient detail to develop a mathematical relationship.
Six-Inch Thick Stainless Steel Bed
The series of experiments preceding the discovery of this concept
suggested that an electrically non-conducting fiber bed was required to
obtain the observed high collection efficiencies. To support this idea,
a conducting bed of stainless steel was tested under equivalent conditions.
As suspected, the efficiencies obtained were very low, between 35 and 43
percent at 350 and 50 ft/min face velocity, compared to the 18 and 35
percent calculated for image force collection under the test conditions.
The higher conductivity of this bed results in charge leakage at a much
lower charge level on the bed and illustrates the importance of the level
of field charge developed in "non-conducting" the bed.
27
-------
Bed Loading Runs^
Data collected during two NH^Cl loading runs at 100 ft/min face
velocity are shown in Figure 9A for a clean bed at the start of the run
and Figure 9B which started with a partially loaded bed from the previous
3
days' tests. Obviously, the scatter in the data was greater at 50 mg/M
aerosol concentration than at twice that concentration. As discussed
previously, this scatter in the data was ascribed to the sensitivity of
the mass measurement for the downstream sampling position which was hover-
ing around the balance sensitivity at these high efficiencies and aerosol
concentrations. Although the averages of the efficiencies appear to be
different, it is thought to be principally an artifact of the sensitivity
rather than a real difference since differences in the efficiency due to
aerosol concentration over the range tested were not discernable. There
appears to be a slight trend for the efficiency to increase with bed
loading, Figure 9B, which was also observed with other loading runs made
at a constant velocity.
Pressure Drop Increase
The increase in pressure drop .with bed loading appears to be pre-
dominantly a function of both the bed void fraction and particle charac-
teristics which apparently influence the predominant plane of deposition
perpendicular to the flow at least for the submicron size range. The
aerosols of 2 ym AMD size and above may also influence depth of penetra-
tion into the bed. The velocity of air flow might also be a factor in
directing the depth of penetration but cursory visual observations indi-
cate that it would be of a second order magnitude. This later impression
was obtained from the NH-C1 aerosols over a flow range of 50-300 ft/min.
For these tests visual inspection did not reveal any dramatic change in
the plane of principal deposition which would have been required to be
discernable. This is also supported to a degree by the much faster pres-
sure drop increase for the higher flow rates under equal aerosol concen-
trations. Specific tests to define the effect of velocity on the pressure
drop-bed loading relationship have not been made. Relative comparisons,
for a bed void fraction of 0.96, 100 ft/min face velocity and an approxi-
mately equal pressure drop increase from 0.08 to 0.3 in. of HgO, gave bed
28
-------
100
95
90
oo
AV = 97.3%
o
o
AEROSOL CONC -50 mq/m3 42 g/ft2 LOADING
>-
o
o
1 2
3 4
5 6 78 9 10 11 12 f
SAMPLE NUMBER
CO
100
95
90
A. CLEAN BED AT START
100 FT/MIN VELOCITY
NH4CI
AMD-0.37
o 0 u
o
o
6g/ft2l_OADING
1 1
° AV = 98.6%
AEROSOL CONC- 100 mg/m3
36g/ft2LOADING
1 1 1 1 1
4 5
SAMPLE NUMBER
B. PARTIALLY LOADED BED AT START
100 FT/MIN VELOCITY
EFFICIENCY VS LOADING - 6" BED
0.96 VOID FRACTION
FIGURE 9.
-------
loadings of about 540, 650 and 970 g/M2 for approximately 108, 1010 and 1012
ohm-cm particle resistivities having nearly equal AMD's.
For non-characteristic aerosols, like the ones used, any tests defin-
ing the load - AP - VF relationship would be of academic interest only.
The information obtained to date suggests that for industrial applica-
tions, particle characteristics, principally resistivity and concentration
and secondarily size as well as the control system operating conditions
are necessary to optimize the efficiency, cleaning frequency and area of
the fiber bed.
Pictures of a loaded bed, Figure 10, show the upstream and downstream
faces of a fiber bed loaded with NH4C1 particles. The pictures are some-
what deceiving since the front face gives the appearance of being plugged.
However, the pressure drop across this 12 in. thick bed under this particle
2
load of ^80 g/ft was about 0.9 in. of water at 300 ft/min face velocity.
Although not readily visible, there was a very slight deposit on the rear
face fibers and support wires.
Electrical Properties of Bed
The electrical properties of the bed could only be inferred from the
leakage current of the bed and from the screens positioned as shown in
Figure 3. The bed leakage current was obtained from a bare wire encompas-
sing the bed at the frame-bed interface. A single wire was found to pro-
duce the same leakage current as a full width aluminum foil. From the
measured leakage current of the bed and the four screens, the space
charge of the bed built instantaneously upon activation of the charger to
about a factor of ten less than that observed during particle generation
and increased proportionally with flow rate. This suggests that molecular
ion or charged Aitken nuclei deposition produces a significant contribu-
tion to the bed field charge and deposition of charged particles is not
required to initiate the collective action. The charge varied with mass
per unit time depositing on the filter but only by about a factor of five
o
over the range tested (10-150 mg/M ). At any one particle deposition rate
the space charge increased immediately by an order of magnitude and then
remained constant indicating that charge leakage from the bed was equal to
that deposited with the particles.
30
-------
LOADED POLYPROPYLENE FIBER BED
12 INCH THICK
VOID FRACTION-0.96
NH4CI PARTICl£S-0.25pm AMD
LOAD ABOUT 150 g/ft2
UPSTREAM FACE
.
•Si;
•
;
• -
DOWNSTREAM FACE
31
-------
By assuming that the charge distribution in the bed is directly pro-
portional to current measurement, some inferences relative to its shape
can be made from the current measurements of the planted screens. For the
6 in. bed, there is an order of magnitude decrease between the front and
rear face which prevailed throughout the run and was present with or
without particle deposition. This suggests essentially an instantaneous
distribution of the charge deposited by the particles. The field charge
distribution transverse to the flow appeared to be uniform within the
rather approximate measurements resulting from the fluctuating meter.
The above observations prevailed also for the 12 in. thick bed except
the average reduction in the inferred field charge front to rear was
about 20 and was more variable than for the 6 in. bed.
CONVERSION FACTORS
j£ Convert T£ Multiply By
feet/minute centimeters/second 5.08 x 10""1
inches centimeters 2.54
inches of H20 atmospheres 3.34 x 10"2
32
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-76-132
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Electrostatic Capture of Fine Particles in
Fiber Beds
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHORIS)
D. L. Reid and L. M. Browne
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING OROANIZATION NAME AND ADDRESS
Battell e Northwest
P.O. Box 999
Richland, Washington 99352
10. PROGRAM ELEMENT NO.
1AB012; ROAP 21ADL-035
11. CONTRACT/GRANT NO.
R801581-02-2
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND
Final; 6/73-6/70
PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA-ORD
is. SUPPLEMENTARY NOTES Project off i cer f or this report is D. C. Drehmel, Mail Drop 61,
Ext 2925.
16. ABSTRACT
repor£ gives results of a study of the collection of charged submicron
sized particles by fiber beds. Removal efficiency was determined as a function of
particle resistivity, fiber bed resistivity, superficial gas velocity, and total concen-
tration of the aerosol. Using a 12-in. thick polypropylene fiber bed, greater than 90%
removal efficiency was reported for superficial bed velocities up to 300 ft/min. For
a 6-in. bed,* 90% removal was possible up to 200 ft/min. In both cases, the pressure
drops for these systems were low (less than 1 in. H2O). A third bed was tested, com-
prised of a 6-in. Teflon fiber bed followed immediately by a 6-in. polypropylene bed.
This composite 12-in. bed did not perform as well as the homogeneous 12-in. bed.
In all tests, the most important parameter was bed face velocity. In general, the
collection efficiency of the fiber bed decreased linearly with increasing velocity.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
cos AT i Field/Group
Air Pollution
Electrostatics
Dust
Filters
Fibers
Resistance
Aerosols
Air Pollution Control
Stationary Sources
Fine Particulate
Fiber Filters
Gas Cleaning
Collection Efficiency
13B
20C
11G
14B
11E
07D
Z. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
42
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (3-73)
33
------- |