US20090277824A1 - Method for reducing performance variability of multi-layer filters - Google Patents

Method for reducing performance variability of multi-layer filters Download PDF

Info

Publication number
US20090277824A1
US20090277824A1 US12387636 US38763609A US2009277824A1 US 20090277824 A1 US20090277824 A1 US 20090277824A1 US 12387636 US12387636 US 12387636 US 38763609 A US38763609 A US 38763609A US 2009277824 A1 US2009277824 A1 US 2009277824A1
Authority
US
Grant status
Application
Patent type
Prior art keywords
capacity
membrane
distribution
membranes
device
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Abandoned
Application number
US12387636
Inventor
Sal Giglia
Mani Krishnan
Nitin Satav
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
EMD Millipore Corp
Original Assignee
EMD Millipore Corp
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D69/00Semi-permeable membranes for separation processes or apparatus characterised by their form, structure or properties; Manufacturing processes specially adapted therefor
    • B01D69/02Semi-permeable membranes for separation processes or apparatus characterised by their form, structure or properties; Manufacturing processes specially adapted therefor characterised by their properties
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61LMETHODS OR APPARATUS FOR STERILISING MATERIALS OR OBJECTS IN GENERAL; DISINFECTION, STERILISATION, OR DEODORISATION OF AIR; CHEMICAL ASPECTS OF BANDAGES, DRESSINGS, ABSORBENT PADS, OR SURGICAL ARTICLES; MATERIALS FOR BANDAGES, DRESSINGS, ABSORBENT PADS, OR SURGICAL ARTICLES
    • A61L2/00Methods or apparatus for disinfecting or sterilising materials or objects other than foodstuffs or contact lenses; Accessories therefor
    • A61L2/0005Methods or apparatus for disinfecting or sterilising materials or objects other than foodstuffs or contact lenses; Accessories therefor for pharmaceuticals, biologicals or living parts
    • A61L2/0011Methods or apparatus for disinfecting or sterilising materials or objects other than foodstuffs or contact lenses; Accessories therefor for pharmaceuticals, biologicals or living parts using physical methods
    • A61L2/0017Filtration
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D29/00Other filters with filtering elements stationary during filtration, e.g. pressure or suction filters, or filtering elements therefor
    • B01D29/50Other filters with filtering elements stationary during filtration, e.g. pressure or suction filters, or filtering elements therefor with multiple filtering elements, characterised by their mutual disposition
    • B01D29/56Other filters with filtering elements stationary during filtration, e.g. pressure or suction filters, or filtering elements therefor with multiple filtering elements, characterised by their mutual disposition in series connection
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D61/00Processes of separation using semi-permeable membranes, e.g. dialysis, osmosis, ultrafiltration; Apparatus, accessories or auxiliary operations specially adapted therefor
    • B01D61/14Ultrafiltration; Microfiltration
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D65/00Accessories or auxiliary operations, in general, for separation processes or apparatus using semi-permeable membranes
    • B01D65/10Testing of membranes or membrane apparatus; Detecting or repairing leaks
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D69/00Semi-permeable membranes for separation processes or apparatus characterised by their form, structure or properties; Manufacturing processes specially adapted therefor
    • B01D69/12Composite membranes; Ultra-thin membranes
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D71/00Semi-permeable membranes for separation processes or apparatus characterised by the material; Manufacturing processes specially adapted therefor
    • B01D71/06Organic material
    • B01D71/66Polymers having sulfur in the main chain, with or without nitrogen, oxygen or carbon only
    • B01D71/68Polysulfones; Polyethersulfones
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D2239/00Aspects relating to filtering material for liquid or gaseous fluids
    • B01D2239/06Filter cloth, e.g. knitted, woven non-woven; self-supported material
    • B01D2239/069Special geometry of layers
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D2319/00Membrane assemblies within one housing
    • B01D2319/02Elements in series
    • B01D2319/022Reject series
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D2325/00Details relating to properties of membranes
    • B01D2325/02Details relating to pores or pososity of the membranes
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D2325/00Details relating to properties of membranes
    • B01D2325/20Specific permeability or cut-off range
    • BPERFORMING OPERATIONS; TRANSPORTING
    • B01PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01DSEPARATION
    • B01D2325/00Details relating to properties of membranes
    • B01D2325/28Degradation or stability over time
    • GPHYSICS
    • G01MEASURING; TESTING
    • G01NINVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N15/00Investigating characteristics of particles; Investigating permeability, pore-volume, or surface-area of porous materials
    • G01N15/08Investigating permeability, pore-volume, or surface area of porous materials
    • G01N15/082Investigating permeability by forcing a fluid through a sample

Abstract

Multi-layer membrane devices with reduced performance variability, and method of reducing performance variability of such devices. The variability is reduced by combining two or more membranes with similar pore sizes, and carefully selecting the upstream membrane based upon its performance rating, in order to control the performance of the overall device. Selective layering reduces the capacity range of the device when compared to random layering, with the mean capacity greater than overall population mean. The flux range also can be reduced, with the mean near the overall population mean. The LRV range also can be reduced, with the mean LRV near or higher than the overall population mean.

Description

  • This application claims priority of U.S. Provisional application Ser. No. 61/127,156 filed May 9, 2008, the disclosure of which is incorporated herein by reference.
  • BACKGROUND OF THE INVENTION
  • Virus contamination poses a threat to the safety of biopharmaceuticals derived from recombinant or human plasma sources. Manufacturing processes must provide clearance of endogenous and adventitious viruses to ensure product safety. To that end, filters have been developed that provide virus removal. To be commercially effective and cost-efficient, such filters must provide effective virus removal while maximizing product recovery, and must be reliable, consistent and capable of validation.
  • Virus removal from liquid streams, particularly process streams in the biotech and pharmaceutical industry, has been practiced for some time. High viral clearance, high product flux, very high protein passage and simplicity of operation are the goals of the end-user, yet the prior art does not provide a solution that satisfies all of these goals. Since high viral clearance is always needed, it is the other process goals that have suffered. Meeting these other goals would substantially increase production efficiency and thus lower processing cost.
  • The prior art provides several membrane types and filtration modes for viral clearance. One such product is the Viresolve NFR filters with Retropore® membrane that is used to remove retroviruses from recombinant protein solutions or human plasma sources. U.S. Pat. No. 7,108,791 B2 discloses a virus removal methodology using Viresolve NFR filters, suitable for conducting a high-flux fluid separation of virus from a protein in the course of biopharmaceutical manufacture, the methodology comprising the steps of: (a) providing a filtration device comprising a housing having a fluid inlet and a filtrate outlet, and containing at least two interfacially contiguous asymmetric membranes, wherein: (i) the asymmetric membranes are each substantially hydrophilic, (ii) at least two of the hydrophilic membranes are each capable of substantially selectively preventing the passage therethrough of said virus and substantially permitting the passage therethrough of said protein, (iii) at least two of the asymmetric membranes have each a tight-side and open-side, (iv) the foremost asymmetric membrane is oriented such that fluid introduced in said housing through the fluid inlet commences passage through said foremost asymmetric membrane through its open side. It also claims that each of said asymmetric membranes are substantially identical in their composition and porosity, and wherein the porosity of each said asymmetric membranes is defined to enable performance of the virus removal methodology, yielding a log reduction value (LRV) greater than 6 and a protein passage greater than 98%.
  • Consistency in device-to-device performance is critically important for users of membrane filtration devices in order to predict filtration performance from run to run, and to scale-up or otherwise engineer and design processes. Users often cite performance consistency as one of the most important factors in filter selection. In the case of bacterial or virus filtration, the performance criteria important to users include throughput capacity, flux (or permeability), and retention of bacteria or viruses.
  • Capacity is important in high value filtration processes such as virus removal in biopharmaceutical manufacturing. Capacity relates to the length of time, or volume of fluid that can be filtered before the filtration rate is reduced through plugging by retained species or fouling, etc., to an uneconomic level. High capacity filtration improves process economics by reducing processing time and the amount of filter area required. High flux becomes critical in medium- and large-scale manufacturing operations where process equipment is required to be turned around rapidly to process the next batch of product. In all these operations, repeatability of filter performance from batch to batch is very important.
  • Performance variability may depend on a number of factors such as membrane variability, lot-to-lot protein variability, variability in operating parameters, etc. Manufacturers of the therapeutic proteins take measures to minimize lot-to-lot protein variability. Membrane variability can be defined by the difference of properties from manufacturing lot to lot. A manufacturing lot, or batch, the terms are interchangeable, is defined by the manufacturer. A lot can be the production output from a single polymer solution volume in immersion casting, or the output from an operating shift. It is common in flat sheet membrane to label each manufactured roll as a lot or sub-lot. A single roll or lot may also be subdivided based on the variability within that roll. There are many factors that influence membrane performance, including the pore size distribution, the membrane chemistry, membrane thickness, membrane porosity, and others. While membrane manufacturing processes are designed to control all of these factors to maximize uniformity and consistency, there will inevitably be some distribution within normal manufacturing conditions for all of these variables. This membrane variability contributes to device-to-device performance consistency. Methods used to reduce variability in a device through selective layering is described herein.
  • In addition to reducing device variability, membrane device manufacturers desire to optimize and/or maximize device properties of a population of devices. As will be described, embodiments herein provide for methods to increase the average capacity (defined below) of a population of manufactured devices through selective layering as compared to a population manufactured by standard non-selective layering.
  • The ability to control device consistency is particularly important for multilayered filtration devices. For these devices, not only is the variability of the membrane important, but membrane variability can affect the interaction between the layers, as will be shown below.
  • It therefore would be desirable to provide a multi-layer membrane device with reduced performance variability, as well as a method of reducing the performance variability of such devices, notwithstanding the inherent variability of the membrane manufacturing process.
  • SUMMARY OF THE INVENTION
  • Aspects of the discoveries of the inventors provide multi-layer membrane filtration devices with reduced performance variability. The variability is reduced by combining two or more membranes with similar pore sizes, and carefully selecting the upstream membrane based upon its performance rating, in order to control the performance of the overall device.
  • The upstream or top membrane layer is the membrane layer that the feed stream fluid initially contacts and passes through. Conversely, the bottom layer is the last layer the fluid contacts and passes through.
  • The membranes do not necessarily need to be stacked. They can be separated within the device, or even contained within separate devices. The membrane/device format is not limited to flat sheets, but can be in any format including pleated, spiral, or hollow fiber and can be applied to devices that are operated either in normal (or dead-ended) flow or tangential flow methods.
  • In an embodiment, selective layering in accordance with the invention reduces the capacity range of the devices when compared to random or non-selective layering. Range is defined as the difference between the high and low value of a property of a population of devices.
  • In an embodiment, the mean capacity of a population of devices is greater than overall population mean of the individual membrane lots used to manufacture the population.
  • In an embodiment, the capacity of a multilayered device produced from layers of different capacity is higher that the arithmetic average of the individual layers.
  • In an embodiment, the flux (or permeability) range is reduced, with the mean near the overall population mean.
  • In an embodiment, the LRV range is reduced, with the mean LRV near or higher than the overall population mean.
  • In an embodiment, the filtration capacity of a population of multilayered filter devices manufactured from membrane lots or a single lot having capacity property variability is maximized.
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • FIG. 1 is a bar graph showing throughput of 2-layer membrane devices containing high and low capacity membranes combined in different positions with respect to top (upstream) and bottom (downstream) position.
  • FIG. 1 a is a graph showing throughput of 3-layer devices containing high, medium, and low capacity membranes combined in different positions with respect to top (upstream), middle and bottom (downstream) position.
  • FIG. 2 is a graph of a hypothetical distribution of membrane capacities;
  • FIG. 3 is a graph of membrane capacity distribution showing a selection of top layer membrane and bottom layer membrane in accordance with certain embodiments;
  • FIG. 4 is a graph of membrane capacity distribution showing another selection of top layer membrane and bottom layer membrane in accordance with certain embodiments;
  • FIG. 5 is a graph showing the effect of layering on buffer permeability;
  • FIG. 6 is a graph showing the effect of layering on LRV;
  • FIG. 7 is a graph of membrane capacity distribution showing a selection of top layer membrane and bottom layer membrane in accordance with certain embodiments;
  • FIG. 8 is a graph of hypothetical capacity,. permeability and LRV distributions and correlations;
  • DETAILED DESCRIPTION OF THE INVENTION
  • Membrane manufacturing processes inherently result in some variability in membrane properties even though materials and process conditions are kept as constant as possible. Manufacturers typically carry out a battery of tests on membrane lots to ensure that the manufacturing process has produced membrane within established limits. Additionally, procedures can also be instituted to classify or “rate” each batch or roll of membranes after manufacture, based on performance. Membrane manufacturers use a variety of methods to rate membranes. For example, for microporous membranes, versions of the well known bubble point test may be used to give a pore size rating. ASTM F-838-05 describes a method of LRV rating using b. diminuta. Ultrafiltration membranes are rated by testing with water soluble polysaccharides as described, for example, in G. Tkacik and S. Michaels, “A rejection profile test for ultrafiltration membranes and devices”, Bio/Technology 9, 941 (1991).
  • Membrane lots to be incorporated into multilayered devices for virus removal in biotherapeutics manufacturing may be tested by normal flow log reduction value (LRV)/capacity tests which challenge the membrane with the bacteriophage, such as φx-174, in a protein solution, such as bovine serum albumin (BSA), or human IgG. Virus retention (LRV) and capacity (volume filtered) at a predetermined end-point, such as 70-75% flow reduction, are measured, and relative performance values are obtained. Typically, water or another appropriate fluid, is used to measure permeability of the membrane.
  • The membranes from the given batch are then performance rated based upon the results obtained. All the membranes are qualified under a single standard of permeability, retention and capacity. That is, all qualified membrane is considered as a single product meeting all specifications with regard to form, function and performance (including pore size rating). As with any product, there are tolerances in the specifications. Regardless of the tightness of the tolerances, there will inevitably be some finite performance variability within the specification tolerance range.
  • It is known that for stacked filters, the resistance to flow is additive, so that the flux of the stack will be approximately the average of the filters used to make the stack. Also, retention, as measured by LRV, is approximately additive, so that the LRV of a stack will be approximately the sum of the LRV's of the individual layers. However, the present inventors have surprisingly found that for multi-layer devices containing similarly rated membranes in each of the layers, the contribution of each layer to performance of the device does not necessarily follow such a predictable response, particularly where performance is defined in terms of capacity.
  • The present inventors have found that by providing multi-layer devices containing similarly rated membranes in each of the layers, each layer within the device contributes to performance. However, the contribution of each layer to performance of the device generally is not equal, particularly where performance is defined in terms of capacity. As used herein, the term “capacity” is the volume of fluid that can be processed by the membrane before reaching a practical endpoint. In the case of constant pressure filtration, for example, that endpoint is defined as the condition at which the flux has decayed due to membrane fouling to a predetermined minimum.
  • The present inventors have found that in a multi-layer device containing similarly rated membranes, the upstream layer controls throughput capacity. Moreover, the inventors were surprised to discover that when selective layering was employed, the average capacity of a device, or of a population of devices, was higher that the average of the individual layers. Accordingly, once the capacity of each membrane of the lots to be used to make a population of devices is known, the upstream membrane of a multi-layer membrane device can be selected according to the methods herein to obtain the highest capacity for a single device, or the highest average for the population manufactured. As well, the range of capacity values for the device population will be smaller than that of the population of membranes from whence the devices were made. Additional performance criteria, such as retention and/or permeability, also can be used to measure performance, and to base membrane selection on.
  • Added benefits are that the while the resultant flux of the selectively layered device is approximately the average of the layers, the range of values is smaller that that of the membranes used. The resulting LRV of selectively layered devices is similar to the average of the membranes used, with a smaller range.
  • By way of example, the effect of layer ordering on capacity can be demonstrated on normal flow virus filtration membranes, although the present invention is not so limited. FIG. 1 shows that a double layer membrane capacity (defined as 75% flow decay from the initial non-fouled value) of an asymmetric polyethersulfone membrane commercially available from Millipore Corporation under the name VIRESOLVE PRO™, is controlled primarily by the capacity of the upstream layer membrane, regardless of the capacity of the bottom layer membrane. Thus, for the membrane roll designated as “low” capacity, two layers of this membrane had a throughput of about 500 L/m2. For the membrane roll designated as “high” capacity, two layers of this membrane had a throughput of about 900 L/m2. Throughput and capacity are equivalent terms. When “high” membranes were positioned upstream over “low” membranes that were placed in the downstream position, the capacity of these devices were nearly equivalent to that of devices that contained two layers of “high” membrane. A similar result is apparent for membranes where the “low” capacity membranes were positioned upstream over “high” capacity membranes that were placed in the downstream position. The capacity of the multi-layer device was not negatively impacted by the relatively low throughput of the bottom layer membrane. The “low/high” combination shows that despite the relatively high capacity of the downstream layer, again the overall capacity of the multi-layer device is a function primarily of the upstream layer. Accordingly, if two or more membranes of known differing capacities are available for assembly into a multi-layer membrane device, the membrane available with the highest capacity can be selected as the upstream layer. To explain the selective layering approach, the capacity properties of membrane lots available to be made into devices is approximately formulated into a distribution as illustrated in FIG. 2. A portion of the membrane capacity distribution is then selected as top layer membrane in order to reduce the range of device capacities compared to layering the membranes randomly, and to maximize capacity and/or minimize device-to-device capacity variability.
  • This phenomenon has also been demonstrated on three-layer Optiscale-25 devices with Viresolve NFP membrane (Millipore Corporation, Billerica, Mass.). Viresolve NFP (Normal Flow Parvovirus) membranes clear parvovirus from recombinant or human plasma sources. U.S. Pat. No. 5,017,292 discloses technology used to produce the Viresolve membrane. It provides a composite PVDF membrane comprising a porous membrane substrate and a tight side (the surface having smaller diameter pores) having ultrafiltration separation properties.
  • To demonstrate the effect on capacity, NFP membrane rolls with a range of average capacities were selected and labeled as high (H), mid (M) and low (L) capacity rolls.
  • Devices were built with various combinations, for example: HHH, HLL, MHH, MLL, LHH, LLL, etc. The first, second and third letter indicate upstream, middle and downstream layers, respectively, as shown below
  • Figure US20090277824A1-20091112-C00001
  • These Optiscales were tested for capacity using a BSA stream spiked with φx-174. FIG. 1A shows relative capacity values for different layering combinations.
  • These results showed that when “Mid” capacity membrane (MMM) was used. as a top layer over “High” capacity (HHH) or “Low” capacity (LLL) membranes, then devices with layered combination (MHH or MLL) showed capacity close to Mid capacity membrane. In other words, variability in relative capacity was reduced as a result of selectively placing “mid” capacity membrane as top layer. In this particular example, the range in relative capacity was reduced from about 1.0 to about 0.2.
  • Accordingly, if two or more membranes of known differing capacities are available for assembly into a multi-layer membrane device, the membrane with the capacity closest to the target capacity of the overall device can be selected as the upstream layer. Using this approach, a portion of the membrane capacity distribution can be selected as top layer membrane, thereby reducing the range of device capacities compared to layering the membranes randomly.
  • In view of the relative influence of the upstream membrane layer, various layering strategies are possible. Selected embodiments of the invention are described based on the FIGS. 2 and 8. FIG. 2 is a Gaussian distribution of capacity values calculated to represent membranes similar to those shown in FIGS. 5 and 6. FIG. 8 represents correlated data of permeability to capacity and LRV to capacity. To make the simulations more realistic, the points of FIG. 8 were generated from an actual linear correlation with a set degree of randomness to give the scatter plot shown. These points were then used to calculate the effects of various layering strategies.
  • To illustrate these strategies, a two-layer membrane device will be exemplified, although it should be understood that similar results can be obtained with devices having three or more layers. In this embodiment, a middle portion (e.g., ±10-20% of the midpoint of the distribution) of the capacity distribution is designated for the top or upstream membrane layer, with the remaining portion of the distribution to be used as the bottom layer membrane. This is illustrated graphically in FIG. 3. By utilizing a top layer membrane from the middle of the capacity distribution curve, and a bottom layer membrane from the left or right side of the capacity distribution curve, the range of device capacities will be reduced compared to random layering, with no effect on mean capacity (Table 1).
  • FIG. 4 illustrates another embodiment of a multi-layered membrane device. In the embodiment of FIG. 4, the top layer membrane is randomly selected from the top half of the capacity distribution curve, and the bottom layer membrane is randomly selected from the bottom half of the capacity distribution curve. In addition to decreasing the capacity range by half as in the embodiment of FIG. 3, the embodiment of FIG. 4 also increases the mean capacity as compared to random layering. Because this embodiment combines the upper portion capacity distribution (and therefore the upper portion permeability distribution and the lower portion retention distribution), with lower portion capacity distribution (and therefore lower portion permeability distribution and upper portion retention distributions, respectively), and because higher capacity membranes tend to also have higher permeability and lower virus retention (but all within the product specifications) variability in permeability and retention is reduced compared to random layering (Table 1).
  • Those skilled in the art will appreciate that to effectively employ the embodiments of FIGS. 3 and 4, the capacity distribution must be well defined and predictable. If the actual distribution does not match the distribution assumed to establish the top and bottom layer specifications, then there will be unequal numbers of top and bottom layer membranes. The top layer capacity range could be expanded to give a safety factor in allowable top layer membrane, but this will diminish the variability reduction advantage of these strategies.
  • Variability in retention and permeability are also important performance factors to filter users. While capacity is controlled primarily by the top membrane layer, both layers in two-layer devices contribute to flux and retention. The resistivity (inverse of permeability) of each layer is additive.
  • The device retention is often expressed by the log reduction value (LRV):
  • LRV = - log 10 [ C p C f ]
  • where C is the concentration of the target species to be retained by the membrane. LRV is generally additive for the two layers. As shown in FIGS. 5 and 6, the order of the membrane layers used in FIG. 1 does not appreciably affect two-layer permeability or the two-layer LRV.
  • In a preferred another embodiment, retention and permeability variability can be minimized along with capacity variability. To that end, if the capacity distribution of FIG. 2 is assumed along with the permeability and retention correlations to capacity as shown in FIG. 8 (typically, membrane capacity correlates positively with permeability and negatively with retention), the membrane layering of FIG. 7 can be employed. Thus, the top and bottom layers are paired sequentially from opposite ends of the capacity distribution curve to minimize capacity variability, permeability variability and retention variability. Specifically, the membrane with a capacity of 750 l/m2 (“Top Layer 1”) is combined with the membrane with a corresponding low capacity of 450 l/m2 (“Bot. Layer 1”) to achieve a capacity of 750 l/m2, etc.
  • The performance variability resulting from the various embodiments was calculated for a two-layer device and compared in Table 1. It can be seen that the embodiment of FIG. 7 allows for the lowest standard deviation in permeability and retention among all of the embodiments, with only slightly higher standard deviation in capacity compared to the embodiment of FIG. 3 or random layering. The FIG. 7 embodiment also allows for higher mean capacity than the FIG. 3 embodiment. Note that the FIG. 3 embodiment can result in slightly increased permeability and retention variability compared to random layering. This occurs because the membranes with the extremes of permeability and retention are never combined with each other. The FIG. 3 embodiment does have the advantage of the lowest capacity standard deviation.
  • TABLE 1
    Capacity LRV Perme-
    Layering Capacity Std. LRV Std. ability Permeability
    Strategy Mean Dev. Mean Dev. Mean Std. Dev.
    Random 600 70 7 0.85 20 1.4
    FIG. 3 600 29 7 0.86 20 1.5
    FIG. 4 656 41 7 0.65 20 1.0
    FIG. 7 656 41 7 0.40 20 0.8
  • In addition to the above advantages, the embodiment of FIG. 7 does not require a priori knowledge of the performance distributions as do the embodiments of FIG. 3 and 4. Membranes can simply be paired according to this ranking system as they are available.
  • This method can be applied in various ways, depending on the desired output. For example, if only membranes at the very high end of capacity are available, the lower capacity membrane of these can be designated for the top layer, to give a device capacity close to the target mean, i.e., smaller range of values. If on the other hand, the maximum capacity is desired, the highest capacity membrane would be used for the top layer.
  • In certain preferred embodiments of carrying out the present invention, each batch of membranes produced is characterized by performance, such as by capacity. The batches are the sorted accordingly, such as in descending order of capacity value. The batches are then matched based upon capacity, such as by pairing batches by first matching those with the highest capacity to those with the lowest capacity, then by matching those with the highest capacity of the available membranes remaining with those with the lowest capacity of the available membranes remaining, etc. (i.e., the embodiment of FIG. 7). The higher capacity layer is then designed as the upstream layer to create devices with reduced capacity variability compared to random layering.
  • For devices with more than two layers, alternative methods are possible. The membranes could be placed into a groups of highest, lowest and a mid-range capacity. The top layer is chosen from the highest capacity group, the bottom layer from the lowest capacity group; and the intermediate layers from the mid-range group. The intermediate layers may be chosen randomly, or the may be arranged in descending order of capacity rating, and paired off as described to fill in the layers required for the device. Alternatively the mid-range group may be sub-divided into sub-groups of descending average rating for each sub-group. The intermediate layers from second layer to second from bottom layer are then chosen from the descending capacity rating groups in order.
  • A calculation relating to selective layering combinations was conducted for a three-layer device using the same population of membranes used for Table 1. Because there are three layers, the capacity distribution was divided into three capacity sections (lower, center, and upper), with each section containing an equal portion of the total population.
  • In what is designated as embodiment A, the top (upstream) layer is selected from the middle portion of the distribution, and the middle layer and bottom (downstream) layers are selected randomly from the rest of the distribution. In what is designated as embodiment B, the top layer is selected from the middle portion of the distribution, the middle layer is selected from the upper section of the distribution, and the bottom- layer is selected from the lower end of the distribution. In what is designated as embodiment C, the top layer is selected from the upper portion of the distribution, the middle layer is selected from the upper section of the distribution, and the bottom layer is selected from the lower end of the distribution. In what is designated as embodiment D, the top and bottom layers are selected as per FIG. 7, and the middle layer is selected from the center portion of the distribution.
  • Table 2 shows the performance variability resulting from the various embodiments. All of the embodiments are advantageous compared to random layering (the existing state of the art) with respect to either mean capacity, performance consistency, or both. Embodiment B offers the lowest capacity variability, along with low LRV and permeability variability, but with the same mean capacity as random layering. Embodiment D offers the lowest LRV and permeability variability, and has lower capacity variability and higher mean capacity than random layering. Either embodiment B or D may be preferred depending on the value placed on mean capacity compared to capacity consistency.
  • TABLE 2
    Capacity LRV Perme-
    Layering Capacity Std. LRV Std. ability Permeability
    Strategy Mean Dev. Mean Dev. Mean Std. Dev.
    Random 600 70 10.5 1.04 13.3 0.80
    A 600 19 10.5 0.63 13.3 0.53
    B 600 19 10.5 1.04 13.3 0.80
    C 678 34 10.5 0.63 13.3 0.53
    D 678 34 10.5 0.52 13.3 0.45

Claims (18)

  1. 1. A filtration device comprising a housing having an inlet and an outlet spaced from said inlet, and a plurality of membranes including at least an upstream membrane and at least a downstream membrane, wherein each said membrane has substantially the same rating, and wherein said upstream membrane has a first predetermined performance characteristic selected such that said filtration device has a second predetermined performance characteristic regardless of the performance characteristics of said downstream membrane.
  2. 2. The filtration device of claim 1, wherein said performance characteristic is capacity.
  3. 3. The filtration device of claim 1, wherein said performance characteristic is permeability.
  4. 4. The filtration device of claim 1, wherein said performance characteristic is retention.
  5. 5. A multi-layered membrane filter device made from at least two membrane layers, said layers having at least one higher and one lower capacity rating, said device having a capacity greater than the average of the capacity ratings of the membranes used to make the device, and less than or equal to the higher capacity membrane.
  6. 6. A population of multi-layered membrane filter devices made from at least two membrane layers, said layers having at least one higher and one lower capacity rating, said devices having an population average capacity greater than the average of the capacity ratings of the membranes used to make the device, and less than or equal to the highest capacity membrane.
  7. 7. A filtration device comprising a housing having an inlet and an outlet spaced from said inlet, and a plurality of membranes including at least an upstream membrane and at least a downstream membrane, wherein each said membrane is selected from a population of membranes having a capacity such that the capacities of said population define a distribution, and wherein said upstream membrane has a capacity at about the median of said distribution, and said at least one downstream membrane has a capacity at either the high end or low end of said distribution.
  8. 8. The filtration device of claim 7, wherein said capacity distribution of said population of membranes is a Gaussian distribution.
  9. 9. The device of claim 7, wherein there are at least three membranes.
  10. 10. The device of claim 9, wherein the membrane between said upstream and downstream membranes has a capacity at either the high end or the low end of said distribution.
  11. 11. The device of claim 7, wherein there are at least four membranes.
  12. 12. A filtration device comprising a housing having an inlet and an outlet spaced from said inlet, and a plurality of membranes including at least an upstream membrane and at least a downstream membrane, wherein each said membrane is selected from a population of membranes having a capacity such that the capacities of said population define a distribution, and wherein said upstream membrane has a capacity at the high end of said distribution, and said at least one downstream membrane has a capacity at the low end of said distribution.
  13. 13. The filtration device of claim 12, wherein said capacity distribution of said population of membranes is a Gaussian distribution.
  14. 14. The device of claim 12, wherein there are at least three membranes.
  15. 15. The device of claim 12, wherein there are at least four membranes.
  16. 16. The filtration device of claim 12, wherein said high end of said distribution is at least as high as the median of said distribution.
  17. 17. The filtration device of claim 12, wherein said low end of said distribution is at least as low as the median of said distribution.
  18. 18. A filtration device comprising a housing having an inlet and an outlet spaced from said inlet, and a plurality of membranes including at least an upstream membrane and at least a downstream membrane, wherein each said membrane is selected from a population of membranes having a capacity such that the capacities of said population define a distribution, and wherein said upstream membrane has a capacity higher than the median capacity of said distribution by a certain amount, and said at least one downstream membrane has a capacity lower than the median capacity of said distribution by about said same certain amount.
US12387636 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters Abandoned US20090277824A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US12715608 true 2008-05-09 2008-05-09
US12387636 US20090277824A1 (en) 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
US12387636 US20090277824A1 (en) 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters

Publications (1)

Publication Number Publication Date
US20090277824A1 true true US20090277824A1 (en) 2009-11-12

Family

ID=40911050

Family Applications (2)

Application Number Title Priority Date Filing Date
US12387576 Active 2030-09-22 US8733556B2 (en) 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters
US12387636 Abandoned US20090277824A1 (en) 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters

Family Applications Before (1)

Application Number Title Priority Date Filing Date
US12387576 Active 2030-09-22 US8733556B2 (en) 2008-05-09 2009-05-05 Method for reducing performance variability of multi-layer filters

Country Status (5)

Country Link
US (2) US8733556B2 (en)
EP (4) EP2123348A1 (en)
JP (1) JP5587560B2 (en)
CN (1) CN101642679B (en)
ES (2) ES2629680T3 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090277835A1 (en) * 2008-05-09 2009-11-12 Sal Giglia Method for reducing performance variability of multi-layer filters
CN101992025A (en) * 2009-08-13 2011-03-30 米利波尔有限公司 Method for improved scaling of filters

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5017292A (en) * 1990-05-10 1991-05-21 Millipore Corporation Membrane, process and system for isolating virus from solution
US5457986A (en) * 1992-06-30 1995-10-17 Millipore Corporation Process for evaluating solute retention characteristics of membranes
US5490937A (en) * 1990-09-17 1996-02-13 Genentech, Inc. Tangential flow filtration process and apparatus
US5968373A (en) * 1997-12-22 1999-10-19 Aaf International Filter arrangement having at least two successive layers having predetermined spacing and its method for making
US6019809A (en) * 1990-10-19 2000-02-01 Donaldson Company, Inc. Filtration arrangement
US20020175124A1 (en) * 1999-09-14 2002-11-28 Gabriel Tkacik High-resolution virus removal methodology and filtration capsule useful therefor
US20030209485A1 (en) * 2001-05-22 2003-11-13 Willem Kools Process of forming multilayered structures
US6669905B1 (en) * 1998-05-21 2003-12-30 Baxter International Inc. Systems and methods for collecting plasma that is free or virtually free of cellular blood species
US6872431B2 (en) * 1995-11-17 2005-03-29 Donaldson Company, Inc. Filter material construction and method
US20070187319A1 (en) * 2000-05-24 2007-08-16 Millipore Corporation High-throughput asymmetric membrane
US20080028834A1 (en) * 2004-07-06 2008-02-07 Lieven Gevers High Throughput Screening for Rapid Development of Membranes and Membrane Processes
US20090277835A1 (en) * 2008-05-09 2009-11-12 Sal Giglia Method for reducing performance variability of multi-layer filters

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2832835B2 (en) 1988-12-21 1998-12-09 旭化成工業株式会社 Virus removal method
JPH07124450A (en) 1993-11-01 1995-05-16 Fuji Photo Film Co Ltd Precision filtration membrane cartridge filter
JP3040298B2 (en) 1993-12-16 2000-05-15 オルガノ株式会社 Filtration membrane for filtration apparatus and microparticles number measurement for particle number measurement in ultrapure water
JPH1024224A (en) 1996-04-10 1998-01-27 Mitsubishi Chem Corp Laminated porous film and its utilization
DE60037262T2 (en) 1999-09-14 2008-11-27 Greg Danvers Kazan Apparatus and membrane for removing viruses
WO2002005934A3 (en) 2000-05-31 2002-06-20 Fendya Thomas J Membrane packets, methods for making membrane packets, and membrane packet assemblies
JP3570713B2 (en) 2001-11-06 2004-09-29 東洋濾紙株式会社 Beer filtration use laminated filter
CN100482316C (en) 2005-07-26 2009-04-29 清华大学 Microbe filter made from micropore ceramics

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5017292A (en) * 1990-05-10 1991-05-21 Millipore Corporation Membrane, process and system for isolating virus from solution
US5490937A (en) * 1990-09-17 1996-02-13 Genentech, Inc. Tangential flow filtration process and apparatus
US6019809A (en) * 1990-10-19 2000-02-01 Donaldson Company, Inc. Filtration arrangement
US5457986A (en) * 1992-06-30 1995-10-17 Millipore Corporation Process for evaluating solute retention characteristics of membranes
US6872431B2 (en) * 1995-11-17 2005-03-29 Donaldson Company, Inc. Filter material construction and method
US5968373A (en) * 1997-12-22 1999-10-19 Aaf International Filter arrangement having at least two successive layers having predetermined spacing and its method for making
US6669905B1 (en) * 1998-05-21 2003-12-30 Baxter International Inc. Systems and methods for collecting plasma that is free or virtually free of cellular blood species
US20020175124A1 (en) * 1999-09-14 2002-11-28 Gabriel Tkacik High-resolution virus removal methodology and filtration capsule useful therefor
US7108791B2 (en) * 1999-09-14 2006-09-19 Millipore Corporation High-resolution virus removal methodology and filtration capsule useful therefor
US20070187319A1 (en) * 2000-05-24 2007-08-16 Millipore Corporation High-throughput asymmetric membrane
US20030209485A1 (en) * 2001-05-22 2003-11-13 Willem Kools Process of forming multilayered structures
US20080028834A1 (en) * 2004-07-06 2008-02-07 Lieven Gevers High Throughput Screening for Rapid Development of Membranes and Membrane Processes
US20090277835A1 (en) * 2008-05-09 2009-11-12 Sal Giglia Method for reducing performance variability of multi-layer filters

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090277835A1 (en) * 2008-05-09 2009-11-12 Sal Giglia Method for reducing performance variability of multi-layer filters
US8733556B2 (en) 2008-05-09 2014-05-27 Emd Millipore Corporation Method for reducing performance variability of multi-layer filters
CN101992025A (en) * 2009-08-13 2011-03-30 米利波尔有限公司 Method for improved scaling of filters
US20110185552A1 (en) * 2009-08-13 2011-08-04 Millipore Corporation Method For Improved Scaling Of Filters
US8387256B2 (en) 2009-08-13 2013-03-05 Emd Millipore Corporation Method for improved scaling of filters
US8904641B2 (en) 2009-08-13 2014-12-09 Emd Millipore Corporation Method for improved scaling of filters

Also Published As

Publication number Publication date Type
EP2460580A1 (en) 2012-06-06 application
EP2457642B1 (en) 2017-01-11 grant
ES2629680T3 (en) 2017-08-14 grant
EP2460580B1 (en) 2017-06-21 grant
JP2009269027A (en) 2009-11-19 application
US20090277835A1 (en) 2009-11-12 application
EP3075439A1 (en) 2016-10-05 application
EP2123348A1 (en) 2009-11-25 application
CN101642679A (en) 2010-02-10 application
JP5587560B2 (en) 2014-09-10 grant
EP2457642A1 (en) 2012-05-30 application
CN101642679B (en) 2013-11-13 grant
ES2621225T3 (en) 2017-07-03 grant
US8733556B2 (en) 2014-05-27 grant

Similar Documents

Publication Publication Date Title
Duclos-Orsello et al. A three mechanism model to describe fouling of microfiltration membranes
US5256294A (en) Tangential flow filtration process and apparatus
Ye et al. Fouling mechanisms of alginate solutions as model extracellular polymeric substances
US5500167A (en) Method of preparing a supported microporous filtration membrane
US6776940B2 (en) Process for making a three-zone microporous membrane
US5417101A (en) Method and apparatus for testing the integrity of filter elements
Ochoa et al. Ultrafiltration of vegetable oils: degumming by polymeric membranes
US5853582A (en) Tubular inorganic filter element having increased mechanical strength and increased filter area
Güell et al. Membrane fouling during microfiltration of protein mixtures
US6054051A (en) Tangential-flow filtration system
JP2008272636A (en) Multilayer microporous membrane
Shukla et al. Performance of ultrafiltration membranes in ethanol–water solutions: effect of membrane conditioning
EP0245863A2 (en) Composite ultrafiltration membranes
EP2113298A1 (en) Hollow fibre membrane with improved permeability and selectivity
Jonsson et al. Fouling phenomena in microporous membranes. Flux decline kinetics and structural modifications
US20020108906A1 (en) Multi-stage filtration and softening module and reduced scaling operation
Cuartas-Uribe et al. A study of the separation of lactose from whey ultrafiltration permeate using nanofiltration
EP0083489B1 (en) Membrane systems for filtration
US5607586A (en) Multichannel inorganic element for filtering a fluid
Wang et al. Characterization and applications of nanofiltration membranes: State of the art
US20120061332A1 (en) Method for retrovirus removal
WO2001014047A1 (en) Filter membranes for physiologically active substances
Luque et al. A new coiled hollow‐fiber module design for enhanced microfiltration performance in biotechnology
Metsämuuronen et al. Critical flux in ultrafiltration of myoglobin and baker’s yeast
De et al. Modeling of ultrafiltration process for a two-component aqueous solution of low and high (gel-forming) molecular weight solutes

Legal Events

Date Code Title Description
AS Assignment

Owner name: MILLIPORE CORPORATION, MASSACHUSETTS

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:GIGLIA, SAL;KRISHNAN, MANI;SATAV, NITIN;REEL/FRAME:022986/0810

Effective date: 20090713

AS Assignment

Owner name: EMD MILLIPORE CORPORATION, MASSACHUSETTS

Free format text: CHANGE OF NAME;ASSIGNOR:MILLIPORE CORPORATION;REEL/FRAME:027620/0891

Effective date: 20120101