&EPA
United States Industrial Environmental Res
Environmental Protection Laboratory
Agency Research Tnanqle Park NC 2771 1
EPA 600 7 78-095
June 1978
Studies of Dust
Cake Formation and
Structure in Fabric
Filtration
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. Approval does not signify that the contents necessarily reflect
the views and policies of the Government, 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-600/7-78-095
June 1978
Studies of Dust Cake Formation
and Structure in Fabric Filtration
by
Bernard Miller, George Lamb. Peter Costanza,
Dan O'Meara, and Janet Dunbar
Textile Research Institute
P.O. Box 625
Princeton, New Jersey 08540
Grant No. R804926
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
Earlier studies have shown that differences in filtration per-
formance produced by certain modifications of nonwoven filter fabrics
are largely due to the development of different dust cake structures.
The chief objectives of the current program are to identify those cake
characteristics affecting performance and in turn to relate the pro-
duction of desirable cake properties to fabric structure and filtration
conditions.
One experimental approach has involved microscopical examination
of flyash deposits on nonwoven fabrics. In a more quantitative approach,
the amount of dust capture as a function of depth into the filter was
measured using nonwovens formed in layers. When filters made of fibers
with trilobal cross sections are compared with those made of round
fibers, the largest advantages in capture efficiency due to trilobal
fibers occur in the upstream layers where the largest amounts of dust
accumulate. Preliminary studies with composite layered filters also
show that filtration performance is dominated by the upstream layer.
For example, a filter of mostly round fibers with only a narrow up-
stream layer of trilobal fibers has the higher efficiency and lower
pressure drop of an all-trilobal filter.
Dust cake structure is influenced not only by fiber geometry but
also by the charge on the particles and by any electric field on the
filter. Neutralizing the naturally charged flyash particles results in
a more compact, less porous cake and in increased pressure drop. When
the particles are precharged, they appear to collect closer to the sur-
face and not to collapse into pores as readily, resulting in increased
efficiency and lower pressure drop. An electric field set up across
the filter perpendicular to the air flow can also produce higher effi-
ciency and reduced pressure drop. When incoming particles are charged,
either naturally or by precharging, more dust is deposited near the
electrodes of opposite charge, creating regions of differing resistance
to air flow, which probably account for some of the observed reduction
in pressure drop.
Theoretical calculations of single-fiber collection efficiencies
support the experimental findings that capture of both charged and
uncharged particles on fibers of any cross-sectional shape is more
likely if an electric field is imposed perpendicular rather than paral-
lel to the flow. In regard to fiber cross-sectional shape, capture
should improve with increasing number of lobes and with the depth of
the lobes.
This report is submitted in fulfillment of Grant No. R804926-1
by Textile Research Institute under the sponsorship of the U. S.
Environmental Protection Agency. This report covers a period from
December 20, 1976 to December 19, 1977, and work was completed in
December 1977.
-------
CONTENTS
Page
Abstract ii
Figures iv
Tables vi
Acknowledgments vii
I. INTRODUCTION 1
II. EFFECT OF FIBER GEOMETRY ON DUST CAKE FORMATION. ... 2
A. TRILOBAL AND ROUND CROSS-SECTIONAL SHAPE 2
1. Introduction 2
2. Photographic Studies . . . , 2
3. Measurements of Cake Density as a Function
of Depth into Fabric 5
4. Composite Filters 9
5. Observations on Fiber Arrays 10
B. ROUGH AND SMOOTH FIBER SURFACES 11
C. EFFECT OF PARTICLE CHARGE 13
1. Neutralized Aerosols 13
2. Effect of Particle Precharging 16
III. STUDIES OF CAPTURE IN ELECTRIC FIELDS 19
A. THEORETICAL STUDIES 19
1. Effect of Field Orientation 19
2. Fibers With Lobed Cross Sections 22
B. EFFECT OF ELECTRIC FIELD ON CAKE STRUCTURE .... 31
1. Introduction 31
2. Experimental Results 31
3. Conclusions 36
IV. ' REFERENCES » . . <, 38
iii
-------
FIGURES
Number
1 Technique for formation of dust cake directly on
SEM specimen 3
2 Loss of particles apparently due to bombardment by
electron beam 4
3 Scanning electron micrographs comparing dust deposits
formed on round and trilobal fiber filters at a dust loading
of 2 mg/cm2. 500X 4
4 Filtration performance of layered bonded nonwovens made
of round and trilobal fibers. Face velocity 12 fpm 6
5 Filtration performance of another pair of layered non-
wovens made of round and trilobal fibers 7
6 Filtration performance of a third pair of layered non-
wovens made of round and trilobal fibers. Note: These
fabrics are almost twice as dense as those of Figures
4 and 5 8
7 Scanning electron micrographs comparinf dust cakes
formed on trilobal and round fiber filters (50X), and
smooth fiber filters (100X) by an untreated flyash aerosol 12
8 Scanning electron micrographs comparing dust cakes
formed on trilobal and round fiber filters (200X), and
rough and smooth fiber filters (100X) by a neutralized
flyash aerosol 14
9 Pressure drop developed by increasing dust cake
deposits of untreated and neutralized flyash aerosols on
trilobal and round fiber filters 15
10 Efficiency of trilobal and round fiber filters with incr-
easing deposits of untreated and neutralized flyash 15
11 Pressure drop developed by increasing dust cake deposits
of untreated and neutralized flyash aerosols on rough and
smooth fiber filters 17
12 Efficiency of rough and smooth surface fiber filters with
increasing deposits of untreated and neutralized flyash 17
13 The two limiting particle trajectories for a round fiber
of radius rc 20
14 Enhancement e of collection efficiency by a field perpen-
dicular to the flow for charged particles as a function
of |G|=qE0B/V0 20
iv
-------
Figures (con.)
Number Page
15 Fiber cross sections for selected values of m (number
of fiber lobes), c (orientation of fiber to air flow), and
size of lobes 22
16 Particle trajectories near a bilobal fiber. Field parallel
to air flow, a = 0. 5, and F = 1.0 23
17 Particulate trajectories near a trilobal fiber. Field
parallel to air flow, a = 0. 5, and F = 1.0 23
18 Comparison of collection efficiencies for bilobal and
trilobal fibers at different orientations of fiber to
electric field 24
19 Comparison of the effective diameters of multilobal
fibers (for charged particles) 28
20 Effective fiber diameter vs. number of lobes (for
charged particles) 28
21 Effect of fiber orientation on the effective diameter of a
bilobal fiber (for charged particles) 29
22 Diagram of tube for precharging flyash aerosol (negative
precharging arrangement) 30
23 Positively precharged particles deposited on filter in
absence of applied field across filter 33
24 Neutralized particles deposited on filter in absence of
applied field across filter 33
25 Naturally charged particles deposited on filter in absence
of applied fielu across filter 33
26 Naturally charged particles deposited on filter with 2 kV/cm
across filter 34
27 Positively precharged particles deposited on filter with
2 kV/cm across filter 34
28 Negatively precharged particles deposited on filter with
2 kV/cm across filter 35
29 Neutralized particles deposited on filter with 2 kV/cm
across filter 35
30 Two sets of two electrodes with 2-kV/cm potential gradient
using naturally charged particles 35
31 Electrical analogy of unequal parallel resistances 37
-------
TABLES
Number Page
1 Performance of Composite Fabrics in Filtering a Fly-
ash Aerosol of 7 g/m3 After 24 Cycles of Conditioning 10
2 Cake Resistance and Efficiency of Round and Trilobal
Fiber Filters when Filtering Aerosols Charged to
Various Levels 18
3 Comparison of Collection Efficiencies for Uncharged
Particles Acted on by Electric Fields Parallel and
Perpendicular to the Flow Direction 21
4 Collection Efficiencies for Uncharged Particles on 1-,
2-, and 3-lobe Fibers in Electric Fields 24
vi
-------
ACKNOWLE DGKENTS
The authors would like to express their appreciation to
Dr. James H. Turner of the Environmental Protection Agency for his
advice and encouragement throughout the period of this research.
They would also like to thank Mr. Harold W. Lambert and Mr. Harry
Buvel of TRI for their invaluable work on apparatus, Mr. John P. Hession
for his assistance with microscopy, and Dr. Harriet G. Heilweil for her
help with reports.
-------
SECTION I
INTRODUCTION
The main object of this work is to find explanations for observed
changes in filtration performance that result from certain modifica-
tions of conventional fabric filtration technology. One such modifi-
cation is the use of fibers having different geometric properties
(cross-sectional shape, surface roughness, crimp, etc.) from those
fibers generally used in filter fabrics. Another is the application
of electric fields to the fabrics. A third, the use of creped fabrics.
In each case, some of the improvements in performance, which have
taken the form of increased capture efficiency, reduced pressure drop,
or both, appear to be coincident with changes in dust cake character-
istics. Therefore, one aspect of the present work is the identification
of certain features of the dust cake or of the fabric-dust cake system,
which may explain the observed changes in performance. A second goal
is to use this information as a basis for the design of improved filter
fabrics.
-------
SECTION II
EFFECT OF FIBER GEOMETRY ON DUST CAKE FORMATION
A. TRILOBAL AND ROUND CROSS-SECTIONAL SHAPE
1. Introduction
A previous study [1] of the differences in capture efficiency
between nonwoven filters made of fibers of round as opposed to tri-
lobal cross section attempted to explain the results as a consequence
of dissimilar interfiber distances. In two fabrics having the same
packing density, if the fibers have equal linear density but different
cross-sectional shapes, the average center-to-center interfiber dis-
tance will be the same, but the effective spacing between fibers will
be different and slightly smaller for trilobal than for round fibers.
It may be argued that the smaller interfiber gaps in trilobal fiber
fabrics would give rise to more readily formed and more stable dust
bridges and in this way to higher collection efficiencies. The de-
pendence of collection efficiency on round-fiber diameter at constant
packing density has been measured, however, and from this the depen-
dence on the interfiber gap. The change in diameter of a round fiber
that would produce the same reduction in interfiber gap as a trilobal
fiber of the same linear density was calculated. The increase in
efficiency corresponding to this smaller interfiber gap with round
fibers is only half of the increase in efficiency obtained experimentally
from the same decrease in interfiber gap due to change from round to
trilobal shape. It is likely, therefore, that trilobal fibers enhance
capture, not only because of reduced interfiber gap, but also by some
other mechanism. It is notable also that, in general, trilobal fibers
do not cause increased pressure drop.
2. Photographic Studies
An initial effort of the present program consisted of a micro-
scopical study of dust formations on nonwoven filters, using at first
a scanning electron microscope (SEM) and later, a binocular light
microscope. The study was begun with the development of a technique
for preparing a filter sample suitable for SEM examination. A 1-in.
diameter disc of filter fabric was glued to an annular aluminum stub
of 1-in. o.d. and 0.75-in. i.d. This stub could be inserted directly
into the SEM sample holder. The stub was then glued to the center
of a patch filter specimen of the same fabric, in which a 0.75-in. hole
had been cut. The arrangement is shown in Figure 1. In this way dust
could be deposited on the test specimen using the TRI patch filter
apparatus with only minor disturbance of the flow pattern. The stub
was then pulled off the patch filter for SEM examination.
-------
Filter Fabric
SEM Stub
Dust Cake -^
f t t t I t
AEROSOL
Figure 1. Technique for formation of dust cake directly on
SEM specimen.
The microscopical study was aimed at identifying features of
the dust deposit which would explain the better performance of tri-
lobal fibers over round. Two levels of dust loading were used:
1) ^2 mg/cm2. a low cake mass to observe incipient bridging, and
2) »\/15 mg/cm , a normal cake mass encountered in practice.
With a light dust loading, problems were caused by the inter-
action of the SEM and the incipient dust cake. Consecutive photo-
graphs taken of the same area of filter with dust deposited on it
showed fewer particles on later photographs. It appeared that the
electron beam bombardment knocked particles off the sample. This
illustrates the fragility of the structures to be studied and also
places in doubt the suitability of scanning electron microscopy as
a method for this type of study. Figure 2 shows an example of such
a sequence of photographs. It can be seen that the lettered parti-
cles are missing in Figure 2b which was taken a short time after
Figure 2a. In Figure 3 are examples of views of light dust deposits
on round (a) and trilobal (b) fiber filter substrates which illustrate
the difficulty of identifying any features that could be correlated
with differences in performance.
When a normal cake mass was deposited on the filters, lower
magnification (SOX) provided some indication that trilobal fibers
form a filter substrate for a more uniform cake with fewer voids and
asperities (Fig. 7a) . This greater uniformity can apparently be
associated with higher capture efficiency and with greater ease of
cleaning resulting in. lower pressure drop. However, since the inter-
face between the dust cake and the substrate could not be seen, an
explanation for the greater uniformity cannot be offered. A top view
under the SEM is not acceptable, and some means of viewing the cake/
fabric interface must be developed before the effect of fiber cross
section can be optimized.
In the proposal submitted originally, several methods for micro-
scopical examination of dust cakes were listed, each of which was
meant to be tried in turn. The methods were themselves not well
established but were implicitly expected to require a certain amount
-------
Figure 2. Loss of particles apparently due to bombardment by
electron beam. Particles A - E are missing from photograph (b)
taken a short time after (a).
Figure 3. Scanning electron micrographs comparing dust deposits
formed on (a) round and (b) trilobal fiber filters at a dust
loading of 2 mg/cm2. 500X.
-------
of development before they proved suitable for the present purpose.
They included various techniques for embedding and sectioning the
fabric/cake combination, as well as other techniques for exposing a
sectional view of the dust cake by removing adjacent areas of the cake
using suction techniques. However, the inconclusive outcome of the
first stage of this work made it desirable to investigate other tech-
niques which might meet with greater success. Therefore the emphasis
was shifted towards what was felt to be a more quantitative analysis,
namely the measurement of cake density with depth into the filter.
3. Measurements of Cake Density as a Function of Depth into Fabric
The word cake tends to mislead because it conveys the image of a
sharp boundary between the deposited dust layer and the fabric. This
is not in fact true, since dust and fabric interpenetrate to an extent
which ultimately determines both efficiency and pressure drop character-
istics.
Card webs were made from round and trilobal 3-denier polyester
(PET) fibers, each containing a minor amount of binder fiber. Pieces
of card web of different thickness were compressed under heat to form
a large number of thin fabric patches. These were trimmed and weighed.
Layered fabrics were then formed from patches selected so that cor-
responding layers in the round and trilobal fabrics would be close in
weight. By controlling the forming pressure, densities of the fabrics
were made as uniform as possible.
After filtering a flyash aerosol through these layered fabrics,
the layers were peeled apart and weighed. The results of three such
experiments, each comparing a round and a trilobal filter, are shown
in Figures 4, 5, and 6. In each case, the penetration (1-E) and the
single fiber efficiency n are plotted vs. the depth into the filter
("a" and "b" plots), and the single fiber efficiency is plotted vs.
the local aerosol concentration ("c" plots) .
As seen in Figures 4c, 5c, and 6c, the collection efficiency is
rather sensitive to both aerosol concentration and packing density a,
and consequently the "a" plots do not show a consistent trend, the
penetration for trilobal fibers being less than for round in Figures 4a
and 6a but more in Figure 5a. Greater consistency is seen in the "b"
figures, which normalize for packing density by plotting the single
fiber efficiency r\. This quantity is calculated from the relationship
„ _
n
_ ftn(l-E)
-
Sx - ' 2ct
where (1-E) is the penetration, Ax the thickness of the layer, R the
fiber radius, and a the packing density- The penetration (1-E) for
each layer used in this calculation is the value given in the "a" plot,
divided by the value for the previous upstream layer. The "b" plots,
being normalized for packing density, give a truer comparison of the
two types of fabric, and indicate a tendency for the efficiency to be
higher for trilobal fibers in the upstream layer, but to be equal in
the downstream layers. This is brought out slightly better in the
"c" figures, which eliminate the effect of concentration by plotting
-------
PENETRATION
1.0
0.8
06
O.4
(0)
I
I
SINGLE FIBER
EFFICIENCY
Q8
0.6
O.4
O.Z
.02 .04 O6 08 .10 .12
DEPTH INTO FILTER (cm)
.14
.16
(b)
o 02
SINGLE FIBER
EFFICIENCY
i.0r
OB
0.6
O.4
X>4
.06 .08 .10 J2
DEPTH INTO FILTER (cm)
.14
.16
(c)
O.OI
0.1 IO
AEROSOL CONC. (g/ms)
10
Figure 4. Filtration performance of layered bonded (15% vinyon)
nonwovens made of round (o) and trilobal (x) fibers. Face
velocity 12 fpm. Points plotted at center of layer thickness.
(a) Penetration vs. depth into filter.
(b) Single fiber efficiency vs. depth into filter.
(c) Single fiber efficiency vs. aerosol concentration.
-------
PENETRATION
05 -
O.4
O3
O2
O.I
.04
SINGLE FIBER
EFFICIENCY
1.0 *•
Q8
0.6
0.4
O.2
.08 .12 .16 .20 .24
DEPTH INTO FILTER (cm)
(b)
-o—x
SINGLE FIBER
EFFICIENCY
1.0I-
O8
O.6
0.4
02
.1 2 .3
DEPTH INTO FILTER (cm)
(c)
•0.080
a -0.088
0.01
O.I 1.0
AEROSOL CONC. (g/m5)
Figure 5. Filtration performance of another pair of layered non-
wovens made of round (°) and trilobal (x) fibers.
(a) Penetration vs. depth into filter.
(b) Single fiber efficiency vs. depth into filter.
(c) Single fiber efficiency vs. aerosol concentration.
-------
PENETRATION
.12
O4
SINGLE FIBER
EFFICIENCY
IX)
OB
O.6
O.4
.06 08 10 J2
DEPTH INTO FILTER (cm)
14
.16
.18
(b)
.02
SINGLE RBER
EFFICIENCY
1-0 r
O.8
06
Q.4
.04 .06 JOB JO
DEPTH INTO FILTER (em)
.12
.14
a. 0.150
•O.ISI
OX) I O.I I.O 10
AEROSOL CONC. (g/hl5)
Figure 6. Filtration performance of a third pair of layered non-
wovens made of round (») and trilobal (x) fibers. Note: These
fabrics are almost twice as dense as those of Figures 4 and 5.
(a) Penetration vs. depth into filter.
(b) Single fiber efficiency vs. depth into filter.
(c) Single fiber efficiency vs. aerosol concentration.
8
-------
single fiber efficiency vs. concentration. The efficiency rises
rapidly with concentration, clearly due to the accumulation of dust
cake, which occurs more rapidly on the upstream layers, rather than
to any change in the capture mechanism. In these "c" plots the
efficiency is greater for trilobal fibers than for round, but only for
the layers on which a dust cake has collected. At lower concentra-
tions (i.e., the downstream layers) efficiencies are equal. This find-
ing is consistent with previous observations of differences in perform-
ance between fabrics made with round and trilobal fibers only when
filtering aerosols of sufficient concentration to deposit a cake. With
very dilute aerosols, no differences were seen.
In explaining the superior performance of one type of fiber over
the other, these findings do not appear to lead to very simple models,
since the downstream layers do not exhibit the better single fiber
efficiency predicted for lobed fibers by theoretical considerations.
However, some mechanism must be acting to cause the differences in
efficiency in the upstream layers.
4. Composite Filters
The results in the previous section prompted initial studies in-
volving composite filters. These would be filters made of more than
one fiber type, with the different fibers arranged in layers and/or in
blends in such a way as to gain the advantage given by each type of
fiber.
The first study involved a set of fabrics made of layers as shown
in Table 1. Two of the fabrics were made up of only either trilobal
or round fibers, the other two had an upstream layer of trilobal fibers
in one case and of round fibers in the other, the remainder being of
the other fiber. The upstream layers amounted to one sixth of the
filter mass. The filters were all conditioned for 24 cycles with a
flyash aerosol of 7 g/m3. Table 1 gives the results. It is clear
that the behavior of the filters is dominated by the nature of the
upstream layer, with the downstream layers playing a minor role. These
results agree with the data (e.g., Fig. 6b) showing that the efficiencies
of downstream layers are equal for round and trilobal fibers. They
also support data obtained previously with filter bags, which showed
that trilobal bags gave higher efficiency and lower pressure drop than
round ones [1].
The reason for the lower pressure drop exhibited by the filters
with trilobal upstream layers is not readily apparent. The lower
pressure drop is observed with trilobal fabric even though the cake
mass is higher in the upstream layer of that fabric. It follows that
the cake density must be less. This however could not be confirmed by
microscopical observation, neither could any special features of the
cakes be discerned that would explain the formation of a less dense
cake.
-------
TABLE 1. PERFORMANCE OF COMPOSITE FABRICS IN FILTERING A
FLYASH AEROSOL OF 7 g/m3 AFTER
24 CYCLES OF CONDITIONING
Upstream layer Downstream layer C (mg/m3) AP(mm
Trilobal Round
(1.2 oz/yd2) (6.2 oz/yd2) 3.44 69.7
All trilobal
(7.4 oz/yd2) 3.33 68.2
Round Trilobal
(1.2 oz/yd2) (6.2 oz/yd2) 6.67 76.8
All round
(7.4 oz/yd2) 6.78 78.9
Filter thickness: 'ul.S mm
Surface layer thickness: ^0.3 mm
In a practical application, greater efficiency might be obtained
by combining an upstream layer that is efficient at the inlet concen-
tration with downstream layers that are more efficient at the lower
concentrations found deeper in the filter; as for example, by utiliz-
ing larger numbers of finer fibers (at the same packing density) for
some of the downstream layers.
5. Observations on Fiber Arrays
The results of the preceding section indicate that efficiency is
controlled by the nature of the fibers (trilobal vs. round) in the
upstream layer but not in the downstream layers. This suggests that
in the upstream layer, even though the efficiency of the bare fibers
may be equa'l, the dust accumulates in such a manner as to facilitate
further collection and the more rapidly formed cake promotes greater
efficiency. Since microscopical examination of dust cakes on random
nonwovens revealed no features that would confirm such an hypothesis,
some time was devoted to attempts to build incipient dust cakes on
regular fiber arrays that might, by occurring in a more regular fasion,
be easier to observe.
The arrays were made by hand and consisted of sets of parallel
fibers separated by a distance of 3 fiber diameters. The arrays were
formed on aluminum annuli which had been made to fit a scanning
electron microscope stage. On each annulus, arrays of trilobal and
round fibers were formed side by side in order to make direct compari-
sons of the features of the deposited dust.
Three kinds of arrays were made, of increasing complexity. The
first consisted of a row of eight parallel fibers. The second con-
sisted of two layers of eight-fiber rows, the rows oriented perpen-
dicular to each other. The third consisted of three eight-fiber sets,
alternating in orientation, such that there were three layers of fibers
at the intersections.
10
-------
In each case, the annulus was placed in a duct and exposed for
five minutes to a 12-fpm stream of flyash aerosol. The annulus was
then examined under a binocular microscope, which gave adequate
magnification for rapid inspection of the dust deposits. It was
planned to use scanning electron microscopy for closer study and for
recording any structure of interest. However, no such structures were
seen. Capture on single rows of fibers revealed a great deal of non-
uniformity between fibers of the same kind and even from place to
place on one fiber. This nonuniform behavior made impossible meaning-
ful comparison of capture between the two kinds of fiber. In any case,
the dust buildup that was the most important feature sought was not
observed. The overlapping rows of fibers, which were expected to give
more stable accumulations, did not do so. In all cases some capture
occurred, but the proximity of adjacent fibers did not affect capture
on any individual one. The uneven deposition on individual fibers
was assumed to be due to areas of electrostatic charge.
This study of model fabrics will continue with some of the
following modifications:
1) Metallizing of the fibers to eliminate uncontrolled static
surface charges, and neutralization of the aerosol particles. This
will be done to ensure a measure of control on electrostatic effects,
which otherwise occur in a random fashion.
2) Changes in the array construction to promote the formation
of dust cakes. Although the fiber-fiber spacings, equal to three
fiber diameters, are about equal to average interfiber spacings in
a nonwoven fabric, the parallel arrangement in the model fabrics may
require smaller spacings in order to form a stable cake.
3) Cake formation in the presence of electric fields. The
effects which are to be studied in these conditions are discussed in
a following section (III).
B. ROUGH AND SMOOTH FIBER SURFACES
Although present studies of fiber geometry effects on dust cake
formation have concentrated on round and trilobal fiber cross sections,
some time has also been devoted to differences due to rough and smooth
fiber substrates. In these studies the rough fibers contained a
greater percentage of Ti02 (2.0%) than the smooth fibers (0.1%), so
that a slightly greater number of surface asperities were present in
the rough fibers. Previous studies indicated that rough fiber sub-
strates produce slightly greater filter efficiencies than round, but
no significant differences were seen in pressure drop after condi-
tioning [1] .
Studies of dust cake formation on rough- and smooth-fiber needled
nonwoven filters thus far mainly consist of microscopical examination.
Studies of layered or composite filters have not yet been conducted.
However, in the next section the effect of particle charge is considered
for rough and smooth fiber filters as well as for trilobal and round.
11
-------
Trilobal
Round
Rough
Smooth
Figure 7. Scanning electron micrographs comparing dust cakes formed
on (a) trilobal and round fiber filters (SOX), and (b) rough and
smooth fiber filters (100X) by an untreated flyash aerosol.
12
-------
Microscopical examination followed the same procedure discussed
in the previous section. High magnifications of light dust loadings
revealed no differences. SEM micrographs of heavier dust loadings
such as the ones shown in Figure 7b also do not reveal any differences
in cake structure formed on the two types of fabric. At SOX, unlike
the comparison between trilobal and round fiber substrates, no dif-
ferences in cake uniformity can be seen. This observation agrees with
the results in Figures 11 and 12, which indicate no differences in
pressure drop or efficiency for the two fabrics on the first cycle.
Since earlier studies indicated slightly greater efficiencies for the
rough fiber fabric after conditioning, further examination of cakes
formed after many cycles will be conducted.
C. EFFECT OF PARTICLE CHARGE
1. Neutralized Aerosols
Further studies of trilobal vs. round and of rough vs. smooth
fiber filters were conducted using the same needled fabrics. The
objective was to observe the differences in performance as a function
of cake mass and as a function of particle charge. In this case the
flyash aerosol was passed through a charge neutralizer before it went
on to the filter.
SEM photographs of the dust cakes obtained with the neutralized
aerosol are shown in Figure 8. Those formed on trilobal fabrics appear
to be more uniform than those on round fabrics (Fig. 8a). The dust
cakes formed on rough and smooth fiber fabrics, on the other hand,
appeared to be very much alike (Fig. 8b). The latter observation is
not unexpected in view of the absence of distinguishing features in
dust cakes formed on rough and smooth fiber fabrics with untreated
aerosol, i.e., with aerosol having more charged particles. If an
important effect of lobes or surface asperities is to produce locally
intensified electric fields due to naturally occurring electric
charges on the fibers, the removal of the particle charges will reduce
any difference between the cake formed on rough and smooth fibers.
Results of pressure drop and efficiency measurements with neutra-
lized aerosol are given in Figures 9 through 12. Each data point was
obtained with the filter freshly cleaned by vacuum cleaning. In Fig-
ure 9 it can be seen that for unneutralized particles pressure drop
increases with cake mass equally for trilobal and for round fiber
filters until an abnormally high cake mass is reached (^30 mg/cm2).
The "trilobal" curve then increases rapidly, whereas the "round" curve
remains approximately linear. The trilobal structure with its smaller
interfiber spacings will therefore eventually cause pressure drop to
rise more rapidly than the round structure. However, at the higher
cake mass efficiency continues to remain higher for trilobal than for
round fiber filters (Fig. 10).
In Figure 9 it can be seen that the same AP relationship between
trilobal and round holds for neutralized particles, although the curves
are shifted in the direction of greater AP at lower cake mass. It
appears that the presence of uncharged particles results in more
13
-------
Trilobal
Round
Rough
Smooth
Figure 8. Scanning electron micrographs comparing dust cakes formed
on (a) trilobal and round fiber filters (200X), and (b) rough and
smooth fiber filters (100X) by a neutralized flyash aerosol.
-------
(mmHjp)
Neutralized Aerosol /
30p Normal Aerosol ^——— n
Trilobal fibers * /
Round " O /
25
20
IS
10
0 10 2O 3O
MASS OF DUST CAKE PER UNIT AREA(mg/cma)
Figure 9, Pressure drop developed by increasing dust cake
deposits of untreated and neutralized flyash aerosols on
trilobal and round fiber filters.
TOTAL EFFICIENCY
1.0
Neutralized Aeroiol-
Normal Aerosol
Trilobal fibers » *
Round * O
099
0.98
as?
096
10 2O 3O 4O
MASS OF DUST CAKE PER UNIT AREA (mg/cm*)
Figure 10. Efficiency of trilobal and round fiber filters with
increasing deposits of untreated and neutralized flyash.
15
-------
compact, less porous cakes. The point at which the "trilobal" pressure
drop curve begins to climb above the "round" curve is at a lower cake
mass (18 mg/cm2) for flyash passed through the neutralizer. This is
still higher than normal operating cake mass. Baghouse comparisons
have verified that trilobal fiber filters do not develop higher pres-
sure drops than round fiber filters up to dust loadings of 24 mg/cm2
[1,2], Therefore, these patch-filter data put a practical limit on
the finding that trilobal fiber filters are not detrimental to pressure
drop even though interfiber spacings are smaller.
The data in Figure 10 show that trilobal fiber filters are more
efficient than round fiber filters, even in this high cake mass region.
For particles passed through the neutralizer, the effect of cake mass
on capture efficiency is more acute, since both lines rise quite
rapidly from lower efficiency values at low cake mass.
In Figure 11 it is seen that pressure drops for rough and smooth
fiber filters behave similarly as a function of cake mass for both
neutralized and unneutralized particles, even at high cake mass. Once
again neutralized particles produce cakes with higher pressure drop.
In Figure 12 no clear difference in capture efficiency between rough
and smooth fiber filters can be seen at any cake mass or for either
aerosol. These findings verify earlier results for overall efficiency-
The previously observed tendency for conditioned rough fiber filters
to capture more submicron particles than smooth ones and the dependence
of this tendency on particle charge remain subjects for further study.
2. Effect of Particle Precharging
A final trial with needled filters involved the deposition of
extra charges on the aerosol particles. A 15-cm long, 3-cm diameter
corona discharge tube was made with a 45-ym diameter axial wire and
grounded wall. The aerosol was passed through the tube before pro-
ceeding to the duct upstream of the patch filter. With an applied
voltage of 10 kV a purple glow could be seen near the wire, and when
the flyash aerosol was passed through the tube it was found that no
dust emerged from the downstream end, all of it having been deposited
in the tube by electrostatic precipitation. The voltage was therefore
reduced to the point where some dust could be seen to emerge from the
tube, even though at this level there was no visible discharge and no
detectable current. Filtration performance of trilobal and round
fiber fabrics for this aerosol were compared, and the results are
given in Table 2 together with those for a neutralized flyash. For
both trilobal and round fibers, efficiency increases and specific
cake resistance decreases as the particle charge increases. This gives
evidence of the strong influence of coulombic forces. The actual
charges on the particles could not be measured at this time. The
ability to measure particle-charge distribution would, however, be an
asset in the quantitative interpretation of electric phenomena in
fabric filtration, and efforts will be made to secure appropriate
instrumentation.
16
-------
30
25
20
15
10
Neutroliied Aerosol /
- Normal Aerosol
Rough fibers x
Smooth "
10 2O 30
MASS OF DUST CAKE PER UNIT AREA (mg/cm2)
Figure 11. Pressure drop developed by increasing dust cake
deposits of untreated and neutralized flyash aerosols on
rough and smooth fiber filters.
TOTAL EFFICIENCY
lOOr
0.99
0.98
0.97
Neutralized Aerosol
Rough Fibers
Smooth •
Normal Aerosol
Rough Fibers
Smooth '
10 20 30
MASS OF CAKE PER UNIT AREA(mg/em»)
Figure 12. Efficiency of rough and smooth surface fiber filters
with increasing deposits of untreated and neutralized flyash.
17
-------
TABLE 2. CAKE RESISTANCE AND EFFICIENCY OF ROUND AND
TRILOBAL FIBER FILTERS WHEN FILTERING AEROSOLS
CHARGED TO VARIOUS LEVELS
Round Trilobal
(IP"* N*s/Kq*nO E(%) (10* N'S/kg-m) E(%)
Neutralized 1.20 98.30 1.17 98.44
As is 0.60 99.69 0.55 99.86
Charged (5 kV) 0.47 99.88 0.47 99.94
18
-------
SECTION III
STUDIES OF CAPTURE IN ELECTRIC FIELDS
A. THEORETICAL STUDIES
1. Effect of Field Orientation
The following section gives an account of computer studies on
the capture efficiency of single fibers and mats in the presence of
electric fields. Several subjects are dealt with. The first con-
cerns the orientation of the electric field with respect to the
flow direction and the fiber axis. In 1965, Zebel [2] set forth a
theory to account for the effect of an electric field in the capture
of particles on cylindrical fibers. Since he envisioned practical
applications that would use an electric field parallel to the flow
direction, he did not consider the effects of field orientation.
What follows is an extension of Zebel's theory to perpendicular
fields.
The cases of charged and uncharged particles must be considered
separately. Taking first the case of uncharged particles that are
acted on by polarization forces caused by the electric field in the
presence of a fiber, we note that Zebel showed the collection effi-
ciency for such uncharged particles to depend on two dimensionless
parameters, F and a. The latter can have values between 0 and 1,
with 0.5 being a typical value. F depends on: the radius of the
fiber r , the particle radius r , the field strength E0, the gas flow
velocity V0, the gas viscosity y, and the dielectric constants of the
fiber and the particle, a and o_, respectively. F is given by
c s
* u _ — j. i u _~j. r
F = *
I
r
c
For field strengths of 2 kv/cm and gas velocities of about 12 fpm, F
will vary typically between 10~3 and 1.0.
For this problem, Zebel's equations do not have an analytical
solution, so that the collection efficiencies were determined by
numerically integrating the particle trajectory equations.
»
Table 3 presents values of the collection efficiency for several
values of F; n|| denotes the efficiency when the field is parallel
to the flow, and nj_ when the electric field is perpendicular. The
collection efficiency is defined as the thickness of the layer in
the undisturbed flow (H in Fig. 13) in front of the fiber from which
all aerosol particles are captured, divided by the fiber diameter
2 r . We note from Table 3 that over the range of values of F that
werS studied, a field that is perpendicular to the flow direction
yields a higher collection efficiency in every case.
19
-------
Figure 13. The two limiting particle trajectories for a round
fiber of radius r^.
e
3.0
2.0
I.O
8.01
O.I
I.O
100
100.0
Figure 14. Enhancement e of collection efficiency by a field
perpendicular to the flow for charged particles as a function
of |G| = qE,B/V8.
20
-------
TABLE 3. COMPARISON OF COLLECTION EFFICIENCIES FOR
UNCHARGED PARTICLES ACTED ON BY ELECTRIC
FIELDS PARALLEL AND PERPENDICULAR TO THE FLOW DIRECTION
Improvement of
F n|| nj. n_[ over n| I (%)
3.2 1.386 1.517 9.5
1.0 0.630 0.778 23.5
0.5 0.390 0.477 22.5
0.2 0.206 0.240 16.5
0.1 0.115 0.136 18.3
0.02 0.0295 0.0337 14.2
0.01 0.0127 0.0164 29.1
We now consider the collection of charged particles on cylindrical
fibers. Zebel showed that in the case where the electric field is
parallel to the flow, the collection efficiency is given as:
n|| = o for -l
-------
typically met in experimental work the preferred orientation of the
electric field is perpendicular to the flow direction.
2. Fibers With Lobed Cross Sections
A second question that was attacked was that of the effect of
fiber cross-sectional shape on capture efficiency when electric fields
are present. Round and multilobal shapes were considered.
In cylindrical polar coordinates (p/4>) the equation that describes
the boundary of a cylindrical fiber is p=l. Multilobal fibers are
appropriately described by the equation
p = 1 + e cos m(+c) ,
where m is the number of fiber lobes, c is a phase angle that deter-
mines the orientation of the fiber to the flow, and e is a parameter
that determines the size of the lobes and is never greater than 1.0.
/>•!•»•< cos m(4>+c)
«-0.2
c -O.O
o O O
m- 3
e-O.2
c •
o
m -3
c-0
« •
D
O.I
O
02
OS
Figure 15. Fiber cross sections for selected values of m
(number of fiber lobes), c (orientation of fiber to air
flow), and c (size of lobes).
Figure 15 shows how the cross sections vary for different values
of these parameters. The equation for the surface of the fiber
suggests that it may be possible to solve for the velocity and
electrical force fields as regular perturbation series in the
parameter a. We have obtained these solutions and have subsequently
solved the trajectory equations for multilobal fibers. Figures 16
and 17 indicate these trajectories for bilobal and trilobal fibers.
22
-------
Figure 16. Particle trajectories near a bilobal fiber. Field
parallel to air flow, o = 0.5, and F = 1.0. All particles
within the region 2t will be captured.
EIIV F-1.O
0-0.3
Figure 17. Particle trajectories near a trilobal fiber.
parallel to air flow, a = 0.5, and F = 1.0.
Field
23
-------
Since these fiber cross sections are not symmetric with respect
to the flow, the collection efficiencies must be averaged over all
possible orientations. Figure 18 shows the collection efficiencies
(evaluated for F = 1.0) for a trilobal and a bilobal fiber in an
uncharged-particle aerosol. Because of the symmetry of the problem,
only values of c between zero and ir/m need be considered. Shown at
the bottom of Figure 18 are the orientations of the fibers for which
the collection efficiencies were calculated. The dotted lines are the
values of the collection efficiencies averaged over all orientations
of the fibers.
COLLECTION
EFFICIENCY
15 k
I.O
05
Trilobal Fiber
EJ.V
Bilobal Fiber
Figure 18. Comparison of collection efficiencies for bilobal
and trilobal fibers at different orientations of fiber to
electric field.
Table 4 compares the average collection efficiencies of circular,
bilobal, and trilobal fibers for F = 1.0. The results indicate that
when the electric field is perpendicular to the fiber and to the gas
flow, the collection efficiency is greater than when the field is
parallel to the gas flow direction.
TABLE 4. COLLECTION EFFICIENCIES FOR UNCHARGED PARTICLES
ON l-,2-^ AND 3-LOBE FIBERS IN ELECTRIC FIELDS
Fiber
cross section
Cylindrical
Bilobal
Trilobal
E||V
0.630
0.668
0.697
E_[V
0.778
1.015
0.953
Improvement of
i over || (%)
23.5
51.9
36.7
24
-------
To calculate the effects on the capture of charged particles we
consider the effective diameter of the fiber. Figure 13 illustrates
the two limiting trajectories for a round fiber. The effective fiber
diameter H is defined to be the distance, measured far upstream from
the fiber, between the two limiting trajectories. The single-fiber
collection efficiency is then defined as the effective fiber diameter
divided by the projected area per unit length of fiber; for the
circular fiber shown above this would be H/2r , where r is the fiber
radius.
A criterion is now determined on which the capture of particles
on multilobal fibers can be compared. Multilobal fibers can have much
larger effective diameters than circular fibers, but, since they can
have larger projected area, the effect on the collection efficiency will
be less pronounced than the effect on the effective diameter.
The single-fiber effective diameter can be related to the collec-
tion efficiency of the fiber mat z; by considering a mass balance
on a differential element of the mat. The cross-sectional area of a
multilobal fiber is Trr 2 (1+^e2) , or TT r2 for round fibers where
e=o. If we define If ai the length of fiber in a unit volume of mat,
then the volume fraction of fibers a is given by
a = irr 2 1- (H-Jse2)
C L
If Q is the flow rate of air into the mat, we expect the air velocity
inside the mat V to be
V ~ A(l-a)
where A is the area of the mat perpendicular to the flow. This velo-
city is greater than the velocity of air which approaches the mat, since
the portion of space in the mat available to accommodate the air is
(1-ct).
Consider now a differential element whose thickness is AZ and
T/hose area is A. For an aerosol of concentration C, the flow rate of
aerosol into this element is CQ, and the flow out is (C+AC)Q . The
total length of fiber in this element is AAZ If; therefore the amount
of aertfsol that gets captured on these fibers is CVHl-AAZ. A mass
balance on the aerosol yields
CQ - (C+AOQ = CVHlfAAZ
(Aerosol In) (Aerosol Out) (Aerosol Captured)
Substituting for lf and letting AZ, AC >° yields
dc _ -Hot
25
-------
Let C be the aerosol concentration entering the mat and Cf the
concentration leaving it. The integration of the above equation
yields
l
where Y = - and L is the mat thickness . The overall mat
irr '(1-Mje2)
collection efficiency is then
Later results may be more easily appreciated after a simple cal-
culation to demonstrate the importance of y. Suppose that a mat has
a collection efficiency of 0.900 when Y = Y0 • Jf* by changing the
field- flow orientation or by adding fiber lobes, y can be increased by
50%, C becomes 0.965; for a 100% increase in y, £=0.990. Since we
wish to compare S for mats of equal volume fraction and thickness, the
best way to compare the collection performance of multilobal fibers is
then to compare values of H/ (l+%e 2) .
It has been shown that the effective fiber diameter for the
collection of charged particles on single fibers is
HI « 2 S Z(a,e,m),
when the field and flow are parallel, and
H2 - -' G
where the field is perpendicular to the flow. The quantity Z is a
function of only fiber geometry and a, where
These values of H are the effective fiber diameter averaged over all
fiber orientations
As in the case of round fibers, we see that the perpendicular
G+l
electric field case provides an enhancement of e = • when all
*^5TT
26
-------
the particles are positively charged, and e = * ' when there are
/G2+1
equal numbers of positively and negatively charged particles. The
enhancement due to the field orientation is therefore independent of
fiber geometry.
It can be shown that for a circular fiber Z = 1 + a; therefore
the largest value of Z occurs at a = 1. Since Z is expected to be a
maximum at a = 1 for multilobal fibers, and since the attempt here
was to investigate the extent to which multilobal fibers provide an
enhancement of the fiber mat collection efficiency, we confine our
attention to the case where a = 1.
As discussed earlier, the appropriate parameter for comparing
multilobal and circular fibers is H/(l+%£2). Since Z does not depend
on the field-flow orientation and therefore isolates the effects of
a, e, and mf then Z/fl+^e2) can be seen to provide the most useful
index for evaluating the effect of fiber geometry on the fiber mat
collection efficiency. Since the value of Z/d+^e2) is (1+a) for
cylindrical fibers, the enhancement over cylindrical shape provided
by a multilobal shape will be given by Z[(l+a)(1+^e2)] or, for
a = 1, Z/(2+e2) .
Figure 19 shows a plot of this latter parameter vs. e for various
multilobal fibers; the integers on each curve are the values of m
along that curve. This figure indicates that for commercially avail-
able bilobal and trilobal fibers (where e £ 0.3) enhancements of up to
50% can be expected. Figure 20 attempts to portray the levelling out
of this enhancement with m. We can extrapolate these curves to
estimate the effect of fiber surface roughness for which, say, e = 0.01
and m = 100. The extrapolated value of the enhancement would be
approximately 3-4%, thus, causing a very small effect on the fiber mat
collection efficiency.
Figure 19 suggests that fiber lobes will always enhance the
collection efficiency. This statement is true except for the case of
bilobal fibers for which the enhancement strongly depends on the
orientation of the fiber to the air flow. Figure 21 illustrates the
wide range of enhancements that are possible with bilobal fibers. The
effect is greatest when the major axis of the fiber is perpendicular to
the flow (2c/ir = 0.5). When the fiber lies flat and the lobes are
aligned"with the flow, the collection efficiency is less than that for
round fibers. For trilobal and other multilobal fibers this orienta-
tion dependence is not important. Therefore our results suggest that,
except for bilobal fibers, fiber lobes will always enhance the col-
lection of charged particles and that this effect levels off after
about five lobes.
The above calculations indicate that in the presence of electric
fields, rough or lobed fibers produce greater collection efficiencies
than smooth or round fibers. This may provide a partial explanation
for the better performance of mats of lobed fibers. If the aerosol
contains charged particles, their deposition will set up electric
fields near the upstream fabric surface. In the presence of these
27
-------
1.6
1.7
16
15
~ I-*
~.
I"
ni 12
1.1
\0
00 OJ05 0.1 QI5 0.2 0.25 O3
Figure 19. Comparison of the effective diameters of multilobal
fibers (for charged particles). Number of lobes indicated on
curves; e indicates depth of lobes.
1.4 r
+
Q
L2
IX)
e-O.I5
e • O.05
345
m. NUMBER OF LOBES
IO
Figure 20. Effective fiber diameter vs. number of lobes
(for charged particles).
28
-------
fields, capture of subsequent particles will be more efficient on
lobed fibers. This hypothesis will be tested in a future grant
period.
1.7
16
1.5
1.4
- 1.3
«i
01
Q 12
CM
ru I.I
I.O
OS
0.8
«-030
e -O25
• 0.20
e>0.l
0.05 OJO O.I5 0.20 025 03O 035 0.4O 045 O50
2C/7T ORIENTATION
Figure 21. Effect of fiber orientation on the effective diameter
of a bilobal fiber (for charged particles).
29
-------
GROUNDED
GRID
TO BAGHOUSE
FINE WIRES
Figure 22. Diagram of tube for precharging flyash aerosol
(negative precharging arrangement).
30
-------
B. EFFECT OF ELECTRIC FIELD ON CAKE STRUCTURE
1. Introduction
Applying a high voltage across wire electrodes embedded 15 mm
apart just below the filter surface has produced not only higher fil-
tration efficiencies but also lower pressure drops. The increase in
efficiency is considered to be a result of the deflection of particles
from the normal air stream under the influence of the electric field.
This increases the probability of impact with a fiber or a particle of
dust in the dust cake. Consequently, a lower concentration of particles
passes through the filter. If incoming particles are already charged
naturally or by corona precharge, they will drift in the direction of
the electrode of opposite charge. If incoming particles are neutral,
they may be influenced by polarization forces acting in the non-uniform
field in the neighborhood of a fiber. Rivers [3] considered the two
mechanisms of coulombic and polarization forces and raised the question
of whether coulombic forces alone may not account for observed effi-
ciency improvement.
The decrease in pressure drop under the influence of the electric
field is more difficult to explain and is the main subject of this
study. The lower pressure drop is most likely due to a different cake
structure. Because of the magnitude of the effect, differences in cake
structure should be detectable.
2. Experimental Results
Flyash cakes formed under the influence of an electric field on
patch filters with wire electrodes embedded below the upstream fabric
surface were examined in a binocular microscope. Incoming particles
were used as they were from the feeder, or else they were passed
through a precharging tube or through a neutralizer. The precharging
tube is shown in Figure 22. It contains an open wire mesh
electrode located upstream from a series of fine wires as the other
electrode. The neutralizer is a tube containing a radioactive
source (krypton 85). The experimental design was as follows:
Precharge Potential gradient
Trial voltage Positive Negative across filter
No. (kV) downstream downstream (kV/cm)
1 ' 11 X 0
2 11 X 2
3 11 X 2
4 14 X 2
5 14 X 2
6 Particles through neutralizer 0
7 Particles through neutralizer 2
8 Untreated particles from feeder 0
9 Untreated particles from feeder 2
10 Two isolated electrodes-particles from 2
feeder
31
-------
The following observations are illustrated in the photographs of
Figures 23-30:
1) No field across filter
If a potential gradient is not applied to the filter (Trial
nos. 1, 6, and 8, Figs. 23-25), the cake deposits uniformly over the
filtering surface. Neutralized particles (Fig. 24) appear slightly
more densely packed than charged particles (Figs. 23 and 25), which
apparently have a tendency to collect closer to the surface and do
not collapse into pores as readily. These observations agree with
previous pressure drop measurements in which the highest value was
obtained for neutralized particles and the lowest values for charged
particles.
2) Potential gradient applied across filter - 2 kV/cm
When a potential gradient of 2 kV/cm is applied across the filter,
the cake no longer deposits uniformly over the filtering surface. Using
untreated particles (Trial no. 9, Fig. 26), the amount of dust is
greater on the negative electrodes with particles depositing on fibers
raised above the surface. The positive electrodes are covered by dust
but deposition is lower. If particles are precharged positively
(Trial no. 2, Fig. 27), cake formation is similar to that of Trial
no. 9. The implication is that untreated dust from the feeder is pre-
dominantly positively charged, so that it behaves like the positively
precharged dust and deposits mainly on the negative electrodes of the
filter.
If the precharge is then reversed so as to charge the particles
negatively (Trial no. 3, Fig. 28), the deposition becomes equal on all
wires. Reversing the charge apparently does not change all the ori-
ginal positive charges to negative but does change enough of them so
that deposition is about equal. Increasing precharge voltage (Trial
nos. 4 and 5) had no observable effect.
If the particles are passed through a neutralizer (Trial no. 7,
Fig. 29), there is still a greater deposition of particles on the
negative electrodes in the filter, but the difference is not as pro-
nounced. Apparently not all particles are neutralized, and some
positively charged particles still remain in the aerosol.
These observations clearly support the dominance of the coulombic
force in the electrical stimulation of fabric filtration (ESFF). If
polarization forces were dominant, the distribution of dust would be
equal on all electrodes. The preferential deposition over certain
electrodes in all these cases is more easily explained by the coulombic
mechanism.
3) Isolated pairs of electrodes - 2 kv/cm
In another experiment two electrodes 10 mil in diameter were
embedded on one half of the filter and two electrodes 2 mil in dia-
meter were embedded on the other half. Any differences in pattern
32
-------
Figure 23. Positively precharged particles Figure 24. Neutralized particles deposited on
deposited on filter in absence of applied filter in absence of applied field across
field across filter. filter.
Figure 25. Naturally charged particles
deposited on filter in absence of applied
field across filter.
33
-------
Figure 26. Naturally charged particles
deposited on filter with 2 kV/cm across
filter.
Figure 27. Positively precharged particles
deposited on filter with 2 kV/cm across
filter.
-------
Figure 28. Negatively precharged particles Figure 29. Neutralized particles deposited
deposited on filter with 2 kV/cm across on filter with 2 kv/cm across filter.
filter.
Figure 30. Two sets of two electrodes with
2-kV/cm potential gradient using naturally
charged particles. (Left side: + -, 10 mil
diam. Right side: - +, 2 mil diam.)
35
-------
due to wire diameter are too slight to be noticed at 2 kv/cm using
naturally charged particles (Fig. 30). However, this sample filter
provides further clues to the mechanism of deposition in an electric
field. In the central region of the filter, there is no field since
the central electrodes (2 mils and 10 mils) are both negative.
Greater penetration into the filter occurs in this region as compared
to the regions between electrodes where dust is closer to the surface.
If only two wires are embedded in the filter surface and one has
a potential of 2 kv/cm, incoming untreated particles will deposit
over the negative wire but will leave the positive wire clean. With
alternating wires (Trial no. 9, Fig. 26), the positive wires received
some deposit. Apparently, the deposition on the positive wires in the
alternating wire arrangement is due to direct impingement in the region
near the positive wire where the field force drawing particles to the
adjacent negative wires is balanced. This effect would support the
supposition that particles coming directly from the feeder are mostly
positive. Measurement of actual particle charge should be performed
to verify this hypothesis. Similar measurements should also be made
on particles passed through the corona and through the neutralizer.
The observation that in the presence of electric fields dust de-
position is not uniform offers a possible mechanism to explain re-
ductions in pressure drop. Uneven dust deposition would create filter
regions with paths of different resistance to air flow. Resorting to
an electrical analogy, we note that the effective resistance of two
parallel resistors is highest when their resistances are equal. Fig-
ure 31 plots the effective resistance of two such resistors as their
values are unbalanced, while the sum of their resistances remains con-
stant, just as the total mass of dust reaching the filter remains
constant. It is seen that for this model a resistance inbalance of
about 6 to 1 cuts the pressure drop (effective resistance) in half.
3. Conclusions
Two mechanisms appear to act in reducing resistance to air flow
through a dust cake formed under the influence of an electric field.
The first is the tendency for particles to deposit closer to the sur-
face and not to collapse into the pores, thereby producing a cake with
a lower packing density. This effect occurs for charged particles even
in the absence of an electric field. The second mechanism is the un-
even distribution of dust on electrodes due to the coulombic attrac-
tion of charged particles to the oppositely charged electrode. The
more uneven the distribution the lower will be the resistance to air
flow. The two mechanisms combined produce a cake having much lower
specific resistance than one formed in the absence of potential grad-
ients .
36
-------
EFFECTIVE
RESISTANCE
I -
10 II 12 13 14 15 16 17 18 19 20 • R,
I09876S432I 0 « R2
Figure 31. Electrical analogy of unequal parallel resistances.
37
-------
SECTION IV
REFERENCES
1. Miller, B., G. Lamb, P. Costanza, and J. Craig, "Nonwoven Fabric
Filters for Particulate Removal in the Respirable Dust Range,"
EPA-600/7-77-115, U. S. Environmental Protection Agency, October
1977.
2. Zebel, G., Deposition of Aerosol Flowing Past a Cylindrical Fiber
in a Uniform Electric Field, J. Colloid. Sci. 20, 522-543 (1965).
3. Rivers, R. D., "Operating Principles of Non-Ionizing Air Filters,"
ASHRAE Journal, February 1962.
38
-------
TECHNICAL REPORT DATA
(Please read liiiimclions on the reverse before completing)
1. REPORT NO.
EPA-600/7-78-095
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Studies of Dust Cake Formation and Structure in
Fabric Filtration
5. REPORT DATE
June 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Bernard Miller, George Lamb, Peter Costanza,
Dan O'Meara. and Janet Dunbar
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Textile Research Institute
P.O. Box 625
Princeton, New Jersey 08540
10. PROGRAM ELEMENT NO.
EHE624
11. CONTRACT/GRANT NO.
Grant R804926
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final; 11/76-11/77
14. SPONSORING AGENCY CODE
EPA/600/13
15.SUPPLEMENTARY NOTES ffiRL-RTP project officer is James H. Turner, Mail Drop 61, 919/
541-2925.
i6. ABSTRACT
gives results of a. study to identify cake characteristics affecting
performance and, in turn, to relate the production of desirable cake properties to
fabric structure and filtration conditions. (Earlier studies showed that differences
in filtration performance produced by certain modifications of nonwoven filter fabrics
are due largely to the development of different dust cake structures. ) The amount of
dust capture as a function of depth into the filter was measured using nonwovens
formed in layers. Comparing trilobal cross -section filter fibers with round fibers,
the largest advantages in capture efficiency due to trilobal fibers are in the upstream
layers where the largest amounts of dust accumulate. Preliminary studies with com-
posite layered filters also show that filtration performance is dominated by the up-
stream layer. Dust cake structure is influenced by fiber geometry, by the charge
on the particles , and by any electric field on the filter. Theoretical calculations of
single-fiber collection efficiencies support the experimental findings that capture of
both charged and uncharged particles on fibers of any cross -sectional shape is more
likely if an electric field is imposed perpendicular rather than parallel to the flow.
As for fiber cross -sectional shape, capture should improve with increasing lobe
number and depth.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATl Field/Group
Pollution
Gas Filters
Fabrics
Nonwoven Fabrics
Dust
Caking
Electric Fields
Pollution Control
Stationary Sources
Fabric Filters
Particulate
13B
13K
11E
11G
07A,13H
20C
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (Tltis Report)
Unclassified
21. NO. OF PAGES
46
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
39
------- |