EP0076321A1 - Method and apparatus for high-efficiency ultrafiltration of complex fluids - Google Patents

Abstract

Method and apparatus for continuous ultrafiltration of complex fluids having components which degrade the characteristics of the filter. The improved apparatus and method involves using one or more techniques to increase the filtration efficiency. The method and the apparatus allow the separation of components of intermediate dimensions or weights of the complex fluid. Using these improvements, continuous filtration of complex fluids and the concentration of intermediate components can be accomplished.

Description

METHOD AND APPARATUS FOR HIGH-EFFICIENCY ULTRAFILTRATION OF COMPLEX FLUIDS

BACKGROUND OF THE INVENTION

1. Field of the Invention

The invention relates general ly to improved efficiency ultrafiltration of a fluid having a broad range of component size distribution and, in addition, to removal of an intermediate-size component from a fluid having a complex composi tion with components which impede filtration.

2. Related Application

This appl ication is in part elosely rel ated to the copending application " Method and Apparatus for Treating Blood and the Like", filed on even date herewith, and having a common inventive entity.

3. Description of the Prior Art

Historically, filtration comprehended removal of the larger particles or compounds from a feed fluid by its passage through a porous f i lter element. Where the feed fluid flow is solely perpendicular to the local plane of the filter, the process is called batch filtration, and the filter eventually clogs due to accumulation of large particles or compounds in or near the pores. When clogging occurs, the filtration becomes inefficient in the sense that either substantially greater pressure drops are required to maintain a given fi ltrate rate or the initial filtrate rate cannot be approached. Inefficiency can also result from osmotieally-derived back pressures in cases where the feedstock fluid comprises a solution of low to moderate molecular weight compounds, some of which are rejected by the filter membrane (i .e. concentration polarization).

A known alternative to batch filtration is to pass at least a portion of the feed fluid parallel to the local plane of the filter. This technique is useful, for example, where the fluid is a mixture of two substances consisting of molecules of different sizes (size being approximately proportional to the molecular weight, MW) the larger of which is substantially rejected by the filter membrane.

Molecular size is collected through A circulation path parallel to the filter membrane while the filtrate comprising the fluid of smal ler melecular size is col lected after passage through the membrane under the influence of a transmembrane pressure di f ference. A second exampl e i s a feed fluid containing cells such as bacteria, white blood cells, red blood cel ls, platel ets, foodstuffs, etc. (not necessari ly in combination) in a solution or solvent where a process requires the separation of the cel l s f rom sol ution or solvent .

Subsequent steps may include the further separation of the compounds i n solution as di scussed therein before and hereinafter. We use the term convective filtration where there is substantial flow parallel to the filter.

For a compl ex feed f luid compr i sing, for exampl e a combination of solutes, suspensions, and perhaps fluids of different molecular sizes, convective and batch filtration both have clogging problems resulting in inefficiency. As wil l be discussed in more detail hereinafter, a second major problems is that the molecular weight distribution of the filtrate may also change with the ineff iciency. Where a filter is required to discriminate at a predetermined molecular or particle size, control over the process is tenuous. Thus, increases in transmembrane pressure to counteract the clogging proclivity and maintain acceptable filtrate rates leads to a modif ied MW distribution in the filtrate which is not acceptable for many applications.

As a single example of a fluid for which fil tration techniques are presently inadequate, consider human (or animal) blood. As shown in Figure 7, blood comprises an extremely brosd molecular weight (MW) distribution of components alone with a distribution of cell sizes. As described in the copending appl ication cited hereinbefore, treatment of a kidney patient requires accurate removal of a minor fraction low and middle molecular weight species at relatively low filtration pressures. Similarly, other common fluids have filtration problems which may be ameliorated by the methods and appraratus described in this application.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide improved apparatus and methods for the efficient ultrafiltration of fluids having a broad range of component sizes whereby reasonable f iltrate rates and stable filtration characteristics may be achieved.

It is the further object of this invention to provide improved apparatus and methods for the efficient ultrafiltration of fluids having a broad range of component sizes whereby reasonable filtrate rates and stable filtration characteristics may be achieved in a continuous filtration process.

It is yet another object of this invention to provide an apparatus and a process for the removal of an intermediate fraction from a complex fluid having a broad range of component sizes by a filtration method and apparatus providing for two convective filters of different rejection characteristics.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

In accordance with one embodiment of this invention, efficient ultrafϊltration is achieved by passing the feed fluid through a first convective filter having an intermediate cutoff characteristic and then passing the fi ltrate from the first convective fi lter through a second convective filter having a smaller cutoff characteristic, whereby a heavy fraction i s obtained from the convective output of the f irst f i lter, an intermediate fraction is obtained from the convective output of the second filter, and the l ight fractin is obtained in the filtrate of the second filter.

In accordance with yet other embodiments of this invention, spectra and pressure-efficient f iltration is achieved in convection filters by geometrical and operational augmentation techniques.

In accordance with yet another embodiment of this invention, at least one of the convective filters has a spiral geometry to help prevent clogging of the filter.

In accordance with yet another embodiment of this invention, there is disclosed a method and apparatus for recirculation of any output fraction through the convective filter to improve its efficiency.

The foregoing and other objects, features, and advantages of the invention will be apparent from the following, more particular description of the preferred embodiments of the invention, as illustrated in the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWING

Figure 1 shows the rejection characteristics of exemplary filters for use in this application.

Figure 2 shows schematically a portion of a filter system for removal, as an example, of middle molecular weight material from a feed fluid.

Figure 3 shows the f i ltrate rate versus pressure characteristics of a filter suitable for removal of low and middle molecular weight materials from a fluid.

Figure 4 shows the clearance or removal rate versus the molecular weight of disolved species as a function of the quantity of rejected materials (e.g. very high molecular weight proteins and/or cells) present on the filter membrane surface.

Figure 5 is a schematic representation of an ultrafiltration system suitable for fluids having a broad range of component distribution.

Figures 6A, 6B and 6C are various views of a filter configuration suitable for use with the present invention.

Figure 7 depicts important components of human blood. including waste materials, as a function of their molecular weight or cell size, as an example of a complex feed fluid. DETAILED DESCRIPTION

Many biological and industr ial f lui ds cons i st of a substantial number of components of widely differing sizes or molecular weights (MW). By way of example, the complex spectrum of human blood is shown in Figure 7, where it may be seen that an extremely large number of identif iable components span a molecular size range substantially in excess of five orders o magnitude. In order to filter such a complex fluid to remove a preselected portion of the spectrum, one must pay careful attention to not only both the gross spectral characteristics of the filter(s) but also to the tendency of certain components to clog the filter(s), rendering them pressure inefficient and/or leading to spectral degradation of the filters. Inabi l i ty to solve these problems for blood has precluded, for example, the practical development of an efficient ultrafiItration system for use as an artifieal kidney. Similar problems exist with respect to a great many other complex fluids.

In the so-called batch filtration process, the feed fluid is passed essentially normal to the plane of the filter. For fluid such as water containing suspended sand, batch filtration is feasible as a semi-continuous process because the water can continue to flow through the sand which builds up on the uostream side of the filter and filtration continues with only moderate increases in pressure. Batch filtration is not suitable for continuous use with a eomolex fluid such as whole blood because the larger constituents effectively clog the Miter and engender large pressure increases for a given filtrate rate. Filtration of such a complex fluid can be made more efficient and continuous by the use of a convective filter where the feed fluid flows approximately parallel to the filter membrane and thus tends to carry off those constituents of the fluid which decrease the filtrate rate. The required flow of filtrate perpendicular to the filter membrane can still cause problems of clogging.

The clogging referred to herein can be two types: surface clogging and membrane pore clogging. Surface clogging is caused by rejected materials which accumulate on the surface (feed fluid side) of the filter membrane. The amount and/or density of this type can be controlled by the methods, devices, and procedures described or referenced in this disclosure. The second type of clogging refers to fluid constituents becoming immeshed within the membrane ultrastrueture. This type is, in general, less affected by convective events within the feed channel although there is still possible minor contribution from events within the feed channel. The basic membrane filtration characteristics would be altered in the latter case wherein a different straight line buffered saline limit could be encountered (e.g. the straight line of Figure 3 would be rotated clockwise). The initial technical concepts on membrane pore clogging as well as other limit phenomena were nresented in the paper "Quantitation of Membrane-Protein-SoIute Interactions during Ultrafiltration" in Transactions of the American Society for Artifical Internal Organs, Vol. 24, Pg. 155, 1978. A more generalized and complete description of multiple limit phenomena supplemental to this disclosure was published in July, 1980 as the chapter entitled, "Ultrafiltration of Plasma and Riood": in the book Adyances in Biomedical Enginnering, Part II, entited by D. O. Cooney (Marcel Dekker, Inc., New York and Basel).

Figure 3 shows how surface clogging affects the efficiency of filtration through its influence on the filtrate rate versus transmembrane pressure relationship. In that figure, filtrate rate is linearly proportional to pressure for a "buffered saline" solution. However, when proteins similar to those found in blood are added to the feed fluid, linearity fails and the pressure deviates from the ideal limit very dramatically at and above a certain filtrate rate determined by the nature of the feed fluid, the filter membrane, and the flow conditions (as examples). Figure 3 also shown graphically the definition of efficiency used herein; efficiency is the ratio of the filtrate rate, N_B, on

the non-linear curve to the filtrate rate, NA ' on the linear buffered saline curve at the same transmembrane pressure. Note that the efficiency decreases as the pressure is increased. Low efficiency conditions at high pressures can result in gel or precipitate formation on the membrane surface as denoted by the indicated forbidden operational area.

Heretofore, attempts to treat certain complex fluids such as blood by ultrafiltration have been either unsuccessful or of limited success because of inefficient filters. There are two major problems. First, while efficiency may be enhanced by moving from say point D to point B in Figure 3 by reducing the pressure and filtrate rate, the rate becomes unacceptably low and may only be increased by making the filter larger. In know configurations of filters for blood applications, area enlargement increases the total amount of protein deposition and in multichannel designs aggravates the degradation of filtrate rate with time due to concentrating effects. As regions of the filter begin to become ineffective, either the filtrate rate drops or the transmembrane pressure (TMP) increases. The second major problem is that when the filter is operated inefficiently, the composition of the filtrate is modified. This is illustrated by Figure 4 where it may be seen that as the conditions change from points A (no protein) to E, C, and D ( i ncreased pressure, protein deposit, and density) in Figure 3 the filtrate includes less and less of the middle mo l ecu l ar we i gh t s pec i es . I n t he cas e of t he ki dney appl i ca t i on , for example, the clearance rate can drop so low (e.g. curve D) that conventional hemodialysis rates (shown for comparison purposes) are more efficacious than hemofiltration clearances.

Figures 3 and 4 typify the problems of maintaining efficiency and spectral integrity in a convective filter for separation of a fluid into two fractions. It is often the case that the end product of the filtration is an intermediate fraction, which then requires a second filter having different characteristics. lt is one of the major features of the present invention to use two or more convective ultrafilters to achieve separation of the intermediate fractions, whereby the intermediate fraction may be removed in a continuous process. Referring now to Figure 1, there is shown the generalized rejection characteristics of two different filter membranes. The type II membrane rejects the heaviest (or largest) particles while passing the intermediate and light fractions, while the type I membrane rejects both the intermediate and largest components and passes the l ightest components. Referring now to Figure 2, if two convective filters are interconnected as shown, the light and intermediate fractions are the filtrate of the primary filter (containing the type II membrane); these fractions are then separated in the secondary filter (containing the type I membrane). In the general case shown, all these fractions are available separately as the outputs of a continuous process. Depending on the components present in the filtrate of the primary filter, there can be a filtration efficiency and spectral integrity problem in the operation of the secondary filter as well as in the operation of the primary filter, as hereinbefore described. Various augmentation techniques presented in more detail hereinafter can ameliorate these problems in one or both of these filters.

In order to achieve and maintain efficient ultrafiltration through either or both of the filters in Figure 2, they must be especially configured and operated using one or more forms of augmentation herein defined as:

(1) surface perturbations in narrow flow channels

(2) irregular but controlled channel geometries

(3) membrane charge characteristic (repellant)

(4) secondary flow induction by channel inserts (screens, ribbons, etc.)

(5) externally applied forces and/or motions (physical movement, ultrasound, electrical potent ial , pressure perturbations, pulse flow, etc.)

(6) staging of devices

(7) independent manipulation of flow rates in the device

(8) preferred geometries in combination with augmenting methods

(9) independent control of biochemical and biophysical conditions during filtration.

Referring now to Figure 6, various views of portions of suitable filter are shown. A complex feed fluid such as blood FF passes through the length L of the filter between the membranes 90. Elements 200 schematically represent a screen which serves to separate the membranes elements 90 by an appropriate distance, to introduce some resistance to flow into the feed fluid path (whereby uniform flow is obtained) and to induce secondary flows which help keep the membrane clean. The model shown contains the membrane cast on a backing 400 sufficiently porous to allow easy flow of the filtrate towards the permeate collecting tub (500). For the kidney machine, the total area of the membrane 9 on Figure 5 is desirably on the order of 0.7m² for average adult intermittent applications. The height H of the feed fluid flow path is desirably in the range 0.25 to 1 mm; too small a value introduces excessive resistance into the feed fluid flow path while too large a value results in ineff icient fi ltration conditions and an impractically large filter.

In order that the convective filters achieve and maintain efficiency, it is imperative that any impediments in the convective path do not appreciably reduce the effective width of the channel (i.e. active membrane) below its nominal value W. For example, if the f ilter consists of multiple hol low fiber membranes in a parallel arrangement, each with a bore diameter U, rapid plugging of a substantial number of the fibers can occur due to feed fluid concentration , and the effective area is unacceptable diminished. Referring to Figure 6A i f a local impediment occurs in the channel, the feed fluid must be able to continue to flow both upstream and downstream of the impediment.

A rough geometrical criterion for such a condition is that W should be at least as large as L. This requi rement is most easi ly met by spiral filters, which are al so compact and relatively easy to fabricate. deferring now to Figures 6B and

6C, there is shown a cross-section of a spiral f i l ter. The membrane 9 (Figure 5) comprises an envelope with the backing 400 from two opposing membrane elements 90 in contact 99 and glued together at the outer edges 66. The envelope and the blood screen 200 are both wound around a central hollow mandrel 500 which serves as a conduit for the filtrate stream. The porous backing 400 from envelope 9 opens only onto holes 300 leading to the hollow portion of the mandrel 500; the filtrate stream passes from the filter unit through the filter perpendicular to the drawing. More details of the construction of a spiral filter may be found in the Westmoreland U.S. Patent 3,367,504, which describes its use for the desalinization of sea water.

Several different combinations of spiral wound construction have resulted in achieving the high efficiency necessary for this application. In looking at the cross section perpendicular to the flow area, prototypes have contained a filtrate mesh spacer. By casting the membrane directly onto a porous, oven, incompressible substrate, the filtrate spacer was eliminated so that existing construction would consist of the feed fluid side spacer and the membrane shown in the drawings herein. The membrane envelope is made by gluing the edges of the porous substrate together with a water-resistant adhesive, such as the urethane glue made by the Hexel Corporation. Other achesives use in the module construction include medical grade silicone (e.g.

Dow Corning Corp.) and possibly polymethylemthacralate or other adhesive strategies common in the field. The substrate materials have been Dacron tricot or sailcloth stiffened with a melamine resin, while other materials, such as the DuPont Peemay, have also been used with success. Two types of membranes have been developed for this purpose with, apparently, equivalent results.

The first type is an asymmetric cellulose acetate somewhat similar to the reverse osmosis membranes developed for desalinization. Unlike the reserve osmosis application, it may be necessary to allow free passage of electrolytes while rejecting the heavy fraction. The membrane may be suitably modified either by formulation and annealing conditions or just by the annealing conditions. Exemplary formulations have been the glycerin perchlorate cellulose acetate formulation with altered annealing and the cellulose acetate annealed for short periods of time at less than or equal to 80. centigrade. The exact annealing conditions will change with different cellulose acetate formulations and still produce an acceptable membrane. The second type of membrane that can be used in hemofiltration is a modification of the newer, thin film composite reverse osmosis technology. The thin film composite reverse osmosis membranes are, typically, a backing similar to the one described above (substate), a polysulfone intermediate membrane, and a thin top film (200-500 Angstroms) on the top of the polysulfone. One top film for reverse osmosis has been a polyamid formulation. The modifications of hemofiltration can be either one of two types. The first is to cast a sufficiently thick polysulfone film with pore sizes to yield the rejection characteristics similar to curve A on Figure 4. Note that these rejection characteristics given as curve A on Figure would represent an acceptable transmission of larger molecules for hemofiltration purposes with the intent for artificial kidney purposes to transmit molecules normally present in urine. A concomitant membrane criteria would be insignificant passage of molecules at and above

45,000 molecular weight. This is better understood with references to Figure 7, which shows the spectrum of molecules and formed elements in blood. The second modification of the thin film composite reverse osmosis technology would allow a thinner easting of the polysulfone base with an even thinner to film than is used in reverse osmosis. Again, the criteria is easy passage of electrolytes and end products of metabolism with insignificant passage of the larger plasma proteins. All of the modifications outlined above are easily accomplished by technical personnel well versed in membrane technology.

In order to achieve efficient hemofiltration, the feed fluid side spacer 200 must have certain characteristics. Many thick commercial screens will not work due to their ineffectiveness in promoting removal of rejected material away from the membrane surface. Conversely, extremely thin screens can result in too much pressure drop, which detracts from the transmembrane pressure differential. One spacer that has worked is the Vexar, made by DuPont (polyethylene), with 12 strands to the inch and measuring a total thickness of approximately 25 mils. (0.025 inches). The preferred orientation is to have the mesh lines at

an approximate angle of 450 to the fIow direction as shown in Figuare 6A. A preferred casting material to encase the spiral filter and direct the feed fluid and filtrate streams is polycarbonate or an equivalent biocompatible material. The same material has been used for the filtrate collection tube onto which the rolled spiral assembly is wound. The wound assembly is sufficiently smaller than the inside diameter of the polycarbonate housing, to enable potting of the wound assembly into the polycarbonate shell using silicon adhesive.

Filters in accordance with the foregoing description may still not operate efficiently, i.e. without clogging, unless one or more fluid feedback paths are used. For example, if the fluid is blood, it has been found essential for maintenance of efficiency to r e c i r c u l a t e a large fraction of the blood exiting the filter at port 5 by reintrodueing it at input port 3 at recirculation rate R times the input blood flow rate FF. R must be substantially larger than 2 with a nominal FF of 200 to 250 c/min.; values on the order of 3-8 are required to assure high efficiency with the filter membranes and devices used hitherto and described hereinbefore. While there is at present no comprehensive and exact theoretical basis for the relation of the value of P to the filter parameters and blood composition, most factors are known and at least two factors are believed substantial. First, the use of large amounts of recirculation P enhance the compositional homogeneity of the blood along the length of its flow path through the channel 4. A typical blood input flow rate range is 200-250 cc/minute with a typical filtrate rate of 80 cc/ minute. Without recirculation, then, the plasma portion of the blood would be depleted of approximately half of its water by the time it reached output port 5. For example, if R = 4, then the filter input flow rate is in the range 1000-1250 cc/minute so that withdrawal of 80-100 cc/minute of water results in a much lower percentage change in blood composition down the length L of the filter. Second, the increased rate of flow through the filter with recirculation apparently results in an increased scrubbing action on the filter membrane whereby its clogging proclivity is reduced. As the blood access limited flow rate is increased, the value of R can be decreased and still achieve high efficiencies.

In a similar manner, either or both the intermediate and light fractions may be recirculated through either or both the convective filters 1 and 11 in order to enhance their efficiency.

For fluids other than blood which may not contain the heavy

(cellular) components which scrub the convective filter to reduce the proclivity to clogging by Intermediate components, additive components may be introduced into the feed fluid stream at the input of the convective filter to promote the scrubbing action. Spherical particles will help achieve high efficiency, but may not be as good as non-spherical particles, or particles which are non-uniform in density, or particles which are flexible, such as the cellular components of blood. In addi tion to or in lieu of the specif ic augmentation techniques descr i bed above, other ef f i ciency-promot ing techniques may be employed. Surface perturbations in narrow flow channels can be achieved in several ways. One is to have the membrane exposed to the feed channel containing surface irregularities which may, as an example, be achieved by easting the membrane over an underlying matrix which would promote the formation of the perturbations in the final membrane product.

Another method is to have the membrane supported by an irregular plastic insert with the transmembrane pressure sufficient to deform the membrane over the perturbation which is typically molded into the plastic support. An example of irregular but controlled channel geometries would include tight coiling of the feed, channel, periodic or asymmetric surface waviness parallel to the low, or folding of the flow channel , al l in a manner to induce flow diversion in the direction of flow. For feed fluids containing charged molecules or particles to be rejected, the membrane can be constructed to contain fixed repellant charges.

In addition to the efficiency induction by screens covered in detail, a tubular blood channel can benefit by using a ribbon to produce spiral flow (secondary flows) in addition to axial flow through the tube. Examples of external ly applied forces can include, but are not restricted to, the application of surface charge (in the absence of signi f icant membrane charge) electrically induced by the insertion of electrodes in either the membrane or support structure. In this way, a polarization parallel to the filtrate flow aids in repel l ing the rejected materials away from the membrane surface. Electrodes have been formed by using metallized screens to support the membrane along with a metallized flow channel bounding surface opposite from the surface of the membrane. Another augmentation technique is the use of ultrasound for improving filtration efficiency. In lieu of the metallization, the material must have, as an example, piezoelectric properties. To achieve effective ultrasonic agitation, discrete crystals integral with the filter would be required rather than single continous sound drivers (e.g., reeds or electromagnetically driven diaphragms) which would produce only low frequencies in the feed channel. Ultrasound may be implemented in several ways, including crystals directly exposed to the feed channel. This is the most electrically efficient way of transmitting ultrasound frequency. It is also the least efficient in promoting filtration efficiency while posing the possibility of "heat" damage to the blood. A less electrically efficient way of producing ultrasound is to have the transducer faces placed parallel to the direction of the feed flow, either in or underneath the membrane structure. Although less electrically eff icient, the augmentation of f i ltration by the membrane is most effective wi th thi s orientation. Ultrasound reacts with any and all acoustic interfaces, one such important interface being the membrane/fluid junction. Ultrasound techniques include the use of a single frequency, frequency spectra, and combination of frequencies dependent upon the application. Examples of physical movement include a "washing machine" agitation, continuous rotation wit special rotating seals or connectors, or l inear vibration, all applied to the entire f iltering module. Staging of devices includes the use of more than one device arranged in a parallel and/or sequential manner. This allows direct introduction of cleansed filtrate into the feed flow between each module. This dilutes the feed flow, allowing more efficient fil tration in each module, but normally at the price of increased total surface area (more modules) with concomitant improvement in total clearance or effective filtration. These trade-offs are inherent in the implementat ion of staging and quantative calculations can be made by individuals versed in control l ing f i l tration phenomena. Staging may be of the macrostage variety, in which selected reintroduction of f iltrate can be achieved by design along an otherwise continuous flow channel . Staging is also meant to imply any method intermittently "mixing up" the feed stream to eliminate any component polarization within the feed stream. Another variation of staging also found to be effective is the alternating of active and inative filtering areas. This concept somewhat accomplishes the sequential mixing alluded to above. Without any other augmenting method, the remixing would be by diffusional processes in the case of rejected molecules. With the simultaneous use of other augmenting methods, convective modes of transport could assist the diffusion. Independent manipulation of flow rates in the device include, generally, any additional pumping or flow action in addition to the s impl throughput required to achieve practical filtration. Details have been given on the use of recirculation in one of the preferred hemofiltration designs, but the invention would also include mechanical oscillatory motions to cause vortex shedding and/or fluid replenishment from grooves perpendicular to the mainstream feed flow, as an example of preferred geometries in combination with other augmentation methods. The more direct example herein is the use of spiral hemofilter modules with screens capable of inducing high efficiency in combination with recirculation of the exiting fluid back to the inlet. Since the hematoerit affects the production of optimum eff iciency, variation of the reintroduction of filtrate between the module Inlet and exit ; also a method of improving the filtering efficiency, considered to be one of the biophysical condition embodiments. In addition to the methods already covered, independent control of biochemical and biophysical conditions includes the pH in the feed channel (more importantly at the membrane surface), control over the charge at the membrane surface, and the fractional filtrate to feed fluid return ratio.

In general if the complex feed fluid itself has particle or cell sizes in excess of 0.1 micron (and preferably in excess of 1 micron), auxiliary particles are unnecessary. That is, there is a category of complex fluids containing natural particles, huge macromolecules, or active or inactive cellular material with which efficient filtration may be achieved by the apparatus and methods described with respect to the blood example hereinbefore without the addition of auxiliary particles. Complex fluids in this category comprise human and animal blood or lymph fluids, microbial or cellular suspensions (e.g. bacterial, plant cells, animal blood or lymph fluids, microbial or cellular suspensions

(e.g. bacterial, plant cells, animal cells, etc.) meat products and by-products, plant extracts, suspensions of algae or fungi, vegetable food and beverages containing particles such as p ulp

(e.g. orange juice), pulp products, activated charcof 1 suspensions, paints, latex suspensions, stanch, photographic emulsions, printer's ink, waste streams such as machine oil, automotive oil and other oily suspensions or emulsions which it is desirable to process or reprocess (e.g. reclaim).

Ideal candidates for the apparatus and techniques described in this application include bacterial or other culture media nothe harvesting of the grown components, including proteins) and the removal of antigens, viruses or bacteria from fluid streams. By way of more specific example, anti-cancersubstances (e.g. interferon) may be produced by stimulating cellular activity. The interferon molecule is sufficiently large to be amenable to isolation by the schematic process shown in Figure 2, where the feed fluid may comprise cell cultures in a broth medium. Interferon is a large, fragi le mol ecule sufficiently bigger in size than the normal nutrients and much smaller than the cells that produce it that it may fit into the strategy of Figure 2 and be removed as an intermediate fraction. The feed fluid in Figure 2 would comprise the live cells or dead cell fragments, the interferon and the nutrient of broth materials. The first filter module in Figure 2 (containing the Type II membrane) would contain, for example, a Nuclepore membrane with a pore size preferably between 0.1 and 0.9 microns. Using recirculation in the spiral wound module cell , as shown in Figure 5, debris accumulation on or near the membrane surface could be reduced and high efficiency filtration conditions established. For this application pulsating flow or pressures may be sl ightly better than constant f low recirculation. Pores of approximately 0.4 microns and above provide for easy transmission of the interferon and nutrient fluid (broth) while the cellular material is rejected. An advantage to this process is the gentle treatment of the interferon molecule. This is a distinct advantage over present methods of concentrating this valuable substance. The filtrate stream from the type II membrane would then be processed by a second filtrate module where the type I membrane could be exceedingly similar to that shown as curve A in figure 4. In this case, the filtrate stream through the type I membrane would contain only the nutrient materials (the broth in a water solution) and could be returned to the cell growth or interferon producing process as part of the overal l process (not shown). The intermediate fraction removal stream from the module with the type I membrane would contain interferon in a more concentrated state than in any other portion of the process shown. This stream might be subject to multi-stage processing with type I modules to gently accompl ish furthe r concentrating of the interferon. A unique advantage of this schema is that the filtrate process is gentle, an absolute requirement in interferon production. Interferon production is presently done in batch systems, whereas the filtration, separation, and concentrating of interferon could be made continuous. This is another distinct advantage. The type I f i lter membrane modul e for thi s application would not be particularly susceptible to the efficiency augmentation methods since neither the interferon nor the nutrient broth would contain macromolecules or particles of sufficient size and concentration to engender the high efficiency potential of the augmentation methods.

In a similar fashion, animals may be harvested for antibody production. An antigen would be introduced into the animal (e.g., cow) and the immune system would produce antibodies to the antigen. Antibodies would be either cel l mediated or in the globulin portion of the plasma protein spectra. Following the immunoglobulin example, Figure 2 is applicable when the cow is producing substantial antibodies. The type II membrane would again be a microporous type for plasmapheres is purposes. There are commercial cellulosic and polysulfone membranes (made by

Millipore, Enka, Gelman Corporations and others) with tortuous path structure that can be used in Place of the Nuclepore membrane given as an example in the interferon process, Plasma as defined in Figures 7 would pass through the cores while al blood cells would be rejected. The type I membrane is a specially designed ultrafiltration type with transmission of molecules up to approximately 100,000 molecular weight. A charged membrane would augment the separation or the albumin molecules from the globulin molecules while a pH change could also shift the isoelectric points of the proteins to enhance selective separation. Another possible co-process would be gelation of the proteins or cryoprecipitation. A perfect membrane or one employing either an electric field or cross-flow fraetionation could accomplish the selective separation of the immunglobulins from the remaining plasma proteins. The unwanted plasma proteins and all lower molecular weight material would be contained in the filtrate stream from the type I membrane and could be returned to the animal. The removal stream would contain concentrated immunoglobulins and could be subject to further processing or concentrating.

Another important application would the harvesting of protein from cellular processes. Again, the vital step is the separation of a large molecular weight protein from huge cells which must be kept in a controlled environment. The inherent advantage is that the intermediate fraction removal stream is already cncentrated in the protein (refer again to Figure 2) and the culture media and cells can be remixed by combining the heavy and light fraction removal streams. The first separation step in the type II module would be subjected to the high efficiency techniques describe hereinbefore and hereinafter, while the second filtrate module containing the type I membrane would not inherently be subject to efficiency augmentation, but could have particles added to it

The same cell separation requirement is inherent in the overall process of using solar energy with photosynthesis to produce high energy molecules which are then useful for food purposes or synfuel (biomass) energy production purposes. Othe r medical applications include the separation of cerebrospinal fluid and the on-line production of cardioplegia solution.

The foregoing material has described complex fluids having sufficiently large particles or cells so that all the augmentation methods are potentially applicable. Where these large particles or cells are not present, certain of the efficiency inducing techniques relating to the dynamics of the feed fluid channel flow are not very helpful. For such complex fluids, other of the augmentation techniques described are still, viable (e.g. ultrasound, charged membranes, etc.). Some complex fluids would be amenable to added particles or cells in order to enable the full range of efficiency augmentation techniques described hereinbefore. After the f i ltration, the auxiliary particles or cells could be removed if desired by additional filtration, centrifugation or other appropriate conventional techniques. Another class of complex feed fluids may have molecules suf f iciently smal l or of such a geometrical configuration that even particle addition could be efficiency inefficacious. Even for this latter class, the double convective filter method and apparatus (possibly in combination with feedback or recirculation techniques, charged membranes, etc.) will provide results superior to existing membrane separation technology. The following is a list of applications or complex fluids containing insufficient cells or particles for the self-induction of all augmentation methods: recombinarit genetic engineering products, part of chemical or clinical analysis systems, waste water treatment, cleaning incoming water streams paints, RNA/DNA processing, enzyme engineering, dairy product processing including whey, food processing such as tomato juice, laundry waste or incoming fluid processing, part of the recycling of dry cleaning fluids, gelatin production, wax and wax product processing, production of liquid foods, reprocessing of steel pickle liquor, polymer processing, de-watering of metallic colloids, processing of oily emulsions, and production or procesing involving venoms or other poisons and toxins. Where particle addition engenders enhanced efficiency, certain parameters are preferred. While homogeneous hard spherical particles will help achieve high efficiency, flexible spherical, particles or flexible particles of non-uniform density or shape are more efficacious.

While the invention has been particularly described and shown in reference to the preferred embodiments thereof, it will be understood by those skilled in the art that various changes in form and detail and omissions may be made therein without departing from the spirit and scope of the invention.

Claims

We Claim:

1. An apparatus for the filtration of predetermined molecular weight components from a complex fluid, comprising convective filtration means for separation of said fluid into a heavyfraction having high molecular weight components and a complex f i ltrate fraction having lower molecular weight components than said heavy fraction, said filtration means including at least one augmentation means for maintaining efficiency during said filtration.

2. An apparatus for the filtration of predetermined molecular weight components from a complex fluid, comprising convective filtration means for separation of said fluid into a first fraction having high molecular weight components and a complex filtrate fraction having lower molecular weight components than said first fraction, said means comprising spiral geometry augmentation means for maintaining efficiency during the period of said filtration.

3. The apparatus of Claim 1 or Claim 2, wherein saidfiltration means includes at least two augmentation means for maintaining the efficiency of said filtration during the period thereof .

4. The apparatus of any of Claims 1-3, further comprising means for recirculating a portion of said heavy fraction through said convective f i l tration means to improve f i l tration efficiency.

5. The apparatus of any of Claims 1-3, further comprising means for recirculating a portion of said filtrate fraction through said convective filtration means to improve filtration efficiency.

6. The apparatus of any of Claims 1-3 , w herein said convective filtration means comprises charged membrane means for repelling selected constituents in said blood.

7. The apparatus of any of Claims 1-3, further comprising first recirculation means for recirculating a portion of said heavy fraction through said convective fi ltration means and second recirculation means for recirculating a portion of said fi ltrate fraction through said convective f i ltration means to improve filtration efficiency.

8. An apparatus for the fil tration of predetermined molecular weight components from a complex fluid comprising, in combination:

first convective filtration means for separation of said fluid into a heavy fraction having high molecular weight components and a filtrate fraction having lower molecular weight components than said heavy fraction; and second convective filtration means for reception of said filtrate fraction and separation of said filtrate fraction into an intermediate fraction having intermediate molecular weight components and a light fraction having lower molecular weight components than said intermediate fraction.

9. The apparatus of Claim. 8, where at least one of said first and second convective filtration means has spiral geometry augmentation means for maintaining efficiency during the period of filtration.

10. The apparatus for Claim 8 or 9, where said first convective filtration means comprises charged membrane means for repelling selected constituents in said blood.

11. The apparatus of any of Claims 8-10, further including recirculation means for recirculating a portion of said heavy fraction through said first filtration means for enhancing efficiency during said separation.

12. The apparatus of any of Claims 8-10, further comprising first recirculation means for recirculation means for recirculating a portion of said heavy fraction through said convective filter means, and second recirculation means for recirculating a portion of said filtrate fraction through said convective filtration means, both of said recirculated portions improving filtration efficiency.

13. The apparatus of any of Claims 8-12, further including means for recombining said heavy and said light fractions.

14. A method for separating predetermined molecular weight components from a complex fluid, comprising the step of filtering said fluid through a convenctive filter for separating said fluid into a heavy fraction having high molecular weight components and a complex filtrate fraction having lower molecular weight component s than said heavy fraction, said filter having at least one augmentation means for maintaining the efficiency of said filtering during the period thereof.

15. A method for separating perdetermined molecular weight components from a complex fluid, comprising the step of filtering said fluid through spiral convective filter means for separating said fluid into a heavy fraction having high molecular Weight components and a complex filtration fraction having lower molecular weight components than said heavy fraction.

16. The method of any of Claims 17-20, further including recirculating a portion of said heavy fraction through said filtration means to improve filtration efficiency.

17. The method of any of Claims 14-16, further including recirculating a portion of said filtrate fraction through said filtration means to improve filtration efficiency.

18. The method of any of Claims 14-17, said method further comprising the step of exposing said complex fluid in said convective filter means to charged membrane means for repelling selected constituents in said blood.

19. A method for separating predetermined molecular weight components from a complex fluid comprising the steps of: passing said fluid through first convective filtration means for separating said fluid into a heavy fraction having high molecular weight components and a filtrate fraction having lower molecular weight components than said heavy fraction; passing at least a portion of said f iltrate fraction through second convective filtration means for separating said f iltrate portion into a n intermediate fraction having intermediate molecular weight components and a light fraction having lower molecular weight components than said intermediate fraetion.

20. The method of Claim 19, wherein said fluid is passed through first convective filtration means having a spiral geometry.

21. Thenethod of Claim 19 or Claim 20, further including step of recirculating a portion of said heavyfraction through said first convective filtration means to enhance the separation effieiency.

22. The method of any of Claims 19-21 , further including the step of recirculating a portion of said filtrate fraction through said first convective filtration means to enhance the separation efficiency.

23. The method of any of Claims 19-22, further including the step of recombining said heavy fraction and said light fraction.

24. The method of any of Claims 19-23, further including the step of removing at least a portion of said intermediate fraction in a continuous process wherein at least portions of said heavy and said light fractions are returned to a generation situs for said complex fluid.

25. An apparatus for the filtration of a complex fluid comprising cellular material, comprising convection filtration means for separation of said fluid into a heavy fraction including said cellular material and a complex filtrate fraction, said filtration means including at least one augmentation means for maintaining efficiency during said filtration.

26. An apparatus for the filtration of a complex fluid comprising cellular material, comprising convection filtration means for separation of said fluid into a heavy fraction including said cellular material and a complex filtrate fraction, said means comprising spiral geometry augmentation means for maintaining efficiency during the period of said fi1lration.

27. A method for filtering cellular components from a complex fluid, comprising the step of filtering said fluid through a convective filter for separating said fluid into a heavy fraction comprising said cellular components and a complex filtrate fraction, said filter having at least one augmentation means for maintaining efficiency of said filtering during the period therefor.