METHOD OF PREPARTING MEMBRANES

Abstract

A method of preparing a membrane comprising the steps of: a) mixing together a membrane-forming polymer, a water-soluble polyetheramine, and a solvent, said mixture containing no component which will react chemically with the polyetheramine; and b) casting said mixture to form the polymer into a solid membrane.

Claims

1-7. (canceled)

8. A method of preparing a membrane comprising the steps of: a) mixing together a membrane-forming polymer, a water-soluble polyetheramine which is a polyether with at least one primary or secondary amine group attached to the polyether backbone, and a solvent, said mixture containing no component which will react chemically with the polyetheramine; and b) casting said mixture to form the polymer into a solid membrane.

9. A method as claimed in claim 8, in which the water solubility of the polyetheramine is at least 0.1% w/v at 21° C.

10. A method as claimed in claim 9, in which the water solubility of the polyetheramine is at least 0.2% w/v at 21° C.

11. A method as claimed in claim 10, in which the polyetheramine is miscible with water at 21° C.

12. A method as claimed in claim 8, in which the polyetheramine has a molecular weight of up to 2,500.

13. A method as claimed in claim 8, in which the polyetheramine contains two or more ethylene oxide and/or propylene oxide monomer units and at least one primary or secondary amine unit —NHX where X is a hydrogen atom or a C.sub.1-4alkyl group.

14. A method as claimed in claim 8, in which the polyetheramine is a mono- or di-amine having the schematic formula:
Y—PAO—Y′  (I) or a mono-, di- or tri-amine having the schematic formula: ##STR00016## in which each of Y, Y′ and Y″ independently represents an end group at least one of which includes a primary or secondary amine group, and PAO represents a polyalkyleneoxide chain consisting of at least two ethylene oxide and/or propylene oxide monomer units.

15. A method as claimed in claim 14, in which the polyetheramine has the schematic formula:
Y—(O—CH.sub.2—CH(CH.sub.3)).sub.a—(O—CH.sub.2—CH.sub.2).sub.b—Y′  (III) in which a represents the number of propylene oxide monomer units present and b represents the number of ethylene oxide monomer units present; or the general formula: ##STR00017## in which R represents a hydrogen atom or a methyl group, R′ represents a hydrogen atom, a methyl or an ethyl group, d is 0 or 1, and c, e and f are the number of propylene oxide and/or ethylene oxide monomer units present.

16. A method as claimed in claim 8, in which the polyetheramine has one of the formulae: ##STR00018## in which R is H or CH.sub.3, and x and y are the numbers of propylene oxide and/or ethylene oxide monomer units in the polyether chain; ##STR00019## in which x is the number of propylene oxide monomer units in the polyether chain; ##STR00020## in which x and z are the number of propylene oxide monomer units in two blocks in the polymer chain, and y is the number of EO monomer units in the polyether chain; ##STR00021## in which x is 2 or 3; ##STR00022## in which x, y and z together represent the total number of propylene oxide monomer units present in the branched chain polymer, n is 0 or 1, and R is hydrogen, methyl or ethyl; or a compound of one of the formulae (V) to (IX) above in which one or more of the NH.sub.2 end groups has been converted into a secondary amine group.

17. A method as claimed in claim 16, in which the polyetheramine has the formula (IX), in which the number of moles of polyethylene oxide is between 5 and 6; or in which the polyetheramine has the formula (VI), in which x in the formula (VI) is on average from 6 to 7.

18. A method as claimed in claim 8, in which the membrane-forming polymer is selected from cellulose acetate/triacetate; polyamide; polypiperazine; polybenzimidazoline; polysulfone; polyol; polyacrylonitrile; polyethersulfone; polysulfone; poly(phthalazinone ether sulfone ketone; poly(vinyl butyral); polyvinylidene fluoride; poly(tetrafluoroethylene); polypropylene; polyethylene; and polyetheretherketone.

19. A method as claimed in claim 8, in which step (b) comprises immersing the mixture produced in step (a) in a medium in which the polymer is insoluble.

20. A method as claimed in claim 8, in which the membrane obtained in step (b) is subsequently provided with a coating.

21. A membrane preparable by the method as claimed in claim 8.

22. A membrane as claimed in claim 21, which is an ultrafiltration membrane, a nanofiltration membrane or a reverse osmosis membrane.

23. A membrane comprising a water-soluble polyetheramine which is a polyether with at least one primary or secondary amine group attached to the polyether backbone, wherein the membrane contains no component which is chemically reactive with the polyetheramine.

24. A membrane as claimed in claim 23, wherein the water solubility of the polyetheramine is at least 0.1% w/v at 21° C.

25. A membrane as claimed in claim 24, which is an ultrafiltration membrane, a nanofiltration membrane or a reverse osmosis membrane.

26. A membrane which has a molecular weight cut off of at least 20 kDa and a flux of at least 20 GFD/psi.

27. A membrane as claimed in claim 26, which is an ultrafiltration membrane, a nanofiltration membrane or a reverse osmosis membrane.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0098] FIGS. 1 to 16 and 18 to 26 show the results of the testing described in the Examples.

[0099] FIG. 1 shows a comparison of flux rates and molecular weight cut-offs of commercially available membranes compared with the membranes of examples 1 to 5.

[0100] FIG. 2 shows a comparison of flux rates and molecular weight cut-offs of the membrane of example 1 and comparative examples 6 and 7.

[0101] FIG. 3 shows the effect of different quantities of polyetheramines added in the membranes of examples 1, 8 and 9.

[0102] FIG. 4 shows the flux decline of a comparative GE Osmonice UF membrane in a BSA test.

[0103] FIG. 5 shows the flux decline of the membrane of example 1 in a BSA test.

[0104] FIG. 6 shows BSA rejection for the membrane of example 1.

[0105] FIG. 7 shows SEM micrographs of the membrane of example 1, surface on the left and cross-section on the right.

[0106] FIG. 8 shows SEM micrographs of the membrane of example 2, surface on the left and cross-section on the right.

[0107] FIG. 9 shows SEM micrographs of the membrane of example 3, surface on the left and cross-section on the right.

[0108] FIG. 10 shows SEM micrographs of the membrane of example 4, surface on the left and cross-section on the right.

[0109] FIG. 11 shows SEM micrographs of the membrane of example 5, surface on the left and cross-section on the right.

[0110] FIG. 12 shows SEM micrographs of the commercially-available membrane Millipore Biomax 30 kDa, pore size 10-20 nm, surface on the left and cross-section on the right.

[0111] FIG. 13 shows SEM micrographs of the commercially-available membrane Sterlitech PVDF, surface on the left and cross-section on the right.

[0112] FIG. 14 shows an SEM of the surface of Sterlitech PES00325100 0.03 micron membrane.

[0113] FIG. 15 shows an SEM of the surface of the membrane of example 8.

[0114] FIG. 16 shows an SEM of the surface of the membrane of example 9.

[0115] FIG. 17 shows a schematic representation of the custom-built pilot casting line used to make the membranes described in the Examples.

[0116] FIGS. 18 to 24 show SEM micrographs of the membranes of examples 15 to 21, respectively, surface on the left and cross-section on the right.

[0117] FIGS. 25 and 26 show SEM micrographs of the comparative membranes of example 22, made using PEO and acrylamide, respectively, as additives, surface on the left and cross-section on the right.

[0118] The following Examples illustrate the invention.

[0119] In the Examples, various polyetheramines were used. In each case, their water solubility was measured using DLS and the following protocol. Malvern ZetaSizer Nano-S light scattering (DLS) equipment was used to observe formation of particles at a given concentration of polyetheramine in water. Combination of count rate and attenuator monitoring was used to determine the increase in number of particles as the concentration of polyetheramine increased to determine solubility. Polystyrene latex was used as the reference material (RI: 1.590; absorption: 0.010 at 633 nm), and water as the dispersant (viscosity: 0.9781 cP; RI: 1.330). The measurements were carried out in Science Brand disposable microcuvettes with a sample volume of 100 at 21° C. Each sample was measured 5 times, each measurement was the average of 11 runs. Where the polyetheramine and water are not completely miscible, DLS measurements were continued up to the point where visible phase separation was observed.

EXAMPLE 1. PS

Membrane (20-30 nm Pore Size)

[0120] Membrane in Example 1 was prepared as follows: 567 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 63 g of polyetheramine (Huntsman JT403, MW 440, miscible in water) was dissolved in 2422 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on a nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm (feet per minute). The membrane was formed in a quench tank, where the dissolved polymer dope is immersed in water (a non-solvent). The casting process was performed using a pilot casting line custom built by Cut Membranes Canada, a schematic representation of which is shown in FIG. 17. In FIG. 17, 1 is the quench tank containing water; 2 is a roll of backing fabric 3 which passes over a granite stone support 4. Dope is added at 6, using doctor blade 5. Additional rollers 8 carry the dope on the fabric through the quench tank, and the finished membrane is taken up by roller 7 powered by an electric motor.

EXAMPLE 2. PS

Membrane (12 nm Pore Size)

[0121] Membrane in Example 2 was prepared as follows: 696.78 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 63 g of polyetheramine (Huntsman JT403) were dissolved in 2292 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on a nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 3. PES

Membrane (20-30 nm Pore Size)

[0122] Membrane in Example 3 was prepared using as follows: 630 g of polyethersulfone (BASF ULTRASON® S6020p), 331 g of isopropanol (Sigma-Aldrich 278475) and 33.18 g of polyetheramine (Huntsman JD400, MW 430, miscible in water) were dissolved in 2321 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on a nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 4. PVDF

Membrane (50-100 nm Pore Size)

[0123] Membrane in Example 4 was prepared as follows: 464 g of polyvinylidene fluoride (Solvay Solef® 1015/1001), 18.2 g of formic acid (Sigma-Aldrich F0507), used to prevent the PVDF from being cross-linked in basic conditions, and 54 g of polyetheramine (Huntsman JT403) were dissolved in 2778 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on the nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 5. PS

Membrane (20-40 nm Pore Size)

[0124] Membrane in EXAMPLE 5 was prepared as follows: 567 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 63.04 g of polyetheramine (Huntsman JM600, MW 600, miscible in water) were dissolved in 2420 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope is degassed and cast on a nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 6 (Comparative). PS Polyethyleneimine 0.8 kDa

Membrane (20-50 nm Pore Size)

[0125] Membrane in Example 6 was prepared using as follows: 630 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 8.29 g of PEI 0.8 kDa (Sigma-Aldrich 408719) were dissolved in 2412 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on a nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 7 (Comparative). PS Polyallylamine 65 kDa

Membrane

[0126] Membrane in Example 7 was prepared using as follows: 630 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 33.18 g of polyallylamine 65 kDa (Sigma-Aldrich 479144) were dissolved in 2387 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on the nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.25 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 8 (20-30 nm Pore Size)

[0127] Membrane in example 8 was prepared using as follows: 567 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 33.15 g of polyetheramine (Huntsman JT403) were dissolved in 2450 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on the nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.30 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 9 (20-30 nm Pore Size)

[0128] Membrane in example 9 was prepared using as follows: 567 g of polysulfone (BASF ULTRASON® S6010), 265 g of 2-Methoxyethanol (Sigma-Aldrich 284467) and 16.6 g of polyetheramine (Huntsman JT403) were dissolved in 2467 g of N,N-dimethylformamide (Sigma-Aldrich D158550) at 70 deg C. under mechanical stirring for 8 h. Upon cooling to room temperature the dope was degassed and cast on the nonwoven polyester backing fabric (Hirose 05TH100) using doctor blade (gap 0.28 micron) at 30 fpm. The membrane was formed in the quench tank, where the dissolved polymer dope is immersed in water. The casting process was performed using the pilot casting line mentioned above.

EXAMPLE 10 (20-30 nm Pore Size)

[0129] The method of Example 1 was followed exactly except that instead of the backing fabric supplied by Hirose, a nonwoven polyester backing fabric supplied by Awa (AWA#2) was used.

EXAMPLE s 11 to 14: Characterization and Testing of Membranes

EXAMPLE 11: Pure Water Flux and Molecular Weight Cut-Off

[0130] Pure water flux tests were carried out using Amicon (EMD Millipore, 5124| Stirred Cell Model 8400, 400 mL) stirred cells. The membrane was fitted into the cell and the cell was filled with deionized water and pressurized with compressed air from 0-5 bar of pressure. Permeate was collected for 12 seconds and permeability GFD/PSI calculated based on known surface area and pressure.

[0131] Molecular weight cut off and rejection properties of the membranes were tested using dextran as a solute. Additionally rejection of Bovine Serum Albumin (BSA) and pepsin was used for rejection and fouling tests.

[0132] Molecular cut off testing was performed using a mixture of dextrans (America Polymer Standards Corporation) at varied molecular weight (1 kDa-10 000 kDa) in PB buffer pH 7.5. The test membrane was placed in an Amicon stirred cell (EMD Millipore, 5121, Model 8010, 10 mL). 5 ml of feed solution containing mixture of dextrans at a concentration of 0.5 g/liter was run through the membrane at 0.094 ml/min using a peristaltic pump (Cole-Palmer Masterflex L/S model no. 4551-10) with the first 0.5 ml discarded and subsequent 0.5 ml of permeate collected for chromatography and comparison with the feed. A low flow rate combined with stirring allows for measurement unaffected by concentration polarization. Permeate and feed were evaluated for the molecular weight distribution using HPLC Agilent Technologies 1260 Infinity using RID detector with three column set up (PL Aquagel-OH 1000-60,000; 20,000-200,000; and 200,000 to 10 million respectively). The distribution of the feed was compared to distribution of the permeate at a given elution volume, resulting in rejection following the formula:

R=RID(feed)−RID(permeate)/RID(feed)

[0133] Molecular cut-off for a particular membrane is defined as the molecular weight of the solute where 90% rejection in observed.

[0134] FIG. 1 presents the membrane performance data comparison of various commercially available membranes, including the market leader Millipore BioMax 30 kD PES membrane, with the membranes where polyetheramines were added to the formulation of different membrane-forming polymers. It is clearly seen that molecular weight cut-offs can be easily manipulated using various polyetheramines and that the resulting membranes are characterized by improved permeability over competitor membranes at the corresponding molecular weight cut off. The commercially available ultrafiltration membranes are designed to be operated at much higher pressure (30PSI+), while membranes according to the invention can operate successfully in the range from 1-100+PSI.

[0135] FIG. 2 presents the membrane performance data comparison of the membrane prepared according to example 1, where polyetheramine was used as a dope additive, with the membranes prepared using other poly-amine containing additives of comparative examples 6 and 7. It is clear that membranes where polyetheramines were used as a dope additive are characterized by improved permeability compared with membranes prepared using polyamine containing additives.

[0136] FIG. 3 shows the effect of the different amounts of polyetheramine added to the membranes prepared in examples 1, 8 and 9.

EXAMPLE 12: BSA Rejection and Fouling Test

[0137] BSA (bovine serum albumin) rejection and fouling studies were performed using a GE Osmonics High Pressure RO Cell (Sterlitech 1230060) under tangential flow in recycling mode. The pressure of the system was adjusted to the permeability of the membrane in order to keep cross-flow at the steady value of 0.15 GPM with 4 liters of solution used in total.

[0138] Feed: 1 g/liter of BSA (Fitzgerald 30-AB70), 50 mM NaCl (Sigma-Aldrich S9888) in deionized water

[0139] The membranes tested for rejection and fouling using BSA were tested in tangential flow configuration under specified pressure. Sampling of permeates and feed was simultaneous, at 30 minutes intervals, and rejection was measured using the Quick Start™ Bradford Protein Assay (BIORAD).

[0140] Flux was measured by collecting the permeate in standardized unit of time using known surface area to calculate GFD/PSI values. Comparative flux decline studies (fouling with BSA) of membranes prepared with the membrane of example 1 and GE Osmonics polysulfone 30 kDa (Sterlitech™) membrane were conducted. Flux decline was monitored over a fixed volume of permeated water (liter/m.sup.2).

Test 1: GE Osmonics 30 KDa Membrane

[0141] Initial flow at 32 psi=0.067 GFD [0142] Testing conditions=50 mM NaCl+1 g/L BSA feed solution, Operating Pressure=32 psi. Total volume of feed was 4 L and a recycle mode of feed fouling was applied to the system. The results are shown in FIG. 4.

Test 2: Polyetheramine Membrane of Example 1 25 kDa

[0143] Initial flow at 4.5 psi=8.4 GFD [0144] Testing conditions: 50 mM NaCl+1 g/L BSA feed solution, Operating Pressure=4.5 psi. Total volume of feed was 4 L and a recycle mode of feed fouling was applied to the system.

[0145] It is evident from the results shown that the pressure requirement for both membranes are different with 32 PSI needed to drive the permeation through GE Osmonics 30 kDa PS membrane and 4 PSI for the membrane from example 1 due to high permeability nature of this membrane.

[0146] Both membranes were characterized by over 90% of BSA rejection with the membrane of example 1 showing 98+% BSA rejection. The results for the membrane of example 1 are shown in FIG. 6. The initial permeability of the membrane from example 1 was 89 times higher than of the GE Osmonics membrane, with both membranes showing flux decline over 100 liters/m.sup.2. The membrane from example 1 showed a steady state flux value 3 times of the GE Osmonics membrane.

EXAMPLE 13: SEM

[0147] Scanning electron microscopy was performed using an FEI XL30 environmental scanning electron microscope and SEM Pin Stubs (Ted Pella, Inc. 16111). Membrane samples were cut and mounted to the stubs using carbon tape (Ted Pella, Inc. 16085-1). Samples were coated with thin layer of gold to avoid charging in the SEM chamber. Gold was spattered for 30 seconds on VG/Polaron SC 7620. Surface and cross-section images were collected for each membrane sample.

[0148] In the membranes studied the benefit of using polyetheramines is illustrated by performance numbers presented in Example 10 and 11 above, and visualized in micrographs showing surface porosity and the cross-section of the membranes. The membranes prepared according to examples 1-5 were further compared to commercial competitive products. SEM images of membranes prepared according to examples 1-5 are presented in FIGS. 7-11 respectively. Each of these figures shows the surface on the left, and the cross-section on the right. The images presented in FIGS. 7-11 show the pore size and pore size distribution improved or competitive, as compared to commercially available membranes shown in FIG. 12 for Millipore Biomax 30 kDa, pore size 10-20nm, and FIG. 13 for Sterlitech PVDF membrane. Additional beneficial characteristics of ultrafiltration membranes prepared according to examples 1-5, namely lack of micro and macro voids, can be observed in the cross-sectional images in FIGS. 7-11. This is important for membrane stability at higher pressures, where compaction is usually observed. This is especially important when using membranes as a support material for coating of NF and RO membranes where higher pressures lead to significant compaction and reduction of flow.

[0149] The beneficial impact of polyetheramine on the structure of the membrane is especially visible when comparing the cross-sectional images of PVDF based membrane of example 4 prepared with polyetheramine shown in FIG. 10 with the cross-section micrograph of PVDF membrane purchased from Sterlitech shown in FIG. 13.

[0150] FIG. 14 shows the commercially available PES based 0.03 micron membrane from Sterlitech (PES00325100). This membrane is said to be characterized by 30 nm pore size, but this is not supported by the surface porosity shown in FIG. 14. FIGS. 9, 15 and 16 show, in comparison, the membranes of examples 3, 8 and 9, respectively, which clearly show superior surface characteristics.

EXAMPLE 14: Bacterial Removal Study: Use of Membrane of Examples 1 and 10 to Remove E. coli BL21 (DE3)

[0151] E. coli suspension was prepared as follows. 500 ml of LB broth (5 g Tryptone, 2.5 g NaCl & 2.5 g yeast in 500 ml of Nano pure water) was prepared. The sample was dispensed into two 1 L culture flasks (250 ml in each flask) and autoclaved. The culture flasks were left to cool down until starting the overnight culture. The overnight culture was prepared by addition of 250u1 of sterilized kanamycin and 10 ul of the Canadian culture seed into each flask using aseptic techniques. The culture was grown at 37° C., 225 rpm for 16hrs. The OD.sub.600 was measured as 2.14. The overnight culture was dispensed into sterilized centrifuge bottle and centrifuged at 5000 rpm for 10 min. The supernatant was poured out. The left-over cell pellet was dispensed in the sterilized PBS using the original volume of culture and centrifuged at 5000 rpm for another 10 min. After disposing of the supernatant the cell pellet was dispensed into PBS. The cells were shaken to get even distribution of cells and PBS was used to bring up the volume to 500m1. The OD.sub.600 at this point was measured as 2.257. This suspension in PBS buffer was used as feed, OD.sub.600=2.14, the estimated number of cells is about 10.sup.9 CFU/ml.

Membrane Testing Procedure:

[0152] 1. 76 mm membrane of example 1 was stamped and assembled in the Amicon cell (model 8400); the pure water flux of the membrane of example 1 was tested at 5 psi [0153] 2. 50 ml of PBS buffer was rinsed through the membrane of example 1 under 5 psi, flow data is measured for 30 s. PBS was collected for optical density (OD) measurement and as cell counting negative control [0154] 3. 200 ml of E-coli was used as feed. The test was performed at 5 psi with 300 rpm stirring, recorded the flow data at the starting point and then collected 1-2 ml of permeate for plating in order to determine the bacterial removal percentage. Agar plates were prepared according to manufacturer instructions. Samples for OD measurement were collected at different time. [0155] 4. Performed the test until the E-coli feed ran out, washed the membrane thoroughly with deionised water, and then retested the pure water flux at 5 psi to compare it with the flux before testing.

[0156] The following results were obtained:

[0157] Feed: CFU/ml=21×10.sup.7/0.2 ml=1.05×10.sup.9

[0158] Permeate: CFU/ml=20.5/0.2 ml=1.025×10.sup.2

[0159] This corresponds to a bacterial rejection % of 1-1.025×10.sup.2/1.05×10.sup.9, or 0.9999999.

[0160] The same procedure was carried out using the membrane of Example 10. In this case, absolute bacterial rejection was obtained.

EXAMPLE s 15-21: Comparison of Different Polyetheramines

[0161] The process of Example 1 was repeated using each of the following polyetheramines manufactured by Huntsman. For each experiment, scanning electron microscopy of the resulting membrane was performed as described in Example 13, and the resulting micrographs (surface of the membrane on the left and cross-section on the right) are shown in FIGS. 18 to 22.

EXAMPLE 15: Jeffamine™ M1000, MW 1,000, miscible in water (FIG. 18)

EXAMPLE 16: Jeffamine™ D400, MW 430, miscible in water (FIG. 19)

EXAMPLE 17: Jeffamine™ ED600, MW 600, miscible in water (FIG. 20)

EXAMPLE 18: Jeffamine™ ED2003, MW 2,000, miscible in water (FIG. 21)

EXAMPLE 19 (comparative): Jeffamine™ T3000, MW 3,000, solubility in water 0.050% w/v (FIG. 22)

EXAMPLE 20 (comparative): Jeffamine™ D4000, MW 4,000, solubility in water 0.0025% w/v (FIG. 23)

EXAMPLE 21 (comparative): Jeffamine™ T5000, MW 5,000, solubility in water 0.040% w/v (FIG. 24).

[0162] The results clearly show that when using the specific polyetheramines according to the invention as additives during the casting process, a high-quality surface with uniform pore size and pore size distribution is obtained, together with a lack of micro and macro voids. In contrast, use of alternative polyetheramines not according to the invention leads to a visibly inferior membrane.

EXAMPLE 22 (Comparative)

[0163] The process of Example 1 was repeated save that the polyetheramine additive was replaced by the same quantity of a different additive. The additives used were of the type listed in WO 2011/069050: a polyalkylene oxide, an acrylamide and a catecholamine. In all cases, it was found that the dope solution either phase separated or micro-phase separated, leading to unstable dope solutions resulting in either membrane not forming or formed membranes characterized by defects.

[0164] In the case of polyalkylene oxide (PEO 100,000, Sigma-Aldrich 181986) the dope solution was very cloudy and micro-phase separated, and the resulting membrane was patchy and defective as shown in FIG. 25. In flux tests the membrane was characterized by unmeasurable molecular cut-off (within the range of the experimental design) and low pure water flow characteristics. In the case of membranes where acrylamide (Sigma-Aldrich A3553) was used as an additive, the formed membrane was characterized by increased pore size and hence a higher molecular cut-off, and lower pure water flux characteristics, as compared to the membrane of example 1. FIG. 26 shows SEMs of the surface and cross-section of the membrane. In the case of catecholamine (dopamine hydrochloride, Sigma-Aldrich H8502) as an additive, the dope was found to phase separate during the degassing stage making it impossible to cast a useful membrane.