This invention relates to a macroscopic composite hollow fiber membrane which is reinforced with multiple multifilament yarns braided into a tubular shape; each such yarn is made by plaiting or twisting multiple strands or monofilaments of a polymer having desired tensile strength. The tubular braid is then coated with a membrane-forming dope (polymer solution or “dope” for brevity) which is coagulated to form a tubular polymer membrane having a lumen diameter of at least 0.5 mm. Such construction is designed for use in demanding applications in which permeate is required to meet strict specifications. For example, for water filtration, one or more skeins are used, each skein comprising a multiplicity of hollow fiber membranes (hereafter “fibers” or “membranes” for brevity) in a module which is deployed in a reservoir of “dirty” water to be filtered. A “skein” of fibers is a bundle of fibers deployed adjacent one and another, all in the same general direction. A “module” is a skein of fibers opposed ends of which are secured in headers, typically by potting. Multiple modules are typically deployed in a reservoir containing a large amount of liquid to be filtered, e.g. in municipal water filtration plants. When, during filtration, the pressure drop through the fibers in a module gets sufficiently high, the module is backwashed with permeate under pressure. Desirable systems provide permeate economically, by providing high permeate flow and by minimizing damage to the hollow fiber membranes.
Braided hollow fiber membranes are commonly used in modules containing from several hundred to several thousand membranes. Damage to a single membrane in a module, which damage results in dirty water contaminating the permeate, is a serious problem which occurs more often than desired. Though the permeate is typically water, the permeate may be any filterable liquid to be separated from a suspension or dispersion of finely divided particles in the liquid.
To date, numerous braided membranes have been disclosed, each of which purports to provide desirable filtration efficiency but offer scant useful knowledge relating to avoiding the damage to a membrane or maximizing permeate efficiency. Emphasis on physical strength of the membrane is embodied in disclosures of U.S. Pat. Nos. 3,644,139; 4,061,821; 5,472,607; 5,914,039; 6,354,444; 7,165,682; 7,172,075; 7,306,105; 7,861,869; 7,909,172 and others.
The requirement of strength decreed that these prior art braids be made by braiding multiple yarns, each comprising lengths of multiple monofilaments (or “filaments” for brevity). The drawbacks of using multifilament yarns were either overlooked or ignored.
The method of making a tubular hollow fiber membrane decreed that the cross-section of the lumen be circular. Whether such a fiber is made by coating a woven tubular reinforcement with polymer, or extruded with a woven circular reinforcement embedded in the fiber's wall, the cross-section of the lumen remained circular.
Membranes such as are disclosed in U.S. Pat. No. 4,061,821 to Hyano et al, (hereafter '821) have braid embedded beneath a thick polymer film to provide a stabilizing effect during use of the membrane. The term “embedded” as used herein describes yarn or monofilament with at least 99% of its surface coated with polymer. Braid having an inner diameter in the range of 0.5-10 mm and unspecified “thin thickness” is made from filaments overlying and randomly overlapping one and another (see FIGS. 4, 5 & 6 in '821) in multiple layers but are preferably made from multifilament yarn. The stabilizing effect of the openings in the reinforcing material was lost (see sentence beginning at the bottom of col 4, and bridging cols 4 and 5) when the braid was coated with polymer, so that their reinforced membrane was not an effective membrane.
The problem of stability was addressed in U.S. Pat. No. 5,472,607 to Mahendran et al, (hereafter '607) which teaches a film having a wall thickness in the range from 0.01 mm to 0.1 mm, supported on the outer circumferential surface of a preshrunk braid; a major portion of the area of the circular cross-section of the porous tubular support, viewed along the longitudinal central axis, is free from film and not embedded in film. Thus, it was not known how embedding the braid in the film affected the performance of the membrane; nor was it established whether failing to embed the braid in the film provided a significant advantage.
Membranes such as are disclosed in U.S. Pat. No. 6,354,444 to Mahendran et al, (hereafter '444) are produced by first weaving a tubular braid of multifilament yarn to have a cylindricity>0.8, preshrinking the braid, then coating the outer circumferential surface of the cylindrical braid with a dope of polymerizable membrane-forming polymer. The term “cylindricity” (sometimes referred to as “roundness”) refers to how perfectly the circular cross-section of the tubular support matches the geometry of a true circle drawn to correspond to the mean diameter of the braid, a perfect match being 1.0. By “weaving” is meant that the filaments are interlaced without being knotted (as they would be if the braid was knit). “Dope” refers to fluid “membrane polymer”, e.g. poly vinylidene fluoride (“PVDF”) whether molten or in a solution. If in solution and coagulated, the dope forms a film having a wall thickness of >0.2 mm and with desired attributes for the filtration of fluid to be filtered, typically dirty water. The '444 braid is relatively dense, having an air permeability in the range of from 1 to 10 cc/sec/cm2 at 1.378 kPa so that the voids in the braid are small enough to provide substantial resistance to the passage of air, and thus inhibit substantial penetration of polymer. The braid is preshrunk to provide stability of the braid. Yarn lying in a generally longitudinal orientation (along the z-axis) provides extension at break of the uncoated braid of at least 10% which extensibility is referred to as “give”.
The weave of the '444 braid is a circularly woven tubular braid. Such braids are tightly woven with at least one circumferential yarn lying in a generally x-y plane (z-axis is longitudinally axial). This orientation necessarily constricts and prevents radial distension of the braid, but the preshrunk braid does have “give” in the longitudinal direction. However, when the braid is coated with a relatively elastic polymer to form the membrane, it is essentially longitudinally non-extensible (along the z-axis). In other words, the '444 membrane, whether pulled in the axial direction or pressured from within during backwashing, has essentially no “give”. The importance of “give” relates particularly to effective backwashing. The higher the backwashing pressure the better, if it does not damage the membranes, because such pressure allows faster and more effective cleansing of contaminated membranes and therefore provides an economic advantage.
Because the '444 braid is deliberately not embedded in the polymer, yarn defining the lumen (bore) of the membrane is not coated with polymer. Other references disclose braids woven to minimize the problem of too-deep penetration of the polymer film. The non-embedded yarn, in all such instances, is prone to damage such as pin-holes. Such damage lessens the initial high bubble point of the freshly deployed membrane. The “bubble point” refers to the pressure under which a stream of air escapes through the largest pore in a wall of a wetted, defect-free membrane which has desirable flux. Further, the importance of stability of the structure of the braid during operation, particularly the effect of shrinkage, was not known.
Though U.S. Pat. No. 7,861,869 discloses a semipermeable capillary membrane reinforced by a braid of monofilament yarn, the yarn is made by bundling multiple monofilaments (36 in example 1). The braid is not made by braiding separate monofilaments. Penetration of the dope into the braid is controlled so that the inner channel (lumen) of the braid is not blocked. The process taught herein prepares an “outer skinned” version of the reinforced membrane, explicitly avoiding embedding the braid.
WO-A-0397221A1 describes a tubular fiber membrane which is longitudinally reinforced by unbraided yarns, not by a braid. The axial bore is formed by injecting an internal coagulation solution in the centre but the thickness of the annular film defining the lumen cannot be controlled.
US 2009/0206026 A1 to Yoon et al, titled “External pressure type hollow fiber membrane having reinforcing supporter with monofilament for gas separation and water treatment, and method and apparatus for preparing the same” states: “The hollow fiber membrane of the present invention has excellent pressure resistance and high tension force by using the rigid and tubular supporter, an improved softness by using the monofilaments, and an increased bonding force between the supporter and the coating layer by increasing the concave-convexo degree of the reinforcing supporter.” (see '026 Abstract, lines 6-11). That the tubular supporter in the described hollow fiber membrane is rigid, is reiterated under “Industrial Capability” (see line 3 of paragraph [0057]. Such rigidity serves to distinguish the '026 membrane over the membrane of '607 to Mahendran et al, discussed in '026 as being the most relevant reference which teaches that “The support itself is so flexible (flaccid) that it does not have a circular cross-section and collapses with finger pressure.” (see Abstract, lines 4-6) “By “flaccid” is meant that the denier of monofilaments used in the yarns or “ends” for carriers which are braided, and the number of picks/unit length of the braid, are such that a tubular braid has very little mechanical strength in a vertical plane normal to its longitudinal central axis, so that it is so flexible that it can be easily manually tied into a knot. A typical braid starts out as multiple filaments which make up a single “end” and two “ends” are plied together in 3.8 twists/25.4 mm to make up a yarn or “carrier”. Multiple carriers, preferably 24, are used to braid a tubular braid.” (see '607, col 3 lines 24-33). Clearly, the '026 statement relating to a rigid and tubular supporter are meant to distinguish over the '607 braid.
Note that though FIGS. 4 and 5 in '026 purport to be photomicrographs of the reinforcing supporter, both woven with monofilaments of 130 deniers and of 32 and 24 yarns respectively, other than stating that the diameter of the supporter can be controlled according to the number of cones (see [0042]), there is no indication in either photomicrograph as to the diameter of the woven braids shown. Neither is there any identification, anywhere, either of the weave, or of the machinery, used to make a braid having any specified diameter, much less a nominal inside diameter in the range from about 1.0 mm (to make a membrane having a nominal outer diameter of 1.5 mm, depending upon the denier of the monofilament to be used), to about 2.5 mm (to make a membrane having a nominal outer diameter of 3.0 mm, depending upon the denier of the monofilament to be used), as specified for the braid and membrane claimed herein. In particular, there is no mention of using a flexible cable, dissolvable in an aqueous solution (referred to as “aqueous-dissolvable”) upon which to weave the braid. By “nominal” is meant “average”.
Particularly noteworthy is that the '026 membrane is woven with both, monofilament and multifilament yarns; this provides convincing evidence that the inventors of the '026 braid failed to realize that “whiskering” and “fuzz” were the root causes of failure in membranes with multifilament braids.
In FIG. 6 of '026 there is illustrated an automatic device in which a perforated wire 2 extends along the central vertical axis of an injector for an internal coagulating solution 4. A high pressure injection nozzle 3 injects the internal coagulating solution onto the wire, and the solution is also squirted through the perforations while the reinforcing supporter passes over and is forwarded by the roller 5 in contact with the wire 2. (see [0045]).
Aside from the problem of axially perforating about a 2.0 mm diameter wire, doing which is beyond the skill of the inventors herein, it will be seen in the test presented in example 1 below, that an open weave tubular braid having the diameter claimed herein, made with woven monofilament in the size range claimed herein, cannot be forwarded (or “passed”) over a wire as described in '026 because the friction is too great, and other reasons. Numerous attempts to forward a tubular braid of monofilaments only (see example 1 below) to make a membrane in the range of nominal outer diameters from 1.5-3.0 mm, fails to produce a usable, undistorted, uniform membrane. The '026 reference is therefore a non-enabling disclosure.
Moreover, manually pulling the braid over the wire after the braid is coated with coagulant polymer, results in destruction of the membrane, again, because of the flaccid membrane and its excessive friction. To make and use the membrane claimed in '026 would require undue experimentation.
US 2004/0197557 to Eshraghi et al teaches (a) providing a molten removable substrate material in the form of an extrudate of a molten polymeric material (see [0011] to make a hollow fiber membrane having a dissolvable core, and the use of reinforcing fibers as follows: “Additionally, one or more reinforcing fibers can be incorporated into such polymeric membrane to form a fiber-reinforced tubular polymeric membrane structure. Preferably, such reinforcing fibers extend continuously along the longitudinal axis of the fibrous core or substrate and therefore provide axial reinforcement to the hollow fibrous membrane. Fiberglass having an average diameter of about 0.1-500 μm is particularly suitable for practice of the present invention, while other fibrous materials, including but not limited to carbon fibers, metal fibers, resin fibers, and composite fibers, can also be employed for reinforcing the hollow fibrous membrane. The reinforcing fiber can either be co-extruded with one of the polymeric membrane-forming layers, or be encapsulated between two polymeric membrane-forming layers, to form an integral part of the hollow fibrous membrane.” (see [0085]) There is no suggestion, beyond the statement that “such reinforcing fibers extend . . . hollow fibrous membrane” how one or more reinforcing fibers are to be incorporated into the polymeric membrane.
The '557 publication states “the solid core fiber itself is formed of a solid-phase removable substrate material, and at least one layer of a polymeric membrane-forming material is coated directly onto such solid core fiber.” (See [0023]). It thereafter states: “the molten removable substrate material is co-extruded with the membrane-forming polymer.” (See [0041]-[0047]).
It is clear that Eshraghi et al did not extrude PVA because it degrades before it can be melt-extruded, irrespective of what grade of PVA is used. As evident in example 2 below, attempts were made to extrude each of three grades of PVA available from Kuraray, namely fully hydrolyzed (F-05 and F-17); intermediate hydrolyzed (M-17); and partially hydrolyzed (P-24, P-20, P-17 and P-05). The temperature at which each of the polymers degrades is lower than its softening temperature. Therefore each attempt resulted in severe degradation of each.
In the description of the process illustrated in FIG. 3A, the '557 publication states “a string or a tow of removable core fiber 122 from a spool 120 is passed through the extrusion die 124. A thin layer of the viscous polymeric solution 101 is therefore applied onto the removable core fiber 122, forming a coated fiber 132.” ([See 0086]) The core fiber 122 could not have been flexible PVA as it necessarily would have to be plasticized with just sufficient plasticizer to provide a core fiber which was not degraded.
A core made from PVA in a solution of hot water has insufficient strength to maintain its cylindrical form—discovered to be a critical requirement for making the open weave braid of this invention. Not being able to make a PVA core negated the suggestion in '557 that PVA may be used for the core.
As for the reinforcing, the '557 publication states “The reinforcing fiber can either be co-extruded with one of the polymeric membrane-forming layers, or be encapsulated between two polymeric membrane-forming layers, to form an integral part of the hollow fibrous membrane.” There is no suggestion that the extrudate be covered with a braid before being coated with membrane polymer, and no way this could be accomplished using the teachings of their disclosure.
In a manner analogous to that stated above, commercially available ethyl vinyl alcohol (EVOH) from Kuraray; commercially available polylactic acid from Nature Works; commercially available nylons from Shakespeare; and no-longer commercially available copolyester from Eastman, failed to produce a usable core despite numerous trials in each of which the conditions of extrusion were changed.
Clearly the disclosure of the '557 publication is not an enabling disclosure.
With respect to the use of monofilaments, apart from a braid thereof, the '557 publication provides no indication that it recognized the ill effects of “whiskering” and “fuzz” in a braid made with at least some multifilament yarns.
Neither did the '557 publication recognize that only a tubular braid embedded near the inside diameter of the membrane, so as to reinforce the lumen, provided the highest peel strength, bubble point and permeate efficiency. There is no suggestion that a braid be woven using only monofilaments woven in a particular way, namely with an open weave so as to avoid having circumferentially restricting filaments which would not have “give” under abnormally high backwashing pressure.
Publication No. WO/2010/148517 to Cote et al (hereafter the “'517 publication”) presents the concept of using a “dissolvable filament (solid or hollow) core” to make a hollow fiber membrane (see [0040]). It states that “the core can be a solid or capillary tube can be later dissolved in a solvent, preferably the solvent used to coagulate the membrane (typically water). Examples of water-soluble polymers include PVA, EVOH (made by Kuraray), as well as some forms of polyester (available from Eastman) and nylon (available from Shakespeare).” (see [0065]). Not mentioned is highly amorphous vinyl alcohol (HAVOH) and more commonly available polylactic acid (PLA), cellulose acetate, hydroxyethyl cellulose, polyethylene oxide (PEO) and polyethylene glycol (PEG), all of which are water-soluble. If he had used PVA as a removable core, he would have realized that despite extended washing with water, more than 10% of the usable pores in the lumen of the membrane remain clogged, and the membrane requires a wash with an aqueous solvent in which the PVA is far more soluble than in water. They would have disclosed such a complex cleaning requirement.
Currently used braided multifilament membranes are more readily susceptible to damage than they should be, resulting in leakage. The inventors herein discovered that such damage, resulting in leaks, typically occurs at vulnerable “thin spots” where yarns overlap near the surface of the braid; further, that such overlap of multifilament yarns or broken yarn is also conducive to formation of “whiskers” or “fuzz” which initiates pin-hole leaks, resulting in a low bubble point. Since in the '444 patent, thin film is deliberately restricted to the upper portion of the tubular braid, the uncoated lumen is formed and reinforced with unprotected, braided yarn, prone to whiskering. Such yarn tends to trap contaminant particles entering with the backwash and broken whiskers contaminate the permeate.
Still further, the relatively greater thickness of multifilament braid, relative to the thin film of polymer overlying the surface of the braid, results in a non-uniform thin polymer film having poor adhesion and a variable, low peel strength. In those instances where the lumen of the braid is deliberately coated with polymer film, the annular thickness of polymer film is uncontrollable (as evident in the '607, '869 and '075 patents inter alia). Though the '517 publication presented the concept of using a dissolvable polymer core such as PVA, what remained to be discovered was (i) how to overcome the degradation problem of PVA, yet make a reliably uniform core, (ii) how to make a flexible non-degraded PVA core which is strong enough to withstand the forces required to make an open weave braid snugly supported on the core, (iii) how to avoid trapping air in the membrane, (iv) how to overcome the problem of dissolving the PVA core within a reasonably short time, and (iv) how to ensure that upon solving the prior two problems, the resulting membrane would have unclogged pores. Clogged pores would greatly diminish the permeate efficiency of the membrane. The goal was to obtain higher permeate efficiency than that obtained with multifilament reinforced membranes, and to remedy the aforementioned problems of braided membranes exemplified by the '517, '444 and '177 membranes.
In addition, particularly because in a skein having reinforced fibers with a nominal diameter>1.5 mm, drawing a vacuum on the skein tends to collapse some fibers, it is also a goal to provide a monofilament reinforced fiber with reinforcement which improves on the strength of prior art reinforced fibers.
It has been discovered that using only monofilaments (“filaments”), and eliminating the use of multifilament yarn, produces an unexpectedly superior braided membrane whether its lumen is cylindrical or not cylindrical. Multiple filaments are woven (interlaced, plaited or braided) directly upon an aqueous-dissolvable solid core or “cable” of necessarily plasticized poly vinyl alcohol (“PVA”), plasticized with a plasticizer in an amount sufficient to yield a homogeneous, solid, flexible extrudate having a density+10% of the density of PVA and sufficient strength to provide an elongated, continuous, solid support for a braid of monofilaments woven on the surface of the extrudate. The PVA is preferably plasticized with from 5-20% by weight of a plasticizer chosen from polyethylene glycol (“PEG”), polypropylene glycol (“PPG”), ethylene oxide capped polypropylene glycol (“EO capped PPG”), sorbitol, glycerol, ethylene glycol, polyvinyl pyrrolidone (“PVP”), pentaerythritol, 1,4-monoanhydrohexitol, 1,4-3,6-dianhydrohexitol and copolymers of poly vinyl acetate. Without the critical amount of plasticizing, the cable would degrade when melt-extruded; would not have the flexibility to survive further processing, or, the requisite strength to withstand the forces of weaving from 6-20 monofilaments on the core's surface; and, if not essentially completely soluble, that is >99% soluble, in an aqueous cleaning bath, the membrane, when formed, could not be parted from the cable.
Whether the lumen is to be cylindrical or non-cylindrical, the cable is a unitary solid preferably formed by extrusion of the plasticized PVA as a continuous extrudate of arbitrary length, having a cross-section which matches that of the desired lumen to be formed by the extrudate. When cylindrical, the solid continuous cable has a diameter “dc” smaller than the nominal outer diameter of the membrane to be made (“dm”) by twice the thickness “dt” of the wall of the membrane; that is, dc=dm−2 dt, and dm is in the range from 0.75-2.6 mm, and dt is in the range from 0.2-0.6 mm. The flexibility of the cable is such that a plasticized cable with dc=2.0 mm can be wrapped around a cylinder 1 m in diameter at 25° C. without breaking.
When non-cylindrical, the lumen may have a scalloped cross-section of plural interconnected arcuate segments including adjacent, outwardly convex arcs, or segments of contiguous circles, interconnected, one to another at their circumferences. Interconnected, closely spaced apart arcs, defining lobes of a lumen, are preferably generated as portions of circles having the same diameter. The arcs are preferably symmetrically disposed about the x-y axes in a vertical plane which is at right angle to the longitudinal axis of the lumen. The minimal distance between adjacent arcs, even in a double-barreled lumen, is less than 25% of the diameter of the circles (and arcs) defining the lobes (see
A lumen with a scalloped cross-section refers to one having a periphery reminiscent of a scallop's shell, comprising plural interconnected circumferential segments. Though such a lumen may be formed by using plural cylindrical elements of arbitrary length, which have been fused, adhered or otherwise secured to each other longitudinally, at points along their circumferential lengths to form a fluid-tight bond, it is most preferred to form a solid unitary cable, with the desired number of radially protruding scallops or lobes, by extruding the plasticized PVA.
When the lumen's cross-section is formed by two interconnected circular segments to form two lobes, they preferably have the same diameter and are diametrically oppositely disposed relative to each other, the lumen so formed being referred to as a double-barreled lumen.
When the lumen's cross-section is formed by three or more, preferably from 3-6, most preferably 3-4, interconnected circular segments, they preferably have the same diameter and are radially equiangularly disposed relative to each other. In a lumen formed with three interconnected circular segments, the diameter of each is located 120 degrees apart relative to another contiguous circular segment, the lumen so formed being referred to as a three-lobe lumen.
In a lumen formed with four interconnected circular segments, the diameter of each is located 90 degrees apart relative to another contiguous circular segment, the lumen so formed being referred to as a four-lobe lumen.
The “lobe profile” of outwardly convex arcs of the lumen is chosen as a function of the desired ratio of (the depth of polymer “hv”, measured as the radial distance from the surface of the membrane) to (the nadir of the valley between convex surfaces, and “hc”, measured as the radial distance from the surface of the membrane to the apex or crest of a lobe's convex surface), or hv/hc. For all fibers having a scalloped lumen, the ratio of hv/hc>1. (see
The filaments are spirally woven on the cylindrical surface of the cable to form an open, tubular braid free of a restricting circumferential filament in the x-y plane. By “open” is meant that the braid has an air permeability much greater than 10 cc/sec/cm2 at 1.378 kPa because the braid has essentially no resistance to air passing through it, therefore ensures thorough infiltration and embedding of the braid in the liquid membrane polymer.
The average diameter of the cable is chosen to match the cross-section, more specifically, to match the hydraulic diameter of the lumen of the membrane to be made. The hydraulic diameter HD=4A/ρ where A is the area of the lumen, measured in very small increments by computer, and ρ is the linear measurement of the periphery of the lumen, measured in very small increments by computer. HD may be in the range from about 0.5-1.5 mm, and is preferably in the range from 0.75-1.5 mm.
Since the hydraulic diameter cannot be known until the lumen is formed, a membrane with a lumen having an average diameter must be made which will match the hydraulic diameter sought. The average diameter of the lumen is defined as: [(the outside diameter of the membrane, “o.d”) minus {twice (the thickness of membrane from surface to crest of lobe, “hc”) plus (the thickness of membrane from surface to nadir of a valley between successive convex arcs, “hv”)}] divided by 2, which may be written as: [o.d.−2(hv+hc)]/2 (see
The outside diameter of the cable is preferably in the range from 0.75 mm-2.5 mm; the average diameter of the lumen chosen to match a desired HD will depend upon the diameter of filaments used, the type of polymer used, the desired wall thickness of the membrane and other factors. The weave of the braid is such that liquid membrane polymer embeds more than 99% of the filaments including those in contact with the surface of the cable.
In practice, after coagulation of the dope, the cross-section of the membrane's lumen is slightly, from 1-10%, greater than the cross-section of the cable on which it was formed. Further, the monofilaments of the open weave occupy less than 50% of the cylindrical surface of the cable, preferably less than 25%, depending upon the diameter of the filaments and the spiral angle of the weave. A typical tubular braid, outer diameter 2.0 mm, which is circumferentially continuous, collapses into a shapeless braid if the cable upon which the braid is woven, is removed before the membrane is formed. Such a braid has no shape, and has no meaningful cylindricity.
Typically, the braid when woven and supported on the cable, has a cylindricity of >0.8. The braid is woven using from 6-24 filaments of the same diameter, each typically having a density in the range from 0.9-1.5 g/ml and a denier in the range from 25-250 denier, approximately equivalent to a diameter of from 50-160 μm. If there is an overriding reason to do so, a mixture of coarse “reinforcing” and relatively fine “mesh” filaments may be used.
The braid, as a monolayer of monofilaments snugly overlying the cable, in turn, provides support for a tubular film of polymer film formed after the braid is infiltrated with dope and coagulated. The membrane so formed has unexpectedly greater peel strength, durability and a higher bubble point than a multifilament braid of comparable nominal diameter. The term “monolayer” refers to a single layer of filament in which layer the only filaments overlying one another are those at locations where two filaments are interlaced over one and another. Interlaced filaments are movable relative to one and another until embedded in coagulated polymer. Except in some instances in which the open weave is relatively closely woven and the thickness of the membrane is only slightly greater, that is, <5% greater than twice the diameter of the reinforcing monofilament used, the volume of the monolayer of filament occupies a minor portion (that is <50%) of the annular volume of the membrane.
A typical monofilament-reinforced membrane has a wall thickness in the range from about 0.2-0.6 mm generated by maintaining a ratio of hv/hc>1, preferably in the range from 1.2-5.
There is no acknowledgment in any reference that producing an open weave braid of the claimed diameter, with monofilament having the aforestated diameter, while the braid is supported on a scalloped cable chosen to match the desired diameter of the lumen of the membrane to be formed, requires modified spinning machinery not known by the inventors to be available in the prior art, anywhere.
The hollow fiber asymmetric microporous membrane (“membrane”) produced as described herein has a relatively thick wall of polymer film, preferably >0.2 mm but <0.5 mm thick, within which the monolayer is embedded near the lumen. A membrane so produced avoids problems associated with braids woven with any multifilament yarn irrespective of the polymer used for the membrane.
A tubular braid having a unique, open, weave (referred to as “open-weave braid”) is woven with multiple monofilaments each in an axially longitudinal helical configuration, in a monolayer with a rhomboid-shaped or diamond pattern (as in a playing card), as shown in
The monolayer can only be embedded so as to define, in combination with coagulated polymer surrounding the monolayer, the lumen of the membrane. The position of the embedded monolayer relative to the polymer above it, can only be manipulated by changing the nominal outer diameter of the membrane.
More specifically, the monolayer of monofilament in the membrane provides rhomboid-shaped zones corresponding to the open-weave braid. The width of each zone is defined by the arcuate (because it is circumferential) rhomboid area within the recurring area confined by intersecting filaments. Since the filaments typically occupy less than 25% of the cylindrical area of the cable, the resulting relatively large filament-free radial zones of polymer formed, provide better (greater) filtration into the lumen than that obtained with a braided multifilament yarn in a comparable prior art membrane. The membrane has an adhesive strength>10 Kgf/cm2, a bubble point>2 bar, and a percent rejection of 0.03 μm polystyrene bead>90%, and more preferably the membrane has an adhesive strength>15 Kgf/cm2, a bubble point>4 bar, and a percent rejection of 0.03 μm polystyrene bead>95%.
Further, high permeate rates are maintained for longer than in a comparable multifilament membrane.
A process is described for embedding a monofilament, tubular open-weave braid of monofilaments comprising, weaving an open tubular braid having a recurring rhomboid pattern of synthetic resinous monofilaments in the denier range from 25-250, directly over the surface of, and snugly contacting a core cable (“cable” for brevity) of poly vinyl alcohol plasticized with from 5-20 weight percent of plasticizer, the cable having a scalloped profile with an average diameter chosen to provide a lumen of desired average diameter in the range from 0.5-2.0 mm; coating the tubular braid with a membrane-forming dope in a coating nozzle until the dope infiltrates into an area below the surface of filaments overlying the cable to form the lumen; pulling the cable and braid together through the coating nozzle; coagulating the dope to form a semipermeable thin polymer film in a membrane embedding the braid as a monolayer which, together with polymer surrounding the braid, defines the lumen of the membrane; washing in hot water until at least 99% of the plasticized PVA is removed, further washing with an oxidizing agent chosen from sodium hypochlorite (NaOCl), hydrogen peroxide and potassium hypochlorite (KOCl) to make a membrane which has a total organic carbon (“TOC”) of <0.5 ppm without damaging the lumen of the membrane formed.
From the foregoing it will be evident that having first made the discovery that “whiskering” and “fuzz” (associated with multifilaments) were the root cause of membrane leakage, what still remained to be discovered was (i) how to make the membrane having the diameter stated, (ii) how to position a monolayer of monofilament so as to define the scalloped lumen of a thin membrane, (iii) how to control the thickness of membrane polymer both above and beneath the monolayer of braid, (iv) how to melt-extrude a PVA cable having the stated diameter with physical properties suitable for the task at hand, and (v) how to rid the membrane of essentially all, that is >90% of trapped air and plasticized PVA and contaminants so that the TOC of the membrane is preferably below 3 ppm.
The complex nature of the solution to the problem cannot reasonably be disregarded in any attempt to use the teachings of the foregoing references as an enabling disclosure.
The foregoing and additional objects and advantages of the invention will best be understood by reference to the following detailed description, accompanied with schematic illustrations of preferred embodiments of the invention, in which illustrations like reference numerals refer to like elements, and in which:
Supporting the braid as it is woven on the cable:
Referring to
The location of the braid relative to the cable, in the polymer defining the lumen formed when the cable is dissolved, is thus fixed. Its location relative to the wall thickness of the membrane, for a stated diameter of the lumen, can only be manipulated by increasing or decreasing the nominal outer diameter of the membrane.
It is seen that the filaments 12 overlap at intersections 15 of the embedded filaments 12 of the braid over which the membrane 20 is formed on the cable 30. When the braid of filaments 12 is woven over the cable 30, there are clearances 14 and 14′, which range in diameter from 5 μm-0.1 mm or more, depending upon how closely the surfaces of the filaments, or whether the filaments overlie a valley 19 between contiguous lobes. The amount of such clearance depends upon the number of lobes of the cable, its diameter and other factors. Polymer infiltrates the clearances and when coagulated forms the film 11.
Though an open-weave filament is expected to be weaker than a prior art weave of multifilament, the resulting open-weave membrane retains its tubular configuration without collapsing under suction pressure exerted during filtration, and has excellent peel strength of at least 15 kgf/cm2. The braid, with the cable removed, has an air permeability >(greater than) 100 cc/sec/cm2 at a ΔP of 1.378 kPa. The monofilament construction of the braid ensures stability and a minimal moisture regain, much less than that of a comparable multifilament braid; and the unique open-weave of the braid 10 makes it unnecessary to preshrink it to ensure its stability.
Referring to
The dope is prepared by mixing from 10 30% by weight of the PVDF with from 70˜90% by weight of N-methylpyrrolidone (NMP) in a dope tank 40 in which the dope is blanketed with an inert gas, preferably nitrogen, from cylinder 41. The dope may be prepared by any conventional method and may include suitable additives, if needed. The dope is prepared by mixing 20 wt % of PVDF (product name: Solef 1015) and 80% of N-methylpyrrolidone (NMP) at 60° C. at a temperature in the range from 30° C.˜100° C., preferably 40° C.˜70° C. so that the viscosity of the dope is in the range from 30,000 cps˜60,000 cps at 30° C.
The covered cable SC is fed from unwinder 23 and over guide rolls 24 and 25 to a coating nozzle 50. Cable 30 is an extrudate of PVA plasticized with 10% glycerine. The cable has a diameter of 0.75 mm; the filaments are woven at a weave angle of 35° to sheathe the cable 30 with the open-weave tubular braid.
The dope is metered through coating nozzle 50 to produce a 400 μm thick film with the braid embedded in the bottom of the film. The dope is then coagulated at a temperature of 30˜50° C. in an aqueous coagulation bath 60 and fed over guide rolls 61 and 62 to a cleaning bath 70. The wash water in a first cleaning bath 70 is maintained at a temperature of 40˜80° C. for from 0.5˜1.5 min to dissolve and wash out the residual NMP from the membrane.
The membrane 20 on cable 30 leaves over guide roll 73 and is cleaned in a second cleaning bath 76 maintained at a temperature of 60˜80° C. after which the cleaned membrane leaves under guide roll 74, and is captured on a winder 80 in a winder bath 81 of diluted 50% aqueous glycerine. The diluted glycerine prevents an upper layer of wrapped membrane from sticking to a contiguous lower layer. The purpose of the winder is to store the washed membrane and the cable still supporting it, until the membrane can be cut into short sections, approximately the length desired for building a desired module, and freed from the plasticized PVA cable.
Accordingly, as illustrated in
When the concentration of PVA in the wash water leaving tank 27 is <0.5% the bundle 26 is removed from the tank 27. Because too many pores of the membranes are still clogged, the bundle 26 is mounted horizontally in a second pore-cleansing tank 28 into which an aqueous solution of from 0.1-0.5% NaOCl solution at from 20° C.-80° C., preferably from 40° C.-60° C., is introduced to remove the remaining PVA and other contaminants which would restrict flow of permeate into the lumen. The solution is continuously recycled by pump 29 through piping 31 overnight, then drained through drain piping 33. The bundle of membranes, each with a lumen having a diameter of 0.8 mm, now free of PVA and other contaminants which clogged the pores of the membrane, is transferred to a module-building facility.
The monolayer of braid 10 is thus embedded in coagulated PVDF film 11 which has excellent permeability and is essentially insoluble in water. The polymer acquires a pattern of rhomboid areas 13 (see
The diameter of plasticized PVA cable with a scalloped periphery is chosen according to the desired diameter of the lumen (inner diameter of the membrane). Typically, the cable, whether one or more, has an average diameter in the range from 0.1˜1.8 mm, preferably 0.5˜1.5 mm, to provide a braid reinforced membrane having an average wall thickness in the range from 0.2-0.5 mm. The resulting lumen is non-circular and the wall thickness of the membrane formed is non-uniform.
Preferred braided membranes have a bubble point of at least 2 bar. The novel membrane has an adhesion strength of more than 15 kgf/cm2 typically from 12 to 20 kgf/cm2.
The open-weave monolayer braid 10, embedded in the polymer film 20, unexpectedly provides excellent permeability and resistance to damage. The embedded monolayer eliminates a “whiskering” problem common to braids woven with one or more multifilament yarns.
The recurring open areas in the open-weave braid provide circumferential, interconnected rhomboid or diamond-shaped loops 13 lying in the vertical (as shown in
The rhomboid pattern 13 is retained when the intersections of filaments 12 are welded. The pattern may be more close-knit (not shown) so that it provides a membrane with smaller radially open passages to the lumen, each of smaller area relative to an area defined by the relatively loosely knit braid (shown).
Prior art braids, woven with multifilament yarn, have at least some of the yarn forming an essentially continuous circle in the x-y plane, thus resulting in constricting any radial distension of the braid. Prior art tubular multifilament braids are therefore inelastic in the radial direction. This constricting property is retained even when the braid is embedded in an elastic polymer film. Pressure exerted from within a prior art multifilament tubular braided membrane, cannot and does not increase the diameter of the braid, thus making it susceptible to damage. In contrast, in addition to the longitudinal extension one would expect of the elastic properties of a polymer-embedded, open-weave monofilament reinforced tubular membrane, such properties allow the membrane to expand or distend radially, when sufficient pressure provides a radial distending force. Consequently, a relatively higher pressure than normally used, sufficient to distend the membrane but insufficient to damage it, may be used to backwash the membrane.
Duplication of feeding a monofilament braid as described in the Yoon et al publication No. US 2009/0206026:
A monofilament braid of 100 denier (0.1 mm) nylon monofilament was spirally woven on a 1.0 mm cable of metal wire, alternate filaments being woven at opposed spiral angles of 30°. The sheathed cable was placed on a 2.54 cm diameter rubber roller rotating at 30 rpm. The braid was crumpled on the cable and could not be advanced. The speed of the roller was reduced to 15 and then 5 rpm. In no case was the braid advanced without damage. The roller speed was then increased to 40 rpm. The braid was crumpled.
The following three grades of PVA available from Kuraray were each melt extruded in a single screw Hankook Model M-65 extruder fitted with a 65 mm diameter screw having a length/diameter ratio of 22. The barrel temperature is 195° C. and the die temperature 160° C. The die is provided with 18 scalloped through-apertures (holes) and the average diameter of each aperture is 1.6 mm. The air quenching length for the PVA cable is 2 m in 25° C. air for 2 seconds. The drawing ratio is 1.5:1.
A braid is formed by weaving 12 filaments, each of 100 denier nylon, at a spiral angle of 35° over a plasticized PVA cable having a diameter of 0.75 mm using a custom-built braiding machine. The sheathed cable is pulled through a coating nozzle into which a dope, prepared as described above to have a viscosity of 43,000 cps at 30° C., is flowed at an output rate of 11 g/min. The dope infiltrates the braid, coats the cable and embeds the braid. The membrane is coagulated in a water bath at 45° C. and washed as shown in
The physical properties of the membrane made in Example 3 above, are given in Table 1 below.
In a manner analogous to that described in Example 3 above, a dope, prepared as described above to have a viscosity of 43,000 cps at 30° C., is flowed at an output rate of 16 g/min a braid is woven at the same spiral angle, over a cable having a diameter of 1.1 mm using 12 nylon monofilaments, each of 100 denier (0.1 mm) to yield a membrane with a 1.25 mm lumen, and a nominal outer diameter of 2.05 mm.
In a manner analogous to that described in Example 3 above, a dope, prepared as described above to have a viscosity of 43,000 cps at 30° C., is flowed at an output rate of 19 g/min a braid is woven at the same spiral angle, over a tri-lobe cable having a average diameter of 0.85 mm using 12 nylon monofilaments, each of 100 denier (0.1 mm) to yield a membrane with the average diameter of the lumen 0.93 mm, and a nominal outer diameter of 1.85 mm.
In a manner analogous to that described in Example 3 above, a braid is woven at the same spiral angle, without using a cable, with twenty four PET multifilament yarns each 300 denier/96 filament (a single filament is superfine, about 3 denier) and having an inner diameter of 0.85 mm; the braid was embedded in the same polymer solution to provide a wall thickness of 650 μm (0.65 mm, but the membrane film thickness is about 100 μm).
Evaluation of Physical Properties:
(i) A membrane having a length of 200 mm is folded in half and inserted in an acrylic tube having a diameter of 10 mm and a length of 100 mm. At one end of the tube, the membrane, near both of its ends is sealed with epoxy (or urethane) leaving the lumen in each end open. The other end of the tube is left open and mounted in a water permeability testing apparatus.
(ii) Distilled water held under pressure, is discharged from a pressurized vessel into the tube at a pressure of 1 bar (14.7 psig or 100 kPa) so as to force water through the wall of the membrane and discharge the permeate into a collection beaker. The water permeability is obtained by weighing of the permeate collected over a specified period of time.
(i) a fiber 50 mm long is inserted for 10 mm of its length near one end, into a 10 mm inside diameter polyurethane tube 50 mm long.
(ii) the 10 mm of fiber in the polyurethane tube is potted using epoxy (or urethane)
(iii) 10 mm of the other end of the fiber is wrapped with paper so as not to damage it and the wrapped end is inserted into one of the jaws of an Instron (UTM) tensilometer, the gage length of which was set at 70 mm. Any material providing suitable gripping without damaging the membrane may be substituted for the paper. When the tube is secured in the other jaw of the machine, the fiber is to be taut so as not to be suddenly elongated when the Instron is in operation.
(iv) The crosshead speed was 50 mm/min, The maximum tensile stress is divided by its unit area, so the maximum tensile stress is registered as the adhesion strength.
The average elongation at break is registered as the elongation.
(i) Use the same membrane sample prepared for the water permeability test, mount the tube in the water permeability testing apparatus, then immerse the tube including the membrane in a water bath.
(ii) The testing apparatus is connected to a source of nitrogen under pressure and, with the membrane immersed, the tube is pressurized with nitrogen.
(iii) Adjust the pressure to the tube in stages with a pressure regulator, through a range from 0 bar (atmospheric) to 8 bar, holding the pressure in increments of 0.5 bar for 60 seconds.
(iv) Record the pressure when a nitrogen bubble forms on the surface of the membrane, or alternatively, the pressure at which the membrane ruptures.
(v) The recorded pressure is the bubble point.
UV [using a Perkin Elmer Lambda 25 UV/vis spectrometer]
(i) preparing two strands of the hollow fiber membrane having a length of 100 mm
(ii) inserting the membrane in an acrylic tube having a internal diameter 10 mm and a length of 100 mm; sealing one end of the membrane with paraffin (or urethane). The other end of the membrane is potted in the acrylic tube to prepare a sample.
(iii) mounting the sample in a water permeability testing apparatus
(iv) preparing a feed solution for measuring rejection ratio, as follows:
A styrene bead dispersion was prepared by mixing polystyrene beads having a size of 0.03 μm in thrice distilled (or reverse osmosis) water to provide a dispersion having a concentration of 0.05% polystyrene beads, with enough surfactant to prevent styrene beads from sticking together, and agitating the mixture for 1 hr.
(v) pouring the dispersion into a pressure vessel and under a pressure of 0.5 bar, forcing the polystyrene beads through the membrane and
(vi) collecting the permeate over a period of 1 minute.
(vii) setting a reference line for the thrice distilled water using a UV-Visible spectrometer and measuring the absorbance of the feed solution, then measuring the absorbance of the permeate collected.
(viii) The % rejection can be obtained by using a UV-Visible spectrometer and be calculated by the following formula:
Rejection (%)=(1−Cpermeate/Cfeed)*100
Cfeed: absorbance of the feed solution:
Cpermeate: absorbance of a sample passed through the membrane
A rejection ratio 90% or more is deemed useful and pore size of the membrane can be estimated indirectly by using dispersions of beads having various diameters ranging from 20 nm-100 nm.
It is evident from the Table 1 above, that the pore sizes for each of the membranes are essentially the same, as one would expect. However, the water permeability of the membrane with multifilament yarn in the braid, is only 75% of the membrane with the monofilament braid, its bubble point is lower than 33%, and its elongation at break is 66% lower than that of the membrane with the monofilament monolayer braid.
Further, the data for the membrane with the tri-lobe lumen, Example 5, show a bubble point about 20% higher than those obtained with membranes having circular lumens, Examples 3 and 4.
Equal lengths (1 m) of a membrane made as described in Example 3, 4 and 5, and a membrane made with a multifilament braid described in the Comparative Example above are dried so as to contain less than 1% by weight of water. Each membrane was then soaked in 30% glycerine solution for 24 hours and dried at 30° C. convection oven for 4 hours and weighed. The membranes were thereafter soaked in water for 24 hours, then weighed. The results are given in Table 2 below.
It is evident from the results above, that the multifilament braid retains more than double the amount of glycerine, and about 43% more water than the monofilament membrane. Such increased weight is magnified when several thousand membranes are assembled in a module, making it more difficult to insert and remove modules in a purification system.
Particularly with respect to the efficacy of removal of the plasticized PVA in the membranes before they are assembled into modules and placed in service, all three membranes test routinely for <0.5 ppm TOC being typically <0.3 ppm TOC using the prescribed KWWA (Korea Water and Wastewater Works Association) F 106 test. This confirms that essentially all the plasticized PVA has been removed.
Having thus described the monofilament membrane having a scalloped lumen, and the process for making the membrane, in detail, and illustrated both with specific examples of the best mode of each, it will be evident that we have provided an effective solution to an unrecognized problem. It is therefore to be understood that no undue restrictions are to be imposed, and our invention not restricted to a slavish adherence to the details set forth herein.
Number | Date | Country | Kind |
---|---|---|---|
10-2011-0039181 | Apr 2011 | KR | national |
10-2011-0134597 | Dec 2011 | KR | national |
PCT/KR2012/003189 | Apr 2012 | KR | national |
This application is a continuation-in-part application of Ser. No. 13/338,557 filed 28 Dec. 2011.
Number | Date | Country | |
---|---|---|---|
Parent | 13659648 | Oct 2012 | US |
Child | 16101369 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 13338557 | Dec 2011 | US |
Child | 13659648 | US |