Reverse osmosis membrane and method of use

Abstract

The invention relates to a membrane, and method of manufacture of a membrane for reverse osmosis having a porous substrate, and a layer adjacent the porous substrate comprising a two dimensional nanosheet material and crosslinked polymer. The two dimensional nanosheet material is preferably chosen from the group comprising graphene oxide including reduced graphene oxide, holey graphene, holey graphene oxide, laminated graphene oxide and holey reduced graphene oxide.

Claims

1. A membrane for reverse osmosis comprising: (i) a porous substrate, and (ii) a layer adjacent the porous substrate and comprising a two dimensional nanosheet material and a crosslinked polymer, wherein the crosslinked polymer is formed from N-isopropylacrylamide and N,N-methylenebisacrylamide.

2. The membrane according to claim 1 wherein the two dimensional nanosheet material is graphene.

3. The membrane according to claim 1 wherein the two dimensional nanosheet material is chosen from the group comprising graphene oxide including reduced graphene oxide, holey graphene, holey graphene oxide, laminated graphene oxide and holey reduced graphene oxide.

4. The membrane according to claim 1 wherein the two dimensional nanosheet material is pristine graphene oxide laminate.

5. The membrane according to claim 1 wherein the two dimensional nanosheet material is chosen from the group comprising MoS2, boron nitride, metal oxides and metal carbides.

6. The membrane according to claim 1 wherein the porous substrate is based on poly(ether sulphone) or poly sulphone.

7. The method of manufacturing the membrane of claim 1 including the steps of: (i) generating the porous substrate, and subsequently (ii) applying a solution of two dimensional nanosheet material and two monomers, wherein the two monomers are N-isopropylacrylamide and N,N-methylenebisacrylamide, to the porous substrate, (iii) crosslinking the monomers to form a polymer.

8. The method according to claim 7 wherein the solution of two dimensional nanosheet material and the two monomers are applied to the porous substrate by spin coating, filtration or reduced pressure casting of the solution.

9. The method according to claim 7 wherein the two dimensional nanosheet material is graphene.

10. The method according to claim 7 wherein the two dimensional nanosheet material is chosen from the group comprising reduced graphene oxide, holey graphene, holey graphene oxide, laminated graphene oxide and holey reduced graphene oxide.

11. The method according to claim 7 wherein the two dimensional nanosheet material is pristine graphene oxide laminate.

12. The membrane according to claim 7 wherein the two dimensional nanosheet material is chosen from the group comprising MoS2, boron nitride, metal oxides and metal carbides.

13. The membrane according to claim 7 wherein the porous substrate is based on poly(ether sulphone).

14. The membrane according to claim 1 when used for reverse osmosis filtration of saline water.

15. The method of filtration comprising the step of passing a fluid through the membrane of claim 1.

16. The filtration apparatus including the membrane of claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Further disclosure, objects, advantages and aspects of preferred and other embodiments of the present application may be better understood by those skilled in the relevant art by reference to the following description of embodiments taken in conjunction with the accompanying drawings, which are given by way of illustration only, and thus are not limitative of the disclosure herein, and in which:

(2) FIG. 1 illustrates a method of manufacturing the reverse osmosis desalination membranes of GO and polymer according to the present invention including obtaining porous substrate (1) dipping one side (2) and spin coating (3) or alternatively forming the precursor solution (4) applying it to the porous substrate by application of vacuum (5) to achieve filtration coating (5) then polymerising (9). During this process the solution of GO and monomers (7) becomes a coating of GO and monomers (8) and the polymerisation creates the GO polymer network (10);

(3) FIG. 2 is a schematic diagram of the polymerisation of the GO-polymer coating (8) on the porous substrate (14), the coating comprising monomers (11) filling up the GO scaffold (13) in the remnants of solution (12). Following polymerisation the GO scaffold (13) is entwined in a cross-linked polymer network (10);

(4) FIG. 3 illustrates scanning electron microscope (SEM) images of the surface (FIG. 3a) and a cross section (FIG. 3b) of a membrane according to the present invention fabricated onto the commercial ultrafiltration membrane substrate. The membrane layer thickness can be seen from the cross-sectional SEM image to be about 30 nm;

(5) FIG. 4 illustrates the structures of preferred monomers and polymers for preparation of the polymer network, namely N, N-methylene-bis-acrylamide (FIG. 4a), N-isopropylacrylamide (FIG. 4b) and poly(N-isopropylacrylamide-co-N,N-methylene-bis-acrylamide) (FIG. 4c) and their reaction (FIG. 4d);

(6) FIG. 5 is a thermogravimetric analysis plot of weight loss (wt %) against temperature ( C.) under inert conditions for GO (20), GO-polymer (1 ml) (21), GO-polymer (2 ml) (22) and monomers (23). The plots for GO-polymer membranes prepared by vacuum filtration at 0.1 MPa and 0.04 MPa are very similar;

(7) FIG. 6 illustrates X-ray diffraction (XRD) patterns of the GO-polymer membranes prepared by vacuum filtration at 0.04 MPa (GOP-L) (FIGS. 6a & 6b) and 0.1 MPa (GOP-H) (FIGS. 6c & 6d). FIGS. 6a and 6c illustrate loadings of 2.0 ml (27), 1.0 ml (28) and the PES substrate (29). The dried GO-polymer membranes prepared with different amounts of GO-monomer solution (25) and wet (26) GO-polymer membrane prepared with 2.0 ml loading of GO-monomer solution;

(8) FIG. 7 illustrates pure water flux (FIG. 7a) at low vacuum (.square-solid.) and high vacuum (.circle-solid.) and water flux and salt rejection (FIG. 7b) of GO-polymer membranes (GOP-L .square-solid. and GOP-H .circle-solid.) measured using 2,000 ppm NaCl solution as a function of solution loading;

(9) FIG. 8 illustrates a plot of water flux of GOP-H1.0 (.circle-solid.) and the commercial PA (.square-solid.) membrane as a function of applied pressure;

(10) FIG. 9 illustrates a plot of water flux (.circle-solid.) and salt rejection (.square-solid.) of GOP-H1.0 as a function of feed solution concentration;

(11) FIG. 10 is a plot illustrating the effect of chlorine exposure on water flux of GO-polymer membranes compared with commercial PA membranes (stripes);

(12) FIG. 11 is a plot illustrating the effect of chlorine on NaCl rejection properties of GO-polymer membranes and commercial PA membranes (stripes);

(13) FIG. 12 is a plot illustrating the effect of chlorine cleaning cycles on water flux (.circle-solid. and NaCl (.square-solid.) rejection of GOP-H1.0.

DETAILED DESCRIPTION

(14) The present invention provides filtration membranes, such as reverse osmosis filtration membranes. In particular, the membranes are graphene-based, combined with a polymer as a thin film layer on a porous substrate.

(15) Typically, the layer comprises GO nanosheets as a scaffold which is combined with a polymer and deposited as a film of tens of nanometers in thickness on a rough substrate surface as depict inn FIG. 1 and FIG. 2. The integration of GO and polymer provides the advantage of high water flux and rejection afforded by GO and good processability afforded by polymers, to provide a platform material suitable for use in a desalination membrane.

(16) The fabrication of these desalination membranes consists of two main steps: (i) preparation of a porous substrate, and (ii) the application of a thin layer of a GO and polymer on the porous substrate.

(17) Typically, the porous substrate can be any readily available commercially available or lab synthesised porous membrane that is appropriate for the type of filtration anticipated. For example, high pressure reverse osmosis filtration requires a porous substrate with low to medium size pores, such as ultrafiltration membranes, to ensure mechanical strength.

(18) GO-polymer membranes according to the present invention are typically formed were formed on a porous substrate by spin-coating or filtration coating from a precursor solution, followed by free radical polymerization of monomers as shown in FIG. 1. A water-based precursor solution is formed from GO nanosheets, monomers and a polymerisation initiator such as ammonium persulfate. The monomers can be, for example, mono-functional or bifunctional terminal alkene groups with CC double bonds such as N-isopropylacrylamide, 3-Buten-2-one, N,N-methylenebis(acrylamide), N,N-ethylenebis(acrylamide), or acrylaldehyde; 2-hydroxyethyl methacrylate and propylene glycol monoacrylate, and triethylene glycol dimethacrylate. Typically the mono-functional monomer forms the polymer backbone and bifunctional monomer acts as a crosslinker during polymerization.

(19) Typically, the GO/polymer ratio in the precursor solution is between 0.1 and 70 wt %.

(20) Typically, the total solid (monomers and GO nanosheets) concentration is between 0.5 to 5 wt % for spin-coating or between 0.01 to 50 wt % for filtration coating.

(21) The precursor solution may be applied to the substrate by dipping and spin-coating or reduced pressure filtration coating as described in FIG. 1 and FIG. 2. For spin-coating, a substrate is typically floated on the precursor solution for a predetermined time, then spin-coated for 10 to 100 seconds at a speed in 500 to 5000 rpm. The spin-coating time and speed can be varied to control the layer thickness.

(22) Subsequently, the substrate with the precursor solution coating is dried, such as in an oven at 50 to 100 C. to promote free-radical polymerization and to form the GO-polymer layer on a substrate. Spin-coating and polymerization can be repeated to ensure a defect-free GO-polymer layer on the substrate.

(23) For filtration coating, a substrate may be located in a filtration set, or the precursor solution poured into a filtration set on a substrate, and a vacuum applied. Subsequently, the substrate with the precursor solution coating was located in an oven at 50 to 100 C. for free-radical polymerization to form a GO-polymer layer on a substrate. Thickness of a GO-polymer layer can be controlled by an amount of solution loaded in a filtration set.

(24) The GO-polymer layer is typically characterised as having an asymmetric microstructure with a dense surface without substantial defects in which highly crosslinked polymers entwine graphene-oxide nanosheets.

EXAMPLES

(25) The invention will be further described with reference to the following non-limiting example in which monofunctional N-isopropylacrylamide (NIPAM) and difunctional N,N-methylenebis(acrylamide) (MBA) are used as monomers, and ammonium persulfate is used an initiator. Highly crosslinked poly(NIPAM-MBA) network is formed on a porous polymer substrate by free radical polymerization of NIPAM and MBA, which is initiated by thermal composition of ammonia persulphate. A defect-free, active layer with a very low thickness of less than 40 nm is successfully fabricated by spin-coating and in situ polymerization. The polymer network in the GO-polymer composite enables excellent membrane mechanical stability and processability while maintaining the physical properties of GO nanosheets that facilitate water transport. GO-polymer thin-film composite membrane endures high chlorine exposure at 1,000 ppm for 24 h without noticeable change, demonstrating outstanding chlorine tolerance.

Example 1: Fabrication of a GO-Poly(NIPAM-MBA) Membrane on a Commercial Ultrafiltration Membrane Support

(26) Fabrication of the GO-Polymer Membrane:

(27) GO was synthesized from graphite powder by modified Hummers method. The concentration of the resulting GO suspension was 3.8 mg/ml in water. The GO concentration was determined by freeze-drying of 10 ml GO suspension, and weighing the residual solid (GO). Monomers to fabricate a polymer network within GO laminate scaffold including N-isopropylacrylamide (NIPAM), N,N-methylenebisacrylamide (MBA), and ammonium persulfate were purchased from Sigma-Aldrich (St Louis, Mo.) and used without further purification. Specifically, 38 mg of NIPAM, 38 mg of MBA, and 7 mg of ammonia persulphate were dissolved into 10 ml GO suspension. To vary the ratio of GO/poly(NIPAM-MBA) from 4.5 to 50% (as per Table 1), the same amounts of NIPAM, MBA and ammonia persulphate were added into 10 ml GO suspension with a different GO loading, 3.4, 4.8, 9.5, 19, 28.5, and 38 mg. The GO suspension with a different GO loading was obtained by diluting 3.8 mg/ml GO suspension.

(28) TABLE-US-00001 TABLE 1 Components used in the preparation: Monomers (mg) GO/ ammonia GO polymer persulph- Water GO conc. ratio Samples NIPAM MBA ate (ml) (mg) (mg/ml) (wt %) GO-P61 38 38 7 10 3.4 0.34 4.5 GO-P65 4.8 0.48 6.3 GO-P72 9.5 0.95 12.5 GO-P79 19 1.9 25.0 GO-P87 28.5 2.85 37.5 GO-P95 38 3.8 50.0

(29) The precursor solution including GO and monomers was sonicated for 30 minutes to completely dissolve. Nylon membrane substrate, obtained from Sterlitech Corporation (product number NY4547100, Kent, Wash., USA), was used as a support layer with a diameter of 47 mm and average pore size of 450 nm. Nylon substrate was floated on 10 ml of precursor solution in a glass petri dish for 30 seconds, set on a spin-coater (WS-650-23B, Laurell Technologies Corporation, PA, USA), and spun for 30 seconds in 1,000 rpm, then located in a convection oven (Thermal Fisher) at 70 C. for 2 h for complete free-radical polymerization of monomers. Coating and polymerization procedures were repeated for 3 times to ensure the formation of a defect-free GO-polymer composite layer on the Nylon substrate.

(30) Fabrication of the GO polymer layer on the ultrafiltration membrane substrate was achieved by floating the ultrafiltration membrane substrate on 10 ml of precursor solution in a glass petri dish for 30 seconds, set on a spin-coater (WS-650-23B, Laurell Technologies Corporation, PA, USA), and spun for 30 seconds at 1,000 rpm. It was subsequently heated in a convection oven (Thermal Fisher) at 70 C. for 2 hrs to achieve free-radical polymerization of monomers. The coating and polymerization procedures were repeated for 3 times to ensure the formation of a defect-free GO-polymer composite layer on the ultrafiltration membrane substrate.

(31) The generalised reaction of the N,N-methylenebisacrylamide and N-isopropylacrylamide monomers to form the cross lined polymer network of poly(N-isopropylacrylamide-co-N,N-methylenebisacrylamide) can be depicted as shown at FIG. 4d:

(32) Characterization.

(33) The morphology of GO-polymer membranes was observed using a field emission scanning electron microscope (FE-SEM) (Nova NanoSEM 450, FEI, Hillsboro, Oreg.). A uniform membrane according to the present invention and without visible defects was fabricated onto the commercial ultrafiltration membrane substrate with rough surface as confirmed by SEM images (FIG. 3). The membrane layer thickness can be seen from the cross-sectional SEM image to be about 30 nm. Without wishing to be bound by theory it is believed that GO nanosheets serve as a scaffold and play a key role in forming such a thin film on the rough substrate surface. During the polymerization process, the monomer solution confined within the GO laminate layer, where the polymerization occurred, leading to defect-free nanocomposite membranes from a highly crosslinked poly(NIPAM-MBA) network that physically entwined GO nanosheets.

(34) The resulting reverse osmosis membrane performance is shown in Table 2.

(35) TABLE-US-00002 TABLE 2 Water flux and salt rejection properties of GO-polymer membranes as a comparison of commercial membranes. Salt Rejection Membrane Water Flux (Lm.sup.2h.sup.1) (%) GO-polymer 43-50 (at 150 psi) 99.2 Commercial CA** 17-30 (at 425 psi) 98.5 (Cellulose triacetate/diacetate Blend from GE) Commercial PA** 15-25 (at 200 psi) 98.6 (DURASLICK from GE) *Feed solution: 2,000 mg/L NaCl solution **G. Infrastructure, Desal(R) PURE WATER Membrane Technology & Applications, in: W.P.T. GE Infrastructure (Ed.) GE Infrastructure, Trevose, PA, 2004.
Results and Discussion

(36) Polymer network entwined GO membranes were fabricated to demonstrate their high performance in reverse osmosis water desalination process. An ultrafiltration membrane was chosen as a substrate. The random copolymer, poly(N-isopropylacrylamide-co-N,N-methylene-bis-acrylamide) (or poly(NIPAM-MBA)), was produced by free radical polymerization of NIPAM and MBA monomers via opening of CC double bond. Poly(NIPAM-MBA) is densely crosslinked and hydrophilic, and firmly becomes entwined around GO laminate scaffold on the ultrafiltration substrate, allowing water permeation and rejecting hydrated ions. Importantly this poly (NIPAM-MBA) polymer network is highly stable in chlorinated water and mechanically robust for membrane fabrication and operation.

(37) The GO-poly(NIPAM-MBA) thin selective layer was prepared by spin coating and subsequent polymerization and drying. The mass ratio of GO and monomers in the precursor solution was varied from 4.5 to 50.0 wt % GO to investigate the effects of membrane composition on the water flux and rejection of the membrane. During the spin-coating, different proportions of the monomer solution and GO nanosheets were lost under centrifugal force due to molecular size and structure. Micro-sized GO nanosheets assembled into a laminate scaffold on the Nylon substrate, where a significant amount of monomers was lost with water during the spin coating and the remained monomer solution was confined within GO interlayer spaces. Consequently, the ratio of GO/poly(NIPAM-MBA) changed after the spin-coating. Without GO nanosheets, all of the monomer solution was lost whilst being spin-coated, and no coating was formed. To determine the actual GO content in the GO-polymer layer, thermogravimetric analysis was conducted and the GO weight percentage was calculated on the basis of relative weight loss of polymer and GO. The actual GO amount ranged from 60.7 to 95.1 wt %.

Example 2: Fabrication of GO Laminate-Poly(NIPAM-MBA)

(38) The following illustrates how the process of free-radical polymerisation in aqueous conditions of a GO suspension enables the formation of a GO-polymer network without covalent bonding between GO and polymer, leading to a combination of mechanical robustness, remarkable water transport properties, and an anti-swelling tendency suitable for the high-pressure operation required for reverse osmosis (RO) membrane process.

(39) The GO-polymer membrane prepared according to the present invention was not swollen during RO operation and remained stable under a harsh oxidising environment. It also demonstrated water flux of 33.5 L m.sup.2 h.sup.1 with 98.5% NaCl rejection at 10 bar when prepared by high-vacuum filtration and subsequent polymerisation of the precursor solution on to a mesoporous substrate. This methodology indicates that large-scalable membrane fabrication is feasible.

(40) Fabrication of the GO Laminate-Polymer Membrane:

(41) Free radical polymerisation has been applied for formation of polymer network such as copolymers, micelles, or hydrogels, due to its rapid reaction under mild conditions. In this experiment, a hydrophilic and highly cross-linked polymer network was synthesised with GO laminates on a porous polymer substrate by vacuum assisted filtration and subsequent polymerisation.

(42) Materials:

(43) GO was synthesised using a modified Hummer's method, [Dikin et al, Nature 2007, 448, 457; Hummers & Offeman, J. Am. Chem. soc 1958, 80, 1339) from graphite powders obtained from Sigma-Aldrich (St Louis, Mo.). For preparation of polymer network, N-isopropylacrylamide (NIPAM), N,N-methylenebisacrylamide (MBA), and ammonium persulfate (APS) were purchased from Sigma-Aldrich and used without further purification. Sodium chloride (NaCl) and sodium hypochlorite for membrane characterisation were purchased from Merck (Kenilworth, N.J.) and Sigma-Aldrich, respectively. Porous poly(ether sulfones) (PES) membranes were used as a substrate. Commercial polyamide (PA) membranes (DOW FILMTEC BW30) were obtained from Stelitech (Kent, Wash.).

(44) Fabrication of the GO-polymer membrane: 45 mg of NIPAM, 45 mg of MBA and 25 mg of APS were dissolved in 10 ml of GO solution (1.0 mg/ml). The precursor solution was filtered through the PES substrate using a glass filtration set with a circulating water vacuum pump (maximum vacuum capacity: 0.098 MPa) at vacuum levels of 0.04 MPa or 0.1 MPa (0.098 MPa), then the substrate was placed in a convection oven (Thermal fisher, Scoresby, Australia) at 60 C. for 1 h for complete polymerisation. The resulting membrane was used for characterisation without further treatment.

(45) Characterisation:

(46) The weight ratio between GO and polymer in the GO-polymer composite was determined by thermogravimetric analysis (TGA, PerkinElmer, MA). The precursor solution containing GO and monomers without initiator (APS) was prepared and filtered on the PES substrate by vacuum at 0.04 and 0.1 MPa with the volume of 1.0 and 2.0 ml, then carefully peeled off the substrate and dried for TGA. The morphology of the GO-polymer membranes was investigated with a field emission scanning electron microscope (FE-SEM, Nova NanoSEM 450, FEI, Hillsboro, Oreg.). The crystalline structure of the membrane samples was determined by using an X-ray diffraction technique (XRD, Miniflex 600 X-ray diffractometer, Rigaku, Japan). The chlorine tolerance of the GO-polymer membranes and the PA commercial membranes were characterised using aqueous sodium hypochlorite (NaClO) solution at concentration of 100, 500, and 1,000 ppm. A coupon of membrane was immersed in the NaClO solution and kept for 24 h at room temperature, then the membrane was taken out and completely washed for following water flux and NaCl rejection measurements.

(47) Water Flux and Salt Rejection:

(48) The water flux and salt rejection properties of the GO-polymer membranes and the PA commercial membranes were measured from deionised (DI) water and 2,000 ppm NaCl aqueous solution using a high-pressure stirred cell (HP4750, Sterlitech, Kent, Wash.). The membrane with an effective area of 14.6 cm.sup.2 was loaded into the RO testing cell and feed solution was pressured at between 5 to 35 bar. To study the effect of NaCl concentration, water flux and salt rejection of the GO-polymer membranes were measured with the NaCl solutions ranged from 2,000 to 15,000 ppm at 20 bar. The water flux of the membrane was calculated from the weight changes and the electrical conductivity of the feed and permeate solutions was measured to determine the salt concentration which was used to calculated the salt rejection.

(49) Discussion:

(50) Vacuum filtration is a widely adopted method for preparing uniform GO laminates, but a large volume of GO solution is often employed, which significantly extends filtration time. Unlike vacuum filtration methods of the prior art, the present invention uses filtration of a small volume of the precursor solution containing GO and monomers and its subsequent polymerisation formed a defect-free GO-polymer membrane with a much shorter filtration time, which is highly feasible for scaling up membrane fabrication for practical application. Moreover, use of the vacuum assisted method to form the polymer composite membrane improves adhesion between the membrane layer and a substrate since the polymer formed penetrates into the nanopores of the skin layer of the substrate under the influence of the vacuum.

(51) The monomers used to produce the polymer network contain a large component of bifunctional monomers and a small part of monofunctional monomers to maximise mechanical strength by balancing flexibleness and modulus, with an aim that the highly cross-linked and dense polymer network-entwined GO layer is mechanically strong and chemically durable. As schematically illustrated in FIG. 2, the cross-linked polymer network tightly entwines the graphene oxide laminates to form a GO-polymer layer, where polymer and GO laminates are physically interconnected.

(52) The degree of vacuum for filtration influences the membrane morphology as it controls the filtration rates of monomers and GO laminates, ultimately determining the interlayer spacing of GO-polymer membranes. After the GO laminates form a dense GO layer on the substrate at the initial stage of filtration, monomer accumulates between GO laminates, forming the asymmetric structure of the GO-polymer layer, with a proportion of denser GO at the bottom of the membrane adjacent the substrate, and a proportion of less dense and looser GO at the top of the GO-polymer layer. This GO-polymer composite membrane also retains the intrinsically desirable properties of pristine GO laminates, that is, fast and selective water transport since there is no chemical reaction between GO and the polymer. It also demonstrates a significant anti-swelling property due to the highly cross-linked structure. The rapid reaction of radical polymerisation also allows for a very rapid and well-defined membrane fabrication.

(53) As an initial step, the water-based precursor solution containing GO laminates and monomers (N-isopropylacrylamide (NIPAM), N,N-methylenebis(acrylamide) (MBA) and ammonium persulfate (APS)) were filtered under vacuum on poly(ether sulfone) (PES) substrates with nanometre-sized pores. The membrane was then placed in a convection oven to initiate free-radical polymerisation and prepare GO-poly(NIPAM-MBA).

(54) During solution filtration, water soluble monomers were able to pass through the substrate, whereas water-dispersed GO laminates were remained on the surface of the substrate. After forming the dense GO layer on the surface, monomers were trapped between GO and formed cross-linked polymer network with GO laminates. Due to the loss of monomers during filtration, the ratio between GO and polymer changed from the initial ratio in the precursor solution. The precursor solution between 0.5 and 3.0 ml was filtered at the low (0.04 MPa) and high (0.1 MPa) vacuum levels on the substrate with effective area of 17.3 cm.sup.2. The weight ratio between GO and monomers in the precursor solution was 1:10, but after filtration and polymerisation, the GO percentages changed to a range of 50 to 70 wt % depending on fabrication conditions, as can be determined from thermogravimetric analysis (TGA) under the inert condition and described in FIG. 5.

(55) Thermal stability of GO laminates and monomers containing NIPAM and MBA were measured as a reference. Thermal decomposition of hydroxyl groups on GO was evidenced with the weight loss around 200 C. and 43% of residual graphene was remained after measurement up to 600 C.

(56) For monomers, the first weight loss between 100 and 250 C. was attributed to the loss of aliphatic alkene groups in MBA and of most NIPAM with a low boiling point. 8% of the carbonized compound was remained at 600 C. Comparing with the residual weight at 600 C., the weight ratio between GO and polymer was confirmed and summarised in Table 3. When high vacuum pressure was applied during filtration, the percentage of GO was higher as compared to low vacuum pressure, but the difference was not significant. In the cases of the high and low levels of applied vacuum, the percentage of GO was lower when more solution was loaded because more monomers were accumulated between GO laminates.

(57) TABLE-US-00003 TABLE 3 GO contents in GO-polymers after filtration: GO/poly(NIPAM-MBA) Applied Solution weight ratio [%] vacuum loading In solution In membrane Samples pressure [MPa] [ml] (GO/monomers) (GO/polymer) GOP-L1.0 0.04 1.0 10 63 GOP-L2.0 2.0 54 GOP-H1.0 0.01 1.0 10 65 GOP-H2.0 2.0 58

(58) The morphology of the effective layer of the GO-polymer membrane was also influenced by vacuum levels applied for filtration and volume of solution loaded. Scanning electron microscopy showed that thin, aligned and uniform GO-polymer layers resulted when the high-level vacuum of 0.1 MPa was applied.

(59) The GO-polymer membrane formed has a layer thickness of few-hundreds nanometres, depending on preparation conditions, such as a volume of solution loading or filtration vacuum levels. The thickness of the effective layer gradually increased as a function of volume of solution loading ranging from 0.1 to 1.3 m from precursor solution loading between 0.5 to 3.0 ml. The effective layer was somewhat irregular due to the rough surface of the substrate. When a solution less than 1.0 ml was loaded, the layer thickness was around 150 nm, meaning that GO and monomers were instantly filtered under 0.1 MPa and formed a thin layer. However, the layer from loading of 3.0 ml was much thicker than expected (1.1-1.3 m), indicating that early on in the process, the GO laminates blocked pores of a substrate and prevented the application of a greater vacuum, leading to a compact GO-polymer layer at the later stage of filtration.

(60) When the low level vacuum was used, the solution loading volume dependency was not significant. Compared to membranes prepared at 0.04 MPa and different solution loading of 0.5 and 2.0 ml, the thickness of layers only increased from 500 to 600 nm, meaning that the volume of solution loading effects compactness of the effective layer, but not on the layer thickness itself. Moreover, when comparing with the membranes prepared by 1.0 ml solution loading at high and low vacuums, the membrane prepared at low vacuums were some three times thicker, although their GO percentages were similar to each other.

(61) High-resolution SEM images are often used to confirm graphene or GO laminates aligned in graphene-based membranes. Graphene laminates in the polymer-graphene composite were also observed in SEM images when polydiacetylene and graphene were stacked within a composite film. The parallel, aligned GO laminates in the polymer network were clearly observed close to the substrate which led to the mechanically robust membrane effective layer and the formation of a defect-free membrane. However, GO laminates in SEM images are not as clear as GO in pure GO membranes because the polymer network is entwined and covers GO laminates. In the case of GOP-L1.0, the aligned GO layer was not clearly recognised, meaning that GO laminates were rather randomly stacked on a substrate and formed a less dense and thick GO-polymer network which retained the ability to become further compacted should more solution be introduced.

(62) FIG. 6 shows X-ray diffraction (XRD) spectra of the GO-polymer membranes and the PES substrate. The PES substrates have a typical amorphous structure with scattered XRD bands. After forming the GO-polymer layer on the substrate, a broad or sharp band around 8 was observed, indicating an interlayer spacing of GO in the polymer composites of around 0.9 to 1.0 nm.

(63) It was found that preparation conditions affected the interlayer spacing of the membranes. When the low vacuum of 0.04 MPa was applied for the membrane preparation, the XRD band was broad. The intensity was relatively weak when 1.0 ml solution was loaded due to the thin layer on the substrate, but after loading 2.0 ml, the band became more resolved as the full width at half maximum (FWHM) of 2.58 and 1.47 degree were recorded for the GO peak after loading 1.0 and 2.0 ml. However, when a higher vacuum of 0.1 MPa was applied, a sharp and strong peak at 8.5 was observed from the more tightly packed GO-polymer structures.

(64) Clearly the high vacuum helps to firmly align GO laminates during filtration and the resulting GO-polymer network is highly ordered and compact. The FWHM is estimated to be 0.115 and 0.068 degree when 1.0 and 2.0 ml solutions were loaded. It is known that the separation capability of GO membranes is significantly limited in an aqueous phase due to the high swelling as the interlayer spacing between GO laminates are enlarged and salt permeability is increased, which can be a major drawback in using GO membranes for water processing. However, the swelling can be confined by the combination of GO with a cross-linked polymer network since the amount of water entrapped inside the interlayer space between GO nanosheets is limited to crosslinking of GO.

(65) The X-ray diffraction patterns of the GO-polymer composite membranes were observed in dry (FIGS. 6a & 6d) and wet (FIGS. 6b & 6c) phases to analyse the interlayer spacing and investigate their swelling behaviours. After hydration, the broad peak shift of the GO-polymer composite was almost negligible, demonstrating that the cross-linked polymer tightly fixed GO laminates within its network, preventing swelling when hydrated.

(66) The hydrophilic and ultrathin GO-polymer membranes were capable of achieving fast and selective water transport with high salt rejection. Water flux and salt rejection properties of GO-polymer membranes prepared under low and high vacuum have been investigated at RO processes from deionised (DI) water and 2,000 ppm sodium chloride (NaCl) solution using a dead-end filtration set-up at 10 bar. As illustrated in FIG. 7, both DI water and salt water flux were linearly decreased as a function of a volume of solution loading at preparation and salt rejection was slightly increased.

(67) Water transport is an intrinsic property of membrane materials, but water flux can be determined by membrane thickness as well as compactness of effective layers. In case of the membrane prepared at 0.1 MPa, water flux was gradually decreased from 39.3 to 11.3 L m.sup.2 h.sup.1 when solution loading was increased from 0.5 to 3.0 ml, which also corresponds to the increase in effective layer thickness observed in SEM images.

(68) When 2,000 ppm NaCl solution was used as a feed, salt water flux was slightly lower than pure water flux ranged between 32.9 and 8.3 L m.sup.2 h.sup.1 resulting from the lower driving force due to osmotic pressure of salt solution. The membranes exhibited high salt rejection properties which increased from 98.1 to 98.9% as membrane layer became thicker. Note that thin effective layer could benefit high water flux, but it can also lead to defects on membranes and reduced salt rejection. Moreover, GO laminates in the effective layer of the membrane easily form nanochannels during the membrane preparation for precise water transport, which greatly improve separation property of the GO-polymer membrane. However, when the more than 1.0 ml solution was loaded, the membranes demonstrated both high water flux over 30 L m-2 h-1 and high salt rejection over 98.5%. In case of the membrane prepared at 0.04 MPa, water flux was ranged from 37.1 to 8.3 L m.sup.2 h.sup.1 depending on solution loading which was a bit lower than the membrane prepared under high level vacuum. In addition, a similar trend due to the volume of solution loaded on salt water flux and salt rejection properties was observed, as salt water flux was ranged between 29.7 and 6.1 L with salt rejection ranging between 97.5 and 98.7%.

(69) Thickness of the effective layers is one of the major property that determines a molecular transport property of composite membranes. However, the density, or compactness of the effective layer also influences a water flux of RO membranes because the less-compact layer allows molecules to permeate through the layer resulting in high flux and low rejection properties. Although the effective layer thickness of the GO-polymer membranes was around 200 nm regardless of solution loading for membranes prepared at 0.04 MPa, the water flux and the salt rejection were very different. In this case, salt rejection improved when more solution was loaded during preparation, which means the membrane effective layer became more compact.

(70) Table 4 summarises water flux and NaCl rejection properties of PA-based thin film composite (TFC) membranes with high water flux as a comparison of the GO-polymer membranes. Commercial RO membranes in general exhibit over 98% NaCl rejection, but there is a trade-off relationship between water flux and salt rejection as RO membranes with higher rejection demonstrate relatively lower water flux.

(71) Many approaches have been conducted to enhance RO performances by incorporating nanomaterials into PA. GO incorporated PA membranes with a small amount of GO loading showed 50% increase in water flux from the pristine PA membrane with relatively low NaCl rejection of 93.8%, but a large amount of GO loading significantly reduced water flux while increasing NaCl rejection to 99.5%. [14e] Introduction of nanomaterials, such as carbon nanotubes (CNTs) and zeolitic imidazolate frameworks (ZIFs) nanoparticles effectively improved water flux and/or NaCl rejection by providing microporous structure from nanoparticles in PA matrix. Although their high water flux and salt rejection can improve productivity and efficiency of RO processes, the drawbacks of poor chemical stability still remain in these PA-based membranes because of degradation of amide bonds in oxidising agents. The GO-polymer membrane in this study exhibited high water flux and salt rejection property as compared to PA-based RO membranes; this study provided a good strategy for developing RO membrane to overcome the stability issue of commercial membranes since the GO-polymer exhibits excellent chemical stability to resist oxidising agents.

(72) TABLE-US-00004 TABLE 4 Comparison of RO performances of this work with the literature data related to high flux RO membranes. Trans- membrane NaCl Water Membrane pressure concentration permeance NaCl materials [bar] [ppm] [L m.sup.2 h.sup.1/ bar] rejection [%] Ref. GOP_H1.0 10 2,000 3.01 98.5 This invention PA 20 2,000 2.77 98.1 g PA-GO 20.7 2,000 2.87 93.8 a PA on 15.5 2,000 3.03 97.7 b CNT/PSf PA-ZIF8 15.5 2,000 3.35 98.5 c PA-zeolite 15.5 2,000 4.79 98.8 d PA-GO 15.5 2,000 1.07 99.5 e PA-POSS 55 32,000 0.892 99.6 f

(73) Refs: a) J. Yin, G. Zhu, B. Deng, Desalination 2016, 379, 93; b) T. H. Lee, M. Y. Lee, H. D. Lee, J. S. Roh, H. W. Kim, H. B. Park, J. Membr. Sci. 2017, 539, 441; c) J. Duan, Y. Pan, F. Pacheco, E. Litwiller, Z. Lai, I. Pinnau, J. Membr. Sci. 2015, 476, 303; d) H. Dong, L. Zhao, L. Zhang, H. Chen, C. Gao, W. W. Ho, J. Membr. Sci. 2015, 476, 373; e) H.-R. Chae, J. Lee, C.-H. Lee, I.-C. Kim, P.-K. Park, J. Membr. Sci. 2015, 483, 128; f) J. H. Moon, A. R. Katha, S. Pandian, S. M. Kolake, S. Han, J. Membr. Sci. 2014, 461, 89; g) H. Wang, L. Li, X. Zhang, S. Zhang, J. Membr. Sci. 2010, 353, 78.

(74) Pressure dependence on water flux and salt rejection can be an important parameter since industrial RO processes are operated at high pressure to overcome the osmotic pressure of seawater and to generate pure water.

(75) Pure water flux of the commercial PA membrane and the GO-polymer membrane prepared under 0.1 MPa vacuum and 1.0 ml solution loading, which demonstrated high water flux and salt rejection, were measured in the pressure ranged between 5 and 35 bar using a dead-end filtration set-up. For the PA commercial membrane, the water flux showed a concave increase to the pressure axis meaning that the PA layer of the membrane suffered from the compaction effect illustrated in FIG. 8.

(76) At high pressures, the PA layer became compressed and dense, as a result water permeability decreased. This compaction effect can be disadvantageous to the RO process with the loss of productivity on the operational conditions. However, in the case of the GO-polymer membranes, water flux almost linearly increased as input pressure increased. The mechanically-robust effective layer of the GO and polymer network improved anti-compaction properties and the membrane structure was retained, even at high pressures.

(77) Considering practical RO operation, seawater contains around 35,000 ppm NaCl, having an osmotic pressure of 27.6 bar. Therefore, high feed pressure is required to overcome the high osmotic pressure in the RO process. The effects of feed solution concentration were investigated by measuring the water flux and salt rejection of the GO-polymer membranes at 20 bar at the feed NaCl concentration ranging from 2,000 to 15,000 ppm (osmotic pressure ranged from 1.6 to 11.8 bar) as shown in FIG. 9.

(78) To accurately measure the water flux by minimising the effect of the concentration polarisation, where the salt concentration on the membrane surface at the high pressure side becomes excessive and water flux is decreased due to greater osmotic pressure on the feed side. Therefore, the maximum concentration of the feed was limited to 15,000 ppm to ensure accurate measurement of salt rejection, although the GO-polymer membrane was mechanically stable under the test with 35,000 ppm NaCl solution (FIG. 8). The GOP-H1.0 membrane showed 52.3 L m-2 h-1 when 2,000 ppm NaCl solution was used. But the water flux gradually decreased when NaCl concentration increased. This was because the difference between the operating pressure and the osmotic pressure of the feed decreased. However, the GOP-H1.0 membrane demonstrated a stable NaCl rejection of around 98.3%.

(79) When considering practical RO applications, chemical durability in oxidising environments, especially chlorine tolerance, is an important property. In a current PA-based RO system, chlorine in the feed water is commonly introduced to minimise membrane biofouling, however, dechlorination and rechlorination steps are required before and after feed water passing though the RO membrane to protect the PA membranes which suffer from poor chlorine stability, which leads to additional processes and costs.

(80) Since commercial RO membranes are continuously operated for 3-5 years and require cleaning once the module is installed, the chlorine stability in the repeated exposure cycles is essential. To characterise chlorine stability of GO-polymer membranes, sodium hypochlorite (NaClO) solution was introduced in 100, 500, and 1,000 ppm concentrations. The GO-polymer membranes and the commercial PA membranes were immersed in NaClO solution for 24 h, then water flux and NaCl rejection measured at 10 bar. As shown in FIGS. 10 and 11, the active layer of the PA membrane was damaged and resulted in increase in water flux and dramatic decrease in NaCl rejection because amide bonds in PA are readily cleaved in the chlorine-rich conditions. However, the GO-polymer membrane showed a remarkable chlorine tolerance due to its amide-bond-free structure, as well as the strong sp2 structure of GO, meaning that changes in water flux and NaCl rejection were negligible.

(81) The effect of the repeated chlorine exposure on membrane performance of the GO-polymer membrane was investigated using 1,000 ppm NaClO aqueous solution. After being immersed in the NaClO solution for 24 h and washed with DI water, water flux and NaCl rejection of the membrane measured at 10 bar and then the GO-polymer membrane was immersed again for 24 h. After 6 cycles of the test, the GO-polymer membrane had the stable water flux and rejection properties as shown in FIG. 12. Note that chlorine exposure at 1,000 ppm is a more extreme situation than is normally encountered in commercial practice as only a few ppm of chlorine is used in RO processes. Our study demonstrates the potential long-term durability of the GO-polymer membranes for the practical RO operations.

(82) While this invention has been described in connection with specific embodiments thereof, it will be understood that it is capable of further modification(s). This application is intended to cover any variations uses or adaptations of the invention following in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains and as may be applied to the essential features hereinbefore set forth.

(83) As the present invention may be embodied in several forms without departing from the spirit of the essential characteristics of the invention, it should be understood that the above described embodiments are not to limit the present invention unless otherwise specified, but rather should be construed broadly within the spirit and scope of the invention as defined in the appended claims. The described embodiments are to be considered in all respects as illustrative only and not restrictive.

(84) Various modifications and equivalent arrangements are intended to be included within the spirit and scope of the invention and appended claims. Therefore, the specific embodiments are to be understood to be illustrative of the many ways in which the principles of the present invention may be practiced. In the following claims, means-plus-function clauses are intended to cover structures as performing the defined function and not only structural equivalents, but also equivalent structures.

(85) Comprises/comprising and includes/including when used in this specification is taken to specify the presence of stated features, integers, steps or components but does not preclude the presence or addition of one or more other features, integers, steps, components or groups thereof. Thus, unless the context clearly requires otherwise, throughout the description and the claims, the words comprise, comprising, includes, including and the like are to be construed in an inclusive sense as opposed to an exclusive or exhaustive sense; that is to say, in the sense of including, but not limited to.