Note: Descriptions are shown in the official language in which they were submitted.
j L~
This Inven~ion relates to cyllndrlcal flbrous struc-
tures, more par~lcularly to cyllndrlcal Fibrous structures com-
~rlslng nonwoven, synthetlc, po I ymcrlc mlcroflbers p~rtlcularly
useful as depth fllters ~or a varlety o~ fluld clarlflcatlon
5 appllcatlons.
Thls applIcatlon Is a dlvlslonal applIcat!on of copend-
lng applicatlon No. 471,536 flled January 4, 1985.
Nonwoven structures formed from a varlety of materlals,
Includlng natural and synthetlc flbers In both staple and contln-
uous -Form, have long been known and used in depth fllter opera-
tlons. Such depth fllters generaliy have a range of pore dlame-
ters. If the fllter medlum Is thln, the larger partlcles In the
fluld belng flltered wlll pass through those areas havlng the
larger pores. If the effluent passlng through the fllter medium
Is then passed through a second equal layer, some of the larger
partlcles remalnlng In the fluld wlll be removed as they
encounter more flnely pored areas. Slmlla~ly, use of a thlrd
equal fllter layer wlll remove addltlonal large partlcles, fur-
ther Increaslng the flltratlon efflclency. Use of a thlc~ layer
of f 1 Iter medlum wlll have the same effect as uslng multlple lay-
ers of equal total thlckness. The Increased efflclency 50
obtalned is one of the motlvations for uslng depth filtratlon.
To be useful for a glven appllcatlon, a depth
;3
Lilter m~s~ provide the requisite level of effici-
ency, that is, an acceptable level of removal of
particles of a specified size present in a fluid
! being filtered. ~nother important measure of the
5performance of a filter is the time to clogging in a
given type of service, that is, the time at which the
pressl-re across the filter has either reached a level
at which an undesirable or unacceptable power input
is required to maintain adequate flow, or the poten-
lOtial Eor filter collapse with the accompanying lossof integrity and effluent contamination is too high.
To extend filter life, it has long been the
practice to design depth filters such that their
density is lower in the upstream portions, thus pro-
]5viding relatively larger pores upstream and smallerpores downstream. By virtue of the graded density,
the contaminated fluid passes through progressively
smaller pores, and particulate material being fil-
tered from the incident fluid penetrates to varying
20depths according to its size, thereby allowing the
- filter element to accomodate more solids (a higher
dirt capacity) without affecting flow and consequent-
ly providing a longer eEfective life for the depth
filter. Stated otherwise, in theory, the larger
25upstream pores remove larger particles which would
otherwise clog the downstream, finer pores and filter
life is thereby extended.
The density oE the Eilter meclium is, however, in
itself, an important determinant of the medium's
30behavior in service. The op-timum density of a filter
medium is determined by two factors:
(1) In order to have a high dirt capacity, the
percent voids volume in the depth filter should be as
high as possible. The reasons for this may be seen
35by comparing a gravel screen made using woven wire
3L25~ 3
--3--
with a metal plate e~ual in size to the screen but
containing a single hole. The metal plate will be
clogged by a single oversized particle, while the
screen, re~uixing a large number of particles to
5become clogged, will remain in service longer.
(2) In a fibrous depth filter, there is an
upper limit beyond which Eurther increasing the per-
cent voids volume becomes undesirable. As the voids
volume is increased, the fibrous depth filter is more
10 readily compressed by the pressure drop generated by
the fluid passing through it; this is particularly
troublesome when the fluid is viscous where, if the
percent voids volume is too high, the filter medium
will collapse at a very low differential pressure.
]5 As it collapses, the pores become smaller and the
differential pressure increases, causing still more
compression. The resulting rapid increase in pres-
sure drop then tends to reduce life rather than - as
might otherwise be expected with a high voids volume
20 filter - extending it. Use of a very low density
- (high voids volume) can also ma)ce the filter very
soft and thereby easily damaged in normal handling.
Thus, there is a practical upper limit to voids
volume, the value of which depends on the clean dif-
25 ferential pressure at which the filter is to be used.For any given type of service there will be an opti-
mum percent voids volume at which filter life will be
at a maximum.
~s noted above, attempts have previously been
30 made to provide depth ilters from fibrous materials
and to extend their effective life by providing a
graduated porosity, accomplished by a density profile
with the density increasing in the direction of flow
of the fluid being filtered. These attempts have met
35 with some success but the filter structures have
_ --4--
5 ~3 6 L~
substantial limitations. These include relatively
short life due to the limited range through which
pore diameters can be changed, and reduction in pore
diameters ~ue to compression when used with viscous
5 fluids or at very high liquid flow ra-tes.
The present invention, then, is directed to
cylindrical fibrous structures, particularly useful
as depth filters, and a method of manufacturing them
which substantially overcomes the shortcomings of the
10 cylindrical fibrous depth filters which have hereto-
fore been used. ~s will become apparent from the
following description of the invention, the cylin-
drical fibrous structures in accordance with this
invention typically have, relative to fibrous cylin-
]5 drical depth filters of the type previously avail-
able, extended filter life, i.e., higher dirt capac-
ity at e~ual efficiency, or better efficiency at
equal lifel or both better efficiency and higher dirt
capacity. They also have the ability to remove much
20 finer particulate contaminants -than have heretofore
been capable of being removed by previously available
commercial fibrous cylindrical depth filters.
In the method in accordance with this invention,
synthetic ~olymeric material is fiberized by extru-
25 sion into a high velocity gas stream and collected asa mass of mechanically entangled or intertwined fi-
bers in the form of a hollow or annular cylindrical
structure, as will be described in more detail below.
In the course of investigating this process, and the
30 products thereof, several surprising observations
were made:
(a) Increasing the percent voids volume (by
decreasing the density) generally yielded only a
small increase in filter life.
(b) When the project was initiated, it was
--5
~25~3~43
assume~ th~t iE two filters differing in fiber dia-
meter but otherwise equal were compared, the filter
with the finer ~iber would be more compressible.
Contrary to this assumption, it was found that the
5 filter with finer fibers had a resistance to com-
pression that was substantially equal to that with
coarser fibers, providing that the densities of the
two filters, i.e., the percent voids volume, were the
same. As a consequence of this discovery, cylin~
lO drical filter structures using fibers as fine as
about 1.5 micrometers were prepared and found to have
satisfactory resistance to compression. Cylindrical
filter structures, or cylindrical filter elements as
they are sometimes referrcd to herein, made with
]5 fibers in the range of from about 1.5 to about 2.5
micrometers and with annular thicknesses of the fib-
rous mass of about 0.6 inch (1.5 cm), had extraor-
dinarily high efficiencies, for example, in excess of
99.9999 percent for removal of bacteria organisms as
20 small as 0.3 microme-ter in diameter.
(c) secause of -the desirable characteristics
obtainable with depth filters prepared from these
very fine fibers, cylindrical filter structures with
fine fibers in their downsteam portions and coarser
25 fibers upstream but with a constant voids volume
throughout were prepared. These filter elements
combined extraordinary efficiencies, e.g., 9~.999
percent, removal for bacteria organisms as small as
0.3 micrometer in diameter with relatively high dirt
30 capacities comparable to dirt capacities of much
coarser conventional cylindrical depth filter ele-
ments.
The cylindrical fibrous structures in accordance
with this invention comprise a fibrous mass of non- -
35 woven, synthelic, polymeric microfibers, the fibrous
-6- ~ 5 ~6L~-~
mass having a substantialiy constant voids volume
over at least a substantial portion of the fibrous
mass, preferably at least the major portion, as meas-
ured in the radial direction. The microfibers are
5substantially free of fiber-to-fiber bonding and are
secured to each other by mechanical entanglement or
intertwining. Filter structures in accordance with
the subject invention are preferably supported by the
incorporation of a hollow, open, relatively rigid,
lOcentral suppor-t member or core, with the fibrous mass
of microfibers on the exterior of the support member.
~lso, for most applications, it is preferred that the
fibrous mass have a substantially constant -voids
volume and a graded fiber diameter structure over at
]5least a portion thereof as measured in the radial
direction, obtained by progressively varying the
fiber diameter as the cylindrical fibrous structure
is built up while simultaneously holding the voids
volume cons-tant.
The method of manufacturing the cylindrical
fibrous structures in accordance with the subject
invention comprises the steps of:
(a) extruding synthetic, polymeric Inaterial from
a fiberizing die and attenuating the e~truded poly-
25meric material to form synthetic, polymeric micro-
fibers by -the application of one or more gas streams
directed toward a rotating mandrel and a forming roll
in operative relationship with tlle mandrel;
(b) cooling the synthetic, polymeric microEibers
30prior to their collection on the mandrel to a temper-
ature below that at which they bond or fuse to each
other to substantially eliminate fiber-to-fiber bond-
ing; and
(c) collecting the cooled microfibers on the
35mandrel as a nonwoven, fibrous mass while applying a
64:~
force on the exterior surface of the collected micro-
fibers on the mandrel by the forming roll to form the
cylinclrical structure, wherein the process variables
are controlled to provide the collected fibrous mass
5 with a substantially constant voids volume over a
substantial part thereof, pre-Eerably at leas-t the
major portion, as measured in the radial direction.
It is preferred, especially or coarser fibers,
that cooling of the microfibers be enhanced by the
10 injection oE a cooling fluid into the stream of the
microfibers prior to their impingement on the mandrel
or the forming roll to assist in eliminating fiber-
to-fiber bonding.
~dditionally, it is preferred that the attenu-
- ]5 ated microfibers impinge on the forming roll which is
held at a temperature substantially belo~ the melting
or softening point of the fibers to further enhance
cooling prior to -the microfibers being transferred to
and collected on the rotating mandrel, thereby pro-
20 viding additional cooling and further reducing the
likelihood of undesirable fiber-to-fiber bonding.
Preferably, the stream of microfibers is directed
-toward the forming roll and mandrel in such a manner
that at least the major portion contact the forming
25 roll first (where they are cooled further) and from
where -they are then transferred to the rotating man-
drel. Also, if the forming roll is wet, particularly
when microfiber collection on the mandrel is initi-
ated, more consistent start-ups are obtained due to
30 better (more uniform) transfer of the microfibers to
the mandrel, i.e., the potential Eor undesirable
layer-to-la~er bonding is reduced and a smoother
wrapping with minimized clumping and a more regular
or uniform laydown of fibers is obtained.
In addition, the apparatus is preferably de-
-8- 1~ 3
signed so as to allow free access of secondary air in
order to assist in the rapid cooling of the hot,
freshly formed fibers.
Figure 1 is a perspective view of an apparatus
5 which can be used to form the c~lindrical filter
structures in accordance with this invention;
Figure 2 is a perspective view showing an an~
cil]ar~ collection means which can be used with the
apparatus oE Figure l;
Figure 3 is a partially cut away perspective
view of a cylindrical filter structure in accordance
with this invention;
Figure ~ is a yraph of resin (or polymeric mate-
rial) pressure versus fiber diameter;
]5 Figure 5 is a graph of fiberizing air pressure
versus fiber diameter;
Figure 6 is a graph of forming roll air pressure
versus fiber diameter;
Figure 7 is a graph of resin rate versus resin
20 pressure; and
Figure 8 is a graph of particle diameter for
which the removal rating equ~ls 99.9 percent versus
fiber diameter.
The subject invention will be better understood
25 by reference to the drawings. Turning first to Fig-
ure l, there is shown an apparatus useful for forming
the cylindrical filter structures in accordance with
the subject invention comprising a fiberizer or fi-
berizing die lO to which molten resin is delivered by
30 a motor-driven extruder 11 and to which hot com-
pressed gas, preferably air, is delivered Erom a
heater 12. The fiberizer lO contains a multiplicity
of individual extrusion nozzles 13 by which the mol-
ten resin is converted to fibers. In the preferred
35 mode illustra-ted in Figure 1, the hot resin ~or poly-
_9_ ~5~ 3
meric material) stream delivered from the extruder ]1to the fiberizer 10 issues from each nozzle under
pressure (fiberiziny air pressure). The molten,
thermoplastic pol~meric microfibers generally desig-
5 nated 1~ are formed as the resin is extruded from thenozzles 13 and attenuated b~ the jets of hot gas
referred to above which carry the microfibers upward
in the direction of a cylindrical Eorming roll 15
which is in operative, rotating relationship with the
10 power driven rotating, and preferably also recipro-
cating, man~rel 16. The Eorming roll 15 may be
cooled, e.g., by passing unheated ambient air through
i-ts internal portions. When a supported filter struc-
ture is being formed, prior to initiation of col-
]5 lection of the microfibers on the mandrel, one ormore open, relatively rigid central support members
or filter cores 17 ~such as that shown in detail in
Figure 3) are placed on the mandrel. Alternatively,
as discussed below, for some low pressure applica-
20 tions a central support member or core may not berequired, in which event the cylindrical filter struc-
ture can be formed directly on a solid mandrel. The
mandrel 16 and the filter cores 17 are designed such
that the cores ro-tate with the mandrel, either by
25 means of friction between the fil-ter cores and the
mandrel or by use of springs or other arrangemen-t.
The forming roll 15 is preferably mounted on
bearings so that it is freely rotatable, i.e., it
rotates freely when in contact with the mandrel 16 or
30 with fibrous material collected on the mandrel 16 or
with filter support cores 17 (as illustrated in Fig-
ure 3) which may be placed on the mandrel 16. Ad-
ditionally, the forming roll 15 is preferably biased,
for example, by an air cylinder 18, operating on the
35 shaft 9 which is rotatably mounted Oll bearings. The
--10-- ~ LP~t3
air cylinder 1~ through the shaft 9 applies bias to
the orming roll 15 in a controlled manner towards or
away from the mandrel 16. Depending on the fric-
tional characteristics of the air cylinder and the
5 charactex of the fibers being collected, damping of
the shaft 9 may be desirable to prevent vibration of
the forming roll 15.
The mandrel 16 is rotated by a motor (not shown),
generally at a rate oE from about 50 to about 500
10 rpm, and, in the preferred embodiment shown in Figure
1, is reciprocated axially at a rate generally be-
-tween about 10 (3.0 meters) and about 300 feet (91.4
meters) per minute. The length of stroke of the
reciprocating mandrel will depend on the desired
]5 length of the cylindrical filter structure or struc-
tures being formed.
Especially when making relatively coarse fibers,
a suspension of finely divided water droplets 19, or
other cooling fluid, is preferably injected into the
20 stream of fibers 14 from one or both sides by the
nozzles 20, impinging on the stream of fibers a short
distance above the extrusion nozzles 13, e.g., 1 to 5
inches (2.5 to 12.7 cm) to cool the microfibers and
help prevent fiber-to-fiber bonding.
In operation, the microfibers are projected
upward in the direction of the forming roll 15 and
the mandrel 16, generally at least in part impinging
on the forming roll 15, from where they are contin-
uously transferred to the filter cores 17 mounted on
30 the rotating, reciprocating mandrel 16. As the man-
drel 16 rotates and reciprocates, the diameter of the
cylindrical mass of flbers collected on the ilter
cores 17 increases.
It is generally preferred that at least the
35 major portion of the fibrous s-tream 1~ imp;n~e on the
5f~L~
forming roll 15 rather than on the mandrel 16 as this
results in a more uniform and more reproducible pro-
duct in whicll the fibers exhibit little or no unde-
sirable interfiber bonding, i.e., they are substan-
5 tially free of fiber-to-fiber bonding.
Under some conditions, particularly when col-
lecting fibers less than about 1.8 to 2 micrometers
in diameter, an auxiliary collection member 22, as is
shown in Figure 2, may be used to advantage. This
10 member can be a flat, stationary - relative to the
forming roll - sheet or plate. Alternatively, it may
have a moderate radius with the concave side facing
downwards toward the fibrous stream 14. It is pre-
ferably mounted such that one edge is about 0.1 inch
]5 (0.25 cm) or less from the forming roll surface. The
collection member 22 is secured by brackets 23 to the
frame 24 supporting the forming roll 15. The func-
tion of member 22 is to collect fine fibers which
would otherwise bypass the forming roll. As rapidly
20 as these fibers collect on member 22, they are trans-
ferred to the forming roll 15 and thence to the ro-
tating reciprocating mandrel 16.
The system used for fiberizing the resin or
polymeric material can take a varie-ty of forms, many
25 of which have been set forth in the patent and jour-
nal literature. See, for example, the paper titled
"Superfine Thermoplastic Fibers" in the ~ugust 1956,
Volume 48, Number 8, edition of Industrial and ~n-
gineeriny Chemistry. The resin stream or streams can
30 be discontinuous ~i.e., delivered by individual noz-
zles) or continuous (i.e., deliverecl through a slot),
and the air stream or streams can be similarly con-
tinuous or discontinuous. ~dditionally, combination
of these design configurations can be used, e.g.,
35 when preparing a filter element Erom two or more
6~1~
di~ferent polymeric materials.
Also, a number of process variables can be con-
trolled to provide any desired combination of fiber
diameter and voids volumes within the limits of the
5 apparatus. As will be evident from consideration of
the Examples below, it is preferred that four var-
iables be used in operating the apparatus of Figure
1. These are:
(1) The Rate Of Delivery Of Resin
(Or Polymeric Material) To
The Fiberizing Die:
This xate is adjusted by increasing or de-
creasing the pressure developea by -the extruder,
which in turn is accomplished b~ changing its speed.
]5 As the rate is increased, coarser fiber is generated
and voids volume of the collected fiber cylinder
tends to decrease.
(2) The Fiberizing Gas Flow ~ate:
This rate is adjusted by altering the pres-
20 sure at which the gas, typically air, is delivered to
the fiberizing die. As the flow rate of the gas
stream (or streams) is increased, the fiber diameter
becomes smaller and the voids volume tends to in-
crease.
(3) The Forming Roll Pressure:
The forming roll pressure is varied as re-
quired to maintain the voids volume constant. For
example, if the fiberizing gas rate is decreased to
increase the diameter of the microfibers, the forming
30 roll pressure must be decreased to maintain a con-
stant voids volume.
(~) The Quantity And Type Of Fiber Cooling:
These include the quantity of secondary air,
- die to collector distance (see below), the tempera-
35 ture of the forming roll on which it is preferred
-13- 1~5~
that the fibers impinge, the quantity and mode of
delivery of liquid coolant, and the rate of rotation
and reciprocation of the mandrel. When fiber dia--
meter is smaller than about 3 to 6 micrometers, wa-
5 ter cooling is not required, although it can be used.The effect of these various cooling means on the
density of the collected fiber varies and must be
determined empirically.
Other process variables which influence the
10 character oE the formed filter cylinder but which
once set - in the preferred mode of opera-tion - no
longer need be altered include:
(1) The fiberizing die to collec-tor distance
(DCD), if too large, permits the fiber to form bun-
]5 dles prior to deposition on the forming roll (a phen-
omenon known as "roping"), causing the formation of a
non-uniform product. If DCD is too small, the fibers
may be insufficiently cooled when collected and this
may result in melting or softening, which tends to
20 close off pores and obstruct free flow of fluids
through the fibrous mass when used as a filtering
device.
The optimum DCD depends upon the diameter and
velocity of the fibers and upon the rapidity with
25 which they are cooled and is best determined by trial
ar;d error.
DCD can be used as a controlling variable and,
indeed, was so used during the early phases of the
development of this invention but was discontinued
30 because it proved easier and adequate to vary the
four variables listed above.
~ 2) The temperature of the resin or polymeric
material supplied to the fiberizing die has a strong
effect on product characteristics. As this tempera-
35 ture is increased, fiber diame-ter decreases, but
-14-
~ 6~3
excessive temperatures cause the production of very
short fibers and shot, as well as significant reduc-
tion in resin molecular weight due to depolymeriza-
tion. 5'he optimum temperature is best determined by
5 trial and error since it depends on a number of fac-
tors, including the particular polymeric material,
the nature of the structure desired and the par-
ticulars of a given apparatus, for example, the ex-
truder size as related to the resin flow rate.
(3) The temperature of the fiberizing air has a
relatively minor effect, provided that it is held
within about 50 degrees F' (2~ degrees C) of the resin
temperature.
(~) The temperature of the forming roll is
]5 preferably low, for example, near ambient, to help
prevent interfiber fusing of the fibers collected on
it prior to transfer to the rotating reciprocating
mandrel.
(S) The rate of rotation of the mandrel; higher
20 rotation rates help to prevent interfiber bonding.
(6) The rate of reciprocation of the mandrel;
higher reciprocation (or axial translation) rates
help to prevent interfiber bonding.
sy the method in accordance with this invention,
25 the fiber diameter of the cylindrical fibrous struc-
tures can be varied in a continuous or step-wise
rnanner from one part of the cylindrical structure of
the fibrous mass to another - as measured in the
radial direction - by varying the resin and fiber-
30 izing air flow rates while the voids volume is main-
tained substantially constant by varying the forming
roll bias force on the cylindrical mass of fibers as
the structure is formed on the rotating mandrel. As
may be seen in Figure 8, if the voids volume is con-
35 stant, the pore size varies with the fiber diameter.
~15- ~2~8~
By the method in accordance with this invention, the
pore diameter can be varied continuously or stepwise
from one part of the filter to another in any desired
manner.
When the desired outside diameter of the cylin-
drical fibrous structure has been reached, the opera-
tion is terminated by discontinuing the flow of resin
and air onto the forming roll 15, discontinuing or
reversing the bias oE the forming roll 15 and stop-
10 ping the mandrel 16, following which the formed cyl-
indrical fiber structure together with the core or
cores is removed from the mandrel 16. The ends of
the resulting cylindrical structure are then cut to
length and if more than one core 17 has been used,
]5 additional cuts are made to separate each section,
thereby forming individual cylindrical filter struc-
tures, sometimes referred to herein as filter cylin-
ders, filter elements or simply as elements. ~ cy-
lindrical filter structure in accordance with this
20 invention is illustrated in Figure ~. The cylin-
drical filter structure generally designated 30 is
comprised of the hollow support core 17 and a fibrous
mass of nonwoven, synthetic, polymeric microfibers
31.
~s noted above, for some applications it may be
desirable to Eorm the cylindrical fibrous structures
in accordance with this invention directly on the
mandrel without -the use of an internal support or
core. For most purposes, however, it is desired that
30 the structure, when used as a filter, be able to
withstand, without collapse or loss of integrity,
diEferential pressures of 40 psi (2.~1 ~g/cm2) or
higher. The voids volumes of the unbonded fibrous
mass of the filter structures in accordance with this
35 invention which yield desirable combinations of high
-16- ~5~3
efficiency and long life in service are, in general,
too higll to withstana pressures of this magnitude and
would collapse if an internal support member were not
provided. Accordingly, for most applications, it is
5 desirable to form the filter on a hollow foraminous,
or open, relatively rigid central support member or
core desi~ned in such a way as to provide support for
the collected fiber or fibrous mass. The central
support mernber or core 17 must be open or foraminous
10 in nature, as illustrated in the perspective view of
a typical supported cylindrical filter structure in
Figure 3, since it must provide adequate passages for
flow of filtered fluid into the central portion of
the core (outside/in filter configuration) or, con-
]5 versely, passage of fluid to be filtered from thehollow cen-ter of the filter structure into the fib-
rous mass (inside/out configuration). Typically, the
core, which is relatively rigid vis-a-vis the mass of
collected fibers on the exterior thereof in order to
20 provide the requisite support, will have openings 32
with spans preferably on the order of one-quarter
inch (0.6 cm) or less and, generally, not more than
one-half inch (1.3 cm).
The central support member or core can be made
25 by a variety of processes and from a variety of mat-
erials, for example, from synthetic resin by injec-
tion molding or extrusion, or from metal by conven-
tional processes. While not required, the core may
have a multiplicity of small protuberances on its
30 exterior to assist in securing the microfibers to the
exterior of the core.
Another alternative is to build up a support
core of self-bonded fibers on the mandrel by operat-
ing under conditions such that fiber-to-fiber bonding
35 occurs during the first part of the formation of the
-17- ~ ~5~
~ibrous structure, e.g., by minimizing the type and
quantity of fiber cooling, following which the method
in accordance with this invention is carried out
under conditions such that fiber-to-fiber bonding is
5 substantially eliminated. The resulting structure
h~s the internal support necessary to prevent col-
lapse of the element under conventional operating
pressures and has the added benefit that the portion
of the structure which is self-bonded (the central
10 support member) has some filtering capability.
For very low pressure service, for example, in
the rarlge of about 5 to about 25 psi (0.35 to 1.76
kg/cm2), the cylindrical depth filters in accordance
with this invention can be made directly on a smooth
]5 mandrel and used without a core. It is, of course,
also possible to make a coreless cylindrical filter
structure and subsequently incorporate a core or
central support member therein.
The preferred fibrous structures prepared by the
20 method in accordance with the subject invention are
comprised of a fibrous mass of nonwoven, synthetic,
polymeric microfibers which are substantially free of
fiber-to-fiber bonding, secured to each other by
mechanical entanglement or intertwining, and wherein
25 the fibrous mass has a substantially constant voids
volume, typically in the range of from about 60 to
about 95 percent, more preferably from about 64 to
about 93 percent and even more preferably from about
75 to about 85 percent. When polypropylene is used
30 as the resin, the most preferred voids volume is
about 82 percent. Typically, the annular thickness
of the cylindrical fibrous structures in accordance
with this invention, particularly when used as depth
filters, is in the range of from 0.4 to 1 inches (1.0
- 35 to 2.5 cm), preferably in the range of 0.5 to 0.8
.25~
inches (1.3 to 2.0 cm), and more preEerably in the
r~nge of 0.6 to 0.7 inches (1.5 to 1.8 cm). ~s will
become more evident from the following Examples, the
combination of these characteristics in the cylin-
5 drical filter structures in accordance with thisinvention result in high filter efficiency and en-
hanced dirt capacity or life.
Polyrneric materials particularly well suited for
use in accordance with this invention are thermo-
10 plastics such as the polyolefins, particularly poly-
propylene and polymethylpentene, polyamides, par-
ticularly nylon 6, nylon 610, nylon 10, nylon ll,
nylon 12, and polyesters, particularl~ polybutylene
tereph-thalate and polyethylene terephthalate. Other
]5 suitable, but less preferxed, polymers are addition
polymers such as polyvinyl fluoride, polyvinylidene
fluoride and their copolymers, and polycarbonates.
The method in accordance with this invention can
also be applied to solutions of resins in appropriate
20 solvents, in which case temperatures can vary down to
ambient or lower. In this mode the solvent must be
at least largely evaporated before the fibers are
collected to avoid fiber-to-fiber bonding.
Thermoset resins in partially polymeri~ed form
25 can be fiberized but are not a preferred starting
material as operation with them is more comple~.
The fiber diameters can be varied from about 1.5
micrometers or less up to about 20 micrometers or
more. However, when the product is made in the pre-
30 ferred voids volume range of 75 to ~5 percent, fiberdiameters above about 20 micrometers make elements so
coarse as to have little use for filtration applica-
tions.
Fiber aspect ratios are large. e.g., l,000 or
35 hlgher. Indeed, it is very difficult even by micro-
-19- ~5~6~
:`
- scopic examination, to determine length to diameter
ratios as fiber ends are difficult to ind.
: Various additives, such as activated carbon, ion
exchange resins, and the like, can be incorporated
5into the cylindrical fibxous structures in accordance
with this invention by, for example, feeding them
; into the stream o fibers prior to laydown. ~lso,
the cylindrical fibrous structures in accordance with
this invention can be formed in any desired length.
10 The cylindrical fibrous structures can be further
processed byr for example, the application of an
external support and the incorporation oE end caps
shaped so as to fit within the particular filter
~' assembly in which the resulting filter element is to
]5 be used.
The term "substantially free of fiber-to-fiber
bonding" r as used hereinr refers to the characteris-
tics of the microfibers ma~ing up the fibrous mass
portion of the cylindrical fibrous structures in
~ 20 accordance with this invention. The microfibers are
!' mechanically entangled or intertwined. It is this
mechanical entanglement which provides the structural
integrity oE the fibrous mass portion of the struc-
ture. Whrn examined under a microscope at lOx to
25 lOOx the fibrous portion of the filter structure may
display random fiber-to-fiber bonding but such bond-
ing is in an amount that would not be significantly
detrimental to filter Eunc-tion nor contribute in any
material way to the structural integrity of the fil-
30 ter. Additionally, it is possible, by the use oftweezersr to separate out fibers which have clean,
smooth proEiles, free of protuberances and oE unsep-
arable clumps of fibers of the type which typically
appear on fibers in structures containing substantial
35 fiber-to-fiber bonding.
-20~ ~2 5~L~3
The term "substantially constant voids volume",
as u~l herein, means the average voids volume o the
fibrous mass portion of the cylindrical filter struc-
ture varies by no more than about 1 to 2 percent.
5 Voids volume determinations or, alternatively, den-
sities, were carried out by use of a series of 5 U-
sh~ped gauges. Diameters oE the gauges were selected
such that the difference between each successive
gauge represented one-fifth of the total volume of
10 the collected fiber on the cylindrical filter struc-
ture as it was being formed. With the polymeric
material or resin delivered at a constant rate, the
time required to reach the diameter of each gauge was
recorded. This procedure was repeated as ten suc-
]5 cessive filter cylinders were prepared under the same
conditions and the times averaged. The percent voids
- volume determined by this procedure was found accur-
ate to within about 2 percent, and in the case of
finer fibers, to within about 1 percent.
As may be seen in Figure 8 and in Table II be-
low, filter elements each having a constant voids
volume of 82 percent and a constant fiber diameter
throughout (but varying from element to element rom
1.9 up to 12.6 mlcrometers) provided removal ratings
25 varying from less than 1 micrometer, e.g., 0.5 micro-
meters or even less r up to over 40 micrometers.
One configuration which is useful because it
provides prefiltration for a very wide range of final
filters is made using a program for forming roll
30 pressure, resin rate, fiberizing air rate, and cool-
ing water Elow, which produces a constant density
element with fibers varying in diameter from about
1.9 micrometer at the id (downstream) to about 12.6
micrometers at the od (upstream). The manner in
35 which the fiber diameters are profiled can be varied
ii, ; ~
. .
-21~ 3
widel~, for e~ample, for some applications a higher
proportion o~ Eine ~ibers could be used while for
others more coarse fibers might be preEerred. For
general prefilter service~ filter elements have beer-
5 made in which the diameters of the fibers ~orm ageometric progression. In such a construction, if
the element is divided into N cylindrical portions,
each containing the same weight (and volume) of fi-
bers, then the fiber diameter of each portion is
10 larger than that of the adjacent downstream portion
by the factor F, where
F = (12.6) N-l
1 . 9
]5 For example, if the number of sections were 20 (N =
20), F would be 1.105.
Another configuration, which is desirable be-
cause it combines absolute filtration with prefiltra-
tion~ is one in which the downstream portion of the
20 filter is made using constant fiber diameter, while
the upstream portion is profiled from the fiber dia-
meter of the downstream portion up to a larger dia-
meter. The constant fiber diameter downstream por-
tion of the filter element may comprise ~rom about 20
25 to about 80 percen-t of the total volume of the fib-
rous filter mass and, correspondingly, the upstream
profiled portion of -the filter element may constitute
from about 80 to about 20 volume percent. One pre-
ferred configuration is a filter element in which
30 about the first 50 percent by volume of the element
has a constant fiber diameter of 1.9 microme-ters
(downstream portion) and the upstream portion has a
graded fiber diameter structure, i.e., it is pro-
filed, with the fiber diameter ranging from 1.9 mi-
35 crometers up to 8 to 12 misrometers. Another con-
-22- ~ S~
: figuration is as that immediately above but the con-
stant Eiher diameter portion has fibers with diameters
of 8 micrometers an~l in the graded fiber diameter
portion they range f~om 8 up to 12 to 16 mic~ometers.
Still another desirable configuration is one in
which the downstream portion of the filter is made
using a constant voids volume, with a constant fiber
diameter, and the upstream portion has both a pro-
filed Eiber diameter and a profiled voids volume.
Filter elements have been made using fibers as
small as 1.6 micrometers in diameter. Still finer
fibers could be used but are not preferred because
production rates become progressively lower and col-
lection oE the fibers becomes more diEficult, with a
; ]5 larger proportion not collected on any o~ the working
parts of the apparatus. Other elernents have been
made with fibers as coarse as 16 micrometers but such
elements have removal ratings so large as to have
limited practical application or, if made quite dense,
20 relatively low dirt capacity. Filter elements made
with their downstream portions composed of fibers of
13 micrometers or smaller can, however, for some
applications, beneEit from upstream layers profiled
up to 16 micrometers or even higher.
Filter elements in accordance with this inven-
tion have been made using polyprop~lene resin with
voids volumes varying from 64 to 93 percent. Voids
volumes above about 85 to 88 percent are not pre-
ferred because they are deformed by relatively low
30 differential pressure, for example, as low as 5 to 10
p~si (0.35 to 0.7 kg/cm2), with consequent change of
pore diameter. Voids volumes below about 75 percent
are not generally preferred for use in the operating
range of most filters, which is up to about 40 to 60
35 psi (2.81 to 4.22 kg/cm2) differential pressure,
-23~ 3
because filter life decreases as the voids volume is
decreased. One exception is the use of lower voids
volumes w}len making filters with fiber diameters at
the low end of the practical fiber diameter range,
5 for example, 1~6 to 2.0 micrometers. Such filters
remove finer particles than would be obtained using
higher voids volume and are useful for that reason.
For filter operations with filter elements made from
fibers having diameters from above about 1.6 to 2.0
10 micrometers and opera~ing at differential pressures
of up to a~ou~ 40 to 60 psi (2.81 to 4.22 Xg/cm2),
the preferred voids volume range is 78 to 85 percent.
For applications in which differential pressures
exceed 60 psi (~.22 kg/cm2), and up to several hun-
]5 dred psi (14 ~g/cm2 or higher), lower voids volumesdown to 60 percent or even less may be needed to
prevent collapse under pressure. Somewhat higher
voids volumes can be used with filter elements pre-
pared from resin materials of relatively higher mod-
20 ulus, such as nylon 6, which have better resistanceto deformation compaxed with polypropylene.
EEficiency, Removal Rating ~nd Dirt Capacity (Life):
These characteristics were determined for 2.5 od
x 1.1 id x 10 inch (6.35 x 2.8 x 25.4 cm) long ele-
ments using a modified version of the F2 test devel-
oped in the 1970s at Oklahoma State University. In
this test a suspension of an artificial contaminant
30 in an appropriate test fluid is passed through the
test filter while continuously sampling the fluid
upstream and downstream of the filtex under test.
The samples are analyzed by automatic particle count-
ers for their contents of five or more different
35 preselected particle diameters and the ratio of the
-2~
upstream count to downstream count is automatically
recorded. T~lis ratio, known in the industry as the
beta (~) ratio, provides the xemoval efficiency at
each oE the preselected particle diameters.
The beta ratio for each oE the five or more dia-
meters tested is plotted as the ordinate against
particle diameter as the abscissa, usually on a graph
in which the ordinate is a logarithmic scale and the
abscissa is a log2 scale. A smooth curve is then
10 drawn between the points. The beta ratio for any
diameter within the range tested can then be read
from this curve. ~fficiency at a particular particle
diameter is calculated from the beta ratio by the
formula:
]5
Efficiency, percent = 100 (1 -l/beta).
As an example, if beta = 1000, efficiency = 99.9
percent.
~nless otherwise stated, the removal ratings
cited in the examples presented below are the parti-
cle diameters at which beta equals 1,000 and the
efficiency is 99.9 percent.
E-fficiencies in the range of from 1 to about 20
25 to 25 micrometers were determined using as the test
contaminant a suspension of AC fine test dust, a
natural silicious dust supplied by the AC Spark Plug
Company. Prior to use, a suspension of the dust in
water was mixed until the dispersion was stable.
30 Test flow rate was 10 liters/minute of the aqueous
suspension. This same procedure was applied to fil-
ters having efficiencies oE less than 1 micrometer by
determining eEficiencies at usually 1, 1.2, 1.5, 2,
2.5 and 3 micrometers and extrapolating the data to
35 under 1 micrometer.
-25~ 86~3
~ fficiencies above about 20 micrometers were
determined using Potter's Industries Incorporated
fl3000 spherical glass beads suspended in MIL-H-5606
hydraulic fluid. These glass beads have a dis-tribu-
5 tion of si~es ranging from less than 15 micrometersup to 50 ~o 55 micrometers and higher. The viscosity
of this fluid is approximately 12 centipoise at the
test temperature of 100 degrees F (37.8 degrees C).
Test flow rate was 20 liters per minute. The higher
10 viscosity and flow rate serve to keep beads up to
about 100 micrometers in diameter in suspension.
Filters in the 20 to 25 micrometer range were
often tested by both methods. The resulting effici-
ency and dirt capacity data were usually comparable.
]5 In both the a~ueous and oil based tests, pres-
sure drop across the test filters was measured as the
test suspension Elowed through the filter and was
recorded as a function of time. The quantity of
contaminant incident on the filter required to devel-
20 op a differential pressure of 60 psi (4.2 kg~cm2) is
recorded as the dirt capacity or "life" of the test
element.
It is characteristic of depth filters, particu-
larly in -the coarser grades, that efficiency tends to
25 be reduced at large differential pressure. Since
filters are rarely exposed to differential pressures
as high as 60 psi (4.2 kg/cm2), efficiency data are
reported as an average of about the initial two-
thirds of the total life of the filter.
As noted above, data reported as less than 1
micrometer are obtained by extrapolation. In order
to provide assurance that the extrapolated data were
reasonably near to correct, or at least conservative,
a number of the filter elements with high efficien-
35 cies at under 1 micrometer were further tested by
, -26- 1~5~
passing suspensions of bacteria of known dimensions
thrQug}l them. The upstream and downstream bacteria
concentrations were used to calculate eEficiency. In
all cases the efficiencies so deterrnined either con-
; 5 firmed the extrapolated ~2 test data or indicated a
still higher efEiciency.
: The test in the finer ranges using the ~C dust
described above showed significant and reproducible
beta ratios as high as 100,000 to 1,000,000 and
10 there~ore permitted measurement of efficiencies of up
; to and over 99.999 percent while the smaller number
of glass beads permi-tted computation of efficiencies
up to about 99.99 percent at up to about 40 micro-
meters and to successively lower efficiencies at
]5 larger diameters.
Filtration Testing Using Bacteria:
Filtration of suspensions of bacteria of known
j20 size is a very useEul high sensitivity method for
determining Eilter efficiency. This test method is
:particularly appropriate for application to filters
made using finer fibers and moderate to high density
because bacteria removal is one of the important
25 prospective applications for the finer grades of
filters in accordance with this invention.
Bacteria removal tests were run in the following
manner:
(a) A suspension in water of a pure strain of a
30 bacterium of known dimensions was prepared at a con-
centration of about 101 to 5 x 1012 organisms per
liter.
(b) The filter element was placed in an appro-
priate housing and 1 liter of the bacteria suspension
35 passed through the element a-t a rate of 0.5 to 1
' ~
'
~ - ;:. .
-~7- ~ 6~
litex per minute.
(c) ~liquots of the effluent from the filter
were collected and diluted with sterile water to 10,
100, 1,000, etcetera, fold. Each such diluted ali-
5 quot was then cultured in a Petri dish in an appro-
priate growth medium. ~ach bacterium present de-
veloped, within 24 to 48 hours, into a colony of
bacteria large enough to be seen at low magnification
using a microscope. The number of colonies in some
10 dilutions was so great that the colonies could not be
counted, while in others there were too few to be
statistically significant. Ilowever, there was always
t at least one dilution providing a useful count, from
which the total number of bacteria in the effluent
. ]5 could be calculated. Knowing the influent count and
the effluent count, efficiency can be calculated.
The bacteria used in developing this invention
included Pseudomonas diminuta (Ps.d) and Serratia
marcescens (Serr. m.), the dimensions of which are
20 respectively 0.3 micrometer diameter x 0.6-0.8 mi-
; crometer long and 0.5 micrometer diameter x 0~8
micrometer long.
The in-vention will be better understood by re-
ference to -the following Examples, which are offered
25 by way of illustration.
- ,
. ,
-~B~ 3
~x~mple 1: Preparation Of ~ Cylindrical
Filter Struct~re Of Uniform ~iber Size
And Uniform Voids VQlume (Ungraded):
_ _ _
The apparatus described above was used to pre-
pare a supported cylindrical filter structure with a
us~ble central section 36 inches long (91.4 cm long~.
The fiberi~ing die length was 6-1/4 inches (15.9 cm),
the str~ke of the reciprocating mandrel was 43-3/4
10 inches (111.1 cm), the mandrel rotation rate was 150
rpm, the axial translation rate was 500 inches (1270
cm) per minute, and the die to collector distance
(DCD) was 12-1/4 inches (31.1 cm). The mandrel was
fitted with three hollow foraminous (latticed) filter
cores, each 1.1 inches (2~.8 cm) inside diameter ~id),
1.3 inches (3.3 cm) outside diameter (OD) by 9.B
inches (24.9 cm) long, of the type illustrated in
Figure 3. Polypropylene resin having a melt flow
index of 30 to 35, was heated to 720 degrees F ~382
degrees C) and the extruder rpm adjusted so as to
give a total resin flow rate of 1.83 grams per second
through the spaced nozzles, each having a gas stream
surrounding the resin extrusion capillary, at a resin
pressure of 625 psi (43.9 kilograms per cm2). (Poly-
propylene resin having a melt flow index of 30 to 35was the resin used in all the Examples herein ~nless
otherwise noted.) ~ fiberizing air pressure of 4 psi
(0.28 kilograms per cm2) was used. The average fiber
diameter produced under these conditions had pre-
viously been determined to be 12.5 micrometers. Themicrofibers having 12.5 micrometer diameters so pro-
duced were directed onto the air cooled Eorming roll
which was biased towards the mandrel by an air cylin-
der pressurized to 8 psi (0.56 ky/cm2) and were thence
transferred to the filter cores on the rotating/
_~9_ ~ ~ 5~
reciprocating mandrel. Fiberizing ~nd collection
were continued until the od of the fibrous cylindri-
cal filter structure (sometimes referred to herein as
a '`filter cylindern~ a "filter element" or simply as
an "element~J reached 2.5 inches (6.35 cm)~
The average density of the fibrous porti~n of
the cylindrical filter structure of this Example was
such as to yield a voids volume of about 81 percent
(voids volume, in percent, equals 100(1-D/d), where D
equals the apparent density and d equals the density
of the resin, which is 0.9 grams per cubic centimeter
for the polypropylene used).
The central section of the cylindrical filter
structure was cut into three sections to make three
filter cylinders, each 9 ~ inches (24.9 cm) long and
each having a corresponding 9.8 inches ~24~9 cm) long
filter core on the interior thereof. The respective
voids volumes of the three filter cylinders were
equal within the measurement error; each had a voids
volume o~ 81.2 percent.
The three filter cylinders denoted A through C
below were assembled into housings which provided
appropriate end sealing means and were tested using
the F2 test method described above, yielding the
results listed below:
Filter Life or Removal Rating,
Dirt Capacity (micrometersj
(qrams)
A 82 40
B 81 41
C 82 39
-30- 1~ 5~
Microscopic examination of the elements was
performed. With the exception of a limited number of
small l~calized areas in which some undesirable fiber
softening had occurred, the individual fibers could
be pulled out of the mass using tweezers with no
evidence of aahesion to neighboring fibers, i.e., the
pr~files ~f the fibers were smooth with no protuber-
ances inaicating fiber-to-fiber bonding.
' It should be noted that in this Example, as in
the following Examples where fibers of 2.5 micrometers
or larger were formed, water spray was used to pro-
vide enhanced cooling of the fibers, thereby assist-
ing in minimizing undesirable fiber-to-fiber bonding.
The water spray was appl-ied in the general manner
illustrated in Figure 1 at application rates suffi-
ciently high, in connect~on with other cooling tech-
niques as descxibed, to provide structures substanti
ally free of fiber-to-fiber bonding, e.~., ;n the
range of about 80 to 140 cubic centimeters per min-
ute.
Example 2: Distribution Of Vo;dsVolume Within Cylindrical Filter Structures
Made Using Constant Forminq Roll Pressure:
Using the apparatus described above and the
general procedure described in Example 1, a series of
elements, each having uniform fiber size~ were pre-
pared directly on a 1.3 inch ~3.3 cm) od solid collec-
tion mandrel. That is, the elements of this Examplediffered from those prepared in Example 1 in that
they did not contain a central support member or
core. The elements, denoted in Table I below as
filters D through H, each had constant fiber dia-
meters but from filter to filter the fiber diameter
L 9~ 3~
~31-
varied from 12.S micrometers down to 2 5 micrometers,
as set out in Table I. A series of five ~U-shapedN
gauges were prepared with the d.iameter of the first
gauge such that t}-e volume of fiber collected re-
presented one-fifth of the total volume of fiber
collected between 1.3 inches (3.~ cm), the id of the
formed filter cylinder, and the 2.5 inch ~6.35 cm) od
of the finished filter cylinder or element. Similar-
ly, the difEerence in diameter between the second and
first gauges - the second having a larger diameter
than the first - represented one-fifth of the volume
of fiber collected between 1.3 inches (3.3 cm) and
2.5 inches (6.35 cm). In li~e manner, the difference
in diameter between the.second and third gauges re-
15 presented one-fifth of the volume of fiber collected
between 1.3 inches (3 3 ~m) and 2.5 inches ~6.35 cm),
etcetera, up through the fifth gauge. Resin flow
rate was held constant and, as the filter cylinder
diameter increased or built up during formation of
20 each of the elements D through H, the time required
to reach the diameter of each of the five gauges was
recorded. These times were then used to determine
the percent voids volume of each of the five sections
of each of the elements, with the results shown in
25 Table I bel~w:
. . _
,' ' ' ~.
- 3 2~ , t~
h o o r~) t`J 0
.
~J O r~
O ~ (~ CO Cl:l ~0
O
~ ~ C~
CO CO
. ~ . . .
C r~ O _1
O ~ ~
_~ . .,
Vv
O t` In O
C ~ ~
~ h
V C
aJ c ~ ~r ~ ~r
~ ~ CO ~
a)
P~ J
a
U~
E i~
~ al u~ O co u- u
,1 V
a
h t) _~
h
E
a
~
3 ~ 5~t~jL~ ~
~ he measurement of voids volume within the
fibrous mass (or, in effect, the average density,
since the density of the fibers is a constant, i e.,
0.9 qrams per cc) by the method described above is
not precise. It i5 believed that the voids volume in
the ~ilter elements is uniform or nearly so through-
out the thickness of the filter cylinder ard that any
errors caused by using the averages of these voids
volumes is small. Thus, in those Examples reported
which are made using a range of fiber diameters on a
single element by adjusting the orming roll pres-
sure, the voids volume is believed to be constant
throughout the fibrous mass within about l to 2 per-
cent.
Examples 3 through 12 Preparation Of A
Constant Voids Volume Cylindrical Filter
Structure Profiled To Provide A Wide Range
f Pore Diameters By Varying_Fiber_Diameter:
Examples 3 through 9 below demonstrate the
preparation of constant or near constant voids volume
filters with removal ratings varying from less than 1
micrometer up to 40 micrometers. Examples lO, 11 and
12 below show how the data generated in Examples 3
through 9 can be used to prepare graded fiber dia-
meter, constant voids volume filter elements.
Step 1:
A series of supported cylindrical filter struc-
tures or elements (Examples 3-9), each with constant
or near constant voids volume of 82 + 1 percent and a
uniform fiber diameter, was prepared. While the
fiber diameter within an individual element was con-
-3~
stant over the serle~ oE seven element~ formed, the
fiber diameters ran~ed from 1.9 to 12.6 micrometers
as set out in Table II below. As also set out in
Table II below, the filter elements provided removal
ratings in the range o~ from less than 1 up to 40
micrometers.
These ~lements were prepared using the general
procedure and apparatus of Example 1 but the resin
pressure, fiberizing air pressure and the forming
roll air pressure were varied in order to obtain
fiber diameters spanning the range from 1.9 to 12.6
micrometers as noted in Table II. The average voids
volume of each test element was held as closely as
possible to 82 percent, the average deviation being
less than 0.4 percent. Conditions were controlled to
substantially eliminate fiber-to-fiber bonding in the
formed elements by the methods described above; most
importantly, the fibers were collected on the forming
roll rather than the mandrel, and water spray was
used when iber diame~er was 2.5 micrometers or
greater. Each element was ~2 tested and the removal
rating (diameter of particles in the incident fluid
at which the removal efficiency equalled 99.9 per-
cent) and the dirt capacit~ (or life) were deter-
mined. The preparation conditions and the test re-
sults obtained are shown in Table II below. Fi~ures
4, 5, 6, 7 and 8 graphically show the relationship of
the important parameters o~ Table II.
~35--
U) '
V ~ ~ CO ~ o ~ o
a c~ ~ o ~1 ~ ~D 0 0 o ~
~ ~ I U~
:~ ~ O ~J
o a) ~ ~ ~ r` o o o
~; E E v ~ ~--I 0 o
I U~
O -
v o Va~ ~ ~ o ~0
a E E E
Cll .
C C
v~ a,
~ O ~ O ~ ~ ~ ~ d'
V C ~ ~
E O (L~ 4 al ~ 1 co o ~ o
,- ~ ~ ~, 1 ~
E _ _ _ ._ _ _ _
E ~ Q~ ~ ~ ~ ~ o~
w a
Q~ , . u~ ~ o u~ In ~n In u~
.rl V ~ D~ ~ O O ~ N N N N
O O r:E3 ~n ~ 0 ~n
J- C~ X _ _ ~ ~ ~ o o
4-~
. . o o o o ~o ~ In
.~
~N
O E ~ ~ ~ ~ 'n o ~D
aJ ~ ~n ~ ~ ~r o r~ m
D,-- ~ ~ ~ _l ~ o o
O ~ u~ o In O O U~ O a:
~ ~ ~ u~
a)
R. QJ
E~
r~ E ~ ~ n ~ 0 O~
WZ
-36
_t~e_~:
Figures 4 through 7 can be used used to prepare
an operating plan which will make ilter ~lements
wit~ any combin~tion of fiber diameters between 1 9
and 12.6 micrometers.
In general, it i~ preferred to construct ele-
ments in which the liquid beinq filtered will flow
from the outside of the elements toward the inside
and then exit through the filter core (an outside/in
configuration). However, in some circumstances, for
example, when it is desired to retain the collected
solids within the ~ilter cartridge, the direction can
be reversed (inside/out configuration). In either
case, it is generally advantageous to have the pores
profiled from large at the upstream side to small at
the downstream side by providing fibers of aecreasing
diameter in a graded or profiled manner in the direc-
tion of fluid flow, i.e., in the radial direction,
while maintaining th~ voids volume substantially
constant. ~
The configuration o~ the profile can vary wide-
ly. In some applications ~t ma~ be desirable to have
the upstream portion of the filter graded with the
downstream portion of uniform pore size. Alterna-
tively, especially if intended for use as a prefil-
ter, the entire thickness of the fibrous portion of
the filter can be varied in an appropriate profile
with the largest pores upstream to the smallest pores
downstream~ Example 10 illustrates a filter of the
latter type in which progressively larger fiber dia-
meters are used as the filter is built up. In Ex-
ample 10, the fiber diameters are varied as a ~eo
metric progression. Varying the fiber diameters
as a geometric progression i5 believed to provide a
37 ~ 5 ~ L~
Eilter;ng element well adapted to a wide variety of
non-speci~ic applications. For any specific applica-
tions, other schemes can be used, for example, linear,
square root, or logarithmic, etcetera. Alternatively,
5 the fi~er diameters can be graded in a continuous
manner without discrete steps in the radial direction,
a form of gradation referred to herein as "continuous-
ly profiled".
Example 10: Element Made With Constant Voids
Volume And Varyinq Fiber Diameter Throuqhout
Using the general procedure and apparatus of
Example 1, the data o~ Figures ~ through 7 (generated
in Examples 3 through 9) were used as follows:
(a) the total volume of the fibrous portion of
the filter element to be formed (2.5 inches (6.~5 cm)
od x l.30 inches (3.30 cm) ia x 9 . 8 inches ~24.9 cm)
long) was divided into 15 equal incremental volumes;
(b) the fiber ~iameter range from l.9 to 12~6
micrometers was then divide~ into 14 steps of in-
creasing fiber diameter, each fiber diameter being
14.447 percent larger than the preceding one, the
first being 1.9 micrometers and the last 12.6 micro-
meters (this forming a geometric progression of fi~er
diameters as set out in Table III below);
(c) the operating conditions required to ob-
tain the 15 flber diameters set out in Table III
below were then read from Figures 4, 5 and 6.
In this Example, the filter is designed to have
an equal incremental volume of the fibrous mass at
each of the selected fiber diameters~ Because the
voids volume is constant within experimental error at
- 82 percent and the density is therefore correspondingly
constant, it was required that an equal weight of the
-3~-
- ~S~
microf;bers be deposited in each of the 15 incrernental
volumes. Since the resin x~t~ i~ a function of the
fiber di~meter, Figure 7 was used to calculate the
time required to deposit an equal weight in each of
the increment~l volumes. The result was the operating
program set out in Table III below.
--3 9--
N ~ ~D ~1 o ~ Cl) r` ~ ~ ~ ~ ~ ~ a~
0 3 ~ u~l ~ ~1 ~ a~ ~D ~ ~`~ ~ ~ CO 1~ U~ ur~ ~r
U1 U ........... o ~ ~ ~
~ ~ ~ ~ ,~ ~ o o o o O O
CJ` ~ ~
C h X
E u~
~ h`~
O ~ o a~ u~ O r~ ~ _ 0 ~ ~cr ~ O
C4 ~s a u ~ ~ ~ ~ ~ ~ ~ ~ .t ~ ~
a)-- __~___________~
h ~ ~ ~r N ~ O ~D t~ ol ~ ~ ~ ~ I` (~) u^l
CU~ ...............
~ ~ oooooC~
N ~J _ _ _ _ _ _, _ _ _ _ _ _ _ _ _
._1 h X
h fl, _
~1) It~ 1~ Lr)
. . .
Ul o ~ o o q~ r ~ ~ ~ o a~
~: a~ u~ ~O ~O ~n
E ~ CD _ a~
O
h ~ ~ ~ `~ ~ ~ ~r ~ ~r Yr ~r ~r ~r ~r ~ ~r
~ ~ ~ _ _ _ ~
c ~n--
W . o o o u) u~
m ~ O _
_.
J-
E ~ ~ ~1~ ~1 ~ t~ ~`
~ ~n
a) c~ ~
E C
,~ o
E~ V
aJ
h
h
a
aJ ~
O o~D O ~D
E h ........ ~ ......
I 1 _I
' 3 ~ 5 t~ L~ ~ ~
The 9.9 inch (24.9 cm) long filter ele~ent,
prepared a5 described above, on a l.l inch (2~79 cm1
ia x 1.3 inch (3.30 cm~ od x 9.~ lnch (24.9 cm~ long
core, had an od of 2-l/2 inches 56.35 cm~, i.e., the
5 fibrous mass had an inside diameter of l.~ inches
(3.30 cm) and an outside diameter of 2-l/2 inches
(6.35 cm). The filter element exhibited the fol-
lowing properties:
Clean pressure drop was l.8 psi (0.13 kg/cm2)
at lO liters o water per rninut~. The life or dirt
capacity was 83 grams to 60 psi (4.22 kg/cm21 dif-
ferential pressure filtration efficiency was in
excess of 90 percent at l.0 micrometer, 99 percent at
3.7 micrometers, 99.9 percent at 5 micrometers and
99.99 percent at 5.6 micrometers; bacteria removal
efficiency tested using 0.~ micrometer diameter Pseu-
domonas diminuta (Ps. d~ organism wa~ 99.997 percent.
Because of its very high dirt capacity (long
life) and removal capability over the full range of
particle diameters f~om O.l to 40 micrometers, this
type oE filter i5 particu ~ rly well suited as a pre-
fllter. For example, it could b~ used to precede an
absolute rated final filter when used in critical
applications, ~uch as ~terilization of parenterals or
for providing water for use in the manufacture o
microelectronic devices. Because of it~ wide range
capability, it would also serve well as a prefilter
for a coarser after filter, ~or example~ one rated at
5 or lO micrometers. It can al50 be used as the only
filter in the system for many other applications.
..... _ _ .
-
~5
-41-
Ex~mple 11: Filter Element With Constant
~iber Di~meter Of 1.~ Micrometers For The
Inner 50 Percent Of The Fibrous Portion
~ The Element And Varying Fiber Diameter
For The Outer 50 Percent, With Constant
Voids Volume Thro qhout:
The filter element of this Example was prepared
- in the same general manner a~ that of Example 10 and,
as with the filter element of Example 10, had a voids
volume of 82 percent but differed from Example 10 in
that the initial 50 percent by weight of the fibrous
portion of the element was made up of fibers having a
constant diameter of 1.9 micrometers with the halance
varied, again as a geometric progression~ from 1.9 to
12.6 micrometers. This was accomplished by the oper-
ating program set out in Table IV below.
. _ . _ _
. .
~, ~ ~ _ _ _ ~ _ _ _ ~ ~ ~ _ _ _ _ ~
O ::1 E u~ ~ n ~
a~ \ _ ~ _ _ _ _ _ _ _ _ _ _ _ _ _
~, _, n
O -~ U~ o ~ u~ o ~ ~ o ~ ~ O ~ ~ I~
r~ ~ Q ~n ~r ~ r~ ~ ~ ~ ~ _l ~ ~ ~
o -- _ _ _ _ _ _ _ _ _ ,_ _ _ _ _ _
~ ~ ~ r~l ~ o ~D c~ ~ ~ 0 ~ n
E u~ n r~
...............
~ ~ _ _ _ _ _ _ _ _ _ _ ~_ _ _ _ _
In n n
o ~ o o er un ~ ~ r ~ o a~ n
~ ~ Q u) ~ \D ~D In ~r ~ ~ ~ ~ _~ ~
H L u7 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
r~ _1 u)
~ u) a~ ~ o o o u~ m un In In un u- u~ m u n In
r~ a) h u~ O _l O s~ N ~ 1 N N N N 1~ N t`J N
V
t~ ~ ~ N O ~ I` ~ ~ o a:~ )~ un
_~ ~ a~ o ~ ~ ~ ~7 ~r m ~ ~ o
~ ~ ~ ~1~ ~ ~ r~
Q) tJ ~
E c
-_~ o
~ u~
v
C
~ a
v E
~ O ~ ~ un ~ ~ r r~ r ~ o ~D
f7 ~ O ,~ P U7 ~ 1~ ~ C~
~ n :~ ~ ,~
-43-
The resultinq element had the same idt od and
length as that of Example 10. It exhibited the fol-
lowing properties:
Clean pressure ~rop: 4.3 psi (0.30 kg/cm2~ at a
S test flow rate of 1~ liters of water per minute
Life or di~t capacity: 36 grams to 60 psi (4.22
kg/cm2) dif~erential pressure;
Filtration efficiency: in excess of 99 percent
at 0.7 micrometer (as estimated by extrapolation),
measured as 99.9 percent at 1.4 micrometers, as 99.99
percent at 2.2 micrometers, and as 99.999 percent at
3 micrometers.
Because of its very high efficiency at 2.2
micrometers, this filter element can, for nearly all
purposes, be rated as 2.2 micrometers absolute. It
also provides very useful levels of removal for par-
ticles as fine as 0.7 micrometer. These high effi-
ciencies, coupled with the very hiyh dirt capacity of
36 grams under the F2 test as described above, pro-
vides a highly usef~l~filter with a long servicelie. ~
It should also be noted that while the efficl-
ency o~ this filter element is f;ner than that of
Example 4, its service life (dirt capacity) is over 4
tirnes higher and, indeed r is equal to the service
life of a uniform pore filter with a removal rating
of about 20 micrometers at 99.9 percent efficiency.
In order to further characterize this ultra-
fine, long lived filter element~ an element made in a
similar manner was tested by passing through it a
suspension of Pseudomonas diminuta bacteria. This
organism is cylindrical in shape, with a diameter of
0.3 micrometer. ~fficiency of removal was 99.997
percent.
Elements made in th~ manner as described above
-44~ 5 3
are well ~u~ted for th~ ~iltr~t~on o a var~ty o
product~ from which yea~t and bacteria are to be
removed, yielding not only a liqu~d e~fluent free of
or greatly reduced in its content of yeast and bac
teria, but also one with high clarity.
Example 12: Filter Element With Constant
Fiber Diameter Of ~.9 Micrometers For The
Inner 59 Percent Of The Fibrou~ Mas-~, And
Varying Fiber Diameters For The outer 41
Percent~ With Constant Voids Volume
Of 82 Percent Throuqhout:
The filter element of this Example was prepared
in the general manner of Example 10. However, the
initial 59 percent by weight of the fibrou~ portion
of the supported filter element of thi~ Example had
fibers with diameters of 2.9 micrometers with the
balance varîed as a geometric progr~ssion from 2.9 to
12.6 micrometer~. This wa~ accomplished by the oper-
ating program set out in ~able V below~
~,
--4 5--
5 ~ ~ j L~ ~3
_~ ~ ~ o a~ ~ t--~ c~ ~ ~ ~ ~D
O ~I E ~ C~
............
.Ioooooo
~ aJ tl _ _ ~_ _ _ _ _ _ _ _ _ _
C ~X
E
~ 1
O ~ ~n o ~ D ~ ~ o a~
1:4 ~ Q r~
~__ ______~______
~ ~ o ~ c~ a~
t7~ ~ E ~ cs7 ~ _ ~D ~ ~ c~ r~ ~D In r~
C ~n t) ............
~ ~oooooo
N a) t~ _ _ _ _ _ _ _ _ _ _ _ _
-,1
u~ In Ll~
~ 1 . . .
.,1 _1 U~ O ~ q- ~ o a~ r~ u~ ' '
O
:~ )~ t~
~ _________.__~_
~ C u~--
m u7 ~ -
~S a) )~ v~ ~3 N N ~ N N N (`i ~1 N N N
E~p:; ~ O
a~ ~ . .
.~ ~ I
0 ~ ~ O cr~ ~ ~D ~ ~ ~
~r ~n ~ r a~ co cr~ o ~ ~ r1
E ~ ~t ~ ~1 ~ ~ ,J ~ ~ N t`J N
V~
C
._1 o
E~ ~ t)
1~ a
U~
~.
a
S~ V
Q) a)
E
O . . . ~ a~ o o ~D
~ n :E
~46~
The ilter element o~ thi~ Example was prepared
by the g~neral proceaure described ~n Example 11
above and had the same id, od and length as the ele-
ment of E~ample 11. It exhibited the following pro-
perties:
Clean pressure drop: 1,5 psi (0.11 kg/cm~) at atest flow .rate o~ 10 liters per minute of water;
Life or dirt capacity: 53 grams;
Filtration efficiency: 90 percent at 1.1 micro-
meters, 99.9 percent at 4.6 micrometers and 9g.99percent at 5.8 micrometers.
Filters of the type of this Example have ap-
plications in.fielcls such as filtration of magnetic
particle suspensions used for video recording tape
manufacture and for processing photographic film
emulsions.
Example 13: Filter With Fiber Diameters
Varyin~ From 8.5 to 12 S Micrometers-
Using a procedure ~i~ilar to that of Example 10
but starting with 8 5 micrometer aiameter fibers, a
filter element was prepared which had the following
properties:
Life or dirt capacity: ~15 grams up to a pres-
sure drop o~ 0.6 p~i (0.042 kg/cm2) with a removal
rating of 24 micrometers. The 115 gram life to 0.6
psi ~0.042 kg/cm2) i~ very much higher wh~n compared
with the best commercially available ilters o equal
removal rating which use a varying density ~tructure,
as opposed to a sub~tantially uniform void~ volume
and corresponding substantially uniform density with
a varying or graded fiber diameter structure.
f
5~
.
Examples 14 throuqh 17:
~ number of filters have been described in the
literature which ;eek to obtain increased life or
dirt capacity using fibers of uniform diameter by
varying ~he pore diameters ~rom larger upstream to
smaller downstream by decreasing the voids volume,
i;e., increasiny the density, of the filter medium in
a progressive manner. The characteristics to be
expected of such a filter can be projected using the
test results obtained in the following Examples 14
through 17
The filter elements of this group of Examples
were made in the same general manner as Example 1.
Each was prepared with uniform fiber diameter of 3.2
micrometers and with uniform voids volume in each
element. However, the voids volume varied from one
element to the next~ as noted in Table VI below.
.
L~ ~3
~ .
.
~ C O ~ ~ ~ ~ CO
O -- ~ a.
E ~ t~ v ~1 ~ ~ ~
a) ~ ~ o
E E
Vl
a~ ~ ~ ~ ~D ~
h r` ~D ~ o
t~
a.
a
~)
O O ~ ~ c~
:>
__
~ ~ _ _ _
O~ ~ ~D O
r~ ~ r~ 1--
a) ~
O
,
.,, U~
~t E ~v~
> ~ ~
O o ~u,o u~ ~n o
a~
~ ~:
_.
a) ~
_ ~ _ _
O
C u~ ~ c~ ~ n~
N ~1) ~ ~`J ~`J ~'J ~ ,,
_ _ _ _
O O O O
r~a ~ _ _ _
a
L~ ~ u~
:~ ~C ~ '7 ~ t'7
U~ _ ~
C~> O O O
~IJ 1`~ 07 O O o O
a~
E
~ E
X ~
Z
3~3~
It may be deduced from the abo~e aata that lf
one were to make a compo~ite filter, wlth voids vol-
ume grad~ated from 79 to 88 percent, its life would
be not better than 10.3 grams under the F2 test.
Furtherl its removal rating wo~ld be somewhere be-
tween 1.7 and 3. a micrometers. These characteris-
tics, when compared with the data of Examples 10-12,
show that an element made with a constant voids vol- -
ume, but with varying fiber diameter, would have at
least four times the life or dirt capacity at equal
efficiency.
The conclusion that far better life can be
obtainea using constant voids volume with varying
fiber diameter, as opposed to constant fiber diameter
and varying voids volume ~or aensity), is also sup-
ported by similar data (not presented herein) at
fiber diameters other than 3.2 micrometers.
Examples 18 throuqh 21:
2 0
The same conclu~ions~regarding the infèriority
of the approach of varying the void~ volume or den-
sity to obtain a graaed pore structure vis-a-vis
varying the fiber diameter while maintaining a sub-
stantially uniform voids ~olume is also supported bythe subject Examples 18-21. The filter elements of
Examples 18-20 were prepared on 1.3 inch (3.3 cm) od
coresO In each of Examples 18~ 19 and 20, the fiber
deposited between 1.3 inches (3.3 cm) diameter and
2.1 lnches (5.3 cm) diameter was identical, IOe., the
fiber diameter used was 3;2 micrometers. The ~oids
volume of this portion of all three Examples was 83
percent.
For the filter element of Example 18, the de-
35 position of fibers was terminated at the 2.1 inches
~50--
(5.3 cm) diameter~
The fllter element of Example 19 wa~ graded or
profiled between 2.1 inches (5.3 cm) diameter and 2.5
inches (6.35 cm) diameter by varying the diameter of
the fiber from 3.2 micrometer to 12.5 micrometer
diameter in the manner of Examples 11 and 12 while
maintaining the voids volume con~tant at 83 percent.
The filter element of Example 20 was profiled
between 2.1 inches (5.3 cm~ diameter and 2.5 inches
(6.35 cm) diameter by increasing the voids volume
from B3 to over 90 percent while maintaining the
fiber diameter constant at 3.2 micrometers.
As indicated in Table VII below, Example 21 was
prepared in the same manner as Ex~mple 20 except that
15 the fiber diameter ~as 3.6 micrometers throughout.
The characteristics of the ~our elements of
Examples 18-21 are set out in Table VII below~
;~0 ! :~
_. ~
-51- 1.25~
. .
U~
~ a~
~E
~ ~ O ~O C~ ~D ~
0-~ ~ . . . .
E~ 1~ Ur ~ (.~1 ~\1 ~r
Q~ n', E
a) ~(:~:) cl:~ ~1 .
~ ~ ~r ~ ~
l ~ ~3
v ~ a~ E ~ o ~)a ~ o Q~
~ ~ o ~ a) a~ a~ ~ v a) ~ ~ v
o IIn O aJ h U U~ V E E E ~ J v F ~ tr~ a) ~
n ro r~ E ~ _ ~J E ~ ~ h C h ~ ::~ O .C C E ~ ~ ~ O .C C E ~; h
a) aJ ,~ o _ _ v 3 --~ ~ o ~ ~ ~
Q) ~ a u~~ v ~ E _l ~ ~I v ~ ~ o o ra~ ~ I ~ ~ O ~ c -~ ~ v
V ~ ~ a a) ~ > ~ C ~ ~ a C ~
t.) O ut (~7 ~J C L: ~ ~ O E C_ c~ C ~1 tU E ~ ~ S ta E
n5 v ~ h l ~1 -1 a h ~-~ v O O ~ u~ u~ ~ ~ ~ O U~ -~ Ql ~ ~ ~ O
C C) ~O E E (V ~u h ~ ~ ~ O C a) Vl h
~ v ~ ~ ~ _ ~ O ~ ~ -I O O ~ ~ U ~ O O ~ ~ ~ C
H C ~ ~ a) 0 -1 ~ ~ ,C C O O 1~ 0 ~1 O ~ h a) ta - o~_~
~> ~ E V a E~ ~ ~ ~ :~: 3 ~ ~ ~ P~ ~ P~ ~ U :~:
m ~ ~ ~ ~ ~ ~ ~ h
~: -- a) a)-- a) a) -- a) a) ~
OP ~ ~ ~P ~ ~d~ ~ ~ dP ~ 1
o I ~ aJ Q) ~a) a) ~ a) ~ aJ a)
_1 ~ ~ a:l ~ E cO E`E ~ E~ E a:~ E3 E
E--_ O r~l _ O ~_ O 1~ ~ O
Q~ ~ a ~ u ~ h--~ h -~ ~ ~ ~ h-~
E u n E ~ E~~ n E~ u
U O ~1 ~ v h ~ç h ,~ 4,i ?,1 ,~
n3 v Ql )~ o u~ ~, o u~ :E 1~ o u~ h o Ul 5 h
a~ u E ~ ~ a~ ~ ~ a~ u~
v V ~a .~ ~ .a ,, ~ ~,q ,. ~ ~-1 ~ r
C ~ Q~ 0 -1 C O ~ ~ o -~ ~: O ~-~ C: O -_~
E ~ cl =) > r~ 1~ ~ ~ ~ :>
,_
E _ u~ u~ _
U ~ . .
1~ _ . U~ .
a) ~ In _ '~_ ~O
U) _ _ _
u~ E E ~:
c) _ u~ u~ n
OQO~-t ~I~I ~ ~
~I)
_ 5
~ a~
E ~ ~ ~ o
Z
-52-
In Examples 19 and 20, ~oiaS volume~ above 90
percent were not used because they were deemed too
soft and compress;ble to be practical.
From the data set out in Table VII and by
interpolating between Examples 20 and 21, it may be
seen that this type of element, if made with a fi~er
dia~eter such as to yield a re~oval rating o~ 2.8
micrometers, i.e.~ equal to that o~ Example 19~ would
have a dirt c~pacity of only about 12 grams or about
one-quarter that of Example 19.
Examples 22 tllrouqh 24:
The filter elemerts of these Examples were made
in a manner similar to that of Example 12 except that
the starting fiber diameters varied from 3,2 to 4.8
micrometers insteaa of the 2.9 micrometers of Example
12. l'he resulting structures had voids volumes of 82
percent and tne characteristics set out in Ta~le VIII
below:
~ .
TABL~ VIII
Exa~ple Starting ~ife Removal
25 Number Fiber Dia-Gr~ms Rating,
meter, Mi- Micrometers
crometers
22 3.2 52 9
23 4.0 67 1~
24 4.8 63 14
-53- ~ S~
Examples 25 through 28: Filter Element Series
Illustr~ting EfEect Of Variation O Voids Volume:
A series of 2-1/2 inch (6.35 cm) od x 1~1 inch
(2.79 cm~ id x 9.8 inch (24.9 cm) long filter ele-
ments were prepared at conditions such as to produce
2.2 micrometer fibers. By varying the water spray
rate and the forming roll pressure, voids volume was
varied from 72.1 percent to 91.8 percent.
The properties of the resulting filters are set
out in Table IX below:
r
~.
~;
- 5 4 -
. . .
O
t)
._~
U~
. ~ ~P
E ~ u~ 'r o
O c~ o
h cn
.
tr~
~J
t,31 E~
~1 ~;
E~
al
w E ~ ~
In ~o o
u~ E ~ `
, ~`I o r~ ,~
O O
:~ ~ ~,
Q~
_ IJ
E a u~ ~ r` CD
)~ ~
2:
` -55~
.
The dlrt cap~clty oE Examples 25 through 28 w~s
plotted ~s the ordinate against the 99.99 perc~nt
removal rating ~ the abscissa ana a llne dxawn
throuqh the four experimental points. On thi~ line
it was seen that the dirt capacity of a filter made
in the rnanner of Examples 25-28 with 99 99 percent
efficiency at 2.2 microme~ers would have a dirt capa-
city of 5.9 grams. This i~ in markea contrast to the
dirt capacity of the filter element of Example 11,
which had a 99.99 percent efficiency at 2.2 micro-
meters but a dirt capacity of 36 grams, i.e., six
times greater.
Examples 29 through 34: Filtering
Element Series Illustrating The Effect Of
Variation OE Voids Volume Or Density:
A ~eries of Eilter elements similar to those of
Examples 25 through 28 was prepared using f ibers
having 12.5 micrometer diameters and with vvid~ vol-
ume~ varied from 63.6 to 89.8 percent.
The characterlstics o~ the elements are set out
` in Table X below:
. .
- 5 6- ~l~ 5~
o
.,.
t,
.,
JJ
~o
~ a~
h
o r~
a~ ~ ~ I
E
h
...~
d~
C~
_~ _
C a~ ~rco In a~ ~ ~ ~
X_ a~ ,~ U
a) u
cr:
m a~ a)
~: ~ 41 v
d .~
E, ~ 1 0
a~ t`3 s~
CL~ ~ E
C~
~, ~ U7
tn n ~
E _ o a
' ~ n~ ~ ~ u~
-_1 h ~ r~
~ ~ ~e; c
~ Q)
a~
C4 ~
QJ U~
. . . . .
r~ r o In ~ O
O O U~
:~ ~ 5:
U~
E t~
r~ h
~ D.
U~ o
o a~,
a
E n
I~ E
X ~ _~ N
X , . ~_ _
.
_57~ ~5~6~3
When tlle above elements axe compared with ele-
ments oE e~ual efficiency made by the method in accord-
ance with this invention, life is seen to be much
highex for the latter. To illustrate this, Exarnple
5 29 above may be compared to Example 13 when both are
tested with a glass bead contaminant suspended in
~IIL-EI-5606 hydraulic fluid. They have virtuall~
iclentical removal ratings of 14 and 15 micrometers,
respe~ively, at 99.9 percent efficiency. However,
10 the dirt capacity of Example 13 is ~ grams versus
only 12 grams for Example 29 (when tested to 60 pSi
t4.2 k~/cm2) pressure drop).
Examples 35 through 37:
]5
This group of Examples compares the collapse
pressures of a filter element made with interfiber
bonding induced by use of molten or softened fibers
(Example 35) with filter elements made using fibers
20 substantially free of this type of interfiber bonding
(Examples 36 and 37).
Example 35 is a purchased specimen of a commer-
cially available ~Iytrex brand filter (available from
Osmonics, Inc.) made using polypropylene fibers and
25 which is characterized by the presence of very strong
interfiber bonding. On examination, the bonding was
seen to be caused by the adhesion to each other of
melted or softened fibers to form a coherent mass.
Example 35 had no internal support core. Examples 36
30 and 37 were prepared by the method in accordance with
this invention, also with no support cores. Collapse
pressures were determined by individually wrapping
the outside of each of the test elements with a thin
water-impervious plastic film, sealing the ends of
35 the element and then applying pressure to the exter-
-58-
ior Q~ the el~ment with water in a transparent hous-
ing so that the Eailure of the element could be ob-
served.
The dimensions, rating, voids volume and col-
5 l~pse pressure for each element are set out below:
Example 35: ~Iytrex 20 micrometer element, 2.75
inch (7 cm) od x 1-3~8 inch (3.5 cm) id x 10 inch
(25.~ cm) long, average voids volume 7~.7 percent.
Collapse pressure was 80 psid (5.63 kg/cm2).
Example 36: A twenty micrometer rated element
was made by the method ln accordance with this inven-
tion, 2.75 inch (7 cm) od x 1-3/8 inch (3.5 cm) id x
10 inch (2S.4 cm) long with no support core. Voids
volume was 75 percent. Collapse pressure was 16 psid
]5 (1.13 kg/cm2).
Example 37: The element of this Example was
similar to that of Example 36 except that the voids
volume was 81.5 percent. Collapse pressure was S
psid (0.35 kg/cm2).
The much lower collapse pressures of the ele-
ments in accordance with this invention are due to
the substantial absence of interfiber bonding. Con-
versely, the Elytrex element was sufficiently
strengthened by interfiber bonding that it had the
25 necessary strength to withstand up to 80 psid (5.63
kg/cm2 ) .
Examples 38 and 39:
Nylon 6 resin was fed into the same apparatus
and processed in the same general manner as previous-
ly described using polypropylene. The operating
conditions and properties o~ the resulting elements
(when tested using the F2 test~ are described below:
_59_ ~ 5~L~3
Example 38 Example~39
Resin Temperature
Degrees F 693 693
(Degrees C)(417, (417)
5 Resin Pressu~e, psi 300 300
(kg/cm2) (21) (21)
Eiberizing Air
Pressure, psi 40 12
(kg/cm2) (2.8) (0.84)
10 Fiber Diameter,
micrometers 2.3 4.0
Voids Volume, Percent 80.5 74.9
Removal Rating,
micrometers 5.0 6.1
]5 Life, grams 12.2 10.2
Example 40:
Filter elements are prepared in accordance with
20 thisinvention in a similar manner to the polypropy-
lene elements previously described, but the resin
used is polymethylpentene. The properties and char-
acteristics of the products are very similar to those
obtained using polypropylene.
Exam~le 41:
-
A filter element made in the manner of Example25 was tested by passing through it 1,000 ml of a
30 suspension of Pseudomonas diminuta (Ps d), a 0.3
micrometer bacterium, in water. The effluent was
analyzed for its content of this bacterium. Whereas
the total number of bacteria in the influent was 2.3
x 1012, the effluent content was found to be 1.6 x
35 105, indicating a reduction by a factor of 1.4 x 107.
-~io-
This corresponds to an efficiency of 99.99999 per-
cent.
E.Yam~~l 2:
Two filter elements (42A and 42B) were made in
a manner generally similar to Example 11 except that
the inner 50 percent of this element consisted of 1.7
micrometer diameter fibers and the outer 50 percent
10 was proEiled up to 12.5 micrometer fiber diameter.
Both Eilters were tested with Ps d in the same manner
as in Example 41 and, in addition, were resterilized
and retested using Serratia marcescens (Serr m) as
the test organism. The results are shown in Table XI
]5 below.
~2~
o~
C~
~ o~ a~
:~ ~ .
U Q) ~ ar~
C U~ s~
u
. ~ ~D
'~ ~D
C~ ~ cr
D~ a~
r~ ~
o o
E
X X
E
~ ~ ~ a~
a~ ~ .
V ~ ~
C
C~
o o
X X
X ~ U~
3 ~ ~
o o
o C~
E ~ _~
X X
~,. a: 0
Q)
E u~
oJ
v
C~ ~ ~
:~ o o
~ X ~C
a. ~tc m
E
X
: ' '
., " , ,,
-62-
E~amE~ 43:
The general procedure of Example 10 is repeated
except that the variation of the resin flow rate, of
5 the fiberizing air flow rate and of the forming roll
pressure is continuous as opposed to stepwise, pro-
ducing a filter element with a continuously graded
fiber diameter structure, i.e., a continuously pro-
filed structure, with characteristics comparable to
10 those of the filter element of Example 10.
Eilter elements are prepared in accordance with
]5 this invention and in a similar manner to the poly-
propylene elements previously described, but the
resin used is polybutylene terephthalate (PBT). The
proper-ties and characteristics of the products are
similar to those obtained using polypropylene. How-
20ever, because of the higher melting point of PBT andits resistance to hydrocarbons, filter elements pre-
paxed from PBT will be useful at higher temperatures
and in service where they will come in contact with
hydrocarbons which might cause the polypropylene
~5fibers to swell.
Example ~5 Relative Compressibility
Of Course And Fine Fibers: -
Filter elernents were prepared with essentially
uniform fiber diameter throughout in the manner of
Fxample 1, differing only with respect to fiber dia-
meter. Using a tool resembling a laboratory cork
borer having an inside diameter of 0.58 inches (1.473
35cm), specimens were cut from the fibrous portion of
.
~3L~D r
-63-
eacil ~ilter element perpendicular to tlle longitudinal
axis of the element, forming a generall~ cylindrical
specimen ~bo~t 0.6 inches (1.52 cm) in length and
about 0.58 inches (1.47 cm) in diameter.
Measured forces of 10 psi tO.7 kg/cm2), 20 psi
(1 4 kg/cm2), 60 psi ~4.22 kg/cm2), 90 psi (6.33
kg/cm2) and 120 psi (8.44 kg/cm2) were individually
and sequentially applied to the ends oE each of the
cylinders while the thickness of the cylinders at
10 e~ch level of applied force was simultaneously mea-
sured.
Three elements with Eiber diameters of 2.0, 6.8
and 12 micrometers were each individually tested in
this manner. The decrease in thickness of each, when
~5 plotted against force applied, was very similar Eor
all three fiber diameters~
Example 46: Collapse Of Filter Flement When
Vsed With High Viscosity_Fluids At Hi~ Flow Rates~
. ~
Two filter elements 46~ and 46B were prepared
using identical procedures in a manner generally
similar to Example 24 except that the average voids
volume was 2 percent lower (80 percent cf 82 per-
25 cent) Element 46A was tested using the F2 methoddescribed above at 10 liters/minute of water. It had
a removal rating of 1l.2 micrometers and a dirt capa-
city of 53 grams. The clean pressure drop, prior to
the test, was 0.7 psi (.05 kg/cm2) a-t am~ient temper-
30 ature of 20 to 25 degrees F (6.67 to 3.89 degrees C).
Element 46B was placed in an F2 test standwhich used hydraulic fluid MIL-H-5606 at 100 degrees
F (37.8 degrees C). At this temperature the viscos-
ity of MIL-H-5606 is 12.7 centipoise or 12.7 times
35 that of water. No test contaminant was added to the
5~
-64-
system, instead clean fluid was flowed through the
element at the flow rates set out in Table XII below.
T~BLE XII
Flow Of MIL-H-5606 Pressure Drop psi
Liters/Minute (kg/cm2)
1 1.3 (.09)
~ 11.0 (.77)
7 25.5(1.79)
46.5(3.27~
13 cartridge failed
due to core collapse
]5 at approximately 80
psi (5.63 kg~cm2)
The pressure drop through the filter elements
in accordance with this invention is proportional to
20 flow and to viscosity. Based on the aqueous pressure
drop at lO liters/minute of 0~7 psi (.05 kg/cm2), the
calculated pressure drops for MIL-H-5606 are 8 psi
~.56 kg/cm2) at 10 liters/minute (vs. ~6.5 psi (3.27
kg/cm2) measured) and about 11 psi (.77 kg/cm2) at 13
25 liters/minute (vs. approximately 80 p5i (5.62 kg/cm2)
measured).
The much higher pressure drops when using the
viscous fluid are due to the compression of the fil-
ter medium. This example illustrates that use of
30excessively high voids volumes is not desirable for
applications in which high flow of viscous fluids at
high pressure drop are involved, particularly with
the finer grades of filter element.
-
-65- ~ 5~L~ 3
-
Example 47 Filter Element With The Inner
T~ `hirds With Voids Volume O~ 74 Percent
~nd The O-lter One-Third Profiled In Fiber
Diameter To 12.5 ~icrometers Also At 74
Percent Voids Volume
~ filter ele~ent is prepared using the general
procedure of Example 11 modified as follows:
The initial 67 percent by weight of the fibrous
10 mass of the element is made up of 1.6 micrometer
diameter fibers prepared using a water spray coolant
and with the forming roll air pressure adjusted to
obtain a voids volume of 74 percent. The outer 33
percent by weight of the fibrous mass of the element
is applied also using a water spray coolant but while
adjusting the resin rate, fiberizing air pressure and
the forming roll pressure in a manner such as to
profile the fiber diameter in a continuous manner
from 1.6 to 12.5 micrometers while maintainins a
uniform voids volume of 74 percent. The resulting
filter element will have a ~itre reduction in excess
of 107( 99.99999 percent efficiency) when tested
using 0.3 micrometer diametèr Psuedomonas diminuta
organisms and will have a much higher dirt capacity
than a similar cylindrical ilter element in which
100 percent of the weight of the weight of the
fibrous mass is made up of 1.6 micrometer fibers with
a 74 percent voids volume.
- - - . . .
~ 6~ 3
E~ample ~8: Filter element With The Inner Two-
Thirds With ~oids Volume Of 74 Percent And The
Out~r One-Third Profiled In Both Fiber Diameter
~nd Voids Volume, ~espectively, Up To 12.5
~~icromete~s And 85 Percent Voids Volume:
A filter element is prepared using the general
procedure of Example 47 but modified as follows:
The outer 33 percent by weight of the fibrous
mass of the element is applied also using a water
spray coolant while adjusting the resin rate, fiber-
izing air pressure, and forming roll yressure in a
manner such as to profile both the fiber diameter and
the voids volume simultaneously, both in continuous
fashion. The Eiber diameter is profiled from 1.6 up
to 12.5 micrometers at the od and the voids volume is
pro~ilea from 74 percent up to 55 percent.
The resulting filter element, when tested for
efficiency using the Pseudomonas diminuta bacteria,
will have an efficiency essentially equal to that of
Example 4? but with a some~hat higher dirt capacity.
-67- ~ ~5~3
The cylindrlcal fibrous structures in accord~
ance with tl~e subject invention Eind use in a variety
o~ filtration applications. The filter elements in
accordance with the subjeet invention combine extend-
5 ed ~ilter life, i.e., higher dirt capacity, at e~ualefficiency, or better efficiency at equal life, or
both better efficiency and higher dirt capacity than
previously available commercial fibrous cylindrical
depth filters. A combination oE high dirt capacity
10 (lon~ life) and removal capacilities over a wide
range of particle diameters ma]ces filter elements in
accordance with this invention useful as prefilters,
for example, to precede an absolute rated final fil-
ter when used in critical applications, SUC}I as ster-
]5 ilization of parenterals or for providing water foruse in the manuEacture of microelectronic devices~
Filter elements in accordance with this inven-
tion are also well suited for the filtration of a
wide variety of products from which yeast and bac-
20 teria are to be removed, yielding not only a liquideffluent Eree of or greatly reduced in its content of
yeast and bacteria, but also one with high clarity.
~ilter elements in aecordanee with this invention may
also be used where high titre reduetions, coupled
25 with high dirt capaeities, are required for removal
of bacteria.
In addi-tion to their primary use as depth fil-
ters with high effieieney and extended life, the
eylindrical fibrous struetures in aecordanee with
30 this invention also find use as eoalescers and in
insulation applications.
